Kinetics and mechanism of ketone enolization mediated by magnesium bis(hexamethyldisilazide)

Article abstract of DOI:10.1021/ja064927w

Magnesium bis(hexamethyldisilazide), Mg(HMDS)₂, reacts with substoichiometric amounts of propiophenone in toluene solution at ambient temperature to form a 74:26 mixture of the enolates (E)- and (Z)-[(HMDS) ₂Mg₂(μ-HMDS){μ-

Full text of DOI:10.1021/ja064927w

Published on Web 09/26/2006
Kinetics and Mechanism of Ketone Enolization Mediated by
Magnesium Bis(hexamethyldisilazide)
Xuyang He, J. Jacob Morris, Bruce C. Noll, Seth N. Brown,* and
Kenneth W. Henderson*
Contribution from the Department of Chemistry and Biochemistry, UniVersity of Notre Dame,
Notre Dame, Indiana 46556-5670
Received July 18, 2006; E-mail: khenders@nd.edu
Abstract: Magnesium bis(hexamethyldisilazide), Mg(HMDS)2, reacts with substoichiometric amounts of
propiophenone in toluene solution at ambient temperature to form a 74:26 mixture of the enolates (E)- and
(Z)-[(HMDS)2Mg2(µ-HMDS){µ-OC(Ph)dCHCH3}], (E)-1 and (Z)-1, which contain a pair of three-coordinate
metal centers bridged by an amide and an enolate group. The compositions of (E)-1 and (Z)-1 were
confirmed by solution NMR studies and also by crystallographic characterization in the solid state. Rate
studies using UV-vis spectroscopy reveal the rapid and complete formation of a reaction intermediate, 2,
between the ketone and magnesium, which undergoes first-order decay with rate constants independent
of the concentration of excess Mg(HMDS)2 (∆Hq ) 17.2 ( 0.8 kcal/mol, ∆Sq ) -11 ( 3 cal/mol‚K). The
intermediate 2 has been characterized by low-temperature 1H NMR, diffusion-ordered NMR, and IR
spectroscopy and investigated by computational studies, all of which are consistent with the formulation of
2 as a three-coordinate monomer, (HMDS)2Mg{η1-OdC(Ph)CH2CH3}. Further support for this structure is
provided by the synthesis and structural characterization of two model ketone complexes, (HMDS)2Mg-
(η1-OdCtBu2) (3) and (HMDS)2Mg{η1-OdC(tBu)Ph} (4). A large primary deuterium isotope effect (kH/kD )
18.9 at 295 K) indicates that proton transfer is the rate-limiting step of the reaction. The isotope effect
displays a strong temperature dependence, indicative of tunneling. In combination, these data support the
mechanism of enolization proceeding through the single intermediate 2 via intramolecular proton transfer
from the R carbon of the bound ketone to the nitrogen of a bound hexamethyldisilazide.
Introduction
complexity underlying these seemingly simple transformations.
In this regard, the systematic spectroscopic and mechanistic
The development of synthetic methods for the selective
generation of enolate anions is an area of active interest due to
the great utility of these synthons in carbon-carbon bond
forming reactions.1 Lithium amides, such as lithium diisopropyl-
amide (LDA), lithium 2,2,6,6-tetramethylpiperidide (LTMP),
and lithium hexamethyldisilazide (LiHMDS), have evolved to
become the dominant bases of choice to mediate kinetic enoliza-
tion reactions.2,3 These reagents are highly selective in the
deprotonation of ketonic substrates as they combine the proper-
ties of relatively large steric encumbrance,4 strong Brønsted
basicity,5 and low nucleophilicity. Numerous theoretical,6
solid-state,7,8 solution,9 and kinetic investigations10 of lithium-
mediated enolizations have revealed an astonishing degree of
studies of lithium amide-mediated enolization reactions by
Collum are particularly revealing.11-13 The picture that has
emerged from this work is that no single mechanism is adequate
(4) For solution aggregation studies of lithium amides, see: (a) Collum, D. B.
Acc. Chem. Res. 1993, 26, 227. (b) Collum, D. B. Acc. Chem. Res. 1992,
25, 448. (c) Lucht, B. L.; Collum, D. B. Acc. Chem. Res. 1999, 32, 1035.
(d) Hilmersson, G. Chem.sEur. J. 2000, 6, 3069. (e) Zune, C.; Dubois,
P.; Grandjean, J.; Kriz, J.; Dybal, J.; Lochmann, L.; Janata, M.; Vlcek, P.;
Jerome, R. Macromolecules 1999, 32, 5477.
(5) For data related to quantitative ion-pair acidity measurements using lithium
amides, see: (a) Furlong, J. J. P.; Lewkowicz, E. S.; Nudelman, N. S. J.
Chem. Soc., Perkin Trans. 2 1990, 1461. (b) Streitwieser, A.; Juaristi, E.;
Nebenzahl, L. In ComprehensiVe Carbanion Chemistry; Buncel, E., Durst,
T., Eds.; Elsevier: New York, 1980; Chapter 7. (c) Fraser, R. R.; Baignee,
A.; Bresse, M.; Hata, K. Tetrahedron Lett. 1982, 23, 4195. (d) Fraser, R.
R.; Bresse, M.; Mansour, T. S. J. Chem. Soc., Chem. Commun. 1983, 620.
(e) Fraser, R. R.; Bresse, M.; Mansour, T. S. J. Am. Chem. Soc. 1983, 105,
7790. (f) Fraser, R. R.; Mansour, T. S. J. Org. Chem. 1984, 49, 3442. (g)
Fraser, R. R.; Mansour, T. S.; Savard, S. J. Org. Chem. 1985, 50, 3232.
(h) Fraser, R. R.; Mansour, T. S. Tetrahedron Lett. 1986, 27, 331.
(6) (a) Jemmis, E. D.; Gopakumar, G. In The Chemistry of Organolithium
Compounds; Rappoport, Z., Patai, S., Eds; Wiley: New York, 2004; Chapter
1. (b) Moreland, D. W.; Dauben, W. G. J. Am. Chem. Soc. 1985, 107,
2264. (c) Pratt, L. M.; et al. J. Org. Chem. 2003, 68, 6387. (d) Yakimansky,
A. V.; Muller, A. H. E. J. Am. Chem. Soc. 2001, 123, 4932. (e) Weiss, H.;
Yakimansky, A. V.; Muller, A. H. E. J. Am. Chem. Soc. 1996, 118, 8897.
(f) McKee, M. L. J. Am. Chem. Soc. 1987, 109, 559.
(1) (a) Heathcock, C. H. In ComprehensiVe Organic Synthesis; Trost, B. M.,
Fleming, I., Eds.; Pergamon: Oxford, 1991; Vol. 2, p 181. (b) Caine, D.
In ComprehensiVe Organic Synthesis; Trost, B. M., Fleming, I., Eds.;
Pergamon: Oxford, 1991; Vol. 3, p 1. (c) d’Angelo, J. Tetrahedron 1976,
32, 2979. (d) Morrison, J. D., Ed. Asymmetric Synthesis; Academic: New
York, 1983; Vols. 2 and 3. (e) Snieckus, V. Chem. ReV. 1990, 90, 879.
(2) (a) Berrisford, D. J. Angew. Chem., Int. Ed. Engl. 1995, 34, 178. (b)
Clayden, J. Organolithiums: SelectiVity for Synthesis; Pergamon: Oxford,
2002. (c) Schlosser, M., Ed. Organometallics in Synthesis; Wiley: New
York, 2002; Chapter 1. (d) Wakefield, B. J. Organolithium Methods;
Academic: London, 1988.
(7) For reviews on lithium enolate structural chemistry, see: (a) Seebach, D.
Angew. Chem., Int. Ed. Engl. 1988, 27, 1624. (b) Williard, P. G. In
ComprehensiVe Organic Synthesis; Trost, B. M., Fleming, I., Eds.;
Pergamon: Oxford, 1991; Vol. 1, p 1. (c) Fromm, K. M.; Gueneau, A. D.
Polyhedron 2002, 23, 1479.
(3) For reviews on enantioselective deprotonations mediated by chiral lithium
amide bases, see: (a) O’Brien, P. J. Chem. Soc., Perkin Trans. 1 1998,
1439. (b) O’Brien, P. J. Chem. Soc., Perkin Trans. 1 2001, 95. (c) Cox, P.
J.; Simpkins, N. S. Tetrahedron: Asymmetry 1991, 2, 1.
9
10.1021/ja064927w CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 13599-13610
13599

