Activation parameters for the reactive intermediates relevant to carbonylation catalysts based on cobalt carbonyls

Article abstract of DOI:10.1021/ic020567b

Time-resolved spectroscopic techniques have been used to prepare and to interrogate transient species that are models for reactive intermediates in cobalt-catalyzed hydroformylation. Flash photolysis of acetylcobalt carbonyl complexes of the type RC(O)Co(CO)₃(PR′₃) (A; R = CH₃, CD₃, or C₂H₅; R′ = Ph or ⁿBu) leads to CO photodissociation to give the unsaturated intermediate [RC(O)Co(CO)₂(PR′₃)] (I), which decays by two competitive pathways, alkyl migration to the cobalt to give RCo(CO)₃PR′₃ (M) and reaction with CO to re-form A. With the perdeuterioacetyl complex (R = CD₃, R′ = Ph), rate constants both of CO trapping (k_CO) and of methyl migration (k_M) were just slightly smaller than those of the perprotio analogue (k^h/k^d = 1.04 ± 0.01 and 1.07 ± 0.09, respectively). Thus, any stabilization of the vacant coordination site of I by agostic interactions with the acetyl methyl group appears to be kinetically insignificant, consistent with the previous conclusion (Inorg. Chem. 2000, 39, 3098-3106) that this site is stabilized by an η²-coordinated carbonyl. Changing the phosphine ligand has a greater influence on the kinetics of I. The species generated by the flash photolysis of the trialkyl phosphine complex CH₃C(O)Co(CO)₃(P(ⁿBu₃)) exhibited a much larger k_M than was the case for the PPh₃ analogue, although there was little difference in the k_CO values. Similarly, k_M proved to be sensitive to the nature of R as demonstrated by the slower alkyl migration (at 298 K) for the intermediate formed by CO photodissociation from the propionyl complex C₂H₅C(O)Co(CO)₃PPh₃ relative to the acetyl analogue. Nonetheless, all these intermediates displayed analogous time-resolved infrared spectra and general kinetics behavior in benzene solution (implying common mechanisms for decay), so it is concluded that all are present as the η²-chelated acyl structure under these conditions.

Full text of DOI:10.1021/ic020567b

Inorg. Chem. 2003, 42, 575−580
Activation Parameters for the Reactive Intermediates Relevant to
Carbonylation Catalysts Based on Cobalt Carbonyls1
Steve M. Massick, Torsten Bu1ttner, and Peter C. Ford*
Department of Chemistry and Biochemistry, UniVersity of California, Santa Barbara,
Santa Barbara, California 93106
Received September 18, 2002
Time-resolved spectroscopic techniques have been used to prepare and to interrogate transient species that are
models for reactive intermediates in cobalt-catalyzed hydroformylation. Flash photolysis of acetylcobalt carbonyl
n
complexes of the type RC(O)Co(CO)3(PR′3) (A; R ) CH , CD , or C H ; R′ ) Ph or Bu) leads to CO
3
3
2 5
photodissociation to give the “unsaturated” intermediate [RC(O)Co(CO)2(PR′3)] (I), which decays by two competitive
pathways, alkyl migration to the cobalt to give RCo(CO)3PR′3 (M) and reaction with CO to re-form A. With the
perdeuterioacetyl complex (R ) CD , R′ ) Ph), rate constants both of CO trapping (kCO) and of methyl migration
3
(kM) were just slightly smaller than those of the perprotio analogue (kh/kd ) 1.04 ± 0.01 and 1.07 ± 0.09, respectively).
Thus, any stabilization of the “vacant” coordination site of I by agostic interactions with the acetyl methyl group
appears to be kinetically insignificant, consistent with the previous conclusion (Inorg. Chem. 2000, 39, 3098−3106)
that this site is stabilized by an η2-coordinated carbonyl. Changing the phosphine ligand has a greater influence
on the kinetics of I. The species generated by the flash photolysis of the trialkyl phosphine complex CH C(O)Co-
3
(CO)3(P(nBu3)) exhibited a much larger kM than was the case for the PPh3 analogue, although there was little
difference in the kCO values. Similarly, kM proved to be sensitive to the nature of R as demonstrated by the slower
alkyl migration (at 298 K) for the intermediate formed by CO photodissociation from the propionyl complex
C H C(O)Co(CO)3PPh3 relative to the acetyl analogue. Nonetheless, all these intermediates displayed analogous
2
5
time-resolved infrared spectra and general kinetics behavior in benzene solution (implying common mechanisms
for decay), so it is concluded that all are present as the η2-chelated acyl structure under these conditions.
Introduction
CO can be considered the most important C1 building block
of the chemical industry.
