Rhodocenium complexes bearing the 1,2,3-Tri-tert-butylcyclopentadienyl ligand: Redox-promoted synthesis and mechanistic, structural and computational investigations

Article abstract of DOI:10.1021/om9707735

Single-electron oxidation of rhodium complexes containing the 1,2,3,5-η-penta-2,4-dienediyl ligand was conducted by electrochemical and chemical means. In all cases, rhodocenium complexes bearing the η⁵-1,2,3-tri-tert-butylcyclopentadienyl ligand were produced in good yield. The results of single-crystal X-ray diffraction studies revealed the solid-state structures of [Rh(η⁵-C₅H₅)(η⁵-C ₅^tBu₃H₂)][PF₆] (1Oa⁺) and [Rh(η⁵C₅H₅)(η⁵-C ₅^tBu₃H₂)[BF₄] (10b⁺). The steric strain in these molecules is apparently relieved by in-plane distortions of the bond lengths and angles of the tri-tert-butylcyclopentadienyl ligand. The results of a deuteriumlabeling study revealed the stereochemistry of the ring-closure reaction. Computational studies using the PM3 semiempirical Hamiltonian suggest that oxidation of the pentadienediyl complexes involves removal of an electron from a molecular orbital centered on the pentadienediyl ligand.

Full text of DOI:10.1021/om9707735

1716
Organometallics 1998, 17, 1716-1724
Rh od ocen iu m Com p lexes Bea r in g th e
1,2,3-Tr i-ter t-bu tylcyclop en ta d ien yl Liga n d :
Red ox-P r om oted Syn th esis a n d Mech a n istic, Str u ctu r a l
a n d Com p u ta tion a l In vestiga tion s
Bernadette T. Donovan-Merkert,* C. Reid Clontz, Leslie M. Rhinehart,
Howard I. Tjiong, and Clifford M. Carlin
Department of Chemistry, The University of North Carolina at Charlotte,
9201 University City Boulevard, Charlotte, North Carolina 28223-0001
Thomas R. Cundari*
Department of Chemistry, University of Memphis, Memphis, Tennessee 38152
Arnold L. Rheingold* and Ilia Guzei
Department of Chemistry, University of Delaware, Newark, Delaware 19716
Received September 2, 1997
Single-electron oxidation of rhodium complexes containing the 1,2,3,5-η-penta-2,4-dienediyl
ligand was conducted by electrochemical and chemical means. In all cases, rhodocenium
complexes bearing the η5-1,2,3-tri-tert-butylcyclopentadienyl ligand were produced in good
yield. The results of single-crystal X-ray diffraction studies revealed the solid-state structures
t
t
of [Rh(η5-C5H5)(η5-C5 Bu3H2)][PF6] (10a +) and [Rh(η5-C5H5)(η5-C5 Bu3H2)][BF4] (10b+). The
steric strain in these molecules is apparently relieved by in-plane distortions of the bond
lengths and angles of the tri-tert-butylcyclopentadienyl ligand. The results of a deuterium-
labeling study revealed the stereochemistry of the ring-closure reaction. Computational
studies using the PM3 semiempirical Hamiltonian suggest that oxidation of the pentadi-
enediyl complexes involves removal of an electron from a molecular orbital centered on the
pentadienediyl ligand.
In tr od u ction
possessing more than one tert-butyl substituent are
those which contain 1,3-di- or 1,3,5- tri-tert-butylcyclo-
pentadienyl ligands.2 Reports on the synthesis of the
1,2-di-tert-butylcyclopentadienyl ligand and complexes
thereof have also appeared.3
Transition-metal, main-group, and f-element com-
plexes bearing cyclopentadienyl ligands with sterically
demanding substituents have attracted considerable
interest. The syntheses and properties of complexes
possessing more than one bulky substituent on a cyclo-
pentadienyl ligand were recently reviewed.1 Most ex-
amples of compounds bearing cyclopentadienyl ligands
Hughes and co-workers recently reported the first
example of a complex bearing the novel 1,2,3-tri-tert-
butylcyclopentadienyl ligand.4 The crystallographically
characterized4b rhodocenium complex 5a + (Scheme 1)
was synthesized by a metal-promoted ring-expansion
reaction of tri-tert-butylvinylcyclopropene 1a . After
cleavage of the vinylcyclopropene ring, the initially
formed pentadienediyl complex 2a was converted to its
indenyl derivative 3a . The conversion of 3a to η4-
cyclopentadiene complex 4a was accomplished by heat-
ing the product in benzene solution or by chromatog-
raphy on cold Florisil. 4a reacted with a variety of
reagents to produce 5a + in good yield. An analogous
reaction starting with deuterium-labeled 1b resulted in
5b+, thereby establishing the stereochemistry of the
ring-closure reaction.
(1) (a) Okuda, J . Top. Curr. Chem. 1992, 160, 97. (b) J aniak, C.;
Schumann, H. Adv. Organomet. Chem. 1991, 33, 291.
(2) For example, see: (a) J ibril, I.; Abu-Nimreh, O. Synth. React.
Inorg. Met.-Org. Chem. 1996, 26, 1409. (b) Sofield, C.; Anderson, R. A.
J . Organomet. Chem. 1995, 501, 271. (c) Schneider, J . J .; Specht, U.
Z. Naturforsch., B 1995, 50, 684. (d) Sitzmann, H.; Wolmershaeuser,
G. Chem. Ber. 1994, 127, 1335. (e) Cotton, J . D.; Byriel, K. A.; Kennard,
C. H.; Scheck, T.; Lynch, D. E. J . Organomet. Chem. 1993, 462, 243.
(f) Scheer, M.; Schuster, K.; Becker, U.; Krug, A.; Hasrtung, H. J .
Organomet. Chem. 1993, 460, 105. (g) Boehme, U.; Langhof, H. Z.
