Synthesis and characterization of water-soluble phenylene-vinylene-based singlet oxygen sensitizers for two-photon excitation

Article abstract of DOI:10.1021/jo050507y

The synthesis and characterization of water-soluble singlet oxygen sensitizers with a phenylene-vinylene motif is presented. The principal motivation for this study was to better understand specific features of a water-soluble molecule that influence the photosensitized production of singlet oxygen upon nonlinear, two-photon excitation of that molecule. To achieve water solubility, sensitizers were synthesized with ionic as well as nonionic substituents. In the ionic approach, salts of N-methylated pyridine, benzothiazole, and 1-methyl-piperazine moieties were used, as were aryl-substituted sulfonic acid moieties. In the nonionic approach, aryl-substituted triethylene glycol moieties were used. Selected photophysical properties of the compounds synthesized were determined, including singlet oxygen quantum yields. Of the molecules examined, the most efficient singlet oxygen sensitizers had triethylene glycol units as the functional group that imparted water solubility. Molecules containing the ionic moieties did not make singlet oxygen in appreciable yield nor did they efficiently fluoresce. Rather, for these latter molecules, rapid charge-transfer-mediated nonradiative processes appear to dominate excited state deactivation.

Full text of DOI:10.1021/jo050507y

Synthesis and Characterization of Water-Soluble
Phenylene-Vinylene-Based Singlet Oxygen Sensitizers for
Two-Photon Excitation
Christian B. Nielsen,†,‡ Mette Johnsen,‡ Jacob Arnbjerg,‡ Michael Pittelkow,§
Sean P. McIlroy,‡ Peter R. Ogilby,*,‡ and Mikkel Jørgensen*,†
Polymer Department, Risø National Laboratory, Frederiksborgvej 399, DK-4000 Roskilde, Denmark,
Department of Chemistry, University of Aarhus, Langelandsgade 140, DK-8000, Århus C, Denmark, and
Department of Chemistry, University of Copenhagen, Universitetsparken 5, DK-2100,
Copenhagen Ø, Denmark
progilby@chem.au.dk
Received March 14, 2005
The synthesis and characterization of water-soluble singlet oxygen sensitizers with a phenylene-
vinylene motif is presented. The principal motivation for this study was to better understand specific
features of a water-soluble molecule that influence the photosensitized production of singlet oxygen
upon nonlinear, two-photon excitation of that molecule. To achieve water solubility, sensitizers
were synthesized with ionic as well as nonionic substituents. In the ionic approach, salts of
N-methylated pyridine, benzothiazole, and 1-methyl-piperazine moieties were used, as were aryl-
substituted sulfonic acid moieties. In the nonionic approach, aryl-substituted triethylene glycol
moieties were used. Selected photophysical properties of the compounds synthesized were
determined, including singlet oxygen quantum yields. Of the molecules examined, the most efficient
singlet oxygen sensitizers had triethylene glycol units as the functional group that imparted water
solubility. Molecules containing the ionic moieties did not make singlet oxygen in appreciable yield
nor did they efficiently fluoresce. Rather, for these latter molecules, rapid charge-transfer-mediated
nonradiative processes appear to dominate excited state deactivation.
Introduction
a given electronic excited state is achieved via the
simultaneous absorption of two photons of comparatively
low energy. As such, the two-photon experiments are
performed by irradiating the sample at wavelengths
where dominant one-photon transitions are generally
absent. This means that one can achieve better depth
penetration in a sample. Moreover, because the prob-
ability of absorbing two photons increases quadratically
Nonlinear, two-photon excitation of organic molecules
has received a great deal of recent attention as an
alternative to linear, one-photon excitation due to prom-
ising spatially resolved applications in 3D microstruc-
turing, imaging, and photodynamic therapy.1-9 In the
two-photon absorption (TPA) process, the production of
† Risø National Laboratory.
(4) Iwase, Y.; Kamada, K.; Ohta, K.; Kondo, K. J. Mater. Chem.
2003, 13, 1575-1581.
‡ University of Aarhus.
§ University of Copenhagen.
(5) Pyun, O. K.; Yang, W.; Jeong, M.-Y.; Lee, S. H.; Kang, K. M.;
Jeon, S.-J.; Cho, B. R. Tetrahedron Lett. 2003, 44, 5179-5182.
(6) Strehmel, B.; Sarker, A. M.; Detert, H. ChemPhysChem 2003,
4, 249-259.
(1) Albota, M.; Beljonne, D.; Bre´das, J.-L.; Ehrlich, J. E.; Fu, J.-Y.;
Heikal, A. A.; Hess, S. E.; Kogej, T.; Levin, M. D.; Marder, S. R.;
McCord-Maughon, D.; Perry, J. W.; Ro¨ckel, H.; Rumi, M.; Subrama-
niam, G.; Webb, W. W.; Wu, X.-L.; Xu, C. Science 1998, 281, 1653-
1656.
(7) Kawata, S.; Sun, H.-B.; Tanaka, T.; Takada, K. Nature 2001,
412, 697-698.
(2) Denk, W.; Strickler, J. H.; Webb, W. W. Science 1990, 248, 73-
76.
(3) Porres, L.; Mongin, O.; Katan, C.; Charlot, M.; Pons, T.; Mertz,
J.; Blanchard-Desce, M. Org. Lett. 2004, 6, 47-50.
(8) Reinhardt, B. A.; Brott, L. L.; Clarson, S. J.; Dillard, A. G.; Bhatt,
J. C.; Kannan, R.; Yuan, L.; He, G. S.; Prasad, P. N. Chem. Mater.
1998, 10, 1863-1874.
(9) King, B. A.; Oh, D. H. Photochem. Photobiol. 2004, 80, 1-6.
10.1021/jo050507y CCC: $30.25 © 2005 American Chemical Society
Published on Web 08/11/2005
J. Org. Chem. 2005, 70, 7065-7079
7065

