Mass spectrometric studies of fast pyrolysis of cellulose

Article abstract of DOI:10.1255/ejms.1335

A fast pyrolysis probe/linear quadrupole ion trap mass spectrometer combination was used to study the primary fast pyrolysis products (those that first leave the hot pyrolysis surface) of cellulose, cellobiose, cellotriose, cellotetraose, cellopentaose, and cellohexaose, as well as of cellobiosan, cellotriosan, and cellopentosan, at 600°C. Similar products with different branching ratios were found for the oligosaccharides and cellulose, as reported previously. However, identical products (with the exception of two) with similar branching ratios were measured for cellotriosan (and cellopentosan) and cellulose. This result demonstrates that cellotriosan is an excellent small-molecule surrogate for studies of the fast pyrolysis of cellulose and also that most fast pyrolysis products of cellulose do not originate from the reducing end. Based on several observations, the fast pyrolysis of cellulose is suggested to initiate predominantly via two competing processes: The formation of anhydro-oligosaccharides, such as cellobiosan, cellotriosan, and cellopentosan (major route), and the elimination of glycolaldehyde (or isomeric) units from the reducing end of oligosaccharides formed from cellulose during fast pyrolysis.

Full text of DOI:10.1255/ejms.1335

J.C. Degenstein et al., Eur. J. Mass Spectrom. 21, 321–326 (2015)
321
Received: 1 March 2015
n
Revised: 26 March 2015 n Accepted: 30 March 2015 n Publication: 19 June 2015
EUROPEAN
JOURNAL
OF
MASS
SPECTROMETRY
Special Issue Celebrating the 20th Anniversary of EJMS—European Journal of Mass Spectrometry
Mass spectrometric studies of fast pyrolysis
of cellulose
John C. Degenstein,a Matt Hurt,b Priya Murria,b Mckay Easton,a Harshavardhan Choudhari,a Linan Yang,b
James Riedeman,b Mark S. Carlsen,b John J. Nash,b Rakesh Agrawal,a W. Nicholas Delgass,a Fabio H. Ribeiroa
and Hilkka I. Kenttämaab,*
aSchool of Chemical Engineering, Purdue University, West Lafayette, Indiana 47907, USA
bDepartment of Chemistry, Purdue University, West Lafayette, Indiana 47907, USA. *E-mail: hilkka@purdue.edu
A fast pyrolysis probe/linear quadrupole ion trap mass spectrometer combination was used to study the primary fast pyrolysis products
(those that first leave the hot pyrolysis surface) of cellulose, cellobiose, cellotriose, cellotetraose, cellopentaose, and cellohexaose, as
well as of cellobiosan, cellotriosan, and cellopentosan, at 600°C. Similar products with different branching ratios were found for the
oligosaccharides and cellulose, as reported previously. However, identical products (with the exception of two) with similar branching
ratios were measured for cellotriosan (and cellopentosan) and cellulose. This result demonstrates that cellotriosan is an excellent small-
molecule surrogate for studies of the fast pyrolysis of cellulose and also that most fast pyrolysis products of cellulose do not originate
from the reducing end. Based on several observations, the fast pyrolysis of cellulose is suggested to initiate predominantly via two com-
peting processes: the formation of anhydro-oligosaccharides, such as cellobiosan, cellotriosan, and cellopentosan (major route), and the
elimination of glycolaldehyde (or isomeric) units from the reducing end of oligosaccharides formed from cellulose during fast pyrolysis.
Keywords: ion trap mass spectrometry, linear quadrupole ion trap, fast pyrolysis, cellulose, cellotriosan
Introduction
Fast pyrolysis (rapid heating in an inert atmosphere) is an
attractive alternative for the conversion of biomass into fuels
or valuable chemicals because it is a relatively simple and
scalable process.1–4 Most studies thus far have focused on
fast pyrolysis of cellulose since it is the simplest polymer in
biomass.1–4 Unfortunately, even for cellulose, the final fast
pyrolysis products (often referred to as biooil in the literature)
are a complex unstable mixture of molecules having an oxygen
content too high to be used (directly) as a fuel.1–4 Upgrading
this mixture is hindered by its extreme complexity, which
arises from numerous competing and consecutive reactions
both during and after pyrolysis.1–4 Currently, no agreement
general, have been proposed as intermediates several
times.1–6
With the aim of gaining a deeper understanding of the fast
pyrolysis of cellulose, which may allow better control over the
final products, the primary products of the fast pyrolysis of
cellulose, cellobiose, cellotriose, cellotetraose, cellopentaose,
and cellohexaose (Scheme 1), as well as of cellobiosan, cello-
triosan, and cellopentosan, were determined using a previously
described mass spectrometry methodology.7 Primary products,
as considered here, are the very first products to evaporate
from the hot surface (at 600°C) where pyrolysis occurs.
The reactor configuration utilized in this work was specifi-
exists in the literature on the mechanisms (e.g., radical, ionic, cally designed7 to detect the primary products and not to allow
or neither) of the fast pyrolysis reactions of cellulose, the
sequence of these reactions, or the identity of the primary
products of fast pyrolysis, although anhydro-oligomers, in
them to undergo further reactions. It has been previously
reported that the pyrolysis reactor configuration is of critical
importance to the product distribution produced from the
ISSN: 1469-0667
doi: 10.1255/ejms1335
© IM Publications LLP 2015
All rights reserved

