Hydrogen production by selective dehydrogenation of HCOOH catalyzed by Ru-biaryl sulfonated phosphines in aqueous solution

Article abstract of DOI:10.1021/cs500655x

The selective dehydrogenation of aqueous solutions of HCOOH/HCOONa to H₂ and CO₂ gas mixtures has been investigated using RuCl₃·3H₂O as a homogeneous catalyst precursor in the presence of different monoaryl-biaryl or alkyl-biaryl phosphines and aryl diphosphines bearing sulfonated groups. All catalytic systems were used in water without any additives and proved to be active at 90 °C, giving high conversions and good TOF values. As an alternative Ru(II) metal precursor, the known dimer [Ru(η⁶-C₆H₆)Cl₂]₂ was also tested as in situ catalyst with selected phosphines as well as an isolated Ru(II)-catalyst with one of them. By using high-pressure NMR (HPNMR) techniques, indications on the nature of the active species involved in the catalytic cycles were obtained.

Full text of DOI:10.1021/cs500655x

Research Article
pubs.acs.org/acscatalysis
Hydrogen Production by Selective Dehydrogenation of HCOOH
Catalyzed by Ru-Biaryl Sulfonated Phosphines in Aqueous Solution
Antonella Guerriero,† Herve Bricout,‡ Katerina Sordakis,§ Maurizio Peruzzini,† Eric Monflier,‡
́
Frederic Hapiot,*,‡ Gabor Laurenczy,*,§ and Luca Gonsalvi*,†
́
́
́
†Consiglio Nazionale delle Ricerche, Istituto di Chimica dei Composti Organometallici (ICCOM-CNR), Via Madonna del Piano 10,
50019 Sesto Fiorentino, Firenze, Italy
‡Universite
rue Jean Souvraz, SP18, F-62307 Lens Ced
§Institut des Science et Ingen
́
Lille Nord de France, CNRS UMR 8181, Unite
́
de Catalyse et de Chimie du Solid, UCCS U Artois, Faculte
́
Jean Perrin,
́
ex, France
́
́
ierie Chimiques, Ecole Polytechnique Fed
́
erale de Lausanne (EPFL), SB ISIC, Station 6, CH-1015
́
Lausanne, Switzerland
S
* Supporting Information
ABSTRACT: The selective dehydrogenation of aqueous solutions of
HCOOH/HCOONa to H2 and CO2 gas mixtures has been
investigated using RuCl3·3H2O as a homogeneous catalyst precursor
in the presence of different monoaryl-biaryl or alkyl-biaryl phosphines
and aryl diphosphines bearing sulfonated groups. All catalytic systems
were used in water without any additives and proved to be active at 90
°C, giving high conversions and good TOF values. As an alternative
Ru(II) metal precursor, the known dimer [Ru(η6-C6H6)Cl2]2 was also
tested as in situ catalyst with selected phosphines as well as an isolated
Ru(II)-catalyst with one of them. By using high-pressure NMR
(HPNMR) techniques, indications on the nature of the active species involved in the catalytic cycles were obtained.
KEYWORDS: aqueous-phase catalysis, formic acid, hydrogen generation, ruthenium, sulfonated phosphines
were recently obtained by Beller and co-workers who
investigated the ruthenium-catalyzed dehydrogenation of
HCOOH under mild conditions in the presence of different
amines and HCOOH/amine ratios.12 An optimized catalytic
system was obtained by combining the [Ru(η6-C6H6)Cl2]2
precursor with diphosphines such as 1,2-bis-
(diphenylphosphino)methane (dppm) and 1,2-bis-
(diphenylphosphino)ethane (dppe).13 Remarkable results
were presented by the same research group concerning the
use of iron(II) tetrafluorborate together with tris[(2-
diphenylphosphino)ethyl]phosphine [P(CH2CH2PPh2)3, PP3]
in propylene carbonate, which afforded TOFs (turnover
frequencies) up to 9425 h−1 and TONs (turnover numbers)
higher than 92 000 at 80 °C without the need of an added
base.14 Good catalytic performances in HCOOH dehydrogen-
ation were also reported by some of us with the octahedral
complexes [Ru(κ4-NP3)Cl2] and [Ru(κ3-triphos)(MeCN)3]-
(PF6)2 (NP3 = N(CH2CH2PPh2)3, triphos = MeC-
(CH2PPh2)3) in the presence of n-octylamine at 80 °C
together with a detailed computational study which clarified
some mechanistic details of these systems.15 All the
aforementioned catalysts operate in organic solvents.
INTRODUCTION
■
In the quest for efficient hydrogen sources1 from cheap and safe
materials, for on-demand use in devices such as fuel cells,
producing only water as byproducts,2 the activation of
abundant hydrogen-rich organic compounds has been proposed
as a possible solution.3 In this context, formic acid (HCOOH),
which contains 4.4 wt % of hydrogen, as well as its conjugate
base, formate anion, are receiving considerable attention as
promising materials for H-storage and delivery.4 The selective
decomposition of HCOOH through dehydrogenation yields
only H2 and CO2 gas mixtures. Furthermore, the CO2 derived
from this reaction can be recycled by hydrogenation to give
HCOOH, resulting in a reversible hydrogen storage−release
system with potentially zero carbon emission.5 Both dehydro-
genation of HCOOH and hydrogenation of CO2 need
appropriate catalysts to provide selective reactions under mild
conditions, including both homogeneous and heterogeneous
catalysis.3d,6
The first study on the homogeneous dehydrogenation of
formic acid (FA) dates back to 1967, when Coffey reported a
highly active iridium phosphine complex for this reaction, with
TOF (turnover frequency) of 1187 h−1 albeit at elevated
temperatures.7 Afterward, many research groups contributed to
this field of research and several iron,8 ruthenium,9 rhodium,10
and iridium11 complexes were described in the literature as FA
dehydrogenation homogeneous catalysts. Important results
Received: May 13, 2014
Revised: July 24, 2014
Published: August 5, 2014
© 2014 American Chemical Society
3002
dx.doi.org/10.1021/cs500655x | ACS Catal. 2014, 4, 3002−3012

ACS Catalysis
Research Article
Chart 1. Sulfonated Aryl-Biaryl and Alkyl-Biaryl Monophosphines
Table 1. Water Solubility, 31P{1H} NMR Chemical Shifts of Sulfonated Monophosphines and 1JP−Se Coupling Constants of the
Corresponding Selenide Derivatives, Giving a Basicity Scale of the Parent Compound22
a
b
phosphine
L1 (MBMS)
S(H2O)25 °C
31P{1H} (δ, ppm)
1JP−Se [Hz]
ref
280 g/L
1.0 kg/L
1.0 kg/L
850 g/L
1.0 kg/L
50 g/L
−5.9
−5.5
−6.4
−7.7
−6.9
0.4
739
752
746
739
739
707
734
754
745
20b
20a
L2 (MBTS)
L3 (BBTS)
20a
L4 (Ph(P(bisbiph)DS)
L5 (P(biph)3TS)
L6 (di-CyMBMS)
L7 (CyBBDS)
m-TPPTS
20b
20a
20b
20b
19a, b
19c
520 g/L
1.1 kg/L
12 g/L
−6.5
−5.1
−3.9
m-TPPMS
a
b
Chemical shift values in D2O. Coupling constant values obtained in absolute EtOH.
A different approach in this field was followed by Laurenczy
temperatures, making this catalytic system suitable for fuel cell
applications. The influence of different hydrophilic ligands on
the FA dehydrogenation in the presence of RuCl3·3H2O has
also been investigated in detail. These studies showed that
ligand basicity, its hydrophilic properties, and its steric effects
were the main parameters which influenced the catalytic
activity.17
et al., who performed the selective FA dehydrogenation in
aqueous phase by using hydrophilic ruthenium-based catalysts,
generated in situ from [Ru(H2O)6](tos)2 (tos = toluene-4-
sulfonate) or commercial hydrated RuCl3 in the presence of the
water-soluble ligand meta-trisulfonated triphenylphosphine (m-
TPPTS).16 The resulting Ru(II) species catalyzed the
dehydrogenation reaction already at 25 °C, reaching a TOF
of 460 h−1 at 120 °C. Notably, this system was shown to be
active in a wide temperature range between 25 and 170 °C,
giving in all cases conversions of 90−95%. Furthermore, the
hydrogen produced was of high purity, and no CO formation
was observed by FTIR (detection limit 2 ppm) even at high
The most common method to convey active catalysts in
water is to design ligands containing hydrophilic substituents,
−
and the anionic sulfonate group (−SO3 ) is one of the most
attractive because of its stability in a wide range of reaction
conditions.18
3003
dx.doi.org/10.1021/cs500655x | ACS Catal. 2014, 4, 3002−3012

ACS Catalysis
Research Article
Hereby, we present results on catalytic FA dehydrogenation
using a small library of monodentate alkyl-biaryl, aryl-biaryl
sulfonated phosphines, and selected tetrasulfonated diphos-
phines with Ru(III) and Ru(II) metal precursors. Mechanistic
studies using high-pressure (HP) NMR techniques gave
indications on the role of stable Ru-hydrido species which are
formed during catalysis.
