Concurrent synergism and inhibition in bimetallic catalysis: Catalytic binuclear elimination, solute-solute interactions and a hetero-bimetallic hydrogen-bonded complex in Rh-Mo hydroformylations

Article abstract of DOI:10.1021/ja909696b

Hydroformylations of cyclopentene and 3,3-dimethylbut-1-ene were performed using both Rh₄(CO)₁₂ and (η⁵-C ₅H₅)Mo(CO)₃H as precursors in n-hexane at 298 K. Both stoichiometric and catalytic hydroformylations were conducted as well as isotopic labeling experiments. Six organometallic pure component spectra were recovered from the high-pressure FTIR experiments, namely the known species Rh₄(CO)₁₂, (η⁵-C₅H ₅)Mo(CO)₃H, RCORh(CO)₄, and the new heterobimetallic complexes RhMo(CO)₇(η⁵-C ₅H₅), a weak hydrogen bonded species (η⁵- C₅H₅)Mo(CO)₃H-C₅H ₉CORh(CO)₄, and a substituted RhMo(CO) _7-y(η⁵-C₅H₅)L_y, where y = 1 or 2 and L = (--C₅H₈). The main findings were (1) catalytic binuclear elimination (CBER) occurs between (η⁵-C ₅H₅)Mo(CO)₃H and RCORh(CO)₄ resulting in aldehyde and RhMo(CO)₇(η⁵-C ₅H₅), and this mechanism is responsible for ca. 10% of the product formation; (2) molecular hydrogen is readily activated by the new heterobimetallic complex(es); (3) FTIR and DFT spectroscopic evidence suggests that the weak hydrogen bonded species (η⁵-C₅H ₅)Mo(CO)₃H-C₅H₉CORh(CO)₄ has an interaction of the type η⁵-C₅H₄-HO - C; and (4) independent physicochemical experiments for volumes of interaction confirm that significant solute-solute interactions are present. With respect to the efficiency of the catalytic cycle, the formation of a weak (η⁵-C₅H₅)Mo(CO)₃H-C ₅H₉CORh(CO)₄ complex results in a significant decrease in the measured turnover frequency (TOF) and is the primary reason for the inhibition observed in the bimetallic catalytic hydroformylation. Such hydrogen bonding through the η⁵-C₅H₅ ring might have relevance to inhibition observed in other catalytic metallocene systems. The present catalytic system is an example of concurrent synergism and inhibition in bimetallic homogeneous catalysis.

Full text of DOI:10.1021/ja909696b

Published on Web 03/12/2010
Concurrent Synergism and Inhibition in Bimetallic Catalysis:
Catalytic Binuclear Elimination, Solute-Solute Interactions
and a Hetero-Bimetallic Hydrogen-Bonded Complex in Rh-Mo
Hydroformylations
Chuanzhao Li,* Shuying Cheng, Martin Tjahjono, Martin Schreyer, and
Marc Garland*
Institute of Chemical and Engineering Sciences, A*STAR (Agency for Science, Technology and
Research), 1 Pesek Road, Jurong Island, Singapore, 627833
Received March 19, 2009; E-mail: marc_garland@ices.a-star.edu.sg; li_chuanzhao@ices.a-star.edu.sg
Abstract: Hydroformylations of cyclopentene and 3,3-dimethylbut-1-ene were performed using both
5
Rh
4
(CO)12 and (η -C
5
H
5
)Mo(CO)
3
H as precursors in n-hexane at 298 K. Both stoichiometric and catalytic
hydroformylations were conducted as well as isotopic labeling experiments. Six organometallic pure
component spectra were recovered from the high-pressure FTIR experiments, namely the known species
5
5
Rh
4
(CO)12, (η -C
5
H
5
)Mo(CO)
3
H, RCORh(CO)
4
, and the new heterobimetallic complexes RhMo(CO)
7
(η -
5
C
5
H
5
), a weak hydrogen bonded species (η -C
RhMo(CO)7-y(η -C
5
H
5
)Mo(CO)
). The main findings were (1) catalytic binuclear
H and RCORh(CO) resulting in aldehyde and
3
H-C
5 9 4
H CORh(CO) , and a substituted
5
5
H
5 y
)L , where y ) 1 or 2 and L ) (π-C
H
5 8
5
elimination (CBER) occurs between (η -C
5
H
5
)Mo(CO)
3
4
5
RhMo(CO)
7
(η -C
5
H
5
), and this mechanism is responsible for ca. 10% of the product formation; (2) molecular
hydrogen is readily activated by the new heterobimetallic complex(es); (3) FTIR and DFT spectroscopic
5
evidence suggests that the weak hydrogen bonded species (η -C
5
H
5
)Mo(CO)
3
5 9 4
H-C H CORh(CO) has an
5
interaction of the type η -C
5
H
4
-H· · ·OdC; and (4) independent physicochemical experiments for volumes
of interaction confirm that significant solute-solute interactions are present. With respect to the efficiency
5
of the catalytic cycle, the formation of a weak (η -C
5
H
5
)Mo(CO)
3
5 9 4
H-C H CORh(CO) complex results in a
significant decrease in the measured turnover frequency (TOF) and is the primary reason for the inhibition
5
observed in the bimetallic catalytic hydroformylation. Such hydrogen bonding through the η -C
H
5 5
ring might
have relevance to inhibition observed in other catalytic metallocene systems. The present catalytic system
is an example of concurrent synergism and inhibition in bimetallic homogeneous catalysis.
5
Introduction
in some cases, preventative action can be taken. However, more
subtle inhibition in homogeneous catalyzed systems is certainly
less understood. Some undesirable changes might be induced
Various mechanisms can be responsible for deactivation in
transition-metal homogeneous catalyzed systems. Some of the
1
by relatively weak solute-solute interactions, such as hydrogen
more dramatic changes such as ligand degradation, that is,
6
bonding, ion-pairing, dipole-dipole interactions, etc. The
2
phosphine oxidation and intrametalation, reduction of com-
detection of subtle deactivation phenomenon, as well as the
confirmation of relatively weak solute-solute interactions,
represents a significant challenge. Moreover, detailed under-
standing of these effects would surely provide new opportunities
and directions for research in homogeneous catalysis.
3
plexes to metal, reactions with impurities to form ultrastable
4
and nonlabile species, etc. are relatively well understood, and
(
1) Detlef, H.; de Vries, A. H. M.; de Vries, J. G. In Handbook of
Homogeneous Hydrogenation; de Vries, J. G., Cornelis, J., Eds.;
Elsevier VCH: New York, 2007; Vol. 3, pp 1483-1516.
In the past two decades, in situ spectroscopic methods have
been increasingly applied to transition-metal homogenously
(
2) (a) Garrou, P. E. Chem. ReV. 1985, 85 (3), 171–185. (b) Parshall,
G. W.; Knoth, W. H.; Schunn, R. A. J. Am. Chem. Soc. 1969, 91
7
catalyzed reactions. This has resulted in a considerably better
(18), 4990–4995. (c) Hage, R.; Iburg, J. E.; Kerschner, J.; Koek, J. H.;
mechanistic understanding of many systems, by facilitating the
Lempers, E. L. M.; Martens, R. J.; Racherla, U. S.; Russell, S. W.;
Swarthoff, T.; van Vliet, M. R. P.; Warnaar, J. B.; Wolf, L. V. D.;
Krijnen, B. Nature 1994, 369 (6482), 637–639. (d) Collins, T. J. Acc.
Chem. Res. 1994, 27 (9), 279–285. (e) Horwitz, C. P.; Fooksman,
D. R.; Vuocolo, L. D.; Gordon-Wylie, S. W.; Cox, N. J.; Collins,
T. J. J. Am. Chem. Soc. 1998, 120 (19), 4867–4868. (f) Kong, K. C.;
Cheng, C. H. J. Am. Chem. Soc. 1991, 113 (16), 6313–6315. (g) Van
Leeuwen, P. W. N. M.; Roobeek, C. F.; Orpen, A. G. Organometallics
8
identification of new chemistry and intermediates and allowing
(5) van Leeuwen, P. W. N. M. Appl. Catal., A: Gen. 2001, 212 (1-2),
61–81.
(6) (a) Cesarotti, E.; Ugo, R.; Kaplan, L. Coord. Chem. ReV. 1982, 43,
275–98. (b) Huang, Y.; Rawal, H. J. Am. Chem. Soc. 2002, 124, 9662–
9663. (c) Breit, B.; Seiche, W. J. Am. Chem. Soc. 2003, 125, 6608–
6609. (d) Fulton, J. R.; Holland, A. W.; Fox, D. J.; Bergman, R. G.
Acc. Chem. Res. 2002, 35, 44–56. (e) Casey, C. P.; Clark, T. B.; Guzai,
A. J. Am. Chem. Soc. 2007, 129, 11821–11827.
1
990, 9 (8), 2179–2181.
(
(
3) Lewis, L. N. J. Am. Chem. Soc. 1990, 112 (16), 5998–6004.
