Reaction of singlet oxygen with thioanisole in ionic liquids: A solvent induced mechanistic dichotomy

Article abstract of DOI:10.1021/ol900140w

A study of the reaction of thioanisole with singlet oxygen in pyrrolidinium-and imidazolium-based ionic liquids has been carried out. In these solvents, thioanisole shows a strongly enhanced reactivity with respect to molecular aprotic solvents, probably

Full text of DOI:10.1021/ol900140w

ORGANIC
LETTERS
2009
Vol. 11, No. 6
1413-1416
Reaction of Singlet Oxygen with
Thioanisole in Ionic Liquids: a Solvent
Induced Mechanistic Dichotomy
Enrico Baciocchi,*,† Cinzia Chiappe,*,‡ Tiziana Del Giacco,§ Chiara Fasciani,¶
Osvaldo Lanzalunga,¶,† Andrea Lapi,*,†,¶ and Bernardo Melai‡
Dipartimento di Chimica and Istituto CNR di Metodologie Chimiche-IMC, Sezione
Meccanismi di Reazione c/o Dipartimento di Chimica, Sapienza UniVersita` di Roma,
P.le A. Moro 5, 00185 Rome, Italy, Dipartimento di Chimica e Chimica Industriale,
UniVersita` di Pisa, Via Bonanno 33, 56126 Pisa, Italy, and Dipartimento di Chimica
and Centro di Eccellenza Materiali InnoVatiVi Nanostrutturati, UniVersita` di Perugia,
Via Elce di sotto 8, 06123 Perugia, Italy
andrea.lapi@uniroma1.it; enrico.baciocchi@uniroma1.it; cinziac@farm.unipi.it
Received January 29, 2009
ABSTRACT
A study of the reaction of thioanisole with singlet oxygen in pyrrolidinium- and imidazolium-based ionic liquids has been carried out. In these
solvents, thioanisole shows a strongly enhanced reactivity with respect to molecular aprotic solvents, probably due to a stabilization of the
persulfoxide intermediate in the ionic medium. Product isotope effects suggest a mechanistic change ongoing from pyrrolidinium to imidazolium
solvents.
Room-temperature ionic liquids (ILs) continue to deserve great
attention as solvents for a large variety of applications as a result
of the fact that the great number of combinations of anions and
cations make them flexible and tunable media for inorganic and
organic reactions, catalysis, and electrochemical studies.1 Al-
though in recent years the possible use of ILs in photochemical
reactions has been explored,2 very few studies concerning the
reactivity of singlet oxygen (1O2) in ILs have been carried out.3
To investigate O2 chemistry in ILs, we felt it worthwhile to
1
(1) (a) Ionic Liquids IIIB: Fundamental, Progress, Challenges, and
Opportunities- Transformation and Processes; Rogers, R. D., Seddon, K. R.,
Eds.; ACS Symposium Series 902; American Chemical Society: Washing-
ton, DC, 2005. (b) Ionic Liquids IIIA: Fundamental, Progress, Challenges,
and Opportunities- Properties and Structure; Rogers, R. D., Seddon, K. R.,
Eds.; ACS Symposium Series 901; American Chemical Society: Washing-
ton, DC, 2005. (c) Ionic Liquids in Synthesis, 2nd ed.; Wasserscheid, P.,
Welton, T., Eds.; Wiley-VCH: Weinheim, 2008. (d) Electrochemical Aspects
of Ionic Liquids; Ohno, H., Ed.; Wiley-Interscience: Hoboken, 2005. (e)
Zhao, H. Chem. Eng. Commun. 2006, 193, 1660.
† Istituto CNR di Metodologie Chimiche, IMC
‡ Universita` di Pisa
§ Universita` di Perugia
¶ Universita` di Roma
10.1021/ol900140w CCC: $40.75
Published on Web 02/19/2009
2009 American Chemical Society

focus our attention on its reactivity with sulfides, since this class
of reactions exhibits quite a complex mechanism where the
solvent exerts a major role by influencing the fate of the
persulfoxide intermediate (2).4 Thus, as shown in Scheme 1,
Figure 1. Stuctures of the ions forming the ionic liquids under
investigation.
