The kinetics, stereochemistry, and deuterium isotope effects in the α-pinene pyrolysis. Evidence for incursion of multiple conformations of a diradical

Article abstract of DOI:10.1016/S0040-4020(02)00676-2

Pyrolysis of optically active α-pinene gave 95% racemic limonene (dipentene), alloocimine, racemic α-pinene, α-pyronene. Activation parameters are reported. Pyrolysis of (S) syn-6-trideuteriomethyl α-pinene at 256.7°C for 2400s gave dipentene with twice as much deuterium as hydrogen transfer with k^H/k^D=1.49 and alloocimine with a Z and E trideuteriomethyl ratio of ca. 5 with k^H/k^D=0.89. The isotope effect on loss of starting material was 1.16. Separation of the enantiomers of α-pinene from 3600s pyrolyses at 256.7°C followed by NMR analysis revealed that the ratio of the R-syn to R-anti to S-anti isomers is 4.6:3.7:1 at roughly two half-lives. Kinetic analysis reveals that the previously proposed mechanism for all conversions involving slow interconversion of two diradicals with C_s symmetry is not consistent with the distribution of the ??-pinene isomers, particularly the formation of more suprafacial-retention product (R-anti) than bond-rotated isomer (S-anti). Inclusion of another C_s species (ignoring the deuteriums) that would be intermediate between the originally proposed C_s species, appears more consistent with the observations.

Full text of DOI:10.1016/S0040-4020(02)00676-2

Tetrahedron 58 (2002) 6943–6950
The kinetics, stereochemistry, and deuterium isotope effects in the
a-pinene pyrolysis. Evidence for incursion of multiple
conformations of a diradical
*
Joseph J. Gajewski, Ilya Kuchuk, Christopher Hawkins and Robert Stine
Department of Chemistry, Indiana University, Bloomington, IN 47405, USA
Received 4 April 2002; accepted 20 June 2002
Abstract—Pyrolysis of optically active a-pinene gave 95% racemic limonene (dipentene), alloocimine, racemic a-pinene, a-pyronene.
Activation parameters are reported. Pyrolysis of (S ) syn-6-trideuteriomethyl a-pinene at 256.78C for 2400 s gave dipentene with twice as
H
D
H
D
much deuterium as hydrogen transfer with k /k ¼1.49 and alloocimine with a Z and E trideuteriomethyl ratio of ca. 5 with k /k ¼0.89.
The isotope effect on loss of starting material was 1.16. Separation of the enantiomers of a-pinene from 3600 s pyrolyses at 256.78C followed
by NMR analysis revealed that the ratio of the R-syn to R-anti to S-anti isomers is 4.6:3.7:1 at roughly two half-lives. Kinetic analysis reveals
that the previously proposed mechanism for all conversions involving slow interconversion of two diradicals with C symmetry is not
s
consistent with the distribution of the a-pinene isomers, particularly the formation of more suprafacial-retention product (R-anti) than bond-
s
rotated isomer (S-anti ). Inclusion of another C species (ignoring the deuteriums) that would be intermediate between the originally proposed
C
s
species, appears more consistent with the observations. q 2002 Elsevier Science Ltd. All rights reserved.
1
. Introduction
concerted pathway is competing with a non-concerted
diradical pathway or whether the overall pathway is totally
diradical with some stereochemical bias resulting from
recombination steps being competitive with bond rotation
that might randomize the stereochemistry. For instance, the
vinylcyclopropane rearrangement occurs with almost, but
a-Pinene has been subjected to pyrolysis long ago and found
to give the retro-ene product limonene, the retro 2þ2
product alloocimine after a 1,5-hydrogen shift in the
initially produced ocimine, as well as racemization
presumably via a 1,3-sigmatropic shift of carbon. Among
the minor products formed in the reaction is a-pyronene,
which is formed by electrocyclization of the Z isomer of
1
1
a,b
alloocimine. Because the limonene produced is racemic,
it has been postulated that homolysis of the C1–C6 bond
gives a diradical which can be responsible for all the
1
c
products (Scheme 1).
