Photochemistry of thiophene-S-oxide derivatives

Article abstract of DOI:10.1002/poc.1384

Photolysis of substituted thiophene-S-oxides in solution results in the formation of either the corresponding thiophene or furan, in addition to uncharacterized materials. No good rationalization is available for the choice of which pathway may predominat

Full text of DOI:10.1002/poc.1384

Research Article
Received: 26 February 2008,
Revised: 1 April 2008,
Accepted: 3 April 2008,
Published online in Wiley InterScience: 2 June 2008
(www.interscience.wiley.com) DOI 10.1002/poc.1384
Photochemistry of thiophene-S-oxide
derivatives
a
a
a
*
Melanie J. Heying , Mrinmoy Nag and William S. Jenks
Photolysis of substituted thiophene-S-oxides in solution results in the formation of either the corresponding
thiophene or furan, in addition to uncharacterized materials. No good rationalization is available for the choice
of which pathway may predominate, but it is demonstrated that the photolysis of 2,5-bistrimethylsilylthiopene-
S-oxide produces O(3P) in the same manner as the well-established photolysis of dibenzothiophene-S-oxide.
Copyright ß 2008 John Wiley & Sons, Ltd.
Supplementary electronic material for this paper is available in Wiley InterScience at http://www.mrw.interscience. wiley.com/
suppmat/0894-3230/suppmat/
Keywords: thiophene-S-oxides; thiophenes; photochemistry; deoxygenation
rather than polarity, viscosity or other ‘physical’ parameters.[7] The
INTRODUCTION
solvents and substituents that led to higher quantum yields
contained functionalities that shared in common a high reactivity
with O(3P). Other workers showed that the reactivity of the
oxidizing species followed an electrophilic trend in the oxidation
of sulfides and alkenes that was consistent with known reactivity
from gas phase chemistry.[8,9]
We believed that DBTO or a related derivative would make an
excellent source for flash photolysis experiments and/or other
work in which photochemical release of O(3P) was desired, but
this was hampered by the low quantum yield (generally < 0.01).
Thus, we began to look for ways of increasing the quantum yield
of deoxygenation. One successful approach was to substitute Se
for S, making the corresponding selenoxide.[10] A much more
modest degree of success was found with heavy atom
substitution on the arene.[11]
The photochemical deoxygenation of dibenzothiophene-
S-oxides (DBTOs) to dibenzothiophenes (DBTs) and the corre-
sponding reactions of thiophene-S-oxide (TO) derivatives has a
history going back into the 1970s, when first reported by Gurria
and Posner.[1] Deoxygenation of other sulfoxides is known,[2–4]
but it appears that the cases with the greatest chemical yield are
generally limited to those in which the sulfur atom resides within
a formal thiophene ring.
A much earlier hypothesis, though, was that derivatives of TO
might be particularly good at deoxygenation because of their
lower S—O bond dissociation enthalpies[12,13] and higher excited
state energies (Table 1). (A potential flaw in this reasoning is the
low triplet energy of cyclopentadiene, which might reasonably
used as a first estimate for the triplet energy of TO.) This made
them an attractive target, even though the anticipated
absorption spectrum took them out of the range that might
be useful with a 355 nm laser line.[14] An important issue,
however, was that TO derivatives must be substituted with
relatively bulky groups to kinetically stabilize them against
dimerization. Furthermore, it was only in the mid 1990s that the
synthesis of TO derivatives became practical.[15,16]
In 1997, we reported a mechanistic study on DBTO photolysis,
in which we suggested that the primary mechanism for
photochemical deoxygenation of DBTO was direct dissociation
of the molecule to DBT and atomic oxygen, O(3P).[5,6] This
proposal was based mainly on experiments that were thought to
exclude other mechanisms – including transient dimer formation
and hydrogen abstraction – and on the observation of oxidized
solvent in a pattern that seemed appropriate for triplet atomic
oxygen. For example, benzene is oxidized to phenol and alkanes
are hydroxylated with some selectivity. Although the chemical
yield of DBT is high, the photochemical efficiency is poor and
subject to solvent and wavelength effects. More recent work on
DBTO derivatives from our laboratories has shown that solvent
effects in the quantum yield were due to solvent functionality,
In work unpublished from our laboratories, save for a PhD
dissertation,[17] we made the observation that photolysis of
2,5-diphenylthiophene-S-oxide did yield the corresponding
*
Correspondence to: W. S. Jenks, Department of Chemistry, Iowa State
University, Ames, IA 50011-3111, USA.
a
M. J. Heying, M. Nag, W. S. Jenks
Department of Chemistry, Iowa State University, Ames, Iowa 50011-3111,
USA
J. Phys. Org. Chem. 2008, 21 915–924
Copyright ß 2008 John Wiley & Sons, Ltd.

M. J. HEYING, M. NAG AND W. S. JENKS
thiophene in some measurable yield.[20] Those that produce
other products (e.g., the furan) can be induced to yield mainly
the thiophene by addition of several equivalents of an amine,
triphenylphosphine or an electron rich thiol. This may be
mechanistically related to the observation of electron-transfer-
sensitized photoreductions of a number of sulfoxides.[24,25]
Table 1. Relevant energies for thiophene sulfoxides
S–O bond dissociation
energy, kcal/mol
ES
kcal/mol
ET
kcal/mol
DBTO
TO
73a
61a
ca. 85b
78.6c
ca. 60b
ca. 59d
‘Benzyl’ functionalization
a Reference[13].
b Reference[56].
Thiemann examined a series of TO derivatives that contained
methyl groups in the 2 and 5 positions. In the absence of
hydroxylic solvent, the thienyl alcohol and/or thienyl ether is
observed as part of the product mixture; ethanol was used to
form the ethyl thienyl ether in at least one case.[19,20,26] This
reactivity was attributed to secondary reactions between water
or an alcohol and the corresponding 2-methylene-3H-thiophene-
S-oxides or 2-methylene-5H-thiophene-S-oxides, though neither
of these intermediates has been observed directly.
