Ni/HZSM-5 catalyst preparation by deposition-precipitation. Part 2. Catalytic hydrodeoxygenation reactions of lignin model compounds in organic and aqueous systems

Article abstract of DOI:10.1016/j.apcata.2018.06.012

Nickel metal supported on HZSM-5 (zeolite) is a promising catalyst for lignin depolymerization. In this work, the ability of catalysts prepared via deposition-precipitation (DP) to perform hydrodeoxygenation (HDO) on two lignin model compounds in organic and aqueous solvents was evaluated; guaiacol in dodecane and 2-phenoxy-1-phenylethanol (PPE) in aqueous solutions. All Ni/HZSM-5 catalysts were capable of guaiacol HDO into cyclohexane at 523 K. The role of the HZSM-5 acid sites was confirmed by comparison with Ni/SiO₂ (inert support) which exhibited incomplete deoxygenation of guaiacol due to the inability to perform the cyclohexanol dehydration step. The catalyst prepared with 15 wt% Ni, a DP time of 16 h, and a calcination temperature of 673 K (Ni(15)/HZSM-5 DP16_Cal673), performed the guaiacol conversion with the greatest selectivity towards HDO products, with an intrinsic rate ratio (HDO rate to conversion rate) of 0.31, and 90% selectivity to cyclohexane. Catalytic activity and selectivity of Ni/HZSM-5 (15 wt%) in aqueous environments (water and 0.1 M NaOH solution) was confirmed using PPE reactions at 523 K. After 30 min reaction time in water, Ni/HZSM-5 exhibited ～100% conversion of PPE, and good yield of the desired products; ethylbenzene and phenol (～35% and 23% of initial carbon, respectively). Ni/HZSM-5 in NaOH solution resulted in significantly higher ring saturation compared to the Ni/HZSM-5 in water or the NaOH solution control.

Full text of DOI:10.1016/j.apcata.2018.06.012

Accepted Manuscript
Title: Ni/HZSM-5 catalyst preparation by
deposition-precipitation. Part 2. Catalytic hydrodeoxygenation
reactions of lignin model compounds in organic and aqueous
systems
Authors: R.R. Barton, M. Carrier, C. Segura, J.L. Garcia
Fierro, S. Park, H.H. Lamb, N. Escalona, S.W. Peretti
PII:
S0926-860X(18)30288-6
DOI:
Reference:
https://doi.org/10.1016/j.apcata.2018.06.012
APCATA 16700
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
5-2-2018
5-6-2018
7-6-2018
Please cite this article as: Barton RR, Carrier M, Segura C, Garcia Fierro JL,
Park S, Lamb HH, Escalona N, Peretti SW, Ni/HZSM-5 catalyst preparation by
deposition-precipitation. Part2. Catalytichydrodeoxygenationreactionsofligninmodel
compounds in organic and aqueous systems, Applied Catalysis A, General (2018),
https://doi.org/10.1016/j.apcata.2018.06.012
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.

Ni/HZSM-5 Catalyst Preparation by Deposition-Precipitation. Part 2.
Catalytic Hydrodeoxygenation Reactions of Lignin Model Compounds in Organic and
Aqueous Systems
R. R. Bartona, M. Carrierb,1, C. Segurac, J. L. Garcia Fierrod, S. Parke, H. H. Lamba, N.
Escalonaf,g,h, S.W. Perettia
aNorth Carolina State University, Department of Chemical and Biomolecular Engineering, Raleigh, NC
bUniversidad de Concepción, Unidad de Desarrollo Tecnológico, Coronel, Chile.
1European Bioenergy Research Institute, Aston University, Birmingham B4 7ET, UK.
c Unidad de Desarrollo Tecnológico, Universidad de Concepción, Coronel, Chile.
dInstituto de Catálisis y Petroquímica, CSIC, Cantoblanco, 28049 Madrid, Spain.
eNorth Carolina State University, Department of Forest Biomaterials, Raleigh, NC
fPontificia Universidad Católica de Chile, Departamento de Ingeniería Química y Bioprocesos, Santiago, Chile.
gPontificia Universidad Católica de Chile, Facultad de Químicas, Santiago, Chile.
hCentro de Investigación en Nanotecnología y Materiales Avanzados (CIEN-UC), Pontificia Universidad Católica
de Chile, Santiago, Chile.
Ryan R. Barton:
rrbarton@ncsu.edu
Marion Carrier:
m.carrier@aston.ac.uk
Cristina Segura:
c.segura@udt.cl
Jose Luis Garcia Fierro:
jlgfierro@icp.csic.es
Sunkyu Park:
sunkyu_park@ncsu.edu
H. Henry Lamb:
lamb@ncsu.edu
Néstor Escalona:
neescalona@ing.puc.cl
Corresponding Author:
Steven W. Peretti
Department of Chemical and Biomolecular Engineering, NC State University
Campus Box 7905
Raleigh, NC 27695-7905
peretti@ncsu.edu
1

Graphical abstract
Highlights
 Effect of catalyst properties of Ni/HZSM-5 synthesized via excess deposition-
precipitation times on guaiacol and 2-phenoxy-1-phenylethanol hydrodeoxygenation was
studied.
 Ni(15)/HZSM-5 with 16 h DP time and calcined at 673 K had high intrinsic rates and
high selectivity.
 Selectivity towards aromatic products was highest in neutral aqueous environments.

Ni/HZSM-5 from deposition-precipitation is a promising catalyst for lignin cleavage at
low temperature.
Abstract
Nickel metal supported on HZSM-5 (zeolite) is a promising catalyst for lignin
depolymerization. In this work, the ability of catalysts prepared via deposition-precipitation (DP)
to perform hydrodeoxygenation (HDO) on two lignin model compounds in organic and aqueous
solvents was evaluated; guaiacol in dodecane and 2-phenoxy-1-phenylethanol (PPE) in aqueous
solutions. All Ni/HZSM-5 catalysts were capable of guaiacol HDO into cyclohexane at 523 K.
The role of the HZSM-5 acid sites was confirmed by comparison with Ni/SiO2 (inert support)
which exhibited incomplete deoxygenation of guaiacol due to the inability to perform the
cyclohexanol dehydration step. The catalyst prepared with 15 wt% Ni, a DP time of 16 h, and a
calcination temperature of 673 K (Ni(15)/HZSM-5 DP16_Cal673), performed the guaiacol
conversion with the greatest selectivity towards HDO products, with an intrinsic rate ratio (HDO
rate to conversion rate) of 0.31, and 90% selectivity to cyclohexane. Catalytic activity and
selectivity of Ni/HZSM-5 (15 wt%) in aqueous environments (water and 0.1 M NaOH solution)
was confirmed using PPE reactions at 523 K. After 30 min reaction time in water, Ni/HZSM-5
exhibited ~100% conversion of PPE, and good yield of the desired products; ethylbenzene and
phenol (~35% and 23% of initial carbon, respectively). Ni/HZSM-5 in NaOH solution resulted in
2

significantly higher ring saturation compared to the Ni/HZSM-5 in water or the NaOH solution
control.
Keywords: hydrodeoxygenation; lignin; depolymerization; HZSM-5; deposition-precipitation
3

