Electron-transfer properties of Cp*FeP5: Evidence for dimerization reactions following both oxidation and reduction

Article abstract of DOI:10.1021/om9809724

The redox reactions of Cp*Fe(η⁵-P₅) (1; Cp= η⁵-C₅Me₅) have been characterized in nonaqueous solvents by electrochemical methods. As anticipated by analogy with ferrocene, 1 may be both oxidized and reduced in one-electron processes. Both processes are irreversible by cyclic voltammetry but reversible by bulk electrolysis. In CH₂Cl₂ complex 1 oxidizes initially to 17-electron 1+ (E_p,a = 0.57 V vs Fc), which rapidly equilibrates to give the dimeric dication [1₂]²⁺. An ESR spectrum attributed to 1⁺ is consistent with a d⁵ iron sandwich complex. A dimerization rate constant for 1⁺ of k_D(17) = 1.4 × 10⁴ M^-1 s^-1 was determined from cyclic voltammetry (CV) data. The dimeric dication quantitatively re-forms neutral 1 upon rereduction. Complex 1 undergoes reduction (E_1/2 = -2.00 V) to 19-electron 1^-, which also appears to dimerize in THF; k_D(19) = ca. 6 × 10⁵ M^-1 s^-1. Reoxidation of the diamagnetic dimer [1₂]^2- regenerates 1. The shifts in potential induced when replacing a cyclopentadienyl ring by a pentaphosphacyclopentadienyl ring, explicable in terms of the weaker electron-donating ability of the latter, are greater for the reductions than the oxidations, implying an increased P₅ character to the LUMO of 1 compared to the HOMO. Possible structures of the dimeric ions are discussed in terms of known structural analogues and previously published molecular orbital descriptions.

Full text of DOI:10.1021/om9809724

Organometallics 1999, 18, 1827-1833
1827
Articles
Electr on -Tr a n sfer P r op er ties of Cp *F eP 5: Evid en ce for
Dim er iza tion Rea ction s follow in g both Oxid a tion a n d
Red u ction
Rainer F. Winter† and William E. Geiger*
Department of Chemistry, University of Vermont, Burlington, Vermont 05405
Received December 2, 1998
The redox reactions of Cp*Fe(η5-P5) (1; Cp* ) η5-C5Me5) have been characterized in
nonaqueous solvents by electrochemical methods. As anticipated by analogy with ferrocene,
1 may be both oxidized and reduced in one-electron processes. Both processes are irreversible
by cyclic voltammetry but reversible by bulk electrolysis. In CH2Cl2 complex 1 oxidizes
initially to 17-electron 1+ (Ep,a ) 0.57 V vs Fc), which rapidly equilibrates to give the dimeric
dication [12]2+. An ESR spectrum attributed to 1+ is consistent with a d5 iron sandwich
complex. A dimerization rate constant for 1+ of kD(17) ) 1.4 × 104 M-1 s-1 was determined
from cyclic voltammetry (CV) data. The dimeric dication quantitatively re-forms neutral 1
upon rereduction. Complex 1 undergoes reduction (E1/2 ) -2.00 V) to 19-electron 1-, which
also appears to dimerize in THF; kD(19) ) ca. 6 × 105 M-1 s-1. Reoxidation of the diamagnetic
dimer [12]2- regenerates 1. The shifts in potential induced when replacing a cyclopentadienyl
ring by a pentaphosphacyclopentadienyl ring, explicable in terms of the weaker electron-
donating ability of the latter, are greater for the reductions than the oxidations, implying
an increased P5 character to the LUMO of 1 compared to the HOMO. Possible structures of
the dimeric ions are discussed in terms of known structural analogues and previously
published molecular orbital descriptions.
In tr od u ction
enyl analogues have appeared,7,8 no information has
been published concerning the ionization potentials and/
or the redox potentials of 1 or other mononuclear
complexes containing the pentaphosphacyclopentadi-
enyl ligand9a (the electron-transfer properties of the
triple-decker complexes Cp*2M2(P5) (M ) Cr, Mo, W, V)
were characterized by Kaim and co-workers).9b
The redox properties of Cp*Fe(η5-P5) (1; Cp* )
C5Me5)1-4 are of interest because of the close analogy
between this compound and ferrocene.5 Molecular or-
The present study shows that complex 1 undergoes
one oxidation and one reduction, each involving one
electron, within the usual potential window in nonaque-
ous solutions. In terms simply of the number of one-
electron processes, its behavior mirrors that of ferrocene
and a number of sandwich complexes containing one or
two monophosphacyclopentadienyl complexes (vide in-
fra). The chemical fates of the one-electron products are
strikingly different, however. Both the 17-electron
cation 1+ and the 19-electron anion 1- are short-lived
and undergo reactions that are second-order in the iron
complex. The apparently dimeric products can each be
re-electrolyzed to give back the starting material 1. It
bital calculations indicate that the HOMO of 1 is ag, in
contrast to the e2g HOMO of ferrocene.6 Although
electrochemical studies of monophosphacyclopentadi-
† Present address: Institut fur Anorganische Chemie, University
of Stuttgart, Pfaffenwaldring 55, D-70569 Stuttgart, Germany.
