CO2 activation. 7. Formation of the catalytically active intermediate in the hydrogenation of carbon dioxide to formic acid using the [{(COD)Rh(μ-H)}4]/ Ph2P(CH2)4PPh2 catalyst: First direct observation of hydride migration from rhodium to coordinated 1,5-cyclooctadiene

Article abstract of DOI:10.1021/om950913f

The nature of the catalytically active intermediate formed in situ from the tetrameric cluster [{(COD)Rh(μ-H)}₄] (COD = 1,5-cyclooctadiene; 1) and the bidentate phosphane Ph₂P(CH₂)₄PPh₂ (dppb) during hydrogenation of CO₂ to formic acid was investigated. Kinetic measurements suggest the initial formation of a catalyst precursor that reacts with dihydrogen to give the actual active species. NMR spectroscopic investigations of the reaction of 1 with dppb in THF-d₈ reveal three phosphorus-containing products that were fully characterized by one- and two-dimensional techniques, including 2D-(³¹P,¹H)-COLOC spectra. The tetrameric hydride cluster [{(dppb)Rh(μ-H)}₄] (2) and the double-phosphane-substituted monomeric rhodium hydride [(dppb)₂RhH] (3) are formed as byproducts in low yield. The (phosphane)rhodium η³-cyclooctenyl complex [(dppb)Rh(η³-C₈H₁₃)] (4), arising via hydride transfer from rhodium to coordinated COD, is the major product, containing about 80% of the dppb. Complex 4 was isolated from the mixture of products, and its molecular structure was determined by X-ray crystal diffraction. Hydrogenolysis of the allyl moiety in the presence of excess dppb was shown to yield 3 presumably via the 14e species [(dppb)RhH]. The results are most consistent with the formation of 4 as the actual precursor for the active species [(dppb)RhH] in the rhodium-catalyzed hydrogenation of CO₂ to formic acid using in situ catalysts consisting of 1 and dppb.

Full text of DOI:10.1021/om950913f

2078
Organometallics 1996, 15, 2078-2082
CO2 Activa tion . 7.† F or m a tion of th e Ca ta lytica lly
Active In ter m ed ia te in th e Hyd r ogen a tion of Ca r bon
Dioxid e to F or m ic Acid Usin g th e [{(COD)Rh (µ-H)}4]/
P h 2P (CH2)4P P h 2 Ca ta lyst: F ir st Dir ect Obser va tion of
Hyd r id e Migr a tion fr om Rh od iu m to Coor d in a ted
1,5-Cycloocta d ien e
Franz Gassner, Eckhard Dinjus, Helmar Go¨rls, and Walter Leitner*,‡
Arbeitsgruppe CO2-Chemie der Max-Planck-Gesellschaft an der Friedrich-Schiller-Universita¨t
J ena, Lessingstrasse 12, 07743 J ena, FRG
Received November 27, 1995X
The nature of the catalytically active intermediate formed in situ from the tetrameric
cluster [{(COD)Rh(µ-H)}4] (COD ) 1,5-cyclooctadiene; 1) and the bidentate phosphane
Ph2P(CH2)4PPh2 (dppb) during hydrogenation of CO2 to formic acid was investigated. Kinetic
measurements suggest the initial formation of a catalyst precursor that reacts with
dihydrogen to give the actual active species. NMR spectroscopic investigations of the reaction
of 1 with dppb in THF-d8 reveal three phosphorus-containing products that were fully
characterized by one- and two-dimensional techniques, including 2D-(31P,1H)-COLOC spectra.
The tetrameric hydride cluster [{(dppb)Rh(µ-H)}4] (2) and the double-phosphane-substituted
monomeric rhodium hydride [(dppb)2RhH] (3) are formed as byproducts in low yield. The
(phosphane)rhodium η3-cyclooctenyl complex [(dppb)Rh(η3-C8H13)] (4), arising via hydride
transfer from rhodium to coordinated COD, is the major product, containing about 80% of
the dppb. Complex 4 was isolated from the mixture of products, and its molecular structure
was determined by X-ray crystal diffraction. Hydrogenolysis of the allyl moiety in the
presence of excess dppb was shown to yield 3 presumably via the 14e species [(dppb)RhH].
The results are most consistent with the formation of 4 as the actual precursor for the active
species [(dppb)RhH] in the rhodium-catalyzed hydrogenation of CO2 to formic acid using in
situ catalysts consisting of 1 and dppb.