4) Walczuk, E. B.; Kamer, P. C. J.; van Leeuwen, P. W. N. M. Angew.
Chem., Int. Ed. 2003, 42 (38), 4665–4669.
10.1021/ja909696b  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 4589–4599 9 4589

A R T I C L E S
Li et al.
9
the evaluation of detailed kinetics and turnover frequencies.
that these interactions are responsible for the inhibitory effects
on the catalytic cycle.
In addition, advances in signal processing and chemometric
10
techniques have significantly improved the limits of detection.
Although these developments have largely focused on under-
standing the primary characteristics of quite active catalytic
systems, potential certainly exists for using these tools to better
understand more subtle aspects of the chemistry, for example
changes in selectivity and activity patterns.
Experimental Section
General. All solution preparations and transfers were carried
out under purified argon (99.9995%, Saxol, Singapore) atmosphere
using standard Schlenk techniques or conducted in a glovebox
(especially when handling (η -C
5
5
H
5
)Mo(CO)
3
H). The argon was
further purified before use by passing it through a deoxy and zeolite
packed column. Carbon monoxide (research grade, 99.97%, Saxol,
Singapore), hydrogen (99.9995%, Saxol, Singapore), and deuterium
Bimetallic and multimetallic homogeneous catalyzed reactions
have attracted considerable attention, due in part to dramatically
1
1
increased reaction rates observed in some systems. The
mechanistic basis for synergism has been clearly identified for
a few systems using in situ spectroscopic techniques. These
include (1) the Ir/Ru catalyzed carbonylation of methanol, where
NMR has shown that Ru aids in the abstraction of iodine from
(
99.8%, Spectra Gases, Branchburg, NJ) were also further purified
through deoxy and zeolite columns before they were used in the
hydroformylation experiments. Purified air was used to purge the
Bruker Vertex70 FT-IR spectrometer system.
The chemicals used in the present study include n-hexane (99.5%
1
2
5
Fluka puriss), Rh
(
4
(CO)12 (98% Strem), and (η -C
5
H
5
)Mo(CO)
3
H
a crucial iridium intermediate, and (2) the Rh/Mn and Rh/Re
hydroformylations of alkenes, where in situ FTIR has shown
Sigma Aldrich), cyclopentene (99%, Fluka), 3,3-dimethylbut-1-
9
d,e,g,h,13
ene (99%, Fluka, 33DMB), and cyclopentane carboxaldehyde. The
n-hexane was purified by distillation from NaK under argon. The
cyclopentene was purified by distillation from maleic anhydride to
that catalytic binuclear elimination occurs.
In the present contribution, the hydroformylation of cyclo-
pentene and 3,3-dimethylbut-1-ene starting with Rh (CO)12 and
H as catalyst precursors was performed, and
9
e
remove dienes, followed by distillation from CaH
2
under argon.
3
4
5
5
The organometallics Rh
4
(CO)12 and (η -C
H
5 5
)Mo(CO) H were used
(
5 5 3
η -C H )Mo(CO)
without further purification.
in situ FTIR was used as the quantitative spectroscopic method.
In this heterobimetallic system, both synergistic and inhibitory
effects were observed. Consequently, stoichiometric and labeling
experiments were performed to clarify mechanistic issues,
particularly the presence of catalytic binuclear elimination. The
spectroscopic, kinetic, and physicochemical measurements
together with DFT calculations confirm that solute-solute
interactions are occurring, in particular hydrogen bonding, and
Apparatus. Hydroformylations were performed in an in-house
designed 100 mL high-pressure SS316 reactor, which was connected
with an injection block, and a high-pressure magnetically driven
gear pump (Model GAH-X21, Micro pump, Vancouver, WA). The
high-pressure infrared cell was situated in a Bruker Vertex-70 FT-
-
1
IR spectrometer, and the spectral resolution was 2 cm with an
interval of 0.12 cm- for the range 1000-2500 cm . A detailed
1
-1
description of the reactor, pump, injection block and mass transfer
issues can be found elsewhere. In addition, a high-pressure syringe
pump (PHD 4400, Harvard Apparatus) was used for injecting (η -
5 5 3
C H )Mo(CO) H/n-hexane under high-pressure reaction conditions.
14
5
(
7) (a) Laurenczy, G.; Helm, L. In Mechanisms in Homogeneous Catalysis:
A Spectroscopic Approach; Heaton, B., Ed.; Wiley: Germany, 2005;
pp 81-106. (b) Haynes, A. In Mechanisms in Homogeneous Catalysis:
A Spectroscopic Approach; Heaton, B., Ed.; Wiley: Germany, 2005;
pp 107-150. (c) Viviente, E. M.; Pregosin, P. S.; Schott, D. In
Mechanisms in Homogeneous Catalysis: A Spectroscopic Approach;
Heaton, B., Ed.; Wiley: Germany, 2005; pp 1-80. (d) Casey, C. P.;
Beetner, S. E.; Johnson, J. B. J. Am. Chem. Soc. 2008, 130, 2285–
Examples of the use of this experimental system can be found
9
elsewhere, and further details concerning the present experimental
setup can be found in the Supporting Information.
In-Situ Spectroscopic and Kinetic Measurements. All of the
catalytic hydroformylation experiments were performed in a similar
manner. First, a background spectrum of the IR sample chamber
5
2
295.
was recorded. Rh
(CO)12 and/or (η -C
5
H
5
)Mo(CO)
3
H were dis-
4
(
8) (a) Liu, J.; Heaton, B.; Iggo, J. A.; Whyman, R. Angew. Chem., Int.
Ed. 2004, 43 (1), 90–94. (b) Haynes, A.; et al. J. Am. Chem. Soc.
solved in 50 mL of n-hexane and then transferred under argon to
the reactor. Under 0.2 MPa CO pressure, infrared spectra of the
solution in the high-pressure cell were recorded. The total system
pressure was raised to 1.7 MPa CO, and the stirrer and high-pressure
gear pump were started. After equilibration, infrared spectra of the
solution in the high-pressure cell were recorded. Then 1 mL of
cyclopentene was injected via the injection block and 1.7 MPa
hydrogen was added to start the reactions. The experimental design
2
004, 126, 2847–2861.
(
9) (a) Garland, M.; Pino, P. Organometallics 1990, 10, 1693–1704. (b)
Liu, G.; Volken, R.; Garland, M. Organometallics 1999, 18, 3429–
3
4
1
2
436. (c) Feng, J.; Garland, M. Organometallics 1999, 18 (3), 417–
27. (d) Li, C.; Widjaja, E.; Garland, M. J. Am. Chem. Soc. 2003,
25, 5540–5548. (e) Li, C.; Widjaja, E.; Garland, M. Organometallics
004, 23, 4131–4138. (f) Liu, G.; Li, C.; Guo, L.; Garland, M. J.
Catal. 2006, 237, 67–78. (g) Li, C.; Chen, L.; Garland, M. J. Am.
Chem. Soc. 2007, 129, 133327–13334. (h) Li, C.; Chen, L.; Garland,
M. AdV. Synth. Catal. 2008, 350, 679–690.
is listed in Table 1. The high-pressure syringe pump was used for
5
injecting (η -C
1
1
5
H
5
)Mo(CO)
3
H/n-hexane in semibatch Exp 5 at ca.
(
(
10) Garland, M. In Mechanisms in Homogeneous Catalysis: A Spectro-
scopic Approach; Heaton, B., Ed.; Wiley: Germany, 2005; pp 151-
60 min. Spectra were recorded at 1 min intervals in the range
-1
000-2500 cm . In every 4-h experiment, ca. 250 spectra were
1
93.
taken. Note that the gas pressures were changed in Exp 7-9 to
11) (a) Guo, N.; Stern, C. L.; Marks, T. J. J. Am. Chem. Soc. 2008, 130
7), 2246–2261. (b) Sammis, G. M.; Danjo, H.; Jacobsen, E. N. J. Am.
promote a higher yield of RCORh(CO)
labeled experiments.
4
for the stoichiometric and
(
Chem. Soc. 2004, 126 (32), 9928–9929. (c) Li, H.; Marks, T. J. Proc.
Natl. Acad. Sci. U.S.A. 2006, 103 (42), 15295–15302. (d) Adams,
R. D., Albert Cotton, F., Eds. Catalysis by di-and polynuclear metal
cluster complexes; Wiley: New York, 1998.
Volume of Interaction Measurements. Recently, a method has
been developed to determine the volumes of interaction between
15
solutes in dilute multicomponent solutions. This method was
(
(
12) (a) Sunley, G. J.; Watson, D. J. Catal. Today 2000, 58, 293–307. (b)
Whyman, R.; Wright, A. P.; Iggo, J. A.; Heaton, B. T. J. Chem. Soc.,
Dalton Trans. 2002, 771–777.
applied to the present system, during low-pressure noncatalytic
conditions, to determine the degree of interaction present in
13) (a) Collman, J. P.; Hegedus, L. S.; Norton, J. R.; Finke, R. G. Principles
and Applications of Organotransition Metal Chemistry; University
Science Books: Mill Valley, CA, 1987. (b) Norton, J. R. Acc. Chem.