It seemed appropriate to examine the stereochemistry and
deuterium isotope effects in the reactions to distinguish
between concerted and diradical pathways and to better
characterize the diradical responsible for dipentene
2
(
racemic limonene). It was also of concern to determine
the stereochemistry of the 1,3-shift of carbon because these
have been a central focus in organic thermal chemistry.
Forbidden (by the Woodward Hoffmann rules) to occur in a
concerted manner if the overall stereochemical pathway is
the suprafacial use of the allylic moiety with retention at the
3
migrating atom. Nonetheless, this pathway often occurs
along with the usually dominant ‘W-H allowed’ suprafacial-
4
inversion pathway. So the question often asked is whether a
Keywords: pyrolysis; alloocimine; dipentene; pinene; 1,3-sigmatropic shift.
*
Corresponding author. Tel.: þ1-812-855-1192; fax: þ1-812-855-8300;
e-mail: gajewski@indiana.edu
Scheme 1. Pyrolysis products of a-pinene and mechanistic rationale.
0040–4020/02/$ - see front matter q 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(02)00676-2

6944
J. J. Gajewski et al. / Tetrahedron 58 (2002) 6943–6950
Table 1. Pyrolysis of (S)-(2)-a-pinene (91.2% ee)—percentages of products
Temperature (8C)
Time (s)
a-Pinene
Limonene
Alloocimine
a-Pyronene
R-a-pinene
S-a-pinene
2
2
2
2
56.7
56.7
41.7
26.7
2400
4800
7800
37.0
13.4
36.9
37.7
35.9
49.2
36.8
37.2
24.1
32.1
23.6
22.6
1.2
1.7
1.0
0.9
17
27
17
15
81
66
81
83
27,000
5
not completely random stereochemistry in the parent case,
however, suprafacial inversion stereochemistry dominates
on an a-cyclodextrin on Celite packed g.c. column while the
former determinations were made on a DB-5 capillary
column from pyrolyses conducted over a 308C temperature
range measured accurately with a Platinum resistance
thermometer. When a sample of (2)-S-a-pinene (91.2%
ee) was pyrolyzed for 2400 s at 256.78C, which represents
roughly 25% enantiomerization, the limonene isolated and
6
the rearrangement of the alpha-tert-butyl material. In cases
where steric effects might reasonably raise the energy of the
suprafacial-inversion pathway over that of the suprafacial-
retention pathway, the ‘W-H forbidden’ pathway is utilized
7
to a great extent.
found to be 99.1% pure had [a ] ¼23.988 (cyclohexane)
D
Further complicating the mechanistic picture is the sugges-
tion that dynamical effects may promote formation of the
which represents roughly 5% preservation of optical purity
considering the extent of racemization of the starting
material. Thus at most, only 5% of the limonene can be
formed by a concerted retro-ene mechanism; the rest of the
retro-ene product, properly termed dipentene, must be
‘
allowed’ suprafacial-inversion product in less substituted
8
cases. The origin of these effects in 1,3-shifts is argued to
be the continuing rotation about a bond after cleavage that
which keeps spinning in one direction until it is stopped by
product formation. Herein we report the full details of
pyrolysis of a-pinene and more recent studies that bear on
the question of the mechanism of the 1,3-shift.
formed from a C symmetric intermediate or equilibrating
s
set of intermediates of effectively C symmetry. All of the
s
data thus far is consistent with the simple mechanism
1
previously postulated involving mostly C1–C6 bond
homolysis to a C symmetric 1,4-diradical followed by
s
either hydrogen atom shift, central bond cleavage, or
recombination to racemic a-pinene (Scheme 1).
2
. Results and discussion
To further define the mechanism of reaction optically pure
(2)-S-syn-6-trideuteriomethyl-a-pinene was pyrolyzed at
It was found that in a well-conditioned 2 L glass bulb
pretreated with dimethyldichlorosilane then diisopropyl-
amine, gas phase pyrolysis of a-pinene proceeded to give
the products previously reported. Table 1 lists the
percentages of the major products determined from multiple
runs at various times and temperatures starting with 99þ%
pure, 91.2% (ee) S-a-pinene.
