Synthesis, characterization, photophysics and photochemistry of pyrylogen electron transfer sensitizers

Article abstract of DOI:10.1111/php.12174

A series of new dicationic sensitizers that are hybrids of pyrylium salts and viologens has been synthesized. The electrochemical and photophysical properties of these "pyrylogen" sensitizers are reported in sufficient detail to allow rationale design of new photoinduced electron transfer reactions. The range of their reduction potentials (+0.37-+0.05 V vs SCE) coupled with their range of singlet (48-63 kcal mol^-1) and triplet (48-57 kcal mol^-1) energies demonstrate that they are potent oxidizing agents in both their singlet and triplet excited states, thermodynamically capable of oxidizing substrates with oxidation potentials as high as 3.1 eV. The pyrylogens are synthesized in three steps from readily available starting materials in modest overall 11.4-22.3% yields. These sensitizers have the added advantages that: (1) their radical cations do not react on the CV timescale with oxygen bypassing the need to run reactions under nitrogen or argon and (2) have long wavelength absorptions between 413 and 523 nm well out of the range where competitive absorbance by most substrates would cause a problem. These new sensitizers do react with water requiring special precautions to operate in a dry reaction environment. Hybrids of pyrylium salts and viologens photochemically generate radical-cation/radical-cation pairs with substantial intra-ion repulsion that increases the rate constant of separation, k_SEP, and competitively inhibits energy wasting return electron transfer. These first representatives of dicationic charge-shift sensitizers generate radical cations that do not react with oxygen on the CV timescale and absorb between 413 and 523 nm well outside the range where competitive absorbance by most substrates would cause a problem.

Full text of DOI:10.1111/php.12174

Photochemistry and Photobiology, 2014, 90: 344–357
Synthesis, Characterization, Photophysics and Photochemistry
of Pyrylogen Electron Transfer Sensitizers†
Edward L. Clennan*1 and Chen Liao2
1Department of Chemistry, University of Wyoming, Laramie, WY
2Chemical Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN
Received 8 July 2013, accepted 5 September 2013, DOI: 10.1111/php.12174
ABSTRACT
act as the oxidant and an energetic electron that can function as a
reductant.
The efficiency of green PET reactions critically depend upon
A series of new dicationic sensitizers that are hybrids of
pyrylium salts and viologens has been synthesized. The elec-
trochemical and photophysical properties of these “pyrylo-
gen” sensitizers are reported in sufficient detail to allow
rationale design of new photoinduced electron transfer reac-
tions. The range of their reduction potentials (+0.37–+0.05 V
vs SCE) coupled with their range of singlet (48–63 kcal
mol−1) and triplet (48–57 kcal molꢀ1) energies demonstrate
that they are potent oxidizing agents in both their singlet and
triplet excited states, thermodynamically capable of oxidizing
substrates with oxidation potentials as high as 3.1 eV. The
pyrylogens are synthesized in three steps from readily avail-
able starting materials in modest overall 11.4–22.3% yields.
These sensitizers have the added advantages that: (1) their
radical cations do not react on the CV timescale with oxygen
bypassing the need to run reactions under nitrogen or argon
and (2) have long wavelength absorptions between 413 and
523 nm well out of the range where competitive absorbance
by most substrates would cause a problem. These new sensi-
tizers do react with water requiring special precautions to
operate in a dry reaction environment.
ꢁꢂ
ꢁꢂ
the photocatalyst regeneration partition ratio (jReg ¼ NP =NR ),
ꢁꢂ
ꢁꢂ
where NP and NR are, for the reaction shown in Scheme 1,
the number of product radical cations (Pꢁꢂ) and reactant radical
cations (Rꢁꢂ), respectively, that oxidize CATꢀ˙ to regenerate the
photocatalyst CAT. Unfortunately, a significant number of the
oxidations of CATꢀ˙ by the reactant radical cation occur in a
tight radical-cation/radical-anion pair (Rꢁꢂ || CATꢀ˙) that forms
in the initial electron transfer event. This nondiffusive return
electron transfer (RET) is entropically favored with respect to
oxidation with Pꢁꢂ and, as a consequence, is a major source of
reaction inefficiency in many PET reactions.
In order to realize the full synthetic potential of PET reactions
a significant amount of effort has been expended to minimize
energy wasting RET in the tight ion pair (Rꢁꢂ || CATꢀ˙). Several
clever methods have been devised to discriminate against RET
including the use of cosensitizers (5), the generation of triplet
tight ion pairs where RET is spin forbidden (6), the use of sub-
stituent effects in order to place the RET deep in the Marcus
inverted region where the rate constants are diminishingly small
(7), and the use of charge shift (CS) reactions (8). Unfortunately,
none of these methods are universally applicable and they have
all generated mixed results. For example, the quantum yield of
ion pair separation in the PET reaction of biphenyl (Scheme 2)
exhibited only a modest increase from 0.18 to 0.33 as the attrac-
tive interaction in ion pair I was decreased by forming the nonat-
tractive CS pair II (9) (Scheme 2).
In this manuscript, we describe in detail our synthesis, charac-
terization, electrochemical and photophysical studies of new
pyrylium-cation/viologen hybrid CS sensitizers, the pyrylogens,
XPY2+ (Scheme 1b) (10). These new photocatalysts were
designed in anticipation that the repulsive interaction in the ini-
tially formed CS radical-cation/radical-cation pair would enhance
kSEP (Scheme 2) above that observed in ion pair I and CS pair
II and completely inhibit RET (kRET in Scheme 2).
INTRODUCTION
The ostensibly minor modification of an organic substrate that
occurs by addition or removal of an electron can open new reac-
tion channels leading to remarkable bonding changes that are
difficult or impossible to achieve by other methods. Conse-
quently, these reactions have received considerable attention from
both physical organic and synthetic organic chemists (1). Photoin-
duced electron transfer (PET) (2) potentially provides a conve-
nient and green method to initiate these reactions since they can
be engineered in favorable cases to bypass the use of stoichiome-
tric methods employing chemical or photochemical oxidants or
reductants whose remnants need to be separated from the desired
reaction product at the end of the reaction. The green variant of
the PET reaction requires regeneration of the electron-transfer
photocatalyst (CAT) concomitant with product formation as illus-
trated in Scheme 1a. Photocatalysts (3,4) can either donate or
remove an electron from a suitable substrate since upon promo-
tion to an excited state they generate a hole that can potentially
MATERIALS AND METHODS
General aspects. 4-Pyridine carbaldehyde, acetophenone, p-chloroace-
tophenone, p-methoxyacetophenone, p-(trifluoromethyl)acetophenone,
triphenyl carbinol, methyl iodide, trifluoroacetic acid, boron trifluoride
dietherate, trimethyloxonium tetrafluoroborate, sodium hydride, and
2,4,6-triphenylpyrylium tetrafluoroborate were all obtained from Sigma-
Aldrich and used without further purification. Tetra-n-butyl ammonium
perchlorate was obtained from Fluka and recrystallized twice from
*Corresponding author email: Clennane@uwyo.edu (Edward L. Clennan)
†This paper is part of the Special Issue honoring the memory of Nicholas J. Turro.
© 2013 The American Society of Photobiology
344

Photochemistry and Photobiology, 2014, 90 345
Scheme 1. (a) Catalyzed photoinduced electron-transfer reactions. (b) Structures of pyrylogen photoinduced electron transfer catalysts.
Scheme 2. Photoinduced electron transfer reactions of biphenyl and structures of electron transfer initiated tight ion pairs.
ethanol and then dried at 40°C in a vacuum oven. Potassium tetrakis(pen-
tafluorophenyl)borate was obtained from Boulder Scientific Company and
used as received. 1-Iodobutane was obtained from Sigma-Aldrich and
purified by passing it through a column of activated alumina. HPLC
Grade acetonitrile was purified by refluxing with a small amount of CaH2
for a few hours followed by distillation. THF was first distilled after
refluxing with sodium for 2 h then redistilled from sodium/benzophenone
ketyl. Methylene chloride was shaken with several portions of sulfuric
acid until the acid layer remained colorless, then washed with water, 5%
aqueous sodium bicarbonate, sodium hydroxide and finally with water. It
was then predried with anhydrous calcium chloride and distilled from
calcium hydride. Chloroform was washed with water to remove the
ethanol stabilizer, dried over calcium chloride and distilled from P2O5.
Tetra-n-butyl ammonium perchlorate is recrystallized from absolute
ethanol and dried in vacuo at 75°C.
IR spectra were recorded on a Nicolet Nexus 470 FT-IR using
OMNIC ESP 5.2 software. ElectroSpray Ionization mass spectra and
MS/MS spectra were obtained with a Thermo-Finnigan LCQ mass spec-
trometer. The samples were prepared by addition of the compound of
interest to methanol. UV–Visible spectra were collected on either an Agi-
lent 8453 or on a Cary 50 spectrometer. Fluorescence and phosphores-
cence spectra were collected on a Cary Eclipse using concentrations that
had less than 0.1 absorbance at the excitation wavelength. Cyclic Vol-
tammetry was done with a CH Instruments Inc. CHI 900 electrochemical
analyzer. Nanosecond laser flash photolysis was done using a Luzchem
miniaturized laser flash photolysis system coupled to a Spectra-Physics
10 Hz pulsed Nd:YAG laser. A Bruker EMX EPR spectrometer was used
to acquire electron spin resonance (ESR) spectra. Proton and carbon
NMR were obtained on a Bruker 400 MHz NMR and referenced to
either the solvent or TMS.
Fluorescence lifetimes. Time resolved fluorescence lifetime measure-
ments were made using a time-correlated single photon counting appara-
tus. The samples were irradiated at 400 or 420 nm obtained by
frequency-doubling the 800 or 840 nm output of a Ti:Sapphire laser
(Spectra Physics, Tsunami 3941) operating at
a repetition rate of
80 MHz. Fluorescence was detected over 60–90 s intervals at the emis-
sion maximum (5 nm slit widths) and the traces were deconvoluted from
the detector response function using the CFS_LS program. The detector
response function was recorded using a piece of glass to direct some of
the laser beam onto the detector and recording the subsequent signal at
405 and 425 nm using a slit width of 1 nm.
Cyclic voltammetry. All CV experiments were done in a three-elec-
trode cell using glassy carbon as the working electrode, Ag/AgNO3
(0.01 M in CH3CN) as a nonaqueous reference electrode, and a platinum
wire as the counter electrode. Samples were 2 9 10ꢀ3 M in the substrate
and ferrocene, which was used as an internal standard, and 0.1 M in the
supporting electrolyte, Bu4N+ ClO4ꢀ. Samples were run at a variety of
scan rates in both argon and oxygen atmospheres.
Computational chemistry. Geometry optimizations were performed
with the Gaussian 03 program using the Becke/Stephens three parameter

