Chemical hermaphroditism: The potential of the Cr(CO)3 moiety to stabilize transition states and intermediates with anionic, cationic, or radical character at the benzylic position

Article abstract of DOI:10.1021/ja983934k

It is known that both benzylic cations and anions are stabilized by Cr(CO)3 complexation. This unusual characteristic of chromium arenes has been the subject of many synthetic, spectroscopic, and physical organic studies over the last four decades. The ef

Full text of DOI:10.1021/ja983934k

3596
J. Am. Chem. Soc. 1999, 121, 3596-3606
Chemical Hermaphroditism: The Potential of the Cr(CO)3 Moiety To
Stabilize Transition States and Intermediates with Anionic, Cationic,
or Radical Character at the Benzylic Position
Craig A. Merlic,* Joseph C. Walsh, Dean J. Tantillo, and K. N. Houk*
Contribution from the Department of Chemistry and Biochemistry, UniVersity of California, Los Angeles,
405 Hilgard AVenue, Los Angeles, California 90095-1569
ReceiVed NoVember 12, 1998
Abstract: It is known that both benzylic cations and anions are stabilized by Cr(CO)3 complexation. This
unusual characteristic of chromium arenes has been the subject of many synthetic, spectroscopic, and physical
organic studies over the last four decades. The effect of Cr(CO)3 on benzylic radicals has received comparatively
little attention, however. In this report, cyclopropylcarbinyl anions, cations, and radicals substituted with both
phenyl and Cr(CO)3-phenyl groups are shown to rearrange via ring-opening to produce, selectively, Cr(CO)3-
stabilized benzylic anions, cations, and radicals, implying that the Cr(CO)3 moiety is capable of stabilizing
transition states with ionic or radical character at the benzylic position. The highest selectivity (>99:<1) was
observed in the anionic reaction, slightly lesser selectivity (95:5) was observed in the cationic reaction, and
only modest selectivity (2.5:1) was observed in the radical reaction. A parallel trend in ground-state stabilities
of Cr(CO)3-complexed benzyl anion, cation, and radical is predicted by density functional theory calculations.
These calculations reveal that considerable structural distortions of both benzyl anion and cation occur upon
complexation, but that little distortion occurs for benzyl radical. The connections between Cr(CO)3 complexation
and the stability of the complexed species are explained in terms of interactions between frontier molecular
orbitals of the Cr(CO)3 and benzyl fragments.
Introduction
Despite sustained interest in Cr(CO)3-complexed benzylic
ions, little progress has been made in elucidating the effect of
Cr(CO)3 on the corresponding radical species. Several reports
mention the possibility of interactions between chromium and
benzylic radicals, yet no conclusive evidence has been offered
concerning the stability of such species.5 One report does suggest
that Cr(CO)3 complexation slightly destabilizes an intermediate
with radical character at the benzylic position arising from a
trimethylenemethane structure.5i
The detailed nature of the interactions between the Cr(CO)3
moiety and complexed benzyl reactive intermediates remains
to be firmly established. The key issues are how the free and
chromium-complexed species differ in structure and whether
the Cr(CO)3 group interacts directly with the benzylic position
of each reactive intermediate. Are the chromium complexes best
represented as a Cr(CO)3 group complexed to a phenyl ligand
with a neighboring p-orbital (Scheme 1, structure A) or as a
Cr(CO)3 group complexed to a pentadienyl moiety with or
without complexation to an exocyclic alkene (Scheme 1,
structures B and C)? For the cation, the chromium may interact
directly with the benzylic position and still maintain an 18-
electron configuration at the metal, but maintenance of an 18-
The Cr(CO)3 moiety has been described as “hermaphroditic” 1
because of its ability to stabilize both benzylic cations and
anions. It has been shown that Cr(CO)3 complexes of benzylic
halides and alcohols solvolyze approximately 103-105 times
faster than their noncomplexed counterparts.2 pKR values
+
suggest that the Cr(CO)3-complexed benzyl cation is nearly as
stable as the benzhydryl cation and at least 5 pK units more
stable than free benzyl cation.2d However, the Cr(CO)3 moiety
can also stabilize complexed anions. The pKas of benzylic acids,
anilines, and phenols are known to be lowered by several pKa
units upon complexation to Cr(CO)3.3 These large pKa differ-
ences have been attributed to the net electron-withdrawing ability
of the Cr(CO)3 moiety. Since the ability of the Cr(CO)3 moiety
to stabilize conjugated ions is well-known and plays an
important role in the formation of new bonds,4 we were curious
about the extent to which Cr(CO)3 can interact with and stabilize
neighboring radicals.
(1) Jaouen, G.; Top, S.; McGlinchey, M. J. J. Organomet. Chem. 1980,
195, C5-C8 and references therein.
(2) (a) Ceccon, A. J. Organomet. Chem. 1971, 29, C19-C21. (b) Ceccon,
A.; Venzo, A. J. Organomet. Chem. 1978, 162, 311-321. (c) Seyferth, D.;
Merola, J. S.; Eschbach, C. S. J. Am. Chem. Soc. 1978, 100, 4124-4131.
(d) Wells, D.; Trahanovsky, W. S. J. Am. Chem. Soc. 1969, 91, 5870-
5872. (e) Holmes, J. D.; Jones, D. A. K.; Pettit, R. J. Organomet. Chem.
1965, 4, 324-331.
(5) (a) Taniguchi, N.; Uemura, M. Tetrahedron 1998, 54, 12775-12788.
(b) Taniguchi, N.; Uemura, M. Tetrahedron Lett. 1998, 39, 5385-5388.
(c) Merlic C. A.; Walsh, J. C. Tetrahedron Lett. 1998, 39, 2083-2086. (d)
Salazar, K. L.; Khan, M. A.; Nicholas, K. M. J. Am. Chem. Soc. 1997,
119, 9053-9054. (e) Taniguchi, N.; Uemura, M. Tetrahedron Lett. 1997,
38, 7199-7202. (f) Taniguchi, N.; Kaneta, N.; Uemura, M. J. Org. Chem.
1996, 61, 6088-6089. (g) Schmalz, H. G.; Siegel, S.; Schwartz, A.
Tetrahedron Lett. 1996, 37, 2947-2950. (h) Schmalz, H.-G.; Siegel, S.;
Bats, J. Angew. Chem., Int. Ed. Engl. 1995, 34, 2383-2385. (i) Creary,
X.; Mehrshiekh-Mohammadi, M. E.; McDonald, S. J. Org. Chem. 1989,
54, 2904-2910.
(3) (a) Wu, A.; Biehl, E. R.; Reeves, P. C. J. Chem. Soc., Perkin Trans.
2 1972, 449-451. (b) Ashraf, M.; Jackson, W. R. J. Chem. Soc., Perkin
Trans. 2 1972, 103-106. (c) Nichols, B.; Whiting, M. C. J. Chem. Soc.
1959, 551-556. (d) Fischer, E. O.; Ofele, H.; Essler, H.; Frohlich, W.;
Mortensen, J. P.; Semmlinger, W. Chem. Ber. 1958, 91, 2763-2772.
(4) See Davies, S. G.; Donohoe, T. J. Synlett 1993, 323-332 and
references therein.
10.1021/ja983934k CCC: $18.00 © 1999 American Chemical Society
Published on Web 04/02/1999

Chemical Hermaphroditism in the Cr(CO)3 Moiety
J. Am. Chem. Soc., Vol. 121, No. 15, 1999 3597
Scheme 1
carbon bond formation by radical processes.5 Although steric
effects of the transition metal moiety are undoubtedly influenc-
ing the stereochemical outcome of these radical reactions,
electronic interactions between the metal and the radical
intermediates may also contribute. This issue has not been
adequately addressed for arene or other complexes.
We now report experiments and theoretical calculations that
demonstrate the ability of Cr(CO)3 to stabilize both transition
states and intermediates having ionic or radical character at the
benzylic position. The origins of the “hermaphroditic” nature
of the Cr(CO)3 group are discussed.
Results
Competition Experiments. Previous work from one of our
laboratories involved attempts to observe a stabilized chromium
arene radical.12 These attempts proved unsuccessful due to the
sensitive nature of these radicals, so a competition experiment
was designed to help elucidate the nature of such species. We
envisioned formation of a cyclopropylcarbinyl radical with two
different aryl substituents at the 2- and 3-positions of the
cyclopropane (Scheme 2). The ratio of homoallylic radicals E
electron configuration in the anion would preclude such an
interaction. In the case of the radical, the chromium may remain
in an 18-electron configuration by interacting only with the
phenyl ring (in A), become 17-electron by interacting with only
the pentadienyl moiety (in B), or become 19-electron by
interacting with both the pentadienyl moiety and the exocyclic
alkene (in C). The best representation of each reactive inter-
mediate remains to be determined.
