MOF-derived hcp-Co nanoparticles encapsulated in ultrathin graphene for carboxylic acids hydrogenation to alcohols

Article abstract of DOI:10.1016/j.jcat.2021.05.013

Highly efficient conversion of carboxylic acids to valuable alcohols is a great challenge for easily corroded non-noble metal catalysts. Here, a series of few-layer graphene encapsulated metastable hexagonal closed-packed (hcp) Co nanoparticles were fabricated by reductive pyrolysis of metal-organic framework precursor. The sample pyrolyzed at 400 °C (hcp-Co@G400) presented outstanding performance and stability for converting a variety of functional carboxylic acids and its turnover frequency was one magnitude higher than that of conventional facc-centered cubic (fcc) Co catalysts. In situ DRIFTS spectroscopy of model reaction acetic acid hydrogenation and DFT calculation results confirm that carboxylic acid initially undergoes dehydroxylation to RCH₂CO* followed by consecutive hydrogenation to RCH₂CH₂OH through RCH₂COH*. Acetic acid prefers to vertically adsorb at hcp-Co (0 0 2) facet with a much lower adsorption energy than parallel adsorption at fcc-Co (1 1 1) surface, which plays a key role in decreasing the activation barrier of the rate-determining step of acetic acid dehydroxylation.

Full text of DOI:10.1016/j.jcat.2021.05.013

Journal of Catalysis 399 (2021) 201–211
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
MOF-derived hcp-Co nanoparticles encapsulated in ultrathin graphene
for carboxylic acids hydrogenation to alcohols
a
a,
Xiaoqing Gao a,b, Shanhui Zhu a, , Mei Dong , Weibin Fan
⇑
⇑
a State Key Laboratory of Coal Conversion, Institute of Coal Chemistry, Chinese Academy of Sciences, Taiyuan 030001, PR China
b University of Chinese Academy of Sciences, Beijing 100049, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
Highly efficient conversion of carboxylic acids to valuable alcohols is a great challenge for easily corroded
non-noble metal catalysts. Here, a series of few-layer graphene encapsulated metastable hexagonal
closed-packed (hcp) Co nanoparticles were fabricated by reductive pyrolysis of metal-organic framework
precursor. The sample pyrolyzed at 400 °C (hcp-Co@G400) presented outstanding performance and sta-
bility for converting a variety of functional carboxylic acids and its turnover frequency was one magni-
tude higher than that of conventional facc-centered cubic (fcc) Co catalysts. In situ DRIFTS
spectroscopy of model reaction acetic acid hydrogenation and DFT calculation results confirm that car-
boxylic acid initially undergoes dehydroxylation to RCH2CO* followed by consecutive hydrogenation to
RCH2CH2OH through RCH2COH*. Acetic acid prefers to vertically adsorb at hcp-Co (002) facet with a
much lower adsorption energy than parallel adsorption at fcc-Co (111) surface, which plays a key role
in decreasing the activation barrier of the rate-determining step of acetic acid dehydroxylation.
Ó 2021 Elsevier Inc. All rights reserved.
Received 6 March 2021
Revised 1 May 2021
Accepted 7 May 2021
Available online 19 May 2021
Keywords:
Carboxylic acid
Hydrogenation
Co
Facet effect
DFT calculation
1. Introduction
challenge to effectively perform hydrogenation of carboxylic acids
by cheaper, more abundant and less toxic base metals (such as Co,
With the rapid development of biomass conversion in recent
years, numerous bio-based carboxylic acids, such as levulinic acid
(LA), succinic acid, lactic acid and itaconic acid have been emerged
as versatile platform chemicals [1,2]. Reduction of carboxylic acids
to alcohols, particularly for diols, is industrially important because
the generated alcohols are intermediates for a variety of polyesters,
pharmaceuticals, agrochemicals and fine chemicals [3,4]. e.g. 1,4-
butanediol is an important commodity chemical to produce valu-
able polymers over 2.5 million tons annually, and it is currently
manufactured through the feedstocks from fossil resources [5,6].
Thus, the development of a green and cost-effective approach is
highly desirable to synthesize alcohols via the reduction of
biomass-derived carboxylic acid.
Currently, the production of alcohols from carboxylic acids has
been performed using stoichiometric quantities of strong reducing
agents including toxic metal hydrides or silanes [7], which inevita-
bly results in serious environmental issues. In recent years, much
progress has been made in the use of homogenous Ru and Ir or
heterogeneous Ru, Rh, Pd and Pt-based catalysts for carboxylic
Fe, Ni and Cu). Compared to ketones and esters, carboxylic acids
are difficult to reduce by H2 owing to the low electrophilicity of
carbonyl carbon [7]. Additionally, the strong acidity of carboxylic
acids can seriously corrode and damage base metals, leading to
detrimental catalyst deactivation. Very recently, Bruin et al. [13]
have reported the first earth-abundant homogenous Co(BF4)2ꢀ6H2-
O catalyst capable of hydrogenating a series of functionalized car-
boxylic acids to alcohols with superior activity. Encouragingly, it is
anticipated to design a highly active heterogeneous Co-based cat-
alyst to overcome strong acidity and inert carboxyl groups.
Metal-organic frameworks (MOFs) are usually consisted of base
metal and organic ligands with porous and tuneable structure.
Recently, MOFs have emerged as promising self-sacrificing tem-
plates for synthesizing some metals nanoparticles (NPs), oxides,
carbon or their composites via simple pyrolysis [14–20], enabling
to tailor their compositions, shapes and phase structures by tuning
MOFs sources, preparation conditions or postsynthetic modifica-
tions. For example, Beller et al. [18] have fabricated Co and Co3O4
composite NPs encapsulated by graphitic shell via the pyrolysis
of cobalt-diamine-dicarboxylic acid MOFs under inert Ar atmo-
sphere, which are broadly effective in reductive amination. Addi-
tionally, more complex MOFs-related Co-based catalysts have
been created, such as Co-CoO@N-doped porous carbon [21], CoS2
acids hydrogenation [8–12]. Nevertheless, it is still
a grand
⇑
Corresponding authors.
E-mail addresses: zhushanhui@sxicc.ac.cn (S. Zhu), fanwb@sxicc.ac.cn (W. Fan).
https://doi.org/10.1016/j.jcat.2021.05.013
0021-9517/Ó 2021 Elsevier Inc. All rights reserved.

X. Gao, S. Zhu, M. Dong et al.
Journal of Catalysis 399 (2021) 201–211
nanobubble hollow prisms [22], or Co3O4@NiCo2O4 double-shelled
nanocages [23], which may offer us potential opportunities to fab-
ricate highly stable Co NPs to tolerate carboxylic acid.
The Co contents were measured by inductively coupled plasma
atomic emission spectroscopy (ICP-AES) apparatus (Optima
2100DV, PerkinElmer). Prior to the measurements, the samples
were heated at 700 °C for 2 h in air, and treated with hydrochloric
acid at 120 °C for 12 h.
Powder X-ray diffraction (XRD) patterns and in situ XRD were
conducted on a D8/max-X-ray diffractometer (Brucker, Germany)
operating with Cu Khcp radiation. In situ XRD patterns of Co3[Co
(CN)6]2 were recorded at different temperature and time with a
heating rate of 5 °C/min in flowing 10 vol% H2/Ar gas.
In situ diffuse reflectance ultraviolet–visible (UV–vis) spectra
were performed at Agilent Cary 5000 UV–Vis-NIR spectropho-
tometer equipped with a poly(tetrafluoroethylene) integrating
sphere. During the measurements, Co3[Co(CN)6]2 was heated to
400 °C with a heating rate of 5 °C/min and kept at this temperature
for 6 h in flowing 10 vol% H2/Ar gas.
