This	document	is	the	Accepted	Manuscript	version	of	a	Published	Work	that	appeared	in	final	form	in	ACS	Catalysis,	copyright	©	
American	Chemical	Society	aAer	peer	review	and	technical	ediCng	by	the	publisher.	To	access	the	final	edited	and	published	work	see	
hEps://pubs.acs.org/doi/abs/10.1021/acscatal.8b00067
Research Article

Cite This: ACS Catal. 2018, 8, 4230−4240

pubs.acs.org/acscatalysis

Chirality, Rigidity, and Conjugation: A First-Principles Study of the
Key Molecular Aspects of Lignin Depolymerization on Ni-Based
Catalysts
Qiang Li* and Núria López*
Institute of Chemical Research of Catalonia (ICIQ), The Barcelona Institute of Science and Technology, Avgda. Països Catalans 16,
43007 Tarragona, Catalonia, Spain

Downloaded via UNIV ROVIRA I VIRGILI on April 15, 2019 at 09:59:56 (UTC).
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

S
* Supporting Information

ABSTRACT: Efficient lignin depolymerization is crucial for achieving a positive economical return in biorefineries. However,
because of the complex cross-polymerized network formed by the aromatic lignols, catalysts that can work at low temperatures
perform insufficiently in terms of both activity and selectivity. Recently, Ni-based catalysts have been reported to exhibit a
superior catalytic behavior in lignin valorization, but the mechanism of the depolymerization remains unclear. In this study, we
employed density functional theory to investigate lignin decomposition on pure and Ru-doped Ni(111) surfaces to unravel the
key issues that limit performance. The reaction network was screened by using complex coniferyl dimer models containing
different stereocenters to simulate the most abundant β-O-4 linkages in lignin, thus presenting a better representation of the
complexity of the parent compound. Our results show that on nickel, both adsorption geometries and the preferred reaction
paths are different depending on the chirality of the dimer reactants. In addition, the rigidity of the β-O-4 link is also responsible
for the reactivity found. Finally, the presence of small amounts of Ru on the surface simplifies the reaction as they act as
preferential points for β-O-4 bond cleavage via the weakened C−O bond that is induced by increasing metal−O bond strength.
KEYWORDS: lignin, DFT, β-O-4, chirality, rigidity, conjugation, Ni, Ru

1. INTRODUCTION
Lignin is a complex cross-linked phenolic polymer that widely
exists in the cell walls and supports plants via its strength and
rigidity. It is one of the main components in lignocellulosic
biomass resources, accounts for 15−30% of the weight, and
stores 40% of the energy.1 However, because of its high
stability, efficient depolymerization methods are lacking,2−4 and
98% of the lignin produced by the paper and pulp industry was
considered a low-value material and burned for heating at the
beginning of the 21st Century.5
In the near future, because of the depletion of oil reserves,
lignin will be seen as a new source of the major aromatic
compounds (such as the BTX fraction of benzene, toluene, and
the three xylene isomers)6 or will directly provide some
synthetic intermediates that are used in the industry.
Altogether, the derived platform chemicals could be used as
renewable materials in the polymer industry with great
potential in food and medical applications and as carbon fibers
to name just a few.7,8 However, the optimal processing of lignin
© 2018 American Chemical Society

is far from being realized. Hence, developing catalysts with high
efficiency and selectivity for depolymerization is the key to
allowing application in high value-added chemicals and
improving the economics of the biorefinery process.9−11
One of the added difficulties in the chemical treatment of
lignin is its variable composition when different raw materials,
such as hardwood, softwood, and grass, are fed in the
bioconfinery process.3 The proportions of the monomeric
units and linkages depend on the origin of lignin.3 The most
common monolignols (see Figure 1) from the depolymerization are p-coumaryl (H), coniferyl (G), and sinapyl alcohols
(S)12 and different linkage topologies connecting these
monomers have been identified: β-O-4, β-5, α-O-4, β-β, 5-5,
4-O-5, β-1, and others (see Figure S1). The most representative
linkage is β-O-4, which accounts for 43−50% in softwood and
Received: January 6, 2018
Revised: March 4, 2018
Published: April 4, 2018
4230

DOI: 10.1021/acscatal.8b00067
ACS Catal. 2018, 8, 4230−4240

ACS Catalysis

Research Article

and the steric effect of methoxy or other groups in the meta and
para aromatic substitution positions would also have an impact
on the interactions between molecules and the surface and
further affect the reaction network.25
Besides, to improve the catalyst design, a further understanding of the reaction process at the atomic level is needed.
Knowledge of the depolymerization mechanisms of lignin on
the Ni-based catalysts is lacking. This is due to several factors,
including: (i) the large size of lignin that has prompted the use
of small surrogates4 that does not account for key aspects like
the chirality of the linkers, the rigidity of the molecules, and the
conjugation of intermediates, (ii) the complex reaction
networks that involve various products, and (iii) the variety
of metal catalysts and the complexity of advanced materials like
TiN-Ni.16 In view of these fundamental complexities, we have
employed DFT calculations on slab models representing the
simplest of these materials (Ni and Ru-doped Ni) but
considered large molecular surrogates that have stereocenters
[dimers (C17H20O6) of the most abundant monolignols
coniferyl alcohol (G)] and an extensive reaction network that
accounts for many of the potential transformations taking place.
Our results first reveal the effects of the stereocenters on the
decomposition mechanism and the synergistic effects on the
Ru-doped Ni catalyst, thus providing guidelines for the design
of the incoming catalysts.

Figure 1. Structures of three lignin monomers.

50−65% in hardwood and happens to be the most labile.
Therefore, breaking the β-O-4 bond is the highest priority in
catalytic depolymerization and valorization processes.
Some compounds have been proposed as homogeneous
depolymerization catalysts, but heterogeneous catalytic lignin
transformation has advantages in flow processing,13,14 production separation, and catalyst recycle ability and is suitable for
large industry scale processes. Comprehensive reviews of the
advances in the area were presented by Calvo-Flores,7
Zakzeski,2 Zhang,3 and Behling et al.,4 focusing on: (i) the
basic lignin compositions, (ii) various lignin valorization
approaches in both heterogeneous and homogeneous systems,
and (iii) wide potential uses of its products in chemicals, fuels,
and materials. A majority of catalysts investigated are transition
metal-based, including those based on Mn, Co, Ni, Ru, Pd, and
Pt. Multicomponent catalysts like transition metal (Ti, Mo, Nb,
and W) nitrides have been applied by Chen et al.,15 and Ti−N
was identified as a promising catalyst for partial lignin
depolymerization. The titanium−nitride−nickel catalyst (TiNNi) has also been used under mild reaction conditions by
Molinari et al.16 and exhibited a performance comparable to
that of the Raney Ni catalyst, which is a well-known catalyst for
hydrogenation processes.
Among the transition metal catalysts, Ni exhibits superior
activity in aryl−alkyl bond activations in lignin hydrogenolysis,
as the aromatic moieties are kept intact in both homogeneous14
and heterogeneous17 processes. Still, Ni-based catalysts exhibit
poor activity below 120 °C,18 and thus, improvements in
performance to reduce working temperatures are desirable. By
combining Pd and Ni into a bimetallic nanoparticle catalyst,
Zhang et al.19 observed a remarkable catalytic enhancement
compared with the results with only Ni and Pd. A previous
study by Yan et al.18 showed that a Ni−Ru bimetallic catalyst
exhibits excellent behavior in aqueous lignin hydrogenolysis,
compared with that of Ni−Rh and Ni−Pd catalysts. Both
studies show the synergistic effects of two metals.
Density functional theory (DFT) approaches to understanding the reaction network for depolymerization are rather
scarce. By combining DFT calculations and experiments, Wang
et al.20 investigated the β-O-4 bond scission on Pd(111). In this
study, the lignin model was represented by 2-phenoxy-1phenylethanol and its β-O-4 bond cleavage was studied.
Recently, adsorption structures of the same simple model on
Ni(111) and a MoS2 surface have been optimized.21 In
addition, Hamou et al. 22 studied the adsorption and
decomposition mechanism of lignin on Pt(111) by using 2phenoxyethanol. The use of surrogates has provene to be
interesting for small molecules;23 however, lignin models
employed in theoretical and experimental studies are exceedingly too simple to fully represent and describe the real lignin
adsorption configurations.4 For example, numerous lignin
linkages with stereocenters are formed between monolignols,24