322
Mass Spectrometric Studies of Fast Pyrolysis of Cellulose
attachment). The structures of the ions were examined by CAD
in MS2 and MS3 experiments and their elemental compositions
were determined by high-resolution measurements using a
Thermo Scientific LQIT/Fourier-transform ion cyclotron reso-
nance mass spectrometer. We are currently unable to deter-
mine mass balance for this pyrolysis experiment. However,
quantitation of the pyrolysis products was performed using a
pyrolysis–gas chromatography/mass spectrometry (GC/MS)
set-up (see the Supporting Information).
Scheme 1. The carbohydrate molecules studied here contain
two to six d-glucose units coupled to each other in a linear
manner via b(1®4) linkages, just as in cellulose.
Results and discussion
fast pyrolysis of cellulose.1b,4 Hence, the primary products
detected here may not be the same as detected in a reactor
of a different design. The gaseous molecules were ionized via
atmospheric pressure chemical ionization (APCI) using either
chloroform in methanol (negative ion mode, Cl– attachment)
The primary products of the fast pyrolysis of cellobiose, the
simplest compound studied, are shown in Figure 1 (top).7,8
The relative abundances of the ions reflect the relative abun-
dances of the products that produced them, as verified earlier
by using authentic compounds.7 Only ten major products were
observed (with an abundance of at least 10% compared to the
+
or ammonium hydroxide in water (positive ion mode, NH4
or proton attachment) in order to ensure that all the prod- most abundant product) and they are consistent with those
ucts were ionized and detected.7 Based on previous model
compound studies, both methods have been found7 to ionize
all the major pyrolysis products of cellobiose (some minor
recently reported in the literature.7,8 The unambiguously iden-
tified7 products include hydroxymethylfurfural (protonated
+
molecule, m/z 127), levoglucosan (NH4 adduct, m/z 180),
products were ionized by only one or the other methods). glucose (NH4+ adduct, m/z 198), b-d-glucopyranosylglycolalde-
Hence, only the positive ion mode results are discussed below. hyde [NH4+ adduct, m/z 240; note that this product is formed8
The structures of most of the ions formed from cellobiose and
cellulose have been examined7,8 previously by using tandem
mass spectrometry (MS2) experiments [i.e., by isolating them
by the loss of two glycolaldehyde units (or isomers) from cello-
+
biose], and cellobiosan (NH4 adduct, m/z 342). Based on our
preliminary computational studies,8 the formation of cellobi-
and subjecting them to collision-activated dissociation (CAD)]. osan has the lowest energy barrier of these reactions. Two
+
When necessary, the structures of the fragment ions were
examined by isolating them and subjecting them to CAD (MS3
experiment).7 Where possible, structures were confirmed7 by
levoglucosan isomers were generated, one that forms an NH4
adduct (m/z 180) like the authentic compound and one that
does not (protonated molecule, m/z 163).7,8 This finding is in
analyzing authentic compounds. High-resolution mass spec- agreement with an earlier report wherein the structure (anhy-
tral data needed to determine the elemental compositions of
droglucofuranose) was proposed for the second isomer.9 It is
the ions were collected using a Thermo Scientific linear quad- also noteworthy that levoglucosan is not the major primary
rupole ion trap (LQIT)/Fourier-transform ion cyclotron reso- product of fast pyrolysis, although it is a major final product.1,2
nance mass spectrometer.7 Similar studies were carried out
In prior reports that utilized on-stream fast pyrolysis-GC/MS,
in this research for the fast pyrolysis products of cellotriose, the largest molecules that were observed for cellulose (and
cellotetraose, cellopentaose, and cellohexaose, as well as of
cellobiosan, cellotriosan, and cellopentosan.
oligosaccharide) pyrolysis were levoglucosan and its isomers.2
This may be explained by the high final temperature of about
300°C typically used in GC analysis, which opens the possibility
for secondary reactions of larger primary products. Further,
the pyrolysis-MS reactor discussed here achieves both pyrol-
ysis and downstream analysis in as little as 125 + 57 ms (see
the Supporting Information for details), whereas a pyrolysis-
GC/MS reactor requires from two up to 30 minutes, depending
on the elution times of the products.
All the major products observed upon the fast pyrolysis of
cellobiose were also observed for cellotriose, cellotetraose,
cellopentaose, cellohexaose, and even cellulose, albeit with
different relative abundances. Specifically, the abundances of
Experimental methods
The pyrolysis method7 employed here is based on the coupling
of a very fast-heating (up to 20,000°Cs–1) Pyroprobe 5200 (CDS
Analytical, Oxford, PA) to a Thermo Scientific LTQ LQIT mass
spectrometer (Waltham, MA) through a custom-built adaptor.
The pyrolysis probe uses a resistively heated platinum ribbon
(2.1mm × 35mm × 0.1 mm). The pyrolysis probe was placed
inside the APCI source of the linear quadrupole ion trap and
the ribbon was heated up to 600°C at a rate of 1000°C s–1. b-d-glucopyranosylglycolaldehyde [m/z 240, dominant product
The primary products of pyrolysis evaporated into a nitrogen
for cellobiose formed via the loss of two glycolaldehyde mole-
atmosphere at 100°C in the ion source and were quenched. cules (or isomers)], glucose (m/z 198), cellobiose that has lost
The gaseous molecules were ionized via APCI using ammo- a glycolaldehyde molecule (or isomer, m/z 300), and cellobiose
+
nium hydroxide in water (positive ion mode; NH4 or proton
that has lost both a glycolaldehyde molecule (or isomer) and a

