Catalytic hydroprocessing of lignin under thermal and ultrasound conditions

Article abstract of DOI:10.1016/j.cattod.2012.02.051

Lignin isolated from Miscanthus × giganteus using acidic (FAL) and basic (AL) conditions was thereafter subjected to catalytic depolymerization under thermal or ultrasonic activation. The characterization of lignin samples was achieved by thermogravimetric analysis and FTIR. Three different classes of catalysts containing nickel as the active species have been prepared in this respect: (i) nano-Ni(0) by reduction of NiCl₂ with NaBH₄ under ultrasonication, (ii) Fe₃O₄-(NiMgAlO)_x and (NiAlO)_x by calcination of Mg(Ni)-Al hydrotalcite incorporating Fe₃O₄, followed by reduction with hydrogen, and (iii) NiO(1 1 1) nanosheets by reduction of Ni(NO₃)₂ with urea using benzyl alcohol as a structure directing agent. The catalysts were characterized by XRD and XPS techniques. Reduced mixed oxides displayed a moderate activity while a significant increase in conversion (up to 90%) was observed with nano-Ni(0) and NiO(1 1 1) nanosheets catalysts. The conversion and the mass distribution of the reaction products were strongly related to the procedure used for the extraction of lignin. In all the reactions AL samples had a greater extent of depolymerization. The performance from the tested catalysts under ultrasonic conditions was inferior to those tested under conventional heating conditions. The nature of the solvent was also found to be very important to this process, with the ionic liquid, 1-butyl-3-methylimidazolium acetate ([BMIM]OAc) demonstrating the best results in autoclave conditions and methanol under ultrasonic conditions.

Full text of DOI:10.1016/j.cattod.2012.02.051

Catalysis Today 196 (2012) 3–10
Contents lists available at SciVerse ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Catalytic hydroprocessing of lignin under thermal and ultrasound conditions
a
a
b
c
Kenneth B.H. Finch , Ryan M. Richards , Aurore Richel , Andrei V. Medvedovici ,
d
e
e
e,∗
Nicoleta G. Gheorghe , Marian Verziu , Simona M. Coman , Vasile I. Parvulescu
Department of Chemistry and Geochemistry, Colorado School of Mines, 304 Coolbaugh Hall, 1012 14th St, Golden, CO 80401, USA
Unit of Biological and Industrial Chemistry, University of Liege – Gembloux Agro-Bio Tech, Passage des Déportés, 2 – B-5030 Gembloux, Belgium
Department of Analytical Chemistry, Faculty of Chemistry, University of Bucharest, Bdul Regina Elisabeta, 4-12, Bucharest 030016, Romania
National Institute of Materials Physics, Atomistilor 105b, 077125 Magurele-Ilfov, Romania
a
b
c
d
e
Department of Organic Chemistry, Biochemistry and Catalysis, Faculty of Chemistry, University of Bucharest, Bdul Regina Elisabeta, 4-12, Bucharest 030016, Romania
a r t i c l e i n f o
a b s t r a c t
Article history:
Lignin isolated from Miscanthus × giganteus using acidic (FAL) and basic (AL) conditions was thereafter
subjected to catalytic depolymerization under thermal or ultrasonic activation. The characterization of
lignin samples was achieved by thermogravimetric analysis and FTIR. Three different classes of catalysts
containing nickel as the active species have been prepared in this respect: (i) nano-Ni(0) by reduction of
NiCl2 with NaBH4 under ultrasonication, (ii) Fe3O4–(NiMgAlO)x and (NiAlO)x by calcination of Mg(Ni)–Al
hydrotalcite incorporating Fe3O4, followed by reduction with hydrogen, and (iii) NiO(1 1 1) nanosheets
by reduction of Ni(NO3)2 with urea using benzyl alcohol as a structure directing agent. The catalysts were
characterized by XRD and XPS techniques. Reduced mixed oxides displayed a moderate activity while a
significant increase in conversion (up to 90%) was observed with nano-Ni(0) and NiO(1 1 1) nanosheets
catalysts. The conversion and the mass distribution of the reaction products were strongly related to
the procedure used for the extraction of lignin. In all the reactions AL samples had a greater extent of
depolymerization. The performance from the tested catalysts under ultrasonic conditions was inferior
to those tested under conventional heating conditions. The nature of the solvent was also found to be
very important to this process, with the ionic liquid, 1-butyl-3-methylimidazolium acetate ([BMIM]OAc)
demonstrating the best results in autoclave conditions and methanol under ultrasonic conditions.
Received 31 October 2011
Received in revised form 6 January 2012
Accepted 12 February 2012
Available online 21 March 2012
Keywords:
Lignin
Alkaline treatment
Acidic treatment
Depolymerization
Nano-catalysts
Low-molecular compounds
©
2012 Elsevier B.V. All rights reserved.
1
. Introduction
thermochemical treatments as pyrolysis, hydrogenolysis, hydrol-
ysis, gasification and oxidation [2–7]. Pyrolysis of lignin leads to
oils, gases as well as solid char. The proportion of these products
depends on several factors such as feedstock type [8], heating rate,
reaction temperature [2] and the presence or absence of a catalyst
[9]. Mullen and Boateng [9] showed that in the presence of the cat-
alysts HZSM-5 or CoO/MoO3 the aromatic fraction is increased at
The continual consumption of petroleum reserves and the
growing concerns over the emission of greenhouse gases have stim-
ulated intense research to investigate the potential of biomass as a
renewable alternative for fuel and/or chemical production.
