Lithium Hexamethyldisilazide-Mediated Ketone Enolization: The Influence of Hindered Dialkyl Ethers and Isostructural Dialkylamines on Reaction Rates and Mechanisms

Article abstract of DOI:10.1021/jo030221y

Mechanistic studies of the enolization of 2-methylcyclohexanone mediated by lithium hexamethyldisilazide (LiHMDS; TMS₂NLi) solvated by hindered dialkyl ethers (ROR′) are described. Rate studies using in situ IR spectroscopy show that enolizatio

Full text of DOI:10.1021/jo030221y

Lith iu m Hexa m eth yld isila zid e-Med ia ted Keton e En oliza tion : Th e
In flu en ce of Hin d er ed Dia lk yl Eth er s a n d Isostr u ctu r a l
Dia lk yla m in es on Rea ction Ra tes a n d Mech a n ism s
Pinjing Zhao, Brett L. Lucht, Sarita L. Kenkre, and David B. Collum*
Department of Chemistry and Chemical Biology, Baker Laboratory, Cornell University,
Ithaca, New York 14853-1301
dbc6@cornell.edu
Received J uly 10, 2003
Mechanistic studies of the enolization of 2-methylcyclohexanone mediated by lithium hexameth-
yldisilazide (LiHMDS; TMS
studies using in situ IR spectroscopy show that enolizations in the presence of i-Pr
tetramethyltetrahydrofuran, and cineole proceed via dimer-based transition structures [(TMS
2
NLi) solvated by hindered dialkyl ethers (ROR′) are described. Rate
2
O, 2,2,5,5-
2
-
‡
2
NLi) (ROR′)(ketone)] . Comparing the relative solvation energies and the corresponding solvent-
dependent activation energies shows that the highly substituted ethers accelerate the enolizations
by sterically destabilizing the reactants and stabilizing the transition structures. Comparisons of
hindered dialkyl ethers with their isostructural dialkylamines reveal that the considerably higher
rates elicited by the amines derive from an analogous relative destabilization of the reactants and
relative stabilization of the transition structures.
In tr od u ction
We began studying how solvation influences the reac-
tivity of LiHMDS by focusing on one of the most impor-
tant reactions in organolithium chemistry, ketone eno-
High thermal stability, reactivity, and selectivity com-
bine to make lithium hexamethyldisilazide (LiHMDS)
one of the most important Br o¨ nsted bases in organic
synthesis.1 These properties also make LiHMDS an
excellent template for investigating structure-reactivity
relationships within organolithium chemistry. Structural
studies of LiHMDS solvated by nearly 100 mono- and
1
lization. Early efforts revealed that poorly coordinating
4a,5
trialkylamines
can cause up to 3000-fold accelerations
,2
6
compared to enolizations in neat toluene (eq 1). The rates
3
bidentate ethers and amines provided key insights into
lithium ion coordination chemistry and established a firm
basis to understand solvent-dependent reactivities and
mechanisms.4
(
1) For selected examples in which LiHMDS is used on large scale,
see: (a) Parsons, R. L., J r. Curr. Opin. Drug Discovery Dev. 2000, 3,
83. (b) Kauffman, G. S.; Harris, G. D.; Dorow, R. L.; Stone, B. R. P.;
7
Parsons, R. L., J r.; Pesti, J . A.; Magnus, N. A.; Fortunak, J . M.;
Confalone, P. N.; Nugent, W. A. Org. Lett. 2000, 2, 3119. (c) Boys, M.
L.; Cain-J anicki, K. J .; Doubleday, W. W.; Farid, P. N.; Kar, M.;
Nugent, S. T.; Behling, J . R.; Pilipauskas, D. R. Org. Process Res. Dev.
increase with increasing steric demand for a considerable
range of amines but fall off sharply using the most
hindered amines. The surprising amine-dependent rates
7
1
997, 1, 233. (d) Ragan, J . A.; Murry, J . A.; Castaldi, M. J .; Conrad,
A. K.; J ones, B. P.; Li, B.; Makowski, T. W.; McDermott, R.; Sitter, B.
J .; White, T. D.; Young, G. R. Org. Process Res. Dev. 2001, 5, 498. (e)
Rico, J . G. Tetrahedron Lett. 1994, 35, 6599. (f) DeMattei, J . A.; Leanna,
M. R.; Li, W.; Nichols, P. J .; Rasmussen, M. W.; Morton, H. E. J . Org.
Chem. 2001, 66, 3330.
were traced to a dimer-based mechanism in which
destabilizing interactions in LiHMDS dimer 3 are at-
tenuated in a putative open-dimer-based transition struc-
8,9
ture 4 (Scheme 1).
