Reductively induced homolytic carbon-carbon bond cleavage in Co(CO)3(PPh3)(COCF3)

Article abstract of DOI:10.1016/j.jorganchem.2006.12.027

The chemical or electrochemical reduction of the trifluoroacetyl complex Co(CO)₃(PPh₃)(COCF₃) involves a single electron transfer yielding trifluoromethyl radical and an anionic cobalt carbonyl complex. The mechanism is proposed to involve electron transfer followed by initial dissociation of either a carbonyl or phosphine ligand from the 19-electron [Co(CO)₃(PPh₃)(COCF₃)]^- anion. The resulting 17-electron intermediate undergoes subsequent one-electron reductive elimination of trifluoromethyl radical by homolytic cleavage of the carbon-carbon bond of the trifluoroacetyl group. The CF₃^{{radical dot}} radical can be trapped by either benzophenone anion, forming the anion of α-(trifluoromethyl)benzhydrol, or Bu₃SnH, yielding CF₃H. The ultimate organometallic product is an 18-electron anion, either [Co(CO)₄]^- or [Co(CO)₃(PPh₃)]^-, depending upon which ligand is initially lost. Fluorine-containing products were identified and quantitated by ¹⁹F NMR while cobalt-containing products were determined by IR.

Full text of DOI:10.1016/j.jorganchem.2006.12.027

Journal of Organometallic Chemistry 692 (2007) 3231–3235
www.elsevier.com/locate/jorganchem
Reductively induced homolytic carbon–carbon bond cleavage
in Co(CO)3(PPh3)(COCF3)
*
Nanda Gunawardhana, Stephen L. Gipson
Chemistry Department, Baylor University, One Bear Place #97348, Waco, TX 76798, United States
Received 20 November 2006; received in revised form 18 December 2006; accepted 20 December 2006
Available online 9 January 2007
Abstract
The chemical or electrochemical reduction of the trifluoroacetyl complex Co(CO)3(PPh3)(COCF3) involves a single electron transfer
yielding trifluoromethyl radical and an anionic cobalt carbonyl complex. The mechanism is proposed to involve electron transfer fol-
lowed by initial dissociation of either a carbonyl or phosphine ligand from the 19-electron [Co(CO)3(PPh3)(COCF3)]ꢀ anion. The result-
ing 17-electron intermediate undergoes subsequent one-electron reductive elimination of trifluoromethyl radical by homolytic cleavage of
the carbon–carbon bond of the trifluoroacetyl group. The CFÅ3 radical can be trapped by either benzophenone anion, forming the anion
of a-(trifluoromethyl)benzhydrol, or Bu3SnH, yielding CF3H. The ultimate organometallic product is an 18-electron anion, either
[Co(CO)4]ꢀ or [Co(CO)3(PPh3)]ꢀ, depending upon which ligand is initially lost. Fluorine-containing products were identified and quan-
titated by 19F NMR while cobalt-containing products were determined by IR.
ꢀ 2007 Elsevier B.V. All rights reserved.
Keywords: Cobalt; Reduction; Carbon–carbon bond cleavage; Trifluoroacetyl; Trifluoromethyl radical
1. Introduction
to participate in migratory insertion of CO because of their
relatively strong metal–carbon bonds [9], we felt that com-
pound 1 was a good candidate for the reverse. The fact that
the thermal decarbonylation of 1 is slow makes it possible
to investigate the initiation and catalysis of this reaction by
electron transfer. The reduction of the related manganese
compound, Mn(CO)5(COCF3), has been reported to occur
at significantly less negative potentials than that of the cor-
responding trifluoromethyl complex, Mn(CO)5CF3 [10]. If
the same held true for compound 1, then it should be capa-
ble of participating in an electron transfer chain catalyzed
decarbonylation reaction as follows:
We have had a long-standing interest in the reactivity of
odd-electron organotransition metal complexes produced
by redox reactions, and in particular those reactions which
are catalytic with respect to electrons (electron transfer
chain catalysis or ETC [1]). Recently, we have been study-
ing the reduction of the trifluoroacetylcobalt complex
Co(CO)3(PPh3)(COCF3) (1). This particular complex is
one of the most thermally stable cobalt acyl complexes
[2] and the trifluoromethyl group provides a convenient
NMR label for following reactivity. While several examples
of oxidatively and reductively induced migratory insertion
reactions have been reported [3–7], so far as we are aware
there are no examples of redox initiated decarbonylation
reactions other than that of one rhenium formyl complex
[8]. Though perfluoroalkyl ligands are considered unlikely
CoðCOÞ3ðPPh3ÞðCOCF3Þ þ eꢀ
! ½CoðCOÞ ðPPh3ÞðCOCF3Þꢁꢀ
ð1Þ
ð2Þ
3
½CoðCOÞ ðPPh3ÞðCOCF3Þꢁꢀ
3
! ½CoðCOÞ ðPPh3ÞðCF3Þꢁꢀ þ CO
3
½CoðCOÞ ðPPh3ÞðCF3Þꢁꢀ þ CoðCOÞ ðPPh3ÞðCOCF3Þ
3
3
! CoðCOÞ ðPPh3ÞðCF3Þ þ ½CoðCOÞ ðPPh3ÞðCOCF3Þꢁꢀ
3
3
*
Corresponding author. Tel.: +1 254 710 4555; fax: +1 254 710 4272.
