Unmodified Rhodium-Catalyzed Hydroformylation of Alkenes Using Tetrarhodium Dodecacarbonyl. The Infrared Characterization of 15 Acyl Rhodium Tetracarbonyl Intermediates

Article abstract of DOI:10.1021/om990230q

The homogeneous catalytic hydroformylation of 20 alkenes was studied, starting with Rh₄(CO)₁₂ as catalyst precursor in n-hexane as solvent, using high-pressure in-situ infrared spectroscopy as the analytical tool. Five categories of alkenes were studied, namely, cycloalkenes (cyclopentene, cycloheptene, cyclooctene, and norbornene), symmetric internal linear alkenes (3-hexene, 4-octene, and 5-decene), terminal alkenes (1-hexene, 1-octene, 1-decene, 1-dodecene, and 1-tetradecene), methylene cycloalkanes (methylene cyclopropane, methylene cyclobutane, methylene cyclopentane, and methylene cyclohexane), and branched alkenes (2-methyl-2-butene, 2-methyl-2-pentene, 2-methyl-2-heptene, and 2,3-dimethyl-2-butene). The typical reaction conditions were T = 293 K, P_H2 = 2.0 MPa (0.018 mol fraction), P_CO = 2.0 MPa (0.033 mol fraction), [alkene]₀ = 0.1-0.02 mol fraction, and [Rh₄(CO)₁₂]₀ = 6.6 × 10^-5 mol fraction. In each experiment, with the exception of those involving methylene cyclopropane and the branched alkenes, the precursor Rh₄(CO)₁₂ was converted in good yield to the corresponding observable mononuclear acyl rhodium tetracarbonyl intermediate RCORh(CO)₄. Due to the spectral characteristics, the intermediate RCORh(CO)₄ is assigned a trigonal bipyrimidal geometry in all cases with C_s symmetry, with the acyl group taking an axial position. Under the present conditions, the cycloalkenes result in one acyl complex, the symmetric internal linear alkenes result in two acyl stereoisomers, the terminal alkenes result in three acyl complexes (two are stereoisomers), and the methylene cycloalkanes result in two acyl complexes. The first four categories of alkenes gave rise to slightly different spectral wavenumbers and relative intensities for the complexes, namely, cycloalkenes {2109 (0.41), 2063 (0.46), 2037 (0.72), 2019 (1.0), 1699 cm^-1 (0.16)}, symmetric internal linear alkenes {2108 (0.43), 2061 (0.45), 2037 (0.84), 2019 (1.0), 1693 cm^-1 (0.12)}, terminal alkenes {2110 (0.35), 2064 (0.46), 2038 (0.72), 2020 (1.0), 1703 cm^-1 (0.16)}, and methylene cycloalkanes {2110 (0.33), 2064 (0.46), 2038 (0.72), 2020 (1.0), 1704 cm^-1 (0.24)}. Finally, the approximate turnover frequencies (TOF) for each system were also calculated. It was found that the TOFs vary from 0.04 to 0.11 min^-1 between alkene categories. Thus, to a first approximation, the primary differences in rates of hydroformylation are due to the conversion of Rh₄(CO)₁₂ and not TOFs. This answers a long-standing question concerning hydroformylation rates.

Full text of DOI:10.1021/om990230q

Organometallics 1999, 18, 3429-3436
3429
Un m od ified Rh od iu m -Ca ta lyzed Hyd r ofor m yla tion of
Alk en es Usin g Tetr a r h od iu m Dod eca ca r bon yl. Th e
In fr a r ed Ch a r a cter iza tion of 15 Acyl Rh od iu m
Tetr a ca r bon yl In ter m ed ia tes†
Guowei Liu, Romeo Volken, and Marc Garland*,‡
Laboratorium fu¨r Technische Chemie, Universita¨tstrasse 6, ETH-Zu¨rich Eidgeno¨ssische
Technische Hochschule, CH-8092 Zu¨rich, Switzerland
Received April 5, 1999
The homogeneous catalytic hydroformylation of 20 alkenes was studied, starting with
Rh4(CO)12 as catalyst precursor in n-hexane as solvent, using high-pressure in-situ infrared
spectroscopy as the analytical tool. Five categories of alkenes were studied, namely,
cycloalkenes (cyclopentene, cycloheptene, cyclooctene, and norbornene), symmetric internal
linear alkenes (3-hexene, 4-octene, and 5-decene), terminal alkenes (1-hexene, 1-octene,
1-decene, 1-dodecene, and 1-tetradecene), methylene cycloalkanes (methylene cyclopropane,
methylene cyclobutane, methylene cyclopentane, and methylene cyclohexane), and branched
alkenes (2-methyl-2-butene, 2-methyl-2-pentene, 2-methyl-2-heptene, and 2,3-dimethyl-2-
butene). The typical reaction conditions were T ) 293 K, PH2 ) 2.0 MPa (0.018 mol fraction),
PCO ) 2.0 MPa (0.033 mol fraction), [alkene]0 ) 0.1-0.02 mol fraction, and [Rh4(CO)12]0 )
6.6 × 10-5 mol fraction. In each experiment, with the exception of those involving methylene
cyclopropane and the branched alkenes, the precursor Rh4(CO)12 was converted in good yield
to the corresponding observable mononuclear acyl rhodium tetracarbonyl intermediate
RCORh(CO)4. Due to the spectral characteristics, the intermediate RCORh(CO)4 is assigned
a trigonal bipyrimidal geometry in all cases with Cs symmetry, with the acyl group taking
an axial position. Under the present conditions, the cycloalkenes result in one acyl complex,
the symmetric internal linear alkenes result in two acyl stereoisomers, the terminal alkenes
result in three acyl complexes (two are stereoisomers), and the methylene cycloalkanes result
in two acyl complexes. The first four categories of alkenes gave rise to slightly different
spectral wavenumbers and relative intensities for the complexes, namely, cycloalkenes {2109
(0.41), 2063 (0.46), 2037 (0.72), 2019 (1.0), 1699 cm-1 (0.16)}, symmetric internal linear
alkenes {2108 (0.43), 2061 (0.45), 2037 (0.84), 2019 (1.0), 1693 cm-1 (0.12)}, terminal alkenes
{2110 (0.35), 2064 (0.46), 2038 (0.72), 2020 (1.0), 1703 cm-1 (0.16)}, and methylene
cycloalkanes {2110 (0.33), 2064 (0.46), 2038 (0.72), 2020 (1.0), 1704 cm-1 (0.24)}. Finally,
the approximate turnover frequencies (TOF) for each system were also calculated. It was
found that the TOFs vary from 0.04 to 0.11 min-1 between alkene categories. Thus, to a
first approximation, the primary differences in rates of hydroformylation are due to the
conversion of Rh4(CO)12 and not TOFs. This answers a long-standing question concerning
hydroformylation rates.
