An efficient binuclear catalyst for decomposition of formic acid

Article abstract of DOI:10.1039/a805789c

The complex [Ru₂(μ-CO)(CO)₄(μ-dppm)₂] is more active than mononuclear ruthenium complexes as a catalyst for decomposition of formic acid to CO₂ and H₂; under conditions of highest activity, a coordinatively unsaturated diruthenium dihydride [Ru₂H(μ-H)(μ-CO)(CO)₂(μdppm)₂] is present and can be isolated from solution.

Full text of DOI:10.1039/a805789c

An efficient binuclear catalyst for decomposition of formic acid
Yuan Gao,a Joshi Kuncheria,a Glenn P. A. Yapb and Richard J. Puddephatta
a Department of Chemistry, University of Western Ontario, London, Canada N6A 5B7. E-mail: pudd@julian.uwo.ca
b Department of Chemistry and Biochemistry, University of Windsor, Sunset Drive, Windsor, Canada N9B 3P4
Received (in Bloomington, IN, USA) 24th July 1998, Accepted 25th September 1998
The complex [Ru2(m-CO)(CO)4(m-dppm)2] is more active
than mononuclear ruthenium complexes as a catalyst for
decomposition of formic acid to CO2 and H2; under
conditions of highest activity, a coordinatively unsaturated
expected from statistical considerations (H2 :HD:D2 = 1:2:1,
determined by MS) but in the very early stages of reaction H2 is
predominant. Under these reaction conditions, there was no
evidence for reversibility which would lead to isomerization
between HCO2D and DCO2H. Decomposition of H13CO2H
gave only H2 and 13CO2, with no free or coordinated 13CO
detectable by either IR or 13C NMR.
diruthenium
dihydride
[Ru2H(m-H)(m-CO)(CO)2(m-
dppm)2] is present and can be isolated from solution.
This paper reports the first study of the decomposition of formic
acid to hydrogen and carbon dioxide using a locked binuclear
catalyst and presents evidence that this is an unusual catalytic
reaction in which two metals can act cooperatively to give
enhanced activity.† The reversible reaction between HCO2H
and H2 + CO2 has been the subject of considerable interest,
either for catalytic transfer hydrogenation using formic acid or
for utilization of CO2 as a reagent in organic synthesis.1,2
The reaction of [Ru2(m-CO)(CO)4(m-dppm)2] 1,3 dppm =
Ph2PCH2PPh2, in acetone (5 3 1023 m) with formic acid (1021
m) in a closed tube at 20 °C was monitored by NMR. The
products were H2 [d(1H) 4.5] and CO2 [d(13C) 125.8] only and
the reaction was complete in 0.3 h, corresponding to a mean
turnover rate of ca. 70 h21, and considerably higher activity was
observed when the reaction was carried out in an unsealed
vessel, for reasons discussed below. The activity is significantly
higher than for comparable mononuclear ruthenium complex
catalysts: for example, [RuHBr(CO)(PEt2Ph)3] gives a turnover
rate of ca. 4 h21 in refluxing acetic acid (117 °C).1a
The high catalytic activity of 1 prompted a more detailed
study and some important features of the binuclear catalysis
have been elucidated. The catalysis is more efficient in the
dipolar aprotic solvent acetone than in solvents such as toluene
or dichloromethane. The catalytic reaction is strongly or
completely inhibited by the presence of excess CO; in a sealed
vessel, CO dissociates from 1 at intermediate stages of reaction
and then acts as inhibitor, whereas, in a vessel in which evolved
gases sweep CO from the system, this effect cannot occur and so
the catalysis is faster. The reaction appears not to be wholly
intramolecular since decomposition of either HCO2D or
DCO2H leads to formation of a mixture of H2, HD and D2 rather
than HD alone, but the conclusion is weakened by the
observation that decomposition of HCO2H in the presence of D2
gives both H2 and HD. The final product mixture from
decomposition of HCO2D or DCO2H is approximately that
There were interesting changes in the ruthenium complexes
present at various stages of the catalytic reaction and several of
these could be isolated or identified spectroscopically. When
the reaction was carried out in a sealed tube, the only ruthenium
complex present when reaction was complete was unchanged 1
but other complexes were present during catalysis (Scheme 1).
The first complex formed at 230 °C was [Ru2(m-H)(m-
CO)(CO)4(m-dppm)2]+ 2, as the formate salt.‡ This complex is
formed by protonation of the Ru–Ru bond of 1 and the same
cation is formed by protonation with other acids such as
H[BF4].4 Next to be formed was the cation [Ru2(m-HCO2)-
(CO)4(m-dppm)2]+ 3 (also as the formate salt), whose spectro-
scopic properties‡ are very similar to the known m-acetate
analogue;4,5 the first formation of H2 could be detected at this
stage. When most formic acid was consumed, two more
complexes were formed transiently. The major complex was
characterized as [Ru2H2(CO)4(m-dppm)2] 5, by the following
spectroscopic data.‡6 In the 1H NMR spectrum, a hydride
resonance at d(1H) 29.25 [qnt, J(PH) 9 Hz] integrated for two
protons and a single resonance at d(CH2P2) 4.6 were observed,
in the 13C NMR spectrum a single terminal carbonyl resonance
was observed at d(CO) 196.8, and in the 31P NMR spectrum a
singlet was observed at d(P) 34.3. These data indicate structure
5, having D2h symmetry; a less symmetrical but fluxional
structure is also possible6 though no change in the NMR spectra
was observed at 270 °C. The second transient complex was
tentatively identified as [Ru2H(HCO2)(CO)4(m-dppm)2] 4.‡ It is
characterized in the 1H NMR by resonances at d 26.7 (br s, 1H,
RuH) and at d 8.5 (s, 1H, HCO2), in the 13C NMR (in a reaction
using H13CO2H) by d 165 (s, CH, HCO2) (the concentration of
4 was never great enough to allow identification of the metal
carbonyl resonances even using 13C enriched starting material
1) and in the 31P NMR by a single resonance at d 39.8.5 If
structure 4 is correct, it is required to be fluxional in order to
give a single resonance in the 31P NMR spectrum.4 Further
+
+
P
P
CO
P
Ru
P
P
Ru
P
P
Ru
P
P
Ru
P
CO
i
OC
OC
CO
CO
OC
OC
CO
CO
ii
H
Ru
O
Ru CO
OC
C
O
C
O
O
P
C
P
H
3
2
1
iv
iii
viii
vi
P
P
P
P
P
P
CO
Ru CO
HCO2
H
H
OC
OC
H
H
CO
Ru
Ru
Ru
OC
OC
Ru
Ru
vii
CO
CO
v
OC
H
C
O
P
P
P
P
P
P
4
5
6
+
Scheme 1 Reagents: i, H+; ii, HCO2H, 2H2; iii, H2, 2H+; iv, HCO2 , 2CO; v, 2CO2; vi, CO, 2H2; viii, 2CO; viii, H , CO, 2H2.
2
Chem. Commun., 1998, 2365–2366
2365

