Electron-transfer cycloreversion of 2,3-diaryloxetanes: Influence of the substitution and the photosensitizer on the regioselectivity

Article abstract of DOI:10.1002/ejoc.200300576

The regioselectivity in the oxidative electron-transfer cycloreversion (CR) of 2,3-diaryloxetanes depends on the substitution of the aryl groups and on the nature of the electron-transfer photosensitizer. The reaction occurs with fragmentation of the C2-C

Full text of DOI:10.1002/ejoc.200300576

FULL PAPER
Electron-Transfer Cycloreversion of 2,3-Diaryloxetanes: Influence of the
Substitution and the Photosensitizer on the Regioselectivity
M. Angeles Izquierdo[a] and Miguel A. Miranda*[a]
Keywords: Small ring systems / Cleavage reactions / Time-resolved spectroscopy / Photochemistry / Radical cations
The regioselectivity in the oxidative electron-transfer cyclo-
reversion (CR) of 2,3-diaryloxetanes depends on the substitu-
thole radical cation has been detected as transient interme-
diate by means of laser flash photolysis. Likewise, CR of 1a,
tion of the aryl groups and on the nature of the electron- photosensitized by chloranil, results in the formation of prod-
transfer photosensitizer. The reaction occurs with fragmenta- ucts arising from the trapping of the trans-β-methylstyrene
tion of the C2−C3 and C4−O bonds either in the presence of
electron-releasing substituents in the 3-aryl groups, or when
chloranil is used as photosensitizer. Thus, CR of the methoxy-
substituted derivative trans,trans-3-(4-methoxyphenyl)-4-
methyl-2-phenyloxetane (1b) results in the production of
radical cation by chloranil-derived intermediates. The re-
verse regioselectivity has been previously found in the ET
cycloreversion of 4-methyl-2,3-diphenyloxetane (1a) using
(thio)pyrylium salts as photosensitizers.
( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
trans-anethole and benzaldehyde. In this case, the trans-ane- Germany, 2004)
Introduction
the intermediacy of metallacyclobutanes (Scheme 1).[6,7]
Ring-splitting of oxetanes can be achieved by chemical
means[8] (hydrogenolysis or nucleophilic attack), thermoly-
sis[8] and by photochemical activation.[3,4] The latter in-
volves the use of ET photosensitizers and has been little ex-
plored.
CarbonϪcarbon bond breaking involving radical cations
has been studied in a number of organic compounds con-
taining strained rings, such as cyclopropanes, oxiranes,
cyclobutanes and spiroalkanes.[1] Trapping of the ring-
opened distonic radical cations with nucleophiles, electron-
rich olefins, NO or oxygen may have interesting synthetic
applications.[2] However, there are very few literature re-
ports describing the cycloreversion (CR) of oxetanes under
oxidative electron-transfer (ET) conditions.[3,4]
Oxetanes can be easily obtained by the photocycloaddi-
tion of carbonyl compounds to alkenes (PaternoϪBüchi re-
action).[5] Their CR involves cleavage of two bonds (CϪC
and CϪO) and can yield formal olefin metathesis products.
In this context, the CR of oxetanes presents clear analogies
with the Chauvin mechanism involved in the interconver-
sion of olefins and (alkylidene)metal compounds through
Theoretical calculations on the ET cycloreversion of oxe-
tane radical cations point to an initial CϪC bond cleav-
age.[9] By contrast, the reductive ET cycloreversion should
start with CϪO bond breaking, according to the free energy
changes calculated for the gas-phase reaction.[9] Control of
the regioselectivity in the CR of oxetanes by photoinduced
ET is synthetically interesting and constitutes the key step
in the enzymatic repair of the (6Ϫ4) photoproducts of the
DNA dipyrimidine sites by photolyase.[10Ϫ12]
Prior to our work,[13,14] only two reports have appeared
on the CR of oxetane radical cations.[3,4] They have pro-
posed that the cycloreversion of 2-aryl- and 2,2-diaryloxet-
Scheme 1
[a]
´
´
´
Departamento de Quımica, Instituto de Tecnologıa Quımica
´
UPV Ϫ CSIC, Universidad Politecnica de Valencia
Camino Vera s/n, Apdo. 22012, 46022, Valencia, Spain
Fax: (internat.) ϩ 34-96-387-9349
E-mail: mmiranda@qim.upv.es
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
Scheme 2
1424
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/ejoc.200300576
Eur. J. Org. Chem. 2004, 1424Ϫ1431

Electron-Transfer Cycloreversion of 2,3-Diaryloxetanes
FULL PAPER
anes using cyanoaromatic compounds as ET photosensitiz- diaryloxetanes. As a result of this effort, clear evidence for
ers proceeds with initial C2ϪC3 bond breaking.[3,4] The C2ϪC3 bond cleavage has been obtained in the oxidative
products of these reactions are the same as the reagents CR of oxetane 1b using the (thio)pyrylium salts 2a and 2b
used to make the oxetanes. Recent experiments performed (where trans-anethole radical cation is detected) and in the
in our group have found the opposite regioselectivity in the CR of 1a using chloranil as alternative photosensitizer
irradiation of trans,trans-4-methyl-2,3-diphenyloxetane (1a) (which leads to β-methylstyrene and/or its trapping prod-
photosensitized by 2,4,6-triaryl(thio)pyrylium salts. The ucts).
radical cation of 1a undergoes stepwise CR with initial
OϪC2 bond cleavage;[14] this results in the production of
trans-stilbene and acetaldehyde (Scheme 2, pathway b).