J.C. Degenstein et al., Eur. J. Mass Spectrom. 21, 321–326 (2015)
323
Figure 1. Positive ion mode mass spectra showing the primary products of the fast pyrolysis of cellobiose (top) and cellohexaose
(bottom) ionized by APCI with ammonium hydroxide. The ions whose relative abundances differ the most are indicated with dotted
lines. Two isomers of levoglucosan were observed, levoglucosan itself (ion of m/z 180) and another isomer (ion of m/z 163). All the
structures shown for cellobiose have been identified7 by CAD mass spectral comparison using authentic compounds.
water molecule (m/z 282) decrease systematically proceeding
from the dimer to trimer, tetramer, pentamer, hexamer, and
cellulose. For example, Figure 1 shows a comparison of the
primary products of the fast pyrolysis of cellobiose (top) and
cellohexaose (bottom). It is obvious that the four products
listed above (highlighted with dotted lines in Figure 1) have
substantially lower abundances for cellohexaose than for
In addition to the products observed for cellobiose, the
oligomers studied generated two larger products: a molecule
likely to be b-d-cellobiopyranosylglycolaldehyde (NH4+ adduct,
+
m/z 402) and cellotriosan (NH4 adduct, m/z 504, verified by
comparison of its CAD mass spectrum to that of an authentic
sample). Cellulose also produced a very small amount of
+
cellotetrosan [NH4 adduct, m/z 666, verified by comparison
cellobiose. For cellulose, their abundances are even lower, of its CAD mass spectrum to an authentic sample (Figure
as shown in Figure 2 (bottom). These findings suggest that 2, bottom)]. Hence, cellotriosan appears to be the largest
the four products described above are somehow associ- product with a significant abundance that is efficiently able to
ated with the terminal glucose units because the ratio of the
terminal units to the total number of glucose units decreases
as the size of the oligomer increases. Indeed, the ratio of the
end group to the monomer has been reported3 to be a vital
descriptor of cellulose pyrolysis chemistry.
escape the hot pyrolysis surface for cellotriose and the larger
oligosaccharides as well as cellulose during the fast pyrolysis
at 600°C.
Inspired by the above observations, the fast pyrolysis of
cellotriosan was also performed. Cellobiosan dominates this
Figure 2. Positive ion mode mass spectra showing the primary products of the fast pyrolysis of cellotriosan (top) and cellulose (bottom)
ionized by APCI with ammonium hydroxide. Some of the ions having the same m/z ratio are indicated with a dotted line.