A R T I C L E S
He et al.
to describe these proton transfer processes. Almost every aspect
of an enolization may affect its mechanism, including the steric
and electronic nature of the base and the substrate,14 the aggre-
gation state of the metal amide, the temperature, the solvent,
and the presence of additives, such as salts or cosolvents.15,16
Another complication is that the mechanism may change as the
reaction progresses since the composition of the species present
in solution can vary with the extent of reaction.11g These
complexities are characteristic of organolithium chemistry,
where the relatively ionic bonding, combined with the metal’s
monovalency, leads to intra- and intermolecular fluxionality.17,18
Our group has had a longstanding interest in the use of
magnesium bis(amides), Mg(NR2)2, as alternatives to lithium
amides in enolization reactions.19,20 Using a divalent metal
reagent allows one greater latitude in tuning the ancillary ligands.
Judicious choice of ancillary ligands, coupled with the greater
covalency of bonds to magnesium compared to lithium,21 allows
access to systems with simpler solution aggregation behavior
and higher thermal stability.22 We and others have demonstrated
the utility of magnesium bis(amide) bases in the regio- and
stereoselective deprotonation of ketones,23,24 as well as with
other substrates, including substituted aromatics,25 cubanes25a,26
and indoles.27 Eaton has shown that alkylmagnesium amides
are useful in the regioselective deprotonation of weakly acidic
substrates, such as cyclopropane and cyclobutane carboxam-
ides.28 We have also developed a series of homochiral magne-
sium bis(amide) reagents that have proven to be highly selective
in the enantioselective deprotonation of conformationally locked
ketones.29 Other emerging applications of homochiral magne-
sium amide reagents in asymmetric synthesis include amina-
tions,30 conjugate additions,31 alkylations,32 and reductions.33
Mixed alkali/alkaline earth metal amide complexes34 have also
very recently been highlighted as regioselective bases in the
deprotonation of ketones,35 arenes,36 and metallocenes.37 In
addition, magnesium amides have been employed with some
success as catalysts in anionic38 and ring-opening polymerization
reactions.39
(8) (a) Williard, P. G.; Carpenter, G. B. J. Am. Chem. Soc. 1985, 107, 3345.
(b) Seebach, D.; Amstutz, R.; Laube, T.; Schweizer, W. B.; Dunitz, J. D.
J. Am. Chem. Soc. 1985, 107, 5403. (c) Laube, T.; Dunitz, J. D.; Seebach,
D. HelV. Chim. Acta 1985, 68, 1373. (d) Jastrzebski, J. T. B. H.; Van Koten,
G.; Christophersen, M. J. N.; Stam, C. H. J. Organomet. Chem. 1985, 292,
319. (e) Amstutz, R.; Dunitz, J. D.; Laube, T.; Schweizer, W. B.; Seebach,
D. Chem. Ber. 1986, 119, 434. (f) Williard, P. G.; Hintze, M. J. J. Am.
Chem. Soc. 1987, 109, 5539.
(9) For solution studies of lithium enolates, see: (a) Sato, D.; Kawasaki, H.;
Shimada, I.; Arata, Y.; Okamura, K.; Date, T.; Koga, K. J. Am. Chem.
Soc. 1992, 114, 761. (b) Suzuki, M.; Koyama, H.; Noyori, R. Bull. Chem.
Soc. Jpn. 2004, 77, 259. (c) Suzuki, M.; Koyama, H.; Noyori, R.
Tetrahedron 2004, 60, 1571.
(21) Wakefield, B. J. Organomagnesium Methods in Organic Synthesis;
Academic: London, 1994.
(10) (a) Held, G.; Xie, L. Microchem. J. 1997, 55, 261. (b) Xie, L.; Saunders,
W. H., Jr. J. Am. Chem. Soc. 1991, 113, 3123. (c) Beutelman, H. P.; Xie,
L.; Saunders, W. H., Jr. J. Org. Chem. 1989, 54, 1703. (d) Majewski, M.;
Nowak, P. Tetrahedron Lett. 1998, 39, 1661.
(22) (a) Markies, P. R.; Akkerman, O. S.; Bickelhaupt, F.; Smeets, W. J. J.;
Spek, A. L. AdV. Organomet. Chem. 1991, 32, 147. (b) Westerhausen, M.
Angew. Chem., Int. Ed. 2001, 40, 2975. (c) Lindsell, W. E. In Compre-
hensiVe Organometallic Chemistry; Wilkinson, G., Stone, F. G. A., Eds.;
Pergamon: Oxford, 1982; Vol. 1, Chapter 4.2.
(11) For kinetic studies, see: (a) McNeil, A. J.; Collum, D. B. J. Am. Chem.
Soc. 2005, 127, 5655. (b) Zhao, P. J.; Condo, A.; Keresztes, I.; Collum, D.
B. J. Am. Chem. Soc. 2004, 126, 3113. (c) Zhao, P. J.; Lucht, B. L.; Kenkre,
S. L.; Collum, D. B. J. Org. Chem. 2004, 69, 242. (d) Zhao, P. J.; Collum,
D. B. J. Am. Chem. Soc. 2003, 125, 14411. (e) Zhao, P. J.; Collum, D. B.
J. Am. Chem. Soc. 2003, 125, 4008. (f) Sun, X. F.; Collum, D. B. J. Am.
Chem. Soc. 2000, 122, 2452. (g) Sun, X. F.; Collum, D. B. J. Am. Chem.
Soc. 2000, 122, 2459. (h) Sun, X. F.; Kenkre, S. L.; Remenar, J. F.;
Gilchrist, J. H.; Collum, D. B. J. Am. Chem. Soc. 1997, 119, 4765.
(12) For solution NMR studies, see: (a) Kim, Y. J.; Bernstein, M. P.; Roth, A.
S. G.; Romesberg, F. E.; Williard, P. G.; Fuller, D. J.; Harrison, A. T.;
Collum, D. B. J. Org. Chem. 1991, 56, 4435. (b) Galiano-Roth, A. S.;
Kim, Y. J.; Gilchrist, J. H.; Harrison, A. T.; Fuller, D. J.; Collum, D. B. J.
Am. Chem. Soc. 1991, 113, 5053. (c) McNeil, A. J.; Toombes, G. E. S.;
Chandramouli, S. V.; Vanasse, B. J.; Ayers, T. A.; O’Brien, M. K.;
Lobkovsky, E.; Gruner, S. M.; Marohn, J. A.; Collum, D. B. J. Am. Chem.
Soc. 2004, 126, 5938.
(13) For selectivity and theoretical studies, see: (a) Romesberg, F. E.; Collum,
D. B. J. Am. Chem. Soc. 1995, 117, 2166. (b) Hall, P. L.; Gilchrist, J. H.;
Collum, D. B. J. Am. Chem. Soc. 1991, 113, 9571.
(14) (a) Xie, L. F.; Vanlandeghem, K.; Isenberger, K. M.; Bernier, C. J. Org.
Chem. 2003, 68, 641. (b) Xie, L. F.; Isenberger, K. M.; Held, G.; Dahl, L.
M. J. Org. Chem. 1997, 62, 7516.
(15) (a) Seebach, D.; Beck, A. K.; Studer, A. Mod. Synth. Methods 1995, 7, 1.
(b) Loupy, A.; Tchoubar, B. Salt Effects in Organic and Organometallic
Chemistry; VCH: New York, 1991. (c) Pratt, L. M. Mini-ReViews in
Organic Chemistry 2004, 1, 209.
(23) He, X.; Allan, J. F.; Noll, B. C.; Kennedy, A. R.; Henderson, K. W. J. Am.
Chem. Soc. 2005, 127, 6920.
(24) (a) Bonafoux, D.; Bordeau, M.; Biran, C.; Cazeau, P.; Dunogues, J. J. Org.
Chem. 1996, 61, 5532. (b) Lesse`ne, G.; Tripoli, R.; Cazeau, P.; Biran, C.;
Bordeau, M. Tetrahedron Lett. 1999, 40, 4037. (c) Bonafoux, D.; Bordeau,
M.; Biran, C.; Dunogues, J. Synth. Commun. 1998, 28, 93. (d) Bonafoux, D.;
Bordeau, M.; Biran, C.; Dunogues, J. J. Organomet. Chem. 1995, 493, 27.
(25) (a) Eaton, P. E.; Lee, C. H.; Xiong, Y. H. J. Am. Chem. Soc. 1989, 111,
8016. (b) Kano, T.; Takai, J.; Tokuda, O.; Maruoka, K. Angew. Chem.,
Int. Ed. 2005, 44, 3055. (c) Ooi, T.; Uematsu, Y.; Maruoka, K. J. Org.
Chem. 2003, 68, 4576.
(26) Eaton, P. E.; Xiong, Y. H.; Gilardi, R. J. Am. Chem. Soc. 1993, 115, 10195.
(27) Kondo, Y.; Yoshida, A.; Sakamoto, T. J. Chem. Soc., Perkin Trans. 1 1996,
2331.
(28) (a) Zhang, M. X.; Eaton, P. E. Angew. Chem., Int. Ed. 2002, 114, 2273.
(b) Eaton, P. E.; Zhang, M. X.; Komiya, N.; Yang, C. G.; Steele, I.; Gilardi,
R. Synlett 2003, 9, 1275.
(29) (a) Henderson, K. W.; Kerr, W. J.; Moir, J. H. Tetrahedron 2002, 58, 4573.
(b) Anderson, J. D.; Garcia, P. G.; Hayes, D.; Henderson, K. W.; Kerr, W.
J.; Moir, J. H.; Fondekar, K. P. Tetrahedron Lett. 2001, 42, 7111. (c)
Henderson, K. W.; Kerr, W. J.; Moir, J. H. Synlett 2001, 1253. (d)
Henderson, K. W.; Kerr, W. J.; Moir, J. H. Chem. Commun. 2001, 1722.
(e) Henderson, K. W.; Kerr, W. J.; Moir, J. H. Chem. Commun. 2000,
479. (f) Carswell, E. L.; Hayes, D.; Henderson, K. W.; Kerr, W. J.; Russell,
C. J. SynLett 2003, 1017. (g) Bassindale, M. J.; Crawford, J. J.; Henderson,
K. W.; Kerr, W. J. Tetrahedron Lett. 2004, 45, 4175.
(16) For studies of mixed anions, see: (a) Romesberg, F. E.; Collum, D. B. J.
Am. Chem. Soc. 1994, 116, 9187. (b) Gilchrist, J. H.; Harrison, A. T.; Fuller,
D. J.; Collum, D. B. Magn. Reson. Chem. 1992, 30, 855. (c) Hall, P. L.;
Gilchrist, J. H.; Harrison, A. T.; Fuller, D. J.; Collum, D. B. J. Am. Chem.
Soc. 1991, 113, 9575. (d) Henderson, K. W.; Dorigo, A. E.; Liu, Q. Y.;
Williard, P. G.; Schleyer, P. v. R.; Bernstein, P. R. J. Am. Chem. Soc.
1996, 118, 1339. (e) Sugasawa, K.; Shindo, M.; Noguchi, H.; Koga, K.
Tetrahedron Lett. 1996, 37, 7377. (f) Henderson, K. W.; Walther, D. S.;
Williard, P. G. J. Am. Chem. Soc. 1995, 117, 8680. (g) Henderson, K. W.;
Dorigo, A. E.; Liu, Q. Y.; Williard, P. G.; Bernstein, P. R. Angew. Chem.,
Int. Ed. Engl. 1996, 35, 1322. (h) Kim, Y. J.; Streitwieser, A. Org. Lett.
2002, 4, 573.
(17) (a) Stey, T.; Stalke, D. In The Chemistry of Organolithium Compounds;
Rappoport, Z., Patai, S., Eds.; Wiley: New York, 2004; Chapter 2. (b)
Gregory, K.; Schleyer, P. v. R.; Snaith, R. AdV. Inorg. Chem. 1991, 37,
47. (c) Weiss, E. Angew Chem., Int. Ed. Engl. 1993, 32, 1501. (d) Beswick,
M. A.; Wright, D. S. In ComprehensiVe Organometallic Chemistry; Abel,
E. W., Stone, F. G. A., Wilkinson, G., Eds.; Elsevier: Oxford, 1995; Vol.
1, Chapter 1. (e) Boche, G. Angew. Chem., Int. Ed. Engl. 1989, 28, 277.
(f) Mulvey, R. E. Chem. Soc. ReV. 1998, 27, 339.
(30) Evans, D. A.; Nelson, S. G. J. Am. Chem. Soc. 1997, 119, 6452.
(31) (a) Sibi, M. P.; Asano, Y. J. Am. Chem. Soc. 2001, 123, 9708. (b) Bunnage,
M. E.; Davies, S. G.; Goodwin, C. J.; Walters, J. A. S. Tetrahedron:
Asymmetry 1994, 5, 35.
(32) Yong, K. H.; Taylor, N. J.; Chong, J. M. Org. Lett. 2002, 4, 3553.
(33) Yong, K. H.; Chong, J. M. Org. Lett. 2002, 4, 4139.
(34) (a) Mulvey, R. E. Chem. Commun. 2001, 1049. (b) Mulvey, R. E.
Organometallics 2006, 25, 1060. (c) Clegg, W.; Henderson, K. W.; Mulvey,
R. E.; O’Neil, P. A. J. Chem. Soc., Chem. Commun. 1993, 969.
(35) (a) He, X.; Noll, B. C.; Beatty, A.; Mulvey, R. E.; Henderson, K. W. J.
Am. Chem. Soc. 2004, 126, 7444. (b) Hevia, E.; Henderson, K. W.;
Kennedy, A. R.; Mulvey, R. E. Organometallics 2006, 25, 1778.
(36) (a) Krasovskiy, A.; Krasovskaya, V.; Knochel, P. Angew. Chem., Int. Ed.
2006, 45, 2958. (b) Armstrong, D. R.; Kennedy, A. R.; Mulvey, R. E.;
Rowlings, R. B. Angew. Chem., Int. Ed. 1999, 38, 131. (c) Andrews, P.
C.; Kennedy, A. R.; Mulvey, R. E.; Raston, C. L.; Roberts, B. A.; Rowlings,
R. B. Angew. Chem., Int. Ed. 2000, 39, 1960.
(37) (a) Clegg, W.; Henderson, K. W.; Kennedy, A. R.; Mulvey, R. E.; O’Hara,
C. T.; Rowlings, R. B.; Tooke, D. M. Angew. Chem., Int. Ed. 2001, 40,
3902. (b) Hevia, E.; Honeyman, G. W.; Kennedy, A. R.; Mulvey, R. E.;
Sherrington, D. C. Angew. Chem., Int. Ed. 2005, 44, 68. (c) Andrikopoulos,
P. C.; Armstrong, D. R.; Clegg, W.; Gilfillan, C. J.; Hevia, E.; Kennedy,
A. R.; Mulvey, R. E.; O’Hara, C. T.; Parkinson, J. A.; Tooke, D. M. J.
Am. Chem. Soc. 2004, 126, 11612.
(18) Sapse, A. M., Schleyer, P. v. R., Eds. Lithium Chemistry, A Theoretical
and Experimental OVerView; John Wiley & Sons: New York, 1995.
(19) Henderson, K. W.; Kerr, W. J. Chem. Eur. J. 2001, 7, 3430.
(20) Henderson, K. W.; Allan, J. F.; Kennedy, A. R. J. Chem. Soc., Chem.
Commun. 1997, 1149.
(38) Harder, S.; Feil, F. Organometallics 2002, 21, 2268.
9
13600 J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006

