Role of a metal-metal bonded dimer dication in the one-electron oxidation of Rh(η5-C5H5)(CO)(PPh3) and related compounds

Article abstract of DOI:10.1021/om3003976

The anodic oxidation mechanism of RhCp(CO)(PPh₃), 1, has been studied in CH₂Cl₂/0.1 M [NBu₄][PF₆]. This complex and its analogue RhCp(PPh₃)₂ had been previously shown to form the fulvalenyl dirhodium complexes [Rh ₂FvL₂(PPh₃)₂]²⁺ (Fv = (η⁵,η⁵-C₁₀H₈), L = CO (2²⁺) or PPh₃) upon chemical oxidation. The present work investigated the reaction of 1 by variable-temperature electrochemistry and IR spectroelectrochemistry. The radical cation 1⁺ initially undergoes a radical-radical coupling reaction, giving the metal-metal bonded dimer dication [Rh₂Cp₂(CO)₂(PPh₃)₂] ²⁺ (5²⁺), which dominates at low temperatures. The room-temperature products are best accounted for by hydrogen atom transfer reactions of the dimer dication, affording 2²⁺ and the metal hydride [RhCp(CO)(PPh₃)H]⁺. The dimetalate complex [Rh ₂(σ: η⁵-C₅H₄) ₂(CO)₂(PPh₃)₂]²⁺ (7) may also be formed. The radical cation of the analogue RhCp(CO)(PPh₂Me) (3) undergoes very rapid formation of a similar metal-metal bonded dimer. A derivative with a large cone angle phosphine, RhCp(CO)(PⁱPr ₃) (4), does not show the same tendency toward oxidative dimerization. The monomer/dimer equilibrium [RhCp(CO)L]⁺ 1/2 [Rh ₂Cp₂(CO)₂L₂]²⁺ increasingly favors the dimer in the sequence L = PⁱPr₃ < PPh₃ < PPh₂Me < PMe₃, P(OPh) ₃, the latter two being based on earlier work. The implied dinuclear hydrogen atom transfer reactions are not mechanistically well understood, but find analogies in the chemistry of second- and third-row early transition metal complexes.

Full text of DOI:10.1021/om3003976

Article
pubs.acs.org/Organometallics
Role of a Metal−Metal Bonded Dimer Dication in the One-Electron
Oxidation of Rh(η5‑C5H5)(CO)(PPh3) and Related Compounds
Sabrina Trupia,*,† Thomas E. Bitterwolf,*,‡ and William E. Geiger*,†
†Department of Chemistry, University of Vermont, Burlington, Vermont 05445, United States
‡Department of Chemistry, University of Idaho, Moscow, Idaho 83844, United States
S
* Supporting Information
ABSTRACT: The anodic oxidation mechanism of RhCp(CO)(PPh3),
1, has been studied in CH2Cl2/0.1 M [NBu4][PF6]. This complex and
its analogue RhCp(PPh3)2 had been previously shown to form the
fulvalenyl dirhodium complexes [Rh2FvL2(PPh3)2]2+ (Fv = (η5,η5-
C10H8), L = CO (22+) or PPh3) upon chemical oxidation. The present
work investigated the reaction of 1 by variable-temperature electro-
chemistry and IR spectroelectrochemistry. The radical cation 1+ initially
undergoes a radical−radical coupling reaction, giving the metal−metal
bonded dimer dication [Rh2Cp2(CO)2(PPh3)2]2+ (52+), which
dominates at low temperatures. The room-temperature products are
best accounted for by hydrogen atom transfer reactions of the dimer
dication, affording 22+ and the metal hydride [RhCp(CO)(PPh3)H]+. The dimetalate complex [Rh2(σ:η5-C5H4)2(CO)2-
(PPh3)2]2+ (7) may also be formed. The radical cation of the analogue RhCp(CO)(PPh2Me) (3) undergoes very rapid formation
of a similar metal−metal bonded dimer. A derivative with a large cone angle phosphine, RhCp(CO)(PiPr3) (4), does not show
the same tendency toward oxidative dimerization. The monomer/dimer equilibrium [RhCp(CO)L]+ ⇌ 1/2 [Rh2Cp2(CO)2L2]2+
increasingly favors the dimer in the sequence L = PiPr3 < PPh3 < PPh2Me < PMe3, P(OPh)3, the latter two being based on earlier
work. The implied dinuclear hydrogen atom transfer reactions are not mechanistically well understood, but find analogies in the
chemistry of second- and third-row early transition metal complexes.
