Reactivity of (η6-arene)tricarbonylchromium complexes toward additions of anions, cations, and radicals

Article abstract of DOI:10.1021/ja000600y

A computational and experimental study of additions of electrophiles, nucleophiles, and radicals to tricarbonylchromium-complexed arenes is reported. Competition between addition to a complexed arene and addition to a noncomplexed arene was tested using 1

Full text of DOI:10.1021/ja000600y

4904
J. Am. Chem. Soc. 2001, 123, 4904-4918
Reactivity of (η6-Arene)tricarbonylchromium Complexes toward
Additions of Anions, Cations, and Radicals
Craig A. Merlic,* Michael M. Miller, Bruce N. Hietbrink, and K. N. Houk*
Contribution from the Department of Chemistry and Biochemistry, UniVersity of California, Los Angeles,
Los Angeles, California 90095-1569
ReceiVed February 18, 2000
Abstract: A computational and experimental study of additions of electrophiles, nucleophiles, and radicals to
tricarbonylchromium-complexed arenes is reported. Competition between addition to a complexed arene and
addition to a noncomplexed arene was tested using 1,1-dideuterio-1-iodo-2-((phenyl)tricarbonylchromium)-
2-phenylethane. Reactions under anionic and cationic conditions give exclusive formation of 1,1-dideuterio-
1-((phenyl)tricarbonylchromium)-2-phenylethane arising from addition to the complexed arene. Radical condi-
tions (SmI2) afford two isomeric products, reflecting a 2:1 preference for radical addition to the noncomplexed
arene. In contrast, intermolecular radical addition competition experiments employing ketyl radical addition to
benzene and (benzene)tricarbonylchromium show that addition to the complexed aromatic ring is faster than
attack on the noncomplexed species by a factor of at least 100 000. Density functional theory calculations
using the B3LYP method, employing a LANL2DZ basis set for geometry optimizations and a DZVP2+ basis
set for energy calculations, for all three reactive intermediates showed that tricarbonylchromium stabilizes all
three types of intermediates. The computational results for anionic addition agree well with established chemistry
and provide structural and energetic details as reference points for comparison with the other reactive
intermediates. Intermolecular radical addition leads to exclusive reaction on the complexed arene ring as predicted
by the computations. The intramolecular radical reaction involves initial addition to the complexed arene ring
followed by an equilibrium leading to the observed product distribution due to a high-energy barrier for homolytic
cleavage of an exo bond in the intermediate cyclohexadienyl radical complex. Mechanisms are explored for
electrophilic addition to complexed arenes. The calculations strongly favor a pathway in which the cation
initially adds to the metal center rather than to the arene ring.
Introduction
While there is no reduction in the aromaticity,7 there is a
remarkable enhancement in reactivity of the arene ring toward
direct anti nucleophilic addition due to the electron-withdrawing
character of the tricarbonylchromium fragment. Nucleophilic
additions by anions are extremely facile, the reactivity patterns
The chemistry of chromium tricarbonyl-complexed arenes1
continues to fascinate researchers more than 40 years after the
initial discovery of (benzene)tricarbonylchromium,2 with ap-
plications in diastereoselective synthesis,3 enantioselective
synthesis,4 chiral ligands,5 and even total synthesis.6 Complex-
ation of an arene ring by tricarbonylchromium imparts dramatic
changes to the reactivity of the arene moiety and is summarized
by Figure 1. Most significant, considering the long history of
aromaticity and aromatic chemistry, is that chromium complex-
ation of aryl rings dramatically alters their electronic nature.
(3) For recent work, see: (a) Del Buttero, P.; Baldoli, C.; Molteni, G.;
Pilati, T. Tetrahedron: Asymmetry 2000, 11, 1927-1941. (b) Moser, W.
H.; Endsley, K. E.; Colyer, J. T. Org. Lett. 2000, 2, 717-719. (c) Kamikawa,
K.; Uemura, M. Synlett 2000, 938-949. (d) Koide, H.; Uemura, M. Chirality
2000, 12, 352-359. (e) Watanabe, T.; Shakadou, M.; Uemura, M. Inorg.
Chim. Acta 1999, 296, 80-85. (f) Schmalz, H.-G.; de Koning, C. B.;
Bernicke, D.; Siegel, S.; Pfletschinger, A. Angew. Chem., Int. Ed. 1999,
38, 1620-1623. (g) Koide, H.; Uemura, M. Tetrahedron Lett. 1999, 40,
3443-3446. (h) Baldoli, C.; Del Buttero, P.; Perdicchia, D.; Pilati, T.
Tetrahedron 1999, 55, 14089-14096. (i) Mandal, S. K.; Sarkar, A. J. Org.
Chem. 1999, 64, 2454-2458. (j) Caro, B.; Senechal-Tocquer, M. C.; Robin-
Le Guen, F.; LePoul, P. J. Organomet. Chem. 1999, 585, 43-52.
(4) For recent work, see: (a) Ratni, H.; Crousse, B.; Ku¨ndig, E. P. Synlett
1999, 626-628. (b) Schnell, B.; Bernardinelli, G.; Ku¨ndig, E. P. Synlett
1999, 348-350. (c) Quattropani, A.; Bernardinelli, G.; Ku¨ndig, E. P. HelV.
Chim. Acta 1999, 82, 90-104. (d) Taniguchi, N.; Hata, T.; Uemura, M.
Angew. Chem., Int. Ed. 1999, 38, 1232-1235. (e) Ariffin, A.; Blake, A. J.;
Ewin, R. A.; Li, W. S.; Simpkins, N. S. J. Chem. Soc., Perkin Trans. 1
1999, 3177-3189. (f) Davies, S. G.; Loveridge, T.; Fatima, M.; Teixeira,
C. C.; Clough, J. M. J. Chem. Soc., Perkin Trans. 1 1999, 3405-3412.
(5) For recent work, see: (a) Jones, G. B.; Guzel, M. Tetrahedron Lett.
2000, 41, 4695-4699. (b) Tranchier, J. P.; Chavignon, R.; Prim, D.;
Auffrant, A.; Plyta, Z. F.; Rose-Munch, F.; Rose, E. Tetrahedron Lett. 2000,
41, 3607-3610. (c) Vasen, D.; Salzer, A.; Gerhards, F.; Gais, H. J.; Sturmer,
R.; Bieler, N. H.; Togni, A. Organometallics 2000, 19, 539-546. (d)
Chapman, B. J.; Jones, G. B.; Pennington, W. T. J. Chem. Crystallogr.
1999, 29, 383-389. (e) Bolm, C.; Muniz, K. Chem. Soc. ReV. 1999, 28,
51-59. (f) Son, S. U.; Jang, H. Y.; Han, J. W.; Lee, I. S.; Chung, Y. K.
Tetrahedron: Asymmetry 1999, 10, 347-354.
(1) For reviews, see: (a) Rose-Munch, F.; Rose, E. Curr. Org. Chem.
1999, 3, 445-467. (b) Semmelhack, M. F. In ComprehensiVe Organo-
metallic Chemistry II; Abel, E. W., Stone, F. G. A., Wilkinson, G., Eds.;
Pergamon: New York, 1995; Vol. 12, pp 979-1015. (c) Semmelhack, M.
F. In ComprehensiVe Organometallic Chemistry II; Abel, E. W., Stone, F.
G. A., Wilkinson, G., Eds.; Pergamon: New York, 1995; Vol. 12, pp 1017-
1038. (d) Davies, S. G.; McCarthy, T. D. In ComprehensiVe Organometallic
Chemistry II; Abel, E. W., Stone, F. G. A., Wilkinson, G., Eds.;
Pergamon: New York, 1995; Vol. 12, pp 1039-1070. (e) Morris, M. J. In
ComprehensiVe Organometallic Chemistry II; Abel, E. W., Stone, F. G.
A., Wilkinson, G., Eds.; Pergamon: New York, 1995; Vol. 5, pp 471-
549. (f) Davies, S. G.; Donohoe, T. J. Synlett 1993, 323-332. (g) Davies,
S. G.; Coote, S. J.; Goodfellow, C. L. AdV. Met.-Org. Chem. 1991, 2, 1-57.
(h) Pape, A. R.; Kaliappan, K. P.; Ku¨ndig, E. P. Chem. ReV. 2000, 100,
2917-2940.
(2) (a) Fischer, E. O.; O¨ fele, K. Chem. Ber. 1957, 90, 2532-2535. (b)
Fischer, E. O.; O¨ fele, K.; Essler, H.; Fro¨lich, W.; Mortensen, J. P.;
Semmlinger, W. Z. Naturforsch. B 1958, 13, 458. (c) Fischer, E. O.; O¨ fele,
K.; Essler, H.; Frohlich, W.; Mortensen, J. P.; Semmlinger, W. Chem. Ber.
1958, 96, 2763-2772.
10.1021/ja000600y CCC: $20.00 © 2001 American Chemical Society
Published on Web 05/03/2001

