Mechanistic and kinetic insights into the thermally induced rearrangement of α-pinene

Article abstract of DOI:10.1021/jo8012995

(Chemical Equation Presented) The thermal rearrangement of α-pinene (1) is interesting from mechanistic as well as kinetic point of view. Carrier gas pyrolyses with 1 and its acyclic isomers ocimene (2) and alloocimene (3) were performed to investigate the thermal network of these hydrocarbons. Kinetic analysis of the major reaction steps allows for a deeper insight in the reaction mechanism. Thus it was possible to explain the racemization of 1, the formation of racemic limonene (4), and the absence of the primary pyrolysis product 2 in the reaction mixture resulting from thermal rearrangement of 1. Results supported the conclusion that the reactions starting with 1 involve biradical transition states.

Full text of DOI:10.1021/jo8012995

Mechanistic and Kinetic Insights into the Thermally Induced
Rearrangement of r-Pinene
Achim Stolle,*,† Bernd Ondruschka,† and Matthias Findeisen‡
Institute for Technical Chemistry and EnVironmental Chemistry, Friedrich-Schiller UniVersity Jena,
Lessingstr. 12, D-07743 Jena, Germany, and Institute for Analytical Chemistry, UniVersity Leipzig,
Linne´str. 3, D-04103 Leipzig, Germany
achim.stolle@uni-jena.de
ReceiVed June 17, 2008
The thermal rearrangement of R-pinene (1) is interesting from mechanistic as well as kinetic point of
view. Carrier gas pyrolyses with 1 and its acyclic isomers ocimene (2) and alloocimene (3) were performed
to investigate the thermal network of these hydrocarbons. Kinetic analysis of the major reaction steps
allows for a deeper insight in the reaction mechanism. Thus it was possible to explain the racemization
of 1, the formation of racemic limonene (4), and the absence of the primary pyrolysis product 2 in the
reaction mixture resulting from thermal rearrangement of 1. Results supported the conclusion that the
reactions starting with 1 involve biradical transition states.
SCHEME 1. Reaction Products Resulted from Thermal
Isomerization of r-Pinene (1)
Introduction
The thermal isomerization of R-pinene (1) is a long-known,
deeply investigated reaction.1 First studies were published in
the 1930s by Smith and Carlson, who independently observed
racemization of the pyrolysis products yielded by thermal
treatment of 1.2 The authors believed that this was due to the
formation of racemic limonene (4, dipentene; Scheme 1). More
intensive studies revealed that beside the rearrangement of 1 to
alloocimene (3) and 4 the pyrolysis of 1 yields nonconverted
R-pinene with a lower optical activity as found for the starting
material.2-5 Apparently, the initial formation of a biradical was
† Friedrich-Schiller University Jena.
‡ University Leipzig.
assumed as explanation of these observations (Scheme 1).6 Due
to the fact that this mechanism has to lead to the formation of
ocimene (2) and the absence of this product in the pyrolysis
mixtures was proved in different studies, a fast isomerization
of 2 to 3 was proposed by various authors.5-7 R-Pinene (1) has
been intensively subjected to kinetic studies in both gaseous
(1) (a) Egloff, G.; Herrman, M.; Levinson, B. L.; Dull, M. F. Chem. ReV.
1934, 14, 287–383. (b) Frost, A. A.; Pearson, R. G. Die Thermische Isomerisation
von R- und ꢀ-Pinen. In Kinetik und Mechanismus homogener chemischer
Reaktionen; Verlag Chemie: Weinheim, 1964; pp 349-354. (c) Banthorpe, D. V.;
Whittaker, D. Quart. ReV. 1966, 20, 373–387. (d) Naves, Y. R. Russ. Chem.
ReV. 1968, 37, 779–788.
(2) (a) Smith, D. F. J. Am. Chem. Soc. 1927, 49, 43–50. (b) Conant, J. B.;
Carlson, G. H. J. Am. Chem. Soc. 1929, 51, 3464–3469.
8228 J. Org. Chem. 2008, 73, 8228–8235
10.1021/jo8012995 CCC: $40.75 2008 American Chemical Society
Published on Web 09/30/2008

Thermally Induced Rearrangement of R-Pinene
and condensed phase,8-10 but the consecutive reactions of 2
and 3 are rarely investigated.9,11 Deuterium-labeling experiments
performed by Gajewski and co-workers support the hypothesis
that a biradical is responsible for the formation of the primary
pyrolysis products (2, 4, racemic 1) resulting from rearrangement
of optically active 1.10 Recently, studies were published showing
that isomerization of 1 can be carried out in supercritical alcohols
without lost of the initial optical activity of 1 but accompanied
by the formation of dipentene.12-14
Although many results have been published in the literature
concerning the thermal isomerization of 1 already, it seems to
be appropriate to investigate the pyrolysis once more from a
mechanistic and kinetic point of view. Performing pyrolysis
experiments with 1, 2, and 3 under identical conditions allows
for a detailed study of the mechanisms of interconversions of
these compounds. The calculation of activation parameters
according to Arrhenius and Eyring on the basis of kinetic
pyrolyses forces the understanding of the transition states the
reactions pass through. Performing the pyrolysis experiments
in a flow system at ambient pressure using nitrogen as diluting
agent and carrier gas allows for calculation of rate constants
apart from the falloff region.
FIGURE 1. Yield of pyrolysis products of R-pinene pyrolysis and
content of remaining R-pinene (1) depending on reaction temperature
(T). 2, ocimene; 3, alloocimene; 4, limonene; cp, consecutive products.