1. Introduction
Pulp and paper plants, and lignocellulosic biorefineries, generate lignin residues as a
byproduct stream. Rather than burn it to power industrial processes [1-7], if successfully
depolymerized, it could be used to produce aromatic compounds, platform chemicals, and other
high value, petroleum-based products [2-5]. The strategy taken in this work is the
depolymerization of lignin into simple aromatic monomers through the use of a bifunctional
catalyst, one capable of hydrogenolysis of both C-O and C-C linkages in lignin, under mild
reaction conditions in an aqueous environment. Two model compounds were chosen to probe
catalyst hydrodeoxygenation (HDO activity); guaiacol and 2-phenoxy-1-phenylethanol (PPE).
Guaiacol exhibits an aromatic ring with C-O bonds in the form of methoxy and hydroxy moieties
which are more accessible than the β-O-4 linkage present in PPE. Successful
hydrodeoxygenation of these compounds would establish this catalyst as a promising candidate
for depolymerization of the more complex lignin molecule.
To evaluate the catalyst performance, a preliminary high-throughput screening based on
guaiacol HDO assays is used. The C-O bonds in the form of methoxy and hydroxy moieties all
can undergo hydrogenolysis; cleavage via hydrodeoxygenation is an effective form of
hydrogenolysis for the depolymerization of lignin. In this work, a series of Ni/HZSM-5 catalysts
which were synthesized via deposition-precipitation (DP) and characterized previously [8] are
evaluated vis-a-vis the hydrodeoxygenation (HDO) of guaiacol, a lignin model compound. In an
additional study, the developed catalyst was evaluated with a lignin model compound in various
aqueous media which serve as environmentally friendly alternatives to organic solvents.
This strategy has been proven to be successful as reported in two recent studies using
Ni/HZSM-5, prepared via incipient wetness impregnation (IWI), in an aqueous environment for
the HDO of phenol (5 MPa H2, 473 K, 3 h reaction time, using 7-9 wt% nickel), and for the
upgrading of pyrolysis oil and various aromatic monomers and dimers (5 MPa H2, 523 K, 2 h for
phenol and 4 h for pyrolysis oil, using 20 wt% nickel) [9, 10]. For all the reactions performed in
the aqueous environment, a high extent of HDO was observed, with mainly C-O cleavage and
some aromatic ring saturation. Reactions involving phenol over 3 h resulted in 80-100%
conversion, with 90-100% selectivity to cyclohexane. Other aromatic monomers had 90-100%
conversion, with approximately 90% selectivity to hydrocarbons and about 10% to methanol. Of
the 90%, 75-90% were cycloalkanes and 5-15% were aromatics. For the dimer reactions, 100%
conversion was achieved, with only 18-37% of aromatics yielded, and for the pyrolysis oil,
100% deoxygenation was achieved, with 10% paraffins, and 90% cycloalkanes and aromatics.
As shown from the studies of nickel catalyzed HDO reactions in aqueous environments,
supported nickel catalysts, such as Ni/HZSM-5, exhibit promise for the selective cleavage of C-
O linkages that represent the dominant bonds in lignin [2]. However, the studies mentioned
above only used nickel catalysts prepared with IWI techniques and not DP, which is the focus of
this work.
Song et al. [11] compared 5wt% Ni/Hβ catalysts prepared by 4 different techniques,
including IWI and DP. They found that the Ni/Hβ prepared by the DP method not only had the
smallest particle size and tightest particle size distribution (highest dispersion nickel catalyst), it
also had the fastest rate of stearic acid HDO and displayed high stability after multiple reaction
runs [11]. He et al. [12, 13] produced Ni/SiO2 using the DP method for the cleavage of β-O-4, α-
4

O-4, and 4-O-5 aromatic ether dimers in an aqueous phase (393 K, 0.6 MPa H2, 40-57 wt%
nickel, and 90 min reaction time). Both studies observed selective cleavage of the ether bonds,
along with some aromatic ring saturation. These studies were only looking at Ni/SiO2, and not
the multifunctional Ni/HZSM-5 system [12, 13].
Song et al. [14] prepared Ni/HZSM-5 catalyst by the DP method for the HDO of several
substituted phenols (473 K, 3 MPa H2, 20 wt% nickel, and 0-2 h reaction time) in an aqueous
environment and proposed several reaction pathways. That catalyst system successfully
performed HDO, and, more specifically, C-O hydrogenolysis. The role of the acidic support in
the adsorption of the reactants for hydrogenolysis instead of aromatic ring hydrogenation was
also reported [14]. Each study used the DP procedure that was developed and characterized by
Burattin et al. [15, 16] to prepare a Ni/SiO2 catalyst. In this laboratory [8], a different DP
procedure was utilized. Nickel loading was manipulated by varying the concentration of nickel
nitrate, rather than the DP time. The effects of catalyst preparation conditions and resulting
catalyst properties on HDO activity were then characterized.
Catalyst activity and selectivity were evaluated using a guaiacol HDO reaction in n-
dodecane solvent, a commonly used reaction assay [17-20]. Guaiacol serves as a simple model
compound for depolymerized lignin due to the presence of the aromatic ring and the substituent
methoxy and hydroxy groups, which are common in the lignin structure. This model reaction
system was used for several reasons: a) to avoid complicated effects of a multiphase liquid
system, b) to avoid effects of the aqueous phase on the support and on the activity of the nickel
catalyst [21, 22], and c) to simplify the product analysis by limiting the number of chemical
species present in the reaction system. Once kinetic and catalytic, properties were assessed, it
was necessary to establish the role of the HZSM-5 support and the global catalytic activity
Ni/HZSM-5 in diverse aqueous reaction environments (e.g. neutral and alkaline aqueous
solutions). Aqueous environments have been observed to preserve aromaticity (at the cost of
slower reaction rates) with similar kinds of lignin model compounds [21, 23]. Water will not
solubilize lignin, but it prevents significant ring saturation. Using a NaOH solution as a solvent
has also been shown to have positive effects on lignin depolymerization. It solubilizes lignin,
allowing the lignin to be more dispersed in the reaction mixture, reducing char formation, and it
hydrolyzes lignin through base catalysis [24, 25]. For these reasons, 2-phenoxy-1-phenylethanol
(PPE) was used as a lignin model compound in aqueous environments to further demonstrate the
activity of the Ni/HZSM-5 produced by this modified DP procedure; specifically, its ability to
cleave β-O-4 linkages found in lignin.
Effective catalysts exhibit high activity and high selectivity for HDO products. This
should coincide with favorable nickel particle characteristics, including high dispersion. The
catalyst with the greatest selectivity for HDO from this simple model system will be used against
other model lignin compounds and lignin in aqueous solvents in subsequent studies. Nickel
loading and catalyst preparation were varied to determine whether the optimal nickel loading
found in Barton et al. [8] correlates with the highest intrinsic rate and/or selectivity toward
complete deoxygenation. The optimal loading (15 wt% Ni) was defined as the highest loading
without significant reduction in dispersion and increase in particle size. In addition, degradation
mechanisms were examined and discussed.
5

2. Experimental
2.1 Catalyst Preparation and Characterization
The Ni/HZSM-5 was synthesized using a DP method similar to that of Burattin et al. [15,
16]. For each catalyst preparation, 5 g of catalyst was synthesized with varying nickel loading, 5,
10, 12.5, 15, and 20 wt% nickel. All catalysts are referred to by their nominal loading, based on
the initial mass of nickel introduced during synthesis. The resulting nickel levels, measured using
ICP-OES, are reported in Table 1A, and vary less than 20% from the nominal loading levels.
DP method parameters, including concentration of NH4-ZSM-5 support (Zeolyst
International, CBV 2314), urea, nitric acid, temperature, stirring, and DP time, were kept
constant. Unlike the standard method, which uses excess amounts of nickel nitrate and varies the
DP time to control the nickel loading, the nickel loading was controlled by using an initial
concentration of nickel nitrate corresponding to the total amount of nickel desired for each
catalyst, and then using a constant, excess DP time to incorporate the nickel. The following
concentrations were used to synthesize the catalyst samples: 7.9 g NH4-ZSM-5·L-1 water, 0.42
mol·L-1 urea, 0.02 mol·L-1 nitric acid, and 0.01-0.03 mol·L-1 Ni(NO3)2·6H2O. All concentrations
are based on the total amount of water, which depends on the required amount of support to
make 5 g of catalyst of a specific nickel loading. The catalysts discussed in this work were all
prepared using the methodology previously developed in this laboratory [8].
The DP method is capable of forming a combination of nickel hydroxide and 1:1 nickel
phyllosilicate species on the support [13-16, 26]. During calcination, the nickel hydroxide
decomposes into NiO, and the 1:1 nickel phyllosilicate forms 2:1 nickel phyllosilicate. He et al.
[12] related TOF by Ni/SiO2 to calcination temperature and particle size. The data suggests that
calcination temperature may contribute to a more complete transformation of a nickel-support
species during calcination, followed by a greater extent of reduction to nickel metal prior to the
reaction. For this reason, all catalyst precursors in this study were calcined either at 673 K or 873
K. After calcination, the sample was cooled and then stored for activation. Calcined catalyst
samples were then reduced in H2 gas environment at 733 K. Given that a batch reactor system
was used for these experiments, the Ni/HZSM-5 catalyst was exposed to air during loading of the
reaction vessel, causing re-oxidation of the nickel metal [27, 28]. To prevent extensive oxidation
and allow for a systematic and consistent oxygen exposure for all catalyst samples, catalysts
underwent passivation to form an oxidized surface layer on the nickel metal [8], which could
then be reduced in the reactor vessel in-situ under the reaction conditions. After passivation, the
catalysts were stored in a desiccator at room temperature for future reactions and catalyst
characterization All synthesized catalysts were characterized via various techniques as detailed
in Barton et al. [8]. The techniques applied include:
(1) Nitrogen adsorption-desorption isotherms to determine BET surface areas, total pore
volumes, micropore volumes (Dubinin-Radushkevich model), micropore size distributions
(carbon dioxide adsorption isotherms, DFT model), and mesopore size distributions (BJH
model);
(2) Inductively coupled plasma-optical emission spectrometer (ICP-OES) to determine Si/Al
ratios of the HZSM-5 support and nickel loading of the catalysts;
6

