Rh4(CO)12-catalyzed hydroformylation of cyclopentene promoted with HMn(CO)5. Another example of Rh4(CO) 12/HMn(CO)5 bimetallic catalytic binuclear elimination

Article abstract of DOI:10.1021/om0342661

A detailed in situ high-pressure FTIR spectroscopic study was performed to investigate the bimetallic origins of catalytic synergism in homogeneous catalysis. The reaction chosen was the homogeneous catalyzed hydroformylation of cyclopentene to cyclopentanecarbox-aldehyde, starting with unmodified rhodium and manganese carbonyls as catalyst precursors in n-hexane as solvent. The spectra were analyzed by an advanced signal processing and statistical technique. Only four organometallic spectra could be found. These were RCORh-(CO)₄, Rh₄(CO)₁₂, HMn(CO)₅, and Mn₂(CO)₁₀. A very significant increase in aldehyde formation was observed in the experiments when both rhodium carbonyl and manganese carbonyl complexes were used simultaneously. The kinetics of product formation shows a distinct linear-bilinear form in observable organometallics; k₁[RCORh(CO)₄][H₂][CO]^-1 + k2[RCORh(CO)₄] [HMn(CO)₅] [CO]^-1.6. The determined activation parameters for the bilinear term were ΔH ^? = 47 ± 8 kJ/mol and ΔS^? = -88 ± 28 J/(mol K). The results, particularly the bilinear term and the negative entropy of activation, suggest that the origin of synergism is the presence of bimetallic catalytic binuclear elimination: namely, the hydride attack on the acyl species. The present results are consistent with observations and conclusions reached for a previously studied Rh₄(CO) ₁₂/HMn(CO)₅/3,3-dimethylbut-1-ene hydroformylation system.

Full text of DOI:10.1021/om0342661

Organometallics 2004, 23, 4131-4138
4131
Rh (CO) -Ca ta lyzed Hyd r ofor m yla tion of Cyclop en ten e
4
12
P r om oted w ith HMn (CO) . An oth er Exa m p le of
5
Rh (CO) /HMn (CO) Bim eta llic Ca ta lytic Bin u clea r
4
12
5
Elim in a tion
Chuanzhao Li, Effendi Widjaja, and Marc Garland*
Department of Chemical and Biomolecular Engineering, 4 Engineering Drive 4,
National University of Singapore, Singapore 119260
Received October 29, 2003
A detailed in situ high-pressure FTIR spectroscopic study was performed to investigate
the bimetallic origins of catalytic synergism in homogeneous catalysis. The reaction chosen
was the homogeneous catalyzed hydroformylation of cyclopentene to cyclopentanecarbox-
aldehyde, starting with unmodified rhodium and manganese carbonyls as catalyst precursors
in n-hexane as solvent. The spectra were analyzed by an advanced signal processing and
statistical technique. Only four organometallic spectra could be found. These were RCORh-
4 4 5 2
(CO) , Rh (CO)12, HMn(CO) , and Mn (CO)10. A very significant increase in aldehyde
formation was observed in the experiments when both rhodium carbonyl and manganese
carbonyl complexes were used simultaneously. The kinetics of product formation shows a
-
1
distinct linear-bilinear form in observable organometallics; k
1
[RCORh(CO)
4
][H
2
][CO]
+
-1.6
k [RCORh(CO) ][HMn(CO) ][CO] . The determined activation parameters for the bilinear
2 4 5
q
q
term were ∆H ) 47 ( 8 kJ /mol and ∆S ) -88 ( 28 J /(mol K). The results, particularly the
bilinear term and the negative entropy of activation, suggest that the origin of synergism is
the presence of bimetallic catalytic binuclear elimination: namely, the hydride attack on
the acyl species. The present results are consistent with observations and conclusions reached
for a previously studied Rh
tem.
4 5
(CO)12/HMn(CO) /3,3-dimethylbut-1-ene hydroformylation sys-
In tr od u ction
constitute a well-defined reaction topological basis for
synergism. In such a context, synergism refers to the
combined application of more than one metal, leading
to regio-, chemo-, and stereoselectivities and/or activities
which differ significantly from a strictly additive effect.
7
Stoichiometric binuclear elimination is the term used
to describe the reaction between two organometallic
complexes yielding a higher nuclear complex and an
organic product. Thus, in the case of two substituted
Hidai and co-workers indicated that the synergistic
effect for cobalt and ruthenium in the hydroformylations
1
1
2
2
mononuclear complexes R M Ln and R M L′n, the di-
1
2
1
2
nuclear complex M M LnL′n and product R R result.
1
2
When M and M are the same element, the stoichio-
metric binuclear elimination is homometallic. In per-
haps the first known case, Heck and Breslow identified
such a reaction pathway involving an intermolecular
hydride transfer between HCo(CO)4 and CH3COCo(CO)3
to eliminate the dinuclear complex Co2(CO)8 and alde-
(
2) (a) Ungvary, F.; Marko, L. J . Organomet. Chem. 1969, 20(1),
2
05-9. (b) Norton, J . R.; Carter, W.; Kelland, J .; Okrasinski, S. Adv.
Chem. Ser. 1978, No. 167, 170-180. (c) Norton, J . R. Acc. Chem. Res.
1979, 12(4), 139-145. (d) Nappa, M. J .; Santi, R.; Diefenbach, S.;
Halpern, J . J . Am. Chem. Soc. 1982, 104(2), 619-621. (e) Nappa, Mario
J .; Santi, R.; Halpern, J . Organometallics 1985, 4(1), 34-41. (f) J ones,
W. D.; Bergmann, R. G. J . Am. Chem. Soc. 1979, 79, 5447-5449. (g)
J ones, W. D.; Huggins, J .; Bergman, R. J . Am. Chem. Soc. 1981,
1
hyde product. Additonal examples of homometallic
stoichiometric binuclear elimination have been identi-
1
03(15), 4415-4423. (h) Alemdaroglu, N. H.; Penninger, J . M. L; Oltay,
E. Monatsh. Chem. 1976, 197, 1103-1153. (i) Mirbach, M. F. J .
Organomet. Chem. 1984, 205-213. (j) Brandes, K. H.; J onassen, H.