15 µL starting material; carrier gas, N2; flow rate, 1.0 L min-1; τ,
0.51-0.71 s.
within the temperature range of 350-500 °C (Figure 1).
Therefore, it can be concluded that the reaction pathways starting
from biradical intermediate sketched in Scheme 1 are not
affected by changes in reaction temperature, which is according
to Carpenter not atypical for reactions passing through a joint
transition state or reaction intermediate.15 Subjecting limonene
(4) to pyrolysis studies at temperatures higher than 550 °C
yielded aromatized products.16,17
The influence of reaction temperature on the enantiomeric
ratio of both 1 and 4 was subject of these studies also. Pyrolysis
experiments with optically pure (1S,5S)-(-)-1 and (1R,5R)-(+)-1
were conducted, whereby the initial enantiomeric excess (ee)
values were 80% and 95%, respectively. Temperature plots of
the corresponding yields of both enantiomers of 1 for pyrolysis
of either (-)-1 and (+)-1 are pictured in Figure 2a and b,
respectively, clearly indicating that pyrolyses at elevated tem-
peratures lead to a racemization of the starting materials.
Contrary to previous studies on the loss of optical activity of 1
while heating,2 a complete racemization was not observed in
both cases, because of the fact that complete conversion was
reached first (425 °C). Additionally, the enantiomeric ratio of
optically active pyrolysis product 4 was calculated based on
GC runs using the same ꢀ-cyclodextrine column as used for
the determination of relative enantiomeric yields in the case of
1. Results shown in Figure 2 for the pyrolysis of both
enantiomers of 1 reveal that in both cases a racemic mixture of
(-)-4 and (+)-4 was formed.
Results and Discussion
Pyrolysis of r-Pinene. In order to unravel the complex
product mixture resulting from thermal rearrangement of
R-pinene (1), flow-type pyrolysis experiments at ambient
pressure were conducted with 1 and its main isomerization
products ocimene (2) and alloocimene (3). Comparative chro-
matograms for these experiments are given in Supporting
Information. The plot of conversion of 1 and of the yields of
pyrolysis products versus pyrolysis temperature (T) is depicted
in Figure 1. With respect to the amounts of products initially
formed from 1 (2, 4) significant differences were found. Whereas
4 is formed with a selectivity of approximately 45%, the yield
of the acyclic primary pyrolysis product 2 does not exceed 2%,
pointing out that 2 is a reactive molecule that rapidly rearranges
to 3. At a pyrolysis temperature of 412 °C the yield of 3 passes
through a maximum and further increase of temperature gives
rise to the formation of consecutive products (cp, e.g. py-
ronene).10 Decomposition and aromatization products of py-
ronenes constitute the main fraction resulting from experiments
at temperatures higher than 525 °C. Interestingly, the ratio of
primarily formed products 4 and 2 (including consecutive
reaction products such as 3 or cp) practically remains constant
Pyrolysis of Ocimene. Pyrolysis experiments with a mixture
of 3Z- (2a) and 3E-ocimene (2b; Chart 1) were conducted to
unravel the products formed from their thermal treatment. GC
analyses showed that the ratio of the two stereoisomers of 2
was approximately 0.5 in favor of 2b. The same results were
found in the 1H NMR spectrum of the mixture. Structures were
confirmed by one-dimensional NOESY experiments explained
in Supporting Information. Explorative pyrolyses investigating
the temperature dependency of the conversion of 2 reveal
differences in the thermal behavior of 2a and 2b pictured in
Figure 3. Whereas the Z-isomer (2a) is completely converted
at a pyrolysis temperature of 375 °C, 2b is thermally stable up
to 525 °C, thus indicating that pyrolysis of 1 mainly yields 2a.
(3) (a) Goldblatt, L. A.; Palkin, S. J. Am. Chem. Soc. 1941, 63, 3517–3522.
(b) Savich, T. R.; Goldblatt, L. A. J. Am. Chem. Soc. 1945, 67, 2027–2031.
(4) Fuguitt, R. E.; Hawkins, J. E. J. Am. Chem. Soc. 1945, 67, 242–245.
(5) Hawkins, J. E.; Hunt, H. G. J. Am. Chem. Soc. 1951, 73, 5379–5381.
(6) Burwell, R. L. J. Am. Chem. Soc. 1951, 73, 4461–4462.
(7) Hawkins, J. E.; Burris, W. A. J. Org. Chem. 1959, 24, 1507–1510.
(8) (a) Fuguitt, R. E.; Hawkins, J. E. J. Am. Chem. Soc. 1947, 69, 319–322.
(b) Hunt, H. G.; Hawkins, J. E. J. Am. Chem. Soc. 1950, 72, 5618–5620. (c)
Hawkins, J. E.; Vogh, J. W. J. Phys. Chem. 1953, 57, 902–905.
(9) Riistama, K.; Harva, O. Finn. Chem. Lett. 1974, 4, 132–138.
(10) (a) Gajewski, J. J.; Kuchuk, I.; Hawkins, C. M.; Stine, R. Tetrahedron
2002, 58, 6943–6950. (b) Gajewski, J. J.; Hawkins, C. M. J. Am. Chem. Soc.
1986, 108, 838–839.
(11) Sasaki, T.; Eguchi, S.; Yamada, H. Tetrahedron Lett. 1971, 12, 99–
103.
(12) Anikeev, V. I.; Yermakova, A.; Chibiryaev, A. M.; Kozhevnikov, I. V.;
Mikenin, P. E. Russ. J. Phys. Chem. A 2007, 81, 711–716.
