Selective hydrogenolysis of Α–O–4, Β–O–4, 4–O–5 C–O bonds of lignin-model compounds and lignin-containing stillage derived from cellulosic bioethanol processing

Article abstract of DOI:10.1016/j.apcata.2017.04.022

Benzyl phenyl ether (BPE), phenethyl phenyl ether (PPE) and diphenyl ether (DPE) have been selected as model compounds of the most abundant and significant ether linkages found within the complex structure of lignin (e.g. α–O–4, β–O–4, and 4–O–5, respectively). The catalytic hydrogenolysis of these compounds has been carried out using several Ru, Pd and Ni catalysts supported over different metal oxides (e.g. Al₂O₃, ZrO₂, TiO₂) and carbon materials (e.g. active carbon, multiwall carbon nanotubes). The conversion of these compounds at relevant hydroprocessing conditions (150?°C, 25?bar-g in H₂ atmosphere) is much dependent on the labile nature of the relevant ether bonds of the selected model compounds. Conversion levels for the three compounds increases in the following order: DPE (4–O–5 linkage)??Pd?>?Ru). The catalytic systems with easier reducible species performes better in the conversion of the dimer models. Besides, a preliminary study on the catalytic depolymerization of a real lignin stream (lignin-containing 2G bioethanol plants stillage) has been carried out. Detailed characterization by ¹³C-¹H heteronuclear single-quantum correlation spectroscopy (HSQC) showed that on 5?wt.% Ru/C it is possible to achieve 50% desapearance of the most abundant lignin ether bond (i.e. β–O–4) and liquid yields above 50?wt.%.

Full text of DOI:10.1016/j.apcata.2017.04.022

Accepted Manuscript
Title: Selective hydrogenolysis of ␣-O-4, ␤-O-4, 4-O-5 C-O
bonds of lignin-model compounds and lignin-containing
stillage derived from cellulosic bioethanol processing
´
Authors: B. Gomez-Monedero, M.P. Ruiz, F. Bimbela, J. Faria
PII:
S0926-860X(17)30189-8
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2017.04.022
APCATA 16221
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
16-2-2017
12-4-2017
29-4-2017
´
Please cite this article as: B.Gomez-Monedero, M.P.Ruiz, F.Bimbela, J.Faria, Selective
hydrogenolysis of ␣-O-4, ␤-O-4, 4-O-5 C-O bonds of lignin-model compounds
and lignin-containing stillage derived from cellulosic bioethanol processing, Applied
Catalysis A, Generalhttp://dx.doi.org/10.1016/j.apcata.2017.04.022
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Selective hydrogenolysis of -O-4, -O-4, 4-O-5 C-O bonds of lignin-model compounds and lignin-containing
stillage derived from cellulosic bioethanol processing
B. Gómez-Monedero,a M. P. Ruiza, F. Bimbela,b† and J. Fariaa*
aAbengoa Research, Campus Palmas Altas, c/ Energía Solar 1, 41014 Seville, Spain.
*Corresponding author: Jimmy Faria, ing.jimmyfaria@gmail.com
bGrupo de Procesos Termoquímicos (GPT), Aragón Institute for Engineering Research (I3A), Universidad Zaragoza,
Mariano Esquillor s/n, 50018 Zaragoza, Spain.† Present address: Grupo de Reactores Químicos y Biorreactores,
Applied Chemistry Department, Universidad Pública de Navarra, 31006 Pamplona, Spain.

Conversion of Model Compounds
Energy of C-O Dissociation
BPE
PPE
DPE
100%
75%
50%
25%
0%
DPE 314 KJ/mol
PPE 289 KJ/mol
BPE 218 KJ/mol
Ruthenium
Palladium
Nickel

Highlights
-
Selective cleavage of C-O bonds of -O-4 ether bonds present in lignin and lignin model compounds have
been achieved on Ru supported catalysts
-
-
Ru catalysts showed a higher activity for hydrogenolysis in comparison to Pd and Ni catalysts
Catalyst support plays an important role in the hydrogenolysis reaction, being carbon-supported catalysts
the most active ones towards C-O cleavage
Abstract
Benzyl phenyl ether (BPE), phenethyl phenyl ether (PPE) and diphenyl ether (DPE) have been selected as model
compounds of the most abundant and significant ether linkages found within the complex structure of lignin (e.g.
-O-4, -O-4, and 4-O-5, respectively). The catalytic hydrogenolysis of these compounds has been carried out
using several Ru, Pd and Ni catalysts supported over different metal oxides (e.g. Al2O3, ZrO2, TiO2) and carbon
materials (e.g. active carbon, multiwall carbon nanotubes). The conversion of these compounds at relevant
hydroprocessing conditions (150 oC, 25 bar-g in H2 atmosphere) is much dependent on the labile nature of the
relevant ether bonds of the selected model compounds. Conversion levels for the three compounds increases in
the following order: DPE (4-O-5 linkage) < PPE (β-O-4) < BPE (α-O-4). Product distribution is also dependent on the
chemical structure of the raw material subjected to catalytic hydroprocessing, with BPE and DPE mainly yielding
aromatic monomers, whilst fully saturated monomers are the main reaction products when PPE is
hydroprocessed. The reaction mechanisms upon hydroprocessing vary from one model compound to another,
being hydrogenolysis the prevailing route in the catalytic hydroprocessing of BPE and DPE, whereas competing
reaction pathways between dimer hydrogenation and hydrogenolysis of the C-O bond occur for the PPE molecule.
The conversion and selectivity to relevant hydrogenolysis products are also dependent on the catalyst used. Ru-
based catalysts are the most active followed by Pd -based and Ni-based, whereas selectivity to aromatic
monomers increases in the reverse order (Ni > Pd > Ru). The catalytic systems with easier reducible species
performes better in the conversion of the dimer models. Besides, a preliminary study on the catalytic
depolymerization of a real lignin stream (lignin-containing 2G bioethanol plants stillage) has been carried out.
Detailed characterization by 13C-1H heteronuclear single-quantum correlation spectroscopy (HSQC) showed that
on 5 wt. % Ru/C it is possible to achieve 50 % desapearance of the most abundant lignin ether bond (i.e. -O-4)
and liquid yields above 50 wt. %.
Keywords
Lignin; hydrogenolysis; lignin model compounds; depolymerization; Ru catalyst;
1. Introduction
Mitigation of greenhouse emissions (GHE), energy independence, and rural development are the key drivers for
biomass-conversion technologies. Moreover, according to reports from the International Energy Agency (IEA),
between 2007 and 2030, global total primary energy demand is expected to increase by 40 %, in which biomass
conversion into liquid transportation fuels will play an important role. Hence, it is clear that biomass will be a
2

critical part in the production of CO2-neutral fuels and chemicals [1,2]. This fact has stimulated the development
of advanced biofuels such as cellulosic ethanol, which are now becoming a reality with the commissioning of large
scale plants in Italy and USA [3-6]. This new generation of biofuels plants can convert agricultural wastes (e.g.
wheat straw and corn stover) and organic fractions of municipal solid wastes into ethanol. These recalcitrant
feedstocks require harsh chemical and thermal pretreatments that have been extensively optimized to increase
accessibility and reactivity of the cellulose fibrils towards the enzymatic cocktail [7]. Likewise, intensive research
has been focused on the development of recombinant enzymes and fermentation yeast by carefully tailoring
genetics of enzyme-producing microorganisms and C5-fermenting yeasts to maximize ethanol production [7,8]. A
major challenge that remains unsolved, however, is the revalorization of lignin, the second most abundant
naturally occurring polymer which accounts for 10-35 % of biomass by weight and around 40 % of its energy
[9,10]. In second generation (2G) biorefineries, lignin can be produced at the pretreatment stage or at the stillage
recovery, where the mixture is concentrated to ~40 wt. % total solids using press-filtration before entering a
biomass-boiler to generate heat. Nonetheless, this approach has significant limitations, as lignin boilers are
expensive and energy content of this material is significantly lower than natural gas. Additionally, the volume of
lignin residue that will be produced readily exceeds the energy requirements of the biorefinery complicating its
utilization as energy source, unless electricity is generated and transferred to the electrical grid [11].
Hence, economical conversion of this byproduct into valuable chemicals, fuels, polymers, and carbon-based
structural materials is germane to guarantee the long-term profitability and sustainability of the biorefineries.
Similarly, in the pulp and paper industry, black liquor produced after the purification of the cellulose pulp is used
for energy recovery, in which the carbonaceous residues (i.e. lignin) are combusted and the mineral oxides are
recovered for reuse in the caustic treatment of the lignocellulose. This process is also under significant political
pressure as new caps of energy co-generation are imposed to pulp-manufacturers [12]. Thereby, it is clear that
valorizing lignin-containing streams into added value products is highly desirable to improve economics of current
2G biofuels and pulping plants, and enhance CO2 abatement. In this context, different possibilities and approaches
can be found in the scientific literature, targeting at breaking down the complex structure of lignin into simpler
molecules that could further serve for producing valuable bio-based chemicals and fuels. Different processes,
such as oxidation [13] or acid-base catalysis [14,15], have been studied as a means to catalytically depolymerize
lignin, as thoroughly reviewed by several authors in recent times [16-20]. Thermochemical processing of lignin via
fast pyrolysis for producing bio-oil, catalytic depolymerization, hydroprocessing and oxidation are different
alternatives for lignin conversion, currently being intensively researched by many scientists [16]. Other routes
include gasification and reforming processes, using steam or in liquid phase. Among these strategies, reductive
lignin depolymerization, i.e. hydroprocessing, is one of the most widely studied routes nowadays to convert lignin
and related model compounds. In this sense, approaches such as hydrodeoxygenation [21-23] or hydrogenolysis
[24-31] under different solvent media have been studied to transform and/or remove functionality of lignin and
model compounds to produce simpler but valuable chemicals [16-17].
Lignin polymeric structure is mainly based on aromatic rings linked to each other by means of C-O ether bonds
and C-C bonds. Among them, β-O-4 linkages are the most abundant (up to 50 % of total), followed by α-O-4 (12
%) and 4-O-5 (8 %) linkages (Figure 1) [32,33].
Bearing in mind that C-C bonds are much more stable, and that upon dissociation of the C-O-C bonds the total
amount of produced monomers will surpass that of dimers [18], selective depolymerization of these linkages is of
utmost importance for lignin valorization. In this sense, among the proposed alternatives to valorize lignin,
catalytic depolymerization offers the possibility to cleave specific bonds in the lignin structure at high-rates. This
process will generate oligomers and monomers, which can be further upgraded to fuels and chemical building
blocks [17,34,35].
Since various factors can dramatically alter the structure and properties of lignin, including the source and
isolation procedure for obtaining lignin, some recent studies have carried out hydrogenolysis of different types of
lignin [36,44]. Moreover, in order to simplify the structural intricacy and variability of lignin, the use of lignin
model compounds has been widely reported in the literature for hydroprocessing of lignin and lignin-derived
3