Cleavage along pathway a is not observed. This mechanism
Results and Discussion
is supported by detection of the relevant transient species
(trans-stilbene radical cation and pyranyl radical) by means
Effect of Substitution on the CR Pathway
of laser flash photolysis (LFP).
Irradiation of the oxetane 1b was carried out in the pres-
ence of catalytic amounts of the photosensitizers 2a and 2b.
In both cases, trans-anethole and benzaldehyde were de-
Attachment of electron-releasing or -withdrawing sub-
stituents to the aryl groups could be a possible strategy to
control the reaction pathway, by favouring location of the
charge and the spin at different positions of the oxetane
radical cation or the intermediates derived thereof. Thus,
methoxy substituents are known to stabilize the positive
charge of aromatic radical cations.[3,15Ϫ18] For example, in
the oxidative ring opening of 2,2-diaryloxetanes,[3] the elec-
tron appears to be removed initially from an electron-rich
aryl group. As a result, the strength of the C2ϪC3 σ bond
decreases through σϪπ interaction.
1
tected as the sole photoproducts by direct H NMR analy-
sis of the reaction mixtures (Scheme 2, pathway a). These
results contrast with the behaviour observed previously for
the oxetane 1a, where CR resulted in the production of
trans-stilbene (5a) and acetaldehyde (Scheme 2, pathway b).
The CR reaction was found to be exergonic, starting
from the triplet state of the pyrylium salts, according to the
Weller equation[24] [Equation (1)], where the redox poten-
tials (measured by cyclic voltammetry in acetonitrile) vs.
SCE, were: E(Dϩ·/D) (1b) ϭ 1.48 V,[25] E(Aϩ/A·) ϭ Ϫ0.29 V
(2a) and Ϫ0.21 V (2b).[26]
With this in mind, a methoxy substituent has been intro-
duced to the 3-phenyl group of 1a, the only oxetane known
to undergo CϪO cleavage upon oxidative ET photosensit-
ized by (thio)pyrylium salts. This could lead to a reversal in
regioselectivity, i.e. C2ϪC3 bond cleavage (Scheme 2, path-
way a). In addition to product studies, LFP of 1b could
allow for the detection of either the radical cation of 4-
∆GET ϭ 23.06 ϫ [E(Dϩ·/D) Ϫ E(A/A-·) ϩ Ecoul.] Ϫ E*A
(1)
The triplet energies for 2a and 2b are 53 and 52 kcal/mol,
methoxystilbene or anethole, providing direct mechanistic respectively,[27] and the coulombic factor (Ecoul.) can be neg-
evidence for the ring-splitting of 1bϩ·.[19]
lected in this case. According to simple calculations based
On the other hand, the choice of photosensitizer and the on this equation,[24] the ET process between the ground
experimental conditions may determine the reaction path- state of 1b and the triplet excited states of 2a and 2b should
ways for the radical cations generated.[20] Hence, the second be thermodynamically favourable (∆GET ഠ Ϫ7 and Ϫ13
objective of the present work was to check whether a kcal/mol, respectively).
change of the photosensitizer could result in a reversal of
As expected for a reaction involving the triplet excited
the regioselectivity in the CR of 1a and 1b, leading to pref- state of the (thio)pyrylium salts,[25] sensitizer 2b (with a
erential cleavage along C2ϪC3 and OϪC4 bonds higher intersystem crossing quantum yield)[28] was found to
(Scheme 2, pathway a).[19]
Chloranil is
be more efficient than 2a.
To elucidate the mechanism of the observed cyclorever-
a
well-established electron-transfer
photosensitizer[21Ϫ23] that has the correct redox and spec- sion, LFP experiments were conducted under different con-
tral characteristics to enable investigation of ET processes ditions. In a typical experiment, LFP (355 nm) of 2b in the
with different excitation methods.[22] Triplet chloranil is presence of 1b, using acetonitrile as solvent, gave rise to two
known to react very nearly at the diffusion-controlled rate intense signals with absorption maxima centred at ca. 380
with a number of organic substrates,[21] and has been used and 600 nm, as shown in Figure 1 (A). These bands corre-
to explore the reactivity of strained radical cations of small- spond to the same transient intermediate, and disappear
ring compounds.[23] In the case of chloranil a radical ion with the same rate constant (Figure 1, B). They were as-
pair is formed,[21] in contrast with (thio)pyrylium salts, cribed to the radical cation of trans-anethole on the basis
where there is no net charge separation associated with the of literature data.[29] This assignment was confirmed by
ET step.[20] Hence, the coulombic factor may be important, LFP of trans-anethole in the presence of 2b, which gave rise
making the reaction more sensitive to the influence of the to the same intermediate.
solvent.
Another band peaking at 550 nm was also clearly ob-
Thus, the aim of the present work was to study the influ- served in the LFP experiments performed with 2a as photo-
ence of the substitution and of the nature of the photosensi- sensitizer (Figure 2, A). This band can be safely assigned to
tizer on the regioselectivity of the ET cycloreversion of 2,3- the pyranyl radical resulting from one-electron reduction of
Eur. J. Org. Chem. 2004, 1424Ϫ1431
www.eurjoc.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1425

M. A. Izquierdo, M. A. Miranda
FULL PAPER
Figure 2. (A) Transient spectra obtained upon LFP (λ ϭ 355 nm)
of 2a (1.25 ϫ 10Ϫ4 ) in acetonitrile under argon, in the presence
of 1b (6.25 ϫ 10Ϫ4 ); spectra recorded 1 µs (solid squares), 3 µs
(open squares) and 6 µs (solid circles) after the laser pulse; (B)
growth and decay of the 550-nm (open squares) and 600-nm (solid
circles) bands
Figure 1. (A) Transient spectra obtained upon LFP (λ ϭ 355 nm)
of 2b (0.75 ϫ 10Ϫ4 ) in acetonitrile under argon, in the presence
of 1b (1.2 ϫ 10Ϫ3 ); spectra recorded 0.3 µs (open squares), 1 µs
(solid triangles) and 5 µs (solid circles) after the laser pulse; (B)
growth and decay of the 380-nm (open circles) and 600-nm (solid
circles) bands
2a.[30] Moreover, the 4-methoxystilbene radical cation
(λmax. ϭ 500 nm)[31] was not detected; this is consistent with
cleavage occurring along pathway a rather than pathway
b (Scheme 2).
