Metallabenzenes and valence isomers. 3. Unexpected rearrangement of two regioisomeric iridabenzenes to an (η5-cyclopentadienyl)iridium(I) complex

Article abstract of DOI:10.1021/om0201563

The reaction of (Z)-1-phenyl-2-(trimethylsilyl)-3-(2-lithiovinyl)cyclopropene (4) with Vaska's complex generates the iridabenzvalene 8, the iridabenzene 9, and the cyclopentadienyl complex 10 in a 10:2:3 ratio. Heating this mixture to 75 °C converts 8 and 9 to 10. NMR studies over a 24 h period at 75 °C show that samples containing pure 8 isomerize to 10 in high yield and generate regioisomeric iridabenzene 11 as an intermediate.

Full text of DOI:10.1021/om0201563

2824
Organometallics 2002, 21, 2824-2826
Meta lla ben zen es a n d Va len ce Isom er s. 3. Un exp ected
Rea r r a n gem en t of Tw o Regioisom er ic Ir id a ben zen es to
a n (η5-Cyclop en ta d ien yl)ir id iu m (I) Com p lex
He-Ping Wu, Seren Lanza, Timothy J . R. Weakley, and Michael M. Haley*
Department of Chemistry, University of Oregon, Eugene, Oregon 97403-1253
Received February 25, 2002
Summary: The reaction of (Z)-1-phenyl-2-(trimethylsi-
lyl)-3-(2-lithiovinyl)cyclopropene (4) with Vaska’s com-
plex generates the iridabenzvalene 8, the iridabenzene
9, and the cyclopentadienyl complex 10 in a 10:2:3 ratio.
Heating this mixture to 75 °C converts 8 and 9 to 10.
NMR studies over a 24 h period at 75 °C show that
samples containing pure 8 isomerize to 10 in high yield
and generate regioisomeric iridabenzene 11 as an inter-
mediate.
isomers, namely iridabenzvalenes (e.g., 3), can be gener-
ated and isomerized cleanly to iridabenzenes.5 Exten-
sion of this new route to other metals and other cyclo-
propenes should permit access to numerous metalla-
benzenes bearing a variety of substituents on the carbon
backbone. In this latter regard, we present herein the
reaction of lithiate 4 with Vaska’s complex, resulting
in formation of an iridabenzvalene and two regioiso-
meric iridabenzenes, and their facile rearrangement to
an (η5-cyclopentadienyl)iridium(I) complex.
Metallabenzenes,1 organic/transition-metal “hybrids”
which possess aromatic properties, have been shown to
exhibit many similarities to heterobenzenessdownfield
chemical shifts for ring protons, planarity of the six-
membered metallacycle, no alternation of bond lengths,
arene displacement from (arene)Mo complexes, and even
electrophilic aromatic substitution.2,3 The transition-
metal center, however, exerts significant influence on
the chemistry of the system by undergoing oxidative-
addition, ligand-dissociation, and electron-transfer pro-
cesses. Unfortunately, the limited number of metalla-
benzenes and the lack of a general synthetic route to
this class of metallacycles have restricted detailed
systematic studies. Recently, we described a direct path
to generate metallabenzenes, which involved an in-
tramolecular, metal-promoted cyclopropene rearrange-
ment. This method has been applied successfully to
generate iridabenzenes (e.g., 1) by reactions of nucleo-
The choice of cyclopropene 5 (Scheme 1) as the ligand
for this study was inspired by three factors. We wished
to see (1) how replacement of one of the phenyl units
with a trimethylsilyl group would influence the overall
chemistry of metallabenzene formation, (2) what would
be the regiochemical outcome using an unsymmetrical
cyclopropene, and (3) if removal of the trimethylsilyl
group by fluoride ion could lead to monosubstituted
metallaaromatics. Compound 5 was prepared by reduc-
tion of ester 66 with DIBAL-H and subsequent oxidiza-
tion with Dess-Martin periodinane.7 The Wittig reac-
tion of aldehyde 7 with Ph3PdCHI8 gave 5 with high
stereoselectivity (>15:1 Z:E). Lithium-iodine exchange
with 1 equiv of butyllithium in Et2O at -78 °C gener-
ated lithiate 4. Addition of this intermediate to (Ph3P)2-
Ir(CO)Cl at -78 °C and then warming to ambient
temperature over a 3 h period resulted in a mixture of
products. Fast chromatography on silica gel gave a 47%
yield of purified mixture of 8, 9, and 10 in a ratio of
10:2:3. Complexes 8 and 10 could be isolated indepen-
dently through chemical methods (by treatment of the
purified mixture with MeI and with heat (vide infra),
respectively); nevertheless, samples of 9 always con-
tained significant amounts of either 8 or 10. Solutions
of 8 and 10 are stable at room temperature for several
weeks, while solutions of 10 “enriched” with 9 show that
the signals for 9 slowly disappear over 1 week at 0 °C.