Nielsen et al.
with the photon flux, excited state production can be
limited to a small volume defined by a focused laser
where a sufficiently high fluence is obtained.
charge transfer (CT) upon excitation (i.e., large transition
moments are desirable).1,20,23-25 As such, the judicious
introduction of electron donor and acceptor groups in the
molecule can significantly enhance the two-photon ab-
sorption probability. Longer conjugation lengths that
result in larger charge separation in the molecule are also
beneficial for obtaining large TPA cross sections.1,23,26
It is acknowledged that singlet molecular oxygen (a1∆g)
is an important reactive intermediate in many photo-
induced oxidations,10,11 including those involved in pho-
todynamic therapy.12 We have embarked on an extensive
program to develop methods by which singlet oxygen can
be generated in, and optically detected from, small
spatially resolved domains in heterogeneous samples.13-18
To this end, the two-photon photosensitized production
of singlet oxygen and the associated development of the
appropriate sensitizer molecules are important issues.19-22
A substantial amount of work has been done in the
recent past on the development of molecules that have
comparatively large two-photon absorption probabilities
(i.e., the so-called TPA cross section).1,8 Most of these
molecules, however, have been designed for fluorescence
experiments and are not suitable for use as singlet
oxygen sensitizers. Molecules used as singlet oxygen
sensitizers generally undergo efficient intersystem cross-
ing to produce the lowest triplet state, which in turn must
efficiently transfer its excitation energy to ground state
oxygen to yield singlet oxygen. As such, one design
strategy to make singlet oxygen sensitizers is to introduce
functional groups into the chromophore that facilitate
intersystem crossing (e.g., carbonyl groups or heavy
atoms such as bromine). Other features of a good singlet
oxygen sensitizer may include stability upon prolonged
irradiation and the lack of appreciable luminescence in
the near-infrared that could interfere with the optical
detection of the weak singlet oxygen phosphorescence
signal at 1270 nm.
We have shown that some OPV molecules can also
generate singlet oxygen.19-22 It has also been shown,
however, that the extent of CT character, both within a
given molecule and in a given molecule-oxygen complex,
can have a large effect on the efficiency with which
singlet oxygen is generated.27-30 In general, CT character
in a sensitizer-oxygen complex facilitates nonradiative
deactivation of the excited state at the expense of energy
transfer from the sensitizer to produce singlet oxygen.
As such, and in the present context of developing two-
photon sensitizers, CT character in a molecule is not
desired (in this argument, we assume that the extent of
intermolecular CT character in the sensitizer-oxygen
complex will be larger for a sensitizer with a proclivity
for intramolecular charge transfer). The extent to which
this CT effect is manifested also depends significantly
on the polarity of the solvent in which the system is
dissolved.22,28,29,31 Generally, CT character is enhanced in
more polar solvents such as water or acetonitrile, which,
in turn, results in comparatively low singlet oxygen
yields. It is with these points in mind that we have
investigated the extent to which singlet oxygen can be
produced in a two-photon photosensitized process in
aqueous systems.
We have demonstrated that the two-photon photosen-
sitized production of singlet oxygen in water can be a
complicated problem.22 Specifically, functional groups on
the sensitizer that impart water solubility and that give
rise to larger TPA cross sections are, in many cases, not
conducive to the production of singlet oxygen in high
yield. In part, this issue involves the competing influence
of charge transfer, not just in the sensitizer itself but in
the sensitizer-oxygen complex as well. As already men-
tioned, molecular features that increase CT character in
the chromophore facilitate two-photon absorption, whereas
these same features contribute to a low singlet oxygen
yield.
A host of oligo phenylene vinylenes (OPVs) with
comparatively large TPA cross sections have recently
been prepared.1,23 It has been ascertained that, with these
compounds, the two-photon transition probability in-
creases with an increase in the degree of intramolecular
(10) Foote, C. S. Acc. Chem. Res. 1968, 1, 104-110.
(11) Singlet Oxygen; Frimer, A. A., Ed.; CRC Press: Boca Raton,
FL, 1985; Vols. I-IV.
(12) Dougherty, T. J.; Gomer, C. J.; Henderson, B. W.; Jori, G.;
Kessel, D.; Korbelik, M.; Moan, J.; Peng, Q. J. Natl. Cancer Inst. 1998,
90, 889-905.
(13) Skovsen, E.; Snyder, J. W.; Lambert, J. D. C.; Ogilby, P. R. J.
Phys. Chem. B 2005, 109, 8570-8573.
(14) Snyder, J. W.; Gao, Z.; Ogilby, P. R. Rev. Sci. Instrum. 2005,
76, 013701.
In this present article, we continue our study of issues
pertinent to the development of water-soluble, two-
photon singlet oxygen sensitizers.22 In particular, we now
focus on the synthesis and characterization of some
phenylene-vinylene-based molecules.
(15) Snyder, J. W.; Zebger, I.; Gao, Z.; Poulsen, L.; Frederiksen, P.
K.; Skovsen, E.; McIlroy, S. P.; Klinger, M.; Andersen, L. K.; Ogilby,
P. R. Acc. Chem. Res. 2004, 37, 894-901.
(16) Zebger, I.; Snyder, J. W.; Andersen, L. K.; Poulsen, L.; Gao, Z.;
Lambert, J. D. C.; Kristiansen, U.; Ogilby, P. R. Photochem. Photobiol.
2004, 79, 319-322.
(17) Zebger, I.; Poulsen, L.; Gao, Z.; Andersen, L. K.; Ogilby, P. R.
Langmuir 2003, 19, 8927-8933.
(18) Andersen, L. K.; Gao, Z.; Ogilby, P. R.; Poulsen, L.; Zebger, I.
J. Phys. Chem. A 2002, 106, 8488-8490.
(24) Wang, C.-K.; Macak, P.; Luo, Y.; Ågren, H. J. Chem. Phys. 2001,
114, 9813-9820.
(19) Frederiksen, P. K.; Jørgensen, M.; Ogilby, P. R. J. Am. Chem.
Soc. 2001, 123, 1215-1221.
(25) Norman, P.; Luo, Y.; Ågren, H. J. Chem. Phys. 1999, 111, 7758-
7765.
(20) Poulsen, T. D.; Frederiksen, P. K.; Jørgensen, M.; Mikkelsen,
K. V.; Ogilby, P. R. J. Phys. Chem. A 2001, 105, 11488-11495.
(21) McIlroy, S. P.; Clo´, E.; Nikolajsen, L.; Frederiksen, P. K.;
Nielsen, C. B.; Mikkelsen, K. V.; Gothelf, K. V.; Ogilby, P. R. J. Org.
Chem. 2005, 70, 1134-1146.
(26) Drobizhev, M.; Karotki, A.; Rebane, A.; Spangler, C. W. Opt.
Lett. 2001, 26, 1081-1083.
(27) Schweitzer, C.; Schmidt, R. Chem. Rev. 2003, 103, 1685-1757.
(28) Kristiansen, M.; Scurlock, R. D.; Iu, K.-K.; Ogilby, P. R. J. Phys.
Chem. 1991, 95, 5190-5197.
(22) Frederiksen, P. K.; McIlroy, S. P.; Nielsen, C. B.; Nikolajsen,
L.; Skovsen, E.; Jørgensen, M.; Mikkelsen, K. V.; Ogilby, P. R. J. Am.
Chem. Soc. 2005, 127, 255-269.
(23) Pond, S. J. K.; Rumi, M.; Levin, M. D.; Parker, T. C.; Beljonne,
D.; Day, M. W.; Bre´das, J.-L.; Marder, S. R.; Perry, J. W. J. Phys. Chem.
A 2002, 106, 11470-11480.
(29) McGarvey, D. J.; Szekeres, P. G.; Wilkinson, F. Chem. Phys.
Lett. 1992, 199, 314-319.
(30) Schmidt, R.; Shafii, F.; Schweitzer, C.; Abdel-Shafi, A. A.;
Wilkinson, F. J. Phys. Chem. A 2001, 105, 1811-1817.
(31) Abdel-Shafi, A. A.; Wilkinson, F. Phys. Chem. Chem. Phys. 2002,
4, 248-254.
7066 J. Org. Chem., Vol. 70, No. 18, 2005

Water-Soluble Singlet Oxygen Sensitizers
SCHEME 1. Synthesis of the MTEG-Substituted OPV Sensitizers
CHART 1. General Structure of the Sensitizers
Examined in the Present Worka
SCHEME 2. Preparation of
3,4,5-Triacetoxy-benzaldehyde
a
The letters A and D refer to electron acceptors and donors,
respectively, and illustrate methods by which substituents can be
used to influence the electronic character of the chromophore.
the 1- and 4-positions. In this study, the functional groups
used to impart solubility in water are attached to the
chromophoric π system. As such, their electron-donating
or -accepting properties also influence the extent to which
CT character plays a role in the photophysics of the
molecule. The substituents employed include nonionic
monomethyl-ether triethylene glycol moieties (MTEGs)
and ionic N-methylated species. Bromine atoms were also
incorporated into most of the molecules studied to
facilitate intersystem crossing to the triplet state.
Synthesis. For the nonionic sensitizers, MTEG moi-
eties were used to impart water solubility. The synthesis
of these molecules is outlined in Scheme 1. Briefly,
Horner-Wadsworth-Emmons (HWE) reactions were
used to couple the core aromatic ring of the OPV to the
terminal, MTEG-substituted aryl moieties. The MTEG-
substituted benzaldehydes were prepared from the hy-
droxy-substituted benzaldehydes in an SN2 reaction with
the tosylated glycol.
Results and Discussion
All the molecules examined in the present work were
prepared following design principles that have evolved
in the development of OPVs with large TPA cross sections
(Chart 1).1,6,23,32-34
The molecules are centrosymmetric and are built
around a core aromatic ring with vinyl substituents in
(32) Beljonne, D.; Wenseleers, W.; Zojer, E.; Shuai, Z.; Vogel, H.;
Pond, S. J. K.; Perry, J. W.; Marder, S. R.; Bre´das, J.-L. Adv. Funct.
Mater. 2002, 12, 631-641.
(33) Rumi, M.; Ehrlich, J. E.; Heikal, A. A.; Perry, J. W.; Barlow,
S.; Hu, Z.; McCord-Maughon, D.; Parker, T. C.; Ro¨ckel, H.;
Thayumanavan, S.; Marder, S. R.; Beljonne, D.; Bre´das, J.-L. J. Am.
Chem. Soc. 2000, 122, 9500-9510.
(34) Zojer, E.; Beljonne, D.; Kogej, T.; Vogel, H.; Marder, S. R.; Perry,
J. W.; Bre´das, J. L. J. Chem. Phys. 2002, 116, 3646-3658.
J. Org. Chem, Vol. 70, No. 18, 2005 7067