The ET behaviour of 1b can be justified according to
Scheme 3. Ionization would be likely to occur at the meth-
oxy-substituted aromatic ring, to give 1bϩ·. Subsequent
C2ϪC3 bond cleavage would generate a distonic 1,4-radical
cation; the observed intermediate would be formed after a
second (CϪO) bond splitting.
Figure 1 (B) shows the formation and decay of the 600-
nm band, corresponding to the trans-anethole radical cat-
ion. The rate constant estimated for the initial growth is 2.5
ϫ 106 sϪ1, and is independent of the concentration of 1b.
Therefore, it must be related to the splitting of the oxetane
radical cation. Electron transfer from 1b to triplet 2b is
much faster, as evidenced by the quenching rate constant
(4.0 ϫ 109 Ϫ1sϪ1). Hence, fragmentation of the undetect-
able oxetane radical cation 1b؉·, generating the trans-ane-
thole radical cation 4b؉· and neutral benzaldehyde, must
occur in the sub-microsecond time domain. Similar results
were obtained with 2a as the photosensitizer; in this case
4b؉· and the pyranyl radical 2b؊· decayed with similar rates
(Figure 2, B).
The last step (formation of neutral 4b) involves back elec-
tron transfer (BET) from 2؊· to 4b؉·. Taking into account
the redox potentials of 4b, 2a and 2b, which are E(Dϩ·/D)
ϭ
1.33 V, E(A/A-·) ϭ Ϫ0.29 V and Ϫ0.21 V vs. SCE,[32,26]
respectively, BET is highly exergonic in both cases [∆GET
(4b؉·/2a؊·) ϭ Ϫ37.4 kcal/mol and ∆GET (4b؉·/2b؊·) ϭ
Ϫ35.5 kcal/mol].
Influence of the ET Photosensitizer on the CR
Regioselectivity
Oxetane 1a was synthesized by the PaternoϪBüchi cyclo-
addition, according to a literature procedure.[33] Irradiation
1426
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 1424Ϫ1431

Electron-Transfer Cycloreversion of 2,3-Diaryloxetanes
FULL PAPER
Table 1. Electron transfer reactions of oxetane 1a and alkene 4a
photosensitized by chloranil 3
Solvent[a]
Time
[min]
Unchanged[b]
1a or 4a
Products[b]
5a
6
7
8
1a[c]
C6H6
MeCN
MeOH
180
180
60
120
60
60
60
100
60
89
77
64
48
7
Ϫ
Ϫ
Ϫ
2
Ϫ
Ϫ
Ϫ
Ϫ
31
Ϫ
Ϫ
36
48
Ϫ
Ϫ
9
Ϫ
Ϫ
Ϫ
4
Ϫ
Ϫ
11
21
Ϫ
Ϫ
93
4a[c]
C6H6
MeCN
MeOH
Ϫ
[a]
Samples prepared in pyrex tubes; inert gas: argon, irradiation
system: Luzchem multilamp photoreactor, 8 W lamps (4 ϫ) with
emission maximum at λ ϭ 350 nm, filter: λ Ͼ 340 nm. Samples
were concentrated after irradiation and then submitted to 1H NMR
[b]
[c]
analysis, using CDCl3 as solvent.
Given in %. Concentration
of 1a (or 4a) and 3: 10Ϫ2 .
Scheme 3
(Scheme 4, pathway c). Similar nucleophilic additions have
been described previously for the irradiation of 3 in the
presence of alkenes.
of equimolar solutions of 1a and 3 (ca. 0.01 ) was per-
formed using benzene, acetonitrile and methanol as sol-
vents. The results are shown in Scheme 4 and Table 1, to-
gether with those obtained for the related alkene 4a.
Formation of 7 (observed only in acetonitrile) may be
explained by nucleophilic attack of residual water on the β-
position of the trans-β-methylstyrene radical cation (4a؉·)
and subsequent reaction of the resulting benzylic radical
with the hydroquinone radical arising from protonation of
the chloranil radical anion (Scheme 4, pathway d). These
results are consistent with LFP observations of the chlor-
anil radical anion (λmax. ϭ 450 nm) and the hydroquinone
radical (λmax. ϭ 430Ϫ440 nm).[22,23a]
When methanol was used as the solvent, irradiation of
equimolar solutions of 1a and 3 led to 8 as the major pho-
toproduct (Scheme 4, pathways a ϩ d); after prolonged ir-
radiation (2 h) traces of other photoproducts such as trans-
stilbene (5a) and a methanol adduct[35,36] were also detected
in the photomixtures. In this solvent, the spirooxetane 6
was not observed.
