Formic acid dehydrogenation catalysed by ruthenium complexes bearing the tripodal ligands triphos and NP3

Article abstract of DOI:10.1039/c2dt32043f

The selective formic acid dehydrogenation to a mixture of CO₂ and H₂ was achieved with moderate to good productivities in the presence of homogeneous Ru catalysts bearing the polydentate tripodal ligands 1,1,1-tris-(diphenylphosphinomethyl)ethane (triphos) and tris-[2- (diphenylphosphino)ethyl]amine (NP₃), either made in situ from suitable Ru(iii) precursors or as molecular complexes. Preliminary mechanistic studies highlighting subtle differences due to ligand effects in the corresponding systems under study are also presented.

Full text of DOI:10.1039/c2dt32043f

Dalton
Transactions
View Article Online
View Journal | View Issue
PAPER
Formic acid dehydrogenation catalysed by ruthenium
complexes bearing the tripodal ligands triphos and
NP3†‡
Cite this: Dalton Trans., 2013, 42, 2495
Irene Mellone,a Maurizio Peruzzini,*a Luca Rosi,a,b Dörthe Mellmann,c Henrik Junge,c
Matthias Bellerc and Luca Gonsalvi*a
The selective formic acid dehydrogenation to a mixture of CO2 and H2 was achieved with moderate to
good productivities in the presence of homogeneous Ru catalysts bearing the polydentate tripodal
ligands 1,1,1-tris-(diphenylphosphinomethyl)ethane (triphos) and tris-[2-(diphenylphosphino)ethyl]-
amine (NP3), either made in situ from suitable Ru(III) precursors or as molecular complexes. Preliminary
mechanistic studies highlighting subtle differences due to ligand effects in the corresponding systems
under study are also presented.
Received 4th September 2012,
Accepted 14th November 2012
DOI: 10.1039/c2dt32043f
www.rsc.org/dalton
hydrogen storage density. Currently, none of the existing
methods, such as compressed gas containers, liquid hydrogen
Introduction
One of the most important scientific and technological chal- containers, or the use of interstitial hydrides and porous
lenges nowadays is finding new energy vectors, possibly materials as physical adsorbents for hydrogen, are entirely satis-
derived from renewable sources that can contribute to the factory options either in terms of costs or safety for widespread
worldwide demand for energy in a sustainable manner with use. Recently, the use of organic compounds as “chemical
advantageous costs. In this respect, hydrogen is a promising reservoirs”, from which hydrogen can be released on-demand
energy carrier and is considered as a clean alternative to fossil and safely under mild conditions, has emerged as a promising
fuels, especially if used as combustible for fuel cells. The alternative to other methods.3
development of efficient technologies for hydrogen generation
Among the different organic hydrogen storage molecules,
from renewable energy sources and for hydrogen storage in a in particular, formic acid (HCOOH) has recently received con-
safe and reversible manner is a pre-requisite for the utilization siderable attention, as it contains 4.4 wt% of hydrogen, is a
of hydrogen on a large scale in small- and middle-size mobile cheap liquid under ambient conditions and can be handled,
or stationary devices.1,2
stored and transported safely. Following a dehydrogenation
Compared to other fuels, hydrogen has a high gravimetric rather than a decarbonylation pathway, in the presence of suit-
energy density, but its volumetric energy density at atmos- able catalysts, HCOOH can be selectively decomposed to a
pheric pressure is too low for most envisioned applications. mixture of CO2 and H2 which, upon purification, can give fuel
Therefore, it is necessary to obtain an improved volumetric cell grade hydrogen. In principle, by a subtle choice of the cata-
lytic system, the reaction can be reverted and CO2 obtained as
a by-product can be re-hydrogenated to HCOOH. In this way
CO2 can be used as a renewable feedstock and a zero-carbon
demand and emission catalytic cycle can be invoked, avoiding
the use of fossil fuels for energy production.4
aConsiglio Nazionale delle Ricerche, Istituto di Chimica dei Composti
Organometallici (CNR-ICCOM), Via Madonna del Piano 10, 50019 Sesto Fiorentino,
Italy. E-mail: l.gonsalvi@iccom.cnr.it, maurizio.peruzzini@iccom.cnr.it;
Fax: +39 055 5225203; Tel: +39 055 5225251
Various combinations of transition metals and stabilizing
ligands have been tested to achieve efficient homogeneous cata-
lysts able to promote the selective dehydrogenation of
HCOOH. Among the different transition metals, such as Fe,
Co, Ir, Rh, ruthenium has been so far the most widely investi-
gated metal for this reaction, often in association with ancil-
lary ligands such as phosphines. It is worth mentioning the
results obtained by Wills,5 Puddephatt6 and co-workers
and later on by Laurenczy et al. who carried out the
bDipartimento di Chimica, Università degli Studi di Firenze, Via della Lastruccia,
3-13, 50019 Sesto Fiorentino, Italy
cLeibniz-Institute for Catalysis (LIKAT) at the University of Rostock, Albert-Einstein
Strasse 29a, 18059 Rostock, Germany
†Dedicated to Professor David Cole-Hamilton, an inspiring scientist, on the
occasion of his retirement and for his outstanding contribution to transition
metal catalysis.