B. Z. Anorg. Allg. Chem. 1966, 343(3-4), 215-219. (k) Ungvary, F.;
Marko, L. Organometallics 1982, 1(9), 1120-1125. (l) Hoff, Carl;
Ungvary, F.; King, R. B.; Marko, L. J . Am. Chem. Soc. 1985, 107(3),
666-671. (m) Kovacs, I.; Ungvary, F.; Marko, L. Organometallics 1986,
2
fied.
When M1 and M2 are the different elements, the
stoichiometric binuclear elimination is heterobimetallic.
Approximately 13 heterobimetallic stoichiometric bi-
nuclear eliminations are presently known. An extensive
list is provided in Table 1.
The possible existence of heterobimetallic binuclear
eliminations during bimetallic catalytic syntheses has
been a subject of considerable interest. Indeed, the
bimetallic catalytic binuclear elimination reaction would
5
1
(2), 209-215. (n) Azran, J .; Orchin, M. Organometallics 1984, 3(2),
97-199.
(3) Mitchell, C. M.; Stone, F. G. A. J . Chem. Soc., Dalton Trans.
972, (1), 102-107.
1
1
1
1
(4) Renaut, P.; Tainturier, G.; Gautheron, B. J . Organomet. Chem.
978, 150(1), C9-C10.
(5) Martin, B.; Warner. D. K.; Norton, J . R. J . Am. Chem. Soc. 1986,
08(1), 33-39.
(6) Hoxmeier, R. J .; Blickensderfe, J . R.; Kaesz, H. D. Inorg. Chem.
979, 18(12), 3453-3461.
(7) (a) Golodov, V. A.; Kashnikova, L. V. Kinet. Katal. 1981, 22(3),
*
To whom correspondence should be addressed. E-mail: chemvg@
nus.edu.sg. Phone: +65-6-874-6617. Fax: +65-6-779-1936.
1) (a) Breslow, D. S.; Heck, R. F. Chem. Ind. 1960, 467. (b) Heck,
R. F.; Breslow, D. S. J . Am. Chem. Soc. 1961, 83, 4033-407.
793-794. (b) Golodov, V. A. J . Res. Inst. Catal., Hokkaido Univ. 1981,
29(1), 49-60. (c) Golodov, V. A.; Sokolsky, D. V. Eurasian Chemi.-
Technol. J . 2000, 2(1), 29-34.
(
1
0.1021/om0342661 CCC: $27.50 © 2004 American Chemical Society
Publication on Web 07/22/2004

4
132 Organometallics, Vol. 23, No. 17, 2004
Ta ble 1. Bim eta llic Stoich iom etr ic Bin u clea r Elim in a tion Rea ction s
Li et al.
no.
reacn
ref
1
2
3
4
5
6
7
8
9
1
1
1
1
HMn(CO)5 + CH3AuPPh3 f Au(PPh3)Mn(CO)5 + CH4
3
3
2c
4
5
5
6
6
6
2m
2m
2m
2m
HOs(CO)4Si(CH3)3 + CH3AuPPh3 f Au(PPh3)Os(CO)4Si(CH3)3 + CH4
Os(CO)4(H)CH3 + HRe(CO)5 f HOs(CO)4Re(CO)5 + CH4
5
2
5
(η -C5H5)2Zr(CH3) + (η -C5H5)Mo(CO)3H f (C5H5)2CH3ZrMo(C5H5)(CO)3 + CH4
EtRe(CO)5 + HMn(CO)5 f EtCHO + Re(CO)4Mn(CO)5
EtRe(CO)5 + HW(CO)3CpEtCHO + Re(CO)5W(CO)3Cp
5
5
5
1
(η -C5H5)MoH2 + CH3Mn(CO)5 f (η -C5H5)(CO)Mo(η :η -C5H4)Mn(CO)4 + H2 + CH4
5
5
5
1
(η -C5H5)WH2 + CH3Mn(CO)5 f (η -C5H5)(CO)W(η :η -C5H4)Mn(CO)4 + H2 + CH4
5
5
5
1
(η -C5H5)ReH + CH3Mn(CO)5 f (η -C5H5)(H)Re(η :η -C5H4)Mn(CO)4 + CO + CH4
HMn(CO)5 + EtOC(O)Co(CO)4 f EtOC(O)H + MnCo(CO)9
0
1
2
3
HMn(CO)5 + EtOC(O)CH2Co(CO)4 f EtOC(O)CH3 + MnCo(CO)9
HCo(CO)4 + EtOC(O)Mn(CO)5 f EtOC(O)H + MnCo(CO)9
HCo(CO)4 + EtOC(O)CH2Mn(CO)5 f EtOC(O)CH3 + MnCo(CO)9
of cyclohexene, 1-hexene, and styrene might be ex-
plained by such a dinuclear elimination between a cobalt
acyl and a ruthenium hydride. In addition, the cobalt-
5
HMn(CO) was prepared using a modified literature prepa-
ration.12 The reaction involves the interaction of H
(CO)10 according to eq 1.
2
2
with Mn -
8
ytterbium-meditated syntheses of aldehydes by the
Russian group of Beletskaya deserve special attention.
It was shown that lanthanum hydrides react in a
stoichiometric manner with acyl cobalt tetracarbonyls
Mn (CO) + H + hν f 2HMn(CO)
5
(1)
2
10
2
The standard procedure is as follows. A 100 mg portion of
Mn (CO)10 was dissolved in 50 mL of n-hexane in a Schlenk
tube. The solution was first saturated with H and a small
9
to give aldehyde. Furthermore, under catalytic reaction
2
conditions a large increase in activity is observed. This
observation was interpreted to mean that the bimolecu-
lar reaction of acyl cobalt carbonyl RCOCo(CO)x with
the ytterbium hydride HYb(C5H5)y represents the larg-
est single contribution to aldehyde formation. However,
in situ spectroscopic and kinetic information was not
available.