256.78C for 2400 s, Table 2.
1
c
The rate constant for of loss of starting material allowed
H
D
H
D
determination of k /k ¼1.16. However, k /k for for-
mation of dipentene is 1.49 and that for formation
alloocimine is 0.89. Further, by deuterium NMR twice as
much deuterium relative to hydrogen was transferred to the
six membered ring in the retro ene reaction to form
dipentene, and the Z to E ratio of trideuterio methyl group in
alloocimine is roughly 5:1 by deconvolution of the Z and E
The Arrhenius parameters for loss of racemic a-pinene are
log k (s2 )¼14.2242,700/2.303RT. The Arrhenius para-
meters for the three reactions, retro-ene, retro 2þ2, and
1
2
1
enantiomerization are log k (s )¼13.6241,800/2.303RT,
10
methyl peaks and integration (Scheme 2).
1
4.1243,300/2.303RT, and 14.4245,000/2.303RT, respec-
9
tively, with standard deviations of the order of 0.5% in
every value except for those for the enantiomerization
which are roughly 5%. The latter determination was made
The Z and E methyl proton chemical shifts of alloocimine
were assigned by first identifying the carbon and proton
resonances of each of the four methyl groups by a 2D
correlation experiment. The 6.9 Hz doublet methyl at d 1.74
having a carbon chemical shift of 14.7 ppm must be C8. The
methyl group at d 1.85 with a carbon shift of 28.0 ppm is
assigned C10 attached to C6 because the proton intensity is
the same in both protio alloocimine and in the alloocimine
obtained from pyrolysis of the syn-6-trideuteriomethyl
a-pinene. The methyl group at d 1.80 and a carbon shift
Table 2. Pyrolysis of syn-6-trideuteriomethyl-a-pinene—percentages of
products
Temperature Time a-Pinene Limonene Alloocimine a-Pyronene
(
8C)
(s)
256.7
2400 42.4
25.4
29.1
1.56
Scheme 2.

J. J. Gajewski et al. / Tetrahedron 58 (2002) 6943–6950
6945
s
Scheme 3. C diradical pathway.
of 20.1 ppm was assigned to be C9, the Z methyl group,
because the upfield carbon in such terminal geminal
dimethyl olefinic systems is the Z carbon due to shielding
by the attached chain. Therefore the E methyl group must
give rise to the proton resonance at 1.81 ppm.
pathway, it would have the trideuteriomethyl group syn to
the allylic moiety and be poised to transfer deuterium to
either allylic site to give dipentene (racemic limonene) and
undergo cleavage to ocimine with a Z trideuteriomethyl
group. Competing with these reactions would be rotation of
the radical carbon bearing the geminal dimethyl groups to
1
1
The simplest mechanistic rationalization for all of these
observations (Scheme 3) revolves about initial formation, as
the dominant rate determining step, of a diradical of Cs
symmetry having a half-chair cyclohexenyl moiety with an
axial dimethylcarbinyl radical positioned along the sym-
metry plane. If this species were formed by a least motion
give the stereoisomeric anti C symmetric diradical with the
s
perprotiomethyl group arrayed to transfer hydrogen to either
side of the allylic radical and undergo cleavage to ocimine
with an E-trideuteriomethyl group. This rotation necessarily
must be a slower reaction that directs product formation
from the original diradical. The greater ratio of Z to E
24
Scheme 4. Calculated product fractions from S-a-pinene after 3600 s at 256.78C with the rate constants indicated (in units of 10 /s) assuming five times as
much reaction via the syn diradical as via the anti diradical in H case. The CD values are in parentheses. All other pinene isomers are reacting with the
appropriate rate constants which are not shown for sake of clarity. The calculated percentages after 2400 s at 256.78C provide the following ratios: k /k
3
H
D3
H
D3
H
D3
(
ocimine)¼0.91, found 0.89; k /k (ocimine)¼1.48, found 1.49; Z/E CD
3
ocimine¼4.6, found ,5; D/H transfer (dipentene)¼2.5, found 2.0; k /k (-total
a-Pinene)¼1.24, found 1.16.