(13) Chibiryaev, A. M.; Yermakova, A.; Kozhevnikov, I. V.; Sal’nikova,
O. I.; Anikeev, V. I. Russ. Chem. Bull. Int. Ed. 2007, 56, 1234–1238.
(14) Yermakova, A.; Chibiryaev, A. M.; Kozhevnikov, I. V.; Anikeev, V. I.
Chem. Eng. Sci. 2007, 62, 2414–2421.
(15) Carpenter, B. K. J. Am. Chem. Soc. 1985, 107, 5730–5732.
(16) Stolle, A.; Brauns, C.; Nu¨chter, M.; Ondruschka, B.; Bonrath, W.;
Findeisen, M. Eur. J. Org. Chem. 2006, 3317–3325.
(17) Pines, H.; Ryer, J. J. Am. Chem. Soc. 1955, 77, 4370–4375.
J. Org. Chem. Vol. 73, No. 21, 2008 8229

Stolle et al.
FIGURE 2. Conversion of R-pinene (X), relative content of (-)-R-pinene [(-)-1], relative content of (+)-R-pinene [(+)-1], and ratio of (-)- and
(+)-limonene (4) depending on the reaction temperature (T) for the pyrolyses of (-)-1 (Figure 2a) and (+)-1 (Figure 2b). 15 µL starting material;
carrier gas, N2; flow rate, 1.0 L min-1; τ, 0.60-0.70 s; enantiomeric ratios measured with a ꢀ-cyclodextrine column.
CHART 1. Structures of 3Z-Ocimene (2a), 3E-Ocimene
(2b), 4E,6Z-Alloocimene (3a), and 4E,6E-Alloocimene (3b)
allowed for structural elucidation of the two isomers (cf.
Supporting Information).
Pyrolysis of Alloocimene. Pyrolysis experiments with a
mixture of 3a and 3b were conducted in order to investigate
content of ocimene isomers depending on reaction temperature (T).
FIGURE 3. Yield of pyrolysis products of ocimene pyrolysis and
the thermal behavior of these two diastereomers (3a/3b: 9.0).
As described, the pyrolysis of 2a yields those two isomers in
favor of 3a. Figure 4 reports the dependency between reaction
temperature and the molar ratio of 3a and 3b, revealing that
both were converted into other C10H16 hydrocarbons (cp; e.g.,
pyronenes).10 Because of the fact that aromatization and
decomposition of those consecutive products occurred at tem-
peratures higher than 525 °C, this region was not investigated
in detail.18 Contrary to 3a, the content of 3b seems to increase
slightly until 450 °C and after passing through this maximum
decreases in the same way as observed for the 4E,4Z-isomer
3a, thus indicating that isomerization of 3a to the other isomer
takes place.
2a, 3Z-ocimene; 2b, 3E-ocimene; 3a, 4E,6Z-alloocimene; 3b, 4E,6E-
alloocimene; cp, consecutive products. 15 µL starting material; carrier
gas, N2; flow rate, 1.0 L min-1; τ, 0.51-0.78 s.
Pyrolysis experiments with 1 and its acyclic pyrolysis
products 2 and 3 reveal significant differences with respect to
the temperature region they are converted. Comparison of these
results allows for the conclusion that 2a is the most reactive
isomer and 3 the most thermally stable C10H16 hydrocarbon.
Results confirm the findings that 2a is only present in very low
concentrations in the mixture of pyrolysis products yielded from
thermal isomerization of 1.4,5,9,10,19
Elucidation of Kinetic Steps in the Pyrolysis of r-Pinene.
R-Pinene (1) as well as its acyclic main isomerization products
3Z-ocimene (2a), 4E,6Z-alloocimene (3a), and 4E,6E-alloo-
cimene (3b) were subjected to kinetic studies. The simplified
reaction scheme used for calculation of the rate constants ki is
depicted in Scheme 2. Explorative pyrolysis studies revealed
that 1 primarily isomerizes to give 2a and 4, whereby minor
amounts of 3E-ocimene (2: < 2.5%; Figure 1) were neglected
FIGURE 4. Content of alloocimene isomers and yield of its pyrolysis
products depending on reaction temperature (T). 3a, 4E,6Z-alloocimene;
3b, 4E,6E-alloocimene; cp, consecutive products. 15 µL starting
material; carrier gas, N2; flow rate, 1.0 L min-1; τ, 0.51-0.71 s.
Therefore the low yields reported for 2 (Figure 1) in the case
of the pyrolysis of 1 have to be assigned to the formation of
2b, because at those temperatures the Z-isomer is not present
any more. Additionally the temperature dependence of the yields
of pyrolysis products of 2 are shown, indicating that alloocimene
(3) is formed in two stereoisomeric forms, 4E,6Z- (3a) and
4E,6E-alloocimene (3b, Chart 1), whereby 3a is the main
product. After passing through a maximum at 375 °C the yield
of 3a decreases in favor of the formation of consecutive products
(cp), whereas the amount of 3b formed from 2a remains
constant. Structures of the two alloocimenes were confirmed
by comparison of retention times and mass spectra with those
(18) (a) Parker, E. D.; Goldblatt, L. A. J. Am. Chem. Soc. 1950, 72, 2151–
2159. (b) Crowley, K. J.; Traynor, S. G. Tetrahedron Lett. 1975, 16, 3555–
3558. (c) Crowley, K. J.; Traynor, S. G. Tetrahedron 1978, 34, 2783–2789.
(19) Pascual Teresa, J. de.; Sanchez Bellido, I.; Alberdi Albistegui, M. R.;
Feliciano, A. san.; Grande Benito, M. Anal. Quim. 1978, 74, 301–304; CAN
89:163752.
