Stabilized open metal sites in bimetallic metal-organic framework catalysts for hydrogen production from alcohols

Article abstract of DOI:10.1039/d1ta00222h

Liquid organic hydrogen carriers such as alcohols and polyols are a high-capacity means of transporting and reversibly storing hydrogen that demands effective catalysts to drive the (de)hydrogenation reactions under mild conditions. We employed a combined theory/experiment approach to develop MOF-74 catalysts for alcohol dehydrogenation and examine the performance of the open metal sites (OMS), which have properties analogous to the active sites in high-performance single-site catalysts and homogeneous catalysts. Methanol dehydrogenation was used as a model reaction system for assessing the performance of five monometallic M-MOF-74 variants (M = Co, Cu, Mg, Mn, Ni). Co-MOF-74 and Ni-MOF-74 give the highest H2 productivity. However, Ni-MOF-74 is unstable under reaction conditions and forms metallic nickel particles. To improve catalyst activity and stability, bimetallic (NixMg1-x)-MOF-74 catalysts were developed that stabilize the Ni OMS and promote the dehydrogenation reaction. An optimal composition exists at (Ni0.32Mg0.68)-MOF-74 that gives the greatest H2 productivity, up to 203 mL gcat-1 min-1 at 300 °C, and maintains 100% selectivity to CO and H2 between 225-275 °C. The optimized catalyst is also active for the dehydrogenation of other alcohols. DFT calculations reveal that synergistic interactions between the open metal site and the organic linker lead to lower reaction barriers in the MOF catalysts compared to the open metal site alone. This work expands the suite of hydrogen-related reactions catalyzed by MOF-74 which includes recent work on hydroformulation and our earlier reports of aryl-ether hydrogenolysis. Moreover, it highlights the use of bimetallic frameworks as an effective strategy for stabilizing a high density of catalytically active open metal sites. This journal is

Full text of DOI:10.1039/d1ta00222h

Journal of
Materials Chemistry A
View Article Online
View Journal | View Issue
PAPER
Stabilized open metal sites in bimetallic metal–
organic framework catalysts for hydrogen
production from alcohols†
Cite this: J. Mater. Chem. A, 2021, 9,
0869
1
a
b
c
a
Jonathan L. Snider, ‡ Ji Su,‡ Pragya Verma,
Farid El Gabaly,
a
b
a
a
Joshua D. Sugar, Luning Chen, Jeffery M. Chames,
Chaochao Dun, Jeffrey J. Urban,
Gabor A. Somorjai and Mark D. Allendorf
A. Alec Talin,
c
c
a
c
Vitalie Stavila,
David Prendergast,
bd
a
*
Liquid organic hydrogen carriers such as alcohols and polyols are a high-capacity means of transporting and
reversibly storing hydrogen that demands effective catalysts to drive the (de)hydrogenation reactions under
mild conditions. We employed a combined theory/experiment approach to develop MOF-74 catalysts for
alcohol dehydrogenation and examine the performance of the open metal sites (OMS), which have
properties analogous to the active sites in high-performance single-site catalysts and homogeneous
catalysts. Methanol dehydrogenation was used as
performance of five monometallic M-MOF-74 variants (M ¼ Co, Cu, Mg, Mn, Ni). Co-MOF-74 and Ni-
MOF-74 give the highest H productivity. However, Ni-MOF-74 is unstable under reaction conditions
and forms metallic nickel particles. To improve catalyst activity and stability, bimetallic (Ni Mg1ꢀx)-MOF-
4 catalysts were developed that stabilize the Ni OMS and promote the dehydrogenation reaction. An
a model reaction system for assessing the
2
x
7
optimal composition exists at (Ni0.32Mg0.68)-MOF-74 that gives the greatest H
2
productivity, up to
catꢀ
1
ꢀ1
ꢁ
2
ꢁ
203 mL g
min at 300 C, and maintains 100% selectivity to CO and H
between 225–275 C. The
optimized catalyst is also active for the dehydrogenation of other alcohols. DFT calculations reveal that
synergistic interactions between the open metal site and the organic linker lead to lower reaction
barriers in the MOF catalysts compared to the open metal site alone. This work expands the suite of
hydrogen-related reactions catalyzed by MOF-74 which includes recent work on hydroformulation and
our earlier reports of aryl-ether hydrogenolysis. Moreover, it highlights the use of bimetallic frameworks
as an effective strategy for stabilizing a high density of catalytically active open metal sites.
Received 9th January 2021
Accepted 13th April 2021
DOI: 10.1039/d1ta00222h
rsc.li/materials-a
7,8
frameworks (MOFs),
liquid–organic hydrogen carriers
Introduction
(LOHCs) can provide high reversible capacities, chemical
Hydrogen has long been regarded as a promising sustainable stability, low-cost transport, and would be compatible with
fuel and energy carrier; however, its direct utilization is a chal- existing infrastructure for fossil fuels as liquid energy
9
–12
lenge because of the low volumetric energy density of hydrogen vectors.
This concept involves storing hydrogen as hydrogen-
4,5
gas and high cost of storage and transportation. In contrast to rich liquid molecular carriers that are converted to H (gas) for
2
other hydrogen storage approaches, such as compressed gas or stationary applications by breaking chemical bonds, typically in
6
solid-state storage in metal hydrides and metal–organic the presence of a catalyst.
Current research efforts focus on developing efficient
carriers and advanced catalysts to increase reaction rates and
provide a more stable hydrogen generation and storage process.
a
Sandia National Laboratories, 7011 East Ave MS 9161, Livermore, CA, 94550, USA.
E-mail: mdallen@sandia.gov
b
To achieve a high hydrogen capacity, researchers have investi-
gated a number of promising LOHCs, including several alcohol-
Materials Sciences Division, Lawrence Berkeley National Laboratory, Building 66, One
Cyclotron Road, Berkeley, CA 94720, USA
c
13–15
Molecular Foundry, Lawrence Berkeley National Laboratory, Building 67, One based molecules: methanol (12.1 wt% H
2
),
ethylene glycol
16
17
Cyclotron Road, Berkeley, CA, 94720, USA
d
2
(6.5 wt% H ), and other alcohols. Although methanol is
Department of Chemistry, University of California – Berkeley, D56 Hildebrand Hall,
a high capacity hydrogen carrier, it requires the separation of
CO to produce high purity hydrogen gas, making it less attrac-
tive for use with fuel cells where even minor CO contamination
Berkeley, CA, 94720, USA
†
Electronic supplementary information (ESI) available: Additional experimental
methods, catalyst characterization, dehydrogenation results, methanol
conversion, and DFT calculations. See DOI: 10.1039/d1ta00222h
18
can lead to catalyst poisoning. Polyols, such as ethylene glycol,
could avoid these concerns by reacting to form pure gaseous
‡
J. L. S. and J. S. contributed equally to this work.
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. A, 2021, 9, 10869–10881 | 10869