In tr od u ction
cal8 studies have shown that CO2 insertion into the
rhodium-hydride bond of the 14e species [(P2)RhH] (P2
) chelating bisphosphane) is most likely the key step
during the catalytic cycle of CO2 hydrogenation in
dipolar nonprotic solvents under the conditions sum-
marized in eq 1. The 14e species [(P2)RhH] tend to form
The hydrogenation of CO2 to formic acid is a promis-
ing approach to the use of CO2 as a raw material in
chemical synthesis.1,2 Late-transition-metal complexes
act as homogeneous catalysts for this process, and
efficient catalytic systems are known to operate in
organic solvents,3,4 in aqueous solutions,5 and under
supercritical conditions.6 Experimental3b,7 and theoreti-
solvent, NEt3
(1)
HCO2H
H2 + CO2
T = 25 °C, p0 = 40 atm
† Part 6: Reference 3d.
oligomers of type [{(P2)Rh(µ-H)}x], and the catalytic
behavior of these clusters in hydrogenation reactions
has been studied in detail, namely by the groups of
Muetterties9 and Fryzuk.10
We have recently reported on the use of the tetrameric
rhodium hydride cluster [{(COD)Rh(µ-H)}4] (COD ) 1,5-
cyclooctadiene; 1) in the presence of the chelating
phosphane Ph2P(CH2)4PPh2 (dppb) as a very efficient
homogeneous catalyst for the hydrogenation of CO2 to
‡ Current address: Max-Planck-Institut fu¨r Kohlenforschung, Kaiser-
Wilhelm-Platz 1, 45470 Mu¨lheim/Ruhr, FRG; Fax: +49-208-306 2980.
E-mail: leitner@mpi-muelheim.mpg.de.
X Abstract published in Advance ACS Abstracts, March 15, 1996.
(1) (a) Behr, A. Carbon Dioxide Activation by Metal Complexes;
VCH: Weinheim, Germany, 1988. (b) Ayers, W. M., ed. Catalytic
Activation of Carbon Dioxide; ACS Symposium Series 363; American
Chemical Society: Washington, DC, 1988. (c) Aresta, M.; Schloss, J .
V., Eds. Enzymatic and Model Carboxylation and Reduction Reactions
for Carbon Dioxide Utilization; NATO ASI Series C, No. 314; Kluwer
Academic Press: Dordrecht, The Netherlands, 1990.
(2) (a) J essop, P. G.; Ikariya, T.; Noyori, R. Chem. Rev. 1995, 95,
259. (b) Leitner, W. Angew. Chem. 1995, 107, 2391; Angew. Chem.,
Int. Ed. Engl. 1995, 34, 2207. (c) Leitner, W. Coord. Chem. Rev., in
press.
(3) (a) Graf, E.; Leitner, W. J . Chem. Soc., Chem. Commun. 1992,
623. (b) Leitner, W.; Dinjus, E.; Gassner, F. J . Organomet. Chem. 1994,
475, 257. (c) Fornika, R.; Go¨rls, H.; Seemann, B.; Leitner, W. J . Chem.
Soc., Chem. Commun. 1995, 1479. (d) Graf, E.; Leitner, W. Chem. Ber.
1996, 129, 91.
(4) (a) Inoue, Y.; Izumida, H.; Sasaki, Y.; Hashimoto, H. Chem. Lett.
1976, 863. (b) Tsai, J .-C.; Nicholas, K. M. J . Am. Chem. Soc. 1992,
114, 5117. (c) Lau, C. P.; Chen, Y. Z. J . Mol. Catal. A: Chem. 1995,
101, 33.
(5) (a) Taqui Khan, M. M.; Halligudi, S. B.; Shukla, S. J . Mol. Catal.
1989, 57, 47. (b) Gassner, F.; Leitner, W. J . Chem. Soc., Chem.
Commun. 1993, 1465.
(6) a) J essop, P. G.; Ikariya, T.; Noyori, R. Nature 1994, 368, 231.
(b) J essop, P. G.; Ikariya, T.; Noyori, R. Science 1995, 269, 1065. (c)
J essop, P. G.; Hsiao, Y.; Ikariya, T.; Noyori, R. J . Am. Chem. Soc. 1996,
118, 344.
(7) Burgemeister, T.; Kastner, F.; Leitner, W. Angew. Chem. 1993,
105, 781; Angew. Chem., Int. Ed. Engl. 1993, 32, 739.