Raman spectroscopy was carried out on a Renishaw-UV–vis
Raman System 1000 instrument using a CCD detector at room tem-
perature. The 532 nm of Nd:YAG laser was employed as the excita-
tion source with a power of 30 MW.
Transmission electron microscopy (TEM) and high resolution
TEM (HRTEM) measurements were taken on a field-emission
transmission electron microscopy (FETEM, JEM-2011F) operating
at 200 kV voltages.
The catalytic performances of metal NPs are highly dependent
on their crystal structure, so-called structure sensitivity [24,25].
Metal Co possesses two crystallographic structures, i.e. metastable
hexagonal closed-packed (hcp) phase and thermodynamically
stable facc-centered cubic (fcc) phase. The hcp and fcc Co have dif-
ferent bulk symmetries and structure, and thus lead to expose dis-
tinct facets, which directly influences catalytic performance. Using
conventional approaches such as impregnation, precipitation or
sol-gel, the synthesized Co-based catalysts are composed of fcc
Co or their mixed phases after the reduction of their precursors
Co3O4, CoO or spinel by H2 [26,27]. Although many researchers
argue that hcp Co is more active than fcc phase in their mixed cat-
alysts, such as well-known Fischer-Tropsch synthesis, the intrinsic
active phase is still in great debate for decades owing to the signif-
icant challenge for the synthesis of single hcp Co phase [25,28,29].
In this context, we made a try to pyrolyze MOF precursor of
Co3[Co(CN)6]2 in reductive atmosphere. Unexpectedly, the single
phase hcp-Co NPs were obtained and encapsulated by few-layer
graphene. The as-prepared hcp-Co exhibited enhanced perfor-
mance in one magnitude in comparison to fcc-Co in carboxylic acid
hydrogenation. Combined with in situ diffuse reflection Fourier
transform infrared spectra (DRIFTS) model reaction of acetic acid
hydrogenation and density functional theory (DFT) calculation
results, the reaction mechanism of carboxylic acid hydrogenation
have been disclosed, in which dehydroxylation is identified as a
key step in the formation of alcohol.
X-ray photoelectron spectroscopy (XPS) was measured on a VG
MiltiLab 2000 spectrometer with Mg K
a radiation and a multi-
channel detector. The binding energies were calibrated using the
C1s peak at 284.5 eV. The reduced catalyst is obtained by in situ
treating the passivated sample at 300 °C for 2 h in a 10 vol% H2/
Ar flow.
Co K-edge X-ray absorption spectroscopy (XAS) was performed
in fluorescence mode at the beam line of Beijing Synchrotron Radi-
ation Facility (BSRF), Institute of High Energy Physics (IHEP), Chi-
nese Academy of Science (CAS). XAS spectra was processed in
Athena XAS version 0.8.056 software, in which background sub-
traction and normalization were fitted linear polynomials to the
pre-edge and post-edge regions of absorption spectra, respectively.
The E0 value was calculated based on the maximum value in the
first derivative of the edge region.
In situ DRIFTS was carried out in a FTIR spectrophotometer
(Nicolet Is10) operating with a SMART collector and an MCT detec-
tor cooled in liquid N2. 40 mg catalyst was initially pretreated at
300 °C for 2 h in H2/Ar gas (20 mL/min), and cooled down to room
temperature in Ar gas. Subsequently, acetic acid was introduced
into the catalyst surface and gradually reached in saturated state,
which was monitored by DRIFTS spectra. With increasing temper-
ature to 220 °C, the adsorbed acetic acid began to desorb from the
catalyst. Once the DRIFTS spectra of adsorbed acetic acid kept
stable, H2 (20 mL/min) was introduced into the system and the
DRIFTS spectra were simultaneously collected.
2. Experimental section
2.1. Catalyst preparation
0.04 mmol K3[Co(CN)6] (J&K Scientific Ltd.) and 0.3
g
polyvinylpyrrolidone (PVP, Aladdin) were first dispersed in 10 mL
water under vigorous stirring. Then, 10 mL 0.0075 mol/L Co(NO3)2-
ꢀ6H2O (Aladdin) aqueous solution was dropwisely added at room
temperature. After stirring for 10 min, the mixed solution was
transferred into Teflon-lined stainless steel autoclave and kept at
room temperature for 24 h. The solid was collected by centrifuga-
tion, washed with water and ethanol, and dried at 80 °C overnight.
Finally, these collected solids were pyrolyzed at 300, 400, 500, 600
or 700 °C in 10 vol% H2/Ar gas for 6 h, and passivated in in 1 vol%
O2/Ar gas at room temperature for 2 h. The as-prepared catalysts
were designated as hcp-Co@GT, where T refers to annealing
temperature.
G400 was synthesized by the treatment of hcp-Co@G400 with
98% sulfur acid at 180 °C for 24 h, followed by centrifugation,
washing and drying at 80 °C overnight.
The fcc-Co@G400 catalyst was prepared as follows: certain
amounts of Co(NO3)2ꢀ6H2O (Sinopharm Chemical Reagent Co.,
Ltd., China (SCRC)), 0.3 g PVP, and 0.8 g G400 were successively
added into 10 mL water under vigorous stirring for 1 h. The col-
lected solid was separated by centrifugation, dried and annealed
in 10 vol% H2/Ar at 400 °C for 6 h. The Cu@G400, Ni@G400 or
Fe@G400 catalysts were prepared with the same method as above.
2.3. Catalytic tests
Before the reaction test, the catalyst was activated in 10 vol%
H2/Ar at 20 mL/min in 300 °C for 2 h. Unless otherwise specified,
1 mmol substrate, 0.05 g catalyst and 5 mL H2O were charged into
a stainless steel autoclave (NS-10316L, Anhui Kemi Machinery
Technology Co., Ltd.). The reactor was sequentially purged with
H2 for 5 times, pressured to 5.0 MPa H2, heated into 220 °C, and
kept at this temperature for 24 h. After the reaction, the reaction
system was quickly cooled to room temperature in an ice-water
bath. After separation, the liquid products were analyzed by gas
chromatography (Shimadzu GC-2010) with a flame ionization
detector. 1,2-butanediol was used as an internal standard for the
formation of diol, while 1-butanol was applied in the case of mono-
hydric alcohol. The insoluble reactants and products in water were
2.2. Catalyst characterization
N2 adsorption–desorption isotherms were recorded at ꢁ196 °C
on a Micromeritics TriStar 3000 instrument. Before the measure-
ments, all the catalysts were pretreated under vacuum at 250 °C
for 8 h under high vacuum condition.
202

X. Gao, S. Zhu, M. Dong et al.
Journal of Catalysis 399 (2021) 201–211
extracted by ethyl acetate, and analyzed with ethyl benzoate as
internal standard. Diethyl ether was used as extract in the conver-
sion of stearic acid and palmitic acid. The gas products were ana-
lyzed by Agilent 7890A gas chromatograph equipped with a TCD
detector. All the products were identified by a Shimadzu GCMS-
QP2010 gas chromatogram–mass spectrometer (GC-MS) and
NMR spectroscopy (Bruker AV-III 400 MHz NMR spectrometer).
The turnover frequency (TOF) in the conversion of acetic acid
was calculated based on surface Co atoms at low conversion. The
surface Co atoms were estimated on total Co atoms and Co disper-
sion. According to previous report [30], Co dispersion (D) can be
calculated by the equation D = 96/d, in which d refers to Co mean
particle size in nm from TEM counting.
phene layers outside Co NPs are not totally closed, and some cracks
in 0.2–0.8 nm can be found. The annealed samples were fiercely
oxidized once it exposed to air, implying that O2 molecules can
freely pass through graphene layers via cracks to react with hcp-
Co species. Notably, these cracks play a decisive role in exposing
active Co facets for catalytic reaction.
