Catalytic depolymerization of alkali lignin in subcritical water: Influence of formic acid and Pd/C catalyst on the yields of liquid monomeric aromatic products

Article abstract of DOI:10.1039/c4gc00854e

Alkali lignin was subjected to depolymerization in subcritical water at 265 °C, 6.5 MPa for reaction times between 1-6 h in a batch reactor and in the presence of formic acid (FA) and Pd/C catalyst. The oil products were extracted into diethyl ether and contained >90% of single-ring phenolic compounds. The reaction of lignin in subcritical water alone yielded 22.3 wt% oil containing 56% guaiacol as the main product. A maximum oil yield of 33.1 wt% was obtained when the lignin was reacted in the presence of formic acid alone. In the presence of FA, catechol became the predominant compound, with more than 80% of the ether extract after 6 h. The conversion of guaiacol to catechol in the presence of formic acid suggested the hydrolysis of O-CH₃ ether bonds. In addition, the yields of alkyl guaiacols increased in the presence of FA. The use of 5 wt% Pd/C catalyst with FA slightly decreased the yields of oil products but led to compounds indicative of hydrogenolysis of aryl-O ether bonds as well as hydrogenation of CC bonds.

Full text of DOI:10.1039/c4gc00854e

Green Chemistry
View Article Online
View Journal | View Issue
PAPER
Catalytic depolymerization of alkali lignin in
subcritical water: influence of formic acid and
Pd/C catalyst on the yields of liquid monomeric
aromatic products†
Cite this: Green Chem., 2014, 16,
4740
Jude A. Onwudili* and Paul T. Williams
Alkali lignin was subjected to depolymerization in subcritical water at 265 °C, 6.5 MPa for reaction times
between 1–6 h in a batch reactor and in the presence of formic acid (FA) and Pd/C catalyst. The oil pro-
ducts were extracted into diethyl ether and contained >90% of single-ring phenolic compounds. The
reaction of lignin in subcritical water alone yielded 22.3 wt% oil containing 56% guaiacol as the main
product. A maximum oil yield of 33.1 wt% was obtained when the lignin was reacted in the presence of
formic acid alone. In the presence of FA, catechol became the predominant compound, with more than
80% of the ether extract after 6 h. The conversion of guaiacol to catechol in the presence of formic acid
suggested the hydrolysis of O–CH3 ether bonds. In addition, the yields of alkyl guaiacols increased in the
presence of FA. The use of 5 wt% Pd/C catalyst with FA slightly decreased the yields of oil products but
led to compounds indicative of hydrogenolysis of aryl–O ether bonds as well as hydrogenation of CvC
bonds.
Received 9th May 2014,
Accepted 13th August 2014
DOI: 10.1039/c4gc00854e
www.rsc.org/greenchem
of organic solvents as depolymerization media.5,13 The appli-
cation of ionic liquids and organic solvents as the reaction
1. Introduction
Lignin makes up to 30 wt% of lignocellulosic biomass and medium for lignin depolymerization has been reported to give
nearly 40% of its energy content, hence its conversion and utili- high conversion rates, with good yields of aromatic chemicals.
zation will contribute to the efficiency, sustainability and viabi- However, there are issues regarding the recovery of these sol-
lity of the lignocellulosic biorefinery industry. Besides, based vents and the use of fossil-based solvents which may impact
on the structure of lignin, it is potentially a likely renewable on the sustainability of the process.
source of aromatic chemicals. This is probably driving the
Hydrothermal conversion involving water and added acid,
increasing research on the valorization of lignin for chemical base and metal catalysts has been proposed for lignin depoly-
and liquid fuel production around the world in recent years.1–8 merization since most lignin streams come with very high
In this regard, different thermochemical methods of depoly- moisture contents, thus avoiding a drying step; and depending
merizing lignin into high-value aromatic chemicals are being on the reaction conditions, char formation may be
investigated, including fast pyrolysis, wet oxidation and hydro- suppressed.14–18 Hydrothermal media presents
a benign
thermal processing. One of the challenges of thermochemical environment for green chemistry and lignin depolymerization
processing of lignin is the rampant re-polymerization of pro- research has been progressing in this direction. In hydrother-
ducts of depolymerization, which are themselves highly reac- mal conditions, water can act as a reaction medium, catalyst
tive under reaction conditions. This often results in extensive and reactant. In particular, lignin depolymerization in the sub-
char formation, leading to poor yields of aromatic chemicals. critical region of water is seen as a greener process and has
A number of processes have been investigated to minimize become increasingly more attractive to researchers than super-
char formation including the use of ionic liquids9 and capping critical water due to the lower temperatures and pressures
agents such as boric acid10 and phenol8,11,12 as well as the use used which may reduce costs and offer thermodynamic
limitations to char formation. Results from hydrothermal
depolymerization in literature show that base-catalyzed depoly-
merization (BCD) using alkalis can effectively convert lignin
Energy Research Institute, School of Process, Environmental and Materials
Engineering, University of Leeds, Leeds, LS2 9JT, UK.
into a variety of aromatic chemicals but with the yields of
E-mail: j.a.onwudili@leeds.ac.uk
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
monomeric aromatic products are still at around 20%.14 As a
result, hydrothermal lignin conversion is now being investi-
c4gc00854e
4740 | Green Chem., 2014, 16, 4740–4748
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
gated with or without metal-based catalysts, with water alone each test have been carried out in this regard. The liquid pro-
or with organic co-solvents, particularly ethanol. For instance, ducts here refer to contents of the liquid residual extracted
Wahyudiono et al.,15 obtained up to 35 wt% of monomeric into diethyl ether solvent.
