Catalytic Hydrogenation of Azobenzene in the Presence of a Cuboidal Mo3S4Cluster via an Uncommon Sulfur-Based H2Activation Mechanism

Article abstract of DOI:10.1021/acscatal.0c05299

Azobenzene hydrogenation is catalyzed under moderate conditions by a cuboidal Mo3(μ3-S)(μ-S)3 diamino complex via a cluster catalysis mechanism. Dihydrogen activation by the molecular [Mo3(μ3-S)(μ-S)3Cl3(dmen)3]+ cluster cation takes place at the μ-S bridging atoms without direct participation of the metals in clear contrast with classical concepts. The reaction occurs with the formation of 1,2-diphenylhydrazine as an intermediate with similar appearance and disappearance rate constants. On the basis of DFT calculations, a mechanism is proposed in which formation of 1,2-diphenylhydrazine and aniline occurs through two interconnected catalytic cycles that share a common reaction step that involves H2 addition to two of the bridging sulfur atoms of the catalyst to form a dithiolate Mo3(μ3-S)(μ-SH)2)(μ-S) adduct. Both catalytic cycles have similar activation barriers, in agreement with the experimental observation of close rate constant values. Microkinetic modeling of the process leads to computed concentration-time profiles in excellent agreement with the experimental ones providing additional support to the calculated reaction mechanism. Slight modifications on the experimental conditions of the catalytic protocol in combination with theoretical calculations discard a direct participation of the metal on the reaction mechanism. The effect of the ancillary ligands on the catalytic activity of the cluster fully agrees with the present mechanistic proposal. The results herein demonstrate the capability of molybdenum sulfide materials to activate hydrogen through an uncommon sulfur based mechanism opening attractive possibilities toward their applications as catalysts in other hydrogenation processes.

Full text of DOI:10.1021/acscatal.0c05299

pubs.acs.org/acscatalysis
Letter
Catalytic Hydrogenation of Azobenzene in the Presence of a
Cuboidal Mo3S4 Cluster via an Uncommon Sulfur-Based H2
Activation Mechanism
§
§
́
́
́
Eva Guillamon, Monica Oliva, Juan Andres, Rosa Llusar,* Elena Pedrajas, Vicent S. Safont,
Andres G. Algarra,* and Manuel G. Basallote*
Cite This: ACS Catal. 2021, 11, 608−614
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Azobenzene hydrogenation is catalyzed under moderate
conditions by a cuboidal Mo3(μ3-S)(μ-S)3 diamino complex via a cluster
catalysis mechanism. Dihydrogen activation by the molecular [Mo3(μ3-S)(μ-
S)3Cl3(dmen)3]+ cluster cation takes place at the μ-S bridging atoms without
direct participation of the metals in clear contrast with classical concepts. The
reaction occurs with the formation of 1,2-diphenylhydrazine as an intermediate
with similar appearance and disappearance rate constants. On the basis of DFT
calculations, a mechanism is proposed in which formation of 1,2-
diphenylhydrazine and aniline occurs through two interconnected catalytic
cycles that share a common reaction step that involves H2 addition to two of the
bridging sulfur atoms of the catalyst to form a dithiolate Mo3(μ3-S)(μ-SH)2)(μ-
S) adduct. Both catalytic cycles have similar activation barriers, in agreement
with the experimental observation of close rate constant values. Microkinetic modeling of the process leads to computed
concentration−time profiles in excellent agreement with the experimental ones providing additional support to the calculated
reaction mechanism. Slight modifications on the experimental conditions of the catalytic protocol in combination with theoretical
calculations discard a direct participation of the metal on the reaction mechanism. The effect of the ancillary ligands on the catalytic
activity of the cluster fully agrees with the present mechanistic proposal. The results herein demonstrate the capability of
molybdenum sulfide materials to activate hydrogen through an uncommon sulfur based mechanism opening attractive possibilities
toward their applications as catalysts in other hydrogenation processes.
KEYWORDS: cluster catalysis, density functional theory, azobenzene hydrogenation, hydrogen activation, microkinetic modeling,
molybdenum sulfide, reaction mechanism
atalytic hydrogenation of organic compounds lies at the
foundations of modern chemistry due to its great
During the past decade, we have exploited this approach by
studying the behavior of trinuclear molybdenum sulfide
clusters containing a cuboidal Mo3(μ3-S)(μ-S)3 core. Specif-
ically, we have shown that these species are able to catalyze the
reduction of nitroarenes5 and the semihydrogenation of
alkynes.6 These results prompted us to extend our studies to
other unsaturated substrates, and in particular, we focused on
those containing N−N bonds.