Res. 1979, 12, 139–145. (c) Jones, W. D.; Huggins, J.; Bergman, R. G.
J. Am. Chem. Soc. 1981, 103, 4415–4423. (d) Martin, B.; Warner,
D. K.; Norton, J. R. J. Am. Chem. Soc. 1986, 108 (1), 33–39. (e)
Nappa, M. J.; Santi, R.; Halpern, J. Organometallics 1985, 4, 34–41.
(14) Gao, F.; Ng, K.; Li, C.; Krummel, K.; Allian, A.; Garland, M. J. Catal.
2006, 237, 49–57.
(15) (a) Tjahjono, M.; Guo, L.; Garland, M. Chem. Eng. Sci. 2005, 60
(12), 3239–3249. (b) Tjahjono, M.; Garland, M. Chem. Eng. Sci. 2007,
62, 3861–3867.
4
590 J. AM. CHEM. SOC. 9 VOL. 132, NO. 13, 2010

Concurrent Synergism and Inhibition in Bimetallic Catalysis
A R T I C L E S
1
7
Table 1. Experimental Design for Rh/Mo Hydroformylations of
evaluation has been carefully evaluated. The rates of the product
formation of cyclopentane carboxaldehyde were calculated in terms
of turn over frequency (TOF) by using eq 1.
Cyclopentene (CP) and 3,3-Dimethylbut-1-ene (33DMB)
5
Rh
4
(CO)12 (η -C
5
H
5
)Mo(CO)
3
H
Mo/Rh CO
ratio (MPa)
H
2
(MPa)
exp no loading (mg)
loading (mg)
alkene T (K)
1
2
3
4
5
6
7
8
9
14.5
14.0
14.0
15.0
14.7
0
20.1
14.6
20.2
0
0
1.7 1.7
CP
298
298
298
298
298
298
298
298
d[aldehyde] /dt
t
15.0
25.0
37.0
0.82 1.7 1.7
1.36 1.7 1.7
1.88 1.7 1.7
1.82 1.7 1.7
CP
CP
CP
CP
CP
CP
CP
TOF )
(1)
t
[
C H CORh(CO) ]
5 9 4 t
a
35.0
Volume of Interactions. The investigation of solute-solute
interaction via volumetric measurements involved 4 steps: (1)
application of a linear-bilinear response-surface model for the total
35.0
60.5
52.7
-
1.7 1.7
b
c
2.29 2.7 0, 1.0
2.75 2.7 0, 1.0
d
d
15
45.2
1.71 2.8 1.2 (D2) 33DMB 298
volume; (2) determinations of volumetric standard states, namely
partial molar volumes for solutes at infinite dilution and pure molar
volume of the solvent; (3) application of an excess volumetric model
based on the McMillan-Mayer theory of solutions; and (4)
calculation of the homotactic and heterotactic interaction volumetric
coefficients using multilinear regression and validation of the
obtained coefficients using statistical tests. The details of the above-
a
η5-C
b
(
5
H
5
)Mo(CO)
3
H was added at ca. 160 min. Stoichiometric
2
hydroformylation was carried out initially, and then H was added at ca.
70 min. Stoichiometric hydroformylation was carried out initially,
and then
deutero-formylation was carried out initially, then (η -C
was added at ca. 320 min.
c
2
d
H
2
was added at ca. 280 min. Rhodium catalyzed
5
5
H )Mo(CO)
5
3
H
1
5
mentioned procedure can be found elsewhere.
DFT Calculations. All DFT calculations were performed with
Gaussian 03 at the DFT/PBEPBE level. DFT computations were
carried out in n-heptane solvent using the SCRF model IEFPCM.
The organic molecule cyclopentane carboxaldehyde was modeled
at the PBEPBE/6-31G (d, p) level. The effective core potentials
(
ECP) basis set was used to represent the innermost electrons of
the molybdenum and rhodium atoms, and the basis set of double-ꢀ
LANL2DZ was employed for molybdenum and rhodium atoms.
The basis set of 6-31G(d, p) was used to model the carbon and
hydrogen atoms of the Cp rings and carbon, oxygen atoms of
5
carbonyl groups in the (η -C
5 5 3 5 9 4
H )Mo(CO) H, C H CORh(CO) , and
the associated hydrogen bonded complexes. The basis function
-311G (d, p) was used for metal-bound hydrogen. The present
6
study takes a similar computational approach, using the same level
of theory and basis sets, as previously used to understand
intermolecular hydrogen bonding in systems containing metal-
5
18
5 5 3
locenes, and particularly (η -C H )Mo(CO) H. The optimized
geometries and parameters as well as free energies can be found
in the Supporting Information.
Results
Figure 1. BTEM spectral estimates of the primary observable solutes during
the mixed-metal Rh-Mo hydroformylations of cyclopentene. Note that
species Y and Z only appeared under stoichiometric conditions in Exp 7-8.
Pure Component Spectra. BTEM analysis of the entire set
of experimental in situ FTIR hydroformylation (Exp 1-8)
spectra provided spectral estimates for the hexane solvent,
dissolved CO, as well as the observable organic and organo-
5
solutions of (η -C
5
H
5
)Mo(CO)
3
H and aldehyde. Details of the
experiments and the data are provided in the Supporting Informa-
tion.
metallic solutes present. Figure 1 shows the spectral estimates
5
for the precursors Rh
4
(CO)12 and (η -C
5
H
5
3
)Mo(CO) H, the
organics cyclopentene and cyclopentane carboxaldehyde, as well
as an intermediate C CORh(CO) and three new species X,
Y, and Z. The species X has vibrational maxima at 2111, 2064.
Computational Section
5
H
9
4
Pure Component Spectra, Concentrations, TOF. The in situ
reaction spectra were consolidated to give a single absorbance
matrix. The spectra were then subjected to singular value decom-
position (SVD). The basis vectors were transformed into pure
component spectral estimates using the Band-Target Entropy
Minimization (BTEM) algorithm. Full-range BTEM spectral
estimates are presented (i.e., Figure 1), as well as partial-range
BTEM spectral estimates (i.e., Figures 6 and 8), to better resolve
-1
2
038.5, 2029, 2020.5, 1945, and 1700 (broad) cm . The other
two species Y and Z were only observed in the stoichiometric
experiments. Species Y has vibrations at 2086, 2034, 2010,
-
1
1
6
1996, 193,1 and 1940 (shoulder) cm whereas species Z has
-1
vibrations at 1989, 1965, 1950, 1920, and 1913 (shoulder) cm .
All of the spectral estimates exhibit very good signal-to-noise
ratios and possess few spectral artifacts (note that the signal
the band shifts occurring and reduce artifacts/colinearities. The
5
moles of alkene, aldehyde, RCORh(CO)
)Mo(CO)
4
, Rh
4
(CO)12, and (η -
intensity of X was particularly low in the catalytic experiments).
5
C
5
H
5
3
H for all experiments could be obtained using a
The spectral estimates of Rh
4
(CO)12, (η -C
5
H
5
)Mo(CO)
3
H, and
multiple linear regression/least-squares approach using the pure
absorptivities. The mathematical procedures have been discussed
in detail elsewhere.
cyclopentene obtained from the reaction spectra are almost
identical to the authentic references. The spectral estimate of
the nonisolatable complex C H CORh(CO) compares favorably
5 9 4
1
0
The appropriateness of various finite difference schemes for the
evaluation of in situ homogeneous catalytic rate data and TOF
(
17) Shirt, R.; Garland, M.; Rippin, D. W. T. Anal. Chim. Acta 1998, 374,
7–91.
6
(
16) (a) Chew, W.; Widjaja, E.; Garland, M. Organometallics 2002, 21,
(18) (a) Belkova, N. V.; Gutsul, E. I.; Filippov, O. A.; Levina, V. A.;
Valyaev, D. A.; Epstein, L. M.; Lledos, A.; Shubina, E. S. J. Am.
Chem. Soc. 2006, 128 (11), 3486–3487. (b) Private communication
with Shubina, E. S. clarifying the basis sets used.
1882–1990. (b) Widjaja, E.; Li, C.; Garland, M. Organometallics 2002,
21, 1991–1997. (c) Widjaja, E.; Li, C.; Garland, M. J. Catal. 2004,
223, 278–289.
J. AM. CHEM. SOC. 9 VOL. 132, NO. 13, 2010 4591

A R T I C L E S
Li et al.
5
Figure 3. Time-dependent concentrations of (η -C
5
H
5
)Mo(CO)
3
H during
Figure 2. Time-dependent formation of C
lytic hydroformylation experiments starting with Rh
5
H
9
CORh(CO)
4
during cata-
5
the catalytic hydroformylation experiments. The arrow corresponds to the
introduction of (η -C H )Mo(CO) H in experiment 5.
5 5 3
4
(CO)12 and (η -
5
5
C
C
5
H
H
5
)Mo(CO)
)Mo(CO)
3
H. The arrow corresponds to the introduction of (η -
H in experiment 5.