H2
HCO2H
CO
P
P
P
P
H
H
H
HCO2
OC
Ru
Ru
Ru
Ru
CO
OC
C
O
C
O
P
P
P
P
6
CO2
P
P
H
H
CO
CO
Ru
Ru
P
(CO2)
C
O
P
Scheme 2 A possible mechanism of catalysis.
than in the trans terminal [d(13C) 200.8] or bridging [d(13C)
278.6] carbonyl sites (exchange detected after one day).§ While
the reactions of Scheme 1 provide a viable route for the catalytic
reaction in the presence of CO, the data suggest that CO-
deficient complexes, such as 6, are most active and it is likely
that other key intermediates are too short-lived to be detected. A
reasonable catalytic cycle involving 6 is shown in Scheme 2.
In summary, this article describes a novel binuclear catalytic
system for formic acid decomposition, in which the major
ruthenium complexes in solution are dependent on both reaction
conditions and the stage of the catalytic reaction. It suggests that
binuclear and cluster complexes, especially those that can easily
achieve coordinative unsaturation, may have distinct advan-
tages over mononuclear transition metal catalysts for this and
related catalytic reactions.7
Fig. 1 A view of the structure of [Ru2H(m-H)(m-CO)(CO)2(m-dppm)2].
Distances (Å): Ru(1)–Ru(2) 2.8769(5), Ru(1)–C(1) 1.857(5), Ru(1)–C(2)
2.198(5), Ru(2)–C(2) 2.006(4), Ru(2)–C(3) 1.841(5). The hydride H-atoms
were located but not refined; approximate distances Ru(1)–H(1) 2.16,
Ru(2)–H(1) 2.16, Ru(1)–H(2) 1.66.
study indicated that the concentration of 5 with respect to 2 was
pH dependent, since addition of Et3N led to an increase in the
relative concentration of 5. Overall then, at high [HCO2H],
complex 1 reacted to give 2 and 3 and, as HCO2H was
consumed, the concentrations of these complexes decreased and
the transient complexes 4 and 5 appeared, quickly followed by
reformation of 1. These observations are readily rationalized
since 5 is expected to be very reactive towards formic acid, and
formate probably dissociates easily from 4 in the presence of
formic acid to give the less nucleophilic anion [H(O2CH)2]2;
hence the concentration of these complexes 4 and 5 only builds
up to detectable levels when the concentration of formic acid is
low.
When the reaction was carried out in an unsealed vessel, the
initial reactions were similar but a new complex [Ru2H(m-H)(m-
CO)(CO)2(m-dppm)2], 6, was formed in the later stages rather
than 4 or 5. Complex 6 is a unique example of a coordinatively
unsaturated binuclear ruthenium dihydride;4,6 it could be
crystallized from the reaction mixture and was fully charac-
terized by an X-ray structure determination (Fig. 1) as well as by
spectroscopic methods.‡ Complex 6 was stable in the solid state
but in solution it was stable only in the presence of hydrogen
and in the complete absence of oxygen; it reacted rapidly with
CO to give 1 with loss of H2. Clearly, this high reactivity of 6
with CO explains why no 6 is formed when the reaction is
carried out in a sealed tube; 6 reacts with one equivalent of CO
to give 5 and then with a second equivalent of CO to give 1 and
H2. Solutions containing the coordinatively unsaturated com-
plex 6 were particularly active for further catalytic decomposi-
tion of formic acid.
It is interesting to speculate on why the binuclear system
described above is so reactive for decomposition of formic acid.
The key steps in the initial catalytic reaction are likely to be the
overall oxidative addition of formic acid to ruthenium(0) to give