Previous studies by our group, in collaboration with Beller’s
group, have shown that hydrogenation of nitroarenes to afford
aniline derivatives catalyzed by diamino and diimino Mo3(μ3-
S)(μ-S)3 complexes occurs through the two competing paths
C
practical significance, and transition metal catalysts have been
the most effective in achieving this transformation.1
Dihydrogen activation is a key step of these catalytic processes
and its mechanism has been intensively studied using both
experimental and theoretical methods.2 Molybdenum sulfide-
based materials have recently emerged as an alternative to
noble metals for the hydrogenation of unsaturated substrates.3
However, the mechanism of hydrogen activation at these
compounds is still under debate. Indeed, a number of studies
have shown that H2 adsorption characteristics of MoS2 depend
on the structure (basal-plane vs edge sites), its degree of sulfur
saturation, hydrogen coverage, and temperature.4
Given the difficulties associated with obtaining mechanistic
information on heterogeneous MoS2 based catalysts, various
research groups have tackled the chemistry of homogeneous
catalysis mediated by molybdenum sulfide complexes. In
particular, molecular clusters represent simplified structural
models of MoS2 that often emulate their catalytic properties.
Received: December 3, 2020
Revised: December 23, 2020
Published: December 29, 2020
https://dx.doi.org/10.1021/acscatal.0c05299
© 2020 American Chemical Society
ACS Catal. 2021, 11, 608−614
608

ACS Catalysis
pubs.acs.org/acscatalysis
Letter
proposed by Haber more than a century ago, i.e., the direct and
the condensation routes.7 The latter leads to aniline through
azoxy, azo, and hydrazo derivatives, so azobenzene catalytic
hydrogenation to afford aniline using Mo3S4 clusters seems
feasible. However, the mechanism of such transformation
remains unclear, although experimental evidence give support
to the preservation of the cluster complex integrity during the
catalytic transformation.
Herein, we present an efficient protocol for the azobenzene
hydrogenation catalyzed by the cuboidal [Mo3(μ3-S)(μ-
S)3Cl3(dmen)3]+ (1) complex (see Scheme 1) (dmen =
Scheme 1. Structure of the [Mo3S4Cl3(dmen)3]+ (1) Cluster
Cation (Top) and Sequential Conversion of Azobenzene to
1,2-Diphenylhydrazine and Aniline (Bottom)
Figure 1. Time course for the different species observed in the
catalytic hydrogenation of azobenzene: azobenzene (circles), 1,2-
diphenylhydrazine (triangles), and aniline (squares). The line
corresponds to the fit to a single exponential of the data for the
disappearance of azobenzene.
azobenzene is initially converted into 1,2-diphenylhydrazine,
which then undergoes hydrogenation to afford 2 equiv of
aniline (Scheme 1). The absence of an induction period in
Figure 1 suggests that the catalyst maintains its integrity and
pinpoints the [Mo3(μ3-S)(μ-S)3Cl3(dmen)3]+ cluster as the
active species, which was confirmed by the absence of
noticeable changes in the ESI-MS spectra at the end of the
reaction (Figure S2). Notably, no deuterated aniline was
formed by using CD3OD as solvent, thus precluding a solvent
mediated transfer hydrogenation process. Attempts to fit the
kinetic profiles in Figure 1 to a model involving two
consecutive kinetic steps (A → B → C) with rate constants
k1 and k2 were unsuccessful because the values of both rate
constants are close to each other, and this makes the
preexponential terms that include the k2 − k1 difference in
the denominator undefined.10 Nevertheless, the value of k1 can
be easily derived from a fit of the data for the disappearance of
the starting compound (A), and yields k1 = 0.42 0.03 h−1.
The value of k2 can then be estimated by considering that at
the time when the intermediate species (B) reaches its
maximum concentration d[B]/dt = 0 and so k1[A] = k2[B].11
Inspection of Figure 1 indicates that the intermediate reaches
its maximum concentration at ca. 2.4 h, and at that time, the
percentages of A and B are both close to 35%, and therefore k2
≈ k1.
To gain further insight into the reaction mechanism, a
theoretical study on the interaction between azobenzene and
the cluster was carried out. Notably, all attempts to obtain
stable adducts resulting from the interaction between the
cluster and azobenzene failed. Based on this outcome, the
existence of cluster−hydrogen interactions was explored.
However, all studied mechanisms involving H2 activation
with participation of the metal centers require a large energy
input, unattainable under our experimental conditions (see
below). This is in fact not surprising given the stability of these
clusters, which feature Mo centers with an octahedral
coordination environment. On the contrary, the H2 molecule
was found to react with two of the bridging sulfide ligands of
N,N′-dimethylethylenediamine). In our study, we combine
experimental and DFT methods to unravel the kinetics of the
process as well as its reaction mechanism. Our work pinpoints
a homolytic H2 activation mechanism at two of the bridging
sulfur atoms followed by two consecutive hydrogen transfers to
azobenzene and 1,2-diphenylhydrazine to afford aniline.
In this work, the hydrogenation of azobenzene is used as a
probe reaction to unveil the hydrogenation mechanism
mediated by cuboidal Mo3(μ3-S)(μ-S)3 cluster compounds.