Ketone Enolization Mediated by Mg(HMDS)2
A R T I C L E S
Despite the increasing interest in these reagents, only very
limited mechanistic investigations have appeared,20,40 and these
have mainly focused on the structural characterization of the
metal enolate products of the reactions.23,41 We were therefore
interested in conducting a detailed kinetic and mechanistic
investigation of magnesium bis(amide)-mediated enolization
reactions, using magnesium bis(hexamethyldisilazide), Mg-
(HMDS)2, as a prototypical magnesium amide.
Mg(HMDS)2 has previously been used as a selective Brønsted
base in enolization reactions23,24a,40,41b and is readily prepared
in a highly pure crystalline form.42 Westerhausen has established
that this compound exists as a mixture of monomers and amide-
bridged dimers in arene solutions, with the dimer predominating
at ambient temperatures (eq 1).43 The existence of a well-
characterized monomer-dimer equilibrium makes this system
particularly amenable to investigating any influence of base
aggregation in the deprotonation reaction. Propiophenone was
chosen as a representative ketone since it has been used
extensively for lithium-mediated methodology development and
because stereochemical information may be obtained using this
substrate.14,44 Indeed, we recently reported in a preliminary
communication that Mg(HMDS)2 reacts with propiophenone at
ambient temperature in toluene solution to give predominantly
the (E)-enolate (70:30 E/Z).23 This is an intriguing result as the
corresponding lithium base LiHMDS is known to be highly (Z)-
selective (2:98 E/Z).44a We also reported that the molecular
structure of the amidomagnesium enolate [(HMDS)Mg{µ-OC-
(Ph)dCHMe}‚THF]2 is a centrosymmetric dimer with a central
Mg2O2 ring.23
Figure 1. 1H NMR spectra in toluene-d8 at 298 K of (E)-1 and (Z)-1
prepared in the presence of excess Mg(HMDS)2. The inset shows the tri-
methylsilyl region from the stoichiometric preparation of (E)-1 and (Z)-1.
The enolates formed are not the previously characterized bis-
(enolate)-bridged dimers [(HMDS)Mg{µ-OC(Ph)dCHMe}]2.23
Instead, the presence of excess Mg(HMDS)2 appears to favor
formation of the bimetallics bridged by one enolate and one
amide, as judged by the appearance of signals due to terminal
and bridging HMDS groups in a 2:1 ratio for each isomer. The
presence of two isomers is attributed to (E)- and (Z)-enolates,
in a 74:26 ratio, similar to the ratio of 70:30 measured using
0.75 molar equiv of ketone.23 The identity of the stereoisomers
is assigned by analogy with nOe experiments on other Mg
enolates23 and is confirmed by trapping the enolates as the silyl
enol ethers. Direct trapping of the magnesium enolates with Me3-
SiCl proved to be very slow in toluene solution. However,
transmetalation using BuLi followed by the addition of Me3-
SiCl and THF results in quantitative formation of the silyl enol
ethers. The stereochemistry of the enolates was confirmed by
comparison with authentic samples, and the E/Z ratio was
determined to be 72:28 by GC analyses.44a
Reaction of 2 molar equiv of Mg(HMDS)2 with propiophe-
none in toluene-d8 was consistent with the formation of (E)-1
and (Z)-1 (inset of Figure 1). Furthermore, this system proved
to be amenable to crystallization, and subsequent analysis by
X-ray diffraction confirmed their identities in the solid state
(Figure 2). The stereoisomers cocrystallize within single crystals
giving rise to disorder of the enolate fragment. Similarly,
cocrystallization of the (E)- and (Z)-enolates was found
for the bis(enolate)-bridged dimer [(HMDS)Mg{µ-OC-
(Ph)dCHMe}‚THF]2.23 In the present instance, the best fit
We now report on the mechanistic details of the deprotonation
reaction using a combination of spectroscopic, kinetic, compu-
tational, and crystallographic methods. These studies provide
an unusually coherent and complete description for a metal base-
mediated enolization reaction.
Results
Reaction Product Studies. Propiophenone reacts quantita-
tively with excess Mg(HMDS)2 in toluene solution within 15
min at ambient temperature to form a mixture of two magnesium
enolates, (E)- and (Z)-[(HMDS)2Mg2(µ-HMDS){µ-OC(Ph)d
CHCH3}], (E)-1 and (Z)-1 (eq 2 and Figure 1).
(42) No¨th, H.; Schlosser, D. Inorg. Chem. 1983, 22, 2700.
(39) (a) Chisholm, M. H.; Gallucci, J.; Phomphrai, K. Chem. Commun. 2003,
48. (b) Chisholm, M. H.; Gallucci, J. C.; Phomphrai, K. Inorg. Chem. 2005,
44, 8004. (c) Dove, A. P.; Gibson, V. C.; Marshall, E. L.; White, A. J. P.;
Williams, D. J. Dalton Trans. 2004, 570.
(43) (a) Westerhausen, M. Coord. Chem. ReV. 1998, 176, 157. (b) Westerhausen,
M.; Schwarz, W. Inorg. Chem. 1991, 30, 96.
(44) (a) Heathcock, C. H.; Buse, C. T.; Kleschick, W. A.; Pirrung, M. C.; Sohn,
J. E.; Lampe, J. J. Org. Chem. 1980, 45, 1066. (b) Brown, H. C.; Dhar, R.
J.; Bakshi, R. K.; Pandiarajan, P. K.; Singaram, B. J. Am. Chem. Soc. 1989,
111, 3441. (c) Fataftah, Z. A.; Kopka, I. E.; Rathke, M. W. J. Am. Chem.
Soc. 1980, 102, 3959. (d) Heathcock, C. H.; Davidsen, S. K.; Hug, K. T.;
Flippin, L. A. J. Org. Chem. 1986, 51, 3027. (e) House, H. O.; Czuba, L.
J.; Gall, M.; Lomstead, H. D. J. Org. Chem. 1969, 34, 2324.
(40) Allan, J. F.; Henderson, K. W.; Kennedy, A. R. Chem. Commun. 1999,
1325.
(41) (a) Allan, J. F.; Clegg, W.; Henderson, K. W.; Horsburgh, L.; Kennedy,
A. R. J. Organomet. Chem. 1998, 559, 173. (b) Allan, J. F.; Henderson,
K. W.; Kennedy, A. R.; Teat, S. J. Chem. Commun. 2000, 1059.
9
J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006 13601