phenols with catechol as the major product during the
decomposition of lignin in near- and supercritical water from
350–400 °C in a batch reactor. Although, they were unable to
measure the gas products, they found that changes in water
2. Results and discussion
2.1. Hydrothermal reactions of formic acid
density affected the yields and composition of the phenolic
compounds with increasing reaction times. Pińkowska et al.6
reported a maximum combined yield of 11.23 wt% for guaia-
col, catechol, phenol and cresol isomers at 280 °C (zero
holding time) in the absence of any catalysts. Zakzeski et al.16
obtained 17 wt% yield of monomeric aromatic oxygenates
such as guaiacols during liquid-phase reforming of Organosolv
lignin in a water–ethanol mixture at 225 °C, 58 bar with
Pt/Al2O3 catalyst. They further hydrotreated the liquid products
with external oxygen supply to obtain 6% aromatic and cyclic
hydrocarbons.
The use of biomass-derived simple carboxylic acids for
hydrothermal lignin conversion has also been receiving atten-
tion. Formic and acetic acids can be produced from biomass
and lignin under certain hydrothermal conditions16–18 and
using them for lignin processing into valuable products will
improve sustainability. Besides, formic acid can easily be
reformed into hydrogen19 required for the hydrogenolysis
and/or hydrodeoxygenation of lignin into valuable chemi-
cals.20 Xu et al.20 used formic acid (FA) for the depolymeriza-
tion and hydrodeoxygenation of Organosolv lignin obtained
from switch grass in the presence of 20 wt% Pt/C catalysts at
280 °C. They obtained 21 wt% yields of seven monomeric
phenols after 4 h of reaction and concluded that reaction time
was critical to lignin conversion under the investigated
conditions.
In this present study, alkali lignin has been subjected to
subcritical water depolymerization reactions with the appli-
cation of formic acid (FA) as capping agent and hydrogen
source and 5 wt% palladium supported on carbon (Pd/C) as
hydrodeoxygenation (HDO) catalyst at 265 °C, 6.5 MPa. Pd/C
has been used in literature for hydrodeoxygenation of bio-oils
and oxygenates in alcohols21 and a recent publications showed
that Pd/C can be effective for deoxygenation of triglycerides in
subcritical water with external hydrogen supply.22 At first, the
hydrothermal reaction of FA was investigated at the given reac-
tion conditions to determine the influence of reaction time
and the presence of Pd/C catalyst on its ultimate conversion
according to eqn (1):19
It was important to understand the hydrothermal reactions of
formic acid with or without the Pd/C in order to understand
its activities in lignin conversion. The results of the hydro-
thermal reactions of 1.0 g of FA at 265 °C, with reaction time
from zero to 6 h, both in the presence and absence of Pd/C are
presented in Table 1. The results show that both reaction time
and Pd/C catalyst influenced the hydrothermal conversion of
formic acid into hydrogen and carbon dioxide according to
eqn (1).19
Eqn (1) shows that the standard enthalpy change (ΔH2°98 K
)
.
for the hydrothermal reaction of formic acid is +32.1 kJ mol−1
This indicates that the position of equilibrium would shift to
favour the yield of products with increase in temperature.
The yields of CO2 and hydrogen have been used to estimate
the conversion of FA during these reactions, since literature
and this work show that complete reaction of FA would yield
about 50 vol% each of CO2 and H2. Therefore, barring experi-
mental errors, and some dissolution of gases (especially CO2)
in the aqueous residuals, the experimental yields of both gases
can be directly used to determine FA conversion. As shown in
Table 1, the decomposition of FA increased with time, with
only 24.9% conversion when the reaction was stopped immedi-
ately the reaction temperature was reached (zero min hold
time). However, after 1 h hold time at the reaction tempera-
ture, the conversion reached 89%, followed by complete con-
version after 3 h of reacting FA alone.
Conversely, the presence of Pd/C led to complete conversion
of FA even at zero minute reaction time. The results in Table 1
also show that yields of gases in the presence of Pd/C at 0 min,
1 h and 3 h reaction times, respectively were similar, indicat-
ing that the Pd/C catalyst catalyzed FA conversion in subcritical
water but did not consume any hydrogen produced from
it. The solid catalyst was separated by filtration, dried and
Table 1 Yields of hydrogen and CO2 in relation to time during hydro-
thermal reactions of formic acid at 265 °C, 6.5 MPa
Hold
CO2
H2
CO2
H2
Conversion
time (min) (vol%) (vol%) (wt%) (wt%) (wt%)
HCOOH ! CO2 þ H2
ð1Þ
FA
FA + Pd/C
FA
FA + Pd/C
FA + Pd/Ca
FA
FA + Pd/C 180
FA 360
0
0
60
60
60
180
66.1
51.5
47.4
50.3
50.2
50.8
51.1
50.3
50.2
31.0
48.5
52.6
49.4
49.8
49.2
48.9
49.7
49.8
21.3
94.3
84.7
94.8
94.5
94.5
94.4
95.4
96.9
0.45
3.97
4.28
4.10
4.11
4.16
4.12
4.27
4.31
24.9
98.3
89.0
98.9
98.6
98.7
98.5
99.7
Thereafter, the separate and combined influence of the FA
and Pd/C on the yields and variety of liquid depolymerization
products from the lignin sample was investigated. This work
was aimed at preliminary understanding of the interactions
between the lignin and the additives which influenced the
yields of liquid products under the hydrothermal conditions
studied. Hence, detailed analyses of the liquid products from
FA + Pd/C 360
101
a Catalyst re-used once.
