Hydrogen bonding and solvent effects in heteroaryldi(1-adamantyl)methanols: An NMR and IR spectroscopic study

Article abstract of DOI:10.1039/a806485g

Reaction ot the organolithium derivatives of certain heteroaromatics [2-furanyl, 2-thienyl, 2-thiazolyl, 2-pyridyl and 2-(3-methylpyndyl)] with di(1-adamantyl) ketone gives potentially rotameric tertiary alcohols With 2-pyndyl- and 2-(3-methylpyridyl)lithium only the syn isomer is obtained. The syn isomer makes up 85-100% of (2-furanyl)diadamantylmethanol and 80-90% of the 2-thienyl derivative, depending on the NMR solvent. In chloroform or benzene the 2-thiazolyl derivative is a 2:1 mixture, the isomer with the sulfur atom syn to the OH group predominating; in DMSO or in the solid state this is the sole species. The IR absorption frequency for OH stretching correlates with the corresponding proton NMR shift in chloroform and with its temperature dependence Δδ/ΔT. In pyridine Δδ/ΔT is either large (-20 ppb/°C) or small (-1 to -2 ppb/°C) for intermolecular and intramolecular hydrogen-bonded species, respectively. Semi-empirical calculations (AM1 and PM3) suggest that the anti alcohols should be the more stable in the gas phase, but solvent effects and hydrogen bonding in the case of the pyridyl derivatives, appear to reverse this situation, making the isomer with OH syn to the heteroatom the principal, and sometimes the only, species observed in solution.

Full text of DOI:10.1039/a806485g

View Article Online / Journal Homepage / Table of Contents for this issue
Hydrogen bonding and solvent effects in heteroaryldi(1-
adamantyl)methanols: an NMR and IR spectroscopic study
John S. Lomas,* Alain Adenier, Christine Cordier and Jean-Christophe Lacroix
Institut de Topologie et de Dynamique des Systèmes, Université de Paris 7, associé au C.N.R.S.,
1
rue Guy de la Brosse, 75005 Paris, France
Received (in Cambridge) 17th August 1998, Accepted 23rd September 1998
Reaction of the organolithium derivatives of certain heteroaromatics [2-furanyl, 2-thienyl, 2-thiazolyl, 2-pyridyl
and 2-(3-methylpyridyl)] with di(1-adamantyl) ketone gives potentially rotameric tertiary alcohols. With
2
-pyridyl- and 2-(3-methylpyridyl)lithium only the syn isomer is obtained. The syn isomer makes up 85–100%
of (2-furanyl)diadamantylmethanol and 80–90% of the 2-thienyl derivative, depending on the NMR solvent. In
chloroform or benzene the 2-thiazolyl derivative is a 2:1 mixture, the isomer with the sulfur atom syn to the OH
group predominating; in DMSO or in the solid state this is the sole species. The IR absorption frequency for OH
stretching correlates with the corresponding proton NMR shift in chloroform and with its temperature dependence,
∆
δ/∆T. In pyridine ∆δ/∆T is either large (Ϫ20 ppb/ЊC) or small (Ϫ1 to Ϫ2 ppb/ЊC) for intermolecular and
intramolecular hydrogen-bonded species, respectively. Semi-empirical calculations (AM1 and PM3) suggest that
the anti alcohols should be the more stable in the gas phase, but solvent effects and hydrogen bonding, in the case
of the pyridyl derivatives, appear to reverse this situation, making the isomer with OH syn to the heteroatom the
principal, and sometimes the only, species observed in solution.
Introduction
Results and discussion
It was shown recently that 2-lithio-N-alkylpyrroles (alkyl = Me
or Et), 1, are sufficiently nucleophilic to attack the very con-
gested carbonyl group in di(1-adamantyl) ketone, with the
Alcohol synthesis
3
4
Treatment of 2-bromopyridine, 2-bromo-3-methylpyridine, 2-
5
6
bromothiophene or 2-bromothiazole with n-butyl- or tert-
butyllithium under various conditions gave the corresponding
lithio compounds. 2-Lithiofuran was prepared directly by
treatment of furan with n-butyllithium in the presence of
formation of the thermodynamically less stable anti isomer of
1
[
2-(N-alkylpyrrolyl)]di(1-adamantyl)methanol, 1A. Rotation
2
3
about the sp –sp C–C bond has an activation energy of about
1 kcal mol † and gives the syn isomer, 1S, which according to
Ϫ1
3
7
TMEDA. Starting materials were used in excess, in order to
AM1 and ab initio calculations is more stable by about 5 kcal
avoid dilithiation. Generally, a large excess of organolithium
reagent over di(1-adamantyl) ketone was used. No attempt was
made to optimize yields by systematic variation of the con-
ditions, but progress of the reaction was monitored by sampling
the reaction mixture from time to time after addition of the
ketone to the organolithium compound.
The 2-lithio derivatives of the selected heteroaromatic sys-
tems react with diadamantyl ketone giving satisfactory yields of
the corresponding alcohols. The aromatic carbon or hydrogen
NMR signals were inspected to determine whether one or
two isomers are formed. In the case of (2-pyridyl)di-
adamantylmethanol, 4, and the 3-methyl derivative, 5, the
product is a single isomer. Attempts to generate a second
rotamer or modify the rotamer ratio by heating at 150 ЊC
resulted in 10–17% conversion of 4, depending on the solvent,
while 5 decomposed under these conditions. The corresponding
Ϫ1
mol . Although there are some minor quantitative differences,
doubtless due to the smaller size of the aromatic ring, the prop-
erties of 1A and 1S are closely analogous to those of o-tolyl-
2
diadamantylmethanols, 3A and 3S, the chemistry being to a
large extent controlled by the N-methyl substituent. It was
of interest, therefore, to examine a series of heteroaryldi-
adamantylmethanols without this dominant feature (with one
exception), in order to study structural and spectroscopic
properties more characteristic of the heterocycle. The systems
studied here were chosen essentially for the accessibility of the
appropriate organolithium derivatives and the proximity of the
heteroatom to the OH group in the target alcohols.
R
N
N
R
R
2
-furanyl compound, 6, is apparently a single isomer in DMSO
Ad
OH
Ad
OH
Ad
OH
or pyridine but two isomers in a ratio of about 9:1 or 6:1 in
benzene or chloroform, respectively. In the same way, the 2-
thienyl compound, 7, is a mixture of two isomers in a ratio of
about 9:1 or 4:1, depending on the NMR solvent. According
to its NMR spectra in chloroform or benzene the 2-thiazolyl
derivative, 8, is a 2:1 isomer mixture; the proton NMR spec-
trum in pyridine indicates about 4% of the second isomer and
in DMSO only one isomer can be detected. Alcohol 8 slowly
decomposed to diadamantyl ketone when heated in chloroform
at 150 ЊC.