The composition of complexes 8 and 9 was determined
by comparison of their NMR chemical shifts with those
of structurally similar analogues 3 (R ) Me)5a and 1 (R
) Ph),4 respectively. Resonance peaks in the 1H/13C
NMR spectra corresponding to the three CH groups in
iridabenzvalene 8 are observed at δH/δC 6.66, 6.33, 3.08/
philic (Z)-1,2-diphenyl-3-(2-lithiovinyl)cyclopropene (2)
with Vaska’s complex4 and several Vaska-type com-
plexes.5 In addition, the first metallabenzene valence
* To whom correspondence should be addressed. E-mail: haley@
oregon.uoregon.edu. Fax: 541-346-0456
(1) Bleeke, J . R. Chem. Rev. 2001, 101, 1205-1227.
(2) Inter alia: (a) Thorn, D. L.; Hoffman, R. Nouv. J . Chim. 1979, 3,
39-45. (b) Elliott, G. P.; Roper, W. R.; Waters, J . M. J . Chem. Soc.,
Chem. Commun. 1982, 811-813. (c) Elliot, G. P.; McAuley, N. M.;
Roper, W. R. Inorg. Synth. 1989, 26, 184-189. (d) Bleeke, J . R.; Xie,
Y.-F.; Peng, W.-J .; Chiang, M. Y. J . Am. Chem. Soc. 1989, 111, 4118-
4120. (e) Bleeke, J . R.; Behm, R.; Xie, Y.-F.; Chiang, M. Y.; Robinson,
K. D.; Beatty, A. M. Organometallics 1997, 16, 606-623.
(3) Richard, C. E. F.; Roper, W. R.; Woodgate, S. D.; Wright, L. J .
Angew. Chem., Int. Ed. 2000, 39, 750-752.
(6) Cyclopropene 6 was prepared in 30% yield using copper powder
instead of the reported rhodium complex as the catalyst for the reaction
of 1-phenyl-2-(trimethysilyl)acetylene with ethyl diazoacetate: Mu¨ller,
P.; Gra¨nicher, C. Helv. Chim. Acta 1993, 76, 521-534.
(7) Dess, D. B.; Martin, J . C. J . Am. Chem. Soc. 1991, 113, 7277-
7287.
(8) (a) Stork, G.; Zhao, K. Tetrahedron Lett. 1989, 30, 2173-2174.
(b) Bestmann, H. J .; Rippel, H. G.; Dostalek, R. Tetrahedron Lett. 1989,
30, 5261-5262.
(4) Gilbertson, R. D.; Weakley, T. J . R.; Haley, M. M. J . Am. Chem.
Soc. 1999, 121, 2597-2598.
(5) (a) Gilbertson, R. D.; Weakley, T. J . R.; Haley, M. M. Chem. Eur.
J . 2000, 6, 437-441. (b) Gilbertson, R. D.; Lau, T. L.; Lanza, S.;
Weakley, T. J . R.; Haley, M. M. Manuscript in preparation.
