Selected Technical Papers

STP1589

Editors: John Sebroski and Mark Mason

Developing Consensus
Standards for Measuring
Chemical Emissions from Spray
Polyurethane Foam (SPF)
Insulation

ASTM STOCK #STP1589

DOI: 10.1520/STP1589-EB

ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959
Printed in the U.S.A.
Library of Congress Cataloging-in-Publication Data
Names: Sebroski, John, editor. | Mason, Mark 1947-, editor.
Title: Developing consensus standards for measuring chemical emissions from
spray polyurethane foam (SPF) insulation / editors: John Sebroski, Mark
Mason.
Description: West Conshohocken, PA : ASTM International, [2016] | Series:
Selected technical papers ; STP 1589 | “ASTM Stock #STP1589.” | Includes
bibliographical references.
Identifiers: LCCN 2016046094 (print) | LCCN 2016046208 (ebook) | ISBN
9780803176232 | ISBN 9780803176249 (ebook)
Subjects: LCSH: Insulating materials–Standards–United States. | Insulation
(Heat)--Standards--United States. | Polyurethanes–Environmental aspects.
| Aerosols--Environmental aspects.
Classification: LCC TH1715 .D4115 2016 (print) | LCC TH1715 (ebook) | DDC
691/.95--dc23
LC record available at https://lccn.loc.gov/2016046094

Copyright © 2017 ASTM INTERNATIONAL, West Conshohocken, PA. All rights reserved. This material may
not be reproduced or copied, in whole or in part, in any printed, mechanical, electronic, film, or other
distribution and storage media, without the written consent of the publisher.

Photocopy Rights
Authorization to photocopy items for internal, personal, or educational classroom use, or the internal,
personal, or educational classroom use of specific clients, is granted by ASTM International provided that
the appropriate fee is paid to the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923,
Tel: (978) 646-2600; http://www.copyright.com/

The Society is not responsible, as a body, for the statements and opinions expressed in this publication.
ASTM International does not endorse any products represented in this publication.

Peer Review Policy

Each paper published in this volume was evaluated by two peer reviewers and at least one editor. The
authors addressed all of the reviewers’ comments to the satisfaction of both the technical editor(s) and
the ASTM International Committee on Publications.

The quality of the papers in this publication reflects not only the obvious efforts of the authors and the
technical editor(s), but also the work of the peer reviewers. In keeping with long-standing publication
practices, ASTM International maintains the anonymity of the peer reviewers. The ASTM International
Committee on Publications acknowledges with appreciation their dedication and contribution of time
and effort on behalf of ASTM International.

Citation of Papers
When citing papers from this publication, the appropriate citation includes the paper authors, “paper
title,” STP title, STP number, book editor(s), ASTM International, West Conshohocken, PA, year, page
range, paper doi, listed in the footnote of the paper. A citation is provided on page one of each paper.

Printed in Bay Shore, NY

February, 2017
Foreword
THIS COMPILATION OF Selected Technical Papers, STP1589, Developing Con-
sensus Standards for Measuring Chemical Emissions from Spray Polyurethane Foam
(SPF) Insulation, contains peer-reviewed papers that were presented at a sympo-
sium held April 30–May 1, 2015, in Anaheim, California, USA. The workshop was
sponsored by ASTM International Committee D22 on Air Quality and Subcommit-
tee D22.05 on Indoor Air.

Symposium Chairpersons and STP Editors:

John Sebroski
Covestro LLC
Pittsburgh, PA, USA

Mark Mason
US Environmental Protection Agency
Research Triangle Park, NC, USA
Contents

Overview vii

Evaluation of Micro-Scale Chambers for Measuring Chemical Emissions

from Spray Polyurethane Foam Insulation 1
John Sebroski, Jason W. Miller, Carl P. Thompson, and Elizabeth Roeske

Measurement of Chemical Emissions from Spray Polyurethane Foam Insulation

Using an Automated Micro-Scale Chamber System 27
Yunyun Nie, Eike Kleine-Benne, and Kurt Thaxton

VOC Analysis of Commercially Available Spray Foam Products 43

J. Paul Duffy and Richard Wood

Flame Retardant Emissions from Spray Polyurethane Foam Insulation 57

Dustin Poppendieck, Matthew Schlegel, Angelica Connor, and Adam Blickley

Glass Chamber Method for Screening of 4,4'- MDI and TCPP Emissions from
Foam Joint Sealant 77
Doyun Won, Angelika Zidek, Gang Nong, and Ewa Lusztyk

Prioritizing Chemical Emissions from Closed-Cell Spray Polyurethane Foam:

Utilizing Micro-Scale Chamber Emission Factors and Field Measurement Data 98
Scott Ecoff, Shen Tian, and John Sebroski

Computer Simulation of Peak Temperatures in Spray Polyurethane Foam

Used in Residential Insulation Applications 119
Richard S. Duncan

Assessment and Remediation of Misapplied Spray Polyurethane Foam 138

Ed Light

Estimating Re-Entry Times for Trade Workers Following the Application of

Three Generic Spray Polyurethane Foam Formulations 148
Richard Wood

v
Predicting TCPP Emissions and Airborne Concentrations from Spray
Polyurethane Foam Using USEPA i-SVOC Software: Parameter
Estimation and Result Interpretation 167
Shen Tian, John Sebroski, and Scott Ecoff

A Modeling Approach for Quantifying Exposures from Emissions of

Spray Polyurethane Foam Insulation in Indoor Environments 199
Charles Bevington, Zhishi Guo, Tao Hong, Heidi Hubbard, Eva Wong,
Katherine Sleasman, and Carol Hetfield

Investigating Sampling and Analytical Techniques to Understand Emission

Characteristics from Spray Polyurethane Foam Insulation and Data Needs 228
Katherine Sleasman, Carol Hetfield, and Melanie Biggs

VOC Emissions from Spray Foam Insulation Under Different Application Conditions 278
Doyun Won, Joan Wong, Gang Nong, and Wenping Yang

vi
Overview
The Symposium on Developing Consensus Standards for Measuring Chemical Emis-
sions from Spray Polyurethane Foam (SPF) Insulation was held on April 30–May 1,
2015. Sponsored by ASTM Committee D22 on Air Quality, the symposium was held
in Anaheim, CA, in conjunction with the standards development meetings of the
Committee. ASTM D22.05 is developing tools to assist decision makers in answering
fundamental questions, such as: What is emitted from SPF, how long do the emis-
sions persist, and how does ventilation impact concentrations and potential expo-
sures? How can we model these processes to address the multiplicity of products,
applications, and environmental conditions that may impact exposure to emissions
over the life cycle of the material? These are complex and interrelated questions that
have challenged the indoor environments research community for many years.
Objectives of Symposium

SPF insulation is manufactured on-site by mixing and spraying chemicals that react to
form an effective insulating material. Standardized methods are needed to assess the
potential impacts of SPF insulation products on indoor air quality, establish re-entry
times for trade workers or ­re-occupancy times for building occupants after product
installation and to evaluate post-occupancy ventilation. The objective of the sympo-
sium was to provide a forum for SPF manufacturers, regulatory agencies, indoor air
quality professionals, testing labs, air quality consultants, instrument vendors and
other stakeholders to exchange information. After a series of presentations on the
current status of measuring emissions from SPF insulation, participants discussed
paths forward for research, method development and development of standards.
The chairs of the symposium distributed a broad call for papers on the f­ ollowing
topics, designed to represent the scope of complex challenges that diverse stakehold-
ers, including industry and government, must address regarding the application and
use of SPF insulation products:

• Research and method development for measuring potential SPF emissions of

semi-volatile and volatile organic compounds used in the formulation (e.g.,
isocyanates, blowing agents, amine catalysts and flame retardants) and from
potential reaction or byproducts;
• Federal and other governmental agencies’ regulatory approaches and support-
ing investigation, assessment and research needs;
• Modeling, scaling up from lab to large scale chambers or buildings; 

vii
 • International perspective on regulation and testing of SPF insulation emissions; 
• Industry perspective/needs and product stewardship activities; 
• Field investigations or large-scale chamber/spray booth studies to evaluate
emissions or ventilation rates; and
• Applying the knowledge from product emissions data/research to practice
(e.g., stewardship commitment, green building practices, codes for residential
ventilation and global leadership).

The collaboration and exchange of information during the symposium and the cor-
responding research papers will support the development of standards at ASTM
D22.05 on Indoor Air for measuring emissions from SPF. New standards are being
developed to estimate the emissions of volatile and semi-volatile organic compounds
(e.g., blowing agents, catalysts, flame retardants, byproducts) with micro- and large-
scale chambers. Analytical methods must be developed to measure emissions from
the chambers. Specialized chambers must be evaluated for measuring isocyanate
emissions such as methylene diphenyl diisocyanate (MDI) to avoid adhesion to the
chamber’s surfaces. The data generated from the new ASTM standards may be useful
as input parameters in computer simulation modelling software to help manufactur-
ers and distributors, researchers, and government agencies assess exposure potential
and control mechanisms for SPF products. In the following paragraphs, the sym-
posium co-chairs summarize key presentations, findings, and knowledge gaps and
identify new standard development activities that have been introduced in ASTM
D22.05 as outcomes of the symposium.
Selected Technical Papers with Excerpts from Abstracts

Several topics are covered in the selected technical papers (STP) resulting from the
symposium including: field investigation studies, emissions measured in various-
sized test chambers, emissions from misapplied material, computer simulation mod-
elling of emissions and peak SPF temperatures in residential homes. According to
the paper “Investigating Sampling and Analytical Techniques to Understand Emis-
sion Characteristics from Spray Polyurethane Foam Insulation and Data Needs,”
reliable and validated emission test methods and sampling, and analytical protocols
are needed to understand the variables that affect emissions and curing in order to
develop and assess residential exposure scenarios.
The paper “Evaluation of Micro-Scale Chambers for Measuring Chemical Emis-
sions from Spray Polyurethane Foam Insulation” evaluates the use of micro-scale
chambers for measuring emissions and compares the results with conventional small-
scale chambers. The authors also investigated the effect of the chamber’s temperature
and trimming samples prior to testing. Automating the micro-scale ­chamber testing
was demonstrated in the paper “Measurement of Chemical Emissions from Spray
Polyurethane Foam Insulation Using an Automated Micro-Scale Chamber System.”
An automated dynamic headspace system was used for on-line, fully automated

viii
micro-scale chamber measurements of SPF to evaluate sampling time, volume and
temperature. The paper “Prioritizing Chemical Emissions from Closed-Cell Spray
Polyurethane Foam: Utilizing Micro-Scale Chamber Emission Factors and Field
Measurement Data” compares emission factors from the m ­ icro-scale chambers in
conjunction with a screening model to emissions measured in a residential home
after application of SPF insulation.
The paper “Flame Retardant Emissions from Spray Polyurethane Foam Insula-
tion” evaluates emissions of a commonly used flame retardant in SPF, tris (1-chloro-
2-propyl) phosphate (TCPP), with micro-scale chambers and a full scale Net Zero
Energy Residential Test Facility. The authors measured emissions in the test house
without the installation of furniture, carpeting, or other household goods to deter-
mine if SPF in the facility was the primary source of the airborne concentrations of
TCPP. This flame retardant was also investigated in the paper “Glass Chamber Meth-
od for Screening of MDI and TCPP Emissions from Foam Joint Sealant.” The goal of
this study was to develop a glass chamber method to examine the emissions of MDI
and TCPP, which were measured during a 24-hour chamber test in a 3-L chamber.
There is a great need to determine when it is safe for trade workers to re-enter a
work area where SPF was recently applied during retrofit or new construction. Emis-
sions from three generic SPF formulations were evaluated in a room-size chamber at
ventilation rates ranging from 1 to 10 air changes per hour in the paper “Estimating
Re-Entry Times for Trade Workers Following the Application of Three Generic Spray
Polyurethane Foam Formulations.” Chemicals selected for evaluation were MDI,
amine catalysts, blowing agent and flame retardant that were used in the formula-
tions. The room-size chamber test was also utilized to evaluate commercial products
in the paper “VOC Analysis of Commercially Available Spray Foam Products.” The
study was aimed at determining if worker reentry times could be reduced from the
industry practice of 24 hours if specific rates of workplace ventilation were employed.
According to the paper “Assessment and Remediation of Misapplied Spray Polyu-
rethane Foam,” the misapplication of SPF may result in occupant complaints asso-
ciated with persistent odor, and that SPF installed in homes may fail to cure and
perform as anticipated when the contractor does not follow the distributors speci-
fied pre-application and installation procedures. This paper discusses strategies for
resolving odor complaints and suggests an assessment and mitigation protocol for
field use. Emissions from misapplied SPF are also investigated in “VOC Emissions
from Spray Foam Insulation under Different Application Conditions.” The research-
ers compare chemical emissions from SPF insulation applied in four different ways in
an attempt to simulate normal and abnormal applications. Application temperature
and A to B-side ratios are investigated to determine the effect of emissions.
To begin to understand exposure to emissions from SPF and to identify and
characterize uncertainty in assessing chemical exposures, a proof-of-concept multi-
zone indoor model to estimate indoor air concentrations of chemicals is described

ix
in the paper “A Modeling Approach for Quantifying Exposures from Emissions of
Spray Polyurethane Foam Insulation in Indoor Environments.” A recently developed
­model, i-SVOC, was applied to estimate TCPP emitted from SPF and its fate and
transport in a modeled indoor environment in the paper “Predicting TCPP Emis-
sions and ­Airborne Concentrations from Spray Polyurethane Foam Using USEPA
i-SVOC software: Parameter Estimation and Result Interpretation.” In order to
evaluate the temperature gradient of SPF in buildings, a one-dimensional transient
dynamic n ­ umerical simulation was evaluated in the paper “Computer Simulation
of Peak Temperatures in Spray Polyurethane Foam Used in Residential Insulation
Applications.”

What We Have Learned and What We Need to Learn

The research methods and results reported in this STP provide a reference point from
which to evaluate progress towards development of reliable methods, data, and mod-
els that inform our understanding of the relationships between use of on-site applied
spray polyurethane foam insulation materials and potential exposures to chemicals
emitted from the insulation material. The following paragraphs attempt to capture
some of the key observations and identify the research needs that will better inform
development of tools and knowledge for managing emissions. The summary is in-
tended to stimulate the reader to probe the manuscripts and to contribute to the
discussion of progress and research needs.

Key observations and findings

What is emitted and how long do emissions persist? The chemicals identified in emis-
sions include those used to produce the insulation material, i.e., MDI and polymeric
MDI, additives to the polyols such as amine catalysts, blowing agents, flame retard-
ants, as well as chemicals that may be related to production of the primary ­ingredients,
reaction products, or minor constituents. The papers by Nie, Klein-Benne, and Thax-
ton, Poppendieck et al., and Won et al. provide insight into the range of compounds
identified in emissions reported in the course of methods development experiments
reported herein. Most of the studies reported in the STP are of relatively short dura-
tion and focus on post application emissions over periods of 2 to 336 hours. Ecoff,
Tian, and Sebroski quantified emissions of blowing agents in the attic and occupied
areas of a residence for 16 months. Poppendieck et al. observed airborne concentra-
tions of TCPP attributable to SPF that was applied in the crawl space of a research
house two years prior to testing. Ecoff, Tian, and Sebroski quantified MDI emissions
during a­ pplication in an attic, and did not detect post application emissions. Won,
Zidek, et  al. observed that MDI emissions in a small glass chamber peaked within
about 11  minutes and fell below quantification levels within about an hour. These
­observations are consistent with post application spray room results reported by Wood
and with reports in the literature as summarized by Sleasman, Hetfield, and Biggs.

x
What factors impact emissions?

Sebroski et al. demonstrate that temperature strongly impacts emission rates for vola-
tile and semivolatile emissions. Won, Wong, et al. demonstrated that environmen-
tal conditions (temperature and relative humidity) impact amount and duration of
post application MDI emissions. Poppendieck et al. demonstrated that TCPP emis-
sion rates are impacted by external mass transfer factors (gas phase concentration
in the boundary layer, air speed and turbulence) whereas more volatile emissions
may be controlled by diffusion rates within the material. Duffy and Wood demon-
strate that amine catalyst emissions vary across product types and formulations, and
Won, Wong, et al. demonstrate that application outside of recommended conditions
results in increased emissions. Sebroski et al. demonstrate that trimming the skin
affects emissions from open cell foam and that aldehyde emissions correlate better
with mass than sample surface area.
How mature are the test methods?

Sleasman, Hetfield, and Biggs provide an overview of the sampling and analysis
methods that may be employed to characterize SPF emissions. Although there are
many methods that have been validated for specific compounds or classes of com-
pounds for specific uses, e.g., occupational exposure a­ ssessment, the adaptability and
reliability of these methods for characterization of SPF insulation emissions during
and following application have not been well-demonstrated. Two of the approxi-
mately eight to ten isocyanate derivatization methods were employed in studies that
­determined MDI emissions (see papers by Won et al., Wood, and Ecoff, Tian, and
Sebroski). Won, Zidek, et al. demonstrated an emissions characterization approach
that generated model inputs for post application MDI and TCPP emissions; how-
ever, this approach has not been demonstrated for application phase MDI e­ missions.
VOC collection on single or multi-sorbent traps followed by thermal desorption with
GC/MS quantification was employed by several researchers. The methodology pro-
vides adequate detection limits for many SPF VOC emissions where sample sizes
are appropriate to the concentrations. However, the TD-GC/MS approach does not
provide adequate quantification levels for some of the amine catalysts, and sampling
media and sample volumes may need to be carefully selected.
Micro chambers, small glass and stainless steel chambers, room-sized spray
booth–spray room chambers, and residential structures were employed in emissions
tests. Uncertainty regarding interactions between reactive and semivolatile emis-
sions, MDI and TCPP, with chamber surfaces provide challenges to data interpreta-
tion in all test systems. Attempts to compare emissions across scales of test systems
were characterized by uncertainty due to many factors, including sorption of semi-
volatile compounds by surfaces, uncertainty of concentration measurements and
mixing in full-scale chambers, lack of data to characterize interzonal flows and air
exchange rates in buildings.

xi
Interpretation of results

Two general approaches were employed to interpret test results. Duffy and Wood
compare concentrations of specific emissions in a defined test environment to
­occupational exposure values. Won et al. employ the more typical indoor envi-
ronment approach where chamber emissions data are used to construct a source
model that is used in a room or building simulation model to predict potential indoor
exposure.
Model system development

Bevington presents a new modeling approach based upon two models developed
by EPA, iSVOC and IAQx, to quantify exposures due to application of SPF. Based
upon results of a sensitivity analysis, the authors concluded that experimental data
are needed to calibrate the model, particularly in the first 24 hours following appli-
cation, characterize longer term mass transfer, and characterize flow paths and air
exchange rates. Tian, Sebroski, and Escoff use literature-based parameters in the in-
door simulation system, iSVOC, to calculate TCPP concentrations in a model room
due to application of SPF insulation. Tian, Sebroski, and Escoff compare model
predictions with predictions based upon micro chamber results and conclude that
the micro chamber results were close to the upper end of the iSVOC predictions.
Sensitivity analysis suggested that the material–air partition coefficent was the most
important parameter for explaining differences between Tian’s iSVOC predictions
and micro chamber based predictions.
Summary of Research Needs

Research is needed to:

• Better characterize emission factors during the first 24 hours following appli-
cation of SPF to provide data to calibrate models and to improve ventilation
guidance for a range of scenarios.
• Characterize interactions of flame retardants with surfaces, characterize inter-
zonal flows and air exchange rates in SPF insulated buildings.
• Improve sampling and analysis of reactive amines compounds that behave
poorly in thermal desorption systems.
• Improve parameter estimation techniques and conduct experiments to validate
parameters and modeling results.
• Develop a better understanding of how micro chamber derived emission
factors can be used to predict concentrations and potential exposures in
buildings.
• Develop a better understanding of how off-ratio or application or application
outside of specified conditions impacts emissions.
• Better understand how automated and integrated chamber/sampling and
analysis systems might improve sensitivity and reliability of test methods.

xii
 • Develop an improved understanding of the impact of environmental factors
on MDI emissions, reactions and products to improve confidence in modeling
potential exposures.
• Better understand how well thermal energy models predict temperatures in
applied foam and in attics and crawl spaces.
• Develop full-scale test methods that provide emission factor inputs to models.
• Improve quality assurance procedures to ensure characterization of uncertainty
of test systems and test protocols.

ASTM Activities Related to Symposium

Knowledge learned from the symposium has been applied towards the d ­ evelopment
of standards at ASTM Subcommittee D22.05 on Indoor Air. The subcommittee
has continued to develop work item, WK40293, New Test Method for Estimating
­Chemical Emissions from Spray Polyurethane Foam (SPF) Insulation Using Micro-Scale
­Environmental Test Chambers. Several of the papers included micro-scale chamber
evaluations that examined method parameters (e.g., temperature, area specific flow
rate, etc.) and compared the emissions data to small-scale chambers or residential
buildings. Isocyanate emissions will require the use of specialized chambers to avoid
­significant bias from adhesion to the chamber walls; therefore, work item WK51589
was created, New Test Method for Estimating Emissions of Methylene diphenyl diisocy-
anate (MDI) from Spray Polyurethane Foam (SPF) Insulation using Emission Cells or
Micro-Scale Environmental Test Chambers.
Two of the symposium papers discussed the use of a large-scale spray room at
various ventilation rates as an approach to evaluate potential impacts on the use of
protective equipment and controls and re-entry times for trade workers without the
use of protective equipment after application of the material. This approach may
be useful to provide a better understanding of the emissions that occur during and
post application under controlled conditions. As a result, a new work item has been
created, WK51588, New Test Method for Measuring Chemical Emissions from Spray
Polyurethane Foam (SPF) Insulation Samples in a Large-Scale Spray Room. The emis-
sions data from the large-scale spray room may also be used in conjunction with
computer simulation models as a means of calibration or verification. Two of the
papers from the symposium discussed estimating emissions in buildings using such
models. The input parameters and the use of models to estimate SPF emissions are
being addressed in a new work item, WK52052, New Practice for Application of
Indoor Emission and Fate Modeling for Spray Polyurethane Foam (SPF) Insulation. In
the future, additional standards may be necessary to measure key input parameters
for the simulation models.

xiii
 DEVELOPING CONSENSUS STANDARDS FOR MEASURING CHEMICAL EMISSIONS 1

STP 1589, 2017 / available online at www.astm.org / doi: 10.1520/STP158920150039

John Sebroski,1 Jason W. Miller,1 Carl P. Thompson,1 and

Elizabeth Roeske1

Evaluation of Micro-Scale
Chambers for Measuring
Chemical Emissions from
Spray Polyurethane Foam
Insulation
Citation
Sebroski, J., Miller, J. W., Thompson, C. P., and Roeske, E., “Evaluation of Micro-Scale
Chambers for Measuring Chemical Emissions from Spray Polyurethane Foam Insulation,”
Developing Consensus Standards for Measuring Chemical Emissions from Spray Polyurethane
Foam (SPF) Insulation, ASTM STP1589, J. Sebroski and M. Mason, Eds., ASTM International,
West Conshohocken, PA, 2017, pp. 1–26, http://dx.doi.org/10.1520/STP1589201500392

ABSTRACT
Consensus standards are being developed at ASTM International to identify
potential emissions from spray polyurethane foam (SPF) insulation products and
to evaluate their impact on indoor air quality after installation. A method is
currently being developed to measure emissions of SPF insulation using micro-
scale chambers. In order to evaluate emission decay rates and the effects of
temperature, generic SPF formulations (closed-cell and open-cell) were tested
for emissions in micro-scale chambers (0.114 L) at ambient (23 C) and two
elevated temperatures (40 and 65 C) for 20 days. Small-scale chamber (36-L)
tests were conducted simultaneously to correlate the emission factors that
were measured with the micro-scale chambers. Thermal desorption gas
chromatography-mass spectrometry was used to capture and quantitate target
volatile organic compounds (blowing agents, amine catalysts, etc.) and the flame
retardant used in both formulations. Aldehydes were captured onto silica gel
sorbent tubes with dinitrophenylhydrazine and analyzed with high-performance

Manuscript received May 14, 2015; accepted for publication August 28, 2016.
1
Covestro LLC, 1 Covestro Circle, Pittsburgh, PA 15205
2
ASTM Symposium on Developing Consensus Standards for Measuring Chemical Emissions from Spray
Polyurethane Foam (SPF) Insulation on April 30–May 1, 2015 in Anaheim, CA.

Copyright V
C 2017 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959.
2 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

liquid chromatography. After the application of open-cell SPF, the outer skin layer
of the foam may be trimmed to align the surface with wall studs or other structural
elements. In this case, it may be necessary to mimic this practice when preparing
samples for analyses to measure emissions. In order to investigate the impact of
emissions on trimmed SPF samples, micro-scale chamber studies were conducted
on both trimmed and untrimmed generic open-cell SPF material for comparison.
The data from this study demonstrate that micro-scale chambers can be used to
identify and quantitate potential emissions from SPF insulation. The findings from
this research will be used to support the development of consensus standards
in ASTM Committee D22 on Air Quality, Subcommittee D22.05 on Indoor Air.

Keywords
spray polyurethane foam insulation, SPF, emissions, VOCs, micro-scale chamber,
small-scale chamber, temperature, indoor air quality

Introduction
PURPOSE OF STUDY
Consensus standards are being developed in ASTM International’s Committee D22 on
Air Quality to identify potential emissions from spray polyurethane foam (SPF) insula-
tion products and to evaluate their impact on indoor air quality after installation. There
are special considerations for evaluating chemical emissions of SPF because they are
produced by means of a chemical reaction within the building. There is a wide range
of volatility and reactivity of compounds that are used to produce SPF insulation
(e.g., isocyanates, blowing agents, amine catalysts, flame retardants), and there are
currently no standardized methods to measure the emissions of many of these com-
pounds. Methods are needed to measure chemical emissions from SPF insulation to
help manufacturers evaluate their products and regulatory agencies to make risk assess-
ments. Micro-scale chambers were evaluated during this study at selected temperatures
to measure chemical emissions from SPF insulation, and results were compared with
conventional small-scale chambers. The effect of removing the surface skin of SPF was
also examined. The findings from this research will be used to support the develop-
ment of consensus standards in ASTM Subcommittee D22.05 on Indoor Air.

BACKGROUND AND OBJECTIVES

During the last few years, an ASTM task group created several work items to develop
standards to measure chemical emissions from SPF insulation. The American
Chemistry Council, Center for the Polyurethanes Industry (CPI), supported research
for the development of ASTM standards for measuring emissions from SPF insula-
tion [1]. Part of this research led to the development of ASTM D7859-13e1,
Standard Practice for Spraying, Sampling, Packaging, and Test Specimen Preparation
of Spray Polyurethane Foam (SPF) Insulation for Testing of Emissions Using Environ-
mental Chambers [2], to standardize the way samples are sprayed, sampled, packaged,
 SEBROSKI ET AL., DOI 10.1520/STP158920150039 3

and prepared for environmental chamber testing. The ASTM practice describes speci-
men preparation for both micro- and small-scale test chambers.
Small-scale chambers, described in ASTM D5116-10, Standard Guide for Small-Scale
Environmental Chamber Determinations of Organic Emissions from Indoor Materials/
Products [3], range in size from a few liters to a few cubic meters and can be used to deter-
mine emissions of volatile organic compounds (VOCs) from various types of building
materials and consumer products. Micro-scale test chambers range in size from a few
milliliters to 150 mL. According to ASTM D7706-11, Standard Practice for Rapid Screen-
ing of VOC Emissions from Products Using Micro-Scale Chambers [4], micro-scale cham-
bers can be operated at moderately elevated temperatures to enhance analytical signals,
reduce the test time, boost emission rates, and facilitate the screening of emissions from
semivolatile organic compounds (SVOCs). Other chambers, such as the full-scale
chambers described in ASTM D6670-13, Standard Practice for Full-Scale Chamber
Determination of Volatile Organic Emissions from Indoor Materials/Products [5], permit
the testing of complete assemblages or may be used to evaluate activities such as spraying.
According to ASTM D7859-13e1 [2], the surface skin of the SPF should not be
removed for chamber testing unless it is specified by the sample submitter. When
open-cell SPF is sprayed at a construction site, the material may be over-sprayed
beyond the wall studs or other structural elements. When this occurs, the SPF insu-
lation’s surface skin must be trimmed off of the material, exposing the open cells of
the foam. It is not clear how the emissions would be affected if the surface skin is
removed during chamber testing. One possibility is that the sample’s skin helps to
seal the surface of the SPF, reducing the emissions measured in the micro-scale
chamber; however, there are no known data to support this hypothesis.
To further investigate the use of micro-scale chambers for measuring chemical
emissions from SPF insulation, 20-day emissions studies were conducted with two
SPF formulations with the following objectives:
• To compare emissions measured with micro-scale chambers to small-scale
chambers at ambient temperatures
• To evaluate how elevated temperatures affect the emissions of SPF insulation
in micro-scale chambers and the appropriate time(s) to collect emissions sam-
ples from the micro-scale chambers by observing the decay curves
• To compare emissions of open-cell SPF with surface skin removed to those
with the skin intact

Apparatus and Instrumentation

INSTRUMENTATION
The following instrumentation was used during the study:
• Markes International (Llantrisant, Wales, UK) TD-100 thermal desorber con-
nected to Agilent Technologies (Santa Clara, CA) 7890A gas chromatography
(GC) and 5975C Inert XL mass selective detector (MSD), and
• Agilent Technologies 1100 Series LC/MSD and diode array detector
4 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

ENVIRONMENTAL TEST CHAMBERS AND ACCESSORIES

The following environmental rest chambers and accessories were used during the study:
• Markes International (Llantrisant, Wales, UK) M-CTE250 micro-chamber/
thermal extractor, 114-mL capacity, and
• Small-scale chamber, Eagle Stainless (Warminster, PA) CTH-36-316L-J, 36-L
R
capacity, electropolished stainless steel with VitonV gasket

LAB EQUIPMENT, SAMPLING PUMPS, AND FLOW CONTROLLERS

The following lab equipment, sampling pumps and flow controllers were used
during the study:
• Ultra zero grade air, compressed, air source for micro chambers
• Parker Filtration, Balston (Lancaster, NY) 75-83N zero-air generator, air
source for small-scale chambers
• Parker Porter (Hatfield, PA) VCD-ABF-110 low-flow controller with optional
flow element, 0-535 cc/min
• Markes International (Llantrisant, Wales, UK) TC-20 tube conditioner
• Markes International (Llantrisant, Wales, UK) calibration solution loading rig
• DryCal (Butler, NJ) BIOS Defender 510 volumetric primary flow standard, cali-
brator, and
R
• Thermo Fisher Scientific (Waltham, MA) TraceableV long-stem thermometer,
Cat. No. S40799-3

SORBENT TUBES AND MEDIA FOR SAMPLE COLLECTION

The following sorbent tubes and media for sample collection were used during this
study:
• Markes International (Llantrisant, Wales, UK) thermal desorption tubes,
R TM
stainless steel, 5-mm quartz wool, TenaxV TA 35/60, Carbograph 5TD 40/60
• Markes International (Llantrisant, Wales, UK) thermal desorption tubes, stainless
R TM R
steel, TenaxV TA 35/60, Carbograph 1TD 40/60, CarboxenV 1003 40/60, and
• SKC Inc. (Eighty Four, PA) treated DNPH silica gel, 8 by 110 mm, Cat.
No. 226-119

SPRAYING EQUIPMENT
TM
A Graco (Minneapolis, MN) H-XP2 proportioner with T2 2:1 transfer pumps was
used to spray the samples during the study.

SAMPLE COLLECTION EQUIPMENT

The following sample collection equipment was used during the study:
• Sheets of high-density polyethylene, 0.95-cm thick, cut into 30.5 by 30.5-cm pieces
• Aluminum foil, heavy-gage roll, approximately 0.024 mm thick
• Electric knife and band saw to scarf and cut SPF samples
• Circular foam coring tool constructed of steel to cut SPF insulation samples to
fit tightly into micro-scale chambers
R
• MylarV bags with aluminum foil layer, light-resistant, composite layer approxi-
mately 0.127 mm thick; for closed-cell SPF and the kit formulation, a bag with
a zipper seal with dimensions of approximately 46 by 71 cm was used; for
 SEBROSKI ET AL., DOI 10.1520/STP158920150039 5

open-cell SPF, a bag size of approximately 51 by 76 cm was used (zipper seal not
available in this size)

Sample Matrices and Target Compounds

GENERIC SPF FORMULATIONS
Two generic high-pressure SPF formulations were evaluated during this study: low-
density open-cell SPF and medium-density closed-cell SPF (Table 1). The generic
formulations were developed by the CPI for research purposes. The formulations
do not reveal confidential information of formulations sold in the marketplace
today; rather, they are meant to represent typical commercial systems in terms of
their density, reactivity, and volume ratios. Although not completely optimized or
code compliant, the formulations are meant to be representative of commercial for-
mulations and are suitable for the purpose of method development. The two generic
SPF formulations represent the following sample types:
• Open-cell, low-density (8 kg/m3 or 0.5 lb/ft3), high-pressure SPF
• Closed-cell, medium-density (32 kg/m3 or 2 lb/ft3), high-pressure SPF

TABLE 1 Generic SPF formulations.

Low-Density, High-Pressure, Open-Cell SPF Medium-Density, High-Pressure, Closed-Cell SPF

Formulation Formulation

A side (50 % of formulation sprayed at a 1:1 A to B ratio)

PMDI (100 %) PMDI (100 %)
B side (50 % of formulation sprayed at a 1:1 A to B ratio)
Polyether polyol (34 %) Aromatic polyester polyol (36.4 %)
Aromatic amino polyether polyol (33.6 %)
NPE emulsifier (11.9 %)
Blowing Agent
Water (20 %) HFC-245fa (6.97 %)
Water (2.55 %)
Flame Retardant
TCPP (25.2 %) TCPP (15.91 %)
Silicone surfactant (1.0 %) Silicone surfactant (1.0 %)
Catalyst
BDMAEE (0.9 %) BDMAEE (0.7 %)
TMIBPA (3.0 %) DAPA (2.59 %)
TMAEEA (4.0 %) TMAEEA (0.3 %)

Note: BDMAEE ¼ bis(2-dimethylaminoethyl) ether; DAPA ¼ bis(dimethylaminopropyl) methylamine;

HFC-245fa ¼ 1,1,1,3,3-pentafluoropropane; NPE ¼ nonylphenol ethoxylate; PMDI ¼ polymeric methy-
lene diphenyl diisocyanate; TCPP ¼ tris(1-chloro-2-propyl) phosphate; TMAEEA ¼ N,N,N-trimethyla-
minoethylethanolamine; TMIBPA ¼ tetramethyliminobispropylamine.
6 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 2 List of target compounds.

Target Compound List CAS Number Description Analytical Method

HFC-245fa 460-73-1 Blowing agent TD-GC-MS

BDMAEE 3033-62-3 Catalyst TD-GC-MS
TMIBPA 6711-48-4 Catalyst TD-GC-MS
TMAEEA 2212-32-0 Catalyst TD-GC-MS
DAPA 3855-32-1 Catalyst TD-GC-MS
TCPP 13674-84-5 Flame retardant TD-GC-MS
Formaldehyde 50-00-0 Aldehyde (not in formulation) DNPH-LC
Acetaldehyde 75-07-0 Aldehyde (not in formulation) DNPH-LC
Propionaldehyde 123-38-6 Aldehyde (not in formulation) DNPH-LC

Note: BDMAEE ¼ bis(2-dimethylaminoethyl) ether; DAPA ¼ bis(dimethylaminopropyl) methyl-

amine; HFC-245fa ¼ 1,1,1,3,3-pentafluoropropane; TCPP ¼ tris(1-chloro-2-propyl) phosphate;
TMAEEA ¼ N,N,N-trimethylaminoethylethanolamine; TMIBPA ¼ tetramethyliminobispropylamine.

TARGET COMPOUNDS
The blowing agent, amine catalysts, and flame retardants used in the generic formula-
tions described in Table 2 were treated as the primary compounds to evaluate the cham-
ber test methods during this study. Although aldehydes were not used in either of the
product formulations, emissions of aldehydes are usually monitored from building
products; therefore, aldehydes observed during previous research with the generic for-
mulations were included in this study [1]. The isocyanate compounds (e.g., methylene
diphenyl diisocyanate [MDI], polymeric MDI) were not monitored during this study
because specialized chambers are necessary to monitor emissions in order to minimize
the adsorption of isocyanates to the chamber surfaces [6].

EXPERIMENTAL DESIGN
SAMPLE PREPARATION
TM
The generic SPF was sprayed in a spray room with a Graco Reactor H-XP2 propor-
tioner at a 1:1 ratio, and samples were prepared for chamber testing as described in
ASTM D7859-13e1 [2]. After the SPF samples were sprayed, they were allowed to cure
for 1 h and then were wrapped in aluminum foil prior to placing each sample into indi-
R
vidual MylarV bags, which were stored for approximately 24 h at room temperature
(approximately 23 C) prior to chamber testing. Immediately after opening the bags,
specimens of SPF were cored to fit tightly into the micro-chambers (6.4-cm diameter;
approximately 3 cm thick). The bottom of the samples was trimmed so that there was
minimal headspace in the micro-scale chambers (approximately 5 mm). The open-cell
samples were also prepared with approximately 2 mm of the top surface (skin)
removed but with the same thickness (3 cm) for comparison. Three sets of samples
R
from separate MylarV bags were prepared to evaluate three test temperatures. Duplicate
specimens from the same block of foam were prepared for every test sample.
 SEBROSKI ET AL., DOI 10.1520/STP158920150039 7

The samples were also prepared for testing in 36-L electropolished stainless
steel small-scale chambers. Specimens of open-cell SPF were cut 13 by 13 by 9 cm.
Specimens of closed-cell SPF were cut 13 by 13 by 6 cm. Open-cell SPF was
trimmed at the top with the surface skin removed so that there would be a smooth
exposed surface. The cut specimens were tightly fitted into a steel box of the same
inner dimensions. Duplicate specimens were prepared for each test sample.

CHAMBER TESTING
The loading factor (L), which is described in ASTM D5116-10 [3], was calculated to
be 200 m2/m3 in the micro-scale chambers and 0.490 m2/m3 in the small-scale
chambers. The air change rate (N) was 188 air exchanges per hour in the micro-
scale chambers and 0.522 air exchanges per hour in the small-scale chambers. The
36-L small-scale chambers were constructed of electropolished stainless steel with
three 0.25-in. (0.635-cm) stainless steel compression fitting ports (inlet, outlet, and
sampling port). To ensure adequate mixing in the small-scale chambers, a polyte-
trafluoroethylene drop tube from the air inlet port was placed near the sample sur-
face; another drop tube was placed midway in the chambers to collect air samples
from the chambers. The air velocity at the air inlet above the sample was calculated
to be approximately 15.8 cm/s in the small-scale chambers and 10.5 cm/s in the
micro-scale chambers. The higher loading factor and air change rate in the micro-
scale chambers may improve the recovery of reactive compounds and SVOCs.
However, the area-specific flow rate (N/L) for the micro- and small-scale chambers
was similar (0.938 and 1.07 m/h, respectively). For consistency during this study,
dry air was used as the carrier gas for both chambers, and the relative humidity was
<1 %. The chamber parameters are summarized in Table 3.

TABLE 3 Chamber parameters.

Chamber Parameter Units Micro-scale Chamber Small-scale Chamber

Flow rate mL/min 50 6 2 300

Sampling rate mL/min Same as flow rate 100
Chamber headspace volume m3 0.0000161 0.0345
Loading factor (L) m2/m3 200 0.490
Air change rate (N) h1 188 0.522
Area-specific flow rate (N/L) m/h 0.938 1.07
Relative humidity % <1 <1
Sample area m2 0.00322 0.0169
Sample mass g 0.7–4.0 12–30
Estimated air velocity at inlet cm/s 10.5 15.8

Note: Micro-scale chamber headspace assumes a planar surface and approximately 5-mm head-
space. The sample mass varies and depends on the formulation and trimming. The small-scale
chambers were operated only at ambient temperature, which was 23 6 2 C during this study. Three
micro-chamber apparatuses, each containing four micro-scale chambers, were operated indepen-
dently of one another at 23 (ambient), 40, and 65 C. The elevated temperatures were controlled
with the micro-scale chamber apparatus and verified to be accurate within 62 C with a National
Institute of Standards and Technology traceable long-stem thermometer.
8 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

SAMPLING FOR EMISSIONS

Airborne concentrations of amine catalysts, blowing agents, and flame retardants
were collected from the outlet of chambers using thermal desorption (TD) tubes.
The TD tubes used to collect VOCs and flame retardants were stainless steel tubes
R TM
containing quartz wool, TenaxV TA 35/60, and Carbograph 5TD 40/60. Highly
volatile organic compounds (blowing agents) were captured with stainless steel TD
R TM
tubes containing TenaxV TA 35/60, Carbograph 1TD 40/60, and Carboxen 1003
40/60. The TD tubes selected for this study were previously evaluated for measuring
emissions from SPF [1,7]. The carbonyl compounds (aldehydes) were sampled
with glass sorbent tubes containing silica gel coated with dinitrophenylhydrazine
(DNPH) as described in ASTM D5197-09e1, Standard Test Method for Deter-
mination of Formaldehyde and Other Carbonyl Compounds in Air (Active Sampler
Methodology) [8]. The sampling flow rate for the micro-scale chambers was 50 mL/
min; the sampling flow rate for the small-scale chambers was 100 mL/min. The
sampling time and volume to collect air samples from the chambers were based
on previous research at 23 C [1,7]. Sample volumes for sampling TD tubes from
micro-scale chambers at elevated temperatures were decreased to avoid possibly
overloading the tubes.
Samples were collected at various times after being loaded into the chambers
when the chamber testing concluded. The emissions from the closed-cell SPF were
collected at 2, 24, 48, 72, 168, 216, 336, and 480 h. The emissions from the open-cell
SPF were collected at 2, 24, 48, 120, 168, 336, and 480 h. Maximum sampling vol-
umes with TD tubes were established for target compounds during previous
research [7]. The detection limits were dependent on the duration of emission
sampling from the chambers with TD tubes, which ranged from 1 min to 2 h
depending on the time each sample resided in the chamber and the chamber’s
temperature.
Aldehydes were collected on DNPH tubes for 16 h to optimize detection lim-
its. The sampling rate for micro- and small-scale chambers was 50 and 100 mL/min,
respectively. Sorbent tubes were placed directly on the outlet of the micro-scale
chamber; small-scale chambers were sampled with a personal sampling pump. All
flows were verified with a volumetric primary flow standard.

ANALYTICAL METHODS
A Markes International TD-100 thermal desorber coupled to an Agilent 7890/5975
gas chromatograph-mass spectrometer was used to analyze the tubes with method-
ology based on U.S. EPA TO-17 and ISO 16000, Part 6 [9,10]. Based on previous
research, the base-deactivated transfer line, analytical column, and method
conditions were optimized for the recovery of target compounds [1]. Formaldehyde,
acetaldehyde, and propionaldehyde were analyzed with an Agilent 1100 series high-
performance liquid chromatograph with the use of the methodology based on
ASTM D5197-09e1 [8].
 SEBROSKI ET AL., DOI 10.1520/STP158920150039 9

Results and Discussion

REPORTING AND DUPLICATE SAMPLES
The emissions results are reported as emission factors, which were calculated by direct
calculation from individual concentration data points as described in ASTM D5116-
10, Section 9.4.1.1 [3]. According to ASTM D5116, calculating emission factors prior
to reaching steady-state conditions can increase error; however, this approach is con-
sistent with the draft ASTM method being investigated. Unless otherwise specified,
the emission factors are reported as area-specific emission rates in lg/m2 per hour.
To be consistent with the draft ASTM method, all test results are reported as the
mean of duplicate chamber samples with a relative percent difference (RPD) <25 %.
If the RPD exceeded 25 %, the higher of the two values were reported as a conserva-
tive approach to reporting the emissions from the sample. Reporting limits were
established with the analytical method detection limits from previous research [7].
The reporting limits for emission factors are based on these detection limits and the
sample collection times onto the sorbents that varied depending on the expected
emissions. The reporting limits are reported in Table 4 and Table 5. The decay of target
compounds was monitored during a 20-day test period. The reported emissions rates
are those measured at the time emissions were collected from the test chambers.

TABLE 4 Generic open-cell SPF: micro- versus small-scale chambers.

Hours in Chamber
Chamber
Target Compound Chamber Blank 2 24 48 120 168 336 480

TMAEEA Micro-scale <74.4 <893 <446 <446 <223 <223 <99.2 <74.4
Small-scale <42.5 <1,020 <510 <510 <255 <255 <113 <85
BDMAEE Micro-scale <5.00 2,985 1,957 1,948 929 689 395 244
Small-scale <3.00 1,125 1,189 1,466 821 625 646 551
TCPP Micro-scale <4.00 <48.5 <24.2 25.0 19.0 23.0 25.7 25.2
Small-scale <2.30 <55.4 <27.7 <27.7 <13.8 <13.8 11.2 14.7
Formaldehyde Micro-scale <2.26 3.28 3.03 <2.23 <2.30 1.96 <1.67 <2.11
Small-scale <1.18 8.75 8.04 7.29 8.09 6.53 6.82 7.43
Acetaldehyde Micro-scale <2.26 <2.73 <2.37 <2.23 <2.30 <1.88 <1.67 <2.11
Small-scale <1.18 5.88 3.00 2.19 2.31 1.38 1.28 1.40
Propionaldehyde Micro-scale <2.26 <2.73 <2.37 <2.23 <2.30 <1.88 <1.67 <2.11
Small-scale 13.0 21.3 B 7.69 B 2.96 B <1.31 <1.08 <0.952 <1.21

Note: BDMAEE ¼ bis(2-dimethylaminoethyl) ether; TCPP ¼ tris(1-chloro-2-propyl) phosphate;

TMAEEA ¼ N,N,N-trimethylaminoethylethanolamine. Emission factors are reported as lg/m2 per
hour. Emission factor reporting limits were calculated with the detection limit on the sorbent tube/
cartridge, sampling time/volume, and surface area of samples. The reporting limits shown with a
“less than” symbol vary depending on the sampling conditions. Results qualified with a B indicate
not significantly greater than the corresponding chamber blank.
10 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 5 Generic closed-cell formulation: micro- versus small-scale chambers.

Hours in Chamber
Chamber
Target Compound Chamber Blank 2 24 48 72 168 216 336 480

HFC-245fa Micro-scale <7.61 2,164 1,408 1,004 948 457 408 328 334
Small-scale <4.35 2,112 818 533 526 269 353 257 156
TMAEEA Micro-scale <74.4 <149 <149 <149 <149 <74.4 <74.4 <74.4 <149
Small-scale <42.5 <170 <170 <170 <85.0 <85.0 <85.0 <85.0 <85.0
BDMAEE Micro-scale <4.97 <9.94 <9.94 <9.94 <9.94 <4.97 <4.97 <4.97 <9.94
Small-scale <2.84 <11.4 <11.4 <11.4 <5.68 <5.68 <5.68 <5.68 <5.68
DAPA Micro-scale <34.5 <68.9 <68.9 <68.9 <68.9 <34.5 <34.5 <34.5 <68.9
Small-scale <19.7 <78.8 <78.8 <78.8 <39.4 <39.4 <39.4 <39.4 <39.4
TCPP Micro-scale <4.04 13.1 36.0 17.2 8.94 4.19 <4.04 <4.04 <8.07
Small-scale <2.30 <9.23 12.0 <9.23 <4.62 <4.62 <4.62 <4.62 <4.62
Formaldehyde Micro-scale <2.10 <2.22 <1.90 <1.90 <1.85 <1.94 <1.90 <2.27 <1.67
Small-scale <1.18 4.92 6.44 3.18 3.17 3.82 3.35 2.96 3.58
Acetaldehyde Micro-scale <2.10 30.0 15.8 8.91 8.50 5.51 4.57 3.72 3.07
Small-scale <1.18 42.3 24.0 12.6 11.7 9.41 7.44 5.74 6.19
Propionaldehyde Micro-scale <2.10 <2.22 5.40 <1.90 <1.85 <1.94 <1.90 <2.27 <1.67
Small-scale 13.0 28.1 B 15.3 B 10.0 B 7.61 B 4.88 B 4.74 B 4.20 B 3.43 B

Note: BDMAEE ¼ bis(2-dimethylaminoethyl) ether; DAPA ¼ bis(dimethylaminopropyl) methyl-

amine; HFC-245fa ¼ 1,1,1,3,3-pentafluoropropane; TCPP ¼ tris(1-chloro-2-propyl) phosphate;
TMAEEA ¼ N,N,N-trimethylaminoethylethanolamine. Emission factors are reported as lg/m2 per
hour. Emission factor reporting limits were calculated with the detection limit on the sorbent tube/
cartridge, sampling time/volume, and surface area of samples. The reporting limits shown with a
“less than” symbol vary depending on the sampling conditions. Results qualified with a “B” indicate
not significantly greater than corresponding chamber blank.

QUALITY CONTROL
Prior to chamber testing, each chamber configuration was evaluated for contamination
by testing emissions from empty chambers and sample holders. The emissions collected
from the blank chambers were analyzed for target compounds concurrently with the
SPF samples. Blank chamber data are reported along with the corresponding samples.
Calibration check samples were analyzed during each analytical sequence. A summary
of the performance of the calibration checks is reported in Table 6. The amine catalyst
tetramethyliminobispropylamine (TMIBPA) showed relatively poor precision (29.7 %
RSD) and bias (36.7 %) compared to the other target compounds; although not
detected in any of the samples, this compound was not reported. Additional method
development may be necessary for this compound and possibly other reactive amine
catalysts that may be challenging to measure with the TD-GC-MS method.

MICRO-SCALE CHAMBERS COMPARED TO SMALL-SCALE CHAMBERS

The emissions measured from open- and closed-cell SPF at ambient temperature are
reported as area-specific emission factors in micro- and small-scale chambers.
Duplicate chamber results are reported as mean values in Table 4 and Table 5.
 SEBROSKI ET AL., DOI 10.1520/STP158920150039 11

TABLE 6 Summary of calibration checks.

Compound Mass (ng) Bias (%) Precision (% RSD) Replicates (n)

HFC-245fa 1,630 6.94 12.7 5

TMAEEA 1,589 3.50 16.5 5
BDMAEE 1,138 3.86 11.6 9
TMIBPA 1,321 36.7 29.7 9
DAPA 1,347 4.97 3.72 5
TCPP 1,598 5.27 13.4 9
Formaldehyde 2,500 1.62 3.93 5
Acetaldehyde 2,500 1.91 3.45 5
Propionaldehyde 2,500 1.13 4.94 5

Note: BDMAEE ¼ bis(2-dimethylaminoethyl) ether; DAPA ¼ bis(dimethylaminopropyl) methylamine;

HFC-245fa ¼ 1,1,1,3,3-pentafluoropropane; TCPP ¼ tris(1-chloro-2-propyl) phosphate; TMAEEA ¼
N,N,N-trimethylaminoethylethanolamine; TMIBPA ¼ etramethyliminobispropylamine. Calibration
checks were analyzed at the beginning and end of each analytical sequence.

Bis(2-dimethylaminoethyl) ether (BDMAEE) was the primary compound to emit

from the generic open-cell SPF decaying from 2,985 lg/m2 per hour on the day
the chambers were loaded to 244 lg/m2 per hour at the end of the 20-day study
(reported values from micro-scale chambers); however, no emissions of this com-
pound were detected from the closed-cell formulation. This compound is a non-
reactive amine catalyst. Because it is not consumed during the foaming process
and has a relatively low molecular weight, this compound is highly emissive [11].
Conversely, no emissions of the reactive catalysts bis(dimethylaminopropyl)
methylamine (DAPA) and N,N,N-trimethylaminoethylethanolamine (TMAEEA)
were detected from the generic foams in either chamber during the entire study.
Method detection limits (MDLs) for the reactive catalysts TMAEEA, TMIBPA,
and DAPA were 479, 654, and 222 ng/sample, respectively, whereas the MDL for
the nonreactive catalyst BDMAEE was much lower (32.2 ng/sample).
Based on duplicate chamber testing data from this study, the BDMAEE emis-
sions in the open-cell SPF appeared to be higher in the micro-scale chambers for the
first 2 days of monitoring than the small-scale chambers, and then the decay rate was
similar throughout the duration of the study. The best correlation of the data
occurred during days 5 and 7, when the emissions factors were 12 % and 9.7 % RPD,
respectively. The decay curves for BDMAEE are shown in Fig. 1 for comparison. The
reduced initial concentration of BDMAEE in the small-scale chamber may have been
from the catalyst adsorbing to the surface of the chamber’s walls or an indication that
the mass transfer conditions are not the same for the two systems. Conversely, the
adsorbed compound could have re-emitted at a later time, which may have contribut-
ed to the higher results in the small-scale chambers at days 14 and 20.
The blowing agent 1,1,1,3,3-pentafluoropropane (HFC-245fa) was observed
in both the small- and micro-scale chambers. The decay curve is shown in Fig. 2.
12 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 1 Comparison of BDMAEE emissions from open-cell SPF in micro- and small-scale
chambers.

3,500

3,000

2,500
Emission Factor, µg/m2 h

2,000

Micro-Scale
1,500
Small-Scale
12 % RPD
1,000
10 % RPD 48 % RPD
500

0
0 100 200 300 400 500
Hours in Chamber

FIG. 2 Comparison of HFC-245fa emissions from closed-cell SPF in micro- and

small-scale chambers.

2,500
Micro-Scale
Small-Scale
2,000
Emission Factor, µg/m2 h

1,500

1,000

500 14 % RPD 24 % RPD

0
0 100 200 300 400 500
Hours in Chamber
 SEBROSKI ET AL., DOI 10.1520/STP158920150039 13

The HFC-245fa emissions decayed from 2,164 to 334 lg/m2 per hour over the 20-
day test period (reported values from micro-scale chambers). The HFC-245fa emis-
sion factors were consistently higher in the micro-scale chambers than the small-
scale chambers, although their decay curves were similar. The emissions data ap-
pear to correlate best after 7 days (168 h) in the chambers.
The emission factors for TCPP reported in Table 4 and Table 5 appeared to be
higher in the micro-scale chambers compared to the small-scale chambers. Sorption
of TCPP by small-scale chamber interior surfaces may be the reason the emission
factor appeared to increase from 13.1 lg/m2 per hour at 2 h to 36.0 lg/m2 per hour
at 24 h in the open-cell SPF in the micro-scale chamber. In addition, because the
emissions of TCPP may be externally controlled by diffusion through the boundary
layer at the sample surface, this difference may be due to the air velocity or higher
air exchange rate used in the micro-scale chambers [12].
The aldehyde emissions appeared to be greater in the small-scale chambers,
although the measured amounts in the micro-scale chambers were near the detec-
tion limits, thus making comparisons difficult. Propionaldehyde was detected in the
small-scale chamber blank, although the source of the contamination is unknown;
results are reported without background subtraction. Results that are qualified with
a B indicate not significantly greater than the corresponding chamber blank. The
detection limits for aldehydes were a factor of 2 times higher with micro-scale
chambers due to decreased sample volume collected and the smaller surface area in
the micro-scale chambers. There was a greater height and mass of sample per unit
surface area in the small-scale chambers, which could contribute to increased
emissions of aldehydes if the compounds readily diffused through the material. To
evaluate this possibility, the emissions factors were recalculated based on the mass
of the sample in the chamber rather than the exposed surface area. The mass-
specific emissions are reported as lg/kg per hour for comparison in Table 7 and
Table 8. This appeared to have improved the correlation between chambers for

TABLE 7 Mass-specific emissions for aldehydes in generic open-cell SPF.

Hours in Chamber
Chamber
Target Compound Chamber Blank 2 24 48 120 168 336 480

Formaldehyde Micro-scale <7.27 14.2 13.1 <9.61 <9.90 8.50 <7.18 <9.09
Small-scale <1.70 12.2 11.4 10.3 11.4 10.7 9.62 10.6
Acetaldehyde Micro-scale <7.27 <11.8 <10.3 <9.61 <9.90 <8.12 <7.18 <9.09
Small-scale <1.70 8.19 4.25 3.10 3.31 1.94 1.80 1.97
Propionaldehyde Micro-scale <7.27 <11.8 <10.3 <9.61 <9.90 <8.12 <7.18 <9.09
Small-scale 18.4 29.7 B 10.9 B 4.23 B <1.88 B <1.50 <1.33 <1.73
2
Note: Emission factors are reported as lg/m per hour. Emission factor reporting limits were calcu-
lated with the detection limit on the sorbent tube/cartridge, sampling time/volume, and surface area
of samples. The reporting limits shown with a “less than” symbol vary depending on the sampling
conditions. Results qualified with a B indicate not significantly greater than the corresponding
chamber blank.
14 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 8 Mass-specific emissions for aldehydes in generic closed-cell formulation.

Hours in Chamber
Chamber
Target Compound Chamber Blank 2 24 48 72 168 216 336 480

Formaldehyde Micro-scale <1.69 <1.81 <1.55 <1.55 <1.51 <1.59 <1.55 <1.85 <1.36
Small-scale <0.665 2.73 3.57 1.76 1.76 2.12 1.86 1.56 1.88
Acetaldehyde Micro-scale <1.69 25.6 13.4 7.60 7.24 4.69 3.73 3.31 2.20
Small-scale <0.665 23.5 13.3 7.75 6.47 5.22 4.13 3.03 3.26
Propionaldehyde Micro-scale <1.69 <1.81 4.81 <1.55 <1.51 <1.59 <1.55 <1.85 <1.36
Small-scale 7.34 15.6 B 8.47 B 5.55 B 4.22 B 2.71 B 2.63 B 2.33 B 1.90 B
2
Note: Emission factors are reported as lg/m per hour. Emission factor reporting limits were calculated
with the detection limit on the sorbent tube/cartridge, sampling time/volume, and surface area of sam-
ples. The reporting limits shown with a “less than” symbol vary depending on the sampling conditions.
Results qualified with a B indicate not significantly greater than the corresponding chamber blank.

aldehydes. In Fig. 3, acetaldehyde decay curves from mass-specific emission factors

are compared in micro- and small-scale chambers. These data indicate that mass-
specific emission factors should be reported for aldehydes.

TEMPERATURE EVALUATION WITH OPEN-CELL SPF

Conventional small-scale chambers are normally maintained between approximately
23 to 25 C; however, micro-scale chambers can also be operated at elevated

FIG. 3 Comparison of acetaldehyde emissions from closed-cell SPF in micro- and

small-scale chambers using mass-specific emission factors.

30
Micro-Scale
Small-Scale
25
Emission Factor, µg/kg h

20

15

10
11 % RPD
9 % RPD
5

0
0 100 200 300 400 500
Hours in Chambers
 SEBROSKI ET AL., DOI 10.1520/STP158920150039 15

temperatures to simulate warm environments or to enhance the detection of chemi-

cals. There are no standardized methods to our knowledge to measure emissions of
VOCs and SVOCs, specifically for SPF; however, there are test methods that specify
elevated temperatures when measuring chemical emissions from polyurethane
foams. For example, ULC-S774-09 directs users to test polyurethane foam samples
for VOC emissions at 40 C, and ISO 12219-2-2012 specifies 65 C for screening
VOC emissions from vehicle interior parts and materials (which may contain poly-
urethane foam) [13,14].
The emission factors are reported as lg/m2 per hour for open-cell SPF at three
temperatures in Table 9 and Table 10. The data from this study show that temperature
does have an effect on chemical emissions of SPF insulation; however, we observed
unexpected fluctuations in the decay of emissions during the first few days of chamber
testing. As shown in Fig. 4, at 40 and 65 C, BDMAEE emissions from generic open-
cell SPF were greatest during the day the samples were loaded in the chambers (2 h).

TABLE 9 Generic open-cell SPF in micro-scale chambers: temperature and skin-trimming evaluations.

Hours in Chamber
Skin Temp. Chamber
Compound Presenta ( C) Blank 2 24 48 120 168 336 480

TMAEEA yes 23 <74.4 <893 <446 <446 <223 <223 <99.2 <74.4
no <893 <446 <446 <223 <223 <99.2 <74.4
yes 40 <74.4 <1,785 <1,785 <1,785 <893 <893 <595 <298
no <1,785 <1,785 <1,785 <893 <893 <595 <298
yes 65 <74.4 <8,925 <8,925 <8,925 <4,463 <4,463 <1,785 <893
no <8,925 <8,925 <8,925 <4,463 <4,463 <1,785 <893
BDMAEE yes 23 <5.00 4,443 2,740 2,461 1,353 998 607 391
no 2,985 1,957 1,948 929 689 395 244
yes 40 <5.00 15,453 7,807 250 1,383 590 231 88.2
no 13,955 5,163 270 593 328 123 58.0
yes 65 <5.00 76,714 2,534 20,702 619 <298 255 142
no 65,245 1,273 13,006 840 664 269 <59.6
TCPP yes 23 <4.00 <48.5 33.8 38.6 30.8 34.8 34.3 37.7
no <48.5 <24.2 25.0 19.0 23.0 25.7 25.2
yes 40 <4.00 374 354 1,494 359 357 416 390
no 492 284 2,188 290 273 375 321
yes 65 <4.00 5,972 7,509 1,873 4,868 3,866 4,763 3,290
no 6,019 4,752 1,440 6,028 3,629 4,433 3,142

Note: BDMAEE ¼ bis(2-dimethylaminoethyl) ether; TCPP ¼ tris(1-chloro-2-propyl) phosphate;

TMAEEA ¼ N,N,N-trimethylaminoethylethanolamine. Emission factors are reported as lg/m2
per hour. Emission factor reporting limits were calculated with the detection limit on the sorbent
tube/cartridge, sampling time/volume, and surface area of samples. The reporting limits shown with
a “less than” symbol vary depending on the sampling conditions.
a
The word “no” indicates that the top surface of the sample was trimmed; “yes” indicates that the
skin was left intact.
16 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 10 Generic open-cell SPF in micro-scale chambers: temperature and skin-trimming evaluations.

Hours in Chamber
Skin Temp. Chamber
Compound Presenta ( C) Blank 2 24 48 120 168 336 480

Formaldehyde yes 23 <2.26 3.72 3.69 2.99 <2.30 <1.88 <1.67 <2.11
no 3.28 3.03 <2.23 <2.30 1.96 <1.67 <2.11
yes 40 <2.26 7.98 6.44 4.37 5.33 4.07 2.30 2.62
no 9.62 6.20 4.64 3.81 2.26 2.20 <2.11
yes 65 <2.26 23.6 9.80 5.62 4.50 3.51 3.53 3.38
no 21.4 9.09 6.25 5.15 4.03 4.40 3.88
Acetaldehyde yes 23 <2.26 <2.73 <2.37 <2.23 <2.30 <1.88 <1.67 <2.11
no <2.73 <2.37 <2.23 <2.30 <1.88 <1.67 <2.11
yes 40 <2.26 5.14 <2.37 <2.23 2.48 2.00 3.30 3.71
no 4.16 <2.37 <2.23 <2.30 1.96 2.40 2.83
yes 65 <2.26 21.4 30.5 30.5 30.1 27.7 12.9 10.7
no 24.3 32.2 24.4 23.3 21.3 11.4 10.0
Propionaldehyde yes 23 <2.26 <2.73 <2.37 <2.23 <2.30 <1.88 <1.67 <2.11
no <2.73 <2.37 <2.23 <2.30 <1.88 <1.67 <2.11
yes 40 <2.26 <2.73 <2.37 <2.23 3.12 3.32 8.53 8.19
no 3.72 <2.37 <2.23 <2.30 3.39 8.50 7.38
yes 65 <2.26 115 130 93.7 49.6 39.9 20.0 11.7
no 74.9 120 96.8 64.5 55.0 21.4 17.7

Note: Emission factors are reported as lg/m2 per hour. Emission factor reporting limits were calculated
with the detection limit on the sorbent tube/cartridge, sampling time/volume, and surface area of sam-
ples. The reporting limits shown with a “less than” symbol vary depending on the sampling conditions.
a
The word “no” indicates that the top surface of the sample was trimmed; “yes” indicates that the
skin was left intact.

There appeared to be a relation between the temperature and unexplained fluctuations

in emissions during this time (Fig. 5). That is, greater fluctuations were observed at
higher temperatures. After a drop, surge, and another drop in emissions during days 2
(48 h) to 5 (120 h) at an elevated temperature, the emissions decayed during days 5
(120 h) to 20 (480 h), with no apparent up and down fluctuations in the decay pattern
of emissions. The reason for the apparent fluctuations in BDMAEE emissions factors
may be due to BDMAEE evaporating from the surface of the sample during the first
few days at an elevated temperature or other factors.
As shown in Fig. 6, TCPP emissions in generic open-cell SPF at an elevated
temperature increased during the first few days of testing followed by a sharp
decrease of emissions, and then a quasi-steady state appeared to be reached after
approximately 5 days (120 h). The initial spike and drop of emissions at elevated
temperatures may be due to the compound rapidly emitting from the surface of the
sample followed by the recovery time to reach equilibrium in the chamber or other
factors. Upon reaching quasi-steady-state conditions, the TCPP emission factors
were approximately 10 times higher at 60 C compared to 40 C and at 40 C
 SEBROSKI ET AL., DOI 10.1520/STP158920150039 17

FIG. 4 BDMAEE emissions in open-cell SPF at three temperatures in micro-scale

chambers.

1,00,000
23°C
40°C
65°C
10,000
Emission Factor, µg/m2 h

1,000

100

10

1
0 100 200 300 400 500
Time in Chamber, H

FIG. 5 BDMAEE emissions from open-cell SPF after 2 h in micro-scale chambers.

1,00,000
Emission Factor, µg/m2 h

y = 974.7e0.0675x
R² = 0.9986

10,000

1,000
0 10 20 30 40 50 60 70
Temperature,°C
18 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 6 TCPP emissions from open-cell SPF at three temperatures in micro-scale

chambers.

8,000
23°C
40°C
7,000
65°C
6,000
Emission Factor, µg/m2 h

5,000

4,000

3,000

2,000

1,000

0
0 100 200 300 400 500
Time in Chamber, H

compared to 23 C. This trend continued throughout the duration of the 20-day
study. The apparent relation between TCPP emissions from generic open-cell SPF
versus temperature are plotted in Fig. 7.

FIG. 7 TCPP emissions from open-cell SPF after 480 h in micro-scale chambers.

10,000.0

1,000.0
Emission Factor, µg/m2 h

y = 4.1541e0.1049x
R² = 0.9816

100.0

10.0

1.0
0 10 20 30 40 50 60 70
Temperature, °C
 SEBROSKI ET AL., DOI 10.1520/STP158920150039 19

Aldehyde emissions were higher at elevated temperatures; however, a tempera-

ture relation was not established. As shown in Fig. 8, formaldehyde emissions were
higher at 40 and 65 C; however, there were no apparent differences in emissions at
these elevated temperatures. In this case, formaldehyde emissions were not
detected, or the result was close to the detection limit at ambient temperature. Test-
ing at elevated temperatures could potentially be used to enhance the detection of
aldehydes, although a temperature relation was not established.

TEMPERATURE EVALUATION WITH CLOSED-CELL SPF

The emission factors for closed-cell SPF are reported in Table 11. As shown in Fig. 9, emis-
sions of the blowing agent HFC-245fa at elevated temperatures also showed a drop, surge,
and another drop pattern in the decay curve during the first few days in the chamber.
After the emissions stabilized, the compound continued to decay, approaching steady-
state conditions at 9 days (216 h) at 23 C; the compound continued to decay through
20 days (480 h) at 40 and 65 C. The relation between temperature and HFC-245fa emis-
sions appeared to be linear rather than exponential (Fig. 10). At 40 and 65 C, TCPP emis-
sions were above the reporting limit for closed-cell SPF at 2 h, and then the emission
factor dropped below the reporting limit for the duration of the 20-day study.
No emissions of formaldehyde or propionaldehyde were detected at 23 and
40 C; however, both of these compounds were observed at 65 C. As shown in Fig. 11,


acetaldehyde was detected at 23, 40, and 65 C, but after 72 h, there was no apparent

FIG. 8 Formaldehyde emissions from open-cell SPF at three temperatures in micro-

scale chambers.

25.00
23°C
40°C
65°C
20.00
Emission Factor, µg/m2 h

15.00

10.00

5.00

0.00
0 100 200 300 400 500
Time in Chamber, H
20 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 11 Generic closed-cell SPF in micro-scale chambers: temperature evaluations.

Hours in Chamber
Chamber

Target Compound Temp. ( C) Blank 2 24 48 72 168 216 336 480

HFC-245fa 23 <7.61 2,164 1,408 1,004 948 457 408 328 334
40 <7.61 3,337 1,146 1,659 1,628 798 732 522 418
65 <7.61 5,519 1,654 3,191 3,050 1,543 1,190 1,077 868
TMAEEA 23 <74.4 <149 <149 <149 <149 <74.4 <74.4 <74.4 <149
40 <74.4 <298 <298 <298 <149 <149 <149 <149 <149
65 <74.4 <893 <893 <595 <298 <298 <298 <298 <298
BDMAEE 23 <4.97 <9.94 <9.94 <9.94 <9.94 <4.97 <4.97 <4.97 <9.94
40 <4.97 <19.9 <19.9 <19.9 <9.94 <9.94 <9.94 <9.94 <9.94
65 <4.97 321 <59.6 46.2 36.4 <19.9 <19.9 <19.9 <19.9
DAPA 23 <34.5 <68.9 <68.9 <68.9 <68.9 <34.5 <34.5 <34.5 <68.9
40 <34.5 <138 <138 <138 <68.9 <68.9 <68.9 <68.9 <68.9
65 <34.5 <414 <414 <276 <138 <138 <139 <138 <138
TCPP 23 <4.04 13.1 36.0 17.2 8.94 4.19 <4.04 <4.04 <8.07
40 <4.04 143 <16.1 <16.1 <8.07 <8.07 <8.07 <8.07 <8.07
65 <4.04 188 <48.4 <32.3 <16.1 <16.1 <16.1 <16.1 <16.1
Formaldehyde 23 <2.10 <2.22 <1.90 <1.90 <1.85 <1.94 <1.90 <2.27 <1.67
40 <2.10 <2.22 <1.90 <1.90 <1.85 <1.94 <1.90 <2.27 <1.67
65 <2.10 10.6 11.6 8.98 9.76 2.87 5.71 4.81 4.67
Acetaldehyde 23 <2.10 30.0 15.8 8.91 8.50 5.51 4.57 3.72 3.07
40 <2.10 55.4 28.1 20.1 12.1 5.59 5.10 4.04 <1.67
65 <2.10 135 46.9 26.8 19.4 10.3 11.5 9.35 10.4
Propionaldehyde 23 <2.10 <2.22 5.40 <1.90 <1.85 <1.94 <1.90 <2.27 <1.67
40 <2.10 15.9 8.60 <1.90 <1.85 <1.94 <1.90 <2.27 <1.67
65 <2.10 74.0 25.4 12.7 12.5 4.43 3.96 <2.27 6.93

Note: BDMAEE ¼ bis(2-dimethylaminoethyl) ether; DAPA ¼ bis(dimethylaminopropyl) methylamine;

HFC-245fa ¼ 1,1,1,3,3-pentafluoropropane; TCPP ¼ tris(1-chloro-2-propyl) phosphate; TMAEEA ¼
N,N,N-trimethylaminoethylethanolamine. Emission factors are reported as lg/m2 per hour. Emission
factor reporting limits were calculated with the detection limit on the sorbent tube/cartridge, sampling
time/volume, and surface area of samples. The reporting limits shown with a “less than” symbol vary
depending on the sampling conditions.

difference in the emissions at 40 C compared to those measured at 23 C. At 65 C,

emissions of formaldehyde, propionaldehyde, and acetaldehyde were all observed
above the detection limits. These data suggest that operating the chambers at 65 C
may enhance the detection of potential aldehyde emissions.

OPEN-CELL SKIN TRIMMING

After application in a building, open-cell SPF insulation may be trimmed to align
with wall studs or other structural elements prior to installing wallboard. To test
the effect of trimming the surface on emissions, the generic open-cell SPF samples
were evaluated with and without the surface skin present at the same time the
 SEBROSKI ET AL., DOI 10.1520/STP158920150039 21

FIG. 9 HFC-245fa emissions in closed-cell SPF at three temperatures in micro-scale

chambers.

6,000
23°C
40°C
5,000
65°C
Emission Factor, µg/m2 h

4,000

3,000

2,000

1,000

0
0 100 200 300 400 500
Hours in Chamber

FIG. 10 HFC-245fa emissions in closed-cell SPF versus temperature after 216 h in

micro-scale chambers.

1,400

1,200

1,000
Emission Factor, µg/m2 h

y = 18.581x - 15.983
R² = 0.9999
800

600

400

200

0
0 10 20 30 40 50 60 70
Temperature, °C
22 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 11 Acetaldehyde emissions in closed-cell SPF at three temperatures in micro-scale

chambers.

140
23°C
40°C
120
65°C
100
Emission Factor, µg/m2 h

80

60

40

20

0
0 100 200 300 400 500
Hours in Chamber

FIG. 12 BDMAEE emissions at 23 C in generic open-cell SPF: skin-trimming comparison.

5,000
Skins
4,500
No Skins
4,000
Emission Factor, µg/m2 h

3,500

3,000

2,500

2,000
37 % RPD
1,500
37 % RPD 42 % RPD
1,000

500

0
0 100 200 300 400 500
Hours in Chamber
 SEBROSKI ET AL., DOI 10.1520/STP158920150039 23

FIG. 13 TCPP emissions in generic open-cell SPF at 40 C: skin-trimming comparison.

2,500
Skins
No Skins
2,000
Emission Factor, µg/m2 h

1,500

1,000

27 % RPD 10 % RPD
500

0
0 100 200 300 400 500
Hours in Chamber

FIG. 14 Acetaldehyde emissions at 65 C in generic open-cell SPF: skin-trimming

comparison.

25
Skins
No Skins
20
Emission Factor, µg/m2 h

15

10

25 % RPD
12 % RPD
5

0
0 100 200 300 400 500
Hours in Chamber
24 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

temperature evaluation study was performed at 23, 40, and 65 C. Emission results
for the open-cell SPF are reported in Table 9 and Table 10. As shown in Fig. 12, the
BDMAEE emissions were consistently higher with the skin intact and lower
with the skin removed. Similarly, as shown in Fig. 13, after the chambers reached
a quasi-steady state, TCPP emissions at elevated temperatures were also lower
with the skins removed, although there was a smaller difference. When cut to
the same dimensions, samples with the skin intact weighed more than samples
with the skin removed. Therefore, samples with skins or samples that were
sprayed with more than one pass during application (internal skins) have a
greater density. This may explain why BDMAEE and TCPP emissions were low-
er with the skins removed. It is also feasible that during the exothermic foaming
process, these compounds may have coalesced at the skin that was removed with
trimming. Conversely, the aldehyde emissions did not show any apparent differ-
ence with the skins removed (Fig. 14), possibly because of the higher vapor pres-
sure of these compounds.

Conclusions
Micro-scale chambers may be useful in evaluating SPF insulation for chemical
emissions of VOCs under controlled conditions. The higher air exchange rate and
loading factor of the micro-scale chambers produced higher emissions of TCPP, an
SVOC. In general, micro-chamber emissions tend to be within an order of magni-
tude compared to small-scale chamber data. Some emission factors were higher for
small-scale chambers (acetaldehyde for closed-cell SPF), and some were lower in
small-scale chambers (TCPP). The aldehyde emissions may be dependent on the
mass of the sample; therefore, mass-specific emission factors should be reported in
addition to area-specific emission factors.
The data from this study demonstrate that SPF samples can be tested in micro-
scale chambers at temperatures ranging from 23 to 65 C; however, additional re-
search is needed to evaluate the relation between sample collection times and the
variability of emission factors of some compounds during the first 5 to 7 days of the
tests. In order to show the decay pattern, several samples may need to be collected
from the chambers over various times; otherwise, careful attention to the time of air
sample collection may be necessary to avoid significant spikes and drops of emis-
sions of some compounds prior to reaching a quasi-steady-state condition in the
chambers. At the time of this research, the draft ASTM method specified collecting
samples from micro-scale chambers after 2 and 24 h; this time may need to be ex-
tended to reach a quasi-steady state and to evaluate emissions from SVOCs such as
TCPP. VOC emissions may initially be emitted at a higher rate when tested at
elevated temperatures; however, emission rates may decay faster because of the
depletion of the compound in the source. Elevated temperatures may be useful in
enhancing the detection of potential aldehyde or other emissions that are not
detected at ambient temperature.
 SEBROSKI ET AL., DOI 10.1520/STP158920150039 25

BDMAEE emissions were reduced when the skin of open-cell SPF was removed to
simulate trimming of the insulation against wall studs or other structural elements in a
building. There was no apparent difference of aldehyde or flame-retardant emissions
when the skin was removed. When the surface skin was removed from the samples,
the density of the samples decreased, which may have reduced the emissions of the
catalyst. Furthermore, the catalyst may have coalesced at the skin during the foaming
process that was removed with trimming. The aldehydes may readily diffuse through
the material regardless of the presence of the surface skin.
TMIBPA showed poor precision (29.7 % RSD) and a negative bias (36.7 %)
compared to other compounds that were evaluated with the TD-GC-MS method.
Although the amine catalyst was not detected in any of the samples, this compound
was not reported during this study because of its poor analytical performance.
Further method development may be necessary for challenging compounds such as
reactive amine catalysts. These reactive compounds can be used in the formulations
to reduce emissions; however, they can be difficult to quantitate with existing
methodology. The ASTM D22.05 task groups that are developing standards to mea-
sure emissions from SPF may need to develop additional analytical methods to
quantify these challenging compounds.
To improve the data capture rate (less nondetects), a longer sample duration
may be necessary when collecting emissions onto the TD tubes. Pretesting samples
can be useful in establishing the scale for sampling times and volumes to improve
the quality of the data. Sampling very volatile compounds onto separate TD tubes
containing stronger sorbents (e.g., Carboxen) helps to avoid breakthrough of the
blowing agent and allows the less volatile compounds (e.g., amine catalysts and
flame retardants) to be sampled longer without exceeding maximum sampling
volumes. Future research should examine extending the maximum sampling times
and volumes to optimize the detection limits of target compounds.

ACKNOWLEDGMENTS
The authors thank Ken Riddle, Tim Takah, Scott Ecoff, Shen Tian, Brian Karlovich,
Richard Romero, and Eloy Martinez at Covestro LLC; Todd Wishneski at BASF Cor-
poration; CPI SPF Emissions Task Force; and ASTM International Subcommittee
D22.05 on Indoor Air.

References

[1] Sebroski, J. R., “Research Report for Measuring Emissions from Spray Polyurethane
Foam (SPF) Insulation,” http://polyurethane.americanchemistry.com/Resources-and-
Document-Library/Research-Report-for-Measuring-Emissions-from-Spray-Polyurethane-
Foam-SPF-Insulation.pdf (accessed November 7, 2016).

[2] ASTM D7859-13e1, Standard Practice for Spraying, Sampling, Packaging, and Test
Specimen Preparation of Spray Polyurethane Foam (SPF) Insulation for Testing of
26 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

Emissions Using Environmental Chambers, ASTM International, West Conshohocken, PA,

2013, www.astm.org

[3] ASTM D5116-10, Standard Guide for Small-Scale Environmental Chamber Determinations
of Organic Emissions from Indoor Materials/Products, ASTM International, West Consho-
hocken, PA, 2010, www.astm.org

[4] ASTM D7706-11, Standard Practice for Rapid Screening of VOC Emissions from Products
Using Micro-Scale Chambers, ASTM International, West Conshohocken, PA, 2011,
www.astm.org

[5] ASTM D6670-13, Standard Practice for Full-Scale Chamber Determination of Volatile
Organic Emissions from Indoor Materials/Products, ASTM International, West Conshohocken,
PA, 2013, www.astm.org
R
[6] Sebroski, J. R., Miller, J., and Spence, M., “Evaluation of Modified FLECV Cell and Micro
Chamber Prototype for Monitoring Methylene Diphenyl Diisocyanate (MDI) Emissions,”
presented at the 13th International Conference on Indoor Air Quality and Climate, Hong
Kong, July 6–12, 2014.

[7] Sebroski, J. R., “Method Development and Research for Measuring Emissions from Spray
Polyurethane Foam (SPF) Insulation,” presented at the 2013 Polyurethanes Technical
Conference, Phoenix, AZ, September 23–25, 2013.

[8] ASTM D5197-09e1, Standard Test Method for Determination of Formaldehyde and Other
Carbonyl Compounds in Air (Active Sampler Methodology), ASTM International, West
Conshohocken, PA, 2009, www.astm.org

[9] U.S. EPA, “Compendium Method TO-17, Determination of Volatile Organic Compounds
in Ambient Air Using Active Sampling onto Sorbent Tubes,” Compendium of Methods
for the Determination of Toxic Organic Compounds in Ambient Air, 2nd ed., U.S. EPA,
Cincinnati, OH, 1999.

[10] ISO 16000-6:2011, Indoor Air—Part 6: Determination of Volatile Organic Compounds in

Indoor and Test Chamber Air by Active Sampling on Tenax TA Sorbent, Thermal Desorp-
tion and Gas Chromatography Using MS or MS-FID, International Organization for Stan-
dardization, Geneva, Switzerland.

[11] Ozbas, B., Vincent, J. L., Tobias, J. D., Rogers, J. R., and Burdeniuc, J. J., “Low- or Non-
Emissive Amine Catalysts for Low Density, Open-Cell Spray Polyurethane Foam,” 2012
Polyurethanes Technical Conference, Atlanta, GA, September 24–26, 2012.

[12] Ni, Y., Kumagai, K., and Yanagisawa, Y., “Measuring Emissions of Organophosphate
Flame Retardants Using a Passive Flux Sampler,” Atmos. Environ., Vol. 41, No. 15, 2007,
pp. 3235–3240.

[13] CAN/ULC-S774-09, Standard Laboratory Guide for the Determination of Volatile

Organic Compound Emissions from Polyurethane Foam, Underwriters’ Laboratories of
Canada, Richmond, BC, 2009.

[14] ISO 12219-2:2012, Interior Air of Road Vehicles—Part 2: Screening Method for the Deter-
mination of the Emissions of Volatile Organic Compounds from Vehicle Interior Parts
and Materials—Bag Method, International Organization for Standardization, Geneva,
Switzerland, 2012.
 DEVELOPING CONSENSUS STANDARDS FOR MEASURING CHEMICAL EMISSIONS 27

STP 1589, 2017 / available online at www.astm.org / doi: 10.1520/STP158920150040

Yunyun Nie,1 Eike Kleine-Benne,1 and Kurt Thaxton1

Measurement of Chemical
Emissions from Spray
Polyurethane Foam Insulation
Using an Automated Microscale
Chamber System
Citation
Nie, Y., Kleine-Benne, E., and Thaxton, K., “Measurement of Chemical Emissions from Spray
Polyurethane Foam Insulation Using an Automated Microscale Chamber System,” Developing
Consensus Standards for Measuring Chemical Emissions from Spray Polyurethane Foam (SPF)
Insulation, ASTM STP1589, J. Sebroski and M. Mason, Eds., ASTM International, West
Conshohocken, PA, 2017, pp. 27–42, http://dx.doi.org/10.1520/STP1589201500402

ABSTRACT
Off-line microscale chamber measurements of spray polyurethane foam require
settings to be changed manually and an analyst to start and stop sampling
times. An automated dynamic headspace system was used for on-line, fully
automated microscale chamber measurements of SPF. In this way, precise
measurement and computerized control of sampling times, volumes, and
temperature would be possible. After the system was qualified analytically,
timed measurements were carried out automatically over a period of
approximately 15 h to understand off-gassing mechanisms of the foams. In a
separate experiment, the effect of sample temperature was probed using an
automated sequence in which temperatures were changed at defined times to
simulate the heating of a foam sample in an attic or ceiling. Both experiments
provided unattended and precise control and measurement of the three principle
variables in microscale chamber experiments: time, sampling volume, and
temperature. The system should prove useful for mathematical modeling and for
sharing procedures between researchers.