(3) Temperature programmed reduction (TPR) to determine H2/Ni uptake over increasing
temperature and confirm complete reduction of the nickel catalysts;
(4) X-ray diffraction (XRD) to determine the types of nickel species present in the nickel
catalysts and to estimate nickel particle diameter via the Scherrer equation (Eq. (1)), where
Dp is particle diameter (Å), λ is CuKα radiation wavelength (Å), β1/2 is the full width at half
maximum (radians), and θ (Bragg angle) is the position (1/2 value of 2θ position) (radians);
(5) Transmission electron microscopy (TEM) to determine mean nickel particle sizes and
distributions;
(6) H2-chemisorption to determine available nickel sites, dispersion (%) calculated by Eq. (2)
where D is the dispersion in %, H atoms is the amount of hydrogen atoms in mmol based on
the volume of H2 chemisorbed, and Ni is the amount of nickel in mmol based on the nickel
loadings determined by ICP-OES, and approximate nickel particle size using Eq. (3), where
dp is particle diameter in nm and D is the dispersion in %, which assumes that a spherical
particle that is less than 1 nm in diameter exhibits 100% dispersion [29];
(7) Acid strength of the HZSM-5 support with and without nickel was measured using a
potentiometric titration method [30]; and
(8) X-ray photoelectron spectroscopy (XPS) was used to observe surface species of the
supported nickel catalysts.
Important characteristic data relevant to this work have been included in Tables 1 – 2 [8].
0.94휆
퐷푝 =
(1)
훽
푐표푠휃
1
⁄
2
퐷 = 100 ∗ 퐻 푎푡표푚푠 (푚푚표푙)
(2)
(3)
푁푖 (푚푚표푙)
101
푑푝 =
퐷 (%)
While the catalyst preparation for the guaiacol HDO reactions is identical to the methods
used in Barton et al. [21], the catalyst applied to PPE model compound reactions had slightly
different preparation conditions, which include: 7.6 g·L-1 HZSM-5 (instead of 7.9 g·L-1), 0.02 M
Ni(NO3)2·6H2O (within the range used for previously) in 560 mL of water, and a DP time of 16
h was used. Due to use of an alternate lab with different equipment and capabilities, calcination
was performed at 673 K for 4 h with a 5 K·min-1 ramp. Approximately 0.5 g of calcined catalyst
was reduced in a quartz reactor tube connected to a flow system. The tube was first purged with
N2 at 63 mL·min-1 for 10 min. H2 gas was then flown into the reactor at 7 mL·min-1 for a total of
70 mL· min-1 (10% v/v H2). After flowing the mixture for 5 min, the reduction oven program was
initiated. The catalyst was reduced at 733 K for 4 h with a 2.5 K·min-1 ramp. After reduction, the
reactor tube was purged with N2 gas flow at 63 mL·min-1 and allowed to cool to room
temperature. Once the system cooled, the gas flows were shutoff, and the reactor tube fittings
were loosened to allow slow exposure to air overnight. After passivation, the catalyst was
removed from the apparatus and stored for reactions and catalyst characterization. The properties
of the Ni/HZSM-5 catalyst used in this work are listed in Table 3 and its TPR profile, XRD
7

patterns, and TEM/SEM images are presented in Figs. S1, S2, S3 respectively (this catalyst
corresponds to the Ni(15)/HZSM-5 - DP16_Cal673 catalyst in Tables 1B and 2B).
2.2 Reagents and Model Compound Preparation
The catalysts were evaluated with three different reaction environments. The first reaction
assay was a guaiacol HDO reaction using a batch reactor system with the non-polar solvent, n-
dodecane, and n-hexadecane as an internal standard. All these compounds were purchased (n-
dodecane (Merck Millipore, ≥99.0%), guaiacol (Merck Millipore, ≥98%), and n-hexadecane
(Sigma Aldrich, 99%).
The second and third reaction assays were a PPE β-O-4 cleavage reaction using a batch
reactor system with water and 0.1 M sodium hydroxide (Fisher Chemical, ≥97%) solution as the
solvent. The PPE was prepared via a synthesis method consisting of two main steps. The first
step was a combination reaction between 2-bromoacetophenone (Acros Organics, 98%) and
phenol (Aldrich, 99.99%) to form 2-phenoxyacetophenone. The second step reduced the 2-
phenoxyacetophenone to form PPE. The complete synthesis procedure is provided in Fig. S4
[31]. This procedure produced a white solid with an overall yield of 70% (1:1 stoichiometry
between moles of initial limiting reagent, 2-bromoacetophenone, and PPE). The identity was
confirmed by GC-MS (Fig. S5), and by comparison with a commercially purchased PPE using
GC-FID (Fig. S6). To also confirm the purity of the synthesized PPE, TGA-DSC (SDT Q600 TA
Instruments) was used to compare heat flow versus temperature data which confirmed that both
the purchased PPE and the synthesized PPE exhibited a phase change (melting point) at the same
temperature and were calculated to have similar purities >97% (based on TA Universal Analysis
software) (Fig S7).
2.3 Catalytic Performance
2.3.1 Guaiacol/Dodecane System
The reactions were performed in a 300 mL Parr reactor equipped with a gas entrainment
impeller and baffle, thermocouple, heating jacket, inlet gas line valve, and outlet liquid sampling
valve with a filter to prevent loss of solid catalyst. Each reaction used dodecane (80 mL),
guaiacol (0.28 mol·L-1), n-hexadecane (0.03 mol·L-1, internal standard), and Ni(x)/HZSM-5
catalyst (0.05 g, calcined, reduced, and passivated). The reaction was performed at 523 K under
5 MPa of H2 gas for 4 h. Samples were taken at the start of the reaction (time = 0 min is when H2
gas was introduced to the reactor after reaching the reaction temperature), 10 min, 20 min, 40
min, 1 h, 1.5 h, 2 h, 3 h, and 4 h. Along with reactions using Ni/HZSM-5, a reaction with
Ni/SiO2 was also performed to observe the influence of the HZSM-5 acid support versus the
inert SiO2 (Strem, 99+%) support.
Samples were quantified using GC-FID. Product identification was determined using a
combination of GC-MS results and reference samples in the GC-FID to confirm residence times.
A PerkinElmer Clarus 400 GC-FID was utilized, and equipped with a CP-SIL 5 Agilent column
(30 m x 0.53 mm x 1.0 μm). The GC-MS utilized was a PerkinElmer Clarus 680 equipped with
an Elite-1 PerkinElmer column (30 m x 0.32 mm x 0.25 μm). The carrier gas flow was 50
8

mL/min, with an injector temperature of 548 K and split ratio of 20:1. The GC oven program
started with a 6 min hold at 305 K, followed by heating to 343 K at 30 K/min and a hold for 22
min, then heating to 548 K at 30 K/min and a hold at for 3 min. After peak assignments using
GC-MS and reference samples, the n-hexadecane internal standard was used to quantify
confirmed HDO products. The FID detector temperature was set at 453 K with a hydrogen flow
rate of 45 mL/min and an air flow rate of 450 mL/min. The detector response is proportional to
the concentration of the compounds observed and therefore used directly to determine yields. In
addition, the internal standard peak and total peak areas were used to confirm consistency among
samples and track the fate of the starting guaiacol.
Results from GC-FID analysis of reaction time course samples were used to calculate
specific rates for guaiacol conversion and HDO rates (rate of producing completely
deoxygenated products; i.e., cyclohexane, cyclohexene, methylcyclopentane, and trace benzene).
Specific rates are normalized per gram of nickel, as determined from ICP-OES results, which are
displayed in Table 1A and B. Specific (per gram of nickel basis) and intrinsic (per nickel site
basis) initial rates are calculated from the slope of the linear region of guaiacol conversion (per
catalyst) over time (within the first 50min of reaction) and are synonymous with guaiacol
conversion rates. The turnover frequencies (TOF) for guaiacol conversion and HDO generation
were made intrinsic using the calculated specific rates and the number of active nickel sites, as
determined from H2-chemisorption dispersion results (which gives the total number of sites if
100% of nickel is reduced). As discussed in Barton et al. [8], the XPS results in Table 2A and B
show incomplete reduction of the nickel sites; however this was attributed to minor exposure of
the nickel metal to oxygen prior to the XPS analysis. Given the use of passivation to allow easy
in-situ re-reduction, the temperature of the reactions, and the high pressure of H2 gas utilized, it
is reasonable to assume that all the catalysts were fully reduced during the reactions. The
intrinsic rate (IR) of guaiacol conversion in the presence of Ni/HZSM-5 was calculated as
follows:
푆 ∗푁
푅
푎푣
퐼푅 =
(4)
푁푖
푆퐴
with the specific rate represented by SR; Avogadro’s number represented by Nav; and the number
of available nickel sites per gram of catalyst (reduced nickel sites) by NiSA. Calculation of NiSA is
as follows:
푁푖
푆퐶
푁푖푆퐴
=
(5)
푁푖
퐿
where NiSC represents the number of nickel sites per gram of catalyst, as determined by H2-
chemisorption, and NiL is the nickel loading as determined by ICP-OES.
2.3.2 PPE/Water System
The reactions were performed in a 50 mL Parr reactor equipped with an impeller, a
thermocouple, a heating jacket, inlet and outlet gas valves, and a safety rupture disc. Each
reaction used 20 mL of water and 0.535 g of PPE (0.125 M). HZSM-5 reactions used 0.0535 g of
9