ð3Þ
E-mail address: stephen_gipson@baylor.edu (S.L. Gipson).
0022-328X/$ - see front matter ꢀ 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2006.12.027

3232
N. Gunawardhana, S.L. Gipson / Journal of Organometallic Chemistry 692 (2007) 3231–3235
Initiation by reduction in Eq. (1) would be followed by a
was obtained from Alfa Aesar and was dried at 100 ꢁC
under vacuum before use. All other reagents were obtained
commercially and were used as received. All chemical reac-
tions were carried out in a nitrogen-atmosphere glove box.
catalytic cycle consisting of Eqs. (2) and (3). While this par-
ticular application of ETC might not be exceptionally use-
ful given the ease of thermal decarbonylation of compound
1 at elevated temperatures, the proof of concept would be
significant.
2.2. Instrumentation
Unfortunately, preliminary studies of the reduction of
compound 1 failed to show any evidence for the proposed
ETC decarbonylation reaction. Denisovich et al. have pro-
posed that reduction of the Mn trifluoroacetyl complex
yields, at least initially, the [Mn(CO)5]ꢀ anion and trifluo-
roacetyl radical [10]. Since we observed metal carbonyl
anions in the products from both the chemical and electro-
chemical reduction of compound 1, we spent some time
searching for products derived from the trifluoroacetyl rad-
ical, but also without success. On the other hand, Cordaro
and Bergman have reported that the thermal decomposi-
tion of electron-rich trifluoroacetyl iridium compounds
yields trifluoromethyl anion through heterolytic cleavage
of the carbon–carbon bond of the trifluoroacetyl group
[9c]. While we did observe CF3H and CF3D among the
products of the reduction of compound 1 in deuterated sol-
vents, other evidence, including the electron stoichiometry,
argued against the production of trifluoromethyl anion.
We have ultimately determined that the organic fragment
being produced by reduction of 1 is trifluoromethyl radical.
Below we will present evidence that reduction of com-
pound 1 proceeds through a 19eꢀ anionic intermediate
which, after ligand dissociation, undergoes homolytic car-
bon–carbon bond cleavage in the trifluoroacetyl group.
The cobalt-containing products have been identified and
quantitated by infrared spectroscopy and the organic prod-
ucts by 19F NMR.
NMR spectra were acquired on either a Bruker DPX
300 or a Varian VNMRS 500. Proton chemical shifts were
referenced to residual protons in the solvent, 19F chemical
shifts to external CFCl3, and 31P chemical shifts to external
85% H3PO4. IR spectra were obtained using a Mattson
Instruments Genesis II FTIR and a cell with CaF2 win-
dows separated by a 0.1 mm spacer. Cyclic voltammetry
was performed with a Bioanalytical Systems BAS 100B/
W electrochemical analyzer using a 0.5 mm Pt disc working
electrode, Pt wire auxiliary electrode, and Ag wire quasi-
reference electrode. Solutions contained approximately
0.1 M [Bu4N]PF6 supporting electrolyte. All potentials
are expressed relative to the formal potential of the ferroce-
nium–ferrocene couple (Fc+/Fc).