(1) Scherer, O. J .; Brueck, T. Angew. Chem. 1987, 99, 59; Angew.
Chem., Int. Ed. Engl. 1987, 26, 59.
(2) Scherer, O. J . Angew. Chem. 1990, 102, 1137; Angew. Chem.,
Int. Ed. Engl. 1990, 29, 1104.
(3) Detzel, M.; Mohr, T.; Scherer, O. J .; Wolmershaeuser, G. Angew.
Chem. 1994, 106, 1142; Angew. Chem., Int. Ed. Engl. 1994, 33, 1110.
(4) Rink, B.; Scherer, O. J .; Wolmershaeuser, G. Chem. Ber. 1995,
128, 71.
(5) The coordination chemistry of phospholyl ligands is discussed
in: Dillon, K. B.; Mathey, F.; Nixon, J . F. Phosphorus: The Carbon
Copy; Wiley: Chichester, U.K., 1998; Chapter 9.
(6) (a) Kerins, M. C.; Fitzpatrick, N. J .; Nguyen, M. T. Polyhedron
1989, 8, 1135 (b) Chamizo, J . A.; Ruiz-Mazon, M.; Salcedo, R.; Toscano,
R. A. Inorg. Chem. 1990, 29, 879.
(7) Lemoine, P.; Gross, M.; Braunstein, P.; Mathey, F.; Deschamps,
B.; Nelson, J . H. Organometallics 1984, 3, 1303.
(8) Ashe, A. J ., III; Al-Ahmad, S.; Pilotek, S.; Puranik, D. B.;
Elschenbroich, C.; Behrendt, A. Organometallics 1995, 14, 2689.
(9) (a) CV data on 1 was obtained elsewhere: Elvers (Elvers, A.
Ph.D. Dissertation, Erlangen, Germany, 1998) reports Ep,a ) 1.19 V
for oxidation and Ep,c ) -1.67 V for reduction in CH2Cl2 (vs SCE). (b)
Scherer, O. J .; Schwalb, J .; Swarowsky, H.; Wolmershaeuser, G.; Kaim,
W.; Gross, R. Chem. Ber. 1988, 121, 443.
10.1021/om9809724 CCC: $18.00 © 1999 American Chemical Society
Publication on Web 04/21/1999

1828 Organometallics, Vol. 18, No. 10, 1999
Winter and Geiger
is significant that throughout both chemically reversible
redox sequences, namely the oxidation of 1 followed by
rereduction of the product and the reduction of 1
followed by reoxidation of the product, the cyclic P5 ring
remains coordinated to the metal.
Exp er im en ta l Section
All operations were conducted under an atmosphere of
dinitrogen using standard Schlenk and drybox procedures.
Solvents were distilled from drying agents before use, and
electrochemical experiments were conducted with solvents that
were vacuum-distilled from drying agents into flasks that were
transferred into the drybox. Complex 1 was obtained from Dr.
T. Mohr and Prof. O. J . Scherer (University of Kaiserslautern)
and used as received. Tris(p-bromophenyl)ammonium hexaflu-
orophosphate was prepared according to the literature.10
Reagent-grade dichloromethane was twice distilled from
CaH2, and THF was subjected to successive distillations from
CaH2, K, and (deep purple) K/benzophenone. 1,2-Difluoroben-
zene (Strem) was stirred over and vacuum-distilled from
activated alumina. The supporting electrolyte was generally
0.1 M [NBu4][PF6], recrystallized from 95% EtOH and dried
at 373 K under vacuum for at least 2 days. For subambient-
temperature work in THF, [NBu4][CF3SO3] was employed as
supporting electrolyte for solubility reasons. Prepared by the
procedure of Dolphin et al.,11 the salt was successively
recrystallized from Et2O and a 3/1 CH2Cl2/Et2O mixture to
yield a waxlike white solid which was finely crushed and dried
at 351 K under vacuum. Although a silver/silver chloride
electrode was used as the experimental reference electrode,
all potentials in this paper are referred to the ferrocene/
F igu r e 1. Cyclic voltammogram of 0.5 mM 1 in CH2Cl2/
0.1 M [NBu4][PF6] at 260 K (v ) 0.5 V/s, Pt electrode).
ESR spectra were measured with a modified Varian E-4
spectrometer using dpph as a standard. Samples were trans-
ferred from electrolyzed solutions into ESR tubes, capped with
a rubber septum, removed from the drybox, and immediately
frozen at 77 K until ESR analysis.
Resu lts a n d Discu ssion
It will be shown below that both the anodic and
cathodic reactions of 1 proceed through ECEC mecha-
nisms; that is, fast chemical reactions follow each
electron-transfer step observed for this complex and its
electrode products. Peaks A and B (Figure 1) comprise
a chemically reversible oxidation/reduction couple, and
peaks C and D constitute a chemically reversible reduc-
tion/oxidation couple. The fully oxidized and fully
reduced products are very reactive and have resisted
our efforts to isolate them. ESR measurements and
voltammetric mechanistic studies shed light, however,
on their probable structures.