In that the ‘benzyl’ functionalization was avoidable by not
using certain substituents, we undertook a study of some of these
reactions that was meant to be more mechanistic in nature and
that we report here. We now report the photochemistry of TOs
1a–1c, which shows both deoxygenation and the furan formation
without the complication of benzyl functionalization. We also
report a brief computational investigation on the mechanism of
furan formation.
c Apparent 0,0 band from absorption spectrum of 1c (this
work).
d Triplet energy of cyclopentadiene.
sulfide in low quantum yield, along with other products. Since we
understood that the phenyl groups also would affect the
chromophore and we were also interested in doing compu-
tational chemistry on the system, we did not pursue this TO
derivative further. However, we also observed at this time that
photolysis of 2,5-di-t-butylthiophene-S-oxide resulted in the
surprising formation of the corresponding furan.[17] We were thus
relieved and pleased to see the work of Thiemann et al.,[18,19] in
which they more thoroughly documented this transformation.
They have recently elaborated these studies with several more
sulfoxides.
Thiemann has recently reviewed his group’s work in this
area,[20] but a short synopsis is useful here.
RESULTS
We wished to study the photochemistry of a series of thiophene
oxides that did not include the potentially complicating
substituents of Br (photochemical cleavage) or 2,5-dimethyl
(isomerization). Ideal substrates seemed to be 3,4-di-
t-butylthiophene-S-oxide or 3,4-di-neopentylthiophene-S-oxide,
which are prepared by oxidation of the appropriate thiophene.
The former of these has been prepared (via a McMurry coupling
for the key cyclization step)[27–30] and its chemistry has been
Furan formation
Certain TO derivatives produce the corresponding furan on
photolysis, in minor to major yield, depending on the structure.
There is not yet a good rationalization for the choice of which
sulfoxides produce the furan. Among them are 2,4- or 2,5-di-
t-butylthiophene-S-oxide and tetraphenylthiophene-S-oxide. No
firm mechanism can yet be established. However, Thiemann has
proposed a mechanism passing through the cyclic oxathiin
isomer (Scheme 1, illustrated for 2,5-di-t-butylthiophene-S-
oxide).[20] The first step, a-cleavage, is a mechanism that is well
established in sulfoxide photochemistry.[3,4] There is ample
precedent for analogous dithiins photochemically desulfurizing
to produce thiophene.[21–23] There is nothing published that
reported.[31–33] We do not dispute these reports, but in our hands,
the McMurry coupling was extremely problematic for this
substrate and did not result in synthetically useful product
mixtures.[34] We had similar difficulties preparing the di-
neopentylthiophene and were also unsuccessful in coupling
neopentyl groups to dibromothiophene via the Kumada method.
A second synthetic procedure toward di-t-butylthiophene, in
which 2,5-dibromothiophene was subjected to Friedel-Crafts
alkylation with tBuCl and AlCl3, followed by debromination,
resulted in the isolation of 2,5-di-t-butylthiophene.[35] Ultimately,
we prepared and studied the photochemistry of three sulfoxides:
3,4-diphenylthiophene-S-oxide, 1a, 3,4-dibenzylthiophene-S-
oxide, 1b, and 2,5-bis(trimethylsilyl)thiophene-S-oxide, 1c. These
sulfoxides were prepared by the oxidation of the corresponding
distinguishes whether
a second photon is required for
desulfurization of TOs, but the general pathway of going through
the oxathiin seems quite reasonable.
Deoxygenation
Nearly every sulfoxide examined by Thiemann, save 2,5-di-t-
butylthiophene-S-oxide, undergoes deoxygenation to form the
www.interscience.wiley.com/journal/poc
Copyright ß 2008 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2008, 21 915–924

THIOPHENE-S-OXIDE PHOTOCHEMISTRY
Scheme 1.
thiophenes, whose individual preparations are described in the
Experimental section, by oxidation by mCPBA in the presence
of BF3.[15,29]
retention times and apparent masses identical to 3a (GCMS).
In methanol, an unidentified precipitate was formed if solutions
of ca. 20 mM initial concentration were photolyzed.
Photolysis of dibenzyl derivative 1b also led to furan (3b) in low
yield without observation of the thiophene. A small quantity of
bibenzyl was detected in all three solvents, but no other
identifiable compounds were noted.
In contrast to 1a and 1b, photolysis of 2,5- bis(trimethylsi-
lyl)thiophene-S-oxide (1c) led to deoxygenation (2c) to the
exclusion of furan formation. Consistent with our observation on
DBTO,[5] excitation at a shorter wavelength increased the
quantum yield of sulfide formation.[36] No product corresponding
to the mass or 1H NMR of 3c[37] was obtained.
In order to probe for O(3P) formation, a suitable probe that did
not absorb the incident light was sought; alkane hydroxylation
seemed ideal. Photolyses of 1c or DBTO in 2-methylbutane led to
very similar quantities of alkane hydroxylation products.
Adjusting for the number of hydrogens, a hydroxylation
selectivity of 3.0:1.7:1 was observed for tertiary:secondary:prim-
ary positions (Scheme 2). This agrees well with our previously
reported result for DBTO.[6]
Direct photolysis
Table 2 summarizes the results obtained on photolysis of
sulfoxides 1a, 1b, and 1c. Photolyses were carried out in
Ar-flushed solutions with initial concentrations of 1–5 mM.
Modest yields of 2 or 3 were obtained. Neither identifiable
products were obtained, nor were any other soluble products
ever formed as major products. The quantum yield for loss is
defined as the observed quantum yield for all processes leading
to the loss of the respective starting material.
Photolysis of 3,4-diphenylthiophene-S-oxide in either aceto-
nitrile or methanol resulted in modest amounts of furan (3a)
formation, near 22%. Yields were determined by 1H NMR and GC.
No thiophene was observed, regardless of the degree of
conversion. Several minor, unidentifiable products were formed.
At least some appeared to be isomeric with 3a, in that they had
aromatic protons in the NMR, differing chromatographic
Table 2. Results of direct photolysis of thiophene sulfoxidesa
Sulfoxide
Solvent
lex, nm
FLoss
Yield 2, (%)
Yield 3, (%)
1a
CH3CN
CH3OH
CH3CN
iPrOH
CDCl3
CH3CN
CH3CN
C6H6
300
300
325
325
325
260
325
325
0.006
0.0072
0.064
0.042
0.021
0.080
0.11
0
0
0
0
0
26
12
17
22
23
17
17
12
0
1b
1c
0
0
0.13
a Quantum yield is for loss of starting sulfoxide by all processes. Yields are relative to the consumed starting material.
Scheme 2.