A R T I C L E S
He et al.
Figure 3. Plot of kobs for the disappearance of 2 versus total concentration
of magnesium bis(hexamethyldisilazide) in toluene at 21.8 °C.
Figure 2. Ellipsoid view (50% probabilities) of 1 with hydrogen atoms
omitted for clarity. The methine C(20′) and methyl C(21′) positions for the
minor component (E)-1 are shown as hollow ellipsoids.
polarity has little effect on the rate of decay of 2 with kobs
)
(4.93 ( 0.07) × 10-3 s-1 at 21.8 °C in CH2Cl2 (ꢀ ) 8.9)
identical within experimental error to the value obtained in
toluene (ꢀ ) 2.4).
model for the disorder indicates a 22:78 ratio of the (E)- and
(Z)-isomers within the crystal studied. This ratio differs from
that observed from the in situ preparation, suggesting that
preferential crystallization is occurring. Indeed, NMR analysis
of a representative sample of crystals deposited indicated an
approximately 20:80 ratio for the (E)- and (Z)-enolates. The
crystals were isolated in an 11% overall yield, suggesting that
the (E)-enolate is the more soluble isomer and that it is retained
in solution.
Deuterium Isotope Effect and Activation Parameters. To
measure a kinetic isotope effect on the enolization reaction,
R-deuterated propiophenone was prepared by base-catalyzed
exchange with CH3OD.46 Propiophenone-d2 reacts much more
slowly with Mg(HMDS)2 than the protio compound, with kH/kD
) 18.9 ( 0.6 at 21.8 °C. This corresponds to a change in half-
life from approximately 2.3 to 42 min. (The stereoselectivity
of enolization is scarcely affected by deuteration, with PhCOCD2-
CH3 reacting to give a 77:23 mixture of (E)- and (Z)-enolates
by NMR.) The primary isotope effect clearly indicates that C-H
bond cleavage is involved in the rate-determining step and is
large enough to suggest that tunneling is important in this
reaction.47
The composition of 1 is somewhat unusual, containing two
different types of bridging groups between the metals. However,
the previously characterized amido(alkoxide) tBuMg(µ-OtBu)-
(µ-TMP)Mg(TMP) also contains three-coordinate magnesium
centers with a central Mg2ON ring core.45 The Mg2ON ring in
1 is slightly puckered, with a Mg(1)-O(1)-Mg(2)-N(3)
dihedral angle of 6.8°, and all four ligand groups are tilted out
of the mean ring plane. The metrical parameters for 1 are in
accord with expectations (Table S1).
Further evidence for tunneling is provided by the temperature
dependence of the isotope effect.48 Experiments were run over
the temperature range of 21.8-65.4 °C, and the isotope effect
decreases markedly with increasing temperature (for example,
to 10.1 at 45.2 °C), as expected for a reaction involving tunnel-
Rate Studies. Mixing the colorless reactants, propiophenone
and Mg(HMDS)2, in toluene at room temperature immediately
produces a bright yellow solution, due to a significant increase
in absorbance in the near-UV region. The appearance of color
is consistent with the formation of a pre-enolization intermediate,
hereafter referred to as 2, and the color decays over the course
of a few minutes, concomitant with formation of the enolate
product. The absorbance due to 2 is proportional to the con-
centration of added propiophenone in the presence of excess
Mg(HMDS)2. The absorbance at 330 nm decays exponentially
(Figure S1), with a half-life of 139 s (kobs ) (4.99 ( 0.11) ×
10-3 s-1) under these conditions (0.048 M Mg(HMDS)2, 4.8 ×
10-3 M propiophenone, 21.8 °C).
D
H
ing. Arrhenius plots (Figure 4) give Ea - Ea ) 2.6 ( 0.3
kcal/mol and AH/AD ) 0.20 ( 0.08. The fact that the EaD - Ea
H
value significantly exceeds the classical maximum of approx-
imately 1.2 kcal/mol and that the prefactor ratio is significantly
smaller than the classical minimum of 0.547b is strong evidence
that tunneling is in fact occurring in the proton transfer step.47,48
Activation parameters for the enolization reaction of the protio
compound were derived from an Eyring plot (Figure S2), giving
∆Hq ) 17.2 ( 0.8 kcal/mol, ∆Sq ) -11 ( 3 cal/mol‚K, and
∆Gq (295 K) ) 20.4 ( 1.7 kcal/mol.
IR and NMR Spectroscopic Characterization of the
Enolization Intermediate 2. Addition of propiophenone (R-
deuterated to retard enolization) to a toluene solution of Mg-
(HMDS)2 leads to the appearance of a single IR band in the
CdO bond stretching region at 1657 cm-1, which diminishes
with time as the enolate is formed. In comparison, the CdO
As expected for a first-order reaction, the observed rate
constant is insensitive to the initial concentration of the limiting
reagent, propiophenone, in the range 4.87 × 10-4 to 2.92 ×
10-3 M. More surprisingly, kobs is also insensitive to the
concentration of the excess reagent, Mg(HMDS)2 (Figure 3).
These data indicate that 2 decays to enolate in a unimolecular
process that is independent of the presence of both monomeric
and dimeric forms of Mg(HMDS)2 in the mixture. Solvent
bond stretch in free propiophenone-d2 appears at 1693 cm-1
.
(46) Maestri, A. G. PhD Thesis, University of Notre Dame, 2003, Chapter 4.
(47) (a) Bell, R. P. The Proton in Chemistry, 2nd ed.; Cornell University: Ithaca,
New York, 1973; Chapter 12. (b) Melander, L.; Saunders, W. H. Reaction
Rates of Isotopic Molecules; John Wiley & Sons: New York, 1980; Chapter
5.2.
(48) (a) Kaldor, S. B.; Saunders, W. H. J. Am. Chem. Soc. 1979, 101, 7594. (b)
Garcia-Garibay, M. A.; Gamarnik, A.; Bise, R.; Pang, L.; Jenks, W. S. J.
Am. Chem. Soc. 1995, 117, 10264.
(45) Conway, B.; Hevia, E.; Kennedy, A. R.; Mulvey, R. E.; Weatherstone, S.
Dalton Trans. 2005, 1532.
9
13602 J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006

Ketone Enolization Mediated by Mg(HMDS)2
A R T I C L E S
ometry of ketone and Mg(HMDS)2 present in complex 2. The
upfield shift in the proton signals was somewhat unexpected
considering that the ketone is acting as a Lewis base. However,
running the same NMR experiment in dichloromethane-d2 in
place of toluene-d8 leads to all the ketone signals shifting
downfield upon complexation, as expected (Figure S3). It
therefore appears that the magnetic anisotropy of the aromatic
solvent strongly influences the observed chemical shifts in 2.
To address the aggregation state of 2, its molecular size
was investigated using diffusion-ordered NMR spectroscopy
(DOSY).51 This technique has found increasing use in determin-
ing aggregation states of organometallic species; for example,
Williard has recently demonstrated the use of DOSY in differen-
tiating between dimeric and tetrameric forms of nBuLi in THF.52
In this case, we used dimeric {Mg(HMDS)2}2 as an internal
standard, allowing comparison of complex 2 to a known species
under identical conditions. This standardization is essential
because the outcomes from triplicate measurements on two
separate NMR samples (Table S4) show that, while the absolute
diffusion coefficients are not reproducible (presumably because
of their strong sensitivity to conditions such as temperature
gradients),51 the relatiVe diffusion constants of dimeric {Mg-
(HMDS)2}2 and 2 are reproducible, with the ratio of hydro-
dynamic volumes of {Mg(HMDS)2}2:2 of 1.25 ( 0.05. Thus,
qualitatively, 2 appears to be significantly smaller than {Mg-
(HMDS)2}2, in effect limiting its nuclearity to that of a
monomer. For more quantitative comparison, the relative
volumes of {Mg(HMDS)2}2 (I) and the monomeric complex
(HMDS)2Mg{η1-OdC(Ph)CH2CH3} (II) were estimated from
the optimized geometries of these compounds obtained by
density functional theory (DFT) calculations (see below). Using
this method, a theoretical VI/VII value of 1.43 was obtained,
which is in reasonable agreement with the experimentally
derived value of 1.25. Given the stoichiometry established by
NMR, the presence of coordinated ketone established by IR,
and mononuclearity indicated by DOSY, 2 is concluded to be
a monomer with a three-coordinate metal center, namely,
(HMDS)2Mg{η1-OdC(Ph)CH2CH3}.
Figure 4. Arrhenius plots of the decay of intermediate 2 formed from the
reaction of Mg(HMDS)2 with OdC(Ph)CH2CH3 (squares) and OdC(Ph)CD2-
CH3 (circles) in toluene.
Figure 5. 1H NMR spectra in toluene-d8 at -86 °C of (a) free propiophe-
none and (b) the 1:1 pre-enolization complex 2 formed in the presence of
excess of Mg(HMDS)2 (0.03 M).
Computational Studies. A computational study was used
to further investigate the identity of the reaction intermediate 2
and to probe the energetics of the system.53 Geometry optimiza-
tions and frequency analyses were successfully carried out on
the full molecules {Mg(HMDS)2}2 (I),54 (HMDS)2Mg{η1-
OdC(Ph)CH2CH3} (II), (HMDS)2Mg{η1-OdC(Ph)CH2CH3}2
(III), and OdC(Ph)CH2CH3 (IV) at the HF/6-31G* level of
theory.55 In addition, calculations were attempted on two types
of dimers containing a pair of metal-bound propiophenones,
one where the ketones bridge the metals, [(HMDS)2Mg{µ-Od
C(Ph)CH2CH3}]2,56 and a second with the ketones acting as
terminal Lewis base donors, [(HMDS){CH3CH2C(Ph)dO}Mg-
(µ-HMDS)]2. Neither dimeric complex optimized to a satisfac-
The observation of a carbonyl stretch confirms that enolization
has not yet taken place in 2, and the decrease in energy of
the stretch is consistent with complexation of the ketone to
magnesium.49,50
The composition of intermediate 2 was further investigated
by low-temperature NMR. NMR samples were prepared by the
addition of approximately 0.25 equiv of propiophenone to 0.03
M solutions of Mg(HMDS)2 in toluene-d8 at ambient temper-
ature, followed by immediate cooling of the tube in a dry ice/
acetone bath. The samples were then inserted into the spectrom-
eter, which was pre-cooled to -86 °C. The 1H NMR spectrum
(Figure 5b) indicates quantitative formation of a complex
between the ketone and Mg(HMDS)2. A single set of ketone
resonances is present, with all of the signals showing an upfield
shift compared with the free ketone. The methylene and methyl
signals of the ketone move significantly upon complexation,
shifting from δ 2.13 to 1.77 ppm, and from 1.10 to 0.90 ppm,
respectively. In addition, a new Me3Si singlet appears at δ 0.50
ppm. Integration of the relevant signals reveals a 1:1 stoichi-
(51) Cohen, Y.; Avram, L.; Frish, L. Angew. Chem., Int. Ed. 2005, 44, 520.
(52) (a) Keresztes, I.; Williard, P. G. J. Am. Chem. Soc. 2000, 122, 10228. (b)
Jacobson, M. A.; Keresztes, I.; Williard, P. G. J. Am. Chem. Soc. 2005,
127, 4965.
(53) Frisch, M. J.; et al. Gaussian 03; Gaussian, Inc.: Wallingford CT, 2004.
(54) The calculations for {Mg(HMDS)2}2 have previously been reported:
Wendell, L. T.; Bender, J.; He, X.; Noll, B. C.; Henderson, K. W.
Organometallics, published online Sep 7, 2006 http://dx.doi.org/10.1021/
om060521g.
(55) (a) Dill, J. D.; Pople, J. A. J. Chem. Phys. 1975, 62, 2921. (b) Hariharan,
P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28, 213. (c) Hehre, W. J.;
Ditchfield, R.; Pople, J. A. J. Chem. Phys. 1972, 56, 2257.
(56) Mayes, J. M.; Greer, J. C.; Mair, F. S. New J. Chem. 2001, 25, 262.
(49) Lochmann, L.; Trekoval, J. J. Orgonomet. Chem. 1975, 99, 329.
(50) Klumpp, G. W. Recl. TraV. Chim. Pays-Bas 1986, 105, 1.
9
J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006 13603