Currently, lignocellulose is the most abundant component of
biomass (20–28% tree composition) and is composed of three
types of polymers: cellulose, hemicelluloses and lignin. Lignin
is potentially an important renewable raw material for the pro-
duction of multiple products including fuels and high-value
chemicals. However, depolymerization of lignin (composed from
different phenylpropene monomer units as coniferyl, sinapyl and
p-coumaryl alcohol [1]) is not an easy task and many efforts
are focused on this challenge. The reported methods to convert
this natural aromatic polymer into smaller molecules utilize such
◦
650 C. In supercritical water with Ni/MgO or Co/MgO the selectiv-
ity of the reaction changes in favor of the production of hydrogen
and the performance depends on the catalyst calcination tempera-
ture [10]. Depolymerization of wheat stalk powder was also carried
out in supercritical ethanol using CaO/Al O and CaO/Al O /Fe O
2
3
2
3
3
4
catalysts [11]. The presence of these catalysts increased the conver-
sion from 12.0% without a catalyst to 29.0% using CaO/Al O3 and
2
42.4% using CaO/Al O /Fe O .
2
3
3
4
Oxidation of lignin may also offer a versatile route to aromatic
aldehydes. A LaFe1 CuxO3 catalyst was shown to enhance the
−x
selectivity and could be recycled for several runs without signif-
icant deactivation [12]. In the absence of a heterogeneous catalytic
species, a pH close to 14 must be employed to obtain similar results
[13]. However, such conditions are difficult from both an economic
∗ Corresponding author. Tel.: +40 21 4100241; fax: +40 21 4100241.
E-mail addresses: simona.coman@g.unibuc.ro (S.M. Coman),
vasile.parvulescu@g.unibuc.ro (V.I. Parvulescu).
0
920-5861/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.cattod.2012.02.051

4
K.B.H. Finch et al. / Catalysis Today 196 (2012) 3–10
and an environmental point of view. Dissolving lignin in ionic liq-
uids [14] followed by oxidation over transition metals [15] may
present an alternative route to aromatic aldehydes.
FAL sample, centrifugation followed by lyophilisation. The obtained
sample is named AL (Ammonia Lignin).
Noble metals can also catalyze the conversion of lignin into
alkanes and methanol in a two step process. First, the conversion of
lignin to monomers and dimers was achieved by a catalytic cleav-
age of the C–O–C bonds (without disrupting the C–C linkages) using
a series of catalysts carbon supported including Ru/C, Pd/C, Rh/C
and Pt/C, under modest H2 pressures. Subsequently, the obtained
monomers and dimers were further hydrogenated over Pd/C with
good yields of alkanes and methanol [16]. The main drawback of
this process is the high price of the noble metals.
2.2. Catalyst preparation
The nano-Ni catalyst was prepared by adding 21.5 g NiCl2·6H2O
dissolved in 300 mL methanol to a mixture of 27.5 mL N2H4·H2O,
26 mL distilled water and 2.5 g NaOH. The obtained mixture was
ultrasonicated for 1.5 h in the presence of 0.8 g of NaBH4. The
formed black precipitate was thereafter washed with distilled
water and dried at room temperature.
The solvation of lignin is important to achieve a high conver-
sion rate. It has been demonstrated that certain ionic liquids can
be used as “green” aprotic solvents for lignin. For the [BMIM]
Fe3O4–(NiMgAlO)x and (NiAlO)x particles containing 30 wt%.
Ni were prepared by calcination of Mg(Ni)–Al hydrotalcite pre-
cursors. An aqueous solution containing 1.92 g of Ni(NO3)2·6H2O,
2.56 g Mg(NO3)2·6H2O and 2.06 g of Al(NO3)3·9H2O or 5.82 g of
Ni(NO3)2·6H2O and 2.47 g of Al(NO3)3·9H2O was added to a Fe3O4
suspension with a Ni/Fe molar ratio of 50. The obtained slurry was
+
–
containing ionic liquids, Ragauskas et al. [14] have shown
that the solvation of lignin varies with the anion in the order:
−
−
−
−
6
[
MeOSO3] > Cl > Br ≫ [PF ] indicating that the solubility of
◦
lignin was principally influenced by the nature of the anions. Ionic
liquids containing large, non-coordinating anions [PF4] and [PF6]
stirred at 60 C and a 1 M solution of NaOH/Na2CO3 was added
−
−
dropwise until a pH of 10 was reached. The resultant mixture was
◦
were unsuitable as a solvent for lignin.
transferred into an autoclave and hydrothermally treated at 200 C
The aim of the present study was to depolymerize lignin under
mild conditions (thermal (e.g., autoclave) and ultrasound condi-
tions), using nickel as the active species in both the reduced (Ni)
and oxidized state ((NiAlO)x and (NiMgAlO)x). For comparison,
NiO(1 1 1) nanosheets with hexagonal holes [17] have been tested
in the same reaction. The catalytic experiments were carried out
in autoclave and ultrasounds activation conditions using as sol-
vents water, organic solvents and ionic liquids. Another scope of
this study was to optimize the analytical conditions for an accurate
characterization of the process.
for 10 h. After cooling to room temperature, the obtained hydrotal-
◦
cite was washed with distilled water to neutral pH, dried at 80 C
◦
◦
for 24 h, at 130 C for 4 h and then calcined at 500 C for 4 h in air
[20]. Finally, the resulted mixed oxides were treated in a hydrogen
−
1
◦
flow (30 mL min ) at 180 C for 3 h.
The NiO(1 1 1) nanosheets were prepared accordingly to a
methodology described by Hu et al. [17]. 9 g of Ni(NO3)2·6H2O was
dissolved in 100 mL absolute methanol, then 1 g urea and 6.7 g ben-
zyl alcohol were added, in the ratio Ni:urea:benzyl alcohol = 1:0.5:2
(molar ratio). After stirring for 1 h, the solution was transferred to
an autoclave, purged with 7500 Torr Argon 5 times, before a pres-
sure of 7500 Torr Argon was imposed prior to heating. The mixture
◦
◦
2
. Experimental
was heated to 200 C for 5 h, and then increased to 265 C for 1.5 h.
Finally, the reaction mixture was supercritically dried by allowing
◦
2.1. Lignin preparation
the solvent to vent at 265 C. A green powder was collected and
subsequently calcined in air at 500 C for 6 h.