Ad
Ad
Ad
1
A: R = Me or Et
1S: R = Me or Et
2: R = H
3
A: R = Me
R
Ad
OH
Ad
Structure determination by NMR
3S: R = Me
1
13
The H and C NMR spectra of all the new alcohols could be
fully attributed on the basis of the proton–proton coupling
†
1 cal = 4.184 J.
J. Chem. Soc., Perkin Trans. 2, 1998, 2647–2652
2647

View Article Online
doubts as to the reality of this type of hydrogen bonding which
must in any case be very weak. Similar two-fold absorptions in
sterically constrained dialkylbenzyl alcohols, such as 2 and 3A,
where the C–O bond is almost coplanar with the benzene ring,
are better attributed to different conformations with respect to
O
N
N
R
R
Ad
OH
Ad
OH
Ad
OH
17
Ad
Ad
Ad
the C –C–O–H torsion angle. There are no crystallographic
ar
data on any of the compounds in the present study but the
4S: R = H
6A
4
A: R = H
related [2-(N-methylpyrrolyl)]di(1-adamantyl)methanols, 1A
5A: R = Me
5S: R = Me
1
and 1S (R = Me), are structurally similar to the congested
18
dialkylbenzyl methanols. Molecular mechanics and semi-
empirical quantum chemical calculations (vide infra) indicate
O
S
S
small C –C–O–H torsion angles.
ar
In this work the IR spectra of the alcohols (Table 1) were
measured in carbon tetrachloride, except for the 2-thiazolyl
mixture, 8A and 8S, which was insufficiently soluble and was
therefore studied in chloroform. The 2-furanyl and 2-thienyl
derivatives, 6S and 7S, show strong absorptions at 3620 and
Ad
OH
Ad
OH
Ad
OH
Ad
7A
Ad
7S
Ad
6
S
Ϫ1
Ϫ1
3
624 cm , respectively, with a strong shoulder at 3608 cm .
Ϫ1
S
S
The 2-pyridyl analogue, 4S, has a single peak at 3313 cm ,
N
N
consistent with intramolecular hydrogen bonding. Treatment at
1
50 ЊC, giving presumably the anti isomer, 4A, is associated
OH
OH
Ad
Ad
Ϫ1
with the appearance of a very weak absorption at 3637 cm .
This is a clear indication that any hydrogen bond formed in
the syn isomer is specific to the nitrogen atom rather than to
the π-electron system as a whole. The 2-(3-methylpyridyl) deriv-
ative, 5S, absorbs at even lower frequency, 3165 cm . Alcohol
8 has absorptions at 3619 and 3608 cm and at 3427 and
Ad
Ad
8S
8A
8
Ϫ1
constants, spectrum simulation by means of the gNMR pro-
9
1
Ϫ1
gram and heteronuclear correlation. H NMR NOE experi-
ments in chloroform were used in some cases to determine spa-
tial relationships and thereby to establish the conformation of
the isomeric alcohols. They consisted essentially in irradiating
the various peaks of the adamantyl group signal and observing
the effects on the other proton signals. Since C NMR shifts
are much less solvent-sensitive than proton shifts they were
used systematically to identify rotamers in solvents other than
chloroform.
NOE study of the 2-pyridyl, 2-(3-methylpyridyl), 2-furanyl
and 2-thienyl alcohols (4, 5, 6 and 7) indicates that the sole or
major product (in chloroform) is the syn isomer. NOE cannot
be used to differentiate the isomers of the 2-thiazolyl derivative,
Ϫ1
3378 cm ; these can be attributed to the 8S and 8A rotamers,
respectively, the latter low-frequency absorption being char-
acteristic of hydrogen bonding to nitrogen, as in 4S. This
absorption is absent when the spectrum is run in KBr, suggest-
ing that in the solid, as in DMSO, the stable species is the syn
isomer.
13
Intramolecular hydrogen bonding has been claimed for (2-
19
pyridyl)-substituted alcohols, though the IR frequency shifts
are substantially smaller than found here, of the order of
190 cm for 2-pyridylmethanol, for example. Later work has
shown that the NMR shift of the hydroxy proton is highly
concentration-dependent, moving upfield with increasing dilu-
tion. Clearly, both inter- and intramolecular hydrogen bond-
ing are involved. The IR absorption for 4S is almost 100 cm
lower than for the intramolecular hydrogen-bonded species
(about 80–90% of the total) in the related (2-pyridyl)-
arylmethanols and about 150 cm lower than in some
quinoxaline diols of similar geometry. 2,3-Di(2-pyridyl)-
butane-2,3-diols and related compounds also show large νOH
shifts and high δOH, but the crystallographic evidence
indicates that the hydrogen bonding involves a six-membered
rather than a five-membered ring, which occurs in compounds
examined here (4S, 5S and 8A).
Ϫ1
20
8
, since there is no aromatic hydrogen close to the adamantyls.
Ϫ1
However, the high chemical shift (5.22 ppm) of the OH proton
in the minor component, by analogy with the 2-pyridyl com-
pound, 4, strongly suggests that it is the isomer with nitrogen
close to the OH group, that with sulfur syn to the OH group
21
Ϫ1
22
(
2.57 ppm) predominating.‡ In DMSO only one isomer is found;
13
the similarity of the C NMR spectra, particularly the C2 and
C4 shifts, in this solvent and chloroform indicates that it is the
syn isomer, 8S.
20
Hydrogen bonding in constrained alcohols
A good IR–NMR correlation [δOH = (50 ± 2) Ϫ (0.0135 ±
Among the many methods for the detection of hydrogen bond-
0
.0006)νOH; r = 0.9945] is found for alcohols 2–8 with, how-
ever, marked dispersion at the low shift/high frequency end
weighted means were taken for the IR values). This means
1
1
ing, either inter- or intramolecular, the best established is
12
doubtless IR spectroscopy but in certain areas of chemistry
(
1
13
H NMR spectroscopy is increasingly used either alone or in
that, while there is clear evidence for hydrogen bonding in the
nitrogen-containing compounds, for the derivatives where
sulfur or oxygen would be involved the situation is much less
clear, the IR absorption being virtually indistinguishable from
that of phenyldiadamantylmethanol, 2. The NMR data, but
not the IR, argue in favour of very weak H-bonding in 7S and
conjunction with IR. Hydrogen bonding is typically associated
with IR absorptions at lower-than-usual frequencies and
higher-than-usual NMR shifts. Linear correlations between
these two types of data have been reported.
