Lignin solubilization and aqueous phase reforming for the production of aromatic chemicals and hydrogen

Article abstract of DOI:10.1002/cssc.201000299

The solubilization and aqueous phase reforming of lignin, including kraft, soda, and alcell lignin along with sugarcane bagasse, at low temperatures (T≤498 K) and pressures (P≤29 bar) is reported for the first time for the production of aromatic chemicals and hydrogen. Analysis of lignin model compounds and the distribution of products obtained during the lignin aqueous phase reforming revealed that lignin was depolymerized through disruption of the abundant β-O-4 linkages and, to a lesser extent, the 5-5' carbon-carbon linkages to form monomeric aromatic compounds. The alkyl chains contained on these monomeric compounds were readily reformed to produce hydrogen and simple aromatic platform chemicals, particularly guaiacol and syringol, with the distribution of each depending on the lignin source. The methoxy groups present on the aromatic rings were subject to hydrolysis to form methanol, which was also readily reformed to produce hydrogen and carbon dioxide. The composition of the isolated yields of monomeric aromatic compounds and overall lignin conversion based on these isolated yields varied from 10-15% depending on the lignin sample, with the balance consisting of gaseous products and residual solid material. Furthermore, we introduce the use of a high-pressure autoclave with optical windows and an autoclave with ATR-IR sentinel for on-line in situ spectroscopic monitoring of biomass conversion processes, which provides direct insight into, for example, the solubilization process and aqueous phase reforming reaction of lignin.

Full text of DOI:10.1002/cssc.201000299

DOI: 10.1002/cssc.201000299
Lignin Solubilization and Aqueous Phase Reforming for
the Production of Aromatic Chemicals and Hydrogen
Joseph Zakzeski and Bert M. Weckhuysen*[a]
The solubilization and aqueous phase reforming of lignin, in-
cluding kraft, soda, and alcell lignin along with sugarcane bag-
asse, at low temperatures (Tꢁ498 K) and pressures (Pꢁ29 bar)
is reported for the first time for the production of aromatic
chemicals and hydrogen. Analysis of lignin model compounds
and the distribution of products obtained during the lignin
aqueous phase reforming revealed that lignin was depolymer-
ized through disruption of the abundant bÀOÀ4 linkages and,
to a lesser extent, the 5À5’ carbon-carbon linkages to form
monomeric aromatic compounds. The alkyl chains contained
on these monomeric compounds were readily reformed to
produce hydrogen and simple aromatic platform chemicals,
particularly guaiacol and syringol, with the distribution of each
depending on the lignin source. The methoxy groups present
on the aromatic rings were subject to hydrolysis to form meth-
anol, which was also readily reformed to produce hydrogen
and carbon dioxide. The composition of the isolated yields of
monomeric aromatic compounds and overall lignin conversion
based on these isolated yields varied from 10–15% depending
on the lignin sample, with the balance consisting of gaseous
products and residual solid material. Furthermore, we intro-
duce the use of a high-pressure autoclave with optical win-
dows and an autoclave with ATR-IR sentinel for on-line in situ
spectroscopic monitoring of biomass conversion processes,
which provides direct insight into, for example, the solubiliza-
tion process and aqueous phase reforming reaction of lignin.
Introduction
With the depletion of fossil fuels as a source of fuels, chemi-
cals, and energy, the fraction of energy and chemicals supplied
by renewable resources such as biomass can be expected to
increase in the foreseeable future.[1,2] One particular opportuni-
ty to produce these products arises through the catalytic valor-
ization of lignin. Lignin is a natural amorphous three-dimen-
sional polymer consisting of methoxylated phenylpropane
structures that composes a significant portion of lignocellulosic
biomass. As of 2004, the pulp and paper industry alone pro-
duced 50 million tons of extracted lignin, but only approxi-
mately 2% of the lignin available is used commercially, with
the remainder used as a low-value fuel.[3] With its unique struc-
ture and chemical properties, the catalytic valorization of lignin
represents a potentially useful method to obtain bulk and fine
chemicals. The composition of lignin varies from plant to plant,
with lignin from softwood consisting mostly of coniferyl alco-
hol units whereas hardwoods consist largely of syringol units,
although exceptions are known.[1] The coniferyl alcohol units
contain a single methoxy group while the syringol alcohol unit
contains two methoxy groups. In some plant species p-cou-
maryl units, which lack methoxy groups on the aromatic ring,
are found in the lignin structure. In the lignin polymer, these
coniferyl, syringol, and p-coumaryl units are connected
through various linkages.[1] The bÀOÀ4 and 5À5’ linkages are
the most abundant, constituting approximately 50–60% and
20–25%, respectively, of all linkages in lignin.[1] The remaining,
less-abundant linkages present in lignin include the 4ÀOÀ5, b-
b, b-5, spirodienone, dibenzodioxocin, and phenylcoumaran
linkages. A schematic depiction of lignin showing these linkag-
es is given in Figure 1.
The disruption of the linkages in lignin represents a poten-
tial route for the production of a wide range of aromatic com-
pounds. Indeed, the depolymerization of the complicated
lignin polymer into smaller platform molecules is an important
aspect of lignin valorization. Previous methods to depolymer-
ize lignin include pyrolysis, catalytic hydrogenation, oxidation,
or hydrocracking.[1,4–7] US Patent Nos. 0218061A1 (2009) and
0218062A1 (2009) to Schinski et al. disclose a process for the
hydrotreatment of lignin to yield aromatic products, requiring
the use of a hydrogen feedstock.[8,9] Lignin treatment and
gasification using supercritical water (Tc =647.3 K, Pc =221 bar)
have also been reported to primarily form light alkanes and hy-
drogen,[10–17] although the disadvantages of these processes in-
clude the need for high reaction temperatures and pressures
(often Tꢀ673 K and Pꢀ250 bar).