The formation of 8 can be explained in the same manner
as that for 7 (see above), but with methanol instead of water
acting as nucleophile. The same product 8 was obtained
from the irradiation of 4a in methanol using 3 as a photo-
sensitizer (Scheme 4, pathways b ϩ d). The different reactiv-
Scheme 4
Whereas no photoproducts were found upon the ir- ities in acetonitrile and methanol can be explained taking
radiation of 1a in benzene, formation of benzaldehyde and into account the possibility of nucleophilic attack by meth-
two new photoproducts, 6 (major) and 7 (minor) was ob- anol on the radical ion pair. This allows trapping of the
served in acetonitrile. trans-Stilbene (5a), acetaldehyde or olefin radical cation and prevents its reaction with the chlo-
stilbeneϪchloranil adducts were not detected in the ir- ranil radical anion.
radiated mixture.[22] In the case of trans-β-methylstyrene
The methoxy-substituted oxetane 1b was also irradiated
(4a), the results were similar under the same experimental in the presence of 3 in benzene and acetonitrile. In all cases,
conditions, although shorter irradiation times were re- CR resulted in the formation of benzaldehyde and trans-
quired. This was expected in view of the much higher reac- anethole 4b. trans-4-Methoxystilbene (5b), acetaldehyde or
tivity of aryl olefins. In addition, the spirooxetane 6 was chloranil adducts were not found in the irradiation mix-
found even with benzene as the solvent. Other possible re- tures. The absence of chloranil-derived products may be due
actions of alkenes,[34] such as isomerization or dimerization, to the greater delocalization of the radical cation in the
did not take place.
presence of the methoxy substituent. This is consistent with
Formation of 6 may be explained by ring-splitting of 1a؉· the fact that no product of this type was found upon ir-
(or ionization of 4a) followed by reaction of the chloranil radiation of trans-anethole and chloranil under the same
radical anion 3؊· with the alkene radical cation 4a؉· conditions.[22a]
Eur. J. Org. Chem. 2004, 1424Ϫ1431
www.eurjoc.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1427

M. A. Izquierdo, M. A. Miranda
FULL PAPER
Time-resolved studies were conducted on mixtures of 3
and 1a in acetonitrile, benzene and/or methanol. In all
cases, the decay of triplet chloranil was faster in the pres-
ence of oxetane. In a typical experiment, LFP (355 nm) of
3 in the presence of 1a, using acetonitrile as solvent, re-
sulted in the formation of two bands with absorption max-
ima at ca. 450 nm and 430 nm, corresponding to chloranil
radical anion and the hydroquinone radical, respectively
(Figure 3).[22,23a]
Figure 4. Transient spectra obtained upon LFP (λ ϭ 355 nm) of 3
(5 ϫ 10Ϫ4 ) in acetonitrile under argon, in the presence of 1b (2.9
ϫ 10Ϫ3 ), spectra recorded 0.3 µs (open squares) and 1 µs (line)
after the laser pulse
Figure 3. Transient spectra obtained upon LFP (λ ϭ 355 nm) of 3
(5 ϫ 10Ϫ4 ) in acetonitrile under argon, in the presence of 1a (1.2
ϫ 10Ϫ2 ); spectra recorded 0.4 µs (solid squares), 0.08 µs (straight
line), 5 µs (open circles) and 8 µs (solid triangles) after the laser
pulse
No band related to the oxetane or the arylalkene radical
cations (resulting from cleavage of 1a) was observed. This
may be due to the fact that the radical cations derived from
1a are very short-lived and cannot be detected under our
reaction conditions. Actually, LFP of 3 in the presence of
4a did not lead to any signal attributable to 4a؉· either. As
stated above, trans-anethole radical cation 4b؉· gives rise to
a typical transient absorption at ca. 600 nm and is produced
during the oxidative CR of 1b photosensitized by (thio)pyr-
ylium salts. For this reason, oxetane 1b was submitted to
parallel studies using chloranil as ET photosensitizer. In
these experiments, transient absorption spectra with two
bands at 450 nm and 600 nm were clearly observed (Fig-
ure 4). They correspond to the chloranil radical anion plus
the trans-anethole radical cation; the assignment was made
on the basis of the literature data.[22,23a,29] This is consistent
with the results of preparative photolysis, showing that the
chloranil-photosensitized CR of 1b proceeds with the same
regioselectivity as that of 1a, leading to trans-anethole.
The triplet quenching rate constants kq(T1) were deter-
Figure 5. Decay traces of the TϪT absorption of 3 (10Ϫ3 ) meas-
ured at 510 nm in the presence of increasing amounts of 1a, using
acetonitrile as solvent
By plotting the reciprocal lifetimes (1/τ) against the con-
centration of oxetane (or arylalkene), linear relationships
were obtained, as shown in Figure 6.
The slopes of the straight lines correspond to kq(T1), as
deduced from Equation (2). Data are presented in Table 2,
and good correlation coefficients were found in all cases.
1/τ ϭ ko ϩ kq (T1) [quencher]
(2)
Results showed that triplet quenching in benzene is very
mined by LFP from the time profiles obtained for the decay slow compared to that found in acetonitrile and methanol,
of the triplet-triplet absorption of 3, measured at and the reaction rate of the oxetane is lower than that of
510 nm[22,23a] in the presence of increasing amounts of 1a, the olefin. Moreover, in the presence of the methoxy sub-
1b, 4a or 4b in benzene, methanol and acetonitrile. A typical stituent in the aryl group, electron transfer is faster. This is
example is shown in Figure 5.
related to the free energy changes associated with electron
1428
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org Eur. J. Org. Chem. 2004, 1424Ϫ1431

Electron-Transfer Cycloreversion of 2,3-Diaryloxetanes
FULL PAPER
the results are also given in Table 2. According to these re-
sults, the reaction should be possible in all cases, except for
the oxetane 1a in benzene. This is in agreement with the
observation that no photoproduct was obtained in the ir-
radiation of mixtures of 1a and 3 in benzene. Thus, the
coulombic factor appears to be very important in this case,
in order to make the electron transfer process favourable.