6946
J. J. Gajewski et al. / Tetrahedron 58 (2002) 6943–6950
Figure 1. Chromatogram of solution of syn-6-trideuteriomethyl-a-pinene pyrolysate in pentane/cyclohexane obtained at 658C on column filled with
a-cyclodextrin on celite.
trideuteriomethyl ocimine (ca. 5) than that for dipentene
resulting from deuterium rather than hydrogen atom transfer
rate of formation of the suprafacial-retention 1,3-shift
product, R(þ) with an anti CD group, is the same as
3
(
2.0) is not inconsistent with this explanation.
formation of the bond-rotated enantiomer, S(2) with an anti
CD3 group since both should be formed with equal
The activation parameters are consistent with rate determin-
ing formation of a diradical, but more importantly, the
different activation energies for formation of the various
products indicate partitioning of the diradical via steps that
are governed by statistical, i.e. Boltzmann, factors rather
than dynamic ones.
probability from the anti C diradical. It should be noted
s
that the rate constant for formation of the suprafacial-
inversion product, R(þ) with a syn CD group is five times
3
that for formation of the former stereoisomers since it is
formed from the syn C diradical. Also reported in Scheme 4
s
are the kinetic isotope effects and deuterium product ratios
expected at 2400 s, the time period for the determinations.
The agreement is remarkable considering that the only rate
constant changed upon deuteration is that for formation of
the deuterium transferred limonene, and this was adjusted to
reproduce the observed kie for dipentene formation.
Further, the numerical integration scheme does not consider
intermediates, so the larger calculated kie for loss of
a-pinene results from a primary isotope effect on the
formation of the deuterium transferred limonene product,
yet this product must be formed from an intermediate
because it is mostly racemic. Given diradical intermediate
and that reformation of a-pinene is slow compared with
other product forming processes, the rate determining step
should be ring opening which should result in no primary
isotope effect on the loss of a-pinene. There may, however,
be a hyperconjugative secondary isotope effect on the ring
opening of a-pinene which may account for the kie of 1.16.
Nonetheless, at least some of the deuterated a-pinene that
To model the kinetics and isotope effects in system, a
numerical integration program was used that calculates the
mole fractions of stable species as a function of time from
assumed first order rate constants. These rate constants, of
course, must reflect the observed rate constants and the
deuterium distributions observed. However, such programs
cannot easily include unstable species like diradicals
because the large rate constants then require very short
time steps which make difficult achieving the overall
reaction. Nonetheless, the relative utilization of the syn
and anti Cs symmetric diradicals becomes a crucial
consideration. The results from various attempts suggested
that the best choice was 5:1. Scheme 4 depicts the products
derived from the syn and anti C symmetric diradicals, the
s
rate constants and the distribution of products expected at
2
56.78C after pyrolysis of the D -a-pinene for 3600 s
3
(
roughly two half-lives). Further, it was assumed that the

J. J. Gajewski et al. / Tetrahedron 58 (2002) 6943–6950
6947
1
Figure 2. Samples of many H NMR spectra to determine the ratio syn- to anti-CD
of starting material (3).
1
in R (1) and S (2) a-pinene after 60 min pyrolysis at 256.78C and H NMR
3
does not react according to Scheme 4 should be giving the
products from the anti C diradical, namely hydrogen
transferred dipentene and E-trideuteriomethyl ocimine. This
should alter some the ratios calculated in Scheme 4 to more
nearly those observed.
b-pinene, and 80–120 mL samples were pyrolyzed at
256.78C for 3600 s in vacuum in a 2 L glass bulb which
had been ‘conditioned’ with dichlorodimethylsilane. The
samples were removed and separated on a 18 ft£1/4 in. GC
column packed with a-cyclodextrin on 80–100 mesh celite,
Fig. 1.
s
Missing from this analysis is an experimental determination
of the distribution of a-pinene isomers involved in the
racemization reaction, the 1,3-sigmatropic shift. If the
a-pinene isomers were formed from the same two diradicals
postulated above, initially there would be a 5:1:1 distri-
bution of isomers resulting from the three stereopathways:
si, sr, and reformation of the same enantiomer of a-pinene
with interchange of the trideuteriomethyl and perprotio-
methyl groups, i.e. the bond-rotated isomer. The rest of the
a-pinene would, of course, be of the configuration of the
starting material. Interestingly, the ratio of the three
geometric isomers changes only slowly with time according
to the calculation since each isomer is also undergoing the
same reactions as starting material with overall rates that
not only differ only by a small isotope effect on the loss of
a-pinene, but all are substantially faster than racemization.
