Phototransposition Chemistry of 4-Substituted Isothiazoles. The P4 Permutation Pathway

Article abstract of DOI:10.1021/jo980936e

Upon irradiation in the presence of a small quantity of base, 4-substituted isothiazoles undergo photocleavage to yield substituted cyanosulfides, which can be trapped as their benzyl thioether derivatives, and substituted isocyanosulfides. Both products are suggested to arise via initial photocleavage of the sulfur-nitrogen bond, resulting in the formation of a substituted β-thioformylvinyl nitrene, which can rearrange to the observed cyanosulfide, or cyclize to an undetected thioformylazirine. Deprotonation of the azirine leads directly to the isocyanosulfide. The plight of the isocyanosulfide depends on the C-4 substituent. If the substituent is aromatic, the isocyanosulfide is reprotonated at the isocyanide carbon and spontaneously cyclizes to a 4-substituted thiazole, the observed transposition product. If the substituent is not aromatic, the isocyanosulfide is reprotonated at sulfur and the resulting species has a higher energy barrier to cyclization. In these cases, the isocyanosulfides can be observed spectroscopically and can be trapped as their N-formylaminobenzyl thioether derivatives.

Full text of DOI:10.1021/jo980936e

5
592
J . Org. Chem. 1998, 63, 5592-5603
P h ototr a n sp osition Ch em istr y of 4-Su bstitu ted Isoth ia zoles. Th e
4
P P er m u ta tion P a th w a y
J ames W. Pavlik,* Pakamas Tongcharoensirikul, and Kathleen M. French
Department of Chemistry and Biochemistry, Worcester Polytechnic Institute,
Worcester, Massachusetts 01609
Received May 18, 1998
Upon irradiation in the presence of a small quantity of base, 4-substituted isothiazoles undergo
photocleavage to yield substituted cyanosulfides, which can be trapped as their benzyl thioether
derivatives, and substituted isocyanosulfides. Both products are suggested to arise via initial
photocleavage of the sulfur-nitrogen bond, resulting in the formation of a substituted â-thio-
formylvinyl nitrene, which can rearrange to the observed cyanosulfide, or cyclize to an undetected
thioformylazirine. Deprotonation of the azirine leads directly to the isocyanosulfide. The plight
of the isocyanosulfide depends on the C-4 substituent. If the substituent is aromatic, the
isocyanosulfide is reprotonated at the isocyanide carbon and spontaneously cyclizes to a 4-substituted
thiazole, the observed transposition product. If the substituent is not aromatic, the isocyanosulfide
is reprotonated at sulfur and the resulting species has a higher energy barrier to cyclization. In
these cases, the isocyanosulfides can be observed spectroscopically and can be trapped as their
N-formylaminobenzyl thioether derivatives.
In tr od u ction
Phenylisothiazoles and phenylthiazoles undergo pho-
ous solvents and the effects of added bases or acid on
the reaction pathways. In particular, these results show
that the addition of acid or base to the reaction media
can dramatically affect the yield of phototransposition
via the P pathway and can substantially alter the ratio
4
of phototransposition to photocleavage.
totransposition in benzene solution via the P
5
7
-P per-
mutation pathways.1 These permutations are consis-
tent with mechanistic pathways involving initial
electrocyclic ring closure and one or two subsequent
,2
4
sigmatropic shifts of sulfur around the azetine ring. A
Resu lts a n d Discu ssion
perplexing feature of these reactions is the almost
4
complete absence of reaction by the P phototransposition
process, a reaction involving interchange of N-2 and C-3
of the heterocyclic ring. Thus, although it was observed
By analogy with the known photochemistry of pyra-
zoles and isoxazoles, 4-substituted isothiazoles are ex-
pected to be the most reactive isothiazoles via the P
4
that 5-phenylthiazole is formed in less than 1% yield from
pathway. Despite this expectation, Vernin and col-
leagues irradiated 4-phenylisothiazole (1a ) in benzene
solution but were unable to detect the formation of any
5
5
-phenylisothiazole, presumably via a P
4
pathway, nei-
ther 3-phenylisothiazole nor 4-phenylisothiazole was
observed to yield P products in greater than trace
quantities.4 These results are unusual since the P
permutation process is known to be a major pathway for
10
4
phototransposition product. In our laboratory, when a
4
pyrazoles3
,6-8
and isoxazoles which also have two het-
9
eroatoms in adjacent ring positions.
In this paper, we present the results of our study of
the photochemistry of 4-substituted isothiazoles in vari-
(
1) For a discussion of permutation pattern analysis in aromatic
phototransposition chemistry, see: (a) Barltrop, J . A.; Day, A. C. J .
Chem. Soc., Chem. Commun. 1975, 177-179. (b) Barltrop, J . A.; Day,
A. C.; Moxon, P. D.; Ward, R. W. J . Chem. Soc., Chem. Commun. 1975,
7
86-787. (c) Barltrop, J . A.; Day, A. C.; Ward, R. W. J . Chem. Soc.,
Chem. Commun. 1978, 131-133.
2) For five-membered heterocycles containing two heteroatoms,
there are 12 different ways of transposing the five ring atoms resulting
solution of 1a in benzene was irradiated for 30 min, gas
liquid chromatography (GLC) analysis showed that 82%
(
in 12 permutation patterns identified P
permutation patterns, see ref 3.
1
-P12. For a table showing these
(9) (a) Kurtz, D. W.; Schechter, H. J . J . Chem. Soc., Chem. Commun.
1966, 689-690. (b) Ullman, E. G.; Singh, B. J . Am. Chem. Soc. 1966,
88, 1844-1845. (c) Ullman, E. G.; Singh, B. J . Am. Chem. Soc. 1967,
89, 6911-6916. (d) Singh, B.; Zweig, A.; Gallivan, J . B. J . Am. Chem.
Soc. 1972, 94, 1199-1206. (e) Nishiwaki, T.; Nakano, A.; Matsuoka,
J . J . Chem. Soc. C 1970, 1825-1829. (f) Nishiwaki, T.; Fujiyama, F.
J . Chem. Soc., Perkin Trans. 1972, 1456-1459. (g) Wamhoff, H. Chem.
Ber. 1972, 105, 748-752. (h) Good, R. H.; J ones, G. J . Chem. Soc. C
1971, 1196-1198. (i) Goeth, H.; Gagneux, A. R.; Eugster, C. H.;
Schmid, H. Helv. Chim. Acta 1967, 50, 137-142. (j) Padwa, A.; Chen,
E.; Ku, A. J . Am. Chem. Soc. 1975, 97, 6484-6491. (k) Dietliker, K.;
Gilgen, P.; Heimgartner, H.; Schmid, H. Helv. Chim. Acta 1976, 59,
2074-2099.
(
3) Pavlik, J . W.; Kurzweil, E. M. J . Org. Chem. 1991, 56, 6313-
320.
4) Pavlik, J . W.; Tongcharoensirikul, P.; Bird, N. P.; Day, A. C.;
Barltrop, J . A. J . Am. Chem. Soc. 1994, 116, 2292-2300.
5) The yield of 5-phenylthiazole was too low to be able to confirm
6
(
(
the permutation pathway. Since C-2 and C-4 in 5-phenylthiazole are
both bonded to hydrogen, it is not possible to determine their origins
in 5-phenylisothiazole. Accordingly, the transposition could have
occurred by either a P
6) Pavlik, J . W.; Connors, R. E.; Burns, D. S.; Kurzweil, E. M. J .
Am. Chem. Soc. 1993, 115, 7645-7652.
7) Pavlik, J . W.; Kebede, N.; Bird, N. P.; Day, A. C.; Barltrop, J . A.
J . Org. Chem. 1995, 60, 8138-8139.
8) Pavlik, J . W.; Kebede, N. J . Org. Chem. 1997, 62, 8325-8334.
4 9
or P permutation pathway.
(
(
(10) Vernin, G.; Riou, C.; Dou, H. J . M.; Bouscasse, J .; Metzger, J .;
Loridan, G. Bull Soc. Chim. Fr. 1973, 1743-1751.
(
S0022-3263(98)00936-0 CCC: $15.00 © 1998 American Chemical Society
Published on Web 07/23/1998

The P
4
Permutation Pathway
J . Org. Chem., Vol. 63, No. 16, 1998 5593
of the reactant had been consumed and that 3% of the
product, 4-phenylthiazole (2a ), had been formed. No
Sch em e 1
P
4
other volatile product could be detected by GLC analysis.
The photoreaction of 1a in benzene was also monitored
by UV absorption spectroscopy. As expected, the absorp-
tion maximum of 1a at 269 nm was obscured by the
intense absorption of the benzene solvent. When a
-
5
solution of 1a (3.0 mL, 6.2 × 10 M) was irradiated, UV
analysis revealed the formation of a new absorption at
2
90 nm which appeared as a shoulder on the benzene
absorption.
To avoid competition with the solvent for the incident
light, the photoreaction of 1a was also carried out in
diethyl ether solvent. Maeda and Ohashi have previously
reported that in this solvent 1a is converted to â-cy-
anostyryl disulfide 3.11 However, no spectral or analytical
Sch em e 2
data were presented to support this structural assign-
ment.
Either structure 3 or 4 suggests that it was formed
from 2-cyano-2-phenylethene-1-thiol (5a H), the antici-
pated photocleavage product from 4-phenylisothiazole
(1a ).12 The formation of either disulfide 3 or thioether 4
Irradiation of 1a in diethyl ether was monitored by
GLC and, after dilution, by UV absorption spectroscopy.
After irradiation for 30 min, GLC analysis showed the
consumption of 89% of 1a without the formation of any
volatile product. During this irradiation, UV analysis
showed the consumption of 1a and an increase in the
optical density at 290 nm from 0.242 to 0.779 due to the
formation of a photoproduct. This absorption maximum
is at the same wavelength as was observed after irradia-
tion in benzene, indicating that the same nonvolatile
product that formed in diethyl ether is also formed upon
irradiation of 1a in benzene.
1
3
from thiol 5a H can occur either by air oxidation or via
1
4
an aldol condensation of the enethiol and its thioalde-
hyde tautomer15 with elimination of H S as shown in
2
Scheme 1.
In an attempt to trap the suggested photocleavage
product, a solution of 1a in diethyl ether was irradiated
until after 30 min the optical density at 290 nm was
maximized. Triethylamine (TEA) was added to the
solution. This caused the absorption maximum to shift
from 290 to 340 nm which is consistent with deprotona-
tion of cyanothiol 5a H, resulting in the formation of its
conjugate base, cyanosulfide 5a (Scheme 2). As expected,
this change was reversible. Thus, if the TEA was
neutralized with HCl, the absorption maximum shifted
back to 290 nm. This confirms that the species absorbing
at 290 and 340 nm are the acidic and basic forms of the
same compound, respectively. To trap cyanosulfide 5a
as its benzyl thioether derivative, benzyl bromide was
added to the solution absorbing at 340 nm. This resulted
in the disappearance of the absorption band at 340 nm
and the formation of a new absorption band with a
maximum at 318 nm confirming that a reaction had
taken place.
On a preparative scale, a solution of 1a in diethyl ether
was irradiated for 30 min. Evaporation of the diethyl
ether provided a viscous yellow oil which exhibited an
-1
infrared absorption at 2212 cm indicating the presence
of a nitrile functional group. This oil solidified after
standing briefly at room temperature to yield a solid
which was purified by repeated recrystallization. The
1
13
H and C NMR and IR spectral data are consistent with
the â-cyanostyryl moiety of the disulfide structure 3
1
suggested by Maeda and Ohashi. Thus, the H NMR
spectrum shows only a one-proton singlet at δ 7.66 due
to the vinyl proton and a five-proton multiplet at δ 7.41-
1
3
7
.58 due to the five phenyl protons. The C NMR
1
spectrum exhibits signals at δ 115.32, 137.77, and 114.37
due to the cyano and vinyl carbons and signals at δ
The H NMR spectrum of the residue obtained from
the above reaction mixture revealed the presence of
unreacted 4-phenylisothiazole (1a ) and unreacted benzyl
bromide. The spectrum also exhibited two singlets at δ
4.14 and 4.00 in a ratio of 3.5:1 for two sets of benzyl
protons. The infrared spectrum exhibited a strong
1
25.69, 129.63, 129.97, and 132.62 for the carbon atoms
of the phenyl ring. Furthermore, the infrared spectrum
-1
exhibited an intense absorption at 2208 cm character-
istic of a nitrile functional group.
-
1
Although these spectral data are consistent with the
â-cyanostyryl structure, the mass spectrum exhibited a
molecular ion at m/z 288, consistent with thioether
structure 4, rather than at m/z 320 as required by the
disulfide structure 3. Elemental analyses, however,
repeatedly gave results that were slightly low for carbon
and slightly high for sulfur, suggesting that the thioether
absorption at 2207 cm indicating the presence of a
-1
nitrile functional group and a weaker signal at 2103 cm
where isocyanides are known to absorb. Preparative
(12) Although isothiazoles have not been reported to undergo
photocleavage to R,â-unsaturated cyanothiols, the analogous photo-
cleavage of pyrazoles to enaminonitriles is well-known. See ref 3 and
literature cited therein.
4
was contaminated with a small amount of the disul-
(13) Tarbell, D. S. In The Chemistry of Organic Sulfur Compounds;
Kharasch, N., Ed.; Pergamon Press: Oxford, 1961; Vol. 1, pp 97-102.
fide 3.
(
14) Campaigne, E.; Moss, R. D. J . Am. Chem. Soc. 1954, 76, 1269-
1
271.
(15) Reyes, A.; Silverstein, R. M. J . Am. Chem. Soc. 1958, 80, 6367-
6372, 6373-6379.
(11) Ohashi, M.; Iio, A.; Ezaki, A.; Yonezawa, T. Heterocycles 1975,
3
, 69.