Manuscript received May 15, 2015; accepted for publication November 24, 2015.
1
GERSTEL GmbH & Co. KG, Eberhard-Gerstel-Platz 1, 45473 Mülheim an der Ruhr, Germany
2
ASTM Symposium on Developing Consensus Standards for Measuring Chemical Emissions from Spray
Polyurethane Foam (SPF) Insulation on April 30–May 1, 2015 in Anaheim, CA.

Copyright V
C 2017 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959.
28 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

Keywords
spray polyurethane foam, SPF, microscale chamber, thermal desorption, gas
chromatography, mass spectrometry, emission factors, temperature, mass transfer

Introduction
Sprayed polyurethane foam (SPF) insulation is widely used in new building con-
struction and the renovation of existing structures. Potentially harmful effects from
emitted chemical components such as amine catalysts [1] and flame retardants [2]
are a focus of current research. In addition to exposure to occupants, determining
the emission behavior of SPF insulation can be used by SPF manufacturers to estab-
lish safe reentry and reoccupancy times for both workers and occupants [3].
Emission test chamber methods are widely used for determining the emission of
compounds from building products [4]; air is sampled through tubes (typically) con-
taining a sorbent, and the tubes are then analyzed using thermal desorption (TD) and
gas chromatography-mass spectrometry (GC-MS) [5]. However, poor analytical re-
covery due to significant sink effects of high-boiling-point compounds (flame retard-
ants, amine catalysts) has been observed when using conventional test chambers
[3,6]. For this reason, further development of consensus standards for measuring
chemical emissions from SPF samples in smaller microscale chambers is required be-
cause sink effects are greatly diminished in the much smaller chambers [6].
Like conventional chambers, common microscale chambers used for SPF emis-
sion measurements are off-line devices that are separate from analytical instrumen-
tation (although the data generated are comparative and not used to estimate or
predict emissions in actual buildings). As a result, there are limitations to this
approach for some types of experiments in which many multiple measurements or
changes in conditions are desired. An analyst or technician must manually set tem-
peratures and flows as well as remove and replace the thermal desorption tubes
(which effectively defines the sampling time).
Microscale chamber measurements can also be thought of as dynamic headspace
(DHS) measurements. In both cases, gas is pumped into a known volume (which
may or may not be heated) and for a known time. The same gas passes through a sor-
bent trap as it leaves that volume. Compounds entrained in the gas are concentrated
on the sorbent trap, which is subsequently analyzed by TD-GC-MS.
In this work, a large-volume dynamic headspace autosampler was used as a fully
automated on-line microscale chamber analysis system for the analysis of SPF. As with
stand-alone systems, the microscale chambers used have similar or improved analytical
performance. However, on-line microscale chamber measurements permit automated
sampling of SPF under a computerized method control of flows, temperatures, and
sampling duration/time. In this way, emission behavior versus time is easily but also
more precisely and reliably obtained, and experiments at different flows or tempera-
tures (e.g., to simulate different installation conditions) are also greatly simplified by
automation and precise measurements of temperatures and volumes sampled.
 NIE ET AL., DOI 10.1520/STP158920150040 29

Experimental
CHEMICALS
Target compounds are listed in Table 1 with names, acronyms, CAS numbers,
and boiling points. Tetramethyliminobispropylamine (TMIBPA), bis(2-dimethyla-
minoethyl)ether (BDMAEE), pentamethyldiethylenetriamine (PMDTA), n,n,n-
trimethylaminoethylethanolamine (TMAEEA), and toluene were in GC-MS grade
and purchased from Sigma-Aldrich. Bis(dimethlyaminopropyl)methylamine (DAPA)
and tris-(1-chloro-2-propyl)phosphate (TCPP) were purchased from abcr GmbH in
GC-MS grade.

GENERIC SPF INSULATION SAMPLES

Two SPF insulation samples were obtained from the Center for the Polyurethanes
Industry, which is a part of the American Chemistry Council. One sample type was an
open-cell foam; the other was a closed-cell foam. Both samples were packaged and
shipped according to ASTM D7859-13e1, Standard Practice for Spraying, Sampling,
Packaging, and Test Specimen Preparation of Spray Polyurethane Foam (SPF) Insula-
tion for Testing of Emissions Using Environmental Chambers [7]. The foams were about
50 by 50 cm and approximately 10-cm thick. They were wrapped in two sheets of clean
R
aluminum foil and sealed in MylarV bags with a zipper seal. Prior to testing, the sam-
R
ples were stored in unopened MylarV bags in the laboratory. The laboratory tempera-
ture was controlled at 23 C, and no direct light contacted the samples. The samples
were opened directly before testing. After testing they were wrapped again and sealed
R
again in MylarV bags. Using a custom-made coring tool, SPF samples 92 mm in diame-
ter and of varying thicknesses (3, 5, and 8 cm) were prepared. All samples were cored
from the surface to the bottom and cut from the bottom to obtain the required thick-
ness (i.e., the top surface was kept on the samples). The cut was started on the top side
to make sure the top surface of the samples did not crack during coring. After coring,
the samples were pushed out of the tool and reinserted in reverse into the coring tool.
This was done to position the sample surface at the top of the tool.

TABLE 1 Target analytes.

Boiling Point
Target Compound Acronym CAS No. ( C at 1 atm)

1,1,1,3,3-Pentafluoropropane HFC-245fa 460-73-1 15

Bis(dimethlyaminopropyl)methylamine DAPA 3855-32-1 102
Tetramethyliminobispropylamine TMIBPA 6711-48-4 128-131
Bis(2-Dimethylaminoethyl)ether BDMAEE 3033-62-3 189
Pentamethyldiethylenetriamine PMDTA 3030-47-5 198
N,N,N-Trimethylaminoethylethanolamine TMAEEA 2212-32-0 207
Tris-(1-chloro-2-propyl)phosphate TCPP 13674-84-5 244
Toluene 108-88-3 111
30 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

According to ASTM D7859-13e1, the sample should be tested within 48 h from

the time the sample was sprayed. This was not possible for this work because the
shipping time was over 2 days and method development took a few days beyond
that. Therefore, the emission factors obtained herein should not be compared to
those from “fresh” samples. The aim of this work was to develop automated micro-
scale chamber methods for SPF foam; future work will be carried out with fresh
samples.

INSTRUMENTATION
Microscale Chambers
An Automated Dynamic Headspace DHSlarge (DHS-L) autosampler together with DHS
vessels were used as a microscale chamber system (GERSTEL GmbH & Co. KG). The
vessels (microscale chambers) are made of electropolished stainless steel and coated
with an inert coating. Three sizes of chambers are possible: 250, 500, and 1000 mL. All
have the same circular cross-section, use the same lids, and vary only in depth. The
chambers are air tight and have very low blank values and greatly diminished sink
effects (more in “Background and Sink Effects”). Gas flows into the chambers are con-
trolled via a mass flow controller (5–100 mL/min; accuracy, 62 %). Outlet gas flows are
also recorded to verify chamber integrity and to measure the actual sampling volume.
Different types of inlet gases are possible: synthetic air, nitrogen, helium, and so on. Dry
synthetic air was used for this work. The temperature in microscale chambers can be set
from room temperature (23 C) to 200 C with an accuracy of 61.5 C. A multipurpose
sampler (GERSTEL) enables automated air sampling from microscale chambers with
sorbent tubes at user-selected time intervals followed by TD-GC-MS analysis.
Chamber parameters used for SPF insulations were as follows: inlet air flow rate,
50-mL/min; chamber temperature, 23, 40, or 65 C (elevated temperatures were used to
simulate field conditions such as a warm attic in the summer). Samples were measured
at 23 C every 1 h for 12 to 24 h (12–24 sampling points). The chamber temperature
was then increased via method-driven computer control to 40 and 65 C for three
measurements, respectively. Between each analytical measurement, a nonsampled
purge flow (50 mL/min) was applied. Sampling flow was also 50 mL/min with a sam-
pled volume of 200 mL for open-cell SPF and 1,000 mL for closed-cell SPF. Different
volumes were used to accommodate the different emission levels between two samples.
R
In all cases, 80 mg of TenaxV TA (Buchem BV) was used as the sorbent.
DHS-L vessels together with the coring tool inserted were conditioned at 150 C
for 2 h in the oven before use. Normally after measurements are made vessels are
washed with diluted commercial cleaning solution and then thermally conditioned
at different temperatures according to the application. Common standard methods
require that either background levels for volatile organic compounds (VOCs) of in-
terest should be 10 % or less of the lowest levels of interest for actual samples [5,8],
or total VOC (TVOC) values should be less than 20 ng/L and individual VOC con-
centrations less than 2 ng/L [4]. Both requirements were met with cleaning and
conditioning.
 NIE ET AL., DOI 10.1520/STP158920150040 31

TD-GC-MS
R
After sampling, the TenaxV TA sorbent was thermally desorbed using a thermal de-
sorption unit (TDU) (GERSTEL). TD was performed with ultrahigh-purity helium
(purity 6.0) carrier gas at a flow rate of 51.5 mL/min in splitless mode. The temperature
program was as follows: initial temperature 20 C for 0.20 min; ramp at 720 C per
minute to 270 C; hold for 8 min. The TDU transfer heater was set at 280 C. A pro-
grammed temperature vaporization-type cooled injection system (CIS 4; GERSTEL)
was used as a concentrator trap for the TDU. The trap temperature was set at 40 C
during primary desorption and heated for analyte transfer at 12 C per second to 275 C
with a hold time of 5 min for injection onto the GC column. Compounds were separat-
ed by a gas chromatograph (6890 GC; Agilent) and detected using a mass spectrometric
detector (5973 MSD; Agilent). The GC was provisioned with a Restek Rtx-5 amine
column (length, 30 m; inner diameter, 0.25 mm; film thickness, 0.5 lm). The following
GC oven temperature program was used: initial temperature 40 C for 2 min; ramp at
20 C per minute to 280 C; final hold for 4 min (total run time of 170 min). Mass
spectrometric detection temperatures were set to 230 C (source) and 150 C (quad).
MS spectra were acquired in scan mode in a range from 40 to 550. A constant column
flow of 1.5 mL/min was used. The overall split ratio was 34:1.

CALCULATION OF EMISSION FACTORS

Emission factors of target compounds were quantitatively determined using calibra-
tion curves and emission equations Eq 1 and Eq 2. Emission factors for TVOC and
the blowing agent 1,1,1,3,3-pentafluoropropane (HFC-245fa) were estimated using
toluene equivalents according to DIN EN ISO 16000-6 and Eq 1 and Eq 2.
The results of the microchamber tests are reported as specific emission factors
(EFs) normalized to the exposed surface area of the specimen according to ISO
standards [4] using Eq 1:

EF ¼ qx  q (1)

where:
qx, concentration of the analyte in test chamber, lg m3, as determined from
the amount of analyte (lg) divided by the air sample volume, m3, and
q, area-specific air flow rate (m3 m2 h1) calculated using Eq 2:
n
q¼ (2)
L

where:
n, air change rate in test chamber (h1), and
L, sample loading factor (m2 m3) obtained by dividing the sample surface
(m2) area A with the chamber volume above the sample (m3) V.
All of the parameters specific to the DHS-L microscale test chamber are listed in
Table 2.
32 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 2 Parameter values for DHS-L microscale chambers.

Parameter Description Value

2
A Sample surface area (m ) 6.65 E-3
V Chamber volume above the sample (m3) 1.05E-4
n Air exchange rate (h1) 28.6
Q Area-specific air flow rate (m3 m2 h1) 0.45

QUANTITATION AND SEMIQUANTITATION

R
Target compounds detected in TenaxV TA sorbent were identified and quantified
through standard calibration solutions using Agilent MassHunter Qualitative Anal-
ysis (B.07.02). Calibration solutions with different concentration levels were spiked
R
into tubes containing TenaxV TA sorbent and then thermally desorbed into the
GC-MS system. Because of chromatographic coelution between BDMAEE and
TMAEEA and between DAPA and TMIPBA, two sets of calibration standards were
prepared: Standard 1 (BDMAEE, PMDTA, DAPA, and TCPP) and Standard 2
(TMIBPA and TMAEEA). Each calibration level, in either set, was run three times.
Quantifier/qualifier MS ions, linearity, and calibration ranges are listed in Table 3.
Good linearity was shown for all target compounds within the calibration range
and had correlation coefficients ranging from 0.986 to 0.997.
For nontargeted VOC compounds, an 80 % or better match was necessary for tenta-
tive identification using National Institute of Standards and Technology spectra library
Version 2.2. Any peak that was above the baseline, had a reasonable chromatographic
peak shape, and met the library match minimum percentage was reported.

BACKGROUND AND SINK EFFECTS

In this application, the lowest-level standard is 25 ng. Compared to this level, the
background of a DHS-L vessel together with a coring tool is very low, with no
direct interference with target compounds (Fig. 1). The toluene equivalent TVOC
value of the background was 9.5 ng/L, which was lower than the 20-ng/L limit.

TABLE 3 MS quantifier and qualifier ions, linearity, and calibration ranges for target compounds.

Target Compound MS Ion Linear Regression (R2) Calibration Range (ng)

BDMAEE 58/2071/42 0.997 25–1,000

PMDTA 72/58/115 0.996 25–500
DAPA 58/85/70 0.989 25–500
TCPP 125/99/157 0.986 25–1,000
TMIBPA 58/85/70 0.992 200–2,500
TMAEEA 88/58/44 0.997 200–5,000

Note: BDMAEE ¼ bis(2-dimethylaminoethyl)ether; DAPA ¼ Bis(dimethlyaminopropyl)methylamine;

PMDTA ¼ pentamethyldiethylenetriamine; TCPP ¼ tris-(1-chloro-2-propyl)phosphate; TMAEEA ¼
n,n,n-trimethylaminoethylethanolamine; TMIBPA ¼ Tetramethyliminobispropylamine.
 NIE ET AL., DOI 10.1520/STP158920150040 33

FIG. 1 Lowest level of SPF standard (25 ng per TD tube; top) compared with the
background of DHS-L 1-L vessel together and coring tool (65 C; 1-L sampling
volume).

The background peaks observed correspond to low-molecular-weight siloxanes,

some of which could originate as a column or septum bleed, or both.
“Sink effect” is a term that denotes the ability of a test chamber system to
adsorb and desorb a substance during the measurement, possibly leading to changes
in the concentration of the substances inside the chamber. The sink effect can be
related to the materials used to construct the chamber and is also related to the boil-
ing point and vapor pressure of substances measured [6]. This means very volatile
compounds have no/very low sink effects, whereas less volatile compounds show
more sink effects that might be also affected by the construction of the chamber.
A high sink effect is observed for amine catalysts and flame retardants in SPF
insulation measurements using conventional emission chambers [3]. Amine cata-
lysts are highly active polar basic compounds, which are adsorbed by the chamber
surface through chemical interactions, whereas flame retardants are more inert but
also higher-boiling-point compounds. The sink effect for amine catalysts can be
minimized using stainless steel chamber materials smoothed through electropolish-
ing and coated with inert materials [3].
DHS-L vessels are made of stainless steel, electropolished, and coated with ei-
ther a commercial (Sulfinert) coating or a proprietary inert coating; the former was
used for the data presented herein. Because of the commonality of these steps with
sink-effect countermeasures employed by others, the sink effects herein were as-
sumed to be minimized.

Results and Discussion

EFFECTS OF SAMPLE THICKNESS
The chambers used herein were constructed of varying depths (3, 5, and 8 cm) to
achieve varying volumes. (The lids were the same, so the sample surface area was
34 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

constant regardless of volume.) According to ASTM D7859-13e1 [7], the SPF specimen
should be cut between 8 and 9 cm in thickness to simulate the typical thickness of the
product type when installed in buildings. This is done because the sample thickness
will influence the emission factors due to internal diffusion within the sample.
However, those samples are usually cut again at the time of analysis to fit smaller
microscale chambers that typically accommodate samples of approximately 3 cm in
length/depth. The longer two depths (5 and 8 cm) have not been used with SPF sam-
ples before.
On this system, sample length is a variable that can be chosen, possibly to more
closely match the longer lengths of SPF used at installations. Consequently, mea-
surements of longer-length samples were needed to understand the effect of sample
length (depth). Open-cell SPF samples of different thicknesses (3, 5, and 8 cm) were
prepared. The top-surface “skin” was always maintained and was positioned at the
top (sampling end) of the microscale chamber. Open-cell foam was chosen for this
experiment because closed-cell foam is known to have a stronger barrier at the sur-
face that should make the effect less observable.
The results are shown in Fig. 2. At 23 C, BDMAEE and TCPP emission factors
increased with the sample thickness. Higher-boiling-point/high-molecular-weight
compounds, such as TCPP, show smaller enhancements than lower-boiling-point/low-
molecular-weight compounds, such as BDMAEE. Emission factors of TMIBPA were
also compared—but only at 65 C. (TMIBPA is not easily observable at lower tempera-
tures.) TMIBPA also showed the same increasing trend with sample thickness. In-
creased sample thickness provides more signal and analysis sensitivity.
Correlation coefficients obtained from linear regression of emission factors ver-
sus sample thickness were >0.99, indicating a linear relation between sample thick-
ness and emission values for open-cell samples. For open-cell foam, an 8-cm length
would more closely approximate emissions at a typical installation; however, the
linear relation of signal versus foam length/depth makes measurements at 3 or 5 cm

FIG. 2 Emission factor (lg/m2h) of three open-cell samples with three different
thicknesses but the same surface area at 23 or 65 C.

4,000
3,500
Average Emission Factor

3,000
2,500
[µg/m2.h]

2,000 3 cm

1,500 5 cm
1,000 8 cm
500
0
BDMAEE (23°C) TMIBPA (65°C) TCPP (23°C)
 NIE ET AL., DOI 10.1520/STP158920150040 35

relatable to longer lengths (and to measurements taken with other microscale

chambers at approximately 3 cm in length).

VOCS AND SVOCS EMITTED FROM OPEN- VERSUS CLOSED-CELL SPF FOAM
Open- and closed-cell foams were analyzed repeatedly for an approximately 15-h time
study (see “Emission Factors Versus Time”). Chromatograms of the first data point in
the series for open- and closed-cell foams are shown in Fig. 3 and Fig. 4, respectively.
The primary VOCs and semivolatile VOCs (SVOCs) emitted by the open-cell
SPF samples at 23 C were BDMAEE, tetramethylpropanediamine, and 2,2,4,6,6-
pentamethylheptane (Fig. 3; target analytes noted alphabetically). Other compounds
noted in Fig. 3 were also found but at much lower emission factors. The flame retar-
dant TCPP and amine catalyst TMIBPA were also detected at 23 C. Neither
TMIBPA nor TCPP was detected at 23 C in previous work, perhaps because of the
smaller sample volumes used [3].
The primary VOCs emitted by closed-cell SPF samples at 23 C were trans-1,2-
dichloroethylene and other similar chlorinated alkenes or alkanes, such as 1-chlroro-1-
propene, 3-chloro-1-propene, and 1,2-dichloropropane (Fig. 4). Other VOCs, SVOC
amine catalysts BDMAEE and DAPA, and flame retardants show much lower
emissions. The blowing agent HFC-245fa was also detected in the closed-cell sample
because it was not easily released from closed-cell foams. Low-molecular-weight

FIG. 3 Total ion chromatograms (top) and enlarged view (bottom) obtained from
the open-cell SPF sample at 23 C. Sampling volume, 200 mL. 1: trans-1,2-
dichloroethylene; 2: 1,4-dioxane; 3: chlorobenzene; 4: N,N0 -dimethylpiperazine;
5: tetramethylpropanediamine; 6: 2,2,4,6,6-pentamethylheptane. A: DMAEE;
B: TMIBPA; C: TCPP. Si, siloxanes.
36 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 4 Total ion chromatograms (top) and enlarged view (bottom) obtained from the
closed-cell SPF sample at 23 C. Sampling volume, 1,000 mL. 1: propylene oxide;
2: 1-chloro-1-propene; 3: 3-chloro-1-propene; 4: trans-1,2-dichloroethylene;
5: 1,2-dichloropropane; 6: 1,4-dioxane; 7: 2-methyl-1,3,6-trioxocane;
8: 4-methylmorpholine; 9: chlorobenzene; 10: N,N0 -dimethylpiperazine; 11: nonanal.
A: HFC-245fa; B: BDMAEE; C: DAPA; D: TCPP. Si, siloxanes.

siloxanes were found and are typical of emissions due to the presence of silicon
stabilizers [9].
Table 4 lists species detected in both SPF sample types at 23 C. The VOCs iden-
tified in the closed-cell foam differ substantially from the open-cell foam, although
compounds found in both foam types can be emitted into indoor air [10]. Chlori-
nated VOCs (e.g., 1,2-dichloropropane) are thought to originate from flame retard-
ants and their impurities [11]. N,N0 -Dimethylpiperazine is a thermal degradation
product of TMAEEA. This amine catalyst was not detected in either sample, possi-
bly due to its thermal liability. Detecting these target compounds at low levels is im-
portant because amine catalysts used to produce SPF insulation may have harmful
effects [1], and chlorinated alkanes and alcohols are known hazards. (1,2-Dichloro-
propane is classified as a possible carcinogen [11].) Tertiary amine catalysts have
unique odors with typically parts per billion odor thresholds [12], so low-level emis-
sions of amines may lead to complaints about unpleasant odors.
Overall, the closed-cell SPF sample delivered lower emissions than the open-
cell sample at the same temperature. In the closed-cell sample, the emission of the
 NIE ET AL., DOI 10.1520/STP158920150040 37

TABLE 4 Compounds and emission factors (lg/m2h) observed in open- and closed-cell SPF sam-
ples at 23 C at time point 0.

Compound Retention Time (min) Open-Cell EF Closed-Cell EF

HFC-245fa 1.654 12
Propylene oxide 2.096 8
3-Chloro-1-propene 2.207 44
1-Chloro-1-propene 2.253 69
trans-1,2-Dichloroethylene 2.530 7 602
1,2-Dichloropropane 3.843 43
1,4-Dioxane 3.936 4 27
2-Methyl-1,3,6-trioxocane 4.489 9
4-Methylmorpholine 4.821 3
Chlorobenzene 5.522 2 0.4
N,N0 -Dimethylpiperazine 5.557 4 0.6
Tetramethylpropanediamine 6.017 72
2,2,4,6,6-Pentamethylheptane 6.885 49
BDMAEEa 7.796 3,203 19
Nonanal 7.848 0.2
DAPAa 9.630 27
TMIBPAa 10.509 NQ
TCPPa 12.561 190 71

Note: BDMAEE ¼ bis(2-dimethylaminoethyl)ether; DAPA ¼ Bis(dimethlyaminopropyl)methyl-

amine; NQ ¼ not quantitated (the calculated concentration fell below the lowest calibration point);
TCPP ¼ tris-(1-chloro-2-propyl)phosphate; TMIBPA ¼ Tetramethyliminobispropylamine.
a
Estimated using a calibration curve; all others estimated using toluene equivalence.

amine catalyst BDMAEE reached a value of 16 lg/m2h, much lower than in the
open-cell sample (3,381 lg/m2h). In addition, for TCPP, closed-cell emission fac-
tors were only approximately 30 to 40 % of the open-cell values.

EMISSION FACTORS VERSUS TIME

Up to this point the focus of this discussion has been on looking at the performance
of the chambers rather than aspects of the overall automated system; repeated
measurements over time are one use of such a system. Both open- and closed-cell
SPF samples were measured repeatedly over time to observe changes in emission
factors. The results can be found in Table 5 and Fig. 5.
The open-cell SPF sample showed relatively constant emissions over 12 h, with
a 3.0 % relative standard deviation (RSD) for BDMAEE and 2.5 % RSD for TCPP.
For closed-cell samples, the emission factors of HFC-245fa and TCPP fell over
time. After approximately 8 h, the emissions of HFC-245fa reached approximately
stable values. However, the flame retardant TCPP still did not reach a constant
emission value within 12 h, possibly due to a lower diffusion coefficient and higher
molecular weight/boiling point. BDMAEE and DAPA had more or less constant
38 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 5 Emission factors (lg/m2h) of target compounds emitted from SPF insulation samples
over approximately 15 h obtained at 23 C.

Open Cell Closed Cell

Time (s) BDMAEE TCPP Time (s) HFC-245faa BDMAEE DAPA TCPP

0 3,203 190 0 12 19 27 71
3,901 3,351 180 4,861 9 18 27 77
7,802 3,518 180 9,719 8 17 26 74
11,702 3,383 189 14,579 7 16 26 67
15,604 3,492 182 19,436 6 16 26 65
19,506 3,523 179 24,297 5 16 26 61
23,408 3,413 180 29,154 5 16 26 59
27,313 3,376 191 34,014 5 15 26 55
31,216 3,423 184 38,873 4 15 26 63
35,116 3,350 190 43,730 4 16 26 63
39,020 3,246 184 48,625 4 16 26 61
42,925 3,293 181 53,482 4 16 26 58
RSD (%) 3.0 2.5 38.8 6.8 1.2 10.4
MW (g/mol) 160.3 327.5 134.0 160.3 201.3 327.5
BP (  C) 189 244 16 189 102 244

Note: BDMAEE ¼ bis(2-dimethylaminoethyl)ether; BP ¼ boiling point; DAPA ¼ Bis(dimethlyami-

nopropyl)methylamine; MW ¼ molecular weight; RSD ¼ relative standard deviation; TCPP ¼
tris-(1-chloro-2-propyl)phosphate. RSDs are over the same period (sample size Ø 92 by 80 mm).
Sampling times (volumes) were reduced for the open-cell sample to reduce overloading. Additional
physical data have been added for convenience.
a
Quantified as toluene equivalent.

emission factors over 12 h—similar to open-cell SPF. For closed-cell samples, the
relatively low emission measurements of amines and other target compounds might
be improved with a higher sampling volume (perhaps 2 L could be used).
Open-cell SPF insulation is more porous than closed-cell SPF insulation be-
cause the cellular structure is porous; consequently, chemical compounds release
more rapidly from open-cell SPF than closed-cell SPF. A constant emission value
indicates a source-phase mass transfer (internal diffusion) emission process for the
remaining unreacted ingredients used during production [13]. Such emissions will
stay constant until the source has depleted or is inactivated [14]. Higher emission
factors obtained for the open-cell sample were probably due to a more porous cellu-
lar structure than closed-cell SPF; consequently, chemical compounds migrated
more rapidly from open-cell SPF than closed-cell SPF.
The closed-cell sample behaved differently. The emission values decreased im-
mediately from initial levels but over time fell to near constant levels. The closed
cells/pores hampered diffusion, particularly through the cured top surface. A gas-
phase mass transfer (external diffusion) could be responsible for this phenomenon.
During storage the chemical ingredients diffused onto the closed-cell sample surface
 NIE ET AL., DOI 10.1520/STP158920150040 39

FIG. 5 Emission factors as a function of time for closed-cell SPF. The right ordinate is
for TCPP only. Time was measured in integer seconds but is shown in hours
along the top of the plot for convenience. Open-cell emission factors were
essentially constant over the same time period and were not plotted.

Time (hours)
0 5 10 15
30 90
80

Emission Factor (µg/m2·h) - TCPP

25
70
Emission Factor (µg/m2·h)

20 60
50
15
40
10 30
20
5
10
HFC-245fa BDMAEE DAPA TCPP
0 0
0 10,000 20,000 30,000 40,000 50,000 60,000
Time (seconds)

and collected there. After opening the sample, the VOCs/SVOCs left the surface for
the air flowing across the surface. After that, a source-phase mass transfer (internal
diffusion)-controlled emission process will followed.
Because these measurements were carried out in an automated system, an ana-
lyst did not have to insert and remove TD tubes every hour for more than 12 h.
Also, because the measurements were carried out by the system, measurement
times were known to a 61-s time resolution; in addition, purge volumes were mea-
sured, not estimated, with an error of 62 % and a precision of 61 mL. All data
were electronically recorded.

EMISSION FACTORS AT ELEVATED TEMPERATURES

One convenience of an on-line system is the method-level control of temperature. One
type of experiment that is more easily done is one in which emissions as a function of
temperature needs to be carried out. For example, SPF emissions as an attic heats up dur-
ing the day can be both simulated and automated. As an example of this concept, both
types of SPF samples were measured at 23 C but again at 40 and 65 C as well. Three rep-
licate measurements of each sample were made at each temperature using an automated
sequence. These samples had been previously run repeatedly for approximately 15 h be-
fore this experiment so that the emission factors versus time had largely stabilized; in this
way, the time-related effects were diminished. The results can be found in Table 6.
40 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 6 Emission factors (lg/m2h) of target compounds emitted from SPF insulation samples at
three different temperatures.

Open Cell Closed Cell

Temperature BDMAEE TCPP TMIBPA HFC-245fa BDMAEE DAPA TCPP

a

23 C: average 3381 184 NQ 12 16 26 71a
% RSD (n 5 12) 3.0 2.5 6.8 1.2
40 C: average OL 793 NQ 10 59 45 260
% RSD (n 5 3) 7.1 1.6 15.8 2.8 6.1
65 C: average OL 3416 1826 24 OL OL OL
% RSD (n 5 3) 2.1 0.9 1.7

Note: BDMAEE ¼ bis(2-dimethylaminoethyl)ether; DAPA ¼ Bis(dimethlyaminopropyl)methylamine;

NQ ¼ not quantitated (the calculated concentration fell below the lowest calibration point); OL ¼ over-
load; TCPP ¼ tris-(1-chloro-2-propyl)phosphate; TMIBPA ¼ Tetramethyliminobispropylamine. The
sample was measured three times at each temperature (sample size Ø 92 by 80 mm).
a
Because of the declining behavior of HFC-245fa and TCPP, only the first emission factor is listed.

The emission factors of the amine catalysts, flame retardant, and blowing agent
(closed-cell sample only) increased dramatically at elevated temperatures. In the
open-cell sample, TMIBPA could only be quantitatively determined at 65 C. At 23
and 40 C, the chromatographic peak area fell below the lowest calibration level
(200 ng). The BDMAEE peak was overloaded at both elevated temperatures. Conse-
quently, TMIBPA and BDMAEE could not be simultaneously determined. The
emission of TCPP increased from 184 lg/m2h (23 C) to 793 lg/m2h (40 C) and
then to 3,416 lg/m2h (65 C), a total increase of 1,350 %. At each temperature, rep-
licate measurements were stable over the measurement period (3 h). This constant
emission suggests emission controlled by source-phase mass transfer (internal dif-
fusion) in the open-cell sample, as was also observed in the previous emission ver-
sus time work previously discussed.
In the closed-cell sample, emissions of all target compounds also increased at
elevated temperatures. Replicate measurements were constant at 40 C during three
measurements (3 h), which also suggest internal diffusion-controlled emission.
BDMAEE, DAPA, and TCPP were overloaded at 65 C, although replicate measure-
ments over 3 h showed a decreasing trend. This trend may indicate exhausting of
emissions from inside the sample. If this is true, heating a closed-cell sample to
65 C for several hours may reduce emissions below observable levels.

Conclusions
A microscale chamber system was developed for determining chemical emissions
from SPF insulation using a DHS-L autosampler. A blowing agent, amine catalysts,
and a flame retardant used in open- and closed-cell SPF samples were well detected
at 23 C and with higher emission factors at elevated temperatures of 40 and 65 C.
 NIE ET AL., DOI 10.1520/STP158920150040 41

A lower sampling volume was needed for higher-emitting open-cell samples than
with lower-emitting closed-cell samples. Emission behavior and emission factors
over time were easily obtained for approximately 15 h as monitored by an auto-
mated chamber sampling system. The observed emission behavior indicates a
source-phase mass transfer process (internal diffusion) occurs in foam samples
(particularly open-cell SPF insulation) and is the controlling factor for emission
values. Closed-cell samples undergo two emission processes: a gas-phase mass
transfer (external diffusion) process at the surface followed by a source-phase
mass transfer process from the bulk.
Automating these experiments has the obvious advantage of less involvement
of the analyst and better reliability due to the reduction of human error. However,
in addition, the major variables in a microscale chamber experiment—flow, temper-
ature, and time—are all under computer control in the system presented herein.
Sampling times, temperatures, and volumes are usually determined with a very
high degree of precision, which can be important for mathematical modeling of air
concentration decay [5]. Computer control can also be used for transmitting meth-
od parameters between systems.
In the future, sink effects (analyte recovery), reproducibility, durability, and
other data of merit will be more fully studied. The relation between sample volume/
surface size and emission factors will be considered in more detail.

ACKNOWLEDGMENTS
We thank John R. Sebroski for supplying SPF samples and for comments on this
work.

References

[1] Dernhl, C. U., “Health Hazards Associated with Polyurethane Foam,” J. Occup. Med.,
Vol. 8, No. 2, 1965, pp. 59–62.

[2] Umweltbundesamt, “Substituting Environmentally Relevant Flame Retardants: Assess-

ment Fundamentals,” Report No. 25, Vols. I–III, Umweltbundesamt, Berlin, Germany,
2000.

[3] Sebroski, J., “Research Report for Measuring Emissions from Spray Polyurethane Foam
(SPF) Insulation,” American Chemistry Council, Washington, DC, 2012.

[4] ISO 1600-9, Determination of the Emission of Volatile Organic Compounds from Building
Products and Furnishings—Emission Test Chamber Method, International Organization
for Standardization, Geneva, Switzerland, 2006.

[5] ISO 1600-6, Determination of Volatile Organic Compounds in Indoor and Test Chamber
Air by Active Samping on Tenax TA Sorbent, Thermal Desorption and Gas Chromatog-
ramphy Using MS/FID, International Organization for Standardization, Geneva, Switzer-
land, 2010.
42 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

[6] Uhde, E. and Salthammer, T., “Influence of Molecular Parameters on the Sink Effect in
Test Chambers,” Indoor Air, Vol. 16, No. 2, 2006, pp. 158–165.

[7] ASTM D7859-13e1, Standard Practice for Spraying, Sampling, Packaging, and Test Speci-
men Preparation of Spray Polyurethane Foam (SPF) Insulation for Testing of Emissions
Using Environmental chambers, ASTM International, West Conshohocken, PA, 2013,
www.astm.org

[8] ASTM D7706-11, Standard Practice for Rapid Screening of VOC Emissions from Products
Using Micro-Scale Chambers, ASTM International, West Conshohocken, PA, 2011,
www.astm.org

[9] Terheiden, A., Landers, R., Schloens, H., and Hubel, R., “Novel Amine Catalyst for Low
Emission Flexible Polyurethane Foam,” presented at Polyethanes Technical Conference
2008, San Antonio, TX, September 28–30, 2008, American Chemical Council, Washing-
ton, DC.

[10] Zhao, D., Little, J. C., and Cox, S. S., “Characterizing Polyurethane Foam as a Sink for or
Source of Volatile Organic Compounds in Indoor Air,” J. Environ. Eng., Vol. 130, No. 9,
2004, pp. 983–989.

[11] Salthammer, T., Fuhrmann, F., and Uhde, E., “Flame Retardants in the Indoor Enviro-
ment—Part II: Release of VOCs (Triehtylphosphasate and Halogenated Degradation
Products) from polyurethane,” Indoor Air, Vol. 13, No. 1, 2003, pp. 49–52.

[12] Albrecht, W. N. and Stephenson, R. L., “Health Hazards of Tertiary Amine Catalysts,”
Scand. J. Work Environ. Health, Vol. 14, No. 4, 1988, pp. 209–219.

[13] Salthammer, T., “Environmental Test Chamber and Cells,” Organic Indoor Air Pollutants,
Salthammer, T. and E. Uhde, Eds., Wiley-VCH, Weinheim, Germany, 2007, pp. 101–115.

[14] Brown, S. K., “Building Products as Sources of Indoor Organic Pollutants,” Organic In-
door Air Pollutants, Salthammer, T. and E. Uhde, Eds., Wiley-VCH, Weinheim, Germany,
2007, pp. 373–404.
 DEVELOPING CONSENSUS STANDARDS FOR MEASURING CHEMICAL EMISSIONS 43

STP 1589, 2017 / available online at www.astm.org / doi: 10.1520/STP158920150047

J. Paul Duffy1 and Richard Wood2

VOC Analysis of Commercially

Available Spray Foam Products
Citation
Duffy, J. P. and Wood, R., “VOC Analysis of Commercially Available Spray Foam Products,”
Developing Consensus Standards for Measuring Chemical Emissions from Spray Polyurethane
Foam (SPF) Insulation, ASTM STP1589, J. Sebroski and M. Mason, Eds., ASTM International,
West Conshohocken, PA, 2017, pp. 43–56, http://dx.doi.org/10.1520/STP1589201500473

ABSTRACT
A working group formed under the auspices of the American Chemistry Council
Center for the Polyurethanes Industry (CPI) recently developed a procedure for
using a room-size chamber to assess the impact of ventilation on trade worker
re-entry times for work environments after the application of various generic
spray foam systems. Key to the assessment was the fact that the procedures also
demonstrated techniques for evaluating the volatile and semivolatile organic
compound emissions from these spray polyurethane foam (SPF) systems. Thus
far, the working group has only reported results from generic high-pressure
low- and medium-density spray foam systems and low-pressure systems.
No commercially available formulations have been reported as yet. This paper
presents the evaluation of two commercially available high-pressure low- and
medium-density SPF systems using the CPI re-entry study test protocol. It also
suggests how the chemical emissions data can be interpreted by risk assessment
professionals to establish trade worker re-entry time guidance for contractors
spraying these systems. This study aimed to determine whether worker re-entry
times could be reduced from the industry practice of 24 h if specific rates of
workplace ventilation were employed. Ventilation rates up to 10 ACH were
attempted and results are presented herein. Ultimately, the methods were
adapted and used for guiding improvements to the design of formulations. This
paper provides the rationale behind what was done and the results achieved.

Manuscript received May 20, 2015; accepted for publication September 30, 2016.
1
Icynene Inc., 6747 Campobello Rd., Mississauga, ON, Canada, L5N 2L7
2
Air Products and Chemicals, Inc., 7201 Hamilton Blvd., Allentown, PA 18195
3
ASTM Symposium on Developing Consensus Standards for Measuring Chemical Emissions from Spray
Polyurethane Foam (SPF) Insulation on April 30–May 1, 2015 in Anaheim, CA.

Copyright V
C 2017 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959.
44 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

Keywords
low-density spray polyurethane foam, medium-density spray polyurethane
foam, worker re-entry time

Introduction
Volatile organic compound (VOC) emissions are of interest to many in the spray
polyurethane foam (SPF) industry for a number of reasons:
1. They have been raised as an issue by some regulatory authorities as part of the
approval of products.
2. They may have a potential negative impact on the health of certain adult workers.
3. They may affect the re-entry guidance given to trades and other workers in the
period immediately after the application of spray foam.
As part of an ongoing effort in association with the United States, a task force
of the American Chemistry Council Center for the Polyurethanes Industry (CPI)
recently completed several research projects on SPF aimed at developing large-scale
methods for testing emissions from SPF products. The CPI work was aimed at
developing consensus around test procedures, equipment specifications, test setup,
and measurement. Thus far, what has been reported are results for noncommercial
generic SPF formulations with ingredients disclosed as part of the work. Specifically,
the CPI work has described a procedure simulating the behavior of spray foams in
a room-size chamber (approximately 2.4 by 2.4 by 2.4 m). Emission decay is mea-
sured in the presence of a calibrated ventilation rate to simulate the effect of spray
and ventilation of the work area [1].
Icynene was interested in using the CPI re-entry study test protocol to assess
emissions from its SPF systems to help determine the appropriateness of its
guidance regarding combination(s) of workplace ventilation versus trade worker
re-entry time for its portfolio of products. Thus far, several products have been
studied. This paper presents the findings of studies on two products: Icynene Clas-
sic Max low-density open-cell foam and Icynene ProSeal Eco medium-density
closed-cell foam. The objective of the research was to determine whether increased
ventilation could be used to document reduced trade worker re-entry times versus
the 24 h suggested by the CAN/ULC S774 method and industry practice [2].
Previous studies conducted by CPI [2] have concluded that methylene
bisphenyl isocyanate is a concern during application and measurable for up to an
hour after application; however, it quickly reacts and/or dissipates thereafter. This
typically is not an issue because installers and others in the work area are required
to wear full personal protective equipment during and immediately after foam is
installed. Other components, such as fire retardants, blowing agents, and amine cat-
alysts, may not be bound to foam, resulting in measurable emissions after applica-
tion. This study focused on amine catalyst emissions and reports concentrations of
the fire retardant tris(1-chloro-2-propyl) phosphate (TCPP) and the blowing agent
1,2-dichloroethylene (1,2-DCE) detected. The five commercially available tertiary
 DUFFY AND WOOD, DOI 10.1520/STP158920150047 45

amine catalysts evaluated during this study are designated as Catalysts A–E to pro-
tect Icynene propriety information.

Methodology
The Icynene formulations were sprayed in the same facility used for the CPI re-entry
study: the Air Products and Chemicals, Inc. air product ventilated spray room located
in Allentown, PA (Fig. 1) [1]. The air sampling and analytical methods used for evalu-
ating Icynene formulation emissions were the same or equivalent to those used for the
CPI re-entry study. The ventilated room is 2.4 by 2.4 by 2.4 m (13.8 m3) and capable
of air turnover rates ranging from 1 to 13 air changes per hour (ACH). The true venti-
lation flow rate was measured 1 day before testing. Two measured air exchange rates
were selected for the first test series: 10.0 ACH (9.6 ACH actual) for the initial
measurements and 1.0 ACH (1.07 ACH actual) for testing beginning at Hour 3 and
through the end of the test period at Hour 48. The higher ventilation rate of 9.6 ACH
was selected for the first segment of air monitoring because it could represent mini-
mal mechanical air ventilation offered by commercial ventilation equipment. In
contrast, the Dutch Implementation Directive URL 27-101 [3] requires an air
exchange rate of 30 ACH or greater using commercially available mechanical venti-
lation for crawlspace SPF application. The 1.07 ACH rate would be considered

FIG. 1 Air products: SPF spray room.

Air Products – SPF Spray Room

2.4 m MDI
1.07 ACH drum
Substrate
Air Flow

Air Flow

9.60 ACH
Spray Polyol
2.4 m 3.0 m
High
2.4 m Spray Spray Booth
Spray Room
Machine

Plexiglas 3.2 m
1.8 m Spray Rm
Entry
6.7 m
1.4 m Breathing
Air

7.6 m
46 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

passive ventilation through windows, etc., representative of a sprayed building

space without the higher-volume supplemental mechanical ventilation.
The two formulations were sprayed on 31-cm stud cavity cardboard inserts for
15 min. Each insert was 198 cm tall and 36 cm wide. The ventilation operated at
9.6 ACH during SPF application. The temperature measured in the ventilated room
was approximately 22 C during each session. The formulations were sprayed “on ratio”
per Icynene recommendations (i.e., A and B sides were sprayed at a 1:1 ratio at approxi-
mately 68 atm). The dynamic pressure range was 58 to 61 atm, and the hose tempera-
ture range was 49 to 57 C for the two formulations. The ProSeal Eco inserts had a foam
thickness of approximately 2 in. The Classic Max insert foam thickness was 7.5 to 9 cm.
After SPF application, the panels were arranged around the ventilated room to
expose the surface of all 12 panels. The test panels remained in the room with the
ventilation operating at 9.6 ACH. Air samples were then collected according to the
schedule listed in the section that follows. Air sampling duration was 20 min for
each sample, with the sample time bracketing the hour (10 min before the hour and
10 min after the hour). This approach, also used for the CPI ventilation study, pro-
vided an average 20-min concentration centered on the hour monitored. Back-
ground samples were collected prior to the test sessions, and field blanks were
submitted with each group of samples.

Air Sampling and Analytical Methods

Stationary/area air samples were collected in the center of the spray room, sus-
pended at a height of approximately 5 ft (Fig. 1) to represent a worker’s breathing
zone. All samples were collected by the same method for the same duration and in
the same location. Variation in pre- and postpump calibrations was <2 %, which
was well within acceptable limits.
Air samples were collected by drawing air through 5-mm quartz-wool stainless-
R TM
steel TenaxV TA 35/60 Carbograph 5TD 40/60 thermal desorption tubes prepared by
Markes International (Llantrisant, Wales) using calibrated MSA ELF pumps equipped
with Flow Gemini sample tube holders. The air sampling tubes contained several pro-
prietary sorbents designed to collect key polyurethane VOC and SVOC components,
including amine catalysts, fire retardants, and blowing agents. At the completion of the
air monitoring sessions, samples were submitted to the Bayer MaterialScience Environ-
mental Analytics Laboratory, where they were analyzed using thermal desorption-
capillary gas chromatography-mass spectrometry in accordance with the U.S. EPA’s
Compendium Method T0-17 [4]. High-purity standards were prepared and analyzed
to identify sample components and to determine instrument response factors.
• Hour 1: Air samples were collected for 20 min 1 h after application. Ventilation
remained at 9.64 ACH.
• Hour 2: Air samples were collected for 20 min 2 h after application. Ventilation
remained at 9.64 ACH.
• Hour 3: At the completion of second hour of sampling, ventilation was
reduced from 9.6 to 1.07 ACH. This was done to simulate the impact of
 DUFFY AND WOOD, DOI 10.1520/STP158920150047 47

discontinuing aggressive ventilation with high-capacity fans. The ventilation

continued to operate at a reduced rate of 1 ACH for the remainder of the test-
ing. No air sampling was conducted during Hour 3.
• Hour 4: Air samples were collected for 20 min 4 h after application. Ventilation
remained at 1.07 ACH.
• Hour 8: Air samples were collected for 20 min 8 h after application. Ventilation
remained at 1.07 ACH.
• Hour 12: Air samples were collected for 20 min 12 h after application. Ventila-
tion remained at 1.07 ACH.
• Hour 24: Air samples were collected for 20 min 24 h after application. Ventila-
tion remained at 1.07 ACH.
• Hour 48: Air samples were collected for 20 min 48 h after application. Ventila-
tion remained at 1.07 ACH.

Discussion
The results of the catalyst analysis are listed in Table 1 and Table 2. Catalyst results are
also presented in Fig. 2 and Fig. 5. Fire retardant and blowing agent results were provid-
ed by the laboratory and are summarized in Table 3 and Table 4. Fire retardant results
were also displayed in Fig. 3 and Fig. 6. Blowing agent results were also displayed in

TABLE 1 ProSeal Eco medium-density closed-cell formulation: tertiary amine catalysts.

Concentration (ppm)

Sample Description Sample Volume (L) Catalyst A Catalyst B Catalyst C

Ventilated room 4.63 0.08 <0.001 0.03

1 h postspray
(9.64 ACH)
Ventilated room 5.04 0.04 <0.001 0.01
2 h postspray
(9.64 ACH)
Ventilated room 5.06 0.03 <0.001 0.01
4 h postspray
(1.07 ACH)
Ventilated room 5.17 0.02 <0.001 0.01
8 h postspray
(1.07 ACH)
Ventilated room 5.03 0.03 <0.001 0.01
12 h postspray
(1.07 ACH)
Ventilated room 5.04 0.01 <0.001 0.01
24 h postspray
(1.07 ACH)
Ventilated room 5.00 <0.001 <0.001 0.004
48 h postspray
(1.07 ACH)
Level of quantitation 0.001 0.001 0.003
48 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 2 Classic Max low-density open-cell formulation: tertiary amine catalysts.

Concentration (ppm)

Sample Description Sample Volume (L) Catalyst D Catalyst E

Ventilated room 5.05 0.24 0.007

1 h postspray
(9.64 ACH)
Ventilated room 5.04 0.22 0.007
2 h postspray
(9.64 ACH)
Ventilated room 5.01 0.23 0.005
4 h postspray
(1.07 ACH)
Ventilated room 5.04 0.24 0.011
8 h postspray
(1.07 ACH)
Ventilated room 4.99 0.21 0.005
12 h postspray
(1.07 ACH)
Ventilated room 4.99 0.21 <0.003
24 h postspray
(1.07 ACH)
Ventilated room 5.03 0.18 0.004
48 h postspray
(1.07 ACH)
Level of quantitation 0.003 0.003

FIG. 2 ProSeal Eco medium-density closed-cell formulation: Catalysts A–C.

0.09
Venlaon Rate Venlaon Rate
0.08
Start – Hour 3 Hour 3 – Hour 48
= 9.6 ACH = 1.07 ACH
0.07
Catalyst A
Catalyst B
Concentraon (ppm)

0.06
Catalyst C
0.05

0.04

0.03

0.02

0.01

0
1 2 4 8 12 24 48
Hours Aer Applicaon
 DUFFY AND WOOD, DOI 10.1520/STP158920150047 49

FIG. 3 ProSeal Eco medium-density closed-cell formulation: TCPP.

0.025
Venlaon Rate Venlaon Rate
Start – Hour 3 Hour 3 – Hour 48 TCPP
= 9.6 ACH = 1.07 ACH
0.02
Concentration (ppm)

0.015

8 hr average = 0.011 ppm TCPP

0.01

0.005

0
1 2 4 8 12 24 48
Hours Aer Applicaon

FIG. 4 ProSeal Eco medium-density closed-cell formulation: 1,2-DCE.

0.6
1,2 – DCE
OEL = 200 ppm 8 hr TWA
0.5
Venlaon Rate Venlaon Rate
Start – Hour 3 Hour 3 – Hour 48
= 9.6 ACH = 1.07 ACH
0.4
Concentration (ppm)

0.3
1,2-DCE

0.2

0.1

0
1 2 4 8 12 24 48
Hours Aer Applicaon
50 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 5 Classic Max low-density open-cell formulation: Catalysts C and D.

TABLE 3 ProSeal Eco medium-density closed-cell formulation: TCPP and 1,2-DCE.

Concentration (ppm)

Sample Description Sample Volume (L) TCPP 1,2-DCE

Ventilated room 4.63 0.008 0.02

1 h postspray
(9.64 ACH)
Ventilated room 5.04 0.003 0.02
2 h postspray
(9.64 ACH)
Ventilated room 5.06 0.018 0.17
4 h postspray
(1.07 ACH)
Ventilated room 5.17 0.005 0.12
8 h postspray
(1.07 ACH)
Ventilated room 5.03 0.003 0.10
12 h postspray
(1.07 ACH)
Ventilated room 5.04 0.002 0.05
24 h postspray
(1.07 ACH)
 DUFFY AND WOOD, DOI 10.1520/STP158920150047 51

TABLE 3 (Continued)

Concentration (ppm)

Sample Description Sample Volume (L) TCPP 1,2-DCE

Ventilated room 5.00 <0.001 0.23

48 h postspray
(1.07 ACH)
Level of quantitation 0.001 0.001

Note: A background sample was collected prior to spraying the formulation. A background concen-
tration of 0.01 ppm 1,2-DCE was detected. The background 1,2-DCE concentration was not sub-
tracted from the reported findings.

Fig. 4 and Fig. 7. Small quantities (<0.003 ppm) of chlorobenzene were also detected in
several of the open-cell formulation samples.
Background air sampling was conducted in the spray room prior to spraying
each formulation. All chemicals were below detection limits with the exception of

TABLE 4 Classic Max low-density open-cell formulation: TCPP and 1,2-DCE.

Concentration (ppm)

Sample Description Sample Volume (L) TCPP 1,2-DCE

Ventilated room 5.05 0.002 0.55

1 h postspray
(9.64 ACH)
Ventilated room 5.04 0.002 0.36
2 h postspray
(9.64 ACH)
Ventilated room 5.01 0.002 0.31
4 h postspray
(1.07 ACH)
Ventilated room 5.04 0.002 0.34
8 h postspray
(1.07 ACH)
Ventilated room 4.99 0.002 0.25
12 h postspray
(1.07 ACH)
Ventilated room 4.99 0.002 0.29
24 h postspray
(1.07 ACH)
Ventilated room 5.03 <0.002 0.21
48 h postspray
(1.07 ACH)
Level of quantitation 0.002 0.21

Note: A background sample was collected prior to spraying the formulation. A background concen-
tration of 0.26 ppm 1,2-DCE was detected. The background 1,2-DCE concentration was not sub-
tracted from the reported findings.
52 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 6 Classic Max low-density open-cell formulation: TCPP.

FIG. 7 Classic Max low-density open-cell formulation: 1,2-DCE.

0.6
Ven laon Rate Ven laon Rate 1,2- DCE
Sta rt – Hour 3 Hour 3 – Hour 48
= 9.6 ACH = 1.07 ACH OEL =200 ppm 8 hr TWA
0.5

0.4
Concentration (ppm)

0.3

0.2

0.1

1,2-DCE
0
1 2 4 8 12 24 48
Hours Aer Applicaon
 DUFFY AND WOOD, DOI 10.1520/STP158920150047 53

1,2-DCE. The DCE background concentration on Day 1 was 0.01 ppm. The concen-
tration on Day 2 was 0.26 ppm. The higher concentration detected on Day 2 was
likely due to a reduced ventilated room purge time between spray formulations that
resulted in cross-contamination. The background contamination concentrations
were not subtracted from the final results and the reduced purge had no impact on
the other chemicals measured.

TERTIARY AMINES
There are occupational exposure limits (OELs) for one of the amine catalysts evalu-
ated during this study. The province of Ontario has set OELs of 3 ppm as an 8-h
time-weighted average (TWA) concentration and a 15-min short-term exposure
limit (STEL) of 6 ppm for Catalyst D [5]. The remaining catalysts measured during
this study do not have published exposure limits; however, a surrogate OEL for
bis(2-dimethylaminoethyl) ether (BDMAEE) may be used to assess the relative risk
of the remaining four catalysts [6]. The American Conference of Governmental
Industrial Hygienists has set a threshold limit value for BDMAEE of 0.05 ppm as an
8-h TWA and a 15-min STEL of 0.15 ppm [7].
The results of this analysis indicate that amine catalyst concentrations were
below published OELs. All catalysts showed a decrease in concentrations between
Hours 1 and 2 when the spray room ventilation operated at 9.6 ACH. Concentra-
tions remained low or below the level of quantitation when the ventilation was re-
duced to 1.07 ACH 3 h after spraying. Catalyst concentrations generally remained
steady between Hours 4 and 12 but began to decline significantly between Hours
12 and 48 (Fig. 2 and Fig. 5).

TCPP
TCCP concentrations (Tables 3 and 4) (Fig. 3 and Fig. 6) followed a pattern similar
to the amine catalysts. Concentrations for the ProSeal Eco closed-cell formulation
declined from Hour 1 to Hour 2, then rose at Hour 4 when the ventilation de-
creased, and then steadily declined from Hour 4 to Hour 48. TCPP concentrations
remained steady at 0.002 ppm for the Classic Max open-cell formulation until Hour
48 when the concentration fell below the detection limit.
No OELs have been established for TCPP. TCPP animal toxicity data, however,
have been evaluated by researchers for the purpose of determining potential health
risks to humans. European researchers [8] developed a chronic toxicity reference
dose for TCPP based on the lowest observed adverse effects level (LOAEL) of
52 mg/kg/day after a 13-week rat exposure study where the authors observed
increased liver weights in male rats. The LOAEL was adjusted to 42 mg/kg/day
assuming 80 % absorption. Dividing by 10 to adjust for animal to human doses and
another 5 for intraspecies results in an absorbed reference dose of 0.84 mg/kg/day.
Based on the 8-h TWA for the ProSeal Eco results (Fig. 3) of approximately
0.011 ppm and assuming a 70-kg worker inhales approximately 10 m3 of air over
an 8-h day, the absorbed dose, assuming 100 % absorption by inhalation, would be
54 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 8 SPF application (left); Air sampling equipment (right).

0.02 mg/kg/day. The calculated absorbed dose would be well below the reference
dose of 0.84 mg/kg/day [9].

1,2-DCE
1,2-DCE concentrations (Tables 3 and 4) (Fig. 4 and Fig. 7) decreased from Hour
1 to Hour 2 and then slightly decreased from Hour 4 to Hour 48 but never fell
below detection limits at any time during the study. Concentrations were well with-
in the 8-h TWA OEL of 200 ppm throughout the study.

Conclusions
Five tertiary amine catalysts were evaluated during this study: however, only one
has been assigned an OEL. A surrogate amine catalyst with a published OEL was
selected to represent the other four catalysts detected. All concentrations measured
during this study were well below the OELs; therefore, the concentrations detected
after the application of the closed- and open-cell formulations under controlled
conditions indicate there is no significant risk to informed workers from catalyst
emissions after SPF application. Concentrations of TCPP and 1,2-DCE were also
below levels that would pose a risk to the health of exposed workers [9].
The spray room parameters and ventilation rates selected for this study may
not represent all commercial SPF field applications; therefore, the data may not
represent actual environmental conditions in all instances. Based on the Dutch
30-ACH requirement for a small confined space during application, the 10-ACH
rate selected for a small space in this study would appear to represent a similar but
poorer ventilation control scenario.
Following a professional health risk assessment based on the emission data pre-
sented in this paper, it may be concluded that residual odors from the spraying
operation would be expected to be the driving force in determining exclusion
 DUFFY AND WOOD, DOI 10.1520/STP158920150047 55

timelines and ventilation needs. Concentrations were generally stable after 4 to 8 h,

suggesting air changes selected for this study reduced potential health risks and pro-
duced stable conditions. On this basis, a re-entry time for informed workers of 4 to
8 h after installation could be employed given similar environmental conditions [9].
This represents a significant reduction from what was previously believed necessary
(i.e., 24 h). This trade worker re-entry time may not be appropriate for all popula-
tions, especially for more sensitive individuals, such as building residents. For this
reason, the re-occupancy guidance for building tenants has not been changed.

Future Research Needs

Although this study followed the protocol developed for the CPI trade worker
re-entry project, further refinements to the protocol would be appropriate. The
authors recommend the following:
Consideration should be given to instituting robust quality assurance procedures.
Although the CPI study demonstrated acceptable precision between duplicate sam-
ples, obtaining one or two duplicate samples during each study and one or two air
samples at a second location and/or height in the spray room to evaluate air mixing
would be advisable.
The chemicals evaluated in this study include amine catalysts and the fire retardant
TCPP. These chemicals are added to the resin side of the formulation and have been
shown by this study and others to emit from SPF over time. Other chemicals may be
formed or be present as impurities; therefore, future studies should be expanded to
evaluate the emissions of known or suspected impurities or contaminants.
Last, future studies should be structured to gather information required for the
determination of chemical emission factors as well as air concentrations. In order
to properly evaluate worker exposure scenarios, emissions are best expressed as a
concentration (ppm/ppb/mg/m3) to compare results to OELs. Data generated for
the purpose of predicting indoor air quality for extended periods of time are better
reported as chemical emission factors. Emission factors can be used for calculating
long-term emissions or entered into computer modeling programs to provide a
wider range of long-term emission information.

References

[1] Wood, R., “CPI Ventilation Research Project for Estimating Re-Entry Times for Trade
Workers Following Application of Three Generic Spray Polyurethane Foam For-
mulations,” presented at the Center for the Polyurethanes Industry 2014 Polyurethanes
Technical Conference, Dallas, TX, September 22–24, 2014—unpublished.

[2] CAN/ULC S774, Standard Laboratory Guide for the Determination of Volatile Organic
Compound Emissions from Polyurethane Foam, Standards Council of Canada, Ottawa,
Canada, https://www.scc.ca
56 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

[3] The Netherlands Implementation Directive, “Implementation Guidelines for Crawlspace

Spraying with Polyurethane Foam,” URL 27-101 dd 02/2009/2013.

[4] U.S. EPA, “Compendium of Methods for the Determination of Toxic Organic Compounds
in Air. Second Edition. EPA/625/12.96/0106. Compendium Method TO-17. Determination
of Volatile Organic Compounds in Ambient Air Using Active Sampling On to Sorbent
Tubes,” January 5, 1999.

[5] Ontario Ministry of Labour, “Occupational Health and Safety Act R.R.O. 1990, Regulation
833—Control of Exposure to Biological or Chemical Agents,” https://www.ontario.ca/
laws/regulation/900833 (accessed October 11, 2016).

[6] Consumer Product Safety Commission, “Status Report: Staff Review of Five Amine
Catalysts in Spray Polyurethane Foam (CPSC-D-07-0006),” https://www.cpsc.gov/s3fs-
public/amine.pdf (accessed October 11, 2016).

[7] American Conference of Governmental Industrial Hygienists (ACGIH), The 2014 Supple-
ment to the Documentation of Threshold Limit Values and Biological Exposure Indices,
ACGIH, Cincinnati, OH, 2014.

[8] European Union, “Risk Assessment Report: Tris (2-Chloro-1-Methylethyl) Phosphate

(TCPP),” 2008, https://echa.europa.eu/documents/10162/13630/trd_rar_ireland_tccp_
en.pdf (accessed October 11, 2016).

[9] DeMott, R. and Gauthier, T., “Potential Risks from Exposure to Tertiary Amine Catalysts
Following Application of Spray Polyurethane Foam Insulation,” ENVIRON International
Corp., 2014—unpublished.
 DEVELOPING CONSENSUS STANDARDS FOR MEASURING CHEMICAL EMISSIONS 57

STP 1589, 2017 / available online at www.astm.org / doi: 10.1520/STP158920150044

Dustin Poppendieck,1 Matthew Schlegel,2

Angelica Connor,2 and Adam Blickley2

Flame Retardant Emissions from

Spray Polyurethane Foam
Insulation
Citation
Poppendieck, D., Schlegel, M., Connor, A., and Blickley, A., “Flame Retardant Emissions from
Spray Polyurethane Foam Insulation,” Developing Consensus Standards for Measuring
Chemical Emissions from Spray Polyurethane Foam (SPF) Insulation, ASTM STP1589,
J. Sebroski and M. Mason, Eds., ASTM International, West Conshohocken, PA, 2017, pp. 57–76,
http://dx.doi.org/10.1520/STP1589201500443

ABSTRACT
The desire to build more energy-efficient homes in the United States has led to
the expansion of the residential spray polyurethane foam (SPF) insulation
industry. Upon application of SPF, reacting chemicals form expanding
polyurethane foam that fills cracks and gaps, reducing infiltration and thermal
conductivity of the building envelope. However, more information is being
sought on the chemical emissions from SPF to better understand occupant
exposures and any potential impacts on health. The objective of this
investigation was to investigate the emission of flame retardant tris(1-chloro-2-
propyl) phosphate (TCPP) from SPF using both microchambers and a full-scale
residential test facility. Two high-pressure open-cell foams and one high-pressure
closed-cell foam were tested using microchambers. After 100 h, TCPP
concentrations from the open-cell samples were 100 times higher than TCPP
concentrations from the closed-cell SPF. TCPP emissions from open-cell foam
were found to correlate exponentially with temperature and vary with flow rate,
indicating emission factors from SPF microchamber experiments may not
directly predict TCPP concentrations in buildings without consideration of

Manuscript received May 15, 2015; accepted for publication November 10, 2015.
1
National Institute of Standards and Technology, Indoor Air Quality and Ventilation Group, Energy and
Environment Division, Engineering Laboratory, Gaithersburg, MD 20899-8633
2
Drexel University, Materials Science and Engineering Dept., 3141 Chestnut St., Philadelphia, PA 19104
3
ASTM Symposium on Developing Consensus Standards for Measuring Chemical Emissions from Spray
Polyurethane Foam (SPF) Insulation on April 30–May 1, 2015 in Anaheim, CA.

Copyright V
C 2017 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959.
58 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

material mass transfer properties. Due to the use of TCPP in furniture, SPF has
not previously been positively identified as a primary source of indoor TCPP
concentrations in actual buildings. This research measured airborne TCPP
concentrations in the furniture-free National Institute of Standards and
Technology Net Zero Energy Residential Test Facility (NZERTF) that contained
15 m2 of exposed 2-year-old open-cell SPF. The measured NZERTF TCPP
emission rates were not directly predicted by emission factors from the
microchamber measurements, which suggests a mass transfer-based modeling
approach is needed for predicting TCPP concentrations from open-cell SPF. More
research is needed to determine how data from microchamber studies can be
used to predict exposures of residential occupants to emissions from SPF foam.

Keywords
spray polyurethane foam, TCPP, flame retardant, emissions

Introduction
Spray polyurethane foam (SPF) insulation increases building energy efficiency by
reducing conductive and convective heat losses through the building envelope and
is used in both new construction and retrofit applications. The U.S. government’s
efforts toward creating more efficient buildings, such as tax incentives and pro-
grams such as the U.S. Environmental Protection Agency (EPA) Energy Star and
Design for the Environment programs [1,2] have increased SPF applications. In
2012, over 61 million kg of SPF were installed in residential applications in the
United States [3], and the industry expects to see significant growth in the use of
their products over the next few years [4].
SPF is produced on site via an exothermic reaction of two sets of chemicals.
The A-side chemicals typically include methylene diphenyl diisocyanate. The
B-side chemicals include polyols, amines, blowing agents, surfactants, and flame
retardants. Flame retardants are present in the reacted polyurethane foam at up to
12 % by mass [5]. The reactions can be designed so the resulting foam is open cell
(low density) or closed cell (medium density). SPF can be installed using high-
pressure systems by professional contractors or with low-pressure do-it-yourself
kits. Whether the SPF is used to insulate a whole house by professionals or retrofit
by homeowners, there is potential for the building occupants to be exposed to SPF-
associated emissions.
Several studies have investigated emissions of flame retardants from SPF [6,7].
The most common identified flame retardant is tris(1-chloro-2-propyl) phosphate
(TCPP). TCPP is not exclusive to SPF and has also been used in mattresses, elec-
tronics, and upholstery [7]. TCPP has been measured in homes at airborne concen-
trations ranging from 2.4 ng m3 to 1,260 ng m3 [7–9] and found in similar
concentrations in cars, offices, and furniture stores [10]. TCPP is persistent in the
environment [11], readily absorbed through skin, and breaks down rapidly into
 POPPENDIECK ET AL., DOI 10.1520/STP158920150044 59

metabolites in the body [12]. Although there are limited data, TCPP is classified by
the EPA Design for the Environment program as having a high hazard for repro-
ductive and developmental effects [12].
The Consumer Product Safety Commission, along with the EPA, have received a
number of complaints regarding health effects that are potentially associated with SPF
applications. Residents have complained of health impacts, including severe respirato-
ry irritation, breathing difficulties, dizziness, and nausea, in the time frame of days to
months after SPF installation in a home. In some cases, the effects are so severe that
consumers report that they can no longer live in their homes [13]. This time frame of
health complaints is longer than the suspected time frame for the presence of isocya-
nates, suggesting that the emission of other chemicals (amines, blowing agents, surfac-
tants, flame retardants, or byproducts of the reactions) from the SPF may be of
concern. However, no direct connection between these health symptoms and SPF
emissions has yet to be established. In order to support such studies, standardized
measurement protocols are needed to determine emission rates of chemicals from SPF
as well as methodologies to relate those emission rates to occupant exposure.
This research was designed to contribute to the development and evaluation of
voluntary standards for testing emissions from SPF. Specifically, the work was per-
formed to support ASTM Subcommittee D22.05 on Indoor Air Quality efforts to stan-
dardize test methods and protocols relating to the measurement of emissions from SPF
using microenvironmental test chambers. Some of the measurement parameters in
this study varied from the values that the subcommittee is considering. Thus, compari-
sons to other data should note the specific experimental parameters employed.

Methods
TCPP concentrations were measured in two environments: microchambers and a resi-
dential test facility. The objective of the work was to quantify TCPP emissions from
various SPF samples and to better understand these emissions based on the measure-
ment results. The specific objectives were to (1) evaluate sampling times for TCPP in
various SPF formulations using the microchamber method; (2) use microchamber
data to assess whether TCPP emissions are likely controlled by internal material diffu-
sion or external mass transfer; and (3) compare TCPP emission rates from microcham-
ber experiments to emission measurements in a full-scale residential test facility.

FOAMS TESTED
Three different foams were tested (Table 1). Samples Open 1 and Closed 1 were pro-
vided by the American Chemistry Council’s Center for the Polyurethanes Industry.
These research formulations were developed in 2011 to be representative of SPF
then available in the marketplace. The formulations were created for research pur-
poses only and were not optimized to meet the specifications of commercial pro-
ducers and therefore may not reflect formulations currently available in the
marketplace. Therefore, conclusions about how these foams would perform outside
60 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

a laboratory setting are speculative. Foams Open 1 and Closed 1 were sprayed in
factory settings under controlled conditions. The foams were packaged and shipped
overnight to the National Institute of Standards and Technology (NIST) in an insu-
lated cooler in accordance with ASTM D7859, Standard Practice for Spraying, Sam-
pling, Packaging, and Test Specimen Preparation of Spray Polyurethane Foam (SPF)
Insulation for Testing of Emissions Using Environmental Chambers [14]. Emission
testing on Open 1 was started within 24 h of spraying. Emission testing on Closed 1
was delayed as noted in Table 1. Closed 1 was stored at room temperature (approxi-
mately 20 C) during storage.
Open 2 was a high-pressure, open-cell, low-density SPF that was applied during
the construction of the NIST Net Zero Energy Residential Test Facility (NZERTF)
in the summer of 2012. The NZERTF was built to support the development and
adoption of cost-effective net-zero energy designs and technologies, construction
methods, and building codes. The design and construction of the NZERTF are
described in Pettit et al. [15]. The NZERTF is a two-story detached house with an
unfinished basement and attic within the building thermal envelope. The garage is
not attached. The house is similar in size (242 m2 for occupied floors and 485 m2
inside the building envelope, including the attic and basement) and aesthetics to
homes in the surrounding communities. To achieve the net-zero energy goals, several
technologies are employed, including a high-efficiency heat pump, a solar hot-water
system, a 10.2-kW photovoltaic system, and a heat recovery ventilator (HRV). To
comply with the outdoor air requirements in ASHRAE Standard 62.2-2010, Ventila-
tion and Acceptable Indoor Air Quality in Low-Rise Residential Buildings [16], the
HRV was sized to continuously deliver 137 m3 h1 of outdoor air. Special attention
was paid to the design and construction of a highly insulated and airtight building
envelope. Approximately 15 m2 of high-pressure open-cell SPF was used to insulate
the basement rim joists. The basement is unfinished, and the SPF is not covered by
any finishing material. The house has no carpet and is not furnished other than per-
manently installed cabinetry. Hence, if TCPP is present in the indoor air of the house
and not measured in the outdoor air, then it can likely be attributed to the SPF.
This work included measurements of airborne TCPP concentrations both in
the NZERTF and in a microchamber for the same foam. In January 2014, three
approximately 0.8-g samples were cored from SPF in the basement rim joists of the

TABLE 1 SPF samples tested.

Foam ID Type Density (kg/m3)a Spray Date Test Date

Open 1 Open cell 12 2/26/2014 2/27/2014

Open 2 Open cell 7 Summer 2012b 1/28/2014
Closed 1 Closed cell 30 11/4/2014 4/13/2015
a
Density determined by measured initial mass and approximate volume.
b
Open 2 sample was taken from the NZERTF.
 POPPENDIECK ET AL., DOI 10.1520/STP158920150044 61

NZERTF and tested in microchambers. In July and August 2014, airborne TCPP
concentrations were measured in the basement and first floor of the NZERTF.

MICROCHAMBER EXPERIMENTS
A cutting tool was machined to precisely cut SPF to fit tightly within Markes 250
microchambers (Fig. 1) according to ASTM D7859-13e1. Unless otherwise noted,
the chambers were operated at a temperature of 40 C with a 100-mL min1 airflow
of ultrahigh-purity air. The temperature of 40 C was chosen to be consistent with
other standards [17], increase emissions to reduce problems associated with low de-
tection limits, and capture the performance of foam in environments with elevated
temperatures such as attics and exterior wall cavities. To investigate the impact of
temperature on the TCPP emission rate, samples of Open 1 were also run at 28, 50,
and 60 C. The impact of flow was investigated by running samples of Open 1 at 50,
100, and 200 mL min1. The airflow contained an absolute humidity of 8.8 g m3
(equivalent to a relative humidity of 38 % at 25 C) for the Open 1 and 2 samples.
The airflow contained zero humidity for the Closed 1 sample based on discussions
within Subcommittee D22.05 on humidity values for a proposed standard.
R
TenaxV TA sorbent tubes were attached to the effluent flow of each microcham-
ber for 2 min to 4 h (0.2–24 L of sample) depending on the concentration of the flame
R
retardant. For samples longer than 2 h, a second TenaxV tube was placed in series
with the first tube and analyzed for breakthrough. For each foam, three microcham-
bers were run concurrently. The fourth chamber was typically used as a control. Each
foam was analyzed for at least 300 h. Prior to sampling, the tubes were spiked with an
internal standard (1.0 lL of 1.25-mg toluene D-8 mL1 of methanol) using a liquid
methanol solution injected into a Talboys standard heat block.

NZERTF AIR SAMPLING

The indoor air was sampled from the first floor and basement of the NZERTF for
TCPP over a period of 2 months in the summer of 2014. Temperature values shown

FIG. 1 Sampling of foam Open 1 for microchamber analysis (left). Sample installed in
microchamber (right).
62 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

in the tables and figures are 12-h average readings from a thermocouple located in
the center of the open basement or in the open floor plan area of the first floor.
R
The NZERTF TCPP sampling involved two TenaxV sorbent tubes in series.
The first tube was used to quantify the TCPP concentration and the second to eval-
uate whether there was breakthrough through the first tube. If TCPP breakthrough
to the second tube was found, the data were not used. For each sampling event
three sets of tubes were prepared. Each tube set was sampled at 50 mL min1 using
a mass flow controller sampling system. Sampling times varied from 52 to 216 min
(average of 155 min). The tubes were separated and spiked with internal standard
(1.0 lL of 1.25-mg toluene D-8 mL1 of methanol).

TUBE ANALYSIS
Samples were analyzed using a thermal desorption gas chromatography-mass spec-
trometer system (GC-MS). A nonactivated guard column was used in the transfer
line from the thermal desorption unit to the GC-MS. An Rtx-5 amine column
(30.0 m by 250 lm by 0.50 lm) was used for compound separation in the GC-MS.
Table 2 and Table 3 document the operational settings used on the thermal desorp-
tion system and GC-MS, respectively. For all open-cell samples, each sample tube
was followed by a blank tube to check for carryover between samples.
In the tested foams, TCPP typically consisted of three isomers: tris(1-chloro-2-
propyl) phosphate (approximately 66 % based on GC-MS area response), bis(1-
chloro-2-propyl) (2-chloropropyl) phosphate (approximately 30 %), and (1-chloro-
2-propyl) bis(2-chloropropyl) phosphate (approximately 4 %). In this research,
only the first two isomers were consistently detected above quantification limits.
The response ratios of the three isomers on the tubes with TCPP and any carryover
on subsequent blanks were summed. The response ratio for each isomer was inte-
grated using a five-point standard curve. The total reported TCPP concentration
was determined as the sum of the isomer concentrations.
Instrument detection limits for TCPP were determined by multiplying three
times the standard deviation of seven replicates at a concentration that was less
than five times the determined method detection limit [18]. The instrument TCPP
detection limit was 8.65 ng, and the method detection limit was 0.71 to 2.86 lg m3
depending on the sample volume. Only values above the method detection limit for
the corresponding sampling volume are shown in the figures and tables.
R
TenaxV sorption tubes also captured amine catalysts, byproducts, and other
volatile organic compounds (VOCs). These other chemicals were analyzed in simi-
lar manners to those described for TCPP previously, and some are presented in the
“Results and Discussion” section to illustrate additional concepts and trends.

OTHER TCPP ANALYSES

To ensure that there were no sources of TCPP other than the SPF in the NZERTF
basement, small samples of a variety of materials from the NZERTF basement with
foam components were placed in a microchamber at 40 C and sampled for TCPP
 POPPENDIECK ET AL., DOI 10.1520/STP158920150044 63

TABLE 2 Thermal desorption system settings.

Phase and Condition Value

Purge
Prepurge time 1 min
Trap in line No
Split On
Flow rate 20 mL min1
Tube desorption
Time 8 min
Temperature 300 C
Split Off
Trap desorption
Trap low temperature 10 C
Trap high temperature 330 C
Trap hold time 3 min
Split On
Trap heating rate Maximum ( C s1)
Split flow rate 50 mL min1
Split ratios
Inlet No split
Outlet 27.3:1
Total 27.3:1
Other
Flow path temperature 200 C
GC cycle time 20 min

R
using the same TenaxV sorbent tubes and thermal desorption GC-MS analysis. The
sampled materials included rigid expanded polystyrene insulation, duct insulation,
and two varieties of pipe insulation. No TCPP was detected from any of these mate-
rials (method detection limit of 2.0 lg TCPP g1 material m3 air to 6.3 lg TCPP g1
material m3 air).

TRACER GAS MEASUREMENTS OF AIR CHANGE RATES

Air change rates in the NZERTF were measured by tracer gas decay (using sulfur
hexafluoride; SF6) following ASTM E-741, Standard Test Method for Determining
Air Change in a Single Zone by Means of a Tracer Gas Dilution [19], in July 2014.
These rates reflect the combination of mechanical HRV ventilation associated with
the HRV operation and infiltration due to building envelope leakage. Leakage rates
were determined by subtracting the measured HRV flow from the total measured
air change rate. An automated tracer gas system with SF6 injection and sampling at
multiple locations (eight indoor and one outdoor) was employed. Measurements
were made at each location once every 27 min after a 1-h mixing period. The esti-
mated uncertainty in the measured air change rates was 10 %.
64 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 3 GC-MS settings.

Phase and Condition Value

Helium flow
Pressure 10.2 psi
Flow 1.3 mL min1
Mode Constant flow
Average velocity 41 cm s1
Temperature profile
Initial time 2 min
Initial temperature 40 C
Ramp 15 C mL min1


Final time 3 min

Final temperature 315 C
Detector
Temperature 250 C
Mode SCAN

EMISSION RATE CALCULATIONS

TCPP emission rates were calculated from both the microchamber and NZERTF
experiments. For the microchamber, a single-zone mass balance model was used to
determine emission rates:

dC
Vchamber ¼ Qin Cin  Qin C þ Ef; chamber Achamber (1)
dt

where:
Qin ¼ airflow rate into the chamber, m3 h1,
Vchamber ¼ total chamber volume, m3,
C ¼ average chamber concentration, lg m3,
Cin ¼ concentration in the inlet air, lg m3,
Ef, chamber ¼ area-specific emission rate, lg m2 h1, and
Achamber ¼ surface area of the SPF sample, m2.
Sorption and desorption of TCPP to the exposed microchamber walls were
assumed to be in equilibrium at the time scales of sampling and are not included in
the data analysis. The green data points (Open Cell 2) in Fig. 2 show the TCPP con-
centration in the microchamber reaches a steady-state value of approximately 300
to 400 lg/m3 by the 4-h sample. This indicates that the exposed surface of the
microchamber is in equilibrium with TCPP by 4 h (sorption and desorption rates
cancel each other). Hence, for the time scales of sampling (daily), steady-state con-
ditions can be assumed and Eq 1 simplifies to:

Qin ðCss  Cin Þ

Ef; chamber ¼ (2)
Achamber
 POPPENDIECK ET AL., DOI 10.1520/STP158920150044 65

FIG. 2 TCPP concentrations from open- and closed-cell SPF tested in 40 C
microchambers. Error bars show standard errors in triplicate data.

The NZERTF basement was modeled using a single-zone mass balance with no net
sorption or reaction losses. The exposed SPF in the NZERTF was sprayed at least
2 years prior to sampling. In this time frame it is likely that the building materials
reached an equilibrium with TCPP at the ambient temperatures (20–22 C) at which
the house was operated for 2 years before sampling. Hence, sorption and desorption
rates will likely cancel each other. For this model, basement air was exchanged with
the first floor and outside air. TCPP concentrations were measured on the first floor
and assumed to be constant:

dC
VBasement ¼ Q1st in C1st þ Qoutside in Cout  Q1st out C  Qoutside out C þ Ef NZERTF ASPF (3)
dt

where:
Q1st in ¼ airflow rate from the first floor to the basement, m3 h1,
Q1st out ¼ airflow rate from the basement to the first floor, m3 h1,
Qoutside in ¼ airflow rate from the outside to the basement, m3 h1,
Qoutside out ¼ airflow rate from the basement to the outside, m3 h1,
VBasement ¼ basement volume, m3,
C ¼ average indoor basement concentration, lg m3,
Cfirst ¼ TCPP concentration measured on the first floor, lg m3,
66 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

Cout ¼ outdoor TCPP concentration, lg m3,

Ef, NZERTF ¼ NZERTF area-specific emission rate (lg m2 h1), and
ASPF ¼ surface area of the SPF in the basement of the NZERTF, m2.

The outdoor TCPP concentration (Cout) was below detection limits.

The airflow between the basement, the outdoors, and the rest of the house is
dominated by the HVAC system. The HVAC total supply airflow rate into the base-
ment, based on airflow rate measurements in the system, was 103 m3 h1. All the
return vents for the HVAC system are located on the first and second floors. The
infiltration rate for the entire house measured by SF6 tracer decay was approximate-
ly 30 m3 h1, with some dependence on outdoor weather conditions. Given the ra-
tio of the external surface area of the basement to the entire house and that Qoutside
in is part of the infiltration rate for the entire house, Qoutside in is at least an order of
magnitude smaller than the total supply airflow rate to the basement. As a result,
for the purpose of emission calculations, Qoutside out and Qoutside in were assumed to
be zero. Given that there are no HVAC return vents in the basement, Q1st out can be
assumed to equal the airflow from the HVAC supply. Hence, Q1st in and Q1st out
were set equal to the total supply airflow rate to the basement adjusted for the
system-operating fraction during the TCPP sampling period. Based on these
assumptions, the area-specific emission rate for foam in the NZERTF at steady state
(Ef, NZERTF) can be calculated from Eq 3:

Q1st in
Ef; NZERTF ¼ ðCss  C1st Þ (4)
ASPF

Results and Discussion

All reported concentration and emission rates in this document only apply to the
tested conditions and foams. The mass transfer processes in building systems may
not be the same as those in a microchamber. Hence, emission rates in the two sys-
tems may be different. TCPP was detected emitting from all tested foams.

MICROCHAMBER EXPERIMENTS
A series of microchamber experiments was conducted to determine the impact of
foam type, chemical type, flow rate, temperature, and humidity.

