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

net/publication/254908764

The Chemistry and Technology of Furfural Production in Modern

Lignocellulose-Feedstock Biorefineries

Article

CITATIONS READS

19 5,217

1 author:

Gianluca Marcotullio
Delft University of Technology
10 PUBLICATIONS   463 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Gianluca Marcotullio on 07 January 2014.

The user has requested enhancement of the downloaded file.

 T H E C H E M I S T RY A N D T E C H N O L O G Y O F
FURFURAL PRODUCTION IN MODERN
LIGNOCELLULOSE-FEEDSTOCK BIOREFINERIES

gianluca marcotullio

Furfural, the sleeping beauty

Process and Energy Department

3ME Faculty

Delft University of Technology

 The Chemistry and Technology of Furfural Production in
Modern Lignocellulose-Feedstock Biorefineries

Proefschrift

ter verkrijging van de graad van doctor

aan de Technische Universiteit Delft,
op gezag van de Rector Magnificus prof. ir. K.C.A.M. Luyben,
voorzitter van het College voor Promoties,
in het openbaar te verdedigen op maandag 19 december 2011 om 10:00 uur

door

Gianluca MARCOTULLIO

Ingegnere Meccanico, Università degli studi dell’Aquila

geboren te L’Aquila, Italië

iii
 Dit proefschrift is goedgekeurd door de promotor:
Prof. dr. ir. A. H. M. Verkooijen
copromotor:
Dr. ir. W. de Jong

Samenstelling promotiecommissie:
Rector Magnificus voorzitter
Prof. dr. ir. A. H. M. Verkooijen Technische Universiteit Delft, promotor
Dr. ir. W. de Jong Technische Universiteit Delft, copromotor
Prof. dr. J. Clark University of York, UK
Prof. dr. H. J. Heeres Rijksuniversiteit Groningen
Prof. dr. I. W. C. E. Arends Technische Universiteit Delft
Prof. dr. G. J. Witkamp Technische Universiteit Delft
Dr. ir. R. R. C. Bakker Wageningen University and Research Centre
Prof. dr. ir. B. J. Boersma Technische Universiteit Delft, reservelid

This research was funded by the European Commission’s Marie Curie EST
project INECSE (MEST-CT-2005-021018), and the European FP6 Integrated Project
Biosynergy (038994-SES6), which are gratefully acknowledged.

ISBN 978-88-95207-49-0

Copyright © 2011 by Gianluca Marcotullio

All rights reserved. No part of the material protected by this copyright no-
tice may be reproduced or utilized in any form or by any means, electronic or
mechanical, including photocopying, recording or by any information storage
and retrieval system, without the prior permission of the author. An electronic
version of this dissertation is available at http://repository.tudelft.nl.

Published by Arkhé Edizioni - L’Aquila, Italy

iv
- Noodles, what have you been doing all these years?

- Been going to bed early...

Sergio Leone, Once upon a time in America

 to my dear

family

friends

Laura

in the order they prefer

vii
prologue

Getting to the end of this 4+ years experience gives conflicting feelings, and this
little piece of work is not the only thing I am left with after leaving Delft. I’ll
take with me a trunk full of people, places, and feelings I will never forget.
The people I shared this time with made it special, and for this reason I would
like to thank them all.
Special thanks to Wiebren, who allowed all this to happen in the first place.
Wiebren, I will always be immensely grateful to you for this great opportunity,
for the fair mix of independence and responsibility you gave me, for encouraging
me to a more critical scientific attitude, and for trusting my work. Thank you to
my promotor, Ad Verkooijen, a real Prof.. Thank you Henk, a cheerful man and
a special colleague who introduced me to the fascinating world of furfural; who
left us unexpectedly for a sudden and unfortunate chance.
Lab work takes numberless unsuccessful experiments, calibrations, duplicates,
repetitions, hiccups, and big headaches. If I still managed to put together few
valuable results during these years it is surely thanks to the man who gets the
lab going at P&E, bedankt Michel. Many thanks also to the whole P&E technical
staff.
Most of my work was done in the framework of the EU research project Biosyn-
ergy. The involvement in Biosynergy has been for me an unmatched working
and learning experience. For all this I would like to thank Hans Reith, a wise
but passionate coordinator; Boris and Rob, two WP leaders who always showed
a very encouraging appreciation of my work; and the many colleagues from 17
different institutions I shared this great experience with.
Being a Marie Curie fellow was another great privilege I had when joining TU
Delft. It must have been a secret selection method, or maybe an experiment on
human relations we weren’t informed about, but how easily the INECSE bunch
got along in the few occasions we gathered together goes beyond imagination.
Especially after working hours ... and after after working hours, till slightly be-
fore working hours. Thank you Adrian, Mariusz, Paula, Dorota, Kasia, Bruno,
Senthoor, Piotr, Audrius, Rastko, Elisa, Catarina, Jacopo and (although only at
the very end) Martina, with you guys going to sleep has always been a remote
option. I wish you all the best.
Of course there was much beside work during my stay in The Netherlands.
At the beginning discovering Rotterdam by night in all its freaky corners with
Ninuzzo, Jean, Pietro and the many occasional guests has been a thrilling hobby.
Thank you guys for being rotterdammers with me. Thanks also to my flatmates
at Hillevliet and Pannekoekstraat, Bo, a superb chinese cook; Erik, a real Dutch
doctor (to be); Anders and Arjan, the acrobat and the juggler.
Sincere thanks to my travel companions at P&E and around, I don’t think I
could have ever made it without your company and friendship. Thank you Ja-
copo, my first nurse at Delft, and his beloved Mahsa; Mattia, the terror of any
forward on the football pitch, I’ll surely see you and Luisella again; Eleonora and
Emile, looking forward to seeing you on the snowy peaks of Switzerland; my of-
ficemates, Atif, Richard T., Stefano, Xiangmei, FanFan and Ming Liu. Thank you
Ryan, a kaleidoscope of drinking events, gossip, and equations of state; John
and Bart, the Italian-speaking Dutch; Miguel, Carsten, Emiliano, Ernesto, Marta,
Albert, Richard L., Michel, Martina, Lawien, Bernardo, Sergio, Helene for shar-
ing countless, laid-back, lunch breaks and those silly discussions; Elsa, for your
precious work in the lab; and thanks to the many other P&E colleagues which I
don’t even try to list.
... and thanks to Laura for being always there, for your supernatural talent
for enduring me, for your counter-engineering mind-set, for your “stupido mas-
chio!”, and for many other things I won’t mention here. (e adesso traducilo...).
Grazie infine alla mia famiglia, ai parenti, agli amici, e a tutte le persone che
ho lasciato in Italia nell’unico posto che possa chiamare casa.

Gianluca
S U M M A RY

This dissertation deals with biorefinery technology development, i.e. with the de-
velopment of sustainable industrial methods aimed at the production of chemi-
cals, fuels, heat and power from lignocellulosic biomass. This work is particularly
focused on the production of furfural from hemicellulose-derived pentoses.
The possibility of producing materials, chemicals, and fuels from biomass has
a long history. Unfortunately fossil resources, in particular oil, have dominated
last century both in the chemical and energy sectors due to their enormous
availability at relatively low cost. Nowadays, bio-based chemicals and fuels are
gaining a much more favorable competitive position, not only by virtue of high
oil prices and the forecasts on their future availability, but also because of the
increasing public environmental awareness.
Although the use of biomass for energy, and in particular for biofuels produc-
tion, has been greatly challenged recently in terms of real benefits to the environ-
ment, the technological advances in the biorefinery field offer many reasons to
believe that, in the near future, biomass could be definitely the sustainable alter-
native to fossil resources in the chemical and also fuels industries. By a careful
review of the scientific literature, and observing the recent research trends, it is
noticeable that a relatively short list of high-potential biomass-derived building
blocks has the potential for replacing or substituting fossil resources in nearly
every industrial field.
The broad family of furan compounds represents certainly an intriguing selec-
tion of biomass derivatives for many industrial applications. Furfural is nowa-
days the only starting material for the production of nearly all the furan com-
pounds. Furfural industry exists since almost a century, but it is nowadays facing
a major renovation challenge in order to meet the global trend toward bio-based
products, and the consequent increased demand for furfural and its derivatives.
The majority of current furfural production is still based on more or less mod-
ified versions of the original Quaker Oats process (1921). For reasons that can
be related to their limited technological evolution, the production processes in
use today generally suffer from low yields (around 50%), besides significant eco-
nomical and environmental concerns. All these reasons hindered the expansion
and modernization of the furfural industry below its actual potential. A pro-
found technological development is a priority in order to upgrade furfural to a
large-volume bio-based commodity. The integrated production of furfural within
modern biorefineries is a big opportunity, and it will most probably represent
the next cornerstone in the development of furfural industry.

xi
 In Chapter 1 the opportunities offered by the modern biorefinery in the broader
context, and the importance of furan compounds are highlighted. In this con-
text the enormous potential of furfural and its derivatives, both in the chemical
and energy sector, is discussed. Recent advances in furfural technology are sum-
marized, both regarding furfural synthesis and applications, eventually stating
the motivation behind this dissertation, and the main achievements herein con-
tained.
In Chapter 2 the experimental methods used in this work are carefully de-
scribed. A new lab-scale titanium reactor was built in order to investigate several
aspects related to the furfural formation and related reactions, and to enable liq-
uid phase reactions under a relatively broad range of pressure, temperature and
pH conditions. This test rig has allowed most of the experimental work behind
this dissertation, and thus it is thoroughly described in this chapter in all the
relevant aspects typical of chemical reactor engineering. The analytical and ex-
perimental methods employed in the several experimental campaigns described
in this dissertation are also extensively described.
Chapter 3 concerns the reaction kinetics of furfural formation. Even consider-
ing the number of relevant works on the topic of furfural formation in acidic
media, a general expression for the reaction kinetics, its dependence on the acid
nature and concentration, and the potential effect of other species present in so-
lution, is yet to be defined. Results of reaction kinetics studies related to furfural
formation from xylose, xylose side reactions, and furfural destruction in acidic
aqueous media are thus studied and reported.
In Chapter 4 some particular aspects of the chemistry of xylose reaction into
furfural are addressed with the aim to clarify the reaction mechanism and to
define new green catalytic pathways for its production. Specifically the reduc-
tion of mineral acids utilization is addressed by the introduction of alternative
catalysts. In this sense the effect of chloride salts in dilute acidic solutions at
temperatures between 170 and 200 °C is described. Results indicate the Cl− ions
to promote the formation of the 1,2-enediol from the acyclic form of xylose, and
thus the subsequent acid catalyzed dehydration to furfural. For this reason the
presence of Cl− ions led to significant improvements with respect to the H2SO4
base case. The addition of NaCl to a 50mM HCl aqueous solution (0.18 wt%)
allows to attain 90% selectivity to furfural. Among the salts tested FeCl3 shows
very interesting preliminary results, producing exceptionally high xylose reac-
tion rates.
Starting from the results discussed in chapter 4 on the effects of Cl− ions on
furfural formation in aqueous acid solution, the general effect of different halides
is addressed in Chapter 5. Experimental results show the halides to influence at
least two distinct steps in the reaction leading from xylose to furfural under
acidic conditions, via different mechanisms. The nucleophilicity of the halides
appears to be critical for the dehydration, but not for the initial enolization reac-
tion. By combining different halides synergic effects become evident resulting in
very high selectivities and furfural yields.
In Chapter 6 the combined production of hemicellulose-derived carbohydrates
and an upgraded solid residue from wheat straw using a dilute-acid pretreat-
ment at mild temperature is described. Dilute aqueous HCl solutions were stud-
ied at temperatures of 100 and 120 °C, and they were compared to dilute FeCl3
under the same conditions. Comparable yields of soluble sugars and acetic acid
were obtained, affording an almost complete removal of pentoses when using
200mM aqueous solutions at 120 °C. The solid residues of pretreatment were
characterized showing a preserved crystallinity of the cellulose, and a almost
complete removal of ash forming matter other than Si. Results showed upgraded
characteristic of the residues for thermal conversion applications compared to
the untreated wheat straw.
Chapter 7 deals with the industrial processes for the production of furfural,
describing in particular an innovative process patented by Delft University of
Technology and based on the results contained in this dissertation. As already
mentioned, the integrated production of furfural within modern biorefineries
will most probably represent the next cornerstone in the development of furfural
industry. The innovative process concept described in this chapter is aimed at an
economically viable and environmentally sound furfural production, with re-
duced energy and chemicals consumption. The evaluation of process economics
shows encouraging results compared to existing processes, making this concept
economically attractive.
Finally, in Chapter 8 main concluding remarks are provided, as well as recom-
mendations for future research.

Gianluca Marcotullio
S A M E N VAT T I N G

Dit proefschrift behandelt de ontwikkeling van bioraffinage technologieën, te

weten de ontwikkeling van duurzame industriële methodes gericht op de pro-
ductie van chemicaliën, brandstoffen, warmte en elektriciteit op basis van ligno-
cellulose biomassa. Dit onderzoekswerk is in het bijzonder gericht op de produc-
tie van furfural uit pentose suikers, afkomstig van het hemi-cellulose deel van
biomassa.
De mogelijkheid om materialen, chemicaliën en brandstoffen te produceren
uit biomassa heeft al een lange geschiedenis. Helaas hebben fossiele bronnen,
met name olie, de afgelopen eeuw gedomineerd in zowel de chemische als
de energie sector door hun enorme beschikbaarheid tegen relatief lage kosten.
Tegenwoordig verkrijgen biomassa gebaseerde chemicaliën en brandstoffen een
gunstiger competitieve positie, niet alleen door de hoge olieprijzen en de voor-
spellingen voor wat betreft hun beschikbaarheid in de toekomst, maar ook door
het toenemende milieubewustzijn van de maatschappij.
Hoewel het gebruik van biomassa voor energievoorziening, en in het bijzon-
der voor bio-brandstof productie, recentelijk behoorlijk is uitgedaagd in termen
van werkelijke voordelen voor het milieu, bieden de technologische vorderin-
gen op het gebied van bioraffinage veel aanleiding om te geloven dat in de
nabije toekomst biomassa zeker een duurzaam alternatief kan bieden voor fos-
siele bronnen in de chemische en brandstofproducerende industrie. Door een
zorgvuldige overzichtsstudie van de wetenschappelijke literatuur en observatie
van de recente onderzoekstrends wordt duidelijk dat een relatief korte lijst van
biomassa afgeleide bouwstenen met een hoog potentieel in potentie zou kunnen
zorgen voor de vervanging of gedeeltelijke substitutie van fossiele bronnen in
bijna elke industriële sector.
De grote familie van furaanverbindingen vertegenwoordigt zeker een interes-
sante selectie van biomassa afgeleide componenten voor veel industriële toepassin-
gen. Furfural is tegenwoordig het enige uitgangsmateriaal voor de productie
van bijna alle furaan-verbindingen. De furfural producerende industrie bestaat
al sinds bijna een eeuw, maar staat vandaag de dag voor de uitdaging om gro-
tendeels te vernieuwen om tegemoet te kunnen komen aan de globale trend in
de richting van biomassa gebaseerde producten en de bijbehorende toegenomen
vraag naar furfural en haar afgeleide producten.
Het grootste deel van de huidige furfural productie is nog steeds gebaseerd
op min of meer gemodificeerde versies van het originele Quaker Oats proces
(1921). Vanwege hun beperkte technologische ontwikkeling worden de produc-
tieprocessen van vandaag gelimiteerd door lage opbrengsten (rond 50%) en is

xiv
er zorg wat betreft zowel economische en ecologische aspecten. Al deze pun-
ten hebben de expansie en modernisering van de furfural industrie zodanig
gehinderd dat zij haar actuele potentieel niet haalt. Een forse technologische
ontwikkeling is nodig als prioriteit om furfural een stap verder te brengen als
grootschalige biomassa gebaseerde sleutelcomponent. De geïntegreerde produc-
tie van furfural in moderne bioraffinaderijen is een veelbelovende mogelijkheid
en zal zeer waarschijnlijk de volgende hoeksteen vormen in de verdere uitbouw
van de furfural producerende industrie.
In hoofdstuk 1 worden de mogelijkheden geboden door de moderne bio-
raffinaderij in een bredere context geplaatst en wordt het belang van furaan-
verbindingen toegelicht. In dit verband wordt het enorme potentieel voor zowel
de chemische als de energie sector van furfural en haar afgeleide componen-
ten bediscussieerd. Recente ontwikkelingen in de furfural technologie worden
samengevat, zowel met betrekking tot synthese als toepassingen, waarbij tenslotte
de motivatie voor dit proefschrift wordt gegeven met het oog op de beoogde re-
sultaten.
In hoofdstuk 2 worden de experimentele methodes die zijn toegepast in dit
werk uitgebreid beschreven. Een nieuwe titanium reactor is gebouwd op lab-
schaal om verschillende aspecten van furfural vorming en gerelateerde reacties
te bestuderen en om vloeistoffase reacties mogelijk te maken in een relatief brede
range van druk, temperatuur en pH-waarden. Deze testopstelling heeft het mo-
gelijk gemaakt om het grootste deel van de experimenten in het kader van dit
proefschrift uit te voeren, en dus wordt zij grondig beschreven in dit hoofdstuk
gerelateerd aan alle relevante aspecten van de chemische reactor technologie.
De analytische en experimentele methodes die zijn toegepast in de verschillende
experimentele campagnes die in dit proefschrift zijn weergegeven staan ook uit-
gebreid in dit hoofdstuk beschreven.
Hoofdstuk 3 behandelt de reactiekinetiek van furfural vorming. Zelfs als wordt
gekeken naar het aantal relevante studies naar furfural vorming in zure media,
is het nog steeds nodig een algemene uitdrukking te definiëren voor de reac-
tiekinetiek, waarbij de afhankelijkheid van de aard en concentratie van het zuur,
alsmede het potentiële effect van andere componenten in de oplossing wordt
meegenomen. Daarom worden de resultaten van de reactiekinetiek studies gere-
lateerd aan furfural vorming uit xylose, xylose nevenreacties en furfural afbraak
in waterige, zure media bestudeerd en gerapporteerd.
In hoofdstuk 4 wordt een aantal bijzondere aspecten van de chemie van de re-
actie van xylose naar furfural behandeld met als doel om het reactiemechanisme
op te helderen en om nieuwe groene, katalytische productieroutes te ontsluiten.
In het bijzonder wordt aandacht besteed aan de reductie van mineraal zuur
gebruik door de introductie van alternatieve katalysatoren. In dit kader wordt
het effect van chloride zouten in verdunde zure oplossingen beschreven bij tem-
peraturen tussen 170 en 200 °C. De resultaten wijzen op een promotie van de
vorming van het “1,2-enediol” uit de acyclische vorm van xylose en de opvol-
gende zuur gekatalyseerde dehydratatie tot furfural. Vanwege dit effect leidde
de aanwezigheid van Cl− ionen tot significante verbeteringen in vergelijking
met de H2SO4 basis case. De toevoeging van NaCl aan een 50 mM HCl waterige
oplossing (0.18 massa%) leidt tot de realisatie van 90% selectiviteit naar furfural.
Van de toegevoegde zouten vertoont FeCl3 zeer interessante initiële resultaten,
waarbij uitzonderlijk hoge xylose reactiesnelheden zijn behaald.
Met de besproken resultaten uit hoofdstuk 4 betreffende de effecten van Cl−
ionen op de vorming van furfural in zure, waterige oplossingen als uitgangspunt,
wordt in hoofdstuk 5 het algemene effect van verschillende halides bestudeerd
in hoofdstuk 5. De experimentele resultaten tonen aan dat de halides minstens
twee afzonderlijke stappen in het reactieproces van xylose naar furfural beïnvloe-
den middels verschillende mechanismen. De mate van nucleofiel gedrag van de
halides blijkt kritisch te zijn voor de dehydratatie, maar niet voor de initiële
enol-vormingsreactie. Door verschillende halides te combineren komen synergie
effecten aan het licht, die resulteren in zeer hoge selectiviteiten en furfural op-
brengsten.
In hoofdstuk 6 wordt de gecombineerde productie van hemicellulose afgeleide
suikers en een verbeterd vast residu op basis van tarwestro beschreven, gebruik-
makend van een verdund zure voorbehandeling bij milde temperaturen. Ver-
dunde, waterige HCl oplossingen zijn bestudeerd bij temperaturen van 100 en
120 °C en deze zijn vergeleken met verdunde FeCl3 oplossingen onder dezelfde
reactieomstandigheden. Vergelijkbare opbrengsten van oplosbare suikers en azi-
jnzuur zijn verkregen, waarbij een bijna complete verwijdering van pentose suik-
ers is gerealiseerd bij toepassing van 200 mM waterige oplossingen bij 120 °C.
De vaste residuen van de voorbehandeling zijn hierop gekarakteriseerd, waarbij
is aangetoond dat de kristalliniteit van cellulose wordt behouden en dat asvor-
mende elementen met uitzondering van Si praktisch compleet worden verwi-
jderd. De resultaten tonen verbeterde karakteristieken van de residuen met be-
trekking tot thermische conversie toepassingen in vergelijking met onbehandeld
tarwestro.
Hoofdstuk 7 behandelt de industriële processen voor de productie van fur-
fural, waarbij in het bijzonder wordt stilgestaan bij een innovatief proces dat is
gepatenteerd door de Technische Universiteit Delft en dat is gebaseerd op de re-
sultaten die staan beschreven in dit proefschrift. Zoals reeds is genoemd, zal de
geïntegreerde productie van furfural in moderne bioraffinaderijen zeer waarschi-
jnlijk de volgende hoeksteen vormen in de verdere ontwikkeling van de furfural
industrie. Het innovatieve procesconcept dat is beschreven in dit hoofdstuk is
gericht op een economisch haalbare en voor het milieu acceptabele furfural pro-
ductie met gereduceerde energie en toegevoerde chemicaliën consumptie. De
evaluatie van de proceseconomie toont veelbelovende resultaten in vergelijk-
ing met de bestaande processen, hetgeen dit concept economisch aantrekkelijk
maakt.
 Tenslotte worden in hoofdstuk 8 de conclusies gegeven, alsmede aanbevelin-
gen voor toekomstig verder onderzoek.

Gianluca Marcotullio
CONTENTS

List of Figures xxi

List of Tables xxiii
Nomenclature xxv
1 introduction 1
1.1 Biorefinery systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.1.1 The lignocellulose-feedstock Biorefinery . . . . . . . . . . . 5
1.1.2 Aspects of biorefinery integration . . . . . . . . . . . . . . . 7
1.1.3 The Furans . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.2 Furfural . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.2.1 Sources and production . . . . . . . . . . . . . . . . . . . . . 12
1.2.2 Advances in the synthesis of furfural from pentoses . . . . 14
1.3 Applications of furfural and its derivatives . . . . . . . . . . . . . . 16
1.3.1 Furfural as solvent . . . . . . . . . . . . . . . . . . . . . . . . 17
1.3.2 Tetrahydrofuran synthesis . . . . . . . . . . . . . . . . . . . . 19
1.3.3 Furfural as precursor for resins and polymers . . . . . . . . 20
1.3.4 Furfural as agricultural nematocide . . . . . . . . . . . . . . 21
1.4 Future perspectives of furfural as liquid fuels precursor . . . . . . 22
1.5 Economical aspects of furfural industry . . . . . . . . . . . . . . . . 24
1.6 Motivation and Scope . . . . . . . . . . . . . . . . . . . . . . . . . . 26
1.7 Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2 experimental methods 29
2.1 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.1.1 Analysis of the reaction products . . . . . . . . . . . . . . . 30
2.2 Experimental setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.3 Methods used for wheat straw pretreatment . . . . . . . . . . . . . 35
2.3.1 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.3.2 Wheat straw pretreatment procedures . . . . . . . . . . . . . 35
2.3.3 Analytical methods . . . . . . . . . . . . . . . . . . . . . . . . 36
3 reaction kinetics in furfural production 39
3.1 Furfural formation and destruction in acidic conditions . . . . . . . 40
3.2 Kinetics of furfural destruction . . . . . . . . . . . . . . . . . . . . . 41
3.3 Kinetics of furfural formation from xylose . . . . . . . . . . . . . . 45
3.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4 the effect of aqueous chlorides on furfural formation 49
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.2 Experimental results and discussion . . . . . . . . . . . . . . . . . . 51
4.2.1 Mechanism of furfural formation from pentoses . . . . . . . 54

xix
 4.3 Effect of chlorides addition on furfural kinetics . . . . . . . . . . . . 56
4.4 Catalytic requirements in furfural production . . . . . . . . . . . . 59
4.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
5 the effect of aqueous halides on furfural formation 61
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
5.2 Experimental results and Discussion . . . . . . . . . . . . . . . . . . 62
5.2.1 Sugar alcohols dehydration . . . . . . . . . . . . . . . . . . . 67
5.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
6 hemicellulose-derived carbohydrates via dilute-acid hy-
drolysis 69
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
6.1.1 Hemicellulose dilute-acid hydrolysis . . . . . . . . . . . . . 70
6.1.2 Dilute-FeCl3 for biomass hydrolysis . . . . . . . . . . . . . . 71
6.1.3 Characterization of the residues . . . . . . . . . . . . . . . . 71
6.2 Soluble carbohydrates recovery . . . . . . . . . . . . . . . . . . . . . 72
6.2.1 Crystallinity of the solid residues . . . . . . . . . . . . . . . 76
6.3 XRF and thermo-gravimetric analysis . . . . . . . . . . . . . . . . . 77
6.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
7 a novel process for making furfural 83
7.1 Current furfural production processes . . . . . . . . . . . . . . . . . 84
7.1.1 Operations of existing furfural production processes . . . . 84
7.2 Furfural production in modern biorefineries . . . . . . . . . . . . . 87
7.3 An innovative process for furfural production . . . . . . . . . . . . 88
7.4 Catalyst choice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
7.4.1 Homogeneous catalysis and recirculation . . . . . . . . . . . 90
7.4.2 Halides addition for optimal yields and separation . . . . . 92
7.5 Process simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
7.5.1 Results for a relevant process configuration . . . . . . . . . 94
7.6 Process economics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
7.6.1 Total cost of production estimate . . . . . . . . . . . . . . . . 97
7.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
8 concluding remarks 103
8.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
8.2 Recommendations for future research . . . . . . . . . . . . . . . . . 105
Bibliography 109
About the author 127
LIST OF FIGURES

Figure 1.1 Bioethanol production in Mm3 . Raw data from [1]. . . . . . 3

Figure 1.2 CO2 emission reduction potential and related cost. Adapted
from [2] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Figure 1.3 The sustainable biorefinery concept, adapted from [3] . . . 6
Figure 1.4 Simplified reaction pathway from biomass to furans, adapted
from [4]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Figure 1.5 Possible furfural conversion routes according to Kamm et al.
[5] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Figure 1.6 Competing furfural and petroleum-based routes to THF . . 19
Figure 1.7 Oil exchange price (CIF WTI), and production cost of gaso-
line and diesel expressed in US$/ton, source [6]. . . . . . . 25
Figure 1.8 Oil exchange price (CIF WTI), and production cost of gaso-
line, diesel and bioethanol in US$/GJ (LHV basis), source
[1, 6]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Figure 2.1 Representation of the experimental setup. The refractive
index detector, injector and the computer were only used
for the testing purposes described in this chapter, but not
during the experimental campaigns. . . . . . . . . . . . . . . 31
Figure 2.2 Dimensionless E(ϑ ) for different flow-rates, at T = 180o C . 32
Figure 2.3 Theoretical and real conversion rates when k = 0.1[min−1 ] . 33
Figure 2.4 Apparent reactor volume V ∗ against solution density ρ H .
(*) Averaged measured values at constant T; (· · · ) measure-
ment dispersion; (-) model used in this work . . . . . . . . . 34
Figure 2.5 TG-FTIR set-up at TU Delft Process and Energy Laboratory 37
Figure 3.1 Simplified scheme for furfural formation from D-xylose . . 41
Figure 3.2 k∗3 [s−1 ] Arrhenius plot in the temperature range 150-200 °C 44
Figure 4.1 Effect of NaCl addition on xylose reaction rate and furfural
yield. Xylose (dotted line) and furfural dimensionless con-
centration (solid line) as from equation (3.2) after fitting to
the experimental results. . . . . . . . . . . . . . . . . . . . . . 51
Figure 4.2 Effect of temperature on xylose reaction rate and furfural
yield in 150mM HCl. Xylose (dotted line) and furfural di-
mensionless concentration (solid line) as from equation (3.2)
after fitting to the experimental results. . . . . . . . . . . . . 52

xxi
Figure 4.3 Effect of different chloride salts on xylose reaction rate and
furfural yield. Xylose (dotted line) and furfural dimension-
less concentration (solid line) as from equation (3.2) after
fitting to the experimental results. . . . . . . . . . . . . . . . 54
Figure 4.4 Simplified reaction scheme involving chlorides . . . . . . . 57
Figure 4.5 ln(k X/[ H + ]) plot against temperature. . . . . . . . . . . . . . 58
Figure 5.1 Effect of different halides salts on xylose reaction rate and
furfural yield. Xylose (dotted line) and furfural dimension-
less concentration (solid line) as from equation 3.2 after fit-
ting to the experimental results. . . . . . . . . . . . . . . . . 64
Figure 5.2 Reaction mechanism leading from D-xylose to furfural in
acidic solutions. Aqueous halides are indicated as X − . . . . 66
Figure 5.3 Main products from the acid catalyzed dehydration of sugar
alcohols at high temperature. . . . . . . . . . . . . . . . . . 67
Figure 6.1 Representation of a portion of xylopyranose polymer, in-
cluding acetyl and arabinofuranose substituents. . . . . . . 71
Figure 6.2 Soluble carbohydrates production from wheat straw. In the
sample codes notations H and Fe indicate respectively di-
lute HCl and FeCl3 , and the following number their con-
centration in mM; LT and HT indicate the temperature of
pretreatment, respectively 100 and 120◦ C. . . . . . . . . . . 74
Figure 6.3 DTG plots of wheat straw residues of pretreatments com-
pared to the untreated, at 10 °C/min under nitrogen. . . . . 82
Figure 7.1 Schematic of the batch Quacker Oats process, adapted from
[7]. HPS, LPS: High and Low Pressure Steam. . . . . . . . . 85
Figure 7.2 Schematic of a Chinese furfural plant with capacity of 2.5
kton/y (6 reactors), adapted from [7]. . . . . . . . . . . . . . 86
Figure 7.3 Furfural-Water T-x diagram at 1 atm., adapted from [8] . . 89
Figure 7.4 Process sketch. The grey shaded region includes the equip-
ment parts exposed to a particularly corrosive environ-
ment, requiring thus appropriate corrosion-resistant con-
struction materials. . . . . . . . . . . . . . . . . . . . . . . . . 91
Figure 7.5 Furfural-water vapor liquid equilibrium in the dilute re-
gion. Measurement points for vapor (squares) and liquid
(circles) composition against the NRTL model (solid line)
at 1, 6 and 9.5 atm. Experimental data from [8, 9]. . . . . . . 93
Figure 7.6 Furfural partition coefficient in dilute furfural-water mix-
tures. Experimental data compared to the model. The grey
shaded region includes data of minor interest for furfural
distillation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
Figure 7.7 Main process streams of a relevant process configuration.
Negligible pressure drop over the columns and heat ex-
changers is assumed at this stage. . . . . . . . . . . . . . . . 95
L I S T O F TA B L E S

Table 1.1 Chemical composition (on dry basis) of common lignocel-

lulose feedstock. Source [5]. . . . . . . . . . . . . . . . . . . . 5
Table 1.2 Top chemical opportunities from integrated biorefinery sys-
tems, source [10, 11]. . . . . . . . . . . . . . . . . . . . . . . . 8
Table 1.3 Furfural production potential as estimated from the A.O.A.C.
method [8, 12]. . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Table 1.4 Comparison of heterogeneous catalysts used in the dehy-
dration of xylose to furfural. . . . . . . . . . . . . . . . . . . 15
Table 3.1 H3 O+ molar concentration [ H3 O+ ] and relative activity co-
efficient γ H3 O+ at various temperature and initial acid con-
centration as estimated by the eNRTL model . . . . . . . . . 43
Table 4.1 Condensed results from kinetic experiments. . . . . . . . . 55
Table 5.1 Results from kinetic experiments, all reaction were carried
out at 200 °C. . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Table 6.1 Biochemical composition of the untreated wheat straw, data
from [13]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Table 6.2 Sugars recovery in the filtrate expressed as wt% d.b. of
initial wheat straw. . . . . . . . . . . . . . . . . . . . . . . . . 75
Table 6.3 ash composition distribution in wt% of the most impor-
tant elements making up for >99 wt% of the total inorgan-
ics. C, H, O, and N are excluded. Values in italic indicate
the amount left compared to the untreated wheat straw.
These values are calculated on SiO2 basis, assuming that
no SiO2 is removed after pretreatments. This is a conser-
vative hypothesis and thus these values represent rather
an upper limit of the residual compounds rather than the
actual residue. . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Table 6.4 Thermo-Gravimetric Analysis of the solid residues. . . . . 80
Table 7.1 Summary of installed equipment cost . . . . . . . . . . . . . 97
Table 7.2 Breakdown of Fixed Capital Investment (FCI)a . . . . . . . . 98
Table 7.3 Summary of annual production cost . . . . . . . . . . . . . . 100

xxiii
N O M E N C L AT U R E

Abbreviations

a.r. As received
AOAC Association of Official Agricultural Chemists
CrI Crystallinity Index
CS Carbon steel
d.b. Dry basis
DTG Differential Thermogravimetric Plot
eNRTL electrolyte Non-Random Two-Liquids
FCI Fixed Capital Investment
FEMA Flavor and Extract Manufacturers Association
HPLC High Performance Liquid Chromatography
LCF Lignocellulose-feedstock
mM millimolar
M molar
[X] molar concentration of X
RON Research Octane Number
RTD Residence Time Distribution
SS Stainless Steel
TGA Thermo-Gravimetric Analysis
XRD X-ray diffraction
XRF X-ray fluorescence
wt Weight

Chemicals
DMSO Dimethylsulfoxide
FA Furfuryl alcohol
FDCA 2,5-Furandicarboxylic acid
HMFA 5-Hydroxymethyl furoic acid
HMF 5-Hydroxymethyl furfural
MF 2-Methyl furan
MTHF 2-Methyl tetrahydrofuran
THF Tetrahydrofuran
THFA Tetrahydrofurfuryl alcohol

xxv
Symbols

t Time [min]
a Ion activity in aqueous solution [mol l−1 ]
A Debye-Hückel costant [l0.5 mol−0.5 ]
C Concentration [mol l−1 ]
Ea Activation Energy [kJ mol−1 ]
∆H ‡ Enthalpy of activation [kJ mol−1 ]
I Ionic strength [mol l−1 ]
k Reaction rate constant [s−1 ]
ṁ Mass flow-rate [g min−1 ]
MO1 First momentum of the distribution curve E (t) [min]
Pe Péclet number [-]
pKa Sulphuric acid second dissociation constant [-]
R Universal gas constant [kJ mol−1 K−1 ]
∆S‡ Entropy of activation [J mol−1 K−1 ]
t Time [min]
T Temperature [K]
V Volume [l]
zi Ion charge

Greek Symbols
γ Ion activity coefficient [-]
ϑ Normalized residence time [-]
ρ Density [g l−1 ]
σϑ Variance of the distribution curve E (ϑ ) [-]
τ Average residence time [min]
 1
INTRODUCTION

In this chapter the opportunities offered by modern biorefinery in the broader context are
introduced, and the importance of furanic compounds within the biorefinery is pointed
out. In this context the enormous potential of furfural and its derivatives, both in the
chemical and energy sector, is discussed. Recent advances in furfural technology are
summarized, both regarding furfural synthesis and applications, eventually stating the
motivation behind this dissertation, and the main achievements therein contained.