1
of an authentic mixture of 3a and 3b. H NMR spectra in
combination with NOE-experiments with this authentic sample
8230 J. Org. Chem. Vol. 73, No. 21, 2008

Thermally Induced Rearrangement of R-Pinene
TABLE 2. Rate Constants for the Formation of Ocimene (k2) at
350 °C Calculated from the Arrhenius Activation Parameters
SCHEME 2. Simplified Reaction Network Used for
Calculation of Rate Constants k1-7
Ea
log A k2
reaction conditions
(kJ mol-1) (s-1) (s-1
)
ref
glass ampoules, liquid phase
glass ampoules, liquid phase, N2
batch reactor, gas phase, vacuum
178.6 15.6 4.261 8a
171 13.30 0.093 9
181.3 14.1 0.080 10
TABLE 1. Kinetic Dataa for the Gas-Phase Isomerization
Reactions of r-Pinene (1; k1-3), 3Z-Ocimene (2a; k4), and
Alloocimene (3; k5-7; cf. Scheme 2)
flow-type reactor, supercritical ethanola 136.6
14, 24
flow-type reactor, N2
178.5 14.11 0.137 this work
a Activation energies calculated on the basis of a model of the reactor
used for the experiments.
Ea (kJ mol-1
)
log A (s-1
)
∆Hq (kJ mol-1
)
∆Sq (J K-1 mol-1
)
k1
k2
k3
k4
k5
k6
k7
170.0 ( 1.7 13.70 ( 0.14
178.5 ( 1.8 14.11 ( 0.15
160.8 ( 2.4 12.61 ( 0.19
165.2 ( 1.6
173.7 ( 1.8
155.9 ( 2.3
120.8 ( 2.1
175
144.9 ( 7.3
236.4 ( 7.6
3.0 ( 0.04
11.0 ( 0.2
-17.7 ( 0.4
-42.4 ( 1.0
-23
-46.6 ( 3.2
74.3 ( 2.9
movement.20 Hence, “normal” transition states are reported to
be typical for fragmentations of four-membered rings the results
found herein for k1 are in accordance to those from studies
investigating the thermal decomposition of cyclobutane de-
rivatives.20-23 Similar conclusions can be drawn for the
formation of 2a from 1 (k2; Table 1), but the positive activation
entropy in this case is an indicator for a “loose” transition state
being typical for biradical intermediates the reaction passes
through. Therefore, it seems to be appropriate assuming that
both reactions k1 and k2 are reactions involving biradical
intermediates. A high Arrhenius factor and an activation entropy
of 74 J K-1 mol-1 were found for the reaction describing the
disappearance of 3b (k7) being an unequivocal indicator for the
fact that this reaction passes through (a) radical reaction
intermediate(s). In contrast to 3b the activation parameters for
the disappearance of the 4E,6Z-isomer of 3a (k6; Table 1) are
typical for reactions passing through rigid six-membered ring
transition states. These are characterized by a lower degree of
freedom for torsional and/or rotational movement.20 Therefore
it can be concluded, that beside k7 the formation of 4 from 1
(k3), the isomerization of 2a to 3a (k4), and the reaction of 3a
leading to 3b (k5) are reactions passing through six-membered
ring transitional states (e.g., [1,5]H-shift, ene reaction).
The thermal isomerization reaction of 1 was subject of various
kinetic pyrolysis studies performed in gaseous, condensed, or
supercritical phase.5,8a,9,10,14,24 Comparison of the kinetic data
presented in Table 1 with those reported in the literature reveals
a high accordance. Although the network of reactions based on
pyrolysis of 1 has been subjected to kinetic studies since the
1950s, there is only one study investigating the isomerization
of 1 in the gas phase using a flow-type reactor.5 Unfortunately,
this study reported data for the isomerization of 2a leading to
3 (k4) only. Nevertheless, these results are in agreement with
those reported herein and found for pyrolysis experiments in
the liquid phase.9,11 Respective rate constants for the formation
of 2a from 1 (k2) on the basis of the reported Arrhenius
parameters for 350 °C are listed in Table 2. Only the studies of
Fugitt and Hawkins performed in glass ampoules under noninert
conditions led to a higher rate constant, which might indicate
that the reaction is not independent from pressure under those
conditions applied. This fact seemed to be proveed with respect
to the studies of Anikeev and co-workers using supercritical
ethanol in a flow system as fluid.14,24 Activation energies, Ea,
125.72.1
181
11.31 ( 0.18
b
12.4
150.9 ( 7.4 11.18 ( 0.54
242.4 ( 7.6 17.49 ( 0.55
a Error limits are 95% certainty limits. b Based on experiments at only
two different temperatures.
due to its thermal stability within the temperature range
investigated (250-525 °C; Figure 3). It has to be pointed out
that the network presented in Scheme 2 does not include the
reaction(s) leading to racemized 1. Consideration of the gas-
phase racemization of both enantiomers of 1 within the
simplified model described in Scheme 2 is difficult from various
points of view. As reported in Figure 2, both (-)-1 and (+)-1
undergo beside their racemization isomerization reactions lead-
ing to 2 and racemic 4 as primary pyrolysis products with
selectivities being independent from the enantiomer used. This
allows for the conclusion that both molecules take part in the
isomerization reaction, and therefore it seems to be impossible
to combine the kinetics for rearrangement and racemization.
Combination of both processes would lead to rate equations
with feedback routes because if molecules of (+)-1 are formed
from the respective (-)-enantiomer those can either undergo
isomerization or can be reconverted to give (-)-1.