(
2) For reviews of structural investigations of lithium amides, see:
a) Gregory, K.; Schleyer, P. v. R.; Snaith, R. Adv. Inorg. Chem. 1991,
7, 47. (b) Mulvey, R. E. Chem. Soc. Rev. 1991, 20, 167. (c) Beswick,
(
3
(5) For evidence that the steric hindrance of amines can make them
poor ligands for lithium, see ref 4.
M. A.; Wright, D. S. In Comprehensive Organometallic Chemistry II;
Abel, E. W., Stone, F. G. A., Wilkinson, G., Eds.; Pergamon: New York,
(6) (a) Zhao, P.; Collum, D. B. J . Am. Chem. Soc. 2003, 125, 4008.
(b) Zhao, P.; Collum, D. B. J . Am. Chem. Soc., in press.
(7) For an early suggestion that steric effects are major determinants
of solvation, see: Settle, F. A.; Haggerty, M.; Eastham, J . F. J . Am.
Chem. Soc. 1964, 86, 2076.
1
2
995; Vol. 1, Chapter 1. (d) Collum, D. B. Acc. Chem. Res. 1993, 26,
27.
(
3) Lucht, B. L.; Collum, D. B. Acc. Chem. Res. 1999, 32, 1035.
(4) For leading references to structural and mechanistic studies of
LiHMDS, see: (a) Lucht, B. L.; Collum, D. B. J . Am. Chem. Soc. 1996,
(8) (a) Remenar, J . F.; Collum, D. B. J . Am. Chem. Soc. 1997, 119,
5573. (b) Haeffner, F.; Sun, C. Z.; Williard, P. G. J . Am. Chem. Soc.
2000, 122, 12542. (c) Remenar, J . F.; Collum, D. B. J . Am. Chem. Soc.
1998, 120, 4081. (d) Remenar, J . F.; Collum, D. B. J . Am. Chem. Soc.
1998, 120, 4081.
1
1
1
1
18, 2217. (b) Romesberg, F. E.; Collum, D. B. J . Am. Chem. Soc. 1992,
14, 2112. (c) Romesberg, F. E.; Collum, D. B. J . Am. Chem. Soc. 1994,
16, 9187. (d) Romesberg, F. E.; Collum, D. B. J . Am. Chem. Soc. 1995,
17, 2166.
1
0.1021/jo030221y CCC: $27.50 © 2004 American Chemical Society
Published on Web 11/11/2003
2
42
J . Org. Chem. 2004, 69, 242-249

Li Hexamethyldisilazide-Mediated Ketone Enolization
SCHEME 1
CHART 1. Rela tive Ra te Con sta n ts (kr el, in p a r en th eses) for th e En oliza tion of 2-Meth ylcycloh exa n on e (1)
a
by LiHMDS (eq 1) in th e P r esen ce of 1.5 Equ iv of Eth er ea l Liga n d in Tolu en e a t -78 °C
a
The values are relative to enolizations in neat (ether-free) toluene (krel ) 1).
We now describe investigations of LiHMDS-mediated
ketone enolizations that were guided by two central
questions: Is there an analogous dimer-based enolization
for LiHMDS in the presence of hindered ethers, and, if
so, how do the influences of amines and ethers differ?
order rate constants (kobsd) are independent of the initial
ketone concentration (0.004-0.04 M), confirming the
first-order dependence on the concentration of 3. Zeroing
the IR baseline and monitoring a second injection afford
no significant change in the rate constant, showing that
conversion-dependent autocatalysis or autoinhibition are
unimportant under these conditions.16 Comparisons of 3
Resu lts
3
and 3-d provided large kinetic isotope effects, confirming
Kin etics: Gen er a l Meth od s. Ketone 1 was added to
17-19
rate-limiting proton transfers.
LiHMDS in hydrocarbon/R O mixtures at -78 °C. In situ
2
IR spectroscopy10 reveals that 1 (1722 cm ) is quanti-
-1
tatively converted to LiHMDS-ketone complexes (1706-
-
1
6
1
708 cm ) shown to be of general structure 3 by Li and
1
5
3,6
N NMR spectroscopies (Supporting Information).
Pseudo-first-order conditions were established by main-
taining low concentrations of ketone 1 (0.004-0.01 M)
and high, yet adjustable, concentrations of recrystal-
lized11 LiHMDS (0.05-0.40 M) and ether (0.15-1.80 M)
in toluene as the cosolvent. At such low ketone concen-
trations, only dimer 3, bearing a single coordinated
ketone, is formed.6
,10a,12-14
In all cases, the loss of 3 (or
1
5
Rela tive Ra te Con sta n ts. The rate constants for
enolization depend on the structure of the ethereal
solvent as listed in parentheses in Chart 1. The rate
3
the less reactive deuterated analogue 3-d ) follows first-
order decays to >5 half-lives. The resulting pseudo-first-
(
9) (a) Remenar, J . F.; Lucht, B. L.; Kruglyak, D.; Romesberg, F.