Kristallogr. 1993, 206, 281. (h) Boese, R.; Blaeser, D.; Kuhn, N.;
Stubenrauch, S. Z. Kristallogr. 1993, 205, 282. (i) Reddy, A. C.; J emmis,
E. D.; Scherer, O. J .; Winter, R.; Heckmann, G.; Wolmershaeuser, G.
Organometallics 1992, 11, 3894. (j) Schneider, J . J .; Krueger, C. Chem.
Ber. 1992, 125, 843. (k) Knyazhanskii, S. Y.; Lobkovskii, E. B.;
Bulychev, B. M.; Bel’skii, V. K.; Soloveichik, G. L. J . Organomet. Chem.
1991, 419, 311. (l) Bel’skii, V. K.; Strel’tsova, N. R.; Gun’ko, Y. K.;
Knyazhanskii, S. Y.; Bulychev, B. M.; Soloveichik, G. L. Metalloorg.
Khim. 1991, 4, 1139. (m) Scheer, M.; Schuster, K.; Schenzel, K.;
Herrmann, E.; J ones, P. G. Z. Anorg. Allg. Chem. 1991, 600, 109. (n)
Recknagel, A.; Knoesel, F.; Gornitzka, H.; Noltmeyer, M.; Edelmann,
F. T.; Behrens, U. J Organomet. Chem. 1991, 417, 363. (o) Okuda, J .
J . Organomet. Chem. 1990, 385, C39. (p) Is this just the ligand?
Sitzmann, H. Z. Naturforsch., B 1989, 44, 1293. (q) J utzi, P.; Dick-
breder, R. J . Organomet. Chem. 1989, 373, 301.
(3) (a) Hughes, R. P.; Lomprey, J . R.; Rheingold, A. L.; Yap, G. P.
A. J . Organomet. Chem. 1996, 517, 63. (b) Hughes, R. P.; Lomprey, J .
R.; Rheingold, A. L.; Haggerty, B. S.; Yap, G. P. A. J . Organomet. Chem.
1996, 517, 89. (c) Hughes, R. P.; Lomprey, J . R. Inorg. Chim. Acta 1995,
240, 653. (d) Hughes, R. P.; Kowalski, A. S.; Lomprey, J . R.; Rheingold,
A. L. Organometallics 1994, 13, 2691.
(4) (a) Donovan, B. T.; Hughes, R. P.; Trujillo, H. A. J . Am. Chem.
Soc. 1990, 112, 7076. (b) Donovan, B. T.; Hughes, R. P.; Trujillo, H.
A.; Rheingold, A. L. Organometallics 1992, 11, 64.
S0276-7333(97)00773-5 CCC: $15.00 © 1998 American Chemical Society
Publication on Web 03/31/1998

Rhodocenium Complexes
Organometallics, Vol. 17, No. 9, 1998 1717
Sch em e 1
Interestingly, other related rhodium pentadienediyl
compounds did not convert to complexes bearing the
1,2,3-tri-tert-butylcyclopentadienyl ligand.5-7 As shown
in eq 1, chloride-bridged dimer 2a converted to η4-
η5-1,2,3-tri-tert-butylcyclopentadienyl ligand. In a re-
cent communication,8 we demonstrated that single-
electron chemical or electrochemical oxidation of 9a
rapidly afforded the new rhodocenium complex 10a +
(see eq 4) and that the conversion 3a f5a + could also
cyclopentadiene compound 6 on prolonged standing in
methylene chloride solution.5 Pentamethylcyclopenta-
dienyl derivative 7 (eq 2) converted to η4-cyclobutadienyl
be carried out using the redox-promoted method. In the
present paper, we report the full details of the conver-
sions 9a f 10a + and 3a f 5a +, 5c+, including the
results of single-crystal X-ray diffraction studies of 10a +
and 10b+. We also show that pentamethylcyclopenta-
dienyl complex 7 undergoes an analogous oxidatively
induced transformation to form rhodocenium complex
11. Computational studies of a model compound using
the PM3 semiempirical Hamiltonian suggest that the
oxidation process involves removal of an electron from
a molecular orbital centered on the pentadienediyl
ligand. Finally, we present the results of a deuterium-
labeling study which demonstrates the stereochemistry
of the oxidatively induced reaction.
compound 8 (plus [Rh(C5Me5)Cl2]2) on standing in
halogenated solvents.6 Cyclopentadienyl derivative 9a
(eq 3) did not react when heated in refluxing toluene
Resu lts a n d Discu ssion
for months, although the syn and anti positions of its
deuterium-labeled isotopomer 9b did exchange after
prolonged heating.7
Given the current interest in the synthesis, proper-
ties, and reactions of substituted cyclopentadienyl
ligands, we thought it would be of interest to develop a
method of facilitating the conversion of pentadienediyl
complexes such as 3, 7, and 9 to molecules bearing the
I. Electr och em istr y In vestiga tion s of th e P en -
ta d ien ed iyl Com p lexes. (i) Th e Cyclop en ta d ien yl
Com p lex 9a . Typical cyclic voltammetry (CV) scans
of 9a are shown in Figure 1. The scan represented by
the solid line, initiated in the positive potential direc-
(8) Donovan-Merkert, B. T.; Tjiong, H. I.; Rhinehart, L. M.; Russell,
R. R.; Malik, J . M. Organometallics 1997, 16, 819.
(9) The ratio (ip /ip )of the peak currents of the anodic (ip ) and
a
cathodic (ip ) portioans cof wave B varies with scan rate. For example,
c
(5) Donovan, B. T.; Hughes, R. P.; Kowalski, A. S.; Trujillo, H. A.;
Rheingold, A. L. Organometallics 1993, 12, 1038
(6) Hughes, R. P.; Kowalski, A. S.; Donovan, B. T. J . Organomet.