2. COMPUTATIONAL DETAILS
In this work, slab model calculations were performed by using
the Vienna Ab-initio Simulation Package (VASP).26,27 The
exchange and correlation energies were obtained via the PBE
functional. 28 We also included van der Waals (vdW)
corrections by applying the Grimme’s DFT-D2 method,29,30
with the C6 reparametrized by our group.31 The inner electrons
were represented by projector augmented wave (PAW)
pseudopotentials with cutoff energies of 450 eV.32,33 The
Brillouin zone was sampled by a Γ-centered k-point mesh
generated through the Monkhorst−Pack method.34 The
calculated Ni and Ru lattice parameters of 3.517 and 2.712 Å,
respectively (c/a = 1.581), agree well with previous
experimental results of 3.524 and 2.706 Å, respectively (c/a =
1.582).35
For small species such as ethylene glycol (EG), H2O, H, CO,
O, and OH, adsorptions were optimized on a p-(3 × 3)
supercell with four layers and Γ-centered 5 × 5 × 1 k-point
sampling was used. The energies are listed in Table S1. In larger
slab optimizations, the Ni(111) and Ru(0001) surfaces were
modeled by a four-layer slab with p-(7 × 4) supercells, the two
uppermost layers were fully relaxed, and the remaining bottom
atoms were fixed to the bulk distances. Then, the atoms in the
last bottom layer were removed and the three remaining layers
were fixed and used for adsorption and transition state
calculations. As the adsorption was performed on one side of
the slab, a 15 Å thick vacuum region was set between each slab
to prevent their interaction, and dipole correction along the z
direction36 was also applied to eliminate the spurious
contributions arising from the system’s asymmetry. To reduce
the computational burden, we used Γ-point sampling to obtain
the adsorption and transition state structures first. Then, they
were taken as the initial inputs for reoptimization via the denser
k-point mesh of 2 × 3 × 1. Because of the large sizes of slabs
and adsorbates, the level of accuracy in our model has been
tested by comparing the adsorption and reaction energies on a
three-layer slab Ni(111) surface with those on the four-layer
4231

DOI: 10.1021/acscatal.8b00067
ACS Catal. 2018, 8, 4230−4240

ACS Catalysis

Research Article

Figure 2. Adsorptions of four β-O-4 bond isomers, two structures from the A isomer after rotation of the Cα−Cβ bond and using the other face on
the Ni(111) surface. ArRu and ORu are two surface A isomer structures on the Ru−Ni(111) surface. Ni, Ru, C, H, and O atoms are colored blue,
yellow, gray, white, and red, respectively.

Scheme 1. Reaction Network of Isomer A Depolymerization and Hydrogenation on the Ni(111) Surfacea

a

Blue and red pathways are preferred reactions from α-C and β-C sides, respectively.

step of 0.02 Å and its diagonalization that rendered a single
imaginary frequency.
All molecules in the gas phase were relaxed in a box with
dimensions of 20 × 20 × 20 Å3. Four coniferyl alcohol dimers
with different chiralities [named A (α-S, β-S), B (α-R, β-S), C
(α-R, β-R), and D (α-S, β-R)] were used as the lignin β-O-4
model in this study, as shown in Figure 2 and Figure S2. The
aliphatic side chain on the right aromatic moiety was removed
to reduce the length of the molecule and also the model surface
length along the y direction. Solvation effects have been
considered by calculating solvation energies of those closedshell molecules both in ethylene glycol and in water via the
MGCM methodology (multigrid continuum model).40,41
To study the Ru doping effects, the system stability was
investigated first via a few aspects following the procedures as in

slab. The tested results are listed in Tables S2 and S3. The
adsorption energy differences of four β-O-4 models and 10
depolymerization products are in the range of 0.14−0.23 eV,
which is within the traditional accuracy of the GGA-PBE level
we used.37 Eight reactions on two slabs were tested, and their
reaction energy differences are from 0.00 to 0.05 eV,
demonstrating the reliability and accuracy by using a threelayer slab to describe the surface reactions.
We employed the climbing image nudged elastic band
method (CI-NEB),38 the improved dimer method (IDM),39
and the quasi-Newton algorithm to locate and optimize the
transition states in the potential surfaces. The thresholds were
10−5 eV and 0.03 eV/Å for electronic and ionic relaxations,
respectively. The saddle point nature of the transition states
was assessed by the calculation of the numeric Hessian with a
4232

DOI: 10.1021/acscatal.8b00067
ACS Catal. 2018, 8, 4230−4240

ACS Catalysis

Research Article
a

Scheme 2. Reaction Network of Isomer B Depolymerization and Hydrogenation on the Ni(111) Surface

a

Blue and red pathways are preferred reactions from α-C and β-C sides, respectively.

Table 1. Adsorption Energies (Eads in electronvolts) of Closed-Shell Species and Small Molecules on the Ni(111), Ru(0001),
H2O
and Ru−Ni(111) Surfaces and Their Solvation Energies in EG (EEG) and Water (Esol )a
sol
species

Ni(111)

Ru(0001)

Ru−Ni(111)

EEG
sol

EH2O
sol

Ncovered

Eads(EG)

A
B
C
D
E
G0
G1
G2
G3
G4
G5
G6
G7
G8
G9
G10

−2.28
−2.29
−1.76
−1.86
−2.33
−1.24
−2.34
−2.07
−1.78
−1.96
−1.62
−1.57
−1.44
−1.20
−1.29
−1.53

−3.59
−3.42
−3.12
−3.16
−3.43
−1.96
−3.38
−3.28
−3.07
−2.99
−2.47
−2.29
−2.30
−2.02
−1.97
−2.17

−2.28
−2.35
−1.83
−1.97
−2.39
−1.18
−2.58
−2.24
−2.30
−2.23
−1.78
−1.76
−1.58
−1.31
−1.43
−1.53

−0.60
−0.60
−0.60
−0.62
−0.56
−0.32
−0.49
−0.39
−0.45
−0.36
−0.49
−0.44
−0.35
−0.48
−0.35
−0.26

−1.00
−1.02
−1.02
−1.04
−0.59
−0.47
−0.75
−0.61
−0.73
−0.73
−0.79
−0.69
−0.60
−0.49
−0.60
−0.26

20
20
20
20
20
12
13
13
13
13
13
13
13
13
13
13

−0.21
−0.23
0.31
0.20
−0.25
−0.04
−0.90
−0.80
−0.33
−0.62
−0.22
−0.26
0.01
0.19
0.01
−0.32

a

Ncovered stands for the surface metal atoms covered by the adsorbates. Eads(EG) is the solvation effect-corrected adsorption energy of a species on
the Ni(111) surface in ethylene glycol.

ref 42. The solubility was calculated in a bulk model containing
32 Ni atoms. Segregation and near surface energies were
obtained via a p-(2 × 2) slab with six layers, and the top four
layers were optimized. The island formation energy was
calculated on a p-(3 × 3) slab with four layers, and the top
two layers were relaxed. As listed in Table S5, the energy of
dissolving one Ru atom into the Ni bulk is 0.53 eV, implying
the a relative instability in forming the alloy. However, the
endothermic energy would be compensated by the entropic
contributions as Ru can replace Ni atoms from three directions
with higher disorder, which corresponds to the higher entropy
and therefore reduces the dissolving energy.42,43 A small
exothermic segregation energy (−0.05 eV) implies that
penetration of Ru into the bulk is slightly preferable and can
occur both on the surface and in the bulk. From the positive
island formation energy (0.12 eV), it is safe to derive that Ru
dispersion on the surface is preferred. According to this
analysis, Ru effects have been investigated by substituting the
Ni atom under the breaking bonds with Ru on the top layer. In
other words, the active site Ni atoms for a variety of types of
bond breakages as well as adsorption structures have been

replaced by the Ru atoms specifically, depending on the
reaction types.
All energies presented in the schemes and diagrams were
obtained from the gas phase and slab calculations without
solvent or zero-point energy contributions except as indicated
otherwise. All coordinates of these structure geometries as well
as the related transition states in this work are listed in the
Supporting Information by using VASP POSCAR’s format. In
addition, the data are also freely accessible from the ioChemBD database,44,45 and the instructions to use this platform have
also been described in the Supporting Information.