streams. Phenol and phenolic compounds such as o-cresol (2-methylphenol) and guaiacol (2-methoxyphenol)
have been widely selected as model compounds in studies related to thermal processing of lignin followed by
catalytic upgrading via hydrodeoxygenation (HDO) [45-51]. Regarding hydrogenolysis, model compounds that
have been typically selected to mimic characteristic linkages from lignin are aryl ethers, such as diphenyl ether
(DPE), phenethyl phenyl ether (PPE), and benzyl phenyl ether (BPE) [32,52-58].3
In this scenario, it was considered of interest to study the performance of several catalytic materials using model
compounds of relevant ether bonds in lignin structure (BPE, DPE and PPE). As commented before, the rupture of
such bonds is crucial for lignin valorization, not only to reduce over-saturation of the aromatic ring, but also to
increase the yield of depolymerization products. Thus, the present work investigates how product distribution
and selectivity towards the cleavage of the different C-O bonds is affected by the characteristics of metal catalyst
and support (i.e. metal oxides and carbon-based materials) together with the nature of the molecular structure of
the selected model compounds. One of the most promising catalytic material (5 wt. % Ru/C) was tested as
catalyst for the depolymerization of a real lignin stream (lignin-containing stillage obtained from 2G
bioethanol plants).
2. Experimental
Materials and methods. Diphenyl Ether (DPE) was purchased from ACROS Organics (Phenyl Ether, 99%) whereas
benzyl Phenyl Ether (BPE) was purchased from Sigma-Aldrich (Benzyl Phenyl Ether, 98 %) and from ACROS
Organics (Benzyl Phenyl Ether, 98 %). Phenethyl Phenyl Ether (PPE) was synthesized according to the procedure
detailed in the synthetic procedures subsection. Methanol used to prepare the stock solutions of the three model
compounds was provided by VWR (Methanol GPR RECTAPUR, ≥99.5 %). Ni precursor was purchased from ACROS
Organics (Ni(NO3)2.6H2O, 99%) and Ru and Pd precursors from Sigma Aldrich (RuCl3, 45-55 % Ru content and
PdCl2, Reagent plus®, 99%). HCl used to prepare the solution to dissolve PdCl2 was obtained from Sigma Aldrich
(Hydrochloric acid, ACS reagent, 37%). In addition, the following supporting materials were used in this study
without further purification: activated carbon (Activated Charcoal DARCO®, ~ 100 mesh particle size, Sigma-
Aldrich), titanium oxide (TiO2-P25, Evonik), zirconium oxide (ZrO2, Sigma-Aldrich), aluminum oxide (Al2O3,
Martinswer (Albemarle), COMPALOX AN/V-813, provided by Sasol), purified multiwall carbon nanotubes
(MWCNT-purified, kindly provided by SouthWest NanoTechnologies) and multiwall carbon nanotubes partially
purified (MWCNT41, 41% content of SiO2, also provided by SouthWest NanoTechnologies).
2.1.Experimental installation and procedures.
Liquid-phase batch catalytic tests for hydroprocessing of BPE, PPE and DPE were carried out in a high-pressure
stainless steel autoclave reactor (Berghof Highpreactor™ High-Pressure Laboratory Reactor BR100), equipped
with a 50-mL teflon liner, a pressure transducer or two manometers (one for pressures up to 10 bar, the other for
pressures up to 250 bar), a stainless steel deposit for liquids, a thermocouple connected to a temperature
controller, and a magnetic stirrer. Figure S6 shows a scheme of the reaction system.
To carry out the catalytic runs, firstly, three stock solutions were prepared, one for each model compound,
dissolving the necessary amount of BPE, DPE or PPE in methanol, in order to obtain solutions with a concentration
of the corresponding model compound equal to 16 mM.
Prior to reaction, 70 mg of the desired catalyst were heated up to 250 °C (Ru catalysts), 400 °C (Ni-based catalysts
except Ni/Al2O3), 650 °C (Ni/Al2O3), or 190 °C (Pd catalysts), for 1 hour and under 15 mL/min of H2 in a tubular
quartz reactor. Once the target temperature was reached for each case, the catalyst was reduced in situ at these
conditions for 3 hours. Then, the catalysts were passivated in an air flow (15 mL/min) at room temperature for 30
min. After this, 50 mg of the selected catalyst together with a magnetic stirrer were placed inside the 50 mL
Teflon liner, the stainless-steel batch reactor was sealed, and a leak test was carried out at 50 bar-g in N2
atmosphere. Then the reactor was flushed 3 times with pure H2 in order to remove any remaining N2 from the
leak test and after this the reactor was pressurized up to 7-8 bar-g using H2. The reaction system was heated up to
4

the desired temperature (150 °C) with a heating rate of approximately 1.5 °C /min and, once the target
temperature was achieved, the system was maintained at this temperature and pressure for 30 min to activate
the catalyst.
Afterwards, 20 mL of stock solution of the desired model compound were placed into the liquid’s vessel of the
reactor, which was subsequently pressurized to 25 bar-g using H2. The discharge valve of the vessel was opened
and the solution together with H2 were introduced into the reactor. The latter procedure was repeated until the
pressure inside the reactor vessel reached 25 bar-g and after this, a stirring speed of 750 rpm was set. At this
moment, reaction time was set to zero. After two hours of reaction, the heating and stirring were stopped and
the reactor was cooled down in an ice-cooled bath. When the reactor temperature was below 20 °C it was
carefully depressurized.
Liquid products were filtered and analyzed using gas chromatography. The products obtained after reaction were
analyzed by GC-MS for identification (Agilent 7890 GC-system, model G3440A, equipped with a 5975C mass
spectrometer detector. Column: Agilent HP5-ms, 0.250 mm inner diameter, 30 m long, 0.25 µm film thickness)
and by GC-FID for quantification (Agilent 7890 GC-system, model G3440A, equipped with a 5975C flame
ionization detector. Column: Agilent HP5, 0.320 mm inner diameter, 30 m long, 0.25 µm film thickness).
After identification of the products, the sample was injected in the GC-FID system to quantify the obtained
products with the aid of calibration curves previously prepared. Each curve contained eight concentration levels,
from 1.2 mM to 50 mM, and response factors for each compound were obtained by adjusting the areas obtained
at each concentration level, assuming that for concentration equal to 0 mM the response area will be 0 as well.
Concentration levels were prepared by producing first a stock solution with a concentration 50 mM for each of
the calibrated compounds in methanol, and diluting selected volumes of this solution to obtain the rest of
concentrations. Each solution was injected three times, and the mean value of the obtained areas was calculated
and used for fitting the curve.
Hydroprocessing of model compounds yielded some chemicals that could not be calibrated due to either their
toxicity (i. e. benzene) or not being commercially available. In this sense, response factors for these non-calibrated
compounds were approximated to those of the most similar compounds calibrated. Calibrated compounds
included the three model compounds of ether bonds (BPE , PPE and DPE), toluene (99.9 % purity, Sigma-Aldrich),
ethyl benzene (99.8 % purity, ACROS Organics), phenol (99 % purity, ACROS Organics), benzyl alcohol (minimum
purity 99.0 %, Sigma-Aldrich), cyclohexane (minimum purity 99 %, Sigma-Aldrich), methylcyclohexane (minimum
purity 98.0 %, Sigma-Aldrich), cyclohexanol (98 % purity, ACROS Organics), 2-phenylethanol (minimum purity 99
%, Sigma-Aldrich), cyclohexylphenyl ether (95 % purity, Sigma-Aldrich), and cyclohexanone (minimum purity 99.5
%, Sigma-Aldrich).
Once the concentration of each product before (in the case of the model compounds) and after reaction was
calculated according to calibration curves, the yield to each product was determined. As several reactions were
taking place at the same time (hydrogenation of the dimer, hydrogenolysis, etc.) conversion, selectivity and yields
were calculated on a molar C basis. The defined equation for each parameter can be seen below (Eqs. 1-3).
푚표푙 퐶 표푓 퐷푖푚푒푟 −푚표푙 퐶 표푓 퐷푖푚푒푟
0
% 퐶표푛푣푒푟푠ꢀ표푛 =
푓 ∙ 100 (퐸푞. 1)
푚표푙 퐶 표푓 퐷푖푚푒푟
0
푚표푙 퐶 표푓 푝푟표푑푢푐푡 퐴
푆퐴 =
(퐸푞. 2)
푇표푡푎푙 푚표푙 퐶 표푓 푝푟표푑푢푐푡푠
( )
( )
푌 % = 푆퐴 ∙ 퐶표푛푣푒푟푠ꢀ표푛 % (퐸푞. 3)
퐴
where;
5

-
-
-
-
-
mol C of dimer0: the moles of carbons in the form of BPE, PPE or DPE in the solution before reaction.
mol C of dimerf: the moles of carbon in the form of BPE, PPE or DPE in the solution after reaction.
SA: selectivity to product A, i. e. toluene.
Total mol C of products: the sum of the moles of carbon of all the products detected by GC-FID.
YA: yield to product A.
In addition, to ease comparison, some results were presented in the form of yield to four families of compounds:
aromatic monomers, saturated monomers, partially saturated dimers and fully saturated dimers.
2.2.Catalyst preparation.
Catalysts were prepared using aqueous solutions of Ru, Ni and Pd metal salt precursors for impregnation onto the
selected supports. Materials acting as supports of the desired metals were either commercial supports or
materials synthesized or modified in the laboratory. Commercial supports have been presented in the materials
subsection. In addition, the following supports were prepared in the laboratory: titanium oxide, zirconium oxide,
and TiO2/C (see Supplementary Information for details on preparation methods).
Ru catalysts (supported on active carbon, Al2O3, ZrO2 commercial and from laboratory, TiO2 commercial and from
laboratory, and TiO2/C) were synthesized by excess impregnation. The necessary amounts of Ru precursor salt
were weighed and dissolved in the corresponding volume of deionized water to produce a catalyst with a content
of the active metal equal to 5 wt. % in the final solid, following the proportion of 500 mL of water for 0.54 g of
RuCl3. The solution of the precursor in water was stirred at room temperature using a magnetic stirrer, until
complete dissolution of the salt. Once the solution of RuCl3 was homogeneous, the desired amount of support
was added to the solution and the mixture was maintained under stirring overnight. Afterwards, the solution was
heated to evaporate the water, and the solid obtained was dried in an oven at 100 °C overnight.
Ru/MWCNT catalyst was synthesized by incipient wetness impregnation. A certain amount of the precursor salt
(RuCl3) was dissolved in a certain amount of a 1:1 (v/v) mixture of methanol and deionized water, in order to
obtain a solution that allowed producing a catalyst with a 5 wt. % of the active metal in the final solid. The
solution was impregnated into the desired support and then the catalyst was dried overnight at 100 °C in an oven.
Ni catalysts were synthesized by incipient wetness impregnation. The precursor nitrate salt was dissolved in a
certain amount of deionized water (approximately equal to the pore volume of the support (measured with
water)), to obtain a solution that allowed producing a catalyst with a 5 wt. % of the active metal in the final solid.
The solution was impregnated onto the desired support and then the catalyst was dried overnight at 100 °C in an
oven.
Pd catalysts were also synthesized by incipient wetness impregnation. The precursor salt was dissolved in the
necessary amount of deionized water to fill the pore volume of the support, though in this case, a few droplets (3-
5) of a 0.2 M hydrochloric acid solution were added to the deionized water in order to facilitate the dissolution of
the palladium salt. The resulting solution was impregnated onto the desired support and then the catalyst was
dried overnight at 100 °C in an oven.
After impregnation and drying in an oven overnight, the catalysts were calcined as follows. In the case of samples
including Al2O3, ZrO2, and TiO2 as supports, the prepared catalysts were calcined in a muffle furnace in air
atmosphere, with a temperature ramp of 3 °C/min up to 400 °C, and maintained at such temperature for 4 hours.
On the other hand, when active carbon, MWCNT-purified, MWCNT41, or TiO2/C were used as supports, the
calcination of the catalysts was carried out in a tubular quartz reactor, with a vertical flow of nitrogen of 20 N
mL/min. A temperature ramp of 3 °C/min was set to reach 400 °C and then the solid was maintained at this
temperature for 4 hours under nitrogen flux.
2.3.Catalyst characterization.
6