Scheme 1. General Mechanisms for the Reaction of Singlet
Oxygen with Sulfides in Aprotic and Protic Solvents
over, very interestingly, our results also show that the solvent-
induced mechanistic dichotomy illustrated above also holds
in ILs.
In a typical reaction, a 1 mL oxygen-saturated solution of
1 (5 × 10-2 M) and methylene blue (5 × 10-4 M) in
anhydrous ionic liquid was irradiated for 1 h (400-600 nm)
at 20 °C. After extraction with diethyl ether, the mixture
was analyzed by GC, GC-MS, and HPLC (comparison with
authentic specimens). Under the same conditions, some
experiments were also carried out with diphenyl sulfide as
the substrate and, for comparison, in MeCN. Blank experi-
ments, carried out under the same reaction conditions but in
the absence of methylene blue or oxygen, showed the absence
of a significant amount of products. In all cases, the mass
balance was >95%. Moreover, laser flash photolysis experi-
ments allowed us to measure the 1O2 self-decay rate constant
in aprotic solvents (e.g., benzene, acetonitrile) the persulfoxide
intermediate (a nucleophilic species unreactive toward sulfides)
can undergo both intersystem crossing, to give O2 and the
3
1
(k0) as well as the O2 total quenching rate constant by 1
sulfide (path b), or when R-hydrogen atoms are present, a
rearrangement to a hydroperoxysulfonium ylide (3) via an
intramolecular hydrogen atom abstraction (path c). This second
intermediate can then react with the sulfide to give two
molecules of sulfoxide (path d) or rearrange to give sulfone
(path e). The key point in this process is represented by the
competition between the hydroperoxysulfonium ylide formation
(path c) and the unproductive intersystem crossing (path b). A
different situation holds instead in protic solvents or in the
presence of protic additives.5 In this case the intersystem
crossing process (path b) is significantly reduced by the
persulfoxide stabilization by hydrogen bond formation (path f).
The complex thus formed is then able to react with the sulfide
substrate to give two molecules of sulfoxide (path g) without
involving the formation of the hydroperoxysulfonium ylide.
We have carried out a detailed investigation of the singlet
oxygen promoted oxidation of thioanisole (1) in a series of
imidazolium- and pyrrolidinium-containing ILs6 (Figure 1)
and present evidence showing that the ILs can be very
convenient solvents for this reaction, in contrast with aprotic
solvents where thioanisole is practically unreactive. More-
and diphenyl sulfide (kT) in the solvents examined. All results
are reported in Table 1.
Table 1. Self Decay and Quenching Rate Constants of Singlet
Oxygen, Products and Yields in the Oxidation of Thioanisole
and Diphenyl Sulfide in ILs
yields (%)a
entry
solvent
k0 (s-1
)
kT (M-1s-1
)
SO
SO2
Thioanisole
1
2
3
4
5
6
7
8
[Bmpy][Tf2N]
[Empy][Tf2N]c
[Emim][Tf2N]
[Bmim][Tf2N]
[Emim][TfO]
MeCN
2.1 × 104
3.8 × 106
5.7 × 106
9.0 × 106
32
53
3.5
2.3
b
3.6 × 104
>98
15
9.8 × 104
b
12
b
b
2.4
b
2.0 × 104
4.7 × 106
b
[Bmpy][Tf2N]d
[Emim][Tf2N]d
26
>98
Diphenyl Sulfide
9
[Bmpy][Tf2N]
[Emim][Tf2N]
2.1 × 104
3.6 × 104
7.4 × 104
1.0 × 105
b
6
b
b
10
a Referred versus the initial amount of substrate and determined by GC
analysis with the exception of methyl phenyl sulfone in [Bmpy][Tf2N],
which was quantified by HPLC analysis because of peak overlapping due
to traces of IL present in the ethereal phase. b Not detected. c 10% (v/v)
d
MeCN. In the presence of benzoquinone (5 × 10-3 M). Benzoquinone
(2) (a) Alvaro, M.; Ferrer, B.; Garcia, H.; Narayana, M. Chem. Phys.
Lett. 2002, 362, 435. (b) Vieira, R. C.; Falvey, D. E. J. Phys. Chem. B
2007, 111, 5023. (c) Paul, A.; Samanta, A. J. Phys. Chem. B 2007, 111,
1957. (d) Vieira, R. C.; Falvey, D. E. J. Am. Chem. Soc. 2008, 130, 1552.