catalyst (Mobil HZSM-5, 10:1 w/w feed to catalyst), while Ni/HZSM-5 reactions used 0.027 g of
Ni/HZSM-5 catalyst (Ni supported on Zeolyst HZSM-5, 20:1 w/w, feed to catalyst). Reactions
were performed at 523 K under 5 MPa of H2 gas for 0-2 h. The H2 gas was introduced prior to
the temperature ramp to 523 K, and the start time was the moment the reactor reached the
reaction temperature (autogenous pressure was about 8 MPa). The HZSM-5 reactions were
performed for 5 min, 15 min, 0.5 h, 1 h, and 2 h. The Ni/HZSM-5 reactions were performed for
0.25 h, 0.5 h, and 1 h, and a p-hydroxyacetophone internal standard was added to the reaction
mixture after the reaction, but prior to collection (approximately 8 mg).
For the HZSM-5 reactions, the product mixtures were collected with 20 mL of methanol
(Fisher Chemical, HPLC grade), and the products were further diluted using a 1:1 v/v ratio of
sample to methanol. The Ni/HZSM-5 reaction mixtures were collected with acetone (EMD
Millipore, HPLC grade) (approximately 20 mL, enough acetone added until reaction mixture was
a single phase with known total volume). For both reactions, the product mixtures were
centrifuged to remove catalyst, and then a 1.5 mL aliquot of the single phase reaction products
was taken directly, filtered through a 0.22 μm nylon syringe filter (FisherbrandTM) and then
injected into a GC for analysis. The reaction products at each time interval (15-30 min) were
quantified using GC-FID. Product identification was determined using a combination of GC-MS
results and standards in the GC-FID to confirm compound residence times. The GC-FID used
was a Shimadzu GC-2010, with a ZB-5MS Zebron column (30 m x 0.25 mm x 0.25 μm), which
is a G27 USP phase column (5% phenyl, 95% dimethylpolysiloxane). The GC-MS used was a
Shimadzu GC-2010 paired with a quadrupole mass spectrometer (GCMS-QP2010S), also
equipped with a ZB-5MS Zebron column (30 m x 0.25 mm x 0.25 μm). The analysis methods for
the GC-FID and GC-MS are provided in Figs. S8 and S9, respectively.
2.3.3 PPE/NaOH Solution System
The reactions were performed in a 50 mL Parr reactor equipped with an impeller, a
thermocouple, a heating jacket, inlet and outlet gas valves, and a safety rupture disc. Each
reaction used 20 mL of 0.1 M NaOH solution and 0.535 g of PPE (0.125 M). Ni/HZSM-5
reactions used 0.027 g of Ni/HZSM-5 catalyst (Ni supported on Zeolyst HZSM-5, 20:1 w/w,
feed to catalyst). Reactions were performed at 523 K under 5 MPa of H2 gas for 0-1 h. The H2
gas was introduced prior to the temperature ramp to 523 K, and the start time was the moment
the reactor reached the reaction temperature (autogenous pressure was about 8 MPa). The
reactions were performed for 0.25 h, 0.5 h, and 1 h, and a dodecane internal standard was added
to the reaction mixture after the reaction, but prior to collection (0.1 mL).
The alkaline reaction mixtures were collected with water (20 mL), acidified with 2 mL
hydrochloric acid (Fisher Chemical, Technical grade), centrifuged to remove catalyst, and
extracted with ethyl ether (Fisher Chemical, Laboratory grade) (20 mL, three times) to recover
the organics. A 1.5 mL aliquot from the ethyl ether phase (with a known total volume) was
filtered through a 0.22 μm nylon syringe filter (FisherbrandTM) and then injected into a GC for
analysis using the GC methods stated in section 2.3.2.
10

3. Results and Discussion
3.1 Guaiacol HDO Activity in Dodecane
3.1.1 Effect of Nickel Loading on Guaiacol HDO
The three main products from the conversion of guaiacol, regardless of nickel loading,
were 2-methoxycyclohexanol, cyclohexanol, and cyclohexane, as shown by the reaction progress
plots in Fig. 1. There were trace amounts of various HDO products (other completely
deoxygenated products) as well as partially hydrodeoxygenated compounds containing only
methoxy or hydroxy groups. Of all the catalyst loadings, only the 5 wt% nickel catalyst did not
reach 100% conversion of guaiacol after 4 h of reaction. As nickel loading increased from 10-20
wt% nickel, the time required to reach 100% conversion decreased. In addition, the amount and
rate (as indicated by the slopes of the yield plots) at which cyclohexane and other HDO products
(i.e., methylcyclopentane, cyclohexene, and benzene) are formed also increases with nickel
loading, and are accompanied by a decrease in the amount and a faster rate of disappearance of
2-methoxycyclohexanol, cyclohexanol, and other oxygenated compounds.
The specific rates for guaiacol conversion and HDO production both reach a maximum at
15 wt% nickel loading, as shown in Fig. 2a. The decrease in rate for the 20 wt% catalyst
suggests that less of the nickel is available for reaction sites, consistent with the hypothesis that
large aggregate nickel particles were formed at higher loading, resulting in lower surface area
(Table 1A) and fewer active sites. This is also in accordance with the H2-chemisorption results
(Table 1A) showing that the 20 wt% catalyst has lower dispersion than the catalysts with lower
nickel loading [8].
The intrinsic rates for all Ni(x)/HZSM-5 catalysts should be unchanging if the structure
and local environment of the active nickel sites are equivalent at all nickel loading levels. The
intrinsic rates, shown in Figs. 2b and 3, suggest that there are at least two different active site
configurations available for HDO activity, and possibly as many as three for guaiacol
conversion. In Fig. 3a, the intrinsic rate data for the guaiacol conversion indicate a dramatic
change in the activity between the 5-10 wt% region and the 12.5-15 wt% region. Above the 12.5-
15 wt% region, the activity of the nickel sites noticeably decreases. This suggests that the active
sites are similar for 12.5-15 wt% catalysts (for at least hydrogenation activity), with less active
sites on the 5, 10, and 20 wt% nickel catalysts.
In Fig. 3b, the low loading asymptote and the apparent high loading level asymptote
suggest, qualitatively, the presence of two different catalytic nickel sites. The results at 12.5 wt%
Ni could suggest the occurrence of a mixture of the two hypothesized sites. This suggests that the
5-10 wt% catalysts had similar HDO activity, and the 15-20 wt% catalysts had similar HDO
activity. The 15 wt% catalyst was in the higher activity region for both guaiacol conversion and
HDO intrinsic rates.
Previously hypothesized reaction networks surrounding hydrodeoxygenation with Ni-
HZSM-5 catalysts [14, 32, 33] were adapted to reflect the observed products (including trace
products) and are presented in Fig. 4. The main reaction pathways are presented in part a) and
alternate pathways in part b); both are comprised of nickel catalyzed and acid catalyzed reaction
steps.
11

The dominant pathway is highlighted by the dark red pathway in Fig. 4a.
Hydrogenation/saturation of the aromatic ring occurs first, converting guaiacol to 2-
methoxycyclohexanol. The 2-methoxycyclohexanol then undergoes demethoxylation to form
cyclohexanol. The cyclohexanol is then converted to cyclohexene via dehydration, followed by
hydrogenation to cyclohexane. This pathway is proposed due to the large amounts of 2-
methoxycyclohexanol, cyclohexanol, and cyclohexane observed in the reaction products, as well
as the reaction progress over time, which first shows formation of 2-methoxycyclohexanol and
cyclohexanol, then formation of cyclohexane. The other reactions in the network are not
included, as most of the other products are present only in trace amounts.
Fig. 5 shows the product distribution at 20% conversion. The 5 wt% nickel catalyst is not
shown because the reaction did not reach 20% conversion within the initial rate region of the
reaction progress. As the nickel loading increases, the product distribution shifts, exhibiting a
decrease in the formation of 2-methoxycyclohexanol and an increase in the formation of
cyclohexane, while the yield of cyclohexanol remains relatively similar between the catalysts.
This suggests that increased nickel loading impacts the chemistry of the active site, which is
directly reflected by different selectivity towards various types of hydrodeoxygenation and
hydrogenolysis (i.e., demethoxylation and direct deoxygenation).
This change in the nature of the active site agrees with the change observed in the
intrinsic rate (Fig. 3b). For lower nickel loading, the reaction pathway mostly follows the main
reaction pathway shown in Fig. 4a (dark red pathway). However, as the nickel loading increases,
the amount of 2-methoxycyclohexanol formed is reduced. This indicates two possibilities: (1) the
conversion rate of 2-methoxycyclohexanol on the active site is increased; or (2) there are other
pathways that bypass the formation of 2-methoxycyclohexanol, but still form cyclohexanol and
cyclohexane. The data would be consistent with (1) only if going from 12.5% loading to higher
loadings caused all of the activities downstream of methoxycyclohexanol (DMO, DH, and HYD)
to increase in a coordinated manner, while not affecting the initial hydrogenation activity to the
same extent. Since the methoxycyclohexanol production reaches a lower maximum and a higher
rate of disappearance with increasing nickel loading (Fig. 1), along with changes in product
distribution at the same extent of conversion (Fig. 5), the existence of alternate pathways is more
likely. One potential pathway that does not form 2-methoxycyclohexanol is shown in Fig. 4b
(purple pathway). This reaction pathway starts with demethylation (DME) of guaiacol to
catechol, followed by a hydrogenation (HYD) to 1,2-cyclohexanediol. The cyclohexanediol then
undergoes direct deoxygenation (DDO) to cyclohexanol. From cyclohexanol, the reaction
follows the same reaction pathway as the main reaction pathway going through a dehydration
(DH) of cyclohexanol to form cyclohexene followed by hydrogenation to cyclohexane. As
mentioned previously, such a pathway still produces cyclohexanol and cyclohexane without the
initial formation of 2-methoxycyclohexanol.
Another possible reaction pathway is shown in Fig. 4b (green pathway). In this pathway,
guaiacol undergoes direct deoxygenation (DDO) to form anisole; the anisole is hydrogenated
forming methoxycyclohexane, and the methoxycyclohexane forms cyclohexane via
demethoxylation (DMO). This pathway bypasses 2-methoxycyclohexanol and cyclohexanol
formation and could occur along with the main reaction mechanism in Fig. 4a. The fact that at
20% conversion the cyclohexanol yield remains constant, while there is a decrease in 2-
methoxycyclohexanol yield and an increase in cyclohexane, is consistent with a shift from the
12