Oxid a tion of 1. Complex 1 has a single anodic wave
(peak A in Figure 1) in CH2Cl2 which is irreversible at
slow scan rates, n (except at very low concentrations,
vide infra), apparently involving one electron (constant
current function from 0.1 to 1 V/s; diffusion coefficient
at 269 K 1.80 × 10-5 cm2/s; cf. diffusion of ferrocene in
CH2Cl2 at 1.4 × 10-5 cm2/s;12b Ep - Ep/2 ) 46-48 mV
at 250 K compared to 48 mV expected for a Nernstian
one-electron process).13 Qualitatively similar behavior
was observed when the solvent was 1,2-difluorobenzene.
The presence of a coupled cathodic wave (peak A′,
Figure 1) at higher scan rates established the E1/2 value
as 0.57 V vs Fc and gave evidence for at least transient
1+. Since the single product wave (B, Ep,c ) -0.03 V)
found at lower scan rates is ascribed to reduction of the
dimer [12]2+ on the basis of concentration studies, the
oxidation mechanism of 1 is written as a dimerization
following reversible electron transfer (eqs 1 and 2; the
dimerization rate constant is designated as kD(17) since
the reaction involves the 17-electron cation 1+).
ferrocenium couple; on
a practical basis, the latter was
obtained by adding ferrocene as an internal standard at an
appropriate point in each experiment. Since the electrochemi-
cal behavior of 1 appeared to be identical on Pt, Au, and glassy
C electrodes, only the results on Pt are presented here.
Homemade Pt disks of diameter 125, 250, or 625 µm were
employed for cyclic voltammetry (CV) scans, the choice of
electrode depending on scan rate. Steady-state voltammograms
were recorded at a Pt electrode rotating at 1800 rpm. The area
of a larger Pt disk (A ) 0.35 cm2) used for chronoamperometry
studies was calibrated using the diffusion coefficient reported
for ferrocene in CH3CN by Hershberger et al.12a
Voltammetric measurements were performed using a PARC
Model 173 potentiostat interfaced to home-written software.
Along with the electrochemical cell, the electrometer monitor
was housed inside the drybox in order to minimize noise.
Solution temperatures were controlled to better than 1 °C by
immersion of the electrochemical cell in
a temperature-
controlled (FTS Systems) heptane bath. Bulk electrolyses were
conducted in a cell having two compartments separated by a
20 mm fine frit; a Pt-gauze cylinder was used as the working
electrode, and coulometry was monitored with a PARC Model
276 plug-in. Mechanistic criteria for diffusion control, revers-
ibility, etc, were applied as described earlier.13Measurement
of the peak potential shifts for 1 in THF as a function of the
CV scan rate utilized an internal standard of decamethylfer-
rocene in order to account for resistive losses. The anodic peak
height of the standard was equal to the cathodic peak height
of 1 in these experiments.
(10) (a) Schmidt, W.; Steckhan, E. Chem. Ber. 1980, 113, 577. (b)
See also references in: Connelly, N. G.; Geiger, W. E. Chem. Rev. 1996,
96, 877.
(11) Rosseau, K.; Farrington, G. C.; Dolphin, D. J . Org. Chem. 1972,
37, 3968.
(12) (a) Hershberger, J . W.; Klingler, R. J .; Kochi, J . K. J . Am. Chem.
Soc. 1983, 105, 61. (b) Bond, A. M.; Henderson, T. L. E.; Mann, D. R.;
Mann, T. E.; Thormann, W.; Zoski, C. G. Anal. Chem. 1988, 60, 1878.
(13) Geiger, W. E. In Laboratory Techniques in Electroanalytical
Chemistry, 2nd ed.; Kissinger, P. T., Heineman, W. R., Eds.; Marcel
Dekker: New York, 1996; p 683ff.
1 h 1+ + e-
(1)
(2)
kD(17)
2 1+ y z [12]2+
kM(17)
The potentials for oxidation and reduction of 1 are
given in Table 1 along with those for related complexes
that are compared below.

Electron-Transfer Properties of Cp*FeP5
Organometallics, Vol. 18, No. 10, 1999 1829
Ta ble 1. Red ox P oten tia ls (E1/2, V vs F er r ocen e) of
P h osp h a fer r ocen es a n d Selected F er r ocen es
a
complex
0/1+
0/1-
∆E1/2
conditions
Cp*FeP5 (1)
(PC4H4)2Fe
(2,5-Me2PC4H2)2Fe
(3,4-Me2PC4H2)2Fe
CpFe(3,4-Me2PC4H2)
(PC4Me4)2Fe
Cp2Fe
Cp(Me2C5H3)Fe
Cp*CpFe
0.57 -2.00
0.38 -2.55
0.23 -2.84
0.17 -2.73
0.12 -2.95
0.08 -3.00
2.57 this work
2.93 PC, room tempb
3.07 glyme, 233 Kc
2.90 PC, room tempb
3.07 PC, room tempb
3.08 glyme, 233 Kc
3.43
0
-3.43d
-0.10 not obsd
-0.25 not obsd
e
f
F igu r e 3. ESR spectrum at 77 K of 1+ in 1/1 CH2Cl2/C2H4-
Cl2 produced by reaction of 1 with equimolar [(p-BrC6-
H4)3N]+.
