Addition of olefins to aromatic ketones catalyzed by Rh(I) olefin complexes

Article abstract of DOI:10.1021/ja990702s

The rhodium bis-olefin complex [C₅Me₅Rh(C₂H₃SiMe₃) ₂], 1, has been shown to be a catalyst for the selective addition of olefins to the ortho position of aromatic ketones. The addition of vinyltrimethylsilane to benzophenone was studied by NMR spectroscopy, which indicated that 1 was the catalyst resting state for this process. This reaction was applied to a series of olefins (allyltrimethylsilane, 1-pentene, norbornene, 2,2′-dimethyl-3-butene, cyclopentene, and vinyl ethyl ether) and aromatic ketones (benzophenone, 4,4′-dimethoxybenzophenone, 3,3′-bis(trifluoromethyl)benzophenone, dibenzosuberone, acetophenone, p-chloroacetophenone, and p-(trifluoromethyl)acetophenone). The dependence of the turnover frequency on substrate concentration was investigated. In the presence of excess ketone the formation of a benzophenone complex, 5, [(C₅Me₅-Rh)₂-η⁴-η ⁴-C₆H₅C(O)Ph] was observed after consumption of olefin. Active catalyst is regenerated upon addition of olefin to 5. On the basis of kinetic experiments and labeling studies using acetophenone-d₈, a mechanism for this catalysis is proposed. In the reaction of acetophenone with vinyltrimethylsilane and 1, the formation of ethylene and the trimethylsilyl enolate of acetophenone was observed together with the alkyl-aryl coupling product. This side reaction is specific for vinyltrimethylsilane.

Full text of DOI:10.1021/ja990702s

6616
J. Am. Chem. Soc. 1999, 121, 6616-6623
Addition of Olefins to Aromatic Ketones Catalyzed by Rh(I) Olefin
Complexes
Christian P. Lenges and Maurice Brookhart*
Contribution from the Department of Chemistry, Venable and Kenan Laboratories,
UniVersity of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599-3290
ReceiVed March 4, 1999
Abstract: The rhodium bis-olefin complex [C5Me5Rh(C2H3SiMe3)2], 1, has been shown to be a catalyst for
the selective addition of olefins to the ortho position of aromatic ketones. The addition of vinyltrimethylsilane
to benzophenone was studied by NMR spectroscopy, which indicated that 1 was the catalyst resting state for
this process. This reaction was applied to a series of olefins (allyltrimethylsilane, 1-pentene, norbornene, 2,2′-
dimethyl-3-butene, cyclopentene, and vinyl ethyl ether) and aromatic ketones (benzophenone, 4,4′-dimethoxy-
benzophenone, 3,3′-bis(trifluoromethyl)benzophenone, dibenzosuberone, acetophenone, p-chloroacetophenone,
and p-(trifluoromethyl)acetophenone). The dependence of the turnover frequency on substrate concentration
was investigated. In the presence of excess ketone the formation of a benzophenone complex, 5, [(C5Me5-
Rh)2-η4-η4-C6H5C(O)Ph] was observed after consumption of olefin. Active catalyst is regenerated upon addition
of olefin to 5. On the basis of kinetic experiments and labeling studies using acetophenone-d8, a mechanism
for this catalysis is proposed. In the reaction of acetophenone with vinyltrimethylsilane and 1, the formation
of ethylene and the trimethylsilyl enolate of acetophenone was observed together with the alkyl-aryl coupling
product. This side reaction is specific for vinyltrimethylsilane.
Introduction
substrates by ruthenium(0) followed by the insertion of an olefin
and reductive elimination to generate a new carbon-carbon
bond has been the focus of recent efforts to develop catalytic
applications. As a major advance in this area, Murai and co-
workers in a series of papers have reported the regioselective
addition of olefins to aromatic carbonyl compounds to generate
ortho-alkylation products mediated by a ruthenium catalyst14-22
(eq 1).
Transition metal catalyzed carbon-carbon bond formation
based on the initial oxidative addition of carbon-halogen or
carbon-heteroatom bonds has been developed to a remarkable
degree of sophistication and applicability with respect to
potential use in organic synthesis. On the other hand, catalytic
processes which allow the formation of carbon-carbon bonds
based on the oxidative addition of carbon-hydrogen bonds are
still rare.1-3 Reactions of low-valent complexes of iridium,
rhodium, and ruthenium traditionally have been used to gain a
detailed mechanistic understanding of transition metal mediated
carbon-hydrogen bond activation. The application of these
results in new catalytic carbon-carbon bond forming reactions
has been mainly developed with ruthenium complexes.4-13 For
example, the activation of sp2-C-H bonds in aromatic or vinylic
The catalysis is performed at 90-140 °C in aromatic solvents
and generally exhibits good turnover numbers. A wide range
of aromatic ketones can be used in this process. The presence
of the carbonyl moiety diverts the metalation to the ortho
position and is believed to result in a chelating interaction that
assists the reductive elimination of product from a RuII aryl-
(1) Shilov, A. E.; Shul’pin, G. B. Chem. ReV. 1997, 97, 2879-2932.
(2) Jones, W. D. In SelectiVe Hydrocarbon ActiVation Principles and
Progress; Davies, J. A., Watson, P. L., Greenberg, A., Liebman, J. F., Eds.;
VCH Publishers: New York, 1990; p 113.
(3) Ryabov, A. D. Chem. ReV. 1990, 90, 403.
(4) Fukuyama, T.; Chatani, N.; Tatsumi, J.; Kakiuchi, F.; Murai, S. J.
Am. Chem. Soc. 1998, 120, 11522.
(5) Naota, T.; Takaya, H.; Murahashi, S.-I. Chem. ReV. 1998, 98, 2599-
2660.
(6) Kosar, W. P.; Jones, W. D. J. Am. Chem. Soc. 1986, 108, 5640.
(7) Hsu, G. C.; Kosar, W. P.; Jones, W. D. Organometallics 1994, 13,
385.
(8) Kondo, T.; Akazome, M.; Tsuji, Y.; Watanabe, Y. J. Org. Chem.
1990, 55, 1286.