(8) Hutschka, F.; Dedieu, A.; Leitner, W. Angew. Chem. 1995, 107,
1905; Angew. Chem., Int. Ed. Engl. 1995, 34, 1742.
(9) (a) Day, V. W.; Fredrich, M. F.; Reddy, G. S.; Sivak, A. J .; Pretzer,
W. R.; Muetterties, E. L. J . Am. Chem. Soc. 1977, 99, 8091. (b) Sivak,
A. J .; Muetterties, E. L. J . Am. Soc. Chem. 1979, 101, 4878.
(10) (a) Fryzuk, M. D. Can. J . Chem. 1983, 61, 1347. (b) Fryzuk, M.
D.; Piers, W. E.; Einstein, F. W. B.; J ones, T. Can. J . Chem. 1989, 67,
883.
0276-7333/96/2315-2078$12.00/0 © 1996 American Chemical Society

CO2 Activation
Organometallics, Vol. 15, No. 8, 1996 2079
F igu r e 2. Activation of the in situ catalyst 1/dppb in
acetone/NEt3 by pre-hydrogenation. For reaction condi-
tions, see Figure 1.
F igu r e 1. Catalytic formation of formic acid by hydroge-
nation of CO2 in acetone/NEt3 solution using the in situ
catalyst 1/dppb. Reaction conditions: solvents acetone/NEt3
(5:1); c(Rh) ) 2.47 × 10-3 mol L-1; dppb/Rh ) 1.2; room
temperature; p0 ) 40 atm.
bers up to 2200 have been reported, and a lower limit
for the turnover frequency of 375 h-1 has been estimated
in this case.3b
formic acid under the conditions summarized in eq 1.3b
One possible explanation for the much higher activity
of the hydride-bridged precursor 1 compared to that of
its chloride analog [{(COD)Rh(µ-Cl)}2] is the direct
formation of the catalytically active 14e species [(dppb)-
RhH] or its oligomers [{(dppb)Rh(µ-H)}x] by COD
substitution from 1.3b However, derivatives of the
species [{(P2)Rh(µ-H)}x] containing dppb or other bi-
dentate phosphanes bearing aryl groups at phosphorus
have not been described up to now. Furthermore, no
information was available on the reactivity of 1 toward
phosphanes. We therefore decided to investigate the
pathway of formation of the actual catalytically active
species from the in situ catalyst consisting of 1 and dppb
in more detail.
Similar to other rhodium catalysts,3b,4b the in situ
system 1/dppb can be activated by treatment with
dihydrogen prior to reaction, as illustrated in Figure 2.
After pre-hydrogenation of the acetone/NEt3 solution
containing 1 and dppb with 20 atm of H2 for 1 h,
catalysis starts immediately with a maximum turnover
frequency of 354 h-1, i.e. 50 times faster than without
pre-hydrogenation under otherwise identical conditions.
These findings lead to the conclusion thatsat least in
acetone/NEt3sthe catalytically active species for CO2
hydrogenation is not formed directly from 1 and dppb
under catalytic conditions.
Further information on the formation of the active
species comes from stoichiometric reactions of 1 and
dppb as investigated by NMR spectroscopy. THF-d8
was used as a solvent in these experiments for solubility
Resu lts a n d Discu ssion
1
reasons. The characteristic high-field H NMR signal
One criterion for the direct formation of a catalytically
active species from in situ catalysts is the absence of
an induction period; i.e., catalysis must start im-
mediately with the maximum turnover frequency as
soon as the reactants are introduced. The kinetics of
the catalytic hydrogenation of CO2 to formic acid in
dipolar aprotic solvent systems can be followed readily
by NMR spectroscopy.3a-c Acetone has been suggested
as the preferred solvent for this type of experiment, as
it yields homogeneous reaction solutions in the presence
of trialkylamines under all conditions.3b Figure 1 shows
the catalytic formation of formic acid when a solution
1
(δ -12.05, qt, J RhH ) 14.2 Hz) for the hydride 111
disappears immediately when 4 equiv of dppb (1 equiv
per rhodium atom) is added as a solid in one portion to
a THF-d8 solution of 1 at room temperature. The 31P
NMR spectrum shows three new sets of signals indicat-
ing the formation of (phosphane)rhodium complexes.
The main product exhibits a doublet at δ 38.2 with a
1
coupling constant J RhP of 199.5 Hz. The second com-
1
plex is characterized by a doublet at δ 27.0 with J RhH
) 149 Hz. A third, very weak signal consists of a
doublet of badly resolved multiplets centered at δ 35.5.