Conclusion
Overall, the above results are consistent with C2ϪC3
bond breaking of the oxetane radical cations. This contrasts
with the previous findings on the cycloreversion of the
trans,trans-4-methyl-2,3-diphenyloxetane (1a) photosensit-
ized by pyrylium salts, where initial OϪC2 bond cleavage
was found to occur. The reaction mechanism is supported
by formation of β-methylstyrenes (or photoproducts de-
rived thereof) and by the detection of anethole radical cat-
ions rather than 4-methoxystilbene radical cations.
In the case of pyrylium salts, the reaction would be ex-
pected to occur at the free ion stage; in contrast, the contri-
bution of the solvent-separated ion pairs should be higher
when using chloranil as the photosensitizer. Regarding this
point, it is interesting to make a comparison with the split-
ting modes of the molecular ions in the mass spectrometer;
this could be related to cleavage of the oxetane radical ions
in the gas phase in the absence of a counter anion, with a
high excess of energy. Indeed, the radical cations of both
4a/b and 5a/b are formed upon electron impact mass spec-
trometry of 1a and 1b. Finally, methoxy substitution at the
3-position of the aryl group induces C2ϪC3 cleavage of 2,3-
diaryloxetane radical cations, which is independent of the
photosensitizer employed. Thus, the regioselectivity of ET-
mediated CR of oxetanes can be controlled either by mod-
ifying the substitution pattern or by changing the nature of
the electron transfer photosensitizer.
Figure 5. Plot to obtain kq(T1) for quenching of triplet 3 by 1a
(top) and 4a (bottom) in several solvents: C6H6 (squares), MeCN
(triangles) and MeOH (circles)
Table 2. Thermodynamics and kinetics of the ET reaction of 1a,
1b, 4a and 4b photosensitized with the triplet excited state of 3
Solvent
10Ϫ9 ϫ kq(T1)[a]
∆GET(T1)[b]
1a
1b
4a
4b
C6H6
Ͻ 0.1
0.3
0.4
0.9
6.5
3.9
1.9
7.2
4.7
ϩ13
Ϫ7
Ϫ7
MeCN
MeOH
C6H6
MeCN
MeOH
C6H6
MeCN
MeOH
C6H6
MeCN
MeOH
ϩ3.8
Ϫ16
Ϫ16
ϩ7
Ϫ12
Ϫ13
ϩ0.3
Ϫ19
Ϫ19
Experimental Section
Chemicals: Pyrylium salt 2a, chloranil (3), trans-β-methylstyrene
(4a) and trans-anethole (4b) are commercially available. The thiapy-
rylium salt 2b was synthesized according to the procedure de-
scribed by Wizinger and Ulrich.[39] The oxetanes 1a and 1b were
prepared by the PaternoϪBüchi photocycloaddition of benzal-
dehyde and trans-β-methylstyrene[33] or trans-anethole,[40] accord-
3.6
10.7
9.6
1
[a]
[b]
Given in sϪ1Ϫ1
.
Given in kcal/mol.
ing to literature procedures. Characterization was achieved by H
and 13C NMR spectra, which were recorded with a Varian Gemini
300 spectrometer at 300 MHz and 75 MHz, respectively. Data for
the known compounds was consistent with that found in the litera-
ture.
transfer from the triplet excited state of the sensitizer, calcu-
lated using the Weller’s semiempirical equation[24] [see
Equation (1) above].
In this case, the redox potentials EDϩ·/D of 1a, 1b, 4a and
4b have been previously reported.[25,32] For 3, it is known[21]
that EA/A-· ϭ Ϫ0.02 V vs. SCE and ET ϭ 49 kcal/mol. The
coulombic interaction is ϩ0.79 eV in benzene, Ϫ0.06 eV in
acetonitrile and Ϫ0.05 eV methanol.[37,38] The ∆GET(T1)
Time-Resolved Absorption Spectroscopy: The laser flash photolysis
system was based on a pulsed Nd (YAG SL404G-10, Spectron
Laser Systems), using 355 nm radiation as the excitation wave-
length. The single pulses were of ca. 10 ns duration and the energy
was ca. 20 mJ/pulse. A Lo255 Oriel xenon lamp was employed as
the detecting light source. The laser flash photolysis apparatus con-
values were obtained taking the above data into account; sisted of the pulsed laser, the Xe lamp, a 77200 Oriel monochroma-
Eur. J. Org. Chem. 2004, 1424Ϫ1431
www.eurjoc.org  2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 1429

M. A. Izquierdo, M. A. Miranda
FULL PAPER
tor, an Oriel photomultiplier (PMT) system (made up of a 77348
side-on PMT tube, a 70680 PMT housing and a 70705 PMT power
supply). A TDS-640A Tektronix oscilloscope was used. The output
signal from the oscilloscope was transferred to a personal computer
for study.
CD3COCD3, 25 °C): δ ϭ 19.7 (CH3), 70.9 (CH), 91.1 (CH), 120.7
(C), 127.5 (CH), 128.8 (CH), 129.3 (CH), 138.8 (C), 146.2 (C),
147.7 (C) ppm. C15H12Cl4O2 (366.0598): calcd. C 47.16, H 3.17, Cl
37.12; found C 46.56, H 2.96, Cl 36.98.
2,3,5,6-Tetrachloro-4-(2Ј-methoxy-1Ј-phenylpropoxy)phenol
(8): Solutions of 4a (0.43 g, 3.6 mmol) and 3 (0.30 g, 1.2 mmol) in
methanol were placed in Pyrex tubes and irradiated in the multil-
amp photo reactor (λmax. ϭ 350 nm, filter λ Ͼ 340 nm) for 3 h.