Carbon monoxide “migratory insertion” into metal-alkyl
bonds is the key C-C bond formation pathway in catalytic
carbonylations such as acetic acid synthesis from methanol,
alkene hydroformylation, etc.2 Since the discovery of cobalt
carbonyl hydroformylation catalysts in 1938, homogeneous
carbonylation has grown to major economic importance, and
Alkylmetal carbonyl complexes have been extensively
probed as models for this fundamental class of organome-
tallic reactions.3-5 Such studies suggest a mechanism in
which reversible alkyl migration to a cis-carbonyl leads to a
reactive intermediate Ith, which subsequently reacts with a
ligand L in solution (eq 1).
* Author to whom correspondence should be addressed. E-mail: ford@
chem.ucsb.edu.
(1) (a) Taken in part from the Diplomarbeit, by T.B., submitted in June
2000 to the Chemisch Geowissenschaftliche Fakulta¨t, Institut fu¨r
Anorganische und Analytische Chemie, Friedrich-Schiller-Universita¨t
Jena, Germany. (b) Reported in part at the 217th ACS National
Meeting, Anaheim, CA, March 1999; INORG 168.
(2) (a) Parshall, G. W.; Ittel, S. D. Homogeneous Catalysis; John Wiley
& Sons: New York, 1992. (b) Cornils, B.; Herrmann, W. A. Applied
homogeneous catalysis with organometallic compounds: a compre-
hensiVe handbook in two Volumes; VCH: Weinheim, New York, 1996.
(3) 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; Chapter 6.
Intermediates such as Ith are rarely observable directly
owing to their low steady-state concentrations. The strategy
used in this laboratory for characterizing the structure and
reactivity of such transient species starts with an acyl
complex, i.e., the product of the thermal reaction. Photodis-
10.1021/ic020567b CCC: $25.00 © 2003 American Chemical Society
Published on Web 12/24/2002
Inorganic Chemistry, Vol. 42, No. 2, 2003 575

Massick et al.
sociation of a ligand L from the acyl complex A (eq 2)
and substituted cobalt carbonyl catalysts.11 A key step is the
migratory insertion of a CO into a Co-R bond (eq 3).
R-Co(CO)3L + CO f RC(O)Co(CO)3L
(3)
Described in the present paper are time-resolved spectro-
scopic investigations of the modified cobalt carbonyl com-
plexes RC(O)Co(CO)3(PR′3), where R is CH3, CD3, or Et
and PR′3 is PPh3 or P(nBu)3. The goal of the present study
was to survey the possible effects of changing the alkyl group
or the phosphine modifier on the reaction dynamics of the
intermediate I. One concern was the possible role of agostic
bonding between alkyl group C-H(D) bonds and Co in
stabilizing I. In addition P(nBu)3 is a closer approximation
(than is PPh3) to the trialkylphosphines used in the modified
cobalt carbonylation catalysts.2 These studies were carried
out using a high-pressure variable-temperature (HP/VT) flow
cell reactor7c to examine the reactivity of I under a wide
range of CO partial pressures (PCO) and temperatures in a
manner that allows the accurate determination of activation
parameters for the competitive decay steps.
prepares a reactive intermediate I with the same composition
as that proposed for the thermal reaction intermediate. Time-
resolved spectroscopic studies are used to interrogate the
nature of I as well as the dynamics of the reactions to give
the stable acyl products and of reverse alkyl migration to
give the metal alkyl complex M. The resulting time-resolved
optical (TRO) and time-resolved infrared (TRIR) spectra
under various conditions are then interpreted in terms of
potential mechanisms. For example, in this manner we have
probed the model systems Mn(CO)5(C(O)R) and CpFe(CO)2-
(C(O)R) (Cp ) η5-C5H5).6,7 Described here is an extension
of our recent study of cobalt systems8 related to the
phosphine-modified cobalt carbonyl catalysts used in certain
industrial hydroformylation processes.2
Experimental Section
Phosphine-modified cobalt carbonyl catalysts used for
carbonylation of higher olefins have more favorable linear-
to-branched selectivity but require higher temperatures than
do catalysts based on the simple cobalt carbonyls.2 The
former also have greater tolerance of feedstock impurities
and better thermal stability than do rhodium catalysts now
predominant for propene hydroformylation. As a conse-
quence, there is continuing interest in new applications of
cobalt carbonyls as carbonylation catalysts.9 Pioneering
mechanistic studies by Heck and Breslow10 of cobalt
carbonyl catalyzed hydroformylation of alkenes led to the
proposed catalytic cycle that has been the starting point for
quantitative reaction mechanism studies with unsubstituted
Materials. All syntheses were carried out on a vacuum line using
Schlenk and cannula techniques or in an inert atmosphere box under
argon. The argon used in the vacuum line was dried and deoxy-
genated by passage over columns of activated molecular sieves and
an oxygen scavenger (Phillips catalyst on silica). Dicobalt octac-
arbonyl was purchased from Strem Chemicals, and the phosphines
PPh3 and P(nBu)3 and acid chlorides CD3C(O)Cl and C2H5C(O)Cl
were purchased from Aldrich. All solvents were from Fischer
Chemicals and dried by refluxing with sodium and distilling under
a N2 atmosphere.12 Research grade carbon monoxide gas (99.995%
purity from Spectra Gases) was used without further purification.
