Electrochemical Oxidation of W(CO)4(LL): Generation, Characterization, and Reactivity of [W(CO)4(LL)]+ (LL≤α-diimine ligands)

Article abstract of DOI:10.1071/CH17256

The electrochemistry of a series of W(CO)4(LL) complexes, where LL is an aromatic α-diimine ligand, was examined in coordinating and weakly coordinating media using several techniques. These compounds undergo metal-centred one-electron oxidations and the electrogenerated radical cations undergo a range of subsequent chemical steps, the nature of which depends on the substituents of the α-diimine ligand and the presence of coordinating species. In CH2Cl2/TBAPF6, where TBAPF6 is n-tetrabutylammonium hexaflurophosphate, the bulk oxidations are partially reversible at scan rates of 0.25Vs-1; the resulting tungsten(i) radicals react via disproportionation and loss of carbonyl, the rate constants for which were measured by double-potential step chronocoulometry. Large-amplitude a.c. voltammetry experiments suggest that the one-electron oxidized species are in equilibrium with the corresponding disproportionation products. Steric crowding of the metal centre prolongs the lifetime of the radical cations, allowing the infrared spectroelectrochemical characterization of two [W(CO)4(LL)]+ species. Electrogenerated [W(CO)4(LL)]+ cations are highly susceptible to attack by potential ligands; oxidations performed in CH3CN/TBAPF6, for example, were chemically irreversible. Kinetic studies in weakly coordinating media show that near-stoichiometric amounts of added pyridine and acetonitrile are enough to greatly diminish the reversibility of the bulk oxidations; the dominant path of the coupled chemistry depends on the ligand strength, with substitution being the major reaction with added pyridine, whereas disproportionation is favoured by the presence of acetonitrile. A reaction scheme that provides an overall framework of the reactions followed by the radical cations is presented and discussed in the context of the previously observed chemistry of the molybdenum analogues.

Full text of DOI:10.1071/CH17256

RESEARCH FRONT
CSIRO PUBLISHING
Aust. J. Chem. 2017, 70, 1006–1015
Full Paper
https://doi.org/10.1071/CH17256
Electrochemical Oxidation of W(CO)4(LL): Generation,
Characterization, and Reactivity of [W(CO)4(LL)]1
(LL ¼ a-diimine ligands)
A C
,
B
A
John P. Bullock,
Chong-Yong Lee, Brian Hagan,
A
A
Humair Madhani, and John Ulrich
A
Division of Natural Science and Mathematics, Bennington College,
Bennington, VT 05201, USA.
B
Australian Research Council Centre of Excellence for Electromaterials Science,
Intelligent Polymer Research Institute, Australian Institute for Innovative Materials,
University of Wollongong, Innovation Campus, Wollongong, NSW 2522, Australia.
C
Corresponding author. Email: jbullock@bennington.edu
The electrochemistry of a series of W(CO)4(LL) complexes, where LL is an aromatic a-diimine ligand, was examined in
coordinating and weakly coordinating media using several techniques. These compounds undergo metal-centred one-
electron oxidations and the electrogenerated radical cations undergo a range of subsequent chemical steps, the nature of
which depends on the substituents of the a-diimine ligand and the presence of coordinating species. In CH2Cl2/TBAPF6,
where TBAPF6 is n-tetrabutylammonium hexaflurophosphate, the bulk oxidations are partially reversible at scan rates of
0.25 V sꢀ1; the resulting tungsten(I) radicals react via disproportionation and loss of carbonyl, the rate constants for which
were measured by double-potential step chronocoulometry. Large-amplitude a.c. voltammetry experiments suggest that
the one-electron oxidized species are in equilibrium with the corresponding disproportionation products. Steric crowding
of the metal centre prolongs the lifetime of the radical cations, allowing the infrared spectroelectrochemical characteriza-
tion of two [W(CO)4(LL)]þ species. Electrogenerated [W(CO)4(LL)]þ cations are highly susceptible to attack by potential
ligands; oxidations performed in CH3CN/TBAPF6, for example, were chemically irreversible. Kinetic studies in weakly
coordinating media show that near-stoichiometric amounts of added pyridine and acetonitrile are enough to greatly
diminish the reversibility of the bulk oxidations; the dominant path of the coupled chemistry depends on the ligand
strength, with substitution being the major reaction with added pyridine, whereas disproportionation is favoured by the
presence of acetonitrile. A reaction scheme that provides an overall framework of the reactions followed by the radical
cations is presented and discussed in the context of the previously observed chemistry of the molybdenum analogues.
Manuscript received: 12 May 2017.
Manuscript accepted: 27 June 2017.
Published online: 20 July 2017.
Introduction
exhibited by these and related compounds, many of which are
chemically reversible. In contrast, the oxidative chemistry of
such species has received far less attention. Several reports
indicate that they tend to undergo one-electron oxidations, the
reversibility of which is highly dependent on the metal.[4–6] In
the case of tungsten, the chemical reactions coupled to the
electrode process are rapid enough to make characterization of
the electrogenerated species difficult, which has no doubt
contributed to the paucity of studies of the 17-electron radical
cations, [W(CO)4(LL)]þ. The chromium and molybdenum
analogues, however, have been examined to a greater extent.
The chromium compounds, in particular, have been shown to
undergo chemically reversible oxidations, and at least one such
compound, Cr(CO)4(tmp), where tmp ¼ 3,4,7,8-tetramethyl-
1,10-phenanthroline, has been spectroscopically characterized
by several techniques.[6] Oxidation of Mo(CO)4(bpy) was
shown to be chemically reversible only in non-coordinating
media.[7]
Group 6 tetracarbonyl complexes of the type M(CO)4(LL), where
LL is an aromatic a-diimine ligand, such as 2,20-bipyridine (bpy),
1,10-phenanthroline (phen), or a substituted derivative thereof,
comprise one of the most extensively studied classes of organ-
ometallic compounds.[1] Their intense luminescence, arising
from metal-to-ligand charge-transfer excited states, makes them
promising candidates for a variety of sensing and energy-
conversion schemes. They also have rich electrochemistry and
more recently have been examined as electrocatalysts for the
reduction of carbon dioxide.[2] Moreover, their structural sim-
plicity makes them useful for examining the more general
reactivity pathways of the ubiquitous M(CO)4 moiety.
Because the excited states of these compounds formally
involve an oxidized metal centre and a reduced a-diimine
ligand, electrochemical techniques have yielded significant
insights into various aspects of their chemistry.[3] Of particular
interest have been the one-electron reduction processes
Journal compilation Ó CSIRO 2017
www.publish.csiro.au/journals/ajc

Electrochemical Oxidation of W(CO)4(LL)
1007
Our laboratory previously investigated the oxidative electro-
chemistry of a variety of Mo(CO)4(LL) compounds to charac-
terize the reactivity of the corresponding electrogenerated radical
cations, [Mo(CO)4(LL)]þ.[8] The present paper describes our
extension of that work to the tungsten analogues and delineates
the chemistry of electrogenerated [W(CO)4(LL)]þ species. As
expected, these compounds share some similarities, but they also
exhibit distinct electrochemical responses under some conditions.
This work sheds light on the underlying reactivity patterns
exhibited by these compounds and provides a comprehensive
framework for the reaction pathways followed by the oxidation
products of this important class of Group 6 tetracarbonyl
compounds.
each a-diimine ligand as specified in the Materials section and in
Table 1. For clarity, references in the text to either 1 or 1þ should
be assumed to be representative of the entire class of com-
pounds, whereas those to individual compounds, e.g. 1a, are
provided in the context of experimental data or to properties
thought to be unique to a particular member of the series.
