Preparation of σ- and π-allylcopper(III) intermediates in S N2 and SN2′ reactions of organocuprate(I) reagents with allylic substrates

Article abstract of DOI:10.1021/ja801186c

The first π-allyl complexes of Cu^III have been prepared and characterized by using rapid injection nuclear magnetic resonance spectroscopy (RI-NMR). The prototype, (η³-allyl)dimethylcopper(III), was prepared by injection of allyl chloride into a THF-d₈ solution of iodo-Gilman reagent, Me₂CuLi·LiI (A), spinning in the probe of an NMR spectrometer at -100 °C. A σ-allyl ate complex, lithium (η¹-allyl)trimethylcuprate(III), was prepared in high yield by including 1 equiv of tributylphosphine in the reaction mixture or by using allyl acetate as the substrate. Cyano ate complex, lithium cis-(η¹-allyl)cyanodimethylcuprate(III) was obtained in high yield by injecting allyl chloride or allyl acetate into the cyano-Gilman reagent, Me₂CuLi·LiCN (B), in THF-d₈ at -100 °C. Reactions of A with allylic substrates show a definite dependence on leaving group (chloride vs acetate), whereas those of B do not. Moreover, these reagents have different regioselectivities, which in the case of A vary with temperature. Finally, the exclusive formation of cis-cyano σ-allyl Cu^III intermediates in both the 1,4-addition of B to α-enones and its S_N2α reaction with allylic substrates now makes sense in terms of π-allyl intermediates in both cases, thus unifying the mechanisms of these two kinds of conjugate addition. Copyright

Full text of DOI:10.1021/ja801186c

Published on Web 08/01/2008
Preparation of σ- and π-Allylcopper(III) Intermediates in SN2 and SN2′
Reactions of Organocuprate(I) Reagents with Allylic Substrates
Erika R. Bartholomew, Steven H. Bertz,*,† Stephen Cope, Michael Murphy, and Craig A. Ogle*
Department of Chemistry, UniVersity of North CarolinasCharlotte, Charlotte, North Carolina 28223
Received February 17, 2008; E-mail: cogle@uncc.edu; sbertz@complexitystudycenter.org
Chart 1. New Compounds Prepared in This Study with NMR
Chemical Shifts for 13C (Red) and 1H (Blue) in THF-d8 at -100 °C
(The Li+ counterions in 2 and 3 have been omitted for clarity.)
The reactions of organocopper(I) reagents with allylic substrates
have been among their most valuable synthetic applications,1 and
consequently, they have been scrutinized from experimental2 as
well as theoretical3 points of view. Nevertheless, the outcome is
difficult to predict for each new case, as the partition between SN2
and SN2′ products is a complex function of the experimental
parameters. Both σ- and π-allylcopper(III) intermediates have been
proposed, and a detailed understanding of their roles can only
accelerate progress in this important area, for example, in the
development of enantioselective reactions.
Copper(III) intermediates have been shown to play a central role
in the conjugate addition4 and SN2 reactions5–7 of organocopper(I)
reagents. σ-Allyl CuI compounds are well-known,8 but π-allyl CuI
compounds appear to be unknown;9 thus, the corresponding CuIII
complexes have theoretical as well as practical interest. We can
now report that, by using rapid injection NMR spectroscopy
(RI-NMR),10,11 we have been able to prepare and characterize the
first examples of both η1 σ-allyl and η3 π-allyl CuIII complexes
and to study their reactions.
Prototypical π-allyl CuIII complex, (η3-allyl)dimethylcopper(III)
1a (Chart 1), was prepared by injecting a solution of allyl chloride
in THF-d8 into a solution of iodo-Gilman reagent Me2CuLi·LiI
(A) in THF-d8, spinning in the probe of an NMR spectrometer at
-100 °C. A σ-allyl CuIII species, lithium (η1-allyl)trimethylcu-
prate(III) 2a, also appeared rapidly (∼50% maximum yield, 0.1 h),
but then disappeared by the time 1a reached its maximum yield
(∼80%, 0.5 h). Complex 1a slowly decomposed to 1-butene.