1
The distance of the two multiplets corresponds to a J RhP
of 1 (2.5 mmol L-1 in rhodium) and dppb (3.0 mmol L-1
)
coupling constant of 169 Hz. At the same time, two new
in a 5:1 mixture of acetone and triethylamine is stirred
under a roughly equimolar atmosphere of CO2 and H2
at a total pressure of 40 atm at room temperature. The
initial rate of formic acid formation is very slow (turn-
over frequency (tof) ) 7.2 h-1) but increases during the
course of reaction. After 20 h, the equilibrium concen-
tration of formic acid under the given conditions (1.75
mol L-1 3b) is not yet reached and a total of 480 catalytic
1
signals in the high-field region of the H NMR spectrum
of the reaction mixture indicate the formation of rhod-
ium hydride complexes. A quintet of doublets is cen-
tered at δ -11.2, and a multiplet consisting of at least
seven lines appears at δ -11.8 ppm. Corresponding
cross-peaks in the 2D-{31P,1H}-COLOC spectrum un-
equivocally demonstrate that these hydride signals arise
from the same compound as the phosphorus signals at
27.0 and 35.5 ppm, respectively.
cycles are performed during this period of time.
A
maximum turnover frequency of 44 h-1 is obtained from
the slope of the concentration profile shown in Figure
1. These results are surprising in view of the excellent
performance of the same catalyst in batch reactions
using the DMSO/NEt3 solvent system. Turnover num-
The 31P broad-band-decoupled 1H{31P} NMR spec-
trum gives further evidence for the characterization of
(11) Kulzick, M.; Price, R. T.; Muetterties, E. L.; Day, V. W.
Organometallics 1982, 1, 1256.

2080 Organometallics, Vol. 15, No. 8, 1996
Gassner et al.
Sch em e 1. P h osp h or u s-Con ta in in g P r od u cts fr om th e Rea ction of [{(COD)Rh (µ-H)}4] (1) w ith d p p ba
a
The ratios were obtained from 31P NMR spectra of the crude reaction mixture. No other phosphorus-containing compounds
were detectable.14
the two hydride complexes. The signal at δ -11.2 ppm
collapses to a doublet under these conditions, indicating
that the original quintet structure arises from coupling
with four equivalent P atoms. In agreement with NMR
data for the known complexes [(P2)2RhH],12b we assign
these signals to [(dppb)2RhH] (3). It is noteworthy in
this context that 3 is not accessible by common routes
used for the synthesis of complexes [(P2)2RhH].12 Our
results show that this is due to limitations of these
synthetic methodologies rather than the instability of
3.
The formula [{(dppb)Rh(µ-H)}4] (2) can be assigned
to the second hydride-containing complex on the basis
of its NMR data. The hydride signal of 2 appears as a
quintet under 31P-decoupled conditions, and this pattern
is only explained if formation of a tetrameric cluster
with fluxional hydrides similar to 1 is assumed.11 The
multiplet structure observed in the 31P{1H} NMR of this
compound may arise from the higher order spin system
formed by chemically equivalent, but magnetically
inequivalent, 31P and 103Rh nuclei. Similar spectro-
scopic properties were reported for a tetrameric hydride-
bridged cluster bearing a chelating phosphite ligand.9a
The major product of the reaction of 1 with dppb
contains no hydride ligand, as is obvious from the
absence of any cross-peak in the 2D-(31P,1H)-COLOC
spectrum. A characteristic set of proton resonances
consisting of a slightly broadened triplet at δ 5.30 and
a quartet at δ 3.71 suggests the formation of the η3-
cyclooctenyl complex [(dppb)Rh(η3-C8H13)] (4).13 This
assignment was confirmed by the 2D-(1H,1H)-COSY
spectrum and 13C-DEPT and 2D-(13C,1H)-HETCOR
experiments. Scheme 1 summarizes the product dis-
tribution of the reaction of [{(COD)Rh(µ-H)}4] (1) with
dppb as revealed by NMR spectroscopy.14
F igu r e 3. Molecular structure of [(dppb)Rh(η3-C8H13)] (4)
as determined by X-ray diffraction of 4‚0.5C5H12. Hydro-
gens were omitted for clarity. Selected distances (Å) and
angles (deg): Rh-C1, 2.219(10); Rh-C2, 2.122(10); Rh-
C3, 2.209(10); Rh-P1, 2.237(3); Rh-P2, 2.228(3); C1-C2,
1.409(14); C2-C3, 1.424(14); C3-C4, 1.491(14); C4-C5,
1.54(2); C5-C6, 1.51(2); C1-Rh-C3, 68.8(4); P1-Rh-P2,
98.63(10); C2-C1-C3, 124.0(10); [P1-Rh-P2]/[C1-Rh-
C3], 6.0; [P1-Rh-P2]/[C1-C2-C3], 113.1.
structure of 4 together with some characteristic dis-
tances and angles is shown in Figure 3.
