The conformational composition of 3-buten-1-ol, the importance of intramolecular hydrogen bonding

Article abstract of DOI:10.1016/S0022-2860(98)00366-4

The conformations of 3-buten-1-ol (1), and its model compounds cis-6- methyl-3-cyclohexen-1-ol (6), 3-cyclopenten-1-ol (7) and epicholesterol (9) have been investigated by FT-IR and ¹H NMR spectroscopy. The energies and geometries of 1, 6 and 7 were also investigated by molecular mechanics, semiempirical molecular orbital and ab initio calculations, while 9 was investigated by molecular mechanics only. The objective of the work was to study the conformational composition and importance of intramolecular OH...π hydrogen bonding for this composition in 1. Only two conformers of 1 have a geometrical possibility for intramolecular hydrogen bonding: Conformers 12 and 13 (Fig. 1). Compounds 6 and 7 were used as models for Conformer 12, while 9 was used as a model for Conformer 13. The investigations showed that Conformer 13 is the only hydrogen-bonded conformer, and that Conformer 12 is not intramolecularly hydrogen bonded. Conformer 13 was the most populated conformer, while Conformer 12 was hardly populated. The combination of experimental and theoretical data, and the use of model compounds was found necessary to obtain this conclusion.

Full text of DOI:10.1016/S0022-2860(98)00366-4

Journal of Molecular Structure 470 (1998) 247–263
The conformational composition of 3-buten-1-ol, the importance of
intramolecular hydrogen bonding
,
a
Jan M. Bakkea *, Leif H. Bjerkeseth1
aOrganic Chemistry Laboratories, Norwegian University of Science and Technology, N-7034 Trondheim, Norway
Received 27 November 1997; revised 25 February 1998; accepted 25 February 1998
Abstract
The conformations of 3-buten-1-ol (1), and its model compounds cis-6-methyl-3-cyclohexen-1-ol (6), 3-cyclopenten-1-ol (7)
1
and epicholesterol (9) have been investigated by FT-IR and H NMR spectroscopy. The energies and geometries of 1, 6 and 7
were also investigated by molecular mechanics, semiempirical molecular orbital and ab initio calculations, while 9 was
investigated by molecular mechanics only. The objective of the work was to study the conformational composition and
importance of intramolecular OH…p hydrogen bonding for this composition in 1. Only two conformers of 1 have a geometrical
possibility for intramolecular hydrogen bonding: Conformers 12 and 13 (Fig. 1). Compounds 6 and 7 were used as models for
Conformer 12, while 9 was used as a model for Conformer 13. The investigations showed that Conformer 13 is the only
hydrogen-bonded conformer, and that Conformer 12 is not intramolecularly hydrogen bonded. Conformer 13 was the most
populated conformer, while Conformer 12 was hardly populated. The combination of experimental and theoretical data, and the
use of model compounds was found necessary to obtain this conclusion.᭧ 1998 Elsevier Science B.V. All rights reserved.
Keywords: 3-Buten-1-ol; Homoallylic alcohols; Intramolecular hydrogen bonding; Conformational composition
¯
1. Introduction
lecular hydrogen bond [8]. One year later, Oki and
Iwamura reported the same two bands, but in addition
pointed out the possibility of a third one [9]. The
presence of this band has later been supported by
other workers [10–13]. From the results of the IR
spectroscopy investigations, especially from the pre-
sence of the low frequency band, 3-buten-1-ol
appeared to have at least one conformer stabilised
by an intramolecular hydrogen bond. The geometry
of the hydrogen-bonded conformers could not be
determined by IR spectroscopy.
The conformational composition of the homoallylic
alcohol 3-buten-1-ol (1) has been the subject of a
number of investigations. The question of intramo-
lecular OH…p hydrogen bonding [1–7] has been cen-
tral to these studies. From the presence of two bands
in the OH region in the IR spectrum of 3-buten-1-ol,
one at ca. 40 cm−1 lower frequency than the other,
Schleyer et al. proposed the existence of an intramo-
In 1979, Trætteberg and Østensen reported an elec-
tron diffraction (ED) investigation of the conforma-
tions of 3-buten-1-ol. They found that in the gas
phase, the major conformer (69%) had a geometry
* Corresponding author.
1 Present address: Forsvarets Forskningsinstitutt (FFI, Norwegian
Defence Research Establishment), Division for Environmental
Toxicology, P.O. Box 25, N-2007 Kjeller, Norway.
0022-2860/98/$ - see front matter ᭧ 1998 Elsevier Science B.V. All rights reserved
PII S0022-2860(98)00366-4

248
J.M. Bakke, L.H. Bjerkeseth / Journal of Molecular Structure 470 (1998) 247–263
ꢀ Ͼ 80%† with a weak intramolecular hydrogen
bond [15].
The conformational composition has been investi-
gated by molecular mechanics calculations [16]. Nine
of the conformers have also been examined by ab
initio calculations (HF/3-21G*//HF/3-21G*, HF/6-
31G*//HF/6-31G* and MP2/6-31G*//MP2/6-31G*)
[17]. Conformer 13 was found to be the conformer
lowest in energy. However, Conformer 12 was not
considered (see Fig. 1 for notation).
We have studied the conformation of several un-
saturated alcohols and other compounds with a geo-
metrical possibility for an intramolecular hydrogen
bond, and have discussed the importance of intramo-
lecular hydrogen bonding in such compounds [18–
23]. The investigations have been performed by a
combination of IR and 1H NMR spectroscopy to
give experimental data, and by calculations at differ-
ent levels of sophistication to provide theoretical data.
We believe that the combination of experimental and
theoretical data is necessary to understand the confor-
mational behaviour of these molecules.
Fig. 1. Ball and stick models of the 14 3-buten-1-ol (1) conformers
(HF/6-31G** geometries).
2. Synthesis of model compounds
The model compound cis-6-methyl-3-cyclohexen-
1-ol (6) was obtained from 1,4-cyclohexadiene (2) in
a four-step synthesis, as shown in Fig. 2. 1,4-Cyclo-
hexadiene was first reacted with m-chloroperbenzoic
acid (MCPBA) to give 92% pure 1,4-cyclohexadiene
monoepoxide (3) in 68% yield. Further reaction with
lithium dimethylcuprate gave Ͼ 97% pure trans-6-
methyl-3-cyclohexen-1-ol (4) in 37% yield. To obtain
the desired product 6 with a cis configuration, its trans
that made intramolecular hydrogen bonding possible
(close to Conformer 13 in Fig. 1). The minor confor-
mer found did not have a geometry with staggered
C-C and C-O bonds, or eclipsed H/C-C-CyC bonds
[14]. Two years later, Marstokk and Møllendal
reported that the results from a microwave spectro-
scopy (MW) investigation agreed with the main result
from the electron diffraction investigation: Con-
former 13 was the dominant one in the gas phase
Fig. 2. Syntheses of model compounds.

249
J.M. Bakke, L.H. Bjerkeseth / Journal of Molecular Structure 470 (1998) 247–263
[33]), were distilled (kugelrohr) before use. All
these compounds were kept (neat) over activated
˚
3 A molecular sieves prior to the preparation of high
dilution samples. Cholesterol (8) and epicholesterol
(9) were used directly without purification and predry-
ing. Both were commercially available. Epicholesterol
was supplied by Steraloids, New Hampshire, USA.
1
3.2. H NMR
The 1H NMR spectra were recorded at high dilution
at room temperature in CCl3F on Jeol JNM-FX 100
FT-NMR, Jeol JNM-EX400 and Bruker WM 400
spectrometers, using a sealed capillary tube or coaxial
tube with deuterated benzene for locking purposes.
The numerical resolution was 0.1 Hz. Chemical shifts
are reported as ppm from TMS. The CCl3F solvent
˚
was distilled from freshly activated 3 A molecular
˚
sieves and kept over activated 3 A molecular sieves
in a cold room. The high dilution CCl3F solutions for
NMR (c Ͻ 5 mM) were prepared in the NMR tubes in
the cold room and were kept over freshly activated
Fig. 3. Notations and Newman projections of the rotamers of 3-
buten-1-ol (1) around the C1-O5, C1-C2 and C2-C3 bonds, and con-
formational composition from 1H NMR.
˚
3 A molecular sieves for one–three days before mea-
surement. Spin simulation and spin iteration were per-
formed with the Jeol program COMIC on a Digital
PDP-11 computer connected to the Jeol JNM-EX400
spectrometer. Spin simulations were also performed
on a PC with the program PMR from Serena Software,
Bloomington, Indiana, USA.
isomer (4) was subjected to a Mitsunobu reaction [24]
to give the benzoic acid ester of the inverted alcohol
(5). Transesterification of 5 with sodium methoxide in
methanol gave the product 6 with a purity of 99%. The
yield for the last two steps was 11%. As the aim was to
obtain pure model compounds, the yields were not
optimized.
3.2.1. 3-Buten-1-ol (1)
1H NMR (400 MHz): d 0.88 (1H, t, JH-D,E
=
3. Experimental
5.66 Hz, HH), 2.26 (2H, q, JF,G-D,E = 5.98 Hz, HF,G),
3.56 (2H, q, HD,E), ca. 5.1 (1H, m, JC-F,G = 1.20 Hz,
HC), ca. 5.1 (1H, m, JB-F,G = 1.20 Hz, HB), 5.75 (1H,
3.1. Materials
ddt, JA-B = 10.20 Hz, JA-C = 17.16 Hz, JA-F,G
=
1,4-Cyclohexadiene monoepoxide (3) was prepared
according to Svendsen et al. The spectroscopic data
were in accordance with the literature [25–29]. trans-
(4) and cis-6-Methyl-3-cyclohexen-1-ol (6) were
prepared as reported [23] with spectroscopic data in
accordance with those reported before [30,31].
3-Buten-1-ol (1) and 4-penten-2-ol (10) were com-
mercially available. Both compounds, together with
the synthesised trans- (4) and cis-6-methyl-3-cyclo-
hexen-1-ol (6), and 3-cyclopenten-1-ol (7, 89%
pure, prepared [32] according to Hess and Brown
6.98 Hz, HA). Not quite stopped exchanging. 1H
NMR (100 MHz): d 0.91 (1H, t, JH-D,E = 5.86 Hz,
HH). Stopped exchanging. For notation, see Fig. 3.
3.2.2. cis-6-Methyl-3-cyclohexen-1-ol (6)
1H NMR (400 MHz): d 0.89 (1H, d, JC-I = 6.83 Hz,
HI), 0.97 (3H, d, JH-J = 6.84 Hz, CH3 (J)), 1.73–1.93
(2H, m, HG or F, H), 2.00–2.12 (2H, m, HE or D, F or G),
2.20–2.32 (1H, dm, Jd Ϸ 18 Hz, HD or E), 3.76 (1H,
ddt, JC-I = 7.0 Ϯ 0.2 Hz, JC-H = 2.2 Ϯ 0.2 Hz, JC-E
=
4:5 Ϯ 0:2 Hz, JC-D = 4.5 Ϯ 0.2 Hz, HC), 5.45–5.55