2
amount of CO for ca. 15 min at room temperature and then
irradiated with an immersion lamp (450 W mercury-vapor
lamp, Model 7825-34, ACE Glass, Vineland, NJ ). Trace CO
was needed during photoirradiation in order to prevent total
decomposition of the manganese carbonyl and hydride. After
ca. 10 min, infrared spectroscopy indicated the formation of
-1
HMn(CO) (νCO 2014, 2007 cm ). Under our experimental
5
Recently, the Rh4(CO)12/HMn(CO)5 bimetallic cata-
lyzed hydroformylation of 3,3-dimethylbut-1-ene (33DMB)
was studied systematically.10 The spectroscopic and
kinetic analyses provided evidence for the existence of
bimetallic CBER. The present contribution extends our
understanding of the Rh4(CO)12/HMn(CO)5 bimetallic
system by application to the homogeneous catalyzed
hydroformylation of cyclopentene. New details about the
mechanism have been obtained, and support for a
second example of bimetallic CBER is presented.
conditions, it was found that the maximum yield (ca. 25%) of
HMn(CO) appeared around 70 min and further irradiation
5
would give some yellow precipitate. It was also found that the
reproducibility of this photoreaction was not ideal; therefore,
the HMn(CO) concentrations for each experiment are some-
5
what different from run to run. (Note that this has no
detrimental effect on the quality of the catalytic data or their
analysis, since all concentrations can be measured in situ.)
Since the HMn(CO) is very unstable, the HMn(CO) solution
5 5
was immediately transferred under CO/argon to the autoclave
after photoirradiation.
The puriss quality cyclopentene (Fluka AG) obtained for this
study had a purity of 98.5%. To remove the possible trace of
dienes in it, the cyclopentene was first refluxed with 2 equiv
of maleic anhydride (99%, Fluka AG) for 2 h. The distilled
cyclopentene was repeatedly washed with distilled water,
passed over 4-A molecular sieves, distilled from calcium
hydride (95%, Merck), and stored under argon at -20 °C. The
resulting purity of the cyclopentene was 99.99%, as determined
by GC analysis (HP6890; HP-1 methylsiloxane capillary
column, 100 °C; FID, 250 °C). The puriss quality n-hexane
(99.6%, Fluka AG) was distilled from sodium-potassium under
argon for ca. 5 h to remove the trace water and oxygen.
Ap p a r a tu s. The apparatus used in this study was identical
Exp er im en ta l Section
Gen er a l In for m a tion . All solution preparations and trans-
fers were carried out under a purified argon (99.9995%, Saxol,
Singapore) atmosphere using standard Schlenk techniques.11
The argon was further purified before use by passing it
through deoxy and zeolite columns. Purified carbon monoxide
(
research grade, 99.97%, Saxol, Singapore) and purified hy-
drogen (99.9995%, Saxol, Singapore) were also further purified
through deoxy and zeolite columns before they were used in
the hydroformylation experiments. Purified nitrogen (99.9995%,
Saxol, Singapore) was used to purge the Perkin-Elmer FT-IR
spectrometer system.
with that used in our previous study of the Rh-Mn-catalyzed
Rh
4
(CO)12 (98%) was purchased from Strem Chemicals
(CO)10 (98%)
hydroformylation of 3,3-dimethylbut-1-ene.10
(
Newport, MA) and was used as obtained. Mn
2
In Situ Sp ectr oscop ic a n d Kin etic Stu d ies. The general
procedures used are that same as those used in ref 10. The
experimental design of the experiments involved 300 mL of
solvent and the intervals 281.5-308.9 K, PH2 ) 1.0-4.0 MPa,
was also purchased from Strem Chemicals and was used
without further purification.
(
8) (a) Koyasu, Y.; Fukuoka, A.; Uchida, Y.; Hidai, M. Chem. Lett.
985, (7), 1083-1086. (b) Hidai, M.; Fukuoka, A.; Koyasu Y.; Uchida,
Y. J . Mol. Catal. 1986, 35(1), 29-37.
9) (a) Beletskaya, I. P.; Voskoboinikov, A. Z.; Magomedov, G. K.
Metalloorg. Khim. 1989, 2(4), 810-813. (b) Beletskaya, I. P.; Mago-
medov, G. K.; Voskoboinikov, A. Z. J . Organomet. Chem. 1990, 385,
P
CO ) 1.0-4.0 MPa, initial alkene 2-10 mL, initial Rh
4
(CO)12
1
2
4.05-106.47 mg, and initial Mn (CO)10 25.0-229.0 mg. The
2
detailed experimental design for this study is shown in Table
2.
All hydroformylation experiments exhibited product forma-
tion rates belonging to infinitely slow reaction compared to
(
2
5
89-295.
(
540-5548.
(
10) Li, C, Widjaaj, E.; Garland, M. J . Am. Chem. Soc. 2003, 125(18),
11) Shriver, D. F.; Drezdzon, M. A. The Manipulation of Air-
Sensitive Compounds; Wiley: New York, 1986.
(12) Byers, B. H.; Brown, T. L. J . Am. Chem. Soc. 1977, 99(8), 2527-
2532.

Rh
4
(CO)12-Catalyzed Hydroformylation
Organometallics, Vol. 23, No. 17, 2004 4133
in the wavenumber range from 1000 to 1550 cm-1, all
experimental spectroscopic data were accordingly truncated.
The consolidated data gave rise to an absorbance matrix with
the following size: rows, 1461 vectors (spectra); columns, 4751
Ta ble 2. Exp er im en t Design for th e Bim eta llic
Rh
4
(CO)12/HMn (CO)
5
-Ca ta lyzed Hyd r ofor m yla tion
of Cyclop en ten e
amt of
Mn (CO)10
mg
amt of amt of amt
amt of
4
-
1
2
,
CO,
H
2
,
of CP, Rh
mL
(CO)12
mg
,
temp,
K
channels (wavenumber range is 1550-2500 cm and data
expt
MPa
MPa
-1
interval 0.2 cm ). Discussion of the mathematical procedures
standard
blank
pure Rh
pure Mn
CO variation
117.36
0
0
122.0
103.07
2.0
2.0
2.0
2.0
1.0
3.0
4.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
1.0
1.0
3.0
4.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
5
5
5
5
5
5
5
5
5
5
2
10
5
5
5
5
5
5
51.60
0
51.89
0
289.7
289.7
289.7
289.7
289.7
289.7
289.7
289.7
289.7
289.7
289.7
289.7
289.7
289.7
289.7
289.7
281.5
293.7
299.4
308.9
289.7
289.7
289.7
used are discussed in detail in ref 15.
