Conformational studies by dynamic NMR. 76.1 stereodynamics of ring inversion of bicyclo[3.3.1]nonan-9-one

Full text of DOI:10.1021/jo000049n

J . Org. Chem. 2000, 65, 3563-3565
3563
Sch em e 1. Com p u ted Str u ctu r es of th e Th r ee
Con for m a tion a l Stu d ies by Dyn a m ic NMR.
76.1 Ster eod yn a m ics of Rin g In ver sion of
Bicyclo[3.3.1]n on a n -9-on e
P ossible Con for m er s of 2. Th e Rela tive En er gy
Va lu es Ar e in k ca l m ol-1
Stefano Grilli,2 Lodovico Lunazzi,* and
Andrea Mazzanti*
Department of Organic Chemistry “A.Mangini”,
University of Bologna, Risorgimento, 4 Bologna 40136, Italy
Received J anuary 12, 2000
The chemistry of bicyclo[3.3.1]nonane, 1 has been ex-
tensively investigated both for what concerns the syn-
thetic3 and conformational properties.4-6
This confirms the negative result previously reported10
and made impossible the measurement of the ring
inversion barrier in 1, even though a theoretical calcula-
11
tion had suggested a value (7.6 kcal mol-1
)
which, in
principle, should be accessible to NMR determination.
The situation appears more promising with regard to
the possibility of measuring the ring inversion barrier of
the corresponding ketone, i.e., bicyclo[3.3.1]nonan-9-one,
2, which, at the best of our knowledge, has never been
determined. Also compound 2 can adopt three conforma-
tions, having the same symmetry of 1, as displayed in
Scheme 1.
In this case, too, the twisted twin boat form 2c has a
computed energy12 much too high to be appreciably
populated, but the difference between the energies of the
BC (2b) and of the most stable CC (2a ) form is signifi-
cantly smaller (0.95 kcal mol-1) than in 1. Our value is
at variance with that reported by Osawa et al. that
predicted11 a quite larger difference (2.61 kcal mol-1) but
essentially agrees with that (1.1 kcal mol-1) computed
by Raber et al.13
In addition to calculations, these authors carried out
also a lanthanide-induced shift (LIS) investigation and
found that their data could be better interpreted by
assuming a 22% proportion of the BC conformer (2b) at
ambient temperature.13 This entails8 a ∆G° value of 1.16
kcal mol-1 (very close to the one they had computed) at
+25 °C which, assuming the invariance of ∆G° with
temperature, corresponds to an amount of 0.9% at a
temperature (e.g., -165 °C) where the ring inversion
process of 2 is likely to be slow in the NMR time scale:
if our computed energy difference (Scheme 1) is used, the
BC proportion at that temperature is expected to be even
higher (2.4%).14 Although quite small, this amount should
not escape detection by 13C NMR, as reported for a
number of analogous biased equilibria.9a,15 It has to be
mentioned, however, that a previous attempt to observe
This hydrocarbon has three possible conformational
minima, corresponding to a twin chair (CC, point group
C2v), boat chair (BC, point group Cs), and twisted twin
boat (BB, point group C2).4-6 All the calculations indicate
that the latter form has too high an energy to be
sufficiently populated and that the CC form corresponds
to the most stable conformer. The energy difference
between the two forms CC and BC is also relatively high;
nonetheless the presence of the latter could be experi-
mentally identified (about 25%) in an electron diffraction
investigation carried out at +400 °C.7 This proportion
entails an energy difference (∆G°) of 2.3 kcal mol-1 which
is, unfortunately, too large a value for detecting this
minor conformer at a temperature low enough to render
the ring inversion process sufficiently slow for NMR
detection. In the equations of ref 7, statistical factors were
introduced to account for the double probability that
conformer BC has to occur with respect to conformer CC,8
and, on this basis, the proportion of the BC conformer is
expected to lie between 2 × 10-3 and 5 × 10-5 in the
temperature range -100° to -170 °C. Such an amount
is clearly too small to be observed or even to affect
significantly the line width of the NMR signals of the
major conformer.9 Accordingly, we did not find any
evidence of a line broadening due to a slow dynamic
equilibrium between these two forms in the 13C spectra
of 1 taken at 100.6 MHz in the range -100° to -170 °C.