5
5
3
Table 2. Initial Rates of Formation of C
for the First 40 Minutes) and the Maximum Yields of
CORh(CO) as a Function of Mo/Rh Ratio in the Catalytic
Hydroformylations
5 9 4
H CORh(CO) (Evaluated
C
H
5 9
4
Mo/Rh
loading(mole ratio)
rate of C
formation(mole fraction /min)
5
H
9
CORh(CO)
4
maximum yield of
exp no
5 9 4
C H CORh(CO) (%)
-
-
-
-
7
7
7
6
1
2
3
4
0
7.1 × 10
8.2 × 10
9.3 × 10
1.2 × 10
40.1
48.2
59.2
72.5
0.82
1.36
1.88
9
d,e,g,h,19
to previous studies.
Since only a rough spectral estimate
2
0
6
of the large cluster Rh (CO)16 could be obtained via BTEM
analysis, it is apparently present at only trace concentration
levels in these experiments. The spectral estimate of new species
5
X appears quite similar to a superposition of the spectra of (η -
C
5
H
5
)Mo(CO)
shifts are apparent.
Concentration Profiles. Figure 2 shows the time dependence
of the intermediate C CORh(CO) in the catalytic experi-
3 5 9 4
H and C H CORh(CO) although noticeable band
5
H
9
4
Figure 4. Time-dependent formation of cyclopentane carboxaldehyde in
ments. As seen in this figure, the base-case pure-rhodium
experiment shows (1) a relatively low initial rate of
the catalytic hydroformylation experiments. The arrow corresponds to the
5
5 5 3
introduction of (η -C H )Mo(CO) H in experiment 5.
C
C
5
H
H
9
CORh(CO)
CORh(CO)
4
formation and (2) a relatively low yield of
Figure 2 provides information not only on the rate of
formation and the yield of C H CORh(CO) but also on
5
9
4
. In contrast, the experiments with increasing
5
initial concentrations of (η -C
H
5 5
)Mo(CO)
CORh(CO)
increasingly higher yields of C
3
H show (1) increas-
5
9
4
deactivation phenomenon. Close inspection shows that the
concentration of C CORh(CO) remains constant after 120
min when only Rh (CO) was used as a catalyst precursor, but
ingly higher rates of C
5
H
9
4
formation and (2)
H
9
CORh(CO)
)Mo(CO) H in a different manner,
one experiment was started with only Rh (CO)12 and then at
H was injected. This experiment
4
. To confirm the
5
H
9
4
5
5
unusual effect of (η -C
5
H
5
3
4
12
that the concentration of C
5
H
3
9
CORh(CO)
4
declines after ca. 120
4
5
5
min when (η -C H )Mo(CO) H is present (Exp 5 is an exception
ca. 160 min (η -C
shows that the late addition of (η -C
4
in an increased yield of the C CORh(CO) .
5 5
H
)Mo(CO)
3
5
5
5
due to Mo injection at ca. 160 min).
The time-dependent concentrations of (η -C
are shown in Figure 3. As seen in this figure, the concentrations
5 5 3
H )Mo(CO) H also results
5
H
9
5
5 3
H )Mo(CO) H
5
The observations seen in Figure 2 can be further quantified.
The initial rate of formation and maximum yields of
5
of (η -C
5 5 3
H )Mo(CO) H are almost constant, although a small
decreasing trend is also apparent for the mixed Rh-Mo
C
5
H
9
CORh(CO)
increase for both the rate of formation as well as the yield of
CORh(CO) as a function of increasing concentration of
η -C )Mo(CO) H.
4
are presented in Table 2. There is a consistent
5
5 5 3
experiments (Exp 2-5). Since the complex η -C H )Mo(CO) H
appears stable under hydroformylation conditions (Exp 6), the
C
5
H
9
4
5
decreasing trends exhibited in Figure 3 appear to be due to a
(
5
H
5
3
5
change in the absorptivity of (η -C
5
H
5
)Mo(CO)
3
H (vide infra).
The time-dependent concentrations for the product cyclo-
pentane carboxaldehyde are shown in Figure 4. It is clear from
(
(
19) Garland, M.; Bor, G. Inorg. Chem. 1989, 28 (3), 410–413.
20) Chini, P. Chem. Comm. 1967, 440–441.
4
592 J. AM. CHEM. SOC. 9 VOL. 132, NO. 13, 2010

Concurrent Synergism and Inhibition in Bimetallic Catalysis
A R T I C L E S
hydroformylation data (Exp 1-6, and latter part of Exp 7). In
particular, BTEM analysis was performed separately on (a)
5
the pure (η -C
5
H
5
)Mo(CO)
3
H experiment, (b) the first hour,
(
c) the second hour, (d) the third hour, and (e) the fourth
hour of the mixed metal Rh-Mo hydroformylations. The
5
resulting pure component spectral estimates of (η -
C
5
H
5
3
)Mo(CO) H for the carbonyl region are shown in Figure
6
. There is a change in both the position and the shape of the
carbonyl bands. The CO vibrations are shifted to lower
wavenumbers. In addition, analysis of different hydroformylation
runs provided dramatically different pure component spectral
5
estimates for (η -C
5
H
5
)Mo(CO)
3
H (Figure 6, inset). Together,
these observations strongly suggest that some sort of solute-solute
interaction is occurring. Since the carbonyl vibrational positions
systematically change as a function of reaction time as well as
5
Mo-Rh loading, molecular recognition between (η -
Figure 5. Decrease of TOF with increasing Mo/Rh ratio for the hydro-
formylation experiments. The error bars correspond to the upper and lower
bound for each experiment.
C
5
H
5
)Mo(CO)
3 5 9 4
H and aldehyde and/or C H CORh(CO) appears
to occur.
To study any spectral changes in the carbonyl vibration of
the aldehyde, spectra from the pure rhodium experiment (Exp
5
this figure that (1) pure (η -C
5
H
5
)Mo(CO)
3
H experiments show
no hydroformylation activity and (2) experiments conducted with
1
(
) were compared to spectra from the Rh/Mo hydroformylations
Exp 2-7). BTEM analysis was performed on both sets to
recover the pure component spectrum of aldehyde. The results
5
Rh
4
5 5 3
(CO)12 and increasing (η -C H )Mo(CO) H concentrations
result in increased rates of aldehyde formation. For example,
at 180 min, the concentrations of aldehyde were 0.0048 mol
fraction (Exp 1), 0.0059 mol fraction (Exp 2), 0.0063 mol
fraction (Exp 3), and 0.0072 mol fraction (Exp 4).
-
1
are shown in Figure 7 where a ca. 2 cm shift to lower
wavenumbers is observed. Since the main difference is the
presence or absence of Mo, the spectral change is consistent
Analysis of Turnover Frequency. Although the addition of
5
5
with an interaction between (η -C
5 5 3
H )Mo(CO) H and aldehyde.
(
η -C
5
H
5
)Mo(CO)
3
H to the unmodified rhodium catalyzed
hydroformylation of cyclopentene increased the rate of formation
and yield of C CORh(CO) , it did not increase the turnover
frequency of the system. Figure 5 shows the average turnover
frequencies for the 5 experiments containing rhodium (see
Experimental Section for details), as well as the highest and
lowest values. There is a clear and quite significant decrease of
BTEM analysis was also performed separately on Exp 1-5
to obtain the pure component spectra of C CORh(CO) at
different Mo loadings. The BTEM estimates of the acyl
H
9
4
5
H
9
4
5
-
1
vibration, with a peak maximum at 1696 cm , are shown in
Figure 8. The spectral estimate of C CORh(CO) obtained
from Exp 1 (Rh only) differs dramatically from the spectral
estimates of C CORh(CO)
5
H
9
4
TOF with increasing Mo/Rh ratio. This result suggests that the
5
H
9
4
obtained from Exp 2-5 (both
5
presence of (η -C
5
H
5
)Mo(CO)
3
H is interfering with the hydro-
Rh and Mo present). In addition, BTEM analysis was reapplied
genolysis of C
5
H
9
CORh(CO)
4
.
to the combined data set from Exp 2-5, and a new acyl
5
-1
Spectral Changes for (η -C
5
H
5
)Mo(CO)
3
H. BTEM was reap-
vibration is recovered with peak maximum at 1684 cm , Figure
plied to the molybdenium carbonyl range of the catalytic
8. This result suggests that a solute-solute interaction is
5
Figure 6. BTEM spectral estimates of (η -C
5
H
5
)Mo(CO)
3
H obtained from 5 different spectroscopic data sets in a time series analysis of one hydroformylation
-
1
experiment. There is a continuous shift to lower wavenumbers for both the bands at 2030 and 1947 cm at increased time (also increased aldehyde
concentration). (Inset) Large changes in BTEM spectral estimates observed for (η -C
5
5 5 3
H )Mo(CO) H between different experimental runs.
J. AM. CHEM. SOC. 9 VOL. 132, NO. 13, 2010 4593

A R T I C L E S
Li et al.