A R T I C L E S
He et al.
Scheme 1. DFT (B3LYP/6-311++G**) Computed Geometry Optimized Structures of I-III and the Energies (kcal/mol) of the
Transformations between the Complexes (selected bond lengths (Å) and angles (deg in italics) are also shown)
tory geometry minimum but rather resulted in dissociation of
the dimer or expulsion of ketone.
that subsequently deposited high quality crystals on cooling.
The IR spectra of both complexes indicated the presence of
ketone, with CdO stretches at 1661 and 1651 cm-1, respec-
tively. As expected, these signals are shifted to lower frequency
To obtain more accurate energies, the Hartree-Fock geom-
etries of I-IV were used as the starting positions for DFT
geometry optimizations run at the B3LYP/6-311++G** level
of theory.57 A summary of these studies is given in Scheme 1.
First, the geometries obtained for I using both levels of theory
compare very well with the crystal structure data for this
compound.54,58 For example, the individual Mg-N distances
and the N-Mg-N angles differ by less than 0.05 Å and 1.5°,
respectively, between theory and experiment. The absolute
energies of the complexes clearly indicate that coordination of
propiophenone and deaggregation of dimer I to two monomers
II is highly favorable (∆E ) -15.7 kcal/mol per Mg). Further
coordination of ketone to the monomer to give the four-
coordinate species III is also thermodynamically favorable, but
less so (∆E ) -4.5 kcal/mol). Preliminary experimental studies
do indicate that III can form in the presence of excess ketone,
but the calculated energetics clearly indicate that only II should
be present in the presence of excess Mg(HMDS)2.
Analyses of the bond lengths within the complexes give some
insights into their relative stabilities. Going from dimer I to
monomer II leads to shortening of the terminal and bridging
Mg-N bonds from 2.01 and 2.18 Å, respectively, to 1.98 Å,
along with formation of a relatively short Mg-O bond of 2.05
Å. However, further coordination of propiophenone to II, to
form III, results in slight lengthening of both the Mg-N bonds
to 2.05 Å and the Mg-O bonds to 2.10 and 2.15 Å. This is a
consequence of increasing the coordination number of the metal
from three to four and also increasing the steric encumbrance
within the complex. The Mg-O-C angles of 157.5° within II
and 144.2 and 153.8° within III show somewhat bent coordina-
tion of the ketone to the metals, although the angles are much
more obtuse than would be expected at sp2-hybridized oxygen.
The magnesium centers in II and III lie close to the plane of
the ketone, with Mg-O-C-C(H2) dihedral angles of 32.3°,
and 30.5 and 28.0°, respectively.
compared with those of the free ketones at 1680 and 1676 cm-1
,
respectively, similar to the decrease of 36 cm-1 found between
propiophenone-d2 and 2-d2.49 1H NMR spectroscopic studies
indicate a 1:1 ratio of metal to ketone present in each case,
and single-crystal X-ray diffraction analyses confirm their
compositions to be (HMDS)2Mg(η1-OdCtBu2) (3, Figure 6) and
(HMDS)2Mg{η1-OdC(tBu)Ph} (4, Figure 7).
The complexes adopt monomeric structures with three-
coordinate Mg centers. Complexes 3 and 4 are the first examples
of magnesium bis(amides) solvated by ketone to be characterized
in the solid state.59 The closest known magnesium analogues
to 3 and 4 are the alkoxymagnesium amide [(HMDS)Mg{µ-
OC(H)Ph2}(η1-OdCPh2)]2 (5)20 and the magnesium bis(enolate)
Mg4{OC(Mes)dCH2}8{η1-OdC(Mes)Me}2 (6, where Mes )
2,4,6-Me3C6H2),40 both containing terminal ketone bound to
four-coordinate magnesium centers. The dimeric ester-solvated
lithium amides [{η1-iPr(tBuO)CO}Li(µ-HMDS)]2 (7) and [{η1-
tBu(tBuO)CO}Li(µ-HMDS)]2 (8), which have three-coordinate
M(N)2O coordination spheres similar to those of 3 and 4, have
been reported by Williard.60 The metrical parameters of 3 and
4 are very similar to each other and are also in reasonable
agreement with the calculated geometry of II (Table 1). One
significant difference between the crystal structures and the
Structural Characterization of Model Complexes. In an
attempt to prepare model complexes of intermediate 2, 1 molar
t
t
equiv of the non-enolizable ketones Bu2CO and BuC(O)Ph
were added to hexane solutions of Mg(HMDS)2. The reactions
resulted in the immediate formation of bright yellow solutions
(57) (a) Becke, D. A. J. Chem. Phys. 1993, 98, 5648. (b) Miehlich, B.; Savin,
A.; Stoll, H.; Preuss, H. Chem. Phys. Lett. 1989, 157, 200. (c) Lee, C.;
Yang, W.; Parr, R. G. J. Chem. Phys. ReV. 1988, 157, 200.
Figure 6. Ellipsoid view of 3 (50% probabilities) with hydrogen atoms
removed for clarity.
(58) Westerhausen, M.; Schwarz, W. Z. Anorg. Allg. Chem. 1992, 609, 39.
9
13604 J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006

Ketone Enolization Mediated by Mg(HMDS)2
A R T I C L E S
anions [Mg(HMDS)3]- and [Ca(HMDS)3]-.23,62 The complexes
thus have helical chirality, though racemization is undoubtedly
facile (for example, the methylene protons of the propiophenone
in 2 are not diastereotopic, even at -86 °C).
Discussion
Magnesium bis(hexamethyldisilazide), Mg(HMDS)2, reacts
with propiophenone in noncoordinating solvents, such as toluene
or dichloromethane, over the course of a few minutes at room
temperature to form a mixture of (E)- and (Z)-magnesium
enolates. Enolization is therefore substantially slower than with
most lithium amides, which typically react within a few seconds
at -78 °C.10d If the Mg:ketone ratio exceeds 2:1, then the enolate
forms dimetallic species (E)-1 and (Z)-1, which are bridged by
one enolate and one amide (eq 2 and Figures 1 and 2). In the
presence of more ketone, further deprotonation can take place,
leading to amido(enolate) or bis(enolate) formation.23 Thus, use
of excess magnesium appeared likely to simplify the kinetics
by isolating the first deprotonation of the ketone by the bis-
(amide) from reactions involving mixed amido(enolates).
UV-visible spectroscopy reveals that addition of propiophe-
none to an excess of Mg(HMDS)2 at ambient temperatures
leads to an immediate jump in absorbance in the near-UV and
concomitant development of a yellow color. That color fades
over the course of several minutes, at which point the enolate
complex can be observed by NMR. The decay of the near-UV
absorbance follows first-order kinetics, and as expected, the
observed rate constant is independent of the initial concentration
of the deficient reagent, propiophenone. Unexpectedly though,
the observed rate constant is also independent of the concen-
tration of the excess reagent, Mg(HMDS)2 (Figure 3). The
appearance of optical bands characteristic of neither reactants
nor products indicates that a new species, 2, must be formed
by interaction of the ketone with the magnesium amide. The
insensitivity of the rate of decay to the concentration of magne-
sium indicates that the propiophenone is essentially completely
converted to 2. If 2 were in equilibrium with propiophenone,
with significant amounts of free ketone still present, then
the concentration of 2 would increase with increasing Mg-
(HMDS)2 concentration, leading to a positive reaction order in
[Mg(HMDS)2].
Many lines of evidence indicate that 2 is a monomeric, three-
coordinate ketone complex, (HMDS)2Mg{η1-OdC(Ph)CH2-
CH3}: (1) Low-temperature NMR spectra establish that the
ketone forms a single complex with a 2:1 ratio of HMDS to
ketone, and with the ethyl group still intact, not deprotonated.
(2) IR spectra indicate a moderate drop in the carbonyl stretch-
ing frequency (∆ν ) -36 cm-1 for the deuterated ketone),
consistent with η1 coordination. (3) Measurement of diffusion
constants of 2 using pulsed field-gradient NMR spectroscopy
indicates that the molecular size of 2 is most consistent with a
monomeric formulation, rather than that of a dimer or higher
oligomer. (4) Theoretical calculations are consistent with highly
exothermic dissociation of dimeric {Mg(HMDS)2}2 by pro-
piophenone to form three-coordinate monomers, while attempts
to find stable minima for ketone-coordinated dimeric species
failed. (5) The Mg(HMDS)2 complexes of the non-enolizable
Figure 7. Ellipsoid view of 4 (50% probabilities) with hydrogen atoms
removed for clarity.
Table 1. Key Bond Lengths (Å) and Angles (°) in 3 and 4 and
Comparisons with the Calculated Values for II and V
3
4
II
V
Mg(1)-O(1)
1.970(2)
1.960(2)
1.965(2)
1.224(3)
136.88(10)
112.31(10)
110.80(9)
177.1(2)
123.93(13)
124.04(13)
1.9630(7)
1.9627(8)
1.9665(7)
1.2321(11)
140.99(3)
106.14(3)
111.83(3)
175.16(7)
121.65(4)
130.89(5)
2.054
1.987
1.989
1.237
142.42
109.42
108.15
157.48
124.29
124.67
2.044
1.987
1.998
1.235
141.46
109.02
109.52
172.40
124.94
122.99
Mg(1)-N(1)
Mg(1)-N(2)
O(1)-C(1)
N(1)-Mg(1)-N(2)
N(1)-Mg(1)-O(1)
N(2)-Mg(1)-O(1)
Mg(1)-O(1)-C(1)
Si(1)-N(1)-Si(2)
Si(3)-N(2)-Si(4)
calculations are the Mg(1)-O(1)-C(1) angles, which are almost
linear at 175.16(7) and 177.2(2)° in 3 and 4, whereas a more
acute angle of 157.5° is found in II. To see whether this reflects
an inadequacy of the calculation or a real difference between
the less hindered propiophenone and the tert-butyl-substituted
ketones, a direct comparative calculation was completed for
(HMDS)2Mg{η1-OC(tBu)Ph} (V) at the B3LYP/6-311++G**
level of theory. This gave a minimum with metrical parameters
very similar to those obtained experimentally for 4, including
a Mg-O-C(1) angle of 172.40° (Table 1). Note that a wide
range of M-O-C angles is also observed in the ester-solvated
lithium complexes 7 and 8 (144.63-165.15°).60 These observa-
tions indicate that ketone coordination to electropositive metals,
such as lithium or magnesium, is highly flexible, with obtuse
M-O-C angles common, and with more sterically hindered
ketones giving more linear coordination.61
The SiN2 planes of the amides and the R2CO plane of the
ketones form a three-bladed propeller around the central atom,
with these planes forming angles of 44.1, 69.0, and 50.5° in 3,
and 50.2, 50.2, and 45.3° in 4, with respect to the Mg(N)2O
coordination plane. This arrangement is typical of metal tris-
(hexamethyldisilazide) complexes, such as that found in the
(59) For other examples of monosolvated three-coordinate magnesium bis-
(amides), see: (a) Tang, Y.; Zakharov, L. N.; Rheingold, A. L.; Kemp, R.
A. Organometallics 2005, 24, 836. (b) Kennedy, A. R.; Mulvey, R. E.;
Schulte, J. H. Acta Crystallogr., Sect. C 2001, 57, 1288. (c) Sebestl, J. L.;
Nadasdi, T. T.; Heeg, M. J.; Winter, C. H. Inorg. Chem. 1998, 37, 1289.
(60) Williard, P. G.; Liu, Q. Y.; Lochmann, L. J. Am. Chem. Soc. 1992, 114,
348.
(61) A similar observation has been made for the coordination of HMPA to
magnesium bis(amides): Clegg, W.; Craig, F. J.; Henderson, K. W.;
Kennedy, A. R.; Mulvey, R. E.; O’Neil, P. A.; Reed, D. Inorg. Chem.
1997, 36, 6238.
(62) (a) Forbes, G. C.; Kennedy, A. R.; Mulvey, R. E.; Rodger, P. J. A. Chem.
Commun. 2001, 1400. (b) Honeyman, G. W.; Kennedy, A. R.; Mulvey, R.
E.; Sherrington, D. C. Organometallics 2004, 23, 1197.
9
J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006 13605