Chem. 1994, 472, C18.
(7) Donovan, B. T.; Egan, J . W.; Hughes, R. P.; Spara, P. P.; Trujillo,
H. A.; Rheingold, A. L. Isr. J . Chem. 1990, 30, 351.
at 200 mV/s, the scan rate employed in the voltammograms of Figure
1, ip /ip is 0.80 but at 1000 mV/s the ratio is 0.97. This indicates that
a
the procduct generated during the cathodic process of wave B undergoes
a homogeneous reaction whose rate is competitive with the time scale
of the slower scan experiment.10 We are currently investigating the
consequences of this homogeneous reaction.

1718 Organometallics, Vol. 17, No. 9, 1998
Donovan-Merkert et al.
Sch em e 2
F igu r e 1. Cyclic voltammetry scans of 0.7 mM 9a in THF
solution. The scan represented by the solid line was
initiated in the positive potential direction while the scan
represented by the dashed line was initiated in the negative
potential direction. The lettered waves arise from the
following reactions: (A) 9a f 10a + + e-, (B) 10a + + e-
S
10a . Conditions: glassy carbon working electrode; scan rate
) 200 mV/s.
show the absence of wave A and the presence of wave
B regardless of the initial scan direction.
Oxidation of 9a was also accomplished by chemical
means, using 1 equiv of ferrocenium hexafluorophos-
phate as the oxidizing agent. The product of the
chemical oxidation was identified unambiguously as
rhodocenium complex 10a + by NMR spectroscopy and
the results of a single-crystal X-ray diffraction study.
The 1H NMR spectrum (CD3CN solution) displays
singlets at 1.45 ppm (18H) and 1.56 ppm (9H) for the
lateral and central tert-butyl groups, respectively, of the
substituted cyclopentadienyl ligand. A singlet at 5.85
ppm (2H) is assigned to the two ring protons of the tri-
tert-butylcyclopentadienyl ligand. Likewise, the singlet
at 5.89 ppm (5H) is assigned to the five protons of the
unsubstituted cyclopentadienyl ligand. The cyclic vol-
tammetry of 10a + in THF solution (Figure 2) is char-
acterized by a wave of limited chemical reversibility at
E1/2 ) -1.83 V.9 This is the same potential observed
for wave B in the voltammetry of both 9a and the
product of electrochemical oxidation of 9a . Similar
results were obtained when 9a was treated with ferro-
cenium tetrafluoroborate.
F igu r e 2. Cyclic voltammetry of 0.7 mM 10a + in THF
solution. Conditions: glassy carbon working electrode; scan
rate ) 300 mV/s.
tion, shows the presence of a chemically irreversible
oxidation wave labeled A (Ep ) +0.02 V) and a nearly
a
chemically reversible wave labeled B (E1/2 ) -1.83 V).9
The chemical irreversibility of wave A persists at a scan
rate of 2 V/s, indicating that oxidation of 9a results in
a rapid homogeneous reaction.10 The product of the
reaction was identified as rhodocenium complex 10a +
by NMR and a single-crystal X-ray diffraction study
(vide infra). Wave B appears only when wave A is
traversed first and hence is absent in the scan depicted
by the dashed line. Additionally, the cathodic portion
of wave B appeared larger in experiments employing a
10 s holding time immediately after traversing wave A.
These observations indicate that wave B arises from the
reversible reduction of the rhodocenium complex gener-
ated at wave A. The events giving rise to the cyclic
voltammetry of 9a are depicted in Scheme 2.
X-r a y Diffr a ction Stu d ies of 10a + a n d 10b+. The
solid-state structures of 10a + and 10b+ were determined
by the results of single-crystal X-ray diffraction studies.
The crystallographic data for these complexes are listed
in Table 1. Figure 3 displays an ORTEP plot of the
cationic portion of the tetrafluoroborate salt 10b+.
Selected bond distances and angles for 10b+ are pro-
vided in Figure 4. The structure assumes a sandwich
arrangement with both rings binding to the metal in
Exhaustive bulk electrochemical oxidation of 9a was
accomplished by poising the working electrode potential
at +0.15 V. The electrolysis required 1.1 F/mol to come
to completion and caused the originally yellow solution
to become colorless. CV scans obtained after electrolysis
(10) For a discussion of homogeneous reactions following heteroge-
neous electron transfer, see: Bard, A. J .; Faulkner, L. R. Electrochemi-
cal Methods; Wiley: New York, 1980; pp 429-487.

Rhodocenium Complexes
Organometallics, Vol. 17, No. 9, 1998 1719
Ta ble 1. Cr ysta llogr a p h ic Da ta for 10a + a n d 10b+
10a +
10b+
formula
fw
C22H34F6PRh
546.37
C22H34BF4Rh
488.21
space group
a, Å
b, Å
P63/ m
28.185(12)
P21/ c
13.398(6)
15.400(3)
10.754(4)
100.35(3)
2183.0(16)
4
c, Å
15.532(3)
â, deg
V, Å3
10685(7)
18
Z
crystal color, habit
D(calc), g cm-3
µ(Mo KR), cm-1
temp, K
diffractometer
R(F), %a
Rw(F), %a
Rw(F2), %b
light brown prism
1.528
8.40
298(2)
Siemens P4
7.64
colorless block
1.486
8.20
243(2)
Siemens P4
5.54
9.19
14.36
a
Quantity minimized ) ∑∆2; R ) ∑∆/∑(Fo); Rw ) ∑∆w1/2
/
2
∑(Fow1/2), ∆ ) |(Fo - Fc)|. b Quantity minimized ) Rw(F2) ) ∑[w(Fo
- Fc2)2]/∑[w(Fo2)2]1/2; R ) ∑∆/∑(Fo), ∆ ) |(Fo - Fc)|.