250
J.M. Bakke, L.H. Bjerkeseth / Journal of Molecular Structure 470 (1998) 247–263
2.3 Hz, HC), 5.64 (2H, br. s, HA,B). For notation, see
Fig. 5.
3.2.4. Epicholesterol (9)
3
1H NMR (400 MHz): d 0.97 (1H, d, JCH-OH
=
7.57 Hz, OH), 3.84 (1H, m, HCOH). The OH signal
and its coupling were identified by a J-resolved NMR
experiment and from its disappearance on shaking
with D2O.
3.3. IR
The IR spectra were recorded at high dilution at
room temperature in CCl4 on a single beam Nicolet
20SXC FT-IR spectrometer equipped with a Nicolet
620 Data Station and TGS detector using IR quartz
cells with 10 mm path length (transparent in the
hydroxyl stretch frequency region in IR), 256 scans
and 2 cm−1 resolution. The instrument was purged
with dry air. The CCl4 solvent was refluxed over
and subsequently distilled from P2O5 under N2 atmo-
Fig. 4. Notations, and ball and stick models of the two Anti ring
conformers of cis-6-methyl-3-cyclohexen-1-ol (6) as examples of
the 6ax and 6eq conformers (HF/6-31G** geometries).
(1H, m, HB), 5.55–5.65 (1H, m, HA). For notation, see
Fig. 4.
˚
sphere, and kept over activated 3 A molecular sieves
in a dry box under N2 atmosphere. The high dilution
CCl4 solutions for IR (c Ͻ 5 mM) were prepared and
added to the IR cells in the dry box, and were kept
3.2.3. 3-Cyclopenten-1-ol (7)
1H NMR (400 MHz): d 0.94 (1H, d, JC-H = 6.59 Hz,
HH), 2.24 (2H, br. d, HF,G), 2.57 (2H, br. dd, JD-F = 16–
˚
over freshly activated 3 A molecular sieves for 24 h
17 Hz, JC-D = 6.34 Hz, HD,E), 4.40 (1H, dtt, JC-F,G
=
before measurement. Overlapping bands were
resolved by the FOCAS 2.1 program from Nicolet
using a linear baseline approach. The number of
bands and their positions were assessed by use of
Fourier Self Deconvolution.
Note: keeping a dilute solution over activated mo-
lecular sieves will reduce the concentration and make
it impossible to measure absolute values of integrated
absorptions. However, relative values are valid.
3.4. Theoretical calculations
Molecular mechanics calculations were performed
with the program MMPMI 1.0 [34] on a PC using the
default dipole–dipole electrostatic mode, or, where
noted, charge–charge interaction mode. MMPMI
consists of MM2 (N.L. Allinger, QCPE Program no.
395) and MMP1 (N.L. Allinger, QCPE Program no.
318), as implemented by Serena Software, Blooming-
ton, Indiana, USA. Input files for MMPMI were made
on a PC using PCMODEL version 1.0 or 4.0 from
Serena Software. The individual point charges for
Fig. 5. Notations, and ball and stick models of the two Anti ring
conformers of 3-cyclopenten-1-ol (7) as examples of the 7ax and 7eq
conformers (HF/3-21G geometries).

251
J.M. Bakke, L.H. Bjerkeseth / Journal of Molecular Structure 470 (1998) 247–263
MMPMI in charge–charge mode were calculated with
CHARGE2, version 1.0 [35–39] on a PC.
with the increase in temperature from 0 K to
298.15 K; and (iv) Gibbs free energy based on (iii)
and the calculated total entropy Stot at 298.15 K. The
calculations using the options (ii)–(iv) were per-
formed both with and without scaling of the frequen-
cies. The GAUSSIAN program package uses the
Rigid Rotor-Harmonic Oscillator approximation to
calculate the thermodynamical properties (no special
treatment for low-frequency modes).
The semiempirical MNDO, AM1 and PM3 MO
calculations were performed with MOPAC 3.13 [40]
on a Digital MicroVax II, or with MOPAC 6.00 [41]
on a Digital Vaxstation 3100 M38. Input files were
made by PCMODEL.
Ab initio calculations were performed with
GAUSSIAN 90, Revision I [42], GAUSSIAN 92,
Revision C or G.2 [43], or with GAUSSIAN 94, revi-
sion B.1 [44], on the Cray X-MP/216 and Cray
Y-MP4D/464 supercomputers in Trondheim. The cal-
culations were done using Restricted Hartree–Fock
(RHF) and Restricted Møller–Plesset of second
order (RMP2) theory (closed-shell species, HF and
MP2). Input files were made by PCMODEL as
MOPAC input files. MOPAC without geometry opti-
mization was used to transform MOPAC style Z-
matrixes to GAUSSIAN style Z-matrixes. Standard
GAUSSIAN basis sets were used (with 6 d polariza-
tion functions on heavy atoms for basis sets derived
from 6-31G). Geometries were gradient-optimised
without symmetry restrictions using the Berny opti-
miser and default convergence criteria. For the MP2
calculations, both the valence and core orbitals were
included in the calculations. Analytical second deri-
vatives were used in the geometry optimisations for
all compounds using 3-21G and 6-31G** basis sets
(except for 7), and for the calculations of hydroxyl
stretch IR frequencies. All vibrational frequencies
for all calculated conformers were positive, confirm-
ing that stationary points were all local minima. For
the geometry optimisation of 7 and MP2 calculations
of 1, numeric force constants were calculated at each
point, using gradient information from previous points.
No numeric vibrational frequencies were calculated.
For calculations of thermodynamical properties
based on scaled vibration frequencies with
GAUSSIAN 92, revision C, a modified version of
link 716, which includes a scale factor, was used in
the recalculations.
4. Results and discussion
If one assumes that staggered conformations are the
stable ones for sp3-sp3 C-C and C-O bonds, and for
sp3-sp2 C-CyC bonds the eclipsed ones [45], there are
27 possible conformers of 3-buten-1-ol (1). Of these,
26 make up 13 enantiomeric pairs. The 14 conformers
are shown in Fig. 1, while notation for the numbering
of atoms and naming of the rotamers is shown in Fig. 3.
Only two conformers have the geometrical possibi-
lity for intramolecular hydrogen bonding: Conformers
12 and 13. We have studied the conformational
composition of 3-buten-1-ol and model homoallylic
1
alcohols by IR and H NMR spectroscopy, and by
theoretical calculations. For Conformer 12 we chose
cis-6-methyl-3-cyclohexen-1-ol (6) and 3-cyclopen-
ten-1-ol (7) as models, and for Conformer 13 the
OH epimer of cholesterol (8): epicholesterol (9).
From inspection of models and the theoretical calcu-
lations, these molecules have geometries of the homo-
allylic alcohol fragments close to those of Conformers
12 and 13 of 3-buten-1-ol.
cis-6-Methyl-3-cyclohexen-1-ol (6) has six poss-
ible conformers. Two Anti conformers are shown in
Fig. 4. In these, HI and HC have an anti conformation
around C1-O8, and the OH group has either an axial
(6ax) or equatorial position (6eq). The shown example
of 6ax (Anti-Ax) is the conformer that models Con-
former 12 of 3-buten-1-ol. 3-Cyclopenten-1-ol (7) has
four possible conformers, and these are shown with
two Anti conformers as examples of the axial (7ax) and
equatorial (7eq) conformers of the ring in Fig. 5. The
conformer that models Conformer 12 of 3-buten-1-ol is
the shown example of 7ax (Anti-Ax). Fig. 6 shows
the Anti conformer of epicholesterol that acts as a
model for Conformer 13 of 3-buten-1-ol. Anti
denotes that the hydroxyl proton is antiperiplanar
The conformational compositions were calculated
from the relative energies, using the Boltzmann’s law
of distribution at 298 K. For the ab initio calculations,
the compositions were calculated based on: (i) total
energy; and also for one molecule on (ii) total energy
together with zero point energy; (iii) total energy
together with zero point energy and DHvib connected