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 4740–4748 | 4741

View Article Online
Paper
Green Chemistry
weighed. Up to 99.3 wt% of the catalyst was recovered indicat- product mass balances gave very good closures of >94 wt% in
ing little losses during these reactions. Furthermore, results of all cases. Mass balances >100 wt% were only obtained during
the reaction of FA alone and FA in the presence of Pd/C for 6 h the reaction of lignin in the presence of both FA and Pd/C at
also gave very similar yields of CO2 and hydrogen. This is extended hold times, which may indicate the participation of
further evidence of the stability of the Pd/C catalyst; in particu- water during the reactions, especially the water–gas shift reac-
lar, it can be concluded that the Pd/C did not contribute to gas tion.23 A test with only Pd/C (without lignin and FA) was
yields. The recovered catalyst for the 1 h run was re-used with carried out for 6 h and neither gas nor liquid products were
results showing similar gas yields. More future work will be obtained, instead 98.5 wt% of the catalyst was recovered.
carried out on the structural stability of the Pd/C catalyst
under hydrothermal conditions.
In Table 2, liquid products refers the combined yields of oil
(ether-extractable liquid products) and water-soluble products.
Considering that it took 12 min to heat the reactor to The water-soluble products were in the form of semi-solids,
265 °C, the results from the reactions at zero min holding time which only appeared when the aqueous residuals had been
can be used to estimate the specific rate of conversion of the shaken with diethyl ether during liquid–liquid extraction. They
formic acid without and with the Pd/C;
extracted into the ether but remained as dissolved solids and
could easily re-dissolve in water. These were separated from
the ether phase by filtration, air-dried and weighed. The
extracted aqueous phases were clear and nearly colourless
after 6 h but at lower reaction times, particularly after 1 h, the
extracted aqueous phases had light brown colours.
Specific rate of FA conversion ðmol hꢀ1
Þ
Mole of FA converted
Reactor Heat-up Time ðhÞ
¼
The reactions of lignin alone at 265 °C, 6.5 MPa for 1–6 h
reaction hold times were first studied for reference. In this
tests, the yields of liquid products decreased while the yields
of solid residues and gas increased with increasing reaction
time. The formation of solid residue was evidence of the rapid
polymerization of lignin and possibly, its degradation products
to form char. Wahyudiono et al.15 obtained up to 25 wt%
solid residue yields from alkali lignin after 4 h, which is
lower than the yields of solid residues (31 wt% after 1 h to
Clearly, without the Pd/C the specific rate of FA conversion
was 0.024 mol h−1, while in the presence of the catalyst, the
specific rate increased 4.5 times to 0.108 mol h−1. Hence, the
Pd/C has very high catalytic effect on FA conversion and this
may affect the depolymerization of lignin with these additives.
2.2. Product yields from lignin depolymerization
A total of 12 experiments were carried out with lignin as feed.
In these tests, the lignin sample was reacted alone and with 46 wt% after 6 h) in this present study. The differences in solid
the addition of formic acid and Pd/C catalyst for reaction hold
times of 1, 3 and 6 h, respectively. The yields of products desig-
residue yields may be attributed to the types and sizes of reac-
tors used. Wahyudiono et al.15 used a 5 ml Hastelloy C-276
nated as liquid, solid residue and gas from these tests are pre- reactor with water densities of 0.16–0.67 g ml−1, while in this
sented in Table 2. The percentage yields of products were
calculated based on the amount of lignin and FA used, where
work, a 500 ml stainless steel reactor was used, with a fixed low
water density of 0.12 g ml−1. The highest gas yield of 17 wt%
applicable, since the preceding section 2.1 showed that FA was obtained after reacting for 6 h and the gas comprised
reacted to produce H2 and CO2 under the test conditions. The
mainly carbon dioxide (83.4 vol%), methane (10.2 vol%) and
hydrogen (5.84 vol%). The high yield of methane suggested the
cracking of C–C linkage in lignin. While the yield of liquid pro-
ducts decreased, the proportion of ether-extractable oils in the
liquid products increased with reaction time, indicating the
possible refining of the liquid products by heat. On the basis of
the lignin feed, extracted oil was 5.23 wt% after 1 h, increased
to 13.9 wt% after 3 h and finally to 22.3 wt% after 6 h. The
yields of the main components in the oil products are presented
and discussed in section 2.3.