For that purpose, we selected the diamino [Mo3S4Cl3(dmen)3]
Cl complex salt and used a protocol similar to that employed
for the catalytic direct hydrogenation of nitroarenes.5c In a
typical experiment, azobenzene (0.1 mmol) and the cluster
catalyst (0.005 mmol) were dissolved in 2 mL of methanol and
the vial introduced in an autoclave. Full conversion and up to
94% yield of aniline was achieved by using a milder protocol
(10 bar of H2, 60 °C for 6 h) than the one needed for the
catalytic transformation of nitrobenzene to aniline (20 bar of
H2, 70 °C for 18 h).5e These conditions are for instance less
demanding than those employed by Suzuki et al. using
trinuclear ruthenium pentahydride clusters, while conversion
and selectivity are significantly improved.8 Moreover, these
results demonstrate that Mo3S4 clusters are capable of
promoting the cleavage of the N−N bond in hydrazine
derivatives, a process of relevance to nitrogen fixation.9
The kinetics of the [Mo3(μ3-S)(μ-S)3Cl3(dmen)3]Cl cata-
lyzed hydrogenation of azobenzene in CH3OH was then
monitored in a series of batch experiments. The time course of
the reaction under standard conditions (60 °C, 10 atm H2) is
illustrated in Figure 1, which shows the typical profile for a
reaction occurring via two consecutive kinetic steps, in which
609
https://dx.doi.org/10.1021/acscatal.0c05299
ACS Catal. 2021, 11, 608−614

ACS Catalysis
pubs.acs.org/acscatalysis
Letter
the cluster to form a [Mo3(μ3-S)(μ-SH)2(μ-S)Cl3(dmen)3]+
(1-H2) adduct with a reasonable energy requirement, as shown
in the corresponding energy profile (Figure 2). The process is
diphenylhydrazine. This can be viewed as two independent
catalytic cycles in which each dihydrogen transfer regenerates
the [Mo3(μ3-S)(μ-S)(μ-S)2Cl3(dmen)3]+ cluster (Figure 3).
Figure 2. Gibbs energy profile, in kcal/mol, relative to reactants (1)
for the addition of hydrogen to the cluster. For simplicity, dmen and
Cl ligands at the Mo centers are not drawn. Blue, yellow, and white
spheres represent the Mo, S, and H atoms, respectively.
Figure 3. Mechanism proposed for the catalytic conversion of
azobenzene into aniline. The figure illustrates the operation of two
interconnected catalytic cycles, one for the hydrogenation of
diazobenzene (cycle I) and another for the hydrogenation of 1,2
diphenylhydrazine (cycle II).
endergonic by 14.7 kcal/mol and leads to the homolytic
cleavage of the H−H bond with formation of two (μ-S)−H
bonds with a Gibbs energy barrier of 16.9 kcal/mol.
Interestingly, this mechanism shares common features with
the [3 + 2] cycloaddition reaction between Mo3(μ3-S)(μ-S)3
clusters and alkynes to form dithiolene adducts, as in both
cases, a bond (either a π C−C bond of the alkyne or the σ H−
H bond of dihydrogen) is broken and two new σ bonds formed
(S−C and S−H, respectively).12 In addition, there is a
shortening of one Mo−Mo bond from 2.759 Å in 1, typical of
a single metal−metal bond, to 2.687 Å in 1-H2, characteristic
of a double metal−metal bond (Table S1). The same tendency
is observed upon alkyne insertion into the Mo3S4 unit both in
the experimental and in the optimized structures.12 The
geometry of the TS, depicted in Figure 2, shows similar (μ-S)−
H bond distances as expected for a homolytic dihydrogen
activation process.
It is worth highlighting at this point the similarities in the
reactivity of [Mo3(μ3-S)(μ-S)3Cl3(dmen)3]+ and the dinuclear
[Cp*Mo]2(μ-S)2(μ-S2) cluster13 (Cp* = pentamethylcyclo-
pentadienyl).The latter also reacts with H2 through its bridging
sulfide ligands to form the species containing two hydrosulfido
ligands [Cp*Mo]2(μ-S)2(μ-SH)2, although in this case, the
process is slightly exergonic.14 Moreover, [Cp*Mo]2(μ-S)2(μ-
SH)2 reacts at room temperature with azobenzene by
transferring those two H atoms, which generates 1,2-
diphenylhydrazine and regenerates the [Cp*Mo]2(μ-S)2(μ-
S2) cluster. Interestingly, at least under the milder experimental
conditions employed by Dubois et al., this process does not
result in the formation of aniline, and only 1 equiv of H2 is
incorporated into the azobenzene molecule. In spite of such a
difference, the data on [Cp*Mo]2(μ-S)2(μ-SH)2 is in line with
the reaction mechanism suggested by our DFT calculations for
the azobenzene hydrogenation in the presence of 1. These
results indicate that the sequential formation of 1,2-
diphenylhydrazine and aniline takes place through two
interconnected catalytic cycles which share the formation of
1-H2 as a common step. This species, resulting from the
homolytic activation of a dihydrogen molecule, is able to
transfer its two hydrogen atoms to either azobenzene or 1,2-
The calculated DFT profile for both cycles is shown in
Figure 4. In the first cycle, the reaction between 1-H2 and
azobenzene through transition state TS-1 directly transforms
the latter species into 1,2-diphenylhydrazine. Interestingly, the
two H atoms transferred in TS-1 feature distinct S−H (1.549
and 1.400 Å) and H−N (1.495 and 2.166 Å) distances, which
is indicative of an asynchronous process (Figure S8).