252
J.M. Bakke, L.H. Bjerkeseth / Journal of Molecular Structure 470 (1998) 247–263
The band at 3624.9 cm−1 (n2) only showed as a
shoulder on the large n1 band in the experimental
spectrum (Fig. 7). The most striking part of the spec-
trum is the band at 3596.2 cm−1, 42 cm−1 lower than
the major band and with 37% of the total OH band
area. 3-Buten-1-ol is the only open chain olefinic alco-
hol that, in addition to two OH stretch bands at stan-
dard frequencies [48–51] in IR, has an extra band at
such a low frequency and with a significantly greater
band half width than all the other OH bands of open
chain olefinic alcohols [18]. This band has accord-
ingly been assigned to an intramolecularly hydro-
gen-bonded conformer [8–13]. The identity and
population of the hydrogen-bonded conformer(s)
are, however, somewhat uncertain. It has been
reported [7] that hydrogen bonding increases the
hydroxyl stretch molecular absorption coefficient
(Band 3). But it has also been reported that the other
two bands (Bands 1 and 2) may not have the same
hydroxyl stretch molecular absorption coefficients
due to differences in delocalisation possibilities [49].
cis-6-Methyl-3-cyclohexen-1-ol (6) and 3-cyclo-
penten-1-ol (7) have geometries of the homoallyl
alcohol fragment close to that of Conformer 12,
while trans-6-methyl-3-cyclohexen-1-ol (4) has a geo-
metry in its dominant conformation that makes an
intramolecular hydrogen bond impossible. The low
Fig. 6. Ball and stick model of the Anti conformer of epicholesterol
(9, MMPMI geometry).
to the vicinal proton. The other conformers of the
model compounds have a Gauche C–O rotamer.
1
The IR and H NMR spectra were obtained at con-
centrations lower than 5 mM in CCl4 or CCl3F. Under
these conditions, intermolecular hydrogen bonding is
not important, and interactions with the solvent are at
a minimum since both CCl4 and CCl3F are non-polar
non-hydrogen bonding solvents [2,3,18–22,46]. CCl4
and CCl3F have equivalent solvent properties [47].
4.1. IR spectroscopy
The pertinent IR data for 3-buten-1-ol (1) and the
model compounds are given in Table 1. The data for
3-buten-1-ol were close to those reported earlier [8–
13]. However, this appears to be the first time the
relative areas of the three OH stretch bands have
been reported as the result of a numerical analysis.
frequency hydroxyl stretch IR band of
6 at
3599.1 cm−1 (Fig. 8) was only 7 cm−1 lower than
that of the trans isomer (4), and 3-cyclopenten-1-ol
(7) had its lowest band at 3601.3 cm−1 (Fig. 8).
Neither the low frequency band of 6 nor that of 7 had
such a large band half width as that of 3-buten-1-ol
Table 1
IR spectroscopic data for 3-buten-1-ol (1) and some other homoallylic alcohols in the OH stretch region
Frequencies
Band areas/%
Compound
n1
n2
n3
n1
n2
n3
3-Buten-1-ol (1)a
3638.0 (19.8)
3633.0 (15.3)
3626.1 (23.8)
3624.6 (18.1)
3623.1 (15.0)
3621.6 (20.1)
3626.7 (18.0)
3624.9 (21.0)
3621.9 (17.9)
3599.1 (17.6)
3601.3 (17.0)
3611.6 (18.5)
3589.0 (32.1)
3610.4 (22.4)
3596.2 (32.8)
3605.7 (17.5)
48
55
67
71
73
36
45
15
23
33
29
27
64
13
37
23
trans-6-Methyl-3-cyclohexen-1-ol (4)
cis-6-Methyl-3-cyclohexen-1-ol (6)b
3-Cyclopenten-1-ol (7)b
Cholesterol (8)b
Epicholesterol (9)b
4-Penten-2-ol (10)
3590.0 (26.3)
42
All spectra recorded at c Ͻ 5 mM in CCl4. All frequencies in cm−1. Numbers in parentheses are band half widths.
aSee Fig. 7 for experimental spectrum with resolved bands and the reconstructed curve (sum of the resolved bands).
bSee Fig. 8 for experimental spectra and reconstructed curves.

253
J.M. Bakke, L.H. Bjerkeseth / Journal of Molecular Structure 470 (1998) 247–263
Fig. 7. IR spectrum of the hydroxyl stretch region of 3-buten-1-ol (1) at high dilution (c Ͻ 5 mM) in CCl4. The uppermost curve is the
experimental followed by the reconstructed curve which is the sum of the three shown resolved bands at 3638.0, 3624.9 and 3596.2 cm−1. See
Table 1.
(1). These data did not indicate the presence of intra-
molecular hydrogen bonds in these compounds.
On the other hand, epicholesterol (9) with a geome-
try of the homoallyl alcohol fragment close to that of
Conformer 13, had its lowest hydroxyl stretch IR fre-
quency band (3589.0 cm−1, Fig. 8) 23 cm−1 lower than
that of cholesterol (8, Fig. 8), 7 cm−1 lower than the
low frequency band of 3-buten-1-ol and at approxi-
mately the same frequency as the low frequency band
of the homoallylic alcohol 4-penten-2-ol (10) with a
secondary OH group. This low frequency epicholes-
terol band also had a band half width comparable to
that of the low frequency band of 3-buten-1-ol. Cho-
lesterol has no possibility for an internal hydrogen
bond. This suggests that epicholesterol (9) has an
intramolecularly hydrogen-bonded conformer. Infra-
red hydroxyl stretch frequencies for 7 [52], 8, 9 [8,53]
and 10 [11,13,54] have been reported earlier.
Fig. 3), while the conformers of the model compounds
with geometrical possibility for intramolecular hydro-
gen bonding have an Anti C–O rotamer (Type III)
ꢀnII Ͼ nIII †.
In conclusion on this point, the model compounds
with geometries close to that of Conformer 12 of
3-buten-1-ol showed IR spectra without low fre-
quency IR hydroxyl stretch bands, indicative of intra-
molecular hydrogen bonding. This suggests that this
conformer was not the source of the low frequency
band in 3-buten-1-ol (1). On the other hand, the model
for Conformer 13, epicholesterol (9), showed such a
band. The IR study therefore indicated that this con-
former gave rise to the hydroxyl stretch band at
3596.2 cm−1 in the IR spectrum of 3-buten-1-ol.
1
4.2. H NMR spectroscopy
¯
It is to be expected according to the rules of Oki et
The vicinal coupling constants contain information
on the conformational composition. In 3-buten-1-ol
(1), there are three single bonds. Different relation-
ships exist between the dihedral angles and vicinal
coupling constant for each of these bonds. The
observed coupling constants are the weighed average
of the coupling constants for each of the three
al. [50,51] that if the model compounds are intramo-
lecularly hydrogen bonded, their IR spectra should
have a hydroxyl stretch IR band at a lower frequency
than the low frequency band of 3-buten-1-ol (1). The
reason for this is that both Conformers 12 and 13 of 3-
buten-1-ol have a gauche C–O rotamer (Type II,