Textural properties of some of these catalysts were analyzed by some of the following techniques: temperature
programmed reduction (TPR), high resolution transmission electron microscopy (HRTEM), nitrogen adsorption
isotherms, and X-ray photoelectron spectroscopy (XPS).
Temperature programmed reduction (TPR) analyses were carried out in a Micromeritics Chemisorption Analyzer
(AutoChem II), equipped with TCD detectors. Samples were heated at 10 °C /min from room temperature to 800
°C or 900 °C, while a stream of 50 mL/min of 10% H2-Ar mixture circulated through the system. High resolution
TEM images were acquired in a Tecnai F30 (FEI company) high resolution Transmission Electron Microscope that
can work in either TEM or STEM (Scanning-Transmission) modes. The working voltages in this microscope are 200
and 300 kV. Isotherms were determined in an ASAP 2020 system (Micromeritics). Prior to the analyses, samples
were degasified under vacuum applying a temperature program (10 °C/min ramp from room temperature to 200
°C, holding the latter temperature during 360 min). N2-adsorption was carried out at 77 K by dosing growing
amounts of N2 to cover the whole relative pressures interval, until reaching a point close to saturation (P/P0 =
0.995). Then vacuum was applied to progressively reduce pressure, producing desorption of the gas. Surface area
was estimated using the Brunauer-Emmett-Teller (BET) model, applied to the adsorption branch at the selected
partial pressures range for each catalyst, thus capillary condensation in mesopores was avoided. The samples
analyzed by X-ray photoelectron spectroscopy (XPS) included 5 % Ru/C, 5 % Ru/Al2O3, 5 % Ni/C, 5 % Ni/Al2O3, 5 %
Pd/C, and 5 % Pd/Al2O3. All these powder materials needed to be compacted before being introduced in the XPS
equipment. Given the impossibility of compacting active carbon-supported materials, they were mixed with
calcined Al2O3 (the same used as support for the other catalysts), in a 1:1 proportion (C-based catalyst: Al2O3). XPS
analyses were carried out in a PHOIBOS HSA3500 150 R6 spectrometer, with Al as source (250 W). Spectra were
calibrated by using the C 1s peak at 284.6 eV. Analyses were first carried out on fresh samples, which were then
treated under H2 and temperature for 1 hour, acquiring the corresponding spectra afterwards. Pd/C and Pd/Al2O3
were treated at 190 °C, Ru/C and Ru/Al2O3 at 250 °C, Ni/C at 400 °C and Ni/Al2O3 at 600 °C, as the latter was the
maximum temperature allowed by the equipment. The binding energies of Ru 3p3/2, Ni 2p3/2, and Pd 3d5/2 have
been studied and compared for the different catalysts before and after the H2 treatment (Table 1).
2.4.Synthetic procedures.
PPE was synthesized in the laboratory according to the procedure described by Rensel et al.[59] Concisely, 4 g of
phenol (99 % purity, ACROS Organics) were added to a round bottom flask together with 6 g of K2CO3 (99 %
minimum purity, ACROS Organics) and 33 mL of anhydrous acetone (synthesis grade, Fisher Chemical). Then, after
1 h of reflux, 11 g of anhydrous (2-bromethyl)-benzene (Sigma Aldrich) were added to the mixture. After 24 h of
reaction, 25 mL of ethyl ether and 25 mL of water were added. Ethyl acetate (99.8 % purity, Sigma Aldrich) was
used to extract the organic compounds, which were washed three times with a solution 1 M of sodium hydroxide
(NaOH, reagent grade, 98 % minimum purity, Sigma Aldrich) and another 1 M of sodium chloride (NaCl, 99.5 %,
ACROS Organics). Then, the ethyl acetate phase was evaporated in a rotary evaporator until only styrene, phenol
and phenethyl phenyl ether were left. The desired compound (PPE) was separated by flash chromatography using
a mixture of hexane and ethyl acetate as eluent.
2.5.Depolymerization of 2G lignin.
Catalytic depolymerization of the stillage cake was carried out in the same installation described for
hydrodeoxygenation of the lignin model compounds. The catalyst used for this preliminary test was 5 % Ru/C,
described in the Catalyst preparation section.
To carry out 2D NMR (HSQC-2D) analyses, ligninic samples were acetylated following this procedure: the
acetylation process was carried out in an acetic anhydride/pyridine solution (1:1, v/v) stirred at room
temperature for 48 hours. 1 mL of solution was used for every 50 mg of ligninic sample. After acetylation, the
reaction mixture was concentrated under vacuum, the crude residue was dissolved in dimethyl sulfoxide (DMSO,
99.5 % minimum purity (GC), plant cell culture tested, Sigma-Aldrich), centrifuged and then the acetylated lignin
7

dissolved in DMSO was recovered and finally lyophilized. The product obtained after lyophilization was dissolved
in deuterated dimethyl sulfoxide (DMSO-d6, “100%”, 99.96 atom % D, Sigma-Aldrich) to perform HSQC-2D
analyses.
NMR spectra were run in DMSO-d6 on acetylated samples, to avoid fractionation of the material before NMR
analysis and to increase both the solubility and the chemical shift dispersion of the side chain units. The inverse
detected 1H–13C correlation spectra (HSQC) were measured at 25 °C on a Bruker AVANCE III 700 MHz instrument
equipped with a cryogenically cooled 5mmTCI gradient probe with inverse geometry (proton coils closest to the
sample). HSQC experiments used Bruker’s “hsqcetgpsisp2.2” pulse program (adiabatic-pulsed version) with
spectral widths of 5000 Hz (from 10 to 0 ppm) and 20,843 Hz (from 165 to 0 ppm) for the 1H- and 13C
dimensions. The number of collected complex points was 2048 for the 1H-dimension with a recycle delay of 1.5 s.
The number of transients was 64, and 256 time increments were always recorded in the 13C dimension. The 1JCH
used was 145 Hz. Processing used typical matched Gaussian apodization in the 1H dimension and squared cosine-
bell apodization in the 13C dimension. Prior to Fourier transformation, the data matrixes were zero-filled up to
1024 points in the 13C-dimension. The central solvent peak was used as an internal reference (δC 39.5; δH 2.49).
Long range J-coupling evolution times of 66 and 80 ms were used in different heteronuclear multiple bond
correlation (HMBC) acquisition experiments. As 2G lignin was obtained from the stillage of 2nd generation
bioethanol production from corn stover, HSQC correlation peaks were assigned mainly by comparing with Min et
al. [60] study on the structural changes of lignin and lignin-carbohydrate complexes in corn stover, though other
references were also considered [61,62].
3. Results and Discussion
BPE, DPE and PPE (Figure 2) have been chosen as the model compounds of the most abundant lignin linkages, β-
O-4, α-O-4 and 4-O-5, respectively, (Figure 1) to perform the hydroprocessing studies over Ru, Ni and Pd catalysts.
3.1.Catalyst screening.
Firstly, several catalysts were tested for the hydroprocessing of benzyl phenyl ether (BPE). BPE was selected as
the benchmark model compound to test the catalytic activity of several materials for two main reasons: being
commercially available (when compared to PPE) and having a lower energy of C-O bond dissociation32 (when
compared to PPE and DPE). A total of nineteen catalysts were tested for this purpose, with Ru, Ni and Pd selected
as active metals. Supporting materials included: active carbon, multiwall carbon nanotubes, Al2O3, ZrO2, TiO2 and
TiO2/C. Reactions were carried out at the following conditions: 150 °C, 2 h of reaction, 25 bar-g of H2, 20 mL of
stock solution (16 mM, BPE in methanol) and 750 rpm (RCF = 4.55·g). After this screening, the most promising
catalysts were selected to be further tested in the hydroprocessing of phenethyl phenyl ether (PPE) and diphenyl
ether (DPE).
Figure 3 shows the possible reaction pathways in the hydroprocessing of BPE in methanol. As it can be observed,
the BPE molecule can undergo hydrogenation of one of the rings until the complete saturation of the dimer (steps
1, 2, 5 and 8); a cleavage of the partially-hydrogenated dimers into the corresponding monomers (steps 6, 7, 9
and 10) and subsequent hydrogenation of those (steps 11 and 14- 17); or the direct cleavage of the C-O-C bond to
the corresponding monomers and subsequent hydrogenation of those (steps 3, 4 and 11-17). The main products
observed after reaction of BPE (identified by GC-MS) are shown in Figure S1. Compounds originated from both
the hydrogenolysis of the C-O bonds and the hydrogenation of the dimer were observed. In addition, it was
observed that, depending on the catalyst used, aromatic monomers produced from the hydrogenolysis of the C-O
bonds underwent further hydrogenation. Regarding cyclohexanone, the currently discussed pathways [63]
suggest that, using similar catalysts as the ones selected here and in the liquid phase, phenol hydrogenates
sequentially. First, phenol hydrogenates to cyclohexanone, and then cyclohexanone is further hydrogenated to
8

cyclohexanol, therefore explaining the presence of both products in the mixture after reaction [64]. In addition,
methoxycyclohexane and 1,1-dimethoxycyclohexane were also detected as products. They were generated due
to the acetalyzation of cyclohexanone with methanol, which has been reported in the presence of a solid acid
catalyst [48,65]. Consequently, both compounds have been considered as cyclohexanone in carbon mass balances
and for yield calculations, i. e. 6 carbons instead of 7 or 8, as the extra carbons come from the reaction solvent
(methanol).
Results for each tested catalyst are presented in Figure 4 in terms of conversion of BPE and yield. To ease the
comparison between catalysts, the different products were grouped in four families of molecules. These four
families included: fully saturated dimers (FSD), partially saturated dimers (PSD), saturated monomers (SM), and
aromatic monomers (AM). Taking as reference Figure S1, the first line of compounds (toluene, phenol and
phenylmethanol) was considered under the group of aromatic monomers (AM). The saturated monomers
included cyclohexanone, cyclohexanol, methylcyclohexane, cyclohexane, cyclohexylmethanol,
methoxycyclohexane, and 1,1-dimethoxycyclohexane. Finally, (cyclohexylmethoxy)-benzene and
(cyclohexyloxy)methylbenzene were the partially saturated dimers (PSD) and (cyclohexylmethoxy)-cyclohexane
was the fully saturated dimer (FSD).
Based on the activity and product yield values obtained in this screening study, the following catalysts were
selected to carry out further catalytic tests with PPE and DPE:
-
-
5 % Pd/C, due to its high activity (100 % conversion in 2 h) and also high selectivity to aromatics, as the
yield to aromatic monomers was over 90 %.
5 % Ru/C. A high conversion value was also obtained with this catalyst, though the yield to aromatic
monomers was lower than that obtained with 5 % Pd/C. Nevertheless, the energy of C-O bond
dissociation is greater for PPE and DPE and thus it was interesting to keep this catalyst as candidate to
further evaluate its performance.
-
-
5 % Ru/Al2O3 and 5 % Ru/MWCNT. These two materials present similar conversion levels, but the yield to
aromatics was greater with 5 % Ru/Al2O3. Although the conversion achieved with them was lower than
with the previous ones, they presented a notable ratio between AM and the rest of products.
5 % Ni/C and 5 % Pd/Al2O3. They attained a considerable level of conversion, bearing in mind the change
of active metal from Ru, but more importantly, their selectivity to aromatic monomers was high. None or
almost negligible yield to saturated products was observed, either to the saturated monomer family or to
any of the saturated dimers.
-
5 % Ni/MWCNT and 5 % Ni/Al2O3, in order to compare the effect of changing the active metal while
keeping the same support (Ni/MWCNT versus Ru/MWCNT and Ni/Al2O3 versus Ru/Al2O3 or Pd/Al2O3).
The rest of catalysts were discarded for further experiments with model compounds. For some of them, the
observed conversions were significantly low (5 % Ni/ZrO2 comm, 5 % Ni/TiO2/C or Ru/TiO2 lab). For others, better
results regarding yield and/or conversion were observed when using the same active metal in a different support.
Therefore, from the 19 catalysts screened in the hydroprocessing of BPE, 8 were finally selected to be tested with
the model compounds of the other two ether bonds in lignin. These were: 5 % Pd/C, 5 % Ru/C, 5 % Ru/Al2O3, 5 %
Ru/MWCNT, 5 % Ni/C, 5 % Pd/Al2O3, 5 % Ni/MWCNT and 5 % Ni/Al2O3.
3.2.Hydroprocessing of model compounds (BPE, PPE and DPE).
Figure 5 presents the results obtained with each of the 8 selected catalysts after reaction with PPE and DPE. In
addition, results obtained with the same catalysts in the hydroprocessing reaction of BPE are also included for
comparison. The average mass balance closure for the presented experiments was 93  3%, which is assumed as
9