3. Results and discussion
3.1. Electrochemistry of 1
Fig. 1 shows the cyclic voltammogram of compound 1 in
CH3CN. It displays a single chemically irreversible reduc-
tion at ꢀ2.43 V vs. Fc+/Fc. On the reverse scan, two ano-
dic peaks are seen at ꢀ0.64 and ꢀ0.27 V, corresponding to
the oxidation of products formed by the chemical reaction
which follows the forward electron transfer. By compari-
son with authentic samples and with literature values
[13], these anodic peaks were assigned to [Co(CO)3(PPh3)]ꢀ
2. Experimental
2.1. Reagents
2.5
2.0
1.5
1.0
0.5
0.0
Co2(CO)8 was obtained from Strem Chemicals and was
used to synthesize Co(CO)3(PPh3)(COCF3) [11] and
Co2(CO)6(PPh3)2 [12] by published procedures. Decameth-
ylchromocene (DMC) was also obtained from Strem
Chemicals. Stock solutions of the [Co(CO)4]ꢀ and
[Co(CO)3(PPh3)]ꢀ anions in 1:1 THF/CH3CN were pre-
pared by reduction of the cobalt carbonyl dimers with
1% Na(Hg). Solutions of K2BP (potassium salt of benzo-
phenone dianion [15]) were prepared by reaction of benzo-
phenone in THF with an excess of K metal and were
standardized by reaction with aqueous ethanol followed
by titration with standardized HCl. Solutions of KBP
(potassium salt of benzophenone anion) were prepared
by 1:1 dilution of K2BP with a THF solution of excess ben-
zophenone and were also standardized with HCl. The THF
was dried with CaH2 and then distilled under nitrogen
from sodium benzophenone anion before use. CH3CN
was distilled from CaH2 under nitrogen. The supporting
electrolyte for electrochemical experiments, [Bu4N]PF6,
0.0
-0.5
-1.0
Potential (V vs. Fc+/Fc)
Fig. 1. Cyclic voltammogram of 3 · 10ꢀ3 M Co(CO)3(PPh3)(COCF3) in
CH3CN/[Bu4N]PF6 (0.1 M) at a Pt disc electrode at 0.2 V sꢀ1
-1.5
-2.0
-2.5
.

N. Gunawardhana, S.L. Gipson / Journal of Organometallic Chemistry 692 (2007) 3231–3235
3233
and [Co(CO)4]ꢀ, respectively. Comparison of the peak cur-
rent for the reduction of compound 1 with that of the one-
electron reduction of tris(dibenzoylmethanato)iron and the
two-electron reduction of [CpFe(CO)2]2 at the same con-
centration indicates that the reduction of compound 1 is
a one-electron process. Note that the relatively small sizes
of the cobalt anion peaks on the reverse scan of the cyclic
voltammogram are as would be expected for oxidations so
far separated from the corresponding reduction on the for-
ward scan and are therefore not to be taken as evidence for
low yields of the anions. This effect is similar to that
observed for chemically reversible but electrochemically
irreversible redox couples in cyclic voltammetry [14]. The
cyclic voltammetric behavior of 1 in THF is essentially
the same as that in CH3CN.
NMR spectra. An internal standard of a,a,a-trifluorotolu-
ene was used to quantitate the fraction of the original fluo-
rine atoms accounted for by each observed product. In all
experiments, greater than 90% of the starting material’s
fluorine atoms were accounted for in the product solution
by comparison with the internal standard. Table 1 lists
the results of these experiments, showing the relative per-
centage yield of each product as determined by integration
of the 19F NMR signals. Some decarbonylation to
Co(CO)3(PPh3)(CF3), either thermally or perhaps via
ETC, occurred but typically no more than 5% of the start-
ing material was converted to this product. Using either
K2BP or KBP as reductant, approximately 20–30% of the
fluorine-containing products consisted of a mixture of
CF3D (1:1:1 triplet at ꢀ79.0 ppm) and CF3H (doublet at
ꢀ78.3 ppm). Most of the fluorine, 70–80%, showed up ini-
tially in the 19F NMR spectrum as a very broad peak cen-
tered at ꢀ74.7 ppm. Addition of a small amount of water
converted this peak to a sharp singlet at the same chemical
shift. The 31P NMR spectrum showed a small quartet for
the decarbonylation product at 52.2 ppm and singlets for
triphenylphosphine (ꢀ5.4 ppm) and triphenylphosphine
oxide (26.0 ppm). The triphenylphosphine oxide peak was
minimized by careful drying of the solvent with molecular
sieves, demonstrating that it most likely originated from a
reaction involving hydroxide produced by reaction of
traces of water with the reducing agent. No attempt was
made to quantitate the yields of phosphorus-containing
products.