J. Phys. Org. Chem. 2008, 21 915–924
Copyright ß 2008 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/poc

M. J. HEYING, M. NAG AND W. S. JENKS
We wished to probe the mechanism for furan formation. As
suggested by Thiemann, a mechanism for formation of the
furan can be elaborated from the now well-documented
mechanism for thiophene formation from dithiins.[21–23]
Substituting oxygen for sulfur results in the pathway illustrated
in Scheme 3. Although this varies from the Thiemann pathway
shown in Scheme 1, it shares in common the key oxathiin
intermediate. Certainly troubling in both pathways is the
unspecified mechanism for loss of sulfur in the final step, but
this process is well documented for thiiranes.[21–23] Our primary
interest was thus in the initial photochemical step(s) postulated
to provide the isomers of 5.
We sought direct evidence for the oxathiin intermediate.
Photolyses of 1a were carried out at ꢀ20 8C in CD3CN to low and
then higher conversion, with analysis by NMR at the same
temperature, to try to observe any evidence for the oxathiin 5a.
The only identifiable protons, aside from starting material, were
attributable to 3a, the furan. No protons from any asymmetric
compounds, which would display 1H-1H coupling, were observed.
The low-temperature reaction was repeated with O2 saturation
instead of Ar saturation in an unsuccessful attempt to trap
intermediates in the formation of the furan by oxidation.[10]
(See below for a precedent for O2 as a trap for an intermediate
analogous to this with selenoxide photochemistry.) If the
unstable oxathiin 5a had been formed, there would have been
potential to trap it as the more stable sultine. No difference was
observed between this and the Ar-saturated case.
Computations regarding furan formation
The inability to isolate intermediates as illustrated in Scheme 3
does not necessarily eliminate this general pathway for the
formation of furan. In particular, if the intermediates are
destroyed more rapidly than they are formed under the reaction
conditions, they will never accumulate. We thus undertook a brief
computational study to investigate the relative energies of the
various isomers.
Computations were carried out at the B3LYP and MP2 levels of
theory with the GAMESS[39] suite of programs. Geometries were
optimized with the 6-31G(d) basis set, and the respective
vibrational matrices showed that each structure was a minimum.
It is well known that fairly large basis sets that contain tight d
polarization functions are required to get good relative energies for
the oxides of sulfur[13,40–52] We thus used the aug-cc- pV(T þ d)Z
basis set to obtain single point energies at the so-obtained
geometries (i.e., MP2/aug-cc-pV(T þ d)Z//MP2/6- 31G(d) or B3LYP/
aug-cc-pV(T þ d)Z//B3LYP/6-31G(d)). We did not attempt to find
transition structures connecting the compounds.
The free energies (G) in Fig. 1 include zero point energies and a
temperature correction to 298 K. The energy of the episulfide (7)
is arbitrarily set to zero. In Fig. 1a, the energies of the sulfoxide,
oxathiin, and episulfide are all given. Additionally, the isomeric
epoxide is shown.
Three conformations for compound 6 were obtained. The
lowest energy conformation had both the C S and C
—
—
O
—
—
rotated away (‘‘exo’’) from the internal pi bond. The ‘‘O-exo’’
conformation has the sulfur atom canted out of plane by ꢁ158
and the highest energy. The S-exo conformation was approxi-
mately planar, and the di-exo conformation was rigorously planar.
For oxathiin, the O and S are displaced above and below the
plane of the carbons, as expected. Figure 2 illustrates the CC, CS,
CO, and SO bond lengths calculated at the MP2/6-31G(d) and
B3LYP/6-31G(d) levels.
By way of comparison, three corresponding structures
corresponding to the C4H4S2 energy surface were optimized
and their energies were calculated in the same way. They are
plotted on the same energy scale in Fig. 1b. The thiosulfoxide is
Photolysis of 1a was carried out in a perdeuterated ethanol/
methanol glass at 77 K in an NMR tube. The frozen glass was
warmed to ꢀ50 8C in the NMR probe, and analysis was carried
out. Again, the only identifiable product was the furan 3a at both
ꢁ5% conversion and complete conversion.
We have recently shown[38] that malonate S,C-sulfonium ylides of
thiophene generate carbenes, but we thought there might be a
chance of observing an isomeric product analogous to the
proposed oxathiin with the correct substitution pattern. We chose
8, based on the sulfoxide results. However, irradiation of 8 in
acetonitrile resulted only in thiopehene 2b and carbene-derived
products.[38] The thiophene was formed nearly quantitatively
relative to consumption of 8. NMR indicated the presence of
carbene-solvent adduct; aside from protons attributable directly to
2b, neither alkenyl nor aromatic hydrogens were observed. By HPLC
analysis, the only benzyl-containing product was 2b.
—
not a relevant structure. The larger C S substituent in the acyclic
—
isomer makes only the single diexo conformation a relevant
minimum, and of course, there is not a second bicyclic analog,
since both heteroatoms are sulfur.
Scheme 3. A potential pathway for furan formation from thiophene oxides
www.interscience.wiley.com/journal/poc
Copyright ß 2008 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2008, 21 915–924

THIOPHENE-S-OXIDE PHOTOCHEMISTRY
aug-cc-pV(T þ d)Z//MP2/6-31G(d) energies, that is, not including
any correlation energy, followed the same pattern as the B3LYP
calculations, rather than the MP2 calculations. (All of the absolute
and relative energies are given in the Supporting Information.)
This kind of huge discrepancy between B3LYP and MP2
calculations is not expected. Moreover, the kind of bonding
involved in the various molecules in the series is so different that
these isomerizations are far from isodesmic reactions. Any
method-dependent errors that are sensitive to functional groups
will be highlighted, rather than cancelled out. We thus face the
reality that at least one of these two methods is highly flawed for
this set of compounds. As a check for the MP2 data, we examined
the natural orbital occupation numbers, and all were in the
natural range of 0–2.
We are unaware of a similar straightforward diagnostic for
troublesome B3LYP calculations. Furthermore, while many
density functionals are available, there is no straightforward,
rational sequence that one can follow to ensure an improvement
of the quality of the calculations. However, it is now fairly well
accepted that the sequence of HF ! MP2 ! CCSD(T) is a
rigorously improving sequence of methodology for single-
reference problems. While CCSD(T) computations are expensive,
we felt the need to try to resolve the strong disagreement
between the B3LYP and MP2 methodologies.