ReactiVity of (η6-Arene)tricarbonylchromium Complexes
J. Am. Chem. Soc., Vol. 123, No. 21, 2001 4905
deactivates attached arene rings toward electrophilic substitution
to a small extent. Rosca and co-workers explored Friedel-Crafts
acetylation reactions in the same systems as those of von
Rosenberg and Pinder and found similar, though not identical,
results and also cited the deactivating nature of the tricarbonyl-
chromium moiety.14 They proposed a mechanism wherein
electrophilic addition on chromium was followed by rate
determining transfer of the electrophile to the aryl ring. Like
most previous researchers, low yields of acetylated product were
obtained. There are no reports of Friedel-Crafts alkylation
reactions of arene complexes. Electrophilic mercuration of
complexed benzene has also been reported, but attempts on
substituted analogues failed.15 Electrophiles such as nitrosonium
ion and benzenediazonium ion react at the metal of arene
complexes, leading to ligand substitution or oxidation proc-
esses.16 Arene complexes have been activated toward electro-
philic addition through reduction to the anionic η4 complex17
or through nucleophilic catalysis of electrophilic desilylation.18
There are several reports on protonation of (benzene)tricar-
bonylchromium.14,19 Protonation occurs at the metal and ex-
change with ring protons occurs, but at a slower rate than that
for free benzene.
The mechanistic question of the regioselectivity of initial
electrophilic addition by a carbocation to the ring (exo or endo)
or the metal in (arene)tricarbonylchromium complexes was
addressed by Bly, though he was not looking at arene substitu-
tion reactions, but rather neighboring group participation by
arenechromium complexes in solvolytic reactions.20 In a com-
plex series of papers, Bly examined solvolysis reactions of tri-
carbonylchromium-complexed substituted phenethyl and benzo-
norbornenyl methanesulfonates. Both d orbital (C-Cr bond) and
σ-π-type (endo C-C bond) participation were evaluated based
on reaction rates and stereochemistry, though most of the data
could not differentiate between them. He eventually concluded
that direct metal-carbon interactions were not likely to be
involved and attributed differing rate effects to σ-π-type and
ion-dipole interactions. Wells and Trahanovsky also tested the
solvolysis of benzonorbornenyl complexes.21 While their data
Figure 1. Effects of complexation of arenes by tricarbonylchromium.
are well defined, and the chemistry is widely utilized in
synthesis.1a,b,e The paucity of reports on electrophilic and radical
addition reactions stands in stark contrast to the extensive body
of literature on nucleophilic additions.
Electrophilic reactions of arene complexes were examined
soon after the discovery of these complexes and, given the rich
chemistry of electrophilic additions to aromatic compounds, this
was a natural course to pursue. However, only sporadic
investigations have been reported since, and often with conflict-
ing results. In their classic paper on nucleophilic substitution
reactions of arene complexes, Nicholls and Whiting reported
that arene complexes failed to undergo Friedel-Crafts acetyl-
ation and concluded that it was “unlikely that electrophilic
substitution will prove useful synthetically in this field”.8
However, the first successful Friedel-Crafts acetylation reac-
tions were reported the same year, but few details were
provided.9 A subsequent paper reported that the isomeric ratios
obtained from the toluene complex (o:m:p ) 39:15:46) were
different from those of the free ligand (o:m:p ) 9:2:89) and
that the complexed arenes were less reactive.10 Brown and co-
workers examined the rate of acetylation of the benzene complex
and found it to be more reactive than the free ligand.11 They
also suggested that electrophilic addition may involve dual attack
at the ring and at the metal. In an interesting series of papers,
Jackson and co-workers found that the isomer ratios in Friedel-
Crafts acetylation reactions of arene complexes are different
than those of the free arenes and that the ratios could be
rationalized based upon preferred conformations of the tricar-
bonylchromium moiety.12 Electrophilic addition occurred on
carbons not eclipsed by metal carbonyl bonds. They also found
that complexed arenes reacted much more slowly than the free
arenes. Von Rosenberg and Pinder examined the acetylation
reaction more carefully and definitively concluded that com-
plexed benzene was less reactive than free benzene.13 They went
on to test interesting monocomplexed R,ω-diphenylalkane
substrates and concluded that the tricarbonylchromium group
(14) Rosca, S.; Chiraleu, F.; Rosca, S. I. ReV. Roum. Chim. 1991, 36,
693-705.
(15) (a) Magomedov, G. K.-I.; Skyrin, V. G.; Frenkel, A. S. J. Gen.
Chem. USSR (Engl. Transl.) 1972, 42, 2443-2444; Zh. Obshch. Khim. 1972,
42, 2450-2452. (b) Magomedov, G. K.-I.; Skyrin, V. G.; Frenkel, A. S.;
Nekrasov, Y. S.; Yu, S.; Zavina, T. A. J. Gen. Chem. USSR (Engl. Transl.)
1975, 45, 2487-2490; Zh. Obshch. Khim. 1975, 45, 2533-2536.
(16) Connelly, N. G.; Demidowicz, Z.; Kelly, R. L. J. Chem. Soc., Dalton
Trans. 1975, 2335-2340.
(17) Leong, V. S.; Cooper, N. J. J. Am. Chem. Soc. 1988, 110, 2644-
2646.
(18) Effenberger, F.; Scho¨llkopf, K. Angew. Chem., Int. Ed. Engl. 1981,
20, 266.
(6) For recent work, see: (a) Kamikawa, K.; Watanabe, T.; Daimon,
A.; Uemura, M. Tetrahedron 2000, 56, 2325-2337. (b) Ratni, H.; Ku¨ndig,
E. P. Org. Lett. 1999, 1, 1997-1999. (c) Horstermann, D.; Schmalz, H.-
G.; Kociok-Kohn, G. Tetrahedron 1999, 55, 6905-6916. (d) Jones, G. B.;
Guzel, M.; Mathews, J. E. Tetrahedron Lett. 2000, 41, 1123-1126.
(7) Schleyer, P. v. R.; Kiran, B.; Simion, D. V.; Sorenson, T. S. J. Am.
Chem. Soc. 2000, 122, 510-515.
(8) Nicholls, B.; Whiting, M. C. J. Chem. Soc. 1959, 551-556.
(9) (a) Ercoli, R.; Calderazzo, F.; Mantica, E. Chim. Ind. (Milan) 1959,
41, 404. (b) Riemschneider, R.; Becker, D.; Franz, K. Monatsh.Chem. 1959,
90, 571-572.
(19) (a) Davison, A.; McFarlane, W.; Pratt, L.; Wilkinson, G. J. Chem.
Soc. 1962, 3653-3666. (b) Kursanov, D. N.; Setkina, V. N.; Baranetskaya,
N. K.; Zdanovich, V. I.; Anisimov, K. N. Dokl. Chem. (Engl. Transl.) 1970,
190, 127-129; Dokl. Akad. Nauk SSSR 1970, 190, 1103-1105. (c) Lillya,
C. P.; Sahatjian, R. A. Inorg. Chem. 1972, 11, 889-891. (d) Setkina, V.
N.; Zdanovich, V. I.; Baranetskaya, N. K.; Yurtanov, A. I.; Anisimov, K.
N.; Kursanov, D. N. J. Gen. Chem. USSR (Engl. Transl.) 1971, 41, 1345-
1348; Zh. Obshch. Khim. 1971, 41, 1338-1342. (e) Lokshin, B. V.;
Zdanovich, V. I.; Baranetskaya, N. K.; Setkina, V. N.; Kursanov, D. N. J.
Organomet. Chem. 1972, 37, 331-337. (f) Kursanov, D. N.; Setkina, V.
N.; Petrovskii, P. V.; Zdanovich, V. I.; Baranetskaya, N. K.; Rubin, I. D.
J. Organomet. Chem. 1972, 37, 339-345.
(20) (a) Bly, R. S.; Veazey, R. L. J. Am. Chem. Soc. 1969, 91, 4221-
4235. (b) Bly, R. S.; Strickland, R. C.; Swindell, R. T.; Veazey, R. L. J.
Am. Chem. Soc. 1970, 92, 3722-3729. (c) Bly, R. S.; Strickland, R. C. J.
Am. Chem. Soc. 1970, 92, 7459-7461. (d) Bly, R. S.; Mateer, R. A.; Tse,
K.-K.; Veazey, R. L. J. Org. Chem. 1973, 38, 1518-1527. (e) Bly, R. S.;
Maier, T. L. J. Org. Chem. 1978, 43, 614-621. (f) Bly, R. S.; Ni, E. K.;
Tse, A. K. K.; Wallace, E. J. Org. Chem. 1980, 45, 1362-1366.
(21) Wells, D. K.; Trahanovsky, W. S. J. Am. Chem. Soc. 1970, 92,
7461-7463.
(10) Herberich, G. E.; Fischer, E. O. Chem. Ber. 1962, 95, 2803-2809.
(11) (a) Brown, D. A.; Raju, J. R. J. Chem. Soc. A 1966, 40-43. (b)
Brown, D. A.; Hughes, F. J. Inorg. Chim. Acta 1967, 1, 448-450.
(12) (a) Gracey, D. E. F.; Jackson, W. R.; Jennings, W. B. J. Chem.
Soc., Chem. Commun. 1968, 366-367. (b) Jackson, W. R.; Jennings, W.
B. J. Chem. Soc. B 1969, 1221-1228. (c) Jackson, W. R.; Rae, I. D.; Wong,
M. G.; Semmelhack, M. F.; Garcia, J. N. J. Chem. Soc., Chem. Commun.
1982, 1359-1360. (d) Jackson, W. R.; Rae, I. D.; Wong, M. G. Aust. J.
Chem. 1986, 39, 303-315.
(13) von Rosenberg, J. L.; Pinder, A. R. J. Chem. Soc., Perkin Trans. 1
1987, 747-752.