main mechanism in Fig. 4a (dark red pathway) to the alternate mechanism in Fig. 4b (purple
pathway). This shift occurs with increasing nickel loading and corresponds to a shift in nickel
species formed during synthesis (Fig. 6). Barton et al. [8] previously reported that as nickel
loading increased, the nickel species formed shifted from a mixture of NiO and nickel
phyllosilicate species to predominantly the NiO species that are present after calcination at 873
K.
Besides the qualitative differences in the local environment of the nickel site that result in
differing intrinsic rates and reaction pathways, the change in nickel loading had an effect on the
specific rates, which are directly related to the dispersion of the nickel particles. The 20 wt%
Ni/HZSM-5 catalyst had the fastest rate of formation and highest yield of cyclohexane per mass
of catalyst (Fig. 1e). However, based on specific rates and, more importantly, on the intrinsic
rates, 20 wt% Ni/HZSM-5 catalyst actually had lower guaiacol conversion rates than the 15 wt%
Ni/HZSM-5 catalyst, while HDO rates were similar to that of 15 wt%. The decrease in guaiacol
conversion rate exhibited is consistent with the idea that the 20 wt% catalyst was composed of
larger nickel particles with lower dispersion than the 15 wt% catalyst. Even though there is more
nickel present with the higher loading, less nickel surface area is accessible. Therefore, much of
the nickel does not participate in the Ni surface-catalyzed reactions. For the nickel loadings less
than 20 wt%, which all have similar dispersions, the percentage of total nickel loaded that is
available to participate in reactions is similar. When dispersions are similar, the specific rate
should increase with nickel loading, since there is proportionally more nickel surface sites
available for reaction.
Overall, the Ni(15)/HZSM-5 catalyst had the greatest conversion of guaiacol and rate of
HDO. This was not only due to its high metal loading, high dispersion, and small particle size, as
was expected from the properties determined from the Barton et al. [8] study, but the nickel sites
themselves were more active, thus making this concentration of nickel precursor in combination
with the corresponding DP method conditions used in this study, the optimal concentration. This
observed difference in metallic sites cannot be attributed to electronic effects and varying nickel
site coordination which are associated with nickel sites from 1-5 nm, since these nickel particles
are in the 10-15 nm size range [34]. An alternative explanation involves the difference in nickel
site to acid site ratio which would be expected with varying nickel loading, with an optimal ratio
occurring at 15 wt%. The influence of nickel site to acid site ratio on HDO activity is not
surprising since previous literature has shown that for supported nickel catalyst systems, HDO
efficiency is directly affected by the presence of acidity, and that presence of acid sites adjacent
to nickel sites influences the adsorption orientation of substrate molecules on the catalytic sites
and consequently the selectivity of the catalyst [14, 35]. Further exploration of the influence of
nickel to acid site ratio would require additional quantification of acid sites, however the
analytical instrumentation required for this analysis was not available.
3.1.2 Effect of Preparation Conditions on Guaiacol HDO
Fig. 7 shows the reaction progress over time for the four different preparations of
Ni(15)/HZSM-5 catalyst. Similar to the Ni(x)/HZSM-5 catalyst experiments, the experiments for
Ni(15)/HZSM-5 catalyst with varying preparation conditions all generated the same three main
products: cyclohexane, 2-methoxycyclohexanol and cyclohexanol. All experiments reached
13

~100% conversion after 1 h, with the product distribution changing over the remaining 3 h of
reaction.
Catalysts DP5_Cal673, DP5_Cal873, and DP16_Cal873 generated similar results (Figs.7a, 7c,
and 7d), in that a large amount (40-45% yield) of 2-methoxycyclohexanol was formed (along
with a minor amount of cyclohexanol), and then these products were gradually consumed and the
cyclohexane was gradually formed throughout the reaction. The DP16_Cal673 catalyst, on the
other hand, performed differently. The main difference was the much sharper increase and
decrease in 2-methoxycyclohexanol, accompanied by a much faster production of cyclohexane
(Fig. 7b). Other differences include the greater production of HDO products other than
cyclohexane, and the reaction being completed within 2 h as opposed to the product distribution
changing over the entire 4 h.
The difference in reaction progress between the DP16_Cal673 catalyst and the other
catalysts is reflected in Fig. 8, which is the product distribution at 20% conversion. The amount
of cyclohexane is much higher, while the production of cyclohexanol is much lower, which
indicates that an alternate pathway is being favored. In general, all the reactions appear to be
following the same main pathway as indicated in Fig. 4a. Catalyst DP16_Cal673 however
appears to catalyze the alternate pathway shown in Fig. 4b (green pathway) to a greater extent,
producing cyclohexane without producing cyclohexanol. In addition, more HDO products other
than cyclohexane are being formed towards the end of the reaction, which suggests that the acid
catalyzed step in Fig. 4a toward methylcyclopentane is also more favorable.
The specific initial rates for the four Ni(15)/HZSM-5 catalysts, shown in Fig. 9a, were
within 2.5-3.5x10-3 mol·g Ni-1·s-1, which is reflected by their similar reaction progress over time
(Fig.7). However, as reflected by the reaction progress of the DP16_Cal673 catalyst, the specific
HDO rate for this catalyst was slightly higher. Unexpectedly, the DP5_Cal673 catalyst had a
higher specific initial rate, but this can be attributed to the slightly faster production of 2-
methoxycyclohexanol and cyclohexanol during the beginning of the reaction. Despite this, the
DP16_Cal673 catalyst still had the highest HDO rate.
The DP16_Cal673 catalyst exhibits the greatest selectivity towards HDO activity of the
catalysts evaluated in this study can be inferred from two distinct data sets. The DP16_Cal673
catalyst clearly exhibits the most rapid production and highest yield of cyclohexane (90%), as
shown in Fig. 7b. While it had the lowest overall intrinsic initial rate, as shown in Fig. 9b., its
HDO rate was similar to the other catalysts, resulting in the largest ratio of HDO rate to
conversion rate (Table 4).
As mentioned in section 2.3.1, the extent of reduction was considered 100% for these
catalysts. This is a valid assumption despite what the XPS data in Table 2B indicate, because as
can be seen in Fig. 10, the DP16 catalysts require less severe conditions to be reduced than the
DP5 catalysts. Also, the catalysts were passivated, which facilitates in-situ re-reduction. The
lower overall intrinsic rate of the DP16 catalysts compared to the DP5 catalysts can be attributed
to the fact that the DP16 catalysts were more selective and proceeded through alternate
pathways, as described in this section. Overall, the DP16_Cal673 catalyst condition led to the
highest dispersion and smallest particle size (Table 1B), as well as the most selective catalyst
site for HDO. This time the nickel site to acid site ratio should be similar for these catalysts, but
in this case the DP16_Cal673 catalyst had nickel particle sizes in the upper limit of 5 nm where
electronic effects and different nickel atom coordination may differ from the other 15 wt%
14