E.; Gilchrist, J . H.; Collum, D. B. J . Org. Chem. 1997, 62, 5748. (b)
Williard, P. G.; Liu, Q.-Y. J . Am. Chem. Soc. 1993, 115, 3380.
(13) For leading references and recent examples of detectable
organolithium-substrate complexation, see: (a) Klumpp, G. W. Recl.
Trav. Chim. Pays-Bas 1986, 105, 1. (b) Andersen, D. R.; Faibish, N.
C.; Beak, P. J . Am. Chem. Soc. 1999, 121, 7553. Also, see ref 4.
(14) For a general discussion of ketone-lithium complexation and
related ketone-Lewis acid complexation, see: Shambayati, S.; Schreiber,
S. L. In Comprehensive Organic Synthesis; Trost, B. M., Fleming, I.,
Eds.; Pergamon: New York, 1991; Vol. 1, p 283.
(
10) (a) Sun, X.; Collum, D. B. J . Am. Chem. Soc. 2000, 122, 2452.
b) Review: Rein, A. J .; Donahue, S. M.; Pavlosky, M. A. Curr. Opin.
Drug Discovery Dev. 2000, 3, 734.
11) Romesberg, F. E.; Bernstein, M. P.; Gilchrist, J . H.; Harrison,
A. T.; Fuller, D. J .; Collum, D. B. J . Am. Chem. Soc. 1993, 115, 3475.
12) Williard, P. G.; Liu, Q. Y.; Lochmann, L. J . Am. Chem. Soc.
992, 114, 348.
(
(
(
1
(15) Peet, N. P. J . Labelled Compd. 1973, 9, 721.
J . Org. Chem, Vol. 69, No. 2, 2004 243

Zhao et al.
F IGURE 2. Plot of kobsd vs [LiHMDS] for the enolization of
F IGURE 1. Plot of kobsd vs [Me
enolization of 1 (0.004 M) by LiHMDS (0.10 M) at -78 °C.
The line depicts an unweighted least-squares fit to kobsd
ax + b (a ) (1 ( 2) × 10 , b ) (2.0 ( 0.2) × 10 ).
4
THF] in toluene for the
1
4
(0.004 M) in 1.2 M Me THF/toluene at -78 °C. The line
depicts an unweighted least-squares fit to kobsd ) ax + b
)
(a ) -(1 ( 3) × 10 , b ) (2.03 ( 0.05) × 10-3).
-5
-
4
-3
SCHEME 2
constants do not, however, depend on the concentration
of the solvent. This behavior is often emblematic of the
dimer-based mechanism in Scheme 1. Casual inspection
reveals that the accelerations caused by even the most
hindered ethers are modest when compared with the
dramatic (up to 3000-fold) accelerations observed for Et
3
N
and related trialkylamines.6
Cor r ela tion of Ra tes a n d Solven t Bin d in g Con -
sta n ts. To show that the modest rate accelerations
correlate inversely with solvent-binding constants, we
used a variant22 of a J ob plot to construct a thermo-
chemical description of solvation as follows.
Ra te La w s: Hin d er ed Dia lk yl Eth er s. Detailed rate
studies of LiHMDS-mediated enolizations (eq 1) in the
presence of commercially available hindered ethers 2,2,5,5-
23
tetramethyltetrahydrofuran (Me
4 2
THF, D), i-Pr O (G),
and cineole (L) provided analogous rate laws; the results
Enolizations of 1 by LiHMDS in mixtures of two
for Me
4
THF are representative. Plots of kobsd vs Me
4
THF
solvents, S
A
and S
B
, at fixed molarity ([S
a
] + [S ] )
b
concentration (Figure 1) and kobsd vs LiHMDS20 concen-
tration (Figure 2) both display zeroth-order dependencies.
The rate law (eq 2) indicates that the observable ether-
solvated complex 3 neither gains nor loses a molecule of
constant) are depicted in Scheme 2. The rate constants
measured in the presence of each solvent alone provide
q
q
the free energies of activation, ∆G (S
Equation 4 describes kobsd in terms of mechanistic con-
stants and mole fractions24 of S
and S (X and 1 - X
respectively). The value of K provides a measure of the
relative free energies of the two ground states, ∆G°GS
corresponding to dimer-based complexes i and ii. In turn,
a
) and ∆G (S
b
).
Me
4
THF or a LiHMDS subunit on proceeding to the rate-
a
b
a
a
,
2
1
limiting transition structure, consistent with the dimer-
based mechanism described generically in eq 3 and with
structural details in Scheme 1.