In tr od u ction
extensive catalytic and kinetic work performed by the
group at Veszpre´m,6 the discovery of [Rh(CO)2(µ-O2-
CR)]2 from a sample of the hydroformylation reaction,7
Chini’s famous stoichiometric hydroformylations of pro-
The unmodified homogeneous rhodium-catalyzed hy-
droformylation of alkenes has a rather long history.
Some of the most pivotal developments have included
the first patents, which appeared in the late1950s,1 the
first syntheses/discoveries of the rhodium carbonyls
HRh(CO)4,2 Rh2(CO)8,3 Rh4(CO)12,4 and Rh6(CO)16,5 the
(2) Vidal, J . L.; Walker, W. E. Inorg. Chem. 1981, 20, 249.
(3) (a) Whyman, R. J . Chem. Soc., Dalton Trans. 1972, 1375. (b)
Oldani, F.; Bor, G. J . Organomet. Chem. 1983, 246, 309.
(4) (a) Chini, P.; Heaton, B.T. Tetranuclear Carbonyl Clusters. Top.
Curr. Chem. 1977, 71, 1. (b) Martinengo, S.; Giordano, G.; Chini, P.
Inorg. Synth. 1980, 20, 209.
(5) (a) Chini, P. J . Chem. Soc., Chem. Commun. 1967, 440. (b)
J ames, B. R.; Rempel, G. L.; Teo, W. K. Inorg. Synth. 1976, 16, 49.
(6) (a) Marko, L. In Aspects of Homogeneous Catalysis; Ugo R., Ed.;
Reidel: Dorrecht, 1974; Vol II. (b) Heil, B.; Marko, L. Chem. Ber. 1968,
101, 2209. (c) Heil, B.; Marko, L. Chem. Ber. 1969, 102, 2238. (d)
Csontos, G.; Heil, B.; Marko, L. Ann. N. Y. Acad. Sci. 1974, 239, 47.
(e) Toros, S. Magy. Kem. Lapja 1974, 29, 543; Chem Abstr. 1975, 83,
177 870.
* To whom correspondence should be sent. E-mail: chemvg@
nus.edu.sg.
† Experimental research performed at ETH-Zurich, 1992-1995.
Analysis performed at NUS, 1998.
‡ Present address: Department of Chemical and Environmental
Engineering, The National University of Singapore.
(1) (a) Schiller, G. (Chem. Verwertungsges. Oberhausen) Ger. Pat.
965.605, 1956; Chem Abstr. 1959, 53, 11226. (b) Hughes, V. L. (Esso
Res. Eng. Co) U.S. Pat. 2.880.241, 1959; Chem. Abst. 1959, 53, 14938.
(c) Hughes, V. L. (Esso Res. Eng. Co.) Brit. Pat. 801.734, 1958; Chem.
Abst. 1959, 53, 7014.
(7) Heil, B.; Marko, L; Bor, G. Chem. Ber. 1971, 104, 3418.
10.1021/om990230q CCC: $18.00 © 1999 American Chemical Society
Publication on Web 08/16/1999

3430 Organometallics, Vol. 18, No. 17, 1999
Liu et al.
pylene using Rh4(CO)12 under hydrogen to give the
aldehydes and Rh6(CO)16,8 and the H2/D2 scrambling
experiments of Pino and Consiglio using Rh4(CO)12
under catalytic conditions.9 The overall stoichiometry
of the hydroformylation reaction is illustrated by eq 1,
where cyclopentene is the substrate.
Over the past decade, numerous in-situ infrared
spectroscopic and kinetic studies of the unmodified
homogeneous rhodium-catalyzed hydroformylation of
alkenes have been performed. These studies have
focused on a number of issues. These have included (i)
the extension of quantitative in-situ measurements to
the ppm level under high pressure,13 (ii) the identifica-
tion of the observable intermediate acyl rhodium tet-
racarbonyl during the syntheses14 [also see eq 2], (iii)
(1)
These early results clearly indicated some of the
particularly troublesome problems present in studies of
the unmodified rhodium-catalyzed hydroformylation
reaction. One problem was the extraordinarily rapid
interconversion of rhodium carbonyl species under syn-
gas. Thus sampling of the reaction mixtures usually
indicated Rh4(CO)12 and Rh6(CO)16 as the only rhodium
species present. A second problem was the extreme
sensitivity of the unmodified rhodium-catalyzed hydro-
formylation to diene and alkyne impurities.10 A third
problem was the sensitivity of the syntheses to trace
oxygen and water. Clusterification, particularly to Rh6-
(CO)16, readily occurred, and system deactivation was
the result.11 Taken together, these problems affected the
reproducibility of rhodium hydroformylation kinetics
and made analyses very difficult.
(2)
quantifying the activity of the systems in terms of the
instantaneous turnover frequencies based on the con-
centrations of intermediates and thereby separating
induction, deactivation, and product formation kinet-
ics,15 (iv) the identification of equilibrium precursor
conversion as a serious issue concerning metal utiliza-
tion,15b (v) statistical treatments of turnover frequencies
and studies of the large variance in such kinetic
measurements,16 (vi) in-situ spectroscopic and kinetics
investigations of heterometallic clusters and inorganic
precursors in the rhodium-catalyzed hydroformylation,17
(vii) the detailed kinetic analysis of the regioselective
hydroformylation of styrene,15c and (viii) statistical tests
aimed to resolve the possible existence and contribution
of a catalytic binuclear elimination reaction.18 With the
aforementioned issues addressed, we have shifted our
attention to a few remaining outstanding issues. In
particular, we have been concerned with the chemistry
associated with deactivation, especially with regard to
the use of alkynes and dienes.19
Despite all these complications, a viable mechanistic
explanation developed. This mechanism suggested the
involvement of only mononuclear rhodium intermedi-
ates in the transformation of alkenes to aldehydes, a
simple unicyclic reaction topology, and the existence of
a mononuclear acyl species as the predominant inter-
mediate. The strongest support for this assertion came
from Marko’s kinetic experiments with cyclohexene,
where it was shown that the activity was proportional
to [Rh4(CO)12]0.25 and [Rh6(CO)16]0.17 6d
Bor’s identifica-
,
tion of [Rh(CO)2(µ-O2-CR)]2,7 and the similarities in H2/
D2 scrambling between rhodium and cobalt hydroformy-
lation systems.9 The Veszpre´m group also showed the
interesting result that the reactivity of the systems
followed a reproducible systematic trend, namely, sty-
rene > terminal alkenes > internal alkenes > cyclic
alkenes.
