Vapor-phase photochemistry of dimethylpyridines

Article abstract of DOI:10.1021/ja990773r

Irradiation of dimethylpyridine vapors (2-5 Torr) at 254 nm results in the formation of two sets of isomerization products. One set, formed in the larger yield, is substantially quenched when the irradiations are carried out in the presence of 15-21 Torr of nitrogen and is not formed when the irradiations are carried out with light of λ > 290 nm. In addition, a second set of reactions, which involve the interconversion of 2,3- and 2,5- dimethylpyridines, is enhanced by the addition of nitrogen, and these reactions are the only photoisomerization reactions observed when the irradiations are carried out with light of λ > 290 nm. In addition to the photoisomerizations, four of the dimethylpyridines also undergo demethylation to yield monomethylpyridines, and 2,6-dimethylpyridine undergoes methylation to yield a trimethylpyridine product. A variety of crossover experiments confirmed that the photoisomerizations are intramolecular. Based on the major phototransposition products, the six dimethylpyridines can be divided into two triads. Interconversion of the three members of each triad results in the major phototransposition products. These intra-triad interconversions are suggested to occur via 2,6-bonding, originating in a vibrationally excited S₂ (π,π*) state of the dimethylpyridine, followed by nitrogen migration and rearomatization. This allows nitrogen to insert within each carbon- carbon bond. Phototransposition of 2,6-dideuterio-3,5-dimethylpyridine to a mixture of 5,6-dideuterio-2,4-dimethylpyridine and 3,4-dideuterio-2,5- dimethylpyridine is consistent with this mechanism. In addition to these intra-triad reactions, 2,5-dimethylpyridine, a member of triad 1, was observed to interconvert with 2,3-dimethylpyridine, a member of triad 2. These inter-triad reactions are suggested to occur via interconverting Dewar pyridine intermediates, formed from the triplet state of the dimethylpyridines. These Dewar pyridine intermediates were also observed by ¹H NMR spectroscopy after irradiation of the dimethylpyridines in CD₃CN at -30 °C.

Full text of DOI:10.1021/ja990773r

5666
J. Am. Chem. Soc. 1999, 121, 5666-5673
Vapor-Phase Photochemistry of Dimethylpyridines
James W. Pavlik,*,† Naod Kebede,† Michael Thompson,‡ A. Colin Day,‡ and
John A. Barltrop‡
Contribution from the Department of Chemistry and Biochemistry, Worcester Polytechnic Institute,
Worcester, Massachusetts 01609, and The Dyson Perrins Laboratory, Oxford UniVersity,
Oxford OX1 3Q4, England
ReceiVed March 9, 1999. ReVised Manuscript ReceiVed April 30, 1999
Abstract: Irradiation of dimethylpyridine vapors (2-5 Torr) at 254 nm results in the formation of two sets of
isomerization products. One set, formed in the larger yield, is substantially quenched when the irradiations are
carried out in the presence of 15-21 Torr of nitrogen and is not formed when the irradiations are carried out
with light of λ > 290 nm. In addition, a second set of reactions, which involve the interconversion of 2,3- and
2,5-dimethylpyridines, is enhanced by the addition of nitrogen, and these reactions are the only photoisomer-
ization reactions observed when the irradiations are carried out with light of λ > 290 nm. In addition to the
photoisomerizations, four of the dimethylpyridines also undergo demethylation to yield monomethylpyridines,
and 2,6-dimethylpyridine undergoes methylation to yield a trimethylpyridine product. A variety of crossover
experiments confirmed that the photoisomerizations are intramolecular. Based on the major phototransposition
products, the six dimethylpyridines can be divided into two triads. Interconversion of the three members of
each triad results in the major phototransposition products. These intra-triad interconversions are suggested to
occur via 2,6-bonding, originating in a vibrationally excited S2 (π,π*) state of the dimethylpyridine, followed
by nitrogen migration and rearomatization. This allows nitrogen to insert within each carbon-carbon bond.
Phototransposition of 2,6-dideuterio-3,5-dimethylpyridine to a mixture of 5,6-dideuterio-2,4-dimethylpyridine
and 3,4-dideuterio-2,5-dimethylpyridine is consistent with this mechanism. In addition to these intra-triad
reactions, 2,5-dimethylpyridine, a member of triad 1, was observed to interconvert with 2,3-dimethylpyridine,
a member of triad 2. These inter-triad reactions are suggested to occur via interconverting Dewar pyridine
intermediates, formed from the triplet state of the dimethylpyridines. These Dewar pyridine intermediates
were also observed by 1H NMR spectroscopy after irradiation of the dimethylpyridines in CD3CN at -30 °C.
When dimethylpyridines are irradiated in the vapor phase,
Scheme 1
they undergo photoisomerization as well as demethylation and
methylation reactions. To account for the photoisomerizations
they reported, Caplain and Lablache-Combier suggested a ring
transposition mechanism involving azaprismane intermediates.1
Such species could arise via Dewar isomers, initially formed
by 2,5- (or 3,6-) bonding in the photoexcited dimethylpyridine.
Indeed, the formation of Dewar pyridine was already known to
result from irradiation of pyridine in liquid solution.2
A perplexing feature of the suggested transposition mecha-
nism is the selectivity required to explain the products reported
by Caplain and Lablache-Combier.1 Thus, as shown in Scheme
1, the reported conversion of 2,6-dimethylpyridine (4) to 2,4-
dimethylpyridine (2) requires that the reactant first undergoes
2,5- (or 3,6-) bonding (but not N1-C4 bonding) and regiospecific
opening of the subsequently formed azaprismane via cleavage
of the C3-C4 and C5-C6 bonds, but not of the C2-C3 and N1-
C6 bonds, which would have resulted in the formation of 2,5-
dimethylpyridine (3).
mane via both possible pathways. These examples illustrate the
arbitrary selectivity that must be imposed upon the possible
modes of formation of the initially formed Dewar pyridines as
well as on the rearomatization of the subsequently formed
azaprismanes.