346 Edward L. Clennan and Chen Liao
Lee–Yang–Parr correlation hybrid functional B3LYP in conjunction with
the 6-311+G(2d,p) basis set. Frequency calculations were used to verify
location of energy minima. The absence of spin contamination was veri-
fied in each calculation by examination of <S2> that showed values
acceptably close to 0 for all singlets and to 0.75 for all doublets.
TD-DFT calculations were done with the Gaussian 09 program using
6-311+G(2d,p) optimized geometries as input (11).
(d, J = 6.8 Hz, 2H). 13C NMR (CD3CN) d 49.9(q, J = 146 Hz), 57.3(q,
J = 146 Hz), 116.0(d, J = 176 Hz), 117.1(d, J = 164 Hz), 121.9(s),
128.0(d, J = 176 Hz), 132.9(d, J = 164 Hz), 147.8(d, J = 194 Hz), 156
(s), 158.3(s), 167.8(s), 172.3(s). Elemental analysis: Found (C, 52.0; H,
4.22; N, 2.43), Calculated for C25H23B2F8NO3 (C, 53.7; H, 4.1; N, 2.5),
.
Calculated for C25H23B2F8NO3 H2O (C, 52.03; H, 4.36; N, 2.43). IR
(KBr): 3125, 3086, 1606, 1626, 1415–1507, 1375, 1120–1270 cmꢀ1
.
1,5-Diphenyl-3-(4-pyridyl)pentane-1,5-dione
1
(X = H). Sodium
ESI-MS m/z (%): 416.5 (M+˙ + CH3Oꢀ, 100). ESI-MSMS m/z (%):
hydride (0.15 g; 6.00 mmol) was added to 15 mL of vigorously stirred
methanol at 0°C followed by 5.0 g (41.7 mmol) of acetophenone. Pyri-
dine-4-carbaldehyde (2.0 g; 19.0 mmol) was then added slowly in order
to maintain the temperature between 0°C and 10°C. After addition was
complete water was then added dropwise until the solution turns milky.
A viscous oil settles to the bottom of the flask after stirring for 4 h. This
mixture was then stored at 0°C for 12 h and the resulting solid filtered
and recrystallized from 95% ethanol to afford 3.11 g (50% yield) of
creamy white crystals. m.p. 125–126°C. 1H NMR (CDCl3) d 3.4(dd,
J = 7.3, 17.2 Hz, 2H), 3.53(dd, J = 6.7, 17.2 Hz, 2H), 3.89(tt, J = 6.7,
7.3 Hz, 1H), 7.36(m, 2H), 7.52(m, 4H), 7.65(m, 2H), 7.95(m, 4H), 8.41
(m, 2H). 13C NMR (CDCl3) d 36.2, 43.9, 123.8, 128.4, 129.2, 133.8,
137, 150, 153.7, 198.7.
3-(1-Methylpyridium-4-yl)-1, 5-diphenylpentane-1, 5-dione Iodide 2
(X = H). Methyl iodide (1.0 g; 0.0068 mol) was added to a solution of
2.78 g (0.00850 mol) of 1, 5-diphenyl-3-(4-pyridyl) pentane-1, 5-dione
in 5 mL of chloroform. The solution is stirred for half an hour at room
temperature. After evaporation of chloroform at 20 mmHg, the green oil
is triturated with diethyl ether and the yellow crystals are recrystallized
from ethanol. Yield(1 g, 30%), m.p. 180–182°C. 1H NMR (CF3COOH/
385.5 (M+˙, 100).
1, 5-Di-p-chlorophenyl-3-(4-pyridyl) pentane-1, 5-dione 1(X = Cl).
p-Chloroacetophenone (3.2 g, 21 mmol) was added to a solution of
0.15 g sodium hydride in 15 mL of anhydrous methanol at 0°C. The
cloudy solution is stirred vigorously. Pyridine-4-carbaldehyde 1.50 g
(14.2 mmol) is added slowly maintaining the temperature at 0–10°C.
Water was added dropwise until the solution became milky. After stirring
the solution for one hour, white crystals were formed. The solution was
placed in ice for one hour and the resulting crystals were filtered and
recrystallized from 90% ethanol. After being dried in a vacuum oven at
room temperature, white crystals were obtained (1.24 g, 55%), m.p. 163–
165°C. 1H NMR (CDCl3) d 3.35(dd, J = 11, 7 Hz, 2H), 3.49(dd, J = 11,
7 Hz, 2H), 4.03(m, 1H), 7.2(d, J = 5 Hz, 2H), 7.4(dd, J = 5, 2 Hz, 4H),
7.8(dd, J = 5, 2 Hz, 4H), 8.5(d, J = 5 Hz, 2H). 13C NMR (CDCl3) d
36.5, 43.4, 128.4, 129.3, 129.8, 134.5, 140.3, 144.3, 165.8, 196.5.
3-(1-Methylpyridinium-4-yl)-1,5-di-p-chlorophenylpentane-1,5-dione
tetrafluoroborate 2(X = Cl). 1,5-di-p-Chlorophenyl-3-(4-pyridyl) pentane-
1,5-dione (1.5 g, 3.8 mmol) was dissolved in 30 mL of chloroform.
Trimethyloxonium tetrafluoroborate (0.59 g, 4.0 mmol) was added to the
mixture and stirred for 30 min. The solvent was evaporated in vacuo.
Ethanol (95%) was added and gently heated in order to dissolve the
entire residue. The cooled solution was refrigerated for 25 min inducing
precipitation of white crystals. The crystals were filtered and dried in a
vacuum oven at room temperature to afford 1.25 g, 65.8% yield of prod-
uct. m.p. 107–108°C. 1H NMR (CDCl3) d 8.5(d, J = 7 Hz, 2H), 8.1(d,
J = 7 Hz, 2H) 7.85(dd, J = 9, 8 Hz, 4H), 7.35(dd, J = 9, 8 Hz, 4H),
4.25(s, 3H), 4.10(m, 1H), 3.65(dd, J = 10, 7 Hz, 2H), 3.55(dd, J = 10,
7 Hz, 2H) 13C NMR (CDCl3) d 36.5, 43.4, 48.0, 128.4, 129.3, 129.8,
134.5, 140.3, 144.3, 165.8, 196.5.
4-(1-Methylpyridinium-4-yl)-2,6-di-p-chlorophenylpyrylium bistetraflu-
oroborate (ClPY2+). Triphenyl methanol (1.3 g, 5.0 mmol) was added to a
40 mL solution of 1.0 g (2.0 mmol) of 3-(1-methylpyridium-4-yl)-1,5-di-
p-chlorophenylpentane-1,5-dione tetrafluoroborate (1.0 g, 2.0 mmol) in
acetic anhydride. A 50% HBF4 solution, prepared by adding HBF4 drop-
wise to cold acetic anhydride, was then added dropwise to the stirred dione
solution. The mixture was then allowed to stir at 50°C for an additional 30
min. Diethyl ether was added to the cooled solution to induce formation of
a precipitate. The precipitate was filtered and recrystallized from glacial
acetic acid. After drying, it afforded orange crystal (0.56 g, 46% yield).
m.p. 303–305°C. Elemental Analysis: Found (C, 48.35; H, 3.1; N, 2.5),
Calculated for C23H17B2F8NOCl2, (C, 48.64; H, 3.0; N, 2.5). 1H NMR
(CDCl3/CF3COOH) d 4.6(s, 3H), 7.7(d, J = 9 Hz, 4H), 8.3(d, J = 9 Hz,
4H), 8.7(d, J = 7 Hz, 2H), 8.8(s, 2H), 9.0(d, J = 6 Hz, 2H). 13C NMR
(CDCl3/CF3COOH) 49.4(q, J = 147 Hz), 117.7(d, J = 177 Hz), 126.6(s),
128.3(d, J = 176 Hz), 130.7(d, J = 165 Hz), 131.8(d, J = 171 Hz), 146
(s), 147.3(d, J = 194 Hz), 149.6(s), 160.9(s), 173.5(s). IR (KBr) 3428,
1629, 1591, 1513, 1468, 1272, and 1088 cmꢀ1. UV–Vis kmax 270, 345 and
455 nm. ESI-MS m/z (%): 424.1(M+˙ + CH3Oꢀ, 100), 426.1 (63), 427.1
(18), 393.2. ESI-MSMS m/z (%): 393.1(M+˙, 100), 395.1(63), 394.1(25).
1,5-Di-4-trifluoromethylphenyl-3-(4-pyridyl)pentane-1,5-dione 1 (X =
CF3). Sodium hydride (0.2 g, 6 mmol) was added to 20 mL of vigor-
ously stirred methanol at 0°C followed by 0.7 g of 4-trifluoromethyl ace-
tophenone. Pyridine-4-carbaldehyde (0.535 g) is then added slowly
maintaining the temperature at 0–10°C. Water was added dropwise until
the solution became milky. The solution was kept stirring at room tem-
perature overnight and a solid precipitated out during the reaction. The
solution was filter and the filtrate was discarded. Crude product was puri-
fied by precipitation with MeOH/H2O to give a creamy soft powder
(0.3 g, 50%). 1H NMR (CDCl3) d 3.42(dd, J = 6.8, 17.5 Hz, 2H), 3.56
(dd, J = 6.9, 17.5 Hz, 2H), 4.09(quintet, J = 6.8 Hz, 1H), 7.24(d,
J = 6 Hz, 2H), 7.74(d, J = 8.1 Hz, 4H), 8.05(d, J = 8.1 Hz, 4H), 8.54(d,
J = 6 Hz, 2H).
CDCl3 (1:2))
d 3.45(dd, J = 7.8, 17.3 Hz, 2H), 3.52(dd, J = 6.3,
17.3 Hz, 2H), 4.28(s, 3H), 4.25(quintet, J = 6.5 Hz, 1H), 7.45(m, 4H),
7.65(m, 2H), 7.95(m, 4H), 8.25(m, 2H), 8.95(m, 2H). 13C NMR
(CF3COOH/CDCl3 (1:2)) d 36.4, 43.6, 47.6, 127.5, 128.4, 129.3, 134.0,
136.6, 145.1, 164.9, 198.0.
4-(2,6-Diphenylpyrylium-4-yl)-1-methylpyridinium
bistetrafluoro-
borate. (HPY2+
)
Boron trifluoride (10 mL) is added dropwise with
stirring to a solution of benzalacetophenone(0.5 g) and 3-(1-methyl-4-yl)-
1,5-diphenylpentane-1,5-dione iodide (0.84 g, 0.0020 mol) in 5 mL of
hot acetic acid. The mixture is refluxed for 12 h. After cooling to room
temperature and addition of diethyl ether (20 mL), crystallization
occurred. Recrystallization from acetic acid/acetonitrile afforded yellow
crystals. Yield: 0.66 g, 76%. 1H NMR (CF3COOH/CDCl3 (1:2)) d 4.58
(3H, s), 7.8(t, J = 8.4 Hz, 4H), 7.9(t, J = 7.6 Hz, 2H), 8.45(d,
J = 7.5 Hz, 4H), 8.71(d, J = 6.8 Hz, 2H), 8.81(s, 2H), 9.02(d,
J = 6.8 Hz, 2H). 13C NMR (CD3CN) d 50.0, 119, 128.4, 129.6, 130.4,
131.4, 137.6, 147.9, 149.4, 161, 174.0. Elemental analysis: Found (C,
53.7; H, 4.0; N, 2.8%), Calculated for C23H19B2F8NO (C, 55.4; H 3.8;
N, 2.8), Calculated for C23H19B2F8NOꢃH2O (C, 53.4; H, 4.1; N, 2.7%).
IR (KBr) 3090, 1630, 1520, 1430, 1270, 1210, 1090 cmꢀ1. ESI-MS m/z
(%): 356.4 (M+˙ + CH3Oꢀ, 100). ESI-MSMS m/z (%): 325.4(M+˙, 100).
1,5-Di-4-methoxyphenyl-3-(4-pyridyl)pentane-1,5-dione 1 (X = MeO).
Pyridine-4-carbaldehyde (0.75 mL) was added to a warm solution of
p-methoxy acetophenone (3.125 g) in 80% EtOH (12.5 mL) containing
potassium hydroxide (0.375 g). The solution turns dark immediately and
was heated under reflux (80°C) for 6 h. The solid product, which sepa-
rated upon cooling, was recrystallized twice from EtOH to give white
crystals. Yield: 1.3 g, 43%. 1H NMR (CD3CN) d 3.31(dd, J = 7.4,
16.9 Hz, 2H), 3.46(dd, J = 8, 16.9 Hz, 2H), 3.87(s, 6H), 4.06(quintet,
J = 7 Hz, 1H), 6.91–6.94(m, 4H), 7.25–7.27(m, 2H), 7.91–7.94(m, 4H),
8.49–8.50(m, 2H).
3-(1-Methylpyridium-4-yl)-1, 5- di-4-methoxyphenylpentane-1, 5-dione
Iodide 2 (X = MeO). To a solution of 0.595 g (0.00850 mol) of 1, 5-di-4-
methoxyphenyl-3-(4-pyridyl) pentane-1, 5-dione in chloroform (1 mL),
0.22 g (0.0068 mol) of methyl iodide was added. The solution is stirred for
half an hour at room temperature. After evaporation of chloroform at
20 mmHg, the red oil is triturated with diethyl ether and the crystals recrys-
tallized from ethanol. Yield: (0.53 g, 90%). 1H NMR (CD3CN) d 3.61–
3.62(m, 4H), 3.88(s, 6H), 4.21(small quintet under one big singlet, 4H),
7.00–7.03(m, 4H), 7.94–7.96(m, 2H), 8.03–8.05(m, 4H), 8.48–8.50(m,
2H).
4-(2,6-Di-4-methoxyphenyl pyrylium-4-yl)-1-methylpyridinium bistet-
rafluoroborate (MeOPY2+). Prepared by the method used to prepare
HPY2+ (vide supra). Yield: 0.3 g, 55%. 1H NMR (CD3CN) d 4.01 (s,
6H), 4.45 (s, 3H), 7.28–7.31 (d, J = 9.1 Hz, 4H), 8.41–8.43(d,
J = 9.1 Hz, 4H), 8.47 (s, 2H), 8.57–8.59(d, J = 6.8 Hz, 2H), 8.94–8.95
3-(1-Methylpyridium-4-yl)-1, 5- di-4-trifluoromethylphenyl -1, 5-dione
tetrafluoroborate 2 (X = CF3). A mixture of 1,5-di-4-trifluoromethylphe-
nyl-3-(4-pyridyl)pentane-1,5-dione (320 mg, 0.688 mmol) and trim-
ethyloxonium tetrafluoroborate (101 mg, 0.688 mmol) in 20 mL