‡Electronic supplementary information (ESI) available: Reaction profiles (gas
volume vs. time) for selected HCOOH dehydrogenation runs; VT-NMR spectra;
additional experimental details. See DOI: 10.1039/c2dt32043f
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 2495–2501 | 2495

View Article Online
Paper
Dalton Transactions
tris-[2-(diphenylphosphino)ethyl]amine (NP3) have been used
in the past for stabilisation of unusual transition metal moi-
eties and geometries as well as in catalytic applications.15
Although a variety of metal complexes with multidentate phos-
phine ligands have been shown to be efficient catalysts for
HCOOH dehydrogenation and CO2 hydrogenation, to the best
of our knowledge, neither triphos nor NP3 have been reported
so far in combination with Ru precursors for the former reac-
tion, whereas the use of an in situ catalytic system based on
Ru(acac)3 and triphos for the reduction of CO2 to methanol
under acidic conditions has been recently disclosed by Leitner
and co-workers.16 Thus, we decided to investigate the
behaviour of such complexes and ligands for HCOOH
Scheme 1
dehydrogenation of HCOOH–HCOONa (9 : 1) in water using Ru dehydrogenation.
sources together with the water soluble TPPTS ligand [TPPTS =
tris-(meta-sulfonatophenyl)phosphine], obtaining a TOF of
Catalytic formic acid dehydrogenation tests
460 h−1 working at 120 °C.7 In parallel, some of us showed We have tested five variations on the catalytic system for the
that the combination of Ru(η6-arene)Cl2 precursors and diphos- HCOOH dehydrogenation to CO2 + H2, namely (i) Ru(acac)3 +
phines such as dppm and dppe [dppm = bis(diphenylphos- triphos (1 : 1); (ii) Ru(acac)3 + NP3 (1 : 1); (iii) [Ru(κ3-triphos)-
phino)methane; dppe = 1,2-bis(diphenylphosphino)ethane] in (MeCN)3](OTf)2 (1); (iv) [Ru(κ4-NP3)Cl2] (2); [Ru(κ3-triphos)-
the presence of organic amines was able to give efficient (CO)(H)2] (3), running the tests under atmospheric pressure,
HCOOH dehydrogenation at 25 °C. TON values of 260 000 were measuring the development of gas with either an automatic17
obtained, corresponding to TOF > 900 h−1 8
. More recently, or a manual gas-burette. Aliquots of the gas mixtures produced
under optimized conditions the same catalyst system showed were analyzed off-line either by GC or FT-IR methods. The
turnover numbers up to 800 000 and a TOF of 48 000 h−1 at effect of the temperature on the productivity of the catalytic
80 °C, setting the benchmark test for Ru catalysts for this reac- systems was also investigated, running the tests both at 40 °C
tion so far.4h
and 80 °C, using either the azeotropic mixture HCOOH–NEt3
The beneficial stabilizing effect of tetradentate ligands such (5 : 2) or the combination HCOOH–OctNMe2 (11 : 10, OctNMe2
as PP3 (PP3
tris-[2-(diphenylphosphino)ethyl]phosphine) = n-octyldimethylamine), respectively.18
together with Fe(BF4)2·6H2O to promote the catalytic activation At 40 °C using the commercial azeotrope HCOOH–NEt3
=
of HCOOH, originally reported in 1991,9 was recently reinvesti- (5 : 2), the reaction proceeds with moderate TOFs using both
gated, showing an outstanding performance with TON = 92 000 molecular complexes 1 and 2 (see Table 1). GC analysis of the
and high TOF = 9425 h−1 at 80 °C in propylene carbonate as a gas mixtures showed the presence of a 1 : 1 mixture of CO2 +
solvent.10 Interestingly, the same ligand was used together with H2, hence complete selectivity for HCOOH dehydrogenation
Co(BF4)2·6H2O for the reduction of CO2 and bicarbonate with vs. decarbonylation is obtained as no signal corresponding to
hydrogen to formates at 60–80 °C under a total pressure CO2 + CO was detected, implying a quantity lower than 10 ppm
H2 = 60–100 bar, giving a maximum 94% yield after 20 h.11
which is the sensitivity limit for this technique. For a
Here we report on the dehydrogenation of formic acid/ maximum expected gas volume of ca. 2.91 L, under these con-
amines in the presence of the facially capping 1,1,1-tris-(diphe- ditions 1 produced ca. 454 mL after 18 h, whereas 2 gave ca.
nylphosphinomethyl)ethane (triphos) and related tetradentate 564 mL after 45 h, corresponding to final TONs of 487 and
tris[2-(diphenylphosphino)ethyl]amine (NP3), either used in 604, respectively. Disappointingly, using the in situ generated
situ together with Ru(acac)3 as a ruthenium source or as catalysts prepared according to methods (i)–(ii), the reactions
ligands for known complexes such as [Ru(κ3-triphos)(MeCN)3]- did not show any significant activity under the conditions
(OTf)2 (1),12 [Ru(κ4-NP3)Cl2] (2)13 and cis-[Ru(κ3-triphos)(CO)- described in Table 1. This suggests that under such mild con-
(H)2] (3,14 see Scheme 1). The catalytic tests are complemented ditions the metal cannot be efficiently reduced from Ru(III) to
by 31P{1H} NMR studies based on stoichiometric reactions, the active Ru(II) form and/or be stabilized by the ancillary
which gave information on the nature of some of the species ligand to form catalytically active species.
which are formed upon reaction with formate and formic acid/
Improved catalytic performances were observed running the
amine adduct, respectively, under different experimental HCOOH dehydrogenation tests at 80 °C, in the presence of
conditions.
n-octyldimethylamine (Table 2), using HCOOH : Ru = 1000 and
HCOOH : Ru = 10 000 ratios. In all cases the gas volume
expected for complete conversion (630 mL) was obtained after
different reaction times depending on the catalytic system
used.