For other samples annealing at 300, 500, 600 and 700 °C
(Figs. S1–S4), all the hcp-Co NPs are well covered by graphene
shell, despite with many cracks. The increased pyrolysis tempera-
ture improves Co particle size, graphene layers and graphitization
degree, but correspondingly declines graphene defects and cracks.
For Co@G300, the presence of both hcp-Co and Co3[Co(CN)6]2 spe-
cies in XRD patterns indicates that Co3[Co(CN)6]2 is not completely
decomposed at 300 °C, as supported by its low Co content but high
N content (Table 1). The BET surface area of hcp-Co@G300 reaches
332.4 m2/g and is much higher than other high temperature sam-
ples, also suggesting that it contains some remaining Co3[Co(CN)6]2
species.
As displayed in Fig. 2a, the diffraction reflections at 17.4°, 24.7°,
35.2°, 39.5° and 51.4° are ascribed to Co3[Co(CN)6]2 over hcp-
Co@G300 [34]. The reflections at 41.4°, 44.5° and 47.4° are
assigned to (010), (002) and (011) planes of hcp-Co phase [26].
Both Co3[Co(CN)6]2 and hcp-Co species are co-existed in hcp-
Co@G300, suggesting the incomplete decomposition of Co3[Co
(CN)6]2. The diffraction peaks of hcp-Co gradually increase in inten-
sity with increasing annealing temperature, indicating a steady
growth of Co particle size. The increase of Co particle size is also
supported by TEM statistical analysis. Additionally, TEM and XRD
results show that (002) plane is the most exposed facet in all the
hcp-Co@G catalysts.
Mole number of con
Number of surface Co atoms ꢂ reation time
verted acetic acid
TOF ¼
In the catalytic reusability tests, the used hcp-Co@G400 catalyst
was magnetically separated from the reaction mixture. The col-
lected catalyst was directly reused for the next run.
2.4. DFT calculation
Periodic DFT calculations were performed in Vienna ab initio
simulation package (VASP) with projector-augmented wave
(PAW) method [31]. Exchange and correlation energies were trea-
ted with spin-polarized generalized gradient approximation and
Perdew, Burke, and Ernzcrhof (PBE) functional [32]. A plane wave
energy cutoff was set as 450 eV, and a 4 ꢂ 4 ꢂ 1 Monkhorst-
Pack k-points were used.
The hcp Co (002) and fcc Co (111) surfaces were modeled by a
four-layer slab with 3 ꢂ 3 unit cell. A 15 Å vacuum layer was set
between the slabs to avoid interlayer interactions along z-
direction. The top two layers of Co atoms and adsorbates were fully
relaxed, while the bottom two layers were fixed during the calcu-
lation. The adsorption energy (Eads) are defined as the equation
Eads = Eads/slab ꢁ Egas ꢁ Eslab, in which Eads/slab, Eads and Eslab are
the energies of the adsorbate with slab, adsorbate in the gas phase,
and the clean slab, respectively.
The transition states calculations were implemented by climb-
ing image nudge elastic band (CI-NEB) method [33]. All the transi-
tion states were validated by vibrational frequency analysis. The
activation energy barrier (Ea) and reaction energy (Er) were deter-
mined by Ea = ETS ꢁ EIS and Er = EFS ꢁ EIS, wherein EIS, ETS, and EFS
are the energies of the initial state (IS), transition state (TS) and
final state (FS), respectively.
The Raman spectra of all the samples are illustrated in Fig. 2b.
The well-resolved two bands at 1344 and 1574 cmꢁ1 are assigned
to D and G bands, respectively [15,21].The D band is associated
with defects or discorded carbon, while G band refers to sp2-
hybridized graphitic carbon [15]. The hcp-Co@G300 catalyst show
very low intensity of D and G bands owing to the preservation of
substantial Co3[Co(CN)6]2. The relative intensity of IG/ID is indica-
tive of graphitization degree. For other samples, the IG/ID ratio
increases with increasing annealing temperature and indicates an
improved graphitization degree, coincident with the enhanced gra-
phene layer number [21].
Unexpectedly, pure hcp Co NPs are successfully fabricated and
remain stable even at high temperature up to 700 °C, which has
never been reported. The hcp Co is only stable below the allotropic
transformation temperature(420–450 °C), whereas the fccstructure
is thermodynamically stable above this temperature [36]. Previous
studies [28,29] have shown that the control synthesis of pure hcp-
Co is a great challenge when the preparation process needs high
temperature thermal treatment. Thus, the formation mechanism
of graphene encapsulated hcp-Co species was necessarily probed
by in situ UV–vis spectra and in situ XRD (Fig. 3a,b and Fig. S5). In
comparison with XRD, UV–vis spectroscopy is more sensitive to
detect surface or minor metal cation species. Fig. 3a in UV–vis spec-
tra show that the as-prepared Co3[Co(CN)6]2 displays many peaks at
250, 320, 528, 600, 638 and 690 nm, which are associated with Co2+
and Co3+. When the annealing temperature enhanced from 30 °C to
100 °C, the presence of very strong adsorption band at 587 nm is
ascribed to the isolated mononuclear octahedral Co2+ [37], originat-
ing from the decomposition of Co3[Co(CN)6]2. A moderate decline in
characteristic peak intensity of Co3[Co(CN)6]2 is also observed in
in situ XRD pattern (Fig. 3b) at 100 °C. The characteristic band of iso-
lated Co2+ is still detected up to 400 °C at UV–Vis spectra. Simultane-
ously, new XRD diffraction reflections at 41.3°, 42.6° and 45.2°
appear, which are assigned to Co2C crystalline [38]. With increasing
annealing time (Fig. S5) or temperature (Fig. 2b), the Co2C character-
istic peaks disappear both in UV–vis spectra and XRD patterns. Since
3. Results and discussion
3.1. Fabrication of hcp-Co@G and structure characterization
Prussian blue analogues (PBA) are one of the most extensively
explored MOFs for their uniform sizes, special thermal properties
and unique reactivity [34,35]. A series of core-shell structure cata-
lysts were prepared by pyrolyzing PBA-structured Co3[Co(CN)6]2
NPs at different temperature in 10% H2/Ar flow. As illustrated in
Fig. 1, the TEM and HRTEM images show that all the uniform Co
NPs in hcp-Co@G400 are well encapsulated by ultrathin graphene
shell. Considering its ultrahigh Co content up to 63.1% (Table 1),
the average Co particle size of 8.8 nm indicates its formation of
rather homogenous Co NPs. The formation of metastable hcp-Co
can be identified by its (111) lattice fringe of 0.202 nm [26], and
no fcc Co is detected.
Statistical analysis of more than 300 core-shell nanocomposites
confirms that the number of graphene shell is only consisted of 2–5
layers. Careful observation in HRTEM images show that the gra-
203

X. Gao, S. Zhu, M. Dong et al.
Journal of Catalysis 399 (2021) 201–211
Fig. 1. (a, b, d) TEM and HRTEM images of hcp-Co@G400; (c) hcp-Co NPs size distribution; (e) statistical analysis of graphene layer number distribution in the graphene shell;
(f) schematic illustration of hcp-Co@G400 structure.
Table 1
Textural and physicochemical properties for various catalysts.