a
b
Difference of E1/2 potentials of the two columns. Reference
7: PC ) propylene carbonate, room temp ) ambient temperature.
c Reference 8. Reference 23b.
f The potential of this compound
d
e
is extrapolated from the value for Cp(Me2C5H3)Fe, subtracting 50
mV per methyl group added.
observed at about 0.5 V. As will be shown below, these
data are consistent with the presence of 1+ as the minor
component and the dimeric dication [12]2+ as the major
component. An estimate of the equilibrium constant for
eq 2 can be obtained if it is assumed that the cathodic
peak heights reflect a “frozen” condition, Kdim ) ca. 2 ×
104 M-1
.
The oxidized solution displayed an ESR signal at 77
K with a rhombic g tensor (g1 ) 2.1961, g2 ) 2.1095, g3
) 2.0071) which was also observed when 1 was chemi-
cally oxidized with [(p-BrC6H4)3N]+ in 1/1 CH2Cl2/C2H4-
Cl2 at 196 K and then frozen (Figure 3). Although these
solutions could be repeatedly warmed to room temper-
ature and refrozen to give virtually the same signal
intensity, fluid solutions were always ESR-silent.
The marginal stability of the oxidation product frus-
trated our attempts to isolate it. For example, oxidation
of 1 by [(p-BrC6H4)3N]+ gave solutions that appeared
to be stable for 24 h at 243 K, but workup always gave
solids contaminated significantly with 1 (NMR and
voltammetry evidence) even when the isolation proce-
dures were conducted below 250 K. The oxidation
product was found to be unreactive to either hexam-
ethylbenzene or triphenyl phosphite (20- to 100-fold
excesses) over 30 min, suggesting that it does not
contain a coordinatively unsaturated metal.
Relatively few reactions can be envisioned which may
give rise to an efficient chemically reversible process
coupled to the oxidation of 1 in these media. The most
likely of these are a structural reorganization, H-atom
abstraction and loss, and dimerization.15Since only the
last of these is higher order in reactant, a test of the
anodic response as a function of the concentration of 1
was undertaken.
F igu r e 2. CV scans taken after bulk anodic oxidation of
1 (scan rate 0.2 V/s). Scans were initiated positive of 0.9 V
and taken to increasingly negative values. The reduction
waves for 1+ (wave A′) and [12]2+ (wave B) are observed
along with a small unassigned cathodic feature at about
0.3 V.
Bulk oxidation of 1 was accomplished in several trials
at temperatures between 250 and 260 K with Eappl
)
0.8 V. The original pale green solution turned orange-
brown in the process, and the major electroactive
product had a reduction wave (Figure 2, peak B) at
-0.03 V. Exhaustive reduction of the orange-brown
solution negative of peak B (Eappl ) -0.5 V) regenerated
the original green solution of 1 (coulometry: 1.1 fara-
days/equiv) with overall yields of 85-93% (CV peak
heights, steady-state voltammograms).14
Cyclic voltammograms were obtained on a series of
solutions at five different concentrations of 1 between
the limits of 0.11 and 1.47 mM at 252 K. A large effect
of concentration on the reversibility of the anodic wave
was observed, with a scan rate of 75 V/s being required
to achieve full chemical reversibility at the highest
concentration, in contrast to 2 V/s at the lowest.
Representative CV curves from these experiments are
given in Figure 4. Table 2 gives the ratio of the reverse
to forward peak currents16 for a single scan rate (1 V/s)
at various concentrations. These ratios as well as those
A closer inspection of CV scans taken after the bulk
oxidation reveals the apparent presence of a minor
quantity of 1+ in solution. In Figure 2, scans initiated
at a quiescent potential positive of the initial wave A
are taken to increasingly negative values. The small
cathodic wave A′ arising from reduction of 1+ is clearly
(14) Bulk anodic coulometry on 1 was complicated by the slow
regeneration of 1 under the conditions of the oxidation. The electrolysis
current did not drop below about 5% of the initial value even after
long electrolysis times. After about
1 h, the charge passed was
consistent with approximately 2 faradays/equiv. This type of behavior
is frequently encountered when oxidations are performed at very
positive potentials and is usually ascribed to the tendency of the
oxidized form to react with adventitious nucleophiles such as trace
water. The electrolytic behavior is consistent with our observation that
attempts to isolate the oxidation product always led to solids highly
contaminated with 1.
(15) (a) Trogler, W. C. Organometallic Radical Processes; Elsevier:
Amsterdam, 1990. (b) Astruc, D. Electron Transfer and Radical
Processes in Transition Metal Chemistry; VCH: New York, 1995;
Chapter 5.