Influence of the Foam

For the tested open-cell foams (Open 1 and Open 2), concentrations of the flame
retardant TCPP tended to be constant over time throughout the duration of the
experiments (Fig. 2). The two open-cell foam TCPP concentrations were not statisti-
cally different (p ¼ 0.06), even though Open 1 was freshly sprayed and Open 2 was
applied over 1.5 years before sampling. All statistical comparisons in this document
used a one-way ANOVA, Tukey-Kramer analysis with a ¼ 0.05. The average TCPP
 POPPENDIECK ET AL., DOI 10.1520/STP158920150044 67

concentration over more than 400 h of sampling for Open 1 was 400 lg m3
(n ¼ 98; SE ¼ 23 lg m3) and 314 lg m3 (n ¼ 78; SE ¼ 23 lg m3) for Open 2.
The Open 2 results show that for this sample and for insulation temperatures of
40 C, flame retardants can be emitted at measurable concentrations for time frames
greater than 18 months after application.
Unlike the tested open-cell foams, the TCPP concentrations from the closed-cell
sample (Closed 1) decreased exponentially during the first 100 h (Fig. 2). Despite the
exponential decrease in TCPP concentration emitted from the Closed 1 sample, the total
TCPP mass was not appreciably depleted. Over the course of the 400-h experiment, less
than 8 lg of TCPP was emitted from the approximately 300,000 lg of TCPP present ini-
tially in the 3.75-g SPF sample (TCPP was approximately 8 % of the mass of ingredients
used to make Closed 1) [5]. One hypothesis to explain the decrease in the emission rate
is that the TCPP initially located near the surface emits over a short time frame and that
long-term emission from the bulk of the foam is limited by the diffusion of TCPP to the
surface. The two different emission profiles for TCPP from the open- and closed-cell
foams suggest that the limiting mass transport mechanism for TCPP is fundamentally
different for the two types of tested foams. It is possible that the emission of TCPP from
open-cell foam is controlled by the mass transfer coefficient of the airflow above the
foam surface (flow dependent), whereas the emission from the closed-cell foam is
controlled by the diffusion of TCPP through the closed cells. These data highlight the
importance of determining foam-specific emission parameters (initial concentration,
diffusion coefficient, partitioning coefficient, and mass transfer coefficient) for each
chemical before attempting to model the TCPP in full-scale systems.
For both open- and closed-cell foams, it seems that the concentrations of TCPP
did not appreciatively change after approximately 100 h. For microchamber experi-
ments, sampling before 100 h should be sufficient to determine differences in TCPP
emissions from different samples of SPF.

Influence of the Chemical

Fig. 2 illustrates that the type of foam will influence the emission profile and likely
the controlling mass transfer mechanism of a given chemical from SPF. In addition,
the emission profile may be different for different chemicals in a single foam. Fig. 3
shows a decaying concentration for bis(2-dimethylaminoethyl) ether (BDMAEE)
and the steady emission profile for TCPP from the Open 1 sample. BDMAEE is a
catalyst with an initial maximum concentration in this foam of less than 1 %, much
smaller than the foam concentrations of TCPP [5]. Over the course of the 400-h ex-
periment, more than 5,000 lg of BDMAEE was emitted from the approximately
7,200 lg of BDMAEE present initially in the 0.8-g SPF sample (BDMAEE was ap-
proximately 0.9 % of the mass of ingredients used to make Open 1) [5]. Hence, the
depletion of BDMAEE may have dominated the emission profile, whereas the large
source of TCPP resulted in no reduction in the TCPP concentration.
Fig. 4 illustrates emissions from the Closed 1 sample. The 1,4-dioxane, 1,2-
dichloropropane, and TCPP concentrations all decayed rapidly in the first 100 h
68 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 3 TCPP and BDMAEE concentrations from open-cell SPF tested in 40 C
microchambers. Error bars show standard errors in triplicate data.

before decreasing at a slower rate. Whether this decay for 1,4-dioxane and 1,2-
dichloropropane illustrates source depletion or diffusion limitation is unknown be-
cause determining the initial concentrations in the foam was beyond the scope of
this research. Salthammer, Fuhrmann, and Uhde [20] identified 1,2-dichloropro-
pane as a degradation product of TCPP in soft polyurethane foam. Regardless of
the exact mechanisms, Fig. 3 and Fig. 4 illustrate that the emissions of chemicals
from a specific foam can be controlled by a range of mechanisms and should not be
assumed to be consistent across the range of chemicals present in the foam.

Influence of the Microchamber Flow Rate

TCPP has a wide range of reported values for vapor pressures [21,22]. Depending
on definitions and reported chemical properties, TCPP could be defined as either a
VOC or semivolatile organic chemical (SVOC). The European Union Paints Direc-
tive [23] defines a VOC as any chemical with a boiling point below 250 C. TCPP
has reported boiling points ranging from 235 to >270 C [21,24]. In general, the
emissions of VOCs from materials are controlled by the diffusion of the VOC
through the material matrix and are not highly chamber dependent. This can allow
the direct use of VOC chamber emission factors to predict emissions in full-scale
buildings. In contrast, the emission of SVOCs tends to be chamber dependent and
 POPPENDIECK ET AL., DOI 10.1520/STP158920150044 69

FIG. 4 TCPP, 1,2 dichloropropane, and 1,4 dioxane concentrations from closed-cell SPF
tested in 40 C microchambers. Error bars show standard errors in triplicate
data.

controlled by the airflow above the material. Because the airflow above a material
can be significantly different in a chamber and in a full-scale building, SVOC cham-
ber emission factors from microchamber studies should not be directly applied to
full-scale buildings. Emission parameters (mass transfer coefficient, initial concen-
tration, partition coefficient, and diffusion coefficient) are needed to accurately pre-
dict the concentrations of SVOCs in buildings.
A series of experiments was conducted to determine whether the flow rate
(a surrogate for the mass transfer coefficient) influenced the emission rate
(lg m2 h1) of TCPP from open-cell foam. Samples of Open 1 foam were run at
50, 100, and 200 mL min1. Fig. 5 shows the TCPP concentrations at the three flow
rates. There was a significant difference between the emission rate for TCPP con-
centrations at 50 compared with 200 mL min1 (p ¼ 0.0001) and 100 compared
with 200 mL min1 (p ¼ 0.0033). Ni, Kumagai, and Yanagisawa [25] used a passive
sampler to demonstrate that the emission rate of TCPP from wallpaper depended
upon the diffusion length to the passive sampler. This indicates that the TCPP
emission rate from the wallpaper depended primarily on external mass transfer
limitations and not internal diffusion within the wallpaper, which suggests that
the flow field around the SPF may affect the TCPP concentration measured.
70 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 5 Average TCPP emission rates for experiments run at increasing flow rates. Error
bars show standard errors in triplicate data.

Combined with the data in Fig. 5, this suggests that TCPP is likely behaving as
an SVOC in open-cell foam, and direct predictions of building-scale TCPP con-
centrations using microchamber emission factors will not be accurate. Therefore,
building-scale TCPP modeling must consider the mass transport parameter prop-
erties of the foam.

Influence of the Microchamber Temperature

In field applications, SPF experiences a range of temperatures. Hence, it is impor-
tant to understand the temperature dependence of TCPP emissions. A triplicate set
of Open 1 SPF was analyzed at three temperatures (28, 40, and 50 C) without re-
moving the foam from the chambers. All samples were taken at least 24 h after the
temperature change. Temperatures typically reached steady-state values within 1 h
of the change. Over the course of the 550-h experiment, less than 1,400 lg of TCPP
was emitted from the approximately 140,000 lg of TCPP present initially in the
1.2-g SPF sample (TCPP was approximately 12 % of the mass of ingredients used to
make Open 1). A separate test on a separate sample of Open 1 was conducted at
60 C. Fig. 6 shows the average emission rates for TCPP at the four temperatures for
Open 1 foam. For comparison purposes, data from Ni, Kumagai, and Yanagisawa
[25] of measured TCPP emission rates from wallpaper at various temperatures
using a passive sampling system are also shown. These data indicate that TCPP
 POPPENDIECK ET AL., DOI 10.1520/STP158920150044 71

FIG. 6 Average TCPP emission rates for experiments run at increasing temperatures.

emissions are exponentially dependent upon temperature for open-cell foam.

Hence, a small change in building temperature may have a relatively large impact
on the TCPP concentration in the building.

Influence of the Microchamber Humidity

A sample of Open 1 was tested at 40 and 60 C with and without water vapor present.
The foam was sampled 7 times over 6 days in triplicate with 0 % relative humidity.
The same foam was also sampled 4 times over 3 days with an absolute humidity of
9.4 g m3 (SE ¼ 0.1 g m3; relative humidity ¼ 40.6 % at 25 C). The foam was then
raised to 60 C with an absolute humidity of 8.9 g m3 (SE ¼ 0.1 g m3; relative
humidity ¼ 38.4 % at 25 C) and sampled over 5 days. The foam was then tested at
0 % relative humidty over 7 days. The average TCPP concentration with humidity
present was higher than without humidity (Table 4). The difference was not statisti-
cally significant at 40 C but was significant at 60 C (p ¼ 0.009).

TABLE 4 Impact of humidity on average TCPP concentration in microchambers.

Temperature Absolute Standard Error Average TCPP Standard Error Number of

( C) Humidity (g m3) (g m3) Concentration (mg m3) (mg m3) Samples

40 0 NA 308 16 21
40 9.4 0.1 355 16 12
60 0 NA 2,630 96 25
60 8.9 0.1 3,500 156 21
72 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

NZERTF MEASUREMENTS
A series of measurements was conducted in the NZERTF to determine the source
of the TCPP and whether microchamber emission rates could be directly used to
predict TCPP concentrations in building-scale environments.

Source of TCPP
The NZERTF has approximately 15 m2 of open-cell SPF sprayed to seal the base-
ment rim joists. The SPF is exposed to the basement air with no covering finish
material. The basement ceiling is not finished, and the basement has no internal
dividing walls. SPF concentrations were measured over a period of 43 days in the
summer of 2014, 2 years after the foam was installed. TCPP concentrations were
measured in both the first floor and the basement (Table 5). Average TCPP concen-
trations were significantly higher in the basement than the first floor (p ¼ 0.008)
despite the fact that the temperature was higher in the first floor. These results sug-
gest that the source of the TCPP is located in the basement of the NZERTF. If the
TCPP source was uniformly distributed throughout the NZERTF, one would expect
higher TCPP concentrations in the first floor associated with the higher tempera-
tures (as seen in Fig. 6 for microchamber data).
TCPP can often be found as a flame retardant in furniture and insulation mate-
rials. However, as noted earlier, there is no furniture in the NZERTF. Five samples
of insulation materials from the walls, pipes, and ductwork were tested at 40 C in
the microchamber, and no TCPP was detected in any of the materials. Given the
higher concentrations of TCPP in the basement and the fact that no other sources
of TCPP were found, it can be reasonably concluded that the TCPP measured in
the upstairs living area was the result of emissions from the 15 m2 of open-cell SPF
sprayed in the rim joists located in the basement.
TCPP emissions are temperature dependent. The temperature in the NZERTF
was raised for 4 days. The average TCPP concentration at the elevated temperature
(28.5 C) was 3.4 times higher than the average TCPP concentration at typical base-
ment temperatures (21.5 C) (Fig. 7).

Comparison of Microchamber Emission Rates to NZERTF Emission Rates

Emission rates were estimated for the SPF in the basement of the NZERTF using
the assumptions described in the “Methods” section. For the ambient temperature
days (21.5 C), the estimated TCPP emission rate was 6.7 lg m2 h1. A sensitivity

TABLE 5 Average TCPP concentrations measured in the NZERTF.

Location Average Number of Average TCPP Standard Error

Temperature ( C) Samples Concentration (mg m3) (mg m3)

First floor 23.7 9 1.5 0.1

Basement 21.0 12 2.8 0.1
 POPPENDIECK ET AL., DOI 10.1520/STP158920150044 73

FIG. 7 Average TCPP concentration in the basement of the NZERTF at various

temperatures. Error bars show standard errors in triplicate data.

analysis on the emission rate calculation, based on varying the estimated airflows
by 50 %, resulted in a variation of emission rates from 3.3 to 10 lg m2 h1. These
NZERTF emission rates are 4 to 12 % of the emission rates predicted in the micro-
chamber at 21.6 C (80 lg m2 h1) (Fig. 8). A previous study found TCPP emission
rates in the range of 50 to 140 lg m2 h1 for one-component foams tested in 0.02-
m3 test chambers operated at 23 C, a 0.5-h1 air change rate, and 50 % relative hu-
midity [6]. However, that SPF was not 2 years old prior to testing.
Taken together, the data in Fig. 5 and Fig. 8 illustrate that the emission rates for
TCPP from SPF are a function of airflow conditions, which indicates that gas-phase
mass transfer limitations are likely controlling the release of TCPP from the foam.
Because flow conditions in the microchamber and real building spaces are different,
TCPP emission rates from the microchamber should not be used to directly predict
concentrations in real-world environments. TCPP emission rates from SPF will
likely be more accurately predicted using mass transfer-based approaches that
include foam-specific measurements of mass transfer parameters.

LIMITATIONS AND IMPLICATIONS

The primary purpose of this effort was to collect information to support the develop-
ment of protocols for using microchambers for evaluating emissions from SPF,
including the development of ASTM standards. The data from this work only apply
to the tested foams. Foam with different constituents or applied in a different manner
may have different emission profiles. Each foam should be tested in microchambers
before drawing conclusions about its emissions. Emission factors from this work
74 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 8 Comparison of TCPP emission rates in the basement of the NZERTF and
microchambers. The error bars on the basement emission values show a 50 %
variation in interzone flow rates.

should not be used to predict emissions from other chambers or to predict real-world
exposures until scaling between microchambers and other systems has been estab-
lished for chemicals similar to those in this study. Finally, this work should be repli-
cated at other laboratories to demonstrate the consistency of the methods employed.
Despite the limitations of this work, however, several conclusions can be drawn
from these results:
• TCPP emissions from SPF are temperature, flow, and foam dependent. Studies
done at lower temperatures and flow rates may not quantify emissions of
chemicals for which the concentrations are below or close to the detection
limits.
• Microchamber data can be used to compare emission profiles from various
foams, but TCPP microchamber emission rates cannot yet be directly applied
to full-scale emissions in a building.
• Occupants may be exposed to measureable concentrations of TCPP 2 years
after application of open-cell foam.

ACKNOWLEDGMENTS
This work was funded in part by the U.S. Consumer Product Safety Commission un-
der Interagency Agreement CPSC-I-13-0016.
 POPPENDIECK ET AL., DOI 10.1520/STP158920150044 75

Disclaimer
Certain trade names or company products are mentioned in the text to specify
adequately the experimental procedure and equipment used. In no case does such iden-
tification imply recommendation or endorsement by NIST or imply that the equipment
is the best available for the purpose.

References

[1] U.S. EPA Energy Star, https://www.energystar.gov (accessed May 14, 2015).

[2] U.S. EPA Design for the Environment Alternatives Assessments, http://www2.epa.gov/
saferchoice/design-environment-alternatives-assessments (accessed May 15, 2015).

[3] Center for the Polyurethanes Industry, 2012 End-Use Market Survey on the Polyur-
ethanes Industry in the United States, Canada, and Mexico, American Chemistry Council,
Washington, DC, 2013.

[4] Kouteren, S. V., “SPF Insulation Demand Growth Creating New Investment Opportuni-
ties; Principia Forecasts 15 % Annual Growth Through 2016,” http://www.businesswire.
com/news/home/20130717005300/en/SPF-Insulation-Demand-Growth-Creating-Investment-
Opportunities#.U6NATPldV8G (accessed June 19, 2016).

[5] Sebroski, J. R., Research Report for Measuring Emissions from Spray Polyurethane Foam
(SPF) Insulation, Center for the Polyurethanes Industry, Pittsburgh, PA, 2012, p. 52.

[6] Kemmlein, S., Hahn, O., and Jann, O., “Emissions of Organophosphate and Brominated
Flame Retardants from Selected Consumer Products and Building Materials,” Atm. Environ.,
Vol. 37, No. 39–40, 2003, pp. 5485–5493.

[7] Marklund, A., Andersson, B., and Haglund, P., “Organophosphorus Flame Retardants and
Plasticizers in Air from Various Indoor Environments,” J. Environ. Monitor., Vol. 7, No. 8,
2005, pp. 814–819.

[8] Bergh, C., Torgrip, R., Emenius, G., and Ostman, C., “Organophosphate and Phthalate
Esters in Air and Settled Dust—A Multi-Location Indoor Study,” Indoor Air, Vol. 21, No. 1,
2011, pp. 67–76.

[9] Saito, I., Onuki, A., and Seto, H., “Indoor Organophosphate and Polybrominated Flame
Retardants in Tokyo,” Indoor Air, Vol. 17, No. 1, 2007, pp. 28–36.

[10] Hartmann, P. C., Burgi, D., and Giger, W., “Organophosphate Flame Retardants and Plas-
ticizers in Indoor Air,” Chemosphere, Vol. 57, No. 8, 2004, pp. 781–787.

[11] Möller, A., Sturm, R., Xie, Z., Cai, M., He, J., and Ebinghaus, R., “Organophosphorus Flame
Retardants and Plasticizers in Airborne Particles over the Northern Pacific and Indian
Ocean Toward the Polar Regions: Evidence for Global Occurrence,” Environ. Sci. Tech.,
Vol. 46, No. 6, 2012, pp. 3127–3134.

[12] U.S. EPA, “Flame Retardants Used in Flexible Polyurethane Foam: An Alternatives As-
sessment Update,” https://www.epa.gov/sites/production/files/2015-08/documents/
ffr_final.pdf (accessed May 13, 2016).
76 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

[13] U.S. Consumer Product Safety Commission, http://www.saferproducts.gov/Search/

Result.aspx?dm¼0&q¼Sprayþfoamþinsulation&srt¼0 (accessed January 28, 2014).

[14] ASTM D7859-13e1, Standard Practice for Spraying, Sampling, Packaging, and Test Speci-
men Preparation of Spray Polyurethane Foam (SPF) Insulation for Testing of Emissions
Using Environmental Chambers, ASTM International, West Conshohocken, PA, 2013,
www.astm.org

[15] Pettit, B., Gates, C., Fanney, A. H., and Healy, W., Design Challenges of the NIST Net Zero
Energy Residential Test Facility, National Institute of Standards and Technology, Gai-
thersburg, MD, 2014.

[16] ASHRAE 62.2-2010, Ventilation and Acceptable Indoor Air Quality in Low-Rise Residen-
tial Buildings, ASHRAE, Atlanta, GA, 2010.

[17] ULC S718, Site Quality Assurance Program for Spray Polyurethane Foam, ULC, Ottawa,
ON, Canada.

[18] U.S. EPA, “Definition and Procedure for the Determination of the Method Detection Lim-
it—Revision 1.11,” https://www.gpo.gov/fdsys/pkg/CFR-2011-title40-vol23/pdf/CFR-
2011-title40-vol23-part136-appB.pdf (accessed May 13, 2016).

[19] ASTM E741-00, Standard Test Method for Determining Air Change in a Single Zone by
Means of a Tracer Gas Dilution, ASTM International, West Conshohocken, PA, 2000,
www.astm.org

[20] Salthammer, T., Fuhrmann, F., and Uhde, E., “Flame Retardants in the Indoor Environ-
ment—Part II: Release of VOCs (Triethylphosphate and Halogenated Degradation Prod-
ucts) from Polyurethane,” Indoor Air, Vol. 13, No. 1, 2003, pp. 49–52.

[21] ATSDR, Toxicological Profile for Phosphate Ester Flame Retardants, U.S. Department of
Health and Human Services, Atlanta, GA, 2012, p. 250.

[22] Verbruggen, E. M. J., Rila, J. P., Traas, T. P., Posthuma-Doodeman, C. J. A. M., and Posthu-
mus, R., “Environmental Risk Limits for Several Phosphate Esters, with Possible Applica-
tion as Flame Retardant,” http://rivm.openrepository.com/rivm/bitstream/10029/7383/
1/601501024.pdf (accessed May 13, 2016).

[23] EU Directive 2004/42/CE of the European Parliament and of the Council, http://faolex.-
fao.org/docs/pdf/eur43014.pdf (accessed May 13, 2016).

[24] WHO, “Environmental Health Criteria 209: Flame Retardants: Tris(Chloropropyl Phos-
phate and Tris(2-Chloroethyl) Phosphate,” http://www.who.int/ipcs/publications/ehc/
who_ ehc_209.pdf (accessed May 13, 2016).

[25] Ni, Y., Kumagai, K., and Yanagisawa, Y., “Measuring Emissions of Organophosphate
Flame Retardants Using a Passive Flux Sampler,” Atm. Environ., Vol. 41, No. 15, 2007,
pp. 3235–3240.
 DEVELOPING CONSENSUS STANDARDS FOR MEASURING CHEMICAL EMISSIONS 77

STP 1589, 2017 / available online at www.astm.org / doi: 10.1520/STP158920150038

Doyun Won,1 Angelika Zidek,2 Gang Nong,1 and

Ewa Lusztyk1

Glass Chamber Method for

Screening of 4,40 -MDI and TCPP
Emissions from Foam Joint
Sealant
Citation
Won, D., Zidek, A., Nong, G., and Lusztyk, E., “Glass Chamber Method for Screening of 4,40 -
MDI and TCPP Emissions from Foam Joint Sealant,” Developing Consensus Standards for
Measuring Chemical Emissions from Spray Polyurethane Foam (SPF) Insulation, ASTM
STP1589, J. Sebroski and M. Mason, Eds., ASTM International, West Conshohocken, PA, 2017,
pp. 77–97, http://dx.doi.org/10.1520/STP1589201500383

ABSTRACT
One-component foam joint sealant is widely used to seal air leaks by
homeowners. Various chemicals can be emitted from a foam joint sealant.
The goal of this study was to develop a glass chamber method to examine
the emissions of 4,40 -methylenediphenyl diisocyanate (4,40 -MDI) and tris(2-
chloroisopropyl)phosphate (TCPP) from foam joint sealants. The concentrations
of 4,40 -MDI and TCPP were measured during a 24-h chamber test that
involved a 3-L chamber operated at 40 C, 20 % relative humidity (RH),
22.2 air changes per hour, and 2 sampling media (glass filter coated with
9-methylaminomethyl anthracene for 4,40 -MDI and sorbent tube filled with
TenaxV TA for TCPP). The 4,40 -MDI concentration peaked within 11 min and
R

decayed to below the lowest limit of quantification within 1 h. The TCPP

concentration reached a maximum value at approximately 4 h and decayed
relatively slowly or stayed almost constant afterward. The 4,40 -MDI concentration
after applying foam joint sealant to all the windows of a small house was
predicted to be much lower than the reference value of 0.6 lg/m3 by the U.S.

Manuscript received May 12, 2015; accepted for publication October 5, 2015.
1
National Research Council Canada, Intelligent Building Operations, 1200 Montreal Rd., Ottawa, ON, Canada, K1A 0R6
2
Health Canada, Existing Substances Risk Assessment Bureau, 269 Laurier Ave. West, Ottawa, ON, Canada, K1A 0K9
3
ASTM Symposium on Developing Consensus Standards for Measuring Chemical Emissions from Spray
Polyurethane Foam (SPF) Insulation on April 30–May 1, 2015 in Anaheim, CA.

Copyright V
C 2017 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959.
78 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

Environmental Protection Agency. Conversely, the higher predicted concentrations

of TCPP than the measured indoor levels may imply the potential of foam joint
sealants as an important source of TCPP in homes. In addition, the test with
4,40 -MDI for method optimization showed that the effect of environmental
factors (temperature and RH) as well as the sink effect by interior surfaces could
be significant. When a test method is standardized for 4,40 -MDI emissions, these
influential factors should be investigated.

Keywords
4,40 -methylenediphenyl diisocyanate, tris(2-chloroisopropyl)phosphate, one-
component foam joint sealant, chamber method, CAS No. 101-68-8, CAS No.
13674-84-5

Introduction
4,40 -Methylenediphenyl diisocyanate (4,40 -MDI) belongs to a family of highly reac-
tive isocyanate compounds known for respiratory disorders and dermatitis [1,2]. It
is classified as a Category 1 respiratory sensitizer and Category 1 skin sensitizer [3].
It is used in the industrial production of rigid polyurethane foams as well as in the
fields of coatings, adhesives, sealants, and elastomers [4]. Methylenediphenyl diiso-
cyanates may also be used in do-it-yourself products such as adhesives, insulation
foam, and sealants [5]. Consumer products that are sold in an uncured applied
form (e.g., spray polyurethane foam) may cause health concerns if the unreacted
isocyanates become airborne and consumers and bystanders are exposed [6].
Monomeric or “pure” MDI usually refers to the 4,40 -methylenediphenyl diiso-
cyanate isomer (CAS No. 101-68-8) [7]. However, the principal material of com-
merce is not pure 4,40 -methylenediphenyl diisocyanate but is a mixture containing
4,40 -methylenediphenyl diisocyanate, other methylene diphenyl diisocyanate iso-
mers (2,40 - and 2,20 -MDI), and low molecular oligomers [4]. Polymeric MDI (CAS
No. 9016-87-9) is typically composed of approximately 45 to 50 % of the 4,40 -MDI
monomer, approximately 1 to 2 % of 2,40 -MDI, and varying amounts of MDI
oligomers with n ¼ 1 to approximately 10 [7,8]. Polymeric MDIs, which are lower
in cost and easier to handle than MDI monomers, are widely used as adhesives in
the foundry core binder area, in oriented strand board or particle board, and
between rubber products and fabric or cord [9].
Sealing air leaks is a practical energy conservation measure that homeowners
can use as a do-it-yourself project. One-component polyurethane-based expandable
foam, which is also known as foam joint sealant (FJS), is one of many caulking
products that can be used for this purpose [10]. These FJSs can contain MDI.
Despite the increased usage of such consumer products containing MDI, there has
been little research on emissions of MDI during the application of FJSs.
FJSs can also contain tris(2-chloroisopropyl)phosphate (TCPP), which is a
flame retardant mostly used in rigid polyurethane foams and is a possible
 WON ET AL., DOI 10.1520/STP158920150038 79

carcinogen [11,12]. As with MDI, TCPP has been prioritized for risk assessment
under the Chemicals Management Plan in Canada [13,14]. While TCPP levels have
been measured in indoor environments [15–17] and in several indoor sources
[11,18], there has been limited information on the emission of TCPP from foam
joint sealants.
The purpose of this research was to develop a chamber method that can exam-
ine the emissions of 4,40 -MDI and TCPP from consumer/building products. The
method was applied to measure emissions of 4,40 -MDI and TCPP from three com-
mercial FJS products.

Method
TEST MATERIALS
Three FJSs were purchased from local retail outlets in Ottawa, Canada. According
to the material safety data sheet, the 4,40 -MDI contents were 5 to 10 % (FJS1 and
FJS2) and 7 to 13 % (FJS3). TCPP (5–10 %) was identified only in one product
(FJS3). Approximately 10 g of a joint sealant was sprayed into a Petri dish (diame-
ter, 6 cm; 0.0064 m2) for testing. After application, the specimen was tested immedi-
ately in a chamber as described in the next section.

CHAMBER TEST AND CONDITIONS

The emissions testing was conducted over 24 h at 40 C, 20 % relative humidity
(RH), and 22.2 air changes per hour in a cylindrical glass chamber (approximately
3 L), which is similar to that in Wirts and Salthammer [19]. The main difference
between the two methods is the location of the sampling filter, which was inside of
the chamber in Wirts and Salthammer and outside of the chamber in this study.
The setup in this study is likely to result in an emission rate lower than the actual
rate due to the sink effects (surface adsorption) associated with 4,40 -MDI’s low vola-
tility and reaction with water. Therefore, the sink rate was measured and taken into
account when the emission rate was estimated in this study.
The chamber was enclosed in an environmental chamber maintained at 40 C,
which was the temperature chosen based on the observation that only small emis-
sions of MDI isomers were seen at typical room temperatures of 20 to 30 C [20].
The chosen temperature is within the range of the melting point of 4,40 -MDI and
therefore is expected to ensure adequate concentrations for the analytical method
used in this study. The emission rates measured at 40 C can be considered as a
worst-case scenario because the indoor temperature can be as high as 40 C in the
summer in buildings or homes without air conditioning. The effects of temperature
on 4,40 -MDI emissions were investigated with an 4,40 -MDI source.
The high air exchange rate was mainly due to the high sampling volume
required with the 4,40 -MDI analysis. It is also expected to promote better mixing
and to reduce the adsorption of 4,40 -MDI to the chamber wall. Because moisture
reacts with 4,40 -MDI, 0 % RH is likely to be the best choice for improving the test
80 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

sensitivity. However, the sealants tested in this study require moisture for curing.
Therefore, 20 % RH was chosen for the test because the absolute humidity of 20 %
at 40 C is similar to that of 50 % at 23 C (10.3 vs. 10.2 g/m3). The effects of RH on
4,40 -MDI emission were also determined in this study.

AIR SAMPLING AND CHEMICAL ANALYSIS

A total of 7 to 9 air samples were taken over 24 h for the test of each product. More
frequent samples were taken during the first 2 h to capture the peak emissions after
the application of the product. The sampling duration, which ranged between 4 and
200 min, was adjusted according to the expected concentration. The sampling rate
was 1.1 and 0.3 L/min for 4,40 -MDI and TCPP, respectively.
For the 4,40 -MDI sampling, glass filters coated with 9-methylaminomethyl an-
thracene (MAMA) were used. The cassette contained two 37-mm glass filters and
was prepared and analyzed at L’Institut de Recherche Robert-Sauvé en Santé et en
Sécurité du Travail. The collected 4,40 -MDI was desorbed with a solution of derivative
HMDI-MAMA (4,40 -diisocyanato-methylenedicyclohexane as internal standard) at
0.1 lg/mL dimethyl sulfoxide/acetonitrile (20 : 80) and then analyzed with a liquid
chromatography-mass spectrometry (LC-MS) system equipped with an electrospray
ionization source. The LC-MS system was an Agilent 1100 Series coupled with an Ion
Trap VL model mass spectrometer. The LC column was a Zorbax RX-C18 (150 mm
by 2.1 mm by 5 lm). The mobile phase was made up with LC-MS-grade acetonitrile
and a buffer of 2 mM ammonium acetate in water whose pH was adjusted to 3.0 with
acetic acid. The lowest limit of quantification was 0.47 ng for 4,40 -MDI, which was
determined by analyzing ten replicates of a solution of low concentration of 4,40 -MDI
representative of an extract generated by the full implementation of the method of
analysis. Therefore, the reporting limit was 0.75 ng per sample.
R
Tenax TA tubes were used for the TCPP sampling. TCPP on TenaxV TA tubes
was thermally desorbed using a Gerstel thermal desorption unit. The gas chroma-
tography (GC)-MS system was an Agilent 6890 GC equipped with an HP-5MS
capillary column (30 m by 0.25 mm by 0.25 lm) and a 5973N mass selective detec-
tor. The detection limit, which was determined in accordance with guidance pub-
lished by the U.S. Environmental Protection Agency (EPA) [21], was 3.21 ng for
TCPP.

Results and Discussion

SURFACE SINK EFFECTS
The sink effects were determined by measuring the 4,40 -MDI concentration before
and after the empty glass chamber (Fig. 1, #1 and #2). The 4,40 -MDI vapor was gen-
erated by passing the clean air stream through a glass impinger (Fig. 1, #3) contain-
ing three diffusion vials (5-mm capillary and 7.62-cm diffusion length) from VICI
Metronics. 4,40 -MDI (purity of 98 %) was placed inside each diffusion vial. The
chamber used for the sink test had a volume of 2.7 L and an air change rate of 26.7
 WON ET AL., DOI 10.1520/STP158920150038 81

FIG. 1 One-chamber setup with a glass impinger as a 4,40 -MDI source.

per hour. To obtain a quasi-steady state, the generation started about 1 day before
the sampling under the condition of 40 C and 20 % RH. The collection of paired
samples was conducted twice.
The measured upstream concentration was 0.055 lg/m3 on average, which was
71 % lower than the expected concentration in the chamber according to the diffu-
sion rate calculated by Eq 1 [22].

qd ¼ DMPA=ðLRTÞ ln½P=ðP  Pv Þ (1)

where:
qd ¼ emission (diffusion) rate, g/s,
D ¼ diffusion coefficient of the diffusing vapor, cm2/s,
M ¼ molecular weight of the diffusing vapor, g/mol,
P ¼ pressure—usually atmospheric—in the diffusion tube, mmHg,
A ¼ diffusion tube cross-sectional area, cm2,
L ¼ length of the diffusion tube, cm,
R ¼ molar gas constant, 62.363E3 mL mmHg/mol K,
T ¼ absolute temperature, K, and
Pv ¼ partial pressure of the diffusing vapor, mmHg.
There could be several potential reasons for the low upstream concentration.
The theoretical equation assumes that the concentration of vapor at the diffusion
vial exit is at nearly zero and the vapor inside the capillary is saturated [22]. If these
two assumptions were not met, the 4,40 -MDI generation rate could be reduced.
Second, the estimated vapor pressure (6.07E-5 mmHg) [23] and air diffusion coeffi-
cient (0.058 cm2/s) of 4,40 -MDI [24], which were required in the calculation, could
82 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

also contribute to the discrepancy. Third, the interior surfaces of the impinger and
the connecting tube before the upstream sampling location could act as sinks, lead-
ing to a reduced concentration at location #1. Fourth, the reduced concentration
might be due to the reaction between 4,40 -MDI and water vapor, for which more
detailed discussion is given in the next section.
The sink effect was assumed to be irreversible. This assumption is based on the
observation that no 4,40 -MDI was recovered from a chamber containing a piece of
memory foam that was exposed to 4,40 -MDI vapor for 48 h (adsorption stage) in the
same setup as the chamber sink test. At the end of the adsorption stage, the vapor
injection was stopped, and the chamber was flushed with MDI-free air for 30 h (desorp-
tion stage). Two samples were taken in sequence during the desorption stage for 6 and
24 h at 1.1 L/min. No 4,40 -MDI was detected in both samples. The strong and irre-
versible sink effects of 4,40 -MDI by chamber walls were also noted in Wirts et al. [20].
Assuming the irreversible sink rate is proportional to the chamber concentra-
tion, the mass balance in the chamber air becomes:

V dC=dt ¼ QCin  QC  ka Aa C (2)

where:
V ¼ chamber volume, 0.0027 m3,
C ¼ 4,40 -MDI concentration in the chamber, lg/m3,
Q ¼ chamber flow rate, 0.072 m3/h,
Cin ¼ 4,40 -MDI concentration of the air entering the chamber, lg/m3,
Aa ¼ surface area of the glass chamber, 0.137 m2, and
ka ¼ apparent sink rate coefficient, m/h.
The solution of Eq 2 is:
0 0
C ¼ NCin ½1  expðka tÞ=ka (3)

where:
0
ka ¼ N þ ka Aa =V and
N ¼ air change rate, 26.7 per hour.
Because the air change rate is large, the exponential term of Eq 3 becomes close to
zero within 1 h. Therefore, ka can be calculated using the following equation:

ka ¼ ðQ=Aa ÞðCin  CÞ=C (4)

The measured downstream concentration (0.017 lg/m3) was approximately 30 % of the

upstream concentration for both pairs of samples. This shows that approximately 70 %
of 4,40 -MDI was removed from the air phase inside the glass chamber, possibly by the
inner surfaces of the glass chamber. The resulting ka was estimated to be 1.22 m/h.

RH EFFECTS ON 4,40 -MDI EMISSIONS

To investigate the effect of RH on 4,40 -MDI emissions, the same diffusion vial con-
taining the MDI monomer was used at different RH levels (0, 20, and 40 %) and at
 WON ET AL., DOI 10.1520/STP158920150038 83

40 C. The main difference from the sink test was that one diffusion vial containing
the MDI monomer was placed inside the testing chamber (Fig. 2). The measured
4,40 -MDI concentrations are given in Fig. 3. The concentration decreased with in-
creasing RH. The concentration at 40 % RH was about 5 and 2 times lower than
that at 0 and 20 % RH, respectively.
MDI isomers are known to react with water, even atmospheric moisture, to
form carbon dioxide and an insoluble urea compound [20], particularly in the pres-
ence of catalysts/basic materials or at high temperatures (>50 C) [25]. However,
there have been conflicting findings with respect to the question of whether MDI
isomers can hydrolyze in humid air under atmospheric conditions. A significant
loss of toluene diisocyanate, which in general hydrolyzes more readily than
TM
4,40 -MDI in water, was observed in a Teflon -lined chamber at 27 C and between
7 and 70 % RH [26]. However, the loss was independent of RH levels and was most-
ly attributed to surface reactions due to a lack of evidence for reaction products
[26]. On the other hand, earlier studies reported reduced toluene diisocyanate con-
centrations with increasing absolute humdity at 24 C [27] and the presence of hy-
drolysis products such as toluene diamine [26].
Because the observations in this study clearly show that the 4,40 -MDI concen-
tration depended on the humidity level, the mass balance of 4,40 -MDI in the cham-
ber considered two removal mechanisms, i.e., adsorption to chamber surfaces and
reaction with water:

V dC=dt ¼ E  QC  ks Aa C  kw Cw C (5)

FIG. 2 Chamber setup with a diffusion vial containing the MDI monomer inside the
chamber.
84 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 3 4,40 -MDI chamber concentrations at three RH levels.

where:
E ¼ the emission rate of 4,40 -MDI from one diffusion vial, lg/h,
ks ¼ sink rate coefficient, m/h,
kw ¼ reaction coefficient with water vapor, m6/g h, and
Cw ¼ absolute water concentration in air [28], g/m3.
The reaction with water was assumed to be proportional to the concentration of
water vapor and 4,40 -MDI [29]. A new term ks was introduced in Eq 5 to differenti-
ate it from ka used in the previous section. It can be deduced that ka reflects the
combined effects of the surface adsorption and reaction with water. The relation
between ka and ks can be determined by comparing Eq 2 with Eq 5:

ka Aa ¼ ks Aa þ kw Cw (6)

The solution of Eq 5 is:

0 0
C ¼ E½1  expðkw tÞ=ðVkw Þ (7)

where:
0
kw ¼ N þ ks Aa =V þ kw Cw =V:

Because the air change rate is large, the exponential term of Eq 7 also becomes close
to zero within 1 h. Therefore, Eq 7 can be simplified as Eq 8. The left-side term of
Eq 8 is for the generation of 4,40 -MDI; the right-side terms are for the removal of
4,40 -MDI by ventilation, sink to chamber walls, and reaction with water.
 WON ET AL., DOI 10.1520/STP158920150038 85

E ¼ QC þ ks Aa C þ kw Cw C (8)

From two RH tests of 0 and 20 %, ka and kw of Eq 8 were estimated as 0.19 m/h and
0.014 m6/g h, respectively. The ka value obtained in the previous section was also
used in the calculation. The resulting emission rate was 0.021 lg/h. Assuming ks
was constant for different RHs, a second kw of 0.02 m6/g h was calculated from the
40 % RH test. The difference in two kws may stem from errors in the experiment
and/or assumptions used in Eq 5.
Eq 8 was divided by E to determine the relative importance of each removal mech-
anism, including ventilation (QC=E), adsorptive sink (ks Aa C=E), and reaction with wa-
ter (kw Cw C=E). Fig. 4 shows that the relative importance of the reaction with water was
0.59, which was highest among three removal mechanisms at 20 % RH and was further
increased to 0.81 at 40 % RH. The results imply that the RH level is an impotant factor
that needs to be standarized in a test method. However, it should be noted that the rela-
tion in Fig. 4 is valid only within the assumptions used in this study. More research is
required to confirm whether the increased removal of 4,40 -MDI with increasing RH is
due to the reaction between water vapor and 4,40 -MDI.

TEMPERATURE EFFECTS ON MDI EMISSIONS

The effects of temperature on 4,40 -MDI emissions were also investigated with a dif-
fusion vial containing the MDI monomer (Fig. 2). The 4,40 -MDI concentrations

FIG. 4 Relative importance of three removal mechanisms for MDI.

86 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

were measured while the temperature of the chamber was varied (23, 27, 32, 37, 40,
and 45 C) with an equilibrium time of approximately 1 day at each temperature.
RH was set at 20 % for all tests. Except at 23 C, the chamber concentrations were
above the detection limit. The natural logarithm of the concentration (Ln C) was
plotted against the reciprocal of temperature (1/T) in Fig. 5 to determine the enthal-
py of emission (DH) based on the Clausius-Clapeyron relation (Eq 9). The R2 value
of 0.9064 indicates that the relation can explain the temperature dependence of the
emission factor reasonably well.

C ¼ Ac expðDH=RTÞ (9)

where:
Ac ¼ a constant,
DH ¼ enthalpy to be supplied for the emission of the compound from the test
specimen (enthalpy of emission), kJ/mol,
R ¼ universal gas constant, 8.314E-3 kJ/K mol, and
T ¼ air temperature, K.
The resulting enthalpy (DH) was determined to be 150 kJ/mol, which is approxi-
mately 70 % greater than the enthalpy of vaporization of 91.1 kJ/mol estimated
for the MDI monomer [30] and that of 88.7 kJ/mol measured with an MDI
prepolymer-based resin [20]. The discrepancy may be explained partly by the dif-
ference between enthalpy of sublimation (DHs) and enthalpy of vaporization (DHv).

FIG. 5 Logarithm of C versus reciprocal of temperature.

 WON ET AL., DOI 10.1520/STP158920150038 87

4,40 -MDI was in solid form up to 40 C in this experiment, whereas the MDI resin
used in Wirts et al. [20] was in liquid form, and the correlations in Yaws [30]
mostly covered liquid state above 38.2 C. It takes more energy to excite a solid than
a liquid to its gaseous phase. However, the measured DH in this study is still higher
than the DHs of approximately 120 kJ/mol, which was obtained by the summation
of DHv and enthalpy of fusion (DHf) [31].
Despite the discrepancy, the temperature dependence is expected to be useful
for extrapolating the emission factors measured at higher temperatures to that at
room temperature. Using Eq 9, the ratio of the emission factor at 23 C to that at
40 C would be 0.04 for DH ¼ 150 kJ/mol and 0.07 for DHs ¼ 120 kJ/mol.

REPEATABILITY OF 4,40 -MDI AND TCPP

To test the repeatability of the 4,40 -MDI test, identical foam joint sealants purchased
on two different dates (FJS3_a in August 2012 and FJS3_b in December 2012) were
tested under the same conditions as those of the main test. Fig. 6 presents the emis-
sion factor normalized to the sample weight. The maximum difference was 25 %,
indicating the repeatability of the 4,40 -MDI test is acceptable.
TCPP concentrations were also measured from two FJS3s purchased on two
different dates (FJS3_a and FJS3_b). In addition, each product was tested one more
time (FJS3_a_2nd and FJS3_b_2nd). Fig. 7 shows the emission factors of TCPP nor-
malized to the specimen weight. The difference associated with the same product
tested twice was as high as 100 % at the beginning (t < 1 h), but it decreased to

FIG. 6 Chamber concentrations of 4,40 -MDI from two FJS3s purchased on two
different dates.
88 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 7 Chamber concentrations of TCPP from FJS3s purchased on two different dates.

approximately 23 % between 1 and 4 h and to <5 % after 4 h. The variation in the

early emission stage was likely due to the instability of the chamber conditions.
This indicates that the repeatability of the TCPP test is acceptable at steady state.
On the other hand, the differences from the products purchased on two different
dates were bigger, ranging between 25 and 45 % for the emission factors measured
after 4 h. This difference may stem from the different TCPP contents in different
batches of products.

CHAMBER CONCENTRATIONS OF 4,40 -MDI AND TCPP

Fig. 8 shows the chamber concentrations of 4,40 -MDI for three products. The cham-
ber concentration from foam joint sealants (FJS1 to 3) peaked within 11 min and
decayed to below the limit of quantification between 30 min and 1 h. The fast decay
is likely due to the decreasing 4,40 -MDI content caused by polymerization reactions
[20]. The observation that FJS3 produced the highest peak 4,40 -MDI air concentra-
tion (Fig. 8) is consistent with the fact that it had the highest 4,40 -MDI content of
the three sealants tested.
Fig. 9 presents the chamber concentrations of TCPP for FJS1 to 3. The chamber
concentration reached a maximum value at 4 h and decayed slowly afterward for
FJS1 and FJS2. Conversely, the concentration reached a plateau around 4 h and
maintained a relatively constant concentration afterward for FJS3. The maximum
concentration was 4.02 and 5.61 lg/m3 for FJS1 and FJS2, respectively. The maxi-
mum concentration for FJS3 was 256 lg/m3, which is greater than FJS1 and FJS2 by
 WON ET AL., DOI 10.1520/STP158920150038 89

FIG. 8 Chamber concentrations of 4,40 -MDI.

FIG. 9 Chamber concentrations of TCPP.

90 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

approximately 50-fold. This observation agrees with the content information found
on the material safety data sheet associated with these products. No information on
TCPP content was available for FJS1 and FJS2, likely because of its low content. On
the other hand, FJS3 was reported to contain 5 to 10 % TCPP. The relatively constant
TCPP concentrations of FJS3 compared to those of 4,40 -MDI reflect the fact that there
is no chemical reaction that will deplete TCPP over time as there is for 4,40 -MDI.

EMISSION FACTORS OF 4,40 -MDI AND TCPP

The mass balance of 4,40 -MDI in the foam phase can be described as follows:

Vf dCf =dt ¼ kf Cf Cwf  hm ðCf  CÞAf (10)

where:
Vf ¼ volume of the foam sealant, m3,
Cf ¼ 4,40 -MDI concentration in the foam, lg/m3,
Cwf ¼ water concentration in the foam, g/m3,
C ¼ 4,40 -MDI concentration in the air, lg/m3,
hm ¼ mass transfer coefficient between the foam surface and air, m/h,
kf ¼ reaction coefficient between 4,40 -MDI and water, m6/g h, and
Af ¼ surface area of the foam, m2.
Eq 10 is based on several assumptions introduced to simplify the complex reactions
during the curing stage. First, the 4,40 -MDI concentration was assumed to be
homogenous within the foam. Second, the reaction between free 4,40 -MDI and
water vapor within the foam was assumed to depend on the concentration of free
4,40 -MDI (Cf) and water vapor (Cw) [29]. Because moisture in the foam can come
from the air as well as from the substrate, the moisture concentration was assumed
to be abundant and relatively constant during a testing period. Third, the emission
of 4,40 -MDI from the foam surface was assumed to be proportional to the concen-
tration difference between the foam (Cf) and air phase (C).
Based on the reported 4,40 -MDI content of 5 to 13 % in each FJS, Cf was
assumed to be much bigger than C. Therefore, Eq 10 can be simplified accordingly
and solved as follows:

Cf ¼ Cf 0 expðk0f tÞ (11)

where:
Cf0 ¼ initial concentration of 4,40 -MDI in the foam, lg/m3, and
0
kf ¼ ðkf Cwf þ hm Af Þ=Vf :

Consequently, the following mass balance equation can be applied to the chamber
air phase:

V dC=dt ¼ QCin  QC  ks Aa C  kw Cw C þ hm ðCf  CÞAf (12)

 WON ET AL., DOI 10.1520/STP158920150038 91

Assuming Cf is much bigger than C, Eq 12 can simplified as:

V dC=dt ¼ QCin  QC  ks Aa C  kw Cw C þ Eo Af expðbtÞ (13)

where:
Eo ¼ hm Cf0, which is termed as the initial emission strength, lg/m2 h,
0
b ¼ kf , which is termed as the decay constant, 1/h, and
t ¼ elapsed time, h.
Solving Eq 13 with the assumption of the negligible concentration of the incoming
air (Cin approximately 0) leads to the following equation:

Eo L 0
C¼ ½expðbtÞ  expðN tÞ (14)
ðN 0  bÞ

where:
N0 ¼ N þ ks Aa/V þ kw Cw/V ¼ N þ ka La,
N ¼ air exchange rate of the chamber, 1/h,
L ¼ loading ratio of the foam (L ¼ Af/V), m2/m3,
La ¼ loading ratio of the chamber surfaces (La ¼ Aa/V), m2/m3,
Cw ¼ absolute water concentration in air [28], 10.2 g/m3,
ks ¼ sink rate coefficient, 0.19 m/h,
kw ¼ reaction coefficient with water, 0.014 m6/g h, and
ka ¼ apparent sink rate coefficient, 1.22 m/h.
To back-calculate Eo and b of Eq 14 from the measured data sets of t and C, a
curve-fitting technique based on nonlinear regression (SigmaPlot Version 11.0) was
used.
2
Table 1 presents the curve-fitting results. The R values >0.88 imply that the first-
0
order decay model can express the 4,4 -MDI emission rates reasonably well. Wirts et al.
[20] reported emission factors ranging from 200 to 1,000 lg/m2 h for four one-
component moisture-curing adhesives. The lower level agrees with Eo of FJS2 and FJS3.
Because the emission rate of TCPP of FJS3 appears to be relatively constant
during the test period of 24 h, it was considered as a constant source (b ¼ 0 for Eq 5).

TABLE 1 Coefficients of an emission model for 4,40 -MDI and TCPP.

4,40 -MDI TCPP

a b a,c
Eo Eo Eo Eob,c
(lg/m2 h) ba (1/h) (lg/g h) bb (1/h) R2 (lg/m2 h) (lg/g h)

FJS1 77.71 3.392 0.052 3.351 0.8810 38.2 0.020

FJS2 231.99 4.987 0.141 4.987 0.9883 60.7 0.039
FJS3 275.38 3.364 0.184 3.387 0.9147 2705.7 1.935
a
Normalized to specimen area.
b
Normalized to specimen weight.
c
From Cmax*N/L.
92 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

The area- and weight-specific emission rates were estimated based on the chamber con-
centration, air change rate, and loading ratio (Table 1). The emission factor for FJS3 is
about two to three times higher than those for wallpapers with 5 and 10 % TCPP [11].
Considering the differences in the test material and conditions, the degree of the differ-
ence may not be significant, thus supporting the validity of the current test method.

PREDICTED INDOOR AIR CONCENTRATION FOR 4,40 -MDI AND TCPP

Based on the emission factors expressed by the first-order decay (4,40 -MDI) and the
constant equation (TCPP), the indoor air concentrations in a small house (251.6 m3)
was estimated, assuming each product was applied instantaneously around all win-
dows with a total perimeter of 59.02 m and a gap of 0.635 cm under an air exchange
rate of 0.21 per hour. It was assumed that the wall (258 m2), floor (96 m2), and ceil-
ing (96 m2) worked as sinks. The scenario is for the smallest house among reference
houses provided by Natural Resources Canada to guide energy simulations in
Canada [32]. The sink area and the apparent sink rate coefficient for 4,40 -MDI were
assumed to be 450 m2 and 1.22 m/h, respectively. The air change rate is the mean
value for 39 control homes measured in a field-monitoring study in Quebec City
[33]. Eq 14 was used to predict the indoor 4,40 -MDI concentration.
Because sources and sinks of 4,40 -MDI were expected to behave differently
from those of TCPP, a new mass balance equation was introduced for TCPP:

V dC=dt ¼ Eo Af  QC  As dCs =dt (15)

where:
Eo ¼ emission factor of a constant source (Table 1), lg/m2 h,
Af ¼ area of the foam sealant applied, 0.37 m2,
hs ¼ convective mass transfer coefficient near the sorptive surface, 0.48 m/h,
which is the average of 0.54 m/h (floor and ceiling) and 0.42 m/h (wall) at the air ve-
locity of 0.15 m/s based on the correlations between Sherwood, Schmidt, and Reyn-
olds numbers [34], and
As ¼ area of the sorptive surface, m2.
A constant emission source was assumed for Eq 15. The sorptive sink rate was
assumed to be equal to the convective mass-transfer rate through the boundary lay-
er above the sorptive surface [35]. The partition to airborne particles that can affect
the fate of TCPP indoors [36] was not considered.
Assuming there is a linear equilibrium between the sorptive surface and the air adja-
cent to the surface [36], the mass balance in the sorptive surface phase can be written as:

dCs =dt ¼ hs ðC  Cs =Ks Þ (16)

where:
Ks ¼ partition coefficient between the sorptive surface and air, 100 m, which is
the average of 110 m (floor/carpet) and 89 m (wall and ceiling) [36] at a TCPP
vapor pressure of 1.4E-3 Pa [37].
 WON ET AL., DOI 10.1520/STP158920150038 93

Solving Eqs 15 and 16 gives:

C ¼ Eo Af f1  ða2  Q=VÞ=ða2  a1 Þexpða1 tÞ þ ða1  Q=VÞ=ða2  a1 Þexpða2 tÞg=Q

(17)

where:
 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
a2 ; a1 ¼ 0:5 ðQ=V þ hs As =V þ hs =K s Þ6 ðQ=V þ hs As =V þ hs =K s Þ2  4hs Q=ðKs VÞ :

The resulting maximum indoor air concentration after applying FJS3 was estimated
to be 0.052 lg/m3 for 4,40 -MDI (Fig. 10) and 4.1 lg/m3 for TCPP (Fig. 11) in 1 day
after application. Due to the constant emission rate, the TCPP concentration
increased to a steady-state concentration of 19.2 lg/m3 in 1 year. The estimated 4,40 -
MDI concentration is much lower than the reference concentration of 0.6 lg/m3
recommended for inhalation exposure by the U.S. EPA [38] and the occupational
exposure limit of 50 lg/m3 (8-h time-weighted average) [39], even under worst-case
conditions (40 C). However, the expected TCPP concentration is greater than the
indoor air concentration measured in homes: 0.038 to 0.21 lg/m3 [17] and 0.0024 to
0.064 lg/m3 [15], which may suggest that foam joint sealants may need further atten-
tion as a potential source that can increase the indoor TCPP concentration. How-
ever, it should be noted that the indoor air quality simulation results of 4,40 -MDI
and TCPP apply to three foam joint sealants and application conditions employed

FIG. 10 Predicted indoor air concentration of 4,40 -MDI from one-component foam joint
sealant applied around all windows in a small house.
94 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 11 Predicted indoor air concentration of TCPP from one-component foam joint
sealant applied around all windows in a small house.

in this study. Generalization of the simulation results is not recommended, and vari-
ous assumptions used for indoor air quality predictions need further validation.

Conclusions
The test method in this study with 4,40 -MDI showed that the effect of environmental
factors (temperature and RH) as well as the sink effect by interior surfaces could be sig-
nificant. When a test method is standardized for 4,40 -MDI emissions, these influential
factors should be thoroughly investigated. Despite the low vapor pressure of 4,40 -MDI
and TCPP, the airborne concentrations were above the detection limit at 40 C. The
chamber test at 40 C and 20 % RH appears to be a suitable method for determining
emission rates of 4,40 -MDI and TCPP from do-it-yourself consumer products. Although
the test repeatability was good for TCPP and 4,40 -MDI for an identical product, attention
needs to be paid to the batch-to-batch variation for TCPP. Based on the emission rates
measured from foam joint sealants under a worst-case scenario, the indoor air concentra-
tion of 4,40 -MDI after the application of foam joint sealants to all the windows of a small
house was predicted to be much lower than the reference concentration of 0.6 lg/m3
recommended by the U.S. EPA. Conversely, the higher predicted concentrations of
TCPP than those found in the literature may imply the potential of foam joint sealants as
an important source of TCPP in homes.
More efforts are needed to verify the assumptions used for the simulations. For
example, the sink coefficient measured for 4,40 -MDI in this study is for glass and
 WON ET AL., DOI 10.1520/STP158920150038 95

those for TCPP are based on correlations developed for phthalates [36]. The sink
coefficients that can reflect the properties of an actual compound (e.g., TCPP) and
actual building materials need to be investigated for a more realistic prediction of
indoor air quality. The potential reaction between water vapor and 4,40 -MDI needs
further attention because this study showed RH could be an important factor affect-
ing the fate of 4,40 -MDI. A more direct approach that can prove the presence of
reaction products would be valuable.

ACKNOWLEDGMENTS
We thank Alex Wang for help with the chamber tests. The quality assurance and con-
trol information for the LC-MS analysis was provided by Simon Aubin.

References

[1] Henneken, H., Vogel, M., and Karst, U., “Determination of Airborne Isocyanates,” Anal.
Bioanal. Chem., Vol. 387, No. 1, 2007, pp. 219–236.

[2] Krone, C. A., “Diisocyanates and Nonoccupational Disease: A Review,” Arch. Environ.
Health, Vol. 59, No. 6, 2004, pp. 306–316.

[3] ECHA, “Summary of Classification and Labelling: Harmonised Classification—Annex VI of

Regulation (EC) No 1272/2008 (CLP Regulation)—4,40 -Methylenediphenyl Diisocyanate,”
http://echa.europa.eu/home

[4] IARC, “4,40 -Methylenediphenyl Diisocyanate and Polymeric 4,40 -Methylenediphenyl

Diisocyanate,” http://monographs.iarc.fr/ENG/Monographs/vol71/mono71-47.pdf (accessed
November 4, 2016).

[5] Government of Canada, “Methylenediphenyl Diisocyanates and Diamine (MDI/MDA)

Substances,” http://www.chemicalsubstanceschimiques.gc.ca/fact-fait/glance-bref/dii-
socyanate-eng.php (accessed August 5, 2015).

[6] U.S. EPA, “Methylene Diphenyl Diisocyanate (MDI) and Related Compounds Action
Plan,” https://www.epa.gov/sites/production/files/2015-09/documents/mdi.pdf (accessed
November 4, 2016).

[7] Booth, K., Cummings, B., Karoly, W. J., Mullins, S., Robert, W. P., Spence, M., Lichtenberg, F. W.,
and Banta, J., “Measurements of Airborne Methylene Diphenyl Diisocyanate (MDI) Concen-
tration in the U.S. Workplace,” J. Occup. Env. Hyg., Vol. 6, No. 4, 2009, pp. 228–238.

[8] Hext, P. M., Booth, K., Dharmarajan, V., Karoly, W. J., Parekh, P. P., and Spence, M.,
“A Comparison of the Sampling Efficiencies of a Range of Atmosphere Samplers When
Collecting Polymeric Diphenylmethane Di-Isocyanate (MDI) Aerosols,” Appl. Occup.
Environ. Hyg., Vol. 18, No. 5, 2003, pp. 346–357.

[9] Pizzi, A. and Mittal, K. L., Handbook of Adhesive Technology, 2nd Edition, CRC Press,
New York, 2003.

[10] U.S. Department of Energy, “Caulking,” http://energy.gov/energysaver/articles/caulking

(accessed November 4, 2016).
96 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

[11] Ni, Y., Kumagai, K., and Yanagisawa, Y., “Measuring Emissions of Organophosphate
Flame Retardants Using a Passive Flux Sampler,” Atmos. Environ., Vol. 41, No. 15, 2007,
pp. 3235–3240.

[12] World Health Organization, “Environmental Health Criteria 209: Flame Retardants: Tris-
(chloropropyl) Phosphate and Tris(2-chloroethyl) Phosphate,” http://www.inchem.org/
documents/ehc/ehc/ehc209.htm (accessed November 4, 2016).

[13] Health Canada, “Certain Organic Flame Retardants Substance Grouping,” http://www.
chemicalsubstanceschimiques.gc.ca/group/flame_retardant-ignifuges/index-eng.php
(accessed November 4, 2016).

[14] Health Canada, “Methylenediphenyl Diisocyanate and Diamine (MDI/MDA) Sub-

stance Grouping,” http://www.chemicalsubstanceschimiques.gc.ca/fact-fait/glance-bref/
diisocyanate-eng.php (accessed November 4, 2016).

[15] Bergh, C., Torgrip, R., Emenius, G., and Östman, C., “Organophosphate and Phthalate Esters
in Air and Settled Dust—A Multi-Location Indoor Study,” Indoor Air, Vol. 21, No. 1, 2011,
pp. 67–76.

[16] Brommer, S., Harrad, S., Van den Eede, N., and Covaci, A., “Concentrations of Organo-
phosphate Esters and Brominated Flame Retardants in German Indoor Dust Samples,”
J. Environ. Monit., Vol. 14, No. 9, 2012, pp. 2482–2487.

[17] Marklund, A., Andersson, B., and Haglund, P., “Organophosphorus Flame Retardants and
Plasticizers in Air from Various Indoor Environments,” J. Environ. Monit., Vol. 7, No. 8,
2005, pp. 814–819.

[18] Kemmlein, S., Hahn, O., and Jann, O., “Emissions of Organophosphate and Brominated
Flame Retardants from Selected Consumer Products and Building Materials: Indoor Air
Chemistry and Physics: Papers from Indoor Air 2002,” Atmos. Environ., Vol. 37,
No. 39–40, 2003, pp. 5485–5493.

[19] Wirts, M. and Salthammer, T., “Isocyanate Emission from PUR Adhesives: Influence of
Temperature, Monomer Content, and Curing Mechanism,” Environ. Sci. Technol., Vol. 36,
No. 8, 2002, pp. 1827–1832.

[20] Wirts, M., Grunwald, D., Schulze, D., Uhde, E., and Salthammer, T., “Time Course of
Isocyanate Emission from Curing Polyurethane Adhesives,” Atmos. Environ., Vol. 37,
No. 39–40, 2003, pp. 5467–5475.

[21] U.S. EPA, “Definition and Procedure for the Determination of the Method Detection Limit—
Revision 1.11,” https://www.gpo.gov/fdsys/pkg/CFR-2011-title40-vol23/xml/CFR-2011-
title40-vol23-part136-appB.xml (accessed November 4, 2016).

[22] Nelson, G. O., Gas Mixtures: Preparation and Control, Lewis Publishers, Boca Raton,
Florida, 1992.

[23] Yaws, C. L., Yaws’ Handbook of Thermodynamic and Physical Properties of Chemical
Compounds, Knovel, Norwich, N.Y., 2003.

[24] Lyman, W. J., Reehl, W. F., and Rosenblatt, D. H., Handbook of Chemical Property Esti-
mation Methods: Environmental Behavior of Organic Compounds, American Chemical
Society, Washington, D.C., 1990.
 WON ET AL., DOI 10.1520/STP158920150038 97

[25] BASF, “Polyurethane MDI Handbook,” http://www2.basf.us/urethanechemicals/Specialty_

Systems/pdfs/mdihandbook.pdf, 2009.

[26] Holdren, M. W., Spicer, C. W., and Riggin, R. M., “Gas Phase Reaction of Toluene Diisocya-
nate with Water Vapor,” Am. Ind. Hyg. Assoc. J., Vol. 45, No. 9, 1984, pp. 626–633.

[27] Dyson, W. L. and Hermann, E. R., “Reduction of Atmospheric Toluene Diisocyanate by

Water Vapor,” Am. Ind. Hyg. Assoc. J., Vol. 32, No. 11, 1971, pp. 741–744.

[28] Vaisala, “Humidity Calculator 5.0,” http://go.vaisala.com/humiditycalculator/5.0/?

utm_campaign¼&utm_medium¼email&utm_source¼Eloqua&utm_content¼CEN-TIA-G-
Humidity%20Calculator%20Autoresponder (accessed June 29, 2015).

[29] Saunders, J. H. and Frisch, K. C., Polyurethanes: Chemistry and Technology, Part I—
Chemistry, Interscience Publishers, New York, 1962.

[30] Yaws, C. L., Chemical Properties Handbook, McGraw-Hill, New York, 1999.

[31] Acree, W. and Chickos, J. S., “Phase Transition Enthalpy Measurements of Organic and
Organometallic Compounds. Sublimation, Vaporization and Fusion Enthalpies from
1880 to 2010,” J. Phys. Chem. Ref. Data, Vol. 39, No. 4, 2010, pp. 1–942.

[32] Lio & Associates, “A Study of Prescriptive Requirements for EnerGuide 80 in Ontario’s
Building Code,” The Ontario Ministry of Municipal Affairs and Housing, 2010, http://
govdocs.ourontario.ca/node/6412

[33] Lajoie, P., Aubin, D., Gingras, V., Daigneault, P., Ducharme, F., Gauvin, D., Fugler, D.,
Leclerc, J.-M., Won, D., Courteau, M., Gingras, S., Héroux, M.-È., Yang, W., and Schlei-
binger, H., “The IVAIRE Project—A Randomized Controlled Study of the Impact of Venti-
lation on Indoor Air Quality and the Respiratory Symptoms of Asthmatic Children in
Single Family Homes,” Indoor Air, Vol. 25, No. 6, 2015, pp. 582–597.

[34] Guo, Z., “Review of Indoor Emission Source Models. Part 2. Parameter Estimation,” Envi-
ron. Pollut., Vol. 120, No. 3, 2002, pp. 551–564.

[35] Liu, Z., Ye, W., and Little, J. C., “Predicting Emissions of Volatile and Semivolatile Organic
Compounds from Building Materials: A Review,” Build. Environ., Vol. 64, June 2013,
pp. 7–25.

[36] Xu, Y., Hubal, E. A. C., Clausen, P. A., and Little, J. C., “Predicting Residential Exposure to
Phthalate Plasticizer Emitted from Vinyl Flooring: A Mechanistic Analysis,” Environ. Sci.
Technol., Vol. 43, No. 7, 2009, pp. 2374–2380.

[37] European Commission, European Union Risk Assessment Report: Tris(2-chloro-1-meth-

ylethyl) phosphate (TCPP), 2008, http://echa.europa.eu/documents/10162/13630/trd_
rar_ireland_tccp_en.pdf

[38] U.S. EPA, “Toxicological Review of Methylene Diphenyl Diisocyanate (MDI),” 1998,
https://cfpub.epa.gov/ncea/iris/iris_documents/documents/toxreviews/0529tr.pdf

[39] Centers for Disease Control and Prevention, “Methylene Bisphenyl Isocyanate,” http://
www.cdc.gov/niosh/npg/npgd0413.html (accessed August 18, 2015).
98 DEVELOPING CONSENSUS STANDARDS FOR MEASURING CHEMICAL EMISSIONS

STP 1589, 2017 / available online at www.astm.org / doi: 10.1520/STP158920150034

Scott Ecoff,1 Shen Tian,1 and John Sebroski1

Prioritizing Chemical Emissions

from Closed-Cell Spray
Polyurethane Foam: Utilizing
Micro-Scale Chamber Emission
Factors and Field Measurement
Data
Citation
Ecoff, S., Tian, S., and Sebroski, J., “Prioritizing Chemical Emissions from Closed-Cell Spray
Polyurethane Foam: Utilizing Micro-Scale Chamber Emission Factors and Field Measurement
Data,” Developing Consensus Standards for Measuring Chemical Emissions from Spray
Polyurethane Foam (SPF) Insulation, ASTM STP1589, J. Sebroski and M. Mason, Eds., ASTM
International, West Conshohocken, PA, 2017, pp. 98–118, http://dx.doi.org/10.1520/
STP1589201500342

ABSTRACT
We conducted field and laboratory chamber studies in March 2013 to assess
airborne concentrations of chemicals emitted from a rigid closed-cell medium-
density spray polyurethane foam (SPF) insulation material. For the field study,
SPF was installed by high-pressure spray application in the main attic and in four
smaller crawlspaces of a home built in 1975. The field study involved the
collection of area air samples in the home before, during, and after SPF
application. We also evaluated chemical emissions from SPF using laboratory
micro-scale chambers specifically optimized to evaluate postspray chemical
emissions from SPF. Micro-scale chambers are a common tool used by labs to
assess volatile organic compound (VOC) emissions from a source, and standard
test methods are being developed by ASTM specifically for SPF. To prioritize
VOC emissions from SPF, a simple well-mixed, single-zone box model and the
emission factors generated from a micro-scale chamber study were used to

Manuscript received April 21, 2015; accepted for publication September 5, 2016.
1
Covestro LLC, 1 Covestro Circle, Pittsburgh, PA 15205
2
ASTM Symposium on Developing Consensus Standards for Measuring Chemical Emissions from Spray
Polyurethane Foam (SPF) Insulation on April 30–May 1, 2015 in Anaheim, CA.

Copyright V
C 2017 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959.
 ECOFF ET AL., DOI: 10.1520/STP158920150034 99

evaluate the screening-level VOC airborne concentration in the home. These

screening-level values were then compared to air sampling data obtained during
the field study. For two VOCs that were the predominant chemicals emitted from
the SPF, the airborne VOC concentrations were within an order of magnitude
between modeling and actual field measurement at Days 7 and 28. This study
indicates that emission factors obtained through micro-scale chamber testing
together with a screening-level model could serve as a prioritization tool in
evaluating indoor VOC airborne concentrations for SPF applications. More
sophisticated models are needed to better characterize emission mechanisms,
indoor fate, transport, and airborne concentration of all chemicals, including
semivolatile VOCs that may be emitted from SPF.

Keywords
spray polyurethane foam (SPF), chemical emissions, micro-scale chamber,
closed-cell polyurethane foam insulation, volatile organic compounds (VOCs),
semivolatile organic compounds (SVOCs)

Introduction
Spray polyurethane foam (SPF) can help improve a home’s energy efficiency.
Because it is sprayed directly into the gaps, cracks, and other surfaces that contrib-
ute to heat loss, it both insulates and air seals, offering one of the most effective
ways of weatherizing existing homes and new construction. Because SPF insulation
minimizes air infiltration, it assists in preventing moisture vapor from entering and
escaping the home, which in turn reduces the load on heating and cooling systems
[1]. There are two types of high-pressure SPF insulation: open and closed cell. The
specific product evaluated during this survey was a medium-density closed-cell SPF
insulation material.
The application of medium-density closed-cell SPF insulation involves the use
of various chemicals that are contained within the A-side (isocyanate) and B-side
(polyol) components. When the A- and B-side components are combined, they
form polyurethane foam within a few seconds. The A side is polymeric methylene
diphenyl diisocyanate (MDI) and contains approximately equal amounts of mono-
meric MDI (predominantly 4,40 -MDI, a chemically bonded two-ring structure) and
higher-molecular-weight oligomers of MDI (three-ring bonded structures and
larger). The B side is a blend of chemicals that typically contain polyols, flame retar-
dants, catalysts, blowing agents, surfactants, and other additives.
The primary focus of this study in March 2013 was to evaluate airborne con-
centrations of chemicals that off-gas from one SPF insulation product manufac-
tured by Covestro. The purpose was to compare laboratory chamber chemical
emission study work to air sampling results measured in a home after installation
of the product. The scope of this study covers only the gas-phase airborne concen-
tration of chemicals while previous studies have shown indoor particulates could be
100 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

an important exposure pathway for semivolatile organic compounds (SVOCs)

[2–7]. However, no samples were collected to characterize the particle size distribu-
tion or chemical species absorbed to the particles because it was out of the scope of
the study when it was planned. This study helped determine whether laboratory
chamber study work can be used as a tool to prioritize and evaluate SPF airborne
chemical concentrations in a structure. This study provides a pilot test data set, dis-
cusses the strengths and limitations of this approach, and provides a basis for
designing future studies and a vehicle to stimulate further discussion. This study is
our first attempt to evaluate an approach using micro-scale chambers to prioritize
chemical emissions from SPF.

Study Methodology
The field study was performed in a two-story wood-frame home built in 1975 with
exterior brick and wood siding. It was built on a concrete slab and did not have a
basement or crawlspace below the structure. The home had four bedrooms and a
detached garage. The approximate square footage of the living space of the home
was 205 m2 (2,200 ft2).
The family vacated the home during and for 24 h after SPF application in
accordance with the insulation manufacturer’s recommendation. The SPF product
was installed by high-pressure spray application primarily in the main attic space
above the second-floor bedrooms and applied to the underside of the roof deck.
Two side walls of the attic were also sprayed to create an unvented attic.
The main attic space was accessible by pull-down stairs located in the hallway
of the second floor. Four small crawlspaces were also sprayed with insulation mate-
rial, and each was accessible through small doorway openings located within two
bedrooms on the second floor. The attic door and crawlspace access doors were
made of quarter-inch plywood. No sealants (i.e., weather stripping or foam) were
used around the attic or crawlspace doorways. All air sampling conducted in the
home after SPF application was completed with the access doors to the four small
crawlspaces closed. The access door to the attic was not closed until the postspray
test period of 24 h. This was because of the use of portable exhaust ventilation oper-
ating during SPF application and up to the postsampling test period of 24 h.
Key information about the spray application and home are outlined as follows:
• Total weight of SPF sprayed: 567.5 kg (1,230 lb)
• Spray parameters: 8,280 kpa (1,200 psi) at 49 C (120 F)
• Total area sprayed in main attic: 84 m2 (900 ft2)
• Total area sprayed within the four crawlspaces: 28 m2 (300 ft2)
• Two SPF applicators: one sprayed in the main attic and one in the four
crawlspaces
• Volume of main attic space: 71 m3 (2,500 ft3)
• Volume of living space of home: 521 m3 (18,400 ft3)
• Outdoor ambient temperature: 18 C (65 F) at the start of spray application
• Spray application time: 11:00 a.m. to 5:00 p.m.
 ECOFF ET AL., DOI: 10.1520/STP158920150034 101

• Portable ventilation: portable ventilation (one exhaust fan) was used to

exhaust air from the main attic during and after SPF application for 24 h
• House ventilation: heating, ventilation, and air-conditioning (HVAC) system
located in the main attic; this system was shut off during SPF application and
up to 48 h postapplication
A portable exhaust fan rated at 18.5 m3/min (650 ft3/min) was used to exhaust
air from the main attic during SPF application. This fan operated for 24 h after SPF
application and then was removed from the home. The number of air changes per
hour (ACH) achieved in the attic with this fan was approximately 10 ACH. The fan
was positioned in the hallway on the second floor of the home. Flexible ductwork
was used to draw air into the fan from the main attic (the end of the exhaust duct
or suction side of the duct was positioned near the center of the main attic above
the access opening). All of the air from the main attic was exhausted outside the
home using flexible ductwork that was secured to an exterior wall window located
within one of the second-floor bedrooms. None of the crawlspaces was vented dur-
ing this study.
Five area air sample locations within the home were established and maintained
throughout the study. One of the area sample locations was in the center of the
main attic. The remaining four locations were within the living space of the home
and included two bedrooms on the second floor and the living room (piano room),
kitchen, and family room on the first floor.
Air samples were collected for aldehydes before the day of SPF application.
This was done because formaldehyde and acetaldehyde are commonly detected in
the indoor air of residential structures. There are many sources of formaldehyde
and other carbonyl compounds, including adhesives, glues, caulks, wall sheathing,
particle board, and other pressed wood products.

CHEMICALS EVALUATED
The specific chemicals selected for evaluation during this study included the following:
• 4,40 - and 2,40 -MDI (monomeric MDI represents approximately 50 % of the
A-side raw material); MDI was evaluated only during the field study and not
in the laboratory micro-scale chambers
• Amine catalysts (three amines are included in the B side of this SPF formula-
tion; the specific amine compounds are considered proprietary information by
the manufacturer and cannot be identified)
• 1,1,1,3,3-Pentafluoropropane (HFC-245fa), a blowing agent and B-side
ingredient
• trans-1,2-Dichloroethylene (1,2-DCE or TDCE), a co-blowing agent and
B-side ingredient
• Tris(1-chloro-2-propyl) phosphate (TCPP), a flame retardant and B-side
ingredient
• 2-Butoxyethanol (ethylene glycol monobutyl ether), a B-side ingredient
• Aldehydes (formaldehyde, acetaldehyde, propionaldehyde); none of these
chemicals was intentionally added into the SPF formulation
102 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

Of the chemical compounds evaluated, the blowing agents and 2-butoxyethanol

are considered volatile organic compounds (VOCs), whereas the other chemicals
(i.e., MDI, TCPP) are generally considered SVOCs.
Tables A.1–A.5 in the appendix at the end of the paper illustrate the field sample
results, analytical limits of quantitation (LOQs) for the chemicals, and sample flow
rates/sample times (each of these parameters is noted in the table notes). Additional
information about field air sampling and laboratory analytical methods is summa-
rized in the following section.

FIELD AIR SAMPLING AND ANALYTICAL TECHNIQUES

General Background Information
Because of the number of different chemicals evaluated, several industrial hygiene
sampling and laboratory analytical methods were used during this study. Treated
filters, glass sorbent tubes, and thermal desorption (TD) tubes were used to assess
the airborne chemical concentrations. The sorbent tubes and TD tubes were
mounted within Zefon Gemini twin-port low-flow samplers. The Gemini was
placed in the sampling train between the sample media and the air sampling pump.
Air was drawn through the twin-port low-flow sampler and the sample media using
SKC Airchek 52 air sampling pumps. All air sampling pumps were calibrated before
sample collection to the appropriate sampling rate, as specified in the sampling and
R
analytical method provided by the analytical laboratory using a Bios DryCalV
DC-Lite Primary Flow Meter. The sampling rate was measured after each sam-
pling event, and the average of the pre- and postvalues was used to calculate the
total sample volume. To assess airborne concentrations at specific locations within
the home, tripods were used to mount the sample media at breathing zone height
(1.5 m above the floor). One blank was submitted to the laboratory with each set
of 10 field samples or less. All field air samples were analyzed at the Covestro
Industrial Hygiene and Environmental Analytics Laboratory, which is accredited
by the American Industrial Hygiene Association (AIHA). The following sections
briefly summarize the sampling and analytical methods used for each chemical
evaluated.

MDI
Airborne concentrations of MDI monomer (2,40 - and 4,40 -MDI) vapor were
collected using 13-mm glass-fiber filters mounted in a Swinnex cassette holder
in accordance with Covestro Industrial Hygiene Laboratory Method 1.7.7. This
method is very similar to OSHA Method 47 for MDI. Prior to mounting the
filter into the Swinnex holder, each filter was coated with 2 mg 1-(2-pyridyl)
piperazine and diethyl phthalate. When airborne isocyanate is drawn through
the filter, it is converted to a stable urea derivative that is quantitatively analyzed
with the use of high-performance liquid chromatography (HPLC) with fluores-
cence detection. Immediately after sampling, each 13-mm filter was removed
 ECOFF ET AL., DOI: 10.1520/STP158920150034 103

from the Swinnex holder and field-desorbed in 2 mL of 90:10 acetonitrile:

dimethylsulfoxide.

Amine Catalysts, Blowing Agents (HFC-245fa/TDCE), Flame Retardant

(TCPP), and 2-Butoxyethanol
Airborne concentrations of amine catalysts, blowing agents, a flame retardant, and
2-butoxyethanol were collected using Markes International Ltd. TD tubes. The TD
R
tubes were constructed of stainless steel and contained quartz wool, TenaxV TA,
and CarbopackTM X (35/60 and 40/60 mesh). Sample volumes ranging from 0.25 to
1 L were selected to avoid overloading the sampling media or analytical system with
blowing agents. A Markes International TD-100 thermal desorber coupled to an
Agilent 7890/5975 gas chromatograph-mass spectrometer was used to analyze the
tubes with methodology based on U.S. EPA TO-17 and ISO 16000 Part 6. The
transfer line, analytical column, and method conditions were optimized for the
recovery of target compounds (e.g., amine catalysts, TCPP).

Aldehydes
Airborne concentrations of formaldehyde, acetaldehyde, and propionaldehyde were
evaluated using glass sorbent tubes (SKC Inc. 226-119; 8 by 110 mm) containing
silica gel coated with dinitrophenylhydrazine. The analysis was performed using
Agilent 1100 series HPLC.

SAMPLING LOCATIONS WITHIN THE HOME AND POSTSAMPLE TEST PERIODS

Five area sampling locations were selected before the study and maintained
throughout each air sampling test period. The air sampling locations were as
follows:
• Center of main attic: 2.5 m from access opening to attic
• Second-floor west bedroom (two crawlspaces accessible from this room)
• Second-floor east bedroom (two crawlspaces accessible from this room)
• First-floor living (piano) room (located in front half of home)
• First-floor kitchen/family room (one large open area in back half of home)

The sampling pumps were mounted on tripods with the sample media placed
at approximately breathing zone height. Air samples were collected at specific time
periods after the completion of SPF application, including at 12, 24, 48, 72, and
168 h (Day 7), as well as 1, 4, and 16 months. Once a chemical component was
below the LOQ on at least two postsampling periods (e.g., not detectable in the air
based on the sampling and analytical method used), air sampling for that particular
chemical was discontinued.

SPF CHEMICAL CRITERIA VALUES

Finding appropriate criteria values to compare air sampling results to after SPF
application in a home study is challenging. Several chemicals used in SPF insulation
do not have published indoor air criteria values. For this study, most of the chemi-
cals did have published occupational exposure limits (OELs), such as OSHA’s
104 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 1 Criteria values for the SPF chemicals evaluated during this study.

EPA Chronic ANSI/ASHRAE USGBC LEED California OEHHA

Chemical Inhalation RfC [8] E 62.1-2013 [9] v4 [10] CREL [11] OEL-TLV [12]

TDCE NV NV NV NV 793,000
HFC-245fa NV NV NV NV 1,644,000
TCPP NV NV NV NV NV
Formaldehyde NV 9a 33 9 (chronic) 370
55b 55 (acute)
Acetaldehyde 9 300 (8 h) 70 140 45,000
Propionaldehyde 8 NV NV NV 47,400
2-Butoxyethanol 1,600 967 (chronic) NV 14,000 (acute) 96,500
MDI 0.6 NV NV 0.7c 51
3
Note: All values are in lg/m . CREL ¼ chronic reference exposure level; NV ¼ no value to reference;
RfC ¼ reference concentration.
a
Chronic never to exceed guideline to avoid irritant effect in sensitive individuals.
b
Acute and 8-h noncancer inhalation CREL based on the California OEHHA CREL.
c
California OEHHA CREL value was lowered in 2016 to 0.08.

permissible exposure limits, American Conference of Governmental Industrial

Hygienists threshold limit values (TLVs), or AIHA workplace environmental expo-
sure levels. In many instances, consultants preparing indoor air reports frequently
refer to environmental regulations (e.g., the U.S. EPA, state environmental agencies
such as California’s Office of Environmental Health Hazard Assessment [OEHHA],
etc.) and organizations such as the American Society of Heating, Refrigerating,
and Air-Conditioning Engineers (ASHRAE) and the U.S. Green Building Council
(USGBC) Leadership in Energy and Environmental Design (LEED). The criteria value
for the same chemical can vary between different agencies/organizations because the
intended use of the value and how it was established may not be the same (Table 1).

MICRO-SCALE CHAMBER SETUP AND ANALYTICAL PARAMETERS

The micro-scale chambers described in ASTM D7706-11, Standard Practice for
Rapid Screening of VOC Emissions from Products Using Micro-Scale Chambers [13],
were used to evaluate postspray SPF chemical emissions in the laboratory. Micro-
scale chambers used in this study were small 114-mL cylindrical-shaped chambers
constructed of stainless steel with an inert coating. The micro-scale chambers were
operated with a relatively high loading factor (L ¼ 62.5 m2/m3) and air change rate
(58.3 changes/h). The area-specific flow rate (calculated with the head space vol-
ume), which is the ratio of the loading factor and air exchange rate in the chamber,
was 0.932 m h1. Samples of SPF insulation were sprayed and packaged as
described in ASTM D7859-13e1, Standard Practice for Spraying, Sampling, Packag-
ing, and Test Specimen Preparation of Spray Polyurethane Foam (SPF) Insulation
for Testing of Emissions Using Environmental Chambers [14], and then shipped via
overnight delivery to the laboratory. Upon receipt (within 48 h of spraying), the
 ECOFF ET AL., DOI: 10.1520/STP158920150034 105

SPF insulation samples were placed in the micro-scale chambers to collect air
(emission) samples at 4 and 24 h and then periodically for up to 35 days.
Released VOCs and SVOCs were collected onto sorbent tubes for analysis by
TD and gas chromatography-mass spectrometry (GC-MS), which is similar to the
methodology described in EPA Method TO-17 and ISO 16000 Part 6 [15,16]. Alde-
hyde emissions were evaluated using sorbent tubes containing treated silica gel, fol-
lowed by analysis with HPLC with ultraviolet detection as described in ASTM
D5197-09e1, Standard Test Method for Determination of Formaldehyde and Other
Carbonyl Compounds in Air (Active Sampler Methodology) [17]. See Table 2 for the
TD and GC-MS analytical parameters used for the laboratory chamber study work.
MDI emissions were not evaluated using the micro-scale chambers because previ-
ous research has shown that MDI adheres to the chamber surfaces, which signifi-
cantly biases the recovery of MDI emissions [18].
The emission factor reporting limits were calculated using the exposed surface
area of the sample (0.00322 m2), time of sample collection (20 min for TD tubes

TABLE 2 TD and GC-MS analytical parameters.