(3) (a) Swiderski, K.; McLean, A.; Gordon, C. M.; Vaughan, D. H.
Chem. Commun. 2004, 2178. (b) Gandra, N.; Frank, A. T.; Le Gendre, O.;
Sawwan, N.; Aebisher, D.; Liebman, J. F.; Houk, K. N.; Greer, A.; Gao,
R. Tetrahedron 2006, 62, 10771.
was recovered unreacted.
Looking first at the k0 and kT values reported in Table 1,
it appears clear that both of these constants are quite close
to those observed in MeCN as well as in other aprotic
molecular solvents7 and that, under our reaction conditions
(4) Clennan, E. L. Acc. Chem. Res. 2001, 34, 875.
(5) (a) Clennan, E. L.; Greer, A. J. Org. Chem. 1996, 61, 4793. (b)
Bonesi, S. M.; Albini, A. J. Org. Chem. 2000, 65, 4532.
1
([1] ) 5 × 10-2 M), in all cases most of the O2 generated
(6) Because [Empy][Tf2N] is solid at room temperature, 10% (v/v) dry
MeCN was added to obtain a liquid mixture.
is intercepted by 1. It can also be observed that, whereas
1414
Org. Lett., Vol. 11, No. 6, 2009

thioanisole is unreactive in MeCN (entry 6),8 the yields of
sulfoxide (4) in ILs range from appreciable to excellent, the
latter situation being observed in [Emim][Tf2N] (entry 3)
where the conversion into sulfoxide is practically quantitative.
Small amounts of sulfone (5) were formed in pyrrolidinium-
based ILs, whereas 4 is the exclusive reaction product in
imidazolium-based ILs (entries 3-5). Moreover, the ef-
ficiency of the reactions in both [Bmpy][Tf2N] and
[Emim][Tf2N] was not suppressed by the presence of
yield.11 However, when the reaction of 1 in [Bmpy][Tf2N]
was carried out in the presence of Ph2S the additive was
cooxidized to give Ph2SO.12
Reactivity. Even though the simple comparison of the
yields does not strictly represent a quantitative comparison
of the singlet oxygen chemical reactivity toward thioanisole
1
in the different solvents given the differences in the O2
lifetime as well as in the rate constant of total quenching by
thioanisole, there is no question that all ILs examined exhibit
a very strong enhancing effect on the chemical reactivity of
1O2 toward thioanisole when compared with MeCN. A
reasonable rationalization is that in ILs the thioanisole
persulfoxide intermediate may be better stabilized than in
aprotic solvents by a more efficient coordination exerted by
the ionic solvent, which may interact with the negatively
charged peroxy appendage of the persulfoxide as well as with
the positively charged sulfur atom. Thus, with respect to
molecular aprotic solvents, in ILs the unproductive inter-
system crossing step (path b in Scheme 1) should be
significantly slowed down relative to the chemical path.13,14
On the other hand, the good coordinating power of ILs is
clearly indicated by their ENT values which are higher than
those for polar aprotic solvents.15 Moreover, it should also
be considered that imidazolium-type solvents have a rela-
tively acidic C-H bond,16 thus in these solvents the
coordination might also assume the character of a hydrogen
bond (see below).
-•
benzoquinone, a very efficient O2 trapping agent (entries
7 and 8),9 thus allowing us to exclude the occurrence of a
photooxygenation mechanism not involving singlet oxygen
but an initial electron transfer oxidation of 1 by the excited
sensitizer, followed by the reaction of 1+• with O2-• formed
by reaction of the reduced sensitizer with O2.10
To obtain mechanistic information, we have also inves-
tigated the reaction of an equimolar mixture of PhSCH3 and
PhSCD3 (1-d3). The results, displayed in Table 2, show that
the pyrrolidinium ILs exhibit a significant product isotope
effect (ranging from 1.2 and 1.3) on sulfoxide formation (i.e.,
product ratio 4/4-d3). Such an isotope effect is instead absent
in imidazolium ILs where 1 and 1-d3 are oxidized at the same
rate, the 4/4-d3 product ratio being unity within experimental
error (Table 2). The same result was obtained for the reaction
in MeOH that was studied for comparison.