INTRODUCTION
■
As a family of molecules that displays a diverse set of properties
and reactions, 17-electron cyclopentadienyl complexes of the
cobalt group offer the opportunity for a deeper understanding of
the behavior of half-sandwich organometallic cation radicals. The
lifetimes of the radicals, their tendencies to undergo dimerization
by radical−radical or radical−substrate1 coupling, and their abilities
fulvalenyl-based dication [Rh2Fv(CO)2(PPh3)2]2+, Fv = (η5,η5-
to take part in substitution and atom-transfer reactions are
C10H8), structure A, L = PPh3 (22+).4 Similarly, RhCp(PPh3)2
gives the analogous dication [Rh2Fv(PPh3)4]2+.5 Dimerization
of the mononuclear radical cation [RhCpL(PPh3)]+ at either
the Cp ring or the metal center, followed by various options
for loss of hydrogen or H+, were discussed as mechanistic
possibilities.4,5 On the basis of recent studies that have added
to the understanding of the preferred reaction routes of
radical cations of cyclopentadienyl metal complexes,3,7−12 we
now report experiments that shed light on the mechanism
of formation of dinuclear products from 1+. The anodic
behavior of two other phosphine-substituted complexes,
RhCp(CO)(PPh2Me) (3) and RhCp(CO)(PiPr3) (4), was
included in the investigation. An overall scheme interrelating
the dominant reactions of [MCpLL′]+ (M = Co, Rh) is now
possible.
sensitive to their metal and ligand makeup.2 In an earlier paper,3
we described how a radical−substrate reaction between [CoCp-
(CO)(PPh3)]+ (Cp = η5-C5H5) and its 18-electron precursor gives
quantitative yields of the ligand-transfer products [CoCp(PPh3)2]+
and CoCp(CO)2. The present work details the anodic behavior of
the second-row congener RhCp(CO)(PPh3), 1, a complex known
to experience a quite different oxidative fate.4
Radical cations derived from RhCp(CO)L complexes are not
known to undergo the kind of ligand-transfer reaction
observed3 for [CoCp(CO)(PPh3)]+. Rather, they generally
form dinuclear products containing two electron-saturated
metals, based on formation of a bond between either the five-
membered rings4,5 (structure A) or the unbridged metals
(structure B).6 When L is a smaller cone angle phosphine or
phosphite, a relatively strong metal−metal bond may be
formed (dRh−Rh = 2.814 Å in [Rh2Cp2(CO)2(P(OPh)3)2]2+),6a
allowing the simple dimer dication of type B to dominate.
For 1, however, one-electron oxidation gives primarily the
Received: May 10, 2012
Published: August 21, 2012
© 2012 American Chemical Society
6063
dx.doi.org/10.1021/om3003976 | Organometallics 2012, 31, 6063−6070

Organometallics
Article
Table 1. Potentials versus Ferrocene of Redox Processes in CH2Cl2/0.1 M [NBu4][PF6]
a
compound
E1/2 or Ep
assignment
reference
b
RhCp(CO)(PPh3), 1
−0.01 V
−0.47 V
−0.08 V (Epa)
−0.40 V (Epc)
−0.12 V
−0.39 V (Epc)
0.24 V (Epa)
1/1+
this work
this work
this work
this work
this work
this work
6a
c
Rh2Fv(CO)2(PPh3)2, 2
RhCp(CO)(PPh2Me), 3
[Rh2Cp2(CO)2(PPh2Me)2]2+, 32
RhCp(CO)(P(iPr)3), 4
[Rh2Cp2(CO)2(PPh3)2]2+, 52+
RhCp(CO)(P(OPh)3), 9
RhCp(CO)(PMe3), 10
2/22+
2+
3 → 32
2+
2+
32 → 3
4/4+
52+ → 1
2+
9 → 92
2+
−0.12 V (Epa)
b
10 → 102
6a
a
c
Anodic or cathodic peak potential at scan rate of 0.1 V s−1. Value reported in ref 5a is −0.03 V when converted from SCE to FeCp20/+. Value
0/+
reported in ref 5a is −0.45 V when converted from SCE to FeCp2
.
employed [NBu4][PF6] as the supporting electrolyte. Table 1
lists the potentials relevant to this study.
RhCp(CO)(PPh3), 1. Cyclic voltammetry scans at fairly slow
scan rates and modest analyte concentrations (ca. 1 mM) are
consistent with those previously reported.4a Examples recorded
at two different temperatures are given in Figure 1. There is
EXPERIMENTAL SECTION
■
Chemicals. The compounds used or generated in this study were
handled under nitrogen, using either Schlenk or drybox conditions.
RhCp(CO)L (L = PPh3, 1; L = PPh2Me, 3; L = P(iPr)3, 4) and
Rh2Fv(CO)2(PPh3)2 (2)4b (Fv = (η5,η5-C10H8) were prepared by
literature methods.4,13 The dication [2][BF4]2 was prepared by
oxidation of 2 with acetylferrocenium tetrafluoroborate.14 Reagent-
grade dichloromethane was twice distilled from CaH2, the second
distillation being carried out, bulb-to-bulb, under static vacuum. The
glassware used for electrochemical experiments was cleaned in a No-
Chromix (Godax Laboratories, Inc.) solution for at least 12 h, followed
by copious rinsing with nanopure water. It was then dried overnight in
a 120 °C oven and transferred to the drybox antechamber while still
warm. [NBu4][PF6] (Acros) was recrystallized from ethanol and
vacuum-dried. NMR spectra were recorded using a Bruker ARX 500
MHz spectrometer. Elemental analyses were performed by Robertson
Laboratories.
Electrochemistry. For electrochemistry, a standard three-
electrode cell was employed in conjunction with a PARC 273A
potentiostat interfaced to a personal computer. The temperature of the
electrochemical experiment was regulated by immersion of the cell in a
controlled temperature hydrocarbon solution in the drybox.