Nielsen et al.
SCHEME 3. Synthesis of a MTEG-Substituted OPV, Where the MTEG Groups Are Attached to an
Amino-ethanol
SCHEME 4. Synthesis of the Sulfonic Acid End-Capped OPV Using a HWE Strategy
All of the MTEG-substituted benzaldehydes were made
from commercially available hydroxy-substituted ben-
zaldehydes, except 3,4,5-trihydroxy-benzaldehyde. To the
best of our knowledge, only two methods have been
reported for preparing this compound: (1) the seminal
work of Rosenmund and co-workers on gallous acid,35,36
and (2) a newer route from the acetyl-protected acid that
proceeds by reduction to the alcohol with diborane,
followed by reoxidation to the aldehyde with pyridinium
dichromate.37 In our work, an alternative to the Rosen-
mund reduction was attempted using (Ph3P)2CuBH4 as
the reducing agent. But this proceeded in only moderate
yield (56%).
biologically active compounds contain a trihydroxy-func-
tionalized aromatic ring.40
As outlined in Scheme 1, the MTEG-substituted OPVs
were then prepared from the MTEG-substituted benzal-
dehydes using a HWE reaction with the diphosphonate
ester 7.
In addition to using MTEG groups directly attached
to the terminal aryl rings of the OPV system, we also
prepared OPV 17 with MTEG-substituted amino groups
as outlined in Scheme 3. The yield of the HWE reaction
for the preparation of 17 was relatively low, however.
This is ascribed to the workup, where it was difficult to
elute 17 from the chromatography column.
In one group of ionic OPV-based sensitizers, the goal
was to use the anions of sulfonic acid moieties to impart
water solubility. As shown in Scheme 4, one such acid
was prepared by an HWE reaction between the phos-
phonate ester 19 and 2,5-dibromo-terephthalaldehyde.
The phosphonate ester 19 was prepared by brominating
toluene-4-sulfonic acid methyl ester and then converting
the resultant 4-bromomethyl-benzenesulfonic acid methyl
ester (18) into the phosphonate ester (19) using the
Michaelis-Arbuzov reaction.41 In this reaction, the meth-
yl sulfonic ester transesterified forming the ethyl ester,
which was later hydrolyzed during workup after the
HWE reaction. The monopotassium salt of the sulfonic
acid was isolated.
A more practical and atom-economical way of prepar-
ing the aldehyde without using chromic reagents is to
reduce the acetyl-protected gallous acid to the alcohol
with BH3‚Me2S, followed by reoxidation to the aldehyde
using a Swern oxidation38 (Scheme 2). The acetyl groups
were then removed with sodium methoxide in methanol
as described previously.35,36 Going beyond the framework
of our present work, it should be noted that the resultant
trihydroxy-benzaldehyde is an important building block
in the synthesis of certain dendrimers39 and that many
(35) Rosenmund, K. W.; Zetzsche, F. Chem. Ber./Recl. 1918, 51, 594-
602.
(36) Rosenmund, K. W.; Pfannkuch, E. Chem. Ber./Recl. 1922, 55,
2357-2372.
(37) Schmidt, U.; Wild, J. Liebigs Ann. Chem. 1985, 9, 1882-1894.
(38) Mancuso, A. J.; Swern, D. Synthesis 1981, 3, 165-185.
(39) Brouwer, A. J.; Mulders, S. J. E.; Liskamp, R. M. J. Eur. J.
Org. Chem. 2001, 10, 1903-1915.
(40) Ito, C.; Itoigawa, M.; Mishina, Y.; Tomiyasu, H.; Litaudon, M.;
Cosson, J. P.; Mukainaka, T.; Tokuda, H.; Nishino, H.; Furukawa, H.
J. Nat. Prod. 2001, 64, 147-150.
(41) Bhattacharya, A. K.; Thyagarajan, G. Chem. Rev. 1981, 81,
415-430.
7068 J. Org. Chem., Vol. 70, No. 18, 2005

Water-Soluble Singlet Oxygen Sensitizers
SCHEME 5. Synthesis of the N-Methylated Benzothiazole Derivatives
The typical way of introducing sulfonic acid function-
alities on an aromatic ring is by direct sulfonation with
HSO3Cl or a reaction with SO3 in concentrated sulfuric
acid. These are not viable strategies for functionalizing
acid-sensitive molecules. Our approach constitutes a mild
method for introducing the sulfonic acid functionality. To
the best of our knowledge, this is the first example of an
HWE reaction involving a sulfonic acid-substituted phos-
phonate ester. Unfortunately, the sulfonic acid-deriva-
tized OPV 20 turned out to be sparingly soluble in water.
However, we did manage to get enough into solution in
a 1 M sodium hydroxide solution to record an absorption
spectrum.
Our quest to make ionic water-soluble OPV derivatives
was more successful with the use of substituents based
on positively charged salts of N-methyl-pyridinium, 1,1-
dimethyl-piperazinium, and N-methyl-benzothiazolium
moieties. An A-π-D-π-A architecture using methoxy
groups as donors was employed in the design of the
N-methyl-benzothiazolium derivative 23 (Scheme 5).
Although initial attempts were made to condense N-
methyl-benzothiazolium iodide with 2,5-dimethoxyben-
zene-1,4-dicarbaldehyde, we were unable to isolate 23
with iodide as the counterion using this approach.
Benzothiazolium derivatives are normally prepared by
a Knoevenagel condensation with a N-methyl-benzothia-
zolium salt, but problems are usually encountered with
this strategy as two benzothiazole molecules can con-
dense with one aldehyde.42 Problems have also been
reported with attempts to condense benzothiazole salts
with ortho-substituted benzaldehydes.42 All of these
difficulties can be avoided by using the HWE methology
illustrated in Scheme 5, and 22 was prepared by reacting
2,5-dimethoxybenzene-1,4-dicarbaldehyde with the phos-
phonate ester 21. The N-methyl-benzothiazolium salt 23
was made by heating a solution of 22 with dimethyl
sulfate in DMF.
In an attempt to prepare a sensitizer with a longer
conjugated bridge between the terminal electron accep-
tors and one that would have a larger probability for
intersystem crossing, we made 26 (Scheme 5). In this
case, the aldehyde end-capped OPV (24) was prepared
by reacting monoacetal protected terephthalaldehyde
with 7 followed by hydrolysis of the acetal protecting
groups during aqueous workup. Compound 25 was then
prepared by reacting the dialdehyde 24 with the phos-
phonate ester 21. The N-methyl-benzothiazolium salt 26
(42) Jørgensen, M.; Krebs, F. C. J. Org. Chem. 2004, 69, 6688-6696.
J. Org. Chem, Vol. 70, No. 18, 2005 7069

Nielsen et al.
SCHEME 6. Preparation of the Tetrabromo Phosphonate Ester (28)
SCHEME 7. Synthesis of the Pyridinium OPV Salts
SCHEME 8. Preparation of a Water-Soluble 1,1-Dimethyl-piperazine Derivative (37)
was likewise made by heating a solution of 25 with
dimethyl sulfate in DMF.
of whether the HWE reaction was quenched with water
or dilute hydrochloric acid.
The preparation of the pyridinium salts 32, 33, 34, and
35 is outlined in Scheme 7. In these cases, reaction of
the appropriate phosphonate ester with 4-pyridine-car-
boxaldehyde was followed by conversion to the N-methyl-
pyridinium salt upon treatment with dimethyl sulfate in
DMF. The tetrabromo phosphonate ester (28) used for
making 31 was prepared as shown in Scheme 6. To
investigate the effect of having a different counterion, in
this case iodide, 30 was also treated with methyl iodide
in a Menschutkin reaction43 to yield 34.
The free base 30 was obtained when the HWE reaction
was quenched with water, whereas the dihydrochloride
salt (30‚2HCl) was obtained when the reaction was
quenched with dilute hydrochloric acid. For 29 and 31,
however, only the free bases were isolated irrespective
A crystal structure of 35 shows that the end methyl
pyridinium rings are almost completely perpendicular
to the tetrabromo-substituted center ring and that the
double bonds are not coplanar with the center ring
(Figure 1, the crystallographic data are presented in the
Supporting Information).
For the cationic salts prepared in Schemes 5 and 7,
the cation is an integral part of the chromophore. We also
prepared a piperizinium derivative in an attempt to
isolate the cation from the π-conjugated system of the
sensitizer (Scheme 8). In this case, the HWE reaction was
quenched with dilute hydrochloric acid and the dihydro-
chloride 36 was isolated. Interestingly, reaction of this
dihydrochloride with methyl iodide yielded the 1,1-
dimethyl-piperazinium salt (37). In principle, the Men-
schutkin reaction with methyliodide can result in the
methylation of either of the two different amine func-
(43) Menschutkin, N. Ber. Dtsch. Chem. Ges. 1895, 28, 1398-1407.
7070 J. Org. Chem., Vol. 70, No. 18, 2005