1
2 introduction

To what extent is agriculture likely in the future to supply the raw materials for
industry in addition to food and textiles, her main customers today? The question arose
in a lecture given by Sir Harold Hartley to the Textile Institute in June 1937, later
published on Nature [14]. It is remarkable to find such a present-day question
in an almost 80 year-old lecture. In that same lecture the competition between
biomass derived and fossil resources was also very current: Of greater significance
from the point of view of displacement of plant products, is the use of cracked oil gas
and natural gas in the United States as starting-points for the manufacture of organic
chemicals [. . . ]. Here there is a direct competition with the products of the fermentation
industry and wood distillation. [. . . ] On the other hand, the more detailed knowledge
of plant products has led to great developments in their extraction and utilization, for
example, the process of fat-hardening, the conversion of molasses into alcohol, glycerin,
acetone and other solvents, and the production of furfural from cereals wastes for use as
solvent or as a constituents of plastics.
The terms of the discussion around bio-fuels, in particular on ethanol high cost
of production compared to gasoline, were also very clear back then: [bioethanol]
use today as motor fuel is justifiable on strategic grounds, but when petroleum supplies
begin to contract, alcohol will be in a better competitive position.
On the perspectives of bio-based material as substitutes for oil derivatives and
metals: The extent to which agricultural materials may replace minerals and metals will
depend on physical properties and cost. Some plastics are suitable for replacing metals;
the use of synthetic resins as adhesives, and the improvements in the manufacture of
laminated wood already make it a successful rival to metal in many fields, [. . . ]. In con-
clusion: Such a development of the use of agricultural products would be most fortunate
in view of the gradual exhaustion of our capital assets, coal and oil.
It is clear that the possibility of producing materials, chemicals, and fuels from
biomass had been envisioned already long ago. Unfortunately fossil resources,
in particular oil, have dominated last century both in the chemical and energy
sectors due to their enormous availability at relatively low cost. Hence only lim-
ited and specific bio-based sectors could survive.
In nowadays scenario bio-based chemicals and fuels are gaining a much more
favorable competitive position, not only by virtue of high oil prices and the
forecasts on its future availability, as predicted by Sir Harold Hartley, but also
because of the increasing environmental awareness and the demand for of a
secure, possibly domestic, energy supply.

1.1 biorefinery systems

The use of biomass for energy applications has been greatly promoted in the
recent years in many countries. As far as the biomass transformations to fu-
els and chemicals is concerned, the production of bioethanol in particular has
shown a very sharp increase [1], especially in the USA, next to Brazil which has
 1.1 biorefinery systems 3

100

90 Others
India
80
China
70 EU 27
Brasil
60
US

50

40

30

20

10

0
2003 2004 2005 2006 2007 2008 2009 2010

Figure 1.1: Bioethanol production in Mm3 . Raw data from [1].

a longer history of ethanol production from sugarcane molasses, see Figure 1.1.
Such massive production of the so called first generation biofuels contributed
also to inflame the debate on the competition between food and fuel, and more
in general on the sustainability of biomass use for energy applications, which
peaked together with food prices between 2007 and 2008. Important organiza-
tions such as the World Bank argued that the increase in internationally traded
food prices was caused by a confluence of factors, but the most important was
the large increase in biofuels production from grains and oilseeds in the US and
EU [15].
Furthermore, first generation bioethanol has been seriously challenged in terms
of actual CO2 emission reductions. An article appeared on Science in 2006 showed
that, in some particular cases where coal was employed to power the processes
for corn-based ethanol production and purification, the associated well-to-wheels
CO2 emissions resulted even higher than regular gasoline [16]. If, on the one
hand, the use of a domestic carbon-intensive energy resources, such as US coal,
for the production of ethanol could be justified on the strategic grounds of the
reduction of the dependence on imported oil, it turned out to be in clear contra-
diction with the environmental basis behind biofuels.
Apart from the particularly fortunate Brazilian case, it has been often recog-
nized that first generation biofuels, besides raising concerns on the competition
with food, may bring very limited environmental benefits [2, 16–18]. The proven
limited reduction of greenhouse gases emissions, next to the predicted increas-
ing pressure on food prices, water resources, biodiversity, forests and vulnerable
4 introduction

Figure 1.2: CO2 emission reduction potential and related cost. Adapted from [2]

lands, brought up the question whether, in some cases, biofuels, and in particu-
lar first generation bioethanol, could become a cure worse than the disease [19].
Concerning the EU context, the European Commission Joint Research Centre
studied the CO2 reduction potential for a variety of alternative solutions in the
transportation sector, showing that the impact on CO2 emissions and the associ-
ated costs spanned over a wide range [2], as shown in Figure 1.2.
In order to address all these issues, and drive the market towards the best
practices available, it became important to introduce sustainability criteria associ-
ated to the biomass use in transportation. The European Commission recently
approved the directive 2009/28/EC, which prescribes a minimum target of 35%
CO2 emissions reduction for biofuels and bioliquids put on the market, which
is set to increase to 50% in 2017 and 60% in 2018. In view of these very clear
sustainability targets the development of future biofuels productions shall prob-
ably shift towards new generation and more sustainable solutions, like cellulosic
ethanol, syn-diesel, furanic fuels, besides hydrogen and electricity.
Besides biofuels, also bio-based chemicals in the recent years gained a better
competitive position against oil derivatives. The spectrum of possible biomass
transformations routes into chemicals is very wide [4]. If on the one hand it is
mostly accepted that a selection of primary biomass-derived building blocks and
secondary intermediates could either replace or substitute oil derivatives in most
of existing industrial applications, on the other hand it is also commonly agreed
 1.1 biorefinery systems 5

Table 1.1: Chemical composition (on dry basis) of common lignocellulose feedstock.
Source [5].
LCF Cellulose (%) Hemicellulose (%) Lignin (%)
Hexoses Pentoses
Softwood 40-48 12-15 7-10 26-31
Hardwood 30-43 2-5 17-25 20-25
Cereal straw 38-40 2-5 17-21 6-21
Maize straw 35-41 2 15-28 10-17
Rape straw 38-41 - 17-22 19-22
Recovered paper 50-70 - 6-15 15-25

that conversion and separation technologies are mostly lagging far behind the
established oil-based refining industry [4, 10, 11, 20–22].
In view of all this, the attention is focused nowadays on the emerging field of
BIOREFINERY systems, aimed at an economically and environmentally sound
use of the biomass feedstock for the production of a whole range of products:
form high-volume commodities, such as fuels, to high added value specialty
chemicals [21].

1.1.1 The lignocellulose-feedstock Biorefinery

The term Biorefinery in general represents a complex system able to process dif-
ferent biomass feedstocks, via different technologies, in order to produce a multi-
plicity of products [23]. Among the potential large-scale industrial biorefineries,
the so-called lignocellulose-feedstock (LCF) biorefinery holds many characteris-
tics for future success. Lignocellulosic raw material, i.e. straw, reed, grass, wood,
paper-waste, etc., is nearly ubiquitous, and normally available at relatively low
cost. Carbohydrates, namely cellulose and hemicellulose, make up for about
70% of the dry weight of lignocellulosic biomass, see Table 1.1, and represent
the feedstock for nearly all of the most promising bio-based building blocks
and chemical intermediates, both deriving from biological and thermo-chemical
conversion routes [4, 10, 11, 21]. Lignin, another important constituent of ligno-
cellulosic biomass making up for about 25% of its weight, is by far the most
important natural resource of aromatics, beside a good solid biofuel.
Global photosynthetic production of biomass is enormous, in the order of 1.7-
2.0×1011 tons per year [24]. Needless to say, only a small fraction of it is nowa-
days harvested and used for food, feed and industrial applications. According to
the United Nations Food and Agriculture Organization statistic data, the world
production of cereals in 2008 accounted for 2.5×109 tons, sugarcane production
reached 1.7×109 tons, whereas the forestry-derived products, such as various
kind of industrial wood and wood fuel, accounted for 3.4×109 m3 . Noteworthy,
6 introduction

Figure 1.3: The sustainable biorefinery concept, adapted from [3]

such figures are comparable to the annual production of oil on weight basis,
3.9×109 tons (163 EJ) in 2008.
From the purely energetic point of view, Kim and Dale [25] estimated that
about 1.5×109 tons (25.5 EJ) of lignocellulosic residues associated to the main
agricultural crops (which are nowadays only partly used and mostly left behind
during harvesting) are globally available for the production of bio-based fuels
and chemicals, corresponding to a potential bioethanol production of 0.4×109
tons (10.7 EJ), in addition to substantial amounts of electricity and steam.
Next to agricultural residues, there is still a huge growth potential for par-
ticularly promising crops like sugarcane. The Brazilian Sugarcane Industry As-
sociation (UNICA) claims that only 10% of the suitable land for sugarcane is
presently used, the remaining being mainly pastures, primarily located in South
American regions not constituting a potential threat to the Amazons, and sub-
Saharan Arica.
Although sugarcane holds an enormous potential as future energy crop, this
is limited to specific regions of the world with appropriate climate conditions. In
temperate regions like Europe, North America and the Mediterranean area, se-
lected perennial rhizomatous grasses hold the highest potential as energy crops,
namely miscanthus (Miscanthus), reed canarygrass (Phalaris arundinacea), giant
reed (Arundo donax) and switchgrass (Panicum virgatum) [26]. These crops present
very high yields per hectare, up to 40 ton/ha annum of dry matter, and, unlike
annual crops, their need for soil tillage is limited to the year in which the crops
 1.1 biorefinery systems 7

are established. Due to the fast growth and reduced tilling, and also because of
a significant below ground carbon sequestration, they present significant benefi-
cial aspects firstly in terms of CO2 balance, but also on preventing soil erosion
and degradation. Furthermore, due to the recycling of nutrients by their rhizome
systems, perennial grasses have low demand for nutrient inputs, and little or no
pesticide use [26, 27], besides being able of growing on a wide range of soils,
including marginal or set-aside lands normally not suitable for agriculture.
Global biomass supply potential has been carefully estimated in the frame-
work on the Netherlands Research Program on Scientific Assessment and Policy
Analysis for Climate Change [28]. By taking into account the complex relation-
ships between food, energy, and water demand, beside land uses, biodiversity,
conversion processes, and many others, it was shown that biomass could con-
tribute up to 510 EJ/y to the global primary energy demand in 2050 (estimated
between 600 and 1040 EJ/y), considering:
• Agricultural and forestry residues (100 EJ/y).
• Forestry surplus (80 EJ/y).
• Energy crops production on good quality surplus agricultural and pasture
lands (120 EJ/y).
• Energy crops production on water scarce, marginal and degraded lands
(70 EJ/y).
• Learning in agricultural technology (140 EJ/y).
The selection of feedstock for currently developing LCF biorefineries projects
reflects the geographic location, with corn stover being a preferred choice in
the US, bagasse in the regions where sugarcane is already well established like
Brazil, and mostly cereal straw in EU. First movers are already looking at en-
ergy crops, and at the beginning of 2011 the northern-Italy based company
M&G started building a cellulosic ethanol pilot plant based of the conversion
of locally grown Arundo Donax. They claim a productivity of about 11 m3 of
ethanol per ha-annum, which is about double the present productivity of Brazil-
ian sugarcane-based ethanol [29, 30].
In view of what mentioned above, the availability of lignocellulosic feedstock
does not represent the bottle-neck for the medium-term development of biore-
fineries. The development effort is currently focused rather on suitable conver-
sion technologies, starting from the pre-treatment of raw biomass for its fraction-
ation into its main constituents: cellulose, hemicellulose and lignin.

1.1.2 Aspects of biorefinery integration

If on the one hand LCF biorefineries at this stage are focusing on the bulk pro-
duction of cellulosic ethanol as main output and source of revenues, on the other
 Table 1.2: Top chemical opportunities from integrated biorefinery systems, source [10, 11].
Alcohols Ethanol Besides the large-scale use as biofuel, ethanol is gaining interest as chemical platform, in particular as precursor
of ethylene.
Furans Furfural The potential of furfural as platform chemical will be discussed more extensively in this chapter.
HMF HMF is an appealing starting material for various chemical transformations.
2,5-FDCA 2,5-furandicarboxylic acid has been suggested as a potential substitute for terephthalic acid in the production of
polyesters. It is preferably produced by HMF oxidation.
Glycerol and derivatives Readily available from the biodiesel industry, glycerol offers many conversion opportunities to interesting
products via reduction, dehydration and fermentation.
Hydrocarbons Isoprene Isoprene is a high value hydrocarbon with a significant market. There are biochemical routes being developed
and brought to market for producing (bio)isoprene.
Biohydrocarbons Production of long chain hydrocarbons from microalgae is being explored, next to biological synthesis using
appropriate bacteria. Hydrocarbons in general will be of high importance for the development of the
biorefinery, providing a direct drop-in interface with the petrochemical industry.
Organic acids Lactic acid Lactic acid is a well-recognized commercial bio-chemical produced by glucose fermentation. Its primary use is
the production of polylactic acid.
Succinic acid Succinic acid production via biochemical transformation of sugars has been widely studied, and is close to
commercial scale. Succinic acid is a high potential platform chemical for the production of C4 intermediates.
Hydroxypropanoic 3-Hydroxypropanoic acid and 3-hydroxypropionaldehyde (HPA) may be produced from glycerol via biological
acid / aldehyde routes and offer several conversion opportunities to valuable products, such as 1,3-propanediol and acrylic acid.
Levulinic acid Of interest as derived with relatively high yields from C6 sugars (via HMF). Transformation of levulinic acids
into substituted pyrrolidones; lactones; levulinate esters; diphenolic acid; and ketals is being studied.
introduction

Sugar alcohols Sorbitol Chemical reduction of glucose is well-established, although the biochemical route is investigated. Of interest for
the production of drop-in hydrocarbons via aqueous phase reforming, but also for the production of isosorbide.
Xylitol Preferably prepared by catalytic hydrogenation of xylose, although biochemical reduction is being investigated.
It is of interest the aqueous phase reforming to hydrocarbons and the transformations to polyols.
8
 1.1 biorefinery systems 9

hand their integration with value-added chemicals production does not present
yet a winning set of technologies, and operators are mostly screening for the
most promising options.
Bio-based chemical production is challenged by an overabundance of possible
products. Integrated biorefinery development has yet to identify a core group of
primary chemicals and secondary intermediates analogous to those used by the
petrochemical industry. A valuable effort was done in this direction by the US
Department of Energy, which produced a systematic study aimed at selecting
the most promising building blocks based on the following criteria [10]:

1. The compound or technology has received significant attention in the lit-

erature. A high level of reported research identifies both broad technology
areas and structures of importance to the biorefinery.

2. The compound illustrates a broad technology applicable to multiple

products. As in the petrochemical industry, the most valuable technolo-
gies are those that can be adapted to the production of several different
structures.

3. The technology provides direct substitutes for existing petrochemicals.

Products recognized by the chemical industry provide a valuable interface
with existing infrastructure and utility.

4. The technology is applicable to high volume products. Conversion pro-

cesses leading to high volume functional equivalents or utility within key
industrial segments will have particular impact.

5. A compound exhibits strong potential as a platform. Compounds that

serve as starting materials for the production of derivatives offer important
flexibility and breadth to the biorefinery.

6. Scale-up of the product or a technology to pilot, demo, or full scale is

underway. The impact of a biobased product and the technology for its
production is greatly enhanced upon scaleup.

7. The biobased compound is an existing commercial product, prepared

at intermediate or commodity levels. Research leading to production im-
provements or new uses for existing biobased chemicals improves their
utility.

8. The compound may serve as a primary building block of the biorefinery.

The petrochemical refinery is built on a small number of initial building
blocks: olefins, BTX, methane, CO. Those compounds that are able to serve
an analogous role in the biorefinery will be of high importance.
10 introduction

Figure 1.4: Simplified reaction pathway from biomass to furans, adapted from [4].

9. Commercial production of the compound from renewable carbon is well

established. The potential utility of a given compound is improved if its
manufacturing process is already recognized within the industry.
The screening resulted in a first selection of 15 candidates in 2004, starting from
a longer list of 30 [11]. In 2010 the list was revised with the elimination of some
products which turned out to be of limited interest during these years, and the
introduction of others, resulting in a new list of 13 compounds based on more
updated developments [10]. The 2010 list is depicted in Table 1.2.

1.1.3 The Furans

As shown in Table 1.2, the family of furans is of high importance for the inte-
grated biorefinery. The dehydration of 5- and 6-carbon sugars to give furfural
and (HMF) is a well-known transformation since 1832 (although the first obser-
vations may be dated back in 1821) when Dobereiner first reported the isolation
of a few drops of yellow oil from the distillate obtained during the preparation
of formic acid from sugar by the action of MnO2 and H2SO4. Later such “oil”
was assigned the correct formula C5H4O2, but only in 1845 it was named Fur-
furol by the English chemist G. Fownes (Furfur-bran; oleum-oil), then changed
to Furfur-al in recognition of its aldehydic nature [8].
The chemistry of the broader family of furan compounds is even older than
furfural, as furoic acid was first obtained in 1780 by dry distillation of galactaric
 1.1 biorefinery systems 11

acid [8]. Already in 1953 an encyclopedic survey on the chemistry of the furans
was written by Dunlop and Peters who had been involved in the field for many
years within The Quaker Oats Company, first industrial producer of furfural [8].
Although many compounds of the furan series have been encountered in con-
nection to studies of natural resources, such as furfural, HMF, furan, 5-methyl
furfural, furoic acid, 5-methyl furan and others, only furfural, along with its
main by-product 5-methyl furfural, is currently commercially produced from
biomass. For this reason furfural represents nowadays the starting material in
the manufacture of most of the furan compounds [31].

5-hydroxymethyl furfural (hmf) HMF is the furfural 6-carbon equiv-

alent. The presence of two functional groups, combined with the furan ring,
makes HMF an appealing starting material for various chemical transformations.
Although the chemistry behind HMF formation is comparable to furfural, and
hexoses are even more abundant and easily available than pentoses, contrary to
furfural HMF has never been produced at industrial scale.
Interestingly, since the earliest observations, furfural could be easily recovered
in a steam distillate and thus purified by virtue of its limited solubility in wa-
ter. It is a matter of fact that the peculiar thermodynamics of the furfural-water
system made it possible to discover furfural in the first place - from a milky dis-
tillate -, and to produce it at industrial scale by rapidly distilling the aldehyde from
the reaction chamber during the period of its formation [8]. Unlike furfural, HMF is
hydrophilic and only slightly volatile in aqueous solutions, therefore its recovery
via steam distillation is practically impossible, and also liquid-liquid extraction
results unfavorable. This fundamental difference, and the fact that HMF, like
furfural, undergoes further reactions under the conditions necessary for its for-
mation, makes any attempt of producing and purifying HMF from raw biomass
inherently challenging [32].
During the last years an enormous research effort has been devoted to the
synthesis and isolation of HMF, which may be grouped by catalysis type: metal
catalysis and acid catalysis, the latter including homogeneous liquid, heteroge-
neous liquid–liquid, solid-liquid and gas-liquid. Such effort resulted in consider-
able improvements for the conversion of fructose to HMF, whereas the transfor-
mation of glucose, sucrose and cellulose remains challenging. For an extensive
coverage of the topic of HMF synthesis and applications the reader might refer
to the comprehensive reviews recently published [32, 33].
Among the vast work on HMF synthesis and isolation, remarkable results have
been achieved by following alternative approaches, i.e. obtaining selected furan
derivatives presenting good chemical stability and favorable purification options.
Mascal and co-workers obtained the substitute furan 5-(chloromethyl)furfural
with relatively high yields directly from cellulosic material, and from this in-
termediate could then isolate HMF and other furan derivatives [34, 35]. Fol-
lowing a similar strategy, the Dutch company Avantium has developed a very
12 introduction

promising process aimed at forming a stable HMF ether derivative, namely 5-

(alkoxymethyl)furfural, reacting hexoses in presence of an alcohol and an acid
catalyst [36, 37]. They firstly highlighted the remarkable properties of such ethers
as liquid fuels or fuel additives for transportation and aviation, although inter-
esting options also exist for bio-based polymers applications [38]. Avantium is
presently building a pilot plant in the Netherlands for demonstrating their tech-
nology [39].

1.2 furfural

As already mentioned, furfural is the only compound of the furan series being di-
rectly obtained from biomass at industrial scale. Furfural production is generally
carried out by hydrolysis of hemicellulose-derived pentosans into monomeric
pentoses, and their subsequent acid-catalyzed dehydration into furfural. An ex-
tensive discussion on the mechanistic aspects of furfural formation from pen-
toses is presented in chapters 4 and 5.
The first time furfural was produced in relatively large amounts was at the
beginning of 1922 in the USA by The Quaker Oats Company, and the conse-
quent development of the furfural industry achieved a relative maturity already
before the first half of the century [40–50]. With a global production of about 300
kton/y furfural is still currently the sole precursor of furyl, furfuryl, furoyl, or
furfurylidene compounds in the chemical industry [31], and actually one of the
most important biomass derived chemicals.

1.2.1 Sources and production

Furfural can be obtained by every pentosan-containing material, i.e. by cellulosic

material of all kind. Curiously, being formed on heating carbohydrates, furfural,
together with HMF and other furans, may be easily found in cooked, roasted
or baked goods, resulting in an appreciable daily human intake. The highest
concentrations of natural occurring furfural are found in cocoa and coffee (55-255
ppm), alcoholic beverages (1-33 ppm), and in wholegrain bread (26 ppm). The
total potential furfural daily intake from food has been estimated in the order
of 0.3 mg/kg/day [51]. Being used as flavor ingredient, furfural underwent a
FEMA assessment in 1997, and it was determined to be generally recognized as
safe (GRAS) [51].
Even though pentosans, namely xylans and arabinans, are present in virtually
all plant materials as constituents of hemicellulose, their concentration varies
depending upon the hemicellulose typology, covering a relatively wide range,
see also Table 1.1.
 1.2 furfural 13

Arabinoxylans Typical of annuals and grasses, contain a predominant xylan

sugar residue in the backbone, with minor amounts of arabinofura-
noside and ester-linked acetyl group substituents [52].

Glucoronoxylans Mostly found in hardwoods, have also a predominant pres-

ence of xylans, together with significant concentrations of glucuronic
acids and acetyl substituents, with a molar ratio of about 10:1:7 [52].

Galactoglucomannans In contrast to hardwoods and herbaceous, softwood hemi-

cellulose is dominated by hexosans rather than pentosans, with galac-
toglucomannans representing roughly 15–20% of the biomass dry
weight, and xylans comprising only 7–10% [52, 53].

Therefore, arabinoxylans containing materials, such as annuals and grasses, are

more suitable for furfural production, see Table 1.3. Today, like in 1922, the large
availability, low cost and relatively high pentosan content of agricultural residues
made them the feedstock of choice in the furfural industry. Currently corncobs
and sugarcane bagasse are the dominating feedstock, the former mainly used
in China and Thailand, whereas the latter is used in Dominican Republic and
South Africa [31, 54].
The majority of current production is still based on more or less modified
versions of the original Quaker Oats process. They mostly consist of suitable di-
gesters loaded with acid-impregnated biomass, where steam is injected through
the biomass bed at an appropriate pressure and temperature, while a furfural
enriched vapor stream is constantly withdrawn. The furfural-containing stream
is subsequently concentrated via azeotropic distillation and rectified via vacuum
distillation, whereas the solid biomass residues are normally incinerated or dis-
posed of. Most of the processes are run in batch mode. A more extensive discus-
sion on furfural production process is presented in Chapter 7.
Currently the largest plant for the production of furfural (35 kton/y) is lo-
cated in Dominican Republic, contiguous to a sugar mill owned by The Central
Romana Corporation, a big sugar producer; whereas most of the world produc-
tion capacity is located China, mainly consisting of a large number of producers
with a capacity in the order of only few kilotons per year.
Many new concepts for the production of furfural have been developed and
described in literature [7, 55], demonstrating a substantial research effort, al-
though only limited progresses may be recorded on the ground of operating
industrial plants.
For reasons that can be related to their limited technological evolution, the pro-
duction processes in use today generally suffer from low yields (around 50%),
besides significant economical and environmental concerns. All these reasons
certainly contributed to the shrinking of furfural production capacity in EU and
USA down to an only residual presence, and hindered the expansion and mod-
ernization of the furfural industry below its actual potential. A profound techno-
14 introduction

Table 1.3: Furfural production potential as estimated from the A.O.A.C. method [8, 12].
Raw material Furfural % wt dry basis

Agricultural residues and herbaceous

Corncobs 23.4
Oat hulls 22.3
Bagasse 17.4
Olive press cake 16.6
Straw 16.0
Beet pulp 17.0
Arundo Donax 17.9

Hardwoods
Birch (paper) 13.2
Maple (red) 10.1
Beech (American) 11.5

Softwoods
Spruce (white) 7.0
Pine (jack) 7.4
Fir (balsam) 6.3

logical development is a priority in order to upgrade furfural to a large-volume

bio-based commodity.

1.2.2 Advances in the synthesis of furfural from pentoses

The innovation effort in the field of furfural production covers few main aspects,
with reaction catalysis certainly being the most active area, especially focused
on the use of solid acids, but also non-aqueous reaction media are being tested,
besides innovative options for furfural separation.
Many researchers consider the use of homogeneous catalysts, such as aqueous
sulfuric or hydrochloric acid, to be the main disadvantage in furfural production.
In particular, many environmental and economic concerns are associated to the
use of large amounts of strong mineral acids, along with neutralizing agents
and the consequent spent wastes [38, 54]. In this view several attempts have
been made towards the introduction of heterogeneous catalysis in furfural pro-
duction, with the particularly active contribution of the group of prof. Valente
and co-workers at University of Aveiro. A variety of solid acid catalysts, namely
 1.2 furfural 15

Table 1.4: Comparison of heterogeneous catalysts used in the dehydration of xylose to

furfural.
Catalyst Solvent Temp. Res. time S /% Y /% Ref.
MCM-41-SO3Hc W-T 140°C 24 h 83 76 [60]
eHTiNbO5-MgO W-T 160°C 4h 60 55 [61]
PSAZ-MCM-41 W-T 160°C 4h 49 39 [64]
ZSM-5 W 200°C 0.3 h n.a. 46 [57]
Nafion-117 DMSO 150°C 2h 66 60 [66]
Dealumin. HNu-6(2) W-T 150°C 4h 53 48 [62]
H-mordenite 13 W-T 260°C 0.05 h 98 98 [58]
H-mordenite 11 W-T 170°C 0.5 h 96 26 [56]
SO42− /12%ZrO2-Al2O3/SBA-15 W-T 160°C 4h 53 53 [65]
Amberlyst-15 / Hydrotalcite N,N-DMF 100°C 1h 51 37 [67]
Furfural selectivity (S) and yield (Y) are highlighted. Water-Toluene (W-T), Water (W),
Dimethylsulfoxide (DMSO), and N,N-dimethylformamide (N,N-DMF) are used as sol-
vent systems.

zeolites [56–58]; microporous and mesoporous niobium silicalites [59]; micro-

mesoporous sulfonic acids [60]; layered titanates, niobates and titanoniobates
[61]; delaminated aluminosilicates [62]; cesium salts of 12-tungstophosphoric
acid and mesoporous silica-supported 12-tungstophosphoric acid [63], bulk and
mesostructurated sulfated zirconia [64, 65], nafion 117 [66] and a combination
of different acid and basic solid catalyst [67], have been tested for the dehy-
dration of xylose to furfural. In general, the results achieved so far in the area
of solid catalysis in furfural production are quite promising as shown in Table
1.4, especially in view of the future improvements expected by fine-tuning of
the catalysts properties and reaction conditions [38]. In particular, outstanding
results have been reported when using zeolites at relatively high temperatures
(Table 1.4, entries 4,7 and 8), both in terms of furfural selectivity, yield, and short
residence time. Nevertheless research in this field is developing quickly, and,
although strong indications already exist, it is too early to select the best solid
catalyst. Moreover, in view of the industrial implementation of solid catalysts,
many aspects are still to be carefully evaluated, such as: catalysts deactivation
and regeneration; water compatibility; leaching and poisoning; and economic
viability. For a more extensive discussion on this topic the reader might refer to
the published reviews [38, 68].
When solid acids are employed the reaction is usually carried out using an
aqueous–organic biphasic solvent system. In the aqueous phase the reaction
of sugars takes place, whereas the organic phase is used, as alternative to the
traditional steam distillation, to readily extract the furfural formed so to avoid
its uncontrolled decomposition. Toluene is normally a preferred choice as or-
16 introduction

ganic extracting agent due to its high affinity with furfural, although methyl-
isobutylketone (MIBK), isobutyl acetate, ethyl acetate, THF and a series of C5
alcohols have been tested for this purpose [69–72]. Supercritical CO2 has also
been tested for furfural extraction with interesting results [73–76], whereas or-
ganic [77, 78], polar aprotic solvents [79], and ionic liquids [80] have been studied
in replacement of the aqueous phase. Furfural formation in uncatalyzed liquid
hot water [81, 82], and with the application of microwave has also been recently
reported [83, 84].

1.3 applications of furfural and its derivatives

The combination of the furan ring and the aldehyde function, make furfural
a very versatile chemical with peculiar solvent, resin precursor and biological
properties. In all these applications furfural may be used as such or, more often,
in the form of its derivatives of hydrogenation, mainly furfuryl alcohol (FA). The
number of potential uses of furfural, as reported in literature, is gigantic [5, 7, 8,
31, 85], see Figure 1.5. The reason why current uses of furfural are limited to few
main applications is the combined result of several factors, certainly including
the enormous expansion of the petrochemical industry during last century, but
also the limited evolution of the furfural industry.
The furan chemistry lived its golden age during the first half of the 20th cen-
tury, and an unmatched review book published in the monographic series of
the American Chemical Society in 1953, entitled The Furans [8], testifies to the
surprising activity in this field. Such volume, regrettably not printed anymore,
contained thousands of references to scientific publications, and a vast collec-
tion of patents covering the field of furans, which in 1953 had already reached
the number of 3500. Despite the enormous effort resulting in a more than 800
pages, the authors, by their own admission, could not achieve a full coverage
on all the aspects regarding the furans. Yet they achieved an impressive collec-
tion of synthesis methods and properties regarding furfural; furan and homol-
ogous; halogen and nitro derivatives; furanols and furylamines; furan metallic
compounds; furan alcohols, acids, aldehydes and ketones; and an enormous
number of derivatives of hydrogenation. Moreover they reported many fields of
application where furans showed interesting results: as chemical intermediates;
pharmaceuticals; fungicides and preservatives; insecticides; herbicides; selective
solvents; and resins and polymers precursors. Although out-of-date, every re-
searcher involved with furans today should look through this very intriguing
book, at least to avoid the very likely and frustrating chance of repeating some-
thing already done before 1953.
A clear sign of decline of the furan chemistry came in 1961, when Du Pont
abandoned its furfural-based process for the production of nylon, preferring the
more economical petroleum derivatives.
 1.3 applications of furfural and its derivatives 17

Interest around furans slowly picked up during the last couple of decades,
and it is now increasing due to the high oil prices, the environmental drive,
and the widely proven versatility of this series of chemical compounds. A very
interesting and updated book on furfural and its derivatives [7], and a huge
number of recent scientific contributions testify to the renewed interest.
In the following paragraphs an attempt is made to summarize the more im-
portant current uses of furfural and derivatives, and the upcoming fields of ap-
plication.

1.3.1 Furfural as solvent

Furfural has excellent solvent characteristics due to some very important aspects
descending from its chemical structure:

- The aromatic character of the furan ring and its polarity gives furfural
good solvent selectivity toward aromatics, and in general unsaturated
compounds.

- Furfural has an intermediate polarity, and for this reason it is only

partially soluble both in highly polar substances like water, and in
highly apolar substances like saturated hydrocarbons. Furfural pres-
ents a miscibility gap with many C6 or higher paraffins and olefins
[8]. Because of this peculiar property furfural is a versatile extracting
agent easy to recover by steam distillation.

- Despite the misleading darkening of furfural at room conditions (which

is due to the formation of very minor amounts of coloring com-
pounds), furfural exhibits very good thermal stability at most operat-
ing temperature levels [7, 8].