Syntheses. The phosphine-modified acylcobalt carbonyl com-
plexes RC(O)Co(CO)3(PR′3) (A) used in the flash photolysis studies
were prepared from Co2(CO)8 in overall yields near 50% according
to published methods.13 The products were characterized by NMR
and FTIR spectroscopy.
Solutions for TRIR Experiments. The calculated amounts of
cobalt complex were transferred into a Schlenk-type flask in the
drybox, where they were stored. Following evacuation for a few
minutes to remove residual solvents or gases, the desired solvent
was added using a cannula. Prior to that, the solvent was dried and
distilled under argon and afterward degassed either by the freeze-
pump-thaw (f-p-t) technique or simply by repeated evacuation.
The solution concentration was generally about 3 mM. With a path
length of 0.5 mm, this gave an optical absorbance at the 355 nm
(4) (a) Mawby, R. J.; Basolo, F.; Pearson, R. G. J. Am. Chem. Soc. 1964,
86, 3994-3999. (b) Calderazzo, F. Angew. Chem., Int. Ed. Engl. 1977,
16, 299-311. (c) Cawse, J. N.; Fiato, R. A.; Pruett, R. L. J. Organomet.
Chem. 1979, 172, 405-413. (d) Flood, T. C. Top. Stereochem. 1981,
12, 37-118. (e) Webb, S.; Giandomenico, C.; Halpern, J. J. Am. Chem.
Soc. 1986, 108, 345-347. (f) Cavell, K. J. Coord. Chem. ReV. 1996,
155, 209-243.
(5) Jordon, R. B. Reaction Mechanisms of Inorganic and Organometallic
Systems, 2nd ed.; Oxford University Press: New York, 1998.
(6) (a) Belt, S. T.; Ryba, D. W.; Ford, P. C. J. Am. Chem. Soc. 1991,
113, 9524-9528. (b) Boese, W. T.; Lee, B. L.; Ryba, D. W.; Belt, S.
T.; Ford, P. C. Organometallics 1993, 12, 4739-4741. (c) Boese, W.
T.; Ford, P. C. Organometallics 1994, 13, 3525-3531. (d) Boese, W.
T.; Ford, P. C. J. Am. Chem. Soc. 1995, 117, 8381-8391.
(7) (a) McFarlane, K. L.; Ford, P. C. Organometallics 1998, 17, 1166-
1168. (b) McFarlane, K. L.; Lee, B.; Fu, W. F.; van Eldik, R.; Ford,
P. C. Organometallics 1998, 17, 1826-1834. (c) Massick, S. M.; Ford,
P. C. Organometallics 1999, 18, 4362-4366.
(11) (a) Martin, J. T.; Baird, M. C. Organometallics 1983, 2, 1073-1078.
(b) Kovacs, I.; F. Ungvary; Marko, L. Organometallics 1986, 5, 209-
215. (c) Roe, D. C. Organometallics 1987, 6, 942-946. (d) Pino, P.;
Major, A.; Spindler, F.; Tannenbaum, R.; Bor, G.; Horvath, I. T. J.
Organomet. Chem. 1991, 417, 65-76. (e) Borovikov, M. S.; Kovacs,
I.; Ungvary, F.; Sisak, A.; Marko, L. Organometallics 1992, 11, 1576-
1579. (f) Bartik, T.; Krummling, T.; Happ, B.; Sieker, A.; Marko, L.;
Boese, R.; Ugo, R.; Zucchi, C.; Palyi, G. Catal. Lett. 1993, 19, 383-
389. (g) Klingler, R. J.; Rathke, J. W. J. Am. Chem. Soc. 1994, 116,
4772-4785. (h) Rathke, J. W.; Klinger, R. J., II; R. E. G.; Kramarz,
K. W.; Woelk, K. Prog. NMR Spectrosc. 1997, 30, 209-253.
(12) Riddick, J. A.; Bunger, W. B.; Sakano, T. K. Organic SolVents Physical
Properties and Methods of Purification, 4th ed.; John Wiley & Sons:
New York, 1986; Vol. II.