254
J.M. Bakke, L.H. Bjerkeseth / Journal of Molecular Structure 470 (1998) 247–263
Fig. 8. IR spectra of the hydroxyl stretch region of (a) cis-6-methyl-3-cyclohexen-1-ol (6), (b) 3-cyclopenten-1-ol (7), (c) cholesterol (8) and (d)
epicholesterol (9) at high dilution (c Ͻ 5 mM) in CCl4. The uppermost curves are the experimental ones, while the curves slightly below are the
reconstructed ones made up of the sum of the resolved bands given in Table 1.
rotamers around each of these bonds (Fig. 3). To
determine the composition around the C1-O5 bond
only conformers that can be intramolecularly hydro-
gen bonded. Both have a gauche relationship around
C1-O5 and a gauche one around C1-C2 (Fig. 3). The
vicinal coupling constants indicate the gauche rota-
mers to be the major ones around both C1-O5 (71%)
and C1-C2 (75%). However, the two conformers have
different rotamers around C2-C3. Conformer 12 has a
cis relationship, i.e. with C1 in the plane of the double
bond. Conformer 13, on the other hand, has a gauche
rotamer around this bond with one of the hydrogens
on C2 in the plane of the double bond. The vicinal
3
3
we have used J60Њ = 2.2 Hz and J180Њ = 12.5 Hz
[19,20,46]. This gives Janti = 2.2 Hz and Jgauche
=
7.35 Hz. For the H2C1-C2H2 fragment, we have
used the values Jgauche = 5.1 Hz and Janti = 8.6 Hz
which we used for the analysis of the CH2CH2
fragment of 2-phenylethanol [20]. For the H2C2-
C3H fragment we used the Garbisch equations devel-
oped for this system [55], and obtained Jgauche = 7.6 Hz
and Jcis = 3.6 Hz. From these relationships and the
observed vicinal coupling constants, the rotamer
distributions were obtained (Fig. 3). It should be
noted that only the average A2X2 spectrum was
observed for the CH2CH2 fragment (not the
AAЈXXЈ one) and an A2X for the CH2CHy fragment.
Conformers 12 and 13 in 3-buten-1-ol (1) are the
3
coupling constant JA-F,G indicates the major rotamer
to be the gauche one (84%, 16% anti). From the IR
spectrum it was evident that the intramolecularly
hydrogen-bonded conformer(s) had a relative large
population. This excludes Conformer 12 (maximum
population 16%) as a major component. Therefore,

255
J.M. Bakke, L.H. Bjerkeseth / Journal of Molecular Structure 470 (1998) 247–263
Conformer 13 appears to be the major intramolecu-
larly hydrogen-bonded conformer. This result is
consistent with the result from IR spectroscopy on
the model compounds. Whether the less populated
Conformer 12 is also intramolecularly hydrogen
bonded can not be decided from 1H NMR on 3-
buten-1-ol (1) alone.
2.2 Hz for secondary alcohols, see above), the com-
position around the CH–OH bond is close to 50% anti
and 50% Gauche (45/55). From these results, there is a
maximum of ca. 45% and a minimum of ca. 20% of
the conformer of 6 with a possibility for an internal
hydrogen bond from the hydroxyl group to the p-bond
(Anti-Ax). Even the lowest estimate would give a
high enough concentration to have its OH stretch
band observed by IR spectroscopy. No low frequency
IR OH stretch band was observed for 6 (Table 1).
3-Cyclopenten-1-ol (7) also has two ring system
conformations, one with the hydroxyl group in an
axial position (7ax) and one with the hydroxyl group
in an equatorial position (7eq, Fig. 5). Only 7ax has a
geometry which makes an internal hydrogen bond
possible if the hydroxyl proton points towards the p-
1
The H NMR results from the model compounds 6,
7 and 9 gave further information on this. In these
compounds, the major conformers have geometries
of the homoallyl alcohol fragments close to those of
Conformers 12 or 13 of 3-buten-1-ol. We will first
discuss the geometry and intramolecular hydrogen
bonding of the models for Conformer 12. cis-6-
Methyl-3-cyclohexen-1-ol (6) has two stable ring con-
formations, one with the hydroxyl group in an axial
position and the methyl group equatorial, the other
with an equatorial hydroxyl group (6ax and 6eq, Fig.
4). From data for conformations of the cyclohexane
system [45], we expected 6ax to be the more stable one
of these two. In this ring conformer, the hydroxyl
group can occupy a position with a geometrical pos-
sibility for an intramolecular hydrogen bond to the p-
1
bond. The H NMR spectrum contains information on
both the conformational composition and the flip
angle of the cyclopentene ring. The observed two vic-
3
3
inal coupling constants JC-D and JC-F (for notation,
see Fig. 5) are the weighted average of those of the
two ring conformations 7ax and 7eq. If we assume the
flip angle of the ring to be the same in the two ring
1
3
electron system of the double bond. The H NMR
conformations, the cis vicinal coupling constant, JC-
spectrum at 400 MHz did not permit an estimate of
the dihedral angles of the ring due to overlapping
bands. However, the signal from HC was well sepa-
D, will be the same in both ring conformations. We
have earlier used Eq. (1), a modified Karplus equa-
tion, for these systems [19,47].
3
rated, and we used the JC-E (see Fig. 4 for notation)
3J = 12:9 cos2v − 0:32 cosv
(1)
for an estimate of the 6ax/6eq ratio. For 6ax, both HC
3
3
and HE are equatorial, and a small JC-E would be
From Eq. (1) and the observed JC-D (6.34 Hz), we
obtain a dihedral angle HC-C1-C5-HD = 44Њ (f). For
angles of this magnitude, the dihedral angle is close to
expected. For 6eq, they would both be axial with a
corresponding large coupling constant. The observed
coupling constant 4.5(2) Hz showed ring conformer
6ax to be the dominant one. To obtain a more accurate
estimate, the HF/6-31G** geometries for 6ax and 6eq
1
the angle of pucker [47] and the H NMR results thus
indicate this to be ca. 40Њ. This is somewhat larger
than that reported for 2-cyclopenten-1-ol (32Њ, [19])
and a few other cyclopentene systems (27–37Њ,
[47,57,58]). From the estimated dihedral angle HC-
C1-C5-HD (f), the observed trans vicinal coupling
3
were used as inputs for PCMODEL 4.0, and JC-E was
evaluated by Haasnoot’s equations implemented in
3
that program [56]. This gave JC-E = 2.3 Hz for 6ax
3
constant JC-F (2.3 Hz) and Eq. (1), a 7ax population
and 10.4 Hz for 6eq. From these values and the
3
of 86% is obtained when assuming that the dihedral
angle HF-C-C-HC = 120Њ − f for 7ax and 120Њ + f for
7eq. This is the same population (85%) as obtained
observed JC-E, the conformational composition of
the ring of 6 was found to be 72(3)% 6ax and 28% 6eq.
For the ring conformer 6ax, one of the rotamers
around the C-O bond (Anti) points towards the double
bond with a geometry favourable for an intramolecu-
lar hydrogen bond. This is the conformer Anti-Ax
given in Fig. 4 as an example of 6ax. From the
3
using the measured JC-F and the Haasnoot’s equation
3
[56] to obtain values for JC-F from HF/3-21G geo-
metries for 7ax and 7eq. The major ring conformation
of the cyclopentene ring therefore places the hydroxyl
group in the axial position. From the vicinal coupling
3
3
observed JC-I = 6.83 Hz and the Js for the rotamers
of the CH-OH fragment (JAnti = 12.5 Hz and JGauche
3
constant JCH-OH, the rotamer composition around
=

256
J.M. Bakke, L.H. Bjerkeseth / Journal of Molecular Structure 470 (1998) 247–263
Table 2
Calculated conformer populations (%) for the 14 3-buten-1-ol (1) conformers using Hartree–Fock theory
3-21G//
3-21G
based on
6-31G**//6-31G**
based on
Total
Total
Total
Total
DGЊb
energy
energy
energy
energy
+ ZPE
+ DHbvib
+ ZPEb
Conformer MMPMI
MMPMIa MNDO
AM1
PM3
1
4
3
0
0
0
0
1
1
1
1
2
3
21
2
22
3
12
1
1
6
1
3
5
1
8
1
10
1
9
1
12
1
4
5
6
6
6
4
11
12
1
10
10
1
20
22
0
17
17
0
5
6
1
7
8
0
8
9
0
7
8
1
9
10
0
7
8
9
17
14
1
3
8
1
13
11
1
0
1
2
0
0
1
0
21
5
2
11
1
2
13
1
2
12
1
3
13
1
10
11
12
13
14
5
4
1
9
4
4
2
18
8
10
8
2
13
8
4
10
5
16
13
7
8
4
27
15
2
3
8
2
4
8
2
36
5
3
8
2
40
5
5
8
1
30
6
18
1
31c
4
43d
5
6
aCharge–charge interaction modus.
bZPE denotes the zero point energy. DHvib denotes the extra vibrational energy connected with increasing the temperature from 0 K to
298.15 K. DGЊ is calculated from the calculated entropy Stot and total energy + ZPE + DHvib. Scaled frequencies were used for the calculations
of ZPE, DHvib, Stot and DGЊ. HF/6-31G** scale factor: 0.87800 [48]. The adjustments are based on the rigid rotor-harmonic oscillator
approximation as implemented in the GAUSSIAN program package after scaling of the frequencies.
cThe zero of energy was: − 229.6746618 hartrees.
dThe zero of energy was: − 230.9725784 hartrees.
the CH-OH bond can be estimated. The observed
3JCH-OH = 6.59 Hz together with the vicinal coupling
constants for each rotamer (see above) indicate that
the rotamer with the hydroxyl proton pointing towards
the p-bond constitutes 43% of this rotamer composition.
From this and the ring conformational composition with
86% 7ax, at least 28% and maximum 43% of the con-
formational population will have a geometrical possi-
bility for an internal hydrogen bond. This is the
conformer Anti-Ax shown in Fig. 5 as an example
of 7ax. The IR OH stretch band of this conformer
would certainly be observable in the spectrum. No
low frequency OH stretch IR band was observed for
7 (Table 1).
frequency indicative of internal hydrogen bonding
(Table 1), the H NMR results indicate that the hydro-
gen-bonded hydroxyl stretch IR band in 1 was not
caused by Conformer 12.
1
For Conformer 13, epicholesterol (9) was chosen as
a model (Fig. 6). In this compound, the geometry of
the homoallyl fragment is close to that of the C-C-C-O
skeleton of Conformer 13 (see Section 4.3.2 for
3
details) and the JCH-OH would give information on
the rotamer composition around the C-O bond. The
3
observed JCH-OH = 7.57 Hz shows 52% of the Anti
rotamer with the hydroxyl proton pointing towards
the 5,6-double bond and 48% of the Gauche one
with the hydroxyl group pointing away (JAnti
=
The theoretical calculations indicated the two mod-
els for Conformer 12 of 3-buten-1-ol, cis-6-methyl-3-
cyclohexen-1-ol (6) and 3-cyclopenten-1-ol (7), to
have geometries of the homoallyl fragment close to
that of Conformer 12 (see Section 4.3.2 for details).
As the two models did not show an IR band at a
12.5 Hz and JGauche = 2.2 Hz). This is consistent with
the results from the IR spectroscopy. The IR spectrum
of epicholesterol (9) had a low frequency hydroxyl
stretch band at 3589 cm−1 with 64% of the band area
(Table 1). From its position and also its bandwidth, this
was assigned to an intramolecularly hydrogen-bonded