A
,
q
q
∆
G (S
a
), ∆G (S ), and ∆G°GS provide the difference in the
b
transition-state energies, ∆G°TS. Results from a number
of binary comparisons are summarized in Table 1 and
are discussed in the context of the generic free energy
diagram illustrated in Figure 3.
-
d[3]/dt ) k [3]
(2)
1
k1
[
(TMS NLi) (R O)(ketone)] 98
2 2 2
(3)
By example, a plot of kobsd versus the mole fraction of
‡
24
[
(TMS NLi) (R O)(ketone)] (3)
i-Pr
i-Bu
2
O
2
2
for the enolization in mixtures of i-Pr O and
2
2
2
O is shown in Figure 4. The upward curvature in
(4)
Figure 4 is emblematic of a higher rate correlating with
the poorly coordinating solvent and is consistent with the
(
16) (a) Seebach, D. Angew. Chem., Int. Ed. Engl. 1988, 27, 1624.
b) Sun, X.; Collum, D. B. J . Am. Chem. Soc. 2000, 122, 2459. (c)
McGarrity, J . F.; Ogle, C. A. J . Am. Chem. Soc. 1984, 107, 1810. (d)
Thompson, A.; Corley, E. G.; Huntington, M. F.; Grabowski, E. J . J .;
Remenar, J . F.; Collum, D. B. J . Am. Chem. Soc. 1998, 120, 2028. (e)
Hall, P.; Gilchrist, J . H.; Harrison, A. T.; Fuller, D. J .; Collum, D. B.
J . Am. Chem. Soc. 1991, 113, 9571.
steric acceleration noted in previous studies of LiHMDS/
(
6
R
3
N-mediated enolizations. The nonlinear least-squares
fit to eq 4 affords the free energies listed in Table 1, which
(21) Edwards, J . O.; Greene, E. F.; Ross, J . J . Chem. Educ. 1968,
45, 381.
(
17) Isotope effects for LiHMDS-mediated ketone enolizations: (a)
Held, G.; Xie, L. F. Microchem. J . 1997, 55, 261. (b) Xie, L. F.;
Saunders: W. H. J . Am. Chem. Soc. 1991, 113, 3123.
(22) For a closely related application of this particular variant of
the J ob plot, see refs 6b and 8a.
(23) (a) J ob, P. Ann. Chim. 1928, 9, 113. For more recent examples
and leading references, see: (b) Huang, C. Y. Methods Enzymol. 1982,
87, 509. (c) Hubbard, R. D.; Horner, S. R.; Miller, B. L. J . Am. Chem.
Soc. 2001, 123, 5810. (d) Potluri, V.; Maitra, U. J . Org. Chem. 2000,
65, 7764.
(
18) The isotope effects fall within the range k
in Supporting Information.
19) The regioselectivity indicated in eq 1 was shown to be >20:1
by trapping with Me SiCl/Et N mixtures.
20) The concentration of the lithium amide, although expressed in
units of molarity, refers to the concentration of the monomer unit
normality).
H
/k
D
) 10-20 as noted
(
3
3
(
(24) Mole fraction is defined as the fraction of the total number of
moles of coordinating solvent: [i-Pr NH] + [i-Pr O] ) 1.2 M.
2 2
(
2
44 J . Org. Chem., Vol. 69, No. 2, 2004

Li Hexamethyldisilazide-Mediated Ketone Enolization
TABLE 1. F r ee En er gies for th e LiHMDS-Med ia ted En oliza tion of 1 in Bin a r y Solven t Mixtu r es (Sch em e 2, F igu r e 3)a
q
q
entry
Sa
Sb
∆G (Sa)
∆G (Sb)
∆G°GS
∆G°TS
1b
2
3
4
5
i-Bu2O
i-Pr2O
i-Bu2O
i-Pr2O
i-PrOEt
i-Pr2O
15.4 ( 0.1
14.9 ( 0.1
15.4 ( 0.2
14.9 ( 0.1
15.3 ( 0.1
14.9 ( 0.1
13.5 ( 0.1
13.5 ( 0.1
13.0 ( 0.1
14.0 ( 0.1
0.3 ( 0.1
0.27 ( 0.08
0.83 ( 0.02
0.79 ( 0.02
0.04 ( 0.09
-0.2 ( 0.1
-1.1 ( 0.1
-1.1 ( 0.2
-1.1 ( 0.1
-1.28 ( 0.09
Me4THF
Me4THF
i-Pr2NH
i-PrNHEt
a
Conditions: [LiHMDS] ) 0.10 M; [solvent]total ) 1.2 M; toluene cosolvent. b Nearly equivalent rates attenuates the accuracy of the
method.