Arrhenius as well as Eyring activation parameters were
calculated according to eqs 1 and 2, respectively. Pyrolysis
experiments with pure 1 and with mixtures of diastereomeric 2
and 3 were conducted to determine the rate constants k for the
disappearance of 1 (k1), 2a (k4), 3a (k5, k6), and 3b (k7). On the
basis of these data it was possible to calculate the other rate
constants using the model of competitive parallel first-order
reactions. Activation data are summarized in Table 1. Hence,
in the case of calculation of k5, experiments at only two different
temperatures were able to be considered and no error limits are
given. Rate constants and Arrhenius plots are given in Sup-
porting Information.
kT ) A · e- Ea
(1)
(2)
RT
∆Hq-T∆Sq
kBT
h
kT )
· e-
RT
Consideration of log A together with ∆Sq allows for charac-
terization of the transition states the reactions presumably pass
through. Typical parameters for “normal” transition states are
approximately 13 in log A and activation entropies of around 0
J K-1 mol-1 as found for k1 describing the disappearance of 1.
Literature assumes that those transition states have a similar
geometry as the ground-state geometry of the starting molecule
with minor differences in torsional and rotational degrees of
(20) (a) Gowenlock, B. G. Quart. ReV. 1960, 14, 133–145. (b) Walsh, R.
Chem. Soc. ReV. 2008, 37, 686–698.
(21) Wellman, R. E.; Walters, W. D. J. Am. Chem. Soc. 1957, 79, 1542–
1546.
(22) Frey, H. M.; Walsh, R. Chem. ReV. 1969, 69, 103–124.
(23) Chickos, J. S.; Frey, H. M. J. Chem. Soc., Perkin Trans. 2 1987, 365–
370.
(24) Yermakova, A.; Chibiryaev, A. M.; Kozhevnikov, I. V.; Anikeev, V. I.
J. Supercrit. Fluids 2008, 45, 74–79.
J. Org. Chem. Vol. 73, No. 21, 2008 8231

Stolle et al.
SCHEME 3. Reaction Network of r-Pinene Thermal
Isomerization
calculated on this basis showed significant differences from all
other values reported in Table 2. Obviously, other reaction
mechanisms are present when the reaction is carried out in a
high-pressure regime. Differences of the rate constant calculated
on the basis of the vacuum pyrolysis studies by Gajewski and
co-workers (0.0013 mbar, static system) may be due to the fact
that k are not pressure-independent anymore.10 In general rate
constants were observed to be generally independent from total
pressure at pressure greater than 1 atm, whereas at lower
pressures the rate constants “fall off” and the activation
mechanism is changed. Reactions performed under (V)LPP-
conditions without an additional fluent have a different activation
procedure, because the collisions between reactant molecules
can not be ignored. Therefore, often the rate constant of the
rate-determining step is of second-order for the falloff region
rather than (pseudo)-first-order.25 Activation parameters avail-
able from literature for the disappearance of 39,10 indicate that
mostly 3a is formed from 2a, thus being in agreement with the
pyrolysis studies on 2 and 3 presented in Figures 3 and 4.
Pyrolysis Mechanisms of r-Pinene, Ocimene, and Alloo-
cimene. Pyrolysis experiments as well as kinetic analysis of
the most important reactions in the thermal reaction network
based on 1 allow for mechanistic evaluation of the initial
reaction steps. Pyrolyses with 1 (Figure 1) reveal that the ratio
of primary formed pyrolysis products 2 and 4 remains constant
while increasing temperature and conversion of 1 from 300 to
500 °C or 0 to 100%, respectively. This behavior is reported to
be not atypical for reactions passing through the same inter-
mediate or transition state.15 Additionally, enantiomerically pure
(-)-1 as well as (+)-1 were subjected to pyrolysis studies in
order to investigate the temperature dependency of the enan-
tiomeric excess (Figure 2), revealing that racemization occurred
independently from the enantiomer used as starting material.
This allows for the conclusion that three major reactions
contribute to the disappearance of 1 during its thermally induced
rearrangement: racemization of 1, fragmentation of the cyclobu-
tane ring (2, 3, cp), and the formation of dipentene (4).10 Those
set of reactions are reported to be typical for isomerization of
vinylcylobutanes.23,26 According to the theoretical study of
Northrop and Houk on vinylcyclobutane rearrangement, only
the syn-conformer has the ability to undergo [1,3] carbon shift
to cyclohexene and fragmentation to cisoid-butadiene and
ethylene, whereas the anti-isomer formed transoid-butadiene
and ethylene only. Additionally cyclohexene is able to form
butadiene and ethylene via retro-Diels-Alder reaction.26g Hence,
1 is a bridged bicyclic vinylcyclobutadiene with cis-orientation
both reaction pathways are virtually possible. Kinetic analysis
of the reactions responsible for disappearance of 1 (k1) and the
formation of 2a (k2) and 4 (k3; cf. Scheme 2 and Table 1) allows
for the conclusions that k1 and k2 pass through “loose” transition
states, whereas the reaction leading to product 4 runs through
a six-membered ring transition state. Since [2 + 2]-cycloaddi-
tions as well as [2 + 2]-cycloreversions are thermally forbidden
reactions with respect to the Woodward-Hoffmann rules,27
formation of 2a has to proceed stepwise including a biradical
intermediate depicted in Scheme 3.6,10,28
On the basis of the initially formed biradicals, three possible
reaction pathways exist leading to the formation of products
2a, 4, and to racemized 1. Reconnection of the bond initially
been broken in the case of TS1a leads to 1 with the initial
configuration, whereas bond reformation in TS1b yields the
opposite enantiomer. This racemization reaction can be classified
as 1,3-sigmatropic rearrangement reaction. Literature reports
many examples for similar thermally as well as photochemically
induced isomerizations.26d,e,29
Analysis of the resulted pyrolysis products from thermal
isomerization experiments of both enantiomers of 1 reveals that
racemic 4 (dipentene) is formed independently from temperature,
conversion of 1, and from the initial enantiomeric excess (Figure
2). Obviously the mesomeric delocalization of the initially
formed biradical is fast compared to its consecutive reactions,
implying that the change of hybridization of C(1) is not rate-
(27) (a) Woodward, R. B.; Hoffmann, R. Angew. Chem., Int. Ed. 1969, 8,
781–853. (b) Hoffmann, R.; Woodward, R. B. Science 1970, 167, 825–831. (c)
Ponec, R. Top. Curr. Chem. 1995, 174, 1–26.