Results and discussion
The facially capping triphosphine 1,1,1-tris-(diphenylphosphi-
From Table 2, it can be concluded that the best catalytic
nomethyl)ethane (triphos) and the tripodal aminotrisphosphine system under the conditions applied was complex 1, as would
2496 | Dalton Trans., 2013, 42, 2495–2501
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
Table 1 Formic acid dehydrogenation catalysed by complexes 1 and 2 in the presence of NEt3 at 40 °C
Entry
Catalyst
Vol (mL) after 180 min
TON after 180 min
Max vol (mL) [t(h)]
Final TON
Final conversion (%)
1
2
1
2
65
20
69
21
454 [18]
564 [45]
487
604
16
19
Reaction conditions: catalyst, 19.1 μmol; HCOOH–NEt3 (5 : 2), 5 mL; HCOOH : catalyst = 3117; gas measured via automatic burette and analyzed by
GC (H2 : CO2 = 1 : 1).
Table 2 Formic acid dehydrogenation catalysed by Ru/Ln (Ln = tripodal ligand:
triphos, NP3) and 1–3 in the presence of OctNMe2 at 80 °C
Table 3 Recycling tests for formic acid dehydrogenation catalysed by 1 in the
presence of OctNMe2 at 80 °C
Max vol
(mL)
[t(min)]
Vol (mL) after
Cycle 60 min
TOF (h−1) after Max vol (mL)
Final TOF
(h−1
Vol (mL) after TON after
Final
TON
60 min
[t(min)]
)
Entry Catalyst 60 min
60 min
I
II
630
600
430
350
340
330
300
240
1000
944
676
550
535
519
472
377
630 [60]
630 [65]
630 [90]
630 [110]
630 [120]
630 [130]
630 [140]
630 [150]
1000
923
666
545
500
461
428
400
1
2
3
4
5
6
7
ia
210
70
630
100
130
60
314
105
1000
1504
204
902
373
630 [180] 1000
630 [490] 1000
iia
1b
1c
2b
2c
3b
III
IV
V
VI
VII
VIII
630 [60]
1000
630 [320] 10 000
630 [230] 1000
630 [420] 10 000
630 [220] 1000
250
Reaction conditions: a Ru(acac)3, 12.9 μmol; ligand, 12.9 μmol; HCOOH– Reaction conditions: complex 1, 12.9 μmol; HCOOH–OctNMe2 (11 : 10),
OctNMe2 (11 : 10), 2.73 mL (density at 25 °C = 0.894 mg mL−1); 2.73 mL (density at 25 °C = 0.894 mg mL−1); HCOOH : catalyst = 1000;
HCOOH : catalyst = 1000; 80 °C. Gas measured via a manual gas 80 °C; from run 2, 12.9 mmol of neat HCOOH were added by
burette and analyzed via FT-IR spectroscopy. b As above, preformed microsyringe for each run. Gas measured via a manual gas burette and
catalyst 12.9 μmol. c As above, preformed catalyst 1.29 μmol.
analyzed via FT-IR spectroscopy.
and 3 evolve toward the same (or at least very similar) catalyti-
cally active species, perhaps [Ru(κ3-triphos)(H)(η2-H2)]+ upon
loss of CO and protonation of one hydrido ligand. This species
has already been proposed in the literature as an active catalyst
be expected for a Ru(II) species bearing a strongly coordinating
tripodal polyphosphine and three labile solvent ligands, able
in principle to make up to three coordination sites available
for substrate coordination and activation. In contrast to the
results reported by Leitner and co-workers for the Ru(triphos)-
catalysed CO2 hydrogenation to methanol under acidic con-
ditions, which were however run at higher temperature
(140 °C),16 the system based on in situ formation of the puta-
tive active catalyst (run 1) did not perform as efficiently as
complex 1. Complete conversion of HCOOH was reached only
after 3 h instead of 1 h as for the molecular Ru-triphos
complex 1. In the case of the NP3-based system, both methods
gave more sluggish conversions than triphos (Table 2, runs 2
and 5; see also reaction profiles in Fig. S1–2, ESI‡). Both com-
plexes 1 and 2 were active at higher substrate to catalyst ratios,
reaching final TONs of 10 000 after 320 and 420 min,
respectively.
for hydrogenation reactions.19 However,
3 was neither
observed to be formed during the HCOOH dehydrogenation
tests when complex 1 was used as a pre-catalyst, nor during
NMR mechanistic studies starting from an analogue of 1 as a
ruthenium source (vide infra).
