Mechanisms of singlet-oxygen and superoxide-ion generation by porphyrins and bacteriochlorins and their implications in photodynamic therapy

Article abstract of DOI:10.1002/chem.201000111

New halogenated and sulfonated bacteriochlorins and their analogous porphyrins are employed as photosensitizers of singlet oxygen and the superoxide ion. The mechanisms of energy and electron transfer are clarified and the rates are measured. The intermed

Full text of DOI:10.1002/chem.201000111

FULL PAPER
DOI: 10.1002/chem.201000111
Mechanisms of Singlet-Oxygen and Superoxide-Ion Generation by
Porphyrins and Bacteriochlorins and their Implications in Photodynamic
Therapy
[
a]
[a]
[b]
[a]
Elsa F. F. Silva, Carlos Serpa, Janusz M. D a˛ browski, Carlos J. P. Monteiro,
[
a]
[b]
[c]
[d]
Sebasti¼o J. Formosinho, Grazyna Stochel, Krystyna Urbanska, Sꢀrgio Simꢁes,
[
a]
[a]
Mariette M. Pereira, and Luis G. Arnaut*
Abstract: New halogenated and sulfo-
nated bacteriochlorins and their analo-
gous porphyrins are employed as pho-
tosensitizers of singlet oxygen and the
superoxide ion. The mechanisms of
energy and electron transfer are clari-
fied and the rates are measured. The
intermediacy of a charge-transfer (CT)
complex is proved for bacteriochlorins,
but excluded for porphyrins. The ener-
gies of the intermediates and the rates
of their interconversions are measured,
and are used to obtain the efficiencies
of all the processes. The mechanism of
formation of the hydroxyl radical in
the presence of bacteriochlorins is pro-
posed to involve a photocatalytic step.
The usefulness of these photosensitiz-
ers in the photodynamic therapy
(PDT) of cancer is assessed, and the
following recommendations are given
for the design of more effective PDT
protocols employing such photosensi-
tizers: 1) light doses should be given
over a more extended period of time
when the photosensitizers form CT
complexes with molecular oxygen, and
2) Fe may improve the efficiency of
such photosensitizers if co-located in
the same cell organelle assisting with
an in vivo Fenton reaction.
2
+
Keywords: charge transfer · photo-
dynamic therapy · photosensitizers ·
porphyrinoids · radical ions · singlet
oxygen
Introduction
ent), high singlet-oxygen-producing ability, and some ten-
dency to accumulate in solid tumours. These properties have
made them the most studied class of molecules for applica-
tion as photosensitizers in the photodynamic therapy (PDT)
of cancer. The design and synthesis of molecules exhibiting
increased infrared absorptions and higher singlet-oxygen
Porphyrins, phthalocyanines, and related macrocycles have
substantial electronic absorptions in the “phototherapeutic
window” (650–900 nm, in which tissues are most transpar-
[1–5]
quantum yields (F ) is a very active field of research,
D
[
a] E. F. F. Silva, Dr. C. Serpa, C. J. P. Monteiro, Prof. S. J. Formosinho,
Prof. M. M. Pereira, Prof. L. G. Arnaut
Chemistry Department, University of Coimbra
004-535 Coimbra (Portugal)
Fax : (+351)239-827-703
and shows promise for becoming a key treatment method in
[6–8]
oncology.
Further progress in this field will benefit from
3
a better understanding of the interaction between the photo-
sensitizer, light, and oxygen. This work addresses quantita-
tively that interaction when the photosensitizer is a porphy-
rin or a bacteriochlorin.
E-mail: lgarnaut@ci.uc.pt
[
b] Dr. J. M. D a˛ browski, Prof. G. Stochel
Faculty of Chemistry, Jagiellonian University
Ingardena 3, 30-060 Krakꢁw (Poland)
ꢀ
Photofrin (a mixture of purified fractions of hematopor-
ꢀ
[
c] Prof. K. Urbanska
phyrin derivatives) and Foscan (a synthetic chlorin) are the
Faculty of Biochemistry, Biophysics and Biotechnology
Jagiellonian University
Gronostajowa 7, 30-349 Krakꢁw (Poland)
most widely used PDT photosenstitizers. More recently, bac-
teriochlorins have been enrolled in clinical trials of photody-
[9,10]
namic therapy for various types of cancers,
because of
[
d] Prof. S. Simꢂes
their enhanced absorptions in the phototherapeutic window.
Some reports have suggested that the efficacy of bacterio-
chlorins in PDT is due not only to their strong infrared ab-
sorbance and their ability to transfer a large part of their
Faculty of Pharmacy, University of Coimbra
3
000 Coimbra (Portugal)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201000111.
Chem. Eur. J. 2010, 16, 9273 – 9286
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
9273