After this time, the solutions were partially concentrated to precipi-
tate unchanged chloranil, which was filtered off. Subsequent con-
centration to dryness gave a dark orange precipitate containing 7,
which was purified by recrystallization from benzene/petroleum
ether and then from benzene, until 8 (25 mg, 5.3%) was obtained
Cycloreversion Procedure Using (Thio)pyrylium Salts as PET Sensit-
izers: The following conditions were used: oxetane 1b: 4 ϫ 10Ϫ2 ;
2a/b: 2 ϫ 10Ϫ3 ; solvent: CDCl3 (0.8 mL); inert gas: argon; time:
15 min. Irradiation was performed in a photoreactor, using 4 ϫ
8 W lamps with an emission maximum λmax. ϭ 350 nm and filter
λ Ͼ 340 nm. There was no reaction in the dark or in the absence
1
of the photosensitizer. Reactions were monitored by H NMR, re-
corded before and after irradiation.
1
as a white solid. H NMR (300 MHz, TMS as internal standard,
PET-Irradiation Procedure Using Choranil as the Electron Acceptor:
Solutions of 1a (0.01 ) and 3 (0.01 ) in the required solvent
(benzene, acetonitrile and methanol) were placed in Pyrex tubes.
Argon was then bubbled through the solutions for 30 min. They
were irradiated in a multilamp photo reactor, using 4 ϫ 8 W lamps
with an emission maximum λmax. ϭ 350 nm and filter λ Ͼ 340 nm.
There was no reaction in the dark or in the absence of the photo-
sensitizer. After irradiation, reaction mixtures were concentrated
CDCl3, 25 °C): δ ϭ 0.90 (d, J ϭ 6.3 Hz, 3 H, CH3), 3.36 (s, 3 H,
OCH3), 3.97 (dq, J1 ϭ 6.3, J2 ϭ 7.4 Hz, 1 H, 2Ј-H), 5.35 (d, J ϭ
7.4 Hz, 1 H, 1Ј-H), 5.95 (s, 1 H, OH), 7.32Ϫ7.50 (m, 5 H, CH,
phenyl) ppm. 13C NMR (75 MHz, CDCl3, 25 °C): δ ϭ 15.4 (CH3),
56.9 (CH3), 80.2 (CH), 88.4 (CH), 118.7 (C), 126.6 (CH), 128.0
(CH), 128.2 (CH), 136.9 (C), 145.1 (C), 146.5 (C) ppm.
C16H14Cl4O3 (396.0806): calcd. C 48.52, H 3.56; found C 48.44,
H 3.56.
1
and analyzed by H NMR in deuterated chloroform. Results were
Supporting Information: Copies of the 1H and 13C NMR spectra
for the new photoproducts 6, 7 and 8 (Figures S.I.1, 2 and 3,
respectively). Representative structural data used for the tentative
assignment of 3-hydroxy-1-methoxy-1,2-diphenylbutane, formed in
small amounts. Laser flash photolysis spectra of trans-anethole (4b)
and (thio)pyrylium salts 2a,b (Figures S.I.4 and 5).
compared with those obtained for the irradiation of 4a and 3 under
the same conditions. A similar procedure was applied for the ir-
radiation of 1b (0.007 ) and 3 (0.0032 ) in benzene and aceto-
nitrile. Results were compared with those obtained for the ir-
radiation of 4b and 3 under the same conditions. Structural assign-
ment was performed by comparison with authentic samples ob-
tained by an alternative synthesis, namely irradiating chloranil 3 in
the presence of trans-β-methylstyrene 4a.
Acknowledgments
Financial support by the Spanish MCYT (grant BQU2001-2725)
is gratefully acknowledged.
5,6,8,9-Tetrachloro-3-methyl-4-phenyl-2-oxaspiro[3,5]nona-5,8-dien-
7-one (6): Solutions of trans-β-methylstyrene (4a, 0.53 g, 4 mmol)
and chloranil (3, 0.53 g, 2 mmol) in benzene (100 mL) were placed
in Pyrex tubes and irradiated for 4 h in the multilamp photo reactor
(λmax. ϭ 350 nm, filter λ Ͼ 340 nm). After this time, the solution
was concentrated (to 10 mL) to precipitate unchanged chloranil,
which was filtered off. The filtrate containing 6 was concentrated
to dryness and the resulting dark orange solid was repeatedly
recrystallized from benzene/petroleum ether, until a white solid 6
(30 mg, 4.1%) was obtained. 1H NMR (300 MHz, TMS as internal
standard, CDCl3, 25 °C): δ ϭ 1.78 (d, J ϭ 6.2 Hz, 3 H, CH3), 4.60
(d, J ϭ 8.3 Hz, 1 H, 4-H), 5.80 (m, 1 H, 3-H), 6.89 (m, 2 H, CH,
phenyl), 7.29 (m, 3 H, CH, phenyl) ppm. 13C NMR (75 MHz,
CDCl3, 25 °C): δ ϭ 21.8 (CH3), 57.3 (CH), 79.0 (CH), 86.8 (C),
125. 9 (CH), 128.1 (CH), 128.8 (CH), 130.5 (C), 131.7 (C), 133.7
(C), 151.4 (C), 152.0 (C), 169.7 (C) ppm. C15H10Cl4O2 (364.044):
calcd. C 49.50, H 2.75, Cl 38.96; found C 49.00, H 2.66, Cl 38.71.
[1]
M. Schmittel, A. Burghart, Angew. Chem. Int. Ed. Engl. 1997,
36, 2250Ϫ2589.