Recycling experiments using the best catalyst found in this
study, i.e. complex 1, were also carried out. After evolution of
the gas volume expected in the first run (630 mL), neat
HCOOH (1000 equiv. to Ru originally used) was added to the
thermostated reaction vessel kept at 80 °C and the next run
was started. In this way, it was possible to recycle the mixture
containing the catalyst and the amine for up to eight consecu-
tive runs (Table 3, Fig. S5‡ for reaction profile). As can be seen
from Table 3, complete conversion of the substrate was invari-
ably reached after each recycle, albeit with a decrease of TOFs
already noticeable after 1 h.
Interestingly, a few nicely shaped crystals formed at the end
of a catalytic experiment using Ru(acac)3 and triphos. 31P{1H}
and 1H NMR spectra of the crystals dissolved in CDCl3 showed
unequivocally that the complex formed under these conditions
was the known [Ru(κ3-triphos)(CO)(H)2] (3).14 Thus, we
thought it could be of interest to synthesise 3 (see ESI‡ for
VT-NMR mechanistic studies on model reactions
details) and test it as the preformed catalyst for HCOOH dehy- In order to gain a better insight into the mechanism of sub-
drogenation under the same conditions. Indeed, also this strate activation and pre-catalyst modifications during the cata-
complex was endowed with catalytic activity that closely lytic cycle, we decided to test the reactivity of complexes 1 and
resembles that observed for an in situ generated Ru-triphos 2 with formate and formic acid/amine. At first, the behaviour
system (cf. runs 1 and 7, Table 2, and Fig. S3 in ESI‡). This of complex 1 was studied. Prior to the mechanistic studies,
observation may in turn suggest that either 3 is formed under anion metathesis on 1 was carried out to exchange the
the conditions applied in run 1 or that both Ru(acac)3/triphos potentially coordinating triflate with hexafluorophosphate
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 2495–2501 | 2497

View Article Online
Paper
Dalton Transactions
concentration at 293 K. The situation at 313 K showed that the
pattern due to 4 was still present, together with a broad hump
at ca. 41.0 ppm. This behaviour again suggests a rapid formate
exchange at higher temperatures.
13C{1H} NMR spectra in acetone-d6 of a mixture of 1′ with 2
equiv. of HCOONHEt3 and 1 equiv. of HCOOH, taken at 233 K,
showed a small singlet at 178.7 ppm, at values expected for
coordinated formate.20a Finally, FT-IR spectra of the final
solution at room temperature showed νCO formate stretching
modes at 1697 and 1546 cm−1, in the range expected for η1-O
(O)CH and η2-O(O)CH coordination, respectively.20,22 These
spectroscopic investigations, although preliminary, suggest the
presence of Ru-coordinated formate species equilibrating
between η2- to η1-fashion in this system. Attempts to isolate 4
by carrying out the reactions at preparative scale have unfortu-
nately failed so far due to decomposition of the complex
during workup.
Scheme 2
(Experimental section), obtaining the new complex [Ru-
(κ3-triphos)(MeCN)3](PF6)2 (1′).
During all experiments, no 31P{1H} signals at negative
chemical shift values attributable to triphos partial de-chela-
tion (i.e. κ3- to κ2-coordination switch), proposed by some
authors to occur during catalytic reactions,23 were visible
throughout the experiments. Moreover, complex 3 was never
observed to be formed during the VT NMR experiments. We
could safely rule out the presence of Ru carbonyl species such
as [Ru(κ3-triphos)(MeCN)x(CO)y]2+ (x = 0–2; y = 1–3; x + y = 3)
by comparison with NMR spectra obtained by saturating an
acetone-d6 solution of 1′ with CO gas at 1 atmosphere (see ESI‡).
Although at this stage the experiments described above do
not allow a full interpretation of the mechanism responsible
for the catalytic dehydrogenation of HCOOH promoted by 1,
on the basis of all the evidence so far obtained it can be con-
cluded that for these systems triphos behaves as a κ3-P facial
ligand, and that any metal-hydrido species which may form
are quickly converted on an NMR timescale in the temperature
range 233–313 K using 1′ as a starting complex.
A series of variable temperature (VT) NMR experiments
involving 1′ was carried out as follows. In the first experiment,
2.58 mg of HCOONHEt3 (0.1 mL of a solution of 25.8 mg of
reagent dissolved in 1.0 mL of acetone-d6), freshly prepared by
the reaction of HCOOH and NEt3 in a 1 : 1 ratio, were added
under nitrogen to a solution of 1′ (20 mg, 1.76 × 10−2 mmol) in
0.6 mL of acetone-d6 at 233 K. The first 31P{1H} NMR spectrum
recorded at this temperature immediately after mixing the
reagents showed that the original singlet due to 1′ at 27.2 ppm
was now accompanied (ca. 1 : 1 ratio) by a new AM2 system at
42.7 ppm (d, 2JPP = 42 Hz) and 29.5 ppm (t), which we assigned
to the diformato complex [Ru(κ3-triphos)(η1-OOCH)(η2-OOCH)]
(4) (Scheme 2). In keeping with formate coordination, the cor-
responding 1H NMR spectrum showed a broad resonance at
8.34 ppm, at chemical shift values expected for Ru-coordinated
formate.20 The appearance of a new singlet at 2.18 ppm con-
firms that MeCN has been partially released in solution.