Table 2 Product yields from lignin depolymerization with additives at
265 °C, 6.5 MPa from 1–6 h hold time
Solid
residue
(wt%)
Hold
time (h)
Liquid
(wt%)
Gas
(wt%)
%
Balance
Sample
Lignin
Lignin
Lignin
1.0
3.0
6.0
58.2
51.6
33.3
30.6
35.9
46.0
5.56
9.12
17.0
94.4
96.6
96.3
2.2.1. Effect of FA. From the preceding section, 89% of FA
was converted after 1 h of reaction hold time at 265 °C,
6.5 MPa, therefore experiments were carried out with lignin
and formic acid from 1–6 h reaction hold times under the
same conditions. The amounts of char/solid residue recovered
from these increased slightly with reaction times from 0.64 wt%
after 1 h to 1.12 wt% after 6 h. As the hydrothermal reaction
of FA in section 2.1 did not produce any solid residue, any
solid residue obtained during the reaction of lignin and FA
was therefore attributed to lignin. Compared to the reaction of
lignin alone, the presence of FA decreased the yield of solid
Lignin/FA
Lignin/FA
Lignin/FA
1.0
3.0
6.0
61.6
52.3
41.1
0.64
0.19
1.12
36.5
46.1
58.3
98.7
98.6
101
Lignin/Pd/C
Lignin/Pd/C
Lignin/Pd/C
1.0
3.0
6.0
41.3
36.1
35.6
54.4
57.3
59.1
0.84
1.25
2.25
96.5
94.7
97.0
Lignin/FA-Pd/C
Lignin/FA-Pd/C
Lignin/FA-Pd/C
1.0
3.0
6.0
45.8
38.5
31.6
16.3
18.3
22.9
38.7
49.8
53.4
101
107
108
Pd/C
6.0
—
98.5
—
98.5
4742 | Green Chem., 2014, 16, 4740–4748
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
residue by up to 98%, showing that FA helped in solubilizing hydrogen utilization during lignin depolymerization. Hydro-
lignin, so that the water-soluble products could contain both gen consumption was possibly via the participation of FA in
depolymerized and unreacted lignin. This indicated the ability lignin hydrolysis as well as in other possible reactions requir-
of formic acid to act as a capping agent to prevent or delay re- ing hydrogen, such as hydrogenation and hydrogenolysis.
polymerization of products. Hence, nearly 98 wt% conversion However, after reacting for 6 h, the yield of hydrogen increased
of the original lignin mainly into liquid and gas products, was to 0.032 g, which was still 26.4% less than its theoretical yield
obtained in these tests, which agreed with some reported from FA but about 28% more than its yield during the experi-
results in literature.24
ment for 1 h. This may indicate hydrogen production from the
The yield of oil products increased with time and was reforming of small organic molecules such as methanol that
highest after 6 h. Again based on the dry weight of lignin, oil could form from the hydrolysis of O–CH3 ether bonds.
yields more than doubled from 15.3 wt% to 33.1 wt% between
2.2.2. Effect of Pd/C. The use of Pd/C alone during lignin
1 h and 6 h, which compares well to the work of Xu et al.20 depolymerization produced the lowest yields of gas and the
who used formic acid and Pt/C for lignin depolymerization, highest yields of solid residues. CO2 was the only gas detected
although at a slightly higher temperature of 280 °C. In the even after 3 h, and after 6 h the gas composition changed
opposite direction, the yield of water-soluble products slightly to 97 vol% CO2, 0.37 vol% H2 and 2.23 vol% methane.
decreased with time, showing that some of the water soluble Yet the total gas yield was only 2.25 wt% of lignin feed
products were being converted into char and extractable oil after 6 h. Apparently, Pd/C catalyzed the polymerization of
products.
lignin to form char, so that the oil products obtained was only
The gas products from these tests mainly composed of CO2, 0.21 wt% after 1 h and increased to just 5.10 wt% after reacting
CO and hydrogen. The trends in the yields of gases are pre- for 6 h.
sented in Fig. 1, with CO2 as the predominant gas product.
2.2.3. Effect of combining formic acid and Pd/C catalyst.
A large proportion of the yields of CO2 and H2 must have When Pd/C and formic acid were combined for lignin depoly-
arisen from the hydrothermal reaction of formic acid.19 merization, oil yield was much higher than with Pd/C alone
However, compared to the gas yields from FA and lignin when but was lower than when formic acid alone was used, still indi-
reacted separately, the total gas yields obtained when FA and cating that Pd/C influenced the formation of char. Oil yield
lignin reacted together were consistently lower at the different increased from 7.60 wt% at 1 h to 25.7 wt% after 6 h reaction
residence times. This observation indicated the occurrence of time.
liquid–liquid reactions between FA and lignin, thereby reduc-
Total gas yields were similar to that obtained during the
ing the overall gas-forming potential. With increasing reac- reaction of lignin in the presence of FA alone. However, there
tion times, the yields of hydrogen and CO increase slightly but were slight differences in the overall gas compositions, which
consistently, while the yields of CO2 gave the most dramatic suggested evidence of the participation of Pd/C in the yields
increase, indicating possible reforming reactions. CO yields and compositions of the gas products. Fig. 2 shows that there
increased from about 3 wt% after 1 h to 5.23 wt% after 6 h. was a consistent decrease in CO yields which may indicate its
Considering that no CO was produced during the reactions of conversion to hydrogen via water–gas shift (WGS) reaction.
FA alone under similar conditions, it can be concluded that Indeed, CO yields decreased from 2.84 wt% after 1 h to 0.98 wt%
the interaction between FA and lignin led to CO production. after 6 h, while hydrogen yield increased from 0.71 wt% to
The yield of hydrogen was consistently≈1 wt% of the gas 1.2 wt% after 6 h. Compared to the reaction of lignin with FA
product. After 1 h and 3 h, hydrogen yields were 0.023 and alone, which produced increasing yields of CO, which reached
0.028 g, respectively, which were much lower than its theo- 5.23 wt% at 6 h, these results thus show that the presence of
retical yield (0.0435 g) from FA based on eqn (1), suggesting Pd/C influenced WGS reaction.