(1) Part 75. Gasparrini, F.; Lunazzi, L.; Mazzanti, A.; Pierini, M.;
Pietrusiewicz, K. M.; Villani, C. J . Am. Chem. Soc. 2000, 122, in press
(Ms. J A 9941779).
(2) In partial fulfillment of the requirements for the Ph.D. in
Chemical Sciences, University of Bologna.
(3) Peters, J . A. Synthesis 1979, 321 and references quoted therein.
(4) Allinger, N. L.; Tribble, M. T.; Miller, M. A.; Wertz, D. H. J . Am.
Chem. Soc. 1971, 93, 1637.
(5) Engler, E. M.; Andose, J . D.; Schleyer, P.v. R. J . Am. Chem. Soc.
1973, 95, 8005.
(10) Schneider, H.-J .; Lonsdorfer, M.; Weigand, E. F. Org. Magn.
Reson. 1976, 8, 363.
(11) J aime, C.; Osawa, E.; Takeuchi, Y.; Camps, P. J . Org.Chem.
1983, 48, 4514.
(12) The energies refer to Molecular Mechanics calculations (MMX
force field) as implemented in the program PCModel, Serena Software,
Bloomington, IN.
(6) Osawa, E.; Aigami, K.; Imamoto, Y. J . Chem. Soc., Perkin Trans.
2 1979, 172.
(7) Mastryukov, V. S.; Popik, M. V.; Dorofeeva, O. V.; Golubinskii,
A. V.; Vilkov, L. V.; Belikova, N. A.; Allinger, N. L. J . Am. Chem. Soc.
1981, 103, 1333.
(8) The equation takes the form [BC]/[CC] ) 2 exp (-∆G°/RT)
(9) (a) Anet, F. A. L.; Basus, V. J . J . Am. Chem. Soc. 1973, 95, 4424.
(b) Anet, F. A. L.; Basus, V. J . J . Magn. Reson. 1978, 32, 339. (c)
Okazawa, N.; Sorensen, T. S. Can. J . Chem. 1978, 56, 2737.
(13) Raber, D. J .; J anks, C. M.; J ohnston, M. D., J r.; Raber, N. K.
Tetrahedron Lett. 1980, 21, 677.
(14) A X-ray diffraction study carried out at -173 °C indicates that
2a (CC conformer) is the only species observable in the solid state
(Mora. A. J .; Fitch, A. N. Z. Kristallogr. 1999, 214, 480). Solely the
most stable of the various possible conformers is usually present in
the crystalline state, so that this result agrees with the theoretical
prediction that the CC conformer of Scheme 1 is the one having the
lowest energy.
10.1021/jo000049n CCC: $19.00 © 2000 American Chemical Society
Published on Web 05/10/2000

3564 J . Org. Chem., Vol. 65, No. 11, 2000
Notes
Ta ble 1. 13C Ch em ica l Sh ifts (p p m ) of 2 a t -90 °C a n d
-165 °C, Wh er e th e Tw o Con for m er s CC (2a ) a n d BC (2b)
Wer e Detected in a 96.4 to 3.6 P r op or tion
C9
C1,C5
47.9
47.2 (CC)
45.6 (BC)
C2,C4,C6,C8
C3,C7
21.7
-90 °C
-165 °C
225.7
227.1
-
35.3
34.6 (CC)
21.2 (CC)
29.8 (BC)a
14.9 (BC)b
a
Shift corresponding to the pair of CH2 carbons in the boat
b
arrangement of 2b. Shift corresponding to the single CH2 carbon
in the boat arrangement of 2b.