◦
Miscanthus × giganteus from a crop cultivated in spring 2007,
and harvested after two years (Tournai, Belgium) was used as the
raw material after the stems were stripped of leaves, air-dried
and ensiled. The material was ground into 1–2 mm particles. The
untreated Miscanthus was mainly composed of 24.5% lignin, 48.4%
glucose, 15.7% xylose, 1.9% arabinose, 1.2% galactose, 0.2% mannose,
2.3. Catalyst and lignin characterization
The prepared catalysts were characterized using several tech-
niques. XPS spectra were obtained at room temperature using
a PHI 5600ci Multi Technique Instrument. The pressure in the
analysis chamber was 1.33 mPa. Monochromatic AlK␣ radiation
(hꢀ = 1486.6 eV) was generated by a beam current of 12 mA with
an acceleration voltage of 10 kV. X-ray diffraction (XRD) patterns
were obtained with a Shimadzu XRD 7000 diffractometer using
6
.4% extractives and 2.4% ash [18].
The extraction of lignin from Miscanthus × giganteus was car-
ried out using either a formic acid/acetic acid/water mixture
(
acidic medium) or an aqueous ammonia solution (25 wt%) (basic
medium). The protocol used for the extraction of lignin in acidic
◦
◦
medium was based on a methodology described by Lam et al.
CuK␣ radiation, from 5 to 90 . The crystallite size was deter-
mined from the X-ray line broadening using the Scherrer formula
given by D = 0.9␭/ˇ cos ꢁ, where D is the average crystallite size,
◦
[
19] with some modifications: the dry lignin was soaked at 50 C
for 30 min in a formic acid/acetic acid/water mixture (30/50/20,
◦
˚
vol/vol/vol ratio) and then the temperature was increased to 107 C
ꢂ is the X-ray wavelength used (1.5406 A), ˇ is the angular line
width of half maximum intensity and ꢁ is the Bragg’s angle in
degree.
for 3 h with continuous agitation. The ratio of dry matter to soaking
liquid was 1/25. After cooling, the pulp was filtered out with the aid
of a vacuum filter funnel (500 mL, 95 mm diameter) assembled with
a fritted disk (40–100 ␮m pore size Robu Glasfilter-Geräte GmbH)
to black liquor. Recovery of lignin was done by precipitation from
The extracted lignin samples (FAL and AL) were characterized
by TG/TDA and FTIR techniques in order to investigate the effect of
the pretreatments on the chemical structure and physico-chemical
characteristics. Thermogravimetric analysis of the lignin samples
was performed using a SDT Q600 instrument supplied by TA Instru-
ments. The samples were placed in an alumina sample holder and
black liquor by adding H SO4 (72%, w/w) until a pH of 1.5 was
2
reached. The precipitated lignin was then isolated by centrifugation
and dried using a lyophilization technique. The obtained sample is
named FAL (Formic Acid Lignin).
◦
−1
◦
from room temperature to 600 C
heated at a rate of 5 C min
under N2 atmosphere with a rate flow of 20 mL min
−
1
The extraction of lignin in basic medium occurred by soaking
.
◦
raw material with an ammonia aqueous solution (25 wt.%) at 60 C
About 5 mg of dried lignin was mixed and ground with potas-
sium bromide and then mechanically pressed into pellet form. FTIR
spectra were obtained using a Bruker IFS 48 spectrometer from
for 6 h in a water bath [18]. The ammonia from black liquor obtained
◦
after pulp filtration was vacuum-evaporated at 40 C. The recov-
−
1
−1
ery of lignin was accomplished by using a method similar to the
2000 to 400 cm with 16 scans at a resolution of 4 cm .

K.B.H. Finch et al. / Catalysis Today 196 (2012) 3–10
5
2.4. Catalytic tests
Typical depolymerization experiments were carried out using
.25 g lignin in 30 mL water or organic solvent (methanol, ethyl
0
acetate) and 0.03 g catalyst while bubbling hydrogen under
ultrasonic conditions for 6 h, at room temperature, using an
Elma Transsonic 460/H bath at a frequency of 35 kHz. Alter-
natively, 0.16 g of lignin dissolved into 2 g [BMIM]OAc ionic
◦
liquid was heated to 180 C under 1–10 atm of H2 for 24 h with
0.02 g catalyst in an autoclave. [BMIM][OAc] was selected from
a preliminary lignin solubility screening using the following
IL’s: 1-hexyl-3-methylimidazolium trifluoromethanesulfonate
(
[HMIM][CF3SO3]), 1-butyl-2,3-dimethylimidazolium tetrafluoro-
borate ([BM2IM][BF4]), 1-butyl-4-methylpyridinium hexafluoro-
phosphate ([BMPY][PF6]), 1-butyl-3-methylimidazolium hex-
afluorophosphate ([BMIM][PF6]), 1-butyl-3-methylimidazolium
chloride ([BMIM][Cl], 1-butyl-3-methylimidazolium bromide
Fig. 1. X-ray powder diffraction patterns of the prepared catalysts: (a) NiO(1 1 1);
(
b) nano-Ni; (c) Fe3O4–(NiAlO)x; (d) Fe3O4–(NiMgAlO)x, P–MgO (periclase), B–NiO
([BMIM][Br], and 1-butyl-3-methylimidazolium acetate ([BMIM]
OAc]).
(
bunsenite).
[
After the ultrasonic experiments in water, the unreacted lignin
was recovered by centrifugation of the mixture at 6000 rpm for
0 min. In the case of ionic liquid and organic solvents, after reaction
the non-transformed lignin was precipitated by adding distilled
water and separated by centrifugation.
The conversion of lignin to low molecular weight compounds
was calculated using Eq. (1):
Miller indices (1 1 1), (2 0 0) and (2 2 2) [21] and reveal the exist-
ence of a pure face-centered cubic (fcc) phase (JCPDS-004-0850).