Voltammetry scans were recorded using glassy carbon working
electrode disks of 1−2 mm diameter (Bioanalytical Systems), the
effective areas of which were measured by chronoamperometry on
ferrocene in acetonitrile/0.1 M [NBu4][PF6] (diffusion coefficient =
2.4 × 10−5 cm2 s−1). The disks were pretreated using a sequence
of polishings with Buehler diamond pastes of decreasing sizes (3 to
0.25 μm), interspersed by washings with nanopure water, and final
vacuum drying. Bulk electrolyses were carried out in an “H-type” cell
having counter and working compartments separated by a fine glass
frit. The working electrode was made of Pt gauze that had been treated
with nitric acid, washed copiously with distilled water, and dried
overnight at 120 °C. Potentials in this paper are referenced to the
ferrocene/ferrocenium couple14 obtained by the in situ method.15
Mechanistic aspects of the cyclic voltammetry (CV) data were probed
by applying the appropriate diagnostic criteria.16
IR Spectroelectrochemistry. Two different approaches were used
in obtaining IR spectra of anodic reaction products. For most
experiments, the fiber-optic “dip” approach initially reported by
Shaw17 was employed. Low-temperature experiments were generally
carried out using the IR-transparent thin layer electrode (IRTTLE)
cell described by Atwood.18
Figure 1. Background-subtracted CV scans of 1.04 mM 1 in CH2Cl2/
0.1 M [NBu4][PF6] at a scan rate of 0.2 V s−1 using a 1.5 mm diameter
glassy carbon electrode at two different temperatures.
limited chemical reversibility for the one-electron oxidation of 1
(E1/2 = −0.01 V vs ferrocene), with less reversibility observed at
lower temperatures and higher concentrations. It is important
to note that the follow-up product responsible for the cathodic
wave at Epc = −0.39 V is not the major long-term electrolysis
product (the fulvalenyl complex 22+), which has a similar
but measurably different potential and is chemically reversible
(Epc = −0.49 V, E1/2 = −0.47 V for 2/22+). Furthermore, the
CV product wave at Epc = −0.39 V is broad, chemically and
electrochemically irreversible, and shifts negatively with
RESULTS
■
Electrochemistry and IR Spectroscopy. Each of the
anodic electrochemical reactions obeyed the standard mecha-
nistic criteria of diffusion-controlled processes16 over the CV
scan rate range of 0.05 to 1 V s−1. Unlike a number of
organometallic cationic systems that are sensitive to the supporting
electrolyte anion,7−12,19 we observed no substantial difference in
the anodic voltammetry of 1, 3, or 4 when [B(C6F5)4]− was used
in place of [PF6]−. Hence, all work reported in this paper
6064
dx.doi.org/10.1021/om3003976 | Organometallics 2012, 31, 6063−6070

Organometallics
Article
increasing scan rates, none of which is true for the highly
reversible 2/22+ couple4a,20 (shown below) or, for that matter,
with other cobalt-group fulvalenyl complexes.21
The short-term product responsible for the cathodic wave
near −0.39 V is almost certainly the rhodium−rhodium bonded
dimer [Rh2Cp2(CO)2(PPh3)2]2+, 52+. The voltammetric
characteristics of this system match well with those of a
number of precedents6a,8,11 in which a sandwich or half-sandwich
radical cation dimerizes via metal−metal bond formation:
Figure 2. Steady-state (linear scan) voltammograms tracking solutions
during a sequence of electrolyses carried out at 233 K. The dashed line
marked (a) is the initial anodic scan of 1.0 mM 1 in CH2Cl2/0.1 M
[NBu4][PF6]. Scan (b) was recorded after exhaustive anodic
electrolysis at Eappl = 0.3 V. The solid line (c) was recorded after
exhaustive “reverse” electrolysis at Eappl = −0.6 V. Scan rate was 2 mV
s−1 at a 1.5 mm diameter glassy carbon electrode.
oxidation of the neutral monomer displays decreased
chemical reversibility at lower temperatures and higher
concentrations; the dimer product has a broad, electrochemi-
cally irreversible, cathodic reduction approximately 0.3 to 0.5 V
negative of the oxidation of the original 18-electron complex;
the starting material is efficiently regenerated by cathodic
reduction of the product.
This interpretation is supported by bulk electrolysis experi-
ments carried out between room temperature and 223 K. One-
electron coulomb counts (napp = 0.95 0.04 e−) were observed
in all cases, but the major product was temperature dependent.
Below 253 K, the electrolysis solution stayed dark blue for at
least an hour. Only one product wave was seen in the linear
scan voltammogram (LSV) (see (b) in Figure 2), at a potential
consistent with the reduction of 52+. When the electrolyzed
solution was rereduced at Eappl = −0.6 V, complex 1 was
regenerated in about 90% yield (Figure 2, solid line, c). This
overall chemically reversible redox behavior is reminiscent of
other oxidative dimerization products that regenerate the
starting material upon back-reduction.8,11,22 The reader may
note that in Figure 2 the limiting cathodic current for 52+ is
only about half the amount of the limiting anodic current of 1,
despite the overall faradaic equivalency of the two processes
(half an equivalent of a two-electron process for 52+). This is
ascribed to a lower diffusion coefficient for 52+, which has
ample literature precedent when redox pairs display significant
changes in charge and/or size.23 CV scans confirmed that the
major oxidation product was 52+ (Epc = −0.39 V, Figure 3) and
that only minor side products were present. Equation 1
accounts for the low-temperature electrochemical behavior of 1.