(16) Nicholson, R. S. Anal. Chem. 1966, 38, 1406.

1830 Organometallics, Vol. 18, No. 10, 1999
Winter and Geiger
F igu r e 5. Comparison of experimental (solid line) and
calculated (dashed line) cyclic voltammograms for oxidation
of 1 in CH2Cl2 at 252 K: (top) concentration 0.42 mM, v )
2.5 V/s; (bottom) concentration 1.47 mM, v ) 50 V/s.
Simulation parameters were E1/2(1) ) 0.57 V, ks(1) ) 1 cm
s-1, 1 - R ) 0.32; E1/2(122+) ) 0.02 V, ks ) 0.03 cm s-1, R )
F igu r e 4. CV scans of oxidation of 1 in CH2Cl2/0.1 M
[NBu4][PF6] at 252 K (Pt electrode, v ) 0.5 V/s) at two
different concentrations: (top) 1.47 mM; (bottom) 0.42 mM.
0.25, dimerization Keq ) 100, kD(17) ) 1.28 × 104 M-1 s-1
Ru ) 8000 Ω, Cdl ) 0.02 µF.
,
Ta ble 2. Rever sibility Va lu esa for th e Cou p le 1/1+
a n d Dim er iza tion Ra te Con sta n ts kD(17) a s a
F u n ction of Con cen tr a tion (v ) 1.0 V/s)a
wave (peak C) is observed (Ep,c ) -2.05 V at v ) 0.2
V/s) with a single anodic product wave (peak D) on the
reverse scan at Ep,a ) -1.34 V (Figure 1). Identical
results are found in THF, dichloromethane, and 1,2-
difluorobenzene. A number of facts establish that the
cathodic wave is a one-electron process, most notably
bulk coulometry (1.01 faradays/equiv), peak height (0.9
times that of the one-electron oxidation wave of 1,
c
concn (mM)
reversibilityb
10-4kD(17) (M-1 s-1
)
0.11
0.27
0.42
0.54
0.77
0.88
1.47
0.96
0.91
0.71
0.69
0.62
0.58
0.48
0.9(0.2)
1.3(0.2)
1.4(0.4)
1.4(0.2)
1.3(0.2)d
1.7(0.3)
2.1(0.5)
discussed above, in CH2Cl2), and peak breadth (Ep/2
-
a
Conditions: 252 K, CH2Cl2/0.5 M [NBu4][PF6], Pt electrode.
Ep ) 45 mV at 236 K compared to 47 mV for ferrocene).
Assuming a one-electron process, a diffusion coefficient
of 8.6 × 10-6 cm2 s-1 was calculated from chronoamper-
ometry data for 1 in THF at 238 K.
b
The value ic/ia for the couple A/A′ of Figure 4 was measured by
the method of Nicholson.16 These are given for illustrative purposes
to show how the chemical reversibility of the oxidation of 1
decreases with increasing concentration. c Rate constant calculated
from reversibility values as a function of scan rate within a range
in which ic/ia varied between 0.5 and 0.9, using the method of
Lasia.17 The values in parentheses are standard deviations from
The reduction wave at -2.05 V (peak C) and the
oxidation wave at -1.34 V (peak D) constitute a chemi-
cally reversible couple. Double-potential step chrono-
amperometry (5 s step time) between step potentials of
-1.1 and -2.3 V gave reverse-to-forward current ratios
exactly the same as that of ferrocene/ferrocenium, and
bulk electrolysis demonstrated the overall reversibility
on a longer time scale. Concerning the latter, coulom-
etric reduction at Eappl ) -2.3 V gave a deep green
d
determinations at a minimum of six scan rates. This value was
obtained in an experiment separate from the others in the table.
at other scan rates were used to extract the second order
rate constant, kD(17) in eq 2, according to the procedure
of Lasia,17 the results being collected in Table 2. The
average value of kD(17) was [1.4 ((0.3)] × 104 M-1 s-1
.
With this estimate of kD(17) in hand we then calculated
theoretical voltammograms for comparison with the
entire CV curve, getting good overall agreement (Figure
5) with kD(17) ) 1.28 × 104 M-1 s-1 over the entire
concentration range. The voltammetric data are there-
fore entirely consistent with the oxidation product being
a dimeric dication. which we designate [12]2+. We defer
a discussion of its probable structure until later in the
paper.
solution with the major anodic product wave at Ep,a
)
- 1.3 V, consistent with CV experiments (wave D), and
a smaller wave at Ep,a ) - 1.05 V. Scanning negative
to positive through the product wave(s) (Figure 6)
demonstrates that 1 is regenerated from the reoxidation
scans. Bulk reoxidation at Eappl ) -0.7 V gave back the
starting material 1 in 80-90% yield in several replicates
at ca. 270 K. The final reoxidized solution also showed
a small reversible wave at E1/2 ) -0.5 V which is most
likely due to Cp*2Fe. The reduced solutions were ESR-
silent when frozen at 77 K, and they did not react with
an excess of hexamethylbenzene.