4906 J. Am. Chem. Soc., Vol. 123, No. 21, 2001
Merlic et al.
could not differentiate between d orbital and σ-π-type partici-
pation, they favored the former. We recently reexamined these
benzonorbornenyl systems with density functional theory meth-
ods. Computations strongly favored chromium-carbon inter-
actions in the cationic intermediates rather than σ-π-type
interactions.22 A recent solvolysis study on an adamantyl
complex found no interaction of the chromium or the arene and
a â-cation, though this system was sterically congested.23
Surprisingly, only two systems have been explored in attempts
to exploit electrophilic addition reactions for synthetic purposes.
Jaouen examined intramolecular Friedel-Crafts acylation reac-
tions of chiral-complexed â-phenylpropionic acid derivatives
leading to diastereomeric indanone complexes and obtained
modest levels of selectivity and low yields of products.24 Uemura
used this chemistry to prepare tetralone derivatives in good
yields and excellent diastereoselectivity.25
Radical processes are even less tested than electrophilic
reactions. Only five examples of radical addition have been
reported and the mechanisms and scope of transformations are
far from being delineated.26 Schmalz discovered that ketyl and
azaketyl radicals add intramolecularly to chromium arene
complexes, but the mechanistic details of this new addition
reaction were not explored. Significantly, the relative reactivity
of arenes and arene complexes toward radical addition is not
known.
for electrophilic and radical addition were ambiguous.28 He
concluded that the rates of the latter two reactions should not
differ significantly between benzene and its complex. Albright
and Carpenter used extended Hu¨ckel molecular orbital theory
to examine nucleophilic and electrophilic, but not radical,
addition reactions.29 They mainly explored the regioselectivity
of addition to alkyl-substituted complexes (Jackson experiments,
vide supra) which, unlike complexes with electron-donating or
-withdrawing substituents, is not well explained by resonance
theory arguments.30 Instead, the regioselectivity is controlled
by the conformation of the Cr(CO)3 unit, which in turn depends
on the steric size of the alkyl substituent. Nucleophilic addition
occurs preferentially at eclipsed arene carbons while electrophilic
addition occurs at staggered arene carbons as a consequence of
intermixing between π* levels induced by the Cr(CO)3 orbitals.
They commented that a small electrophile like a proton should
add directly to the metal. Weber and co-workers used semi-
empirical quantum chemical methods derived from extended
Hu¨ckel molecular orbital theory to examine nucleophilic addition
reactions to indole complexes and were able to reproduce
experimental trends.31 Recently, Koga and co-workers employed
HF, B3LYP, and MP2 levels of theory to study molecular
electrostatic potentials and electron density topographies of arene
chromium complexes.27a Although they confirmed the results
of Albright and Carpenter, they did not consider electrophilic
addition to the metal and did not examine radical addition at
all.
Many theoretical calculations, initially employing Hu¨ckel
molecular orbital theory and recently ab initio methods, on arene
complexes have focused on bonding, structure, electron densi-
ties, and spectral properties,27 but there have been a few
computational studies on addition reactions to complexed arene
rings. Brown used Hu¨ckel molecular orbital theory and found
that nucleophilic addition should be enhanced, but the results
Given our long-term interest in metal-templated radical
reactions,27c,32 we initiated studies on the unexplored chemistry
of radical additions to arene chromium complexes. Our primary
focus was on radical chemistry, but we expanded the scope of
these investigations to include anionic and cationic addition
reactions (eq 1) since anionic reactions would provide calibration
(22) Tantillo, D. J.; Hietbrink, B. N.; Merlic, C. A.; Houk, K. N. J. Am.
Chem. Soc. 2000, 122, 7136-7137.
(23) Badejo, I. T.; Choi, H.; Hockensmith, C. M.; Karaman, R.; Pinkerton,
A. A.; Fry, J. L. J. Org. Chem. 1991, 56, 4688-4695.
(24) (a) Jaouen, G. Tetrahedron Lett. 1973, 1753-1756. (b) Dabard, R.;
Jaouen, G. Bull. Soc. Chim. Fr. 1974, 1639-1645. (c) Jaouen, G.; Dabard,
R. Bull. Soc. Chim. Fr. 1974, 1946-1950. (d) Caro, B.; Jaouen, G. J.
Organomet. Chem. 1982, 228, 87-106.
(25) (a) Uemura, M.; Isobe, K.; Take, K.; Hayashi, Y. J. Org. Chem.
1983, 48, 3855-3858. (b) Uemura, M.; Take, K.; Isobe, K.; Minami, T.;
Hayashi, Y. Tetrahedron 1985, 41, 5771-5778.
(26) (a) Schmalz, H.-G.; Siegel, S.; Bats, J. W. Angew. Chem., Int., Ed.
Engl. 1995, 34, 2383-2385. (b) Schmalz, H.-G.; Siegel, S.; Schwarz, A.
Tetrahedron Lett. 1996, 37, 2947-2950. (c) Hoffmann, O.; Schmalz, H.-
G. Synlett 1998, 1426-1428.
points against known benchmarks and cationic reactions would
allow entry into the unexplored Friedel-Crafts alkylations of
arene complexes. We report herein on (a) an intramolecular
rearrangement test that uses a single compound to directly
examine the relative reactivities of free and tricarbonyl-
chromium-complexed phenyl rings toward addition of an alkyl
anion, cation, and radical, (b) intermolecular radical addition
competition experiments, and (c) density functional theory
calculations on intra- and intermolecular addition of anions,
cations, and radicals to chromium-complexed aryl rings. Cou-
pling the experimental and theoretical results to provide a
mechanistic framework, we find that a Cr(CO)3 moiety can
activate arene rings toward anion, cation, and radical addition
reactions.
(27) (a) Suresh, C. H.; Koga, N.; Gadre, S. R. Organometallics 2000,
19, 3008-3015. (b) Schleyer, P. v. R.; Kiran, B.; Simion, D. V.; Sorenson,
T. S. J. Am. Chem. Soc. 2000, 122, 510-515. (c) Merlic, C. A.; Walsh, J.
C.; Tantillo, D. J.; Houk, K. N. J. Am. Chem. Soc. 1999, 121, 3596-3606.
(d) Pfletschinger, A.; Dargel, T. K.; Bats, J. W.; Schmalz, H.-G.; Koch,
W. Chem. Eur. J. 1999, 5, 537-545. (e) Elass, A.; Brocard, J.; Surpateanu,
G.; Vergoten, G. THEOCHEM 1999, 466, 35-48. (f) Baldridge, K. K.;
Siegel, J. S. J. Phys. Chem. 1996, 100, 6111-6115. (g) Simion, D. V.;
Sorenson, T. S. J. Am. Chem. Soc. 1996, 118, 7345-7352. (h) Rabaa, H.;
Saillard, J.-Y.; Le Beuze, A. J. Organomet. Chem. 1993, 463, 109-114.
(i) Kanis, D. R.; Ratner, M. A.; Marks, T. J. J. Am. Chem. Soc. 1992, 114,
10338-10357. (j) Williamson, R. L.; Hall, M. B. Int. J. Quantum Chem.,
Quantum Chem. Symp. 1987, 21, 503-512. (k) Kok, R. A.; Hall, M. B. J.
Am. Chem. Soc. 1985, 107, 2599-2604. (l) Kahn, S. D.; Hehre, W. J.;
Bartmess, J. E.; Caldwell, G. Organometallics 1984, 3, 1740-1743. (m)
Fitzpatrick, N. J.; Gogan, N. J.; King, I. J. J. Organomet. Chem. 1979,
166, 45-51. (n) Fitzpatrick, N. J.; Savariault, J. M.; Labarre, J. F. R. J.
Organomet. Chem. 1977, 127, 325-335. (o) Albright, T. A.; Hofmann, P.;
Hoffmann, R. J. Am. Chem. Soc. 1977, 99, 7546-7557. (p) Brown, D. A.;
Chester, J. P.; Fitzpatrick, N. J.; King, I. J. Inorg. Chem. 1977, 16, 2497-
2500. (q) Guest, M. F.; Hillier, I. H.; Higginson, B. R.; Lloyd, D. R. Mol.
Phys. 1975, 29, 113-128. (r) Brown, D. A.; Fitzpatrick, N. J.; Mathews,
N. J. J. Organomet. Chem. 1975, 88, C27-C29. (s) Saillard, J. Y.;
Grandjean, D.; Chopplin, F.; Kaufmann, G. J. Mol. Struct. 1974, 23, 363-
375. (t) Brown, D. A.; Rawlinson, R. M. J. Chem. Soc. A 1969, 1534-
1537. (u) Carroll, D. G.; McGlynn, S. P. Inorg. Chem. 1968, 7, 1285-
1290.
(28) Brown, D. A. J. Chem. Soc. 1963, 4389-4394.
(29) Albright, T. A.; Carpenter, B. K. Inorg. Chem. 1980, 19, 3092-
3097.
(30) Semmelhack, M. F.; Clark, G.; Farina, R.; Saeman, M. J. Am. Chem.
Soc. 1979, 101, 217-218.
(31) Furet, E.; Savary, F.; Weber, J.; Ku¨ndig, E. P. HelV. Chim. Acta
1994, 77, 2117-2124.
(32) (a) Merlic, C. A.; Xu, D. Q. J. Am. Chem. Soc. 1991, 113, 9855-
9856. (b) Merlic, C. A.; Xu, D.; Nguyen, M. C.; Truong, V. Tetrahedron
Lett. 1993, 34, 227-230. (c) Merlic, C. A.; Walsh, J. C. Tetrahedron Lett.
1998, 39, 2083-2086.

ReactiVity of (η6-Arene)tricarbonylchromium Complexes
Scheme 1. Rearrangement of the Diphenylethyl System
J. Am. Chem. Soc., Vol. 123, No. 21, 2001 4907
Results
Scheme 2. Synthesis of Competition Substrate
Experimental Design and Competition Experiments. We
sought a single substrate that could be used to study cationic,
anionic, and radical additions to chromium-complexed arenes
and to compare directly the relative reactivities of aryl rings
and chromium tricarbonyl-complexed aryl rings. Diphenylethyl
iodide proved to be an ideal candidate for our study due to the
possibility of having one complexed and one noncomplexed
arene in the same substrate. System 1 can give rise to two
rearranged regioisomers, 2 and 3, via an internal competition
experiment between the two arenes (Scheme 1). This allows
for facile determination of relative reactivities as well as
providing a means of direct comparison among the three possible
types of reactive intermediates. In path I, initial addition of the
anion, cation, or radical occurs on the complexed phenyl ring
anti to the chromium atom.33 Cleavage of a cyclopropyl bond
in the hexadienyl complex leads to a benzylic reactive inter-
mediate. Reaction quenching with a proton, hydride, or hydrogen
atom provides substituted diphenylethane 2. Path II proceeds
through ion or radical addition to the noncomplexed ring. Bond
cleavage provides the reactive intermediate stabilized by the
complexed phenyl ring which, upon quenching, yields comple-
mentary isomer 3. In our experimental design, the relative
amounts of 2 and 3 indicate which path is preferred for each of
the reactive intermediates. Quenching the initial reactive
intermediate without rearrangement yields reduction product 4
(path III).
the existence of this species, the “phenonium ion” is now a well-
accepted carbocation rearrangement intermediate.37 Even though
σ-bridged ethylenebenzenium ions have been subject to such
intense investigation over the last five decades,34 no clear study
on the effects of Cr(CO)3 complexation of the phenonium ion
has been reported.
Synthesis of the desired competition substrate 1 proved
straightforward (Scheme 2). Reduction of methyl diphenyl-
acetate with lithium aluminum deuteride provided deuterium-
labeled alcohol 5, which was then monocomplexed with
38
Cr(CO)3(CH3CN)3 in 70% yield. The complexed alcohol 7
was converted into iodide 1 in 87% yield by treatment with
freshly recrystallized PPh3‚I2 and imidazole in CH2Cl2.
With the key substrate in hand, we then explored the anionic,
cationic, and radical competition rearrangements. Upon treat-
ment of iodide 1 with tert-butyllithium at -78 °C, allowing
the reaction mixture to warm slowly to 25 °C, and quenching
with dilute HCl, the anion was found to rearrange in 40%
conversion to isomer 2 (eq 3). The remaining material was 1,1-
This test system harks back to the classic work on the
phenonium ions.34 Cram demonstrated in 1949 that solvolysis
of a â-substituted ethylbenzene leads to a Wagner-Meerwein
rearrangement via the phenyl-bridged species, rather than the
classical “open” carbonium ion (eq 2).35 Since Cram’s original
proposal, numerous investigators have examined this system,
concentrating on the nature of the intermediate ion.36 Although
a number of researchers became embroiled in a controversy over
(33) Endo addition or addition to the metal will be addressed later.
(34) Lancelot, C. J.; Cram, D. J.; Schleyer, P. v. R. In Carbonium Ions;
Olah, G. A., Schleyer, P. v. R., Eds.; Wiley: New York, 1972; Vol. 3,
Chapter 27, pp 1347-1483 and references therein.
(35) (a) Cram, D. J. J. Am. Chem. Soc. 1949, 71, 3863-3870, 3871-
3875, 3875-3883. (b) Cram, D. J. J. Am. Chem. Soc. 1952, 74, 2129-
2137, 2137-2148, 2149-2151, 2152-2159, 2159-2165. (c) Cram, D. J.
J. Am. Chem. Soc. 1964, 86, 3767-3772.
(36) (a) Fornarini, S.; Muraglia, V. J. Am. Chem. Soc. 1989, 111, 873-
877. (b) Olah, G. A.; Porter, R. D. J. Am. Chem. Soc. 1971, 93, 6877-
6887. (c) Olah, G. A.; Porter, R. D. J. Am. Chem. Soc. 1970, 92, 7627-
7629.
dideuterio-2-((phenyl)tricarbonylchromium)-2-phenylethane (4),
resulting from simple reduction of iodide 1. When the com-
plexed diphenylethyl iodide was treated under radical initiating