Many synthetic endeavors have taken advantage of Cr(CO)3-
complexed arenes due to the ability of Cr(CO)3 to facilitate
formation of anionic and cationic benzylic intermediates and
to direct the stereoselective formation of new bonds at the
benzylic position.4,6,7 Chromium arene complexes have also been
prepared in enantiomerically pure form, allowing for the
enantioselective synthesis of tetrahydronaphthalene-, tetrahy-
droisoquinoline-, tetrahydrobenzazepine-, and diphenylmethane-
derived natural and biologically active products.4,6a,c,7
Scheme 2
In addition, transition metal-templated radical reactions are
providing new opportunities for radical chemistry and stereo-
selective synthesis.5,8-11 Examples have been reported which
demonstrate the remarkable ability of certain transition metals
to provide complete control of stereoselectivity in carbon-
and F, produced by rearrangement of the cyclopropylcarbinyl
radical, would be determined by the relative abilities of each
arene to stabilize radical character at the benzylic position.13 It
has been demonstrated that cyclopropylcarbinyl ring-openings
can show remarkable selectivities in the formation of regioiso-
meric ring-opened products despite early transition states.13d
Enhanced stabilization of radical character by Cr(CO)3 com-
plexation would selectively weaken bond a of D (see Scheme
2) and increase the rate of formation of product E. Quenching
of radicals E and F would then lead to 3,4-diaryl-1-butene
products. The beauty of this internal competition is that a single
experiment will yield the kinetic ratio of ring-opened radicals
E and F.
(6) Recent examples include: (a) Corey, E. J.; Helal, C. J. Tetrahedron
Lett. 1996, 4837-4840. (b) Uemura, M.; Nishimura, H.; Minami, T.;
Hayashi, Y. J. Am. Chem. Soc. 1991, 113, 5402-5410. (c) Coote, S. J.;
Davies, S. G.; Middlemiss, D.; Naylor, A. Tetrahedron: Asymmetry 1990,
1, 33-56. (d) Uemura, M.; Nishimura, H.; Hayashi, Y. Tetrahedron Lett.
1990, 31, 2319-2322.
(7) Recent examples include: (a) Geller, T.; Schmalz, H.-G.; Bats, J.
Tetrahedron Lett. 1998, 39, 1537-40. (b) Majdalani, A.; Schmalz, H. G.
Synlett 1997, 1303-1305. (c) Davies, S. G.; Loveridge, T.; Clough, M.
Synlett 1997, 66-68. (d) Schmalz, H.-G.; Arnold, M.; Hollander, J.; Bats,
J. Angew. Chem., Int. Ed. Engl. 1994, 33, 109-111. (e) Schmalz, H.-G.;
Hollander, J.; Arnold, M.; Durner, G. Tetrahedron Lett. 1993, 6259-6262.
(f) Davies, S. G.; Goodfellow, C. L.; Peach, J.; Waller, A. J. Chem. Soc.,
Perkin Trans. 1 1991, 1009-1017. (g) Tanaka, T.; Mikamiyama, H.; Maeda,
K.; Iwata, C.; In, Y.; Ishida, T. J. Org. Chem. 1998, 63, 9782-9793. (h)
Volk, T.; Bernicke, D.; Bats, J. W.; Schmalz, H.-G. Eur. J. Inorg. Chem.
1998, 1883-1905.
(8) Cr carbene complexes: (a) Merlic, C. A.; Xu, D.; Nguyen, M. C.;
Truong, V. Tetrahedron Lett. 1993, 34, 227-230. (b) Merlic, C. A.; Xu,
D. J. Am. Chem. Soc. 1991, 113, 9855-9856.
(9) Co alkyne complexes: (a) Melikyan, G. G.; Khan, M. A.; Nicholas,
K. M. Organometallics 1995, 14, 2170-2172. (b) Melikyan, G. G.;
Vostrowsky, O.; Bauer, W.; Bestmann, H. J.; Khan, M.; Nicholas, K. M.
J. Org. Chem. 1994, 59, 222-229. (c) Melikyan, G. G.; Combs, R. C.;
Lamirand, J.; Khan, M.; Nicholas, K. M. Tetrahedron Lett. 1994, 35, 363-
366. (d) Melikyan, G. G.; Vostrowsky, O.; Bauer, W.; Bestmann, H. J. J.
Organomet. Chem. 1992, 423, C24-C27.
(10) Fe diene complexes: (a) Bendorf, H. D. Ph.D. Thesis, University
of California, Los Angeles, 1994. (b) Hwang, W.-S.; Liao, R.-L.; Horng,
Y.-L.; Ong, C. W. Polyhedron 1989, 8, 479-482. (c) Pearson, A. J.;
Connelly, N. G.; Kitchen, M. D.; Stansfield, R. F. D.; Whiting, S. M.;
Woodward, P. J. Organomet. Chem. 1978, 155, C34-C36. (d) Mahler, J.
E.; Gibson, D. H.; Pettit, R. J. Am. Chem. Soc. 1963, 85, 3959-3963.
(11) Co and Fe metallocenes: (a) Herberich, G. E.; Carstensen, T.; Klein,
W.; Schmidt, M. U. Organometallics 1993, 12, 1439-1441. (b) Geiger,
W. E.; Gennett, T.; Lane, G. A.; Salzer, A.; Rheingold, A. L. Organome-
tallics 1986, 5, 1352-1359.
In analogy to the cyclopropylcarbinyl radical rearrangement,
cyclopropylcarbinyl anions and cations also rearrange to homal-
lylic anions and cations. While rearrangements in anionic
(12) Mercier, A. H. M.S. Thesis, University of California, Los Angeles,
1995.
(13) Recent reviews, experiments, and calculations on cylopropylcarbinyl
radical reactions include: (a) Martinez, F. N.; Schlegel, H. B.; Newcomb,
M. J. Org. Chem. 1998, 63, 3618-3623. (b) Engel, P. S.; He, S.-L.; Banks,
J. T.; Ingold, K. U.; Lusztyk, J. J. Org. Chem. 1997, 62, 1210-1214. (c)
Martinez, F. N.; Schlegel, H. B.; Newcomb, M. J. Org. Chem. 1996, 61,
8547-8550. (d) Le Tadic-Biadatti, M. H.; Newcomb, M. J. Chem. Soc.,
Perkin Trans. 2 1996, 1467-1473. (e) Martinesker, A. A.; Johnson, C. C.;
Horner, J. H.; Newcomb, M. J. Am. Chem. Soc. 1994, 116, 9174-9181.
(f) Beckwith, A. L. J.; Bowry, V. W. J. Am. Chem. Soc. 1994, 116, 2710-
2716. (g) Newcomb, M. Tetrahedron 1993, 49, 1151-1176. (h) Nonhebel,
D. C. Chem. Soc. ReV. 1993, 22, 347-359. (i) Newcomb, M. Tetrahedron
1993, 49, 1151-1176. (j) Newcomb, M.; Johnson, C. C.; Manek, M. B.;
Varick, T. R. J. Am. Chem. Soc. 1992, 114, 10915-10921. (k) Tanko, J.
M.; Drumright, R. E. J. Am. Chem. Soc. 1990, 112, 5362-5363.

3598 J. Am. Chem. Soc., Vol. 121, No. 15, 1999
Merlic et al.
Scheme 3
systems are relatively facile,14 those in cationic systems are more
complex. In fact, the structure of cyclopropylcarbinyl cations
and their potential equilibria with not only homoallylic cations
but also cyclobutyl (or nonclassical bicyclobutonium) cations
has received much attention over the past half century.15
However, aryl substitution at the 2- and 3-positions of the
cyclopropane is expected to greatly stabilize the homallylic
cation resulting from ring-opening relative to the cyclopropyl-
carbinyl cation precursor. In addition, the pseudosymmetry of
the cyclopropylcarbinyl ions and radicals in our design should
result in product distributions upon rearrangement that reflect
the relative influence of phenyl and Cr(CO)3-phenyl on ring-
opening transition states. The fact that the ionic and radical
reactions utilize identical substrates and lead to identical products
upon quenching simplifies characterization and allows for facile
comparison of all three reaction types.
order to explore the rearrangement selectivities. When 4 was
reacted with t-BuLi at -78 °C in THF and then quenched with
dilute HCl, alkene 5 was isolated as a single isomer arising from
rearrangement to the chromium-stabilized benzylic anion; no
competing product 6 was detected (Scheme 4). When cyclo-
Scheme 4
Synthesis of the rearrangement substrate 4 was straightfor-
ward (Scheme 3). Cyclopropanation of cis-stilbene with ethyl
diazoacetate utilizing a copper(I) triisopropyl phosphite catalyst16
in refluxing cyclohexane afforded the 2,3-cis-diphenyl-1-trans-
carboethoxycyclopropane (1) in 42% yield after recrystallization.
The ester was reduced with lithium aluminum hydride to give
the cyclopropane alcohol 2. Metalation of the alcohol with Cr-
(CO)3(MeCN)317 smoothly afforded the cyclopropane chromium
arene complex 3 in 98% yield. Interestingly, the ortho and meta
positions of the metalated ring are diastereotopic, so five separate
propyl iodide 4 was treated with AgBF4 and Et3SiH in
dichloromethane, alkenes 5 and 6 were formed in a 95:5 ratio,
again favoring ring-opening on the chromium-bound side, but
via the cation. When iodide 4 was subjected to radical-forming
conditions using SmI2/HMPA18 and t-BuOH in THF, products
5 and 6 were produced in a ratio of 2.5:1, only modestly favoring
the product arising from the benzylic radical of the Cr(CO)3-
complexed arene.
1
signals for those aryl protons are observed in the H NMR.