As a result of these mechanistic ambiguities, we have re-
investigated the photochemistry of the six isomeric dimethyl-
pyridines in the vapor phase. In this paper we report the results
of this investigation which, to our initial surprise, are totally at
variance with the results previously published by Caplain and
Lablache-Combier.1 Although we cannot account for the dif-
ferences between our results and those previously published,
Alternatively, the reported conversion of 2,3-dimethylpyridine
(1) to a mixture of 3,4-dimethylpyridine (5) and 2,5-dimethyl-
pyridine (3) requires (Scheme 2) initial N1-C4 bonding (but
not 2,5- or 3,6-bonding), followed by cleavage of the azapris-
† Worcester Polytechnic Institute.
‡ Oxford University.
(1) Caplain, S.; Lablache-Combier, A. J. Chem. Soc., Chem. Commun.
1969, 1247-1248.
(2) Wilzbach, K. E.; Rausch, D. J. J. Am. Chem. Soc. 1970, 92, 2178-
2179.
10.1021/ja990773r CCC: $18.00 © 1999 American Chemical Society
Published on Web 06/05/1999

Vapor-Phase Photochemistry of Dimethylpyridines
J. Am. Chem. Soc., Vol. 121, No. 24, 1999 5667
Table 1. Photolysis Productsa
Scheme 2
our results are highly reproducible, and we have no reason to
doubt their validity.3
Results and Discussion
Each dimethylpyridine (1-6) vapor, with a total pressure of
2-5 Torr, was irradiated through quartz at 254 nm or through
Pyrex with light of wavelength >290 nm. In addition, a partial
pressure of 2-5 Torr of each dimethylpyridine vapor (1-6)
was diluted with 15-21 Torr of nitrogen, and each mixture was
also irradiated through quartz at 254 nm. After each irradiation,
the volatile materials were pumped out of the reaction flask,
condensed, and dissolved in a standard volume of ether for
quantitative GLC analysis. The recovered material accounted
for approximately 45% of the mass of the original reactant.
Products were identified by GLC retention times, by GLC co-
injections with authentic samples, and, when possible, by mass
spectroscopy. Photolyses were also accompanied by the forma-
tion of variable amounts of nonvolatile polymeric material which
built up on the sides of the reaction flask and is responsible for
1
the low mass balances shown in Tables 1 and 2. H NMR
analysis of this material showed the presence of aliphatic
hydrogen but no absorptions due to aromatic protons, suggesting
that polymerization is accompanied by ring opening. Photolyses
of dimethylpyridines 3 and 6 were accompanied by the
formation of substantially less polymeric material, and accord-
ingly these reactions exhibited higher mass balances.
a Numbers are percent of reactant consumed or the percent yields
of products formed after irradiation. b Irradiation at 254 nm. c Irradiation
at 254 nm in the presence of added N2. d Irradiation at λ > 290 nm.
Tables 1 and 2 show the products formed from each dimethyl-
pyridine under each set of conditions. As noted earlier, the pho-
totransposition products observed in this study are totally dif-
ferent than those previously reported by Caplain and Lablache-
Combier.1 As Tables 1 and 2 show, both phototransposition and
photodemethylation and methylation products were observed.
Although Caplain and Lablache-Combier reported that methyl-
pyridines undergo phototransposition as well as demethylation
and methylation upon irradiation in the vapor phase, they re-
ported that dimethylpyridines undergo only photoisomerization.
Although these irradiations result in a variety of phototrans-
position products, inspection of Tables 1 and 2 reveals that they
can be divided into two classes. For example, irradiation of pure
2,3-dimethylpyridine (1) vapor at 254 nm results in the
formation of 2,6-dimethylpyridine (4), 3,4-dimethylpyridine (5),
and 2,5-dimethylpyridine (3) in yields of 5.7, 0.9, and 0.8%,
respectively.5 Upon dilution of the reactant with N2, the yields
of 4 and 5 were substantially decreased to 1.8 and 0%,
respectively, while the yield of 2,5-dimethylpyridine (3) in-
creased from 0.8 to 1.5%. Furthermore, 3 is the only phototrans-
position product formed when 1 is irradiated with light of
wavelength longer than 290 nm. These results indicate the
operation of two mechanistically different phototransposition
pathways. One pathway, leading to the formation of 4 and 5, is
substantially quenched upon dilution of the reactant with N2
and does not operate upon irradiation with λ > 290 nm. In
addition, a second pathway, leading to the formation of 3, is
also implicated, which is enhanced by dilution with N2 and is
the only pathway in operation upon irradiation at λ > 290 nm.
Irradiation of pure 2,5-dimethylpyridine (3) vapor at 254 nm
also leads to the formation of three phototransposition products,
2,4-dimethylpyridine (2), 3,5-dimethylpyridine (6), and 2,3-
dimethylpyridine (1), in yields of 13.7, 3.7, and 3.5%, respec-
tively. Again, two mechanistic pathways are implicated since
the yields of 2 and 6 are either reduced or unaffected by dilution
with N2, and neither product is formed upon irradiation of 3 at
(3) Caplain and Lablache-Combier also reported that 2-methylpyridine
undergoes photoisomerization in the gas phase to yield 4-methylpyridine.
This is in contrast to the report by Roebke that the gas-phase photolysis of
2-methylpyridine yields both 3- and 4-methylpyridine in a ratio of 10:1.4
(4) Roebke, W. J. Phys. Chem. 1970, 74, 4198-4203.