1
triplet-state energy to molecular oxygen, with the conse-
kdiff
k
kdiff
9
D
kD ¼
ꢁ
ð1Þ
1
1
quent production of singlet oxygen ( Dg O ), but also to
9ðkꢀdiffþ k Þ
2
D
their ability to transfer an electron to molecular oxygen and
ꢀ
[11–13]
generate the superoxide ion (O C ).
It has been noted
In the presence of charge transfer, the CT channel, and
for small energetic differences between the singlet and trip-
let states of the intermediates, the rate-determining step is
the parallel formation of singlet and triplet exciplexes; the
limiting rate increases to k
tum yield decreases to F
fication to this mechanism, based on detailed data on
energy-transfer rates to higher electronic states of molecular
oxygen, for which the intersystem crossing equilibrium is
fully established between singlet and triplet pathways at the
encounter complexes rather than at the charge-transfer com-
2
that tumour cells have diminished amounts of the superox-
ide dismutase, and that superoxide radical production may
[14]
be involved in PDT apoptosis. This low level of superox-
ide dismutase in tumour cells suggests that photosensitizers
capable of producing superoxide may be particularly effec-
tive in killing such cells. In view of the roles played by sin-
glet oxygen and the superoxide ion in the oxidative stress
and death of cells, it is particularly important to establish
the mechanisms of energy transfer and electron transfer
from the photosensitizers to molecular oxygen.
4
ꢁ /
kdiff, and the limiting quan-
9
=0.25. Schmidt proposed a modi-
D
D
[
19,20]
The ubiquity of molecular oxygen and its low-energy sin-
plexes.
ꢀ
1
glet excited state (E =22.5 kcalmol ), have made energy
This work presents a consistent mechanistic interpretation
of singlet oxygen and superoxide generation from the triplet
states of the porphyrins and bacteriochlorins illustrated
here, based on kinetic modelling of spectroscopic, photo-
D
transfer from a photosensitizer to molecular oxygen one of
the most prevalent and intensively investigated photochemi-
cal processes. The mechanism of this energy transfer is now
[15]
close to reaching a consensus, but disparate mechanisms
have been proposed for the concomitant generation of su-
peroxide. The current understanding of triplet-state quench-
ing by molecular oxygen is based on the mechanism pro-
ACHTUNGTRENNUNaG coustic, and chemical data. Energy and electron transfers
are exquisitely entangled in such systems, which are a fasci-
nating playground for energy and electron-transfer models.
We show that the quenching of porphyrin or bacteriochlorin
triplet states by oxygen follows different reaction mecha-
nisms. The bacteriochlorin triplet quenching involves
charge-transfer species, and the quantum yield of reactive
oxygen species (ROS: singlet oxygen, superoxide ion, hydro-
gen peroxide, hydroxyl radical) is increased when a given
light dose is delivered at lower laser intensities over a more
extended period of time. Additionally, the mechanism of hy-
droxyl radical formation suggests that the presence of fer-
rous ions may increase the yield of this ROS through the
Fenton reaction. On the other hand, the quenching of por-
phyrin triplets by molecular oxygen leads essentially to sin-
glet oxygen, and the quantum yield of its generation is less
sensitive to laser intensities or adjuvants. The conclusion of
this mechanistic study is that PDT with bacteriochlorins
[16–18]
posed by Wilkinson, Scheme 1.
This mechanism shows
that the diffusion-controlled encounter of the excited triplet
sensitizer and the triplet ground-state O leads to the rever-
2
sible formation of excited encounter complexes of singlet,
triplet, and quintet multiplicities. Statistically, there will be
one singlet, three triplets, and five quintet complexes. The
quintet pathway has no spin- or energy-allowed products,
and quenching occurs only through internal conversion pro-
cesses in the singlet and triplet pathways. In the absence of
charge-transfer interactions, the nCT channel, the triplet
pathway is unproductive, the rate of formation of singlet
oxygen is given by Equation (1), and the quantum yield of
singlet-oxygen generation may reach F =1.
D
Scheme 1. Wilkinson mechanism for singlet-oxygen generation from the triplet state of a sensitizer, simplified under the assumption that intersystem
crossing between encounter complexes is not efficient.
9274
www.chemeurj.org
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 9273 – 9286