In this study, the above-mentioned sulfonated phosphines
were applied as stabilizing ligands in the aqueous-phase
dehydrogenation of formic acid, by using either the
commercially available RuCl3·3H2O or the [Ru(η6-C6H6)Cl2]2
dimer as catalyst precursors. The latter was also used as
synthon to obtain an isolated Ru(II)-arene phosphine complex,
which was also tested as catalyst. Several tests were performed
under different reaction conditions and the recycling of the
catalysts was also investigated. For the first time, the bidentate
phosphines L10, L9, and L8 were used with Ru(III) or Ru(II)
precursors for catalytic FA dehydrogenation reactions.
Formic Acid Dehydrogenation Tests with RuCl3·3H2O/
Monophosphines. On the basis of the results obtained with
Ru/m-TPPTS systems (Ru = [Ru(H2O)6](tos)2 (tos =
toluene-4-sulfonate) or RuCl3·3H2O),16 we decided to test
the water-soluble mono- and diphosphines listed in Charts 1
and 2 under comparable catalytic conditions.
The dehydrogenation of a standard concentration formic
acid/sodium formate solution (4 M, 9:1) was carried out in
water without any other additive. The presence of a small
amount of formate is useful for the activation of the catalytic
species, and the 9:1 ratio was chosen as it was shown previously
to be a good compromise between conversion and reaction
rate.5b
RESULTS AND DISCUSSION
■
Among sulfonated ligands,19 the phosphines shown in Chart 1
were synthesized by some of us20 and previously used in the
Pd-catalyzed cleavage of allylcarbonate (Tsuji−Trost reaction)
and in the aqueous biphasic Rh-catalyzed hydroformylation of
terminal alkenes. All these ligands contain at least one biaryl
−
group and differ in the number and position of −SO3
substituents. Besides the influence on the water solubility and
basicity properties (Table 1), these structural differences can
have a significant impact on their coordination behavior and,
therefore, on the catalytic activity of the corresponding
complexes. On the other hand, the water-soluble diphosphines
1,2-bis[bis(m-sodiosulfonatophenyl)phosphino]ethane
(DPPETS, L10), 1,2-bis[bis(m-sodiosulfonatophenyl)-
phosphino]propane (DPPPTS, L9), and 1,2-bis[bis(m-sodio-
sulfonatophenyl)phosphino]butane (DPPBTS, L8) (see Chart
2) were reported as useful ligands to synthesize several Rh, Pt,
and Pd coordination compounds which were able to catalyze
different reactions in aqueous phase.21
The selective dehydrogenation of HCOOH into H2 and CO2
was run under atmospheric pressure in thermostated gas
reactors and monitored by gas volume evolution collected in
gas burets. Using 9.0 mmol of FA, complete conversion is
expected to develop 0.440 L of gas mixture, as calculated from
the following equation:
Chart 2. Sulfonated Aryl Diphosphines
V
= x mol · 24.48 L mol−1 + x mol · 24.36 L mol−1
298K
(eq 1)
where x = mol FA, 24.48 L mol−1 = estimated molar volume of
H2 at 298 K, 24.36 L mol−1 = estimated molar volume of CO2
at 298 K.
a
Table 2. Formic Acid Dehydrogenation Catalyzed by Ru/Aryl-Biaryl Monophosphine Systems (First Catalytic Cycle)
d
f
d
g
h
f
entry
ligand
Ru/L vol (mL) after 5 min
TOF (h−1) after 5 min
max vol (mL) [t(min)]
max conv %
final TON
final TOF (h−1
)
b
1
L1
1/2
1/2
1/4
1/2
1/2
1/4
1/2
1/2
1/4
1/2
1/2
1/4
1/2
1/2
1/4
1/2
1/2
−
270
80
1188
348
385 [10]
395 [80]
400 [10]
380 [5]
87.5
89.9
91
141
144
146
139
133
154
141
139
143
150
144
148
152
121
150
139
141
33
846
108
876
1668
160
616
846
139
858
450
108
1776
700
242
300
93
c
2
L1
b
3
L1
390
380
45
1716
1668
192
b
4
L2
86.4
83
c
5
L2
365 [50]
420 [15]
385 [10]
380 [60]
390 [10]
410 [20]
395 [80]
405 [5]
b
6
L2
385
380
75
1692
1668
324
95.5
87.5
86.4
88.7
93.3
89.9
92.1
94.4
75.1
93.3
86.4
87.6
20.5
b
7
L3
c
8
L3
b
9
L3
L4
385
360
80
1692
1584
348
b
10
c
11
L4
b
e
12
L4
405
405
60
1776
1776
264
b
e
13
L5
415 [13]
330 [30]
410 [30]
380 [90]
385 [15]
90 [330]
c
14
L5
b
e
15
L5
405
1776
1092
1644
60
b
16
m-TPPMS
m-TPPTS
no ligand
250
375
15
b
17
563
6
b
18
a
b
c
d
Reaction conditions: RuCl3·3H2O, 0.056 mmol; HCOOH, 9 mmol; HCOONa, 1 mmol; H2O, 2.5 mL; 750 rpm. At 90 °C. At 60 °C The values
e
f
are averaged from three to six measurements with a reproducibility of 10%. Volume reached after 3 min. Calculated as (mmol produced gas/mmol
g
h
initial Ru) × h−1. Defined as (mmol produced gas/mmol initial HCOOH) × 100. Calculated as (mmol produced gas/mmol initial Ru).
3004
dx.doi.org/10.1021/cs500655x | ACS Catal. 2014, 4, 3002−3012

ACS Catalysis
Research Article
The precatalysts were formed in situ by using the
commercially available RuCl3·3H2O (22.5 mM solution in
water) as catalyst precursor together with at least two
equivalents of the appropriate phosphine ligand, using literature
reaction conditions,16 namely, 90 °C with ratios Ru/HCOOH
= 1:161 and Ru/ligand = 1:2. In selected cases, the Ru/ligand
ratio was increased to 1:4 to verify the effect of a higher ligand
concentration on the overall stability and long-term activity of
the catalytic systems, for example, upon recycles. Catalytic runs
were carried out in the presence of the sulfonated mono-
phosphines (see Chart 1) and selected results are reported in
Table 2. In the case of L1−L5, conversions in the range of 86−
93% were obtained within few minutes for the first catalytic
cycle (Table 2, entries 1, 4, 7, 10, and 13). As expected by
comparison with the known RuCl3/m-TPPTS system, a 100%
conversion was never reached as the formate salt is not
dehydrogenated under these conditions.16
In the presence of two equivalents of L2, 86.4% conversion
was achieved in 5 min, giving the highest TOF = 1668 h−1
under these conditions. For all the other monophosphines
tested, with the exception of L1 (Table 2, entry 1), initial TOF
values higher than 1500 h−1 were calculated after 5 min. The
highest conversion (ca. 92.1%) for the first catalytic cycle was
observed in the presence of L5, which produced 405 mL of gas
after only 3 min (initial TOF = 2960 h−1). Gas evolution
stopped after 13 min, reaching 94.4% conversion corresponding
to a final TOF of 700 h−1.