View Article Online
Journal of Materials Chemistry A
Paper
16
hydrogen and a condensable hydrogen-lean byproduct. Ulti- OMS, such as oxygen-containing carboxylate groups, provides
mately, the selection of LOHC will be dependent on the needs of active atoms adjacent to the single metal atom that promote the
the end use application; however, each will require a catalyst for adsorption, activation, reaction, and desorption of molecules.
the cycling reactions.
Catalysts based on the MOF-74 structure have been demon-
44–48
Supported transition metal catalysts have been developed for strated for a variety of reactions, including C–H oxidation,
alcohol steam reforming, but poor hydrogen selectivity and NO
deactivation remain challenging.
4
9
50
51
x
reduction, hydroformulation, water dissociation, CO
2
19,20
52,53
54,55
Alternatively, acceptorless conversion,
and ether hydrogenolysis.
This demonstrates
dehydrogenation simplies the reaction chemistry by removing that the divalent OMS in MOF-74 has great potential to combine
21
the need for an external oxidant. However, the activity of the the best of homogeneous and heterogeneous catalysis by
supported transition metal catalysts is signicantly lower providing a tunable and specic catalyst with an extended and
2
2,23
without water in the feed.
High activity and selectivity can be uniform array of single metal sites.
Here, we demonstrate that OMS in the isostructural series M-
homogeneous metal complexes. These catalysts have disad- MOF-74 (M
(dobdc), where M ¼ Co, Cu, Mg, Mn, Ni and dobdc
achieved for acceptorless dehydrogenation through the use of
24
2
vantages as well, including high cost, poor stability, and diffi- ¼ 2,5-dioxido-1,4-benzenedicarboxylate) catalyze the dehydro-
cult reusability. To access the full hydrogen capacity under mild genation of alcohols. The MOF-74 topology has the highest-
reaction conditions, researchers are seeking catalysts with lower known volumetric density of OMS, and the nickel version,
2
+
costs while being highly active, selective, and stable.
which contains square pyramidal Ni OMS, is an unexplored
Coordinatively unsaturated metal sites, like the catalytically coordination environment for nickel-based dehydrogenation
active centers present in many metal complexes, have been catalysts. A key issue, however, is that Ni-MOF-74 has poor
56,57
shown to trigger different reaction pathways and improve thermal and structural stability.
To improve stability,
performance over traditional metallic catalysts. For example, bimetallic MOF-74 structures have previously been found to
2
+
58
the introduction of phosphate ligands to silica-supported Ni
stabilize Mg-MOF-74 against hydrolysis. We hypothesize that
improves selectivity toward propane dehydrogenation by stabi- a partial substitution of Ni into the nodes of Mg-MOF-74 will
2
+
2+
lizing the single-site Ni and preventing the formation of give rise to catalytically active Ni OMS while also maintaining
0
25
methane-selective metallic Ni . Similarly, homogeneous cata- structural stability, thereby maximizing activity and selectivity
lysts based on Ni complexes are found to be highly selective for alcohol dehydrogenation.
for the dehydrogenation of alcohols.
2
+
26,27
Toward the goal of creating an improved catalyst, single
metal site catalysts are attracting great interest due to their
atom-utilization efficiency, unique structure, and improved
Methods
(Ni Mg1ꢀx)-MOF-74 synthesis
x
28–30
performance.
However, because traditional heterogeneous
catalyst supports (e.g. metal oxides) can exhibit facets and M-MOF-74 (M
2
(dobdc) where M ¼ Ni, Mg, or Ni
x
Mg1ꢀx and
anchoring sites with vastly different surface energies, it is (dobdc) ¼ 2,5-dioxido-1,4-benzenedicarboxylate) was synthe-
difficult to achieve structural homogeneity or a high density of sized using a solvothermal reaction of metal salts with 2,5-
31–33
active sites.
Moreover, the lack of an ensemble of active dihydroxyterephthalic acid. This procedure is slightly modied
54
atoms adjacent to the metal atom in most single metal site from one reported previously. In a 250 mL glass bottle, 160 mL
catalysts prevents surface reactions that involve large molecules N,N-dimethylformamide (DMF), 8 mL ethanol, and 8 mL
34,35
or multiple adsorbates.
Finding an ideal support to permit deionized water were mixed. To this, 0.41 g (2.1 mmol) of 2,5-
a high loading of structurally homogeneous and highly active dihydroxytherephthalic acid was dissolved via sonication. Next,
single metal sites is one of the present challenges in heteroge- 6.2 mmol of a mixture of anhydrous magnesium(II) chloride and
neous catalysis.
anhydrous nickel(II) chloride was added (about a 50% excess
Metal–Organic Frameworks (MOFs) have emerged as relative to dobdc). The metal chlorides were added in the
a tunable platform to create coordinatively unsaturated single synthesis ratios prescribed in Table S4.† In all cases, the bottles
metal sites, or open metal sites (OMS), within a uniform porous were then tightly capped, and the mixtures were sonicated until
support. MOFs feature a well-dened structure, chemically and well dissolved. In some cases, the solutions remained turbid,
structurally uniform active sites, and rational design via judi- but turned clear upon mild heating. The bottles were heated in
3
6–41
ꢁ
cious choice of molecular building blocks.
within MOFs can be engineered as immobilized elements, on from the oven and cooled. A solid material was recovered via
Single sites an oven at 120 C. Aer 16 hours, the samples were removed
37
the node or linker, or as encapsulated species. Creating reac- centrifugation, washed with 50 mL DMF, and soaked for several
tive OMS in MOFs is among the most attractive strategies as hours. This washing procedure was repeated once more with
these can present similar local environments to those of high- DMF and three additional times with methanol. The solid was
42,43
performance homogeneous catalysts.
The long-range order then collected by centrifugation and dried under a nitrogen
of MOFs eliminates the structural variability present in tradi- ow. Aer drying, the MOF was activated under dynamic
ꢁ
tional solid-phase supports and offers precise control of vacuum at 170 C for 16 hours and transferred, air-free, to
chemical functionality to modulate the reaction selectivity and a nitrogen glovebox for storage. Synthetic methods used for the
delineate structure–property relationships. Furthermore, monometallic MOFs, similar to those above, are detailed in the
chemical functionalities in the local environment around the ESI, Section S1.†
10870 | J. Mater. Chem. A, 2021, 9, 10869–10881
This journal is © The Royal Society of Chemistry 2021

View Article Online
Paper
Journal of Materials Chemistry A
Catalytic testing
the peaks within a given spectral region, with multiple peaks
being included to t the complex satellite peak regions. Spectra
were calibrated to the adventitious C 1s peak (284.8 eV), which
allows for internal referencing of the binding energies.
The catalytic activity of all MOF-74 catalysts toward hydrogen
production from methanol was evaluated in a continuous ow
reactor. Methanol (CH
n-propanol (n-C OH, 99.7%), isopropanol (CH
9.5%), butyl alcohol (C OH, 99.5%), benzyl alcohol
CH OH, 99.8%), and ethylene glycol (CH (OH)CH
9.8%) were purchased from Sigma-Aldrich. All reagents were radiation generated at 40 kV and 44 mA. Measurements were
collected in air on powder samples that were spread onto glass
3
OH, 99.9%), ethanol (C
2
H
5
OH, 99.8%),
3
H
7
3
CH(OH)CH
3
,
X-ray diffraction (XRD)
9
4 9
H
(C
6
H
5
2
2
2
(OH), XRD patterns were collected on a Rigaku SmartLab using Cu Ka
9
used without further purication.
In a typical catalytic experiment, 100 mg of catalyst was rst slides. Scans were conducted between 4 and 80 degrees, with
mixed with 500 mg of glass beads (50–70 mesh particle size). a step size of 0.05 degrees, and a scan rate of about 5 degrees per
The mixture was packed in a xed-bed ow reactor with quartz minute.
wool plugs. Then methanol liquid was pumped into the heating
ꢀ
1
chamber with a feeding rate of 0.1 mL min . In the heating Nitrogen isotherms
chamber, methanol vapor mixes with carrier gas N (35
2
Isotherms and Brunauer, Emmett and Teller (BET) surface area
measurements were carried out at 77 K using either a Micro-
meritics ASAP 2020 porosimeter or a Quantachrome Autosorb
ꢀ1
mL min ) before entering the reactor tube. Pressure was
maintained at 1 atm. The temperature of the bed was monitored
by a K-type thermocouple, and the furnace was controlled by
a PID 679 controller. The products were analyzed on-stream by
a HP 5890 gas chromatograph (Hayesep D column and Hayesep
iQ porosimeter with liquid N cooling. Prior to each measure-
2
ꢁ
ment, samples were degassed at 170 C under vacuum for 16
hours.
2
Q column, N carrier gas) equipped with a thermal conductivity
detector and a ame ionization detector. Any liquid products
were also collected and measured by NMR (AV-700). Metrics for
catalytic performance were calculated as shown in the ESI,
Section S1.†
Aer catalysis, the catalyst bed was removed from the reactor
tube. To remove contamination by quartz wool and glass, the
samples were cleaned prior to further characterization. Samples
were rst sieved with a 140 mesh sieve to remove the grains of
sand, yielding a small amount of usable powder. When char-
acterization required additional powder, the sand grains were
washed with methanol to give a colored liquid which was dried
under N2 to yield additional spent catalyst. Since BET
Transmission electron microscopy (TEM)
TEM images were collected on a ThermoFisher Titan Themis Z
operated at 300 kV. The Themis Z is equipped with a probe Cs-
corrector and a large-angle four quadrant SuperX EDS detector
(
ꢂ0.7 sr). Energy dispersive spectrum (EDS) images were
collected with 10–20 ms pixel times, a probe current of approx-
imately 150–200 pA, and the application of automatic dri
correction. Samples were prepared in air by sonicating the
sample materials in methanol and drop casting the solution
onto Cu lacey carbon grids.
measurements would be especially sensitive to low surface area Scanning electron microscopy (SEM)
contaminants, such as the quartz wool and sand, the methanol-
washed material was used for N isotherms.
2
SEM images were collected with a JEOL JSM7600F Thermal
Field Emission Electron Microscope. EDS analysis was per-
formed with an Oxford Instruments X-MaxN 80 Detector
combined with Oxford's Aztec soware. Microscope conditions
for both imaging and EDS spectrum capture were: accelerating
voltage of 15 kV, probe current setting of 10, and a working
distance of 8 mm. Samples were prepared for SEM by pressing
the powders into indium foil, to avoid carbon contamination
from other adhesive techniques. Spectra were collected at
positions and magnications to yield sufficient powder thick-
ness to block signal from the underlying indium.
Sample collection was generally performed with air expo-
ꢁ
sure, except for the catalysts collected aer reaction at 250 C.
For these samples, the reactor tube was sealed under nitrogen
and opened in a nitrogen-containing glovebox. Samples were
mounted for TEM and XPS within the glovebox and transferred
air-free to the characterization instruments.
X-ray photoelectron spectroscopy (XPS)
XPS data was collected at Sandia National Laboratories at
ambient temperature, using samples adhered to carbon tape or
dropcast onto Au foil from a suspension in methanol. The
Theoretical approach
samples were mounted into a steel stage that was then affixed to Density functional theory (DFT) calculations were performed on
a custom-designed clean transfer system that allows air- the unit cell of M-MOF-74 with periodic boundary conditions as
59,60
sensitive samples to be sealed within either an inert atmo- implemented in Vienna Ab Initio Simulation Package (VASP).
sphere or vacuum, and which can be attached to the XPS UHV For the periodic model, a unit cell consisting of 54 atoms (3
chamber for insertion without exposure to air or moisture. The formula units of M (dobdc)), of which six are metal atoms, is
2
X-ray source was an Al K-alpha anode (PE ¼ 1486.7 eV). Analysis used. The metal atoms are arranged in two chains (Fig. S9†)
of all XPS spectra was performed using the CasaXPS peak-tting with three in each chain. The DFT methods used are the
61
soware, using Shirley backgrounds and appropriate generalized gradient approximation, PBE, the van der Waals
62
constraints for each element. FWHM were held constant across density functional, rev-vdW-DF2, and the +U corrected rev-
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. A, 2021, 9, 10869–10881 | 10871