No lines assigned to oxidized phases or other nickel compounds
have been detected. The crystallinity is enough high to eliminate
the presence of amorphous phases. Sonochemistry is known to pro-
vide a green alternative to the synthesis of nanomaterials with a
good control of particle size and morphology [22].
1
ꢀ
ꢁ
m − m
i
f
C% = 100 −
× 100
(1)
The structure of the NiO nanosheets was also confirmed by
the XRD patterns (Fig. 1a). The diffraction lines of (1 1 1), (2 0 0),
(2 2 0), (3 1 1), and (2 2 2) faces were identified from the d spac-
m
i
where m , mass of lignin introduced in reaction; m , mass of lignin
i
f
ings of 2.4049, 2.0826, 1.4742, 1.2584, and 1.2051
A˚ , respectively
recovered after reaction.
The analysis of the reaction products was made using LC–MS
technique using an Agilent 1100 Series LC/MS equipped with
a degasser (model G1379A), quaternary pump (model G1311A),
automated injector (model G1329A), autosampler thermostat
[17] (JCPDS-001-1239). They correspond to a single phase of well-
crystallized NiO with the face-centered cubic structure (space
group: Fm3m) [23]. Coherent scattering regions for nickel oxide
were determined by the Scherrer equation [24] based on broaden-
ing of diffraction lines (2 0 0), (1 1 1) for NiO. Average size of NiO
particles was found as being 19 nm.
The powder XRD patterns of Fe3O4–(NiAlO)x (Fig. 1c) and
Fe3O4–(NiMgAlO)x (Fig. 1d) show diffraction lines correspond-
ing to Fe3O4 (2ꢁ = 30.19, 35.57, 43.17, 57.12 and 62.79) and NiO
(2ꢁ = 37.11, 43.64, 63.09, 75.67 and 79.89). Magnesium and alu-
minum are silent in these patterns [25].
(
model G1330BA), column thermostat (model G1316A), detector
DAD (model G1315B), detector Triple Quad LC/MS (model G2445D),
interface APCI (model G1947A), interface ESI (model G1948A), and
column: Zorbax Eclipse XDB C18, lengh 50 mm, 4.6 mm i.d., 1.8 ␮m
d.p. (sterically protected with di-isobutyl n-octadecylsilane). The
◦
operating conditions were temperature 25 C, mobile phase: sol-
vent A: aq. 0.1% formic acid and solvent B: 0.1% formic acid in
acetonitrile, gradient profile time 0 min: 5% solvent B; linear to
The reduction of the mixed metal catalysts incorporating Fe3O4
with hydrogen does not change the XRD patterns.
−
1
time 20 min: 0% solvent B, flow rate: 1 mL min , and injection
volume: 1 ␮L. MS analysis conditions were: APCI: polarity posi-
tive, corona voltage 4000 V, nebulising gas pressure 60 psi, drying
However, XPS analysis provided advanced information about
the nature of the samples after the reduction in hydrogen. Fig. 2
shows the XPS spectra in the region of the Ni2p binding energy
for nano-Ni(0) and NiO(1 1 1). While the energy determined for
NiO(1 1 1) corresponds to Ni2+ (854 eV) [26] the binding ener-
gies determined for the nano-Ni(0) indicated the presence of
both Ni2+ (854 eV) and Ni(0) (852.5 eV) levels. The reduction of
Fe3O4–(NiAlO)x and Fe3O4–(NiMgAlO)x in hydrogen led only to a
partial reduction of nickel, and the intensity of the Ni2p3/2 bands
−
1
◦
gas flow 5 L min , vaporizer temperature 400 C, drying gas tem-
◦
perature 350 C, capillary voltage 3500 V; ESI: polarity negative,
capillary voltage 3500 V, nebulising gas pressure 65 psi, drying gas
−1
◦
flow 12 L min , drying gas temperature 350 C; Ion trap: scan
1
1
00–3000 m/z; maximum accumulation time, 200 ms, ICC target
0,000. The content of different fractions was quantified as a ratio
of the intensity of the areas of the peaks corresponded to m/z in the
indicated limits for reacted lignin to the fresh one. Since the conver-
sion of the fractions with m/z higher that 1000 was always smaller
than 5%, we will report only the evolution of the fractions with m/z
smaller than 1000. Due to some similitude the report was divided
in three m/z range categories: 100–500, 500–700 and 700–1000.
2+
assigned to Ni(0)-to-Ni had approximately a ratio of 1:3. The
deconvoluted XPS spectra for Ni(0) and Ni2+ of the two samples
are shown in Figs. 1S and 2S (supporting information).
3.2. Lignins extraction and characterization
3
. Results and discussion
As its polymeric structure is composed of a variety of monomer
units, lignin is considered recalcitrant against most chemical
treatment and a major obstacle in processing of plant biomass.
Moreover, the isolation method of lignin has an influential role in
determining its nature and structure.
3.1. Catalysts characterization
XRD patterns of nickel catalysts indicated that the preparation
procedure led to pure 10 nm nanoparticles (Fig. 1b). The charac-
teristic diffraction lines at 2ꢁ = 44.4 , 51.7 and 76.5 correspond to
Vanderghem et al. [18] have shown that depending on the condi-
tions of the pretreatments, different degrees of delignification were
◦
◦
◦

6
K.B.H. Finch et al. / Catalysis Today 196 (2012) 3–10
composition of the extracted lignin samples. The exceptions being
−
1
the bands located in the carbonyl region at 1760 and 1650 cm
,
where FAL exhibits a higher relative intensity compared to AL. This
difference was already assigned to the acid hydrolysis of lignin
generating new carbonyl functional groups [30].
Combining these characterizations with HSQC analysis Van-
derghem et al. [18] proposed different linkages for AL and FAL. The
weight average (Mw) of acetylated lignin samples was reported to
−
1
−1
be 2760 g mol for FAL lignin and 3140 g mol for AL lignin. Tak-
ing into account that the molecular weight of a phenylpropane
−
1
unit is 180 g mol , the polymerization degree of FAL can be
approximated at 15, while the polymerization degree of AL was
approximated at a value of 17.