3. RESULTS AND DISCUSSION
3.1. Adsorption. Adsorption energies of reactants (A−D)
and products (G0−G10 in Schemes 1 and 2) on Ni(111),
Ru(0001), and Ru-doped Ni(111) surfaces and their solvation
energies in water and EG are listed in Table 1. Adsorption
geometries of these four lignin β-O-4 models are shown in
Figure 2. Isomers A and B with β-S carbons bind to the surface
at least 0.40 eV stronger than β-R configurations do (C and D),
because of the chirality difference at β-C caused by changing
4233

DOI: 10.1021/acscatal.8b00067
ACS Catal. 2018, 8, 4230−4240

ACS Catalysis

Research Article

adsorption energies depend on their motifs above the Ru
atom. Two A structures with an aromatic part and O (from βO-4) above the Ru atom, denoted as ArRu and ORu, respectively,
were optimized, and their adsorption energies are −2.28 and
−2.13 eV, respectively. The local surface environment and
lignin rigidity would account for the weaker ORu adsorptions.
The Ni atoms surrounding Ru have been increased by 0.02 Å,
and Ru is 0.06 Å above the surface. The local surface roughness
pushes the adsorbate up in the z direction slightly, compared to
the flat Ni(111) surface. For example, the carbon atoms from
the benzene ring connecting the O atom shift upward slightly
with a range from 0.003 to 0.098 Å, weakening the aryl−metal
interaction. In addition, the aryl far from the O atom nearly
keeps its structure unaltered, and this rigidity might bring out
the intramolecule strain, further reducing its adsorption
strength.
3.2. Reaction Network. Because of their lower adsorption
energies, C and D were not further considered in the
depolymerization and hydrogenation network. For A and B
conversions on the Ni(111) surface, as reported experimental
results show that aromatic parts are undisturbed during the
catalytic process by using an activated carbon-supported Ni
(Ni/AC) catalyst,17 only reactions related to α- and β-C were
considered, and the full decomposition pathways are outlined
in Schemes 1 and 2 and Figure S3 and S4. Two C−C bonds,
Cα−Caromatic and Cα−Cβ, were assumed to require high barriers
to proceed because of interactions lacking between these bonds
and the surface metals. The reactants can be treated as EG with
its H atoms being replaced by aromatic moieties and CH2OH;
a previous study of EG dehydrogenation shows that the C−C
bond breaks in the late reaction steps.50 In addition, in the
previous experimental studies over the Ni catalyst,17 Caromatic−
O, Cα−Caromatic, and Cα−Cβ bond scissions were not observed
and the Cβ−Cγ bond breaking product accounts for only 1% of
the product distribution. Therefore, aromatic hydrogenation
and Caromatic−O and C−C bond breaking are not further
considered. Calculated reaction barriers and energies from A
and B isomers are also shown in Schemes 1 and 2, and the
energy diagrams and geometries are plotted in Figures 3 and 4.
To describe the complex reaction network more clearly, all

the relative positions of the H and CH2OH groups. On the
contrary, the chirality at α-C has little effect on the adsorption
energies. For instance, isomers A and B adsorb to the surface
with similar exothermic energies of −2.28 and −2.29 eV,
respectively, on Ni(111).
To ensure the stabilities of A and B isomers, we also tried
two geometries derived from the A isomer by rotating the Cα−
Cβ bond and using the other face to adsorb, see Figure 2. For
these alternative structures, the molecule adsorbs on the surface
with only one aromatic unit connecting to the β-O-4 bond,
leaving the other aromatic unit intact. When A adsorbs to the
surface by using the other face, the aryl−surface interactions are
blocked for both sides by the γ-CH2OH group and A adsorbs
on the surface in a tilted way. As the aryl−surface interactions
make the main contributions to the adsorption energy, these
two geometries with only one aryl−surface interaction and
without any aryl−surface interactions have small adsorption
energies of −1.08 and −0.69 eV, respectively, showing the
stereoeffects in the adsorption energies.
There are two important aspects to consider about the
adsorption of very large molecules from the liquid phase. For
molecules like the lignin A model, a sergeant-and-soldier
principle can be applied to understand adsorption.46 The
largest driving force is the adsorption of the aryl groups that act
as directing groups for adsorption; this fact coupled to the
rigidity of the Ar−Cα−Cβ−O4−Ar structure implies that Cβ
stands relatively far from the surface, and this will prevent an
easy activation. This can be seen from the fact that molecular
adsorption for the configurations with two rings on the surface
is double that of only one ring interacting with the catalyst. In
addition, the large adsorption energy arising from the large
molecule sizes of the β-O-4 model would also lead to a large
energy span in the reaction profile. However, adsorption
occupies a large number of sites, and thus, the adsorption
energy per area is more meaningful for assessing the real
strength of the binding at the interface. For instance, hydrogen
has been reported to affect co-adsorption;47 however, in our
case, the reactions are performed in solvent (EG), and thus, a
competition among hydrogen, solvent, and lignin for the active
sites will appear. The corresponding adsorptions are −0.60 eV/
H atom, −0.71 eV for EG, and around −2.28 eV for A, but the
corresponding adsorption energy densities (divided by the
number of metal atoms in the ensembles, Ncovered in Table 1)
are quite similar. In addition, the reactions are performed in
solvent. Solvent contributions are difficult to assess48 and would
require extensive first-principles molecular dynamics sampling,49 which are not possible for the lignin dimers presented
here. Instead, we have followed two methods to report the
contributions: (i) adsorption/desorption events calculated in
the gas phase (rendering large values, around 2 eV for the
model A lignin) and (ii) introduction of the replacement
energy that implies removal of some solvent from the surface
and desolvation of the reactants from our continuum model
(see the Born−Haber cycle in section S3.2). The discussion in
terms of only the competition between the solvent and lignin
model A is sufficient to identify potential bottlenecks.
On the Ru(0001) surface, our results show that adsorptions
for all species are much stronger than those on Ni(111).
Therefore, the adsorption of intermediate species on a Rudecorated Ni(111) surface would be between those on pure
Ni(111) and Ru(0001) surfaces. This has been confirmed by
the calculated adsorption energies on the Ru−Ni(111) surface.
When A and B approach the Ru−Ni(111) surface, the

Figure 3. Reaction profile and intermediates’ structures of isomer A
conversion on the Ni(111) surface. For smaller molecules on a p-(7 ×
4) Ni(111) surface, some of the surface metal atoms are not shown in
the figures. Same color code as in Figure 2.
4234