Because of its properties, furfural has been used as selective solvent in many
applications [7, 31, 47, 85, 86]. The knowledge on the processes involving furfural
or furan compounds as solvents in industry is vast, and reviewing it goes far
beyond the scopes of this thesis. Nevertheless a short collection of the most
common applications may be exposed here.

furfural in the refining industry The extraction of aromatic compo-

nents by using selective solvents is currently one of the basic operations in the
process of obtaining lubricating oils, and furfural is the solvent most commonly
used [87]. Such practice was started in 1933 by a subsidiary of The Texas Com-
pany [8], and it is aimed at achieving a low viscosity index, i.e. low viscosity
variations of the oil with temperature, by reducing its aromatic content via fur-
fural extraction. The content of sulfur compound was also reduced.
18 introduction

Figure 1.5: Possible furfural conversion routes according to Kamm et al. [5]
 1.3 applications of furfural and its derivatives 19

Based on the same principle, solvent extraction of aromatics and sulfur com-
pounds may be applied for diesel fuel upgrading. It has been reported that fur-
fural extraction reduced extensively virgin gas-oils sulfur content and improved
its burning characteristics such as the cetane number [8].
The selective action of furfural towards unsaturated hydrocarbons is exploited
also in the refining of lighter hydrocarbons. Namely, in the purification of 1,3-
butadiene from a complex mixture of C4 hydrocarbons, where simple distillation
would not be practical due to the very similar boiling points of mixture compo-
nents. The addition of a selective solvent, such as furfural, acts changing the
relative volatility of the different compounds in the mixture, enabling their sep-
aration by extractive distillation [7, 8, 31].
Due to its solvent characteristics, furfural may also be used in vegetable oil
refining, carboxylic acids extraction from aqueous solutions, and also as reactive
solvent and wetting agent in the manufacture of phenolic resins, abrasive wheels,
brake linings and refractory products [31].
Some furfural derivatives of hydrogenation are also important industrial sol-
vents. In particular tetrahydrofurfurul alcohol (THFA) is used as water-miscible,
high-boiling, biodegradable solvent for dyes, printing inks, pesticides and her-
bicides [31]. Another interesting product of furfural hydrogenation is methylte-
trahydrofuran (MTHF), which is gaining interest as solvent in organometallic
reactions where a strong Lewis base such as tetrahydrofuran (THF) is required.
Due to the partial mutual solubility and the low boiling azeotrope of MTHF/wa-
ter mixtures, MTHF can be used to conveniently recover and dry the products
of reaction, and be easily recycled [88].

1.3.2 Tetrahydrofuran synthesis

Figure 1.6: Competing furfural and petroleum-based routes to THF

Among all the compounds of the furan series, tetrahydrofuran (THF) deserves
a particular mention, as, contrary to what could be expected, it is not derived
from furfural or any other furan intermediate.
With an annual production in the order of one Mton, THF largely exceeds
furfural production, and it is mainly used as precursor of polytetrahydrofu-
ran (PTHF), important in the production of thermoplastic polyurethanes, elastic
fibers e.g. Spandex or Lycra, molded elastomers, and copolyesters or copolyamides.
20 introduction

A smaller proportion finds use as a solvent, as an extracting agent, and as the

preferred medium for organometallic syntheses [7, 89].
Preferential industrial processes for obtaining THF involve the catalytic dehy-
dration and cyclization of 1,4-butanediol to THF, see Figure 1.6. The key interme-
diate 1,4-butanediol may be obtained via different processes, all involving fossil
resources [89].
The more logical, bio-based, process involving the decarbonylation of fur-
fural to furan and its subsequent hydrogenation to THF, is very well-known
as it was employed between 1949 and 1961 by Du Pont in the manufacture of
nylon [8]. The Du Pont process involved four steps with furan, THF, and 1,4-
dichlorobutane as intermediates to adiponitrile, but was abandoned in 1961 be-
cause tetrahydrofuran could be made more economically from petrochemicals
[31].
Interestingly in the last years, due to high oil prices and the local availability of
raw material, the furfural-based process for making THF is being reintroduced
in China with an initial capacity of 35 kton/y and substantial expansion perspec-
tives [90].

1.3.3 Furfural as precursor for resins and polymers

Polymerization of furfural is a well-known reaction, but it has no industrial

importance [31]. On the other hand, furfural copolymers with phenol, phenol-
formaldehyde resins or ketones have drawn significant interest [7, 91]. In particu-
lar, when furfural is used in substitution of formaldehyde in phenol-formaldehyde
formulations, the ensuing resins exhibit interesting properties such as improved
flow/cure characteristic, solvent tolerance and reduced brittleness, although the
polymerization reaction, especially when acid-catalyzed, is more difficult to con-
trol [7, 8].

furfuryl alcohol resins When considering the current production of

resins from furan compounds, FA is by far the dominating precursor used for
numerous industrial applications. The mechanism of FA acid-catalyzed poly-
merization has been studied extensively for decades, and the complex structure
of poly-FA has been elucidated by Gandini and co-workers [92]. The peculiar
characteristic of FA-derived resins is a highly cross-linked structure giving them
good mechanical and thermal properties, alongside an outstanding chemical re-
sistance. Such resins are used in a variety of applications including sand cores
and molds for metal casting; corrosion-resistant fiberglass-reinforced plastics;
low flammability and low smoke generating composites and foams; carbona-
ceous products; corrosion-resistant polymer concretes and impregnating agents
for wood modification [91]. Although the patent literature covering the subject
of FA- and furfural-based resin formulations for different uses is vast, the use of
 1.3 applications of furfural and its derivatives 21

FA resins in the worldwide foundry industry remains by far the leading applica-
tion. The production of FA-based resins for use as binders to produce sand cores
and molds for metal castings is currently responsible for the majority of furfural
consumption [31].

wood modification Next to the traditional use in the foundry industry,

another application for FA resins is recently gaining commercial interest, i.e. its
use as impregnating agent for the production of modified wood. Research con-
cerning modification of wood with FA, sometimes referred to as “furfurylation”,
was initiated by Dr. Alfred Stamm in the early 1950’s, but only recently such
technique was brought to commercial scale thanks to the pioneer work of few
companies, such as the Norwegian Kebony ASA [93, 94].
Furfurylation, originally developed to substitute the toxic - in some countries
now banned - CCA (Copper Chromium Arsenic) treatments, gives to wood an
improved dimensional stability, hardness, moisture barrier and resistance to mi-
crobial and insect decay, by direct grafting of the FA polymer onto the wood
cell walls [95, 96]. The result is a durable, non-toxic, heavy metals-free, directly
disposable wood with a very good environmental profile. European or North
American wood species treated in this way could be conveniently employed in
substitution of tropical teak. The commercial development of furfurylated wood
is at its infancy, with an annual production in the order of few thousand tons,
but the potential growth is enormous as industrial wood consumption is in the
order of 0.9×109 m3 [97].

furan-based polymers Besides resins, furans are increasingly seen as build-

ing blocks for the production of copolymers, polyesters and Diels-Alder systems
from renewable resources, capable of replacing fossil-based conventional and
high-tech materials. The possibilities of furan chemistry are vast, and the ex-
ploitation of furan derivatives for the synthesis of macromolecular materials can
be planned strategically in a fashion that resembles closely the approach adopted
in the petrochemical industry. A whole original area of polymer science can be
built starting from the first-generation furan building blocks, i.e. furfural and
HMF. For an overview of the most recent advances in furan polymers the reader
may refer to the work of Gandini and co-workers [92, 96, 98, 99].

1.3.4 Furfural as agricultural nematocide

It is known from literature that furfural may control plant-parasitic nematodes

without adversely affecting crop growth and yield [100]. Very interestingly, it has
been shown that furfural do not kill nematodes directly, but act stimulating the
development of antagonistic bacteria, thus controlling nematodes in a biological
fashion [7]. In view of the phasing out of methyl bromide, and considering that
22 introduction

furfural has low acute and ecological toxicity, is non-systemic (not taken up
by plants) and safely applicable to soils via water solutions, it is likely to gain
a very interesting position as biomass-derived active ingredient in nematocide
formulations for agriculture.
Based on these results, the South African sugar and furfural producer Illovo
Sugar Ltd. introduced a furfural-based agricultural nematocide with the trade
name of Crop GuardTM , which is already registered at the South African Depart-
ment of Agriculture for different crops. Data have also been submitted to the US
Environmental Protection Agency for the same formulation under the name of
Multiguard ProtectTM by Agriguard Company LLC, USA.

1.4 future perspectives of furfural as liquid fuels precursor

The focus on biofuels for transportation covers a much wider spectrum than
ethanol and common biodiesel, incorporating virtually the whole range of bio-
mass transformation technologies [101]. Furfural and furans are no exception,
and they have been often addressed as interesting sugar-derived intermediates
for the synthesis of biofuels or fuel additives.
In a purely chemical view, biomass derived sugars are highly oxygenated
molecules, and need necessarily to be dehydrated, condensed to larger molecules
and thus hydrogenated to higher hydrocarbons similar to the constituents of
gasoline, diesel and jet fuel. In this perspective, it is noteworthy that both pen-
toses and hexoses lose three water molecules upon reaction to furfurals, at the
same time totally preserving their carbon content (opposite to ethanol), and heat-
ing value. The resulting unsaturated, lower oxygen content (furan) molecules,
although not indicated for direct use as transportation fuels, are very interesting
starting materials for intriguing synthetic pathways to compounds with superior
fuel characteristics [102–104].

furfural derivatives of hydrogenation A straightforward method

for upgrading furfurals to suitable liquid fuels is by direct hydrogenation/hy-
drogenolysis. In this way molecules with higher chemical stability may be at-
tained, presenting also higher heating values due to the hydrogen uptake and
eventual oxygen elimination.
The technical knowledge on catalytic hydrogenation of furfural is well estab-
lished, and by a careful selection of the catalysts and operating conditions quan-
titative yields of selected molecules may often be approached [7, 8, 31, 45, 58].
Thanks to such selectivity, a whole set of ensuing molecules may be obtained,
with a variety of key properties such as octane number, energy density, volatility,
polarity, viscosity, water solubility and oxygen content, that might be rationally
exploited when formulating novel and optimized fuel blends.
 1.4 future perspectives of furfural as liquid fuels precursor 23

Apart from the odd use of FA as main component of the rocket propellant
called “furaline” [7, 8], the first reported tests on furan derivatives as regular
gasoline additives may be dated back to 1941, when furfural and few selected
derivatives of hydrogenation, namely FA, 2-methylfuran (MF), THFA and MTHF
were tested in blends with gasoline [105]. MTHF resulted the only derivative pre-
senting a proper chemical stability and an octane number comparable to - 1940’s
- gasoline, thus eligible to be used as liquid fuel. Nevertheless, the MTHF unsatu-
rated homologue MF exhibited higher octane number and no appreciable gums
formation up to about 20 vol% mixture with gasoline, whereas FA, despite an
octane rating comparable to MF, presented an objectionable (but predictable) ten-
dency to polymerization. THFA showed an octane number comparable to MTHF,
but, unexpectedly, a higher polymerization tendency was reported, probably due
to its lower volatility rather than its reactivity.
Furans use in liquid fuel blends remained apparently dormant until recently,
when MTHF was picked up as one of the main components of the so-called P-
series alternative fuels [106–109]. Despite its relatively low RON (86), MTHF has
been found to favor a better blending of ethanol with C5 hydrocarbon mixtures,
resulting in an alternative fuel formulation with optimized properties that could
be directly used in flexi-fuel vehicles [110, 111]. MF has also been reconsidered
as bio-based octane enhancer due to its remarkable RON (131), low water solu-
bility and good energy density [58, 104], whereas THFA has been used in low
concentrations as anti-icing agent in jet fuels in Russia, due to its melting point
below -80°C [112, 113].

synthetic fuels derived from furans Besides direct hydrogenation,

furfurals well-known reactivity offers more options to obtain synthetic liquid
fuels. By controlled reactions involving furfurals, such as aldol condensation,
alkylation and etherification, larger molecules may be obtained, that can un-
dergo subsequent hydrogenation/hydrogenolysis to higher alkanes. Due to the
variety of possible synthetic pathways, a whole range of ensuing molecules may
be obtained for direct replacement of gasoline, diesel and also jet-type fuels.
Prof. Dumesic and co-workers at University of Wisconsin have done extensive
work in this sense, testing numerous controlled condensation reactions of fur-
furals with acetone, or their self-condensation after ring hydrogenation [104, 114–
118]. The resulting higher-carbon content molecules underwent a subsequent hy-
drogenation/hydrogenolysis step to yield higher alkanes with interesting overall
yields.
A similar approach has been recently used by prof. Corma and co-workers
[119], who exploited the reactivity of the furan ring at the 5th position for the
hydroxyalkylation/alkylation of MF with butanal or other aldehydes. The result-
ing molecules underwent subsequent hydrogenation/hydrodeoxygenation steps
to obtain branched alkanes with comparatively high yields. The organic fraction
24 introduction

obtained after such multi-step process presented an excellent pour point (-90°C)
and cetane number (70.9), thus excellent diesel-range characteristics.
Similarly, the Amsterdam based company Avantium developed an interesting
approach to furanic biofuels, in particular focusing on several 2- or 2,5-ring-
hydrogenated furan ethers. The compounds selected present very interesting
properties in terms of solubility with regular diesel, cetane number, heating val-
ues and good results on engine tests. Due to their properties they also represent
an option as bio-based jet fuels [36, 37, 39]. Of particular interest is the furfural
and ethanol derivative ethyltetrahydrofurfuryl-ether, presenting superior cetane
number, high heating value and good results on engine tests in terms of exhaust
emissions [39].

1.5 economical aspects of furfural industry

furfural supply and demand In the last decades, the global furfural pro-
duction capacity has been shrinking or stagnating in most of the countries, with
the only exception of China, which currently holds the leadership in furfural and
furfuryl alcohol production. As already mentioned, furfural current production
generally suffers from low yields, and significant economical and environmen-
tal concerns mainly due to its limited technological evolution, furthermore it is
characterized by a high number of relatively small producers. As a result, cur-
rent furfural supply is unstable and volatile, with exchange prices topping 2500
$/ton in the last years mainly because of supply shortages [90]. Yet, investing in
new furfural capacity using current technologies is considered a risky enterprise
by most operators.
The demand for furfural comes from established businesses, and more re-
cently from few emerging applications such as wood modification, and THF
synthesis in China. As mentioned above, there is a multitude of potential down-
stream applications, but the expansion of furfural demand has to face a stag-
nating supply. In a scenario of increasing furfural availability experts envision a
substantial growth of furfural demand up to 1 Mton/y by 2020, only considering
the expansion of traditional businesses [90]. In addition, the global trend toward
bio-based products, and rising cost of oil derivatives, is stimulating the demand
for furfural for new applications. As shown in Figure 1.7, the cost of production
of the main oil derivatives in the last years has been in the range of 1 k$/ton,
making furfural an appealing candidate as platform chemical when available at
lower and stable prices. The challenge is now in the hands of innovative furfural
producers willing to implement novel and disruptive technologies.
Concerning the potential of furfural derivatives as transportation fuels, the
competition with oil derivatives is more severe considering the lower energy
density of furfural. When comparing the production cost of current fuels on $/GJ
basis, as shown in Figure 1.8, furfural would be a competitive alternative in a cost
Figure 1.7: Oil exchange price (CIF WTI), and production cost of gasoline and diesel expressed in US$/ton, source [6].
1.5 economical aspects of furfural industry
25
26 introduction

range of 500 $/ton, whereas its fully hydrogenated derivative THFA around 600
$/ton. Current market prices for furfural and derivatives are significantly higher,
although they could reach those cost levels when produced in a considerably
larger scale, and adopting innovative technologies, as discussed in Chapter 7.

1.6 motivation and scope

Despite the relative simplicity of current production processes, there are still
some questions that must be addressed concerning the chemistry of C5 sugars
which might lead to significant improvements in furfural production. Although
it is well known that C5 sugars like xylose can be converted quantitatively into
furfural under acidic conditions [7, 120–122], there are many aspects which de-
serve further investigation, especially concerning the reaction mechanisms and
kinetics of both furfural formation and destruction under the same conditions
of pH and temperature.
The objective of this research is to address some unsolved issues regarding the
chemistry and technology of furfural, which may be summarized as follows:

1. Better understanding of the chemistry of furfural formation, in particular

addressing the contradicting theories reported by different authors in the
scientific literature on the mechanistic aspects of xylose dehydration.

2. To seek in the chemistry of furfural formation the answer to the so-called

“paradox of furfural yields” pointed out by Zeitsch [7], concerning the
discrepancy in the furfual yields from pentoses obtained in industrial pro-
cesses (50-60%), in laboratory tests (70%), and in the analytical chemistry
(close to 100%).

3. The development of optimal xylose production conditions from raw bio-

mass in view of the development an integrated biorefinery based on fur-
fural production.

4. Finally, developing an innovative furfural production process in order to fit

the economical and environmental requirements of the modern biorefinery
industry.

1.7 outline

This dissertation is divided in 8 chapters and is organized in the following way:

In Chapter 2 the experimental methods used in this work are carefully de-
scribed. A new dedicated test rig is described in all the relevant aspects typical
of chemical reactor engineering, and the analytical and experimental methods
employed in the several experimental campaigns are also extensively described.
Chapter 3 concerns the reaction kinetics of furfural formation. The dependence
Figure 1.8: Oil exchange price (CIF WTI), and production cost of gasoline, diesel and bioethanol in US$/GJ (LHV basis), source [1, 6].
1.7 outline
27
28 introduction

of furfural formation reactions on the acid nature and concentration is discussed,

and a suitable reaction kinetic model is introduced and described.
In Chapter 4 some particular aspects of the chemistry of xylose reaction into
furfural are addressed with the aim to clarify the reaction mechanism and to
define new green catalytic pathways for its production. In particular the effect of
chloride salts in dilute acidic solutions at temperatures between 170 and 200 °C
is described, obtaining a better understanding of furfural formation mechanism.
Starting from the results discussed in chapter 4, the general effect of different
halides is addressed in Chapter 5, showing interesting synergistic effects, and
revealing more about the furfural mechanism of formation.
In Chapter 6 the combined production of hemicellulose-derived carbohydrates
and an upgraded solid residue from wheat straw using a dilute-acid pretreat-
ment at mild temperature is described. The solid residues of pretreatment are
also characterized showing a preserved crystallinity of the cellulose, and up-
graded characteristics for thermal conversion applications.
Chapter 7 deals with the industrial processes for the production of furfural,
describing in particular an innovative process patented by Delft University of
Technology based on the results contained in this dissertation. The innovative
process concept described in this chapter is aimed at an economically viable
and environmentally sound furfural production from biomass hydrolisates, with
reduced energy and chemicals consumption.
Finally in Chapter 8 an overview of the main conclusions is presented, and
some recommendation are provided for future research.
 2
E X P E R I M E N TA L M E T H O D S

In order to investigate several aspects related to the furfural formation and related reac-
tions, a new lab-scale titanium reactor has been designed and built in order to enable
liquid phase reactions under a relatively broad range of pressure, temperature and pH
conditions. Such test rig has allowed most of the experimental work behind this disser-
tation, and it is thoroughly described in this chapter in all the relevant aspects typical
of chemical reactor engineering. The analytical and experimental methods employed in
the several experimental campaigns described in this dissertation are also extensively
described in this chapter.

The contents of this chapter have been adapted from:

Marcotullio, G.; Cardoso, M. A. T.; De Jong, W. and Verkooijen, Ad H.M. (2009) Bioenergy II:
Furfural Destruction Kinetics during Sulphuric Acid-Catalyzed Production from Biomass,
International Journal of Chemical Reactor Engineering: Vol. 7: A67

29
30 experimental methods

2.1 materials

Xylose reagent grade (Sigma-Aldrich, ≥99%) and Furfural reagent grade (Sigma-
Aldrich, 99%) were used as model compounds in the experiments, and for HPLC
calibration. Furfural was further purified via vacuum distillation. Concentrated
HCl (36.5 – 38%wt), H2 SO4 (95-98%wt), and acetic acid (99-100%wt) were pur-
chased from J.T.Baker. Pure FeCl3 hexahydrate, NaCl, CaCl2 dihydrate, KCl, KBr,
KI, KHSO4 and CaCO3 were purchased from Merck.
For the experiments discussed in chapters 3, 4 and 5, acidic aqueous solutions
of xylose (or furfural) were prepared with demi-water, then added with known
amounts of inorganic salts, and fed to the tube reactor after stirring to ensure a
complete dissolution of the solids.

2.1.1 Analysis of the reaction products

Analysis of the reaction products was carried out by means of an HPLC ap-
paratus equipped with a Resex ROA-Organic acid column, 8% cross linked
H+ , 300×7.80 mm, or alternatively with a Resex RHM-Monosaccharide column,
8% cross linked H+ , 300×7.80 mm (Phenomenex Inc., Torrance, CA, USA). A
Marathon XT auto-sampler (Separations, Ambacht, NL) was used to enhance re-
producibility. Quantification of the components was carried out by means of both
Refractive Index detector (Varian Model 350) and UV detector (Varian Model 310
Pro Star) in series.
The mobile phase consisted of a 0.005 N H2 SO4 solution in demineralized
water. The HPLC was operated at a flow rate of 0.6 ml/min, at a column temper-
ature of 80 ◦ C.

2.2 experimental setup

The experimental setup consists of a coiled tube reactor immersed in a thermo-

static oil bath. The reactor is made of titanium grade 2 (Merinox, Alblasserdam,
NL) to ensure high resistance to corrosion and minimal catalytic effects. The tube
reactor has an external diameter of 3.2 mm, internal diameter of 1.7 mm, and a
length of 4.40 m. The reactants are fed to the reactor by a Waters HPLC pump,
while the reaction temperature is precisely controlled and kept constant by the
oil bath. The reaction solution is rapidly heated up to the set temperature by an
electrical heater consisting of an insulated aluminum block which encloses the
tube reactor for a length of approximately 15 cm. Downstream the reactor the
solution is cooled down by means of a double pipe heat exchanger fed with tap
water, see figure 2.1.
In the range of process conditions considered, the flow in the tube reactor is
within the laminar region as the maximum Reynolds number achieved in this
 2.2 experimental setup 31

Figure 2.1: Representation of the experimental setup. The refractive index detector, injec-
tor and the computer were only used for the testing purposes described in
this chapter, but not during the experimental campaigns.

work is around 320. This implies the need of a careful evaluation of the residence
time distribution (RTD) along the reactor. To accurately determine the RTD, tests
were carried out injecting a small volume of a tracer solution (50%vol formic
acid in water) in the proximity of the reactor inlet through an in-line injector,
see figure 2.1, and detecting its concentration distribution in time at the reactor
outlet by means of a refractive index detector. This procedure was repeated for
various conditions of flow rate and temperature and the detector and the injector
were connected to a terminal for data collection and processing.
32 experimental methods

Figure 2.2: Dimensionless E(ϑ ) for different flow-rates, at T = 180o C

In figure 2.2 some experimental RTD curves E(ϑ ) are depicted. All the curves
are normalized, ϑ being a dimensionless time scale:
� ∞ � ∞
t
ϑ= where MO1 = tE(t)dt and E(t)dt = 1
MO1 0 0

The quantity MO1 is the first momentum of the normalized RTD curve E(t),
which represents the average residence time of a molecule in the reactor. Nor-
malized variance σϑ can be easily calculated by integration of the RTD curves. It
is relevant to estimate Péclet (Pe) number for all curves using equation (2.1):

2Pe − 2 + 2e− Pe 2
σϑ2 = ⇒ Pe � f or Pe > 100 (2.1)
Pe2 σϑ2

as the Pe number gives an indication of how axial dispersion counteracts the

spreading effect due to the laminar velocity profile. In the cases investigated
Pe varies from 164 to 775, while MO1 varies from 2.76 to 37.21 min. A higher
Pe number means a behavior closer to an ideal plug-flow reactor, as shown in
figure 2.2 where RTD curves present an asymmetrical shape which approaches
a sharper Gaussian distribution when the residence time becomes longer and Pe
higher, [123].
Considering a first order reaction, and knowing E(t), real conversion rates can
be calculated by applying equation (2.2):
� ∞
C
= e−kt E(t)dt (2.2)
C0 0
 2.2 experimental setup 33

Figure 2.3: Theoretical and real conversion rates when k = 0.1[min−1 ]

These values are shown to very closely approach the theoretical conversion
∗
rates calculated as (C/C0 ) = e−kt when t∗ = MO1 as shown in figure 2.3 for
a hypothetical case. Thus, when MO1 is taken into account as the average resi-
dence time in the reactor, uncertainties due to the laminar flow spreading effect
can be neglected even at higher flow-rates and shorter residence times. In this
case the reactor can be considered as an ideal plug-flow reactor.

2.2.0.1 Main reactor characteristics validation

Since the entire length of the tube is not at reaction temperature and in order to
precisely assess the average time of the reactant under actual reaction conditions,
the length of the tube at those conditions needed to be estimated. This was done
by residence time measurements as described below.
The tube reactor was considered to be the sum of a two portions of tube in se-
ries at different temperatures, being the volume VH at reaction temperature and
VR at room temperature. Temperature transitories were neglected and instanta-
neous temperature increments were assumed to occur within the length of the
two heat exchangers. For simplicity VH and VR are assumed to be constant for all
conditions of temperature and flow rates. Thus the total volume between the in-
jector and the detector is Vtot = VH + VR ; in the same way the average residence
time in the reactor can be considered to be the sum of the average residence time
in the high temperature portion of the reactor, which is the one of interest for
the reaction, and the average residence time in the rest of the volume, being then
τtot = τH + τR . Considering ṁ to be the mass flow rate and ρ the density of the
solution, by the mass conservation one derives:
34 experimental methods

13,50

13,00

12,50

Volume [ml]
12,00

11,50

11,00

10,50
880 900 920 940 960 980 1000 1020
Density [g/l]

Figure 2.4: Apparent reactor volume V ∗ against solution density ρ H . (*) Averaged mea-
sured values at constant T; (· · · ) measurement dispersion; (-) model used in
this work

ρ R VR ρ H VH
τR = ; τH = (2.3)
ṁ ṁ
which results in:
τtot ṁ = Vtot ρ R + VH (ρ H − ρ R ) (2.4)
When the reactor is at room temperature, for every ṁ:

Vtot
ṁ = 0
ρR (2.5)
τtot
0 is the average residence time in the reactor when the reaction temper-
where τtot
ature equals the room temperature. τtot 0 can be measured directly by detecting

the tracer solution at the outlet of the reactor. Combining equations (2.4) and
(2.5) results in the following equation:
τtot ρ
V∗ = 0
Vtot = VR + VH H (2.6)
τtot ρR

The volume V ∗ in equation (2.6) represents the total apparent volume of the
reactor which decreases as the solution density ρ H decreases with increasing
reaction temperature. Experimental assessment of V ∗ variation with ρ H allows
the estimation of VH and VR . In this way the average residence time τH can be
calculated by equation (2.3) for every ṁ and temperature. Operating with very
diluted solutions, density ρ H was assumed to closely approach that of water
and it was computed, given pressure and temperature, using the highly accu-
rate thermodynamic model for water included in the RefProp package by NIST,
[124], where the FluidProp package was used as interface, [125]. Experimental
results are shown in figure 2.4 where V ∗ is plotted against ρ H . The solid line in
 2.3 methods used for wheat straw pretreatment 35

the figure 2.4 was drawn given VH , which was derived by estimating the temper-
ature profile along the tube, and deriving VR by least square fitting. As shown
in figure 2.4 the average error introduced by using the calculated reaction resi-
dence time is always less than 3%, and can be mostly regarded as the result of
the pump flow rate variation.

2.2.0.2 Operations
After achieving the desired temperature in the oil bath and pre-heater, the flow-
rate ṁ was set in order to get the desired τH . The operating pressure was kept
always around 60 bar by means of a back-pressure regulator, which is higher
than the saturated pressure of water in the temperature range considered to en-
sure entirely liquid phase operation in the reactor. For each set of temperature
and flow rate ṁ conditions the system was run for a time longer than the cor-
responding ϑ = 1 (see figure 2.2) to ensure that steady state was achieved and
then a sample was recovered in a vial at the reactor outlet and analyzed. The
operation was carried out for every desired τH , and every set of temperature
and initial pH.

2.3 methods used for wheat straw pretreatment

2.3.1 Materials

The wheat straw used in this work came form Spain, and it was made avail-
able within the FP6 European project Biosynergy. The straw was ground, passed
through a 1.4 mm sieve and stored in a Polyethylene container at room temper-
ature. Pure monomeric sugars (99%), D-xylose, L(+)arabinose, and D(+)glucose
used as reference materials were purchased from Sigma-Aldrich. Concentrated
HCl (36.5 – 38%wt), H2 SO4 (95-98%wt), and acetic acid (99-100%wt) were pur-
chased from J.T.Baker. Pure FeCl3 hexahydrate, NaCl, CaCl2 dihydrate, and
CaCO3 were purchased from Merck.

2.3.2 Wheat straw pretreatment procedures

Borosilicate glass bottles of 50 ml internal volume with membrane screw cap

(Duran) were used in this work to carry out the reactions. Wheat straw samples
of 2g (a.r.) were mixed with 10ml of solution directly inside the glass reactor.
The solutions employed in this work were pure water and dilute solutions of
HCl or FeCl3 (100-200 mM in pure water). After the preparation of the sample
the bottle was sealed and placed in a ventilated oven, which was then set to the
desired temperature (100 - 120◦ C). For all the pretreatments the reaction time
was 120 min, and it was counted starting when the oven reached the aimed tem-
perature. The actual temperature inside the glass bottles could not be measured.
36 experimental methods

The reactions were carried out in static conditions as no stirring was employed.
After the desired residence time, the reaction was ended by quenching the bot-
tle in a water bath at room temperature. The contents of the bottle were thus
collected using a spatula into a vacuum filtering device. The bottle was repeat-
edly cleaned from the remaining traces of solids using ultra-pure water, which
was thus poured onto the filter cake in order to ensure a better washing of the
solids. The liquid filtrate was collected and made up to 50 ml with ultra-pure
water using an appropriate volumetric flask and, after pH measurement, it was
stored in a refrigerator for further analysis. The solid residue was dried at room
temperature and weighed after 24 hours.
All pretreatments were performed in duplicates and averaged data are pre-
sented. Raw measurement data such as carbohydrates concentration in the fil-
trates, filtrates pH and sample weight loss showed limited deviations between
duplicates, the average deviation being in the range of ±8%.

2.3.3 Analytical methods

2.3.3.1 Liquid filtrates analysis

Monomeric sugars (MS) and acetic acid concentrations in the filtrates were mea-
sured by means of an HPLC apparatus equipped with a Resex RHM - Monosac-
charide column, 8% cross linked H+ , 300×7.80 mm, (Phenomenex Inc., Torrance,
CA, USA). A Marathon XT auto-sampler (Separations, Ambacht, NL) was used
to enhance reproducibility. All components were quantified by means of a Re-
fractive Index detector (Varian Model 350). A 0.005 N H2 SO4 solution was used
as the eluent at a flow rate of 0.6 ml/min with a column temperature of 80◦ C
and a run time of 25 min. No appreciable amounts of furfurals were found in
the filtrates.
In order to quantify the soluble sugars in oligomeric form (OS), 10 ml of liq-
uid filtrate was added with 0.4g HCl solution (36.5 – 38wt%) and underwent a
second hydrolysis step at 120 °C for one hour. Sugar standard samples of known
concentration were also subjected to the same treatment at the same conditions
to quantify the sugar loss (SL) during such step (always in the order of few
percent). The hydrolysate was added with CaCO3 until neutral pH and then
analyzed. The total pentoses content measured after the second hydrolysis step
was corrected for the sugar loss considered as the fraction of sugars disappeared
after the second hydrolysis step on the initial. For this correction a conservative
hypothesis was taken considering only the monomeric fraction MS to be sub-
jected to loss reactions. The OS were thus calculated as the measured sugars in
the hydrolysate minus MS*(1-SL), whereas the total sugars were MS+OS. Such
method employed for the quantification of soluble oligomeric sugars did not
give any information on their degree of polymerization.
 2.3 methods used for wheat straw pretreatment 37

Figure 2.5: TG-FTIR set-up at TU Delft Process and Energy Laboratory

The yield of total sugars was calculated on the basis of the untreated wheat
straw composition. The composition analysis of the wheat straw used in this
work was made available within the framework of the European FP6 research
project Biosynergy and fully presented by Huijgen et al. [13]. The reported con-
tent of the main polysaccharides is: xylan 21.5, arabinan 2.1, glucan 34.6 wt%
d.b.. No biochemical composition analysis of the untreated wheat straw, nor of
the residues of pretreatments, was directly carried out in this work.

2.3.3.2 Thermo-gravimetric analysis of the solid residues

The untreated wheat straw and the solid residues of the pretreatments were
analyzed using a Thermo-Gravimetric setup SDT Q600 produced by TA Instru-
ments, USA, used in a configuration similar to what was reported in previous
works from our group [126], and shown in Figure 2.5 in its complete configura-
tion including online FTIR detector (not used in this work). The residues were
manually ground and placed in an alumina cup in amounts varying from 4 to
6 mg. The samples were kept at 50 °C for 45 min under a flow of 100 ml/min
of nitrogen. The temperature was then ramped up at a rate of 10 °C/min to 120
°C and kept constant for 15 min in order to thoroughly dry the sample. Mois-
ture was quantified as the weight loss at the end of this period. The temperature
was then increased again at 10 °C/min to 550 °C where the sample was kept
isothermal for 20 min, and it was finally combusted in air (100 ml/min) at this
temperature for 20 min. The residue after the isothermal period at 550 °C is re-
garded as char, the residue after combustion is regarded as “ash” whereas the
fixed carbon is quantified as the difference between the char and the ash.
38 experimental methods

2.3.3.3 X-ray analysis of the solid residues

X-ray diffraction (XRD) and X-ray fluorescence (XRF) were used to examine the
physical characteristic of untreated wheat straw, as well as pretreated samples.
X-ray diffraction analysis of untreated and pretreated solid samples can give
information regarding the crystallinity index of the cellulose in the wheat straw
samples, whereas XRF was used for elemental analysis of the solid residues.
XRD analysis was carried out using a Bruker D5005 diffractometer equipped
with Huber incident-beam monochromator and Braun PSD detector. The X-ray
powder diffraction patterns were recorded in a Bragg-Brentano geometry. The
samples of untreated wheat straw and the pretreated solid residues were scanned
from 2θ=5° to 60° with a step size of 0.038°. All samples were measured under
identical conditions. Cu Kα radiation source (α = 0.154056 nm) was used and
around 20 mg of sample was deposited on a Si<510> wafer and was rotated
during measurement. Data evaluation was done with the Bruker program EVA.
The crystallinity index (CrI) was defined as follows:

I002 − Iam
CrI = 100
I002

where I002 and Iam are the intensity of diffraction at 2θ=22.6° (crystalline region)
and at 2θ=16,2° (amorphous region) respectively, as already described in litera-
ture [127].
Semi-quantitative multi-element analysis of all inorganics (C, H, O, and N
are not measured) was carried out using a Philips PW2400 X-ray wavelength
dispersive Fluorescence Spectrometer. Data evaluation was done with Uniquant
(5.0) software. Samples of wheat straw solid residues of pretreatment, of about 1
gram each, were pressed into pellets of 27mm diameter and analyzed.
All the X-ray analysis were carried out at the Department of Materials Science
and Engineering of the Delft University of Technology.
 3
REACTION KINETICS IN FURFURAL PRODUCTION

Even considering the number of relevant works on the topic of furfural formation in
acidic media, a general expression for the reaction kinetics, its dependence on the acid
nature and concentration, and the potential effect of other species present in solution,
is yet to be established. Results of reaction kinetics studies related to furfural formation
from xylose, xylose side reactions, and furfural destruction in acidic aqueous media are
reported in this chapter.