Hydrogen transfer in the second cycle occurs in a stepwise
manner through the I-2 intermediate connected with 1,2-
phenylhydrazine and aniline via transition states TS-2A and
TS-2B, respectively. Note that the dispersion and standard
state corrections required to obtain relative free energies within
this cycle results in an unrealistic free energy value for TS-2A,
as it is computed to be lower than those of 1-H2 and I-2. This
is an artifact that appears when neighboring stationary points
have similar energies, and it is not present in the potential
energy profile (Figure S7). In any case, intermediate I-2 is not
relevant from a thermodynamic viewpoint, nor it is expected to
be experimentally observed.
The data in Figure 4 indicates that overcoming transition
states TS-1 and TS-2B represent the rate-determining steps of
each cycle. This results in Gibbs energy barriers of 18.2 and
17.7 kcal/mol, respectively, when those of the cluster, H2, and
nitrogen-containing substrate are taken as relative zero. This is
indeed in very good agreement with the experimental
observation of close rate constants for the formation of 1,2-
diphenylhydrazine and its conversion to aniline. It is worth
highlighting that each of these barriers results from the
combination of two different processes: (a) the formation of
intermediate 1-H2, which in both cases requires 16.9 kcal/mol
to overcome TS-H2 and reach 1-H2, which lies 14.7 kcal/mol
above the starting point of each cycle; (b) the subsequent
transfer of the H atoms to the substrates, which requires 3.5
and 3.0 kcal/mol in cycles I and II, respectively, to climb from
1-H2 to each limiting TS. Based on this Gibbs energy profile,
the formation of 1-H2 is relatively difficult; hence, the required
high H2 pressure and temperature. Nevertheless, once it is
formed, it will react with azobenzene or 1,2-diphenylhydrazine
with quite similar energy barriers. This suggests that 1-H2 will
610
https://dx.doi.org/10.1021/acscatal.0c05299
ACS Catal. 2021, 11, 608−614

ACS Catalysis
pubs.acs.org/acscatalysis
Letter
Figure 4. Gibbs energy profile (kcal/mol) for the two-cycle catalytic conversion of azobenzene to aniline. Cl and dmen ligands are omitted for
simplicity. N−N distances at the TSs are given in Å.
have an almost negligible discrimination capability, with the
experimental observation of sequential formation of 1,2-
diphenylhydrazine and aniline being just a consequence of
the relative concentrations throughout the experiment.
Specifically, the absence of 1,2-diphenylhydrazine at the early
stages of reaction will favor the reaction with azobenzene
(cycle I), but as the reaction goes on and the ratio between the
concentrations of 1,2-diphenylhydrazine and azobenzene
increases, cycle II will be favored against cycle I.
The complexity of the proposed mechanism is, evidently,
associated with the presence of at least five different species in
the reaction mixture (cluster, H2, azobenzene, 1,2-diphenylhy-
drazine, and aniline). In order to evaluate how the Gibbs
energy profile in Figure 4 is affected by the varying
concentrations of these species, the system has been studied
via microkinetic modeling.15 For that purpose, the rate
constants for the different steps were estimated from their
Gibbs energies of activation in Figure 4 using the Eyring
equation, which allowed the concentration−time profiles for
each species to be computed. The experimental kinetic data in
Figure 1 were found to be well reproduced (Figure 5) by
simply introducing minor changes in the rate constant for the
initial H2 addition to the cluster. Actually, a rate constant of
1.31 0.06 M−1 s−1 was obtained by fitting the experimental
data to the computed mechanism, which corresponds to a
Gibbs energy of activation of 19.4 kcal mol−1. This value differs
from that in Figure 4 (16.9 kcal mol−1) only by 2.5 kcal mol−1,
which is surely within the experimental and computational
errors.
Figure 5. Concentration−time profiles for azobenzene (circles), 1,2-
diphenylhydrazine (triangles), and aniline (squares) during the
cluster-catalyzed hydrogenation process. The symbols correspond to
the experimental data (Figure 1 with the percentages converted to
concentrations and times in seconds), and the lines to the
concentrations calculated from the computed mechanism in Figure
4 using the refined value of the rate constant for H2 addition to the
cluster (1.31 M−1 s−1).
imental conditions used is estimated to be lower than 10−14
M
both in the presence and in the absence of azobenzene (Figure
S10), which makes it undetectable. In agreement with these
expectations, the NMR spectra of the cluster under H2
pressure (Figure S5), failed to observe the formation of
significant amounts of any other species.