Indeed, Scheme 4 also includes the calculated amount of the
various a-pinene isomers after roughly two half-lives.
It therefore was important to determine the ratio of the
a-pinene isomers to verify the proposed diradical pathway.
The ratio of syn and anti CD3 methyl groups in each
1
enantiomer could be determined from the H NMR peak
ratios at d 0.85 (syn methyl) and 1.28 (anti methyl). The
ratio of syn to anti in the R isomer was 1.24^0.07 (average
of eight runs), and that in the S isomer was 19^0.6, Fig. 2.
The ratio of S to R a-pinene recovered was determined to be
2.4^0.2 which allowed calculation of the ratio of si to sr to
rotation products of 4.6:3.7:1, respectively. The error in
these values is estimated to be 20% due to incursion of
minor unknown impurities and tailing of the solvent peak in
the GC. Extraordinary means were taken to provide accurate
NMR integrations (see Section 4).
The results of the pyrolyses, GC separation of a-pinene
enantiomers, and NMR analysis of each enantiomer make it
clear that the mechanism of Scheme 3 proposed previously
cannot be correct for the formation of the geometric isomers
of a-pinene. Schemes 3 and 4 require a 4.1:1:1 ratio of
a-pinene stereoisomers from suprafacial-inversion, supra-
facial-retention, and bond-rotated isomer pathways, respec-
tively. The ratio of si to bond-rotated product observed
(4.6:1) might be within experimental error of the calculated
Thus 99% optically pure (1S)-(2)-a-pinene-9-(syn )-d was
3
prepared as described previously from 99% optically pure

6948
J. J. Gajewski et al. / Tetrahedron 58 (2002) 6943–6950
Scheme 5. Mechanistic rationale for the pyrolysis products of a-pinene.
value since it proved difficult to integrate a small peak
relative to a large one in the NMR. Further, the comparison
required good integration of enantiomers from the optically
active GC column, a task that was complicated by base-line
drift due to solvent. However, there is no way to manipulate
the rate constants of Scheme 4 to provide more suprafacial-
retention product without ignoring the symmetry of the
diradical intermediates. The ratio of si to sr product involves
NMR integration of two nearly identical sized peaks so this
value, 1.24, has a standard deviation of only 3%. Thus the
hypothesis of Schemes 3 and 4 is inconsistent with the data
unless there is an additional, separate suprafacial retention
pathway that intervenes to the extent of 25–30% of the total
pathway responsible for geometric isomerization.
ments for the retro 2þ2 reaction to form ocimine and the
ene-like reaction to form the dipentene. It is important to
recognize that both types of diradicals, the syn and anti Cs
species and the S and R C species (in the absence of D) must
s
be formed directly from the appropriate a-pinene. However,
the species S need provide only 1/4 of the a-pinene product.
The mechanistic Schemes 3 and 5 have focused on a
conformation of the diradicals that would place the
dimethylcarbinyl radical group on the axial position in a
half-chair geometry for the cyclohexenyl radical. Unknown
is the extent to which the half-chair undergoes a ‘flip’ to
place the dimethylcarbinyl radical equatorial. However,
where this to occur, no product could be formed from it
recognizing the geometric requirements for formation of
the observed products. Nonetheless, it would randomize the
stereochemistry of the dimethylcarbinyl group and could be
responsible for all products attributed to the S and R and anti
diradicals of Scheme 5 as well as an equal amount of the
products attributed to the syn diradical of Scheme 3. Thus,
up to 50% of the total product might be attributed to this
conformational interconversion with the rest proceeding via
the syn species. However, if this were the case then an equal
amount of sr and bond rotated a-pinene isomers should be
formed, and this is not the case. Perhaps, at best, the extent
of involvement of the equatorial substituted half chair
diradical is measurable by the relative extent of formation of
the bond rotated a-pinene isomer. This is probably less than
10% once there is consideration of the observed reformation
of the S enantiomer of a-pinene.