257
J.M. Bakke, L.H. Bjerkeseth / Journal of Molecular Structure 470 (1998) 247–263
1
conformer. The H NMR results confirm this, ca. 50%
but overestimated the gauche rotamer around the C1-
O5 bond (see below), as did the AM1 calculations.
We also used ab initio calculations at various
levels. The conformer populations from some of the
Hartree–Fock calculations using the same basis sets
for both geometry optimisation and energy calcula-
tions are given in Table 2. By using the small basis
set 3-21G, a population of 31% of Conformer 13 was
obtained. The largest basis set used was 6-31G**. For
this, we have tabulated the populations based on total
energy and total energy corrected for thermodynami-
cal parameters. The tabulated values are based on
scaled frequencies. However, the difference between
calculations using scaled and non-scaled frequencies
were hardly visible. The HF/6-31G** calculations
indicate Conformer 13 with a possibility for an inter-
nal bond to be the dominant one (30–43%). It is
important to note that Conformer 12, the other con-
former with a geometry which makes an internal
hydrogen bond possible, is only populated at 1–2%
by the HF/6-31G** calculations. AM1 was the
method that gave the highest population for this con-
former, only 5%. Of these calculations, the HF/6-
31G** results are believed to be the most reliable
and the HF/6-31G** geometries of the conformers
of 1 are given in Fig. 1. Thus, the calculations support
that Conformer 13 is the major conformer of 1 that
gives rise to the low frequency IR hydroxyl stretch
band. Whether the less populated Conformer 12 is
also intramolecularly hydrogen bonded can not be
decided by energy calculations alone.
of the C-O rotamers have hydroxyl groups pointing
towards the p-bond. The integrated molecular absorp-
tion coefficient for Band 2 was 64% larger than for
Band 1. It is well known that the IR band intensity
may increase on hydrogen bonding [7]. However,
other factors also influence the relative molecular
absorption coefficients [48–51].
1
The conclusion from the IR and H NMR spectro-
scopic investigations of 3-buten-1-ol (1) and the
model compounds is that Conformer 13 of 3-buten-
1-ol is the intramolecularly hydrogen-bonded conformer.
This conformer gives rise to the IR band at 3596 cm−1
for 3-buten-1-ol. Conformer 12 of 3-buten-1-ol (1) is
not intramolecularly hydrogen bonded.
4.3. Theoretical calculations
We have performed calculations on the geometries
and energies of 3-buten-1-ol (1) and its model com-
pounds cis-6-methyl-3-cyclohexen-1-ol (6), 3-cyclo-
penten-1-ol (7) and epicholesterol (9). The calculations
were effected at various levels, depending on the number
of atoms in the molecules. 3-Buten-1-ol (1) itself was
the smallest of the studied compounds and was
accordingly calculated at the highest level.
4.3.1. Theoretical calculations on 3-buten-1-ol (1)
In Table 2, we present the calculated populations of
the conformers of 1 from the obtained relative ener-
gies and Boltzmann’s law of distribution. Conformers
12 and 13 (see Fig. 1) are the only ones with a possi-
bility for an internal hydrogen bond. From the spec-
troscopic investigations (see above), Conformer 13
appeared to be the one giving rise to the low
frequency hydroxyl stretch band in the IR spectrum.
The calculations on 1 were performed by molecular
mechanics, semiempirical molecular orbital and ab
initio methods. The results were very dependent on
the method of calculation. The molecular mechanics
calculation in the dipole–dipole interaction mode
gave Conformers 2, 7 and 8 as the most stable ones
with only a minor population of Conformer 13. In the
charge–charge interaction mode this changed, and this
calculation indicated Conformer 13 to be a major one
together with Conformers 2, 4 and 5. The semiempiri-
cal molecular orbital calculations by the PM3 method
gave Conformer 13 as the most populated one, 27%,
The optimized HF/6-31G** geometries for these
two conformers are given in Table 3. For reference,
we also give the MP2/6-31++G** geometry of Con-
former 13 in this table. From the optimised geometries
and energies of all 14 conformers, the rotamer com-
positions around the three bonds C1-O5, C1-C2 and
C2-C3 (see Fig. 3 for notation) were calculated.
These are given in Table 4 together with those calcu-
1
lated from the observed vicinal H NMR coupling
constants. Except for the C2-C3 composition for
which the gauche rotamer was slightly overestimated,
the compositions calculated by HF/6-31G**//HF/6-
1
31G** were close to those found from H NMR spec-
troscopy. For the calculations performed on a lower
level, there were larger discrepancies.
From the HF/6-31G** calculations of 1, the hydroxyl
group geometry was available (Table 5). The O-H

258
J.M. Bakke, L.H. Bjerkeseth / Journal of Molecular Structure 470 (1998) 247–263
Table 3
Ab initio HF/6-31G** and MP2/6-31++G** optimised geometries for the two conformers of 3-buten-1-ol (1) with geometrical possibility for
internal hydrogen bonding (Conformers 12 and 13)a
HF/6-31G**
12
MP2/6-31++G**
13
13b
˚
Bond distances/A
C3yC4
1.3204
1.5110
1.5238
1.4005
1.0768
1.0755
1.0793
1.0896
1.0872
1.0830
1.0889
0.9432
1.3198
1.5056
1.5298
1.3983
1.0772
1.0760
1.0806
1.0884
1.0855
1.0832
1.0906
0.9441
1.3418
1.4974
1.5275
1.4254
1.0820
1.0805
1.0863
1.0929
1.0910
1.0878
1.0952
0.9677
C2-C3
C1-C2
O5-C1
HC-C4
HB-C4
HA-C3
HG-C2
HF-C2
HE-C1
HD-C1
HH-O5
Angles/Њ
C2-C3yC4
C1-C2-C3
O5-C1-C2
HC-C4yC3
HB-C4yC3
HA-C3-C2
HG-C2-C3
HF-C2-C3
HE-C1-C2
HD-C1-C2
HH-O5-C1
Dihedral angles/Њ
C1-C2-C3yC4
O5–C1-C2-C3
HC-C4yC3-C2
HB-C4yC3-C2
HA-C3-C2-C1
HG-C2-C3yC4
HF-C2-C3yC4
HE-C1-C2-C3
HD-C1-C2-C3
HH-O5-C1-C2
127.35
125.17
124.16
115.80
112.76
122.96
120.92
114.72
108.46
108.62
109.15
110.45
109.48
111.96
112.43
121.67
121.60
115.96
109.58
110.14
109.76
109.83
109.24
111.02
111.90
121.08
121.53
116.79
110.25
110.00
110.48
109.95
107.48
− 16.3
− 66.5
0.1
− 179.1
164.6
106.7
− 138.3
175.8
58.4
108.9
− 65.2
2.0
− 178.8
− 69.3
− 130.4
− 12.2
176.7
58.8
106.9
− 63.8
3.3
− 177.9
− 70.2
− 132.5
− 13.4
179.0
59.6
54.4
54.2
51.2
aAtom numbering given in Fig. 3.
bThe zero of energy was: − 231.7570757 hartrees.
bond lengths for the anti C1-O5 conformers were all
gauche C1-O5 rotamers. For Conformer 12 the O-H
bond length was calculated to be 0.9432 A, very close
to those of the other conformers with a gauche con-
formation. However, the O-H bond length of Confor-
mer 13 deviated from this pattern with a bond length
˚
˚
within the interval 0.9423–0.9425 A, while the etha-
˚
nol anti conformer had a O-H bond length of 0.9424 A
[48]. For conformers of 1 with a gauche conformation
around the C1-O5 bond, the O-H bond length was
˚
˚
˚
calculated to be 0.9430–0.9432 A. The corresponding
of 0.9441 A, but only by 0.0010 A. Although the
effect is significant, it is small, and alone it would
not strongly indicate the presence of an internal
hydrogen bond. However, the same calculation gave
˚
value for the ethanol gauche conformer was 0.9431 A
[48]. Both Conformers 12 and 13 with geometrical
possibilities for an internal hydrogen bond have