In the present contribution we address issues sur-
rounding the formation of the acyl rhodium tetracar-
bonyl intermediate in hydroformylation reactions and
the approximate comparative turnover frequencies. In
other words, is the acyl rhodium tetracarbonyl the only
observable intermediate for all major categories of
alkene hydroformylations, and is the relative reactivity
In addition, it can noted that the unmodified rhodium-
catalyzed hydroformylation exhibited two very nice
features compared to the unmodified cobalt-catalyzed
hydroformylation. These features were the strong sup-
pression of hydrogenation products and the strong
suppression of CdC bond isomerization, particularly
under mild reaction conditions.12 Both of these proper-
ties allow potentially more straightforward kinetic
analyses.
(12) (a) Pino, P.; Piacenti, F.; Bianchi, M. In Organic Syntheses via
Metal Carbonyls Vol II; Wender, I., Pino. P., Eds.; Wiley: New York,
1977. (b) Dickson, R. S. Homogeneous Catalysis with Compounds of
Rhodium and Iridum; Reidel: Dordecht, 1986.
(13) Garland, M Dissertation 8585, ETH-Zurich, 1988.
(14) Garland, M., Bor, G. Inorg. Chem. 1989, 28, 410.
(15) (a) Garland, M.; Pino, P. Organometallics 1991, 10, 1693. (b)
Fyhr, Ch; Garland, M. Organometallics 1993, 12, 1753. (c) Feng, J .;
Garland, M.; Organometallics 1999, 18, 417.
(8) Chini, P.; Martinengo, S.; Garlaschelli, G. J . Chem. Soc., Chem.
Commun. 1972, 709.
(9) Pino, P.; Oldani, F.; Consiglio, G. J . Organomet. Chem. 1983,
250, 491.
(10) Cornils, B. In New Syntheses with Carbon Monoxide; Falbe, J .,
Ed.; Springer: Berlin, 1980.
(16) Shirt, R.; Garland, M.; Rippin, D. W. T. Anal. Chim. Acta 1998,
374, 67.
(17) Garland, M. Organometallics 1993, 12, 535.
(11) Kagan, Y. B.; Slivinskii, E. V.; Kurkin, V. I.; Korneeva, G. A.;
Aranovich, R. A.; Fal′kov, I. G.; Rzhevskaya, N.; Loktev, S. M.
Neftekhimiya 1985, 25, 791; Chem. Abstr. 1986, 104, 56974d.
(18) Feng, J .; Garland, M. Organometallics 1999, 18, 1542.
(19) (a) Liu, G.; Garland, M. Organometallics, in press. (b) Liu, G.;
Garland, M. In preparation.

Rh-Catalyzed Hydroformylation of Alkenes
Organometallics, Vol. 18, No. 17, 1999 3431
cells described in the literature (for a review, see Whyman24).
The high-pressure cell was situated in a Perkin-Elmer 983
infrared spectrometer. The resolution was set to 4 cm-1 for
all spectroscopic measurements. A schematic diagram of the
experimental setup can be found in ref 15c.
of alkenes due to turnover effects or due to systematic
changes in the conversion of the rhodium precursor?
Exp er im en ta l Section
In -Situ Sp ectr oscop ic a n d Kin etic Stu d ies. The experi-
ments were performed in a similar manner. First, single-beam
background spectra of the IR sample chamber were recorded.
Then 5-20 mL of alkene was dissolved in 150 mL of n-hexane,
and this solution was transferred under argon to the autoclave.
Under 0.2 MPa CO pressure, infrared spectra of the alkene
solution in the high-pressure cell were recorded. The total
system pressure was raised to 2.0 MPa CO, and the stirrer
and high-pressure membrane pump were started. After equili-
bration, infrared spectra of the alkene/n-hexane/CO solution
in the high-pressure cell were recorded. A solution of 100 mg
of Rh4(CO)12 dissolved in 100 mL of n-hexane was prepared,
transferred to the high-pressure reservoir under argon, pres-
sured with CO, and then added to the autoclave. Infrared
spectra of the alkene/n-hexane/CO/Rh4(CO)12 solution in the
high-pressure cell were recorded. Hydrogen (2.0 MPa) was
then added to initiate the syntheses. Spectra were recorded
at 15 min intervals in the range 1600-2200 and 1100-1300
Gen er a l In for m a tion . All solution preparations were
carried out under argon (99.999% Pan Gas AG, Luzern,
Switzerland) using standard Schlenk techniques.20 The argon
was further purified prior to use by passage through a column
containing 100 g of reduced BTS-catalyst (Fluka AG Buchs,
Switzerland) and 100 g of 4 Å molecular sieves to adsorb trace
oxygen and water, respectively. All reactions were carried out
under carbon monoxide (99.997% Messner Griesheim GmbH,
Germany) and hydrogen (99.999% Pan Gas AG, Luzern,
Switzerland) after further purification through deoxy and
zeolite columns.
The precious metal complex Rh4(CO)12, with stated purity
of 98% min was obtained from Strem Chemicals SA (Bis-
chheim, France) and was used without further purification,
although trace quantities of the high nuclearity cluster
Rh6(CO)16 are virtually always present. The complex Rh4(CO)12
is known to be oxygen, water, and light sensitive.4 The
n-hexane solvent (stated purity >99.6%, Fluka AG) was
refluxed over sodium potassium alloy under argon. Weights
were measured with a precision of 0.1 mg. Volumes were
measured with a precision of 0.045 mL. Further microanalytic
techniques were not employed.21 All alkenes used in this study
were of the highest quality commercially obtainable, usually
99.0+% (obtained either from Wiley Organic Chemicals, Lan-
caster Chemicals, U.K., Fluka AG, Switzerland, or Merck,
Germany). Concerning further purification, the alkenes were
simply degassed before use.