Analysis Parameters

TD
Flow path temperature 160 C
Split in standby 10 mL/min
Material emissions trap (quartz wool, Tenax TA, CarbographTM 5TD)
R
V
Cold trap
Dry purge 1 min; 20 mL/min flow to split
Prepurge 0.1 min (default)
Primary desorption 270 C for 8 min; 35 mL/min trap flow; no split flow
Pretrap fire purge 1 min; 35 mL/min trap flow; 50 mL/min split flow
Focusing trap conditions Trap low: 25 C; trap high: 300 C; heating rate: MAX (>40 C/s); hold
time: 3 min; 50 mL/min split flow
Overall TD split 34.3:1
GC-MS
Transfer line Base-deactivated transfer line
Column Low-polarity phase; 5% diphenyl/95% dimethyl polysiloxane;
amine optimized; 30 m; 0.25 mm by 0.5 lm
Column flow 1.5 mL/min; constant flow
Temperature program 40 C (2 min); 20 C/min to 300 C (2 min)
Total run time 17 min
Carrier gas Helium
GC inlet temperature 200 C
MS source temperature 230 C
MS quad temperature 150 C
MS transfer line
temperature 250 C
Mass scan range m/z ¼ 40–550 amu
106 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

and approximately 20 h for aldehydes), and the analytical method detection limits
or reporting limits as shown in Table 3.

Results and Discussion

All of the air sampling results from the field study are illustrated in the tables
found at the end of this paper. Table A.1 illustrates the airborne concentrations of
monomeric MDI (2,40 - and 4,40 -MDI). Table A.2 illustrates the airborne concentra-
tions of TDCE and HFC-245fa (blowing agents). Table A.3 illustrates the airborne
concentrations of TCPP. Table A.4 illustrates the airborne concentrations of the
three aldehyde compounds. Table A.5 illustrates the airborne concentrations of
2-butoxyethanol. The following sections summarize the airborne concentrations
found in the home and the applicable criteria levels that may be used to compare
results.

FIELD STUDY RESULTS

MDI
Airborne concentrations of MDI were detected only during SPF application
within the main attic and in one of the second-floor bedrooms. The airborne
concentrations of MDI in these areas during application were 2.0 lg/m3 (center
of main attic approximately 8 ft from access opening) and 1.0 lg/m3 (second-
floor west bedroom). These airborne concentrations were less than the OEL
based on an 8-h time-weighted average of 51 lg/m3. The other three area sample
locations within the home were less than the LOQ during spray application. The
LOQ refers to a minimum concentration of a chemical that can be measured
within the specified limits of precision and accuracy set by the laboratory analyt-
ical method (instrumentation). The LOQ is also dependent upon the sample
duration and flow rate, which are used to calculate the volume of air sampled.

TABLE 3 Analytical method detection limits.

Chemical Method Detection Limit, ng/tube

HFC-245fa 49.4
TDCE 0.819
TCPP 26.0
2-Butoxyethanol 22.0
Formaldehyde 160
Propionaldehyde 160
Acetaldehyde 160

Note: Detection limits for HFC-245fa, TDCE, and TCPP were determined on TD tubes with the
procedures described in Ref. [19]. All other compound detection limits were estimated with the low-
est calibration standard.
 ECOFF ET AL., DOI: 10.1520/STP158920150034 107

When reporting a nondectable (ND) or less than (<) value, it means that the
chemical is not present in a sufficient amount to be reliably quantified below
this value. The airborne concentrations of MDI were less than the LOQ or ND
at each of the five test locations in the home during the postspray sample test
periods evaluated. All the MDI values were at or below the EPA chronic inhala-
tion reference concentration (RfC) of 0.6 lg/m3 and the California OEHHA
chronic reference exposure level (CREL) of 0.7 lg/m3 for this chemical at the
time this study was completed.

Amine Catalysts
All three amine compounds in the SPF formulation are proprietary ingredients.
Because the three amines cannot be identified by name, no field results were pre-
sented in this paper. Although the specific results are not illustrated, at each post-
spray test period the airborne amine concentrations were less than the analytical
LOQ both within the living spaces of the home and in the main attic. The use of
larger sampling volumes may have enhanced the detection of the amine catalysts;
however, larger sampling volumes were not obtained to avoid overloading the tubes
with blowing agents. Larger sampling volumes may be necessary, especially for
reactive catalysts, which can be used in SPF formulations to minimize their emis-
sions [20]. None of the amine catalysts in the formulation has a published OEL,
odor threshold, or any other criteria value from the organizations/agencies previ-
ously referenced.

Blowing Agents
Airborne concentrations of the two blowing agents (HFC-245fa and TDCE) were
detected in all five areas tested within the home during SPF application and all post-
spray air sampling test periods. Both of these chemicals appeared to reach a quasi-
steady-state concentration in the home on Day 7. This is believed to have been
caused by the HVAC system that was located in the attic. Because of warmer out-
door air temperatures beginning on Day 3 after SPF application, the homeowner
shut all windows, and the air-conditioning system began to operate on a regular
basis. More consistent operation of the HVAC may have helped to distribute the
airborne chemicals throughout the home.
During the postspray test period of 7 days, the airborne concentrations of
HFC-245fa in the living spaces of the home ranged from 3,300 to 3,540 lg/m3. Dur-
ing the postspray test period of 1 month, the HFC-245fa airborne concentrations
decreased, with the values ranging from 1,610 to 1,750 lg/m3.
During the postspray test period of 7 days, the airborne concentrations of TDCE
within the living spaces of the home ranged from 3,510 to 3,990 lg/m3. During the
postspray test period of 1 month, the TDCE airborne concentrations within the living
spaces of the home ranged from 2,280 to 2,370 lg/m3. None of the agencies/organiza-
tions referenced has a published indoor air criteria value for either of the blowing
agents.
108 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TCPP
Airborne concentrations of TCPP were detected above its LOQ during SPF application
and within 6 of 30 samples collected postspray. The highest airborne concentration of
TCPP (1,290 lg/m3) was measured within the main attic during SPF application. Postap-
plication airborne concentrations of TCPP in the main attic decreased from 1,300 lg/m3
to airborne concentrations of 72, 89, 119, 68, and 19 lg/m3 at the 12, 24, 48, 72, and
168-h (7 days) postspray sample test periods, respectively. The only postspray test period
in which TCPP was measured in the living spaces of the home above its LOQ was at
24 h. The airborne TCPP concentration in a bedroom on the second floor was 20 lg/m3.
All the other postspray airborne concentrations of TCPP were ND (<17 lg/m3) in the
living spaces of the home. None of the agencies/organizations referenced has a published
indoor air criteria value for TCPP.

Aldehydes
Airborne concentrations of formaldehyde, acetaldehyde, and propionaldehyde were
evaluated during this study. Background air samples were collected within the home the
day before SPF application. Preapplication air sampling was completed because both
formaldehyde and acetaldehyde are ubiquitous chemicals (e.g., they are commonly found
in the indoor air of most structures). Formaldehyde is a component of many household
building products. Acetaldehyde is used as a food preservative and an ingredient in per-
fumes, polyester resins, and dyes. The following list summarizes the aldehyde findings:
• Formaldehyde: Preapplication airborne concentrations in the home averaged
31 lg/m3. This average background level decreased before eventually returning
to near the preapplication background airborne concentration of 31 lg/m3
four months after application. The airborne concentration of formaldehyde in
the attic was approximately 20 lg/m3 during the postapplication sample peri-
ods of 48 h and 4 months. The California OEHHA has a CREL for formalde-
hyde of 9 lg/m3. LEED has a criteria value for this chemical of 33 lg/m3.
ANSI/ASHRAE has a maximum criteria value of 55 lg/m3 for this chemical.
• Acetaldehyde: Preapplication airborne concentrations in the home averaged
34 lg/m3. This background level decreased immediately after application before
eventually reaching a peak airborne concentration of approximately 84 lg/m3 at
72 h postspray in the living spaces of the home. This concentration decreased to
an average level of 58 lg/m3 at the postspray test period of 4 months. The air-
borne concentration of acetaldehyde in the attic was 37 lg/m3 during the post-
application sampling periods of 48 h and 4 months. The California OEHHA
CREL for acetaldehyde is 140 lg/m3. The LEED criteria value for this chemical
is 70 lg/m3. The EPA has a chronic inhalation RfC value of 9 lg/m3.
• Propionaldehyde: Preapplication airborne concentrations in the home were all
ND (<9.6 lg/m3). During SPF application, airborne concentrations within the liv-
ing spaces of the home were found as high as 70.5 lg/m3 (second-floor bedroom)
but decreased immediately after the spray period to where all sample locations,
including the attic, were ND (from < 5.4 to < 7.3 lg/m3) at 12 h postspray. The
EPA has a chronic inhalation RfC value of 8 lg/m3 for this chemical.
 ECOFF ET AL., DOI: 10.1520/STP158920150034 109

2-Butoxyethanol
Airborne concentrations of 2-butoxyethanol were detected during SPF application
and at each postsample test period except Day 120. The highest airborne concentra-
tion of 2-butoxyethanol was 7,800 lg/m3 within the main attic space during SPF
application. Airborne concentrations of this chemical decreased in the main attic
space with each subsequent sample period, decreasing from a value in the attic of
115 lg/m3 at 12 h postspray to 67 lg/m3 at Day 7 and 23 lg/m3 at 1 month. The
airborne concentrations of 2-butoxyethanol in the living spaces during all the
postspray test periods ranged from ND (<15 lg/m3) to 49 lg/m3. ANSI/ASHRAE
62.1-2013 has an indoor air quality criteria value for this chemical of 967 lg/m3.
The EPA has a chronic inhalation RfC value of 1,600 lg/m3. All postspray airborne
concentrations of 2-butoxyethanol were below these criteria values.

CHAMBER EMISSION RESULTS AND CALCULATED CONCENTRATIONS

The micro-scale chamber results are reported as area-specific emission factors,
which are the units specified in the draft ASTM micro-scale chamber method; the
emission factors were calculated as described in ASTM D5116-10, Standard Guide
for Small-Scale Environmental Chamber Determinations of Organic Emissions from
Indoor Materials/Products [21]. The results are expressed as lg/m2 h in Table A.6.
The loading factor of SPF installed in the home was determined to be 0.2 m2/m3.
The loading factor was calculated by comparing the ratio of the surface area of exposed
SPF insulation in the attic to the actual volume of the interior living space. A standard
air exchange rate of a building with SPF insulation is specified in CAN/ULC-S774-09
(0.3 air changes/h) [22]. This air exchange rate is considered to be a conservative value
that results in an area-specific flow rate (N/L) equal to 1.43 m/h in the home compared
to the test conditions of the draft ASTM micro-scale chamber method at 0.93 m/h.
This external factor may affect the rate of emissions for SVOCs such as TCPP.
A well-mixed, single-zone, mass-balance model was used to calculate the
steady-state HFC-245fa and 1,2-DCE airborne concentrations. Because of the
screening nature and assumptions made in this model, building sink effects and
inter-room air flow were not accounted for during this study. Therefore, the model
was assumed to be applicable only for a well-mixed scenario of airborne VOC
concentrations in the home at steady state. The following equation can be used to
calculate the indoor air concentrations of VOC chemical emissions from applied
SPF in the home when the steady-state condition is reached:

C ¼ EF ðLÞð1=N Þ (1)

where:
C ¼ indoor air concentration, lg/m3,
EF ¼ emission factor, lg/m2 h,
L ¼ loading factor of product in building (surface area of applied SPF per air
volume of building), m2/m3, and
110 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

N ¼ number of outdoor air exchanges in structure per hour (also referred to as

the ventilation rate).
Table A.7 illustrates the estimated airborne concentrations in the home based on the
average emission factor of triplicate samples taken of all chemicals evaluated during
the micro-scale chamber work. During micro-scale chamber work, air samples were
collected at more frequent time periods than the field measurement data. For exam-
ple, micro-scale chamber data are shown from the time period when the samples
were received by the laboratory (approximately 36–48 h after spray application)
through Day 35.
All SPF chemical components except the two blowing agents (HFC-245fa and
1,2-DCE) were calculated to be at or below their LOQ in the home. In some cases,
because of the volume of air sampled during the field study, values above the
LOQ for chemicals such as 2-butoxyethanol and acetaldehyde were measured
within the living space of the home. For this study, we only compared mod-
eled and field measurement data to the chemicals that were consistently mea-
sured above the LOQ during micro-scale chamber testing and in the home.
This was because after the retrofit SPF installation in the attic of this home, it
took a few days before normal activities were established. For example, during
the first 2 days after installation, mechanical ventilation (i.e., a portable fan) was
present in the home and windows and doors were opened by the homeowner to
increase air exchange. Therefore, we decided to compare the micro-scale chamber
results with field measurement data starting from Day 7 when the indoor activities
were returned to normal. Any results before Day 7 were not included in the com-
parison because the variations of indoor activities were not well characterized. The
comparison results may not be used for prioritizing VOCs from SPF.
The primary chemicals measured both in the micro-scale chamber and in the
home, with calculated airborne concentrations well above their LOQ during all
test periods, were HFC-245fa and 1,2-DCE. These two chemicals were predicted
to be measured during all postspray test periods. The ratio of the modeled to field
measured airborne concentrations of these two VOCs at Days 7 and 28 are shown
in Table 4.
The airborne concentrations of HFC-245fa and 1,2-DCE are within an order of
magnitude between modeling and field measurement at Days 7 and 28. This study
illustrates that utilizing the micro-scale chamber emission factor and a screening-
level model could potentially serve as a preliminary assessment of chemical

TABLE 4 Chemical airborne concentration ratio between model and field measurement.

Chemical Day 7 Day 28

HFC-245fa 1.04 1.48

1,2-DCE 0.53 0.59
 ECOFF ET AL., DOI: 10.1520/STP158920150034 111

emissions from an SPF product. However, because of the assumptions made in this
screening-level model and that only two of the targeted chemicals (i.e., HFC-245fa
and 1,2-DCE) were detected throughout the chamber and field work, it may not be
applicable for every time point after SPF application or for all chemicals in this
study. More sophisticated models are needed to employ the micro-scale chamber
study results so that the fate and transport of chemicals from SPF can be better
characterized in a residential structure.
Fig. 1 illustrates the calculated airborne concentration values with variable ACH
and actual values measured in the home for the two blowing agents.

FIG. 1 Modeled and measured airborne concentrations of HFC-245fa and 1,2-DCE.

112 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

Considerations for Future Studies

The single-zone, well-mixed model employed in this study has several limitations.
First, it assumes steady-state conditions when the chemical emission rate from the
source equals the chemical removal rate from ventilation. However, such conditions
may not exist at every single time point at which the field monitoring is conducted.
Second, no indoor sinks or chemical reactions were considered, which may under-
estimate the indoor sink effects. Third, this model only has one zone that is well
mixed. In a residential home, multiple rooms exist, and inter-room air flow patterns
are complicated. To better characterize SPF emissions in an indoor residential envi-
ronment, more sophisticated indoor models that consider indoor sinks, air flow
patterns, and human activities are needed. An ASTM work item has been created to
address this effort. Future field studies are being planned to perform long-term
monitoring of indoor VOC/SVOC concentrations, characterize indoor air flow pat-
terns, and evaluate the impact of human activity. These studies and more sophisti-
cated model development would provide an opportunity to advance the field of
indoor air quality monitoring and modeling.
In addition to model development, field measurement studies also play an
important role in indoor air quality evaluations. They can be used to validate the
application of emission factors obtained from chamber studies and subsequently
applied in fate and transport models to estimate SVOC airborne concentrations. In
terms of field measurement studies, opportunities to improve future SPF chemical
emission study work include collecting duplicate field samples to help quantify and
clarify any uncertainties that cannot be explained with the data, documenting air
flow patterns/air exchange rates in the structure, and using more sensitive sam-
pling/analysis methods that will provide better data to evaluate model predictions
for all chemicals, including flame retardants. To improve the data capture rate with
less nondetects, separate TD tubes containing a stronger sorbent for very volatile
compounds such as the blowing agents should be used in addition to utilizing long-
er sampling times for amine catalysts and flame retardants. Micro-scale chamber
testing could also be performed at an elevated temperature (e.g., 35 C) to enhance
the detection of these compounds.

Conclusions
The key findings from this study are as follows:
• VOCs contained within the SPF insulation product, such as the two blowing
agents (HFC 245-fa and TDCE) and 2-butoxyethanol, were measured in the
home above their limits of detection after installation. The airborne concentra-
tions of the two blowing agents decreased with time but were still detected in
the home at approximately 1/1000th of their OEL (TLV) 16 months after the
installation of the product.
• Micro-scale chamber emission factors may be used together with a
screening-level, well-mixed, single-zone model to identify potential VOCs
 ECOFF ET AL., DOI: 10.1520/STP158920150034 113

of concern from SPF. With the assumption that steady state may be
reached during certain time periods 7 days after installation of the product,
the calculated airborne concentrations of VOCs (HFC 245-fa and TDCE)
were within an order of magnitude compared to the field measurement
data. Because of the nature of activities that occur in a home such as
HVAC unit cycles and the opening and closing of doors, steady-state con-
ditions may not be reached at all times. Therefore, using emission factors
combined with a well-mixed single-zone model is limited to prioritization
purpose, and further evaluation is needed.
• Amines and TCPP were not measured above the reported limit of detection
during the chamber emission study work. These same chemicals, as well as
MDI, were not detected within the living spaces of the home beyond the post-
spray test period of 24 h. Therefore, the simple, well-mixed single-zone model
for predicting airborne concentrations could not be applied for these
compounds.
• Airborne concentrations of MDI were less than the limit of detection (<0.4 to
<0.7 lg/m3) during all postspray air sampling test periods evaluated.
• Average background airborne concentrations of formaldehyde measured in
the home prior to SPF application were equal to or higher than during each
post air sampling test period.
• Airborne concentrations of acetaldehyde did increase in the living spaces
of the home after SPF application, but at no time did the acetaldehyde
concentrations exceed the California OEHHA CREL value. The airborne
concentrations of acetaldehyde were just above the USGBC LEED criteria
value of 70 lg/m3 at Day 3 after SPF application. The airborne con-
centrations of acetaldehyde were below this value at the postspray test
period of 4 months.

TABLE A.1 Total airborne MDI results: SPF attic installation.

Fixed Location During Spray After 12 h After 24 h After 48 h

Main attic (center, near access opening) 2.0 <0.4 <0.7 <0.6
West bedroom: 2nd floor 1.0 <0.5 <0.7 <0.6
East bedroom: 2nd floor <0.6 <0.5 <0.7 <0.6
Living (piano) room: 1st floor <0.6 <0.5 <0.7 <0.6
Kitchen: 1st floor <0.6 <0.5 <0.7 <0.6
3 0 0
Note: Reported values are in lg/m and are a combination of 2,4 - and 4,4 -MDI found on the
13-mm filter. The sample and analytical method used was Covestro Method 1.7.7 HPLC (modified
OSHA Method 47). The method sample flow rate was 1.0 L/min. Samples were collected for approx-
imately 140 min. A less than (<) sign means the concentration reported was less than the LOQ. The
LOQ for the analytical method was 0.1 lg MDI/sample. The main attic sample was collected near
the center of the attic, approximately 8 ft from the main access opening.
114 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE A.2 HFC-245fa and TDCE results using TD tubes: SPF attic installation.

During After After After After After 7 After 1 After 4 After 16

Fixed Location Spray 12 h 24 h 48 h 72 h Days Month Months Months

HFC-245fa
Main attic 106,000 6,200 9,980 7,780 5,050 3,900 2,690 NS 500
West bedroom 4,140 3,110 3,170 2,210 1,500 3,540 1,700 NS 1,130
East bedroom 2,750 1,920 3,760 3,150 2,100 3,300 1,630 NS 160
Living room 124 900 2,400 1,880 1,410 3,430 1,610 1,460 700
Kitchen 160 1,020 2,300 1,700 — 3,370 1,750 2,420 1,200
TDCE
Main attic 183,000 5,540 5,960 4,450 4,050 8,280 4,360 NS 1,220
West bedroom 2,260 1,700 1,700 1,150 815 3,990 2,350 NS 1,200
East bedroom 1,290 850 2,140 2,020 1,350 3,730 2,280 NS 1,360
Living room 11 205 995 765 625 3,670 2,280 1,680 1,170
Kitchen 18 245 935 685 — 3,510 2,370 2,760 1,440

Note: All values are in lg/m3. Samples were collected using TD tubes. The sample flow rate was
0.05 L/min. Samples collected in the attic were collected for 5 min. Samples collected in the living
space of the home were collected for 20 min. The analytical method LOQ for HFC-245fa was
0.05 lg. The LOQ for TDCE was 0.0008 lg. NS ¼ no sample obtained.

TABLE A.3 TCPP results using TD tubes: SPF attic installation.

TCPP During Spray 12 h 24 h 48 h 72 h 7 Days 1 Month

Main Attic 1,300 72 89 119 68 19 < 17

West BR 25 <17 20 <17 <17 <17 <17
East BR 23 <17 <17 <17 <17 <17 <17
Piano Room <17 <17 <17 <17 <17 <17 <17
Kitchen <17 <17 <17 <17 NS <17 <17
Crawl Space NS <68 <68 <68 NS NS NS
3
Note: All values are in lg/m . A less than (<) sign means the concentration reported was less than
the LOQ. The LOQ for the analytical method was 0.026 lg. The sample flow rate was 0.05 L/min.
Samples collected in the attic were collected for 5 min. Samples collected in the living space of the
home were collected for 20 min. NS ¼ no sample obtained.

TABLE A.4 Aldehyde results using silica gel sorbent tubes: SPF attic installation.

Sample Prespray During Spray After

Location Backgrounda App. After 12 h After 24 h After 48 h After 72 h 4 Months

Formaldehyde
Main attic NS 16.0 <3.3 9.5 20.8 NS 19.3
West bedroom 32.7 8.9 <4.4 12.5 22.9 26.8 18.6
East bedroom 36.5 9.3 <4.4 <6.9 25.0 25.0 43.3
Living room 23.1 7.2 <4.6 6.4 <6.3 26.8 NS
Kitchen NS <6.5 7.3 <6.7 10.2 21.4 NS
 ECOFF ET AL., DOI: 10.1520/STP158920150034 115

TABLE A.4 (Continued)

Sample Prespray During Spray After

Location Backgrounda App. After 12 h After 24 h After 48 h After 72 h 4 Months

Acetaldehyde
Main attic NS 31.1 10.9 14.8 37.5 NS 36.7
West bedroom 34.6 16.8 12.5 14.1 39.6 83.9 53.0
East bedroom 34.6 13.5 16.0 <11.5 43.8 78.6 63.3
Living room 32.7 <10.9 <7.6 12.8 <10.4 83.9 NS
Kitchen NS <10.9 13.1 <11.2 27.1 76.8 NS
Propionaldehyde
Main attic NS 126.8 <5.4 <11.4 <10.4 NS NS
West bedroom <9.6 70.5 <7.3 <9.9 <10.4 <8.9 <8.9
East bedroom <9.6 23.9 <7.3 <11.4 <10.4 < 8.9 <8.9
Living room <9.6 <10.9 <7.6 < 10.6 <10.4 < 8.9 NS
Kitchen NS <10.9 <7.3 <11.2 <10.4 < 8.9 NS
3
Note: All values are in lg/m . A less than (<) sign means the concentration reported was less than
the analytical LOQ. The LOQ for each aldehyde was 0.16 lg. The samples were collected using
dinitrophenylhydrazine-treated silica gel tubes. The sample flow rate was 0.2 L/min. The sample
time was approximately 120 min. NS ¼ no sample obtained.
a
Pre-SPF application air samples were obtained to assess ambient or background levels in the home
prior to the start of the SPF installation.

TABLE A.5 2-Butoxyethanol results using TD tubes: SPF attic installation.

Sample During After 7 After 1 After 4

Location Spray After 12 h After 4 h After 48 h After 72 h Days Month Months

Main Attic 7,800 118 90 115 90 67 23 NS

West bedroom 65 39 39 25 17 21 <15 NS
East bedroom 54 20 36 49 28 19 20 NS
Living room <15 <15 <15 <15 <15 17 <15 <15
Kitchen <15 <15 <15 <15 NS 16 22 <15
Crawlspace NS NS 73 50 28 NS NS NS
3
Note: All values are in lg/m . A less than (<) sign means the concentration reported was less than
the analytical LOQ. The LOQ for the method used was 0.022 lg. The sample flow rate was
0.05 L/min. Samples collected in the attic were collected for 5 min. Samples collected in the living
space of the home were collected for 20 min. NS ¼ no sample obtained.

TABLE A.6 Area-specific emissions factors with micro-scale test chambers.

Target Compound Day 0 Day 1 Day 2 Day 3 Day 4 Day 7 Day 11 Day 14 Day 21 Day 28 Day 35

HFC-245fa 92,960 101,400 50,900 34,070 20,170 5,080 3,840 3,090 3,310 3,520 2,370
1,2-DCE 110,310 35,470 21,260 15,030 9,830 2,850 2,880 2,120 1,690 1,980 1,050
Chlorobenzene <19 <19 <19 <19 <19 <19 <19 <19 <19 <19 <19
2-Butoxyethanol <34 <34 <34 <34 <34 <34 <34 <34 <34 <34 <34
116 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE A.6 (Continued)

Target Compound Day 0 Day 1 Day 2 Day 3 Day 4 Day 7 Day 11 Day 14 Day 21 Day 28 Day 35

Amine #1 <30 <30 <30 <30 <30 <30 <30 <30 <30 <30 <30
Amine #2 <58 <58 <58 <58 <58 <58 <58 <58 <58 <58 <58
Amine #3 <333 <333 <333 <333 <333 <333 <333 <333 <333 <333 <333
TCPP <84 <84 <84 <84 <84 <84 <84 <84 <84 <84 <84
Formaldehyde <3.1 <3.1 <3.1 <3.1 <3.1 NA NA NA <2.5 NA NA
Propionaldehyde 14.7 5.5 4.2 3.9 <3.1 NA NA NA <2.5 NA NA
Acetaldehyde 9.6 <3.0 <2.5 <2.9 <3.9 NA NA NA <2.9 NA NA

Note: All values are in lg/m2 h. All tests were done in triplicate. Mean results are reported. Relative
standard deviations ranged from 2 % to 23 %. Because HFC-245fa and 1,2-DCE exceeded the cali-
bration curve at Day 0, results have been estimated. NA ¼ not analyzed.

TABLE A.7 Estimated concentrations in the home.

Target Compound Day 0 Day 1 Day 2 Day 3 Day 4 Day 7 Day 11 Day 14 Day 21 Day 28 Day 35

HFC-245fa 65,070 70,940 35,640 23,850 14,120 3,560 2,690 2,170 2,310 2,470 1,660
1,2-DCE 77,220 24,830 14,880 10,520 6,880 1,990 2,020 1,480 1,190 1,380 736
Chlorobenzene <13 <13 <13 <13 <13 <13 <13 <13 <13 <13 <13
2-Butoxyethanol <15 <15 <15 <15 <15 <15 <15 <15 <15 <15 <15
Amine #1 <21 <21 <21 <21 <21 <21 <21 <21 <21 <21 <21
Amine #2 <39 <39 <39 <39 <39 <39 <39 <39 <39 <39 <39
Amine #3 <233 <233 <233 <233 <233 <233 <233 <233 <233 <233 <233
TCPP <59 <59 <59 <59 <59 <59 <59 <59 <59 <59 <59
Formaldehyde <2 <2 <2 <2 <2 NA NA NA <2 NA NA
Propionaldehyde 10.3 3.9 3.0 2.7 <2 NA NA NA <2 NA NA
Acetaldehyde 6.7 <2 <2 <2 <2 NA NA NA <2 NA NA
3
Note: All values are in lg/m . Results are based on the average emission factor of triplicate micro-
scale chamber tests (loading factor ¼ 0.2 m2/m3; air exchange rate ¼ 0.3 exchanges/h). Because
HFC-245fa and 1,2-DCE exceeded the calibration curve at Day 0, reported results have been esti-
mated. NA ¼ not analyzed.

References

[1] American Chemistry Council, Center for the Polyurethanes Industry, “Why Select Spray
Polyurethane Foam (SPF) Insulation?,” https://spraypolyurethane.org/Main-Menu-Cate-
gory/Consumers/Why-SPF (accessed December 31, 2015).

[2] Shin, H. M., McKone, T. E., Tulve, N. S., Clifton, M. S., and Bennett, D. H., “Indoor
Residence Times of Semivolatile Organic Compounds: Model Estimation and Field
Evaluation,” Environ. Sci. Tech.,” Vol. 47, No. 2, 2013, pp. 859–867.
 ECOFF ET AL., DOI: 10.1520/STP158920150034 117

[3] Weschler, C. J. and Nazaroff, W. W., “Semivolatile Organic Compounds in Indoor Envi-
ronments,” Atm. Environ., Vol. 42, No. 40, 2008, pp. 9018–9040.

[4] Xu, Y. and Little, J. C., “Predicting Emissions of SVOCs from Polymeric Materials and
Their Interaction with Airborne Particles,” Environ. Sci. Tech., Vol. 40, No. 2, 2006,
pp. 456–461.

[5] Weschler, C. J. and Nazaroff, W. W., “SVOC Partitioning Between the Gas Phase and Set-
tled Dust Indoors,” Atm. Environ., Vol. 44, No. 30, 2010, pp. 3609–3620.

[6] Weschler, C. J., Salthammer, T., and Fromme, H., “Partitioning of Phthalates Among the
Gas Phase, Airborne Particles and Settled Dust in Indoor Environments,” Atm. Environ.,
Vol. 42, No. 7, 2008, pp. 1449–1460.

[7] Liu, C., Morrison, G. C., and Zhang, Y., “Role of Aerosols in Enhancing SVOC Flux
Between Air and Indoor Surfaces and Its Influence on Exposure,” Atm. Environ., Vol. 55,
August, 2012, pp. 347–356.

[8] U. S. EPA, “IRIS Assessments,” http://www.epa.gov/iris (accessed December 31, 2015).

[9] American National Standards Institute, Ventilation for Acceptable Indoor Air Quality,
American Society of Heating, Refrigerating and Air-Conditioning Engineers, Atlanta, GA,
2013.

[10] U. S. Green Building Council, “LEED IDþC: Commercial Interiors, Version 4, Indoor Air
Quality Assessment,” http://www.usgbc.org/credits/commercial-interiors-retail-com-
mercial-interiors-hospitality-commercial-interiors/v4/eq114, 2015 (accessed February
20, 2015).

[11] California Office of Environmental Health Hazard Assessment, “Acute, 8-Hour and
Chronic Reference Exposure Levels (RELs) 2014,” http://www.oehha.ca.gov/air/all-
rels.html (accessed February 22, 2015).

[12] American Conference of Governmental Industrial Hygienists (ACGIH), TLVs and BEIs,
ACGIH, Cincinnati, OH, 2015.

[13] ASTM D7706-11, Standard Practice for Rapid Screening of VOC Emissions from Products
Using Micro-Scale Chambers, ASTM International, West Conshohocken, PA, 2011,
www.astm.org

[14] ASTM D7859-13e1, Standard Practice for Spraying, Sampling, Packaging, and Test Speci-
men Preparation of Spray Polyurethane Foam (SPF) Insulation for Testing of Emissions
Using Environmental Chambers, ASTM International, West Conshohocken, PA, 2013,
www.astm.org

[15] EPA/625/R-96/010b, Compendium of Methods for the Determination of Toxic Organic

Compounds in Ambient Air, Second Edition, Compendium Method TO-17, Determination
of Volatile Organic Compounds in Ambient Air Using Active Sampling onto Sorbent
Tubes, EPA, Cincinnati, OH, 1999.

[16] ISO 16000-6:2011, Indoor Air—Part 6: Determination of Volatile Organic Compounds in

Indoor and Test Chamber Air by Active Sampling on Tenax TA Sorbent, Thermal Desorp-
tion and Gas Chromatography Using MS or MS-FID, ISO, Geneva, Switzerland, 2011.
118 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

[17] ASTM D5197-09e1, Standard Test Method for Determination of Formaldehyde and Other
Carbonyl Compounds in Air (Active Sampler Methodology), ASTM International, West
Conshohocken, PA, 2009, www.astm.org

[18] Sebroski, J., “Research Report for Measuring Emissions from Spray Polyurethane Foam
(SPF) Insulation,” presented at the 2012 Polyurethanes Technical Conference, Atlanta,
GA, September 24–26, 2012, http://polyurethane.americanchemistry.com/Resources-
and-Document-Library/Research-Report-for-Measuring-Emissions-from-Spray-Polyure-
thane-Foam-SPF-Insulation.pdf (accessed September 1, 2016).

[19] U. S. Government Publishing Office (GPO), “40 CFR Appendix B to Part 136—Definition
and Procedure for the Determination of the Method Detection Limit—Revision 1.11,” 2011,
GPO, Washington, DC.

[20] Ozbas, B., Vincent, J. L., Tobias, J. D., Rogers, J. R., and Burdeniuc, J. J., “Low- or Non-
Emissive Amine Catalysts for Low Density, Open-Cell Spray Polyurethane Foam,” pre-
sented at the 2012 Polyurethanes Technical Conference, Atlanta, GA, September 24–26,
2012.

[21] ASTM D5116-10, Standard Guide for Small-Scale Environmental Chamber Determinations
of Organic Emissions from Indoor Materials/Products, ASTM International, West Consho-
hocken, PA, 2010, www.astm.org

[22] CAN/ULC-S774-09 (R2014), Standard Laboratory Guide for the Determination of Vola-
tile Organic Compound Emissions from Polyurethane Foam, Standards Council of Cana-
da, Ottawa, ON.
 DEVELOPING CONSENSUS STANDARDS FOR MEASURING CHEMICAL EMISSIONS 119

STP 1589, 2017 / available online at www.astm.org / doi: 10.1520/STP158920150035

Richard S. Duncan1

Computer Simulation of Peak

Temperatures in Spray
Polyurethane Foam Used in
Residential Insulation
Applications
Citation
Duncan, R. S., “Computer Simulation of Peak Temperatures in Spray Polyurethane Foam Used
in Residential Insulation Applications,” Developing Consensus Standards for Measuring
Chemical Emissions from Spray Polyurethane Foam (SPF) Insulation, ASTM STP1589,
J. Sebroski and M. Mason, Eds., ASTM International, West Conshohocken, PA, 2017, pp. 119–137,
http://dx.doi.org/10.1520/STP1589201500352

ABSTRACT
Test methods to measure emissions of specific chemical compounds for spray
polyurethane foam (SPF) insulation are currently under development within
the ASTM D22.05 Subcommittee on Indoor Air. Although most of this work has
focused on small chamber emissions testing at room temperatures,
supplemental emissions testing at elevated temperatures of 40 C (104 F) and
65 C (150 F) are being considered. The applicability of these elevated
temperatures needs to be determined based on in situ temperatures
experienced by SPF. In lieu of costly, time-consuming field measurements of
SPF temperatures, this work used a one-dimensional transient dynamic
numerical simulation to estimate SPF temperatures. This simulation was
conducted over a 1 year period at hourly increments for SPF installed in wall
and roof assemblies located in Miami and Phoenix. The results show that the
outer layer of SPF in these assemblies can exceed 40 C for 8 to 10 h during
each day of the summer months. However, the inner layers of the SPF remain
within 1 to 2 C of room temperature. Due to these temperature variations with

Manuscript received May 2, 2015; accepted for publication March 7, 2016.

1
Spray Polyurethane Foam Alliance, 3927 Old Lee Hwy., Fairfax, VA 22030
2
ASTM Symposium on Developing Consensus Standards for Measuring Chemical Emissions from Spray
Polyurethane Foam (SPF) Insulation on April 30–May 1, 2015 in Anaheim, CA.

Copyright V
C 2017 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959.
120 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

time and position in these assemblies exposed to real-world conditions,

elevated temperature emission data from conventional test chambers may
have limited use for contaminant modeling, but elevated temperature emission
data may have other uses, such as for comparison testing or advanced
modeling.

Keywords
spray polyurethane foam, SPF, insulation, maximum temperature, building
assemblies, computer simulation, WUFI, attic insulation, unvented attics

Introduction
EMISSIONS TESTING FOR SPRAY POLYURETHANE FOAM
Since the late 1970s, increasing energy costs have had a significant influence on the
evolution of building enclosure design and energy efficiency. Changes to building
energy codes have increased the efficiency of the building enclosure by requiring
the use of more insulation while simultaneously reducing infiltration of uncondi-
tioned air. Although these changes to building enclosure design save energy,
reduced air infiltration can have a negative impact on indoor air quality. Reduction
in air leakage reduces the air infiltration and exfiltration in buildings, which, with-
out a well-designed and properly operated ventilation system, can cause a variety of
issues such as mold and mildew from accumulated moisture as well as retention of
a wide range of chemical emissions from furniture, electronics, appliances, interior
finishes, and building products as well as inhabitants (people and pets) and process-
es (manufacturing, cooking, etc.).
Concerns with the accumulation of chemical emissions inside the built environ-
ment have resulted in the development of several test methods to measure levels of
volatile organic compounds (VOCs) emitted from a variety of materials. Several
small chamber test methods, such as CDPH/EHLB Std. Method V1.1, Standard
Method for the Testing and Evaluation of Volatile Organic Chemical Emissions from
Indoor Sources Using Environmental Chambers, Version 1.1 [1], have been devel-
oped to measure general VOC and organic emissions from a wide range of building
materials. Small-chamber methods have been further refined to measure specific
emissions from a specific set of materials. ASTM D6007-14, Standard Test Method
for Determining Formaldehyde Concentrations in Air from Wood Products Using a
Small-Scale Chamber [2], measures formaldehyde emissions from wood products.
ASTM D6803-13, Standard Practice for Testing and Sampling of Volatile Organic
Compounds (Including Carbonyl Compounds) Emitted from Paint Using Small Envi-
ronmental Chambers [3], targets preparation and testing of VOCs from paints and
coatings. CAN/ULC-S774, Standard Laboratory Guide for the Determination of
Volatile Organic Compound Emissions from Polyurethane Foam [4], is a guide
describing measurement protocols for VOCs specifically from spray polyurethane
 DUNCAN, DOI 10.1520/STP158920150035 121

foam (SPF). These small-chamber test methods provide VOC data that can be used
for the following purposes:
• To screen or rank different products based on organic emissions
• To measure the effect of external variables (temperature, humidity, ventilation
rate) on emission rates
• To provide compound-specific emission data to models to predict indoor con-
centrations of these compounds
• To assist building designers and manufacturers on the selection of products
and ingredients
These tests have provided the basis for VOC qualification programs such as
UL Environment’s GREENGUARD Certification program [5]. However, small-
chamber methods can be somewhat limited because they use analytical chemistry
techniques that target a fixed set of VOCs and hence can be specific and nontrans-
ferrable across different classes of materials. These general guides and methods do
not focus on those compounds that are specifically used in the preparation of SPF,
such as methylene diphenyl diisocyanate, fluorinated blowing agents (hydrofluoro-
carbons, hydrofluoroolefins), amine and metal catalysts, and flame retardants.
During the last few years, the ASTM International D22.05 Subcommittee on
Indoor Air initiated several work items to develop standard methods to measure
the emissions from SPF, and in 2013 ASTM published a guide for spraying, sam-
pling, packaging, and test specimen preparation of SPF [6]. To date the develop-
ment of these new test methods has focused on emissions measured at room
temperature (23 C). However, microscale chambers allow the emissions to be mea-
sured at elevated temperatures (e.g., 40 C and 65 C).
The emission rates of organic compounds are influenced by temperature; higher
temperatures will generally result in higher emission rates because the temperatures
approach the boiling point of the respective compounds. Measuring emissions at
extreme elevated temperatures, although possibly useful for material comparison pur-
poses, can also create challenges for equipment and procedures that have been developed
for room temperature applications. Testing protocols should be developed at real-world
elevated temperatures to minimize systemic measurement errors that may arise from
using equipment outside their calibrated temperature ranges. In addition, severe elevated
temperature emissions rates must be used prudently with emission modeling, and service
temperatures of the materials being tested should be carefully considered.

BUILDING ASSEMBLIES USING SPF

SPF is used as a key component of a building enclosure as an insulation to reduce
heat flow. When SPF is installed at typical insulation thicknesses, it is an air-
impermeable material [7]. Because it expands and cures in place, it seals cracks and
gaps in the structure to significantly lower building air leakage (i.e., undesirable or
unintended natural ventilation), thereby improving energy efficiency. Closed-cell
foam is also water and moisture resistant and may be used as a water-resistant barri-
er (exterior roofing) or as a code-compliant vapor retarder in cold climates. SPF
insulation provides distinctly different performance benefits than fibrous insulations [8].
122 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

Insulation materials inherently develop a temperature gradient throughout their

installed thickness compared to noninsulating building materials in a building assembly.
This temperature gradient is inversely proportional to the thermal conductivity of the
material. A high thermal gradient indicates low heat transfer through that material.
Building products can also store or retain heat, as measured by their thermal capaci-
tance. As indoor and outdoor conditions change by day and by season, the temperature
gradients and peak temperature through the building assembly can change by the hour.
In addition, materials used in building enclosure assemblies can be subjected to wide
fluctuations in temperature and moisture content depending upon outdoor conditions.
Similarly, materials used in the building assembly may contain and retain different
levels of moisture as a result of processing, storage, construction, and environmental
conditions. Using different building materials will change thermal and moisture trans-
mission and storage capabilities of the assembly and will impact modeling results.
These variations in temperatures are highest near the outside of the building enclosure.

OBJECTIVE
The objective of this paper is to estimate peak temperatures, temperature profiles,
and time-at-temperature of SPF insulation encountered in typical real-world build-
ing assemblies located in hot climates. These estimated peak temperatures can be
used by researchers to define realistic elevated test temperatures for SPF emission
testing and to develop strategies for indoor emission modeling.
Although insulation temperature profiles can be measured in the field, the process is
costly as well as dependent upon climate and construction practices. There are several
simulation tools used to predict the transient temperature and moisture levels within
different materials in a building assembly. A computer-based simulation tool will be used
to estimate peak temperatures and temperature profiles within typical building enclosure
assemblies (walls and roofs) for both open- and closed-cell SPF used in hot climates.
It should be noted that the framed roof and wall assemblies selected for this study do
not include every building assembly, but it is believed that they provide typical examples
of U.S. residential construction practices used in the corresponding climate zones.

Simulation and Modeling

PHYSICAL PRINCIPLES
The transient dynamic behavior of energy (temperature) and moisture levels within
materials of a building enclosure assembly is governed by diffusion processes based on
fundamental physical principals. Because energy transfer may be moisture dependent
and moisture transfer may be temperature (energy) dependent, the behavior of tem-
perature and moisture is often a coupled phenomenon. From basic principles, the gen-
eral equations for coupled moisture and heat transfer are provided in Eqs 1 and 2 [9].
It is important to note that these coupled equations do not account for fluid flows,
such as movement of air or bulk water. Fluid flow and its associated energy transport
is minimal in most properly constructed building assemblies. For the purpose of this
 DUNCAN, DOI 10.1520/STP158920150035 123

study, any air movement through the modeled building assemblies would only serve
to lower the peak temperatures of the SPF insulation.
It should also be noted that the material properties of these equations vary
through the assembly. For example, a typical wall is made from siding, sheathing,
insulation, and gypsum board, so that the variables and material properties vary by
position as well as temperature and moisture content and are not constant. When
these equations are solved simultaneously, a hygrothermal analysis is completed:

Energy transfer
 
@H @h d
 r  ðkrhÞ  hv r  rð/psat Þ ¼ 0 (1a)
@h @t l

The first term is the energy storage term, where:

@H
@t ¼ specific heat.
The second term is heat conduction, and the third term is latent heat from water
phase change. For one-dimensional models, Eq 1a becomes:
   
@H @h @ @h @ d @p
 k  hv ¼0 (1b)
@h @t @x @x @x l @x

Mass transfer
 
@w @h d
qw  r  qw Dw r/ þ rð/psat Þ ¼ 0 (2a)
@h @t l

The first term is the moisture storage term, and the second term contains the liquid
conduction (capillary) and vapor diffusion terms, respectively. For one-dimensional
models, Eq 2a becomes:
 
@w @h @ @/ d @p
qw  q Dw þ ¼0 (2b)
@h @t @x w @x l @x

where:
h ¼ temperature; K;
k ¼ thermal conductivity; mW K ;
J
hv ¼ latent heat of phase change; kg ;
kg
d ¼ water vapor diffusion in air; m  s  Pa ;
l ¼ water vapor diffusion factor of dry material;
d kg
l ¼ dd ¼ water vapor permeance of material; m  s  Pa ;
psat ¼ saturation pressure; Pa;
p ¼ /psat ¼ water vapor partial pressure; Pa;
H ¼ enthalpy of moist material; mJ3 ;
124 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

J
c ¼ @H
@t ¼ specific heat of material; kg  K ;
kg
w ¼ moisture content; m2 ;
/ ¼ relative humidity;
qw ¼ density of water; mkg3 ;
kg
Dw ¼ liquid transport coefficient of material; ms :

WUFI SIMULATION SOFTWARE

When these coupled differential equations for energy (Eq 1) and moisture (Eq 2)
are solved simultaneously for every material in the assembly, the spatial and tempo-
ral values of temperature and moisture can be estimated. There are several software
programs designed to numerically solve these coupled equations. One of the most
widely used and trusted simulation tools for this application is WUFI [9]. WUFI is
a suite of software simulation tools jointly developed by the U.S. Department of
Energy’s Oak Ridge National Laboratory (ORNL) and Germany’s Fraunhofer Insti-
tute for Building Physics (IBP). WUFI ORNL/IBP numerically estimates the tran-
sient coupled one- and two-dimensional heat and moisture transport in multilayer
building assemblies exposed to natural weather. WUFI is based on the water vapor
and liquid transport in building materials and is compliant with ISO standards for
energy simulation software [10]. It is reported that WUFI thermal modeling is typi-
cally accurate to within 1 C based on communications with WUFI experts [11].
The WUFI ORNL/IBP software functionality can best be described by Fig. 1
[12]. The calculation process begins with the user defining the construction (order
and type of building materials), along with the orientation (one of eight cardinal
compass points) and the inclination of the assembly (angle from vertical position,
e.g., roofs). Initial conditions for temperature and moisture content are established
for each material. Material properties are either taken from the material property
library or input by the user as a custom material. In this study, all material proper-
ties were taken from the WUFI material data library, which is compliant with ISO
10456, Building Materials and Products—Hygrothermal Properties—Tabulated
Design Values and Procedures for Determining Declared and Design Thermal Values
[13]. Exterior climate conditions are applied from a meteorological database includ-
ed in the software. The user also selects indoor conditions that may vary by season
as well as the physical orientation (north-facing, south-facing, etc.) of the assembly.
Once the assembly, initial conditions, and climate data are defined, the software
uses numerical finite-difference algorithms to solve the energy and mass equations
at each time increment (typically 1 h). It uses the updated results from the previous
time increments to update any boundary conditions as well as temperature- and
moisture-dependent material properties.
Output includes animated profiles of temperature and moisture content at 1-h
increments throughout the duration of the simulation. In addition, the user may
define monitoring locations throughout the assembly. Plots of temperature and
moisture content versus time can also be developed for each location. Heat and
moisture fluxes can also be estimated at any desired location in the assembly.
 DUNCAN, DOI 10.1520/STP158920150035 125

FIG. 1 WUFI ORNL/IBP process flowchart.

WUFI has undergone rigorous evaluations using numerous laboratory and

field trials. It is frequently used by architects and engineers to perform hygro-
thermal modeling of building assemblies during building design and in some
cases forensic failure investigations of buildings with mold, mildew, and water
damage.

CONSTRUCTION
For this modeling study, the wall and roof assemblies were modeled with a southern
exposure to obtain the maximum solar heat gain. Fig. 2 shows the details of the roof
and wall assemblies used in this study. These assemblies are representative of typical
and common framed roof and wall designs; however, they do not represent each
and every building enclosure system. It should be noted that from 2009 to 2014, 93
to 95 % of all new single-family homes were constructed with wood framing and
cavity insulation [14]. These percentages vary by region and were lowest in the
southern United States (87–91 %).
126 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 2 Wall and roof assemblies used for WUFI modeling.

Wall Assembly
Residential wall assemblies are commonly wood-framed using 50.8 by 101.6-mm (2 by
4-in.) 2 by 4 studs on 406.4-mm (16-in.) centers with plywood or oriented strand board
(OSB) exterior sheathing and gypsum wallboard on the interior. SPF insulation is
applied directly to exterior sheathing to provide the cavity insulation R-value required
by energy codes. For R-13 cavity insulation, a 2 by 4 cavity will be completely filled with
open-cell SPF, whereas a 2 by 4 cavity will only need about 53.3 mm (2.1 in.) of closed-
cell foam, leaving a 35.6-mm (1.4-in.) air gap. Hot-climate claddings are typically light-
colored stucco instead of vinyl or wood siding due to high heat loads. The materials used
for this wood-framed wall assembly, from inside to outside, consisted of the following:
1. 12.7-mm (1=2-in.) gypsum board
2. Air space (note that R-13 in a 50.8 by 101.6-mm [2 by 4-in.] nominal stud
wall does not require full-cavity fill of closed-cell SPF)
 DUNCAN, DOI 10.1520/STP158920150035 127

3. SPF insulation to a thickness meeting R-13

4. 12.7-mm (1=2-in.) nominal thickness OSB
5. Weather-resistant barrier
6. 6.4-mm (1=4-in.) acrylic stucco
R-13 is the precise R-value defined in the U.S. Energy Code. The R-value is
expressed in IP units as h-ft2- F/Btu. There is an equivalent RSI in the SI system of
units: R-13 ¼ RSI 2.29. All material properties for this assembly are based on default
values provided by the WUFI software. To include the effects of convective heat
transfer of exterior wall surface, a film heat resistance coefficient of 0.059 m2-K/W
(0.334 h-ft2- F/Btu) was used. To model radiant heat transfer of the exterior wall
surface to the surroundings, a short-wave radiation absorptivity factor of 0.4 was
assumed for the bright stucco wall cladding. The short-wave reflectivity of the
surroundings was assumed to be 0.2. For convective heat transfer of the interior
wall surface to the indoor environment, a film heat resistance coefficient of
0.126 m2-K/W (0.71 h-ft2- F/Btu) was used.

Roof Assembly
Homes built in hot climates are typically one story. These homes have sloped roofs
with low roof pitches—typically less than about a 3:12 pitch (15 )—for two reasons.
First, a high-pitch roof is not needed to shed snow like roofs in colder climates.
Second, a low pitch minimizes the roof area and thus reduces construction costs and
lowers the total solar heat load on the building. Sloped roofs in extremely hot climates
are typically made of clay tile to provide better resistance to solar heating. If asphalt
shingles are used, the colors must be light to reduce solar heat gain. In this analysis, a
3:12 pitch was used. A red clay tile roof material was assumed for all roof models.
Roof decks are typically made from plywood, with an underlayment between
the wood deck and shingles to help protect the wood deck from water in the event
that the shingles are damaged or otherwise leak. When an unvented attic design is
used, SPF insulation is applied directly to the underside of the roof deck. Unvented
attics are used to cost effectively bring HVAC ductwork and equipment into pas-
sively conditioned space. It is estimated that more than 100,000 homes are con-
structed with unvented attics. The inside surface of SPF may or may not include a
fire-protective coating to meet building code requirements for 15-min thermal bar-
riers or ignition barriers. The materials used for this roof assembly, from inside to
outside, consisted of the following:
1. SPF insulation to a thickness meeting R-30 or R-38
2. 14.3-mm (9/16-in.) plywood roof deck
3. 30-min asphalt impregnated felt paper
4. Shingles
To include the effects of convective heat transfer of exterior roof surface, a film
heat resistance coefficient of 0.052 m2-K/W (0.297 h-ft2- F/Btu) was used. To model
radiant heat transfer of the exterior roof surface to the surroundings, a short-wave
radiation absorptivity factor of 0.67 was assumed for red tile shingles. The short-wave
128 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

reflectivity of the surroundings was assumed to be 0.2. For convective heat transfer of
the interior foam surface to the indoor environment of the attic, a film heat resistance
coefficient of 0.126 m2-K/W (0.71 h-ft2- F/Btu) was used.
All materials and their respective properties were taken from the WUFI materi-
al database. Initial values of temperature and moisture content were set to the
default values. All assemblies were assumed to have a southern orientation to maxi-
mize solar heat gain. Wall assemblies were inclined 90 from horizontal, and roof
assemblies were 15 from horizontal in all WUFI models. Temperatures at the inner
and outer surfaces of the foam were monitored in WUFI.

CLIMATE ZONES
In this project, the peak temperatures of SPF insulation in wall and roof temperatures
were estimated. Climates with high outdoor summer temperatures and high solar
heat loads will result in the highest SPF temperatures inside the wall and roof assem-
blies. For this reason, U.S. climates with the highest summer design temperature
and/or the lowest latitude for maximum solar heat loading were used where needed.
A summary of the climatic conditions for major U.S. cities is shown in Table 1
[15]. Based on these data, Miami and Phoenix were selected based on highest solar
heat load (lowest latitude) and highest summer design temperatures from all candi-
date cities listed in Table 1.
WUFI applied outdoor temperatures, humidity, rainfall, solar exposure, and
wind loads at hourly increments based on built-in weather data files. The annual
temperature profiles for Miami and Phoenix are shown in Fig. 3.

Results and Discussion

The WUFI simulation software was used to estimate the temperature variations through
the thickness of the SPF insulation at hourly intervals throughout the year. Annual varia-
tions in temperature for the inner and outer SPF surfaces in the wall assemblies are
shown in Fig. 4. Fig. 5 shows annual variations in temperature for the inner and outer
SPF surfaces in the roof assemblies. Fig. 6 shows the temperature and moisture profiles
through wall and roof assemblies. Tables 2 and 3 provide a summary of the WUFI results.
The exterior climate conditions in Miami and Phoenix are considerably differ-
ent, as shown by the annual temperature and relative humidity data in Fig. 3. The
hot and moist climate of Miami tends to moderate annually, and there are more
daily variations in temperature compared to the hot and dry climate of Phoenix.
These differences affect the maximum temperatures experienced by SPF insulation
in both wall and roof assemblies.

WALLS
Fig. 4 shows the annual variations in the inner (red) and outer (blue) temperatures of
the SPF insulation layer of the wall assemblies. In all four cases, it can be observed that
the temperatures of the inner surface of the SPF closely follow the interior room
 DUNCAN, DOI 10.1520/STP158920150035 129

TABLE 1 Average climatic conditions for several U.S. cities.

Winter
Design
IECC Temperature Heating Summer Design Temperature Latitude
Climate Degree Days,
      
City Zone C F HDD-18 C Dry, C Dry, F Wet, C Wet, F Degrees Minutes

Atlanta 3A 6 22 2,961 33 92 24 76 33 40
Boston 5A 13 9 5,634 31 88 23 74 42 20
Charlotte 4A 6 22 3,181 34 93 24 76 35 10
Chicago 5A 17 2 5,882 33 91 25 77 42 0
Dallas 3A 6 22 2,363 38 100 26 78 32 50
Denver 5B 17 1 6,283 34 94 17 63 39 10
Detroit 5A 14 6 6,232 29 85 22 72 42 20
Houston 2A 0 32 1,396 34 94 26 79 29 40
Jacksonville 2A 0 32 1,239 34 94 26 79 30 30
Los Angeles 3B 6 43 2,061 27 80 21 69 34 0
Miami 1A 8 47 214 32 90 26 79 25 50
Minneapolis 6A 24 12 8,382 32 89 24 75 44 50
New Orleans 2A 1 33 1,385 33 92 27 80 30 0
New York 4A 9 15 4,871 32 89 24 75 40 50
Phoenix 2B 1 34 1,765 42 107 24 75 33 30
Pittsburgh 5A 14 7 5,053 31 88 23 73 40 30
San 3B 3 38 3,015 25 77 18 64 37 40
Francisco
Seattle 4C 3 27 4,424 28 82 19 67 47 40
St. Louis 4A 14 6 4,900 34 94 25 77 38 50
Washington, 4A 8 17 4,224 33 91 25 77 38 50
DC

temperature shown in Fig. 3. In both climates the solar heat gain effect appears to be
greatest in late summer and early fall, where the warm air temperatures combined
with the lower angle of the sun provide a maximum solar heat gain for walls. As stated
previously, the outdoor temperatures of Miami have less daily and seasonal variation
than Phoenix, and these same variations are observed for the outer layer of the SPF.
In both cities, the use of open- versus closed-cell SPF has little effect on the peak
temperatures of the outer layer of SPF. However, there is about a twofold difference
between the daily variations of inner layer temperature. This higher daily variation
of inner layer temperature was caused by the air gap present when closed-cell foam
was used. This air gap provided a modest amount of thermal resistance that
separated the inner surface of the closed-cell SPF from that of the room interior.
From the results in Fig. 4, there were significantly greater outer-layer peak
temperatures in Phoenix compared to Miami. In Miami, the peak outer-layer
temperature was about 42 C (107 F) for both SPF product types. In Phoenix,
the peak outer-layer temperature was about 49 C (120 F) for closed-cell SPF
130 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 3 (a) Annual outdoor conditions for Miami; (b) annual outdoor conditions for
Phoenix; (c) annual indoor conditions for Miami and Phoenix.

and 52 C (126 F) for open-cell SPF. The duration of exposure to these elevated
temperatures was significantly different between the Miami and Phoenix walls.
In Miami, the outer-layer SPF temperature was above 40 C (104 F) at most 3 to
4 h per day. In Phoenix, the outer-layer SPF temperature exceeded 40 C for 8 to
11 h per day during most of the summer. The total exposure time of the outer
layer of SPF above 40 C was about 30 h per year in Miami and 800 to 1,300 h
per year in Phoenix. The peak temperature of the SPF wall insulation never
exceeded 65 C (150 F) in either Miami or Phoenix.

ROOFS
Fig. 5 shows the annual variations in the inner (red) and outer (blue) tempera-
tures of the SPF insulation layer of the roof assemblies. The annual and daily
variations of temperature were similar to those of the wall assemblies, but the
outer layer SPF of the roof assemblies reaches greater minimum and maximum
values than SPF in the walls. This difference was attributed to different solar in-
cidence and solar heat gain, as well as more significant radiative heat loss and
gain for the roofing surfaces.
 DUNCAN, DOI 10.1520/STP158920150035 131

FIG. 4 (a) Inner and outer SPF surface temperatures over 1-year simulation of Miami
open-cell SPF wall assembly; (b) inner and outer SPF surface temperatures over
1-year simulation of Phoenix open-cell SPF wall assembly; (c) inner and outer
SPF surface temperatures over 1-year simulation of Miami closed-cell SPF wall
assembly; (d) inner and outer SPF surface temperatures over 1-year simulation
of Phoenix closed-cell SPF wall assembly.

Fig. 6 shows the daily variation of the outer and inner SPF surfaces of an open-
cell SPF roof on a typical summer and winter day in Phoenix. These simulation
results follow the same approximate form and magnitude as roof assembly simula-
tions performed using finite-element analysis [16,17].
In both cities, the use of open- versus closed-cell SPF had little effect on the
temperatures of the outer layer of SPF in the roof. Unlike the wall assemblies, there
was no difference between the daily variations of inner-layer temperature. This is
because roof assemblies using closed-cell SPF did not have an air gap on the inside
surface, making the temperatures of the inside surface of both open- and closed-cell
SPF nearly identical to room temperature.
From the results in Fig. 5, there were significantly greater outer-layer peak tem-
peratures in Phoenix compared to Miami. In Miami, the peak outer-layer tempera-
ture was about 58 C (136 F) for both foams. In Phoenix, the peak outer-layer
132 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 5 (a) Inner and outer SPF surface temperatures over 1-y simulation of Miami open-
cell SPF roof assembly; (b) inner and outer SPF surface temperatures over 1-y
simulation of Phoenix open-cell SPF roof assembly; (c) inner and outer SPF
surface temperatures over 1-y simulation of Miami closed-cell SPF roof
assembly; (d) inner and outer SPF surface temperatures over 1-y simulation of
Phoenix closed-cell SPF roof assembly.

temperature was about 68 C (155 F) for both open- and closed-cell SPF. The dura-
tion of exposure to these elevated temperatures was similar between the Miami and
Phoenix walls. In Miami, the outer-layer SPF temperature was above 40 C at most
8 h per day. In Phoenix, the outer-layer SPF temperature exceeded 40 C for 11 h
per day during most of the summer. During select summer days, the outer-layer
SPF temperature exceeded 65 C for 2 to 3 h per day. The total exposure time of the
outer layer of SPF above 40 C was about 1,850 h per year in Miami and 2,030 h per
year in Phoenix. The peak temperature of the SPF roof insulation exceeded 65 C in
Phoenix for about 3 h per day or between 85 to 110 h per year.
These simulation results predict conservative, high-peak SPF temperatures. Red
tile shingles are often installed with ventilation air space between the shingles and
roof deck. Natural ventilation of the underside of the shingle will dramatically lower
the outer SPF surface temperatures.
 DUNCAN, DOI 10.1520/STP158920150035 133

FIG. 6 Inner and outer SPF surface temperatures over 1-day period during summer and
winter (Phoenix open-cell SPF roof assembly).

In addition, there are several design strategies that can reduce peak tempera-
tures in the SPF. The first is to use a very light-colored shingle. Simulations per-
formed for a roof assembly in Las Vegas have shown using a white shingle in place
of a black shingle can lower peak temperature by almost 40 C [11]. In this study, a

TABLE 2 WUFI results summary at day of maximum SPF temperature on roof deck.

Maximum SPF Temperature

Outer Surface Inner Surface

Time Time
Location SPF Insulation 
C 
F 
C 
F >40 C, h >65 C, h

Wall Miami (R-13) 89-mm (3.5-in.) open-cell 41.6 106.9 22.5 72.4 2 0
53-mm (2.1-in.) closed-cell 41.4 106.6 24.3 75.8 2 0
Phoenix (R-13) 89-mm (3.5-in.) open-cell 52.4 126.3 23.2 73.7 8 0
53-mm (2.1-in.) closed-cell 49.4 120.9 26.9 80.5 9 0
Roof Miami (R-30) 210-mm (8.3-in.) open-cell 58.0 136.3 22.6 72.6 8 0
122-mm (4.8-in.) closed-cell 57.5 135.6 22.7 72.9 8 0
Phoenix (R-38) 269-mm (10.6-in.) open-cell 68.7 155.6 22.8 73.0 11 3
155-mm (6.1-in.) closed-cell 68.2 154.8 22.8 73.1 11 3
 134
STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions
TABLE 3 WUFI results summary at day of longest time above 40 C and annual cumulative time above set temperature on roof deck.

Day of longest time >40 C Cumulative Time Per Year

Max SPF Temp. Outer Surface Max SPF Temp. Inner Surface
Time Time Time Time
Location SPF Insulation 
C 
F 
C 
F >40 C, h >65 C, h >40 C, h >65 C, h

Wall Miami (R-13) 89-mm (3.5-in.) open-cell 41.1 106.0 22.1 71.9 4 0 28 0
53-mm (2.1-in.) closed-cell 40.8 105.4 23.9 75.1 3 0 34 0
Phoenix (R-13) 89-mm (3.5-in.) open-cell 47.8 118.0 23.8 74.8 11 0 1278 0
53-mm (2.1-in.) closed-cell 48.6 119.5 26.1 78.9 9 0 848 0
Roof Miami (R-30) 210-mm (8.3-in.) open-cell 56.3 133.3 22.2 72.0 8 0 1866 0
122-mm (4.8-in.) closed-cell 56.6 133.9 22.7 72.9 8 0 1824 0
Phoenix (R-38) 269-mm (10.6-in.) open-cell 68.7 155.6 22.8 73.0 11 3 2043 109
155-mm (6.1-in.) closed-cell 68.2 154.8 22.8 73.1 11 3 2010 86
 DUNCAN, DOI 10.1520/STP158920150035 135

red shingle color was used. Another strategy to reduce SPF temperatures is to ventilate
the roof deck. In this approach, attic baffles or vent chutes are installed between the
wood roof deck and the SPF insulation. When these vent chutes are installed continu-
ously between the soffit and ridge vents, natural convection will be used to cool both
the underside of the roof deck/shingle assembly and the outer SPF surface. This natu-
ral cooling cannot be readily simulated using WUFI, but previous finite-element simu-
lations [16] have shown that ventilated attic roof decks and shingles operate about 7 to
10 F cooler than the unvented attic roof assembly modeled in this paper.
Fig. 7 shows the instantaneous temperature and moisture distributions predicted
by WUFI for a wall assembly with open-cell SPF insulation at 2:00 p.m. on June 23
for a south-facing wall in a home located in Miami. Although there was no specific
meaning in this particular result, it does clearly illustrate the typical temperature dis-
tribution in an assembly. It is important to note that the interior surface of the
foam, closest to the gypsum wall board, was within about 0.5 C of the interior
room temperature. The exterior surface of the foam, in contact with the exterior
OSB sheathing, was about 35 C (95 F), which was about the same temperature as
the OSB sheathing and the exterior surfaces of the wall. This result clearly showed
a linear temperature gradient across the SPF insulation. This temperature gradient
should be considered when modeling SPF emissions to the interior of the building.

FIG. 7 WUFI diagram of temperature and moisture content gradient through open-cell
SPF wall assembly in Miami.
136 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

Conclusions
The temporal results in Fig. 4 and Fig. 5 show that temperatures in the outer layer of
SPF can vary significantly, whereas the temperatures of the SPF inner layer closely
follow that of the interior conditioned space. Fig. 6 shows the variation in tempera-
ture through the Miami open-cell SPF wall assembly. The red line indicates the
temperature at midnight on June 21. The pink shading shows the variation of the
temperature through the entire year.
Although there is interest in characterizing SPF emissions at elevated tempera-
tures, it should be noted that the temperatures of the SPF insulation adjacent to the
interior space will remain close to interior room temperatures. The temperature pro-
file through the foam thickness will vary in a near linear fashion from 1 to 2 F above
room temperature on the inner layer to extreme high or low temperatures in the
outer layer, provided that the thermal conductivities of these materials are constant.
The high temperatures proposed for elevated temperature emissions testing
will only be reached inside the foam, near the outer layer. Although outer layers of
the SPF will exceed 40 C or 65 C (Phoenix roofs), these sections of foam are not
exposed to the interior air, and the emissions rate for these “encapsulated” regions
of elevated-temperature foam will not be as high as the emissions rate measured for
elevated-temperature SPF exposed to air (i.e., not encapsulated by cooler materials)
in the emissions testing chambers. It should be noted that emission chambers typi-
cally heat the specimen to a uniform temperature throughout, which is very differ-
ent than the temperature experienced by insulation products used in buildings. It is
important to recognize the temperature variations through the thickness of the SPF
layer when taking inventory of emission sources and using emissions data for
room-scale modeling.
The thermal gradients of the SPF in real-world applications (Fig. 6), coupled
with the fact that the high-temperature regions of the foam are surrounded by air-
impermeable materials to the outside and cooler foam to the inside, limit the utility
of elevated temperature emission data for estimating SPF emissions using conven-
tional small- or microscale chambers. Elevated temperature emission data may be
useful for the purposes of comparison in a laboratory setting, but the data should
not be used to predict indoor concentrations of chemical emissions. More sophisti-
cated modeling will be required to predict indoor concentrations of the emitted
compounds from insulation.

References

[1] CDPH/EHLB California Specification 01350, Standard Method for the Testing and Evalua-
tion of Volatile Organic Chemical Emissions from Indoor Sources Using Environmental
Chambers, Version 1.1, Environmental Health Laboratory Branch, California Department
of Public Health, Sacramento, CA, 2010.
 DUNCAN, DOI 10.1520/STP158920150035 137

[2] ASTM D6007-14, Standard Test Method for Determining Formaldehyde Concentrations
in Air from Wood Products Using a Small-Scale Chamber, ASTM International, West Con-
shohocken, PA, 2014, www.astm.org

[3] ASTM D6803-13, Standard Practice for Testing and Sampling of Volatile Organic Com-
pounds (Including Carbonyl Compounds) Emitted from Paint Using Small Environmental
Chambers, ASTM International, West Conshohocken, PA, 2013, www.astm.org

[4] CAN/ULC-S774, Standard Laboratory Guide for the Determination of Volatile Organic
Compound Emissions from Polyurethane Foam, Underwriters Laboratories of Canada,
Ottawa, 2014.

[5] UL Environment GREENGUARD Certification Program website, greenguard.org/en/

CertificationPrograms.aspx (accessed January 6, 2015).

[6] ASTM D22.05 Subcommittee WK40193, New Test Method for Estimating Chemical Emis-
sions from Spray Polyurethane Foam (SPF) Insulation Using Micro-Scale Environmental
Test Chambers, ASTM International, West Conshohocken, PA, 2013, www.astm.org

[7] Air Barrier Association of America, “Closed Cell, Medium-Density Spray Polyurethane
Foam Air Barrier Specification,” ABAA, Wapole, MA, 2013.

[8] “Spray Foam,” www.whysprayfoam.org (accessed January 7, 2015).

[9] “WUFI,” Fraunhofer Institute for Building Physics, www.wufi.de/index_e.html (accessed

January 6, 2015).

[10] “WUFI Documentation,” http://web.ornl.gov/sci/buildings/tools/wufi/ (accessed

October 29, 2015).

[11] Andre DeJarlais, private communication, 2015.

[12] “WUFI ORNL/IBP Hygrothermal Model,” web.ornl.gov/~webworks/cppr/y2001/pres/

111509.pdf (accessed January 6, 2015).

[13] ISO 10456, Building Materials and Products—Hygrothermal Properties—Tabulated

Design Values and Procedures for Determining Declared and Design Thermal Values,
ISO, Geneva, 2007.

[14] U.S. Census Bureau, “Characteristics of New Housing: Framing,” Washington, DC, 2015.

[15] Air Conditioning Contractors Association, “Manual J—Residential Load Calculation

v 2.10,” www.acca.org/technical-manual/manual-j (accessed January 2, 2015).

[16] Rudd, A. and Lstiburek, J., “Vented and Sealed Attics in Hot Climates,” ASHRAE Transac-
tions, Vol. 104, No. 2, 1998, pp. 1199–1210.

[17] Parker, D. S., Fairey, P. W., and Gu, L., “A Stratified Air Model for Simulation of Attic
Performance,” Insulation Materials: Testing and Applications, 2nd Vol., ASTM STP1116,
R. S. Graves and D. C. Wysocki, Eds., ASTM International, Philadelphia, PA, 1991,
pp. 44–71, http://dx.doi.org/10.1520/STP1116-EB

[18] ASTM D5116-10, Standard Guide for Small-Scale Environmental Chamber Determinations of
Organic Emissions from Indoor Materials/Products, ASTM International, West Conshohocken,
PA, 2010, www.astm.org
138 DEVELOPING CONSENSUS STANDARDS FOR MEASURING CHEMICAL EMISSIONS

STP 1589, 2017 / available online at www.astm.org / doi: 10.1520/STP158920150043

Ed Light1

Assessment and Remediation

of Misapplied Spray Polyurethane
Foam
Citation
Light, E., “Assessment and Remediation of Misapplied Spray Polyurethane Foam,” Developing
Consensus Standards for Measuring Chemical Emissions from Spray Polyurethane Foam (SPF)
Insulation, ASTM STP1589, J. Sebroski and M. Mason, Eds., ASTM International, West
Conshohocken, PA, 2017, pp. 138–147, http://dx.doi.org/10.1520/STP1589201500432

ABSTRACT
Misapplication of spray polyurethane foam (SPF) insulation may result in
occupant complaints associated with persistent odor. SPF installed in homes
may fail to completely cure when the contractor does not follow specified
procedures (e.g., with respect to the depth of individual layers, timing between
layer application, ratio, temperature, and mixing of SPF components). Few data
are available on emissions from misapplied spray polyurethane foam (MSPF),
and field practices used to control odors have not been validated. This paper
discusses strategies for resolving MSPF odor concerns and suggests an
assessment/mitigation protocol for field use pending further research. MSPF is
suggested by a persistent “fishy” type of odor after installation. A visual
inspection looking for discoloration and discontinuities may be helpful in
confirming the presence of MSPF and estimating its extent. Limitations in the
sensitivity and selectivity of air sampling methods available to field practitioners
may preclude the identification of contaminants associated with MSPF
emissions. Emissions testing of bulk samples facilitates the identification of
airborne contaminants under more concentrated conditions, but data
interpretation is subject to considerable uncertainty. Interim exposure reduction
pending remediation can be achieved by site isolation and ventilation. A
mitigation process has been suggested for resolving odors associated with MSPF

Manuscript received May 15, 2015; accepted for publication August 4, 2016.
1
Building Dynamics, LLC, 1216 Ashton Rd., Ashton, MD 20861
2
ASTM Symposium on Developing Consensus Standards for Measuring Chemical Emissions from Spray
Polyurethane Foam (SPF) Insulation on April 30–May 1, 2015 in Anaheim, CA.

Copyright V
C 2017 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959.
 LIGHT, DOI 10.1520/STP158920150043 139

that involves the removal of MSPF, cleanup, resealing the substrate with properly
cured SPF, and ventilation. Verification of odor control can be based on
systematic evaluation under specified conditions. Recommendations for refining
and validating assessment and mitigation protocols are presented. Additional
research is needed to better understand and resolve potential health risks.
Tertiary amines associated with SPF catalysts are significant contributors to
MSPF odor. Lower-emitting catalysts are now being introduced into SPF
products with a potential to reduce MSPF odors.

Keywords
spray polyurethane foam, indoor air quality, remediation, odor, amine,
assessment

Introduction
This paper addresses situations in which the misapplication of spray polyure-
thane foam (SPF) interferes with the curing process and unreacted components
are released that produce persistent odors. This is an infrequent event, occurring
when product application requirements are not followed (e.g., layers of SPF
exceed a specified thickness, insufficient time is allowed between layer applica-
tion, the ratio between the two product components is incorrect, the temperature
is outside the specified range, etc.) [1]. Slight odors may be detected from prop-
erly applied SPF immediately after installation. MSPF is suggested when a “fishy”
type of odor is initially strong or continues to be detected beyond a few weeks.
This odor gradually diminishes over time but has been reported to persist for
more than a year.
Although protocols have been established for resolving sources of indoor air
pollutants such as asbestos, mold, and corrosive drywall [2–4], efforts to control
odors from misapplied spray polyurethane foam (MSPF) have not been standard-
ized or validated. Protocols for restoring sites affected by other sources of contami-
nation have included the following elements:
• Problem identification based on observation of suspect conditions
• Inspection to characterize site conditions
• Project objectives based on target contaminant concentrations, odor elimina-
tion, etc.
• Interim site management to minimize exposure pending mitigation
• Mitigation measures to restore site
• Inspection or sampling to clear site for reoccupancy, or both

SPF is widely used to insulate and air seal new structures and weatherize older
buildings. Diisocyanates, the component presenting the most widely recognized
health concern, are only briefly airborne and have not been detected within a few
hours after application [5–7]. Industry guidance requires that occupants be protected
from exposure by vacating the structure during and immediately after application
[8–10]. Similar to other building materials, installed SPF emits various volatile
140 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

organic compounds (VOCs) that diminish over time [5–7]. SPF installation may
increase preexisting ambient concentrations of VOCs by reducing air infiltration.

Methodology
The scope of this paper is limited to odor concerns associated with MSPF (potential
health effects are not addressed). Findings are based on published literature and the
author’s field experience evaluating indoor air quality (IAQ) complaints attributed
to SPF. There are few published journal papers on this subject, with conference
papers and industry publications providing much of the available information.
Recommended procedures for assessment and mitigation are modeled after proto-
cols established for the restoration of other types of indoor contamination. The dis-
cussion is generally applicable to most open- and closed-cell SPF products but not
to single-component SPF sealants.

Emissions from Misapplied SPF

MSPF emissions may include unreacted catalysts (e.g., amines), flame retardants
(e.g., organophosphates), blowing agents (e.g., hydrofluorocarbons), and polyols
[1,7,9]. Because incomplete curing increases heat generated by the process, other
volatile reactants may also be emitted [1].
Preliminary results from two limited studies of MSPF emissions have been
reported. Genz et al. [1] compared emissions from foam applied per specifications
(SPF) to misapplied foam (MSPF). MSPF was created under laboratory conditions
using various combinations of deliberately created thick layers, off-specification
ratios, and damp conditions. Foam odor was strongest when layers exceeded speci-
fied thickness and components were off-ratio and persisted several days after appli-
cation. Tertiary amines appeared to be the primary source of odor, and the authors
suggested that added catalyst compounds may have reacted to form other amines at
the elevated temperature of MSPF (foam temperature measured up to 180 C com-
pared to 80 C in SPF applied per specification).
Genz et al. measured emissions from foam samples immediately after applica-
tion in a test facility at 140 air exchanges per hour. Diisocyanates were sampled per
ISO 14382, and VDA 278 was utilized to test for polyols, catalysts (amines), flame
retardants (triethyl phosphate), and blowing agents (hydrofluorocarbons). Emissions
were then tracked over time [1]. Measured emissions are summarized in Table 1.
Although no diisocyanates or polyols were detected in any samples, MSPF
emitted higher concentrations of amines, triethyl phosphate, and hydrofluorocar-
bons. After 10 days, all SPF emissions were below detection limits, whereas MSPF
emissions continued [1]. This study was limited in scope, with only one SPF prod-
uct tested and the analysis limited to listed product components.
The only other available study of MSPF emissions was summarized by a manufac-
turer that tested off-ratio foam (methodology not specified). No isocyanates were
 LIGHT, DOI 10.1520/STP158920150043 141

TABLE 1 Emissions from foam applied per specification (SPF) versus misapplied foam (MSPF).

Immediately After Application (mg/m3) 10 Days After Application (mg/m3)

1
SPF MSPF SPF MSPF1

Diisocyanates ND 10 ND ND
Polyols ND ND ND ND
Amines 2,500 5,000 ND ND
Triethyl phosphate 2,000 3,500 ND 1,000
Hydrofluorocarbons 1,500 1,500 ND 45,000

Note: From Ref. [1]. ND ¼ not determined.

1
Maximum concentration from various treatments (e.g., ratio, thickness).

detected within a few hours of application. Several VOCs were found at 24 h but were
nondetectable at 72 h. A tertiary amine continued to be measured up to 35 lg/m3
at 72 h [9].
Available information suggests that tertiary amines are important contributors
to SPF odor [1,6,7]. A wide variety of these compounds may be present depending
on product formulation. Amine catalysts SPF have traditionally been nonreactive,
leaving excess free gas after the A- and B-side SPF components were mixed. Some
SPF formulations have recently substituted “reactive” amine catalysts that reduce
odor by increased bonding with the other SPF reactants [11–13]. Studies have
shown amine emissions from reactive catalysts quickly decrease below the detection
limit, whereas emissions continued from nonreactive catalysts [14]. The use of non-
reactive catalysts may reduce detectable odor when misapplied.

Recognition of Misapplied SPF

MSPF odor varies in intensity and may not be readily identified without a systemat-
ic odor survey. The recognition of MSPF odor is facilitated by closing and warming
up the structure before inspection and enlisting an informal panel of individuals
capable of detecting common odors. Locating the odor source may be facilitated by
noting where suspect odor is strongest after isolating spaces (e.g., closing interior
doors or covering an insulated surface with plastic sheeting) and then noting odor
intensity.
Odors other than the fishy type of smell may originate from non-MSPF sour-
ces. Odor patterns can help distinguish MSPF from other sources. For example, the
odor should not have a preexisting SPF application. When used to weatherize
existing structures, preexisting odors may increase in intensity because of reduced
ventilation. Other odor complaints may be associated with construction off-gassing
unrelated to SPF or from increased dampness.
Although ambient air sampling at SPF sites measures the concentration of
some contaminants, it cannot be relied on to conclusively identify MSPF or quantify
occupant exposure. Indoor air sampling in noncompliant buildings typically
142 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

detects >100 compounds [15,16], and contaminants emitted by MSPF may not be
readily distinguished from normal background. Furthermore, MSPF emissions with
potential IAQ significance may not be detected because of limitations in the sensitivi-
ty and selectivity of analyses available from commercial laboratories (e.g., amines). In
contrast, the human nose is often capable of detecting and recognizing low concen-
trations of amines.

Site Inspection
Where persistent odor is suspected of originating from MSPF, visual inspection can
be helpful in confirming the source. MSPF may be identified visually by an inspec-
tor familiar with SPF characteristics. Cured SPF has uniform texture and coloration,
whereas MSPF may be distinguished by dark stains produced by elevated reaction
temperature, discontinuities (e.g., bubbles), or failure to solidify (A and B compo-
nents applied off-ratio) [17].
Indicators of misapplication may be visible on the surface, but core sampling
may be necessary to confirm the presence of discoloration and discontinuities [17].
Core sampling sites should include areas that appear to have the strongest odor.
Sealing core samples in plastic bags and then evaluating the headspace for the char-
acteristic odor can be useful in confirming the location of MSPF. From core sample
observations, the inspector can evaluate additional surfaces for darkening, bubbles,
or failure to solidify and may also be able to determine if layers exceeded specified
thickness.
Chamber testing of core samples facilitates the identification of compounds
associated with emissions by increasing contaminant concentrations. However, it
should be recognized that concentrations measured in chamber tests are not repre-
sentative of occupant exposure. In addition, contaminants from renovation activi-
ties and other sources may be adsorbed and then re-emitted from SPF.
Investigators have analyzed MSPF emissions for isocyanates, amines, VOCs, and
aldehydes, but interpreting emissions data is subject to the same limitations noted
previously for ambient air sampling.
In cases in which MSPF is localized, an inspection may suggest boundaries for
foam removal. In other situations, there may not be a clear boundary separating
MSPF from properly cured foam, and all SPF in areas with a persistent fishy type of
odor may need to be included in the scope of removal. Adjacent porous insulation
(e.g., batt fiberglass) may also be designated for removal when it is in contact with
MSPF. An MSPF site survey should also note structural and ventilation characteris-
tics of the affected area for developing a mitigation plan.

Interim Management
Careful planning is needed to effectively restore sites affected by MSPF, and occu-
pants may be exposed to emissions until the mitigation process is initiated. MSPF
 LIGHT, DOI 10.1520/STP158920150043 143

areas should be isolated and ventilated during this period. Areas adjacent to MSPF
can be separated from occupied space by sealing doors or erecting barriers. Plastic
sheeting can also be sealed directly over MSPF to retard emissions. Exhaust fans
can be operated to prevent odor migration [10], and energy recovery ventilators
can be installed to increase air exchange. This reduces exposure and may allow
occupants to remain in the home until mitigation (note that additional research is
needed to assess occupant health risks). In some cases, this additional period of
enhanced ventilation may sufficiently resolve odor concerns. However, if odor per-
sists, removing misapplied foam is necessary.

Mitigation
Some SPF manufacturers and contractors recommend removing and replacing
MSPF to resolve odor concerns [1,17]. There are no published procedures for
accomplishing this, and the efficacy of foam removal/replacement has not been
documented. Techniques have been developed by the industry for identifying defec-
tive foam [18] and removing overspray or foam not applied to specifications [17].
Similar to initial foam application, the site must be contained and ventilated,
followed by a detailed site cleanup. Substrates are then resealed with properly cured
SPF [17].
Protocols used to restore IAQ contaminated by other sources of material off-
gassing have employed a variety of steps to facilitate odor elimination, such as:
• Replacement of materials (e.g., removing as much of the contaminant as possi-
ble, followed by a complete cleanup)
• Surface treatment (e.g., disinfection or oxidation)
• Ventilation (e.g., air out to accelerate off-gassing)
• Encapsulation (e.g., apply sealant)
• Pathway elimination (e.g., permanently seal off or depressurize source area, or
both)
A combination of these steps is often used to ensure restoration, and some may
be useful for mitigating MSPF.