F igu r e 4. Selected bond lengths (Å) (top) and bond angles
(deg) (bottom) for the substituted cyclopentadienyl ligand
of complex 10b+.
the value of approximately 1.40 Å observed for rhodium
complexes containing cyclopentadienyl ligands.11 The
bond angles (C(7)-C(6)-C(11) and C(7)-C(8)-C(19))
involving the lateral tert-butyl substituents are signifi-
cantly larger than the analogous bond angles (C(16)-
C(7)-C(6) and C(16)-C(7)-C(8)) encompassing the
central tert-butyl group. These observations are in
agreement with the results of the previously reported
crystal structure of 5a +.4b
F igu r e 3. ORTEP plot of the cationic portion of 10b+.
10a + cystallizes in the hexagonal space group, P63/
m, with three symmetry-independent cations (Figure 5),
each situated on a crystallographic mirror plane (1/2
occupancy), and four independent hexafluorophosphate
anions, with one on a 3h site (1/6 occupancy), one on a
3-fold site (1/3 occupancy), and two on mirror planes. The
ring systems on Rh(1) are perfectly eclipsed, on Rh(3)
perfectly staggered, and on Rh(2) perfectly disordered.
an η5 fashion. The angle between the Rh-ring centroid
vectors is 177.2°. The Rh-centroid distance of the
substituted ligand (1.795 Å) is slightly shorter than the
analogous distance of the unsubstituted cyclopentadi-
enyl ligand (1.819 Å), which can be attributed to the
t
greater inductive effect of the BuCp ligand vs that of
the C5H5 ligand. The cyclopentadienyl ligands are
positioned in an eclipsed conformation.
The disorder was modeled approximately as a hexagon
5
with
/ occupancy carbon and hydrogen atoms but
6
The presence of tert-butyl groups on adjacent carbon
atoms of the substituted cyclopentadienyl ligand in 10b+
raises the question of whether the steric repulsion of
the tert-butyl substituents will result in any deforma-
tions of the molecule. Both rings are nearly planar; out-
of-plane displacement of the tert-butyl substituents is
not observed. The steric strain in the molecule is
apparently relieved by in-plane deformation of the bond
lengths and angles of the tri-tert-butylcyclopentadienyl
ligand. The bond angles of the ring differ from 108°,
the value expected for a pentagon. The bond angles
C(9)-C(8)-C(7) (106.7°), C(8)-C(7)-C(6) (107.0°), and
C(7)-C(6)-C(10) (105.8°), where each of the central
carbon atoms contains a tert-butyl substituent, are
somewhat smaller than 108°, presumably to allow more
room between adjacent tert-butyl groups. As a result,
the remaining angles of the ring are somewhat larger
than 108°. The bond lengths between the ring carbon
atoms bearing tert-butyl substituents are longer than
appears to be more nearly a featureless torus of electron
density. Selected bond distances and angles are given
in Figures 6 and 7, respectively, and in Table 2. The
three cations are bent such that the planes open slightly
on the substituted side; the centroid-Rh-centroid
angles vary from 175.7 to 177.1°. As with 10b+, the
Rh-centroid distances are somewhat shorter to the
substituted ring (average 1.78 Å) as compared to the
unsubstituted ring (average 1.83 Å). Both rings are
planar and bind to the metal in an η5 manner. The tri-
tert-butylcyclopentadienyl ligand experiences in-plane
distortions similar to those described for 10b+ and 5a + 4b
apparently to relieve the steric strain imposed on the
molecule by the adjacent tert-butyl substituents.
(11) For examples, see: (a) Bruce, M. I.; Rodgers, J . R.; Walton, K.
J . Chem. Soc., Chem. Commun. 1981, 1253. (b) Restivo, R.; Ferguson,
G.; O’Sullivan, D. J .; Lalor, F. J . Inorg. Chem. 1975, 14, 3046. (c) Cash,
G. C.; Helling, J . F.; Matthew, M.; Palenik, G. J . J . Organomet. Chem.
1973, 50, 277.

1720 Organometallics, Vol. 17, No. 9, 1998
Donovan-Merkert et al.
F igu r e 5. Comparison of the three crystallographically independent cations of 10a +. At Rh(1) the rings are eclipsed, at
Rh(2) the unsubstituted ring is rotationally disordered and approximately modeled as a six-membered ring, and at Rh(3)
the rings are staggered.
F igu r e 6. Selected bond distances (Å) for the substituted cyclopentadienyl ligands of the three crystallographically
independent cations of 10a +.
(ii) Th e P en ta m eth ylcyclop en ta d ien yl Com p lex,
7. A typical CV scan of 7 is shown in Figure 8.
Initiation of the scan in the positive potential direction
reveals a chemically irreversible oxidation wave (labeled
tained after electrolysis is shown in Figure 9. The
absence of wave A indicates complete consumption of
7.
Two chemically reversible waves are present, a small
one at E1/2 ) +0.15 V (wave F) and a larger one at E1/2
) -2.04 V (wave B). Waves B and F are present
regardless of the direction in which the CV scan is
initiated, indicating that one wave is not a product of
the other. Wave F was determined to be due to the
chemically reversible oxidation of η4-cyclobutadienyl
complex 8 by synthesizing the compound according to
the literature procedure6 and investigating its cyclic
voltammetry behavior: the authentic sample displays
a chemically reversible wave at E1/2 ) +0.15 V, which
matches the half-wave potential of wave F.