Equ ip m en t. In-situ spectroscopic studies were performed
in a 1.5 L stainless steel (SS316) autoclave (Bu¨chi-Uster,
Switzerland), which was connected to a high-pressure infrared
cell. The autoclave (Pmax ) 22.5 MPa) was equipped with a
packed magnetic stirrer with six-bladed turbines in both the
gas and liquid phases (Autoclave Engineers, Erie, PA) and was
constructed with a heating/cooling mantle. A high-pressure
membrane pump (Model DMK 30, Orlita AG, Geissen, Ger-
many) with a maximum rating of 32.5 MPa and a 3 L/h flow
rate was used to circulate the n-hexane solutions from the
autoclave to the high-pressure IR cell and back to the autoclave
via jacketed 1/8 in. (SS316) high-pressure tubing (Autoclave
Engineers). The entire system, autoclave, transfer lines, and
infrared cell, was cooled using a Lauda cryostat Model RX20
and could be maintained isothermal (∆T e 0.5 °C) at 298-
313 °C. Temperature measurements were made at the cry-
ostat, autoclave, and IR cell with PT-100 thermoresistors. The
necessary connections to vacuum and gases were made with
1/4 in. (SS316) high-pressure tubing (Autoclave Engineers),
and 1.0, 5.0, and 10.0 piezocrystals were used for pressure
measurements (Keller AG Winterthur, Switzerland). The
entire system was gastight under vacuum as well as at 20.0
MPa, the maximum operating pressure.
cm-1
.
A considerable number of spectral subtractions were per-
formed on each reaction spectrum in order to subtract the
absorbance of the n-hexane solvent, dissolved CO, the absor-
bance of the Rh4(CO)12, the absorbance of the alkenes, and the
absorbance of the aldehydes.
The solubility of CO under these reaction conditions was
approximately 0.033 mole fraction, and the solubility of H2 was
approximately 0.018 mole fraction.17 Consequently, the moles
in each experiment were 1.51 mol of n-hexane, 0.05 mol of CO,
and 0.03 mol of H2. Since the amount of Rh4(CO)12 used in the
experiments was ca. 100 mg, and since the maximum amount
of alkenes was ca. 20 mL, the corresponding in-situ measured
initial concentrations of Rh4(CO)12 in the experiments was ca.
6.6 × 10-5 mole fraction.
The present hydroformylation reactions were performed
under negligible gas-liquid mass transfer resistance. The
experimentally measured overall mass transfer coefficients
KLa for hydrogen and carbon monoxide into n-hexane at 200
rpm were approximately 0.1 and 0.6 s-1, respectively, as
determined using the method of Deimling.25 Since the maxi-
mum observed rate of hydroformylation in this study was 4 ×
10-7 mol s-1, all experiments belong to the category of infinitely
slow reaction with respect to gas-liquid mass transfer (i.e.,
the kinetic regime, Hatta category H).26 The liquid phase of
each experiment became essentially saturated with dissolved
CO and H2 in the first 60 s. Mass transfer effects are known
to severely complicate the interpretation of kinetic data from
hydroformylation reactions.27
In any single 8 h kinetic experiment, less than 0.01 mol
conversion (or 3%) of alkene to aldehyde occurred. Therefore,
the partial pressures of hydrogen and carbon monoxide in the
closed-batch autoclave changed less than 1% during the each
8 h experiment. This partial pressure change was considered
negligible, and hence, the liquid-phase concentrations of the
two gaseous components were treated as constants for the
duration of each experiment.
The high-pressure infrared cell was constructed at the ETH-
Zu¨rich of SS316 steel and could be heated and cooled. The CaF2
single-crystal windows (Korth Monokristalle, Kiel, Germany)
had dimensions of 40 mm diameter by 15 mm thickness. Two
sets of Viton and silicone gaskets provided sealing, and Teflon
spacers were used between the windows. The construction of
the flow-through cell22 is a variation on a design due to Noack23
and differs in some respects from other high-pressure infrared
The approximate mole fractions of the acyl rhodium tetra-
carbonyl species RCORh(CO)4 and the aldehydes were deter-
(24) Whyman, R. In Laboratory Methods in Vibrational Spectros-
copy, 3rd ed.; Willis H. A., van der Maas, J . H., Miller, R. G. J ., Eds.;
Wiley: New York, 1987; Chapter 12.
(25) Deimling, A.; Karandikar, B. M.; Shah, Y. T.; Carr, N. L. Chem.
Eng. J . 1984, 29, 127.
(26) Levenspiel, O. Chemical Reaction Engineering; Wiley: New
York, 1972; p 418.
(20) Shriver, D. F..; Drezdzon, M. A. The Manipulation of Air-
Sensitive Compounds; Wiley: New York, 1986.
(21) (a) Cheronis, N.D. Micro and Semimicro Methods; Inter-
science: New York, 1954. (b) Ma, T.S.; Horak, V. Microscale Manipula-
tions in Chemistry; Wiley: New York, 1976. (c) Kirk, P. L. Quantitative
Ultramicro Analysis; Wiley: New York, 1954.
(22) Dietler, U. K. Dissertation 5428, ETH-Zurich, 1974.
(23) Noack, K. Spectrochim. Acta 1968, 24A, 1917.
(27) Bhattacharya, A.; Chaudari, R. V. Ind. Eng. Chem. Res. 1987,
26, 1168.

3432 Organometallics, Vol. 18, No. 17, 1999
Liu et al.
mined by measuring the absorbance at 2020 cm-1 and at the
aldehyde maximum, and by using the absorptivity 2500 L/(mol
cm) at 2020 cm-1, the absorptivity 375 L/(mol cm) for alde-
hydes, and the absorptivity 1.95 L/(mole cm) at 1138 cm-1 for
hexane. Calculations were made using the dimensionless
Lambert-Beer law (eq 3).28 The term xC6 is the mole fraction
of n-hexane at the initial reaction conditions.
xi ) xC6(Aiꢀ1138)/(A1138ꢀi)
(3)
The rates of reaction for the transformation of alkenes to
aldehydes were calculated from the concentrations using a
central difference approximation. The central difference ap-
proximation was chosen since it provides an accurate ap-
proximation of derivatives from sets of smooth monotonically
increasing or decreasing experimental data.29 In addition,
extensive investigations concerning the numerical evaluation
of turnover frequencies in homogeneous catalytic systems
support the central difference approximation.16 Turnover
frequencies were calculated using the formula
∆[aldehyde]t
TOFt )
(4)
[acyl]t∆t
Resu lts
Cycloa lk en es. Four cycloalkenes were used in these
experiments: cyclopentene, cycloheptene, cyclooctene,
and norbornene. The first three cyclic alkenes result in
only one aldehyde each, namely, cyclopentane carbox-
aldehyde, cycloheptane carboxaldehyde, and cyclooctane
carboxaldehyde. Therefore, only one acyl intermediate
is expected in the corresponding hydroformylation
experiments. However, these aldehydes occur in more
than one conformation; for example, in the case of
cyclohexane carboxaldehyde, both the chair and boat
conformations exist. In the case of norbornene, two
different product isomers occur, namely, the endo and
exo configurations. Consequently, in these hydroformy-
lations, two acyl isomers also occur.