Photochemistry and Photobiology, 2014, 90 347
anhydrous CH2Cl2 is stirred overnight at room temperature. After the
removal of the solvent, the crude product is used directly in the next step
(purity >95%). (360 mg, 95% yield). 1H NMR (CD3CN) d 3.66(dd,
J = 18.7, 8.0 Hz, 2H), 3.73(dd, J = 18.8, 5.4 Hz, 2H), 4.19(s, 3H), 4.22
(quintet, J = 5.8 Hz, 1H), 7.83(d, J = 8.3 Hz, 4H), 8.00(d, J = 6.6 Hz,
4H), 8.08(d, J = 8.3 Hz, 2H), 8.46(d, J = 6.8 Hz, 2H).
Benzalacetophenone, 3, in the cyclization step shown in
Scheme 3 functions as a hydride acceptor (oxidant) that is neces-
sary in order to convert a 2H-pyran, or possibly 4H-pyran,
intermediate into the pyrylium ring found in HPY2+. When this
reaction was utilized to make MeOPY2+ a peak at 460 nm in
the fluorescence spectrum that accompanied the emission of
MeOPY2+ at 653 nm was observed. We tentatively identify this
spurious emission to 2,4,6-triphenylpyrylium tetrafluoroborate
(kem = 465 nm) (13) that presumably forms by Michael addition
of acetophenone to benzalacetophenone followed by oxidative
cyclization. Apparently traces of acetophenone, only detectable
in the emission and not NMR spectrum, formed under the reac-
tion conditions by a minor background hydrolysis of benzalace-
tophenone, 3. On the other hand, a peak at approximately
540 nm was not observed for 2,4-diphenyl-6-(p-methoxyphenyl)
pyrylium tetrafluoroborate (kem = 538 nm) (13) even though
p-methoxyacetophenone is a potential product of retro-Michael
addition of either the 1,5-diketone or alkylated 1,5-diketone
intermediate in the formation of MeOPY2+. In order to prevent
formation of 2,4,6-triphenylpyrylium tetrafluoroborate we
replaced benzalacetophenone, 3, with trityl cation (Ph3C ⊕) as
the hydride acceptor for all the studies that required spectroscopi-
cally pure pyrylogen. The trityl cation was generated in situ by
treatment of triphenylcarbinol with HBF4 during the last (oxida-
tive cyclization) step (14). This strategy completely eliminated
the emission peak at 460 nm and provided verification of its ten-
tative assignment to 2,4,6-triphenylpyrylium tetrafluoroborate.
Another advantage of this oxidative cyclization procedure is that
it is readily adaptable to form pyrylogens with perchlorate
(HClO4/Ph3COH) or hexafluorophosphate (HPF6/Ph3COH)
counterions.
4-(2,6-Di-4-trifluoromethylphenyl pyrylium-4-yl)-1- methyl- pyridini-
um bistetrafluoroborate (CF3PY2+). 3-(1-Methylpyridium-4-yl)-1,5-di-p-
trifluoro-methylphenylpentane-1,5-dione
tetrafluoroborate
(0.38 g,
0.67 mmol) is dissolved in 10 mL of acetic anhydride and 0.436 g of tri-
phenyl methanol (1.67 mmol) is added to the solution. A 50% HBF4
solution, which was prepared by slowly adding 48% HBF4 in H2O to
10 mL of cold acetic anhydride, was then added dropwise to the rapidly
stirred dione solution. The mixture was kept at 80°C for half an hour fol-
lowed by addition of 100 mL diethyl ether to the cooled solution to
induce precipitate formation. The precipitate was filtered and dried under
vacuum at 100°C. (200 mg, 47% yield). 1H NMR (CDCl3/CF3COOH) d
4.59(s, 3H), 8.05(d, J = 8.4 Hz, 4H), 8.49(d, J = 8.4 Hz, 4H), 8.74(d,
J = 6.4 Hz, 2H), 8.99(s, 2H), 9.00(d, J = 6.4 Hz, 2H). 13C NMR
(CDCl3/CF3COOH;decoupled and gated coupled data)
d 49.4(q,
J = 146 Hz), 119.6(d, J = 178 Hz), 123.3(q, 273 Hz), 128(d,
3
J = 169 Hz, JCF = 3 Hz), 128.2(d, J = 167 Hz), 130(d, J = 166 Hz),
131(s), 138.9(q, J = 34 Hz), 147.1(d, J = 195 Hz), 149(s), 162.4(s),
173.7(s). Elemental analysis: found (C, 46.7; H, 2.95; N, 2.16), Calc. for
C25H17B2F14NO (C, 47.3; H, 2.7; N, 2.5), Calc. for
C25H17B2F14NOꢃH2O (C, 45.98; H, 2.93; N, 2.14). IR(KBr): 3124, 1631,
1528, 1322 cmꢀ1. ESI-MS m/z (%): 492.1 (M+˙ + CH3Oꢀ, 100). ESI-
MSMS m/z (%): 461.1(100).
3-(1-methylpyridium-4-yl)-1, 5- diphenylpent-2-ene -1, 5-dione tetra-
fluoroborate. An aqueous solution of 10ꢀ4 M NaOH was added to a stir-
red mixture of PY2+ until the pH reaches 6. The mixture is then stirred
for an additional 0.5 h and the pale yellow precipitate was filtered and
recrystallized twice in hot water to give very low yield (15%) of the
PY2+–H2O-adduct. 1H NMR (CD3CN) d 8.63(d, 2H, J = 6.7 Hz), 8.17
(d, 2H, J = 6.7 Hz), 8.05–8.01(m, 4H), 7.77(s, 1H), 7.71–7.65(m, 2H),
7.59–7.53(m, 4H), 4.77(s, 2H), 4.29(s, 3H). 13C NMR (CD3CN;
decoupled and gated coupled data) d 197.33(s), 191.87(s), 158.33(s),
146.43(d, J = 192 Hz), 145.93(s), 138.61(s), 137.44(s), 135.02(d, J =
162 Hz), 134.87(d, J = 162 Hz), 132.11(d, J = 159 Hz), 130.03(d,
J = 164 Hz), 130.00(d, J = 164 Hz), 129.77(d, J = 161 Hz), 129.32(d,
J = 161 Hz), 126.54(d, J = 173 Hz), 48.90(q, J = 145 Hz), 42.15(t,
J = 130 Hz). Elemental analysis: C23H20BF4NO2, (C, 64.36%; H,
4.70%; N, 3.26%), Found: (C, 64.35%; H, 4.65%; N: 3.22%).
General sensitized irradiation conditions. Acetonitrile solutions con-
taining between 5 9 10ꢀ4 and 1 9 10ꢀ3 M of HPY2+BF4ꢀ˙and 0.1 to
0.02 M substrate (see Fig. 7) were placed in pyrex flasks open to the
atmosphere. These flasks were then irradiated with UVA lamps, tungsten
halogen lamps, or medium-pressure Hanovia lamps, through 1 cm of a
75% aqueous NaNO2 solution to insure elimination of all photons higher
in energy than 420 nm. The reactions were then monitored by thin layer
chromatography, gas chromatography, or NMR and generally were com-
plete within 10 min to an hour. Preparative reactions were worked up by
first washing with water to remove the charged hydrolysis product of the
sensitizer. Then the organic layer was dried (MgSO4) and removed by
rotovap to give the products that were characterized by standard analyti-
cal techniques.
The structure of HPY2+ was confirmed with a crystal structure
of its bis-tetra-perfluorophenyl borate, (C6F5)4Bꢀ salt (Fig. 1).
This counterion was used sinceꢀrepeated attempts to get X-ray
quality crystals of HPY2+ꢃ2BF4 failed. The tetra-perfluorophe-
nyl borate counterion was not incorporated in the pyrylogen dur-
ing the oxidative cyclization step as described above. Instead, an
ion exchange process that takes advantage of the enhanced solu-
bility of the (C5F6)4Bꢀ pyrylogens was used. When a mixture of
RESULTS AND DISCUSSION
Synthesis and structural characterization
We initially adapted
a method used extensively for the
synthesis of pyrylium cations, as shown in Scheme 3, for the
construction of HPY2+ (12). The reaction is initiated by a base
catalyzed condensation between acetophenone and pyridine-
4-carbaldehyde to give an a-, b-unsaturated ketone that
immediately undergoes
a Michael addition with a second
equivalent of acetophenone to give 1,5-diketone 1 (Scheme 3).
Alkylation of this ketone with methyl iodide gives a pyridini-
um cation, 2, that is subsequently oxidatively cyclized to give
the pyrylogen.
Scheme 3. Synthesis of pyrylogens.