acceptable. The calculated standard deviation of the experiments was bellow 5 % for the model compounds
studies.
Deviations may be a consequence of two factors: the presence of compounds which cannot be quantified due to
their concentration being under the detection limit of the GC-FID system; or to errors in the calculations derived
from the use of response factors estimated by correlating the structure of the unknown compound with
commercially available standards [66].
Figures 6 and 7 show the proposed reaction networks in the hydroprocessing of PPE and DPE, respectively.
Figures S2 and S3 show the products identified by GC-MS after catalytic reaction of PPE and DPE, respectively.
Similar to the reported mechanism for BPE, observed products can be related to the hydrogenolysis of the C-O-Ar
bonds (in the form of aromatic or saturated monomers) or to the hydrogenation of the dimer (PPE or DPE in this
case). In addition, after reaction of PPE and DPE, cyclohexanone, methoxycyclohexane and 1,1-
dimethoxycyclohexane were also observed. The same explanation applies as the one for the case of reaction with
BPE for the presence of these compounds.
According to the results presented in Figure 5, some general conclusions can be drawn regarding the effect of the
type of ether linkage processed, the effect of the active metal and the effect of the support on the conversion
levels and on the product distribution. When comparing the energy values of the bonds of the three selected
model compounds (BPE, PPE and DPE), it can be noted that the most labile bond is the α-O-4 (BPE, 218 kJ/mol), as
commented before, followed in terms of energy by β-O-4 and 4-O-5 linkages (PPE, and DPE, 289 and 314 kJ/mol,
respectively) [32]. In general, it was observed that, for any given catalyst, the achieved conversion level (keeping
constant the rest of operating parameters) decreases from BPE to PPE to DPE. Taking Pd/C as example, a
conversion of ca. 100 % was obtained in reaction with BPE. The conversion value decreased to approximately 50
% for PPE, and finally the obtained conversion of DPE was 19.3 %. In fact, the three catalysts selected from the
screening that produced the lowest conversions of BPE, i. e. Pd/Al2O3, Ni/MWCNT and Ni/Al2O3, were not able to
convert neither PPE nor DPE under the studied conditions. This can be clearly connected to the aforesaid bonding
energy of the different model compounds. Thus, the dimer with the most labile linkage (α-O-4) was the one to
undergo the greatest conversion. Conversely, the conversion of the dimer with the strongest 4-O-5 bond was the
lowest.
Regarding the product distribution, it can be observed that it differed from one model compound to another.
Looking at the results obtained with Ru/C for the three model compounds, in the case of BPE the yield to
aromatic monomers was around 53 %, leaving the rest of the products to fully saturated monomers (26 % of
yield) and partially saturated dimers (19 % of yield). For PPE, though the conversion of the model compound was
within the same range, the distribution to the product families was different: 59 % of the obtained products
corresponded to fully saturated monomers whereas aromatic monomers represented only 20 % of the final
products. The remaining products from PPE conversion were distributed as follows: 3 % were partially saturated
dimers and 18 % were fully saturated dimers. Finally, in the case of carrying out the reaction with DPE as model
compound, it can be seen that the main products obtained were aromatic monomers (37 % of yield), followed by
partially hydrogenated dimers (18 %), fully hydrogenated monomers (4.8 %) and fully hydrogenated dimers (1%).
On one hand, the higher C-O energy bond dissociation for DPE may explain why the total yield to monomers, both
aromatic and saturated, is lower (nearly 42 %) than that obtained with BPE and PPE. On the other hand, the
results also seem to indicate that the reaction mechanism varies from one compound to another, therefore
changing the distribution of products. For example, for PPE it seems that hydrogenation of the dimer and
hydrogenolysis of the C-O bond are competing pathways, whereas for BPE it seems that the hydrogenolysis of the
α-O-4 ether bond is more important than dimer hydrogenation. It is also interesting to notice that if selectivity to
fully and partially saturated dimers is calculated, an increase in the sum of both selectivities can be observed
when changing from BPE to PPE and to DPE for some catalysts. Following the same Ru/C example, selectivity to
fully or partially saturated dimers was 18 % when the model compound was BPE. In contrast, selectivity to these
10

families was 21 % after PPE reaction, and increased up to 31 % for DPE. The latter tendency was observed for
reactions with all active carbon-supported catalysts. In addition, in the case of Ru/Al2O3 and Ru/MWCNT,
selectivity to fully or partially saturated monomers increased from BPE to PPE, but then decreased from PPE to
DPE, in contrast with the observed behavior in active carbon-supported catalysts.
These results seem to indicate that conversion and product distribution to the different families is dependent on
the type of ether compound but also on the nature of the catalyst. In this sense, it was observed that, for any of
the tested catalyst and under the same reaction conditions, achieved conversion of the model compounds
decreased as follows: BPE > PPE > DPE. Interestingly, the energy necessary for scission of the C-O bond increases
as BPE < PPE < DPE. In addition, for those catalysts with active carbon as support, the selectivity to saturated
dimers increased in the order BPE<PPE<DPE. Thus, it can be said that, when the catalyst was not able to quickly
break the C-O-C bond, saturation of the rings of the dimers gained importance. Nevertheless, the absence of the
latter behavior when using Ru/Al2O3 or Ru/MWCNT as catalysts seems to indicate that product distribution is also
affected by the nature of the selected catalyst. A possible explanation might be that the way in which each dimer
is absorbed on the catalyst surface varies, depending on the dimer itself, but also on the characteristics of the
catalyst surface (metal particle size, support affinity, etc.), which will be discussed in depth later.
Regarding the effect of the active metal, when comparing the results obtained for the same model compound,
operating conditions and catalyst support, it can be observed that conversion of dimers decreased from Ru to Pd
and to Ni. Interestingly, this decrease in conversion caused also an increase in the selectivities to aromatic
monomers. For example, if reaction results of PPE hydroprocessing over Ru/C, Pd/C and Ni/C were compared, it
could be observed that conversion decreased in the order Ru/C (100 %)>Pd/C (50 %)>Ni/C (18 %) while selectivity
to aromatic monomers increased as Ru/C (20 %) < Pd/C (30 %) < Ni/C (75 %). In this sense, Zhang et al. [67]
reported the hydrogenolysis of 2-phenoxy-1-phenylethan-1-ol using unsupported NiM (M = Ru, Rh and Pd)
catalysts. When studying the reaction on monometallic catalysts (Ru, Pd, and Ni) it was observed that noble
metals (Ru and Pd) achieved higher conversions (ca. 100 %) but with low selectivity to monomers (0.3 and 2.3 %
for Ru and Pd, respectively), whereas only 30 % of conversion was obtained with monometallic Ni, but with a 57 %
of selectivity towards aromatic monomers.
3.3.Catalysts characterization.
The results seem to indicate that the characteristics of the selected catalysts highly influenced the obtained
product distributions. Different techniques, such as BET surface area, temperature programmed reduction (TPR),
high-resolution transmission electron microscopy (HRTEM) and X-ray photoelectron spectroscopy (XPS) were
used to characterize the catalysts.
N2-physisorption. The surface topology of the different catalysts was characterized using nitrogen physisorption.
The Brunauer–Emmett–Teller (BET) theory was employed to fit the N2-physisorption isotherms obtained for the
different catalysts and determine the surface area (see Table 1). As expected, catalysts supported on active
carbon had the highest surface areas, with values around 800-860 m2/g, followed by those obtained using carbon
nanotubes, that ranged between 200 and 300 m2/g. It is also remarkable the relatively high surface area obtained
on the alumina-supported catalysts, which had surfaces areas around 215-230 m2/g in all the cases.
3.3.1. High-resolution transmission electron microscopy (HR-TEM). Table 1 also summarizes the average particle
size values of Ru, Ni and Pd in the analyzed catalysts by HRTEM, as well as the metal dispersion value. The
procedure used for the metal dispersion calculation can be found in the Supplementary Information (Table S1).
From the images of HRTEM of Ru catalysts (Figure 8), it can be observed that Ru particles are homogeneously
dispersed and present smaller particle size (2-4 nm) when they are supported in carbonaceous materials (active
carbon and multiwall carbon nanotubes) than when the support is Al2O3 (50 nm). When Ru is supported in the
11

latter material, the average particle size is greater and particles tend to agglomerate in regions of the support,
rather than being dispersed along the entire surface. In addition, it was observed that the active carbon selected
as support contained other elements as impurities, such as Fe, Cu, S and Ca. From Figure S4, a similar tendency
can be noticed when comparing Ni and Pd catalysts supported on active carbon and Al2O3. Roughly, when Ni or Pd
were supported on active carbon, particle size was smaller (3-5 nm, Table 1) than that observed for each metal on
Al2O3 (60 nm). Therefore, the metal dispersion highly decreases for the three metals when changing the support
to Al2O3 (Table 1). The metal dispersion values for Al2O3 supported catalysts are lower than 2 % for Ru and Ni, and
almost 10 % for Pd. However, when supported on active carbon, the dispersion increases to over 20 %, and
reaches a value of 47 % for Ru/MWCNT catalyst. Moreover, the tendency of metal particles to agglomerate
when the support was Al2O3 could also be detected. Finally, HRTEM images for 5 % Ni/Al2O3 where much more
difficult to interpret than the rest. Ni particles were difficult to identify due to low contrast of the obtained
images, which complicated the calculation of the mean particle size of this catalyst. This parameter was
nevertheless calculated using the particles more clearly identified, to reasonably compare the value with those of
the rest of the catalysts.
3.3.2. Temeprature programmed reduction (TPR). Figure 9 shows the TPR profiles of Ru and Ni catalysts supported
on Al2O3, active carbon and MWCNT. Ru main reduction peaks were observed at varying temperatures depending
on the support. For Ru/Al2O3, the main reduction peak appeared at 208 ᵒC and could be assigned to the reduction
of Ru oxides [68]. A second peak was observed at ca. 532 ᵒC, which could be the consequence of the reduction of
bulk Ru atoms having stronger interaction with the Al2O3 support [69]. Ru/MWCNT profile showed a peak at 98 ᵒC
with a shoulder at 149 ᵒC, and a wide peak at around 606 ᵒC. The first two peaks might be matched to the
reduction of ruthenium, as reported elsewhere [70,71]. Moreover, the reduction of surface oxygen complexes
(incorporated as a consequence of the purification process of CNT) has been observed from 397 to 997 ᵒC, thus
explaining the wide peak observed with the maximum at 606 ᵒC [72]. Finally, Ru/C presented a double peak with
maximums at 171 and 244 ᵒC which could be attributed to the reduction of RuOx species and also to the reduction
of bulk RuCl3 [73]. The broad peak with maximum at 512 ᵒC was attributed to the partial gasification of the
carbonaceous support [74,75].
Likewise, the TPR profiles of Ni-based catalysts showed different reduction peaks depending on the supporting
material. For the Ni/Al2O3 sample, a small peak at 435 ᵒC was observed, together with a wider peak with
maximum at 651 ᵒC and a shoulder at ca. 737 ᵒC. The peak at the lowest temperature could be attributed to the
reduction of NiO species with weak interaction with the support [76]. The peak at 651 ᵒC could be connected to
the reduction of non-stoichiometric nickel aluminate species [76], and the shoulder at 737 ᵒC could be related to
the reduction of NiO surface species with strong interaction with the support [77] and to the partial reduction of
the NiAl2O4 spinel [78]. For the peak at 510 °C in the Ni/MWCNT sample, highly dispersed NiO particles with
strong interaction with the support have been reported to produce peaks in the 420-560 °C range [79], but also
peaks at around 550 °C have been related to partial methanation of CNT [80]. Thus, the observed peak at 510 °C
can be a consequence of both contributions. In addition, the small peak observed at ca. 99 ᵒC in the TPR profile of
Ni/MWCNT could be connected to water desorption. MWCNT were not dried before performing the TPR analysis,
thus it is possible that some water adsorbed in the material could have been desorbed upon heating, producing
the observed signal. Finally, Ni/C TPR profile presented a sharp peak at 398 °C and a wide peak with maximum at
552 °C. The former might be connected to the reduction of NiO species with weak interaction with the support,
whereas the second can be attributed to a mixture of the reduction of NiO species with stronger interaction with
the support together with partial gasification of the carbonaceous support as seen with the Ru/C catalyst.
Focusing on the Ni-based catalysts, it was observed that overall conversion of any of the dimers and product
distribution to the different families of compounds were different when the support was changed (Figure 5). In
the results obtained using BPE as model compound, it was observed that conversion decreased as follows: Ni/C >
Ni/MWCNT > Ni/Al2O3. According to the temperature programmed reduction (TPR) characterization results
(Figure 9), Ni supported on active carbon presented the reducible NiO species at the lowest temperature (400
ᵒC), which seems to indicate that, under the selected activation conditions, most of the nickel in the catalyst was
12