DOI: 10.1021/acscatal.8b00067
ACS Catal. 2018, 8, 4230−4240

ACS Catalysis

Research Article

For reactions starting from the β-C side, β-C−H (A02) and
direct β-O-4 bond breaking (A04), were first investigated. Our
results show that the A02 step is blocked by its high barrier and
endothermic energy of 1.92 and 1.49 eV, respectively, because
of the rigidity of the lignin structure arising from the strong
interactions between two aromatic units that prevents the
rotation of the Cα−Cβ bond. Hence, the possible reactions after
the β-C−H step have not been considered in the next step. For
direct β-O-4 bond breaking (A04), this step needs a relatively
low barrier of 1.07 eV and is exothermic by −0.66 eV.
Therefore, this step would compete with the C−OH and O−H
bond breakings from the α-C side. One of its products, 2methoxy-phenolate, would be hydrogenated to G0 and then
desorb. For the other product, there are three possible reactions
that can proceed; one is the direct hydrogenation step (A17)
that takes place with a comparable energy of 0.82 (−0.05) eV,
while its product G5 requires 1.62 (0.22) eV to desorb in the
gas phase (or EG solvent). The other two are C−OH bond
breakings from α-C (A15) and β-C (A16) sides with energies
of 0.86 (−0.45) and 0.76 (−0.23) eV, respectively. Among
them, step A16 with the lowest barrier is the most kinetically
favorable. The A15 step produces the coniferyl alcohol (G1),
and the A16 step produces another unsaturated intermediate
(G2) that needs a high energy of 2.07 (0.80) eV to desorb in
the gas phase (or EG solvent). From G2, two hydrogenation
steps proceed on the β-C (A34) and γ-C (A44) sides and the
product G7 needs 1.44 (0.01) eV to desorb.
For the analysis presented above, we noticed that C−OH
and β-O-4 scissions can affect each other when they proceed in
a different sequence order. For example, C−OH bond breaking
in A01 facilitates the subsequent β-O-4 bond rupture (A11) by
forming the conjugated structures. On the other hand, direct βO-4 bond cleavage (A04) facilitates C−OH bond ruptures
(A15 and A16) by ∼0.10 eV, compared with the A01 step.
3.2.2. Reactions Starting from Isomer B. The B isomer can
be regarded as the product of changing the H and OH group
positions (epimers) at α-C in the A isomer, and the reaction
network and profile for its conversion are outlined in Scheme 2
and Figure 4. For reactions starting at the α-C side, only the
C−H bond interacts with the surface; thus, the α-C−H bond
breaking proceeds first with a barrier of 0.67 eV, and this step is
almost thermodynamically neutral, −0.06 eV. Then, β-O-4
bond scission (B11) is greatly promoted with energies of 0.15
(−0.78) eV. This indicates that conjugation again plays a
significant role in reducing the barrier to breaking the β-O-4
bond. In addition, the barrier in B11 is 0.24 eV lower than that
in the A11 step (0.39 eV). The differences between initial
structures of A11 and B11 are the H and OH groups at the α-C
side. As oxygen has a higher electron affinity, more electrons
would be transferred from the β-O-4 bond to the α-C−OH
bond and the β-O-4 bond in B11 is more activated than that in
A11, resulting in a lower barrier. 2-Methoxy-phenolate and G3
are the products in this step. From G3, two sequential C
hydrogenation steps (B33 and B43) proceed with barriers of
0.85 (0.55) and 1.14 (−0.03) eV, respectively, and G8 is the
product that needs less energy (1.20 eV) to desorb. The
solvent-corrected adsorption energy of G8 is 0.19 eV; this
means that it will desorb once it is formed.
For reactions from the β-C side, similar to those from A, βC−H bond rupture (B02) again requires the highest barrier
among all reactions, showing that the change in chirality from
that of α-C has little effect on reducing the structure’s rigidity.
Thus, reactions after the B02 step are no longer considered. For

Figure 4. Reaction profile and intermediate structures of isomer B
conversion on the Ni(111) surface. Same color code as Figure 2.

steps in the schemes are labeled with notations such as A01,
A11, etc. Here, A01 stands for the first possible elementary
reaction in the first reaction stage from the A isomer. Because
of the different chirality of A and B isomers, the reactions that
take place at the same α- or β-C positions at the same stage in
two isomer conversions are labeled with the same number even
though the bond breaking types are different. For instance, A01
and B01 are C−OH and C−H bond breakings, respectively.
C−OH from the B isomer is not labeled as B01 because this
bond is not activated in the adsorbed B isomer.
3.2.1. Reactions Starting from Isomer A. The reaction
network and the energy profile starting from isomer A are
shown in Scheme 1 and Figure 3. For reactions from the α-C
side of this isomer, there are two competing reactions, C−OH
(A01) and O−H (A03) bond ruptures. The barriers (reaction
energies) for them are 0.87 (−0.36) and 1.00 (0.38) eV,
respectively. After the A03 step, α-C−O breaking (A12)
proceeds with lowest barriers of 0.38 eV in three possible steps
(A12−A14), and this step is highly exothermic with an energy
of −1.04 eV. As a result, a common intermediate appears from
the A01 step or the combined A03 and A12 steps. Then, it is
interesting to find that the β-O-4 bond is very easy to break
(A11) both kinetically and thermodynamically with energies of
0.39 (−0.82) eV. Such a lower barrier for β-O-4 bond rupture is
attributed to the formation of a highly conjugated coniferyl
alcohol (G1). After A01, another α-C hydrogenation step
(A22) to produce E has been calculated, and this step has an
activation barrier of 0.85 eV and is nearly thermoneutral by
0.06 eV. The inverse reaction of A01 is difficult because of its
higher barrier of 1.23 eV. Hence, once the C−OH bond breaks,
both the hydrogenation step (A22) and the reverse reaction of
A01 cannot compete with β-O-4 bond breakings. Another
product of the β-O-4 bond cleavage is 2-methoxy-phenolate,
which requires 0.82 eV (A21) to be hydrogenated into 2methoxy-phenol (G0). The desorption energies for G0 and G1
are 1.24 and 2.34 eV, respectively. When solvation effects are
considered, their desorption energies decrease to 0.04 and 0.90
eV, respectively, in the EG solvent. Compared to the higher
barrier to desorb for G1, two hydrogenation steps at the CC
bond (A33 and A43) would take place, and the hydrogenation
product is G6. In the last step, G6 would desorb in the gas
phase (EG solvent) with a lower barrier of 1.57 (0.26) eV.
4235

DOI: 10.1021/acscatal.8b00067
ACS Catal. 2018, 8, 4230−4240

ACS Catalysis

Research Article

direct β-O-4 bond breaking (B04), the activation barrier (1.15
eV) is 0.08 eV higher than that in the A04 step. One product is
2-methoxy-phenolate, and the other one would undergo three
possible reactions, α-C−H (B15) and γ-C−OH bond breakings
(B16) and β-C hydrogenation (B17), the products being G3,
G4, and G8, respectively. The energies in the B15−B17 steps
are 0.89 (−0.33), 0.90 (−0.42), and 0.77 (0.15) eV,
respectively. From G3, two hydrogenation steps would
proceed, and the final product is G8 as stated before. From
G4, both β-C (B34) and γ-C (B35) hydrogenations take place
with energies of 0.84 (0.40) and 0.72 (0.24) eV, respectively. In
the next steps, the energies are 0.88 (−0.34) and 0.71 (−0.18)
eV for the B44 and B45 steps, respectively. The hydrogenation
product G9 needs 1.29 (0.01) eV to desorb in the gas phase (or
EG solvent). Considering the energies in the three initial steps
(B01−B03), it is safe to infer that the reaction proceeds via αC−H scission first, followed by the β-O-4 bond cleavage and
hydrogenation steps. The products should be G0 and G8.
3.2.3. Comparison of the Reactivity for A and B Isomers.
As described above, the most preferred reaction pathways from
A and B isomers on the Ni(111) surface have been selected and
the energy diagrams are shown in Figure 5. G0, G6, G7, etc., in

