Anodic reaction of CO(η5-C5H5)(CO) (PPh3): An oxidatively induced ligand exchange involving a 17 e -/18 e- redox pair

Article abstract of DOI:10.1021/om100631k

Unlike a number of other monocarbonyl-substituted complexes CoCp(CO)L (Cp = (η⁵-C₅H₅)), chemical oxidation of the L = triphenylphosphine derivative CoCp(CO)(PPh₃), 1, is known to give, in the absence of added PPh₃, the disubstituted phosphine complex [CoCp(PPh₃)₂]⁺ rather than the simple 17-electron radical cation 1⁺. Electrochemistry of 1 in CH ₂Cl₂/[NBu₄][B(C₆F₅) ₄] (E_1/2(1/1⁺) = -0.38 V vs ferrocene) shows that the favored anodic products depend on the concentration of 1. At low concentrations (e.g., 10^-4 M) and on the cyclic voltammetry time scale, the radical cation 1⁺ dominates, but at higher concentrations and longer reaction times, a quantitative mixture of CoCp(CO)₂ and [CoCp(PPh₃)₂]⁺ is produced in an overall half-electron process. On the basis of cyclic voltammetry, chronoamperometry, bulk electrolysis, and IR spectroelectrochemistry, a radical-substrate mechanism is proposed involving the reaction of 1⁺ with 1 to give transient [Co₂Cp₂(CO) ₂(PPh₃)₂]⁺. This putative dimer radical cation intermediate may be viewed as an odd-electron analogue of the intermediates that have previously been invoked by Kochi, Atwood, and others to explain ligand-transfer reactions between cation/anion pairs of organometallic complexes.

Full text of DOI:10.1021/om100631k

4276 Organometallics 2010, 29, 4276–4281
DOI: 10.1021/om100631k
Anodic Reaction of Co(η5-C5H5)(CO)(PPh3): An Oxidatively Induced
Ligand Exchange Involving a 17 e-/18 e- Redox Pair
Ayman Nafady† and William E. Geiger*
Department of Chemistry, University of Vermont, Burlington, Vermont 05405. †Present address: School of
Chemistry, Monash University, Clayton, Victoria 3800, Australia.
Received June 30, 2010
Unlike a number of other monocarbonyl-substituted complexes CoCp(CO)L (Cp=(η5-C5H5)),
chemical oxidation of the L=triphenylphosphine derivative CoCp(CO)(PPh3), 1, is known to give, in
the absence of added PPh3, the disubstituted phosphine complex [CoCp(PPh3)2]þ rather than the
simple 17-electron radical cation 1þ. Electrochemistry of 1 in CH2Cl2/[NBu4][B(C6F5)4] (E1/2(1/1þ)=
-0.38 V vs ferrocene) shows that the favored anodic products depend on the concentration of 1.
At low concentrations (e.g., 10-4 M) and on the cyclic voltammetry time scale, the radical cation
1þ dominates, but at higher concentrations and longer reaction times, a quantitative mixture of
CoCp(CO)2 and [CoCp(PPh3)2]þ is produced in an overall “half-electron” process. On the basis of
cyclic voltammetry, chronoamperometry, bulk electrolysis, and IR spectroelectrochemistry, a “radical-
substrate” mechanism is proposed involving the reaction of 1þ with 1 to give transient [Co2Cp2(CO)2-
(PPh3)2]þ. This putative dimer radical cation intermediate may be viewed as an odd-electron analogue of
the intermediates that have previously been invoked by Kochi, Atwood, and others to explain ligand-
transfer reactions between cation/anion pairs of organometallic complexes.
Introduction
are employed as oxidants.6,7 Here we probe the curious
oxidative behavior of CoCp(CO)(PPh3), 1, drawing on the
recent discovery11 of the “radical-substrate” reactions of
17 e-/18 e- redox pairs to explain its oxidative fate.
The properties and reactions of 17-electron half-sandwich
transition metal complexes have been broadly studied.1-5
The chemistry of the cobalt-group radicals [MCp(CO)L]þ
(M = Co, Rh, Ir; Cp = (η5-C5H5), L=CO or other two-
electron ligand) is particularly rich. Depending on the metal,
the ligand L, and possible substitution in the Cp ring, the
radical cation varies in longevity (some have been isolated)6,7
and may undergo dimerization at either the metal8 or the
five-membered ring.9 The radicals are also implicated as
intermediates in the oxidative addition reactions of their
18-electron precursors when halogens are used as oxidants,10
and M-Ag complexes have been isolated when silver salts
As shown independently by two groups,6,7 chemical oxi-
dation of 1 gives primarily the disubstituted radical cation
[CoCp(PPh3)2]þ, with no direct (e.g., spectral) evidence for
1þ. Employing ferrocenium as the oxidant, and perform-
ing the reaction in the presence of an equivalent of PPh3,
Broadley et al.6 isolated [CoCp(PPh3)2][PF6] in good yield.