Thus CCSD(T)/aug-cc-pV(T þ d)Z//MP2/6-31G(d) calculations[54]
were run on the episulfide and diexo conformation of 6 as well as
on 1 and 5. The temperature correction obtained from the MP2
calculations was added to obtain the G(298 K) values. The CCSD(T)
energy differences, which are clearly the most reliable of the three
types obtained, remarkably reflect almost the average of the
widely differing MP2 and B3LYP values. As can be seen in Fig. 1 and
Table 3, the relative energy of the diexo ring-opened compound is
slightly below that of the episulfide. The relative energies of the
conformers of 6 were not explored by the high level calculations, as
the pattern was in agreement in both lower level methods.
Sensitized photolysis
Benzophenone-sensitized photolysis of 1a or 1b in acetonitrile
resulted in a complex product mixture that contained neither
the corresponding thiophenes nor furans. No isolable products
were identified. Both low and high conversion experiments were
run. The benzophenone concentration was approximately
30 mM, near the minimum required to ensure absorption of at
least 99.9% of the light by benzophenone in the presence of
0.5 mM of the respective sulfoxide.
Figure 1. Calculated free energies for isomeric structures of (a) C4H4OS
and (b) C4H4S2. Energies are MP2/aug-cc-pV(Tþ d)Z//MP2/6-31G(d),
CCSD(T)/aug-cc-pV(Tþ d)Z//MP2/6-31G(d), or B3LYP/aug-cc-pV(Tþ d)Z//
3LYP/6-31G(d) with ZPE and 298K temperature corrections taken from
MP2/6-31G(d) or B3LYP/6-31G(d) vibrational calculations, as appropriate
Benzophenone-sensitized photolysis of 1c was not practical
because the absorption spectrum of 1c extended too far to the
red to exclusively irradiate benzophenone.
As can be seen in Fig. 1a, the relative energies of
the C4H4OS species differ dramatically, depending on the
computational method. As a check against great sensitivity to
the geometries, additional single point energies were calculated
at the B3LYP/aug-cc-pV(T þ d)Z//MP2/6-31G(d) level. Only small
differences between these energies and those obtained at B3LYP/
aug-cc-pV(T þ d)Z//B3LYP/6-31G(d) were observed, so the qual-
itative picture in Fig. 1 would not change. Similarly, the use of
Jensen’s a-PC2 basis set (i.e., B3LYP/a-PC2//B3LYP/6-31G(d)),
specifically optimized for use in density functional appli-
cations,[53] gave very similar results to the B3LYP results with
the aug-cc-pV(T þ d)Z calculations. Qualitatively, the HF//
DISCUSSION
Deoxygenation
Of the three sulfoxides examined, only 1c showed significant
yields of deoxygenation, that is, the thiophene 2c. The observed
hydroxylation selectivity of 2-methylbutane is quite similar to
that we reported previously for DBTO,[6] and we thus tentatively
conclude that the mechanism of deoxygenation is the same as
that for DBTO, that is, direct dissociation to form O(3P). Although
the quantum yield is higher than for DBTO, the chemical yield is
J. Phys. Org. Chem. 2008, 21 915–924
Copyright ß 2008 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/poc

M. J. HEYING, M. NAG AND W. S. JENKS
Figure 2. Optimized geometries for C4H4OS isomers. Bond lengths shown in normal text are from MP2/6-31G(d) and those in italics are from B3LYP/
6-31G(d) calculations. CH bond lengths are not illustrated
poor enough so that 1c is probably not an especially useful
compound for the intentional generation of oxygen atoms.
A full study on the sensitized deoxygenation of DBTO is being
reported elsewhere,[55] but it can briefly be stated that
benzophenone-sensitization of DBTO does not produce O(3P),
and the subsequent deoxygenation certainly does not go by the
same mechanism as is observed for direct photolysis of DBTO.
Thus, particularly considering the low triplet energy of TO we
expect, it does not come as a surprise that no evidence of
unimolecular deoxygenation was found on sensitization of either
1b or 1c. (We routinely examine sulfoxides for phosphorescence
to determine the triplet energies, but it is only the exceptional
case in which any is observed, e.g., DBTO itself or methanesulfi-
occurs regardless of whether direct photolysis in neat solvent
produces mainly furan, thiophene, or other products. The clear
implication is that there is a second reduction mechanism in
those instances. We believe that it is quite likely that this is related
to the electron-transfer-mediated sulfoxide reduction previously
demonstrated by Kropp and ourselves.[24,25] Thus, while of
interest, we did not believe it was critical to our current
investigation to reproduce such results for these particular
substrates.
Furan formation
We have assumed from the beginning that photochemical furan
formation from thiophene-based sulfoxides is ‘‘unimolecular,’’ in
that it involves only a single TO molecule. That assumption is due
to two lines of thought. First, although we do not report a
systematic study here, we have never seen evidence for a
nylpyrene.[56,57]
)
Thiemann has reported[20] that addition of easily oxidizable
substrates (e.g., amines and sulfides) to solutions containing TOs
increases the yield of thiophene after photolysis. This effect
Table 3. Relative energies of C4H4OS isomers at various computational levelsa
Oxiraneb
298K
rel
G
1
5
6 O exo
6 S exo
6 di exo
7
B3LYP/aug-cc-pV(T þ d)Z//B3LYP/6-31G(d)
MP2/aug-cc-pV(T þ d)Z//MP2/6-31G(d)
HF/aug-cc-pV(T þ d)Z//MP2/6-31G(d)
B3LYP/aug-cc-pV(T þ d)Z//MP2/6-31G(d)
CCSD(T)/aug-cc-pV(T þ d)Z//MP2/6-31G(d)
10.8
13.3
16.4
10.7
6.6
13.2
7.8
6.6
9.7
ꢀ3.9
5.5
ꢀ6.3
3.8
ꢀ7.7
2.7
ꢀ9.4
ꢀ7.5
ꢀ1.7
0.0
0.0
0.0
0.0
0.0
7.4
6.3
6.8
5.9
ꢀ5.5
ꢀ8.1
ꢀ3.7
ꢀ6.0
a All energies in kcal/mol. Unscaled ZPE and temperature correction to 298 K, calculated at MP2/6-31G(d) [or B3LYP/6-31G(d) for the
density functional calculations] are included.
b Oxirane refers to the structure obtained by exchanging the O and S in structure 7.