The contents of this chapter have been adapted from:

Marcotullio, G.; Cardoso, M. A. T.; De Jong, W. and Verkooijen, Ad H.M. (2009) Bioenergy II:
Furfural Destruction Kinetics during Sulphuric Acid-Catalyzed Production from Biomass,
International Journal of Chemical Reactor Engineering: Vol. 7: A67
Marcotullio G., Heidweiller H., de Jong W., Reaction kinetics assessment for selective production of
furfural from C5 sugars contained in biomass, in proceedings of 16th European Biomass Conference and
Exhibition - From Research to Market, pp. 1-6, ETA, June 2008, Valencia - Spain.

39
40 reaction kinetics in furfural production

3.1 furfural formation and destruction in acidic conditions

Many relevant studies on the chemistry of furfural, its mechanism and kinetics
of formation and destruction in aqueous acidic solution, have been published
from the 1940s up to now [7, 50, 120–122, 128–137]. By reviewing the literature
available on this topic, a limited agreement emerges between different studies,
which are often carried out under different temperature conditions, using dif-
ferent catalysts in a wide concentration range, or using different raw materials
at different initial concentrations. In this dissertation it will be shown that the
chemistry of furfural formation has a complex dependence on catalyst type, and
secondary reactions may become relevant when increasing the initial concentra-
tions of reactants.
In this chapter experimental results are presented concerning furfural forma-
tion in aqueous acidic solutions from pure D-xylose, and furfural destruction
under similar conditions. A relatively narrow range of conditions is investigated,
which is considered relevant both from the analytical point of view, and for in-
dustrial application. In particular, reaction temperature is varied between 150
and 200 °C, and a dilute-acid media is employed in order to resemble industrial
operations. Sulfuric acid is chosen as model catalyst, at concentrations between
145.5 and 36.4 mM (pH between 0.81 and 1.36 at 298 K). Furthermore initial
reactants concentration is kept in the dilute range - the initial concentration of
furfural varied between 60 and 72 mM, and that of D-xylose varied between 7
and 133 mM - in order to limit second order reactions.
Different reaction kinetic mechanisms leading from xylose to furfural have
been proposed in literature, from more simple ones to more complicated [7, 8,
120–122, 133], but almost all of them can be expressed in a simplified way as
depicted in Figure 3.1. Even if simple such a scheme of reaction enables to accu-
rately fit the experimental results under the conditions of interest for this work,
thus it is chosen for describing the reaction. As the initial sugar concentration
is kept in the dilute range, second order reactions between furfural and inter-
mediates can be ruled out; the effect of increasing initial sugar concentration is
presently being studied and it will be discussed elsewhere. Such approximation
allows for describing the reaction rates by a relatively simple analytical expres-
sion: 
dCX
 r D− Xylose = dt = −(k1 + k2 )CX

r Intermediate = dCdt = k 1 CX − k 1b Ci
i (3.1)

 dCF
r Fur f ural = dt = k1b Ci − k3 CF
Reaction intermediate(s) have never been clearly identified in literature, never-
theless, judging from experimental results, their concentration can be assumed
to be low and not varying significantly with time. Hence steady-state approxima-
tion can be introduced for the intermediate resulting in dCi/dt = k1 CX − k1b Ci ∼
=0
 3.2 kinetics of furfural destruction 41

Figure 3.1: Simplified scheme for furfural formation from D-xylose

and then dCF/dt ∼= k1 CX − k3 CF . Introducing this assumption makes the integra-

tion easier and the concentration of xylose and furfural can be written as:
�
CX (t) = CX0 e�−(k1 +k2 )t � � � (3.2)
CF (t) = CX0 k −kk1 −k e−(k1 +k2 )t − e−k3 t
3 1 2

where the initial furfural concentration CF0 = 0 and CX0 is the initial xylose
concentration. The maximum molar yield of furfural can also be analytically
derived as a function of the three main kinetic parameters as:
 
� � k3
CFmax k 1  k 1 + k 2 k3 −k1 −k2 
= (3.3)
C X0 k1 + k2 k3

3.2 kinetics of furfural destruction

Furfural is a very reactive compound and is known to easily polymerize in

both acid and basic environment following various routes. Moreover, in aque-
ous acidic media, the furan heterocycle is reported to undergo ring opening re-
sulting in aliphatic open-chain products [98]. For these reasons, under the same
conditions of temperature and acidity employed for its production, furfural de-
grades at an appreciable rate, posing a major issue when designing industrial
production processes.
In order to better quantify the furfural rate of destruction under specific condi-
tions, specific experimental studies have been performed using pure furfural as
model compound in acidic aqueous solutions. Under the conditions considered
in this work the reaction rate of furfural destruction presents a first order behav-
ior with respect to furfural concentration, hence furfural-furfural polymerization
reactions leading to resins seem unlikely to take place, in agreement with earlier
observations [50, 120].
42 reaction kinetics in furfural production

The model proposed here postulates the protonation of the furfural molecule
and its subsequent reaction, resulting in the following simple first order rate
equation:
dCF
= − k 3 CF (3.4)
dt
The kinetic parameter k3 may be easily derived by least square fitting starting
from experimental measurements. As expected, the k3 show a clear dependence
on temperature as well as on H2 SO4 initial concentration. Nevertheless, a simple
correlation with initial H3 O+ aq concentration resulted inadequate to predict the
measured results.
Thermodynamics of electrolytes in liquid hot water need to be considered in
the first place. It is known from literature that the second dissociation constant of
sulphuric acid varies significantly with temperature and with the ionic strength
of the solution [138–140], and neglecting such effects might lead to significant un-
certainties, as observed by some authors [7, 133]. In particular, within the range
of conditions of interest for this work, the bisulfate ion becomes a very weak
acid (pKa2 � 4) [138], thus aqueous sulphuric acid can be considered with good
approximation monoprotic and only dissociated into H3 O+ and HSO4− . For this
reason the H3 O+ concentration at reaction conditions may be assumed to equal
the initial acid concentration, see Table 3.1. Nevertheless, when considering the
formulation k3 = k∗3 [ H3 O+ ], and an Arrhenius temperature correlation for k∗3 , the
model results inadequate when compared to experimental results. In particular,
at constant temperature, the reaction rate presents a less than linear dependence
on [H3 O+ ] (H3 O+ molar concentration), thus the correlation between k3 and
acid concentration has an higher complexity.
When studying acid catalyzed reactions of organic compounds, a common
approach is to make use of the acidity function H◦ proposed by Hammett in
1934 [141], which takes into account a complex correlations derived from the
equilibrium between a weak base and its protonated form in an acidic solutions.
On the other hand, the complicated Hammett acidity function for low acid con-
centration approaches the pH = − Log( a H3 O+ ) [142]. Considering the range of
conditions employed in this work in terms of temperature, pressure and ionic
strength, the general formulation for dilute solutions may be considered accept-
able. Based on these considerations, the ion activity a H3 O+ was used in the kinetic
formulation instead of the simple [H3 O+ ] , which possibly accounts for the lower
“availability” of the ions with increasing ionic strength as well as for the chang-
ing dielectric characteristics of the solvent induced by temperature [142, 143].
When using the formulation k3 = k∗3 a H3 O+ , the resulting rate constant k∗3 shows
a clear Arrhenius dependence on temperature in the range of conditions con-
sidered, as depicted in Figure 3.2. Thus a H3 O+ may be considered an adequate
measure of the acid catalytic activity on furfural destruction. Deviation of two
measurements at lower temperature can be explained by the amplification of
 3.2 kinetics of furfural destruction 43

Table 3.1: H3 O+ molar concentration [ H3 O+ ] and relative activity coefficient γ H3 O+ at

various temperature and initial acid concentration as estimated by the eNRTL
model
Initial H2 SO4 molar concentration
0.145 M 0.109 M 0.073 M 0.036 M
T /°C [H3 O+ ] (γ H3 O+ ) [H3 O+ ] (γ H3 O+ ) [H3 O+ ] (γ H3 O+ ) [H3 O+ ] (γ H3 O+ )
190 0.146 (0.566) 0.109 (0.607) 0.073 (0.660) 0.036 (0.739)
180 0.146 (0.573) 0.110 (0.614) 0.073 (0.667) 0.036 (0.744)
170 0.146 (0.578) 0.110 (0.620) 0.073 (0.673) 0.036 (0.750)
160 0.147 (0.583) 0.110 (0.625) 0.073 (0.678) 0.036 (0.754)

measurement errors in a range of conditions where furfural concentration decay

is more difficult to quantify. In this temperature range the reaction is very slow
and thus k3 values rather small (∼ 10−5 s−1 ).
Individual ion activity for the ionic species i may be defined as: ai = γi [Ci ],
where γi represents the individual activity coefficient and [Ci ] the molar concen-
tration. In general, from the Debye-Hückel theory, the ions activity coefficient
can be related to the ionic strength I in dilute solutions according to the equa-
tion:
log10 γi = − Az2i I 1/2
where z is the ion charge and A is a constant dependent on temperature, density
and dielectric constant of the solvent. Although simple, such correlation might
be inadequate in a wider range of conditions, and especially at ionic strength
higher than about 0.1 M. For this reason, some authors have made use of either
the Pitzer model [138, 144], or the electrolytes NRTL (eNRTL) model [145]. In this
work the latter model was chosen, where the commercial software Aspen PlusTM
was used for its implementation. Individual ion activity coefficients under actual
reaction conditions could thus be estimated, see Table 3.1.
The solid line in figure 3.2, was drawn by least square fitting of the experimen-
Ea
tal measurements and it refers to the Arrhenius formulation k∗3 = Aexp(− RT )
− 1
where ln(A)=26.64 [s ] and the activation energy Ea =125.1 [kJ/mol]. From these
kinetic parameters enthalpy and entropy of activation � could easily be derived, re-� ‡
sulting in ∆H ‡ =Ea -RT=121.1 [kJ/mol] and ∆S‡ = R ln k∗3 − ln T − 23.76 + ∆HRT
= -35.7 [J/molK] at 473K. The negative value for the activation entropy shows
a decrease in entropy across the transition state, which in this case could mean
the association of the furfural molecule with the proton forming a more ordered
intermediate [143].
As far as the activation energy is concerned, these results are in disagreement
with earlier studies, which reported much lower values for this reaction, namely
44 reaction kinetics in furfural production

Figure 3.2: k∗3 [s−1 ] Arrhenius plot in the temperature range 150-200 °C

20 kcal/mol (83.7 kJ/mol) [50]. Rather than to differences in reaction rate mea-
surements, such disagreement might be explain due to the different modeling
done by the authors. In particular, major discordances are generated when con-
sidering the non-ideality of the acid solution, i.e. ions activity, as well as the
sulphuric acid second dissociation constant variation with temperature. Never-
theless, despite the different modeling, good agreement is observed when com-
paring the experimental measurement reported in the same work [50] with those
predicted by using the model proposed here under the same conditions (0.1 N
H2 SO4 at 160 °C for 180 min). In this case the deviation in terms of predicted
and measured furfural concentration in time is always lower than 7%.
With respect to the furfural destruction reaction products, in the same work
[50] it was claimed that formic acid had been identified, which was suggested
to result from the hydrolytic fission of the furfural aldehyde group. In this work
formic acid could not be clearly identified and quantified, and by a qualitative
analysis of the HPLC-UV chromatograms, its presence, if cannot be excluded,
can be regarded as marginal under the tested conditions. Formic acid detected
during furfural production from pentosans might rather be considered as the
result of parallel sugar reactions [133], whereas a stronger evidence of rehydra-
tion of the furfural molecule to components like reductic acid has been reported
[130].
Preliminary subsequent studies using HCl as acid catalyst instead of H2 SO4 ,
have shown some inconsistencies namely in the rate of furfural destruction,
which appears to be faster than what could be expected by using the reaction
kinetics proposed in this work. Nevertheless a similar dependence on tempera-
ture was evidenced. Such results suggest a more complex mechanism involving
not only the H3 O+ ion but also the anions HSO4− and Cl− , which might play a
role in the catalysis of the reaction.
 3.3 kinetics of furfural formation from xylose 45

3.3 kinetics of furfural formation from xylose

Similarly to furfural, D-xylose reaction rate presents a first order reaction rate
with respect to its initial concentration under the conditions considered. Hence
the expected second order loss reactions, such as condensation reactions involv-
ing xylose, furfural and the intermediates of reaction, may be neglected in the
range of dilution considered in this work, and this is in agreement with earlier
observations [50, 121].
In order to derive the raw values of the main kinetic parameters k1 , k2 , k3 in
equations (3.2), a least-square minimization fitting to experimental results was
performed. The model showed very good agreement with the experimental mea-
surements in the whole range of conditions considered.
Remarkably, the three kinetic parameters showed a similar dependence on
acid concentration, thus, as evident from eq. 3.3, no significant improvements in
terms of maximum furfural yield may be achieved increasing acid concentration
within the range of interest for this work. This is an important observation when
considering the acid use in furfural production, evidencing that, in the dilute
range, acid concentration has a direct influence only on xylose rate of reaction,
but not on selectivity toward furfural. Different catalysts influence on furfural
selectivity will be discussed more in details in the following chapters.
Similarly to what shown in the previous section for furfural reaction, when
describing the acid influence on xylose reaction rate H3 O+ activity a H3 O+ is
preferably considered, resulting in the formulation:

k X = k1 + k2 = k∗X a H3 O+

Using such formulation for k∗X , similarly to k∗3 , shows a clear Arrhenius depen-
dence on temperature in the range of conditions considered.

133.3 [kJ/mol ] −1
ln(k∗X ) = 31.86 − [s ]
RT
125.1 [kJ/mol ] −1
ln(k∗3 ) = 26.64 − [s ]
RT
The activation energy of both reactions is of similar magnitude, thus no signifi-
cant furfural yield increase may be achieved by varying the reaction temperature
in the range considered. When considering the transition state model for k∗X , the
calculated values of ∆H ‡ and ∆S‡ are respectively 129.4 [kJ/mol] and 7.8 [J/-
molK] at 473K. Remarkably, the positive value of ∆S‡ indicates the increased
degrees of freedom of the activated state, which is a strong indication of xy-
lose ring opening being involved in the transition state. Hence protonation, ring
opening and subsequent dehydration are seemingly the key steps in the reaction
46 reaction kinetics in furfural production

of xylose to furfural. More indications in this sense will be provided in chapter

4.
If defining a reliable formulation for xylose and furfural reaction rates (k∗X
and k∗3 ) is of major interest for reactor engineering purposes in the furfural in-
dustry, the analysis of k1 and k2 is important for the optimization of furfural
selectivity and yield. In particular, experimental results showed that xylose side-
reactions rate (k2 ) and xylose-to-furfural reaction rate (k1 ) are related to acidity
and temperature in a very similar manner. Consequently, the resulting selec-
tivity toward furfural, expressed as k1 /k2 , remains relatively unchanged across
the whole range of conditions considered - around 77% -, displaying an objec-
tionable limit to furfural yield from xylose. Similar results were reported earlier
[122]. Such “selectivity limit” is in apparent disagreement with the commonly
reported quantitative furfural yields from pentoses when performing the classic
standard method for biomass pentosans-content estimation [7, 8, 12]. Such dis-
agreement suggests that different reaction mechanisms might take place under
different conditions, indicating that xylose dehydration to furfural goes beyond
the commonly accepted specific-acid catalysis. This observation stimulated fur-
ther experimental work, which led to original results regarding furfural forma-
tion mechanism from xylose, selectivity and overall yield, that will be exhaus-
tively discussed in the following chapters.
Regarding the influence of initial xylose concentration, it is important to re-
mind that higher sugar concentrations may lead to a significant decrease in over-
all furfural yield due to the occurrence of second order loss reactions. Such effect
is known from past studies, but, although some reaction mechanisms have been
proposed [7, 31], the products of such reactions have never been isolated. The in-
fluence of second order reactions have been quantified only in terms of furfural
yield losses compared to the dilute conditions. In this study some tests have
been performed in this sense, confirming earlier observations [122]. In particular
it was shown that an increase in initial xylose concentration from 0.1 to 1.5 wt%
led to a maximum furfural yield drop from about 70% to 58%. Furthermore, an-
alyzing the experimental results it emerged that the lower furfural yield was the
result of an increased apparent furfural destruction rate, i.e. an increased appar-
ent k3 compared to the actual parameter estimated by using furfural as model
compound in the absence of other species (sugars or derivatives). Such devi-
ation appeared only when increasing initial xylose concentration. Considering
also that xylose reaction rate remained virtually unchanged during those tests,
and that under similar conditions furfural showed a first order reaction rate, the
increased furfural rate of disappearance might be explained by its reaction with
other compounds than xylose or furfural itself, possibly involving intermediates
of reaction or side products.
 3.4 conclusions 47

3.4 conclusions

The results reported in this chapter regard xylose reaction and furfural destruc-
tion reaction kinetics in the temperature range 150 – 200 °C and H2 SO4 concen-
tration between 36 and 145 mM. Initial reactants concentration was kept in a
relatively narrow range of dilution in order to observe the rate of furfural and
xylose reaction in dilute acid solutions in the absence of second order reactions
effects, mostly occurring at higher reactants concentrations.
The kinetic models proposed, although simple, present a good agreement with
the measured reaction rates. The kinetic constants are related to the acid concen-
tration via the ion activity a H3 O+ to better account for the thermodynamics of
the electrolytes in solution, which may be accurately estimated by means of the
eNRTL model. Moreover, no significant formation of formic acid from furfural
from furfural decay was observed, contrary to earlier studies.
The influence of temperature on the rate constants k∗X and k∗3 showed a clear
agreement with the Arrhenius law, the activation energies being respectively
133.3 and 125.1 kJ/mol and the pre-exponential factors 68.65·1012 and 3.71·1011
s−1 . From the transition state theory a strong indication emerges of ring opening
being involved in the activated state of xylose reaction to furfural.
 4
C H L O R I D E I O N S E N H A N C E F U R F U R A L F O R M AT I O N F R O M
XYLOSE IN DILUTE AQUEOUS ACIDIC SOLUTIONS

In this chapter some particular aspects of the chemistry of D-xylose reaction into furfural
are addressed with the aim to clarify the reaction mechanism and to define new green cat-
alytic pathways for its production. Specifically the reduction of mineral acids utilization
is addressed by the introduction of alternative catalysts. In this sense chloride salts were
tested in dilute acidic solutions at temperatures between 170 and 200 ºC. Results indi-
cate the Cl− ions to promote the formation of the 1,2-enediol from the acyclic form of
xylose, and thus the subsequent acid catalyzed dehydration to furfural. For this reason
the presence of Cl− ions led to significant improvements with respect to the H2 SO4 base
case. The addition of NaCl to a 50mM HCl aqueous solution (0.18 wt%) allows to at-
tain 90% selectivity to furfural. Among the salts tested FeCl3 shows very interesting
preliminary results, producing exceptionally high xylose reaction rates.

The contents of this chapter have been adapted from:

Marcotullio, G. and de Jong, W. Chloride ions enhance furfural formation from D-xylose in dilute
aqueous acidic solutions, Green Chem., 2010, 12, 1739-1746

49
50 the effect of aqueous chlorides on furfural formation

4.1 introduction

Novel and green concepts for furfural production are required in order to mini-
mize the carbon footprint and waste streams related to furfural industry. To this
aim a deeper comprehension of the chemistry of pentoses is required in the first
place, as well as the establishment of new catalytic pathways for their conversion
into furfural.
In view of the understanding of the mechanisms behind the quantitative con-
version of pentoses into furfural, it is interesting to recall the old standard meth-
ods for the estimation of pentosans content of plant material [7, 8, 12], where
nearly 100% of the initial pentosans can be recovered as furfural by distilling
atmospherically the raw material in a 12 wt% HCl solution saturated with NaCl.
Evidently such reaction conditions are just prohibitive for any industrial scale
application due to the high chemicals utilization. Nevertheless a deeper under-
standing of the reaction mechanism taking place under those conditions is neces-
sary in order to approach the same yields under significantly milder conditions.
The acid choice was not a case, since it is known from literature that differ-
ent acids catalyze differently the dehydration of pentoses into furfural [8], and
among strong acids HCl is reported to be the preferable choice [7, 8]. Such ob-
servation suggests that the mechanism of furfural formation might go beyond
specific acid catalysis, i.e. involving ionic species other than H3 O+ . Also very
important is the role played by NaCl which, besides the salting-out effect en-
hancing the separation of furfural during distillation, was reported to increase
sugars reaction rate. Such effect was thought to be the combined result of the
higher boiling temperature and of the increased activity coefficient of the acid
[8, 146].
More recently it was shown that the addition of KCl, NaCl, CaCl2 , MgCl2 or
FeCl3 to pure water increased the reaction rate of both xylose and xylotriose at
180ºC. This effect was more pronounced for xylotriose than for xylose [136]. Sim-
ilar results were reported in recent works when treating corn stover with diluted
inorganic salts [147, 148]. A peculiar catalytic effect of FeCl3 on hemicellulose
hydrolysis was also evidenced in these works, and FeCl3 in water was shown to
be significantly more effective than a strong acid solution of the same pH. In par-
ticular it was shown that Fe3+ and Cl− ions are both responsible for this effect
[148]. Gravitis et al. [149] reported the metal cations to catalyze the reaction of
biomass derived carbohydrates to furfural proportionally to their ionization po-
tential, mentioning an increasing effectiveness for K+ , Na+ , Ca2+ , Mg2+ , Fe3+ .
Furthermore metal salts, and especially metal chlorides, have been shown to
have significant effects in microwave aided hydrolysis of chitosan [150], the hy-
drolysis of ethers and amines in near and supercritical water [151], but also on
hemicellulose and cellulose hydrolysis both in solid state [152, 153], and in dilute
aqueous solution [136, 147, 148]. Some metal chlorides have also been shown to
affect the dehydration of glucose and fructose to HMF in ionic liquids [154].
 4.2 experimental results and discussion 51

#!!"
!,,$!!5-,.,'!/0,1-2,.,#)!!/0,34-2
8,,$!!5-,.,'!/0,1-2,.,,*'!/0,34-2
+!"
$,,$!!5-,.,'!/0,1-2,.,,(!!/0,34-2

*!"

)!"

(!"

'!"
!,,$!!°-,.,'!/0,1-2,.,%&!/0,34-2
&!" ",,$!!5-,.,'!/0,1-2,
#,,$!!5-,.,'!/0,1$67&

%!"

$!"

#!"

!"
! $!! &!! (!! *!! #!!!
Residence time /sec

Figure 4.1: Effect of NaCl addition on xylose reaction rate and furfural yield. Xylose (dot-
ted line) and furfural dimensionless concentration (solid line) as from equation
(3.2) after fitting to the experimental results.

In view of what mentioned above, in this work the use of different chloride
salts for furfural production is addressed, evaluating the role of the different
electrolytes in solution in terms of xylose rate of reaction and furfural selectivity
and yield, with the aim of exploring new catalytic pathways leading to high
furfural yields with a reduced mineral acid usage.

4.2 experimental results and discussion

Figure (4.1) shows the results of xylose dehydration in dilute aqueous acid solu-
tions with the addition of different amounts of NaCl. Initial xylose concentration
is constant for all the experiments in this work, and equal to CX0 = 33.3 mM. The
reaction rate of xylose in two equimolar solutions of HCl and H2 SO4 at 200 ºC
does not show a significant deviation, nevertheless furfural yield is significantly
higher in the former case with respect to the latter. The second dissociation con-
stant of H2 SO4 is known to decrease markedly with temperature [138, 139], to
such an extent that H2 SO4 can be considered with very good approximation to
be monoprotic at 200 ºC (pKa2 � 4), and consequently the H+ concentration of
two equimolar solutions of HCl and H2 SO4 can be assumed to be equal at those
conditions. The different results in terms of furfural yield can thus be ascribed
52 the effect of aqueous chlorides on furfural formation

#!!"
!,#+!,°-,.,#'!,/0,1-2
+!" ",#*!,3-,.,#'!,/0,1-2
#,#)!,3-,.,#'!,/0,

*!"

)!"

(!"

'!"

&!"

%!"

$!"

#!"

!"
! $!! &!! (!! *!! #!!!
Residence time /sec

Figure 4.2: Effect of temperature on xylose reaction rate and furfural yield in 150mM
HCl. Xylose (dotted line) and furfural dimensionless concentration (solid line)
as from equation (3.2) after fitting to the experimental results.

to the presence of different anions in solution, in particular Cl− having a more

beneficial effect than HSO4− .
In the presence of increasing amounts of NaCl in dilute aqueous HCl a clear
increase in xylose reaction rate is immediately evident and surprisingly higher
furfural yields are achieved, see Figure (4.1). Thus NaCl is confirmed [146] to
affect xylose reaction rate already at relatively low concentrations in acidic so-
lutions, although, contrary to what was believed before [8, 146], this effect can
neither be ascribed to the increased activity of the acid, nor to the increased
boiling temperature or the salting-out effect. From the Debye-Hückel theory the
ions activity coefficient γ can be related to the ionic strength I in dilute solutions
according to the equation log10 γi = − A z2i I 1/2 , in which the expression z is the
ion charge and A a constant dependent on temperature, density and dielectric
constant of the solvent [143].
Thus, under these conditions of dilution, increasing the salt concentration
leads to a decrease of the ions activities, and consequently to a decreased acid ac-
tivity. Furthermore the reaction is carried out at constant and controlled temper-
ature in liquid solution without any separation by steam stripping, thus boiling
point raise or salting-out effect do not play any role in this case.
 4.2 experimental results and discussion 53

In the previous chapter, where the action of H2 SO4 catalyst was studied in the
same range of conditions used in this work, increasing the temperature led to
negligible variations in terms of furfural selectivity and yield. Contrary to this
result, in Figure (4.2) it is shown how a significant increase of furfural yield can
be obtained raising the temperature from 170 to 190 ºC when dilute HCl is used
instead of H2 SO4 .
Considering these results, it can be postulated that the ionic species in solution
other than H+ induce a change in the reaction mechanism, and Cl− ions seem
to be the main responsible for this effect.
Different salts, namely KCl, NaCl, CaCl2 and FeCl3 , were also tested in 50
mM aqueous HCl, keeping a constant Cl− molarity in every solution. Results
showed very similar reaction behavior for the three salts KCl, NaCl and CaCl2 ,
both in terms of xylose reaction rate and furfural yield. For these three salts the
concentration of Cl− in solution seems to be the main feature for the kinetics of
xylose reaction, the different cations seemingly playing only a minor role, Figure
(4.3).
A significantly different behavior was observed when adding FeCl3 , where the
xylose reacted so fast that no sugar could be detected already after a residence
time of 124 seconds, thus, even if the exact value could not be derived, the
reaction rate appeared to be more than 10 times larger than in the other cases.
Hence the Fe3+ ion itself can be considered to have a relevant catalytic effect on
the reaction of xylose. On the other hand, furfural maximum yield was lower
when adding FeCl3 compared to the other cases, which accounted for 62% of
the initial xylose, in agreement with what was shown earlier when treating corn
stover under very similar conditions [148].
The addition of electrolytes to the water can generally induce significant changes
by attracting or orienting its molecules, or change its ordered short range struc-
ture and in general its interaction with the solute. Similar effects are responsible
for the vapor pressure increase of aqueous electrolytes solutions, for salting-
out or salting-in effects [155], or colloids precipitation as first studied by Franz
Hofmeister more than a hundred years ago [156]. In the same way these effects
could be regarded as responsible for the reported results in terms of xylose re-
action rates and furfural yield. Nevertheless, especially when considering dilute
solutions, Hofmeister effects are normally not so dependent on the nature of the
electrolytes but rather on the total ion concentration, whereas, by carefully con-
sidering the results shown in Figures (4.1), (4.2) and (4.3), it becomes evident the
the nature of the ions is very important even in very dilute solutions. Further-
more when comparing the results for CaCl2 with those for KCl and NaCl we can
notice that, even if the total ions concentration in solution is 25% lower in the
first case if compared to the other two, the results are very similar to each other.
Thus a direct contribution of the ions to the chemistry of the reaction, rather
than a change of the solvent-reactant interactions, can be considered to play the
major role.
54 the effect of aqueous chlorides on furfural formation

#!!"
,-$!!./-0-'!12-3/4-0-##%12-56/4%
7-$!!./-0-'!12-3/4-0-#)!12-/8/4$
+!" !-$!!°/-0-'!12-3/4-0-%&!12-98/4
"-$!!./-0-'!12-3/4-0-%&!12-:;<
*!" #-$!!./-0-'!12-3/4-0-%&!12-:/4

)!"

(!"
$-$!!./-0-'!12-3/4-
'!"

&!"

%!"

$!"

#!"

!"
! $!! &!! (!! *!! #!!!
Residence time /sec

Figure 4.3: Effect of different chloride salts on xylose reaction rate and furfural yield.
Xylose (dotted line) and furfural dimensionless concentration (solid line) as
from equation (3.2) after fitting to the experimental results.

Xylose reaction in two equimolar solutions of KCl and KBr was also studied
at the same conditions of temperature and acidity, Figure (4.3). The reaction re-
sulted to be about 15% slower in KBr with respect to KCl, even though faster
than in the absence of any added salt, while the furfural maximum yield was
comparable in both cases, see also Table (4.1). This last result denotes the addi-
tion of metal halides in general to aqueous acidic solutions to potentially have
surprisingly beneficial effect in terms of xylose conversion into furfural, and fur-
ther results in this sense will be exposed in the next chapter where where the
effect of different ions of the series I, Br and Cl will be thoroughly discussed.

4.2.1 Mechanism of furfural formation from pentoses

Contradictory theories exist in literature concerning the mechanism of furfural

formation from pentoses. On the one hand many authors considered the reaction
to proceed via the acyclic form of the pentoses, through a first 1,2-enediol for-
mation and subsequent three dehydration steps yielding furfural [131, 135]. On
the other hand, other authors believed the reaction to take place starting from
the pyranose form of the pentoses [121], which recently was postulated to take
place by the action of H+ on a specific position of the pyranose ring, leading
 Table 4.1: Condensed results from kinetic experiments.
Entry Acid Acid Salt Salt Total Cl− Temp. k X /10−4 k1 /10−4 k2 /10−4 k3 /10−4 Selectivity± Max. Yield∗
/ mM wt% / mM / ºC [s−1 ] [s−1 ] [s−1 ] [s−1 ] [%] [%]
1 HCl 150 - - 150 170 18.5 13.68 4.83 1.5 73.9 59.2
2 HCl 150 - - 150 180 42.0 31.5 10.5 2.5 75.0 62.7
3 HCl 150 - - 150 190 95.0 77.9 17.1 5.3 82.0 69.1
4 H2 SO4 50 - - - 200 46.2 32.6 13.6 2.0 70.6 61.4
5 HCl 50 - - 50 200 46.7 37.3 9.4 2.2 79.9 68.7
6 HCl 50 KCl 2.5 390 200 67.3 54.9 12.4 1.7 81.6 74.1
7 HCl 50 NaCl 2.0 390 200 64.5 52.84 11.7 2.2 81.9 72.8
8 HCl 50 KBr 4.1 390• 200 57.0 46.7 10.3 2.2 82.0 72.0
9 HCl 50 CaCl2 2.5� 390 200 63.0 51.0 12.0 2.0 81.0 72.3
10 HCl 50 FeCl3 3.1∓ 390 200 > 600 > 400 > 200 8.0 65.7 62.0
11 HCl 50 NaCl 3.5 650 200 74.4 63.1 11.3 2.2 84.8 76.1
12 H2 SO4 50 NaCl 3.5 600 200 78.7 65.1 13.6 2.0 82.7 75.3
13 HCl 50 NaCl 3.5 390 210 175.0 144.4 30.6 4.0 82.5 75.5
14 HCl 50 NaCl 5.0 906 200 119.1 107.4 11.6 3.4 90.2 81.3
15 HCl 50 NaCl 10.0 1762 200 196.8 159.4 37.4 4.3 81.0 74.4
16 Formic acid 220 NaCl 10.0 1712 200 66.5 42.2 24.3 0.2 63.5 62.1
17 HCl 50 NaCl 5.0 906 170 11.7 9.1 2.6 0.6 77.7 65.7
18 HCl 50 NaCl 5.0 906 185 39.5 32.4 7.0 1.1 82.2 74.3
19 HCl 50 - - 50 170 9.6 7.1 2.4 0.4 74.5 65.6
20 HCl 50 - - 50 185 29.5 21.0 8.5 0.9 71.2 63.7
21 HCl 100 NaCl 5.0 956 200 184.1 161.4 22.7 5.1 87.7 79.2

X CaCl2 dihydrate. FeCl3 hexahydrate.

± Calculated as k1/k .∗ Maximum furfural yield calculated from (3.3). • Considered as [Cl− ] + [Br− ]. � ∓
4.2 experimental results and discussion
55
56 the effect of aqueous chlorides on furfural formation

to a first dehydration and structure rearrangement yielding the furanose ring

and subsequent dehydration to furfural [133, 137]. It has been demonstrated
that after reacting xylose-1-14 C in 12% aqueous HCl the “aldehydic” carbon of
the pentose at C-1 position was found nearly completely at the aldehyde group
of the final product, 2-furaldehyde-α-14 C [129]. Such result is perfectly aligned
with the former theory, whereas it cannot be explained according to the mech-
anism proposed in the latter. Furthermore, in analogy to the reaction of xylose
to furfural, an isomerization reaction of glucose to fructose has been shown to
take place in acidic conditions at high temperature, proving this to be the first
step in the reaction leading from glucose to HMF [132]. Besides, ketopentoses,
which present a significantly higher proportion of acyclic form in water solution
if compared to aldopentoses [157, 158], react much faster in water yielding fur-
fural proportionally to the acidity of the solution [135]. From all these evidences
provided by different authors throughout the years, the former theory appears
to be strongly supported, whereas the latter fails to comply with some of the
earlier works. For these reasons, in this work, the first theory is considered for
the interpretation of the results.
A strong indication was also provided of the 1,2-enediol intermediate to be
irreversibly formed from the aldopentoses in acidic conditions, showing on the
other hand a partial equilibration with the keto form [135]. Such indication is
in agreement with the general aldoses reluctance to isomerization in acidic con-
ditions [132], and thus, since enolization reactions are normally reversible, the
1,2-enediol formation from xylose in acidic conditions can be considered to be
the rate limiting step in the formation of furfural. In analogy with this, HMF is
known to be formed much faster from fructose than from glucose in acidic condi-
tions, and fructose has often been measured in the reaction from glucose to HMF,
indicating aldose isomerization to take place as intermediate step [132, 159].