D.C. Cyclic Voltammetry Studies
Cyclic voltammograms of 1a and its molybdenum analogue,
Mo(CO)4(phen), each in two different solvents, are shown in
Fig. 1. As can be seen, the voltammetric responses of these
compounds depend on the media in which the experiments are
performed. In 0.1 M n-tetrabutylammonium hexafluorophosphate
(TBAPF6) dissolved in dichloromethane, a weakly coordinating
medium, the tungsten and molybdenum compounds exhibit
quasi-reversible one-electron oxidations (Fig. 1, top), whereas in
the more highly coordinating acetonitrile/TBAPF6, the bulk
Results and Discussion
Ten W(CO)4(LL) compounds were examined in this study and
are designated 1a–j herein, with the letters corresponding to
Table 1. Summary of electrochemical characterization of W(CO)4(LL) compounds
Cyclic voltammograms obtained using CH2Cl2/TBAPF6 at scan rates of 0.25 V sꢀ1; analyte concentrations were 1.0 ꢁ 0.1 mM
LL
W(CO)4(LL) letter designation
E8 [V versus Ccþ/0
]
ic/ia
EAp [V versus Ccþ/0
]
kBdisp [Mꢀ1 sꢀ1
]
phen
bpy
dmp
tmp
bph
ncp
bcp
nop
clp
1a
1b
1c
1d
1e
1f
1558
1564
1522
1500
1565
1528
1544
1629
1591
1514
0.65
0.78
0.58
0.66
0.73
0.92
0.85
0.65
0.74
0.66
0.349
0.386
370
240
C
0.307
–
0.360
780
710
0.319
D
–
ꢀ0.015
ꢀ0.057
0.487
1g
1h
1i
120
C
–
C
0.445
–
dmb
1j
0.293
790
ACathodic peak maximum assigned to the disproportionation product of 1þ
.
BRate constant for the disproportionation of 1þ as determined by double potential-step chronocoulometry; uncertainties for these values are estimated to
be ꢁ10 %.
CNot measured.
DChronocoulometric data did not satisfactorily fit the theoretical response for a disproportionation mechanism.
W(CO)4(phen)
Mo(CO)4(phen)
14 μA
20 μA
CH2Cl2/TBAPF6
CH2Cl2/TBAPF6
20 μA
20 μA
CH3CN/TBAPF6
CH3CN/TBAPF6
0
0.5
1.0
1.5
2.0
0
0.5
1.0
1.5
2.0
Potential [V vs Cc0/ϩ
]
Fig. 1. Cyclic voltammograms of 1a (left) and Mo(CO)4(phen) (right) in CH2Cl2/TBAPF6 (top row), and CH3CN/TBAPF6
(bottom row) obtained at a 3-mm diameter glassy carbon electrode at scan rates of 0.25 V sꢀ1; concentrations in each case are
2.0 mM, except for that shown in the top right panel, which was 1.4 mM.

1008
J. P. Bullock et al.
(a)
oxidation is irreversible for complexes of both metals (Fig. 1,
bottom). This behaviour is consistent with observations made
previouslyina thoroughstudy ofthe electrochemical oxidation of
Mo(CO)4(bpy) in several solvent/electrolyte systems[7] and also
parallels that seenamong the structurally analogous molybdenum
and tungsten tetracarbonyl diazadiene[9] complexes.
25 μA
The cyclic voltammogram of 1a in CH2Cl2/TBAPF6 shown
in Fig. 1 is typical for most of the complexes examined in the
present study measured under similar conditions; data for other
compounds are summarized in Table 1. The peak current ratios
of the reverse to forward scans, ic/ia, for the bulk oxidations
ranged from 0.6 to slightly more than 0.9 at scan rates of
0.250 V sꢀ1 and the oxidation potentials show expected trends
based on the nature of the substituents on the a-diimine ligands.
For example, the standard oxidation potentials, E8, are shifted
negatively, relative to unsubstituted diimine ligands, when
electron-donating methyl groups are present, such as with di-
and tetramethyl substituted phenanthroline, 1c and 1d, whereas
those with electron-withdrawing groups, such as nitro or choro,
1h and 1i, are shifted positively.
0
0.5
1.0
1.5
2.0
(b)
100 μA
We found that the tungsten and molybdenum M(CO)4(LL)
compounds show similar oxidative electrochemistry in weakly
coordinating media, but with some notable differences. For
example, the bulk oxidations of M(CO)4(LL) compounds for
both metals are coupled to multiple chemical processes. At a
scan rate of 0.250 V sꢀ1, the voltammogram of Mo(CO)4(phen)
exhibits three small coupled reduction peaks in addition to that
of the radical cation (Fig. 1, top right). In previous work,[8] we
assigned these peaks to reductions of a Mo2þ disproportionation
product and two different Moþ compounds with the general
formula [Mo(CO)3(LL)D]þ, formed via the loss of a CO ligand,
where D is the weakly coordinating PFꢀ6 supporting anion[10] in
one case and adventitious water in the other. Tungsten com-
pounds, in contrast, typically show only one coupled reduction
peak at this scan rate, other than that due to the radical cation.
We assign this peak to the reduction of a W2þ disproportionation
product, [W(CO)4(LL)(D)]2þ; potentials of the coupled reduc-
tion peak maxima are included in Table 1 and correlate with the
electron-donating or releasing nature of the a-diimine substitu-
ents. The generation of this compound likely takes place via the
reactions described in Eqns 1 and 2; a disproportionation of 1þ
initially yields 12þ, a 16-electron octahedral species that associ-
ates with a weak donor species, probably PFꢀ6 from the electro-
lyte. The existence of seven-coordinate tungsten(II) carbonyl
compounds such as these is well established,[11] and their
generation via electrochemical processes has previously been
reported.[12] In addition, the above reaction pathway is consis-
tent with the well-established tendency of d5 carbonyl species to
undergo disproportionation reactions, especially when stabili-
zation via isomerization is not possible.[13] Finally, it is worth
noting that voltammograms of the structurally similar cis-W
(CO)4(Z2-alkene) compounds measured in weakly coordinating
media bear a striking resemblance to that shown in Fig. 1 and
may be indicative that similar homogeneous coupled chemistry
is operating in those systems.[14]
0
0.5
1.0
1.5
2.0
Potential [V vs Cc0/ϩ
]
Fig. 2. (a) Cyclic voltammograms of 2.1 mM 1a obtained at 0.25 V sꢀ1
;
when the anodic sweep extends beyond the first oxidation process (black),
there is substantial current enhancement, relative to when the switching
potential is between the two anodic processes (red), of the peak we assign to
the tungsten(II) disproportionation product, at 0.32 V; (b) a scan rate study of
2.1 mM 1a; scan rates employedwere 0.05 (green), 0.25 (blue), 1.0 (red), and
5.0 (black) V sꢀ1; two small reduction processes, at ,0.3 and ,0.7 V, are
discernible at 1.0 and 5.0 V sꢀ1. The increased peak separation for the 1a0/þ
couple observed at higher scan rates is due to uncompensated resistance.
that is broad and irreversible. The species formed by the second
process, presumably 12þ, would be expected to be very electron-
deficient and highly Lewis acidic, making its subsequent asso-
ciation with an electron donor quite favourable, even with a
weakly coordinating species such as PFꢀ6 . Significantly, the
current of the coupled reduction peak at ,0.3 V versus the
cobalticenium/cobaltocene internal standard, Cc0/þ, is markedly
enhanced when the initial anodic sweep extends to the second
irreversible oxidation peak (Fig. 2a); [W(CO)4(LL)D]2þ can,
therefore, be generated via chemical and electrochemical routes.