McKinney’s use of Ag[BF4] as oxidant without the addition
of PPh3 again gave [CoCp(PPh3)2]þ, but only in about 35%
yield.7 The side-products in these reactions were not identified.
In contrast, the one-electron oxidation of other members of this
mixed-ligand family proceeded without complication, allowing
the isolation of [CoCp(CO)(PCy3)]þ (Cy=cyclohexyl; [FeCp2]þ
as oxidant)6 and [CoCp(CO)(P(OiPr)3)]þ (Agþ as oxidant).7
Oxidation of the pentamethylcyclopentadienyl complex CoCp*-
(CO)(PPh3) (Cp*=η5-C5Me5) also gave its simple 17-electron
counterpart.6
*To whom correspondence should be addressed. E-mail: william.geiger@
uvm.edu.
(1) Connelly, N. G.; Geiger, W. E. In Advances in Organometallic
Chemistry; Stone, F. G. A.; West, R., Eds.; Academic Press: Orlando, 1984;
pp 1-93.
(2) Astruc, D. Electron Transfer and Radical Processes in Transition-
Metal Chemistry; VCH Publishers: New York, 1995.
(3) Baird, M. C. In Organometallic Radical Processes; Trogler, W. C.,
Ed.; Elsevier: Amsterdam, 1990; pp 49-66.
(4) Trogler, W. C. In Organometallic Radical Processes; Trogler,
W. C., Ed.; Elsevier: Amsterdam, 1990; pp 306-337.
(5) Geiger, W. E. Organometallics 2007, 26, 5738.
(6) Broadley, K.; Connelly, N. G.; Geiger, W. E. J. Chem. Soc.,
Dalton Trans. 1983, 121.
(7) McKinney, R. J. Inorg. Chem. 1982, 21, 2051.
(8) Fonseca, E.; Geiger, W. E.; Bitterwolf, T. E.; Rheingold, A. L.
Organometallics 1988, 7, 567.
(9) (a) McKinney, R. J. Chem. Commun. 1980, 603. (b) Freeman, M. J.;
Orpen, A. G.; Connelly, N. G.; Manners, I.; Raven, S. J. J. Chem. Soc.,
Dalton Trans. 1985, 2283.
The anodic oxidation of 1 gave clear evidence for 1þ in
cyclic voltammetry (CV) scans, but electrochemical studies
were hindered owing to the poor solubility of [BF4]- and
[PF6]- salts of 1þ in the low-donor solvents necessary to
avoid solvent attack on the radical cation.6 Thus, CV scans
of 1 in CH2Cl2/0.1 M [NBu4][PF6] revealed severe specific
adsorption of [CoCp(CO)(PPh3)][PF6] at a Pt electrode,
which obstructed the voltammetric tools12 ideally available
for diagnosis of the anodic reaction. Even the E1/2 of the
(10) (a) King, R. B. Inorg. Chem. 1966, 5, 82. (b) Oliver, A. J.; Graham,
W. A. G. Inorg. Chem. 1970, 9, 243.
(11) Nafady, A.; Costa, P. J.; Calhorda, M. J.; Geiger, W. E. J. Am.
Chem. Soc. 2006, 128, 16587.
r
pubs.acs.org/Organometallics
Published on Web 09/10/2010
2010 American Chemical Society

Article
Organometallics, Vol. 29, No. 19, 2010 4277
system (ca. -0.33 V vs ferrocene in CH2Cl2/0.1 M [NBu4]-
[PF6]) could only be estimated.6,13 We decided to reinvesti-
gate the anodic oxidation of 1 in CH2Cl2, using tetrakis-
(pentafluorophenyl)borate, [B(C6F5)4]- (TFAB), as the
electrolyte anion to mitigate the electrode adsorption problem,14
with the goal of better understanding why the redox behav-
ior of PPh3-substituted 1 differs from that of its PCy3 and
P(OiPr)3 counterparts. The heterogeneous versus homogeneous
nature of electrochemical versus chemical redox processes
had to be kept in mind in this study. Since a combination of
these methods is often used for organic,15 inorganic,16 and
organometallic17 redox reactions, a better overall under-
standing of the oxidation of 1 may aid how these comple-
mentary methods are employed in synthetic-level redox
chemistry.
The present work suggests that the ligand-exchange pro-
cess of the radical cation 1þ occurs through its interaction
with the 18-electron precursor, 1, in a “radical-substrate”
reaction. Variations in the production of the disubstituted
product, [CoCp(PPh3)2]þ (2þ), can be understood in terms of
differences in the relative “localized” homogeneous concen-
trations of 1 and 1þ generated by either electrochemical or
chemical oxidation.