Red u ction of 1. Remarkably, the reduction of 1
mirrors the behavior of its oxidation: a single cathodic
(17) Lasia, A. J . Electroanal. Chem. Interfacial Electrochem. 1983,
146, 413.

Electron-Transfer Properties of Cp*FeP5
Organometallics, Vol. 18, No. 10, 1999 1831
12 h 2 1 (6)
breadth shows that a single electron is involved in its
rate-determining step.
The dimerization of the 19-electron system 1- was
further probed by investigating the concentration de-
pendence of the chemical reversibility of the couple 1/1-
(eq 3). These experiments verified the dimerization
hypothesis, albeit less quantitatively than achieved in
the case of the 17e- system. Some chemical reversibility
was observed for 1/1- at v ) 10 V/s (E1/2 ) -2.00 V)
with a concentration of 0.2 mM. Because of the combi-
nation of high scan rates, low concentrations, low
temperature, and a highly resistive solvent, we per-
formed only semiquantitative simulations of the CV
curves. These calculations allow a conclusion that the
dimerization rate for 1- must be between 5 × 105 and
6 × 105 M-1 s-1, i.e., about 40 times higher than
observed for the 17-electron system 1+.
2-
F igu r e 6. CV scans of 12 in THF/0.1 M [NBu4][PF6] (v
) 0.5 V s-1) from -2.45 V to increasingly positive potentials
at 252 K. 122- was prepared from bulk electrolytic reduction
of a solution of 1 at Eappl ) -2.3 V.
An a lysis of Red ox P oten tia ls. The E1/2 potentials
for the oxidation and reduction processes of 1 may be
compared with those observed for analogues including
ferrocene derivatives and sandwich complexes with one
or two monophosphacyclopentadienyl rings (Table 1).
Although the values were obtained under different
conditions, the potentials of 1 appear to be explicable
in terms of a more or less additive substituent effect
when P is substituted for CH in the five-membered
cyclic ligand.
Mechanistic information is obtained from the manner
in which the peak potentials of the cathodic wave of 1
(peak C) and of the anodic product wave D vary with
CV scan rate at 236 K. Over the range 0.05 V/s < v <
10 V/s, Ep,c (peak C) shifted negative by 20 mV per 10-
fold increase in v and Ep,a (peak D) shifted positive by
56 mV (T ) 240 K). The cathodic behavior is consistent
with a fast chemical reaction following electron transfer
(i.e., an EC mechanism), but the data are not sufficiently
precise to distinguish between a first-order reaction
(expected shift 24 mV)18 and a dimerization reaction
(expected shift ca. 20 mV).19 (We show below that
concentration variations support a dimerization model).
The anodic behavior (peak D) is diagnostic, however,
of a one-electron irreversible charge transfer (expected
peak potential shift vs log v for R ) 0.5 system, 48 mV;
observed shift of 56 mV consistent with system having
1 - R ) 0.42).20 The increased breadth of the product
wave D (Ep,a - Ep,a/2 ) 70 mV at room temperature) is
also consistent with irreversible electron transfer, (1 -
R ) ca. 0.4).
All electrochemical data on the reduction of Cp*FeP5
may therefore be understood in terms of eqs 3 and 4,
which describe a reversible one-electron reduction of 1
(eq 3), followed by a fast dimerization reaction (rate
constant kD(19), eq 4) giving the dianion [12]2- for the
cathodic process; the coupled anodic process involves an
irreversible (slow charge transfer) oxidation of the
dimeric dianion (eq 5) followed by rapid cleavage of the
neutral dimer (eq 6) to give back 1.
When the oxidation potential of 1 (E1/2 ) 0.57 V) is
compared with that of Cp*CpFe (E1/2 ) -0.25 V), a
positive shift of 820 mV is noted with the change of five
groups (P for CH), an average of 164 mV per substitu-
tion. Previous work by other groups demonstrated
similar shifts for single substitutions in monophos-
phacyclopentadienyl complexes. Thus, Lemoine et al.
found the oxidation of (PC4H4)2Fe at E1/2 ) 0.38 V vs
ferrocene, a shift of 190 mV per P-for-CH substitution.7
Similarly, the potential for oxidation of (2,5-Me2PC4H2)2-
Fe measured by Elschenbroich et al. (Table 1) is shifted
by 430 mV from a ferrocene with four Me groups, a
change of 215 mV per P-for-CH substitution.8 Lemoine
et al.’s measurement on the isomer (3,4-Me2PC4H2)2Fe
gives a value of 185 mV per substitution. The average
shift for a single substitution from these previous
studies is 196 mV. We conclude therefore that the more
positive oxidation potential of 1 compared to that of
ferrocene can be understood qualitatively as arising
from the weaker electron-donating ability of phospholyl
rings compared to Cp rings21,22 and quantitatively as
an approximately additive effect of this electron-donat-
ing ability as several ring substitutions are involved.