Water-Soluble Singlet Oxygen Sensitizers
TABLE 1. Photophysical Properties of Selected OPVsa
abs
λ
λfmax
(nm)
τs
max
log ꢀb
(ns) φfc
φ∆
c
compd
solvent
toluene
toluene
(nm)
8
9
354
4.37 426, 450 0.2 0.12 0.33
4.62 430, 454 0.2 0.11 0.35
370
10
toluene
water
0.32
0.20
11
12
13
17
20
23
toluene
water
364
348
443, 462
483
0.19 0.26
0.04 0.14
toluene
water
352
337
415, 436
442
0.18
0.10
FIGURE 1. ORTEP-III plot of the crystal structure of 35 (50%
ellipsoids) including one CH3SO4- counterion. The end-methyl
pyridinium rings are almost perpendicular to the center ring.
The connecting double bonds are also not in the same plane
as the center ring.
toluene
water
362
347
0.35
0.14
toluene
water
335, 426
352, 418
496, 526
557
0.16 0.33
0.04 0.11d
toluene
1 M NaOH 280
338
3.67 413, 438 0.3 0.12 0.22
water
491
395
4.65 639
4.51
1.9 0.18 0.01
26
30
water
509e
723e
toluene
397
326
4.48 457
4.38
1.9 0.01 0.31
32
33
water
water
360
4.49 466
0.01 0.03
446
360
4.60 604
4.48
1.5 0.14 0.02
FIGURE 2. Jablonski diagram illustrating the triplet-
sensitized production of singlet oxygen in both one- and two-
photon excitation schemes.
34
water
445
358
4.82 604
4.71
1.6 0.14 0.02
35
37
water
water
315
364
4.53 438
4.39 569
0.00 0.00
tionalities in 36. The 1H- and 13C NMR spectra obtained,
however, showed only one type of N-methyl group in
accordance with what is shown for 37 in Scheme 8.
Methylation of 36 with dimethyl sulfate was attempted,
but the product isolated was not stable upon exposure
to air (i.e., with CH3SO4- as the counterion, the product
isolated by filtration rapidly changed color from yellow
to brown and then to black).
Photophysical Characterization. The photophysical
processes that occur upon both one- and two-photon
irradiation of a singlet oxygen sensitizer are shown in
Figure 2. For centrosymmetric molecules such as ours,
selection rules dictate that the state initially populated
upon one-photon irradiation is different from that ini-
tially populated upon two-photon irradiation.44-46 How-
ever, we assume that, irrespective of the excitation
method and which state is populated first, Kasha’s rule47
will hold and that all subsequent photophysics and
photochemistry originate from one common state: the
vibrationally relaxed, lowest-energy excited singlet state,
S1. As such, measurements of fluorescence and singlet
oxygen yields performed under one-photon conditions can
be compared to those obtained under two-photon condi-
tions.
1.0 0.07 0.00
a Singlet oxygen quantum yield, φ∆; one-photon absorption band
abs
max
maximum,
λ
; molar extinction coefficient, ꢀ; fluorescence
band maximum, λfmax; fluorescence quantum yield, φf; S1 lifetime,
τs. b ꢀ has the units cm-1 M-1 c Errors are (10%. d Measured in
.
D2O. e Broad bands.
improbable. Rather, singlet oxygen is more efficiently
produced upon oxygen quenching of the longer-lived
triplet state of the sensitizer. As such, singlet oxygen
quantum yields, φ∆, are generally the largest for mol-
ecules that have large quantum yields of intersystem
crossing and, correspondingly, small quantum yields of
fluorescence, φf. Pertinent photophysical data for some
of the compounds prepared in the present work are
summarized in Table 1.
A. Absorption and Fluorescence Spectra. All the
MTEG-substituted OPV molecules (8-13) and the OPV
salts with two bromine atoms on the central aromatic
ring (20, 32, and 37) have a single broad one-photon
absorption band between ∼300 and 400 nm (Figure 3a,d).
On the other hand, the compounds with comparatively
strong electron-donating methoxy groups on the central
aromatic ring (i.e., 23, 33, and 34) have a second discrete
absorption band whose maximum is red-shifted with
respect to that of the first band (Figure 3b). We assign
this second band to a transition that populates a com-
paratively low-lying CT state. It was difficult to dissolve
the more extensively conjugated N-methyl-benzothiazo-
lium derivative 26 in water, and the absorption maxi-
mum reported in Table 1 for this compound may be that
for an aggregated species. This is consistent with the
rather broad spectral profile observed for 26.
Although singlet oxygen can be produced upon oxygen
quenching of S1,48 the S1 lifetime in most molecules is
generally so short as to render this bimolecular process
(44) McClain, W. M. Acc. Chem. Res. 1974, 7, 129-135.
(45) McClain, W. M. J. Chem. Phys. 1971, 55, 2789-2796.
(46) Monson, P. R.; McClain, W. M. J. Chem. Phys. 1970, 53, 29-
37.
(47) Gilbert, A.; Baggott, J. Essentials of Molecular Photochemistry;
CRC Press: Boca Raton, FL, 1991.
(48) Dobrowolski, D. C.; Ogilby, P. R.; Foote, C. S. J. Phys. Chem.
1983, 87, 2261-2263.
The piperazine salt 37 has an absorption maximum
(364 nm) very similar to that of the N-methylpyridyl salt
J. Org. Chem, Vol. 70, No. 18, 2005 7071