Figure 3. CV scan at 0.2 V s−1 of a 0.95 mM solution of 1 that had
been anodically electrolyzed (Eappl = 0.3 V) at 233 K in CH2Cl2/0.1 M
[NBu4][PF6].
At room temperature there was a more complex product
mixture, whether the anodic electrolysis was carried out at that
temperature or a low-temperature electrolysis solution was
allowed to warm. CV scans (Figure 4) showed a reversible
reduction at E1/2 = −0.45 V, within experimental error of that
measured for a bona fide sample of the fulvalenyl complex 22+.
Two other cathodic waves presented themselves near −1 V,
with LSV scans establishing roughly comparable amounts of the
sum of these side products and 22+. We again note that the
more easily reduced product at −0.45 V is not the simple dimer
52+. Furthermore, when the solution is rereduced at Eappl
=
−0.6 V, neutral 2 is formed without regeneration of the starting
material 1. However, when a rereduction potential negative of
the second set of products is employed (Eappl = −1.3 V), about
50% of the original amount of 1 is regenerated.
We therefore conclude that at room temperature (i) neither
the radical cation 1+ nor the simple dimer 52+ is present in
detectable amounts after exhaustive oxidation of 1, (ii) there is
about a 50% yield of the fulvalenyl complex 22+, and (iii) the
remaining group of oxidation products may be reduced
−2e−
+2e−
21 Hoooo=oo0o.o4ooVI 21+ ⇌ 52+ ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯→ 21
Eappl=−0.6V
Eappl
(1)
6065
dx.doi.org/10.1021/om3003976 | Organometallics 2012, 31, 6063−6070

Organometallics
Article
Figure 4. CV scan after anodic electrolysis of 2.3 mM 1 at 298 K
(0.2 V s−1 at 1.5 mm diameter Pt disk).
efficiently back to the starting material, 1. For the purpose of
discussion, we refer to the second set of products as side
products, although their yields are not small at room
temperature. Although we lack direct evidence of their
structures, literature precedents and IR data (vide infra)
provide likely candidates. A strong possibility is [RhCp(CO)-
(PPh3)H]+, similar to the hydrides produced when the simple
Figure 5. IR difference spectrum comparing absorptions before and
after anodic electrolysis of 2.6 mM 1 at 223 K using a fiber-optic
probe. The band at 1940 cm−1 is for the starting material, and those at
2032 and 2069 cm−1 are assigned to the dimer dication 52+.
8
dimers of either ruthenocenium ion11 or [ReCp(CO)3]+ are
allowed to warm to room temperature. Indeed, the hydride
[RhCp(PPh3)2H]+ was observed when RhCp(PPh3)2 was
chemically oxidized.5 The 17 e− rhenium complex [ReCp-
(CO)3]+ has also been reported to abstract a chlorine atom
from dichloromethane under similar conditions.8 Thus, 18-
electron cationic complexes such as 6 (X = H, Cl) are very
likely to be the major side products. Such complexes would
return cathodically to the starting material, consistent with
experiment, owing to loss of the X atom upon one-electron
reduction of 6 to the neutral 19 e− complex.
An additional possibility for a side product is the dimetalate
complex [Rh2(σ:η5-C5H4)2(CO)2(PPh3)2]2+ (7), analogues of
which are the major room-temperature oxidation products for
ruthenocene11 and osmocene24 under non-nucleophilic reac-
tion conditions.
Figure 6. Deconvolution of IR spectrum of room-temperature anodic
electrolysis products of 2.2 mM 1.
major products are identical at room temperature whether the
oxidation of 1 is carried out by electrolysis or by a chemical
oxidant.4 The fact that those two bands have the same relative
intensity when recorded on a sample of pure 22+ implies that the
comparatively more intense 2078 cm−1 band in Figure 6 arises
from overlapping absorptions.25 On that basis, the IR data
establish the presence of at least two metal−carbonyl containing
products in the room-temperature electrolysis solution, both (or
all) of which have one positive charge per metal atom.
Cathodic re-electrolysis of the room-temperature solution
regenerated the original 1940 cm−1 band, but did not allow for
determination of the relative amounts of 1 and 2, both of which
absorb at this energy. Thus, the IR data are consistent with the
electrochemical data, while not shedding new light on the identity
of the side products, except to confirm that they are cationic.