The aqueous-phase reforming (APR) of biomass-derived oxy-
genated compounds, such as methanol, glycols, glycerol, sorbi-
tol, xylose, and glucose, has been reported for the production
[a] Dr. J. Zakzeski, Prof. Dr. B. M. Weckhuysen
Inorganic Chemistry and Catalysis group
Debye Institute for NanoMaterials Science
Utrecht University
Sorbonnelaan 16, 3584 CA Utrecht (The Netherlands)
Fax: (+31)30 251 1027
E-mail: b.m.weckhuysen@uu.nl
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cssc.201000299.
ChemSusChem 2011, 4, 369 – 378
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
369

J. Zakzeski and B. M. Weckhuysen
until this work, there have been
no reports of the APR of pure
lignin samples at low tempera-
tures and pressures. The purpose
of this work is therefore to dem-
onstrate a process for the solva-
tion and APR of lignin, exempli-
fied with the use of soda lignin,
kraft lignin, alcell lignin, and
lignin from sugarcane bagasse,
for the production of hydrogen
and aromatic platform chemi-
cals. Through the use of lignin
model compounds and analysis
of actual lignin samples, we
demonstrate the capabilities of
the system and the types of
products obtainable by the APR
of pure lignin samples. Addition-
ally, information on these pro-
cesses will be obtained with spe-
cially-designed high-pressure au-
toclaves equipped with optical
windows and an ATR-IR sentinel
providing direct in situ mecha-
nistic insight.
Figure 1. Schematic depiction of lignin, showing various linkages and lignin model compounds used in this study
to model (A) phenol and methoxy functionality, (B) bÀOÀ4 linkages, (C) 5À5’ linkages, (D) propyl side chain, and
(E) benzylic groups. Lignin structure reproduced from Ref. [1]. Copyright 2010, American Chemical Society.
of hydrogen at temperatures (T<538 K) and pressures consid-
erably lower than those required for gasification or pyrolysis.[18]
Several patents and manuscripts have been published in the
field of APR of biomass-derived oxygenated compounds for
the production of hydrogen and chemicals. The preferred cata- Results and Discussion
lysts for these systems comprise Group VIII transition metals,
Lignin solubilization in water
including alloys and mixtures, with platinum, ruthenium, or
rhodium giving the most favorable results. The catalyst sup-
port is preferably selected from the group consisting of alumi-
na, boron nitride, carbon, ceria, silica, silica-alumina, silica ni-
tride, titania, zirconia, or mixtures thereof, with silica the pre-
ferred support.[19,20] In most reports of APR reactions, fluidized-
bed tubular reactors were used, and a decrease in void-space,
defined as portions of the reactor that contain no solid cata-
lyst, resulted in higher hydrogen production.[21] In the case of
actual biomass, however, the presence of solid feeds and resi-
dues increases the difficulty of developing continuous process-
es; thus, batch or semi-batch systems are more readily applied
to the reactions.
We began our investigation by conducting the solubilization
of the lignin samples under APR conditions. As indicated
above, the structural composition varies considerably from
plant to plant and also depends on the pretreatment method
used to obtain the lignin. The kraft lignin used in this study,
for example, contains sulfur, whereas the soda lignin and sug-
arcane bagasse have reduced sulfur content, and the organo-
solv alcell lignin and is relatively sulfur-free. For the purpose of
analyzing the solvation behavior of the various lignin samples
used in this study, we employed a specially constructed reactor
equipped with quartz windows (see Supporting Information
for reactor schematic) and also a reactor equipped with a sili-
con sentinel for in-situ ATR-IR spectroscopy measurements.
Figure 2 depicts the visual images for the exemplary case of
organosolv lignin heated from room temperature to 498 K. Ini-
tially, the organosolv lignin was insoluble in water and formed
a slurry upon vigorous stirring. The color of this slurry became
gradually darker as the temperature increased. The biggest
change in the slurry occurred at 388 K, where a large portion
of the lignin began to solubilize in the aqueous phase. Upon
reaching 403 K, nearly all of the solid mass was dissolved,
which resulted in a yellow-colored solution. As the tempera-
ture was increased to 498 K (the reaction temperature used for
APR) the solution color turned crimson-red, possibly attributa-
ble to soluble quinones.[25] At the conclusion of the reaction,
To date, all studies of APR have focused on the use of model
compounds that could be derived from biomass, such as sorbi-
tol and glucose, and only recently the APR of actual biomass
consisting of Southern pine sawdust was investigated for the
production of hydrogen.[22] Only a small fraction of the lignin
in the Southern pine sawdust was acid-soluble, and depolyme-
rization occurred to a limited extent due to the formation of
some acid-labile bonds, such as a- and b-ether linkages.[22]
Earlier reports of condensation caused by the acid hydrolysis
of lignin is due to the intermolecular dehydration between
benzylic carbon and guaiacyl aromatic rings, and sulfuric acid
caused a cross-linking effect that resulted in the formation of
higher molecular weight polymeric products.[22–24] Nevertheless,
370
www.chemsuschem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2011, 4, 369 – 378

Lignin Solubilization and Aqueous Phase Reforming
Figure 2. Visual images for the solubilization of organolv lignin in water from 298 K to 498 K, measured with the high-pressure autoclave equipped with an
optical window.
the solution color reverted back to the yellow color originally
observed at 403 K.
matic CÀH breathing species were detected in the fingerprint
region. Quinones tend to absorb strongly between 1690 and
1655 cmÀ1,[26] so the appearance of the vibration at 1695 cmÀ1
may be attributed to quinone species. The vibrations at 1508,
1460, and 1426 cmÀ1 correspond to ring carbon-carbon skeletal
vibrations, which usually give three bands with the strongest
The ATR-IR spectra of the organosolv lignin dissolution are
depicted in Figure 3a, while the fingerprint regions of the
spectra are given in Figure 3b. At room temperature, the dom-
inant vibrations observed originated from the water stretching
and bending modes, and vibrations attributed to the insoluble
lignin were not readily observed. The most significant change
in the IR spectra occurred between 388 and 403 K, which cor-
responded to the appearance of several vibrations detected by
ATR-IR. Two aliphatic CÀH stretch vibrations were observed at
2945 and 2852 cmÀ1, and several vibrations attributed to aro-
near 1500 cmÀ1 [26] The vibrations in the range 1030–1150 cmÀ1
.