Analogously, six resonances are observed for the complexed
arene ring in the 13C NMR. Conversion of the alcohol to the
corresponding iodide was accomplished in 84% yield using
freshly recrystallized Ph3P‚I2 and imidazole in dichloromethane;
other methods resulted in low yields of mixtures of products
and demetalated material.
With cyclopropyl iodide chromium arene adduct 4 in hand,
we then tested it under ionic and radical reaction conditions in
The excellent selectivity of the anionic ring-opening is in
agreement with previous reports on the strong stabilizing effect
of Cr(CO)3 on complexed anions.3 In examining the ring-
opening of the cyclopropyl carbocation, we were somewhat
surprised by the 95:5 ratio of ring-opened products. Indeed, the
major product arises from enhanced stabilization of cationic
character by the Cr(CO)3 moiety, but it appears that this
stabilization is not at great as that in the anionic case. More
surprising, however, was the modest selectivity exhibited in the
radical rearrangement.
(14) A recent application of cyclopropylmethyllithiums: (a) Charette,
A. B.; Naud, J. Tetrahedron Lett. 1998, 39, 7259-7262. Seminal work on
cyclopropylmethyllithiums: (b) Lansbury, P. T.; Pattison, V. A.; Clement,
W.; Sidler, J. D. J. Am. Chem. Soc. 1964, 86, 2247-2251. Seminal work
on cyclopropylmethyl Grignard reagents: (c) Maercker, A.; Roberts, J. D.
J. Am. Chem. Soc. 1966, 88, 1742-1759. (d) Silver, M. S.; Shafer, P. R.;
Ru¨chardt, C.; Roberts, J. D. J. Am. Chem. Soc. 1960, 82, 2646-2647. (e)
Roberts, J. D.; Mazur, R. H. J. Am. Chem. Soc. 1951, 73, 2509-2520.
(15) Recent reviews include: (a) Olah, G. A.; Laali, K. K.; Wang, Q.;
Prakash, G. K. S. Onium Ions; John Wiley & Sons: New York, 1998; 359-
362. (b) Olah, G. A.; Reddy, V. P.; Rakash, G. K. S. Chem. ReV. 1992, 92,
69-95. Seminal work on cyclobutylcarbinyl cations: (c) Mazur, R. H.;
White, W. N.; Semenow, D. A.; Lee, C. C.; Silver, M. S.; Roberts, J. D. J.
Am. Chem. Soc. 1959, 81, 4390-4398. See also ref 14e.
It is known that the 2-phenylcyclopropylcarbinyl radical
rearranges with a rate constant of 3 × 1011 s-1 13j
,
which is 103
(18) (a) Hasegawa, E.; Curran, D. P. Tetrahedron Lett. 1993, 34, 1717-
1720. (b) Totleben, M. J.; Curran, D. P.; Wipf, P. J. Org. Chem. 1992, 57,
1740-1744. (c) Inanaga, J.; Ishikawa, M.; Yamiguchi, M. Chem. Lett. 1987,
1485-1486. (d) Girard, P.; Namy, J. L.; Kagan, H. B. J. Am. Chem. Soc.
1980, 102, 2693-2698.
(16) Nishizawa, Y. Bull. Chem. Soc. Jpn. 1961, 34, 1170-1178.
(17) (a) Ross, B. L.; Jeanette, G. G.; Ritchey, W. M.; Kaesz, H. D. Inorg.
Chem. 1963, 2, 1023-1030. (b) Tate, D. P.; Knipple, W. R.; Augl, J. M.
Inorg. Chem. 1962, 1, 433-434.

Chemical Hermaphroditism in the Cr(CO)3 Moiety
J. Am. Chem. Soc., Vol. 121, No. 15, 1999 3599
times faster than the rearrangement of the parent cyclopropyl-
carbinyl radical.19 This rate enhancement is due to stabilization
of radical character in the transition state by the conjugated
aromatic ring. Because there is an enhancement in the ring-
opening rate of D leading to the benzylic radical that is
conjugated to the Cr(CO)3-complexed arene, it can be concluded
that the Cr(CO)3 group does, indeed, stabilize radical character
in the transition state. However, the ring-opening of cyclopro-
pylcarbinyl radical D does not demonstrate the same selectivity
as seen in the corresponding ionic rearrangements. It can
therefore be concluded that, although the Cr(CO)3 moiety
stabilizes the benzylic radical, the extent of stabilization is not
nearly as great as that in the anionic and cationic cases.
Computational studies were undertaken to provide a better
understanding of the nature of the interactions between the metal
fragment and the benzylic reactive intermediates.
Computed Structures of Free and Cr(CO)3-Complexed
Benzyl Species. (i) Cation. A wealth of implicit structural
information on Cr(CO)3-complexed benzylic cations is available.
It was observed in the 1960s that Cr(CO)3 complexation of a
benzylic halide or alcohol led to increased rates of solvolysis.2e,20
It was also observed that such substitution reactions proceed
with retention of stereochemistry at the benzylic position,4 due
to stereocontrol in the three fundamental reaction steps. First,
ionization proceeds with the leaving group anti to chromium.
Second, the Cr(CO)3 group protects the back side of the cation
and increases the barrier to rotation around the exocyclic
carbon-carbon bond. Third, attack on the benzylic cation occurs
anti to the bulky Cr(CO)3 group, preserving the stereochemistry
of the starting material. 1H and 13C NMR experiments demon-
strated hindered rotation about the exocyclic carbon-carbon
bond in 1-p-tolylethyl- and di-p-tolylmethyltricarbonylchromium
cations.21a,b Hindered rotation can arise from an increase in
double bond character of the exocyclic bond or direct interaction
between the chromium and the benzylic carbon. Such an
interaction could also provide anchimeric assistance to departure
of benzylic leaving groups. The involvement of similar factors
has also been suggested in reactions of ferrocenyl cations.22
Attempts to isolate Cr(CO)3 complexes of simple benzylic
cations led to decomposition via loss of carbon monoxide.2e
This observation suggests that the Cr-CO bonds were made
more labile by complexationsan effect that should manifest
itself spectroscopically and in increased Cr-CO and decreased
C-O bond lengths relative to those in normal Cr(CO)3-arene
complexes. In fact, increased CO stretching frequencies in IR
experiments2c and altered chemical shifts in NMR experiments2c,21
have been observed for Cr(CO)3-complexed cations relative to
those of related neutral species, indicating decreased back-
donation of chromium electron density to the CO ligands and
a corresponding shift in positive charge from the benzyl ligand
to the chromium.
differences between the free and complexed structures reflect
the changes in physical properties observed upon complexation.
An increase in the Cr-CO and a decrease in the C-O bond
lengths in the complexed cation relative to those in the Cr(CO)3-
toluene complex are predicted, consistent with the spectroscopic
information described above and the tendency of cationic
complexes to decompose through loss of CO. The expected
redistribution of positive charge from the benzyl ligand to the
Cr(CO)3 moiety is reflected in computed Mulliken atomic
charges (B3LYP/LANL2DZ level, not shown) and can be seen
clearly in the electrostatic potential surfaces shown in Figure
2.
The structure of the cationic benzyl ligand in the Cr(CO)3
complex is also highly distorted from the planar structure of
the free ligand. The complexed species can be thought of as a
pentadienyl cation fragment and an exocyclic double bond,
which serve as 4- and 2-electron ligands, respectively, for the
chromium. As such, the exocyclic carbon-carbon bond is
inclined toward the chromium (Figure 1b, i and ii), and the
chromium is not centered below the centroid of the phenyl ring
(Figure 1b, iii) as it is in neutral Cr(CO)3-arene complexes such
as Cr(CO)3-toluene (Figure 1a, iii). An earlier study utilizing
semiempirical INDO calculations noted a similar distortion but
suggested that specific bond formation between chromium and
the benzylic center was not involved.23
(ii) Anion. Less experimental information on the structure
of Cr(CO)3-complexed benzylic anions is available. It was
shown around 1970 that base-catalyzed hydrogen-deuterium
exchange of benzylic protons in Cr(CO)3-indan complexes was
stereoselective; only the benzylic protons anti to the metal were
exchanged (in a solution of 0.5 M potassium tert-butoxide in
DMSO).24 It was proposed that this selectivity arose from
anchimeric assistance by the chromium, although a steric effect
could not be excluded.24 IR, 1H NMR, and 13C NMR spectros-
copy of the bis(tricarbonylchromium)diphenylmethyl and free
diphenylmethyl carbanions indicated that the benzylic center
was most likely sp2-hybridized and that electron density from
the benzyl ligand was redistributed through the chromium onto
the carbonyl ligands, but also that there was no direct interaction
between the chromium and the benzylic carbon.1
Geometry optimizations on the free and Cr(CO)3-complexed
benzyl anions (B3LYP/LANL2DZ, see Methods section) were
performed in order to rationalize these experimental results, and
the fully optimized structures are shown in Figure 1c. A decrease
in the Cr-CO and an increase in the C-O bond lengths in the
complexed anion relative to those in the toluene complex are
predicted, consistent with the spectroscopic observations de-
scribed above. In addition, negative charge is clearly redistrib-
uted from the benzyl framework onto the carbonyl ligands, as
shown in the electrostatic potential surfaces displayed in Figure
2 (and this is reflected in Mulliken atomic charges computed
at the B3LYP/LANL2DZ level, not shown). A significant
increase in bond order for the exocyclic carbon-carbon bond
and an increase in bond length alternation/localization in the
complexed anion relative to those in the free system are also
predicted.