in its metallic reduced form. In fact, XPS analyses (Table 1) showed that treating Ni/C in H2 atmosphere at 400 ᵒC
was enough to convert NiO in metallic Ni. Similarly, TPR profile of Ni/MWCNT showed reduction peaks at higher
temperatures (550 ᵒC) than those for Ni/C, and finally XPS and TPR characterization of Ni/Al2O3 showed the
presence of nickel aluminate and NiAl2O4 spinel, which are species difficult to reduce. This indicates that as the
reducibility of the catalyst increases the conversion of the dimer is favored. It is also worth noting that, according
to HRTEM results, metal particle size increased from 5 to 62 nm when comparing Ni/C and Ni/Al2O3 (Table 1,
Figure S4), which may indicate also the importance of the effect of metal particle size. One can assert that the
lower activity of the Ni/Al2O3 catalysts compared to the C could be associated to lower metal dispersions and the
formation of partially unreduced Ni+ species and nickel aluminates that have a lower catalytic activity.
3.3.3. X-ray photoelectron spectroscopy (XPS). The samples analyzed by X-ray photoelectron spectroscopy (XPS)
included 5 % Ru/C, 5 % Ru/Al2O3, 5 % Ni/C, 5 % Ni/Al2O3, 5 % Pd/C, and 5 % Pd/Al2O3. Analyses were first carried
out on fresh samples, which were then treated under H2 and temperature for 1 hour, acquiring the corresponding
spectra afterwards. Pd/C and Pd/Al2O3 were treated at 190 ᵒC, Ru/C and Ru/Al2O3 at 250 ᵒC, Ni/C at 400 ᵒC and
Ni/Al2O3 at 600 ᵒC, as the latter was the maximum temperature allowed by the equipment. Binding energies of Ru
3p3/2, Ni 2p3/2, and Pd 3d5/2 have been studied and compared for the different catalysts before and after the H2
treatment.
Ru-based catalysts. The Ru 3p3/2 binding energies (BE) at which maxima of the primary peaks were found before
and after H2 activation in Ru/C and Ru/Al2O3 spectra are collected in Table 2. Values at 463.15 and 462.07 eV for
the Ru/C catalyst were attributed to RuO2 and metallic Ru, respectively, by comparison with literature [81].
Therefore, it can be said that, under the selected H2 treatment, RuO2 was reduced to metallic Ru, which was
assumed to be the active phase for the hydroprocessing of the lignin model compounds. According to Shen et al.
[82], increasing oxidation of a Ru wafer surface caused the presence of a second peak (after deconvolution of
3p3/2 spectrum) at higher binding energies that grew with increasing oxidizing treatment. Such peak, that
appeared at 465.0 eV was observed on Ru/Al2O3 after reduction in H2, was attributed to RuOx/Ru species.
Literature regarding XPS spectra for the ruthenium 3p level is scarce when compared to available data for 3d
level, which is the primary XPS region for this metal. Nevertheless, the spectra for this level presents strong
overlapping with C 1s level, and thus it was considered that, given one of the samples was supported on carbon,
elucidation of spectra would be complicated. In turn, reported binding energies (BE) for 3p3/2 level are mostly in
the 460-464 eV range and a match for the peak at 466.25 eV in Ru/Al2O3 could not be found in literature. Thus,
following the aforesaid discussion by Shen et al. [82] it was proposed that this peak corresponded to a mixture of
RuOx/Ru species, with a higher degree of oxidation than the peak observed after the H2 treatment of the sample.
These results seem to indicate that, under the selected pretreatment conditions (H2 at 250 ᵒC), oxidized
ruthenium particles could not be completely reduced.
Ni-based catalysts. In the case of the Ni/C catalyst, analyzing the Ni 2p3/2 spectra, the peak having a maximum at a
BE of 855.14 eV before the H2 treatment of the catalyst was attributed to Ni2+ of NiO [83]. After H2 activation of
the catalyst at 400 ᵒC, the maximum of the primary peak in the Ni 2p3/2 spectrum shifted to a BE of 852.05 eV,
which could correspond to metallic Ni [84,85]. The shift of the primary peaks BE found on Ni indicated that, under
the selected activation conditions for this catalyst, NiO is reduced to metallic Ni. Concerning the Ni/Al2O3 catalyst,
the Ni 2p3/2 spectrum revealed a primary peak having a maximum before H2 treatment at a BE of 859.15 eV, which
was attributed to NiO [86]. Interestingly, after the H2 treatment, the corresponding BE for the maximum of the
main Ni 2p3/2 peak increased to 861.90 eV, instead of decreasing. This could be a consequence of the formation of
13

nickel aluminates. Formation of nickel aluminates after calcination at 623-673 K has been reported in the
literature [87], and as discussed by Salagre et al. [88], the presence of such irreducible species was connected to
the presence of peaks in the Ni 2p3/2 spectra at ca. 863 eV. Moreover, the formation of structures that imply
strong metal-support interactions such as nickel aluminates or nickel silicates have been reported to cause a shift
of O 1s binding energy to higher values [89]. In this sense, the maximum of the primary peak of the O 1s spectra
for Ni/Al2O3 increases from BE of 534.15 to 536.63 eV when the sample is subjected to the H2 treatment at 600 ᵒC.
Pd-based catalysts. Pd/C catalyst presented BE energies for the maxima at 335.23 eV before the H2 treatment
and 335.15 eV after it. Such peaks have been attributed to Pd [90] and bulk Pd metal [91], respectively. In
addition, a secondary peak corresponding to the 3d3/2 level with a maximum at BE of 341.02 eV was found. The
presence of such peak together with a primary peak at BE values of 335.8 eV has also been attributed to Pd0 [92].
Thus, it could be concluded that the Pd in the fresh Pd/C catalyst sample was already present in its metallic form
(Pd0) without needing further activation. Regarding the Pd/Al2O3 catalyst, the maxima of the Pd 3d5/2 primary peak
before and after H2 treatment moved to higher BE when the support was changed from active carbon to Al2O3, as
observed in the cases of Ru- and Ni-based catalysts. The presence of surface Pd(OH)4 was proposed in previous
studies of Pd-based materials by XPS, reporting BE for the maxima of the peaks at around 338.5 eV [93], whereas
other authors endorsed maxima found at BE of 338 eV in Pd 3d5/2 spectra to PdO2 [94]. On the other hand,
similarly to what was discussed previously for Ru- and Ni-based samples, the shift in the BE of the maxima of the
primary and secondary peaks toward lower values (the latter, corresponding to the 3d3/2 level, shifting from ca.
345.29 to ca. 343.70 eV) was observed for the Pd/Al2O3 sample after the H2 treatment. This could be connected to
the fact that surface Pd species would have less degree of interaction with the support because of the H2
treatment, probably due to partial reduction of the surface Pd species. Considering the acidic nature of the
medium in which the Pd/Al2O3 catalyst was prepared and the rather oxidizing conditions of the calcination carried
out during its preparation, it could be concluded that the BE values of the maxima of the peaks obtained for the
samples analyzed before and after the H2 treatment could probably correspond to surface PdO2 with certain
degree of interaction with the Al2O3 support.
A more in-depth analysis of both spectra (including deconvolution of the secondary Pd 3d3/2 peaks at 345.29 and
ca. 343.70 eV, revealed that this peak could actually be the sum of two peaks, one at lower BE values that could
correspond to Pd0 and the other having the maximum at greater BEs, that could be attributed to Pd2+ species, in
agreement with similar results found for other impregnated Pd/Al2O3 catalysts, which also reported the
coexistence of surface Pd species with both electronic states [92]. Despite the use of different metal precursor
and calcination conditions, the fraction of metallic Pd to Pd2+ was also low here, though the differences in the BE
values of the different peak maxima, greater in this study, could be attributed to a stronger interaction of the Pd
species with the support (Pd-Al2O3). Also, when analyzing the spectra of the Pd/Al2O3 sample before the hydrogen
treatment it should also be noted that maxima of primary peaks at 339.2 and 339.9 eV have been previously
reported for the presence of surface Pd(CN)2 and K2(PdCl6) species [95]. Although it is not likely that these specific
structures were present in the sample, it could also be possible that the observed peak for 3d5/2 at 339.63 might
be connected to some PdCl2-related structures in the sample, as a consequence of the metallic precursor used to
prepare the catalyst. It should also be noticed the presence in both reduced samples (Pd/C and Pd/Al2O3) of a
secondary peak with a maximum at BE around 560 eV, corresponding to another region of the spectrum
attributed to the Pd 3d3/2 level, which had previously been assigned to the existence of surface Pd0 [90].
XPS analyses for Ru/C and Ru/Al2O3 indicated that Ru was more easily reducible when supported on active
carbon, in agreement with TPR results (Figure 9). On the other hand, the TPR profile of Ru/MWCNT showed that
14

reduction peaks were obtained at slightly lower temperatures than with Ru/C. In this case, the decrease in
conversion from Ru/C to Ru/MWCNT could have been caused by a fraction of the Ru particles being deposited
inside of the MWCNT tubes and thus being less accessible to the dimers to react. Regarding Pd catalyst, XPS
analyses (Table 1) revealed again that Pd was more difficult to reduce when supported on Al2O3 than on active
carbon, shifting Pd 3d5/2 signal to higher binding energies. In the case of supporting Ru and Pd, differences in
particle size were also observed when changing the supporting material. HRTEM images of Ru catalysts (Figure 8)
showed that Ru particle size (Table 1) increased from Ru/MWCNT (2 nm) to Ru/C (4 nm) and Ru/Al2O3 (52 nm).
For Pd, when it was supported on C mean particle size was 3 nm, whereas this value increased to 8 nm when the
supporting material was Al2O3 (Table 1, Figure S4). Therefore, one could infer that, according to XPS, TPR and
HRTEM characterization, Ru, Pd and Ni were more difficult to reduce when Al2O3 was selected as support than
with carbonaceous materials due to the larger particle size and stronger metal-support interactions between the
metals and Al2O3. This may explain the differences observed in product distribution when changing from one
support to another.
To evaluate the overall performance of each catalyst considering the three model compounds, a weighting factor
was defined. For each model compound, the catalyst that provided the highest AM/(SM+PSD+FSD) ratio was
assigned with a 5, the one that provided the following ratio with 4 and then sequentially down to 1 for the
catalyst yielding the lowest ratio. Then the three values assigned to each catalyst (one for each model compound)
were added, obtaining the weighting values shown in Figure 6. In this way, it is possible to identify the best
catalyst for the three different types of C-O bonds. It was observed that the catalyst that provided the highest
value of this factor was Ru/Al2O3, due to achieving medium values of production of aromatic monomers from the
three dimers when compared to other catalysts. The following catalyst with the highest weighting factor was
Ni/C, as it provided an elevated concentration of aromatic monomers in the hydrogenolysis of BPE and PPE, but
aromatic monomers were not observed when DPE was the model compound. Moreover, Pd/C also provided a
high weighting factor, mostly due to the high selectivity to aromatic monomers achieved for BPE. In the case of
Ru/C and Ru/MWCNT, the lower values of this ratio were mainly related to a greater value of saturated
monomers and partially saturated dimers.
On the other hand, both BPE and PPE present two different C-O linkages to be theoretically broken, whereas the
two DPE C-O linkages are equivalent. Subsequently, the preference of rupture of the C-O bond with the C
belonging to the aromatic ring (Car-O) versus that of the C-O bond with the C in the aliphatic chain (Cali-O) was also
calculated according to Equation 4. The results obtained are presented in Figure 11.
푚표푙 퐶 표ꢁ 푐표푚푝표푢푛푑푠 푝푟표푑푢푐푒푑 푡ℎ푟표푢푔ℎ 푠푝푒푐ꢀꢁꢀ푐 퐶푥푥 − 푂 푟푢푝푡푢푟푒
푅 =
(퐸푞. 4)
푚표푙 퐶 표ꢁ 푐표푚푝표푢푛푑푠 푝푟표푑푢푐푒푑 푡ℎ푟표푢푔ℎ 푎푛푦 퐶 − 푂 푟푢푝푡푢푟푒
15

It was observed that, with BPE as model compound, Cali-O scission occurred preferentially, with
only a minor rupture of the Car-O bond when the catalysts were Pd/C, Ru/Al2O3 or Ru/MWCNT.
The preferential rupture of the Cali-O bond was an expected result, due to its bonding energy
being 96 kJ/mol lower when compared to that of the Car-O bond. For instance, J. He et al.
showed that Ni/SiO2 catalyst preferentially depolymerize the Cali-O bond of the BPE in aqueous
and organic (e.g. hexadecane) environments [32]. The high selectivity observed was attributed
to the differences in bonding strengths of the two bonds. Similar results were reported by Z. Luo
and C. Zhao using Ru and Pd catalysts supported on activated carbon and sulfated zirconia for
the selective hydrogenolysis of BPE [96]. In this kinetic study they showed that the rate on metal
surfaces, in the absence of acid sites, is the cleavage of the C-O bond. Therefore, the stronger
the C-O bond energy the higher the activation barrier. In contrast, when studying this
phenomenon with PPE as model compound, it was observed that while Cali-O scission was still
majoritarian in most cases, an increment of products obtained from the rupture of the Car-O
bond is produced. It should be noticed that, in this case, the bonding energy of the Cali-O bond is
only 25 kJ/mol lower than that of the Car-O bond, which may explain the greater competence
between the rupture of both. Interestingly, whereas with Ni/C no products from Car-O scission of
PPE were observed, when using Pd/C products from such rupture exceeded those obtained from
the hydrogenolysis of the Cali-O bond. This seems to indicate, as commented previously, that
product distribution is affected both from the strength of the C-O bond but also from the specific
characteristics of the catalyst employed.
To rationalize these results first the possible surface reaction mechanism must be taken into
account. Recently, J. He et al. [57] showed through kinetic studies and Density Functional Theory
(DFT) that on Ni (111) catalyst the hydrogenolysis of Car-O bonds of 4-4´-dihydroxydiphenyl ether
starts with the non-dissociative chemisorption of the molecule in a co-planar configuration.
Then, the Car-O bond cleavage takes place leading to the formation of HOC6H4* and OC6H4OH*.
The latter intermediate latter undergoes a second Car-O hydrogenolysis step leading to the
formation of HOC6H4* and O* surface species. Finally, the two surface HOC6H4* undergo
hydrogenation to generate phenol. In the case of the present study, it could be proposed that
hydrogenolysis of the Car-O ether bond of the BPE will require the co-planar adsorption of the
aromatic ring containing the sp carbon bonded to the oxygen. On the contrary, the cleavage of
the Cali-O bond will require the binding of the carbonyl carbon of the aliphatic chain and the
oxygen. Therefore, it is possible to interpret the differences in selectivity observed on BPE
hydrogenolysis in terms of the interaction of the aromatic ring C6H6 and Cali-O to the metal
surface. The interaction of benzene and CO with metal surfaces has been studied in detailed
using experimental [97-99] and theoretical methods [100-102]. While, quantitatively, the
accuracy of first principle calculations is significantly sensitive to the proper description of the
surface, Van der Waals interactions, crystalline structure of the surface, and orientation of the
molecule, the qualitative information can be used to obtain a deeper understanding of the
underplaying phenomena [100,101]. Aromatic molecules tend to adsorb flat-lying onto the
metal surface via covalent bonding of
 orbitals of the aromatic ring with d states of the metal.
When comparing the theoretically calculated adsorption energies of benzene adsorption on
ferromagnetic Ni (111) with platinum group metals such as Ru(0001) and Pd (111) significant
differences arise. While on Ni (111) the adsorption energy of benzene varies between 0.91 and
1.00 eV (20.98 kcal/mol) [103,104],which are not far from experimentally obtained adsorption
energies of 0.78 eV [105], the binding to Ru(0001) and Pd(111) is higher with approximate values
of 1.31 eV (30.21 kcal/mol) [106] and 1.03-1.19 eV (23.75-27.44 kcal/mol) [107,108],
respectively. Thus, the C6H6-metal interaction increases as follows; Ru > Pd > Ni. In contrast, the
16

