Thermal Decomposition of Chroman. Reactivity of o-Quinone Methide

Article abstract of DOI:10.1021/jo9701694

The gas phase thermal decomposition of 3,4-dihydro-2H-1-benzopyran (chroman, 1) has been studied between 760 and 1110 K in different bath gases and hydrogen donors. In nitrogen, the unimolecular rate parameters are k₁ (s^-1) = 10^15.3 exp(-263 (kJ mol^-1)/RT). The activation energy is slightly higher than the bond dissociation energy (BDE) of the phenoxylic C - O bond. The decomposition starts with elimination of ethene and formation of 6-methylene-2,4-cyclohexadien-1-one (o-quinone methide, 2). Quinone methides are important intermediates in the chemistry of lignin. In the high temperature range (860-980 K) 2 decomposes cleanly into CO, benzene, and small amounts of fulvene, obeying k₂ (s^-1) = 10^14.8 exp(-281 (kJ mol^-1)/RT. Reverse radical disproportionation of 2 with toluene is mainly responsible for o-cresol formation. In cis-2-butene at 770 K, exclusively cis-2,3-dimethylchroman is formed. This stereospecificity suggests a concerted retro-Diels - Alder mechanism and is not compatible with the high Arrhenius parameters, indicative of a stepwise, biradical mechanism.

Full text of DOI:10.1021/jo9701694

4
804
J . Org. Chem. 1997, 62, 4804-4810
Th er m a l Decom p osition of Ch r om a n . Rea ctivity of o-Qu in on e
Meth id e
Edwin Dorrestijn, Rapha e¨ l Pugin, M. Victoria Ciriano Nogales, and Peter Mulder*,†
Center for Chemistry and the Environment, Leiden Institute of Chemistry, Gorlaeus Laboratories,
Leiden University, P.O. Box 9502, 2300 RA Leiden, The Netherlands
Received J anuary 30, 1997X
The gas phase thermal decomposition of 3,4-dihydro-2H-1-benzopyran (chroman, 1) has been studied
between 760 and 1110 K in different bath gases and hydrogen donors. In nitrogen, the unimolecular
-
1
15.3
-1
rate parameters are k
1
(s ) ) 10
exp(-263 (kJ mol )/RT). The activation energy is slightly
higher than the bond dissociation energy (BDE) of the phenoxylic C-O bond. The decomposition
starts with elimination of ethene and formation of 6-methylene-2,4-cyclohexadien-1-one (o-quinone
methide, 2). Quinone methides are important intermediates in the chemistry of lignin. In the
high temperature range (860-980 K) 2 decomposes cleanly into CO, benzene, and small amounts
-
1
14.8
-1
of fulvene, obeying k
2
2
(s ) ) 10
exp(-281 (kJ mol )/RT). Reverse radical disproportionation of
with toluene is mainly responsible for o-cresol formation. In cis-2-butene at 770 K, exclusively
cis-2,3-dimethylchroman is formed. This stereospecificity suggests a concerted retro-Diels-Alder
mechanism and is not compatible with the high Arrhenius parameters, indicative of a stepwise,
biradical mechanism.
In tr od u ction
A large variety of aryl ethers is present in natural
atmosphere has been demonstrated, and addition of a
hydrogen atom was suggested as the first step in the
7a
formation of o-cresol. At high temperatures, 2 elimi-
composite materials such as lignin. For a long time, the
thermal properties of lignin have been a challenge for
chemists in order to derive useful feed stock molecules
from this renewable source. In the recent years we have
studied the thermal behavior of acyclic and cyclic aryl
8
nates CO, leading to formation of benzene and fulvene.
9
Turner, in his review, has discussed the susceptibility
in the liquid phase of the terminal methylene group
toward nucleophilic attack.
Different precursors are available for the thermal or
photochemical generation of 2, depending on the tem-
perature region and phase of interest, such as o-(meth-
oxymethyl)phenol,10 o-hydroxybenzyl alcohol, o-[(meth-
ylthio)methyl]phenol,12 and benzofuran-2(3H)-one.
1
2
ethers like methoxybenzene, diphenyl ether, phenyl
2
3
vinyl ether, 2,3-dihydro-1,4-benzodioxin, and 1,4-ben-
zodioxin.4 Under well controlled conditions the structural
influence on the energetics of the phenoxylic C-O bond
could be established. In case of 2,3-dihydro-1,4-benzo-
dioxin a biradical mechanism has been suggested as the
first step in the thermal degradation, to yield o-benzo-
quinone as an intermediate. Recently, the thermal
behavior of another lignin model compound, o-methox-
yphenol, and derivatives has been studied under radical
chain conditions.5 The occurrence of o-cresol in the
pyrolysate suggested that 6-methylene-2,4-cyclohexadien-
11
7e,8,13
3
By analogy with 2,3-dihydro-1,4-benzodioxin, 3,4-dihy-
dro-2H-1-benzopyran7 (chroman, 1) is also a precursor
for 2. Since accurately determined kinetic parameters
for the decomposition of 1, indicative for the followed
route, are not available, we investigated the thermal
behavior of 1 under various reaction conditions. For
formation of 2 at lower temperatures o-hydroxybenzyl
alcohol (3) was used. In this report a complete profile
for the thermolysis of 1 over a wide temperature range
a
1
-one (o-quinone methide, 2) may be a key intermediate.
Despite the fact that quinone methides are known to be
(
760-1110), as well as for the reactivity of 2, is presented
6
important intermediates in the conversion of lignin, only
in order to obtain mechanistic and kinetic information.
a few studies (mostly at one temperature) are dealing
with the chemistry of 2. In the presence of alkenes,
products stemming from cycloaddition (Diels-Alder type)
are reported.7
Resu lts
Th er m a l Decom p osition of 1 a n d 3: P r od u cts.