4908 J. Am. Chem. Soc., Vol. 123, No. 21, 2001
Merlic et al.
conditionssSmI2/HMPA39 and tert-butyl alcohol in THF at
room temperaturesisomers 2 and 3 were formed in a ratio of
1:2 in an 11% conversion. The product ratio and conversion
depended somewhat on the reaction conditions, but this result
was typical. Again, the unrearranged material was reduced to
4. Analogously, the cationic rearrangement, performed by
reaction of the iodide with AgBF4 at room temperature in the
presence of Et3SiH as a hydride trap, afforded 2 in 94% yield.
2-5% rearrangement of 2-phenylethyl radical at 170 °C. Upon
treatment of 8 with AgBF4 under the same conditions as before
for a cationic rearrangement, only starting material was recov-
ered. This result was surprising in light of the reactivity of the
complexed substrate and thus was repeated several times. It
suggests either enhancement of reactivity of the iodide by
chromium complexation or direct participation in the cationic
reaction by the chromium.
1
Isomers 2 and 3 are readily differentiated by H NMR as the
Because of the low conversion to the rearranged product in
the radical reaction of 1, we sought to determine whether our
system was providing an unintentional bias due to the constained
angle of addition. Thus, to remove any bias, we tested
intermolecular radical additions. Remarkably, reaction of acetone
and samarium(II) iodide in the presence of 6 equiv of benzene-
d6 and 1 equiv of (benzene)tricarbonylchromium (11) yielded
a 74:17:9 mixture of 13:14:15 in 58% yield.44 No evidence of
addition to the noncomplexed deuterium-labeled arene was
obtained by mass spectrometric analysis! As a control reaction,
the reverse sequence was carried out using 6 equiv of benzene
and 1 equiv of (benzene-d6)tricarbonylchromium (17) and
similar results were found. Repetition of the reaction shown in
eq 5, but using 100 equivalents of benzene-d6, again resulted
noncomplexed benzylic methylene protons in 2 are observed
as a singlet at 2.36 ppm and the protons benzylic to the
chromium-complexed ring in 3 are seen as a singlet at 2.14
ppm.40 Given the significant chemical shift difference, the
detection limit was less than 1%.
These initial results demonstrated that complexation by
chromium dramatically activates the phenyl ring toward addition
of both anions and cations, but that the Cr(CO)3 entity has a
slight deactivating character for radical addition to the arene
ring in this system. While the results for the anion were
expected, the results for the cation and radical were not. This
led to further analysis of these systems.
To verify that the product ratio reflected the ratio of
cyclization pathways, a control was needed for each of the three
cases, and the obvious standard to use was the noncomplexed
system 8 (eq 4). When this symmetric diphenylethyl iodide was
in no observed addition to the noncomplexed deuterium-labeled
aromatic substrate with a detection limit of 0.1%.
treated under the same anionic reaction conditions, no rearranged
product was observed; only net reduction of the iodide to 10
was seen in 84% yield. This result is consistent with the work
of Zimmerman and Zweig, who found that rearrangement in a
similar system required refluxing conditions for several hours.41,42
The radical reaction, also run under the same conditions
employed for the complexed case, yielded only 12% rearrange-
ment (9), the remaining mass recovery being the reduced product
(10). The low rearrangement conversion for the radical reaction
is consistent with the results of Slaugh,43 who observed only
(37) (a) Okuyama, T.; Ochiai, M. J. Am. Chem. Soc. 1997, 119, 4785-
4786. (b) Goldfinger, M. B.; Crawford, K. B.; Swager, T. M. J. Am. Chem.
Soc. 1997, 119, 4578-4593. (c) Olah, G. A.; Head, N. J.; Rasul, G.; Prakash,
G. K. S. J. Am. Chem. Soc. 1995, 117, 875-882.
Upon completion of the competition experiments, we turned
to computational studies to examine in greater detail the
electronic and structural effects controlling the direct addition
of anions, cations, and radicals to complexed and noncomplexed
arenes. In particular, we wished to answer two intriguing and
nonintuitive questions. First, why is the intramolecular cation
addition from 1 strongly favored on the chromium-complexed
phenyl group, when the tricarbonylchromium moiety is generally
thought of as an electron-withdrawing substituent comparable
to a nitro group and nearly all reports on intermolecular
electrophilic aromatic substitution reactions of arene complexes
find them less reactive than the respective free arenes?9-14
Second, given that complexation by tricarbonylchromium can
activate phenyl rings toward addition of both anions and cations,
(38) (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.
(39) (a) Hasegawa, E.; Curran, D. P. Tetrahedron Lett. 1993, 34, 1717-
1720. (b) Inanaga, J.; Ishikawa, M.; Yamiguchi, M. Chem. Lett. 1987, 1485-
1486. (c) Girard, P.; Namy, J. L.; Kagan, H. B. J. Am. Chem. Soc. 1980,
102, 2693-2698.
(40) The 1H NMR spectrum of an authentic sample of (dibenzyl)-
tricarbonylchromium (see Experimental Section in the Supporting Informa-
tion) exhibits a triplet resonance at 2.37 and 2.15 ppm for the noncomplexed
benzylic methylene protons in 2 and the complexed benzylic methylene
protons in 3, respectively. See also: Traylor, T. G.; Goldberg, M. J. J. Am.
Chem. Soc. 1987, 109, 3968-3973.
(41) Zimmerman, H. E.; Zweig, A. J. Am. Chem. Soc. 1961, 83, 1196-
1213.
(42) For reviews on anionic 1,2-aryl rearrangements, see: (a) Groven-
stein, E., Jr. Aryl Migrations in Organometallic Compounds of the Alkali
Metals. In AdVances in Organometallic Chemistry; Stone, F. G. A., West,
R., Eds.; Academic Press: New York, 1977; Vol. 16, pp 167-210. (b)
Grovenstein, E., Jr. Angew. Chem., Int. Ed. Engl. 1978, 17, 313-332.
(43) Slaugh, L. H. J. Am. Chem. Soc. 1959, 81, 2262-2266.
(44) The ratio of isomers is based on 1H NMR analysis, while the yield
is based on 15% recovered starting material 11.

ReactiVity of (η6-Arene)tricarbonylchromium Complexes
J. Am. Chem. Soc., Vol. 123, No. 21, 2001 4909
Figure 2. (U)B3LYP/LANL2DZ-optimized structures of starting materials and products of CH3* addition reactions. Selected distances are in Å.
Table 1. Heats of Reaction for CH3* Addition to Benzene and
(Benzene)tricarbonylchromiuma
Figure 3. Electrostatic potential surfaces of free and Cr(CO)3-
complexed ions and constant spin density surfaces for free and Cr(CO)3-
complexed radicals formed from CH3* addition reactions.
a All values are in kcal/mol. Energies were calculated using the
DZVP2+ basis set. Values in parentheses are from the LANL2DZ basis
set which is a lower level of theory.
the noncomplexed case. Delocalization of the negative charge
onto the chromium center results in increased back-bonding from
the metal to the CO ligands. This is reflected structurally in a
0.02 Å decrease in the Cr-C bond lengths and a 0.02 Å increase
in the C-O bond lengths in 23a in comparison to 11 (Figure
2). Delocalization of charge onto the chromium moiety is also
apparent upon examination of the electrostatic potential surfaces
shown in Figure 3. In the noncomplexed case, the bulk of the
negative charge lies in the pentadienyl fragment. In the
chromium complex, the six-membered ring is more positive,
as negative charge has been transferred to the chromium
fragment. This charge distribution is also reflected in the
calculated Mulliken charges. In the anionic chromium complex
23a, the sp3 carbon of the six-membered ring is tipped up away
from the metal. This inclination of the substituted carbon away
from the metal compares closely with the X-ray crystal structure
of the product of dithiane nucleophilic addition to (benzene)-
tricarbonylchromium as reported by Semmelhack and co-
workers.45,46 They report an inclination of 38.6° away from the
why is there no significant selectivity in the intramolecular
radical addition tested, but there is in the intermolecular radical
additions?
Computed Structures of Free and Cr(CO)3-Complexed
Arene Species
Intermolecular Additions. Additions of methyl anion, radi-
cal, and cation to the aromatic ring of benzene and of
(benzene)tricarbonylchromium were investigated. Optimized
structures of the starting materials and products were obtained
using density functional theory calculations as discussed in the
Methods Section in the Supporting Information (Figure 2). The
∆H for each reaction was calculated from the energies of these
structures, and a comparison between reactions of free and
chromium-complexed benzene provided the ∆∆H due to
complexation (Table 1). The large calculated exothermicity of
the ionic reaction systems stems, in part, from the inherent
reactivity of the bare gas-phase methyl ions.
Methyl anion addition to benzene, forming product 22a, is
exothermic by 24.1 kcal/mol. Anion addition to (benzene)tri-
carbonylchromium (11) yields compound 23a. The electron-
withdrawing nature of the chromium fragment results in a
greatly enhanced reaction: 54.4 kcal/mol more favorable than
(45) Semmelhack, M. F.; Hall, H. T., Jr.; Farina, R.; Yoshifuji, M.; Clark,
G.; Bargar, T.; Hirotsu, K.; Clardy, J. J. Am. Chem. Soc. 1979, 101, 3535-
3544.
(46) A more recent X-ray crystal structure of an anion adduct was
reported and the structural features are similar, though slightly distorted
due to heteroatom substitution, see: Fretzen, A.; Ripa, A.; Liu, R.;
Bernardinelli, G.; Ku¨ndig, E. P. Chem. Eur. J. 1998, 4, 251-259.