Only ca. 50% conversion of the precursor Rh4(CO)12
was observed in the first 8 h of the hydroformylation
experiments. The mid-infrared spectrum of each acyl
product is shown in Figure 1. It is immediately seen
that the acyl rhodium tetracarbonyls have very similar
spectral characteristics, namely, four terminal CO’s and
one “organic” CO vibration near 1699 cm-1. In these
spectra, the presence of Rh6(CO)16 can be observed in
both the cycloheptene spectrum and the cyclooctene
spectrum at 2073 cm-1. In addition a very weak band
can be observed at ca. 1990 cm-1 in all spectra.
F igu r e 1. Mid-infrared spectra of acyl rhodium tetracar-
bonyl complexes derived from cyclic alkenes after subtrac-
tion of Rh4(CO)12, hexane, CO, alkene, and aldehyde.
Ta ble 1. Rela tive In fr a r ed Absor ba n ce of Acyl
Rh od iu m Tetr a ca r bon yls Der ived fr om Cyclic
Alk en es
ν
CO (cm-1) above,
relative intensity below
νCO
alkene
(cm-1
)
cyclopentene 2109.9
2063.1
0.44
2063.0
0.47
2063.2
0.47
2061.8
0.45
2037.3
0.70
2036.5
0.69
2036.0
0.72
2037.1
0.74
2019.6 1698.2
0.40
cycloheptene 2109.0
0.42
2108.8
0.42
2108.6
0.40
1
0.18
2018.1 1703.2
1
0.13
cyclooctene
norbornene
2017.1 1699.3
1
0.16
2019.8 1696.4
1
0.17
larger than normal. This is probably due to the numer-
ous conformations possible.
The region between 2200 and 2100 cm-1 is noisy due
to the subtraction of dissolved CO. In the spectrum of
the cyclooctene-derived acyl, a sigmoid waveform is seen
at 1885 cm-1 after subtraction of the reference Rh4-
(CO)12. Finally, the “organic” CO vibration is symmetrics
indicating only one isomersat least in the cases of
cyclopentene, cycloheptene, and cyclooctene. The result-
ing aldehyde bands were symmetric and occurred at ca.
1732.8 cm-1, and the width at half-height (WHH) was
about 12 cm-1. It is noted that this WHH is slightly
The detailed spectroscopic characteristics of the acyl
complexes are given in Table 1. The average wavenum-
bers for the all acyl complexes in this category were 2109
(0.41), 2063 (0.46), 2037 (0.72), 2019 (1.0), and 1699
cm-1 (0.16), where the relative intensities are given in
parentheses. It is clear from numerical data that the
spectroscopic characteristics are very similar for each
acyl complex in both the terminal CO region and the
“organic” CO region.
The measured approximate turnover frequencies for
the four systems were as follows: TOF cyclopentene )
0.06 min-1, TOF cycloheptene ) 0.10 min-1, TOF
cyclooctene ) 0.20 min-1, and TOF norbornene ) 0.07
min-1. Thus, the average TOF for cyclic alkenes was
(28) McClure, G. L. In Laboratory Methods in Vibrational Spectros-
copy, 3rd ed.; Willis, H. A., van der Maas, J . H., Miller, R. G. J ., Eds.;
Wiley: New York, 1987; Chapter 7.
(29) Davies, M. E. Numerical Methods and Modelling for Chemical
Engineers; Wiley: New York, 1984.
(30) Volken, R. Diplomarbeit # 4187, ETH-Zurich, 1994.
ca. 0.11 min-1
.

Rh-Catalyzed Hydroformylation of Alkenes
Organometallics, Vol. 18, No. 17, 1999 3433
Ta ble 2. Rela tive In fr a r ed Absor ba n ce of Acyl
Rh od iu m Tetr a ca r bon yls Der ived fr om Sym m etr ic
In ter n a l Lin ea r Alk en es
ν
CO (cm-1) above,
relative intensity below
νCO
alkene
(cm-1
)
3-hexene
2108.9
0.41
2108.2
0.45
2108.2
0.41
2062.3
0.45
2060.9
0.45
2060.9
0.46
2037.3
0.81
2037.3
0.83
2037.3
0.87
2019.2
1696.7
0.13
1690.4
0.12
1690.0
0.12
1
2019.2
1
2019.2
1
4-octene
5-decene
symmetric and occurred at ca. 1732.0 cm-1, and the
width at half-height (WHH) was only 10 cm-1
.
The average frequencies for the resulting acyl com-
plexes were 2108 (0.43), 2061 (0.45), 2037 (0.84), 2019
(1.0), and 1693 cm-1 (0.12), where the relative intensi-
ties are given in parentheses. Detailed spectroscopic
characteristics are given in Table 2. It is clear from the
numerical data that the spectroscopic characteristics are
very similar for each acyl complex.
The measured approximate turnover frequencies for
the three systems were as follows: TOF 3-hexene ) 0.05
min-1, TOF 4-octene ) 0.07 min-1, and TOF 5-decene
) 0.11 min-1. Thus, the average TOF for symmetric
linear alkenes was ca. 0.08 min-1
.
Ter m in a l Lin ea r Alk en es. Five alkenes were used
in these experiments: 1-hexene, 1-octene, 1-decene,
1-dodecene, and 1-tetradecene. Each of these alkenes
results in three aldehydes. For example, in the case of
1-hexene, the expected products are 1-heptanal and rac-
2-methylhexanal.
The reactivity of the rhodium precursor Rh4(CO)12
toward these terminal alkenes was much greater than
that observed for the aforementioned internal alkenes.
In the present series of experiments, nearly 100%
conversion of the precursor Rh4(CO)12 was observed in
the first 8 h of the hydroformylation experiments. The
mid-infrared spectrum of each set of acyl products is
shown in Figure 3. It is seen that all the acyl rhodium
tetracarbonyls have very similar spectral characteris-
tics, namely, the four terminal CO’s and one “organic”
CO vibration. In these spectra, the presence of Rh6-
(CO)16 can be observed in all cases due to the bands
present at both 2073 and 1820 cm-1. Again a very weak
band at ca. 1990 cm-1 is observed in all spectra, and
the region between 2200 and 2100 cm-1 is noisy due to
the subtraction of dissolved CO. The “organic” CO
vibration is clearly seen at ca. 1703 cm-1 for all alkenes,
1-hexene, 1-octene, 1-decene, 1-dodecene, and 1-tet-
radecene, and the “organic” CO vibration is clearly
asymmetric (the asymmetry is certainly not an artifact
of the spectral subtractions). This indicates that the
primary maximum at 1703 cm-1 is due to coordination
at the R carbon atom, and the relatively weak sideband
at 1694 is due to coordination at the â carbon atom.