348 Edward L. Clennan and Chen Liao
ꢀ
the insoluble salts, HPY2+ꢃ2BF4 and (C5F6)4BꢀK+, was stirred
the ferrocene–ferrocenium (16) ion redox couple (17). The first
reduction potentials spanned a range of nearly one-third of a volt
increasing in the following order MeOPY2+ (+0.05 V) < HPY2+
(+0.17 V) < ClPY2+ (+0.19 V) < CF3PY2+ (+0.37 V) all versus
the standard calomel electrode (SCE). Remarkably, in dramatic
contrast to the reduction to the radical-cation, the second reduc-
tion to the neutral redox couple was nearly insensitive to the sub-
stituent effect and only spanned a range of 0.15 V increasing in
the order MeOPY2+ (ꢀ0.42 V) < HPY2+ (ꢀ0.35 V) ClPY2+
(ꢀ0.35 V) < CF3PY2+ (ꢀ0.27C) all versus SCE. These results,
and for comparison, a series of substituted 2,6-diphenyl-4-X-phe-
nyl-pyrylium salts, XTPP+ can be viewed quantitatively by relat-
ing them to the substituent constants rp+ as shown in Fig. 2. The
effect of substituents on the reductions of XTPP+ and
XPY2+ꢃ2BF4 is similar (q+ = 0.13 and 0.16) but more than twice
as large as that observed for reduction of XPY+˙ (q+ = 0.061).
The reduced effect of substituents on the reduction potentials of
under a blanket of CH2Cl2 the methylene chloride gradually
turned yellow as ion exchange occurred. Simple filtering of the
KBF4 and removal of the methylene chloride then gave pure
HPY2+ꢃ2(C6F5)4B−.
The dication is nonplanar with a pyridinium-pyrylium ring
dihedral angle of 53.2° and phenyl-pyrylium ring dihedral angles
of 10.2°. The two counterions sandwich the dication (Fig. 1) and
ꢀ
exhibit close contact distances of 2.66 and 2.68 A between the
hydrogen on C2, and 2.78 A between the hydrogens on C5, and
fluorines on the perfluorophenyl rings. A similar geometry of the
ion pair may also exist in solution since it is the hydrogens on
C2 (8.28 ppm) and C5 (8.36 A) that exhibit the laꢀrgest H NMR
ꢀ
1
ꢀ
upfield shifts of 0.52 and 0.24 ppm when the BF4 counterion is
replaced by (C6F5)4Bꢀ. The crystal structure geometry is remark-
ably well reproduced by density functional theory (DFT) calcula-
tions (15) using the B3LYP/6-311+G(2d,p) basis set with the
exceptions of the pyridinium- (43.1°) and phenyl-pyrylium
(16.2° and 16.7°) ring dihedral angles. The barrier for twisting
around the single bond connecting the pyridinium and pyrylium
rings at the B3LYP/6-31G(d) computational level is only about
2.1 kcal molꢀ1 and as a consequence the electronically preferred
intra-ring dihedral angles could easily be sacrificed for favorable
crystal packing energy. (see Supplementary Material for com-
plete computational details).
XPY+˙ in comparison to those observed for the dication, XPY2+
,
is consistent with a Mulliken population analysis that reveals that
the pyrylium ring in the dication bears a much greater positive
charge than in the radical-cation (vide infra). It is also instructive
to compare the reduction potentials of XTPP+ (X = H;
ꢀ0.35 V) to HPY2+ (+0.17 V), which demonstrates that replace-
ment of a phenyl ring with a pyridinium ring increases the ease
of reduction by over 0. 5 V.
The radical cation formed during reduction of HPY2+ꢃ2BF4
ꢀ
did not react with oxygen on the CV time scale (50–500 mV sꢀ1
)
Pyrylogen redox couple
in contrast to the rapid reaction of the radical formed upon reduc-
tion of 2,4,6-triphenylpyrylium tetrafluoroborate. This is most
likely a result of the enhanced electrophilic character of HPY+˙ in
comparison to XTPP˙ (X = H).
Insight into the electronic character and potential reactivity of
all three redox partners was obtained by partitioning the charge
or charge/spin on the pyrylogens between the diarylpyrylium
(red in Scheme 4) and N-methylpyridinium (blue in Scheme 4)
functional groups using Mulliken population analysis and the
B3LYP/6-311+G(2d,p) minimized geometries. In all the dica-
tions, XPY2+, the positive charge is greater on the diarylpyryli-
um (+1.239 ꢄ 0.083) than on the N-methylpyridinium
(+0.760 ꢄ 0.083) groups. This is consistent with the enhanced
In analogy to the isoelectronic viologens it was anticipated that
pyrylogens would exist in three readily accessible redox states
(Scheme 4). This contention is supported by computational stud-
ies at the B3LYP/6-311+G(2d,p) level that located the dication,
radical cation and neutral redox partners. In the computed HPYx
structures the bond connecting the pyridinium and pyrylium rings
ꢀ
decreases in length from 1.486 to 1.432 to 1.383 A and the
pyridinium–pyrylium ring dihedral angle decreased from 43.1° to
3.2° to 0.4° along the sequence dicationꢃradical-cationꢃneutral as
expected from the Lewis structures shown in Scheme 4.
The reduction potentials for the pyrylogens, XPY2+, were
determined in acetonitrile at a glassy carbon electrode relative to
Figure 1. View of [C23H19NO] 2 [B(C6F5)4]. Thermal ellipsoids are draw at 30% probability. Hydrogen atoms are omitted for clarity.