Table 3. Experimentally Determined Homotactic (νii) and
Heterotactic (νij) Volumetric Interaction Coefficients for the Solutes
5
(
η -C
5 5 3
H )Mo(CO) H (1) and Cyclopentane Carboxaldehyde (2) in
n-Hexane under 0.1 MPa Argon at 298 K
solute-i/solute-j
volumetric interaction coefficient, νij
P value
1/1
2/2
1/2
105.2 ( 1238.5
-62.9 ( 19.6
-817.3 ( 351.6
0.9332
0.0046
0.0313
statistical significance. Accordingly, the P-values 0.0046 and
.0313 associated with ν22 and ν12, respectively, confirm that
0
5
5 5 3
aldehyde-aldehyde and aldehyde-(η -C H )Mo(CO) H inter-
actions indeed exist.
1
7
DFT Spectral Predictions for Interactions. DFT calculations
were performed to (1) better understand the changes observed
in the experimental spectra as well as the apparent decrease
5
in the concentration of (η -C
5
H
5
)Mo(CO)
3
H (Figure 3) and
(
2) ultimately better understand the basis for the pronounced
decrease in TOF with increasing molybdenum loadings
Figure 5). Accordingly, optimized geometries and IR spectral
Figure 7. BTEM spectral estimates of aldehyde obtained from catalytic
runs Exp 1 (pure Rh) and Exp 2-7 (Rh and Mo).
(
5
predictions were carried out for (η -C
5 5 3
H )Mo(CO) H,
C
5
H
9
CORh(CO) and cyclopentane carboxaldehyde as well
4
as a variety of different hydrogen bonding scenarios between
these species.
The experimental results presented in Figure 6 show that
significant spectral changes in the terminal CO vibrations of
5
(
η -C
5 5 3
H )Mo(CO) H occur. Furthermore, these changes correlate
with reaction time and hence increasing aldehyde concentration.
5
Accordingly, DFT calculations were performed on (η -
5
C
5
H
5
)Mo(CO)
3
H and two different (η -C
5
H
5
)Mo(CO)
3
H-alde-
hyde complexes. One of the complexes involved hydrogen
bonding of the type Mo-H· · ·OdC and one involved hydrogen
5 4
bonding through the Cp ring, that is, C H -H· · ·OdC (Please
refer to Supporting Information for details).
Figure 9 shows the optimized geometry as well the spectral
5
prediction for the (η -C
5
H
5
)Mo(CO)
3
H-aldehyde complex where
-H· · ·OdC occurs (bond
hydrogen bonding of the type C
5
H
4
length 2.22 Ångstrom). A ca. 2 wavenumber shift to lower
wavenumbers as well as a decrease in the intensity is seen for
-
1
the bands at both 1947 and 2030 cm and little or no change
-
1
is predicted for the Mo-H vibration at 1820 cm . The first
Figure 8. (a) Normalized BTEM spectral estimates of the acyl vibration
-
1
centered at 1696 cm obtained independently for Exp 1-5. (b) BTEM
two predictions are certainly consistent with the experimental
-1
spectral estimate of the acyl vibration centered at 1684 cm obtained from
the combined data from Exp 2-5. The slightly uneven baseline in the
recovered pure component spectrum (b) is due to the relatively low intensity
of the new signal.
5
occurring, in this case, between (η -C
5
H
5
)Mo(CO)
3
H and
C
5
H
9
CORh(CO)
4
.
Determination of Interaction Volume. Independent experi-
5
ments were conducted on n-hexane solutions of (η -
)Mo(CO) H and cyclopentane carboxaldehyde, using in situ
C H
5 5
3
FTIR spectroscopy and online density measurements (see
Supporting Information). The purpose of these experiments was
5
to determine volumes of interaction ν11, self-association of (η -
C
5
H
5
)Mo(CO)
3
H; ν22, self-association of cyclopentane carbox-
5
aldehyde; and ν12, interaction between (η -C
5 5 3
H )Mo(CO) H and
cyclopentane carboxaldehyde. The results presented in Table 3
5
suggest that (a) there is negligible self-association of (η -
C
5
H
5
3
)Mo(CO) H, (b) there is measurable self-association of
cyclopentane carboxaldehyde, and (c) the volume of interaction
5
5 5 3
between (η -C H )Mo(CO) H and cyclopentane carboxaldehyde
is very substantialsit is an order of magnitude greater than the
self-association of aldehyde. Moreover, the P-value tests (and
specifically P-values below 0.05) provide confirmation of
5
Figure 9. DFT spectral predictions for (η -C
5
H
5
)Mo(CO)
3
H (dashed line)
and its hydrogen bonded complex (solid line) with aldehyde. The optimized
geometry for hydrogen bonding through the Cp ring is also provided.
4
594 J. AM. CHEM. SOC. 9 VOL. 132, NO. 13, 2010

Concurrent Synergism and Inhibition in Bimetallic Catalysis
A R T I C L E S
Figure 11. Relative concentration profiles of the primary solutes from Exp
8
in the stoichiometric hydroformylation of cyclopentene starting with the
Figure 10. DFT spectral predictions for C
5
H
9
CORh(CO)
4
and its hydrogen
5
5
4 5 5 3
complexes Rh (CO)12 and (η -C H )Mo(CO) H. (Inset) Window clarification
that alkene additions were performed at ca. 55, 60, and 70 min. Hydrogen
was added to the system at 280 min to initiate the catalytic hydroformylation.
bonded complex with (η -C H )Mo(CO) H. The optimized geometry for
5 5 3
hydrogen bonding through the Cp ring is also provided.
observations in Figure 6 (the third prediction is quite difficult
to confirm due to the very weak experimental absorptivity of
the Mo-H vibration).
5 9 4
C H CORh(CO) (see Figure 1) and (2) that the experimental
Also, the DFT calculations indicate that the complex involv-
ing hydrogen bonding through the Cp ring provided the lower
free energy (see Supporting Information). In addition, hydrogen
bonding through Mo-H· · ·OdC is predicted to split the band
at 2030 cm- significantly, and this is not seen experimentally.
The experimental results presented in Figure 7 showed that
and DFT predicted spectral shifts for the Mo-CO and acyl
moiety are consistent, the new species X is assigned to the weak
5
hydrogen bonded complex (η -C
Rh(CO)
H· · ·OdC occurs.
5
H
5
)Mo(CO)
3
5 9
H-C H CO-
4 5 4
where hydrogen bonding of the type C H -
1
The complex involving hydrogen bonding through the Cp
ring provided the lowest free energy in the DFT calculations.
The DFT prediction for the complex with bonding through
spectral changes in the CO vibration of cyclopentane carbox-
5
aldehyde occurred with increasing (η -C
5 5 3
H )Mo(CO) H loading.
-
1
A shift to lower wavenumbers for CdO was predicted by DFT
Mo-H· · ·OdC shows a split for the band at 2030 cm (see
for both of the two hydrogen bonding scenarios considered
Supporting Information for both issues). Therefore, hydrogen
5
between aldehyde and (η -C
5 5 3
H )Mo(CO) H. The shift to a lower
bonding through the Cp ring to C
the primary scenario.
5 9 4
H CORh(CO) appears to be
wavenumber is consistent with the decrease in CdO bond order
upon hydrogen bonding.
The experimental results presented in Figure 8 showed a marked
The DFT predictions shown in Figures 9 and 10 both indicate
-
1
a decrease in the absorptivities at 1947 and 2030 cm , when
5
spectral change in the CdO vibration of C
5
H
9
CORh(CO)
H was present in solution. Accordingly, DFT
calculations were performed on C CORh(CO) and two
CORh(CO) complexes.
4
when
hydrogen bonding is occurring between (η -C
and aldehyde or C CORh(CO) . Consequently, a least-squares
fit of the reference spectra for (η -C
5 5 3
H )Mo(CO) H
5
(
η -C
5
H
5
)Mo(CO)
3
H
5 9
4
5
5
H
9
9
4
5 5 3
H )Mo(CO) H onto the
5
different (η -C
5
H
5
)Mo(CO)
3
H-C
5
H
4
reaction spectra will not be as precise when more and more
One of the complexes involved hydrogen bonding through
hydrogen bonding occurs. This mis-match will result in an
5
Mo-H· · ·OdC and one involved hydrogen bonding through
apparent decrease in the (η -C
5 5 3
H )Mo(CO) H concentration with
the Cp ring, that is, C
Information for details).
Figure 10 shows the DFT predicted spectra of (η5-
)Mo(CO) H and C CORh(CO) . The former DFT
5 4
H -H· · ·OdC (please refer to Supporting
time. This is precisely the effect that is seen in Figure 3.
Stoichiometric Hydroformylation. Pino et al were apparently
the first to demonstrate the possibility of stoichiometric hydro-
formylation of alkenes, by reacting Co (CO) , H , alkene to
C
5
H
5
3
5
H
9
4
2
8
2
2
1
spectrum almost matches the experimental measurement.
Although the latter DFT spectrum shows very accurate
relative band intensities, there are noticeable wavenumber
shifts from the experimental measurements (see Figure 1).