(9) Chatani, N.; Ie, Y.; Kakiuchi, F.; Murai, S. J. Org. Chem. 1997, 62,
2604.
(10) Jones, W. D.; Foster, G. P.; Putinas, J. M. J. Am. Chem. Soc. 1987,
109, 5047. This example features a rare iron-catalyzed reaction.
(11) Jones, W. D.; Kosar, W. P. J. Am. Chem. Soc. 1986, 108, 5640.
(12) Hsu, G. C.; Kosar, W. P.; Jones, W. D. Organometallics 1994, 13,
385.
(14) Kakiuchi, F.; Sekine, S.; Tanaka, Y.; Kamatani, A.; Sonoda, M.;
Chatani, N.; Murai, S. Nature 1993, 366, 529.
(15) Murai, S.; Kakiuchi, F.; Sekine, S.; Tanaka, Y.; Kamatani, A.;
Sonoda, M.; Chatani, N. Pure Appl. Chem. 1994, 66, 1527.
(16) Kakiuchi, F.; Sekine, S.; Tanaka, Y.; Kamatani, A.; Sonoda, M.;
Chatani, N.; Murai, S. Bull. Chem. Soc. Jpn. 1995, 68, 62.
(17) Sonoda, M.; Kakiuchi, F.; Chatani, N.; Murai, S. J. Organomet.
Chem. 1995, 504, 151.
(18) Kakiuchi, F.; Yamamoto, Y.; Chatani, N.; Murai, S. Chem. Lett.
1995, 681.
(19) Sonoda, M.; Kakiuchi, F.; Kamatani, A.; Chatani, N.; Murai, S.
Chem. Lett. 1996, 109.
(20) Harris, P. W. R.; Woodgate, P. D. J. Organomet. Chem. 1997, 530,
211.
(21) Harris, P. W. R.; Woodgate, P. D. J. Organomet. Chem. 1996, 506,
339.
(13) Fujii, N.; Kakiuchi, F.; Chatani, N.; Murai, S. Chem. Lett. 1996,
939.
(22) Sonoda, M.; Kakiuchi, F.; Chatani, N.; Murai, S. J. Organomet.
Chem. 1995, 504, 151.
10.1021/ja990702s CCC: $18.00 © 1999 American Chemical Society
Published on Web 06/30/1999

Addition of Olefins to Aromatic Ketones
J. Am. Chem. Soc., Vol. 121, No. 28, 1999 6617
alkyl intermediate.23-27 This chemistry has also been extended
to vinylic C-H bonds in certain R,â-unsaturated ketones and
esters.28,29
Rhodium or iridium complexes, on the other hand, have found
only limited application in catalytic carbon-carbon bond
forming reactions.30-37 While rhodium systems of the type [C5-
Me5RhL] have been crucial in developing an understanding of
the oxidative addition reaction of C-H bonds, they have had
limited application toward the goal of catalytic functionaliza-
tions.38-41
To effect carbon-carbon coupling reactions, the strategy
applied here was to utilize bis-olefin complexes, for example,
[C5Me5Rh(C2H3SiMe3)2], 1, which allow thermal access to the
16-electron species [C5Me5Rh(olefin)]. Oxidative addition of
C-H bonds to this species generates a complex of the type
[C5Me5Rh(olefin)(R)(H)] in which coupling reactions between
olefin and R can be envisioned to occur. This approach is based
on the observation that the parent ethylene complex [C5Me5-
Rh(C2H4)2] will incorporate deuterium into the ethylene ligand
when heated in benzene-d6 (78 °C, 2 h, 25% decrease in
intensity).42,43 This process follows a mechanism that involves
initial ethylene dissociation followed by the reversible activation
of benzene and insertion of olefin which results in scrambling
of the deuterium label from the solvent into the coordinated
ethylene. Increasing the steric bulk of the coordinated olefin in
catalyst 1 results in a decrease of the barrier for olefin loss and
thus more facile generation of the reactive intermediate
[C5Me5RhL]. The increased reactivity of 1 relative to the bis-
ethylene complex is illustrated in a thermolysis experiment in
benzene-d6 that results in selective deuterium incorporation into
coordinated olefin under mild conditions44 (eq 2).
In this paper, we report the use of 1 as a catalyst for the
addition of olefins to aromatic ketones to selectively generate
ortho-alkylated aromatic ketones.
Results and Discussion
A. Alkylation of Benzophenone with Vinyltrimethylsilane
Catalyzed by [C5Me5Rh(C2H3SiMe3)2], 1. Prior to surveying
a range of substrates, we conducted a detailed NMR study of
the addition of vinyltrimethylsilane to benzophenone employing
1 as the catalyst to yield mono- and dialkylated products 3a
and 4a (eq 3). A 1:1 molar ratio of benzophenone and
vinyltrimethylsilane and 1 (0.01 g, 2.3 × 10-5 mol, 5 mol %)
in cyclohexane-d12 was heated in a sealed NMR tube at 120 °C
1
and monitored by H NMR spectroscopy. The formation of
ortho-alkylated product 3a was observed with a turnover
frequency of ca. 0.65 TO/h (Figure 1, Experimental Section).
Characteristic for ketone 3a are resonances for the -CH2-
protons at 2.65 (m, 2H) and 0.84 (m, 2H) and a new -SiMe3
resonance at -0.08 (s, 9H). As in the case of the Murai system,
only alkylation at the ortho position is observed and only the
â-silyl addition product is formed (“anti-Markovnikov” addi-
tion). The only rhodium species observed during catalysis is
complex 1, indicating that this species is the catalyst resting
state.
It was also informative to follow the reaction in toluene-d8
at 120 °C (1:1 molar ratio of substrates, 5 mol % 1). Free
vinyltrimethylsilane showed considerable deuterium incorpora-
tion as the reaction proceeded. After 22 h, both vinylic sites
showed ca. 80% deuteration as judged by use of an internal
standard. At this point 36% conversion of substrates to 3a had
occurred and significant deuteration of the -CH2- groups was
evident (ca. 25%) due to deuteration of the vinylsilane prior to
reaction as well as deuteration of 3a after formation.45 The
deuteration of the olefin and the product 3a are explained with
a reversible H/D exchange sequence as shown in eq 2.44 This
observation is consistent with the assignment of 1 as the catalyst
resting state. Reversible C-H bond activation by [C5Me5Rh-
(C2H3SiMe3)] (in rapid equilibrium with 1), followed by
reversible olefin insertion into Rh-D (see eq 2) must occur on
a faster time scale than productive ketone-olefin coupling.46
Coupling products based on the activation of toluene were not
observed.