Catalyst
BET surface
area (m2/g)
Pore volume
(cm3/g)
Pore size
(nm)
Co content
(wt. %)a
Co particle
size (nm)b
Dispersion (%)
N content (at.%)c
hcp-Co@G300
hcp-Co@G400
hcp-Co@G500
hcp-Co@G600
hcp-Co@G700
20fcc-Co@G400
60fcc-Co@G400
332.4
88.4
56.1
50.2
50.3
75.1
48.9
0.288
0.385
0.297
0.284
0.289
0.367
0.264
15.4
21.4
25.3
25.1
24.4
22.8
24.0
29.8
63.1
63.6
64.0
64.3
19.1
62.5
8.0
8.8
10.5
12.6
13.8
7.0
12.0
10.9
9.1
7.6
7.0
17.6
2.9
2.2
2.1
2.0
3.9
1.7
13.7
7.7
12.4
a
b
c
Tested by ICP.
Determined by TEM.
Obtained by EA.
Fig. 2. (a) XRD patterns and (b) Raman spectra of hcp-Co@GT catalysts annealed at various temperatures.
204

X. Gao, S. Zhu, M. Dong et al.
Journal of Catalysis 399 (2021) 201–211
Fig. 3. (a) In situ UV–vis and (b) in situ XRD patterns for annealing Co3[Co(CN)6]2 in H2/Ar gas with a heating rate of 5 °C/min. (c) Co K-edge XANES and (d) Fourier
transformed EXAFS spectra of hcp-Co@GT catalysts, Co foil and Co3O4.
then, noany obvious peaks are detected by UV–vis spectra, reflecting
that there is no oxidized Co species. Indeed, only reduced hcp-Co
species diffractionreflections are detected by XRD. The above results
indicate that the Co2C from Co3[Co(CN)6]2 decomposition is rapidly
reduced into metastable hcp-Co. Metallic hcp-Co NPs can catalyti-
cally promote the growth of graphene around their edges, and are
further confined in graphene shell. The core-shell structure can inhi-
bit the occurrence of allotropic transformation of hcp-Co into fcc-Co
below 700 °C.
XPS spectroscopy (Fig. S6) analysis, X-ray absorption near-edge
structure (XANES, Fig. 3c), and extended X-ray absorption fine
structure (EXAFS, Fig. 3d) were employed to determine the chem-
ical states and coordination of the Co atoms in hcp-Co@G catalysts.
As displayed in Fig. S6, hcp-Co@G400 and other high temperature
samples present two peaks at around 778.5 and 793.5 eV, which
are assigned to 2p3/2 and 2p1/2 spectra of metallic Co0 [21], respec-
tively. The spent hcp-Co@G400 still displayed only Co0 characteris-
tic peaks, indicating that Co0 active sites kept stable during the
reaction. For hcp-Co@G300, another two broad peaks at 782.1
and 797.3 eV are associated with Co2+/3+ of Co3[Co(CN)6]2. Except
for hcp-Co@G300, all the hcp-Co@G catalysts show similar XANES
spectra as Co foil (Fig. 3c), suggesting that Co species are only
existed as metal state. The EXAFS results show that only CoACo
bonds were detected for the above samples, consistent with XPS
results for the detection of Co0. In the case of hcp-Co@G300, its
pre-edge peak is located at 7710.5 eV, lower than that of Co foil
(7711.4 eV) and higher than that of Co3O4 (7709.0 eV). This sug-
gests that it contains different types of Co species. The preservation
of CoAN bond is verified by its bond length of 1.42 Å over hcp-
Co@G300 in EXAFS [39]. Combined with XANES, EXAFS, and XRD
results, it can be deduced that both undecomposed Co3[Co(CN)6]2
and hcp-Co are present in the hcp-Co@G300.
3.2. Catalytic performance for carboxylic acid hydrogenation
LA and acetic acid are two of the most important biomass-
related carboxylic acids [40,41]. Therefore, these two acids are
selected as model compounds for hydrogenation at 220 °C and
5.0 MPa H2. According to the reaction pathway of LA hydrogena-
tion (Scheme S1), the reduction of carbonyl and carboxylic group
is competitive, and the product distribution is highly dependent
on catalyst composition. As shown in Fig. 4a and Table S1, all the
hcp-Co@G catalysts exhibited complete LA conversion but with
various 1,4-pentanediol (1,4-PDO) yields, because carbonyl group
are much more easily reduced than carboxylic group in LA. Hcp-
Co@G300 gave 29.9% yield of target product 1,4-PDO along with
a variety of by-products, including 19.1%
c-valerolactone (GVL),
11.7% 2-methyl tetrahydrofuran (MTHF), and 3.1% 1-pentanol.
The poor 1,4-PDO yield of hcp-Co@G300 is primarily ascribed to
its low hcp-Co content, originating from the incomplete decompo-
sition of Co3[Co(CN)6]2. Of all the examined catalysts, hcp-
Co@G400 exhibited dramatic enhancement in 1,4-PDO yield,
reaching 89.5%. However, an increase of annealing temperature
gradually declined the yield of 1,4-PDO as a result of increased
GVL yield. Because these catalysts cannot effectively reduce car-
boxylic into hydroxyl group, the intermediate product 4-hydroxyl
pentanoic acid (HPA) cannot be converted into 1,4-PDO but under-
goes lactonization to form GVL.
Additionally, we used sulfuric acid to completely etch Co
species from hcp-Co@G400 in order to prepare G400 support.
205

X. Gao, S. Zhu, M. Dong et al.
Journal of Catalysis 399 (2021) 201–211
Fig. 4. (a) Catalytic performance of LA hydrogenation over various catalysts. Reaction conditions: 1 mmol LA, 0.05 g catalyst, 220 °C, 5 MPa, 24 h. (b) TOF value of acetic
hydrogenation over various catalysts.
ICP-AES result confirmed that the hcp-Co species had been totally
removed. The G400 support was nearly inactive for LA conversion
(Table S1). Subsequently, 20fcc-Co@G400 and 60fcc-Co@G400 cat-
alysts were prepared from Co(NO3)2ꢀ6H2O with similar methods as
hcp-Co@G400. These two catalysts only contain fcc-Co crystalline
phase, as confirmed by TEM (Figs. S7–S8) and XRD (Fig. S9) mea-
surements. Both of them exhibited very poor performance for LA
hydrogenation, and their yields of 1,4-PDO were only 0.4% and
3.6%, respectively. These evaluation results unambiguously con-
firm the superior activity of hcp-Co to fcc-Co. Additionally, other
base metal catalysts, such as Cu, Ni and Fe, also showed very poor
yields of 1,4-PDO.
To better reflect the reduction ability of carboxylic group for
these catalysts, the model reaction of acetic acid hydrogenation
was performed to estimate their intrinsic turnover frequency
(TOF, Fig. 4b and Table S2). All the catalysts obtained high ethanol
selectivity (greater than 93%) along with small amounts of by-
product ethane. The activity trend was similar to their performance
in LA hydrogenation. The TOF value of hcp-Co@G400 reached
10.7 hꢁ1, higher than those catalysts annealed at other tempera-
tures. The remaining Co3[Co(CN)6]2 species in hcp-Co@G300 can-
not contribute to TOF, and decreased its value to 5.2 hꢁ1. For the
catalysts annealed at temperature higher than 400 °C, the
increased graphene layer numbers greatly inhibited the adsorption
of acetic acid (Fig. S10), leading to low reaction performance. As
confirmed by DRIFTS of acetic acid adsorption in Fig. S10, the peak
area of acetic acid adsorption (at 1726, 1468 and 1291 cmꢁ1) of
hcp-Co@G400 was much higher than other samples. Both 20fcc-
Co@G400 and 60fcc-Co@G400 gave very limited TOF in acetic acid
hydrogenation, only 0.3 hꢁ1 and 0.4 hꢁ1, respectively. Their TOF
values were even one magnitude lower than that of hcp-Co@G
countparts. The particle size effects on the TOF can be excluded
since both Co particles are centered at 10 ~ 13 nm for 60fcc-
Co@G400 and hcp-Co@G500, 600 and 700 catalysts with similar
Co content. Thus, the huge difference in reactivity is strongly
related to Co crystalline phase.
hydrogenation, particularly for non-precious metal catalysts. As
displayed in Fig. 5a, no any decrease in LA conversion and 1,4-
PDO yield was observed over hcp-Co@G400 in the reusability tests.