Figure 1. Cyclic voltammogams of 0.1 mM 1 in CH2Cl2/0.1 M
[NBu4][B(C6F5)4] at scan rate of 0.1 V/s, 298 K using a 1.5 mm
diameter GC disk electrode. Extended scan (solid line) shows
virtual absence of follow-up product peaks at 0.37 V (CoCp-
(CO)2) and -1.23 V ([CoCp(PPh3)2]þ).
cm2 s-1). The disks were polished using diamond paste (Buehler)
of decreasing sizes (3 to 0.25 μm), interspersed by washings
with Nanopure water, followed by final vacuum drying. The
working electrode for bulk electrolyses was a basket-shaped
platinum gauze.
Experimental Section
Diagnostic criteria involving shapes and scan rate responses
of the cyclic voltammetry (CV) waves were applied using pro-
cedures described elsewhere.12d Linear scan voltammograms
(LSVs) were obtained under quasi-steady-state conditions using
unstirred solutions and a scan rate of 2 mV s-1. Digital simula-
tions were performed using Digisim 3.0 (Bioanalytical Systems).
All potentials given in this paper are referred to the ferrocene/
ferrocenium reference couple,17 using the in situ method.19
[NBu4][B(C6F5)4] was prepared by metathesis of [NBu4]Br
with M[B(C6F5)4] (M = Li or K, Boulder Scientific) in aqueous
methanol and recrystallized from dichloromethane/ether.20
Spectroscopy. IR spectra were recorded with an ATI-Mattson
Infinity Series FTIR spectrometer operating at a resolution
of 4 cm-1. NMR spectra were recorded using a Bruker ARX
500 MHz spectrometer, and ESR spectra were obtained on a
Bruker ESP 300E spectrometer. IR spectroelectrochemistry was
performed under argon using a mid-IR fiber-optic “dip” probe
(Remspec, Inc.).21 Descriptions of the spectroelectrochemical
procedures used in recording spectra and voltammograms are
available.11,21
Experiments were performed under nitrogen using either
standard Schlenk conditions or a Vacuum Atmospheres drybox.
Reagent-grade solvents were distilled from appropriate drying
agents. Dichloromethane was further distilled from CaH2 under
static vacuum to a flask that was later opened in the drybox.
Glassware used for electrochemical experiments was cleaned by
placing it in a No-Chromix (Godax Laboratories, Inc.) solution
for at least 12 h, followed by copious rinsings with Nanopure
water and subsequent drying for at least 12 h in a 120 °C oven.
The warm glassware was then loaded into the drybox ante-
chamber. CoCp(CO)2 was purchased from Strem Chemicals.
CoCp(CO)(PPh3)10a and [FeCp2][B(C6F5)4]18 were prepared by
the literature methods.
Electrochemistry. A PARC 273A potentiostat linked to
a personal computer was employed, using standard three-
electrode cells for voltammetry and electrolysis experiments,
most of which were carried out inside the drybox. The drybox
was outfitted with a cooling bath capable of controlling solution
temperatures to within 1 °C. Voltammetry scans were recorded
using glassy carbon working electrode disks of 1 to 2 mm
diameter (Bioanalytical Systems), their effective areas being
determined through chronoamperometry experiments using
ferrocene in acetonitrile/0.1 M [NBu4][PF6] (Do = 2.25 ꢀ 10-5
Results
Electrochemical Oxidation of CoCp(CO)(PPh3). A concise
overview of the anodic properties of 1 in CH2Cl2/0.1 M
[NBu4][TFAB] is that, depending on concentrations and reac-
tion times, the oxidation may either follow a simple, Nernstian,
one-electron process (eq 1) or be subject to second-order
reactions
(12) See, for example: (a) Bard, A. J.; Faulkner, L. R. Electrochemical
Methods, 2nd ed.; John Wiley & Sons: New York, 2001; pp 471-531.
(b) Rieger, P. H. Electrochemistry, 2nd ed.; Chapman & Hall: New York,
1994; pp 296-309. (c) Hammerich, O. In Organic Electrochemistry, 4th
ed.; Lund, H.; Hammerich, O., Eds.; Marcel Dekker, Inc.: New York, 2001;
pp 95-182. (d) Geiger, W. E. In Laboratory Techniques in Electroana-
lytical Chemistry; Kissinger, P. T.; Heineman, W. R., Eds.; Marcel Dekker,
Inc.: New York, 1996; pp 683-717.
CoCpðCOÞðPPh3Þ - e- h ½CoCpðCOÞðPPh3Þꢁþ ð1Þ
1þ
1
(13) The estimated E1/2 of 1, 0.13 V vs SCE in ref 6, is converted to the
ferrocene/ferrocenium potential by subtraction of 0.46 V in CH2Cl2/0.1
M [NBu4][PF6].
which give rise to ligand exchange. The former is favored
at lower concentrations, lower temperatures (e253 K), and
shorter electrochemical reaction times (e.g., faster CV scans).