been considered by analyzing the vibration modes under 473 K,
which is the experiment temperature from ref 17. As shown in
Table S7, entropy contributions are very small for the surface
reactions, and this has also been shown in previous studies by
Hamou et al.22 For reactions related to the A isomer, A01, A33,
and B43 have very close barriers, and it is difficult to tell which
one is the rate-limiting step. However, for reactions related to
the B isomer, B43 has the highest barrier of 1.03 eV. This again
implies the chirality effects in the reaction pathways.
3.2.4. Reactions on the Ru-Doped Ni(111) Surface. A
previous experimental study by Yan et al.18 demonstrated that
bimetallic Ru−Ni exhibits superior activity during the hydrogenolysis of lignin in water. By combination of XPS and X-ray
absorption spectroscopy methods, it was found that the catalyst
is surfaced-enriched with Ni. In the study presented here, the
Ru effects were studied via a simple model by replacing one
surface Ni with a Ru atom (impurity model), and then the
adsorption properties and transition states were optimized.
Reaction barriers and energies on Ni(111) and Ru−Ni(111)
surfaces are listed in Table 2, and effects of Ru on the energy
profiles for the selected pathways are plotted in Figure S8.
Table 2. Comparison of Activation Barriers, Ea, and Reaction
Energies, ΔE (both in electronvolts), on Pure Ni(111) and
Ru-Doped Ni(111) Surfaces
Ni(111)

Ru−Ni(111)

label

Ea

ΔE

Ea

ΔE

A01
A04
A11
A15
A16
A21
B01
B04
B11
B15
B16

Figure 5. Comparison between the reaction energy profiles for A and
B conversions on the Ni(111) surface.

type
α-C−OH
β-O-4
β-O-4
α-C−OH
β-C−OH
O4+H
α-C−H
β-O-4
β-O-4
α-C−H
γ-C−OH

0.87
1.07
0.39
0.86
0.76
0.82
0.67
1.15
0.15
0.89
0.90

−0.36
−0.66
−0.82
−0.45
−0.23
0.37
−0.06
−0.40
−0.78
−0.33
−0.42

0.78
0.77
0.29
1.01
0.72
1.06
0.74
0.96
0.05
0.56
0.87

−0.12
−0.39
−1.20
−0.39
−0.04
0.74
−0.09
−0.18
−1.18
−0.68
−0.20

When α-C−OH bond cleaves (A01) over the substituted Ru
atom, the barrier decreases slightly by 0.09 eV, and the process
is less exothermic by 0.24 eV. Large differences are observed in
direct β-O-4 bond scission (A04) with the barrier decreasing
from 1.07 to 0.77 eV, and the reaction is less exothermic by
0.28 eV. Ru has little effect on the barrier of β-O-4 in the
second step (A11) but a strong effect on the reaction energy
(from −0.82 to −1.20 eV). The α-C−OH breaking step (A15)
after A04 needs 0.15 eV more energy to proceed on the Ru−
Ni(111) surface, and the reaction energy decreases slightly by
0.06 eV. The barrier of another α-C−OH scission step (A16)
changes slightly by 0.04 eV, and the process turns out to be
thermoneutral. Furthermore, for hydrogenation of 2-methoxyphenolate to G0 (A21), Ru also increases the barrier by 0.24 eV
and the reaction becomes more endothermic from 0.37 to 0.74
eV.
For the reactions of B, Ru also decreases the barrier by 0.16
eV for direct β-O-4 bond scission (B01). A larger release of
energy from the second β-O-4 bond (B04) has been observed
again, in agreement with reaction from A. It is interesting to
observe that for B15, the C−H bond breaking after the B04
step is also activated by 0.33 eV and is thermodynamically more
favorable by −0.68 eV. However, A15 becomes more difficult

the profiles stand for their desorption steps. For isomer A, the
pathway starting from the α-C side is A01−A11−A21−A33−
A43−G6−G0, with sequential reaction barriers of 0.87, 0.39,
0.82, 0.82, 0.87, 1.57, and 1.24 eV. Similarly, for the other path
starting from the β-C side, the sequential barriers (A04−A16−
A21−A34−A44−G7−G0) are 1.07, 0.76, 0.82, 0.84, 0.99, 1.44,
and 1.24 eV. In the last two steps, product desorptions require
more energy. However, when solvation effects are considered,
the desorption barriers are diminished. For instance, in the EG
solvent, the barriers of G6 and G7 desorptions decrease from
1.57 and 1.44 eV to 0.24 and 0.20 eV, respectively. For the B
isomer, the barriers in the pathway starting from the α-C side
(B01−B11−A21−B33−B43−G8−G0) are 0.67, 0.15, 0.82,
0.85, 1.14, 1.20, and 1.24 eV, respectively. The barriers in the
pathway starting from the β-C side (B04−B16−A21−B35−
B45−G9−G0) are 1.15, 0.90, 0.82, 0.72, 0.71, 1.29, and 1.24
eV. Thus, the reactions starting from the α-C side are
kinetically preferred. For elementary reactions in two pathways
staring from the α-C side for A and B isomers, entropy and
zero-point energy corrections to their activation barriers have
4236

DOI: 10.1021/acscatal.8b00067
ACS Catal. 2018, 8, 4230−4240

ACS Catalysis

Research Article

structures on the Ni(111) surface. In fact, the rigidity inhibits
the β-C−H bond breakings as the A02 and B02 steps need to
overcome extremely large barriers of 1.92 and 1.87 eV,
respectively. The α-C−H bond breaking barrier in the B01
step is only 0.67 eV. As there are two faces of the isomers, we
also tried to put the other face of the A isomer on the surface.
However, because of the stereoeffects from the γ-C side, the
interactions between aromatic rings and the surface are
blocked, which results in the rather weak adsorption.
Second, for chirality effects, the differences from the α-C side
on the adsorption energies are small, −2.25 and −2.27 eV for A
and B, respectively. However, this is not the case for the
reaction networks. From the reaction network discussed before,
the reactions starting from α-C and β-C sides proceed
differently for both A and B isomers, and the pathway starting
at α-C is more preferable than those from the β-C side. As
shown in the two reaction schemes, changing the chirality at
the α-C side (from A to B) would also lead to different product
distributions. For example, C−OH bond scission from the α-C
side in the B isomer is unlikely because the lack of interaction
of C−OH with the surface leads to the unactivated C−OH
bond. Thus, the OH group on the α-C side of reactant B
remains during the reaction network. It should be emphasized
here that the various products from A and B cannot be
observed when the simple β-O-4 surrogates are used in the
study. From the comparison between reaction barriers of the
A01 and B01 steps, it is clear that C−H bond breaking with a
barrier of 0.66 eV proceeds easier than C−OH bond breaking,
which has barrier of 0.86 eV. Because of the lower barriers in
the first two steps (C−H and β-O-4 scissions), lignin
decomposition from reactant B would proceed much faster
than that from reactant A, demonstrating that stereocenters
affect the relative rates.
Third, for both A and B conversions, the formation of
conjugation structures would significantly reduce the reaction
barriers for β-O-4 bond breaking. For example, direct bond
breakings in A04 and B04 need barriers of 1.04 and 1.15 eV,
respectively, and the barriers decreased to 0.39 and 0.11 eV for
the A11 and B11 steps, respectively. This is an indication of a
useful strategy for breaking the β-O-4 bond by breaking one
bond on the α-C side ahead and allowing the formation of
conjugation structures.
3.3.2. Solvation Effects. In the reaction profiles of A and B
conversions (Figure 5), a large energy span from the lowest
structures to the products exists (∼3.0 eV). This is mainly due
to the large size of the molecule when the gas phase and pure
slab energies are used as reference energies. Under real reaction
conditions, reactants and products are surrounded by the
solvent molecules and solvation energies will decrease the initial
energy level (gas phase and the pure slab surface) on a large
scale. On one hand, by using the MGCM method,40 the
computed solvation energies of A and B are −0.60 eV in EG.
This means their adsorption energy would approximately be
reduced to −1.65 and −1.67 eV, respectively. On the other
hand, the surface is also covered by the solvent molecules; if
adsorption were to proceed, the surface solvent species would
be pushed back into the solvent. Therefore, the energy to
remove surface solvent molecules should be paid by the
adsorption energy of these adsorbates. The energy for this
whole replacement process can then be written as
Eads(EG).41,51 For example, for the adsorbed A isomer and
EG on the Ni(111) surface, they roughly cover 20 and 4.5 Ni
atoms, respectively. The monolayer-covered Ni(111) surface