Table 3. Recycling Tests for HCOOH Dehydrogenation by
Using Ru(III)/L2 (1:2) System at 90 °C
a
b
c
d
e
cycle
max vol (mL)[t(min)]
conv %
TON
TOF (h−1
)
I
II
380 [5]
340 [90]
360 [60]
360 [30]
380 [15]
385 [15]
390 [20]
365 [30]
380 [15]
385 [15]
390 [15]
86.4
77.4
81.9
81.9
86.4
87.6
88.7
83.0
86.4
87.6
88.7
139
124
132
132
139
141
143
133
139
141
143
1668
83
III
IV
V
132
264
556
564
429
266
556
564
572
VI
VII
VIII
IX
X
XI
a
Reaction conditions: RuCl3·3H2O, 0.056 mmol; L2, 0.112 mmol;
HCOOH, 9 mmol; HCOONa, 1 mmol; H2O, 2.5 mL; 750 rpm. For
each recycle, 9 mmol of neat HCOOH were added by microsyringe.
b
The values are averaged from three to six measurements with a
c
reproducibility of 10%. Defined as (mmol produced gas/mmol
initial
d
HCOOH) × 100. Calculated as (mmol produced gas/mmol initial Ru).
e
Calculated as (mmol produced gas/mmol initial Ru) × h−1. The number
of recycles was determined by time constraint rather than the effective
loss in activity of the catalyst.
from the fourth cycle on, with average conversions of 86.3%. A
similar behavior was observed with L3, L4 and L5 (see Figures
S1, S2, S3 in Supporting Information (SI)), in accordance with
literature data obtained with the methylammonium-substituted
aryl phosphines.23
For all the catalytic tests, at the end of each cycle, aliquots of
the gas mixtures were collected in a 10 cm gas cell (KBr disks)
and analyzed by FTIR spectroscopy using Wills’ method (see
the Experimental Section and Figures S6−S9 in SI).9c In all
cases, the spectra showed only the peaks corresponding to CO2
with no traces of CO (detection limit of 0.02%), in line with
results obtained by some of us with benchmark sulfonated
ligands.16
Considering the promising results obtained at 90 °C, we
decided to explore the effect of temperature on HCOOH
dehydrogenation using the most active systems by repeating
selected tests at a lower temperature. As reported in Table 2, in
all runs performed at 60 °C (entries 2, 5, 8, 11, and 14), fairly
similar conversions and TON values were obtained at the end
of the first cycle, as expected with a decrease in TOFs (160 h−1
vs 1668 h−1 in the case of L2).
Another important issue to be considered is the possible
phosphine degradation during the catalytic runs and upon
recycling. This was assessed by taking aliquots of the reaction
mixtures after the end of the catalytic runs and carrying out
31P{1H} NMR analyses. It was shown that ligand oxidation
occurred under these conditions in most cases, as witnessed by
the presence of signals in the range 30−40 ppm.24 Although it
is generally accepted that introducing an excess of phosphine at
the beginning of the experiments can have a detrimental effect
on the rate of dehydrogenation, we decided to verify whether a
known excess of ligand could act as a reservoir to stabilize the
metal precursor.
As for m-TPPMS,17 in the case of L1, foam formation was
observed upon gas evolution mainly during the first cycle.
Interestingly, this behavior was not exhibited by the other
monophosphines tested, probably due to a lower surfactant
effect and higher water solubility.
In order to compare the activities of our systems with
benchmark ligands, some catalytic tests were carried out also
with m-TPPMS and m-TPPTS using our experimental setup
(Table 2, entries 16 and 17). During the first cycle, m-TPPMS
and its biaryl analogue L1, reached comparable conversions
after 5 min of reaction. Interestingly, the final TOF of the Ru/
L1 system was significantly higher than for Ru/m-TPPMS (846
h−1 vs 93 h−1, respectively). A similar situation was observed for
Ru/m-TPPTS and Ru/L2 systems, with comparable con-
versions after 5 min but a higher TOF for the latter (Table 2,
entries 4 and 17).
The results obtained are in agreement with a higher ligand
basicity (see Table 1) than m-TPPTS. Recent examples
demonstrating that higher basicity and hence a stronger σ-
donation of ligands can promote HCOOH dehydrogenation
were shown by some of us23 using methylammonium-
substituted aryl phosphines.
In order to check the activities and stabilities of the catalytic
systems for repeated runs, recycling experiments were also
performed, by adding neat HCOOH (9 mmol) to the reaction
mixtures kept at 90 °C at the end of each run, determined by
observation that gas evolution had finished. With the exception
of L1 that catalyzed HCOOH dehydrogenation for only three
consecutive cycles giving decreasing conversions (63%, TOF =
56 h−1 and 68%, TOF = 55 h−1 for second and third cycles,
respectively), the other four ligands tested, resulted to be active
for several cycles. Selected results obtained with L2 are
summarized in Table 3. Eleven consecutive runs were carried
out without loss of catalytic activity, obtaining an overall TON
of 1506. A slightly slower rate was displayed during the second
and third cycles, to reach comparable conversions and TOFs
Selected experiments were repeated using a Ru/ligand ratio
of 1:4 at 90 °C. As shown in Table 2 (entries 3 and 9),
conversions and TOFs were comparable to those obtained with
a 1:2 ratio in the case of L1 and L3. On the other hand, in the
case of L2, an improvement of ca. 10% in terms of conversion
3005
dx.doi.org/10.1021/cs500655x | ACS Catal. 2014, 4, 3002−3012

ACS Catalysis
Research Article
was observed at the end of the first cycle (95.5% vs 86.4%,
Table 2, entries 4 and 6), although the reaction rate was
significantly slower. No improvements were observed for
recycling experiments and, for all the three aryl phosphines
considered, not more than four cycles were obtained under
these conditions (see Table S1 in SI).
A different behavior was observed with the highly active L4
and L5. With a Ru/ligand = 1:4 ratio, both catalytic systems
showed the same initial TOF = 1776 h−1 after 5 min, the
highest produced (Table 2, entries 12 and 15). Upon recycling
(Figure 1), the Ru/L5 system was active for four cycles
maintaining high and constant initial TOFs (5 min) and high
conversions (>90%).
or biphasic catalysis due to their chelating properties,21 but to
the best of our knowledge they have never been tested for FA
dehydrogenation. Selected results are summarized in Table 5.
In the case of L8 at 90 °C, conversions of ca. 92% were
obtained in the first cycle both using a 1:1 or 1:2 ratio Ru/
ligand, albeit at different rates (TOF = 226 h−1 and 115 h−1,
respectively), as shown in Figure 2. The recycling tests gave
comparable results both with one and two equivalents of L8,
providing a final TON after six cycles of 770 and 841,
respectively (see Table S2 in SI). A different behavior was
observed in the case of L9 and L10 ligands, bearing backbones
with shorter alkyl chains. At Ru/L = 1:1 ratio, lower
conversions (less than 80%) and turnover frequencies were
achieved during the first cycle (Figure 2). Under these
conditions, both systems remained active for only three cycles,
giving a final TON of 388 for L9 and 342 for L10, respectively.
Interestingly, using Ru/L = 1:2 ratios, both catalytic systems
gave ca. 89% conversions at comparable rates during the first
cycle but continued to be active for several recycles. In
particular, with L9 (see Figure S4 in SI) more than seven
consecutive cycles were run, resulting in conversions higher
than 80% and a final TON of 943. For L10, more than six
catalytic cycles were possible with constant conversions and
TOFs and a final TON of 804.
On the basis of these experiments, it can be concluded that
the use of Ru/diphosphines in 1:2 ratios gives high conversions
and TONs but lower TOFs than the corresponding Ru/aryl-
biarylphosphine analogues, so a compromise between pre-
catalyst stability and high reaction rates must be considered in
the presence of this series of ligands.
Finally, in order to exclude the formation of soluble
ruthenium colloidal particles, mercury poisoning test reactions
were performed in the presence of L2 and L8. As illustrated in
Table 6, when some drops of Hg(0) were added to catalytic
tests using Ru/L2 = 1:2 and Ru/L8 =1:1 at 90 °C, comparable
conversions and rates as observed without the addition of
mercury were obtained, suggesting the absence of active metal
particles in solution.
Figure 1. Initial TOFs calculated for conversions after 5 min upon
recycling of catalytic system Ru/L5 (1:4) at 90 °C.
Some biaryl phosphines bearing cycloalkyl instead of aryl
groups on P, namely, L6 and L7 (Chart 1), were also tested in
catalysis under the usual conditions (Table 4). Using a Ru/
phosphine = 1:2 ratio, low conversions were observed during
the first cycle (61.4% after 5 h for L6 and 54.6% after 4 h with
L7, respectively). Although these ligands are more basic than
their aryl analogues (Table 1), lower water solubilities and
different steric effects could play a detrimental role for these
systems. Despite the lower activity in HCOOH dehydrogen-
ation compared to aryl biaryl analogues, the FTIR analysis of
the gas mixtures again showed only the signals corresponding
to CO2 (Figure S9 in SI).