are likely attributed to hydroxyphenyl vibrations from guaiacol
and syringol groups, as supported by the appearance of similar
vibrations observed in pure guaiacol samples.[26] The vibrations
at 1358, 1314, and 1213 cmÀ1 may be attributed to phenolic
OÀH deformation and CÀO stretching vibrations, which tend
to appear in the ranges 1390–1330 cmÀ1 and 1260–
1180 cmÀ1.[26] The appearance of these vibrations corresponded
to the formation of dissolved organic species from the lignin
sample, as described above.
The behavior of the kraft lignin contrasted distinctly with
the organosolv lignin. Figure 4 depicts the visual images cap-
tured during the similar treatment of kraft lignin. As with the
organosolv lignin, the kraft lignin was initially water-insoluble,
and a slurry formed upon vigorous stirring. In contrast to the
organosolv lignin slurry, which tended to darken with increas-
ing temperature, the kraft slurry became lighter in color as the
reaction temperature was increased from 293 to 403 K. A dis-
tinct change occurred at 482 K, which corresponded to the ag-
glomeration of solid materials in the aqueous solution. This
material began to deposit near the windows at 487 K, and gas
bubbles were observed to form on the solid surface at 498 K.
Similarly to the organosolv lignin, the solution reverted back
to a yellow color upon cooling.
The visual images and IR spectra for the other two lignin
samples, namely the sugarcane bagasse and soda lignin, are
given in the Supporting Information. These samples resembled
the kraft and organosolv lignin in that the solution color
turned red at elevated temperatures. In both cases the onset
of dissolution began at 464 K. Very few solids were observed in
the reactor in the case of the soda lignin, whereas black parti-
cles were readily observed in the case of the sugarcane bag-
asse.
Figure 3. In situ ATR-IR spectra of the dissolution of organosolv lignin in
water: a) full range spectra, and b) fingerprint region.
ChemSusChem 2011, 4, 369 – 378
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
371

J. Zakzeski and B. M. Weckhuysen
Figure 4. Visual images for the solubilization of kraft lignin in water from 298 K to 498 K, measured with the high-pressure autoclave equipped with an optical
window.
Aqueous phase reforming of lignin
The total masses of the solids recovered at the conclusion of
the reaction (at 293 K) for the different lignin samples are sum-
marized in Table 1. In accordance with the visual observations
and ATR-IR spectra (Figure 5), nearly 93% of the starting orga-
nosolv lignin was solubilized with the remainder recovered as
solid residue. Slightly lesser quantities were dissolved with the
soda lignin, corresponding to 83% of the initial mass dissolved.
As also indicated by the visual observations, solids were pre-
dominant in both the case of the sugarcane bagasse and kraft
lignin, corresponding to only 51% and 38% of the initial mass
dissolved, respectively. Some of the differences in solubility be-
tween the lignin samples may be attributed to the lignin parti-
cle sizes, which tended to be larger for the kraft and sugarcane
bagasse lignin relative to the organosolv and soda lignin (see
SEM images in the Supporting Information).
With the promising results obtained with respect to the lignin
solubilization in water, we next turn our attention to the APR
of the lignin samples for the production of hydrogen and aro-
matics. As discussed in further detail below, sulfuric acid was
used in these studies, and the corrosive conditions precluded
the use of the windowed reactor and in situ ATR-IR sentinel; in-
stead, smaller, simpler autoclaves had to be used. The catalytic
results of the APR of various lignin samples are given in
Table 2. The columns indicate the lignin sample identity, the
average starting mass of the dry lignin loaded into the auto-
clave, and the weight of dry solids recovered at the conclusion
of the reaction not including the catalyst, which adhered to
the autoclave walls and was recovered as a separate, grey-col-
ored solid (see Supporting Infor-
mation for SEM images of the
Table 1. Solids recovered during the dissolution of the various lignin samples. Reaction conditions: 200 mL
H2O, P=29 bar He, T=498 K, t=1.5 h.
solids before and after reaction).
The total isolated products are
the products extracted from the
aqueous phase using dichloro-
methane and then isolated by
removal of the dichloromethane
solvent by rotary evaporation
(see Supporting Information for
NMR spectra of these isolated
products). The GC-detected iso-
lated products indicate the
quantity of total isolated prod-
ucts detected by gas chromatog-
raphy (see Supporting Informa-
tion for mass spectra of materi-
als unidentified by GC/MS). The
balance between the GC detect-
ed isolated products and the
total isolated products likely
constitute higher molecular
weight products undetectable
by GC. Finally, the percent con-
Entry
Lignin
Lignin starting
mass [g]
Total recovered
solid mass [g]
Pt/Al2O3
mass [g]
Solids from
lignin [g]
Solubilized
lignin [%]
1
2
3
4
organosolv
kraft
soda
2.002
2.001
1.998
1.999
0.344
1.450
0.548
1.182
0.201
0.202
0.200
0.200
0.144
1.248
0.348
0.983
93
38
83
51
sugarcane bagasse
Table 2. Solids and liquids obtained from the aqueous phase reforming of lignin samples. Reaction conditions:
10.98g H2O, 0.58g H2SO4, 0.1245g 1% Pt/Al2O3, P=29 bar He, T=498 K, t=1.5h.
Entry
Lignin
Avg. start
mss [g]
Recovered
solids[a] [g]
Total isolated
Products [g]
GC-detected
isolated products [g]
Conv.