Geometry optimizations on the free and Cr(CO)3-complexed
benzyl cations (B3LYP/LANL2DZ, see Methods section) were
performed in an effort to rationalize these experimental results.
Fully optimized structures are shown in Figure 1b. The
(19) Beckwith, A. L.; Moad, G. J. Chem. Soc., Perkin Trans. 2 1980,
1473-1482.
(20) Gubin, S. P.; Khandkarova, V. S.; Kreindlin, A. Z. J. Organomet.
Chem. 1974, 64, 229-238.
(21) (a) Acampora, M.; Ceccon, A.; Dal Farra, M.; Giacometti, G.;
Rigatti, G. J. Chem. Soc., Perkin Trans. 2 1977, 483-486. (b) Acampora,
M.; Ceccon, A.; Dal Farra, M.; Giacometti, G. J. Chem. Soc., Chem.
Commun. 1975, 871-872. (c) Olah, G. A.; Yu, S. H. J. Org. Chem. 1976,
41, 1694-1697.
(22) (a) Pettit, R.; Haynes, L. W. In Carbonium Ions; Olah, G. A.,
Schleyer, P. v. R., Eds.; Wiley-Interscience: New York, 1976; Vol. 5, pp
2263-2302. (b) Cais, M. Organomet. Chem. ReV. 1966, 1, 435-454.
The anionic benzyl ligand is distorted toward a pentadienyl
moiety plus an exocyclic double bond to an even larger extent
than in the case of the cation. Since the chromium is most stable
as an 18-electron species, and a pentadienyl anion is a 6-electron
(23) Clack, Q. W.; Kane-Maguire, L. A. P. J. Organomet. Chem. 1978,
145, 201-206.
(24) (a) Trahanovsky, W. S.; Card, R. J. J. Am. Chem. Soc. 1972, 94,
2897-2898. (b) Gracey, D. E. F.; Jackson, W. R.; McMullen, C. H.;
Thompson, N. J. Chem. Soc. B 1969, 1197-1203.

3600 J. Am. Chem. Soc., Vol. 121, No. 15, 1999
Merlic et al.
Figure 1. (U)B3LYP/LANL2DZ-optimized geometries of free and Cr(CO)3-complexed toluene, benzyl cation, benzyl anion, and benzyl radical.
Selected bond lengths are in angstroms. Values in parentheses are from the X-ray structure of Cr(CO)3-toluene.26,39 (i) Side view of preferred
rotamers of Cr(CO)3 complexes. (ii) Elevated view of complexes in (i). (iii) Top view of complexes in (i). (iv) Elevated view of disfavored rotamers
of Cr(CO)3 complexes. (v) Free toluene and benzyl species.
ligand, the exocyclic double bond tilts away from the chromium
(Figure 1c, i and ii), and the chromium moves closer to the
pentadienyl fragment (Figure 1c, iii). Thus, the Cr(CO)3
fragment again moves away from the centroid of the phenyl
ring, but in a direction opposite to that observed in the cationic
system. As predicted by experiment,1 there is no direct interac-
tion between the chromium and the benzylic carbon. The relative
increase in kinetic acidity of the benzylic protons anti to the
Cr(CO)3 group is, therefore, likely a result of steric blocking
of the endo face of the exocyclic double bond by the CO ligands
of the Cr(CO)3 fragment.
that reactions of such radicals occur anti to the Cr(CO)3 moiety.5
To obtain detailed structural information on such species,
geometry optimizations on the free and Cr(CO)3-complexed
benzyl radicals (UB3LYP/LANL2DZ, see Methods section)
were performed. Fully optimized structures are shown in Figure
1d. Unlike the cationic and anionic species, the benzyl radical
experiences little distortion upon complexation. The chromium
remains centered beneath the ring, which remains nearly planar
(Figure 1d, i-iii). The benzylic methylene group does bend
slightly away from the chromium, suggesting that the metal
center does not interact with all seven electrons of the benzyl
ligand (which would make it a 19-electron species) but rather
interacts solely with the phenyl moiety as a 6-electron ligand.
(iii) Radical. Very little is known about the structure and
reactivity of Cr(CO)3-complexed benzylic radicals other than

Chemical Hermaphroditism in the Cr(CO)3 Moiety
J. Am. Chem. Soc., Vol. 121, No. 15, 1999 3601
Figure 2. Electrostaticpotential surfaces of free and Cr(CO)3-complexed benzyl ions and constant spin density surfaces of free and Cr(CO)3-
complexed benzyl radicals. The electrostatic potential is graphed over the ranges shown, in units of kcal/mol. See Methods section for details.
As shown in Figure 2, the unpaired spin density in the Cr-
(CO)3-complexed benzyl radical remains mostly on the benzyl
ligand. Computed Mulliken atomic spin densities (UB3LYP/
LANL2DZ) suggest that the spin density that resides on the
benzylic carbon in the free species (+0.81) decreases only
slightly upon complexation (+0.73 for the benzylic carbon,
+0.31 for the chromium). Although some spin density is
delocalized onto the chromium, the complexed ligand is
predicted to behave similarly to an uncomplexed benzyl radical
from an electronic standpoint.
(iv) Rotamers. Rotation of the Cr(CO)3 moiety relative to
the arene ligand in chromium arene complexes is known to
occur. As previously explained on the basis of orbital consid-
erations,25 complexes with electron-donating substituents on the
aromatic ring tend to prefer syn-eclipsed conformations (in
which a carbonyl ligand eclipses the ring-substituent bond when
viewed down the axis between the metal and the ring center),
while those with electron-withdrawing substituents tend to prefer
anti-eclipsed conformations (in which a carbonyl ligand is anti
to the ring-substituent bond). In the case of toluene, for example,
the syn-eclipsed conformation is preferred so that vacant metal
hybrid orbitals can interact with the ortho and para positions of
the aromatic ring positions that bear larger orbital coefficients
than the meta and ipso positions, respectively, in the toluene
HOMO (see Figure 3).
kcal/mol less stable than the fully optimized structure (Figure
1a, iv), suggesting that the rotational potential is relatively shal-
low for the toluene complex, consistent with the fact that a meth-
yl group is only slightly electron-donating. Fully optimized
structures ((U)B3LYP/LANL2DZ) for all syn- and anti-eclipsed
rotamers of the Cr(CO)3-complexed benzyl cation, anion, and
radical are shown in Figure 1b-d (iv). The following confor-
mational preferences are predicted: the cation prefers to be anti-
eclipsed by 11.7 kcal/mol,27 the anion prefers to be syn-eclipsed
by 6.3 kcal/mol, and the radical prefers to be anti-eclipsed by
2.8 kcal/mol. These preferences are predicted by orbital analysis,
since these conformations involve maximum bonding between
metal orbitals and those of the benzyl fragments (see below).
The considerable difference in the magnitude of the energetic
preferences of the cation and radical is most likely due to the
direct interaction of chromium with the exocyclic bond of the
benzyl cation. Since the benzylic methylene is brought close to
the chromium in this case, unfavorable steric interactions with
the carbonyl groups in the syn-eclipsed rotamer are amplified,
as are favorable orbital interactions in the anti-eclipsed rotamer
(see below).
Computed Stabilization Energies of Cr(CO)3 Complexes.
The gas-phase hydride affinity, proton affinity, and bond
dissociation energy of toluene are lower than those of methane
by approximately 80, 40, and 20 kcal/mol, respectively.28 It is
also known that the Cr(CO)3 moiety can provide additional
stabilization in the case of complexed benzylic anions, cations,
and related transition states.1-4 Our experiments support these
observations and provide evidence that the Cr(CO)3 group can
also provide minor stabilization of a benzylic radical. Although
the “hermaphroditic” nature of Cr(CO)3-complexed arenes was
revealed by the early experiments, the details of how this group
provides stabilization were not clearly defined. In an effort to
Various rotamers of the Cr(CO)3-benzyl species described
above were examined, and the expected preferences were ob-
served. First, unconstrained optimizations (B3LYP/LANL2DZ)
on the Cr(CO)3-toluene complex led to a slightly distorted syn-
eclipsed conformer (Figure 1a, i-iii). This conformer is similar
to that observed in the X-ray structure of Cr(CO)3-toluene.26
A
pure syn-eclipsed conformer is most likely disfavored by small
steric interactions with the methyl group. Constraining the com-
plex to be anti-eclipsed led to a structure which was only 0.6
(27) This is in contrast to earlier semiempirical INDO calculations that
suggested that the relative energies of the syn- and anti-eclipsed rotamers
of the cation are similar and that a staggered conformation is actually
preferred by approximately 6 kcal/mol.23
(25) Albright, T. A. Tetrahedron 1982, 38, 1339-1388.
(26) van Meurs, F.; van Koningsveld, H. J. Organomet. Chem. 1977,
131, 423-428.

3602 J. Am. Chem. Soc., Vol. 121, No. 15, 1999
Merlic et al.