The overall shift for five P atoms (820 mV) is somewhat
less than 5 times the single shift (5 × 196 ) 980 mV),
but the point of deviation from additivity cannot be
discerned from the presently available data.
1 + e- h 1-
(3)
(4)
(5)
kD(19)
2 1- y z [12]2-
kM(19)
[12]2- f 12 + 2e-
Similar arguments are useful to rationalize the
potentials of the reductions of phosphacyclopentadienyl
vs cyclopentadienyl complexes. We assume a reduction
potential for Cp*CpFe of E1/2 ) -3.68 V by taking into
account the observed value of -3.43 V for ferrocene (in
Although the anodic reaction of eq 5 has a net
stoichiometry of two electrons, the voltammetric peak
(18) Nicholson, R. S.; Shain, I. Anal. Chem. 1964, 36, 706.
(19) (a) Saveant, J . M.; Vianello, E. Electrochim. Acta 1967, 12, 1545.
(b) Andrieux, C. P.; Nadjo, L.; Saveant, J . M. J . Electroanal. Chem.
Interfacial Electrochem. 1970, 26, 147.
(20) We write the transfer coefficient as 1 - R for an oxidation and
as R for a reduction.
(21) Poizat, O.; Sourisseau, C. J . Organomet. Chem. 1981, 213, 461.
(22) Waluk, J .; Klein, H.-P.; Ashe, A. J ., III; Michl, J . Organome-
tallics 1989, 8, 2804.

1832 Organometallics, Vol. 18, No. 10, 1999
Winter and Geiger
glyme)23b and subtracting 0.25 V for the substitution of
Cp* for Cp. Then the reduction of 1 (E1/2 ) -2.00 V) is
seen to be easier by 1680 mV, or -336 mV per P-for-
CH ring substitution. This compares to values of about
-390 mV for the mono-P complexes in Table 1 arrived
at by considerations similar to those in the previous
paragraph. The fact that the P-for-CH shifts are sub-
stantially greater for the reductions as compared to the
oxidations is consistent with molecular orbital calcula-
tions that indicate increased P5 character to the LUMO
of 1 as compared to its HOMO.6,24
It is also worth noting that the separation of E1/2
oxidation (18e-/17e-) and reduction (18e-/19e-) poten-
tials decreases steadily from 3.4 V for ferrocene to ca.
3.0 V for monophosphacyclopentadienyl complexes to
2.57 V for 1 (Table 1). This is consistent with a
diminished HOMO/LUMO gap for complexes of in-
creased P-for-CH substitution.
metal or the ligands.30,31 Metal-metal coupling is
extremely unlikely in the present system. The isoelec-
tronic osmocene monocation dimerizes at the metal,
forming a long Os-Os bond,30 but the smaller radius of
Fe compared to Os is expected to preclude this route,
owing to steric hindrance. It also appears that coupling
involving the Cp* rings is unlikely, owing to an apparent
lack of C5Me5 character in the SOMO of 1+. Calculations
have consistently indicated that the ag-type orbitals in
mixed-sandwich phosphaferrocenes (with C5R5 and
CnRnP5-n ligands) are devoid of cyclopentadienyl char-
acter.6,7,24
From these considerations and since mixing of both
2
πP and σP occurs with the metal dz orbital in the SOMO
of 1+, the coupling reactions of the 17e- system likely
involve the two P5 rings. Coupling might occur through
πP-πP, σP-πP, or σP-σP interactions, an idealized draw-
ing of the latter being shown as structure 2. In this case,
ESR Ch a r a cter iza tion of 1+. The ESR character-
istics of anodically electrolyzed 1 (vide supra) confirm
the presence of the 17-electron system 1+ in solutions
containing [12]2+ as the major product. The two most
likely electronic configurations for a d5 metal in a
sandwich ligand field are (a1g)2 (e1g)1 and (e1g)2(a1g)1.25
The ferrocenium ion has the former electronic configu-
ration,26 whereas some ferrocenium analogues (e.g.,
[(2,4-dimethylpentadienyl)CpFe]+) are known to possess
the latter.27 The symmetry labels used here are not
strictly correct, of course, since they refer to the ideal-
ized cylindrical symmetry of sandwich complexes. In
actuality the e1g pair is split by ligand field distortion
and is usually referred to as a “quasi-degenerate” pair.28
removal of σP electron density from a phosphorus sp2
orbital leads to a fulvalenyl-type structure. The appar-
ent diamagnetism of [12]2+ (recall that the ESR activity
of its solutions may be ascribed simply to 1+) is
consistent with 2, since efficient electronic coupling of
the two metal centers is reasonably expected in such a
structure.