1) is expected for carbons in a boat10,16 with respect to
those in a chair conformation. The integrated intensities
for the minor methylene lines are 1.8% of the correspond-
ing major partners, but in the BC conformer 2b only half
of the CH2 carbons are in a boat arrangement, the other
half still remaining in the chair arrangement (Scheme
1). The lines for the latter carbons are expected to have
shifts very similar to those of the major CC conformer
2a and cannot be thus observed, being overlapped by the
corresponding intense signals. As a consequence, the
population of the BC conformer 2b is twice as large (i.e.,
3.6%) than the measured proportion.
The smaller line broadening experienced by the CH
line with respect to the two CH2 lines indicates that the
CH shift of the minor BC conformer is much closer to
that of its major partner.17 Indeed such a minor CH line
(Table 1) was barely resolved at -165 °C (Table 1) and
could not be properly integrated. Nonetheless its relative
intensity, with respect to the other two small CH2 lines,
could be reasonably estimated by computer simulation
and was found to be about twice as intense as the line at
14.9 ppm (corresponding to a single CH2 carbon) and
equally intense as the line at 29.8 ppm (corresponding
to two CH2 carbons). Since in conformer BC both the CH
carbons are equivalent, thus yielding a single line, the
factor of 2 in its intensity confirms the proportion of 3.6%
for this conformer.
F igu r e 1. Aliphatic region of the 13C spectrum (100.6 MHz)
of 2 as function of temperature. The inset (50-fold amplified)
of the bottom trace shows two of the methylene lines due to
the minor conformer BC (2b).
Line shape simulation at the temperatures of maxi-
mum broadening yielded rate constants of 130 s-1 (at
-148 °C) and 110 s-1 (at -150 °C) for the line corre-
sponding to two and for that corresponding to four
methylene carbons, respectively. These values translate
into a free energy of activation (∆Gq) equal to 5.7 ( 0.15
kcal mol-1 for the CC to BC interconversion (in the
Eyring equation a transmission coefficient of 1/2 was
employed, to account for the possibility of CC to exchange
with the BC conformer by inverting either of the two
rings18).
The two small CH2 lines were found broader than their
major partners at -165 °C, indicating that even at such
a low temperature the exchange rate is not yet completely
negligible (on further cooling to -170 °C the product
precipitates). The theory of the dynamic equilibrium
between two very biased species predicts, in fact, that
the smaller lines broaden more than the intense ones,9
thus displaying a noticeable width even for quite small
rate constants. In the present case the computer simula-
tion indicated a rate constant of 3 s-1 at -165 °C, which
directly the BC conformer 2b by this method was unsuc-
cessful,10 most likely because the attained temperature
was not sufficiently low.
We obtained the 13C spectrum of 2 at 100.6 MHz in a
2:1 mixture of CHF2Cl and CD2Cl2 as solvent: on
lowering the temperature below -120 °C, the three lines
of the aliphatic carbons (Figure 1) broadened much more
than that of the carbonyl carbon. The lines of the two
CH2 carbons (positions 3, 7) reached a maximum width
(25 Hz) at -148 °C, that of the four CH2 carbons
(positions 2, 4, 6, 8) at -150 °C (22 Hz) and that of the
two CH carbons (positions 1, 5) at -156 °C (12 Hz). On
further lowering, the temperature the lines sharpened
again, despite the increased viscosity, and at the lowest
attained temperature (-165 °C) their widths became 12.5
Hz for both the CH2 lines and 10 Hz for the CH line
(Figure 1). This feature clearly points out to an exchange
between two very biased conformers, and indeed, by
amplifying the vertical scale, two small peaks, lying
upfield to the major CH2 lines, could be clearly detected
(bottom trace, Figure 1). The shift at higher field (Table
(16) (a) Dalling, D. K.; Grant, D. M. J . Am. Chem. Soc. 1974, 96,
1827. (b) Wehrly, F. W.; Marchand, A. P.; Wehrly, S. Interpretation of
13C NMR spectra, 2nd ed.; Wiley: Chichester, New York, 1988; p 61.
(17) The line of the carbonyl carbon does not display appreciable
line broadening effects, thus indicating that the corresponding shifts
in the CC and BC conformers 2a and 2b are very similar.