(23) Guari, Y.; Sabo-Etienne, S.; Chaudret, B. J. Am. Chem. Soc. 1998,
120, 4228.
(24) Matsubara, T.; Koga, N.; Musaev, D. G.; Morokuma, K. J. Am.
Chem. Soc. 1998, 120, 12092.
(25) In a reaction of benzophenone with RuH2(PPh3)4 an ortho-metalated
ruthenium hydride was isolated in which chelation of the carbonyl group
to ruthenium was observed in the position trans to the hydride ligand, see
ref 27.
(26) Halpern, J. Pure Appl. Chem. 1987, 59, 173.
(27) Cole-Hamilton, D. J.; Wilkinson, G. NouV. J. Chim. 1977, 1, 141.
(28) Kakiuchi, F.; Tanaka, Y.; Sato, T.; Chatani, N.; Murai, S. Chem.
Lett. 1995, 679.
(29) Trost, B. M.; Imi, K.; Davies, I. W. J. Am. Chem. Soc. 1995, 117,
5371.
(30) Ghosh, C. K.; Graham, W. A. G. J. Am. Chem. Soc. 1989, 111,
375-376.
(31) Fujii, N.; Kakiuchi, F.; Yamada, A.; Chatani, N.; Murai, S. Bull.
Chem. Soc. Jpn. 1998, 71, 285-298.
(32) Fujii, N.; Kakiuchi, F.; Yamada, A.; Chatani, N.; Murai, S. Chem.
Lett. 1997, 425.
(33) Jun, C. H.; Lee, H.; Hong, J. B. J. Org. Chem. 1997, 62, 1200.
(34) Bosnich, B. Acc. Chem. Res. 1998, 31, 667 and references therein.
(35) Jun, C. H.; Huh, C. W.; Na, S. J. Angew. Chem., Int. Ed. Engl.
1998, 37, 145.
(36) Jones, W. D.; Hessell, E. T. Organometallics 1990, 9, 718.
(37) Suggs, J. W.; Wovkulich, M.; Cox, S. D. Organometallics 1985, 4,
1101.
(38) Ezbiansky, K.; Djurovich, P. I.; LaForest, M.; Sinning, D. J.; Zayes,
R.; Berry, D. H. Organometallics 1998, 17, 1455. This example discusses
the dehydrogenative coupling of arenes with silanes to generate arylsilanes
with [C5Me5Rh(H)2(SiR3)2] as a catalyst.
(39) Marder, T. B.; Roe, C. D.; Milstein, D. Organometallics 1988, 7,
1451-1453.
(40) (a) Foo, T.; Bergman, R. G. Organometallics 1992, 11, 1801-1810.
This example discusses Ir-indenyl complexes as a precursor for function-
alization reactions. (b) Perthuisot, C.; Edelbach, B. L.; Zubris, D. L.; Jones,
W. D. Organometallics 1997, 16, 2016.
(41) Duckett, S. B.; Perutz, R. N. Organometallics 1992, 11, 90. This
example utilizes complexes of the type [C5H5Rh(olefin)2] for catalytic
hydrosilations.
(42) Seiwell, L. P. J. Am. Chem. Soc. 1974, 96, 7134-7135.
(43) Jones, W. D.; Duttweiler, R. P.; Feher, F. J.; Hessell, E. T. New J.
Chem. 1989, 13, 725-236.
(44) For details concerning catalytic H/D exchange reactions using 1
see: C. P. Lenges, P. S. White, M. Brookhart, J. Am. Chem. Soc. 1999,
121, 4385-4396.
(45) The turnover frequency is 0.7 TO/h and compares well with the
results for the reaction in cyclohexane.
(46) The degree of deuterium incorporation into free olefin and product
3a are on the same order of magnitude, which indicates that activation of
toluene-d8 followed by reversible H/D exchange are faster than productive
aryl-alkyl coupling.

6618 J. Am. Chem. Soc., Vol. 121, No. 28, 1999
Lenges and Brookhart
When the catalysis is carried out to high conversions, the
double alkylation product 4a47 is observed in addition to the
expected product 3a (eq 3). For example, after 18 h in a reaction
with 5 mol % catalyst at 120 °C, 83% conversion to products
was observed. Of these products 18% corresponded to 4a.
as a brown solid by column chromatography. In addition to the
Cp* resonances, signals in the aryl region accounting for five
protons are observed as well as resonances at 2.82 (dd, 1H),
3.18 (d, 1H), 3.28 (d, 1H), 5.60 (m, 1H), and 5.81 ppm (m,
1H). No -SiMe3 resonances are present. Clearly the deactivation
product is composed of 1 equiv of benzophenone and two
nonequivalent [C5Me5Rh] moieties. To date we have been
unsuccessful in obtaining crystals of this species. A proposed
structure that is consistent with the NMR data is shown below
(5).
1
Indicative for 4a was an H NMR resonance at 2.76 ppm (m,
4H, -CH2-). Synthesis of the mono-addition product 3a was
best accomplished in the presence of excess benzophenone,
which prevented formation of 4a. On the other hand, 4a could
be selectively prepared in a reaction employing excess vinyl-
trimethylsilane with 2a (2 mol % 1; olefin-to-ketone ratio of
2.5, cyclohexane-d12). After one week at 120 °C ketone 4a was
the sole product with 3a being formed initially in the catalysis.48
The second equivalent of olefin added regioselectively to the
unsubstituted aromatic ring. This observation was supported by
using 4,4′-dimethoxybenzophenone, 2b, as the substrate. The
first alkylation generates an unsymmetrical ketone with two
inequivalent -OMe groups (1H NMR). Addition of the second
equivalent of vinyltrimethylsilane generates a benzophenone
derivative with equivalent methoxy groups indicating addition
to the second aryl group.49
Catalyst lifetimes are long provided excess olefin is present.