The graphene encapsulated hcp-Co nanostructures were still well
preserved after six reaction cycle tests, as confirmed by TEM
images (Fig. S12). Moreover, ICP-AES test verified that only 1.1%
Co content was leached, suggesting that the graphene shell can
protect Co from leaching and corrosion caused by LA. In a control
experiment, the supported 60fcc-Co@G400 (Fig. 5b) without gra-
phene overcoat exhibited irreversible deactivation due to substan-
tial Co leaching (18.7%).
Considering the excellent performance in LA and acetic acid
hydrogenation over hcp-Co@G400, we investigated its substrate
scope for other typical carboxylic acid, particularly for biomass-
derived acids (Table 2). More than 90% yields of alcohols were
obtained in the conversion of short-chain monoaliphatic acids
and challenging long-chain acids, such as palmitic and stearic acid.
Diols are of special interest because of their vital role in manufac-
turing polyesters [5,10], particularly using the biomass-related car-
boxylic acids as feeds. We were pleased to find that these typical
bio-based carboxylic acids, including LA, lactic, succinic, adipic, ita-
conic acids, were smoothly converted into desired diols with good
to excellent yields. Moreover, many aromatic acids, such as furoic,
benzoic and functionalized benzoic acids, can also been reduced
efficiently in 75.6–94.6% yields. Notably, the aromatic rings,
ANO2, and aromatic ether were initially reduced, suggesting that
the reduction of carboxylic group is very challenging relative to
other reductive groups. However, there are still some limitations
for this catalytic system. The selective hydrogenation of carboxylic
moieties over other functional groups is very important for the
synthesis of some special chemicals [9,42]. The aromatic C@C
bonds in furan or benzene rings are preferentially reduced over
hcp-Co@G400, which results in total disappearance of aromaticity
in the target alcohol.
3.3. Reaction mechanism for carboxylic acid hydrogenation
The time course of LA hydrogenation over hcp-Co@G400 is dis-
played in Fig. S11. LA conversion sharply increased to 100% in 4 h.
With increasing reaction time to 24 h, the yield of 1,4-PDO gradu-
ally improved at the expense of GVL. The reaction trend confirms
that LA preferentially undergoes the hydrogenation of carbonyl
group in comparison to carboxylic group. 1,4-PDO yield reached
the maximal value of 89.5% at 24 h. The reaction stability is a very
important criterion to evaluate strong acidic carboxylic acid
So far, the reaction mechanism of carboxylic acid (RCOOH)
hydrogenation still remains great debate on the initial dehydro-
genation to form RCOO* and then deoxygenation to RCO* or direct
dehydroxylation to generate RCO* [13,43,44]. Another key ques-
tion is whether aldehyde (RCHO) is involved by RCO* hydrogena-
tion [45,46]. To explore possible reaction mechanism, we initially
performed the reaction kinetics of acetic acid hydrogenation over
206

X. Gao, S. Zhu, M. Dong et al.
Journal of Catalysis 399 (2021) 201–211
Fig. 5. Reusability tests for LA hydrogenation over (a) hcp-Co@G400 and (b) 60fcc-Co@G400. Reaction conditions: 1 mmol LA, 0.05 g catalyst, 220 °C, 5 MPa, 24 h.
Table 2
Reaction results of various functional carboxylic acids hydrogenation to alcohols over hcp-Co@G400.a
Substrate
Alcohol product
Time (h)
Con. (%)
Yield (%)
Alcohol
93.4
Substrate
Alcohol product
Time (h)
24
Con. (%)
100
Yield (%)
Alcohol
81.7
Alkane
5.0
Alkane
1.1
2
100
2
2
100
100
95.1
98.4
3.2
0.5
24
24
100
100
94.4
73.2
0.7
1.2
2
2
100
100
99.5
96.8
0.2
2.7
24
12
100
100
81.9
94.6
8.7
2.3
24
24
99.8
98.9
90.8
92.3
3.4
3.0
24
24
100
100
89.4
86.8
8.9
9.4
24
2
100
100
89.5
91.4
2.8
0.9
24
24
100
100
75.6
90.1
4.2
3.6
a
Reaction conditions: 1 mmol substrate, 0.05 g catalyst, 220 °C, 5 MPa H2, 2 or 24 h. Conversion and yield were determined by GC with internal standard method. 1,2-
butanediol and 1-butanol were generally used as internal standards for the formation of diol and monohydric alcohol, respectively. The detailed analysis procedure was
described in Section 2.3.
hcp-Co@G400 catalyst. As illustrated in Fig. 6a, the reaction order
with respect to H2 pressure was calculated to be 0.58, indicating
that the reaction rate was moderately affected by H2 pressure. As
the dissociation of H2 is very fast on Co surface, the further reaction
of dissociated H* may require high reaction energy barrier. The
reaction order with respect to acetic acid concentration (Fig. 6b)
was estimated to be 0.08, suggesting that acetic acid is strongly
adsorbed on hcp-Co surface, in well line with DFT calculation
results (see below).
cannot chemically adsorbed on support alone (Fig. S10). As dis-
played in Fig. 7a, three typical peaks cantered at 1726, 1468 and
1291 cmꢁ1 are assigned to C@O stretching, CH3 bending, CAO
stretching vibrations of adsorbed acetic acid over hcp-Co@G400
[47,48], respectively. After introduction of flowing H2 at 220 °C,
the peak intensity of acetic acid gradually decreased with increas-
ing reaction time. Meanwhile, a new peak corresponding to C@O
stretching vibration of acetyl group appeared at 1630 cmꢁ1 after
10 min and steadily grew up to 15 min [47]. The varying trends
of these peaks indicate that CH3COOH undergoes dehydroxylation
to form CH3CO* over hcp-Co sites. With further increasing time to
20 min, the peak of acetyl gradually disappeared accompanying
To deeply understand the huge difference in reactivity of hcp-
Co and fcc-Co, the surface reaction process was monitored by
in situ DRIFTS of acetic acid hydrogenation. Notably, acetic acid
207

X. Gao, S. Zhu, M. Dong et al.
Journal of Catalysis 399 (2021) 201–211
Fig. 6. Effect of (a) H2 pressure and (b) acetic acid concentration on the reaction rate (r) of acetic acid hydrogenation over hcp-Co@G400. Reaction conditions: 0.05 g catalyst,
220 °C, 2 h.