F IGURE 3. Thermochemical description of LiHMDS-mediated enolizations in binary solvent mixtures (Scheme 2, Table 1).
tive ∆G°TS). Additional binary comparisons listed in Table
1
also prove instructive.
A plot of kobsd vs the mole fraction of Me
4
THF24 for the
4 2
LiHMDS-mediated enolizations in Me THF/i-Pr O mix-
tures (Figure 5) displays a slight upward curvature
consistent with nearly identical solvation of the reactants
(
∆G°GS ) 0). Thus, the approximate 20-fold higher rate
for the LiHMDS/Me THF-mediated enolization derives
4
almost entirely from a net stabilization of the transition
structure. The nonlinear least-squares fit to the data
affords the free energies listed in Table 1 (Table 1, entry
2
). We will discuss the effects of increasing substitution
within the ethereal ligand on the reactant and transition
structure in the context of competing steric and electronic
effects.
kobsd ) [k + (k - k K )X /K ]/[1 + (1 - K )X /K ]
b
a
b
A
a
A
A
a
A
F IGURE 4. Plot of kobsd vs i-Pr
2
O mole fraction in i-Pr
for the enolization of 1 (0.004 M) by LiHMDS (0.10 M)
at -40 °C. The curve depicts an unweighted least-squares fit
2
O/i-
2
4
Bu
2
O
(4)
-
3
to y ) (a + bx)/(1 + cx), where a ) (2 ( 2) × 10 , b )
such that X
constant molarity.
a
) [S
a
]/([S
a
] + [S
b
]) and ([S
a
] + [S
b
]) )
-
2
(
1
1.2 ( 0.3) × 10 , c ) -0.905 ( 0.009. K
A 2 2
(i-Pr O/i-Bu O) )
+ c ) 0.095 ( 0.009.
The veracity of the data in Table 1 can be supported
by simple cross checks. By example, the binary compari-
are expressed by the free energy diagram shown in
q
Figure 3. As expected, the acceleration reflected by ∆G -
sons of i-Bu
(entry 2) allow one to predict the result from a compari-
son of i-Bu O vs Me THF and then compare it to the
2 2 2 4
O vs i-Pr O (entry 1) and i-Pr O vs Me THF
q
(i-Pr
2
O) < ∆G (i-Bu
2
O) derives in part from a relative
destabilization of the reactant (reflected in ∆G°GS) by the
sterically more demanding ligand. It was altogether
unexpected, however, that the more sterically congested
ligand would also impart a net stabilization of the open-
dimer-based transition structure (reflected by the nega-
2
4
experimental result (entry 3), showing strong self-
consistency within a reasonable experimental error.
Com p a r ison of R NH vs R O. Table 2 lists relative
2 2
rate constants for a select group of isostructural dialkyl-
J . Org. Chem, Vol. 69, No. 2, 2004 245

Zhao et al.
F IGURE 6. Plot of kobsd vs i-Pr
i-Pr O
2
2
NH mole fraction in i-Pr
for the enolization of 1 (0.004 M) by LiHMDS (0.10
M) at -78 °C. The curve depicts an unweighted least-squares
2
NH/
2
4
-
5
fit to y ) (a + bx)/(1 + cx), where a ) (1 ( 1) × 10 , b )
F IGURE 5. Plot of kobsd vs Me
4
THF mole fraction in Me
for the enolization of 1 (0.004 M) by LiHMDS (0.10
M) at -78 °C. The curve depicts an unweighted least-squares
4
THF/
-
3
2
4
(1.2 ( 0.2) × 10 , c ) -0.871 ( 0.008. K (i-Pr NH/i-Pr O) )
i-Pr
2
O
A
2
2
1
+ c ) 0.129 ( 0.008.
-
5
fit to y ) (a + bx)/(1 + cx), where a ) (9 ( 9) × 10 , b )
-
3
(
1
1.3 ( 0.3) × 10 , c ) -0.5 ( 0.1. K
+ c ) 0.5 ( 0.1.
A 4 2
(Me THF/i-Pr O) )
TABLE 2. Rela tive Ra tes (kr el) for th e En oliza tion by
LiHMDS Solva ted by Isostr u ctu r a l Dia lk yla m in es
(
R1NHR2) a n d Dia lk yl Eth er s (R1OR2)
R1XR2
kamine/kether
R1 ) R2 ) i-Bu
R1 ) R2 ) i-amyl
R1 ) Et, R2 ) i-Pr
R1 ) R2 ) i-Pr
2
10
35
170
60
R1 ) R2 ) s-Bu
a
Enolization rates were measured with 0.1 M LiHMDS, 0.004
M 2-methylcyclohexanone (1), and 1.2 M ligand in a toluene
solution at -78 °C.
amines and dialkyl ethers. Comparing the hindered
i-Pr
tion by the i-Pr
stants measured as a function of solvent mole fraction.