4.3 effect of the cl − ion on furfural kinetics of formation

Based on the results obtained in this work, it seems reasonable to postulate that
one particular step among the multi-step reaction from xylose to furfural to be
catalyzed by both H + and Cl − , this particular step being the formation of the
1,2-enediol intermediate. Thus the mechanism depicted in Figure (4.4) can be
drawn, which appears in agreement with the results obtained in this work and
with what was observed in previous works. In such mechanism the presence of
Cl − favors the formation of the 1,2-enediol 2, which can equilibrate with both
the aldo 1 and keto 3 form of the sugar, and reacts to furfural 4 in presence of
an acid.
Accordingly, it is evident from the experimental results that, with increasing
chloride salts concentration, xylose reacts significantly faster to furfural, while
no significant effect is observed on the side reactions, resulting in a remarkable
 4.3 effect of chlorides addition on furfural kinetics 57

Figure 4.4: Simplified reaction scheme involving chlorides

selectivity improvement. When the salt concentration becomes higher than a cer-
tain threshold with respect to [H + ], the increasing selectivity trend to furfural is
inverted and a small drop is observed, Figure (4.1). This can be due to the larger
concentration of the intermediate(s) which are prone to undergo loss reactions,
including possible condensation reactions with furfural. In the simplified kinetic
model proposed here this effect results in an increase of k 2 , as it is observed for
10 wt% NaCl , see Table (4.1); entry 15. As mentioned before, 2 and 3 are more re-
active than 1, also in absence of acids [135, 159], yielding furfural proportionally
to the acidity. Accordingly, the faster xylose reaction and the poor furfural yields
obtained using a weak acid in presence of 10 wt% NaCl (Table (4.1); entry 16)
can be explained due to the relatively high pH in presence of a relatively large
concentration of chloride salts. In addition, comparing the results when using
different strong acids like HCl and H 2 SO 4 led to different results as discussed
before, Figure (4.1), whereas in the same conditions, when a larger concentration
of chloride salt is also present, results are very similar (Table (4.1); entries 11-12),
meaning the H + source does not significantly influence the reaction when a
larger concentration of Cl − is also present.
Hence, even if the reaction of xylose can be catalyzed by chloride salts by favor-
ing the 1,2-enediol formation, the presence of a strong acid is necessary in order
to favor the selectivity toward furfural. This conclusion is in agreement with
what was previously shown for ketopentoses, where higher acidity influenced
the selectivity to furfural not influencing so much the sugars rate of reaction
[135].
The mechanism in Figure (4.4) helps also explaining the Arrhenius plot de-
picted in Figure (4.5). In this plot the natural logarithm of the total xylose re-
action rate kinetic parameter k X = k1 + k2 divided by [H+ ] is plotted against
58 the effect of aqueous chlorides on furfural formation

!(#$
!**"$*+,*-./01")*+,*23./
–*($$*+,*-./01")*+,*23./
!(#"
%*"$*+,*-./0
**************)$$*+,*23./
!'#$
#***"$*+,*-./
$*("$*+,*-./
!'#"
ln (k X /[H+ ])
!&#$

!&#"
y = -16.244x + 32.903
4**5'*+,*-'67%
0**"$*+,*-'67%
!%#$ "**&)*+,*-'67%
8*8*8*9:;/<=*>(?"&@ y = -16.788x + 33.482

!%#"

y = -17.389x + 34.299
!"#$
'#$) '#(( '#() '#'( '#')
>'($A.@ >'$$A.@ >(?$A.@ >(1$A.@ >(5$A.@

1000/T /K -1

Figure 4.5: ln(k X/[ H + ]) plot against temperature.

the inverse of the absolute temperature, in order to show the rate of reaction of
xylose excluding the effect of acid concentration.
The bottom line in Figure (4.5) shows the H2 SO4 catalyzed reaction, i.e. the
well known H+ catalyzed route, where formation of 2 through reaction (I) is the
rate limiting step according to the scheme in Figure (4.4). The linear relation-
ship between k X and [H+ ] can clearly be observed, accordingly the points for
different H2 SO4 concentrations are reasonably aligned in the plot in Figure (4.5).
This results are also in very good agreement with the know kinetics provided in
previous works [8] for the same range of acid concentration and temperature.
When a relatively large Cl− concentration is present, k X/[ H + ] values are larger,
and accordingly they shift up in the Arrhenius plot. Such shift can be explained
with a switch in the rate limiting step, namely from reaction (I), which is by-
passed due to the Cl− catalyzed formation of 2, to the subsequent H+ catalyzed
dehydration reaction (III). The 50mM HCl curved line is particularly revealing
the transition: at the higher temperatures it tends to the H+ catalyzed mecha-
nism (I); lowering the temperature the Cl− catalyzed reaction (II) becomes faster,
having seemingly a lower activation energy, leading to a switch to reaction (III)
becoming therefore limiting step.
When reaction (III) is the rate limiting step, the kinetic parameter k X should
be directly proportional to [H+ ], as previously shown, and indirectly to [Cl− ]
through 2 formation. Contrary to this no linearity is evident between k X and
 4.4 catalytic requirements in furfural production 59

[Cl− ] or [H+ ], the data showing a marked less than proportional relationship.
It can be analytically demonstrated, and it is also qualitatively intuitive, that
such relationship indicates reaction (II) to be reversible, contrary to (I), being the
equilibrium between 1 and 2 only partly attained. This means that the rates of
reaction (II) and (III) are comparable, thus concentration of 2 varies accordingly.
Due to this, an increased acidity at fixed [Cl− ] would cause reaction (III) to be
faster, at the same time lowering to some extent the mean concentration of 2,
and resulting in a less than proportional observed increase of xylose reaction
rate. In the same way, an increase of [Cl− ] at fixed [H+ ] would tend to increase
the average concentration of 2, resulting again in only a less than proportional
observed increase of furfural rate of formation because of the combined effect
of reaction (III) and (II). In this way the 150 mM HCl straight line intermediate
position in the Arrhenius plot can also be explained.
For these conditions the competing side reactions also become more complex,
involving 1, 2 and possibly 3, being ketopentoses very reactive also at lower acid
concentrations as mentioned before.
The peculiar Cl− behavior in this reaction is particularly interesting by itself.
As it has been shown, there is good reason to believe Cl− ions to act as a cat-
alyst in the enolization reaction of xylose, which is normally catalyzed in basic
environment even at lower temperatures. Alternatively the Cl− ions could be
considered to favor the presence of the 1,2-enediol with respect to both the al-
dose and ketose by a stabilizing effect, thus favoring the subsequent dehydration
reactions starting from such form. Anyway, based on the results presented in this
work, the 1,2-enediol promoting effect induced by the Cl− remains a hypothesis
consistent with the reaction kinetics of xylose to furfural, although no ultimate
evidence is provided in support of this theory.
As far as the peculiar effect of FeCl3 is concerned, it shows a significantly
more vigorous action on xylose than H+ . Although, in a first instance, Fe3+ ions
in water were believed to have a direct influence on xylose reaction, as it will be
explained in chapter 6, such action may be ascribed to the indirect formation of
HCl as a consequence of iron oxides precipitation at higher temperatures.

4.4 catalytic requirements in furfural production

Contrary to the dehydration reactions yielding furfural from pentoses and HMF
from hexoses, which require strong acids, aldoses isomerization is normally fa-
vored by basic conditions [132, 159], and proved to proceed through the 1,2-
enediol formation [160]. Moreover, in the past, it was also observed that furfural
formation from pentoses was not a simple dehydration reaction, since only poor
furfural yields were obtained when xylose was reacted with dehydrating cat-
alysts like pure ZnCl2 or phosphorous pentoxide, whereas, when an aqueous
solution of ZnCl2 was employed, furfural yield approached that obtained using
60 the effect of aqueous chlorides on furfural formation

strong acids solutions [161]. Thus a discrepancy in terms of “catalytic require-

ments” becomes evident in the process of furfural formation from pentoses, pre-
senting, in turn, an enolization followed by three dehydration reactions, being
the two reactions normally favored by different catalysts.
Thus, when thinking at industrial furfural production, it should be borne in
mind that the high severity normally required in terms of acid concentration
and temperature is mainly needed to overcome the first enolization step, as this
is normally not favored by acid. Whereas an acid catalyst is required for the
following dehydration steps. Thus, in order to achieve better furfural yields un-
der milder conditions and in a greener manner, the combination of two different
catalysts, having in turn basic and acidic character, would be probably the key.
On a similar observation a recent remarkable work is based[67, 162] where the
combined use of acid and basic solid catalysts in a one-pot reaction allowed for
good HMF and furfural yields, although improvements are still needed.
In the present work it has been shown that such double catalytic effect can
seemingly be achieved by the addition of metal chlorides to aqueous acidic solu-
tions, allowing for high furfural yields and relatively high xylose reaction rates.
In terms of furfural yield maximization an optimum for H+ and Cl− con-
centrations exists, also in relation with the loss reactions, involving a relatively
large Cl− excess and a relatively low acid concentration. Noteworthy seawater
typical composition with the addition of a small amount of strong acid could be
a suitable solvent for furfural production.

4.5 conclusions

Chloride salts were shown to enhance the reaction of xylose to furfural in aque-
ous acidic solution at temperatures between 170 and 200 ºC. Experimental re-
sults indicated the Cl− ions to promote the formation of the 1,2-enediol from the
acyclic form of the aldose, which undergoes subsequent acid catalyzed dehydra-
tion reactions to furfural.
The effect caused by the chloride salts addition to aqueous acid solutions was
studied under various conditions. Significant improvements were observed with
respect to the H2 SO4 -based case, furfural yield and selectivity respectively in-
creasing by 18% and 28%, selectivity in particular attaining 90%. At the same
time a 4-fold xylose reaction rate increase was possible by 1.7 M NaCl addition
keeping the acid concentration as low as 50 mM (0.18 wt% HCl), and initial
sugar concentration 33.3 mM.
 5
THE USE OF DIFFERENT HALIDES IN DILUTE AQUEOUS
A C I D I C S O L U T I O N S A L L O W S F O R E X C E P T I O N A L LY H I G H
FURFURAL YIELDS

Starting from the results discussed in the previous chapter on the effects of Cl− ions
on furfural formation in aqueous acid solution, the general effect of different halides is
addressed. Experimental results show the halides to influence at least two distinct steps
in the reaction leading from D-xylose to furfural under acidic conditions, via different
mechanisms. The nucleophilicity of the halides appears to be critical for the dehydration,
but not for the initial enolization reaction. By combining different halides synergic effects
become evident resulting in very high selectivities and furfural yields.

The contents of this chapter have been adapted from:

Marcotullio, G. and de Jong, W. Furfural formation from D-xylose: the use of different halides in
dilute aqueous acidic solutions allows for exceptionally high yields, Carbohydr. Res., 2011, 346,
1291-1293

61
62 the effect of aqueous halides on furfural formation

5.1 introduction

The influence of halides on the dehydration reaction of sugars to furfurals has

been increasingly addressed, with particular attention regarding the selection
of different ionic liquids showing optimal characteristics for this reaction [154,
163, 164]. The simultaneous use of solid basic and acidic catalysts has also been
explored in order to foster the isomerization of glucose to fructose and the sub-
sequent dehydration to 5-hydroxymethyl furfural [162]. In the previous chapter
the peculiar influence of Cl− ions on the formation of furfural from D-xylose in
aqueous acidic solutions was shown, reporting very high yields and selectivities.
It has been postulated that the Cl− ions promote the formation of the 1,2-enediol
from the acyclic form of the aldose, favoring thus the subsequent acid-catalyzed
dehydration. Nonetheless, the experimental evidence is based on reaction rates
studies, and thus insufficient to unravel the mechanism involving Cl− in this
reaction.
In order to further clarify these aspects, a systematic study has been carried
out observing the variations in the kinetics of D-xylose reaction in acid solution
brought about by adding different potassium halides. In particular KCl, KBr and
KI were used, whereas KF was not included in the series due to the relevant dif-
ference of HF Brønsted acidity compared to the rest of the hydrogen halides.
Using KF in acidic solution would decrease considerably the acidity of the solu-
tion making it impossible to compare the effect of this salt with the others on
similar basis.

5.2 experimental results and discussion

For consistency with the results shown in the previous chapter, initial xylose con-
centration was kept at 33.3 mM. Experimental results summarized in Table 5.1
exhibit two distinct and competing effects when adding in turn KCl, KBr and
KI to aqueous acidic solutions: on the one hand xylose reaction rate increases
after the addition of all potassium halides, but to a different extent depending
on the halide, and in particular following the order Cl− >Br− >I− ; on the other
hand, selectivity, and thus furfural yield, are also improved by all potassium
halides, but following the opposite order I− >Br− >Cl− . Moreover a synergic ef-
fect becomes evident, in terms of selectivity towards furfural, when combining
the halides at the two extremes of the series, i.e. Cl− and I− , Figure 5.1. Accord-
ingly the reaction rate of xylose when adding KI and KCl salts in equal molar
amount (Table 5.1, entry 6), approaches the average reaction rate when using
only KCl or KI at the same total concentration (Table 5.1, entries 2 and 4), but,
on the other hand, it clearly displays higher selectivity and yield. In particular,
the highest selectivity and furfural yield - 95.3% and 87.5% respectively - was
measured when using a combination of KCl and KI, with the concentration of
 Table 5.1: Results from kinetic experiments, all reaction were carried out at 200 °C.
Entry Acida Salt(s) Salt(s) kX k1 k2 k3 k1/k2 Selectivity Max. yieldb
/mM /mM /10−4 [s−1 ] /10−4 [s−1 ] /10−4 [s−1 ] /10−4 [s−1 ] [-] [%] [%]
1 50 - - 43.2 34.5 8.7 1.9 4.0 79.7 69.1
2 50 KCl 500 73.9 62.8 11.1 1.7 5.7 85.0 77.7
3 50 KBr 500 63.0 55.4 7.7 1.4 7.2 87.8 80.6
4 50 KI 500 60.0 53.6 6.5 1.2 8.2 89.2 82.6
5 50 KHSO4 500 57.7 47.2 10.5 2.1 4.5 81.8 72.2
6 50 KI-KCl 250-250 67.6 61.9 5.7 2.0 10.6 91.6 82.4
7 50 KI-KCl 500-250 70.3 64.6 5.7 1.5 11.3 91.8 84.4
8 50 KI-KCl 250-500 78.2 74.6 3.7 1.7 20.2 95.3 87.5
9 50 KI-KCl 500-500 82.3 75.9 6.4 1.6 11.9 92.2 85.4
10 50 KBr 750 66.9 62.1 4.9 1.9 12.7 92.7 83.4
a H SO was employed in every experiment. The pH at room conditions was constant for every solution (1.25), and it was controlled before salts addition.
2 4
b Calculated based on eq. (3.3).
5.2 experimental results and discussion
63
64 the effect of aqueous halides on furfural formation

Figure 5.1: Effect of different halides salts on xylose reaction rate and furfural yield. Xy-
lose (dotted line) and furfural dimensionless concentration (solid line) as from
equation 3.2 after fitting to the experimental results.
 5.2 experimental results and discussion 65

KCl being double the KI concentration (Table 5.1, entry 8). This synergic effect
of I− and Cl− , besides allowing for remarkable furfural yields, offers the ground
for a deeper comprehension of the underlying mechanism of reaction.
Potassium bisulphate KHSO4 was also tested in addition to the potassium
halides. Adding KHSO4 caused an increase in the xylose reaction rate, but only
minor effects on furfural selectivity and yield if compared to no salts addition.
The minor change of the selectivity to furfural compared to the change of k X ,
suggests that the residual acidity of the HSO4− ion, or its capacity to donate the
proton, is responsible for the increase of all of the reactions rates k1 , k2 and k3 ,
leading thus to only a minor selectivity improvement. This confirms that not
all the anions show the same effect on this reaction, and that the halides in
particular show exceptional properties.
Analyzing the results it appears clear that the halides in acidic solution play a
role in at least two different and consecutive steps in the reaction leading from
xylose to furfural. In the scheme depicted in Figure 5.2 a reaction mechanism is
proposed based on the results presented in this chapter and those of previous
works [135, 165]. In this mechanism more than one option for side reactions is
assumed, whereas the formation of furfural (F) entails a distinctive sequence of
one enolization followed by three dehydration reactions, always started by proto-
nation at specific positions. Noteworthy, the kinetics for such mechanism can still
be described with very good accuracy by the scheme in Figure 3.1, considering
k2 as the sum of the possible side reactions, and considering the intermediates
to occur in very low concentrations. As far as the role of halides is concerned,
in first instance they tend to promote the formation of the 1,2-enediol 2 from
the protonated acyclic xylose (X), secondly they act promoting the first and sec-
ond dehydration steps forming 5 and 6, and the last intramolecular dehydration
and ring closure leading to F. The first enolization step is considered rate lim-
iting, as discussed before in this dissertation [165], and thus every change in
xylose reaction rate, and especially in k1 , can be ascribed to a rate enhancing
action on this particular reaction. Adding Cl− shows here the most pronounced
effect, followed by Br− and I− in the last place; thus electronegativity, and/or
the smaller size of the ion, or the - least weak - basicity in water, appear to be the
driving forces favoring the formation of 2. Regarding the mechanistic aspects of
this reaction, C-2-C-1 intramolecular hydrogen transfer has been proven the pre-
vailing mechanism for glucose dehydration in 1 M aqueous H2 SO4 , although of
minor importance at lower acid concentration [132]. Considering the relatively
low acid concentration and the significant presence of halides, the mechanism
proposed in scheme depicted in Figure 5.2 seems more realistic under the con-
ditions employed in this work. In such scheme the halides act as weak bases
assisting the enolization reaction via proton transfer, hence with a decreasing
efficacy in the order Cl− >Br− >I− . Nevertheless, this specific aspect may be clari-
fied only with further mechanistic studies under similar conditions, maybe using
isotope-exchange techniques.
the effect of aqueous halides on furfural formation

Figure 5.2: Reaction mechanism leading from D-xylose to furfural in acidic solutions. Aqueous halides are indicated as X − .
66
 5.2 experimental results and discussion 67

Figure 5.3: Main products from the acid catalyzed dehydration of sugar alcohols at high
temperature.

After the enolization reaction, three dehydration steps are required to yield F,
in particular at C-3 and C-4, followed by the final ring closure and intramolec-
ular dehydration. Analyzing the results in Table 5.1, the selectivity to F is not
only improved as a result of the increased k1 , as it was noticed earlier for Cl−
[165], but also because of a reduction in the side reactions rate k2 . This becomes
evident especially when using I− , followed by Br− and lastly Cl− , denoting a
correlation with the nucleophilicity of the halides in aqueous solution, which
derives from their size, solvation and polarizability. In view of these results, as
illustrated in Figure 5.2, the halides are suggested to assist the dehydration re-
actions by stabilizing the transition states leading to 5 and 6, where I− results
more effective because more polarizable and less strongly solvated than Br− and
lastly Cl− . The dehydration reactions are not rate limiting, thus they do not influ-
ence the xylose overall rate of reaction. The mechanism proposed in Figure 5.2
finds also confirmation in the synergic effect evidenced when using KCl and KI
in combination. These two halides have been shown to have their major effects
respectively in the first enolization step, and in the following dehydrations, and
consequently, combining the two, leads to the higher selectivities and yields.

5.2.1 The analogy with sugar alcohols dehydration

As far as the dehydrations steps are concerned, the presence of an adjacent

electron-rich enol function at 2 and 5 is also relevant for understanding this
mechanism. In this respect it is clarifying to consider the acid-catalyzed dehy-
dration of sugar alcohols, and to remind that the dehydration rate of different
pentoses in acidic solution can be directly related to their acyclic proportion
in aqueous solution [135]. The dehydration of sorbitol and xylitol, despite the
fact that they exist only in acyclic form, requires more severe conditions to take
place compared to glucose and xylose, and they do not yield the corresponding
68 the effect of aqueous halides on furfural formation

furfuryl-alcohols as one could wrongly expect. Their reaction in acid solution

is normally initiated with protonation at C-1 leading to 1,4 cyclization and in-
tramolecular dehydration, yielding relatively stable products [166–168], see Fig-
ure 5.3. Sorbitol in particular can undergo a further dehydration and cyclization
yielding a commercially valuable bio-based diol commonly known as isosorbide
(1,2:3,6-dihydrosorbitol) [169, 170].
The incomplete dehydration of sugar alcohols, and their relative reluctance to
such reaction, is indicative of the importance of the enol function present at 2
and 5 for the dehydrations to F. An enol is a very good electron donor, favoring
in this case the elimination of water molecules as depicted in Figure 5.2. The nu-
cleophilic character of the halides, especially of I− , make them good candidates
for providing similar assistance further promoting the selective dehydration to
F.

5.3 conclusions

Halides salts were shown to enhance the reaction of xylose to furfural in aqueous
acidic solution by two distinct and consecutive effects. The formation of the 1,2-
enediol from the acyclic form of the aldose was preferably catalyzed by the
halides in the order Cl− >Br− >I− , whereas selective dehydration to furfural is
assisted by the halides in the opposite order I− >Br− >Cl− . Synergic effects are
also evident when using a combination of I− and Cl− ions, with selectivity of
reaction attaining 95.3% and furfural yield 87.5%.
 6
PRODUCTION OF HEMICELLULOSE-DERIVED
C A R B O H Y D R AT E S F R O M W H E AT S T R AW V I A
D I L U T E - A C I D H Y D R O LY S I S

In the present chapter the combined production of hemicellulose-derived carbohydrates

and an upgraded solid residue from wheat straw using a dilute-acid pretreatment at mild
temperature is exposed. Dilute aqueous HCl solutions were studied at temperatures of
100 and 120 °C, and they were compared to dilute FeCl3 under the same conditions.
Comparable yields of soluble sugars and acetic acid were obtained, affording an almost
complete removal of pentoses when using 200 mM aqueous solutions at 120 °C. The solid
residues of pretreatment were characterized showing a preserved crystallinity of the cellu-
lose, and a almost complete removal of ash forming matter other than Si. Results showed
upgraded characteristic of the residues for thermal conversion applications compared to
the untreated wheat straw.

The contents of this chapter have been adapted from:

Marcotullio, G., Krisanti, E., Giuntoli, J. and de Jong, W. Selective production of
hemicellulose-derived carbohydrates from wheat straw using dilute HCl or FeCl3 solutions under
mild conditions. X-ray and thermo-gravimetric analysis of the solid residues, Biores. Tech., 2011, 102,
5917-5923.

69
70 hemicellulose-derived carbohydrates via dilute-acid hydrolysis

6.1 introduction

Hemicelluloses, together with cellulose and lignin, represent one of the three
main constituents of lignocellulosic biomass, making up for about 30% of its
structure. The predominant form of hemicelluloses are the xylans, i.e. the wide
range of polysaccharides that have a β-(1→4)-D-xylopyranose backbone with a
variety of side chains [52]. D-Xylose, the main monomeric component of such
polysaccharides, is an industrially valuable feedstock for the production of xyli-
tol, and it is gaining interest as additional feedstock, combined with glucose, for
the production of bio-ethanol via fermentation.
As largely recalled in this dissertation, although mostly as an intermediate
of reaction, xylose is the main feedstock for the production of furfural, mainly
from corncobs and sugarcane bagasse. In the view of the development of novel
solutions for the production of furfural, a particular attention is devoted to the
optimized production of xylose and arabinose from lignocellulosic biomass.
Beside xylose, acetic acid, an industrially largely employed chemical, can be
directly obtained by the hydrolysis of the hemicelluloses acetyl groups, which
can account up to a few weight percent on the initial dry biomass.

6.1.1 Hemicellulose dilute-acid hydrolysis

Hemicelluloses, due to their amorphous structure, are comparatively easily hy-

drolyzed and dissolved in water. Both cellulose hydrolysis and furfurals forma-
tion require in fact more severe conditions to take place appreciably, and they are
normally hindered at low temperatures [171, 172]. In xylose production patent
literature temperatures between 100-150 °C are reported to be a preferred op-
tion, see [173], and in a recently published work an optimal temperature of 130
°C was indicated for xylose production from rice husks under similar condi-
tions [174]. Dilute-acid pretreatment of biomass has been extensively studied,
although mostly within the context of cellulosic bio-ethanol production via fer-
mentation [175–177], resulting among the best options in terms of economic via-
bility [178, 179]. In these cases relatively high temperatures of pretreatment were
normally preferred resulting in a reduction of the cellulose crystallinity and, con-
sequently, in an improved digestibility of the solid residue. On the other hand
the reported hemicellulose recovery was not always remarkable. Only a minority
of the available studies dealing with dilute-acid biomass pretreatment focused
on the production of xylose, and nearly all of them explored a relatively low
range of temperatures, [172–174, 180–183].
 6.1 introduction 71

Figure 6.1: Representation of a portion of xylopyranose polymer, including acetyl and

arabinofuranose substituents.

6.1.2 Dilute-FeCl3 for biomass hydrolysis

In previous works and in chapter 4 it has been shown that FeCl3 catalyzes the
hydrolysis of carbohydrates and also the dehydration of monomeric sugars into
furfurals [136, 147, 165]. Aqueous solutions of FeCl3 present a Brønsted acid
character due to the hydrolysis of Fe3+ resulting in the formation of different
kinds of complexes. The main reactions taking place in this range of dilution are
[184]:
[ Fe( H2 O)6 ]3+ � H + + [ Fe(OH )( H2 O)5 ]2+
2 [ Fe(OH )( H2 O)5 ]2+ � [( H2 O)5 FeOFe( H2 O)5 ]4+ + H2 O
The equilibrium of these reactions is obviously influenced by pH and by the ini-
tial iron concentration. Hence, the H+ molar concentration of a 100 mM FeCl3
aqueous solution at ambient conditions is about 17 mM, i.e. 17% of an equimolar
HCl solution. Even then, in previous works it has been reported that the effect
of aqueous FeCl3 on carbohydrates goes beyond its H+ production at room tem-
perature, and that the Fe3+ ion itself could play an important catalytic role [136].

6.1.3 Characterization of the residues

The use of biomass residues, including agricultural residues such as wheat straw,
has great potential for thermal conversion, but it is nowadays still hindered by
some issues: the heterogeneous nature of many such fuels [185], the unknown
release chemistry of pollutants such as fuel-bound nitrogen and sulfur [186], and
72 hemicellulose-derived carbohydrates via dilute-acid hydrolysis

the quantity and quality of ash forming matter. The presence of high concentra-
tions of alkali metals combined with Si, Cl and sulfur in agricultural residues
is known to favor the formation of alkali silicates and alkali sulfates with melt-
ing points lower than most thermal conversion applications. The presence of a
molten phase, then, gives rise to several problems such as slagging, fouling, cor-
rosion and loss of fluidization [187]. In a previous work from our group [185] we
have shown how a simple water leaching pretreatment could have beneficial ef-
fects for the pyrolysis of wheat straw in terms of the quantity and quality of ash
forming matter but also showing a higher reactivity and lower char production.
In this respect, the solid residues of acid hydrolysis retain a high potential for
thermal conversion applications due to an expected reduced inorganic content
and a partially degraded structure which could facilitate pyrolytic decomposi-
tion.
In view of what exposed above, this research aims at investigating the com-
bined production of hemicellulose-derived carbohydrates and an upgraded solid
residue from wheat straw using a dilute-acid pretreatment. Mild temperatures
are explored in order to minimize sugars degradation reactions. Dilute solu-
tions of HCl are used as representative for strong monoprotic acids, and these
are compared to equimolar solutions of FeCl3 as an interesting option for acid
replacement [148]. Enzymatic degradability of the residues of dilute-acid hy-
drolysis has been extensively studied in literature, for this reason the residues
of pretreatment are analyzed in this work using X-ray and thermo-gravimetric
analysis in order to characterize their crystallinity, minerals content and their
behavior under thermal conversion conditions.

6.2 hemicellulose hydrolysis and soluble carbohydrates recov-

ery

In this work wheat straw was treated in FeCl3 or HCl dilute solutions for 120
min at 100 and 120 °C, refer to chapter 2 for a more detailed description of the
experimental methods. As a result of these treatments, increasing fractions of
biomass could be dissolved with increasing severity of reaction, reaching more
than 37 wt% at 120 °C in 200 mM HCl or FeCl3 , (Table 6.2). A large part of
the dissolved material was recovered in the filtrates as hemicellulose derived
carbohydrates, the rest being mainly unidentified extractives and partly minerals.
We report surprisingly high yields of pentoses when treating wheat straw for
120 min at 120 °C in 200 mM HCl or FeCl3 , respectively accounting for 26.5
and 26.0 wt% d.b. of the initial wheat straw, i.e. about 99% and 97% of the
potential. More than 85% of this fraction is represented by monomeric sugars.
Analyzing the effect of temperature, at 100 °C the fraction of sugars recovered in
monomeric form is always less if compared to the treatments at 120 °C, figure 6.2,
and treating wheat straw at 120 °C and 100 mM HCl or FeCl3 produced similar
 6.2 soluble carbohydrates recovery 73

[wt% d.b.]
Lignin ASL 1.0±0.0
All 15.1±0.1
Hemicellulose Arabinan 2.1±0.1
Xylan 21.5±0.7
Mannan 0.2±0.1
Galactan 0.5±0.0
Cellulose Glucan 34.6±1.0
Extractives H2 O 11.2±1.2
EtOH 2.0±0.0

Table 6.1: Biochemical composition of the untreated wheat straw, data from [13].

amounts of total sugars but much more monomeric sugars than at 100 °C in 200
mM solutions, Table 6.2. Hence catalyst concentration can be well compensated
by temperature in this range, also affording a significantly higher production of
monomers. Using pure water under the same conditions of temperature leads
to hardly any carbohydrates recovery in the liquid filtrate, so the presence of
a strong acid, even at relatively low concentrations, is crucial to hemicellulose
hydrolysis under such mild conditions.
The yield of arabinose, glucose and acetic acid is almost constant after all
the treatments, and immediately evident even after the treatments under milder
conditions. Evidently the α-L-arabinofuranoside or ester-linked acetyl group sub-
stituents are readily hydrolyzed yielding monomeric arabinose or acetic acid,
whereas the depolymerization of the longer xylopyranose needs longer reaction
times, [171, 182]. Wheat straw hemicellulose is known to contain glucose in the
form of galactoglucomannans and/or β-glucans substituents, accordingly galac-
tose and mannose are reported in minor amounts in the composition of the
wheat straw employed in this work [13]. Results presented in Table 6.2 show the
concentration of glucose to follow a trend very similar to arabinose and acetic
acid, i.e. less depending on reaction severity than xylose. Moreover, as it will be
shown below, the crystallinity of the solid residues, besides their TGA results,
indicate that cellulose is hardly affected by the pretreatments. From these indica-
tions, it can be derived that the glucose present in the hydrolysates might mainly
originate from hemicellulose side groups hydrolysis, although a partial hydrol-
ysis of cellulose cannot be entirely excluded. Galactose and mannose detection
and quantification might partly clarify this aspect, but, unfortunately, they could
not be measured in this work due to the insufficient resolution of the analytical
method employed. The use of sugars-specific HPLC configuration may be sug-
gested for this purpose.
74 hemicellulose-derived carbohydrates via dilute-acid hydrolysis

Figure 6.2: Soluble carbohydrates production from wheat straw. In the sample codes nota-
tions H and Fe indicate respectively dilute HCl and FeCl3 , and the following
number their concentration in mM; LT and HT indicate the temperature of
pretreatment, respectively 100 and 120◦ C.

The exceptionally high yields of pentoses obtained in this work are in fact
hardly reported in literature when higher temperatures or much higher acid
concentrations are employed [182, 188], mainly because of sugar reactions to
furfurals and/or other degradation products. The use of mild temperatures at
nearly atmospheric pressure, although for longer residence times, is thus a realis-
tic option for the pentoses production oriented biomass pretreatments, provided
that the larger size of the equipments can be economically outbalanced by better
selectivity of reaction, process simplification and energy savings. Besides, higher
temperatures of reaction, even if attractive for the shorter residence times and
reduced equipment size, present scale-up issues regarding biomass continuous
feeding at high pressures, material resistance and thus investment costs.
It is clearly shown in this work that equimolar aqueous solutions of HCl and
FeCl3 catalyze the hydrolysis of hemicellulose in a comparable way. The only rel-
evant difference appears at the lower concentrations (100 mM) more evidently
at 100 °C than at 120 °C, where FeCl3 seems more active than HCl. It is also
remarkable that the pH of the filtrates after the dilute FeCl3 treatments are un-
expectedly lower than after the HCl treatments at the same concentration. The
explanation for this result relies on the peculiar chemistry of iron(III) salts in
aqueous solutions. When a dilute aqueous solution of FeCl3 is boiled at normal
conditions a dark precipitate is formed, which consists mainly of iron oxides in
 Table 6.2: Sugars recovery in the filtrate expressed as wt% d.b. of initial wheat straw.