(8) Massick, S.; Rabor, J.; Elbers, S.; Marhenke, J.; Bernhard, S.;
Schoonover, J.; Ford, P. C. Inorg. Chem. 2000, 39, 3098-3106.
(9) (a) Rathke, J. W.; Klinger, R. J.; Krause, T. R. Organometallics 1991,
10, 1350-1355. (b) Knifton, J. F.; Lin, J. J. J. Mol. Catal. 1993, 81,
27-36. (c) Piotti, M. E.; Alper, H. J. Am. Chem. Soc. 1996, 118,
111-116. (d) Kramarz, K. W.; Klingler, R. J.; Fremgen, D. E.; Rathke,
J. W. Catal. Today 1999, 49, 339-352. (e) Miguel, S.; Zeigler, T.
Organometallics 1996, 15, 2611 and references therein. (f) Rossi. L.,
Piacenti, F. Bianchi, M.; Frediani, P.; Salvini, A. Eur. J. Inorg. Chem.
1999, 67-68. (g) Goh, S. K.; Marynick, D. S. Organometallics 2002,
22, 2262-2267. (h) Allmendinger, M., Eberhardt, R.; Luinstra, G.;
Reiger, B. J. Am. Chem. Soc. 2002, 124, 5646-5647.
(13) Lindner, E.; Zipper, M. Chem. Ber. 1974, 107, 1444-1455. (b) Roe,
(10) Heck, R. F.; Breslow, D. S. J. Am. Chem. Soc. 1962, 84, 2499-2502.
D. C. Organometallics 1987, 6, 942-946.
576 Inorganic Chemistry, Vol. 42, No. 2, 2003

Carbonylation Catalysts Based on Cobalt Carbonyls
laser excitation wavelength of ∼0.75 and initial IR absorbances of
0.3-0.4 for the most intense of the terminal carbonyl stretching
bands of the parent complex. The sample was transferred into the
Parr autoclave of the HP/VT flow system7c using a gastight syringe
under a CO atmosphere. The gas mixture at a specific pressure
was introduced, and the solution was stirred for at least 1 h to
guarantee complete equilibration. Concentrations of CO were
calculated from published solubility data.14 TRIR experiments were
carried out using flowing solutions to minimize complications
arising from product accumulation. The temperature throughout the
flow system and sample cell was regulated.
TRIR Instrumentation. Time-resolved infrared studies were
conducted on an experimental apparatus described previously.6d,8,15
The probe system is based upon diode IR lasers mounted in a
Spectra Physics/Laser Analytics model SP5731 laser head and a
CVI Digikrom model 240 monochromator and provides a tunable
IR probe laser source in the frequency range 1500-2200 cm-1
.
The probe beam was focused to a 7 mm diameter spot and
overlapped at the plane of the sample cell with the 355 nm pump
pulse from a Lumonics HY600 Nd:YAG laser. The transmitted
intensity of the IR probe beam was focused to fill the 1 mm2 active
area of a Fermionics model PV-8-1 photovoltaic Hg/Cd/Te detector.
A Fermionics model PVA 500-50 preamplifier enhanced the
detector signal prior to digitization by a LeCroy 9400 oscilloscope.
The average intensity of the pretrigger region (-2 to 0 µs) was
taken as I0 to calculate absorbance according to Abs(t) ) -log-
(I(t)/I0).
Figure 1. Temporal IR absorbance change at 1915 cm-1 (milli absorbance
units, mOD) after signal-averaged 40-shot acquisition for the 355 nm flash
photolysis of ACD in benzene solution under 100 psig of CO and at 45 °C.
3
For this particular experiment kobsd was calculated to be (9.73 ( 0.03) ×
105 s-1 from the exponential fit, which is also shown. The quality of the fit
is illustrated by the residual plotted at the top of the figure at a scale of
(0.5 mOD.
kobsd was determined to be (9.63 ( 0.11) × 105 s-1. The
relative standard deviations of the kobsd values were generally
lower than 2%.