The rhodium center in complex 4 is symmetrically
bound to the two phosphorus atoms of the dppb ligand.
The Rh-P distances and the P-Rh-P angle are in the
typical range for rhodium dppb complexes.3c,15 The
seven-membered chelate ring has a twisted-boat con-
formation. The four atoms P1, P2, C1, and C3 form a
nearly ideal square-planar arrangement around the
rhodium center, the angle between P1-Rh-P2 and C1-
Rh-C3 being 6°. The η3-cyclooctenyl ligand adopts the
expected16 boatlike conformation and is symmetrically
bound to the rhodium atom via the three allylic carbon
atoms C1, C2, and C3. The hydrogen atoms on these
allylic carbons and on the adjacent carbons C4-C8 are
among those located on the difference map and do not
exhibit any interaction with the rhodium center. The
angle between the plane P1-Rh-P2 and the plane
formed by the allylic carbon atoms is 113.1°. Nearly
the same value is found for a bis(triphenylphosphane)-
The X-ray crystal structure analysis of complex 4
verified the NMR spectroscopic assignments (see Ex-
perimental Section for details). Crystals suitable for
X-ray diffraction analysis of the pentane solvate
4‚0.5C5H12 were obtained by slow crystallization from
light petroleum ether (bp 40-60 °C). The molecular
(12) (a) Sacco, A.; Ugo, R. J . Chem. Soc. 1964, 3274. (b) J ames, B.
R.; Mahajan, D. Can. J . Chem. 1979, 57, 180.
(13) Fornika, R.; Dinjus, E.; Go¨rls, H.; Leitner, W. J . Organomet.
Chem. 1996, 511, 145.
(14) The formation of the bisphosphane complex 3 together with the
stoichiometry of the reaction requires the presence of phosphane-free
rhodium species. The presence of weak multiplets between 5.6 and
5.4 ppm in the 1H spectrum indeed indicates the presence of such
species, together with some free cyclooctene. The low concentration of
these compounds together with severe overlap of signals in other parts
of the proton spectrum did not allow a final assignment, however.
(15) Anderson, P. M.; Pignolet, L. H. Inorg. Chem. 1981, 20, 4101.
(16) (a) J onas, K. Angew. Chem. 1985, 97, 292. (b) Braunstein, P.;
Faure, T.; Knorr, M.; Stra¨hlfeldt, T.; DeCian, A.; Fischer, A. Gazz.
Chim. Ital. 1995, 125, 35 and references therein.

CO2 Activation
Organometallics, Vol. 15, No. 8, 1996 2081
rhodium complex containing the cyclooctadienyl ligand
cally active species, as shown by the high turnover
frequency of 354 h-1, which corresponds well to the
lower limit of 375 h-1 estimated for the same catalyst
in DMSO solution.
17
C8H11
.
In order to obtain further insight into the formation
of 2-4, complex 1 was reacted subsequently with 4
equiv of dppb per cluster unit. All three complexes 2-4
are formed in a constant ratio from the beginning, with
the remaining cluster 1 staying intact. If a 2-fold excess
of dppb per rhodium is employed, the three complexes
are formed again and free dppb is observed in the 31P-
{1H} NMR spectrum. These results suggest that the
reaction of 1 with dppb leads to the COD substitution
product 2 only via a minor reaction pathway. Cleavage
of the tetrameric core of 1 predominates, resulting in
formation of the bisphosphane hydride complex 3 and
the cyclooctenyl complex 4, presumably via the common
intermediate [(dppb)Rh(H)(COD)] (5). Formation of 3
and 4 via 2 can be excluded, as 2 has been found to be
present in reaction mixtures containing free COD and
dppb. The low reactivity of tetrameric clusters [{(P2)-
Rh(µ-H)}4] toward excess P2 has been noted also by
Fryzuk et al.10a A similar pathway has been suggested
most recently for the formation of cyclooctenyl com-
plexes of rhodium by hydride transfer from formate to
coordinated COD.13
Similarly to the tetrameric cluster, the cyclooctenyl
complex 4 was found to be stable in the presence of
excess dppb.18 However, both compounds are readily
converted to complex 3 when the reaction mixture is
pressurized with 10 atm of dihydrogen in a high-
pressure NMR tube. In the case of complex 4, this
reaction most likely involves removal of the η3-allyl
moiety by hydrogenolysis10 whereafter the resulting 14e
species [(dppb)RhH] is trapped immediately by excess
dppb to yield 3 (eq 2).