The production of CF3H and CF3D could be accounted
for by either deprotonation of solvent and/or trace water by
CFꢀ3 anion or by hydrogen abstraction by CFÅ3 radical.
Unfortunately, because of the strongly reducing conditions
of the reaction, we could not use CH3OD as solvent to
distinguish between these two possibilities as others have
done [9c,18]. A survey of the literature on the reactivity of
CFꢀ3 anion revealed that it can react with benzophenone
to produce the anion of a-(trifluoromethyl)benzhydrol
[19]. The reported 19F NMR chemical shift of ꢀ74.5 ppm
matches our value extremely well, and additionally we have
obtained a commercial sample of the benzhydrol which gives
the same NMR spectrum. Deprotonation of the commercial
3.2. Chemical reduction of 1
The chemical reduction of compound 1 was performed
in both CH3CN and THF using either solutions of KBP
0
0
or K2BP in THF (E0 ꢂ ꢀ2.5 V) or solid DMC (E0 ꢂ
ꢀ1.6 V) as the reducing agent. Note that DMC is likely
able to reduce compound 1 despite its relatively low redox
potential because the rapid chemical reactions of 1 shift the
redox equilibrium toward the products. The IR spectrum
of the product solution in CH3CN showed carbonyl
stretching bands at 1926, 1891, 1837, and 1695 cmꢀ1, while
in THF bands were observed at 1926, 1886, and 1846 cmꢀ1
.
The bands at 1926 and 1837 cmꢀ1 correspond to indepen-
dently synthesized K[Co(CO)3(PPh3)] and the one at
1891 cmꢀ1 to K[Co(CO)4], both in CH3CN. The relatively
weak band observed at 1695 cmꢀ1 in CH3CN could not be
identified, but is in the range reported for bridging car-
bonyl ligands on anionic cobalt carbonyl clusters [16]. Sim-
ilar results have been obtained in controlled potential
electrolyses of compound 1 [17].
In order to quantitatively account for the cobalt-contain-
ing products of the reduction of compound 1, series of stan-
dard solutions of K[Co(CO)3(PPh3)] and K[Co(CO)4] were
prepared in 1:1 THF/CH3CN. Plots of absorbance as a
function of concentration at 1837 and 1891 cmꢀ1, respec-
tively, were used to estimate the concentrations of the
cobalt carbonyl anions in the mixture produced by the
reduction of 10 mM 1 in CH3CN using K2BP in THF.
The mixed solvent system was used for the calibration
curves in order to most closely model the solution resulting
from the chemical reduction of compound 1. This analysis
showed that the product solution contained 3.2 mM
K[Co(CO)4] and 4.8 mM K[Co(CO)3(PPh3)], accounting
for 80% of the cobalt from the starting material. The pro-
duction of these two anions in similar amounts agrees well
with the cyclic voltammetry results. The missing cobalt may
be accounted for by either precipitation of organometallic
anions or the formation of higher nuclearity cobalt cluster
Table 1
Percentage yields of products from the reduction of Co(CO)3(PPh3)
(COCF3) in d8-THF
Reductant Radical
Trapa
K2BP KBP KBP
Bu3SnH
DMC DMC
Bu3SnH
CF3H
CF3D
[Ph2C(O)CF3]ꢀ
Co(CO)3(PPh3)(CF3)
Unassigned
18
0
80
2
25
2
66
3
3b
58
10
24
0
4
62d
78
0
0
0
22d
0
37
0
6c
0
a
10 equiv. vs. 1.
anions as evidenced by the unassigned band at 1695 cmꢀ1
.
b
d
d
d
19F = ꢀ83.1(s).
c
The fate of the trifluoroacetyl group was determined by
19F = ꢀ84.1(s), ꢀ76.9(s).