adsorption energy of CO on Ni, Pd, and Ru follows a completely different trend. Theoretical and
experimental data show that among the three metals Ru (001) has the higher adsorption
interaction with CO with values of approximate 48.6 [109] and 41.8 kcal/mol [110], respectively.
On Ni (111) the calculated adsorption energy of CO varies between 34.0 and 51.3 kcal/mol
depending on the cluster size [109], while the experimental value is around 36.6 kcal/mol [111].
Notably, calculated adsorption energies of CO on Pd decreased to values ranging from 25.5 to
37.7 [109], which were similar to experimental adsorption heats obtained by temperature
programmed desorption (30 kcal/mol) [112]. Therefore, the CO-metal surface chemisorption
energy increases as follows; Ru > Ni > Pd. In consequence, the differences in selectivity observed
on BPE and PPE may be attributed to both the lower bond strength of Cali-O compared to Car-O
and to the affinity of the metal surface to the aromatic ring and carbonyl moiety. The latter
effect becomes more pronounced when the differences in bond energies of C-O are smaller, like
the one observed on BPE. For instance, when the reaction is performed on Ni/C the ratio of Cali-O
to total is nearly one. In this case the interaction of the aromatic ring and the surface is
significantly lower than that of the Cali-O with surface and as result the sp carbon of the aromatic
ring bonded to the ether oxygen is not activated. The opposite occurs on Ru/C and Pd/C where
the ratios of Car-O to the total were 0.35 and 0.55, respectively. Here, the interaction of the
aromatic ring with the Ru and Pd surfaces facilitated the activation of the aromatic ring. Notably,
on Pd the lower interaction with CO could explain the higher selectivity to Car-O hydrogenolysis
observed when compared to Ru and Ni. These observations will be in line with the previously
reported results. For instance, S. Sitthisa and D. E. Resasco studied that the hydrodeoxygenation
of furfural over Pd, Cu, and Ni [113]. In that case, the authors observed that on Ni catalyst the
ring-opening reaction (i.e. C-O hydrogenolysis) was favored, while on the Pd catalyst
decarbonylation and hydrogenation reactions were predominant. The authors found that while
the aromaticity of the furan makes the furural more stable the activity for ring-opening is much
higher with furan than with tetrahydrofuran. They attributed this enhancement to the strong
interaction of the
 electrons of the furan aromatic ring with the metal surface [114].
3.4.
Hydroprocessing of lignin-containing 2G stillage.
A preliminary study on the influence of the presence of a catalyst in the depolimerization of
lignin-containing 2G stillage (named as 2G lignin) was carried out. Several tests were undertaken
to determine more clearly the actual influence of the catalyst on the reaction. For this purpose,
experiments 1, 2, and 3 were carried out (see Table 3 for operating conditions), each of them
having a total running time equal to 4 h. Experiments 1 to 3 were carried out in the laboratory-
scale set-up described in the experimental section. The catalyst used in reaction 3 was 5 wt. %
Ru/C. Figure 12 shows the overall results in terms of yield to solid and liquid products for these
experiments.
17

Mass balance closures were in the 90-100 wt. % range, as can be seen in Figure 12. Variations
were attributed to experimental deviations in the quantification of solid and liquid products. It
was observed that, even from just heating 2G lignin in methanol at atmospheric pressure
(Blank), some yield to liquid products was obtained (30 wt. %). Most likely, these products may
have originated from small molecules contained within the stillage cake, that under heating and
stirring were dissolved into the methanol used as solvent. The source of such molecules can be
found in the process from which 2G lignin was recovered from the raw stillage sample. As 2G
lignin was the cake obtained after filtration, it is plausible that during filtration small molecules
dissolved in the liquid fraction of stillage could be retained in the cake, thus being dissolved
afterward when heated in methanol. Moreover, it was observed that, when 2G lignin was
heated in N2 atmosphere and 200 ᵒC (Experiment 1), the recovered liquid fraction increased
from 30 to 37 wt. %.
Regarding Experiment 2, when the reaction was carried out in the presence of hydrogen, it was
observed that liquid fraction accounted for up to 48 wt. % of the recovered products, which
clearly indicates the cracking ability of H2 in the studied conditions. Finally, when a catalyst (5 %
Ru/C) was included in the process (Experiment 3), the highest yield to liquid products (ca. 56 wt.
%) was observed.
To have a deeper insight in the depolymerization effect of each of the studied factors, solid
samples after experiments 2, and 3 were analyzed by HSQC-2D. Figure 13 shows the HSQC-2D
spectra of acetylated 2G lignin, together with the identification of relevant regions attributed to
several substructures and types of linkages in lignin. In this sense, regions corresponding to
guaiacyl, syringyl and hydroxyphenyl units can be observed. Also, p-coumarate and ferulate
structures were identified in the spectra of 2G lignin. Regions denoting the presence of methyl
and methoxy groups can also be observed, the latter coming from the lignin structure itself and
from the acetylation process. Deuterated dimethyl sulfoxide (DMSO-d6), the solvent for the
HSQC analysis, was also detected. Moreover, some of the types of linkages that bond these units
were identified such as β-β or β-5, as well as β-O-4 ether linkages. For analysis of the samples
after catalytic depolymerization of 2G lignin, the regions corresponding to guaiacyl and syringyl
units (Figure 14), together with β-O-4 linkages are of particular relevance.
-O-4 linkages were selected as the target bonds for depolymerization of 2G lignin, since they are the
most abundant ether bonds in lignin, accounting from 40 to 60 % [115-117]. Thus, the determination of
their abundance in the native sample, as well as in the samples after reaction, is necessary. Several
quantitative and semi-quantitative methods have been proposed to determine the abundance of
relevant units and inter-unit linkages in different lignin samples [61,118-120]. Most of them are based
on the integration of the areas corresponding to the C9-units (hydroxyphenyl (H), guaiacyl (G) and
syringyl (S) structures presented in Figure 14) and the relative abundance of the desired inter-unit
linkage was calculated referring it to the total integrated area of C9-units. Differences between methods
are mostly in connection with the relative weight of each C9-unit in the total integrated area. In some
cases, S2/6 (syringyl) regions are corrected by a factor of 0.5,388 whereas in some others all the areas
are added without further correction [60]. Given the characteristics of the material, it was proposed to
calculate the proportion of β-O-4 linkages by adding the total areas of G and S subunits identified in
Figure 13 and assume them to stand for 100 % (100% of C9). The selection of these two subunits alone
was justified by being the most easily identifiable and abundant in the obtained spectrum. Then, the
area corresponding to the Hα in β-O-4 linkage was integrated and related to that of total C9, thus
18

calculating the percentage of this linkage in the sample. Figure 15 shows the relevant regions integrated
for the determination of the share of β-O-4 linkages in 2G lignin.
Following the procedure described above, it was calculated that the raw material used in
catalytic depolymerization studies contained ca. 46.6 % of β-O-4 bonds. The latter value is within
the range reported by Min et al. [60], who observed that β-O-4 were the most abundant
linkages, with a share between 40 % and 60 %. The same calculation procedure was followed
with the HSQC spectra of the acetylated samples after reaction, obtaining the percentage of β-O-
4 linkages in each sample. Then, by comparison with the initial results for acetylated 2G lignin,
the percentage of disappearance of such bond was calculated. This value was selected as the
factor to describe lignin depolymerization achieved after each reaction. Figure 16 shows the
integrated HSQC spectra from acetylated lignin after Experiment 3.
Based on the results observed in HSQC-2D, it was possible to quantify the degree of
depolymerization of the β-O-4 linkages after the different reactions performed. The reported
results (Table 4) indicate that disappearance of β-O-4 linkages increases from 7 % to 10 % when
comparing the blank experiment and Experiment 2, but more importantly, this percentage rises
to 53.1 % when using 25 bar of H2 combined with a catalyst (Experiment 3). Therefore, the effect
of a catalyst was proven to be positive for the depolymerization of β-O-4 bonds in 2G lignin. In
addition, to further analyze the effect of the different conditions applied in experiments 1 to 4,
liquid products from these experiments were analyzed by GC-MS. The corresponding
chromatograms (together with a blank) are depicted in Figure S5.
As a first appreciation, it can be observed that GC-MS analyses of the liquid samples
corresponding to experiments 1 to 3 presented their main peaks at almost equal retention
times, which indicated that the most important compounds for each sample were the same.
Nevertheless, some differences can be observed between the four chromatograms. Firstly, the
number of detected compounds grew in the following order: Experiment 1 < Experiment 2 <
Experiment 3. This is in accordance with the already observed behavior of increasing liquid yield
and depolymerization of the β-O-4 linkage. Therefore, it seems that even though some
compounds were extracted or dissolved by only heating the cake in methanol (blank), an
increase in temperature (experiment 1) but most importantly, the use of H2 and H2 plus a
catalyst (experiments 2 and 3, respectively), improve all the parameters that indicate a higher
degree of depolymerization of the sample. In this sense, it is significant the increase of
compounds detected with retention times above 19 minutes, which present higher molecular
weights. This indicates that lignin structure is being attacked yielding aromatic species such as 1-
methoxy-2-(methoxymethyl)benzene or methyl (E)-3-(4-hydroxy-3-methoxyphenyl)acrylate. In
addition, the most important detected compounds of the samples in Figure S5 are presented in
Table S1.
The tabulated compounds (Table S1) can be grouped into four main families: short-chain linear
compounds, nitrogen-containing compounds, benzene-based monomers, and esters of fatty
acids. Benzene-based monomers, which represented the greatest share of detected compounds
(> 66 wt. % of GC-FID-detectable compounds in all the cases), were clearly derived from the
polyphenolic structure of lignin. Monomers presenting syringylic, ferulic, hydroxyphenolic and
coumaric structures were detected and such monomers can be matched to the already
identified sub-structures in 2G lignin by HSQC. Several studies have reported similar aromatic
monomers after lignin depolymerization reactions [16,24,44,121,122]. For example, Barta et al.
19