Figure 3. (a) Hybrid fragment orbitals of Cr(CO)3. (b) π molecular orbitals of toluene. Frontier orbitals that are of appropriate symmetry to
interact, and whose interactions constitute the key features of chromium-arene bonding, are connected by dashed lines. (c) π molecular orbitals of
benzyl. The asterisk indicates that this orbital may contain zero, one, or two electrons, depending on the reactive intermediate in question. “S” and
“A” indicate whether the orbitals are symmetric or antisymmetric with respect to the plane of symmetry of the complexes.
Scheme 5
assess the relative energetic influence of the Cr(CO)3 moiety
on different benzylic reactive intermediates, the enthalpies for
several homodesmotic equations (these represent the enthalpy
differences between the gas phase hydride affinities of free and
complexed benzyl cation, proton affinities of free and complexed
benzyl anions, and bond dissociation energies of free and
complexed benzyl radicals; see Scheme 5 and Table 1) were
calculated using a hybrid HF-DFT method. (U)B3LYP/
LANL2DZ-optimized geometries were used, and single-point
energies were calculated at the (U)B3LYP/LANL2DZ and (U)-
B3LYP/DZVP2+ levels (see Table 1 and Methods section).
These calculations predict, in accord with experiment, that both
the anion and cation should be strongly stabilized by the Cr-
(CO)3 moiety (Table 1) but that the Cr(CO)3 group should have
Table 1. ∆E (kcal/mol) for the Homodesmotic Reactions Shown
in Scheme 5 Based on (U)B3LYP/LANL2DZ Geometries and
Single-Point Energies at the Levels Shown Below
little effect on the stability of the radical.
The hydride affinity of the benzyl cation is reduced by
approximately 4-12 kcal/mol upon Cr(CO)3 complexation,
while the proton affinity of the benzyl anion is decreased by
approximately 30-40 kcal/mol (Table 1). The lesser stabiliza-
tion of the cation is reflected experimentally in the fact that a
significant amount (5%) of the competing product is detected
level of theory
cationic
anionic
radical
(U)B3LYP/LANL2DZ
(U)B3LYP/DZVP2+
-4.3
-11.7
-41.3
-34.0
1.3 (0.784, 0.780)a
1.0 (0.782, 0.777)
a S2 values for free and complexed radicals, respectively, are given
in parentheses.
(28) (a) March, J. AdVanced Organic Chemistry: Reactions, Mechanisms,
and Structure, 4th ed.; Wiley-Interscience: New York, 1992; pp 171 and
191. (b) Carey, F. A.; Sundberg, R. J. AdVanced Organic Chemistry, Part
A: Structure and Mechanisms, 3rd ed.; Plenum Press: New York, 1990; p
273. (c) Carroll, F. A. PerspectiVes on Structure and Mechanism in Organic
Chemistry; Brooks/Cole: Pacific Grove, CA, 1998; pp 268, 286, and 407.
(d) Aue, D. H.; Bowers, M. T. In Gas-Phase Ion Chemistry; Bowers, M.
T., Ed.; Academic Press: New York, 1979; Vol. 2, p 1. (e) Lias, S. G.;
Bartmess, J. E.; Liebman, J. F.; Holmes, J. L.; Levin, R. D.; Mallard, W.
G. J. Phys. Chem. Ref. Data 1988, 17, Supplement 1, 1; Gas-Phase Ion
and Neutral Thermochemistry; American Chemical Society and American
Institute of Physics for the National Bureau of Standards: New York, 1988.
(f) Hrovat, D. A.; Borden, W. T. J. Phys. Chem. 1994, 98, 10460-10464.
(g) Graul, S. T.; Squires, R. R. J. Am. Chem. Soc. 1990, 112, 2517-2529.
(h) DePuy, C. H.; Gronert, S.; Barlow, S. E.; Bierbaum, V. M.; Damrauer,
R. J. Am. Chem. Soc. 1989, 111, 1968-1973.
for the cationic ring-opening of substrate 4 described above,
while no competing product is observed for the analogous
anionic ring-opening reaction (Scheme 4).
In the radical ring-opening reaction, a 2.5:1 ratio of products
5 and 6 is observed, corresponding to a ∆∆Gq of approximately
0.5 kcal/mol. This result also agrees (within experimental and
computational error) with the small difference in bond dissocia-
tion energies (approximately 1 kcal/mol, but favoring the
uncomplexed species) predicted by the calculations for the free
and complexed radical species. Several caveats should be noted
at this point. First, the competition experiments reflect kinetic

Chemical Hermaphroditism in the Cr(CO)3 Moiety
J. Am. Chem. Soc., Vol. 121, No. 15, 1999 3603
Figure 4. Selected stabilizing orbital interactions between Cr(CO)3 and benzyl species. The position of the Cr(CO)3 fragment, deduced from
geometry optimizations, is indicated by an arrow for each case.
ratios while the calculations predict thermodynamic values;
nonetheless, the stabilization of the reactive intermediates should
be reflected in transition-state energies. Second, calculations
were performed only on the parent benzyl system; substitutions
at the benzylic position and on the aromatic ring are likely to
alter the magnitude of the Cr(CO)3-induced stabilization.2b,d
orbitals, since these orbitals are of the wrong symmetry to
interact with the benzylic p-orbital. In addition, symmetry-
allowed interactions between Cr(CO)3 orbitals and π orbitals
of toluene and benzyl are very similar, except for the new
interactions involving the benzyl cation LUMO, anion HOMO,
and radical SOMO. These interactions are shown schematically
in Figure 4 for the preferred rotamers of the benzyl complexes
and are described below. The position of the chromium in each
complex is indicated in Figure 4 on the benzyl orbitals by arrows
and is related to the degree of spatial overlap between the metal
and ligand orbitals.29
Discussion
The large stabilization of anionic and cationic, and almost
negligible stabilization of radical-like, transition states and
intermediates, demonstrated by our experiments and computa-
tions, can be rationalized in terms of changes in orbital
interactions between the benzylic species and Cr(CO)3 relative
to those in typical Cr(CO)3-complexed arenes such as Cr(CO)3-
toluene. Symmetry-allowed interactions between the π molecular
orbitals of toluene and the hybrid fragment orbitals25 of Cr-
(CO)3 are shown schematically in Figure 3. The three occupied
and three vacant hybrid orbitals of the Cr(CO)3 fragment are
shown, along with the π orbitals of toluene. All are labeled
according to their symmetry with respect to the plane of
symmetry of the complex (S, symmetric, and A, antisymmetric,
correspond to A′ and A′′ in the Cs point group). The dashed
lines represent stabilizing interactions between filled and vacant
orbitals of the two fragments. There is strong complementarity
between the HOMOs and LUMOs of the two systems. The small
syn-eclipsed preference of toluene can be attributed to the more
favorable interactions of the S HOMO and LUMO of the metal
with the S LUMO and HOMO of toluene. The methyl group
enhances the para orbital coefficient in the S HOMO and the
ipso orbital coefficient in the S LUMO of toluene, and the Cr-
(CO)3 fragment aligns to maximize interaction with these sites.
The π molecular orbitals of the benzyl ligands (Figure 3)
involve a p-orbital at the benzylic position mixing with the π
molecular orbitals of the phenyl group; doing so results in a
new frontier molecular orbital (FMO). Due to electron-electron
repulsion, this orbital is higher in energy for the anion than for
the radical and is higher in energy for the radical than for the
cation. The result is a new low-lying LUMO for the benzyl
cation, a new high-lying HOMO for the benzyl anion, and a
SOMO for the benzyl radical of intermediate energy.
In the case of the cation, the low-lying LUMO of the benzyl
fragment interacts more strongly with the occupied symmetric
hybrid metal orbitals than does the lowest unoccupied symmetric
toluene π orbital (Figure 4). Distortion of the benzyl ligand away
from planarity and shifting of the chromium away from the
center of the aromatic ring increase the spatial overlap between
2
the metal and ligand orbitals (in particular, with the dz -like metal
orbital) and leads to further stabilization. In fact, single-point
energy calculations (B3LYP/LANL2DZ and B3LYP/DZVP2+)
on a cationic species constrained to the planar geometry of the
radical complex predict that planarization of the benzyl cation
increases the energy of the complex by 9-15 kcal/mol. The
direct interaction of the chromium with the exocyclic carbon-
carbon bond accompanying ligand distortion is, therefore,
essential for stabilization.
In the case of the anion, the high-lying HOMO of the free
benzyl fragment interacts more strongly with the vacant
symmetric hybrid metal orbital shown than does the highest
occupied symmetric toluene orbital, again leading to enhanced
stabilization through a two-electron interaction (Figure 4). In
this case, distortion of the benzyl ligand away from planarity
and shifting of the chromium away from the center of the
aromatic ring also lead to improved overlap and increased
stabilization. Single-point energy calculations (B3LYP/LANL2DZ
and B3LYP/DZVP2+) on an anionic species constrained to the
planar geometry of the radical complex predict that planarization
of the benzyl anion increases the energy of the complex by
approximately 14-17 kcal/mol. Ligand distortion is clearly
essential for maximum stabilization.
Interactions of the antisymmetric benzyl orbitals with the Cr-
(CO)3 fragment are expected to be of comparable magnitude to
the analogous interactions involving the antisymmetric toluene
(29) In this treatment, we make the simplifying approximation that the
different benzyl orbitals of the cation, anion, and radical will be mixed
with identical Cr(CO)3 fragment orbitals.