Formic Acid Dehydrogenation Catalyzed by RuCl3·
3H2O and Sulfonated Diphosphines. The decomposition
of formic acid into H2 and CO2 was also investigated using
RuCl3·3H2O in the presence of tetrasulfonated diphosphines
(NaSO3C6H4)2P(CH2)nP(C6H4SO3Na)2 (n = 2−4) with
different lengths of the carbon atoms chain in the backbone
(Chart 2). These ligands have found applications in water phase
Formic Acid Dehydrogenation Using [Ru(η6-C6H6)Cl2]2
as Ru(II) Precursor. In a second series of experiments, the
catalytic performances were investigated using a preformed
Ru(II) complex, namely [Ru(η6-C6H6)Cl2]2, instead of RuCl3·
3H2O. This Ru(II) η6-arene dimer has been previously tested
by Beller and co-workers as catalytic precursor for hydrogen
generation from neat HCOOH/NEt3 (5:2) adducts, obtaining
good results at 40 °C in the presence of PPh3 (TON = 452
after 3 h) and dppe (dppe =1,2-bis(diphenylphosphino)ethane)
as stabilizing ligands (TON = 1376 after 3 h).12b Further
studies showed that the catalytic activity can be improved using
a [Ru(η6-C6H6)Cl2]2/ligand = 1:6 ratio in the presence of the
adduct HCOOH/NMe2(n-C6H11) (5:4).25
a
Table 4. HCOOH Dehydrogenation with Ru/Alkyl-Biarylphosphines at 90 °C (First Catalytic Cycle)
b
c
d
e
phosphine
Ru/L
max vol (mL) [t(min)]
conv %
TON
TOF (h−1
)
L6
L6
L7
L7
1/2
1/4
1/2
1/4
320 [300]
150 [250]
225 [240]
275 [240]
72.8
34.1
54.6
62.5
117
55
23
13
20
25
82
100
a
b
Reaction conditions: RuCl3·3H2O, 0.056 mmol; HCOOH, 9 mmol; HCOONa, 1 mmol; H2O, 2.5 mL; 90 °C; 750 rpm. The values are averaged
c
d
from three to six measurements with a reproducibility of 10%. Defined as (mmol produced gas/mmol initial HCOOH) × 100. Defined as (mmol
e
produced gas/mmol initial Ru). Defined as (mmol produced gas/mmol initial Ru) × h−1.
3006
dx.doi.org/10.1021/cs500655x | ACS Catal. 2014, 4, 3002−3012

ACS Catalysis
Research Article
a
Table 5. HCOOH Dehydrogenation with Ru/Diphosphines at 90 °C (First Catalytic Cycle) Using Ru/L = 1:1 and 1:2
b
c
d
e
phosphine
Ru/L
max vol (mL) [t(min)]
conv %
TON
TOF (h−1
)
L8 (DPPBTS)
L8 (DPPBTS)
L9 (DPPPTS)
L9 (DPPPTS)
L10 (DPPETS)
L10 (DPPETS)
1/1
1/2
1/1
1/2
1/1
1/2
413 [40]
395 [75]
335 [85]
400 [45]
345 [105]
390 [40]
93.9
89.9
76.2
91.0
78.5
88.7
151
144
122
146
126
143
226
115
86
195
72
214
a
b
Reaction conditions: RuCl3·3H2O, 0.056 mmol; HCOOH, 9 mmol; HCOONa, 1 mmol; H2O, 2.5 mL; 90 °C; 750 rpm. The values are averaged
c
d
from three to six measurements with a reproducibility of 10%. Defined as (mmol produced gas/mmol initial HCOOH) × 100. Defined as (mmol
e
produced gas/mmol initial Ru). Defined as (mmol produced gas/mmol initial Ru) × h−1.
By comparing these results with those obtained with RuCl3·
3H2O under the same reaction conditions (for comparison, see
Table 2), it is observed that during the first cycle all systems
gave almost the same conversions and TONs, but lower final
TOFs were obtained with the Ru(II) precursor. During
recycling experiments, all Ru(II)/monophosphine systems
showed a constant activity from the first cycle (Figure 3),
suggesting, as expected, that the active species are formed
efficiently from the beginning, without the need for initial metal
reduction from Ru(III) to Ru(II). The two systems containing
L2 and L4 ligands were more active, with more than seven
cycles obtained (TON > 1000 after seven cycles).
Among bidentate ligands, L8 and L10 were selected and used
in 1:1 ratio with [Ru(η6-C6H6)Cl2]2 at 90 °C, obtaining similar
data both in the first cycle (Table 8, entries 5 and 6) and in the
recycles. Although with L8 comparable conversions and TONs
were obtained with both Ru(III) and Ru(II) precursors, a
Figure 2. HCOOH dehydrogenation reaction profiles (conversion %
vs time) for the first cycle runs using Ru/diphosphines at different Ru/
remarkable difference was observed for the L10 ligand. Using
RuCl3, two equivalents per ruthenium atom of this diphosphine
were needed to have a satisfactory activity, whereas with
[Ru(η6-C6H6)Cl2]2 good results were achieved, both in the first
cycle and during recycles, with one equivalent of the ligand per
ruthenium atom (Table 8, entry 6), with slightly better final
conversions and TONs than those observed with L8. This
observation is in agreement with data reported by Beller on
FA/amine dehydrogenation using the [Ru(η6-C6H6)Cl2]2/
diphosphine systems where the presence of longer carbon
chains in the bidentate ligand led to a decrease in the catalytic
activity.12b The FTIR spectra of gas mixtures run at the end of
each experiment showed again the presence of peaks as a result
of CO2 only.
L ratios, at 90 °C. Legenda: Ru/L8 = 1:1 (black squares), Ru/L8 = 1:2
(red circles), Ru/L9 = 1:1 (green triangles), Ru/L9 = 1:2 (yellow
circles), Ru/L10 = 1:1 (pink crosses), Ru/L10 = 1:2 (blue triangles).
The trend line is shown as a guide and should not be intended as a
mathematical fitting of the data.
Table 6. Comparison of Selected Catalytic Tests and
a
Recycling with and without Hg(0)
d
d
max vol mL [t(min)]
max vol mL [t(min)]
b
b
c
c
cycle
[Ru]/L2
[Ru]/L2/Hg
[Ru]/L8
[Ru]/L8/Hg
I
380 [5]
340 [90]
360 [60]
360 [30]
380 [15]
390 [5]
345 [90]
325 [120]
380 [30]
390 [15]
400 [35]
360 [45]
340 [60]
360 [60]
375 [60]
395 [35]
390 [45]
385 [60]
380 [60]
385 [60]
II
III
IV
V
[Ru(η6-C6H6)Cl2]2 was also tested in association with m-
TPPTS and m-TPPMS, as to the best of our knowledge, this
combination has not yet been described. Both systems showed
comparable results of conversion in the first catalytic cycle
(Table 7, entries 7 and 8), although the latter ligand was less
active in recycles (see Figure S5 in SI). Finally, a blank test was
performed with [Ru(η6-C6H6)Cl2]2 in the absence of
phosphines (entry 9). In this run, ca. 90% conversion was
obtained in the first cycle, but the reaction was very slow (6 h,
TOF = 24 h−1). Furthermore, a black precipitate was formed at
the end of the first cycle, and the catalytic activity dropped
severely already in the first recycle (6.8% conversion after 60
min). Although a detailed study of the mechanism active in the
presence of this phosphine-free complex was beyond the
purpose of the present study, it cannot be excluded that
different Ru-arene active species are formed during the catalytic
runs under these conditions, perhaps involving polymetallic
assemblies such as Ru-arene hydrido clusters (vide infra),26 or
(in the case of arene decoordination at higher temperatures)
a
Reaction conditions: RuCl3·3H2O, 0.056 mmol; HCOOH, 9 mmol;
b
HCOONa, 1 mmol; H2O, 2.5 mL; 750 rpm; 90 °C. Ru/L2 = 1:2.
c
d
Ru/L8 = 1:1. The values are averaged from three to six
measurements with a reproducibility of 10%.
Thus, in order to compare the effect of the choice of this
metal precursor in water-phase FA dehydrogenation, selected
sulfonated ligands were tested together with [Ru(η6-C6H6)-
Cl2]2 at 90 °C using a Ru/ligand = 1:2 ratio (monophosphines)
and 1:1 ratio (diphosphines), keeping the other reaction
parameters constant. Among the monophosphine ligands,
catalytic tests were performed with L1, L2, L4, and L5. As
reported in Table 7, although ligands L1, L4 and L5 gave
similar conversions at the same rate in the first cycle (entries 1,
3, and 4), L2 showed also with this metal precursor the highest
final TOF value (429 h−1), giving a 88.7% HCOOH conversion
in 20 min.