Kin etics a n d Activa tion P a r a m eter s. Mole fraction
concentrations were used throughout for all species including
dissolved gases. Rate expressions were developed using only
the mole fractions as concentrations.
Arrhenius and Eyring type activation parameters were
determined for the rate constants obtained as a function of
temperature. Accordingly, the rate constants were modeled as
either
53.73
54.32
52.58
49.46
52.48
47.74
49.53
50.63
51.04
49.80
51.84
50.63
48.12
49.00
52.05
50.07
24.05
82.11
1
1
08.45
08.97
H
2
variation
106.30
1
1
01.86
04.57
cyclopentene
variation
100.37
109.36
25.00
k ) A exp(-∆Ea /RT)
(3)
Mn
2
(CO)10
i
i
variation
50.00
1
2
50.25
29.00
or
temp
variation
104.79
108.28
q
q
k ) κT/h exp(-∆H /RT + ∆S /R)
i i i
(4)
1
1
04.50
03.11
5
5
5
5
where κ is Boltzmann’s constant, h is Planck’s constant, A is
Rh
4
(CO)12
variation
103.29
103.68
the Arrhenius preexponential factor, Ea is the apparent
1
00.30
5
106.47
activation energy, ∆H is the standard enthalpy of activation,
q
q
and ∆S is the standard entropy of activation.
gas-liquid mass transfer (category H of the Hatta classifica-
1
3
tions). Since the maximum rate of the hydroformylation in
this study was 10-6 mol/s, gas-liquid mass transfer control
Resu lts
can be ignored in all the experiments.
Sp ect r oscop ic Asp ect s. The in situ spectroscopic
Blank experiments, in the presence of organic reactants but
without added complexes, were performed throughout the
study. A very small residual activity of the cleaned reactor
with blank experiments was occasionally observed.
The in situ spectra were taken every 10 min during the each
data were subjected to band-target entropy minimiza-
1
5
tion (BTEM). BTEM is an advanced and totally blind
deconvolution technique based on Shannon entropy
1
6
criteria. Hydroformylation of alkene in the presence
-
1
of tetrarhodium dodecacarbonyl alone has been exten-
6
h experiment in the range of 1000-2500 cm
with a
-1
sively studied at low temperatures.17
resolution of 4 cm . A total of 1461 spectra were obtained for
the further spectroscopic and kinetic analyses.
In the present Rh4(CO)12/HMn(CO)5 bimetallic cat-
alyzed hydroformylation of cyclopentene (CP), the
chemometric analysis of the in situ infrared spectra of
the reaction recovered the following: the precursors Rh4-
The appropriateness of various finite difference schemes for
the evaluation of in situ homogeneous catalytic rate data and
TOF evaluation has been carefully evaluated. The rates of
the product formation of cyclopentanecarboxaldehyde were
calculated in terms of turnover frequency (TOF) by using a
finite difference equation:
1
4
(CO)12, HMn(CO)5, and Mn2(CO)10, the organic reagent
cyclopentene, the rhodium intermediate RCORh(CO)4,
and the organic product cyclopentanecarboxaldehyde.
Similar to the results of the Rh4(CO)12/HMn(CO)5-
catalyzed hydroformylation of 33DMB, no Rh-Mn
complex analogous to MnCo(CO)9
d[aldehyde] /dt
t
1
0
TOF )
(2)
t
[
RCORh(CO)4]
t
2m
was observed in
Rh4(CO)12/HMn(CO)5-catalyzed hydroformylation of the
cyclopentene system. Furthermore, GC analysis (HP6890;
HP-1 methyl siloxane capillary column, 90 °C; FID, 250
°C) did not find the possible hydrogenation product
cyclopentane in the reaction mixture.
This central difference expression provides a numerical
value for the reaction rate at time t, based on the measured
in situ concentrations of the aldehyde at time t - 1 and t + 1.
The time interval between t and t + 1 or t - 1 was 20 min.
A total of 1461 in situ FTIR spectra were acquired in
this experimental study, over the spectral interval
Com p u ta tion a l Section
Sp ectr oscop ic Asp ects. The newly developed algorithm
-1
-1
1
550-2500 cm
with data intervals of 0.2 cm
.
for algebraic system identification was employed to analyze
Notable vectors from the singular value decomposition
of the spectroscopic matrix A1461×4751 are shown in
Figure 1, and the recovered pure component spectra
obtained from the BTEM analysis are shown in Figure
1
5
the in situ IR spectra. Since no special features were observed
(
13) Garland, M. Transport Effects in Homogeneous Catalysis. In
Encyclopedia of Catalysis; Horvath, I. T., Ed.; Wiley: New York, 2002.
14) Shirt, R.; Garland, M.; Rippin, D. W. T. Anal. Chim. Acta 1998,
74, 67-91.
15) (a) Chew, W.; Widjaja, E.; Garland, M. Organometallics 2002,
(
2
(solvent hexane, atmospheric moisture, CO2, and
3
(
dissolved CO are omitted). From a spectroscopic/observ-
able viewpoint, the reactive system is very well defined
and the signal-to-noise level is very high. No further
2
2
1, 1882-1990. (b) Widjaja, E.; Li, C.; Garland, M. Organometallics
002, 21, 1991-1997. (c) Li, C.; Widjaja, E.; Chew, W.; Garland, M.
Angew. Chem., Int. Ed. 2002, 20, 3785-3789. (d) Widjaja, E.; Li, C.;
Garland, M. In Proceeding of the International Conference on Scientific
&
Engineering Computation (IC-SEC), Recent Advances in Computa-
(16) Shannon, C. E. Bell Syst. Technol. J . 1948, 27, 379-423.
(17) (a) Garland, M.; Bor, G. Inorg. Chem. 1989, 28, 410-413. (b)
Garland, M.; Pino, P. Organometallics 1990, 10, 1693-1704. (c) Fyhr,
C.; Garland, M. Organometallics 1993, 12, 1753-1764. (d) Liu, G.;
Volken, R.; Garland, M. Organometallics 1999, 18(17), 3429-3436.
tional Sciences and Engineering; Lee, H. P., Kumar, K., Eds.; Imperial
College Press: London, 2002; pp 36-40. (e) Widjaja, E.; Li, C.; Chew,
W.; Garland, M. Anal. Chem. 2003, 75, 4499-4507. (f) Widjaja, E.; Li,
C.; Garland, M. J . Catal. 2004, 223(2), 278-289.