A) at Ep ) -0.10 V. Reversal of the scan direction after
a
traversing wave A shows three small chemically ir-
reversible reduction waves (labeled C, D, and E) and a
much larger chemically reversible reduction wave la-
beled B at E1/2 ) -2.03 V. Waves B-E are only
observed when wave A is traversed first. Additionally,
wave B shows a considerable increase in current height
when a 10 s delay is imposed immediately after scan-
ning through wave A. We will demonstrate that wave
A arises from a 1 e- chemically irreversible oxidation
of 7 to form rhodocenium complex 11+ and that wave B
arises from the reduction and reoxidation of 11+.12
To identify the process giving rise to wave B, a sample
of 7 was treated with 1 equiv of ferrocenium hexafluo-
rophosphate. The product was identified as rhodoce-
nium complex 11+ on the basis of 1H NMR data obtained
from a CDCl3 solution of the compound. Singlets
observed at 1.45 ppm (18H) and 1.50 ppm (9H) were
assigned to the lateral and central tert-butyl groups,
respectively, of the tri-tert-butylcyclopentadienyl ligand;
a singlet at 2.07 ppm (15H) was assigned to the methyl
protons of the pentamethylcyclopentadienyl ligand. The
Exhaustive electrochemical oxidation of 7 was con-
ducted by applying a potential of -0.04 V to a platinum
basket working electrode. The electrolysis required 0.92
F/mol and caused the solution to change from orange
to dark pink-orange. A cyclic voltammetry scan ob-
(12) Wave B is completely chemically reversible at all scan rates
employed (50-1000 mV/s). We are currently investigating the conse-
quences of reducing 11+ in bulk electrolysis experiments.

Rhodocenium Complexes
Organometallics, Vol. 17, No. 9, 1998 1721
F igu r e 7. Selected bond angles (deg) for the substituted cyclopentadienyl ligands of the three crystallographically
independent cations of 10a +.
Ta ble 2. Selected Bon d Dista n ces (Å) a n d An gles
(d eg) for 10a +
mol A (Rh1) mol B (Rh2) mol C (Rh3)
ct (C5H5)-Rha
1.82(1)
1.79(1)
176.5(8)
5.1
1.83(1)
1.78(1)
175.7(8)
3.3
1.83(1)
1.76(1)
177.1(9)
2.3
t
ct (C5 Bu3H2)-Rh
ct-Rh-ct
C5 planes, dihedral
a
ct ) centroid.
F igu r e 9. Cyclic voltammetry scan obtained after exhaus-
tive electrochemical oxidation of 0.5 mM 7 in THF solution.
The lettered waves arise from the following processes: (B)
11+ + e- S 11; (F) 8 S 8+ + e-. A platinum disk electrode
was employed with scan rate ) 200 mV/s. The scan was
initiated in the positive potential direction.
oxidation of 7. The electrochemical conversion 7 f 11+
is described in Scheme 2.
F igu r e 8. Cyclic voltammetry scan of 0.99 mM 7 in THF
solution. The scan was initiated in the positive potential
direction. The lettered waves arise from the following
reactions: (A) 7 S 11+ + e-; (B) 11+ + e- S 11. A platinum
disk working electrode was used with scan rate ) 200 mV/
s. The scan was initiated in the positive potential direction.
(iii) Th e In d en yl Der iva tive, 3a . Typical CV scans
of 3a are shown in Figures 10 and 11.
The scan in Figure 10 was initiated in the positive
potential direction whereas the scan in Figure 11 was
initiated in the negative potential direction. A com-
parison of the two voltammograms aids in understand-
ing the electrochemistry of the compound. We turn our
attention first to Figure 11, which shows the simpler of
the two voltammograms. Wave D arises from a chemi-
cally irreversible reduction of 3a ; this wave is present
regardless of whether any other waves have been
traversed first. Waves E and F appear only when D
has been scanned first and hence are products of D.13
Turning now to Figure 10, we see that waves A, G, B,
and C are present in addition to waves D, E, and F.
remaining signal in the spectrum, a singlet at 5.89 ppm
(2H), was assigned to the two equivalent protons of the
tri-tert-butylcyclopentadienyl ligand. Cyclic voltamme-
try scans of 11+ in THF solution show the presence of
a chemically reversible wave at E1/2 ) -2.04 V, the same
potential observed for wave B in CV scans obtained
before and after oxidation of 7. Hence rhodocenium
complex 11+ is produced by chemical or electrochemical

1722 Organometallics, Vol. 17, No. 9, 1998
Donovan-Merkert et al.
F igu r e 12. Cyclic voltammogram obtained after bulk
electrochemical oxidation of 0.86 mM 3a in THF solution.
Conditions: platinum disk woking electrode; scan rate )
200 mV/s. The scan was initiated in the positive potential
direction.
F igu r e 10. Cyclic voltammogram of 1.53 mM 3a obtained
in THF solution. The scan was initiated in the positive
potential direction. Conditions: platinum disk working
electrode; scan rate ) 200 mV/s.
The 1H NMR data of the product matched those
reported in the literature.4b A CV scan of 5c+ in THF
solution shows the presence of two chemically reversible
waves having E1/2 values identical to those of waves B
and C in Figure 10. Scheme 2 describes the events
giving rise to waves A and B in the electrochemistry of
3a . At wave A, the complex undergoes a chemically
irreversible oxidation to form rhodocenium complex 5a +.