[g]
1[b]
2[c]
3
4
5
kraft
kraft
kraft
alcell
sugar-cane
soda
1.384
1.383
1.386
1.384
1.387
1.387
1.302
1.157
1.219
1.099
0.935
0.963
0.026
0.069
0.136
0.203
0.174
0.159
0.012
0.026
0.020
0.038
0.028
0.025
1.9
4.9
9.8
14.6
12.6
11.4
6
[a] After subtraction of catalyst weight. [b] No H2SO4. [c] No Pt/Al2O3.
372
www.chemsuschem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2011, 4, 369 – 378

Lignin Solubilization and Aqueous Phase Reforming
Figure 5. In situ ATR-IR spectra of the dissolution of kraft lignin in water:
a) full spectra, and b) fingerprint region.
Figure 6. Composition of GC-detected isolated products from kraft lignin
APR in the absence or presence of H2SO4 or Pt/Al2O3. Reaction conditions:
10.98 g H2O, 0.58 g H2SO4, 1.385 g kraft lignin, 0.1245 g 1% Pt/Al2O3,
P=29 bar He, T=498 K, t=1.5 h.
version corresponds to the percent of lignin converted based
on the recovered weight of isolated products.
Although differences exist between the lignin samples, we
begin our study of the actual lignin samples using kraft lignin
as an exemplary feedstock and analyze the role of the reaction
components on the products obtained, particularly H2SO4 and
Pt/Al2O3. The results of solids and liquids obtained from kraft
lignin in the absence of H2SO4 are given in Table 2, entry 1. Ap-
proximately 94% of the lignin was recovered as a dark solid
material, and a small number of isolated products were ob-
tained, corresponding to 1.9% of the lignin. Discrepancies in
the mass balance occur because of gas products formed and
small losses attributed to solids tightly bound to the autoclave
stirrer, although the latter losses are minimal. The composition
of the isolated products is given in Figure 6. The other prod-
ucts were often complicated, high-molecular-weight com-
pounds that were not identified by GC/MS (see Supporting In-
formation). A majority of the isolated products consisted of
coniferyl-derived products with considerably less p-coumaryl-,
syringol-derived, and other products, as expected since the
kraft lignin was obtained from pine, a softwood consisting
mostly of coniferyl units. The results of kraft lignin in the pres-
ence of H2SO4 but in the absence of Pt/Al2O3, are given in
Table 2, entry 2. The presence of H2SO4 resulted in reduced
quantities of solid products recovered at the conclusion of the
reaction (corresponding to 84% of the starting lignin by mass)
and increased isolated yields relative to the reaction in the ab-
sence of H2SO4. As indicated by Figure 6, the product distribu-
tion differed considerably, with the majority of the isolated
products consisting of unidentified products. Table 2 entry 3
gives the results of the APR of kraft lignin with all components
present. In this case, the highest isolated yields were obtained,
which corresponded to 9.8% conversion of the dry lignin to
isolated products. The product distribution also varied consid-
erably in that fewer higher-weight unidentified products were
detected, which corresponded to an increase in guaiacol-de-
rived products. These results indicate that although aromatic
products are obtained in the absence of either H2SO4 or Pt/
Al2O3, the highest yields of lower-molecular-weight aromatic
products are obtained in the presence of both components.
As indicated by Table 2, entries 4, 5, and 6, similar results to
the case of kraft lignin with all reaction components were ob-
tained with the alcell, sugarcane bagasse, and soda lignin, in
which between 73%, 68%, and 69% of the dry lignin was re-
covered as solids, respectively, which also corresponded to
14.6%, 12.6% and 11.4% conversion of lignin to isolated prod-
ucts. As shown in Figure 7, which gives the composition of the
isolated products, the product distributions varied depending
on the lignin samples. Whereas predominantly coniferyl-de-
rived products were obtained from kraft lignin, including
guaiacol and catechol, in the case of alcell lignin, the most
abundant product observed were sinapol-derived products in-
cluding syringol and the related product 1,2-diphenyl-3-me-
ChemSusChem 2011, 4, 369 – 378
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
373

J. Zakzeski and B. M. Weckhuysen
with the balance gas consisting mostly of the He initially
charged to the autoclave. For more details of the gas composi-
tion in the autoclave at various reaction times we refer to the
Supporting Information. We again begin our analysis using the
exemplary case of kraft lignin and analyze the importance of
the H2SO4 and Pt/Al2O3 components on the gas production.
Similar results (not shown) were obtained for the other lignin
samples. As indicated by Table 3, entry 1, only traces of H2
Table 3. Autoclave gas composition following the aqueous phase reform-
ing of the lignin samples. Reaction conditions: 10.98 g H2O, 0.58 g H2SO4,
0.1245 g 1% Pt/Al2O3, P=29 bar He, T=498 K, t=1.5 h.
Entry
Lignin
H2
[%]
CO2
[%]
CH4
[%]
C2H6
[%]
C3H8
[%]
1[a]
2[b]
3
4
5
kraft
kraft
kraft
alcell
sugarcane
bagasse
soda
0.03
0.47
4.28
8.83
5.03
0.00
0.00
0.95
1.52
1.14
1.05
1.10
2.55
3.06
1.98
0.55
4.36
3.38
5.52
4.26
0.03
0.00
0.02
0.06
0.03
6
5.51
2.35
3.74
1.12
0.07
[a] No H2SO4. [b] No Pt/Al2O3.
were formed when H2SO4 was omitted from the system, al-
though small quantities of lighter hydrocarbons were detected.
When instead Pt/Al2O3 was omitted, increased quantities of H2
were produced (Table 3, entry 2) relative to the case in which
H2SO4 was omitted, indicating that the Pt/Al2O3 is not solely re-
sponsible for the production of hydrogen, an effect similarly
observed during the APR of wood and other biomass.[22] The
proportion of other light gases, particularly ethane, was also
higher. The highest H2 yields were obtained in the presence of
both H2SO4 and Pt/Al2O3 (Table 3, entry 3). Similar results were
obtained with the other lignin samples (Table 2, entries 4, 5,
and 6) with the highest hydrogen yields were obtained from
the alcell organosolv lignin (8.83%). In the case of all lignin
samples, the hydrogen production ceased after 1.5 h of reac-
tion (see Supporting Information).