With use of a 1:1 molar ratio of substrates (5 mol % 1, 120 °C,
C6D12) the turnover/time plot is linear to 95% conversion (0.65
TO/h, Figure 1, Experimental Section). This behavior suggests
that the turnover frequency is first order in ketone and inverse
order in olefin, which is consistent with the proposal of
reversible olefin loss from the catalyst resting state followed
by reaction of [C5Me5Rh(C2H3SiMe3)] with ketone. While
quantitative kinetics were not carried out, qualitative experiments
verified that the TOF was suppressed in the presence of excess
vinyltrimethylsilane and accelerated with added ketone. For
example, by using a 1:2.5 molar ratio of vinylsilane:ketone the
initial TOF was observed to increase to ca. 2 TO/h. To further
probe catalyst stability a reaction employing 1 mol % 1 and a
1:1 molar ratio of substrate was carried out (C6D12, 120 °C).
After 110 h, ca. 95% conversion to products had occurred and
NMR analysis of the solution indicated that 1 is still present
and accounts for 90% of the original loading.
A notable feature of the above reaction is the very mild
conditions (55 °C, 8 h) under which product is formed when
no free olefin is present. This is clearly a consequence of the
high concentration of the intermediate [C5Me5Rh(C2H3SiMe3)]
in solution in the absence of added free vinyltrimethylsilane.
This result also suggests that slow addition of olefin to a reaction
mixture generated in this fashion can result in catalytic ketone
formation at these low temperatures compared to the temper-
atures required in the catalytic process described above.
Complex 5 is not a true deactivation product unlike the rhodium
complexes generated in the presence of excess benzophenone
under higher temperature thermolysis conditions. The addition
of excess vinyltrimethylsilane to a cyclohexane solution of 5 at
50 °C results in the formation of 1 without the formation of
additional ketone 3a; after 2 h slow catalytic olefin addition to
generate 3a is observed in this reaction if the olefin concentration
remains low. Complex 5 essentially functions as an alternative
to 1 to deliver the 14-electron fragment [C5Me5Rh] for further
catalysis.
B. Addition of Olefins to Benzophenones and Acetophe-
nones. A survey of the reactivity of several olefins toward
benzophenone (and derivatives) and acetophenones was carried
out. Results are summarized in Table 1 and correspond to
reactions run under standard, unoptimized conditions. Reactions
were conducted at 120 °C, 5 mol % catalyst loading, and with
a 1:1 molar ratio of substrates. In selected cases it was possible
to reduce the catalyst loading significantly (vide infra).
Entry 1 involving benzophenone and vinyltrimethylsilane has
been discussed extensively above. The use of the bulkier 3,3-
dimethyl-1-butene (entry 2) results in more rapid turnover as
might be expected due to more facile dissociation from the
Rh(I) center.50 After 1 h 65% conversion is observed with 3,3,-
dimethyl-1-butene whereas with vinyltrimethylsilane 11 h are
required for 49% conversion.
Thermolysis in the presence of excess benzophenone under
conditions of catalysis (120 °C) results in catalyst deactivation
after complete conversion of olefin. Analysis of the H NMR
1
spectrum of a reaction mixture generated in this fashion showed,
in addition to ketone 3a and excess benzophenone, a series of
resonances in the Cp* region indicating an ill-defined deactiva-
tion process to generate a series of rhodium-containing species.
This was further investigated in a reaction of 1 with excess
benzophenone (10 equiv) in cyclohexane-d12 at lower temper-
ature (55 °C). After 4 h, 65% of 1 was converted to generate
3a (92%) and 4a (8%) and a single rhodium-containing species,
complex 5, with Cp* resonances at 1.62 and 1.91 ppm in the
ratio of 1:1. No free olefin remained in the reaction mixture,
which changed color from orange to brown during this process.
After 8 h at 55 °C complete conversion of 1 to 5 was observed
with nearly all olefin converted to 3a. Complex 5 was isolated
Another particularly reactive olefin is norbornene, which is
both bulky (disubstituted) and strained. In this case the reaction
(47) Organic products have been isolated by chromatography and
characterized by NMR spectroscopy and elemental analysis or HRMS; see
Supporting Information.
(48) The formation of even higher alkylation products is observed if
reaction times are increased, indicating that the catalyst is long-lived even
though additional alkylation is slower. Indicative for these products are the
1H NMR resonances for the methylene groups R to the aryl ring (3a, δ
2.65, 4a, δ 2.83; higher alkylation products, δ 3.11).
(49) The relatively high ratio of higher alkylated products suggests that
the second alkylation might occur without prior release of the initially
alkylated 2a. Also, the observation of secondary alkylation products in the
deactivation process (vide infra) suggests this notion.
1
was followed by H NMR spectroscopy (1:1 substrate ratio, 5
mol % 1). Complete conversion to 3g was observed after 3.5 h
at 70 °C in cyclohexane-d12 (for a turnover-time plot see the
Experimental Section). The TOF for the formation of 3g was
moderate in the initial stage of catalysis (1 TO/h for 1.5 h) and
increased by an order of magnitude to 10 TO/h (1.5-2.5 h) as
(50) Independent synthesis of a rhodium olefin complex with 3,3-
dimethyl-1-butene was not successful. In situ NMR studies indicate the
formation of such complexes as intermediates.

Addition of Olefins to Aromatic Ketones
J. Am. Chem. Soc., Vol. 121, No. 28, 1999 6619
Table 1. Catalytic Addition of Olefins to Aromatic Ketones
Catalyzed by 1 [5 mol % 1 (0.01 g, 2.3 × 10-5 mol), 120 °C,
cyclohexane, ketone:olefin ratio 1:1]a
the -CF3-substituted system (entry 9) showed somewhat
increased reactivity. With reductive elimination being the
turnover-limiting step the electron-deficient benzophenone
derivative is expected to accelerate this process. In entry 10,
however, the conformational flexibility has been significantly
reduced and a decrease in the rate of addition to olefin was
observed. The required geometry for the transition state of the
reductive elimination reaction might be less accessible with this
substrate. Even though this reaction is slow, high conversions
can be obtained at longer times.