Nielsen et al.
the intensity of the 1270 nm singlet oxygen phosphores-
cence produced upon one-photon irradiation of the given
sensitizer to that obtained upon one-photon irradiation
of a standard sensitizer (phenalenone with φ∆ ) 1.00 (
0.0549). The concentrations of the dyes used were adjusted
such as to yield an absorbance at the irradiation wave-
length of ∼0.2-0.3. In general, this corresponds to [dye]
< 2 × 10-5 M. With respect to the experiments performed
in water, we ascertained that the Lambert-Beer rela-
tionship was valid over a range of dye concentrations
whose upper limit exceeded the concentration used for
the quantum yield measurements (i.e., absorbance was
linearly dependent on dye concentration). Thus, dye
aggregation does not appear to be a problem for the
molecules studied.
In a triplet state photosensitized process, the evolution
of singlet oxygen in time is described in the sequence of
reactions shown in Scheme 9.
SCHEME 9. Evolution of Singlet Oxygen in a
Triplet Photosensitized Processa
a Singlet oxygen (O2(a1∆g)) is formed in a collision between the
photosensitizer triplet state and ground state oxygen (O2(X3Σg-)).
Thereafter, singlet oxygen is removed from the system by a
number of processes (i.e., chemical reaction, bimolecular quench-
ing, solvent-induced photophysical decay). For our present pur-
poses, we identify only one of these channels: decay to ground
state oxygen with a rate constant kdecay
.
On the basis of the kinetic sequence shown in Scheme
9, the concentration of singlet oxygen will evolve in time
according to:
FIGURE 3. Representative examples of absorption and
fluorescence spectra recorded in H2O. (a) 12. (b) 23. (c) 35. (d)
37.
kform[O2(X3Σg-)]
[O2(a1∆g)]t )
[T1]0(exp(-kTt) -
kdecay - kT
32 (360 nm). In contrast, the free-base amino analogue
17 has an absorption maximum at 418 nm. We infer from
these data that the lone-pair electrons on the phenyl-
substituted nitrogens in 37, assumed to be part of the
chromophore, are in fact decoupled from the π-system.
Indeed, as outlined below in our discussion of singlet
oxygen yields, our data are consistent with a π-system
in 37 characterized by an appreciable amount of positive
charge. These observations suggest that we have failed
in our attempt to isolate the terminal cationic site from
the π-system in 37.
Fluorescence data were also recorded for the com-
pounds prepared (Table 1, Figure 3). Even though
quantum yields of fluorescence were very small in many
cases (e.g., φf ≈ 0 for 35), we were able to record data
with appreciable signal-to-noise ratios (e.g., see spectrum
for 35 in Figure 3c). With compound 23, for which we
recorded a distinct absorption band assigned to a CT
transition (vide supra), the same fluorescence spectrum
was obtained irrespective of the irradiation wavelength
(Figure 3b). More specifically, when monitoring emission
at 639 nm, the excitation spectrum of 23 was identical
to the absorption profile independently recorded. These
data are consistent with a process in which emission in
23 occurs from the low-lying CT state.
exp(-kdecayt)) (1)
where kT is the rate constant that expresses all channels
for the decay of the photosensitizer triplet state, including
the oxygen-dependent bimolecular channel kform[O2(X3Σg-)],
and [T1]0 is the initial concentration of sensitizer triplet
state. In a typical aqueous system, the rate constants
kdecay and kT will be roughly equal,22 and as a conse-
quence, the time evolution of the singlet oxygen signal
should appear as the difference of two exponential
functions, showing distinct rising and falling components.
In Figure 4, we show time-resolved singlet oxygen
phosphorescence signals produced upon irradiation of 10
in water.
Upon irradiation of the sensitizer at comparatively low
powers (e.g., less than ∼1 mW or 1 µJ/pulse with our
femtosecond system), the data obtained correspond to
those expected on the basis of eq 1 (Figure 4a). Moreover,
in a fit of eq 1 to the data, we obtain a singlet oxygen
lifetime (i.e., 1/kdecay) that is consistent with that expected
for H2O (∼3.5 µs50,51). Upon irradiation of 10 at higher
laser powers (greater than ∼1 mW or 1 µJ/pulse with our
femtosecond system), the time-resolved singlet oxygen
phosphorescence signal shows an increased deviation
B. Singlet Oxygen Quantum Yields. The singlet
oxygen quantum yields, φ∆, reported in the present work
were obtained in time-resolved experiments by comparing
(49) Schmidt, R.; Tanielian, C.; Dunsbach, R.; Wolff, C. J. Photo-
chem. Photobiol., A 1994, 79, 11-17.
7072 J. Org. Chem., Vol. 70, No. 18, 2005

Water-Soluble Singlet Oxygen Sensitizers
observed, indicating that singlet oxygen is more ef-
ficiently produced in toluene than it is in the more polar
solvent water. Although this observation could reflect a
number of solvent-dependent phenomena (vide infra), the
data are at least consistent with the thesis that CT
character in a sensitizer and/or sensitizer-oxygen com-
plex can adversely affect singlet oxygen yields by provid-
ing an independent pathway for the deactivation of
sensitizer excited states that competes with energy
transfer to O2(X3Σg-).28,29 Keeping the framework of this
model in mind, we now focus on several specific points.
Within the MTEG-substituted series, the singlet oxy-
gen quantum yields consistently reflect the substitution
pattern of the pendant ether. Thus, for compounds in
which the alkoxy groups are ortho and/or para to the
vinyl moiety, the singlet oxygen yield is comparatively
large. In contrast, for the compounds in which the alkoxy
groups are meta to the vinyl moiety, the singlet oxygen
yield is noticeably smaller. These observations are con-
sistent with the expectation that, in contrast to the effect
of withdrawing electron density from the chromophore,
electron-donating groups cause, through resonance ef-
fects, an increase in φ∆.
It is clear that, for this particular series of OPV-based
sensitizers, addition of positively charged electron-
withdrawing substituents to each end of the π-system
adversely affects the singlet oxygen yield. This is seen
not only in the comparison between the MTEG- and ionic
N-methylated species, but in the comparison between the
ionic N-methylated species and the corresponding non-
methylated amines. For example, for the pyridine-
terminated OPV 30, φ∆ ≈ 0.3, whereas for the corre-
sponding N-methylpyridinium-terminated OPVs 33 and
34, φ∆ ≈ 0. Similarly, φ∆ values for the amino-terminated
OPV 17 are noticeably larger than φ∆ values for the
corresponding N-methylated compounds 32 and 37.
Indeed, the observation of φ∆ ≈ 0 for the N-methylpip-
erazine (37) suggests an appreciable 1,4-transannular
interaction in the terminal piperazinyl groups that, in
turn, couples positive charge onto the π-framework.
Singlet oxygen yields can be influenced by many
features of a given sensitizer, only one of which is the
extent of CT character. The latter can manifest its effects
in the sensitizer-oxygen contact complex that necessarily
precedes singlet oxygen formation.28-30,53,54 Alternatively,
prior to any interaction with oxygen, CT character in the
sensitizer can influence the energy, yield, and lifetime
of the sensitizer triplet state,55 all of which can, likewise,
ultimately be reflected in the singlet oxygen yield.
FIGURE 4. Time-resolved 1270 nm phosphorescence signal
from singlet oxygen in H2O using 10 as the photosensitizer.
Data were recorded after pulsed femtosecond irradiation of the
sensitizer, and singlet oxygen was monitored in a single photon
counting experiment. (a) Low laser power, i.e., 0.4 mW or 0.4
µJ/pulse. (b) High laser power, i.e., 4 mW or 4 µJ/pulse. The
intense “spike” preceding the singlet oxygen signal is due to a
combination of sensitizer fluorescence, scattered laser light,
and luminescence from the optics used.
from first-order single-exponential decay kinetics (Figure
4b). Such behavior is commonly observed and has been
interpreted to reflect the photoinduced production of a
transient species that can quench singlet oxygen and
whose lifetime is somewhat shorter than that of singlet
oxygen.28,52 On the basis of these latter observations, all
singlet oxygen quantum yields were obtained using low
laser powers. (As an aside, it is important to note that,
in a typical two-photon experiment, where the sample
may be exposed to comparatively high laser powers, the
number of photons actually absorbed is significantly less
than that in the one-photon case. As such, all of the two-
photon-induced singlet oxygen signals we have observed
thus far in our work decay with single-exponential
kinetics.)
For all sensitizers in which electron-withdrawing ionic
groups are an integral part of the chromophore, a
comparatively small yield of singlet oxygen was obtained.
In contrast, the non-ionic, MTEG-substituted sensitizers
produce singlet oxygen in greater yield. For the latter
molecules, an appreciable solvent effect on φ∆ was
C. Sensitizer Triplet State Yields and Lifetimes.
To provide a more satisfying explanation for the singlet
oxygen yields reported in Table 1, it is necessary to
examine aspects of the triplet states of the respective
sensitizers. For example, the relative energy of the
sensitizer triplet state will influence the efficiency of
singlet oxygen production,30 and this energy is certainly
expected to change upon methylation of a nitrogen atom
whose lone-pair electrons are an integral part of the
chromophore. This phenomenon could certainly influence
(50) Egorov, S. Y.; Kamalov, V. F.; Koroteev, N. I.; Krasnovsky, A.
A.; Toleutaev, B. N.; Zinukov, S. V. Chem. Phys. Lett. 1989, 163, 421-
424.
(51) Schmidt, R.; Afshari, E. Ber. Bunsen-Ges. Phys. Chem. 1992,
(53) Stevens, B.; Small, R. D. Chem. Phys. Lett. 1979, 61, 233-238.
(54) Garner, A.; Wilkinson, F. Chem. Phys. Lett. 1977, 45, 432-435.
(55) McGlynn, S. P.; Azumi, T.; Kinoshita, M. Molecular Spectros-
copy of the Triplet State; Prentice Hall: Englewood Cliffs, NJ, 1969.
96, 788-794.
(52) Gorman, A. A.; Rodgers, M. A. J. J. Am. Chem. Soc. 1986, 108,
5074-5078.
J. Org. Chem, Vol. 70, No. 18, 2005 7073