Verification
Areas in which indoor contaminants have been mitigated are often cleared for reoc-
cupancy based on inspection or sampling, or both. Inspection generally includes
visual confirmation that all specified material has been removed or treated, surfaces
are dust-free, and no odor associated with the contaminant is detected.
Limited sampling may not be representative of ongoing conditions, and detected
contaminants may originate from sources other than MSPF. There are no generally
accepted contaminant standards for IAQ, and clearance is often based on target
values suggested by various organizations and individuals. Because of these limita-
tions, the use of sampling to clear mitigation sites is subject to both false-negatives
(samples are in compliance but the odor persists) and false-positives (site is restored
144 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

but test criteria are exceeded). Verifying odor elimination after MSPF removal
should recognize that a slight odor may be detected from properly applied SPF for
several days.

Conclusions
1. Following misapplication, SPF may fail to completely cure, producing a persis-
tent odor often described as fishy.
2. Detection of a strong fishy odor continuing for more than a few days after
application is generally the initial indicator of MSPF.
3. Visual and tactile inspection, including examining core samples, may be help-
ful in confirming the presence of MSPF and estimating its extent. Site inspec-
tion is also needed to develop a mitigation plan.
4. Limitations in the sensitivity and selectivity of air sampling methods available
to field practitioners may preclude the identification of contaminants associat-
ed with MSPF emissions. Contaminants detected by VOC sampling may be
produced by a variety of sources, making it very difficult to determine whether
they are associated with MSPF.
5. Although testing foam emissions in a laboratory chamber facilitates the identi-
fication of compounds under more concentrated conditions, data interpreta-
tion relative to occupant exposure is subject to considerable uncertainty.
6. Interim exposure reduction pending remediation can be accomplished by
using plastic barriers and exhaust ventilation.
7. A mitigation process is suggested by some manufacturers and contractors
involving the removal of misapplied foam, cleanup, ventilation, and sealing
the substrate with properly cured SPF. The efficacy of this process has not
been documented.
8. Verifying odor mitigation can be accomplished by evaluation under worst-
case conditions.
9. Amines associated with product catalysts appear to be the primary contribu-
tors to MSPF odor. Some SPF product formulations have recently been modi-
fied, substituting reactive for nonreactive catalysts, potentially reducing the
potential for detectable odor when misapplied.

Recommendations
1. Additional research is needed to characterize MSPF emissions and health risks.
2. Pending further research, experience with restoration at sites with similar con-
tamination suggests that the following steps may be effective in controlling
odors from problem foams:
(a) Conduct a systematic odor survey to facilitate the identification of
MSPF. For at least a day prior to the survey, windows and exterior
doors should remain closed, and the home or building should be
warmed up several degrees above normal temperature. Interior
spaces should be isolated by closing doors and sealing areas with
plastic sheeting.
 LIGHT, DOI 10.1520/STP158920150043 145

(b) Delineate areas for MSPF removal based on the identification of

spray-foamed areas in which a fishy odor is detected, observation
of foam characteristics indicative of unreacted foam, and detection
of a fishy odor in the headspace of foam core samples.
(c) Isolate and ventilate areas with MSPF pending mitigation (i.e.,
HVAC off, vents sealed, exhaust fans operating, plastic sheeting
over contents, tack mats leaving work area). When depressurizing
an area, take care not to draw combustion gas back into the
building.
(d) Before MSPF removal is initiated, anticipate objects (e.g., fasteners,
wires, conduit, plumbing, cross-bracing, ductwork) that may be
hidden inside the SPF. After these have been located, turn off the
breakers to the AC circuits and power to other wiring and depres-
surize plumbing and gas lines.
(e) Remove foam in specified areas to the extent feasible.
(f) Clean all potentially affected surfaces until free of visible dust (e.g.,
consider HEPA vacuuming followed by damp wiping).
(g) Exhaust air until there is no detectable MSPF odor with windows
closed and HVAC operating with a normal thermostat setting. If
odor is still detected, repeat the above steps as needed.
(h) Reapply SPF (use formulation with reactive catalysts) in a manner
that ensures complete curing and seals all substrates previously
covered by MSPF.
(i) Re-clean and ventilate the work area.
3. Additional research is needed to refine and validate the assessment and mitiga-
tion protocols. The development of a standardized protocol for MSPF mitiga-
tion should address the following:
• Scope of removal. Is removing as much MSPF as possible, leaving resi-
due in cracks and crevices, sufficient?
• Work site containment. Can the home remain occupied during
mitigation?
• Ventilation design. Is a simple exhaust fan sufficient?
• Cleanup. Is HEPA vacuuming needed?
• Surface treatment. Is substrate encapsulation needed?
• Re-insulation. How should new SPF be applied to ensure odor
elimination?
• Secondary odor sources. How should nearby porous surfaces that may
have adsorbed MSPF emissions be addressed?
• Postremoval ventilation. How long should affected areas be aired out?
• Clearance. How should mitigation be confirmed?
4. MSPF that generates IAQ complaints can be prevented by adherence to prod-
uct installation requirements. Under current training and certification pro-
grams, the vast majority of SPF applications do not cause IAQ complaints.
Expanded efforts to train workers and ensure quality control should be
encouraged.
146 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

5. Substitution of reactive for nonreactive amine catalysts should be encouraged

to reduce amine emissions where SPF is misapplied. Additional research is
needed to verify the efficacy of reactive catalysts in reducing odor-causing
emissions from SPF and reducing occupant complaints.

References

[1] Genz, M., Schilling, U., Krasnow, J., Bockhoff, M., and Jansen, R., “Impact of Process
Parameters on Emissions from SPF,” presented at the Polyurethanes 2014 Technical Con-
ference, American Chemistry Council, Washington, DC, 2014.

[2] Light, E., Bailey, J., and Gay, R., “New Protocol for the Assessment and Remediation of
Indoor Mold Growth,” presented at the Proceedings of 12th International Conference of
Indoor Air Quality and Climate, ISIAQ, Austin, TX, July 20, 2011.

[3] Light, E., “Assessment and Mitigation of Corrosive Drywall,” presented at the Proceed-
ings of 12th International Conference of Indoor Air Quality and Climate, ISIAQ, Austin, TX,
July 20, 2011.

[4] U.S. EPA, Guidance for Controlling Asbestos-Containing Materials in Buildings, U.S. EPA,
Washington, DC, 1985.

[5] Lesage, J., Stanley, J., Karoly, W. J., and Lichtenberg, F. W., “Airborne MDI Concentra-
tions Associated with the Application of Polyurethane Spray Foam in Residential Con-
struction,” J. Occ. Env. Hyg., Vol. 4, No. 2, 2007, pp. 145–155.

[6] Havermans and Houtzager, “Emissions of Volatiles from Spray Polyurethane Foam Insu-
lated Crawl Spaces,” presented at the Proceedings of Indoor Air 2014, ISIAQ, Hong
Kong, July 7, 2014.

[7] Karlovich, J., Thompson, C., and Lambach, J., “A Proposed Methodology for Develop-
ment of Building Re-Occupancy Guidelines Following Installation of Spray Polyurethane
Foam Insulation,” presented at the 2012 Center of Polyurethanes Conference, Fairfax,
VA, September 20, 2012.

[8] Bayer MaterialScience, “Frequently Asked Questions About High-Pressure Application of

SPF,” https://sweets.construction.com/swts_content_files/153465/804691.pdf (accessed
May 1, 2015).

[9] Icynene, “Icynene Spray Foam Insulation,” http://www.icynene.com/sites/default/files/

icynene_over_25_years_of_proven_experience_2013_ebook.pdf (accessed May 20,
2015).

[10] American Chemistry Council, Ventilation Considerations for Spray Polyurethane Foam,
American Chemistry Council, Washington, DC, 2013.

[11] Casati, F., Sonney, J., Mispreuve, H., Fanget, A., Herrington, R., and Tu, J., “Elimination of
Amine Emissions from Polyurethane Foams: Challenges and Opportunities,” http://
msdssearch.dow.com/PublishedLiteratureDOWCOM/dh_003c/0901b8038003cbe3.pdf?
filepath=/polyurethane/pdfs/noreg/109-01552.pdf&fromPage=GetDoc (accessed April
30, 2015).
 LIGHT, DOI 10.1520/STP158920150043 147

[12] Wood, R., “CPI Ventilation Research Project Update,” presented at the Proceedings 2013
Polyurethanes Technical Conference, Fairfax, VA, September 12, 2013.

[13] Ozbas, B., Tobias, J., Rogers, J., and Burdeniuc, J., “Low- or Non-Emissive Amine Cata-
lysts for Low-Density, Open Cell SPF,” presented at the 2012 Center of Polyurethanes
Conference, September 20, 2012.

[14] Wood, R., “CPI Ventilation Research Project for Estimating Re-Entry Times for Trade
Workers Following Application of Three Generic Polyurethane Foam Formulations,”
presented at the Proceedings 2014 Polyurethanes Technical Conference, Fairfax, VA,
October 1, 2014.

[15] Hippelein, M., “Background Concentrations of Individual and Total VOCs in Residential
Indoor Air of Scheswig-Holstein, Germany,” J. Environ. Monitor., Vol. 6, No. 9, 2004,
pp. 745–752.

[16] Hodgson, A. and Levin, H., Volatile Organic Compounds in Indoor Air: A Review of Con-
centrations Measured in North America Since 1990, Lawrence Berkeley Laboratories,
Berkeley, CA, 2003.

[17] Knowles, M., “Troubleshooting Spray-Foam Insulation,” JLC, September, 2010, pp. 161–165.

[18] American Chemistry Council, “Spray Polyurethane Foam: Guidance on Sampling Tech-
niques for the Inspection of Installed SPF,” https://polyurethane.americanchemistry.
com/Spray-Foam-Coalition/Guidance-on-Sampling-Techniques-for-the-Inspection-of-
Installed-Spray-Polyurethane-Foam.pdf (accessed December 16, 2016).
148 DEVELOPING CONSENSUS STANDARDS FOR MEASURING CHEMICAL EMISSIONS

STP 1589, 2017 / available online at www.astm.org / doi: 10.1520/STP158920150008

Richard Wood1

Estimating Re-Entry Times for

Trade Workers Following the
Application of Three Generic
Spray Polyurethane Foam
Formulations
Citation
Wood, R., “Estimating Re-Entry Times for Trade Workers Following the Application of Three
Generic Spray Polyurethane Foam Formulations,” Developing Consensus Standards for
Measuring Chemical Emissions from Spray Polyurethane Foam (SPF) Insulation, ASTM
STP1589, J. Sebroski and M. Mason, Eds., ASTM International, West Conshohocken, PA, 2017,
pp. 148–166, http://dx.doi.org/10.1520/STP1589201500082

ABSTRACT
There is a great need for emissions information that can be used to estimate
when it is safe for trade workers to re-enter a work area in which spray
polyurethane foam (SPF) was recently applied to surfaces during retrofit and
new construction. The Center for the Polyurethanes Industry (CPI) Ventilation
Research Task Force conducted experiments to determine the decay rate of
chemical vapor after SPF application. The purpose of the study was to estimate
the time required to restrict unprotected trade workers, such as plumbers,
electricians, dry wall installers, etc., from the work area to minimize their
exposure. Emissions from CPI generic high-pressure medium-density, high-
pressure low-density, and low-pressure two-component kit formulations were
evaluated in a ventilated room operating at a rate of 10 air changes per hour
(ACH) for 3 h and lowered to 1 ACH for 9 h to evaluate decay rates for a total of
12 h. Chemical ingredients selected for evaluation represent those typically
present in SPF formulations. Substances evaluated included methylene diphenyl
diisocyanate, amine catalysts, blowing agents, and flame retardants. This paper

Manuscript received January 29, 2015; accepted for publication December 14, 2015.
1
Air Products and Chemicals, Inc., Global Industrial Hygiene Services, 7201 Hamilton Blvd., Allentown, PA 18195
2
ASTM Symposium on Developing Consensus Standards for Measuring Chemical Emissions from Spray
Polyurethane Foam (SPF) Insulation on April 30–May 1, 2015 in Anaheim, CA.

Copyright V
C 2017 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959.
 WOOD, DOI 10.1520/STP158920150008 149

discusses the experimental protocol, reports results of the air sampling study,
and provides recommendations.

Keywords
SPF application, re-entry time, MDI, amine catalyst, blowing agent, flame retardant

Introduction
Determining trade worker re-entry time after applying spray polyurethane foam
(SPF) in residential and commercial buildings has traditionally been estimated by
SPF manufacturers and contractors. Decisions related to the time required for air-
borne concentrations of SPF chemical components to fall to within acceptable levels
have frequently been based on qualitative information such as comfort and odor.
According to the Center for the Polyurethanes Industry (CPI), other factors may
include building volume and mechanical and natural ventilation. In addition to the
release of airborne SPF chemicals during spray application, certain components can
be liberated from some newly installed SPF products for a short period of time after
installation. CPI recommends contacting the SPF supplier for guidance on ventila-
tion time and reoccupancy for the particular job and SPF in use. For an interior
application using two-component high-pressure SPF, some manufacturers recom-
mend 24 h before reoccupancy, and for an interior two-component, low-pressure
SPF kit application, some manufacturers recommend a 1-h reoccupancy time [1].
This research considers the evaluation and impact of SPF emissions on healthy
adults working an average 8- to 10-h workday. Emission data collected after SPF
application were used to determine chemical degradation times to estimate the time
required for workers to safely re-enter and work without respiratory protection.
The CPI re-entry research builds on the Phase 1 and 2 work completed during
the CPI Ventilation Research Project [2]. During Phase 1 of the ventilation research
study, three generic SPF formulations were developed to represent those currently
available in the SPF marketplace. The generic formulations (summarized in Table 1)
included a high-pressure low-density, high-pressure medium-density, and low-
pressure kit formulation. During Phase 2 of the study, generic formulations (summa-
rized in Table 2) were sprayed at three ventilation rates (10 air changes per hour
[ACH], 233 ACH, and 598 ACH) while the spray applicator was monitored for air-
borne SPF chemical components. Fixed area samples were also collected during and
30 min after application. The generic high-pressure medium-density formulation was
evaluated at 10.4 ACH, 233 ACH, and 598 ACH. All three generic formulations were
evaluated at the 10.4-ACH rate. Because worker exposure information was available
for the three generic formulations at 10.4 ACH, the CPI Ventilation Task Force se-
lected the 10.4-ACH ventilation rate as the starting point for the re-entry research.
The previous CPI ventilation research study and re-entry research summarized in
this paper may be combined to provide a complete view of the time required from
application to re-entry for each generic formulation at the 10-ACH ventilation rate.
150 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 1 Generic SPF formulations.

High-Pressure Low-Density (1=2 lb) High-Pressure Medium-Density (2 lb) Low-Pressure (Two-Component)

SPF Formulation SPF Formulation Kit Formulation

A-side
100 % Polymeric methylene 100 % pMDI 92.5 % pMDI
diphenyl diisocyanate (pMDI) Blowing agent 1,1,1,2-tetrafluoro-
ethane (HFC-134a) (7.5 %)
B-side
Polyether polyol (34 %) Aromatic polyester polyol Polyester polyol (23 %)
(36.39 %) Polyether polyol (23 %)
Aromatic amino polyether
polyol (33.61 %)
Nonylphenol ethoxylate
emulsifier (11.9 %)
Blowing agent Blowing agent Blowing agent
Water (20 %) 1,1,1,3,3-Pentafluoropropane HFC-134a (17 %)
(HFC-245fa) (6.97 %)
Water (2.53 %)
Fire retardant Fire retardant Fire retardant
Tris-(1-chloro-2-propyl) TCPP (15.91 %) TCPP (30 %)
phosphate (TCPP) (25.2 %)
Silicone surfactant (1.0 %) Silicone surfactant (1.0 %) Silicone surfactant (2 %)
Catalyst Catalyst Catalyst
Bis (2-dimethylaminoethyl) BDMAEE (0.7 %) Pentamethyldiethylene
ether (BDMAEE) (0.9 %) triamine (5 %)
Tetramethyliminobispropylamine Bis (dimethylaminopropyl) Potassium 2-ethylhexanoate/2,20 -
(TMIBPA) (3.0 %) methylamine (DAPA) (2.59 %) oxybisethanol
N,N,N-Trimethylaminoethyletha- TMAEEA (0.3 %)
nolamine (TMAEEA) (4.0 %)

Knowledge and experience related to the selection and use of mechanical venti-
lation to control emissions during and after SPF application varies widely among
contractors. Consequently, contractors often rely on government-required personal
protective equipment (PPE) for spray applicators and helpers while mechanical
ventilation may be used to supplement power PPE to reduce emissions in sur-
rounding areas of the building. Due to variations in the use of effective mechanical
ventilation during SPF application, this study was designed to represent a feasible
worst-case scenario for residential or commercial work. It is feasible that most con-
tractors will use barriers and portable exhaust ventilation as recommended by the
U.S. Environmental Protection Agency (EPA) [3] and the SPF industry; however, it
is plausible that some will remove the portable ventilation at the end of the day/
project so that it can be used at another location the following day. This would be
especially true for small retrofit projects. The protocol selected for this study was
designed to simulate this scenario. Exhaust ventilation is supplied to the work area
 WOOD, DOI 10.1520/STP158920150008 151

TABLE 2 Air sampling and analytical methodology for select SPF constituents.

CAS # Analyte Analytical Method Flow Rate Sampling Media

101-68-8 2,4- and 4,4-MDI Modified OSHA 47 1.0 lpm 13-mm glass fiber filter
HPLC using a UV or coated with 1.0 mg 1-2PP
fluorescence detector extracted with 90/10
(v/v) ACN/DMSO
811-97-2 HFC-134a Modified OSHA 7 Diffusive sampler assay
(diffusive sampler) technology
460-73-1 HFC-245fa EPA Compendium 0.25 lpm Multisorbent tubes
Method T0-17
13674-84-5 TCPP Thermal
3033-62-3 BDMAEE desorption/GC/MS
6711-48-4 TMIBPA
2212-32-0 TMAEEA
3855-32-1 DAPA
3030-47-5 Pentamethyldiethylene
triamine

during application and is removed several hours afterward when the contractor
leaves the site and other trade workers arrive to provide services. The generic formu-
lations were sprayed for a 15-min period at 10 ACH. The ventilation continued to
operate at 10 ACH for 3 h after application and was reduced to 1 to 2 ACH for the
remaining 9 h to represent minimal passive ventilation from open doors or windows.
Emissions from the generic high-pressure low-density, high-pressure medium-
density SPF, and low-pressure, two-component kit formulations were evaluated in
a ventilated room operating at an air exchange rate of 10 ACH and 1 to 2 ACH.
Chemical ingredients selected for evaluation represent those typically present in
SPF formulations. Substances evaluated included the following:
• Methylene diphenyl diisocyanate (MDI)
• Amine catalysts
• Blowing agents other than water
• Flame retardants

Facilities, Equipment, and Application

Area air monitoring was conducted while generic formulations were applied to test
panels inside a ventilated spray room. The ventilated spray room was designed to
meet the following specifications:
1. The dimensions of the spray room were approximately 8 by 8 by 8 ft. Make-
up air was supplied to the room and was exhausted through a 4 by 8-ft filter
bank on the opposite wall.
2. The spray substrate consisted of 2 by 4-in. studs that were 7 ft in height,
spaced 16 in. apart, and provided two cavities lined with cardboard inserts.
152 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

Inserts were 16 in. wide, 6 in. deep, and 7 ft in height. The total size of the
spray substrate was approximately 3 by 7 ft, representing a typical wall cavity.
The SPF applicator sprayed the generic high-pressure formulations using a
Graco Fusion Air Purge 01 round-tip spray gun. The generic low-pressure kit
formulation was sprayed using manufacturer-supplied spray equipment. SPF was
applied under the following conditions:
1. Ambient air temperature of 75 to 80 F with 50 % relative humidity.
2. Inserts were sprayed using manufacturer-recommended pressure.
3. The system was heated to the manufacturer-recommended temperature.
4. Inserts were sprayed while maintaining a 12- to 24-in. distance from the sub-
strate being sprayed.
5. When SPF application was complete, inserts were removed and stacked in
the room. The cardboard inserts were then replaced and SPF application
continued.
6. A 3 by 7-ft area was sprayed in one pass at a nominal thickness of the following:
• 2 in. for high-pressure medium-density (2 lb) SPF
• 4 in. for high-pressure low-density (1=2 lb) SPF
• 2 in. for low-pressure kit SPF
7. The area of inserts sprayed, the average density of the sprayed foam, and the
approximate quantity of B-side sprayed are listed in Appendix A for each
generic SPF formulation.

Test Protocol
Air sampling for the chemical ingredients selected for evaluation was conducted by a
third-party contractor using the following applicable sampling and analytical protocols
(summarized in Table 2):
1. Two sampling sessions were conducted for each SPF formulation after applica-
tion to the spray substrate.
2. SPF formulations were applied for 15 min during each session.
3. The spray room exhaust ventilation rate was set at 10 ACH during and after
application. Actual measured ventilation rates varied slightly between air sam-
pling sessions.
4. After application, sprayed inserts remained in the ventilated spray room with
ventilation operating at 10 ACH.
5. Air sampling equipment was suspended near the center of the room, approxi-
mately 6 ft from the floor. After air sampling sessions, samples were submitted
to several qualified laboratories for analysis.
6. Air sampling was conducted according to the following schedule:
a. Hour 1: Air sampling was initiated 1 h after application.
b. Hour 2: At the completion of Hour 1 sampling, air samplers were
replaced with a second set of air samplers.
c. Hour 3: At the completion of Hour 2 sampling, ventilation was re-
duced from 10 ACH to a target rate of 1 ACH. The ventilation con-
tinued to operate at 1 ACH for 1 h. Air sampling was not conducted
during Hour 3 to allow ventilation to stabilize.
 WOOD, DOI 10.1520/STP158920150008 153

d. Hour 4: At Hour 4, a third set of air samplers was placed in the ven-
tilated room. The ventilation rate continued at 1 ACH.
e. Hour 8: At Hour 8, a fourth set of air samplers was placed in the
ventilated room. The ventilation rate continued at 1 ACH.
f. Hour 12: At Hour 12, a fifth set of air samplers was placed in the
ventilated room. Ventilation continued at 1 ACH.

Air Sampling Methodology

MDI
MDI samples were collected by drawing a known volume of air through a glass fiber
filter coated with 1.0 mg 1-(2-pyridyl) piperazine (1-2PP) contained within an open-
face cassette. Samples were then extracted with 90/10 (v/v) acetonitrile/dimethyl sulf-
oxide (ACN/DMSO) and analyzed by high-performance liquid chromatography
(HPLC) using an ultraviolet (UV) or fluorescence detector (OSHA 47) [4]. The air
sampling and analytical method was selected in part with the assumption that residual
MDI vapor may be present in the atmosphere 1 h after application, whereas SPF
aerosol would be minimal. Analytical sensitivity is 12 ng per sample (0.8 lg/m3).

1,1,1,2-TETRAFLUOROETHANE
1,1,1,2-Tetrafluoroethane (HFC-134a) samples were collected with diffusive air
samplers. The passive samples are small badges containing a solid sampling media,
such as charcoal, that can be clipped to an employee’s clothing for personal moni-
toring of halocarbons. Area monitoring can be performed using these samplers if
sufficient air flow is present in the workplace. Samples are analyzed by gas chroma-
tography (GC) according to a modified OSHA 7 method [4].

AMINE CATALYSTS AND 1,1,1,3,3-PENTAFLUOROPROPANE

Air samples were collected by drawing air through thermal desorption tubes prepared
by Markes International using calibrated air pumps. The air sampling tubes contained
several proprietary sorbents designed to collect key polyurethane volatile and semivola-
tile compound components, including amine catalysts, fire retardants, and blowing
agents. At the completion of the air monitoring sessions, samples were submitted to
the Bayer MaterialScience Industrial Hygiene Laboratory, where they were analyzed us-
ing thermal desorption-capillary GC mass spectrometry (GC/MS) in accordance with
the EPA’s Compendium Method T0-17 [5]. High-purity standards were prepared and
analyzed to identify sample components and determine instrument response factors.

Results of Air Sampling

GENERIC HIGH-PRESSURE MEDIUM-DENSITY FORMULATION
Samples were collected at a measured ventilation rate of 9.2 ACH at 1 and 2 h after
SPF application. Beginning 3 h after application, the ventilation rate was reduced to
154 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 1 Air sampling configurations.

a measured air exchange rate of 1.7 ACH. Air samples were again collected at 4, 8,
and 12 h after application at 1.7 ACH. The A-side MDI vapor was monitored dur-
ing Hours 1, 2, and 4 only because no detectable concentrations of MDI were
expected throughout the 12-h study period due to the reactivity of the isocyanate.
The B-side components were monitored at each target hour. All samples were col-
lected for 20 min with the exception of MDI. MDI samples were collected for
60 min to meet acceptable analytical detection limits. Sample times bracketed each
target hour; e.g., 20-min samples began 10 min before the hour and ended 10 min
after the hour. Fig. 1 illustrates the air sampling equipment configuration used
throughout the study.
Air samples were collected prior to spraying the generic formulations to mea-
sure potential background concentrations of the SPF components. Field blanks
were also submitted after each round of testing. Background and field blank sample
analysis indicated small amounts of SPF components in 2 of the 14 samples collect-
ed; however, the quantities were slightly above analytical detection limits.
The results of the analysis are listed in Appendix B and displayed in graphical
form in Figs. 2–4. There were no measureable airborne concentrations of 2,4- or
4,4-MDI in air samples collected 1, 2, and 4 h after application. MDI aerosol and
vapor have been detected during spray application [1] of the generic high-pressure
medium-density formulation; however, in this study, MDI concentrations were
 WOOD, DOI 10.1520/STP158920150008 155

FIG. 2 Amine catalyst—generic high-pressure medium-density formulation.

0.06

0.05
Occupaonal Exposure Limit BDMAEE = 0.05 ppm
Concentraon - (ppm)

0.04 Air Exchnge

Air Exchnge Rate
Rate Hours Hours 3–12 = 1.7ACH
0.03 0–3 = 9.2ACh

0.02

0.01

0
0 1 2 3 4 5 6 7 8 9 10 11 12
Hours Post Spray
DAPA 11/19/2013 BDMAEE 11/19/2013 TMAEEA 11/19/2013
DAPA 11/20/2013 BDMAEE 11/20/2013 TMAEEA 11/20/2013

below detection limits when sampled 1 h after application. The detection limits
ranged from <0.00014 to <0.00016 ppm, which is significantly below the 8-h time-
weighted average (TWA) threshold limit value (TLV) of 0.005 ppm assigned by the
American Conference of Governmental Industrial Hygienists (ACGIH) [6].

FIG. 3 TCPP—generic high-pressure medium-density formulation.

0.005

0.004 Air Exchnge Rate

Air Exchnge Rate
Hours 3-12 = 1.7ACH
Concentraon - (ppm)

Hours 0-3 = 9.2ACH

0.003

0.002

0.001

0
0 1 2 3 4 5 6 7 8 9 10 11 12
Hours Post Spray
TCPP 11/19/2013 TCPP 11/20/2013
156 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 4 HFC-245fa—generic high-pressure medium-density formulation.

1000

Occupaonal Exposure Limit HFC 245fa = 300 ppm

100
Concentraon - (ppm)

Air Exchnge Rate Air Exchnge Rate

Hours 0-3 = 9.2ACH Hours 3-12 = 1.7ACH

10

11/19/2013
11/20/2013
1

0.1
0 1 2 3 4 5 6 7 8 9 10 11 12
Hours Post Applicaon

The amine catalyst concentrations listed in Appendix B ranged from below

detection limits to 0.023 ppm. Fig. 2 displays the data in graphical form and includes
the occupational exposure limit (OEL) for BDMAEE as a point of reference.
BDMAEE is the only amine catalyst in the generic formulation that has been assigned
an OEL. The ACGIH has assigned BDMAEE a TLV of 0.05 ppm as an 8-h TWA
concentration. The highest concentration measured during the 2-day study was
approximately half the ACGIH-recommended value. This occurred within 1 h after
application. The concentration continued to decrease during the second hour and
increased slightly during the fourth hour after the ventilation had been lowered to 1.7
ACH. TMAEEA, a nonemissive catalyst, was not detected in any of the samples; how-
ever, it must be noted that the reported TMAEEA values are approximations because
the compound degrades due to the thermal desorption during sample analysis.
DAPA, an emissive catalyst, followed emission patterns similar to BDMAEE. Con-
centrations decreased from the first to the second hour with a ventilation rate of 9.2
ACH. At the fourth hour after SPF application, concentrations of the catalysts
BDMAEE and DAPA increased or remained the same at 1.7 ACH. Concentrations
then generally remained steady or decreased slightly from Hours 4 through 12.
TCPP concentrations ranged from <1 ppb to 3 ppb (Appendix B; Fig. 3). Con-
centrations were highest 1 h after application and dropped by half by the second
hour. At Hours 4 through 12, concentrations remained at approximately 1 ppb at
the reduced air exchange rate of 1.7 ACH. There are no published OELs for TCPP;
however, a chronic toxicity reference dose based on the lowest observed adverse
effect level (LOAEL) of 52 mg/kg/day has been in an European Union Risk Assess-
ment Report for TCPP (EU 2008) [7]. The LOAEL was based on a 13-week study
that resulted in increased liver weights in male rats. The authors assumed an 80 %
 WOOD, DOI 10.1520/STP158920150008 157

absorption by the oral dosing, which resulted in an absorbed dose of 42 mg/kg/day.

Only one animal species was tested; therefore, a protection factor of 10 is applied
for the potential variability in response between rats and humans. Intraspecies vari-
ability due to increased liver weights would be considered minimal; therefore, a fac-
tor of 5 to 10 could be applied. When divided by a maximum factor of 100 to
account for animal to human doses, the absorbed reference dose is 0.42 mg/kg/day.
The average TCPP concentration for the first 8 h after application was
0.0019 ppm (0.026 mg/m3). Assuming a 70-kg worker inhales 10 m3 during an 8-h
workday, the absorbed dose, assuming 100 % absorption by inhalation, equals
0.0037 mg/kg/day, a dose well below the 0.42 mg/kg/day absorbed reference dose.
Concentrations of the blowing agent HFC-245fa followed the same pattern as the
other B-side components (Appendix B; Fig. 4). Concentrations dropped significantly
from the first to second hour and rose again in the fourth hour after the air exchange
rate was reduced. The OEL of 300 ppm was established as a workplace environmental
exposure level (WEEL) by the American Industrial Hygiene Association (AIHA) [8].
The WEEL is an 8-h TWA concentration. The WEEL for HFC-245fa has been added
to Fig. 4 as a point of reference and indicates acceptable blowing agent emissions in
terms of worker exposure. Although detectable, HFC-245fa concentrations were 200
to 1,000 times lower than the WEEL for the 12-h period after application, even at the
reduced ventilation rate of 1.7 ACH.

Discussion
Concentrations of MDI, BDMAEE, and HFC-245fa were well within OELs begin-
ning 1 to 12 h after application. Although the measurements do not represent actual
worker exposure assessments, based on OELs the measured chemicals emitted
would not present a significant airborne risk to healthy adult trade workers entering
the area after application. All measured chemicals decreased from the first hour
after application, rose again as the air exchange rate was decreased from 9.2 to 1.7
ACH 4 h after application, and either remained steady or decreased until the com-
pletion of the monitoring 12 h after application. Using established OELs as safe 8-h
exposure concentrations for healthy adult trade workers as a benchmark—and con-
sidering 9.2 ACH may have partially controlled emissions during the first through
the third hours after application—trade workers could safely re-enter the work area
sprayed with the generic formulation after 4 h without the use of respiratory protec-
tion. This re-entry time is also based on the assumption that the passive ventilation
rate of 1.7 ACH would provide limited control of SPF emissions. It must also be
noted that manufacturers of high-pressure medium-density SPF formulations and
spray equipment may recommend shorter re-entry times based on internally gener-
ated exposure information that was not included in this study.

GENERIC HIGH-PRESSURE LOW-DENSITY FORMULATION

The study protocol previously described was followed for the generic high-pressure
low-density formulation. Chemicals monitored included MDI, three amine catalysts,
158 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

and the fire retardant TCPP. Blowing agent samples were not collected because water
was used in the low-density formulation. Air samples were collected at a measured ven-
tilation rate of 10 ACH at 1 and 2 h after SPF application. Beginning 3 h after applica-
tion, the ventilation rate was reduced to a measured ventilation rate of 2.1 ACH. Air
samples were again collected at 4, 8, and 12 h after application at 2.1 ACH. A-side MDI
vapor was monitored during Hours 1, 2, and 4. B-side components were monitored at
each target hour. All samples were collected for 20 min with the exception of MDI sam-
ples, which were collected for 60 min. Sample times bracketed each target hour.
Air samples were collected prior to spraying the generic formulation to measure
potential background concentrations of the SPF components. Field blanks were also
submitted after each round of testing. Background and field blank sample analysis
indicated a trace quantity of BDMAEE in one sample.
The results of the analysis are listed in Appendix B and displayed in graphical
form in Fig. 5 and Fig. 6. There were no measureable airborne concentrations of 2,4-
or 4,4-MDI in samples collected 1, 2 and 4 h after application. MDI concentrations
were below analytical detection limits, ranging from <0.00014 to <0.00016 ppm,
significantly below the ACGIH TLV TWA of 0.005 ppm for 4,4-MDI.
One thermal desorption tube sample collected at Hour 4 on the first day of the
study could not be analyzed due to an analytical instrument malfunction; therefore,
BDMAEE and TCPP Hour 4 concentrations have been estimated and displayed as
a dashed line connecting the Hour 2 results to the Hour 8 results in Fig. 5 and Fig. 6.
The sample is listed as “Void” in Appendix B.
BDMAEE was detected in all thermal desorption tube samples, whereas
TMIBPA and TMAEEA were not detected in any air sample. As demonstrated in

FIG. 5 Amine catalyst—generic high-pressure low-density formulation.

0.1
0.09
0.08
0.07
Concentraon - (ppm)

0.06 Hour 4 sample could not be analyzed.

Dashed line esmates BDMAEE
0.05
Occupaonal Exposure Limit BDMAEE = 0.05 ppm
0.04
0.03
Air Exchnge Rate Air Exchnge Rate
0.02 Hours 0-3 = 10 ACH Hours 3-12 = 2.1 ACH

0.01
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Hours Post Spray
TMAEEA 02/25/14 TMIBPA 02/25/14 BDMAEE 02/26/14
BDMAEE 02/25/14 TMAEEA 02/26/14 TMIBPA 02/26/14
 WOOD, DOI 10.1520/STP158920150008 159

FIG. 6 TCPP—generic high-pressure low-density formulation.

0.005
TCPP 02/25/2014
TCPP 02/26/2014
0.004
Air Exchnge Rate Air Exchnge Rate
Concentraon - (ppm)

Hours 0-3 = 10 ACH Hours 3-12 = 2.1 ACH

0.003

0.002

0.001
Hour 4 sample could not be analyzed.
Dashed line esmates TCPP
concentraons between Hour 2 and Hour 8
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Hours Post Spray

Appendix B and Fig. 5, BDMAEE concentrations were generally above the ACGIH
TLV TWA of 0.05 ppm. Only one sample collected during the second hour after
application resulted in a BDMAEE concentration less than the TLV TWA. BDMAEE
concentrations decreased from Hour 1 to Hour 2 after application but rose again at
Hour 4 after the ventilation rate was reduced to 2.1 ACH. Concentrations continued
to rise at Hours 8 and 12 after application on the first day of sampling but decreased
slightly at Hour 12 on the second day.
The quantity of BDMAEE in the high-pressure medium-density formulation
and the low-density formulation are similar: 0.7 % and 0.9 %, respectively; there-
fore, the difference in BDMAEE emissions between the two formulations is likely
due to the open-cell nature of the low-density foam. BDMAEE is an emissive cata-
lyst that is not bound to the cured formulation and will readily emit from an open-
cell foam over time, whereas the closed-cell formulation will encapsulate the cata-
lyst, resulting in reduced emissions.
TCPP emissions varied between Days 1 and 2 of the study; however, concentra-
tions approached the analytical detection limit. On the second day, TCPP concen-
trations rose slightly from the first to second hour and remained at approximately
2 ppb for Hours 4 through 12. Hours 8 through 12 on the first day of testing resulted
in the same TCPP concentration of 2 ppb.

Discussion
Concentrations of 2,2- and 2,4-MDI and the amine catalysts TMEEA and TMIBPA
were below detection limits. TCPP was present at very low part per billion concen-
trations that were slightly above analytical detection limits. However, BDMAEE
concentrations generally exceeded the ACGIH TLV TWA during the 12-h study
160 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

period. Although the catalyst measurements do not represent actual worker expo-
sure, trade workers entering the area where the generic high-pressure open-cell for-
mulation had been sprayed at the reported ventilation rates would require
respiratory protection in addition to any other required PPE. Additional BDMAEE
measurements at 18 or 24 h or greater would be required to establish a re-entry
time during which trade workers would not require respiratory protection in addi-
tion to other required PPE.

GENERIC LOW-PRESSURE TWO-COMPONENT KIT FORMULATION

Test protocol followed for the generic high-pressure medium-density and low-
density formulations was also followed for the low-pressure kit formulation. Chem-
icals measured included MDI, the amine catalyst N,N,N0 ,N00 ,N00 -pentamethyldiethy-
lenetriamine (PMDTA), the fire retardant TCPP, and the blowing agent HFC-134a.
Air samples were collected at a measured ventilation rate of 9.2 ACH 1 and 2 h after
SPF application. Beginning 3 h after application, the ventilation rate was reduced to
a measured air exchange rate of 2.1 ACH, and air samples were again collected at
4, 8, and 12 h after application. A-side MDI vapor was monitored during Hours 1, 2,
and 4. B-side components were monitored at each target hour. Catalyst and fire
retardant samples were collected for 20 min, whereas MDI and HFC-134a were col-
lected for 60 min. Sample times bracketed each target hour.
Air samples were collected prior to spraying the generic kit formulation to mea-
sure potential background concentrations of the SPF components. Field blanks
were also submitted after each round of testing. All chemical concentrations in
background samples and field blanks were below analytical detection limits.
The results of the analysis are listed in Appendix B and displayed in graphical
form in Fig. 7 and Fig. 8. There were no measureable airborne concentrations of
either 2.4- or 4,4-MDI in samples collected 1, 2, and 4 h after application. The MDI
detection limits ranged from 0.00014 to 0.00016 ppm. The blowing agent HFC-
134a was also not detected in any sample collected during the 12-h study period.
The detection limits for the blowing agent were 3.91 to 5.91 ppm, which is far below
the AIHA WEEL of 1,000 ppm.
The pentamethyldiethylene triamine (PMDETA) concentrations listed in
Appendix B and Fig. 7 indicate the amine catalyst was detected in all samples with
the exception of the sample collected at Day 1, Hour 2. Catalyst concentrations
were low on each day of the study, ranging from undetectable to 0.018 ppm.
PMDETA concentrations either decreased slightly or remained the same between
Hour 1 and Hour 2 and then remained the same or increased slightly as the ventila-
tion was decreased at Hour 4. On Day 1, Hour 12, the PMDETA concentration
rose from 0.002 to 0.018 ppm. Although there is no clear reason for the increase, it
is likely due to emissions near the analytical detection limit, where a slight increase
in amine concentration may appear significant.
The fire retardant TCPP varied by several parts per billion between the first
and second day of air sampling. During Day 1 of the study, TCPP concentrations
 WOOD, DOI 10.1520/STP158920150008 161

FIG. 7 Amine catalyst—generic low-pressure two-component kit formulation.

0.05
PMDETA 05/27/2014
PMDETA 05/28/2014
0.04
Concentraon - (ppm)

Air Exchnge Rate Air Exchnge Rate

Hours 0-3 = 9.8 ACH Hours 3-12 = 2.1 ACH
0.03

0.02

0.01

0
0 1 2 3 4 5 6 7 8 9 10 11 12
Hours Post Spray

followed the previously observed pattern of concentrations decreasing from Hour 1 to

Hour 2 and then a gradual increase after Hour 4. TCPP concentrations on the second
day remained steady at approximately 45 ppb and dropped to 15 ppb at Hour 12.
Similar to the amine catalyst results, concentrations were very low, with the variation
attributable to slight changes in environmental conditions or analytical sensitivity.

Discussion
Concentrations of 2,2- and 2,4-MDI and the blowing agent HFC-134a were below
detection limits. TCPP and the amine catalyst PMDETA were present at very low

FIG. 8 TCPP—generic low-pressure two-component kit formulation.

0.005
TCPP 05/27/2014
TCPP 05/28/2014
0.004
Air Exchnge Rate Air Exchnge Rate
Concentraon - (ppm)

Hours 0-3 = 9.8 ACH Hours 3-12 = 2.1 ACH

0.003

0.002

0.001

0
0 1 2 3 4 5 6 7 8 9 10 11 12
Hours Post Spray
162 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

parts per billion concentrations, slightly above analytical detection limits. Consider-
ing the limited information generated by this study, a re-entry time of 4 h or more
after SPF application is recommended for workers entering the sprayed areas with-
out respiratory protection. The recommended re-entry time is also based on the
assumption that the passive ventilation rate of 2.1 ACH provided limited control of
SPF emissions. It must also be noted that manufacturers of two-component kit for-
mulations and spray equipment may recommend different re-entry times based on
internally generated exposure information not included in this study.

Re-Entry Study Summary

2,4-MDI OR 4,4-MDI
There were no measureable airborne concentrations of either 2,4- or 4,4-MDI in
the samples collected 1, 2, and 4 h after application. MDI has been shown to be pre-
sent during spray application; however, MDI vapor could not be detected in any
sample 1 h after application. The detection limits ranged from 0.00014 to
0.00016 ppm, which is significantly below the TLV of 0.005 ppm for 4,4-MDI.

AMINE CATALYSTS
Measurable quantities of amine catalysts were detected in each generic formulation;
however, only one of the five catalysts present in the three formulations, BDMAEE,
has been assigned an OEL. The ACGI has assigned BDMAEE a TLV TWA of
0.05 ppm. In a status report regarding the five amine catalysts in the CPI generic
formulations, the Consumer Product Safety Commission (CPSC) [9] stated that
four “had very limited toxicological or exposure data to review; therefore, data rele-
vant to amine catalyst toxicity is based on the chemical with the most information,
bis (2-dimethylaminoethyl) ether.” Using CPSC’s guide to interpretation, with the
exception of BDMAEE, all catalyst emissions from the high-pressure medium-den-
sity formulation and the low-pressure kit formulation would be considered within
OELs beginning 1 h after application given the ventilation rates used for this study.
BDMAEE emissions measured after application of the high-pressure low-density
open-cell formulation at the ventilation rates used for this study were greater than
the 0.05-ppm TLV TWA and would not be considered safe for unprotected trade
workers after application.

TCPP
The generic formulations contain significant quantities of TCPP ranging from 16 to
30 % of the total B-side formulation. Although the quantity is significant, most fire
retardants remain in the sprayed foam once the material is applied. After applica-
tion, as the foam temperature decreases, low parts per billion concentrations of
TCPP were emitted. Although there are no established exposure limits for TCPP,
the measured concentrations 1 to 12 h after application were near analytical
 WOOD, DOI 10.1520/STP158920150008 163

detection limits and would be well below the LOAEL or 52 mg/kg/day and the EU
absorbed reference dose of 42 mg/kg/day at the ventilation rates used for this study.

HFC-245FA AND HFC-134A

Measurable quantities of HFC-245fa were detected 1 to 12 h after application of the
high-pressure medium-density formulation. All concentrations were <1 ppm, which is
well within the OEL of 300 ppm set by the AIHA WEEL Committee. Concentrations of
HFC-134a used in the low-pressure kit formulation were below analytical detection
limits. The WEEL Committee has set an OEL of 1,000 ppm for HFC-134a. Based on the
findings, blowing agent concentrations emitted from the generic high-pressure medi-
um-density and low-pressure kit formulations would likely not pose a risk to the health
of unprotected trade workers at the ventilation rates used for this study.

Conclusions
There are several conclusions that can be drawn from the data generated. In the
absence of properly designed ventilation, SPF emissions can be expected to vary
between formulations. It is also evident, assuming the CPI generic formulations rep-
resent typical SPF formulations in commercial use today, that MDI vapor does not
present a significant airborne health risk to unprotected trade workers beginning 1 h
after application. Certain emissive catalysts, such as BDMAEE, are not bound to the
sprayed open-cell formulation and will continue to be released at concentrations
above the OEL 12 h after application if ventilation is not sufficient to control emis-
sions. BDMAEE in the closed-cell formulation did not exceed the OEL at any time
during the 12-h period. Nonemissive catalysts, such as TMAEEA and TMIBPA, pre-
sent in the same open-cell formulation, were bound to the spray foam, and were be-
low the detection limits throughout the test period. Given the environmental
conditions and the protocol used in this study, other components such as TCPP and
the blowing agents HFC-245fa and HFC-134a would not pose a significant health
risk for unprotected workers at the ventilation rates used for this study.
Based on the results of this study, it is apparent emissions of SPF chemicals will
vary depending on the type of foam (open- versus closed-cell) and the specific
chemicals comprising the B-side of the formulation. It is therefore essential that
emissions from commercial SPF formulations, in particular open-cell formulations,
be controlled through the use of mechanical ventilation during application and for
12 h after application. Control effectiveness should be evaluated though industrial
hygiene air monitoring. The use of respiratory protection must then be used to sup-
plement ventilation where worker exposure to SPF chemicals cannot be controlled
by mechanical ventilation. In those instances in which SPF producers desire
reduced re-entry times for commercial products, those formulations should be eval-
uated in the laboratory using ventilation rates similar to those used in this study.
Laboratory studies should then be validated in the field with contractor-provided
mechanical ventilation.
164 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

Appendix A

Spray parameters.

Quantity of Quantity of
Average Number of A-side B-side
Density Inserts Sprayed Sprayed Total Spray
Generic formulation Date (lb/ft3) Sprayed (kg) (kg) Time (min)

High-pressure 11/19/2013 1.94 12 10 10 15

medium-density
High-pressure 11/20/2013 2.00 12 10 10 15
medium-density
High-pressure 2/25/2014 0.62 14 10 10 15
low-density
High-pressure 2/26/2014 0.54 14 10 10 15
low-density
Low-pressure 5/27/2014 2.21 6 12.4 12.8 15.75
two-component kit
Low-pressure 5/28/2014 2.21 6 12.3 11.7 14.5
two-component kit

Appendix B

Generic high-pressure medium-density formulation.

Concentration (ppm)
Exposure
Date Analyte Hour 1 Hour 2 Hour 4 Hour 8 Hour 12 Limit

11/19/2013 HFC-245fa 1.3 0.38 0.80 0.54 0.28 300

11/20/2013 1.0 0.26 0.86 0.63 0.38
11/19/2013 TMAEEA <0.006 <0.006 <0.006 <0.006 <0.006 N/A
11/20/2013 <0.26 <0.27 <0.25 <0.26 <0.25
11/19/2013 BDMAEE 0.023 0.0038 0.0043 0.0029 0.0027 0.05
11/20/2013 0.021 0.0065 0.0073 0.0043 0.044
11/19/2013 DAPA 0.019 0.0053 0.010 0.0064 0.0049 N/A
11/20/2013 0.019 0.0089 0.012 0.0100 0.0096
11/19/2013 TCPP 0.0028 0.0011 0.00085 0.0010 0.00087 N/A
11/20/2013 0.0029 0.0015 0.0013 0.0013 0.0011
11/19/2013 2,4-MDI <0.00016 <0.00015 <0.00015 N/A N/A N/A
11/20/2013 <0.00014 <0.00015 <0.00015 N/A N/A
11/19/2013 4,4-MDI <0.00014 <0.00015 <0.00015 N/A N/A 0.005
11/20/2013 <0.00014 <0.00015 <0.00015 N/A N/A
 WOOD, DOI 10.1520/STP158920150008 165

Generic high-pressure low-density formulation.

Concentration (ppm)
Exposure
Date Analyte Hour 1 Hour 2 Hour 4 Hour 8 Hour 12 Limit

2/25/2014 TMAEEA <0.015 <0.016 Void <0.015 <0.016 N/A

2/26/2014 <0.016 <0.016 <0.015 <0.016 <0.016
2/25/2014 BDMAEE 0.063 0.032 Void 0.076 0.081 0.05
2/26/2014 0.065 0.055 0.073 0.094 0.081
2/25/2014 TMIBPA <0.016 <0.016 Void <0.015 <0.017
2/26/2014 <0.17 <0.017 <0.016 <0.017 0.017
2/25/2014 TCPP 0.0015 <0.0011 Void 0.0021 0.0021
2/26/2014 0.0016 0.0022 0.0018 0.0024 0.0020
2/25/2014 2,4-MDI <0.00016 <0.00015 Void N/A N/A N/A
2/26/2014 <0.00015 <0.00014 <0.00014 N/A N/A
2/25/2014 4,4-MDI <0.00016 <0.00015 Void N/A N/A 0.005
2/26/2014 <0.00015 <0.00014 <0.00014 N/A N/A

Generic low-pressure (two-component) kit formulation.

Concentration (ppm)
Exposure
Date Analyte Hour 1 Hour 2 Hour 4 Hour 8 Hour 12 Limit

5/27/2014 HFC-134a <5.87 <5.87 <3.91 <3.91 <3.91 1,000

5/28/2014 <5.87 <5.87 <3.91 <3.91 <3.91
5/27/2014 PMDETA 0.007 <0.002 0.003 0.002 0.018 N/A
5/28/2014 0.011 0.006 0.009 0.011 0.009
5/27/2014 TCPP 0.0008 0.0006 0.0009 0.0013 0.0017 N/A
5/28/2014 0.0011 0.0016 0.0025 0.0027 0.0015
5/27/2014 2,4-MDI <0.00012 <0.00015 <0.00015 N/A N/A N/A
5/28/2014 <0.00015 <0.00015 <0.00015 N/A N/A
5/27/2014 4,4-MDI <0.00012 <0.00015 <0.00015 N/A N/A 0.005
5/28/2014 <0.00015 <0.00015 <0.00015 N/A N/A

References

[1] Center for the Polyurethanes Industry, “Spray Foam Health and Safety,” https://www.epa.
gov/saferchoice/ventilation-guidance-spray-polyurethane-foam-application (accessed
November 22, 2016).

[2] Wood, R., “CPI Ventilation Research Project Update,” presented at the CPI Polyur-
ethanes 2013 Polyurethanes Technical Conference, Phoenix, AZ, September 23–25, 2013.
166 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

[3] U.S. EPA, “Ventilation Guidance for Spray Polyurethane Foam Application,” https://
www.epa.gov/saferchoice/ventilation-guidance-spray-polyurethane-foam-application
(accessed November 22, 2016).

[4] Occupational Safety and Health Administration, “Sampling and Analytical Methods,”
https://www.osha.gov/dts/sltc/methods (accessed November 22, 2016).

[5] U.S. EPA, Compendium of Methods for the Determination of Toxic Organic Compounds in
Air, Second Ed., EPA/625/12.96/0106, Compendium Method TO-17, Determination of Vola-
tile Organic Compounds in Ambient Air Using Active Sampling on to Sorbent Tubes, 1999,
Cincinnati, OH, https://www3.epa.gov/ttnamti1/files/ambient/airtox/tocomp99.pdf

[6] American Conference of Governmental Industrial Hygienists, “2014 Threshold Limit

Values for Chemical Substances and Physical Agents and Biological Exposure Indices.”

[7] European Chemicals Agency, European Union Risk Assessment Report, Tris (2-Chloro-1-
Methylethyl) Phosphate (TCPP), CAS No: 13674-84-5, EINECS No: 237-158-7, May 2008.

[8] American Industrial Hygiene Association, Emergency Response Planning Guidelines

(ERPG) and Workplace Environmental Exposure Level (WEEL) Handbook, 2014.

[9] Biggs, M. B., “Review of Five Amine Catalysts in Spray Polyurethane Foam,” (CPSC-
D07-006), Status Report, Consumer Product Safety Commission, September 19, 2012.
 DEVELOPING CONSENSUS STANDARDS FOR MEASURING CHEMICAL EMISSIONS 167

STP 1589, 2017 / available online at www.astm.org / doi: 10.1520/STP158920150036

Shen Tian,1 John Sebroski,2 and Scott Ecoff1

Predicting TCPP Emissions and

Airborne Concentrations from
Spray Polyurethane Foam Using
USEPA i-SVOC Software:
Parameter Estimation and Result
Interpretation
Citation
Tian, S., Sebroski, J., and Ecoff, S., “Predicting TCPP Emissions and Airborne Concentrations
from Spray Polyurethane Foam Using USEPA i-SVOC Software: Parameter Estimation and
Result Interpretation,” Developing Consensus Standards for Measuring Chemical Emissions from
Spray Polyurethane Foam (SPF) Insulation, ASTM STP1589, J. Sebroski and M. Mason, Eds.,
ASTM International, West Conshohocken, PA, 2017, pp. 167–198, http://dx.doi.org/10.1520/
STP1589201500363

ABSTRACT
Spray polyurethane foam insulation is a widely used insulation material in both
new construction and building renovations. To better understand its impact on
indoor air quality in addition to establishing consensus standards for measuring
chemical emissions experimentally, a recently developed mathematical model
and software tool by the U.S. Environmental Protection Agency, i-SVOC, was
applied to estimate a semivolatile organic compound (SVOC) flame retardant
(tris[2-chloro-1-methylethyl] phosphate [TCPP]) emitted from spray foam and
its fate and transport in a modeled indoor environment. With limited literature
data, we first estimated the key parameters required by i-SVOC software,
including the solid/air partition coefficient (kma), solid-phase diffusion coefficient
(Ds), and gas-phase mass transfer coefficient (hg) of TCPP, and then included
indoor sinks such as dry wall and flooring as well as the interaction between

Manuscript received May 9, 2015; accepted for publication July 16, 2015.
1
Covestro, LLC, Product Safety and Regulatory Affairs, 1 Covestro Cir., Pittsburgh, PA, 15205
2
Covestro, LLC, Environmental Analytical Lab, 1 Covestro Cir., Pittsburgh, PA, 15205
3
ASTM Symposium on Developing Consensus Standards for Measuring Chemical Emissions from Spray
Polyurethane Foam (SPF) Insulation on April 30–May 1, 2015 in Anaheim, CA.

Copyright V
C 2017 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959.
168 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TCPP and airborne particles. Finally, i-SVOC was used to calculate the airborne
TCPP concentration and source emissions/sink sorption factor using a
hypothetical room condition. Additional simulations were conducted using the
microchamber conditions and compared to the experimental results. In addition,
a sensitivity analysis was performed to study the impact of key parameters on
modeling results. Comparing between modeling and experimental results, we
found that at 23 C the emissions factors derived from the microchamber study
were very close to the high-end modeling results, whereas at higher
temperatures, i-SVOC predicted lower emissions factors with current inputs. In
the sensitivity analysis, kma was found to be the most important parameter to
explain such differences between modeling and experimental results. The hg
contributed to the differences as well with a higher impact within the first few
hours. In addition, it is important to include indoor SVOC sinks, which are
ubiquitous in indoor environments. Although i-SVOC is a promising tool for
studying SVOC emissions from building materials, experimental measurements
are essential for validating input parameter estimation and modeling results.

Keywords
semivolatile organic compounds, i-SVOC, emission, sink, airborne particles,
flame retardant, parameter estimate

Introduction
Spray polyurethane foam (SPF) is an effective insulation and air sealant material
commonly used in existing building renovation and new construction [1]. Because
of its high performance in promoting energy efficiency, most recent data show a
29 % increase in sales from the first half of 2010 to the first half of 2012 for the overall
spray foam sector [2]. Indoor volatile organic compounds (VOCs) and semivolatile
organic compounds (SVOCs) emitted from building materials have been widely
studied in the past several decades [3]. Recent research has focused on characteriz-
ing SVOC emissions from products with potential indoor exposure [4–8]. Unlike
other building materials such as flooring, concrete, and particle board, research on
SVOC emissions from SPF was rare. The U.S. Environmental Protection Agency
(EPA) states that “the potential off-gassing of volatile chemicals from SPF is not
fully understood and is an area where more research is needed” [1]. To address the
public concerns about chemical emissions from SPF and promote product steward-
ship, value-chain stakeholders such as industry associations and raw material
suppliers partnered with federal agencies such as the EPA and National Institute
of Standards and Technology. From this partnership, research was conducted to
develop consensus standards for measuring SPF emissions. In addition to experimen-
tal measurements, numerous mathematical models and software tools were developed
to study the physical mechanism of chemical emissions from building materials and
to estimate key parameters such as the solid/air partition coefficient (kma), solid-phase
diffusion coefficient (Ds), and gas-phase mass transfer coefficient (hg) [4,9–12].
 TIAN ET AL., DOI: 10.1520/STP158920150036 169

Using models to predict chemical emissions has its advantages because modeling
results are not limited to the number of chemicals, measurement time, and equipment/
method detection limit. Modeling results are often calculated as a time series instead
of point estimates so that statistical analyses such as Monte Carlo can be performed.
In addition, existing models are able to include indoor sinks and airborne particles
(APs), which could have significant impacts on indoor SVOC concentrations. When
the key parameters are available or can be estimated, the time and cost of using models
to predict chemical emissions are minimal. The objective of this work was to estimate
SVOC emissions from SPF and compare the modeling results with experimental stud-
ies. Tris(2-chloro-1-methylethyl) phosphate (TCPP) was chosen as the test SVOC
because it was observed as one of the most common organophosphate flame retardants
in polyurethane (PU) foam applications [13]. To achieve this objective, the required
input parameters in i-SVOC were derived from existing methods, and their relative
impacts to the modeling results were studied by sensitivity analysis.

Literature Review on Existing VOC and SVOC

Models
Researchers first started to develop mathematical models to predict VOC emissions.
In the past two decades, more than 20 mass transfer models have been published
[7]. One of the pioneer models was developed by Little, Hodgson, and Gadgil [14].
This model marked a significant improvement over previous mass transfer models
[15–17] by including the internal diffusion in solid materials. However, an instanta-
neous equilibrium was assumed between the material surface and ambient air.
Consequently, the external convective or gas-phase mass transfer of chemicals was
ignored. Thus, this model yielded a nonzero airborne concentration at time zero,
which may not represent the reality. This problem was soon resolved by including
the gas-phase mass transfer coefficient (hg) [11,18,19]. More recently, the state-
space (SS) method, often used in the heat transfer field, has been applied to estimate
VOC emissions from building materials [20]. The SS method differs from other
models because it divides a slab of material into a finite number of layers with the
same thickness, and the VOC concentration within each layer is assumed to be uni-
form during the entire emissions process. Mass transfer equations have been devel-
oped in the SS method to characterize chemical movements in the diffusional
source and room air. At material surface, mass transfer between room air and the
surface layer is characterized by hg, kma, and the fugacity difference at the interface.
Within the source, mass transfer is controlled by the fugacity difference between
two adjacent layers (can be different materials) and the solid-phase mass transfer
coefficient (hm), which is a function of Ds and the layer thickness. Fig. 1 illustrates
the SS method graphically. An advantage of the SS method is that it uses a system
of first-order ordinary differential equations (ODEs) to describe the mass balance
of the studied chemical in room air, the exposed surface, and inner material layers.
These first-order ODEs can be solved more easily to calculate the emission rates
170 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 1 Graphical illustration of the SS method. A, solid material surface area (m2);
Ca, indoor/chamber SVOC air concentration (lg/m3); Cmi, chemical
concentration in the ith layer (lg/m3); Cm1, concentration in the surface layer
(lg/m3); Ds, solid-phase diffusion coefficient (m2/h); hg, gas-phase mass transfer
Ds
coefficient (m/h); hm ¼ Dx , solid-phase mass transfer coefficient (m/h); kma,
material/air partition coefficient (unitless); Q, indoor/chamber air flow rate (m3/h);
Rij, mass transfer rate between layers (lg/h); R1a, mass transfer rate from surface
layer to room air (lg/h); V, indoor chamber volume (m3); and Dx, thickness of
each layer.

and airborne concentrations than partial differential equations. Moreover, the SS

method is capable of including multiple sources, sinks, contaminant barriers, or
modeling more realistic scenarios when the initial airborne SVOC concentration is
not zero [3]. Using existing models, researchers found that VOC emissions from
building materials is primarily “internally controlled,” which means the internal dif-
fusion (measured by Ds) is more important than the external convective mass trans-
fer resistance (measured by kma and hg) [3,5,21].
Developing SVOC models is more recent and challenging than VOC models
because of the chemical and physical characteristics of SVOCs [5,6,21–24]. Unlike
VOCs, SVOCs have higher kma values, which lead to lower airborne concentrations,
and the emissions process is primarily externally controlled instead of internally
controlled [5,25]. In addition, chamber surfaces have a strong sorption effect on
SVOCs, which makes emission testing more difficult. Compared to VOCs,
researchers also found SVOCs tend to stick to and migrate into indoor APs and set-
tled dusts, which could significantly increase indirect exposure to those chemicals
[21,26,27]. Despite the challenges, the same modeling scheme can be applied to
 TIAN ET AL., DOI: 10.1520/STP158920150036 171

both VOCs and SVOCs. Xu and Little [5] developed a conceptual model to study
di-2-ethylhexyl phthalate emissions from vinyl flooring and validated it with experi-
mental data. Their model was capable of including SVOC sorption in indoor sinks
and APs, which set a foundation for future model development. Because of the
importance of SVOC’s sorption effect as found in Xu and Little [5], Xu et al. [4]
designed a special stainless steel chamber to maximize the source surface area and
minimize the interior sink area (chamber walls). The authors simplified their emis-
sions model and validated it with experimental data. Similarly, Liu et al. [22] devel-
oped a model to describe SVOC emissions from a polymer slab in a chamber. The
difference between the two models was the surface adsorption mechanism. Liu et al.
assumed the surface concentration to be linear to the air concentration, whereas Xu
modeled it as a nonlinear relation. The previously mentioned SVOC models often
require users to perform complicated and time-consuming numerical computations
that significantly reduce their usability. Most recently, the EPA developed i-SVOC
software using a modified state-space (MSS) method to free users from such
computations [9].

i-SVOC
i-SVOC is a Microsoft Windows-based computer program based upon the MSS
method for simulating the emissions and sorption of SVOCs in indoor environ-
ments. MSS is a nonsteady-state model built upon the original SS method
described previously with a few modifications to improve its performance [7].
First, MSS assumes the exposed layer of building material to be 107-m thick.
Second, MSS divides a building material into ten layers, and the inner layer has
twice the thickness than the adjacent outer layer. Third, the local two-phase
mass transfer theory is used to calculate the mass transfer across the solid
material/air interface, where the mass transfer resistance within the surface layer
of solid material is taken into account. Fourth, the same theory is applied to
mass transfer across the interface of two solid materials. Based on these modifi-
cations, a mass balance equation can be written to calculate airborne concentra-
tion, emissions, and sorption rate. i-SVOC is the first general-purpose tool for
dynamic modeling of indoor SVOCs [9]. In addition to emissions factors, which
are often tested by experimental studies, i-SVOC can include multiple sink
materials (e.g., floor or wall surfaces), APs, settled dusts, and contaminant barriers
in the simulation. The outputs of i-SVOC are a time series of airborne chemical con-
centration and mass flux of each source or sink instead of a steady-state concentra-
tion such as the results of the well-mixed single-zone model specified by Standard
Council of Canada CAN/ULC-S774-09, Standard Laboratory Guide for the Deter-
mination of Volatile Organic Compound Emissions from Polyurethane Foam [28]. In
order to run i-SVOC, a few key parameters are required, including kma, Ds, and hg.
These parameters depend on chemical and building material and should be derived
for each source and sink.
172 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

Methods for Estimating Key Parameters Used

in i-SVOC
Input parameters are important because a model is useful only when the required
parameters can be estimated using readily available chemical/physical properties or
measured by experiments. The accuracy and availability of these parameters signifi-
cantly affect model predictability and usefulness. The most direct approach to esti-
mating these parameters is through laboratory emissions and sorption experiments.
In the past few decades, tremendous efforts have been devoted to designing testing
chambers and developing analytical methods to calculate these parameters by fitting
experimental data to mass transfer models [4,6,12,18,29–32]. However, laboratory
testing is often time-consuming and can only give point estimates. In addition, di-
rect measurements are often subject to the detection limit of analytical methods
and equipment for interested chemicals. An alternative approach is to establish cor-
relations between these parameters and the chemical/physical properties (e.g., vapor
pressure or molecular volume), which are either available or can be estimated (e.g.,
using quantitative structure–activity relationship models [QSARs]). Guo [33]
reviewed 48 methods that can be used for such purposes and developed a computer
program, PARAMS 1.0, to implement those methods [10]. However, existing meth-
ods were originally developed for VOCs, and correlations in PARAMS 1.0 are often
limited to VOCs emitted from building materials, excluding SPF. The applicability
of these methods to SVOCs is unknown. Other relevant studies in the literature
either test VOCs in SPF only [34,35] or test SVOCs from the source materials other
than SPF [30]. To bridge this gap and to study TCPP emissions from SPF, this
work primarily focused on how to use limited literature and experimental data to
derive key parameters required by i-SVOC.
Although other SVOCs can be estimated using the methods summarized in
this section, the focus of this study was the emissions and fate of a flame retardant
(TCPP) in SPF. There are two major types of SPF insulation: open-cell (OC) and
closed-cell (CC). Both foams are made of two components (A- and B-side) that are
sprayed at the application site with a fixed ratio. The A-side chemical is 100 % poly-
meric methylene diphenyl diisocyanate, and the B-side contains polyols, blowing
agents, catalysts, flame retardants, and surfactants. For OC SPF, the polyol used
typically is a polyether polyol, whereas CC SPF typically uses an aromatic polyester
polyol. In addition, the other chemical components in the B-side have different
weight percentages for OC and CC SPF [36]. Structurally, all cell windows in CC
SPF are intact, whereas one or more windows are open in OC SPF. The CC foam is
more rigid and has a higher density and unit R-value (R-value per inch) than the
OC foam; thus, the key parameters are expected to be different for these two types
of foam. Fig. 2 shows microscopic pictures of OC and CC SPF. Due to the limited
data in CC SPF, this work evaluated OC SPF only.
 TIAN ET AL., DOI: 10.1520/STP158920150036 173

FIG. 2 Microscopic pictures of OC SPF (left) and CC SPF (right) [37] (with permission
from SPFA).

SOURCE
Initial Concentration in SPF
The generic SPF formulations from the American Chemistry Council Center for
the Polyurethanes Industry were used to represent the initial TCPP concentration
(C0) in OC SPF [36]. TCPP accounts for 25.2 % of the B-side by weight, which is
equivalent to 1.01 by 109 lg/m3 in OC SPF. TCPP is a nonreactive flame retardant,
and thus the final concentration in the SPF product is expected to be close to the
number calculated based on weight percentages in the formulations.

Vapor Pressure
One of the essential chemical properties needed to estimate parameters for i-SVOC
is vapor pressure (VP). Because of its low volatility, the VP of TCPP is difficult to
measure experimentally. Current literature estimates VP either from high-
temperature zone extrapolation or using QSAR. Table 1 summarizes TCPP VP data
in the literature and calculated VP at 23, 40, and 65 C. The enthalpy of vaporization
was calculated as 81 kJ using the Clausius-Clapeyron equation (Eq 1) with VP data
derived by Price at 25 and 65 C [38]:
   
P2 DH v 1 1
ln ¼  (1)
P1 R T1 T2

where:
P1 and P2 ¼ VP (Pa) of TCPP at temperature T1 and T2 (K), respectively,
DHv ¼ enthalpy of vaporization, J/mol, and
R ¼ ideal gas constant, J/k/mol.

As shown in Table 1, the uncertainty of VP could be as high as a factor of 14 between

literature sources. The highest and lowest VP values at each temperature were se-
lected to estimate other parameters as discussed in the next section.
174 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 1 TCPP vapor pressure (VP) estimation.

Reference
Temperature,

Source C [39] [40] [41] [38]

Literature VP 25 1.40  103 2.69  103 7.52  103 2.00  102

40 N/A 0.1
65 1.0
Estimated VP 23 1.12  103 2.16  103 6.03  103 1.60  102
using Eq 1 at desired 40 6.71  103 1.29  102 3.60  102 9.58  102
temperature 65 6.70  102 1.29  101 3.60  101 9.58  101

Solid/Air Partition Coefficient in SPF

The solid/air partition coefficient (kma) has been identified as the single-most
important parameter that governs the indoor behavior of SVOCs [21]. Therefore, it
is crucial to estimate kma as accurately as possible. VP [10] and the octanol/air
partition coefficient (koa) are often used to establish empirical correlations with kma
[42], which can be either dependent or independent of building materials. The
building material–dependent kma is correlated with the chemical VP or koa, and
such correlation is only applicable to a certain chemical class emitted from a specific
type of building material (e.g., particle board) [35,43,44], whereas the building
material–independent approach correlates kma with VP or koa regardless of the
types of chemicals and building materials [33].
Neither direct measurements nor empirical correlations are available for the
kma of TCPP in OC SPF insulation. The best available approach is to estimate kma
using correlations developed for other types of PU foam (e.g., PU foams in consum-
er products). In one study, Zhao, Little, and Cox [35] experimentally determined
the kma of eight aromatic VOCs in a flexible polyether-type OC PU foam at room
temperature (21 6 1 C). A good correlation between ln(kma) and ln(VP) was
derived using the eight aromatics as shown in Fig. 3, which can be used to calculate
the kma of TCPP in OC SPF.
Similarly, Kamprad and Goss [43] used the Abraham equation to calculate the
kma of hundreds of VOCs in both polyether and polyester polyol–based OC PU foams
with the density ranging from 20 to 40 kg/m3. The Abraham equation relates kma to
the Gibbs free-energy change and can be expressed linearly by Abraham solvation
parameters. Unfortunately, the kma of TCPP was not reported by Kamprad and Goss.
Later, Holmgren et al. [44] expanded Kamprad and Goss’s work by including more
chemicals for PU foam but still did not include TCPP. Using data from Holmgren
et al., a correlation can be established between kma and VP (n ¼ 172; R2 ¼ 0.81).
In another study, Guo [33] developed a building material–independent method
for estimating kma based on VP only. His method was derived using 56 material/
compound combinations for VOC emissions that could be used to roughly estimate
 TIAN ET AL., DOI: 10.1520/STP158920150036 175

FIG. 3 Correlation between kma and VP of chemicals in OC PU foam using data from
Zhao, Little, and Cox [35] (left) and Holmgren et al. [44] (right).

Correlaon between Ln (kma) vs. Ln(VP) Correlaon between ln(kma) and ln(VP)
13
6.0
y = –0.7527x + 9.5434
11
4.0 y = –1.0503x + 7.4476 R² = 0.814

ln (kma)
9
ln(kma)

R² = 0.9754
2.0
7
0.0
5
2.0 4.0 6.0 8.0 10.0
–2.0 3
–4.0 –4 –2 0 2 4 6 8
ln (VP, mm Hg) ln (VP, mmHg)

kma for all material/compound combinations. The derived correlation can be

written as Eq 2, where P is the VP in millimeters of mercury (mmHg). To use this
method, kma has to be adjusted by the density of OC SPF (0.008 g/cm3) by multiply-
ing 0.008 to the kma calculated in Eq 2:

ln kma ¼ 8:76  0:785  ln P (2)

Table 2 gives the estimated kma of TCPP using different methods based on the
selected VP values for OC SPF. The results show Guo’s [33] method gives the low-
est kma, whereas Zhao, Little, and Cox’s [35] method gives the highest. Comparing
these two methods, the correlation between the kma and VP derived from Kamprad
and Goss’s [43] and Holmgren et al.’s [44] data is specific to OC PU foam and
covers more chemicals. Therefore, it was further used for i-SVOC simulations in
this study.
As shown in Table 2, using VP values and empirical methods from different
literature, the estimated kma values range from 6.11 by 104 to 3.7 by 108 at 23 C,
which is more than three orders of magnitude difference. There are three possi-
ble reasons for such differences. First, for the two material-dependent methods,

TABLE 2 Estimated TCPP solid/air partition coefficient kma at different temperatures.

Kamprad and Goss

Zhao, Little, and [43] and Holmgren
Temperature,  C Estimated VP, Pa Cox [35] Guo [33] et al. [44]
3 8
23 1.12  10 3.75  10 5.00  105 9.21  107
2 7 4
1.60  10 2.25  10 6.11  10 1.25  107
3 7 5
40 6.71  10 5.62  10 1.21  10 2.40  107
2 6 4
9.58  10 3.43  10 1.50  10 3.24  106
2 6 4
65 6.71  10 5.00  10 1.98  10 4.24  106
1 5 3
9.58  10 3.05  10 2.45  10 5.72  105
176 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

the studied OC PU foam could be different from SPF insulation in both chemi-
cal and physical properties because each manufacturer may have a different for-
mulation. The correlation derived from a specific literature may only be applicable
to that type of PU foam. Second, the chemical compounds or material/compound
combinations evaluated in each method are limited and often cover only VOCs.
The highest kma reported in those studies is in the range of 104 (with a corre-
sponding VP of approximately 20 Pa at room temperature), whereas the kma of
SVOCs is often in the range of 107 (with a VP in the range of 103 to 102 Pa at
room temperature). Those empirical methods derived from limited VOCs may
not be applicable for SVOCs. Third, among the three methods listed in Table 2,
Guo’s method was derived from material-independent correlation and adjusted
by material density, whereas the other two methods were derived specifically for
OC PU foam. Differences would be expected when using the two approaches to
estimate kma.

Gas-Phase Mass Transfer Coefficient in Ambient Air

The gas-phase mass transfer coefficient (hg) is also called the convective mass trans-
fer coefficient [3], which is used to quantify the diffusion flux at the material sur-
face. As discussed previously, SVOCs are often externally controlled; thus, hg is
more important for characterizing SVOC emissions than VOC. In addition, previ-
ous studies have shown that hg mainly affects the initial period of emissions [11,18];
thus, estimating hg is important for calculating the early-stage concentration of
TCPP in indoor environments. Guo reviewed seven methods for estimating hg. One
method developed specifically for indoor environments, the Sparks method, is
expressed in Eq 3.
 2
1 lq 3
hg ¼ 0:33Da L3 (3)
l

where:
Da ¼ air diffusivity, m2/s,
L ¼ characteristic length of a source, m, calculated as the square root of the
source area,
u ¼ air velocity, m/s,
q ¼ air density, kg/m3, and
l ¼ viscosity of air, kg/m/s.
Table 3 gives the estimated hg values with given air properties and source dimen-
sions. All parameters in this table were estimated using PARAMS 1.0 [10] except
the characteristic length of a source (L) and air diffusivity (Da) at 65 C. Three L val-
ues were derived based on the sample area used in the microchamber emissions
study and two theoretical indoor insulation sources (30 m2 for the wall and 100 m2
for the ceiling) to simulate field applications of SPF. A Da at 65 C was not available
in PARAMS 1.0 and was thus estimated using the Chapman-Enskog theory [45,46].
 TIAN ET AL., DOI: 10.1520/STP158920150036 177

TABLE 3 Estimated hg for TCPP from calculated air and source properties.

Temperature,  C Da, m2/s L, m u, m/s q,a kg/m3 l, kg/m/s hg, m/h

6 5
23 4.57  10 0.057 0.05 1.19 1.85  10 3.07
0.1 4.88
40 5.09  106 0.05 1.12 1.90  105 3.23
0.1 5.13
65 5.90  106 0.05 1.04 2.02  105 3.42
0.1 5.43
23 4.57  106 5.48 0.05 1.19 1.85  105 0.67
0.1 1.07
40 5.09  106 0.05 1.12 1.90  105 0.70
0.1 1.12
65 5.90  106 0.05 1.04 2.02  105 0.75
0.1 1.18
23 4.57  106 10 0.05 1.19 1.85  105 0.55
0.1 0.87
40 5.09  106 0.05 1.12 1.90  105 0.58
0.1 0.92
65 5.90  106 0.05 1.04 2.02  105 0.61
0.1 0.97
a
Air density is calculated for dry air (relative humidity ¼ 0 %), which is consistent with the experi-
mental chamber study [47].

Typical air velocity was between 0.05 and 0.1 m/s in an indoor environment based on
PARAMS 1.0. The estimated hg ranged from 0.55 m/h at 23 C for a 100-m2 source
with an average 0.05-m/s air velocity to 5.43 m/h at 65 C for a 3.2 by 103-m2 source
with an average 0.1-m/s air velocity. Results show that hg increases with temperature
and air velocity but decreases with characteristic length. In terms of hg, the major
difference between microchamber and real field application is the characteristic
length. The hg of a microchamber SPF could be five times higher than a 100-m2
ceiling SPF installed in a house. The impact of hg on modeling results is illustrated
in the sensitivity analysis sections.

Solid-Phase Diffusion Coefficient (Ds)

Another essential parameter for estimating SVOC emissions from SPF is the solid-
phase diffusion coefficient (Ds). Ds characterizes the diffusion within the solid mate-
rials and is both chemical and material dependent.
Existing literature disagrees on which physical property is most predictive in estab-
lishing a correlation with Ds. Zhao, Cox, and Little [48] first correlated Ds with VP line-
arly in a log-log scale for water and eight aromatic hydrocarbons in an OC PU foam.
However, such correlation was found unsatisfactory by other researchers [33]. Zhao,
Cox, and Little [35] then revised their study by using molecular free surface area (SM)
as a parameter to correlate Ds for eight aromatic VOCs as shown in Eq 4. SM could be
178 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

estimated using the number of valence electrons (Zv) provided by Gavezzotti [49] as
shown in Eq 5. The Zv of TCPP is 104, and the corresponding SM is 299 Å2. At the
experimental temperature (21 C), the calculated Ds values ranged from 6.60 by 1019
to 2.35 by 1016 m2/s, which were significantly lower than other SVOCs with a similar
SM. For example, Liu, Guo, and Roache [6] reported 2,20 ,4,5,50 -pentachlorobiphenyl
(SM ¼ 257 Å2) emitted from an amorphous polyolefin-type material (e.g., low-density
polyethylene [LDPE]) has a Ds value in the range of 1013 m2/s.

logðDs Þ ¼ 9:14ð60:38Þ  0:026ð60:003ÞSM (4)

SM ¼ 2:601ZV þ 28:13 (5)

Guo [33] suggested expressing Ds as a function of the chemical molar volume.

However, his method was specifically designed for VOCs; thus, the applicability to
SVOCs is unknown. Using the Ds values experimentally determined by Zhao et al.
[35] for eight aromatic VOCs, a correlation between molar volume (Vs) and Ds can
be established as shown in Eq 6 by excluding naphthalene, which has a similar Vs as
the other seven aromatic VOCs but a much less Ds. At 21 C, the calculated Ds for
TCPP in an OC PU foam was 1.34 by 1014 m2/s (TCPP Vs ¼ 260) [50]. The tem-
perature dependence of a diffusion coefficient is generally well represented by the
Arrhenius equation as Eq 7 [51], where DE is the activation energy and D1 and D2
are diffusion coefficients at temperatures T1 and T2, respectively. Therefore, Ds at
higher temperatures could be calculated based on Ds at 21 C if DE is known. DE of
TCPP was calculated using the emissions study data in the microchamber study as
shown in Fig. 4 by plotting 1/T and the natural logarithm of the emissions factor
[30]. Using the average of the two slopes in Fig. 4, DE was calculated as 95.8 kJ/mol.
Therefore, using Eq 7, Ds at 23, 40, and 65 C was calculated as 1.75 by 1014, 1.45
by 1013, and 2.21 by 1012 m2/s, respectively.

lnðDs Þ ¼ 3:518  lnðVs Þ  12:393 (6)

   
D2 DE 1 1
ln ¼  (7)
D1 R T2 T1

Begley et al. [52] used the Piringer equation, which correlates Ds with chemical
molecular weight (MW) for a specific material as shown in Eq 8, where Ap is a
material-specific coefficient:
 
2 10; 450
Ds ¼ exp Ap  0:1351ðMWÞ3 þ 0:003MW  (8)
T

This method was originally developed to study chemical release from consumer
goods packaging, but the same mass transfer mechanism could be applied to study
emissions from insulation materials. For polyolefins such as LDPE and PU foam,
the average Ap value is 10. Table 4 summarizes estimated Ds values used in this
work.
 TIAN ET AL., DOI: 10.1520/STP158920150036 179

FIG. 4 Microchamber emissions factor measurements from OC SPF (left: with skin;
right: without skin).

TCPP Emission Factor (Skin on) TCPP Emission Factor (Skin off)
9.0 9.0
8.0 8.0

ln (Emission Factor)
ln (Emission Factor)

7.0 7.0
6.0 6.0
5.0 5.0
4.0 y = –11069x + 40.986 4.0 y = –11984x + 43.668
3.0 R² = 0.98 3.0 R² = 0.98
2.0 2.0
1.0 1.0
0.0 0.0
2.8E-03 3.0E-03 3.2E-03 3.4E-03 2.8E-03 3.0E-03 3.2E-03 3.4E-03
1/T (T in K) 1/T (T in K)

The methods used in this section to estimate i-SVOC-required input parame-

ters for indoor chemical sources are summarized in Table 5.

SINKS
Indoor sinks are an important factor to consider in studying the behavior of SVOCs
because when the airborne concentration is high, SVOCs could diffuse or adhere to
indoor sinks such as dry wall and plywood flooring, whereas when the airborne
concentration is low, indoor sinks may become sources. i-SVOC software takes into
account the indoor sink effects by calculating the mass flux of SVOC sinks. For the
purposes of illustration, this study considered two diffusional indoor sinks: gypsum
board walls and plywood flooring. Similar to the parameters needed for sources, the
same set of parameters was needed for sink materials. Current research focuses on
the mass transfer of polychlorinated biphenyl congeners and plasticizers (e.g.,
phthalates) into building materials such as flooring and concrete [4,6]. For flame
retardants such as TCPP, a few emissions studies were conducted, but the data were
not sufficient to derive those key parameters to characterize the mass transfer
process. PARAMS 1.0 can be used as a starting point. Table 6 gives the parameters
estimated for the two diffusional indoor sinks using PARAMS 1.0 at 23 C. It is
assumed that there are five pieces of gypsum board surfaces in the room.

TABLE 4 Estimated Ds at different temperatures.

Temperature,  C Guo [33], m2/s Begley et al. [52], m2/s

14
23 1.75  10 3.51  1014
13
40 1.45  10 3.04  1013
12
65 2.21  10 3.59  1012
180 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 5 Summary of methods used to estimate source parameters in i-SVOC.

Parameters Method/Equation References

kma lnðkma Þ¼1.0503 lnðVPÞ þ 7.4476 [35]

lnðkma Þ ¼ 0.7527 lnðVPÞ þ 9.5434 [44]
lnðkma Þ ¼ 8.76  0.785 lnðVPÞ [33]
hg 1
 23 [10]
hg ¼ 0:33Da L3 uq l

Ds lnðDs Þ ¼ 3:518  lnðVs Þ  12:393 [33,35]

 2

Ds ¼ exp Ap  0:1351ðMWÞ3 þ 0:003MW  10;450
T [52]

One 100-m2 piece covers the ceiling and the other four 30-m2 pieces cover four walls.
One 100-m2 piece of plywood material is assumed to cover the floor. For TCPP, the
kma for both sink materials was calculated based on the empirical equations derived
from aromatics. Since the primary predictor of kma is VP, the lower and higher kma
values were calculated using the VP values in Table 1. Molar volume is the key predic-
tor for Ds, and both materials were calculated using the correlation derived from
mixed-compound classes. For the gas-phase mass transfer coefficient (hg), which is
dependent on sink dimensions and air properties, the lower and higher values were
calculated using an air velocity of 0.05 and 0.1 m/s, respectively.