C. Catalytic Conversion of Acetophenones. Acetophenones
are also substrates for catalytic hydroarylation (entries 11-16).
A noteworthy difference was observed in the addition reaction
of vinyltrimethylsilane to acetophenones 6a-c. The conversion
to hydroarylation products 7a-c was slow and analysis of the
reaction by NMR spectroscopy indicated that additional products
were generated in this process. For example, in the reaction of
p-chloroacetophenone, 6b, with vinyltrimethylsilane (5 mol %
1, cyclohexane-d12, ratio 1:1, 120 °C) 24% conversion to 7b
was observed after 12 h in addition to 22% conversion of 6b to
a new species. In addition to a resonance for free ethlyene in
solution (s, 5.3 ppm, C6D12) two new doublets are observed at
4.38 and 4.85 ppm (J ) 3 Hz) along with resonances in the
aromatic and the -SiMe3 region. Comparison to an authentic
sample54 established this side product as the silyl protected
enolate (eq 4, see the Supporting Information for a sample NMR
spectrum).
a (1) Parentheses: reaction time (hours). (2) Conversion corresponds
to 3a (88%) and 4a (11%). (3) Endo/exo distribution see text. (4) Olefin
addition to 2,5-disubstituted isomer. (5) Conversion of 6a to products;
side product 30%; see text (eq 4).
4
catalysis proceeded. A decrease of the TOF was again observed
close to complete conversion. After 1.5 h, 85% conversion of
vinyltrimethylsilane (introduced with 1) to 3a was observed,
which explains the significant increase in TOF for the formation
of 3g. After complete conversion of norbornene, 16% endo-3g
and 84% exo-3g were observed in the reaction mixture.51 The
high reactivity of norbornene allowed the decrease of catalyst
loading to 0.5 mol % with 92% conversion to exo/endo-3g
(exo: 82%) after 18 h at 80 °C.
Cyclopentene on the other hand was converted less effectively
to 3f (entry 6, Table 1); catalyst deactivation to unassigned
products was observed. The addition of R-olefins to 2a as in
entries 3 and 4 (Table 1) showed fair conversions to ketones
3c and 3d. Exclusive anti-Markovnikov addition was observed
to generate the linear products, even though isomerization to
internal olefins in the case of entry 3 was observed during the
reaction.52 The vinyl ether in entry 5 was converted to product
but conversion was low and early catalyst deactivation was
observed.53
The formation of ethylene during this side reaction decreases
catalytic activity. Ethylene coordinates quite strongly to the
rhodium system and is less labile under the reaction conditions.
With acetophenone, catalysis stops after approximately 60%
conversion. With p-(trifluoromethyl)acetophenone, 6c, the silyl
transfer was of minor importance and high conversion to 7c
was observed (entry 13).
A possible mechanism for catalytic silyl transfer to generate
a silyl enolate and ethylene from vinyltrimethylsilane and
acetophenone is illustrated in Scheme 1. Several significant
features of catalytic hydroarylation with 1 as a catalyst are
illustrated in this side reaction. Olefin complex 1 reacts readily
to activate sp3-CH bonds in substrates such as acetone indicated
by catalytic H/D exchange as illustrated in eq 2 for benzene.44
The reversible activation of the acetophenone -CH3 group was
observed by using labeling experiments (vide infra) and supports
formation of intermediate A via step i in Scheme 1. Reductive
elimination from A of two sp3-alkyl groups to generate a γ-silyl
ketone is not likely in this system. Alternatively â-silyl
elimination from A (step ii) generates intermediate B from which
a lower barrier for the reductive elimination reaction is available
by generating a C-Si bond. Ethylene dissociation in the
presence of excess vinyltrimethylsilane regenerates 1. The
In general, substituted benzophenones showed similar reac-
tivities (entries 8-10) for the addition reaction. The -OMe-
substituted system (entry 8) had somewhat lower reactivity while
(51) It is curious to note that the ratio of endo to exo products in this
reaction increases as the reaction proceeds with exo-3g being the exclusive
addition product in the early stages of catalysis, which suggests that
intermediates such as 5 or mixed olefin complexes generated from 1 and
norbornene show different reactivity as the presumed bis-norbornene
rhodium complex, which is assigned as the resting state during norbornene
conversion.
(54) (a) p-Chloroacetophenone was treated with LDA at low temperatures
and quenched with ClSiMe3. The enolate was isolated by vacuum distillation.
The corresponding SiMe3 protected enolate of acetophenone is commercially
available. (b) A referee has suggested that formation of the silyl enolate
might occur through a mechanism involving the addition of a rhodium
hydride to the vinyl silane followed by â-silyl elimination and release of
ethylene to generate a Rh-silyl complex. Insertion of ketone into the Rh-
silyl bond followed by â-elimination would generate the silyl enolate and
regenerate the rhodium hydride. While we cannot strictly rule out this
mechanism, it is not apparent how a Rh-H species could be generated as
the chain carrier in this system.
(52) Analogous results are observed by Murai and co-workers using the
Ru system in the reaction of 1-hexene (see refs 14-22).
(53) Addition product 3e was characterized in situ by 1H NMR
spectroscopy and mass spectroscopy. 1H NMR (C6D6, 300 MHz, 20 °C) δ
0.98 (t, 3H, -CH3); 3.25 (q, 2H, OCH2); 3.15 (t, 2H, -CH2-); 3.62 (t,
2H, -CH2-); 7.95 (m, 2H, Ar); 7.04-7.18 (m, Ar.). MS/EI: 254.18.

6620 J. Am. Chem. Soc., Vol. 121, No. 28, 1999
Lenges and Brookhart
Scheme 1. Proposed Mechanism for Silylenolate Formation
during Catalytic Hydroarylation of Acetophenone with
Vinyltrimethylsilane with Use of 1 as Catalyst
roughly 30% incorporation of protio label into these sites. H/D
exchange delivered deuterium from acetophenone into free
vinyltrimethylsilane and decreased the 1H NMR signal intensity
1
without converting the olefin to products. All H vinylic sites
were reduced in intensity on the same time scale to generate
perdeuteriovinyltrimethylsilane (eq 5).