3604 J. Am. Chem. Soc., Vol. 121, No. 15, 1999
Merlic et al.
In the case of the radical, the SOMO of the free benzyl
fragment interacts with the occupied and vacant symmetric
hybrid metal orbitals shown in Figure 4. Little stabilization is
expected from these odd-electron interactions due to decreased
orbital overlap and decreased matching of orbital energies
relative to the ionic cases. Distortion of the benzyl fragment to
increase overlap or to allow for additional interactions with the
benzylic center does not provide enough additional stabilization
to compensate for the energy required for distortion. Single-
point energy calculations (UB3LYP/LANL2DZ and UB3LYP/
DZVP2+) on radical species constrained to the nonplanar
geometries of either the cationic or anionic complexes predict
that distortion in either direction will lead to destabilization,
by approximately 18-20 kcal/mol in the cation geometry
(despite increased delocalization of spin density) and ap-
proximately 15-22 kcal/mol in the anion geometry. Being an
odd-electron species dooms the radical to a planar existence.
These experiments and computations also have important
implications for the use of metal-complexed benzylic radicals
in synthesis and physical organic studies. First, the high
stereoselectivities observed, to date, in radical reactions of Cr-
(CO)3-complexed arenes5 must be attributed to the steric effect
of the Cr(CO)3 moiety rather than to a direct bonding interaction
beween chromium and the radical center. Second, the results
of the cyclopropylcarbinyl radical ring-opening reactions suggest
that metal arene complexes may provide faster radical clocks.13i
The parent cyclopropylcarbinyl radical rearranges with a rate
Chart 1
from the benzyl ligands to the Cr(CO)3 fragment. Since
geometric changes or delocalization of the unpaired electron
of a benzylic radical upon complexation would lead to 17- or
19-electron species, enhanced orbital interactions and extensive
spin delocalization are not possible, and little stabilization is
therefore observed.
Computed structures of the various benzyl species demon-
strate that direct interaction between chromium and the benzylic
position is present in cationic complexes but not in the
corresponding complexes of anions and radicals. The structural
representations shown in Chart 1 capture the essential features
of the computed structures. Further experiments and computa-
tions on chromium-complexed benzylic ions and radicals are
underway in our laboratories.
Methods
Computations. All calculations were performed with the Gaussian
94 program.32 Geometry optimizations employed the hybrid HF-DFT
B3LYP method33 with the LANL2DZ basis set.34 Density functional
theory has been used effectively in modeling various organometallic
systems,35,36 including various chromium-containing species such as
chromium carbonyls,36e,k Fischer carbenes,36b-d,h,j and chromium oxides
and halides.36a,f-h It has also been used to model organometallic radical
species.35f-h The LANL2DZ basis set is of double-ú quality and utilizes
an effective core potential (ECP) for chromium.34 Many DFT calcula-
tions on organometallic species have utilized ECP basis sets.35a,b,d,e,g,36f,h,j
(U)B3LYP/LANL2DZ bond lengths and angles for free toluene and
benzyl species are comparable in both relative and absolute magnitudes
with those calculated previously by (U)HF and (U)MP2 methods with
constant of 1.2 × 108 s-1 19
, but the 2-phenylcyclopropylcarbinyl
radical rearranges with a rate constant of 3 × 1011 s-1 13j
.
Based
on the work of Castellino and Bruice,30 it can be estimated that
the 2,3-diphenylcyclopropylcarbinyl radical rearranges with a
rate constant close to 3 × 1011 s-1. Therefore, based on the
2.5:1 ratio of ring-opened products described above, we can
estimate that the rate constant for rearrangement of the Cr(CO)3-
complexed 2,3-diphenylcyclopropylcarbinyl radical is approxi-
mately (7-8) × 1011 s-1. Thus, incorporation of transition metal
complexes into rearrangement substrates may lead to faster
radical clocks. Furthermore, the large difference in regioselec-
tivities observed for the ionic and radical ring-opening reactions
of substrate 4 suggests that systems of this type may find
application as probes for the intermediacy of radical vs ionic
intermediates.31
(32) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.; Petersson, G.
A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski,
V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.;
Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.;
Wong, M. W.; Andres, J. L.; Repogle, E. S.; Gomperts, R.; Martin, R. L.;
Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. P.; Head-
Gordon, M.; Gonzales, C.; Pople, J. A. Gaussian 94; Gaussian, Inc.:
Pittsburgh, PA, 1995.
Conclusions
(33) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652. (b) Becke,
A. D. J. Chem. Phys. 1993, 98, 1372-1377.
Cyclopropylcarbinyl ring-opening competition experiments
demonstrate that transition states bearing anionic, cationic, or
radical character at the benzylic position are stabilized to a
greater extent by phenyl-Cr(CO)3 than by phenyl alone. The
magnitude of this stabilization is greatest for the anionic case
and least in the case of the radical. This hierarchy of transition-
state stabilization energies is mirrored by computed ground-
state stabilization energies for Cr(CO)3-complexed benzyl anion,
cation, and radical.
The factors leading to stabilization of benzylic transition states
and intermediates by Cr(CO)3 can be summarized as follows.
Improved matching of orbital energies, coupled with geometric
changes such as bending of the benzyl framework and shifting
of the Cr(CO)3 moiety, lead to strengthened orbital interactions
in the ions, which are accompanied by delocalization of charge
(34) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270-283.
(35) Recent examples include: (a) Decker, S. A.; Klobukowski, M. J.
Am. Chem. Soc. 1998, 120, 9342-9355. (b) Jemmis, E. D.; Giju, K. T. J.
Am. Chem. Soc. 1998, 120, 6952-6964. (c) Drouin, B. J. J. Am. Chem.
Soc. 1998, 120, 6774-6780. (d) Maseras, F.; Lockwood, M. A.; Eisenstein,
O.; Rothwell, I. P. J. Am. Chem. Soc. 1998, 120, 6598-6602. (e) Niu, S.;
Hall, M. B. J. Am. Chem. Soc. 1998, 120, 6169-6170. (f) Braden, D. A.;
Tyler, D. R. J. Am. Chem. Soc. 1998, 120, 942-947. (g) Smith, K. M.;
Poli, R.; Legzdins, P. J. Chem. Soc., Chem. Commun. 1998, 1903-1904.
(h) Aarnts, M. P.; Wilms, M. P.; Peelen, K.; Fraanje, J.; Goubitz, K.; Hartl,
F.; Stufkens, D. J.; Baerends, E. J.; Vlcek, A., Jr. Inorg. Chem. 1996, 35,
8-5477. (i) Barone, V. J. Phys. Chem. 1995, 99, 11659-11666. (j) Barone,
V. Chem. Phys. Lett. 1995, 233, 129-133.
(36) Examples with Cr-containing species include: (a) Ault, B. S. J.
Am. Chem. Soc. 1998, 120, 6105-6112. (b) Torrent, M.; Duran, M.; Sola`,
M. Organometallics 1998, 17, 1492-1501. (c) Torrent, M.; Duran, M.;
Sola`, M. J. Chem. Soc., Chem. Commun. 1998, 999-1000. (d) Gleichmann,
M. M.; Do¨tz, K. H.; Hess, B. A. J. Am. Chem. Soc. 1996, 118, 10551-
10560. (e) Hu, C.-H.; Chong, D. P. Chem. Phys. Lett. 1996, 262, 733-
736. (f) Torrent, M.; Gili, P.; Duran, M.; Sola`, M. J. Chem. Phys. 1996,
104, 9499-9510. (g) Ziegler, T.; Li, J. Organometallics 1995, 14, 214-
223. (h) Jacobsen, H.; Ziegler, T. Organometallics 1995, 14, 224-230. (i)
Adamo, C.; Lelj, F. J. Chem. Phys. 1995, 103, 10605-10613. (j) Jacobsen,
H.; Schreckenbach, G.; Ziegler, T. J. Phys. Chem. 1994, 98, 11406-11410.
(k) Ziegler, T.; Tschinke, V.; Ursenbach, C. J. Am. Chem. Soc. 1987, 109,
4825-4837.
(30) Castellino, A. J.; Bruice, T. C. J. Am. Chem. Soc. 1988, 110, 7512-
7519.
(31) Design and application of a related "probe/clock" is described in:
(a) Toy, P. H.; Newcomb, M.; Hollenberg, P. F. J. Am. Chem. Soc. 1998,
120, 7719-7729. (b) Newcomb, M.; Chestney, D. L. J. Am. Chem. Soc.
1994, 116, 9753-9754. See also ref 13d.