259
J.M. Bakke, L.H. Bjerkeseth / Journal of Molecular Structure 470 (1998) 247–263
Table 4
Calculated rotamer compositions (%) around the C1-O5, C1-C2 and C2-C3 bonds in 3-buten-1-ol (1) as compared with the compositions from 1H
NMR (c Ͻ 5 mM in CCl3F)
C1-O5
C1-C2
C2-C3
Method
gauche
anti
gauche
anti
gauche
cis
MMPMI
40
63
62
97
99
72
60
37
38
3
1
28
60
49
66
50
63
83
40
51
34
50
37
17
88
90
95
87
92
92
12
10
5
13
8
MMPMIa
MNDO
AM1
PM3
HF/3-21G//HF/3-21G
HF/6-31G**//HF/6-31G**
based on:
8
Total energy
Total energy + ZPEb
Total energy + ZPE + DHbvib
DGЊb
77
73
75
70
71
23
27
25
30
29
75
71
74
67
75
25
29
26
33
25
95
95
95
96
84
5
5
5
4
1H NMRc
16
aCharge–charge interaction modus.
bSee footnote b in Table 2.
cCalculated from measured coupling constants and model coupling constants.
that the C1-O5-HH bond angle for Conformer 13
(109.24Њ, Table 5) is smaller than the corresponding
bond angles for all the other conformers of 1, in accor-
dance with the change in geometry necessary for an
hydroxyl group to form an internal hydrogen bond to
the p-bond. Together, we take these two changes as an
indication of a weak intramolecular hydrogen bond
for Conformer 13. The C1-O5-HH bond angle for Con-
former 12 was only slightly larger than the corre-
sponding angle for Conformer 13 and smaller than
the corresponding bond angles for all the other
conformers of 1. However, the O-H bond length of
Conformer 12 did not deviate from the O-H bond
lengths of the other gauche C1-O5 conformers. This
indicates that the interaction between the OH group
and the p-electron system in Conformer 12 is very
weak, much weaker than for Conformer 13. Conformer
14 also deviated from the bond length pattern, but this
was believed to be caused by a repulsion between the
hydroxyl proton and another proton in the molecule.
Table 5
Ab initio HF/6-31G** calculated hydroxyl group geometries for the conformers of 3-buten-1-ol (1)
C1-O5 bond length/A
˚
Conformer
gauche
anti
C1-O5-HH bond angle/Њ
1
2
3
4
5
6
7
8
0.9424
0.9424
109.84
109.92
109.69
109.73
109.68
109.76
109.90
110.00
109.79
109.86
109.62
109.48
109.24
109.79
0.9430
0.9430
0.9430
0.9425
0.9424
0.9423
9
0.9431
0.9432
0.9432
0.9432
0.9441
0.9424
10
11
12
13
14

260
J.M. Bakke, L.H. Bjerkeseth / Journal of Molecular Structure 470 (1998) 247–263
can reproduce the experimental IR hydroxyl stretch
frequency pattern for an acyclic saturated alcohol
[48]. However, even a medium–large basis set, (6-
31G**) at the Hartree–Fock level, can not fully repro-
duce the effect of intramolecular hydrogen bonding.
A clear sign for the presence of an intramolecular
hydrogen bond is that the distance between the donor
and acceptor functions of the presumed hydrogen
bond is less than the sum of the van der Waals radii
of the contributing atoms, in this case C and O. For
very weak hydrogen bonds, this test criteria may not
be satisfied. Instead, the test criteria is modified to use
the distance between C and H(O) as compared to the
sum of the van der Waals radii of C and H [4,60]. The
sum of the van der Waals radii of C and O is 3.1–
Table 6
Ab initio hydroxyl stretch frequenciesa calculated with the 6-31G**
basis set at the Hartree–Fock level for the conformers of 3-buten-1-
ol (1)
Conformer
C1-O5 gauche
C1-O5 anti
1
2
3
4
5
6
7
8
3680.8
3680.0
3671.5
3670.7
3671.4
3678.0
3680.4
3681.8
9
3669.9
3669.1
3668.6
3672.2
3658.3
3681.9
10
11
12
13
14
˚
3.2 A, while the sum of the van der Waals radii for C
˚
and H is 2.7–2.9 A [4,61,62]. However, the use of this
criteria for a hydrogen bond to CyC is not as straight-
forward as for a hydrogen bond to O, because it is
difficult to define the acceptor point. From the ab
initio HF/6-31G** geometries for 3-buten-1-ol Con-
All frequencies in cm−1
aThe frequencies are scaled. HF/6-31G** scale factor: 0.87800
[48].
.
4
4
˚
˚
We have also calculated the ab initio HF/6-31G**
O–H stretch IR frequency for each of the 14 confor-
mers of 3-buten-1-ol (1, Table 6). These frequencies
follow the same pattern as the O–H bond lengths. For
all gauche conformers (Type II) around the C–O
bond, except for Conformer 13 (and Conformer 14),
a frequency of 3670 Ϯ 2 cm−1 (Band 2) was calculated
after scaling to the gas phase O-H stretch frequency of
methanol, ngas = 3681 cm−1 [59]. However, for
Conformer 13, a frequency of 3658 cm−1 (Band 3)
was found, significantly lower than those of the
other conformers, which together with the small C1-
O5-HH bond angle indicates intramolecular hydrogen
bonding. For all anti conformers (Type I) around the
C-O bond, a frequency of 3680 Ϯ 2 cm−1 (Band 1)
was found. The frequency difference between the cal-
culated highest frequency bands is ca. 10 cm−1, while
the difference between band 2 and the low frequency
band of Conformer 13 (Band 3) is 12 cm−1. The cor-
responding values from high dilution IR spectroscopy
(Table 1) were 13.1 and 28.7 cm−1, respectively. The
calculated difference between Bands 1 and 2 was in
good agreement with the experimental result, while the
calculated frequency difference between Bands 2 and 3
was much smaller than the experimental value. This
was not surprising. We have previously shown that
HF/6-31G** frequency calculations semiquantitatively
former 13, the C -O (3.54 A) and C -H(O) (2.94 A)
distances were larger then the sum of the van der
Waals radii of C and O/C and H, while both the C3-
3
˚
˚
O (3.00 A) and C -H(O) (2.62 A) distances were
˚
0.1 A shorter (see Fig. 3 for notation). We take this
as an indication of a weak intramolecular hydrogen
bond in Conformer 13. For Conformer 12, the calcu-
lated distances were C4-O (3.30 A), C4-H(O)
˚
3
3
˚
˚
˚
(2.70 A), C -O (3.09 A) and C -H(O) (2.72 A).
Some of these distances fell in the intervals of the
van der Waals radii, but none of them was shorter
than the sum of the van der Waals radii. If Conformer
12 is intramolecularly hydrogen bonded, this bond must
be weaker than the hydrogen bond in Conformer 13.
The results from the ab initio HF/6-31G** calcula-
tions on the energy and geometries for 3-buten-1-ol
(1) are in accordance with and support the experimen-
1
tal IR and H NMR results. Only for Conformer 13
did we find a shorter O-H bond length and a smaller
C-O-H angle than for the other conformers, and only
for this conformer did we calculate a significantly
lower O-H stretch frequency than for the other con-
formers and found a short enough hydrogen bond
length for hydrogen bonding to take place.
4.3.2. Theoretical calculations on model compounds
The theoretical calculations on structures and