Con clu sion
In conclusion, we have shown that the η3-cyclooctenyl
complex [(dppb)Rh(η3-C8H13)] (4) is formed as the major
product in the reaction of the hydride cluster 1 and the
chelating phosphane dppb. The molecular structure of
4 was elucidated by an X-ray crystal structure analysis,
and direct transfer of hydride from rhodium to coordi-
nated COD was demonstrated to account most likely for
the formation of 4. The η3-cyclooctenyl ligand from 4
can be removed by reaction with dihydrogen leading to
the 14e species [(dppb)RhH] that was trapped as the
stable 18e rhodium hydride complex [(dppb)2RhH] (3)
in the presence of excess dppb.
The present findings imply that complex 4 is the
actual catalyst precursor during CO2 hydrogenation to
formic acid using the in situ catalyst 1/dppb. The
results obtained from catalytic experiments under vari-
ous conditions are in agreement with the initial forma-
tion of 4 and its hydrogenolysis to yield the 14e species
[(dppb)RhH] that serves as the catalytically active
species.
Exp er im en ta l Section
Gen er a l Com m en ts. All manipulations were carried out
under an atmosphere of argon using standard Schlenk tech-
niques. The complex [{(COD)Rh(µ-H)}4] (1) was synthesized
according to a recently published procedure.20 The phosphane
dppb was obtained from Strem Chemicals and used without
further purification. The high-pressure reactor used for
catalytic experiments is described in detail elsewhere.3b NMR
spectra were recorded on a Bruker AC 200F, operating at
200.13 MHz (1H), 50.26 MHz (13C), and 80.96 MHz (31P).
Broad-band 31P decoupling was achieved using an external
BSV 3BX decoupling unit. Chemical shifts are referenced to
SiMe4 using the solvent resonances as internal standards for
1H and 13C and to H3PO4 as external standard for 31P. The
2D-{1H,31P}-COLOC spectra were carried out using standard
Bruker software based on the pulse sequence reported in ref
21. The mixing time was optimized for a P-H coupling
10 atm of H2
[(dppb)Rh(η3-C8H13)]
THF-d8, room temp
4
dppb
"[(dppb)RhH]"
[(dppb)2RhH]
(2)
3
The possible formation of 4 under catalytic conditions
was verified experimentally when 1 was reacted with
dppb in a two-layer mixture resulting from addition of
a solution of dppb in DMSO to a solution of 1 in NEt3.
The complexes 1 and 4 are contained exclusively in the
NEt3 layer under these conditions, but the system is
known to be homogenized under catalytic conditions.3b
As in THF-d8, hydrogenolysis of 4 may then lead rapidly
to the 14e species [(dppb)RhH] either stabilized as
amine adduct19 or as oligomer 2.
Similarly, the initial formation of 4 readily explains
the marked induction period for CO2 hydrogenation
using the 1/dppb catalyst in acetone/NEt3. In contrast
to the situation in DMSO/NEt3, hydrogenolysis of 4
appears to be sluggish in the presence of CO2 in this
case; the reason for this discrepancy is not yet clear.
However, pre-hydrogenation of the catalyst solution
with 20 atm of H2 for 1 h cleanly leads to the catalyti-
2
constant of ∼10 Hz, as J PH is commonly between 5 and 20
Hz in (phosphane)rhodium hydride complexes.
Rea ction of [{(µ-H)Rh (COD)}4] (1) w ith d p p b in THF -
d 8. A solution of 1 (11.1 mg, 13.1 µmol) in 0.45 mL of THF-d8
was placed in an NMR tube, and the phosphane dppb (22.3
mg, 52.2 µmol) was added as a solid in one portion or in four
equal portions. The mixture was agitated until a clear
1
homogeneous solution was obtained, and H and 31P{1H} NMR
spectra were recorded immediately after each addition. Char-
acteristic spectroscopic data are as follows.