Fig. 7. In situ DRIFTS spectroscopy of acetic hydrogenation over (a) hcp-Co@G400 at 220 °C and (b) 60fcc-Co@G400.
with the presence of ethanol vibration peaks at 1390 (CH3) and
1066 cmꢁ1 (CAO) [45,49]. These evidences confirm that acetic acid
hydrogenation to ethanol is proceeded through the acetyl pathway
on hcp-Co surface. Although a similar phenomenon was observed
for acetic acid hydrogenation over 60fcc-Co@G400 (Fig. 7b), its
activity was much lower than that of hcp-Co@G400. No surface
reaction occurred at 220 °C and it began to react at 250 °C. The very
poor reactivity of 60fcc-Co@G400 is consistent with the evaluation
results.
most exposed crystalline facets, as evidenced by XRD and TEM
results. The optimized structures and adsorption energies for all
the possible species were first calculated and the results are sum-
marized at Fig. 8, Table S3 and Table S4. CH3COOH preferably
adsorbs at Co top site through carbonyl oxygen, in which the
CAC bond is vertical to hcp-Co surface. Its adsorption energy is
ꢁ0.63 eV and CoAO bond length is 2.081 Å, suggesting its strong
chemisorption on hcp-Co surface. Contrarily, the preferred adsorp-
tion of CH3COOH on fcc-Co is very weak with an absorption energy
of 0.08 eV and long CoAO bond (2.208 Å), in which the CAC bond is
parallel to fcc-Co surface. H2, ethanol and all the possible interme-
diate species exhibit similar or minor differences in adsorption
We carried out DFT calculations to investigate the reaction
mechanism on hcp-Co and fcc-Co catalysts. The hcp-Co (002)
and fcc-Co (111) surfaces were modelled (Fig. S13) based on the
208

X. Gao, S. Zhu, M. Dong et al.
Journal of Catalysis 399 (2021) 201–211
Fig. 8. Adsorption configurations and energies of acetic acid on (a and b) hcp-Co (002) and (c and d) fcc-Co (111) facets. The blue, brown, red and white balls represent Co, C,
O and H atoms, respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
configurations and energies on hcp-Co and fcc-Co. H2 preferentially
adsorbs on top sites in hcp-Co, and easily dissociates to H* on 3-
fold hollow sites. The H2 dissociation (Fig. S14) has nearly no reac-
tion barrier and is thermodynamically favourable process. More-
over, the diffusion of H* from a hollow site to a nearby hollow
site has a very low activation barrier of 0.15 eV, suggesting its high
Fig. 9. Potential energy surfaces of acetic acid hydrogenation to ethanol on (a) hcp-Co (002) and (b) fcc-Co (111) surfaces. The blue, brown, red and white balls represent Co,
C, O and H atoms, respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
209

X. Gao, S. Zhu, M. Dong et al.
Journal of Catalysis 399 (2021) 201–211
mobility. Similar behaviour is observed on fcc-Co, confirming that
metal Co possesses excellent H2 activation ability, irrespective of
crystallite phase.
4. Conclusions
In this work, we have successfully synthesized pure hcp-Co NPs
encapsulated in few-layer graphene via the reduction of Co2C
intermediate from MOF precursor. The graphene shell enables
metastable hcp-Co NPs to be strongly confined in the inner space,
which can resist not only their allotropic transformation upon high
temperature treatment but also their leaching against strong car-
boxylic acids. Compared to conventional 60fcc-Co@G400, hcp-
Co@G400 exhibited one magnitude enhancement in reaction activ-
ity of carboxylic acid hydrogenation to alcohol. Moreover, the gra-
phene encapsulated hcp-Co@G400 showed extensive substrate
tolerance and excellent reusability. In situ DRIFTS and DFT calcula-
tion in model acetic acid hydrogenation reaction reveal that car-
boxylic acid hydrogenation initially undergoes dehydroxylation
to RCO* followed by its consecutive hydrogenation to RCH2OH.
The enhanced reaction reactivity of hcp-Co over fcc-Co is mainly
due to the decreasing activation energy from 1.23 eV to 0.86 eV
in acetic acid dehydroxylation, which is associated with their dis-
tinct CH3COOH adsorption configurations.
Based on all the possible elementary reactions, the full potential
energy surface of acetic acid hydrogenation over hcp-Co are dis-
played in Fig. 9a and their transition state (TS) configurations are
listed in Table S5. CH3CO* is the first generated intermediate, no
matter the reaction occurs via the direct dehydroxylation pathway
or the dehydrogenation-deoxygenation pathway. Compared to
direct dehydroxylation pathway, the activation energy of initial
CH3COOH* dehydrogenation reaction is lower, while its consecu-
tive deoxygenation is fairly unfavourable owing to the remarkably
high activation barrier up to 1.27 eV. Therefore, the CH3CO* is most
probably formed from CH3COOH* dehydroxylation with a moder-
ate activation energy of 0.86 eV. The generated CH3CO* species
can proceed to hydrogenation of carbonyl group either by initial
CAH bond formation to CH3CHO* or OAH bond formation to CH3-
COH*. With lower activity energy of 0.33 eV, CAH bond formation
is preferred over OAH bond formation (0.33 eV vs. 0.49 eV), in line
with previous calculation results of the hydrogenation of ketones
[50,51]. Subsequently, CH3CHO* undergoes further hydrogenation
either by CAH bond formation or OAH bond formation. The forma-
tion of CH3CH2O* is kinetically and thermodynamically favourable
(Ea = 0.08 eV and Er = ꢁ0.67 eV). However, the hydrogenation of
CH3CH2O* to CH3CH2OH* needs significantly high activation bar-
rier of 1.09 eV, suggesting that this reaction pathway is probably
not favourable. The formation of CH3CHOH* from CH3CHO* hydro-
genation by CAH bond also needs high activation energy of
0.89 eV. Moreover, CH3CHO species were not detected by DRIFTS
spectroscopy, suggesting that CH3CHO pathway is not favoured
at all. Then, we reconsider that the hydrogenation of CH3CO* to
CH3COH* only has a little higher activation energy (0.49 eV) than
that of CH3CHO* formation (0.33 eV). Further generation of CH3-
CHOH* by CH3COH* hydrogenation is remarkably preferred to CH3-
CHO* hydrogenation (0.50 eV vs. 0.89 eV). Moreover, consecutive
hydrogenation of CH3CHOH* to final product CH3CH2OH* occurs
with nearly no activation barrier (0.05 eV), because this is a ther-
modynamically favourable elementary reaction (ꢁ0.29 eV). Thus,
it can be deduced that CH3COOH hydrogenation to ethanol follows
the pathway of CH3COOH* ? CH3CO* ? CH3COH* ? CH3CHOH* ?
CH3CH2OH*, as supported by the in situ DRIFTS spectra results that
the CH3CHO species were not detected. The produced ethanol can
easily desorb from hcp-Co (002) surface with 0.29 eV. Notably, the
remaining OH* from CH3COOH* dehydroxylation will react with
nearby H* to produce H2O with rather high activation barrier
(1.07 eV), suggesting that this is the determining-rate step for CH3-
COOH hydrogenation over hcp-Co.
Declaration of Competing Interest
The authors declare that they have no known competing finan-
cial interests or personal relationships that could have appeared
to influence the work reported in this paper.
Acknowledgements
We are grateful to the financial support from National Natural
Science Foundation of China (21878321), Science Foundation for
Youth Scholars of State Key Laboratory of Coal Conversion
(2016BWZ002), and Youth Innovation Promotion Association CAS
(2015140).
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.jcat.2021.05.013.
References
[1] P. Gallezot, Conversion of biomass to selected chemical products, Chem. Soc.
Rev. 41 (2012) 1538–1558.
[2] C.-H. Zhou, X. Xia, C.-X. Lin, D.-S. Tong, J. Beltramini, Catalytic conversion of
lignocellulosic biomass to fine chemicals and fuels, Chem. Soc. Rev. 40 (2011)
5588–5617.
Although acetic acid hydrogenation follows the same reaction
pathway in both hcp-Co and fcc-Co (Fig. 9b), large differences in
the reaction and activation energies are clearly observed. The
determining-rate step over fcc-Co becomes the formation of CH3-
CO* from CH3COOH* dehydroxylation with a significantly higher
activation barrier of 1.23 eV, which leads to poor reaction activity
of fcc-Co. The weak parallel adsorption of CH3COOH on fcc-Co
needs to undergo configuration transformation to vertical motif
in the formation of TS1, resulting in higher activation barrier in
dehydroxylation reaction. Conversely, the vertical adsorption of
CH3COOH on hcp-Co facilitates the formation of TS1 by elongating
CAO bond from 1.36 Å to 2.08 Å, which greatly declines the activa-
tion barrier to 0.86 eV. Another elementary reaction of CH3CHOH +
H ? CH3CH2OH also have moderate difference in activation bar-
rier. For CH3CHOH* configuration, the CAC tilt angle relative to
hcp-Co surface is more similar to that of CH3CH2OH* and only a
small distortion in the formation of TS4 geometry, thus leading
to lower activation energy in comparison to fcc-Co.