Given the impossibility of detecting the 1-H2 intermediate,
we directed our efforts to find indirect experimental evidence
which could provide support to the above sulfur centered
The key feature of the proposed mechanism is the activation
of H2 at the bridging sulfur atoms of the cluster to form
intermediate 1-H2. Unfortunately, the energy profiles show
that this species is expected to be formed under steady-state
conditions, and actually its concentration under the exper-
611
https://dx.doi.org/10.1021/acscatal.0c05299
ACS Catal. 2021, 11, 608−614

ACS Catalysis
pubs.acs.org/acscatalysis
Letter
mechanism and simultaneously discard any metal participation
in the process. Clear evidence supporting H2 activation by the
bridging sulfur atoms in 1 was obtained by blocking the three
bridging sulfur atoms of the cluster. For that purpose, CuCl
was added to the reaction mixture during the catalytic protocol
(Table 1, entry 2), which resulted in the inhibition of the
added chloride (Table 1, entries 3 and 4) has no significant
effect in the reaction outcome, discarding a mechanism
involving a reversible dissociation-coordination of the chloride
ancillary ligands that would allow for the heterolytic activation
of H2 at a Mo-(μ-S) bond to form a MoH/SH intermediate. In
addition, no loss of the cluster C3 symmetry is observed by
proton NMR when the cluster catalyst is heated under
hydrogen pressure (Figure S5) as expected for a cluster
containing partially decoordinated ligands. These experimental
results are further supported by theoretical calculations (Figure
S11a). The alternative possibility of creating a vacant
coordination site at the metal through dissociation of one of
the nitrogen atoms of the diamine can also be discarded on the
basis of the aforementioned cluster C3 symmetry preservation
and it is also corroborated by theoretical calculations. (Figure
S11b). On the other hand, HCl elimination to furnish Mo
NCH3 species that could yield MoH/NH species upon
hydrogenation can also be ruled out, as calculations indicate
that HCl elimination to form MoNCH3 species present a
barrier of 39.5 kcal/mol, again much higher than that evaluated
for the proposed mechanism (Figure S11c). In addition, the
effect of added basis (Table 1, entries 5−7) do not support any
mechanism involving HCl elimination.
Table 1. Evaluation of Conditions for Azobenzene (AZ)
a
Hydrogenation
b
yield (%)
b
variation from “initial
conditions”
conversion
(%)
c
entry
DPH AN ΔG
1
2
3
None
83
0
82
25
0
24
54
0
55
14.7
addition of CuCl (2 equiv)
addition of (n-Bu4N)Cl
(3 equiv)
4
5
addition of (n-Bu4N)Cl
(30 equiv)
80
78
19
19
45
51
addition of pyridine (2−12
equiv)
With regard to the effect of added bases, pyridine addition
has no significant effect in conversion and product distribution
(Table 1, entry 5; see also Figure S3). However, a striking
effect is found upon addition of the stronger NEt3 base, which
completely inhibits the catalytic process (Table 1, entries 6−
7). Under these conditions, partial substitution of one or two
of the outer chlorides by methoxo and hydroxo groups is
observed by ESI-MS mass spectrometry (Figure S4). The
calculated ΔG of 15.7 kcal mol−1 for the hydrogenation of the
most abundant substitution product, [Mo3S4Cl-
(OH)2(dmen)3]+, only shows a modest increase with respect
to that of 1 to form 1-H2. Even though these substitution
compounds could have a lower catalytic activity, this would not
explain the complete inhibition of the catalytic process,
because most of the starting cluster remains unreacted under
these conditions. Indeed, a plausible explanation can be given
on the basis of the proposed mechanism, as the S−H groups of
the 1-H2 intermediate are expected to have acidic character
(calculated pKa of 4.1), and therefore, the base would capture
these H atoms, thereby introducing an efficient additional
reaction pathway for this intermediate that would lead to
inhibition of the whole hydrogenation process.
6
7
8
9
addition of Et3N (1 equiv)
addition of Et3N (2−3 equiv)
10
0
78
14
32
4
0
34
3
6
0
36
11
20
d
[Mo3S4Br3(dmen)3]+
14.1
16.2
13.8
d
[Mo3S4Cl3(dnbpy)3]+
d
10
[Mo3S4(tacn)3]4+ + HBF4
12
(1.5 equiv)
d
11
12
13
[Mo3S4Cl3(edpp)3]+
0
0
0
0
0
0
0
0
0
21.7
18.3
23.5
d
[Mo3S4Cl3(dmpe)3]+
d
[Mo3S4H3(dmpe)3]+
a
b
Reaction conditions: AZ (0.1 mmol), CH3OH (2 mL). Determined
by GC analysis using n-hexadecane as an internal standard.
c
Computed free energy of the corresponding Mo3S4−H2 adduct
(kcal mol−1) quoted relative to the Mo3S4 catalyst + H2 (entries 1 and
d
8−13). Outer ligands:
At that point, we decided to explore the effect of the
ancillary ligands on the catalytic activity of a series of Mo3S4
clusters decorated with diamino, diimino, triamino, amino-
phosphino, and diphosphino ligands, aimed to gather addi-
tional arguments to this uncommon hydrogen activation
mechanism at the sulfur centers. A summary of the results is
listed in Table 1 (entries 8−13) together with the calculated
free energy needed for the formation of the corresponding
Mo3S4−H2 intermediates, the first and key step in the
azobenzene hydrogenation reaction. To our delight, we
found a good correlation between the evaluated catalytic
activity and the free energy needed to reach the Mo3S4−H2
intermediate in all complexes containing bidentate ligands,
which provides strong support to the proposed mechanism. In
particular, all clusters containing phosphorus atoms bound to
the metal (Table 1, entries 11−13) are inactive. Diamino
clusters show higher activities than their diimino counterparts
(Table 1, entries 8−10) and substitution of chlorine by
hydrogenation process. It is well-known that cuboidal Mo3(μ3-
S)(μ-S)3 clusters interact with a variety of metals (M′) to form
Mo3M′S4 clusters with cubane-type structures.16 In this case,
copper addition results in the in situ formation of the
heterobimetallic [Mo3(CuCl)S4Cl3(dmen)3]+ cation, which
was isolated as the CuCl2− salt whose crystal structure (Figure
S6) shows the formation of three new (μ3-S)−Cu bonds.