Complex 4 may arise from stepwise ligand substitution from
putative species such as [Ru(κ3-triphos)(MeCN)(η2-OOCH)] (4′)
depending on the Ru/formate stoichiometry.21
In keeping with this hypothesis, addition of 1 equiv. of neat
HCOOH at 233 K caused the complete disappearance of the
signal due to 1′, while maintaining the pattern due to 4. The
VT NMR experiment was continued by raising gradually the
temperature up to 313 K, and a broadening of the signals cor-
responding to 4 was observed, to eventually collapse at 313 K.
A close inspection of the corresponding 1H NMR spectra
showed that no hydride species were formed throughout the
heating sequence.
A VT-NMR study was also carried out starting from complex
2. By dissolving 2 in CD2Cl2, a first 31P{1H} NMR spectrum
showed the expected AM2 spin system with the PA triplet at
45.7 ppm (2JPP = 30 Hz) and the PM doublet at 26.6 ppm. To
this solution, HCOONHEt3 (1 equiv.) was added under nitro-
gen at 233 K. The corresponding 31P{1H} NMR spectrum
showed the appearance of a new set of signals at 47.8 ppm
2
(t, JPP = 29 Hz, PA) and 25.8 ppm (d, PM). The corresponding
1H NMR spectrum showed a new singlet at 8.41 ppm which
can be attributed to coordinated formate. Based on chemical
shift and coupling constant values, we propose for the new
species the tentative formula [Ru(κ4-NP3)Cl(η1-OOCH)] (5).
This species reached a 1 : 1 ratio with 2 at 253 K. At 273 K a
new AM2 system, showing a doublet at 42.4 ppm (2JPP = 18 Hz)
and a triplet at 29.3 ppm, slowly increased at the expense of 5.
This species was identified by comparison with literature
data24 as the hydrochloride complex [Ru(κ4-NP3)Cl(H)] (6), as
confirmed by the presence of the Ru–H resonance at
−6.56 ppm (dt, JHP(trans) = 97 Hz; JHP(cis) = 26 Hz, 1H, RuH) in
the corresponding 1H NMR spectrum. At 283 K the ratio
between 2, 5 and 6 reached 7 : 3 : 1, while at 273 K a further
In a second experiment, 1′ was mixed with 2 equiv. of
HCOONHEt3 at 233 K before starting the VT NMR sequence.
The initial 31P{1H} NMR spectrum measured upon mixing of
the reagents gave the same picture observed for experiment 1,
except that 1′ was much less abundant (ca. 8% left), support-
ing the hypothesis discussed above. The addition of HCOOH
caused the complete disappearance of 1′ accompanied by
release of MeCN in solution as observed in the 1H NMR
spectra. By increasing the temperature, at 273 K two singlets at
42.3 ppm and 42.0 ppm appeared, reaching a maximum
2498 | Dalton Trans., 2013, 42, 2495–2501
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
decrease in the concentration of 5 was accompanied by an ratio to the complexes, and the observations previously
increase of 6 (ratio 2 : 5 : 6 = 3 : 1.5 : 1). At 273 K a new octa- described were consistent with these experiments.
hedral species characterized by a doublet at 56.4 ppm (d, 2JPP
=
In order to assign the signals due to 8, a close inspection of
16 Hz, 2P, PM) and a triplet at 51.6 ppm (PA) slowly increased. the VT NMR experiment sequence (Fig. S6 and S7 in ESI‡)
A broad singlet at −9.99 ppm appeared in the corresponding suggested that species such as 2 and 5, having Cl− trans to
1H NMR spectrum. This new species was identified as the apical P, were characterized by 31P NMR patterns bearing the
highly fluxional complex [Ru(κ4-NP3)(η2-H2)(H)]+ (7) by com- triplet at lower (i.e. more deshielded) and the doublet at
parison with previous data obtained in our laboratories.25 At higher fields, whereas this order reverts in the case of 6 and 7.
this point, a second equivalent of HCOONHEt3 was added. As As the former pattern (deshielded triplet) was observed also
expected, this caused an increase in the concentration of 5 for 8, we reasoned that this species could be structurally
and a decrease of 2. Raising the temperature to 293 K a further related to 2 and 5. For this reason, compound 6 was prepared
increase of 6 and 7 was observed, together with a decrease of 2 and isolated as described in the literature,24 and reacted at
and 5. The probe was then cooled to 233 K and 1 equiv. of 233 K with HCOOH on an NMR tube scale using CD2Cl2 as a
1
HCOOH was added. This in turn caused the complete dis- solvent. The corresponding 31P{1H} NMR and H NMR spectra
appearance of 2 and 5, leaving 7 and a new AM2 set of signals showed unequivocally the formation of 8, to which we assign
with a PA triplet at 50.0 ppm (2JPP = 33 Hz) and a PM doublet at the formula [Ru(κ4-NP3)Cl(η2-H2)]+, as a major product. Inter-
25.5 ppm. At 293 K the latter (8) was the main species present estingly, by leaving the NMR tube at room temperature and
in the 31P{1H} NMR spectrum, with only a small residue of 7. collecting another set of spectra after 1 h at 293 K, formation
The broad singlet at −9.99 ppm in the 1H NMR spectrum, of 2 was observed to form at the expense of 8. From these
typical of Ru(η2-H2) moiety, increased in intensity.