The IR response of the anodic electrolysis solution at 223 K
was obtained as an absorption difference spectrum using a fiber-
optic probe. As shown in Figure 5, the band for 1 at 1940 cm−1
disappears, leaving a new pair of bands at 2069 and 2032 cm−1,
which we assign to the dimer 52+. The average shift in νCO of
about +110 cm−1 is consistent with the dipositive charge of the
dimer. Consistent with the electrochemical data, these bands
were not present after room-temperature electrolysis. Instead, a
different pair of bands was observed, deconvolution of which
(dashed lines in Figure 6) suggested a minimum of three
absorbances, at 2078, 2054, and 2044 cm−1. The two higher
energy bands are within experimental error of those of
electrochemically verified 22+ (Table 2), showing that the
6066
dx.doi.org/10.1021/om3003976 | Organometallics 2012, 31, 6063−6070

Organometallics
Article
Table 2. Solution IR Spectral Data on Pertinent
Cyclopentadienyl and Fulvalenyl Rhodium Complexes
temperature (K),
ν(CO)
(cm−1
acquistion
a
compound
)
method
reference
RhCp(CO)(PPh3), 1
Rh2Fv(CO)2(PPh3)2, 22+
1940
298 or 223
this work
this work
2055, 2075 298, fiber-optic
probe
2049, 2069 ambient
ref 4a
[Rh2Cp2(CO)2(PPh3)2]2+, 52+ 2032, 2069 223, fiber-optic
this work
probe
RhCp(CO)(PiPr3), 4
1922
298 or 223
this work
this work
oxidation products of 4
2005, 2030, 298
2067, 2100
a
Fiber-optic IR probe recorded in situ from electrolysis solutions.
IRTTLE refers to an IR-transparent thin-layer electrode approach. See
Experimental Section for details.
Figure 8. CV scan taken after anodic electrolysis of 1.5 mM 3 at 233 K
(0.2 V s−1, 2 mm diameter Pt disk). The cathodic product wave at
Epc = −0.4 V is assigned to the metal−metal bonded dimer
[Rh2Cp2(CO)2(PPh2Me)2]2+ (322+).
RhCp(CO)(PPh2Me) (3) and RhCp(CO)(PiPr3) (4). Experi-
ments very similar to those described above for 1 were also
carried out on 3 and 4. The major points are given here.
Of the three compounds studied, only RhCp(CO)-
(PPh2Me), 3, showed no direct evidence for the mononuclear
radical cation on the CV time scale. At different concentrations
and scan rates, the only features in the CV scans were an
irreversible oxidation (Epa = −0.08 V) followed by a single
irreversible reduction (Epc = −0.40 V, Figure 7). This behavior
Eappl = −0.7 V regenerated 3 quantitatively. IR spectra of the
oxidized solutions were not recorded.
The triisopropylphosphine complex 4 displayed increased
stability of its radical cation compared to those of 1 and 3. At
low temperatures, CV scans showed reversibility for the 4/4+
couple (E1/2 = −0.12 V, Figure 9). Although cathodic peaks for
Figure 9. CV scan of 1.1 mM 4 at 233 K (1 V s−1, 1.5 mm diameter Pt
disk).
follow-up products were seen in CV scans at room temperature,
none of these waves had the position and the broad,
electrochemically irreversible shape that is now well-established
for reduction of the corresponding dimer dication. Bulk anodic
electrolysis of 4 at room temperature, passing 1 F of charge,
also failed to give any evidence of the dimer dication. At least
three major cathodic waves were observed for products (Figure
S1), and high yields of the starting material 4 were obtained
when these solutions were re-electrolyzed at negative applied
potentials. Under these conditions, therefore, the oxidation of 4
gives products that can be reduced back to the original
Figure 7. CV scan of 2.8 mM 3 at room temperature (0.1 V s−1,
1.5 mm diameter Pt disk).
is typical of a rapid radical−radical dimerization of 3+ giving a
2+
Rh−Rh bonded dimer dication, 32 (eq 2).
2RhCp(CO)(PPh2Me)
3
−2e−
HooooI 2[RhCp(CO)(PPh2Me)]+
3+
fast
complex, but there is no evidence for involvement of the dimer
⎯⎯→ [Rh2Cp (CO)2(PPh2Me)2]2+
2+
2
dication 42
.
2+
32
(2)
IR spectra recorded on the anodically electrolyzed solutions
of 4 were active in the expected cationic region (Supporting
Information Figure S2), revealing major absorptions at 2030
and 2067 cm−1, along with low-intensity features at 2005 and
2100 cm−1. All of these bands disappeared after the back-
electrolysis, giving only the 1922 cm−1 band of 4. Although the
Bulk anodic electrolysis (1 F passed) gave red solutions of
the apparent dimer dication at both room temperature
and 233 K. Side products were absent in this case at low
temperature and barely present at room temperature (see
the postelectrolysis CV of Figure 8). Back-electrolysis at
6067
dx.doi.org/10.1021/om3003976 | Organometallics 2012, 31, 6063−6070

Organometallics
Article
complexity of the product distribution discouraged us from
pursuing details, it appears that the anodic oxidation of the
triisopropylphosphine complex 4 gives positively charged
complexes analogous to 6, which can be reduced back to the
starting material in an overall chemically reversible process.
Mechanism. The radical cation 1+ formed by one-electron
oxidation of RhCp(CO)(PPh3) undergoes rapid dimerization
to give a metal−metal bonded dication, 52+, which is the
favored product at low temperatures. At room temperature, or
when solutions of 52+ are allowed to warm to room
temperature, the best-characterized product is the fulvalenyl
complex, 22+, containing two fewer hydrogen atoms than 52+.