Additional evidence that supports the above reaction scheme
includes the following. Spectroelectrochemical experiments
(see below) indicate that [W(CO)4(ncp)]þ (ncp ¼ 2,9-dimethy-
phenanthroline), 1fþ, decomposes to form the parent compound
1f, and a tungsten carbonyl species with an oxidation state
greater than þ1, presumably via disproportionation. We
also collected kinetic data, discussed in the next section, that
show most 1þ compounds follow a reaction pathway that is
second-order with respect to the radical cation.
Although the simplicity of the voltammogram of 1a in
CH2Cl2/TBAPF6 may suggest that disproportionation is the sole
reaction that 1þ radicals undergo in weakly coordinating media,
scan rate studies (Fig. 2b) reveal that they also follow a parallel
reaction pathway. Voltammograms of 1a obtained at 1.00 V sꢀ1
feature two small coupled reduction peaks in addition to those of
the 1a0/þ couple. We assign these as follows: the peak at the
more negative potential, at 0.32 V in Fig. 2b, is due to the
2þ
2½WðCOÞ4ðLLÞꢂþ Ð WðCOÞ4ðLLÞ þ ½WðCOÞ4ðLLÞꢂ
ð1Þ
ð2Þ
½WðCOÞ4ðLLÞꢂ2þ þ D ! ½WðCOÞ4ðLLÞðDÞꢂ
2þ
Consistent with the above proposed reaction scheme is the
fact that cyclic voltammograms of 1 show a second oxidation

Electrochemical Oxidation of W(CO)4(LL)
1009
disproportionation product [W(CO)4(phen)D]2þ, while that at
0.73 V is due to the reduction of [W(CO)3(LL)D]þ, formed via
loss of a carbonyl ligand and association with a donor species D.
We base the latter assignment on the similarity in the observed
potential to that seen with other ligand substitution products as
discussed below. The precise nature of D is not clear, as both PF6ꢀ
and H2O could plausibly complete the octahedral coordination
sphereofthe tungsten(I) metal centre. The factthat thispeakisnot
seenatlower scanrates indicates thatthese tricarbonyl species are
fairly short-lived; it is likely that if D is a poor ligand, decompo-
sition by additional loss of carbonyl ligands would result owing to
the electron-deficient nature of the metal centre. It is notable that
at higher scan rates, these two coupled reduction peaks have
similar currents, suggesting that the processes that give rise to
those signals occur at similar rates under the conditions
employed. This is consistent with chronocoulometric data,
described below, that indicate 1þ species react via parallel path-
ways at comparable rates, one of which is first-order, and the
other second, with respect to the radical cations.
In contrast to the partial reversibility of the oxidation of1a and
its molybdenum analogue in CH2Cl2/TBAPF6, the corresponding
processes in CH3CN/TBAPF6 are completely irreversible (Fig. 1,
bottom), indicating thatboththe Wþ and Moþ radicalsare rapidly
attacked by coordinating solvent molecules.[15] In addition, the
coupled chemistry of the electrogenerated cations is less similar
in this medium than was seen in the less coordinating solvent.
Specifically, although the oxidation of Mo(CO)4(phen) results in
the formation of the carbonyl substitution product, [Mo
(CO)3(phen)(CH3CN)]þ,[7,8] as evidenced by the large reversible
reduction at 1.2 V in Fig. 1, the peak assignable to the correspond-
ing [W(CO)3(phen)(CH3CN)]þ is detectable but has a much
smaller current. We describe additional experiments that shed
light on these observations later in this paper.
0.6
0.5
0.4
0.3
0.2
0.1
0
1.2
1.0
0.8
0.6
0.4
0.2
0
0
1
2
3
W(CO)4(phen) conc. [ϫ 103 M]
Ϫ2
Ϫ1
0
1
2
log keff [W(CO)4(phen)]τ
Fig. 3. Double potential-step chronocoulometry results obtained for 0.6
(’), 0.9 (¢), 1.6 (r), and 2.5 (ꢃ) mM 1a in CH2Cl2/TBAPF6. Step times, t,
ranged from 0.05 to 5 s. The potential step was 0.5 V and was centred on the
bulk oxidation peak of the corresponding cyclic voltammogram. Charge
ratios measured for each concentration were individually fitted to the
theoretical working curve (solid line) of a disproportionation reaction of
1aþ. Best-fit rate constants for a disproportionation, keff, were 900, 650, 550,
and 500 Mꢀ1 sꢀ1 (ꢁ,5 %, represented by the error bars in the inset figure)
for the four concentrations examined, from lowest to highest. Inset: keff[1a]
values as function of the bulk 1a concentration; the slope and intercept of this
plot yield estimates of kdisp and k1st-order respectively. From these data,
the following rate constants were estimated: kdisp ¼ 370 ꢁ 30 Mꢀ1 sꢀ1 and
k1st-order ¼ 0.28 ꢁ 0.03 sꢀ1
.
contributions from two distinct pathways. The decrease in
best-fit rate constants observed at higher concentrations is an
artefact that arises from fitting the data to an oversimplified
kinetic scheme. Specifically, at lower concentrations, the frac-
tion of 1þ that follows the second-order path is lower than at
higher concentrations; this, in turn, requires a greater rate
constant to match the observed rate if a purely second-order
reaction scheme is assumed. We estimated the rate constants for
Kinetic Studies of [W(CO)4(LL)]1 Disproportionation
To get further insight into the chemical reactions coupled to the
oxidation of 1, we performed a series of kinetic studies using
double potential-step chronocoulometry. In these experiments,
we measured the charge passed in timed steps over the bulk
oxidation process and the reduction of 1þ; potentials were
chosen such that no other coupled reductions would be effected.
Plots of the charge ratios of the reverse to forward steps, Qr/Qf,
against either log kt or log k[1]t, where k is the rate constant for
the operative reaction and t is the step time of the experiment,
were compared against working curves of several likely
mechanisms, including those for a simple first-order pathway
and a second-order disproportionation.[16]
Data obtained for several concentrations of 1a are presented
in Fig. 3 along with the theoretical response for a disproportion-
ation reaction. For a given step time, the charge ratios decreased
with increasing concentration of the bulk species, suggesting
that the coupled homogeneous chemistry is second-order with
respect to the cation radicals. Indeed, experimental data
obtained at each given analyte concentration closely fit the
working curve for a disproportionation. However, the best-fit
rate constants, which correspond to the rate at which 1aþ decays,
for each such set of measurements decreased with increasing
tungsten concentration. This is consistent with a scenario
involving simultaneous reactions involving 1aþ: a dispropor-
tionation and separate first-order reaction. The rate constants
employed in the x axis of the graph shown in Fig. 3, therefore,
represent effective overall rate constants, keff, that have
the disproportionation, kdisp, and the first order process, k1st-order
,
from the slope and intercept of a plot of keff[1a] versus [1a] to be
370 ꢁ 30 Mꢀ1 sꢀ1 and 0.28 ꢁ 0.03 sꢀ1 respectively (see Experi-
mental section for further information). Thus, at the concentra-
tions employed for these studies, 0.5 to 2.5 mM, these are
competitive processes, as suggested by the voltammetric scan
rate study. We suspect that dissociative substitution of carbon
monoxide is the major first-order event, as described earlier. An
important caveat concerning these rate constants should be
mentioned here, however. Additional experiments using large-
amplitude a.c. voltammetry, described below, indicate that 1þ is
in equilibrium with the corresponding disproportionation pro-
ducts; thus, these rate constants are not those for simple, irrevers-
ible processes, but are better described as observed values for a
sequence of reactions in which the rate-determining step follows
the disproportionation.