View Article Online
Journal of Materials Chemistry A
Paper
63,64
vdW-DF2+U method, where on-site Hubbard U correction
is MOF-74 structure (Fig. S1†). Cu-MOF-74, Co-MOF-74, Mg-MOF-
applied to the open-shell 3d e–s of the transition metals, Mn, 74, and Ni-MOF-74 had BET surface areas ranging from 1050–
2
ꢀ1
Co, Ni, and Cu of M-MOF-74. The choice of +U is based on 1510 m g , whereas Mn-MOF-74 had a lower surface area of
65
2
ꢀ1-
published recommendations. For all the periodic DFT calcu- 490 m
g
. X-ray diffraction (XRD) conrmed the MOF-74
lations performed in this work, we optimized the nuclear structure in all ve samples (Fig. 1a). Small peak shis are
positions, cell shape, and unit cell volume using plane-wave observed from the simulated Mg-MOF-74 reference, most
66,67
basis sets and projector-augmented-wave potentials
notably in Mn-MOF-74, consistent with the expanded metal–
provided with the VASP package. These calculations were per- metal distances predicted by DFT (Table S1†) and reported for
7
1
formed using a k-point mesh of 4 ꢃ 2 ꢃ 2 was used to sample previously synthesized MOF-74 frameworks.
Given the
the Brillouin zone, a cutoff energy of 600 eV, a self-consistent constant topology of the ve porous MOF-74 structures
ꢀ
6

eld energy convergence criterion of 10
eV, and a force synthesized, the identity of the OMS is expected to drive
ꢀ
3
ꢀ1
˚
convergence criterion of –10 eV A
.
differences in catalytic performance.
To probe the mechanism of conversion of methanol to
The catalyst performance towards methanol dehydrogena-
ꢁ
carbon monoxide catalyzed by M-MOF-74, we used an 88-atom tion was evaluated in a ow reactor at 1 atm and up to 300 C.
68
cluster model with three metal atoms and six organic ligands The results, shown in Fig. 1b and c, reveal that Mg-MOF-74 and
Fig. S9†), carved from the aforementioned periodic models. Mn-MOF-74 are both inactive for the reaction under these
These were then optimized using either the rev-vdW-DF2 conditions. Cu-MOF-74 exhibited low H productivity (#42 mL
(
2
ꢀ
1
ꢀ1
method (for Mg-MOF-74 and Zn-MOF-74) or the rev-vdW- H2 gcat min ) and high dimethyl ether (DME) selectivity
DF2+U (for Mn-MOF-74, Co-MOF-74, Ni-MOF-74, and Cu-MOF- ($10.9%). Co-MOF-74 and Ni-MOF-74 are signcantly more
7
4) method. Only the central metal atom and the ve oxygen active than the other MOFs, exhibiting the highest H produc-
2
atoms of the cluster were optimized, with the remaining atoms tivity and CO selectivity of the series. The H
2
ꢁ
productivity of Co-
frozen. The cluster is terminated using H atoms and a high-spin MOF-74 increases from near zero at 200 C to 117.1 mL H
2
ꢀ
1
ꢀ1
ꢁ
ground state is used with the three metal ions of the cluster
g
cat
min at 300 C with 100% CO selectivity. Ni-MOF-74
ꢀ
1
ꢀ1
ferromagnetically coupled. For each species involved in the produces even more H
reaction – reactants, intermediates, transition structures, and at 300 C. However, this coincides with a loss in CO selectivity,
2
, as much as 158.1 mL H
2
g
cat min
ꢁ
ꢁ
products – we performed a partial optimization of the cluster, with 5.3% methane selectivity observed at 300 C. Comparing
whereby the central metal atom, its rst coordination sphere, Co-MOF-74 and Ni-MOF-74, the performance is similar until
ꢁ
and any adsorbate interacting with the central metal atom were about 250 C; at this temperature Ni-MOF-74 begins to catalyze
optimized. These calculations were performed with the PBE methane production. In contrast, Co-MOF-74 is nearly 100%
exchange-correlation functional and the def2-SVP or the def2- selective towards syngas (CO + H ). Due to the well-documented
2
6
9
72,73
TZVP basis set using the Gaussian 16 suite of quantum propensity of metallic nickel for methanation
and Ni-MOF-
70
56,58
mechanical programs.
74's poor thermal stability
(Ni nanoparticle formation
ꢁ
74
observed as low as 200 C under vacuum ), this nding
suggests that Ni-MOF-74 is degrading to form nickel
nanoparticles.
Results and discussion
Monometallic M-MOF-74 catalysts
To test this hypothesis, samples of Co-MOF-74 and Ni-MOF-
4 were characterized before and aer performing catalysis at
50 C for one hour. Under these conditions, the Co-MOF-74
catalyst was stable. Prior to reaction, transmission electron
microscopy (TEM) images reveal a rod-like morphology with
7
2
To evaluate the methanol dehydrogenation performance of M-
MOF-74, a series of ve catalysts were prepared (M ¼ Co, Cu,
Mg, Mn, and Ni) by solvothermal synthesis. Most of the catalyst
series was found to feature the high surface areas typical of the
ꢁ
Fig. 1 (a) Normalized XRD patterns of the ‘as synthesized’ M-MOF-74 catalyst series, compared against a simulated Mg-MOF-74 reference
1
(
CSD-VOGTIV). Catalytic performance of the synthesized M-MOF-74 catalysts towards methanol dehydrogenation: (b) hydrogen productivity
normalized to the total mass of catalyst and (c) product selectivity towards CO (solid circles), DME (open triangles) and CH (solid squares).
4
ꢀ1
2
Reaction conditions: 0.1 mL min methanol, 35 sccm N , 100 mg catalyst, 1 atm.
10872 | J. Mater. Chem. A, 2021, 9, 10869–10881
This journal is © The Royal Society of Chemistry 2021

View Article Online
Paper
Journal of Materials Chemistry A
ꢁ
Fig. 2 Characterization of Co-MOF-74 before and after performing methanol dehydrogenation for 1 hour at 250 C and 1 atm. Dark field TEM
images and TEM-EDS maps of the catalyst (a) before testing and (b) after testing, showing cobalt (red), carbon (blue), and oxygen (green). (c)
2+
Normalized Co 2p3/2 XPS spectra of the catalyst before and after testing. Experimental data (grey points) fit with a Co peak (green) and several
satellite peaks (blue) to create the envelope shown (black line). (d) Normalized XRD patterns of the catalyst before and after testing, compared to
2
a simulated Co-MOF-74 reference (CSD-SATNOR).
homogeneous composition throughout (Fig. 2a), consistent methanol dehydrogenation, as shown in Fig. 3b by the
75,76
with previous reports on Co-MOF-74.
A small number of appearance of bright spots in the dark eld TEM and green
clusters with high oxygen content are also observed. Similar clusters in the EDS mapping. These particles were found to
morphology and composition distributions are observed aer agglomerate into large clusters of nickel. Particle formation is
catalysis testing (Fig. 2b). X-ray photoelectron spectroscopy also evidenced by XPS data showing that a shoulder forms on
2
+
2+
(
XPS) indicates that the Co nodes in the MOF (780.7 eV) are the Ni peak of the MOF (856.3 eV) aer catalysis that is
0
unchanged by the reaction (Fig. 2c). Furthermore, the X-ray assigned to metallic Ni at 852.8 eV (Fig. 3c). Comparing the
2
+
0
diffraction (XRD) patterns obtained before and aer reaction areas of the Ni and Ni features, ꢂ18% of the nickel is reduced
ꢁ
indicate that the crystal structure is also intact (Fig. 2d). aer one hour on stream at 250 C. Post-catalysis XRD reveals
ꢁ
Together, the characterization identies Co-MOF-74 as a robust that a broad Ni(111) peak at 44.6 has also formed (Fig. 3d),
material for methanol dehydrogenation at 250 C.
ꢁ
consistent with the small nanoparticles observed in TEM. A
ꢁ
In contrast, Ni-MOF-74 is unstable under the catalysis featureless band is also observed between 10–40 , consistent
conditions, consistent with prior results on Ni-MOF-74 with the formation of amorphous carbon species upon frame-
74
stability. This is rst apparent visually via a darkening of the work degradation. The presence of amorphous carbon is further
initial yellow powder to become brown and eventually black supported by signicant broadening of the Raman bands of the
ꢀ
1
(
Fig. S2†). The initial ake-like morphology of the MOF (Fig. 3a) tested sample between 1200–1700 cm (Fig. S3†). TEM and
ꢁ
is dotted with small nanoparticles of nickel aer performing PXRD data obtained from a sample aer catalysis at 300
C
ꢁ
Fig. 3 Characterization of Ni-MOF-74 before and after performing methanol dehydrogenation for 1 hour at 250 C and 1 atm. Dark field TEM
images and TEM-EDS maps of the catalyst (a) before testing and (b) after testing, showing carbon (blue) and nickel (green). (c) Normalized Ni 2p3/2
0
2+
XPS spectra of the catalyst before and after testing. Experimental data (grey points) fit with a Ni peak (red), Ni peak (green), and several satellite
peaks (blue) to create the envelope shown (black line). (d) Normalized XRD patterns of the catalyst before and after testing, compared to
3
ꢁ
a simulated Ni-MOF-74 reference (CSD-LECQEQ). Inset plot showing the Ni(111) peak observed in the sample after reaction at 250 C.
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. A, 2021, 9, 10869–10881 | 10873