19F = ꢀ74.1(s), ꢀ73.4(s).
performing the reaction in d8-THF and monitoring 19F
d

3234
N. Gunawardhana, S.L. Gipson / Journal of Organometallic Chemistry 692 (2007) 3231–3235
sample of a-(trifluoromethyl)benzhydrol with butyl-
lithium also produces the same peak broadening that we
observe in our initial product solutions. Thus the initially
broad resonance for a-(trifluoromethyl)benzhydrol likely
stems from proton exchange of the alkoxide, an effect which
has been previously reported for the anion of a-(trifluorom-
ethyl)benzyl alcohol [20]. Upon further reflection, we real-
ized that there were two possible routes to the benzhydrol
product when using either K2BP or KBP as the reducing
agent. Reaction of CFꢀ3 anion with benzophenone, Eq. (4),
or reaction of CFÅ3 radical with benzophenone anion, Eq.
(5), would yield the same product.
CFÅ3. Therefore, based on the results of the chemical titra-
tions of compound 1 with both KBP and K2BP as well as
cyclic voltammetry, we are confident that the reduction of 1
is a one-electron process.
3.3. Proposed mechanism for the reduction of compound 1
Scheme 1 shows the proposed mechanism for the reduc-
tion of compound 1. The initial electron transfer produces
an electron-rich 19eꢀ anion, [Co(CO)3(PPh3)(COCF3)]ꢀ.
Like many 19eꢀ organometallic complexes, this species
likely participates in rapid equilibration with 17eꢀ interme-
diates via dissociation of one of the ligands [22]. Based
upon the relative yields of [Co(CO)4]ꢀ and [Co(CO)3-
(PPh3)]ꢀ, it appears that dissociation of a CO ligand is
slightly favored. Homolytic cleavage of the C–C bond of
the trifluoroacetyl group converts the 17eꢀ anion to an
18eꢀ anion and CFÅ3 radical in what is effectively a one-elec-
tron reductive elimination. An alternative mechanism
would involve cleavage of the C–C bond prior to ligand
dissociation, but this would generate a 20eꢀ organometallic
intermediate and so is considered much less likely. Interest-
ingly, comparison can be made between this reaction and
previously reported oxidatively-induced one-electron
reductive eliminations [23,24]. In order for a one-electron
ligand to be lost directly from the metal following oxida-
tion of an 18eꢀ compound, prior coordination of another
two-electron ligand is required. Reductive elimination thus
occurs from a 19eꢀ species. In the present case, the car-
bonyl group of the trifluoroacetyl ligand functions as the
new two-electron ligand after dissociation of the CFÅ3 rad-
ical and reductive elimination occurs from a 17eꢀ species.
Another alternative mechanism proposed by a reviewer
of this manuscript is decarbonylation of the 17eꢀ interme-
diate anion followed by C–Co bond cleavage to yield the
CFꢀ þ Ph2CO ! ½Ph2CðOÞCF3ꢁꢀ
ð4Þ
ð5Þ
3
CFÅ þ Ph2COÅꢀ ! ½Ph2CðOÞCF3ꢁꢀ
3
In fact, using K2BP as the reducing agent would make
the second reaction quite likely if CFÅ3 were formed quickly
following the electron transfer. Using KBP as reductant
yielded the same ambiguous result, though with a some-
what lower yield of the benzhydrol anion.
We then decided to try to test for the intermediacy of
CFÅ3 by adding a hydrogen atom donor. The usual radical
traps with benzylic hydrogens did not significantly affect
the yield of CF3H when 1 was reduced in d8-THF. How-
ever, the literature indicates that while CFÅ3 is highly reac-
tive as a hydrogen atom abstractor, it strongly prefers
hydrogens bound to less electronegative elements [21].
We therefore added 10 equiv. of Bu3SnH as a radical trap
and saw a large increase in the yield of CF3H when using
KBP as the reducing agent. In order to further differentiate
between CFÅ and CFꢀ and increase the yield of CF3H, we
3
3
performed the reduction of compound 1 using DMC as the
reducing agent. In this case also, a very high yield of CF3H
was observed with Bu3SnH compared with a much lower
yield of a mixture of CF3D and CF3H in the absence of
Bu3SnH. This result conclusively identified CFÅ radical as
3
the organic intermediate produced by the chemical reac-
tions following reduction of compound 1.