(5) All yields reported are absolute yields determined from the number
of moles of products formed and the number of moles of reactant con-
sumed.

5668 J. Am. Chem. Soc., Vol. 121, No. 24, 1999
Table 2. Photolysis Productsa
PaVlik et al.
Scheme 3
Scheme 4
products. One set, formed in the larger yield, is substantially
quenched when the irradiations are carried out in the presence
of 15-21 Torr of nitrogen and is not formed when the
irradiations are carried out with light of λ > 290 nm. In contrast,
the yields of the second set of products are lower when the
reactant is irradiated at 254 nm but are increased by the addition
of nitrogen and are the only photoisomerization products
observed when the irradiations are carried out with light of λ
> 290 nm.
a Numbers are percent of reactant consumed or the percent yields
of products formed after irradiation. b Irradiation at 254 nm. c Irradiation
at 254 nm in the presence of added N2. d Irradiation at λ > 290 nm.
Based on these major products, the dimethylpyridines can
be divided into two triads, each containing three interconverting
compounds. Triad 1, shown in Scheme 3, consists of 3,5-di-
methylpyridine (6), 2,4-dimethylpyridine (2), and 2,5-dimethyl-
pyridine (3), while triad 2 (Scheme 3) consists of 2,6-di-
methylpyridine (4), 2,3-dimethylpyridine (1), and 3,4-dimethyl-
pyridine (5).
Since irradiation also results in demethylation-methylation
reactions, the possibility was considered that these phototrans-
positions also occur by way of such a process. Although this
explanation seems unlikely in view of the regioselectivity of
these reactions, various photochemical crossover experiments
were carried out in order to determine if these phototransposition
reactions occur via an intermolecular demethylation-methyla-
tion pathway.
Irradiation of 6-deuterio-2-trideuteriomethyl-3-methylpyridine
(1-d4) vapor at 254 nm gave rise (Scheme 4) to the formation
of 4, 5, and 3 in yields of 1.6, 0.6, and 0.5%, respectively, which
were all shown by mass spectral analysis to be d4-labeled. This
indicates that the transposition reaction occurred by way of an
intramolecular process, since if 4, 5, and 3 were formed by
intermolecular demethylation-methylation reactions the deu-
λ > 290 nm. At the same time, however, the yield of
2,3-dimethylpyridine (1) is enhanced by the addition of N2 and
is the only product formed upon irradiation at λ > 290 nm.
Irradiation of 2,4-dimethylpyridine (2), 2,6-dimethylpyridine
(4), 3,4-dimethylpyridine (5), or 3,5-dimethylpyridine (6)
resulted in the formation of two transposition products in each
case. Furthermore, in each case, the yields of both transposition
products are decreased upon irradiation of the reactant in the
presence of added N2 gas and are not formed when the reactant
is irradiated at λ > 290 nm.
Unsymmetrical dimethylpyridines 1, 2, 3, and 5 each also
underwent photodemethylation to yield two different mono-
methylpyridine derivatives formed by loss of either of the two
nonequivalent methyl groups. As expected, 2,6-dimethylpyridine
(4) provided only one demethylation product, 2-methylpyridine
(7), formed by loss of either of the two equivalent methyl
groups. No demethylation product was detected after irradiation
of 3,5-dimethylpyridine (6). In the case of 2,6-dimethylpyridine
(4), a trimethylpyridine (10) was also detected as a product.
These results show that irradiation of dimethylpyridine vapors
at 254 nm results in the formation of two sets of isomerization

Vapor-Phase Photochemistry of Dimethylpyridines
J. Am. Chem. Soc., Vol. 121, No. 24, 1999 5669
Table 3. Crossover Experiments
Scheme 5
in triad 1, 3,5-dimethylpyridine (6) can be viewed as having
the nitrogen inserted between C-2 and C-6 of the pyridine ring.
Because of the symmetry of this molecule, insertion of nitrogen
between C-2 and C-3 or between C-5 and C-6 results in the
second member of the triad, 2,4-dimethylpyridine (2), while
insertion of nitrogen between C-3 and C-4 or between C-4 and
C-5 will result in the formation of 2,5-dimethylpyridine (3),
the third member of the triad. Consideration of the members of
triad 2 leads to a similar conclusion. Thus, because of the
symmetry of 2,6-dimethylpyridine (4), insertion of nitrogen
either between C-2 and C-3 or between C-5 and C-6 results in
2,3-dimethylpyridine (1), the second member of the triad, while
insertion of nitrogen either between C-3 and C-4 or between
C-4 and C-5 results in 3,4-dimethylpyridine (5), the third
member of the triad. Thus, because of the symmetry of 3,5-
dimethylpyridine (6) and 2,6-dimethylpyridine (4), each of these
two compounds can be converted only into the other two
members of their respective triads by a reaction mechanism
involving nitrogen insertion within the carbon skeleton of the
pyridine molecule.
terium content would range from d1 to d7. In the case of 4-d4,
the ring deuterium was determined to be at the C-3 or C-5
1
position since the H NMR spectrum exhibited a doublet at δ
6.9 (J ) 8.4 Hz), which means that the deuterium cannot be at
C-4, since in that case both remaining ring protons would appear
as a singlet in the spectrum. The position of the ring deuterium
could not be determined in the case of 3-d4 or 5-d4 due to their
low yields.