3007
dx.doi.org/10.1021/cs500655x | ACS Catal. 2014, 4, 3002−3012

ACS Catalysis
Research Article
a
Table 7. Results of Catalytic HCOOH Dehydrogenation (First Cycle) in the Presence of [Ru(η6-C6H6)Cl2]2 at 90 °C
d
e
f
g
e
g
entry
ligand
L1
vol (mL) after 5 min
TOF (h−1) after 5 min
max vol (mL) [t(min)] conv % TON
TOF (h−1
)
TON tot. [no. cycles]
b
1
288
320
140
380
125
47
1264
1404
612
415 [60]
390 [20]
400 [60]
395 [30]
405 [60]
412 [67]
395 [60]
400 [30]
395 [360]
94.4
88.7
91.0
89.9
92.1
93.7
89.9
91.0
89.9
152
143
146
144
148
151
144
146
144
152
429
146
288
148
135
144
292
24
761 [5]
1020 [7]
1027 [7]
578 [4]
712 [5]
738 [5]
730 [5]
428 [3]
155 [2]
b
2
L2
b
3
L4
b
4
L5
1668
552
c
5
L8
c
6
L10
206
b
7
m-TPPTS
m-TPPMS
no ligand
160
380
50
696
b
8
1668
b
9
216
a
d
f
b
c
Conditions: [Ru]/HCOOH = 1:161; 90 °C. [Ru], 0.028 mmol; monophosphine, 0.112 mmol. [Ru], 0.028 mmol; diphosphine, 0.056 mmol.
e
The values are averaged from three to six measurements with a reproducibility of 10%. Defined as (mmol produced gas/mmol initial Ru) × h−1.
Defined as (mmol produced gas/mmol initial HCOOH) × 100. Defined as (mmol produced gas/mmol initial Ru). The number of recycles was determined by
time constraint rather than the effective loss in activity of the catalytic system.
g
catalyst for FA dehydrogenation. The results summarized in
Table 8 showed that 2 was active for several cycles giving an
average 89% conversion and a final TON of 1002 after seven
cycles. This value is very similar to that obtained with [Ru(η6-
C6H6)Cl2]2/L2 in situ system under the same reaction
conditions (total TON = 1020 after seven cycles, see Table
S3 in SI); however, preformed complex 2 gave higher TOFs
during recycling.
Synthesis of [RuCl2(MBMS)2]2 (1) and [Ru(η6-C6H6)Cl-
(MBTS)2]Cl (2). The syntheses of defined Ru complexes to be
used as precatalysts instead of in situ formation were attempted
following various methods.
On the basis of the procedures reported in literature for m-
TPPMS,19c,28 [RuCl2(PPh3)3] and L1 (2.4 equiv) were reacted
in THF under reflux conditions. After 4 h, a brown precipitate
was filtered off, dissolved in CD3OD, and analyzed by 31P{1H}
Figure 3. Reaction profiles for HCOOH dehydrogenation recycling
experiments (conversion % vs time) using [Ru(η6-C6H6)Cl2]2/L2 =
1:4 molar ratio at 90 °C. Cycle I (black squares), II (red circles), III
(green triangles), IV (light blue diamonds), V (pink crosses), VI (dark
blue triangles), VII (gray squares). The trend line is shown as a guide
and should not be intended as a mathematical fitting of the data.
NMR showing a broad singlet at 53.8 ppm in ca. 50% NMR
purity, the rest being free phosphine oxide, at a chemical shift
value compatible with a species such as [RuCl2(MBMS)2]2
(1).28 The complex was found to be poorly stable in solution,
giving undefined decomposition products when left in D2O for
30 min, as observed by 31P{1H} NMR, and therefore, it was not
considered further for catalytic tests.
Table 8. Results of Catalytic HCOOH Dehydrogenation and
Recycles in the Presence of [Ru(η6-C6H6)Cl(MBTS)2]Cl (2)
The reaction of phosphine MBTS (L2) with [Ru(η6-
C6H6)Cl2]2 (L2: Ru = 2) gave complex [Ru(η6-C6H6)Cl-
(MBTS)2]Cl (2) in 83% purity as a microcrystalline powder,
which was sufficiently stable both as a solid and in water
solution and was used as an isolated catalyst (see Table 8).
Complex 2 was also converted selectively to the correspond-
ing monohydride [Ru(η6-C6H6)H(MBTS)2]Cl (3) by two
independent reactions. In one case, it was dissolved in D2O and
pressurized with 40 bar H2. After 3 h at 60 °C, the 31P{1H} and
1H NMR spectra showed a singlet at 53.5 ppm and a triplet at
−9.41 ppm (2JHP = 36 Hz), respectively. Furthermore, in the
positive range of chemical shifts of 1H NMR spectrum, a singlet
at 5.62 ppm suggested that benzene was still coordinated to the
metal.29 The same 31P and 1H NMR signals were also obtained
by reacting 2 (25 mg, 0.016 mmol) with an excess of HCOONa
(11 mg, 0.16 mmol) in degassed D2O (0.8 mL). After stirring
for 2 h at 60 °C, the reaction mixture was analyzed by 31P{1H}
a
at 90 °C
b
c
c
d
cycle
max vol (mL) [t(min)]
conv %
TON
TOF (h−1
)
I
190 [70]
185 [50]
195 [30]
200 [30]
205 [20]
200 [20]
195 [20]
86.2
83.9
88.5
90.8
93.1
90.8
88.5
139
135
143
146
150
146
143
119
162
286
292
450
438
429
II
III
IV
V
VI
VII
a
Reaction conditions: [Ru(η6-C6H6)Cl(MBTS)2]Cl, 0.028 mmol;
HCOOH, 4.51 mmol; HCOONa, 0.50 mmol; H2O, 2 mL; 750
rpm. For each recycle, 4.51 mmol of neat HCOOH were added by
b
microsyringe. Expected volume for 100% conversion, 0.220 mL. The
values are averaged from three to six measurements with a
c
reproducibility of 10%. Defined as (mmol produced gas/mmol
initial
c
HCOOH) × 100. Calculated as (mmol produced gas/mmol initial Ru).
d
Calculated as (mmol produced gas/mmol initial Ru) × h−1.
1
NMR and H NMR, showing again the signals corresponding
to 3 as shown in Figure 4.
Mechanistic Studies by High-Pressure HPNMR Tech-
niques. In order to identify the main species formed in situ
and the possible intermediates involved in the catalytic cycle,
NMR studies were carried out using HPNMR techniques.
formato-bridged dimeric species as described by other
authors.9,27
Complex [Ru(η6-C6H6)Cl(MBTS)2]Cl (2) was independ-
ently synthesized (vide infra) and tested as an isolated Ru(II)-
3008
dx.doi.org/10.1021/cs500655x | ACS Catal. 2014, 4, 3002−3012

ACS Catalysis
Research Article
1
77%), as shown in Figure 5a. In the corresponding H NMR
spectrum, two triplets at −10.48 ppm (2JHP= 36 Hz, ca. 77%)
and −10.88 ppm (2JHP= 36 Hz, ca. 23%) were observed,
turning into two singlets by 31P decoupling (Figure 5b,c).
Again, it is possible to compare these values with literature
data.33 It was reported that cis- and trans-[RuH-
(TPPTS)2(H2O)3]+ gave two 31P{1H} NMR peaks at 51 (cis)
and 55 ppm (trans) with corresponding triplets in the 1H NMR
spectrum at −9.7 (cis) and −9.6 ppm (trans), both with a 2JHP
=
38 Hz. Therefore, we attribute the signals observed to the
complexes cis- and trans-[RuH(MBTS)2(H2O)3]+(5). In
particular, the most abundant isomer (δP 52.4; δH −10.48)
resulted to be stable in solution for several hours both under
hydrogen pressure and upon pressure release.
1
Figure 4. 31P{1H} NMR (161.97 MHz, left) and H NMR (400.13
MHz, negative range, right) spectra of the monohydride [RuH(η6-
C6H6)(MBTS)2]+ (3).
Subsequently, H13COONa (3 equiv respect to Ru) was
1
added to the NMR tube containing 5. In 31P and H NMR
spectra, the signal due to 5 was still present as the major
species, accompanied by a series of minor signals in the range
δP 73.4−70.1 ppm. The corresponding 13C{1H} NMR
spectrum contained small signals at ca. 201 and 172 ppm.
The small broad peaks in the range δP 56.0−57.6 and the two
doublets at δC 172.96 (JHC = 208 Hz) and 172.80 (JHC = 205
Hz) can likely be attributed, in analogy with the TPPTS
system,16b to the formation of cis- and trans-[Ru(MBTS)2(η1-
O-O2CH)(H2O)3]+ (6) by ligand exchange from complex 4
still present in solution. By further heating to 70 °C, signals due
to complex 5 were still visible in the 31P{1H} NMR spectrum,
suggesting high stability of such species. The proposed
geometries of the main species based on NMR data are
shown in Chart 3.