Results and Discussion
Flash Photolysis Studies of CD3C(O)Co(CO)3(PPh3)
(ACD ). An earlier report8 from this laboratory using both
3
Treatment and Quality of TRIR Kinetics Data. Data
were accumulated for flowing solutions by recording the
temporal signal intensity of the IR photodiode immediately
before and after laser pulse excitation of the solution. Usually
data from 40 to 60 shots were acquired and encapsulated to
a single data set that was transformed to absorbance vs time
curves with good signal-to-noise ratios. Under CO, these
curves fit well to simple exponential decays from which first-
order rate constants ki were calculated. Generally, the mean
of 26 ki values so determined was used to provide the
“observed rate constant” kobsd values reported here for a
particular compound at a specific PCO and T (see Figure S-1
in the Supporting Information). An experiment was consid-
ered relevant only if the internal deviation among the 26 ki
values was small and random rather than systematic. This
level of data analysis ensured that no trend caused by
insufficient stirring, which could result in inhomogeneous
CO concentrations, was present. Figure 1 represents one of
the 26 acquisitions used to calculate the kobsd for the decay
of the transient seen at 1915 cm-1 upon 355 nm flash
variable single frequency and step scan FTIR techniques
demonstrated that 355 nm excitation of CH3C(O)Co(CO)3-
(PPh3) (ACH ) leads to CO photodissociation to give a reactive
3
intermediate ICH . Analysis of the TRIR spectra concluded
3
that there were but two detectable products from the decay
of ICH ,8 ACH re-formed by CO trapping of ICH and the alky
3
3
3
complex MCH from the competitive methyl migration from
3
the acyl group to the metal. The analogous scheme (eq 4)
was assumed for the perdeuterio analogue.
According to this model, the rate constant for the decay
of ICD would be expected to follow the behavior kobsd ) kM
3
+ kCO[CO]. Variation of [CO] was accomplished by applying
CO pressures up to 160 psig (11.96 bar absolute) in the HP/
VT flow system. PCO was determined by taking the vapor
pressure of benzene into account, and the mole fractions ø
and concentrations (mol/L) in solution were calculated. The
reaction dynamics were probed for 7-10 different [CO]
values at each temperature studied. Plots of kobsd vs [CO]
obtained for a specific temperature were linear with slopes
photolysis of CD3C(O)Co(CO)3(PPh3) (ACD ) in benzene
3
under 100 psig of CO (PCO ) 7.85 bar) and at 45 °C. For
this particular experiment the decay constant was determined
to be (9.73 ( 0.03) × 105 s-1, while the mean value for
(14) Concentrations of CO in various solvents were corrected for differences
in solubility: IUPAC Solubility Data Series: Carbon Monoxide;
Cargill, R. W., Ed.; Pergamon Press: New York, 1990; Vol. 43.
(15) (a) Ford, P. C.; Bridgewater, J. S.; Lee, B. Photochem. Photobiol.
1997, 65, 57-64. (b) DiBenedetto, J. A.; Ryba, D. W.; Ford, P. C.
Inorg. Chem. 1989, 28, 3503-3507. (c) Bridgewater, J. S.; Schoonover,
J. R.; Netzel, T. L.; Massick, S. M.; Ford, P. C. Inorg. Chem 2001,
40, 1466-1476.
k
CO and intercepts kM as shown in the inset of Figure 2 and
in Figure S-2 of the Supporting Information. Table 1 lists
kM and kCO values determined at six temperatures over the
range 20-45 °C and compares these to the rate constants
Inorganic Chemistry, Vol. 42, No. 2, 2003 577

Massick et al.
the methyl C-H(D) bonds and the metal center (B). If the
latter were the case, then an “inverse” isotope effect (i.e.,
kh/kd < 1.00) would be expected owing to the changes in
the C-H(D) vibrational zero-point energies in going from
the agostic coordinated B to transition states of the pathways
leading either to the η1-coordinated -CH3 of M (kM) or to
the η1-bonded acetyl of A (kCO).16 The small normal kh/kd
values are consistent with modest weakening of methyl group
normal modes as the CH3 moves through the reaction
transition states. Such a result might imply a more negative
charge on the CH3 group at the kM transition state in the
course of moving from C to M. Alternatively, this might
reflect agostic bonding during the course of the migratory
insertion mechanism; however, we are disinclined to place
much emphasis on such a small effect.
Flash Photolysis of CH3C(O)Co(CO)3P(nBu)3 (APBu
)
3
and C2H5C(O)Co(CO)3PPh3 (AEt). The intermediate IPBu
3
was generated by 355 nm flash photolysis of APBu in 308 K
3
Figure 2. Eyring plots for kM (lower) and for kCO (upper) determined for
benzene (PCO ≈ 4 atm), and the carbonyl region transient
the decay of the transient ICD generated by flash photolysis of ACD in
3
benzene. Inset: kobsd vs [CO] f3or individual temperatures ranging from 20
TRIR spectrum was recorded using a 400 ns window for
(lowest) to 45 °C.
data collection. Bleaching of bands characteristic of APBu
3
was noted, as well as the appearance of νCO bands for IPBu
3
at 1903((2) and 1640((2) cm-1, both of which disappeared
with lifetimes of ∼1 ms under these conditions. On this time
frame, a long-lived absorbance at ∼1935 cm-1 characteristic
measured for decay of the perprotio analogue ICH in benzene.