As shown in Scheme 4, irradiation of 6-deuterio-2-trideute-
riomethyl-5-methylpyridine (3-d4) or 2,6-dideuterio-4-trideute-
riomethyl-3-methylpridine (5-d5) vapor at 254 nm led to the
formation of only d4 transposition products 2-d4 (2.2%), 6-d4
(2.8%), and 1-d4 (3.8%) or d5 transposition products 1-d5 (4.9%)
and 4-d5 (2.0%), respectively. In the case of 3-d4, no evidence
could be detected for the formation of any d1 or d7 transposition
products, while in the case of 5-d5, it was not possible to detect
any d2 or d8 phototransposition products that would be expected
if the transpositions involved intermolecular demethylation-
methylation reactions.
A mechanism that allows for this selective nitrogen insertion
involves electrocyclic ring closure, nitrogen migration around
the sides of the cyclopentenyl ring, and rearomatization. Similar
mechanisms have previously been employed to explain the
phototransposition reaction of both five- and six-membered
heteroaromatic compounds,6 and a similar mechanism has been
recently proposed by Pincock et al. to explain the photoisomer-
ization reactions of isomeric cyanotoluenes.7
Additional crossover experiments were carried out by ir-
radiating the deuterated pyridines shown in Table 3 mixed with
equal amounts of the corresponding undeuterated compound.
Analysis of the product mixtures by GLC-MS provided no
evidence for any crossover products that would implicate
intermolecular demethylation-methylation reactions in the
phototransposition reactions. Thus, irradiation of equimolar
mixtures of 1-d4 and 1-d0, 2-d7 and 2-d0, 3-d4 and 3-d0, 4-d6
and 4-d0, 5-d5 and 5-d0, or 6-d2 and 6-d0 led to the detection of
only d4 and d0, d7 and d0, d4 and d0, d6 and d0, d5 and d0, or d7
and d0 phototransposition products, respectively.
This mechanism for triad 1 is shown in Scheme 5. Thus,
electrocyclic ring closure of 3,5-dimethylpyridine (6) results in
the formation of azabicyclohexenyl species 6a, shown here as
a diradical. This structure is analogous to the bicyclic species
originally named prevalene and suggested by Bryce-Smith to
be the photochemically generated precursor of benzvalene.8
(6) See, for example: Pavlik, J. W. Kurzweil, E. M. J. Org. Chem. 1991,
56, 6313-6320. Pavlik, J. W.; Tongcharoensirikul, P.; Bird, N. P.; Day,
A. C.; Barltrop, J. A. J. Am. Chem. Soc. 1994, 116, 2292-2300. Pavlik, J.
W.; Patten, A. D.; Bolin, D. R.; Bradford, K. C.; Clennan, E. L. J. Org.
Chem. 1984, 49, 4523-4531 and references therein.
(7) MacLeod, P. J.; Pincock, A. L.; Pincock, J. A.; Thompson, K. A. J.
Am. Chem. Soc. 1998, 120, 6443-6450.
(8) Bryce-Smith, D.; Longuet-Higgins, H. C. J. Chem. Soc., Chem.
Commun. 1966, 593-594.
An inspection of the isomers in the two triads (Scheme 3)
shows that the compounds of each triad differ in the position
of nitrogen within the carbon skeleton of the pyridine ring. Thus,

5670 J. Am. Chem. Soc., Vol. 121, No. 24, 1999
PaVlik et al.
Scheme 6
product mixture revealed signals at δ 2.61 (CH3 at C-2 for 2
and 3) and singlets at δ 6.9 due to a C-3 or a C-5 ring proton
and at δ 8.3 due to a hydrogen at C-6 of the pyridine ring.
Since the C-3 proton of 3 is known to absorb at δ 7.0 and the
C-3 and C-5 protons of 2 resonate at δ 6.9 and 6.8, respectively,
the observed singlet at δ 6.9 was assigned to the C-3 proton of
2, indicating that the two deuterium atoms are at ring positions
5 and 6, as in 2-d2. This requires that the observed signal at δ
8.3 must be due to the C-6 proton of 3, which places the two
deuterium atoms at positions 3 and 4, as in 3-d2.
The formation of 2-d2 and 3-d2 from 6-d2 is consistent with
the mechanism proposed in Scheme 5. Thus, as shown in
Scheme 7, 2-d2 would be formed by a single sigmatropic shift
of nitrogen, converting 6a-d2 to 2a-d2, followed by rearomati-
zation, while two consecutive shifts of nitrogen followed by
rearomatization will result in the formation of 3-d2 via 3a-d2.
The 2,6-bridging heteroatom migration mechanism accounts
for the interconversions of the three members of each triad but
predicts that there should be no interconversion of the members
of triad 1 with members of triad 2. Experimental results, how-
ever, clearly show that the two triads are connected via the inter-
conversion of 2,5-dimethylpyridine (3), a member of triad 1,
and 2,3-dimethylpyridine (1), a member of triad 2. These inter-
triad reactions clearly require a different mechanistic interpreta-
tion.
Scheme 7
The interconversion of 1 and 3 can be viewed as reversible
1,3-shifts of the methyl group from ring position 3 in 1 to ring
position 5 in 3. Although this interconversion can be rationalized
by a mechanism involving formation and cleavage of an
azaprismane intermediate, such a pathway would allow both
1,2- and 1,3-shifts. Thus, the observed specificity for only a
1,3-shift is inconsistent with an azaprismane mechanism.
A mechanism involving interconverting Dewar pyridine
intermediates, however, is consistent with the observed re-
giospecificity. Scheme 8 shows the products predicted from 2,3-
dimethylpyridine (1) by the interconverting Dewar pyridine
mechanism. Thus, initial 2,5-bridging (Scheme 8a) in the excited
state of 1 would result in the formation of Dewar pyridine 1b.