An HPNMR experiment under quasi-catalytic conditions in
the presence of Ru(III)/L2 was then carried out. To an
aqueous solution containing RuCl3·3H2O/L2 (1:2) and
H13COONa (1 equiv), H13COOH (10 equiv) was added,
and the reaction was monitored by variable-temperature
HPNMR. After heating to 80 °C, complex 5, already observed
in the previous experiment, was formed in a minor amount
during the reaction. The main species detected in solution were
characterized by a broad peak at 43.4 ppm in 31P{1H} NMR, a
signal at −13.87 ppm in 1H NMR, and a series of signals in the
range 198.07−203.06 ppm in the 13C{1H} NMR spectrum. A
These mechanistic studies have been reported in detail for Ru/
m-TPPTS systems, and a reaction mechanism involving two
competing cycles has been proposed.16b By considering the
structural analogy with m-TPPTS and the higher purity of
ligand, phosphine L2 was chosen for these experiments with
both Ru(III) and Ru(II) precursors.
At first, a HPNMR experiment was carried out as follows. An
aqueous solution containing RuCl3·3H2O and L2 (1:2 ratio)
was prepared under nitrogen atmosphere and transferred into a
10 mL medium-pressure sapphire NMR tube. At room
temperature, the first 31P{1H} NMR spectrum showed the
peaks corresponding to the free ligand (−6.9 ppm), phosphine
oxide (33.5 ppm), and a new species characterized by a singlet
at 55.0 ppm. In analogy with data obtained under identical
conditions with m-TPPTS (54.0 ppm),30 we attribute this
signal to the formation of complex trans-[Ru-
(MBTS)2(H2O)4]2+ (4). In the TPPTS-based system, com-
31
32
plexes [Ru(Cl)(μ-Cl)(TPPTS)2]2 and RuCl2(TPPTS)3
were detected in solution at chemical shift values close to
57.0 ppm, albeit these species were obtained in slightly different
reaction conditions.
Subsequently, the NMR tube was pressurized with 100 bar of
H2 and heated to 70 °C. The 31P{1H} NMR pattern changed,
showing two broad singlets at 57.8 (ca. 23%) and 52.4 ppm (ca.
1
1
Figure 5. (a) Room temperature 31P{1H} HPNMR (161.97 MHz), (b) H HPNMR water suppression (400.13 MHz), and (c) H{31P} HPNMR
water suppression (400.13 MHz) spectra of cis- and trans-[RuH(MBTS)2(H2O)3]+ (5).
3009
dx.doi.org/10.1021/cs500655x | ACS Catal. 2014, 4, 3002−3012

ACS Catalysis
Research Article
1
at 4.94 ppm in the H NMR spectrum. After depressurization
and addition of H13COONa (3 equiv with respect to Ru), the
31P{1H} NMR pattern showed the disappearance of the signal
due to 2 and the presence of 3 as main species in solution.
Another small signal at 43.2 ppm was also present.
Interestingly, the same signal was observed as the only species
present in the 31P{1H} NMR spectra obtained by analyzing
aliquots of the solutions taken after recycling tests listed in
Chart 3. Proposed Geometries for Complexes 4−6
1
Table 8. The corresponding H NMR spectrum showed a
triplet at −15.12 ppm (2JHP = 20.01 Hz) and no signals due to
Ru-coordinated benzene. This suggests that under catalytic
conditions and recycling, the arene ligand is released, giving Ru-
hydrido species similar to those observed in the case of systems
using Ru(III) precursors.
CONCLUSIONS
■
In this study, a series of sulfonated biaryl monophosphines and
aryl diphosphines were used as efficient stabilizers for aqueous-
phase formic acid dehydrogenation to H2 + CO2 gas mixtures.
Monodentate aryl-biaryl phosphines gave comparable overall
TONs to m-TPPMS and m-TPPTS under the same
experimental conditions. The choice of Ru/L ratios (1:2 vs
1:4) did not affect significantly the catalytic performances. For
most ligands, after a first fast initial run (highest TOF), a
second run gave a decrease of reaction rate, related to the
reduction from Ru(III) to Ru(II) active species which once
formed gave again fast reactions (from third cycle). Good
recycling capacities of these catalytic systems were also assessed
(up to 11 consecutive recharges for L2). The aryl diphosphines
tested showed high stability of the catalytic system, reflected in
slightly slower initial reaction rates. Cycloalkyl groups on
phosphorus gave less stable systems which in turn resulted in
worse catalytic performance. Both Ru(III) salt and Ru(II) arene
complex were used as metal precursors and proved to be
efficient in the catalytic reaction studied. Mechanistic studies by
HPNMR experiments, although complicated by the presence of
many species formed under the conditions tested, highlighted
the formation of stable water-soluble Ru-hydrido species
mirroring previous studies carried out with aryl sulfonated
phosphine analogues.
set of similar experiments using TPPTS was reported by some
of us16b and showed the formation of complexes with
comparable chemical shift values, which were tentatively
attributed to [RuH(CO2)(TPPTS)2(H2O)2]+. In that system,
it was proposed that such a species is in equilibrium with the
more stable complex [RuH(HCO3)(TPPTS)2(H2O)2], whose
concentration increases after 40% conversion of initial
HCOOH and reaches a maximum at the end of the catalytic
run (basic pH). Although we do not have enough evidence for
a conclusive attribution for the NMR pattern observed under
these conditions, and we cannot in principle exclude the
formation of traces of Ru-carbonyl species described by other
authors,9,27 our proposal agrees with data reported for the
closely structurally related Ru/TPPTS system.
HPNMR experiments were also run in the presence of L2
and [Ru(η6-C6H6)Cl2]2 as metal precursor. In a first experi-
ment, [Ru(η6-C6H6)Cl2]2 and L2 (1:4) were dissolved in water,
and the solution was placed in an NMR sapphire tube. At room
temperature, the 31P{1H} NMR showed the formation of
complex [Ru(η6-C6H6)Cl(MBTS)2]Cl (2, in 1:5 ratio with
remaining free ligand), as confirmed by the presence of a singlet
at 22.3 ppm and comparison with the value observed from the
independent synthesis reported above. The tube was then
pressurized with H2 gas (100 bar) at room temperature, and a
new species was formed as evidenced by a broad signal at 52.7
EXPERIMENTAL SECTION
■
General Information. All manipulations were carried out
under an inert atmosphere of dry nitrogen by standard Schlenk
techniques unless otherwise noted. Doubly distilled water was
used after deoxygenation through a stream of nitrogen for
several hours. RuCl3·3H2O (Pressure Chemicals 99.5%,),
formic acid (Sigma-Aldrich, 98−100%) and sodium formate
(Sigma-Aldrich, 99%) were bought from commercial suppliers
and used without further purification. [Ru(η6-C6H6)Cl2]2 was
prepared as reported in the literature.35 Sulfonated mono-
phosphines20a,b and diphosphines were synthesized according
to literature procedures.21 HPNMR experiments were
performed in 10 mm medium-pressure sapphire tubes, and
NMR spectra were run on a Bruker Avance DRX 400 MHz
spectrometer. H2 (99.95%) and CO2 (99.9%) were acquired
from Carbagas-CH; enriched H13COOH and H13COONa,
(99% in 13C) were obtained from Cambridge Isotope
1
ppm in 31P{1H} NMR. The corresponding H NMR spectrum
displayed an intense broad signal at −8.56 ppm, a value
expected upon formation of ruthenium hydrido-clusters
[Ru4H4(η6-C6H6)4]2+ and [Ru4H6(η6-C6H6)4]2+ as described
earlier by Suss-Fink,26 together with a less intense triplet at
̈
−10.22 ppm (2JHP = 36 Hz). Upon increasing the temperature
to 70 °C, the former signal disappeared with a corresponding
increase of the latter, reaching a 1:8 ratio with 2. In the
corresponding 31P{1H} NMR spectrum, the signal at 52.7 ppm
was the major species observed. On the basis of NMR data for
m-TPPTS analogue,34 and confirmed by independent synthesis
(vide infra), these signals were attributed to the monohydride
cationic complex [RuH(η6-C6H6)(MBTS)2]+ (3). The pres-
ence of Ru-coordinated η6-benzene was confirmed by a singlet
1
Laboratories. The H, 13C{1H}, and 31P{1H} NMR spectra
were recorded on a Bruker Avance II 300 spectrometer
(operating at 300.13, 75.47, and 121.50 MHz, respectively) and
on a Bruker Avance II 400 spectrometer (operating at 400.13,
3010
dx.doi.org/10.1021/cs500655x | ACS Catal. 2014, 4, 3002−3012

ACS Catalysis
Research Article
100.61, and 161.98 MHz, respectively) at room temperature.