10.1021/om0201563 CCC: $22.00 © 2002 American Chemical Society
Publication on Web 06/12/2002

Communications
Organometallics, Vol. 21, No. 14, 2002 2825
Sch em e 1
151.25, 144.15, 56.97 ppm, which are close to those
reported previously for 3 (R ) Me) (δH/δC 7.25, 6.55,
3.82/146.76, 141.79, 60.83 ppm).5a The upfield shift of
ring protons in 8 are attributed to stronger shielding
by the -SiMe3 group. It is noteworthy that iridacycle
8, which has bulkier, less donating phosphines com-
pared to 3 (R ) Me), forms as the major product of this
reaction, whereas the corresponding diphenyl-substi-
tuted molecule from Vaska’s complex (3, R ) Ph) is
never observed. A possible explanation is that the longer
carbon-silicon bond makes the cyclopropene π-bond
accessible for bonding to the electron-deficient Ir center,
as well as increased electron density of the π-bond due
to the electron-donating -SiMe3 group. Typical of met-
allabenzenes, the 1H NMR spectrum of 9 exhibits
downfield resonance peaks of the hydrogens ortho and
para to the iridium at 10.45 and 8.14 ppm, which are
comparable to the analogous signals in 1 (R ) Ph) at
10.79 and 8.44 ppm, respectively.4 The signals for the
meta H and the remaining protons of 9 were not
observed because of the small amount of material, the
instability of the compound, and overlap with the phenyl
resonances of 8 and 10.
F igu r e 1. Molecular structure of complex 10. The thermal
ellipsoids are drawn at the 30% probability level. Se-
lected bond lengths (Å) and angles (deg): Ir-P ) 2.235(1),
Ir-C1 ) 1.814(4), Ir-C2 ) 2.251(3), Ir-C3 ) 2.306(3),
Ir-C4 ) 2.254(4), Ir-C5 ) 2.310(4), Ir-C6 ) 2.284(4),
C3-Si ) 1.872(4), C2-C3 ) 1.433(5), C2-C6 ) 1.437(5),
C2-C7 ) 1.498(5), C3-C4 ) 1.438(5), C4-C5 ) 1.420(6),
C5-C6 ) 1.395(6); P-Ir-C1 ) 90.3(1), P-Ir-Cp ) 133.4-
(1), C1-Ir-Cp ) 136.3(2), Ir-C1-O ) 177.4(3), C3-C2-
C6 ) 108.7(3), C3-C2-C7 ) 125.3(3), C6-C2-C7 )
125.3(3), Si-C3-C2 ) 130.9(3), Si-C3-C4 ) 123.9(3),
C2-C3-C4 ) 105.1(3), C3-C4-C5 ) 109.9(3), C4-C5-
C6 ) 107.6(3), C2-C6-C5 ) 108.5(3).
Sch em e 2
1
Complex 10 was unambiguously characterized by H
and 13C NMR spectroscopy and by single-crystal X-ray
diffraction. The 1H/13C NMR resonances of the three
-CH groups in the cyclopentadienyl ring of 10 appear
at δH/δC 5.40, 4.67, 4.54/91.06, 89.67, 85.47 ppm. Signals
at 111.85 and 92.25 ppm are attributed to the two
quaternary C atoms. The IR spectrum shows a band at
1936 cm-1 for the CO stretching frequency. X-ray-
quality crystals of 10 were grown from hexane. Selected
bond lengths and angles are listed in Figure 1.9 In
comparison to the parent CpIr(CO)PPh3 complex,10
the carbon-carbon bonds in the cyclopentadienyl moiety
are lengthened on average about 0.035 Å. The effect
of steric interactions between the bulky trimethylsi-
lyl and phenyl groups is reflected in a slightly larger
Si-C3-C2 bond angle (130.9(3)°) compared to
Si-C3-C4 (123.9(3)°).