CF3COCo(CO)3PPh3
+e –
One further piece of information determined via chemi-
cal reduction of compound 1 was the electron stoichiome-
try. Previous results from controlled potential electrolyses
had agreed with cyclic voltammetric peak currents in indi-
cating that the reduction of compound 1 is a net one-elec-
tron process [17]. This result was confirmed by titration of
compound 1 using KBP and following the progress of the
reaction by both IR (in THF) and NMR (in d8-THF) spec-
troscopies. Though exact results were difficult to obtain, it
consistently took slightly more than 1 equiv. (average of
1.3) of KBP to completely consume all of the starting mate-
rial. An excess of KBP was most likely required because of
[CF3COCo(CO)3PPh3]–
–L(CO,PPh3)
[CF3COCo(CO)2L]–
[Co(CO)3L]–
CF3•
reaction of CFÅ with KBP. Using K2BP the result was very
3
Bu3SnH
KBP,H2O
HO
close to one mole of K2BP per mole of 1. In this case it
should be recognized that the reduction of compound 1
CF3
C
by K2BP followed by reaction of CFÅ radical with benzo-
3
CF3H
phenone anion, Eq. (5), amounts to a two-electron process,
though only one of the electrons effects reduction of the
cobalt complex. The other electron effectively reduces the
Ph
Ph
Scheme 1.

N. Gunawardhana, S.L. Gipson / Journal of Organometallic Chemistry 692 (2007) 3231–3235
3235
trifluoromethyl radical and anionic cobalt carbonyl com-
Acknowledgements
plex. We have eliminated this possibility by examining
the reduction of the trifluoromethyl complex, Co(CO)3-
(PPh3)(CF3), synthesized by thermal decarbonylation of
compound 1 [25]. Reduction of the trifluoromethyl com-
plex with KBP yielded a mixture of anionic cobalt carbonyl
complexes similar to that obtained with compound 1, but
only about a 5% yield of a-(trifluoromethyl)benzhydrol.
Reduction of the trifluoromethyl complex in the presence
of Bu3SnH yielded less than 10% CF3H. We therefore con-
clude that the reduction of compound 1 is unlikely to pro-
ceed through the anion of the trifluoromethyl complex as
an intermediate. This conclusion is also consistent with
the lack of ETC substitution which would otherwise be
expected to occur if this anion were formed.
We thank the Robert A. Welch Foundation (Grant No.
AA-1083) for support of this research. We also thank the
National Science Foundation (Award #CHE-0420802)
for funding the purchase of our 500 MHz NMR.
References
[1] (a) T.V. Magdesieva, K.P. Butin, Russ. Chem. Rev. 71 (2002) 223;
(b) M.-H. Delville, Inorg. Chim. Acta 291 (1999) 1;
(c) F. Battaglini, E.J. Calvo, F. Doctorovich, J. Organomet. Chem.
547 (1997) 1;
(d) D. Astruc, Angew. Chem., Int. Ed. Engl. 27 (1988) 643.
[2] R.F. Heck, J. Am. Chem. Soc. 85 (1963) 651.
[3] R.H. Magnuson, R. Meirowitz, S.J. Zulu, W.P. Giering, Organo-
metallics 2 (1983) 460.
In CH3CN only, an unassigned CO stretching band is
observed in the product solution at 1695 cmꢀ1. It was orig-
inally suspected that this band originated from an organic
product containing the trifluoroacetyl fragment. However,
19F NMR spectroscopy accounts for essentially all of the
fluorine in the starting material, so it now seems unlikely
this IR band represents such an organic carbonyl com-
pound. This frequency is consistent with that of bridging
carbonyl ligands in some known anionic cobalt carbonyl
clusters [15], and so we are tentatively assigning it to such
a species. However, attempts to isolate this species for
definitive characterization have not been successful.
[4] A. Morneau, B.T. Donovan-Merkert, W.E. Geiger, Inorg. Chim.
Acta 300–302 (2000) 96.
[5] C. Amatore, M. Bayachou, J.-N. Verpeaux, L. Pospisil, J. Fiedler, J.
Electroanal. Chem. 387 (1995) 101.
[6] R.S. Bly, G.S. Silverman, M.M. Hossaln, R.K. Sly, Organometallics
3 (1984) 642.
´
[7] D. Miholova, A.A. Vlek, J. Organomet. Chem. 240 (1982) 413.
[8] B.A. Narayanan, C. Amatore, C.P. Casey, J.K. Kochi, J. Am. Chem.
Soc. 105 (1983) 6351.