The reactivity of 2 in the gas phase in a hydrogen
Thermolysis between 774 and 903 K of 1 (initial
-6
concentration: ca. 3 × 10 M) dissolved in excess toluene
(to serve as a radical scavenging agent) in a nitrogen gas
stream yielded ethene and o-cresol as the main products
†
E-mail: P.Mulder@Chem.LeidenUniv.NL.
X
Abstract published in Advance ACS Abstracts, J une 15, 1997.
1) Arends, I. W. C. E.; Louw, R.; Mulder, P. J . Phys. Chem. 1993,
7, 7914-7925.
2) Scheppingen, W.; Dorrestijn, E.; Arends, I.; Mulder, P.; Korth,
H. G. J . Phys. Chem., in press.
3) Schraa, G. J .; Arends, I. W. C. E.; Mulder, P. J . Chem. Soc.,
(
9
(
(8) (a) Wentrup, C.; M u¨ ller, P. Tetrahedron Lett. 1973, 2915-2918.
(b) Wentrup, C. Tetrahedron Lett. 1973, 2919-2922.
(9) Turner, A. B. Q. Rev. Chem. Soc. 1964, 18, 347-360.
(10) (a) Gardner, P. D.; Sarrafizadeh R. H.; Brandon, R. L. J . Am.
Chem. Soc. 1959, 81, 5515. (b) Cavitt, S. B.; Sarrafizadeh R., H.;
Gardner, P. D. J . Org. Chem. 1962, 27, 1211-1216.
(11) (a) Eck, V.; Schweig, A.; Vermeer, H. Tetrahedron Lett. 1978,
27, 2433-2436. (b) McIntosh, C. L.; Chapman, O. L. J . Chem. Soc.,
Chem. Commun. 1996, 771. (c) Mao, Y. L.; Boekelheide, V. Proc. Nat.
Acad. Sci. U.S.A. 1980, 77, 1732-1735.
(12) Inoue, T.; Inoue, S.; Sato, K. Chem. Lett. 1989, 653-656.
(13) Chapman, O. L.; McIntosh, C. L.; Clardy, J . C. J . Chem. Soc.,
Chem. Commun. 1971, 384-385.
(
Perkin Trans. 2 1994, 189-197.
(
(
(
4) Boels, E.; Epema, O.; Mulder, P. To be published.
5) Dorrestijn, E.; Santos Gonzalez, C.; Mulder, P. To be published.
6) Shevchenko, S. M.; Apushkinskii, A. G. Russ. Chem. Rev. 1992,
6
1, 105-131.
(
7) (a) Paul, G. C.; Gajewski, J . J . J . Org. Chem. 1993, 58, 5060-
5
5
062. (b) Diao, L.; Yang, C.; Wan, P. J . Am. Chem. Soc. 1995, 117,
369-5370. (c) Katada, T.; Eguchi, S.; Esaki, T.; Sasaki, T. J . Chem.
Soc., Perkin Trans. 1 1984, 2649-2653. (d) Yato, M.; Ohwada, T.;
Shudo, K. J . Am. Chem. Soc. 1990, 112, 5341-5342. (e) Chapman, O.
L.; McIntosh, C. L. J . Chem. Soc., Chem. Commun. 1971, 383-384.
S0022-3263(97)00169-2 CCC: $14.00 © 1997 American Chemical Society

Thermal Decomposition of Chroman
J . Org. Chem., Vol. 62, No. 14, 1997 4805
stemming from 1; o-quinone methide (2) could not be
detected. Other compounds such as benzene and styrene
originated from the toluene carrier, as was demonstrated
by switching from toluene to xylene. The yield of ethene
corresponded well with the amount of converted chroman.
The yield of o-cresol increased from 0 to 18% with respect
to ethene. However, the carbon mass balance remained
far below 100%. To explore the fate of 2, thermolysis
experiments were performed using hydrogen instead of
nitrogen as the carrier gas. In this case the yield of
o-cresol increased to 83% (903 K), while the degree of
conversion of 1 (ca. 90%) did not change. With propene
as the bath gas, at 883 K, both isomers 2- and 3-meth-
ylchroman emerged with a total yield of about 40% on
converted 1 (ethene could not be quantified since it was
already present as an impurity in the propene). One
isomer, 2-methylchroman,14 was formed in more than
sixfold excess to the other. Other products were o-cresol
F igu r e 1. Product composition for the thermolysis of chroman
(2): ethene (9), CO (b), and benzene + fulvene + 1-buten-3-
yne ()).
(4%) and benzofuran (12%). To verify if 2 was still
present in the exit stream from an experiment with 1,
toluene, and nitrogen, we added propene directly after
the reactor. No methylchromans were detected, and the
product composition remained unchanged.
These observations suggest that the first step in the
thermolysis of 1 is the elimination of ethene and the
formation of o-quinone methide (2) (eq 1). Subsequently,
2
, which appears to be thermally stable in this temper-
ature region, reacts (partially) with hydrogen, propene
or toluene.
F igu r e 2. Relative (to ethene) amounts of o-cresol (9),
benzofuran ()), 2-methyl-2,3-dihydrobenzofuran (2), and dimer
(b) during the thermolysis of chroman.
To study the reactivity of 2 at a lower temperature
range (674-883 K), 2-hydroxybenzyl alcohol (3), dissolved
in toluene, was used as precursor. The conversion of 3
into 2, under elimination of water,1 was quantitative
under our reaction conditions. In propene a large amount
of 2 was converted into both methylchromans, the
amount decreasing with temperature (80% at 704 K, 1%
at 883 K, based on intake of 3). Over this temperature
range, the ratio 2-methylchroman:3-methylchroman
changed from 19:1 to 6:1, indicating that the isomer ratio
becomes thermodynamically controlled. Experiments
with 3 in a mixture of nitrogen/cis-2-butene at a ratio of
Traces of o-ethylphenol were found as a possible product
from addition of methane to 2.