5
594 J . Org. Chem., Vol. 63, No. 16, 1998
Pavlik et al.
layer chromatography led to the isolation of a white
crystalline nitrile, as evidenced by the strong absorption
band in the infrared spectrum at 2207 cm-1, but did not
allow isolation of an isocyanide. The mass spectrum of
the nitrile exhibited a molecular ion at m/z 251, consis-
tent with benzyl thioether 6a , expected by reaction of
cyanosulfide 5a with benzyl bromide as shown in Scheme
intensity absorption band at 2075 cm-1. Although this
new band may be due to the formation of an isocyanide,
-1
the absorption position at 2075 cm is outside of the
-
1
2100-2150 cm range where unsaturated isocyanides
generally absorb.20 Vinyl isocyanide, for example, ab-
-1 21
sorbs in the infrared at 2125 cm
.
Although both of
these absorption bands were still observed after warming
to 100 K, only the nitrile absorption band, which shifted
2
. Furthermore, the mass spectrum also exhibited an
intense signal at m/z 91 as expected for a benzyl deriva-
tive.
-1
to 2211 cm , was still present at 200 K. During this
warming, the species responsible for the absorption band
The 1H NMR spectrum exhibits a singlet at δ 4.14
assigned to the benzyl protons and a multiplet at δ 7.24-
-1
at 2075 cm could have either reacted or evaporated.
-1
Finally, at 250 K, the absorption band at 2211 cm was
7
.54 due to the aromatic protons with an integral ratio
still present but greatly diminished in intensity.
1
3
of 1:5. The C NMR spectrum also exhibits signals at δ
8.33 and 116.16 consistent with benzyl and nitrile
The photoreaction took a different path when it was
carried out in methanol solvent. Thus, after a solution
3
carbon atoms, respectively. These spectral data are
consistent with the 2-cyano-2-phenylethenylbenzyl thio-
-2
of 1a (10.0 mL, 2.0 × 10 M) in methanol was irradiated
for 30 min, GLC analysis showed that 74% of the reactant
was consumed and that 4-phenylthiazole (2a ) was formed
in 38% yield. In addition to the consumption of 1a and
the formation of 2a , analysis by UV absorption spectros-
copy showed an increase in the optical density at 340 nm
from 0 to 0.287. This absorption maximum is identical
to that of the basic form of the photocleavage product
1
13
ether structure 6a . Since the H and C NMR spectra
each exhibit one signal in the benzyl region at δ 4.14 and
3
8.33, respectively, the isolated thioether 6a apparently
1
consists of a single isomer. Although no H NMR signal
appears at δ 7.04 where the vinyl proton of the E isomer
is estimated to absorb,16 the region at δ 7.27 where the
vinyl proton of the Z isomer is estimated to absorb is
obscured by the signals due to the phenyl protons.
Despite this, the product is tentatively assigned the Z
configuration of 6a due to the absence of a signal for the
E isomer and because the Z isomer is expected to be
formed in the ring cleavage reaction.
Isolation of benzyl thioether 6a is excellent evidence
that irradiation of 4-phenylisothiazole (1a ) in diethyl
ether results in the formation of 2-cyano-2-phenylethene-
5
a H. Thus, in methanol solvent, 4-phenylisothiazole (1a )
undergoes both phototransposition to 2a and photocleav-
age to yield cyanosulfide 5a .
Since C-2 and C-5 of 4-phenylthiazole (2a ) are both
bonded to hydrogen, either of these two ring atoms could
have originated at either ring position 3 or 5 in the
reactant. As a result, this phototransposition can be
4
rationalized by either a P or a P11 permutation pathway.
1
-thiol (5a H). The infrared spectrum of the crude
This ambiguity was removed by studying the phototrans-
position chemistry of 5-deuterio-4-phenylisothiazole (1b),
prepared by the base-catalyzed deuteration of 4-phenyl-
isothiazole (1a ). The mass spectrum of the deuterated
product exhibited a molecular ion at m/z 162, consistent
product mixture, however, also suggested that an iso-
cyanide was also being formed in the photoreaction of
4
-phenylisothiazole (1a ). This possibility is of particular
interest since Ferris and colleagues have detected iso-
4
cyanides at low temperatures in the analogous P pho-
with monodeuteration, and an intense peak at m/z 135
totransposition reactions of 4-substituted isoxazoles1
7,18
due to the loss of HC-3tN,2
2,23
indicating that the
while work in this laboratory has recently shown that
isocyanides are general intermediates in the P pho-
4
1
deuterium is located at ring position 5. The H NMR
spectrum confirmed this by showing a 1-H singlet at δ 8.75
where the C-3 proton is known to absorb but no signal
totransposition of N-methylpyrazoles to N-methylimida-
zoles.7
,8
2
4
at δ 8.65 where the C-5 proton is expected to appear.
In an attempt to detect the isocyanide intermediate in
this reaction, the photochemistry of 4-phenylisothiazole
5
-Deuterio-4-phenylisothiazole (1b) is a reactant in
which each ring position is uniquely substituted. Product
identification thus allows unambiguous assignment of the
permutation pattern. A solution of 5-deuterio-4-phenyl-
isothiazole (1b) in methanol was irradiated for 30 min.
GLC-MS analysis of the 4-phenylthiazole photoproduct
showed a molecular ion at m/z 162 confirming that the
loss of deuterium had not occurred during the photolysis.
The mass spectrum also exhibited an intense signal at
(
1a ) was briefly investigated in an argon matrix at 12
1
9
K. After irradiation, infrared analysis at 12 K showed
the formation of a new absorption band at 2225 cm-1
,
consistent with the formation of a nitrile, as well as new
-
1
absorption bands at 1480, 814, 770, and 722 cm where
-phenylthiazole (2a ) is known to absorb. Thus, even
4
before the matrix was allowed to warm, there is evidence
that some 4-phenylthiazole (2a ) has been formed. On
warming to 50 K, the nitrile band was observed to shift
m/z 135 due to loss of HC-2tN, indicating that the
-
1
22,23
from 2225 to 2213 cm
.
This is most likely due to
deuterium is at C-5 of the 4-phenylthiazole ring.
The
1
evaporation of the matrix rather than reaction since all
other bands in the spectrum were also observed to shift
similarly. More importantly, however, warming to 50 K
was also accompanied by the formation of a new low-
H NMR spectrum confirmed this by exhibiting a sharp
singlet at δ 8.86 where H-2 of 4-phenylthiazole is known
(20) J uchnovski, I. N.; Binev, I. G. in Supplement C: The Chemistry
of the Triple-Bonded Functional Groups; Patai, S., Rappoport, Z., Eds.;
Wiley: New York, 1983; pp 107-135.
(
16) Frieboln, H. Basic One- and Two-Dimensional NMR Spectros-
copy, 2nd ed.; VCH Publishers: New York, 1993; pp 139-143.
17) Ferris, J . P.; Antonucci, F. R. J . Am. Chem. Soc. 1974, 96, 2010-
(21) Bolton, K.; Owen, N. L.; Sheridan, J . Spectrochim. Acta 1970,
26, 909-917.
(
2
014, 2014-2019.
(
(
(22) Salmona, G.; Vincent, E.-J . Org. Mass Spectrom. 1978, 13, 119-
120.
(23) Clarke, G. M.; Grigg, R.; Williams, D. H. J . Chem. Soc. B 1966,
339-343.
(24) Staab, H. A.; Mannschreck, A. Chem. Ber. 1965, 98, 1111-1121.
18) Ferris, J . P.; Trimmer, J . J . Org. Chem. 1976, 41, 13-19.
19) The matrix isolation experiment was carried out by Max P.
Bernstein and Scott A. Sandford, NASA-Ames Research Center,
Moffett Field, CA.