View Article Online
Journal of Materials Chemistry A
Paper
reveals that the decomposition is even more severe at higher as our results (Fig. 1b) demonstrate that Mg-MOF-74 is cata-
temperatures, forming large nickel agglomerates and sharp lytically inactive. Finally, the propensity of metallic Ni for
Ni(111) peaks (Fig. S4†). As will be discussed later, the observed methanation makes methane selectivity a useful diagnostic for
framework degradation correlates with a signicant deactiva- particle formation and MOF stability.
tion of the catalytic activity aer six hours on stream (Fig. S5†).
x
A series of ve bimetallic (Ni Mg1ꢀx)-MOF-74 catalysts were
Consequently, a stabilization strategy must be developed if the synthesized using a modied one-pot solvothermal synthesis
highly active Ni OMS of Ni-MOF-74 is to be used for methanol method whereby the ratio of Ni to Mg precursors was varied
dehydrogenation.
from 0.01 to 0.6. A 50% excess of the metal precursor relative to
the organic linker was used to ensure full conversion of the
linker molecule. This altered the elemental composition ob-
tained in the nal product as the Ni OMS was found to be
preferentially incorporated into the bimetallic framework
Bimetallic (Ni Mg )-MOF-74 catalysts
x
1ꢀx
Building upon these results, we sought to stabilize the Ni OMS
and determine whether its already high activity could be
improved. Since the stability of the MOF-74 framework under
the reaction conditions varies with the metal used, we hypoth-
esized that a partial substitution of the unstable Ni OMS into
a more stable framework would inhibit the precipitation of
metallic nickel nanoparticles. Mixed-metal MOFs have been
previously investigated for a variety of applications, including
catalysis, in which they prove to be a highly tunable approach to
improve performance.
as the target system due to several key advantages. First, Mg-
MOF-74 features much greater thermal and chemical stability
than any of the other versions of MOF-74, up to 400 C.
Second, activity arising from the Ni OMS can be readily assessed
(Table S2†). For example, when a Ni : Mg ratio of 0.25 was used
for solvothermal synthesis, the nal product had a Ni : Mg ratio
of 0.47, as determined by scanning electron microscopy and
energy dispersive spectroscopy (SEM-EDS) and checked with
inductively coupled plasma-optical emission spectroscopy (ICP-
OES). This is consistent with preferential incorporation into the
MOF structure of the excess NiCl in solution. This phenom-
2
2
+
enon can be understood by the higher stability of Mg in
58,77–81
For this study, Ni–Mg was selected
2
+
2+
solution compared to Ni . The Mg ion has a standard heat of
ꢀ
1
ꢀ1
formation of ꢀ462 kJ mol in aqueous media vs. ꢀ64 kJ mol
2+
84
for Ni . A preferential uptake of metals has been observed
ꢁ
82,83
previously in multimetallic MOFs made via one-pot
Fig. 4 (a) Normalized XRD patterns of the ‘as synthesized’ (Ni
reference (CSD-VOGTIV). Catalytic performance of the synthesized (Ni
x
Mg1ꢀx)-MOF-74 catalyst series, compared against a simulated Mg-MOF-74
1
x
Mg1ꢀx)-MOF-74 towards methanol dehydrogenation: (b) hydrogen
productivity normalized to the total mass of catalyst, (c) product selectivity towards CO (solid circles), DME (open triangles) and CH
4
(solid
ꢁ
ꢀ1
squares), and (d) stability of methanol dehydrogenation at 250 C for Ni-MOF-74 and (Ni0.32Mg0.68)-MOF-74. Reaction conditions: 0.1 mL min
2
methanol, 35 sccm N , 100 mg catalyst, 1 atm.
10874 | J. Mater. Chem. A, 2021, 9, 10869–10881
This journal is © The Royal Society of Chemistry 2021

View Article Online
Paper
Journal of Materials Chemistry A
58,85
ꢁ
The nal metal ratios of the catalyst series, as 275 C. By the previously established correlation between Ni
syntheses.
listed in Table S2,† were determined by SEM-EDS and are used particle formation and methane selectivity, we infer that this
in naming the catalyst samples henceforth. A discussion and MOF composition yields the greatest number of active Ni OMS
comparison of the SEM-EDS and ICP-OES results can be found and maintains high stability within the Mg framework.
in the ESI (Section S2†).
Post-catalysis characterization of the (Ni0.32Mg0.68)-MOF-74
The MOF-74 structure was conrmed by PXRD for the cata- catalyst conrms that the bimetallic framework stabilizes the
lyst series (Fig. 4a). No peaks corresponding to metallic nickel Ni OMS. Characterization was performed on the bimetallic
ꢁ
ꢁ
(
e.g. Ni(111) at 44.6 ) were observed. Minor contractions in the catalyst before and aer one hour of testing at 250 C, the
MOF-74 structure were observed in some samples as the nickel temperature at which Ni-MOF-74 was previously found to have
fraction increased, particularly for (Ni0.76Mg0.24)-MOF-74. This signicant particle formation. As was seen with Ni-MOF-74, the
is consistent with the subtle decrease in metal–metal distance “as synthesized” (Ni0.32Mg0.68)-MOF-74 takes the form of akes
predicted by DFT for Ni-MOF-74 relative to Mg-MOF-74. with uniform elemental distribution (Fig. 5a). Aer catalysis at
ꢁ
However, because the “as synthesized” PXRD was collected on 250 C, the morphology of (Ni0.32Mg0.68)-MOF-74 is unchanged,
air-exposed samples, this effect is masked by expansion of the with no evidence for particle formation observed via TEM
frameworks upon adsorption of guest molecules, such as (Fig. 5b). Additional conrmation is provided by XPS analysis
moisture from the air or residual methanol (Table S1†). N
isotherms were also collected for the bimetallic catalyst series (Ni0.32Mg0.68)-MOF-74 catalyst aer testing at 250 C. The
2
(Fig. 5c). Spectra in the Ni 2p region show no change to the
ꢁ
2
+
and again conrm that high BET surface areas were achieved, spectra can be t solely with the MOF-74 Ni peak at 855.8 eV
2
ꢀ1
ranging from 910–1510 m g (Fig. S6†). Together, these results and its satellite peaks. The Mg 1s region is also unchanged at all
establish that our synthetic method produces the MOF-74 temperatures tested (Fig. S7†). Raman spectroscopy of the
topology across the composition series.
sample aer one hour of catalysis is consistent with that of the
The results of methanol dehydrogenation experiments starting material, suggesting that the framework has not
across the (Ni
x
Mg1ꢀx)-MOF-74 catalyst series are presented in degraded (Fig. S3†). Finally, MOF-74 remains the sole crystalline
ꢁ
ꢁ
Fig. 4b–d and show that, between 200-250 C, the highest H
2
phase observed by XRD aer testing at 250 C (Fig. 5d). At the
ꢁ
productivity is found when using MOFs with Ni fractions of x ¼ higher reaction temperature of 300 C, the bimetallic catalyst
0.32 or higher. At higher temperatures, H productivity degrades, forming Ni nanoparticles (Fig. S8a–d†). This coin-
2
increases up to x ¼ 0.32, at which point productivity decreases cides with the methane selectivity of (Ni0.32Mg0.68)-MOF-74
with increasing Ni fraction (Fig. 4b). Excellent CO selectivity, increasing to 0.2%, reinforcing the utility of methane selec-
greater than 99.5%, is observed at all temperatures for Ni frac- tivity as a diagnostic. Despite the appearance of metallic Ni, the
tions up to x ¼ 0.32. As the Ni fraction is increased above x ¼ bimetallic framework remains stable enough to maintain a high
2
ꢀ1
ꢁ
0
.32, DME selectivity and methane selectivity both increase with BET surface area of 850 m g even aer testing at 300 C
increasing nickel content. DME selectivity decreases with (Fig. S8e†). Overall, post-reaction characterization reveals that
increasing temperature, whereas methane selectivity increases embedding the Ni OMS within a stable Mg-MOF-74 framework
with increasing temperature (Fig. 4c). The increasing methane increases the stability of the reactive Ni OMS, creating an
selectivity of the Ni-rich compositions is indicative of nickel improved catalyst for methanol dehydrogenation with high
particle formation in those catalysts. As a result, an optimal productivity and up to 100% CO selectivity.
ꢁ
composition exists at (Ni0.32Mg0.68)-MOF-74 that gives the
A continuous stability test at 250 C, the temperature at
ꢀ
1
ꢀ1
ꢁ
greatest H productivity, up to 203 mL gcat min at 300 C, which selectivity is maximized across the series, clearly reveals
2
while also maintaining 100% selectivity to syngas between 225– the increased stability achieved by the bimetallic catalyst
ꢁ
Fig. 5 Characterization of (Ni0.32Mg0.68)-MOF-74 before and after performing methanol dehydrogenation for 1 hour at 250 C and 1 atm. Dark
field TEM images and TEM-EDS maps of the catalyst (a) before testing and (b) after testing, showing magnesium (red), carbon (blue), and nickel
2+
(
green). (c) Ni 2p3/2 XPS spectra of the catalyst before and after testing. Experimental data (grey points) fit with a Ni peak (green) and several
satellite peaks (blue) to create the envelope shown (black line). (d) XRD patterns of the catalyst before and after testing, compared to a simulated
1
Mg-MOF-74 reference (CSD-VOGTIV).
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. A, 2021, 9, 10869–10881 | 10875