Benzyl alcohol and many related structures show two IR
absorptions in the OH stretching region separated by 10–30
1
4
8
S, but not in 6S. Such small deshielding effects could, however,
Ϫ1
cm , that at the lower frequency being attributed to a species
have causes other than hydrogen bonding. It should be noted
that the anti isomer of the 2-thienyl derivative, 7A, is associated
with a proton signal at 2.17 ppm (that of 7S is at 2.40 ppm) but
it is impossible to locate the corresponding IR absorption.
Generally speaking, although changing the solvent from
chloroform to benzene modifies considerably the aromatic
proton spectrum the effect on the hydroxy proton shift is small,
there being a modest upfield or downfield change (0.0–0.4 ppm)
(Table 1). Going to DMSO, on the other hand, has marked
12
in which there is hydrogen bonding to the π-electron system.
A relationship was established between this frequency and the
15
C –C –C–O torsion angle, orthogonality with the benzene
ar
ar
16
ring being clearly advantageous. Nevertheless, there are some
‡
Since S has priority over N the 2-thiazolyl isomer with OH close
to nitrogen is denoted anti, whereas this situation for the 2-pyridyl
derivatives corresponds to the syn isomer, N having priority over C.
1
0
2
648
J. Chem. Soc., Perkin Trans. 2, 1998, 2647–2652

View Article Online
Table 1 Solvent dependence of hydroxy proton NMR chemical shifts (in ppm) in heteroaryldi(1-adamantyl)methanols; temperature coefficient
(
25–55 ЊC) of chemical shift in chloroform (ca. 0.05 M; in ppb/ЊC); OH stretching frequencies in carbon tetrachloride (ca. 0.01 M)
a
a
b
Compound
δ(CDCl3)
Ϫ∆δ/∆T
δ(C D )
δ(C D N)
Ϫ∆δ/∆T
δ(DMSO-d6)
νOH
6
6
5
5
2
1.97
1.85
2.40
2.17
2.57
5.22
6.51
0.13 ± 0.02
0.47 ± 0.04
0.40 ± 0.01
0.24 ± 0.01
0.57 ± 0.04
1.86 ± 0.08
2.68 ± 0.10
1.66
1.93
2.14
4.93
5.17
5.67
5.27
6.18
5.66
6.74
4.52
8.30
4.94
21.5 ± 0.1
21.5 ± 0.7
20.7 ± 0.4
20.7 ± 0.3
19.9 ± 0.3
3.0 ± 0.1
1.7 ± 0.1
20.5 ± 0.1
1.9 ± 0.1
19.6 ± 0.1
3.81
3.87
4.31
4.10
4.99
3636, 3610
3620, 3608
3624, 3608
c
6
7
7
8
8
4
4
5
S
S
A
S
A
S
A
S
d
e
—
—
f
f
f
2.25
5.63
6.78
1.79
8.32
3619, 3608
3429, 3378
g
f
—
6.30
3.93
7.90
3313
3637
3165
d
—
8.14
3.08 ± 0.02
Water
a
b
c
d
e
2
5 ЊC. 60 ЊC. Ϫ8 to 32 ЊC. The signal fell in the range of the adamantyl protons (1.6–2.1 ppm) but could not be located unambiguously. See
f
g
text. In chloroform. The species is not detected in this solvent.
Ϫ1
Table 2 Experimental free energy differences in various solvents and theoretical gas phase heat of formation differences (kcal mol ) for heteroaryl-
diadamantylmethanols (anti Ϫ syn)
Compound
∆∆GЊ(CDCl3)
∆∆GЊ(C D )
∆∆GЊ(C D N)
∆∆GЊ(DMSO-d6)
>2.7
0.8
>2.7
1.4
∆∆H (AM1)
∆∆H (PM3)
6
6
5
5
f
f
a
6
1.0
1.2
0.4
1.8
1.3
1.0
0.5
1.7
>2.7
1.3
1.9
Ϫ0.1
—
0.2
Ϫ1.7
—
a
7
Ϫ2.0
Ϫ2.1
Ϫ0.7
a
8
b
4
1.5
a
b
At 25 ЊC. At 150 ЊC.
effects on the phenyl, syn-2-furanyl, syn-2-thienyl, syn-2-
thiazolyl and anti-2-pyridyl derivatives (2, 6S, 7S, 8S and 4A),
the downfield displacement being 1.9–2.4 ppm. Changes of this
magnitude are consistent with intermolecular bonding of the
hydroxy hydrogen to the oxygen of DMSO. However, the two
syn-pyridyl derivatives, 4S and 5S, move slightly upfield. This
suggests that in these two alcohols the OH hydrogen is so
strongly intramolecularly bonded to nitrogen that bonding with
the solvent presents no thermodynamic advantage. In the case
of the thiazolyl derivative the hydrogen bond in the anti isomer,
Small amide ∆δ/∆T values in non-hydrogen-bonding solvents
have been taken to mean that the proton is entirely free of
hydrogen bonding or completely locked in an intramolecular
hydrogen bond, though Gellmann has shown that small values
are not incompatible with equilibrium between intramolec-
23a
ularly hydrogen-bonded and non-hydrogen-bonded states.
Given that these hypotheses appear to cover the complete range
of possibilities, the interpretation of small ∆δ/∆T values in such
solvents would appear to be a hopeless task. For the alcohols
examined in the present work, in chloroform the temperature
coefficients range from Ϫ0.1 ppb/ЊC for phenyldiadamantyl-
methanol, 2, to Ϫ3.1 ppb/ЊC for the 2-(3-methylpyridyl) deriv-
ative, 5S. These values are all small, regardless of whether the
hydroxy hydrogen is H-bonded or not, and there is an approxi-
mately linear dependence on the shift at 25 ЊC (Table 1) [∆δ/
∆T = (0.68 ± 0.12) Ϫ (0.48 ± 0.03)δOH; r = 0.9906]. In this case,
therefore, all that can be said is that ∆δ/∆T correlates with the
strength of intramolecular hydrogen bonding, insofar as it is
expressed by the NMR shift data and the IR absorption
frequencies.
8
A, is too weak to make this isomer the more stable in chloro-
form or benzene, and in DMSO it disappears completely, as
does 6A. In pyridine, chosen as a potentially powerful hydrogen
bonding solvent, the shifts of the hydroxy protons are about
3
0% higher than in DMSO, except for 4S and 5S which again
do not move significantly.