1a
Th er m olysis of 1 a n d 2: Kin etics. Thermolysis of
-
6
1, highly diluted (10 M) in nitrogen, over a wide range
of randomly varied temperatures was performed. The
product profile between 750 K and 875 K almost exclu-
sively consisted of ethene (see Figure 1).
Minor products in the thermolysis of 1 were o-cresol,
benzofuran, and 2-methyl-2,3-dihydrobenzofuran. o-
Cresol could be quantified above 830 K (26% conversion
of 1) and reached a maximum of 1.7% on the intake of 1,
at 873 K (76% conversion, the highest temperature used
for determination of the Arrhenius parameters). Over
the same temperature range benzofuran was detected up
to 1.4% and 2-methyl-2,3-dihydrobenzofuran to around
4
/6 showed a complete conversion above 766 K into only
1
one major product (m/ z ) 162). On the basis of H-NMR
analysis (see Experimental Section) it could be concluded
that this was exclusively cis-2,3-dimethylchroman (eq 2).
1
%. Traces (<0.5%) of 2,3-dihydrobenzofuran, 2-meth-
ylbenzofuran, and o-allylphenol (identified by retention
time and/or GC-MS) were found. Starting at 855 K, a
high boiling product (A) was observed (1.5% at 873 K).
The identity could not be established by GC-MS, but
based on retention time a dimerization product of 2 is
the most likely candidate. The ratio of o-cresol/ethene,
benzofuran/ethene, and A/ethene increased almost lin-
early with temperature, but the ratio 2-methyl-2,3-
dihydrobenzofuran/ethene remained almost constant (see
Figure 2).
First-order behavior was verified by stepwise changing
the concentration of 1 at 853 K. After an increase in
concentration of in total a factor of 1.85 the same
conversion (46%) was maintained. At the same time, the
ratios of o-cresol, benzofuran, A, and 2-methyl-2,3-
Only small amounts of o-cresol were formed in hydro-
gen (4% at 674 K, 7% at 704 K, no other products could
be detected) or in methane (0% at 673 K, 6% at 730 K).
(
14) GC-MS showed two compounds (methylchromans) with identi-
cal spectra. According to a Diels-Alder pathway the regiospecificity
frontier orbital theory) would predict that formation of 2-methylchro-
(
1
5
man is favored. When a biradical route is followed the most stable
intermediate arises when the bond is formed between the CH in 2
and the primary CH in propene. This also leads to 2-methylchroman.
15) Fleming, I. Frontier Orbitals and Organic Chemical Reactions;
J ohn Wiley: London, 1976.
2
2
(

4
806 J . Org. Chem., Vol. 62, No. 14, 1997
Dorrestijn et al.
containing oligomers of 2 is responsible for the CO
production.
The correlation between both rate constants (eq 4) for
decomposition of 1 and 2 (k
derived according to a consecutive kinetic scheme (eq 5),
1 2
and k , respectively) can be
1
8
in which a possible dimerization of 2 is ignored in view
of the high reaction temperatures and low concentrations
of 2.
[
products] ) [chroman]
×
t
t)0
F igu r e 3. Arrhenius plots for unimolecular decomposition of
k1k2 exp(-k τ) exp(-k τ)
1 2
-
1
15.3
-1
chroman ([) k
1
(s ) ) 10
exp(-263 (kJ mol )/RT) and
1 -
-
(5)
{
(
)}
(s-1) ) 10
14.8
-1
k
2
- k
1
k
1
k
2
o-quinone methide (b) k
RT).
2
exp(-281 (kJ mol )/
In this expression [chroman]t)0 is the initial concentra-
dihydrobenzofuran over ethene increased by factors of
.6, 1.6, 7, and 1.3, respectively. With the measured
product composition the unimolecular rate constants, k
were derived, assuming plug flow behavior, according to:
tion of 1 (taken equal to the sum of ethene, 2-methyl-
2,3-dihydrobenzofuran, and remaining chroman) and
[products] the final concentration of products, i.e. the
t
sum of the measured concentrations of benzene, fulvene,
and 1-buten-3-yne. Based on these calculated values for
1
1
,
[
chroman]t
2
k , an Arrhenius plot (Figure 3) yielded the rate expres-
k1τ ) -ln
(3)
-1
14.8 ( 0.3
-1
(
)
sion: k
2
(s ) ) 10
exp(-281 ( 6 (kJ mol )/RT).
[
chroman]t)0
This rate of decomposition of 2 is lower than the
comparable CO elimination from o-benzoquinone. In
3
where τ is the residence time, [chroman]
t
the final, and
that case the decomposition was complete at these
temperatures. Both the thermolysis of 1 and 2 have been
repeated and found to be reproducible.
[chroman]t)0 the initial concentration of 1. The latter was
taken equal to the sum of remaining chroman, ethene,
and 2-methyl-2,3-dihydrobenzofuran, in agreement with
control measurements at 523 K. The Arrhenius expres-
sion was obtained from linear regression (Figure 3) of
Discu ssion
-
1
-1
15.3 ( 0.2
log(k
1 1
) vs T and found to obey k (s ) ) 10
The main process in the thermolysis of chroman (1) is
cleavage into ethene and o-quinone methide (2). The
kinetics will be discussed in next section, followed by an
elaboration on the reactivity of 2 under different condi-
tions, to include its thermal decomposition.
-
1
exp(-263 ( 2 (kJ mol )/RT).