5
These results show that activation of the meta and para C-H
bonds followed by olefin insertion is reversible and results in
H/D exchange by a mechanism similar to that shown in eq 2.56
In contrast, no H/D exchange is observed in the ortho position,
yet this is the sole site of alkylation. There are two plausible
explanations for this observation. One possibility is that the
barrier for the overall process of oxidative addition of the ortho
C-D(H) bond and migratory insertion to generate a Rh(III) alkyl
aryl intermediate (see eq 2) is much higher than the correspond-
ing barrier for the meta and para C-D(H) bonds. An alternative,
illustrated in eq 6, is that oxidative addition/migratory insertion
thermal rearrangement of R-silyl ketones to silyl-protected
enolates (step iv) has been reported (Brook rearrangement) and
is likely to occur under these conditions to generate the final
products. Another attractive option involves isomerization of
A to generate an O-bound rhodium enolate complex (species
C) that will, after silyl elimination (step iii) and Si-O reductive
elimination (step v), generate the silyl enolate directly. At this
point it is difficult to distinguish between these mechanistic
alternatives.54b But in favor of pathway iii, it is significant to
point out that in all cases investigated silyl-aryl coupling with
vinyltrimethylsilane and 1 as catalyst has not been observed,
which is notable since the reductive elimination of an aryl-
silyl bond is expected to be more facile as recently observed
by Berry and co-workers.38 The presence of the carbonyl group
in the catalytic process investigated here is not only important
for reducing the barrier for reductive elimination in C-C
coupling and controlling the regioselectivity of alkylation but
also seems to be responsible for the specific side reaction
observed in the addition of vinyltrimethylsilane to acetophenone.
Using 3,3-dimethyl-1-butene as the olefin gave good conver-
sion to products 8a and 8b,55 which confirmed that this side
reaction is specific for vinyltrimethylsilane and acetophenones.
As in the benzophenone case, double alkylation can be observed
and such a product has been isolated from a reaction of 6c with
an excess of vinyltrimethylsilane.
D. Reaction of Acetophenone-d8 with Vinyltrimethylsilane.
Mechanistic information was obtained from the results of the
reaction of acetophenone-d8 with vinyltrimethylsilane (5 mol
% 1, 0.01 g, substrate ratio 1:1, cyclohexane-d12) at 80 °C,
conditions under which product formation is not observed. (For
a plot showing the change of intensity of the residual 1H NMR
resonances of acetophenone-d8 during this process see the
Experimental Section, Figure 3). The residual resonance for the
ortho sites of acetophenone-d8 did not change during the 35 h
this reaction was followed by 1H NMR spectroscopy. In contrast,
the resonances for the meta and para sites were increased 15-
fold in intensity during this process, which corresponds to
6
occurs reversibly but a chelate interaction in the alkyl aryl
Rh(III) intermediate renders the methylene hydrogens (in this
case -CHD-) diastereotopic and therefore the deuterium which
migrates to the olefin must, by microscopic reversibility, also
return to the rhodium in the reverse â-elimination step. Thus,
even though the C-H oxidative addition/migratory insertion
process is reversible, no H/D exchange occurs.
We have no firm evidence to distinguish these possibilities.
However, we have observed that with other monosubstituted
benzenes44 (e.g., toluene, chlorobenzene) H/D exchange in the
ortho position is only slightly slower than in meta and para
positions and the exchange occurs at much lower temperatures
than is required for the catalytic coupling reaction of aceto-
phenone with vinyltrimethylsilane. Thus, we currently favor the
latter explanation, that is oxidative addition/migratory insertion
of the ortho C-H(D) bond is reversible but not detectable via
an H/D exchange experiment. If this explanation is correct, it
suggests that in the catalytic cycle reductive elimination from
the alkyl-aryl intermediate is the turnover limiting step.
On the basis of the results obtained to date, some statements
can be made concerning the overall catalytic cycle (see Scheme
2). The increase in turnover frequency with increase in ketone
concentration and suppression of the turnover frequency with
added olefin suggests a reversible loss of olefin to produce an
intermediate (9) that is trapped by ketone. For simplicity we
show this trapped species as the oxidative addition adduct, 10.
(A simple ketone adduct of 9 may well intervene between 9
and 10). Olefin insertion generates 11 in which chelation of
the carbonyl group provides an 18-electron intermediate. Such
species have precedent. Cyclometalations which afford metal
(56) No site selectivity regarding the olefin was observed, which suggests
that olefin insertion into a Rh hydride can occur with both regiochemistries
under these conditions. This is especially remarkable since the exclusive
products observed are based on anti-Markovnikov olefin addition. Olefin-
hydride insertion might be directed by the carbonyl group with respect to
regiochemistry; no direct evidence for this is available.
(55) Ketone 8b was characterized in situ after catalysis and shows
spectroscopic features comparable to similar fully characterized products
(see ref 16 and Supporting Information). 1H NMR (C6D6, 300 MHz, 20
°C) δ 0.88 (s, 9H, CMe3); 1.32 (m, 2H, CH2); 2.68 (m, 2H, CH2); 1.96 (s,
3H, Me); 7.38 (s, 1H, Ar); 7.11 (d, 1H, Ar); 6.92 (d, 1H, Ar). MS/EI: 272.16.