ꢀ
(14) Geiger, W. E.; Barriere, F. Acc. Chem. Res. 2010, 43, 1030.
th
€
(15) Schafer, H. J. In Organic Electrochemistry, 4 ed.; Lund, H.;
Hammerich, O., Eds.; Marcel Dekker, Inc.: New York, 2001; pp 207-222.
(16) Davidson, A.; Holm, R. H. In Inorganic Syntheses; Muetterties,
E. L., Ed.; McGraw-Hill: New York, 1967; Vol. 10, pp 8-26.
(17) Connelly, N. G.; Geiger, W. E. Chem. Rev. 1996, 96, 877.
(18) (a) Antonelli, D. M.; Leins, A.; Stryker, J. M. Organometallics
1997, 16, 2500. (b) O'Connor, A. R.; Nataro, C.; Golen, J. A.; Rheingold,
A. L. J. Organomet. Chem. 2004, 689, 2411.
(19) Gagne, R. E.; Koval, C. A.; Lisensky, G. C. Inorg. Chem. 1980,
19, 2854.
(20) LeSuer, R. J.; Buttolph, C.; Geiger, W. E. Anal. Chem. 2004, 76,
6395.
(21) Shaw, M. J.; Geiger, W. E. Organometallics 1996, 15, 13.

4278 Organometallics, Vol. 29, No. 19, 2010
Table 1. Redox Potentials (vs FeCp20/þ) Relevant to This Study
Nafady and Geiger
redox couple
E1/2 (V)
medium
reference
CoCp(CO)(PPh3), 1/1þ
1/1þ
-0.38
-0.33b
-1.23
0.37
CH2Cl2/0.1 M [NBu4][TFAB]a
CH2Cl2/0.1 M [NBu4][PF6]
CH2Cl2/0.1 M [NBu4][TFAB]
CH2Cl2/0.1 M [NBu4][TFAB]
CH2Cl2/0.1 M [NBu4][TFAB]
this work
6b
this work
11
11
CoCp(PPh3)2, 2/2þ
[CoCp(CO)2]0/þ, 3/3þ
[CoCp(CO)2]2
þ/2þ
0.47
a TFAB = [B(C6F5)4]. b Estimated potential of 0.13 V vs SCE in ref 6 has been converted to ferrocene reference potential.
Figure 3. Linear scan voltammograms obtained with a scan
rate of 2 mV/s under quasi-steady-state conditions before (solid
line) and after (dotted line) bulk electrolysis of 1.07 mM 1 in
CH2Cl2/0.05 M [NBu4][B(C6F5)4] at 273 K, using a 1 mm dia-
meter GC electrode. The currents B and C are ascribed to the
oxidation of CoCp(CO)2 and reduction of [CoCp(PPh3)2]þ,
respectively.
Figure 2. Cyclic voltammograms obtained at a scan rate of
0.1 V/s for 2.5 mM 1 in CH2Cl2/0.1 M [NBu4][B(C6F5)4] using
a 1.5 mm diameter GC disk electrode at 298 K. Waves B and C
fall at the potentials of the oxidation of CoCp(CO)2 and reduc-
tion of [CoCp(PPh3)2]þ, respectively.
Among the factors favoring the second-order pathway is the
extended reaction time inherent to bulk anodic electrolysis.
Room-temperature CV scans of 1 at a low concentration
(0.1 mM) in CH2Cl2/0.1 M [NBu4][TFAB] (Figure 1) from
0.1 to 1 V s-1 establish that CoCp(CO)(PPh3) undergoes
a diffusion-controlled (Do=1.2 ꢀ10-5 cm2 s-1) one-electron
oxidation (E1/2 = -0.38 V, Table 1) that is reversible in
both the electrochemical and chemical senses.22 Neither 1
nor 1[TFAB] undergoes detectable specific adsorption on
the glassy carbon electrode.23 At higher concentrations, the
oxidation of 1 displays less chemical reversibility at room tem-
perature,24 and two new product waves are seen (Figure 2),
both of which are readily assigned. Wave B falls at the
potential of the oxidation of CoCp(CO)2 (Epa = 0.37 V),11
and the reversible set at E1/2 =-1.23 V (wave(s) C) is that of
the 2/2þ couple.6 The sizes of the features B and C grew
relative to A as the concentration of 1 was increased up to a
maximum of 3.7 mM.
scans identifying only three products: 1þ (wave A), CoCp-
(CO)2 (3, wave B), and 2þ (wave C). Typical results are
shown in the linear scan voltammograms (LSVs) of Figure 3.