Interestingly, a dramatic enhanced reactivity was also
observed by Clennan and co-workers in the reaction of
aromatic sulfides with 1O2 carried out in zeolites.17 In this
case too, the authors attributed such a behavior to the
stabilization of the persulfoxide exerted by the zeolite
countercation. This similarity might lend support to the
suggestion that zeolites may be in some way considered the
solid counterparts of ionic liquids.18
Table 2. Product Isotope Effect in the Singlet Oxygen-Promoted
Oxidation of Thioanisole in ILs
a
a
solvent
4/4-d3
solvent
4/4-d3
[Bmpy][Tf2N]
[Empy][Tf2N]b
[Emim][Tf2N]
1.33(8)
1.25(3)
1.02(2)
[Bmim][Tf2N]
[Emim][TfO]
MeOH
1.04(1)
0.98(1)
1.03(1)
a 4/4-d3 molar ratio determined in the oxidation of an equimolar mixture
of 1 and 1-d3 under the same reaction conditions described above but keeping
the substrate conversion under 15%. The values were determined by GC-
MS analysis (single ion monitoring mode) by the ratio of the intensity of
molecular peaks m/z ) 140 and 143. The values are the average of at least
three independent determinations. The error (standard deviation) in the last
significant digit is given in parentheses. b 10% (v/v) MeCN.
Reaction Mechanisms. The different behaviors observed
in pyrrolidinium and imidazolium ILs (product isotope effect
and reactivity of Ph2S) suggest that in these two classes of
ILs the reaction should follow two different mechanisms.
In pyrrolidinium ILs, the product isotope effect observed on
sulfoxide formation (Table 2) clearly indicates that an
(11) This low sulfoxide yield (with respect to that of 1) is somewhat
expected from the k0 and kT values reported in Table 1. Accordingly, under
our conditions (initial sulfide concentration of 5 × 10-2 M) in [Emim][Tf2N],
Another important difference between pyrrolidinium and
imidazolium ILs was found with respect to the oxidation of
Ph2S. Whereas Ph2S was unreactive in [Bmpy][Tf2N] as
observed in molecular aprotic solvents (it lacks an R-C-H
bond), it exhibited an appreciable reactivity in [Emim][Tf2N],
affording phenyl sulfoxide as the exclusive product in 6%
1
89% of O2 is intercepted by thioanisole, whereas only 12% is by Ph2S.
(12) In [Bmpy][Tf2N], the reaction of 1 (0.05 M) in the presence of
Ph2S (0.25 M) under usual reaction conditions afforded methyl phenyl
sulfoxide (11 mmol) and diphenyl sulfoxide (8.5 mmol).
(13) As kT does not change very much in ILs with respect to molecular
solvents, the strong increase in reactivity observed in the former solvents
should be mainly associated with a decrease in the intersystem crossing
rate.
(14) Preliminary experiments indicate that the rate enhancement effect
observed with thioanisole is not present with dibutyl sulfide. A possibility
is that the persulfoxide of dialkyl sulfide is stable enough also in aprotic
solvents (lone pair on sulfur not delocalized as in an aromatic sulfide) that
the larger coordinating effect of ILs may be practically not felt. However,
it might also be that in the dibutyl persulfoxide the relatively long alkyl
chains hamper the coordination by the IL solvent. In this respect, it may be
noted that steric effects might rationalize the finding that the reactivity of
thioanisole in [Bmim][Tf2N] is much lower than in [Emim][Tf2N] (Table
1). Clearly, further detailed study is necessary to understand how the
structure of the sulfide influence the IL coordination in the persulfoxide.
(7) In benzene: k0 ) 3.3 × 104 s-1, kT ) 8.0 × 105 M-1s-1. In methanol:
k0 ) 1.0 × 105 s-1, kT ) 2.0 × 106 M-1s-1. Wilkinson, F.; Helman, W. P.;
Ross, A. B. J. Phys. Chem. Ref. Data 1995, 24, 663.
(8) This observation is not surprising since Sawaki and co-workers
observed that the conversion of 1 into products was 0.48% after 8 h
irradiation in benzene and 0.8% after 2 h irradiation in MeCN. Ishiguro,
K.; Hayashi, M.; Sawaki, Y. J. Am. Chem. Soc. 1996, 118, 7265.