catalyst with larger particle size and this may contribute to the difference in observed selectivity
[34].
3.1.3 Effect of Active HZSM-5 versus Inert SiO2 Support on Guaiacol HDO
To determine the contribution of HZSM-5 support to the catalyst activity, reaction
experiments with a Ni/HZSM-5 catalyst and a comparable Ni/SiO2 catalyst were performed. The
nickel loading was reduced for the Ni/SiO2 preparation to make the ratio of nickel mass to
available BET surface area equal, since the SiO2 has about half the starting surface area of
HZSM-5, ~198 m2·g-1.
The Ni/HZSM-5 catalyst used was a 15 wt% DP16_Cal673 catalyst, and the Ni/SiO2 was
a 7 wt% DP16_Cal673 catalyst (TPR patterns of Ni/SiO2 compared to Ni/HZSM-5, and BJH
mesopore size distributions for Ni/HZSM-5 and Ni/SiO2 are provided in the Supplementary
Material (Figs. S10 and S11, respectively). As can be seen from the reaction progresses
displayed in Fig. 11, the different supports result in catalysts exhibiting markedly different
behavior. With the Ni/HZSM-5, the main products are cyclohexane and HDO products, along
with production and consumption of 2-methoxycyclohexanol and cyclohexanol. However, with
the Ni/SiO2, the main products are 2-methoxycyclohexanol along with some cyclohexanol; the
extent of HDO of guaiacol does not achieve 100%.
The reaction progress for both reactions is consistent with the proposed mechanism in
Fig. 4a. Both reactions follow the nickel catalyzed step of aromatic ring saturation to 2-
methoxycyclohexanol. While the Ni/HZSM-5 is able to perform acid catalysis to form
cyclohexane or methylcyclopentane from 2-methoxycyclohexanol, the Ni/SiO2 has no acid
functionality as confirmed by the absence of acid catalyzed reaction products as indicated in Fig.
11. This conclusion is supported by the difference in acid strength exhibited in the potentiometric
titration data (Table 5).
The presence of cyclohexanol in the Ni/SiO2 reaction indicates there is an alternate
reaction that forms cyclohexanol without forming 2-methoxycyclohexanol, and that pathway was
shown in Fig. 4b (purple pathway). This pathway starts with 1,2 cyclohexanediol formation,
followed by direct deoxygenation to cyclohexanol. However, unlike the Ni/HZSM-5 catalysts,
where cyclohexanol undergoes dehydration due to acid catalysis, the Ni/SiO2 is unable to further
convert the cyclohexanol. As mentioned previously, in addition to the main reaction mechanism
proposed in Fig. 4a, the Ni/HZSM-5 appears to undergo the mechanism in Fig. 4b (green
pathway), which completely bypasses the formation of 2-methoxycyclohexanol and
cyclohexanol. This mechanism uses solely nickel catalysis, and no direct acid catalysis, therefore
one would expect the Ni/SiO2 to also be able to perform this catalytic step and form
cyclohexane; however that is not the case as shown from the progress of the Ni/SiO2 reaction.
This difference in catalytic activity of the nickel may be attributed to the different
supports. Even though the purple pathway in Fig. 4b does not require acid catalysis, as
mentioned in section 1, the presence of acid sites in HZSM-5 adjacent to nickel sites can
influence the mechanism of adsorption [11]. The presence of the acid support enables guaiacol to
be adsorbed onto the catalytic site in such a way that the catalytic step shown in Fig. 4b (green
pathway) can be performed. Without the adjacent acid sites, the nickel sites in Ni/SiO2 are not
able to perform the reaction in Fig. 4b (green pathway), and instead perform the reactions in Fig.
15

4a and b (dark red and purple pathways; without the acid catalyzed steps). Therefore, not only
does the Ni/HZSM-5 function as a multifunctional catalyst system with both nickel and acid
catalysis, the acid sites adjacent to the nickel sites are able to adsorb guaiacol molecules such
that HDO activity is favored.
After demonstrating Ni/HZSM-5 catalyst activity for guaiacol HDO and optimizing its
preparation via the DP method, the Ni/HZSM-5 was evaluated in various aqueous environments.
The catalytic activities of both, the HZSM-5 support and Ni/HZSM-5 catalyst, were evaluated in
neutral and alkaline solutions at the desired reaction conditions (523 K and 5 MPa H2 gas),
specifically for β-O-4 cleavage of 2-phenoxy-1-phenylethanol (PPE).
3.2 PPE β-O-4 Cleavage Activity in Water
When exposed to water and HZSM-5 zeolite, PPE is transformed into two main products
as a result of β-O-4 linkage cleavage, phenol and phenylacetaldehyde, along with some
dehydration (MW=196 g·mol-1) and condensation (MW=222 g·mol-1) products (Fig. 12). Phenol
and phenylacetaldehyde are desirable products, since their presence indicates β-O-4 cleavage of
PPE has occurred; dehydration and condensation products are undesirable. The conversion of
PPE, yield (wt%) to products, and the selectivity to desired products are presented in Fig. S12. In
addition to the observed products, there was a systematic loss of approximately 10 wt% (~0.05
g) material that occurred. This loss most likely occurred during collection, including any gaseous
products, along with some mass loss due to the dehydration and condensation reactions with the
formation of water.
As can be seen from the reaction progress over time (Fig. 12), phenol,
phenylacetaldehyde, and undesired products are formed during the course of the reaction until
the PPE conversion reaches about 100 wt% at around 30 min. After the PPE is completely
consumed, between 30 min and 1 h, the phenylacetaldehyde concentration starts to decrease
while the undesired product concentration increases. This suggests that there may be a secondary
reaction occurring. To test this, reactions with just phenol, phenylacetaldehyde, and the
combination of the two were performed under the same conditions. While phenol did not react
after 1 h in the reaction system, phenylacetaldehyde did form additional products including the
undesired 222 g·mol-1 product. The combination of the two formed similar products as in the
pure phenylacetaldehyde case, and phenol appeared to be mostly inactive since 95 wt% of the
phenol was recovered in the product mixture.
Based on the range of identified products present in the PPE reaction mixture and the side
reaction experiments, the reaction mechanisms in Fig. 13 were proposed. Fig. 13a shows the
main reaction mechanism where HZSM-5 catalyzes the β-O-4 cleavage of PPE, while Fig. 13b
shows the suggested mechanism for the formation of undesired products. The presence of
phenylacetaldehyde and phenol indicates that acid sites do not simply cleave the ether linkage
and form two products in a single step; otherwise only phenol and acetophenone or 1-
phenylethanol would have been formed. The exclusive presence of phenylacetaldehyde as
opposed to acetophenone suggests that the acid site attacks the ether linkage and releases phenol,
but instead of a one-step reaction, there is the formation of an unstable epoxide type ring due to
electron propagation from the hydroxide species on the PPE. This oxirane intermediate has the
possibility to open up in two ways to form a stable product. The first pathway, the formation of
16

acetophenone, is the expected one because resonance stability is preserved, therefore more stable
than the other possible product of ring opening, phenylacetaldehyde. However, acetophenone is
not formed; instead the ring opens via the second path to form the less stable product,
phenylacetaldehyde. This suggests that this pathway is kinetically favored, an outcome possibly
influenced by the narrow pore size of the catalyst.
The undesired reaction products include a putative result of PPE dehydration (196 g·mol-
1
from the initial 214 g·mol-1), and a higher molecular weight compound (222 g·mol-1).
Dehydration can be the result of the acid site attacking the hydroxyl group of PPE instead of the
ether linkage. The reaction mechanism for the formation of the 222 g·mol-1 product is unknown.
This uncertainty is mostly due to the inability to identify the product using GC-MS (mass spectra
are provided in Fig. S13). Despite the unknown product structure and mechanism, from the side
reactions discussed above, it is known that the phenylacetaldehyde appears to participate in the
reaction to form the 222 g·mol-1 product.
From the model reaction with PPE, it has been shown that the HZSM-5 zeolite performs
acid catalysis, and with the single catalyst functionality, it is able to cleave the β-O-4 linkage
with a selectivity of 68 wt% and yield of 65 wt% to desirable cleavage products (96 wt%
conversion of PPE). It was also shown that the HZSM-5 is able to perform the catalysis in an
aqueous environment and at a relatively low temperature (523 K), which is quite different from
its typical application which is high temperature pyrolysis in the gas phase. The next step was to
use the bifunctional catalyst with the presence of nickel metal particles on the HZSM-5 support
prepared via the DP method described earlier by Barton et al. [8], and compare the activity and
selectivity of the Ni/HZSM-5 in water to that of HZSM-5 in water for PPE model compound
reactions. The bifunctionality of the Ni/HZSM-5 catalyst was confirmed in section 3.1.3 using
guaiacol HDO reactions in dodecane to compare the performance of Ni/HZSM-5 with a Ni/SiO2
catalyst (inert support). The bifunctional Ni/HZSM-5 catalyst was able to convert guaiacol to
cyclohexane, while the Ni/SiO2 catalyst was only able to convert the guaiacol to cyclohexanol
due to inability to perform the acid catalyzed dehydration step from cyclohexanol to
cyclohexane.
The results from the Ni/HZSM-5 catalytic reactions against PPE are presented in Figs.
14a and S14. The Ni/HZSM-5 catalyzed reaction in water exhibited very rapid initial conversion
of PPE. The main cleavage products are the desired products, ethylbenzene and phenol; since
phenol is an important industrial feedstock and ethylbenzene is a completely deoxygenated
aromatic. The production of total monomers again is in the 60-70 wt% range. The control
reaction in water without any catalyst exhibits little to no reaction, as shown in Fig. 14b. Very
little PPE is converted and the reaction is much slower, forming only a small amount of large
undesired dimer products that are mainly dehydration products.
The reaction scheme proposed in Fig. 15 is helpful in understanding the time course of
reaction (Fig. 14a). The carbon balance indicates that 2-phenethyl phenyl ether (PPEther) is a
reaction intermediate; as suggested by Fig. 15, it is the result of PPE undergoing deoxygenation.
Subsequent hydrogenolysis of the ether generates ethylbenzene and phenol. Once produced, the
ethylbenzene remains in the product mixture without further reaction, whereas the phenol is
susceptible to undergo further hydrogenation especially after longer reaction durations. The
Ni/HZSM-5 in an aqueous environment demonstrates clearly its ability to generate highly
desirable aromatic monomer compounds. It also shows that the nickel metal is in fact active
17