The rhodocenium complex then undergoes a chemically
reversible reduction at wave B to form the neutral
rhodocene compound 5. At wave C, 5 undergoes further
reduction.14
II. Deu t er iu m -La b elin g St u d y of t h e R in g-
Closu r e Rea ction . The stereochemistry of the redox-
induced ring-closure reaction was probed by examining
the consequences of oxidizing deuterium-labeled isoto-
pomer 9b in which deuterium is located exclusively in
the syn position of the pentadienediyl ligand. Reaction
of 9b with a stoichiometric amount of ferrocenium
hexafluorophosphate afforded 10c+ which was identified
and shown to contain deuterium on its tri-tert-butylcy-
clopentadienyl ring by H and H NMR studies. A H
NMR spectrum of 10c+ in CDCl3 solution displays
singlet resonances at 1.46 ppm (18H) and 1.55 ppm (9H)
which are assigned to the lateral and central tert-butyl
groups, respectively, of the tri-tert-butylcyclopentadienyl
ligand. The singlet appearing at 5.87 ppm (5H) is
assigned to the protons of the unsubstituted cyclopen-
tadienyl ligand; the remaining singlet at 6.01 ppm (1H)
is attributed to the ring proton of the substituted
cyclopentadienyl ligand. The 2H NMR spectrum dis-
plays only one singlet at 6.03 ppm, indicating the
presence of deuterium on the substituted cyclopentadi-
enyl ligand. Apparently, the redox-promoted ring-
closure reaction proceeds with the same stereochemistry
found for the thermally promoted ring-closure reaction
of indenyl derivative 3b (see Figure 1).4a
F igu r e 11. Cyclic voltammogram of 1.53 mM 3a obtained
in THF solution. The scan was initiated in the negative
potential direction. Conditions: platinum disk working
electrode; scan rate ) 200 mV/s.
Wave A (Ep ) +0.28 V) arises from a chemically
a
1
2
1
irreversible oxidation of 3a . Waves G, B (E1/2 ) -1.41
V) and C (E1/2 ) -2.17 V) are only observed after wave
A has been traversed first. We will demonstrate that
oxidation of 3a at wave A affords 5a +.
3a was oxidized electrochemically by holding the
working electrode potential at +0.50 V. The electrolysis
required 1.33 F/mol and caused the solution to change
from orange-red to bright yellow. Figure 12 shows a
typical CV scan obtained after electrolysis. The absence
of wave A indicates complete consumption of 3a during
the course of the electrolysis. The major waves present
after electrolysis are B and C, which have E1/2 values
identical to those of waves B and C of Figure 10. These
waves arise from sequential one-electron reductions of
rhodocenium complex 5a + (vide infra). Waves H, I, and
J probably arise from redox events of minor products
generated during the electrolysis.
III. Com p u ta tion a l Stu d ies of th e Rin g-Closu r e
Rea ction . Several model rhodium organometallic com-
plexes were studied using the PM3 semiempirical
The reaction of 3a with 1 equiv of ferrocenium
tetrafluoroborate produced rhodocenium complex 5c+.
(14) We are currently studying the consequences of reducing 5a +
in more detail. It is possible that the indenyl ligand of electrogenerated
rhodocene 5 may undergo ring slippage to form an η3-indenyl complex.
(13) The consequences of reducing 3a are currently under investiga-
tion in our laboratory.

Rhodocenium Complexes
Organometallics, Vol. 17, No. 9, 1998 1723
cyclopentadienyl ligand. Oxidation of deuterium-la-
beled isotopomer 9b revealed that the stereochemistry
of the redox-induced ring-closure reaction 9b+ f 10c+
studied in this work is in agreement with the thermally
induced ring-closure reaction 3a f 5a + reported by
Hughes and co-workers.4a We are currently investigat-
ing the redox-induced reactions of other rhodium pen-
tadienediyl complexes and the consequences of reducing
the rhodocenium complexes 5+, 10+, and 11+. The
results of these studies will be reported in due course.
Exp er im en ta l Section
Syn th etic P r oced u r es a n d Ch em ica ls. All reactions
were conducted in oven-dried glassware under a dinitrogen
atmosphere using either standard Schlenk or glovebox tech-
niques unless otherwise noted. Solvents were distilled from
the appropriate drying agents (potassium metal for tetrahy-
drofuran, diethyl ether, and petroleum ether (bp 30-60 °C); 4
Å molecular sieves for acetone; calcium hydride for methylene
chloride) immediately prior to use. Compounds 1a ,7 2,7 3a ,4
7,6 and 9a 7 were prepared according to the literature methods.
Ferrocenium hexafluorophosphate and ferrocenium tetrafluo-
roborate were prepared by treating ferrocene with silver
hexafluorophosphate or silver tetrafluoroborate, respectively,
in acetone solution.17 Tetrabutylammonium hexafluorophos-
phate, purchased from Southwest Analytical Chemicals, was
recrystallized from acetone three times and dried under
vacuum for 24 h prior to use. Silica gel (Davisil 62) was
purchased from Mallinckrodt and used as received. 1H, 2H,
and 13C NMR data were obtained at 300.197, 46.082, and
75.582 MHz, respectively, using a General Electric QE-300
Spectrometer modified with a Tecmag Aquerius software
upgrade.
E lect r och em ica l P r oced u r es. Electrochemical studies
were performed on an EG&G Princeton Applied Research
(PAR) model 273 potentiostat/galvanostat driven using the
EG&G model 270/250 research electrochemistry software
package. All electrochemistry experiments were conducted in
an Mbraun Lab Master 100 glovebox. Experiments employed
a conventional three-electrode configuration. A Luggin capil-
lary probe for the reference electrode (Ag/AgCl of SCE) and
positive-feedback iR compensation were used routinely in
voltammetry experiments. In bulk electrolysis experiments,
the auxiliary (platinum mesh) and reference electrodes were
separated from the working electrode compartment by fine
sintered glass frits. The applied potential was held at the
value specified until the current decayed to less than 5% of
its original value. All potentials are referenced to the fer-
rocene/ferrocenium redox couple, used as an internal standard;
this redox couple has an E1/2 value of +0.56 V vs SCE under
the same experimental conditions employed in this study
(tetrahydrofuran solution containing 0.1 M tetrabutylammo-
nium hexafluorophosphate).
F igu r e 13. Calculated structure A, showing selected bond
distances and ∆Hf(PM3).