a hydrido(formato) complex and then probably a ß-elimination
from the formate to give a transient dihydrido(CO2) complex
which ultimately yields H2 and CO2. Both of these proposed
steps require a vacant coordination site, and the necessary
dissociation of two CO ligands is probably easier to accomplish
in the binuclear system. There is some independent evidence for
CO labilization cis to the bridging hydride ligand in complex 2.
Thus, exposure of 2 to 13CO led to carbonyl exchange but the
substitution in the terminal carbonyl sites cis to the m-H ligand
[d(13C) 198.6] was much faster (exchange detected in < 1 h)
We thank the NSERC (Canada) for financial support.
Notes and references
† Binuclear complexes have been identified in formic acid decomposition
previously but were not thought to be involved in the catalytic cycle.1c
‡ Selected spectroscopic data: 2: d(1H) 28.9 [qnt, 1H, J(PH) 9 Hz, Ru2(m-
H)]; d(13C) 199, 201 (terminal CO), 278.6 (m-CO); d(31P) 27.8 (dppm).
Preliminary X-ray data on the [BF4]2 salt gives d(Ru–Ru) 2.960(3) Å
compared to 2.903(2) Å in 1. 3: d(13C) 188, 206 (terminal CO); 181 (HCO2);
d(31P) 30.9 (dppm). 4: d(1H) 26.7 (m, 1H, RuH), 8.5 (s, 1H, HCO2); d(13C)
165 (HCO2); d(31P) 39.9 (dppm). 5: d(1H) 29.2 (qnt, 2H, RuH) d(13C) 197
(terminal CO); d(31P) 34.3 [dppm]. 6: d(1H) 29.3 [t, 1H, RuH], 29.6
[quint, 1H, Ru2(m-H)]; d(31P) = 42.5, 46.5 (m, dppm). Crystal data for 6:
¯
monoclinic, space group P21/n, a
= 11.583(1), b = 28.557(3), c =
16.783(2) Å, b = 97.817°, V = 5499(1) Å3, T = 296 K, m = 7.1 cm21
7115 reflections, R1 = 0.0466, wR2 = 0.0918. CCDC 182/1033.
,
§ The assignments are based on the observation of J(CC) coupling between
the mutually trans bridging and terminal carbonyl ligands in the fully 13CO
enriched complex.
1 (a) R. S. Coffey, Chem. Commun., 1967, 923; (b) S. H. Strauss, K. H.
Whitmire and D. F. Shriver, J. Organomet. Chem., 1979, 174, C59;
(c) R. S. Paonessa and W. C. Trogler, J. Am. Chem. Soc., 1982, 104,
3529; (d) J.-C. Tsai and K. M. Nicholas, J. Am. Chem. Soc., 1992, 114,
5117.
2 P. G. Jessop, T. Ikariya and R. Noyori, Chem. Rev., 1995, 95, 260.
3 J. Kuncheria, H. A. Mirza, H. A. Jenkins, J. J. Vittal and R. J. Puddephatt,
J. Chem. Soc., Dalton Trans., 1998, 285; G. M. Ferrence, P. E. Fanwick,
C. P. Kubiak and R. J. Haines, Polyhedron, 1997, 16, 1453; H. A. Mirza,
J. J. Vittal and R. J. Puddephatt, Inorg. Chem., 1993, 32, 1327.
4 R. J. Haines, in Comprehensive Organometallic Chemistry II, ed. D. F.
Shriver and M. I. Bruce, Pergamon, Oxford, 1995, vol. 7, ch. 11; R. W.
Hilts, S. J. Sherlock, M. Cowie, E. Singleton and M. M. de V. Steyn,
Inorg. Chem., 1990, 29, 3161.
5 For reversible insertion of CO2 into Ru–H bonds of mononuclear
ruthenium comlexes to give ruthenium formates see for example: M. K.
Whittlesey, R. N. Perutz and M. H. Moore, Organometallics, 1996, 15,
5166; G. Jia and D. W. Meek, Inorg. Chem., 1991, 30, 1953.
6 K. J. Edwards, J. S. Field, R. J. Haines, B. D. Homann, M. W. Stewart,
J. Sundermeyer and S. F. Woollam, J. Chem. Soc., Dalton Trans., 1996,
4171.
7 R. D. Adams, Comprehensive Organometallic Chemistry II, Pergamon,
Oxford, 1995, vol. 10; P. Braunstein and J. Rose, in Catalysis by Di- and
Polynuclear Metal Complexes, ed. R. D. Adams and F. A. Cotton, Wiley,
New York, 1997, p. 346.
Communication 8/05789C
2366
Chem Commun., 1998