261
J.M. Bakke, L.H. Bjerkeseth / Journal of Molecular Structure 470 (1998) 247–263
energies were done at various levels of sophistication
for the models of the homoallylic alcohol 3-buten-1-ol
(1) Conformers 12 and 13. Only the results from the
highest levels of calculations will be given here.
In Conformer 12, the dihedral angle C1-C2-C3yC4
is − 16Њ (HF/6-31G**), while in the Anti-Ax confor-
mers of 6 (Fig. 4) and 7 (Fig. 5), the corresponding
dihedral angles are − 18Њ (HF/6-31G**) and − 17Њ
(HF/3-21G), respectively (see Figs. 3–5 for notation).
For the O-C1-C2-C3 dihedral angle, the corresponding
values were − 67Њ, − 76Њ and − 93Њ, while for the
H-O-C1-C2 dihedral angle, the values were 54Њ, 48Њ
and 57Њ. In Conformer 13, the dihedral angle C1-C2-
C3yC4 is 109Њ (HF/6-31G**), while the correspond-
ing dihedral angle for the Anti conformer of 9 (Fig. 6)
is 129Њ (MMPMI). For the O-C1-C2-C3 dihedral
angle, the corresponding values were − 65Њ and −
68Њ, while for the H-O-C1-C2 dihedral angle, 54Њ and
55Њ. As can be seen, the homoallylic geometries of the
models are close to those of Conformers 12 and 13 of
3-buten-1-ol (1).
indicative of intramolecular hydrogen bonding was
not present in the IR spectrums of 6 and 7, this indi-
cates that the hydrogen-bonded band in 1 was not
caused by its Conformer 12.
From the HF/6-31G** calculated O-H bond lengths
and hydroxyl stretch frequencies for 6, no conclusions
about the presence of intramolecular hydrogen bond-
ing could be drawn due to lack of a pattern. Nor for the
Hartree–Fock calculations on 7 could a conclusion be
drawn, since these calculations were done with the
small basis set 3-21G which is not accurate enough
for this kind of calculation: the pattern for O-H bond
length and hydroxyl stretch frequencies found for 3-
buten-1-ol (1) with HF/6-31G** (Section 4.3.1) was
not that obvious from the corresponding HF/3-21G
geometry and frequency calculations.
The distances C4-O, C3-O, C4-H(O) and C3-H(O)
for the Anti-Ax conformer of cis-6-methyl-3-cyclo-
hexen-1-ol (6) calculated with HF/6-31G** were
˚
3.28, 3.09, 2.79 and 2.71 A, respectively (for notation,
see Fig. 4). None of these distances is shorter than the
sum of the van der Waals radii of the donor and accep-
tor functions for a hydrogen bond, although some of
the distances fell in the intervals of the van der Waals
radii (Section 4.3.1). Thus, if Conformer 12 of
3-buten-1-ol (1) is intramolecularly hydrogen bonded,
this bond must be very weak. The calculated distances
are very close to the corresponding distances calcu-
lated for Conformer 12 (Section 4.3.1).
We will first discuss the geometries and energetics
of the models of Conformer 12. Ab initio Hartree–
Fock 6-31G**//6-31G** calculations on cis-6-
methyl-3-cyclohexen-1-ol (6) gave 83% of the ring
conformer with an axial hydroxyl group, 6ax (Fig.
4), and 56% of the conformer Anti-Ax, while 1H
NMR predicted 72(3)% 6ax and a maximum of 45%
Anti-Ax (minimum ca. 20%) by using the observed
coupling constants together with Haasnoot’s equa-
tions on the HF/6-31G** geometries (Section 4.2).
For 3-cyclopenten-1-ol (7), the Hartree–Fock calcula-
tions predicted 79% (3-21G//3-21G) and 61%
(6-31G**//3-21G) of the ring conformer 7ax with an
axial hydroxyl group (Fig. 5), and 73% (3-21G//3-
21G) and 42% (6-31G**//3-21G) of the conformer
Anti-Ax. For comparison, 1H NMR gave 86% 7ax
from a modified Karplus equation, and 85% 7ax and
a maximum of 43% Anti-Ax (minimum 28%) based
on HF/3-21G geometries and the Haasnoot’s
equations, see Section 4.2. The reason for the discre-
For the other model of Conformer 12, 3-cyclopen-
ten-1-ol (7), the HF/3-21G calculated distances in the
3
˚
symmetrical Anti-Ax conformer were 3.15 A (C -O)
3
˚
and 2.87 A (C -H(O)) [see Fig. 5 for notation]. Both
these values fell in the upper part of the intervals of
the van der Waals radii (Section 4.3.1). An intramo-
lecular hydrogen bond in this model molecule is thus
not very probable.
We will now focus on the model for 3-buten-1-ol
(1) Conformer 13. Epicholesterol (9, Fig. 6) is too
large to allow calculations with HF/3-21G or larger
basis sets, and was accordingly only calculated by
molecular mechanics (MMPMI). For the geometry
of the homoallyl fragment, the only dihedral angle
of interest is C-C-O-H. All the other dihedral angles
of the homoallylic C-C-C-O skeleton are fixed in a
chair form. The MMPMI calculations on 9 gave 24%
of the Anti conformer, which is the conformer respon-
sible for the hydrogen-bonded hydroxyl band at low
1
pancies between the ab initio and H NMR results for
7 might be the use of HF/3-21G instead of HF/6-
31G** for geometry optimisation. This confirms the
1
result from H NMR: the Anti-Ax conformers of 6
and 7 have high enough populations to have their
OH stretch bands observed by IR spectroscopy.
Since an IR hydroxyl stretch band at a frequency

262
J.M. Bakke, L.H. Bjerkeseth / Journal of Molecular Structure 470 (1998) 247–263
1
frequency in IR (Table 1). H NMR, however, gave
52% Anti. The explanation for this discrepancy is that
MMPMI underestimates weak hydrogen bonds. Since
this calculation was only done with molecular
mechanics and not with ab initio using a medium–
large basis set, no O-H bond lengths or hydroxyl
stretch IR frequencies can be safely compared. How-
ever, since the molecule is locked into a ring, the
calculated C4-O, C3-O, C4-H(O) and C3-H(O)
distances are assumed to be reasonably accurate.
Note that the superscripts denote the position in the
homoallylic system (see Fig. 3 for notation) and not
the numbering normally used for the cholesterol ske-
leton. The calculated values for these distances were
NMR spectroscopy) and theoretical methods at differ-
ent levels of sophistication (molecular mechanics,
semiempirical molecular orbital and ab initio meth-
ods) have been used in the study of conformational
composition, and the importance of intramolecular
OH…p hydrogen bonding for this composition in
3-buten-1-ol (1) and model compounds. The combi-
nation of experimental and theoretical data, and the
use of model compounds was necessary to understand
the conformational behaviour of these molecules. All
investigations on both 1 and its model compounds
showed that the Conformer 13 of 1 (see Fig. 1) is
the only intramolecularly hydrogen-bonded confor-
mer, and that the only other conformer that by visual
inspection of models might have a geometrical possi-
bility for an intramolecular hydrogen bond, Confor-
mer 12, is not intramolecularly hydrogen bonded.
Conformer 13 was the most populated conformer,
while Conformer 12 was hardly populated.
˚
3.70, 2.97, 3.15 and 2.58 A, respectively, while the
corresponding values for 3-buten-1-ol (1) Conformer
13 for comparison were 3.54, 3.00, 2.94 and 2.62 A,
˚
respectively (HF/6-31G** geometries). As for Con-
former 13, the C4-O and C4-H(O) distances of epicho-
lesterol’s Anti conformation were larger than the sum
of the van der Waals radii (see Section 4.3.1), while
both the C -O and C -H(O) distances were 0.1 A
shorter. This is an indication of weak intramolecular
hydrogen bonding in the model 9 and in 3-buten-1-ol
Conformer 13.
The results from the theoretical calculations on the
model compounds are also in accordance with and
support the results from experimental IR and 1H
NMR. We have shown that the geometries of the
models are close to those of 3-buten-1-ol Conformers
3
3
˚
Acknowledgements
Financial support for this research from The
Norwegian Research Council for Science and the
Humanities (NAVF) is gratefully acknowledged.
This work has also received support from The
Research Council of Norway (Programme for Super-
computing) through a grant of computing time (Cray
X-MP/216 and Cray Y-MP4D/464 in Trondheim).
1
12 and 13. Furthermore, as also shown by H NMR,
the populations of the conformers with a geometric
possibility for intramolecular hydrogen bonding in the
models for Conformers 12 and 13 were large enough
to be visible in the IR spectrum, indicating that the
hydrogen-bonded band in 1 was not caused by Con-
former 12 but by Conformer 13, since only the model
of Conformer 13 had an intramolecularly hydrogen-
bonded hydroxyl band in IR. From calculations of
bond distances between the acceptor and donor func-
tions for a hydrogen bond, it was shown that only the
model for Conformer 13 had a short enough hydrogen-
bond length for hydrogen bonding to take place.
References
[1] D.A. Smith (Ed.), Modeling the Hydrogen Bond, ACS Sym-
posium Series 569, American Chemical Society, Washington,
DC, Chapter 1, 1994.
[2] H.S. Aaron, Top. Stereochem. 11 (1979) 1.
[3] M. Tichy´, Adv. Org. Chem. 5 (1965) 115.
[4] F. Hibbert, J. Emsley, Adv. Phys. Org. Chem. 26 (1990) 255.
[5] P. Schuster, G. Zundel, C. Sandorfy (Eds.), The Hydrogen
Bond. Recent Developments in Theory and Experiments,
North Holland, Amsterdam, 1976.
[6] T. Dziembowska, Polish J. Chem. 68 (1994) 1455.
[7] G.C. Pimentel, A.L. McClellan, The Hydrogen Bond, W.H.
Freeman, San Francisco, 1960.
[8] P.v.R. Schleyer, D.S. Trifan, R. Bacskai, J. Am. Chem. Soc. 80
5. Conclusions
(1958) 6691.
¯
[9] M. Oki, H. Iwamura, Bull. Chem. Soc. Jpn. 32 (1959) 567.
A combination of experimental (FT-IR and 1H
[10] P. Arnaud, Y. Armand, Compt. Rend. 253 (1961) 1426.