[9] (a) H. Berke, R. Hoffmann, J. Am. Chem. Soc. 100 (1978) 7224;
(b) C.J. Bourgeois, R.P. Hughes, T.L. Husebo, J.M. Smith, I.M.
Guzei, L.M. Liable-Sands, L.N. Zakharov, A.L. Rheingold, Organo-
metallics 24 (2005) 6431;
(c) J.G. Cordaro, R.G. Bergman, J. Am. Chem. Soc. 126 (2004) 16912.
[10] L.I. Denisovich, M.G. Peterleitner, A.A. Stepanor, S.M. Peregudova,
Russ. Chem. Bull. 49 (2000) 375.
[11] R.F. Heck, D.S. Breslow, J. Am. Chem. Soc. 84 (1962) 2499.
[12] A.R. Manning, J. Chem. Soc. (A) (1968) 1135.
[13] (a) Y. Zhen, W.G. Feighery, C.-K. Lai, J.D. Atwood, J. Am. Chem.
Soc. 111 (1989) 7832;
(b) C.-K. Lai, W.G. Feighery, Y. Zhen, J.D. Atwood, Inorg. Chem.
28 (1989) 3929;
(c) K.Y. Lee, J.K. Kochi, Inorg. Chem. 28 (1989) 567.
[14] W.R. Heineman, P.T. Kissinger, in: P.T. Kissinger, W.R. Heineman
(Eds.), Laboratory Techniques in Electroanalytical Chemistry, Mar-
cel Dekker, New York, 1984, p. 92.
[15] T. Yoshimura, Z. Hou, Y. Wakatsuki, Organometallics 14 (1995) 5382.
[16] (a) P. Chini, V. Albano, J. Organomet. Chem. 15 (1968) 433;
(b) P. Chini, V. Albano, S. Martinengo, J. Organomet. Chem. 16
(1969) 471.
[17] (a) C.C. Wilkinson, B.S. Thesis, Baylor University, December 1999;
(b) Z. Lin, M.S. Thesis, Baylor University, December 2003.
[18] P.J. Toscano, H. Brand, S. Liu, J. Zubieta, Inorg. Chem. 29 (1990)
2101.
4. Conclusions
The reduction of Co(CO)3(PPh3)(COCF3) does not lead
to either dissociation or decarbonylation of the trifluoro-
acetyl ligand. Instead, dissociation of one of the neutral
ligands, CO or PPh3, occurs in order to decrease the
electron count at cobalt in accordance with the 18-electron
rule and the known reactivity of other 19eꢀ species. It then
appears that the resulting 17eꢀ anion, [Co(CO)2(L)-
(COCF3)]ꢀ (L = CO or PPh3), undergoes homolytic car-
bon–carbon bond cleavage in the trifluoroacetyl group to
yield trifluoromethyl radical and a stable 18eꢀ anionic
cobalt complex. This reaction amounts to a one-electron
reductive elimination in which the oxidation state of the
cobalt decreases from zero to negative one. Since both
reductive elimination of trifluoromethyl radical and
decarbonylation involve cleavage of the C–C bond of the
trifluoroacetyl group, it appears that the lack of electron
transfer catalyzed decarbonylation likely stems from a lack
of favorable interaction between the CFÅ3 radical and the
18eꢀ cobalt center. The instability of such a F3C–Co bond
is also evidenced by the rapid decomposition of the trifluo-
romethyl complex upon reduction.
[19] G.K.S. Prakash, J. Hu, G.A. Olah, Org. Lett. 5 (2003) 3253.
[20] J. Russell, N. Roques, Tetrahedron 54 (1998) 13771.
[21] W.R. Dolbier Jr., Chem. Rev. 96 (1996) 1557.
[22] (a) S. Sun, D.A. Sweigart, Adv. Organomet. Chem. 40 (1996) 171;
(b) D.R. Tyler, Acc. Chem. Res. 24 (1991) 325;
(c) D. Astruc, Chem. Rev. 88 (1988) 1189.
[23] T. Aase, M. Tilset, V.D. Parker, J. Am. Chem. Soc. 112 (1990) 4974.
[24] J. Kim, S.L. Gipson, Polyhedron 23 (2004) 1371.
[25] W. Hieber, W. Beck, E. Lindner, Z. Naturforsch. Teil. B 16 (1961) 229.