The P
4
Permutation Pathway
J . Org. Chem., Vol. 63, No. 16, 1998 5595
to absorb but there is no signal for H-5 at δ 7.52.25 This
shows that 5-deuterio-4-phenylisothiazole (1b) had been
Sch em e 3
transposed to 5-deuterio-4-phenylthiazole (2b) via a P
4
permutation pathway which involves only the inter-
change of N-2 and C-3 of the isothiazole ring.
The photoreaction in methanol was found to be very
dependent on the purity of the methanol used. The above
results were obtained only when the methanol was
freshly distilled from magnesium methoxide. When the
photoreaction was carried out in unpurified methanol,
4
-phenylthiazole (2a ) could not be detected either by UV
absorption spectroscopy or by GLC analysis. Under these
conditions, UV analysis showed that the optical density
at 269 nm decreased from 0.862 to 0.622 due to consump-
tion of 1a , while the optical densities at 290 and 340 nm
increased from 0.322 to 0.756 and from 0 to 0.427,
respectively. This shows that if acidic impurities are not
removed from the methanol solvent, 4-phenylisothiazole
(1a ) undergoes only photocleavage to yield a mixture of
cyanothiol 5a H and cyanosulfide 5a .
Additional experiments have also shown that the yield
of the phototransposition product is substantially en-
hanced if the irradiation is carried out in methanol
containing a small quantity of an added base. Thus,
when three solutions of 4-phenylisothiazole (1a ) (3.0 mL,
and after irradiation of a solution of 1a in methanol (3.0
-
4
mL, 1.15 × 10 M) containing 1.0 µL of concentrated
HCl showed that consumption of 1a was not accompanied
by an increase in the optical density at 250 nm, indicating
that in acid solution the phototransposition is completely
-
2
-2
26
2
2
.0 × 10 M) in methanol containing 0, 1.0 × 10 , or
suppressed. Instead, irradiation was accompanied by
-
1
.4 × 10 M TEA were simultaneously irradiated on a
an increase in the optical denisty at 290 nm to a value
of 1.18, indicating the formation of cyanothiol 5a H. If
after irradiation the HCl was neutralized by the addition
of aqueous ammonia, the absorption band at 290 nm
disappeared while the optical density at 340 nm due to
the formation of cyanosulfide 5a increased to 1.35. On
the basis of the optical densities, the yield of cyanosulfide
5a is approximately 5 times greater when it is formed
indirectly via cyanothiol 5a H than when it is formed
directly by irradiation of 1a in methanol containing
aqueous ammonia. Finally, experiments show that nei-
ther cyanothiol 5a H nor cyanosulfide 5a is photochemi-
cally converted to either 1a or 2a . These results are
summarized in Scheme 3.
merry-go-round apparatus, GLC analysis showed that
consumption of 4-phenylisothiazole (1a) decreased slightly
from 82% in the absence of TEA to 75 or 76% in the
-2
-2
presence of 1.0 × 10 or 2.4 × 10 M TEA, respectively.
At the same time, however, the yield of 4-phenylthiazole
(2a ) increased dramatically from 37% in the absence of
TEA to 86 or 82% in the presence of TEA. GLC analysis
also showed that TEA is not consumed during these
irradiations. This increase in the yield of the phototrans-
position product is not dependent on the particular base
used. Thus, essentially identical results were observed
when the photolysis was carried out in methanol contain-
ing small quantities of aqueous sodium bicarbonate,
aqueous ammonia, or propylamine.
Convincing evidence for the involvement of an isocya-
Interestingly, the effect of added base is not observed
only in methanol solvent. Thus, when three solutions of
4
nide intermediate in the P isothiazole-to-thiazole pho-
totransposition was obtained by studying the photochem-
istry of 4-benzylisothiazole (1c).
-
2
4
-phenylisothiazole (1a ) (3.0 mL, 2.0 × 10
M) in
-
2
-1
benzene containing 0, 1.0 × 10 , or 2.4 × 10 M TEA
were simultaneously irradiated on the merry-go-round,
GLC analysis showed that the consumption of 4-phenyl-
isothiazole (1a ) increased from 63% in the absence of TEA
Irradiation of a solution of 1c in methanol containing
-
2
ammonia (7.5 × 10 M) was monitored by GLC and,
after dilution, by UV absorption spectroscopy. After
irradiation for 30 min, GLC analysis showed the presence
of unreacted 4-benzylisothiazole (1c) (22%) and the
formation of a single product that was identified by its
spectroscopic properties as 4-benzylthiazole (2c) formed
in 51% yield.
-
2
-1
to 77 or 65% in the presence of 1.0 × 10 or 2.4 × 10
M TEA, respectively. The yield of 4-phenylthiazole (2a )
increased from 3% in the absence of TEA to 72 or 68% in
-
2
-1
the presence of 1.0 × 10
or 2.4 × 10
M TEA,
respectively. Again, TEA was not consumed during these
irradiations.
The mass spectrum of the photoproduct, which shows
a molecular ion at m/z 175, and the elemental analysis
are consistent with C10
9
H NS, confirming that the pho-
Finally, in either methanol or benzene containing TEA,
toproduct is an isomer of 4-benzylisothiazole (1c). The
5
5
-deuterio-4-phenylisothiazole (1b) was also converted to
-deuterio-4-phenylthiazole (2b), showing that the added
1
H NMR spectrum of the photoproduct exhibits signals
at δ 8.75 and 6.86 where the C-2 and C-5 protons of the
base has not changed the permutation pathway for the
phototransposition.
Whereas the presence of base enhances the yield of the
phototransposition product, the addition of acid has the
opposite effect. Thus, the UV absorption spectrum before
thiazole ring are known to absorb, respectively, but no
2
5
signal at δ 7.98 where the C-4 proton is expected.
Furthermore, as anticipated for protons in ring positions
(26) Addition of 1.0 µL of concentrated HCl does not change the UV
absorption spectrum, indicating that 1a is not protonated under these
conditions. This is consistent with the low basicity of isothiazoles.
See: Clementi, S.; Forsythe, P. P.; J ohnson, C. D.; Katritzky, A. A.;
Terem, B. J . Chem. Soc., Perkin Trans. 2 1974, 399-402.
(
25) Aune, J . D.; Dou, H. J . M.; Crousier, J . In Thiazoles and Its
Derivatives, Part One; Metzger, J . V., Ed.; Wiley: New York, 1979;
Vol. 34, Part 1, pp 342-347.

5
596 J . Org. Chem., Vol. 63, No. 16, 1998
Pavlik et al.
2
and 5, the spectrum shows that these protons are spin
converted to 4-benzylthiazole (2c) during refluxing or
under the conditions of our GLC and H NMR analyses.
1
coupled with a J
d
of 2.05 Hz. Finally, the signal at δ
6
.86 due to the C-5 proton is further split into a triplet
To gain further information about the intermediate or
intermediates absorbing at 294 nm, an identical photo-
of doublets (J ) 1.99 Hz) due to its coupling with the
t
-
2
methylene protons of the benzyl group at position 4 of
the thiazole ring. The 13C NMR spectrum is also con-
sistent with the proposed product. The spectrum thus
exhibits signals at δ 152.77 and 114.07 where C-2 and
C-5 of the thiazole ring are known to resonate, respec-
tively.25 The signal of C-4 is shifted from δ 143.7 in
unsubstituted thiazole to δ 157.20 due to the benzyl
group bonded to that carbon. As expected for a quater-
nary carbon, the intensity of this signal is low. Finally,
the spectrum also exhibits a signal at δ 37.69 due to the
methylene group of the benzyl group.
reaction (10.0 mL, 2.0 × 10 M, 50 µL of aqueous
ammonia) was carried out. Aliquots (2.0 mL) of the
resulting solution were analyzed by infrared spectroscopy
immediately after irradiation and after the solution
remained in the dark at room temperature for 12 and 48
h. Immediately after irradiation, the infrared spectrum
-
1
exhibited absorption signals at 2207 and 2104 cm in a
ratio of 1:1.2 indicating the presence of a nitrile and
suggesting that an isocyanide was also present. After
12 and 48 h, the relative intensities of these bands
changed from 1:1.2 immediately after photolysis to 1:0.96
after 12 h and 1:0.5 after 48 h. At the same time, the
spectrum revealed an increase in an absorption band at
The resulting solution was evaporated to dryness and
1
examined by H NMR. The spectrum showed the singlet
-
1
1
407 cm where 4-benzylthiazole (2c) absorbs. These
at δ 4.02 due to the benzyl protons of unreacted 4-ben-
zylisothiazole (1c) and a singlet at δ 4.17 due to the
benzyl protons of the photoproduct, 4-benzylthiazole (2c).
The spectrum was quite clean and showed only the
presence of the unreacted 4-benzylisothiazole (1c) and
the photoproduct 2c.
results indicate that 4-benzylisothiazole (1c) is photo-
chemically converted to a mixture of a nitrile and an
isocyanide and that the isocyanide is thermally converted
to the phototransposition product, 4-benzylthiazole (2c).
To further investigate the structures of the photo-
chemically generated primary products, a solution of
The P
confirmed by showing that 5-deuterio-4-benzylisothiazole
1d ) undergoes photoisomerization in methanol/ammonia
4
permutation pattern for the transposition was
4
-benzylisothiazole (1c) in methanol containing ammonia
was irradiated until after 30 min the absorption at 294
nm was maximized. Benzyl bromide was added to the
solution, resulting in a change in the UV absorption
maximum from 294 to 278 nm. The infrared spectrum
of the residue resulting from this solution showed strong
(
to yield 5-deuterio-4-benzylthiazole (2d ). The location of
the deuterium at ring position 5 in the photoproduct was
confirmed by the H NMR spectrum which exhibited a
one-proton singlet for the C-2 proton at δ 8.75 but no
signal at δ 6.86 where the C-5 proton is known to absorb.
1
-1
absorption bands at 2200 and 2100 cm due to nitrile
and isocyanide functional groups. This residue was
treated with glacial acetic acid. After neutralization,
infrared analysis showed no change in the absorption
1
Although both GLC and H NMR show the clean
photoconversion of 4-benzylisothiazole (1c) to 4-ben-
zylthiazole (2c), analysis by UV absorption spectroscopy
shows that 2c is not a primary product in this reaction.
-1
band at 2200 cm , as expected for a nitrile functional
group. In contrast, the absorption band present before
-
2
Thus, a solution of 1c (10.0 mL, 2.0 × 10
M) in
-1
acid treatment at 2100 cm was replaced by a new
methanol containing aqueous ammonia (50 µL) was
irradiated for 30 min. Analysis by UV absorption spec-
troscopy of an aliquot of the resulting solution after
dilution showed a decrease in the optical density at 251
nm due to the consumption of 4-benzylisothiazole (1c)
and to an increase in the optical density at 294 nm from
-1
strong band at 1683 cm . This is consistent with the
acid-catalyzed hydrolysis of the isocyanide to a forma-
mide functional group.2
7,28
Preparative layer chroma-
tography allowed the isolation of unreacted benzyl bro-
mide, unreacted 4-benzylisothiazole (1c), 4-benzylthiazole
(2c), and two new compounds identified as 2-cyano-3-
phenylpropen-1-ylbenzyl thioether (7c) and 2-(N-formy-
lamino)-3-phenylpropen-1-ylbenzyl thioether (8c).
0
to 0.861. Neither the reactant nor the expected
photoproduct, 4-benzylthiazole (2c), absorbs at 294 nm.
Interestingly, after a portion (2.0 mL) of the irradiated
solution was refluxed for 2 h, UV absorption spectroscopy
showed that the optical denisty at 294 nm decreased from
0
4
0
.861 to 0.369 while the optical density at 248 nm, where
-benzylthiazole (2c) absorbs, increased from 0.632 to
.887. This indicates that a portion of the species
absorbing at 294 nm was converted to 4-benzylthiazole
2c) during the reflux. The remaining portion of the
(
refluxed solution (1.5 mL) and the portion of the irradi-
ated solution that had not been refluxed were each
2
-Cyano-3-phenylpropen-1-ylbenzyl thioether (7c) was
isolated as a pale yellow liquid. As required by the
assigned structure, the mass spectrum exhibited a mo-
lecular ion at m/z 265 while the infrared spectrum
showed an absorption band for the nitrile functional
group at 2200 cm . Although the isolated product 7c
appeared to be homogeneous by GLC and TLC, the H
NMR spectrum revealed the presence of two sets of
signals suggesting the presence of two geometrical iso-
mers. Thus, in addition to a 10-H multiplet absorbing
1
concentrated and analyzed by H NMR. In both cases,
1
the H NMR spectra showed singlets at δ 4.02 and 4.17
due to the benzyl protons of 4-benzylisothiazole (1c) and
4
-benzylthiazole (2c), respectively, in a ratio of 1:2.54
after refluxing) or 1:2.49 (without refluxing). Both
spectra were very clean and showed only the reactant
-benzylisothiazole (1c) and the photoproduct 4-ben-
-1
(
1
4
zylthiazole (2c). These results show that upon irradia-
tion in methanol containing ammonia 4-benzylisothiazole
(
1c) is photochemically converted to a primary product
(
27) Drenth, W. Rec. Trav. Chim. 1962, 81, 319-322.
or products absorbing at 294 nm which are thermally
(28) Lim, Y. Y.; Stein, A. R. Can. J . Chem. 1971, 49, 2455-2459.