View Article Online
Journal of Materials Chemistry A
Paper
(Fig. 4d). Comparing the most active catalyst ((Ni0.32Mg0.68)- the reaction occurs primarily at the surface of the MOF parti-
MOF-74) with monometallic Ni-MOF-74, we nd that the cles, the computed TOF values would be a lower limit of true
stability of the bimetallic catalyst is signicantly higher than TOF of the OMS.
that of the monometallic version. The bimetallic catalyst loses
The comparison in Fig. 6 with RANEY® Ni, which we tested
29% of its initial activity over 24 hours, whereas Ni-MOF-74 under equivalent conditions, demonstrates that the optimal
loses 42% over just six hours. Very high selectivity is also MOF is much more selective than this commercial catalyst.
maintained in the bimetallic case with CO selectivity remaining RANEY® Ni is pyrophoric upon drying, so care was taken to
at or above 99.6% over 24 hours. This is in stark contrast to Ni- handle the material under a N atmosphere while loading and
2
MOF-74, for which the CO selectivity drops to 94.2% aer only unloading the reactor. The Ni-normalized activity of
six hours due to methane selectivity rising to 5.8%. The gradual (Ni0.32Mg0.68)-MOF-74 is comparable to RANEY® Ni at high
degradation of the bimetallic catalyst correlates with the loss of temperatures. As shown in Table S3,† the TOF of RANEY® Ni is
2
+
ꢀ1
ꢁ
ꢀ1
the Ni OMS. Aer one hour, when performance is still at its estimated to be 36.5–41 h at 250 C compared to 31.4 h for
maximum, TEM-EDS, XPS, XRD, and Raman characterizations the bimetallic MOF. However, the OMS of the MOF catalyst has
do not reveal any evidence for framework degradation (Fig. 5). a signicant selectivity advantage when compared to the
Aer 24 hours on stream, the performance has degraded and metallic sites in RANEY® Ni. (Ni0.32Mg0.68)-MOF-74 has
0
ꢁ
ꢁ
characterization by XPS and Raman reveals the formation of Ni
and amorphous carbon respectively (Fig. S9 and S3†).
a methane selectivity of 0% at 250 C and 0.2% at 300 C. In
contrast, RANEY® Ni has 15.3% and 23% methane selectivity,
As our measurements show that the Mg OMS is inactive for respectively, at those conditions (Fig. S10†).
methanol dehydrogenation, the Ni OMS can be identied as the
The (Ni0.32Mg0.68)-MOF-74 catalyst is also active for the
active site for the bimetallic catalyst series. To examine the dehydrogenation of other hydrogen carriers. We evaluated
intrinsic activity of this site, the hydrogen productivity was (Ni0.32Mg0.68)-MOF-74 for hydrogen production from a variety of
normalized to the nickel content of each catalyst as determined alcohol-based carriers: ethylene glycol, ethanol, 1-propanol, and
by SEM-EDS. Even when normalized to Ni, the (Ni0.32Mg0.68)- 2-propanol. As shown in Fig. 7, the catalyst produces hydrogen
MOF-74 catalyst still provides the greatest H
series, 1501 mL H g
2
2
productivity of the at high rates from all of these molecules. This versatility creates
Ni min (Fig. 6). On the basis of meth- the potential for situations in which the selection of carrier
anol consumed, the turnover frequency (TOF) of the Ni OMS in varies, for example depending on cost, seasonal availability, or
ꢀ1
ꢀ1
ꢀ1
ꢁ
Ni0.32Mg )-MOF-74 is calculated to be 31.4 h at 250 C and materials compatibility.
0.68
(
ꢀ
1
ꢁ
105 h at 300 C, competitive with many transition metal
catalysts (Table S3†). Two assumptions were made in
computing this value: (1) the Ni OMS is the active site of the
reaction and (2) all reactive nickel is in the form of an OMS. If
the amount of accessible reactive Ni is less than this, as, for
example, the result of the observed degradation over time or if
DFT investigation of MOF-74 catalysts
To understand the experimental trends and gain insight into
how the OMS participates in the reaction, we performed
a computational study for a series of M-MOF-74 structures (M ¼
Co, Cu, Mg, Mn, Ni, Zn). Density functional theory (DFT)
calculations were applied to two models, as shown in Fig. S11:†
Fig. 6 Hydrogen productivity of the catalyst series for methanol
dehydrogenation, normalized to the nickel content as determined by
SEM-EDS. Commercial RANEY® Ni, dried under vacuum and handled Fig. 7 Hydrogen productivity, averaged over 1.5 hours, from alcohol-
under N , used as a nickel standard and normalized to its total mass. based hydrogen carriers using (Ni0.32Mg0.68)-MOF-74 catalyst. Reac-
2
ꢀ1
ꢀ1
ꢀ1
Reaction conditions: 0.1 mL min methanol, 35 sccm N
MOF catalyst (45 mg RANEY® Ni), 1 atm.
2
, 100 mg tion conditions: 0.1 mL min liquid feed (0.05 mL min for ethylene
ꢁ
glycol), 100 mg catalyst, 35 sccm N
2
, 1 atm, 250 C.
10876 | J. Mater. Chem. A, 2021, 9, 10869–10881
This journal is © The Royal Society of Chemistry 2021

View Article Online
Paper
Journal of Materials Chemistry A
ꢀ
1
ꢀ1
a periodic model with 54 atoms (three formula units of Co and Ni, ꢀ18.8 kcal mol and -15.8 kcal mol respectively,
ꢀ
1
M (dobdc)) and a cluster model with 88-atoms constrained at its are about ꢂ10 kcal mol stronger than the other metals.
2
periphery to reproduce the local structure within the periodic
The cluster model was then used to investigate the reaction
68
model. The cluster model allows calculations to be run on energetics and determine whether the linker plays a role in the
a system that is small enough to perform accurate DFT calcu- catalysis. Two reactions were considered: dehydrogenation of
lations while still being large enough to capture the behavior of methanol to formaldehyde (reaction (1)) and the subsequent
the full catalyst structure (see Table S4† for a comparison dehydrogenation of formaldehyde to carbon monoxide (reac-
86
between cluster and periodic models).
tion (2)).
We rst computed the binding energy of several adsorbates
for each M-MOF-74 structure using the periodic model at 0 K.
The results indicate that all metals readily absorb methanol,
with exoergic adsorption energies between ꢀ11.6 and ꢀ21.7 kca
2
r
98K
¼ 21 kcal molꢀ1
CH OH 4 HCHO + H , DH
(1)
(2)
3
2
2
98K
¼ 2 kcal molꢀ1
HCHO 4 CO + H
2
, DH
r
ꢀ
1
mol
(Table S5†). This behavior is consistent with prior
For the species involved in each reaction – reactants, inter-
mediates, transition structures, and products – a partial opti-
mization of the cluster was performed in which the positions of
the adsorbate, the active metal site, and the ve linker oxygen
atoms in the rst coordination sphere of the OMS were opti-
mized. Frequency calculations were performed in the optimized
degrees of freedom to include temperature effects and the
enthalpies were calculated by adding the zero-point vibrational
observations of methanol adsorption in MOF-74 frame-
51,75
works.
In all cases, the lowest energy conguration for
methanol was found to be binding via the oxygen atom. The
cluster model was also used to predict the energetics of CO
binding at the OMS, which is also predicted to be exoergic,
regardless of whether the CO is bound via carbon or oxygen
(Table S6†). In all cases, binding via the carbon atom yielded the
lowest energy conguration. Notably, the binding energies to
ꢀ1
Fig. 8 Gibbs free energy profile (DG298.15 in kcal mol ) of the conversion of methanol into carbon monoxide and hydrogen on Ni-MOF-74
calculated using PBE/def2-TZVP. The DG298.15 values of each species are relative to the MOF + MeOH complex. For visual clarity, only the first
coordination sphere of Ni-MOF-74 and key bond distances (in A˚ ) are shown for each structure of the catalytic cycle. As shown in the full cluster
model (inset), the linker molecules provide the O atoms shown. TS0 (not given in Table 1) is the first transition structure of the catalytic cycle that
involves the transfer of a H atom from methanol to the O atom of the linker. TSrot (not given in Table 1) is the transition structure for the rotation of
HCHO. For some of the structures more than one conformer were located, which differ in energy by a few kcal mol. [(Color coded such that Ni ¼
gray, O ¼ red, C ¼ brown, and H ¼ white).] Transition structures for other M-MOF-74 structures can be found in Fig. S12.†
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. A, 2021, 9, 10869–10881 | 10877