Chemical Hermaphroditism in the Cr(CO)3 Moiety
J. Am. Chem. Soc., Vol. 121, No. 15, 1999 3605
double-ú and triple-ú basis sets (on average, the (U)B3LYP carbon-
carbon bond lengths are approximately 0.01-0.03 Å longer than those
computed with (U)HF or (U)MP2 methods)28f,37 and with those observed
in the recently published X-ray structure of the cumyl cation.38 In
addition, geometries of Cr(CO)6 and the Cr(CO)3 complex of toluene
were calculated as test cases (see Figure 1). Computed CO bond lengths
are overestimated relative to those observed in X-ray and neutron
diffraction structures (by approximately 0.03-0.05 Å),26,39 while Cr-
CO bond lengths appear to be reproduced very precisely (( ap-
proximately 0.01 Å); Cr-C(ring) distances are overestimated system-
atically by approximately 0.1 Å. Single-point energies were also
calculated for selected stationary points using an all-electron basis set
(the B3LYP/DZVP2+ level).35f,40 Electrostatic potential surfaces
(contoured at 0.002 electrons/au3) and constant spin density surfaces
(contoured at 0.004 electrons/au3) were calculated from (U)HF/3-21G
single-point calculations on (U)B3LYP/LANL2DZ-optimized geom-
etries using SPARTAN V5.0.41
chloride16 (126 mg, 0.54 mmol), and distilled cyclohexane (1 mL) was
heated to a gentle reflux. Ethyl diazoacetate (2.3 mL, 21.5 mmol) was
added as a solution in cyclohexane (3 mL) to the refluxing solution at
a rate of 0.2 mL h-1. After the addition was complete, the reaction
mixture was cooled to room temperature and filtered through a pad of
alumina using Et2O. The solution was evaporated to dryness, taken up
in hot hexanes, decolorized with charcoal, and filtered. The resulting
solution was evaporated to dryness, and the residue was recrystalized
from hexanes to afford 600 mg (42%) of the cyclopropane 1 as white
needles: 1H NMR (500 MHz, CDCl3) δ 1.39 (3H, t, J ) 7.1 Hz), 2.60
(1H, t, J ) 5.3 Hz), 3.11 (2H, d, J ) 5.3 Hz), 4.28 (2H, q, J ) 7.1
Hz), 6.99-7.01 (4H, m), 7.15-7.20 (6H, m); 13C NMR (125 MHz,
CDCl3) δ 14.19, 27.16, 33.16, 60.87, 126.32, 127.83, 128.78, 135.49,
173.24; IR (KBr) 3155, 3030, 1717, 1433, 1304, 1182, 908, 733, 650
cm-1; MS (70 eV) 266 (M+, 30), 221 (20), 193 (100), 178 (30), 115
(85), 91 (35); HRMS (EI) calcd for C18H18O2 266.1307, found 266.1303.
cis-2,3-Diphenyl-trans-1-hydroxymethylcyclopropane (2). A solu-
tion of cyclopropane ester 1 (780 mg, 2.93 mmol), THF (20 mL), and
lithium aluminum hydride (111 mg, 2.93 mmol) was stirred overnight.
The reaction was cooled to 0 °C and quenched by adding water (110
µL), followed by 15% aqueous NaOH (110 µL), followed by water
(330 µL). The resulting mixture was warmed to room temperature,
MgSO4 was added, and the slurry was filtered. The solvent was removed
in vacuo to give 638 mg (97%) of 2 as a white solid: 1H NMR (400
MHz, CDCl3) δ 1.56 (1H, t, J ) 4.9 Hz), 2.10 (1H, pent, J ) 5.9 Hz),
2.42 (2H, d, J ) 5.7 Hz), 3.86 (2H, t, J ) 6.1 Hz), 6.93-6.94 (4H,
m), 7.04-7.13 (6H, m); 13C NMR (100 MHz, CDCl3) δ 27.78, 29.60,
66.35, 125.88, 127.81, 128.95, 137.31; IR (KBr) 3359, 1603, 1496,
1446, 1030, 696 cm-1; MS (70 eV) 224 (M+, 45), 206 (30), 193 (100),
151 (50), 91 (25); HRMS (EI) calcd for C16H16O 224.1201, found
224.1206.
cis-2-(Phenyltricarbonylchromium)-3-phenyl-trans-1-hydroxy-
methylcyclopropane (3). A mixture of cyclopropyl alcohol 2 (104 mg,
0.46 mmol) and tris(acetonitrile)tricarbonylchromium (160 mg, 1.34
mmol) was diluted with dioxane (15 mL) and immersed in a preheated
oil bath at 140 °C for approximately 3 min. The reaction was cooled
to room temperature, filtered through a pad of silica gel, washed with
EtOAc, and absorbed onto Celite. The arene complex was isolated by
chromatography using 20% Et2O/hexanes, followed by 50% Et2O/
hexanes as the eluent, to afford 162 mg (98%) of 3 as a yellow solid:
1H NMR (500 MHz, C3D6O) δ 2.15-2.22 (2H, m), 2.67 (1H, dd, J )
9.4, 6.1 Hz), 3.60 (1H, t, J ) 6.6 Hz), 4.04 (2H, dd, J ) 11.0, 3.6 Hz),
5.00 (1H, d, J ) 6.7 Hz), 5.28 (1H, t, J ) 6.3 Hz), 5.44 (1H, t, J ) 6.5
Hz), 5.51 (1H, d, J ) 6.6 Hz), 5.57 (1H, t, J ) 5.9 Hz), 7.14-7.22
(5H, m); 13C NMR (125 MHz, C3D6O) δ 26.19, 26.66, 31.26, 62.18,
90.96, 91.85, 94.39, 94.46, 95.05, 112.58, 126.10, 127.86, 129.51,
136.51, 233.94; IR (KBr) 3406, 1954, 1857, 665 cm-1; MS (70 eV)
360 (M+, 50), 276 (50), 206 (50), 193 (100), 115 (55), 91 (40); HRMS
(EI) calcd for C19H16CrO4 360.0454, found 360.0454.
General Experimental Procedures. All reactions were carried out
in flame-dried glassware in an inert atmosphere of Ar. Reaction solvents
were distilled from the indicated drying agents: THF (Na, benzophe-
none), Et2O (Na, benzophenone), dioxane (Na, benzophenone), cyclo-
hexane (CaH2), and dichloromethane (CaH2). All reagents were purified
before use according to literature procedures. Samarium metal was
purchased from Aldrich and was kept in a drybox under N2 atmosphere.
Samarium diiodide was generated by the procedure given by Wipf et
al.18b The SmI2 solution was concentrated to dryness and kept in a
drybox under N2 atmosphere. HMPA and Et3SiH were distilled from
CaH2. CuCl was recrystallized from concentrated HCl. Ph3P‚I2 was
recrystallized from dried, degassed dichloromethane under Ar and kept
in a drybox under N2 atmosphere. Solvents for extractions and
chromatography were not distilled. Flash column chromatography was
performed by the method of Still et al.,42 using 32-63-µm silica gel
1
(ICN). H NMR and 13C NMR spectra were recorded at either 400 or
500 MHz using CDCl3 (δ 7.26 ppm for 1H, δ 77.0 ppm for 13C), C3D6O
1
(δ 2.05 ppm for H, δ 29.92 and δ 206.68 ppm for 13C), and C6D6 (δ
7.16 ppm for 1H, δ 128.39 ppm for 13C) as solvents and reference
standards. Chemical shifts are given in ppm (δ); multiplicites are
indicted by s (singlet), d (doublet), t (triplet), q (quartet), m (multiplet),
or br (broadened). Coupling constants, J, are reported in hertz. Infrared
spectra (IR) were recorded on a Nicolet 510P spectrometer, and signals
are reported in cm-1. High-resolution electron impact (EI) mass spectra
were obtained with an ionization voltage of 70 eV. Data are reported
in the form of m/z (intensity relative to base peak ) 100).
cis-2,3-Diphenyl-trans-1-carboethoxycyclopropane (1). A solution
of cis-stilbene (968 mg, 5.37 mmol), triethyl phosphite copper(I)
(37) (a) Adamo, C.; Subra, R.; di Matteo, A.; Barone, V. J. Chem. Phys.
1998, 109, 10244-10254. (b) Jones, J.; Bacskay, G. B.; Mackie, J. C. J.
Phys. Chem. A 1997, 101, 7105-7113. (c) Eiden, G. C.; Lu, K.-T.;
Badenhoop, J.; Weinhold, F.; Weisshaar, J. C. J. Chem. Phys. 1996, 104,
8886-8895. (d) Arnaud, R.; Postlethwaite, H.; Barone, V. J. Phys. Chem.
1994, 98, 5913-5919. (e) Amyes, T. L.; Richard, J. P.; Novak, M. J. Am.
Chem. Soc. 1992, 114, 8032-8041. (f) Dorigo, A. E.; Li, Y.; Houk, K. N.
J. Am. Chem. Soc. 1989, 111, 6942-6948. (g) Rice, J. E.; Handy, N. C.;
Knowles, P. J. J. Chem. Soc., Faraday Trans. 2 1987, 83, 1643-1649.