Fig. 2 Gas yields from lignin depolymerization with FA in the presence
Fig. 1 Gas yields from the depolymerization of lignin with FA only.
of Pd/C.
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 4740–4748 | 4743

View Article Online
Paper
Green Chemistry
The yields of CO2 was again much higher, reaching
48.5 wt% after 6 h, also suggesting the possible conversion of
lignin and its decomposition products with increasing hold
time. Hydrogen yield was 0.028 g and 0.035 g at 1 h and 3 h,
respectively but increased to 0.048 g after reacting for 6 h.
Hence, hydrogen yield increased by 9.4% of the theoretical
yield after 6 h in the presence of Pd/C indicating that the cata-
lyst may have promoted both the reforming of small molecules
and the water–gas shift reaction. In addition, hydrogen yields
during lignin depolymerization with the combination of FA
and Pd/C were consistently higher than hydrogen yields when
only FA was reacted with lignin. In section 2.1 Pd/C has been
shown to catalyze FA decomposition but the lower yields of
hydrogen obtained here after 1 h and 3 h may indicate that the
presence of lignin altered the catalytic effect of Pd/C on FA due
to competition for active sites by both FA and lignin. However,
it is clear from these results that the metal catalyst promoted
gas-phase hydrogen production and therefore, could have
reduced the extent of liquid–liquid interactions between FA
and lignin. This would influence the hydrolytic ability of FA
and therefore may affect the yields of oil products.
Fig. 4 GC/MS/MS chromatogram of oil product from lignin depolymeri-
zation with formic acid and Pd/C.
from the uncatalyzed and catalyzed (FA and Pd/C) lignin
depolymerization tests.
The yields of the major components of the oil products
obtained from the depolymerization of lignin after from 1 h to
6 h reaction time are shown in Fig. 5. In all cases, these com-
pounds made up more than 90% of the oil products. In the
absence of any additives, lignin depolymerization in subcriti-
cal water produced only 5.23 wt% of ether-extractable oil after
1 h hold time. This increased to 13.9 wt% after 3 h and finally
to 22.3 wt% oil after 6 h. Guaiacol was consistently the
dominant product and after 6 h, the oil comprised of 56%
guaiacol, 25% catechol and 8% phenol and 5% alkyl guaiacols.
This corresponded to a guaiacol/catechol yield ratio of 2.2.
While, it was earlier observed that Pd/C catalyzed the
decomposition of FA, these results showed that the catalysis of
Pd/C on FA may have been inhibited by the presence of lignin.
In this sense, the presence of lignin could block some active
sites in Pd/C, preventing the immediate catalytic decompo-
sition of FA. However, the oil yields from lignin in the presence
of FA and Pd/C were consistently lower compared with those
obtained with lignin and FA alone. The oil yields on lignin
feed-basis increased from 7.60 wt% after 1 h to 20.7 wt% after
3 h and then to 25.7 wt% after 6 h. Hence, the rate of depoly-
merization increased dramatically between 1 h and 3 h but
increased at slower rate from 3 h up to 6 h. With the combi-
nation of FA and Pd/C, the pattern of product yields similar to
when only FA was used, was obtained. Therefore, results in
Fig. 5[B] and 5[C] suggest that the rate of guaiacol conversion
to catechol only increased between 3–6 h. This may indicate
that the hydrogen produced from FA became more active in
the presence of Pd/C. The yield ratio of guaiacol/catechol after
3 h was 1 : 1.28. After 6 h reaction time, the yield ratio of cate-
chol and guaiacol was in favour of catechol at 1 : 4.6.
2.3. Detailed analysis of the oil products
A clearer picture of the predominant reaction pathway for the
effect of FA could be seen from compositional analysis of the
oil products. In this section, the yields of oils and their com-
ponents have been calculated on the basis of the dry weight of
lignin feed (1.0 g). The major compounds in the ether extracts
were guaiacol, 1,2-benzenediol (catechol), methylguaiacol, ethyl-
guaiacol and propylguaiacol. In addition, minor compounds
identified from the GC/MS/MS analysis included phenol,
dimethoxybenzene, dimethoxytoluene, eugenol, 2,3,6-tri-
methylguaiacol, vanillin and 2′,4′-dihydroxy-3′-methylpropiophe-
none. Indeed, some of the oil products gave a pleasant smell as
a result of the presence of some of these compounds which are
commonly used in flavourings. Fig. 3 and 4 show the anno-
tated GC/MS/MS chromatogram for the ether extractable oil
From the results, it is apparent that the presence of formic
acid altered the reaction pathway in favour of catechol over
guaiacol after long reaction hold times. This agrees with the
recent work of Forchheim et al.25 who studied the kinetic model-
ling of hydrothermal lignin depolymerization. They found
Fig. 3 GC/MS/MS chromatogram of oil product from lignin depolymer-
ization without any additives.