(9) Manring, L. E.; Kramer, M. K.; Foote, C. S. Tetrahedron Lett. 1984,
25, 2523.
(10) Baciocchi, E.; Del Giacco, T.; Elisei, F.; Gerini, M. F.; Guerra,
M.; Lapi, A.; Liberali, P. J. Am. Chem. Soc. 2003, 125, 16444.
Org. Lett., Vol. 11, No. 6, 2009
1415

R-hydrogen atom abstraction step should take place during
the oxidation process, as predicted by the mechanism
proposed for the reaction in aprotic solvents in which an
R-hydrogen is removed during the product determining step
to form the hydroperoxysulfonium ylide and ultimately the
sulfoxide product (Scheme 1, path c).19 Moreover, it is
noteworthy that the KIE values observed in our experiments
are very close to those observed by Clennan in the reaction
of 1O2 with 2,2,8,8-tetradeuterio-1,5-dithiacyclooctane
(1.21)19a or with a series of 1,3-dithianes (1.31-1.62).19b
In both cases, the observation of a product isotope effect
was considered strong support for the mechanism via
hydroperoxysulfonium ylide (Scheme 1, path c). Additional
important evidence in this respect is that whereas diphenyl
sulfide is not directly oxidized by singlet oxygen in pyrro-
lidinium solvents (which is expected because, as stated
before, it lacks R-hydrogens and cannot form the ylide), its
oxidation to diphenyl sulfoxide takes place when it is reacted
with singlet oxygen in the presence of 1. In this case it is
the ylide formed from thioanisole persulfoxide that can
oxidize the diphenyl sulfide as it oxidizes the thioanisole
itself.20 In conclusion, both the presence of a product isotope
effect and the co-oxidation of Ph2S strongly support the
consistent with our finding that the sulfoxidation of thioani-
sole promoted by O2 in MeOH (Table 2) does not exhibit
1
a product isotope effect as well as with the observation that
Ph2S can be sulfoxidated in this solvent.22
The reason for the mechanistic change ongoing from
pyrrolidinium to imidazolium solvents is probably related
to the fact that, as mentioned before, imidazolium solvents
have an acidic C-H bond and may be able to coordinate
the persulfoxide oxygen by hydrogen bonds. Accordingly,
in the scale of hydrogen bond acidity (Kamlet-Taft R
parameters) the imidazolium ILs show an R value moderately
high (0.659 and 0.625, respectively, for [Emim][Tf2N] and
[Bmim][Tf2N])23,24 comparable to that of tert-butanol (R )
0.68).25 If this interpretation is correct, it may be that, as
occurs in protic solvents, the nucleophilicity of persulfoxide
is significantly reduced,26 thus allowing it to react with the
sulfide substrate to give the sulfoxide (Scheme 1, path g)
but inhibiting its reaction with a molecule of sulfoxide to
give the sulfone,27 as actually observed in imidazolium ILs
where sulfone formation was not detected.
In conclusion, we have found that ionic liquids are good
solvents for the sulfoxygenation of thioanisole by singlet
oxygen. Particularly, the reaction in [Emim][Tf2N] may have
synthetic interest as it provides complete conversion into
sulfoxide without any overoxidation to sulfone. Probably,
this is due to a stabilizing effect by the IL on the persulfoxide
intermediate, which relatively slows down the intersystem
crossing pathway with respect to chemical reactions. More-
over, through the study of product isotope effect and the
reaction of diphenyl sulfide, it has been discovered that in
ILs there is the same mechanistic dichotomy observed in
molecular solvents. Accordingly, in pyrrolidinium-type sol-
vents the persulfoxide is converted to a hydroperoxysulfo-
nium ylide intermediate as it occurs in conventional aprotic
solvents. With imidazolium-type solvents, which possess an
acidic C-H bond, a hydrogen-bonded persulfoxide is prob-
ably formed and the reaction proceeds through this species
without involving the intermediacy of the hydroperoxysul-
fonium ylide intermediate.