263
J.M. Bakke, L.H. Bjerkeseth / Journal of Molecular Structure 470 (1998) 247–263
[11] Y. Armand, P. Arnaud, Ann. Chim. (Paris) 9 (1964) 433. USA, 1990.
[12] Y. Armand, P. Arnaud, Ann. Chim. (Paris) 9 (1964) 457.
[13] J.F. Bacon, J.H. van der Maas, Spectrochim. Acta 44A (1988)
1215.
[14] M. Trætteberg, H. Østensen, Acta Chem. Scand., Ser. A 33
(1979) 491.
[15] K.-M. Marstokk, H. Møllendal, Acta Chem. Scand., Ser. A 35
(1981) 395.
[16] T.M. Ivanova, S.F. Arkhipova, G.P. Kugatova-Shemyakina,
Acad. of Sci. USSR Bulletin 22 (1973) 1475.
[17] B.W. Gung, R.A. Fouch, D. Zou, J. Org. Chem. 61 (1996)
4200.
[42] M.J. Frisch, M. Head-Gordon, G.W. Trucks, J.B. Foresman,
H.B. Schlegel, K. Raghavachari, M.A. Robb, J.S. Binkley, C.
Gonzalez, D.J. Defrees, D.J. Fox, R.A. Whiteside, R. Seeger,
C.F. Melius, J. Baker, R.L. Martin, L.R. Kahn, J.J.P. Stewart,
S. Topiol, J.A. Pople, GAUSSIAN 90, Revision I, Gaussian,
Pittsburgh, Pennsylvania, USA, 1990.
[43] M.J. Frisch, G.W. Trucks, M. Head-Gordon, P.M.W. Gill,
M.W. Wong, J.B. Foresman, B.G. Johnson, H.B. Schlegel,
M.A. Robb, E.S. Reploge, R. Gomperts, J.L. Andres, K.
Raghavachari, J.S. Binkley, C. Gonzalez, R.L. Martin, D.J.
Fox, D.J. Defrees, J. Baker, J.J.P. Stewart, J.A. Pople, GAUS-
SIAN 92, Revision C and G.2, Gaussian, Pittsburgh, Pennsyl-
vania, USA, 1992.
[18] R.J. Abraham, J.M. Bakke, Acta Chem. Scand., Ser. B 37
(1983) 865.
[19] J.M. Bakke, A.M. Schie, T. Skjetne, Acta Chem. Scand., Ser.
B 40 (1986) 703.
[20] J.M. Bakke, D.J. Chadwick, Acta Chem. Scand., Ser. B 42
(1988) 223.
[21] J.M. Bakke, J. Krane, T. Skjetne, Acta Chem. Scand., Ser. B
43 (1989) 777.
[22] J.M. Bakke, L.H. Bjerkeseth, T.E.C.L. Rønnow, K. Steinsvoll,
J. Mol. Struct. 321 (1994) 205.
[23] L.H. Bjerkeseth, Conformational analysis of alcohols with
weak hydrogen bonds (in Norwegian). Thesis, Norwegian
Institute of Technology, Department of Organic Chemistry,
Trondheim, Norway, 1994.
[44] M.J. Frisch, G.W. Trucks, H.B. Schlegel, P.M.W. Gill, B.G.
Johnson, M.A. Robb, J.R. Cheeseman, T. Keith, G.A.
Petersson, J.A. Montgomery, K. Raghavachari, M.A. Al-
Laham, V.G. Zakrzewski, J.V. Ortiz, J.B. Foresman, J.
Cioslowski, B.B. Stefanov, A. Nanayakkara, M. Challacombe,
C.Y. Peng, P.Y. Ayala, W. Chen, M.W. Wong, J.L. Andres,
E.S. Replogle, R. Gomperts, R.L. Martin, D.J. Fox, J.S.
Binkley, D.J. Defrees, J. Baker, J.P. Stewart, M. Head-Gordon,
C. Gonzalez, J.A. Pople, GAUSSIAN 94, Revision B.1,
Gaussian, Pittsburgh, Pennsylvania, USA, 1995.
[45] E.L. Eliel, S.H. Wilen, Stereochemistry of Organic Com-
pounds, Wiley, New York, 1994.
[24] O. Mitsunobu, Synthesis (1981) 1.
[25] J.S. Svendsen, J.E. Whist, L.K. Sydnes, J. Chromatogr. 337
(1985) 9.
[46] J.M. Bakke, Acta Chem. Scand., Ser. B 40 (1986) 407.
[47] R.J. Abraham, J.M. Bakke, Acta Chem. Scand., Ser. B 38
(1984) 547.
[26] S.G. Davies, M.E.C. Polywka, S.E. Thomas, J. Chem. Soc.,
Perkin Trans I (1986) 1277.
[27] J.K. Crandall, D.B. Banks, R.A. Colyer, R.J. Watkins, J.P.
Arrington, J. Org. Chem. 33 (1968) 423.
[48] J.M. Bakke, L.H. Bjerkeseth, J. Mol. Struct. 407 (1997) 27.
[49] P.J. Krueger, J. Jan, H. Wieser, J. Mol. Struct. 5 (1970) 375.
¯
[50] M. Oki, H. Iwamura, Bull. Chem. Soc. Jpn. 32 (1959) 950.
¯
[51] M. Oki, H. Iwamura, T. Onoda, M. Iwamura, Tetrahedron 24
[28] v.R. Schwesinger, H. Prinzbach, Angew. Chem. (1972) 990.
[29] S.G. Davies, G.H. Whitham, J. Chem. Soc., Perkin Trans II
(1975) 861.
[30] J.A. Staroscik, B. Rickborn, J. Org. Chem. 37 (1972) 738.
[31] T. Ebata, H. Kawakami, K. Matsumoto, K. Koseki, K. Okano,
H. Matsushita, Biosci. Biotech. Biochem. 57 (1993) 515.
[32] K. Grøstad, Diploma thesis, Norwegian Institute of Technol-
ogy, Organic Chemistry Laboratories, Trondheim, Norway,
1985.
[33] H.M. Hess, H.C. Brown, J. Org. Chem. 32 (1967) 4138.
[34] N.L. Allinger, H.L. Flanagan, J. Comput. Chem. 4 (1983) 339
(as adapted by Serena Software, Bloomington, Indiana, USA).
[35] R.J. Abraham, G.H. Grant, B. Hudson, P.H. Smith,
CHARGE2, University of Liverpool, UK. Modified by D.J.
Chadwick for input of modified MNDO files.
[36] R.J. Abraham, L. Griffiths, P. Loftus, J. Comput. Chem. 3
(1982) 407.
[37] R.J. Abraham, B. Hudson, J. Comput. Chem. 5 (1984) 562.
[38] R.J. Abraham, B. Hudson, J. Comput. Chem. 6 (1985) 173.
[39] R.J. Abraham, P.E. Smith, J. Comput. Chem. 9 (1987) 288.
[40] J.J.P. Stewart, MOPAC ver. 3.13, Frank J. Seiler Research Labora-
tory, Air Force Systems Command, United States Air Force.
[41] J.J.P. Stewart, MOPAC ver. 6.00, Frank J. Seiler Research
Laboratory, United States Air Force Academy, Colorado,
(1968) 1905.
[52] E.F. Healy, J.D. Lewis, A.B. Minniear, Tetrahedron Lett. 35
(1994) 6647.
¯
[53] M. Oki, H. Iwamura, Bull. Chem. Soc. Jpn. 32 (1959) 306.
[54] P. Arnaud, Y. Armand, Compt. Rend. 253 (1961) 1547.
[55] E.W. Garbisch Jr, J. Am. Chem. Soc. 86 (1964) 5561.
[56] C.A.G. Haasnoot, F.A.A.M. De Leeuw, C. Altona, Tetrahe-
dron 36 (1980) 2783.
[57] M.I. Davis, T.W. Muecke, J. Phys. Chem. 74 (1970) 1104.
[58] D.A. Usher, E.A. Dennis, F.H. Westheimer, J. Am. Chem.
Soc. 87 (1965) 2320.
[59] W.J. Hehre, L. Radom, P.v.R. Schleyer, J. A. Pople, Ab Initio
Molecular Orbital Theory, Wiley, New York, 1986.
[60] S. Scheiner, in: K.B. Lipkowitz, D.B. Boyd (Eds.), Reviews in
Computational Chemistry II, Chapter 5, VCH Publishers, New
York, 1991.
[61] T. Zeegers-Huyskens, P. Huyskens, in: P.L. Huyskens, W.A.P.
Luck, T. Zeegers-Huyskens (Eds.), Intermolecular Forces. An
Introduction to Modern Methods and Results, Chapter 1,
Springer, Berlin, 1991.
[62] L. Pauling, The Nature of the Chemical Bond, 3rd edn, Cornell
University Press, New York, 1960, p. 260.