4744 | Green Chem., 2014, 16, 4740–4748
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
However, the increase in the yield of alkylguaiacols
suggested that the mechanism of lignin depolymerization
alone and with FA may also involve hydrogenolysis or cracking
of C–C bonds along the propane units in lignin. The higher
yields of these compounds in the presence of FA at longer
reaction times suggested possible significant contributions
from hydrogenolysis. This is underscored by the improved
yield in oil, indicating that formic acid also influenced the
overall depolymerization process. The decrease in the total
yield of oil in the presence of Pd/C and FA may be explained by
the influence of the catalyst on the conversion of FA as
shown in Table 1 and the promotion of re-polymerization reac-
tions by Pd/C. The faster rate of conversion of FA may prevent
its hydrolytic ability in depolymerizing the lignin. It may also
decrease its capping ability to prevent re-polymerization of
products.
In essence, the hydrolytic depolymerization of the lignin by
FA would occur as long as FA remained and could provide H+
ions. Hence, its hydrolytic ability would be impaired as soon
as it is converted to hydrogen and CO2. It would therefore
mean that lignin conversion to water-soluble products resulted
from hydrolysis, and this was followed by the conversion of the
water-soluble products to oil products. The conversion of the
water-soluble products was time-dependent as shown by
results of increased oil yield with reaction time. The formation
of alkyl guaiacols suggested that during the reaction, the
hydrogen gas produced was able to participate in the conver-
sion of the water-soluble products to oil products even without
a hydrogenation catalyst. These results indicate that substan-
tial hydrothermal depolymerization of lignin can be achieved
via hydrolysis but the use of hydrothermally reactive additives
that can also generate in situ hydrogen for HDO of liquid pro-
ducts would require careful reactor design. This will be necess-
ary to allow them sufficient time to act as depolymerization
agents before acting as hydrogen donors for HDO.
2.4. Hydrothermal reactions of guaiacol
To further understand the reaction schemes of the two addi-
tives with guaiacol five sets of experiments were carried out for
6 h reaction time. From the reactions of lignin, 6 h reaction
time gave the highest yields of oil and highest yields of cate-
chol in the presence of FA. Hence, a reference experiment was
carried out with 1.0 g of guaiacol for 6 h, in the absence of any
additives. In this test, 97% of the original guaiacol was re-
covered. Only small amounts of phenol and dimethoxybenzene
(ca. 0.2 wt% each) were identified but no catechol was found.
Then the same amount of guaiacol was reacted with formic
acid for 6 h and this yielded 82.5 wt% of catechol and 15.3 wt%
guaiacol. Further, the combination of formic acid and Pd/C
was used to investigate the reaction of guaiacol from 1–6 h as
presented in Fig. 6. Results show that guaiacol was highly
resistance to degradation after 1 h of reaction even in the pres-
ence of these additives, so that 95 wt% guaiacol was recovered.
It was only after 3 h of reacting that the formation of catechol
was observed. However, the yield of catechol again increased
dramatically beyond 3 h reaction time, so that the composition
Fig. 5 Yields of aromatic compounds during lignin depolymerization in
relation to reaction hold time [A] lignin alone; [B] lignin with FA [C] lignin
with FA and Pd/C.
that, the yields of catechols increased with increasing tempera-
ture and residence time. It would appear that the active reac-
tion pathway involving formic acid was via hydrolytic cleavage
of the O–CH3 ether linkage to produce catechol. The alterna-
tive mechanism of hydrogenolysis using in situ produced
hydrogen from formic acid would invariably lead to methane
formation if the O–CH3 bond was cleaved or to substantial
phenol formation on cleaving the aryl–OCH3 bond. Little or no
methane was observed during the reactions of the alkali lignin
in the presence of formic acid.
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 4740–4748 | 4745

View Article Online
Paper
Green Chemistry
Though the yields of these hydrogenation products were low,
their presence indicated the possibility of catalytic hydrogen-
ation or hydrodeoxygenation of oxygenates such as guaiacol.
3. Conclusion
The results presented here show the possibility of alkali lignin
depolymerization under acidic conditions using a carboxylic
acid that can be derived from biomass. The yields of mono-
meric phenolic compounds are considerable compared with
what is presently possible in literature. The presence of formic
acid influenced both the extent of depolymerization and the
composition of the oil products. Without any additive/catalyst,
lignin depolymerized in subcritical water to produce mainly
guaiacol. However, in the presence of formic acid, catechol
became the predominant phenolic compound possibly due to
hydrolysis of the O–CH3 ether bond in guaiacol. In addition,
Pd/C appeared to catalyse the hydrogenolysis of the aryl–O
ether bond resulting in significant yield of phenol from guaia-
col also promoted char formation. The continued presence of
significant amount of water-soluble products after 6 h indi-
cated that the reaction time could further be manipulated and
the use of appropriate catalysts can improve yields. Moreover,
there are many variables still to be considered in future
research including the use of different lignin samples based
on sources and methods of isolation, use of different operating
conditions of temperature, pressure and reaction time as well
as the application of different catalysts/additives for hydrother-
mal lignin conversion.
Fig. 6 Yield of aromatic compounds in relation to time from the hydro-
thermal reaction of guaiacol in the presence of FA and Pd/C.
Fig. 7 GC/MS/MS chromatograms of oil products from the hydrother-
mal reactions of (a) guaiacol alone; (b) guaiacol with formic acid and Pd/
C (6 h hold time).