to reach as the energy requirement increases by 0.15 eV on the
Ru−Ni(111) surface. Compared to those of A01 and A15, the
lower barrier in the first C−H breaking step (B01) due to its
chirality at α-C and the barrier lowered by the doped Ru in the
second step (B15) demonstrate the synergistic effects of
stereocenters and Ru on the promotion of lignin conversion.
For the C−OH bond in B16, the barrier is slightly reduced
from 0.90 to 0.87 eV and the reaction is less exothermic by 0.23
eV, which is similar to the A situation. Thus, for C−OH bond
breaking, Ru has weak (for A16 and B16) or negative (for A15)
effects on the reaction barriers and reactions become more
endothermic by 0.20−0.30 eV. For β-O-4 bond breaking in the
first steps (A04 and B04), doped Ru facilitates the reaction by
reducing the barrier by ≤0.30 eV, with reactions becoming less
exothermic by 0.20 eV. For β-O-4 bond breaking in the second
steps (A11 and B11), Ru has little effect on the barriers but the
reaction becomes more exothermic by 0.40 eV. In the final step,
the introduced Ru increases the barrier and energy for
phenolate hydrogenation into phenol by 0.27 and 0.35 eV,
respectively. This can be attributed to the O affinity of Ru being
stronger than that of Ni, which stabilizes the phenolate species.
In spite of that, this step can be promoted by the high surface H
concentration and H2O via the formation of hydrogen bonds.
3.3. Discussion. 3.3.1. Chirality, Rigidity, and Conjugation Effects. As stated above, three stereoeffects were found in
this study: rigidity, chirality, and conjugation. First, the most
stable adsorption geometries would be affected by the rigidity
of the dimer molecules. For example, when the H and CH2OH
group positions are switched at the β-C side in A and B
isomers, the products would bind to the surface more weakly by
∼0.50 eV (see Table 1). In addition, when the Cα−Cβ bond
rotates, because of the increased strain in the structures, the
adsorption energy decreases significantly (Figure 2). To further
clarify the effects of rigidity on the adsorption energies, the
aromatic units connecting to α-C were replaced by H atoms
from A and D isomers to eliminate the rigidity (see Figure 6).

Figure 6. Adsorption of rigidity-eliminated isomers from A and D by
substitution of the aromatic units bound to α-C with H atoms on the
Ni(111) surface.

Calculated adsorption energies are −1.42 and −1.58 eV,
respectively. It should be emphasized here that the adsorption
energies of A and D on the Ni(111) surface are −2.28 and
−1.86 eV, respectively. Thus, the replaced D isomer becomes
more stable upon removal of the structural rigidity. The rigidity
also has an impact on the reactions from α- and β-C sides. A
previous study that used 2-phenoxyethanol as a surrogate22
showed that the α-C−H and β-C−H bond breakings have
barriers of 1.21 and 1.14 eV, respectively, on Pt(111). However,
these very similar barriers were not observed in the A and B
4237

DOI: 10.1021/acscatal.8b00067
ACS Catal. 2018, 8, 4230−4240

ACS Catalysis

Research Article

was optimized by adding up to eight EG molecules on a p-(6 ×
6) supercell (see Figure S7); 20/4.5 surface EG molecules
would be substituted by only one A isomer. Thus, the
replacement on the surface in the solvent goes as
Asolvated +
solvated

20
20 solvated
EG* → A* +
EG
4.5
4.5

in return weakens the β-O-4 linkage and promotes C−O bond
cleavage.
In addition, adsorption energies are more exothermic on
Ru(0001) than on the Ni(111) surface (Table 1). Thus,
employing too much Ru on the surface would lead to surface
poisoning by impeding the refreshing of the catalyst surface
during the desorption step of the catalytic cycle. In addition, it
has to be emphasized that Ru also has a strong ability to
hydrogenate arenes.52 Therefore, a small percentage of Ru
doped into the Ni catalyst would improve the catalytic process.
The structure of this active site then recalls single atom
catalysts.Our results are in excellent agreement with the
experimental observations of Yan et al.,18 who found a
volcano-like behavior exists upon variation of the Ru
percentages in the Ni-based catalyst.

(1)

solvated

where A
and EG
are species in the solvation and A*
and EG* are surface-adsorbed species. The adsorption energy
in this step can be obtained via a Born−Harber cycle as shown
in Figure S6 and its corresponding eq S8. For all the reactants
and products, their adsorption energies are reduced to a large
degree when solvation effects are employed, and this also
decreases the large energy span in the energy profiles. The
A01−A11−A21−A33−A43−G6−G0 pathway on Ni(111) is
taken as an example to illustrate solvation effects. As shown in
Figure 7, when the solvation energy is corrected, the large
energy span when taking the gas phase energies as references is
reduced to more representative values.

4. CONCLUSIONS
In this study, DFT calculations of complex lignin dimer β-O-4
bond models over Ni(111) and Ru-doped surfaces were
performed. The reaction network accompanied by the
stereoeffects and Ru doping effects has been investigated.
Our results show that stereocenters have three effects: on the
adsorption structures, on the reaction pathways in the network,
and on product distributions. The chirality at the β-C side and
the rigidity are important factors for the different adsorption
energies of four reactant candidates and for the reactions from
both α-C and β-C sides, thus determining the preferable
reaction path from A and B isomers. For the favorable reactions
from both A and B, it is chirality at the α-C side that leads to
the different kinetic results and product distributions. Rigidity
implies that adsorption becomes dominated by the large aryl
rings, in the form of the sergeants-and-soldiers principle. In
addition, it blocks the reactions from the β-C side at the
beginning of the reaction networks. Computed barriers and
energies imply that reactions from the B isomer should be faster
than those from the A isomer. We also found that β-O-4 bond
cleavage would be promoted or accelerated by the unsaturated
α-C side (α-C−H and α-C−OH breaks ahead) because this
would benefit the formation of conjugated products. Ru has
both positive and negative effects on the catalytic behavior. It
promotes direct β-O-4 bond scission by lowering the reaction
barriers via the strengthened metal−O bond and makes it
competitive with the first C−H or C−OH bond scissions. As
Ru has a stronger adsorption toward reactants than Ni, too
much Ru would lead to less active surface sites by covering the
intermediates and impede catalysts. Our DFT study rationalizes
the complex reaction network of the most abundant β-O-4
linkage in lignin, showing the role of stereocenters, rigidity, and
conjugation in lignin depolymerization on nickel and Ru-doped
nickel catalysts.

Figure 7. Reaction profiles employing gas phase reference energies
and the corrected solvation energy (replacement energy values) for the
A01−A11−A21−A33−A43−G6−G0 path on the Ni(111) surface.