4910 J. Am. Chem. Soc., Vol. 123, No. 21, 2001
Merlic et al.
plane defined by the pentadienyl fragment while complex 23a
shows an inclination of 38.3°. The distance between chromium
and the plane defined by the pentadienyl fragment in complex
23a is 1.832 Å. This is 0.09 Å longer than the comparable
distance in Semmelhack’s dithiane complex.
etries of the starting materials and products are shown in Figure
4. These structures more closely resemble our experimental
system, and the reactions are considerably less exothermic than
the methyl additions (even endothermic for the radical and anion
cases) due to the strain of the spiro[5.2]octane skeleton.
Cyclization of the noncomplexed phenylethyl anion 24a to
spirocycle 25a is endothermic by 5.4 kcal/mol, but complexation
by the electron-withdrawing chromium moiety makes this
reaction favorable by 19.8 kcal/mol (Table 2). We were unable
to find a fully optimized acyclic chromium-bound phenylethyl
anion as cyclization to 27a is too favorable; every starting
geometry minimized to the cyclic structure. To calculate
structure 26a, the exocyclic Cipso-CR-Câ angle was con-
strained to 110° and the rest of the molecule was allowed to
relax. The calculated structure for cyclized product 27a is
comparable to that of methyl anion addition, 23a. In particular,
the sp3 carbon of the ring is tilted away from the metal center
(36.6° relative to the pentadienyl plane). Examination of bond
lengths and electrostatic potential surfaces again reflects polar-
ization of charge toward the chromium tricarbonyl fragment
(Figure 5).
Radical cyclization was endothermic in both the non-
complexed and complexed cases, by 11.8 and 10.9 kcal/mol,
respectively. The cyclization of phenylethyl radical 24b to
spirocycle 25b was previously examined at the HF/STO-3G54,55
and UMP2/6-31G*56 levels. In the cyclic structure, previous
researchers report distances of 1.525,54 1.521,55 and 1.526 Å56
for the newly formed carbon-carbon bond. Our fully minimized
structure has a distance of 1.572 Å for this bond. The calculated
structures in the complexed case resemble those for anionic
cyclization. The tilt of the substituted carbon away from the
metal in 27b is less pronounced (Figure 5), though, at 24.2°
relative to the pentadienyl fragment plane. In both acyclic and
cyclic radical structures, the chromium tricarbonyl fragment is
rotated 60° relative to anion 27a. Although both cyclizations
are endothermic, the calculated energetic preference for radical
cyclization onto the complexed aryl ring is at odds with the
experimental results.
Methyl radical addition is less favorable than the methyl anion
addition, though still exothermic in both the free and chromium-
bound reactions. In this case, the complexed reaction, forming
23b, is 9.0 kcal/mol more exothermic than the noncomplexed
reaction, forming 22b. The optimized structure of metal complex
23b is qualitatively similar to that of anionic complex 23b. The
sp3 carbon of the six-membered ring is inclined away from the
metal, forming an angle of 33.4° with the pentadienyl plane.
The chromium tricarbonyl fragment is rotated by 60° relative
to 23a, and the chromium is 1.803 Å from the pentadienyl plane.
The spin-density surface in Figure 3 demonstrates the nearly
complete transfer of radical character from the ring to the metal
fragment.
Addition of methyl cation to benzene, forming 22c, is the
most exothermic reaction of this series (86.5 kcal/mol), reflecting
the inherent reactivity of the methyl cation. Wheland intermedi-
ate 22c was previously examined by Raos and co-workers using
RHF calculations and a 6-31G* basis set.47 These researchers
found an optimal distance of 1.5588 Å for the bond between
the methyl and the ring carbon, compared to our calculated
distance of 1.590 Å. An X-ray crystal structure of a related
Wheland intermediate was recently reported.48 Chromium
complexation increases the exothermicity of this reaction by
4.3 kcal/mol. In complex 23c, lengthening of the Cr-C(CO)
bonds by 0.05 Å and slight (0.002 Å) shortening of the C-O
bonds indicates a decrease in back-bonding to the CO ligands
as a result of the withdrawal of electron density from the
chromium fragment. Again, the electrostatic potential surfaces
in Figure 3 reflect the transfer of charge from the six-membered
ring to the chromium tricarbonyl moiety. Cationic chromium
complex 23c also exhibits a very interesting agostic interaction
of the metal center with the hydrogen on the sp3 center. Unlike
the previous two cases, this sp3 carbon is inclined toward the
metal center by 10.3°. This C-H bond is lengthened to 1.245
Å (compared to 1.099 Å in the anionic case) and the Cr-H
distance is 1.838 Å. Known chromium-hydride bonds have
measured lengths ranging from 1.42 to 2.014 Å,49 and a Cr-H
agostic interaction has been noted at a distance of 2.124 Å,50
so the Cr-H distance in 23c is well within this range. This
agostic interaction is not observed in the anion and radical cases,
as the ipso hydrogens of 23a and 23b are tilted away from the
metal. Early studies on protonation of (arene)tricarbonyl-
chromium complexes using NMR analysis suggested predomi-
nant metal protonation,19,51 but agostic structures, and the
resulting spectroscopic consequences, were not considered.52,53
To further elucidate details of the radical cyclizations,
transition states were computed (Figure 6). Phenylethyl radical
24b cyclizes to spirocycle 25b via transition state 29 wherein
the partially formed bond is 1.809 Å. Previous calculations on
this transition state predicted the partially formed bond to be
1.847 Å54 and 1.804 Å.55 Transition state 29 is 16.3 kcal/mol
higher in energy than phenylethyl radical 24b (Figure 7).
Chromium-bound phenylethyl radical 26b cyclizes to complexed
spirocycle 27b via transition state 30 where the partially formed
bond is 1.834 Å long. This transition state is 16.1 kcal/mol
higher in energy than compound 26b. A second transition state,
31, was also found, in which the partially formed bond is endo
to the chromium tricarbonyl moiety. This was much higher in
energy, 21.6 kcal/mol higher than compound 26b. Transition
Intramolecular Reactions. The next reactions to be consid-
ered computationally were the cyclizations of phenylethyl
intermediates, forming spirocyclic products. Optimized geom-
(51) For reviews on protonation of metal centers versus ligand sites,
see: (a) Henderson, R. A. Angew. Chem., Int. Ed. Engl. 1996, 35, 946-
967. (b) Kristja´nsdo´ttir, S. S.; Norton, J. R. In Transition Metal Hydrides;
Dedien, A., Ed.; VCH: Weinheim, 1992; pp 309-359.
(47) Raos, G.; Astorri, L.; Raimondi, M.; Cooper, D. L.; Gerratt, J.;
Karadakov, P. B. J. Phys. Chem. A 1997, 101, 2886-2892.
(48) Reed, C. A.; Fackler, N. L. P.; Kim, K.-C.; Stasko, D.; Evans, D.
R. J. Am. Chem. Soc. 1999, 121, 6314-6315.
(52) Brookhart, M.; Green, M. L. H. J. Organomet. Chem. 1983, 250,
395-408.
(53) For reports on protonation at the metal of electron-rich (η6-arene)-
molybdenum-(phosphine)3 complexes, see: (a) Kowalski, A. S.; Ashby,
M. T. J. Am. Chem. Soc. 1995, 117, 12639-12640. (b) Ashby, M. T.;
Asirvatham, V. S.; Kowalski, A. S.; Khan, M. A. Organometallics 1999,
18, 5004-5016. (c) Asirvathan, V. S.; Gruhn, N. E.; Lichtenberger, D. L.;
Ashby, M. T. Organometallics 2000, 19, 2215-2227.
(49) (a) Petersen, J. L.; Brown, R. K.; Williams, J. M. Inorg. Chem.
1981, 20, 158-165. (b) Girolami, G. S.; Salt, J. E.; Wilkinson, G.; Thornton-
Pett, M.; Hursthouse, M. B. J. Am. Chem. Soc. 1983, 105, 5954-5956. (c)
Heintz, R. A.; Haggerty, B. S.; Wan, H.; Rheingold, A. L.; Theopold, K.
H. Angew. Chem., Int. Ed. Engl. 1992, 31, 1077-1079. (d) Jagirdar, B. R.;
Palmer, R.; Klabunde, K. J.; Radonovich, L. J. Inorg. Chem. 1995, 34,
278-283.
(54) Yamabe, S. Chem. Lett. 1989, 1523-1526.
(55) Leardini, R.; Nanni, D.; Pedulli, G. F.; Tundo, A.; Zanardi, G.;
Foresti, E.; Palmieri, P. J. Am. Chem. Soc. 1989, 111, 7723-7732.
(56) Smith, W. B. J. Phys. Org. Chem. 1995, 8, 171-7732.
(50) Noh, S. K.; Sendlinger, S. C.; Janiak, C.; Theopold, K. H. J. Am.
Chem. Soc. 1989, 111, 9127-9129.

ReactiVity of (η6-Arene)tricarbonylchromium Complexes
J. Am. Chem. Soc., Vol. 123, No. 21, 2001 4911
Figure 4. (U)B3LYP/LANL2DZ-optimized structures of starting materials and products for phenylethyl cyclizations. Selected distances are in Å.
Table 2. Heats of Reaction for Phenylethyl Cyclizatonsa
Figure 5. Electrostatic potential surfaces of free and Cr(CO)3-
complexed ions and constant spin density surfaces for free and Cr(CO)3-
complexed radicals formed from phenylethyl cyclizations.
Unexpectedly, no fully optimized structure could be found
for a chromium-bound cation with a spirocyclic geometry
a All values are in kcal/mol. Energies were calculated using the
DZVP2+ basis set. Values in parentheses are from the LANL2DZ basis
analogous to 27a and 27b. Instead, the primary cation adds to
set. Structure 28c is the optimized cation geometry.
the chromium atom, forming chromacyclic structure 28c.
state 31 does not connect intermediate 27b to primary radical
26b; instead it is the transition state connecting 27b to primary
radical 32, a rotamer of 26b, that is 2.2 kcal/mol higher in energy
(Figure 7). This corresponds to the ring-opening step in our
radical rearrangement reaction and has important consequences
in interpreting our competition results (vide infra).
Formation of a Cr-C bond is indicated by pyramidalization of
this carbon and a Cr-C distance of 2.446 Å. This is longer
than known strain-free chromium-alkyl bonds, which have been
measured from 2.054 to 2.206 Å.57 The Cipso-CR bond exocyclic
to the arene ring is inclined toward the metal by 19.2° from the
arene plane to facilitate direct interaction of the â carbon with