Lithium cis-(η1-allyl)cyanodimethylcuprate(III) 3a was obtained
in essentially quantitative yield (0.3 h) from allyl chloride and
cyano-Gilman reagent Me2CuLi·LiCN (B) in THF-d8 at -100 °C.
At longer times, minor amounts of 2a appeared as Me replaced
CN. Substituted cyano ate complexes 3b-d were prepared analo-
gously from B and the corresponding allyl chlorides. In contrast to
the cis-cyano complexes formed here, trans-EtMe2Cu(CN)Li was
formed in the reaction of B with EtI.5
Likewise, substituted π-allyl CuIII compounds 1b-d and σ-allyl
CuIII ate complexes 2b-d were prepared from A and the corre-
sponding allyl chlorides. As in reactions involving A and EtI,5,6
the main side product with excess A was Ashby’s cuprate,
Me3Cu2Li;12 otherwise it appeared to be MeCu(Cl)Li.
2
Two-bond 13C-13C coupling constants J across copper have
When the organocopper(I) reagent was Me2CuLi ·LiI/PBu3, the
product was 2a (>90%) and the side product was MeCuPBu3. No
1a was present: PBu3 appears to inhibit the conversion of 2a to 1a
by complexing MeCu (vide infra). No copper(III)-phosphine
complex was observed, in contrast to the result with this reagent
and EtI, where the product was trans-EtMe2Cu(PBu3).6
been invaluable for structural studies of organocopper(I)13,16 and
organocopper(III) compounds.4–6 In the case of labeled 1a, (η3-
13CH2CH13CH2)(13CH3)2Cu,17 a single two-bond coupling constant,
2J ) 9.7 Hz,13 was measured between C1/C3 of the π-allyl group
and the carbon atoms of the (equivalent) methyl groups. Its
magnitude lies between typical values for cis and trans two-bond
couplings in square planar complexes of copper(III).4–6 For
example, in labeled 2a, (η1-13CH2CH13CH2)(13CH3)3CuLi,17 the
methylene-methyl trans coupling was 2J ) 34.2 Hz, while the
methylene-methyl cis coupling was too small to be resolved. The
Structures of the new compounds were assigned by using 1D
and 2D NMR (see Supporting Information). Especially noteworthy
are the 13C NMR chemical shifts to higher shielding from TMS in
1a-d (Chart 1). The 13C NMR shifts for the methyl (-3.56 ppm)
and allyl (77.39 ppm for C1/C3) groups in 1a are similar to the
corresponding values for π-complexes of A with R-enones.10,13,14
The lack of a clear demarcation between NMR shifts for CuI and
CuIII compounds is not surprising, as Snyder has calculated that
the charges on Cu in what are formally organocopper(III) complexes
are actually ca. +1.15
2
methyl-methyl cis coupling was J ) 2.3 Hz.
In labeled 3a, (η1-13CH2CH13CH2)(13CH3)2Cu(13CN)Li,17 two-
bond 13C-13C couplings between the cyano group and the methyl
groups cis and trans to it were 2J ) 4.2 and 38.0 Hz, respectively.
The methylene-methyl trans coupling was 2J ) 31.1 Hz. The rest
of the cis couplings were not resolved.
Addition of allyl acetate to A in THF-d8 at -100 °C gave a
good yield (>90%) of 2a (Scheme 1). In the absence of acetate or
† Complexity Study Center, 88 East Main St., Suite 220, Mendham, NJ 07945.
9
11244 J. AM. CHEM. SOC. 2008, 130, 11244–11245
10.1021/ja801186c CCC: $40.75  2008 American Chemical Society

C O M M U N I C A T I O N S
Scheme 1. Alternative Routes to Organocopper(III) Complexes
Upon warming to -90 °C (0.5 h), 90% of 2d had been converted
to 1d and alkenes (1:3), 3-phenylbutene 4 (SN2′ product) and (E)-
1-phenylbutene 5 (SN2 product, ca. 20:1 at -100 °C). Upon
warming to -70 °C, 1d decomposed at a significant rate to afford
the final mixture of 4 and 5 (1:1, 1 h). The ratio of 4 to 5 from 1d
at -70 °C was 1:3.