4
134 Organometallics, Vol. 23, No. 17, 2004
Li et al.
F igu r e 1. Singular value decomposition of the precondi-
tioned in situ spectroscopic data showing the 1st, 2nd, 5th,
F igu r e 3. Time-dependent mole fractions of the four
6
th, and 7th significant vectors and the 200th vector. The
observable organometallics Rh
CO)10, and HMn(CO) during a 350 min hydroformylation
of cyclopentene at ca. 2.0 MPa of CO, 2.0 MPa of H , and
89.7 K with 350 mL of hexane and 5 mL of cyclopentene.
4 4 2
(CO)12, RCORh(CO) , Mn -
marked extrema are those which were used to recover the
organometallic pure component spectra as well as alkene
and aldehyde by BTEM.
(
5
2
2
hydroformylation of cyclopentene are shown in Figure
3
.
It can be noted that after the simultaneous addition
of the manganese carbonyls HMn(CO)5 and Mn2(CO)10
and hydrogen to the systems initially containing Rh4-
(CO)12, alkene, hexane, and CO, the concentrations of
Rh4(CO)12 monotonically decreased and the concentra-
tion of RCORh(CO)4 monotonically increased for the
first 6 h. However, the data presented in Figure 3
indicate that the decrease of Rh4(CO)12 and increase of
RCORh(CO)4 are not simple monotonic functions at
initial reaction times. Indeed, upon addition of Mn2-
(CO)10, HMn(CO)5, and H2, a rapid decline in the
catalyst precursor concentration and a corresponding
increase in acyl complex concentration occur.
The data from all 23 experimental runs suggest that
the concentration of HMn(CO)5 has a big effect on the
transformation of Rh4(CO)12. Generally speaking, the
higher the concentration of HMn(CO)5, the faster the
transformation of Rh4(CO)12 to RCORh(CO)4. Moreover,
the half-life for Rh4(CO)12 in the mixed Rh4(CO)12/HMn-
F igu r e 2. Recovered pure component spectra of the
organic and organometallic species using BTEM: (a)
RCORh(CO)
4
; (b) Rh
4
(CO)12; (c) cyclopentene (CP); (d) Mn
5
; (f) cyclopentanecarboxaldehyde.
2
-
(CO)10; (e) HMn(CO)
(
CO)5 system is ca. 60 min when 150.25 mg of Mn2(CO)10
observable species could be identified in the present
experimental study.
was used, whereas a half-life under similar reaction
conditions for pure Rh systems is ca. 2 h. This atypical
initial time behavior is indicative of the existence of a
second parallel precatalytic pathway for the transfor-
mation of Rh4(CO)12 to RCORh(CO)4 in the presence of
manganese carbonyls: namely, the hydride-facilitated
degradation of the cluster Rh4(CO)12. It should be
mentioned that the effects observed in the present study
were also observed in the Rh4(CO)12/HMn(CO)5/3,3-
H yd r ofor m yla t ion of Cyclop en t en e w it h On ly
Mn 2(CO)10/HMn (CO)5. To determine the activities of
Mn2(CO)10/HMn(CO)5, a blank experiment with only
Mn2(CO)10/HMn(CO)5 as the precursor was carried out.
The 6 h experiment showed there was no observable
reaction under our reaction conditions (2.0 MPa of CO,
2
.0 MPa of H2, 289.7 K). This is consistent with the inert
behavior of manganese toward transformations of al-
1
0
1
8
dimethylbut-1-ene system.
kenes under mild conditions.
In addition, Figure 3 contains a new piece of informa-
tion which was unavailable in our previous study.10 In
the present study, HMn(CO)5 was added and a short
period (5 min) was allowed to transpire before molecular
hydrogen was added to the system and the spectral
acquisition started. As seen from the first data point in
Figure 3, the concentration of the rhodium acyl complex
is nonzero; therefore, the species is present in the first
Hyd r ofor m yla tion of Cyclop en ten e w ith Rh 4-
CO)12 a n d HMn (CO)5. In itia l Rea ction Tim es. The
typical time-dependent concentrations of the precursors
Rh4(CO)12, HMn(CO)5, and Mn2(CO)10 and the only
observable intermediate, RCORh(CO)4, during an active
(
(18) Weil, T. A.; Metlin, S.; Wender, I. J . Organomet. Chem. 1973,
4
9(1), 227-32.

Rh
4
(CO)12-Catalyzed Hydroformylation
Organometallics, Vol. 23, No. 17, 2004 4135
F igu r e 5. Effect of manganese loading on the turnover
frequency of the rhodium-catalyzed hydroformylation of
F igu r e 4. Time-dependent mole fractions of the product
aldehyde during 350 min hydroformylations conducted
with different manganese loadings. Typical reaction condi-
tions during a 350 min hydroformylation of cyclopentene
cyclopentene at ca. 2.0 MPa of CO, 2.0 MPa of H
2
and 289.7
K with 350 mL of hexane and 5 mL of cyclopentene.
5
0.0, 117.36, 150.25, and 229.0 mg, respectively. The
are ca. 2.0 MPa of CO, 2.0 MPa of H
2
, and 289.7 K with
errors are listed as twice the standard deviation: i.e.,
the 95% confidence limit.
3
50 mL of hexane and 5 mL of cyclopentene. (The missing
data at ca. 220 min for 229 mg of Mn
data acquisition interruption.)
2
(CO)10 represent a
The TOF values further demonstrate that manganese
promotes rhodium-catalyzed hydroformylation of cyclo-
pentene. In previous hydroformylations of 33DMB,
cyclohexene, and styrene with pure unmodified rhodium
carbonyl, it has been repeatedly shown that turnover
frequencies (based on instantaneous concentration of
rhodium intermediates) are only dependent on the
temperature, hydrogen pressure, and carbon monoxide
pressure. Our present results clearly show that TOFs
increase with the increased manganese loadings. When
the initial loading of Mn2(CO)10 was 229.0 mg, the TOF
is 230% of that anticipated from pure rhodium-catalyzed
hydroformylation.
spectrum taken. This is rather compelling evidence for
the HMn(CO)5 attack on Rh4(CO)12 in the presence of
substrate to produce RCORh(CO)4. This provides an
independent confirmation of the manganese hydride
facilitated Rh4(CO)12 fragmentation.
