See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/225453863

New approach to calibrating conductivity meters in the low conductivity range

Article  in  Accreditation and Quality Assurance · February 2005

DOI: 10.1007/s00769-004-0880-4

CITATIONS READS

9 478

5 authors, including:

Petra Spitzer
Physikalisch-Technische Bundesanstalt
78 PUBLICATIONS   1,340 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Petra Spitzer on 21 May 2014.

The user has requested enhancement of the downloaded file.

Accred Qual Assur (2005) 10:78–81
DOI 10.1007/s00769-004-0880-4 GENERAL PAPER

Petra Spitzer
Bruno Rossi
New approach
Yves Gaignet to calibrating conductivity meters
St
phane Mabic
Uwe Sudmeier in the low conductivity range

Received: 30 December 2003 Abstract There is currently a major Keywords Electrolytic

Accepted: 25 August 2004 issue with the calibration of conduc- conductivity · High purity water ·
Published online: 12 October 2004 tivity meters used for high purity Calibration · Primary methods ·

Springer-Verlag 2004 water: the lack of availability of a Metrology in chemistry
reference material or reference
Milipore, Milli-Q and Elix are registered methods for low conductivity ranges
trademarks of Millipore Corporation. (conductivity below 1 mS cm1 at
P. Spitzer ()) · U. Sudmeier 25.0 C, resistivity >1 MW cm at
Physikalisch-Technische Bundesanstalt, 25.0 C). This paper describes the
Metrology in Chemistry, current status of conductivity mea-
Bundesallee 100, 38116 Braunschweig, surements in high purity water. A
Germany new and improved approach, cur-
e-mail: petra.spitzer@ptb.de rently being investigated, should al-
Tel.: +49-531-5923130
Fax: +49-531-5923015 low us to make the calibration of
conductivity meters used for low
B. Rossi · Y. Gaignet · S. Mabic conductivity ranges traceable to the
Lab Water Division, Millipore, SI.
St Quentin-Yvelines, France

Introduction and should be compliant with the regulatory require-

ments. Documented evidence and traceability to national
Conductivity is the parameter most commonly used to or international measurement standards are, therefore,
monitor the overall ionic purity of water, and it is the among the mandatory criteria. However, a major issue
major parameter considered for defining and ranking faced today with the calibration of conductivity measur-
grades of purified water. Among the various ways of ing devices used for high purity water is the lack of a
measuring conductivity, in-line monitoring has evolved as reference material or reference methods suitable for low
the preferred technique, for both central loops and labo- conductivity ranges (conductivity below 1 mS cm1 at
ratory water purification systems. Indeed, reducing air- 25.0 C, resistivity >1 MW cm at 25.0 C). Traceable
borne contamination is key to measuring conductivity reference standards delivered by metrological institutes
accurately, in particular for high purity water. Built-in, in- and accredited laboratories are only available for higher
line conductivity meters have been the way to solve issues conductivity ranges. Currently no reference solution is
due to contamination from the environment. available to calibrate conductivity measuring devices used
However, just like any measurement, conductivity at low conductivity ranges with low uncertainties. Hence,
values can only be used and trusted if they are determined no accredited laboratory can actually propose calibration
with a high level of confidence in their reliability. This for low conductivity measurements (below 1 mS cm1 at
implies that calibration of the conductivity measuring 25.0 C) [1].
device should be performed periodically. Such a process To address this need, a collaborative effort was un-
should fit with the laboratory quality management system dertaken to develop a means to perform calibration in the
 79