At higher temperatures (around 870 K) the thermal
degradation of 2 started and carbon monoxide appeared
as the prominent product, together with benzene (Figure
1
) and small amounts of fulvene (based on retention time
Th er m olysis of 1: Kin etics. Two possible routes can
be envisaged for decomposition of chroman: a biradical
pathway similar to that reported for 2,3-dihydro-1,4-
1
6
and GC-MS). The ratio fulvene/benzene remained at
about 10/90 between 900 and 950 K, decreasing to 2/98
at 1006 K and, finally, fulvene disappeared (T > 1040
K). At temperatures above 930 K smaller quantaties
3
19
benzodioxin and 1,2,3,4-tetrahydronaphthalene, and a
concerted, retro-Diels-Alder mechanism.20 In general,
the observed kinetic parameters are indicative for the
preferred pathway. A biradical pathway has been con-
cluded for 2,3-dihydro-1,4-benzodioxin and 1,2,3,4-tet-
rahydronaphthalene because the measured activation
(e2%) of ethyne and a C4-species (based on retention
time), most likely 1-buten-3-yne, were found. The ratio
ethyne/1-buten-3-yne was close to unity and remained
constant with the temperature. Both compounds are
likely to emerge from the decomposition of fulvene.
Conversion of ethene into ethyne is highly unlikely under
these conditions.17 At about 1000 K the product mixture
consisted of almost equimolar amounts of CO, benzene,
and ethene. At higher temperatures the yield of CO is
clearly more than the initial amount of oxygen in the feed
energy (E
a
) was higher than the enthalpy required to
cleave the C-O/C-C bond. The differences (22 and 30
-1
kJ mol ) between the computed bond dissociation energy
BDE) and the E (as derived from product measure-
(
a
ments) were explained in terms of an additional barrier
for rotation around the phenolic O-C or the benzylic
(
see Figure 1). Even after removal of 1 from the
3
,19
C-C bond, before elimination of ethene.
The experi-
evaporation vessel and flushing the system with nitrogen,
CO as the only compound could still be detected for hours
when the reactor temperature was set at 1150 K. The
initial yield amounted to about 15% of the original intake
of 1 in the thermolysis experiments. Clearly, at these
high temperatures gasification of accumulated oxygen
-
1
mental energy of activation (263 kJ mol ) for decomposi-
tion of 1 is slightly above the strength of the phenoxylic
-
1 21
O-C bond, 254 kJ mol
.
This suggests that only a
-
1
small additional rotational energy of 9 kJ mol needs
to be incorporated. Together with the high preexponen-
tial factor, almost equal to that for 2,3-dihydro-1,4-
benzodioxin, it can be concluded that the thermal de-
(16) Fulvene was generated in a separate experiment by recombina-
tion of cyclopentadienyl radicals with chloromethyl radicals and
subsequent elimination of HCl.
(18) Moore, J . W.; Pearson, R. G. Kinetics and Mechanism, 3rd ed.;
J ohn Wiley: New York, 1981; p 291.
(17) Mallard, W. G.; Westley, F.; Herron, J . T.; Hampson, R. F. NIST
Chemical Kinetics Database, version 5.0; NIST Standard Reference
Data, National Institute of Standards and Technology: Gaithersburg,
MD, 1993.
(19) Tsang, W.; Cui, J . P. J . Am. Chem. Soc. 1990, 112, 1665-1671.
(20) Houk, K. N.; Gonzalez, J .; Li, Y. Acc. Chem. Res. 1995, 28, 81-
90.

Thermal Decomposition of Chroman
J . Org. Chem., Vol. 62, No. 14, 1997 4807
Sch em e 1. Ra te P a r a m eter s for th e Th er m olysis
of 1,2,3,4-Tetr a h yd r on a p h th a len e,
substantial increase in freedom of movement of the non-
benzoic ring. From Scheme 1, an average pre-exponen-
tial factor for decomposition of 1 and analogous com-
2
,3-Dih yd r o-1,4-ben zod ioxin ,
,4-Dih yd r o-2H-1-ben zop yr a n , Cycloh exen e,
,3-Dih yd r o-1,4-d ioxin , a n d 3,4-Dih yd r o-2H-p yr a n
3
q
pounds leads to an entropy of activation, ∆S , of 34 J
2
-1
-1
mol
K , still in the wide range reported for other retro-
-
1
-1 35
Diels-Alder reactions (-21 to +46 J mol
K
). Note
q
that for regular bond homolysis ∆S is mostly around 35-
-
1
-1 17
4
5 J mol
K
.
Hence, also no conclusions can be
drawn based on entropy considerations. In any case, the
stereospecificity of the reverse reaction should always be
taken into account. It is remarkable that this has not
1
9
been done in case of 1,2,3,4-tetrahydronaphthalene,
where 5,6-bis(methylene)-1,3-cyclohexadiene (o-quino-
dimethane) is the initial product of decomposition. The
reactivity of o-quinodimethane toward a dienophilic
compound has been investigated extensively. In almost
all cases the conclusion can be drawn, based on product
stereospecificity, that the Diels-Alder addition is the
preferred pathway.36 Efforts to disentangle the course
of reaction for 1,2,3,4-tetrahydronaphthalene, using a
different precursor for the biradical intermediate, were
composition of 1 follows a biradical pathway. On the
other hand, under the same conditions the reverse
reaction, 2 with cis-2-butene, yielded only cis-2,2-di-
methylchroman (eq 2), which leaves no doubt that an
addition according to a concerted Diels-Alder mechanism
is operative. Thus, based on the principle of microscopic
reversibility, these results are contradictive.
37
inconclusive.
Evidence in favor of a biradical pathway for thermoly-
1
9
sis of 1,2,3,4-tetrahydronaphthalene and 2,3-dihydro-
3
1
,4-benzodioxin can be obtained from the product com-
position. In both cases compounds were present with
retention of the CH CH moiety: o-allyltoluene and
-methyl-1,3-benzodioxole. The latter arises through
Our results are consistent with the findings of Paul
and Gajewski.7 In pyrolysis experiments of 1 in excess
cis-2-butene or trans-2-butene they observed addition to
2
2
a
2
bond rupture, followed by an exothermic 1,2-hydrogen
shift, and subsequent ring closure. According to a similar
mechanism 2-methyl-2,3-dihydrobenzofuran is expected
in the thermolysis of 1 and indeed found, albeit at low
yields (maximum 1.3%). This substantiates that a bi-
radical pathway (eq 6) is present, at least to some degree.