4912 J. Am. Chem. Soc., Vol. 123, No. 21, 2001
Merlic et al.
Scheme 3. Heats of Reaction for Rearrangements of
Cationic Complexesa
a All values are in kcal/mol. Energies were calculated using the
DZVP2+ basis set. Values in parentheses are from the LANL2DZ basis
set.
Figure 6. UB3LYP/LANL2DZ-optimized geometries of transition
states and product for radical rearrangements. Selected distances are
in Å.
1.625 Å between each of the methylenes and the arene ring,
compared to a distance of 1.653 Å in our structure. For
comparison with the cyclic anionic and radical cases, a
spirocyclic chromium-bound cation (27c) was obtained by
constraining the interior angles of the three-membered ring and
allowing the rest of the molecule to relax. This structure was
11.6 kcal/mol less favorable than fully optimized chromacycle
28c. Cyclization to this structure would still be exothermic by
20.7 kcal/mol, but would be 6.2 kcal/mol less favorable than
cyclization in the noncomplexed case. This result is important
with regard to the debate on the mechanism of electrophilic
additions to the arene complexes (vide infra).
The preference of the cation to exist as chromacycle 28c led
us to question whether this was a general phenomenon for
carbocations interacting with chromium arene complexes.
Therefore, a series of comparisons was made as shown in
Scheme 3. For all cases except complex 27c (vide supra), fully
optimized minima were found for both starting material and
product. In each case, shifting a carbon from the hexadienyl
ring to the chromium center is strongly favored. These carbon
migrations convert formally 16-electron complexes to more
stable closed shell 18-electron complexes. The agostic interac-
tion in 23c discussed above also demonstrates the preference
of the electron-deficient metal to interact with another ligand
as an additional formal source of two electrons, though the
absence of a complete hydrogen shift may reflect bond energies.
The first suggestion that chromium can interact with remote
cations (γ position) came from calculations on cations 33c and
34c. Closure of the primary cation onto the arene ring forming
spirocycle 33c and onto the chromium atom forming chroma-
cycle 34c are both exothermic, but 34c is more favorable by
26.1 kcal/mol. Structure 34c, like 28c and the known benzylic
cation complex,27c,d,60 has closed shell electron configurations
at both chromium and the reactive carbon. The calculated
chromium-CH2 bond length of 34c is 2.365 Å, which is close
to known chromium-alkyl bond lengths which range from
2.054 to 2.206 Å.57
Figure 7. Radical reaction profiles (all energies in kcal/mol).
the chromium. Bond lengths within the six-membered ring of
28c indicate that it is best viewed as a normal delocalized
aromatic system (in contrast to 27a and 27b which have longer
bonds to the sp3 carbon). In both the free and metal-bound
phenylethyl cations (24c and 26c), the exocyclic Cipso-CR-Câ
angle was constrained to 110°. We were unable to fully optimize
minima for primary cations, which instead prefer to cyclize.
Cyclization to chromacycle 28c from acyclic cation 26c is
exothermic by 32.3 kcal/mol and is more favorable than
cyclization of noncomplexed cation 24c to 25c by 5.4 kcal/
mol. The parent phenonium cation (25c) has been studied
computationally by several researchers and continues to be
explored using different methods and levels of theory.58 Sieber
and Schleyer reviewed prior literature reports on 25c and also
reported an MP2/6-31G* optimization.59 The distance between
the two methylene units is 1.426 Å; our computed structure
has a distance of 1.454 Å. They further report a distance of
(57) (a) Fryzuk, M. D.; Leznoff, D. B.; Rettig, S. J. Organometallics
1997, 16, 5116-5119. (b) Heintz, R. A.; Leelasubcharoen, S.; Liable-Sands,
L. M.; Rheingold, A. L.; Theopold, K. H. Organometallics 1998, 17, 5477-
5485. (c) Fryzuk, M. D.; Leznoff, D. B.; Rettig, S. J.; Young, V. G., Jr. J.
Chem. Soc., Dalton Trans. 1999, 147-154.
(58) del Rio, E.; Menendez, M. I.; Lopez, R.; Sordo, T. L. J. Phys. Chem.
A 2000, 104, 5568-5571.
(59) Sieber, S.; Schleyer, P. v. R.; Gauss, J. J. Am. Chem. Soc. 1993,
115, 6987-6988.
To compare the relative energetics of chromacycles with
pseudo four- or five-membered rings, the constitutional isomers
39c and 40c were examined (eq 7). Isomer 40c, with the less
(60) Jaouen, G.; Top, S.; McGlinchey, M. J. J. Organomet. Chem. 1980,
195, C5-C8 and references therein.

ReactiVity of (η6-Arene)tricarbonylchromium Complexes
J. Am. Chem. Soc., Vol. 123, No. 21, 2001 4913
Figure 8. Cation substitution pattern (distances in Å).
systems, but a weak chromium-carbon interaction in tertiary
compound 41c. First, the Cr-C distance is 2.365 Å in primary
cation 34c and 2.558 Å in secondary cation 40c, but a much
longer 3.492 Å in tertiary cation 41c. Second, as the chromium
donates more electron density to the carbocation, back-bonding
to the CO ligands should diminish, and the distance from the
metal to the CO’s should increase. Indeed, while the average
trans Cr-CCO distance is 1.856 Å in the acyclic complexes,
the distances are 1.879, 1.870, and 1.854 Å in 34c, 40c, and
41c, respectively. Third, the extent of Cr-C interaction is
evidenced by the pyramidalization of the coordinated carbon.
For a simple chromium-carbon bond with a sp3 hybridized
carbon, the sum of the bond angles to the organic substitutents
should be 328.5° (3 × 109.5°) while the sum of such bond
angles would be 360.0° for a noninteracting sp2 hybridized
carbocation. The values for 34c, 40c, and 41c are 335.0°, 340.1°,
and 356.2°, respectively. Finally, the conclusions based upon
geometric considerations are supported by the energies of
cyclization from the acyclic cations. Cyclization of 42c to 34c
and 43c to 40c is highly exothermic at 36.6 and 19.8 kcal/mol,
respectively, while tertiary cation 44c is only 3.7 kcal/mol less
stable than 41c.
strainedpseudo five-membered ring size, is lower in energy by
6.5 kcal/mol. This energy difference is intuitively assigned to
the ring strain and several geometric parameters confirm this.
The Cipso-CR-Câ angle about the methylene connecting the
arene ring to the cationic carbon in 39c is 99.9°, compared to
an angle of 108.2° for the benzylic methylene in 40c. In addition
to the greater Baeyer (angle) strain, 39c also has greater Pitzer
(eclipsing) strain. While complex 40c has substituents in a
staggered conformation, the substituents in 39c are close to
eclipsed. Most importantly, the ethyl group in 39c eclipses one
of the CO ligands of the Cr(CO)3 unit. These geometric features
lead to a Cr-C distance of 2.791 Å in 39c, which shortens to
2.558 Å in 40c. From Benson’s strain corrections for cyclo-
alkanes,61 the energy difference of 6.5 kcal/mol between 39c
and 40c is closer to the 6.3 kcal/mol difference between
cyclopentane and cyclohexane than it is to the 19.9 kcal/mol
difference between cyclobutane and cyclopentane. From the
perspective of ring strain, the chromium phenyl fragment is
roughly equivalent to three methylene groups. Overall, calcula-
tions suggest that interaction of chromium with remote γ cations
is not only a favorable process, but more favorable than
interaction with a â postion.
To explain prior experimental results and guide future studies,
we desired a clearer understanding of the effect of alkyl
substitution on chromium stabilization of the acyclic positive
charge. Compound 34c has chromium bonded to a primary
carbon, while 40c involves a secondary carbon, so fully
optimized structure 41c was calculated in which a tertiary carbon
coordinates to the metal (Figure 8). Acyclic cations without
chromium-carbon bonds were calculated by fixing dihedral
angles about the Cipso-CR and CR-Câ bonds. In the case of
primary cation 42c, it was also necessary to constrain the CR-
Câ bond length and the two Câ-H bond lengths to prevent
cationic rearrangement upon optimization.
Despite evidence for Cr-C interactions in cationic complexes,
we found that direct interaction of the carbon with the chromium
center is not favored in the anionic and radical cases (Table
3).62 We were unable to find optimized minima for anions or
radicals with geometries represented as 28a, 28b, 38a, and 38b.
Instead, geometry optimizations of such species always led to
cleavage of the chromium-carbon bond. Single point energies
were calculated for these anions and radicals by using the
geometries of 28c and 38c and were found to be significantly
higher in energy (up to 99 kcal/mol) than the energies of 27a,
27b, 37a, and 37b.
The parent chromacyclic cation 28c cannot undergo further
rearrangement, as ring opening would lead to a primary
carbocation. To understand the experimentally observed cationic
rearrangment of 1 to 2, we computed the structure of 45, the
chromacyclic intermediate that is actually formed under the
experimental conditions (Figure 9). Appending a phenyl ring
Several geometric features point to relatively strong chro-
mium-carbon interactions in the primary and secondary
(62) Recently, researchers have invoked radical addition to the chromium
atom of bis(benzene)chromium, but this system does not appear to be
comparable with (benzene)tricarbonylchromium systems. Samuel, E.;
Caurant, D.; Gourier, D.; Elschenbroich, C.; Agbaria, K. J. Am. Chem. Soc.
1998, 120, 8088-8092.
(61) Benson, S. W.; Cruickshank, F. R.; Golden, D. M.; Haugen, G. R.;
O’Neal, H. E.; Rodgers, A. S.; Shaw, R.; Walsh, R. Chem. ReV. 1969, 69,
279-324.

4914 J. Am. Chem. Soc., Vol. 123, No. 21, 2001
Table 3. Heats of Reaction for Rearrangementsa
Merlic et al.
Discussion
Anion. The selectivity of the anionic rearrangement via
addition to the complexed ring is expected from a simple
electronics argument based on the known chemistry of chro-
mium-arene complexes.1 The fact that the rearrangement
proceeds exclusively through addition to the complexed ring is
further verified by the control reaction, where we only observe
net reduction of the starting iodide and no addition into the free
arene. Anion addition to simple phenyl rings is known to require
elevated temperatures and prolonged reaction times.41,42
Our computational study of anionic addition neatly fits in
with the current and reported experimental data, in particular
the X-ray crystal structure of an anion adduct reported by
Semmelhack,45 but adds information on the energy changes. As
a general rule, the intramolecular reactions studied are less
favorable than the intermolecular reactions due to the introduc-
tion of strain in the cyclopropyl rings. In the anionic phenylethyl
cyclization, chromium complexation makes the difference
between an endothermic reaction and an exothermic reaction.
This has implications for the competitive cyclization depicted
in Scheme 1. We should expect the exothermic anionic
cyclization to form intermediate A instead of the endothermic
cyclization to form B, and indeed, this is what we observe
experimentally. Overall, the calculations on anionic addition
provide a baseline for evaluating the methodology and for
comparison with the cation and radical results. Further, it is
the necessary prelude to the next level of analysis of predicting
the regiochemistry of addition to unsymmetrical arene com-
plexes.63
a All values are in kcal/mol. Energies were calculated using the
DZVP2+ basis set. Values in parentheses are from the LANL2DZ basis
set.
Radical. Interestingly, we detected no enhancement for
addition of radicals into the complexed arene in our intra-
molecular competition experiment. This is initially surprising
given that chromium complexation activates aryl rings toward
other additions. Further, Schmalz and co-workers demonstrated
that radical addition to metal-bound arenes can in fact occur in
a facially specific manner.26 However, they never examined a
substrate where the intermediate is given a choice between
radical addition to a free or chromium-complexed aryl ring. In
contrast to the intramolecular competition, the intermolecular
competition experiment yielded complete selectivity for radical
addition to the complexed arene ring. In fact, the experiment
employing a 100-fold excess of benzene-d6 found a >99.9:<0.1
ratio of nondeuterated and hexadeuterated products. Ignoring a
possible secondary kinetic isotope effect, that translates into a
greater than 100 000:1 ratio of relative reactivities for complexed
and noncomplexed benzene. These seemingly disparate results
are actually nicely explained by mechanistic rationale developed
from the computational results.
Figure 9. B3LYP/LANL2DZ-optimized geometries of stationary points
in the cationic rearrangement mechanism. Selected distances are in Å.
The computational results for radical cyclization were initially
troubling. Formations of both 25b and 27b are endothermic, so
as neither radical cyclization in Scheme 1 is favorable, the
radical reaction is expected to be fairly unproductive. Indeed,
the radical reaction gives a low conversion to rearranged
products. The problem is that the calculated energies of the
radical spirocyclic intermediates do not explain the observed
experimental product ratios. In the intramolecular radical
competition experiments we observe varying ratios of roughly
1 to 1 up to a maximum of 1 to 4 for products 2 and 3,
respectively. The best ratio would correspond to a 0.8 kcal/mol
preference for addition to the noncomplexed arene, which is in
contrast to our computed preference of cyclization to complexed
does not greatly perturb the structure of 28c. For instance, the
chromium-methylene bond distance in intermediate 45 is 2.436
Å, compared to a distance of 2.446 Å in parent chromacycle
28c. We further found transition state 46 leading smoothly to
benzylic cation 47, which can be quenched to afford observed
product 2. Animation of this transition state demonstrates that
it is a concerted shift of the two-carbon bridge, with a slight
rotation of the phenyl ring to bring it into conjugation with the
forming carbocation. This transition state is 7.0 kcal/mol higher
in energy than chromacyclic intermediate 45 and the overall
reaction is exothermic by 3.7 kcal/mol. Thus, a cationic
intermediate equivalent to A in Scheme 1 was not found in the
computed reaction path, and related structure 46 was actually
calculated to be a transition state, not an intermediate.
(63) Semmelhack, M. F. Nucleophilic Addition to Arene-Metal Com-
plexes. In ComprehensiVe Organic Synthesis; Trost, B. M., Fleming, I.,
Eds.; Pergamon Press: Oxford, England, 1991; Vol. 4, pp 517-549.