In an attempt to prepare a pure sample of 9, we
heated a benzene-d6 solution of 8 at 75 °C for 24 h.
Although valence isomerization to yield 9 was the
expected transformation, 1H NMR analysis of the reac-
tion mixture showed exclusive and essentially quantita-
tive formation of 10. This result is surprising, given the
ease with which benzvalene 3 (R ) Me) can be converted
to 1 (R ) Me). Closer inspection of the thermal chem-
istry of these complexes showed multiple conversions.
Analysis of a sample of 8 heated at 75 °C for 1 h
revealed, in addition to unreacted 8 and a small amount
(9) Crystal data for 10: C33H32IrOPSi, Mr ) 695.9, monoclinic, P21/
a, a ) 9.1428(17) Å, b ) 32.708(3) Å, c ) 9.7927(9) Å, â ) 94.344(12)°,
V ) 2920.0(6) Å3, Z ) 4, Dcalcd ) 1.583 g cm-3, µ ) 47.0 cm-1, F(000)
) 1376, Mo ΚR radiation (λ ) 0.710 73 Å), T ) 300 K, 2θmax ) 50°,
5114 independent reflections scanned, 5114 reflections in refinement
on F2, 463 parameters, R(F) ) 0.022 (I ) σ(I)), Rw(F2) ) 0.052 (all data).
Data were collected on an Enraf-Nonius CAD-4 Turbo diffractometer.
Structure refinement (C atoms anisotropic, H atoms isotropic) was
accomplished with the teXsan program suite (version 1.7 for SGI
workstations).
1
of 10, the characteristic H NMR signals for an irida-
benzene; however, these resonances were at 10.85 and
8.78 ppm, suggesting the formation of a different
iridabenzene. The assignment of 11 as the regioisomer
(10) Bennett, M. J .; Pratt, J . L.; Tuggle, R. M. Inorg. Chem. 1974,
13, 2408-2413.

2826 Organometallics, Vol. 21, No. 14, 2002
Communications
Sch em e 3
depicted in Scheme 2 is based on H-29Si gHMQC NMR
experiments, which indicate that the para proton at 8.78
ppm is coupled to a silicon nucleus.11 Careful monitoring
of the reaction showed that the concentration of 11
peaked after 8-10 h and then steadily decreased.
Benzvalene 8 was completely consumed after 15 h, with
total conversion of 8 and 11 to 10 after 24 h at 75 °C in
89% isolated yield. Heating the purified mixture of 8-10
furnished similar results. After 1 h, the resonances of
9 had disappeared and been replaced by signals for 11.
The intensity of the signals for 8 and 10 had decreased
and increased, respectively, suggesting that 9 goes to
10 and 8 goes to 11. The disappearance of 8 and 11
progressed as described above upon further heating.
After 24 h, transformation to 10 (with concomitant
formation of PPh3) was complete with a color change
from red-brown to yellow.
Once formed, benzvalene 8 is stable at room temper-
ature but slowly isomerizes to iridabenzene 11 upon
heating. Although much less thermally labile than 9,
11 also undergoes carbene migratory insertion at el-
evated temperatures to give 10. It is possible that 9 is
also an intermediate in the conversion of 8 to 10;
however, 9 was not detected during the isomerization
of pure samples of 8. The conversion of 9 and 11 to 10
is quite unexpected, since all other iridabenzenes that
we4,5 and others1,2d,e have previously studied have not
undergone a similar rearrangement. For example,
samples of 3 (R ) Me, Et, i-Bu) can be cleanly and
quantitatively isomerized to 1 with no evidence of
subsequent rearrangement or decomposition.5 In fact,
iridabenzene 1 (R ) Ph) is thermally stable to ca. 200
°C. Nevertheless, metallabenzenes have been implicated
as intermediates in several reactions, yielding cyclo-
pentadienyl-metal complexes.15 The transformations
described above are, to the best of our knowledge, the
first documented examples of iridabenzenes rearranging
to an (η5-cyclopentadienyl)iridium(I) complex.