3.3. Analytic measurements
The RPLC mechanism separates compounds in the increasing
order of their lipophilic character. The gradient profile starts at 5%
acetonitrile in water with 0.1% formic acid and ends at 100% ace-
tonitrile (also containing 0.1% formic acid) over 20 min. Formic acid
suppresses ionization of acidic phenolic moieties, avoiding peak
broadening and symmetry distortion upon chromatographic elu-
tion.
Fig. 2. XPS spectra in the region of the Ni2p binding energy.
When observing the UV traces at 252 nm, the depolymerised
lignin (reaction products) contains a huge number of compounds
eluting up to minute 14 (consequently, such compounds are quite
polar in nature), while the lignin raw material indicates a less
complex pattern, with a group of compounds eluting between
minutes 10 and 12. The “hill” profile of the depolymerised lignin
obtained from a variety pulps. The acidic pretreatment led to exten-
sive, 86.5%, delignification, while the basic pretreatment led to only
3
6.3% delignification. The elevated temperatures used in the acidic
(
reaction products) between minutes 6 and 14 (similar to GC chro-
pretreatment may explain the greater extent of delignification by
allowing more lignin to be solvated. Moreover, lignin obtained from
acidic pretreatment was more pure, likely due to the dissociation
of carbohydrates in the lignin fraction. Therefore, the lignin con-
tent, determined by the Klason procedure, was 80% after the acidic
pretreatment but only 70% after the ammonia pretreatment [18].
In both FAL and AL samples small amounts of carbohydrates (e.g.,
pentoses and hexoses) were evidenced and their content depended
on the applied pretreatment method. In FAL samples, the level of
these compounds was much lower (3.0 wt%) compared to AL sam-
ples (13.7 wt%). These differences were explained before [18] by
intensive hydrolysis of lignin–carbohydrates bonds resulting in low
carbohydrate content for the FAL samples. AL lignin had a notice-
able amount of xylan in composition suggesting that xylose may
link to lignin by bonds that are resistant to cleavage under basic
conditions.
−
1
matograms of petroleum fractions) indicates that the 4.75% min
gradient being applied is still too fast to produce a baseline sep-
aration between compounds generated by depolymerisation. It is
worthwhile to note that in the conditions of the RPLC separation
(even allowing elution with acetonitrile for additional 40 min after
the end of the gradient) it is possible that the main congeners from
lignin with high molecular masses (>3000) are not eluted from the
column at all.
MS detection applied to the lignin raw material and depoly-
merised lignin (reaction products) revealed that intense responses
are obtained under positive (+) APCI and negative (−) ESI con-
ditions. Surprisingly, positive (+) ESI leads to poor responses,
disappearing in the baseline noise. APCI mode indicates that the
separated compounds have less polar moieties, due to either the
loss of phenolic groups or that the ionization is inhibited by the
presence of the formic acid. The (−) ESI probably detects com-
pounds having phenolic groups with a low pKa. Note that the
mass scan was limited to 3000 amu. The averaged mass spectra
obtained over the whole chromatogram (monitored either in (+)
APCI or (−) ESI conditions) indicates that compounds eluting near
Thermogravimetric analysis indicated the three stages of
decomposition reported early by part of the group [18]. First stage,
◦
at temperatures below 200 C, corresponds to evaporation of light
◦
compounds (mainly water); the second stage, between 200 C and
00 C, corresponds to the liberation of hydrocarbons from the
◦
5
2
0 min (almost the end of the gradient profile) have masses in the
50–350 amu range, while compounds eluting during the isocratic,
00% acetonitrile, step exhibited masses greater than 350 amu.
rapid thermal decomposition of hemicelluloses, cellulose and a
◦
portion of lignin, and the third stage, above 500 C, to the steady
1
decomposition of the heavier components of lignin [27]. These
results are consistent with the reactivity of the constituent parts
These affirmations are based on the assumption that little molec-
ular fragmentation arises in the APCI sources due to milder the
ionization conditions. The groups of peaks eluting after 20 min have
similar MS patterns that have characteristic signals separated by
[
28,29]. Hemicellulose is the most reactive compound while lignin
is more thermally stable and starts to decompose at higher tem-
◦
peratures (between 250 and 500 C).
fixed distances (44 corresponding to a CO loss, 28 for CO, or 18 for
FTIR spectra were also very similar to the spectra reported by
Vanderghem et al. [18]. They contain bands characteristic of car-
bonyl functional groups, of un-conjugated carbonyl groups and
conjugated carbonyl groups (due to both acid hydrolysis and ester-
ification that can take place during the delignification process), of
aromatic ring vibrations of the phenylpropene groups, of syringyl
units and guaiacyl units. Despite minor deviations, FAL and AL
spectra were rather similar, indicating a comparable chemical
2
water could be observed). This indicates that compounds within
a group belong to homologous series differing by a carboxyl or
a carbonyl group. Compounds detected under (−) ESI conditions
were lighter ones, as their chromatographic retention time is less
than 20 min. For exemplification, illustrative figures showing APCI
and ESI mass spectra were included in the supplementary material
(Figs. 3S and 4S).

K.B.H. Finch et al. / Catalysis Today 196 (2012) 3–10
7
Fig. 5. Depolymerization of lignin samples under ultrasound conditions (catalyst:
Fig. 3. Depolymerization of lignin under autoclave conditions (catalyst: 0.02 g nano
Ni, 0.16 g lignin, IL: 2 g [BMIM]OAc, 10 atm H2, 180 C, 24 h).
0.27 g nano-Ni, 0.25 g lignin, 30 mL water, 6 h).
◦
3
3
.4. Catalytic behavior
responsible for this behavior. It is worth mentioning that the
amount of the catalyst used under ultrasonic conditions was almost
10 times greater than that used in an autoclave.