(15) (a) Anet, F. A. L.; Bradley, C. H.; Buchanan, G. W. J . Am. Chem.
Soc. 1971, 93, 258. (b) Lunazzi, L.; Placucci, G.; Macciantelli, D.
Tetrahedron 1991, 47, 4765. (c) Wiberg, K. B.; Hammer, J . D.; Castejon,
H.; Bailey, W. F.; DeLeon, E. L.; J arret, R. M. J . Org. Chem. 1999, 64,
2085. (d) Lunazzi, L.; Mazzanti, A.; Casarini, D.; De Lucchi, O.; Fabris,
F. J . Org. Chem. 2000, 65, 883.
(18) The use of a unitary coefficient would lead to a ∆Gq ) 5.9 kcal
mol-1
.

Notes
J . Org. Chem., Vol. 65, No. 11, 2000 3565
yields the same ∆Gq value (5.7 kcal mol-1) as that
determined at higher temperatures from the more in-
tense lines.
calibrated by means of a Ni/Cu thermocouple inserted in the
NMR probe (Varian, Mercury 400) before the measurements.
For a meaningful integration of the 13C signals the pulse
sequence suppressing the NOE effect was used, and a long delay
between the transients was employed. Great care had to be
taken in determining the intrinsic line width of 2 which is quite
dependent on the viscosity at these low temperatures. The line
widths were measured above -120 °C and at -165 °C (i.e., in
the range of fast and slow exchange, where the interconversion
process does not affect noticeably the width of the major lines)
and compared with the line width of the solvent. The corre-
sponding ratio was plotted against the temperature and inter-
polated at the temperature of maximum broadening (the ratio
was 2 for the CH2 lines and 1.6 for the CH line at -148, -150
°C). From the knowledge of the solvent line width that of the
investigated line could be therefore deduced. We also checked
that errors on the line width as large as 50% only affected the
value of the barrier by 0.03 kcal mol-1. The main source of error
thus lies in the temperature calibration, since uncertainty of (2
°C (larger in any case than the error of the thermocouple)
affected the barrier by (0.1 kcal mol-1. The line shape simula-
tion was performed by means of a PC version of the DNMR6
program.21
As anticipated,11 the rigidity of the bicyclic frame
makes the ring inversion barrier of 2 higher than that
of cyclohexanone (4.1 kcal mol-1),19 although our mea-
sured value is somewhat lower than the computed one
(7.1 kcal mol-1).11 On the other hand the proportion of
the minor BC conformer 2b is higher than anticipated
by the indirect LIS determination13 as well as by MM
calculations (Scheme 1). From our direct measurement
at low temperature, the value of ∆G° is 0.86 kcal mol-1
which, on the assumption of invariance with tempera-
ture, corresponds to a 31% proportion8 of 2b at ambient
temperature. Given the approximations involved, how-
ever, the agreement between theory and experiment
seems quite satisfactory.
Exp er im en ta l Section
Derivative 1 was obtained by Clemmensen reduction20 of 2,
which is commercially available. The samples for the low
temperatures determinations were prepared by connecting to a
vacuum line the NMR tubes containing the desired compound
dissolved in CD2Cl2, which also served for locking purpose. The
gaseous CHF2Cl was subsequently condensed by using liquid
nitrogen, and the tubes, sealed in a vacuum, were introduced
in the precooled probe of the spectrometer. The temperature was
Ack n ow led gm en t. Thanks are due to the I.Co-
.C.E.A. Institute of CNR (Bologna) for access to the 400
MHz spectrometer. Financial support has been received
from MURST (national project “Stereoselection in Or-
ganic Synthesis”) and from the University of Bologna
(Finanziamento triennale d’Ateneo 2000-2002).
J O000049N
(19) Anet, F. A. L.; Chmurny, G. N.; Krane, J . J . Am. Chem. Soc.
1974, 96, 2929.
(20) Foote, C. S.; Woodward, R. B. Tetrahedron 1964, 20, 687.
(21) QCPE Program 633, Indiana University, Bloomington, IN.