3
Eyring plots (Figure 2) for the kCO and kM values give the
activation parameters ∆Hq and ∆Sq for the two decay
pathways (Table 1).
of the methyl analogue MPBu was also apparent. The bands
3
The data in Table 1 indicate that there is little isotope effect
on either the kM or kCO pathways of I. A very small “normal”
for IPBu were analogous to those reported previously for the
3
8
PPh3 analogue ICH but were shifted to somewhat lower
3
kinetic isotope effect (i.e., kh/kd > 1.00) was observed at
frequencies (∼10 cm-1) owing to the greater electron-
h
298 K for both kM and kCO (kMh/kMd ) 1.07 ( 0.09 and kCO
/
donating ability of the trialkylphosphine.
kCOd ) 1.04 ( 0.01). This was also reflected in the activation
parameters ∆Hq and ∆Sq, which are (within experimental
uncertainty) the same for the two decay pathways for the
Kinetics data were acquired by following the exponential
disappearance of the 1903 cm-1 band of IPBu to obtain kobsd
3
values at seven CO concentrations over the range 0 to ∼90
mM for each of four temperatures (298, 308, 313, and 318
K). Values of kM and kCO were extracted for each temperature
from the intercepts and slopes, respectively, of linear kobsd
vs [CO] plots. Eyring plots of these data (Figure S-3 in the
Supporting Information) gave the ∆Hq and ∆Sq values. The
kM and kCO values at 298 K and the respective activation
intermediates ICH and ICD (Table 2).
3
3
Three possible structures for ICH are illustrated below as
3
C, S, and B. The absence of significant changes in kCO as
the solvent was varied argued against the ICH being the
3
parameters for the reactions leading to decay of IPBu are
3
summarized in Table 2. While kM is significantly larger than
seen for ICH , consistent with a smaller value of ∆HMq, the
3
rates and activation parameters for the kCO step are little
affected.
The carbonyl region TRIR spectrum of the intermediate
IEt generated by 355 nm flash photolysis of propionyl
complex AEt in benzene was recorded using a 300 ns window
for data collection. Bleaching of the νCO bands characteristic
of AEt and the appearance of new bands for IEt at 1915((2)
and 1632((2) cm-1 were noted. The terminal νCO band for
solvento complex S, especially in a relatively weakly
coordinating solvent such as benzene.8 Shifts in the acetyl
group νCO frequency indicated that this is still intact in ICH
,
3
probably as the η2-coordinated structure as illustrated by C.
A recent density functional calculation9f confirms that C is
likely to be the lowest energy structure of an intermediate
having the related composition CH3C(O)Co(CO)3, barring
strongly donating solvents. The very small normal kinetic
isotope effect for both pathways adds further credence to
the assignment of ICH and ICD as having the η2-acetyl-
IEt matches an analogous band in the TRIR spectrum of ICH
,
3
(16) (a) Calvert, R. B.; Shapley, J. R. J. Am. Chem. Soc. 1978, 100, 7726-
7727. (b) Buchanan, J. M.; Stryker, J. M.; Bergman, R. G. J. Am.
Chem. Soc. 1986, 108, 1537-1550. (c) Piers, W. E.; Bercaw, J. E., J.
Am. Chem. Soc. 1990, 112, 9406-7. (d) Paur-Afshari, R.; Lin, J.;
Schultz, R. H. Organometallics 2000, 19, 1682-1691. (e) Tanner,
Martha J.; Brookhart, M.; DeSimone, J. M. J. Am. Chem. Soc. 1997,
119, 7617-7618.