[44] reported C9 catechols as major products after depolymerization of organosolv lignin in
methanol media using a Cu-PMO catalyst. On the other hand, Parsell et al. [122] observed 2-
methoxy-4-propylphenol and 3-(3-methoxy-4-hydroxyphenyl)-1-propanol as major products
after depolymerization of a synthetic lignin polymer which comprises only the β-O-4 ether
linkage. Such reaction was carried out in methanol at 150 ᵒC and 20 bar H2 using a bimetallic
catalyst system based on metallic Pd nanoparticles on carbon and Zn (II) Lewis acid. It should be
pointed out that, regarding this family of compounds, no saturated compounds derived from
this type of structures were detected. Therefore, lignin structure was depolymerized during
reaction, but saturation of aromatic rings did not take place to such extent to fall beyond the
detection limits of the GC-MS, and thus saturation could be considered as negligible.
While the results using Ru supported on activated carbon showed significant improvement over
the control experiments, it is clear that a significant fraction of the lignin is transformed to
liquids in the presence of methanol. The use of microporous catalytic materials like activated
carbon to perform depolymerization of a macromolecule like lignin is inevitably hindered by
mass transport limitations. While this point has been avoided in the literature, here we see that
a significant fraction of the lignin is depolymerize in the homogeneous phase, and the catalyst
helps to increase this fraction considerably. For this reason, additional studies using
heterogeneous catalysts for lignin depolymerization with an enhanced catalyst/lignin contact are
required to reduce those limitations.
4. Conclusions
Several catalysts have been studied for the hydrogenolysis of three model compounds of ether linkages
in lignin: benzyl phenyl ether (BPE), phenethyl phenyl ether (PPE), and diphenyl ether (DPE). Firstly, up
to 19 catalysts based on different active metals and supports were screened using BPE as model
compound, due to its commercial availability and lower dissociation energy for the C-O bond. From
these tests, 8 catalysts including; Pd/C, Ni/C, Ru/C, Ru/MWCNT, and Ru/Al2O3 were selected to further
study the hydrogenolysis of DPE and PPE. The results indicated that as the energy necessary for the
scission of the C-O bond increased (BPE<PPE<DPE), the conversion of such dimer for the same catalyst
decreased (BPE>PPE>DPE). In addition, for a given model compound, supporting different metals on the
same support caused also differences in the products obtained. In general, the activity of such metals
decreased from Ru to Pd to Ni, which in turn increased the selectivity to aromatic monomers in most
cases from Ru to Pd to Ni. Finally, the selected support also influenced the performance of the catalyst.
From the HRTEM and TPR characterization results it was observed that Ru, Pd and Ni were more difficult
to reduce and presented greater values of metal particle size when they were supported in Al2O3 than in
carbonaceous materials, which seemed to affect the performance of the catalysts. Those with easier
reducible species tended to perform better in terms of conversion of the dimer models.
When comparing the ratio of aromatic monomers to the rest of products obtained with several
catalysts, results observed for Ru/Al2O3 were interesting as it achieved a medium value of such
ratio for the three model compounds studied. If the model compound with the greatest bonding
energy (DPE) was not considered, Pd/C and Ni/C also provided interesting results, mainly due to
the significant yield of aromatic monomers obtained when studying the hydrogenolysis of BPE.
Finally, the preference of rupture of the C-O bond with the C belonging to the aromatic ring (Car-
O) versus that of the C-O bond with the C in the aliphatic chain (Cali-O) for BPE and PPE was
studied. It was observed that, as expected, in the case of BPE hydrogenolysis the rupture of the
Cali-O bond was preferential due to its lower binding energy, with only minor rupture of the Car-
20

O. In contrast, competence in the hydrogenolysis of both linkages was observed when PPE was
used as model compound. Although Cali-O scission was still more important, Car-O was also
present, varying the proportion between them depending on the catalyst employed: with Ni/C
as catalysts Car-O rupture was not detected, but with Pd/C Car-O rupture exceeded that of Cali-O
bond. Thus it was concluded that, in addition to binding energy of the C-O bond, the specific
characteristics of the catalyst also affected the preferential rupture of one of the bonds. Further
studies will be necessary to establish the specific characteristics of the catalyst surface affecting
product distribution.
When using real lignin, it was observed that the presence of hydrogen, and foremost, hydrogen
plus a catalyst is crucial to improve the depolymerization of the lignin. Very promising results of
over 50% of depolymerization degree of the -O-4 linkages were obtained when using a Ru/C
catalyst. Further studies with this real residue must be carried out to analyze the effect of
different parameters, such as temperature, pressure, type of metal and catalyst support, and
metal loading.
Acknowledgements
The authors would like to thank the European Union (Project Valor Plus funded under the EU’s
Seventh Framework Programme, project FP7-KBBE-2013-7-613802). Authors also acknowledge
Prof. J. Arauzo and Prof. L. M. Gandía for their valuable discussions. The authors would like to
acknowledge Southwest Nanotechnologies for the MWCNT samples supplied for the catalyst
testing.
21

References
[1]
J.E. Holladay, J.J. Bozell, J.F. White, D. Johnson, Top value added chemicals from biomass volume
I — Results of screening for potential candidates from sugars and synthesis gas, 2004.
[2]
[3]
J. C. Serrano-Ruiz, J. A. Dumesic, Energy Environ. Sci. 4 (2011) 83–99.
A. Bioenergy, 2014. Project Hugoton
http://www.abengoabioenergy.com/web/en/2g_hugoton_project/general_information/.
[4]
B. Magazine, 2014. Commercial-scale cellulosic ethanol refinery opens in Italy
http://biomassmagazine.com/articles/9540/commercial-scale-cellulosic-ethanol-refinery-opens-
in-italy/.
[5]
[6]
POET-DSM, 2014. Project Liberty http://poetdsm.com/liberty.
DuPont, 2014. DuPont Nevada Site Cellulosic Ethanol Facility
http://biofuels.dupont.com/cellulosic-ethanol/nevada-site-ce-facility/.
[7]
[8]
[9]
W. Kricka, J. Fitzpatrick, U. Bond, Front Microbiol. 5 (2014) 174.
D. Yuan, K. Rao, P. Relue, S. Varanasi, Bioresour. Technol. 102 (2011) 3246-3253.
D. E. Resasco, S. Sitthisa, J. Faria, T. Prasomsri and M. P. Ruiz, in Heterogeneous Catalysis in
Biomass to Chemicals and Fuels, eds. D. Kubicka and I. Kubicková, Research Signpost, Kerala,
2011.
[10] A. Rico Campos, PhD Thesis. University of Seville (Spain), 2014.
[11] F.F. Furlan, R.T. Filho, F.H.P.B. Pinto, C.B.B. Costa, A.J.G. Cruz, R.L.C. Giordano, R.C. Giordano,
Biotech. Biofuels 6 (2013) 142.
[12] E. Hyvärinen, Unasylva 56 (2005) 39–41.
[13] W. Deng, H. Zhang, X. Wu, R. Li, Q. Zhang and Y. Wang, Green Chem. 17 (2015) 5009–5018.
[14] A. Toledano, L. Serrano and J. Labidi, Fuel 116 (2014) 617–624.
[15] X. Erdocia, R. Prado, M. A. Corcuera and J. Labidi, Front. Energy Res. 2 (2014) 1–7.
[16] C. Li, X. Zhao, A. Wang, G. W. Huber and T. Zhang, Chem. Rev. 115 (2015) 11559–11624.
[17] J. Zakzeski, P. C. A. Bruijnincx, A. L. Jongerius and B. M. Weckhuysen, Chem. Rev.110 (2010)
3552–3599.
[18] P. Azadi, O. R. Inderwildi, R. Farnood and D. A. King, Renew. Sustain. Energy Rev. 21 (2013) 506–
523.
[19] M. D. Kärkäs, B. S. Matsuura, T. M. Monos, G. Magallanes and C. R. J. Stephenson, Org. Biomol.
Chem. 14 (2016) 1853–1914.
[20] C. Xu, R. A. D. Arancon, J. Labidi and R. Luque, Chem. Soc. Rev. 43 (2014) 7485–7500.
[21] J. Kong, B. Li and C. Zhao, RSC Adv. 6 (2016) 71940–71951.
[22] D. D. Laskar, M. P. Tucker, X. Chen, G. L. Helms and B. Yang, Green Chem. 16 (2014) 897–910.
[23] S. Kasakov, H. Shi, D. M. Camaioni, C. Zhao, E. Baráth, A. Jentys and J. a. Lercher, Green Chem. 17
(2015) 5079–5090.
22

[24] Q. Song, F. Wang, J. Cai, Y. Wang, J. Zhang, W. Yu and J. Xu, Energy Environ. Sci. 6 (2013) 994–
1007.
[25] K. Barta, T. D. Matson, M. L. Fettig, S. L. Scott, A. V. Iretskii and P. C. Ford, Green Chem. 12 (2010)
1640–1647.
[26] J. A. Onwudili and P. T. Williams, Green Chem. 16 (2014) 4740–4748.
[27] A. Toledano, L. Serrano, A. Pineda, A. a. Romero, R. Luque and J. Labidi, Appl. Catal. B Environ.
145 (2014) 43–55.
[28] H. Wang, L. Zhang, T. Deng, H. Ruan, X. Hou, J. R. Cort and B. Yang, Green Chem. 18 (2016) 2802–
2810.
[29] X. Huang, T. I. Korányi, M. D. Boot and E. J. M. Hensen, Green Chem. 17 (2015) 4941–4950.
[30] Z. Luo, Y. Wang, M. He and C. Zhao, Green Chem. 18 (2016) 433–441.
[31] S. Nanayakkara, A. F. Patti and K. Saito, Green Chem. 16 (2014) 1897–1903.
[32] J. He, C. Zhao and J. A. Lercher, J. Am. Chem. Soc. 134 (2012) 20768–20775.
[33] G. Zhu, X. Ouyang, Y. Yang, T. Ruan and X. Qiu, RSC Adv. 6 (2016) 17880–17887.
[34] A. J. Ragauskas, G. T. Beckham, M. J. Biddy, R. Chandra, F. Chen, M. F. Davis, B. H. Davison, R. a
Dixon, P. Gilna, M. Keller, P. Langan, A. K. Naskar, J. N. Saddler, T. J. Tschaplinski, G. a Tuskan and
C. E. Wyman, Science 344 (2014) 1246843 (1–10).
[35] H. Wang, M. Tucker and Y. Ji, J. Appl. Chem. 2013 (2013) 1–9.
[36] R. Shu, J. Long, Z. Yuan, Q. Zhang, T. Wang, C. Wang and L. Ma, Bioresour. Technol. 179 (2015)
84–90.
[37] R. Shu, J. Long, Y. Xu, L. Ma, Q. Zhang, T. Wang, C. Wang, Z. Yuan and Q. Wu, Bioresour. Technol.
200 (2016) 14–22.
[38] F. Liu, Q. Liu, A. Wang and T. Zhang, ACS Sustain. Chem. Eng. (2016) acssuschemeng.6b00620.
[39] J. Long, Y. Xu, T. Wang, Z. Yuan, R. Shu, Q. Zhang and L. Ma, Appl. Energy 141 (2015) 70–79.
[40] T. Klamrassamee, N. Laosiripojana, D. Cronin, L. Moghaddam, Z. Zhang and W. O. S. Doherty,
Bioresour. Technol. 180 (2015) 222–229.
[41] J. Zhang, H. Asakura, J. van Rijn, J. Yang, P. Duchesne, B. Zhang, X. Chen, P. Zhang, M. Saeys and
N. Yan, Green Chem. 16 (2014) 2432–2437.
[42] X. Erdocia, R. Prado, M. A. Corcuera and J. Labidi, Biomass and Bioenergy 66 (2014) 379–386.
[43] F. P. Bouxin, A. McVeigh, F. Tran, N. J. Westwood, M. C. Jarvis and S. D. Jackson, Green Chem. 17
(2015) 1235–1242.
[44] K. Barta, G. R. Warner, E. S. Beach and P. T. Anastas, Green Chem. 16 (2014) 191–196.
[45] S. Boullosa-Eiras, R. Lødeng, H. Bergem, M. Stöcker, L. Hannevold and E. A. Blekkan, Catal. Today
223 (2014) 44–53.
[46] C. Zhao, J. He, A. A. Lemonidou, X. Li and J. A. Lercher, J. Catal. 280 (2011) 8–16.
[47] Y.-B. Huang, L. Yan, M.-Y. Chen, Q.-X. Guo and Y. Fu, Green Chem. 17 (2015) 3010–3017.
[48] J. He, C. Zhao and J. a. Lercher, J. Catal. 309 (2014) 362–375.
[49] R. C. Runnebaum, T. Nimmanwudipong, D. E. Block and B. C. Gates, Catal. Sci. Technol. 2 (2012)
23