The temperature dependence of the amide proton NMR shift
in strongly hydrogen bonding solvents, usually DMSO, has
1
3
been used as a criterion of hydrogen bonding in peptides,
23
24
model amides and other compounds. Small values (Ϫ∆δ/
∆T ≤ 3 ppb/ЊC) are associated with protons involved in intra-
molecular hydrogen bonds or otherwise shielded from the
medium, while exposed hydrogens exhibit larger values (Ϫ∆δ/
Molecular mechanics and semi-empirical quantum chemical
calculations
∆T > 4 ppb/ЊC). To the best of our knowledge, there has been
no work on N ؒ ؒ ؒ H–O bonding.
The temperature dependence of the hydroxy proton chemical
shift in pyridine, measured from 25 to 50 ЊC on approximately
No detailed structural information was available for any of
these compounds, but it was interesting to see to what extent the
relative stabilities within the various rotamer pairs were repro-
duced by molecular mechanics and/or semi-empirical AM1 or
PM3 calculations (Table 2). The most commonly used force
field, MMP2, has not been parametrized for substituted five-
membered heteroaromatics, and even for pyridine derivatives
two torsion angles have to be parametrized ad hoc. Both types
of calculation gave well defined minima corresponding to the
two rotamers, with the C–O bond generally within 10–12Њ of
the ring plane.
0
.05 M solutions (shifts were not significantly concentration-
dependent), is striking. For all those alcohols (as well as water)
which are hydrogen-bonded to the solvent the coefficient, ∆δ/
25
∆T, is of the order of Ϫ20 ppb/ЊC, while for the two pyridyl
derivatives the value is barely a tenth of this, Ϫ1.7 and Ϫ1.9
ppb/ЊC for 4S and 5S, respectively. Here we would appear to
have a clear-cut distinction between intramolecular and inter-
molecular hydrogen-bonded species. Only one structure, anti-
(
2-thiazolyl)diadamantylmethanol, 8A, shows what appears to
According to MMP2 calculations on (2-pyridyl)diadamant-
ylmethanol, 4, the anti isomer is about 1.2 kcal mol less
Ϫ1
be an anomalous value, Ϫ3.0 ppb/ЊC. A reasonable interpret-
ation would be that there is an equilibrium between intra- and
intermolecular H-bonded species of this rotamer, and that the
greater temperature dependence is due in part to a change in the
equilibrium constant, as has been observed in certain diamides
strained than the syn. This is consistent with the intuitive idea
that the nitrogen atom is less bulky than the cyclic CH, and can
be taken as an indication of the relative stabilities in the absence
of non-steric interactions. The fact that the syn is in reality the
23
Ϫ1
and triamides.
more stable, by 1.4–1.8 kcal mol (depending on the solvent,
J. Chem. Soc., Perkin Trans. 2, 1998, 2647–2652
2649

View Article Online
Table 2) indicates that there is a stabilizing interaction, equiv-
alent to 2.6–3.0 kcal mol . However, the solvent-dependence
7S appears to stabilize both isomers, but by very different
Ϫ1
Ϫ1
amounts, 0.4 and 2.3 kcal mol , respectively. This reduces the
Ϫ1
of the equilibrium shows that this quantity cannot be entirely
attributed to hydrogen bonding but also includes solvation
energy differences. For the 3-methyl-substituted derivative, 5S,
the steric energy difference is substantially greater, 7.6 kcal
difference to 0.1 kcal mol , still in favour of 7A, but tends to
run counter to the observation that 7A is enhanced in DMSO.
However, the opposed effects of pyridine and DMSO serve to
underline the specificity of solvation phenomena and suggest
that this treatment of solvent effects is inadequate to explain
such relatively small free energy differences. In conclusion, the
observation that isomer ratios are sensitive to the NMR solvent
used clearly indicates that there are differential solvation
effects upon the relative stabilities of the isomers, and it is
quite likely that the solvation energies involved are of the same
order of magnitude as the differences in the gas-phase heats
of formation (albeit with their own degree of approximation)
calculated by the semi-empirical methods.
Ϫ1
mol in favour of the isomer with nitrogen anti to the OH
group. This is slightly more than the difference between the
corresponding o-tolyldiadamantylmethanols, 3A and 3S (7.0
Ϫ1
kcal mol ). Since 5S decomposes upon heating rather than
rotates, and also fails to rotate in the presence of an organo-
2b
lithium reagent, it is not possible to estimate either the equi-
librium constant or the contribution of hydrogen bonding to
the stability of this isomer. Calculated distances between the N
and O atoms in 4S and 5S are 2.55 and 2.45 Å, respectively;
between N and H, 2.18 and 2.09 Å, respectively. The distances
between the heteroatoms are well within the limit, the sum of
their van der Waals radii (3.07 Å), below which a hydrogen
Conclusions
1
1,26
The NMR spectra of most of the tertiary alcohols synthesized
here (not the pyridyl derivatives) indicate that they exist in two
rotameric forms, which interconvert rapidly on the laboratory
time-scale and show a significant thermodynamic preference for
the syn isomer, that in which the heteroatom and the OH group
are closest. The IR and NMR spectra of the 2-pyridyl deriv-
atives, which are synthesized as the syn isomers, and one isomer
of the 2-thiazolyl derivative show that intramolecular hydrogen
bonding is important; the data for the other alcohols indicate
little or no hydrogen bonding. In the 2-thiazolyl derivative, the
hydrogen bonding in the anti isomer is not strong enough to
make it the more stable species, though semi-empirical calcu-
lations suggest that it should be, even without hydrogen bond-
ing. Semi-empirical calculations systematically give the anti
alcohol as the more stable in the gas phase. This disagreement
with the experimental (solution) results may be due to deficien-
cies in the parametrization or to the neglect of solvation. The
solvent-dependent variations in the isomer ratios show that the
solvation energies of the two rotamers of a given alcohol can
bond can occur,
far from ideal.
though the N ؒ ؒ ؒ H ؒ ؒ ؒ O bond angles are
27
AM1 calculations on 4A and 4S find a difference in the
heats of formation similar to that of the MMP2 steric energies
and in the same direction, 4A being the more stable by 0.7 kcal
mol , making the possible contribution of hydrogen bonding
and differential solvation 2.1–2.5 kcal mol . The geometry,
Ϫ1
Ϫ1
however, is rather different, with the N ؒ ؒ ؒ O and N ؒ ؒ ؒ H dis-
tances 2.70 and 2.47 Å, respectively. For 5, where the energy
difference is dominated by steric interactions between the ortho-
methyl group and the adamantyls, the anti rotamer is calculated
Ϫ1
to be 9.0 kcal mol more stable than the syn; the N ؒ ؒ ؒ O and
N ؒ ؒ ؒ H distances in 5S are 2.58 and 2.35 Å, respectively. Des-
pite the marked differences between the MMP2 and semi-
empirical calculations, the N ؒ ؒ ؒ O distance is clearly well inside
the range required for hydrogen bonding. The effect of the
ortho-methyl group is simply to push the two heteroatoms
slightly closer.