Figure 7. Composition of GC-detected isolated products from alcell, bagasse,
kraft, and soda lignin APR. Reaction conditions: 10.98 g H2O, 0.58 g H2SO4,
1.385 g lignin, 0.1245 g 1% Pt/Al2O3, P=29 bar He, T=498 K, t=1.5 h.
thoxybenzene followed by coniferyl-derived guaiacol. The
products obtained from sugarcane bagasse, which is the resid-
ual plant material remaining after sugarcane is crushed to ex-
tract the juice, consisted predominantly of p-coumaryl-derived
products, including phenol, although units associated with
guaiacol and some syringol were also detected. Compounds
with remnants of the propyl chain, such as methylphenol and
ethylguaiacol were also detected, as were compounds associ-
ated with disruption of the methoxy group, such as catechol
and 3-methoxycatechol. The soda lignin yielded an approxi-
mate equal mixture of coniferyl- and sinapol-derived products.
As evidenced from Table 2, a significant proportion of the
original lignin was recovered as a dark brown or black solid fol-
lowing the APR reaction. These results contrast with the initial
solubility study indicated above, in which considerably smaller
quantities of material were recovered. Although this discrepan-
cy is partially attributed to differences in reaction conditions
and reactor configuration, the higher lignin starting concentra-
tions used in the latter APR reaction also tended to result in in-
creased solid formation relative to the lignin solubilization
study.
Together with the analysis from the solid and liquid isolated
yields, these results suggest that the H2SO4 aids in the disrup-
tion of the ether linkages to form monomeric aromatic com-
pounds and hydrolysis of the methoxy groups to form metha-
nol. The methanol and alcohol groups present on the alkyl
chain of the compounds formed following the disruption of
the ether linkage are then readily reformed on the Pt/Al2O3 cat-
alyst to produce H2 and other light gases. This reforming of
the alkyl chain ultimately results in the formation of simpler ar-
omatic platform molecules, such as syringol or guiaicol, as re-
action products. In the absence of H2SO4, the lignin polymer
largely remains intact, which results in the decrease in both
isolated product and gas yields. In the absence of Pt/Al2O3, the
alkyl chains formed as a result of the disruption of the ether
linkages by H2SO4 instead remain intact, resulting in reduced
yields of simple molecules such as guaiacol and increased
yields of compounds with more complicated alkyl chains.
In addition to the production of monomeric aromatic com-
pounds indicated above, a significant characteristic of the APR
of lignin involves the production of hydrogen and other useful
light gases. The results of the gas phase analysis of the lignin
samples at the conclusion of the reaction are given in Table 3
374
www.chemsuschem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2011, 4, 369 – 378

Lignin Solubilization and Aqueous Phase Reforming
As indicated above, the Pt/Al2O3 catalyst formed as a sepa-
rate solid from the residual lignin, and it was readily isolated
from the solids attributed to lignin. In order to test the recycla-
bility and stability of the catalyst, a portion of this spent Pt/
Al2O3 was collected at the end of the APR of organosolv lignin,
dried, and immediately placed in a clean autoclave with fresh
organosolv lignin. Although slightly decreased gas formation
was observed, the catalyst was still highly active for the forma-
tion of hydrogen and other APR catalysts, resulting in 8.6% H2
relative to 8.8% H2 observed with the fresh catalyst (see Sup-
porting Information for gas distribution versus time and SEM
images of the fresh and spent catalyst). This decrease is possi-
bly caused by small coke deposits as indicated by the darker
color of the spent catalyst relative to the fresh catalyst or by
pitting as a result of Al2O3 exposure to H2SO4. These results
also suggest that the cessation of gas formation after 1.5 h is
attributed to changes occurring to the lignin rather than deac-
tivation of the catalyst.
importance for the depolymerization of lignin in order to pro-
duce monomeric platform chemicals. We therefore synthesized
the lignin model compound depicted in Figure 1 (species B)
and subjected it to APR reaction conditions to test for possible
disruption of the bÀOÀ4 bond. The results of the reaction are
given in Figure 9, where the percent yield of each product de-
tected at the end of the reaction is indicated by the percentag-
es below the compound.
As indicated in Figure 9, none of the starting material, the
bÀOÀ4 model compound, was detected at the conclusion of
the reaction. The most abundant products detected by GC in-
cluded both guaiacol (22.2% yield by mass) and syringol
(22.9% yield by mass), which indicated that the bÀOÀ4 bond
was readily hydrolyzed under reaction conditions to yield mon-
omeric products. These products were then subject to further
reaction to form a variety of other products. As in the case of
the pure guaiacol sample indicated above, a small portion of
the methoxy groups present in the syringol and guaiacol that
were formed as a result of the disruption of the bÀOÀ4 linkage
were hydrolyzed to form 1,2-dihydroxy-3-methoxybenzene
(6.4%) and catechol (1.4%), respectively, along with a corre-
sponding quantity of methanol. The balance of the bÀOÀ4
compound was recovered as unidentified products and solid
residue (138 mg–125 mg catalyst=13 mg solids), which likely
formed via a repolymerization of the components following
disruption of the bÀOÀ4 bond. In addition, small quantities of
ethyl- and methyl-guaiacol possibly formed as a result of hy-
drogenation of the alcohol group on the alkane chain. H2
(7.50%) and CO2 (0.20%) along with small quantities of meth-
ane and ethane (<1%) were produced during the course of
the reaction. The production of these gases greatly exceeded
the case of pure guaiacol (see above), from which H2 is formed
solely through hydrolysis of the methoxy group and subse-
quent reforming of methanol. These results, and the high
abundance of guaiacol (formed from the bÀOÀ4 compound)
suggest that the alcohol-containing alkyl chains of the latter
products are highly susceptible to APR to form H2, CO2, and
guaiacol. This catalytic reforming also occurs at a higher rate
than does the hydrolysis of the methoxy groups as evident by
the higher abundance of guaiacol and syringol relative to cate-
chol and 1,2-dihydroxy-3-methoxybenzene, respectively.