AIRBORNE PARTICLES
The interaction between APs and indoor pollutants has been studied more recently
and identified as an important pathway for human exposure to SVOCs
[5,21,27,54,55]. The mass transfer process of SVOCs to APs is more complex than
indoor sources and sinks because such a process is also dependent on particle size
and density, release rate, residence time, organic carbon fraction, and particle mor-
phologies. Little information is available on deriving such data for TCPP. Liu et al.
[55] systematically discussed the dynamic interactions between SVOCs and APs.
Correlations presented by Liu et al. were used in this study.

Particle/Air Partition Coefficient

Similar to the kma of building materials, kpa characterizes the chemical partition
between APs and air. EPA EPI SuiteTM software [41] uses the octanol/air partition

TABLE 6 Parameters estimates for indoor sinks (at 23 C)a.

Type of material Area, m2 Thickness, m kma Ds, m2/h hg, m/h

Gypsum board 100 0.0125 7

1.14  10 –8.84  10 7
4.07  109 0.55–0.87
30 0.67–1.07
Plywood 100 0.03 5.57  107–6.84  108 7.43  109 0.55–0.87
a
Assuming only the top side is exposed to indoor air.
 TIAN ET AL., DOI: 10.1520/STP158920150036 181

coefficient (koa) to estimate kpa. For TCPP, log(koa) was estimated to be 8.203,
and the corresponding kpa was 3.92 by 105 m3/lg. Assuming the AP density was
1 g/cm3, the dimensionless kpa was 3.92 by 107. This value is less than other flame
retardant kpa values reported in the literature [55]. In addition to the difference in
koa values, another possible explanation is that both organic and inorganic consti-
tutes of particles contribute to kpa, whereas EPI SuiteTM only considers TCPP sorp-
tion to organics. Liu et al. [55] incorporated both contributions by introducing a
partition coefficient between inorganic elemental carbon in particles and air
(kEC_a). However, the key parameters for estimating kEC_a, including the partition
coefficient between elemental carbon and water, are only available for flame retard-
ants such as brominated diphenyl ethers (BDEs) but not TCPP.

Particle-Phase Diffusion Coefficient

The particle-phase diffusion coefficient (Dsp) is a parameter determined by particle
morphologies. Liu et al. [55] derived two semiempirical equations to estimate Dsp.
This method can differentiate between particles fully consisted of or covered by or-
ganic liquid (Fig. 5, left) and porous particles (Fig. 5, right). Due to limited data,
only particles fully covered by organic liquids (Fig. 5, left) were modeled in this
study to account for the interaction between APs and SVOCs. Therefore, only Dsp
in organic liquid was calculated with an average Ds value of 1.92 by 1010 m2/s by
Eq 9 proposed by Levin [56] and Theis et al. [57]. For the two flame retardants
(BDE-47 and BDE-99) calculated by Liu et al., Dsp in porous particles is less than in
organic liquid but within one order of magnitude.

Dsp  MW0:5 ¼ ð1:1 6 0:11Þ  1010 (9)

FIG. 5 Schematic of morphologies of indoor airborne particles (left: particle fully

covered by organic liquid; right: porous particle).
182 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

Gas-Phase Mass Transfer Coefficient for Particles

Similar to hg, the gas-phase mass transfer coefficient (hp) characterizes the chemical
mass transfer from ambient air to particles. Li and Davis [58] developed a semiem-
pirical equation to calculate the external mass transfer coefficient for particles as
shown in Eq 10 and Eq 11.

Dg 1 þ Kn
hp ¼  (10)
R2 1 þ 1:71Kn þ 1:333Kn 2
rffiffiffiffiffiffiffiffiffiffiffiffiffi
k Dg 8RT
Kn ¼ ; k ¼ 3 ; c ¼ (11)
R2 C p MW

where:
Dg ¼ diffusion coefficient of SVOCs in air as determined by Sparc Performs
Automated Reasoning in Chemistry [59], m2/s,
R2 ¼ outer radius of AP,
Kn ¼ Knudsen number,
k ¼ mean free path of SVOCs, m,
c ¼ mean molecular speed of SVOCs, m/s,
R ¼ ideal gas constant, J/K/mol,
T ¼ temperature, K, and
MW ¼ molecular weight of SVOCs, kg/mol.
Table 7 lists the parameters used to calculate hp (m/s). The hp of TCPP was calculat-
ed as 0.74 m/s or 2,669 m/h.

Other Parameters
In addition to input parameters that depend on the AP characteristics discussed
previously, a few other parameters were needed by i-SVOC to model the interaction
between APs and SVOCs. Table 8 summarizes all the parameters needed to calculate
SVOCs absorption to APs in i-SVOC. It is assumed there is a pulse release of APs
at time zero with zero initial SVOC concentration. The number of APs per cubic
meter can be calculated by Eq 12 [33]. The deposition rate constant was assumed to
be 0.6 h1 per Guo’s study [33].

TABLE 7 Parameters needed to calculate TCPP hp.

Parameter Value Unit

T 296 K
MW 0.327 kg/mol
Dg 3.75  106 m2/s
R2 5 lm
R 8.314 J/K/mol
Kn 0.016 —
 TIAN ET AL., DOI: 10.1520/STP158920150036 183

TABLE 8 Parameters for the AP module in i-SVOC.

Parameter Value Unit

Release mode Pulse —

Diameter of particle 10 lm
Diameter of solid core 1 lm
Initial SVOC concentration 0 lg/g
Density of liquid film 1,000 kg/m3
Deposition rate constant 0.6 h1
Pulse release time 0 h
Density of solid core 1,000 kg/m3
AP number concentration 2.87  105 #/m3
7
kpa 3.92  10 —
Dsp 1.92  1010 m2/s
hp 2,669 m/h

Mp
Np ¼ 6  106 (12)
p  q  dp 3

where:
Np ¼ AP number concentration in room air, #/m3,
Mp ¼ AP mass concentration in room air ¼ 150 lg/m3 [60] (assuming the
indoor PM10 concentration equals the National Ambient Air Quality Standard of
100),
q ¼ density of AP, g/cm3, and
dp ¼ AP diameter, lm.

MODELED ROOM CONDITIONS

It is assumed the ceiling and four walls of the hypothetical room are covered by
gypsum board and the floor by plywood. Table A.1 gives the volume, air change per
hour (ACH), and initial SVOC airborne concentration for the hypothetical room
and microchamber. It also shows the calculated loading factor, which is the area of
SPF installed divided by the volume.

Results and Discussion

The four modeling outputs (see Fig. 7, Fig. 8, and Figs. B.1–B.4 for illustrations)
include TCPP airborne concentration (Ca), source emissions factor (EFsa), sink
sorption factor (SFan), and mass flux of APs (Rap) at three temperatures. Because of
uncertainties of the key modeling parameters, the high emission/low sink (HELS)
and low emission/high sink (LEHS) represent the upper and lower boundaries of
each modeling outputs, respectively. The following sections discuss each result in
detail.
184 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 6 Modeled indoor TCPP airborne concentration.

600

500
Concentraon (μg/m3)
H.E.L.S. (23°C)
400
H.E.L.S. (40°C)
H.E.L.S. (65°C)
300
L.E.H.S. (23°C)
200 L.E.H.S. (40°C)
L.E.H.S. (65°C)
100

0
1 10 100 500
Time (h)

AIRBORNE CONCENTRATION
With the hypothetical room conditions listed in Table A.1, Ca was plotted in Fig. 6.
As listed in Table A.2 and Table A.3, the HELS scenario used the lowest source kma
along with the highest source Ds and hg values to predict the highest emissions fac-
tor while at the same time using the lowest sink kma to model the lowest sink sorp-
tion factor (sink Ds was fixed and sink hg needs to be consistent with source hg).
The HELS scenario resulted in the highest Ca with the given range of input parame-
ters. In contrast, the LEHS scenario resulted in the lowest Ca. The predicted concen-
trations at 500 h for the HELS scenario were 30, 70, and 500 lg/m3 at 23, 40, and
65 C, respectively, which show the importance of temperature because input
parameters such as kma, Ds, and hg all depend on temperature. In comparison, at
500 h, the predicted concentrations were, respectively, 4, 20, and 90 lg/m3 in the
LEHS scenario at the three temperatures. The results also demonstrate that, by in-
creasing the source temperature from 23 to 65 C, Ca increases by approximately
20 times. Between the high- and low-emission scenarios, the difference of Ca is
within a factor of 10, which is roughly proportional to the highest uncertainty in
input parameters that occurs for kma.

SOURCE EMISSIONS FACTOR

Fig. 7 presents the source emissions factor (EFsa) of OC SPF with two source dimen-
sions (100 and 30 m2) at three temperatures. In the HELS scenario, EFsa at 500 h
ranged from 40 to 800 lg/m2/h, whereas in the LEHS scenario, EFsa was between
4 and 100 lg/m2/h depending on the temperature and SPF dimensions. During the
modeling period from 0 to 500 h, EFsa became stable after a few hours and gradually
decreased to 90 % of values in the initial hours. In addition to the temperature
 TIAN ET AL., DOI: 10.1520/STP158920150036 185

FIG. 7 Modeled TCPP emissions factor from OC SPF (left: HELS; right: LEHS; 100 and
30 m2 are the area of the SPF).

900 120

800
100

Emission Factor (μg/m2/h)

Emission Factor (μg/m2/h)

700

600 100 m2 (23°C) 80 100 m2 (23°C)

500 30 m2 (23°C) 30 m2 (23°C)
100 m2 (40°C) 60 100 m2 (40°C)
400 30 m2 (40°C) 30 m2 (40°C)
300 100 m2 (65 C) 40
100 m2 (65°C)
30 m2 (65°C)
200 30 m2 (65°C)
20
100

0 0
1 10 100 500 1 10 100 500
Time (h) Time (h)

effects (20 times between 23 and 65 C) and emission/sink scenarios (within one
order of magnitude) discussed earlier, the SPF source dimension also caused differ-
ences in EFsa. As the result of a higher hg value, the 30-m2 SPF emitted approximately
10 to 20 % more than a 100-m2 SPF per unit area under the same temperature.

SINK SORPTION FACTOR

Two diffusional sink materials, gypsum board and plywood, were considered in this
study. Fig. B.1 and Fig. B.2 present the TCPP sorption factor (SFan) by the sink

FIG. 8 Sensitivity analysis of key input parameters on modeling results. Use the OC SPF
LEHS scenario for the 100-m2 source and sink at 23 C.

Impact of Input Parameters on Modeling Results

1000% 978% 965% 978%
kma decreased by a factor of 10;
900%
hg, Ds and ACH increased by a factor of 10.
800% 890% 890% 890%
Concentration (Ca)
700%
Emission Factor (EFsa)
Percentage

600%
Sorption Factor (SFan)
500%

400%
304% 304% 306%
300%
242%
200% 239% 239% 240%
220% 139% 138%
101% 104% 100% 100%
100% 126%
100% 100% 98% 93% 31% 33% 125% 33%
0% 25% 26% 26%
Ca EFsa SFan Ca EFsa SFan Ca EFsa SFan Ca EFsa SFan Ca EFsa SFan Ca EFsa SFan
Kma (Source) hg (Source) Ds (Source) Kma (Sink) hg (Sink) ACH
186 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

materials, which depend on material dimensions, emission/sink scenarios, and tem-

peratures. When OC SPF existed, gypsum board SFan at 500 h ranged from 2 lg/m2/h
for the 100-m2 material in the LEHS scenario (23 C) and 420 lg/m2/h for the 30-m2
material in the HELS scenario (65 C). The plywood SFan was similar to but slightly
higher than that for the 100-m2 gypsum board. When the temperature increased, the
increase of EFsa and Ca led to a higher SFan. In terms of the emission/sink scenarios,
unlike source EFsa, the sink kma had little effect on sink SFan, which only depends on hg
and Ca as discussed previously. Therefore, the variation of SFan at the two emission sce-
narios was within one order of magnitude as well, which is consistent with the variation
of Ca. Regarding the effect of sink dimensions on EFsa, the trend was similar to SFan.
It is important to note that under these modeling conditions and for the same
dimensions, the sink material SFan was approximately 56 to 61 % of the source EFsa.
When sink material exists, on the one hand, the TCPP emitted is partially absorbed
by the sink, which may decrease Ca, whereas on the other hand, EFsa may increase
because of this sink sorption effect, which may increase Ca. These two opposite effects
of sink material on Ca cancel each other to a certain extent. When the TCPP emis-
sions from OC SPF were modeled by using the same hypothetical room but without
sink materials, the new simulated EFsa was approximately 40 % of the previous EFsa
value, but Ca almost doubled.

MASS FLUX TO AIRBORNE PARTICLES

This work also evaluated TCPP absorptions to APs. The estimated modeling input
parameters are shown in Table 8. In the initial hours, TCPP mass flux to APs (Map)
was in the same range as the EFsa and SFan but quickly decreased by two to four
orders of magnitude. Research shows that for APs with a 0.1-lm diameter, when
kpa is in the range of 108, APs reach saturation within a few minutes [8]. Therefore,
Map is excluded from further discussion.

COMPARISONS BETWEEN MODELING RESULTS AND EXPERIMENTAL

MEASUREMENTS
It is important to validate mass transfer models such as i-SVOC with experimental
studies such as the ones using microchambers. To perform this task, we simulated
source EFsa using the testing conditions in the microchamber study such as ACH,
foam thickness, exposed area, head space volume, and corresponding hg values as
shown in Table 3 and Table A.1 but adopted the same kma and Ds in the HELS and
LEHS scenarios. Because the microchamber was designed to minimize the sink
sorption effect in the chamber, no sink materials or APs were modeled in these spe-
cific simulations. i-SVOC results found that EFsa ranges from 55 to 1,000 lg/m2/h
and 7 to 165 lg/m2/h at 500 h for the high- and low-emission scenarios, respec-
tively. Table 9 presents the ratio of the i-SVOC-modeled EFsa using microchamber
conditions and the lab measurements [47]. The first and second numbers after each
temperature are the LEHS and HELS scenarios, respectively. From these results, the
HELS scenario was roughly in the same order of magnitude as lab measurements.
 TIAN ET AL., DOI: 10.1520/STP158920150036 187

TABLE 9 i-SVOC modeling results compared with microchamber laboratory measurements at

Day 20a.

Temperature,  C EFsa, % Ca , %

23 30–220 11–80
40 9–47 3–14
65 5–32 2–11
a
Compared to OC SPF skin-off results, the percentage varies during the 500 h of simulation time.

However, the predicted EFsa values were higher than measurements at 23 C but
lower at 40 and 65 C by a factor of 2 to 3.
A second comparison was made between the Ca (Fig. 6) in the hypothetical
room (Table A.1) modeled by i-SVOC and that was calculated based on the micro-
chamber EFsa values using a single-zone, well-mixed steady-state model per CAN/
ULC-S774-09 (Eq 13).

EFsa  LF
Ca ¼ (13)
N

where:
EFsa ¼ measured emissions factor, lg/m2/h,
LF ¼ loading factor, which is the ratio of SPF area installed to room volume
(L ¼ 220/300 ¼ 0.73 m2/m3, which is consistent with the hypothetical room), and
N ¼ room air exchange rate at 0.5 h1.
As shown in Table 9, the Ca calculated by the i-SVOC in the HELS scenario was
approximately 80 % of that derived from the measured EFsa at 23 C but only
approximately 10 % at 40 and 65 C. The possible causes of such differences are dis-
cussed herein. From the modeling perspective, a few key parameters such as kma
and hg need to be better studied. For kma, because the literature values used to esti-
mate these parameters were primarily derived from noninsulation PU foams, the
empirical relations may not accurately represent the SPF insulation. The EFsa
results indicate that the true kma might be higher than the current values at 23 C
but lower at 40 and 65 C. In addition to kma, air velocity in the microchamber
could be higher than 0.05 to 0.1 m/s, which leads to higher hg values. Therefore,
EFsa and Ca in the microchamber study were higher than the results predicted by
i-SVOC. In addition, laboratory measurements showed that TCPP EFsa went up
and down during the first few days and became steady after Day 7, whereas the
i-SVOC-predicted results showed a slowly decreasing EFsa after 10 h. The potential
reason could be that i-SVOC assumes SPF properties (e.g., density, kma, and Ds) are
uniform across the entire thickness, but this may not be true in reality. From the ex-
perimental measurement perspective, because of limited measurements during the
20-d period, the experimental values may only represent the SPF emissions rate at a
specific time, which may also lead to results different from model predictions.
188 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

Sensitivity Analysis
It is important to understand the impact of key input parameters (ACH, kma, Ds,
and hg) on modeling results. To quantify how modeling results vary due to a change
of each input parameter, a sensitivity analysis was conducted by establishing the
OC SPF LEHS scenario as the baseline with a few modifications. First, source mate-
rial was limited to only one 100-m2 OC SPF. Second, diffusional sink material was
limited to one 100-m2 gypsum board. Each previously mentioned parameter was
then varied by a factor of 10, one at a time, to generate the new results. A factor of 10
was chosen because it was large enough to generate significant variations in the model-
ing results. Both the source (Table A.2) and sink kma (Table A.3) decreased by a factor of
10, whereas source (Table A.2) and sink hg (Table A.3) and ACH (Table A.1) increased by
a factor of 10. Two temperatures (23 and 65 C) were studied, and the results are shown
in Fig. 8 and Fig. B.5, respectively. In Fig. 8, the y-axis represents the variation of each of
the three modeling results (Ca, EFsa, and SFan) when the input parameters changed.
The highest number of each bar shows that the highest variation occurred in the
elapsed 500 h, whereas the lowest number is the lowest variation.
Fig. 8 clearly shows that the modeling results (Ca, EFsa, and SFan) increase pro-
portionally as the source kma decreases but are almost unchanged when the sink
kma decreases. Similarly, the results are more sensitive to the change in hg for the
source than for the sink. When source hg increases, both EFsa and SFan increase.
Therefore, the TCPP concentration increases accordingly. In comparison, when
sink hg increases, SFan increases at the same ratio as when source hg increases, but
EFsa only increases by approximately 30 %, which leads to a lower Ca. For ACH, an
increase of 10 times only results in about an approximately 30 % higher EFsa and
leads to a much lower Ca and SFan. A similar trend can be found for 65 C in
Fig. B.4, in which the results show less sensitivity to the change of source kma and hg
than at 23 C but are still significant. The impact of sink parameters and ACH do
not change with temperatures.
The effect of source Ds is more complicated because it depends on both the rel-
ative values of source kma and Ds. When kma is high (e.g., approximately 107–8), Ds
has a minimal effect on modeling results, as shown in Fig. 8. However, when kma is
low (e.g., 106), the change of Ds could affect the modeling results more significantly
if Ds is also low (in the range of 1012 to 13 m2/h). For example, at 65 C, kma is in
the range of 106. When Ds decreases from 109 to 1011 (m2/h) by two orders of
magnitude, the three modeling results (Ca, EFsa, and SFan) decrease by up to 18 %,
whereas if Ds were to decrease from 1012 to 1013 m2/h by only one order of mag-
nitude, the three modeling results (Ca, EFsa, and SFan) would decrease by more
than 80 %.
The results shown in this sensitivity analysis can be explained mathematically
based on the MSS method. The mass transfer rate between solid material and air
is expressed in Eq 14, which reveals the quantitative relation between emission/
sorption factor and input parameters.
 TIAN ET AL., DOI: 10.1520/STP158920150036 189

 1  
DL 1 Cm
EFsa or SFan ¼ þ   Ca (14)
2kma Ds hg kma

where:
EFsa or SFan ¼ mass transfer factor between solid material and air or air and
sink material, lg/m2/s,
Ca ¼ indoor/chamber TCPP air concentration, lg/m3,
kma ¼ material/air partition coefficient,
Ds ¼ solid-phase diffusion coefficient, m2/s,
Cm ¼ concentration in the exposed slice of material, lg/m3,
hg ¼ gas-phase mass transfer coefficient, m/s, and
DL ¼ 107 m.
Based on the kma, Ds, and hg values used in the OC SPF LEHS scenario, Eq 14 can
be further reduced to Eq 15 since hg (in the range of 104 m/s) is much smaller
than 2kDL
ma Ds
(in the range of 100 to 101 m/s) at all three temperatures and two emis-
sion scenarios studied.
 
Cm
EFsa or SFan ¼ hg   Ca (15)
kma

For gypsum board selected as the diffusional sink material in this sensitivity analysis,
3
Cm
kma is very small (in the order of 10 lg/m3 in the early hours and then up to 102
lg/m3 at 500 h), whereas Ca is in the order of 100 across the modeled period. There-
fore, practically speaking, SFan is linearly related to Ca but independent of sink kma
because the surface fugacity difference ( kCmam  Ca) is dominated by Ca. When EFsa and
Ca are unchanged, a decrease of sink kma has a minimal effect on modeling results. In
contrast, for the source material, when kma decreases, both kCmam and Ca increase, which
leads to a linear increase of the fugacity difference at the surface. This is the reason
that modeling results increase proportionally with the decrease of source kma but
remain almost unchanged with sink kma.
Unlike source kma, a change in hg or ACH affects modeling results dispropor-
tionally. When source hg increases by 10 times, EFsa is only twice as high as be-
fore. The reason is that, although an increase of hg leads to higher emissions, Ca
increases as well. When source kma remains the same, kCmam is unchanged and
( kCmam  Ca) decreases; therefore, EFsa does not increase as much as hg. A practical
explanation is that because kma is unchanged, TCPP potentially available at the
boundary layer to be transferred by air is not changed. Therefore, even the gas-
phase mass transfer coefficient increases by 10 times, and EFsa does not increase
as much as hg. When source hg increases, SFan changes at the same percentage as
Ca, which can be explained by using the same logic that kCmam is much less than Ca.
Similarly, when sink hg increases by 10 times, because Ca decreases, the increase
of SFan is less than 10 times. At the same time, EFsa increases approximately 30 %
due to a decrease of Ca. Regarding modeling sensitivity to ACH, it has very similar
190 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

results as sink hg. The only difference is that when ACH increases, both Ca and
SFan decrease.
Although Ds is not in Eq 15, it may have effects on modeling results through Cm
as discussed earlier. This is because Eq 15 only characterizes the mass transfer between
the top layer of the solid material and air, whereas mass transfer within solid material
is governed by Ds. When kma is not high enough, which means a higher emission
potential, if Ds is low (approximately 1012 to 13 m2/h) at the meantime, the inner
layer cannot supply the top layer with enough TCPP. The emissions factor then
decreases. In contrast, if kma is high enough, which means lower emission potential,
then the emissions process is less likely to be limited by mass transfer within the solid
materials.

Conclusions
SVOC emissions from SPF is an important indoor air-quality research area. This
study used i-SVOC to evaluate the TCPP emissions from SPF and its potential
interaction with indoor sinks and APs. The preliminary findings indicate that the
modeling results are generally in agreement with experimental measurements
but may underestimate TCPP emission rates at higher temperatures using cur-
rent input parameters. This research demonstrates how to estimate key parame-
ters used in i-SVOC and their impacts to the final results. There are a few
advantages of using mass transfer models to estimate SVOC emissions. First,
sink materials can be included to represent real-world conditions. Second, mod-
els are not limited to the experimental detection limits often encountered in
chamber studies. Third, instead of point measurements, models can generate
time-series data, which give more information on the SPF emissions profile and
SVOC behavior in indoor environments. Last, mass transfer models reveal a clear
physical process of SVOC emissions from building materials, and the under-
standing of such a process is essential for expanding this work to study other
chemicals in addition to TCPP that may emit from a source material. Although
modeling has its advantages, further research is needed to refine the key input
parameters so that the predictability of mass transfer models can be improved.
For this work, due to limited literature data, kma had the highest uncertainty and
greatest impact on TCPP emissions and needs additional research. In addition,
the uniform material property assumption could be revised for future modeling.
For example, SPF could be modeled as two adjacent materials (skin versus main
foam) with different physical characteristics. In addition, more research should
be done to study the behavior of TCPP in indoor sinks and APs by experimental
measurements or by developing empirical equations using readily available
chemical/physical properties. Overall, this work demonstrates that mass transfer
models such as i-SVOC are promising tools for studying SVOC emissions from
SPF, but future collaborations are needed between model developers and experi-
mental scientists to validate and improve these tools, especially experimental
 TIAN ET AL., DOI: 10.1520/STP158920150036 191

measurements of key modeling input parameters. Chemical manufacturers, gov-

ernment agencies, and other value-chain stakeholders should work together to
better understand SVOC emissions from building materials.

ACKNOWLEDGMENTS
We thank Zhishi Guo, Xiaoyu Liu, and Rick Duncan for their help and advice on this
paper.

Appendix A: Parameters Used in i-SVOC

Simulation Scenarios

MODELED ROOM CONDITIONS

TABLE A.1 Modeled room conditions.

Ventilation Rate, Initial TCPP

Volume, m3 h1 Loading Factor, m2/m3 Concentration in Air

Hypothetical room 300 0.5 0.73 0

Microchamber 1.61  105 188 200 0

SOURCE

TABLE A.2 Source parameters and modeling scenarios for TCPP emissions from OC SPFa.

Temperature,  C Area, m2 kma,  107 Ds,  1010 m2/h hg, m/h Emission

23 100 9.36 0.63 0.55 Low

1.25 1.26 0.87 High
30 9.36 0.63 0.67 Low
1.25 1.26 1.07 High
40 100 2.40 5.22 0.58 Low
0.32 10.94 0.92 High
30 2.40 5.22 0.70 Low
0.32 10.94 1.12 High
65 100 0.42 79.56 0.61 Low
0.057 129.24 0.97 High
30 0.42 79.56 0.75 Low
0.057 129.24 1.18 High
a
Foam thickness was assumed to be 0.1 m for OC SPF. In the microchamber study, the thickness
was measured as 0.05 m.
192 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

SINKS

TABLE A.3 Sink parameters and modeling scenarios for TCPP emissions from SPF.

Material Area, m2 kma,  107 Ds,  1010 m2/h hg, m/h Thickness, m Sorption

Gypsum 100 8.84 1.47 0.55 0.0125 High

board 1.14 0.87 Low
30 8.84 0.67 High
1.14 1.07 Low
Plywood 100 6.84 2.67 0.55 0.03 High
5.57 0.87 Low

Appendix B: Result and Sensitivity Analysis

Figures
SINK SORPTION FACTOR

FIG. B.1 Gypsum board TCPP sorption factor when OC SPF is the emission source (left:
HELS; right: LEHS; 100 and 30 m2 are the area of the gypsum board).

–600
–70

–500 –60
Sorpon Factor (μg/m2/h)

Sorpon Factor (μg/m2/h)

–400 –50

100 m2 (23°C) 100 m2 (23°C) 30 m2 (23°C)

–40
–300 30 m2 (23°C) 100 m2 (40°C) 30 m2 (40°C)
100 m2 (40°C) –30 100 m2 (65°C) 30 m2 (65°C)
–200 30 m2 (40°C)
100 m2 (65°C) –20
–100 30 m2 (65°C)
–10

0 0
1 10 100 500 1 10 100 500
Time (h) Time (h)
 TIAN ET AL., DOI: 10.1520/STP158920150036 193

FIG. B.2 Plywood TCPP sorption factor when OC SPF is the emission source.

–450

–400

–350 H.E.L.S. (23°C)

Sorpon Factor (μg/m2/h)
L.E.H.S. (23°C)
–300
H.E.L.S. (40°C)
–250
L.E.H.S. (40°C)
–200 H.E.L.S. (65°C)
–150 L.E.H.S. (65°C)

–100

–50

0
1 10 100 500
Time (h)

COMPARISONS

FIG. B.3 VOC modeling results using microchamber experimental conditions (no sinks).
HE, high-emission scenario; LE, low-emission scenario.

1,200

1,000
Emission Factor (μg/m2/h)

L.E. (23°C)
800 H.E. (23°C)
L.E. (40°C)
600
H.E. (40°C)

400 L.E. (65°C)

H.E. (65°C)
200

0
1 10 100 500
Time (h)
194 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. B.4 OC SPF TCPP emissions factor measured by the microchamber study.

7,000 23°C (w/skin)

23°C (w/o skin)
6,000 40°C (w/skin)
Emission Factor (μg/m2/h) 40°C (w/o skin)
65°C (w/skin)
5,000

4,000

3,000

2,000

1,000

0
0 5 10 15 20
Time (Days)

References

[1] U.S. EPA, “Design for the Environment: Spray Polyurethane Foam Home,” http://
www.epa.gov/dfe/pubs/projects/spf/spray_polyurethane_foam.html (accessed Febru-
ary 21, 2015).

[2] American Chemistry Council, “Latest Spray Foam Data Shows Growth in Sales,” http://
www.americanchemistry.com/Media/PressReleasesTranscripts/ACC-news-releases/
Latest-Spray-Foam-Data-Shows-Growth-in-Sales.html (accessed February 21, 2015).

[3] Liu, Z., Ye, W., and Little, J. C., “Predicting Emissions of Volatile and Semivolatile
Organic Compounds from Building Materials: A Review,” Build. Environ., Vol. 64, 2013,
pp. 7–25.

[4] Xu, Y., Liu, Z., Park, J., Clausen, P. A., Benning, J. L., and Little, J. C., “Measuring and Pre-
dicting the Emission Rate of Phthalate Plasticizer from Vinyl Flooring in a Specially-
Designed Chamber,” Environ. Sci. Technol., Vol. 46, No. 22, 2012, pp. 12534–12541.

[5] Xu, Y. and Little, J. C., “Predicting Emissions of SVOCs from Polymeric Materials and
Their Interaction with Airborne Particles,” Environ. Sci. Technol., Vol. 40, No. 2, 2006,
pp. 456–461.

[6] Liu, X., Guo, Z., and Roache, N. F., “Experimental Method Development for Estimating
Solid-Phase Diffusion Coefficients and Material/Air Partition Coefficients of SVOCs,”
Atmos. Environ., Vol. 89, 2014, pp. 76–84.

[7] Guo, Z., “A Framework for Modelling Non-Steady-State Concentrations of Semivolatile

Organic Compounds Indoors—I: Emissions from Diffusional Sources and Sorption by
Interior Surfaces,” Indoor Built Environ., Vol. 22, No. 4, 2013, pp. 685–700.
 TIAN ET AL., DOI: 10.1520/STP158920150036 195

[8] Guo, Z., “A Framework for Modelling Non-Steady-State Concentrations of Semivolatile

Organic Compounds Indoors—II. Interactions with Particulate Matter,” Indoor Built Envi-
ron., Vol. 23, No. 1, 2014, pp. 26–43.

[9] U.S. EPA, Simulation Program i-SVOC User’s Guide, EPA Office of Research and Develop-
ment, Research Triangle Park, NC, 2013.

[10] U.S. EPA, Program PARAMS User’s Guide 1.0, EPA Office of Research and Development,
Research Triangle Park, NC, 2004.

[11] Xu, Y. and Zhang, Y., “An Improved Mass Transfer Based Model for Analyzing VOC Emis-
sions from Building Materials,” Atmos. Environ., Vol. 37, No. 18, 2003, pp. 2497–2505.

[12] Xiong, J., Liu, C., and Zhang, Y., “A General Analytical Model for Formaldehyde and VOC
Emission/Sorption in Single-Layer Building Materials and Its Application in Determining
the Characteristic Parameters,” Atmos. Environ., Vol. 47, 2012, pp. 288–294.

[13] Kemmlein, S., Hahn, O., and Jann, O., “Emissions of Organophosphate and Brominated
Flame Retardants from Selected Consumer Products and Building Materials,” Atmos.
Environ., Vol. 37, No. 39–40, 2003, pp. 5485–5493.

[14] Little, J. C., Hodgson, A. T., and Gadgil, A. J., “Modeling Emissions of Volatile Organic
Compounds from New Carpets,” Atmos. Environ., Vol. 28, No. 2, 1994, pp. 227–234.

[15] Silberstein, S., Grot, R. A., Ishiguro, K., and Mulligan, J. L., “Validation of Models for Pre-
dicting Formaldehyde Concentrations in Residences Due to Pressed-Wood Products,”
JAPCA, Vol. 38, No. 11, 1988, pp. 1403–1411.

[16] Dunn, J. E. and Tichenor, B. A., “Compensating for Sink Effects in Emissions Test Cham-
bers by Mathematical Modeling,” Atmos. Environ., Vol. 22, No. 5, 1967, pp. 885–894.

[17] Chang, J. C. S. and Guo, Z., “Modeling of the Fast Organic Emissions from a Wood-Finishing
Product—Floor Wax,” Atmos. Environ. A-Gen., Vol. 26, No. 13, 1992, pp. 2365–2370.

[18] Deng, B. and Kim, C. N., “An Analytical Model for VOCs Emission from Dry Building
Materials,” Atmos. Environ., Vol. 38, No. 8, 2004, pp. 1173–1180.

[19] Huang, H. and Haghighat, F., “Modelling of Volatile Organic Compounds Emission from
Dry Building Materials,” Build. Environ., Vol. 37, No. 11, 2002, pp. 1127–1138.

[20] Yan, W., Zhang, Y., and Wang, X., “Simulation of VOC Emissions from Building Materials
by Using the State-Space Method,” Build. Environ., Vol. 44, No. 3, 2009, pp. 471–478.

[21] Guo, Z., “Improve Our Understanding of Semivolatile Organic Compounds in Buildings,”
Indoor Built Environ., Vol. 23, No. 6, 2014, pp. 769–773.

[22] Liu, C., Liu, Z., Little, J. C., and Zhang, Y., “Convenient, Rapid and Accurate Measurement
of SVOC Emission Characteristics in Experimental Chambers,” PLOS ONE, Vol. 8, No. 8,
2013, p. e72445.

[23] Tichenor, B. A., Guo, Z., Dunn, J. E., Sparks, L. E., and Mason, M. A., “The Interaction of Vapour
Phase Organic Compounds with Indoor Sinks,” Indoor Air, Vol. 1, No. 1, 1991, pp. 23–35.

[24] Van Loy, M. D., Lee, V. C., Gundel, L. A., Daisey, J. M., Sextro, R. G., and Nazaroff, W. W.,
“Dynamic Behavior of Semivolatile Organic Compounds in Indoor Air. 1. Nicotine in a
Stainless Steel Chamber,” Environ. Sci. Technol., Vol. 31, No. 9, 1997, pp. 2554–2561.
196 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

[25] Cox, S. S., Zhao, D., and Little, J. C., “Measuring Partition and Diffusion Coefficients for
Volatile Organic Compounds in Vinyl Flooring,” Atmos. Environ., Vol. 35, No. 22, 2001,
pp. 3823–3830.

[26] Bennett, D. H. and Furtaw, E. J., “Fugacity-Based Indoor Residential Pesticide Fate,”
Environ. Sci. Technol., Vol. 38, No. 7, 2004, pp. 2142–2152.

[27] Shin, H.-M., McKone, T. E., Tulve, N. S., Clifton, M. S., and Bennett, D. H., “Indoor Resi-
dence Times of Semivolatile Organic Compounds: Model Estimation and Field Eval-
uation,” Environ. Sci. Technol., Vol. 47, No. 2, 2012, pp. 859–867.

[28] Standards Council of Canada CAN/ULC-S774-09, Standard Laboratory Guide for the
Determination of Volatile Organic Compound Emissions from Polyurethane Foam,
Standards Council of Canada, Ottawa, ON, 2014.

[29] Xiong, J., Huang, S., and Zhang, Y., “A Novel Method for Measuring the Diffusion, Parti-
tion and Convective Mass Transfer Coefficients of Formaldehyde and VOC in Building
Materials,” PLOS ONE, Vol. 7, No. 11, 2012, p. e49342.

[30] Ni, Y., Kumagai, K., and Yanagisawa, Y., “Measuring Emissions of Organophosphate
Flame Retardants Using a Passive Flux Sampler,” Atmos. Environ., Vol. 41, No. 15, 2007,
pp. 3235–3240.

[31] Sebroski, J. R., “Developing ASTM Standards for Test Specimen Preparation and Measur-
ing Emissions of Volatile and Semi-Volatile Organic Compounds (VOCs and SVOCs)
from Spray Polyurethane Foam (SPF) Insulation Products,” presented at the 2013 Poly-
urethanes Technical Conference, Phoenix, AZ, September 23–25, 2013, American Chem-
istry Council, Washington, DC—unpublished.

[32] Sebroski, J. R., “Research Report for Measuring Emissions from Spray Polyurethane
Foam (SPF) Insulation, presented at 2012 Polyurethanes Technical Conference, Atlanta,
GA, September 24–26, 2012, American Chemistry Council, Washington, DC—unpublished.

[33] Guo, Z., “Review of Indoor Emission Source Models. Part 2. Parameter Estimation,”
Environ. Pollut., Vol. 120, No. 3, 2002, pp. 551–564.

[34] Scheutz, C. and Kjeldsen, P., Determination of the Fraction of Blowing Agent Released
from Refrigerator/Freezer Foam After Decommissioning the Product, Technical Universi-
ty of Denmark, Lyngby, 2002, p. 72.

[35] Zhao, D., Little, J., and Cox, S., “Characterizing Polyurethane Foam as a Sink for or Source
of Volatile Organic Compounds in Indoor Air,” J. Environ. Eng., Vol. 130, No. 9, 2004,
pp. 983–989.

[36] Center for the Polyurethanes Industry, CPI Ventilation Project—Spraying of Polyurethane
Foam (SPF) Formulations Phase 2—Scope of Work Statement for Proposed Air Sam-
pling, American Chemistry Council, Washington, DC, 2013.

[37] Duncan, R., “Introduction to Spray Polyurethane Foam (SPF),” presented at the 2011 ACI
National Conference, San Francisco, CA, March 31, 2011, Spray Polyurethane Foam Alli-
ance, San Francisco, CA—unpublished.

[38] Price, D. M., “Vapor Pressure Determination by Thermogravimetry,” presented at The

Twenty-Seventh Conference of the North American Thermal Analysis Society, September
20–22, 1999, Savannah, GA—unpublished.
 TIAN ET AL., DOI: 10.1520/STP158920150036 197

[39] Verbruggen, E. M. J., Rila, J. P., Traas, T. P., Posthuma-Doodeman, C. J. A. M., and
Posthumus, R., Environmental Risk Limits for Several Phosphate Esters, with Possible
Application as Flame Retardant, National Institute for Public Health and the Environ-
ment, Bilthoven, the Netherlands, 2005.

[40] Agency for Toxic Substances and Disease Registry, Toxicological Profile for Phosphate
Ester Flame Retardants, U.S. Department of Health and Human Services, Atlanta, GA,
2012, pp. 245–251.

[41] U.S. EPA, Estimation Programs Interface Suite for Microsoft Windows, v 4.11, EPA, Wash-
ington, DC, 2015.

[42] Balasubramanian, R. and He, J., “Fate and Transfer of Semivolatile Organic Compounds
in a Multi-Compartment Environment,” in Urban Airborne Particulate Matter Origin,
Chemistry, Fate and Health Impacts, Springer, Berlin, Germany, 2011.

[43] Kamprad, I. and Goss, K.-U., “Systematic Investigation of the Sorption Properties
of Polyurethane Foams for Organic Vapors,” Anal. Chem., Vol. 79, No. 11, 2007,
pp. 4222–4227.

[44] Holmgren, T., Persson, L., Andersson, P. L., and Haglund, P., “A Generic Emission Model
to Predict Release of Organic Substances from Materials in Consumer Goods,” Sci. Total
Environ., Vol. 437, 2012, pp. 306–314.

[45] Hirschfelder, J. O., Curtiss, C. F., and Bird, R. B., The Molecular Theory of Gases and
Liquids, Wiley-Interscience, New York, 1964.

[46] Cussler, E. L., Diffusion: Mass Transfer in Fluid Systems, Cambridge University Press, New
York, 2009.

[47] Sebroski, J., Miller, J. W., Thompson, C. P., and Roeske, E., “Evaluation of Micro-Scale
Chambers for Measuring Chemical Emissions from Spray Polyurethane Foam Insulation,”
Developing Consensus Standards for Measuring Chemical Emissions from Spray Polyure-
thane Foam (SPF) Insulation, ASTM STP1589, J. Sebroski and M. Mason, Eds., ASTM Inter-
national, West Conshohocken, PA, 2017, pp. 1–26.

[48] Zhao, D. Y., Cox, S. S., and Little, J. C., “Source/Sink Characterization of Diffusion-
Controlled Building Materials, presented at The Eighth International Conference on
Indoor Air Quality and Climate, Edinburgh, Scotland, August 8–13, 1999, Air Infiltration
and Ventilation Centre—unpublished.

[49] Gavezzotti, A., “Molecular Free Surface: A Novel Method of Calculation and Its Uses in
Conformational Studies and in Organic Crystal Chemistry,” J. Am. Chem. Soc., Vol. 107,
No. 4, 1985, pp. 962–967.

[50] Molinspiration, “Calculation of Molecular Properties and Bioactivity Score,” http://

www.molinspiration.com/cgi-bin/properties (accessed March 12, 2015).

[51] Misra, K. C., Introduction to Geochemistry: Principles and Applications, Wiley, Hoboken,
NJ, 2012, p. 640.

[52] Begley, T., Castle, L., Feigenbaum, A., Franz, R., Hinrichs, K., Lickly, T., Mercea, P., Milana,
M., O’Brien, A., Rebre, S., Rijk, R., and Piringer, O., “Evaluation of Migration Models that
Might Be Used in Support of Regulations for Food-Contact Plastics,” Food. Addit. Con-
tam., Vol. 22, No. 1, 2005, pp. 73–90.
198 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

[53] Mercea, P., “Models for Diffusion in Polymers,” in Plastic Packaging: Interactions with
Food and Pharmaceuticals, 2nd Ed., Wiley-VCH, Weinheim, Germany, 2008, pp. 123–162.

[54] Radonic, J., Miloradov, M., Turk-Sekulic, M., Kiurski, J., Djogo, M., and Milovanovic, D.,
“The Octanol–Air Partition Coefficient, KOA, as a Predictor of Gas–Particle Partitioning of
Polycyclic Aromatic Hydrocarbons and Polychlorinated Biphenyls at Industrial and
Urban Sites,” J. Serb. Chem. Soc., Vol. 76, No. 3, 2011, pp. 447–458.

[55] Liu, C., Shi, S., Weschler, C., Zhao, B., and Zhang, Y., “Analysis of the Dynamic Interaction
Between SVOCs and Airborne Particles,” Aerosol. Sci. Tech., Vol. 47, No. 2, 2012,
pp. 125–136.

[56] Levin, V. A., “Relationship of Octanol/Water Partition Coefficient and Molecular Weight
to Rat Brain Capillary Permeability,” J. Med. Chem., Vol. 23, No. 6, 1980, pp. 682–684.

[57] Theis, A. L., Waldack, A. J., Hansen, S. M., and Jeannot, M. A., “Headspace Solvent Micro-
extraction,” Anal. Chem., Vol. 73, No. 23, 2001, pp. 5651–5654.

[58] Li, W. and Davis, E. J., “Aerosol Evaporation in the Transition Regime,” Aerosol. Sci.
Tech., Vol. 25, No. 1, 1996, pp. 11–21.

[59] ARChem, “SPARC Performs Automated Reasoning in Chemistry,” http://archemcalc.-

com/sparc.html 2015 (accessed March 14, 2015).

[60] The National Ambient Air Quality Standards for Particle Pollution, “Revised Air Quality
Standards for Particle Pollution and Updates to the Air Quanlity Index (AQI),” https://
www3.epa.gov/airquality/particlepollution/2012/decfsstandards.pdf (accessed May 24,
2015).
 DEVELOPING CONSENSUS STANDARDS FOR MEASURING CHEMICAL EMISSIONS 199

STP 1589, 2017 / available online at www.astm.org / doi: 10.1520/STP158920150045

Charles Bevington,1 Zhishi Guo,2 Tao Hong,2

Heidi Hubbard,2 Eva Wong,1 Katherine Sleasman,1 and
Carol Hetfield1

A Modeling Approach for

Quantifying Exposures from
Emissions of Spray Polyurethane
Foam Insulation in Indoor
Environments
Citation
Bevington, C., Guo, Z., Hong, T., Hubbard, H., Wong, E., Sleasman, K., and Hetfield, C.,
“A Modeling Approach for Quantifying Exposures from Emissions of Spray Polyurethane
Foam Insulation in Indoor Environments,” Developing Consensus Standards for
Measuring Chemical Emissions from Spray Polyurethane Foam (SPF) Insulation, ASTM
STP1589, J. Sebroski and M. Mason, Eds., ASTM International, West Conshohocken, PA,
2017, pp. 199–227, http://dx.doi.org/10.1520/STP1589201500453

ABSTRACT
A range of chemicals from very volatile to semivolatile organic compounds are
emitted from spray polyurethane foam (SPF) insulation. SPF installation
procedures and environments can vary widely, and the emissions, transport, and
fate of these chemicals in the indoor environment after SPF installation are not
well characterized. To begin to understand exposure to emissions from SPF and
to identify and characterize the uncertainty in assessing chemical exposures, a
proof-of-concept multizone indoor model was developed to estimate indoor air
concentrations of chemicals emitted from SPF over time. The model supported
the development of different approaches for characterizing the emissions of
volatile and semivolatile organic compounds and for predicting short- and long-
term emissions and subsequent air concentrations. It also incorporated estimates

Manuscript received May 15, 2015; accepted for publication September 29, 2015.
1
U.S. Environmental Protection Agency, 1200 Pennsylvania Ave. NW, Washington, DC 20004
2
ICF International, 2635 Meridian Pkwy. #200, Durham, NC 27713
3
ASTM Symposium on Developing Consensus Standards for Measuring Chemical Emissions from Spray
Polyurethane Foam (SPF) Insulation on April 30–May 1, 2015 in Anaheim, CA.

Copyright V
C 2017 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959.
200 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

for a wide range of parameters that influence emission and subsequent exposure
from SPF. A sensitivity analysis was performed to explore the impact of model
inputs, including those considered the most influential and those for which there is
uncertainty because of a lack of data. Model inputs included the location and type
of SPF foam; the chemical-specific diffusion and partitioning coefficients; the
temperature and ventilation rates of zones within the residence; and the impact of
engineering controls, such as increased ventilation during installation. Sensitivity
analysis results identified trends and relations between model inputs and outputs.
Additional experimental data are needed to calibrate the model and to reduce
uncertainty of model estimates. In particular, information is needed to characterize
emissions within the first 24 h after spray application, to characterize longer-term
mass transfer of chemicals from SPF, and to describe interzonal air flow and
leakage rates between attics, living spaces, and crawl spaces.

Keywords
spray polyurethane foam, SPF, SVOCs, VOCs, indoor exposure

Introduction
Spray polyurethane foam (SPF) is a versatile and effective insulation and sealant
material. To create the final SPF product, chemical reactions of the two component
parts commonly referred to as Sides A and B must occur. In general, SPF products
consist of approximately 50 % Side A and 50 % Side B. Side A contains reactive
chemicals known as isocyanates, and Side B contains a mixture of chemicals, in-
cluding polyols (which react with isocyanates), catalysts, flame retardants, blowing
agents, processing aids, and surfactants [1].
SPF products vary in their composition, structure, and performance [1]. For in-
stance, there is much variation in the initial concentration of individual chemicals
within the SPF components and therefore in the resulting installed SPF insulation
product. In addition, the mass present in each component before spray application
may differ from the mass present after spray application and subsequent curing due
to chemical reactions that occur during application. However, these differences are
not well characterized.
Foam insulation and its chemical constituents can potentially represent a large
exposure source based on the mass of product within a building that is on the order
of several hundred kilograms in a given residence. Depending on where SPF is used
as well as state or local building codes and requirements, SPF insulation may re-
main exposed to open air or may be covered with drywall, barriers, or coatings
depending on its location and whether a thermal or ignition barrier is needed [1].
More common SPF applications include the insulation of attic spaces, crawl spaces,
and exterior walls. However, residences may be insulated in any of the previously
mentioned applications as well as many additional applications such as garage
 BEVINGTON ET AL., DOI 10.1520/STP158920150045 201

ceilings, exposed ductwork, soundproofing below flooring or interior walls, and in

cracks and openings surrounding windows and doors [1].
SPF is sprayed on site under an elevated temperature and pressure. After appli-
cation, the SPF product cures or hardens and may also need to be trimmed. Trim-
ming of excess SPF, particularly open-cell foam, may generate suspended
particulates that can vary in size and quantity based on the type of equipment used.
One study reported short-term airborne concentrations above occupational limits
for inhalable and respirable particulates not otherwise specified [2]. The chemical
composition of these particulates, their ultimate fate, and their overall impact on
exposures is not well characterized.
Potential exposure to isocyanates, a key ingredient in SPF, is highly variable
based on the conditions surrounding application. Short-term elevated concentra-
tions of diisocyanates after spray application of SPF have been well documented for
years [3]. Personal and area industrial hygiene air samples show that concentrations
of an airborne methylene diphenyl diisocyanate (MDI) monomer, a common isocy-
anate, can peak and remain elevated above workplace exposure limits for a short
period of time and that area air samples in other locations can also remain elevated
in the short term, although they are generally lower than personal samples [4–7].
The availability and concentrations of isocyanates is related to the completeness of
the chemical reactions and the vapor pressures of the individual compounds. The
lower-molecular-weight isocyanates pose a vapor inhalation hazard, and higher-
molecular-weight isocyanates pose an inhalation hazard if aerosolized (sprayed) or
heated in the work environment [8]. Several researchers have recommended estab-
lishing a work zone to minimize the exposure of unprotected individuals during ap-
plication. This is supported by data that show a reduction of MDI concentrations
with both distance from and time after the application [5,9]. However, sampling
and analysis of isocyanates, including monomers, prepolymers, polyisocyanates,
and oligomers, present challenges because these chemicals are highly reactive and
have strong adhesion properties. To better understand long-term isocyanate expo-
sure, better sampling and analysis methods are required [10].
Compared to short-term emissions, longer-term emissions of volatile organic
compounds (VOCs) and semivolatile organic compounds (SVOCs) from SPF over
months to years are not as well characterized [11]. Given the wide range in product
formulations and environmental conditions during and after application, there are
many factors that could influence overall magnitude and trends of multiple chemi-
cal compounds from SPF over time. One study reported headspace concentrations
of 20 chemical substances under elevated temperature and relative humidity from
closed-cell foam insulation several days after application [12]. The exact number of
compounds emitted from various kinds of SPF under different conditions over
short and longer timeframes is not well characterized.
Modeling exposure to SPF presents unique challenges because of the high vari-
ability among SPF foam, components, and installation and use conditions coupled
with a lack of monitoring data. Still, exposure models can be useful tools for better
202 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

understanding exposures related to emissions of VOCs and SVOCs from SPF and
can account for some aspects of variability of SPF emissions over time. This paper
introduces a proof-of-concept multizone indoor exposure model for estimating in-
door air concentrations of compounds emitted from SPF that considers the wide
variety of site-specific applications of SPF insulation. For this proof-of-concept
model, a small number of representative chemicals were selected to illustrate trends
of SVOCs and VOCs over both short- and long-term timeframes.

Methodology
CONCEPTUAL MODEL
To quantify potential exposures from SPF insulation, a conceptual modeling frame-
work was developed. This proof-of-concept modeling framework is illustrated in
Fig. 1 and includes many, but not all, of the processes that influence the migration
and transport of VOCs and SVOCs from SPF over time. The modeling framework
and input parameters selected were intended to approximate a generic residential
setting. For the proof-of-concept model, this residential setting should be

FIG. 1 Conceptual model for SPF exposures.

 BEVINGTON ET AL., DOI 10.1520/STP158920150045 203

considered hypothetical. Further parameterization, using data from a variety of resi-

dences, would better approximate residential settings.
The model currently considers open-cell low-density and closed-cell medium-
density foams. It has not yet been parameterized for low-pressure SPF kits, one-
component foams, or roofing foams. The model has characterized emissions from
1,1,1,3,3-pentafluoropropane (HFC-245fa), bis(2-dimethylaminoethyl) ether
(BDMAEE), and tris(1-chloro-2-propyl) phosphate (TCPP). Other chemicals such
as MDI and other isocyanates; 1,4-dioxane; aldehydes, including formaldehyde and
acetaldehyde; tris(2-chloroethyl)phosphate and triethyl phosphate (TEP); and
chlorobenzenes can be considered should short- or long-term emissions data be-
come available. Preliminary model runs to date have not included chemical reac-
tions over time, and the model has not been parameterized to incorporate the
impact of work practices to reduce exposures such as high-efficiency particulate air
vacuuming, filtration, or containment.
SPF can be considered as having two phases: the installation phase, in which the
SPF is “wet” or uncured, and the postinstallation phase, in which the SPF is sufficient-
ly cured or “dry.” Closed-cell foam dries to a semi-rigid state, whereas open-cell foam
remains soft. The A- and B-side components are fed through heated hoses to a spray
gun, where they are mixed and sprayed under pressure, rapidly reacting and expand-
ing in volume to form polyurethane foam. During the installation phase, emissions
are dominated by isocyanate and other VOCs in the form of vapors, aerosols, and
mists from overspray of the SPF ingredients. Emissions may also result from dust
and particulates while trimming or sanding foam. After curing, the SPF insulation is
expected to become a diffusional source that slowly emits SVOCs and residual VOCs
over time. To evaluate the impact of SPF installed adjacent to the living area, such as
an attic, crawlspace, or within the wall cavity, a multizone model was required.
There are currently no existing simulation tools that address diffusional sour-
ces, such as SPF, in a multizone environment. To fill this gap, the proof-of-concept
model combined code from the U.S. Environmental Protection Agency’s Indoor
Air Quality and Inhalation Exposure (IAQX) model for multizone environments
and application-phase simulation with code from the EPA’s i-SVOC model for dif-
fusional sources. Documentation for both programs can be downloaded from the
EPA website [13,14]. To the extent data are available, the model is able to account
for the following:
1. Emissions from exposed SPF to the bulk air
2. Transport between zones and the outdoors through natural and forced
(HVAC) ventilation
3. Molecular diffusion through a barrier material between the SPF and the bulk
air (e.g., gypsum board)
4. Gas-phase chemical reactions within the indoor environment that generate re-
action byproducts
5. Partitioning from the bulk air to indoor particulates, e.g., particulate matter,
total suspended particles, and dust
204 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

6. Partitioning from the bulk air to indoor sinks, such as building materials and
furnishings

The following subsections describe each of these processes in detail; the conceptual
model is shown in Fig. 1.

EMISSIONS FROM SPF

The rate and magnitude of emissions depend on both the physical and chemical
characteristics of SPF and the conditions of the environment. Ensuring the proper
1:1 ratio of A- and B-side chemicals, according to manufacturers’ specifications, is
important in preventing the production of “off-ratio” products [1,15]. Off-ratio
foam may result in different emissions over time compared to on-ratio foam, but
additional data are needed [15]. Relative humidity and temperature are also impor-
tant factors in the production of foam [1,15]. Foam applied too thickly may result
in an exothermic reaction and produce persistent chemical odors after installation
[1,15]. If the SPF product is not heated to the manufacturer’s recommended tem-
perature for application, it may not react fully because of increased viscosity and
poor mixing, resulting in poor foam quality [15]. However, if the chemicals are too
hot, there may be more overspray [15]. The product temperature at the gun affects the
spray pattern and foam texture and quality [1,15]. In addition, the density, porosity,
and cell type will affect the rate that chemicals can be emitted from the foam, and the
overall mass of foam in the residence will affect the magnitude of the emissions.
Generic SPF formulations intended to represent two of the more commonly
used SPF insulation products currently in the marketplace were developed and are
presented in Table 1 [16]. Representative open- and closed-cell foams were selected.
Open versus closed cell structure affects the density and subsequent mass of foam
and associated chemical components for a given volume of applied foam. A volatile
blowing agent (HFC-245fa), an amine catalyst (BDMAEE), and a semivolatile flame

TABLE 1 SPF parameters.

Characteristics Low-density (1=2-lb) High-pressure Medium-density (1=2-lb) High-pressure

SPF Formulationa,b SPF Formulationa,b

Open- or closed-cell Open Closed

foam
Density, g/cm3 0.008 0.032
Side A compositiona MDI/pMDI (100 %) MDI/pMDI (100 %)
Side B compositiona Polyol; nonylphenol ethoxylate Polyol; HFC-245fa (6.97 %)b; water
emulsifier; water (20 %); TCPP (2.55%); TCPP (15.91 %)b; silicone
b
(25.2 %) ; silicone surfactant; BDMAEE surfactant; BDMAEE (0.7 %)b;
b
(0.9 %) ; tetramethyliminobispropylamine; bis(dimethylaminopropyl) methylamine;
N,N,N-trimethylaminoethylethanolamine N,N,N-trimethylaminoethylethanolamine
a
Generic formulations.
b
Selected for proof-of-concept modeling.
 BEVINGTON ET AL., DOI 10.1520/STP158920150045 205

retardant (TCPP) were chosen as representative chemicals for the proof-of-concept

modeling.
For the purposes of the proof-of-concept model, the initial concentrations of
“inert” chemicals were based solely on the generic formulations. The initial concen-
tration of a chemical in the original formulation and in the applied foam may vary
based on several factors, including potential losses due to overspray and chemical
reactions that occur during the spray that may change the composition of the ap-
plied foam. The most notable of these reactions is the reaction of water with isocya-
nates to form amines and carbon dioxide. All of the water in medium-density
closed-cell foam is likely to react, whereas approximately one-third of the water in
low-density open-cell foam is expected to react [17]. In addition, it is possible that
the foam may not be perfectly applied in a 50:50 ratio between Side A and Side B.
The measurement of “initial” concentrations of chemicals within the foam after
spray application would reduce model uncertainty.

Short-Term Emission Parameters

HFC-245fa is a component of SPF that acts as a blowing agent, that is, a chemical
that is designed to aid in the dispersion of SPF during application. As such, HFC-
245fa is designed to have a short life span in the foam. Therefore, this chemical was
chosen as the representative chemical for understanding short-term emissions.
In the proof-of-concept model, short-term emissions use an emission factor (E)
and a first-order decay constant (k) derived from the results of chamber testing.
Based on its volatility, HFC-245fa is expected to follow a first-order decay [18] as
shown in Eq 1:

E ¼ M0 k ekt (1)

where:
E ¼ emission factor at time t, lg/m2/h,
M0 ¼ initial concentration in unit area of foam available for rapid emissions,
lg/m,2 and
k ¼ first-order decay rate constant, h1.
The baseline value for M0 was based on product formulation and SPF thickness.
Given the extremely high vapor pressure of this chemical the decay constant k is
expected to be large (1 h1 or higher).

Long-Term Emission Parameters

Long-term emissions from SPF were estimated for two representative chemicals:
BDMAEE, an amine catalyst, and TCPP, a flame retardant. After application, the
SPF cures and becomes a diffusional source that can be represented by the modified
state-space method [19]. Diffusional sources are characterized by their surface area
and thickness. Because SPF insulation is thicker than most other indoor sources,
the SPF model used 15 state-space layers instead of 10 in program i-SVOC. Key
206 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

parameters governing emissions include the initial concentration of chemical in

SPF (C0), material-air partition coefficient of chemical (K), and the diffusion coeffi-
cient of the chemical in SPF (D) as shown in the following equations:
 
CSPF
E ¼ Ha C (2)
K
Ha ¼ f ðh; K; DÞ (3)

where:
E ¼ emission factor, lg/m2/h,
Ha ¼ overall gas-phase mass transfer coefficient, m/h,
CSPF ¼ concentration in the surface layer of SPF, lg/m3,
C ¼ concentration of chemical in air, lg/m3,
K ¼ material-air partition coefficient (dimensionless),
h ¼ gas-phase mass transfer coefficient, m/h, and
D ¼ solid-phase diffusion coefficient, m2/h.
The mass transfer, diffusion, and partition coefficients depend on both the SPF and
the chemical of interest as well as temperature. None has been measured directly
for the combinations of chemicals of interest and SPF insulation in this modeling
exercise; therefore, for the proof-of-concept modeling, they were estimated using
quantitative structure-activity relationship (QSAR) models.

MASS TRANSFER COEFFICIENTS

The gas-phase mass transfer coefficient was calculated with EPA’s PARAMS [20]
software with the use of the Sherwood method [21], which is based on correlations
between three dimensionless numbers: the Sherwood number (Sh), the Schmidt
number (Sc), and the Reynolds number (Re). Detailed explanations of the method-
ology is provided in the PARAMS user’s guide [20]. The methods for calculating
the solid-phase and overall mass transfer coefficients are described in [19].

PARTITION COEFFICIENTS
Partition coefficient data are available for VOCs in polyurethane foam used for air sam-
pling purposes [22,23]. However, data for SVOC and SPF are currently not available. The
partition coefficient (K) for SPF depends on properties of the foam (e.g., foam type,
density/porosity, and temperature) and those of the chemical (e.g., vapor pressure or
octanol-air partition coefficient). In the absence of experimental data, the partitioning coef-
ficient for solid (neat) polyurethane polymer was estimated based on an empirical relation
with the vapor pressure of the chemical of interest as presented in [24] and shown below:

ln K ¼ 8:76  0:7851 ln P (4)

where:
K ¼ material-air partition coefficient (dimensionless) and
P ¼ vapor pressure, mm Hg.
 BEVINGTON ET AL., DOI 10.1520/STP158920150045 207

This K value was then adjusted by the density of the SPF as follows:

dSPF
KSPF ¼  KSP (5)
d0

where:
KSPF ¼ partition coefficient for SPF with density of dSPF (dimensionless),
dSPF ¼ density of the SPF product, g/cm3,
d0 ¼ density of common building materials (g/cm3), assuming d0 ¼ 1, and
KSP ¼ partition coefficient for solid polyurethane polymer obtained from the
correlation above (dimensionless).

Partitioning is also temperature-dependent; K can be calculated for a range of tem-

peratures using the following relation from Zhang et al. [25]:
 
pffiffiffiffi A2
K ¼ A1 T exp (6)
T

where:
T ¼ temperature, K,
A1 ¼ empirical constant, assuming A1 ¼ 3.22  1018 based on TCPP vapor
pressure data, and
A2 ¼ empirical constant, assuming A2 ¼ 1.58  104 based on TCPP vapor
pressure data.

DIFFUSION COEFFICIENTS
Experimentally determined diffusion coefficients for specific chemicals in SPF are
scarce [26]. Some data are available for polyurethane foam for air sampling pur-
poses [22,27]. The extrapolated diffusion coefficients for SPF based on the previous-
ly presented information varied greatly. A professional judgment was made based
on the available information, including the EPA’s recent research on polychlori-
nated biphenyls (PCBs) [28,29]. For TCPP at 25 C, the following values were used:
D ¼ 5  1013 m2/h for closed-cell medium-density SPF and D ¼ 1  1010 m2/h
for open-cell low-density SPF.
Similar to partitioning, diffusion is also temperature-dependent. The diffusion
coefficient was adjusted for varying temperature using the following equation
from [30]:
 
1 1
lnðD2 Þ ¼ lnðD1 Þ  10450   (7)
T2 T1

where:
T ¼ temperature, K, and
D ¼ diffusion coefficient, m2/h.
208 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

CHEMICAL TRANSPORT IN THE BULK AIR

A simplified three-zone model was developed for the proof-of-concept model.
Although the purpose of the proof-of-concept model was to demonstrate the feasi-
bility of using a multizone model to evaluate SPF emissions under assumed realistic
building conditions, information related to these parameters is highly fragmented
in the literature and highly variable because of the influence of the following
factors:
• Seasonal temperature differences between the zones
• Diurnal temperature differences between the zones
• Exterior meteorological conditions (temperature, wind speed)
• Type, location, and operation of HVAC system (on/off cycles)

Because of both the variability and uncertainty, it is challenging to define a single set of
conditions that is universally applicable for evaluating emissions from SPF. Accepting
these limitations, a simplified three-compartment box model was developed that repre-
sented three main areas of a residence: the living space, the attic, and the crawl space or
basement. Possible representative interzonal air flows are depicted in Fig. 2, and the air
flow patterns for the three zone models are presented in Fig. 3.
Typical ventilation rates for the three zones were obtained from the literature
as cited in Table 2. Two flow rates were particularly underrepresented in the litera-
ture: the air leakage rate from attic to living space and the air leakage rate from the
crawl space to living space. As such, a value of 15 m3/h was used for the air flow
from the vented attic/crawl space to the living area and 5 m3/h for the air flow from
the unvented attic/crawl space to the living area. These values are generally smaller

FIG. 2 Possible interzonal air flows between the attic, living area, and crawl space.
 BEVINGTON ET AL., DOI 10.1520/STP158920150045 209

FIG. 3 Baseline air flows for the three-zone model. The ventilation rate for the living area is
0.5/h; for the vented and unvented attic the ventilation rates are 2 and 0.7/h,
respectively; and for the vented and unvented crawl space, the ventilation rates are
1 and 0.35/h, respectively. The air flows between unconditioned and conditioned
spaces can go in either direction depending on the season (i.e., Q12/Q21 and Q13/Q31).
Time-varying interzonal air flows were calculated based on the temperature
difference between zones using empirical models. Note that the HVAC can be
located in either the attic or crawl space but not both.

and more conservative than the values found in the cited literature [31–33]
and were chosen based on two considerations: (1) air flows can go in either
direction depending on the season (and climate), and (2) when using constant
values for these air flows, the proposed values are more like time-averaged inter-
zonal flows.
Temperature affects the rate of release of chemicals from the SPF insulation;
temperature gradients between zones affect the direction and magnitude of air flow
between zones. There is not enough information at present to characterize the effect
of temperature on emissions from installed SPF.

CHEMICAL TRANSPORT BY MOLECULAR DIFFUSION AND CONVECTIVE MASS

TRANSFER THROUGH WALLS
In addition to transport due to interzonal air flows, VOCs and SVOCs can also be
transported throughout the residence by solid-phase molecular diffusion through walls
210 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

(ceiling and flooring as well) and by convective mass transfer through wall cracks.
The molecular diffusion mechanism is independent of air leakage rates through the
walls but dependent on characteristics of the wall and the microclimate around the resi-
dence. Both mechanisms can be modeled. The molecular diffusion process can be ap-
proximated by the two-layer source model (e.g., SPF þ gypsum board) as described in
[14]. The convective mass transfer through wall cracks is similar to the mass transfer
between the attic/crawl space and living area, as described previously. A preliminary
evaluation suggested that molecular diffusion is much less effective than the convective
mass transfer through wall cracks. Because of the large uncertainties associated with
these mass transfer mechanisms, they were not included in the exposure scenarios pre-
sented in the sections that follow.

REACTION CHEMISTRY
Chemically reactive components of SPF may react with other chemicals after be-
ing emitted. For example, MDI and other isocyanates react with water vapor to
form amines and carbon dioxide. As such, the relative humidity can influence
the rate of this reaction. The proposed model allows for the inclusion of gas-
phase chemical reactions and heterogeneous reactions on interior surfaces. How-
ever, for the representative chemicals presented herein, there were not adequate
data to characterize the rates of reaction, and this mechanism was not further
investigated. The current model does not consider chemical reactions within the
SPF insulation.

LOSSES TO INDOOR PARTICULATES

Indoor particulates can be considered as diffusive sinks, that is, behaving analo-
gously to diffusive SVOC sources, with the concentration gradient and chemical
transport occurring in the opposite direction (from the bulk air to the particulates).
A methodology for estimating instantaneous partitioning was put forth by
Weschler and Nazaroff [34]. A methodology for dynamic transfer (based on mass
transfer, diffusion, and partitioning) was put forth by Shi and Zhao [35]. Both
approaches were incorporated during the exploratory modeling phase and were not
found to affect air concentrations appreciably. Therefore, for the proof-of-concept
modeling, this mechanism was not incorporated for estimating air concentrations.
However, modeling airborne and settled dust concentrations is potentially impor-
tant from an exposure standpoint and can be estimated by the proof-of-concept
model.

LOSSES TO INDOOR SINKS

In this model, the deposition of chemically reactive species is treated as first order,
resulting in a deposition rate (lg/h) that depends on the airborne concentration,
volume of the zone, and the deposition rate constant as shown in Eq 8:

d ¼ V kd C (8)
 BEVINGTON ET AL., DOI 10.1520/STP158920150045 211

TABLE 2 Model parameters for the SPF proof-of-concept model.

Parameter Value Unit

BLDG
Volume
Living areaa 300 m3
a
Attic 150 m3
a
Crawl space 150 m3
Air exchange rate
Living areaa 0.5 h1
Attic (vented) b
2.0 h1
Attic (unvented) c
0.7 h1
Crawl space (vented) d
1 h1
Crawl space (unvented) d
0.35 h1
Air leakage
From vented attic to living spacee 15 m3/h
e
From unvented attic to living space 5 m3/h
f
From vented crawl space to living space 15 m3/h
f
From unvented crawl space to living space 5 m3/h
Air speed
Living area 0.1 m/s
Attic (vented) 0.1 m/s
Attic (unvented) 0.05 m/s
Crawl space (vented) 0.1 m/s
Crawl space (unvented) 0.05 m/s
Exposed gypsum board surface areaa 800 m2
a
Exposed gypsum board thickness 0.01 m
SPF
Exposed surface area
Attic (vented)a 180 m2
a
Attic (unvented) 400 m2
a
Crawl space 120 m2
a
Thickness 0.1 m
Density
Open cell (low density)g 0.008 g/cm3
g
Closed cell (medium density) 0.032 g/cm3
Initial concentrations
HFC-245fa: closed cell (medium density)g 1.12  109 lg/m3
g
HFC-245fa: open cell (low density) NA lg/m3
g 8
BDMAEE: closed cell (medium density) 1.12  10 lg/m3
g 7
BDMAEE: open cell (low density) 3.6  10 lg/m3
g 9
TCPP: closed cell (medium density) 2.55  10 lg/m3
g 9
TCPP: open cell (low density) 1.01  10 lg/m3
212 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 2 (Continued)

Parameter Value Unit

STE
First-order decay constant of HFC-245fah 1 h1
LTE
Partition coefficient of TCPP at 25  Ch Low-density open cell: 3.06  105; Unitless
medium-density closed cell:
1.23  106
 h
Partition coefficient of BDMAEE at 25 C Low-density open cell: 6.50  101; Unitless
medium-density closed cell:
2.60  102
Diffusion coefficient of TCPPh Low-density open cell: 1  1010; m2/h
medium-density closed cell
5  1013
Diffusion coefficient of BDMAEE h
Low-density open cell: 1.8  1010; m2/h
medium-density closed cell:
9.3  1011
Gas-phase mass transfer coefficient of TCPPh 0.4 m/h
Gas-phase mass transfer coefficient of BDMAEEh 0.43 m/h
Partitioning coefficient for TCPP and gypsum boardh 3.95  107 Unitless
Diffusion coefficient for TCPP and gypsum boardh 7.54  1011 Unitless
a
Based on [36].
b
Based on a ventilation rate range of 0.9 to 4.6 from five studies [37].
c
Based on 65 % reduction.
d
Based on [38].
e
Walker et al. [39] give an interior attic exchange rate for attics of approximately 0.1 h1 when the
interior attic temperature difference ranges between approximately 3 and 20 C.
f
Based on a 45-home study [40]. A leakage flow rate through the house floor was 22 m3/h when
DP ¼ 4 Pa.
g
Calculated from generic formulations.
h
Estimated as discussed in “Emissions from SPF” section.

where:
d ¼ deposition rate, lg/h,
C ¼ air concentration, lg/m3,
V ¼ zone volume, m3, and
kd ¼ first-order deposition rate constant, h1.

Indoor materials can also serve as diffusive sinks. As such, they are treated similarly
to diffusive sources with an initial concentration of zero. Thus, material surface
area and diffusion, partition, and mass transfer coefficients are required. In the
proof-of-concept modeling, gypsum wall board was considered as an indoor sink.
As with the diffusional source, the mass transfer coefficient was calculated using the
Sherwood method in PARAMS [20]. The diffusion and partitioning coefficients
were estimated using QSAR models based on experimental data for PCB-52 diffu-
sion and partitioning in gypsum wall board [28].
 BEVINGTON ET AL., DOI 10.1520/STP158920150045 213

GOVERNING MASS BALANCE EQUATION

Combining all processes governing short- and long-term emissions, chemical trans-
port by bulk air and molecular diffusion, and losses to indoor sinks, the resulting
mass balance on a chemical of interest used to estimate time series of indoor con-
centrations is as follows:

dCi Xn1 Xn2 Xn3 Xn4

Vi ¼ AE 
j¼1 j j
Q C þ
k¼0 ik i
Q C 
k¼0 ki k
d
m¼1 m
dt
Xn5 Xn6
 r 6Vi
p¼1 p
x
q¼1 q
(9)

where:
Vi ¼ volume of zone i, m3,
Ci ¼ air concentration in zone i, lg/m3,
t ¼ time, h,
Aj ¼ area of source j in zone i, m2,
Ej ¼ emission factor for source j in zone i, lg/m2/h,
Qik ¼ air flow from zone i to zone k, i = k, m3/h,
Qki ¼ air flow from zone k to zone i, k = i, m3/h,
Ck ¼ air concentration in zone k, lg/m3,
dm ¼ sorption rate onto interior surface m in zone i, lg/h,
rp ¼ rate of sorption by particulate matter p in zone i, lg/h,
xq ¼ rate of gas-phase chemical reaction q in zone i (plus sign for products and
minus sign for reagents),
j, k, l, m, p, and q are summation counters, and
n1 through n6 are item numbers for their respective summations.

Note that when chemical reactions are involved, the mass unit must be in either
moles or molecules. For example, term x in Eq 9 should be either mol/m3/h or
molecules/m3/h.

MODEL INPUT PARAMETERS

To run the simulations, a variety of input parameters were required that can be or-
ganized into three sets of inputs:
• Those used to characterize the exposure environment, such as the building config-
uration, airflows, and temperature profile of given zones (BLDG in Table 2)
• Those used to characterize the SPF itself (SPF in Table 2)
• Those used to characterize chemical components of the SPF and associated
emission, either short or long term (STE or LTE in Table 2)

The input parameters used to build the proof-of-concept model are presented in
Table 2. In some cases, the values have large uncertainty due to a lack of experi-
mental data. The values put forth herein are not recommended default values and
can be updated should additional data sources be available. It is important to con-
sider this uncertainty when interpreting the proof-of-concept model and results.
214 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 3 Parameters for SPF sensitivity analysis.

Range

Parameter Symbol Baseline Low High

SPF thickness, m H 0.1 0.05 0.15

TCPP partition coefficient (dimensionless) K 2  106 5  105 5  107
TCPP diffusion coefficient, m2/h D 1  1011 1  1014 1  109
Gas-phase mass transfer coefficient, m/h h 0.4 0.2 2
Initial concentration, lg/m3 C0 1  109 5  108 5  109
Leakage from attic to living space, m3/h Q21 15 5 30
Air exchange rate in living area, h1 N1 0.5 0.35 2
Air exchange rate in attic, h1 N2 1 0.2 6

Input parameters that had either greater uncertainty or variability were explored in
a sensitivity analysis and are shown in Table 3.

APPROACH FOR SENSITIVITY ANALYSIS

Sensitivity analysis is a useful tool that helps improve our understanding of the rela-
tion between input and output variables of this model, identify key parameters, and
gain insight into ways to effectively reduce the uncertainty in the output. Two
approaches were used for the sensitivity analysis. A one-at-a-time (OAT) sensitivity
analysis varied one parameter at a time while keeping other parameters at baseline
values. This is the simplest and most commonly used method for sensitivity analy-
sis. This method provides direct and intuitive images of how a given input parame-
ter affects the output. As a disadvantage, this method does not consider the
interactions between input variables.
Scatter plots are x-y plots of the output variable(s) against individual input vari-
ables. The data are generated by running a large number of simulations based on
randomly selected independent variables, assuming no correlations existed between
input variables. Unlike the OAT method, in which all but one input parameter are
kept fixed, the scatter plots use arbitrarily picked input variables from either a range
or distribution. The plots give direct visualization of the sensitivity. As a disadvan-
tage, this method is computationally intensive. In this exercise we ran 500 simula-
tions for each case.
A scatter plot analysis was conducted for two key variables: the partitioning
coefficient and the diffusion coefficient for TCPP. The parameters are randomly
picked within the predefined ranges or distributions. A parameter (independent
variable) is more important than others if it imparts more “shape” on the depen-
dent variable, which, in our case, is the air concentration. Scatter plots can reveal
the interactions between independent variables. The results can be further analyzed
by other methods such as nonlinear regression.
 BEVINGTON ET AL., DOI 10.1520/STP158920150045 215

Results
For the proof-of-concept modeling, the intent was to explore major trends and to
identify variables that affect estimated air concentrations resulting from SPF foam.
For this, the following variables were specifically explored:
1. The type of chemical, by performance function, used in the SPF (HFC-245fa,
BDMAEE, and TCPP chosen as representative chemicals)
2. The type of SPF foam, either medium-density closed cell or low-density open cell
3. The location of the SPF, either in the attic or crawl space
4. The impact of vented versus unvented attics and crawl spaces
5. The location of the HVAC system, either in the attic or crawl space
6. The impact of increased ventilation used as an engineering control during in-
stallation of the SPF

IMPACT OF CHEMICAL COMPONENTS OF FOAM

Fig. 4 shows short- and long-term airborne concentrations of HFC-245fa, BDMAEE,
and TCPP in the living area of a residence for SPF installed in the attic. All the simula-
tions in Fig. 4 were done with the modified state-space method except the short-term
HFC-245fa emission, which was represented by the first-order decay model. As shown,
concentrations of HFC-245fa are initially orders of magnitude higher than the other

FIG. 4 Comparison of short- and long-term airborne-modeled concentrations in the

living area for three representative chemicals found within SPF installed in the
attic.
216 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

two chemicals; this is as expected due to the high vapor pressure of HFC-245fa. How-
ever, concentrations of HFC-245fa dropped quickly to the level of the other two chemi-
cals approximately 25 h after installation. When looking at long-term concentrations,
TCPP, the compound with the lowest vapor pressure of the three chemicals modeled,
dominates. Fig. 4 provides insight into the complex nature of SPF emissions and the
changing concentration profiles as the SPF is applied, cures, and hardens.
These results are corroborated by air sampling data collected during and after
the application of SPF. In particular, a 2009 study [6] examined five residences
where SPF was applied. Some homes were finished while others were under con-
struction. Spraying generally lasted for the entire workday, and SPF was applied un-
der the roof deck in the attic and sometimes also to walls in the living space. MDI/
pMDI, HFC-245fa, and an unspecified amine catalyst were analyzed in the study
samples. HFC-245fa levels in two of the homes ranged from nondetectable (ND) to
281 ppm (approximately 1.5 by 106 lg/m3). Concentrations decreased throughout
the day and were higher in the room of application than in other parts of the home.
An amine catalyst, potentially similar to BDMAEE, was detected in all five homes,
and levels ranged from ND to 394 ppb (approximately 2.5 by 103 lg/m3). More
NDs were present overall, but in one home under construction, area samples
collected at the end of the day in the room of application remained elevated at
approximately 200 ppb (approximately 1 by 103 lg/m3), indicating the potential for
longer-term emissions under these conditions. Airborne MDI exceeded occupation-
al exposure limits in most samples collected, and MDI levels had mostly decreased
below the level of detection by the end of the day. Upward and downward migra-
tion of MDI, the amine catalyst, and HFC-245fa across floors were reported.

IMPACT OF PHYSICAL CHARACTERISTICS OF FOAM

In addition to the chemical components of SPF foam, the physical characteristics of
the foam also influence indoor concentrations, specifically for open- versus closed-
cell foams and their associated densities. These characteristics influence the partition-
ing and diffusion of chemical components out of the foam. This is an area that would
benefit greatly from experimental data. For proof-of-concept modeling, the partition
coefficients were estimated using QSAR models and adjusted by density as discussed
in the “Long-Term Emission Parameters” section. Thus, the low-density foam was
associated with smaller partition coefficients and subsequent lower airborne equili-
brium concentrations than the medium-density foam. The open- versus closed-cell
nature of the foams affects diffusion, with smaller diffusion rates and subsequent
lower emission rates for closed-cell foams. These trends are illustrated in Fig. 5. Note
that the two representative chemicals are also present in slightly higher amounts in
the open-cell foam, additionally causing higher airborne concentrations.

IMPACT OF LOCATION OF FOAM INSTALLATION

Chemical concentration profiles in the living space are affected by the location of
SPF foam installation for three overarching reasons: (1) the air exchange or leakage
 BEVINGTON ET AL., DOI 10.1520/STP158920150045 217

FIG. 5 Comparison of airborne-modeled concentrations in the attic for two

representative chemicals and two representative types of SPF installed in a
vented attic with a summer temperature (35  C).

FIG. 6 Comparison of airborne-modeled TCPP concentrations in the source zone for

SPF installed in either an attic or crawl space. Day 0 is May 1.
218 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

rates between either the attic or crawl space and the living space, (2) the
temperature-gradient profiles between either the attic or crawl space, and (3) the
source area in either location. (If a larger area of SPF is typically installed in an attic
versus crawl space, the emissions and resulting concentrations will be larger.) The
other factors can be more complex. Fig. 6 shows the estimated long-term concentra-
tion profile of TCPP in the source zone for an SPF installation in the attic or crawl
space. The peaks centered at 85 elapsed days are modeled as the cooling season
(summer); the “valleys” centered at 270 elapsed days are considered the heating sea-
son (winter). The degree to which the attic and crawl space affect the living space is
related to the magnitude of airflow between the two zones, which in turn depends
on the temperature gradient between the zones and the temperatures in the uncon-
ditioned zones. Because of the stack effect during the heating period, the source in
the crawl space is expected to affect the indoor concentration more than the source in
the attic. On the other hand, because of the reversed stack effect during the cooling sea-
son, the source in the attic becomes a dominant factor [41]. In general, the pollutant
concentrations are higher in summer because of the high temperature in the attic.

IMPACT OF VENTED VERSUS UNVENTED ATTICS AND CRAWL SPACES

Vented attics and crawl spaces have a higher air exchange rate with outdoor air;
this increased ventilation lowers the airborne concentrations in these zones and the

FIG. 7 Comparison of airborne-modeled TCPP concentrations in the source zone for

three SPF installations in either a vented or unvented attic.
 BEVINGTON ET AL., DOI 10.1520/STP158920150045 219

resulting concentrations in the living space. TCPP air concentration profiles for
vented and unvented attics under three different SPF installation scenarios are
shown in Fig. 7 (left: SPF in roofing; middle: SPF in ceiling; right: SPF in both loca-
tions). Results suggest that a vented attic would have a relatively lower air concen-
tration, although the air concentration difference between unvented and vented
attics is small after 20 h. This illustrates an additional complexity in considering
SPF and indoor air quality. Although an unvented attic is beneficial from an energy
conservation perspective, its effect on indoor air quality is rather complex. For ex-
ample, unvented attics tend to have lower air change rates and reduced temperature
fluctuations compared with vented attics. A reduced air exchange rate likely causes
pollutant accumulation in the attic air and thus may decrease indoor air quality un-
der certain, but not all, conditions. The effect of an unvented attic on the air leakage
between attic and living space is not well characterized for homes insulated with
SPF. However, one ventilation study [42] was recently conducted to evaluate
changes in ventilation rates on short-term concentrations of chemicals emitted dur-
ing SPF application. As expected, higher ventilation rates resulted in lower chemical
concentrations. The study notes, however, that there were many factors beyond
ventilation, such as differences between foams, formulations, and application tem-
peratures, that influenced emissions.