[wt% d.b.] 100 °C 120 °C

Water HCl FeCl3 Water HCl FeCl3
- 100mM 200mM 100mM 200mM - 100mM 200mM 100mM 200mM
pH filtrate [-]a 5.96 2.46 1.81 2.25a 1.74a 5.59 2.32 1.80 2.25a 1.73a
Weight loss [%wt a.r.]b 0.8 15.5 27.8 19.5 28.1 8.2 28.3 37.5 31.5 37.6
Tot. Xylose 0.8 5.3 15.0 7.3 14.2 0.9 17.2 23.3 20.0 22.9
Tot. Arabinose 0.4 2.0 3.1 2.2 2.8 0.4 2.8 3.2 2.8 3.1
Tot. Glucose 1.2 1.5 2.4 1.7 2.1 1.5 2.2 2.6 1.9 2.6
Tot. Acetic Acid 0.0 1.0 2.6 1.6 2.6 0.4 3.1 3.3 3.2 3.3
Monom. Xylose 0.1 1.1 6.1 1.4 6.4 0.1 10.0 20.1 14.3 20.6
Monom. Arabinose 0.0 1.7 2.4 1.7 2.6 0.2 2.0 2.6 2.2 2.6
Tot. Pentoses 1.2 7.3 18.1 9.5 17.0 1.3 20.0 26.5 22.8 26.0
Tot. Pentoses yield [%] 4.6 27.1 67.5 35.5 63.3 5.0 74.7 98.9 85.0 96.8

a As explained in the Methods, the pH of all the filtrates is measured at room temperature after dilution to 50ml. It is important to notice

that, due to the equilibrium of Fe3+ hydrolysis reactions the H+ concentration of a 5 times more concentrated solution would be less
than 5 times larger. b Weight loss was measured on the basis of the initial weight a.r., after drying every pretreated sample for 24h at
room temperature.
6.2 soluble carbohydrates recovery
75
76 hemicellulose-derived carbohydrates via dilute-acid hydrolysis

the form of α-Fe2 O3 and β-FeOOH, as reported by Voigt and Göbler [189]. Con-
sequently the H+ concentration of the FeCl3 liquid solution increases according
to the simplified reaction:

3Fe3+ + 5H2 O � 9H + + Fe2 O3 � + FeOOH �

For these reasons aqueous FeCl3 , already at 100 °C, can be regarded as a source
of HCl and iron oxides, lowering accordingly the pH of the solution. This is con-
firmed by the filtrates pH measurements (Table 6.2) and by the analysis of the
inorganic elements distribution in the solid straw before and after the pretreat-
ments, Table 6.3. Clearly Cl and Fe are not completely removed from the solids
with washing, but the Fe/Cl weight ratio in the FeCl3 pretreated wheat straw
is clearly exceeding that of the pure salt. Although the presence of Cl can be
mainly ascribed to the incomplete washing, the higher Fe proportion is due to
its precipitation in solid forms.
Crossing the filtrates pH with the pentoses yield at constant temperature (Ta-
ble 6.2), a relation is evident between the H+ final concentration and the extent
of hemicellulose hydrolysis, whereas, contrary to what was reported before, the
effect of the Fe3+ ions results to be mainly indirect and referable to the H+
production. Hence, treating biomass with FeCl3 or HCl solutions is not really
different from the catalytic point of view. The opportunity of using dilute FeCl3
as an alternative to dilute HCl should be evaluated in a broader perspective,
carefully evaluating environmental and economic benefits and costs. On the one
hand the use of FeCl3 might be less harmful than HCl, moreover, due to its equi-
librium reactions in hot water, it presents a pH buffer effect partially mitigating
the neutralizing capacity of agricultural residues like straw. On the other hand,
FeCl3 contains three times as much Cl compared to HCl on a molar basis, and
the deposition of fine iron oxide particles on the biomass residues, probably also
containing Cl, may have consequences in the downstream applications. In the
second part of this work the effects of such Fe containing particles under thermal
conversion conditions are reported, but their interaction with microorganisms
or enzymes during biological conversions, e.g. in cellulosic ethanol production,
should also be carefully evaluated.

6.2.1 Crystallinity of the solid residues

The residues treated at 120°C and 200mM HCl and FeCl3 show a significantly
higher crystallinity index than untreated wheat straw, Table 6.4. The CrI is a rela-
tive measure of the intensity of diffraction in the crystalline region compared to
the amorphous region, thus the treated samples show an increased proportion
of crystalline cellulose compared to amorphous cellulose and, more importantly,
hemicellulose. A significantly increased CrI is not a common feature to all pre-
treatments aimed at the fractionation of biomass, as shown by Kumar et al. [127],
 6.3 xrf and thermo-gravimetric analysis 77

and it may be considered a relevant aspect of this particular pretreatment. The in-
creased CrI not only indicates an extensive hemicellulose removal, which is clear
from the filtrates analysis, but also shows the limited effect of such pretreatment
on the crystalline structure of cellulose, which can be attributed to the relatively
mild conditions employed. The cellulose content of the residues of pretreatment
was not directly measured, although, considering the glucose concentration in
the filtrates and the samples weight loss reported in Table 6.2, it is expected to
be markedly higher compared to the untreated wheat straw. Based on all these
observations, it will be interesting for future research to test a pretreatment of
this kind, possibly followed by a lignin separation/solubilization step, for the
selective production of cellulose fibers, lignin and hemicellulose-derived carbo-
hydrates. In particular it may be suggested to test a uncatalyzed organosolv type
of treatment for the residues of dilute-acid hydrolysis, preferably using ethanol
as solvent in the same range of temperature, aimed at recovering large fractions
of the water-insoluble lignin, leaving a clean, cellulose-rich residue with high
crystallinity.

6.3 results of xrf and thermo-gravimetric analysis of the resi-

dues of pretreatment

From the distribution of inorganics in all the pretreated samples compared to

the untreated wheat straw (Table 6.3), silica (SiO2 ) appears to be present in sig-
nificantly larger amounts compared to all other elements. Silica is insoluble in
water and thus cannot be expected to be removed in appreciable amounts dur-
ing these kind of pretreatments. This is not true for potassium (K), which is
the compound with the largest concentration in the untreated straw, and it is
largely removed after all pretreatments. The extent of K removal is around 90%
(K2 O basis) already after the pure water pretreatment, and seems to increase
slightly with severity of reaction reaching more than 97% in some cases, Table
6.3. Calcium (Ca) was also largely removed during all treatments, apart from
those using pure water, where only about 50% was removed. Evidently Calcium
compounds show acid preferred solubility, in agreement with the general low
water solubility of calcium alkaline salts like CaO, Ca(OH)2 and CaCO3 . Magne-
sium (Mg) shows a behavior similar to that of Ca, but with much lower concen-
trations. Chlorine (Cl) was removed thoroughly (98%) when using pure water,
although being present in all the other pretreated samples as a result of HCl
or FeCl3 incomplete washing. Looking at the little Cl residues left when using
pure water, it can be expected to be removed thoroughly also when employing
acids not containing Cl. Iron (Fe), and similarly Aluminum (Al), are removed to
a much lesser extent probably because of the limited solubility of their oxides.
Accordingly, the large Fe containing residues in the samples treated with FeCl3
is explained because of the mentioned precipitation of Fe oxides. The presence of
78 hemicellulose-derived carbohydrates via dilute-acid hydrolysis

Cl in the solid Fe precipitates is also not to be excluded. Phosphorus (P), detected

as phosphates, and especially Sulphur (S), present as sulfides, are also removed,
but less than other compounds. They seem to be more difficult to remove both
with pure water and acidic pretreatments, also not showing any clear relation
with the reaction severity. A fraction of P and S could in fact be included in more
complex organic compounds beside soluble salts, similarly to what reported in
a previous work [126]. Anyway the presence of P and S is of minor importance
being less than 1 wt% of the total inorganics.
The results for thermo-gravimetric analysis (see Table 6.4) show that all the
pretreated samples displayed a reduction of both the char residue at 550 °C and
the ash residue if compared to the untreated wheat straw, as it could be expected
mainly as a result of the reduced alkaline minerals content [185, 190, 191]. The
total amount of volatiles released also increased for all the pretreated samples
if compared to the untreated straw (68.7 wt% d.b.), approaching values around
80 wt% d.b., (Table 6.4). Such results have to be evaluated also considering that,
in some cases, the pretreated straw was deprived of significant fractions of its
main constituents. In fact the hemicellulose fraction, which is known to undergo
devolatilization at lower temperatures [190], is removed to significant extents
for all the samples apart from those treated with pure water or under milder
conditions (namely 100mM of HCl and FeCl3 at 100 °C).
In general, looking at the differential thermogravimetric plots (DTG), fig. 6.3,
it is evident that the large presence of alkalies in the untreated wheat straw
catalyzes the reactions of devolatilization, presenting a broad peak of reaction
rate occurring at 323 °C, quite common for wheat straw [185, 190]. In the water
treated samples two peaks become more evident, the first occurring at tempera-
tures slightly lower than 300 °C and the second in the range of 350 °C. Here two
different reactivities for the hemicellulose and cellulose fractions of the straw
become evident due to the leaching of the alkaline minerals, as discussed in pre-
vious works [185]. The dilute HCl pretreated samples show in general highest-
temperature peaks, reaching a maximum of 361 °C for the sample treated at 120
°C in 200 mM HCl. In these cases the increasing removal of hemicellulose, to-
gether with the alkaline minerals, displays a behavior slightly resembling pure
cellulose [190, 192]. All samples treated in FeCl3 present reaction rate peaks at
lower temperatures if compared to the corresponding HCl treated samples. Be-
sides, opposite to the HCl case, increasing FeCl3 concentration appears to lower
such temperature, thus denoting an improved devolatilization, even considering
that larger fractions of hemicellulose have been removed. As anticipated before,
the Fe deposits on the solid residues can be considered to be responsible for this
catalytic effect. Thus, if on the one hand the removal of alkalies and of large part
of the hemicellulose shifts the devolatilization reaction rate towards higher tem-
peratures (in the range of pure cellulose), the presence of iron deposits on the
residues acts as catalyst for the devolatilization reactions. At the same time the
amount of fixed carbon at 550 °C is also increasing for the FeCl3 treated residues,
Table 6.3: ash composition distribution in wt% of the most important elements making up for >99 wt% of the total inorganics. C, H,
O, and N are excluded. Values in italic indicate the amount left compared to the untreated wheat straw. These values are
calculated on SiO2 basis, assuming that no SiO2 is removed after pretreatments. This is a conservative hypothesis and thus
these values represent rather an upper limit of the residual compounds rather than the actual residue.

100 °C 120 °C
Untreated Water HCl FeCl3 Water HCl FeCl3
[wt%] Wheat Straw - 100mM 200mM 100mM 200mM - 100mM 200mM 100mM 200mM
SiO2 54.8 86.0 89.2 93.3 71.0 80.5 87.5 89.8 92.1 87.7 80.5
100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100%
K2 O 28.0 4.9 2.8 1.2 3.0 1.1 3.7 3.0 2.3 1.2 1.4
100% 11.1% 6.2% 2.5% 8.3% 2.6% 8.2% 6.5% 5.0% 2.7% 3.4%
Cl 5.7 0.2 2.6 2.4 9.5 6.1 0.1 3.1 2.0 3.7 6.9
100% 1.7% 27.6% 25.2% 128.8% 73.1% 1.6% 32.8% 21.0% 40.6% 82.1%
Fe2 O3 0.5 0.7 0.7 0.3 13.2 9.8 0.6 0.4 0.2 4.7 8.7
100% 78.5% 76.4% 36.9% 1855.5% 1223.4% 65.7% 39.7% 24.4% 539.3% 1083.2%
CaO 7.0 5.4 1.5 0.4 0.8 0.3 5.4 0.9 0.6 0.3 0.3
100% 48.9% 13.1% 3.5% 8.6% 3.0% 47.8% 8.0% 5.4% 2.6% 3.3%
Al2 O3 0.4 0.5 0.5 0.4 0.3 0.5 0.5 0.4 0.5 0.5 0.4
100% 95.7% 89.7% 62.1% 64.2% 91.9% 83.5% 61.7% 77.0% 82.6% 84.1%
MgO 1.0 0.6 0.3 0.1 0.2 0.2 0.6 0.3 0.4 0.2 0.2
100% 37.5% 16.1% 4.6% 14.6% 10.0% 36.3% 17.8% 21.4% 10.7% 14.5%
P2 O5 0.3 0.6 0.8 0.7 1.0 0.5 0.7 0.7 0.8 0.6 0.4
100% 27.0% 35.4% 29.0% 60.2% 26.9% 31.5% 32.4% 34.8% 29.8% 21.8%
S 0.8 0.5 0.8 0.8 0.6 0.6 0.6 0.9 0.8 0.6 0.5
100% 52% 62.8% 55.3% 58.4% 52.6% 47.5% 68.4% 61.0% 48.7% 43.6%
6.3 xrf and thermo-gravimetric analysis
79
hemicellulose-derived carbohydrates via dilute-acid hydrolysis

Table 6.4: Thermo-Gravimetric Analysis of the solid residues.

[wt% d.b.] Wheat Straw 100◦ C 120◦ C
(Untreated) Water HCl FeCl3 Water HCl FeCl3
- 100mM 200mM 100mM 200mM - 100mM 200mM 100mM 200mM
Moisture [wt% a.r.] 5.42 6.3 6.3 3.4 5.5 5.9 8.5 4.7 5.0 4.0 5.5
Chars 550◦ C 30.3 22.9 19.4 22.3 23.8 23.3 20.9 22.6 23.8 23.8 24.2
Volatiles 550◦ C 68.7 77.1 80.6 77.7 76.2 76.7 79.1 77.4 76.2 76.2 75.8
Fixed carbon 550◦ C 21.8 15.8 13.5 14.2 17.3 16.9 16.8 13.5 14.1 14.6 17.2
Ash 9.5 7.2 5.9 8.0 6.5 6.3 4.1 9.1 9.6 9.2 7.0
Ash on initial [%]a 100 74.7 52.5 61.0 55.0 48.0 39.8 68.5 63.4 66.5 46.1
Peak Temp. [◦ C] 323.3 354.8 355.3 358.1 349.6 345.3 354.7 359.7 361.2 353.7 348.8
Crystallinity Index 47.6 - - - - - 54.8 - 59.5 - 59.5
I002 (Crystalline) 1360 - - - - - 1234 - 1946 - 1809
a Ash present in the residue divided by the initial ash in the untreated wheat straw taking into account the weight loss after the
pretreatment.
80
 6.4 conclusions 81

and this can also be ascribed to the presence of Fe deposits. Similar conclusions
on cellulose reactivity were reported by Varhegyi et al. [192] as a consequence of
the addition of FeSO4 besides other inorganic salts.
Due to the nearly total removal of unwanted alkaline materials like K and
Cl, the residues of pretreatments present generally upgraded characteristics as
solid fuel for combustion or gasification. The benefits from alkaline material re-
moval are a reduced char production, as far as the the primary reactions are
concerned, but also significant improvements in terms of equipment operation.
The potential for slagging and fouling, equipment corrosion and fluidised bed
agglomeration is in fact significantly reduced, becoming comparable to that rel-
ative to higher quality woody biomass [187, 193, 194]. As a matter of facts the
high alkaline minerals content of agricultural residues, and their consequent ash
behavior, is among the main reasons to hinder their use in large scale thermal
applications despite their large availability and relatively low cost. An upgrad-
ing in this sense could increase the value of the residue itself making it attractive
for use in large scale heat and power units in combination, or replacing, solid
fossil fuels.
A thorough removal of alkaline minerals contained in wheat straw is also
afforded by dilute FeCl3 , although accompanied by solid iron oxide particles
deposition on the biomass residue, affecting the devolatilization reaction. The
effect of such micro particles on secondary gasification reactions, for instance in
fluidized bed reactors, would be worth further investigation, considering that
iron containing minerals, like olivine, have been successfully tested as bed ma-
terial for their positive catalytic effects on tars decomposition reactions [195].
Nevertheless, the opportunity of using iron containing salts for the pretreatment
of biomass as an alternative to mineral acids, should be evaluated in a broader
perspective, carefully balancing economic and environmental benefits and costs.

6.4 conclusions

It was shown that minor concentrations of hydrochloric acid in water can be ef-
fective for hemicellulose hydrolysis already at temperatures between 100 and 120
°C. Recovery of hemicellulose derived carbohydrates, mainly considering xylose
and arabinose, approached 100% under specific conditions. Dilute solutions of
FeCl3 were confirmed to be an option for mineral acid replacement, even though
the Fe3+ ions were proved to act only indirectly on the hydrolysis of hemicellu-
lose via their partial precipitation to iron oxides and consequent formation of
HCl.
The residues of pretreatment showed upgraded characteristics with respect to
the untreated wheat straw, presenting improved crystallinity and a significant
reduction of alkaline minerals with direct consequences on their behavior under
thermal conversion conditions.
82 hemicellulose-derived carbohydrates via dilute-acid hydrolysis

(a) Wheat straw untreated and pretreated at 100◦ C (b) Wheat straw untreated and pretreated at 100◦ C
for 120 min. for 120 min.

(c) Wheat straw untreated and pretreated at 120◦ C (d) Wheat straw untreated and pretreated at 120◦ C
for 120 min. for 120 min.

Figure 6.3: DTG plots of wheat straw residues of pretreatments compared to the un-
treated, at 10 °C/min under nitrogen.
 7
A NOVEL PROCESS FOR MAKING FURFURAL

Make things as simple as possible, but not simpler - Albert Einstein

The majority of current production is still based on more or less modified versions of
the original Quaker Oats process (1921). For reasons that can be related to their lim-
ited technological evolution, the production processes in use today generally suffer from
low yields (around 50%), besides significant economical and environmental concerns.
All these reasons hindered the expansion and modernization of the furfural industry
below its actual potential. A profound technological development is a priority in order
to upgrade furfural to a large-volume bio-based commodity. The integrated production
of furfural within modern biorefineries is a big opportunity, and it will most probably
represent the next cornerstone in the development of furfural industry. In this chapter
an innovative process concept is described, aimed at an economically viable and environ-
mentally sound furfural production from biomass, with reduced energy and chemicals
consumption.

The contents of this chapter have been adapted from the patent application:
Marcotullio, G. and de Jong, W., Process for the production of Furfural from pentoses, patent
application PCT/NL2011/050730; October 2010.

83
84 a novel process for making furfural

7.1 current furfural production processes

A variety of processes for the production of furfural from biomass have been
studied and applied across last century. The concept behind all such processes is
impregnating the raw solid biomass feedstock with an acidic solution (normally
consisting of aqueous sulfuric acid), and reacting it at temperatures ranging from
150 to 250 °C. High pressure steam is passed through the reactants in order to
keep them at reaction temperature and to strip out the furfural formed. Furfural
is always recovered in the aqueous distillate at concentrations not exceeding 6%
by weight.
Sulfuric acid use is substantial, up to a rate of 2.2 wt% on initial dry weight
of biomass, therefore the discharged solids can be very acidic. In some process
configurations acid usage is reduced, and the consequent loss of furfural yield
is compensated by the use of higher temperatures and possibly larger amounts
of steam.
A thorough review of existing process concepts was reported by Zeitsch [7],
whereas more recent developments were reported by de Jong and Marcotullio
[55].

7.1.1 Operations of existing furfural production processes

Processes industrially employed nowadays are not very different from the origi-
nal Quaker Oats process dating back to the 1920s, see Figures 7.1 and 7.2. These
processes are described more in details in literature [7, 41] from where some key
figures can be drawn.
Reported furfural yield is about 11% on dry initial weight of biomass, which
corresponds roughly to 50% of the theoretical yield. Oat hulls, which were origi-
nally employed, are quite rich of pentosans, thus furfural mass yield might even
be lower when considering a different feedstock. For sugarcane bagasse, another
common feedstock used for furfural production, the yield of furfural is normally
between 9 and 10 % on initial dry matter.
The sulfuric acid usage is 2.2 wt% on initial dry biomass, and thus in the order
of 20 wt% on recovered furfural basis.
Steam consumption for furfural production is remarkable. Furfural is recov-
ered in a 5.8 wt% solution in water, thus 16.2 kg of steam per kg of furfural
recovered are employed. In the original Quaker Oats process the operating tem-
perature is 153°C with a residence time of 5 hours. Under these conditions the
theoretical energy need for the production of saturated steam is 2.64 MJ/kg, re-
sulting in a specific energy consumption of 42.9 GJ per metric ton of furfural
produced.
The follow-up furfural distillation is also energy intensive, even though, in
more integrated designs, it relies completely on large amounts of heat recovered
Figure 7.1: Schematic of the batch Quacker Oats process, adapted from [7]. HPS, LPS: High and Low Pressure Steam.
7.1 current furfural production processes
85
a novel process for making furfural

Figure 7.2: Schematic of a Chinese furfural plant with capacity of 2.5 kton/y (6 reactors), adapted from [7].
86
 7.2 furfural production in modern biorefineries 87

from the condensation of the high-pressure vapors emerging from the main re-
actor(s). Considering the schematic of an average batch process used in China,
Figure 7.2, the production of high-pressure steam for the reactor is the main
energy input to the process. Hence, it is not surprising that a strong correlation
emerges between furfural prices (strongly influenced by Chinese producers) and
asian coal prices [90].

7.2 the opportunity of furfural production within integrated

biorefineries

Furfural production is normally carried out in dedicated plants, which, despite

being based on residual feedstock like sugarcane bagasse or corncobs, do not
really comply with the requisites of a modern and integrated biorefinery as
described in Chapter 1. The solid residues of furfural production are normally
not further valorized except from incineration.
In a modern lignocellulosic biorefinery, every fraction of the biomass should
be used possibly for added value productions, minimizing the wastes and the
environmental footprint. In the majority of biorefinery concepts cellulose is ad-
dressed as the most valuable constituent of biomass, and for this reason many
biomass pretreatment processes aim at separating cellulose from hemicellulose
and lignin. On the other hand, the hemicellulose fraction of biomass is of interest
for furfural production.
As shown in chapter 6, among the main biomass constituents, hemicellulose
is relatively hydrophilic and, moreover, readily accessible and fast hydrolyzing.
Opposite to hemicellulose, lignin is only marginally soluble in water, and cellu-
lose has a crystalline structure significantly more difficult to hydrolyze than the
amorphous structure of hemicellulose. Because of these characteristics, in many
biomass fractionation processes of interest for the biorefinery industry, somehow
involving the use of water either as solvent or as anti-solvent, hemicellulose-
derived carbohydrates normally end up in an aqueous stream together with
other water-soluble impurities.
In view of this, two avenues can be foreseen for furfural production within
modern biorefineries:

1. Furfural production via traditional processes as an integrated biomass pre-

treatment process: simultaneously converting pentosans into furfural and
modifying the structure of the solid biomass residue in order to prepare it
for further processing.

2. Furfural production as a follow-up of biomass pretreatment and fraction-

ation: where a pentosan-containing aqueous stream is directed to a dedi-
cated furfural production facility.
88 a novel process for making furfural

In the first case, due to the severity of reaction normally employed for furfural
production, the structure of the residue is significantly modified compared to the
raw biomass, and it should be assessed for further conversions and/or separa-
tions. In this perspective the limitations associated to traditional furfural produc-
tion would remain basically unsolved, although advantages might result from
system integration. When considering this option it is also important to remem-
ber that the biomass residue might still contain furfural(s) in amounts incompat-
ible with further biological treatments due to enzymes inhibition (furfural(s) are
well-known enzymes inhibitors).
In the second case soluble pentosans contained in a liquid aqueous stream,
potentially arising from a variety of pretreatment technologies, may be used for
furfural production. The idea of producing furfural via a two-steps process was
already explored in the past, but the solutions proposed never reached commer-
cial scale mainly because of the additional costs associated to the production
of xylose [7]. Higher investment cost is the main limitation to such approach,
although future perspectives might be promising considering the innovation ef-
forts towards the development of new integrated biorefineries and new biomass
pretreatment technologies.
In particular, cellulosic ethanol refineries being set-up in many countries, rep-
resent potentially an enormous source of pentosans, mainly available in side
streams. In this view the current paradigm for furfural production could be
shifted towards innovative, possibly more efficient, processes involving only liq-
uid and vapor streams.

7.3 an innovative concept for furfural production with high

yield and low energy consumption

In view of the reasons mentioned above, and based on the results discussed
in the previous chapters, a novel process concept has been developed at Delft
University of Technology, which is schematically depicted in Figure 7.4. The pro-
cess consists of feeding an aqueous solution containing pentoses at the top of
a reactive distillation column, where furfural production takes place. The liquid
downflow is under optimal conditions for the conversion of monomeric pen-
toses to furfural, which, once formed, is readily transferred to the upflow vapor
stream. Needless to say, soluble pentosans, if present, will be rapidly hydrolyzed
to monomeric pentoses under the conditions required for the formation of fur-
fural.
As depicted in Figure 7.4, vapor recompression is the key solution opted for
energizing the column with minimal energy consumption. Due to the peculiar
furfural-water thermodynamics in the dilute region (Figure 7.3), the temperature
difference between the top and the bottom of such column is expected to be in
the order of 2 °C or lower. For this reason the work required for vapor compres-
 7.3 an innovative process for furfural production 89

Figure 7.3: Furfural-Water T-x diagram at 1 atm., adapted from [8]

90 a novel process for making furfural

sion can be extremely low, allowing for remarkable energy savings compared to
the existing processes.
Vapor compression is also proposed for the azeotropic distillation of furfural,
with the aim of minimizing the energy consumption, and thus the operating
costs. As it will be shown below, encouraging preliminary simulation results
make this process concept an interesting option for furfural production from
very dilute solutions of pentosans. Remarkably, raw biomass hydrolysates emerg-
ing from many biomass pretreatment technologies can be directly fed to such
kind of process without further purification or concentration.

7.4 catalyst choice

In principle this process could be operated with or without the use of an acid
catalyst, as furfural is also formed at neutral pH although in minor amounts.
Nevertheless, in order to achieve optimal yields, the use of a strong acid catalyst
is desirable. This may be a mineral or organic acid dissolved in the aqueous
liquid downflow, but also a solid acid of the kind described in Chapter 1 bound
to the column packing.
The use of suitable solid acids is certainly a preferred choice in the medium-
long term, although many issues regarding the catalyst fouling, deactivation
and regeneration should be carefully evaluated. Moreover, dissolved monova-
lent and divalent alkalis like K, Na, Ca and Mg are always present in biomass
hydrolysates and, unless completely removed, they may likely poison the cata-
lyst by cation exchange with H+ at the acid sites [66].

7.4.1 Homogeneous catalysis and recirculation

In the near term homogeneous catalysis may be considered a viable option for
this process due to its simplicity, no need for regeneration and low cost, pro-
vided that acid use is minimized, and only a part of the equipment is exposed
to corrosive environment. To this aim, the acid containing column bottoms are
recirculated within the process. By doing so, the downsides associated to homo-
geneous acid catalysis like acid consumption, neutralization and disposal of the
spent salts are significantly limited.
On the other hand, recirculation requires dealing with impurities buildup,
mainly represented by unwanted organics present in the feed, beside the prod-
ucts of side reactions. In order to control the composition of the recirculation
stream, it should be partly diverted and replenished with a “clean” solution. A
treatment for organic carbon removal at this stage allows for both limiting the
concentration of impurities in the recirculation stream, and recycling the min-
eral acid catalyst. Such purification step could be based on existing wastewater
treatment technologies, like adsorption or oxidation, or even on upcoming tech-
Figure 7.4: Process sketch. The grey shaded region includes the equipment parts exposed to a particularly corrosive environment,
requiring thus appropriate corrosion-resistant construction materials.
7.4 catalyst choice
91
92 a novel process for making furfural

nologies like aqueous phase reforming, or supercritical-water gasification, where

the dissolved organic carbon is converted to a valuable syngas.
As a consequence of recirculation the dissolved acid will stay confined to a lim-
ited portion of the process and continuously recycled, with no need for replace-
ment. It is noteworthy that, by doing so, very limited or no acid net consumption
would be associated to furfural production.

7.4.2 Halides addition for optimal yields and separation

According to the results shown in Chapters 4 and 5, the use of inexpensive alkali
or earth alkali halides, such as NaCl, in dilute-acid solution is a simple option
to achieve optimal furfural selectivity and yields. Such practice would perfectly
suite this process design, allowing for reaction selectivities up to 95%, Table 5.1.
Beside the mentioned effect on reaction catalysis, the use of dissolved salts is
also expected to enhance the separation of furfural by salting-out effect, further
reducing both furfural losses and energy requirements. Such salting-out effect
is known since the earliest experiments on furfural formation [8], and experi-
mental evidence was provided on the effects of salts on furfural-water mutual
solubility and miscibility temperature [9]. Moreover NaCl has been recently used
to enhance furfural production and separation via solvent extraction using THF
[72].
Although the salting-out effect on vapor-liquid equilibrium was never quanti-
fied, it is expected to be relevant for the separation of furfural via steam strip-
ping.

7.5 process simulation

In order to evaluate the performance of the process both in terms of furfural

yield and mass and energy balances, a model was created using the commercial
process simulation software Aspen PlusTM V7.0 from Aspen Technology Inc.
Reaction kinetics relative to furfural formation from pentoses, as well as side
reactions and furfural loss reactions, were implemented in the model based on
the results exposed Chapter 3.
The knowledge of the thermodynamics of the binary system furfural-water
is also critical for the modeling to be satisfactory, and to this aim a literature
survey was carried out in order to asses the best models available [8, 9, 196–199].
The furfural-water system is well known, and experimental data for vapor-liquid
equilibrium have been reported by several authors, especially in the temperature
range between 40 and 170°C. Mutual solubility of furfural and water is also well
known at 1 atm [9, 200].
The highly non-ideal behavior of such mixture requires a careful evaluation
of the models to employ for the simulations. Especially when modeling vapor
 7.5 process simulation 93

Figure 7.5: Furfural-water vapor liquid equilibrium in the dilute region. Measurement
points for vapor (squares) and liquid (circles) composition against the NRTL
model (solid line) at 1, 6 and 9.5 atm. Experimental data from [8, 9].

recompression, simulation results are particularly dependent on the precise esti-

mation of the temperature at the bubble and dew point of the mixture, besides
furfural vapor-liquid partition coefficients. Suitable thermodynamic models have
been recently suggested in literature. Both NRTL and UNIQUAC binary corre-
lation parameters were provided for estimating the liquid-phase activity coeffi-
cients, NRTL showing somewhat more satisfactory results. After a careful com-
parison with the available experimental data, especially in the dilute region, the
NRTL model was chosen for the estimation of the liquid activity coefficients,
whereas the Redlich-Kwong equation of state was used for the estimation of the
vapor-phase fugacity coefficients.
Experimental data in the dilute region are compared to the model between 1
and 9.5 atm, see Figure 7.5, showing satisfactory results. The azeotrope tempera-
ture is slightly underestimated at the higher pressures, but it does not represent
a problem for the scopes of this work as the uncertainty is in the order of 0.5
°C. On the other hand, furfural vapor-liquid partition coefficient is reasonably
well described, especially within the solubility region (8.3 wt% furfural at 20°C),
which is the region of interest for furfural-water distillation, Figure 7.6. For en-
gineering purposes such model can be considered satisfactory across the whole
temperature range.
94 a novel process for making furfural

Figure 7.6: Furfural partition coefficient in dilute furfural-water mixtures. Experimental

data compared to the model. The grey shaded region includes data of minor
interest for furfural distillation.

7.5.1 Results for a relevant process configuration

The process described here is an intellectual property of TU Delft, and a patent

application has been filed to both the Dutch and US patent office and it is at the
moment under examination. In order not to interfere with the patenting proce-
dures, and because of confidentiality issues, only partial information about the
process conditions may be disclosed here as resulting from preliminary studies.
In Figure 7.7 main information regarding a relevant process configuration are
depicted, including mass rate, pressure, temperature, besides stream composi-
tion limited to pentoses and furfural concentration. The targeted capacity is
about 2.7 ton/h furfural, similarly to current furfural production plants.
The feed solution contains 5 wt% pentoses, and is further diluted after mixing
with the recycle stream. The recycle stream carries specific amounts of an acid
catalyst, in order to obtain optimal furfural yields. The rate of furfural-containing
distillate equals the feed rate, so to maintain a constant rate of recirculation.
As already mentioned, the wastewater treatment is intended for the removal of
unwanted organic impurities, leaving the catalytic mixture unchanged.
Furfural separation and concentration is based on the traditional azeotropic-
distillation, although integrated with the vapor recompression in analogy with
Figure 7.7: Main process streams of a relevant process configuration. Negligible pressure drop over the columns and heat exchangers is
assumed at this stage.
7.5 process simulation
95
96 a novel process for making furfural

the main reactor. An additional furfural purification column may be also added
on the basis of the traditional furfural distillation as described by Zeitsch [7].
As it can be calculated from the main streams composition shown in Figure
7.7 furfural molar yield from pentoses is 83%. Even if the composition of the
catalyst containing streams is not indicated, such yield is in line with the re-
sults exposed in chapters 4 and 5. Because the chemistry of furfural formation
has been largely discussed in the previous chapters, this aspect is not further
exposed here, preferring to focus on process design issues.

aspects related to vapor recompression Vapor recompression repre-

sents the key aspect of the process as far as the energy consumption is concerned,
especially addressing the costly steam stripping of furfural. Mechanical vapor
compression is a known technique largely employed in seawater desalination for
its relatively low cost, compact installation and low energy consumption. Raising
the condensation temperature of the distillate vapor above the reboiler operat-
ing temperature enables heat transfer between the condensing distillate and the
boiling bottoms, making the distillation process practically self-sustaining. This
is accomplished at the expense of the top vapor compression. Hence, the energy
required for vapor compression is the main energy input to the process and,
in analogy with heat pumps, the closer the top and bottom temperatures, the
lower will be the work input. Therefore, when dealing with close boiling-point
mixtures the benefits brought about by this technique may be substantial.
As already shown, the furfural-water system, despite the significant difference
between the normal boiling points of the pure components (161.7 and 100°C re-
spectively), presents peculiar mixture behavior due to the hydrophobic character
of furfural. Especially in the dilute region, up to the azeotropic composition (35
wt% furfural), this mixture shows very little variation of boiling temperature,
resulting particularly suitable for vapor compression distillation. Moreover, due
to the limited liquid solubility of furfural in water (8.3 wt% at 20 °C), concen-
trated furfural spontaneously separates from the condensed distillate, see Figure
7.7. For these reasons the overall furfural-water separation process is particularly
simple and economical compared to most of the biomass derived chemicals.

7.6 process economics

From the analysis of the streams depicted in Figure 7.7 it is evident that the
direct energy use for producing and separating one ton of furfural starting from
an aqueous stream containing 5 wt% pentoses is 0.67 MWhe , representing a
significant improvement compared to the 42.9 GJth per ton of furfural needed
only for steam production in a traditional process. Even considering a reasonably
wide variation of the country-dependent overall thermal conversion efficiency
in power production (from 35 to 50%), the savings in terms of primary energy
 7.6 process economics 97

Table 7.1: Summary of installed equipment cost

Equipment Main characteristics Material kUS $

(2010)
Reactor systema Heat exchange area = 4326 m2 , hold-up Titanium 6509
volume = 35 m3
Vapor compressor #1 Installed power 1300 kWb CS 1195
Heat recovery system #1 Hydrolysate feed/boiler discharge. SS304 1398
Heat exchange area = 2250 m2

Process water treatment Based on activated carbon adsorption, Titanium 2900

and replenishment wastewater rate = 30 ton/h

Stripping column D = 2.5 m; 30 trays, [7] SS304 725

Reboiler #2 Heat exchange area = 1335 m2 SS304 977
Vapor compressor #2 Installed power 500 kWb CS 464
Decanter Volume 40 m3 , includes cooling coil CS 94
Heat recovery system #2 Decanter inlet/outlet, heat exchange SS304 601
area = 675 m2

Total installed equipment cost 14862

a Including reboiler and mass transfer equipment.

b It includes drive, gear mounting, base plate, and normal auxiliaries equipment.

are between 84.0 and 88.7%. It can reasonably be assumed that energy costs
per ton of furfural would be reduced to a similar extent, and so it would the
environmental footprint and CO2 emissions.
A higher initial concentration of pentoses in the feed stream may be consid-
ered in order to reduce both the variable and fixed costs, although care must be
taken considering the consequent increase of the yield-reducing second-order
loss reactions briefly discussed in chapter 3.