P r od u ct F or m a tion . The concentration profiles of
the product aldehyde as a function of increasing man-
ganese loadings are shown in Figure 4. The concentra-
tions of the product aldehyde monotonically increase in
all six experiments. The production of aldehyde could
be identified within the first few minutes when HMn-
(
CO)5/Mn2(CO)10 were added simultaneously.
As shown in Figure 4, a very significant increase in
Moreover, it should be noted that the TOF with
150.25 mg of Mn2(CO)10 is quite close to that of 50.0
mg of Mn2(CO)10 and even smaller than the TOF with
117.36 mg of Mn2(CO)10. This can be traced to the HMn-
(CO)5 concentration profiles shown in Figure 6. Indeed,
the average HMn(CO)5 concentrations of the experi-
ments with 50.0 and 150.25 mg of Mn2(CO)10 are quite
similar, while the average HMn(CO)5 concentrations
with 117.36 mg of Mn2(CO)10 are higher. This suggests
that HMn(CO)5 and not Mn2(CO)10 is associated with
the high activities in the Rh4(CO)12/HMn(CO)5 bimetal-
lic system. The higher the HMn(CO)5 concentration in
an experiment, the higher the TOF.
aldehyde formation was observed in the experiments
when both rhodium carbonyl and manganese carbonyl
complexes were used. Increased concentrations of HMn-
(
CO)5 lead to increased hydroformylation rates. After
the first 100 min, the rates of product formation are
-
6
more or less constant. The rates are ca. 8.2 × 10 , 1.1
-
5
-5
-5
-5
×
10 , 1.3 × 10 , 1.6 × 10 , 1.8 × 10 , and 2.3 ×
5
-
1
0
0
(mol fraction)/min when Mn2(CO)10 amounts are
, 25.0, 50.0, 117.36, 150.25, and 229.0 mg, respectively.
When 229.0 mg of Mn2(CO)10 was used with Rh4(CO)12,
the rate of hydroformylation was 300% of the rate
anticipated from the rhodium loading alone. Since HMn-
The present results are entirely consistent with those
of the Rh4(CO)12/ HMn(CO)5 bimetallic catalyzed hy-
(CO)5/Mn2(CO)10 is nearly inert toward transformations
1
0
of alkenes under the present reaction conditions, it
implies that HMn(CO)5/Mn2(CO)10 is involved in the
rhodium catalytic pathway to produce aldehyde. This
is consistent with the observations in the Rh4(CO)12/
HMn(CO)5/3,3-dimethylbut-1-ene system.10
droformylation of 33DMB. Specifically, the activity of
the systems appears correlated with the HMn(CO)5
concentrations and not the Mn2(CO)10 concentrations.
HMn(CO)5 appears necessary for the final product
formation and is therefore involved in at least one step
of the catalytic cycle of the hydroformylation.
TOF An a lyses. The effect of HMn(CO)5/Mn2(CO)10
loading on the TOF is shown in Figure 5. The turnover
frequencies (TOF) based on rhodium and defined in the
Experimental Section were 0.12 ( 0.01, 0.15 ( 0.01,
Ca ta lysis a n d Kin etics. A total of 23 kinetic experi-
ments in six subseries were performed for the kinetic
study. In each set, one experimental parameter was
systematically varied while the remaining variables
were held essentially constant. The complete experi-
0
.20 ( 0.02, 0.23 ( 0.002, 0.21 ( 0.02, and 0.28 ( 0.02
-
1
min when initial Mn2(CO)10 amounts were 0, 25.0,

4
136 Organometallics, Vol. 23, No. 17, 2004
Li et al.
Ta ble 3. Regr essed k Va lu es a t Differ en t
Tem p er a tu r es
expt no.
temp (K)
k1 (L/min)
k2 (L/min)
T-1
T-2
T-3
T-4
T-5
281.5
289.7
293.7
299.4
308.9
0.095
0.314
0.524
0.995
3.02
21
37
44
71
137
(
CO)12, and 229.0 mg of Mn2(CO)10 (XMnH(CO) ) 4 × 10-5)
5
the contribution of CBER to the final product formation
is 51%.
Rearrangement of the two terms k1[CO]-1/k2[CO]-1.6
leads to the conclusion that under 2.0 MPa of CO, HMn-
(CO)5 is ca. 520 times more efficient than H2, for
reaction with the acyl-rhodium bond, on a mole-to-mole
basis. The exponent x ) -1.6 ( 0.01 suggests that
coordinative unsaturation may occur in both metal
carbonyl complexes.
The temperature was systematically varied in five
unmodified rhodium-catalyzed hydroformylations of cy-
clopentene promoted with HMn(CO)5 at 2.0 MPa of CO
and 2.0 MPa of H2 to regress the reaction activation
parameters. The five temperatures and the correspond-
ing regressed rate constants k1 and k2 at different
reaction temperatures are shown in Table 3.
F igu r e 6. Actual time-dependent concentrations of HMn-
CO) in experiments with different nominal manganese
loadings at ca. 2.0 MPa of CO, 2.0 MPa of H , and 289.7 K
(
5
2
with 350 mL of hexane and 5 mL of cyclopentene.
mental design is documented in the Experimental
Section. The strong effect of HMn(CO)5 on the product
formation rate and TOF was elucidated.
Regression of the rate constants as a function of
temperature provided acceptable estimates of the acti-
As observed in the Rh4(CO)12-catalyzed hydroformy-
q
q
vation parameters ∆Ea, ∆H , and ∆S . For k1 the
1
0
lation of 33DMB promoted with HMn(CO)5/Mn2(CO)10,
q
parameters were ∆Ea ) 91 ( 6 kJ /mol, ∆H ) 88 ( 6
the following two-term linear-bilinear form in organo-
metallic species provided an excellent fit for the observ-
able kinetics of the aldehyde formation:
q
kJ /mol, and ∆S ) 15 ( 20 J /(mol K). For k2 the
q
parameters are ∆Ea ) 49 ( 8 kJ /mol, ∆H ) 47 ( 8
q
kJ /mol, and ∆S ) -88 ( 28 J /(mol K). The errors are
listed as twice the standard deviation: i.e., the 95%
confidence limit.