Photochemistry and Photobiology, 2014, 90 349
Scheme 4. Pyrylogen dication, radical-cation and neutral redox states.
0.5
0.4
0.3
0.2
0.1
0
XTPP Reduction
XPY2+ Reduction
XPY+. Reduction
N
y = 0.1298x + 0.1965
= 0.96494
R
O
X
X
X
2
XPY
X = MeO, H, Cl, CF
3
-0.1
-0.2
-0.3
-0.4
-0.5
y = 0.1631x - 0.2884
R
= 0.98124
O
XTPP
X = H, Me, MeO, CN
y = 0.061x - 0.3468
R
= 0.98905
-1.5
-1
-0.5
0
0.5
1
1.5
p+
Figure 2. Reduction potentials as a function of Σrp+
.
tetrafluoroborate, TPP+ꢃBF4 (13), in Fig. 3 and Table 1. All
the XPY2+ spectra, and for comparison TPP+ were taken in
CH3CN since XPY2+ had limited solubility in less polar sol-
vents. In order to explore the effect of solvent on the XPY2+
ꢀ
opportunity to delocalize the positive charge when placed on the
diarylpyrylium (resonance structures A, B and C; Scheme 5) in
comparison to the N-methylpyridinium (resonance structure D;
Scheme 5) groups. In all the radical cations, XPY+˙, the charge
is greater on the N-methylpyridinium (+0.748 ꢄ 0.056) than on
the diarylpyrylium (+0.252 ꢄ 0.056) group but the spin is
greater on the diarylpyrylium (0.558 ꢄ 0.045) than on the N-
methylpyridinium (0.442 ꢄ 0.045) group. This is consistent with
the greater importance of resonance structure E than F in the
overall electronic description of XPY+˙. In all the neutral redox
ꢀ
absorption spectra the BF4 counterions had to be exchanged
to give XPY2+ꢃ2[B(C6F4)4]ꢀ. For example, HPY2+ꢃ2[B
(C6F4)4]ꢀ is soluble in CH2Cl2 and absorbs at 476 nm but
shifts to 441 nm when dissolved in the more polar CH3CN.
This hypsochromic shift with increased solvent polarity is
consistent with stabilization of a more polar ground state in
preference to a less polar excited state.
partners, XPY0, polarization occurs to place
a negative
ꢀ0.284 ꢄ 0.063 charge on the diarylpyrylium and a positive
+0.284 ꢄ 0.063 charge on the N-methylpyridinium group. This
Mulliken population analysis reflects the greater contribution of
resonance structure G (Scheme 5) than resonance structure H to
the description of the neutral redox partner. This CS (polariza-
tion) is consistent with lower electronegativity of nitrogen than
oxygen and its ability to accommodate a positive charge.
The absorption spectra of alkyl- and aryl-substituted pyrylium
cations were first examined in detail by Balaban et al. in 1960
(18). This study and several subsequent studies (19) that focused
on the effect of substituents (20) on the appearance of the
absorption spectra (21) came to two key conclusions (22): (1)
The absorption spectra of pyrylium cations can be characterized
by two chromophores one with its transition dipole oriented
along the pseudo-x-axis passing through carbons 2 and 6 and the
other lying along the C2 axis passing through the aryl group on
carbon 4 and the oxygen atom (23); (Scheme 6) and (2) Both of
these transitions are charge transfer transitions involving migra-
tion of electron density from the aryl rings to the pyrylium core.
Photophysical properties
ꢀ
The UV–Vis spectra for the pyrylogens, XPY2+ꢃ2BF4 are
compared to the closely related 2,4,6-triphenylpyrylium

350 Edward L. Clennan and Chen Liao
Scheme 5. Resonance structure rationalization of Mulliken charge and charge/spin distributions.
Examples of how substituent effects can provide evidence for
these two conclusions are illustrated in Scheme 6.
methoxy group is added to 2,6-[NMe2]2TPP+ (477–482 nm) to
give 4-MeO-2,6-[NMe2]2TPP+ because the electronic demand
for charge transfer to the pyrylium core is significantly dimin-
ished by the strong electron donating aryl groups at the 2,6-posi-
tions. On the other hand, when there are no electron donating
groups appended to the 2-chromophore addition of the potent
electron donating dimethylamino group to the 4-phenyl of TPP+
to generate 4-NMe2TPP+ generates an enormous 181 nm batho-
chromic shift to 552 nm so that the y-band now becomes the
lowest energy band in the UV–Vis spectrum. Addition of two
less potent electron-donating p-methoxy groups to the 2-chromo-
phore of 4-NMe2TPP+ generates a 61 nm bathochromic shift of
the x-band and a 9 nm hypsochromic shift of the y-band but (see
2,6-diMeO-4-NMe2TPP+) but does not reverse the assignment
of the lowest energy absorbance to the y-chromophore.
(Scheme 6).
The x-band (chromophore) is the lowest energy absorbance in
TPP+ appearing at 404 nm in acetonitrile (Fig. 3) and at 421 nm
in methylene chloride. The y-band is more intense and appears at
357 nm in acetonitrile and 371 nm in methylene chloride. Addi-
tion of the electron donating dimethylamino group to the para-
carbon of the 2-phenyl group to generate 2-NMe2TPP+ raises
the 2-chromophore orbital energy and reduces the HOMO–
LUMO gap shifting the x-band by 156 nm from 421 to 577 nm
(24). The y-band experiences a much smaller hypsochromic shift
from 371 to 359 nm. The bathochromic shift of the x-band is
magnified when a 2nd dimethylamino group is added to the
2-chromophore to generate 2,6-[NMe2]TPP+ (Scheme 6). On the
other hand, addition of a nitro group to the 2-phenyl ring to gen-
erate NO2TPP+, has the opposite effect of lowering the HOMO
energy and increasing the HOMO–LUMO gap and shifting the x-
band hypsochromically to 410 nm (25). Addition of an electron-
donating group to the y-chromophore also has the anticipated
effect of raising the y-chromophore energy and shifting the
y-band to lower energy (26). This effect is modest when a
In the pyrylogens, XPY2+, the donor part of the y-chromo-
phore in TPP+, the 4-phenyl ring, has been replaced with an
electron poor N-methylpyridinum group. Consequently, we antic-
ipated that the low energy band in XPY2+ would correspond to
an x-chromophore transition and that a y-band would be either at