(28) (a) Stolle, A.; Ondruschka, B.; Bonrath, W. Eur. J. Org. Chem. 2007,
2310–2317. (b) Stolle, A.; Ondruschka, B.; Findeisen, M. Proceedings of the
3rd International Conference on EnVironmental Science and Technology,
Houston, TX, August 5-9, 2007; Starrett, S. K.; Hong, J.; Wilcock, R. J.; Li,
Q.; Carson, J. H.; Arnold, S., Eds.; American Science Press: Houston, 2007; pp
325-329.
(25) Gajewski, J. J. Hydrocarbon Thermal Isomerization, 2nd ed.; Elsevier
Academic Press: London, 2004; pp 1-3.
(26) (a) Frey, H. M.; Walsh, R. Chem. ReV. 1969, 69, 103–124. (b) Gajewski,
J. J.; Paul, G. C. J. Am. Chem. Soc. 1991, 56, 1986–1989. (c) Lewis, D. K.;
Charney, D. J.; Kalra, B. L.; Plate, A.-M.; Woodard, M. H.; Cianciosi, S. J.;
Baldwin, J. E. J. Phys. Chem. A 1997, 101, 4097–4102. (d) Leber, P. A.; Baldwin,
J. E. Acc. Chem. Res. 2002, 35, 279–287. (e) Baldwin, J. E. Chem. ReV. 2003,
103, 1197–1212. (f) Baldwin, J. E.; Leber, P. A. Org. Biomol. Chem. 2008, 6,
36–47. (g) Northrop, B. H.; Houk, K. N. J. Org. Chem. 2006, 71, 3–13.
(29) (a) Frey, H. M.; Hopkins, R. G.; O’Neal, H. E. J. Chem. Soc., D 1969,
1069. (b) Roth, W. R.; Friedrich, A. Tetrahedron Lett. 1969, 10, 2607–2610.
(c) Roth, W. R.; Ko¨nig, J.; Stein, K. Chem. Ber. 1970, 103, 426–439. (d) Schuster,
D. I.; Widman, D. Tetrahedron Lett. 1971, 12, 3571–3574. (e) Dietrich, K.;
Musso, H. Chem. Ber. 1974, 107, 731–734. (f) Gajewski, J. J. Hydrocarbon
Thermal Isomerization, 2nd ed.; Elsevier Academic Press: London, 2004; pp
46-50.
8232 J. Org. Chem. Vol. 73, No. 21, 2008

Thermally Induced Rearrangement of R-Pinene
SCHEME 4. Thermal Network of Pyrolysis Products from
Thermolysis of 2,2-Dimethyl-1-vinylcyclobutane (5) and Its
Relationship to r-Pinene (1)a
CHART 2. Possible Transition States (TS3a and TS3b) for
the Formation of 4E,6Z-Alloocimene (3a) from 3Z-Ocimene
(2a)11
missing methyl group. Four major reaction pathways were found
by Chickos and Frey indicated by the respective rate constants
in Scheme 4. Fragmentation of 5 (kf) yields isobutene and 1,3-
butadiene exclusively, whereby the addition of the missing
methylene bridge and methyl group leads to 2a. 4,4-Dimeth-
ylcyclohexene is the 1,3-sigmatropic shift reaction product (k1,3)
from 5 being equal to the racemized form of 1 after pyrolysis.
Both reaction pathways leading to acyclic products for the
pyrolysis of 5 differ in position of one of the two carbon-carbon
double bonds, thus expressing the two different enantiomers of
4 resulting from pyrolysis of 1. Comparison of experimental
results presented herein and those found for thermolysis of 523
showed a good agreement concerning the products formed and
therefore the comparability of the two compounds.
a Asterisks attached to 5 and its pyrolysis products assign connection
points for the cyclohexane ring in 1, whereas the circle assigns the missing
methyl group.23
determining; otherwise optically active 4 has to be formed.
Additionally the syn-configuration of the vinylic system re-
mained unchanged since the change of configuration to anti in
the cyclic system would lead to an intermediate with trans-
cyclohexene structure, which is energetically disfavored.26g
Activation parameters for the formation of dipentene from 1
indicate a “tight” transition states thus implying a [1,5]H-shift
reaction as most plausible reaction (Scheme 3).10 Because
racemic 4 is formed, the retro-ene reaction has to proceed after
the formation of the biradical rather than before. Retro-ene
reaction of the starting material would yield optically active 4.