Therefore we observe all three possible acyl isomers. The
resulting aldehyde bands were visibly asymmetric, at
least in the cases of 1-decene, 1-dodecene, and 1-tet-
radecene. This indicates the formation of both the linear
and branched aldehydes in observable quantities. The
typical aldehyde band had a center at ca. 1734.8 cm-1
and a width at half-height (WHH) of about 14 cm-1. This
very broad WHH is further confirmation of the presence
of both linear and branched aldehydes under these
reaction conditions.
F igu r e 2. Mid-infrared spectra of acyl rhodium complexes
derived from symmetric internal linear alkenes after
subtraction of Rh4(CO)12, hexane, CO, alkene, and alde-
hyde.
Sym m et r ic In t er n a l Lin ea r Alk en es. Three al-
kenes were used in these experiments: 3-hexene,
4-octene, and 5-decene.30 These alkenes result in two
aldehydes each, namely, rac-2-ethylpentanal, rac-2-n-
propylhexanal, and rac-2-n-butylheptanal. Therefore,
two stereoisomeric acyl rhodium tetracarbonyl inter-
mediates are produced in each of the corresponding
hydroformylation experiments. However, using unpo-
larized infrared radiation, each pair of racemic alde-
hydes and each pair of racemic acyl complexes have
exactly the same infrared absorbance spectrum.
Again, only ca. 50% conversion of the precursor Rh4-
(CO)12 was observed in the first 8 h of the hydroformy-
lation experiments. The mid-infrared spectrum of each
acyl product is shown in Figure 2. [Please note that the
uneven quality of this figure is due to the redigitaliza-
tion of the data from hardcopies.] It is immediately seen
that the three racemic acyl rhodium tetracarbonyls have
very similar spectral characteristics, namely, four ter-
minal CO’s and one “organic” CO vibration near 1693
cm-1. In these spectra, the presence of Rh6(CO)16 can
be observed in both the 4-octene and 5-decene spectra
at 2073 cm-1. In addition the very weak band at ca. 1990
cm-1 is again observed in all spectra, and the region
between 2200 and 2100 cm-1 is particularly noisy due
to the subtraction of dissolved CO. Finally, the “organic”
CO band is symmetric, indicating exact superposition
of the isomers. The resulting aldehyde bands were

3434 Organometallics, Vol. 18, No. 17, 1999
Liu et al.
F igu r e 4. Mid-infrared spectra of acyl rhodium complexes
derived from methylene cycloalkanes after subtraction of
hexane, CO, alkene, and aldehyde.
F igu r e 3. Mid-infrared spectra of acyl rhodium complexes
derived from terminal linear alkenes after subtraction of
hexane, CO, alkene, and aldehyde.
Ta ble 3. Rela tive In fr a r ed Absor ba n ce of Acyl
Rh od iu m Tetr a ca r bon yls Der ived fr om Ter m in a l
Lin ea r Alk en es
cyclopropane, methylene cyclobutane, methylene cyclo-
pentane, and methylene cyclohexane. In the latter three
cases, hydroformylation readily occurred. Unusual chem-
istry occurred only in the case of methylene cyclopro-
pane as substrate. In principle, each of the hydroformy-
lated alkenes could result in two aldehydes each. For
example, in the case of methylene cyclobutane, the
expected products are 2-cyclobutyl ethanal and 1-formyl-
1-methylcyclobutane.
Again, the reactivity of the rhodium precursor Rh4-
(CO)12 toward the methylene cycloalkanes was much
greater than that toward the internal alkenes. In the
present series of experiments, nearly 100% conversion
of the precursor Rh4(CO)12 was observed in the first 8
h of the hydroformylation experiments. The mid-
infrared spectrum of each set of acyl products is shown
in Figure 4. It is seen that all the sets of acyl rhodium
tetracarbonyls have very similar spectral characteris-
tics, namely, the four terminal CO’s and one “organic”
CO vibration. In these spectra, the presence of Rh6-
(CO)16 can be observed in all cases due to the bands
present at both 2073 and 1820 cm-1. Again a very weak
band at ca. 1990 cm-1 is observed in all spectra, and
the region between 2200 and 2100 cm-1 is noisy due to
the subtraction of dissolved CO. Finally, the “organic”
CO vibration is clearly seen at ca. 1704 cm-1, and in
the case of all three alkenes, the “organic” CO vibration
is clearly asymmetric. This indicates that the primary
maximum at 1704 cm-1 is due to coordination at the R
carbon atom, and the moderate sideband at 1696 cm-1
ν
CO (cm-1) above,
relative intensity below
νCO
alkene
(cm-1
)
1-hexene
2110.5
0.33
2110.5
0.34
2110.6
0.36
2110.3
0.38
1-tetradecene 2110.7
0.35
2064.2
0.45
2064.1
0.47
2064.5
0.48
2064.3
0.47
2064.2
0.46
2038.2
0.70
2038.0
0.72
2038.4
0.73
2038.0
0.69
2038.0
0.73
2020.2 1702.9
1
0.15
1-octene
2020.0 1703.1
1
0.16
1-decene
1-dodecene
2020.4 1703.2
1
0.16
2020.2 1702.7
1
0.16
2020.1 1702.8
1
0.16
The average wavenumbers for the resulting mixtures
of acyl complexes were 2110 (0.35), 2064 (0.46), 2038
(0.72), 2020 (1.0), and 1703 cm-1 (0.16), where the
relative intensities are given in parentheses. The de-
tailed spectroscopic characteristics are given in Table
3. It is clear from the numerical data that the spectro-
scopic characteristics are very similar for the acyl
complexes.
The measured approximate turnover frequencies for
the five systems were as follows: TOF 1-hexene ) 0.01
min-1, TOF 1-octene ) 0.11 min-1, TOF 1-decene ) 0.05
min-1, TOF 1-dodecene ) 0.05 min-1 and TOF 1-tet-
radecene ) 0.05 min-1 . Thus, the average TOF for the
terminal alkenes was ca. 0.05 min-1
.