FIG. 8 Comparison of airborne-modeled TCPP concentrations in the living space

showing the effect of no HVAC or location in an attic or crawl space. Day 0 is
May 1.
220 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

IMPACT OF HVAC LOCATION

Homes may have HVAC systems installed in the attic or crawl space or in some
cases may not have an HVAC system at all. Air leakages associated with the HVAC
system can be either “leak in” (flowing from indoor air to HVAC duct), “leak out”
(flowing from HVAC duct to indoor air), or both. If SPF is installed in attic or crawl
space, the leak-in flow (i.e., the “return leak”) has the potential to transport SVOCs
and VOCs into the living space.
The location of the HVAC system is subject to the temperature impacts as dis-
cussed in the previous section concerning the location of the SPF. If HVACs are lo-
cated in the attic, their effect on concentrations in the living space are stronger in
the summer, as illustrated in Fig. 8, in which dashed lines are associated with the
highest concentration values at approximately 90 elapsed days (May 1 is Day 0). If
HVACs are located in the crawl space, their effect on concentrations in the living
space are stronger in the winter (at approximately 270 elapsed days).

IMPACT OF INCREASED VENTILATION DURING SPF INSTALLATION AND

LONG-TERM VENTILATION CONSIDERATIONS
During installation of SPF foam, industry best practice suggests artificially increas-
ing the ventilation rate in the zone of installation to establish air flow across the
spraying area and draw overspray away from workers. Ventilation is a method of
controlling worker exposure to hazardous airborne chemicals or flammable vapors
by exhausting contaminated air away from the work area and replacing it with clean
air. To begin to explore the effectiveness of this procedure, the impact of increased
ventilation (10 ACH) during installation and continuing for 2 h after installation

FIG. 9 Comparison of airborne-modeled HCF-245fa concentrations for SPF installed

with either normal or elevated ventilation conditions.
 BEVINGTON ET AL., DOI 10.1520/STP158920150045 221

was examined, as shown in Fig. 9. The increased ventilation lowered estimated air
concentrations during application, providing lower estimated occupational expo-
sures. However, increased short-term ventilation does not lower air concentrations
in the days after installation. In 2014, a study reported short-term concentrations of
a variety of SPF components, including MDI, amine catalysts, TEP (potentially sim-
ilar in behavior to TCPP), and HFC-245fa [43]. Air exchange was increased in all
three homes during the application with air exchange rates ranging from 1 to 3
ACH/min. MDI and TEP levels were low in the first home, with amine catalyst con-
centrations ranging from 0.2 to 1.8 ppm. However, breakthrough was reported for
one amine sample. HFC-245fa levels from personal samplers ranged from 170 to
340 ppm while room air samples were approximately 100 ppm. Samples collected
on other floors after application were much lower and ranged from 3 to 6 ppm.
MDI was not detected in the second home during application. However, TEP was
measured to be above occupational exposure levels in exhaust, indicating that venti-
lation is successful in removing it from the sprayed area. Amines were also detected
in exhaust monitoring at concentrations ranging from 0.1 to 2 ppm (approximately
103 to 104 lg/m3), and HFC-245fa was measured at levels up to 200 ppm (approxi-
mately 106 lg/m3), indicating that ventilation removes these chemicals. Amine con-
centrations reported from personal sampling ranged from 0.8 to 1.8 ppm
(approximately 104 lg/m3). Area sampling of HFC 245fa during installation ranged
from 155 to 300 ppm (approximately 106 lg/m3). Area samples after installation
ranged from 2 to 12 ppm (approximately 105 lg/m3), with no blowing agent
detected 6 months after spray [43].
Tightening a building by installing SPF may affect pressure relations between
occupied spaces, combustion appliance zones, and crawl spaces/basements due to
decreased natural ventilation. Living spaces should be positively pressurized relative
to attics, crawl spaces, and combustion appliance zones to reduce migration of
chemicals into living spaces, which may require considering long-term ventilation
needs [15].

TABLE 4 Effects of eight input parameters on air concentration in living area.

Parameter Air Concentration in Living Area Short or Long Term

Source thickness, h Not correlated NA

Partition coefficient, K Negatively correlated Both
Diffusion coefficient, D Positively correlated Both
Gas-phase mass-transfer coefficient, h Positively correlated Short
Initial concentration in the source, C0 or M0 Positively correlated Both
Air leakage rate from attic to living area, Q21 Positively correlated Both
Ventilation rate in living area, N1 Negatively correlated Both
Ventilation rate in attic, N2 Negatively correlated Both
222 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 10 Sensitivity of airborne-modeled concentrations to the partitioning coefficient

for TCPP: (left) attic; (right) living area.

SENSITIVITY ANALYSIS
An OAT sensitivity analysis was conducted for key variables and provided useful
information on the impacts of these parameters on air concentrations in the living
area. The model output metrics for this analysis were 24-h (short-term) and 1-year
(long-term) averaged air concentrations. The results are presented in Table 4. A neg-
ative correlation means that as the parameter value increased, air concentrations
decreased. Inversely, a positive correlation means that as the parameters value in-
creased, the air concentrations also increased.
A scatter plot analysis was conducted for two key variables: the partitioning co-
efficient K and the diffusion coefficient D for TCPP. In the plots presented in Fig. 10
and Fig. 11, all independent parameters were randomly selected on a logarithmic

FIG. 11 Sensitivity of airborne-modeled concentrations to the diffusion coefficient for

TCPP: (a) attic; (b) living area.
 BEVINGTON ET AL., DOI 10.1520/STP158920150045 223

scale. The scatter plot for K presented in Fig. 10 revealed that the emissions de-
creased as K increased. As the partitioning coefficient becomes larger, however, it is
a key parameter that affects emissions and concentrations, as evidenced by the nar-
row band of dots when K > 1 by 107. Conversely, when the partition coefficient is
small, other parameters have a greater impact on emissions, as indicated by the
scattering of the dots when K < 1 by 107.
The effect of the diffusion coefficient is different than that of the partitioning
coefficient. As shown in Fig. 11, diffusion and emissions are positively correlated.
When the diffusion coefficient is relatively small, it is the key parameter that affects
emissions, as evidenced by the narrow band of dots when D < 1 by 1012 m2/h.
When the diffusion coefficient is large, other parameters also have an impact on
emissions, as indicated by the scattering when D > 1012 m2/h.

Conclusions
A proof-of-concept model was developed that combined code from the EPA’s
IAQX model for multizone environments and application-phase simulation with
code from EPA’s i-SVOC model for diffusional sources and sinks. Model results
shown illustrate that robust and comprehensive modeling of chemical emissions
from SPF requires the following capabilities:
• Incorporating the diffusion-based mass transfer model into a multizone set-
ting to evaluate sources in the attic, living space, and basement or crawl space
• The ability to model the transfer rate from SPF behind a barrier such as gyp-
sum board
• The ability to simulate temperature difference across zones and temperature
gradients within SPF insulation
• The ability to model variable air leakage rates (diurnal and seasonal)
• The ability to simulate the SPF application process, during which the source
area increases from zero to its full size
Additional experimental data are essential for calibrating the model and reduc-
ing the uncertainty of model estimates. Specifically, experimentally determined par-
titioning and diffusion coefficients for typical SPF products are needed. This
information is essential for developing and evaluating QSAR models for estimating
these two parameters for different products and environmental conditions (e.g.,
temperature). These values are informative in characterizing longer-term mass
transfer. Understanding the typical interzonal air flows and the magnitude of air
leakage rate through walls is equally important in predicting the indoor air concen-
trations. Additional chamber data are needed to characterize short-term emissions,
especially within the first 24 h. Prioritizing data needs based on the relative expo-
sure potential of chemicals emitted from SPF would be valuable once the overall
magnitude and trends of chemical emissions from SPF over time are better
characterized.
224 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

ACKNOWLEDGMENTS
We thank the following individuals for helpful feedback in discussions characterizing
modeling approaches and parameters: Mark Mason, Dustin Poppendieck, Mary Cath-
erine Fehrenbacher, Xiaoyu Liu, and Melanie Biggs.

Disclaimer
The views expressed in this manuscript are solely those of the authors and do not rep-
resent the policies of the U.S. Environmental Protection Agency. Mention of trade
names of commercial products should not be interpreted as an endorsement by the
U.S. Environmental Protection Agency.

References

[1] Spray Polyurethane Foam Alliance, “Builder’s Reference Handbook: Reference Materials
for Professional Builders,” Version 1, 2013, http://www.sprayfoam.org/files/docs/news/
557/SPF%20Builders%20Handbook%20v1.pdf (accessed May 1, 2015).

[2] Spence, M. and Graham, C., “Evaluation of Particulates Generated During Trimming and
Cutting of Spray Polyurethane Foam (SPF) Insulation,” 2010, http://polyurethane.ameri-
canchemistry.com/Resources-and-Document-Library/Evaluation-of-Particulates-Gener-
ated-During-Trimming-and-Cutting-of-Spray-Polyurethane-Foam-Insulat.pdf (accessed
May 1, 2015).

[3] Hosein, H. and Farkas, S., “Risk Associated with the Spray Application of Polyurethane
Foam,” Am. Ind. Hyg. Assoc. J., Vol. 42, No. 9, 1981, pp. 663–665.

[4] Crespo, J. and Galan, J., “Exposure to MDI During the Process of Insulating Buildings
with Sprayed Polyurethane Foam” Ann. Occup. Hyg., Vol. 43, No. 6, 1999, pp. 415–419.

[5] Lesage, J., Stanley, J., Karoly, W. J., and Lichtenberg, F. W., “Airborne Methylene Diphenyl
Diisocyanate (MDI) Concentrations Associated with the Application of Polyurethane Spray
Foam in Residential Construction,” J. Occup. Environ. Hyg., Vol. 4, No. 2, 2007, pp. 145–155.

[6] Karlovich, B. and Miller, J., “An Evaluation of Airborne Methylene Diphenyl Diisocyanate
(MDI), Polymeric MDI, and Aliphatic Amines,” Submitted by Bayer Material Science to
EPA, FYI-1209-01618, 2007, https://java.epa.gov/oppt_chemical_search/proxy?filename=
2009-12-FYI-09-01618A_fyi_1209_01618.pdf (accessed May 1, 2005).

[7] Roberge, B., Gravel, R., and Drolet, D., “4,40 -Diphenylmethane Diisocyanate (MDI) Safety
Practices and Concentration During Polyurethane Foam Spraying,” R-629, 2009, http://
www.irsst.qc.ca/media/documents/Pubirsst/r-629.pdf (accessed May 1, 2005).

[8] Streicher, R. P., Reh, C. M., Key-Schwartz, R., Schlecht, P. C., and Cassinelli, M. E.,
“Determination of Airborne Isocyanate Exposure,” NIOSH Manual of Analytical Methods,
1998, CDC, Atlanta, GA, pp. 115–140, http://www.cdc.gov/niosh/docs/2003-154/pdfs/
chapter-k.pdf (accessed May 1, 2005).

[9] Ecoff, S. and Lambach, J., “A Proposed Methodology for Establishing a Safe Work Zone
Around an SPF Application to Exterior Walls of a Commercial Structure,” presented at
 BEVINGTON ET AL., DOI 10.1520/STP158920150045 225

the Polyurethanes Technical Conference, Atlanta, GA, September 24–26, 2012, American
Chemical Council, Washington, DC—unpublished.

[10] Lockey, J. E., Redlich, C. A., Streicher, R., Pfahles-Hutchens, A., Hakkinen, P. B. J., Ellison, G.,
Harber, P., Utell, M., Holland, J., Comai, A., and White, M., “Isocyanates and Human Health:
Multistakeholder Information Needs and Research Priorities,” J. Occup. Environ. Med.,
Vol. 57, No. 1, 2015, pp. 44–51.

[11] Poppendieck, D., Schlegel, M. P., Persily, A. K., and Nabinger, S. J., “Characterization
of Emissions from Spray Polyurethane Foam,” 2014, http://www.nist.gov/manuscript-
publication-search.cfm?pub_id=915634

[12] Kryzmien, M. E., “GC-MS Analysis of Organic Vapours Emitted from Polyurethane Insu-
lation,” Int. J. Environ. Anal. Chem., Vol. 36, No. 4, 1989, pp. 193–207.

[13] Guo, Z., “Simulation Tool Kit for Indoor Air Quality and Inhalation Exposure (IAQX),
Version 1.0 User’s Guide,” 2000, https://nepis.epa.gov/Adobe/PDF/P1000A0G.pdf
(accessed May 1, 2015).

[14] Guo, Z., “Simulation Program i-SVOC User’s Guide, 2013, https://nepis.epa.gov/Adobe/
PDF/P100HYEF.pdf (accessed May 1, 2015).

[15] U.S. EPA, “Self-Evaluation Tool for Professional Spray Polyurethane Foam (SPF) Insu-
lation Application and Related Activities,” 2014, http://www.epa.gov/oppt/spf/spf_
contractor_workplace_practices_evaluation_tool.pdf (accessed May 1, 2015)

[16] Sebrowski, J., “Research Report for Measuring Emissions from Spray Polyurethane
Foam (SPF) Insulation,” presented at the Polyurethanes Technical Conference,
Atlanta, GA, September 24–26, 2012, American Chemical Council, Washington, DC—
unpublished.

[17] Spray Polyurethane Foam Alliance, “Life Cycle Assessment of Spray Polyurethane Foam Insu-
lation for Residential & Commercial Building Applications,” 2012, http://www.sprayfoam.org/
files/docs/SPFA%20LCA%20Long%20Summary%20New.pdf (accessed May 1, 2015).

[18] Clausen, P. A., Laursen, B., Wolkoff, P., Rasmusen, E., and Nielsen, P.A., “Emission of Vola-
tile Organic Compounds from a Vinyl Floor Covering,” Modeling of Indoor Air Quality
and Exposure, ASTM STP1205, N. L. Nagda, Ed., ASTM International, Philadelphia, PA,
1993, pp. 3–13, http://dx.doi.org/10.1520/STP1205-EB

[19] Guo, Z., “A Framework for Modelling Non-Steady-State Concentrations of Semivolatile

Organic Compounds Indoors—I. Emissions from Diffusional Sources and Sorption by
Interior Surfaces,” Indoor Built Environ., Vol. 22, No. 4, 2013, pp. 685–700.

[20] Guo, Z., “Program PARAMS User’s Guide,” 2005, http://www.epa.gov/nrmrl/pubs/

600r05066.html (accessed May 1, 2015).

[21] Bennet, C. O. and Myers, J. E., Momentum, Heat, and Mass Transfer, 3rd Ed., McGraw-Hill,
New York, 1982, p. 504.

[22] Zhao, D., Little, J. C., and Cox, S. S., “Characterizing Polyurethane Foam as a Sink for or Source
of Volatile Organic Compounds in Indoor Air,” J. Environ. Eng., Vol. 130, 2004, pp. 983–989.

[23] Kamprad, I. and Kai-Uwe, G., “Systematic Investigation of the Sorption Properties of Poly-
urethane Foams for Organic Vapors,” Anal. Chem., Vol. 79, No. 11, 2007, pp. 4222–4227.
226 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

[24] Guo, Z., “Review of Indoor Emission Source Models—Part 2. Parameter Estimation,”
Environ. Poll., Vol. 120, No. 3, 2002, pp. 551–564.

[25] Zhang, Y., Luo, X., Wang, X., Qian, K., and Zhao, R., “Influence of Temperature on Formal-
dehyde Emission Parameters of Dry Building Materials,” Atm. Environ., Vol. 41, No. 15,
2007, pp. 3203–3216.

[26] Hulse, R., Bogdan, M., and Iwamoto, N., “Predictive Model for Polyurethane Blowing
Agent Emissions into a House,” 2011, http://www.honeywell-blowingagents.com/?
document=cpi-2011-paper-predictive-model-for-polyurethane-blowing-agent-emissions-
into-a-house&download=1 (accessed May 1, 2015).

[27] Petrich, N., Spak, S., Carmichael, G., Hu, D., Martinez, A., and Hornbuckle, K. C.,
“Simulating and Explaining Passive Air Sampling Rates for Semi-Volatile Compounds on
Polyurethane Foam Passive Samplers,” Environ. Sci. Tech., Vol. 47, No. 15, pp. 8591–8598.

[28] Guo, Z., Liu, X., Krebs, K. A., Greenwell, D. J., Roache, N. F., Stinson, R. A., Nardin, J. A.,
and Pope, R. H., “Laboratory Study of Polychlorinated Biphenyl (PCB) Contamination
and Mitigation in Buildings–Part 2. Transport from Primary Sources to Building Materials
and Settled Dust,” 2012, https://nepis.epa.gov/Exe/ZyNET.exe/P100FA0Z.TXT?ZyActionD=
ZyDocument&Client=EPA&Index=2011+Thru+2015&Docs=&Query=&Time=&EndTime=
&SearchMethod=1&TocRestrict=n&Toc=&TocEntry=&QField=&QFieldYear=&QFieldMonth=
&QFieldDay=&IntQFieldOp=0&ExtQFieldOp=0&XmlQuery=&File=D%3A%5Czyfiles%5
CIndex%20Data%5C11thru15%5CTxt%5C00000005%5CP100FA0Z.txt&User=ANONYMOUS
&Password=anonymous&SortMethod=h%7C-&MaximumDocuments=1&FuzzyDegree=
0&ImageQuality=r75g8/r75g8/x150y150g16/i425&Display=hpfr&DefSeekPage=x&Search
Back=ZyActionL&Back=ZyActionS&BackDesc=Results%20page&MaximumPages=1&ZyEntry=
1&SeekPage=x&ZyPURL (accessed May 1, 2015).

[29] Guo, Z., Liu, X., Krebs, K. A., Roache, N. F., Stinson, R. A., Nardin, J. A., Pope, R. H., Mocka, C. A.,
and Logan, R. D., “Laboratory Study of Polychlorinated Biphenyl (PCB) Contamination and
Mitigation in Buildings–Part 3. Evaluation of the Encapsulation Method,” 2012, https://nepis.
epa.gov/Exe/ZyNET.exe/P100FA5L.TXT?ZyActionD=ZyDocument&Client=EPA&Index=
2011+Thru+2015&Docs=&Query=&Time=&EndTime=&SearchMethod=1&TocRestrict=n&
Toc=&TocEntry=&QField=&QFieldYear=&QFieldMonth=&QFieldDay=&IntQFieldOp=0&
ExtQFieldOp=0&XmlQuery=&File=D%3A%5Czyfiles%5CIndex%20Data%5C11thru15%5
CTxt%5C00000005%5CP100FA5L.txt&User=ANONYMOUS&Password=anonymous&
SortMethod=h%7C-&MaximumDocuments=1&FuzzyDegree=0&ImageQuality=r75g8/r75g8/
x150y150g16/i425&Display=hpfr&DefSeekPage=x&SearchBack=ZyActionL&Back=Zy
ActionS&BackDesc=Results%20page&MaximumPages=1&ZyEntry=1&SeekPage=x&ZyPURL
(accessed May 1, 2015).

[30] Dole, P, Feigenbaum, A. E., De La Cruz, C., Pastorelli, S., Paseiro, P., Hankemeier, T.,
Voulzatis, Y., Aucejo, S., Saillard, P., and Papaspyrides, C., “Typical Diffusion Behaviour in
Packaging Polymers—Application to Functional Barriers,” Food Add. Contam., Vol. 23,
No. 2, pp. 2006, pp. 202–211.

[31] Sextro, R. G., Feustel, H. E., Modera, M. P., Revzan, K. L., and Sherman, M. H., “A Coupled
Multizone Flow and Radon Transport Model of Radon Entry and Concentrations in a
Crawlspace House,” 1992, http://eetd.lbl.gov/publications/coupled-multi-zone-flow-
and-radon-transport-model-radon-entry-and-concentrations-crawlspace-house (accessed
December 16, 2016).
 BEVINGTON ET AL., DOI 10.1520/STP158920150045 227

[32] Rudd, A. and Bergey, D., “Ventilation System Effectiveness and Tested Indoor Air Quality
Impacts,” 2013, https://buildingscience.com/documents/bareports/ba-1309-ventilation-sys-
tem-effectiveness-and-indoor-air-quality-impacts/view (accessed May 1, 2015).

[33] Prahl, D. and Shaffer, M., “Moisture Risk in Unvented Attics Due to Air Leakage Paths, 2014,”
http://www.nrel.gov/docs/fy15osti/63048.pdf (accessed May 1, 2015).

[34] Weschler, C. J. and Nazaroff, W. W., “SVOC Partitioning Between the Gas Phase and Set-
tled Dust Indoors,” Atm. Environ., Vol. 44, No. 30, 2010, pp. 3609–3620.

[35] Shi, S. and Zhao, B., “Comparison of the Predicted Concentration of Outdoor Originated In-
door Polycyclic Aromatic Hydrocarbons Between a Kinetic Partition Model and a Linear In-
stantaneous Model for Gas-Particle Partition,” Atm. Environ., Vol. 59, 2012, pp. 93–101.

[36] U.S. EPA, Exposure Factors Handbook, U.S. EPA, Washington, DC, 2011.

[37] Parker, D., “Literature Review of the Impact and Need for Attic Ventilation in Florida
Homes,” 2005, http://www.fsec.ucf.edu/en/publications/pdf/FSEC-CR-1496-05.pdf
(accessed December 16, 2016).

[38] U.S. EPA, “Assessment of Risks from Radon in Homes,” 2003, https://sosradon.org/
files/sosradon/EPA%20-%20Assessment%20of%20Risks%20from%20Radon%20in%20-
Homes%20-%202003.pdf (accessed May 1, 2015).

[39] Walker, I. S., Forest, T. W., and Wilson, D. J., “An Attic-Interior Infiltration and Interzone
Transport Model of a House,” Building Environ., Vol. 40, No. 5, 2005, pp. 701–718.

[40] Coutler, J., “Liabilities of Vented Crawl Spaces, Their Impacts on Indoor Air Quality in South-
eastern U.S. Homes and One Intervention Strategy,” n.d., http://c.ymcdn.com/sites/
www.nibs.org/resource/resmgr/BEST/BEST1_024.pdf (accessed May 1, 2015).

[41] Quirouette, R. L. and Arch, B., Air Pressure and the Building Envelope, Canada Mortgage
and Housing Corporation, Ontario Association of Architects, 2004, pp. 1–19.

[42] Wood, R., “CPI Ventilation Research Project for Estimating Re-Entry Times for Trade Workers
Following Application of Three Generic Spray Polyurethane Foam Formulations,” presented
at the Polyurethanes Technical Conference, Dallas, TX, September 22–24, 2014, American
Chemical Council, Washington, DC, http://polyurethane.americanchemistry.com/Resources-
and-Document-Library/CPI-Ventilation-Research-Project-for-Estimating-Re-Entry-Times-for-
Trade-Workers-Following-Applicati.pdf (accessed May 1, 2015).

[43] Robert, W., Anderson, J., Wood, R., and Bogdan, M., “Ventilation and Re-Occupancy of a
Residential Home Sprayed with High Pressure Polyurethane Foam,” presented at the
Polyurethanes Technical Conference, Atlanta, GA, September 24–26, 2012, American
Chemical Council, Washington, DC—unpublished.
228 DEVELOPING CONSENSUS STANDARDS FOR MEASURING CHEMICAL EMISSIONS

STP 1589, 2017 / available online at www.astm.org / doi: 10.1520/STP158920150037

Katherine Sleasman,1 Carol Hetfield,1 and Melanie Biggs2

Investigating Sampling and

Analytical Techniques to
Understand Emission
Characteristics from Spray
Polyurethane Foam Insulation
and Data Needs
Citation
Sleasman, K., Hetfield, C., and Biggs, M., “Investigating Sampling and Analytical Techniques to
Understand Emission Characteristics from Spray Polyurethane Foam Insulation and Data
Needs,” Developing Consensus Standards for Measuring Chemical Emissions from Spray
Polyurethane Foam (SPF) Insulation, ASTM STP1589, J. Sebroski and M. Mason, Eds., ASTM
International, West Conshohocken, PA, 2017, pp. 228–277, http://dx.doi.org/10.1520/
STP1589201500373

ABSTRACT
The increasing usage of spray polyurethane foam (SPF) insulation in residential
buildings, coupled with potential adverse health outcomes, warrant consideration
of exposure pathways to the chemicals that may be emitted from SPF, including
unreacted isocyanates, catalysts, blowing agents, flame retardants, and aldehydes.
The generation of new emissions data through chamber testing and modeling can
inform stewardship guidance and practices during the use of SPF insulation,
including application techniques and environmental variables, postapplication
reentry times, personal protective equipment guidelines, and the use of adequate
ventilation. Emissions test methods and protocols that utilize sampling and
analytical protocols are needed to understand the variables that affect emissions
and curing to develop and assess residential exposure scenarios. Factors
important for chamber testing include the type of chamber being used (e.g.,

Manuscript received May 12, 2015; accepted for publication September 29, 2015.
1
U.S. Environmental Protection Agency, 1201 Constitution Ave. NW, Washington, DC 20460
2
U.S. Consumer Product Safety Commission, 5 Research PL., Rockville, MD 20850
3
ASTM Symposium on Developing Consensus Standards for Measuring Chemical Emissions from Spray
Polyurethane Foam (SPF) Insulation on April 30–May 1, 2015 in Anaheim, CA.

Copyright V
C 2017 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959.
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 229

emissions test cell, small chamber, full-scale chamber) and the physical
characteristics of the test chambers. Chamber testing data could then be used to
model product formulation, room size, distance from the source, amount of
product used (grams per event), thickness of application, air exchange and
ventilation rates, equipment-related variables (method of application), temperature,
and humidity. This paper evaluates current analytical and sampling methods
used to determine potential worker and residential exposure to SPF chemicals.
Data gaps in worker and residential exposures will also be discussed to show
where chamber testing and modeling can answer questions such as what is an
appropriate reoccupancy or reentry time and recommended ventilation methods.

Keywords
isocyanates, polyurethane, spray polyurethane foam insulation, SPF, modeling,
microscale chambers, small-scale chambers, full-scale chambers, indoor air, building
performance, consumer/residential, risk management, reoccupancy, reentry

Introduction
Spray polyurethane foam (SPF) insulation used in residential, commercial, and
public buildings is a potential source of exposure from a complex and varied mix-
ture of chemical substances. SPF is produced by a chemical reaction of two compo-
nent parts, commonly referred to as Side A and Side B. In the United States, Side A
typically contains methylene diphenyl isocyanate (MDI), consisting of monomeric
MDI and higher-molecular-weight oligomers. Side B contains polyols and a mixture
of other chemicals, including catalysts, flame retardants, blowing agents, and surfac-
tants. Polyurethane polymers are formed by reacting di- or polyisocyanates with
polyols. Depending on the SPF formulation being installed, the polyol side may in-
clude a complex proprietary formulation of polyols, flame retardants, surfactants,
cross-linking agents, chain extenders, catalysts, blowing agents, pigments, fillers, or
other ingredients [1]. One type of SPF is professionally installed and consists of a
two-component, high-pressure, open- or closed-cell insulation. It is applied on-site
at elevated temperatures through heated spray hoses, exothermic reactions, and
pressures in residential or commercial/institutional buildings [2].
SPF is a complex mixture of chemicals and reactions. During and after curing of
the SPF, chemicals may be emitted at different rates depending on the extent of the
reactions and may lead to adverse health effects, such as respiratory-related prob-
lems (i.e., asthma, coughing); irritation (e.g., eyes, throat); and headaches. Some SPF
ingredients are recognized as presenting respiratory hazards in the workplace, and
those same concerns could apply to other trade workers and residential/commercial
building inhabitants because it is manufactured on-site. Thus, exposures to chemi-
cals emitted from SPF through vapors and particulates during and after application
are possible, and it is especially important to understand the reoccupancy timeframe
for the inhabitants entering residential or commercial/institutional buildings and the
230 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

reentry timeframe for the trade workers who enter these buildings after SPF has
been applied. Analyses of spray foam application, ventilation techniques, sampling
and analytical methods, and industrial hygiene (IH) studies will be presented to un-
derstand better what chemicals residential/commercial building inhabitants could
potentially be exposed to and how to model them.

Occupational Exposure Limits

Due to the hazards posed by SPF chemicals, most chemical substances used in SPF
have recognized occupational exposure limits (OELs) established by the Occupational
Safety and Health Administration (OSHA), the National Institute for Occupational
Safety and Health (NIOSH), the American Conference of Industrial Hygienists
(ACGIH), and GESTIS. OSHA sets enforceable permissible exposure limits (PELs) to
protect workers. Most OSHA PELs have 8-h time-weighted averages (TWAs), although
there are also ceiling and peak limits, and many chemicals include a skin designation to
warn against skin contact. Approximately 500 PELs have been established. However,
there are many chemical substances for which OSHA does not have workplace expo-
sure limits; for these substances, OSHA recommends alternate OELs, such as NIOSH
Recommended Exposure Limits (RELs) or ACGIH Threshold Limit Values (TLVs).
OSHA also concedes that many PELs are outdated and not protective. Therefore,
OSHA’s Hazard Communication Standard requires the development and dissemina-
tion of Safety Data Sheets and training for workers on the hazards of chemicals, how to
handle chemicals safely, including the use of personal protective equipment (PPE), and
medical surveillance. OSHA also encourages the industry to develop guidelines to
ensure worker protection because there may be exposures to chemicals in products,
such as SPF, even when levels are in compliance with the relevant PELs [3].

ISOCYANATE (SIDE A) OELS AND GUIDELINES

Diisocyanates, a key component of SPF, are skin and respiratory sensitizers and irri-
tants and have been documented in the workplace as causing asthma [4–6], lung
damage, and, in severe cases, death [7–10]. A variety of regulatory exposure limits
are available for MDI and other isocyanates because these chemicals have well-
documented health effects [2,8–10] (Table 1). OSHA regulates workplace exposures
to MDI by setting a PEL. To reduce worker exposures when levels exceed PELs,
OSHA requires the use of PPE or engineering controls, such as ventilation or ad-
ministrative controls. OSHA’s PEL for the MDI monomer is 0.2 mg/m3 (0.02 ppm)
as a ceiling limit [11]. OSHA initiated a National Emphasis Program in 2013 to
identify and reduce or eliminate the incidence of adverse health effects associated
with occupational exposure to isocyanates [12].

AMINE CATALYSTS (SIDE B) OELS AND GUIDELINES

Most amine catalysts used in Side B, which constitute 4 % to 10 % of the SPF mix-
ture [16], have not been assigned a regulatory exposure limit. Table 2 represents
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 231

TABLE 1 Exposure limits for MDI [2,8,9,11,13–15].

Agency Exposure Exposure Limit Concentration Method

Limit Type

OSHA PEL MDI monomer: 0.2 mg/m3 (0.02 ppm) as a OSHA 47

ceiling limit
NIOSH REL MDI monomer: TWA of 0.05 mg/m3 (0.005 ppm) NIOSH 5525
for up to a 10-h workday during a 40-h workweek
with a ceiling limit of 0.2 mg/m3 (0.02 ppm) for
any 10-min period
U.S. EPA IRIS Reference 6  104 mg/m3
concentration
ACGIH TLVs MDI monomer: 0.05 mg/m3 as a TWA for 8-h work-
day and 40-h (0.005 ppm) workweek
California Office of California Acute: 12 lg/m3 (0.8 ppb) Chronic: 0.08 lg/m3
Environmental Health reference (0.0074 ppb) 8 h: 0.16 lg/m3 (0.015 ppb)
and Hazard Assess- exposure level
ment (OEHHA)
Austria MDI monomer: 0.05 mg/m3 TWA: 0.005 ppm
STEL: 0.1 mg/m3 (0.01 ppm)
Belgium MDI monomer: 0.052 mg/m3 TWA: 0.005 ppm
China MDI monomer: 0.05 mg/m3 TWA STEL: 0.1 mg/m3
Germany MDI monomer 0.05 mg/m3 TWA STEL: 0.05 mg/m3
Polymeric MDI: 0.05 mg/m3 inhalable aerosol
STEL: 0.1 mg/m3
Hungary MDI monomer: 0.05 mg/m3 TWA STEL: 0.05 mg/m3
Ireland MDI monomer: 0.02 mg/m3 TWA STEL: 0.07 mg/m3
Poland MDI monomer: 0.05 mg/m3 TWA STEL: 0.2 mg/m3
Singapore MDI monomer: 0.051 mg/m3 TWA: 0.005 ppm
South Korea MDI monomer 0.055 mg/m3 TWA: 0.005 ppm
Spain MDI monomer: 0.052 mg/m3 TWA: 0.005 ppm
Sweden MDI monomer: 0.03 mg/m3 TWA: 0.002 ppm
STEL: 0.05 mg/m3 (0.005 ppm)

amine catalysts that have been identified for use in SPF and have regulatory and
nonregulatory OELs. The OELs, PELs, and TLVs have assigned TWAs over an 8-h
workday and assigned short-term exposure limits (STELs) for average exposure
over a 15-min period that should not be exceeded in an 8-h workday. To reduce
worker exposures when levels exceed PELs, OSHA requires the use of PPE or engi-
neering controls, such as ventilation or administrative controls. NIOSH considers
SPF insulation application to present hazards similar to other spray polyurethane
applications and recommends the use of the same safety procedures and PPE.
Exposure to the amine catalysts may cause irritation of the respiratory tract and eyes.
Inhalation of vapors may cause temporary foggy or blurry vision, and halos may
appear around bright objects, conditions known as “halovision” or “blue haze” [17].
232 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 2 Exposure limits for amine catalysts [17–24].

Agency and Exposure Limit

Chemical CAS RN Concentration Method

Bis(2-dimethylaminoethyl) ether 3033-62-3 ACGIH

(BDMAEE)
STEL: 0.15 ppm
TWA: 0.05 ppm
Belgium
STEL: 1 mg/m3 (0.15 ppm)
TWA: 0.33 mg/m3 (0.05 ppm)
Canada
STEL: 0.15 ppm
TWA: 0.05 ppm
Diethanolamine 111-42-2 ACGIH OSHA PV2018
TLV – TWA: 1 mg/m3 (3 ppm)
REL – TWA: 15 mg/m3 (3 ppm)
Australia
TWA: 13 mg/m3 (3 ppm)
Austria
STEL: 4 mg/m3 (0.92 ppm)
TWA: 2 mg/m3 (0.46 ppm)
Belgium
TWA: 2 mg/m3 (0.46 ppm)
Canada – Ontario
TWA: 1 mg/m3
Canada – Quebec
TWA: 13 mg/m3 (3 ppm)
Denmark
STEL: 4 mg/m3 (0.92 ppm)
TWA: 2 mg/m3 (0.46 ppm)
Finland
TWA: 2 mg/m3 (0.46 ppm)
France
TWA: 15 mg/m3 (3 ppm)
Germany (DFG)
TWA: 1 mg/m3 TWA
Ireland
TWA: 1 mg/m3
New Zealand
TWA: 13 mg/m3 (3 ppm)
Poland
TWA: 9 mg/m3
Singapore
TWA: 2 mg/m3 (0.46 ppm)
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 233

TABLE 2 (Continued)

Agency and Exposure Limit

Chemical CAS RN Concentration Method

South Korea
TWA: 2 mg/m3 (0.46 ppm)
Spain
TWA: 2 mg/m3 (0.46 ppm)
Sweden
STEL: 30 mg/m3 (6 ppm)
TWA: 5 mg/m3 (3 ppm)
Switzerland
STEL: 1 mg/m3
TWA: 1 mg/m3
United Kingdom
TWA: 13 mg/m3 (3 ppm)
Dimethylethanolamine 108-01-0 N/A N/A
N-Ethylmorpholine 100-74-3 ACGIH NIOSH S-264
TLV – TWA: 24 mg/m3 (5 ppm)
NIOSH
REL: 5-ppm ceiling
OSHA
PEL: 94 mg/m3 (20 ppm)
Australia
TWA: 24 mg/m3 (5 ppm)
Austria
STEL: 46 mg/m3 (10 ppm)
TWA: 23 mg/m3 (5 ppm)
Belgium
TWA: 24 mg/m3 (5 ppm)
Canada – Ontario
TWA: 5 ppm
Canada – Quebec
TWA: 24 mg/m3 (5 ppm)
China
TWA: 25 mg/m3
Denmark
STEL: 47 mg/m3 (10 ppm)
TWA: 23.5 mg/m3 (5 ppm)
Finland
STEL: 48 mg/m3 (10 ppm)
TWA: 24 mg/m3 (5 ppm)
France
TWA: 23 mg/m3 (5 ppm)
234 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 2 (Continued)

Agency and Exposure Limit

Chemical CAS RN Concentration Method

Ireland
STEL: 95 mg/m3 (20 ppm)
TWA: 23 mg/m3 (5 ppm)
New Zealand
TWA: 24 mg/m3 (5 ppm)
Poland
STEL: 46 mg/m3
TWA: 23 mg/m3
Singapore
TWA: 24 mg/m3 (5 ppm)
South Korea
TWA: 23 mg/m3 (5 ppm)
Spain
TWA: 24 mg/m3 (5 ppm)
Sweden
STEL: 50 mg/m3 (10 ppm)
TWA: 25 mg/m3 (5 ppm)
Switzerland
TWA: 25 mg/m3 (5 ppm)
United Kingdom
STEL: 96 mg/m3 (20 ppm)
TWA: 24 mg/m3 (5 ppm)
N,N-Dimethylaminopropylamine 109-55-7 ACGIH
TWA: 0.5 ppm
Canada – Ontario
TWA: 2 mg/m3 (0.5 ppm)
Triethanolamine 102-71-6 ACGIH OSHA PV2141
STEL: 12.4 mg/m3 (3 ppm)
TLV – TWA: 4.1 mg/m3 (1 ppm)
NIOSH
REL – TWA: 23 mg/m3 (5 ppm)
Australia
TWA: 5 mg/m3
Austria
STEL: 10 mg/m3 (0.16 ppm)
TWA: 5 mg/m3 (0.8 ppm)
Belgium
TWA: 5 mg/m3
Canada – Ontario
TWA: 3.1 mg/m3 (0.5 ppm)
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 235

TABLE 2 (Continued)

Agency and Exposure Limit

Chemical CAS RN Concentration Method

Canada – Quebec
TWA: 5 mg/m3
Denmark
TWA: 3.1 mg/m3 (0.5 ppm)
STEL: 6.2 mg/m3 (1 ppm)
Finland
TWA: 5 mg/m3
Germany (DFG)
TWA: 5 mg/m3
STEL: 20 mg/m3
Ireland
TWA: 5 mg/m3
New Zealand
TWA: 5 mg/m3
Singapore
TWA: 5 mg/m3
Spain
TWA: 5 mg/m3
Sweden
STEL: 10 mg/m3 (1.6 ppm)
TWA: 5 mg/m3 (0.8 ppm)
Switzerland
TWA: 5 mg/m3
STEL: 20 mg/m3
Triethylamine 121-44-8 ACGIH OSHA PV2060
STEL: 3 ppm
OSHA
PEL – TWA: 100 mg/m3 (25 ppm)
Australia
STEL: 17 mg/m3 (4 ppm)
TWA: 8 mg/m3 (2 ppm)
Austria
STEL 12.6 mg/m3 (3 ppm)
TWA: 8.4 mg/m3 (2 ppm)
Belgium
TWA: 4.2 mg/m3 (1 ppm)
STEL: 12.6 mg/m3 (3 ppm)
Canada – Ontario
STEL: 3 ppm
TWA: 1 ppm
236 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 2 (Continued)

Agency and Exposure Limit

Chemical CAS RN Concentration Method

Canada – Quebec
STEL: 61.5 mg/m3 (15 ppm)
TWA: 20.5 mg/m3 (5 ppm)
Denmark
STEL: 8.2 mg/m3 (2 ppm)
TWA: 4.1 mg/m3 (1 ppm)
European Union
STEL: 12.6 mg/m3 (3 ppm)
TWA: 8.4 mg/m3 (2 ppm)
Finland
STEL: 4.2 mg/m3 (1 ppm)
France
STEL: 12.6 mg/m3 (3 ppm)
TWA: 4.2 mg/m3 (1 ppm)
Germany (AGS and DFG)
STEL 8.4 mg/m3 (2 ppm)
TWA: 4.2 mg/m3 (1 ppm)
Hungary
STEL: 12.6 mg/m3
TWA: 8.4 mg/m3
Italy
STEL: 12.6 mg/m3 (3 ppm)
TWA: 8.4 mg/m3 (2 ppm)
Ireland
STEL 12.6 mg/m3 (3 ppm)
TWA: 8.4 mg/m3 (2 ppm)
Latvia
STEL: 12.6 mg/m3 (3 ppm)
TWA: 8.4 mg/m3 (2 ppm)
Netherlands
STEL: 12.6 mg/m3
TWA: 4.2 mg/m3
New Zealand
STEL: 20 mg/m3 (5 ppm)
TWA: 12 mg/m3 (3 ppm)
Poland
STEL: 9 mg/m3
TWA: 3 mg/m3
Singapore
STEL 20.7 mg/m3 (5 ppm)
TWA: 4.1 mg/m3 (1 ppm)
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 237

TABLE 2 (Continued)

Agency and Exposure Limit

Chemical CAS RN Concentration Method

South Korea
STEL 16.6 mg/m3 (4 ppm)
TWA:8.3 mg/m3 (2 ppm)
Spain
STEL 12 mg/m3 (3 ppm)
TWA:8.4 mg/m3 (2 ppm)
Sweden
STEL: 40 mg/m3 (10 ppm)
TWA: 8 mg/m3 (2 ppm)
Switzerland
STEL: 8.4 mg/m3 (2 ppm)
TWA: 4.2 mg/m3 (1 ppm)
United Kingdom
STEL: 17 mg/m3 (4 ppm)
TWA: 8 mg/m3 (2 ppm)
Triethylenediamine 280-57-9 Canada – Ontario
TWA: 4.6 mg/m3 (1 ppm)
Trimethylaminoethylethanolamine 2212-32-0 N/A N/A
(TMAEEA)

BLOWING AGENT (SIDE B) OELS AND GUIDELINES

Most blowing agents, which may constitute up to 9 % of the SPF mixture [16], are pro-
prietary and do not have regulatory exposure limits. However, proprietary formulations
do have set manufacturer acceptable exposure limits (AELs), recommended engineer-
ing controls, and PPE requirements. According to manufacturers, if these AELs are ad-
hered to, risks to workers are significantly decreased. Other blowing agents have
workplace environmental exposure limits (WEELs) assigned by the American Industri-
al Hygiene Association (AIHA), OSHA PELs, and NIOSH RELs, which have TWAs
ranging from 300 to 1,000 ppm. These workplace exposure limits protect workers from
potential inhalation health effects, including drowsiness, dizziness, disorientation, nau-
sea, vomiting, narcosis, asphyxia, and cardiac arrhythmia [25,26].
Previous blowing agents used in SPF and other polyurethane products con-
sisted of chlorinated volatiles, such as trichlorofluoromethane and other chlorinated
fluorocarbons (CFCs). These were commonly used until the late 1990s when CFCs
were phased out due to concerns about their ozone depletion potential (ODP). Fol-
lowing the phase out, the U.S. Environmental Protection Agency (EPA) created the
Significant New Alternatives Policy (SNAP) program under Section 612 of the
Clean Air Act amendments. SNAP evaluates alternatives to ozone-depleting sub-
stances. Substitutes are reviewed on the basis of ODP, global warming potential
(GWP), toxicity, flammability, and exposure potential. As of October 2014, the
238 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

EPA has approved the blowing agents in Table 3 for use in rigid polyurethane foam,
and some of the listed blowing agents are also commonly used in SPF [25], such as
1,1,1,3,3-pentafluoropropane (HFC-245fa), which has a GWP of 950 to 1,020 [25,27].

TABLE 3 Exposure limits for approved blowing agents evaluated under the SNAP program [26].

Blowing Agents for

Rigid Polyurethane Identification Method for
Foam Information Agency Exposure Limit Concentration Detection

Carbon dioxide CAS ID 124-38-9 OSHA PEL – TWA: 9,000 mg/m3 OSHA ID-175
(5y,000 ppm)
REL – TWA: 9,000 mg/m3
(5,000 ppm)
STEL: 54,000 mg/m3
(30,000) ppm
Ecomate Confidential business OSHA PEL: 100 ppm
information (CBI)
Electroset CBI N/A
technology
Exxol blowing Blend contains light Manufacturer Manufacturer acceptable expo-
agent saturated hydrocar- sure limit (AEL): 504 ppm
bons with five
carbons
Formacel B CBI Manufacturer Manufacturer AEL: 1000 ppm
Formacel TI CBI Manufacturer Manufacturer AEL: 1000 ppm
Formic acid CAS ID 64-18-6 OSHA PEL –TWA: 9 mg/m3 (5 ppm) OSHA ID-186SG
HFC-134a CAS ID 811-97-2 AIHA WEEL: 1,000 ppm
HFC-152a CAS ID 75-37-6 AIHA WEEL: 1,000 ppm
HFC-245fa CAS ID 460-58-6 AIHA WEEL: 300 ppm
HFO-1234ze CAS ID 29118-24-9 EPA AEL: 1,000 ppm
Methyl formate CAS ID 107-31-3 OSHA PEL – TWA: 250 mg/m3 OSHA PV2041
(100 ppm)
Saturated light Propane, butane, OSHA PEL – TWA: 300 to 1,000 ppm OSHA PV2077
hydrocarbons isobutene, pentane,
C3–C6 cyclopentane,
hexane, cyclohexane
OSHA PV2010
NIOSH REL – TWA: 300 to 1,000 ppm NIOSH 1500
Solstice trans-1- CBI AIHA WEEL: 800 ppma
chloro-3,3,3-
trifluoropropene
Transcend CBI OSHA PEL: 200 ppm
TechnologiesTM
Water CAS ID 7732-18-5 N/A

Note: Formacel B, Formacel TI, HFC 134a, and HFC 245fa will no longer be acceptable under the
SNAP program as of January 1, 2020. See the following website for the Federal Register Notice:
https://www.epa.gov/snap/snap-regulations
a
SPF is 291 ppm because of high exposure.
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 239

However, two of the next-generation blowing agents for SPF are Solstice and
Formacel, which have very low GWPs but are in various stages of market penetra-
tion [28,29].

FLAME RETARDANTS (SIDE B) OELS AND GUIDELINES

Tris(1-chloro-2-propyl)phosphate (TCPP) (CAS-RN 13674-84-5) is a commonly
used flame retardant in polyurethane products such as SPF and can be emitted
from these products in residential and other buildings, leading to potential exposure
[16,30]. There are no exposure guidelines for workers for TCPP; however, many
manufacturers of flame retardants advise using PPE, ventilation, and general hy-
giene when working with these chemicals [31]. TCPP shows limited evidence of
chronic organ toxicity in animal studies [32] and has been recommended to the
National Toxicology Program for chronic toxicity and cancer assessments [33].
Other phosphate flame retardants used in SPF may include triethylphosphate (TEP)
(CAS-RN 78-40-0) and tris(1,3-dichloroisopropyl)phosphate (CAS-RN 13674-87-8)
[16]. Similarly, there are no exposure limits for these flame retardants, but they also
demonstrate some adverse health effects [30]. Although not identified for SPF use,
OSHA has set a legal limit for other phosphate ester flame retardants of 3 mg/m3
workplace air for triphenylphosphate and 5 mg/m3 for tri-n-butylphosphate over an
8-h workday and 40-h workweek [34,35].

REACTION PRODUCTS, BYPRODUCTS, AND OTHER CHEMICALLY

DETECTED OELS
When SPF is applied, the formation of aldehydes may occur and applicators and
other trade workers may be exposed [36]. These byproducts could lead to adverse
health effects, including irritation of the eyes, nose, and throat; conjunctivitis;
cough; central nervous system depression; and delayed pulmonary edema. These
byproducts have been shown in animals to have kidney, reproductive, and terato-
genic effects. Table 4 list aldehydes that have been measured in SPF studies as
byproducts or product impurities [36–38]. Many of the aldehydes have well-
established regulatory exposure limits [39].
Furthermore, other compounds that have been observed in studies include
1,4-dioxane, chlorobenzene, hexamethylcyclotrisiloxane, octamethylyclotrisiloxane,
tetramethylethylenediamine, 1-chloro-2-propanol, and propylene glycol [36]
(Table 5). Many of these compounds also have established regulatory exposure limits
and have several health effects associated with them. Chlorobenzene, propylene gly-
col, and 1,4-dioxane may cause dizziness, drowsiness, headache, and irritation of
the eyes, skin, nose, and throat. 1,4-Dioxane may also cause liver and kidney dam-
age, and it is a suspected carcinogen [48–50].

Manufacturer Specifications for Applying SPF

SPF should be applied according to manufacturer specifications to ensure worker pro-
tection and product effectiveness. Environmental conditions, such as temperature,
240 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 4 OELs for aldehydes [36–47].

Agency/Country and Exposure Limit

Aldehyde CASRN Concentration Method

Actetaldehyde 75-07-7 OEHHA OSHA 68

CREL: 140 ppm
OSHA
PEL – TWA: 360 mg/m3 (200 ppm)
Australia
STEL: 91 mg/m3 (50 ppm)
TWA: 36 mg/m3 (20 ppm)
Austria
STEL: 90 mg/m3 (50 ppm)
TWA: 90 mg/m3 (50 ppm)
Belgium
TWA: 46 mg/m3 (25 ppm)
Canada – Ontario
STEL: 25 ppm
Canada – Quebec
STEL: 45 mg/m3 (25 ppm)
China
STEL: 45 mg/m3
Denmark
STEL: 45 mg/m3 (25 ppm)
TWA: 45 mg/m3 (25 ppm)
Finland
STEL: 46 mg/m3 (25 ppm)
France
TWA: 180 mg/m3 (100 ppm)
Germany (AGS and DFG)
STEL: 91 mg/m3 (50 ppm)
TWA: 91 mg/m3 (50 ppm)
Hungary
STEL: 25 mg/m3
TWA: 25 mg/m3
Ireland
STEL: 45 mg/m3 (25 ppm)
TWA: 45 mg/m3 (25 ppm)
Latvia
TWA: 5 mg/m3
Netherlands
TWA: 37 ppm
New Zealand
STEL: 50 ppm
TWA: 20 ppm
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 241

TABLE 4 (Continued)

Agency/Country and Exposure Limit

Aldehyde CASRN Concentration Method

Poland
STEL 45 mg/m3
TWA: 5 mg/m3
Singapore
STEL: 45 mg/m3 (25 ppm)
South Korea
STEL: 270 mg/m3 (150 ppm)
TWA: 90 mg/m3 (50 ppm)
Spain
STEL: 46 mg/m3 (25 ppm)
Sweden
STEL: 90 mg/m3 (50 ppm)
TWA: 45 mg/m3 (25 ppm)
Switzerland
STEL: 90 mg/m3 (50 ppm)
TWA: 90 mg/m3 (50 ppm)
United Kingdom
STEL: 92 mg/m3 (50 ppm)
TWA: 37 mg/m3 (20 ppm)
Benzaldehyde 100-52-7 Canada – Ontario N/A
STEL: 17 mg/m3 (4 ppm)
Finland
STEL: 4.4 mg/m3 (1 ppm)
TWA: 17.4 mg/m3 (4 ppm)
Hungary
STEL: 10 mg/m3
TWA: 5 mg/m3
Latvia
TWA: 5 mg/m3
Poland
STEL: 40 mg/m3
TWA: 10 mg/m3
Butyraldehyde 123-72-8 Austria NIOSH 2539
STEL: 64 mg/m3 (20 ppm)
TWA: 64 mg/m3 (20 ppm)
China
STEL: 10 mg/m3
TWA: 5 mg/m3
Finland
TWA: 74 mg/m3 (25 ppm)
242 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 4 (Continued)

Agency/Country and Exposure Limit

Aldehyde CASRN Concentration Method

Germany (AGS)
STEL: 64 mg/m3 (20 ppm)
TWA: 64 mg/m3 (20 ppm)
Latvia
TWA: 5 mg/m3
Crotonaldehyde 123-73-9 OSHA OSHA 81
PEL – TWA: 6 mg/m3 (2 ppm)
Austria
STEL: 4 mg/m3 (1.36 ppm)
TWA: 1 mg/m2 (0.34 ppm)
Denmark
STEL: 12 mg/m3 (4 ppm)
TWA: 6 mg/m3 (2 ppm)
Finland
STEL: 0.87 mg/m3 (0.3 ppm)
TWA: 0.29 mg/m3 (0.1 ppm)
France
TWA: 6 mg/m3 (2 ppm)
Ireland
STEL: 18 mg/m3 (6 ppm)
TWA: 6 mg/m3 (2 ppm)
Spain
STEL: 0.87 mg/m3 (0.3 ppm)
Switzerland
TWA: 1 mg/m3 (0.34 ppm)
4170-30-3 Australia
TWA: 5.7 mg/m3 (2 ppm)
Belgium
STEL: 0.87 mg/m3 (0.3 ppm)
TWA: 5.8 mg/m3 (2 ppm)
Canada – Ontario
STEL: 0.3 ppm
Canada – Quebec
TWA: 5.7 mg/m3 (2 ppm)
China
STEL: 12 mg/m3
Denmark
TWA: 6 mg/m3 (2 ppm)
STEL: 12 mg/m3 (4 ppm)
Finland
STEL: 0.87 mg/m3 (0.3 ppm)
TWA: 0.29 mg/m3 (0.1 ppm)
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 243

TABLE 4 (Continued)

Agency/Country and Exposure Limit

Aldehyde CASRN Concentration Method

Ireland
STEL: 0.3 ppm
New Zealand
TWA: 5.7 mg/m3 (2 ppm)
Poland
STEL: 12 mg/m3
TWA: 6 mg/m3
Singapore
TWA: 5.7 mg/m3 (2 ppm)
South Korea
TWA: 6 mg/m3 (2 ppm)
Formaldehyde 50-00-0 ACGIH NIOSH 2016
STEL: 2 ppm
TWA: 0.3 ppm
NIOSH
REL – TWA: 0.016 ppm and 0.1 ppm
ceiling (15 min)
OEHHA
CREL: 9 ug/m3
OSHA
PEL – TWA: 0.75 ppm
Australia
STEL: 2.5 mg/m3 (2 ppm)
TWA: 1.2 mg/m3 (1 ppm)
Austria
STEL: 0.6 mg/m3 (0.5 ppm)
TWA: 0.6 mg/m3 (0.5 ppm)
Belgium
STEL: 0.38 mg/m3 (0.3 ppm)
Canada – Ontario
STEL: 1 ppm
Canada – Quebec
STEL: 3 mg/m3 (2 ppm)
China
STEL: 0.5 mg/m3
Denmark
STEL: 0.4 mg/m3 (0.3 ppm)
TWA: 0.4 mg/m3 (0.3 ppm)
Finland
STEL: 1.2 mg/m3 (1 ppm)
TWA: 0.37 mg/m3 (0.5 ppm)
244 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 4 (Continued)

Agency/Country and Exposure Limit

Aldehyde CASRN Concentration Method

France
STEL: 1 ppm
TWA: 0.5 ppm
Germany (AGS and DFG)
STEL: 0.74 mg/m3 (0.6 ppm)
TWA: 0.37 mg/m3 (0.3 ppm)
Hungary
STEL: 0.6 mg/m3
TWA: 0.6 mg/m3
Ireland
STEL: 2.5 mg/m3 (2 ppm)
TWA: 2.5 mg/m3 (2 ppm)
Japan
TWA: 0.1 ppm
Latvia
TWA: 0.5 mg/m3
Netherlands
STEL: 0.5 mg/m3
TWA: 0.15 mg/m3
New Zealand
STEL: 1 ppm
TWA: 0.33 ppm
Poland
STEL: 1 mg/m3
TWA: 0.5 mg/m3
Singapore
STEL: 0.37 mg/m3 (0.3 ppm)
South Korea
STEL: 1.5 mg/m3 (1 ppm)
TWA: 0.75 mg/m3 (0.5 ppm)
Spain
STEL: 0.37 mg/m3 (0.3 ppm)
Sweden
STEL: 0.74 mg/m3 (0.6 ppm)
TWA: 0.37 mg/m3 (0.3 ppm)
Switzerland
STEL: 0.74 mg/m3 (0.6 ppm)
TWA: 0.37 mg/m3 (0.3 ppm)
United Kingdom
STEL: 2.5 mg/m3 (2 ppm)
TWA: 2.5 mg/m3 (2 ppm)
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 245

TABLE 4 (Continued)

Agency/Country and Exposure Limit

Aldehyde CASRN Concentration Method

Nonyl aldehyde 75718-12-6 N/A N/A

Propionaldehyde 123-38-6 ACGIH NIOSH 2539
TWA: 47.5 mg/m3 (20 ppm)
Belgium
TWA: 48 mg/m3 (20 ppm)
Canada – Ontario
TWA: 20 ppm
Finland
TWA: 48 mg/m3 (20 ppm)
Latvia
TWA: 5 mg/m3
Tolualdehyde 529-20-4; 620-23-5; N/A N/A
104-87-0
Valeraldehyde 110-62-3 NIOSH OSHA 85
REL – TWA: 175 mg/m3 (50 ppm)
Australia
TWA: 176 mg/m3 (50 ppm)
Austria
STEL: 350 mg/m3 (100 ppm)
TWA: 175 mg/m3 (50 ppm)
Belgium
TWA: 179 mg/m3 (50 ppm)
Canada – Ontario
TWA: 50 ppm
Canada – Quebec
TWA: 176 mg/m3 (50 ppm)
Denmark
STEL: 350 mg/m3 (100 ppm)
TWA: 175 mg/m3 (50 ppm)
Finland
TWA: 110 mg/m3 (30 ppm)
France
TWA: 175 mg/m3 (50 ppm)
Ireland
TWA: 176 mg/m3 (50 ppm)
New Zealand
TWA: 176 mg/m3 (50 ppm)
Poland
STEL: 300 mg/m3
TWA: 118 mg/m3
246 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 4 (Continued)

Agency/Country and Exposure Limit

Aldehyde CASRN Concentration Method

Singapore
TWA: 176 mg/m3 (50 ppm)
South Korea
TWA: 175 mg/m3 (50 ppm)
Spain
TWA: 179 mg/m3 (50 ppm)
Switzerland
TWA: 175 mg/m3 (50 ppm)

humidity, and ventilation rate, should be adhered to according to manufacturer speci-

fications. In general, SPF should not be applied when temperatures are <21 C with
relative humidity >80 %. Lower temperatures will slow the chemical reaction and
affect curing, and moisture will react with isocyanates and adversely affect the proper-
ties of the product [1]. Applicators are cautioned to not apply the SPF below the
temperature or above the humidity specified by the manufacturer because these con-
ditions may vary in accordance with climate zones. Under proper precautions, SPF
may be applied under cold conditions if applicators use drum heaters and heat hoses,
in which case manufacturers will also specify product temperature [1,16]. If the sub-
strate is too cold, the foam may not develop properly; if the substrate is too cold or
wet, the foam may crack or fail to adhere properly [58]. Manufacturers provide guid-
ance on substrate temperatures and environmental conditions, including moisture
levels, to ensure adhesion and proper curing. When SPF is applied at a high pressure
or not under ideal temperatures, there is also the potential for overspray and chemical
compound migration from the intended application site. The ratio between Side A
and Side B should adhere to manufacturer guidelines as well [59]. Improper ratios
can affect the curing process, which, in turn, affects the properties of the resulting
foam, causing the foam to be dry and brittle or soft and spongy. The application of
SPF during high humidity, foam thicknesses >5 cm per sprayed layer, and off-ratio
mixtures may lead to adverse emissions of Side B chemicals, such as amine catalysts,
flame retardants, and blowing agents [59]. Manufacturers will specify the thickness of
layers or lifts and the time in between applying layers. If a second lift is applied too
soon, overheating can result in off-gassing for a period of time after installation and
may even result in fire [1,16].
There are also a variety of foam types: medium- and low-density and closed-
and open-cell [1,16]. The application of these different foam types may lead to dif-
ferences in exposure. For example, closed-cell foams trap and lock blowing agents
in the gas bubbles formed during expansion, increasing the insulation capability,
water resistance, and durability of the foam. However, for open-cell SPF, there is a
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 247

TABLE 5 OELs for other chemicals observed [36–39,51–57].

Chemical CASRN Agency/Country and Exposure Limit Concentration Method

1-Chloro-2-propanol 127-00-4 ACGIH N/A

TWA: 1 ppm
Belgium
TWA: 4 mg/m3 (1 ppm)
Canada – Ontario
TWA: 1 ppm
Latvia
TWA: 2 mg/m3
Spain
TWA: 1 ppm
1,2-Dichloropropane 142-28-9 OEHHA N/A
CREL: 462 ppm
Austria
STEL: 1750 mg/m3 (375 ppm)
TWA: 350 mg/m3 (75 ppm)
Spain
STEL: 517 mg/m3 (110 ppm)
TWA: 325 mg/m3 (75 ppm)
1,4-Dioxane 123-91-1 OSHA NIOSH 1602
PEL – TWA: 360 mg/m3 (100 ppm) for dermal
NIOSH
REL – 3.6 mg/m3 (1 ppm) ceiling (30 min)
OEHHA
CREL: 3,000 ppm
Australia
TWA: 36 mg/m3 (10 ppm)
Austria
STEL: 146 mg/m3 (40 ppm)
TWA: 73 mg/m3 (20 ppm)
Belgium
TWA: 73 mg/m3 (20 ppm)
Canada – Ontario
TWA: 20 ppm
Canada – Quebec
TWA: 72 mg/m3 (20 ppm)
China
TWA: 70 mg/m3
Denmark
STEL: 72 mg/m3 (20 ppm)
TWA: 36 mg/m3 (10 ppm)
European Union
TWA: 73 mg/m3 (20 ppm)
248 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 5 (Continued)

Chemical CASRN Agency/Country and Exposure Limit Concentration Method

Finland
STEL: 150 mg/m3 (40 ppm)
TWA: 36 mg/m3 (10 ppm)
France
STEL: 140 mg/m3 (40 ppm)
TWA: 73 mg/m3 (20 ppm)
Germany (AGS and DFG)
STEL: 146 mg/m3 (40 ppm)
TWA: 73 mg/m3 (20 ppm)
Hungary
STEL: 10 mg/m3
TWA: 10 mg/m3
Ireland
TWA: 73 mg/m3 (20 ppm)
Italy
TWA: 73 mg/m3 (20 ppm)
Japan
TWA: 10 ppm
Latvia
TWA: 20 mg/m3
Netherlands
TWA: 20 mg/m3
New Zealand
TWA: 90 mg/m3 (25 ppm)
Poland
TWA: 50 mg/m3
Singapore
TWA: 90 mg/m3 (25 ppm)
South Korea
TWA: 72 mg/m3 (20 ppm)
Spain
TWA: 74 mg/m3 (20 ppm)
Sweden
STEL: 90 mg/m3 (25 ppm)
TWA: 35 mg/m3 (10 ppm)
Switzerland
STEL: 144 mg/m3 (40 ppm)
TWA: 72 mg/m3 (20 ppm)
United Kingdom
STEL: 366 mg/m3 (100 ppm)
TWA: 91 mg/m3 (25 ppm)
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 249

TABLE 5 (Continued)

Chemical CASRN Agency/Country and Exposure Limit Concentration Method

Chlorobenzene 108-90-7 OSHA NIOSH 1003

PEL – TWA: 350 mg/m3 (75 ppm)
OEHHA
CREL: 1,000 ppm
Australia
TWA: 46 mg/m3 (10 ppm)
Austria
STEL: 70 mg/m3 (15 ppm)
TWA: 23 mg/m3 (5 ppm)
Belgium
STEL: 70 mg/m3 (15 ppm)
TWA: 23 mg/m3 (5 ppm)
Canada – Ontario
TWA: 10 ppm
Canada – Quebec
TWA: 230 mg/m3 (50 ppm)
China
TWA: 50 mg/m3
Denmark
STEL 92 mg/m3 (20 ppm)
TWA: 46 mg/m3 (10 ppm)
European Union
STEL: 70 mg/m3 (15 ppm)
TWA: 23 mg/m3 (5 ppm)
Finland
STEL: 70 mg/m3 (15 ppm)
TWA: 23 mg/m3 (5 ppm)
France
STEL: 70 mg/m3 (15 ppm)
TWA: 23 mg/m3 (5 ppm)
Germany (AGS and DFG)
STEL: 94 mg/m3 (20 ppm)
TWA: 47 mg/m3 (10 ppm)
Hungary
TWA: 23 mg/m3
Ireland
STEL: 70 mg/m3 (15 ppm)
TWA: 23 mg/m3 (5 ppm)
Italy
STEL: 70 mg/m3 (15 ppm)
TWA: 23 mg/m3 (5 ppm)
250 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 5 (Continued)

Chemical CASRN Agency/Country and Exposure Limit Concentration Method

Japan
TWA: 10 ppm
Latvia
STEL: 70 mg/m3 (15 ppm)
TWA: 23 mg/m3 (5 ppm)
Netherlands
TWA: 23 mg/m3
New Zealand
TWA: 46 mg/m3 (10 ppm)
Poland
STEL: 70 mg/m3
TWA: 23 mg/m3
Singapore
TWA: 46 mg/m3 (10 ppm)
South Korea
STEL: 94 mg/m3 (20 ppm)
TWA: 46 mg/m3 (10 ppm)
Spain
STEL: 70 mg/m3 (20 ppm)
TWA: 23 mg/m3 (5 ppm)
Sweden
STEL: 70 mg/m3 (20 ppm)
TWA: 23 mg/m3 (5 ppm)
Switzerland
STEL: 92 mg/m3 (20 ppm)
TWA: 46 mg/m3 (10 ppm)
United Kingdom
STEL: 3 ppm
TWA: 1 ppm
Hexamethylcyclotrisiloxane 541-05-9 N/A N/A
Isopropyl alcohol 67-63-0 ACGIH OSHA 109
STEL: 1,225 mg/m3 (500 ppm)
NIOSH
REL – TWA: 980 mg/m3 (400 ppm)
OEHHA
CREL: 7,000 ppm
OSHA
PEL – TWA: 980 mg/m3 (400 ppm)
Australia
STEL: 1,230 mg/m3 (500 ppm)
TWA: 983 mg/m3 (400 ppm)
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 251

TABLE 5 (Continued)

Chemical CASRN Agency/Country and Exposure Limit Concentration Method

Austria
STEL: 2,000 mg/m3 (800 ppm)
TWA: 500 mg/m3 (200 ppm)
Belgium
STEL: 1,000 mg/m3 (400 ppm)
TWA: 500 mg/m3 (200 ppm)
Canada – Ontario
STEL: 400 ppm
TWA: 200 ppm
Canada – Quebec
STEL: 1,230 mg/m3 (500 ppm)
TWA: 983 mg/m3 (400 ppm)
China
STEL: 700 mg/m3
TWA: 350 mg/m3
Denmark
STEL: 980 mg/m3 (400 ppm)
TWA: 490 mg/m3 (200 ppm)
Finland
STEL: 620 mg/m3 (250 ppm)
TWA: 500 mg/m3 (200 ppm)
France
STEL: 980 mg/m3 (400 ppm)
Germany (AGS and DFG)
STEL: 1,000 mg/m3 (400 ppm)
TWA: 500 mg/m3 (200 ppm)
Hungary
TWA: 500 mg/m3
Ireland
STEL: 400 ppm
TWA: 200 ppm
Japan
TWA: 400 ppm
Latvia
TWA: 350 mg/m3
New Zealand
STEL: 1,230 mg/m3 (500 ppm)
TWA: 983 mg/m3 (400 ppm)
Poland
STEL: 1,200 mg/m3
TWA: 900 mg/m3
252 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 5 (Continued)

Chemical CASRN Agency/Country and Exposure Limit Concentration Method

Singapore
STEL: 1,230 mg/m3 (500 ppm)
TWA: 983 mg/m3 (400 ppm)
South Korea
STEL: 980 mg/m3 (400 ppm)
TWA: 480 mg/m3 (200 ppm)
Spain
STEL: 1,000 mg/m3 (400 ppm)
TWA: 500 mg/m3 (200 ppm)
Sweden
STEL: 600 mg/m3 (250 ppm)
TWA: 350 mg/m3 (150 ppm)
Switzerland
STEL: 1,000 mg/m3 (400 ppm)
TWA: 500 mg/m3 (200 ppm)
United Kingdom
STEL: 1,250 mg/m3 (500 ppm)
TWA: 999 mg/m3 (400 ppm)
Methylpropanamine 78-81-9 Denmark N/A
STEL: 15 mg/m3 (5 ppm)
TWA: 15 mg/m3 (5 ppm)
Germany (DFG)
STEL: 12.2 mg/m3 (4 ppm)
TWA: 6.1 mg/m3 (2 ppm)
Sweden
STEL: 15 mg/m3 (5 ppm)
Switzerland
STEL: 12.2 mg/m3 (4 ppm)
TWA: 6.1 mg/m3 (2 ppm)
Octamethylcyclotrisiloxane 556-67-2 N/A N/A
Phenoxyethanol 122-99-6 N/A N/A
Propylene glycol 57-55-6 Australia OSHA
PV2051
TWA: 474 mg/m3 (150 ppm)
Canada – Ontario
TWA: 155 mg/m3 (50 ppm)
Ireland
TWA: 470 mg/m3 (150 ppm)
New Zealand
TWA: 474 mg/m3 (150 ppm)
United Kingdom
TWA: 474 mg/m3 (150 ppm)
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 253

potential concern for exposure to free isocyanates if trimmed very soon after instal-
lation. Unlike closed-cell foam, open-cell foam is known for its rapid expansion
and typically requires applicators to trim these materials. Trimming operations typ-
ically occur shortly after application [60]. The blowing agent gas for open-cell
foam, typically water, is not trapped by the forming cells and is released to the at-
mosphere during foam expansion, and spaces within the cells are filled with atmo-
spheric air [1]. On the other hand, blowing agents for closed-cell foam can include
a variety of chemicals to which workers may be exposed [61].

Ventilation Techniques for Worker Safety

When Applying SPF
Most SPF manufacturers recommend that applicators wear PPE when installing
SPF, and proper ventilation and controls must also be implemented to protect the
health of trade workers and building occupants. Employing ventilation techniques
during application, such as using local and general ventilation systems, can protect
worker health. Proper ventilation from efficiently supplied fresh air to the applica-
tion area also must be implemented after installation to protect the health of other
trade workers and building occupants from potentially harmful emissions of SPF
chemicals (Fig. 1).
General exhaust ventilation systems typically consist of an exhaust fan that
pulls air out of the room and discharges the air outdoors. General exhaust ventila-
tion, also known as “dilution ventilation,” dilutes contaminated air by mixing the
contaminated air with cleaner room air. However, general exhaust systems are not

FIG. 1 Diagram of air flow [62].

254 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

recommended as the sole source of ventilation when hazardous vapors or mists are
present because, if local exhaust ventilation is not feasible, general exhaust systems
do not immediately remove contaminants. For an SPF installation, a local exhaust
system would be appropriate during the application process to capture vapors,
mists, dusts, and particulates. A particulate filter would be appropriate at the source
because they are emitted during the spraying and trimming processes. When
designing a ventilation system, it is important to consider these factors: (1) the
work space, room, or enclosure to be exhausted; (2) the exhaust system, including a
hood or exhaust vent, ductwork, and fan that captures contaminants at the source
and transports them to a location outside the building away from HVAC air intakes
and occupied areas; and (3) getting fresh air to the room or work area to replace the
exhausted air. Ventilation combined with the use of PPE will result in less worker
exposure to the SPF chemical mixture. After application, general ventilation is ap-
propriate to ventilate the entire work area and building before other workers or
building occupants enter the area; however, other workers entering the space may
still need PPE to prevent exposure [62–65].
Carefully observing best practices and controls, as recommended by the EPA,
to prevent exposures during SPF installation in commercial or residential settings is
paramount because application occurs on-site [62]. The use of proper PPE, contain-
ment, ventilation, and proper clean-up practices are also important in these set-
tings. The polyurethane industry developed worker studies that evaluated the use of
PPE during the application of SPF insulation in a variety of residential settings
[63–65]. Air samples showed that MDI exceeded its OEL, as described in Table 1
[64]. This study also showed that MDI and amines could migrate to other floors
during application [65]. Another industry study evaluated a 25-ft perimeter for PPE
and discovered MDI readings beyond 25 ft, which is important in managing expo-
sures to other workers [66]. These studies directly enforce the need for trade work-
ers and building occupant guidance during SPF installation, specifically vacating
the premises during the application of SPF and to inform reoccupancy and reentry
timelines [36,63–65].

Sampling and Analytical Methods Used to

Measure Isocyanates During Application
Sampling and analytical methods are used for isocyanates to assess compliance with
OELs, to evaluate the effectiveness of controls, and to study the relation between ex-
posure and health effects. Several established sampling and analytical methods for
isocyanates are presented in Table 6.
The accurate measurement of isocyanates relies on implementing appropriate
methodology for both sampling and analysis [77]. Almost all established isocyanate
methods rely on derivatization of the isocyanate with an amine reagent. Derivatiza-
tion of isocyanates at the time of sampling is important because isocyanates are
highly reactive and require stabilization. In addition, often the derivatizing reagent
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 255

TABLE 6 Air sampling methods for isocyanates [67–76].

Method Sampler(s) Reagent Analysis

ISO 16702 (MDHS25/4) GFF or impinger or MP LC-UV-EC

impinger þ GFF
LC-UV-EC or LC-MS-MS
ISO 17734-1 Impinger DBA LC-MS-MS
ISO 17735 (NIOSH 5525) GFF or impinger or MAP LC-UV-FL
impinger þ GFF
LC-MS-MS
ISO 17736 PTFE filter þ GFF MP þ MAMA LC-UV and LC-FL
OSHA 47 GFF PP LC-UV or LC-FL
Bayer Material Science 1.20.1 Impinger PP LC-UV or LC-FL
ASSET EZ4-NCO sampler Denuder/GFF DBA LC-MS-MS
CIP10M sampler CIP10M MP LC-PDA
DAN method Impinger DAN LC-MS-MS

Note: DAN ¼ 1,8-diaminonaphthalene; DBA ¼ dibutylamine; EC ¼ electrochemical detection;

FL ¼ fluorescence detection; GFF ¼ glass fiber filter; LC ¼ liquid chromatography; MAMA ¼
9-(methylaminomethyl)anthracene; MAP ¼ 1-(9-anthracenylmethyl)piperazine; MP ¼ 1-(2-
methoxyphenyl)piperazine; MS ¼ mass spectrometry; PDA ¼ photodiode array; PP ¼ 1-(2-pyridyl)-
piperazine; PTFE ¼ polytetrafluoroethylene; UV ¼ ultraviolet.

is what makes the derivatized isocyanate detectable. In the case of measuring

airborne isocyanates in vapor form, derivatization at the time of sampling enables
collection by converting a volatile or semivolatile isocyanate to a nonvolatile deriva-
tive. Air sampling of isocyanates generally uses a reagent-coated glass fiber filter
(GFF) or an impinger containing a solution of a derivatizing reagent in an organic
solvent. Reagent-coated filters are much more convenient to use in the field, espe-
cially for personal sampling. However, the choice of sampler is greatly influenced
by the type of isocyanate being measured, how that isocyanate is being used, and
the physical state of the isocyanate in air. Although impingers and GFFs are effec-
tive in sampling isocyanate vapors, GFFs are the sampler of choice for sampling a
vapor-only atmosphere. Such atmospheres include most operations using relatively
volatile isocyanates, such as toluene diisocyanate (TDI), as well as the measurement
of off-gassing from partially cured or cured materials, such as SPF. Processes in-
volving heating or spraying may result in isocyanates being present in the air in the
form of aerosols. Although GFFs collect these aerosols efficiently, the reagent on
GFFs is not effective at derivatizing isocyanates present in droplets or particles,
which may result in a measurement that has a low bias [78]. However, reagent-
coated GFFs cannot efficiently derivatize isocyanates present in fast-curing aerosols,
such as those generated in the application of SPF [79]. Efficient derivatization of
fast-curing aerosols typically requires sampling with impingers. A novel isocyanate
air sampler has recently become commercially available. The ASSET EZ4-NCO
sampler is a reagent-impregnated denuder followed by a reagent-coated filter [74].
256 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

The denuder portion serves as a collector of isocyanate vapors and a source of re-
agent to derivatize aerosols collected on the downstream filter. Evaluations are un-
derway to test the effectiveness of the sampler to derivatize unstable isocyanate
aerosols. Finally, the CIP10 sampler recently has been adapted for sampling of MDI
aerosols [75]. The sampler uses a small volume of nonvolatile solvent to facilitate
derivatization of reacting aerosols.
Although most sampling methodology for isocyanates has focused on air
sampling, sampling of surfaces is also important given the potential for dermal
sensitization and the potential for surfaces to serve as a source for airborne isocya-
nates. Sampling of surfaces is generally accomplished by wipes that are then im-
mediately extracted in solutions of derivatizing reagents used in air sampling
methods [80]. Dermal sampling has been accomplished in research settings using
skin wipes, tape stripping, and interception techniques [80–82]. Once an environ-
mental sample has been collected and the isocyanates derivatized, the analysis of
isocyanates is typically accomplished using liquid chromatography with ultravio-
let (UV), fluorescence, electrochemical, or mass spectrometric detection. When
very low levels are anticipated, or when the sample matrix is particularly complex
(e.g., surface sampling or dermal sampling), mass spectrometric detection is rec-
ommended. Some methods measure oligomeric isocyanates in addition to diiso-
cyanate monomers, even though analytical standards are not generally available
for oligomeric species (NIOSH Method 5525; UK Health and Safety Executive’s
Methods for the Determination of Hazardous Substances Guidance 25/4). Indi-
vidual isocyanate species in the sample produce peaks in the chromatogram, and
these peaks can be summed to give a measure of total reactive isocyanate group
(TRIG), which is the basis of the isocyanate OEL in some countries [15]. In SPF,
the MDI monomer is the major isocyanate species in the formulation. However,
the fast-curing nature of the SPF means that the MDI monomer and other
oligomers present in the original isocyanate formulation are rapidly converting to
intermediate isocyanate species [76]. Established chromatographic methods are
not suited for measuring these intermediate species.
The summation of many peaks in a chromatogram is labor-intensive for sam-
ples containing complex isocyanate mixtures. Moreover, not all isocyanate species
in a sample elute from the column or produce a measurable peak in the chromato-
gram. A sampling and analytical method based on derivatization with the reagent
1,8-diaminonaphthalene (DAN) is fundamentally different from conventional
chromatographic methods [76]. Like conventional methods, isocyanate species re-
act with DAN upon sampling. However, rather than analyzing this derivatized iso-
cyanate mixture, the mixture is treated with acid to catalyze an intramolecular
cyclization that converts all DAN-derivatized isocyanate groups to a single molecule
of 1H-perimidin-2(3H)-one (perimidone). This reaction is shown in Fig. 2.
The DAN method has been shown to measure isocyanate groups accurately in
aromatic isocyanate products (MDI- and TDI-based products) [84]. The method is
particularly suitable for measuring isocyanates in SPF because of the rapid curing of
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 257

FIG. 2 Reaction of DAN with isocyanates to yield the analyte perimidone [83].

this material and the generation of intermediate isocyanate species [76]. In a simu-
lated spray of SPF insulation, side-by-side air sampling using NIOSH Method 5525
impingers and DAN impingers found approximately tenfold higher TRIG levels
with the DAN impinger [76]. Future investigation with the DAN method will in-
clude measuring residual isocyanates on surfaces and in dust, distinguishing
between extractable and chemically bound isocyanates and extending/rating the
curing of SPF.
There are also direct-reading methods for isocyanates. Colorimetric paper tape
monitors measure isocyanate vapors as air is drawn through a filter paper contain-
ing a reagent [85]. The collection of isocyanate vapors results in a color change to
the paper, enabling detection and measurement. These monitors can set off an
alarm when a particular air level is detected, such as that corresponding to an OEL.
Paper tape technology is useful for detecting isocyanate vapors; however, it is not
suitable in low- or high-humidity conditions or for measuring aerosols [80,86]. The
detection limit for paper tapes is higher than the detection limit of chromatographic
methods. For the immediate detection of isocyanates on surfaces, SWYPE is a pro-
priety qualitative wipe-test kit specific to isocyanates and is commonly used by the
industry for testing [86].

Air and Wipe Sampling for Amine Catalysts

There are also some recommended methods for sampling and analyzing the repre-
sentative amine catalysts. These sampling and analytical methods for amine
258 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

catalysts are used to assess compliance with OELs established by OSHA and
NIOSH to understand the relation between exposure and health effects. For in-
stance, OSHA Method PV2060 recommends air samples to be drawn through a 10 %
phosphoric acid-coated XAD-7 tube and desorbed and analyzed by gas chromatogra-
phy using a flame ionization detector [87]. For triethanolamine collection, OSHA
Method PV2141 directs that samples be collected by drawing a known volume of air
through a glass fiber filter. Then the sample is extracted with 2 mL of acetone and an-
alyzed by gas chromatography using a flame ionization detector [88]. Likewise,
OSHA Method PV2018 for diethanolamine instructs that air samples be collected by
drawing a volume of air through tubes containing XAD-2 resin coated with 10 % 1-
naphthylisothiocyanate by weight. Samples are then analyzed by desorbing the
adsorbent with dimethylformamide and quantitating the amine derivative by high-
performance liquid chromatography using UV detection [89]. For N-ethylmorpholine,
OSHA recommends using NIOSH Method S-264, in which the air sample is drawn
through a tube containing XAD-2 resin for absorption, desorbed with carbon disul-
fide, and injected into a gas chromatograph [90]. The purpose of a skin notation, as
applied by some existing OEL-setting organizations, is to provide a warning that a
chemical has the potential to be absorbed dermally in sufficient quantities to affect
the interpretation of the risks from inhalation if concomitant dermal contact may
occur. Numerous governmental and nongovernmental organizations (see Table 2)
include skin notations as part of their occupational exposure guidelines for chemi-
cals. The criteria and protocol for the assignment of skin notations vary among the
different organizations, but the systems share the general interpretation that a skin
notation represents the potential for dermal absorption [91].

Air Sampling for Blowing Agents

Because no regulatory limits exist for blowing agents, only a few OSHA or NIOSH
air-sampling techniques have been developed to protect workers from inhaling the
blowing agents. However, HFC-245fa does have associated sampling and analytical
procedures used to detect worker exposures based on the AIHA’s WEEL. Examples
of methods for monitoring workplace emissions of HFC-245fa include the use of
partially evacuated steel cylinders, known as Vacu-Samplers, portable halogen
detectors, diffusive air samplers, charcoal tubes, and thermal desorption tubes.
Vacu-Samplers are steel canisters with nonabsorbing interiors. The Vacu-Sampler
is placed in the area to be sampled and takes a 3- to 5-s air sample commonly
referred to as a “grab sample.” The sample is then analyzed using gas chromatogra-
phy. Halogen and electron capture detectors may also be used to detect gases by us-
ing a near-infrared or photoionization detector. To measure worker exposure, a
diffusive air sampler may be clipped to the worker with a sampling media for work-
place monitoring [60]. In addition, the SPF industry and refrigerant manufacturers
are analyzing emissions rates through modeling [92]. According to the SPF indus-
try, some portion of the blowing agent is retained in the cellular structure of the
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 259

foam and diffuses to the environment over time. The manufacturers of refrigerants
have studied the effects of blowing agents through a diffusion-based model, and
the limited field validation data suggest this model has the potential to provide
reasonable estimates of emissions over time. The model may also provide a meth-
od for estimating diffusion parameters based upon physical properties for other
refrigerants [25,92].