Results similar to those above were obtained for most of the
other compounds so analysed; kdisp values are presented in
Table 1. Of particular note is the observation that electron-
donating substituents onthe a-diimine ligands increase the values
of kdisp; for example, the disproportionation rate constant of [W
(CO)4(dmb)]þ (dmb ¼ 4,40-dimethyl-2,20-bipyridine), 1jþ, is
greater than that of the bpy analogue, 1bþ, by approximately a
factor of three. We speculate that this is a result of a stabilization

1010
J. P. Bullock et al.
lived enough to characterize by room-temperature infrared
spectroelectrochemistry. Infrared spectra collected during a
bulk electrolysis of 1f in a flow-through cell equipped with a
gold-mesh working electrode are presented in Fig. 4a. The
observed shift to higher frequencies of the carbonyl stretches as
well as the decrease in absorption intensities are consistent with
an oxidation that is primarily tungsten-centred.[6,8,14] The fact
that four discernible stretches are observed in the carbonyl
10 μA
(a)
0.2
0.1
Ϫ1
0
1
2
Potential [V vs Cc0/ϩ
]
region, peaks at 2085, 1990, 1960, and a shoulder at 1938 cmꢀ1
,
0
is notable as this shows that none of the carbonyl ligands are lost
by 1fþ under these conditions. Bulk oxidation of 1g yielded very
similar spectra (Fig. S1, Supplementary Material), showing new
peaks at 2084, 1989, 1956, and 1940 cmꢀ1 (s). This is, to our
knowledge, the first spectroscopic characterization of tungsten
tetracarbonyl radical cations of diimine ligands.
(b)
0.08
0.04
0
Useful comparisons can be made between the infrared
spectra 1fþ and 1gþ and those of related compounds. Although
the spectra of [Cr(CO)4(tmp)]þ and [Mo(CO)4(ncp)]þ are very
similartoeachother,[6,8] theydiffersomewhatfromthetungsten(I)
species in the current work. Most significantly, although the
carbonyl peak frequencies of the parent M(CO)4(LL) com-
pounds for all three Group 6 metals are similar, those for 1fþ
and 1gþ are substantially lower than those for the isoelectronic
chromium and molybdenum species; the symmetric peak is
,10 cmꢀ1 lower, while the asymmetric stretches are lower by an
average of ,40 cmꢀ1. This could be indicative of a stronger
interaction between the metal centre and the PFꢀ6 supporting
anion; such an interaction would be expected to be minimal in
the chromium analogue owing to size constraints, but could be
important for tungsten. The carbonyl stretching frequencies
reported for cis-[W(CO)4(Z2-C2H4)]þ are considerably
higher[14] than those of 1fþ and 1gþ but this can be attributed
to differences between the diimine and olefin ligands; the
observed peak shifts between the neutral parents and radical
cations are actually quite similar, averaging ,100 cmꢀ1 in each
case. Finally, we note the pattern of carbonyl peaks observed for
1fþ and 1gþ is distorted relative to the parent compounds in a
manner similar to that observed for other [M(CO)4(LL)]þ
complexes, including the aforementioned [Cr(CO)4(tmp)]þ
and [Mo(CO)4(ncp)]þ, as well as [Cr(CO)4(dppf)]þ, where dppf
is 1,10-bis(diphenylphosphino)ferrocene;[18] a recent quantum
mechanical analysis of the infrared spectrum of the latter
indicates that a re-ordering of the relative energies of the
vibrational modes takes place on oxidation, as well as a signifi-
cant broadening of one vibrational mode in particular.[19] It is
plausible that similar effects are operative with 1fþ and 1gþ.
The spectra in Fig. 4a show isosbestic behaviour until
,70 % of the electrolysis is complete, when a second, more
highly oxidized species becomes evident. We attribute this to a
tungsten(I) disproportionation process that is slow on the cyclic
voltammetry timescale but becomes significant during the bulk
electrolysis. Evidence for this assignment comes from an
experiment in which the applied potential in the spectroelec-
trochemical cell was turned off after nearly exhaustive oxida-
tion of 1f; this allowed the detection of products generated via
homogeneous processes that would otherwise be oxidized at
the electrode. Peaks attributable to two products grew in
intensity, nearly isosbestically, as those due to the radical
cation decreased (Fig. 4b). One set of product peaks was due
to the parent compound, 1f, while the other had carbonyl peaks
at 2101, 2012, 1997, and 1925 cmꢀ1, higher than those assigned
to 1fþ and consistent with a more highly oxidized metal
tetracarbonyl.
1800
1900
2000
2100
Wavenumber [cmϪ1
]
Fig. 4. (a) Changes in infrared spectra of the carbonyl region during room-
temperature thin-layer bulk electrolysis of W(CO)4(ncp), 1f, at 0.2 V
positive of the E1/2 value observed on the voltammogram (inset); (b) infrared
changes observed on reaction of the electrogenerated radical cation 1fþ after
termination of the electrolysis; total elapsed time for the reaction is ,3 min.
Arrows indicate the direction of change for the peak heights.
of the resulting tungsten(II) disproportionation product by the
increased electron donation from the diimine ligand.
An important exception to the reactivity described above was
observed for [W(CO)4(ncp)]þ, 1fþ, which appears to decay via a
simple first-order process, with k1st-order ¼ 0.07 sꢀ1. This is con-
sistent with the greatly attenuated peak on the cyclic voltammo-
gram ascribable to the corresponding disproportionation product
(Fig. 4, inset). The methyl substituents of the ncp ligand are
adjacent to the nitrogen donor atoms and thereby serve to
sterically block the metal centre. This bestows kinetic stability
on otherwise reactive species by reducing access to the metal
centre by other radical cations or potential nucleophiles.[8,17]
Similarly, the rate of disproportionation of [W(CO)4(bcp)]þ (bcp
¼ 2,9-dimethyl-4,7-diphenylphenanthroline), 1gþ, which is also
sterically crowded, is less than that of its unmethylated analogue
[W(CO)4(bph)]þ (bph¼4,7-diphenylphenanthroline), 1eþ.
Interestingly, infrared spectroelectrochemical experiments
(see below) strongly suggest that 1fþ does, in fact, dispropor-
tionate. We attribute this to a ligand-induced pathway, in which
association with a donor species is the rate-determining step in
the mechanism, giving rise to the observed first-order kinetics;
this is discussed in greater detail in the following section.
Finally, the significantly greater rate of disproportionation of
1gþ compared with 1fþ is curious. They both are sterically
hindered by methyl groups in the 2- and 9-positions of the
phenanthroline ligand, so would be expected to show similar
steric effects. An alternative pathway for electron transfer may
therefore exist that involves the phenyl substituents rather than
direct or bridged interaction between the metal centres.