Photochemistry and Photobiology, 2014, 90 351
Figure 3. UV–Vis spectra for TPP+ (dotted line) and XPY2+ (solid lines) in CH3CN.
much higher energy or absent. This speculation was confirmed
by time domain density functional calculations (TD-DFT) in
acetonitrile using B3LYP/6-311+G(2d,p) calculations in conjunc-
tion with the polarization continuum model as shown in
Scheme 7. This model chemistry does an excellent job of repro-
ducing the experimental absorption spectra of TPP+. The HOMO
? LUMO transition in TPP+ is responsible for the low energy
absorption at 410 nm (404 nm experimentally) and is clearly an
x-band transition involving charge transfer from the 2,6-aryl-
rings to the pyrylogen core. The HOMO-1 ? LUMO (y-band)
transition is responsible for the peak at 370 nm (357 nm experi-
mentally) and involves charge transfer from the 4-phenyl ring to
the pyrylium ring. In HPY2+, the low energy band at 467 nm
(441 nm experimentally) is also the HOMO ? LUMO transition
and an x-band. In fact, no low energy y-band was detected
computationally. Two bands with low oscillator strengths
(f = 0.0034, 0.0111) appearing at approximately 400 nm and a
band with modest oscillator strength (f = 0.1896) at 361 nm are
both x-band transitions. The agreement between the computed
and observed wavelength of the low energy x-band, Dnm, is
much better for TPP+ (Dnm = 6 nm) than for HPY2+
(Dnm = 26 nm). We believe this reflects the much greater charge
transfer observed from the 2,6-phenyl-rings to the dicationic
pyrylogen than mono-cationic pyrylium core. It is well estab-
lished that charge transfer bands are poorly predicted with TD-
DFT as a result of their incorrect long-range form of the
exchange potential (27). Despite this problem we do believe that
the TD-DFT calculations for HPY2+ are qualitatively correct.
Table 1. Photophysical data for XPY2+ꢃ2BF4 and TPP+.
ꢀ
HPY2+
ClPY2+
CF3PY2+
413
MeOPY2+
523
TPP+
401
355
24 920
31 700
410
kMAX
e*
*
441
454
21 010
467
24 030
500
21 770
440
20 640
575
0.4225
kMAX
f†
†
370
0.4156
533
578 (565)
92
0.5186
0.4989
494
552 (546)
81
0.4790
0.6415
465
520
64
kF‡
kP§
556
588 (588)
102
659
658 (665)
136
Stokes shift
sFk
3.42 ꢄ 0.23
–
2.1 ꢄ 0.21
–
4.1 (2.9)††
225 ꢄ 3‡‡
sP¶
35.6 ꢄ 0.2
(91 ꢄ 4)
59
38.3 ꢄ 1
(109 ꢄ 3)
56
33.9 ꢄ 0.2
25.4 ꢄ 0.6
(214 ꢄ 57)
(55 ꢄ 1)
E(S1)#
E(T1)#
ΦF**
ΦTkk
63
48
48
§§
–
65‡‡
53‡‡
0.60
–
54
53
57
0.18 ꢄ 0.01
0.43 ꢄ 0.01
0.24 ꢄ 0.01
0.03 ꢄ 0.02
–
–
*Experimental absorption spectrum in CH3CN; e molꢀ1 Lꢀ1 cmꢀ1. †TD-DFT B3LYP/6-311+G(2d,p) calculations in CH3CN; f-oscillator strength. ‡In
nanometers (nm) at 298 K. §In nm in EtOH/HCl(g) and in parenthesis nPrCN/HCl(g) at 77 K. kIn nanoseconds (ns). ¶In milliseconds (ms) at 77 K in
EtOH/HCl(g) and in parenthesis in nPrCN/HCl(g). #In kcal molꢀ1. **Fluorescence quantum yield in CH3CN. ††Haucke, G., P. Czerney, F. Cebulla
(1992) Absorption and Fluorescence of Pyrylium Salts. Ber. Bunsenges. Phys. Chem., 96, 880–886. CH3CN(CH2Cl2). ‡‡Miranda, M. A., H. García
(1994) 2,4,6-Triphenylpyrylium Tetrafluoroborate as an Electron-Transfer Photosensitizer Chem. Rev., 94, 1063–1089. §§Too small to be measured accu-
rately. kkTriplet quantum yield.

352 Edward L. Clennan and Chen Liao
Scheme 6. Effect of substituents on positions of pyrylium cation x- and y-bands.
The calculation successfully predicts the substituent induced shift
in the x-band and a plot of the observed (x-axis) versus the TD-
DFT calculated (y-axis) x-band wavelength gives a straight line
(slope = 1.23; R2 = 0.98604).
increased from 77 to 202 K. We attribute this shift to comparable
rate constants for emission and solvent reorganization leading to
emission from a continuum of slightly divergent solvated excited
states; an explanation that was previously utilized (29) to explain a
similar spectral shift that had been observed for TPP+ (30). Emis-
sion from a Franck–Condon state at low temperature and a relaxed
excited state at higher temperatures is not a viable explanation for
this phenomenon since the peak width at half height and the excita-
tion spectrum did not change between 77 and 202 K.
The fluorescence of all the pyrylogens, XPY2+, at room temper-
ature in acetonitrile is characterized by a single peak devoid of any
fine structure at the wavelengths depicted in Table 1. The fluores-
cence quantum yields are much smaller (Table 1) than the 0.60
reported for TPP+. It is tempting to suggest that the low fluores-
cence quantum yields are due to the absence of the
y-chromophore in XPY2+. This speculation is based on a very
novel study of TPP+ which demonstrated that incarceration in the
supramolecular host cucurbit[8]uril restricted the motion of the
2,6-phenyl groups in the x-chromophore and allowed simultaneous
observation of fluorescence and room temperature phosphores-
cence (28). Furthermore, irradiation into the y-band increased, and
irradiation into the x-band decreased, the room temperature ΦF/ΦP
ratio. On the other hand, intramolecular charge transfer (ICT)
quenching of the fluorescent state, especially in MeOPY2+, cannot
be unambiguously ruled out. The formation of an unusual radical-
cation/cation state, (CT-MeOPY2+)* with a near planar geometry
(Scheme 8) reminiscent of pyrylogen radical-cations (vide supra)
could explain both the unusually small ΦF and unusually large
Stokes shift observed for MeOPY2+ (Table 1).
The emission wavelength for XPY2+ꢃ2[B(C6F4)4]− also shifted
bathochromically by ~30 nm as the solvent polarity was
decreased from acetonitrile to 1,2-dichloroethane. Unfortunately,
XPY2+ꢃ2[B(C6F4)4]− does not exhibit sufficient solubility in the
widespread range of solvents necessary to construct a statistically
meaningful (31) Lippert–Mataga plot (32). Nevertheless, the
bathochromic shift with decreasing solvent polarity is consistent
with the previous conclusion from absorption spectra (vide
supra) that the dipole moment of pyrylogen excited states are
smaller than the dipole moment of their ground state.
−
Phosphorescence was observed at 77 K for XPY2+BF4 in a
variety of low temperature glasses including butyronitrile/HCl(g)
and EtOH/HCl(g) (Table 1). The acid was added to decrease the
nucleophilic character of the medium. The stability of the prylo-
gens under these conditions was demonstrated by UV–Vis spec-
troscopy. Identical phosphorescence wavelengths were observed
for ClPY2+BF4− in both glasses while minor environmental effects
The fluorescence peak for HPY2+ in an EtOH/HCl matrix
shifted bathochromically by 36 nm as the temperature was

Photochemistry and Photobiology, 2014, 90 353
Scheme 7. Low energy electronic transitions in TPP+ and HPY2+
.
Scheme 8. Electronic distribution in the intramolecular charge transfer (ICT) state.
were observed for the other pyrylogens (Table 1). On the other
reactivity of the pyrylogen radical cation with oxygen on the much
longer time scale of a CV experiment eliminates it as a candidate
for the peak at 570 nm. Instead, the transient signal was assigned
to the triplet–triplet absorption spectrum based upon its dramatic
quenching with oxygen (inset in Fig. 4).
hand, triplet lifetimes were dramatically different and systemati-
cally longer in butyronitrile than in EtOH (Table 1) suggesting that
the nucleophilic reaction of ethanol (vide infra) with the pyrylo-
gens was not completely suppressed by the addition of HCl. The
triplet energies in Table 1 were determined from the onset of the
phosphorescence and were identical within ꢄ0.2 kcal molꢀ1 in
the two glasses. Formation of the triplet was independently verified
during a nanosecond laser flash photolysis study of HPY2+BF4− in
1,2-dichloroethane. Irradiation with 10 mJ 355 nm pulses from a
Nd/YAG laser generated bleaching of the pyrylogen and observa-
tion of a transient absorption at 570 nm (Fig. 4). The absence of
Reactions
The pyrylogens are potent oxidants in both the S1 and T1 excited
states. The reduction potentials of the excited states suggest that
the ease of oxidation of a substrate, M, [DGo = E0(M/M+˙)ꢀ
E0(XPY2+/XPY+˙)ꢀE(S1 or T1)] by the S1 or T1 excited state will