ReactiVity of (η6-Arene)tricarbonylchromium Complexes
J. Am. Chem. Soc., Vol. 123, No. 21, 2001 4915
intermediate 27b by 0.9 kcal/mol. This discrepancy demands
that we consider the transition states for cyclization rather than
simply the intermediates (Figure 6). An examination of reaction
energy profiles for the two radical cyclizations tells the story
(Figure 7). Cyclization of phenylethyl radical 24b to intermedi-
ate 25b proceeds with a barrier of 16.3 kcal/mol via transition
state 29. Since the two faces of the six-membered ring are
equivalent, ring opening to form the phenylethyl radical proceeds
via an identical transition state. In the chromium-bound arene,
though, the two faces of the six-membered ring are different.
Cyclization of complexed phenylethyl radical 26b to intermedi-
ate 27b is expected and calculated to occur with radical approach
anti to the chromium tricarbonyl moiety with a barrier of 16.1
kcal/mol. In the experimental system, rearrangement products
would have to come from cleavage of the endo bond of the
three-membered ring, via transition state 31. This step for
intermediate 27b in the phenethyl model has a barrier of 21.6
kcal/mol relative to the starting material, making this the rate-
limiting step for the overall reaction. Therefore, while formation
of complexed intermediate 27b is somewhat preferred relative
to 25b, it is noncomplexed intermediate 25b that has a lower
energy pathway leading to rearranged product. To further refine
the reaction profile, the fact that the experimental system is
substituted with a phenyl ring must be taken into account. We
estimate that the phenyl substituent should decrease the barrier
for ring opening via complexed transition state 31 by 4.6 kcal/
mol.64 Such a correction would make the highest point along
the metal-bound pathway higher than the highest point along
the noncomplexed pathway by <1 kcal/mol, correlating well
with our experimental observations. Thus, while radical addition
to the complexed arene ring is favored kinetically, the observed
product ratio resulted from an equilibration process.
These results have important consequences for the synthetic
design of radical addition reactions to chromium-complexed
arenes. Since loss of an endo radical substituent via a pathway
similar to transition state 31 is disfavored, then radical ipso
substitutions of chromium arene complexes should not be high-
yielding reactions. Instead, a different mechanistic pathway
should be pursued to take advantage of the favored radical
addition to a chromium-complexed aryl ring (27b vs 25b). For
example, electron transfer to intermediate 27b would lead to
stable 18-electron anionic hexadienylchromium complex 27a.
Thus, following radical addition steps with anionic chemistry
should allow one to observe net favorable radical addition
reactions to chromium-complexed aryl rings. Indeed, the few
radical addition reactions of arenechromium complexes reported
by Schmalz26 fortunately take advantage of this exact pathway.
With this mechanistic rationale, it should be possible to design
other radical transformations of arenechromium complexes.
Indeed, the intermolecular radical competition results fit
perfectly within this mechanistic framework. Reduction of
acetone with SmI2 generates a ketyl radical that adds to the
benzene complex anti to the chromium forming a 17-electron
hexadienyl complex (Scheme 4). Next capture by a second SmI2
generates a stable 18-electron hexadienyl complex. Subsequent
protonation at the metal center and reductive elimination
isomerization yields the final products. These latter steps are
in direct analogy to the known arene complex chemistry of
nucleophilic addition followed by protonation as first reported
Scheme 4. Proposed Mechanism for Intermolecular Radical
Addition
by Semmelhack65 and mechanistically elucidated by Ku¨ndig.66
While ketyl-olefin coupling reactions represent one of the most
widely studied SmI2-promoted radical processes,39c,67 ketyl-
arene coupling reactions are rare.68 Use of arene chromium
complexes can now dramatically broaden the scope of these
coupling reactions. The reactivity enhancement by a factor of
100 000 due to complexation with Cr(CO)3 will allow access
to many otherwise unfavorable reactions and will allow facile
discrimination between aryl groups within the same substrate.
The calculated and observed selectivity for radical addition
to complexed arenes is readily understood from a molecular
orbital analysis. The SOMO of nucleophilic radicals such as
ketyl radicals will interact with the arene LUMO. Since
complexation by a tricarbonylchromium moiety lowers the arene
LUMO energy, complexed arenes are more reactive than free
arenes toward radical addition. This picture is in agreement with
the few reported SmI2-promoted ketone-arene coupling re-
actions where intermolecular68b,e and even intramolecular68d
reactions require electron-withdrawing substituents on the arene
ring.
Studies are in progress exploring the range of radical additions
to complexed arenes and regioselectivity for substituted arenes.
As with anionic addition reactions, computational studies based
on the results reported herein will prove useful for explaining
and predicting the regiochemical outcome of the radical
reactions.
Cation. The high selectivity found in the cationic rearrange-
ment case comes as a surprise because the opposite regioisomer
would be expected electronically since the Cr(CO)3 moiety is
electron withdrawing. Indeed, reports on electrophilic aromatic
substitution reactions have found arene tricarbonylchromium
(65) (a) Semmelhack, M. F.; Hall, H. T. J. Am. Chem. Soc. 1974, 96,
7092-7094. (b) Semmelhack, M. F.; Hall, H. T.; Yoshifuji, M. J. Am. Chem.
Soc. 1976, 98, 6387-6389. (c) Semmelhack, M. F.; Cark, G. R.; Garcia, J.
L.; Harrison, J. J.; Thebtaranonth, Y.; Wulff, W. Tetrahedron 1981, 37,
3957-3065.
(66) Ku¨ndig, E. P.; Amurrio, D.; Bernardinelli, G.; Chowdhury, R.
Organometallics 1993, 12, 4275-4277.
(64) This correction comes from calculating the difference in activation
energies between the ring opening of cyclopropylcarbinyl radical and
2-phenylcyclopropylcarbinyl radical. The rate of each of these reactions
has been measured, and the ∆Ea can be computed using the Arrenhius
equation. See: (a) Beckwith, A. L. J.; Moad, G. J. Chem. Soc., Perkin
Trans. 2 1980, 1473-1482. (b) Newcomb, M.; Johnson, C. C.; Manek, M.
B.; Varick, T. R. J. Am. Chem. Soc. 1992, 114, 10915-10921.
(67) Molander, G. A.; Harris, C. R. Chem. ReV. 1996, 96, 307-338.
(68) (a) Williams, D. B. G.; Blann, K.; Holzapfel, C. W. J. Org. Chem.
2000, 65, 2834-2836. (b) Shiue, J.-S.; Lin, M.-H.; Fang, J.-M. J. Org.
Chem. 1997, 62, 4643-4649. (c) Maury, O.; Villiers, C.; Ephritikhine, M.
Tetrahedron Lett. 1997, 37, 6591-6594. (d) Kise, N.; Suzumoto, T.; Shono,
T. J. Org. Chem. 1994, 59, 1407-1413. (e) Shiue, J.-S.; Lin, C.-C.; Fang,
J.-M. Tetrahedron Lett. 1993, 34, 335-338.