Meth ylen e Cycloa lk a n es. Four methylene cycloal-
kanes were used in these experiments: methylene

Rh-Catalyzed Hydroformylation of Alkenes
Organometallics, Vol. 18, No. 17, 1999 3435
Ta ble 4. Rela tive In fr a r ed Absor ba n ce of Acyl
Rh od iu m Tetr a ca r bon yls Der ived fr om Meth ylen e
Cyclo Alk a n es
ν
CO (cm-1) above,
relative intensity below
νCO
alkene
(cm-1
)
methylene
cyclobutane
methylene
cyclopentane
methylene
cyclohexane
2109.9
0.31
2110.7
0.33
2110.3
0.36
2064.3
0.47
2064.8
0.45
2063.9
0.47
2038.3
0.70
2038.2
0.73
2037.8
0.73
2020.4 1703.8
1
0.23
2020.1 1704.0
1
0.25
2020.2 1704.5
1
0.24
is due to coordination at the â carbon atom. Therefore
we observe both possible acyl complexes. The resulting
aldehyde bands were visibly asymmetric, particularly
in the case of methylene cyclohexane. The latter alde-
hyde had a maximum at 1733.5 cm-1, a strong shoulder
at 1730 cm-1, and a width at half-height (WHH) of about
13 cm-1
.
The average frequencies for the resulting mixtures
of acyl complexes were 2110 (0.33), 2064 (0.46), 2038
(0.72), 2020 (1.0), and 1704 (0.24), where the relative
intensities are given in parentheses. The detailed
spectroscopic characteristics are given in Table 4. It is
clear from the numerical data that the spectroscopic
characteristics are very similar for each of the acyl
complexes.
The measured approximate turnover frequencies for
the 3 systems were as follows: TOF methylene cyclobu-
tane ) 0.04 min-1, TOF methylene cyclopentane ) 0.04
min-1, and TOF methylene cyclohexane ) 0.04 min-1
.
Thus, the average TOF for methylene cycloalkanes was
ca. 0.04 min-1
.
F igu r e 5. Mid-infrared spectra of the reaction of Rh4(CO)12
with methylene cyclopropane under hydroformylation con-
ditions: (A) 4.0 MPa CO, 1.0 MPa H2, (B) 2.0 MPa CO and
1.0 MPa H2, (C) 1.0 MPa CO and 1.0 MPa H2 with best
curve fit, (D) deconvolution of spectra.
Resu lts w ith Meth ylen e Cyclop r op a n e. The reac-
tion of methylene cyclopropane with Rh4(CO)12 under
hydroformylation conditions gave unexpected results.
Five different experiments were performed at 20 °C with
250 mL of n-hexane solvent, 5 mL of methylene cyclo-
propane, and 100 mg of Rh4(CO)12. The partial pressures
of CO and H2 used were 0.5 MPa/1.0 MPa, 1.0 MPa/1.0
MPa, 2.0 MPa/1.0 MPa, 2.0 MPa/2.0 MPa, and 4.0 MPa/
1.0 MPa. In each of the five experiments, essentially
the same infrared spectra at 2200-1950 cm-1 were
generated under reaction conditions. This region is very
complex and contains at least 15 predominant bands.
Figure 5 shows spectra from three experiments as well
as a typical deconvolution.
carbonyl complexes have been successfully used to break
cyclopropane structures.32 Finally, we note that there
is considerable similarity in the terminal CO region of
the present experiment and hydroformylations per-
formed with dienes.19b Indeed, if the methylene cyclo-
propane structure is destroyed at the â carbon atom,
followed by hydrogen migration, one obtains 1,3-buta-
diene. This issue will be treated in more detail else-
where.19b
Br a n ch ed Alk en es. The hydroformylations of the
following branched alkenes were also attempted: 2-meth-
yl-2-butene, 2-methyl-2-pentene, 2-methyl-2-heptene,
and 2,3-dimethylbut-2-ene. Most of these experiments
showed negligible conversion of the precursor Rh4(CO)12
and usually little or no product. Thus, it seems that the
present reaction conditions are not optimal for observing
intermediates.
Interestingly, there is a relatively weak band at ca.
1703 cm-1, which suggests an acyl complex, and a very
weak band at 1733 cm-1, indicative of the formation of
a small amount of the linear aldehyde. However, there
are not any bands at ca. 1695 nor 1730 cm-1. Together,
these results strongly suggest that a rhodium carbonyl
hydride can successfully attack the R carbon atom and
form an acyl complex. However, it appears that attack
at the â carbon atom results in ring opening. Ring
opening of the cyclopropane moiety would result in a
host of new reaction products. Indeed, the ring opening
of cyclopropanes is well documented, and rhodium
complexes are particularly suitable for this purpose.31
In the present context it can be noted that rhodium
Discu ssion
Acyl Rh od iu m Tetr a ca r bon yls. Previous to this
study, five acyl rhodium tetracarbonyls formed from 3,3-
dimethylbut-1-ene, cyclohexene, and styrene were spec-
troscopically characterized. In all cases, four terminal
CO vibrations and one “organic” CO vibration were
observed.13,14,15b,c,17
(31) Brandl, M.; Kozhushkov, S. I.; Brase, S.; de Meijere, A. Eur. J .
Org. Chem. 1938, 3, 453.
(32) (a) Chum, P. W.; Roth, J . A. J . Catal. 1975, 39, 198. (b) Powell,
K. G.; McQuillin, F. J . Chem. Commun. 1971 931. (c) Voigt, H. W.;
Roth, J . A. J . Catal. 1974, 33, 91.
In the present spectral results, from the first four
categories of alkenes (and with the exception of meth-

3436 Organometallics, Vol. 18, No. 17, 1999
Liu et al.
ylene cyclopopane), only one type of rhodium intermedi-
ate was ever observed. In all cases, only the acyl
rhodium tetracarbonyl could be identified, and no other
unusual spectral features were present. All the acyl
rhodium tetracarbonyls possessed four strong terminal
vibration is found at ca. 1715 cm-1 37
measured in CCl4 solvent.
.
All three were
Tu r n over F r equ en cies. As mentioned in the Intro-
duction, Marko et al. found a clear correlation between
the category of alkenes used in a hydroformylation and
the rate of aldehyde formation: styrene > terminal >
internal > branched > cyclic at ca. 70 °C. In the present
study, we found that the turnover frequencies follow the
trend cyclic alkenes > symmetric linear alkenes >
terminal alkenes > methylene cycloalkanes at 20 °C.
However, the TOF varied only a little from 0.04 to 0.11.