The P
4
Permutation Pathway
J . Org. Chem., Vol. 63, No. 16, 1998 5597
at δ 7.07-7.36, the spectrum also showed two 1-H triplets
for the vinyl protons at δ 7.05 (J ) 0.90 Hz) and 6.66 (J
The 1H and 13C NMR spectra of this product again
show two sets of signals suggesting the presence of two
)
1.34 Hz), two 2-H broad singlets for the benzyl protons
isomers. Thus, in DMSO-d
6
, the 1H NMR spectrum
at δ 3.46 and 3.51, and two 2-H sharp singlets for the
thiobenzyl protons at δ 3.96 and 4.02.
exhibits signals for the amino proton as a doublet at δ
9.57 (J ) 10.72 Hz) and a broad singlet at δ 9.38 while
the aldehyde proton is observed as two doublets at δ 8.18
(J ) 10.72 Hz) and 8.01 (J ) 1.72 Hz). As required by
The signals at δ 7.05 and 6.66 were assigned to the
vinyl protons of the E and Z isomers, respectively, since
in the E configuration the vinyl proton is situated in the
deshielding zone of the cyano group and is expected to
resonate downfield from the vinyl proton in the Z isomer.
Similarly, the vinyl proton of (E)-3-methoxyacrylonitrile
is reported to appear at δ 7.25, while in the Z isomer,
the vinyl proton is upfield at δ 6.75.29
2
these assignments, upon addition of D O, the signals
assigned to the amino proton disappeared while the
doublets assigned to the aldehyde proton collapsed to two
broad singlets.
1
In CDCl
3
, the aldehyde proton appears in the H NMR
spectrum as two doublets at δ 8.19 (J ) 11.12 Hz) and
1
3
The absorptions at δ 3.51 and 3.46 and at δ 4.02 and
.96 were assigned to the benzyl and thiobenzyl protons
8.08 (J ) 4.0 Hz) and correlate with signals in the
C
3
NMR spectrum at δ 161.34 and 158.94, respectively,
which were therefore assigned to the carbonyl carbons
of the formamide group. These chemical shifts are
similar to those of the analogous carbons in N-vinylfor-
mamide which resonate at δ 163.50 and 159.16.
of the E and Z isomers, respectively, on the basis of the
results of NOE experiments.30 Thus, when the decoupler
was set on the vinyl proton absorption at δ 6.66, the
signal at δ 3.46 due to the benzyl protons was enhanced
due to the NOE effect while the signal at δ 3.96 due to
the thiobenzyl protons was slightly enhanced due to the
decoupling. This confirms that the vinyl and benzyl
protons absorbing at δ 6.66 and 3.46 are situated in close
spatial proximity as required by the Z configuration and
that the signal at δ 3.96 is due to the thiobenzyl protons
of the same isomer. Conversely, when the decoupler was
set on the vinyl proton absorption at δ 7.05, the thioben-
zyl proton signal at δ 4.02 was slightly enhanced, but no
NOE enhancement of the signal at δ 3.51 due to the
benzyl protons was observed. This shows that these sets
of benzyl and vinyl protons are not situated in close
spatial proximity and were accordingly assigned to the
E isomer.
6
In DMSO-d , the benzyl protons, thiobenzyl protons,
and vinyl protons each appear as pairs of singlets at δ
3.59 and 3.52, at δ 3.90 and 3.88, and at δ 6.89 and 5.89,
respectively. In CDCl , the vinyl protons appear at δ 7.14
3
1
3
and 5.77 and correlate with signals in the C NMR
spectrum at δ 112.78 and 109.38 which were therefore
assigned to the C-2 vinyl carbon. The 13C NMR spectrum
also exhibits signals at δ 133.03 and 136.38 which do not
1
correlate with any signals in the H NMR spectrum and
were therefore assigned to the quaternary vinyl carbon
atoms.
1
Substantial changes were observed when the H NMR
6
spectrum was recorded in DMSO-d at 80 °C. First, the
two pairs of singlets due to the benzyl and thiobenzyl
protons each coalesced into two broad singlets at δ 3.58
and 3.89, while the singlets due to the vinyl protons
broaden and shift to δ 6.85 and 5.92. More significantly,
the two doublets due to the aldehyde proton collapsed to
two broad singlets at δ 8.15 and 8.02, while the doublet
and broad singlet due to the amino group changed to two
broad singlets at δ 9.38 and 9.23, respectively. These
changes were reversed upon cooling the solution to 20
°C.
The 13C NMR spectrum is also consistent with the
assigned structure. The spectrum exhibits signals at δ
1
18.87 and 117.02 for the nitrile carbons and at δ 108.37
and 145.56 and at δ 108.73 and 144.45 for the R- and
â-vinyl carbon atoms of the E and Z isomers. These
chemical shifts are consistent with the expected polariza-
tion due to the resonance interaction of the nitrile and
thiobenzyl moieties. Similarly, the E and Z isomers of
1
3
3
-methoxyacrylonitrile in the C NMR spectrum are
reported to resonate at δ 74.84 and 165.23 and at δ 74.27
The temperature-dependent changes show that the
pairs of signals observed for the various sets of protons
are not due to the existence of geometrical isomers but
that the isolated photoproduct 8c consists of a single
stereoisomer with two rotamers due to restricted rotation
and 164.32.29
As demanded by the proposed structure, the 13C
spectrum should also exhibit four signals due to the
benzyl and thiobenzyl carbon atoms. These were ob-
served for the benzyl carbons of the E and Z isomers at
δ 36.49 and 40.16 and for the thiobenzyl carbons of the
E and Z isomers at δ 38.07 and 36.95, respectively.
These signals were confidently assigned on the basis of
32
around the C-N bond in the formamide group. Since
decoupling of the vinyl proton was not accompanied by
an NOE enhancement of the signal due to the benzyl
protons, the structure of the isolated product was as-
signed the E configuration with the two rotamers as
shown below.
1
3
1
31
C- H NMR correlation analysis.
2
-(N-Formylamino)-3-phenylpropen-1-ylbenzyl thioet-
her (8c) was also isolated as a yellow oil. The mass
spectrum, which exhibited a molecular ion at m/z 283 and
a base peak at 91 due to the benzyl cation radical, and
the infrared spectrum, which confirmed the presence of
the formamide functional group by showing an intense
-
1
absorption band at 1683 cm , were consistent with the
assigned structure.
On the basis of the Karplus equation,33 the rotamer in
which the amino and aldehyde protons absorb at δ 9.38
(
29) Pouchet, C. J .; Behnke, J . The Aldrich Library of 13C and 1
H
NMR Spectra, 1st ed.; Aldrich Chemical Co.,1993; Vol. 1, spectrum
no. 1374A.
(30) Neuhaus, D.; Williamson, M. P. The Nuclear Overhauser Effect
(32) Kemp, W. Organic Spectroscopy, 3rd ed.; W. H. Freeman: New
York, 1991; p 130.
(33) Pavia, D. L.; Lampman, G. M.; Friz, G. S. Introduction to
Spectroscopy, 2nd ed.; Saunders: New York, 1996; pp 193-194.
in Structural and Conformational Analysis; VCH Publishers: New
York, 1989; pp 380-386.
(31) See ref 16, pp 203-250.