Addition of Olefins to Aromatic Ketones
J. Am. Chem. Soc., Vol. 121, No. 28, 1999 6621
Scheme 2. Proposed Mechanism for Catalytic Olefin
Addition to Aromatic Ketones
Scheme 3. Murai Mechanism for Ruthenium-Catalyzed
Hydroarylation of Olefins
refluxing toluene, 1-24 h.). While optimization of the catalytic
process in the Ru system has resulted in generally good turnover
numbers, the rhodium system investigated here shows interesting
potential for further developments in this area, especially
considering the mild conditions for the reaction of norbornene
and the reactivity in the presence of excess benzophenone. Both
systems show in general high regioselectivity for addition to
the olefin to generate the anti-Markovnikov product. Olefin
isomerization is observed in situ with the rhodium as well as
with the ruthenium system, but aromatic alkylation generates
only the linear addition products for substrates such as 1-pentene
or 1-hexene. The catalyst activation procedure in the ruthenium
system is believed to be an initial olefin hydrogenation step
that generates a ruthenium(0) olefin complex; in the system
investigated here rhodium bis-olefin complexes are the catalyst
resting states. The proposed mechanism for catalytic alkylation
of aromatic ketones catalyzed by [RuH2(CO)(PPh3)3] is il-
lustrated in Scheme 3.
complexes stabilized by intramolecular coordination of a
carbonyl group have been observed in a variety of ruthenium
complexes.23,58-64
The barrier to reductive elimination is apparently reduced by
this chelate interaction which results in exclusive ortho alky-
lation. This is consistent with proposals by Murai regarding the
ruthenium-catalyzed alkylation and by others.57 This is also
consistent with our recent study of catalytic hydroacylation with
similar cobalt complexes where it was found that reductive
elimination of ketone is most likely mediated by an η2-acyl
interaction (eq 7).65,66 Maitlis and co-workers have reported the
formation of acetophenone from [C5Me5Rh(CO)(Me)(Ph)] in a
reductive elimination reaction that most likely involves η2-acyl
assistance.67
A key difference in the mechanisms of Schemes 2 and 3 is
that in the ruthenium-catalyzed process carbonyl coordination
is presumed to be the first step, which directs the C-H bond
activation to the ortho position of the aromatic substrate.68 H/D
exchange experiments establish that para and meta C-H bonds
are not activated in the Ru system. These rhodium complexes,
on the other hand, are not discriminating in the C-H bond
activation step and activation of all sites (ortho, meta, para) of
the substrate is observed. Murai has established that C-H bond
activation is fast and reversible in the Ru system and that the
reductive elimination of the alkylated product is the turnover-
limiting step. We believe the same feature applies to the rhodium
system, but no conclusive evidence is available yet concerning
this point. In both systems, chelation of the carbonyl group to
the metal center lowers the barrier for reductive elimination to
product.
7
E. Conclusions. The catalytic process reported here intro-
duces rhodium complex 1 as an efficient catalyst for the addition
of a variety of olefins regioselectively to the ortho positions of
aromatic ketones. Mechanistic details of this reaction have been
studied with benzophenone and vinyltrimethylsilane as sub-
strates. The turnover-limiting step is likely the reductive
elimination reaction to generate the alkylated ketone product.
Deactivation routes and side reactions for acetophenones have
been described and support the mechanistic scheme developed
for this reaction.
Several significant differences are apparent in a comparison
of the ruthenium system developed by Murai and co-workers
and the rhodium system introduced here. The reaction conditions
for both catalytic systems are similar (Murai system: 135 °C,
Experimental Section
General Considerations. All manipulations of air- and/or water-
sensitive compounds were performed by using standard high-vacuum
or Schlenk techniques. Argon and nitrogen were purified by passage
through columns of BASF R3-11 catalyst (Chemalog) and 4 Å
molecular sieves. Solid organometallic compounds were transferred in
an argon-filled Vacuum Atmospheres drybox. 1H and 13C NMR
chemical shifts were referenced to residual 1H NMR signals and to the
13C NMR signals of the deuterated solvents, respectively. NMR probe
temperatures were measured by using an anhydrous ethylene glycol
sample. Elemental analyses were performed by Atlantic Microlabs, Inc.,
of Norcross, GA.
(57) Brown, J. M.; Cooley, N. A. Chem. ReV. 1988, 88, 1031.
(58) (a) Komiya, S.; Ito, T.; Cowie, M.; Yamamoto, A.; Ibers, J. A. J.
Am. Chem. Soc. 1976, 98, 3874. (b) Tolman, C. A.; Ittel, S. D.; English, A.
D.; Jesson, J. P. J. Am. Chem. Soc. 1979, 101, 1742
(59) Rothwell, I. P. Acc. Chem. Res. 1988, 21, 153.
(60) Ryabov, A. D. Synthesis 1985, 233.
(61) Newkome, G. R.; Puckett, W. E.; Gupta, V. K.; Kiefer, G. E. Chem.
ReV. 1986, 86, 451.
(62) McGuiggan, M. F.; Pignolet, L. H. Inorg. Chem. 1982, 21, 2523.
(63) McKinney, R. J.; Kaesz, H. D. J. Am. Chem. Soc. 1975, 97, 3066.
(64) Mawby, R. J.; Saunders: D. R. J. Chem. Soc. 1984, 2133.
(65) Lenges, C. P.; White, P. S.; Brookhart, M. J. Am. Chem. Soc. 1998,
120, 6965.
Materials. All solvents used for synthesis were deoxygenated and
dried via passage over a column of activated alumina.69 Tetrahydrofuran
(66) Bergman, R. G. Acc. Chem. Res. 1980, 13, 113.
(67) Sunley, G. J.; Fanizzi, F. P.; Saez, I. M.; Maitlis, P. M. J. Organomet.
Chem. 1987, 330, C27.
(68) Labeling studies have shown that C-H bond activation is only
observed in the ortho positions of acetophenone with use of the ruthenium
catalyst; the other sites are not activated (see ref 16).

6622 J. Am. Chem. Soc., Vol. 121, No. 28, 1999
Lenges and Brookhart
Figure 1. Conversion of benzophenone 2a with Vinyltrimethylsilane
(ratio 1:1) to 3a catalyzed by 1 (0.01 g, 2.3 × 10-5 mol, 5 mol %) at
120 °C. kTOF ) 0.6 TO/h.
Figure 2. Catalytic conversion of norbornene and benzophenone to
3g at 70 °C in cyclohexane-d12 (0.01 g 1, 2.3 × 10-5 mol, 5 mol %,
substrate ratio 1:1).
was distilled from sodium benzophenone-ketyl prior to use. [C5Me5-
Rh(C2H3SiMe3)2] was prepared following a previously reported pro-
cedure.44 The procedure is also described in the Supporting Information.