4 12
produce aldehyde and Co (CO) . Subsequently, Chini et al
showed a similar stoichiometric hydroformylation with rhodium
at 293K, using 3 Rh (CO) + 4 H + 4 alkene to give 4
4
12
2
2
2
aldehyde and 2 Rh (CO)
6
16
.
In both cases, the starting metal
5
In addition, the optimized geometry for the (η -
carbonyls provide the needed CO for the aldehyde product.
The first half of Exp 7 and 8 represents a new variation on
the theme of stoichiometric hydroformylation. In this new
variation, the starting metal hydride provides all the needed
hydrogen for the aldehyde product. Figure 11 shows that the
reaction was very fast upon addition of alkene to the system,
then appeared to equilibrate, as suggested by the sustained and
C
5
H
5
)Mo(CO)
DFT spectrum are shown in Figure 10, where hydrogen bonding
of the type C -H· · ·OdC occurs (bond length 2.32 Ång-
strom). Again, the most intense carbonyl band for the molyb-
3
H-C
5 9 4
H CORh(CO) complex as well as the
5
H
4
-1
denium at 1947 cm is shifted to lower wavenumbers. Also,
the acyl vibration is shifted ca. 12 cm- to lower wavenumbers,
which is in agreement with the experimentally observed
difference in Figure 8. Taking into consideration (1) that the
1
(
(
21) Pino, P.; Ercoli, R.; Calderazzo, F. Chim. Ind. 1955, 37, 782.
22) Chini, P.; Martinengo, S.; Garlaschelli, G. J. Chem. Soc., Chem. Comm.
1972, 709.
BTEM spectral estimate of new species X appears quite similar
5
to a superposition of the spectra of (η -C
5
H
5
)Mo(CO)
3
H and
J. AM. CHEM. SOC. 9 VOL. 132, NO. 13, 2010 4595

A R T I C L E S
Li et al.
Scheme 1. Possible Set of Reactions Consistent with
Stoichiometric Hydroformylation Starting with Rh
4
(CO)12 and
5
(
5 5 3
η -C H )Mo(CO) H
5
(
CO)
3
-Rh(CO)
3
L and (η -C
5
H
5
)Mo(CO)
2
3
L-Rh(CO) L where L
)
(π-C H ); however, none of the DFT-predicted spectra closely
matched the BTEM spectral estimate.
5 8
4
The presence of detectable quantities of RCORh(CO) in the
stoichiometric hydroformylation indicates that the formation of
a rhodium hydride occurred. One possible set of stochiometric
reactions, which would be consistent with both the formation
of a rhodium hydride and the formation RCORh(CO)
4
is given
in Scheme 1, where the apparent overall reaction is Rh
4
(CO)12
5
+
4 (η -C
and 4 RhMo(CO)
The stochiometric hydroformylation results also help to clarify
the issue of increased rate and yield of RCORh(CO) , which
5
H
5
)Mo(CO)
3
H + 6 CO + 2 alkene f 2 aldehyde
5
7
(η -C ).
5 5
H
Figure 12. Comparison of the BTEM spectral estimate and DFT-predicted
5
spectrum of (η -C
5
H
5
)Mo(CO)
3
-Rh(CO)
4
.
4
was presented earlier in Figure 2 and Table 2 for the catalytic
experiments. Indeed, eqs 1 and 2 of Scheme 1 represent a
simultaneous presence of the metal carbonyls Rh
)Mo(CO) H, RCORh(CO) , and new Species Y and Z, as
4
(CO)12, (η5-
C
5
H
5
3
4
plausible route for the formation of RCORh(CO)
5 5 3 4
the attack of (η -C H )Mo(CO) H on Rh (CO)12. The issue of
4
arising from
well as aldehyde. In Exp 8, multiple perturbations of alkene
were introduced, and the species Z showed an immediate and
pronounced increase in concentration upon each addition of
cyclopentene to the system, and a new equilibrium state was
rapidly established as seen by the step-function increases in
aldehyde concentration.
Figure 11 also shows that upon addition of molecular
hydrogen to the system at ca. 280 min, catalytic hydroformy-
lation commenced very quickly and the Species Y and Z very
5
metal hydride assisted conversion of metal carbonyl clusters
9d-h
has been discussed for other cases elsewhere.
Catalytic Binuclear Elimination. Since the stoichiometric
hydroformylation resulted in the formation of the heterobime-
5
5
tallic complex (η -C )Mo(CO)
5
H
5
3 4
-Rh(CO) , and since (η -
C H
5
5
)Mo(CO) -Rh(CO) readily activates molecular hydrogen,
3
4
a catalytic binuclear elimination reaction is probably occurring
in the catalytic system. To confirm this possibility, a deutero-
formylation (Exp 9) was performed. For the initial 320 min,
the pure rhodium deutero-formylation resulted in d-aldehyde
rapidly disappeared, giving rise to increased concentrations of
5
both Rh
4
(CO)12 and (η -C
5
H
5
)Mo(CO)
3
H. This indicates that
the species Y and Z can rapidly activate molecular hydrogen
under high CO pressure and that Y and Z are heterobimetallic.
Given the precedence of other heterobimetallic dinuclear
complexes to readily activate molecular hydrogen under high
-
1
CDOCDHCH
this experiment, using the more reactive 3,3-dimethylbut-1-ene
as substrate, ca. 90% conversion of Rh (CO)12 to RCORh(CO)
)Mo(CO) H was then
2 3 3
C(CH ) with a CO vibration at 1723 cm . In
4
4
5
occurred in these initial 320 min. (η -C
5
H
5
3
2
3c
23a,b
CO pressures, particular CoMo(CO)
RhRe(CO) ,
9
7
Cp, CoRh(CO)
7
and
added to the system. The new FTIR spectra taken immediately
thereafter indicated the presence of 2 CO vibrations centered
9
g
DFT calculations were performed on numerous
possible heterobimetallic Rh-Mo complexes to identify species
Y and Z. In particular, very good agreement was obtained
-1
at 1723 and 1733 cm . This observation confirms the formation
5
5 5 3
of RCHO and the role of (η -C H )Mo(CO) H as the hydrogen
between the BTEM spectral estimate of species Y and the DFT
source.
5
predicted spectrum of (η -C
5
H
5
)Mo(CO)
3
-Rh(CO)
4
(Figure 12
In addition, mass balances were performed. At 320 min, the
and Supporting Information). It can be mentioned that this
-
5
amount of RCORh(CO)
while the Rh
(CO)12 was ca. 2.0 × 10 moles. Upon addition
)Mo(CO) H, the product RCHO was formed at a
rather high rate. The total amount of RCHO produced was ca.
4
present was ca. 9.7 × 10 moles,
5
species has a spectrum similar to the known species (η -
-6
4
C
(
5
H
5
)Mo(CO)
η -C )Mo(CO)
structure of (η -RC
3
-Co(CO)
-Rh(CO)
3 4
)Mo(CO) -Co(CO) .
4
and that the DFT-predicted geometry
5
of (η -C
5
H
5
3
5
5
H
5
3
4
is similar to the determined crystal
5
23c,d
5
H
4
-
4
-4
1
.5 × 10 moles, in close agreement with the 1.7 × 10 moles
The BTEM spectral estimate of species Z indicates that (i)
there are no bridging CO and (ii) the symmetry at Mo has been
dramatically changed as evidenced by the shift in the highest
5
of (η -C
5
H
5
)Mo(CO)
3
H added at 320 min. In Table 4, three
synthetic pathways are considered for the formation of aldehyde
5
from (η -C
the moles of RCHO formed and the moles of (η -
)Mo(CO) H introduced confirms the presence of catalytic
5 5 3
H )Mo(CO) H. The close correspondence between
-1
C-O vibration from 2030 to 1989 cm . In previous studies of
5
2
3c
5
the Mo-Co complexes, both (η -C
5
H
5
)Mo(CO)
L where L ) solvent, P(n-
have been identified. The former species has its highest
3 3
-Co(CO) L
C
5
H
5
3
5
and (η -C
Bu)
5 5 2 3
H )Mo(CO) L-Co(CO)
binuclear elimination in this bimetallic system. However, the
initial rate of RCHO formation versus the rate of RCDO
formation suggests that only ca. 10% of the aldehyde formation
occurs via a CBER mechanism.
3
-
1
C-O vibration at 2024 cm , and the latter has its highest C-O
vibration at 1993 or 1990 cm- depending on the isomer
1
considered. Since the species Z appears to be in equilibrium
5
exchange with RhMo(CO)
7
(η -C
5
H
5
) and alkene (see Figure 11),
Discussion
5
we tentatively assign the stoichiometry RhMo(CO)7-y(η -C
where y ) 1 or 2, and L ) (π-C ) to species Z. Numerous
DFT calculations were attempted for isomers of (η -C
5 5 y
H )L
5
H
8
General. The in situ measured solute concentrations provided
a number of important observations. For example, the rate of
5
5
H
5
)Mo-
4
596 J. AM. CHEM. SOC. 9 VOL. 132, NO. 13, 2010

Concurrent Synergism and Inhibition in Bimetallic Catalysis
A R T I C L E S
5
Table 4. Possible Synthetic Pathways for the Formation of Aldehyde from (η -C
5
5 3
H )Mo(CO) H in the Catalytic Deutero-Formylation (Exp 9)
type of aldehyde
produced
moles RCHO produced
5
reaction
5 5 3
per (η -C H )Mo(CO) H
CHOCH
2
CH
2
C(CH
3
)
)
3
3
1/2
5
1
2
. Rh (CO) + 4(η -C H )Mo(CO) H + 6 CO + 2 alkene f
4
12
5
5
3
5
aldehyde and 4RhMo(CO) (η -C H ).