A R T I C L E S
He et al.
Scheme 2. Possible Enolization Mechanisms
t
ketones Bu2CdO and PhC(O)tBu are shown to be three-
metallic intermediate. The net process (Scheme 3) would thus
involve initial deaggregation of dimeric {Mg(HMDS)2}2 to
form the monomeric, monoketone complex 2. Complex 2 then
undergoes intramolecular proton transfer via a six-membered
ring transition state to produce, initially, a three-coordinate
amidomagnesium enolate containing a coordinated hexamethyl-
disilazane. Dissociation of HN(SiMe3)2 and aggregation of the
amido(enolate) with Mg(HMDS)2 leads to the observed products
(E)-1 and (Z)-1.
The intramolecular proton transfer reaction in the rate-
determining step is reminiscent of the classic Ireland model of
lithium amide-mediated enolization reactions.63 While Ireland
proposed a chair-like structure for the transition state, compu-
tational studies support a much less puckered structure.6f,13a The
observation of a very large primary isotope effect strongly
supports such a modified Ireland mechanism. Achieving such
a large isotope effect is generally accepted to require a nearly
linear C-H-N trajectory, as well as a central transition state
(since early or late transition states lead to reduced isotope
effects). In fact, large isotope effects, and significant tunneling
such as is observed here, are frequently observed for deproto-
nations of carbon acids with nitrogen bases, with enhanced
isotope effects often observed on increasing the steric bulk of
the base.64 The steric enhancement of tunneling is generally
ascribed to the bulkier bases requiring a long distance between
carbon and nitrogen in the transition state.65
coordinate monomers by X-ray crystallography.
Since (HMDS)2Mg{η1-OdC(Ph)CH2CH3} (2) is the domi-
nant form of propiophenone in the presence of excess Mg-
(HMDS)2, the decay of the optical transient corresponds to the
decay of the complex. The very large rate retardation observed
for the PhCOCD2CH3 complex (kH/kD ) 18.9 at 21 °C) clearly
indicates that the rate-determining step for the decay involves
proton transfer. Four conceivable mechanisms for the enolization
step are shown in Scheme 2. Deprotonation of complexed ketone
by exogenous magnesium amide (Mechanism A), whether mon-
omeric or dimeric, is ruled out by the observed insensitivity of
the rate to the concentration of free Mg(HMDS)2. Pre-equilib-
rium dissociation of ketone from the complex, followed by
deprotonation of the free ketone by monomeric Mg(HMDS)2
(Mechanism B), in contrast, is consistent with the observed rate
law since the first-order dependence of the deprotonation step
on [Mg(HMDS)2] would be canceled by the inverse dependence
of the concentration of free ketone on [Mg(HMDS)2]. (Depro-
tonation by dimeric {Mg(HMDS)2}2 would not exhibit such a
cancellation and so is ruled out by the observed rate law.)
However, if free ketone is deprotonated by free magnesium
amide, the transfer of H+ would result in net charge separation
to form an enolate ion pair. The substantial solvent dependence
expected for such a reaction is not observed, however, with
reactions in toluene (ꢀ ) 2.4) and dichloromethane (ꢀ ) 8.9)
proceeding at identical rates. Analogous arguments mitigate
against dissociation of an amido group and deprotonation of
bound ketone within the ion pair (Mechanism C). In contrast,
intramolecular proton transfer from coordinated ketone to coor-
dinated amide (Mechanism D) would be expected to show little
change in polarity in the transition state. Such an intramolecular
reaction is also consistent with the measured moderate negative
entropy of activation (∆Sq ) -11 cal/mol‚K) of the enolization.
The experimental data thus strongly suggest that enolization
takes place intramolecularly within a three-coordinate mono-
The proposed transition structure bears on two important
issues of selectivity. First, a cyclic transition state offers an
(63) Ireland, R. E.; Mueller, R. H.; Willard, A. K. J. Am. Chem. Soc. 1976, 98,
2868.
(64) (a) Lewis, E. S.; Funderburk, L. H. J. Am. Chem. Soc. 1967, 89, 2322. (b)
Caldin, E. F.; Mateo, S. J. Chem. Soc., Faraday Trans. 1 1975, 71, 1876.
(c) Caldin, E. F.; Mateo, S. J. Chem. Soc., Faraday Trans. 1 1976, 72,
112. (d) Pruszynski, P.; Jarczewski, A. J. Chem. Soc., Perkin Trans. 2 1986,
1117. (e) Pruszynski, P. Can. J. Chem. 1987, 65, 2160-2163. (f) Leffek,
K. T.; Pruszynski, P. Can. J. Chem. 1988, 66, 1454.
(65) (a) Wolfe, S.; Hoz, S.; Kim, C.-K.; Yang, K. J. Am. Chem. Soc. 1990,
112, 4186. (b) Kim, Y.; Kreevoy, M. M. J. Am. Chem. Soc. 1992, 114,
7116.
9
13606 J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006

Ketone Enolization Mediated by Mg(HMDS)2
A R T I C L E S
Scheme 3. Proposed Mechanism of the Enolization of Propiophenone by Mg(HMDS)2
Scheme 4. Cyclic Transition State Structures Leading to (E)- and
(Z)-Enolates
Scheme 5. Possible Conformations of 4-Substituted
Cyclohexanones Coordinated to a Magnesium Bis(amide)
Conclusions
appealing explanation for the relatively high (E)-selectivity of
enolization of propiophenone by Mg(HMDS)2. Typically, eno-
lizations of propiophenone give predominantly (Z)-enolates
because of developing A1,3 strain between the methyl group and
the phenyl group during formation of the (E)-enolate.14,44,63
However, in a cyclic transition state, the methyl group would
interact with the trimethylsilyl groups on the amide in the
transition state leading to the (Z)-enolate, disfavoring it relative
to the (E)-transition state, where the methyl group is nearly anti
to the amide (Scheme 4).24b
Second, this transition state model also offers an intriguing
mechanism for stereoselection in the enantioselective enolization
of substituted cyclohexanones by chiral bis(amido)magnesium
reagents.29 The Mg{N(SiMe3)2}2 fragment is intrinsically chiral
in the reactive three-coordinate complex since the planes of the
amides are canted around the coordination plane like two blades
of a three-bladed propeller. Furthermore, the tilt of the amide
orients the nitrogen lone pair correctly for proton transfer, as
judged from the calculated structures for intramolecular lithium-
mediated enolizations.6f,13a The helical sense of chirality of the
Mg(amide)2 fragment determines which enantiotopic axial
proton can be deprotonated; the mismatched proton cannot align
with the lone pair on nitrogen (Scheme 5). Thus, any bias
provided by the chiral amides that favors one helical conforma-
tion over the other could play a determining role in the observed
selectivity. This explanation differs significantly over the usual
rationalizations in that it emphasizes the interactions within the
reagent rather than direct interactions between the amide groups
of the reagent and the substrate.66
Data from reaction kinetics, spectroscopic characterization
of intermediates, and computational and crystallographic struc-
ture analyses paint a detailed picture of the mechanism of
enolization of propiophenone by the prototypical magnesium
bis(amide), magnesium bis(hexamethyldisilazide). The reaction
proceeds via initial quantitative formation of the three-coordinate
ketone complex (HMDS)2Mg{η1-OdC(Ph)CH2CH3} (2). This
complex undergoes rate-limiting intramolecular proton transfer
from the ketone to a coordinated amide in a reaction in which
tunneling plays a significant role (kH/kD ) 18.9 at 21.8 °C). An
intrinsically chiral, envelope-shaped transition state is proposed,
which has important implications for the stereo- and enantio-
selectivity of magnesium amide-mediated enolization reactions.
Experimental Section
General. All operations were carried out using Schlenk techniques
or inside an argon-filled glovebox.67 All glassware was flame-dried
under vacuum before use. Toluene and hexane were dried by passage
through copper-based catalyst and molecular sieve columns (Innovative
Technology). The ketones were purchased from commercial sources
and were distilled over CaH2 under N2 prior to use.68 Me3SiCl was
distilled under a N2 atmosphere before use. Mg(HMDS)2 was prepared
as described previously.20 Butyllithium was purchased from Aldrich
as a 1.6 M solution in hexanes and was standardized by titration against
salicylaldehyde phenylhydrazone.69 Deuterated solvents were purchased
from Cambridge Isotope Laboratories and were dried by storage over
4 Å molecular sieves. NMR data were recorded on a Bruker Avance
(67) Shriver, D. F.; Drezdzon, M. A. Manipulation of Air SensitiVe Compounds;
John Wiley and Sons: New York, 1986.
(68) Amarego, W. L. F.; Perrin, D. D. The Purification of Laboratory Chemicals,
4th ed.; Butterworth Heinemann: Bath, UK, 2002.
(66) Majewski, M.; Wang, F. Tetrahedron 2002, 58, 4567.
(69) Love, B. E.; Jones, E. G. J. Org. Chem. 1999, 64, 3755.
9
J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006 13607