1
[{(dppb)Rh(µ-H)}4] (2): 1H NMR δ -11.8 (m, J RhH ) 13.1
Hz, Rh-H); 31P{1H} NMR δ 35.5 (dm, 1J RhP ) 169 Hz, Rh-P).
1
[(dppb)2RhH] (3): 1H NMR δ -11.2 (dqt, J RhH ) 7.5 Hz,
1
2J PH ) 19.9 Hz, Rh-H); 31P{1H} NMR δ 27.0 (d, J RhP ) 146
Hz, Rh-P).
3
[(dppb)Rh(η3-cyclooctenyl)] (4): 1H NMR δ 5.30 (t, br, J HH
1
3
) 7.9 Hz, H, ≈CH≈), 3.71 (q, br, J HH ) 7.9 Hz, 2H, dCH-);
(17) Vitulli, G.; Raffaelli, A.; Costantino, P. A.; Barberini, C.;
Marchetti, F.; Merlino, S.; Skell, P. S. J . Chem. Soc., Chem. Commun.
1983, 232.
31P{1H} NMR 38.2, d, J RhP ) 199 Hz.
1
(18) Note that the related complex [(dppp)Rh(η3-C8H13)] reacts with
excess dppp to give [(dppp)2RhH] and 1,3-cyclooctadiene: Fornika, R.
Ph.D. Thesis, J ena, Germany, 1994.
(20) Bo¨nnemann, H.; Brijoux, W.; Brinkmann, R.; Dinjus, E.;
Fretzen, R.; J oussen, T.; Korall, B. J . Mol. Catal. 1992, 74, 323.
(21) Kessler, H.; Griesinger, C.; Zarbock, J .; Loosli, H. R. J . Magn.
Reson. 1984, 57, 331.
(19) Hofmann, P.; Meier, C.; Englert, U.; Schmidt, M. U. Chem. Ber.
1992, 125, 353.

2082 Organometallics, Vol. 15, No. 8, 1996
Gassner et al.
Ta ble 1. Cr ysta llogr a p h ic Da ta for 4‚0.5C5H12
Rea ction of [{(µ-H)Rh (COD)}4] (1) w ith Excess d p p b
a n d Dih yd r ogen in THF -d 8. 1 (7.3 mg, 8.6 µmol) was
dissolved in 0.30 mL of THF-d8, and dppb (29.3 mg, 68.7 µmol)
was added as a solid in one portion. The solution was
transferred to a thick-walled NMR tube (Wilmad 522-PV) by
color
yellow
formula
mol wt
C
674.6
36H41P2Rh‚0.5C5H12
cryst syst
space group
monoclinic
P21/n (No. 14)
1
syringe, and H and 31P{1H} NMR spectra were recorded. The
lattice param at 183 K
tube was then pressurized to 10 atm with H2 and carefully
agitated, and spectra were recorded again. Ca u tion ! Han-
dling of glass apparatus under high pressure is hazardous and
has to be carried out under appropriate safety conditions.
a (Å)
b (Å)
c (Å)
10.252(1)
19.627(2)
18.053(1)
92.97(1)
3629.4(6)
1.235
4
5.82
7110
6733
â (deg)
Rea ction of [{(µ-H)Rh (COD)}4] (1) w ith d p p b in
a
vol (Å3)
density (calcd), Dc (g cm-3
no. of formula units Z
)
DMSO/NEt3 Tw o-P h a se System . A solution of dppb (29.1
mg, 68.3 µmol) in 15 mL of DMSO was layered with a solution
of 1 (14.5 mg, 17.1 µmol) in 3.0 mL of NEt3. After 15 h,
samples were withdrawn from each layer and analyzed by 31P-
{1H} NMR spectroscopy (unlocked). Complex 4 was identified
as the only phosphorus-containing species in the NEt3 phase,
whereas only free dppb could be detected in the DMSO layer
(δ -15.2, s).
abs coeff µ(Mo KR) (cm-1
no. of collcted data
no. of unique data
)
no. of obsd data with I e 2σ(I)
no. of variables
4032
397
R1
0.074
wR2
0.184
1.72
Rea ction of [{(µ-H)Rh (COD)}4] (1) w ith d p p b in THF
an d Isolation of [(dppb)Rh (η3-C8H13)]‚0.5C5H12 (4‚0.5C5H12).