AM1 calculations on the gas-phase stabilities of the other
alcohol pairs generally predict the anti isomer to be the more
Ϫ1
differ by at least 2 kcal mol .
Ϫ1
Ϫ1
stable, by 0.1 kcal mol for 6A, 2.0 kcal mol for 7A and 2.1
Ϫ1
kcal mol for 8A, in contradiction with the solution results. On
Experimental
General methods
the other hand, PM3, which is better parametrized for sulfur,
Ϫ1
makes 7S 0.2 kcal mol more stable than 7A, but 8S still 1.7
Ϫ1
kcal mol less stable than 8A. 8A is calculated (AM1, PM3) to
NMR measurements were performed on a Bruker AS 200
FT instrument operating at 200 MHz (proton) or 50 MHz
(carbon). Chemical shifts are given in ppm and J values in Hz.
Unless otherwise noted all measurements were made in hexa-
deuteriobenzene, deuteriochloroform, pentadeuteriopyridine
or hexadeuteriodimethyl sulfoxide at 25 ЊC (internal scaling/
TMS: δ = 7.16, 7.26, 8.71 and 2.50; δ = 128.0, 77.0, 149.9 and
have N ؒ ؒ ؒ O and N ؒ ؒ ؒ H distances of 2.77, 2.70 and 2.56, 2.54
Å, respectively, obviously slightly greater than for the six-
membered pyridine heterocycles but still well within the limit.
In none of these calculations is there any explicit indication of
hydrogen bonding, such as lengthening of the O–H bond, for
example.
H
C
If we consider the “best” values for the gas-phase energy
differences, in chloroform and benzene the calculations err in
favour of the anti isomer by about 1 kcal mol (Table 2). The
free energy difference in favour of the syn isomer is rather
greater for DMSO and pyridine; since 6A is not seen at all in
39.5). Carbon and hydrogen shifts of the heterocyclic system
are numbered: C2, C3, etc. Generally, the proton signals were
Ϫ1
8
assigned on the basis of coupling constants and spectrum
9
simulation by the gNMR program (Cherwell Scientific). The
13
corresponding C signals were identified by 2D heteronuclear
correlation experiments using the XHCORR sequence.
Samples for NOE experiments were solutions in deuterio-
chloroform degassed by several freeze–pump–thaw cycles
before sealing under vacuum. Measurements were made on a
Bruker AM-500 spectrometer at 500 MHz using the NOE-
MULT pulse sequence. NOE difference spectra were obtained
by subtraction of the off-resonance control FID (16 K) from
the on-resonance FID. The signal of interest was selectively
saturated for 4 s. A cycle of 16 cycles of 16 scans was chosen
with a relaxation delay of 2 s between each irradiation. Free
induction decays were processed using exponential multiplic-
ation with a line broadening of 3 Hz before Fourier transform-
ation. IR spectra were measured in carbon tetrachloride,
chloroform or KBr on a Nicolet 60SX FTIR spectrometer with
these solvents, we can only say that ∆∆GЊ is greater than about
Ϫ1
2
.7 kcal mol , assuming (somewhat optimistically, perhaps)
that 1% of 6A could have been detected. In the case of 8 the
anti isomer is absent in DMSO and just detectable (ca. 4%) in
pyridine, making the free energy difference >2.7 and 1.9 kcal
Ϫ1
mol , respectively. There is no simple way of rationalizing the
fact that, in contrast to the behaviour of the closely related
compounds 6 and 8, where the stability of the syn isomer is
enhanced by hydrogen-bonding solvents, the level of 7A is
slightly greater in DMSO than in the other solvents examined.
Preliminary calculations on 6 and 7 using the AM1/SM2
model for hypothetical solvation by water, taken as a polar
solvent, give rather surprising results. According to this model
both 6A and 6S are destabilized by solvation, to the extent of
Ϫ1
Ϫ1
1
.3 and 0.9 kcal mol , respectively. This now makes the syn
2 cm resolution. Lorentzian deconvolution was used to locate
isomer the more stable. On the contrary, solvation of 7A and
shoulders and to resolve broad absorptions. Gas chrom-
2
650 J. Chem. Soc., Perkin Trans. 2, 1998, 2647–2652

View Article Online
atography was performed on a 30 cm 10% SE30 on Chrompack
column. Column chromatography was performed on silica gel
alcohol was unchanged by treatment with tert-butyllithium in
pentane for 1 h.
6
0 (Merck) in light petroleum (boiling range 35–60 ЊC)–
3
dichloromethane mixtures. Melting points were determined
in capillary glass tubes on a Mettler FP5 instrument with a
heating rate of 3 ЊC min .
(2-Furanyl)di(1-adamantyl)methanol, 6. Furan (0.36 cm , 5
mmol) was lithiated by treatment with n-butyllithium (1.6 M in
Ϫ1
3
hexanes, 3.1 cm , 5 mmol) in the presence of TMEDA (0.68
cm , 5 mol) in diethyl ether (10 cm ) under argon at room tem-
perature. After 30 min a solution of di(1-adamantyl) ketone
0.15 g, 0.5 mmol) in diethyl ether (20 cm ) was added in about
10 min. After 1 h the mixture was quenched with water, the
organic phase washed with water, then dried over MgSO and
the solvent evaporated to yield a light brown solid which was
purified by silica gel chromatography and crystallization from
hexane (0.153 g, 83%; mp 178 ЊC); by NOE the major constit-
uent was identified as 6S: νOH/cm (CCl ) 3620, 3608sh; δ
chloroform) 29.4 (6 CH), 37.1 (6 CH ), 38.7 (6 CH ), 44.5 (2
2 2
C ), 82.6 (COH), 107.9 (C3), 109.2 (C4), 140.1 (C5) and 160.0
C2); δH (chloroform, Ϫ8 ЊC) 1.6–2.1 (br m, Ad), 1.84 (OH),
.17 (H3, J 1.0 and 3.2), 6.32 (H4, J 1.8 and 3.2) and 7.38 (H5,
J 1.0 and 1.8). A minor component, ca. 15% of the total, gave
H NMR peaks, located in part by gNMR simulation, at 6.15
H3), 6.42 (H4) and 7.36 (H5) ppm and C signals at 105.5,
3
3
Synthesis of heteroaromatic di(1-adamantyl)methanols
3
(
(
2-Pyridyl)di(1-adamantyl)methanol, 4. 2-Bromopyridine
3
(
1 cm , 10.3 mmol) was stirred in sodium–dry diethyl ether
3
(
10 cm ) under argon at Ϫ75 ЊC. A solution of n-butyllithium in
4
3
cyclohexane (2 M, 5 cm , 10 mmol) was added dropwise in
about 5 min. After stirring for 30 min, a solution of di(1-
adamantyl) ketone (0.5 g, 1.7 mmol) in diethyl ether (40 cm )
was added in about 15 min, the cooling bath removed and the
mixture allowed to warm to about 0 ЊC. It was then quenched
with water, and the organic product was extracted with a mix-
ture of hexane and dichloromethane and twice rinsed with
3
Ϫ1
4
C
(
q
(
6
water. After drying (MgSO ) the volatile solvents were largely
4
evaporated at reduced pressure until solid started to appear. At
this point the solution was refrigerated to complete crystalliz-
ation. Filtration and washing with cold hexane gave a light
brown product, identified by NOE, νOH and the chemical shift
of the OH proton as 4S, which was further purified by column
1
13
(
1
10.8 and 138.5 ppm. These were attributed to the anti isomer,
A; this isomer was not found in DMSO at 25 or 60 ЊC (Found:
6
Ϫ1
C, 81.7; H, 9.3. C H O requires C, 81.92; H, 9.35%).