The simplest scheme required by an extra sr reaction
pathway would be exactly like Scheme 3 with the addition
of an equilibrium pathway connecting the starting a-pinene
and its sr 1,3-shift isomer as well as one connecting the si
1
,3-shift isomer with the bond-rotated isomer. Justification
of such a pathway is not provided by the Principle of
Conservation of Orbital Symmetry which would favor a
suprafacial-inversion pathway for a 1,3-shift of carbon. A
more reasonable mechanistic formulation would recognize
that the sr product could arise from the species responsible
for interconversion of the two diradicals of Scheme 3,
namely, the other C symmetric diradical (ignoring the
s
deuterium labels). This species exists as enantiomers due to
the presence of the CD group; however, one of them, S (in
3
Scheme 5), is more accessible from the a-pinene starting
material of the S configuration via a least motion pathway. If
these latter species were responsible only for additional
a-pinene product and not the other isomeric products, then
Scheme 5 is a possible scenario. That these chiral species,
S and R, could not easily form the other products is not
unreasonable considering the geometry and overlap require-
Finally, it should be noted that the diradicals drawn in
Scheme 5 may not be the actual structures responsible for
products. They may merely represent a set of rapidly
interconverting conformations of the diradicals. Hopefully
detailed quantum mechanical calculations will reveal such

J. J. Gajewski et al. / Tetrahedron 58 (2002) 6943–6950
6949
1
behavior. Lastly, there might be concern that the product
distributions obtained in the pyrolyses are controlled by
dynamic as opposed to statistical factors. The nearly
order of increasing retention times the H NMRs are:
a-pyronene, d 0.86 (d, J¼7.0 Hz, 3H), 0.96 (s, 3H), 0.97 (s,
3H), 1.69 (q, J¼7.0 Hz, 1H), 1.76 (s, 3H), 5.25 (d, J¼
9.4 Hz, 1H), 5.51 (d, J¼5.0 Hz, 1H), 5.72 (dd, J¼9.4,
5.0 Hz, 1H); a-pinene, d 0.84 (s, 3H), 1.17 (d, J¼8.5 Hz,
1H), 1.27 (s, 3H), 1.64–1.70 (m, 3H), 1.90–1.98 (m, 1H),
2.04–2.38 (m, 1H), 2.17–2.20 (m, 1H), 2.20–2.25 (m, 1H),
2.30–2.38 (m, 1H), 5.17–5.22 (m, 1H); dipentene, d 1.40–
1.55 (m, 2H), 1.66 (s, 3H), 1.75–2.20 (m, 5H), 1.73 (s, 3H),
4.7 (m, 2H), 5.41 (d, J¼2.1 Hz, 1H); alloocimine, d 1.75 (d,
J¼7 Hz, 3H), 1.80 (s, 3H), 1.81 (s, 3H), 1.85 (s, 3H), 5.39
(q, J¼7 Hz, 1H), 5.95 (dd, J¼9.4, 1.2 Hz, 1H), 6.35–6.55
(m, 2H).
8
complete formation of racemic limonene would suggest a
diradical intermediate which can partition to both enantio-
mers of limonene argues against dynamical control at least
in the formation of dipentene. Further, there is no evidence
for dynamical control on the ratio of retro-ene to retro 2þ2
to racemized a-pinene products since differences in
activation energies are primarily responsible for the
distribution. Nonetheless, it could be argued that the
a-pinene stereoisomer distribution is dominated by
dynamical factors since it has not be ruled out by a direct
experiment which, unfortunately, would require examining
the stereoisomer distribution as a function of temperature.
A sufficiently precise distribution has been difficult to
obtain, vide infra.
A 0.081 g sample of dipentene obtained from a pyrolysis of
91.2% ee (1S )-(2)-a-pinene at 256.78C for 2400 s was
diluted to 1 mL with cyclohexane. GC analysis showed the
dipentene to be 99.1% pure, and [a] ¼23.988.