Isomerization via 1,3-shifts of nitrogen or carbon would convert
1b to the isomeric Dewar pyridines 3b, 3c, 5b, and 5c, and
rearomatization would result in the formation of 2,5-dimeth-
ylpyridine (3) and 3,4-dimethylpyridine (5).
Alternatively, 3,6-bonding in excited 1 (Scheme 8b) would
lead to the formation of Dewar pyridine 1c, which is predicted
to isomerize via 1,3-shifts of nitrogen and carbon, followed by
rearomatization to 2,3-dimethylpyridine (1) and 2,5-dimethyl-
pyridine (3).
This mechanistic pathway thus predicts that 2,3-dimethylpy-
ridine (1), a member of triad 2, should transpose to 2,5-
dimethylpyridine (3), a member of triad 1, by either initial 2,5-
or 3,6-bonding and accordingly accounts for the observed inter-
triad conversion. Interestingly, this mechanism also predicts that
1 should also transpose to 3,4-dimethylpyridine (5), which is
also a member of triad 2 and which is, therefore, also predicted
to result from 1 via the 2,6-bridging heteroatom shift mechanism
(Scheme 6) that accounts for the intra-triad interconversions. It
is possible, therefore, that 1 and 5 interconvert by both
mechanistic pathways, which in the absence of suitable labeling
cannot be distinguished.
Sigmatropic shift of the nitrogen around the five sides of the
cyclopentenyl ring thus allows interconversion of 6a with
azabicyclohexenyl species 2a and 3a which, after rearomatiza-
tion, allows the interconversion of 6 with 2 and 3, the two other
members of the triad. It is not known whether the nitrogen
actually migrates around all five sides of the cyclopentenyl ring.
Indeed, conversion of 2a to 6a can occur by four consecutive
migrations in a counterclockwise manner or by a one-step
clockwise migration of nitrogen. The analogous mechanism for
the interconversion of 1, 4, and 5, the three members of triad 2,
is shown in Scheme 6.
Nitrogen migration around the cyclopentenyl ring could occur
via an azaniabenzvalene intermediate. Although such species
have been implicated in the phototransposition reactions of
N-methylpyridinium cations,9 there is no compelling evidence
to claim their intermediacy in the present case.
Deuterium labeling studies are consistent with the proposed
mechanism. Thus, as shown in Scheme 7, irradiation of 2,6-
dideuterio-3,5-dimethylpyridine (6-d2) in the vapor phase at 254
nm yielded products that were identified by GLC and GLC-
The inter-triad interconversion of 2,3-dimethylpyridine (1)
and 2,5-dimethylpyridine (3) was enhanced when either isomer
was irradiated at 254 nm in the presence of N2 and was the
only phototransposition reaction observed when the reactant was
irradiated with light of wavelengths greater than 290 nm. This
1
MS as 2-d2 and 3-d2. Furthermore, H NMR analysis of the
(9) Kaplan, L.; Pavlik, J. W.; Wilzbach, K. E. J. Am. Chem. Soc. 1972,
94, 3283-3284.

Vapor-Phase Photochemistry of Dimethylpyridines
J. Am. Chem. Soc., Vol. 121, No. 24, 1999 5671
Scheme 8
a
Figure 1. 1H NMR spectrum from δ 3.5 to 7.0 of 2,3-dimethylpyridine
(1) in CD3CN (a) at -30 °C before irradiation and (b) after irradiation
at 254 nm at -30 °C for 2 h.
b
Figure 2. 1H NMR spectrum from δ 3.7 to 7.0 of 2,5-dimethylpyridine
(3) in CD3CN (a) at -30 °C before irradiation and (b) after irradiation
at 254 nm at -30 °C for 2 h.
suggested that the inter-triad reactions might also occur in
condensed media at low temperature, where the proposed Dewar
pyridine intermediates might be observed. Indeed, Wilzbach and
Rausch observed the parent Dewar pyridine by 1H NMR
spectroscopy at -25 °C after irradiation of pyridine at 254 nm
in butane at -15 °C.2
well as the demethylation products, 2-methylpyridine (7) and
3-methylpyridine (8). Neither 2,6-dimethylpyridine (4) nor 3,4-
dimethylpyridine (5), the intra-triad products which were the
major products formed when 1 vapor was irradiated at 254 nm,
could be detected in the present solution.
In an attempt to observe the Dewar pyridine isomers of
dimethylpyridines 1 and 3, each isomer was irradiated at 254
nm in CD3CN solution at -30 °C and examined at that
1
Figure 2 shows a portion of the H NMR spectrum from δ
3.7-7.0 of 2,5-dimethylpyridine (3) in CD3CN at -30 °C (a)
before irradiation and (b) after irradiation of the sample at -30
°C for 2 h. The latter spectrum shows that irradiation results in
the formation of three new signals at δ 4.0, 4.8, and 6.4 that
are of greater intensity than the signals observed from the
irradiation of 1.
These signals are consistent with Dewar pyridine 3c (Scheme
8b), since this isomer would be expected to show one-proton
signals near δ 4.0 and 5.0 for the C-4 and C-1 bridgehead
protons, respectively, and a one-proton signal below δ 6.0 for
the C-5 vinyl proton. The other possible Dewar pyridine 3b
can be excluded since this isomer would exhibit signals only
in the vinyl region between δ 6.0 and 6.5. These new signals
disappeared upon warming of the solution to room temperature.
GLC analysis of the resulting solution showed the formation
of 2,3-dimethylpyridine (1), the inter-triad product, as the major
product, much lower yields of 2,4-dimethylpyridine (2) and 3,5-
dimethylpyridine (6), the intra-triad products, and small quanti-
ties of 2-methylpyridine (7) and 3-methylpyridine (8), the
demethylation products.