.4.1. Depolymerization on Ni(0)
Blank experiments in the absence of the catalysts showed that
the lignin was fragmented neither under autoclave nor under the
ultrasound activation.
Fig. 3 shows the conversion of AL and FAL, and distribution of
products with masses below 1000 amu on the nano-Ni(0) catalyst.
Under the indicated conditions, the conversion of AL reached a
higher value (around 90% of the initial amount of lignin was depoly-
merized) compared to FAL. 60% of the products shown are between
However, it is difficult to compare the two reaction method-
ologies. Under autoclave conditions no or very poor results were
obtained by using an organic solvent in which lignin is not soluble
or only partly soluble. In both cases, in the reaction medium, only
a dispersion of two solids (lignin and catalyst) has been observed.
Therefore, almost no interactions between lignin and catalyst will
be possible, because the solid particles of the lignin, consisting of
polymeric chains, are not able to enter the pore system. Under auto-
clave conditions the interaction of lignin with the catalyst become
possible since the lignin is soluble in IL. Contrarily, under ultra-
sounds conditions, solvents in which lignin is only partly soluble
or totally insoluble can be used as long as the lignin particles are
fragmented under the ultrasounds effects.
On the other side, it is well known that in sonochemistry, in
most of the cases, high temperatures are not necessary to acceler-
ate a chemical process. Under autoclave conditions the activation is
different, and therefore, because of the high cross-linked polymer
form, the lignin breakdown requires elevated temperatures. Work-
ing under ultrasounds conditions, our experimental set-up does not
support a high pressure as it is the case of autoclave.
1
00 and 500 m/z. This behavior is supported by the structures pro-
posed by Vanderghem et al. [18] for AL and FAL samples. Nickel
may more easily catalyze the hydrogenolysis of ethers in AL, than
the ␣,␤-unsaturated esters in FAL. The high conversion rate could
also be a function of particle size, the nanoparticles may be able
to penetrate deeper into the polymeric network of lignin and, con-
sequently, could more efficiently disrupt the linkages between the
lignin monomers.
Figs. 4–6 present the catalytic behavior of the same catalyst
under ultrasonic activation in different solvents. Like in the auto-
clave experiments, in all the cases the conversion of AL was superior
to the conversion of FAL, supporting the hypothesis the reac-
tion mechanism does not depend on the way the reaction has
been activated. However, irrespective of the nature of the solvent
that has been employed, all the experiments show the extent of
depolymerization was greatest under autoclave conditions. The
advanced solubilization of lignin and the hydrogen pressure may be
Generally, sonochemistry in heterogeneous systems is the result
of a combination of chemical and mechanical effects of cavitation,
and it is very difficult to ascribe sonochemistry to any single global
origin, other than the overriding source of activity, namely, cavita-
tion [31]. Whatever the mechanism is (i.e. chemical and mechanical
Fig. 4. Depolymerization of lignin samples under ultrasound conditions (catalyst:
Fig. 6. Depolymerization of lignin samples under ultrasound conditions (catalyst:
0
.27 g nano-Ni, 0.25 g lignin, 30 mL methanol, 6 h).
0.27 g nano-Ni, 0.25 g lignin, 30 mL ethyl acetate, 6 h).

8
K.B.H. Finch et al. / Catalysis Today 196 (2012) 3–10
Fig. 9. Depolymerization of FAL under ultrasound conditions (catalyst: 0.27 g nano-
Ni or 0.27 g NiO, 0.25 g FAL lignin, 30 mL methanol, 6 h).
Fig. 7. Depolymerization of FAL and AL samples under autoclave conditions (cata-
lyst: 0.02 g (NiAlMgO)x, 0.16 g lignin, IL: 2 g [BMIM]OAc, 10 atm H2, 180 C, 24 h).
◦
between the two catalysts. Such a behavior may account for at least
a partial reduction of NiO(1 1 1) under hydrogen bubbled atmo-
sphere and ultrasound activation.
effects of cavitation) the real benefit of using ultrasound lies in its
unique selectivity and reactivity enhancement.
Data presented in Figs. 4–6 shows changes in selectivity com-
pared to the experiments carried out in autoclave conditions. Under
ultrasonic activation the distribution of the products between 100
and 1000 m/z is not influenced by any of the solvents used.
The nature of the solvent was also found to exhibit a small effect
on these reactions, except for ethyl acetate, the conversion of AL
samples was greater. The extent of lignin depolymerization varied
in the order methanol > ethyl acetate > water. This could imply that
methanol interacts with the disrupted fragments and favors a more
advanced depolymerization. Transesterification of ethyl acetate is
also plausible.
3.6. General assessments
Reaction of lignin under pressurized hydrogen at elevated tem-
peratures and in the presence of a suitable catalyst is one of
the most intriguing approaches for lignin depolymerization and
hydrodeoxygenation [1]. Shabtai et al. [32] proposed a three-
step process for converting lignin into reformulated gasoline.
Including base-catalyzed depolymerization (BCD) followed then by
hydrodeoxygenation and hydrocracking. When chemicals are the
target products, the process may be ended before the final step and
the phenolics can be extracted from the product after hydrodeoxy-
genation [32].
3.5. Depolymerization on Ni oxide catalysts
The proposed BCD process has two distinct disadvantages:
the use of strong homogeneous base catalysts and the energy
input required. Replacing the homogeneous catalyst with a
heterogeneous solid base catalyst, such as those used in this
work (i.e. (NiAlMgO)x, (NiAlO)x, NiO(111), nano-Ni(0)) could have
the advantage of avoiding waste productions and the sepa-
ration of the homogeneous catalyst. Moreover, the magnetic
Figs. 7 and 8 show catalytic results of Fe O –(NiAlO)x and
3
4
Fe O –(NiMgAlO)x catalysts under autoclave conditions. Although
3
4
AL still shows higher conversions than FAL, the difference between
the conversion of AL and FAL is less than the difference measured
using nano-Ni(0) catalysts. This is also apparent in the distribution
of the products between 100 and 1000 m/z. The presence of mag-
nesium (Fig. 7) led to an increase in the conversion but with this
catalyst the products between 100 and 500 m/z was decreased.