chromatography (0.34 g, 54%): mp 215 ЊC; νOH/cm (CCl4)
25 34 2
3
3
1
313, (KBr) 3277; δ (chloroform) 29.2 (6 CH), 37.1 (6 CH ),
C
2
(
2-Thienyl)di(1-adamantyl)methanol, 7. 2-Bromothiophene
3
9.3 (6 CH ), 44.5 (2 C ), 82.2 (COH), 121.8 (C5), 124.0 (C3),
2
q
(
0.95 cm , 9.9 mmol) was stirred in sodium–dry diethyl ether (10
34.4 (C4), 145.7 (C6) and 161.0 (C2); δ (chloroform) 1.58 (br
H
3
cm ) under argon at 0 ЊC. A solution of n-butyllithium in cyclo-
s, Ad), 1.6–2.0 (br m, Ad), 6.51 (OH), 7.21 (H5, J 1.0, 5.5 and
.6), 7.58 (H3, J 1.0, 1.0 and 8.0), 7.65 (H4, J 1.3, 7.6 and 8.0)
3
hexane (2 M, 5 cm , 10 mmol) was added dropwise in about 5
min. After stirring for 30 min, a solution of di(1-adamantyl)
ketone (0.50 g, 1.7 mmol) in diethyl ether (30 cm ) was added in
about 15 min. After a further 30 min the reaction mixture was
quenched with water, and the organic phase was washed with
water, then dried over MgSO and the solvent evaporated. Crys-
tallization from n-hexane gave a solid (0.50 g, 78%; mp 205 ЊC)
apparently containing two products in a ratio of 7:1. Major
product, identified by NOE, 7S: νOH/cm (CCl ) 3624, 3608sh;
7
and 8.49 (H6, J 1.0, 1.3 and 5.5) (Found: C, 82.8; H, 9.4; N, 3.6.
C H NO requires C, 82.71; H, 9.34; N, 3.71%).
3
26
35
Heating a solution of the alcohol in CDCl in a sealed tube
3
for 2 h at 150 ЊC gave about 10% of an isomeric material, pre-
sumably the anti rotamer, 4A, revealed by its IR spectrum and
4
13
C NMR (chloroform) peaks at 39.1 (CH ), 44.9 (C ), 120.7
2
q
(
CH), 122.7 (CH), 134.8 (CH) and 146.3 (CH) ppm; signals
Ϫ1
corresponding to two of the aromatic protons were detected
at ca. 7.05 and 8.6 ppm; νOH/cm (CCl ) 3637, (KBr) 3624.
Replacing the solvent by benzene for NMR analysis greatly
improved the separation of two of the aromatic proton signals.
4
Ϫ1
δ (chloroform) 29.0 (6 CH), 36.9 (6 CH ), 38.8 (6 CH ), 45.0 (2
C
2
2
4
C ), 84.7 (COH), 121.6 (C5), 122.5 (C3), 126.3 (C4) and 151.4
q
(
(
(
4
3
1
C2); δH (chloroform) 1.60 (br s, Ad), 1.7–2.1 (br m, Ad), 2.40
OH), 6.92 (H3, J 1.0 and 3.6), 7.01 (H4, J 3.6 and 5.1) and 7.16
H5, J 1.0 and 5.1). In DMSO at 60 ЊC the 7S/7A isomer ratio is
:1. Major product, 7S: δ (DMSO) 28.3 (6 CH), 36.3 (6 CH ),
8.6 (6 CH ), 44.2 (2 C ), 83.9 (COH), 121.6 (C5), 121.6 (C3),
4
S: δ 6.60 (H5), 7.05 (H4), 7.33 (H3) and 8.20 (H6); 4A: δ 6.65
H
H
(
H5), 7.21 (H4), 7.74 (H3) and 8.51 (H6). The same isomeriz-
ation experiment run in benzene, pyridine and DMSO gave 4A
to the extent of 11, 14 and 17%, respectively.
C
2
2
q
26.7 (C4) and 152.1 (C2); δ (DMSO) 1.56 (br s, Ad), 1.7–2.1
H
[
2-(3-Methylpyridyl)]di(1-adamantyl)methanol, 5. 2-Bromo-
(br m, Ad), 4.31 (br, OH), 6.90 (H3, J 1.0 and 3.6), 6.97 (H4,
J 3.6 and 5.2) and 7.20 (H5, J 1.0 and 5.2); minor product, 7A:
δ (DMSO) 28.4 (6 CH), 36.4 (6 CH ), 38.0 (6 CH ), 43.6 (2 C ),
3
3
-methylpyridine (0.25 cm , 2.2 mmol) was stirred in sodium–
dry diethyl ether (10 cm ) under argon at Ϫ75 ЊC. A solution of
tert-butyllithium in pentane (1.7 M, 2.5 cm , 4.2 mmol) was
3
C
2
2
q
3
84.6 (COH), 123.3 (CH), 123.7 (CH), 126.6 (CH) and 146.5
added dropwise in about 2 min. After stirring for 30 min at
(C2); δ (DMSO) 1.56 (br s, Ad), 1.7–2.1 (br m, Ad), ca. 4.1 (br,
H
the same temperature, a solution of di(1-adamantyl) ketone
OH), 6.93 (H4, J 3.6 and 5.1), 7.06 (H3, J 1.4 and 3.6) and 7.31
(H5, J 1.4 and 5.1) (Found: C, 78.6; H, 9.1; S, 8.3. C H OS
3
(
0.15 g, 0.5 mmol) in diethyl ether (20 cm ) was added in about
25
34
1
5 min. The reaction mixture was allowed to rise slowly to room
requires C, 78.48; H, 8.96; S, 8.38%).