Once again, the free energies listed in Table 1 are
qualitatively consistent with the free-energy diagram in
2
O/i-Pr
2
NH pair reveals a markedly greater accelera-
F IGURE 7. Plot of kobsd vs i-PrNHEt mole fraction in
i-PrNHEt/i-PrOEt24 for the enolization of 1 (0.004 M) by
LiHMDS (0.10 M) at -78 °C. The curve depicts an unweighted
least-squares fit to y ) (a + bx)/(1 + cx), where a ) (4 ( 2) ×
2
NH. Figure 6 illustrates the rate con-
2
4
-
5
-4
1
0
, b ) (5.7 ( 0.9) × 10 , c ) -0.1 ( 0.1. K
A
(i-PrNHEt/
i-PrOEt) ) 1 + c ) 0.9 ( 0.1.
Figure 3. In particular, the higher rates using i-Pr
2
NH
compared to the isostructural i-Pr O derive from a
2
that the ketone enolization in eq 1 is dramatically
combination of relative destabilization of the hindered
lithium amide dimer 3 and relative stabilization of open-
dimer-based transition structure 4. A different result is
obtained by comparing i-PrOEt and i-PrNHEt (Figure
accelerated by poorly coordinating trialkylamines due to
6
the intervention of a dimer-based pathway (Scheme 1).
The correlation of high rates with solvation of LiHMDS
by hindered di- and trialklyamines was traced to steric
relief that occurs when the reaction proceeds from the
congested dimer-based LiHMDS-ketone complex 3 to the
relatively less congested8 open-dimer-based transition
7
). The higher rates for i-PrNHEt derive almost entirely
from stabilization of the transition state (∆G°GS ) 0). The
third and possibly most aberrant result derives from a
comparison of i-Bu
equivalent rates and binding constants are observed;
G°GS and ∆G°TS are too small to measure using this
2 2
O and i-Bu NH in which nearly
structure 4 as depicted using the generic free-energy
diagram in Figure 8.2
6,27
These results offer a tribute to
∆
method.
(25) Collum, D. B. Acc. Chem. Res. 1992, 25, 448.
(
26) The relative solvation energies of the reactant (∆G°GS) and
transition structure (∆G°TS) for a series of amines display a linear free
energy relationship of the form:
Discu ssion
Conventional wisdom suggests that the LiHMDS-
mediated enolizations of ketones are accelerated by
strongly coordinating ligands due to intervening deaggre-
gation.25 Nonetheless, previous investigations showed
∆
G°TS ) 0.6∆G°GS
The high linearity and simplicity of the relationship is remarkable
given that specific through-space (van der Waals) interactions can
render steric effects very complex.
27
2
46 J . Org. Chem., Vol. 69, No. 2, 2004

Li Hexamethyldisilazide-Mediated Ketone Enolization
F IGURE 8. Thermochemical description of amine-mediated enolizations noted previously.7a
computational chemistry in that a facile open-dimer-
based ketone enolization and an accompanying steric
acceleration emerged quite unexpectedly from semiem-
pirical calculations in 1992.4
We prefaced the current study with two questions: (1)
Is there an analogous dimer-based enolization for Li-
HMDS in the presence of hindered ethers, and (2) if so,
how do the influences of amines and ethers differ? Rate
studies reveal that LiHMDS solvated by sterically de-
the reactant and stabilizing the transition structure.
Once again, however, the mechanistic picture is not
altogether simple. Comparison of less hindered isostruc-
tural ligands i-PrOEt and i-PrHNEt (Table 1, entry 6)
reveals that i-PrHNEt accelerates the enolization almost
entirely due to a net stabilization of the transition
structure.
Competing steric effects that are dominant in the
ground state and electronic effects that can be dominant
in the transition state offer an adequate model for the
solvent effects as follows.