2
with high stereoselectivity. However, it has to be
noticed that these results were obtained after thermolysis
for 4.5 h at 686 K. With their reported conversions of 1,
an average rate constant can be calculated of around 8
-
6
-1
×
10
s , which is in reasonable accordance with our
-
6
-1
kinetic study, yielding k
conditions.
1
) 3.2 × 10
s
under those
In the last decades the transition state and mechanism
of the (retro-)Diels-Alder reaction have been subject to
extended experimental and theoretical studies and dis-
The decomposition rate parameters of 1 and structur-
ally related compounds are compiled in Scheme 1.
The decomposition of compounds without the aromatic
ring, like cyclohexene and 3,4-dihydro-2H-pyran, has
been reported to follow a retro-Diels-Alder mechanism
(
22) Colussi, A. J .; Zabel, F.; Benson, S. W. Int. J . Chem. Kinet. 1977,
, 161-178.
23) Suryan, M. M.; Kafafi, S. A.; Stein, S. E. J . Am. Chem. Soc.
1989, 111, 4594-4600.
24) Stein, S. E.; Rukkers, J . M.; Brown, R. L. NIST Structures and
9
(
a
based on energy considerations (E lower than BDE of
(
interest, see Table 1) and the stereospecificity of the
reverse reactions. From Scheme 1 and Table 1 a clear
difference can be noticed in behavior between the ben-
zocyclic and cyclic compounds. This may be due to the
rigidity of the aromatic ring which prevents the formation
of an intermediate state necessary for the concerted
decomposition.
Properties Database, version 2.0; NIST Standard Reference Data,
National Institute of Standards and Technology: Gaithersburg, MD,
1994.
(
25) Wayner, D. D. M.; Lusztyk, E.; Ingold, K. U.; Mulder, P. J . Org.
Chem. 1996, 61, 6430-6433.
26) J onsson, M.; Lind, J .; Eriksen, T. E.; Mer e´ nyi, G. J . Chem. Soc.,
Perkin Trans. 2 1993, 1567-1568.
27) Hansch, C.; Leo, A.; Taft, R. W. Chem. Rev. 1991, 91, 165-
95.
(
(
1
The current study shows that the magnitude of the
activation energy as such cannot be used as conclusive
evidence as to the followed mechanism. For a concerted,
retro-Diels-Alder process, only a small entropy change
may be expected, while a biradical pathway involves a
(
28) Cohen, N. J . Phys. Chem. Ref. Data. 1996, 25, 1411-1481.
(29) Tsang, W. Int. J . Chem. Kinet. 1973, 5, 651-662.
(
30) This work.
(
31) Taylor, R. J . Chem. Soc., Perkin Trans. 2 1988, 183-189.
(32) Chirico, R. D.; Archer, D. G.; Hossenlopp, I. A.; Nguyen, A.;
Steele, W. V.; Gammon, B. E. J . Chem. Thermodyn. 1990, 22, 665-
82.
6
(
33) Dorofeeva, O. V. Thermochim. Acta 1992, 200, 121-150.
(
21) BDE(C-O) in 1 can be calculated using BDE(C-O) in ethyl
(34) ∆
f
H(2) can be estimated according to increment rules: The
by a CdO in cyclohexane to cyclohexanone or in
f
2,4-cyclohexadienone to give o-benzoquinone changes ∆ H by -103 and
-1 24
-1
22
-1
phenyl ether (265 kJ mol
the ortho CH
)
and substracting 11 kJ mol , due to
replacement of a CH
2
2
3
2
CH
2
substituent, resulting in BDE(C-O) ) 254 kJ
-1
-1
mol in 1. As an alternative ∆
f
H(o-propylphenol) can be estimated to
-105 kJ mol , respectively.
f
∆ H(2) can be estimated to 68 kJ mol
1
-1
-
171 kJ mol- , applying the increment (-75 kJ mol ) for replacement
by applying this increment to 5-methylene-1,3-cyclohexadiene. Second
2
4
-1
of hydrogen by propyl in benzene. The O-H bond strength in
approach: A change of 67 kJ mol can be derived for the replacement
of a CH
cyclohexadiene. Applying that replacement to 2,4-cyclohexadienone
2
5
substituted phenols can be estimated using
∆BDE(O-H)
)
2
by a CdCH in 1,3-cyclohexadiene to give 5-methylene-1,3-
2
+
+
+
-1
+
+ 26
24
3
0.63{Σ(σ
o
+ σ
m
+ σ
p
)} - 2.68 kJ mol ; and σ
o
) 0.66 σ
p
.
With
+
27
-1
-1
σ
p
(propyl) ) -0.29, BDE(O-H) becomes 356 kJ mol . BDE(C-H)
f
yields ∆ H(2) ) 67 kJ mol . AM1 calculations yielded a value of 64.5
-1 -1
-
1
24
•
is taken equal to that in ethane: 423 kJ mol
.
Thus ∆
f
H(C
6
H
4
(O )-
f
kJ mol . We use ∆ H(2) ) 66 kJ mol .
•
-1
-1
C
3
H
6
) becomes 172 kJ mol and with ∆
f
H(1) ) 82.4 kJ mol (see
(35) Sauer, J .; Sustmann, R. Angew. Chem. 1980, 92, 773-801.
(36) (a) Charlton, J . L.; Alauddin, M. M. Tetrahedron 1987, 43,
2873-2889. (b) Sato, H.; Isono, N.; Okamura, K.; Date, T.; Mori, M.
Tetrahedron Lett. 1994, 35, 2035-2038. (c) Fleming, I.; Gianni, F. L.;
Mah, T. Tetrahedron Lett. 1976, 881-884.
-1
Table 1), BDE(C-O) in 1 254 kJ mol . It should be noted that the
error associated with these type of BDE calculations is in the order of
1
7
kJ mol- . Since both approaches resulted in the same value it can
be concluded that an additional ring strain component is negligible,
in accordance with ref 28.