2
7
D
3. Summary
4.1.2. Pyrolysis of a-pinene-9(syn )-d . Pyrolysis of
3
1
2
a-pinene-9(syn )-d3 at 256.78C for 2400 s followed by
preparative GC separation provided dipentene and allo-
ocimine. Dipentene DMR (carbon tetrachloride): d 4.67 (s,
4.00D), 1.90 (d, J¼1.4 Hz, 1.92D), 1.64 (s, 2.88D). In the
deuterated alloocimine, the 1.80 ppm resonance line is a
shoulder on the 1.81 ppm signal which, with Gaussian
deconvolution, provides a integral of ca. 1:5, respectively,
but with an uncertainty estimated to be 40%. The proton
resonance lines were assigned by a 2D correlation with the
carbon-13 resonances of the methyls which could be
assigned—see text.
The stereochemistry of the 1,3-shift which thermally
racemizes a-pinene is most easily rationalized by simul-
taneous formation of two extreme conformations of a
diradical. One diradical is responsible for inversion at the
migrating carbon in the 1,3-shift as well as the formation of
racemic dipentene and of ocimine. The other is responsible
for retention at the migrating carbon in the 1,3-shift as well
as for the interconversion of the two stereoisomers of the
other diradical when substituted with both methyl and
trideuteriomethyl groups but not for formation of dipentene
and ocimine due to geometric constraints (see Scheme 5).
4.1.3. Pyrolysis of optically active S(2)-a-pinene-9(syn )-
d3. Pyrolysis of optically pure S(2)-a-pinene-9(syn )-d3
1
2
4
. Experimental
was conducted on 50–100 mL samples at 256.78C which
were removed after 3600 s. The samples were then diluted
with 100 mL of cyclohexane and separated on a GC column
packed with a-cyclodextrin on celite prepared by the
4
.1. General
1
13
H NMR spectra were obtained on a Nicolet 360 MHz and a
Varian VXR 400 (400 MHz) spectrometer. Chemical shifts
are reported in delta units relative to TMS in chloroform-d
solution where chloroform-h is the reference, which is
assumed to be at delta 7.28 ppm. Capillary gas chroma-
tography was performed on a Varian 3700 FID gas
chromatograph equipped with a 30 m£0.25 mm DB-5
fused silica capillary column. A Hewlett Packard 3390A
was used for electronic integrations. Preparative GC was
performed using a Varian 4700 gas chromatograph
equipped with a thermal conductivity detector and a 15 ft,
/4 in. column packed with 20% Carbowax 20 M on
0/80 mesh Chromosorb P. Starting 1S-(2)-b-pinene
99%) and other reagents were purchased from commercial
sources and used without further purification. The solvents
used were dried and distilled by standard procedures. All
glassware for addition reactions was flame-dried under
method of Sybilska and Koscielski as follows: a-Cyclo-
dextrin hydrate (4.77 g) (Aldrich) was suspended in 10 mL
of distilled water, then mixed with 11.7 g freshly distilled
formamide and stirred. Distilled water was added until the
a-cyclodextrin completely dissolved. The solution was
added slowly to 45 g 80–100 mesh celite (Alltech Gas
chrom S) in a 250 mL flask under vigorously shaking. After
being subjected to rotatory evaporation at 508C overnight
under aspirator vacuum, the solid was used to pack an
18 ft£1/4 in. copper column which was conditioned over-
night at 708C. GC separation was accomplished by injecting
70 mL samples of the pyrolysate in cyclohexane into the
column which was operated at 658C. The injector
temperature was 1608C and the detector temperature was
2008C. Near baseline separation between the pinene
enantiomers was achieved (at 20c He flow, the retention
times were 27 and 46 min for the R and S enantiomers,
respectively; however, the peaks were broad).
1
6
(
vacuum with an N purge.