Infrared Spectroelectrochemistry of W(CO)4(ncp), 1f, and W
(CO)4(bcp), 1g
The kinetic stability afforded by the methyl groups of ncp and
bcp, described above, suggested to us the possibility that the
corresponding 1þ radical cations, 1fþ and 1gþ, may be long-

Electrochemical Oxidation of W(CO)4(LL)
1011
To reconcile the spectral changes presented in Fig. 4b with
the chronocoulometric data described earlier, we propose that a
ligand-induced disproportionation is operative with 1fþ, as
explained below. First, 1fþ undergoes a slow associative reac-
tion with a donor species D, such as PFꢀ6 , thereby converting the
17-electron electrogenerated species to a 19-electron species
(Eqn 3) capable of serving as a reducing agent. The subsequent
electron transfer (Eqn 4) then yields two 18-electron species. We
have previously observed similar behaviour by other electro-
generated species, including [Mo(CO)3(LL)(CH3CN)]þ and
[[CpFe(CO)2]2]þ.[8,20]
Large-Amplitude A.C. Cyclic Voltammetric Studies
Large-amplitude a.c. voltammetry has been shown to be an
exceptionally useful technique in mechanistic studies involving
electrogenerated species.[21] Given its potential to provide
additional insight into the reactivity of 1þ, as well as to further
characterize the heterogeneous electron transfer associated with
the oxidation of the parent compounds 1, we performed a series
of such experiments on 1a. The third through sixth harmonics
obtained for 0.3 mM 1a are shown in Fig. 5 (black traces, left and
centre columns). These data are consistent with a fast electron
transfer; simulations suggest that the standard rate constant for
the electron transfer has a lower limit of ,0.2 cm sꢀ1, similar to
that we observed for ferrocene under similar conditions.[22]
An analysis of the a.c. voltammetric data was particularly
illuminating in light of some of our other observations. Specifi-
cally, we initially attempted to fit the data in Fig. 5 with the
scheme suggested by the chronocoulometric data, that is, to an
½WðCOÞ4ðncpÞꢂþ þ D Ð ½WðCOÞ4ðncpÞðDÞꢂþ
½WðCOÞ4ðncpÞꢂþ þ ½WðCOÞ4ðncpÞðDÞꢂþ
ð3Þ
ð4Þ
2þ
! WðCOÞ4ðncpÞ þ ½WðCOÞ4ðncpÞðDÞꢂ
0.3 μA
0.5 μA
1.0 μA
3.0 μA
5
10
15
20
5
10
15
20
5
10
15
20
Time [s]
Fig. 5. Third (bottom row) through sixth (top row) harmonics of large-amplitude a.c. voltammetry results of 0.28 mM 1a; (left) data
obtained in the presence of no added acetonitrile (black traces) and 3.2 mM acetonitrile (blue traces); (middle) black traces are the same as
those shown in the left column, and red traces show simulated fits; (right) blue traces are the same as those shown in the left column, and
red traces show simulated fits. The switching time in each experiment was 13.4 s; peaks observed at earlier times correspond to those
occurring during the anodic sweep, those at longer time are from the cathodic sweep. Experimental parameters are: scan rate ¼ 0.0894
V sꢀ1, electrode area ¼ 0.071 cm2, a.c. amplitude, DE ¼ 0.100 V, frequency ¼ 13.49 Hz. Simulation parameters that were kept constant
in the two simulations were the diffusion coefficient ¼ 1.2 ꢄ 10ꢀ5 cm2 sꢀ1, k8 ¼ 0.30 cm sꢀ1, and the homogeneous rate constants (see
Scheme 1) k1 ¼ 2 ꢄ 104 Mꢀ1 sꢀ1, kꢀ1 ¼ 1 ꢄ 105 Mꢀ1 sꢀ1, k2 ¼ 0.10 sꢀ1, k3 ¼ 0.20 sꢀ1, k4 ¼ 500 Mꢀ1 sꢀ1, kꢀ4 ¼ 50 sꢀ1, k5 ¼ 0.05 sꢀ1 and
k6 ¼ 8 ꢄ 103 Mꢀ1 sꢀ1; uncompensated resistance varied slightly in the two simulations (250 and 220 O in the absence and presence of
acetonitrile respectively).

1012
J. P. Bullock et al.
irreversible disproportionation taking place alongside a parallel
first-order decay. Interestingly, this model did not yield accept-
able agreement with experimental data in that the ratio of peak
heights within the envelopes for the forward and reverse sweeps
could not be satisfactorily emulated (Fig. S2, Supplementary
Material). Good agreement was achieved, however, by employing
a scheme in which the disproportionation was reversible (as
written in Eqn 1); in previous work, we demonstrated that the
response of a.c. voltammetry is highly sensitive to the existence
of equilibria involving electrogenerated species.[23] The fits
shown in Fig. 5 (red traces, middle column) were generated
using a mechanistic scheme that included, in addition to the
reversible disproportionation, an irreversible first-order (or
pseudo-first-order) follow-up reaction involving the tungsten(II)
disproportionation product that gives rise to coupled reduction
peak in the d.c. voltammograms, e.g. that seen at 0.32 V in
Fig. 2a, and a first-order decay of 1aþ; the rate constants for
these steps are designated k2 and k3 respectively. The simula-
tions in shown in Fig. 5 were obtained using forward and reverse
rate constants for the disproportionation reaction, k1 and kꢀ1, of
2 ꢄ 104 and 1 ꢄ 105 Mꢀ1 sꢀ1 respectively; these rate constants
translate to an equilibrium constant, K1, of 0.2; the thermody-
namic driving force for the disproportionation, therefore,
appears to be modest; the other rate constants are given in the
figure caption.
no added ligand
20 μA
with 1.1 equiv. pyridine
with 1.6 equiv. CH3CN
0.5
1.0
1.5
2.0
Potential [V vs Cc0/ϩ
]
Fig. 6. Cyclic voltammograms of: (top) 1.4 mM 1b in CH2Cl2/TBAPF6
with no added ligand; (middle) 1.4 mM 1b in CH2Cl2/TBAPF6 with1.1
equivalents of pyridine (scan initiated at 1.3 V); and (bottom) 1.2 mM 1b in
CH2Cl2/TBAPF6 with1.6 equivalents of acetonitrile; scan rates in each case
Reactivity of [W(CO)4(LL)]1 Towards Added Ligands
are 0.25 V sꢀ1
.
As exemplified by the voltammograms in Fig. 1, the homoge-
neous coupled chemistry followed by 1þ in coordinating media
differs substantially from that of the molybdenum analogues
under similar conditions. It is likely that these divergent beha-
viours have a kinetic basis, that is, whereas the essential reaction
schemes followed by these species are probably similar, the
relative rates of the competing reactions are not, manifesting in
their distinctive electrochemical responses. To get further
insight into the competing reactions followed by radical cations,
we titrated 1a and 1b with pyridine and acetonitrile and moni-
tored the chemistry of the corresponding 1þ species with d.c.
cyclic voltammetry (Figs S3 and S4 respectively, Supplemen-
tary Material).
Voltammograms of 1b obtained in the presence of pyridine
(1.1 equivalents) and acetonitrile (1.6 equivalents), a less
strongly coordinating ligand, are presented in Fig. 6. These
results show that the rate of attack by the added ligands on 1þ, as
well as the yield of the corresponding substitution product, [W
(CO)3(LL)(L0)]þ, where L0 is the substituting ligand, both
increase with ligand nucleophilicity. With respect to the first
point, the reversibility of the bulk oxidation process for 1b is
essentially lost in the presence of near-stoichiometric amounts
of pyridine, whereas a detectable return peak due to the reduc-
tion of 1bþ is discernible in the presence of more than 3
equivalents of acetonitrile.
other 17-electron species[24] and is supported by double poten-
tial-step chronocoulometry studies of 1a and 1b (data not
shown) that indicate the reaction between 1þ and added ligands
is first-order in both species.
k4
½WðCOÞ4ðLLÞꢂþ þ L0 Ð ½WðCOÞ4ðLLÞðL0Þꢂþ
k
ꢀ4
ð5Þ
k5
ꢀ! ½WðCOÞ3ðLLÞðL0Þꢂþ þ CO
As a function of pyridine concentration, the extent of
substitution reaches a limiting value that corresponds to approx-
imately 0.7 equivalents of the tricarbonyl substitution product
per equivalent of 1bþ generated at the electrode; this ratio was
estimated by comparing the measured peak current with that
obtained via simulation of a reaction scheme in which 100 % of
1bþ is converted to the substitution product. This limiting yield
is attained after less than 10 equivalents of pyridine are added.