The approximately 1:6 ratio of limiting currents for wave A
versus wave B implies a yield of about 7-8% 1þ and 46%
each of CoCp(CO)2 and (by balance) 2þ. The fact that the
LSV limiting current for wave C is slightly less than that of
wave B is likely due to a lower diffusion coefficient for the
considerably larger disubstituted PPh3 complex.
On the basis of the historical precedent for the anodic
oxidation of CoCp(CO)2 under similar conditions,11 we
believe that a “radical-substrate” reaction is likely to be
responsible for the electrolytic behavior of 1. In this mechanism,
the radical cation 1þ produced in the initial oxidation (eq 1)
reacts with the starting material 1 (eq 2), producing a transient
dimer radical, 12þ, which then undergoes ligand exchange to
give the final products (eq 3).
Bulk anodic electrolyses were carried out for 1-3 mM
solutions of 1 between 273 K and ambient temperatures.
In each case, exhaustive electrolysis at Eappl = 0 V resulted in
passage of only 0.5 F of charge as the solution changed color
from light to dark yellow, with postelectrolysis voltammetry
½CoCpðCOÞðPPh3Þꢁþ þ CoCpðCOÞðPPh3Þ
1þ
1
h }½Co2Cp2ðCOÞ ðPPh3Þ ꢁþ}
ð2Þ
2
2
þ
12
}½Co2Cp2ðCOÞ ðPPh3Þ ꢁþ} h ½CoCpðPPh3Þ ꢁþ þ CoCpðCOÞ
(22) ΔEp = 64-69 mV; irev/ifd = 0.91 to 0.96. See sources in ref 12 for
the relevance of these mechanistic criteria.
2
2
2
2
2þ
þ
3
12
(23) This conclusion is based on Anson plots obtained for double
potential step chronocoulometry data for 1.1 mM 1 and step times of 1 s.
For data, see Figure S1 in the Supporting Information. For theoretical
treatment, see: (a) Anson, F. C. Anal. Chem. 1966, 38, 54. (b) Anson,
F. C.; Osteryoung, R. J. Chem. Educ. 1983, 60, 293.
ð3Þ
þ
Possible structures of the transient dimer radical 12
,
as well as alternative mechanisms, will be treated in the Dis-
cussion section. Concentrating here on corroborating the
dominant overall chemical reaction as that given in eq 4,
(24) At 253 K and a scan rate of 1 V s-1, the follow-up product waves
B and C are virtually absent for a 1.1 mM solution of 1. This CV scan is
available in the Supporting Information (Figure S2).

Article
Organometallics, Vol. 29, No. 19, 2010 4279
Figure 5. Experimental (solid line) and simulated (circles)
background-subtracted cyclic voltammograms of 1.04 mM 1
in CH2Cl2/0.1 M [NBu4][B(C6F5)4] at a scan rate of 1 V/s, 298 K,
and 1.5 mm diameter GC disk electrode. Relevant simula-
tion parameters are as follows: for 1/1þ couple (eq 5), E1/2
=
-0.38 V, ks = 0.05 cm s-1, R = 0.5; for R-S coupling and
ligand exchange of eq 6, kf = 6 ꢀ 102 M-1 s-1 and Keq g 1. A
follow-up reaction was added for [CoCp(CO)2]þ (see ref 11) in
order to fit the chemical irreversibility of the anodic wave
at 0.37 V.
verified small amount of 1þ (vide infra), representing a shift
of þ144 cm-1 from the νCO frequency of 1. Relevant spectral
data are collected in Table 2.
Figure 4. (a) Square-wave voltammograms of 2.5 mM 1 in
CH2Cl2/0.05 [NBu4][B(C6F5)4] recorded during bulk electro-
lysis at Eappl = 0 V and 273 K using a frequency of 25 Hz and 1 mm
GC disk electrode. (b) In situ IR spectroelectrochemistry moni-
toring of the electrolysis products.
Bulk oxidation of 1 was also carried out in the presence of
an equivalent of PPh3. In this case, a full one-electron process
was observed (0.97 F measured) and the only product formed
was the disubstituted product 2þ. This was confirmed by
several measurements on the electrolysis solution: voltam-
metric scans, an IR spectrum showing the absence of metal-
carbonyl bands, and a room-temperature ESR spectrum
having giso=2.162, in agreement with literature for 2þ.6
The voltammetric features of the CV scans were digitally
simulated (Figure 5; simulations at other scan rates in the
Supporting Information as Figure S4) using the reactions in
eqs 5-8, the first of which is a rewrite of eq 1. Equation 6
compresses eqs 2 and 3 into a single follow-up reaction, and
eqs 7 and 8 account for currents from the electroactive pro-
ducts. Changing the equilibrium and rate constants for the
gross reaction of eq 6 affected both the reverse current for the
1/1þ process and the currents calculated
Table 2. Spectral Information Relevant to This Studya
compound
ν
CO (cm-1
)
giso
reference
CoCp(CO)(PPh3), 1
[CoCp(CO)(PPh3)]þ, 1þ
[CoCp(PPh3)2]þ, 2þ
CoCp(CO)2, 3
1919
2063
n.a.
n.a.
this work
this work
6
2.134
2.162
n.a.