324
Mass Spectrometric Studies of Fast Pyrolysis of Cellulose
Figure 3. A positive ion mode mass spectrum showing the primary products of the fast pyrolysis of cellobiosan (m/z 342) ionized by
APCI with ammonium hydroxide. The ions of m/z 85, 127, 145, 163, 180, and 195 are also formed for cellotriosan, cellopentosan and cel-
lulose. The ion of m/z 324 results from the loss of H2O from the ion of m/z 342.
product distribution. However, all the major products observed
for cellobiose and the oligomers were also observed, with
cellobiose that has lost a glycolaldehyde molecule (or isomer,
m/z 300), and the product corresponding to cellobiose that has
the following exceptions: glucose (m/z 198), b-d-glucopyra- lost both a glycolaldehyde (or isomer) and a water molecule
nosylglycolaldehyde (m/z 240), the product corresponding to (m/z 282). These are the four products mentioned above as
Scheme 2. A simple schematic of the major fast pyrolysis pathways proposed for oligosaccharides formed from cellulose during fast
pyrolysis upon addition of water. The cleavages indicated in red are thought to occur in the middle of a cellulose chain. The cleavages
indicated in blue and green likely occur only at the reducing terminals, which for long chains of cellulose represent a small overall
fraction of the total units. Hence, they are minor pathways.

J.C. Degenstein et al., Eur. J. Mass Spectrom. 21, 321–326 (2015)
325
being somehow associated with the terminal glucose units. As
the reducing end in cellotriosan has been modified compared
to that in cellotriose, and cellotriose yields these four prod-
ucts but cellotriosan does not, the formation of these four
products is likely to depend on the presence of the reducing
end in cellotriose. The same is true for cellobiosan (Figure 3)
except that no ions larger than m/z 342 (ionized cellobiosan)
were observed. The mechanisms of these fragmentations are
why levoglucosan is a major final fast pyrolysis product but not
a major primary product.
Supporting information
Detailed descriptions of the pyrolysis probe/tandem mass
spectrometry and gas chromatography/mass spectrometry
currently being investigated using quantum chemical calcu- (GC/MS) experiments and quantitative product distribution
lations. The most surprising finding made in this study is
that the fast pyrolysis product distribution of cellotriosan (and
cellopentosan) is nearly identical to that of cellulose (Figure
2). The most significant differences are the formation of b-d-
glucopyranosylglycolaldehyde (m/z 240) and b-d-cellobiopyra-
nosylglycolaldehyde (m/z 402) for cellulose but not for cello-
triosan (or cellopentosan).These two products correspond
to cellotriose that has lost two glycolaldehyde molecules (or
isomers) (m/z 402) and cellobiose that has undergone the
same losses (m/z 240). These findings suggest that the fast
pyrolysis of cellulose may be initiated predominantly via two
determined using GC/MS experiments can be found in the
online version at http://dx.doi.org/10.1255/ejms.1335.
Notes
The authors declare no competing financial interest.
Acknowledgments
competing processes—the formation of anhydro-oligosac- The authors gratefully acknowledge partial financial support
charides, such as cellobiosan, cellotriosan, and cellopen- by the Center for Direct Catalytic Conversion of Biomass to
tosan (major route), and the elimination of glycolaldehyde
Biofuels (C3Bio), an Energy Frontier Research Center funded
(or isomeric) units from the reducing end of oligosaccha- by the US Department of Energy, Office of Science, Office of
rides formed from cellulose during fast pyrolysis [minor route
leading to products of m/z 240 and m/z 402 (Scheme 2)].
Basic Energy Sciences, under Award Number DE-SC0000997.
The work by John C. Degenstein was supported by the National
Science Foundation EFRI (0938033-DGE).
Conclusions
References
The results suggest that molecules larger than cellotriosan
are not able to leave the heated surface efficiently during fast
pyrolysis of oligosaccharides at 600°C. Instead, they undergo
further degradation on the hot surface. The observation of very
similar primary product distributions for the fast pyrolysis of
cellotriosan and cellulose suggests that cellotriosan presents
an excellent small-molecule surrogate for cellulose, and it
should be a much better choice than glucose, which has been
considered previously.3
1. (a) D. Shen, R. Xiao, S. Gu and K. Luo, “The pyrolytic
behavior of cellulose in lignocellulosic biomass: a
review”, RSC Adv. 1, 1641 (2011). doi: http://dx.doi.
org/10.1039/c1ra00534k (b) G. Huber, S. Iborra and A.
Corma, “Synthesis of transportation fuels from biomass:
chemistry, catalysts, and engineering”, Chem. Rev. 106,
4044 (2006). doi: http://dx.doi.org/10.1021/cr068360d
2. (a) R. Vinu and L.J. Broadbelt, L. “A mechanistic model of
fast pyrolysis of glucose-based carbohydrates to predict
bio-oil composition”, J. Energy Environ. Sci. 5, 9808
(2012). doi: http://dx.doi.org/10.1039/c2ee22784c (b) H.B.
Mayes and L.J. Broadbelt, “Unraveling the reactions that
unravel cellulose”, J. Phys. Chem. A 116, 7098 (2012). doi:
http://dx.doi.org/10.1021/jp300405x
Based on the primary products observed for cellotri-
osan and cellulose, the fast pyrolysis of cellulose under the
conditions used here may be initiated predominantly via two
competing pathways. One involves the formation of small
anhydro-oligosaccharides (but not predominantly levoglu-
cosan, as suggested in the literature1–3) that either evaporate
from the hot pyroprobe surface or degrade to yield most of the
other primary products (major pathway). The other involves
the elimination of glycolaldehyde (or isomeric) molecules
from the reducing end of small oligosaccharides formed from
cellulose during the pyrolysis to yield volatile b-d-cellobiopyra-
nosyl-glycolaldehyde and b-d-glucopyranosylglycolaldehyde
molecules (ions of m/z 240 and m/z 402 (Scheme 1)). Reactions
of the primary products of fast pyrolysis are currently under
investigation in order to understand better how the final fast
pyrolysis products are formed. These reactions may explain
3. (a) M.S. Mettler, S.H. Mushrif, A.D. Paulsen, A.D.
Javadekar, D.G. Vlachos and P.J. Dauenhauer, “Revealing
pyrolysis chemistry for biofuels production: conversion
of cellulose to furans and small oxygenates”, Energy
Environ. Sci. 5, 5414 (2012). doi: http://dx.doi.org/10.1039/
C1EE02743C (b) M.S. Mettler, A.D. Paulsen, D.G. Vlachos
and P.J. Dauenhauer, “The chain length effect in
pyrolysis: bridging the gap between glucose and
cellulose”, Green Chem. 14, 1284 (2012). doi: http://dx.doi.
org/10.1039/c2gc35184f