4916 J. Am. Chem. Soc., Vol. 123, No. 21, 2001
Merlic et al.
Scheme 5. Three Modes of Electrophilic Addition to Arene
Complexes
tional predictions. First, all three modes are feasible, so the
specific pathway will likely depend on the substituents. Second,
there are stereochemical implications (vide infra).
Thirty years ago, Bly and co-workers considered an inter-
mediate such as 28c in their study of the solvolyses of
methanesulfonates â to chromium-complexed arenes.20 Upon
further kinetic studies, they rejected direct chromium involve-
ment with the cation (as in 28c), reasoning that it did not explain
all of their rate data. Instead, they argued that the rate
enhancement by chromium was due to an electrostatic effect,
with the partially negative end of the arene-to-chromium dipole
stabilizing the forming positive charge as solvolysis proceeds.20f
The current computational results favoring direct carbon-
chromium interaction call for a reconsideration of their conclu-
sion. Chromium-complexed phenonium 27c is a 16-electron
metal complex, while chromacycle 28c is a more favorable 18-
electron species. Formation of the chromium-complexed phe-
nonium structure also involves breaking the aromatic π system,69
an energetic price that does not have to be paid in the formation
of 28c.
Scheme 6. Energetics of Methyl Cation Addition to
(Benzene)tricarbonylchromiuma
In addition, consideration of the frontier molecular orbitals
of the fragments involved also points toward the formation of
the chromacycle. Figure 10 depicts the combinations of the
HOMO of a chromium tricarbonyl fragment with the LUMO’s
of the phenonium ion (25c) and phenylethyl cation (24c) to form
the complexed phenonium (27c) and chromacycle species (28c),
respectively. A cursory examination of the LUMO of the
phenylethyl cation, which can be considered as largely a p orbital
on the primary cation, shows that this orbital has a greater spatial
overlap with the chromium fragment HOMO than does the
phenonium LUMO. Shifting the two ethyl carbons down toward
the metal increases this overlap still further. The calculated
energy of the phenylethyl cation LUMO is lower than the energy
of the phenonium LUMO. This leads to a more favorable
energetic matching between this LUMO and the chromium
fragment HOMO. Both of these effects lead to a greater
stabilization of the HOMO of the chromacycle relative to the
complexed phenonium, favoring the chromacyclic structure.
The chromacycle-type intermediate provides a clear mecha-
nistic understanding for an intriguing stereochemical experiment
reported by Bly and co-workers.20a Solvolysis of L-threo-3-
[(phenyl)tricarbonylchromium]-2-butyl methanesulfonate oc-
curred with net retention of stereochemistry. Invoking partici-
pation by chromium, anchimeric assistance to the ionization
would generate a chromacyclic intermediate with inversion of
the carbon stereochemistry. Subsequent nucleophilic displace-
ment of chromium with inversion at the carbon would lead to
net retention of stereochemistry. Although Bly considered such
a mechanism, he was unable to differentiate it from an arene
participation pathway. The large calculated energy difference
between the two pathways strongly argues for the chromium
participation pathway.
The rearrangements examined in Scheme 3 demonstrate that
this direct interaction of cations with the metal center can be
preferred.22,70 In all cases this can be rationalized by the
preference for transition metals to form 18-electron complexes.
In the cyclic cases, shifting the carbon to the metal center also
involves release of ring strain, increasing the exothermicity of
these rearrangements. Direct interaction of the chromium with
distant cations, as in the formation of 34c, suggests that
chromium complexation of arenes could influence solvolysis
a All values are in kcal/mol. Energies were calculated using the
DZVP2+ basis set.
complexes to be less reactive than the respective free
arenes.8-10,12-16 Here our calculations point to an important
alternate mechanistic pathway.
There are three possible modes of electrophilic addition to
arene complexes: addition to the metal atom, endo addition to
the arene ring, and exo addition to the arene ring (Scheme 5).
Previous researchers have considered these pathways, but never
together in a systematic treatment examining all three. Net
electrophilic aromatic substitution arises from proton loss from
the ring adducts while the metal-bound electrophile (e.g., 48c)
would first have to rearrange to the endo adduct (49c).
In the specific case where the electrophile is a methyl cation,
addition in all three modes is calculated to be exothermic, but
endo addition is less favored by >9 kcal/mol (Scheme 6).
Surprisingly, the exo adduct is most favorable of the three due
to an agostic interaction with the endo proton. In the absence
of the agostic interaction, as in the geminal dimethyl cation 35C,
the metal-bound electrophile, 36C, is strongly favored. Similarly,
with methylene groups as the substitutents (27c and 33c vs 28c
and 34c), the metal-coordinated cation is again highly favored.
There are important chemical consequences of these computa-
(69) Cr(CO)3 complexation does not reduce the aromaticity of benzene,
see: ref 27b.
(70) Merlic, C. A.; Miller, M. M. Organometallics 2001, 20, 373-375.

ReactiVity of (η6-Arene)tricarbonylchromium Complexes
J. Am. Chem. Soc., Vol. 123, No. 21, 2001 4917
Figure 10. Frontier orbital interactions showing overlap between the HOMO of Cr(CO)3 and the LUMOs of cations 25c and 24c to form chromium
complexes 27c and 28c, respectively.
reactions beyond the presently known R- and â-positions,1b,23,71
to the γ-position and perhaps beyond. Indeed, the computational
comparison between pseudo four-membered ring 39c and
pseudo five-membered ring 40c indicates that the larger ring is
actually favored (eq 7). We recently reported experimental
evidence of such a phenomenon (eqs 8 and 9).70 Solvolysis of
(S)-52 with 2 equivs of sodium acetate in acetic acid at 90 °C
gave acetate (S)-53 (eq 8) and analysis72 showed that substitution
and radicals do not interact directly with the metal center. Again,
this can be explained in terms of the 18-electron rule. Direct
anion or radical addition to the metal would lead to a 20- or
19-electron species, respectively.
Friedel-Crafts acylations most likely proceed via exo addi-
tion. All the examples reported would benefit from an agostic
interaction with the endo proton in the complexed Meisenheimer
intermediate and none would suffer from ring strain. Then the
observed reduced reactivity of complexed arenes would be a
manifestation of electron withdrawal by the tricarbonyl-
chromium moiety. This mechanistic rationale is analogous to
ferrocene chemistry wherein small electrophiles add to iron, but
acylation occurs anti to iron.74
The computational and experimental results for cationic and
radical reactions call for revisions to our initial scheme for
competitive cyclization (Scheme 1 vs Scheme 7). It is apparent
that the anionic rearrangement proceeds along pathway I,
starting with nucleophilic addition to the complexed arene.
Radical rearrangement is slightly more complex. Initial radical
addition via pathway I addition is calculated to be favored.
However, the highly disfavored endo homolytic bond cleavage
in the test substrate suggests a reversible addition leading to
equilibration via pathway II to intermediate B that undergoes
homolytic bond cleavage more readily, leading ultimately to
favored product 3. Finally, the cation proceeds through an
intriguing mechanism, pathway IV. Initial cyclization is not to
either arene ring, but to the metal center through a d orbital
interaction, forming chromacyclic intermediate C. This inter-
mediate rearranges via transition state D leading to product 2
upon hydride quenching. This metal-assisted pathway leading
to rearranged product is also supported by our control reaction
using 1,1-dideuterio-1-iodo-2,2-diphenylethane where we see
a complete recovery of the starting iodide under the same
reaction conditions, consistent with the idea that the cation
rearrangement proceeds via neighboring group participation with
formation of the metallacycle. We also computationally exam-
ined a pathway leading from chromacycle C toward product 3
via a 1,2-phenyl shift in the chromacycle,75 but found it to be
significantly higher in energy.
occurred with 74% net retention of configuration, employing
the gem-dialkyl effect73 to enhance the chromium participation.
In contrast to the noncomplexed control reaction (eq 9), we
propose that the 74% of substitution with retention in (S)-52
occurs by ionization with inversion using neighboring group
participation from chromium followed by displacement of
chromium by the nucleophile, also with inversion. Thus,
chromium does participate in the substitution reactions at remote
centers. Calculations also point to a limit to metal-cation
interactions in the case of tertiary carbons. Tertiary carbocations
are not expected to interact strongly with chromium atoms and
this is supported by the experimental work of Fry and co-
workers.23 The comparisons in Table 3 demonstrate that anions
(71) Hegedus, L. S. Transition Metals in Synthesis of Complex Organic
Molecules; University Science Books: Mill Valley, CA, 1994; Chapter 10,
pp 307-333.
As noted for radicals, the results from the cationic processes
have important implications for synthetic applications. First, the
(72) The alcohol resulting from decomplexation and hydrolysis was
analyzed by 19F NMR of the Mosher’s ester.
(73) (a) Jung, M. E. Synlett 1999, 843-846. (b) Jung, M. E.; Gervay, J.
J. Am. Chem. Soc. 1991, 113, 224-232.
(74) Cunningham, A. F., Jr. Organometallics 1994, 13, 2480-2485.
(75) Transition state 56. Included in the Supporting Information.

4918 J. Am. Chem. Soc., Vol. 123, No. 21, 2001
Merlic et al.
Scheme 7. Rearrangements of Chromium-Complexed 2,2-Diphenylethyl Anion, Cation, and Radical
direct carbon-chromium interactions illustrated in Scheme 3
suggest that there might be a general reaction rate enhancement
(anchimeric assistance) for appropriately placed leaving groups
beyond these reported for the benzylic1f and homobenzylic20,23
positions. Second, there should be stereochemical consequences
for neighboring group participation by chromium. Indeed,
substitution with retention is well-known at the benzylic
position1f and a few examples at the homobenzylic position have
been reported.20 However, this stereochemical control might
extend to the third carbon out from the aryl ring70 and the results
from eqs 8 and 9 provide the first evidence for this.70 Finally,
there might be new manifolds of chemical reactivity to exploit,
such as Friedel-Crafts alkylations,76 based on electrophilic
addition to the chromium atom.
premise found (benzene)tricarbonylchromium to be more reac-
tive than benzene by a factor of more than 100 000:1, in strong
agreement with the calculated energy differences. Finally,
complexation leads to three mechanistic pathways for addition
of electrophiles to arenes. Electrophilic addition to the ring endo
relative to the metal is calculated to be significantly less
favorable than addition exo to the ring or directly to the metal.
Instead, electrophilic aromatic substitution in arene complexes
is expected to occur via initial exo addition of the electrophile
to the arene ring or initial addition to the metal followed by
endo transfer to the arene ring. In an intramolecular experiment
employing a primary carbon electrophile, the latter mechanism
predominated in contrast with known intermolecular reactions
with acylium ions. Calculations predict strong carbon-
chromium bonding for primary and secondary, but not tertiary,
cations and predict interactions with electrophilic sites at the γ
position. Efforts to exploit the synthetic potential of radical and
carbocation additions to arene complexes are currently underway
in our laboratories.70
Conclusions
In conclusion, we examined computationally and experimen-
tally the reactivity of tricarbonylchromium-complexed aryl rings
toward direct addition of reactive intermediates and found that
chromium complexation can enhance the addition of anions,
cations, and radicals to arene rings. Anion addition proceeds
via a chromium-complexed analogue of the Meisenheimer
intermediate and calculations provide structural details, charge
distributions, and energy changes for the reaction. As expected,
charge is localized on the chromium atom and addition is more
favorable than in the noncomplexed arene by more than 30 kcal/
mol. Likewise, complexation enhances direct addition of radicals
to arene rings by about 8 kcal/mol, depending on the system,
creating 17-electron intermediates structurally similar to anionic
addition. Subsequent, homolytic cleavage of a bond endo to
the metal is highly unfavorable due to large structural changes
required. Instead, reduction of the intermediate radical complex
resulting from radical addition results in a net favorable reaction.
An intermolecular radical competition reaction designed on this
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 (NCSA). We
also acknowledge Joseph C. Walsh and Dean J. Tantillo for
many helpful discussions and Andrea L. Zechman for prepara-
tion of Cr(CO)3(CH3CN)3.
Supporting Information Available: Full experimental
procedures and spectral characterization for compounds reported
in the text and computational methods and structural data for
compounds 11 through 47 and 56 computed at the B3LYP/
LANL2DZ level (PDF). This material is available free of charge
via the Internet at http://pubs.acs.org.
(76) Conversion of 1 to 2 is but one example.
JA000600Y