We also found that the degree of conversion of Rh4(CO)12
followed the trend terminal alkenes ≈ methylene cy-
cloalkanes > cyclic alkenes ≈ symmetric linear alkenes
> branched alkenes. Thus, it seems that TOF is not the
primary controlling factor for the rates of aldehyde
formation. Instead, it seems that the degree of precursor
conversion and hence the amount of acyl complex formed
is the most important factor with regard to the rates of
product formation between categories of alkenes.
Oth er In ter m ed ia tes. The present contribution
indicates that the acyl rhodium tetracarbonyl is the only
observable rhodium intermediate in hydroformylation
experiments conducted with four major categories of
unsubstituted simple alkenes. One final important
category of alkenes remains to be explored, namely,
ethylene and substituted ethylenes. In these experi-
ments, a new rhodium species is observed. This issue
will be taken up further in a later contribution.38
CO vibrations (and a small overtone at ca. 1990 cm-1
)
and one “organic” CO vibration. No bridging CO ligands
were observed. All the spectral data are consistent with
only one type of geometry, namely, a trigonal bipyramid
mononuclear species with an axial acyl group and with
Cs symmetry. As Bor has pointed out on many occasions,
the introduction of the acyl moiety causes a loss of C3v
symmetry in many mononuclear transition metal com-
plexes.33 The strong vibration at ca. 2108-2110 cm-1
,
which belongs to the totally symmetric in-phase stretch-
ing mode, is indicative of the low symmetry of this
mononuclear species.34
The small vibration that appears at ca. 1990 cm-1 is
almost certainly due to the 13CO-substituted isoto-
pomers. Indeed, this band has ca. 1% of the integral
absorbance of the most prominent terminal CO vibra-
tion, which appears at ca. 2020 cm-1. These data are
consistent with the natural abundance of 1.1% 13C and
a shift to lower wavenumbers of up to 45 cm-1 35
.
A correlation could be found between the highest cm-1
vibrations of the acyl complexes and the category of
alkene used. In the cases of the cycloalkenes and the
symmetric internal linear alkenes, the highest vibration
occurred at 2108-2109 cm-1, and in the cases of
terminal alkenes and methylene cycloalkanes, the high-
est vibration occurred at 2110-2111 cm-1
.
Con clu sion s
The “organic” CO vibration is also clearly correlated
with the nature of the adjacent carbon atom in the acyl
structure. Accordingly, for the terminal alkenes and the
methylene cycloalkanes, the primary acyl structure was
one in which the R carbon atom had been attacked and
the resulting adjacent carbon atom was a primary
carbon atom. This consistently gave a wavenumber of
1703-1704 cm-1. In the case of the internal alkenes,
namely, the cycloalkenes and the symmetric linear
alkenes, attack occurred exclusively at a secondary
carbon atom. These acyl structures consistently gave a
The in-situ infrared spectroscopic results from five
categories of alkenes showed only one observable rhod-
ium intermediate during unmodified homogeneous hy-
droformylations. This acyl rhodium tetracarbonyl was
consistently present in the active hydroformylation
experiments, and its mid infrared spectral pattern was
very similar between categories of alkenes. All acyl
rhodium tetracarbonyl intermediates observed in this
study have trigonal bipyramid geometry, with Cs sym-
metry and the acyl group in the axial position. As a
generalization, it appears that under most reactions
conditions, and using simple alkenes or simple aliphatic-
substituted alkenes, the acyl rhodium tetracarbonyl
complex is by far the predominant intermediate present
in solution. This implies that CO insertion is very rapid
and the possibility of observing an alkyl rhodium
tetracarbonyl is remote.
wavenumber of 1694-1696 cm-1
.
As a final note in this regard, it is well-known that
the vibrations of the CO moiety in aldehydes correlate
with the type of carbon atom adjacent to the carbonyl
group. Thus, in the case of a primary adjacent carbon
atom, the carbonyl vibration is found at ca. 1738 cm-1 36
,
in the case of a secondary adjacent carbon atom the
The approximate turnover frequencies for these sys-
tems were also evaluated. It was found that the TOFs
do not vary much. Between categories of alkenes, TOFs
of 0.04-0.11 min-1 were measured. The higher TOFs
were observed for the acyl complexes formed at second-
ary carbon atoms, and the lower TOFs were observed
for acyl complexes formed at primary carbon atoms. In
addition it was observed that the conversion of the
precursor Rh4(CO)12 was lower for the former and higher
for the latter systems. This strongly implies that the
acyl complexes formed from secondary carbon atoms
have a higher free energy than the acyl complexes
formed from primary carbon atoms. In other words the
acyl complexes formed at primary carbon atoms are
thermodynamically more stable.
carbonyl vibration is found at ca. 1729 cm-1 37
,
and in
the case of a tertiary adjacent carbon atom the carbonyl
(33) (a) Marko, L.; Bor, G.; Almasy, G.; Szabo, P. Brenstoff Chem.
1963, 44, 184. (b) Bor, G. Inorg. Chim. Acta 1967, 1, 81.
(34) Bor, G. Spectrochim. Acta 1963, 19, 1209.
(35) (a) Bor, G. Inorg. Chim. Acta 1967, 1, 81. (b) Bor, G. Chem.
Commun. 1969, 641. (c) Bor, G.; J ung, C. Inorg. Chim. Acta 1969, 3,
69. (d) Braterman, P. S.; Harrill, R. W.; Kaesz, H. D. J . Am. Chem.
Soc. 1967, 89, 2851. (e) Cotton, F. A.; Musco, A.; Yagupsky, G. Inorg.
Chem. 1967, 6, 1357. (f) Harrill, R. W.; Kaesz, H. D. J . Am. Chem.
Soc. 1968, 90, 1449. (g) Kaesz, H. D.; Bau, R.; Hendrickson, D.; Smith,
J . M. J . Am. Chem. Soc. 1968, 90, 1449. (h) Lewis, J .; Manning, A. R.;
Miller, J . R. J . Chem. Soc. A 1966, 843. (i) Loutellier, A.; Bigorgne, M.
Bull. Soc. Chim. Fr. 1965, 3186. (j) Noack, K. J . Organomet. Chem.
1968, 12, 181. (k) Poletti, A.; Catiotti, R.; Foffani, A. Inorg. Chim. Acta
1968, 2, 157.
(36) Cook, D. J . Am. Chem. Soc. 1958, 49, 9.
(37) Applequist, D. E.; Kaplan, L. J . Am. Chem. Soc. 1965, 2194.
(38) Liu, G.; Garland M. Manuscript in preparation.
OM990230Q