temperature over a period of about 2 h, then quenched with
water. The organic phase was washed with water, then dried
(2-Thiazolyl)di(1-adamantyl)methanol, 8. 2-Bromothiazole
3
over MgSO and the solvent evaporated under vacuum to give a
(0.5 cm , 5.6 mmol) was stirred in sodium–dry diethyl ether (10
4
3
yellowish paste. Chromatography on silica gel gave a white solid
cm ) under argon at Ϫ75 ЊC. A solution of n-butyllithium
3
(
179 mg, 91%) consisting of 5S (as shown by NOE, ν and δH
in hexanes (1.6 M, 3 cm , 4.8 mmol) was added dropwise.
OH
Ϫ1
of the OH proton): mp 131 ЊC, decomp.; νOH/cm (CCl ) 3165;
After stirring for 1 h at the same temperature, a solution of
di(1-adamantyl) ketone (0.15 g, 0.5 mmol) in diethyl ether (20
4
δ (chloroform) 26.4 (Me), 29.3 (6 CH), 37.1 (6 CH ), 39.4 (6
C
2
3
CH ), 45.4 (2 C ), 86.5 (COH), 121.4 (C5), 132.4 (C3), 141.6
cm ) was added in about 10 min. The reaction mixture was
2
q
(
C4), 142.2 (C6) and 160.7 (C2); δ (chloroform) 1.5–2.1 (br m,
allowed to warm slowly to room temperature and was left over-
night. It was then quenched with water and a mixture of light
petroleum and dichloromethane, an abundant black precipitate
filtered off, the organic phase washed with water, then dried
H
Ad), 2.74 (Me), 7.12 (H5, J 4.6 and 7.5), 7.52 (H4, J 0.6, 1.75
and 7.5), 8.14 (OH) and 8.37 (H6, J 0.6, 1.75 and 4.6); homo-
nuclear decoupling experiments indicate that H4 and H6 are
coupled with the methyl group (Found: C, 83.0; H, 9.6; N, 3.4.
C H NO requires C, 82.81; H, 9.52; N, 3.58%).
over MgSO and the solvent and other volatiles evaporated.
4
1
After chromatography H NMR showed the product to consist
27
37
Ϫ1
Attempts to convert this material to the anti isomer by heat-
of a 2:1 mixture [167 mg, 87%; mp 199 ЊC; νOH/cm (CHCl3)
ing resulted only in decompostion to diadamantyl ketone. The
3619, 3608sh and 3429, 3378, (KBr) 3588, 3581] of isomeric
J. Chem. Soc., Perkin Trans. 2, 1998, 2647–2652
2651

View Article Online
alcohols. Major product, syn with respect to sulfur, 8S: δC
chloroform) 29.0 (6 CH), 36.9 (6 CH ), 38.7 (6 CH ), 45.1 (2
4 S. Nomura and T. Hidaka, Jap P 02,300,171 [90,300,171] (Chem.
Abstr., 1990, 114, 185292k); S. Nomura and T. Hidaka, Jap P
(
2
2
0
3,56,464 [91,56,464] (Chem. Abstr., 1991, 115, 158974y).
C ), 85.2 (COH), 117.5 (C5), 141.6 (C4) and 178.3 (C2); δH
q
5
G. Van Zyl, R. J. Langenberg, H. H. Tan and R. N. Schut, J. Am.
Chem. Soc., 1956, 78, 1955.
(
chloroform) 1.5–2.2 (br m, Ad), 2.57 (OH), 7.21 (H5, J 3.3)
and 7.78 (H4, J 3.3); δ (DMSO) 28.4 (6 CH), 36.6 (6 CH ),
C
2
6 A. Dondoni, G. Fantin, M. Fogagnolo, A. Medici and P. Pedrini,
J. Org. Chem., 1988, 53, 1748; A. Dondoni and D. Perrone, J. Org.
Chem., 1995, 60, 4749.
3
8.0 (6 CH ), 44.3 (2 C ), 84.5 (COH), 118.3 (C5), 141.3 (C4)
2 q
and 178.8 (C2); δ (DMSO) 1.5–2.2 (br m, Ad), 4.99 (OH), 7.47
H
7
D. J. Chadwick and C. Willbe, J. Chem. Soc., Perkin Trans. 1, 1977,
87.
(
H5, J 3.3) and 7.74 (H4, J 3.3); minor product, 8A: δ (chloro-
C
8
form) 29.0 (6 CH), 37.0 (6 CH ), 38.6 (6 CH ), 44.4 (2 C ), 85.3
2
2
q
8
E. Pretsch, W. Simon, J. Seibl and T. Clerc, Tables of Spectral Data
for Structure Determination of Organic Compounds, Springer-
Verlag, Berlin-Heidelberg, 2nd edn., 1989.
(
COH), 119.2 (C5), 137.9 (C4) and 172.8 (C2); δ (chloroform)
H
1
.5–2.2 (br m, Ad), 5.22 (OH), 7.33 (H5, J 3.2) and 7.70 (H4,
J 3.2). 8A was not found in DMSO (Found: C, 74.6; H, 8.6; N,
9 gNMR from Cherwell Scientific Publishing Limited, The Magdalen
Centre, Oxford Science Park, Oxford OX4 4GA.
0 R. S. Cahn, C. K. Ingold and V. Prelog, Angew. Chem., 1966, 78,
413.
1 M. D. Joesten and L. J. Schaad, Hydrogen Bonding, Marcel Dekker,
New York, 1974; F. Hibbert and J. Emsley, Adv. Phys. Org. Chem.,
3
.9; S, 8.1. C H ONS requires C, 75.15; H, 8.67; N, 3.65; S,
24 33
1
1
8
.36%).
Alcohol 8 (65 mg) in chloroform (1.5 cm ) was heated in a
3
sealed tube at 150 ЊC for 5 h. Evaporation of the solvent fol-
lowed by chromatography on alumina in light petroleum–
diethyl ether gave di(1-adamantyl) ketone (42 mg, 83%).
1
990, 26, 255.
12 H. S. Aaron, Top. Stereochem., 1979, 11, 1; M. C. R. Symons, Chem.
Soc. Rev., 1983, 12, 1.