Figs. 9 and 10 show comparative results in depolymerization
of FAL and AL using nano-Ni(0) and NiO(1 1 1) under ultrasound
activation in methanol. These data show rather small differences
properties of materials incorporating Fe O4 provide an easier
3
separation method and allow for recycling of the catalyst. The
use of nanostructured catalysts may also have other advantages
such as higher reactivity or better dispersion into the reaction
mixture.
Fig. 8. Depolymerization of FAL and AL samples under autoclave conditions (cata-
lyst: 0.02 g (NiAlO)x, 0.16 g lignin, IL: 2 g [BMIM]OAc, 10 atm H2, 180 C, 24 h).
Fig. 10. Depolymerization of AL under ultrasound conditions (catalyst: 0.27 g nano-
Ni or NiO, 0.25 g AL lignin, 30 mL methanol, 6 h).
◦

K.B.H. Finch et al. / Catalysis Today 196 (2012) 3–10
9
The characterization of the catalysts indicated that nickel can
exist in nanoparticles in a zero-valent oxidation state (nano-Ni(0)),
oxidized state (NiO(1 1 1)), or a mixture of reduced and oxidized
states ((NiAlMgO)x and (NiAlO)x)).
The currently accepted structure of lignin consists of three basic
monomers (coniferyl, sinapyl and p-coumaryl alcohols) connected
by ␤-O-4, ␣-O-4, ␤-5, 5-5, 4-O-5, ␤-1 and ␤-␤ linkages [33]. How-
ever, as the characterization of the lignin samples has shown, the
proportion of these monomers is different as a function of the
applied procedure (FAL or AL). Previous studies have shown that
ultrasound conditions were inferior to those tested under con-
ventional heating conditions. The nature of the solvent was also
found to be very important in this process, with the ionic liquid
[BMIM]OAc leading to the best results under autoclave conditions,
and methanol under ultrasonic conditions. In all the cases the con-
version of the fractions with m/z higher than 1000 was smaller than
5%.
Acknowledgements
␤
-O-4 and ␣-O-4 linkages in lignin are easily cleaved, while the 5,5-
The authors kindly acknowledge the American Chemical Soci-
ety (ACS) for the ACS GREET pilot program grant in 2011 awarded
to Kenneth Finch to study at the Department of Organic Chemistry,
Biochemistry and Catalysis of University of Bucharest, under host
Professor Dr. Simona Coman. Part of the work was supported by
the strategic grant POSDRU/89/1.5/S/58852, Project “Postdoctoral
programme for training scientific researchers” co-financed by the
European Social Foundation within the Sectorial Operational Pro-
gram Human Resources Development 2007–2013. Aurore Richel is
grateful to the “Région Wallonne” (Belgium) for its financial support
(
biphenyl)-type and aromatic ring structures are more stable [34].
This data agrees with the results presented here. Showing that irre-
spective of the catalyst, the extent of depolymerization in AL was
higher than the extent of depolymerization in FAL. These results
are more interesting from a practical point of view taking into con-
sideration the molecular weight of AL is higher than that of FAL,
the low-molecular weight fraction (m/z = 100–500) preponderated,
and it was followed by the fraction with m/z of 700–1000.
The activation of the catalysts is also very important. The
chemical effects of a sonochemical reaction result from acoustic
cavitation, defined as the formation, expansion, and rapid implo-
sion of bubbles. The bubbles create transient localized hot spots
upon implosion with extremely high temperature and pressure
(“Technose” Excellence Programme). The authors are also grateful
for the support of COST (UBIOCHEM action) and the STSM of Dr.
Marian Verziu.
[
35]. The experiments carried out in an autoclave led to better
Appendix A. Supplementary data
results compared to ultrasonic activation. Under ultrasonic acti-
vation the solid surfaces (especially powders) collapse because
enough energy is produced to cause fragmentation, even for nanos-
tructured materials. Such an effect takes place in these reactions
for both the lignin and the catalyst. In the case of the catalyst
this effect may lead to an increased active surface area and may
favor efficient mixing and enhanced mass transport. For very fine
powders, as in the case of nano-Ni catalyst, the particles are accel-
erated to high velocity by cavitational collapse and may collide
to cause surface abrasion. These phenomena may generate par-
ticles with different sizes, as well as the agglomeration of small
particles. The rearrangement of solid particles (i.e. fragmentations,
agglomeration) may affect the catalytic performance leading, in
many cases, to a decrease in the catalytic activity. In the case of
lignin, irrespective of the preparation procedure that has been
employed, the low molecular weight monomers (m/z = 100–500)
may suffer a polymerization to higher molecular weight monomers
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.cattod.2012.02.051.
References
[
[
1] M.P. Pandey, C.S. Kim, Chem. Eng. Technol. 34 (2011) 29–41.
2] S. Baumlin, F. Broust, F. Bazer-Bachi, T. Bourdeaux, O. Herbinet, F. Toutie Ndiaye,
M. Ferrer, J. Lédé, Int. J. Hydrogen Energy 31 (2006) 2179–2192.
[3] B.N. Kuznetsov, M.L. Shchipko, Bioresour. Technol. 52 (1995) 13–19.
[4] H. Yang, R. Yan, H. Chen, D. Ho Lee, C. Zheng, Fuel 86 (2007) 1781–1788.
[5] S. Nenkova, T. Vasileva, K. Stanulov, Chem. Nat. Compd. 44 (2008) 182–185.
[6] H. Deng, L. Lin, Y. Sun, C. Pang, J. Zhuang, P. Ouyang, Z. Li, S. Liu, Catal. Lett. 126
2008) 106–111.
7] C. Crestini, M.C. Caponi, D.S. Argyropoulosc, R. Saladino, Bioorg. Med. Chem. 14
2006) 5292–5302.