View Article Online
Journal of Materials Chemistry A
Paper
energy and thermal contributions at 298.15 K and 1 atm pres- CO is endergonic. The computed barriers for reaction (2) are
sure. The intermediate structures were characterized by the similar in magnitude across the metal series, with Cu and Mg
ꢀ
1
presence of only real frequencies and the transition structures giving the highest barriers at 63.1 kcal mol
and
ꢀ1
were characterized by the presence of one imaginary frequency. 61.4 kcal mol respectively and Ni and Co having the lowest
ꢀ
1
ꢀ1
This analysis provided the transition state for specic steps barriers at 47.3 kcal mol and 48.3 kcal mol , respectively.
along the dehydrogenation pathway. Given the large multi- Based on these results we expect that reaction (2) will be rate-
plicity of possible landing points for reactants and products in limiting, requiring high temperatures to achieve a signicant
the MOF framework around each metal site, we limited our turnover. The experimental reaction results weakly correlate
analysis to selected key transition states, shown in the free with the activation barriers shown in Table 1. The catalysts with
energy diagram in Fig. 8 for Ni-MOF-74. Transition states for the highest barriers (Cu-MOF-74, Mg-MOF-74) have little to no
other M-MOF-74 structures can be found in the ESI (Fig. S12 activity and those with the lowest barriers (Co-MOF-74, Ni-MOF-
and S13†).
74) being the most active of the series. The hydrogen produc-
When the reaction proceeds via the OMS alone (i.e. without tivity also correlates with the calculated CO binding energy, with
adsorbate interaction with the linker oxygen atom), highly the strongest binding Ni-MOF-74 and Co-MOF-74 catalysts
unfavorable reaction energies (Table 1) are observed, with giving the greatest hydrogen productivity of the series.
‡
ꢀ1
barrier heights (DG ) >70 kcal mol for both reaction steps
The free energy diagram in Fig. 8 provides a mechanistic
across all the metals surveyed. These energies are comparable to picture of the reaction using Ni-MOF-74 as a representative
the uncatalyzed (gas-phase) reaction. However, if the adsorbate example. Before reaction (1), there is rst adsorption of meth-
is permitted to interact with both the OMS and the neighboring anol onto the metal site via its oxygen atom, followed by transfer
organic linker via its oxygen atom, the transition state orienta- of a hydrogen atom to a neighboring linker oxygen atom (TS0).
tions shown in Fig. 8 are obtained. The resulting barrier heights The methoxy adsorbate on the OMS is bent over toward the
ꢀ1
are dramatically reduced to a few kcal mol for reaction (1) and linker oxygen (TS1), giving a geometry that facilitates the release
ꢀ
1
4
8–64 kcal mol for reaction (2). The high barriers for reaction of H
2
by a heterolytic recombination (proton from the linker
(2) are consistent with the elevated temperatures required in the and hydride from the methyl) to produce formaldehyde (reac-
catalysis tests. Additional information, including reaction tion (1)). The formaldehyde is adsorbed to the metal site via its
enthalpies (Table S7†) and visualizations of the reduced barrier oxygen atom and undergoes rotation (TSrot), leading to a struc-
heights (Fig. S14†), can be found in the ESI.†
ture in which the hydrogen atom interacts with the metal site.
ꢀ1
Looking closer at the individual reaction steps, the linker- This rotation, with a barrier of 7.1 kcal mol
is not rate-
assisted dehydrogenation of methanol to formaldehyde is limiting. Finally, the oxygen atom of the linker is again able
found to be exergonic and exhibits barriers less than to assist by interacting with a formaldehyde hydrogen (TS2),
ꢀ
1
1
2 kcal mol for all metals surveyed (Table 1). Ni and Mg have facilitating the release of carbon monoxide and the recombi-
ꢀ
1
ꢀ1
the lowest barriers at 1.9 kcal mol
respectively. In contrast, the subsequent dehydrogenation to unit of H
and 1.3 kcal mol
,
nation of the two hydrogen atoms at the OMS to form the nal
(reaction (2)). This is consistent with prior work
showing that linker functionalities can play an active role in
2
87
catalysis within MOFs, such as controlling product selectivity
‡
ꢀ1 a
88,89
Table 1 Barrier heights (DG , kcal mol
)
and reaction energies or enantioselectivity.
DG, kcal mol ) of (1) the dehydrogenation of methanol to formal-
ꢀ
1 b
(
dehyde and (2) the dehydrogenation of formaldehyde to carbon
monoxide for the reaction catalyzed by each M-MOF-74 and refer-
enced against the uncatalyzed reaction. Lower barrier heights are
observed when the reaction at the OMS is assisted by the linker, as
Conclusions
In this report we describe MOF-based catalysts developed for
shown in Fig. 8. Gibbs free energies calculated by PBE/def2-TZVP hydrogen production from alcohol-based hydrogen carriers
using the cluster model
using a combined experimental and theoretical approach. Five
monometallic MOF-74 catalysts were evaluated for methanol
dehydrogenation. Co-MOF-74 and Ni-MOF-74 were found to be
Catalysis at unassisted Catalysis at OMS
OMS
assisted by linker
ꢁ
ꢀ1
ꢀ1
2
the most active at 300 C (117.1 and 158.1 mL H gcat min
‡
‡
‡
1
‡
Metal in M-MOF-74 DG
1
DG
1
DG
2
DG
2
DG
DG
1
DG
2
DG
2
respectively). For Ni-MOF-74, however, decreasing activity and
increasing methane selectivity were observed over time due to
framework degradation and formation of metallic nickel parti-
cles. To improve performance and stabilize the Ni OMS, bime-
tallic frameworks were synthesized with partial substitution of
Ni OMS into the inactive but stable Mg-MOF-74 framework.
Cu
Co
Mg
Mn
Ni
Zn
75.3 18.9 73.1 2.7 11.5 ꢀ20.7 63.1 4.2
71.8 18.1 78.9 6.1 11.2 ꢀ17.9 48.3 5.3
68.5 22.2 80.0 8.4 1.3
76.7 20.6 78.6 7.5 5.0
77.2 20.6 81.4 10.3 1.9
76.9 18.3 78.3 5.0 3.6
ꢀ29.0 61.4 3.9
ꢀ22.7 49.9 3.3
ꢀ26.8 47.3 4.9
ꢀ24.7 57.0 4.2
Bimetallic
(Ni
x
Mg1ꢀx)-MOF-74
catalysts
demonstrated
Uncatalyzed
80
13
74
ꢀ1
—
—
—
—
improved activity, selectivity and stability compared to Ni-MOF-
a
The barrier heights are calculated with respect to the immediate 74. An optimal composition was found at (Ni Mg0.68)-MOF-
0
.32
ꢁ
b
intermediate preceding the transition structure. Reactant adsorption
ꢀ1
ꢀ1
7
4, with up to 203 mL H g
min at 300 C and 100%
selectivity to syngas between 225-275 C. TEM, XPS, XRD, and
2
cat
and product desorption are not considered in the reaction energies.
The reaction energies are calculated as the difference in energies
between the product complex and the reactant complex.
ꢁ
Raman characterization conrm that the (Ni0.32Mg0.68)-MOF-74
10878 | J. Mater. Chem. A, 2021, 9, 10869–10881
This journal is © The Royal Society of Chemistry 2021

View Article Online
Paper
Journal of Materials Chemistry A
catalyst improves the stability of the Ni OMS. The bimetallic
catalyst is also found to be broadly active for alcohol dehydro-
genation and compatible with a number of hydrogen carriers.
Author contributions
J. L. S. conducted synthesis, characterization, and analysis. J. S.
The Ni-normalized activity of the bimetallic catalyst series and L. C. conducted catalytic testing. P. V. performed DFT
was maximized at (Ni0.32Mg0.68)-MOF-74. At high temperatures, calculations. F. E. G. performed XPS characterization. J. D. S.
the Ni-normalized activity decreased with more Ni-rich performed TEM characterization. J. M. C. performed SEM
compositions, consistent with the increased methane selec- characterization. C. D. performed ICP-OES analysis. A. A. T.
tivity and framework instability observed in those samples. performed Raman characterization. J. L. S., J. S., V. S., and M. D.
Conversely, the intrinsic activity increased from x ¼ 0.01 to x ¼ A. discussed and planned the project. D. P., G. A. S., and M. D. A.
0.32 indicating that composition tuning has promotional supervised the project. J. L. S. wrote the manuscript including
effects beyond simply stabilizing the framework; highlighting signicant contributions from J. S., P. V., and V. S. All authors
the tunability of MOF catalysts. However, even though activity discussed the results and reviewed the nal manuscript.
was maximized at the (Ni0.32Mg0.68) composition, gradual
degradation was still observed in the catalyst with high
temperatures and extended time on stream. While the stability
Conflicts of interest
of the bimetallic catalyst is a marked improvement over the Ni- There are no conicts to declare.
MOF-74 catalyst and demonstrates the effectiveness of this
strategy, additional stabilization is still required. Future work is
needed to understand the tunability of these bimetallic frame-
Acknowledgements
works and improve their stability.
The TOF for methanol dehydrogenation at the Ni OMS in Materials-Advanced Research Consortium (HyMARC), estab-
The authors gratefully acknowledge support from the Hydrogen
ꢀ
1
ꢁ
(
1
Ni0.32Mg0.68)-MOF-74 was estimated to be 32 h at 250 C and lished as part of the Energy Materials Network under the U.S.
ꢀ1
ꢁ
05 h at 300 C. This value assumes all of the nickel is Department of Energy, Office of Energy Efficiency and Renew-
accessible to the reaction. Consequently, the computed TOF able Energy, Hydrogen and Fuel Cell Technologies Office, under
should be considered a lower limit. In future work we plan to Contract Number DE-AC04-94AL85000. Sandia National Labo-
use the tunable pore diameter of the structurally analogous ratories is a multimission laboratory managed and operated by
IRMOF-74(n) series to determine whether mass transport limi- National Technology and Engineering Solutions of Sandia,
tations within the pore network are limiting the reaction LLC., a wholly owned subsidiary of Honeywell International,
kinetics.
Inc., for the U.S. Department of Energy's National Nuclear
Our DFT calculations for monometallic M-MOF-74 catalysts Security Administration under contract DE-NA-0003525. We
allow us to propose a reaction mechanism in which interactions thank High Performance Computing Services at Lawrence Ber-
between reacting species coordinated to the Ni OMS and the keley National Laboratory and Minnesota Supercomputing
adjacent oxygen atom of the organic linker signicantly reduce Institute at the University of Minnesota for computational
the reaction barriers. When using this linker-assisted mecha- resources. Work at the Molecular Foundry was supported by the
nism, Ni-MOF-74 was found to have the lowest barrier for the Office of Science, Office of Basic Energy Sciences, of the U.S.
rate-limiting step of the reaction (the dehydrogenation of Department of Energy under Contract No. DE-AC02-
formaldehyde). Given the large number of potential landing 05CH11231. The authors gratefully acknowledge both Dr Mat-
points for reaction species in the MOF, however, we do not thijs van Spronsen at Lawrence Berkeley National Laboratory
consider this mechanism denitive. The lack of detectable and Dr Michael E. Foster for their time and fruitful discussions
formaldehyde in the product stream and the observed correla- of the project. The views and opinions of the authors expressed
2
tion between strong CO binding and high H productivity also herein do not necessarily state or reect those of the United
could be consistent with alternative methanol dehydrogenation States Government or any agency thereof. Neither the United
pathways. Future studies will examine the full range of path- States Government nor any agency thereof, nor any of their
ways for methanol dehydrogenation, including mechanisms for employees, makes any warranty, expressed or implied, or
reabsorption and poisoning, which may give an improved assumes any legal liability or responsibility for the accuracy,
alignment with the reaction trends observed among the completeness, or usefulness of any information, apparatus,
monometallic catalyst series.
product, or process disclosed, or represents that its use would
Overall, this work demonstrates that MOF-74 is an effective not infringe privately owned rights.
platform for developing single-site alcohol dehydrogenation
catalysts. The activity, selectivity, and stability of monometallic
MOF-74 can be enhanced by engineering a bimetallic frame-
References
work. This provides another illustration of the tremendous
tunability of MOF catalysts. It also motivates future studies to
fully understand the observed performance enhancements,
further improve the stability of the catalysts, and extend this
strategy to other OMS catalysts such as the Co-MOF-74 reported
here.
1 P. D. C. Dietzel, R. Blom and H. Fjellv ˚a g, Eur. J. Inorg. Chem.,
2008, 2008, 3624–3632.
2 P. D. C. Dietzel, Y. Morita, R. Blom and H. Fjellv ˚a g, Angew.
Chem., Int. Ed., 2005, 44, 6354–6358.
3 P. D. C. Dietzel, B. Panella, M. Hirscher, R. Blom and
H. Fjellv ˚a g, Chem. Commun., 2006, 959–961.
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. A, 2021, 9, 10869–10881 | 10879