Hamiltonian after calibration of the method by com-
parison of known and predicted structures for prototypi-
cal rhodium complexes.15 An extensive survey of over
40 plausible complexes with formula CpRh(C5H6) yielded
two structures that have calculated energies signifi-
cantly (i.e., g4 kcal mol-1) lower than all others.
Calculated structure A (CpRh(pentadienediyl)) (see
Figure 13) is in very good agreement with experimental
structure 9a and is 0.5 kcal mol-1 more stable than the
other low-energy isomer, CpRh(η4-cyclopentadiene).
It is of interest to probe the molecular and electronic
structural consequences resulting when complexes such
as 9a are converted to rhodocenium derivatives by
oxidation. Removal of an electron from A followed by
geometry optimization resulted in radical cation A+•
.
The change in the geometry of the C5H6 moiety upon
oxidation of A to A+• hinted of incipient ring closure;
this group became more planar, and the distance
between nonbonded carbons (C(22) and C(14)) decreased
from 2.76 Å (A) to 2.60 Å (A+•). The changes in
geometry of the pentadienediyl ligand upon oxidation
of A suggest that the electron is removed from a
molecular orbital (MO) centered on the C5H6 (pentadi-
enediyl) ligand. This inference is supported by the
observation that the two highest MOs in A are centered
on the C5H6 ligand and that the majority of the spin
density in A+• is located on the C5H6 ligand.16
Con clu sion s
The results presented herein demonstrate that al-
though their thermally promoted reactions differ greatly
from one another, pentadienediyl complexes 3, 7, and
9 all undergo single-electron oxidation to afford rhodoce-
nium complexes bearing the novel 1,2,3-tri-tert-butyl-
The cyclic voltammetry of 3a , 7, and 9a was investigated
in tetrahydrofuran solution using either glassy carbon, plati-
num, or gold disk working electrodes. The voltammetry
responses of the compounds were unaffected by the choice of
electrode material. Linear plots of peak potential vs (scan
rate)1/2 were obtained for the oxidation wave of each of the
compounds (wave A in Figures 1, 8, and 10), indicating
diffusion control of mass transfer.
Rea ction of 9a w ith F cP F 6. 9a (0.334 g, 0.829 mmol) and
ferrocenium hexafluorophosphate (0.284 g, 0.858 mmol) were
allowed to react in acetone (25 mL) solution under a nitrogen
atmosphere for 1 h. After removal of the solvent under
reduced pressure, the product was purified in the air by
(15) Calculations were performed using the program MacSpartan
Plus (Wavefunction Inc., Irvine, CA, 1996) and the PM3 semiempirical
Hamiltonian. The PM3 method is described in: Stewart, J . J . P. J .
Comput.-Aided Mol. Des. 1990, 4, 1.
(16) Although this computational investigation does not indicate that
an η4-cyclopentadiene complex actually forms as a result of removing
an electron from CpRh(C5H6), it should be pointed out that electro-
chemical oxidation of an η4-cyclopentadiene rhodium complex results
in loss of H• and formation of an η5-cyclopentadienyl complex. See:
Ustynyuk, N. A.; Pererleitner, M. G.; Gusev, O. V.; Denisovich, L. I.
Russ. Chem. Bull. 1993, 1727. For an analogous reaction involving an
(η4-pentamethylcyclopentadiene)platinum complex, see: Gusev, O. V.;
Morozova, L. K.; Peganova, T. A.; Peterleitner, M. G.; Peregudova, S.
M.; Denisovich, L. I.; Petroskii, P. V.; Oprunenko, Y. F.; Ustynyuk, N.
A. J . Organomet. Chem. 1995, 493, 181.
(17) For an excellent review on chemical redox reagents, see:
Connelly, N. G.; Geiger, W. E. Chem. Rev. 1996, 96, 877.

1724 Organometallics, Vol. 17, No. 9, 1998
Donovan-Merkert et al.
chromatography on a silica gel column (5.5 cm × 2.54 cm)
packed in petroleum ether (bp 30-60 °C). The product was
introduced into the column by dissolving it in a minimum of
methylene chloride. Ferrocene was removed from the column
by elution with petroleum ether. Once the eluent became
completely colorless, acetone was introduced into the column,
resulting in elution of 10a +. Solvent was removed under
reduced pressure to afford 0.342 g of 10a + as a tan solid (78%
yield). An analytically pure sample was obtained by adding
0.6 mL of chloroform to 0.0770 g of the product and sonicating
the resulting mixture at 45 °C. This left a brown insoluble
product at the bottom of the flask. The colorless solution was
decanted into a clean flask. Solvent was removed, leaving a
white powder. Yield (based on crude product): 0.038 g (50%).
Anal. Calcd for C22H34F6PRh: C, 48.36; H, 6.27. Found: C,
48.39; H, 6.33. Crystals suitable for single-crystal X-ray
analysis were obtained by vapor diffusion of diethyl ether into
a saturated solution of 10a + (obtained from chromatographic
purification) in methanol. 1H NMR (acetone-d6): δ 1.53 (s,
with acetone. Evaporation of the solvent afforded a tan powder
(0.049 g, 0.008 mmol, 41%). 1H NMR (CDCl3): δ 1.45 (s, 18H,
t
t
tBu), 1.50 (s, 9H, Bu), 2.07 (s, 15H, C5Me5), 5.89 (s, 2H, C5 -
Bu3H2).
Rea ction of 3a w ith F cBF 4. 3a (0.156 g, 0.344 mmol),
ferrocenium tetrafluoroborate (0.108 g, 0.94 mmol), and ac-
etone (15 mL) were placed in a Schlenk flask under a nitrogen
atmosphere, and the resulting solution was allowed to stir for
30 min. The volume was then reduced in vacuo to ca. 1-2
mL to obtain a viscous brown solution. The residue was rinsed
with petroleum ether to remove ferrocene. Recrystallization
of the product from acetone afforded bright yellow crystals of
5c+ (0.076 g, 41% yield based on the original amount of 3a
used).