Lignin model compounds
In order to elucidate the possible changes that occur to the
specific linkages and functionality contained in lignin under
the reaction conditions, we extend our study to the APR of
several lignin model compounds. Schematic depictions of the
model compounds are given in the right portion of Figure 1.
The simplest model compound, guaiacol (Figure 1, species A),
contains both hydroxyl and methoxy functionality, the latter of
which is observed in high abundance on the aromatic constit-
uents of lignin. Figure 8 depicts the products observed as a
Figure 8. Aqueous phase reforming of guaiacol by a Pt/Al2O3 catalyst. Reac-
tion conditions: 1.385 g guaiacol, 10.98 g H2O, 0.58 g H2SO4, 0.1245 g 1% Pt/
Al2O3, 29 bar He, 498 K, and 1.5 h.
The appearance of guaiacol formed as a result of the reform-
ing of the alkyl chain of the bÀOÀ4 model compound provid-
ed an indication for the disruption of carbon-carbon bonds
during the APR reaction. As indicated above, a carbon-carbon
bond, in the form of the 5À5’ linkage in lignin, is the second-
most abundant linkage in lignin, and its disruption is therefore
also important for the formation of monomeric aromatic com-
pounds. In order to test the susceptibility of 5À5’ linkages in
lignin to disruption, the model compound depicted in Figure 1
species C was synthesized and subjected to APR. The results
are depicted in Figure 10, with the relative yield percentages
by mass given below the compounds. Analysis of the solution
at the conclusion of the reaction revealed the presence of
methylguaiacol (7%) and guaiacol (14%), which indicated that
the 5À5’ bond was susceptible to disruption under the APR
conditions. This result is significant because the 5À5’ bond is
result of guaiacol APR under standard conditions for 90 min. In
addition to unreacted guaiacol, catechol (5.8% by mass) and a
small quantity of methanol were detected by GC. Similarly,
along with the He initially charged to the autoclave, analysis of
the gas phase at the conclusion of the reaction revealed the
presence of hydrogen and carbon dioxide, which constituted
0.71% and approximately 0.12% of the total gases, respective-
ly. These results suggest that the methoxy groups of guaiacol
and, by extension and as demonstrated above, of lignin are
susceptible to hydrolysis to form methanol and to produce a
hydroxyl group on the aromatic ring. The methanol that forms
as a result is readily reformed to form hydrogen and carbon di-
oxide, as previously demonstrated for pure methanol feeds.[18]
As indicated above, the bÀOÀ4 bond constitutes the most
abundant linkage in lignin, and its disruption is thus of integral
ChemSusChem 2011, 4, 369 – 378
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
375

J. Zakzeski and B. M. Weckhuysen
starting material). The most
abundant compound, however,
was obtained as a result of the
hydrolysis of one or both of the
methoxy-groups (in 42% and
9% yield, respectively) of the 5À
5’ compound to form methanol.
In contrast to the model com-
pounds discussed above, in
which the ether bonds were de-
composed to form methanol or
monomeric compounds, the re-
action products obtained from
both cinnamyl alcohol and ve-
ratryl alcohol (Figure 1, species D
and E, respectively) displayed
evidence of coupling reactions.
In the case of veratryl alcohol,
less than 1% of the original ve-
ratryl alcohol was recovered as
monomeric reaction products,
such as guaiacol, 1,2-dimethoxy-
phenol, and catechol. Instead,
the remainder of the material
formed a black, solid material,
which, as discussed below, re-
sembled in appearance the solid
material collected at the conclu-
sion of the reaction of the differ-
ent lignin compounds. No solids
Figure 9. Products obtained during the aqueous phase reforming of the bÀOÀ4 model compound by a Pt/Al2O3
catalyst. Reaction conditions: 0.043 g bÀOÀ4 compound, 10.98 g H2O, 0.58 g H2SO4, 0.1245 g 1% Pt/Al2O3, 29 bar
He, 498 K, and 1.5 h.
very difficult to break. In contrast to the bÀOÀ4 model com-
pound, a small proportion of unreacted starting material (16%)
was recovered at the conclusion of the reaction along with
some heavy, unidentified products (estimated at 14% of the
were formed when cinnamyl alcohol was used as a substrate.
Instead, a large oily substance was recovered. GC/MS analysis
indicated the presence of several diaryl isomers coupled by
propyl chains completely devoid of oxygen-containing func-
tionality. Since cinnamyl alcohol
only contains a single alcohol
group that was consumed
during the course of the reac-
tion, the reaction stopped at
these dimeric products rather
than the polymeric products ob-
tained with veratryl alcohol. A
large quantity of H2 and CO2
(8.03% and 0.12%, respectively)
was obtained relative to either
guaiacol (0.71% H2 and 0.12%
CO2) or veratryl alcohol (1.89%
H2 and 0.11% CO2), which again
suggests that the reforming of
the side chain in lignin occurs at
a faster rate than does the hy-
drolysis of the methoxy groups
and subsequent reforming of
the resulting methanol. In addi-
tion, the high reactivity of these
Figure 10. Products obtained during the aqueous phase reforming of the 5À5’ model compound by a Pt/Al2O3
catalyst. Reaction conditions: 0.042 g 5À5’ compound, 10.98 g H2O, 0.58 g H2SO4, 0.1245 g 1% Pt/Al2O3, 29 bar He,
498 K, and 1.5 h.
model compounds to form poly-
meric compounds in the case of
376
www.chemsuschem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2011, 4, 369 – 378

Lignin Solubilization and Aqueous Phase Reforming
The lignin solubilization studies were conducted in a semi-batch
200 mL autoclave equipped with quartz windows, thermocouple,
pressure gauge and transducer, magnetic driver (750 rpm), and
back-pressure regulator set at 29 bar. Details of this set-up can be
found in the Supporting Information. Lignin samples were stored
in a desiccator prior to use. During a typical treatment, 2.000 g
lignin (either kraft, alcell, sugarcane bagasse, or soda) was added
to the autoclave with 0.200 g Pt/Al2O3 (1% Pt) and 200 mL H2O.