www.interscience.wiley.com/journal/poc
Copyright ß 2008 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2008, 21 915–924

THIOPHENE-S-OXIDE PHOTOCHEMISTRY
concentration-dependence on a product distribution. We are
confident that under our conditions, deoxygenation is unim-
olecular (based on the analogous studies of DBTO). If furan
formation involved two TO derivatives, then we would likely have
seen more thiophene formation at low initial concentrations and
more furan formation at higher concentrations, or perhaps a
conversion-dependent product distribution. Nothing of the sort
has been observed in the concentration range of
ca. 10ꢀ4–10ꢀ2 M. The second line of reasoning is that the
reaction would have to be rather complex. The initial step might
be oxygen atom transfer, but this should lead to the observation
of a thiophene derivative in the mixture. The initial step might
also be electron transfer (single-electron disproportionation), but
again, the best analogous evidence is that this should ultimately
lead to observation of thiophene derivative as part of the product
mixture. This said that the large amount of uncharacterized
product material implied by the low yields does leave the door at
least a little open to such a possibility. Nonetheless, we will
proceed further with the assumption that the initial steps of
photochemical furan formation do not require a second TO
nucleus.
Given a fundamentally unimolecular initial step, the formation
of the furan due to irradiation of thiophene oxides seemingly
‘‘must’’ proceed through the oxathiin in the sense that it is hard to
come up with a sensible alternative. The known chemistry of the
dithiins[21–23] that leads to thiophenes seems quite analogous. In
the current sulfoxide-based case, the initial bond C—O bond
formation by way of the oxathiin 5 seems inescapably logical
and inescapably photochemical in origin, due to a-cleavage and
isomerism.
under the experimental conditions. It is possible that narrow
irradiation at a well-chosen wavelength (where absorption by the
starting sulfoxide is much greater than that of 5) could allow for
accumulation of 5, but unfortunately, we do not know the
absorption spectrum of 5, a requirement to design that
experiment properly.
The carbon analog 8 would have yielded 9, which should not
have had extra low-energy bands of the dithiin sort. However, it
simply does not undergo any process besides loss of the carbene
to a substantial degree.
We thus sought further information on the relative stabilities of
the structures in Scheme 3 by computational methods. In
principle, the postulated ring opening from 5 to 6 could be
thermal, but it is photochemical in the dithiin cases.[21-23]
—
However, formation of 6 involves the formation of a C O pi
—
—
bond in lieu of a C S pi bond for the dithiin analog. Given basic
—
—
—
knowledge of the relative strengths of C S versus C O pi
—
—
bonds and S—S versus S—O sigma bonds,[51] it seemed
reasonable to expect that the conversion of 5 into 6 would be
more exergonic than the corresponding opening of a dithiin to a
dithial. This expectation is clearly borne out in the computational
results, though the latter are hardly quantitatively definitive.
Although the conflict between the MP2 and B3LYP calculations
cannot be explained fully at this time, there is still worthwhile
information that can be extracted. The CCSD(T) calculations
clearly show that the ring opening of the oxathiin to 6 is
exothermic by about 10 kcal/mol. Assuming this general pathway
is followed, we cannot at this time distinguish secondary
photolysis of the oxathiin or a low-barrier thermal process. The
relative stability of these isomers is apparently reversed in the
case of the dithiin, and thus it is probably not surprising that a
limited number of simple dithiins have been observed at room
temperature, while no simple stable oxathiins are known.
The transformation of 6 into 7 is endergonic, according to
the CCSD(T) calculations, by a couple of kcal/mol, starting from
the lowest energy conformation of 6. The process will be
energy-neutral or slightly exergonic from the initial conformation
of 6 formed from 5. However, the final equilibrium favoring
formation of thiophene is almost certainly driven by an
unspecified, but exergonic desulfurization mechanism.
However, direct evidence for 5 is lacking to the best of our
knowledge. The closest analog of which we are aware is our own
assignment of the analogous structure during the photolysis of
dibenzoselenophene-Se-oxide, 11.[10] In addition to deoxygenat-
ing (12) on direct irradiation, 11 yields an isomeric compound
1
whose H NMR is indicative of an asymmetric compound with
two aromatic ABCD spin systems and whose mass is the same as
11. We assigned the structure 13 to that compound, in part due
to chemical deduction, but also in part because the presence
of O2 in the sample converted that compound (assigned to 13) to
another compound. The new product also had two aromatic
ABCD 1H spin systems, but had a mass 16 g/mol higher than that
of 11 and 13 and was assigned to the sultine analog 14.
The lack of observation of any derivative of 6 or 7 at low
temperature suggests (but does not prove) that the conversion of
derivatives of
6 to furans is photochemical. The bulky
Given the room-temperature existence of certain dithiins, we
believed that observation of the oxathiin might be possible,
though we failed in that attempt. However, it is possible that the
strategy was doomed to fail to begin with because of greater
absorption by 5 (which would prevent the accumulation of 5,
regardless of temperature). Dithiins are characterized by a low
substituents will surely change the relative energies of 6 and
7 a little bit,[60] so it is hard to say which of these two will be of
lower energy. Regardless, it should also be noted that
desulfurization must also take place at low temperature.[61] (It
should also be pointed out that the ‘‘-S’’ notation is commonly
used in this literature, but should not be taken to imply that
atomic sulfur is necessarily produced.)
Clearly, a more extensive computational investigation is
required to give convincing quantitative data on these reactions,
and we did not pursue the transition state between 5 and 6 at this
time. However, what can be concluded is that (1) the thermal
[58,59]
*
energy absorption in the visible, assigned to an n ! s band.
The lamp systems used by the Thiemann group and ourselves in
the great majority of experiments provide relatively broad
irradiation, which implies that a compound with an absorption to
the red of the starting material is likely to absorb light efficiently
J. Phys. Org. Chem. 2008, 21 915–924
Copyright ß 2008 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/poc

M. J. HEYING, M. NAG AND W. S. JENKS
barrier to ring opening will certainly be lower for the oxathiin
than for the dithiin; and (2) if the product of photochemical ring
opening of 5 is mainly determined by product stability, the
quantum yield of oxathiin ring opening should also be higher
than that of dithiin.
J ¼ 7.6 Hz), 7.06 (dd, 4H, J ¼ 7.2 Hz, 1.6 Hz), 6.63 (d, 2H, J ¼ 0.8 Hz).