1
temperature by H NMR spectroscopy.10
1
Figure 1 shows a portion of the H NMR spectrum from δ
3.5 to 7.0 of 2,3-dimethylpyridine (1) in CD3CN at -30 °C (a)
before irradiation and (b) after the sample had been irradiated
at 254 nm at -30 °C for 2 h. The latter spectrum shows the
formation of three new low-intensity signals of equal area due
to the formation of Dewar pyridine 1c resulting from 3,6-
1
bonding (Scheme 8b) in excited 1. Based on the H NMR
spectrum of the unsubstituted Dewar pyridine,2 1c would be
expected to exhibit two signals near δ 6.5 for the C-5 and C-6
vinyl protons and one signal near δ 5.0 due to the C-1
bridgehead proton. The other possible Dewar pyridine isomer
1b (Scheme 8a), resulting from 2,5-bonding, can be excluded
since this isomer would be expected to show a signal at δ 4.0
due to the C-4 bridgehead proton and only one signal in the
vinyl region near δ 6.5. All of the observed signals disappeared
after the sample was warmed to room temperature. GLC analysis
of the resulting solution showed the formation of a very small
quantity of 2,5-dimethylpyridine (3), the inter-triad product, as
Although both 2,5- and 3,6-bonding were considered in
Scheme 8, the H NMR spectral data in Figure 1 shows that
(10) We would like to thank Dr. Pakamas Tongcharoensirikul for
technical assistance in recording these low-temperature H NMR spectra.
1
1

5672 J. Am. Chem. Soc., Vol. 121, No. 24, 1999
PaVlik et al.
Scheme 9
The sudden decrease in φF suggests the existence of a decay
channel, which has been referred to as “channel three”.15 In
benzene, this decay pathway, which originates in a vibrationally
excited π,π* state, results in the formation of the ground-state
diradical prefulvene, the presumed precursor of benvalene.16
Although the analogous meta-bonded valence isomer of
pyridine, azabenzvalene, has never been isolated or definitively
detected, theoretical calculations predict that the S2 (π,π*) state
of pyridine crosses both the S1 (n,π*) and So states along a
concerted pathway, leading to the ground-state azaprefulvene
diradical,17 the suggested intermediate responsible for intra-triad
isomerizations (Schemes 5 and 6). Quenching of these intra-
triad reactions by added N2 is consistent with their origin from
an excited state possessing excess vibrational energy.
In contrast to the intra-triad reactions, the inter-triad inter-
conversions of 2,3- and 2,5-dimethylpyridines (1) and (3) are
enhanced by the addition of N2, are the major products observed
upon irradiation in condensed media, and are the only photoi-
somerization products formed upon irradiation with light of
wavelength >290 nm. Furthermore, the suggested Dewar
pyridine intermediates (Scheme 8) can also be detected upon
irradiation of 1 or 3 in the condensed phase at low temperature.
Accordingly, these reactions occur from an excited state of lower
energy. A number of facts indicate that this lower energy state
is the T1 state of the dimethylpyridine.
In the case of pyridine and methylpyridines, the quantum yield
for intersystem crossing is known to increase substantially as
the excitation energy is decreased from SofS2 (π,π*) absorption
to the SofS1 (n,π*) region.18,19 Thus, the net effect of decreasing
the excitation energy is to increase the triplet population.
Although no evidence has been detected for any triplet state,
gas-phase chemistry in pyridine,20 photolysis of pyridine at 254
nm in an argon matrix has been shown to result in the formation
of Dewar pyridine.21 Furthermore, this reaction was enhanced
when the matrix was changed to xenon. This enhancement in
xenon indicates strongly that the triplet state of pyridine is
involved in the isomerization.21
2,3-dimethylpyridine (1) undergoes only 3,6-bonding to yield
1c. 1,3-Sigmatropic shift of nitrogen from C-6 to C4 (path B,
Scheme 8b) or of C-4 from C-3 to N (path C, Scheme 8b),
followed by rearomatization of 1c′ (an enantiomer of 1c) or
1d, results only in the regeneration of 1, the reactant. Alterna-
tively, 1c could also rearrange via 1,3-sigmatropic shift of C-2
from C-3 to C-5 (path A, Scheme 8b) or of C-5 from C-6 to
C-2 (path D, Scheme 8b), followed by rearomatization of 3c or
3d, to result in the 2,5-dimethylpyridine (3), the inter-triad
product.
1H NMR spectroscopy (Figure 2) also revealed that 2,5-
dimethylpyridine (3) undergoes 3,6-bonding to form 3c′. 1,3-
Sigmatropic shift of nitrogen from C-6 to C-4 (path A, Scheme
9) would convert 3c′ into its enantiomer 3c. Rearomatization
thus regenerates 2,5-dimethylpyridine (3). 1,3-Sigmatropic shift
of C-2 from C-3 to C-5 (path B, Scheme 9), followed by
rearomatization of 1c′, results in the formation of 2,3-dimeth-
ylpyridine (1), the inter-triad product.
1
The relative intensities of the H NMR spectral signals for
Dewar pyridines 1c and 3c, formed from dimethylpyridines 1
and 3, indicate that, after the same duration of irradiation, the
yield of 3c from 3 is larger than the yield of 1c from 1. This
may be due to the greater stability of 3c, which bears methyl
substituents at both the CdC and CdN double bonds as
compared to 1c, which has one of its methyl groups at a
bridgehead position. These relative yields of the Dewar pyridines
are consistent with the yields of the inter-triad reactions
observed. Thus, as shown in Table 1, 2,5-dimethylpyridine (3),
which in the condensed phase gives Dewar pyridine in the
greater yield, is also converted in the gas phase to the inter-
triad product 2,3-dimethylpyridine (1) in the greater yield.