113.
[50] M. Saidi, F. Samimi, D. Karimipourfard, T. Nimmanwudipong, B. C. Gates and M. R. Rahimpour,
Energy Environ. Sci. 7 (2014) 103–129.
[51] Q. Bu, H. Lei, A. H. Zacher, L. Wang, S. Ren, J. Liang, Y. Wei, Y. Liu, J. Tang, Q. Zhang and R. Ruan,
Bioresour. Technol. 124 (2012) 470–477.
[52] A. G. Sergeev and J. F. Hartwig, Science 332 (2011) 439–43.
[53] S. Jia, B. J. Cox, X. Guo, Z. C. Zhang and J. G. Ekerdt, Holzforschung 64 (2010) 577–580.
[54] A. G. Sergeev, J. D. Webb and J. F. Hartwig, J. Am. Chem. Soc. 134 (2012) 20226–20229.
[55] A. Wu, B. O. Patrick, E. Chung and B. R. James, Dalt. Trans. 41 (2012) 11093–11106.
[56] V. Roberts, S. Fendt, A. A. Lemonidou, X. Li and J. A. Lercher, Appl. Catal. B Environ. 95 (2010) 71–
77.
[57] J. He, C. Zhao, D. Mei and J. a. Lercher, J. Catal. 309 (2014) 280–290.
[58] C. M. Bernt, G. Bottari, J. A. Barrett, S. L. Scott, K. Barta and P. C. Ford, Catal. Sci. Technol. 6
(2016) 2984–2994.
[59] D.J. Rensel, S. Rouvimov, M.E. Gin and, J.C. Hicks, J. Catal. 305 (2013) 256–263.
[60] D.Y. Min, H. Jameel, H.M. Chang, L. Lucia, Z.G. Wang, Y.C. Jin, RSC Adv. 4 (2014) 10845–10850.
[61] J.L. Wen, S.L. Sun, B.L. Xue, R.C. Sun, Materials (Basel) 6 (2013) 359–391.
[62] T.Q. Yuan, F. Xu, R.C. Sun, J. Chem. Technol. Biotechnol. 88 (2013) 346–352.
[63] Y. Yoon, R. Rousseau, R. S. Weber, D. Mei and J. A. Lercher, J. Am. Chem. Soc. 136 (2014) 10287–
10298.
[64] C. Zhao, Y. Kou, A. A. Lemonidou, X. Li, J. A. Lercher, Angew. Chem. Int. Ed. 48 (2009) 3987–3990.
[65] B. Thomas, S. Prathapan and S. Sugunan, Microporous Mesoporous Mater. 80 (2005) 65–72.
[66] P. Zapata, J. Faria, M. P. Ruiz, D. E. Resasco, Topic Catal. 55 (2012) 38-52.
[67] J. Zhang, J. Teo, X. Chen and H. Asakura, ACS Catal. 4 (2014) 1574–1583.
[68] V. Mazzieri, F. Coloma-Pascual, A. Arcoya, P.C. L’Argentiꢀre, N.S. Fígoli, Appl. Surf. Sci. 210 (2003)
222–230.
[69] S. Gaur, D. Pakhare, H. Wu, D.J. Haynes, J.J. Spivey, Energy Fuel 26 (2012) 1989–1998.
[70] W. Deng, X. Tan, W. Fang, Q. Zhang, Y. Wang, Catal. Letters 133 (2009) 167–174.
[71] A.K. Hill, L. Torrente-Murciano, Int. J. Hydrogen Energy 39 (2014) 7646–7654.
[72] A.K. Hill, L. Torrente-Murciano, Appl. Catal. B Environ. 172-173 (2015) 129–135.
[73] C. Liang, Z. Wei, M. Luo, P. Ying, Q. Xin, C. Li, Hydrogen spillover effect in the reduction of barium
nitrate of Ru- Ba(NO3)2/AC catalysts for ammonia synthesis. In Studies in Surface Science and
Catalysis; Guerrero-Ruiz, A.; Rodriguez-Ramons, I., Eds.; Elsevier Masson SAS, 2001; Vol. 138, pp.
283– 290.
[74] L. Hu, S. Yin, Q. Yin, H. Wang, S. Wang, Energy Sci. Eng. 3 (2015) 126–134.
[75] J. Xiong, X. Dong, L. Li., J. Nat. Gas Chem. 21 (2012) 445–451.
[76] M.H. Youn, J.G. Seo, K.M. Cho, J.C. Jung, H. Kim, K.W. La, D.R. Park, S. Park, S.H. Lee, I.K. Song,
Korean J. Chem. Eng. 25 (2008) 236–238.
24

[77] G. García, A. Monzon, F. Bimbela, J.L. Sanchez, J. Abrego, Energy Fuel 27 (2013) 2846–2856.
[78] J. Remꢁn, J. Medrano, F. Bimbela, L. García, J. Arauzo, J. Appl. Catal. B Environ. 132-133 (2013)
433–444.
[79] M. Zhou, L. Tian, L. Niu, C. Li, G. Xiao, R.U. Xiao, Fuel Process. Technol. 126 (2014) 12–18.
[80] S. Wang, Q. Yin, J. Guo, L. Zhu, Energy Fuel 27 (2013) 3961–3968.
[81] Folkesson, B. ESCA studies on the charge distribution in some dinitrogen complexes of rhenium,
iridium, ruthenium, and osmium. Acta Chem. Scand. 27 (1973) 287–302. ꢀ
[82] Shen, J. Y.; Adnot, A.; Kaliaguine, S. An ESCA study of the interaction of oxygen with the surface of
ruthenium. Appl. Surf. Sci. 51 (1991) 47–60.
[83] Bimbela, F.; Chen, D.; Ruiz, J.; García, L.; Arauzo, J. Ni/Al coprecipitated catalysts modified with
magnesium and copper for the catalytic steam reforming of model compounds from biomass
pyrolysis liquids. Appl. Catal. B Environ. 119-120 (2012) 1–12.
[84] Beccaria, A. M.; Poggi, G.; Castello, G. Influence of passive film composition and sea water
pressure on resistance to localised corrosion of some stainles steels in sea water. Br. Corros. J. 30
(1995) 283–287.
[85] Wagner, C. D.; Riggs, W. M.; Davis, L. E.; Moulder, J. F. Handbook of X-ray Photoelectron
Spectroscopy; Muilenberg, G. E., Ed.; 1st ed.; Perkin-Elmer Corporation (Physical Electronics),
1979.
[86] Upare, D. P.; Yoon, S.; Lee, C. W. Selective ring opening of methylcyclohexane over Ni / SiO 2. J
Porous Mater 20 (2013) 1129–1136.
[87] Ahmed, K.; Mistry, P. Effect of calcination temperature on nickel/alumina catalysts. Collect.
Czech. Chem. Commun 57 (1992) 2073–2077.
[88] Salagre, P.; Fierro, J. L. G.; Medina, F.; Sueiras, J. E. Characterization of nickel species on several y-
alumina supported nickel samples. J. Mol. Catal. A Chem. 106 (1996) 125–134.
[89] Shalvoy, R. B.; Reucroft, P. J.; Davis, B. H. Characterization of coprecipitated nickel on silica
methanation catalysts by X-ray phtoelectron spectroscopy. J. Catal. 56 (1979) 336–348.
[90] Hüfner, S.; Wertheim, G. K.; Wernick, J. H. XPS core line asymmetries in metals. Solid State
Commun. 17 (1975) 417–422.
[91] Sarapatka, T. J. XPS-XAES study of charge transfers at Ni/Al2O3 systems. Chem. Phys. Lett. 212
(1993) 37–42.
[92] Hinokuma, S.; Murakami, K.; Uemura, K.; Matsuda, M.; Ikeue, K.; Tsukahara, N.; MacHida, M. Arc
plasma processing of Pt and Pd catalysts supported on γ-Al2O3 powders. Top. Catal. 52 (2009)
2108–2111.
[93] Barr, T. L. An ESCA study of the termination of the passivation of elemental metals. J. Phys. Chem.
82 (1978) 1801–1810.
[94] Guimaraes, A. L.; Dieguez, L. C.; Schmal, M. The effect of precursors salts on surface state of
Pd/Al2O3 and Pd/CeO2/Al2O3 catalysts. An Acad Bras Cienc 76 (2004) 825–832.
[95] Kumar, G.; Blackburn, J. R.; Jones, M. M.; Albridge, R. G.; Moddeman, W. E. Photoelectron
spectroscopy of coordination compounds. II. Palladium complexes. Inorg Chem 11 (1972) 296–
300.
[96] Z. Luo and C. Zhao, Catal. Sci. Technol. 6 (2016) 3476–3484.
25

[97] M. A. Filler and S. F. Bent, Prog. Surf. Sci. 73 (2003) 1–56.
[98] H. Search, C. Journals, A. Contact and M. Iopscience, J. physics. Condens. Matter 15 (2003)
R1501–R1516.
[99] F. Zaera, Chem. Rev. (1995) 2651–2693.
[100] S. J. Jenkins, Proc. R. Soc. A (2009) 2949–2976.
[101] H. Yildirim, T. Greber and A. Kara, J. Phys. Chem. C (2013) 20572–20583.
[102] C. Chiu, A. Genest, A. Borgna and N. Ro, ACS Catal. (2014) 4178–4188.
[103] S. Yamagishi, S. J. Jenkins and D. A. King, J. Chem. Phys. 114 (2001) 5765–5773.
[104] F. Mittendorfer and J. Hafner, Surf. Sci. 472 (2001) 133–153.
[105] A. K. Myers, G. R. Schoofs and J. B. Benziger, J. Phys. Chem. (1987) 2230–2232.
[106] G. Held, W. Braun, H. Steinrück, S. Yamagishi, S. J. Jenkins and D. A. King, Phys. Rev. Lett. 87
(2001) 216102-1-216102–4.
[107] C. Morin, D. Simon and P. Sautet, J. Phys. Chem. 108 (2004) 5653–5665.
[108] H. Orita and N. Itoh, Appl. Catal. A Gen. 258 (2004) 17–23.
[109] C. D. Zeinalipour-Yazdi, A. L. Cooksy and A. M. Efstathiou, Surf. Sci. (2008) 1858–1862.
[110] O. Dulaurent, M. Nawdali, A. Bourane and D. Bianchi, Appl. Catal. A Gen. 201 (2000) 271–279.
[111] S. Derrouiche and D. Bianchi, Appl. Catal. A Gen. 313 (2006) 208–217.
[112] J. Szanyi, W. K. Kuhn, D. W. Goodman, J. Szanyi, W. K. Kuhn and D. W. Goodman, J. Vac. Sci.
Technol. A 11 (1993) 1969–1974.
[113] S. Sitthisa and D. E. Resasco, Catal. Letters 141 (2011) 784–791.
[114] M. K. Bradley, J. Robinson and D. P. Woodruff, Surf. Sci. 604 (2010) 920–925.
[115] C. Zhang, J. Lu, X. Zhang, K. MacArthur, M. Heggen, H. Li, F. Wang, Green Chem. (2016) DOI:
10.1039/C6GC01456A.
[116] S. Dabral, J. Mottweiler, T. Rinesch, C. Bolm, Green Chem. 17 (2015) 4908-4912.
[117] Z. Strassberger, A.H. Alberts, M.J. Louwerse, S. Tanase, G. Rothenberg, Green Chem. 15 (2013)
768-774.
[118] A.T. Martínez, J. Rencoret, G. Marques, A. Gutiꢂrrez, D. Ibarra, J. Jimꢂnez- Barbero, J.C. del Río,
Phytochemistry 69 (2008) 2831–2843.
[119] J.L. Wen, Z. Sun, Y.C. Sun, S.N. Sun, F. Xu, R.C. Sun, Biobased Mater. Bio. 4 (2010) 408–425.
[120] M. Sette, R. Wechselberger, C. Crestini, Chem. Eur. J. 17 (2011) 9529–9535.
[121] W. Xu, S.J. Miller, P.K. Agrawal, C.W. Jones, ChemSusChem 5 (2012) 667–75.
[122] T.H. Parsell, B.C. Owen, I. Klein, T.M. Jarrell, C.L. Marcum, L.J. Haupert, L.M. Amundson, H.I.
Kenttämaa, F. Ribeiro, J.T. Miller, M.M. Abu-Omar, Chem. Sci. 4 (2013) 806–813.
26

Figure 1. β-O-4, α-O-4, and 4-O-5 ether bonds in lignin structure.
Figure 2. Model compounds for ether bonds in lignin: BPE (휶-O-4), PPE (휷-O-4), and DPE (4-O-5).
4
11
12
BPE
3
1
2
14
13
6
7
9
16
15
5
8
10
(12,13)
(11,14-17)
17
Figure 3. Proposed reaction network for hydroprocessing of BPE in methanol.
27

Figure 4. Yields to families of compounds for the catalysts used in the initial screening. Reaction
conditions: 150 °C, 2 h of reaction, 25 bar-g of H2 and 16 mM of BPE. FSD: fully saturated dimers; PSD:
partially saturated dimers, SM: saturated monomers; AM: aromatic monomers
a)
b)
c)
Figure 5. Conversion and yield to families of compounds after reaction of a) BPE, b) PPE and c) DPE.
Conditions: 150 °C, 2 h of reaction, 25 bar of H2, 16 mM of model compound and RCF = 4.55·g.
28