Air Sampling for Aldehydes

Sampling methods for aldehydes, identified as a potential reaction product of SPF
installation, are used to assess compliance with OELs and to study worker exposure
and health effects. For acetaldehyde, OSHA Method 68 is used to measure this
common indoor air pollutant by drawing samples of air through a sampling tube
containing XAD-2 adsorbent that has been coated with 2-(hydroxymethyl)piperi-
dine. The tube contains a 450-mg sampling section and a 225-mg backup section
desorbed with toluene, and derivatized acetaldehyde is analyzed by gas chromatogra-
phy using a nitrogen-selective detector [93]. For butyraldehyde, propionaldehyde,
and tolualdehyde, NIOSH Method 2539 is used by drawing air through a solid
sorbent tube containing XAD-2 adsorbent coated with 10 % 2-(hydroxymethyl)-
piperidine and analyzed by gas chromatography [94]. For crotonaldehyde, OSHA
Method 81 is used by drawing air through an open-face cassette containing two glass
fiber filters, each coated with 2,4-dinitrophenylhydrazine (DNPH) and phosphoric
acid. The sample is then extracted with acetonitrile and analyzed by high-
performance liquid chromatography (HPLC) using a UV detector [95]. NIOSH
Method 2016 is commonly used to measure formaldehyde in air samples pulled
through a cartridge containing silica gel coated with DNPH. The sample is extracted
with acetonitrile and analyzed by HPLC using a UV detector [96]. Valeraldehyde is
measured by using OSHA Method 85, which specifies that air be drawn through an
open-face air-monitoring cassette containing three glass fiber filters. Each filter is
coated with 2,4-dinitrophenylhydrazine and phosphoric acid. The filters are
extracted with acetonitrile and analyzed by HPLC using a UV detector [97].

Air and Wipe Sampling for Other Compounds

Observed
Because of the known health effects from the observed chemicals associated with
SPF installation, sampling and analytical methods are used to assess compliance
with OELs. The sampling procedures for 1,4-dioxane include following NIOSH
Method 1602 by drawing air through a charcoal tube with carbon disulfide desorp-
tion (100/50-mg sections; 20/40 mesh) and analyzing by a gas chromatograph. 1,4-
Dioxane may also be sampled using a charcoal pad and a glass vial for shipment
[48]. Chlorobenzene can be sampled by following NIOSH Method 1003 by also
drawing air through a charcoal tube with carbon disulfide desorption (100/50-mg
260 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

sections; 20/40 mesh) and analysis by gas chromatography [49]. Isopropyl alcohol
can be sampled by OSHA Method 109 by drawing air through two Anasorb 747
tubes (400 mg for the first and 200 mg for the second) and desorbed with a mixture
of 60:40 dimethylformamide/carbon disulfide and analyzed by gas chromatography.
Isopropyl alcohol may also be sampled on-site with the use of detector tubes [98].
For propylene glycol, OSHA Method PV2051 is used by drawing air through an
OSHA Versatile Sampler (OVS-7) 13-mm glass fiber filter followed by a XAD-7
tube (270/140-mg sections; 20/60 mesh). Then it is extracted with methanol and
analyzed using gas chromatography [50].

Industrial Hygiene Studies Estimating Worker

Exposure During and After SPF Application
Recent IH studies have used these established sampling and analytical methods,
when available, to estimate worker exposure from chemicals of concern during and
after application [36,61,64]. It should be noted, however, that these IH studies are
representative of the potential exposures following manufacturer guidance for SPF
application and are not representative of the generation of off-ratio SPF, misapplied
SPF, or the potential for emissions from installed SPF over time. One study was
conducted to estimate the effectiveness of ventilation and the 24-h reentry time-
frames for other trades for residential homes sprayed with closed- and open-cell
foams [36]. Emissions from both were analyzed for MDI, blowing agents (such as
HFC-245fa), aldehydes, volatile organic compounds (VOCs), and amine catalysts.
Samples were taken during and after application for 4, 12, and 24 h to determine
possible exposures to workers and other trades. This study concluded that ventila-
tion and proper PPE for the applicator and possibly other trade workers are vital
for protection during and for a time after application. Based on these concerns for
workers, some companies suggest a 24-h timeframe for residents before reoccupy-
ing their residence and longer timeframes for sensitive populations, such as children
and the elderly [36]. Another study analyzed emissions from an array of chemicals
in both closed- and open-cell SPF, including MDI, catalysts (dimethylethanolamine
and N,N,N0 ,-trimethylaminoethylethanolamine), flame retardants (TEP and TCPP),
blowing agents (HFC-245fa), and VOCs through field sampling, spray booths, and
small environmental chambers [61]. Samples were taken before, during, and after
application. In general, this study concluded that levels of MDI, amines, and flame
retardants decrease after application; blowing-agent chemicals may not pose a
health risk to workers via inhalation; and, based on the type of foam, a range of lev-
els were observed for total VOCs per EPA Method TO-15 [61]. A third study ana-
lyzed three generic formulations, including a low-density high-pressure, medium-
density high-pressure, and low-pressure kit formulation with a ventilation rate of
10 ACH. Samples were taken during application and 30 min after application to
measure emissions of MDI, amine catalysts, blowing agents, and flame retardants.
This study concluded that (1) there were no measureable airborne concentrations
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 261

of MDI, (2) there were low concentrations of TCPP, (3) concentrations of blowing
agents were within the AIHA WEEL, and (4) most amine concentrations were
within OELs except for bis(2-dimethylaminoethyl) ether, which was above the
ACGIH TLV [64]. All studies concluded that ventilation and proper PPE for the
applicator and possibly other trade workers are vital for protection during and after
application [36,61,64]. For real-time monitoring, paper tape instruments are typical-
ly used to monitor airborne isocyanates. However, isocyanate aerosols, particularly
rapidly curing aerosols generated in SPF, cannot be measured using paper tape tech-
nology. A manufacturer of paper tape instruments suggests these instruments should
only be used as a rough indication of polymeric MDI airborne concentration [99].

Estimating Nonoccupational Indoor Air

Exposures
Residents have SPF applied in their homes, but many residents are not aware that
SPF contains diisocyanates, flame retardants, amines, and other chemicals that may
cause adverse health effects if SPF is not applied correctly, does not cure properly,
or if there is inadequate ventilation. Residential exposures are not well characterized
but are expected to vary with the product type and quantity used, application meth-
od, particle size distribution, and end-product curing time [100]. Even after SPF ful-
ly cures, gases may be released into the home, resulting in inhalation. Limited
information exists on what factors, such as temperature, humidity, and application
technique, affect curing times and safe reoccupancy time after application. Howev-
er, advances in indoor air quality focus on the techniques used to measure and
model emissions of indoor air chemical contaminants. Small and large environmen-
tal test chambers have been adopted and used for characterizing emissions from in-
door materials and products [101–110]. Exposures to emissions from chemicals in
SPF and their byproducts include a range of VOCs, such as aldehydes, and semivo-
latile organic compounds (SVOCs), such as flame retardants and amines. Because
SVOCs tend to bind to interior surfaces and particulate matter, indoor exposure to
these chemicals can take place via multiple routes, including ingestion of dust and
inhalation of emissions [100]. For example, TCPP is an additive flame-retardant
chemical that is not bound to SPF and, over time, will diffuse out of SPF and into
the surrounding environment. It has a high molecular weight, low vapor pressure,
and high octanol-water partition coefficient [111–113]. These properties determine
the fate of TCPP in the environment, which affects the potential for human expo-
sure. Therefore, TCPP will readily partition into and be present in organic sub-
strates, such as dust, which can then be inhaled, ingested, or dermally contacted
[114,115].
Modeling of the emissions, transport, sorption, and distribution of SVOCs in
the indoor environment has the potential to cover a wide range of indoor media, in-
cluding air, sources, sinks, contaminant barriers, suspended particles, and settled
particles. Input parameters may include solid-air partition coefficients, solid-phase
262 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

diffusion coefficients, gas-phase mass transfer coefficients, and initial SVOC con-
centrations in indoor media [100]. To estimate residential exposure potential to
SPF chemicals, chamber test methods and SVOC modeling approaches should in-
clude the chemicals in the formulated mixture, byproducts, and other observed
chemicals. These emission measurements can then be used to determine the overall
magnitude and trends of chemical emissions over time to help inform safe reentry
for workers and reoccupancy after application.

Residential/Commercial Building Inhabitant

Exposures and Modeling
Application of SPF in a building may lead to residential exposure to chemicals in
the SPF, such as MDI, flame retardants, and the amine catalysts, particularly if in-
dustry guidelines regarding workplace isolation and reoccupancy are not followed.
For instance, a study assessed two building occupants who reentered their residence
4 h after application. The residents were diagnosed with reactive airway dysfunction
syndrome, attributed by the authors as exposure to MDI [37]. Likewise, amine cata-
lysts may continue to be emitted from the installed foam and cause health effects.
Health effects associated with amine exposure include ocular, dermal, and respiratory
tissue irritation. The Consumer Product Safety Commission (CPSC) has received a
number of incident reports from residents after SPF was installed in their residences.
These complaints include lingering odors, respiratory-related problems, irritation,
and headaches, which are similar to health effects associated with amine exposure
[17]. Building inhabitants may also be exposed to blowing agents over time. As the
foam ages, the blowing agent may leak from the foam and expose occupants to very
low levels, which may or may not present a significant toxicity risk [25].
The multitude of SPF formulations results in varying manufacturer specifica-
tions for application and cure times [61]. Furthermore, product testing by the man-
ufacturers and review by regulatory agencies and other professional organizations,
such as the AIHA and ACGIH, should be undertaken to understand potential SPF
exposure to residential/commercial building inhabitants for product safety [80].
Available data focus on workplace studies that cannot be easily extrapolated for res-
idential exposures. There may be aspects of these studies in which do-it-yourself
applicators are applying polyurethane materials that are equivalent to the activities
of a professional building contractor using the same products. However, there are a
lack of studies focused on residential use, including data on indoor air emissions of
polyurethane products [80].
Indoor air modeling for SPF emissions in a building could provide screening-
level estimates of dose and potential exposure to residential/commercial building
inhabitants. Before modeling can occur, emissions data need to be determined from
samples collected from an on-site SPF application. These samples can be placed in
environmental chambers, particularly micro- or small chambers, to provide emis-
sions data that can then be used for calculations of possible air concentrations.
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 263

However, these samples are seldom collected over a long timeframe (months) and
often do not include important information, such as the application rate, amount of
foam applied, thickness, and air exchange rates. Modeling emissions could lead to a
better understanding of the emissions over time, initial chemical concentrations,
diffusion coefficients, and how these parameters are influenced by environmental
conditions. In addition, information on interzonal air flows between conditioned
and unconditioned spaces is needed to determine how chemicals are transported
within buildings over time, and this information can also be useful for determining
whether ventilation may be a way to isolate an SPF work area from the rest of the
building during application. Given the wide variety of possible SPF application sce-
narios, a model should consider the following important parameters: SPF formula-
tion, loading rates, air exchange rates, and mechanisms for interzonal transfer of air
particulates within buildings. Other parameters may also be identified. It is also crit-
ical for a model to consider the role of dust, gases, and vapors as sources of expo-
sure. An adequately parameterized model could serve to analyze the impact of
sources, sinks, and ventilation [1,16,116,117].

Environmental Chamber Testing to

Estimate Exposure
Environmental chamber testing is used under controlled conditions to estimate ex-
posure and can be used for modeling. For SPF testing, samples are placed in a
chamber and air is passed over the surface to measure chemical emissions under
defined environmental parameters. Chamber tests are intended to determine spe-
cific emission rates of chemicals during specified constant conditions for tempera-
ture, relative humidity, ventilation rates, and product loading factors. Current
chamber testing standards are used for polyurethane and test for VOCs, which con-
sist of a broad range of organic chemicals that vaporize at room temperature. VOCs
contained in these polyurethane products tend to diffuse to the surface and emit
into the air over time, potentially exposing occupants and other receptors via
inhalation.
To reduce exposure to VOCs, the California Department of Health Services de-
veloped the “Standard Method for the Testing and Evaluation of Volatile Organic
Chemical Emissions from Indoor Sources Using Environmental Chambers, Version
1.1,” under California Section 01350. This standard is based on the California refer-
ence exposure level for a list of VOCs, including aldehydes and 1,4-dioxane. Indus-
try and third-party testing laboratories use this method to comply with building
code standards, including the American Society of Heating, Refrigerating and Air
Conditioning Engineers (ASHRAE) Standard 189.1, Standard for the Design of
High-Performance Green Buildings, and product certification programs, such as the
GREENGUARD indoor air quality certification, Leadership in Energy and Environ-
mental Design indoor environmental quality criteria, and the National Green Build-
ing Standard. California Standard 01350, General Requirements–Special Project
264 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

Procedures, is widely applied by the industry because it is a flexible, relatively low-

cost, and health-based standard. It has been updated to provide better modeling
parameters, reference to other comparable standards for specific product types, and
a lowered concentration limit for formaldehyde of 9.0 lg/m3. The standard speci-
fication requires that samples be taken within 24 h of product installation and
that appropriate conditioning is used [14]. Guidance from ASTM D5116-06,
Standard Guide for Small-Scale Environmental Chamber Determinations of
Organic Emissions from Indoor Materials/Products [118], and ISO 16000-9:2006,
Indoor Air—Part 9: Determination of the Emission of Volatile Organic Compounds
from Building Products and Furnishing—Emission Test Chamber Method, was
incorporated into this California standard. Modeling techniques using emission
rates relative to chronic health guidelines, as determined by the California Office
of Environmental Health and Hazard Assessment, are preferred for use in the
California standard. This standard method is used for a variety of polyurethane
products applied in buildings, including SPF, flooring, and paints and coatings
[14]. However, current methods do not validate for MDI, amine catalysts, or
flame retardants emitted from SPF [119].
In addition, a method similar to the California standard has been developed in
Canada: Underwriters Laboratories of Canada document “Standard Laboratory
Guide for the Determination of Volatile Organic Compound Emissions from Poly-
urethane Foam” (CAN/ULC-S774-09). This method uses a two-dynamic-chamber
procedure to characterize emissions starting 24 h after installation and continuing
to Day 30 to determine a time period previously defined as “airing out time” and
now defined as “time to occupancy.” However, like the California standard, the Ca-
nadian standard also does not address nontarget VOCs, unreacted isocyanates,
amine and other catalysts, flame retardants, and many of the blowing agents [120].
Since 2010, ASTM Subcommittee D22.05 on Indoor Air has been working col-
laboratively to create standardized methods for measuring chemical emissions from
SPF insulation. D22.05 is focused primarily on developing methods for measuring
SPF emissions using micro- and small-scale chambers. These methods could be
used for screening for emissions from SPF in residential or commercial buildings
and for identifying the production of off-specification foam. The final and proposed
standards considered by Subcommittee D22.05 include ASTM D7859, WK43872,
WK46527, WK40292, and WK40293. The list of proposed and final standards rep-
resents the issues raised by the committee and the range of compounds of interest
that include isocyanates, blowing agents, amine catalysts, flame retardants, and
aldehydes. The ongoing work of ASTM D22.05 aims to develop consensus methods
to address fundamental issues such as what is emitted, what the time course of
emissions is, what factors affect emissions and potential exposures, and how useful
emission factors developed in microchambers might be for predicting concentra-
tions in full-scale environments. [121,122].
Although the measurement of VOCs in chambers has been standardized, the
measurement of amines in chambers requires special sampling and analytical
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 265

considerations. Chemical-specific emission rates standardized by mass or surface

area may need to be considered in coordination with mass transfer rates via diffu-
sion and subsequent partitioning to particles. Likewise, converting chamber measure-
ments to be representative indoor air and dust concentrations requires special
consideration for amines because of their strong preference to partition to suspended
particles and dust [120].

EPA Request for Curing Information

Because of the uncertainty regarding residential exposure to polyurethane products,
including SPF, the EPA requested information in September 2012 from nine domes-
tic manufacturers and processors [123] of polyurethane chemicals regarding curing
rates for polyurethane products intended to further react and thus undergo curing as
part of their commercial use, such as use in adhesives, sealants, foams, coatings, and
SPF. The curing rates for these types of products can affect the exposure potential
for chemicals that may be formed as reaction products. Specifically, the EPA
requested information related to isocyanates and wanted to know the curing time re-
quired to chemically react all diisocyanate functional groups and the amount of time
required, before and after a product has fully cured, to safely reoccupy or use an area
where diisocyanates have been reacted, including expected exposures to diisocya-
nates, other formulation additives, and the reaction products formed.
In response to this request, the EPA received IH studies, proprietary informa-
tion on formulations, and a risk assessment related to reoccupancy [124]. However,
based on the information collected, there remains uncertainty around curing time,
the multitude of other chemicals that may be emitted from SPF during and after
curing, and residential reoccupancy. For data received specific to SPF, chamber
study results were submitted that tested for only a limited set of VOCs based on
California Standard 01350 and ASTM D5116-06 [14]. Emission testing for residen-
tial reoccupancy was rarely addressed in manufacturers’ responses, and there was
only one human health risk assessment specific to reoccupancy for residents. This
assessment based its conclusion on a limited set of VOCs and recommended a reoc-
cupancy time of 2 h for foam insulation [125,126]. There may be additional indus-
try studies that have performed emission testing on a multitude of SPF components
compared with available health-based criteria, but this information was not provid-
ed to the EPA per the request.

Manufacturer Reoccupancy Guidelines

As the demand for energy efficiency grows, workers and residential/commercial
building inhabitants need clear information on the hazards and safe handling of in-
sulation materials, including reoccupancy and reentry guidances for installation and
postapplication. Sampling and analytical methods will need to be developed and
evaluated to measure emissions from these products from the time of application to
266 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

curing and over time. For example, indoor air quality must be considered as a factor
for the safe reoccupancy of a building in which SPF has been applied. Many manu-
facturers of SPF recommend a 24-h reoccupancy time [1]. However, a standardized
method for establishing this timeframe does not exist. Developing an ASTM stan-
dard to address the issue of postapplication reoccupancy times would allow the de-
velopment of practices that protect returning residents and building inhabitants
from chemical emissions from SPF and possible byproducts. SPF product manufac-
turers suggest a reoccupancy timeframe should be adhered to for building occupants
because the chemicals in SPF continue to emit after installation [36]. Factors such as
the type of product, product formulation and proportioning, ratio of coverage, tem-
perature, humidity, and moisture affect an SPF installation. Reoccupancy should not
take place until chemical emissions from the SPF product are below levels of concern
for building occupants. If SPF is not applied correctly or under ideal environmental
conditions (e.g., temperature and humidity), as determined by the manufacturer’s
specifications, they could continue to off-gas VOCs and SVOCs, which could lead to
exposure [59].

Building Codes and Residential Ventilation

Needs for Indoor Air Quality
Many building codes promote using insulation to improve energy efficiency, and
those establishing these codes should also consider the health effects associated
with the insulation. Building codes such as the International Energy Conservation
Code (IECC) with the U.S. Department of Energy (DOE) created a climate map of
the United States to assist with dictating appropriate R-values [127]. An insulating
material’s resistance to conductive heat flow is measured or rated in terms of its
thermal resistance or R-value. The higher the R-value, the greater the insulating ef-
fectiveness. The R-value depends on the type of insulation and its thickness and
density. Insulation barriers contribute to the building’s R-value [128]. These bar-
riers include SPF, cellulose, fiberglass, mineral wool, and polystyrene that are
applied or installed to increase the R-value, and, of these barriers, SPF provides the
highest R-values [16,128–130]. The DOE conducts analyses of building energy
codes and assists in formulating methodologies that set energy efficiency standards
for commercial and residential buildings. According to the DOE and building code
requirements, humid or cold climates require higher R-value for insulation material
to maintain comfort levels inside the building and to conserve heating and cooling
costs. The DOE also supports building energy codes and standards administered
by ASHRAE and the International Code Council, which also recommend high
R-values [127]. However, these building codes do not examine the insulation mate-
rials and the potential for harmful chemicals to be emitted during and after applica-
tion, which can contribute to poor indoor air quality for building occupants.
In addition, building codes should consider ventilation needs because, even if
these products are applied correctly, indoor air issues may occur. Insulation should
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 267

be coupled with suitable mechanical ventilation devices to promote air flow

throughout the building. Air flows throughout buildings, either through mechanical
ventilation mechanisms, such as HVAC systems, or through air pressure differ-
ences. The rate at which outdoor air replaces indoor air is called the air exchange
rate. When there is little infiltration, natural ventilation, or mechanical ventilation,
the air exchange rate is low, and pollutant levels may increase. For instance, if venti-
lation is not taken into consideration after SPF is applied, emission rates could in-
crease the pollutant levels inside the building from sources other than SPF [131].
For a typical residence in which SPF is on the exterior walls and in the attic, SPF
reduces the air exchange rate to an average of 0.1 to 0.2 ACH [132]. This reduction
can restrict the flow of air in the building, leading to indoor air issues if HVAC sys-
tems are not adjusted to promote an increased air exchange rate. Inadequate venti-
lation can increase indoor pollutant levels by not bringing in enough outdoor air to
dilute emissions from indoor sources. Lack of ventilation may cause chemicals to
accumulate [133]. Effects of poor indoor air quality may result in irritation of the
eyes, nose, and throat; headaches; dizziness; fatigue; and asthma [131]. The 2012
IECC requires mechanical ventilation [134], but guidelines need to ensure that
insulation and mechanical ventilation are coupled to facilitate adequate air flow
[135]. As additional states adopt more energy-efficient building codes, residential
ventilation needs should also be considered.

Conclusion
Most IH methodologies, including personal monitoring and area sampling, are typ-
ically used to measure the potential exposure to a single or limited class of chemi-
cals in the workplace for a short period of time. These IH methodologies do not
typically evaluate the long-term emissions from a formulated product, such as SPF
insulation and sealants, which are complex chemical mixtures. Products such as
SPF are also formulated with proprietary ingredients and are applied and reacted
on-site. IH studies use established sampling and analytical methods. However,
some chemicals in SPF have more developed sampling and analytical methods than
others; therefore, there needs to be a better understanding of which chemicals have
the greatest potential for resulting in harmful exposures. This is a challenge because
many chemicals in SPF-formulated mixtures do not have well-documented expo-
sure limits. A more thorough understanding of potential worker exposures could
help inform the potential for residential exposures as residents reenter their build-
ing immediately after application or if they remain on-site during installation of
SPF, which is not recommended. This should also be taken into consideration when
evaluating other trade worker exposures for SPF application. Indoor air quality
(IAQ) methods for evaluating material and product emissions in a building, includ-
ing environmental chamber studies, are also critical for understanding short- and
long-term residential exposures. The range of chemicals emitted, including any
byproducts or reaction products, should also be taken into consideration when
268 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

developing modeling scenarios based on environmental chamber testing. Advances

in environmental chamber technologies and in modeling indoor air, particulate,
and dust concentrations should help inform the relation between IAQ and the in-
stallation and use of SPF. Robust sampling and analytical methods are needed to
understand the variables that affect short- and long-term emissions via all mecha-
nisms, such as volatilization, diffusion, and absorption, to assess worker and resi-
dential exposure scenarios. Chamber testing can be useful for investigating
exposure potential and to determine parameters that may have an impact on the
emissions. Identifying the impact that changes to these parameters have on emis-
sions profiles should be determined. Chamber testing should also explore the use of
microchambers for quality control to identify off-ratio foam, which may lead to
exposures. It is also important to be able to characterize the profile of SPF emissions
over time, including from the time of application through the first 24 h, 48 h post-
application, and over a longer timeframe, such as 1 year, to understand exposure to
residential/commercial building inhabitants.
Modeling should consider product formulation, size of the room, distance
from the source, amount of product used (grams per event), thickness of applica-
tion, air exchange and ventilation rates, equipment-related variables (method of
application/type of applicator), temperature, and humidity. For SPF insulation,
there are multiple challenges from the standpoint of sampling and analysis to
support and enable robust emissions characterization. The continued develop-
ment of standardized test methods, modeling approaches, and generation of
reliable data will aid in understanding exposure scenarios. This will fill in infor-
mation gaps crucial to exposure assessment and identify areas requiring future
research. Information is needed to understand the complex exposure scenarios
that may affect residents and workers. Information on potential exposure
scenarios and uses will aid in conducting robust exposure assessments for SPF to
use for guidance for worker and residential/commercial building inhabitant
protection.

Disclaimer
These comments are those of the CPSC staff and have not been reviewed or
approved by, and may not necessarily reflect the views of, the Commission.

ACKNOWLEDGMENTS
We thank Robert Streicher, Dustin Poppendeick, Eva Wong, Charles Bevington, and
Amy Benson for their review and expertise.

References

[1] EPA, “Design for the Environment Best Practices,” http://www.epa.gov/dfe/pubs/proj-

ects/spf/spf_chemicals.html (accessed January 20, 2015).
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 269

[2] EPA, “MDI Action Plan,” http://www.epa.gov/oppt/existingchemicals/pubs/action-

plans/mdi.html (accessed December 29, 2014).

[3] OSHA, “Safety and Health Topics,” https://www.osha.gov/SLTC/hazardoustoxicsubstances

(accessed March 25, 2015).

[4] Fisseler-Eckhoff, A., Bartsch, H., Zinsky, R., and Schirren, J., “Environmental Isocyanate-
Induced Asthma: Morphologic and Pathogenetic Aspects of an Increasing Occupational
Disease,” Int. J. Environ. Res. Public Health, Vol. 8, No. 9, 2011, pp. 3672–3687.

[5] Redlich, C. A., Stowe, M. H., Wisnewski, A. V., Eisen, E. A., Karol, M. H., Lemus, R., Holm, C. T.,
Chung, J. S., Sparer, J., Liu, Y., Woskie, S. R., Appiah-Pippim, J., Gore, R., and Cullen, M. R.,
“Subclinical Immunologic and Physiologic Responses in Hexamethylene Diisocyanate-
Exposed Auto Body Shop Workers,” Am. J. Ind. Med., Vol. 39, No. 6, 2001, pp. 587–597.

[6] Tarlo, S. M. and Liss, G. M., “Diisocyanate-Induced Asthma: Diagnosis, Prognosis, and Effects of
Medical Surveillance Measures,” App. Occup. Environ. Hyg., Vol. 17, No. 12, 2002, pp. 902–908.

[7] Fabbri, L. M., Danieli, D., Crescioli, S., Bevilacqua, P., Meli, S., Saetta, M., and Mapp, C. E.,
“Fatal Asthma in a Subject Sensitized to Toluene Diisocyanate,” Am. Rev. Respir. Dis.,
Vol. 137, No. 6, 1988, pp. 1494–1498.

[8] NIOSH, “Report No. HETA 91-0386-2427,” 1994, http://www.cdc.gov/niosh/nioshtic-2/

00222308.html (accessed August 7, 2015).

[9] NIOSH, “Preventing Asthma and Death from MDI Exposure During Spray-on Truck Bed
Liner and Related Applications,” http://www.cdc.gov/niosh/docs/2006-149 (accessed
August 7, 2015).

[10] NIOSH, “Alert: Preventing Asthma and Death from Diisocyanate Exposure,” http://
www.cdc.gov/niosh/docs/96-111 (accessed August 7, 2015).

[11] Occupational Safety and Health Standards–Toxic and Hazardous Substances, 29 CFR §
1910.1000

[12] OSHA, https://www.osha.gov/OshDoc/Directive_pdf/CPL_03-00-017.pdf (December

29, 2014).

[13] EPA, “Toxicological Review of Methylene Diphenyl Diisocyanate (MDI) (CAS No. 101- 68-
8 and 9016-87-9,” http://www.epa.gov/ncea/iris/toxreviews/0529tr.pdf (accessed
August 7, 2015).

[14] OEHHA, “Air Toxics Hot Spots Program–Proposed Revised Reference Exposure Levels
for Toluene Diisocyanate and Methylene Diphenyl Diisocyanate,” http://oehha.ca.gov/
air/hot_spots/061710tdimdd.html viewed (accessed August 7, 2015).

[15] GESTIS, “Isocyanates,” http://limitvalue.ifa.dguv.de/WebForm_ueliste2.aspx_viewed

(accessed April 3, 2015).

[16] Spray Polyurethane Foam Alliance, “Builder’s Reference Handbook V1.0 IBS,” http://www.
sprayfoam.org/files/docs/news/557/SPF%20Builders%20Handbook%20v1.pdf (accessed
February 27, 2015).

[17] CPSC, “Status Report Staff Review of 5 Amine Catalysts Used in Spray Polyurethane
Foam,” http://www.cpsc.gov/PageFiles/129845/amine.pdf (accessed January 20, 2015)
270 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

[18] GESTIS, “Bis(2-Dimethylaminoethyl)oxide,” http://limitvalue.ifa.dguv.de/WebForm_ue-

liste2.aspx (accessed April 3, 2015).

[19] GESTIS, “2,20 -Iminodiethanol,” http://limitvalue.ifa.dguv.de/WebForm_ueliste2.aspx

(accessed August 7, 2015).

[20] GESTIS, “4-Ethylmorpholine,” http://limitvalue.ifa.dguv.de/WebForm_ueliste2.aspx

(accessed August 7, 2015).

[21] GESTIS, “3-(Dimethylamino)propylamine,” http://limitvalue.ifa.dguv.de/WebForm_ue-

liste2.aspx (accessed August 7, 2015).

[22] GESTIS, “Triethanolamine,” http://limitvalue.ifa.dguv.de/WebForm_ueliste2.aspx (accessed

August 7, 2015).

[23] GESTIS, “Triethylamine,” http://limitvalue.ifa.dguv.de/WebForm_ueliste2.aspx (accessed

August 7, 2015).

[24] GESTIS, “Triethylenediamine,” http://limitvalue.ifa.dguv.de/WebForm_ueliste2.aspx

(accessed August 7, 2015).

[25] EPA, “Protection of Stratospheric Ozone: Determination 29 for Significant New Alterna-
tives Policy Program 79 Vol. 62863, October 21, 2014,” https://www.gpo.gov/fdsys/pkg/
FR-2014-10-21/pdf/2014-24989.pdf (accessed February 24, 2015).

[26] EPA, “Substitutes in Rigid Polyurethane: Spray,” https://www.epa.gov/snap/substitutes-

rigid-polyurethanespray (accessed August 7, 2015).

[27] Honeywell, “HFC-245fa Product Stewardship Summary,” 2007, http://www51.honeywell.-

com/sm/common/documents/Public-Ris-Summary-HFC-245fa.pdf (accessed March 12,
2015).

[28] Honeywell, “Spray Polyurethane Foam Insulation,” http://www.honeywell-blowinga-

gents.com/applications/spray-polyurethane-foam-insulation (accessed February 27,
2015).

[29] DuPont, “Formacel Foam Blowing Agents,” http://www2.dupont.com/Formacel/en_US

(accessed February 27, 2015).

[30] Marklund, A., Andersson, B., and Haglund, P., “Organophosphorus Flame Retardants and
Plasticizers in Air from Various Indoor Environments,” J. Environ. Monitor., Vol. 7, No. 8,
2005, pp. 814–819.

[31] National Institute of Environmental Health Sciences, “Testing Status of Tris (2-Chloroiso-
propyl) Phosphate,” http://ntp.niehs.nih.gov/testing/status/agents/ts-m20263.html
(accessed August 11, 2015).

[32] European Union, “TCPP Risk Assessment,” https://echa.europa.eu/documents/10162/

13630/trd_rar_ireland_tccp_en.pdf (accessed February 24, 2015).

[33] Freudenthal, R. I. and Henrich, R. T., “Chronic Toxicity and Carcinogenic Potential of Tris-
(1,3-Dichloro-2-Propyl) Phosphate in Sprague-Dawley Rat,” Int. J. Toxicol., Vol. 19, 2000,
pp. 119–125.

[34] OSHA, “Triphenylphosphate,” https://www.osha.gov/dts/chemicalsampling/data/

CH_274400.html (accessed August 7, 2015).
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 271

[35] OSHA, “Tributyl Phosphate,” https://www.osha.gov/dts/chemicalsampling/data/

CH_272700.html viewed (accessed August 7, 2015).

[36] Karlovich, B. F., Thompson, C., and Lambach, J., “A Proposed Methodology for Develop-
ment of Building Re-Occupancy Guidelines Following Installation of Spray Polyurethane
Foam Insulation,” presented at the Center for the Polyurethanes Industry Technical Con-
ference, Nashville, TN, September 26–28, 2011, American Chemistry Council, Washington
DC—unpublished.

[37] Tsuang, W., “Asthma Induced by Exposure to Spray Polyurethane Foam Insulation in a
Residential Home,” J. Occup. Environ. Med., Vol. 54, No. 3, 2012, pp. 272–273.

[38] Huang, Y. T. and Tsuang, W., “Health Effects Associated with Faulty Application of Spray
Polyurethane Foam in Residential Homes,” Environ. Res., Vol. 134, 2014, pp. 295–300.

[39] California Department of Public Health, “Standard Method for the Testing and
Evaluation of Volatile Organic Chemical Emissions from Indoor Sources Using Environ-
mental Chambers, Version 1.1,” https://www.scsglobalservices.com/files/standards/
CDPH_EHLB_StandardMethod_V1_1_2010.pdf (accessed December 29, 2015).

[40] GESTIS, “Acetaldehyde,” http://limitvalue.ifa.dguv.de/WebForm_ueliste2.aspx (accessed

August 7, 2015).

[41] GESTIS, “Benzaldehyde,” http://limitvalue.ifa.dguv.de/WebForm_ueliste2.aspx (accessed

August 7, 2015).

[42] GESTIS, “Butyraldehyde,” http://limitvalue.ifa.dguv.de/WebForm_ueliste2.aspx (accessed

August 7, 2015).

[43] GESTIS, “2-Butenal,” http://limitvalue.ifa.dguv.de/WebForm_ueliste2.aspx (accessed

August 7, 2015).

[44] GESTIS, “2-Butenal—All Isomers,” http://limitvalue.ifa.dguv.de/WebForm_ueliste2.aspx

(accessed August 7, 2015).

[45] GESTIS, “Formaldehyde,” http://limitvalue.ifa.dguv.de/WebForm_ueliste2.aspx (accessed

August 7, 2015).

[46] GESTIS, “Propionaldehyde,” http://limitvalue.ifa.dguv.de/WebForm_ueliste2.aspx (accessed

August 7, 2015).

[47] GESTIS, “Valeraldehyde,” http://limitvalue.ifa.dguv.de/WebForm_ueliste2.aspx (accessed

August 7, 2015).

[48] GESTIS, “Dioxane,” https://www.osha.gov/dts/chemicalsampling/data/CH_237200.html

(accessed February 9, 2015).

[49] OSHA, “Chlorobenzene,” https://www.osha.gov/dts/chemicalsampling/data/CH_227000.

html (accessed February 9, 2015).

[50] OSHA, “Propylene Glycol,” https://www.osha.gov/dts/chemicalsampling/data/CH_

264480.html (accessed February 9, 2015).

[51] GESTIS, “1-Chloro-2-Propanol,” http://limitvalue.ifa.dguv.de/WebForm_ueliste2.aspx

(accessed August 7, 2015).
272 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

[52] GESTIS, “1,3-Dichloropropane,” http://limitvalue.ifa.dguv.de/WebForm_ueliste2.aspx

(accessed August 7, 2015).

[53] GESTIS, “1,4-Dioxane, Tech. Grade,” http://limitvalue.ifa.dguv.de/WebForm_uelis-

te2.aspx (accessed August 7, 2015).

[54] GESTIS, “Chlorobenzene,” http://limitvalue.ifa.dguv.de/WebForm_ueliste2.aspx (accessed

August 7, 2015).

[55] GESTIS, “Isopropyl Alcohol,” http://limitvalue.ifa.dguv.de/WebForm_ueliste2.aspx

(accessed August 7, 2015).

[56] GESTIS, “Isobutylamine,” http://limitvalue.ifa.dguv.de/WebForm_ueliste2.aspx (accessed

August 7, 2015).

[57] GESTIS, “Propane-1,2-Diol, Total Vapour, and Particulates,” http://limitvalue.ifa.dguv.de/

WebForm_ueliste2.aspx (accessed August 7, 2015).

[58] IRSST, “Studies and Research Projects Report R-629,” http://www.irsst.qc.ca/media/

documents/Pubirsst/r-629.pdf (accessed March 1, 2015).

[59] Genz, M. Schilling, U., Krasnow, J. J., Bockhoff, M., and Jansen, R., “Impact of Process
Parameters on Emissions of SPF,” presented at the Center for the Polyurethanes Indus-
try Technical Conference, Dallas, TX, September 22–24, 2014, American Chemistry Coun-
cil, Washington DC—unpublished.

[60] Spence, M. and Graham, C., “Evaluation of Particulates Generated During Trimming and
Cutting of Spray polyurethane Foam Insulation,” presented at the Center for the Polyur-
ethanes Industry Technical Conference, Houston, TX, October 11–13, 2010, American
Chemistry Council, Washington DC—unpublished.

[61] Robert, W., Andersen, J., Wood, R., and Bogdan, M., “Ventilation and Re-Occupancy of a
Residential Home Sprayed with High Pressure Polyurethane Foam,” presented at the
Center for the Polyurethanes Industry Technical Conference, Dallas, TX, September
22–24, 2014, American Chemistry Council, Washington DC—unpublished.

[62] EPA, “Ventilation Guidance for Spray Polyurethane Foam Application,” http://www.epa.-
gov/dfe/pubs/projects/spf/ventilation-guidance.html (accessed December 29, 2015).

[63] Wood, R., “CPI Ventilation Research,” presented at the Center for the Polyurethanes In-
dustry Technical Conference, Phoenix, AZ, September 23–25, 2013, American Chemistry
Council, Washington DC—unpublished.

[64] Wood, R., “CPI Ventilation Research Project for Estimating Re-Entry Times for Trade
Workers Following Application of Three Generic Spray Polyurethane Foam For-
mulations,” presented at the Center for the Polyurethanes Industry Technical Confer-
ence, Dallas, TX, September 22–24, 2014, American Chemistry Council, Washington
DC—unpublished.

[65] Robert, W., Wood, R., and Andersen, J., “Spray Polyurethane Foam Monitoring and Re-
Occupancy of High Pressure Open Cell Applications to New Residential Constructions,”
presented at the Center for the Polyurethanes Industry Technical Conference, Dallas, TX,
September 22–24, 2014, American Chemistry Council, Washington DC—unpublished.
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 273

[66] Ecoff, S. and Lambach, J., “A Proposed Methodology for Establishing a Safe Work Zone
Around an SPF Application to Exterior Walls of a Commercial Structure,” presented at
the Center for the Polyurethanes Industry Technical Conference, Atlanta, GA, September
24–26, 2012, American Chemistry Council, Washington DC—unpublished.

[67] ISO 16702:2007, Workplace Air Quality—Determination of Total Organic Isocyanate

Groups in Air Using 1-(2-Methoxyphenyl)piperazine and Liquid Chromatography, ISO,
Geneva, Switzerland, 2007, www.iso.org

[68] ISO 17734-1:2013, Determination of Organonitrogen Compounds in Air Using Liquid

Chromatography and Mass Spectrometry—Part 1: Isocyanates Using Dibutylamine Deriv-
atives, ISO, Geneva, Switzerland, 2013, www.iso.org

[69] ISO 17735:2009, Determination of Total Isocyanate Groups in Air Using 1-(9-Anthracenyl-
methyl)piperazine (MAP) Reagent and Liquid Chromatography, ISO, Geneva, Switzer-
land, 2009, www.iso.org

[70] CDC, “Isocyanates, Total (MAP),” http://www.cdc.gov/niosh/docs/2003-154/pdfs/

5525.pdf (accessed May 22, 2015).

[71] Wilder, L.C., Langley, R., Middleton D., Ernst, K., Lummusd Z., Streicher, R., Campbell, R.,
Wattigney, W., Bernstein, J., Bernstein, D., and Dearwent, S., “Communities Near Toluene
Diisocyanates Sources: An Investigation of Exposure and Health,” J. Exposure Sci. Envi-
ron. Epidemiol., Vol. 21, 2011, pp. 587–594.

[72] ISO 17736:2010, Workplace Air Quality—Determination of Isocyanate in Air Using a

Double-Filter Sampling Device and Analysis by High Pressure Liquid Chromatography,
ISO, Geneva, Switzerland, 2010, www.iso.org

[73] OSHA, “Method 47: Methylene Bisphenyl Isocyanate (MDI),” https://www.osha.gov/dts/

sltc/methods/organic/org047/org047.html viewed (accessed April 3, 2015).

[74] Sigma-Aldrich, “Air Monitoring,” http://www.sigmaaldrich.com/analytical-chromatogra-

phy/air-monitoring/asset-nco-sampler.html viewed (accessed April 3, 2015).

[75] Puscasu, S., Aubin, S., Cloutier, Y., Sarazin, P., Tra, H. V., and Gagné, S., “CIP10 Optimiza-
tion for 4,4-Methylene Diphenyl Diisocyanate Aerosol Sampling and Field Comparison
with Impinger Method,” Ann. Occup. Hyg., Vol. 59, No. 3, 2015, pp. 347–357.

[76] Bello, D., Nourian, F., Ernst, M. K., Steinmetz, M. A., and Streicher, R. P., “Measuring Total Re-
active Isocyanate Group Using 1,8-Diaminonaphthalene (DAN),” presented at Isocyanates
and Health: Past, Present and Future Conference, April 4, 2013, Bethesda, MD—unpublished.

[77] UK Health and Safety Executive, “Organic Isocyanates in Air,” http://www.hse.gov.uk/

pubns/mdhs/pdfs/mdhs25-4.pdf (accessed August 11, 2015).

[78] Streicher, R. P., Reh, C. M., Key-Schwartz, R. J., Schlecht, P. C., Cassinelli, M. E., and
O’Connor, P. F., “Determination of Airborne Isocyanate Exposure: Considerations in
Method Selection,” Am. Ind. Hyg. Assoc. J., Vol. 61, No. 4, 2000, pp. 544–556.

[79] Lesage, J., Stanley, J., Karoly, W. J., and Lichtenberg, F. W., “Airborne Methylene
Diphenyl Diisocyanate (MDI) Concentrations Associated with the Application of Poly-
urethane Spray Foam in Residential Construction,” J. Occup. Environ. Hyg., Vol. 4, No. 2,
2007, pp. 145–155.
274 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

[80] Lockey, J., Redlich, C., Streicher, R., Pfahles-Hutchens, A., Hakkinen, P. J., Ellison, G.,
Harber, P., Utell, M., Holland, J., Comai, A., and White, M., “Isocyanates and Health:
Knowledge Gaps and Research Priorities Across Stakeholders,” J. Occup. Environ. Med.,
Vol. 57, No. 1, 2015, pp. 44–51.

[81] Bello, D., Herrick, C., Smith, T., Woskie, S., Streicher, R., Cullen, M., Liu, Y., and Redlich, C.,
“Skin Exposure to Aliphatic Polyisocyanates in the Auto Body Repair and Refinishing
Industry: II. A Quantitative Assessment,” Ann. Occup. Hyg., Vol. 52, No. 2, 2008,
pp. 117–124.

[82] Fent, K. W., Javari, K., Gold, A., Ball, L. M., and Nylander-French, L. A., “Tape-Strip Sam-
pling for Measuring Dermal Exposure to 1,6-Hexamethylene Diisocyanate,” Scand. J.
Work Environ Health, Vol. 32, No. 3, 2006, pp. 225–240.

[83] Bello, D., Grote, A. A., Streicher, R. P., and Woskie, S. R. 2005. “Measurement of total re-
active isocyanate groups in samples using bifunctional nucleophiles such as 1,8-diamino-
naphthalene (dan). Publication number EP 1579207 A2, filed December 1, 2003, and
published September 28, 2005.

[84] Bello, D., Woskie, S. R., and Streicher, R. P., “Final Report: Development and Evaluation
of a Method for Determination of the Total Reactive Isocyanate Group (TRIG) Using the
Derivatizing Reagent 1,8-Diaminonaphthalene (DAN). Phase II—Analytical Finish,”
reported to the International Isocyanate Institute, 2002.

[85] Center for the Polyurethanes Industry, “Occupational Hygiene Air Monitoring for MDI
and TDI Guidance,” http://polyurethane.americanchemistry.com/Resources-and-Docu-
ment-Library/3822.pdf (accessed August 11, 2015).

[86] Allport, D., Gilbert, D., and Outterside, S., MDI and TDI: Safety, Health and the Environ-
ment, John Wiley & Sons, West Sussex, England, 2003, pp. 365–396.

[87] OSHA, “Triethlyamine,” https://www.osha.gov/dts/sltc/methods/partial/pv2060/

2060.html (accessed February 9, 2015).

[88] OSHA, “Triethanolamine,” https://www.osha.gov/dts/sltc/methods/partial/t-pv2141/

pv2141.html (accessed February 9, 2015).

[89] OSHA, “Diethanolamine,” https://www.osha.gov/dts/sltc/methods/partial/pv2018/

2018.html (accessed February 9, 2015).

[90] OSHA, “N-Ethylmorpholine,” https://www.osha.gov/dts/chemicalsampling/data/

CH_240640.html (accessed February 9, 2015).

[91] Dotson, S. G., Chen, P. H, Gadaqbvi, B., Maier, A., Ahlers, H. W., and Lentz, T. J., “The
Evolution of Skin Notations for Occupational Risk Assessment: A New NIOSH Strategy,”
Reg. Toxicol. Pharmacol., Vol. 61, 2011, pp. 53–62.

[92] Pilon, L., Fedorov, A. G., and Viskanta, R., “Gas Diffusion in Closed-Cell Foams,” J. Cell.
Plastics, Vol. 36, No. 6, 2000, pp. 451–474.

[93] OSHA, “Acetaldehyde,” https://www.osha.gov/dts/chemicalsampling/data/CH_216300.

html (accessed February 9, 2015).

[94] OSHA, “Butyraldehyde,” https://www.osha.gov/dts/chemicalsampling/data/CH_

223790.html (accessed February 9, 2015).
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 275

[95] OSHA, “Crotonaldehyde,” https://www.osha.gov/dts/chemicalsampling/data/

CH_230000.html (accessed February 9, 2015).

[96] OSHA, “Formaldehydes,” https://www.osha.gov/dts/sltc/methods/studies/srvassay/

srvassay.html (accessed February 9, 2015).

[97] OSHA, “n-Valeraldehyde,” https://www.osha.gov/dts/chemicalsampling/data/

CH_274960.html (accessed February 9, 2015).

[98] OSHA, “Isopropyl Alcohol,” https://www.osha.gov/dts/sltc/methods/organic/org109/

org109.html (accessed February 9, 2015).

[99] AFC International, Inc., “Measuring the Level of Polymeric MDI Vapors in the Workplace
Using GMD Systems MDI Monitors,” http://www.afcintl.com/pdfs/applications/mdi.pdf
(accessed August 24, 2015).

[100] EPA, “Characterizing Air Emissions from Indoor Sources,” http://www.epa.gov/nrmrl/

appcd/mmd/i-sovc.html (accessed March 15, 2015).

[101] ASTM D5116, Standard Guide for Small-Scale Environmental Chamber Determinations of
Organic Emissions from Indoor Materials/Products, ASTM International, West Consho-
hocken, PA, 2010, www.astm.org

[102] ASTM D6670, Standard Practice for Full-Scale Chamber Determination of Volatile Or-
ganic Emissions from Indoor Materials/Products, ASTM International, West Consho-
hocken, PA, 2013, www.astm.org

[103] ASTM D7706, Standard Practice for Rapid Screening of VOC Emissions from Products
Using Micro-Scale Chamber, ASTM International, West Conshohocken, PA, 2011,
www.astm.org

[104] ASTM D6330, Standard Practice for Determination of Volatile Organic Compounds (Ex-
cluding Formaldehyde) Emissions from Wood-Based Panels Using Small Environmental
Chambers Under Defined Test Conditions, ASTM International, West Conshohocken, PA,
2014, www.astm.org

[105] ISO 16000-9, Indoor Air—Part 9: Determination of the Emission of Volatile Organic Com-
pounds from Building Products and Furnishing—Emission Test Chamber Method, ISO,
Geneva, Switzerland, 2006, www.iso.org

[106] ISO 16000-10, Indoor Air—Part 10: Determination of the Emission of Volatile Organic
Compounds from Building Products and Furnishing—Emission Test Cell Method, ISO, Ge-
neva, Switzerland, 2006, www.iso.org

[107] ISO 16000-11, Indoor Air—Part 11: Determination of the Emission of Volatile Organic Com-
pounds from Building Products and Furnishing—Sampling, Storage of Samples and
Preparation of Test Specimens, ISO, Geneva, Switzerland, 2006, www.iso.org

[108] ISO 16000-25, Indoor Air—Part 25: Determination of the Emission of Semi-Volatile Or-
ganic Compounds by Building Products—Micro-Chamber Method, ISO, Geneva, Switzer-
land, 2006, www.iso.org

[109] ISO 16000-38, Indoor Air—Part 38: Determination of Amines in Indoor and Test Chamber
Air—Active Sampling on Samplers Containing Phosphoric Acid Impregnated Filters, ISO,
Geneva, Switzerland, 2015, www.iso.org
276 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

[110] ISO 16000-39, Indoor Air—Part 39: Determination of Amines in Indoor and Test Chamber
Air—Analysis of Amines by Means of High-Performance Liquid Chromatography (HPLC)
Coupled with Tandem Mass Spectrometry (MS MS), ISO, Geneva, Switzerland, 2015,
www.iso.org

[111] International Programme on Chemical Safety, “Flame Retardants: A General Introduction,”

http://www.inchem.org/documents/ehc/ehc/ehc192.htm (accessed August 11, 2015).

[112] U.S. National Library of Medicine, “Tris(chloropropyl)phosphate (TCPP), CASRN: 26248-

87-3,” https://toxnet.nlm.nih.gov/cgi-bin/sis/search2/f?./temp/tHbGNA:1 (accessed
May 22, 2016).

[113] EPA, “Exposure Factors Handbook,” http://www.epa.gov/ncea/efh/pdfs/efh-comple-

te.pdf (accessed August 11, 2015).

[114] Hoffman, K., Garantziotis, S., Birnbaum, L., and Stapleton, H., “Monitoring Indoor Expo-
sure to Organophosphate Flame Retardants: Hand Wipes and House Dust,” Environ.
Health Persp., Vol 123, No. 2, 2015, pp. 160–165.

[115] Stapelton, H., Klosterhaus, S. Eagle, S. Fuh, J., Meeker, J. D., Blum, A., and Webster, T. F.,
“Detection of Organophosphate Flame Retardants in Furniture Foam and U.S. House
Dust,” Environ. Sci. Tech., Vol. 43, No. 19, 2009, pp. 7490–7495.

[116] Oak Ridge National Laboratory, “The Science of Foam Aging,” http://info.ornl.gov/sites/
publications/files/Pub40530.pdf (accessed March 25, 2015).

[117] NIST, “Long-Term Emissions from Spray Polyurethane Foam Insulation,” http://
www.nist.gov/customcf/get_pdf.cfm?pub_id=915781 (accessed March 25, 2015).

[118] ASTM D5116-10, Standard Guide for Small-Scale Environmental Chamber Determinations
of Organic Emissions from Indoor Materials/Products, ASTM International, West Consho-
hocken, PA, 2010, www.astm.org

[119] Sebroski, J., “Research Report for Measuring Emissions from Spray Polyurethane Foam
(SPF) Insulation,” presented at the Center for the Polyurethanes Industry Technical Con-
ference, Atlanta, GA, September 24–26, 2012, American Chemistry Council, Washington
DC—unpublished.

[120] IHS Engineering, “Standard: ULC - ULC - CAN/ULC-S774-09,” http://standards.global-

spec.com/std/1659211/ulc-can/ulc-s774-090 (accessed March 25, 2015).

[121] ASTM D22.05 Subcommittee on Indoor Air WK40293, “New Test Method for Estimating
Chemical Emissions from Spray Polyurethane Foam (SPF) Insulation Using Micro-Scale
Environmental Test Chambers,” http://www.astm.org/DATABASE.CART/WORKITEMS/
WK40293.htm (accessed May 27, 2015).

[122] ASTM D22.05 Subcommittee on Indoor Air WK43872, “New Test Method for Estimating
Emissions of Methylene Diphenyl Diisocyanate (MDI) from Spray Polyurethane Foam
(SPF) Insulation Using Emission Cells or Micro-Scale Environmental Test Chambers,”
http://www.astm.org/DATABASE.CART/WORKITEMS/WK43872.htm (accessed May 27,
2015).

[123] EPA, “Letter to Companies,” (2012), http://www.regulations.gov/#!documentDetail;

D=EPA-HQ-OPPT-2011-0182-0103 (accessed August 7, 2015).
 SLEASMAN ET AL., DOI: 10.1520/STP158920150037 277

[124] EPA, “MDI Action Plan Docket,” http://www.regulations.gov/#!docketDetail;D=EPA-HQ-

OPPT-2011-0182 (accessed August 7, 2015).

[125] Saskatchewan Research Council Building Performance Unit, “V.O.C. Emissions Profiling
of a Polyurethane Foam Sample,” http://www.regulations.gov/#!documentDetail;
D=EPA-HQ-OPPT-2011-0182-0068 (accessed February 24, 2015).

[126] Icynene, http://www.regulations.gov/#!documentDetail;D=EPA-HQ-OPPT-2011-0182-

0065 (accessed February 24, 2015).

[127] DOE, “Building Energy Codes Program,” (accessed December 29, 2015).

[128] DOE, “Insulation,” http://energy.gov/public-services/homes/home-weatherization/

insulation (accessed August 7, 2015).

[129] DOE, “Insulation Materials,” http://energy.gov/energysaver/articles/insulation-materials

(accessed August 7, 2015).

[130] EPA, “Energy Star Qualified Homes Building Science Introduction,” https://www.energystar.
gov/ia/partners/bldrs_lenders_raters/downloads/ENERGY_STAR_V3_Building_Science.
pdf (accessed December 29, 2014).

[131] EPA, “An Introduction to Indoor Air Quality,” http://www.epa.gov/iaq/ia-intro.html#

Causes (accessed January 20, 2015).

[132] Garrett, D., “Proper Design of HVAC Systems for Spray Foam Homes,” http://www.icynene.
com/sites/default/files/US%20content%20uploads/builders/ProperDesignHVACSystems.
pdf (accessed February 27, 2015).

[133] EPA, “Remodeling Your Home? Have You Considered Indoor Air Quality?” http://www.
epa.gov/iaq/homes/hip-ventilation.html (accessed March 25, 2015).

[134] IECC, “Residential Provisions of the 2012 International Energy Conservation Code,”
http://www.energycodes.gov/sites/default/files/becu/2012iecc_residential_BECU.pdf
(accessed February 24, 2015).

[135] EPA, “Healthy Indoor Environment Protocols for Energy Upgrades,” http://www.epa.
gov/iaq/pdfs/epa_retrofit_protocols.pdf (accessed February 24, 2015).
278 DEVELOPING CONSENSUS STANDARDS FOR MEASURING CHEMICAL EMISSIONS

STP 1589, 2017 / available online at www.astm.org / doi: 10.1520/STP158920150042

Doyun Won,1 Joan Wong,2 Gang Nong,1 and

Wenping Yang1

VOC Emissions from Spray Foam

Insulation Under Different
Application Conditions
Citation
Won, D., Wong, J., Nong, G., and Yang, W., “VOC Emissions from Spray Foam Insulation Under
Different Application Conditions,” Developing Consensus Standards for Measuring Chemical
Emissions from Spray Polyurethane Foam (SPF) Insulation, ASTM STP1589, J. Sebroski and
M. Mason, Eds., ASTM International, West Conshohocken, PA, 2017, pp. 278–290, http://dx.doi.org/
10.1520/STP1589201500423

ABSTRACT
Spray polyurethane foam (SPF) insulation has been reported to emit hazardous
compounds if not applied correctly. Because two-component low-pressure spray
foam insulation is more readily available for a do-it-yourself project in homes,
the importance of optimal application is getting attention. We compared
chemical emissions from spray foam insulations applied in four different ways in
an attempt to simulate normal and abnormal applications. The normal
application (SPF1c) involved adhering to the manufacturer’s application
instructions and assumed identical amounts of Components A and B (1 : 1) and a
final foam thickness of 2 in. in two passes applied at room temperature
(22–23 C). The Component A opening was reduced to a quarter turn in attempt
to achieve a 0.25 : 1 ratio with the intention of generating a nonoptimal ratio of
two components in the next application (SPF1d). SPF1e and SPF1f were applied
at 16 and 5 C, respectively, which are suboptimal compared to the recommended
application temperatures of 21 to 32 C. After application, the specimens were
tested for 4 days in 50-L chambers at 23 C, 50 % relative humidity, and 1 air
change per hour. In general, the emission factors were higher if the foam was

Manuscript received May 15, 2015; accepted for publication August 1, 2016.
1
National Research Council Canada, Intelligent Building Operations, 1200 Montreal Rd., Ottawa, ON K1A 0R6, Canada
2
Health Canada, Water and Air Quality Bureau, Indoor Air Contaminants Assessment, 269 Laurier Ave., West,
Ottawa, ON K1A 0K9, Canada
3
ASTM Symposium on Developing Consensus Standards for Measuring Chemical Emissions from Spray
Polyurethane Foam (SPF) Insulation on April 30–May 1, 2015 in Anaheim, CA.

Copyright V
C 2017 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959.
 WON ET AL., DOI: 10.1520/STP158920150042 279

applied below the manufacturer’s recommended application temperature. More

specifically, the emission factor was the highest when the foam was applied
either at 5 C for the most volatile organic compounds (VOCs) or at 16 C for
some VOCs, followed by SPF1d. For example, the emission factor of triethyl
phosphate increased by a factor of 2, 8, and 12 for SPF1d, SPF1e, and SPF1f,
respectively. This demonstrates that the VOC emissions can increase significantly
when the spray foam is not applied according to the manufacturer’s application
instructions.

Keywords
spray foam insulation, chamber test, abnormal application, VOC emissions

Introduction
The demand for reducing energy consumption is growing for the building energy
sector because it accounts for a significant portion of total energy both in the Unit-
ed States and around the world [1]. Thermal insulation is one of the most effective
and practical ways of achieving energy efficiency in homes [2]. Among various
insulation options, spray polyurethane foam (SPF) insulation has gained in popu-
larity because of its performance and cost-effectiveness [3]. Although the insulation
is relatively inert when it is cured, there have been concerns about odor [4] and
occupant health [3], possibly as a result of improperly cured SPF insulation.
We sought to determine VOC emissions from SPF specimens applied outside
of manufacturer-recommended conditions that may reflect faulty applications. We
compared the chamber concentrations of selected VOCs from test specimens,
which were applied either with strict adherence to the manufacturer’s instructions
or with modifications to the application temperature and opening ratio of two
components.

Methods
PREPARATION OF TEST SPECIMENS
A spray kit of a two-component low-pressure SPF product was purchased from a
local home improvement store. According to the manufacturer, the product can
cover up to 200 board ft per kit and has an R value of 5.48. The recommended tem-
perature ranges were 16 to 32 C for storage and 21 to 32 C for application. The
two-component spray foam insulation (SPF1) was applied into a galvanized metal
specimen holder (approximately 0.2 by 0.254 by 0.05 m) lined with aluminium foil
(Fig. 1) in four different ways in an attempt to simulate normal and abnormal appli-
cations. The resulting specimens had an uneven top surface (Fig. 1) that was not cut
in order to minimize any disturbance.
The application conditions can be found in Table 1. SPF1c represents a normal
application that follows the manufacturer’s application instructions and assumes an
280 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 1 Test specimen contained in a metal box and tested in a 50-L chamber.

identical amount of Components A and B (1 : 1), 2-in. foam thickness, and a two-
pass application at room temperature (22–23 C). There was an attempt to achieve a
0.25 : 1 ratio of Components A and B with the intention of generating a nonoptimal
ratio of two components to make SPF1d. This was attempted by partially opening
the Component A tank, i.e., a quarter opening compared to the full opening of the
Component B tank. SPF1e and SPF1f were applied at 16 and 5 C, respectively,
which are suboptimal compared to the recommended application temperature of
21 to 32 C. The 5 C application represented one of the extreme temperature cases.
Air temperature was monitored during the application using an Omega HH311

TABLE 1 Application and test conditions.

Application Conditiona Test Condition

Assumed A : B Temperature Temperature

ID Ratio ( C) Weight (g)b ( C) Humidity (%) Weight (g)c

SPF1c 1:1 22.5 6 0.5 106.5 23.2 6 0.0 50 6 0.02 107.1

SPF1d 0.25 : 1 22.5 6 0.5 102.3 22.1 6 0.1 50 6 0.05 103.0
SPF1e 1 : 1d 16 98.6 23.2 6 0.3 50 6 0.01 99.2
SPF1f 1 : 1d 5 109.2 23.0 6 0.0 50 6 0.30 110.2
a
The relative humidity during the application was not controlled and was measured to be approxi-
mately 20 %.
b
Weight of foam after application.
c
Weight of foam after 4-day test.
d
Although the two tanks were fully opened (1:1), the pressure of the two tanks could be different be-
cause of the pressure disturbance caused during the application of SPF1d (see the “Limitations and
Recommendations” section for further discussion).
 WON ET AL., DOI: 10.1520/STP158920150042 281

humidity temperature monitor. The humidity level during the application was mea-
sured to be approximately 20 %. Two tanks of Components A and B were stored at
a corresponding temperature for 3 days before application.

CHAMBER TEST
The specimen was introduced into a 50-L chamber within 1 h after application with
the intention of capturing early emissions. The chamber was subjected to a 4-day
test under standard conditions of 50 % relative humidity (RH), a temperature of 22
to 24 C, and 1 air change per hour. SPF1c and SPF1d were sprayed and tested on
the same day (Week 1). Two electropolished stainless-steel 50-L chambers were op-
erated simultaneously. One chamber (Chamber 1) was enclosed in an environmen-
tal chamber, and the other chamber (Chamber 2) was placed in a room; both were
maintained at a temperature of 23 6 1 C. One week after the test of SPF1c and
SPF1d, SPF1e was sprayed and tested in Chamber 1 (Week 2). During the following
week, the test for SPF1f was performed in Chamber 1 (Week 3).
Clean air for the chambers was supplied with an Aadco Model 737 Pure Air
Generator. The humidity of the supply air was maintained at a constant level of
50 6 0.3 % RH with the use of mass flow controllers to blend separate streams of
Aadco air (one dry and one saturated by bubbling air through high-performance
liquid chromatography [HPLC]-grade water). According to Table 1, the weight of a
specimen was slightly increased after a test. The increase of specimen weight
typically occurs when the test RH is higher than the application RH; i.e., specimens
seemed to absorb moisture because they were tested at 50 % RH after being
applied at 20 % RH. The pressure inside the chamber was maintained at approxi-
mately 60 Pa.

AIR SAMPLING AND CHEMICAL ANALYSIS

During a 4-day test, air samples were collected at 0 h, 3 h, 6 h, 1 day, 2 days, 3 days,
and 4 days on sorbent tubes and 2,4-dinitrophenylhydrazine (DNPH) cartridges. The
R
two-layer sorbent tube (178 by 6 by 4 mm) packed with 160 mg TenaxV TA and
R
180 mg CarbographV 5TD was purchased from Camsco. The DNPH cartridge was a
Sep-Pak XPoSure from Waters. The sampling volume was approximately 2 to 6 L at
100 to 200 mL/min for sorbent sampling and approximately 25 L at 250 mL/min for
DNPH sampling. Additional samples were taken on tubes packed with approximately
R
300 mg TenaxV TA for approximately 21 h at 100 mL/min, leading to a sampling
volume of 90 to 130 L. This was to check for the presence of semivolatile organic
compounds (SVOCs) such as flame retardants. The large sampling volume was
required for SVOCs because the concentrations of SVOCs in the 50-L chambers at
23 C were expected to be low partly because of chamber sink effects.
Air samples collected on two-layer sorbent tubes were thermally desorbed and
analyzed with a gas chromatography-mass spectrometry (GC-MS) system, includ-
ing a Gerstel thermal desorption system (TDS) connected with an autosampler, an
Agilent 6890 gas chromatograph equipped with a DB-624 capillary column (length,
282 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

30 m; inner diameter, 0.25 mm; thickness, 1.4 lm), and a 5973N Mass Selective De-
tector. The desorbed analytes were injected with a programmable temperature va-
porizer called a cooled injection system (CIS) that concentrated the sample before
being injected onto the column of the gas chromatograph. The GC system was op-
erated in the TDS splitless/CIS split mode (split ratio of 20:1), and the MS system
R
was operated in the full-scan mode (m/z ¼ 35–300). The TenaxV tubes were ana-
lyzed with a similar GC-MS system with a DB-5 capillary column (length, 30 m; in-
ner diameter, 0.25 mm; thickness, 0.25 lm) and a mass scanning range of 35 to 800.
The temperature profiles of the VOC and SVOC methods are provided in Table 2.
Carbonyl compounds were analyzed in accordance with ASTM D5197-09e1,
Standard Test Method for Determination of Formaldehyde and Other Carbonyl
Compounds in Air (Active Sampler Methodology) [5]. Exposed DNPH cartridges
were washed with acetonitrile, and the eluate was then analyzed by reverse-phase
HPLC with ultraviolet (UV) detection at 360 nm. The HPLC-UV system consisted
of a Prostar 410 autosampler, Varian 9012 solvent delivery system with two Supel-
cosil LC-18 columns (length, 25 cm; inner diameter, 4.6 mm; thickness, 5 lm) in a
series maintained at 30 C, and Varian 9050 variable-wavelength UV-visible spec-
trophotometer. A gradient of acetonitrile in water from 60 % to 100 % was used for
separating carbonyl compounds.
Six-point calibration was performed with the use of three commercial liquid
calibration mixtures (Supelco) for 120 VOCs and a commercial DNPH derivative
mixture (Supelco) for carbonyl compounds. The method detection limit (MDL),
which was determined in accordance with guidance published by the US Environ-
mental Protection Agency [6], ranged from 0.4 to 14.6 ng with an average of 3.7 ng
for 120 VOCs and 7 to 13 ng for carbonyl compounds. Abundant compounds,
which have a relatively large peak and are not included in the target VOC list, were
quantified as a toluene equivalent. The National Institute of Standards and

TABLE 2 Temperature profiles of the GC/MS method (VOC and SVOC).

VOC SVOC

Desorption Initial temperature 30 C 30  C

Ramp 60 C/min 60  C/min
 
Final temperature 300 C (holding 5 min) 320 C (holding 5 min)
Cooled injection system Initial temperature 90  C 20  C

Ramp 12 C/s 12  C/s
 
Final temperature 300 C (holding 3 min) 350 C (holding 3 min)
GC Initial temperature 10  C (holding 4 min) 70  C (holding 1 min)

Ramp 1 6 C/min 20  C/min

Middle temperature 230 C —
Ramp 2 15  C/min —
Final temperature 250  C (holding 2 min) 320  C (holding 6 min)
 WON ET AL., DOI: 10.1520/STP158920150042 283

Technology Mass Spectral Library (NIST98 version 1.7a) was used for the chemical
identification of abundant compounds.
Background and blank samples were analyzed as a quality-control measure.
R R R
Background samples (TenaxV TA/CarbographV 5TD, TenaxV TA, and DNPH)
were collected in an empty chamber before each test to verify chamber cleanliness.
In addition, blank samples were analyzed for each run of the GC-MS and HPLC-
UV systems. In general, the background levels were higher than MDL for all three
sampling methods. Therefore, background levels were subtracted from chamber
concentrations when emission factors were estimated (Eq 1). Because the blank lev-
els were also higher than MDL for DNPH, the blank correction was made for car-
bonyl compounds collected on the DNPH cartridge. The blank levels of sorbent
tubes were typically lower than MDLs except for benzene. Therefore, the blank cor-
rection was done for benzene only.

Results and Discussion

CHAMBER CONCENTRATIONS
The chamber concentrations of VOCs emitted from SPF1c applied according to the
manufacturer’s instructions are shown in Table 3. VOCs are organized in three
groups based on three sampling media. The reported time corresponds to the mid-
point of the sampling time.
R R
Among the VOCs collected on the TenaxV/CarbographV, six were detected
above MDLs. The dominant VOC that was not on the target list was trans-1,2-
dichloroethene (TDCE), which is typically used as a blowing agent enhancer. The
concentration of TDCE was 6,339 lg/m3 at 3 h and 2,282 lg/m3 at 96 h. Other
detected VOCs had lower concentrations than TDCE by 2 to 4 orders of magnitude.
For example, triethyl phosphate, which was likely to be used as a flame retardant,
had a concentration of 17.2 lg/m3 at 3 h and 11.4 lg/m3 at 96 h. Hydrofluorocarbon
blowing agents were not detected because of the standard practice of monitoring
the mass starting at 2 min as a precaution to remove any solvent peak.
Table 3 also shows two low-molecular-weight aldehydes (acetaldehyde and
propanal) sampled on the DNPH cartridges. Formaldehyde levels are not shown
because they were below MDL. Aldehydes are known to be formed from oxidative
degradation of tertiary amine catalysts used in polyurethane foams [7]. It is interest-
R R
ing to see there were no amine catalysts detected in in the TenaxV/CarbographV
samples despite the presence of aldehydes. This may be because of a low sampling
volume or the presence of reactive catalysts that tend to be entrained in the polyure-
thane matrix [8].
R
SVOCs collected on the TenaxV TA tubes are also summarized in Table 3. The
Tenax TA tubes for SVOC sampling did not reveal the presence of more chemicals
in addition to those detected by VOC sampling. One exception was tris(1-chloro-2-
propyl) phosphate (TCPP) (Table 3). TCPP was not detected by VOC sampling
most likely because of the low sampling volume. Although two siloxane compounds
284 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 3 VOC and SVOC chamber concentrations from SPF1c.

Chamber Concentration by Tenax /Carbograph (lg/m3)

R
V R
V

Compound CAS # 0h 3h 6h 23.9 h 48 h 72 h 96 h

2-Ethyl-1-hexanol 104-76-7 0.6 0.7 0.6 0.4 0.5 0.5

Acetone 67-64-1 0.9 0.8 0.4 0.3 0.2 0.2
Texanol 25265-77-4 1.2 1.8 1.4 0.7 0.8
1,2-Dichloroethane 107-06-2 1.9 1.8 0.9 0.7 0.6 0.4
1,2-Dichloropropane 78-87-5 1.0 0.9 0.5 0.4 0.4
Triethyl phosphate 78-40-0 17.2 26.4 48.4 22.0 14.2 11.4
trans-1,2-Dichloroethenea 156-60-5 0.0 6338.9 5387.9 3819.6 2877.0 2511.6 2282.0

Chamber Concentration by DNPH Sampling (lg/m3)

0h 4.1 h 7.1 h 25 h 49.1 h 73.1 h 97.1 h

Acetaldehydeb 75-07-0 0.6 4.4 3.3 1.9 1.2 1.0 1.2

Propanalb 123-38-6 14.0 10.1 5.1 3.0 2.1 1.5

Chamber Concentration by Tenax Sampling (lg/m3)

R
V

0h 15.7 h 38.1 h 62.1 h 86.3 h

Tris(1-chloro-2-propyl) 13674-84-5 0.0 1.2 11.7 11.7 10.0

phosphatea
D5a 541-02-6 0.6 10.2 10.2 10.5 12.3
D6a 540-97-6 0.4 4.7 2.9 2.4 2.1

a
Abundant compound quantified as a toluene equivalent.
b
DNPH sampling.

were detected in the VOC samples, they were not quantified because they were out-
side of the 120 target compounds. Instead, they were quantified as a toluene equiva-
lent in the SVOC sample. It is not clear whether siloxane compounds are by-
products of the GC column bleed or siloxane surfactants used in the foam.
In general, the chamber concentrations decreased over time. One exception
was TCPP, for which the concentration increased from 15.6 to 38 h and was main-
tained at a relatively constant level afterward. This was likely because TCPP belongs
to a group of SVOCs. The steady increase of a concentration in a chamber is one of
the characteristics of SVOC emissions, likely because of slow emissions and sub-
stantial sink effects affected by low vapor pressure. Similar observations were made
in other tests (SPF1d, 1e, and 1f).
Whereas dodecamethylcyclohexasiloxane (D6) showed a decreasing trend with
time, decamethylcyclopentasiloxane (D5) had relatively constant emission factors
over time. No clear reason can be found to explain the difference in the decay pat-
terns of D5 and D6. However, it may be speculated that the decreasing trend of D6
supports the proposition that it is a by-product of siloxane surfactants rather than
the GC column bleed. Further information is required to validate this assumption.
 WON ET AL., DOI: 10.1520/STP158920150042 285

FIG. 2 Emission factors for acetone and Texanol.

COMPARISON OF EMISSION FACTORS

The emission factors were calculated from the measured chamber concentrations
normalized to the specimen weight (Eq 1).

EFW ¼ 1; 000ðC  Cin ÞQ=W (1)

where:
EFw ¼ emission factor normalized to specimen weight, ng/gh,
C ¼ VOC concentration in the chamber, lg/m3,
Cin ¼ VOC background concentration in the empty chamber, lg/m3,
Q ¼ chamber flow rate, m3/h,
W ¼ weight of a specimen, g, and
1,000 ¼ a conversion factor from lg to ng.
The emission factors normalized to the specimen weight are compared for acetone
and Texanol in Fig. 2, chlorinated compounds in Fig. 3, aldehydes in Fig. 4, phospho-
rous compounds in Fig. 5, and siloxanes in Fig. 6. In general, the emission factors

FIG. 3 Emission factors for chlorinated compounds.

286 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

FIG. 4 Emission factors for aldehydes.

were the highest with the specimen applied at 5 C (SPF1f), followed by the speci-
men applied at 16 C (SPF1e), specimen SPF1d, and the specimen applied according
to the manufacturer’s instructions (SPF1c). Exceptions were aldehydes and silox-
anes. For example, acetaldehyde (Fig. 4a) showed the highest emission factor with
the specimen applied at 16 C (SPF1e). Propanal (Fig. 4b) also had the highest emis-
sion factor at 16 C (SPF1e) at the beginning. However, the emission factors of
propanal became similar for all specimens after 24 h. Two siloxanes (D5 and D6)
(Fig. 6) had the highest emission factor at 16 C.
To obtain a quick overview of how much the emission rates were increased for
suboptimal specimens, the emission factors from suboptimal applications were nor-
malized to those of the normal application (SPF1c). The ratios were slightly differ-
ent at each sampling time. Table 4 presents the ratios averaged for 48, 72, and 96 h;
the ratios >1.5 or <0.5 are indicated, and the difference smaller than 650 % is

FIG. 5 Emission factors for phosphorous compounds.

 WON ET AL., DOI: 10.1520/STP158920150042 287

FIG. 6 Emission factors for siloxanes.

insignificant in the light of the potential uncertainness associated with the test and
chemical analysis.
Overall, the chemicals analyzed with the VOC method showed the highest
emission factors at 5 C for most VOCs or at 16 C for some VOCs (e.g., Texanol
and acetaldehyde). The ratio was up to 9.2 at 16 C and 15.4 at 5 C. On the other
hand, compounds analyzed with the SVOC method (TCPP, D5, and D6) tended to
show smaller differences between specimens with a maximum ratio of 2:1. SPF1d
showed similar or slightly higher emission factors than the normal application
(SPF1c). The increased emissions in SPF1d were observed with Texanol, triethyl
phosphate, and acetaldehyde by a factor of 2 (Table 4).

LIMITATIONS AND RECOMMENDATIONS

Although this study provides insights into emissions of VOCs and SVOCs for SPFs
applied under nonoptimal conditions that can occur during do-it-yourself applica-
tions, there are several limitations that need to be recognized.
First, it is not clear why emission factors were increased with abnormally ap-
plied specimens. One potential explanation is that the increased emissions resulted
from improper ratios of Components A and B, possibly leading to inadequate cur-
ing reactions, an abnormal release of additives, or both. When an SPF is sprayed,
Components A and B are typically introduced in a 1:1 ratio by volume into a small
mixing chamber of a spray gun [9,10]. Off-ratio foams can also occur when the
tanks are not stored within the recommended temperature range [9].
There is also a possibility that off-ratio foams might be generated in the 16 C
(SPF1e) and 5 C (SPF1f) test as a result of the testing order. SPF1e and SPF1f were
applied in sequence with a 1-week interval after the application of SPF1d, which
was intended as an off-ratio specimen. The unbalanced opening during the prepara-
tion of SPF1d might disturb the pressure balance of the two tanks, leading to off-
ratio specimens for SPF1e and SPF1f. Nevertheless, the possibility needs to be
288 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

TABLE 4 Ratio of emission factors to those of the normal specimen (mean values: 48–96 h).

SPF1c (Normal) SPF1d (Off-Ratio) SPF1e (16  C) SPF1f (5  C)

2-Ethyl-1-hexanol 1.0 0.6 1.5 2.7

Acetone 1.0 0.9 3.0 15.4
Texanol 1.0 1.9 5.6 4.8
1,2-Dichloroethane 1.0 1.2 2.1 3.3
1,2-Dichloropropane 1.0 1.0 1.8 3.6
trans-1,2-Dichloroethene 1.0 1.0 1.2 2.3
Triethyl phosphate 1.0 2.2 8.3 11.9
Acetaldehyde 1.0 2.3 9.2 5.0
Propanal 1.0 1.5 1.4 1.1
Tris(1-chloro-2-propyl) 1.0 0.9 0.4 1.2
phosphate
D5 1.0 0.7 1.3 0.7
D6 1.0 0.9 2.1 1.3

Note: Ratios greater than 1.5 are indicated in bold; a ratio less than 0.5 is indicated in bold and
italic.

investigated in a future study by changing the order of tests or using a new tank sys-
tem in each test.
Second, no attempt was made to confirm the actual nature of nonoptimal appli-
cation conditions and cured specimens. For example, SPF1d was accomplished by a
quarter opening of the Component A tank and full opening of the Component B
tank. However, there were no additional steps to confirm that a quarter opening
corresponded to 25 % flow of Component A. Therefore, it is not clear whether the
application method resulted in an off-ratio specimen for SPF1d. Better specimen
preparation methods need to be developed in future studies if the goal is to study
emissions from an off-ratio specimen. For example, weighing each tank during the
specimen preparation can be used to confirm the applied mass or volume. In addi-
tion, examining cured specimens for properties such as cell structure, density, and
adhesion can shed light on the curing quality [9]. This may be helpful for elucidat-
ing the influencing factors and mechanisms for increased emissions of nonoptimal
specimens. Despite this limitation, the preparation method used in this study can
still provide valuable information in the sense that it reflects one of the conditions
that can occur when the foam is not applied according to the manufacturer’s appli-
cation instructions.
Third, the test results are based on one particular SPF product. The test method
and results need to be replicated and verified in tests with more diverse products. In
addition, more application conditions need to be investigated. For example, one
set of RH levels (approximately 20 % during application and 50 % during testing)
was used in this study. Because water can be involved in SPF installation as a
co-blowing agent [11,12], RH is most likely to affect the curing process and therefore
 WON ET AL., DOI: 10.1520/STP158920150042 289

VOC emissions. This type of research would be helpful for better understanding the
likelihood and extent of chemical emissions from nonoptimal applications of SPF
products in diverse environments and for developing more suitable guidelines for the
SPF applications.

Conclusions
VOC emissions were measured in SPF specimens applied under four different con-
ditions, attempting normal and off-ratio openings of Components A and B, and
two suboptimal temperatures (16 and 5 C) using a two-component low-pressure
spray kit. In general, the VOC emission factors were the highest with the applica-
tion at 5 C, followed by the one applied at 16 C, the attempted off-ratio and normal
specimen. For example, the emission of triethyl phosphate increased by a factor of
2, 8, and 12 for the attempted off-ratio, 16 C, and 5 C applications, respectively,
compared to the normal application. The increased emissions from the attempted
off-ratio application were modest (up to a factor of 2) compared to those at subopti-
mal temperature applications. The increased emissions were less pronounced with
SVOCs (e.g., TCPP) compared to those of VOCs (2-ethyl-1-hexanol; acetone;
triethyl phosphate; 1,2-dichloroethane; 1,2-dichloropropane; TDCE; and acetalde-
hyde). Although the mechanism that led to the increased emissions is not clear, it
was likely caused by improper reactions of foam ingredients, nonoptimal curing, or
the abnormal release of additives. The outcomes of this research need to be validat-
ed in at least two ways. First, more diverse products and application conditions
need to be investigated to better understand the likelihood and magnitude of the in-
creased emissions in actual environments. In addition, more fundamental research
is needed to shed light on the factors and mechanisms involved in the increased
emissions from suboptimally applied specimens.

References

[1] Jelle, B. P., “Traditional, State-of-the-Art and Future Thermal Building Insulation Materi-
als and Solutions—Properties, Requirements and Possibilities,” Energ. Buildings, Vol. 43,
No. 10, 2011, pp. 2549–2563.

[2] Al-Homoud, M. S., “Performance Characteristics and Practical Applications of Common

Building Thermal Insulation Materials,” Build. Environ., Vol. 40, No. 3, 2005,
pp. 353–366.

[3] Huang, Y.-C. T. and Tsuang, W., “Health Effects Associated with Faulty Application
of Spray Polyurethane Foam in Residential Homes,” Environ. Res., Vol. 134, 2014,
pp. 295–300.

[4] Holiday, M., “Spray Foam Jobs with Lingering Odor Problems,” 2011, http://www.
greenbuildingadvisor.com/blogs/dept/musings/spray-foam-jobs-lingering-odor-problems
(accessed May 14, 2013).
290 STP 1589 On Developing Consensus Standards for Measuring Chemical Emissions

[5] ASTM D5197-09e, Standard Test Method for Determination of Formaldehyde and Other
Carbonyl Compounds in Air (Active Sampler Methodology), ASTM International, West
Conshohocken, PA, 2009, www.astm.org

[6] U.S. EPA, “Definition and Procedure for the Determination of the Method Detection
Limit— Revision 1.11, https://www.gpo.gov/fdsys/pkg/CFR-2011-title40-vol23/xml/CFR-
2011-title40-vol23-part136-appB.xml (accessed November 4, 2016).

[7] Barman, B. N., “Accurate Determination of Aldehydes in Amine Catalysts or Amines by

2,4-Dinitrophenylhydrazine Derivatization,” J. Chromatogr. A, Vol. 1327, 2014, pp. 19–26.

[8] Wegener, G., Brandt, M., Duda, L., Hofmann, J., Klesczewski, B., Koch, D., Kumpf, R.-J.,
Orzesek, H., Pirkl, H.-G., Six, C., Steinlein, C., and Weisbeck, M., “Trends in Industrial Ca-
talysis in the Polyurethane Industry,” Appl. Catalysis A: General, Vol. 221, No. 1–2, 2001,
pp. 303–335.

[9] Knowles, M., “Troubleshooting Spray-Foam Insulation,” J. Light Construct., September

2010, pp. 1–6.

[10] Rashidi, M. 2002. Two component spray gun and nozzle attachment. U.S. Patent
6,375,096, filed Mar. 1, 2000, and issued Aug. 23, 2002.

[11] Bogdan, M., Williams, D., and Verbiest, P., “HFC-245fa Spray Polyurethane Foam
Systems Co-Blown with Water: A Quality, Cost Effective, Safe Substitute for HCFC-141b,”
J. Cell. Plastics, Vol. 37, 2001, pp. 58–71.

[12] Brilhante, A. M. F., “Re-Formulation and Cost Optimisation of One Component PU

Foams,” Universidade Técnica de Lisboa, Lisbon, Portugal, 2008.
 | SEBROSKI
MASON
Emissions from Spray Polyurethane Foam (SPF) Insulation
Developing Consensus Standards for Measuring Chemical
ASTM INTERNATIONAL
Selected Technical Papers

Developing Consensus
Standards for Measuring
Chemical Emissions
from Spray Polyurethane
Foam (SPF) Insulation
STP1589
STP 1589

ASTM INTERNATIONAL Editors:

Helping our world work better John Sebroski
Mark Mason
ISBN: 978-0-8031-7623-2
Stock #: STP1589 www.astm.org
www.astm.org