1958, 2024
11, this work
a IR and ESR spectra were recorded in dichloromethane at room
temperature.
we turn to IR spectroelectrochemical results. The fiber-optic
method21 of recording IR spectra is
CoCpðCOÞðPPh3Þ - 0:5e- h 0:5½CoCpðPPh3Þ2ꢁþ
þ 0:5CoCpðCOÞ2
ð4Þ
CoCpðCOÞðPPh3Þ - e- h½CoCpðCOÞðPPh3Þꢁþ ð5Þ
attractive for following complex redox processes, especially
those that proceed somewhat slowly on the CV time scale.25
In the present case, CO-region IR spectra and square-wave
voltammograms were recorded as a function of electrolysis
time (Figure 4). The square-wave voltammograms (top of
Figure 4) show the rise in concentration of the disubstituted
product 2þ as the reactant 1 is depleted. The IR spectra
½CoCpðCOÞðPPh3Þꢁþ þ CoCpðCOÞðPPh3Þ
Keq,kf
½CoCpðPPh3Þ ꢁþ þ CoCpðCOÞ
ð6Þ
ð7Þ
ð8Þ
F
R
2
2
CoCpðCOÞ2 - e- h ½CoCpðCOÞ2Þꢁþ
(bottom of Figure 4) show the growth of CoCp(CO)2 (νCO
1958, 2024 cm-1) as CoCp(CO)(PPh3) disappears (νCO
=
=
½CoCpðPPh3Þ2ꢁþ þ e- h CoCpðPPh3Þ2
1919 cm-1). The isosbestic point near 1945 cm-1 is consistent
with the cleanliness of the reaction. The very small feature
close to 2063 cm-1 is consistent with the electrochemically
for CoCp(CO)2 and [CoCp(PPh3)2]þ. For the sake of simpli-
city, equal diffusion coefficients were employed for all the
species involved. We note that the simulation’s Keq and kf values
(see caption of Figure 5) are a calculational tool meant to
demonstrate the overall conversion from 1 to [CoCp(PPh3)2]þ
(25) For an example of such an application, see: Stoll, M. E.; Geiger,
W. E. Organometallics 2004, 23, 58.

4280 Organometallics, Vol. 29, No. 19, 2010
Nafady and Geiger
Scheme 1
known about the only related dimer radical of which we aþre
aware, namely, the all-carbonyl analogue [CoCp(CO)2]2þ, 32
,
Figure 6. IR spectra recorded at ambient temperature before
(solid line) and after (dashed line) chemical oxidation of 0.0045
mmol of 1 with 1 equiv of [FeCp2][B(C6F5)4] in cold CH2Cl2.
this is not surprising. The dimer 32þ complex undergoes partial
dissociation to the parent mononuclear radical ([CoCp(CO)2]þ,
3þ) and the 18-electron substrate (CoCp(CO)2, 3) (see eq 9).11
Although the associative process in eq 9 is
and CoCp(CO)2 on the CV time scale and are not likely to be
unique.
CoCpðCOÞ þ ½CoCpðCOÞ ꢁþ h ½Co2Cp ðCOÞ ꢁþ ð9Þ
Chemical Oxidation of CoCp(CO)(PPh3) by Ferrocenium
Ion. With the intention of approximating homogeneous
aspects of the electrochemical experiments, the chemical
oxidation of 1 by [FeCp2][TFAB] was carried out with either
a half-equivalent or a full equivalent of ferrocenium ion.
When 0.5 equiv of [FeCp2]þ was added to a solution of 1
in CH2Cl2, followed by 5 min of stirring, IR spectra (Figure
2 þ
2
2
4
3þ
32
3
favored (Keq = 3 ꢀ 104 at room temperature), the replace-
ment of a CO ligand by PPh3 is expected to significantly shift the
equilibrium in favor of the monomers. Strong literature pre-
cedents exist for the weakening of metal-metal bonded dimers
by ligands of decreasing acceptor ability and increasing cone
angles in [CrCp(CO)2L]n and [Fe(η5-C5Ph5)(CO)2]n (n = 1 or
2) complexes.26,27 For this reason, the dimer radical
[Co2Cp2(CO)2(PPh3)2]þ is unlikely to be found in high con-
centrations at any point in the oxidation process.