Next, a combined experimental and theoretical approach was
undertaken to analyze any potential mechanisms which involve
the formation of vacant sites at the metal center such as
dissociation of the outer halide, partial decoordination of the
diamino bidentate ligand, or HCl elimination with the
concomitant formation of MoNCH3 bonds induced by a
base. Performance of the catalytic protocol in the presence of
612
https://dx.doi.org/10.1021/acscatal.0c05299
ACS Catal. 2021, 11, 608−614

ACS Catalysis
pubs.acs.org/acscatalysis
Letter
bromine in [Mo3S4Cl3(dmen)3]+ (Table 1, entries 1 and 8)
decreases by ca. 10% the azobenzene conversion. The catalytic
protocol for the triazacyclononane Mo3S4 derivative (Table 1,
entry 10) was performed in the presence of acid to avoid
deprotonation of the coordinated tacn ligands, as it has been
shown that cluster salts of the [Mo3S4(tacn)3]4+ cation can
only be crystallized from acidic solutions (pH < 2).17 Under
these conditions, conversions lye between those obtained for
the diamino and the diimino cluster catalysts.
A definite proof in favor of our sulfur centered mechanistic
proposal comes from the observed catalytic activity when
capping the molybdenum atoms with tacn ligands. In the
structure of [Mo3S4(tacn)3]4+, the three outer positions on
each metal are completed by facially coordinating tridentate
macroclyclic tacn amine ligands. This circumstance indicates
that formation of a Mo-based reaction site is highly unlikely. In
contrast, as detailed before, blocking of the sulfur atoms (Table
1, entry 2) results in the inhibition of the catalytic
hydrogenation of azobenzene.
optimized structures of TS1, TS-2A, I-2, and TS-2B,
microkinetic modeling, computational tests on alter-
native paths, and Cartesian coordinates of all optimized
structures (PDF)
AUTHOR INFORMATION
■
Corresponding Authors
Rosa Llusar − Departament de Química Física i Analítica,
Universitat Jaume I, 12071 Castelló, Spain; orcid.org/
0000-0002-3539-7269; Email: rosa.llusar@uji.es
Andres G. Algarra − Departamento de Ciencia de los
Materiales e Ingeniería Metalúrgica y Química Inorgánica,
Facultad de Ciencias, Universidad de Cádiz, 11510 Puerto
Real, Cádiz, Spain; orcid.org/0000-0002-5062-2858;
Email: andres.algarra@uca.es
Manuel G. Basallote − Departamento de Ciencia de los
Materiales e Ingeniería Metalúrgica y Química Inorgánica,
Facultad de Ciencias, Universidad de Cádiz, 11510 Puerto
Real, Cádiz, Spain; orcid.org/0000-0002-1802-8699;
Email: manuel.basallote@uca.es
In conclusion, the [Mo3S4Cl3(dmen)3]+ cuboidal cluster
efficiently catalyzes the reduction of azobenzene to aniline with
high yields under relatively mild conditions. The reaction
occurs with formation of 1,2-diphenylhydrazine as a detectable
intermediate, which is formed and disappears with close rate
constants. Experimental outcomes, computational studies and
microkinetic modeling support a reaction mechanism in which
H2 is homolytically activated at two of the three bridging sulfur
atoms of the cluster to form a dithiol [Mo3(μ3-S)(μ-S)(μ-
SH)2Cl3(dmen)3]+ intermediate under steady-state conditions.
This intermediate is able to transfer both hydrogen atoms to
azobenzene and to 1,2-diphenylhydrazine through two
interconnected catalytic cycles. Although all processes take
place without direct participation of the Mo centers or their
ancillary ligands, the latter are found to significantly modulate
the ability of the cluster to activate the H2 molecule and
ultimately catalyze the hydrogenation of azobenzene. The
proposed mechanism represents a new example of activation of
small molecules at the bridging sulfur atoms triangle present in
cuboidal Mo3S4 clusters. Interestingly, whereas the previously
reported semihydrogenation of alkynes involves the activation
of the substrate at the bridging sulfide ligands,6 in the present
case those ligands are employed to initially cleave the σ H−H
bond of H2. All in all, the results herein highlight the capability
of these clusters to catalyze the hydrogenation of substrates
with initial activation of either H2 or the substrate. In light of
these new findings, we are currently investigating the
hydrogenation of other unsaturated substrates, and the results
obtained will be reported in due form. Finally, it is worth
noting that based on the structural similarity between Mo3S4
clusters and the basal plane of MoS2, which also display sulfur
atoms triangles linked through Mo atoms, the catalytic
potential of MoS2-based materials in hydrogenation processes
deserves further studies.