View Article Online
Journal of Materials Chemistry A
Paper
4
5
6
F. Dawood, M. Anda and G. M. Shaullah, Int. J. Hydrogen 26 S. Chakraborty, P. E. Piszel, W. W. Brennessel and
Energy, 2020, 45, 3847–3869. W. D. Jones, Organometallics, 2015, 34, 5203–5206.
S. Koohi-Fayegh and M. A. Rosen, J. Energy Storage, 2020, 27, 27 Z. Dai, Q. Luo, H. Jiang, Q. Luo, H. Li, J. Zhang and T. Peng,
01047. Catal. Sci. Technol., 2017, 7, 2506–2511.
A. Schneemann, J. L. White, S. Kang, S. Jeong, L. F. Wan, 28 L. Liu and A. Corma, Chem. Rev., 2018, 118, 4981–5079.
E. S. Cho, T. W. Heo, D. Prendergast, J. J. Urban, 29 A. Wang, J. Li and T. Zhang, Nat. Rev. Chem., 2018, 2, 65–81.
B. C. Wood, M. D. Allendorf and V. Stavila, Chem. Rev., 30 S. Ji, Y. Chen, X. Wang, Z. Zhang, D. Wang and Y. Li, Chem.
1
2
018, 118, 10775–10839.
M. D. Allendorf, Z. Hulvey, T. Gennett, A. Ahmed, T. Autrey, 31 H. Ha, S. Yoon, K. An and H. Y. Kim, ACS Catal., 2018, 8,
J. Camp, E. Seon Cho, H. Furukawa, M. Haranczyk, M. Head- 11491–11501.
Gordon, S. Jeong, A. Karkamkar, D.-J. Liu, J. R. Long, 32 J. H ¨o cker, J.-O. Krisponeit, T. Schmidt, J. Falta and J. I. Flege,
K. R. Meihaus, I. H. Nayyar, R. Nazarov, D. J. Siegel, Nanoscale, 2017, 9, 9352–9358.
V. Stavila, J. J. Urban, S. P. Veccham and B. C. Wood, 33 M. Nolan and G. W. Watson, J. Phys. Chem. B, 2006, 110,
Energy Environ. Sci., 2018, 11, 2784–2812. 16600–16606.
M. P. Suh, H. J. Park, T. K. Prasad and D. W. Lim, Chem. Rev., 34 H. Jeong, G. Lee, B.-S. Kim, J. Bae, J. W. Han and H. Lee, J.
012, 112, 782–835. Am. Chem. Soc., 2018, 140, 9558–9565.
M. Markiewicz, Y. Q. Zhang, A. B ¨o smann, N. Br u¨ ckner, 35 Y. J. Hwang, Y. Kwon, Y. Kim, H. Sohn, S. W. Nam, J. Kim,
Rev., 2020, 120, 11900–11955.
7
8
9
2
J. Th ¨o ming, P. Wasserscheid and S. Stolte, Energy Environ.
Sci., 2015, 8, 1035–1045.
T. Autrey, C. W. Yoon, Y. S. Jo and H. Jeong, ACS
Sustainable Chem. Eng., 2020, 8, 9846–9856.
1
1
1
0 M. Niermann, A. Beckendorff, M. Kaltschmitt and 36 L. Jiao and H.-L. Jiang, Chem, 2019, 5, 786–804.
K. Bonhoff, Int. J. Hydrogen Energy, 2019, 44, 6631–6654.
1 P. Preuster, C. Papp and P. Wasserscheid, Acc. Chem. Res.,
37 M. C. Wasson, C. T. Buru, Z. Chen, T. Islamoglu and
O. K. Farha, Appl. Catal., A, 2019, 586, 117214.
2
017, 50, 74–85.
38 S. M. J. Rogge, A. Bavykina, J. Hajek, H. Garcia, A. I. Olivos-
Suarez, A. Sep u´ lveda-Escribano, A. Vimont, G. Clet,
P. Bazin, F. Kapteijn, M. Daturi, E. V. Ramos-Fernandez,
F. X. Llabr ´e s i Xamena, V. Van Speybroeck and J. Gascon,
Chem. Soc. Rev., 2017, 46, 3134–3184.
2 P. M. Modisha, C. N. M. Ouma, R. Garidzirai,
P. Wasserscheid and D. Bessarabov, Energy Fuels, 2019, 33,
2778–2796.
1
1
3 Y. Xie, P. Hu, Y. Ben-David and D. Milstein, Angew. Chem.,
Int. Ed., 2019, 58, 5105–5109.
4 T.-W. Liao, A. Yadav, P. Ferrari, Y. Niu, X.-K. Wei,
39 V. Pascanu, G. Gonz ´a lez Miera, A. K. Inge and B. Mart ´ı n-
Matute, J. Am. Chem. Soc., 2019, 141, 7223–7234.
J. Vernieres, K.-J. Hu, M. Heggen, R. E. Dunin-Borkowski, 40 D. Yang and B. C. Gates, ACS Catal., 2019, 9, 1779–1798.
R. E. Palmer, K. Laasonen, D. Grandjean, E. Janssens and 41 C. Wang, B. An and W. Lin, ACS Catal., 2019, 9, 130–146.
P. Lievens, Chem. Mater., 2019, 31, 10040–10048.
5 L. N. Chen, K. P. Hou, Y. S. Liu, Z. Y. Qi, Q. Zheng, Y. H. Lu, 43 U. Kokçam-Demir, A. Goldman, L. Esrali, M. Gharib,
42 J. N. Hall and P. Bollini, React. Chem. Eng., 2019, 4, 207–222.
¨
¨
1
J. Y. Chen, J. L. Chen, C. W. Pao, S. B. Wang, Y. B. Li,
S. H. Xie, F. D. Liu, D. Prendergast, L. E. Klebanoff,
A. Morsali, O. Weingart and C. Janiak, Chem. Soc. Rev.,
2020, 49, 2751–2798.
V. Stavila, M. D. Allendorf, J. Guo, L. S. Zheng, J. Su and 44 D. J. Xiao, J. Oktawiec, P. J. Milner and J. R. Long, J. Am.
G. A. Somorjai, J. Am. Chem. Soc., 2019, 141, 17995–17999. Chem. Soc., 2016, 138, 14371–14379.
6 Y.-Q. Zou, N. von Wolff, A. Anaby, Y. Xie and D. Milstein, Nat. 45 P. Verma, K. D. Vogiatzis, N. Planas, J. Borycz, D. J. Xiao,
1
1
Catal., 2019, 2, 415–422.
J. R. Long, L. Gagliardi and D. G. Truhlar, J. Am. Chem.
Soc., 2015, 137, 5770–5781.
7 P. J. Bonitatibus Jr, S. Chakraborty, M. D. Doherty,
O. Siclovan, W. D. Jones and G. L. Soloveichik, Proc. Natl. 46 E. D. Bloch, W. L. Queen, S. Chavan, P. S. Wheatley,
Acad. Sci. U. S. A., 2015, 112, 1687–1692.
8 X. Cheng, Z. Shi, N. Glass, L. Zhang, J. Zhang, D. Song,
Z.-S. Liu, H. Wang and J. Shen, J. Power Sources, 2007, 165,
J. M. Zadrozny, R. Morris, C. M. Brown, C. Lamberti,
S. Bordiga and J. R. Long, J. Am. Chem. Soc., 2015, 137,
3466–3469.
1
7
39–756.
9 S. D. Davidson, H. Zhang, J. Sun and Y. Wang, Dalton Trans.,
014, 43, 11782–11802.
47 J. Borycz, J. Paier, P. Verma, L. E. Darago, D. J. Xiao,
D. G. Truhlar, J. R. Long and L. Gagliardi, Inorg. Chem.,
2016, 55, 4924–4934.
1
2
2
2
0 A. Kubacka, M. Fern ´a ndez-Garc ´ı a and A. Mart ´ı nez-Arias, 48 D. J. Xiao, E. D. Bloch, J. A. Mason, W. L. Queen,
Appl. Catal., A, 2016, 518, 2–17.
1 K.-i. Shimizu, K. Kon, M. Seto, K. Shimura, H. Yamazaki and
J. N. Kondo, Green Chem., 2013, 15, 418–424.
M. R. Hudson, N. Planas, J. Borycz, A. L. Dzubak,
P. Verma, K. Lee, F. Bonino, V. Crocell `a , J. Yano,
S. Bordiga, D. G. Truhlar, L. Gagliardi, C. M. Brown and
J. R. Long, Nat. Chem., 2014, 6, 590–595.
22 Y. Choi and H. G. Stenger, Appl. Catal., B, 2002, 38, 259–269.
23 N. Takezawa and N. Iwasa, Catal. Today, 1997, 36, 45–56.
2
4 M. Trincado, D. Banerjee and H. Gr u¨ tzmacher, Energy
Environ. Sci., 2014, 7, 2464–2503.
49 H. Jiang, Q. Wang, H. Wang, Y. Chen and M. Zhang, ACS
Appl. Mater. Interfaces, 2016, 8, 26817–26826.
50 G. Bauer, D. Ongari, D. Tiana, P. Gaumann, T. Rohrbach,
G. Pareras, M. Tarik, B. Smit and M. Ranocchiari, Nat.
Commun., 2020, 11, 1059.
25 G. Zhang, C. Yang and J. T. Miller, ChemCatChem, 2018, 10,
961–964.
10880 | J. Mater. Chem. A, 2021, 9, 10869–10881
This journal is © The Royal Society of Chemistry 2021