S3, top) indicated yields of about 45% CoCp(CO)2 (νCO
=
1958, 2024 cm-1) and 10% 1þ (νCO = 2063 cm-1), consistent
with the bulk electrolysis yields described above. The implied
presence of radical 2þ was confirmed by ESR spectroscopy
(Figure S3, bottom), the major peak at giso = 2.1626 having a
weak higher-field shoulder most likely arising from the small
amount of 1þ.
A radical-radical coupling mechanism of 1þ with another
1þ must also be considered, especially since the metal-metal
2þ
bond of the dimer dication [CoCp(CO)L]2
should be
When the same experiment was conducted using one equiv-
alent of [FeCp2]þ, however, this situation reversed itself,
with 1þ being the major product (ca. 85%, assuming equal
extinction coefficients for the νCO bands of 1 and 1þ). A
relatively small amount of CoCp(CO)2 was formed (15%
based on CoCp(CO)2 IR intensity, Figure 6). Presumably,
the νCO-silent 2þ was also present. A room-temperature ESR
stronger than that of the monocation owing to removal of
an additional electron from an antibonding orbital. In terms
of voltammetry, diagnostic criteria exist that, in principle,
allow differentiation between radical-substrate (R-S) and
radical-radical (R-R) dimerization reactions.28 However,
the coupling reaction of the present system occurs too slowly
to take advantage of these diagnostics, requiring less direct
arguments to be made on the question of R-S versus R-R
coupling. A prime reason for favoring R-S coupling is
found in the electrochemically based stoichiometry of 0.5
electron required to complete the oxidation of 1. As the radi-
cal 1þ is produced, it diffuses away from the electrochemical
reaction layer into the bulk of solution, where, at least until
later in the bulk electrolysis, the neutral complex is in excess
and available for coupling.
spectrum of this solution gave a strong resonance at giso
=
2.134, which was assigned to 1þ (the similar radical [CoCp-
(CO)(PCy3)]þ has the same giso6). Thus, the dominant
chemical oxidation product of 1 in CH2Cl2 is shown to vary
with the relative amount of the oxidizing agent, 1þ and 2þ
being favored, respectively, for full-equivalent and half-
equivalent stoichiometries of oxidant.
Discussion
(26) (a) Morton, J. R.; Preston, K. F.; Cooley, N. A.; Baird, M. C.;
Krusic, P. J.; McLain, S. J. J. Chem. Soc., Faraday Trans. 1987, 83, 3535.
(b) O'Callaghan, K. A. E.; Brown, S. J.; Page, J. A.; Baird, M. C.; Richards,
T. C.; Geiger, W. E. Organometallics 1991, 10, 3119. (c) Kuksis, I.; Baird,
M. C. Organometallics 1994, 13, 1551.
(27) For leading references to other CrCp(CO)2L monomer/dimer
reactions see: Weng, Z.; Goh, L. Y. Acc. Chem. Res. 2004, 37, 187.
(28) (a) Parker, V. D. Acta Chem. Scand. 1998, 52, 154. (b) Andrieux,
Scheme 1 shows a radical-substrate reaction mechanism
that accounts for the oxidation products of 1 under the
range of experimental conditions that we and others have
employed. The first line in the Scheme, wherein 1þ undergoes
a second substitution process to furnish 2þ, applies only
to cases in which the oxidation of 1 is carried out in the
presence of additional PPh3. The putative dimer radical
intermediate, [Co2Cp2(CO)2(PPh3)2]þ (12þ), formed in the
radical-substrate reaction of eq 2, has not been detected by
either voltammetry or spectroscopy. On the basis of what is
ꢀ
C. P.; Grzeszczuk, M.; Saveant, J.-M. J. Electroanal. Chem. 1991, 318, 369.
(c) Alden, J. A.; Cooper, J. A.; Hutchinson, F.; Prieto, F.; Compton, R. G.
ꢀ
J. Electroanal. Chem. 1997, 432, 63. (d) Amatore, C.; Pinson, J.; Saveant,
J.-M. J. Electroanal. Chem. 1982, 137, 143. (e) Andrieux, C. P.; Nadjo, L.;
ꢀ
Saveant, J.-M. J. Electroanal. Chem. 1973, 42, 223.

Article
It is possible to get a half-electron stoichiometry from a
Organometallics, Vol. 29, No. 19, 2010 4281
Scheme 2
radical-radical coupling reaction if ligand transfer occurs in
the dimer dication (eq 10), giving a half-equivalent of
[CoCp(CO)2]þ, which would be reduced back to the neutral
complex as it re-entered the electrode reaction layer (eq 11).