7.6.1 Total cost of production estimate

In order to fully evaluate the economic profitability of the process, a comprehen-

sive cost-of-production estimate has been performed using both suitable cost-
engineering models [201, 202], and real cost indications of actual furfural process
equipment provided by experts in the field [90].
The cost model is based on the process configuration depicted in Figure 7.7,
and on the following assumptions:
98 a novel process for making furfural

Table 7.2: Breakdown of Fixed Capital Investment (FCI)a

Investment item Factor Cost kUS$ % of FCI

Installed equipment 1.47 14862 35 %
Piping 0.66 6673 16 %
Instrumentation and control 0.18 1820 4%
Electrical 0.11 1112 3%
Buildings 0.18 1820 4%
Yard Improvement 0.10 1011 2%
Service facilities 0.70 7077 17 %
Total Direct costs 34374 81 %

Engineering and supervision 0.33 3336 8%

Construction and fee 0.41 4145 10 %
Contingency 0.07 708 2%
Total indirect costs 8189 19 %

Total Fixed Capital Investment (FCI) 42564 100 %

a Estimated using the Peters and Timmerhaus factors [201, 202]

• All costs and revenues are expressed in US$, which is the currency globally
adopted in the furfural market.

• Pentoses are available in a liquid hydrolysate at a relatively low concentra-

tion of 5 wt%, at price of 130 US $/ton, dry basis [72].

• The plant is operated for 8150 hours per year, assuming that furfural pro-
duction is integrated within a non-seasonal industry, such as pulp mills or
perennial-crop based biorefinery.

• The total furfural capacity is 22 kton/y.

In Tables 7.1, 7.2 and 7.3, a summary of estimated installed equipment cost, total
fixed capital investment (FCI), and total cost of production are reported.
Investment cost are relatively high for this process. The FCI of 42 M$ is more
than double with respect to a traditional process with equivalent capacity [90].
This is to be expected as the whole process concept is designed with the aim
of reducing variable costs, especially for energy, raw material and chemicals,
at the expense of higher capital investment. The main titanium-made reactor
system, with a significant heat exchange area needed for reducing the direct cost
of vapor compression, represents alone about 44% of the installed equipment
cost. Significant costs are also associated to the process-water purification and
 7.7 conclusions 99

recycling system, needed to nearly eliminate the consumption of chemicals, in

particular of strong acids. Various heat recovery systems also contribute to the
installed equipment cost, to the benefit of a variable cost abatement.
Even considering the high FCI, the most important cost item in the annual
production-cost is raw material, similarly to existing processes. The raw mate-
rial cost depends directly on the cost of pentoses divided by the relatively low
furfural mass yield - 54% in this case -, making pentoses cost the most critical
variable in furfural production. On the other hand, the cost for utilities is signifi-
cantly reduced compared to existing processes, accounting for 11.4% of the final
furfural production cost of 833 $/ton.
It is evident from this analysis that the access to inexpensive pentoses streams
is the most important issue in furfural production using this process.
The abatement of the FCI, and thus of the related cost items, is also critical,
and could be pursued by the integration to an existing biorefinery facility, prefer-
ably non-seasonal, such as pulp mills, and by upscaling to a significantly higher
capacity, in the order of 100 kton/y.
By carefully selecting an appropriate source for raw material, and an optimal
production capacity, the unit cost of production might easily drop in the range
of 650 US$/ton.

7.7 conclusions

Furfural production paradigm should be revised in order to reduce the energy

consumption, increase the yields and better fit within modern biorefineries. The
process described in this chapter is a result of the entire research effort reported
in this dissertation, and it tends towards this aim. In particular it addresses the
following key aspects regarding furfural production:

• The significant reduction of direct costs related to steam (and thus energy)
consumption, leading to improved economics and a reduced environmen-
tal footprint.

• The consideration of the latest findings on the mechanism of furfural for-

mation in dilute acid/halides solutions allowing for optimal yields and
separation.

• The drastic reduction of the specific acid consumption by virtue of the con-
tinuous recirculation of the catalyst-containing solution. The use of suitable
solid acids is also foreseen in the medium term.

• The possibility of feeding the process with a dilute aqueous stream con-
taining pentoses, opening up various opportunities for process integration
within modern biorefineries.
100 a novel process for making furfural

Table 7.3: Summary of annual production cost

Cost item Factor kUS$/y % of total Unit cost

US$/ton
Raw material Pentoses cost = 130 US$/ton 5298 28.9 % 241
Utilities Compressors + 20% for auxiliaries 2090 11.4 % 95
(120 US$/MWhe )
Process water treatment 35% water treatment investment 888 4.8 % 40
Maintenance 4% of FCI 1703 9.3 % 77
Supply 0.6 of FCI 255 1.4 % 12
Royalties 3% cost of production 660 3.6 % 30
Total variable cost 10894 59.4 % 495

Direct labor 45 employee-hours/day 517 2.8 % 23

Supervision 20% of direct labor 103 0.6 % 5
Lab cost 15% of direct labor 77 0.4 % 4
Taxes and insurance 3% of FCI 1277 7.0 % 58
S.A.R.E.a 2.5% of revenue 660 3.6 % 30
Overheads 72% of labor + 2.4% of FCI 1394 7.6 % 63
Interest 8% of FCI 3405 18.6 % 155
Total fixed cost 7433 40.6 % 338

Total annual cost 18327 100 % 833

Annual revenue Furfural unit price 1200 US$ 26406

Gross earnings 8079

a Sales, Administration, Research and Engineering

• Encouraging results from cost analysis, with a predicted cost of production

significantly lower than average furfural exchange prices.

Although encouraging preliminary results have been achieved, several remain-

ing issues are still to be addressed. In particular a more detailed process analysis
is required including several by-products typical of furfural production, such as
5-methyl furfural, acetic and formic acids, and methanol among others. Espe-
cially when considering a woody feedstock the rate of formation of such by-
products may be very important, and by choosing appropriate solutions they
might bring significant benefits to the process, as recently reported by Xing et al.
[72].
 7.7 conclusions 101

Furthermore, only after a more detailed analysis of the process, including a

careful estimation of pressure drops and heat losses, and a detailed analysis of
the process water treatment and recycling, it will be possible to better evalu-
ate many aspects regarding the technical feasibility, materials choice, equipment
sizing - especially regarding heat-exchange areas -, and process conditions opti-
mization based on total cost minimization.
 8
CONCLUDING REMARKS

This final chapter summarizes the conclusions drawn in this dissertation in view of the
scopes of the research stated in chapter 1. Furthermore, suggestions and recommenda-
tions are given for future research .

103
104 concluding remarks

8.1 conclusions

In line with the scope of this research as stated in chapter 1, the main findings
contained in this dissertation may be summarized as follows:
1. By carefully studying the reaction behavior of D-xylose and furfural in the
selected range of conditions, new kinetic models have been proposed, pre-
senting a good agreement with the measured reaction rates. The kinetic
constants have been related to the acid concentration via the ion activity
a H3 O+ to better account for the thermodynamics of the electrolytes in so-
lution, which may be accurately modeled by means of the eNRTL model.
The influence of temperature on the main rate constants showed a clear
agreement with the Arrhenius law, furthermore, from the transition state
theory, and from a careful literature review, a strong indication emerges
of ring opening being involved in the activated state of xylose reaction to
furfural. In this sense Cl− ions were shown to promote the formation of
the 1,2-enediol from the acyclic form of D-xylose, and thus the subsequent
acid catalyzed dehydration to furfural.
2. Significant improvements were observed with respect to the H2SO4-based
case in terms of furfural yield and selectivity from xylose when adding
NaCl to a dilute-acid solution. A 4-fold xylose reaction rate increase was
possible by 1.7 M NaCl addition keeping the acid concentration as low as
50 mM (0.18 wt% HCl), and initial sugar concentration 33.3 mM. General
acid-base catalysis, rather than specific acid catalysis, is involved in the
formation of furfural. In particular, a discrepancy in terms of “catalytic
requirements” becomes evident in the process of furfural formation from
pentoses, which involves an enolization followed by three dehydration re-
actions, two types of reactions favored by basic and acid conditions. Based
on these results the “paradox of furfural yields” may be explained consid-
ering the complex chemistry of furfural formation in the presence of ionic
species other than H+ in aqueous solution. In this view, when thinking at
industrial furfural production, it should be borne in mind that the high
severity normally required is mainly needed to overcome the first enoliza-
tion step, as this is not favored by acid. On the other hand the acid catalyst
is indeed required in the following dehydration steps. Thus, in order to
achieve better furfural yields under milder conditions and in a greener
manner, the combination of two different catalysts, having in turn basic
and acidic character, would be probably the key. Halides were shown to
achieve this aim. In particular they were shown to enhance the reaction of
xylose to furfural in aqueous acidic solution by two distinct and consecu-
tive effects: the formation of the 1,2-enediol from the acyclic form of the
aldose was preferably catalyzed by the halides in the order Cl− >Br− >I− ,
whereas selective dehydration to furfural is assisted by the halides in the
 8.2 recommendations for future research 105

opposite order I− >Br− >Cl− . Synergic effects are also evident when using a
combination of I− and Cl− ions, with selectivity of reaction attaining 95.3%
and furfural yield 87.5%.

3. It was shown that minor concentrations of hydrochloric acid in water can

be effective for hemicellulose hydrolysis already at temperatures between
100 and 120 °C. Recovery of hemicellulose derived carbohydrates, mainly
considering xylose and arabinose, approached 100% under specific condi-
tions. Dilute solutions of FeCl3 were confirmed to be an option for mineral
acid replacement, even though the Fe3+ ions were proved to act only in-
directly on the hydrolysis of hemicellulose via their partial precipitation
to iron oxides and consequent formation of HCl. The residues of pretreat-
ment showed upgraded characteristics with respect to the untreated wheat
straw, presenting improved crystallinity and a significant reduction of al-
kaline minerals with direct consequences on their behavior under thermal
conversion conditions.

4. The process exposed in chapter 7 is a result of the entire effort described

in this research work. It addresses several issues regarding furfural pro-
duction: 1. The significant reduction of direct costs related to steam (and
thus energy) consumption, leading to improved economics and a reduced
environmental footprint. 2. The inclusion of the latest findings on fur-
fural mechanism of formation in dilute acid/halides solutions allowing
for optimal yields and separation. 3. The drastic reduction of the specific
acid consumption by virtue of the continuous recirculation of the catalyst-
containing solution, besides the possibility of using suitable solid catalysts
in the medium term. 4. The possibility of feeding the process with a dilute
aqueous stream containing pentoses, opening up various opportunities for
process integration within modern biorefineries. 5. Interesting economical
outlook as resulting from cost analysis, with a predicted cost of production
of 833 US$/ton, significantly lower than average furfural exchange prices.

8.2 recommendations for future research

The research work described in this dissertation left many open issues of interest
for future research.

furfural salting-out effect As already mentioned in chapter 7, the ad-

dition of salts is expected to enhance furfural separation by salting-out effect.
Future work in this sense is suggested in order to quantify the effect of salts on
VLE of furfural-water-salts system. An attempt was made in this sense during
this research work (results not shown), with the aim of determining the bubble
point of water-salt-furfural systems by means of a Cailletet tube experimental
106 concluding remarks

setup. Regrettably, the results of such experimental campaign were unsatisfac-

tory because of the occurrence of unacceptable systematic errors presumably
due to furfural decomposition during the time required for the measurements.
The release of minor amounts of volatiles consequent to furfural thermal decom-
position gave rise to a slight pressure increase in the Cailletet tubes, the extent
of which made the measured data unusable for VLE purposes. Such decomposi-
tion could not be prevented under the measurement conditions. For this reason
it is recommended to use different methods than Cailletet tubes for VLE mea-
surement involving water-salt-furfural systems, such as appropriate equilibrium
apparatus enabling both vapor and liquid sampling under the target conditions.

sugars and sugar derivatives halide-enhanced dehydration The

effect shown in this work concerning the use of halides salts on xylose dehydra-
tion is of general interest for the dehydration of sugars and also sugars deriva-
tives. In the first place, a similar approach using different halides in aqueous
solution is suggested for HMF production from hexoses. In this sense it would
be interesting to study the different effect on glucose and fructose dehydration.
Based on this same consideration it is also recommended to study the dehydra-
tion of sugar acids, such as gluconic acid, for the production of hydroxymethyl-
furoic acid (HMFA), or even the dehydration of galactaric (glucaric) acid into
2,5-furandicarboxylic acid (FDCA).

raw biomass pretreatment Based on the observations exposed in chap-

ter 6, it will be interesting for future research to test a pretreatment method
involving a dilute-acid treatment of the raw biomass under mild conditions, fol-
lowed by a lignin separation/solubilization step, for the selective production of
cellulose fibers, lignin and hemicellulose-derived carbohydrates. In particular it
is recommended to test an uncatalyzed organosolv type of treatment for the res-
idues of dilute-acid hydrolysis, preferably using ethanol as solvent in the same
range of temperature used in the dilute-acid treatment. By using a further treat-
ment of this kind large fractions of the water-insoluble lignin are expected to be
recovered, leaving a clean, cellulose-rich residue with high crystallinity. Such pre-
hydrolysis and lignin removal could be of particular interest for the integrated
furfural (and lignin) production within cellulose fibers industries.
The thorough removal of alkaline minerals contained in wheat straw afforded
by dilute HCl and FeCl3 may be of interest for biomass gasification. If the ef-
fect of a reduced alkaline minerals content has been already discussed, the ef-
fect of the iron containing micro particles on the biomass residues treated with
dilute FeCl3 on secondary gasification reactions, for instance in fluidized bed
reactors, would be worth further investigation. In fact, iron containing miner-
als, like olivine, have been successfully tested as bed material for their positive
catalytic effects on tars decomposition reactions. Hence, it is recommended to
test dilute-FeCl3 pretreated agricultural residues for gasification, with a partic-
 8.2 recommendations for future research 107

ular attention to slagging, fouling, bed agglomeration, and tar decomposition.

Nevertheless, the opportunity of using iron containing salts for the pretreatment
of biomass as an alternative to mineral acids, should be evaluated in a broader
perspective, carefully balancing economic and environmental benefits and costs.

furfural catalysis Regarding the development of suitable solid catalysts

for the production of furfural (and furan compounds in general) it is recom-
mended to take into account the discussion on the catalytic requirements for
furfural production exposed in chapters 4 and 5. It is important for the catalyst
to enhance the enolization reaction, next to the subsequent dehydrations. The
purely acidic character given by the H+ moiety has often been proven not suffi-
cient to this aim. The simultaneous use of acid and basic solid catalysts is also
not sufficient due to mass transfer limitations and to the instability of the inter-
mediates of reaction. A single catalyst bearing a double function could be an in-
teresting solution, as well as a combination of solid acids and soluble bases/salts
in aqueous solutions. Further research is recommended in this direction.
BIBLIOGRAPHY

[1] OECD-FAO Agricultural outlook 2010-2019. www.agri-outlook.org, 2011.

(Cited on pages xxi, 2, 3, and 27.)

[2] EUCAR, CONCAWE, and JRC. Well-to-wheels analysis of future automo-

tive fuels and powetrains in the European context. Technical report, Euro-
pean Commission Joint Research Centre, March 2007. (Cited on pages xxi,
3, and 4.)

[3] Biosynergy - Biorefinery Development for Europe. Results of the inte-

grated project Biosynergy 2007-2010 - www.biosynergy.eu, 2011. (Cited on
pages xxi and 6.)

[4] A. Corma, S. Iborra, and A. Velty. Chemical routes for the transformation
of biomass into chemicals. Chem. Rev., 107(7):2411–2502, 2007. (Cited on
pages xxi, 4, 5, and 10.)

[5] B. Kamm, P.R. Gruber, and M. Kamm, editors. Biorefineries – Industrial

Processes and Products: Status Quo and Future Directions. Wiley-VCH Verlag
GmbH & Co. KGaA, 2006. (Cited on pages xxi, xxiii, 5, 16, and 18.)

[6] Italian Ministry of Economical Development - Department of En-

ergy - Energy and Mining Statistics and analysis. Available at
http://dgerm.sviluppoeconomico.gov.it. (Cited on pages xxi, 25, and 27.)

[7] K. J. Zeitsch. The chemistry and technology of furfural and its many by-products,
volume 13 of Sugar Series. Elsevier, 2000. (Cited on pages xxii, 13, 16, 17,
19, 20, 21, 22, 23, 26, 40, 42, 46, 50, 84, 85, 86, 88, 96, and 97.)

[8] A. P. Dunlop and F. N. Peters. The Furans. ACS Monograph Series. Rein-
hold Publishing Corporation, New York, 1953. (Cited on pages xxii, xxiii,
10, 11, 14, 16, 17, 19, 20, 22, 23, 40, 46, 50, 52, 58, 89, 92, and 93.)

[9] R.G. Curtis and H.H. Hatt. Equilibria in Furfural-Water systems under
increased pressure and the influence of added salts upon the mutual sol-
ubilities of furfural and water. Australian Journal of Scientific Research, 1:
213–235, 1948. (Cited on pages xxii, 92, and 93.)

[10] J. J. Bozell and G. R. Petersen. Technology development for the production

of biobased products from biorefinery carbohydrates - the US Department
of Energy’s "top 10" revisited. Green Chem., 12:539–554, 2010. (Cited on
pages xxiii, 5, 8, 9, and 10.)

109
110 Bibliography

[11] T. Werpy and G. R. Petersen. Top value added chemicals from biomass
- Volume I: Results of screening of potential candidates from sugars and
synthesis gas. Technical report, US Department of Energy, 2004. (Cited on
pages xxiii, 5, 8, and 10.)

[12] E. E. Hughes and S. F. Acree. Journal of Research of the National Bureau of

Standards, 21:327–336, 1938. (Cited on pages xxiii, 14, 46, and 50.)

[13] W.J.J. Huijgen, H.J. Reith, and H. den Uil. Pretreatment and fractionation
of wheat straw by acetone-based organosolv process. Ind. Eng. Chem. Res.,
49:10132–10140, 2010. (Cited on pages xxiii, 37, and 73.)

[14] Agricultural products as raw materials for industry. Nature, 140:221–222,

1937. (Cited on page 2.)

[15] D. Mitchell. A note on rising food prices. Policy research working paper
4682, The World Bank Development Prospects Group, July 2008. (Cited on
page 3.)

[16] A.E. Farrell, R.J. Plevin, B.T. Turner, A.D. Jones, M. O’Hare, and D.M. Kam-
men. Ethanol can contribute to energy and environmental goals. Science,
311:506–508, 2006. (Cited on page 3.)

[17] R.E.H. Sims, W. Mabee, J.N. Saddler, and M. Taylor. An overview of second
generation biofuel technologies. Bioresour. Technol., 101:1570–1580, 2010.

[18] BioEthanol for sustainable transport - results and recommendations from

the European BEST project. http://www.best-europe.org/, 2009. (Cited
on page 3.)

[19] R. Doornbosch and R. Steenblik. Biofuels: Is the cure worse than the dis-
ease? Round table on sustainable development SG/SD/RT/(2007)3/REV1,
OECD, September 2007. (Cited on page 4.)

[20] M. Patel, M. Crank, V. Dornburg, B. Hermann, L. Roes, B. Husing, L. Over-

beek, F. Terragni, and E. Recchia. The BREW project - medium and long-
term opportunities and risks of the biotechnological production of bulk
chemicals from renewable resources - the potential of white biotechnol-
ogy. Final report, European Commission GROWTH Programme, Septem-
ber 2006. (Cited on page 5.)

[21] L.R. Lynd, C.E. Wyman, and T.U. Gerngross. Biocommodity engineering.
Biotechnol. Prog., 15:777–793, 1999. (Cited on page 5.)

[22] F.W. Lichtenthaler and S. Peters. Carbohydrates as green raw materials for
the chemical industry. Comptes Rendus Chimie, 7:65 – 90, 2004. (Cited on
page 5.)
 Bibliography 111

[23] B. Kamm and M. Kamm. Principles of biorefineries. Appl. Microbiol. Biotech-

nol., 64:137–145, 2004. (Cited on page 5.)

[24] B. Kamm and M. Kamm. Biorefineries - multi product processes. Adv.

biochem. Engin/Biotechnol., 105:175–204, 2007. (Cited on page 5.)

[25] S. Kim and B.E. Dale. Global potential bioethanol production from wasted
crops and crop residues. Biomass Bioenerg., 26:361–375, 2004. (Cited on
page 6.)

[26] I. Lewandowski, J.M.O. Scurlock, E. Lindvall, and M. Christou. The de-

velopment and current status of perennial rhizomatous grasses as energy
crops in the US and Europe. Biomass Bioenerg., 25:335–361, 2003. (Cited on
pages 6 and 7.)

[27] C. Mariani, R. Cabrini, A. Danin, P. Piffanelli, A. Fricano, S. Gomarasca,

M. Dicandilo, F. Grassi, and C. Soave. Origin, diffusion and reproduction
of the giant reed (Arundo donax L.): a promising weedy energy crop. Ann.
Appl. Biol., 157:191–202, 2010. (Cited on page 7.)

[28] V. Dornburg, A. Faaij, P. Verweij, H. Langeveld, G. van de Ven, F. Wester,

H. van Keulen, K. van Diepen, M. Meeusen, M. Banse, J. Ros, D. van Vu-
uren, G.J. van den Born, M. van Oorschot, F. Smout, J. van Vliet, H. Aik-
ing, M. Londo, H. Mozaffarian, and K. Smekens. Assessment of global bio-
mass potentials and their links to food, water, biodiversity, energy demand
and economy. Technical Report WAB500102012, Netherlands Research Pro-
gramme on Scientific Assessment and Policy Analysis for Climate Change
(WAB), 2008. (Cited on page 7.)

[29] M&G news. http://www.gruppomg.com/news.php?newsid=25. (Cited

on page 7.)

[30] D. Chiaramonti. Lignocellulosic ethanol production at industrial scale (BI-

OLYFE). In European Biomass Conference - BIOLYFE workshop, Berlin, June
2011. (Cited on page 7.)

[31] H. E. Hoydonckx, W. M. Van Rhijn, W. Van Rhijn, D. E. De Vos, and P. A. Ja-

cobs. Furfural and derivatives. In Ullmann’s Encyclopedia of Industrial Chem-
istry. Wiley-VCH Verlag GmbH & Co. KGaA, 2000. (Cited on pages 11, 12,
13, 16, 17, 19, 20, 21, 22, and 46.)

[32] A.A. Rosatella, S.P. Simeonov, R.F.M. Frade, and C.A.M. Afonso. 5-
Hydroxymethylfurfural (HMF) as a building block platform: Biological
properties, synthesis and synthetic applications. Green Chem., 13:754–793,
2011. (Cited on page 11.)
112 Bibliography

[33] J. Lewkowski. Synthesis, chemistry and applications of 5-

hydroxymethylfurfural and its derivatives. ARKIVOC, i:17–54, 2001.
(Cited on page 11.)

[34] M. Mascal and E. B. Nikitin. Direct, high yield conversion of cellulose into
biofuel. Angew. Chem. Int. Ed., 47:7924–7926, 2008. (Cited on page 11.)

[35] M. Mascal and E. B. Nikitin. High-yield conversion of plant biomass into

the key value-added feedstocks 5-(hydroxymethyl)furfural, levulinic acid,
and levulinic esters via 5-(chloromethyl)furfural. Green Chem., 12:370–373,
2010. (Cited on page 11.)

[36] G.-J. Gruter and E. de Jong. Furanics: novel fuel options from carbohy-
drates. Biofuels Technol., (1):11–17, 2009. (Cited on pages 12 and 24.)

[37] G.-J. Gruter and F. Dautzenberg. Method for the synthesis of 5-

alkoxymethyl furfural ethers and their use. US Patent Application
20090131690, May 2009. (Cited on pages 12 and 24.)

[38] A.S. Dias, S. Lima, M. Pillinger, and A.A. Valente. Furfural and Furfural-
Based Industrial Chemicals. In Bruno Pignataro, editor, Ideas in Chemistry
and Molecular Sciences, pages 165–186. Wiley-VCH Verlag GmbH & Co.
KGaA, 2010. (Cited on pages 12, 14, and 15.)

[39] E. de Jong. XYX Building blocks: Biorefinery Approach towards Fuels

and Plastic Applications. In World Biofuel Markets, Rotterdam, March 2011.
(Cited on pages 12 and 24.)

[40] F.B. Laforge and G.H. Mains. Furfural from corncobs. Ind. Eng. Chem., 15:
823–829, 1923. (Cited on page 12.)

[41] H.J. Brownlee. Furfural manufacture from oat hulls. Ind. Eng. Chem., 19:
422–424, 1927. (Cited on page 84.)

[42] F.N. Peters. The Furans - Fifteen years of progress. Ind. Eng. Chem., 28:
755–759, 1936.

[43] F.N. Peters. Furan Chemistry. Ind. Eng. Chem., 40:200, 1948.

[44] H.J. Brownlee and C.S. Miner. Industrial development of furfural. Ind. Eng.
Chem., 40:201–204, 1948.

[45] B.H. Wojcik. Catalytic hydrogenation of furan compounds. Ind. Eng. Chem.,
40:210–216, 1948. (Cited on page 22.)

[46] O.W. Cass. Chemical intermediates from furfural. Ind. Eng. Chem., 40:
216–219, 1948.
 Bibliography 113

[47] L.C. Kemp, G.B. Hamilton, and H.H. Gross. Furfural as a selective sol-
vent in petroleum refining. Ind. Eng. Chem., 40:220–227, 1948. (Cited on
page 17.)
[48] S.W. Gloyer. Furans in vegetable oil refining. Ind. Eng. Chem., 40:228–236,
1948.
[49] A.J. Norton. Furan resins. Ind. Eng. Chem., 40:236–238, 1948.
[50] D. L. Williams and A. P. Dunlop. Kinetics of furfural destruction in acidic
aqueous media. Ind. Eng. Chem., 40(2):239–241, 1948. (Cited on pages 12,
40, 41, 44, and 45.)
[51] T. B. Adams, J. Doull, J. I. Goodman, I. C. Munro, P. Newberne, P. S. Por-
toghese, R. L. Smith, B. M. Wagner, C. S. Weil, L. A. Woods, and R. A. Ford.
The FEMA GRAS assessment of furfural used as a flavour ingredient. Food
and Chemical Toxicology, 35:739–751, 1997. (Cited on page 12.)
[52] C.E. Wyman, S.R. Decker, M.E. Himmel, J.W. Brandy, C.E. Skopec, and
L. Viikari. Chapter 43. Hydrolysis of Cellulose and Hemicellulose. In Sev-
erian Dumitriu, editor, Polysaccharides. Structural Diversity and Functional
Versatility, II Edition. CRC Press, 2004. (Cited on pages 13 and 70.)
[53] H.H. Nimz, U. Schmitt, E. Schwab, O. Wittmann, and F. Wolf. Wood. In
Ullmann’s Encyclopedia of Industrial Chemistry. Wiley-VCH Verlag GmbH &
Co. KGaA, 2000. (Cited on page 13.)
[54] A.S. Mamman, J.-M. Lee, Y.-C. Kim, I.T. Hwang, N.-J. Park, Y.K. Hwang,
J.-S. Chang, and J.-S. Hwang. Furfural: Hemicellulose/xylose-derived bio-
chemical. Biofuels, Bioprod. Bioref., 2:438–454, 2008. (Cited on pages 13
and 14.)
[55] W. de Jong and G. Marcotullio. Overview of biorefineries based on co-
production of furfural, existing concepts and novel developments. Int. J.
Chem. React. Eng., 8:A69, 2010. (Cited on pages 13 and 84.)
[56] C. Moreau, R. Durand, D. Peyron, J. Duhamet, and P. Rivalier. Selective
preparation of furfural from xylose over microporous solid acid catalyst.
Ind. Crop. Prod., 7:95–99, 1998. (Cited on page 15.)
[57] R. O’Neill, M.N. Ahmad, L. Vanoye, and F. Aiouache. Kinetics of aqueous
phase dehydration of xylose into furfural catalysed by ZSM-5 zeolite. Ind.
Eng. Chem. Res., 48:4300–4306, 2009. (Cited on page 15.)
[58] J. Lessard, J.-F. Morin, J.-F. Wehrung, D. Magnin, and E. Chornet. High
yield conversion of residual pentoses into furfural via zeolite catalysis and
catalytic hydrogenation of furfural to 2-methylfuran. Top. Catal., 53:1231–
1234, 2010. (Cited on pages 15, 22, and 23.)
114 Bibliography

[59] A.S. Dias, S. Lima, P. Brandão, M. Pillinger, J. Rocha, and A.A. Valente.
Liquid-phase dehydration of D-xylose over microporous and mesoporous
niobium silicates. Catal. Lett., 108:179–186, 2006. (Cited on page 15.)
[60] A.S. Dias, M. Pillinger, and A.A. Valente. Dehydration of xylose into fur-
fural over micro-mesoporous sulfonic acid catalysts. J. Catal., 229:414–423,
2005. (Cited on page 15.)
[61] A.S. Dias, S. Lima, D. Carriazo, V. Rives, M. Pillinger, and A.A. Valente.
Exfoliated titanate, niobate and titanoniobate nanosheets as solid acid cat-
alysts for the liquid-phase dehydration of D-xylose into furfural. J. Catal.,
pages 230–237, 2006. (Cited on page 15.)
[62] Sérgio Lima, Martyn Pillinger, and Anabela A. Valente. Dehydration of
D-xylose into furfural catalysed by solid acids derived from the layered
zeolite Nu-6(1). Catal. Commun., 9:2144–2148, 2008. (Cited on page 15.)
[63] A.S. Dias, S. Lima, M. Pillinger, and A.A. Valente. Acidic cesium salts of
12-tungstophosphoric acid as catalysts for the dehydration of xylose into
furfural. Carbohydr. Res., 341:2946–2953, 2006. (Cited on page 15.)
[64] A.S. Dias, S. Lima, M. Pillinger, and A.A. Valente. Modified versions of
sulfated zirconia as catalysts for the conversion of xylose to furfural. Catal.
Lett., 114:151–160, 2007. (Cited on page 15.)
[65] X. Shi, Y. Wu, P. Li, H. Yi, M. Yang, and G. Wang. Catalytic conversion of
xylose to furfural over the solid acid SO42/ZrO2–Al2O3/SBA-15 catalysts.
Carbohydr. Res., 346:480–487, 2011. (Cited on page 15.)
[66] E. Lam, E. Majid, A.C.W. Leung, J.H. Chong, K.A. Mahmoud, and J.H.T.
Luong. Synthesis of furfural from xylose by heterogeneous and reusable
nafion catalysts. ChemSusChem, 4:535–541, 2011. (Cited on pages 15
and 90.)
[67] A. Takagaki, M. Ohara, S. Nishimura, and K. Ebitani. One-pot formation
of furfural from xylose via isomerization and successive dehydration reac-
tions over heterogeneous acid and base catalysts. Chem. Lett., 39:838–840,
2010. (Cited on pages 15 and 60.)
[68] M.J. Climent, A. Corma, and S. Iborra. Converting carbohydrates to bulk
chemicals and fine chemicals over heterogeneous catalysts. Green Chem.,
13:520–540, 2011. (Cited on page 15.)
[69] F. Trimble and A. P. Dunlop. Recovery of furfural from aqueous solutions.
Ind. Eng. Chem., 12:721–722, 1940. (Cited on page 16.)
[70] J.R. Croker and R.G. Bowrey. Liquid extraction of furfural from aqueous
solution. Ind. Eng. Chem. Fundam., 23:480–484, 1984.
 Bibliography 115

[71] J.L. Cabezas and L.A. Bárcena. Extraction of furfural from aqueous solu-
tions using alcohols. J. Chem. Eng. Data, 33:435–437, 1988.
[72] R. Xing, W. Qi, and G.W. Huber. Production of furfural and carbohylic
acids from waste aqueous hemicellulose solutions from the pulp and paper
and cellulosic ethanol industries. Energy Environ. Sci., 4:2193–2205, 2011.
(Cited on pages 16, 92, 98, and 100.)
[73] T. Sako, T. Sugeta, N. Nakazawa, T. Okubo, M. Sato, T. Taguchi, and T. Hi-
aki. Phase equilibrium study of extraction and concentration of furfural
produced in reactor using supercritical carbon dioxide. J. Chem. Eng. Jpn,
24:449–455, 1991. (Cited on page 16.)
[74] T. Sako, T. Sugeta, N. Nakazawa, T. Okubo, M. Sato, T. Taguchi, and T. Hi-
aki. Kinetic study of furfural formation accompanying supercritical carbon
dioxide extraction. J. Chem. Eng. Jpn., 25:372–377, 1992.
[75] T. Gamse, R. Marr, F. Fröshi, and M. Siebenhofer. Extraction of furfural
with carbon dioxide. Separ. Sci. Tech., 32:355–371, 1997.
[76] W. Sangarunlert, P. Piumsomboon, and S. Ngamprasertsith. Furfural pro-
duction by acid hydrolysis and supercritical carbon dioxide extraction
from rice husk. Korean J. Chem. Eng., 24:936–941, 2007. (Cited on page 16.)
[77] S. Abad, J.L. Alonso, V. Santos, and J.C. Parajó. Furfural from wood in
catalyzed acetic acid media: A mathematical assessment. Bioresour. Technol.,
62:115–122, 1997. (Cited on page 16.)
[78] R. Lehnen, B. Saake, and H.H. Nimz. Furfural and hydroxymethylfurfural
as by-products of FORMACELL pulping. Holzforschung, 55:199–204, 2001.
(Cited on page 16.)
[79] J.N. Chheda, Y. Román-Leshkov, and J.A. Dumesic. Production of 5-
hydroxymethylfurfural and furfural by dehydration of biomass-derived
mono- and poly-saccharides. Green Chem., 9:342–350, 2007. (Cited on
page 16.)
[80] J. B. Binder, J.J. Blank, A.V. Cefali, and R. T. Raines. Synthesis of fur-
fural from xylose and xylan. ChemSusChem, 3:1268–1272, 2010. (Cited on
page 16.)
[81] C. Usuki, Y. Kimura, and S. Adachi. Degradation of pentoses and hex-
ouronic acids in subcritical water. Chem. Eng. Technol., 31:133–137, 2008.
(Cited on page 16.)
[82] S.B. Kim, M.R. Lee, E.D. Park, S.M. Lee, H.K. Lee, K.H. Park, and M.-J.
Park. Kinetic study of the dehydration of D-xylose in high temperature
water. Reac Kinet Mech Cat, 103:267–277, 2011. (Cited on page 16.)
116 Bibliography

[83] R. Weingarten, J. Cho, W.C. Conner, and G.W. Huber. Kinetics of furfural
production by dehydration of xylose in a biphasic reactor with microwave
heating. Green Chem., 12:1423–1429, 2010. (Cited on page 16.)
[84] O. Yemiş and G. Mazza. Acid-catalyzed conversion of xylose, xylan and
straw into furfural by microwave-assisted reaction. Bioresour. Technol., 102:
7371–7378, 2011. (Cited on page 16.)
[85] B. Sain, A. Chaudhuri, J.N. Borgohain, B.P. Baruah, and J.L. Ghose. Fur-
fural and Furfural-Based Industrial Chemicals. J. Sci. Ind. Res., 41:431–438,
1982. (Cited on pages 16 and 17.)
[86] J.P. Trickey. Certain solvent properties of furfural and its derivatives. Ind.
Eng. Chem., 19:643–644, 1927. (Cited on page 17.)
[87] A. De Lucas, L. Rodríguez, P. Sánchez, and A. Carnicer. Extraction of
aromatic compounds from heavy neutral distillate lubricating oils by using
furfural. Separ. Sci. Tech., 28:2465–2477, 1993. (Cited on page 17.)
[88] D.F. Aycock. Solvent applications of 2-Methyltetrahydrofuran in
organometallic and biphasic reactions. Org. Process Res. Dev., 11:156–159,
2007. (Cited on page 19.)
[89] H. Müller. Tetrahydrofuran. In Ullmann’s Encyclopedia of Industrial Chem-
istry. Wiley-VCH Verlag GmbH & Co. KGaA, 2005. (Cited on page 20.)
[90] DalinYebo Trading & Development Pty Ltd,
www.dalinyebo.co.za/dalinyebo-trading-development-pty-ltd. Personal
communication, 2011. (Cited on pages 20, 24, 87, 97, and 98.)
[91] G. Collin, R. Mildenberg, M. Zander, H. Höke, W. McKillip, W. Freitag, and
W. Imöhl. Resins, synthetic. In Ullmann’s Encyclopedia of Industrial Chem-
istry. Wiley-VCH Verlag GmbH & Co. KGaA, 2000. (Cited on page 20.)
[92] M. Chours, M. N. Belgacem, and A. Gandini. Acid-catalyzed polycon-
densation of furfuryl alcohol: mechanisms of chromophore formation and
cross-linking. Macromolecules, 29:3839–3850, 1996. (Cited on pages 20
and 21.)
[93] S. Lande, M. Westin, and M. Schneider. Development of modified wood
products based on furan chemistry. Mol. Cryst. Liq. Cryst., 484:367–378,
2008. (Cited on page 21.)
[94] S. Lande, M. Eikenes, M. Westin, and M. Schneider. Furfurylation of wood:
Chemistry, properties, and commercialization. In T.P. Schultz, M. Militz,
M.H. Freeman, B. Goodell, and D.D. Nicholas, editors, Development of Com-
mercial Wood Preservatives, number 982 in ASC symposium series, pages
337–355. American Chemical Society, 2008. (Cited on page 21.)
 Bibliography 117

[95] S. Lande, O. Høibø, and E. Larnøy. Variation in treatability of scots pine

(pinus sylvestris) by the chemical modification agent furfuryl alcohol dis-
solved in water. Wood Sci. Technol., 44:105–118, 2010. (Cited on page 21.)