-
1
rtotal ) k [RCORh(CO) ][H ][CO]
+
1
4
2
The present activation parameters for k1, correspond-
ing to the classic unicycle hydroformylation mechanism,
are consistent with previous work (pure rhodium sys-
x
k [RCORh(CO) ][HMn(CO) ][CO] (5)
2
4
5
Approximately 1200 sets of mole fraction data were
used in the regression. The activity of the system (rate
of product formation) did not correlate with the presence
of any other observables.
Using the nonlinear regression program “nonlinfit”
in Matlab, the regressed numerical values of the rate
17
tem and HMn(CO)5/Mn2(CO)10 bimetallic catalyzed
10
hydroformylation of 33DMB ). As for k2, the negative
entropy clearly indicates that the degrees of freedom
have decreased in the transition state.
Discu ssion
-
1
constants at 289.7 K were k1 ) 0.314 ( 0.08 min , k2
-
1
Evid en ce for Existen ce of CBER. Similar to the
)
37 ( 15 min , and x ) -1.6 ( 0.01. The errors are
Rh4(CO)12/HMn(CO)5 bimetallic catalyzed hydroformy-
listed as twice the standard deviation: i.e., the 95%
confidence limit. The rate constant k1 is consistent with
the previous studies. The rate constants and the expo-
nent for CO are in agreement with the results of the
Rh4(CO)12-catalyzed hydroformylation of 33DMB pro-
moted with HMn(CO)5/Mn2(CO)10, where k1 ) 0.446 (
1
0
lation of 33DMB, the present kinetic results for
cyclopentene hydroformylation strongly suggest the
simultaneous existence of a unicyclic catalytic topology
and at least one bimetallic catalytic elimination reac-
tion. Accordingly, the unicyclic catalysis occurs exclu-
sively on mononuclear rhodium intermediates, and the
bimetallic CBER occurs on a set of rhodium, manganese,
and rhodium-manganese intermediates.
-
1
-1
0
.018 min and k2 ) 44.5 ( 14 min at 298 K and the
1
0
exponent was x ) -1.5 ( 0.1. In a broader context,
the hydroformylation rate differences between the cyclic
alkene (cyclopentene) and the branched alkene (33DMB)
are small. This result is consistent with the extensive
In pure rhodium carbonyl systems, RCORh(CO)4 has
been repeatedly shown to be the predominant coordi-
natively saturated rhodium intermediate under hydro-
formylation conditions and RCORh(CO)3 is the equili-
brated, coordinately unsaturated complex. Mechanis-
tically, hydrogen activation on the acyl complex RCORh-
rate studies by Marko,19 as well as in situ studies
17d
using unmodified rhodium carbonyl and many different
alkenes.
The contributions of these two terms under the
reference reaction conditions of this experimental study
were ca. 1:1. In particular, at 289.7 K, 2.0 MPa of CO,
(CO)3 is the rate-limiting step in the unicyclic rhodium
catalysis. In the bimetallic CBER, the attack of (a) HMn-
(CO)5 on RCORh(CO)3 and/or (b) HMn(CO)4 on RCORh-
2
.0 MPa of H2, 5 mL of cyclopentene, ca. 50 mg of Rh4-
(
CO)3 are the rate-limiting steps. The exponent x ) -1.6
(
19) Heil, B.; Marko, L. Chem. Ber. 1969, 102, 2238-2240.
( 0.1 arises from this latter competing situation.

Rh
4
(CO)12-Catalyzed Hydroformylation
Organometallics, Vol. 23, No. 17, 2004 4137
Sch em e 1. P r op osed Rea ction Top ology for th e Sim u lta n eou s In ter con n ected Un icyclic Rh a n d Bim eta llic
Rh -Mn CBER Cyclop en ten e Hyd r ofor m yla tion Rea ction s
In the bimetallic Rh4(CO)12/HMn(CO)5 systems, the
attack of HMn(CO)5 on RCORh(CO)3 or of HMn(CO)4
on RCORh(CO)3 with elimination of aldehyde implies
the existence of a transient, coordinately unsaturated
bimetallic rhodium manganese dinuclear complex with
the probable stoichiometry RhMn(CO)8 (or possibly
RhMn(CO)7). However, such a species (or its coordina-
tively saturated analog) was not observable in these
experiments or in the Rh4(CO)12/HMn(CO)5/33DMB
existence of the bimetallic CBER. The R step denotes
the simultaneous generation of the crucial mononuclear
hydrides HRh(CO)3 and HMn(CO)5 from the dinuclear
complex H2RhMn(CO)8. The â step is the binuclear
elimination step. The steps -R and â are bimolecular
in metal complexes.
The possible existence of isomers of the intermediates
is not the emphasis of Scheme 1, and accordingly, we
have assumed the most common geometry for five-
coordinate rhodium structures: namely, axially substi-
tuted trigonal-bipyramidal geometries. The intercon-
nected topology of a unicyclic mechanism and a CBER
mechanism as shown in Scheme 1 directly leads to the
linear-bilinear mathematical form for the observed
kinetics of product formation as shown in eq 5.
The binuclear elimination step â has been drawn as
if the process proceeds through a concerted four-center
transition state for convenience. If the elimination does
not proceed through a four-center transition state but
instead through two or more simpler transition states,
additional intermediates exist on the reaction pathway.
Although the assumption of a four-center transition
state is not essential for the existence of the bimetallic
CBER, it perhaps helps to explain the rate increase
observed. In this regard, it is worthwhile to note that
1
0
system. This inability to observe a dinuclear bimetallic
complex is probably due to the very rapid activation of
molecular hydrogen to regenerate HRh(CO)3 and HMn-
(CO)5. Hydrogen activation, which is normally very
difficult to achieve, particularly under CO, is known to
occur exceptionally rapidly, even at low temperatures,
2
0
on heterometallic dinuclear carbonyl complexes. Hy-
2
1
drogen activation on CoRh(CO)7 is a good example.