Nielsen et al.
SCHEME 10. Kinetic Scheme for the Photosensitized Production of Singlet Oxygen
the singlet oxygen yields observed for 30 and 33 as well
as 37, for example. Sensitizer triplet yields and lifetimes
will likewise have a direct effect on singlet oxygen
production, and we set out to examine these issues in
detail for several of our compounds.
When comparing photophysical data obtained in tolu-
ene with those obtained in water, it is essential to account
for the comparatively low solubility of oxygen in the latter
solvent.56 Specifically, differences in the concentration of
ground state oxygen, O2(X3Σg-), in these solvents can
have a pronounced effect on the respective yields of
singlet oxygen, particularly if the inherent lifetime of the
sensitizer triplet state is comparatively short. To facili-
tate discussion of this point, and related material, we
refer to the kinetic model shown in Scheme 10.
Compound 17. Upon irradiation of compound 17 in
water and toluene solutions, the fluorescence intensities
and lifetimes obtained in oxygen-saturated solutions were
the same as those in nitrogen-saturated solutions. These
data indicate that oxygen does not quench the S1 state,
an observation that is consistent with the comparatively
short singlet state lifetimes of these OPV molecules
(Table 1).
Upon 355 nm pulsed laser irradiation of 17 in toluene,
we observed a transient absorption signal with λmax at
630 nm. This signal decayed in a first-order process,
yielding a lifetime of 2.9 ( 0.1 µs in a nitrogen-saturated
solution (Figure 5a). In an air-saturated solution, the
lifetime of this transient decreased to 0.18 ( 0.01 µs
(Figure 5b). With an oxygen concentration of 1.7 × 10-3
M in air-saturated toluene,56 these data yield a rate
constant of (3.1 ( 0.2) × 109 M-1 s-1 for oxygen-induced
deactivation of the transient. These oxygen-dependent
results that yield a quenching rate constant approxi-
mately 10 times smaller than the diffusion-limited rate
constant are consistent with assigning this transient
signal to the triplet state of 17.57
The initial absorbance of this triplet signal (i.e., at time
) 0) was independent of oxygen. This observation is
consistent with an S1 lifetime in 17 that is short enough
to preclude quenching by oxygen (vide supra). Most
importantly, however, the lifetime of the triplet state in
the absence of oxygen, τT ) 2.9 ( 0.1 µs, is likewise short,
certainly in comparison to triplet lifetimes for many other
organic molecules.58 This latter observation was verified
in an independent photoacoustic experiment (Figure 6).
For photoacoustic experiments performed against a
calibrated reference standard, one can accurately quan-
tify the fraction of “fast” heat released by a given
compound, R, and distinguish this from heat released
over longer periods of time (i.e., “slow” heat release).59,60
For our experiment performed in the presence of oxygen,
fast heat release derives from Sn and Tn deactivation in
the sensitizer, whereas slow heat release reflects O2(a1∆g)
deactivation (see Scheme 10 and Figure 2). For compound
17 in air-saturated toluene, the photoacoustic data yield
R ) 0.77 ( 0.06. Using all of the photophysical param-
eters determined thus far for this particular system, we
found that the calculated expected value of R is 0.80 (
0.02.61 This independent exercise confirms the accuracy
of data reported in Table 1.
Upon 355 nm pulsed laser irradiation of 17 in water,
we were unable to detect a transient absorption signal
that could be assigned to the triplet state. From photo-
(57) Gijzeman, O. L. J.; Kaufman, F.; Porter, G. J. Chem. Soc.,
Faraday Trans. 2 1973, 69, 708-720.
(58) Carmichael, I.; Hug, G. L. In CRC Handbook of Organic
Photochemistry; Scaiano, J. C., Ed.; CRC Press: Boca Raton, FL, 1989;
Vol. 1, pp 369-403.
(59) Heihoff, K.; Redmond, R. W.; Braslavsky, S. E.; Rougee, M.;
Salet, C.; Favre, A.; Bensasson, R. V. Photochem. Photobiol. 1990, 51,
635-641.
(56) IUPAC Solubility Data Series. Volume 7: Oxygen and Ozone;
Battino, R., Ed.; Pergamon Press: Oxford, 1981.
(60) Braslavsky, S. E.; Heibel, G. E. Chem. Rev. 1992, 92, 1381-
1410.
7074 J. Org. Chem., Vol. 70, No. 18, 2005

Water-Soluble Singlet Oxygen Sensitizers
FIGURE 6. (Top) Photoacoustic signal observed after pulsed
laser irradiation of 17 (solid line), and independently, the
signal observed from the reference standard, 2-hydroxyben-
zophenone (dashed line). Superimposed on, and indistinguish-
able from, the data from 17 is a simulated acoustic signal
obtained using a time constant for fast heat release of 1 ns
and a time constant for slow heat release of 2.8 µs. The latter
is consistent with the triplet lifetime obtained in the time-
resolved optical absorption measurements (Figure 5a). (Bot-
tom) Residuals plot for the difference between the simulated
and actual data on 17. Data were recorded in nitrogen-
saturated toluene.
where φT is the triplet quantum yield. Thus,
φw∆ater
φwTaterf wTaterS∆water
)
(4)
FIGURE 5. Transient absorption signals recorded at 630 nm
upon 355 nm pulsed laser irradiation of compound 17 in
toluene. (a) Nitrogen-saturated solution. (b) Air-saturated
solution.
φt∆oluene φtToluenef tTolueneSt∆oluene
We ascertained that, in toluene, the inherent triplet
lifetime of 17 in the absence of oxygen, τT, is 2.9 µs.
Moreover, the rate constant for oxygen quenching of the
triplet state of 17 is 3.1 × 109 M-1 s-1. If we assume that
these values are equally valid in water, and given that
[O2]toluene ) 1.73 mM and [O2]water ) 0.278 mM in air-
saturated solutions,56 we can use the expression shown
in Scheme 10 to quantify the relative fraction of triplet
states quenched by oxygen, fT, in these respective solvents
(eq 5).
acoustic measurements, however, we obtained R ) 0.94
( 0.08. Using the photophysical parameters indepen-
dently obtained for this system (Table 1),62 we calculated
an expected value of 0.95 ( 0.01 for R.
We now examine the relative magnitudes of the singlet
oxygen quantum yield for compound 17 in water and
toluene.
From Table 1, we see that
fwTater
f tToluene
φw∆ater
φt∆oluene
0.71 ( 0.08
0.94 ( 0.08
0.11 ( 0.01
0.33 ( 0.03
)
) 0.76 ( 0.11
(5)
)
) 0.33 ( 0.05
(2)
Substituting eq 5 into eqs 2 and 4 yields
From Scheme 10, we see that, in the triplet sensitized
production of singlet oxygen,
φwTaterSw∆ater
) 0.44 ( 0.08
φtTolueneSt∆oluene
(6)
φ∆ ) φTfTS∆
(3)
Thus, taking into account the “trivial” effect of solvent-
dependent changes in the oxygen concentration, we
nevertheless find that the change in solvent from toluene
to water has a pronounced adverse effect on the efficiency
with which compound 17 can produce singlet oxygen.
This decrease in production efficiency derives from
solvent-dependent changes in two parameters, φT and S∆.
Thus, with the change from a nonpolar to polar solvent,
the changes observed in the singlet oxygen yield remain
consistent with those expected for a solvent-dependent
(61) The following expression was used to calculate R, the fraction
of heat released as “fast” heat: Elaser(1 - R) ) φfEf + φ∆E∆ exp(-τa/τ∆)
where Elaser is the energy imparted by the pulsed laser used to irradiate
the system at 355 nm (337 kJ/mol), φf is the quantum efficiency of
fluorescence (0.16), Ef is the energy of the fluorescent state (241 kJ/
mol, obtained from λmax fluorescence), φ∆ is the quantum yield of singlet
oxygen (0.33), E∆ is the excitation energy of singlet oxygen (94 kJ/
mol), τ∆ is the lifetime of singlet oxygen in toluene (30 µs), and τa is
the effective acoustic transit time in toluene (500 ns). The latter
parameter is determined by the laser beam radius and the speed of
sound in toluene.
(62) In this case, the following parameters were used in the
calculation: φf ) 0.04, Ef ) 215 kJ/mol, φ∆ ) 0.11, τ∆ ) 4 µs.
J. Org. Chem, Vol. 70, No. 18, 2005 7075