[96] A. Gandini. The irruption of polymers from renewable resources on the

scene of macromolecular science and technology. Green Chem., 13:1061–
1083, 2011. (Cited on page 21.)

[97] Food and Agriculture Organization Statistics - FAOSTAT. Available at

http://faostat.fao.org. (Cited on page 21.)

[98] A. Gandini and M. N. Belgacem. Furans in polymer chemistry. Prog. Polym.

Sci., 22:1203–1379, 1997. (Cited on pages 21 and 41.)

[99] C. Moreau, M. N. Belgacem, and A. Gandini. Recent catalytic advances in

the chemistry of substituted furans from carbohydrates and in the ensuing
polymers. Top. Catal., 27:11–30, 2004. (Cited on page 21.)

[100] R. Rodríguez-Kábana, J.W. Kloepper, C.F. Weaver, and D.G. Robertson.

Control of plant-parasitic nematodes with furfural- A naturally occurring
fumigant. Nematropica, 23:63–73, 1993. (Cited on page 21.)

[101] G.W. Huber, S. Iborra, and A. Corma. Synthesis of transportation fuels

from biomass: Chemistry, catalysts, and engineering. Chem. Rev., 106:4044–
4098, 2006. (Cited on page 22.)

[102] L. Petrus and M.A. Noordermeer. Biomass to biofuels, a chemical perspec-

tive. Green Chem., 8:861–867, 2006. (Cited on page 22.)

[103] A.J. Ragauskas, C.K. Williams, B.H. Davison, G. Britovsek, J. Cairney, C.A.
Eckert, W.J.Jr Frederick, J.P. Hallett, D.J. Leak, C.L. Liotta, J.R. Mielenz,
R. Murphy, R. Templer, and T. Tschaplinski. The path forward for biofuels
and biomaterials. Science, 311:484–489, 2006.

[104] Yuriy Román-Leshkov, Christopher J. Barrett, Zhen Y. Liu, and J.A.A.

Dumesic. Production of dimethylfuran for liquid fuels from biomass-
derived carbohydrates. Nature, 447:982–986, 2007. (Cited on pages 22
and 23.)

[105] S. Bayan and E. Beati. Il furfurolo ed i suoi derivati come carburanti. Cim.
Ind. (in Italian), 23:432–434, 1941. (Cited on page 23.)

[106] Y Kar and H Deveci. Importance of P-series fuels for flexible-fuel vehicles
(FFVs) and alternative fuels. Energ. Source., 28(9-12):909–921, 2006. (Cited
on page 23.)

[107] S.F. Paul. Alternative fuel. US Patent 5,697,987, 1997.

118 Bibliography

[108] S.F. Paul. Alternative fuel. US Patent 6,712,866, 2004.

[109] Alternative Fuel Transportation Program; P-series Fuels; Final Rule. Fed-
eral register Vol. 64, No. 94, US Department of Energy, May 17 1999. (Cited
on page 23.)

[110] S.F. Paul. An optimized alternative motor fuel formulation: natural gas
liquids, ethanol, and a biomass-derived ether. In ACS Division of Fuel Chem-
istry, volume 43, pages 373–377, Boston, August 1998. (Cited on page 23.)

[111] T.W. Rudolph and J.J. Thomas. NOx, NMHC and CO emissions from
biomass derived gasoline extenders. Biomass, 16:33–49, 1988. (Cited on
page 23.)

[112] N. I. Uryanskaya, N. A. Surovtsev, V. T. Vasilenko, P. A. Mikheichev, and

E. Ya. Yunitskaya. Surface activity of anti-icing additives. Chemistry and
Technology of Fuels and Oils, 25:162–167, 1989. ISSN 0009-3092. (Cited on
page 23.)

[113] B. N. Klopov and R. N. Plakhova. Rapid determination of content of anti-

icing additives in jet fuels under airfield conditions. Chemistry and Technol-
ogy of Fuels and Oils, 15:68–70, 1979. (Cited on page 23.)

[114] G.W. Huber, J.N. Chheda, C. J. Barrett, and J.A. Dumesic. Production of
liquid alkanes by aqueous-phase processing of biomass-derived carbohy-
drates. Science, 308:1446–1450, 2005. (Cited on page 23.)

[115] J.N. Chheda and J.A. Dumesic. An overview of dehydration, aldol-

condensation and hydrogenation processes for production of liquid alka-
nes from biomass-derived carbohydrates. Catal. Today, 123:59–70, 2007.

[116] J.N. Chheda, G.W. Huber, and J.A. Dumesic. Liquid-phase catalytic pro-
cessing of biomass-derived oxygenated hydrocarbons to fuels and chemi-
cals. Angew. Chem. Int. Ed., 46:7164–7183, 2007.

[117] E.L. Kunkes, D. Simonetti, R.M. West, J.C. Serrano-Ruiz, C.A. Gärtner, and
J.A.A. Dumesic. Catalytic conversion of biomass to monofunctional hydro-
carbons and targeted liquid-fuel classes. Science, 322:417–421, 2008.

[118] R. Xing, A.V. Subrahmanyam, H. Oclay, W. Qi, G.P. van Walsum,

H. Pendse, and G.W. Huber. Production of jet and diesel fuel range alka-
nes from waste hemicellulose-derived aqueous solutions. Green Chem., 12:
1933–1946, 2010. (Cited on page 23.)

[119] A. Corma, O. de la Torre, M. Renz, and N. Villandier. Production of high-

quality diesel from biomass waste products. Angew. Chem. Int. Ed., 50:1–5,
2011. (Cited on page 23.)
 Bibliography 119

[120] A. P. Dunlop. Furfural formation and behavior. Ind. Eng. Chem., 40(4):
204–209, 1948. (Cited on pages 26, 40, and 41.)

[121] E. R. Garret and B. H. Dvorchik. Kinetics and mechanisms of acid degrada-

tion of the aldopentoses to furfural. J. Pharm. Sci., 58:813–820, 1969. (Cited
on pages 45 and 54.)

[122] D. F. Root, J. F. Saeman, J. F. Harris, and W. K. Neill. Kinetics of the acid

catalyzed conversion of xylose to furfural. Forest Prod. J., 9:158–165, 1959.
(Cited on pages 26, 40, and 46.)

[123] K. R. Westerterp, W. P. M. Van Swaaij, and A. A. C. M. Beenackers. Chemical

reactor design and operation. John Wiley and Sons, second edition, 1987.
(Cited on page 32.)

[124] E. W. Lemmon, M. O. McLinden, and M. L. Huber. Refprop - reference

fluid properties software. NIST, 2002. (Cited on page 34.)

[125] P. Colonna and T. Van der Stelt. Fluidprop: A program for the estimation
of thermophysical properties of fluids. Energy Technology Section - Delft
University of Technology, The Netherlands - http://fluidprop.tudelft.nl/,
2005. (Cited on page 34.)

[126] J. Giuntoli, W. de Jong, S. Arvelakis, H. Spliethoff, and A. H. M. Verkooijen.

Quantitative and kinetic (TG-FTIR) study of biomass residue pyrolysis:
Dry distiller’s grains with solubles (DDGS) and chicken manure. J. Anal.
Appl. Pyrolysis, 85:301–312, 2009. (Cited on pages 37 and 78.)

[127] R. Kumar, G. Mago, V. Balan, and C.E. Wyman. Physical and chemical
characterizations of corn stover and poplar solids resulting from leading
pretreatment technologies. Bioresour. Technol., 100(17):3948 – 3962, 2009.
(Cited on pages 38 and 76.)

[128] J.F. Seaman. Kinetics of wood saccharification - hydrolysis of cellulose

and decomposition of sugars in dilute acid at high temperature. Ind. Eng.
Chem., 37:43–52, 1945. (Cited on page 40.)

[129] W. A. Bonner and M. R. Roth. The conversion of D-Xylose-1-C14 into

2-Furaldehyde-α-C14 . J. Am. Chem. Soc., 81(5454-5456), 1959. (Cited on
page 56.)

[130] M. S. Feather. The Reductic acid-14 C derived from D-xylose-1-14 C and 2-

furaldehyde-α-14 C. J. Org. Chem., 34(6):1998–1999, 1969. (Cited on page 44.)

[131] M. S. Feather, D. W. Harris, and S. B. Nichols. Routes of convertion of

D-Xylose, Hexuronic acids, and L-Ascorbic acid to 2-Furaldehyde. J. Org.
Chem., 37(10):1606–1608, 1972. (Cited on page 54.)
120 Bibliography

[132] D. W. Harris and M. S. Feather. Evidence for a C-2-C-1 intramolecular

hydrogen-transfer during the acid-catalyzed isomerization of D-glucose
to D-fructose. Carbohydr. Res., 30:359–365, 1973. (Cited on pages 56, 59,
and 65.)

[133] M. J. Antal Jr, T. Leesomboon, W. S. Mok, and G. N. Richards. Mechanism

of formation of 2-furaldehyde from D-xylose. Carbohydr. Res., 217:71–85,
1991. (Cited on pages 40, 42, 44, and 56.)

[134] P.J. Oefner, A.H. Lanziner, G. Bonn, and O. Bobleter. Quantitative stud-
ies on furfural and organic acid formation during hydrothermal, acidic
and alkaline degradation of D-xylose. Monatshefte für Chimie, 123:547–556,
1992.

[135] T. Ahmad, L. Kenne, K. Olsson, and O. Theander. The formation of 2-

furaldehyde and formic acid from pentoses in slightly acidic deuterium
oxide studied by 1 H NMR spectroscopy. Carbohydr. Res., 276:309–320, 1995.
(Cited on pages 54, 56, 57, 65, and 67.)

[136] C. Liu and C.E. Wyman. The enhancement of xylose monomer and xy-
lotriose degradation by inorganic salts in aqueous solutions at 180 ◦ C. Car-
bohydr. Res., 341:2550–2556, 2006. (Cited on pages 50 and 71.)

[137] R. M. Nimlos, X. Quian, M. Davis, M. E. Himmel, and D. K. Johnson. En-

ergetics of xylose decomposition as determined using quantum mechanics
modeling. J. Phys. Chem. A, 110(42):11824–11838, 2006. (Cited on pages 40
and 56.)

[138] A. G. Dickinson, D. J. Wesolowsky, D. A. Palmer, and R. E. Mesmer. Dis-

sociation constant of bisulfate ion in aqueous sodium chloride solutions to
250 ◦ C. J. Phys. Chem., 94:7978–7985, 1990. (Cited on pages 42, 43, and 51.)

[139] W. L. Marshall and E. V. Jones. Second dissociation constant of sulfuric

acid from 25 to 350 ◦ C evaluated from solubilities of calcium sulfate in
sulfuric acid solution. J. Phys. Chem., 70(12):4928–4040, 1966. (Cited on
page 51.)

[140] A. S. Quist, W. L. Marshall, and H. R. Jolley. Electrical conductances of

aqueous solutions at high temperature and pressure. II. the conductances
and ionization constants of sulfuric acid-water solutions from 0 to 800 ◦ C
and pressures up to 4000 bars. J. Phys. Chem., 69(8):2726–2735, 1965. (Cited
on page 42.)

[141] L. P. Hammett and M. A. Paul. The relation between the rates of some
acid catalyzed reactions and the acidity function, H◦ . J. Am. Chem. Soc., 56
(4):830–832, 1934. (Cited on page 42.)
 Bibliography 121

[142] L. Arnaut, S. Formosinho, and H. Burrows. Chemical Kinetics - From molec-

ular structure to chemical reactivity. Elsevier, first edition, 2007. (Cited on
page 42.)

[143] W.J. Moore. Physical Chemistry. Longman Group Ltd., fifth edition, 1974.
(Cited on pages 42, 43, and 52.)

[144] J. Haubrock, J.A. Hogendoorn, and G.F. Versteeg. The applicability of

activities in kinetic expressions: a more fundamental approach to represent
the kinetics of the system CO2 -OH- in terms of activities. Int. J. Chem. React.
Eng., 3:article A40, 2005. (Cited on page 43.)

[145] A. Haghtalab, V. G. Papangelakis, and X. Zhu. The electrolyte NRTL model

and speciation approach as applied to multicomponent aqueous solutions
of H2 SO4 , Fe2 (SO4 )3 , MgSO4 and Al(SO4 )3 at 230-270 ◦ C. Fluid Phase Equi-
libr., 220(2):199–209, 2004. (Cited on page 43.)

[146] E. I. Fulmer, L. M. Christensen, R. M. Hixon, and R. L. Foster. The pro-

duction of furfural from xylose solutions by means of hydrochloric acid-
sodium chloride systems. J. Phys. Chem., 40(133), 1936. (Cited on pages 50
and 52.)

[147] L. Liu, J. Sun, M. Li, S. Wang, H. Pei, and J. Zhang. Enhanced enzymatic
hydrolysis and structural features of corn stover by FeCl3 pretreatment.
Bioresour. Technol., 100:5853–5858, 2009. (Cited on pages 50 and 71.)

[148] L. Liu, J. Sun, C. Cai, S. Wang, H. Pei, and J. Zhang. Corn stover pre-
treatment by inorganic salts and its effects on hemicellulose and cellulose
degradation. Bioresour. Technol., 100:5865–5871, 2009. (Cited on pages 50,
53, and 72.)

[149] J. Gravitis, N. Vedernikov, J. Zandersons, and A. Kokorevics. Furfural

and levoglucosan prduction from deciduous wood and agricoltural wastes.
ACS symposium series, 784:110–122, 2001. (Cited on page 50.)

[150] R. Xing, S. Liu, H. Yu, Z. Guo, P. Wang, C. Li, Z. Li, and P. Li. Salt-assisted
acid hydrolysis of chitosan to oligomers under microwave irradiation. Car-
bohydr. Res., 340:2150–2153, 2005. (Cited on page 50.)

[151] L. A. Torry, R. Kaminsky, M. T. Klein, and M. R. Klotz. The effect of

salts on hydrolysis in supercritical and near-critical water: Reactivity and
availability. J. Supecrit. Fluids, 5(3):163–168, 1992. (Cited on page 50.)

[152] N. Shimada, H. Kawamoto, and S. Saka. Solid-state hydrolysis of cellu-

lose and methyl α- and β-D-glucopyranosides in presence of magnesium
chloride. Carbohydr. Res., 342:1373–1377, 2007. (Cited on page 50.)
122 Bibliography

[153] A. S. Amarasekara and C. C. Ebede. Zinc chloride madiated degradation

of cellulose at 200◦ C and identification of the products. Bioresour. Technol.,
100:5301–5304, 2009. (Cited on page 50.)
[154] H. Zhao, J. E. Holladay, H. Brown, and Z. C. Zhang. Metal chlorides in
ionic liquid solvents convert sugars to 5-Hydroxymethylfurfural. Science,
316:1597–1600, 2007. (Cited on pages 50 and 62.)
[155] J. Prausnitz and J. Targovnik. Salt effects in aqueous vapor-liquid equilib-
ria. Ind. Eng. Chem. Chem. Eng. Data Series, 3(2):234–239, 1958. (Cited on
page 53.)
[156] W. Kunz, J. Henle, and B.W. Ninham. ’Zur Lehre von der Wirkung der
Salze’ (about the science of the effect of salts): Franz Hofmeister’s historical
papers. Curr. Opin. Colloid Interface Sci., 9:19–37, 2004. (Cited on page 53.)
[157] J. Wu, A. S. Serriani, and T. Vuorinen. Furanose ring anomerization: ki-
netic and thermodynamic studies of the D-2-pentuloses by 13 C-n.m.r. spec-
troscopy. Carbohydr. Res., 206:1–12, 1990. (Cited on page 56.)
[158] K. N. Drew, J. Zajicek, G. Bondo, B. Bose, and A. S. Serriani. 13 C-labeled
aldopentoses: detection and quantitation of cyclic and acyclic forms by
heteronuclear 1D and 2D NMR spectroscopy. Carbohydr. Res., 307:199–209,
1998. (Cited on page 56.)
[159] M. Watanabe, Y. Aizawa, T. Iida, T. M. Aida, C. Levy, K. Sue, and H. In-
omata. Glucose reactions with acid and base catalysts in hot compressed
water at 473 K. Carbohydr. Res., 340:1925–1930, 2005. (Cited on pages 56,
57, and 59.)
[160] S. J. Eitelman and D. Horton. Studies on enolization of aldehydo-aldose
derivatives. Carbohydr. Res., 341:2658–2668, 2006. (Cited on page 59.)
[161] C. D. Hurd and L. L. Isenhour. Pentose reactions. I. furfural formation. J.
Am. Chem. Soc., 54:317–330, 1932. (Cited on page 60.)
[162] A. Takagaki, M. Ohara, S. Nishimura, and K. Ebitani. A one-pot reac-
tion for biorefinery:combination of solid acid and base catalysts for direct
production of 5-hydroxymethylfurfural from saccharides. Chem. Commun.,
pages 6276–6278, 2009. (Cited on pages 60 and 62.)
[163] J. B. Binder and R. T. Raines. Simple chemical transformation of lignocel-
lulosic biomass into furans for fuels and chemicals. J. Am. Chem. Soc., 131:
1979–1985, 2009. (Cited on page 62.)
[164] E. A. Pidko, V. Degirmenci, R. A. van Santen, and Hensen E. J. M. Glucose
activation by transient Cr2+ dimers. Angew. Chem. Int. Ed., 49:2530–2534,
2010. (Cited on page 62.)
 Bibliography 123

[165] G. Marcotullio and W. de Jong. Chloride ions enhance furfural formation

from D-xylose in dilute aqueous acidic solutions. Green Chem., 12:1739–
1746, 2010. (Cited on pages 65, 67, and 71.)
[166] A. Wisniewski, J. Szafranek, and J. Sokolowski. Isomerization during dehy-
dration of pentitols in acid media. Carbohydr. Res., 97:229–234, 1981. (Cited
on page 68.)
[167] A. Wisniewski, J. Sokolowski, and J. Szafranek. Products from the action
of hydrochloric acid on pentitols. J. Carbohyd. Chem., 2(3):293–304, 1983.
[168] M. Kurszewska, E. Skorupowa, J. Madaj, and A. Wisniewski. Solvent-free
thermal dehydration of pentitols on zeolites. J. Carbohyd. Chem., 23(4):169–
177, 2004. (Cited on page 68.)
[169] F. Fenouillot, A. Rousseau, G. Colomines, R. Saint-Loup, and J.-P. Pascault.
Polymers from renewable 1,4:3,6-dianhydrohexitols (isosorbide, isoman-
nide and isoidide): A review. Prog. Polym. Sci., 35:578–622, 2010. (Cited on
page 68.)
[170] G. Flèche and M. Huchette. Isosorbide. Preparation, properties and chem-
istry. Starch/Stärke, 1:26–30, 1986. (Cited on page 68.)
[171] R. Aguilar, J.A. Ramírez, G. Garrote, and M. Vázquez. Kinetic study of
the acid hydrolysis of sugar cane bagasse. J. Food Eng., 55:309–318, 2002.
(Cited on pages 70 and 73.)
[172] T. Carrasco and C. Roy. Kinetic study of dilute-acid prehydrolysis of
xylan-containing biomass. Wood Sci. Technol., 26:189–208, 1992. (Cited on
page 70.)
[173] P.R. Fields and R.J. Wilson. Process for the production of xylose. US Patent
5,340,403, 1994. (Cited on page 70.)
[174] H. Zhang, X. Zhao, X. Ding, H. Lei, X. Chen, D. An, Y. Li, and W. Zichen.
A study on the consecutive preparation of D-xylose and pure superfine
silica from rice husk. Bioresour. Technol., 101:1263–1267, 2010. (Cited on
page 70.)
[175] Y. Sun and J. Cheng. Hydrolysis of lignocellulosic materials for ethanol
production: a review. Bioresour. Technol., 83:1–11, 2002. (Cited on page 70.)
[176] B. Yang and C.E. Wyman. Pretreatment: the key to unlocking low-cost
cellulosic ethanol. Biofuels, Bioprod. Bioref., 2:26–40, 2008.
[177] N. Mosier, C.E. Wyman, B.E. Dale, R.T. Elander, Y.Y. Lee, M. Holtzapple,
and M. Ladisch. Features of promising technologies for pretreatment of
lignocellulosic biomass. Bioresour. Technol., 96(6):673–686, 2005. (Cited on
page 70.)
124 Bibliography

[178] C.E. Wyman, B.E. Dale, R.T. Elander, M. Holtzapple, M. Ladisch, and Y.Y.
Lee. Comparative sugar recovery data from laboratory scale application
of leading pretreatment technologies to corn stover. Bioresour. Technol., 96:
2026 – 2032, 2005. ISSN 0960-8524. doi: DOI:10.1016/j.biortech.2005.01.018.
(Cited on page 70.)
[179] T. Eggeman and R.T. Elander. Process and economic analysis of pretreat-
ment technologies. Bioresour. Technol., 96:2019 – 2025, 2005. (Cited on
page 70.)
[180] I.C. Roberto, S.I. Mussatto, and R.C.L.B. Rodrigues. Dilute-acid hydrolysis
for optimization of xylose recovery from rice straw in semi-pilot reactor.
Ind. Crop. Prod., 17:171–176, 2003. (Cited on page 70.)
[181] Y. Chen, B. Dong, W. Qin, and D. Xiao. Xylose and cellulose fractionation
from corncob with three different strategies and separate fermentation of
them to bioethanol. Bioresour. Technol., 101:7005–7010, 2010.
[182] A. Herrera, S.J. Téllez-Luis, J.A. Ramírez, and M. Vázquez. Production
of xylose from sorghum straw using hydrochloric acid. J. Cereal Sci., 37:
267–274, 2003. (Cited on pages 73 and 74.)
[183] G. González, J. López-Santín, G. Caminal, and C. Solà. Dilute acid hydrol-
ysis of wheat straw hemicellulose at moderate temperature: A simplified
kinetic model. Biotechnol. Bioeng., XXVIII:288–293, 1986. (Cited on page 70.)
[184] F.A. Cotton and G. Wilkinson. Advanced Inorganic Chemistry. Wiley-
Interscience, fifth edition, 1988. (Cited on page 71.)
[185] J. Giuntoli, S. Arvelakis, H. Spliethoff, W. de Jong, and A. H. M. Verkooi-
jen. Quantitative and kinetic thermogravimetric fourier transform infrared
(TG-FTIR) study of pyrolysis of agricultural residues: Influence of differ-
ent pretreatments. Energ. Fuel, 23:5695–5706, 2009. (Cited on pages 71, 72,
and 78.)
[186] X. Wang, J. Si, H. Tan, L. Ma, M. Pourkashanian, and T. Xu. Nitrogen,
sulfur, and chlorine transformations during the pyrolysis of straw. Energ.
Fuel, 24:5215–5221, 2010. (Cited on page 71.)
[187] S. Arvelakis and E.G. Koukios. Physicochemical upgrading of
agroresidues as feedstocks for energy production via thermochemical con-
version methods. Biomass Bioenerg., 22:331–348, 2002. (Cited on pages 72
and 81.)
[188] T.A. Lloyd and C.E. Wyman. Combined sugar yields for dilute sulfu-
ric acid pretreatment of corn stover followed by enzymatic hydrolysis of
the remaining solids. Bioresour. Technol., 96:1967 – 1977, 2005. (Cited on
page 74.)
 Bibliography 125

[189] B. Voigt and A. Göbler. Formation of pure haematite by hydrolysis of iron

(III) salt solutions under hydrothermal conditions. Cryst. Res. Technol., 21
(9):1177–1183, 1986. (Cited on page 76.)

[190] Anker Jensen, Kim Dam-Johansen, Marek A. Wójtowicz, and Michael A.

Serio. TG-FTIR study of the influence of potassium chloride on wheat
straw pyrolysis. Energ. Fuel, 12(5):929–938, 1998. (Cited on page 78.)

[191] K. Ravendraan, A. Ganesh, and K.C. Khilar. Influence of mineral matter

on biomass pyrolysis characteristics. Fuel, 74:1812–1822, 1995. (Cited on
page 78.)

[192] G. Varhegyi, M. J. Antal Jr, T. Szekely, F. Till, and E. Jakab. Simultane-

ous thermogravimetric-mass spectrometric studies of the thermal decom-
position of biopolymers. 1. avicel cellulose in the presence and absence of
catalysts. Energ. Fuel, 2(3):267–272, 1988. (Cited on pages 78 and 81.)

[193] S. Arvelakis, P. Vourliotis, E. Kakaras, and E.G. Koukios. Effect of leaching

on the ash behavior of wheat straw and olive residue during fluidized bed
combustion. Biomass Bioenerg., 20:459–470, 2001. (Cited on page 81.)

[194] M. Zevenhoven-Onderwater, R. Backman, B.J. Skrivfars, and M. Hupa. The

ash chemistry in fluidised bed gasification of biomass fuels. Part II: Ash
behavior prediction versus bench scale agglomeration tests. Fuel, 80:1489–
1502, 2001. (Cited on page 81.)

[195] S. Rapagnà, N. Jand, A. Kiennemann, and P.U. Foscolo. Steam-gasification

of biomass in a fluidised-bed of olivine particles. Biomass Bioenerg., 19:
187–197, 2000. (Cited on page 81.)

[196] E.J. Pearce and J.A. Gerster. Furfural-water system. experimental and the-
oretical vapor-liquid relationships. Ind. Eng. Chem., 42:1418–1424, 1950.
(Cited on page 92.)

[197] L. Fele and V. Grilic. Separation of furfural from ternary mixtures. J. Chem.
Eng. Data, 48:564–570, 2003.

[198] M.S. Sunder and D.H.L. Prasad. Phase equilibria of Furfural + Water and
Dichloromethane + n-Hexane. J. Chem. Eng. Data, 48:221–223, 2003.

[199] Aspen Plus Databank Version 7.0. Aspen Technology Inc., 2007. (Cited on
page 92.)

[200] W.V. Evans and M.B. Aylesworth. Some critical costants of furfural. Ind.
Eng. Chem., 18:24–27, 1926. (Cited on page 92.)
126 Bibliography

[201] M. S. Peters and K.D. Timmerhaus. Plant design and economics for chemical
engineers, fourth edition. McGraw-Hill International Editions, 1991. (Cited
on pages 97 and 98.)

[202] J.R. Couper. Process engineering economics. CRC Press, 2003. (Cited on
pages 97 and 98.)
gianluca marcotullio

Born on 15 July 1979, in L’Aquila, Italy. He obtained his MSc degree in mechani-
cal engineering at the University of L’Aquila in 2004. During his master studies
he has been exchange student at Instituto Superior Técnico (Lisbon), and guest
student at ENEA Casaccia Research Centre (Rome). He followed a post-graduate
master program on energy and utilities management between 2005 and 2006. He
started his doctoral research at the Process and Energy Department of Delft Uni-
versity of Technology on December 2006 in the field of biorefinery. Between 2006
and 2009 he has been Marie Curie fellow within the EST program INECSE, and
worked between 2006 and 2010 in the EU FP6 research project Biosynergy. He is
presently researcher and design engineer at SEA Servizi Energia e Ambiente srl,
Italy, and associate consultant at DalinYebo, South Africa.
peer-reviewed publications

• Marcotullio, G. and De Jong, W. Furfural formation from D-xylose: the use

of different halides in dilute aqueous acidic solutions allows for exception-
ally high yields. Carbohydr. Res., 346: 1291-1293, 2011.
• Marcotullio G.; Krisanti E.; Giuntoli J.; de Jong W. Selective production of
hemicellulose-derived carbohydrates from wheat straw using dilute HCl
or FeCl3 solutions under mild conditions. X-ray and thermo-gravimetric
analysis of the solid residues. Biores. Tech. 102: 5917-5923, 2011.
• Marcotullio, G. and De Jong, W. Chloride ions enhance furfural formation
from D-xylose in dilute aqueous acidic solutions, Green Chem., 12: 1739-
1746, 2010.
• De Jong, W. and Marcotullio G. Overview of biorefineries based on co-
production of furfural, existing concepts and novel developments. Int. J.
Chem. React. Eng., 8:A69, 2010.
• Marcotullio, G.; Cardoso, M.A.T.; De Jong, W. and Verkooijen, A.H.M. Bioen-
ergy II: Furfural Destruction Kinetics during Sulphuric Acid-Catalyzed
Production from Biomass. Int. J. Chem. React. Eng., Vol. 7: A67, 2009.
• Marcotullio G. Sorprendente, fluorescente! Energia, 27: 80-83, 2006.

international conferences

• Marcotullio G., “A novel concept for the production of furfural from pen-
toses with high yields and exceptionally low energy consumption” Bio-
mass derived pentoses: from biotechnology to fine chemistry, 14-16 Novem-
ber 2010, Reims - France. (Oral presentation).
• Marcotullio, G., de Jong, W., “Selective Acid-Catalyzed Furfural Genera-
tion from C5-Sugars Contained in Biomass - A Reaction Kinetics Optimiza-
tion Assessment” 18th European Biomass Conference and Exhibition, May
2010, Lyon - France. (Oral presentation and proceedings paper pp. 1454 -
1460).
• Tavares Cardoso, M.A., Marcotulio, G., Van Spronsen, J., Witkamp, G.-J., De
Jong, W., Van Ruud Ommen, J., “Dissolution and fractionation of wood and
straw using ionic liquids”, Conference Proceedings - 2009 AIChE Annual
Meeting, 09AIChE.
• Marcotullio, G. “Selective Acid-Catalyzed Production of Furfural From
C-5 Sugars Contained in Biomass. Reaction Kinetic Assessment.” Fifth In-
ternational Conference on Renewable Resources and Biorefineries (RRB5),
June 2009, Ghent - Belgium. (Oral presentaion).
 • Marcotullio, G. “Furfural Destruction Kinetics during Sulphuric Acid- Cat-
alyzed Production from Biomass” ECI Engineering Conference Interna-
tional – Bioenergy II: Fuels and Chemicals from Renewable Resources,
March 2009, Rio De Janeiro – Brasil. (Oral presentation).

• Marcotullio G., Heidweiller H., de Jong W., “Reaction kinetics assessment

for selective production of furfural from C5 sugars contained in biomass”
in proceedings of 16th European Biomass Conference and Exhibition -
From Research to Market, pp. 1-6, ETA, June 2008, Valencia - Spain. (Best
visual presentation award for the topic Biological conversion/Biorefiner-
ies).

• W. de Jong, G. Marcotullio, H. Heidweiller, and P. Jansens, ”A novel green

production process for cheaper furfural from biomass,’” in proceedings
of 15th European Biomass Conference and Exhibition - From Research to
Market, pp. 1-6, ETA, 7-11 May 2007, Berlin - Germany.

• R. Carapellucci, G. Girardi, G. Marcotullio, Coproduction of Hydrogen and

Electric Power Using Coal Hydrogasification, in proceedings of Clean Air
2005, 27-30 June 2005, Lisbon - Portugal.

patents

• Marcotullio, G. and de Jong, W., “Process for the production of Furfural

from pentoses”, patent application PCT/NL2011/050730; October 2010.

View publication stats