All substrates listed in Table 1 were commercially available from
Aldrich. Benzene-d6, toluene-d8, and cyclohexane-d12 were dried over
potassium benzophenone-ketyl, acetone-d6, and acetophenone-d8 over
CaH2, vacuum transferred, and degassed by repeated freeze-pump-
thaw cycles.
Catalytic Alkylation of Benzophenone: Kinetic Analysis. A
solution of 1 (0.01 g, 2.28 × 10-5mol) in cyclohexane-d12 (0.6 mL)
was treated with 20 equiv of benzophenone and 20 equiv of vinyltri-
methylsilane. Both reagents were added in a drybox to an NMR tube
equipped with a Teflon screw top. This tube was inserted into the heated
probe of an AMX 300 NMR spectrometer and spectra were taken with
time as the reaction proceeded. A plot of turnover number vs time is
illustrated in Figure 1. Complete conversion was observed. An
alternative procedure involved the same sample preparation followed
by thermolysis in a temperature-controlled oil bath. The sample tube
was removed at various intervals and analyzed by NMR spectroscopy
at 20 °C.
In a series of similar experiments turnover frequencies on the order
of 1 TO/h were observed. In all cases the reaction mixtures remained
homogeneous and no deactivation was observed. The only rhodium-
containing species present during these reactions was 1.
Analogously, reactions were carried out with different ratios of
vinyltrimethylsilane to benzophenone. The appropriate amount of
substrate was added by weight and the turnover process was analyzed
by NMR spectroscopy. TOF ) 0.6 TO/h (5 mol % 1, 1:1 substrate
ratio) and TOF ) 0.03 TO/h (5 mol % 1, 1:2.5 ratio of ketone to olefin);
TOF ) 1.4 TO/h (5 mol % 1, 1.4:1 ratio of ketone to olefin); TOF )
2.0 TO/h (5 mol % 1, 2.5:1 ratio of ketone to olefin).
Catalytic Addition of Norbornene to Benzophenone. In a similar
reaction norbornene and benzophenone, 2a, were heated at 70 °C in
cyclohexane-d12 with 1 (0.01 g, 2.3 × 10-5 mol, 5 mol %) as catalyst.
The reaction was followed by NMR spectroscopy and the conversion
to endo/exo-3g was plotted vs time in addition to the conversion to
only exo-3g. After complete conversion, endo-3g accounts for 16% of
the product. In addition, vinyltrimethylsilane was converted to 3a during
this process. The initial turnover frequency of 1 TO/h increased as
catalysis progressed and an intermediate TOF of 10 TO/h was observed.
Close to complete conversion the turnover frequency decreased.
H/D Exchange with Acetophenone-d8. Acetophenone-d8 and vin-
yltrimethylsilane were mixed with 1 (0.01 g, 2.3 × 10-5 mol, 5 mol
%) in cyclohexane-d12 and heated to 80 °C and followed with time by
1H NMR spectroscopy. Conversion to ketone 8a or side products
according to eq 4 was not observed under these conditions. H/D
exchange based on reversible acetophenone activation and olefin
insertion resulted in an increase of the residual acetophenone resonances
Figure 3. Increase in intensity of the residual resonances for acetophe-
none-d8 during thermolysis at 80 °C in the presence of vinyltrimeth-
ylsilane and 1 (5 mol %, ratio of olefin to acetophenone 1:1).
and a decrease of the olefinic resonances of vinyltrimethylsilane.
Resonances were normalized based on the integration of 1 (relative to
one proton) and plotted vs time in Figure 3 for the change of the
acetophenone resonances. The change of the integration for the
decreasing olefinic resonances showed that no site selectivity is observed
and all three vinylic protons of vinyltrimethylsilane were reduced to
65% after 30 h under the indicated reaction conditions. The solvent
was not activated during this process and was therefore not a source
for deuterium exchange.
The residual resonance for the ortho sites of acetophenone did not
change in intensity during this reaction, while the para and meta sites
showed considerable H/D exchange that resulted in an increase that
accounted for roughly three protons based on the original intensity after
30 h. The residual resonance of the -CD3 group in the 1H NMR
spectrum increased in intensity during this process (3-fold based on
the original intensity).
Catalytic Alkylation of Aromatic Ketones. Complex 1 (0.01 g,
2.3 × 10-5 mol) was added to an NMR tube equipped with a Teflon
screw cap or a thick-walled glass tube with a Teflon screw top and
dissolved in cyclohexane (-d12 or -h12, if solubility required a different
solvent (entry 8, Table 1) benzene was used). To this was added the
required amount of ketone and olefin (in general 4.6 × 10-4 mol each)
by weight wiht a syringe. The sealed containers were removed from
the drybox and kept at 120 °C immersed in an oil bath. Samples
prepared in NMR tubes were monitored during catalysis. The residual
solvent resonance was used as an integration standard. After the
reactions were terminated all volatiles were removed and the reaction
mixture was analyzed by NMR spectroscopy to determine conversion.
Column chromatography (flash-chromatography) on alumina gave in
general pure material as judged by NMR spectroscopy. When vinyl-
trimethylsilane was not used as the olefin, small impurities based on
(69) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.;
Timmers, F. J. Organometallics 1996, 15, 1518.

Addition of Olefins to Aromatic Ketones
J. Am. Chem. Soc., Vol. 121, No. 28, 1999 6623
the initial conversion of vinyltrimethylsilane originating from complex
1 were present in some samples and proved difficult to separate by
chromatography. NMR data (1H and 13C) and data from elemental
analysis or mass spectroscopic analysis are summarized in the Sup-
porting Information.
Supporting Information Available: Spectroscopic and
analytical data for the organic products of Table 1; H NMR
spectrum for complex 5; sample NMR spectrum for the side
reaction in the conversion of acetophenones (PDF). This material
is available free of charge via the Internet at http://pubs.acs.org.
1
Acknowledgment is made to the National Institutes of Health
(Grant GM 28938) for financial support. C.P.L. thanks the Fonds
der Chemischen Industrie, Germany, for a Kekule´ fellowship.
JA990702S