(
[
(
[8] S. Wang, K. Wang, Q. Liu, Y. Gu, Z. Luo, K. Cen, T. Fransson, Biotechnol. Adv. 27
(2009) 562–567.
[
9] C.A. Mullen, A.A. Boateng, Fuel Process. Technol. 91 (2010) 1446–1458.
[
[
10] T. Furusawa, T. Sato, H. Sugito, Y. Miura, Y. Ishiyama, M. Sato, N. Itoh, N. Suzuki,
Int. J. Hydrogen Energy 32 (2007) 699–704.
11] W.-J. Xu, W. Zhao, C. Sheng, S.-T. Zhong, X.-N. Wu, C.-H. Yan, S. Bai, Z.-M. Zong,
X.-Y. Wei, Energy Fuels 24 (2010) 250–252.
(
m/z = 700–1000) (see Figs. 4–6).
[
[
12] J. Zhang, H. Deng, L. Lin, Molecules 14 (2009) 2747–2757.
13] J.D.P. Araújo, C.A. Grande, A.E. Rodrigues, Chem. Eng. Res. Des. 88 (2010)
4
. Conclusions
1024–1032.
[
[
14] Y. Pu, N. Jiang, A.J. Ragauskas, J. Wood Chem. Technol. 27 (2007) 23–33.
15] J. Zakzeski, A.L. Jongerius, B.M. Weckhuysen, Green Chem. 12 (2010)
Lignin isolated from Miscanthus × giganteus using acidic (FAL)
and alkali (AL) conditions was thereafter subjected to catalytic
depolymerization under thermal or ultrasonic activation. The char-
acterization of lignins was achieved by thermogravimetric analysis
and FTIR. Three different classes of catalysts, containing nickel
as active species, have been prepared in this scope: (i) nano-
Ni by reduction of NiCl2 with NaBH4 under ultrasonication, (ii)
Fe O –(NiMgAlO)x and (NiAlO)x by calcination of Mg(Ni)–Al hydro-
1
225–1236.
[16] N. Yan, C. Zhao, P.J. Dyson, C. Wang, L.-tao Liu, Y. Kou, ChemSusChem 1 (2008)
26–629.
17] J. Hu, K. Zhu, L. Chen, H. Yang, Z. Li, A. Suchopar, R. Richards, Adv. Mater. 20
2008) 267–271.
6
[
(
[18] C. Vanderghem, A., Richel, N., Jacquet, Ch. Blecker, M. Paquot, Polym. Degrad.
Stab. doi:10.1016/j.polymdegradstab.2011.07.022.
[
19] H.Q. Lam, Y. Le Bigot, M. Delmas, G. Avignon, Ind. Crop. Prod. 4 (2001) 139–144.
20] J. Wang, J. You, P. Yang, C. Zhong, Z. Li, M. Zhang, X. Jing, Mater. Sci.: Poland 26
(2008) 591–599.
3
4
[
talcite incorporating Fe O4 followed by reduction with hydrogen,
3
and (iii) NiO(1 1 1) nanosheets by reduction of Ni(NO ) with urea
[21] Z.-F. Zhao, Z.-J. Wu, L.-X. Zhou, M.-H. Zhang, W. Li, K.-Y. Tao, Catal. Commun. 9
2008) 2191–2194.
22] J.A. Dahl, B.L.S. Maddux, J.E. Hutchison, Chem. Rev. 107 (2007) 2228–2269.
[23] Y. Diao, W.P. Walawender, C.M. Sorensen, K.J. Klabunde, T. Ricker, Chem. Mater.
14 (2002) 362–368.
3
2
(
using benzyl alcohol as a structure directing agent. The catalysts
were characterized by XRD and XPS techniques. Reduced mixed
oxides displayed a moderate activity while a significant increase in
conversion (up to 90%) was observed in the presence of nano-Ni(0).
NiO(1 1 1) nanosheet catalysts performed very close to nano-Ni(0).
The conversion and the mass distribution of the reaction prod-
ucts were strongly related to the procedure used for the extraction
of lignin samples. In all the cases AL led to a greater extent of
depolymerization. The performances of the tested catalysts under
[
[
[
24] R.T.K. Baker, Carbon 27 (1989) 315–323.
25] L. Obalova, K. Jiratova, F. Kovanda, M. Valaskova, J. Balabanova, K. Pacultova, J.
Mol. Catal. A: Chem. 248 (2006) 210–219.
[26] M.G. Cook, N.S. McIntyre, Anal. Chem. 47 (1975) 2208–2213.
[
27] S.S. Abdullah, S. Yusup, M.M. Ahmad, A. Ramli, L. Ismal, Int. J. Chem. Biol. Eng.
(2010) 137–141.
28] Y. Wu, Z.L.H. Zhao, F. He, J. Fuel. Chem. Technol. 37 (2009) 427–432.
3
[
[29] A. Uslu, A.P.C. Faaij, P.C.A. Bergman, Energy 33 (2008) 1206–1223.

1
0
K.B.H. Finch et al. / Catalysis Today 196 (2012) 3–10
[
[
[
30] L. Serrano, I. Egües, M. González Alriols, R. Llano-Ponte, J. Labidi, Chem. Eng. J.
56 (2010) 49–55.
31] M. Doble, A.K. Kruthiventi, Green Chemistry & Engineering, Elsevier Science &
Technology Books, 2007.
[33] J. Zakzeski, P.C.A. Bruijnincx, A.L. Jongerius, B.M. Weckhuysen, Chem. Rev. 110
(2010) 3552–3599.
[34] A. Tejado, C. Pena, J. Labidi, J.M. Echeverria, I. Mondragon, Bioresour. Technol.
98 (2007) 1655–1663.
1
32] J.S. Shabtai, W.W. Zmierczak, E. Chornet, US Patent 5959167 (1999).
[35] K.S. Suslick, T. Hyeon, M. Fang, Chem. Mater. 8 (1996) 2172–2179.