experiments it can be proposed that the stable complex 6 may
The 31P{1H} and H NMR variable temperature experiments follow two different activation pathways, shown in Scheme 3.
were repeated reacting 1′ and 2 with the azeotrope HCOOH– In path A, chloride exchange with formate in 6 leads to an
NEt3 (5 : 2) which was used for the catalytic tests, using a 5 : 1 (unobserved) hydrido-formate species [Ru(κ4-NP3)H(η1-OOCH)]
which rapidly evolves upon protonation to 7, which then regen-
1
erates 6 by hydrogen release and chloride coordination, in line
with the mechanism reported for the Fe/PP3 system.10 In path
B, protonation of 6 leads to 8, which then gives back 2 releas-
ing hydrogen. The coexistence of these two pathways may have
an implication in the lower activity of 2 compared to 1
observed in formic acid dehydrogenation tests.
Conclusions
The activity of Ru(II) complexes containing the tripodal phos-
phine triphos and the amino(tris)phosphine NP3 in the selec-
tive catalytic dehydrogenation of formic acid to carbon dioxide
and hydrogen has been assessed under different conditions.
The complex [Ru(κ3-triphos)(MeCN)3](OTf)2 (1) showed
superior performances with a TON of 10 000 after 6 h using
0.01 mol% of the catalyst and allowed for recycling up to eight
times (0.1 mol% catalyst) giving a total TON of 8000 after ca.
14 h of continuous reaction at 80 °C in the presence of
OctNMe2. Preliminary mechanistic NMR studies using Ru-
(κ3-triphos)(MeCN)3](PF6)2 (1′) and [Ru(κ4-NP3)Cl2] (2) are pre-
sented demonstrating that the NP3 ligand helps to stabilise
Ru-hydrido species. This kind of intermediate must be short-
lived in the triphos-based system. Further work is currently
being carried out to establish a rationale for such behaviour,
also by means of DFT theoretical calculations.
Experimental
General methods and materials
Complexes 1 and 2 were prepared according to literature pro-
Scheme 3
cedures.12,13 Formic acid and N,N-dimethyloctylamine were
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 2495–2501 | 2499

View Article Online
Paper
Dalton Transactions
distilled in vacuum prior to use and stored under nitrogen. 0.38 g) in warm EtOH (20 mL) was added. After 15 min stirring
The formic acid to amine ratio was determined by 1H NMR at room temperature, the mixture was concentrated under
spectroscopy on a Bruker Avance 400 spectrometer. Ru(acac)3 vacuum and cold EtOH was added to induce precipitation.
and the HCOOH–NEt3 (5 : 2) mixture were purchased from
The desired complex [Ru(κ3-triphos)(MeCN)3](PF6)2 (1′) was
commercial suppliers and used without further purification. obtained in almost quantitative yield, recrystallised from
Deuterated solvents for NMR measurements were purchased MeCN–EtOH (1 : 1) giving a white microcrystalline product
from commercial suppliers and stored onto activated 4A mo- (1.06 g, 0.93 mmol).
lecular sieves under N2 before use. The variable temperature
1H and 31P{1H} NMR spectra were recorded on a Bruker
Avance 400 spectrometer (operating at 400.13 and 100.61 MHz,
respectively). Peak positions are relative to tetramethylsilane
Acknowledgements
and were calibrated against the residual solvent resonance (1H) This work was financially supported by CNR, MATTM and
or the deuterated solvent multiplet (13C). 31P{1H} NMR were MIUR through projects EFOR, PIRODE and PRIN 2009 which
referenced to 85% H3PO4, with the downfield shift taken as are gratefully acknowledged. Thanks are also expressed to
positive.
ECRF through the project Firenze Hydrolab-2. Regione
Toscana is thanked for granting mobility funding. Additional
support by the BMBF and the state of Mecklenburg-Vorpom-
Catalytic tests
All catalytic experiments were carried out under an inert argon mern is gratefully acknowledged. G. Manca, A. Rossi and
or nitrogen atmosphere in a flask heated either by oil bath or A. Rossin (ICCOM-CNR) are thanked for help and useful
by external heating liquid circulation. The flask was connected discussions.
to a reflux condenser, which was connected either to a setup
of two manual burettes or to an automatic burette to measure
the amounts of gas generated. The mixture of gas produced
during the catalytic runs was either analyzed by FTIR-ATR Shi-
Notes and references
madzu IRAffinity-1 with a gas cell (path length 10 cm) or by
GC using a gas chromatograph HP 6890N (permanent gases:
Carboxen 1000, TCD, external calibration; amines: HP Plot Q,
30 m, FID) and a hydrogen sensor (Hach Ultra Analytics GmbH).
Typical procedure for the HCOOH dehydrogenation tests: A
solution of HCOOH–amine (2.5 mL) was pre-warmed to 80 °C
(HCOOH–OctNMe2 11 : 10) in a three-necked round bottomed
flask with a thermostated oil bath. The flask was purged with
argon or nitrogen before the starting reaction by addition of
the solid catalyst. In the case of experiments run at 40 °C
(using the mixture HCOOH–NEt3 5 : 2), the automatic burette
setup previously described was used.17
Typical procedure for recycling experiments at 80 °C: As
soon as the gas volume evolution reached the theoretical value
expected for 100% HCOOH conversion, the apparatus was
flushed with nitrogen to remove any traces of hydrogen and
carbon dioxide. The remaining reaction mixture containing
the catalyst and the amine, kept at 80 °C, was reutilized as
such for the following run by addition of neat HCOOH
(0.49 mL, 12.9 mmol) through a syringe.
