Meyer and Kass
JOCArticle
TABLE 1. Relative Stabilities of the r-, β-, γ-, and δ-Enolate Anions of
interconversion of the βeq anion to its axial and cyclic
counterparts should occur readily since the isomerization
barriers are small (i.e., 0.3 and 10.4 kcal mol-1, respectively).
To further examine the deprotonation site of 2-adaman-
tanone, the β-anion (1β) was selectively generated via the
fluoride-induced desilylation of an ∼1:1 mixture of axial and
equatorial 4-trimethylsilyl-2-adamantanone (eq 4).15 The
resulting anion (which presumably consists of a
2-Adamantanone Computed at the B3LYP, M06-2X, and G3 Levelsa
M06-2X
anion
B3LYP DZb
DZb
TZc
G3
R
6.33
0.00
5.08
2.46
2.30
8.16
4.27
0.00
4.46
-1.58
2.28
6.17
4.25
0.00
4.60
-2.08
2.21
6.45
5.60
0.00
-
0.56
3.40
7.55
βax
βeq
βcy
γ
d
δ
aAll values are at 298 K and are in kcal mol-1. bDZ = aug-cc-pVDZ.
cTZ = maug-cc-pV(Tþd)Z (see ref 14). dThe 1βeq anion cyclizes upon
optimization at the MP2(full)/6-31G(d) level, which precludes a G3
energy determination.
To address this issue, DFT calculations were carried out
on 2-adamantanone and its conjugate bases. B3LYP was
employed because of its successful track record and popu-
larity, but the more recent M06-2X functional developed by
Truhlar et al.13 was also used because in its early testing it was
found to be more accurate. All five deprotonation sites in 1
were examined and six different anions were located since the
β-anion can cyclize to form a three-membered ring. The
structures of the R- and β-enolates are given in Figure 1 and
all of the computed geometries can be found in the Support-
ing Information. G3 energies were also computed since this
model chemistry has been extensively benchmarked and is
often accurate to within 2 kcal mol-1.8 The relative energies
of the six enolate anions are given in Table 1 and all three
approaches indicate that the stability order is β > γ > R >
δ. These predictions are in accord with the H/D exchange
results of Stothers in t-BuOD in that the βax anion is found to
be more stable than the R and βeq anions. In addition, the
M06-2X calculations (but not B3LYP) indicate that the
R-position in 1 is more acidic than the βeq hydrogens, which
reproduces the observed βax > R > βeq H/D exchange
behavior.7 All of the calculations indicate that the γ-anion
is more stable than the R-anion, however, but the former site
was not observed to undergo H/D exchange.
The relative instability of the R-enolate of 1 is due to its
inability to delocalize the charge via resonance, and the
resulting electron-electron repulsion between the lone pairs
of electrons on the carbon at the carbanion center and the
oxygen atom. The latter interaction results in the anion
breaking symmetry (CS f C1) to diminish this repulsion.
Since H/D exchange is observed at the R-position in solution,
this presumably is due to the interaction of the carbonyl
oxygen atom with the cation associated with the base.
As for the preferred structure of the β-anion, it may well be
the cyclized species as predicted with the M06-2X functional,
but the energy difference between the βcy and βax ions is small
and their relative stability is reversed by B3LYP and G3
calculations. Our computations do not lead to a clear pre-
diction for the preferred structure of the β-enolate anion,
consequently, but they do indicate that the β-position is the
most acidic site in 2-adamantanone. Additional calculations
at the M06-2X/aug-cc-pVDZ level also indicate that the
rapidly interconverting mixture of 1βax and 1βcy) was
transferred to the second reaction cell in the FTMS where
it was collisionally cooled with a pulse of argon and then
allowed to react with a variety of neutral reagents. Water
rapidly protonated 1β, dimethylamine reacted slowly (k =
(3.14 ( 0.14) ꢀ 10-10 cm3 molecules-1 s-1), and ethylamine
and ammonia did not undergo proton transfer at all, but
the addition - H2O products shown in eqs 1b and 2b were
produced. These results are qualitatively the same as for
deprotonated 2-adamantanone, which indicates that de-
pronation takes place primarily, if not exclusively, at the
β-position. In addition, the rate constants for the depro-
tonation of 1 with Me2N- and the protonation of 1β with
Me2NH can be combined to derive an equilibrium constant
since K = k1/k-1 as illustrated in eq 5.
The resulting value for K is 6.05 ( 0.49, which leads
to ΔG°acid(1 - Me2NH) = -1.1 ( 0.6 kcal mol-1 if a
conservative uncertainty of (100% for K is used in the
error analysis. This free energy acidity difference can be
combined with the literature value for dimethylamine
(ΔG°acid(Me2NH) = 388.2 ( 0.9 kcal mol-1) to afford
ΔG°acid(1) = 387.1 ( 1.1 kcal mol-1. To obtain the depro-
tonation enthalpy, the entropies for 1 and 1β are needed.
These quantities were computed by using unscaled B3LYP
and M06-2X vibrational frequencies with the aug-cc-
pVDZ basis set. Both methods provide TΔS°acid(1) terms
that differ by only 0.4 kcal mol-1 for the βax and βcy
anions, and only 0.1 kcal mol-1 when the same ion is used.
If one assumes uncertainties of (2 eu for the individual
entropies, then M06-2X leads to TΔS°acid(1) = 7.6 (
0.9 kcal mol-1 for the cyclic ion. This enables us to assign
ΔH°acid(1) = 394.7 ( 1.4 kcal mol-1, which is in good accord
with the bracketing results. It also is well-reproduced by
B3LYP/aug-cc-pVDZ, M06-2X/aug-cc-pVDZ, M06-2X/
maug-cc-pV(Tþd)Z,14 and G3 calculations which give depro-
tonation enthalpies of 394.0, 393.8, 394.1, and 396.8 kcal mol-1
,
respectively. Our equilibrium acidity for 1 indicates that it is a
(13) (a) Zhao, Y.; Truhlar, D. G. J. Phys. Chem. A 2008, 112, 1095–1099.
(b) Zhao, Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215–241. (c) Zhao,
Y.; Truhlar, D. G. Acc. Chem. Res. 2008, 41, 157–167.
(14) Papajak, E.; Truhlar, D. G. J. Chem. Theory Comput. 2010, 6, 597–
601.
(15) (a) Duddeck, H.; Islam, M. R. Chem. Ber. 1984, 117, 554–564. (b)
Vodicka, L.; Triska, J.; Hajek, M. Collect. Czech. Chem. Commun. 1980, 45,
2670–2674.
4276 J. Org. Chem. Vol. 75, No. 12, 2010