Since in the biradical intermediates TS1a and TS1b the one
radical is delocalized as allyl-type radical, TS1 loses completely
its configurational integrity as a result of the existence of a
mirror plane (point symmetry group: CS). Due to the findings
presented in Figure 2 it is obvious that the formation of 4 is
faster than racemization of 1. This leads to the conclusion that
the reconnection of the carbon bond is unfavorable comparing
to the hydrogen shift, thus indicating that the isopropyl group
changes its confirmation during the formation of the radical
intermediates.
The formation of 2a from 1 by thermally induced rearrange-
ment follows the known mechanisms for ring opening of
cyclobutanes.21-23,29f,30 Rupture of the second carbon-carbon
bond in intermediate TS1 and subsequent radical recombination
yielded 3Z-ocimene (2a) exclusively. The formation of 2b is
suppressed because of the rigid cyclic structure in the case of
TS1. The small amounts of the E-isomer (2b) probably result
from transition state TS2, whereby rotation around the C(2)-C(3)
single bond seems to be the most plausible explanation, which
indicated that the mesomeric delocalization is not complete,
since a rotation around a C-C-bond with sp2-hybridized atoms
is impossible. However, a photochemical rearrangement of 2a
to 2b31 can be ruled out due to the absence of any quenchers
and the stability of a mixture of both diastereomers.
Results of pyrolysis experiments with a mixture of 3Z- (2a)
and 3E-ocimene (2b) depicted in Figure 3 revealed that
predominately the Z-isomer was isomerized, whereas the
concentration of 2b remains constant within the temperature
range investigated. Analysis of the pyrolysis products revealed
that the formation of 4E,6Z-alloocimene (3a) is favored,
whereby thermolysis experiments with a mixture of 3a and 3b
allows for the conclusion that the 4E,6E-isomer (3b) results from
reconfiguration of the 6Z-double bond in 3a. Therefore, the
conclusion can be drawn that both the formation of 2a and its
disappearance are highly diastereospecific reactions. Liquid-
phase pyrolysis studies with 2a reported similar activation
parameters found for the gas-phase isomerization presented in
Table 1 (k4).11 Activation entropy ∆Sq and frequency factor log
A indicate a “tight” transition state as reaction intermediate,
being typical for reactions passing through a six-membered ring
transition state.20 Probably, the rearrangement of 2a to 3a
proceeds as [1,5]H-shift reaction pictured in Scheme 3. Ac-
cording to the Woodward-Hoffmann rules, this reaction is
thermally allowed if it proceeds in the suprafacial version.11,22,27,32
Whereas the stereospecific formation of the 6Z-double bond in
3a is due to the rigid cyclic transition state, the stereochemistry
of the double bond at C(4)-C(5) is controlled by the nature of
the reaction intermediate. Chart 2 pictures two possible inter-
mediates (TS3a and TS3b) for the suprafacial [1,5]H-migration
leading to either 3a or to 4Z,6Z-alloocimene, respectively. It is
evident that the thermal isomerization of 2a exclusively yields
3a and the reaction has to pass through intermediate TS3a (Chart
2). The preference of TS3a to TS3b could be explained by less
steric repulsion of the large isobutenyl substituents in the less
crowded equatorial position than in the more crowded axial.11
The formation of 3b from 3a can be explained by acid- or
base-catalyzed rearrangement of the respective double bond
affected by interactions with the reactor walls. This is evident
with respect to the kinetic analysis of the temperature-dependent
yields of both stereoisomers of 3 reported in Figure 4. On the
basis of the activation parameters listed in Table 1 for the rate
Results reported herein on the pyrolysis of 1 are in good
agreement with thermolysis studies on 2,2-dimethyl-1-vinylcy-
clopropane (5) as nonbridged cyclobutane model for 1.23 The
relationship between 1 and 5 is described in Scheme 4. Asterisks
attached to 10 and its pyrolysis products assign connection points
for the cyclohexane ring in 1, whereas the circle assigns the
(30) Srinivasan, R.; Kellner, S. M. E. J. Am. Chem. Soc. 1959, 81, 5891–
5893.
(32) Mironov, V. A.; Fedorovich, A. D.; Akhrem, A. A. Russ. Chem. ReV.
1981, 50, 666–681.
(31) Frank, G. J. Chem. Soc., B 1968, 130–132.
J. Org. Chem. Vol. 73, No. 21, 2008 8233

Stolle et al.
FIGURE 5. Comparison of data for the temperature-dependent contents of 4E,6Z-alloocimene (Y3a; Figure 5a) and of 4E,6E-alloocimene (Y3b;
Figure 5b) from experiment ([; cf. Figure 4a) and for data based on kinetic parameters listed in Table 1.
constants k5, k6, and k7 (Scheme 2) and with respect to eq 1 it
is possible to calculate the mole fractions of 3a and 3b.
Assuming that a mixture of both isomers is used (3a/3b ) 9.0),
eqs 3 and 4 allow for the calculation of their concentrations at
a desired temperature and residence time, respectively. The
additional formation of 3b from 3a is neglected in this first step
of approximation.
stable, and only minor amounts found in the pyrolysis product
of 1 seem to result from undesirable side reactions. Reactions
of 2a directly leading to 2b or 4E,6E-alloocimene (3b) were
not found. Results are summarized in the thermal reaction
network presented in Scheme 3. Products primarily formed from
1 are in agreement with those resulting from pyrolysis of 2,2-
dimethyl-vinylcyclobutane.
Kinetic analysis of the reactions allowed for determination
of Arrhenius as well as Eyring activation parameters for the
initial reaction steps. The use of nitrogen as carrier gas allows
for performing the reaction apart from the falloff region. Because
of the high extent of this diluting agent, its concentration relative
to substrate molecules is constant, allowing for the assumption
that the rate-determining reaction steps are unimolecular. The
model of competitive first-order parallel and consecutive reac-
tions was used to describe the reaction network. Activation
entropies and frequency factors for the formation of 2a and the
racemization of 1 clearly indicate a biradical as transition state.