[3] X. Cui, Y. Li, C. Topf, K. Junge, M. Beller, Direct Ruthenium-catalyzed
hydrogenation of carboxylic acids to alcohols, Angew. Chem. Inter. Ed. 54
(2015) 10596–10599.
[4] F.M.A. Geilen, B. Engendahl, M. Hölscher, J. Klankermayer, W. Leitner, Selective
homogeneous hydrogenation of biogenic carboxylic acids with [Ru(TriPhos)
H]+: A mechanistic study, J. Am. Chem. Soc. 133 (2011) 14349–14358.
[5] H. Yim, R. Haselbeck, W. Niu, C. Pujol-Baxley, A. Burgard, J. Boldt, J. Khandurina,
J.D. Trawick, R.E. Osterhout, R. Stephen, J. Estadilla, S. Teisan, H.B. Schreyer, S.
Andrae, T.H. Yang, S.Y. Lee, M.J. Burk, S. Van Dien, Metabolic engineering of
escherichia coli for direct production of 1,4-butanediol, Nat. Chem. Biol. 7
(2011) 445.
[6] Z. Huang, K.J. Barnett, J.P. Chada, Z.J. Brentzel, Z. Xu, J.A. Dumesic, G.W. Huber,
Hydrogenation of c-butyrolactone to 1,4-butanediol over CuCo/TiO2 bimetallic
catalysts, ACS Catal. 7 (2017) 8429–8440.
[7] T.P. Brewster, A.J.M. Miller, D.M. Heinekey, K.I. Goldberg, Hydrogenation of
carboxylic acids catalyzed by half-sandwich complexes of iridium and
rhodium, J. Am. Chem. Soc. 135 (2013) 16022–16025.
[8] J. Pritchard, G.A. Filonenko, R. van Putten, E.J.M. Hensen, E.A. Pidko,
Heterogeneous and homogeneous catalysis for the hydrogenation of
carboxylic acid derivatives: history, advances and future directions, Chem.
Soc. Rev. 44 (2015) 3808–3833.
[9] J. Ullrich, B. Breit, Selective hydrogenation of carboxylic acids to alcohols or
alkanes employing a heterogeneous catalyst, ACS Catal. 8 (2018) 785–789.
210

X. Gao, S. Zhu, M. Dong et al.
Journal of Catalysis 399 (2021) 201–211
´
´
[10] Y. Takeda, T. Shoji, H. Watanabe, M. Tamura, Y. Nakagawa, K. Okumura, K.
Tomishige, Selective hydrogenation of lactic acid to 1,2-propanediol over
[30] A.n. Martınez, C. López, F. Márquez, I. Dıaz, Fischer-Tropsch synthesis of
hydrocarbons over mesoporous Co/SBA-15 catalysts: the influence of metal
loading, cobalt precursor, and promoters, J. Catal. 220 (2003) 486–499.
[31] G. Kresse, D. Joubert, From ultrasoft pseudopotentials to the projector
augmented-wave method, Phys. Rev. B 59 (1999) 1758–1775.
highly active ruthenium–molybdenum oxide catalysts, ChemSusChem
(2015) 1170–1178.
8
[11] H.G. Manyar, C. Paun, R. Pilus, D.W. Rooney, J.M. Thompson, C. Hardacre,
Highly selective and efficient hydrogenation of carboxylic acids to alcohols
using titania supported Pt catalysts, Chem. Commun. 46 (2010) 6279–6281.
[12] Y. Takeda, M. Tamura, Y. Nakagawa, K. Okumura, K. Tomishige, Hydrogenation
of dicarboxylic acids to diols over Re-Pd catalysts, Catal. Sci. Technol. 6 (2016)
5668–5683.
[32] J.P. Perdew, K. Burke, M. Ernzerhof, Generalized gradient approximation made
simple, Phys. Rev. Lett. 77 (1996) 3865–3868.
[33] G. Henkelman, B.P. Uberuaga, H. Jónsson, A climbing image nudged elastic
band method for finding saddle points and minimum energy paths, J. Chem.
Phys. 113 (2000) 9901–9904.
[13] T.J. Korstanje, J. Ivar van der Vlugt, C.J. Elsevier, B. de Bruin, Hydrogenation of
carboxylic acids with a homogeneous cobalt catalyst, Science 350 (2015) 298–
302.
[14] H. Bin Wu, X.W. Lou, Metal-organic frameworks and their derived materials
for electrochemical energy storage and conversion: Promises and challenges,
Sci. Adv. 3 (2017) 9252.
[15] L. Fan, P.F. Liu, X. Yan, L. Gu, Z.Z. Yang, H.G. Yang, S. Qiu, X. Yao, Atomically
isolated nickel species anchored on graphitized carbon for efficient hydrogen
evolution electrocatalysis, Nat. Commun. 7 (2016) 10667.
[16] B.Y. Xia, Y. Yan, N. Li, H.B. Wu, X.W. Lou, X. Wang, A metal-organic framework-
derived bifunctional oxygen electrocatalyst, Nat. Energy 1 (2016) 15006.
[17] J. Deng, P. Ren, D. Deng, X. Bao, Enhanced electron penetration through an
ultrathin graphene layer for highly efficient catalysis of the hydrogen
evolution reaction, Angew. Chem. Inter. Ed. 54 (2015) 2100–2104.
[18] K.M. Rajenahally, V. Jagadeesh, A.S. Alshammari, H. Neumann, M.-M. Pohl, J.
Radnik, 1 Matthias Beller, MOF-derived cobalt nanoparticles catalyze a general
synthesis ofamines, Science 358 (2017) 326–332.
[19] V.P. Santos, T.A. Wezendonk, J.J.D. Jaén, A.I. Dugulan, M.A. Nasalevich, H.-U.
Islam, A. Chojecki, S. Sartipi, X. Sun, A.A. Hakeem, A.C.J. Koeken, M. Ruitenbeek,
T. Davidian, G.R. Meima, G. Sankar, F. Kapteijn, M. Makkee, J. Gascon, Metal
organic framework-mediated synthesis of highly active and stable Fischer-
Tropsch catalysts, Nat. Commun. 6 (2015) 6451.
[34] L. Han, X.-Y. Yu, X.W. Lou, Formation of prussian-blue-analog nanocages via a
direct etching method and their conversion into Ni–Co-mixed oxide for
enhanced oxygen evolution, Adv. Mater. 28 (2016) 4601–4605.
[35] Y. Yang, Z. Lin, S. Gao, J. Su, Z. Lun, G. Xia, J. Chen, R. Zhang, Q. Chen, Tuning
electronic structures of nonprecious ternary alloys encapsulated in graphene
layers for optimizing overall water splitting activity, ACS Catal. 7 (2017) 469–
479.
[36] V.A. de la Peña O0Shea, P.R. de la Piscina, N. Homs, G. Aromí, J.L.G. Fierro,
Development of hexagonal closed-packed cobalt nanoparticles stable at high
temperature, Chem. Mater. 21 (2009) 5637–5643.
[37] N. Osakoo, R. Henkel, S. Loiha, F. Roessner, J. Wittayakun, Palladium-promoted
cobalt catalysts supported on silica prepared by impregnation and reverse
micelle for Fischer-Tropsch synthesis, Appl. Catal. A: Gen. 464–465 (2013)
269–280.