Nielsen et al.
increase in the extent of CT character, both in the
sensitizer itself as well as in the sensitizer-oxygen
complex.
Compound 11. Upon 355 nm pulsed laser irradiation
of 11 in toluene, we observed a transient absorption
signal with λmax at 580 nm. This signal decayed in a first-
order process, yielding a lifetime of 560 ( 30 ns in a
nitrogen-saturated solution. In an air-saturated solution,
the lifetime of this transient decreased to 160 ( 8 ns.
With an oxygen concentration of 1.7 × 10-3 M in
air-saturated toluene,56 these data yield a rate constant
of (2.6 ( 0.2) × 109 M-1 s-1 for oxygen-induced deactiva-
tion of the transient. These results are consistent with
assigning this transient signal to the triplet state of 11.
In a photoacoustic study of 11 in toluene, we obtained
a value for “fast” heat release, R, of 0.77 ( 0.06. Using
data in Table 1, we calculated an expected value of R )
0.77 ( 0.02. Thus, once again, the photoacoustic experi-
ment independently confirms our optical data.
FIGURE 7. Transient absorption signals recorded at 550 nm
upon 355 nm pulsed laser irradiation of compound 32 in water
under both nitrogen- and air-saturated conditions.
The comparatively short inherent triplet lifetime of 11
means that, in this case, the yield of singlet oxygen
produced from the triplet state will be particularly
sensitive to the effects of a solvent-dependent change in
oxygen concentration. Once again, if we assume that the
values of τT and kqT obtained for 11 in toluene are equally
valid in water, and given that [O2]toluene ) 1.73 mM and
[O2]water ) 0.278 mM in air-saturated solutions,56 we
obtain
numbers). On this basis, we assign this signal to the
triplet state of 32.
The assignment of the rapidly decaying optical signal
in Figure 7 to the triplet state is also consistent with the
photoacoustic data recorded for this compound. Specifi-
cally, we obtain an experimental R value of 1.00 ( 0.05
and a calculated R value of 0.98 ( 0.02, indicating that
essentially all the energy absorbed upon irradiation of
32 is released as “fast” heat. This result is in keeping
with the fact that this molecule does not produce singlet
oxygen in appreciable yield (φ∆ ) 0.03) nor does it
fluoresce (φf ) 0.01).
f wTater
0.29 ( 0.04
)
) 0.40 ( 0.06
(7)
f tToluene
0.72 ( 0.04
The data obtained from compound 32 further reinforce
our working hypothesis that substantial CT character in
a given molecule can contribute to rapid nonradiative
deactivation channels that compete with channels that
eventually result in the production of singlet oxygen.
D. Sensitizer Stability. Although the water-soluble
MTEG-substituted molecules produce singlet oxygen in
moderate yields, these molecules were unfortunately not
particularly stable upon irradiation in aerated H2O. This
is illustrated in Figure 8 with data recorded from 11 and
17. In both cases, the degradation reactions involved
appear to shorten the chromophore as seen by a decrease
in the intensity of the longest wavelength absorption
band. Moreover, distinct isosbestic points in the spectra
indicate that, under the conditions of these experiments,
only one degradation product is formed in each case.
Similar behavior has been observed by Strehmel et al.63
upon irradiation of other OPV systems.
Despite the apparent similarities in the data shown
in Figure 8, the kinetics of degradation for 11 and 17
were different. It is important to first note that quantify-
ing the kinetics of self-sensitized decay reactions can be
complicated. Specifically, even over comparatively short
periods of irradiation, the species absorbing light at a
given wavelength can change, and in turn, different
reactions can come into play.
This result indicates that, indeed, differences in oxygen
solubility have a pronounced effect on the fraction of
compound 11 triplet states quenched by oxygen.
In this light, when we consider the ratio of singlet
oxygen quantum yields obtained for compound 11 in
these respective solvents,
φw∆ater
φt∆oluene
0.14 ( 0.01
0.26 ( 0.03
)
) 0.54 ( 0.08
(8)
the relative effect of the change in solvent from toluene
to water on the product φTS∆ is comparatively small for
compound 11. It is important to keep in mind, however,
that the origin of this solvent effect ultimately derives
from the uniquely short triplet lifetime of compound 11.
As already discussed, this in itself could be a consequence
of a number of factors, including the extent of CT
character in the molecule.
Compound 32. Recall that this molecule is a pyri-
dinium salt, and the expectation is that that there is an
appreciable amount of CT character in the system. Upon
355 nm pulsed laser irradiation of 32 in water, we
observed a transient absorption signal with λmax at 550
nm. This signal decayed in a first-order process, yielding
a lifetime of 160 ( 8 ns in a nitrogen-saturated solution
(Figure 7). In an air-saturated solution, the lifetime
obtained was 150 ( 8 ns (Figure 7). These data yield a
rate constant of 1.5 × 109 M-1 s-1 for the oxygen-induced
deactivation of the transient (the errors on this number
are large, however, because the calculation involves
taking the difference between two large and similar
With this caveat in mind, we note that the overall rate
for photoinduced changes in a solution of 11 was slightly
(63) Strehmel, V.; Sarker, A. M.; Lahti, P. M.; Karasz, F. E.;
Heydenreich, M.; Wetzel, H.; Haebel, S.; Strehmel, B. ChemPhysChem
2005, 6, 267-276.
7076 J. Org. Chem., Vol. 70, No. 18, 2005

Water-Soluble Singlet Oxygen Sensitizers
FIGURE 10. Photoinduced changes in an aqueous solution
of compound 17 as monitored by recording the absorbance of
the solution at 418 nm as a function of the elapsed photolysis
time at 400 nm. A second-order decay function has been
applied to the experimental data.
such data are generally interpreted to indicate that
singlet oxygen is involved in the degradation process. In
contrast, we did not observe a D2O/H2O solvent isotope
effect on the rate of photoinduced change in a solution
of 17. Thus, it appears that, for these two compounds,
different reactions may be involved in the degradation
process. In support of this, we found that the decay of
11 did not follow either first- or second-order kinetics over
the reaction period. On the other hand, the decay of 17
followed second-order kinetics over a comparatively long
period of irradiation (Figure 10). These preliminary data
indicate that a complete study of these degradation
mechanism(s) exceeds the scope of the present report.
We conclude this discussion by noting that singlet
oxygen quantum yields obtained using these molecules
as sensitizers were recorded under conditions in which
negligible bleaching had occurred. Nevertheless, in a
series of experiments performed in toluene, we ascer-
tained that the quantum yield of singlet oxygen produc-
tion upon irradiation at 355 nm remains unchanged with
the extent of degradation. In the least, these latter results
should establish confidence in the data shown in Table
1.
FIGURE 8. UV-vis absorption spectra of two MTEG-
substituted sensitizers recorded as a function of elapsed
exposure to 400 nm light in aerated H2O (CW laser operated
at 30 mW, irradiated spot size ∼5 mm in diameter). Arrows
on the respective spectra indicate whether the absorbance
increases (up arrow) or decreases (down arrow) upon prolonged
irradiation. Spectra were recorded at 1 min intervals during
irradiation. (a) 11. Note that the data are consistent with the
presence of a band at ∼290 nm whose intensity arguably
increases in time. (b) 17.
E. Two-Photon Photosensitized Production of
Singlet Oxygen. In our previous work,22 we demon-
strated that we could produce and optically detect singlet
oxygen in a time-resolved experiment upon two-photon
irradiation of 5,10,15,20-tetrakis(N-methyl-4-pyridyl)-
21H,23H-porphine dissolved in water. In our present
study, it remains for us to demonstrate that singlet
oxygen can similarly be generated and detected using one
of our phenylene-vinylene-based sensitizers. To this end,
17 was irradiated at 800 nm using the output of our
femtosecond laser system. Experiments were performed
in both D2O and H2O, using a dye concentration of 4 ×
10-4 M. Under these conditions, we detected the 1270 nm
phosphorescence of singlet oxygen in single photon
counting experiments. As required for a process of two-
photon excitation, the intensity of these singlet oxygen
phosphorescence signals scales according to the square
of the power of the irradiating laser (Figure 11). Thus,
we can indeed produce singlet oxygen in water upon two-
photon excitation of an OPV sensitizer with an electron
donor-acceptor architecture.
FIGURE 9. Photoinduced changes in aqueous solutions of
compound 11 as monitored by recording the absorbance of the
solution at 364 nm as a function of the elapsed photolysis time
at 400 nm. The data indicate that the degradation reaction
proceeds at a slightly faster rate in D2O than in H2O. An
arbitrary fitting function has been applied to the data (i.e.,
neither first- nor second-order kinetics are observed).
faster in D2O than in H2O (Figure 9). Because the lifetime
of singlet oxygen is longer in D2O than it is in H2O,50,64
(64) Ogilby, P. R.; Foote, C. S. J. Am. Chem. Soc. 1983, 105, 3423-
3430.
J. Org. Chem, Vol. 70, No. 18, 2005 7077