3.3.3. Ru Doping Effects. As stated in the adsorption section,
on the Ru-doped Ni(111) surface, ArRu binds to the surface
more strongly (−2.28 eV) than ORu does (−2.15 eV).
However, the β-O-4 bond breaking barrier of ArRu is 0.13 eV
higher than that of ORu. This implies that Ru participates in
lowering the C−O bond breaking barriers. To confirm this
assumption, the adsorptions of a single O atom atop Ni and Ru
atoms in pure Ni(111) and Ru-doped Ni(111) surfaces were
calculated. The Ni−O and Ru−O distances are 1.678 and 1.734
Å, respectively, with adsorption energies for the gas phase O2 of
−0.54 and −1.51 eV, respectively. The more negative O
adsorption energy atop Ru indicates that the Ru−O bond is
much stronger than the Ni−O bond. This shows that more
electron density is localized between O and Ru, thus weakening
the β-O-4 bond when A and B adsorb on the surface, in
agreement with the earlier transition state C−O bond length of
2.103 Å on Ru−Ni(111), compared with the longer bond
distance of 2.242 Å on Ni(111). Stronger Ru−O interaction
can also be implied via comparison of the O adsorption
energies on Ru(0001) and Ni(111) (−2.98 versus −2.45 eV).
Therefore, the doped Ru strengthens the metal−O bond, which

■

ASSOCIATED CONTENT

S
* Supporting Information

The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acscatal.8b00067.
Structures and their coordinates, decomposition networks, computed adsorption energies of closed-shell
structures, reaction barriers, reaction energies, energy
profiles, stability calculation details, and equations for
solvation correction (PDF)
4238

DOI: 10.1021/acscatal.8b00067
ACS Catal. 2018, 8, 4230−4240

ACS Catalysis

■

Research Article

AUTHOR INFORMATION

(13) Roberts, V.; Stein, V.; Reiner, T.; Lemonidou, A.; Li, X.;
Lercher, J. A. Towards Quantitative Catalytic Lignin Depolymerization. Chem. Eur. J. 2011, 17, 5939−5948.
(14) Sergeev, A. G.; Hartwig, J. F. Selective, Nickel-Catalyzed
Hydrogenolysis of Aryl Ethers. Science 2011, 332, 439−443.
(15) Chen, L.; Korányi, T. I.; Hensen, E. J. Transition Metal (Ti, Mo,
Nb, W) Nitride Catalysts for Lignin Depolymerisation. Chem.
Commun. 2016, 52, 9375−9378.
(16) Molinari, V.; Clavel, G.; Graglia, M.; Antonietti, M.; Esposito, D.
Mild Continuous Hydrogenolysis of Kraft Lignin over Titanium
Nitride−Nickel Catalyst. ACS Catal. 2016, 6, 1663−1670.
(17) Song, Q.; Wang, F.; Xu, J. Hydrogenolysis of Lignosulfonate
into Phenols over Heterogeneous Nickel Catalysts. Chem. Commun.
2012, 48, 7019−7021.
(18) Zhang, J.; Teo, J.; Chen, X.; Asakura, H.; Tanaka, T.; Teramura,
K.; Yan, N. A Series of NiM (M = Ru, Rh, and Pd) Bimetallic Catalysts
for Effective Lignin Hydrogenolysis in Water. ACS Catal. 2014, 4,
1574−1583.
(19) Zhang, J. W.; Cai, Y.; Lu, G. P.; Cai, C. Facile and Selective
Hydrogenolysis of β-O-4 Linkages in Lignin Catalyzed by Pd−Ni
Bimetallic Nanoparticles Supported on ZrO2. Green Chem. 2016, 18,
6229−6235.
(20) Lu, J.; Wang, M.; Zhang, X.; Heyden, A.; Wang, F. β-O-4 Bond
Cleavage Mechanism for Lignin Model Compounds over Pd Catalysts
Identified by Combination of First-principles Calculations and
Experiments. ACS Catal. 2016, 6, 5589−5598.
(21) Zhang, C.; Li, H.; Lu, J.; Zhang, X.; MacArthur, K. E.; Heggen,
M.; Wang, F. Promoting Lignin Depolymerization and Restraining the
Condensation via an Oxidation-Hydrogenation Strategy. ACS Catal.
2017, 7, 3419−3429.
(22) Ould Hamou, C. A.; Réocreux, R.; Sautet, P.; Michel, C.; Giorgi,
J. B. Adsorption and Decomposition of a Lignin β-O-4 Linkage Model,
2-Phenoxyethanol, on Pt (111): Combination of Experiments and
First-Principles Calculations. J. Phys. Chem. C 2017, 121, 9889−9900.
(23) Li, Q.; García-Muelas, R.; López, N. Microkinetics of Alcohol
Reforming for H2 Production from a FAIR Density Functional Theory
Database. Nat. Commun. 2018, 9, 526.
(24) Calvo-Flores, F. G.; Dobado, J. A.; Isac-García, J.; MartíMartínez, F. J. Lignin and Lignans as Renewable Raw Materials:
Chemistry, Technology and Applications; John Wiley & Sons, 2015.
(25) He, J.; Zhao, C.; Lercher, J. A. Ni-Catalyzed Cleavage of Aryl
Ethers in the Aqueous Phase. J. Am. Chem. Soc. 2012, 134, 20768−
20775.
(26) Kresse, G.; Furthmü J. Efficiency of ab-initio Total Energy
ller,
Calculations for Metals and Semiconductors Using a Plane-wave Basis
Set. Comput. Mater. Sci. 1996, 6, 15−50.
(27) Kresse, G.; Furthmü J. Efficient Iterative Schemes for ab
ller,
initio Total-energy Calculations Using a Plane-wave Basis Set. Phys.
Rev. B: Condens. Matter Mater. Phys. 1996, 54, 11169−11186.
(28) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient
Approximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865−3868.
(29) Grimme, S. SemiEmpirical GGA-type Density Functional
Constructed with a Long-range Dispersion Correction. J. Comput.
Chem. 2006, 27, 1787−1799.
(30) Buč T.; Hafner, J.; Lebegue, S.; Angyán, J. G. Improved
ko,
Description of the Structure of Molecular and Layered Crystals: Ab
initio DFT Calculations with van der Waals Corrections. J. Phys. Chem.
A 2010, 114, 11814−11824.
(31) Almora-Barrios, N.; Carchini, G.; Błoński, P.; López, N. Costless
Derivation of Dispersion Coefficients for Metal Surfaces. J. Chem.
Theory Comput. 2014, 10, 5002−5009.
(32) Blö P. E. Projector Augmented-wave Method. Phys. Rev. B:
chl,
Condens. Matter Mater. Phys. 1994, 50, 17953−17979.
(33) Kresse, G.; Joubert, D. From Ultrasoft PseudoPotentials to the
Projector Augmented-wave Method. Phys. Rev. B: Condens. Matter
Mater. Phys. 1999, 59, 1758−1775.
(34) Monkhorst, H. J.; Pack, J. D. Special Points for Brillouin-zone
Integrations. Phys. Rev. B 1976, 13, 5188−5192.

Corresponding Authors

*E-mail: qli@iciq.es. Phone: +34 977920237. Fax: +34
977920231.
*E-mail: nlopez@iciq.es.
ORCID

Qiang Li: 0000-0001-5568-2334
Núria López: 0000-0001-9150-5941
Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS

The authors thank MINECO (CTQ2015-68770-R) for
financial support and the Barcelona Supercomputing Centre
(BSC-RES) for providing generous computer resources via
Centro de Supercomputación y Visualización de Madrid,
Universidad Politécnica de Madrid. The authors also appreciate
the kind discussion with Dr. Frank Abild-Pedersen from SLAC.