The required loss of hydrogen may well occur by an
intramolecular H-atom transfer, as indicated in Scheme 1.
Scheme 2
Scheme 1
this would give the radical 8, the latter providing the neutral
fulvalenyl complex 2, which would undergo oxidation to give
the observed dication 22+. The overall electron transfer
stoichiometry would remain as one electron. However,
although there is considerable literature precedent26−28 for
thermally and photochemically driven cyclopentadienyl C−H
activation, we are unaware of such a reaction being induced by
an electron-transfer process.
DISCUSSION
■
The primary reaction of the 17 e− cation [RhCp(CO)L]+
involves radical−radical coupling29,30 to give a rhodium−
rhodium bonded unbridged dimer dication. Although only a
limited number of two-electron ligands (L) have been probed,
it appears that the monomer/dimer equilibrium (eq 3 below) is
quite sensitive to steric constraints. Combining the present data
with earlier reports,4−6a the dimer is increasingly favored in the
sequence L = PiPr3 < PPh3 < PPh2Me < PMe3, P(OPh)3. This
sequence qualitatively parallels that of isoelectronic neutral
complexes of the chromium group, [MCp(CO)2L]n (M = Cr,
Mo, W, n = 1, 2).31
This reaction, which is a combined reductive elimination/
oxidative addition process, gives the metal hydride 6, which was
suggested by postelectrolysis CV results, along with the formal
diradical [Rh(η5-C5H4)(CO)(PPh3)]+, 8+. The latter could
react to form a stable dinuclear complex, either via ligand−
ligand dimerization giving the observed 22+ or via metal−ligand
dimerization giving the postulated bis(σ:η5-C5H4) dimetalate
isomer 7. The hydride 6 (X = H) can also be formed in a
homolytic side reaction with the solvent.
Although there are precedents for intramolecular hydrogen-
transfer reactions for other dinuclear cyclopentadienyl metal
complexes,11,24,26,27 the transition-state structures for such
processes have not been investigated. One possibility on
going from 52+ to 8+ and 6 is the asymmetric structure 9, which
facilitates the single H-atom transfer required to give the ligand-
based radical 8+, antecedent to formation of the fulvalenyl
complex 22+.
A good number of the types of reactions that have been
identified for 17 e− cyclopentadienyl half-sandwich radical
cations (Scheme 3) lead to dinuclear products, the diversity of
which is especially rich for heavier metals, owing no doubt to
their greater tendency toward increased coordination numbers.
The processes described in the present work have parallels in
the oxidation of other heavier-row metal complexes, including
MCp*(Me2SO)Me2 (M = Rh, Ir),32 IrCp*(CO)2,33 ReCp-
( C O ) 3 , 8 R e C p ( N O ) ( P P h 3 ) M e , 3 4 a n d R h ( η 5 -
C5H4(anthracenyl))(CO)2.35 The recently described Ir-
(CO)2(acetylacetonate)36 oxidation may also be viewed in
this context. In each of those cases, a dinuclear product having
an unsupported metal−metal bond is formed.
Thought may also be given to an intermolecular hydrogen transfer
between the radical 1+ and the starting material 1 (Scheme 2), which
could come into play under some experimental conditions,
most notably at room temperature. Along with the hydride 6,
To some extent, the scenarios in Scheme 3 can be extended
to the reactions of electron-deficient second- and third-row
metallocenes, wherein unsupported metal−metal bonded
6068
dx.doi.org/10.1021/om3003976 | Organometallics 2012, 31, 6063−6070

Organometallics
Article
a
(4) (a) Connelly, N. G.; Lucy, A. R.; Payne, J. D.; Galas, A. M. R.;
Geiger, W. E. J. Chem. Soc., Dalton Trans. 1983, 1879. (b) Freeman, M.
J.; Orpen, A. G.; Connelly, N. G.; Manners, I.; Raven, S. J. J. Chem.
Soc., Dalton Trans. 1985, 2283.
Scheme 3
(5) (a) McKinney, R. J. Chem. Commun. 1980, 603. (b) McKinney,
R. J. Inorg. Chem. 1982, 21, 2051.
(6) (a) Fonseca, E.; Geiger, W. E.; Bitterwolf, T. E.; Rheingold, A. L.
Organometallics 1988, 7, 567. Unbridged Ir−Ir dimers of this general
type are also known. See: (b) Einstein, F. W. B.; Jones, R. H.; Zhang,
X.; Yan, X.; Nagelkerke, R.; Sutton, D. J. Chem. Soc., Chem. Commun.
1989, 1424. (c) Heinekey, D. M.; Fine, D. A.; Barnhart, D.
Organometallics 1997, 16, 2530.
(7) Camire, N.; Nafady, A.; Geiger, W. E. J. Am. Chem. Soc. 2002,
124, 7260.
(8) (a) Chong, D.; Nafady, A.; Costa, P. J.; Calhorda, M. J.; Geiger,
W. E. J. Am. Chem. Soc. 2005, 127, 15676. (b) Chong, D.; Laws, D. R.;
Nafady, A.; Costa, P. J.; Rheingold, A. L.; Calhorda, M. J.; Geiger, W.