Peak positions are relative to tetramethylsilane and were
calibrated against the residual solvent resonance (1H) or the
deuterated solvent multiplet (13C). Electrospray mass spec-
trometry (ESI-MS) analysis was measured on a LCQ Orbitrap
mass spectrometer (ThermoFisher, San Jose, CA, U.S.A.)
equipped with a conventional ESI source by direct injection of
the sample solution and are reported in the form m/z (intensity
relative to base = 100). Infrared spectra of gas mixtures were
measured on a PerkinElmer FT-IR Spectrum BX II instrument
using a 10 cm gas phase cell equipped with KBr windows.
Formic Acid Catalytic Dehydrogenation Tests. All
catalytic tests were performed under an inert atmosphere of
nitrogen in a 25 mL glass jacketed reactor heated by external
liquid circulation. The three-necked round bottomed flask was
connected to a reflux condenser, which was in turn connected
to a setup of two manuals burets to measure the amount of gas
generated during the catalytic reaction. In a typical experiment,
the flask reactor was charged under a nitrogen atmosphere with
either RuCl3·3H2O (14.7 mg, 0.056 mmol) or [Ru(η6-
C6H6)Cl2]2 (14.0 mg, 0.028 mmol), the water-soluble
phosphine (0.112 or 0.224 mmol, depending on the Ru/ligand
ratio) and sodium formate (68 mg, 1 mmol). Degassed H2O
(2.5 mL) was then added by syringe, and the resulting solution
was heated to the desired temperature and kept stirring at ca.
750 rpm. Pure HCOOH (0.34 mL, 9.00 mmol) was then
added, and the course of dehydrogenation reaction was
monitored via gas evolution. Aliquots of gas mixtures were
collected for FTIR analyses in a 10 cm gas phase cell (KBr
windows). For recycling experiments, when no more gas
generation was observed, the apparatus was flushed with
nitrogen to remove any traces of H2 and CO2, then neat
HCOOH (0.34 mL, 9.00 mmol) was added by syringe, and a
new run was started. In the case of catalytic tests performed
with [Ru(η6-C6H6)Cl(MBTS)2]Cl (2), the procedures de-
scribed above were followed by maintaining the same Ru/
HCOOH ratio (in details: 2, 50 mg, 0.028 mmol; HCOOH,
0.17 mL, 4.51 mmol; HCOONa, 34.1 mg, 0.50 mmol). The
values of % conversion summarized in tables were calculated
from gas volume evolution as the average of three or four
measurements with a reproducibility of 10%.
syringe. The reaction was monitored by NMR taking 31P{1H},
1H, and 13C{1H} spectra at different temperatures.
Synthesis of [Ru(η6-C6H6)Cl(MBTS)2]Cl (2). The method
described by Tokunaga et al.36 for [Ru(η6-C6H6)Cl(TPPTS)2]
Cl was adapted for this synthesis. In a three-necked flask with
reflux condenser, under an inert atmosphere of nitrogen,
[Ru(η6-C6H6)Cl2]2 (30 mg, 0.06 mmol) and MBTS (L2, 154.7
mg, 0.24 mmol) were dissolved in degassed H2O (12 mL).
When the temperature reached 90 °C, both solids were
completely dissolved, and the reaction solution was stirred at
this temperature for 1 h. The solvent was then removed under
reduced pressure, yielding 120 mg of a crystalline brown
powder (83% pure on the basis of 31P{1H} NMR). 31P{1H}
NMR (161.97 MHz, D2O, 20 °C): δ (ppm) 23.2 (s). 1H NMR
(400.13 MHz, D2O, 20 °C): δ 8.31−7.43 (m, 32 H, Ar-H);
5.86 (s, 6H, C6H6). 13C{1H} NMR (75.47 MHz, D2O, 20 °C):
δ 143.33 (Ar-C), 143.04 (Ar-C), 142.02 (Ar-C), 141.88 (Ar-C),
136.74 (Ar-C), 136.34 (Ar-C), 134.45 (Ar-C), 131.52 (Ar-C),
130.63 (Ar-C), 128.48 (Ar-C), 128.16 (Ar-C), 127.56 (Ar-C),
127.18 (Ar-C), 126.07 (Ar-C), 96.88 (η6-C6H6). MS (nESI+):
m/z 1502.67 (100) [Ru(η6-C6H6)Cl(MBTS)2]+.
ASSOCIATED CONTENT
■
S
* Supporting Information
General methods and materials; additional tables and reaction
profiles for FA dehydrogenation reactions; and FT-IR spectra
of gas mixtures analyses. This material is available free of charge
via the Internet at http://pubs.acs.org
AUTHOR INFORMATION
■
Corresponding Authors
*E-mail: frederic.hapiot@univ-artois.fr.
*E-mail: gabor.laurenczy@epfl.ch.
*E-mail: l.gonsalvi@iccom.cnr.it.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Financial contributions by CNR and ECRF projects EFOR and
Firenze Hydrolab2 are gratefully acknowledged. Italian
Ministries MIUR and MATTM are thanked for supporting
this research through projects PRIN2009, Premiale 2011, and
PIRODE, respectively. This work was supported by COST
Action CM1205 CARISMA (Catalytic Routines for Small
Molecule Activation), also by granting a STSM of A.G. to
EPFL. L.G. thanks the CNR Short Term Mobility program for
covering the costs of a STM to EPFL. Swiss National Science
Foundation and EPFL are thanked for financial support.
HPNMR Mechanistic Experiments. Method A. A typical
experiment to explore the reactivity of Ru/L under hydrogen
pressure and further reaction with sodium formate is hereby
described. In a Schlenk tube, RuCl3·3H2O (38.7 mg, 0.148
mmol) or [Ru(η6-C6H6)Cl2]2 (37.0 mg, 0.074 mmol) and the
water-soluble phosphine L2 (190.8 mg, 0.296 mmol) were
dissolved in degassed water (2 mL) under a nitrogen
atmosphere. The resulting red solution was transferred to a
10 mm medium-pressure sapphire NMR tube and pressurized
with 100 bar of H2. HPNMR experiments were run at several
temperatures by heating the sample directly in the NMR probe.
Before the addition of H13COONa (30.6 mg, 0.444 mmol), the
solution was cooled to room temperature and depressurized,
REFERENCES
■
(1) Zuttel, A.; Borgschulte, A.; Schlapbach, L. Hydrogen as a Future
Energy Carrier; Wiley VCH: Weinheim, 2008.
1
then 31P{1H}, H, and 13C{1H} spectra were taken at different
(2) (a) Steele, B. C. H.; Henzel, A. Nature 2001, 414, 345−352.
(b) Sorensen, B. Hydrogen and Fuel Cells, Emerging Technologies and
Applications; Elsevier: Amsterdam, 2005.
(3) (a) Grochala, W.; Edwards, P. P. Chem. Rev. 2004, 104, 1283−
1315. (b) Graetz, J. Chem. Soc. Rev. 2009, 38, 73−82. (c) Hamilton, C.
W.; Baker, R. T.; Staubiz, A.; Manners, I. Chem. Soc. Rev. 2009, 38,
279−293. (d) Dalebrook, A. F.; Gan, W.; Grasemann, M.; Moret, S.;
Laurenczy, G. Chem. Commun. 2013, 49, 8735−8751.
́
(4) (a) Joo, F. ChemSusChem 2008, 1, 805−808. (b) Johnson, T. C.;
Morris, D. J.; Wills, M. Chem. Soc. Rev. 2010, 39, 81−88.
temperatures.
Method B. A typical experiment to test the behavior of Ru/L
systems under quasi-catalytic conditions of FA dehydrogen-
ation is hereby described. In a Schlenk tube, RuCl3·3H2O (67.9
mg, 0.26 mmol), L2 (335.1 mg, 0.52 mmol), and H13COONa
(20 mg, 0.29 mmol) were dissolved in H2O (2 mL). The
obtained solution was transferred to a 10 mm sapphire NMR
tube, and H13COOH (0.1 mL, 2.6 mmol) was added by
3011
dx.doi.org/10.1021/cs500655x | ACS Catal. 2014, 4, 3002−3012

ACS Catalysis
Research Article
(c) Grasemann, M.; Laurenczy, G. Energy Environ. Sci. 2012, 5, 8171−
8181.
(5) (a) Leitner, W. Angew. Chem., Int. Ed. Engl. 1995, 34, 2207−2221.
́
(b) Jessop, P. G.; Joo, F.; Tai, C. C. Coord. Chem. Rev. 2004, 248,
2425−2442. (c) Enthaler, S. ChemSusChem 2008, 1, 801−804.
(d) Fukuzumi, S. Eur. J. Inorg. Chem. 2008, 1351−1362. (e) Laurenczy,
G. Chimia 2011, 65, 663−666.
(6) (a) Gan, W.; Dyson, P. J.; Laurenczy, G. React. Kinet. Catal. Lett.