4 2
manding ethers, Me THF (D), i-Pr O (G), and cineole (L),
enolizes ketone 1 via the dimer-based pathway shown
in Scheme 1. The accelerations shown in Chart 1,
however, are muted compared to the 3000-fold accelera-
(1) Ster ic Effects. Trialkylamines are functionally
larger than even the most hindered dialkyl ethers since
ethers coordinated to lithium are trigonal planar at
oxygen28 whereas the corresponding amines are neces-
3
tion observed for LiHMDS/Et N mixtures. Do the attenu-
ated rates for the ethers simply reflect the lesser steric
demands of even the most hindered dialkyl ethers when
compared with simple trialkylamines? To address this
point, we turned to a variant22 of a J ob plot to correlate
the ligand-dependent rates with the steric and electronic
demands of the coordinated solvent. For example, a plot
2
9
sarily tetrahedral at nitrogen (cf. 5 and 6). Because open
dimers and open-dimer-based transition structures are
23
4b
suggested to be less congested than the cyclic dimers,
extremely hindered ligands such as trialkylamines mark-
edly destabilize the congested dimeric reactant 3 more
than the open-dimer-based transition structure 4, result-
ing in sterically driven accelerations. By contrast, the
sterically less demanding dialkyl ethers and their di-
alkylamine counterparts should display attenuated steric
accelerations and be nearly indistinguishable provided
that the alkyl groups are not large.4 Apparently, the
smallest ligands studied, i-PrNHEt and i-PrOEt, are
below what Brown refers to as the “minimum steric
threshold”30 as evidenced by the absence of measurable
differences in the energies of the reactants.4a
2 2 2
of kobsd vs mole fraction of i-Pr O in i-Pr O/i-Bu O
mixtures (Figure 4) afforded the thermochemical data in
Table 1 (entry 1) and the accompanying free-energy
diagram depicted generically in Figure 3. The higher
rates for LiHMDS solvated by i-Pr
destabilization of the reactant dimer as reflected by
G°GS. Surprisingly, however, the more sterically de-
manding i-Pr O also imparts a net stabilization of the
2
O derive from a
a
∆
2
open-dimer-based transition structure. Although we were
initially tempted to ascribe the odd reversal in solvation
to the unusual steric demands of isobutyl groups noted,4
a,6b
(2) Electr on ic Effects. Cyclic dimers of lithium
amides appear to be less electrophilic (toward coordinat-
the same pattern, a positive ∆G°GS and a negative ∆G°TS
,
observed for other binary comparisons of hindered ethers
ing ligands) than their open dimer or monomer counter-
parts.3
,4b,8a,31
Therefore, increased Lewis basicity would
(Table 1) suggests a greater generality.
Dialkylamines show more dramatic accelerations that
preferentially stabilize the open-dimer-based transition
appear to derive from similar effects. Thus, the sterically
demanding i-Pr NH accelerates the enolization when
compared with the isostructural i-Pr O by destabilizing
2
(28) Chakrabarti, P.; Dunitz, J . D. Helv. Chim. Acta 1982, 64, 1482.
(
29) Impey, R. W.; Sprik, M.; Klein, M. L. J . Am. Chem. Soc. 1987,
09, 5900.
30) Choi, M.-G.; Brown, T. L. Inorg. Chem. 1993, 32, 1548.
2
1
(
(27) Brown, T. L.; Lee, K. J . Coord. Chem. Rev. 1993, 128, 89.
(31) Ram ´ı rez, A.; Collum, D. B. J . Am. Chem. Soc. 1999, 121, 11114.
J . Org. Chem, Vol. 69, No. 2, 2004 247

Zhao et al.
F IGURE 9. Effect of decreasing steric demands of the coordinating ligand on the relative activation energies. Curve a corresponds
to a benchmark dimer solvated by an unhindered ethereal ligand. Curve b corresponds to more highly substituted ether or amine.
The dotted arrows indicate decreasing steric demands of the more congested solvate.
the open-dimer-based transition structure become domi-
nant (Figure 9C). Further removal of substituents would
not influence the steric environment but would decrease
the stabilization of the transition structure.
There is an outlying data point, however, we simply
2 2
do not understand; i-Bu NH and i-Bu O were found to
structures. In addition, the functional Lewis basicity of
a ligand depends on a number of factors, including (a)
the steric effects noted above, (b) the electrophilicity of
the Lewis acid,32 (c) inductive electron donation resulting
be nearly indistinguishable not only in the ground state
but also in the transition state. Although equivalent
binding in the ground state was anticipated from previ-
4a
ous spectroscopic studies, it is unclear why the en-
from substitution on the alkyl groups of the Lewis base
hanced basicity of the amine does not cause a pronounced
stabilization of the transition structure.
(
Et versus i-Pr),3 and (d) the inherent nucleophilicity of
2
3
4
the coordinating atom (N vs O). Although the steric
demands of Me THF are a little unclear, the stabilization
4
Su m m a r y a n d Con clu sion s
of the transition structure due to putative electronic
effects is consistent with the inherently high Lewis
basicity of tetrahydrofurans.3 In the absence of mitigat-
ing steric effects, the Lewis basicity of the ligand should
be greater for amines than for ethers34 and should be
enhanced by increased substitution on the ligands due
to inductive electron donation.33
Previous work showed that LiHMDS solvated by di-
and trialkylamines effects ketone enolization via a steri-
cally accelerated dimer-based enolization shown in Scheme
5
6
1. Analogous accelerations by hindered dialkyl ethers
are traced to sterically derived destabilization of the
reactant in conjunction with stabilization of the transition
structure due to putative electronic effects. In general,
however, the ethers are not sterically demanding enough
to elicit high reactivities. Dialkylamines also destabilize
the reactant and stabilize the transition structures when
compared with the hindered ethers, eliciting higher
overall rates.