1
13 H. Kessler, Angew. Chem., Int. Ed. Engl., 1982, 21, 512. See also:
M. P. Williamson and J. P. Waltho, Chem. Soc. Rev., 1992, 21, 227.
14 L. W. Reeves, E. A. Allan and K. O. Strømme, Can. J. Chem., 1960,
Temperature coefficient of the H NMR chemical shift
Approximately 0.05 M solutions of the various alcohols in
deuteriochloroform or pentadeuteriopyridine (chosen in pref-
erence to DMSO because of the low solubility of the alcohols
in the latter) were examined at 5 ЊC intervals in the 25–50 ЊC
range (Table 1). The temperature variation of the hydroxy
proton shift is based on the assumption that the solvent refer-
ence is constant. Values of ∆δ/∆T are expressed in ppb/ЊC.
3
8, 1249; I. Gränacher, Helv. Phys. Acta, 1961, 34, 272; J. R. Merrill,
J. Phys. Chem., 1961, 65, 2023; G. O. Dudek, J. Org. Chem., 1965,
30, 548; A. D. Sherry and K. F. Purcell, J. Phys. Chem., 1970, 74,
535.
3
1
5 H. Iwamura, Tetrahedron Lett., 1970, 2227.
16 H. S. Im, E. R. Bernstein, H. V. Secor and J. I. Seeman, J. Am. Chem.
Soc., 1991, 113, 4422.
1
7 F. H. Hon, H. Matsumura, H. Tanida and T. T. Tidwell, J. Org.
Chem., 1972, 37, 1778; J. S. Lomas, P. K. Luong and J. E. Dubois,
J. Org. Chem., 1977, 42, 3394.
Molecular mechanics and semi-empirical quantum mechanical
calculations
1
8 J. S. Lomas and J. Vaissermann, Bull. Soc. Chim. Fr., 1996, 133, 25;
J. Vaissermann and J. S. Lomas, Acta Crystallogr., Sect. C, 1997, 53,
25
To handle pyridyl derivatives with MMP2(85), parameters
for two torsion angles (types 1-1-2-37 and 6-1-2-37) have to be
1
341.
1
2
9 L. P. Kuhn, R. A. Wires, W. Ruoff and H. Kwart, J. Am. Chem. Soc.,
Ϫ1
supplied. The steric energies (kcal mol ) are based on the
1
969, 91, 4790.
assumption that they can be treated as types 1-1-2-2 and 6-1-
0 J. N. Brown, R. M. Jenevein, J. H. Stocker and L. M. Trefonas,
2
-2, respectively. 4A 58.6; 4S 59.8; 5A 61.8; 5S 69.4.
The Spartan package with AM1 and PM3 (for sulfur-
J. Org. Chem., 1972, 37, 3712.
27
21 S. E. Bojadziev, D. T. Tsankov, P. M. Ivanov and N. D. Berova, Bull.
Chem. Soc. Jpn., 1987, 60, 2651.
containing species) was used for semi-empirical calculations.
Preliminary calculations with the AM1/SM2 model for solv-
2
2
2 L. Re, B. Maurer and G. Ohloff, Helv. Chim. Acta, 1973, 56, 1882.
3 (a) S. H. Gellmann, G. P. Dado, G. B. Liang and B. R. Adams,
J. Am. Chem. Soc., 1991, 113, 1164; (b) G. P. Dado and S. H.
Gellmann, J. Am. Chem. Soc., 1993, 115, 4228; (c) G. P. Dado and
S. H. Gellmann, J. Am. Chem. Soc., 1994, 116, 1054; (d) B. W. Gung
and Z. Zhu, J. Org. Chem., 1996, 61, 6482; (e) B. W. Gung and
Z. Zhu, J. Org. Chem., 1997, 62, 2324; ( f ) B. W. Gung, Z. Zhu and
B. Everingham, J. Org. Chem., 1997, 62, 3436.
ation by water were run on 6 and 7. The heats of formation
Ϫ1
(
∆∆H /kcal mol ) listed are those for the lowest-energy con-
f
formations within the different conformers: 4A (AM1) Ϫ58.2;
S Ϫ57.5; 5A (AM1) Ϫ60.1; 5S Ϫ51.1; 6A (AM1) Ϫ91.6; 6S
Ϫ91.5; 7A (AM1, PM3) Ϫ66.2, Ϫ48.4; 7S Ϫ64.2, Ϫ48.6; 8A
AM1, PM3) Ϫ54.3, Ϫ38.4; 8S Ϫ52.2, Ϫ36.7. AM1/SM2
4
(
2
4 Cavitands: D. M. Rudkevich, G. Hilmersson and J. Rebek, J. Am.
Chem. Soc., 1997, 119, 9911. Carbohydrates: M. López de la Paz,
J. Jiménez-Barbaro and C. Vicent, J. Chem. Soc., Chem. Commun.,
results: 6A Ϫ90.2; 6S Ϫ90.6; 7A Ϫ66.6; 7S Ϫ66.4.
1
998, 465; M. López de la Paz, G. Ellis, S. Penadés and C. Vicent,
Acknowledgements
We are indebted to Ms Claire Mangeney for helpful discussions.
Tetrahedron Lett., 1997, 38, 1659; C. Landersjö, R. Stenutz and
G. Widmalm, J. Am. Chem. Soc., 1997, 119, 8695.
5 N. L. Allinger, Quantum Chemistry Program Exchange, Program
MMP2(85), Indiana University, USA.
2
References
26 R. Taylor and O. Kennard, Acc. Chem. Res., 1984, 17, 320.
2
7 PC Spartan from Wavefunction, Inc., 18401 Von Karman Avenue,
Suite 370, Irvine, CA 92612, USA. AM1: M. J. S. Dewar, E. G.
Zoebisch, E. F. Healy and J. J. P. Stewart, J. Am. Chem. Soc., 1985,
1
J. S. Lomas and J. Vaissermann, J. Chem. Soc., Perkin Trans. 2, 1998,
777.
1
2
(a) J. S. Lomas and J. E. Dubois, Tetrahedron, 1981, 37, 2273; (b) J. S.
Lomas and V. Bru-Capdeville, J. Chem. Soc., Perkin Trans. 2, 1994,
1
07, 3902. PM3: J. J. P. Stewart, J. Comput. Chem., 1989, 10, 209.
4
59.
3
J. A. Lepoivre, Janssen Chim. Acta, 1991, 9, 20.
Paper 8/06485G
2
652
J. Chem. Soc., Perkin Trans. 2, 1998, 2647–2652