5
598 J . Org. Chem., Vol. 63, No. 16, 1998
Pavlik et al.
and 8.01 (J ) 1.72 Hz) was assigned to the structure in
which these protons are syn with respect to each other
while the rotamer in which the amino and aldehyde
protons absorb at δ 9.57 and 8.18 (J ) 10.72 Hz) was
assigned to the structure in which these protons are anti
with respect to each other. NOE experiments confirmed
these assignments.
Sch em e 4
Isolation of nitrile 7c and formamide 8c provides
excellent evidence that the initial photoproducts absorb-
ing at 294 nm are a mixture of cyanosulfide 9c and
isocyanosulfide 10c.
As in the case of 4-phenylisothiazole (1a ), the presence
of acid has a substantial effect on the photochemistry of
4
1
-benzylisothiazole (1c). When a solution of 1c (3.0 mL,
.6 × 10 M) in methanol containing ammonia was
-
4
irradiated for 10 s, the optical density at 251 nm due to
trum of the residue remaining after evaporation of this
solution showed sharp absorption bands at 2207 and 2104
the absorption of 4-benzylisothiazole (1c) decreased from
-
1
0
.962 to 0.550 while the optical density at 294 nm
increased from 0 to 0.930 due to the formation of 9c and
0c. After acidification with concentrated HCl, the
cm in a ratio of 1.77:1 showing the presence of cyano-
sulfide 9c and isocyanosulfide 10c. GLC analysis of the
irradiated alkaline solution, however, showed 88% con-
sumption of the reactant and formation of 2c in 43%
yield, while infrared analysis of the resulting residue
showed the sharp bands due to the 9c and 10c in a ratio
of 1.23:1. In contrast, after neutralization of the irradi-
ated methanol/HCl solution, GLC analysis showed 57%
consumption of 1c but did not reveal the formation of
any volatile photoproduct while infrared analysis of the
residue from this solution showed an absorption band at
1
absorption maximum shifted from 294 to 251 nm. When
this solution was again made basic with ammonia, the
absorption maximum shifted back to 294 nm but with a
decrease in the optical density from 0.930 to 0.626.
Further acidification followed by basification was also
accompanied by changes in the absorption maximum
from 294 to 251 nm and back to 294 nm but without any
additional significant changes in the optical density. This
suggests that a portion of the mixture absorbing at 294
-1
2207 cm , indicating the presence of 9c, but did not show
-1
nm is unstable in acidic solution. Isocyanides are known
absorption at 2104 cm , showing that 10c is not present
in this mixture.
to be quite sensitive to acid.2
7,28
Thus, it seems likely
that although cyanosulfide 9c and isocyanosulfide 10c
are both formed photochemically in methanol/ammonia,
the isocyanide 10c is decomposed upon acidification and
only the cyanothiol 9cH, which is relatively insensitive
to acid, remains in the solution. Further experiments
support this suggestion.
These results demonstrate (Scheme 4) that 4-benzyl-
isothiazole (1c) undergoes photocleavage in methanol
containing HCl to 2-cyano-3-phenylpropene-1-thiol (9cH)
absorbing at 251 nm which can be deprotonated by
addition of ammonia to yield the conjugate base 9c
absorbing at 294 nm. UV and infrared spectroscopy
indicate that 2-isocyano-3-phenylpropene-1-thiol (10cH)
is not formed in this reaction. Furthermore, no 4-ben-
zylthiazole (2c) was observed when 1c was irradiated in
methanol containing HCl or after the solution was made
alkaline by the addition of aqueous ammonia. These
results show that 2-cyano-3-phenylpropene-1-thiol (9cH)
is not an intermediate in the conversion of 1c to 2c.
Alternatively, photolysis of 4-benzylisothiazole (1c) in
methanol containing ammonia leads to the formation of
cyanosulfide 9c and isocyanosulfide 10c as primary
products. Reaction of cyanosulfide 9c with benzyl bro-
mide led to the formation of (E)- and (Z)-2-cyano-3-
phenylpropen-1-ylbenzyl thioether 7c, whereas reaction
of isocyanosulfide 10c with benzyl bromide followed by
hydrolysis led to the formation of (E)-2-(N-formylamino)-
3-phenylpropen-1-ylbenzyl thioether (8c). Isolation of
only the E isomer suggests that (Z)-2-isocyano-3-phenyl-
propene-1-sulfide (10c) undergoes thermal cyclization to
4-benzylthiazole (2c).
A solution of 4-benzylisothiazole (1c) (3.0 mL, 1.6 ×
-
4
1
0
M) in methanol containing concentrated HCl (1.0
µL) was irradiated for 10 s, during which the optical
density at 251 nm increased slightly from 0.948 to 0.960
but no increase in the optical density at 294 nm was
observed. After this solution was made basic with
ammonia, however, the absorption maximum shifted
from 251 to 294 nm with an optical density of 0.867. After
this solution was reacidified, the absorption maximum
shifted back to 251 nm with only a very slight decrease
in the optical density due to dilution of the sample. This
suggests that only the cyanothiol 9cH is formed photo-
chemically in methanol/HCl and is stable with respect
to subsequent changes in acidity.
To confirm this, three solutions of 4-benzylisothiazole
(1c) in methanol, methanol containing ammonia, or
methanol containing HCl were irradiated simultaneously
on a merry-go-round apparatus. GLC analysis of the
resulting neutral methanol solution showed 70% con-
sumption of the reactant and formation of 4-benzylthi-
azole (2c) in 18% yield. In addition, the infrared spec-
The photochemistry of 4-methylisothiazole (1e) was
also investigated and found to be similar to the photo-

The P
4
Permutation Pathway
J . Org. Chem., Vol. 63, No. 16, 1998 5599
chemistry of 4-benzylisothiazole (1c). Thus, GLC analy-
sis showed that the yield of 4-methylthiazole (2e) is
increased from 10% in the absence of ammonia to 32%
when the irradiation is carried out in methanol contain-
ing ammonia. Also, UV absorption analysis again re-
vealed that thiazole 2e is not the primary product in this
reaction but that in methanol containing ammonia 1e
undergoes photocleavage to yield a mixture of cyanosul-
fide 9e and isocyanosulfide 10e, absorbing in the UV at
Ta ble 1. Qu a n tu m Yield Resu lts
reactant (solvent) φ(consumption) P4 product φ(formation)
1
1
a (C6H12)
0.41 ( 0.06
0.40 ( 0.06
0.21 ( 0.03
0.25 ( 0.04
2a
2a
2c
2c
0
a (C6H12/TEA)
0.25 ( 0.04
0.08 ( 0.01
0.13 ( 0.02
1c (CH OH)
3
1c (CH3OH/TEA)
Ta ble 2. Sp ectr oscop ic a n d P h otop h ysica l P r op er ties
λmax (nm)
log ꢀ
φf (×10 ) τf0 (ns) τf (ps)
4
τp (s)
4.20, 4.16
(3.84)
4.00
1.32
1.9
0.25 3.8 × 10-
0.29 5.2 × 10-
1
3
2
90 nm, or in methanol containing HCl to yield only
1a 241, 270
cyanothiol 9eH which absorbs at 250 nm (Scheme 4).
In an attempt to trap 9e and 10e, a solution of
(288 sh)
1
c
251
∼0.4
7.2
4
-methylisothiazole (1e) in methanol containing TEA was
Sp ectr oscop ic P r op er ties. 4-Phenylisothiazole (1a )
irradiated for 30 min. Benzyl bromide was added, and
after 6 h in the dark, the absorption maximum had
shifted from 290 to 272 nm, confirming that the reaction
of 9e and 10e with benzyl bromide was complete. The
infrared spectrum of the resulting residue showed intense
absorption bands at 2209 and 2106 cm confirming the
presence of the benzyl derivatives of 9e and 10e. After
treatment of this residue with acetic acid, infrared
analysis showed that the absorption due to the isocyanide
at 2106 cm was replaced by a new absorption band at
1
derivative of 10e to the formylbenzyl thioether 8e. The
absorption at 2209 cm remained unchanged during
treatment with acetic acid, indicating that 2-cyanopro-
pen-1-ylbenzyl thioether 7e was unchanged by the reac-
tion conditions.
and 4-benzylisothiazole (1c) exhibit structureless absorp-
tion spectra in methanol solvent (Table 2) with extinction
coefficients consistent with π f π* transitions. Both
compounds exhibit very weak fluorescence and phospho-
rescence in methanol/ethanol (1:1) solvent. Discernible
-
1
0
,0 bands in the fluorescence and phosphorescence
spectra of 1a and 1c indicate that ES1 values are 90.8
-
1
T
and 91.1 kcal mol while E 1 values are 65.7 and 71.5
-1
kcal mol , respectively. The quantum yields of fluores-
cence at room temperature were measured relative to
-
1
-
1
684 cm , consistent with the hydrolysis of the benzyl
hexaphenylbenzene35 and are shown in Table 2. Intrinsic
0
fluorescence lifetimes (τ
f
) were determined from Strick-
-
1
ler-Berg calculations in which areas under the absorp-
tion and emission curves are evaluated.3
cence lifetimes (τ ) were then estimated from τ
The phosphorescence lifetimes (τ ) were determined from
6,37
The fluores-
0
f
f
f f
) τ φ .
p
2
-Cyanopropen-1-ylbenzyl thioether 7e was isolated by
plots of the natural logarithm of the phosphorescense
decay versus time. These lifetimes and energy gaps are
consistent with π and π* states.
preparative layer chromatography as a colorless oil which
exhibited an intense absorption in the infrared spectrum
at 2209 cm as required by the nitrile functional group.
The mass spectrum, which exhibited a molecular ion at
m/z 189, and the elemental analysis are consistent with
-
1
Sen sitized Ir r a d ia tion s. The photoreaction of 4-phe-
-
1
nylisothiazole (1a ) (E
T
) 65.7 kcal mol ) could not be
) 74.7
sensitized in acetonitrile by butyrophenone (E
kcal mol ) despite the observation that 1a efficiently
T
the molecular formula of C11H11NS required by 7e. The
-
1
1
H NMR spectrum indicated that the isolated product
was a mixture of two geometrical isomers. Thus, in
quenched the Norrish type II reaction of the sensitizer
9
-1
-1
with a rate constant of 3.2 × 10
M
s . 4-Benzyl-
1
addition to the two 2-H singlets in the H NMR spectrum
-1
isothiazole (1c) (E ) 71.5 kcal mol ) was also unreactive
T
at δ 3.97 and 3.98 due to the thiobenzyl protons of the
two isomers, the spectrum also exhibits two quartets at
δ 6.64 (J ) 1.42 Hz) and 6.94 (J ) 1.34 Hz) in an
integrated ratio of 2:1 assigned to the vinyl protons of
when irradiated in the presence of butyrophenone. In
this case, however, the type II reaction of butyrophenone
was not quenched, indicating that energy transfer had
not occurred.
(Z)- and (E)-7e which are spin coupled to the allylic
Mech a n istic Discu ssion . The experimental results
reported here show that the phototransposition reaction
is greatly enhanced by the presence of small quantities
of bases such as triethylamine in the reaction medium.
Amines are known to be good electron donors in electron
methyl protons which appear as two doublets (ratio of
2
which absorb at δ 20.35 and 16.45 in the C NMR
spectrum. Similarly, the carbons of the methylene and
cyano groups in the Z and E isomers were observed at δ
1
:1) in the H NMR spectrum at δ 1.88 and 1.83 and
1
3
3
8
transfer reactions to excited states of electron acceptors.
3
7.81 and 117.63 and at δ 37.98 and 119.75, respectively.
Considering the reduction potential of TEA (1.15 eV vs
All of these spectroscopic data confirm that the isolated
product is a 2:1 mixture of (Z)- and (E)-2-cyanopropen-
1
SCE in acetonitrile), and the ∆E0,0 for 4-phenylisothiazole
-
1
(
97.9 kcal mol ), the reduction potential of 4-phenyl-
-ylbenzyl thioether (7e) formed by the reaction of (Z)-
isothiazole (1a ) can be no more negative than 3.08 eV
for the electron transfer to be thermodynamically fea-
sible. Although 4-phenylisothiazole (1a ) was not ob-
served to undergo reduction in the range of 0 to -2.2 eV
versus SCE, the reduction could not be evaluated at more
and (E)-9e with benzyl bromide.
Although our results show that isocyanosulfide 10e
was also formed and reacted with benzyl bromide and
then with acetic acid to form 8e, the quantity of 8e was
insufficient to allow isolation by preparative layer chro-
matography.
Qu a n tu m Yield s. The quantum yields for the con-
sumption of isothiazoles 1a and 1c and for the formation
4
of the P phototransposition products 2a and 2c were
determined in triplicate in cyclohexane or methanol
solutions in the absence and presence of TEA and are
shown in Table 1. 1,3-Cycloheptadiene was used as the
actinometer.34
(34) Numao, N.; Hamada, T.; Yonemitsu, O. Tetrahedron Lett. 1977,
1661-1664.
(35) Berlman, I. B. Handbook of Fluorescence Spectra of Aromatic
Molecules, 2nd ed.; Academic Press: New York, 1971; p 237.
(
36) Strickler, S. J .; Berg, R. A. J . Chem. Phys. 1962, 37, 814-822.
(37) We are thankful to Dr. R. E. Connors of this department for
carrying out these calculations.
(38) Pienta, N. J . In Photoinduced Electron Transfer. Part C.
Photoinduced Electron-Transfer Reactions: Organic Substrates; Fox,
M. A., Chanon, M., Eds.; Elsevier: New York, 1988; pp 421-486.