In the most general sense, increasing the reactivities
of lithium amides, whether through amine or ether
additives, is potentially important in synthesis. Although
the trialkylamines impart the largest accelerations, the
ethers may be more appropriate if subsequent reactions
use highly reactive electrophiles that can N-alkylate
secondary amines. Most importantly, however, enolates
solvated by poorly coordinating ligands may display
interesting reactivities.
By placing the thermochemical depictions proximate
(
Figure 9), we have constructed a reasonably self-
consistent model that serves as a useful mnemonic. For
the most hindered ligands, electronic stabilizations of the
transition structures are obscured by dominant steric
effects in the congested cyclic dimer reactants (Figure
9
A). As the steric influences on ∆G°GS and ∆G°TS decrease
(indicated by the dotted arrows), the electronic biases
favoring the open-dimer-based transition structures be-
come observable (Figure 9B). Under these circumstances,
dialkylamines coordinate to the open-dimer-based transi-
tion structure more strongly than do dialkyl ethers and
substituted ethers bind more strongly than lesser sub-
stituted ethers. As the ligands fall below the minimum
steric threshold, steric effects localized primarily in the
reactant disappear and electronic effects influential in
Exp er im en ta l Section
(32) (a) Ivanova, S. M.; Nolan, B. G.; Kobayashi, Y.; Miller, S. M.;
Anderson, O. P.; Strauss, S. H. Chem.-Eur. J . 2001, 7, 503. (b) Reed,
Rea gen ts a n d Solven ts. Amines and hydrocarbons were
routinely distilled by vacuum transfer from blue or purple
solutions containing sodium benzophenone ketyl. The hy-
drocarbon stills contained 1% tetraglyme to dissolve the
ketyl. 6Li metal (95.5% enriched) was obtained from Oak
C. A. Acc. Chem. Res. 1998, 31, 133.
(33) (a) Baeten, A.; De Proft, F.; Langenaeker, W.; Geerlings, P.
THEOCHEM 1994, 112, 203. (b) Safi, B.; Choho, K.; De Proft, F.;
Geerlings, P. Chem. Phys. Lett. 1999, 300, 85. (c) Headley, A. D. J .
Org. Chem. 1991, 56, 3688.
6
(
34) (a) March, J . Advanced Organic Chemistry; Wiley: New York,
Ridge National Laboratory. The LiHMDS, [ Li]LiHMDS, and
1
992; Chapter 8. (b) Gutmann, V. The Donor-Acceptor Approach to
6
15
11
[
Li, N]LiHMDS were prepared and purified as described.
Molecular Interactions; Plenum: New York, 1978. (c) Marcus, Y. J .
Solution Chem. 1984, 13, 599. (d) Kaufmann, E.; Gose, J .; Schleyer,
P. v. R. Organometallics 1989, 8, 2577.
15
Ketone 1-d
3
has been prepared as described. Air- and
moisture-sensitive materials were manipulated under argon
or nitrogen using standard glovebox, vacuum line, and syringe
techniques.
(
35) Berthelot, M.; Besseau, F.; Laurence, C. Eur. J . Org. Chem.
1
940, 5, 925.
2
48 J . Org. Chem., Vol. 69, No. 2, 2004

Li Hexamethyldisilazide-Mediated Ketone Enolization
NMR Sp ectr oscop ic An a lyses. Samples were prepared,
J ohnson, and Schering-Plough for indirect support. P.Z.
thanks Boehringer-Ingelheim for a fellowship. We also
acknowledge the National Science Foundation Instru-
mentation Program (CHE 7904825 and PCM 8018643),
the National Institutes of Health (RR02002), and IBM
for support of the Cornell Nuclear Magnetic Resonance
Facility.
6
15
13
and the Li, N, and C NMR spectra were recorded as
6
described elsewhere. The LiHMDS samples containing both
ketone 1 and ethers were prepared using a specific protocol
6
designed to minimize unwanted enolization.
IR Sp ectr oscop ic An a lyses. In situ IR spectra were
recorded using a 30-bounce silicon-tipped probe using fully
1
0a
established procedures.
Su p p or tin g In for m a tion Ava ila ble: NMR spectra and
rate data. This material is available free of charge via the
Internet at http://pubs.acs.org.
Ack n ow led gm en t. We thank the National Insti-
tutes of Health for direct support of this work as well
as Dupont Pharmaceuticals (Bristol-Myers Squibb),
Merck Research Laboratories, Pfizer, Aventis, R. W.
J O030221Y
J . Org. Chem, Vol. 69, No. 2, 2004 249