along with the HZSM-5 support in water, since the reaction mechanisms observed for the
Ni/HZSM-5 versus the plain HZSM-5 catalyst (Figs. 13 and 15) are quite different. This is in
stark contrast with studies described in the literature where non-polar organic solvents are used
and ring saturation is predominant [21, 23, 36].
3.3 PPE β-O-4 Cleavage Activity in NaOH Solution
The alkaline PPE cleavage reactions exhibited rapid conversion of PPE (within 15 min)
in the presence and absence of catalyst. As shown in Fig. 16, the reaction with Ni/HZSM-5
catalyst was able to almost completely convert the PPE in 15 min, and at the 15 min mark, the
sample had the largest fraction of aromatic monomers (phenol and methyl benzyl alcohol
(MBA)). Following cleavage of the β-O-4 linkage in PPE, the resulting aromatic monomers
undergo ring saturation. Phenol was converted to saturated products, including cyclohexanone
and cyclohexanol, and the trace aromatics present were also mostly saturated. Only a minor
amount of MBA was saturated, but it did appear to undergo a secondary reaction that has not
been explored in detail.
The time course of the PPE cleavage reaction, in alkaline solution without catalyst, is
shown in Fig. 16b. A majority of the PPE was converted within 15 min, though to a lesser extent
and yielding a different product distribution than the Ni/HZSM-5 catalyzed reaction. A more
detailed comparison of the product yields for the PPE cleavage with and without Ni/HZSM-5 is
presented in Fig. S15. Rather than phenol and MBA, followed by saturation products of phenol
and secondary reaction products of MBA, the NaOH-catalyzed system gave primarily phenol
and phenylethanediol, with some undesired dimer products but without any further saturation
products. Monomer levels fluctuated between 60-70 wt% without catalyst. Ni/HZSM-5 catalyzed
reactions generated 75% total monomers early in the reaction, but secondary reactions led to
subsequent monomer consumption. This confirmed that Ni/HZSM-5 is active in this system.
The reaction schemes proposed in Fig. 17 are consistent with the products identified and
suggest two important conclusions. The first being that Ni/HZSM-5 remains active in alkaline
solution, and is able to exhibit catalytic activity characteristic of nickel catalysis, including both
hydrogenolysis and hydrogenation, as demonstrated by the unique presence of un-oxidized
cleavage products (Fig. 17b). The second is that both Ni/HZSM-5 and NaOH participate in the
cleavage of lignin type linkages. This, along with the ability to solubilize the lignin in an alkaline
solution, makes the alkaline Ni/HZSM-5 system a potential candidate for depolymerization of
lignin into monomers.
4. Conclusions
All the Ni(x)/HZSM-5 catalysts in this study were able to perform HDO of guaiacol to
varying extents. Furthermore, the combined functionality of Ni/HZSM-5 was shown to perform
both nickel and acid catalyzed reactions, as well as facilitate reaction pathways that were not
evident with Ni/SiO2. The mixture of nickel species resulting from the DP method allowed for a
few different reaction mechanisms to occur during the guaiacol HDO reaction, while the
presence of HZSM-5 allowed more complete HDO through dehydration reactions and facilitated
an alternate direct deoxygenation pathway.
18

The guaiacol HDO catalyst assays indicated that the Ni(15)/HZSM-5 catalyst with a DP
time of 16 h and calcined at 673 K had the highest selectivity for HDO products, the highest
nickel dispersion, and the smallest particle size. It had higher selectivity toward cyclohexane,
with less accumulation of 2-methoxycyclohexanol, suggesting that it can perform direct
deoxygenation on various lignin-type substituents. While the Ni/HZSM-5 catalysts prepared in
these studies were able to completely deoxygenate the guaiacol model compound, all the
guaiacol molecules also underwent complete saturation of the aromatic ring, which was
undesirable. However, this ring saturation is most likely a result of using the simplified model
compound in an organic solvent system instead of using a larger, more complex compound in an
aqueous environment.
Exposing the Ni(15)/HZSM-5 DP16_Cal673 catalyst to a lignin model compound (PPE)
in aqueous environments produced ethylbenzene and phenol (~35% and 23%, respectively) as its
two main products, with a low extent of ring saturation compared to reactions in a non-polar,
organic solvent, and a low yield of recombination products as compared with HZSM-5 in water.
This selectivity towards aromatic compounds (~60%), with minimal selectivity for ring
saturation and recombination reactions (both less than 10%) is important for the valorization of
lignin since some of the issues that plague many lignin depolymerization technologies include
oversaturation of the aromatic structure and char/coke formation (as a result of uncontrolled
recombination reactions). PPE reactions with Ni/HZSM-5 in the alkaline aqueous system also
exhibited promising results for β-O-4 cleavage but resulted in oversaturation of the aromatic
monomers. While this is undesired, this may not occur with more complex substrates, as in the
case of lignin. Also, the presence of NaOH may be necessary for the solubilization of lignin prior
to reaction. This Ni/HZSM-5 catalyst prepared by a modified DP method, was able to avoid the
mentioned undesired pathways while using a relatively low reaction temperature in aqueous
environments (particularly in pure water), which makes this a promising catalyst system for
lignin depolymerization into aromatic monomers.
5.
Nomenclature
Dp
λ
β1/2
θ
particle diameter (Å)
CuKα radiation wavelength (Å)
full width at half maximum (radians)
Bragg angle (radians)
D(%) dispersion (%)
dp
IR
SR
Nav
particle diameter (nm)
intrinsic rate (molecules of guaiacol per available nickel site) (s-1)
specific rate (mol·g-1·s-1)
Avogadro’s Number
NiSA number of available nickel sites per mass of nickel (g-1)
NiSC number of nickel sites per mass of catalyst (g-1)
NiL
nickel loading (fraction)
Acknowledgments
19

The work was supported by the National Institute of Food and Agriculture, U.S. Department
of Agriculture, under Agreement No. 2010-38420-21828, the Project Basal PFB-27 of
CONICYT, FONDECYT 1140528, and the Program PMI InEs UCO-1302 of the Chilean
Ministry of Education. This work was performed in part at the Analytical Instrumentation
Facility (AIF) at North Carolina State University, which is supported by the State of North
Carolina and the National Science Foundation (award number ECCS-1542015). The AIF is a
member of the North Carolina Research Triangle Nanotechnology Network (RTNN), a site in the
National Nanotechnology Coordinated Infrastructure (NNCI). This work was also performed in
part at the Environmental and Agricultural Testing Service Laboratory (EATS), Department of
Crop and Soil Sciences, at North Carolina State University.
References
[1]
[2]
F. G. Calvo-Flores, J. A. Dobado, ChemSusChem 2010, 3 (11), 1227-1235.
J. Zakzeski, P. C. A. Bruijnincx, A. L. Jongerius, B. M. Weckhuysen, Chem. Rev. 2010,
110 (6), 3552-3599.
[3]
[4]
U.S. Department of Energy. 2011, R.D. Perlack and B.J. Stokes (Leads), ORNL/TM-
2011/224. Oak Ridge National Laboratory, Oak Ridge, TN.
J. J. Bozell, J. E. Holladay, D. Johnson, J. F. White, Top Value Added Chemicals from
Biomass. Volume II, Pacific Northwest National Laboratory, Richland, WA, PNNL-
16983 (2007).
[5]
[6]
M. P. Pandey, C. S. Kim, Chem. Eng. Technol. 2011, 34 (1), 29-41.
P. Azadi, O. R. Inderwildi, R. Farnood, D. A. King, Renewable and Sustainable Energy
Reviews 2013, 21, 506-523.
[7]
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, C. E. Wyman, Science 2014, 344 (6185).
[8]
[9]
R. R. Barton, M. Carrier, C. Segura, J. L. G. Fierro, N. Escalona, S. W. Peretti, Applied
Catalysis A: General 2017, 540, 7-20.
C. Zhao, S. Kasakov, J. He, J. A. Lercher, Journal of Catalysis 2012, 296, 12-23.
[10] C. Zhao, J. A. Lercher, Angewandte Chemie 2012, 124 (24), 6037-6042.
[11] W. Song, C. Zhao, J. A. Lercher, Chemistry – A European Journal 2013, 19 (30), 9833-
9842.
20