4. Experimental
Samples of alkali lignin (low sulfonate grade), formic acid
(99% purity), diethyl ether (solvent) and Pd/C (5 wt%) were
obtained from Sigma-Aldrich UK and used as received. The
proximate composition of the alkali lignin was; 3.23% moist-
ure, 3.57% ash, 60.6% volatiles and 32.6% fixed carbon; while
the elemental analysis results gave 61.7% carbon, 5.11%
hydrogen, 31.3% oxygen, 1.22% nitrogen and 0.67% sulfur.
The reactor used was a stainless steel batch reactor with
500 ml volume capacity and can reach a maximum tempera-
ture and pressure of 500 °C and 30 MPa, respectively.26
of oil was 65.7 wt% yield of catechol, 24.9 wt% of the original
guaiacol and 3.01 wt% yield of phenol after 6 h.
The production of phenol indicated the cleavage of aryl–O
bond via hydrogenolysis possibly catalysed by Pd/C using
the in situ generated hydrogen. Moreover, the high yield of
catechol also suggested the cleavage of O–CH3 bond via acid
catalysed hydrolysis.
These results therefore suggest that while formic acid
favoured hydrolysis of the O–CH3 ether linkage, the Pd/C
catalyst appeared to favour hydrogenolysis to form phenol and
possibly methanol, which would be a substantial component
of the aqueous residuals or could be reformed to hydrogen
and CO2 at extended reaction hold times.
Fig. 7 shows the two chromatograms of oils from these reac-
tions. It can be seen from Fig. 7b that the presence of Pd/C
also influenced the formation of hydrodeoxygenation pro-
ducts. Examples of these included 1,1′-bicyclohexyl, 2-propenyl
cyclohexane and cyclohexyl benzene. However, their yield was
only 0.13 wt% after 1 h and increased to 1.72 wt% after 6 h.
In each experiment, 1.0 g of alkali lignin was added into
the reactor along with 60 ml of deionized water. This amount
of the lignin sample completely dissolved in the water. The
volume of water was chosen to give a pressure of 6.5 MPa at
the chosen reaction temperature of 265 °C. When required,
1.0 g of formic acid and 0.5 g of Pd/C were added either separ-
ately or in pre-determined combinations. The reactor vessel
was closed and purged for 10 minutes with nitrogen gas at a
flow rate of 1 ml min−1. The reactor was then heated to 265 °C
in 12 min and held at this temperature for 1–6 h. At the end of
the heating, the reactor was quickly withdrawn from the
heated and cooled to room temperature with a large fan within
4746 | Green Chem., 2014, 16, 4740–4748
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
one hour. The liquid products were extracted with diethyl
ether and analyzed using a was a Varian CP-3800 gas
chromatograph coupled with a Varian Saturn 2200 mass
spectrometer (GC/MS/MS) (please see ESI† for details). The
weight of solid residue was determined after drying in an oven
at 105 °C and for the experiments involving Pd/C, temperature
programmed oxidation (TPO) method was used to determine
the amount of solid residue produced.
4.1. Analysis of products
4.1.1. Gas analysis. On cooling, the gas samples (especially
when formic acid had been used) were sampled for offline ana-
lysis. The analytical protocol used for gas products analysis
has been widely published.27,28 Briefly, the gas samples were
analysed using three packed column gas chromatographs. The
permanent gases, hydrogen, oxygen, nitrogen and carbon
monoxide, were separated on a molecular sieve column and
analysed using a Varian CP-3380 gas chromatograph with a
Thermal Conductivity Detector (GC/TCD). Hydrocarbon gases,
C1 to C4, were separated on a Haysesp column and analysed
using a second Varian CP-3380 gas chromatograph with a
Flame Ionisation Detector (GC/FID). Carbon dioxide was ana-
lysed using a third gas chromatograph fitted with a TCD. The
results obtained from the GCs were given as a volume percent
and were converted into masses of each gas using the ideal gas
equation.
After gas analysis, the reactor was opened to sample the
contents (liquid and solids) into a beaker and weighed. The
reactor contents were transferred into a clean glass bottle,
followed by rinsing with 30 ml of diethyl ether as solvent and
re-weighed. The mixtures were filtered under vacuum to
separate the solids from the liquids. The solids retained on the
filter paper were dried in an oven at 105 °C for 2 h and
weighed. The weight of solid residue was subtracted from the
weight of the reactor contents to obtain the weight of the
liquid products (less weight of solvent used).
Fig. 8 Picture showing the different fractions of the liquid products;
(a) extracted aqueous phase, (b) clear ether extract and, (c) water-
soluble semi-solid fraction from ether phase.
(GC/MS/MS) fitted with a 30 m DB-5 equivalent column. For
the GC/MS/MS analysis, 2 μl of the diethyl ether extract were
injected into the GC injector port at a temperature of 290 °C;
the oven programme temperature was 35 °C for 8 min, then
ramped to 120 °C at 5 °C min−1 heating rate, held for 1 min
and ramped at 4 °C min−1 to 210 °C and finally ramped at
20 °C min−1 to 280 °C (total analysis time of 55.5 min). The
transfer temperature line was at 280 °C, manifold at 120 °C
and the ion trap temperature was held at 200 °C. The ion trap
was initially switched off for 8 min to allow the elution of the
solvent prior to data acquisition to safeguard the life of the
trap. The compounds present in the extracts were quantified
by internal standard method with 2-hydroxyacetophenone as
internal standard (IS). List of compounds and the properties
used for their quantitative analysis are shown in the ESI.†
References
In all cases, the liquid products were partitioned by liquid–
liquid extraction technique in a separating funnel with the
organic layer separated into diethyl ether. The addition of two
drops of hydrochloric acid (1.0 M) helped to ensure complete
separation of the distinctive phases. The aqueous phases were
subsequently extracted twice more with 15 ml of diethyl ether.