3
3
chelated structure C. An alternative would be a bidentate
configuration stabilized by an agostic interaction between
578 Inorganic Chemistry, Vol. 42, No. 2, 2003

Carbonylation Catalysts Based on Cobalt Carbonyls
Table 1. Values of kM and kCO (with Standard Deviations) for ICD and ICH (from Ref 8) Determined from the Linear kobsd versus [CO] Fits at Each
3
Individual T Used for the Determination of the Eyring Activation P3arameters ∆Hq and ∆Sq
d
d
h
h
T
(°C)
kM
(104 s-1
kCO
kM
(104 s-1
kCO
d
d
)
(107 M-1 s-1
)
)
(107 M-1 s-1
)
kMh/kM
kCOh/kCO
20
25
30
35
40
45
4.5 ( 0.5
5.8 ( 0.5
7.1 ( 0.5
9.8 ( 0.8
13.6 ( 0.6
16.9 ( 1.2
1.09 ( 0.01
1.10 ( 0.01
1.20 ( 0.01
1.24 ( 0.02
1.35 ( 0.01
1.45 ( 0.02
6.2 ( 0.7
7.6 ( 0.5
10.5 ( 0.9
13.5 ( 0.6
18.0 ( 0.8
1.14 ( 0.01
1.19 ( 0.01
1.26 ( 0.02
1.34 ( 0.01
1.40 ( 0.02
1.07 ( 0.09
1.07 ( 0.07
1.07 ( 0.09
0.99 ( 0.05
1.07 ( 0.06
1.04 ( 0.01
0.99 ( 0.01
1.02 ( 0.01
0.99 ( 0.01
0.97 ( 0.02
∆Hq (kJ mol-1
∆Sq (J mol-1 K-1
)
39 ( 3
-23 ( 9
6.6 ( 1.4
-88 ( 5
40 ( 2
-19 ( 6
5.7 ( 0.4
-91 ( 5
)
Table 2. Rate Constants (298 K) and Activation Parameters for the Reactions of Flash Photolysis Generated Intermediates in Benzene Solution
starting
complex
starting
complex
starting
complex
AEt
starting
complex
A
CH3
A
CD3
A
PBu3
k
CO(298 K) (M-1 s-1
∆HCOq (kJ mol-1
∆SCOq (J mol-1 K-1
km(298 K) (s-1
∆Hmq (kJ mol-1
∆Smq (J mol-1 K-1
)
(1.14 ( 0.01) × 107
(1.10 ( 0.01) × 107
(6.1 ( 0.1) × 106
(8.8 ( 0.2) × 106
)
5.7 ( 0.4
6.6 ( 1.4
6.8 ( 0.6
4.7 ( 2.8
)
-91 ( 12
-88 ( 5
-92 ( 2
-96 ( 9
)
(6.2 ( 0.7) × 104
(5.8 ( 0.5) × 104
(2.7 ( 0.5) × 104
(6.3 ( 0.1) × 105
)
40 ( 2
39 ( 3
52 ( 5
34 ( 3
)
-19 ( 6
-23 ( 9
13 ( 15
-21 ( 9
but the acyl νCO bands for AEt (1670 cm-1) and IEt occur at
lower frequencies, respectively, than for the acetyl analogues
ACH (1679 cm-1) and ICH (1639 cm-1), perhaps owing to
Scheme 1 Proposed Mechanisms of Methyl Migration and CO
Addition Steps for I
3
3
Et being more electron donating than Me. The rate constants
determined by monitoring the recovery of the AEt bleach at
1670 cm-1 and decay of the IEt absorption at 1632 cm-1
were in good agreement.
Kinetics data for the decay of IEt were acquired by
following the exponential disappearance of the 1915 cm-1
band to obtain kobsd values at seven CO concentrations over
the range 0 to ∼90 mM for each of four temperatures (298,
308, 313, and 318 K). Values of kM and kCO were extracted
for each temperature from the intercepts and slopes, respec-
tively, of linear kobsd vs [CO] plots. Eyring plots of these
data (Figure S-4 in the Supporting Information) gave the ∆Hq
and ∆Sq values. The kM and kCO values at 298 K and the
respective activation parameters for the reactions leading to
decay of IEt are summarized in Table 2, where it is seen
be relatively independent of the solvent medium, and this
was partially the basis of the conclusion that ICH is
3
principally present as the chelate C. The data in Table 2
show that, for each I studied, kCO (298 K) has a value near
107 M-1 s-1, with a very small ∆HCOq but a sizable, negative
∆SCOq. Such a pattern would be expected for a simple
associative mechanism such as illustrated in Scheme 118 for
the kCO pathway.
q
that, while kM demonstrates a larger value of ∆HM for IEt
than for ICH , the rates and activation parameters for the kCO
3
step are little affected.
Possible Mechanistic Implications. The TRIR data for
the complexes studied are highly consistent with each other.
The alkyl migration rates offer somewhat greater variety,
although they all fit the pattern of displaying a much larger
activation enthalpy but much less negative activation entropy
than does the respective CO trapping reaction. Given that in
C the R group is poorly oriented for migration to the metal,
the larger activation enthalpy may reflect the requirement
of solvent involvement in the reorganization of the structure
to one where migration can occur in a concerted step as
illustrated in Scheme 1. Again, our earlier data8 demonstrated
8
For example, the acyl νCO band in the IR spectra of ICH
,
IEt, and IPBu is considerably less intense than the acyl νC3 O
3
band of the respective starting complex A and is shifted about
40 cm-1 to lower frequency in each case. Furthermore, the
consistency of the kinetics behavior summarized in Table 2
for the four species studied argues that the structures of the
four respective intermediates must be analogous. As we have
argued above and elsewhere,8 these observations indicate that
I is present principally as the chelate C in each case.