354 Edward L. Clennan and Chen Liao
0.05
0.04
0.03
0.02
0.01
0
Series1
Series2
350
400
450
500
550
600
650
700
-0.01
-0.02
-0.03
Wavelength(nm)
Figure 4. Laser flash photolysis of 1.1 9 10ꢀ4 M HPY2+ in 1,2-dichloroethane. Inset: Effect of oxygen on the decay of the peak at 570 nm.
Scheme 9. Hydrolytic ring opening of HPY2+
.
increase in the order 3(TPP+)* [1.95 eV] < 1,3(MeOPY2+)*
1
3
3
[2.13 eV] < (TPP+)*[2.47 eV] < (ClPY2+)*[2.49 eV] < (HPY2+)*
1
1
[2.51 eV] < (ClPY2+)* [2.62 eV] < (HPY2+)* [2.73 eV]
<
3(CF3PY2+)* [2.84 eV] < (CF3PY2+)* [3.10 eV]. The ground states,
So, of the pyrylogens are also very electron deficient and reactive,
consequently, it is important to be aware of potential reactions of
So when designing PET reactions using these sensitizers. For
example, HPY2+ reacts rapidly with water at the pyrylium ring to
generate a ring-opened water adduct (Scheme 9) and care must
be expended to scrupulously dry PET reaction solvents.
1
Figure 5. Charge transfer band formation between CF3PY2+ and
PDMB.
Pyrylogens also form charge transfer complexes (33) (CT) that
could potentially function as precursors to an electron transfer
event (34). Upon addition of 1,4-dimethoxybenzene (PDMB) to
a solution of CF3PY2+ a new peak was observed at 595 nm that
increased in intensity with increasing concentration of PDMB
(Fig. 5). Charge transfer complexes of the other pyrylogens with
PDMB are not as readily apparent and are only observed as a
broad shoulder on the long wavelength side of the pyrylogen
UV–Vis spectrum. This hypsochromic shift with decreasing ther-
modynamic stability of the complex is a well-established phe-
nomenon and serves as confirmation for the assignment of these
bands to CT complexes (35). The long-wavelength edges of the
CT bands in conjunction with a Benesi–Hildebrand analysis (36)
demonstrates that these are weak complexes (0.2 < Keq < 1).
The Keq value for CF3PY2+ꢃꢃꢃPDMB (0.63 ꢄ 0.09) is larger
than that measured for the CT complexes of HPY2+ (0.47 ꢄ
0.02), ClPY2+ (0.40 ꢄ 0.05), or MeOPY2+ (0.56 ꢄ 0.02) and is
probably more accurate (37), despite its larger error limits (38),
because it forms a well defined peak clearly separated from the
CF3PY2+ absorbance (39).
Figure 6. Reduction of HPY2+ with Cd and Ni in acetonitrile.
Pyrylogens will also react with several different chemical
reducing agents. For example, when an acetonitrile solution of
−
HPY2+ꢃ2BF4 is allowed to stir over metallic nickel or cadium a
new peak appears in the UV–Vis spectrum at 614 nm (Fig. 6).

Photochemistry and Photobiology, 2014, 90 355
Figure 7. Photoinduced electron transfer reactions catalyzed by HPY2+
.
We assign this peak to the radical cation, HPY+˙, based upon
several pieces of evidence: (1) the same peak is formed during a
coulometric reduction of HPY2+ꢃ2BF4ꢀ; (2) a single broad peak
is observed in the ESR spectrum of solutions containing the peak
at 614 nm; (3) the closely related dimethyl viologen radical cat-
ion appears at 620 nm (40); and (4) a time dependent density
functional theory calculation for HPY+˙ at the B3LYP/6-311+G
(2d,p) computational level predict a peak at 594 nm with an
oscillator strength of 0.4049. Parenthetically, the appearance of
this peak at 614 nm provides additional evidence for the veracity
of the assignment of the peak at 570 nm in the laser flash pho-
tolysis study as the triplet–triplet absorption.
Finally, we did a brief study of a series of well-established
PET reactions in order to determine if HPY2+ could function
as a catalyst as shown in Fig. 7. The reaction of the 1,3-dithi-
ane, 4 (41), and adamantane adamantylidene, 5 (42), are both
electron transfer oxygenation reactions that gave only the prod-
ucts indicated along with unreacted starting material. These
reactions were also accompanied by extensive decomposition
of the pyrylogen sensitizer. On the other hand the cycloaddi-
tion of 1,3-cyclohexadiene (43) occurred with 95% conversion
in one-hour with less than 5% decomposition of the pyrylogen
sensitizer. The formation of transient HPY+˙ and stilbene radi-
cal cations were directly confirmed by UV–Visible spectros-
copy and by laser flash photolysis in the 1,3-cyclohexadiene
cycloaddition and the cyclobutane cycloreversion reactions
(44), respectively.
enhanced triplet-triplet absorption of 3(HPY2+)* at 570 nm is
observed in the presence of 1-bromobutane but only radical cat-
ion (HPY+˙) formation, at ~620 nm, is observed in the presence
of bromobenzene. This suggests that 1-bromobutane promotes
singlet-triplet intersystem crossing 1(HPY2+)* ? 3(HPY2+)*
while bromobenzene undergoes an PET reaction.
Furthermore, the ability of 1-bromobutane to promote inter-
system crossing provides us with an additional mechanistic tool
to measure the quantum yield of triplet formation, ΦT. This
method, first used by Horrocks et al. (45), compares the ability
of 1-bromobutane to quench S1 to its enhancement of triplet for-
mation as shown in Eq (1). The virtue of this method is that
absolute quantum yields do not need to be measured but only
the ratios of the amount of triplet formed in the presence and
absence of 1-bromobutane, ΦT/U0T, and the quenching ratios in
Stern–Volmer experiments, U0f /Φf, are needed to plot the quan-
tity [U0f /Φf ꢀ 1] versus the term in the parenthesis on the left
side of equation 1 to determine U0T. The ratio ΦT/UT0 is conve-
niently determined by taking the ratio of integrated triplet-triplet
absorption decay curves at 570 nm in the laser flash photolysis
experiment in the presence and absence of the different concen-
trations of 1-bromobutane. In the case of HPY2+, the quantum
yield of triplet formation, U0T = 0.033 ꢄ 0.02, is significantly
smaller than that reported for TPP+, UT0 = 0.48 (12).
ꢀ
ꢁ
UT U0f
Uf0
Uf
x
ꢀ 1 U0T ¼
ꢀ 1
ð1Þ
UT0
Uf
Mechanistic tools
CONCLUSION
Both spectral identification of the pyrylogen radical cation and
triplet formation provides mechanistic tools that can be used to
follow pyrylogen sensitized reactions. For example, both 1-bro-
The synthesis of a series of pyrylogen sensitizers and preliminary
evidence that they catalyze PET reactions has been presented.
These promising results along with electrochemical and photo-
physical characterizations of the pyrylogens should enhance
mobutane and bromobenzene quench the fluorescence of HPY2+
,
however, in the nanosecond time-resolved emission spectra