1 (72.8 mg, 85.8 µmol) was dissolved in 5 mL of THF; a solution
of dppb (0.16 g, 0.38 mmol) in 5 mL THF was added at room
temperature, and the mixture was stirred for 20 min. The
solvent was removed and the residue thoroughly dried under
high vacuum. The remaining solid was dissolved in a mini-
mum of light petroleum ether (bp 40-60 °C), and the solution
was placed in a refrigerator at -35 °C. The yellow crystals
formed after 4 weeks were removed from the solvent and
analyzed by NMR, mass spectroscopy, and X-ray crystal
diffraction. Yield: 73.0 mg (0.10 mmol, 29%). Mp: 231 °C
dec. EI-MS: m/ z 638 (4+), 529 (4+ - C8H13), 425 (4+ - C8H13
- C6H5). 1H NMR (C6D6): δ 7.8-7.4 (m, 8H, PPh, m-H), 7.3-
residual electron density (e Å-3
)
Table 1. Atomic coordinates, bond lengths and angles, and
thermal parameters have been deposited at the Cambridge
Crystallographic Data Centre. See Information for Authors,
Issue No. 1.
Ca ta lytic Hyd r ogen a tion of CO2 to F or m ic Acid Usin g
th e in situ Ca ta lyst [{(µ-H)Rh (COD)}4] (1)/d p p b in Ac-
eton e/NEt3. 1 (12.6 mg, 14.8 µmol) and dppb (30.2 mg, 70.8
µmol) were dissolved in a mixture of 20.0 mL of acetone and
4.0 mL of NEt3, and the mixture was stirred vigorously for 15
min. The solution was transferred to a high-pressure reactor
and pressurized to 20 atm with CO2. A total pressure of 40
atm was adjusted with H2, and the reaction mixture was
stirred with a constant rate of 750 rpm at room temperature.
Samples were withdrawn after given time intervals, frozen in
liquid nitrogen, and diluted with acetone-d6 in air in order to
avoid formic acid decomposition.3b The formic acid concentra-
tion was determined by 1H NMR using NEt3 as an internal
standard.3b In the pre-hydrogenation experiment, the order
of addition of the two gases was reversed and the solution was
stirred under H2 for 1 h before introducing CO2.
3
6.7 (m, 12H, PPh, o,p-H), 5.27 (t, J HH ) 7.9 Hz, 1H, ≈CH≈),
3
3.65 (q, J HH ) 7.9 Hz, 2H, dCH-), 2.2-1.2 (18H, -CH2-).
2
4
13C{1H} NMR (C6D6): δ 132.5/133.0 (t, | J PC + J PC| ) 12 Hz,
3
5
o-C), 127.5/127.6 (t, | J PC + J PC| ) 10 Hz, m-C), 128.0/128.2
(s, p-C), 103.9 (dt, 1J RhC ) 5 Hz, 2J PC ≈ 2 Hz, ≈CH≈), 68.6 (m,
3
1
dCH-), 32.3 (d, J RhC ) 2 Hz, dCHCH2-), 30.6 (t, | J PC
+
3J PC| ) 24 Hz, -CH2P), 30.0 (t, 4J PC(cis) + 4J PC(trans) ) 6 Hz,
2
dCHCH2CH2-), 23.3 (s, dCH(CH2)2CH2-), 24.0 (t, | J PC
+
4J PC| ) 8 Hz, -CH2CH2P), the ipso-C atoms were not detected.
31P{1H} NMR (C6D6): 38.5 (d, J RhP ) 199 Hz).
1
Ack n ow led gm en t. Financial support from the Max-
Planck-Society and the Fonds der Chemischen Industrie
is gratefully acknowledged. We thank Degussa AG for
a generous loan of RhCl3‚3H2O.
X-r a y Cr ysta l Str u ctu r e An a lysis of [(d p p b)Rh (η3-
C8H13)]‚0.5C5H12 (4‚0.5C5H12). X-ray data were collected on
an Enraf-Nonius CAD4 diffractometer using Mo KR radiation
(λ ) 0.710 69 Å). Data were corrected for Lorentz and
polarization effects but not for absorption. The structure was
solved by Patterson methods (SHELXS-86) and refined by full-
Su p p or tin g In for m a tion Ava ila ble: Tables giving crys-
tal data and details of data collection, positional and thermal
parameters, and bond distances and angles and an ORTEP
plot of 4 (8 pages). Ordering information is given on any
current masthead page.
2
matrix least squares against Fo (SHELXL-93). Hydrogen
atoms were located in the difference map. After the hydrogen
atoms were included at calculated positions with fixed thermal
parameters, non-hydrogen atoms were refined anisotropically.
Cell parameters and other pertinent data are summarized in
OM950913F