low conductivity range. The purpose of this paper is (1) to kexp=f(T) displayed by the conductivity meter with the
point out the current status of conductivity measurements theoretical curve kth=f(T).
of high purity water and report on calibration issues, and Even if such an approach is currently used and con-
(2) to describe the metrological approach investigated to sidered to be the most valid approach available, this
afford calibration of conductivity measuring devices used method has some limitations. Mobility values are exper-
in low conductivity ranges and traceable to the SI. imental and their determination is not traceable. The
characterization of ionically pure water is not easy due to
complex and expensive instrumentation (ICP-MS), and it
Low conductivity range calibration: status and issues is difficult to identify and measure all possible ions.
Finally, to prevent potential issues due to other con-
Lack of reference standards taminants present in water, such as particulates, organics,
and gases, it is necessary to ensure they are reduced to the
The instantaneous contamination of high purity water minimum levels. To achieve this, ultrapure water should
with carbon dioxide from the air, with ions from container be produced using multiple purification technologies; for
flasks, and from the environment prevents the prepara- example 0.22 mm membrane filtration, UV photo-oxida-
tion, storage and distribution of low conductivity aqueous tion, and so on. Otherwise, the presence of these con-
standards. Indeed, contamination from the above-cited taminants may potentially impact on the conductivity
sources readily results in an increase in conductivity. In measurements. Despite such weaknesses, this method
water of higher conductivity, where the impact of con- certainly remains, metrologically speaking, the most ac-
taminants on the overall conductivity is proportionally ceptable, until reference solutions can be proposed.
smaller (standard solutions ranging from 1 mS cm1 to
10 mS cm1 at 25.0 C), it is possible to add 1-propanol to
the aqueous standard to reduce the carbon dioxide solu- Conductivity meter and cell calibration
bility and preserve the ionic content of the solution [2].
To compensate for the absence of certified reference Conductivity meters consist of an electronic meter com-
solutions, an ultrapure-water-based calibration method ponent connected to a conductivity cell via a cable.
has become widely accepted by the users of low con- The cell constant Kcell is used to calculate the con-
ductivity measuring devices, and utilized by some man- ductivity from the resistance R of the sample in the
ufacturers to validate the conductivity values displayed by conductivity cell using Eq. 1:
conductivity meters in the low conductivity range. Kcell
k¼ ð1Þ
R
Ultrapure-water-based method For a cylindrical cell with parallel electrodes, the cell
constant Kcell is determined through the effective cross-
The method discussed here relies on the fact that water sectional area A, and the effective length l, of the current
purified by high-quality grade ion exchange resin contains path, as in Eq. 2:
residual ions at trace levels, that is ng l1 or sub-ng l1 l
level for each ion, and below 1 mg l1 NaCl equivalent KCell ¼ ð2Þ
A
overall. Such water can be considered as formally ioni-
cally pure water. To confirm the absence of significant If the cell geometry and the resistance of the sample
concentrations of ions, dosage of residual ions can be solution are known to be traceable to the SI, then the
performed using recognized analytical techniques, such as conductivity of the solution can be determined by a pri-
ICP-MS and ion chromatography. mary method of measurement.
By definition, in ultrapure water, only hydronium In order to ensure the traceability of conductivity
(H3O+) and hydroxide (OH) ions are in solution, in values for routine measurements, a calibration hierarchy
concentrations given by the water dissociation constant. has been established in many countries over the last few
The resulting nominal conductivity is 0.055 mS cm1 at years. Potassium chloride solutions, having SI-traceable
25.0 C [3]. With the current detection sensitivity of conductivity values, measured with primary cells of
conductivity meters, no effect is seen on the conductivity known geometry, are employed as reference solutions to
value as long as the ionic contamination remains below calibrate secondary cells. Secondary cells, together with
1 mg l1 NaCl equivalent overall. Conductivity (k, also properly calibrated conductivity meters, are used for
sometimes symbolised by c) is temperature-dependent via routine measurements [1]. However, as explained above,
the ion mobility parameter mi and a theoretical curve low conductivity standards are subjected to airborne CO2
kth=f(T) can be drawn. The ultrapure-water-based method, contamination, introducing additional uncertainty beyond
then, consists of comparing the experimental curve the target uncertainty for most applications. Therefore,
this procedure does not provide traceability for the con-
80

ductivity of very dilute aqueous solutions with an ap- the PTB primary piston type cell [7] (which has the ad-
propriate uncertainty. Another approach sometimes used vantage that fringe effects can be eliminated) has been
is to calibrate the electronic part and the conductivity cell applied to concentric electrodes. The differential cell
with the cable separately. Calibration of the electronic constant is of the order of 1 m1 with a relative expanded
part is readily done using resistors traceable to national target uncertainty (k=2) smaller than 0.5%. The complete
reference standards [4]. For the cell constant determina- piece of equipment including the impedance bridge con-
tion, different methods exist like those mentioned by nected to the cell, cable and temperature sensors is cali-
regulatory bodies (for example, the United States Phar- brated at different frequencies and at the measurement
macopoeia, USP [5]). Again, one limitation for low temperature. Pt-100 temperature sensors are incorporated
conductivity measurement ranges, is that such calibra- in the process at different places. The resistance of the
tions are, in most cases, done (directly or indirectly) using pure water is measured and converted to conductivity (not
high conductivity standards. Another limitation can be the compensated for temperature). All measurements are
method itself. Two individual, separate calibrations of the done at a temperature that can be accurately determined
conductivity cell and the electronic meter component with and controlled with known uncertainty. Because cell di-
the cable using resistors cannot be considered to be a mensions, resistance and temperature measurements are
metrologically robust calibration of the complete mea- traceable to the SI, the conductivity of the ultrapure water
suring system (conductivity cell and electronic meter). can be determined by a primary method of measurement.
Actually, the use of resistors to calibrate the electronic Furthermore, a secondary cell to be calibrated can be
meter part does not permit simulation of the conditions installed in-line with the primary cell, as close as possible
once the cell is connected. Depending on the treatment of to avoid temperature variations. To increase the accuracy,
the electrical signal, the value displayed on the electronic calibrations will be performed at various temperatures,
meter component could be degraded compared to the using a heat exchanger. At this step, it should be possible
value obtained with a resistor (ideal signal), due to the to get results k=f(T) for ultrapure water traceable to the
capacitive effect of the complete conductivity cell. So SI, with a known uncertainty, via a primary method. The
calibration of the electronic meter part cannot be limited repeatability of such a curve (obtained by using media
to the use of traceable resistors but should be completed from different lots, by modifying the flow rates,