The analogous reaction of cinnamyl chloride with B gave only
3d at -100 °C. It slowly decomposed at -80 °C to 4 and 5 (1:14,
2 h), as a small amount (15% max) of π-allyl complex 1d appeared.
Thus, in the reaction of A with cinnamyl chloride, the π-allyl
CuIII intermediate 1d gave mainly SN2 product, whereas the σ-allyl
CuIII intermediate 2d gave mainly SN2′ product. In contrast, σ-allyl
CuIII intermediate 3d from B and cinnamyl chloride gave predomi-
nantly SN2 product, which can be rationalized by the intermediacy
of 1d, as in the case of the SN2 product from A.
Scheme 2. Dissociative versus Associative Reaction Mechanisms
(The Li+ counterions are not shown for the sake of simplicity.)
In summary, we have confirmed the structures proposed for
π-allyl CuIII complexes by preparing the first examples. On the
other hand, the ate structures that we have established for the first
σ-allyl CuIII complexes are novel, and they are crucial to understand
the reactivity of allylic substrates with organocopper(I) reagents.
Finally, cis-cyano σ-allyl CuIII intermediates are found in both
the 1,4-addition reaction of B/TMSCl with 2-cyclohexenone4 and
the SN2′ reaction of B with dimethylvinylcarbinol acetate; hence,
the two types of conjugate addition are unified by a deep connection,
which we conjecture to be π-allylcopper(III).
Acknowledgment. We thank D. Deadwyler for his assistance
with the rapid injection equipment and the U.S.A. National Science
Foundation (Grant 0718368) for funding.
PBu3, methyl is rapidly removed from 2a, perhaps by a soluble
form of MeCu (e.g., Me3Cu2Li). Substituted ate complexes 2b-d
were prepared analogously from the corresponding allyl acetates.
Supporting Information Available: NMR spectra for compounds
1a, 2a, and 3a. This material is available free of charge via the Internet
at http://pubs.acs.org.
Under the same conditions, allyl acetate and B gave an essentially
quantitative yield of 3a. Substituted cis-cyano complexes 3b-d
were prepared in good yields from B and the corresponding allyl
acetates, although prenyl acetate required higher temperatures. Small
amounts of π-allyl complexes 1 appeared at longer times or higher
temperatures; presumably, 3 loses cyanide in a soluble form such
as Li2CN+.18
References
(1) (a) Harutyunyan, S. R.; den Hartog, T.; Geurts, K.; Minnaard, A. J.; Feringa,
B. L. Chem. ReV. 2008, 108, in press. (b) Woodward, S. Angew. Chem.,
Int. Ed. 2005, 44, 5560–2. (c) Alexakis, A. ACS Symp. Ser. 2004, 880,
43–59. (d) Helmchen, G.; Ernst, M.; Paradies, G. Pure Appl. Chem. 2004,
76, 495–506.
Addition of either prenyl acetate or dimethylvinylcarbinol acetate
to a solution of B in THF-d8 at -60 °C yielded 3c exclusively.
Upon warming the reaction mixture to -40 °C, 2-methyl-2-pentene
appeared as the only alkene.
(2) Karlstro¨m, A. S. E.; Ba¨ckvall, J.-E. Chem.—Eur. J. 2001, 7, 1981–9.
(3) Yamanaka, M.; Kato, S.; Nakamura, E. J. Am. Chem. Soc. 2004, 126, 6287–
93.
(4) Bertz, S. H.; Cope, S.; Murphy, M.; Ogle, C. A.; Taylor, B. J. J. Am. Chem.
Soc. 2007, 129, 7208–9.
In a related experiment, addition of allyl-1-d1 acetate to B in
THF-d8 at -100 °C gave 3a-d1 in which the label was scrambled
between C1 and C3.