■

REFERENCES

(1) Perlack, R. D.; Wright, L. L.; Turhollow, A. F.; Graham, R. L.;
Stokes, B. J.; Erbach, D. C. Biomass as Feedstock for a Bioenergy and
Bioproducts Industry: The Technical Feasibility of a Billion-ton
Annual Supply. Technical Report; U.S. Department of Energy:
Washington, DC, 2005.
(2) Zakzeski, J.; Bruijnincx, P. C.; Jongerius, A. L.; Weckhuysen, B.
M. The Catalytic Valorization of Lignin for the Production of
Renewable Chemicals. Chem. Rev. 2010, 110, 3552−3599.
(3) Li, C.; Zhao, X.; Wang, A.; Huber, G. W.; Zhang, T. Catalytic
Transformation of Lignin for the Production of Chemicals and Fuels.
Chem. Rev. 2015, 115, 11559−11624.
(4) Behling, R.; Valange, S.; Chatel, G. Heterogeneous Catalytic
Oxidation for Lignin Valorization into Valuable Chemicals: What
Results? What Limitations? What Trends? Green Chem. 2016, 18,
1839−1854.
(5) Thielemans, W.; Can, E.; Morye, S.; Wool, R. Novel Applications
of Lignin in Composite Materials. J. Appl. Polym. Sci. 2002, 83, 323−
331.
(6) Delidovich, I.; Hausoul, P. J.; Deng, L.; Pfü
tzenreuter, R.; Rose,
M.; Palkovits, R. Alternative Monomers Based on Lignocellulose and
Their Use for Polymer Production. Chem. Rev. 2016, 116, 1540−1599.
(7) Calvo-Flores, F. G.; Dobado, J. A. Lignin as Renewable Raw
Material. ChemSusChem 2010, 3, 1227−1235.
(8) Kai, D.; Tan, M. J.; Chee, P. L.; Chua, Y. K.; Yap, Y. L.; Loh, X. J.
Towards Lignin-Based Functional Materials in a Sustainable World.
Green Chem. 2016, 18, 1175−1200.
(9) Ragauskas, A. J.; Beckham, G. T.; Biddy, M. J.; Chandra, R.;
Chen, F.; Davis, M. F.; Davison, B. H.; Dixon, R. A.; Gilna, P.; Keller,
M.; Langan, P.; Naskar, A. K.; Saddler, J. N.; Tschaplinski, T. J.;
Tuskan, G. A.; Wyman, C. E. Lignin Valorization: Improving Lignin
Processing in the Biorefinery. Science 2014, 344, 1246843.
(10) Alonso, D. M.; Hakim, S. H.; Zhou, S.; Won, W.; Hosseinaei,
O.; Tao, J.; Garcia-Negron, V.; Motagamwala, A. H.; Mellmer, M. A.;
Huang, K.; Houtman, C. J.; Labbé, N.; Harper, D. P.; Maravelias, C.;
Runge, T.; Dumesic, J. A. Increasing the Revenue from Lignocellulosic
Biomass: Maximizing Feedstock Utilization. Sci. Adv. 2017, 3,
No. e1603301.
(11) Schutyser, W.; Renders, T.; Van den Bosch, S.; Koelewijn, S.-F.;
Beckham, G.; Sels, B. F. Chemicals from Lignin: An Interplay of
Lignocellulose Fractionation, Depolymerisation, and Upgrading. Chem.
Soc. Rev. 2018, 47, 852−908.
(12) Chakar, F. S.; Ragauskas, A. J. Review of Current and Future
Softwood Kraft Lignin Process Chemistry. Ind. Crops Prod. 2004, 20,
131−141.
4239

DOI: 10.1021/acscatal.8b00067
ACS Catal. 2018, 8, 4230−4240

ACS Catalysis

Research Article

(35) Lide, D. CRC Handbook of Chemistry and Physics, 84th ed.; CRC
Press LLC: Boca Raton, FL, 2003−2004; pp 12(19-21).
(36) Makov, G.; Payne, M. C. Periodic Boundary Conditions in ab
initio Calculations. Phys. Rev. B: Condens. Matter Mater. Phys. 1995, 51,
4014−4022.
(37) Nørskov, J. K.; Studt, F.; Abild-Pedersen, F.; Bligaard, T.
Fundamental Concepts in Heterogeneous Catalysis; John Wiley & Sons,
2014.
(38) Henkelman, G.; Uberuaga, B. P.; Jónsson, H. A Climbing Image
Nudged Elastic Band Method for Finding Saddle Points and Minimum
Energy Paths. J. Chem. Phys. 2000, 113, 9901−9904.
(39) Heyden, A.; Bell, A. T.; Keil, F. J. Efficient Methods for Finding
Transition States in Chemical Reactions: Comparison of Improved
Dimer Method and Partitioned Rational Function Optimization
Method. J. Chem. Phys. 2005, 123, 224101.
(40) Garcia-Ratés, M.; López, N. Multigrid-Based Methodology for
Implicit Solvation Models in Periodic DFT. J. Chem. Theory Comput.
2016, 12, 1331−1341.
(41) Garcia-Ratés, M.; García-Muelas, R.; López, N. Solvation Effects
on Methanol Decomposition on Pd(111), Pt(111) and Ru(0001). J.
Phys. Chem. C 2017, 121, 13803−13809.
(42) López, N.; Vargas-Fuentes, C. Promoters in the Hydrogenation
of Alkynes in Mixtures: Insights from Density Functional Theory.
Chem. Commun. 2012, 48, 1379−1391.
(43) Soto-Verdugo, V.; Metiu, H. Segregation at the Surface of an
Au/Pd Alloy Exposed to CO. Surf. Sci. 2007, 601, 5332−5339.
(44) Á lvarez-Moreno, M.; De Graaf, C.; López, N.; Maseras, F.;
Poblet, J. M.; Bo, C. Managing the Computational Chemistry Big Data
Problem: The ioChem-BD Platform. J. Chem. Inf. Model. 2015, 55,
95−103.
(45) Li, Q.; López, N. Dataset for lignin decomposition on Ni and
Ru-Ni. DOI: 10.19061/iochem-bd-1-52.
(46) Green, M. M.; Reidy, M. P.; Johnson, R. D.; Darling, G.;
O’Leary, D. J.; Willson, G. Macromolecular StereoChemistry: The
Out-of-proportion Influence of Optically Active Comonomers on the
Conformational Characteristics of Polyisocyanates. The Sergeants and
Soldiers Experiment. J. Am. Chem. Soc. 1989, 111, 6452−6454.
(47) Wang, S.; Vorotnikov, V.; Vlachos, D. G. Coverage-induced
Conformational Effects on Activity and Selectivity: Hydrogenation and
Decarbonylation of Furfural on Pd(111). ACS Catal. 2015, 5, 104−
112.
(48) Besora, M.; Vidossich, P.; Lledós, A.; Ujaque, G.; Maseras, F.
Calculation of Reaction Free Energies in Solution: A Comparison of
Current Approaches. J. Phys. Chem. A 2018, 122, 1392−1399.
(49) Mellmer, M. A.; Sanpitakseree, C.; Demir, B.; Bai, P.; Ma, K.;
Neurock, M.; Dumesic, J. A. Solvent-enabled Control of Reactivity for
Liquid-phase Reactions of Biomass-derived Compounds. Nat. Catal.
2018, 1, 199−207.
(50) Salciccioli, M.; Vlachos, D. G. Kinetic Modeling of Pt Catalyzed
and Computation-driven Catalyst Discovery for Ethylene Glycol
Decomposition. ACS Catal. 2011, 1, 1246−1256.
(51) Bellarosa, L.; García-Muelas, R.; Revilla-López, G.; López, N.
Diversity at the Water−Metal Interface: Metal, Water Thickness, and
Confinement Effects. ACS Cent. Sci. 2016, 2, 109−116.
(52) Dwivedi, A. D.; Rai, R. K.; Gupta, K.; Singh, S. K. Catalytic
Hydrogenation of Arenes in Water over in situ Generated Ruthenium
Nanoparticles Immobilized on Carbon. ChemCatChem 2017, 9, 1930−
1938.

4240

DOI: 10.1021/acscatal.8b00067
ACS Catal. 2018, 8, 4230−4240