The autoclave was then sealed, purged and charged with 29 bar
He, and finally heated at approximately 4 KminÀ1 to 498 K. In situ
ATR-IR measurements were recorded using a silicon sentinel
equipped at the bottom of the autoclave.
The aqueous phase reforming reactions were conducted in a semi-
batch 40 mL autoclave equipped with thermocouple, pressure
transducer and gauge, magnetic driver (750 rpm), and back-pres-
sure regulator set at 29 bar. Lignin samples were stored in a desic-
cator prior to use. During a typical reaction, lignin (either kraft,
alcell, sugarcane bagasse, or soda) was added to the autoclave
along with 0.125 g Pt/Al2O3 (1% Pt), 10.98 g H2O, and 0.58 g H2SO4.
The autoclave was then sealed, purged with He, and then 29 bar
He was charged to the autoclave. The autoclave was then rapidly
heated to 498 K in the course of about 15 min. Gas sampling was
conducted using a dual-column Galaxie micro gas chromatography
unit. After the designated time (typically 1.5 h), the autoclave was
cooled in an ice water bath and vented.
veratryl alcohol and dimeric compounds in the case of cinnam-
yl alcohol also provides insight regarding the formation of the
black solids obtained during the lignin APR.
Taken together, the results of the APR of the lignin model
compounds revealed several possible reactions expected to
occur with actual lignin samples. The first is that the most im-
portant linkages in lignin, the bÀOÀ4 and 5À5’, are susceptible
to disruption, although disruption of the former occurs much
more readily than the latter. These disruptions are key steps
necessary for the production of monomeric compounds. Fol-
lowing the disruption of the bÀOÀ4 ether linkage, the alcohol-
containing alkyl side chain is readily reformed to produce H2
and CO2, suggesting the possibility of obtaining guaiacol and
syringol platform chemicals from lignin. The methoxy groups
present in the model compounds were also susceptible to hy-
drolysis to form methanol, which was readily reformed. Even in
the case of the simple model compounds, however, reconden-
sation reactions occurred to yield high molecular weight prod-
ucts and other solids.
Conclusions
At the conclusion of the reaction, the liquid phase was separated
from solids, and finely dispersed solids were isolated by centrifuga-
tion if necessary. Products contained in the liquid phase were iso-
lated by three sequential extractions using approximately 9 g di-
chloromethane, and isolated yields were obtained by removal of
the dichloromethane solvent using a rotary evaporator at 310 K.
The extracted products, which were often of a yellow, oily consis-
tency, were weighed and dissolved in ethyl acetate. Chemical com-
position of the isolated yields was determined by Varian GC
equipped with a VF-WAXms capillary column equipped with an
FID detector. Hexadecane was used as an internal standard. The
quantity of unknown products was estimated using the response
factor determined for vanillyl alcohol. Product identification was
conducted using a Shimadzu GCMS-QP2010 equipped with a VF-
WAXms capillary column and by comparison with pure compounds
when available.
Solids, both residual lignin and the catalyst, were often bound to
the stirrer or the autoclave wall. These materials were meticulously
collected from the autoclave, washed with water, and dried in an
oven at 393 K overnight before weighing.
NMR measurements were conducted using Varian 400 MHz or
Varian 600 MHz spectrometers. The isolated solid lignin was dis-
solved in deuterated DMSO, which was used as the lock source.
Scanning electron micrographs were measured using a Panalytical
Phenom SEM in back scatter mode.
The solubilization and APR reaction of pure lignin samples, in-
cluding soda, kraft, alcell, and sugarcane bagasse, represents a
possible method to obtain hydrogen, light gases, and aromatic
platform molecules, such as guaiacol, syringol and other similar
products. The resulting aromatic products are readily separated
from the aqueous phase by extraction. Analysis of lignin
model compounds suggests that the most abundant linkages
in lignin, the bÀOÀ4 ether and, to a lesser extent, the 5À5’
carbon-carbon linkages, are disrupted under the reaction con-
ditions to form monomeric aromatic compounds. Hydrolysis of
the methoxy functionality on the aromatic rings leads to the
formation of methanol, which is readily reformed to produce
hydrogen and carbon dioxide. The highest yields of products
were obtained in the presence of both H2SO4 and Pt/Al2O3, al-
though products were still obtained in the absence of either
component. The aromatic product distribution, in terms of the
number of methoxy group present on the isolated com-
pounds, depended on the lignin source.
Experimental Section
The bÀOÀ4 model compound was synthesized as previously de-
scribed.[29] During the reaction of the lignin model compounds,
1.385 g of substrate was used instead of lignin samples except in
the case of the bÀOÀ4 and 5À5’ model compounds, in which cases
0.042 g of the substrates were used.
Soda lignin (56.15% C, 6.06% H, 0.79% N, 3.77% S) was obtained
from sisal pulping black liquor, which was acidified with sulphuric
acid to pH 1, filtered through laboratory paper, and dried for 2 h at
328 K. The alcell lignin (66.47% C, 5.96% H, 0.15% N, 27.43% O by
difference) was isolated by the Organosolv extraction method. The
INDULIN AT kraft lignin (63.25% C, 6.05% H, 0.94% N, 1.64% S,
28.12% O by difference) was obtained from pine and is free from
all hemicellulosic materials. The lignin from sugarcane bagasse
(58.90% C, 4.90% H, 0.14% N, 1.53% S, 34.53% O by difference)
was derived from Brazilian sugarcane (see the work of Frollini et al.
for additional characterization information regarding this sub-
strate[27,28]). SEM images of the lignin samples before reaction and
the solid residues collected after reaction are given in the Support-
ing Information.