Authors
́
Eva Guillamon − Departament de Química Física i Analítica,
Universitat Jaume I, 12071 Castelló, Spain; orcid.org/
0000-0002-3595-5929
Monica Oliva − Departament de Química Física i Analítica,
Universitat Jaume I, 12071 Castelló, Spain; orcid.org/
́
0000-0001-6651-7852
́
Juan Andres − Departament de Química Física i Analítica,
Universitat Jaume I, 12071 Castelló, Spain; orcid.org/
0000-0003-0232-3957
Elena Pedrajas − Departament de Química Física i Analítica,
Universitat Jaume I, 12071 Castelló, Spain
Vicent S. Safont − Departament de Química Física i Analítica,
Universitat Jaume I, 12071 Castelló, Spain; orcid.org/
0000-0003-2709-4230
Complete contact information is available at:
https://pubs.acs.org/10.1021/acscatal.0c05299
Author Contributions
§These authors contributed equally to this work.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Financial support from the Spanish Ministerio de Ciencia,
́
Innovacion y Universidades for projects PGC2018-094417-B-
I00 (E.G., M.O., J.A., R.L., and V.S.S.) and PID2019-
107006GB-C22 (A.G.A. and M.G.B.) is gratefully acknowl-
edged. E.G., M.O., J.A., R.L., E.P., and V.S.S. acknowledge
Universitat Jaume I for projects UJI-B2017-44 and UJI-B2019-
30. A.G.A. and M.G.B. acknowledge the 2014-2020 ERDF
Operational Programme and the Department of Economy,
Knowledge, Business and University of the Regional Govern-
ment of Andalusia for the project FEDER-UCA18-106840.
ASSOCIATED CONTENT
■
sı
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acscatal.0c05299.
́
The authors also thank the Universidad de Cadiz and the
Universitat Jaume I for computational resources and the SCIC
of the Universitat Jaume I for providing us with NMR, mass
spectrometry, and X-ray diffraction techniques. We acknowl-
Detailed synthetic and experimental procedures, cluster
monitoring, crystal structure determination, computa-
tional details, experimental and optimized bond
distances of 1 and 1-H2, potential energy profile,
́
edge the ICIQ NMR Unit staff and Dr. M. Gimenez for their
assistance with the high pressure NMR experiments.
613
https://dx.doi.org/10.1021/acscatal.0c05299
ACS Catal. 2021, 11, 608−614

ACS Catalysis
REFERENCES
pubs.acs.org/acscatalysis
Letter
́
́
Gushchin, A. L.; Fernandez-Trujillo, M. J.; Hernandez-Molina, R.;
Vicent, C.; Algarra, A. G.; Basallote, M. G. Mechanism of [3 + 2]
Cycloaddition of Alkynes to the [Mo3S4(acac)3(py)3][PF6] Cluster.
Chem. - Eur. J. 2015, 21, 2835−2844.
(13) Casewit, C. J.; Coons, D. E.; Wright, L. L.; Miller, W. K.;
Dubois, M. R. Homogeneous Reductions of Nitrogen-Containing
Substrates Catalyzed by Molybdenum (IV) Complexes with μ-Sulfido
Ligands. Organometallics 1986, 5, 951−955.
(14) Algarra, A. G. Computational Insights into the Mechanisms of
H2 Activation and H2/D2 Isotope Exchange by Dimolybdenum
Tetrasulfide Complexes. Eur. J. Inorg. Chem. 2016, 2016, 1886−1894.
(15) Besora, M.; Maseras, F. Microkinetic modeling in homoge-
neous catalysis. Wiley Interdiscip. Rev.: Comput. Mol. Sci. 2018, 8,
No. e1372.
(16) Hernandez-Molina, R.; Sokolov, M. N.; Sykes, A. G. Behavioral
Patterns of Heterometallic Cuboidal Derivatives of [M3Q4(H2O)9]4+
(M = Mo, W; Q = S, Se). Acc. Chem. Res. 2001, 34, 223−230.
(17) Cotton, F. A.; Dori, Z.; Llusar, R.; Schwotzer, W. Oxidative
fragmentation of the cuboid molybdenum-sulfur Mo4S4 cluster core:
synthesis and structures of [Mo3(μ3-S)(μ-S)3([9]aneN3)3]4+ and
{[MoO([9]aneN3)]2(μ-S)2}2+ ([9]aneN3 = 1,4,7-triazacyclononane).
Inorg. Chem. 1986, 25, 3654−3658.
■
(1) Rylander, P. N. Catalytic Hydrogenation in Organic Syntheses;
Academic Press: New York, 1979.
(2) Kubas, G. J.; Heinekey, D. M. Activation of Molecular
Hydrogen. In Physical Inorganic Chemistry; John Wiley & Sons,
2010; pp 189−245.
(3) (a) Mao, J.; Wang, Y.; Zheng, Z.; Deng, D. The rise of two-
dimensional MoS2 for catalysis. Front. Phys. 2018, 13, 138118.
(b) Drescher, T.; Niefind, F.; Bensch, W.; Grunert, W. Sulfide
̈
Catalysis without Coordinatively Unsaturated Sites: Hydrogenation,
Cis-Trans Isomerization, and H2/D2 Scrambling over MoS2 and WS2.
J. Am. Chem. Soc. 2012, 134, 18896−18899. (c) Li, Z.; Zhang, D.; Ma,
J.; Wang, D.; Xie, C. Fabrication of MoS2 microflowers for
hydrogenation of nitrobenzene. Mater. Lett. 2018, 213, 350−353.