A R T I C L E S
He et al.
DPX-400 instrument at 298 K unless stated otherwise. 1H and 13C
spectra were referenced to the residual solvent signals. GC experiments
were performed on a Shimadzu GC-17A gas chromatograph fitted with
a Rtx-5 fused Crossbond 5% diphenyl/95% dimethyl polysiloxane
column (30 m, 0.25 mm i.d., 0.25 umdf), using N2 as carrier gas.
Detection was by flame ionization, and the chromatograms were
interpreted using GCsolution software. Elemental analyses were
performed by Midwest Microlab, Indianapolis, Indiana.
DPX-400 instrument equipped with a z-axis gradient amplifier. This
1
instrument was operated at 400.13 MHz for H observations using a
1
Nalorac 5 mm inverse detection H/19F probe with a z-axis gradient
coil. The LEDBPGS2S pulse sequence was performed with a spectral
width of 4006 Hz, and 8 transients were collected for each gradient
value. Sinusoidal gradients were used with a total duration of 2.4 ms.
The pulsed field-gradients were incremented in 16 steps from 5 up to
90%. Diffusion time was 0.4 s. The eddy current delay was 5 ms. Data
processing was accomplished using XWINNMR 2.6.
Measured diffusion coefficients (D) of dimeric {Mg(HMDS)2}2 and
2 were converted to idealized spherical volumes for the Mg(HMDS)2
dimer, Vstd, and complex 2, V2, using the respective hydrodynamic radii
(rs) obtained from the Stokes-Einstein equation.51,70 The viscosity of
toluene at -85.4 °C was estimated using a formula from the literature,71
and the viscosity of the solution was assumed to be the same as that of
the neat solvent. The temperature of the NMR probe was calibrated
using 100% methanol.72
Spectroscopic Characterization of (E)- and (Z)-[(HMDS)2Mg2-
(µ-HMDS){µ-OC(Ph)dCHCH3}], (E)-1 and (Z)-1. A 1H NMR sample
was prepared at ambient temperature by adding 1 µL of a 2.0 M toluene-
d8 solution of propiophenone (2.03 × 10-6 mol) into an NMR tube
charged with 7 mg (2.03 × 10-5 mol) of Mg(HMDS)2 and 0.7 mL of
toluene-d8. The spectrum was taken within a few minutes of mixing
the reactants together. Equivalent results are obtained using a stoichio-
metric ratio of ketone to magnesium (1:2). Previously reported nOe
studies were used to assign the stereochemistry of the enolate product;23
1
the (E)- and (Z)-enolates are formed in a 74:26 ratio. (E)-1: H NMR
Quenching Studies. A Schlenk tube under N2 was filled with 0.17
g (0.5 mmol) of Mg(HMDS)2 and 2.5 mL of toluene. Propiophenone
(6.7 µL, 0.05 mmol) was added to the solution and the mixture stirred
for 15 min at ambient temperature. Me3SiCl (0.25 mL, 2 mmol) was
added, and the solution was cooled to -78 °C. BuLi (1 mL of a 1.6 M
solution in hexanes, 1.6 mmol) was added over 1 min followed by 1 mL
of THF. The mixture was allowed to warm to room temperature, and
a 0.5 mL aliquot of the resulting solution was quenched with 5 mL of
saturated aqueous NaHCO3. The quenched reaction mixture was ex-
tracted with 5 mL of ether, diluted by THF, and analyzed by GC. The
conversion of the ketone to the silyl enol ethers was determined to be
>99% with an E/Z ratio of 72:28. The stereochemical assignments for
the silyl enol ethers were made by comparison with authentic samples.44a
Synthesis of PhC(O)CD2CH3. A 100 mL round-bottom flask was
flushed with argon and charged with 7.90 g (0.059 mol) of propiophe-
none, a stirring bar, and 30 mL of CH3OD. Sodium methoxide (200
mg, 3.7 mmol) was added via a solids addition tube, and the flask was
sealed with a rubber septum. The mixture was stirred for 6 days at
ambient temperature, subsequently quenched with cold water (75 mL),
and the organics were extracted with Et2O (3 × 40 mL). The organic
layer was dried with MgSO4, filtered, and the solvent removed using
a rotary evaporator. The ketone was purified by distillation from CaH2
under reduced pressure at 40-42 °C. This cycle was repeated a second
time to improve the isotopic enrichment. Yield: 5.81 g, 73.5%. The
(toluene-d8) δ 0.23 (s, 36 H, terminal (CH3)3Si), 0.33 (s, 18H, bridging
(CH3)3Si), 1.62 (d, 3H, J ) 7.1 Hz, CHCH3), 5.33 (q, 1H, J ) 7.0 Hz,
CHCH3), 7.08 (t, 1H, J ) 7.6 Hz, p-H), 7.21 (t, 2H, J )7.6 Hz, m-H),
7.43 (d, 2H, J )7.6 Hz, o-H); 13C NMR spectra were run in methylene
chloride-d2 due to multiple overlapping signals in toluene-d8; 13C NMR
(methylene chloride-d2) δ 5.95 (terminal (CH3)3Si), 6.46 (bridging
(CH3)3Si), 13.35 (CHCH3), 103.84 (CHCH3), 125.84 (o-C), 128.81 (p-
C), 129.34 (m-C), 139.30 (i-C), 151.39 (OCdCHCH3). (Z)-1: 1H NMR
(toluene-d8) δ 0.18 (s, 36 H, terminal (CH3)3Si), 0.40 (s, 18H, bridging
(CH3)3Si), 1.84 (d, 3H, J ) 6.9 Hz, CHCH3), 5.09 (q, 1H, J ) 6.8 Hz,
CHCH3), 7.04 (t, 1H, J ) 7.6 Hz, p-H), 7.17 (t, 2H, J ) 7.6 Hz, m-H),
7.44 (d, 2H, J ) 7.6 Hz, o-H); 13C NMR (methylene chloride-d2) δ
5.82 (terminal (CH3)3Si), 6.60 (bridging (CH3)3Si), 13.57 (CHCH3),
101.59 (CHCH3), 128.51 (o-C), 128.68 (p-C), 129.39 (m-C), 140.76
(i-C), 152.40 (OCdCHCH3).
Crystallization of (E)-1 and (Z)-1. A Schlenk tube was charged
with 0.73 g (2.1 mmol) of Mg(HMDS)2 and 3 mL of toluene. Dropwise
addition of 0.13 mL (1 mmol) of propiophenone at ambient temperature
resulted in the formation of a bright yellow solution. The solution was
allowed to stir for 30 min, during which time the color faded to become
colorless. The solvent was removed in vacuo, replaced by 1 mL of
hexane, and the mixture filtered through a glass frit. The filtrate was
stored at -20 °C, and a crop of colorless crystals was deposited over
3 days. Yield: 0.07 g, 11%. Identical NMR spectra were obtained from
the crystals compared with the in situ preparation described above with
the exception that the E/Z ratio was determined to be 20:80. Anal. Calcd
for C27H63Mg2N3OSi6: C, 48.99; H, 9.60; N, 6.35%. Found: C, 48.98;
H, 9.23; N, 6.08%.
1
isotopic purity of PhC(O)CD2CH3 was determined by the H NMR
spectrum of the product in CDCl3 at 295 K. This indicated a
composition of 94% PhC(O)CD2CH3 with the remaining 6% being
principally PhC(O)CH(D)CH3, with only a trace of PhC(O)CH2CH3.
IR Spectroscopic Analyses. Spectra were recorded on a Perkin-
Elmer Paragon 1000 FTIR spectrometer. Solid samples were prepared
as Nujol mulls using KBr plates. Solution samples were prepared in a
semi-demountable cell using the deuterated ketone since the half-life
of its reaction with Mg(HMDS)2 is relatively long (approximately 43
min). Samples were prepared by mixing 8 µL of PhC(O)CD2CH3 with
0.5 mL of a 0.145 mol/L solution of Mg(HMDS)2 in a sealed vial under
argon. The reaction mixture was quickly introduced into the IR cell
through a septum and its spectrum recorded. The sample was re-run
several times over a 1 h period, and the carbonyl stretch at 1657 cm-1
was seen to diminish as expected due to conversion to the enolate.
UV-Vis Spectroscopic Analyses. Spectra were recorded on a
Beckman DU 7500 spectrophotometer in a cell block whose temperature
was controlled by circulating a water/ethylene glycol mixture through
it. The temperature was measured using a digital thermometer which
Low-Temperature NMR Spectroscopic Analyses of (HMDS)2Mg-
{η1-OdC(Ph)CH2CH3} (2). Samples for the low-temperature studies
were prepared as follows: an NMR tube with a screw cap fitted with
a Teflon-lined rubber septum was charged with 4.2 mg of Mg(HMDS)2
and 0.6 mL of toluene-d8 inside a glovebox. The tube was removed
from the glovebox and 20 µL of a 0.25 M solution of propiophenone
in toluene-d8 was added via a syringe at ambient temperature. The
reactants were mixed thoroughly, and the tube was immediately cooled
in a dry ice/acetone bath. The sample was then placed in a -86 °C
pre-cooled NMR instrument and its spectrum recorded. 1H NMR
(toluene-d8, 187 K): δ δ 0.50 (s, 36H, SiMe3), 0.90 (t, 3H, J ) 6.6
Hz, CH3), 1.77 (q, 2H, J ) 6.7 Hz, CH2), 6.72 (t, 2H, J ) 7.8 Hz,
m-H), 6.85 (t, 1H, J ) 7.8 Hz, p-H), 7.54 (br, 2H, o-H). A similar
procedure was repeated using dichloromethane-d2 in place of toluene-
1
d8. H NMR (CD2Cl2, 187 K): δ -0.11 (s, 36H, SiMe3), 1.28 (t, 3H,
(70) (a) Crank, J. The Mathematics of Diffusion, 2nd ed.; Clarendon: Oxford,
1975. (b) Cussler, E. L. Diffusion: Mass Transfer in Fluid Systems;
Cambridge University: Cambridge, 1984.
(71) Barlow, A. J.; Lamb, J.; Matheson, A. J. Proc. R. Soc. London, Ser. A
1966, A292, 322.
(72) Braun, S.; Kalinowski, H. O.; Berger, S. 100 and More Basic NMR
Experiments, A Practical Course; VCH: New York 1996.
J ) 6.8 Hz, CH3), 3.41 (q, 2H, J ) 6.7 Hz, CH2), 7.59 (t, 2H, J ) 7.5
Hz, m-H), 7.79 (t, 1H, J ) 7.5 Hz, p-H), 8.19 (d, 2H, J ) 8.2 Hz,
o-H).
Diffusion NMR Analyses of 2. Samples of 2 were prepared as
described above. The NMR spectra were collected on a Bruker Avance
9
13608 J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006