The kinetics results reported in Table 2 summarize the
rate constants for the bimolecular reactions of the intermedi-
ates I with CO to regenerate the respective A complexes
(kCO) as well as the first-order migration of the acyl R group
to generate the respective alkyl complexes M (kM). Our
for ICH that, unlike the relatively solvent insensitive kCO
3
values, kM measured in benzene, dichloromethane, and THF
varied by a factor of 20, the fastest rate being seen in the
strongest donor THF. Although there is little difference
between the ∆HMq and ∆SMq values for ICH and ICD , ∆HM
q
3
3
is measurably larger for IEt and smaller for IPBu . This would
3
earlier studies8 of the reactions of ICH demonstrated kCO to
be consistent with the TRIR spectral data mentioned above.
3
Inorganic Chemistry, Vol. 42, No. 2, 2003 579

Massick et al.
One might expect the carbonyl oxygen of the propionyl group
In summary, further examination of the TRIR spectra of
the reactive intermediates generated by the flash photolysis
of the complexes RC(O)Co(CO)3(PR′3) (A) (R ) CH3, CD3,
to be more basic than that of the acetyl, and hence the η2-
carbonyl group more difficult to displace for IEt than for ICH
.
3
n
However, a peculiar aspect is that ∆SMq is surprisingly more
favorable than for the other analogues, thus partially com-
pensating for the less favorable ∆HMq. In the same vein, the
stronger donor character of PBu3 relative to PPh3 would make
the cobalt center less electron accepting, which would thus
bind the acetyl less strongly. This has little effect on the
associative kCO pathway, but does have the effect of lowering
the enthalpic barrier for CH3 migration.
or Et; R′ ) Ph or Bu) demonstrates that CO photodisso-
ciation leads to intermediates I, for which it is concluded
that the structure is the η2-carbonyl species C in each case.
These decay by two competitive pathways. The first is the
simple associative addition of CO to regenerate the respective
starting compound A. The kCO rates are essentially indepen-
dent of the nature of the solvent medium (within the limits
of the systems studied) and display a small activation
q
enthalpy ∆HCO and large and negative activation entropy
Notably, the sensitivity of the kM pathway to solvent is
consistent with the frequent observation that the microscopic
reverse, namely, the migratory insertion reaction of the type
M + L f A, which forms the acyl group from the metal
alkyl, is promoted by donor solvents.6d,17 Similarly, rates of
alkene hydroformylation catalyzed by phosphine-modified
cobalt carbonyls are increased by addition of polar solvents;2b
however, given the complexity of the catalytic cycles, it
would be difficult to attribute such an effect to any single
step.
∆SCOq consistent with the proposed associative mechanism.
In contrast the kM rates are accelerated by donor solvents
and display much larger activation enthalpies, and we
conclude that this behavior reflects the intimate involvement
of solvent in the transition state of the migratory pathway.
Acknowledgment. This research was sponsored by a
grant (DE-FG03-85ER13317) to P.C.F. from the Division
of Chemical Sciences, Office of Basic Energy Sciences, U.S.
Department of Energy. T.B. acknowledges Dr. Prof. D.
Walther, Friedrich-Schiller-Universita¨t Jena, who served as
Chair of the Diplomarbeit.
(17) (a) Mawby, R. J.; Basolo, F.; Pearson, R. G. J. Am. Chem. Soc. 1964,
86, 3994-3999. (b) Butler, I. S.; Basolo, F.; Pearson, R. G. Inorg.
Chem. 1967, 6, 2074. (c) Noack, K.; Calderazzo, F. J. Organomet.
Chem. 1967, 10, 101-104. (d) Wax, M. J.; Bergman, R. G. J. Am.
Chem. Soc. 1981, 103, 7028-7030. (e) Cotton, J. D.; Bent, T. L.
Organometallics 1991, 10, 3156-3160.
(18) The stereochemistry depicted in Scheme 1 is strictly conjectural for
both reactions. The pathways chosen would have the minimum number
of steps to give the most stable isomers in each case with the phosphine
and acyl (for A) or alkyl (for M) groups in the trans axial positions.
Supporting Information Available: A table of kinetics data
for the transient decay of the reactive intermediate generated by
the flash photolysis of CD3COCo(CO)3(PPh3) and four figures
describing the results of various kinetics experiments. This material
is available free of charge via the Internet at http://pubs.acs.org.
IC020567B
580 Inorganic Chemistry, Vol. 42, No. 2, 2003