1
The observed results and possible conversion path-
ways are summarized in Scheme 3. Addition of 4 to
Vaska’s complex furnishes the η1-vinyl species12 12
which, upon coordination of the cyclopropene π-bond,
yields predominantly benzvalene 8. The small amount
9 and 10 in the initial reaction mixture could likely come
from 13, which may or may not be directly along the
pathway to 8. An intermediate such as 13 seems
plausible, as the carbocation is stabilized by both the
R-phenyl and the â-trimethylsilyl moieties. Subsequent
ring opening results in regioselective formation of
iridabenzene 9. Similar regioselective addition of an
unsymmetrical disubstituted cyclopropene has been
observed in a metal-mediated cyclopropene ring rear-
rangement.13 Complex 9, however, is sterically desta-
bilized by having the trimethylsilyl group adjacent to
the bulky iridium fragment and thus undergoes a facile
carbene migratory insertion14 followed by dissociation
of a Ph3P ligand, leading to the formation of (cyclopen-
tadienyl)iridium complex 10.
Further work will focus on the steric and electronic
influence of substituents on the cyclopropene and on the
phosphine ligands at the Ir center. The results of these
studies will be reported shortly.
Ack n ow led gm en t. This research was supported by
the National Science Foundation (Grant No. CHE-
0075246). We thank Dr. Michael Strain for his help in
obtaining the gHMQC NMR data.
Su p p or tin g In for m a tion Ava ila ble: Text giving experi-
mental details and spectral data for 8-11 and description of
the X-ray structure of 10, including structure refinement
details and tables of atomic coordinates, thermal parameters,
bond lengths, bond angles, torsion angles, and mean planes.
This material is available free of charge via the Internet at
http://pubs.acs.org.
(11) In addition to the NMR experiments, compound 11 can be
intercepted by a second iridium fragment, which is only produced at
elevated temperatures needed to give 10. The structure of the resultant
dinuclear complex has been confirmed by X-ray crystallography: Wu,
H.-P.; Weakley, T. J . R.; Haley, M. M. Unpublished results.
(12) Schwartz, J .; Hart, D. W.; McGiffert, B. J . Am. Chem. Soc. 1974,
96, 5613-5614.
OM0201563
(15) (a) Ferede, R.; Allison, N. T. Organometallics 1983, 2, 463-
465. (b) Schrock, R. R.; Pederson, S. F.; Churchill, M. R.; Ziller, J . W.
Organometallics 1984, 3, 1574-1583. (c) Ferede, R.; Hinton, J . F.;
Korfmacher, W. A.; Freeman, J . P.; Allison, N. T. Organometallics
1985, 4, 614-616. (d) Mike, C. A.; Ferede, R.; Allison, N. T. Organo-
metallics 1988, 7, 1457-1459. (e) Yang, J .; Yin, J .; Abbound, K. A.;
J ones, W. M. Organometallics 1994, 13, 971-978. (f) Yang, J .; J ones,
W. M.; Dixon, J . K.; Allison, N. T. J . Am. Chem. Soc. 1995, 117, 9776-
9777. (g) Hughes, R. P.; Kowalski, A. S. Organometallics 1998, 17,
270-273.
(13) Cho, S. H.; Liebeskind, L. S. J . Org. Chem. 1987, 52, 2631-
2634.
(14) Relevant examples of carbene migratory insertion, inter alia:
(a) Threlkel, R. S.; Bercaw, J . E. J . Am. Chem. Soc. 1981, 103, 2650-
2659. (b) Hoover, J . F.; Stryker, J . M. J . Am. Chem. Soc. 1990, 112,
464-465. (c) Trace, R. L.; Sanchez, J .; Yang, J .; Yin, J .; J ones, W. M.
Organometallics 1992, 11, 1440-1442. (d) Casty, G. L.; Stryker, J . M.
Organometallics 1997, 16, 3083-3085.