(4) (a) Kronberg, R.; Hakala, M.; Holmberg, N.; Laasonen, K.
Hydrogen adsorption on MoS2-surfaces: a DFT study on preferential
sites and the effect of sulfur and hydrogen coverage. Phys. Chem.
Chem. Phys. 2017, 19, 16231−16241. (b) Prodhomme, P.-Y.;
Raybaud, P.; Toulhoat, H. Free-energy profiles along reduction
pathways of MoS2 M-edge and S-edge by dihydrogen: A first
principles study. J. Catal. 2011, 280, 178−195. (c) Sun, M.; Nelson,
A. E.; Adjaye, J. Adsorption and dissociation of H2 and H2S on MoS2
and NiMoS catalysts. Catal. Today 2005, 105, 36−43. (d) Travert, A.;
Nakamura, H.; van Santen, R. A.; Cristol, S.; Paul, J.-F.; Payen, E.
Hydrogen Activation on Mo-Based Sulfide Catalysts, a Periodic DFT
Study. J. Am. Chem. Soc. 2002, 124, 7084−7095.
̈
(5) (a) Sorribes, I.; Wienhofer, G.; Vicent, C.; Junge, K.; Llusar, R.;
Beller, M. Chemoselective Transfer Hydrogenation to Nitroarenes
Mediated by Cubane-Type Mo3S4 Cluster Catalysts. Angew. Chem.,
Int. Ed. 2012, 51, 7794−7798. (b) Pedrajas, E.; Sorribes, I.; Junge, K.;
Beller, M.; Llusar, R. A Mild and Chemoselective Reduction of Nitro
and Azo Compounds Catalyzed by a Well-Defined Mo3S4 Cluster
Bearing Diamine Ligands. ChemCatChem 2015, 7, 2675−2681.
(c) Pedrajas, E.; Sorribes, I.; Gushchin, A. L.; Laricheva, Y. A.;
Junge, K.; Beller, M.; Llusar, R. Chemoselective Hydrogenation of
Nitroarenes Catalyzed by Molybdenum Sulphide Clusters. Chem-
CatChem 2017, 9, 1128−1134. (d) Pedrajas, E.; Sorribes, I.;
́
Guillamon, E.; Junge, K.; Beller, M.; Llusar, R. Efficient and Selective
N-Methylation of Nitroarenes under Mild Reaction Conditions.
Chem. - Eur. J. 2017, 23, 13205−13212. (e) Pedrajas, E.; Sorribes, I.;
Junge, K.; Beller, M.; Llusar, R. Selective reductive amination of
aldehydes from nitro compounds catalyzed by molybdenum sulfide
clusters. Green Chem. 2017, 19, 3764−3768.
́
́
́
(6) Algarra, A. G.; Guillamon, E.; Andres, J.; Fernandez-Trujillo, M.
J.; Pedrajas, E.; Pino-Chamorro, J. A.; Llusar, R.; Basallote, M. G.
́
Cuboidal Mo3S4 Clusters as a Platform for Exploring Catalysis: A
Three-Center Sulfur Mechanism for Alkyne Semihydrogenation. ACS
Catal. 2018, 8, 7346−7350.
̈
(7) Leipzig, B. Uber stufenweise Reduktion des Nitrobenzols mit
begrenztem Kathodenpotential. Z. Elektrochem. Angew. Phys. Chem.
1898, 4, 506−514.
(8) Nakajima, Y.; Suziki, H. Nitrogen-Nitrogen Double Bond
Cleavage of Azobenzene by a Triruthenium Pentahydrido Complex,
(Cp′Ru)3(μ3-H)2(μ-H)3(Cp′ = η5-C5Me5), and Catalytic Hydro-
genation of Azobenzene and 1,2-Diphenylhydrazine. Organometallics
2005, 24, 1860−1866.
(9) (a) Tanifuji, K.; Ohki, Y. Metal-Sulfur Compounds in N2
Reduction and Nitrogenase-Related Chemistry. Chem. Rev. 2020,
120, 5194−5251. (b) Ohki, Y.; Uchida, K.; Tada, M.; Cramer, R. E.;
Ogura, T.; Ohta, T. N2 activation on a molybdenum-titanium-sulfur
cluster. Nat. Commun. 2018, 9, 3200.
(10) Wilkins, R. G. Kinetics and Mechanism of Reactions of Transition
Metal Complexes, 2nd ed.; Wiley-VCH, 1991.
(11) Hendry, P.; Sargeson, A. M. Intramolecular phosphoryl
transfer: chelated phosphoramidate. Inorg. Chem. 1986, 25, 865−869.
(12) (a) Algarra, A. G.; Basallote, M. G. Computational Insights Into
the Reactivity at the Sulfur Atoms of M3S4 (M = Mo, W) Clusters:
The Mechanism of [3 + 2] Cycloaddition With Alkynes. Advances in
Inorganic Chemistry 2017, 70, 311−342. (b) Pino-Chamorro, J. A.;
614
https://dx.doi.org/10.1021/acscatal.0c05299
ACS Catal. 2021, 11, 608−614