A
1H NMR spectrum obtained in CDCl3 solution
matched the literature spectrum.4b
Cr ysta llogr a p h ic Str u ctu r a l Deter m in a tion . Crystal,
data collection, and refinement parameters for 10a + and 10b+
are given in Table 1.
The systematic absences in the diffraction data were
consistent for space groups P63 and P63/m for 10a + and were
uniquely consistent for space group P21/c for 10b+. In the case
of 10a +, the latter centrosymmetric space group was preferred
on the basis of the chemically reasonable and computationally
stable results of refinement. The structures were solved using
direct methods, completed by subsequent difference Fourier
synthesis and refined by full-matrix least-squares procedures.
The absorption corrections were not required because the
variation in the integrated ψ-scan intensities was less than
10% for both 10a + and 10b+.
The asymmetric unit of 10a + consists of three independent
cations residing on mirror planes and four independent anions
(two on mirror planes, one on a 3-fold axis, and one on a 3-fold
inversion axis). Due to limited data, anisotropic refinement
was confined to the Rh, P, and F atoms. Disorder in one of
the Cp rings on Rh(2) is described in the text. In addition,
one of the anions is axially disordered about P(1) which, due
t
t
t
18H, Bu), 1.64 (s, 9H, Bu), 6.09 (s, 2H, C5 Bu3H2), 6.19 (s,
t
5H, C5H5). 1H NMR (CD3CN): δ 1.45 (s, 18H, Bu), 1.56 (s,
t
t
9H, Bu), 5.85 (s, 2H, C5 Bu3H2), 5.89 (s, 5H, C5H5). 13C{1H}
NMR (acetone-d6): δ 33.71 (s), 34.72 (s), 35.39 (s), 35.62 (s),
86.31 (d, J Rh-C ) 9.7 Hz), 88.64 (d, J Rh-C ) 7.3 Hz), 118.30 (d,
J Rh-C ) 7.3 Hz), 123.42 (d, J Rh-C ) 9.7 Hz).
Rea ction of 9a w ith F cBF 4. 9a (0.231 g, 0.573 mmol),
ferrocenium tetrafluoroborate (0.165 g, 0.603 mmol), and
acetone (25 mL) were added to a Schlenk flask. After the
resulting mixture was allowed to stir for 1 h, solvent was
removed under reduced pressure. The product was dissolved
in a minimum amount of methylene chloride, and the solution
was placed on a silica gel column (6 cm × 2.5 cm) packed in
petroleum ether. After removal of all of the ferrocene from
the column by elution with petroleum ether, the desired
product was eluted using acetone. Removal of the solvent
afforded 10b+ (0.230 g, 82% yield). The product was then twice
recrystallized from chloroform as described for 10a + to obtain
white crystals (0.091 g, 33% based on the amount of 9b used).
to its presence on a 3-fold inversion site, contains an equatorial
2
plane with six F atoms, instead of four, each refined with
/
3
t
1H NMR (CDCl3): δ 1.46 (s, 18H, lateral Bu), 1.55 (s, 9H,
occupancy. For 10b+, all non-hydrogen atoms were refined
with anisotropic displacement coefficients. In both structures
all hydrogen atoms were treated as idealized contributions.
All software and sources of the scattering factors are
contained in SHLXTL (version 5.03) and various versions of
the SHLXTL PC and VAX program library (G. Sheldrick,
Siemens XRD, Madison, WI).
central tBu), 5.87 (s, 2H, C5 Bu3H2), 6.01 (s, 5H, C5H5).
t
Crystals suitable for single-crystal X-ray analysis were ob-
tained by dissolving some of the solid in acetone-d6 and
allowing the solvent to evaporate over the course of 4 months.
Rea ction of 9b w ith F cP F 6. 9b (30 mg, 0.074 mmol),
ferrocenium hexafluorophosphate (25 mg, 0.074 mmol) and
acetone (10 mL) were combined in a Schlenk flask. After the
mixture was allowed to stir for 1 h, solvent was removed. The
residue was chromatographed on silica gel packed in petroleum
ether. Ferrocene was removed from the column by elution
with petroleum ether. 10c+ was eluted with acetone. Removal
of the solvent afforded an off-white powder. 1H NMR
Ack n ow led gm en t. B.T.D.-M. is grateful to the
donors of the Petroleum Research Fund, administered
by the American Chemical Society, and The University
of North Carolina at Charlotte for generous financial
support of this research and to Professor Russell P.
Hughes for providing a sample of 2b. T.R.C. acknowl-
edges the National Science Foundation for support
(Grant CHE-9614346).
t
t
(CDCl3): δ1.46 (s, 18H, Bu), 1.55 (s, 9H, Bu), 5.87 (s, 5H,
t
C5H5), 6.01 (s, 1H, C5 Bu3HD). 2H NMR (CHCl3): δ 6.03 (s,
t
1D, C5 Bu3HD).
Rea ction of 7 w ith F cP F 6. Acetone (15 mL) was added
to a mixture of 7 (0.090 g, 0.188 mmol) and ferrocenium
hexafluorophosphate (0.069 g, 0.201 mmol) in a Schlenk flask.
The resulting mixture was allowed to stir at room temperature
for 2 h. After removal of solvent, the product was chromato-
graphed on a silica gel column (14.5 cm × 1.5 cm) packed in
petroleum ether. After removal of ferrocene from the column
using petroleum ether, the rhodocenium complex was eluted
Su p p or tin g In for m a tion Ava ila ble: Tables of crystal
data, structure refinement details, atomic coordinates, bond
distances, bond angles, and thermal parameters (18 pages).
Ordering information is given on any current masthead page.
OM9707735