For extended reaction times, the aqueous phases were largely
clear while the ether phases were unclear and required cen-
trifugation to separate the clear liquid from a dark brown
semi-solid product, most of which dissolved when shaken with
water. Fig. 8 shows the typical phases of the liquid products;
(a) the extracted aqueous phase, (b) the clear ether extract, and
(c) water-soluble semi-solid fraction from ether phase. The
clear ether (oil) portion had a colour range between light
yellow to brown. The weight of the oil fractions were deter-
mined after vacuum evaporation. The ether extracts were com-
bined and dried over anhydrous sodium sulfate prior to
analysis by GC/MS/MS.
1 M. Kleinert and T. Barth, Chem. Eng. Technol., 2008, 31(5),
736–745.
2 M. Kleinert, J. R. Gasson and T. Barth, J. Anal. Appl. Pyro-
lysis, 2009, 85, 108–117.
3 C. A. Mullen and A. A. Boateng, Fuel Process. Technol., 2010,
91, 1446–1458.
4 D. J. Nowakowski, A. V. Bridgwater, D. C. Elliott, D. Meier
and P. Wild, J. Anal. Appl. Pyrolysis, 2010, 88, 53–72.
5 R. J. A. Gosselink, W. Teunissen, J. E. G. van Dam, et al.,
Bioresour. Technol., 2012, 106, 173–177.
6 H. Pińkowska, P. Wolak and A. Złocińska, Chem. Eng. J.,
2012, 187, 410–414.
7 Q. Song, F. Wang and J. Xu, Chem. Commun., 2012, 48(56),
7019–77021.
8 A. Toledano, L. Serrano and J. Labidi, Fuel, 2014, 116, 617–
624.
9 B. J. Cox and J. G. Ekerdt, Bioresour. Technol., 2012, 118,
584–588.
The equipment used was a Varian CP-3800 gas chromato- 10 V. M. Roberts, V. Stein, T. Reiner, A. Lemonidou, X. Li and
graph coupled with a Varian Saturn 2200 mass spectrometer
J. A. Lercher, Chem. – Eur. J., 2011, 17, 5939–5948.
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 4740–4748 | 4747

View Article Online
Paper
Green Chemistry
11 M. Saisu, T. Sato, M. Watanabe, T. Adschiri and K. Arai, 21 T. H. Parsell, B. C. Owen, I. Klein, T. M. Jarrell,
Energy Fuels, 2003, 17, 922–928.
12 K. Okuda, M. Umetsu, S. Takami and T. Adschiri, Fuel
Process. Technol., 2004, 85, 803–813.
C. L. Marcum, L. J. Haupert, L. M. Amundson,
H. I. Kenttämaa, F. Ribeiro, J. T. Miller and
M. M. Abu-Omar, Chem. Sci., 2013, 4, 806–813.
13 Q. Song, F. Wang, J. Cai, et al., Energy Environ. Sci., 2013, 22 S. A. W. Hollak, M. A. Ariëns, K. P. de Jong and D. S. van
6(3), 994–1007. Es, ChemSusChem, 2014, 7(4), 1057–1062.
14 J. E. Miller, L. Evans, A. Littlewolf and D. E. Trudell, Fuel, 23 T. Yoshida, Y. Oshima and Y. Matsumura, Biomass Bio-
1999, 78(11), 1363–1366. energy, 2004, 26, 71–78.
15 Wahyudiono, M. Sasaki and M. Goto, Chem. Eng. Process., 24 J. R. Gasson, D. Forchheim, T. Sutter, U. Hornung, A. Kruse
2008, 47, 1609–1619. and T. Barth, Ind. Eng. Chem. Res., 2012, 51, 10595–10606.
16 J. Zakzeski, A. L. Jongerius, P. C. A. Bruijninc and 25 D. Forchheim, U. Hornung, A. Kruse and T. Sutter,
B. M. Weckhuysen, ChemSusChem, 2012, 5(8), 1602–1609.
17 F. Jin, Z. Zhou, A. Kishita and H. Enomoto, J. Mater. Sci.,
2006, 41, 1495–1500.
18 J. Yun, F. Jin, A. Kishita, K. Tohji and H. Enomoto, J. Phys.:
Conf. Ser., 2010, 215, 012126.
Waste Biomass Valorization, 2014, DOI: 10.1007/s12649-014-
9307-6.
26 R. Muangrat, J. A. Onwudili and P. T. Williams,
Int. J. Hydrogen Energy, 2012, 37(3), 2248–2259.
27 J. A. Onwudili and P. T. Williams, RSC Adv., 2013, 3, 12432–
12442.
19 N. Akiya and P. E. Savage, AIChE J., 1998, 44(2), 405–415.
20 W. Y. Xu, S. J. Miller, P. K. Agrawal and C. W. Jones, 28 J. A. Onwudili and P. T. Williams, Green Chem., 2010, 12,
ChemSusChem, 2012, 5, 667–675. 2214–2224.
4748 | Green Chem., 2014, 16, 4740–4748
This journal is © The Royal Society of Chemistry 2014