Conversely, 2,3-dimethylpyridine (1), which in the condensed
phase gives the Dewar pyridine in the lower yield, is also
converted in the gas phase to the inter-triad product 3 in the
lower yield.
The 0-0 bands for SofS1 and SofS2 absorptions in pyridine
vapor occur at 34 769 and 38 350 cm-1, respectively.11 This
corresponds to Es1 (n,π*) and Es2 (π,π) at 99.4 and 109.7 kcal
mol-1, respectively. Methyl substitution in pyridine significantly
lowers the S2 energy level, while the S1 level is nearly
unchanged.11 The energies of the S1 (n, π*) and S2 (π,π*) states
in 2,6-dimethylpyridine (4), for example, are 100.4 and 105.5
kcal mol-1, and the six dimethylpyridines are expected to have
excited-state energies similar to these values.11 Photochemical
excitation of dimethylpyridines with light of 254 nm is thus
expected to result in the population of S2 (π,π*) excited states
with excess vibrational energy.
Both experimental22 and theoretical studies23 point to the
existence of two triplet forms of pyridine with different intrinsic
lifetimes, brought about by strong pseudo-Jahn-Teller vibronic
coupling of the nearly degenerate 3π,π* and 3n,π* states. This
results in a double minimum in the lowest triplet surface,
yielding a vibrationally relaxed triplet state with a boat-shaped
geometry which is suggested to be involved in the photochemi-
cal generation of Dewar pyridine.22
(13) Yamazaki, I.; Murao, T.; Yamanaka, T.; Yoshihara, K. Faraday
Discuss Chem. Soc. 1983, 75, 395-405.
(14) Kaplan, L.; Wilzbach, K. E. J. Am. Chem. Soc. 1968, 90, 3291-
3292.
(15) Riedle, E.; Weber, T.; Shubert, U.; Neusser, H. J.; Schlag, E. W. J.
Chem. Phys. 1990, 93, 967-978. Suzuki, T.; Ito, M. J. Chem. Phys. 1989,
91, 4564-4570. Hornburger, H.; Sharp, C. M. Chem. Phys. 1986, 101,
67-79. Sobolewski, A. L.; Czerminski, R. Chem Phys. 1989, 130, 123-
128. Sobolewski, A. G. J. Chem. Phys. 1990, 93, 6433-6439. Sobolewski,
A. L.; Lim, E. C.; Siebtand, W. Int. J. Quantum Chem. 1991, 39, 309-
324.
(16) Palmer, I. J.; Ragazos, I. N.; Bernardi, F.; Olivucci, M.; Robb, M.
A. J. Am. Chem. Soc. 1993, 115, 673-682.
(17) Sobolewski, A. L.; Domcke, W. Chem. Phys. Lett. 1991, 180, 381-
386.
(18) Yamazaki, I.; Sushida, K.; Baba, H. J. Chem. Phys. 1979, 71, 381-
387.
The nonradiative decay properties of pyridine are reported
to be similar to those of benzene. Thus, the φF in both
pyridine11-13 and benzene14 shows a characteristic dependence
on excess vibrational energy in the lowest π,π* excited state.
(19) Yamazaki, I.; Murao, T.; Yamanaka, T.; Yoshihara, K. Faraday
Discuss. Chem. Soc. 1983, 75, 395-405.
(20) Lemaire, J. J. Phys. Chem. 1967, 71, 612-615.
(21) John Stone, E. E.; Sodeau, J. R. J. Phys. Chem. 1991, 95, 165-
169.
(11) Yamazaki, I.; Sushida, K.; Baba, H. J. Phys. Chem. 1979, 71, 381-
387.
(12) Yamazaki, I.; Sushida, K.; Baba, H. J. Lumin. 1979, 425-428.
(22) Selco, J. I.; Holt, P. L.; Weisman, R. B. J. Chem. Phys. 1983, 79,
3269-3278.
(23) Naqaoka, S.; Nagashima, U. J. Phys. Chem. 1990, 94, 4467-4469.

Vapor-Phase Photochemistry of Dimethylpyridines
J. Am. Chem. Soc., Vol. 121, No. 24, 1999 5673
Materials. Dimethylpyridines were commercially available and were
purified by distillation. Deuterated dimethylpyridines were prepared
by heating the dimethylpyridines in D2O containing potassium carbon-
ate.24
Irradiation Procedures. The dimethylpyridine (0.185 g, 1.73 mmol)
was placed in a Pyrex tube, attached to the vacuum line, and subjected
to three freeze-thaw cycles. The remaining material was then allowed
to vaporize into a 3-L quartz reaction flask which had been evacuated
overnight. The resulting pressure in the reaction flask ranged from 1.10
Torr for 3,5-dimethylpyridine (6) to 3.62 Torr for 2,6-dimethylpyridine
(4). The flask was irradiated in a Rayonet reactor equipped with 16
2537-Å lamps for 3 h.
Irradiation at 254 nm in the Presence of Added N2. After the
dimethylpyridine was allowed to vaporize into the 3-L quartz flask,
N2 gas was added to the flask until the final pressure in the reaction
flask was approximately 20 Torr. The reaction flask was then removed
from the vacuum line and irradiated as above.
Irradiation of λ > 290 nm. The 3-L quartz reaction flask containing
a low pressure of the dimethylpyridine was removed from the vacuum
line and irradiated through a cylindrical Pyrex sleeve in a Rayonet
reactor equipped with 16 3000-Å lamps for 24 h.