According to the reaction sequence proposed in Scheme 1, the
extent of substitution observed depends on the relative magni-
tudes of two rate constants: those of the decarbonylation of the
pyridine adduct, k5, and a ligand-induced disproportionation, k6,
which we propose as the major competing pathway. The latter is
analogous to that described previously for 1fþ; specifically, the
19-electron intermediate shown in Eqn 5 can serve as a reducing
agent towards 1þ (Eqn 6).
A reaction sequence that accounts for these observations is
presented in Scheme 1 and described briefly below; the scheme
also includes the previously discussed reactivity of electrogen-
erated 1þ. The ligand substitution likely proceeds via an
associative mechanism because the carbonyl is not labilized to
a great extent on oxidation of 1 in the absence of added ligands.
Presumably, attack by the added ligand results in a 19-electron,
7-coordinate species, followed by loss of one CO ligand (Eqn 5;
rate constants are labelled to be consistent with Scheme 1). The
associative nature of the reaction has many precedents with
½WðCOÞ4ðLLÞðL0Þꢂþ þ ½WðCOÞ4ðLLÞꢂþ
ð6Þ
k6
ꢀ! ½WðCOÞ4ðLLÞðL0Þꢂ2þ þ WðCOÞ4ðLLÞ
The fraction of substitution products generated will be
affected by the nature of the added ligand in two ways. First,

Electrochemical Oxidation of W(CO)4(LL)
1013
W(CO)4(LL)
[W(CO)4(LL)]ϩ
W(CO)4(LL)
ϪeϪ
k1
kϪ1
ϪCO
ϩD
[W(CO)3(LL)(D)]ϩ
[W(CO)4(LL)]ϩ
[W(CO)4(LL)]2ϩ
ϪeϪ
ϩD
k3
k2
decomposition
ϩLЈ
[W(CO)4(LL)D)]2ϩ
kϪ4
k4
[W(CO)4(LL)]ϩ
W(CO)4(LL)
[W(CO)4(LL)(LЈ)]2ϩ
ϪCO
k5
[W(CO)4(LL)(LЈ)]ϩ
[W(CO)3(LL)(LЈ)]ϩ
ϩeϪ
k6
W(CO)3(LL)(LЈ)
Scheme 1. Summary of observed chemistry followed by electrogenerated 1þ species; rate constants are discussed
in the text; L0 is an intentionally added ligand (pyridine or acetonitrile in this work) and D is a coordinating species
in the solvent electrolyte such as PFꢀ6 or adventitious water. Values for rate constants that gave satisfactory
simulations of d.c. voltammetric data over the entire range of pyridine concentrations employed in the titration
experiments were: k4 ¼ 5 ꢄ 103 Mꢀ1 sꢀ1, kꢀ4 ¼ 0.1 sꢀ1, k5 ¼ 0.35 sꢀ1, k6 ¼ 3 ꢄ 103 Mꢀ1 sꢀ1
.
the rate of CO loss by the 19-electron ligand adduct will likely be
influenced by structural and electronic factors of the associating
ligand; our results imply that the loss of carbon monoxide is
faster with the pyridine adduct, perhaps because, as a stronger
donor ligand, it more effectively stabilizes the electron-deficient
transition state. Second, the extent of the equilibrium, governed
by K4, will also play a role: if formation of the adduct is strongly
favoured, then the availability of free 1þ will be limited,
suppressing the disproportionation pathway. We estimated the
value of this equilibrium constant to be ,5 ꢄ 104 in the case of
pyridine using rate constants k4 and kꢀ4 we obtained by fitting
the observed voltammograms with simulations generated for the
reactions described by Scheme 1 (Fig. S3, Supplementary
Material); other rate constants employed in the simulations are
specified in the caption for Scheme 1.
The reaction pathways outlined in Scheme 1 are supported by
several additional observations. For example, there is significant
current enhancement of the reduction peak assigned to the
tungsten(II) disproportionation product during the pyridine titra-
tion of 1a and 1b; this supports our contention that the rate of
disproportionation increases in the presence of coordinating
ligands. In addition, Scheme 1 predicts current enhancement
of the bulk oxidation peak, due to regeneration of the parent
compound 1, via the disproportionation step; we observed this
effect and found in the simulations that the magnitude of the
current enhancement is a sensitive function of k6; on this basis,
inherent instability of this species as related experiments in our
laboratory show that the oxidation of W(CO)3(bpy)(CH3CN) is
chemically reversible under similar conditions, albeit accompa-
nied by a disproportionation process analogous to that exhibited
by 1.
There are several plausible explanations for the relatively
low conversion rate to the substitution product. Although the
rate of decarbonylation of the 19-electron adduct could be a
factor, others must also be at play because simulations that
employ values of k5 that are low enough to fit the voltammo-
grams observed at low concentrations of acetonitrile grossly
underestimate the concentrations of the substitution product at
higher levels of added ligand. In fact, no combinations of k4, kꢀ4
,
k5, and k6 provide adequate fits of our experimental voltammo-
grams over the acetonitrile concentration range examined. We
therefore conclude that additional reaction pathways are oper-
ating. Several possibilities exist, some of which we have direct
evidence for. Specifically, as mentioned above, the oxidation of
W(CO)3(bpy)(CH3CN) is chemically reversible, but when per-
formed in the presence of 1b, the reversibility is completely
compromised when the anodic scans extend to positive enough
potentials to generate 1bþ. We propose the disproportionation
reaction described by Eqn 7 to account for this observation; here,
an electron transfer occurs between two 17-electron species,
regenerating 1b and forming a tungsten(II) tricarbonyl com-
pound. Thus, in the titration experiment, any [W(CO)3(bpy)
(CH3CN)]þ produced via substitution will be readily consumed
by reaction with 1bþ. In addition, [W(CO)3(bpy)(CH3CN)]þ
could also react with the 19-electron acetonitrile adduct of 1bþ
in a manner analogous to Eqn 6. Owing to the increasingly
speculative nature of these reactions, we did not attempt to
estimate the various rate constants via simulation. We note that
we estimated that k6 has an upper limit of ,1 ꢄ 104 Mꢀ1 sꢀ1
.