Acknowledgements
The authors thank the National Science Foundation International
Research Fellowship Program for support of this research under
Award No. 0856754, and Wouter Huijgen and Jaap van Hall
(ECN), Richard Gosselink (Wageningen University), and Matthijs
ChemSusChem 2011, 4, 369 – 378
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
377

J. Zakzeski and B. M. Weckhuysen
[14] M. Osada, N. Hiyoshi, O. Sato, K. Arai, M. Shirai, Energy Fuels 2007, 21,
1854–1858.
[15] M. Osada, N. Hiyoshi, O. Sato, K. Arai, M. Shirai, Energy Fuels 2007, 21,
1400–1405.
[16] M. Osada, O. Sato, K. Arai, M. Shirai, Energy Fuels 2006, 20, 2337–2343.
[17] T. Sato, T. Furusawa, Y. Ishiyama, H. Sugito, Y. Miura, M. Sato, N. Suzuki,
N. Itoh, Ind. Eng. Chem. Res. 2006, 45, 615–622.
Ruitenbeek and Rui Cruz (The Dow Chemical Company) for
kindly supplying the lignin samples. We also thank Pieter C. A.
Bruijnincx, Anna L. Jongerius, Ilona van Zandvoort, and Dilek
Boga (Utrecht University) for many useful discussions, model
compounds synthesis, scanning electron micrograph acquisition,
and assistance with the APR equipment, respectively.
[18] R. D. Cortright, R. R. Davda, J. A. Dumesic, Nature 2002, 418, 964.
[19] R. D. Cortright, J. A. Dumesic, Low-Temperature Hydrogen Production
from Oxygenated Hydrocarbons, US 6.699.457 B692, 2004.
[20] R. D. Cortright, J. A. Dumesic, Low-Temperature Hydrogen Production
from Oxygenated Hydrocarbons, US 6.964.758 B962, 2005.
[21] G. W. Huber, R. D. Cortright, J. A. Dumesic, Angew. Chem. 2004, 116,
1575–1577; Angew. Chem. Int. Ed. 2004, 43, 1549–1551.
[22] M. B. Valenzuela, C. W. Jones, P. K. Agrawal, Energy Fuels 2006, 20,
1744–1752.
[23] I. Hasegawa, Y. Fujii, K. Yamada, C. Kariya, T. Takayama, J. Appl. Polym.
Sci. Appl. Polym. Sci. 1999, 73, 1321–1328.
[24] Y. Matsushita, S. Yasuda, Bioresour. Technol. 2005, 96, 465–470.
[25] C. Heitner, D. R. Dimmel, J. A. Schmidt, Lignin and Lignans: Advances in
Chemistry, CRC Press, Boca Raton, 2010.
Keywords: aqueous phase reforming · aromatics · catalysis ·
hydrogen · lignin
[1] J. Zakzeski, P. C. A. Bruijnincx, A. L. Jongerius, B. M. Weckhuysen, Chem.
Rev. 2010, 110, 3552–3599.
[2] F. G. Calvo-Flores, J. A. Dobado, ChemSusChem 2010, 11, 1227–1235.
[3] R. J. A. Gosselink, E. de Jong, B. Guran, A. Abꢁcherli, Ind. Crops Prod.
2004, 20, 121–129.
[4] E. Dorrestijn, L. J. J. Laarhoven, I. W. C. E. Arends, P. Mulder, J. Anal. Appl.
Pyrolysis 2000, 54, 153–192.
[5] P. F. Britt, A. C. Buchanan, M. J. Cooney, D. R. Martineau, J. Org. Chem.
2000, 65, 1376–1389.
[26] G. Socrates, Infrared and Raman Characteristic Group Frequencies: Tables
and Charts, Wiley, New York, 2001.
[27] J. L. Guimar¼es, E. Frollini, C. G. da Silva, F. Wypych, K. G. Satyanarayana,
Ind. Crops Prod. 2009, 30, 407–415.
[6] M. Misson, R. Haron, M. F. A. Kamaroddin, N. A. S. Amin, Bioresour. Tech-
nol. 2009, 100, 2867–2873.
[7] R. W. Thring, J. Breau, Fuel 1996, 75, 795–800.
[8] W. L. Schinski, A. E. Kuperman, J. Han, D. G. Naae, Process for Generating
a Hydrocarbon Feedstock from Lignin, US 2009/0218062A0218061, 2009.
[9] W. L. Schinski, A. E. Kuperman, J. Han, D. G. Naae, Process for Generating
a Hydrocarbon Feedstock from Lignin, US 2009/0218061A0218061, 2009.
[10] K. Ehara, D. Takada, S. Saka, J. Wood Sci. 2005, 51, 256–261.
[11] M. Saisu, T. Sato, M. Watanabe, T. Adschiri, K. Arai, Energy Fuels 2003, 17,
922–928.
[28] W. Hoareau, F. B. Oliveira, S. Grelier, B. Siegmund, E. Frollini, A. Castellan,
Macromol. Mater. Eng. 2006, 291, 829–839.
[29] J. Zakzeski, A. L. Jongerius, B. M. Weckhuysen, Green Chem. 2010, 12,
1225–1236.
[12] M. Osada, T. Sato, M. Watanabe, T. Adschiri, K. Arai, Energy Fuels 2004,
18, 327–333.
[13] T. Furusawa, T. Sato, M. Saito, Y. Ishiyama, M. Sato, N. Itoh, N. Suzuki,
Appl. Catal. A 2007, 327, 300–310.
Received: September 13, 2010
Revised: October 28, 2010
Published online on January 18, 2011
378
www.chemsuschem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2011, 4, 369 – 378