326
Mass Spectrometric Studies of Fast Pyrolysis of Cellulose
4. (a) P.R. Patwardhan, D.L. Dalluge, B.H. Shanks and
R.C. Brown, “Distinguishing primary and second-
ary reactions of cellulose pyrolysis”, Biores. Tech.
102, 5265 (2011) doi: http://dx.doi.org/10.1016/j.
biortech.2011.02.018. (b) P.R. Patwardhan, J.A. Satrio,
R.C. Brown and B.H. Shanks, “Product distribution
from fast pyrolysis of glucose-based carbohydrates”,
J. Anal. Appl. Pyrolysis 86, 323 (2009). doi: http://dx.doi.
org/10.1016/j.jaap.2009.08.007
5. V. Agarwal, P.J. Dauenhauer, G.W. Huber and S.M.
Auerbach, “Ab initio dynamics of cellulose pyrolysis:
nascent decomposition pathways at 327 and 600°C”,
J. Am. Chem. Soc. 134, 14958 (2012). doi: http://dx.doi.
org/10.1021/ja305135u
7. M.R. Hurt, J.C. Degenstein, P. Gawecki, D.J. Borton II,
N.R. Vinueza, L. Yang, R. Agrawal, W.N. Delgass, F.H.
Ribeiro and H.I. Kenttämaa, H. I. “On-line mass spec-
trometric methods for the determination of the primary
products of fast pyrolysis of carbohydrates and for their
gas-phase manipulation”, Anal. Chem. 22, 10927 (2013).
doi: http://dx.doi.org/10.1021/ac402380h
8. J.C. Degenstein, P. Murria, M. Easton, H. Sheng, M. Hurt,
A.R. Dow, J. Gao, J. Nash, R. Agrawal, W.N. Delgass,
F. Ribeiro and H.I. Kenttämaa, “Fast pyrolysis of
13C-labeled cellobioses: gaining insights into the mecha-
nisms of fast pyrolysis of carbohydrates”, J. Org. Chem.
80, 1909 (2015). doi: http://dx.doi.org/10.1021/jo5025255
9. Q. Lu, X. Yang, C. Dong, Z. Zhang, X. Zhang and X. Zhu,
“Influence of pyrolysis temperature and time on the cel-
lulose fast pyrolysis products: analytical Py-GC/MS
study”, J. Anal. Appl. Pyr. 92, 430 (2011). doi: http://dx.doi.
org/10.1016/j.jaap.2011.08.006
6. J. Piskorz, P. Majerski, D. Radlein, A. Vladars-Usas and
D.J. Scott, “Flash pyrolysis of cellulose for production of
anhydro-oligomers”, J. Anal. App. Pyr. 56, 145 (2000). doi:
http://dx.doi.org/10.1016/S0165-2370(00)00089-9