E. J. Am. Chem. Soc. 2008, 130, 2692.
(9) Nafady, A.; Costa, P. J.; Calhorda, M. J.; Geiger, W. E. J. Am.
Chem. Soc. 2006, 128, 16587.
(10) Laws, D. R.; Chong, D.; Nash, K.; Rheingold, A. L.; Geiger,
W.E. . J. Am. Chem. Soc. 2008, 130, 9859.
(11) Swarts, J. C.; Nafady, A.; Roudebush, J.; Trupia, S.; Geiger, W.
E. Inorg. Chem. 2009, 48, 2156.
(12) Geiger, W. E.; Barriere, F. Acc. Chem. Res. 2010, 43, 1030.
̀
a
M is a Co-group metal, L and L′ are two-electron ligands, R = radical,
S = substrate.
(13) Graham, W. A. G.; Hart-Davis, A. J. Inorg. Chem. 1970, 9, 2658.
(14) Connelly, N. G.; Geiger, W. E. Chem. Rev. 1996, 96, 877.
(15) Gagne, R. E.; Koval, C. A.; Lisensky, G. C. Inorg. Chem. 1980,
19, 2854 Owing to the fact that some compounds or reaction
products had potentials near that of ferrocene, cobaltocenium
hexafluorophosphate (Strem) was sometimes used as the in situ
reference couple. In these cases, the potential vs [CoCp2]0/+ was
converted to [FeCp2]0/+ by subtraction of 1.33 V.
(16) Geiger, W. E. In Laboratory Techniques in Electroanalytical
Chemistry; Kissinger, P. T.; Heineman, W. R., Eds.; Marcel Dekker,
Inc.: New York, 1996; pp 683−717. See also: Bard, A. J.; Faulkner, L.
R. Electrochemical Methods, 2nd ed.; John Wiley & Sons: New York,
2001; pp 471− 531.
(17) Shaw, M. J.; Geiger, W. E. Organometallics 1996, 15, 13.
(18) Atwood, C. G.; Geiger, W. E.; Bitterwolf, T. E. J. Electroanal.
Chem. 1995, 397, 279.
dimers or dinuclear metal−ring bonded complexes have been
observed for complexes of Nb,37 Mo,38 W,38 Re,39 Ru,11 and
Os.24 Detailed structures of intermediates or transition states
that facilitate dinuclear hydrogen transfer reactions in this
broad family of complexes do not appear to be available. This is
certainly an aspect of this chemistry that would profit from
computations similar to those reported for ligand−ligand based
H-atom transfer in mononuclear complexes.40
ASSOCIATED CONTENT
■
S
* Supporting Information
One figure showing voltammetry (Figure S1) and two showing
IR spectra (Figures S2 and S3). This material is available free of
charge via the Internet at http://pubs.acs.org.
(19) Hill, M. G.; Lamanna, W. M.; Mann, K. R. Inorg. Chem. 1991,
30, 4687.
(20) The voltammetry of both 2 and 22+ is consistent with a pseudo-
Nernstian two-electron process, with peak shapes and peak separations
at slow scan rates being close to those expected for a two-electron
process (e.g., Epc − Epc/2 = 31 mV at a scan rate of 0.05 V s−1).
(21) (a) Chin, T. T.; Geiger, W. E. Organometallics 1995, 14, 1316.
(b) Chin, T. T.; Rheingold, A. L.; Geiger, W. E. J. Am. Chem. Soc.
1996, 108, 5002. (c) Chin, T. T.; Grimes, R. N.; Geiger, W. E. Inorg.
Chem. 1999, 38, 93. (d) Nafady, A.; Chin, T. T.; Geiger, W. E.
Organometallics 2006, 25, 1654. (e) Nafady, A.; Geiger, W. E.
Organometallics 2008, 27, 5624.
(22) Original example: Dessy, R. E.; Weissman, P. M.; Pohl, R. L. J.
Am. Chem. Soc. 1966, 88, 5117.
(23) See, for example: Kadish, K. M.; LaCombe, D. A.; Anderson, J.
E. Inorg. Chem. 1986, 25, 2246.
(24) Droege, M. W.; Harman, W. D.; Taube, H. Inorg. Chem. 1987,
AUTHOR INFORMATION
Corresponding Author
*E-mail: strupia@gmail.com; bitterte@uidaho.edu; william.geiger@
uvm.edu.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We are grateful for the funding of this work by the National
Science Foundation (most recently, NSF CHE-0808909) and
to a reviewer for suggesting the transition structure 9.
REFERENCES
■
26, 1309.
(1) In this context, “substrate” refers to the neutral 18-electron
complex from which the radical cation has been generated.
(25) As shown in the Supporting Information (Figure S3), the
carbonyl absorption bands of pure 22+ have equal intensity under these
sample conditions.
(26) Alvarez, M. A.; García, M. E.; Riera, V.; Ruiz, M. A. J. Am. Chem.
Soc. 1993, 115, 3786.
(27) Coville, N. J.; du Plooy, K. E.; Pickl, W. Coord. Chem. Rev. 1992,
116, 1.
(28) (a) Lee, S. Y.; Bergman, R. G. J. Am. Chem. Soc. 1995, 117, 5877.