This is in accordance with the rules of Woodward and
Hoffmann, because [2 + 2]-cycloreversions are thermally
forbidden reactions. Therefore, the fragmentation of the cy-
clobutane ring has to proceed stepwise. The formation of
racemic 4 and the fast isomerization of 2a to 3a seem to proceed
as [1,5]H shift reactions. Activation energy for the disappearance
of 2a supports the hypothesis that this is a very reactive
molecule. Pyrolysis of a mixture of 3a and 3b indicate an
isomerization leading from 3a to 3b but not the other way.
6;Tτ
[3a]T ) [3a]0 · e-k τ ) 0.9 · e-k
(3)
(4)
6;T
7;Tτ
[3b]T ) [3b]0 · e-k τ ) 0.1 · e-k
7;T
Figure 5a reveals that this simple kinetic is able to describe
well the disappearance of the 4E,6Z-isomer, whereas it is not
suitable for the modeling in the case of 3b (Figure 5b). It is
evident that consideration of the side reaction describing the
isomerization of 3a to 3b (eq 5) allows for an excellent modeling
of the temperature-dependent content of 3b (Y3b; cf. red line in
Figure 5b). Therefore, this Z/E-isomerization of the two isomers
of alloocimene is approved to proceed as a one-directional
reaction. Other isomers of 3 with a Z-configuration of the
carbon-carbon double bond at C(4)-C(5) were not able to be
identified in the reaction mixture, thus being in accordance with
the proposed mechanism pictured in Scheme 3. Consecutive
reactions of both isomers of 3 yield various monocyclic
products that mainly result from cyclization reactions leading
to substituted cyclohexadienes, thus being in agreement with
the reported activation data listed in Table 1 for the
disappearance of 3 (k6, k7).18
[3a]0 · k5;T
Experimental Section
6;Tτ
τ
[3b]T ) [3b]0 +
(1 - e-k ) · e-k7;T
[
]
k6;T
Pyrolysis Experiments. Dilution gas pyrolyses were carried
out in an electrically heated quartz tube 500 mm in length and
with a pyrolysis zone approximately 200 mm in length, using
the apparatus reported previously.16 Temperatures (T) were
regulated with thermocouples, and in addition the actual T was
measured inside the reactor. In all experiments, oxygen-free
nitrogen with a purity of >99.999% was used as the carrier gas.
The substrates (15 µL) were introduced onto a quartz ladle at
the top part of the pyrolysis apparatus using a glass syringe (50
µL) and carried along with the nitrogen stream into the reactor.
Vaporization of the starting material was supported by heating
the ladle to 200 °C with a hot blast. Pyrolysis products were
collected in a cold trap (liquid nitrogen) and were dissolved in
1.5 mL of ethyl acetate.
0.9 · k5;T
k6;T
6;Tτ
τ
) 0.1 +
(1 - e-k ) · e-k7;T
(5)
[
]
Conclusion
The bicyclic monoterpene R-pinene (1) was subjected to gas-
phase pyrolysis studies. Carrier gas pyrolysis of 1 reveals that
in addition to racemization of the starting material the formation
of alloocimene (3) and racemic limonene (4) was observed. The
expected product from ring opening of the cyclobutane ring in
1, ocimene (2) was not present in the product mixtures, thus
indicating that it is a highly reactive compound. In order to
unravel primary from secondary pyrolysis products, thermal
isomerization experiments with 2 and 3 were conducted.
Obviously the isomerization of 3Z-ocimene (2a) yields 4E,6Z-
alloocimene (3a) exclusively. The 3E-isomer (2b) is thermally
The liquid products obtained were analyzed by GC-FID and
GC-MS, with 5 µL of hexadecane added as internal standard.
Comparison of both residence times and mass spectra with those
of reference compounds allowed for identification of the main
8234 J. Org. Chem. Vol. 73, No. 21, 2008

Thermally Induced Rearrangement of R-Pinene
products. Yields and conversions have been calculated on the
basis of the corrected and normalized peak areas from the GC
analyses.
of various rate constants by performing kinetic pyrolysis experi-
ments at various temperatures allows for the calculation of both
Arrhenius (Ea, log A) as well as Eyring activation parameters (∆Hq,
∆Sq; eqs 1 and 2).
Kinetic Experiments. The kinetic experiments were carried out
in different temperature ranges depending on the starting materials
used for the experiments. Lower and upper boundaries for the
experiments were chosen with respect to those regions in the re-
spective temperature-conversion plots where the slope of the curve
is linear. Variation of the carrier gas velocity in a range from 0.4
to 1.2 L min-1 N2 allowed for the modulation of the average
residence time (τ). Pyrolysis experiments used for calculation of
rate constants were individually carried out three times. For cal-
culation of the parameters, all experimental results were considered.
If outliers were present within a data set the corresponding
experiments were redone individually four times and the data were
recalculated considering the additional experiments. Determination
Supporting Information Available: Analytical details, cal-
culation procedure for residence time, pyrograms of the three
substrates (1, 2, 3), 1D-NOESY data for structural elucidation
of isomeric ocimenes and alloocimenes, rate constants for all
reactions shown in Scheme 2, and Arrhenius plots for the
calculation of the activation parameters. This material is
available free of charge via the Internet at http://pubs.acs.org.
JO8012995
J. Org. Chem. Vol. 73, No. 21, 2008 8235