2009, 98, 205−213. (b) Enthaler, S.; von Langermann, J.; Schmidt, T.
Energy Environ. Sci. 2010, 3, 1207−1217.
and PO bond formation via hydroxy-complexes has been proposed
by different authors, see: (a) Kalck, P.; Monteil, F. Adv. Organomet.
Chem. 1992, 34, 219−284. (b) Larpent, C.; Dabard, R.; Patin, H. Inorg.
Chem. 1987, 26, 2992−2924.
(25) Junge, H.; Boddien, A.; Capitta, F.; Loges, B.; Noyes, J. R.;
Gladiali, S.; Beller, M. Tetrahedron Lett. 2009, 50, 1603−1606.
(26) (a) Suss-Fink, G.; Meister, A.; Meister, G. Coord. Chem. Rev.
̈
1995, 143, 97−111. (b) Suss-Fink, G. J. Organomet. Chem. 2014, 751,
̈
2−19.
(27) (a) Halpern, J.; Kemp, A. L. W. J. Am. Chem. Soc. 1966, 88,
5147−5150. (b) Czaun, M.; Goeppert, A.; Khotandaraman, J.; May, R.
B.; Haiges, R.; Prakash, S. G. K.; Olah, G. A. ACS Catal. 2014, 4, 311−
320.
(7) Coffey, R. S. Chem. Commun. 1967, 923−924.
(8) Bianchini, C.; Peruzzini, M.; Polo, A.; Vacca, A.; Zanobini, F.
Gazz. Chim. Ital. 1991, 121, 543−549.
(28) Tot
161.
́ ́
h, Z.; Joo, F.; Beck, M. T. Inorg. Chim. Acta 1980, 42, 153−
(9) (a) Gao, Y.; Kuncheria, J.; Yap, G. P. A.; Puddephatt, R. J. Chem.
Commun. 1998, 2365−2366. (b) Gao, Y.; Kuncheria, J.; Jenkins, H. A.;
Puddephatt, R. J.; Yap, G. P. A. J. Chem. Soc., Dalton Trans. 2000,
3212−3217. (c) Morris, D. J.; Clarkson, G. J.; Wills, M. Organo-
metallics 2009, 28, 4133−4140. (d) Majewski, A.; Morris, D. J.;
Kendall, K.; Wills, M. ChemSusChem 2010, 3, 431−434.
(10) (a) Foster, D.; Beck, G. R. J. Chem. Soc. D 1971, 1072−1072.
(b) Fukuzumi, S.; Kobayashi, T.; Suenobu, T. ChemSusChem 2008, 1,
827−834.
(29) Loges, B.; Boddien, A.; Junge, H.; Noyes, J. R.; Baumann, W.;
Beller, M. Chem. Commun. 2009, 4185−4187.
(30) Kovac
2336−2340.
́ ́ ́
s, J.; Joo, F.; Benyei, A.; Laurenczy, G. Dalton Trans. 2004,
(31) Hernandez, M.; Kalck, P. J. Mol. Catal. A: Chem. 1997, 116,
117−130.
(32) Parmar, D. U.; Bhatt, S. D.; Bajaj, H. C.; Jasra, R. V. J. Mol.
Catal. A: Chem. 2003, 202, 9−15.
(11) (a) Himeda, Y. Green Chem. 2009, 11, 2018−2022.
(b) Fukuzumi, S.; Kobayashi, T.; Suenobu, T. J. Am. Chem. Soc.
2010, 132, 1496−1497.
(33) Loy, M.; Laurenczy, G. Helv. Chim. Acta 2005, 88, 557−565.
́
́ ́
h, H.; Laurenczy, G.; Katho, A. J. Organomet. Chem.
(34) Horvat
2004, 689, 1036−1045.
(12) (a) Loges, B.; Boddien, A.; Junge, H.; Beller, M. Angew. Chem.,
Int. Ed. 2008, 47, 3962−3965. (b) Boddien, A.; Loges, B.; Junge, H.;
Beller, M. ChemSusChem 2008, 1, 751−758.
(35) (a) Zelonka, R. A.; Baird, M. C. Can. J. Chem. 1972, 50, 3063−
3072. (b) Bennett, M. A.; Smith, A. K. J. Chem. Soc., Dalton Trans.
1974, 233−241.
(36) Tokunaga, M.; Suzuki, T.; Koga, N.; Fukushima, T.; Horiuchi,
A.; Wakatsuki, Y. J. Am. Chem. Soc. 2001, 123, 11917−11924.
(13) (a) Boddien, A.; Loges, B.; Junge, H.; Gartner, F.; Noyes, J. R.;
̈
Beller, M. Adv. Synth. Catal. 2009, 351, 2517−2520. (b) Loges, B.;
Boddien, A.; Gartner, F.; Junge, H.; Beller, M. Top. Catal. 2010, 53,
̈
902−914. (c) Boddien, A.; Gartner, F.; Federsel, C.; Sponholz, P.;
̈
Mellmann, D.; Jackstell, R.; Junge, H.; Beller, M. Angew. Chem., Int. Ed.
2011, 50, 6411−6414.
(14) Boddien, A.; Mellmann, D.; Gartner, F.; Jackstell, R.; Junge, H.;
Dyson, P. J.; Laurenczy, G.; Ludwig, R.; Beller, M. Science 2011, 333,
1733−1736.
̈
(15) (a) Mellone, I.; Peruzzini, M.; Rosi, L.; Mellmann, D.; Junge, H.;
Beller, M.; Gonsalvi, L. Dalton Trans. 2013, 42, 2495−2501.
(b) Manca, G.; Mellone, I.; Bertini, F.; Peruzzini, M.; Rosi, L.;
Mellmann, D.; Junge, H.; Beller, M.; Ienco, A.; Gonsalvi, L.
Organometallics 2013, 32, 7053−7064.
(16) (a) Fellay, C.; Dyson, P. J.; Laurenczy, G. Angew. Chem., Int. Ed.
2008, 47, 3966−3968. (b) Fellay, C.; Yan, N.; Dyson, P. J.; Laurenczy,
G. Chem.Eur. J. 2009, 15, 3752−3760.
(17) Gan, W.; Fellay, C.; Dyson, P. J.; Laurenczy, G. J. Coord. Chem.
2010, 63, 2685−2694.
(18) Shaughnessy, K. H. Chem. Rev. 2009, 109, 643−710.
(19) (a) Bartik, T.; Bartik, B.; Hanson, B. E.; Glass, T.; Bebout, W.
Inorg. Chem. 1992, 31, 2667−2670. (b) Hida, S.; Roman, P. J., Jr.;
Bowden, A. A.; Atwood, J. D. J. Coord. Chem. 1998, 43, 345−348.
́ ́ ́ ́
(c) Joo, F.; Kovacs, J.; Katho, A.; Benyei, A. C.; Decuir, T.;
Darensbourg, D. J. Inorg. Synth. 1998, 32, 1−8.
(20) (a) Ferreira, M.; Bricout, H.; Hapiot, F.; Sayede, A.; Tilloy, S.;
Monflier, E. ChemSusChem 2008, 1, 631−636. (b) Ferreira, M.;
Bricout, H.; Azaroual, N.; Gaillard, C.; Landy, D.; Tilloy, S.; Monflier,
E. Adv. Synth. Catal. 2010, 352, 1193−1203. (c) Ferreira, M.; Bricout,
H.; Azaroual, N.; Landy, D.; Tilloy, S.; Hapiot, F.; Monflier, E. Adv.
Synth. Catal. 2012, 354, 1337−1346.
(21) (a) Bartik, T.; Bunn, B. B.; Bartik, B.; Hanson, B. E. Inorg. Chem.
1994, 33, 164−169. (b) Verspui, G.; Papadogianakis, G.; Sheldon, R.
A. Chem. Commun. 1998, 401−402. (c) Lucey, D. W.; Atwood, J. D.
Organometallics 2002, 21, 2481−2490.
(22) Genin, E.; Amengual, R.; Michelet, V.; Savignac, M.; Jutand, A.;
̂
Neuville, L.; Genet, J. P. Adv. Synth. Catal. 2004, 346, 1733−1741.
(23) Gan, W.; Snelders, D. J. M.; Dyson, P. J.; Laurenczy, G.
ChemCatChem 2013, 5, 1126−1132.
(24) Phosphine oxidation in water is ubiquitous even under rigorous
oxygen-free conditions. Redox chemistry involving metal reduction
3012
dx.doi.org/10.1021/cs500655x | ACS Catal. 2014, 4, 3002−3012