However, this would require a substantial presence of
2½CoCpðCOÞðPPh3Þꢁþ h ½CoCpðPPh3Þ ꢁþ
2
1þ
2þ
þ ½CoCpðCOÞ2ꢁþ
ð10Þ
ð11Þ
3þ
½CoCpðCOÞ2ꢁþ þ e- h CoCpðCOÞ2
3þ
3
[CoCp(CO)2]þ or its own dimeric products11 in the bulk of
solution, neither of which was observed in the IR experi-
ments.
has gone into discerning the mechanism of ligand transfer in
these reactions, in particular on the complex question31 of
“single” electron transfer versus electrophile/nucleophile
([Co(CO)3(PPh3)2]þ/[Co(CO)4]- in eq 12) mechanisms.
The chemical oxidations are also explicable in terms of
R-S coupling. When a half-equivalent of ferrocenium is
employed, mixing of 1þ with unreacted 1 produces the same
complete reaction mixture as observed by exhaustive electro-
lysis. However, when a full equivalent of ferrocenium is
used, 1þ is formed more rapidly due to the fast outer-sphere
reaction between 1 and [FeCp2]þ. This results in higher
localized concentrations of 1þ, discriminating against the
R-S coupling process and producing a high yield of the
primary radical 1þ. We conclude that a radical-substrate
reaction gives a better accounting of the essential experi-
mental data.
Returning to the postulated intermediate [Co2Cp2(CO)2-
(PPh3)2]þ, 12þ, the rarity of information on dimetallic dimer
radicals unsupported by bridging ligands encumbers an expla-
nation of how the ligand transfer occurs between the metal
centers. Perhaps the closest body of literature on ligand-
transfer processes is that involving reactions of cation/anion
pairs.28-30 Equation 12 gives an example of such a process,
sometimes referred to
þ
In the present study 12 differs from the “ion-pair”
dinuclear complexes by its possession of an odd number of
electrons. Scheme 2 offers one possible sequence, wherein
formation and cleavage of the dimer radical cation might lead
to ligand transfer. It begins with a radical-substrate structure
(A) in which the highly electron rich33 1 donates an electron
pair to 1þ, resulting in loss of PPh3, formation of a weak
metal-metal bond, and transfer of the positive charge to the
remaining phosphine-substituted metal (structure B). Transfer
of a carbonyl ligand, either intramolecularly (structure C) or
by CO dissociation, can then occur, giving the final products.
The present study gives credence to growing evidence for
the influence of weak to modest strength metal-metal inter-
actions on the properties and reactions of cation radicals of
metal sandwich34 and half-sandwich11,35,36 complexes.
Acknowledgment. The authors gratefully acknowledge
the support of the National Science Foundation (most
recently, CHE-0808909) and thank Boulder Scientific
Corp. for partial support.
-
½CoðCOÞ ðPPh3Þ ꢁþ þ ½CoðCOÞ ꢁ
3
2
4
f PPh3ðCOÞ3Co- CoðCOÞ3PPh3 þ CO
as an “ion-pair annihilation by electron transfer”.29 This
reaction is essentially the CO-promoted reverse of eq 3, but
with a total of one more valence electron. A significant effort
ð12Þ
Supporting Information Available: Three figures (Figures
S1-S4) showing voltammetry curves, IR and ESR spectra,
and four additional CV simulations. This material is available
free of charge via the Internet at http://pubs.acs.org.
(32) (a) Kochi, J. K.; Bockman, T. M. Adv. Organomet. Chem. 1991,
33, 51. (b) Bockman, T. M.; Kochi, J. K. J. Phys. Org. Chem. 1997, 10, 542.
(33) Werner, H. Angew. Chem., Int. Ed. Engl. 1983, 22, 927.
(34) Swarts, J. C.; Nafady, A.; Roudebush, J.; Trupia, S. Inorg.
Chem. 2009, 48, 2156.
(35) Chong, D.; Laws, D. R.; Nafady, A.; Costa, P. J.; Rheingold,
A. L.; Geiger, W. E. J. Am. Chem. Soc. 2008, 130, 2692.
(36) Brownie, J. H.; Baird, M. C.; Geiger, W. E.; Laws, D. R.
Organometallics 2007, 26, 5890.
(29) (a) Manning, A. R. J. Chem. Soc. (A) 1968, 1135. (b) McCleverty,
J. A.; Davidson, A.; Wilkinson J. Chem. Soc. (A) 1969, 2610.
(30) (a) Masnovi, J. M.; Kochi, J. K. J. Am. Chem. Soc. 1985, 107,
7880. (b) Lee, K. Y.; Kochi, J. K. Inorg. Chem. 1989, 28, 567. (c) Lehman,
R. E.; Bockman, T. M.; Kochi, J. K. J. Am. Chem. Soc. 1990, 112, 458.
(31) (a) Atwood, J. D. Inorg. Chem. 1987, 26, 2918. (b) Zhen, Y.;
Atwood, J. D. J. Am. Chem. Soc. 1989, 111, 1506. (c) Zhen, Y.; Atwood,
J. D. J. Coord. Chem. 1992, 25, 229.