(37) Gajewski, J . J .; Paul, G. C. J . Org. Chem. 1990, 55, 4575-4581.

4
808 J . Org. Chem., Vol. 62, No. 14, 1997
Dorrestijn et al.
Ta ble 1. En th a lp ies, Bon d Dissocia tion En er gies a n d Activa tion En er gies for Con ver sion of Ch r om a n a n d Rela ted
-
1
Com p ou n d s u n d er Elim in a tion of Eth en e (in k J m ol
)
cussions.20 The reasoning described above is based on
the linkage between stereospecifity and a concerted
mechanism. However, even in case of a stepwise (bi-
radical) mechanism the stereochemical retention can be
the same as for a concerted process if the second bond
breaks within the time necessary for C2-C3 rotation.
Very recently, Horn et al.38 demonstrated the existence
of such a process by studying the retro-Diels-Alder
reactions of norbornene and norbornadiene on femto-
second timescale. In our case, the difference between the
Even under our dilute conditions and higher temper-
atures, the low mass balances point to formation of
oligomers. These may stick to the wall of the reactor
and are thermolyzed at elevated temperatures (T > 1000
K) to result in the formation of CO. In the presence of
other reagents this oligomerization competes with other
processes.
7
a
A product which arises in variable amounts is o-cresol,
of which the presence is not affected by the followed
experimental procedures. Thus far, only speculative
-
1
7a
BDE and the experimental E
a
(of 9 kJ mol ) has been
mechanisms have been proposed.
In the absence of
explained as an additional barrier for rotation around
the benzylic C-C axis, prior to ethene elimination. At
hydrogen or a hydrogen donating reagent (toluene) small
amounts of o-cresol were formed. As a function of
temperature, the formation of o-cresol and benzofuran
showed the same behavior (Figure 2), and therefore these
compounds are likely to be formed from the same
precursor. At higher initial concentrations of 1 and at a
constant degree of conversion, the increase in yield of
these compounds is more than proportional, suggesting
that both products originate from a reaction between two
quinone methides. However, it appears to be difficult to
propose the ruling mechanism on thermochemical kinetic
grounds. A suggestion is that first (reversible) dimer-
ization of 2 takes place and, thereafter, thermal rear-
rangement of the dimer partially leads to formation of
o-cresol and benzofuran.
8
0
50 K, the rotational time around this bond will be about
.4 ps,38 which is most probably higher than that for the
(
free) rotation around the C2-C3 axis in 1. Hence, a loss
of stereospecificity is expected. Therefore, only when the
equals the BDE the above mechanism would explain
E
a
our observations. However, the predicted BDE by en-
thalpy assessment will be too high rather than too low,
since minor corrections (e.g. ∆C , see Table 1 footnote o)
p
are not incorporated. Hence, further (computational)
research will be necessary before reaching a final conclu-
sion as to this issue.
Rea ctivity of 2. At temperatures as low as 273 K,
1
0a,11c,39
several authors reported the formation of the dimer,
trimer,1
0,11b,c,39
and even tetramer of 2. These oligomers
39
In the same temperature region and in a large excess
of toluene, the yield of o-cresol is much higher (18% at
are products of the Diels-Alder reaction between two
molecules of 2 (e.g. eq 7).
9
03 K). An explanation for this result is that reverse
radical disproportionation (RRD) of 2 with the hydrogen
(
38) Horn, B. A.; Herek, J . L.; Zewail, A. H. J . Am. Chem. Soc. 1996,
18, 8755-8756.
39) Letulle, M.; Guenot, P.; Ripoll, J . L. Tetrahedron Lett. 1991,
2, 2013-2016.
donor (toluene) takes place (eq 8).
1
3
It should be emphasized that 2 is extremely sensitive
to hydrogen transfer reactions (RRD) due to the strong
(

Thermal Decomposition of Chroman
J . Org. Chem., Vol. 62, No. 14, 1997 4809
(
de)anthracene (BenzH)) resulted in a selectivity toward
4
6
o-cresol of 88% (554 K, 1 h, 3:BenzH:AnH
) 1:4:40).
2
Degr a d a tion of 2: P r od u ct Com p osition . Conver-
sion of o-quinone methide yields predominantly CO and
benzene, together with some fulvene. Whether fulvene
and benzene are formed in a parallel or consecutive
process can be judged from reported Arrhenius param-
C-H bond that will be formed.40 By treating the conver-
sion of 1 into 2 and subsequently of 2 into o-cresol by
RRD with toluene the same way as eq 5, a kRRD,exp is
-
1 24
eters for the exothermic (-141 kJ mol
) isomerization
-
1
-1 42
-1
derived of 12 M
s
.
In reality the concentration of 2
of fulvene into benzene. With a rate constant k (s ) )
1
3,14
-1
47
is lower due to the reversible dimerization. This would
10
exp(-(268-285) (kJ mol )/RT) the rate is too
result in a higher kRRD in accordance with the predicted
low to explain the change in the ratio of these compounds
and hence fulvene and benzene arise from two parallel
pathways, in accordance with a study by Wentrup and
M u¨ ller.8
4
2
value based on energy considerations. A second route
toward o-cresol starts with the addition of a hydrogen
atom, present in the gas mixture, to 2. The hydrogen
a
-
14
atom concentration (5.6 × 10
M) can be calculated
Degr a d a tion of 2: Kin etics. Two possibilities can
be given for the decomposition of o-quinone methide (2):
Firstly, a pathway starting with cleavage of the C(dO)sC-
independently from the amount of methane formed by
demethylation of toluene. With a toluene concentration
4
4
-
3
10
-1 -1 17
of 3 × 10 M and assuming a kadd of 10
RRD, exp × [toluene] ) 3 × 10 is higher than kadd
and therefore RRD with toluene is
M
s
,
2
(dCH ) bond, to give a biradical, followed by elimination
-
2
-1
k
s
×
of CO to form benzene and fulvene (eq 9, route a).