13C: d 145.2, 131.4, 130.2, 128.7, 128.6, 127.6; MS (TOF, EI) m/z
236.07; 3,4-Dibenzylthiophene-S-oxide (1b): Typical yield 18%.
1H NMR (CDCl3, 300 MHz): d 7.34 (m, 4H), 7.15 (m, 6H), 6.06 (s, 2H),
3.63 (s, 4H). 13C: d 146.4, 135.4, 129.3, 129.1, 128.1, 127.8, 34.9; MS
(TOF, EI) m/z 264.10; 2,5-Bis(trimethylsilyl)thiophene-S-oxide[15]
1
(1c): Typical yield 45%. H NMR (CDCl3, 300 MHz): d 6.83 (s, 2H),
0.36 (s, 18H).
CONCLUSIONS
Although we have demonstrated a single thiophene-based
sulfoxide (1c) that does undergo deoxygenation with a higher
quantum yield than does DBTO, this compound would appear to
have limited utility as an O(3P) precursor due to its low chemical
yield and uncharacterizable byproducts.
Furans
3.4-Diphenylfuran (3a) was prepared as described by Fallis.[66]
3,4-Dibenzylfuran (3b) was prepared from 3,4- bis(tributylstan-
nyl)furan[67] by the method of Yang.[68]
More notable, though, in the sum of both this work and that of
others in this field is the unpredictability of the major isolable
product: the corresponding thiophene or the furan (except,
perhaps, in the instances with a methyl group in the 2-position[20]
due to the intervention of a third major process). With the
transformation of the sulfoxide to a S,C-sulfonium ylide, however,
we have seen no process other than the analog of deoxygena-
tion, that is, carbene formation.[38]
Thus, at this stage we are left to tentatively conclude that the
competition between furan formation – which we presume to
begin with the a-cleavage event – and deoxygenation compete
with one another based on closely spaced energy surfaces in the
relevant excited states. Until a more detailed understanding of
the subtleties of the effect of molecular structure on those
dynamics can be obtained, it is likely that the finding of furan or
thiophene formation from the photolysis of thiophene sulfoxides
will remain empirical.
Dibenzylthiophenium bismethoxycarbonylmethylide (8)
In a small vial, 3,4-dibenzylthiophene (204 mg, 0.77 mmol),
dimethyldiazomalonate (132 mg, 0.77 mmol) and a rhodium
catalyst [Rh(OAc)2]2 (2 mg, 0.0045 mmol) were mixed together.
The mixture was allowed to stir for 2 days in the dark, open to the
air, after which the color changed from an emerald green to a
brownish-teal. IR indicated that the diazo compound had been
completely consumed. The mixture was washed with hexane to
leave behind a dark green precipitate. Another washing with 75%
ethyl acetate and 25% hexane removed the residual catalyst,
leaving behind a white residue, which proved to be 3,4-
dibenzylthiophenium bismethoxycarbonylmethylide (136mg,
1
57.4% yield). H NMR (CDCl3, 400 MHz): d 7.32 (t J ¼ 7.2 Hz, 4H),
7.26 (t, J ¼ 7.2 Hz, 2H), 7.15 (d, J ¼ 7.2 Hz, 4H), 6.36 (s, 2H), 3.73 (s,
4H), 3.65 (s, 6H); 13C NMR: d 165.8, 149.6, 136.5, 129.1, 128.9, 127.4,
126.4, 51.3, 35.6; MS (EI) m/z 394.
EXPERIMENTAL
Photolyses
Photolyses were carried out by using spectro grade solvents with
initial concentration in the range of 1–5 mM and all solutions
were purged with Ar to remove O2, unless otherwise noted.
Dodecane was used as an internal standard for GC and dioxane
was used as an internal standard for reactions monitored by NMR.
Valerophenone was used as the actinometer.[69] The light source
was a 75 W Xe lamp coupled to a monochromator with a
cell-holder mounted at the exit. The slit widths allow a linear
dispersion of ꢂ12 nm from the stated wavelength, which was
adjusted for each compound to minimize absorption by the
products. Samples, held either in 1 cm quartz cells or NMR tubes,
depending on the method of analysis, were deoxygenated by
flushing with Ar. Stirring was provided either by a small magnetic
bar or by constant slow Ar bubbling. Analysis was conducted
preliminarily by NMR, and further by HPLC, LCMS, and GC,
including GCMS. GC work was limited by the fact that the
sulfoxides did not survive the chromatography. Quantification
was generally carried out by using HPLC. Mass spectra were
obtained to help confirm product identification, using LCMS.
Analyses and quantifications of alcohols were carried out by GC.
Control experiments showed that all compounds 2a–c and 3a–c
were photostable under the conditions used.
Materials
Unless otherwise noted, all solvents for photochemical exper-
iments were of the highest purity commercially available, and all
reagents were used as received.
Thiophenes
3,4-Diphenylthiophene (2a),[62] 3,4-Dibenzylthiophene (2b),[63,64]
and 2,5-bis(trimethylsilyl)thiophene (2c)[65] were prepared as
previously described.
Thiophene S-oxides: general procedure[15,29]
To
a solution of substituted thiophene (1 eq.) in dry
dichloromethane was added boron trifluoride-diethyl etherate
(4 eq.). The solution was stirred at ꢀ20 8C for 10 min. Then mCPBA
(1 eq.) in dry dichloromethane was added to the stirred solution
drop-wise. The solution was stirred at ꢀ20 8C under argon for
another 2 h. Saturated sodium bicarbonate solution was then
added to the reaction mixture and stirred for half an hour at room
temperature. The aqueous layer was extracted twice with
methylene chloride and the organic layer was washed with
water and brine. The solution was dried over magnesium sulfate
and concentrated in vacuo. The sulfoxide was isolated by
chromatography using 30% ethyl acetate in hexane.
Some preliminary and low-temperature experiments were
carried out by using the broad emission centered at 300 nm from
low-pressure fluorescent tubes in a Rayonet mini-reactor from
Southern New England Ultraviolet Company.