1 (a) N. Armaroli and V. Balzani, Angew. Chem., Int. Ed., 2007,
46, 52; (b) W. Lubitz and B. Tumas, Chem. Rev., 2007, 107,
3900.
2 (a) U. Eberle, M. Felderhoff and F. Schüth, Angew. Chem.,
Int. Ed., 2009, 48, 6608; (b) M. K. Thomas, Catal. Today,
2007, 120, 389.
3 (a) A. W. C. van den Berg and C. O. Areán, Chem. Commun.,
2008, 668; (b) L. J. Murray, M. Dinca and J. R. Long, Chem.
Soc. Rev., 2009, 38, 1294; (c) J. Graetz, Chem. Soc. Rev., 2009,
38, 73; (d) C. W. Hamilton, R. T. Baker, A. Staubiz and
I. Manners, Chem. Soc. Rev., 2009, 38, 279.
4 (a) P. G. Jessop, T. Ikariya and R. Noyori, Chem. Rev., 1995,
95, 261; (b) P. G. Jessop, F. Joó and C. C. Tai, Coord. Chem.
Rev., 2004, 248, 2425; (c) P. G. Jessop, in The Handbook of
Homogeneous Hydrogenation, ed. J. G. de Vries and
C. J. Elsevier, Wiley-VCH, Weinheim, 2007, 489;
(d) W. Leitner, Angew. Chem., Int. Ed., 1995, 34, 2207;
(e) Y. Himeda, Eur. J. Inorg. Chem., 2007, 3927; (f) J. F. Hull,
Y. Himeda, W.-H. Wang, B. Hashiguchi, R. Periana,
D. J. Szalda, J. T. Muckerman and E. Fujita, Nat. Chem.,
2012, 4, 383; (g) G. Papp, J. Csorba, G. Laurenczy and
F. Joo, Angew. Chem., Int. Ed., 2011, 50, 10433;
(h) Y. Himeda, S. Miyazawa and T. Hirose, ChemSusChem,
2011, 4, 487; (i) A. Boddien, C. Federsel, P. Sponholz,
D. Mellmann, R. Jackstell, H. Junge, G. Laurenczy and
M. Beller, Energy Environ. Sci., 2012, 5, 8907.
Variable temperature NMR experiments
Typically, the NMR tube samples were prepared at 233 K
(acetone–liquid nitrogen bath) under nitrogen by dissolving
20 mg of the Ru complex in 0.70 mL of deuterated solvent and
adding either 1 to 2 equiv. of HCO2NHEt3, followed by 1 equiv.
of neat HCOOH, or 5 equiv. of HCOOH–NEt3 (5 : 2). The NMR
spectra were collected in the temperature range 233–313 K in
the case of 2 (CD2Cl2) or 233–323 K in the case of 1 (acetone-d6).
Synthesis of [Ru(κ3-triphos)(MeCN)3](PF6)2 (1′). Complex 1
(1.0 mmol, 1.15 g) was dissolved under nitrogen in warm
MeCN (20 mL), to which a solution of NH4PF6 (2.3 mmol,
5 D. J. Morris, C. J. Clarkson and M. Wills, Organometallics,
2009, 28, 4133.
6 (a) Y. Gao, J. Kuncheria, G. P. A. Yap and R. J. Puddephat,
Chem. Commun., 1998, 2365; (b) Y. Gao, J. K. Kuncheria,
H. A. Jenkins, R. J. Puddephatt and G. P. A. Yap, J. Chem.
Soc., Dalton Trans., 2000, 4703.
2500 | Dalton Trans., 2013, 42, 2495–2501
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
7 (a) C. Fellay, P. J. Dyson and G. Laurenczy, Angew. Chem., 16 S. Wesselbaum, T. vom Stein, J. Klankermayer and
Int. Ed., 2008, 47, 1; (b) C. Fellay, N. Yan, P. J. Dyson and
G. Laurenczy, Chem.–Eur. J., 2009, 15, 3752.
8 (a) B. Loges, A. Boddien, H. Junge and M. Beller, Angew.
W. Leitner, Angew. Chem., Int. Ed., 2012, 51, 7499.
17 A. Boddien, B. Loges, H. Junge and M. Beller, Chem-
SusChem, 2008, 1, 751.
Chem., Int. Ed., 2008, 47, 3962; (b) A. Boddien, B. Loges, 18 (a) H. Junge, A. Boddien, F. Capitta, B. Loges, J. R. Noyes,
H. Junge, F. Gärtner, J. R. Noyes and M. Beller, Adv. Synth.
Catal., 2009, 351, 2517.
9 (a) C. Bianchini, M. Peruzzini, A. Polo, A. Vacca and
S. Gladiali and M. Beller, Tetrahedron Lett., 2009, 50, 1603;
(b) A. Majevski, D. J. Morris, K. Kendall and M. Wills,
ChemSusChem, 2010, 3, 431.