7
5
5
(
Scheme 1)
CDOCH
2
CH
2
C(CH
3
0
5
2
2
a. Rh (CO) + 2(η -C H )Mo(CO) H + 6CO + 2 alkene f
4 12 5 5 3
5
(CH ) CCH CH CORh(CO) +2 RhMo(CO) (η -C H )
3
3
2
2
4
7
5
5
2
2
b. 2(CH ) CCH CH CORh(CO) + 2D f
3 3 2 2 4 2
aldehyde + 2 DRh(CO)4
2 3
CHOCDHCH C(CH )
3
1
5
3
. (CH ) CCH CDHCORh(CO) + (η -C H )Mo(CO) H f
3 3 2 4 5 5 3
5
aldehyde + RhMo(CO) (η -C H )
7
5
5
formation and yield of C
rate of aldehyde formation as a function of molybdenum loading
both indicate a positive effect of (η -C
5
H
9
CORh(CO)
4
as well as the increased
inhibitory effects due to hydrogen bonding and synergistic
effects, that is, (η -C H )Mo(CO) H assisted cluster fragmenta-
5 5 3
tion and catalytic binuclear elimination.
5
5
5
H
5
)Mo(CO)
3
H on the
rhodium catalyzed hydroformylation of cyclopentene. However,
Crystallographic Support. Independent support for an inter-
action of the type C H -H· · ·OdC can be obtained from
the decreasing turnover frequencies with increasing molybdenum
5
4
5
loading strongly suggested that (η -C
5
H
5
)Mo(CO)
3
H plays an
crystallographic databases. A search of the Cambridge Structural
Database (CSD) revealed numerous crystal structures where the
inhibitory role and reduces the efficiency of the catalytic cycle.
5
In addition, the decrease in C
5
H
9
CORh(CO)
)Mo(CO)
4
concentrations at
H is present, but
H---O interatomic distance (between a η -C H moiety and a
5
5
5
long reaction times when (η -C
not when Rh
5
H
5
3
CdO) is in a range consistent with some orbital overlap and
hydrogen bonding. Table S-8 (Supporting Information) lists 29
structures where a distance of less than 2.40 Å occurs between
H on Cp and the O on an organic carbonyl. In addition, Table
S-9 (Supporting Information) lists 13 structures where a distance
of less than 2.40 Å occurs between H on Cp and the O on an
organic carbonyl moiety in an organometallic compound. It can
be noted that there are many other similar structures in the CSD
with somewhat longer H---O distances as well. This solid-state
database search further supports the conclusions reached in the
present liquid-phase catalytic experiments.
4
(CO)12 is used alone, is an observation that
warrants further explanation.
The changes observed for the in situ measured band positions
5
and band shapes for (η -C
5 5 3 5 9 4
H )Mo(CO) H, C H CORh(CO) , and
cyclopentane carboxaldehyde suggested the existence of
5
5
(
C
η -C
)Mo(CO)
during the catalysis. The spectroscopic changes in the acyl
vibration of C CORh(CO) were the most significant (Figure
). Independent physicochemical measurements on solutions of
5
H
5
)Mo(CO)
3
H-cyclopentane carboxaldehyde and (η -
5
H
5
3
H-C
H
5 9
CORh(CO) solute-solute interactions
4
H
5 9
4
8
(
5
η -C
5
H
5
)Mo(CO)
3
H and cyclopentane carboxaldehyde con-
)Mo(CO) H-cyclopentane carboxaldehyde
It should be noted that crystallographic data has been
extensively used to support hydrogen bonding hypotheses. Of
particular relevance is the work on organometallic systems, and
especially hydrogen bonding of the type M-H---OdC and
5
firmed that (η -C
5
H
5
3
interactions are quite significant.
The origin of the solute-solute interactions was further
examined using DFT. It was concluded that interaction through
the Cp ring on molybdenum to the organic carbonyl groups of
aldehyde and C H CORh(CO) are the most likely interactions.
5 9 4
This does not negate the possibility of some hydrogen bonding
C
5
H -H· · ·OdC in metallocene systems where the CdO is a
4
2
5
terminal or bridged metal carbonyl. It has been noted
elsewhere that, in many cases, hydrogen bonding of the type
C H -H· · ·OdC is the predominate scenario in the solid state.
5
26
4
of the type Mo-H· · ·OdC; however, hydrogen bonding of the
5
(
η -C
5 5 3 5 9 4
H )Mo(CO) H and C H CORh(CO) Interactions and
5 4
type C H -H · · · OdC is consistent with all of the available
TOF. It was concluded that hydrogen bonding of the type
experimental results. Given the widespread use of metallocenes
5
C
5
H
4
5 5 3 5 9
-H· · ·OdC between (η -C H )Mo(CO) H and C H CORh-
2
4
in organic syntheses, the present results suggest that similar
Cp hydrogen-bonding interactions might be influencing selectiv-
ity and activity patterns in other catalytic systems.
(
23) (a) Horvath, I.; Garland, M.; Bor, G.; Pino, P. J. Organomet. Chem.
988, 358, C17–C22. (b) Garland, M.; Pino, P. Organometallics 1990,
1
The stoichiometric hydroformylations allowed the observation
9, 1943–1949. (c) Kovacs, I.; Sisak, A.; Ungvary, F.; Marko, L.
Organometallics 1989, 8, 1873–1877. (e) Song, X.; Brown, T. Inorg.
Chem. 1995, 34, 3220–3231. (d) Song, L.; Shen, J.; Wang, J.; Hu,
Q.; Wang, R.; Wang, H. Polyhedron 1994, 13, 3235–3241.
5
of two further heterobimetallic species, namely, (η -C
5
H
5
)Mo-
(
3 4
CO) -Rh(CO) and species Z. In addition, it was found that
these species can rapidly activate molecular hydrogen under
significant CO partial pressure. The existence of the former
species, together with the labeled hydroformylation experiment
Exp 9), confirms that catalytic binuclear elimination is occurring
in this Rh-Mo system; however, this contribution is not the
major synthetic route for aldehyde formation.
(24) (a) Marks, T. J. Chem. ReV. 2000, 100 (4), 1391–1434. (b) Metal-
locenes in Regio- and StereoselectiVe Synthesis; Takahasi, T., Ed.;
Springer: New York, 2005.
(
25) (a) Braga, D.; Grepioni, F.; Biradha, K.; Pedireddi, V. R.; Desiraju,
G. R. J. Am. Chem. Soc. 1995, 117 (11), 3156–66. (b) Braga, D.;
Grepioni, F. Acc. Chem. Res. 1997, 30 (2), 81–87. (c) Choi, J. C.;
Pulling, M. E.; Smith, D. M.; Norton, R. J. J. Am. Chem. Soc. 2008,
(
1
30, 4250–4252.
Therefore, in summary, the present heterobimetallic hydro-
formylation is quite complex, since it involves concurrent
(
26) Braga, D.; Grepioni, F.; Tedesco, E.; Biradha, K.; Desiraju, G. R.
Organometallics 1996, 15 (12), 2692–2699.
J. AM. CHEM. SOC. 9 VOL. 132, NO. 13, 2010 4597

A R T I C L E S
Li et al.
Scheme 2. Proposed Reaction Topology for Simultaneous (1) Rhodium-Catalyzed Hydroformylation, (2) Inhibition of the Rhodium Catalysis
5
a
Due to the Presence of (η -C
5
H )Mo(CO)
5
3
H-C
5
9 4
H CORh(CO) , and (3) Rh-Mo CBER in Rh-Mo Bimetallic Catalysis
a
Blue highlighted species are observable organometallics in the system.