The coupling possibilities for the 19-electron anion,
forming diamagnetic [12]2-, are also limited to the ligand
centers. Since the LUMO of 1 contains significant
amounts of π character from both the C5Me5 and P5
ligands,6 the dimer [12]2- likely has the structure 3 or
4, in which the two metal centers are once again 18-
2
2
The metal orbitals involved are dxy and dx -y for the “e1g”
2
pair and dz for the a1g state.
d5 Complexes having dxy or dx -y ground states in
sandwich-type ligand fields are generally characterized
by very large g-value anisotropies and rapid relaxation
rates, often requiring use of temperatures below 77 K
for spectral detection. By contrast, d5 systems with pure
2
2
2
dz ground states have small g-value anisotropies and
may even display fluid-solution ESR spectra if excited
states are sufficiently removed in energy.29 The data for
1+ are clearly inconsistent with a ground-state having
e1g parentage. Consistent with theoretical calculations
which place the a1g orbital above the e1g for 1,6 the
2
primary metal contribution must be dz . There must be
2
significant mixing, however, of the dz orbital with other
metal orbital(s) to explain the fact that there is no
g-value below the free-spin value.25 Finally we note that
the large deviation from axial symmetry in 1+ suggests
a significant lowering of its effective ligand field sym-
metry as opposed to the C5v symmetry of 1.
Lik ely St r u ct u r es of Dim er ic Dica t ion s a n d
Dia n ion s. Seventeen-electron metal sandwich cations
are known to dimerize through coupling at either the
(23) (a) In DMF: Mugnier, Y.; Moise, C.; Tirouflet, J .; Laviron, E.
J . Organomet. Chem. 1980, 186, C49. (b) In glyme at 228 K: Ito, N.;
Saji, T.; Aoyagui, S. J . Organomet. Chem. 1983, 247, 301.
(24) Kostic, N. M.; Fenske, R. F. Organometallics 1983, 2, 1008.
(25) Warren, K. D. In Structure and Bonding; Dunitz, J . D., Ed.;
Springer-Verlag: Berlin, 1976, Vol. 27, p 45.
electron Fe(II), consistent with the apparent diamag-
netism of the dimer dianion. A considerable number of
analogues of 3 are known. Sandwich complexes having
19 valence electrons have been shown to form similar
structures through coupling at one of their cyclic
(26) Prins, R. Mol. Phys. 1970, 19, 603.
(27) Elschenbroich, C.; Bilger, E.; Ernst, R. D.; Wilson, D. R.; Kralik,
M. S. Organometallics 1985, 4, 2068.
(28) Ammeter, J . H. J . Magn. Reson. 1978, 30, 299.
(29) Rieger, P. H. Coord. Chem. Rev. 1994, 135/ 136, 203.
(30) Droege, M. W.; Harmon, W. D.; Taube, H. Inorg. Chem. 1987,
26, 1309.
(31) Reference 15b, pp 349-357.

Electron-Transfer Properties of Cp*FeP5
Organometallics, Vol. 18, No. 10, 1999 1833
hydrocarbons, representative examples being Cp2Rh,32
CpFe(arene),33 and (η4-C4R4) complexes of CpNi34 and
(arene)Co.35 There is also precedent, however, for the
P5-P5 coupled structure in the complex [(Cp′′Rh)2(P5-
P5)(RhCp′′)2]36 (Cp′′ ) 1,3-di-tert-butylcyclopentadienyl).
ferrocene and analogous phosphacyclopentadienyl com-
plexes: a 19-electron anion and a 17-electron cation are
accessible within normal potential limits. The two odd-
electron species derived from 1 react, however, by a
route unknown in the analogues, namely by dimeriza-
tion. The resulting charged dimeric species re-form the
original complex 1 in reverse electrolyses.
This complex possesses a Rh2P10 core that is isoelec-
2-
tronic with the [Fe2P10
]
core of 4.
Su m m a r y
Ack n ow led gm en t. We are grateful to the National
Science Foundation (Grant No. CHE 97-05763), donors
of the Petroleum Research Fund (Grant No. PRF 28321-
AC3), and the DFG for support of this research. We
thank Dr. Thomas Mohr and Prof. O. J . Scherer for a
sample of 1, Prof. U. Zenneck for informing us of the
CV results described in ref 9a, and Prof. Ch. Elschen-
broich for a number of helpful comments.
The sandwich complex Cp*FeP5 has reduction and
oxidation properties similar in one sense to those of
(32) (a) Fischer, E. O.; Wawersik, H. J . Organomet. Chem. 1966, 5,
559. (b) El Murr, N.; Sheats, J . E.; Geiger, W. E.; Holloway, J . D. L.
1979, 18, 1443.
(33) Hamon, J .-R.; Astruc, D.; Michaud, P. J . Am. Chem. Soc. 1981,
103, 758.
(34) Koelle, U.; Ting-Zhen, D.; Keller, H.; Ramakrishna, B. L.;
Raabe, E.; Krueger, C.; Rabbe, G.; Fleischhauer, J . Chem. Ber. 1990,
123, 227.
(35) Herberich, G. E.; Klein, W.; Koelle, U.; Spiliotis, D. Chem. Ber.
1992, 125, 1589.
(36) Scherer, O. J .; Hoebel, B.; Wolmershaeuser, G. Angew. Chem.,
Int. Ed. Engl. 1992, 31, 1027.
OM9809724