(b) Ho, J.; Hou, Z.; Drake, R. J.; Stephan, D. W. Organometallics 1993,
12, 3145. (c) Ho, J.; Stephan, D. W. Inorg. Chem. 1994, 33, 4595.
(2) For general background concerning the electrochemistry of this
class of compounds, see: (a) Connelly, N. G.; Geiger, W. E. In Stone,
F. G. A.; West, R., Eds. Advances in Organometallic Chemistry;
Academic Press: Orlando, FL, 1984; pp 1−93. (b) Astruc, D. Electron
Transfer and Radical Processes in Transition-Metal Chemistry; VCH
Publishers: New York, 1995. (c) Baird, M. C. In Organometallic
Radical Processes; Trogler, W. C., Ed.; Elsevier: Amsterdam, 1990; pp
49−66. (d) Geiger, W. E. Organometallics 2007, 26, 5738.
(3) Nafady, A.; Geiger, W. E. Organometallics 2010, 29, 4276.
6069
dx.doi.org/10.1021/om3003976 | Organometallics 2012, 31, 6063−6070

Organometallics
Article
(d) Roggan, S.; Schnakenburg, G.; Limberg, C.; Sandhofner, S.;
̈
Pritzkow, H.; Ziemer, B. Chem.Eur. J. 2005, 11, 225.
(29) The dimerization may, in principle, follow either a radical−
radical (R−R, 1+−1+) or radical−substrate (R−S, 1+−1) pathway.
However, if the R−S coupled product 5+ was formed, it would be
immediately oxidized by the electrode to the corresponding dication.
Distinguishing between the R−R and R−S pathways is possible
through careful measurement of peak potential shifts as a function of
CV scan rate (see ref 30).
(30) The measured peak potential shift of 21.8 mV per 10-fold
increase in scan rate is closer to the value predicted for R−R
dimerization (20 mV) than that expected for R−S dimerization (30
mV),31 lending further weight to the assigned radical−radical based
dimerization of 1+.
(31) (a) Savea
(b) Andrieux, C. P.; Nadjo, L.; Savea
1970, 26, 147. (c) Nadjo, L.; Saveant, J.-M. J. Electroanal. Chem. 1973,
48, 113. (d) Amatore, C.; Garreau, D.; Hammi, M.; Pinson, J.; Saveant,
́
nt, J.-M.; Vianello, E. Electrochim. Acta 1967, 12, 1545.
́
nt, J.-M. J. Electroanal. Chem.
́
́
J.-M. J. Electroanal. Chem. 1985, 184, 1. (e) Crooks, R. M.; Bard. J.
Electroanal. Chem. 1988, 240, 253.
(32) The monomer/dimer equilibria of Cr(C5R5)(CO)3−nLn (R = H,
Me) have a rich history beginning with Adams, R. D.; Collins, D. E.;
Cotton, F. A. J. Am. Chem. Soc. 1974, 96, 749. Leading references may
be found in Brownie, J. H.; Baird, M. C.; Laws, D. R.; Geiger, W. E.
Organometallics 2007, 26, 5890.
(33) Fooladi, E.; Graham, T.; Turner, M. L.; Dalhus, B.; Maitlis, P.
M.; Tilset, M. J. Chem. Soc., Dalton Trans. 2002, 975.
(34) (a) Einstein, F. W. B.; Jones, R. H.; Zhang, X.; Yan, X.;
Nagelkerke, R.; Sutton, D. J. Chem. Soc., Chem. Commun. 1989, 1424.
(b) Yan, X.; Batchelor, R. J.; Einstein, F. W. B.; Zhang, X.; Nagelkerke,
R.; Sutton, D. Inorg. Chem. 1997, 36, 1237.
(35) Tilset, M.; Bodner, G. S.; Senn, D. R.; Gladysz, J. A.; Parker, V.
D. J. Am. Chem. Soc. 1987, 109, 7551.
(36) Carano, M.; Cicogna, F.; D’Ambra, I.; Gaddi, B.; Ingrosso, G.;
Marcaccio, M.; Paolucci, D.; Paolucci, F.; Pinzino, C.; Roffia, S.
Organometallics 2002, 21, 5583.
(37) Huang, H.; Rheingold, A. L.; Hughes, R. P. Organometallics
2009, 28, 1575.
(38) (a) Guggenberger, L. J.; Tebbe, F. N. J. Am. Chem. Soc. 1971,
93, 5924. (b) Guggenberger, L. J. Inorg. Chem. 1973, 12, 294.
(39) (a) Barral, M. C.; Green, M. L. H.; Jimenez, R. J. Chem. Soc.,
Dalton Trans. 1982, 2495. (b) Bashkin, J.; Green, M. L. H.; Poveda, M.
L.; Prout, K. J. Chem. Soc., Dalton Trans. 1982, 2485.
(40) Pasman, P.; Snel, J. J. M. J. Organomet. Chem. 1984, 276, 387.
(41) Herbert, B. J.; Baik, M.-H.; Green, J. C. Organometallics 2004,
23, 2658.
6070
dx.doi.org/10.1021/om3003976 | Organometallics 2012, 31, 6063−6070