356 Edward L. Clennan and Chen Liao
Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A.
Liashenko, P. Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T.
Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M. Challa-
combe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C.
Gonzalez and J. A. Pople (2004) Gaussian 03. Revision C.02 ed.,
Gaussian Inc, Wallingford, CT.
design of PET reactions using these new sensitizers. In addition,
several mechanistic tools that have been introduced should assist
in the analysis and optimization of new PET reactions using
these sensitizers.
12. Katritzky, A. R., J. Adamson, E. M. Elisseou, G. Musumarra, R. C.
Patel, K. Sakizadeh and W. K. Yeung (1982) Kinetics and mecha-
nisms of nucleophilic displacements with heterocycles as leaving
groups. Part 4. 2,4,6-Triaryl-N-benzylpyridinium cations: rate varia-
tion with electronic effects in the leaving group. J. Chem. Soc. Per-
kin Trans. II, 1041–1048.
Acknowledgements—We gratefully acknowledge the support of the
National Science Foundation (CHE-0646612 and CHE-1147542) for their
generous support of this research. We also thank Peter Ogilby and Mette
Johnsen (Aarhus University Denmark) for fluorescence lifetime
measurements.
13. Miranda, M. A. (1994) 2,4,6-Triphenylpyrylium tetrafluoroborate
as an electron-transfer photosensitizer. Chem. Rev. 94, 1063–
1089.
14. Branchi, B., G. Doddi and G. Ercolani (2005) Study of the reaction
of bis- and tris-pyrylium and -thiopyrylium cations with methoxide
ion. Electronic effects of the heterocyclic rings as substituents.
J. Org. Chem. 70, 6422–6428.
SUPPORTING INFORMATION
Additional Supporting Information may be found in the online
version of this article:
Data S1: Supplementary Material Computational Results:
Table S1. B3LYP/6-311+G(2d,p) Structural Parameters for 4,4′-
Pyrylogen Dications.; Table S2. B3LYP/6-311+G(2d,p)
Structural Parameters for 4,4′-Pyrylogen Radical Cations.; Table
S3. B3LYP/6-311+G(2d,p) Structural Parameters for Neutral
4,4′-Pyrylogens.; TD-DFT Results.
15. Seminario, J. M. (1996) Recent Developments and Applications of
Modern Density Functional Theory. Elsevier, New York.
16. Pavlishchuk, V. V. and A. W. Addison (2006) Conversion constants
for redox potentials measured versus different reference electrodes in
acetonitrile solutions at 25°C. Inorg. Chim. Acta 298, 97–102.
ꢁ
17. Gagne, R. R., C. A. Koval and G. C. Lisensky (1980) Ferrocene as
an internal standard for electrochemical measurements. Inorg. Chem.
19, 2854–2855.
18. Balaban, A. T., V. E. Sahini and E. Keplinger (1960) Electronic
absorption spectra of alkyl- and phenyl-substituted pyrylium salts.
Tetrahedron 9, 163–174.
19. Alvaro, M., C. Aprile, E. Carbonell, B. N. Ferrer and H. Garcia
(2006) Synthesis and photophysical properties of the 4-(biphenyl-4-
yl)-2,6-bis(4-iodophenyl)pyrylium ion. Eur. J. Org. Chem., 2644–
2648.
20. Manoj, N. and K. R. Gopidas (1997) Structure-photophysics correla-
tion in a series of 2,6-dimethyl-4-arylpyrylium derivatives. Chem.
Phys. Lett. 267, 567–572.
21. Manoj, N., G. Ajayakumar, K. R. Gopidas and C. H. Suresh (2006)
Structure absorption spectra correlation in a series of 2,6-dimethyl-4-
arylpyrylium salts. J. Phys. Chem. A 110, 11338–11345.
22. Balaban, A. T., G. W. Fischer, A. Dinculescu, A. V. Koblik, G. N.
Dorofeendo, V. V. Mezheritskii and W. Schroth (1982) Pyrylium
salts: synthesis, reaction, and physical properties. In Advances in
Heterocyclic Chemistry, Vol. Suppl. 2 (Edited by A. R. Katritzky),
pp. 434. Academic Press, New York.
REFERENCES
1. Kavarnos, G. J. (1993) Fundamentals of Photoinduced Electron
Transfer. VCH Publishers Inc., New York, NY.
2. Fox, M. A. and M. Chanon (1988) Photoinduced Electron Transfer,
Vol. A–D. Elsevier Science Publishers, B. V., Amsterdam.
3. Fagnoni, M., D. Dondi, D. Ravelli and A. Albini (2007) Photocataly-
sis for the formation of the C–C bond. Chem. Rev. 107, 2725–2756.
4. Ravelli, D., D. Dondi, M. Fagnoni and A. Albini (2009) Photocataly-
sis. A multi-faceted concept for green chemistry. Chem. Soc. Rev.
38, 1999–2011.
5. Gould, I. R., D. Ege, J. E. Moser and S. Farid (1990) Efficiencies of
photoinduced electron-transfer reactions: Role of the Marcus inverted
region in return electron transfer within geminate radical-ion pairs. J.
Am. Chem. Soc. 112, 4290–4301.
6. Roth, H. D. (2005) Return electron transfer in radical ion pairs of
triplet multiplicity. Pure Appl. Chem. 77, 1075–1085.
7. Gould, I. R., D. Noukakis, L. Gomez-Jahn, J. L. Goodman and S.
Farid (1993) Explanation of the driving-force dependence of return
electron transfer in contact radical-ion pairs. J. Am. Chem. Soc. 115,
4405–4406.
8. Gould, I. R., J. E. Moser, B. Armitage and S. Farid (1995) Factors
controlling the efficiencies of photoinduced electron-transfer reac-
tions. Res. Chem. Intermed. 21, 793–806.
9. Todd, W. P., J. P. Dinnocenzo, S. Farid, J. L. Goodman and I. R.
Gould (1991) Efficient photoinduced generation of radical cations in
solvents of medium and low polarity. J. Am. Chem. Soc. 113, 3601–
3602.
10. Clennan, E. L., C. Liao and E. Ayokosok (2008) Pyrylogens: Syn-
thesis, structural, electrochemical, and photophysical characterization
of a new class of electron transfer sensitizers. J. Am. Chem. Soc.
130, 7552–7553.
11. Frisch, M. J. T., G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, J. A. Jr Montgomery, T. Vreven,
K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tom-
asi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega,
G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota,
R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O.
Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian,
J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R.
E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J.
W. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salva-
dor, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D.
Daniels, M. C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck,
K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G.
ꢁ
23. Ecoffet, C., D. Markovitsl, P. Millie, C. Jallabert, H. Strzelecka and
M. Veber (1993) Spectroscopic properties of a triarylpyrylium cat-
ion. J. Chem. Soc., Faraday Trans. 89, 457–464.
24. Haucke, G., P. Czerney and F. Cebulla (1992) Absorption and fluo-
rescence of pyrylium salts. Ber. Bunsenges. Phys. Chem. 96,
880–886.
25. Chen, Y., P.-F. Wang and S.-K. Wu (1995) A study on the intramo-
lecular charge transfer and the photophysical behavior of 2,4,6-tri-
substituted pyrylium salt. J. Lumin. 65, 257–262.
ꢁ
26. Markovitse, D., H. Sigal, C. Ecoffet, P. Millie, F. Charra, C. Fiorini,
J.-M. Nunzi, H. Strzelecka, M. Veber and C. Jallabert (1994) Charge
transfer in triaryl pyrylium cations. Theoretical and experimental
study. Chem. Phys. 182, 69–80.
27. Dreuw, A. and M. Head-Gordon (2004) Failure of time-dependent
density functional theory for long-range charge-transfer states: The
zincbacteriochlorin-bacteriochlorin and bacteriochlorophyll-spheroid-
ene complexes. J. Am. Chem. Soc. 126, 4007–4016.
28. Montes-Navajas, P., L. Teruel, A. Corma and H. Garcia (2008)
Specific binding effects for cucurbit[8]uril in 2,4,6- triphenylpyryli-
um-cucurbit[8]uril host-guest complexes: Observation of room-
temperature phosphorescence and their application in electrolumines-
cence. Chem., A Eur. J., 14, 1762.
29. Wintgens, V., J. Pouliquen, P. Valat, J. Kossanyi, S. Canonica and
U. P. Wild (1986) Time-resolved interaction of polar solvents with
excited pyrylium cations. Chem. Phys. Lett. 123, 282–288.
30. Wintgens, V., T. Sadhana, J. Pouliquen and J. Kossanyi (1986)
Temperature and solvent viscosity effects on the absorption and flu-
orescence spectra of a series of pyrylium salts. J. Photochem. 32,
81–93.

Photochemistry and Photobiology, 2014, 90 357
31. Lakowicz, J. R. (1999) Principles of Fluorescence Spectroscopy.
Kluwer Academic/Plenum, New York.
32. Mataga, N., Y. Kaifu and M. Koizumi (1956) Solvent effects upon
fluorescence spectra and the dipole moments of excited molecules.
Bull. Chem. Soc. Jpn. 29, 465–470.
33. Rosokha, S. V., M. D. Newton, A. S. Jalilov and J. K. Kochi (2008)
The spectral elucidation versus the X-ray structure of the critical pre-
cursor complex in bimolecular electron transfers: Application of
experimental/theoretical solvent probes to ion-radical (redox) dyads.
J. Am. Chem. Soc. 130, 1944–1952.
39. Zaini, R., A. C. Orcutt and B. R. Arnold (1999) Determination of
equilibrium constants for weakly bound charge-transfer complexes.
Photochem. Photobiol. 69, 443–447.
40. Monk, P. M. S. (1998) The Viologens. Physicochemical Properties,
Synthesis and Applications of the Salts of 4, 4′-Bipyridine. John
Wiley & Sons, Chichester, England.
41. Kamata, M., M. Sato and E. Hasegawa (1992) Photosensitized oxy-
genation reactions of 1,3-dithianes through cooperative single elec-
tron transfer pathway and singlet oxygen pathway. Tetrahedron Lett.
33, 5085–5088.
34. Rosokha, S. V. and J. K. Kochi (2008) Fresh look at electron-trans-
fer mechanisms via the donor/acceptor bindings in the critical
encounter complex. Acc. Chem. Res. 41, 641–653.
35. Pace, A., E. L. Clennan, F. Jensen and J. Singleton (2004) A novel
zeolite-induced population of a planar viologen conformation. New
viologen charge transfer complexes and alkene/viologen/zeolite
arrays. J. Phys. Chem. B 108, 4673–4678.
36. Benesi, H. A. and J. H. Hildebrand (1949) A spectrophotometric
investigation of the interaction of iodine with aromatic hydrocarbons.
J. Am. Chem. Soc. 71, 2703–2707.
37. Hoenigman, S. M. and C. E. Evans (1996) Improved accuracy and
precision in the determination of association constants. Anal. Chem.
68, 3274–3276.
42. Akaba, R., S. Aihara, H. Sakuragi and K. Tokumaru (1987) Triphe-
nylpyrylium salt as sensitizer for electron transfer oxygenation not
involving superoxide anion. J. Chem. Soc., Chem. Commun., 1262–
1263.
43. Fornes, V., H. Garcia, M. A. Miranda, F. Mojarrad, M.-J. Sabater
and N. N. E. Suliman (1996) Zeolite-based heterogeneous photosen-
sitizers containing triphenylpyrylium and dibenzotropylium cations.
Modifications of the product selectivity in the cyclodimerization of
1,3-cyclohexadiene. Tetrahedron 52, 7755–7760.
44. Miranda, M. A., M. A. Izquierdo and F. Galindo (2002) Involvement
of triplet excited states and olefin radical cations in electron-transfer
cycloreversion of four-membered ring compounds photosensitized by
(thia)pyrylium salts. J. Org. Chem. 67, 4138–4142.
38. Kuntz, I. D., Jr, F. P. Gasparro, M. D. Johnston Jr and R. P. Taylor
(1968) Molecular interactions and the Benesi–Hildebrand Equation.
J. Am. Chem. Soc. 90, 4778–4781.
45. Horrocks, A. R., T. Medinger and F. Wilkinson (1967) Solvent
dependence of the quantum yield of triplet state production of 9-Phe-
nylanthracene. Photochem. Photobiol. 6, 21–28.