Secondly, a concerted process can be envisaged (eq 9,
route b).
•
-4 -1
[H ] ) 6 × 10
s
faster. When the bath gas is changed from nitrogen to
hydrogen the methane and hence hydrogen atom con-
centration increases by a factor 14. Since the rate of the
RRD remains constant the increase in o-cresol by a factor
-6
-6
4
(from 0.5 × 10 M to 2 × 10 M) can be ascribed to
the higher contribution of the hydrogen atom pathway.
Both the RRD rate constant and the hydrogen atom
concentration are temperature dependent. Therefore,
both processes to yield o-cresol will diminish in impor-
tance with decreasing temperatures, in agreement with
our observations. RRD with H
sibility to form o-cresol but is too slow compared with
addition of hydrogen atoms.
2
would be another pos-
Direct CO elimination has also been observed in the
4
8
48a
thermolysis of 3-cyclopentenone, 3,5-cycloheptadienone,
and 2,2,4,4-tetramethylcyclobutanone.49 In all cases, CO
elimination proceeds via a concerted pathway. In case
of 3-cyclopentenone and 3,5-cycloheptadienone the prod-
In a large excess of propene, at 883 K, a competition
takes place between addition of propene to 2 and hydro-
gen transfer. Product analysis reveals a high selectivity
toward addition. The rates of RRD42 for toluene and
propene are quite similar since the benzylic and allylic
C-H BDEs are identical.41 Under these conditions the
rate constant for Diels-Alder addition of propene to 2,
1
4
ucts are acyclic, with preexponential factors of 3.3 × 10
-1
16 -1
s
and 1.6 × 10
s
, respectively. A better comparison
can be made with 2,2,4,4-tetramethylcyclobutanone where
the product of CO elimination is cyclic. The preexpo-
1
4
-1
nential factor is reported as 8 × 10
equals the value we found, namely 6.3 × 10
The BDE for bond homolysis can be estimated to 293
s , which almost
4
5
14 -1
k
DA is around 200 times higher than kRRD
.
s .
Our analyses show that, besides addition of alkenes,
-
1 50
kJ mol
.
Because the observed energy of activation,
81 kJ mol is lower, it may be concluded that the
RRD can be an effective route for scavenging and
quantifying 2. Indeed, preliminary experiments with 3
in a pressurized liquid system in the presence of hydro-
-
1
2
decomposition follows a concerted mechanism. This
pathway (eq 9) suggests that fulvene is the major product.
However, since mainly benzene is observed and the
isomerization of (ground state) fulvene is too slow, it may
be that collisional not stabilized fulvene isomerizes more
easily into benzene.
2
gen donors (9,10-dihydroanthracene (AnH ) and 7H-benz-
(
40) Using the BDE(C-H) in o-cresol (370 kJ mol-1, taken equal to
41 -1
that of toluene) and the BDE(O-H) (355 kJ mol , analogous to ref
+
27
•
1
9 with σ
p
(methyl) ) -0.31),
kJ mol and ∆ H(o-HO-C -CH
a hydrogen atom to 2 yields ∆ H ) -273 kJ mol (addition to CdCH
H ) -258 kJ mol (addition to CdO).
41) Berkowitz, J .; Ellison, G. B.; Gutman, D. J . Phys. Chem. 1994,
∆
f
H(o-CH -C
•
2
) 26.5 kJ mol . Thus addition of
3
6 4
H -O ) becomes 11.5
-1
-1
f
6
H
4
-1
r
2
)
-1
and ∆
(
f
Con clu sion s
9
8, 2744-2765.
On the basis of the activation energy, the thermolysis
of 1 starts with the cleavage of the phenolic C-O bond.
-
1
-7
(
42) Substituting k
1
) 0.36 s (903 K), [o-cresol]
t
) 0.5 × 10 M,
-7
-1 -1
and [chroman]t)0 ) 32.4 × 10 M in eq 5 yields kRRD,exp ) 12 M
s .
For RRD the general Arrhenius parameters are log A ) 8.6 (per H)
-1 43
and E
1
a
) E
0
+ 0.82 × ∆
r
H (with E
0
) 37.66 kJ mol ). The ∆
r
H is
(46) (a) Ciriano Nogales, M. V.; Dorrestijn, E.; Mulder, P. To be
published. For experimental details: (b) Arends, I. W. C. E.; Mulder,
P. Energy Fuels 1996, 10, 235-242.
-
1
01 kJ mol using ∆
f
H of the compounds from refs 24 and 40. The
-
1
-1
9
-1
rate equation becomes kRRD (M
RT), resulting in a predicted value kRRD ) 114 M
s
) ) 1.2 × 10 exp(-121 (kJ mol )/
-1
-1
s
.
(47) Gaynor, B. J .; Gilbert, R. G.; King, K. D.; Harman, P. J . Aust.
J . Chem. 1981, 34, 449-452.
(
43) Savage, P. E. Energy Fuels 1995, 9, 590-598.
•
•
-1 -1
10
(44) For C
6
H
5
CH
3
-
+ H f C
6
H
6
+ CH
3
, k (M
s
) ) 1.55 × 10
(48) (a) Buxton, J . P.; Simpson, C. J . S. M. Chem. Phys. 1986, 105,
307-316. (b) Dolbier, W. R.; Frey, H. M. J . Chem. Soc., Perkin Trans.
2 1974, 1674-1676.
1
17
exp(-24.19 (kJ mol )/RT). At 900 K in toluene/nitrogen, [CH
7
4
] )
-7
.3 × 10 M.
45) The rate parameters for Diels-Alder addition of an alkene to
can be derived for ethene but, in a first approximation, will also be
(
(49) Frey, H. M.; Hopf, H. J . Chem. Soc., Perkin Trans. 2 1973,
2016-2019.