Analysis Procedure. After irradiation by one of the above proce-
dures, the 3-L reaction flask was attached to the vacuum line, and the
volatile contents were recovered by pumping it out through a trap cooled
in an acetone-dry ice bath. The contents of the trap were weighed
and dissolved in a known volume of diethyl ether for GLC analysis.
At 80 °C and a helium flow rate of 10.0 mL/min, the three
methylpyridines and the six dimethylpyridines elute in the order
2-methylpyridine (7), 2,6-dimethylpyridine (4), 3-methylpyridine (8),
4-methylpyridine (9), 2,5-dimethylpyridine (3), 2,4-dimethylpyridine
(2), 2,3-dimethylpyridine (1), 3,5-dimethylpyridine (6), and 3,4-
dimethylpyridine (5), with retentions [relative to 2-methylpyridine (7)]
of 1.00, 1.44, 2.32, 2.44, 2.96, 3.24, 3.92, 5.72, and 9.28, respectively.
Quantitative GLC analysis of reactant consumption and product
formation was accomplished using calibration curves constructed for
each methylpyridine and for each dimethylpyridine by plotting detector
responses vs six standards of known concentrations. Correlation
coefficients range from 0.976 to 1.000. Quantitative results for reactant
consumption and product formation after 3.0 h of irradiation are given
in Tables 1 and 2.
Cross-Over Experiments: Irradiation of Pure Deuterated Com-
pounds. 6-Deuterio-2-trideuteriomethyl-3-methylpyridine (1-d4), 6-deu-
terio-2-trideuteriomethyl-5-methylpyridine (3-d4), or 2,6-dideuterio-4-
trideuteriomethyl-3-methylpyridine (5-d5) vapors were irradiated at 254
nm as described for the undeuterated dimethylpyridines. GLC-MS
analyses of the resulting volatile products showed that 1-d4, 3-d4, and
5-d4 were converted to d4, d4, or d5 products, respectively, as shown in
Scheme 4.
Although the demethylation-methylation reactions were not
studied in detail, it appears that these reactions occur by
extrusion or insertion of a CH2 group from CH3 or a CD2 group
from CD3. Thus, irradiation of 6-deuterio-2-trideuteriomethyl-
5-methylpyridine (3-d4) resulted in the formation of 6-deuterio-
2-trideuteriomethylpyridine (7-d4) and 2,6-dideuterio-3-meth-
ylpyridine (8-d2) as a result of the expulsion of CH2 and CD2,
respectively. Similarly, irradiation of 2,6-dideuterio-4-trideute-
riomethyl-3-methylpyridine (2-d5) led to the formation of 2,6-
dideuterio-4-trideuteriomethylpyridine (9-d5) and 2,4,6-trideuterio-
3-methylpyridine (8-d3), also by expulsion of CH2 and CD2,
respectively.
Irradiation of an equimolar mixture of 2,6-dimethylpyridine
(4-d0) and 2,6-bis(trideuteriomethyl)pyridine (4-d6) resulted in
intramolecular phototransposition to provide 1-d0 and 1-d6 and
5-d0 and 5-d6, loss of CH2 and CD2 to give 7-d0 and 7-d4,
respectively, and the formation of a trimethylpyridine as a
mixture of 10-d0, 10-d2, 10-d6, and 10-d8. These would result
by insertion of CH2 into 4-d0 and 4-d6 or insertion of CD2 into
4-d0 and 4-d6.
The distribution of the demethylation products as a function
of the reaction conditions also suggests that they arise from dif-
ferent excited states. Thus, as shown in Table 1, 2,3-dimeth-
ylpyridine (1) undergoes demethylation upon irradiation at 254
nm to yield 2-methylpyridine (7) and 3-methylpyridine (8) in a
ratio of 8:1. When the irradiation of 1 is carried out in the
presence of N2, the yield of 7 is quenched while the yield of 8
is enhanced and is the only product observed when 1 is irradiated
with light of λ > 290 nm. Thus, the formation of 2-methylpy-
ridine (7) follows the profile of the intra-triad products 4 and
5, while the profile of the formation of 3-methylpyridine (8) is
similar to the formation of the inter-triad product 3.
The situation is not as clear, however, in other cases. Thus,
upon irradiation of 2,5-dimethylpyridine (3), the yields of both
demethylation products 7 and 8 are enhanced by the addition
of N2, but only 2-methylpyridine (7), the isomer which is less
affected by N2, is observed upon irradiation at λ > 290 nm.
Interestingly, in the case of 2,4-dimethylpyridine (2) and 2,6-
dimethylpyridine (4), although irradiation at λ > 290 nm does
not result in any photoisomerization products, demethylation
was still observed. In the case of 3,4-dimethylpyridine (5),
neither photoisomerization nor demethylation was observed
upon irradiation at λ > 290 nm.
Irradiation of Mixtures of Deuterated and Underterated Di-
methylpyridines. Equimolar mixtures of 1-d4 and 1-d0, 2-d7 and 2-d0,
3-d4 and 3-d0, 4-d6 and 4-d0, 5-d5 and 5-d0, or 6-d2 and 6-d0 vapors
were irradiated as previously described. GLC-MS analyses of the
resulting volatile products led to the results shown in Table 3.
Experimental Section
1
Instrumentation. H NMR spectra were recorded at 200 or 400
MHz on a Brucker FT-NMR system. GLC analyses were performed
on a PE-9000 FID instrument equipped with a 15-m × 3-µm Carbowax-
20 M bonded phase capillary column at a temperature of 80 °C. Mass
spectra were recorded with an HP5970 mass-selective detector inter-
faced to an HP 5880 capillary gas chromatograph.
JA990773R
(24) Kebede, N.; Pavlik, J. W. J. Heterocycl. Chem. 1997, 34, 685-
686.