[2]
[2a]
See for example:
H. Tomioka, D. Kobayashi, A. Hashim-
[2b]
oto, S. Murata, Tetrahedron Lett. 1989, 30, 4685Ϫ4688.
Y.
Takahashi, T. Miyashi, T. Mukai, J. Am. Chem. Soc. 1983, 105,
6511Ϫ6513. [2c] D. R. Arnold, X. Du, J. Am. Chem. Soc. 1989,
[2d]
111, 7666Ϫ7667.
K. Mizuno, J. Ogawa, Y. Otsuji, J. Org.
Chem. 1992, 57, 4669Ϫ4675.
[3]
[4]
[5]
K. Nakabayashi, J. I. Kojima, K. Tanabe, M. Yasuda, K.
Shima, Bull. Chem. Soc. Jpn. 1989, 62, 96Ϫ101.
G. Prakash, D. E. Falvey, J. Am. Chem. Soc. 1995, 117,
11375Ϫ11376.
A. G. Griesbeck, in CRC Handbook of Organic Photochemistry
and Photobiology (Eds.: W. M. Horspool, P.-S. Song), Boca Ra-
ton, FL, 1995, pp. 522Ϫ535.
2,3,5,6-Tetrachloro-4-(2Ј-hydroxy-1Ј-phenylpropoxy)phenol
(7): Solutions of 4a (0.35 g, 3 mmol) and 3 (0.62 g, 2.5 mmol) in
acetonitrile (120 mL) were placed in Pyrex tubes and irradiated in
the multilamp photo reactor for 4 h. The volume was concentrated
(to 10 mL) to precipitate unchanged chloranil, which was filtered
off. The filtrate was concentrated to dryness to give a dark orange
precipitate, containing a mixture of 6 and 7. The photoproduct 7
was isolated from the first crop in the recrystallization of the mix-
ture from hot benzene/petroleum ether. It was purified by repeated
recrystallization, to give 7 (12 mg, 1.3%) as a white solid. 1H NMR
(300 MHz, TMS as internal standard, CDCl3, 25 °C): δ ϭ 1.03 (d,
J ϭ 6.4 Hz, 3 H, CH3), 3.00 (s, 1 H, OH), 4.50 (dq, J1 ϭ 6.4, J2 ϭ
7.6 Hz, 1 H, 2Ј-H), 5.29 (d, J ϭ 7.6 Hz, 1 H, 1Ј-H), 5.91 (s, 1 H,
OH), 7.26Ϫ7.31 (m, 5 H, CH, phenyl) ppm. 13C NMR (75 MHz,
[6]
[7]
T. M. Trnka, R. H. Grubbs, Acc. Chem. Res. 2001, 34, 18Ϫ29.
A. Fürstner, Angew. Chem. Int. Ed. 2000, 39, 3012Ϫ3043.
[8] [8a]
[8b]
T. Bach, Synthesis 1998, 683Ϫ703.
T. Bach, Liebigs
Ann./Recueil 1997, 1627Ϫ1634. [8c] G. Jones II, M. A. Aquadro,
M. A. Carmody, J. Chem. Soc., Chem. Commun. 1975, 206.
Y. Wang, P. P. Gaspar, J. S. Taylor, J. Am. Chem. Soc. 2000,
122, 5510Ϫ5519.
[9]
[10]
[11]
[12]
[13]
X. Zhao, J. Liu, D. S. Hsu, S. Zhao, J. S. Taylor, A. Sancar, J.
Biol. Chem. 1997, 272, 32580Ϫ32590.
A. Joseph, D. E. Falvey, Photochem. Photobiol. Sci. 2002, 1,
632Ϫ635.
M. K. Cichon, S. Arnold, T. Carell, Angew. Chem. Int. Ed.
2002, 41, 767Ϫ770.
M. A. Miranda, M. A. Izquierdo, F. Galindo, J. Org. Chem.
2002, 67, 4138Ϫ4142.
1430
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 1424Ϫ1431

Electron-Transfer Cycloreversion of 2,3-Diaryloxetanes
FULL PAPER
F. L. Cozens, R. Bogdanova,
[14]
[29c]
M. A. Miranda, M. A. Izquierdo, J. Am. Chem. Soc. 2002,
Soc. 1993, 115, 6564Ϫ6571.
124, 6532Ϫ6533.
M. Regimbald, H. Garcia, V. Marti, J. C. Scaiano, J. Phys.
Chem. B 1997, 101, 6921Ϫ6928.
[15]
C. Pac, T. Fukunaga, Y. Go-An, T. Sakae, S. Yanagida, J. Pho-
[30] [30a]
Y. Kuriyama, T. Arai, H. Sakuragi, K. Tokumaru, Chem.
tochem. Photobiol., A. 1987, 41, 37Ϫ51.
F. D. Lewis, M. Kojima, J. Am. Chem. Soc. 1988, 110,
8664Ϫ8670.
N. P. Schepp, L. J. Johnston, J. Am. Chem. Soc. 1994, 116,
6895Ϫ6903.
O. Brede, F. David, S. Steenken, J. Chem. Soc., Perkin Trans. 2
[30b]
[16]
Lett. 1988, 1193Ϫ1196.
R. Akaba, K. Oshima, Y. Kawai,
Y. Obuchi, A. Negishi, H. Sakuragi, K. Tokumaru, Tetrahedron
Lett. 1991, 32, 109Ϫ112.
T. Majima, S. Tojo, A. Ishida, S. Takamuku, J. Org. Chem.
1996, 61, 7793Ϫ7800.
N. P. Schepp, L. J. Johnston, J. Am. Chem. Soc. 1996, 118,
2872Ϫ2881.
[17]
[31]
[32]
[33]
[18]
1995, 23Ϫ32.