The simultaneous existence of a unicyclic Rh topology
and a bimetallic Rh-Mn CBER for cyclopentene hydro-
formylation is represented in Scheme 1. All intermo-
lecular transformations in the unicyclic topology are
first order in the corresponding mononuclear metal
complexes. Two elementary steps in Scheme 1 are
highlighted by the labels R and â. These labels are used
to identify the key mechanisms which give rise to the
(1) the hydroformylation is being performed in an apolar
(
20) J ames, B. R. In Comprehensive Organometallic Chemistry: The
solvent (n-hexane), (2) HMn(CO)5 is acidic and hence
quite polarized with partial positive charge on the
hydrogen atom, and (3) RCORh(CO)4 should be have a
partial negative charge on the acyl carbon (the elec-
Synthesis, Reactions, and Structures of Organometallic Compounds;
Wilkinson, G., Stone, F. G. A., Abel, E. W., Eds.; Pergamon Press: New
York, 1982; Vol. 18.
(
21) Garland, M.; Pino, P. Organometallics 1990, 9(6), 1943-1949.

4
138 Organometallics, Vol. 23, No. 17, 2004
Li et al.
Ch a r t 1. P r op osa l for a F ou r -Cen ter Tr a n sition
Sta te in th e Bin u clea r Elim in a tion Step ,
High ligh tin g th e P a r tia l Ch a r ges a n d Hen ce
P r efer r ed Align m en t/Associa tion of Ma n ga n ese
Hyd r id e a n d Acyl-Rh od iu m Com p lexes
in the presence of HMn(CO)5 is due to its efficient attack
on {Rh4(CO)14}.
Redistribution of manganese carbonyl complexes also
occurs at initial reaction times. In the particular experi-
ments shown in Figure 3, the HMn(CO)5 concentrations
generally decrease from their nonzero initial values.
However, the HMn(CO)5 concentration never reached
zero in any of the 6 h hydroformylations. This was also
observed in the previous Rh4(CO)12/HMn(CO)5/33DMB
1
0
study. Redistribution of Mn2(CO)10 and HMn(CO)5 in
the bimetallic system may occur due to hydrogen
activation on Mn2(CO)10 and reaction of HMn(CO)5 with
the trace impurities in the 300 mL reaction autoclave
solution. The manganese hydride is well-known to be
acutely sensitive to trace moisture and oxygen. The
redistribution of Mn2(CO)10 and HMn(CO)5 may also
occur due to other reactions involving rhodium com-
plexes, manganese complexes, and mixed rhodium-
manganese species.
tronegativty of C is 2.55, and that of rhodium is 2.28).
The polarized organometallics HMn(CO)5 and RCORh-
(CO)4 should not be randomly and uniformly distributed
in solution. Instead, the molecules should have a non-
negligible tendency to associate and even align in
solution. This situation is illustrated in Chart 1. In
addition, there is the observation by Norton that metal
hydrides should have an unusually good ability to fill
vacant coordination sites.2 Taken together, these issues
provide a working hypothesis in order to explain the
significantly higher efficiency of HMn(CO)5 over H2 in
the hydrogenolysis of the acyl-rhodium bond.
Con clu sion s
c
Detailed, in situ spectroscopic and kinetic measure-
ments have been performed for the homogeneous rhod-
ium- and manganese-catalyzed hydroformylation of
cyclopentene. A series of well-defined experiments were
conducted to search for bimetallic catalytic binuclear
elimination. The addition of manganese carbonyl hy-
dride to the unmodified rhodium-catalyzed hydroformy-
lation of cyclopentene leads to a significant increase in
system activity. Regression of the kinetic data strongly
suggests that there is a statistically significant contri-
bution to aldehyde formation from the bimetallic cata-
lytic binuclear elimination of manganese hydride com-
plexes with acyl rhodium tetracarbonyl. As secondary
effects, it was observed that manganese carbonyl hy-
dride promotes the precatalytic transformation of rhod-
ium precursor to acyl-rhodium. The present results are
consistent with observations and conclusions reached
for a previously studied Rh4(CO)12/HMn(CO)5/3,3-di-
methylbut-1-ene hydroformylation system. This study
represents a second well-defined example of bimetallic
catalytic binuclear elimination.
In itia l Rea ction Tim es a n d P r eca ta lytic Step s.
The observation of RCORh(CO)4 formation upon addi-
tion of HMn(CO)5 to Rh4(CO)12/alkene in the absence
of added hydrogen suggests that two precatalytic mech-
anisms can be modeled. In the present Rh4(CO)12/HMn-
(CO)5 bimetallic system, a two-term equation providing
a good fit was again obtained from the 21 experiments
at 289.7 K:
2
d[RCORh(CO) ]/dt ) k ′[Rh (CO) ][CO] [H ] +
4
1
4
12
2
2
k ′[Rh (CO) ][CO] [HMn(CO) ] (6)
2
4
12
5
The first term is that which was previously shown to
arise in simple monometallic rhodium hydroformyla-
tions as well as during the Rh4(CO)12/HMn(CO)5/33DMB
1
0
hydroformylations. The common factor [Rh4(CO)12]-
2
[
CO] suggests a rate-limiting step involving the pre-
equilibrated open polyhedral intermediate {Rh4(CO)14}.
Kinetic evidence for open polyhedral clusters as inter-
mediates in cluster fragmentation reactions is avail-
Ack n ow led gm en t. Financial support for this ex-
perimental research was provided by the Academic
Research Fund of the National University of Singapore
2
2
able. The second term, which was also observed in the
(
NUS) under Grant No. R-279-000-089-112 “Bimetallic
1
0
previous Rh4(CO)12/HMn(CO)5/33DMB study, domi-
nates under manganese concentrations far lower than
the dissolved hydrogen concentrations. Thus, the pro-
nounced and accelerated rate of cluster fragmentation
Catalytic Binuclear Elimination”. Research scholarships
for C.L. and E.W. were provided by the Graduate School
of Engineering (NUS). C.L. also thanks the Singapore
Millenium Foundation (SMF) for a postdoctoral fellow-
ship.
(22) Bor, G.; Dietler, U. K.; Pino, P.; Poe, A. J . Organomet. Chem.
1
978, 154, 301-315.
OM0342661