Titration data obtained with acetonitrile were more complex
than with pyridine. The voltammogram in Fig. 6 is intriguing in
that 1.6 equivalents of the weaker ligand is enough to substan-
tially reduce the reversibility of the bulk oxidation, yet virtually
no substitution product is observed. This cannot be due to any

1014
J. P. Bullock et al.
the reactions proposed that consume [W(CO)3(bpy)(CH3CN)]þ
should also be viable in the case of pyridine. As mentioned
previously, however, if K4 is large, then the concentration of free
1bþ will be low enough to suppress the reactions shown in Eqns
6 and 7, resulting in the sort of saturation kinetics we observed in
the pyridine titration.
substitution product [W(CO)3(LL)(L0)]þ. The latter pathway is
favoured by stronger donors such as pyridine, and appears to be
the path followed by molybdenum even in the presence of weak
bases such as acetonitrile. The net conversion of 1þ to the
corresponding tricarbonyl substitution products is diminished if
the added ligand is a poor donor and is attributed to subsequent
electron-transfer reactions that consume the substitution pro-
ducts, although slower decarbonylation from [W(CO)4(LL)
(L0)]þ may also contribute to this effect. Current work in our
laboratory is focussed on investigating additional aspects of the
coupled chemistry described above, especially at low levels of
added ligand. Specifically, we aim to more definitively eluci-
date the reaction pathways followed by the 19-electron adduct
species of weakly donating ligands.
½WðCOÞ4ðLLÞꢂþ þ ½WðCOÞ3ðLLÞðL0Þꢂþ
ð7Þ
2þ
! WðCOÞ4ðLLÞ þ ½WðCOÞ3ðLLÞðL0Þꢂ
We further investigated the nature of the equilibrium
between 1bþ and acetonitrile using large-amplitude a.c. vol-
tammetry (Fig. 5, blue traces). A.c. voltammograms of 1b in the
presence and absence of added acetonitrile are shown in the left
column of Fig. 5; most notable about these results is the decrease
in peak currents in the return sweep. Voltammograms obtained
in the presence of acetonitrile were satisfactorily fitted by
simulations (Fig. 5, right column) of a mechanism involving
an equilibrium between 1aþ and the corresponding acetonitrile
adduct using a forward rate constant of 5 ꢄ 102 Mꢀ1 sꢀ1; this is a
factor of 10 lower than we estimate for the analogous reaction
with pyridine. In addition, the simulations indicate K4, the
equilibrium constant for the reaction between 1bþ and acetoni-
trile, to be ,10, less than that observed with the more strongly
coordinating pyridine by more than three orders of magnitude.
These results are consistent with the more pronounced diminu-
tion of the coupled reduction of 1bþ as well as the greater
prevalence of the substitution pathway observed with pyridine.
Finally, we should note that the coupled chemistry associated
with oxidation of 1 in the presence of added ligands is almost
certainly more complicated than Scheme 1 indicates. The added
complexity of the chemistry exhibited by 1bþ in the presence of
acetonitrile is only one aspect of this point. In addition, Hanzlik
et al. observed ‘curve-crossing’ in the voltammograms of
Mo(CO)4(bpy) in acetonitrile that they attributed to oxidation
of the parent compound by molybdenum(II) species generated
via disproportionation.[7] We observed similar behaviour in
these studies under some conditions, indicating that the tung-
sten(II) products we invoke in the above discussions are not
innocent and will likely introduce additional reactions not
included in Scheme 1.
Experimental
Materials and Methods
Tungsten hexacarbonyl, W(CO)6 (Strem), pyridine (Aldrich),
and all ligands used in this study were used as received (GFS
Chemicals, Aldrich); these included 1,10-phenanthroline, a
(phen), 2,20-byridine, b (bpy), the substituted phenanthroline
ligands 4,7-dimethylphenanthroline, c (dmp), 3,4,7,8-tetra-
methylphenanthroline, d (tmp), 4,7-diphenylphenanthroline, e
(bph), 2,9-dimethyphenanthroline, f (ncp), 2,9-dimethyl-4,7-
diphenylphenanthroline, g (bcp), 5-nitrophenanthroline, h
(nop), 5-chlorophenanthroline, i (clp), and the substituted
bipyridyl ligand, 4,40-dimethyl-2,20-bipyridine, j (dmb). Sol-
vents used for electrochemical experiments, dichloromethane
and acetonitrile (Aldrich), were HPLC grade. TBAPF6 (South-
western Analytical) was used as received. All solvent and
˚
electrolyte solutions were dried over activated 4-A molecular
sieves (Aldrich).
Syntheses
1a–j were prepared photolytically using a procedure adapted
from the literature;[25] the synthesis of 1j is presented below and
typifies all such syntheses performed in this work. W(CO)6
(0.550 g, 1.56 mmol) was dissolved in 50 mL dry tetrahydrofu-
ran (Pharmco) and degassed in a round-bottom flask. The
resulting solution was illuminated with a 100-W mercury arc
lamp until nearly complete conversion to W(CO)5(THF) had
been effected[26] as determined by infrared spectroscopy; this
normally required 3 to 4 h of constant illumination. To the
resulting yellowish-green solution was added a slight molar
excess of dmp (0.347 g, 1.64 mmol) under constant stirring and
nitrogen purging; this resulted in a rapid colour change and
eventual precipitation of the desired product. After 30 min, the
precipitate was isolated and washed with additional THF; the
crude product was purified by recrystallization from dichlor-
omethane/hexanes. The product (0.516 g, 1.02 mmol, 65 %
yield) consisted of brick-red crystals and its identity was con-
firmed by IR spectroscopy. Most of the compounds synthesized
exhibited deep-red luminescence as solids.
Conclusion
Despite the substantially different cyclic voltammograms
exhibited by 1a and its molybdenum analogue in CH3CN/
TBAF6, their respective electrogenerated cations appear to
share the same reaction pathways, albeit with significant dif-
ferences in their relative reaction rates. The general reaction
pathways are illustrated in Scheme 1 and briefly summarized
below. The radical cations 1þ are generated via a fast electrode
process; a.c. voltammetry studies indicate that the radicals exist
in equilibrium with disproportionation products 1 and 12þ. In
weakly coordinating media, 12þ associates with donor species
D, either an anion from the supporting electrolyte or adventi-
tious water, to yield [W(CO)4(LL)(D)]2þ, which is readily seen
on cyclic voltammograms. 1þ can also lose a carbon monoxide
ligand, giving relatively short-lived species [W(CO)3(LL)
(D)]þ, observable via cyclic voltammetry using scan rates of
,1 V sꢀ1 or greater.
Instrumentation
D.c. cyclic voltammetry and double potential-step chron-
ocoulometry experiments were performed using a Bioanalytical
Systems 50B electrochemical workstation using a 3-mm diam-
eter glassy carbon working electrode and a Ag/Agþ pseudo-
reference electrode. Oxidation potentials, E8, were measured
using cobaltocenium hexafluorophosphate, Ccþ (Aldrich), as an
internal standard. Analyte solutions were degassed and kept
In the presence of added ligands, 1þ expands its coordination
sphere, yielding the 19-electron species [W(CO)4(LL)(L0)]þ;
this can reduce 1þ or lose a carbonyl ligand to generate the

Electrochemical Oxidation of W(CO)4(LL)
1015
under nitrogen for all analyses. To facilitate comparisons with
other reports, the Ccþ/0 redox couple was observed at ꢀ1.335 V
versus the ferrocene/ferrocenium couple in CH2Cl2/TBAPF6.
Infrared spectroelectrochemical experiments were performed
using a celldescribed previously[27] and a Thermoelectron Avatar
350 Fourier-transform (FT)IR spectrometer; the Bioanalytical
Systems 50B workstation was used to control and monitor the
electrolyses.
Large-amplitude a.c. voltammetry was performed using a
custom-built instrument described elsewhere.[21] The following
parameters were estimated heuristically[28] using MECSIM
software:[29] uncompensated resistance, R, heterogeneous rate
constant, k8, and relevant homogeneous rate constants; the
transfer coefficient, a, was assumed to be 0.5 in all simulations.