The photolysis at 77 K was carried out with an initial sulfoxide
concentration of 10.1 mM in a 1:9 mixture of perdeuterated
1
3,4-Diphenylthiophene-S-oxide (1a): Typical yield 25%. H NMR
(CDCl3, 400 MHz): d 7.39 (dt, 2H, J ¼ 7.6 Hz, 1.6 Hz), 7.29 (t, 4H,
www.interscience.wiley.com/journal/poc
Copyright ß 2008 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2008, 21 915–924

THIOPHENE-S-OXIDE PHOTOCHEMISTRY
ethanol:methanol solvent mixture. The solution was flushed with
Ar in an NMR tube and then plunged into liquid nitrogen in a
transparent quartz dewar. The light source was the broadly
emitting 300 nm tubes described previously. In separate
experiments, after 2–3 min (6 bulbs) or 40 min (8 bulbs), the
NMR tube was allowed to warm in the dark to ꢀ68 8C in an
acetone/dry ice bath before being lowered into a pre-cooled
NMR probe held at ꢀ50 8C. Spectra were obtained at ꢀ50 8C and
periodically as the sample was allowed to warm slowly to room
temperature. No significant change was observed in the
spectrum over this period of time, save for a small change in
the chemical shift of the thiophene protons (<0.1 ppm). The
short photolysis resulted in an approximate 5% conversion of 1a,
while the longer photolysis converted all of the material. After
withdrawal of the sample from the NMR instrument at room
temperature, a small quantity of solid material was observed in
the high conversion sample, as in the room temperature
photolyses in methanol.
[20] K. Arima, D. Ohira, M. Watanabe, A. Miura, S. Mataka, T. Thiemann, J. I.
Valcarcel, D. J. Walton, Photochem. Photobiol. Sci. 2005, 4, 808–
816.
[21] W. Schroth, R. Spitzner, C. Bruhn, Eur. J. Org. Chem. 1998, 2365–2371.
[22] E. Block, M. Birringer, R. DeOrazio, J. Fabian, R. S. Glass, C. Guo, C. He, E.
Lorance, Q. Qian, T. B. Schroeder, Z. Shan, M. Thiruvazhi, G. S. Wison, S.
Zhang, J. Am. Chem. Soc. 2000, 122, 5052–5064.
[23] E. Block, J. Page, J. P. Toscano, C.-X. Wang, X. Zhang, R. DeOrazio, C.
Guo, R. S. Sheridan, G. H. N. Towers, J. Am. Chem. Soc. 1996, 118,
4719–4720.
[24] P. J. Kropp, G. E. Fryxell, M. W. Tubergen, M. W. Hager, G. D. Harris Jr.,
T. P. McDermott Jr., R. Tornero-Velez, J. Am. Chem. Soc. 1991, 113,
7300–7310.
[25] J. W. Cubbage, T. A. Tetzlaff, H. Groundwater, R. D. McCulla, M. Nag,
W. S. Jenks, J. Org. Chem. 2001, 66, 8621–8628.
[26] T. Thiemann, D. Ohira, K. Arima, T. Sawada, S. Mataka, F. Marken, R. G.
Compton, S. Bull, S. G. Davies, J. Phys. Org. Chem. 2000, 13, 648–653.
[27] J. Nakayama, S. Yamaoka, M. Hoshino, Tetrahedron Lett. 1988, 29,
1161–1164.
[28] J. Nakayama, R. Hasemi, K. Yoshimura, Y. Sugihara, S. Yamaoka, J. Org.
Chem. 1998, 63, 4912–4924.
[29] J. Nakayama, Y. Sugihara, Sulfur Rep. 1997, 19, 349–375.
[30] J. Nakayama, T. Yu, Y. Sugihara, A. Ishii, Chem. Lett. 1997, 499–500.
[31] J. Nakayama, Y. Sano, Y. Sugihara, A. Ishii, Tetrahedron Lett. 1999, 40,
3785–3788.
Computational methods
[32] J. Nakayama, Sulfur Rep. 2000, 22, 123–149.
All computations were done by using the GAMESS suite of
programs[70] and all structures and orbitals were visualized by
using the MacMolPlt[71] application.
[33] J. Nakayama, T. Otani, Y. Sugihara, Y. Sano, A. Ishii, A. Sakamoto,
Heteroat. Chem. 2001, 12, 333–348.
[34] We did successfully prepare 1a by this route.
[35] M. G. Reinecke, P. Pedaja, In Halothiophenes and Their Reactions, (Ed.:
S. Gronowitz), Vol. 44, Part 2, Wiley, New York, 1985, 159–522.
[36] The quantum yield of sulfide formation is given by the product of the
yield of sulfide and the quantum yield for loss of the sulfoxide.
[37] T. J. Barton, G. P. Hussmann, J. Am. Chem. Soc. 1983, 105, 6316–6318.
[38] S. A. Stoffregen, M. Heying, W. S. Jenks, J. Am. Chem. Soc. 2007, 129,
15746–15747.
[39] M. W. Schmidt, K. K. Baldridge, J. A. Boatz, S. T. Elbert, M. S. Gordon,
J. H. Jensen, S. Koseki, N. Matsunaga, N. Nguyen, S. J. Su, T. L. Windus,
M. Dupuis, J. A. Montgomery, J. Comput. Chem. 1993, 14, 1347–1363.
[40] L. A. Curtiss, K. Raghavachari, P. C. Redfern, V. Rassolov, J. A. Pople,
J. Chem. Phys. 1998, 109, 7764–7776.
Acknowledgements
The authors gratefully acknowledge the National Science
Foundation (CHE 0211371) for partial support of this work.
[41] S. S. Xantheas, T. H. Dunning, Jr., J. Phys. Chem. 1993, 97, 18–19.
[42] A. J. Frank, F. Turecek, J. Phys. Chem. A 1999, 103, 5348–5361.
[43] B. S. Jursic, Theochem 1999, 467, 187–193.
REFERENCES
[1] G. M. Gurria, G. H. Posner, J. Org. Chem. 1973, 38, 2419–2420.
[2] J. R. Shelton, K. E. Davis, Int. J. Sulfur Chem. 1973, 8, 217–228.
[3] I. W. J. Still, In Photochemistry of Sulfoxides and Sulfones, (Eds.: S. Patai,
Z. Rappaport, C. J. M. Stirling,), John Wiley & Sons Ltd., New York,
1988, 873–887.
[4] W. S. Jenks, D. D. Gregory, Y. Guo, W. Lee, T. Tetzlaff, Mol. Supramol.
Photochem. 1997, 1, 1–56.
[5] D. D. Gregory, Z. Wan, W. S. Jenks, J. Am. Chem. Soc. 1997, 119,
94–102.