(38) Laube, T.; Olah, G. A.; Bau, R. J. Am. Chem. Soc. 1997, 119, 3087-
3092.
cis-2-(Phenyltricarbonylchromium)-3-phenyl-trans-1-iodometh-
ylcyclopropane (4). Cyclopropane 3 (80 mg, 0.36 mmol) was added
to a solution of Ph3P‚I2 (184 mg, 0.36 mmol), imidazole (24 mg, 0.36
mmol), and CH2Cl2 (3 mL) at 0 °C. The reaction mixture turned color
from yellow to white. The reaction was stirred for 40 min. The
triphenylphosphine oxide was removed by filtration using Et2O to rinse,
and the solute was absorbed onto Celite. The arene complex was isolated
by chromatography using hexane as the eluent to afford 134 mg (84%)
of 4 as a yellow oil: 1H NMR (400 MHz, C3D6O) δ 2.37 (1H, dd, J
) 9.5, 5.4 Hz), 2.45 (1H, dddd, J ) 9.5, 6.1, 2.9, 5.4 Hz), 2.64 (1H,
dd, J ) 9.4, 6.1 Hz), 3.38 (1H, t, J ) 9.5 Hz), 3.75 (1H, dd, J ) 9.8,
5.9 Hz), 5.05 (1H, d, J ) 6.5 Hz), 5.30 (1H, t, J ) 6.2 Hz), 5.42 (1H,
t, J ) 6.4 Hz), 5.46 (1H, d, J ) 6.5 Hz), 5.55 (1H, t, J ) 6.3 Hz),
7.16-7.24 (5H, m); 13C NMR (100 MHz, CDCl3) δ 9.35, 30.18, 34.94,
37.37, 89.89, 90.36, 92.85, 93.29, 93.59, 109.80, 127.07, 128.42, 129.60,
135.13, 233.08; IR (KBr) 1956, 1863, 1601, 1410, 1296, 1172, 1147,
704, 630, 619 cm-1; MS (FAB, Cs) 470 (M+, 75), 386 (95), 307 (100),
289 (40), 207 (40); HRMS (FAB) calcd for C19H15ICrO3 469.9467,
found 469.9471.
(39) (a) Rees, B.; Mitschler, A. J. Am. Chem. Soc. 1976, 98, 7918-
7924. (b) Jost, A.; Rees, B.; Yelon, W. B. Acta Crystallogr. 1975, B31,
2649-2658. (c) Whitaker, A.; Jeffery, J. W. Acta Crystallogr. 1967, 23,
977-984.
(40) The DZVP2 basis set was obtained from the Extensible Computa-
tional Chemistry Environment Basis Set Database, Version 1.0 (http://
www.emsl.pnl.gov:2080/forms/basisform.html), as developed and distributed
by the Molecular Science Computing Facility, Environmental and Molecular
Sciences Laboratory, which is part of the Pacific Northwest Laboratory,
P.O. Box 999, Richland, WA 99352, and funded by the U.S. Department
of Energy. The Pacific Northwest Laboratory is a multiprogram laboratory
operated by Battelle Memorial Institute for the U.S. Department of Energy
under Contract DE-AC06-76RLO 1830. Contact David Feller, Karen
Schuchardt, or Don Jones for further information. The basis set was
augmented as described in ref 35f.
(41) SPARTAN V5.0; Wavefunction, Inc.: 18401 Von Karman Ave.,
#370, Irvine, CA 92612; 1997.
(42) Still, C. W.; Kahn, M.; Mitra, A. J. Org. Chem. 1978, 43, 2923-
2925.
Anionic Rearrangement: 4-(Phenyltricarbonylchromium)-3-
phenyl-1-butene (5) via t-BuLi. To a solution of iodide 4 (100 mg,
0.21 mmol) and THF (5 mL) at -78 °C was added t-BuLi (1.7 M, 313

3606 J. Am. Chem. Soc., Vol. 121, No. 15, 1999
Merlic et al.
µL, 0.53 mmol), and the reaction was stirred for 10 min. The reaction
mixture was poured onto dilute HCl and extracted with Et2O (3×).
The combined organic layers were dried over MgSO4, filtered, and
absorbed onto Celite. The residue was chromatographed, using 5%
Et2O/hexanes as the eluent, to afford 54 mg (74%) of alkene 5 as a
) 6.4 Hz), 4.85 (1H, d, J ) 6.6 Hz), 4.86-4.89 (1H, m) 5.64 (1H, dd,
J ) 13.5, 8.31 Hz), 6.88 (2H, d, J ) 7.4 Hz), 7.02-7.09 (3H, m); 13
C
NMR (125 MHz, C6D6) δ 42.01, 49.70, 91.53, 92.04, 92.13, 92.62,
93.10, 93.40, 115.05, 117.27, 127.00, 128.88, 130.01, 139.10, 233.91;
IR (KBr, mixture) 3235, 2280, 1968, 1892, 1618, 1454, 1331, 1163,
812 cm-1; MS (70 eV, mixture) 344 (M+, 40), 260 (100), 208 (15),
169 (10), 117 (60); HRMS (EI, mixture) calcd for C19H16CrO3 344.0505,
found 344.0508.
Radical Rearrangement: 4-(Phenyltricarbonylchromium)-3-
phenyl-1-butene (5) and 3-(Phenyltricarbonylchromium)-4-phenyl-
1-butene (6) via SmI2. To a solution of SmI2 in THF (0.04 M, 0.38
mmol) and HMPA (190 µL, 1.14 mmol) was added a degassed solution
of iodide 5 (60 mg, 0.13 mmol) and t-BuOH (19 µL, 0.26 mmol) in
THF (10 mL) via cannula. A white precipitate formed, and the reaction
was allowed to stir at room temperature for 1 h. The reaction mixture
was then poured onto dilute HCl and extracted with Et2O (3×). The
combined organic layers were dried over MgSO4, filtered, and absorbed
onto Celite. The residue was chromatographed, using 2.5% Et2O/
hexanes as the eluent, to afford 40 mg (89%) of alkenes 5 and 6 as a
yellow oil. 1H NMR analysis indicated a 2.5:1 ratio of products 5 and
6, respectively.43
1
yellow oil. H NMR analysis indicated a >99:<1 ratio of products 5
and 6, respectively.43 5: 1H NMR (500 MHz, C6D6) δ 2.32 (1H, dd, J
) 13.5, 8.31 Hz), 2.41 (1H, dd, J ) 13.5, 6.8 Hz), 3.04 (1H, q, J )
7.4 Hz), 4.10 (1H, d, J ) 6.4 Hz), 4.20 (1H, t, J ) 6.3 Hz), 4.32-4.34
(2H, m), 4.49 (1H, t, J ) 6.3 Hz), 4.81 (1H, d, J ) 17.1 Hz), 4.89
(1H, d, J ) 10.3 Hz), 5.70 (1H, ddd, J ) 17.4, 10.2, 7.4 Hz), 6.82
(2H, d, J ) 7.1 Hz), 7.00-7.10 (3H, m); 13C NMR (125 MHz, C6D6)
δ 41.91, 52.65, 90.49, 93.54, 93.67, 93.84, 93.97, 111.40, 115.56,
127.34, 128.43, 129.16, 140.91, 142.65, 234.04; IR (KBr) 3235, 2280,
1968, 1892, 1618, 1454, 1331, 1163, 812 cm-1; MS (70 eV) 344 (M+,
40), 260 (100), 208 (15), 169 (10), 117 (60); HRMS (EI) calcd for
C19H16CrO3 344.0505, found 344.0508.
Cationic Rearrangement: 4-(Phenyltricarbonylchromium)-3-
phenyl-1-butene (5) and 3-(Phenyltricarbonylchromium)-4-phenyl-
1-butene (6) via AgBF4. To a 50-mL two-neck round-bottom flask
containing iodide 4 (60 mg, 0.13 mmol), Et3SiH (102 µL, 0.64 mmol),
and CH2Cl2 (5 mL) was added AgBF4 (25 mg, 0.13 mmol) via a solid
addition funnel, causing the reaction mixture to turn a dark brown color.
The reaction stirred for 1 h, whereby the color of the reaction dissipated
to a clear yellow. H2O (500 µL) was added, and the reaction was filtered
through a pad of Celite using Et2O to rinse. The solvent was absorbed
onto Celite, and the crude reaction mixture was chromatographed, using
5% Et2O/hexanes as the eluent, to afford 36 mg (82%) of alkenes 5
and 6 a yellow oil. 1H NMR analysis indicated a 95:5 ratio of products
5 and 6, respectively.43 6: 1H NMR (500 MHz, C6D6) δ 2.56 (1H, dd,
J ) 13.2, 8.5 Hz), 2.61 (1H, dd, J ) 13.2, 6.2 Hz), 2.94 (1H, q, J )
7.8 Hz), 4.31-4.38 (3H, m), 4.44 (1H, t, J ) 6.4 Hz), 4.55 (1H, d, J
Acknowledgment. We are grateful to the National Science
Foundation and the National Institutes of Health for financial
support of this research. We gratefully acknowledge grants of
computer time from the UCLA Office of Academic Computing
and the National Computational Science Alliance (under Grant
MCA93S015N, utilizing the NCSA SGI/CRAY POWER CHAL-
LENGEarray and the NCSA HP-Convex Exemplar SPP-2000).
We acknowledge Bruce Hietbrink for many helpful discussions.
Note Added in Proof. A related paper, which arrives at
similar conclusions, has just been published: Pfletschinger, A.;
Dargel, T. K.; Bats, J. W.; Schmalz, H.-G.; Koch, W. Chem.
Eur. J. 1999, 5, 537-545.
(43) It is possible that a haptotropic shift of the Cr(CO)3 moiety from
one phenyl group to the other in the reactive intermediates could affect the
observed product ratios. However, we are not aware of precedent for this,
and the experimental results are fully in accord with literature precedent1-7
and our computational results.
JA983934K