[19]
[19a]
For preliminary communication see:
M. A. Miranda, M.
S. A. Fleming, J. J. Gao, Tetrahedron Lett. 1997, 38,
A. Izquierdo, Chem. Commun. 2003, 364Ϫ365. [19b] M. A. Mir-
anda, M. A. Izquierdo, Photochem. Photobiol. Sci. 2003, 2,
848Ϫ850.
5407Ϫ5410.
[34] [34a]
H. D. Roth, M. L. M. Schilling, J. Am. Chem. Soc. 1980,
[34b]
102, 4303Ϫ4310.
N. P. Schepp, L. J. Johnston, J. Am.
[20]
M. A. Miranda, H. Garcia, Chem. Rev. 1994, 94, 1063Ϫ1089.
Chem. Soc. 1994, 116, 6895Ϫ6903.
[21] [21a]
[21b]
M. E. Peover, Nature (London) 1961, 191, 702Ϫ703.
[35]
A methanol adduct (tentatively assigned as 3-hydroxy-1-meth-
oxy-1,2-diphenylbutane) was obtained in small amounts as a
mixture of diastereoisomers, resulting from the nucleophilic at-
tack of methanol on the oxetane-derived 1,4-radical cation.
The structural assignment was done by comparison of the spec-
troscopic data (1H and 13C NMR, see Supporting Information)
with those of related compounds.[36] The relative positions of
the methoxy and hydroxy groups was confirmed by GC-MS
R. Gschwind, E. Haselbach, Helv. Chim. Acta 1979, 62,
941Ϫ955. [21c] J. J. Andre, G. Weill, Mol. Phys. 1968, 15, 97Ϫ99.
[22] [22a]
H. Xu, L. C. Wang, J. W. Xu, B. Z. Yan, H. C. Yuan, J.
[22b]
Chem. Soc., Perkin Trans. 1 1994, 571Ϫ577.
S. M. Hubig,
J. K. Kochi, J. Chem. Soc., Perkin Trans. 2 1999, 781Ϫ788.
[23] [23a]
W. Bergmark, S. Hector, G. Jones II, C. Oh, T. Kumagai,
S. I. Hara, T. Segawa, N. Tanaka, T. Mukai, J. Photochem.
[23b]
Photobiol. A Chem. 1997, 109, 119Ϫ124.
T. Miyashi, Y.
(presence of
a peak at m/z ϭ 121, corresponding to
Takahashi, H. Ohaku, K. Yokogawa, A. Morishima, T. Mukai,
1
[PhCHOCH3]ϩ) and by the changes in the H NMR spectrum
Tetrahedron Lett. 1990, 17, 2411Ϫ2414.
upon acetylation with acetyl chloride.
[24]
[25]
A. Weller, Z. Phys. Chem. (Wiesbaden) 1982, 133, 93.
M. A. Miranda, M. A. Izquierdo, R. Perez-Ruiz, J. Phys.
Chem. A 2003, 107, 2478Ϫ2482.
M. Martiny, E. Steckhan, T. Esch, Chem. Ber. 1993, 126,
1671Ϫ1682.
[36] [36a]
A. Oku, M. Abe, M. Iwamoto, J. Org. Chem. 1994, 59,
[36b]
7445.
A. R. Katritzky, S. Rachwal, B. Rachwal, P. J. Steel,
J. Org. Chem. 1992, 57, 4925.
[26]
[27]
[37]
[38]
J. Mattay, J. Runsink, J. Gersdorf, T. Rumbach, C. Ly, Helv.
Chim. Acta 1986, 69, 442Ϫ455.
R. Searle, J. L. R. Williams, D. E. DeMeyer, J. C. Doty, J.
Chem. Soc., Chem. Commun. 1967, 22, 1165.
The coulombic interaction in a given solvent (∆Ecoul.) is de-
rived from the Born equation as shown by Weller. Both the
Coulombic attraction and the solvation entalpy were taken into
account: ∆Ecoul. [eV] ϭ (e/4πεoa)·(1/ε Ϫ 2/37.5), where e ϭ
[28] [28a]
S. Parret, F. Morlet-Savary, J. P. Fouassier, K. Inomata, T.
Matsumoto, F. Heisel, Bull. Chem. Soc. Jpn. 1995, 68,
2791Ϫ2795. [28b] R. Akaba, M. Kamata, A. Koike, K. I. Mogi,
Y. Kuriyama, H. Sakuragi, J. Phys. Org. Chem. 1997, 10,
861Ϫ869. [28c] N. Manoj, R. A. Kumar, K. R. Gopidas, J. Pho-
1.602 ϫ 10Ϫ19 C, εo ϭ 8.854 ϫ 10Ϫ12 CVϪ1mϪ1 and a ϭ 7 A.
˚
The value of ε is known to be 2.28 in benzene, 37.5 in aceto-
nitrile and 32.7 in methanol.
[39]
[40]
tochem. Photobiol. A Chem. 1997, 109, 109Ϫ118.
R. Wizinger, P. Ulrich, Helv. Chim. Acta 1956, 39, 1Ϫ4.
H. Nozaki, I. Otani, R. Noyori, M. Kawanisi, Tetrahedron
1967, 24, 2183Ϫ2192.
[29]
[29a]
See for example:
N. P. Shepp, D. Shukla, H. Sarker, N. L.
Bauld, L. J. Johnston, J. Am. Chem. Soc. 1997, 119,
[29b]
10325Ϫ10334.
L. J. Johnston, N. P. Schepp, J. Am. Chem.
Received September 14, 2003
Eur. J. Org. Chem. 2004, 1424Ϫ1431
www.eurjoc.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1431