Ketone Enolization Mediated by Mg(HMDS)2
A R T I C L E S
Table 2. Crystallographic Data for Compounds 1, 3, and 4
1
3
4
empirical formula
formula weight
temperature, K
wavelength, Å
crystal system
space group
a, Å
C27H63Mg2N3OSi6
662.96
100(2)
0.71073
triclinic
P1h
9.0548(5)
14.5893(7)
15.6973(8)
84.429(2)
83.604(2)
79.376(2)
2019.20(18)
2
C21H54MgN2OSi4
487.33
100(2)
0.71073
monoclinic
C2/c
21.135(3)
8.3722(10)
37.287(5)
90
93.029(2)
90
6588.4(14)
8
0.979
C23H50MgN2OSi4
507.32
100(2)
0.71073
monoclinic
P21/n
a ) 8.3655(2)
b ) 20.6214(6)
c ) 18.7658(5)
90
100.694(1)
90
3181.03(15)
4
1.059
b, Å
c, Å
R, deg
â, deg
γ, deg
volume, Å3
Z
density (calcd), g/cm3
absorption coefficient, mm-1
F(000)
1.090
0.261
724
0.213
2142
0.223
1112
crystal size, mm
φ range, °
index ranges
0.48 × 0.37 × 0.32
1.31 to 27.88
-11 e h e 11,
-19 e k e 19,
-20 e l e 20
131789
0.49 × 0.45 × 0.28
1.93 to 28.88
-28 e he 28,
-10 e k e 11,
-46 e l e 50
27124
0.46 × 0.43 × 0.16
1.48 to 31.56
-12 e h e 12,
-30 e k e 30,
-26 e l e 27
116269
reflections collected
independent reflections
max and min transmission
data/restraints/parameters
Goodness-of-fit on F2
R1 [I > 2σ(I)]
9578 [R(int) ) 0.0430]
0.9212 and 0.8856
9578/2/385
1.108
0.0377
0.1106
8604 [R(int) ) 0.0254]
0.94 and 0.91
8604/1/327
1.325
0.0851
0.2021
10635 [R(int) ) 0.0367]
0.9652 and 0.9044
10635/0/298
1.028
0.0286
0.0791
wR2 (all data)
largest diff. peak and hole, e‚Å-3
0.612 and -0.444
0.427 and -0.419
0.459 and -0.176
was placed in a solvent-filled cuvette within the sample holder. Samples
were contained in quartz cuvettes with screw-caps fitted with PTFE-
faced silicone rubber septa.
crystals formed over a period of 2 days. Yield: 0.20 g, 39%. 1H NMR
(toluene-d8): δ 0.29 (s, 36H, Si(CH3)3), 1.11 (s, 9H, C(CH3)3), 7.03-
7.05 (m, 3H, m- and p-H), 7.87 (d, 2H, J ) 7.2 Hz, o-H). 13C NMR
(toluene-d8): δ 5.91 (Si(CH3)3), 28.10 (C(CH3)3), 45.47 (C(CH3)3),
129.09 (m-C), 130.27 (o-C), 135.07 (p-C), 135.46 (i-C), 219.86 (Cd
O). IR (cm-1): ν 1651 (CdO). Anal. Calcd for C23H50MgN2OSi4: C,
54.51; H, 9.95; N, 5.53%. Found: C, 54.80; H, 9.59; N, 5.26%.
Computational Details. The Gaussian 03 series of programs was
used for the geometry optimization calculations for I-V.53 No
symmetry constraints were imposed, and the molecules were allowed
to freely optimize at the HF/6-31G* level using related crystal structure
data as starting geometries.55 The geometries were verified as true
minima using frequency analyses. The ab initio derived geometries of
I-V were re-optimized at the B3LYP/6-311++G** level of theory to
obtain more accurate absolute energy data.57 Molecular volumes for
I-III were calculated using the Spartan 04 program.73
X-ray Crystallography. Single crystals of 1, 3, and 4 were grown
according to the experimental procedures outlined above and were
examined under Infineum V8512 oil. In each case, the datum crystal
was affixed to either a thin glass fiber atop a tapered copper mounting
pin or a Mitegen mounting loop and transferred to the 100 K nitrogen
stream of a Bruker APEX diffractometer equipped with an Oxford
Cryosystems 700 series low-temperature apparatus. Cell parameters
were determined using reflections harvested from three sets of 12 0.5°
φ scans. The orientation matrix derived from this was transferred to
COSMO74 to determine the optimum data collection strategy requiring
a minimum of 4-fold redundancy. Final cell parameters were refined
using reflections harvested from the data collection with I g 10σ(I).
All data were corrected for Lorentz and polarization effects, and runs
were scaled using SADABS.75
A typical sample was prepared as follows: inside a glovebox, a 50
mL volumetric flask was charged with 0.8270 g of Mg(HMDS)2 and
filled with dry toluene. This was then used as a standard solution for
the reactions. Each run used 3.0 mL samples of this solution, which
were placed into cuvettes, capped, and the caps covered with Parafilm.
The cuvettes were taken out of the glovebox and loaded into the UV-
vis sample holder. The temperature of the circulating bath was then
adjusted and the sample allowed to thermally equilibrate for ap-
proximately 20 min. After this time, 30 µL of a 0.292 M solution of
propiophenone was added to the cuvette, and data were collected
immediately. Spectra were acquired at 330 nm for at least 5 half-lives.
For the reactions of PhC(O)CH2CH3, absorbance versus time data
were fit to the formula A ) Af + (A0 - Af)exp(-kobst) using the program
Origin 6.1 (OriginLab Corporation, Northampton MA). For reactions
of PhC(O)CD2CH3, data were fit to A ) 0.06(A0 - Af)exp(-kHobst) +
0.94(A0 - Af)exp(-kDobst) + Af, where both kHobs and kDobs were allowed
to vary. Values of kH agreed, within experimental error, with the
obs
0.5 × kobs measured under similar conditions for the protio compound
(as expected from a statistical factor of 2).
Synthesis of (HMDS)2Mg(η1-OCtBu2) (3). A Schlenk tube under
argon was charged with 0.69 mL (4 mmol) of 2,2,4,4-tetramethyl-3-
pentanone, and a solution of Mg(HMDS)2 (4 mmol, 1.38 g) in hexane
(10 mL) was added by syringe to give a light yellow solution. The
solution was cooled to -20 °C, and a crop of colorless crystals
1
precipitated overnight. Yield: 0.37 g, 19%. H NMR (toluene-d8): δ
0.30 (s, 36H, Si(CH3)3), 0.90 (s, 18H, C(CH3)3). 13C NMR (toluene-
d8): δ 6.10 (Si(CH3)3), 27.84 (C(CH3)3), 47.44 (C(CH3)3), 236.33 (Od
C). IR (cm-1): ν 1661 (CdO). Anal. Calcd for C21H54MgN2OSi4: C,
51.76; H, 11.17; N, 5.75%. Found: C, 50.75; H, 10.83; N, 5.90%.
Synthesis of (HMDS)2Mg{η1-OC(tBu)Ph} (4). A Schlenk tube was
charged with 0.35 g (1 mmol) of crystalline Mg(HMDS)2 and 1 mL of
hexane. 2,2′-Dimethylpropiophenone (0.17 mL, 1 mmol) was added
dropwise to produce a light yellow solution. After stirring for 15 min
at ambient temperature, the mixture was stored at 0 °C, and colorless
The structures were solved and refined using SHELXTL.76 Structure
solution was by direct methods. Non-hydrogen atoms not present in
(73) Spartan 04; Wavefunction Inc.: Irvine, CA, 2004.
(74) Bruker-Nonius AXS (2005), APEX2, and COSMO; Bruker-Nonius AXS:
Madison, Wisconsin.
(75) Sheldrick, G. M. 2004. SADABS; Bruker-Nonius AXS: Madison, Wisconsin.
(76) Sheldrick, G. M. SHELXTL; University of Go¨ttingen: Go¨ttingen, Germany.
9
J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006 13609

A R T I C L E S
He et al.
the direct methods’ solution were located by successive cycles of full-
matrix least squares refinement on F2. All non-hydrogen atoms were
refined with parameters for anisotropic thermal motion. Hydrogen atoms
were placed at idealized geometries and allowed to ride on the position
of the parent atom. Hydrogen thermal parameters were set to 1.2 times
the equivalent isotropic U of the parent atom, 1.5 times for methyl
hydrogens. Table 2 lists the key crystallographic parameters.
The crystal structure of 1 exhibits disorder in the enolate functional
group. Carbon C(21) presented two sites, the (E)- and (Z)-isomeric
forms. Alternate positions were located and refined for C(20) and C(21)
(C(20′) and C(21′)). The site occupancies were refined so that the total
of each part equaled 1. The site occupancy factor of the major
orientation is 0.779(6). Because of the disorder, thermal factors for
C20 and C20′ were set to be equal using EADP from XL.76 Also, the
C19-C20 and C19-C20′ distances were restrained to be similar to
the SADI command. Distances C20-C21 and C20′-C21′ were
similarly restrained. In compound 3, the carbon atoms of the tert-butyl
groups were disordered. Atoms C(7), C(8), and C(9) were modeled as
having two-site disorder. The site occupancy of the major component
refined to 0.643(12). C(3), C(4), and C(5) were also modeled as having
two-site disorder. The site occupancy of the major component refined
to 0.525(13). Parameters for thermal motion of the methyl carbons of
the ketone were restrained to be equivalent using the EADP function
of XL.76 Compound 4 was well-behaved on refinement.
Acknowledgment. K.W.H. and X.H. gratefully acknowledge
the National Science Foundation for support (CHE03-45713).
We also thank the National Science Foundation for instrumenta-
tion support (CHE04-43233).
Supporting Information Available: Crystallographic data for
1, 3, and 4 in CIF format; the complete list of authors for refs
6c and 53; tables summarizing (i) bond lengths and angles for
1, (ii) the kinetic data, and (iii) the diffusion NMR experiments;
Eyring plot of the reaction of Mg(HMDS)2 with propiophenone;
plot of absorbance versus time for the disappearance of interme-
1
diate 2; H NMR spectra of free propiophenone and interme-
diate 2 in dichloromethane-d2; and details of the calculations.
This material is available free of charge via the Internet at
http://pubs.acs.org.
JA064927W
9
13610 J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006