5
600 J . Org. Chem., Vol. 63, No. 16, 1998
Pavlik et al.
negative potentials since the acetonitrile solvent is
reduced at this potential. Therefore, we cannot exclude
the possibility that it is thermodynamically feasible for
an electron to transfer from TEA to the excited singlet
state of 4-phenylisothiazole (1a ). Kinetically, however,
this transfer does not seem likely. Thus, the calculated
lifetime for the excited singlet of 4-phenylisothiazole (1a )
Sch em e 5
-
13
is 2.5 × 10
10 . If we assume that the diffusion-controlled rate
constant is 1.0 × 10 L mol
s and the rate of unimolecular decay is 4.0
1
2
-1
×
s
1
0
-1 -1
s
, the rate of unimolecu-
-
2
lar decay in the presence of 1.0 × 10
M TEA is
calculated to be 4 orders of magnitude faster than the
rate of bimolecular diffusion-controlled electron transfer.
Clearly, under these conditions, electron transfer would
not be expected to compete with unimolecular decay of
the singlet state.
pected to be in equilibrium with thioformylazirine 14. In
the absence of other reaction pathways, â-thioformylvinyl
nitrene 11c is thus converted totally to cyanothiol 13H.
As a result, upon irradiation in benzene, ether, or
methanol containing HCl, 4-substituted isothiazoles such
as 11c undergo only photocleavage to cyanothiols such
as 13H.
If the photolysis is carried out in the presence of
ammonia or TEA, the base will convert cyanothiol 13H
to cyanosulfide 13. More importantly, however, it is
envisioned that the base will also deprotonate azirine 14,
resulting in its conversion to isocyanide 15. This sug-
gestion is not without precedent. Thus, although ben-
zoylazirine 16 is stable in solution at room temperature,
Isomura and colleagues have shown that upon addition
of pyridine the azirine is converted to oxazole 18 via
In contrast, however, the lifetime of the triplet state
of 4-phenylisothiazole (1a ), measured from the phospho-
rescence decay, is 0.4 s. Thus, even at a TEA concentra-
-
2
tion of 1.0 × 10 M, electron transfer to the triplet state
of 4-phenylisothiazole (1a ) should be kinetically quite
possible. In this case, however, since the triplet energy
is 67.8 kcal mol- , the reduction potential of 4-phenyl-
isothiazole (1a ) can be no more negative than 1.79 eV
versus SCE for the electron transfer to be thermody-
namically feasible. Cyclic voltammetry, however, indi-
cates that the reduction potential is considerably more
negative than this value.
1
In addition to the above arguments, several other
experimental observations are also inconsistent with an
electron transfer mechanism. First, the effect of TEA on
the photoreaction was not sensitive to changing the
solvent polarity from benzene to methanol. Second,
although TEA, propylamine, ammonia, and bicarbonate
have substantially different ionization potentials, they
all affect the reaction to essentially the same extent.
Accordingly, a mechanism involving electron transfer to
either the excited singlet or triplet of the 4-substituted
isothiazole can be excluded.
1
isocyanide 17 which was detected by H NMR and
41
infrared spectroscopy.
It is suggested that the fate of isocyanide 15 depends
on the nature of the substituent originally at C-4 of the
isothiazole ring. If R is aryl, as in 15(RdP h ), the
extended conjugation of the sulfide and the aryl group is
expected to lower the basicity of the sulfide, leaving the
isocyanide carbon as the more basic site. Protonation
thus converts 15 to 15H. The net effect of this protona-
tion is to render the carbon more susceptible to nucleo-
philic attack by the negative sulfur. As a result, these
substituted isocyanides cannot be detected or chemically
trapped because they spontaneously cyclize to the 4-arylth-
iazole 2a .
By analogy with the photochemistry of isoxazoles and
pyrazoles, photoexcitation of 4-substituted isothiazoles
1
1 is suggested to result in cleavage of the S-N bond,
resulting in the formation of a species that can be viewed
as diradical 11a or zwitterion 11b, which are resonance
forms of â-thioformylvinyl nitrene 11c.
In addition to recyclizing to the isothiazole reactant,
â-thioformylvinyl nitrene 11c would be expected to
rearrange as shown in Scheme 5 to the cyanothiol
photocleavage product 13H, plausibly by way of thio-
formylketeneimine 12. Indeed, isomerization to nitriles
is a well-documented reaction of terminal vinyl ni-
trenes.39 In addition, since vinyl nitrenes are also known
to be in thermal equilibrium with their isomeric azir-
ines,40 â-thioformylvinyl nitrene 11c would also be ex-
(39) See Hassner, A. In Azides and Nitrenes. Reactivity and Utility;
Scriven, E. F. V., Ed.; Academic Press: Orlando, 1984; pp 35-94, and
references therein.
(40) Hassner, A.; Wiegand, N. H.; Gottlieb, H. E. J . Org. Chem. 1986,
(41) Isomura, K.; Hirose, Y.; Shuyama, H.; Abe, S.; Ayabe, G.;
5
1, 3176-3180.
Taniguchi, H. Heterocycles 1978, 9, 1207-1216.

The P
4
Permutation Pathway
J . Org. Chem., Vol. 63, No. 16, 1998 5601
If the C-4 substituent is alkyl or a substituted alkyl
as in 10c,e, the reduced conjugation raises the energy of
the sulfide so that sulfur is more basic than the isocya-
nide carbon. Protonation of 10c,e at sulfur reduces the
nucleophilic character of the sulfide and also leaves the
negatively charged carbon less susceptible to nucleophilic
attack. As a result, cyclization requires a higher energy
of activation, and hence, the alkyl-substituted isocyanides
can be detected spectroscopically and trapped by benzyl
bromide.
the cell purged with argon for 15 min and closed with a Teflon
cap. The cell was irradiated with one low-pressure Hg lamp
for six consecutive 5 s irradiations and analyzed by UV
absorption spectroscopy after every 5 s.
GLC An a lysis. A solution of 1a , 1c, or 1e (3.0 mL, 2.0 ×
-
2
1
0
M) in benzene, methanol, or diethyl ether solvent in the
absence or presence of the appropriate quantity of aqueous
ammonia, triethylamine, or concentrated HCl was placed in
a quartz tube (0.70 cm inside diameter × 15 cm long) and the
tube sealed with a rubber septum and purged with argon for
3
0 min. The tubes were then irradiated at 254 nm in a
Rayonet photochemical reactor equipped with eight low-
pressure Hg lamps.
Solutions without added aqueous ammonia or HCl were
analyzed directly by GLC. Solutions containing aqueous
The effects of acid and base added to the photochem-
istry of 3- and 5-substituted isothiazoles are currently
being investigated and will be reported at a later time.
ammonia were dried (Na
containing concentrated HCl were first neutralized (NaHCO
and then dried (Na SO ) prior to GLC analysis.
Quantitative GLC analysis of reactant consumption and
product formation was accomplished using calibration curves
constructed for 4-substituted isothiazoles 1a , 1c, and 1e and
for 4-substituted thiazoles 2a , 2c, and 2e by plotting detector
responses versus 10 standards with known concentrations.
Correlation coefficients ranged from 0.993 to 0.999.
2 4
SO ) prior to analysis, while solutions
Exp er im en ta l Section
3
)
2
4
Gen er a l P r oced u r es. 1H and 13C NMR spectra were
recorded at 200 and 50.3 MHz, respectively. GLC was
performed using a 30 m × 0.25 µm Supelcowax 10 bonded
phase column. Preparative layer chromatography was carried
out on 20 cm × 20 cm glass plates coated with 2 mm Kieselgel
6
0 F254 (Merck). Elemental analyses were determined by
Desert Analytics (Tucson, AZ).
Ir r a d ia tion of 4-P h en ylisoth ia zole (1a ) in Ben zen e. A
Ma ter ia ls. EM hydrochloric acid GH (36.5-38.0%) and EM
ammonium hydroxide GH (28-30%) were used. Benzene was
purified by refluxing over calcium hydride followed by frac-
tional distillation. Diethyl either was purified by refluxing
over a 40% dispersion of sodium in paraffin containing
benzophenone until the benzophenone ketyl radical was bright
blue followed by fractional distillation. Methanol was purified
by refluxing over magnesium methoxide followed by fractional
distillation.
-2
solution of 1a (3.0 mL, 2.0 × 10 M) in benzene was irradiated
for 30 min. GLC analysis showed the consumption of 1a (82%)
and the formation of 4-phenylthiazole (2a ) (3%) with a relative
retention of 1.07.43
Ir r a d ia t ion of 4-P h en ylisot h ia zole (1a ) in Diet h yl
-3
Eth er . A solution of 1a (3.0 mL, 6.7 × 10 M) in diethyl ether
was irradiated for 30 min. GLC analysis showed the con-
sumption of 1a (89%) but did not show the formation of any
volatile product. Analysis by UV absorption spectroscopy after
dilution (1:200) showed a decrease in the optical density at
269 nm from 0.891 to 0.726 and an increase in the optical
density at 290 nm from 0.242 to 0.779.
Syn th esis of Rea cta n ts a n d P r od u cts. 4-Phenylisothia-
zole (1a ) and 4-phenylthiazole (2a ) were prepared by proce-
4
42
dures previously published. 4-Methylisothiazole (1e) was
4
2
prepared from 3-chloro-2-methylpropenal and ammonium
thiocyanate and purified by fractional distillation. 4-Benzyl-
isothiazole (1c) and 4-methylthiazole (1e) were obtained from
Aldrich Chemical Co. and purified by distillation.
Ir r a d ia tion of 4-P h en ylisoth ia zole (1a ) in Meth a n ol.
-
2
A solution of 1a (3.0 mL, 2.0 × 10 M) in methanol was
irradiated for 30 min. GLC analysis showed the consumption
of 1a (74%) and the formation of 4-phenylthiazole (2a ) (38%)
with a relative retention of 1.07. Analysis by UV absorption
spectroscopy after dilution (1:1000) showed an increase in the
optical density at 340 nm from 0 to 0.287.
5
-Deu ter io-4-ph en ylisoth iazole (1b). Sodium metal (0.20
g, 8.8 mmol) was added to CH OD (30 mL). After the reaction
3
was complete, 4-phenylisothiazole (1a ) (0.98 g, 6.1 mmol) was
added and the flask was tightly closed and allowed to stand
at room temperature in the dark for 3 days. The resulting
solution was added to aqueous HCl (5 M, 60 mL), and the
mixture was extracted with chloroform (3 × 100 mL). The
Ir r a d ia tion of 4-P h en ylisoth ia zole (1a ) in Meth a n ol.
Effect of Ad d ed TEA. Three solutions of 1a (3.0 mL, 2.0 ×
-
2
-2
-1
1
0
M) in methanol containing 0, 1.0 × 10 , or 2.4 × 10
M
TEA were simultaneously irradiated on a merry-go-round
apparatus for 30 min. GLC analysis showed the consumption
of 1a (82, 75, or 76%, respectively) and the formation of
4-phenylthiazole (2a ) (37, 86, or 82%, respectively) with a
relative retention of 1.07. GLC analysis at 50 °C showed no
consumption of TEA.
2 4
extract was dried (Na SO ) and concentrated to yield a white
solid that was sublimed (40 °C, 1 Torr) to give 5-deuterio-4-
phenylisothiazole (1b) as white crystals: mp 35-36 °C; yield
1
0
7
1
.90 g (5.57 mmol, 91%); H NMR (CDCl
3
) δ 8.75 (s, 1H), 7.61-
.55 (m, 2H), 7.47-7.24 (m, 3H); MS m/z (relative intensity)
64 (5), 163 (11), 162 (100), 161 (27), 135 (44).
Ir r a d ia t ion of 4-P h en ylisot h ia zole (1a ) in Ben zen e.
Effect of Ad d ed TEA. Three solutions of 1a (3.0 mL, 2.0 ×
5
-Deu ter io-4-ben zylisoth ia zole (1d ). 4-Benzylisothiazole
1c) (0.54 g, 3.1 mmol) was allowed to react with a solution
prepared by dissolving sodium metal (0.1 g, 4.4 mmol) in CH
OD (25 mL) as described above to yield a yellow oil that was
distilled (Kugelrohr) to give 5-deuterio-4-benzylisothiazole (1d )
(
-
2
-2
-1
1
0
M) in benzene containing 0, 1.0 × 10 , or 2.4 × 10
M
3
-
TEA were simultaneously irradiated on a merry-go-round
apparatus for 30 min. GLC analysis showed the consumption
of 1a (63, 77, or 65%, respectively) and the formation of
4-phenylthiazole (2a ) (3, 72, or 68%, respectively) with a
relative retention of 1.07. GLC analysis at 50 °C showed no
consumption of TEA.
as a colorless oil: bp (Kugelrohr oven temperature) 120 °C (0.5
1
Torr); yield 0.50 g (2.8 mmol, 92%); H NMR (CDCl
3
) δ 8.31
(
s, 1H), 7.32-7.13 (m, 5H), 4.02 (s, 2H); MS m/z (relative
intensity) 177 (15), 176 (100), 175 (75), 149 (14), 148 (39), 117
(12), 116 (36).
Ir r a d ia tion of 4-Ben zylisoth ia zole (1c) in Meth a n ol.
Effects of Ad d ed Am m on ia or HCl. Three solutions of 1c
Ir r a d ia tion a n d An a lysis P r oced u r es. Photoreactions
-
2
of 4-substituted isothiazoles 1a , 1c, and 1e on an analytical
scale were monitored by GLC and by UV absorption spectros-
copy.
UV Absor p tion An a lysis. A solution of 1a , 1c, or 1e (3.0
mL) in methanol in the presence or absence of either 1.0 µL
of aqueous ammonia (density of 0.88 g/mL) or concentrated
HCl was placed in a quartz UV cell (1.0 cm path length) and
(3.0 mL, 2.0 × 10 M) in methanol, in methanol containing
aqueous ammonia (10 µL, density of 0.88 g/mL), or in concen-
trated HCl (10 µL) were simultaneously irradiated on a merry-
go-round apparatus for 30 min. GLC analysis showed the
consumption of 1c (70, 88, or 57%, respectively) and the
formation of 4-benzylthiazole (2c) (18, 43, or 0%, respectively)
with a relative retention of 0.86. Each of these solutions was
(
42) Arnold, Z.; Zemlicka J . Collect. Czech. Chem. Commun. 1959,
(43) The retentions of products are given relative to the appropriate
reactant.
2
4, 2385-2392.