[38] Z. Zhao, W. Lu, R. Yang, H. Zhu, W. Dong, F. Sun, Z. Jiang, Y. Lyu, T. Liu, H. Du, Y.
Ding, Insight into the formation of Co@Co2C catalysts for direct synthesis of
higher alcohols and olefins from syngas, ACS Catal. 8 (2018) 228–241.
[39] H. Fei, J. Dong, Y. Feng, C.S. Allen, C. Wan, B. Volosskiy, M. Li, Z. Zhao, Y. Wang,
H. Sun, P. An, W. Chen, Z. Guo, C. Lee, D. Chen, I. Shakir, M. Liu, T. Hu, Y. Li, A.I.
Kirkland, X. Duan, Y. Huang, General synthesis and definitive structural
identification of MN4C4 single-atom catalysts with tunable electrocatalytic
activities, Nat. Catal. 1 (2018) 63–72.
[20] J. Tang, R.R. Salunkhe, J. Liu, N.L. Torad, M. Imura, S. Furukawa, Y. Yamauchi,
Thermal conversion of core-shell metal-organic frameworks: A new method
for selectively functionalized nanoporous hybrid carbon, J. Am. Chem. Soc. 137
(2015) 1572–1580.
[40] W. Luo, M. Sankar, A.M. Beale, Q. He, C.J. Kiely, P.C.A. Bruijnincx, B.M.
Weckhuysen, High performing and stable supported nano-alloys for the
catalytic hydrogenation of levulinic acid to
(2015) 6540.
c-valerolactone, Nat. Commun. 6
[21] X. Ma, Y.-X. Zhou, H. Liu, Y. Li, H.-L. Jiang, A MOF-derived Co-CoO@N-doped
porous carbon for efficient tandem catalysis: Dehydrogenation of ammonia
borane and hydrogenation of nitro compounds, Chem. Commun. 52 (2016)
7719–7722.
[22] L. Yu, J.F. Yang, X.W. Lou, Formation of CoS2 nanobubble hollow prisms for
highly reversible lithium storage, Angew. Chem. Inter. Ed. 55 (2016) 13422–
13426.
[41] D.J.M. de Vlieger, L. Lefferts, K. Seshan, Ru decorated carbon nanotubes – a
promising catalyst for reforming bio-based acetic acid in the aqueous phase,
Green Chem. 16 (2014) 864–874.
[42] T. Toyao, S.M. Siddiki, A.S. Touchy, W. Onodera, K. Kon, Y. Morita, T. Kamachi,
K. Yoshizawa, K. Shimizu, TiO2-supported Re as a general and chemoselective
heterogeneous catalyst for hydrogenation of carboxylic acids to alcohols,
Chem. Eur. J. 23 (2017) 1001–1006.
[23] H. Hu, B. Guan, B. Xia, X.W. Lou, Designed formation of Co3O4/NiCo2O4 double-
shelled nanocages with enhanced pseudocapacitive and electrocatalytic
properties, J. Am. Chem. Soc. 137 (2015) 5590–5595.
[24] W.-Z. Li, J.-X. Liu, J. Gu, W. Zhou, S.-Y. Yao, R. Si, Y. Guo, H.-Y. Su, C.-H. Yan, W.-
X. Li, Y.-W. Zhang, D. Ma, Chemical insights into the design and development
of face-centered cubic ruthenium catalysts for Fischer-Tropsch synthesis, J.
Am. Chem. Soc. 139 (2017) 2267–2276.
[25] J.-X. Liu, H.-Y. Su, D.-P. Sun, B.-Y. Zhang, W.-X. Li, Crystallographic dependence
of CO activation on cobalt catalysts: HCP versus FCC, J. Am. Chem. Soc. 135
(2013) 16284–16287.
[43] H. Olcay, L. Xu, Y. Xu, G.W. Huber, Aqueous-phase hydrogenation of acetic acid
over transition metal catalysts, ChemCatChem 2 (2010) 1420–1424.
[44] V. Pallassana, M. Neurock, Reaction paths in the hydrogenolysis of acetic acid
to ethanol over Pd (111), Re (0001), and PdRe alloys, J. Catal. 209 (2002) 289–
305.
[45] W. Rachmady, M.A. Vannice, Acetic acid reduction by H2 over supported Pt
catalysts: A DRIFTS and TPD/TPR study, J. Catal. 207 (2002) 317–330.
[46] M. Zhang, R. Yao, H. Jiang, G. Li, Y. Chen, Insights into the mechanism of acetic
acid hydrogenation to ethanol on Cu (111) surface, Appl. Surf. Sci. 412 (2017)
342–349.
[26] S. Lyu, L. Wang, J. Zhang, C. Liu, J. Sun, B. Peng, Y. Wang, K.G. Rappé, Y. Zhang, J.
Li, L. Nie, Role of active phase in Fischer-Tropsch synthesis: Experimental
evidence of CO activation over single-phase cobalt catalysts, ACS Catal. 8
(2018) 7787–7798.
[47] L. Chen, Y. Zhu, H. Zheng, C. Zhang, Y. Li, Aqueous-phase hydrodeoxygenation
of propanoic acid over the Ru/ZrO2 and Ru–Mo/ZrO2 catalysts, Appl. Catal. A:
Gen. 411–412 (2012) 95–104.
[48] L. Chen, Y. Zhu, H. Zheng, C. Zhang, B. Zhang, Y. Li, Aqueous-phase
hydrodeoxygenation of carboxylic acids to alcohols or alkanes over
supported Ru catalysts, J. Mol. Catal. A: Chem. 351 (2011) 217–227.
[49] M. Tamura, R. Tamura, Y. Takeda, Y. Nakagawa, K. Tomishige, Insight into the
mechanism of hydrogenation of amino acids to amino alcohols catalyzed by a
heterogeneous MoOx-modified Rh catalyst, Chem. Eur. J. 21 (2015) 3097–3107.
[50] O. Mamun, E. Walker, M. Faheem, J.Q. Bond, A. Heyden, Theoretical
[27] Y.-P. Pei, J.-X. Liu, Y.-H. Zhao, Y.-J. Ding, T. Liu, W.-D. Dong, H.-J. Zhu, H.-Y. Su, L.
Yan, J.-L. Li, W.-X. Li, High alcohols synthesis via Fischer-Tropsch reaction at
cobalt metal/carbide interface, ACS Catal. 5 (2015) 3620–3624.
[28] R.W. Mitchell, D.C. Lloyd, L.G.A. van de Water, P.R. Ellis, K.A. Metcalfe, C.
Sibbald, L.H. Davies, D.I. Enache, G.J. Kelly, E.D. Boyes, P.L. Gai, Effect of
pretreatment method on the nanostructure and performance of supported Co
catalysts in Fischer-Tropsch synthesis, ACS Catal. 8 (2018) 8816–8829.
[29] M. Sadeqzadeh, H. Karaca, O.V. Safonova, P. Fongarland, S. Chambrey, P.
Roussel, A. Griboval-Constant, M. Lacroix, D. Curulla-Ferré, F. Luck, A.Y.
Khodakov, Identification of the active species in the working alumina-
supported cobalt catalyst under various conditions of Fischer-Tropsch
synthesis, Catal. Today 164 (2011) 62–67.
investigation of the hydrodeoxygenation of levulinic acid to
over Ru (0001), ACS Catal. 7 (2017) 215–228.
c-valerolactone
[51] O. Mamun, M. Saleheen, J.Q. Bond, A. Heyden, Importance of angelica lactone
formation in the hydrodeoxygenation of levulinic acid to -valerolactone over
a Ru (0001) model surface, J. Phys. Chem. C 121 (2017) 18746–18761.
c
211