1
hypothesis that the reaction of O2 with 1 in pyrrolidinium
ILs follows the mechanism observed in aprotic solvents
(Scheme 1, path c).21
Different behaviors are instead observed with imidazolium
ionic liquids. In this case the product isotope effects were
unity within experimental error (Table 2), thus showing that
the mechanism involving hydroperoxysulfonium ylide for-
mation is not operating in these solvents. Moreover, in the
imidazolium solvent [Emim][Tf2N] diphenyl sulfide is
oxidized to diphenyl sulfoxide even in the absence of
thioanisole. These results thus suggest that the mechanism
of the singlet-oxygen-promoted sulfoxygenation of thioani-
sole is very likely the same as that observed in protic solvents
(path f in Scheme 1). This mechanistic hypothesis is
(15) (a) The ENT values for [Emim][Tf2N] (0.644) and [Bmpy][Tf2N]
(0.544)15b are in the range associated to those of aliphatic alcohols (0.654
and 0.546 for ethanol and 2-propanol, respectively)15c and higher than that
of MeCN (0.460).15c. (b) Crowhurst, L.; Mawdsley, P. R.; Perez-Arlandis,
L. M.; Salter, P. A.; Welton, T. Phys. Chem. Chem. Phys. 2003, 5, 2790.
(c) Reichardt, C. Chem. ReV. 1994, 94, 2319.
Acknowledgment. MIUR, La Sapienza University of
Rome, University of Pisa and University of Perugia are
thanked for the financial support.
(16) (a) Tait, S.; Osteryoung, R. A. Inorg. Chem. 1984, 23, 4352. (b)
Avent, A. G.; Chaloner, P. A.; Day, M. P.; Seddon, K. R.; Welton, T.
J. Chem. Soc., Dalton Trans. 1994, 3405.
Supporting Information Available: Experimental details,
isotope effects, and product analyses. This material is
available free of charge via the Internet at http://pubs.acs.org.
(17) Clennan, E. L.; Zhou, W.; Chan, J. J. Org. Chem. 2002, 67, 9368.
(18) Marquis, S.; Ferrer, B.; Alvaro, M.; Garcia, H.; Roth, D. H. J. Phys.
Chem. B 2006, 110, 14956.
OL900140W
(19) (a) Clennan, E. L.; Liao, C. Tetrahedron 2006, 62, 10724. (b)
Toutchkine, A.; Clennan, E. L. J. Org. Chem. 1999, 64, 5620.
(20) (a) Liang, J.-J.; Gu, C.-L.; Kacher, M. L.; Foote, C. S. J. Am. Chem.
Soc. 1983, 105, 4717. (b) Gu, C.-L.; Foote, C. S.; Kacher, M. L. J. Am.
Chem. Soc. 1981, 103, 5949.
(22) Bonesi, S. M.; Fagnoni, M.; Monti, S.; Albini, A. Photochem.
Photobiol. Sci. 2004, 3, 489.
(23) Tokuda, H.; Tsuzuki, S.; Susan, M. A. B. H.; Hayamizu, K.;
(21) Rearrangement of the ylide might be taken as the origin of the
small amounts of sulfone observed in the pyrrolidium solvents (Scheme
1, path e). However, we have observed that the formation of sulfone is
subject to the same products KIE observed with the sulfoxide (Supporting
Information). This observation is in line with a sulfone generation from
partial oxidation of the formed sulfoxide, presumably by the persulfox-
ide,20 whereas a higher isotope effect should have been expected in the
sulfone formation via ylide rearrangement (see Supporting Information for
details).
Watanabe, M. J. Phys. Chem. B 2006, 110, 19593.
(24) Moreover, these values are higher than those of pyrrolidinium ILs
(a ) 0.427 for [Bmpy][Tf2N]).15b
(25) Kamlet, M. J.; Addoud, J. L.; Abraham, M. H.; Taft, R. W. J. Org.
Chem. 1983, 48, 2877.
(26) Sawaki, Y.; Ogata, Y. J. Am. Chem. Soc. 1981, 103, 5947.
(27) (a) Gu, C.-L.; Foote, C. S. J. Am. Chem. Soc. 1982, 104, 6060. (b)
Cauzzo, G.; Gennari, G.; Fabrizio, D. R. Gazz. Chim. Ital. 1979, 109, 541.
(c) Foote, C. S.; Peters, J. W. J. Am. Chem. Soc. 1971, 93, 3795.
1416
Org. Lett., Vol. 11, No. 6, 2009