) will
by verification and analysis of the signal treatment (once allow the accuracy of the ultrapure water versus temper-
the cell used is connected) under different electronic ature reference curve obtained to be defined.
conditions (frequency,


). So, for a reliable calibration,
the electronic conductivity meter and the cell used should
be calibrated together however the calibration method is Status of experiments
defined.
The test bench was built with particular attention to the
selection of the components and materials in contact with
New metrological approach water. This selection was, in particular, based on previous
material release studies [8]. Ionically pure water delivered
Method by an ultrapure water purification system (Milli-Q Gra-
dient A10 system fed by water purified by an Elix system)
In the absence of traceable reference standards for circulates in a loop that includes specific mixed-bed ion
low conductivity measurements, developing a calibration exchange resins to consistently ensure the lowest possible
method that holds a metrological significance must rely conductivity. Water quality is defined by a nominal
on calibration traceable to the result of a primary mea- conductivity of 0.055 mS cm1 at 25 C and a total organic
surement method. A project was initiated to develop a carbon content below 5 mg l1.
calibration method based on a primary method that will Tests conducted so far have supported the validity of
provide users with metrologically-recognized calibration the approach investigated. A detailed description of the
for their conductivity measuring device in the low con- experimental method, along with the results, will be re-
ductivity range. The most common design for in-line ported in a further document.
applications in the ultrapure water range is conductivity Other approaches for primary conductivity cells, useful
cells with concentric electrodes. The concentric design in the range of ultrapure water, are being investigated by
minimizes electromagnetic interference [6]. other metrology institutes [9]. Results obtained for low
In collaboration with Millipore, a primary flow- ranges of conductivity by different methods should be
through conductivity cell was constructed at the PTB. The compared as soon as possible to ensure their mutual
constant of this primary cell is determined by a geomet- recognition.
rical approach. The distance between the concentric
electrodes is varied in a controlled and traceable way. For
this new cell design, the reliable differential principle of
 81

Conclusion and long-mastered technique by metrological institutes.

This will provide the ability to calibrate conductivity
Progress has been made to allow the metrological cali- meters employed to assess the conductivity of high purity
bration of conductivity meters in the low conductivity water used in industries and laboratories.
range. This approach is based on geometry, a well-defined

References
1. Brinkmann F, Dam NE, Dek E, 5. U.S. Pharmacopoeia (2004) USP 27-NF 8. Darbouret D, Kano I, Youf E, Stewart B
Durbiano F, Ferrara E, Fuko J, Jensen 22. U.S. Pharmacopoeia, Rockville (1999) The Millipore R&D notebook.
HD, Mrissy M, Shreiner RH, Spitzer 6. Thornton RD, Light TS (1989) Ultra- Millipore Corporation, Billerica, MA
P, Sudmeier U, Surdu M, Vyskocil L pure water, July/August 7 9. Jensen HD, Sørensen J (2000) Electro-
(2003) Accred Qual Assur 8:346–353 7. Spitzer P, Sudmeier U (2000) Electro- lytic conductivity at DFM—results and
2. Wu YC, Berezansky PA (1995) J Res lytic conductivity—a new subject field experiences. In: Spitzer P, Sudmeier U
Natl Inst Stan 100:521–527 at PTB. In: Spitzer P, Sudmeier U (eds) (eds) Bericht PTB-ThEx-15. PTB,
3. Marshall WL (1987) J Chem Eng Data Bericht PTB-ThEx-15. PTB, Braun- Braunschweig, pp 3–13
32:221–226 schweig, pp 39–47
4. Kremer MK (2000) Chem Tech 29:26–
28

View publication stats