5
602 J . Org. Chem., Vol. 63, No. 16, 1998
Pavlik et al.
evaporated to dryness, and the infrared spectrum (KBr) of each
residue was recorded. The spectra showed bands at 2207 and
to dryness. 1H NMR (CDCl
a singlet at δ 8.75 due to the C-3 proton of residual 1b and a
singlet at δ 8.86 due to the C-2 proton of 2b formed in the
reaction.
3
) analysis of the residue showed
-
1
-1
2
104 cm (1.77:1), 2207 and 2104 cm (1.23:1), and 2207
-
1
cm , respectively.
Ir r a d ia tion of 4-Ben zylisoth ia zole (1c) in Meth a n ol
Con ta in in g Am m on ia . 4-Benzylisothiazole (1c) (0.18 g, 1.03
mmol) was dissolved in methanol (50 mL) containing aqueous
ammonia (0.25 mL, density of 0.88 g/mL) and irradiated. The
resulting solution was evaporated to dryness, and the residue
Ir r adiation of 4-Meth ylisoth iazole (1e). Effect of Added
-
2
TEA. Two solutions of 1e (3.0 mL, 2.0 × 10 M) in methanol
or in methanol containing TEA (10 µL, 0.072 mmol) were
simultaneously irradiated on a merry-go-round apparatus for
0 min. GLC analysis showed the consumption of 1e (68 or
2%, respectively) and the formation of 4-methylthiazole (2e)
3
8
(
(
(
)
0
4
0.18 g) was subjected to preparative layer chromatography
silica gel, CH Cl ) which gave two major bands. Band 1 (R
0.50) contained recovered 4-benzylisothiazole (1c) (0.036 g,
.21 mmol, 20.4% recovery). Band 2 (R ) 0.34) contained
-benzylthiazole (2c) (0.073 g, 0.42 mmol, 51% yield): H NMR
) δ 8.75 (d, J ) 2.02 Hz, 1H), 7.35-7.18 (m, 5H), 6.86
triplet of doublet, J ) 1.99 Hz, J ) 2.02 Hz), 4.17 (br s, 2H);
2
2
f
10 or 32%, respectively) with a relative retention of 0.80.
P r ep a r a tive Sca le Ir r a d ia tion s. A solution of the reac-
tant was dissolved in the appropriate solvent (10.0, 50.0, or
0.0 mL) and the mixture placed in a quartz tube [1.45 cm
inside diameter × 13 cm long (for 10.0 mL) or 2.5 cm inner
diameter × 30 cm long (for 50.0 or 90.0 mL)] which was closed
with a rubber septum, purged with argon for 30 min, and
irradiated in a Rayonet reactor equipped with eight low-
pressure lamps for 30 min while the solution was continuously
purged with a fine stream of argon.
f
1
9
(
CDCl
3
(
t
d
13
C NMR (CDCl ) δ 157.20, 152.77, 139.01, 128.96, 128.57,
3
1
1
3
26.47, 114.07, 37.69; MS m/z (relative intensity) 176 (14.6),
75 (100), 174 (75.9), 148 (32.3), 147 (34.6), 115 (69.1); IR (KBr)
-1
107, 3063, 3028, 2909, 2839, 1604 cm
. Anal. Calcd for
C
10
H
9
NS: C, 68.53; H, 5.18; N, 7.96. Found: C, 68.31; H, 5.08;
Ir r a d ia t ion of 4-P h en ylisot h ia zole (1a ) in Diet h yl
Eth er . 4-Phenylisothiazole (1a ) (0.097 g, 0.60 mmol) was
dissolved in purified diethyl ether (90.0 mL) and irradiated.
The solutions from two such reactions were combined and
evaporated to dryness to yield a yellow viscous liquid (0.19 g)
which solidified after standing at room temperature. The
residue was repeatedly recrystallized from diethyl ether and
methylene chloride to give 2-cyano-2-phenylethenyl thioether
N, 7.99.
Ir r a d ia tion of 5-Deu ter io-4-ben zylisoth ia zole (1d ) in
Meth an ol Con tain in g Am m on ia. 5-Deuterio-4-benzylisothi-
azole (1d ) (0.12 g 0.68 mmol) was dissolved in methanol (50
mL) containing aqueous ammonia (0.25 mL, density of 0.88
g/mL) and irradiated. The resulting solution was evaporated
to dryness, and the residue (0.10 g) was subjected to prepara-
tive layer chromatography (silica gel, CH Cl ) which gave two
2
2
(
4) as colorless crystals: mp 192-193 °C; yield 0.012 g (0.042
major bands. Band 1 (R ) 0.50) contained recovered 5-deu-
f
1
mmol, 3.5%); H NMR (CD
1
1
cm ; HRMS calcd for C18
Anal. Calcd for C18
2
Cl
2
) δ 7.66 (s, 2H), 7.41-7.58 (m,
terio-4-benzylisothiazole (1d ) (0.024 g, 0.14 mmol, 19.5%
recovery). Band 2 (Rf ) 0.34) contained 5-deuterio-4-ben-
1
3
0H); C NMR (CD Cl ) δ 137.77, 132.62, 129.97, 129.63,
25.69, 115.32, 114.37; IR (KBr) 2208 (CtN), 680.9 (C-S)
2 2
1
zylthiazole (2d ) (0.044 g, 0.25 mmol, 46% yield): H NMR
-1
H
N
12
N
2
S 288.07226, found 288.07234.
3
(CDCl ) δ 8.75 (s, 1H), 7.35-7.13 (m, 5H), 4.17 (s, 1H); MS
m/z (relative intensity) 177 (14.7), 176 (100), 175 (76.3), 149
(30.4), 148 (33.7).
H
12
2
S: C, 74.97; H, 4.19; N, 9.71.
Found: C, 74.35; H, 4.13; N, 9.60.
Ir r a d ia tion of 4-Ben zylisoth ia zole (1c) in Meth a n ol
Con ta in in g Am m on ia . Tr a p p in g of Cya n osu lfid e 9c a n d
Isocya n osu lfid e 10c. 4-Benzylisothiazole (1c) (0.12 g, 0.69
mmol) was dissolved in methanol (50 mL) containing aqueous
ammonia (0.25 mL, density of 0.88 g/mL) and irradiated.
Benzyl bromide (84 µL, 0.2 g, 0.70 mmol) was added to the
resulting solution, and the mixture was allowed to stand for
Ir r a d ia t ion of 4-P h en ylisot h ia zole (1a ) in Diet h yl
Eth er . Tr a p p in g of Cya n osu lfid e 5a . 4-Phenylisothiazole
1a ) (0.050 g, 0.21 mmol) was dissolved in purified diethyl
ether (50.0 mL) and irradiated. To the resulting solution was
added TEA (5.0 mL) followed by benzyl bromide (0.100 mL,
(
0
.84 mmol). The mixture was allowed to stand overnight at
room temperature. The solutions from two such reactions were
combined and evaporated to dryness, and the residue (0.096
g) was subjected to preparative layer chromatography (silica
3
0 min and then evaporated to dryness. Glacial acetic acid
(1.0 mL) was added to the residue, and the mixture was
allowed to stand at room temperature overnight. It was
subsequently neutralized with saturated aqueous sodium
gel, CH
graphed (silica gel, 1:1 CH
.50 gave 2-cyano-2-phenylethenylbenzyl thioether (6a ) as a
white solid, with a melting point of 58-60 °C and a yield of
.036 g (0.14 mmol, 22.6%) which was further purified by
sublimation (60 °C, 0.2 Torr) to give white crystals: mp 58-
2
Cl
2
). The band at R
f
) 0.76 (0.050 g) was rechromato-
2
Cl
2
/hexane), and the band at R
f
)
bicarbonate and extracted with CH
SO
subjected to preparative layer chromatography (silica gel, CH
Cl ). The band at R
phenylpropen-1-ylbenzyl thioether (7c) as a yellow oil (0.04 g,
2
Cl
2
(3 × 20 mL), dried (Na
2
-
0
4
), and concentrated. The yellow residual oil (0.19 g) was
2
-
0
2
f
) 0.79 gave (E)- or (Z)-2-cyano-3-
1
6
0 °C; H NMR (CDCl
3
) δ 4.14 (s, 2H), 7.24-7.54 (m, 10H);
) δ 143.75, 136.05, 133.11, 129.03, 128.96,
28.91, 128.37, 128.02, 124.79, 116.16, 109.75, 38.33; IR (KBr)
1
1
3
0.15 mmol, 21.4% yield): H NMR (CDCl
1
3
) δ 7.36-7.07 (m,
C NMR (CDCl
3
0H), 7.05 (t, J ) 0.90 Hz, vinyl proton of E isomer), 6.66 (t,
1
2
(
-
1
J ) 1.34 Hz, vinyl proton of Z isomer), 4.02 (s, E isomer), 3.96
207 (CtN) cm ; MS m/z (relative intensity) 251 (16.7), 91
13NS: C, 76.46; H, 5.21; N, 5.57.
1
3
(
s, Z isomer), 3.51 (br s, E isomer), 3.46 (br s, Z isomer);
C
100). Anal. Calcd for C16
H
NMR (CDCl ) δ 145.56, 144.45, 136.21, 136.05, 135.95, 135.90,
3
Found: C, 76.52; H, 5.07; N, 5.52.
1
1
28.83, 128.69, 128.58, 128.40, 123.38, 128.37, 127.82, 127.58,
27.02, 126.96, 118.87 (CtN, E isomer), 117.02 (CtN, Z
Ir r a d ia tion of 5-Deu ter io-4-p h en ylisoth ia zole (1b) in
Met h a n ol Con t a in in g Am m on ia . 5-Deuterio-4-phenyl-
isothiazole (1b) (0.44 g, 0.27 mmol) was dissolved in methanol
isomer), 108.73, 108.37, 40.16 (Z isomer), 38.07 (E isomer),
3
2
6.95 (Z isomer), 36.49 (E isomer); MS m/z (relative intensity)
(
0
10.0 mL) containing aqueous ammonia (0.10 mL, density of
.88 g/mL) and irradiated. The resulting solution was evapo-
rated to dryness, and the residue (0.040 g) was subjected to
preparative layer chromatography (silica gel, CH Cl ) which
gave two bands. Band 1 (R ) 0.2) contained 5-deuterio-4-
phenylthiazole (2b) (0.012 g, 0.076 mmol, 48% yield): H NMR
CDCl ) δ 8.86 (s, 1H), 7.60-7.55 (m, 2H), 7.47-7.24 (m, 3H);
MS m/z (relative intensity) 164 (5), 163 (12), 162 (100), 161
-1
65 (4.8), 91 (100); IR (KBr) 2200 (CtN) cm ; HRMS (FAB)
+
calcd for C17
The band at R
phenylpropen-1-ylbenzyl thioether (8c) as a yellow oil (0.02 g,
H
16NS (M ) 266.1003, found 266.10181.
f
) 0.16 gave (E)-2-(N-formylamino)-3-
2
2
f
.07 mmol, 10.0% yield): 1H NMR (DMSO-d
0
1
)
(
6
) δ 9.57 (d, J )
0.72 Hz, exchangeable), 9.38 (br s, exchangeable), 8.18 (d, J
10.72 Hz), 8.01 (d, J ) 1.72 Hz), 7.32-6.97 (m, 10H), 6.89
s), 5.89 (s), 3.90 (s, 0.75H), 3.88 (s, 1.25H), 3.59 (s, 0.75H),
1
(
3
(
26), 135 (37).
13
3
3
.52 (s, 1.25 H); C NMR (CDCl ) δ 161.34, 158.94, 138.13,
Ir r a d ia tion of 5-Deu ter io-4-p h en ylisoth ia zole (1b) in
Meth a n ol. A solution of 1b (0.044 g, 0.27 mmol) in methanol
10.0 mL) was irradiated for 30 min. GLC-MS analysis of
b showed ions at m/z (relative intensity) 164 (5), 163 (12),
137.57, 136.38, 135.69, 133.03, 128.90, 128.68, 128.54, 127.37,
127.09, 112.78, 109.38, 39.00, 38.34, 37.29, 36.66; IR (KBr)
-
1
(
2
1
1683 cm ; MS m/z (relative intensity) 284 (3), 283 (15.8), 91
+
(100); HRMS (FAB) calcd for C17
284.1092.
H18NSO (M ) 284.1104, found
62 (100), 161 (26), and 135 (37). The solution was evaporated