2
4
.1.1. Pyrolyses of a-pinene. ‘Gold label’ (1S)-(2)-a-
The ratio of enantiomers was achieved by weighing the
peaks in the chromatogram, Fig. 1.
pinene (Aldrich) was distilled from lithium aluminum
hydride under nitrogen and used directly. The pyrolysis
apparatus has been described previously. Analyses were
performed by capillary GC and reported in Table 1. The
products were separated on the Carbowax 20 M column. In
6
4.2. NMR analysis of the enantiomers
The distribution of anti and syn deuterium labeled isomers

6950
J. J. Gajewski et al. / Tetrahedron 58 (2002) 6943–6950
was determined by integration of methyl singlets at 1.28 and
.85 ppm, which are the anti and syn methyls, respectively,
2. A preliminary report on this work can be found in Gajewski,
J. J.; Hawkins, C. M. J. Am. Chem. Soc. 1986, 108, 838.
3. Woodward, R. B.; Hoffmann, R. Angew. Chem., Int. Ed. Engl.
1969, 8, 781.
0
Fig. 2. Spectra were phased, but not base-line corrected. In
the case of overlapping impurity peaks, the integration value
of each peak was determined by deconvolution or by
subtraction of spectra of solvents. The integration results
were compared with values obtained after applying
resolution-enhancing filter (combination of gaussian,
shifted gaussian, exponential, sine, sine squared, and shifted
sine functions) to the Free Induction Decay.
4. Gajewski, J. J. Hydrocarbon Thermal Isomerizations; Peri-
cyclic Reactions, Marchand, A. P., Lehr, R. E., Eds.;
Academic: New York, 1980; Vol. 1. Chapter 1.
5. Baldwin, J. E.; Villarica, K. A.; Freedberg, D. I.; Anet, F. A. L.
J. Am. Chem. Soc. 1994, 116, 10845. In a perhaps more
relevant example, the stereochemistry of the vinylcyclobutane
rearrangement shows little preference for orbital symmetry
control in a system where methyls are stereolabels on the
migrating carbon and the migration terminus: Baldwin, J. E.;
Burrell, J. E. J. Org. Chem. 2002, 67, 3249.
Acknowledgments
We thank the National Science Foundation and the
Department of Energy for support of this work and Dr
Ulrike Werner-Zwanziger for her NMR integration
expertise.
6
. Gajewski, J. J.; Olson, L. P. J. Am. Chem. Soc. 1991, 113,
432. Gajewski, J. J.; Olson, L. P.; Willcott, M. R., III. J. Am.
7
Chem. Soc. 1996, 118, 299.
7
. Berson, J. A.; Nelson, G. L. J. Am. Chem. Soc. 1970, 92, 1096.
. Carpenter, B. K. J. Am. Chem. Soc. 1995, 117, 6336.
Carpenter, B. K. Acc. Chem. Res. 1992, 25, 520.
. These values are altered slightly from those reported in Ref. 2
due to better data reduction techniques.
8
References
9
1
. (a) Conant, J. B.; Carlson, G. H. J. Am. Chem. Soc. 1927, 51,
464. By hydrogenation, dipentene product was postulated to
1
0. This value was reported in Ref. 2 to be more than 9:1 based on
peak heights. Subsequent studies suggest the smaller value of
3
be responsible for the thermally induced loss of optical activity
of a-pinene. Subsequent work established that, in addition,
a-pinene undergoes racemization albeit slower than formation
of dipentene. (b) Fuguitt, R. E.; Hawkins, J. E. J. Am. Chem.
Soc. 1945, 67, 242. Fuguitt, R. E.; Hawkins, J. E. J. Am. Chem.
Soc. 1947, 69, 319. (c) Burwell, R. L. J. Am. Chem. Soc. 1951,
5
but with substantial error—see Section 4.
1. Bellamy, A. J.; Crilly, W J. Chem. Soc., Perkin Trans. 2 1973,
22We thank J. K. Crandall for suggesting this experiment.
2. Borden, W. T.; Lee, J. G.; Young, S. D. J. Am. Chem. Soc.
980, 102, 4841.
3. Koscielski, T.; Sybilska, D.; Jurczak, J. J. Chromatogr. 1983,
80, 131.
1
1
1
1
1
73, 4461. (d) Riistoma, K.; Harva, O. Finn. Chem. Lett. 1974,
2
132.