[6] Z. Wan, W. S. Jenks, J. Am. Chem. Soc. 1995, 117, 2667–2668.
[7] M. Nag, W. S. Jenks, J. Org. Chem. 2005, 70, 3458–3463.
[8] E. Lucien, A. Greer, J. Org. Chem. 2001, 66, 4576–4579.
[9] K. B. Thomas, A. Greer, J. Org. Chem. 2003, 68, 1886–1891.
[10] R. D. McCulla, W. S. Jenks, J. Am. Chem. Soc. 2004, 126, 16058–16065.
[11] M. Nag, W. S. Jenks, J. Org. Chem. 2004, 69, 8177–8182.
[12] W. S. Jenks, N. Matsunaga, M. Gordon, J. Org. Chem. 1996, 61,
1275–1283.
[44] C. W. Bauschlicher Jr., H. Partridge, Chem. Phys. Lett. 1995, 240,
533–540.
[45] R. D. Bell, A. K. Wilson, Chem. Phys. Lett. 2004, 394, 105–109.
[46] J. M. L. Martin, Chem. Phys. Lett. 1999, 310, 271–276.
[47] P. J. A. Ruttink, P. C. Burgers, M. A. Trikoupis, J. K. Terlouw, Chem. Phys.
Lett. 2001, 342, 447–451.
[48] O. N. Ventura, M. Kieninger, P. A. Denis, R. E. Cachau, Chem. Phys. Lett.
2002, 355, 207–213.
[49] A. K. Wilson, T. H. Dunning Jr., J. Chem. Phys. 2003, 119, 11712–11714.
[50] A. K. Wilson, T. H. Dunning Jr., J. Phys. Chem. A 2004, 108, 3129–3133.
[51] D. D. Gregory, W. S. Jenks, J. Org. Chem. 1998, 63, 3859–3865.
[52] D. D. Gregory, W. S. Jenks, J. Phys. Chem. A 2003, 107, 3414–3423.
[53] F. Jensen, T. Helgaker, J. Chem. Phys. 2004, 121, 3463–3470.
[54] P. Piecuch, S. A. Kucharski, K. Kowalski, M. Musial, Comp. Phys.
Commun. 2002, 149, 71–96.
[55] E. M. Rockafellow, D. McCulla Ryan, W. S. Jenks, J. Photochem.
Photobiol. A 2008, doi:10.1016/j.jphotochem.2008.02.014.
[56] W. S. Jenks, W. Lee, D. Shutters, J. Phys. Chem. 1994, 98, 2282–2289.
[57] W. Lee, W. S. Jenks, J. Org. Chem. 2001, 66, 474–480.
[58] R. S. Glass, J. R. Pollar, T. B. Schroeder, D. L. Lichtenberger, E. Block, R.
DeOrazio, C. Guo, M. Thiruvazhi, Phosphorus, Sulfur Silicon Relat. Elem.
1997, 120 440.
[13] S. A. Stoffregen, D. McCulla Ryan, R. Wilson, S. Cercone, J. Miller, W. S.
Jenks, J. Org. Chem. 2007, 72, 8235–8242.
[14] We also believed that we could carry out very reasonable quantum
chemical calculations on thiophene-S-oxide photochemical dis-
sociation, but not on DBTO. Those results will be published separately.
[15] N. Furukawa, S. Zhang, S. Sato, M. Higaki, Heterocycles 1997, 44,
61–66.
[16] P. Pouzet, I. Erdelmeier, D. Ginderow, J.-P. Mornon, P. Dansette, D.
Mansuy, Chem. Commun. 1995, 473–474.
[17] D. D. Gregory, Iowa State University, 1998.
[59] R. S. Glass, N. E. Gruhn, D. L. Lichtenberger, E. Lorance, J. R. Pollard, M.
Birringer, E. Block, R. DeOazio, C. He, Z. Shan, X. Zhang, J. Am. Chem.
Soc. 2000, 122, 5065–5074.
[60] Temperature also has a modest effect. The CCSD(T) enthalpy at O K has
6 an 7 at the same energy, for example.
[18] T. Thiemann, K. Arima, S. Mataka, Kyushu Daigaku Kino Busshitsu
Kagaku Kenkyusho Hokoku 2000, 14, 37–45.
[19] T. Thiemann, K. G. Dongol, J. Chem. Res. (S) 2002, 303–308.
[61] It is possible that the insoluble material in methanol may be S8, but we
were unable to determine that.
[62] J. Nakayama, R. Hasemi, J. Org. Chem. 1998, 63, 4912–4924.
J. Phys. Org. Chem. 2008, 21 915–924
Copyright ß 2008 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/poc

M. J. HEYING, M. NAG AND W. S. JENKS
[63] K. Tamao, S. Kodama, I. Nakajima, M. Kumada, A. Minato, K. Suzuki,
Tetrahedron 1982, 38, 3347–3354.
[64] J. P. Montheard, J. F. Delzant, M. Gazard, Synth. Commun. 1984, 14,
289–292.
[65] N. Furukawa, H. Hoshiai, T. Shibutani, M. Higaki, F. Iwasaki, H. Fujihara,
Heterocycles 1992, 34, 1085–1088.
[66] P. Forgione, P. D. Wilson, A. G. Fallis, Tetrahedron Lett. 2000, 41, 17–20.
[67] S. P. H. Mee, V. Lee, J. E. Baldwin, A. Cowley, Tetrahedron 2004, 60,
3695–3712.
[68] Y. Yang, H. N. C. Wong, Tetrahedron 1994, 50, 9583–9608.
[69] P. J. Wagner, P. A. Kelso, A. E. Kemppainen, J. M. McGrath, H. N. Schott,
R. G. Zepp, J. Am. Chem. Soc. 1972, 94, 7506–7512.
[70] M. W. Schmidt, K. K. Baldridge, J. A. Boatz, T. S. Elbert, M. S. Gordon,
J. H. Jensen, S. Koseki, S. Matsunaga, N. Nguyen, S. J. Su, T. L.
Windus, M. Dupuis, J. A. Montgomery, J. Comput. Chem. 1993, 14,
1347–1363.
[71] B. M. Bode, M. S. Gordon, J. Mol. Graphics. Mod. 1998, 16, 133–
138.
www.interscience.wiley.com/journal/poc
Copyright ß 2008 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2008, 21 915–924