The P
4
Permutation Pathway
J . Org. Chem., Vol. 63, No. 16, 1998 5603
Ir r a d ia t ion of 4-Met h ylisot h ia zole (1e) in Met h a n ol
Con ta in in g TEA. Tr a p p in g of Cya n osu lfid e 9e. A solu-
tion of 1e (0.26 g, 2.67 mmol) in methanol (90.0 mL) containing
TEA (1.0 mL) was irradiated for 30 min. Benzyl bromide (0.3
mL, 2.67 mmol) was added, and the mixture was allowed to
stand at room temperature for 2 h and then evaporated to
dryness. Glacial acetic acid (1.0 mL) was added, and the
mixture was allowed to stand overnight at room temperature.
It was then neutralized with saturated aqueous sodium
ing solutions of 4-phenylisothiazole (1a ) or 4-benzylisothiazole
-2
(1c) (2.0 × 10 M, 3.0 mL) in cyclohexane or methanol solvent
-
3
in the absence or presence of TEA (7.2 × 10 M) and the
actinometer (1,3-cycloheptadiene) in a Rayonet reactor equipped
with a merry-go-round. The consumption of the reactants or
4
actinometer and the formation of the P phototransposition
products (2a or 2c) were measured using GLC.
Sen sitiza tion s. Six solutions of butyrophenone (3.0 mL,
-
2
3
(
7
.0 × 10 M) in cyclohexane containing 4-phenylisothiazole
bicarbonate and extracted with CH
Cl extract was washed with saturated aqueous sodium
bicarbonate (3 × 5 mL), dried (Na SO ), and concentrated. The
residues from two such reactions (yellow oil, 0.56 g) were
subjected to preparative layer chromatography (silica gel, CH
Cl ). The band at R ) 0.80 gave 2-cyanopropen-1-ylbenzyl
2
Cl
2
(3 × 20 mL). The CH
2
-
-3 -3 -3 -6
1a ) (0, 1.52 × 10 , 3.03 × 10 , 4.55 × 10 , 6.06 × 10 , or
2
-
3
.58 × 10 M) were irradiated simultaneously in Pyrex tubes
2
4
in a Rayonet reactor equipped with a merry-go-round and one
50 nm lamp. The concentrations of 1a , butyrophenone, and
acetophenone were measured using GLC. A plot of φ°/φ versus
3
2
-
2
f
-
1
thioether (7e) as a yellow oil (0.14 g, 0.74 mmol, 16% yield).
The total amount of 7e from six photoreactions (0.50 g) was
subjected to preparative layer chromatography (silica gel, 3:4
[1a ] was linear with a slope of 771 M .
A similar irradiation was conducted in CH CN using an
3
initial butyrophenone concentration of 0.20 M and concentra-
CH
oil (0.25 g) which was again subjected to preparative layer
chromatography (silica gel, 1:1 CH Cl /hexane). The band at
2
Cl
2
/hexane). The band at R
f
) 0.25 gave 7e as a yellow
-3
-3
-3
tions of 1a of 11.92 × 10 , 23.85 × 10 , 35.78 × 10 , 47.70
-3
-3
×
10 , and 59.63 × 10 M. A plot of φ°/φ versus [1] was linear
2
2
-1
with a slope of 488 M . In neither case could any reaction of
R ) 0.40 gave 7e as a slightly yellow oil (0.24 g) which was
f
1
a be detected.
further purified by distillation (Kugelrohr) to give 7e as a
Lu m in escen ce Sp ectr a , En er gies, a n d Lifetim es of
colorless oil: bp (Kugelrohr oven temperature) 40 °C (0.2 Torr);
1
Excited Sta tes. The fluorescence spectrum of 1a or 1c was
recorded at room temperature in cyclohexane (A265nm ) 0.345)
or in a 1:1 methanol/ethanol solvent (A250nm ) 0.250) after
excitation at 265 or 250 nm, respectively. Quantum yields of
fluorescence were determined relative to hexaphenylbenzene
yield 0.18 g; H NMR (CDCl
)
3
) δ 7.37-7.24 (m, 5H), 6.94 (q, J
1.34 Hz, vinyl proton of E isomer), 6.64 (q, J ) 1.42 Hz,
vinyl proton of Z isomer), 3.98 (s, E isomer), 3.97 (s, Z isomer),
1
.88 (d, J ) 1.42 Hz, Z isomer), 1.83 (d, J ) 1.34 Hz, E isomer);
1
3
C NMR (CDCl
3
) δ 145.28 (E isomer), 143.56 (Z isomer),
(φ
f
) 0.01).35
1
36.43, 136.20, 129.09, 128.97, 128.85, 128.22, 127.94, 119.75
(
1
CtN, E isomer), 117.63 (CtN, Z isomer), 104.57 (Z isomer),
03.62 (E isomer), 37.98 (E isomer), 37.81 (Z isomer), 20.35
Z isomer), 16.45 (E isomer); MS m/z (relative intensity) 189
The phosphorescence spectrum of 1a or 1c was recorded at
7 K in methylcyclohexane or a 1:1 methanol/ethanol solvent
after excitation at 265 or 250 nm. Decay constants and triplet
lifetimes were obtained from plots of the natural logarithm of
the phosphorescence decay versus time.
7
(
(
-
1
2), 91 (100); IR (KBr) 2209 cm . Anal. Calcd for C11H11NS:
C, 69.80; H, 5.86; N, 7.40. Found: C, 69.99; H, 5.61; N, 7.37.
Qu a n tu m Yield s of Rea ction s. Quantum yields of reac-
tions were determined in triplicate by simultaneously irradiat-
J O980936E