Since 1a and 3a are 16-electron d8 species, either dissociative
or associative mechanisms can account for the results from B
and allylic substrates (Scheme 2). The former involves a trigonal
14-electron species T3,15 and the latter a pyramidal 18-electron
one, P.19,20 Allyl isomerization occurs when C3 in 3a (or T)
attaches to Cu to form a π-allyl moiety in P (or 1a), and then
C1 detaches.
(5) Bertz, S. H.; Cope, S.; Dorton, D.; Murphy, M.; Ogle, C. A. Angew. Chem.,
Int. Ed. 2007, 46, 7082–5.
(6) Bartholomew, E. R.; Bertz, S. H.; Cope, S.; Dorton, D. C.; Murphy, M.;
Ogle, C. A. Chem. Commun. 2008, 1176–7.
(7) Ga¨rtner, T.; Henze, W.; Gschwind, R. M. J. Am. Chem. Soc. 2007, 129,
11362–3.
(8) (a) Lipshutz, B. H.; Ellsworth, E. L.; Dimock, S. H.; Smith, R. A. J. J. Am.
Chem. Soc. 1990, 112, 4404–10. (b) Whitesides, G. M.; Fleming, J. S.
J. Am. Chem. Soc. 1967, 89, 2855–9.
(9) Pavlyuk, O. V.; Mykhalichko, B. M.; Mys’kiv, M. G. Russ. J. Coord. Chem.
2004, 30, 159–83.
(10) Bertz, S. H.; Carlin, C. M.; Deadwyler, D. A.; Murphy, M. D.; Ogle, C. A.;
Seagle, P. H. J. Am. Chem. Soc. 2002, 124, 13650–1. See also ref 4.
(11) Murphy, M. D.; Ogle, C. A.; Bertz, S. H. Chem. Commun. 2005, 854–6.
(12) Ashby, E. C.; Watkins, J. J. J. Am. Chem. Soc. 1977, 99, 5312–7.
(13) Canisius, J.; Mobley, T. A.; Berger, S.; Krause, N. Chem.—Eur. J. 2001,
7, 2671–5. These authors observed a value of 2J ) 12 Hz for an R-enone-
dimethylcuprate(I) π-complex (from B) with a similar C-Cu-C angle.
(14) π-Complexes of R-enones with A or B are usually considered to be η2 CuI
species;10,13 however, some authors consider them η3 CuIII complexes.3
(15) Snyder, J. P. J. Am. Chem. Soc. 1995, 117, 11025–6.
(16) Bertz, S. H. J. Am. Chem. Soc. 1991, 113, 5470–1.
(17) Each end of the allyl group in labeled 1a, 2a, and 3a had 50 atom % 13C.
In principle, 3a can be formed via capture of cyanide by 1a or
via rearrangement of 3a′, produced by the same concerted mech-
anism that yields trans-EtMe2Cu(X)Li from Me2CuLi·LiX and EtI.5
The trans to cis isomerization is observed for some X, but not X
) CN; hence, allyl must play a key role if 3a′ is involved.
Rearrangement could occur by a mechanism analogous to one of
those above. For example, coordination of the double bond in 3a′
gives a pyramidal intermediate with Me in the apical position (i.e.,
P′), which affords 3a when this methyl group replaces the
coordinated double bond trans to the cyano group.
¨
(18) Bertz, S. H.; Nilsson, K.; Davidsson, O.; Snyder, J. P. Angew. Chem., Int.
Ed. 1998, 37, 314–7.
(19) Clark, G. R.; Skelton, B. W.; Waters, T. N. J. Chem. Soc., Dalton Trans.
1976, 1528–36.
(20) Xifra, R.; Ribas, X.; Llobet, A.; Poater, A.; Duran, M.; Sola`, M.; Stack,
T. D. P.; Benet-Buchholz, J.; Donnadieu, B.; Mah´ıa, J.; Parella, T.
Chem.—Eur. J. 2005, 11, 5146–56.
Addition of cinnamyl chloride to A in THF-d8 at -100 °C
resulted in a mixture of 1d and 2d (1:2, 0.5 h; 1:1, 2 h; 2:1, 3.5 h).
JA801186C
9
J. AM. CHEM. SOC. VOL. 130, NO. 34, 2008 11245