F. Zanobini, Gazz. Chim. Ital., 1991, 121, 543; 19 F. M. A. Geilen, B. Engendahl, M. Hölscher,
(b) C. Bianchini and M. Peruzzini, in Recent Advances in
Hydride Chemistry, ed. M. Peruzzini and R. Poli, Elsevier
BV, Amsterdam,NL, 2001. ch. 9, p. 286.
J. Klankermayer and W. Leitner, J. Am. Chem. Soc., 2011,
133, 14349.
20 (a) M. K. Whittlesey, R. N. Perutz and M. H. Moore, Organo-
metallics, 1996, 15, 5166; (b) O. R. Allen, S. J. Dalgarno,
L. D. Field, P. Jensen and A. C. Willis, Organometallics,
2009, 28, 2385.
10 A. Boddien, D. Mellmann, F. Gärtner, R. Jackstell, H. Junge,
P. J. Dyson, G. Laurenczy, R. Ludwig and M. Beller, Science,
2011, 333, 1733.
11 C. Federsel, C. Ziebart, R. Jackstell, W. Baumann and 21 Preliminary DFT calculations at B3LYP/631+G(d,p) level
M. Beller, Chem.–Eur. J., 2012, 18, 72.
12 L. F. Rhodes, C. Sorato, L. M. Venanzi and F. Bachechi,
Inorg. Chem., 1988, 27, 604.
13 L. Dahlenburg, K. M. Frosin, S. Kestan and D. Werner,
J. Organomet. Chem., 1991, 407, 115.
14 V. I. Bakhmutov, E. V. Bakhmutov, N. V. Belkova,
C. Bianchini, L. M. Epstein, D. Masi, M. Peruzzini,
E. S. Shubina, E. V. Vorontsov and F. Zanobini,
Can. J. Chem., 2001, 79, 479.
with SDD pseudo-potential on an Ru atom on a simplified
model (H instead of Ph on phenyl rings), including the
solvent (acetone, CPCM model), showed that the reaction
of [Ru(κ3-triphos)(MeCN)3]2+ with HCO2 to give 4′ is
−
favoured with a ΔG298 = −9.74 kcal mol−1 and ΔG233
=
−8.08 kcal mol−1, respectively. Further stabilisation (ΔG298
= −5.24 kcal mol−1 and ΔG233 = −5.42 kcal mol−1) is gained
by reaction of 4′ with HCO2− to give 4. The full DFT study is
in progress and will be reported in due time.
15 (a) W. O. Siegl, S. J. Lapporte and J. P. Collman, Inorg. 22 (a) I. S. Kolomnikov, A. I. Gusev, G. G. Aleksandrov,
Chem., 1973, 12, 674; (b) F. A. Cotton and B. Hong, Prog.
Inorg. Chem., 1992, 40, 179; (c) H. A. Mayer and
T. S. Lobeeva, Yu. T. Struchkov and M. F. Vol’pin, J. Organo-
met. Chem., 1973, 59, 359.
W. C. Kaska, Chem. Rev., 1994, 94, 1239; (d) C. Bianchini, 23 (a) B. J. Sarmah and D. K. Dutta, J. Organomet. Chem., 2010,
A. Meli, M. Peruzzini, F. Vizza and F. Zanobini, Coord.
Chem. Rev., 1992, 120, 193; (e) H. T. Teunissen and
695, 781; (b) A. B. Chaplin and P. J. Dyson, Eur. J. Inorg.
Chem., 2007, 4973; (c) C. Landgrafe, W. S. Sheldrick and
M. Südfeld, Eur. J. Inorg. Chem., 2007, 4973.
C.
J.
Elsevier,
Chem.
Commun.,
1997,
667;
(f) H. T. Teunissen and C. J. Elsevier, Chem. Commun., 24 X. Chen, P. Xue, H. H. Y. Sung, I. D. Williams, M. Peruzzini,
1998, 1367; (g) M. C. van Engelen, H. T. Teunissen, J.-G. de C. Bianchini and G. Jia, Organometallics, 2005, 24, 4330.
Vries and C. J. Elsevier, J. Mol. Catal. A: Chem., 2003, 206, 25 The complex [Ru(κ4-NP3)(η2-H2)(H)](BPh4) was indepen-
185; (h) L. Rosi, M. Frediani and P. Frediani, J. Organomet.
Chem., 2010, 695, 1314; (i) M. R. L. Furst, R. Le Goff,
D. Quinzler, S. Mecking, C. H. Botting and D. J. Cole-
Hamilton, Green Chem., 2012, 14, 472; ( j) F. M. A. Geilen,
B. Engendahl, A. Harwardt, W. Marquardt, J. Klankermayer
and W. Leitner, Angew. Chem., Int. Ed., 2010, 49, 5510.
dently synthesised by the reaction of [Ru(κ4-NP3)(H)2]
[31P{1H} NMR (acetone-d6: AM2 system, 63.0 PM, 59.3 PA,
2JPP = 18 Hz] with triflic acid and NaBPh4 under an atmos-
phere of hydrogen. M. Peruzzini, et al., unpublished data.
See also; D. Michos, X.-L. Luo and R. H. Crabtree, Inorg.
Chem., 1992, 31, 4245.
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 2495–2501 | 2501