(
CO)
bonded complex (η -C
catalysis has a number of consequences. First, a large portion of
space around the rhodium atom in C CORh(CO) is effectively
4
exists. The existence of a weak heterobimetallic hydrogen-
Representation of Catalytic System. The results of the
catalytic and stoichiometric experiments suggest that the present
bimetallic catalytic system consists of (1) a classic rhodium-
catalyzed hydroformation, (2) a rhodium-catalyzed hydroformy-
5
5 5 3 5 9 4
H )Mo(CO) H-C H CORh(CO) during
5 9
H
4
shielded. Second, the formation of the complex should (a) interfere
with effective dissociation of a terminal CO from Rh, (b) hinder
the approach of molecular hydrogen to the resulting coordinatively
lation that is inhibited by the presence of hydrogen bonding
5
between RCORh(CO)
4
and (η -C
5
H
5
)Mo(CO)
3
H and (3) a
catalytic binuclear elimination reaction involving the slow attack
5
unsaturated {C
energies for oxidative addition of H
5
H
9
CORh(CO)
3
}, and possibly (c) alter the activation
to rhodium as well as the
of (η -C
bimetallic complex and aldehyde. The heterobimetallic complex
then activates molecular hydrogen to reform HRh(CO) (hence
H. A representation of
5 5 3 4
H )Mo(CO) H on RCORh(CO) resulting in a hetero-
2
subsequent hydrogenolysis of the Rh-acyl bond. The first two
situations, a and b, would certainly result in a marked decrease in
the experimental TOF (as observed in Figure 5) for the catalytic
4
5
RCORh(CO)
4
5 5 3
) and (η -C H )Mo(CO)
these combined effects is shown in Scheme 2.
cycle. The main experimental conclusion is that the solute-solute
In Scheme 2, there are two mechanisms for aldehyde
formation (the oxidation addition of hydrogen to RCORh(CO)
3
5
interaction between (η -C
5
H
5
)Mo(CO)
3
H and C
5
H
9
CORh(CO)
4
5
results in the formation of the hydrogen bonded complex (η -
)Mo(CO) H-C CORh(CO) which in turn results in a
decrease in the rate of hydrogen activation and hence a decrease
and the binuclear elimination step ꢁ). However, closer inspection
5
C H
5 5
3
5 9
H
suggests that the heterobimetallic hydrogen-bonded complex (η -
C
5
5 3 5 9 4
H )Mo(CO) H-C H CORh(CO) itself may also be involved
for TOF in the system.
in aldehyde formation. For example, the hydrogen-bonded
complex may undergo CO dissociation followed by intramo-
lecular hydrogen transfer to cleave the Rh-acyl bond (Scheme
3). Alternatively, the hydrogen-bonded complex may undergo
CO dissociation followed by oxidative addition of molecular
hydrogen (Scheme 4). From an experimental viewpoint, the ꢁ
step and Scheme 3 are quite similar, since both begin with the
individual mononuclear complexes and result in the heterobi-
metallic binuclear complex (the only difference being the lack
Decreasing C
Reaction Time. At reaction times greater than ca. 120 min, there
is a small decrease in the concentrations of C CORh(CO)
5 9 4
H CORh(CO) Concentrations at Longer
5
H
9
4
in the bimetallic experiments (Exp 2-4). A relatively straight-
forward explanation is provided by the increasing aldehyde
concentrations. As the aldehyde concentration increases
during the course of a hydroformylation experiment, a small
5
net increase in the concentration of a weak (η -C
5
H
5
)-
Mo(CO)
3
H-aldehyde complex will result. When this occurs,
of hydrogen bonding and presence of hydrogen bonding,
5
5
5
slightly less free (η -C
5
H
5
)Mo(CO)
H-assisted transformation of Rh
CORh(CO) . Consequently, the yield of C CORh(CO)
3
H is available for the (η -
respectively). Since a least-squares fit of the (η -C
5
H
5
)-
5
C
C
5
H
H
5
)Mo(CO)
3
4
(CO)12 to
Mo(CO)
)Mo(CO)
contribute little to aldehyde formation if they exist. Further-
3
H spectral region suggests only ca. 10% of the (η -
5
9
4
5
H
9
4
C H
5 5
3
H is hydrogen bonded, Schemes 3 and 4 might
slowly declines with increasing aldehyde concentrations.
4
598 J. AM. CHEM. SOC. 9 VOL. 132, NO. 13, 2010

Concurrent Synergism and Inhibition in Bimetallic Catalysis
A R T I C L E S
Scheme 3. Possible Intramolecular Route for Aldehyde Formation from the Hydrogen-Bonded Heterobimetallic Complex
Scheme 4. Possible Hydrogen Addition Route for Aldehyde Formation from the Hydrogen-Bonded Heterobimetallic Complex
more, Figure 5 clearly shows a net decrease in TOF with
with supporting thermo-physical measurements and first prin-
increased Mo loading (increased hydrogen bonding), and
cipal calculations might allow considerably greater insight into
subtle but important molecular interactions affecting regio-,
chemo-, and stereoselective syntheses.
5
therefore, the net effect of the presence of (η -C
5
H
5
)-
Mo(CO)
3
5 9 4
H-C H CORh(CO) is inhibitory.
5
A final remark with respect to the thermochemistry of (η -
C H
5
5
)Mo(CO)
3
H-C
5
H
9
CORh(CO)
4
is useful. Since the experi-
Conclusion
mentally obtained equilibrium conversion was ca. 10%, the
experimental Gibbs free energy change corresponds to ap-
proximately 4.0 kcal/mol. In contrast, the DFT-predicted free
energy changes were slightly greater than 0 kcal/mol (corre-
sponding to ca. 0.01% conversion). This discrepancy appears
to be related to the known accuracy/sensitivity of DFT ther-
mochemical calculations, particularly those involving transition
The Rh-Mo catalyzed hydroformylation of cyclopentene
exhibits a number of interesting and unusual spectroscopic and
kinetic effects. A combination of spectroscopic and kinetic
measurements as well as DFT calculations and volume of
interaction determinations were used to identify the primary
mechanistic reasons behind these effects. In summary, two main
mechanistic routes exist for aldehyde formation, and concurrent
synergistic and inhibitory effects exist in this heterobimetallic
system. Catalytic binuclear elimination is one of the primary
2
7
metal carbonyls.
Detection of Solute-Solute Interactions During Catalysiss
A Possible Future Direction. In the past decade, experimental
progress has been made with both 2D Raman spectroscopy and
synergistic effects, and the heterobimetallic complex RhMo-
5
(
CO)
7
(η -C
5
H
5
) was shown to rapidly active molecular hydro-
2
D infrared spectroscopy (vibrational equivalents of 2D
5
2
8
gen. The formation of a heterobimetallic hydrogen-bonded (η -
NMR). In particular, there has been keen interest in the use
of these new techniques to directly identify weak and brief intra-
and intermolecular interactions, even those involved in solvation.
Such interactions are detectable due to the vibrational energy
transfer occurring between the moieties involved. Since vibra-
tional spectroscopies are (1) generally very sensitive, with
relatively short measurement time, (2) not restricted to the
presence/use of any particular isotopes, and therefore applicable
to all atom-combinations/moieties present, and (3) normally
readily applied to reactive systems, 2D vibration techniques
together with signal processing might provide a unique op-
portunity for identifying subtle molecular interactions in ho-
mogeneous catalytic systems. Finally, the combined application
of such new and sophisticated in situ spectroscopic techniques,
C
5
H
5
)Mo(CO)
3
H-C
5
H
9
CORh(CO)
4
complex, where hydrogen
bonding of the type C
5
H
4
-H· · ·OdC occurs, is the primary
reason for inhibition and decreased TOFs. The formation of this
weak hydrogen-bonded complex decreases the effective rate of
hydrogen activation on C H CORh(CO) . Hence, the present
5 9 4
study shows that quite subtle solute-solute interactions can be
indentified during homogeneous catalysis. In addition, the
present study also suggests it may be necessary to consider
competing effects (synergistic as well as inhibitory) when
analyzing the results from complex bimetallic catalytic systems.
Acknowledgment. This work was supported by the Science and
Engineering Research Council of A*STAR (Agency for Science,
Technology and Research), Singapore. We thank reviewers for their
suggestions and thank Professor E. S. Shubina for clarifying the
basis sets used in their DFT calculations (ref 18).
(
27) Proynov, E.; Chermete, H.; Salahub, D. R. J. Chem. Phy. 2000, 113,
0013–10027.
28) (a) Mukamei, S.; Piryatinski, A.; Chernyak, V. Acc. Chem. Res. 1999,
2, 145–154. (b) Tokmakoff, A.; Lang, M. J.; Larsen, D. S.; Fleming,
1
(
Supporting Information Available: Detailed experimental
setup; detailed volumetric measurement results; detailed DFT
prediction results; detailed Cambridge Structural Database
3
G. R. Phys. ReV. Lett. 1997, 79, 2702–2705. (c) Hahn, S.; Kwak, K.;
Cho, M. J. Chem. Phy. 2000, 112, 4553–4557. Nagata, Y.; Tanimura,
Y. J. Chem. Phys. 2006, 124, 024508-1–024508-9. (d) Wright, J. C.;
Zhao, W.; Murdoch, K. M.; Besemann, D. M.; Condon, N. J.;
Meyer, K. A. In Handbook of Vibrational Spectroscopy; Chambers,
J. M., Griffiths, P. R., Eds.; Wiley: New York, 2002; Vol. 1, pp
(
CSD) analysis; complete ref 8b. This material is available free
of charge via the Internet at http://pubs.acs.org.
8
53-865.
JA909696B
J. AM. CHEM. SOC. 9 VOL. 132, NO. 13, 2010 4599