The diffusion coefficient for 1a was calculated from chrono-
coulometric data using the Cottrell equation.[30] Analyses of the
data focussed on the third through sixth harmonics, as these
show little contribution from non-Faradaic current.
References
[1] A. Vlcˇek, Jr, Coord. Chem. Rev. 2002, 230, 225. doi:10.1016/S0010-
8545(02)00047-4
[2] (a) M. L. Clark, K. A. Grice, C. E. Moore, C. P. Kubiak, Chem. Sci.
2014, 5, 1894. doi:10.1039/C3SC53470G
(b) J. Tory, B. Setterfield-Price, R. A. W. Dryfe, F. Hartl, Chem-
ElectroChem 2015, 2, 213. doi:10.1002/CELC.201402282
[3] A. A. Vlcˇek, Chemtracts – Inorg. Chem. 1993, 5, 1.
[4] S. Ernst, W. Kaim, J. Am. Chem. Soc. 1986, 108, 3578. doi:10.1021/
JA00273A005
[5] E. Ioachim, G. S. Hanan, Can. J. Chem. 2005, 83, 1114. doi:10.1139/
V05-127
[6] I. R. Farrell, F. Hartl, S. Za´lisˇ, M. Wanner, W. Kaim, A. Vlcˇek, Jr,
Inorg. Chim. Acta 2001, 318, 143. doi:10.1016/S0020-1693(01)
00421-2
[7] J. Hanzl´ık, L. Posp´ısˇil, A. A. Vlcˇek, M. Krejcˇ´ık, J. Electroanal. Chem.
1992, 331, 831. doi:10.1016/0022-0728(92)85009-R
[8] R. Johnson, H. Madhani, J. P. Bullock, Inorg. Chim. Acta 2007, 360,
3414. doi:10.1016/J.ICA.2007.04.029
[9] H. tom Dieck, E. Ku¨hl, Z. Naturforsch. B: Anorg. Chem., Org. Chem.
1982, 37B, 324.
[10] M. G. Hill, W. M. Lamanna, K. R. Mann, Inorg. Chem. 1991, 30, 4687.
doi:10.1021/IC00025A003
[11] R. Colton, Coord. Chem. Rev. 1971, 6, 269. doi:10.1016/S0010-8545
(00)80042-9
[12] A. M. Bond, R. Colton, K. McGregor, Organometallics 1990, 9, 1227.
doi:10.1021/OM00118A051
[13] A. M. Bond, R. Colton, Coord. Chem. Rev. 1997, 166, 161.
doi:10.1016/S0010-8545(97)00022-2
Kinetic Analysis
Determination of second-order disproportionation rate constants
using chronocoulometric datawascomplicated bythe presence of
a simultaneous first-order reaction as described in the text. To
calculate the net rate of disproportionation, experimental data
obtained at each analyte concentration were initially fitted to a
theoretical working curve of a disproportionation mechanism;[16]
using the corresponding best-fit rate constants, keff, the overall
reaction rate was estimated using a second-order rate law. This
overall rate of conversion of 1þ was set equal to the sum of two
separate process with rate constants kdisp and k1st-order (Eqn 8).
Linearization afforded values of k1st-order from the intercept and
kdisp from the slope of the resulting plot (Eqn 9).
[14] (a) J. Handzlik, F. Hartl, T. Szyman´ska-Buzar, New J. Chem. 2002, 26,
145. doi:10.1039/B104786H
(b) M. Go´rski, F. Hartl, T. Szyman´ska-Buzar, Organometallics 2007,
26, 4066. doi:10.1021/OM700384G
[15] M. C. Baird, Chem. Rev. 1988, 88, 1217. doi:10.1021/CR00089A011
[16] M. K. Hanafey, R. L. Scott, T. H. Ridgway, C. N. Reilley, Anal. Chem.
1978, 50, 116. doi:10.1021/AC50023A030
[17] J. P. Bullock, E. Carter, R. Johnson, A. T. Kennedy, S. E. Key, B. J.
Kraft, D. Saxon, P. Underwood, Inorg. Chem. 2008, 47, 7880.
doi:10.1021/IC800530N
Â
Ã
Â
Ã
2
2
rate ꢅ keff ½WðCOÞ4ðLLÞꢂþ ꢅ kdisp ½WðCOÞ4ðLLÞꢂþ
þ k1st-order½WðCOÞ ðLLÞꢂþ
4
[18] A. C. Ohs, A. L. Rheingold, M. J. Shaw, C. Nataro, Organometallics
2004, 23, 4655. doi:10.1021/OM049735T
ð8Þ
[19] P. A. Eckert, K. J. Kubarych, J. Phys. Chem. A 2017, 121, 2896.
doi:10.1021/ACS.JPCA.6B12898
[20] J. P. Bullock, M. C. Palazotto, K. R. Mann, Inorg. Chem. 1991, 30,
Â
Ã
Â
Ã
keff WðCOÞ4ðLLÞ þꢅ kdisp ½WðCOÞ4ðLLÞꢂþ þ k1st-order ð9Þ
1284. doi:10.1021/IC00006A024
[21] A. M. Bond, N. W. Duffy, S.-X. Guo, J. Zhang, D. Elton, Anal. Chem.
2005, 77, 186A. doi:10.1021/AC053370K
[22] J. P. Bullock, E. Mashkina, A. M. Bond, J. Phys. Chem. A 2011, 115,
Conflicts of Interest
The authors declare no conflicts of interest.
6493. doi:10.1021/JP2021787
[23] C.-Y. Lee, J. P. Bullock, G. F. Kennedy, A. M. Bond, J. Phys. Chem. A
2010, 114, 10122. doi:10.1021/JP105626Z
[24] D. R. Tyler, Prog. Inorg. Chem. 1988, 36, 125. doi:10.1002/
9780470166376.CH2
[25] W. Kaim, S. Kohlmann, Inorg. Chem. 1987, 26, 68. doi:10.1021/
IC00248A016
[26] W. Strohmeier, Angew. Chem. Int. Ed. Engl. 1964, 3, 730. doi:10.1002/
ANIE.196407301
[27] J. P. Bullock, D. C. Boyd, K. R. Mann, Inorg. Chem. 1987, 26, 3084.
doi:10.1021/IC00266A004
[28] A. N. Simonov, G. P. Morris, E. A. Mashkina, B. Bethwaite, K. Gillow,
R. E. Baker, D. J. Gavaghan, A. M. Bond, Anal. Chem. 2014, 86, 8408.
doi:10.1021/AC5019952
[29] G. F. Kennedy, A. M. Bond, A. N. Simonov, Curr. Opin. in Electro-
chem 2017, 1, 140. doi:10.1016/J.COELEC.2016.12.001
[30] See Ch. 5 in A. J. Bard, L. R. Faulkner, Electrochemical Methods:
Fundamentals and Applications (2nd Edn) 2001 (Wiley: New York,
NY).
Supplementary Material
Infrared spectra obtained during spectroelectrochemical oxi-
dation of 1g, optimized a.c. voltammograms (3rd through 6th
harmonics) of the oxidation of 1a using an irreversible dis-
proportionation alongside an irreversible ligand substitution,
and d.c. voltammograms obtained during the titration of 1b
with pyridine and acetonitrile are available on the Journal’s
website.
Acknowledgements
C.Y.L. is grateful to the University of Wollongong (UOW) for the UOW
Vice Chancellor Research Fellowship. J.U. gratefully acknowledges finan-
cial support from the Vermont Science Initiative that made possible his
participation in this work.