View Article Online
Paper
Journal of Materials Chemistry A
5
1 E. M. A. Fuentes-Fernandez, S. Jensen, K. Tan, S. Zuluaga,
H. Wang, J. Li, T. Thonhauser and Y. J. Chabal, Appl. Sci.,
F. Egidi, J. Goings, B. Peng, A. Petrone, T. Henderson,
D. Ranasinghe, V. G. Zakrzewski, J. Gao, N. Rega,
G. Zheng, W. Liang, M. Hada, M. Ehara, K. Toyota,
R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, T. Vreven, K. Throssell,
J. A. Montgomery Jr, J. E. Peralta, F. Ogliaro,
M. J. Bearpark, J. J. Heyd, E. N. Brothers, K. N. Kudin,
V. N. Staroverov, T. A. Keith, R. Kobayashi, J. Normand,
K. Raghavachari, A. P. Rendell, J. C. Burant, S. S. Iyengar,
J. Tomasi, M. Cossi, J. M. Millam, M. Klene, C. Adamo,
R. Cammi, J. W. Ochterski, R. L. Martin, K. Morokuma,
O. Farkas, J. B. Foresman, and D. J. Fox, Gaussian 16,
Revision C.01, Gaussian, Inc., Wallingford CT, 2016.
2018, 8, 270.
5
5
5
2 D.-A. Yang, H.-Y. Cho, J. Kim, S.-T. Yang and W.-S. Ahn,
Energy Environ. Sci., 2012, 5, 6465–6473.
3 I. Choi, Y. E. Jung, S. J. Yoo, J. Y. Kim, H.-J. Kim, C. Y. Lee and
J. H. Jang, J. Electrochem. Sci. Technol., 2017, 8, 61–68.
4 V. Stavila, R. Parthasarathi, R. W. Davis, F. El Gabaly,
K. L. Sale, B. A. Simmons, S. Singh and M. D. Allendorf,
ACS Catal., 2015, 6, 55–59.
5
5 V. Stavila, M. E. Foster, J. W. Brown, R. W. Davis,
J. Edgington, A. I. Benin, R. A. Zarkesh, R. Parthasarathi,
D. W. Hoyt, E. D. Walter, A. Andersen, N. M. Washton,
A. S. Lipton and M. D. Allendorf, Chem. Sci., 2019, 10, 71 W. Zhou, H. Wu and T. Yildirim, J. Am. Chem. Soc., 2008, 130,
880–9892. 15268–15269.
6 M. Mukoyoshi, H. Kobayashi, K. Kusada, M. Hayashi, 72 M. Liu, S. Zhao, X. Xiao, M. Chen, C. Sun, Z. Yao, Z. Hu and
T. Yamada, M. Maesato, J. M. Taylor, Y. Kubota, K. Kato, L. Chen, Nano Energy, 2019, 61, 540–549.
M. Takata, T. Yamamoto, S. Matsumura and H. Kitagawa, 73 M. Matsukata, T. Hayashi, M. Nishiyama, E. Kikuchi and
Chem. Commun., 2015, 51, 12463–12466. Y. Morita, Sekiyu Gakkaishi, 1987, 30, 265.
7 S. Zuluaga, E. M. A. Fuentes-Fernandez, K. Tan, F. Xu, J. Li, 74 T. C. Wang, J. L. White, B. Bie, H. Deng, J. Edgington,
9
5
5
Y. J. Chabal and T. Thonhauser, J. Mater. Chem. A, 2016, 4,
176–5183.
J. D. Sugar, V. Stavila and M. D. Allendorf, ChemPhysChem,
2019, 20, 1305–1310.
5
58 Y. Jiao, C. R. Morelock, N. C. Burtch, W. P. Mouneld, 75 I. Strauss, A. Mundstock, D. Hinrichs, R. Himstedt,
J. T. Hungerford and K. S. Walton, Ind. Eng. Chem. Res.,
015, 54, 12408–12414.
A. Knebel, C. Reinhardt, D. Dorfs and J. Caro, Angew.
Chem., Int. Ed., 2018, 57, 7434–7439.
2
5
6
6
6
6
6
9 G. Kresse and J. Furthm u¨ ller, Comput. Mater. Sci., 1996, 6, 76 J. Chen, X. Mu, M. Du and Y. Lou, Inorg. Chem. Commun.,
5–50. 2017, 84, 241–245.
0 G. Kresse and J. Furthm u¨ ller, Phys. Rev. B: Condens. Matter 77 J. A. Villajos, G. Orcajo, C. Martos, J. A. Botas, J. Villacanas
Mater. Phys., 1996, 54, 11169–11186. and G. Calleja, Int. J. Hydrogen Energy, 2015, 40, 5346–5352.
1 J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 78 M. Y. Masoomi, A. Morsali, A. Dhakshinamoorthy and
996, 77, 3865–3868. H. Garcia, Angew. Chem., Int. Ed., 2019, 58, 15188–15205.
2 I. Hamada, Phys. Rev. B: Condens. Matter Mater. Phys., 2014, 79 S. Abednatanzi, P. Gohari Derakhshandeh, H. Depauw,
1
´
˜
1
8
9, 121103.
F.-X. Coudert, H. Vrielinck, P. Van Der Voort and K. Leus,
Chem. Soc. Rev., 2019, 48, 2535–2565.
80 D. Sun, F. Sun, X. Deng and Z. Li, Inorg. Chem., 2015, 54,
8639–8643.
3 A. I. Liechtenstein, V. I. Anisimov and J. Zaanen, Phys. Rev. B:
Condens. Matter Mater. Phys., 1995, 52, R5467–R5470.
4 S. L. Dudarev, G. A. Botton, S. Y. Savrasov, C. J. Humphreys
and A. P. Sutton, Phys. Rev. B: Condens. Matter Mater. Phys., 81 K. Yuan, T. Song, D. Wang, Y. Zou, J. Li, X. Zhang, Z. Tang
998, 57, 1505–1509. and W. Hu, Nanoscale, 2018, 10, 1591–1597.
5 L. Wang, T. Maxisch and G. Ceder, Phys. Rev. B: Condens. 82 S. R. Caskey, A. G. Wong-Foy and A. J. Matzger, J. Am. Chem.
Matter Mater. Phys., 2006, 73, 195107. Soc., 2008, 130, 10870–10871.
6 P. E. Bl ¨o chl, Phys. Rev. B: Condens. Matter Mater. Phys., 1994, 83 V. Stavila, R. K. Bhakta, T. M. Alam, E. H. Majzoub and
0, 17953–17979. M. D. Allendorf, ACS Nano, 2012, 6, 9807–9817.
7 G. Kresse and D. Joubert, Phys. Rev. B: Condens. Matter Mater. 84 W. L. Masterton, E. J. Slowinski and C. L. Stanitski, Chemical
Phys., 1999, 59, 1758–1775. principles, CBS College Publishing, New York, 5th edn, 1983.
8 P. Verma, X. Xu and D. G. Truhlar, J. Phys. Chem. C, 2013, 85 L. J. Wang, H. Deng, H. Furukawa, F. Gandara, K. E. Cordova,
17, 12648–12660. D. Peri and O. M. Yaghi, Inorg. Chem., 2014, 53, 5881–5883.
9 F. Weigend and R. Ahlrichs, Phys. Chem. Chem. Phys., 2005, 86 E. Jones and G. G. Fowlie, J. Appl. Chem., 1953, 3, 206–213.
1
6
6
6
6
6
7
5
1
7
, 3297–3305.
87 G. Bauer, D. Ongari, X. Xu, D. Tiana, B. Smit and
0 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. Ranocchiari, J. Am. Chem. Soc., 2017, 139, 18166–18169.
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, 88 C.-D. Wu, A. Hu, L. Zhang and W. Lin, J. Am. Chem. Soc.,
G. A. Petersson, H. Nakatsuji, X. Li, M. Caricato, 2005, 127, 8940–8941.
A. V. Marenich, J. Bloino, B. G. Janesko, R. Gomperts, 89 M. Zheng, Y. Liu, C. Wang, S. Liu and W. Lin, Chem. Sci.,
B. Mennucci, H. P. Hratchian, J. V. Ortiz, A. F. Izmaylov,
J. L. Sonnenberg, D. Williams-Young, F. Ding, F. Lipparini,
2012, 3, 2623–2627.
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. A, 2021, 9, 10869–10881 | 10881