Nielsen et al.
FIGURE 12. One-photon absorption spectrum (bottom and
left-side axes) and two-photon action spectrum (top and right-
side axes) for compound 17 in toluene. The two-photon
spectrum was recorded by monitoring the intensity of singlet
oxygen phosphorescence as a function of the wavelength of
light used to irradiate 17, calibrating each point against the
corresponding signal from a reference standard.22 The wave-
length range accessible for the two-photon spectrum is limited
by the range over which the femtosecond laser used can be
tuned.
FIGURE 11. Double logarithmic plot of the singlet oxygen
phosphorescence intensity against the average laser power
incident on the sample. These data were recorded upon 800
nm irradiation of aqueous solutions of 17. (a) Experiments
performed in H2O with 0.44 mM of 17. A slope of 2.1 ( 0.4
was obtained. (b) Experiments performed in D2O with 0.38 mM
of 17. A slope of 2.0 ( 0.2 was obtained.
produced with moderate efficiency. On the other hand,
for sensitizers in which water solubility derived from
ionic substituents (e.g., salts of N-methylpyridyl moi-
eties), singlet oxygen was not produced in appreciable
yield. These data provide insight into features of a given
molecule important in the process of energy transfer to
produce singlet oxygen. Of particular concern in this
regard are the extent to which rapid charge transfer-
mediated processes can deactivate sensitizer excited
states at the exclusion of singlet oxygen production.
The two-photon absorption profile of a compound also
provides important information about the energies of
accessible states. Using a technique described previ-
ously22 in which we monitor the intensity of singlet
oxygen phosphorescence produced as a function of the
wavelength of light used to irradiate a given molecule,
we can generate a two-photon action spectrum. When
calibrated against the corresponding action spectrum
obtained from a standard molecule, one can obtain
the two-photon absorption spectrum. Such a spectrum
is shown in Figure 12 for compound 17 dissolved in
toluene. The reference standard used in this case was
2,5-dicyano-1,4-bis-(2-(4-diphenylaminophenyl)vinyl)ben-
zene, CNPhVB.22 Although values reported for the ab-
solute TPA cross section of CNPhVB at its peak maxi-
mum cover a large range,1,23,65 the data indicate that the
molecule has a comparatively large TPA cross section
(e.g., (1890 ( 280) × 10-50 cm4 s photon-1 molecule-1 at
∼840 nm). As such, the TPA cross section for 17 at 775
nm is appreciable.
Experimental Section
Synthesis. The detailed procedures used to synthesize the
compounds discussed in this work, along with the associated
characterization data, are provided in the Supporting Informa-
tion.
Photophysical Methods. All measurements were per-
formed using spectroscopic grade solvents.
A. Sensitizer Fluorescence Experiments. Solutions were
degassed for 15 min with argon prior to use. Data were
recorded using an instrument composed of a 450 W Xe lamp
for steady-state measurements and a diode laser for lifetime
measurements. The detection system was a single photon-
counting photomultiplier tube in a peltier-cooled housing. All
spectra were measured in a perpendicular geometry using 1
cm quartz cuvettes. Steady-state measurements were obtained
with 1.8 nm band-pass filters and corrected for wavelength-
dependent intensity variation of the excitation light source.
Quantum yields were determined using 9,10-diphenyl-
anthracene in cyclohexane (φf ) 0.93 ( 0.03)66 as the fluores-
cence standard with refractive index and differential absorp-
tion corrections. In all cases, time-resolved fluorescence decay
traces were single-exponential, and lifetimes were determined
using least-squares analysis. All time-resolved measurements
were conducted with 7.3 nm band-pass filters. Experiments
Conclusion
In conclusion, we have shown that viable water-soluble
two-photon singlet oxygen sensitizers can be constructed
using an electron donor-acceptor OPV architecture. In
two different approaches, water solubility was achieved
through the use of specific substituents covalently at-
tached to the sensitizer. In most cases examined, these
substituents were an integral part of the chromophore.
For OPV sensitizers in which solubility derived from the
use of triethylene glycol substituents, singlet oxygen was
were performed using dye concentrations of ∼5 × 10-6
where the sample absorbance was ∼0.2-0.3.
M
(65) Zhang, B.-J.; Jeon, S.-J. Chem. Phys. Lett. 2003, 377, 210-216.
7078 J. Org. Chem., Vol. 70, No. 18, 2005
(66) Meech, S. R.; Phillips, D. J. Photochem. 1983, 23, 193-217.

Water-Soluble Singlet Oxygen Sensitizers
B. Singlet Oxygen Experiments. Details of the instru-
mentation and approach used to produce and optically detect
singlet oxygen in both one- and two-photon experiments are
presented elsewhere.19,22,67 For some of the one-photon experi-
ments, the near-IR output of our femtosecond laser system was
frequency-doubled in a â-barium borate crystal to yield a
suitable UV-vis wavelength with which to irradiate the
sensitizer (e.g., 800 nm f 400 nm; energy < 0.4 µJ/pulse).
Otherwise, the sensitizer was irradiated with the output of a
nanosecond Nd:YAG laser system (energy < 0.3 mJ/pulse). For
data recorded in water (e.g., Figure 4a), the fitting function
derived from eq 1 was extrapolated to time ) 0 and then
integrated to yield an intensity suitable for comparison to that
obtained from the reference standard. For data recorded in
toluene, where kT . kdecay, eq 1 reduces to a single-exponential
decaying function. In this case, both the signal intensity at
time ) 0 as well as the integral of the signal were used in the
comparison to the reference standard.
approach used has been outlined previously.59 Irradiation
energies used were approximately 20 µJ/pulse. Data were
analyzed using a commercially available deconvolution routine
(Sound Analysis 1.50D, Quantum Northwest).
E. Two-Photon Experiments. Details of the instrumenta-
tion and approach used to quantify TPA cross sections and
action spectra are published elsewhere.22
Acknowledgment. This work was funded by the
Villum Kann Rasmussen Foundation and by the Danish
Natural Science Research Council through a block grant
for The Center for Oxygen Microscopy. We thank (1)
Silvia Braslavsky (Max Planck Institute, Mu¨lheim) for
the photoacoustic transducer used in these experiments,
(2) Frederik C. Krebs and Holger Spanggaard (Risø
National Laboratories) for valuable discussions, and (3)
Flemming Hansen for collecting the crystallographic
data and the Center for Crystallographic Studies,
University of Copenhagen, for the use of their equip-
ment.
C. Triplet State Absorption Measurements. Details of
the instrumentation and approach used for the time-resolved
absorption measurements are published elsewhere.67
D. Photoacoustic Experiments. Time-resolved photo-
acoustic signals were monitored after pulsed laser irradiation
using a home-built piezoelectric transducer. The general
Supporting Information Available: Synthetic proce-
dures and characterization of all compounds prepared. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(67) Keszthelyi, T.; Weldon, D.; Andersen, T. N.; Poulsen, T. D.;
Mikkelsen, K. V.; Ogilby, P. R. Photochem. Photobiol. 1999, 70, 531-
539.
JO050507Y
J. Org. Chem, Vol. 70, No. 18, 2005 7079