2
applied to other alkenes with 2. For the activation energy a value can
(50) AM1 calculations provide the C-H BDE in CH dCHsC-
2
-1
-1
-1
be derived of E
a
(DA) ) E
a
(retro-DA) - ∆
r
H ) 263 - 201 ) 62 kJ mol
.
(dO)sH (346 kJ mol ) and in CH
2 2
dCHsC(dCH )sH (409 kJ mol ).
1 24
-
Assuming the same entropy change as for norbornene to cyclopenta-
diene and ethene (142 J mol
derived as 7.6 × 10
These values, with the ∆ H of 2,4,6-heptatrienal (40 kJ mol ), yield
f
-1
-1 17
•
•
f 2
H( C(dO)sCHdCHsCHdCHsC(dCH ) ) ) 359 kJ mol . Apply-
34 -1
-1
K
)
a preexponential factor can be
a ∆
f
ing a ∆ H of 66 kJ mol for 2 results in a BDE of 293 kJ mol .
7
-1 -1
-1
M
s
.

4
810 J . Org. Chem., Vol. 62, No. 14, 1997
Dorrestijn et al.
However, the stereospecificity of the reverse reaction
suggests that a concerted pathway is operative. The
barrier for rotation around the benzylic C-C bond, prior
to the elimination of ethene, is too high to allow a
retention of configuration. The product o-quinone me-
thide (2) decomposes at high temperatures smoothly into
benzene and CO, according to a concerted mechanism.
It has been shown that o-quinone methide is very reactive
toward alkenes (addition) and toluene (hydrogen trans-
fer). Reverse radical disproportionation with toluene has
been shown to be an important pathway in the formation
of o-cresol. Since quinone methides are important in-
termediates in e.g. lignin liquefaction, the application of
hydrogen donors is likely to be beneficial for the forma-
tion of usefull feedstock compounds.
a mass spectrometric detector (HP 5972 MSD) was used. The
residence times were between 7.7 and 9.0 s for the experiment
with 1 and between 7.8 and 10.2 s for the experiments with 3.
On -Lin e Exp er im en ts. The instrument and the analysis
3
have been described before but in contrast to former experi-
ments, monitoring of data (GC signal, temperatures, flow,
pressure) as well as controlling and regulating the equipment
took place by software (Labview 3.1, HP GC Chemstation rev.
A.03.02).51 Nitrogen was used as carrier gas in case of the
thermolysis of 1 and 2. The residence times were between
2
.9 and 4.0 s. When cis-2-butene was present during the
thermolysis of 3 residence times were 3.5 to 5.5 s. On-line
sampling took place on a HP 5890A series II gas chromato-
graph with He as carrier gas.
1
NMR Mea su r em en ts. A H-NMR spectrum of 200 scans
3
in CDCl was measured on a 300 MHz DPX300 from Bruker.
Ch em ica ls. o-Hydroxybenzyl alcohol (Aldrich, 99%), chlo-
robenzene (Aldrich, 99.9%), p-chlorotoluene (Fluka, purum),
toluene (Baker, p.A.), xylene (Baker, 98%), nitrogen (Air
Products, ultrapure (on-line experiments), 99.9% (off-line
experiments)), propene (Air Products, 99%), hydrogen (Air
Products, 99.9%), cis-2-butene (Air Products, 99%), and meth-
ane (Air Products, 99.995%) were used as received. Chroman
Exp er im en ta l Section
In this study two different experimental setups have been
utilized. To investigate the (change in) product distribution
of the thermal decomposition of 1 and 3, an offline apparatus
was used, while detailed kinetic parameters were determined
using the online equipment.
5
2
(1) was prepared according to the synthesis reported earlier.
The mass spectrum was in agreement with that from litera-
ture,53 the GC-purity amounted to 98.5%.
cis-2,3-Dim eth ylch r om a n : MS m/z: 162 (M , 63), 147 (29),
Off-Lin e Exp er im en ts. This technique has been described
3
+
before. In short: a solution of the reagent in toluene (ca.
1
000-fold excess) was lead by the carrier gas (molar ratio gas/
133 (100), 107 (74), 105 (43), 91 (44), 78 (48), 77 (42), 51 (29),
39 (34). NMR: δ 7.12-6.76 (m, 4H), δ 4.24 (dq, J ) 2.56, 6.53
Hz, 1H), δ 2.95 (broad dd, J ) 5.75, 16.31 Hz, 1H), δ 2.49
(broad dd, J ) 5.32, 16.35 Hz, 1H), δ 2.09 (m, 1H), δ 1.28 (d,
J ) 6.51 Hz, 3H), δ 0.96 (d, J ) 9.97 Hz, 3H).
reagent ca. 3500) through a heated quartz tubular flow reactor
with an effective volume of 43.5 mL. Gaseous as well as liquid
samples were obtained. To prevent any hold-up by condensa-
tion, all tubing was kept at 473 K. The gas samples were
analyzed on a Packard 428A gas chromatograph (packed
J O9701694
Carbosphere column 80/100 mesh) for CO, CO
2 4 2 2
, CH , C H ,
C
2
H
4
, and C . Detection took place on a FID after metha-
2 6
H
(
(
51) Dorrestijn, E.; Mulder, P. J . Anal. Appl. Pyrolysis, submitted.
52) Deady, I. W.; Topsom, R. D.; Vaughan, J . J . Chem. Soc. 1963,
nization. The liquid samples were analyzed on a HP 5890A
gas chromatograph (CP-Sil-5-CB capillary column, 50 m × 0.32
2
094-2095.
mm i.d., 0.4 µm film thickness) with H
2
as carrier gas and a
(53) Trudell, J . R.; Sample Woodgate, D.; Djerassi, C. Org. Mass
FID detector. For identification a HP 5890 GC equipped with
Spectrom. 1970, 3, 753-776.