[12] J. He, C. Zhao, J. A. Lercher, J. Am. Chem. Soc. 2012, 134 (51), 20768-20775.
[13] J. He, C. Zhao, D. Mei, J. A. Lercher, Journal of Catalysis 2014, 309, 280-290.
[14] W. Song, Y. Liu, E. Barath, C. Zhao, J .A. Lercher, Green Chem. 2015, 17 (2), 1204-
1218.
[15] P. Burattin, M. Che, C. Louis, The Journal of Physical Chemistry B 1998, 102 (15),
2722-2732.
[16] P. Burattin, M. Che, C. Louis, The Journal of Physical Chemistry B 2000, 104 (45),
10482-10489.
[17] A. B. Dongil, I. T. Ghampson, R. Garcia, J. L. G. Fierro, N. Escalona, RSC Advances
2016, 6 (4), 2611-2623.
[18] A. L. Jongerius, R. W. Gosselink, J. Dijkstra, J. H. Bitter, P. C. A. Bruijnincx, B. M.
Weckhuysen, ChemCatChem 2013, 5 (10), 2964-2972.
[19] Z. He, X. Wang, Frontiers of Chemical Science and Engineering 2014, 8 (3), 369-377.
[20] M. Lu, J. Zhu, M. Li, Y. Shan, M. He, C. Song, Energy Fuels 2016, 30 (8), 6671-6676.
[21] J. He, L. Lu, C. Zhao, D. Mei, J. A. Lercher, Journal of Catalysis 2014, 311, 41-51.
[22] J. Kong, B. Li, C. Zhao, RSC Advances 2016, 6 (76), 71940-71951.
[23] X. Wang, R. Rinaldi, ChemSusChem 2012, 5 (8), 1455-1466.
[24] S. K. Singh, J. D. Ekhe, RSC Advances 2014, 4 (53), 27971-27978.
[25] H. Konnerth, J. Zhang, D. Ma, M. H. G. Prechtl, N. Yan, Chemical Engineering Science
2015, 123, 155-163.
[26] R. Nares, J. Ramírez, A. Gutiérrez-Alejandre, C. Louis, T. Klimova, The Journal of
Physical Chemistry B 2002, 106 (51), 13287-13293.
[27] C. H. Bartholomew, R. J. Farrauto, Journal of Catalysis 1976, 45 (1), 41-53.
[28] J. T. Richardson, R. J. Dubus, Journal of Catalysis 1978, 54 (2), 207-218.
[29] J. S. Smith, P. A. Thrower, M. A. Vannice, J. Catal., 1981, 68 (2), 270-285.
21

[30] R. Cid, G. Pecchi, Applied Catalysis 1985, 14, 15-21.
[31] J. M. Nichols, L. M. Bishop, R. G. Bergman, J. A. Ellman, J. Am. Chem. Soc. 2010, 132
(36), 12554-12555.
[32] N. Escalona, W. Aranzaez, K. Leiva, N. Martínez, G. Pecchi, Applied Catalysis A:
General 2014, 481, 1-10.
[33] K. Leiva, C. Sepúlveda, R. García, J. L. G. Fierro, N. Escalona, Catalysis
Communications 2014, 53, 33-37.
[34] M. Che, C. O. Bennett, In Advances in Catalysis, D.D. Eley, H. P.; Paul, B. W., Eds.
Academic Press: 1989; (36), 55-172.
[35] S. Qiu, Y. Xu, Y. Weng, L. Ma, T. Wang, Catalysts 2016, 6 (9), 134.
[36] S. Kasakov, H. Shi, D. M. Camaioni, C. Zhao, E. Barath, A. Jentys, J. A. Lercher, Green
Chem. 2015, 17 (11), 5079-5090.
22

100
80
60
40
20
0
100
80
60
40
20
0
a)
100
80
60
40
20
0
100
80
60
40
20
0
d)
0
50
100
150
200
250
0
50
100
150
200
250
Time (min)
Time (min)
100
80
60
40
20
0
100
80
60
40
20
0
100
80
60
40
20
0
100
80
60
40
20
0
b)
e)
0
50
100
150
200
250
0
50
100
150
200
250
Time (min)
Time (min)
100
80
60
40
20
0
100
80
60
40
20
0
c)
0
50
100
150
200
250
Time (min)
Fig. 1. Conversion of guaiacol and yield of products as a function of reaction time for a) Ni(5)/HZSM-5, b)
Ni(10)/HZSM-5, c) Ni(12.5)/HZSM-5, d) Ni(15)/HZSM-5, and e) Ni(20)/HZSM-5. List of compounds for Trace
HDO, Methoxy and Hydroxy categories can be found in Fig. S16 in the supplementary material.
(Print with color)
23

3.00x10-3
2.50x10-3
2.00x10-3
1.50x10-3
1.00x10-3
Initial Rate
HDO Rate
a)
5.00x10-4
0.00
-5
SM
-5
-5
SM
-5
SM
-5
SM
Z
H
SM
Z
Z
Z
Z
H
)/
H
)/
H
)/
H
)/
)/
0
0
5
.
5
(5
i
2
(1
(1
(2
i
i
i
N
(1
N
N
N
i
N
6
Intrinsic Rate
HDO Rate
b)
4
2
0
-5
SM
-5
-5
SM
-5
SM
-5
SM
Z
H
SM
Z
Z
Z
Z
H
)/
H
)/
H
)/
H
)/
)/
0
0
5
.
5
(5
i
2
(1
(1
(2
i
i
i
N
(1
N
N
N
i
N
Fig. 2. a) Specific conversion and HDO rates and b) intrinsic conversion and HDO rates for Ni(x)/HZSM-5
catalysts.
(Print with color)
24

6
5
4
3
2
1
0
a)
Conversion
2
3
1
0
2
4
6
8
10 12 14 16 18 20
Ni Loading (wt %)
2.0
b)
HDO
1.5
1.0
0.5
0.0
2
1
0
2
4
6
8
10 12 14 16 18 20
Ni Loading (wt %)
Fig. 3. Plots of intrinsic rate against nickel loading; a) guaiacol conversion, b) HDO.
(Print with color)
25

a)
DDO – Direct Deoxygenation
DH – Dehydration
H+/ – Acidic conditions
DME - Demethylation
DMO – Demethoxylation
HYD – Hydrogenation
b)
Fig. 4. Reaction network for guaiacol conversion; a) main pathways, b) alternate pathways: (
) nickel catalysis,
(
) acid catalysis; Observed Product, Final Product, [Theoretical Intermediate]; Dominant pathway represented
by dark red molecules, alternate pathway 1 is represented by purple molecules, alternate pathway 2 is represented by
green molecules.
(Print with color)
26

20
18
16
14
12
10
8
2-Methoxycyclohexanol
Cyclohexane
Cyclohexanol
Trace HDO
Methoxy
Hydroxy
6
4
2
0
-5
N
-5
SM
-5
SM
-5
SM
Z
H
SM
Z
Z
Z
H
)/
H
)/
H
)/
)/
0
0
5
.
5
2
(1
(1
(2
i
i
i
(1
N
N
N
i
Fig. 5. Product distribution at 20% conversion of guaiacol for Ni(x)/HZSM-5 catalysts.
(Print with color)
27

NiO
Ni phyllosilicate
Ni(20)/HZSM-5
Ni(15)/HZSM-5
Ni(12.5)/HZSM-5
Ni(10)/HZSM-5
Ni(5)/HZSM-5
273 373 473 573 673 773 873 973 1073 1173 1273
Temperature (K)
Fig. 6. Temperature programmed reduction (TPR) for Ni(x)/HZSM-5 catalysts [8].
(Print with color)
28

a)
d)
100
80
60
40
20
0
100
80
60
40
20
0
100
80
60
40
20
0
100
80
60
40
20
0
0
50
100
150
200
250
0
50
100
150
200
250
Time (min)
Time (min)
b)
100
80
60
40
20
0
100
80
60
40
20
0
c)
100
80
60
40
20
0
100
80
60
40
20
0
0
50
100
150
200
250
0
50
100
150
200
250
Time (min)
Time (min)
Fig. 7. Conversion of guaiacol and yield of products as a function of reaction time for Ni(15)/HZSM-5 catalysts a)
DP5_Cal673, b) DP16_Cal673, c) DP5_Cal873, and d) DP16_Cal873. List of compounds for Trace HDO, Methoxy
and Hydroxy categories can be found in Fig. S16 in the supplementary material.
(Print with color)
29