Porphyrins and Bacteriochlorins
FULL PAPER
Figure 1. Absorption and fluorescence spectra of TDCPPSNHethyl and
TDCPBSNHethyl in ethanol.
The intense absorption of the bacteriochlorin at lmax
=
7
45.5 nm is particularly remarkable. The absorption coeffi-
ꢀ
1
ꢀ1
cient in the infrared, emax =97000m cm , is among the
largest for non-aggregating molecules in polar solvents. The
shoulder at 410 nm and the small peak at 658 nm are evi-
dence of a small chlorin contamination, which is present in
less than 5% in our samples. The chlorin red absorption
band is 6–7 nm further displaced into the red than that of
the corresponding porphyrin, which is typical of these mole-
cules and allows for their assignment. With appropriate se-
lection of excitation and emission wavelengths, this contami-
nation does not affect our results.
should be more effective at lower laser intensities, and in
the presence of ferrous ions.
Results
Porphyrins and bacteriochlorins are very useful for address-
ing the mechanism of triplet-state quenching by molecular
Absorption and steady-state fluorescence data are collect-
ed in Table 1, together with literature data on related mole-
cules. Both porphyrins and bacteriochlorins have small
Stokes shifts between absorption and fluorescence emission,
and the fluorescence quantum yields of 5,10,15,20-tetra-
kis(2,6-dichlorophenyl) porphyrins and bacteriochlorins are
oxygen, because their triplet-state energies (E =
T
ꢀ
ꢀ
1
[21]
3
3
3 kcalmol
0 kcalmol
for tetraphenylporphyrin, TPP,
and 25–
1
[4]
for analogous bacteriochlorins ) are lower
than the energy of the second excited state of molecular
1
+
ꢀ1
oxygen ( Sg O , E =37.5 kcalmol ). Hence, energy trans-
rather low, F <0.02. The reduced fluorescence is due to the
2
S
F
fer to higher electronic states of molecular oxygen is not rel-
evant to our systems, in contrast to the case for higher
[22]
Table 1. Absorption and fluorescence data of sulfonyl and sulfamoyl por-
phyrins and bacteriochlorins in ethanol, and of related compounds in tol-
uene.
energy photosensitizers. Additionally, the tetraphenylpor-
phyrins and bacteriochlorins employed in this work have
chlorine atoms in the ortho-positions of the phenyl rings,
which reduce their singlet-state lifetimes to less than 1 ns,
and make energy transfer from the singlet state of the sensi-
tizers uncompetitive with intersystem crossing to the triplet
state. This simplifies the mechanistic studies appreciably, be-
cause only one electronically excited state of the photosensi-
tizer and one electronically excited state of molecular
oxygen have to be taken into account.
l
max
e
max
ꢀ1
l
em
E
S
F
F
ꢀ1
ꢀ1
[
nm]
A[ m cm
H
U
G
R
N
N
]
[nm]
[kcalmol ]
[a]
TPP
TDCPP
TDCPPSO
649.8
660
9600
2000
652/719 44.0
661/719 43.3
654/720 43.8
655/720 43.8
658/724 43.6
748 38.3
0.10
0.005
–
0.0167
0.0172
0.012
0.0062
0.0081
0.0082
[
b]
3
H
TDCPPSNHethyl
TDCPPSNHheptyl
TDCPB
TDCPBSO
TDCPBSNHethyl
652
5000
[c]
747.0 126000
744.5
745.5
3
H
61000
97000
76000
748 38.3
749 38.3
751 38.2
Figure 1 shows the absorption and fluorescence spectra of
TDCPPSNHethyl and TDCPBSNHethyl in ethanol, which
are representative of our porphyrins and bacteriochlorins.
TDCPBSNHheptyl 746
[a] Reference [21]. [b] References [24] and [26]. [c] Reference [4].
Chem. Eur. J. 2010, 16, 9273 – 9286
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9275

L. G. Arnaut et al.
internal heavy-atom effect produced by the chlorine atoms
in the ortho-positions of the phenyl rings. As discussed else-
where, such chlorine atoms provide enough spin-orbit cou-
pling to enhance the intersystem crossing to the triplet
8 ns), and Ehn is the laser energy at the excitation wave-
length (E355 =80.5 kcalmol ).
ꢀ
1
E
F
T
¼ ð1ꢀꢀ
ÞEhnꢀF
E
ð2Þ
T
1
F
S
[23,24]
state.
The singlet-state energy was obtained from the in-
tersection of normalized absorption and emission bands.
The 1-octanol/water partition coefficients, KOW, of these
photosensitizers range from log KOW =ꢀ1.80 for
The values for the triplet quantum yields of 5,10,15,20-tet-
rakis(2,6-dichlorophenyl) porphyrins, chlorins, and bacterio-
chlorins are, to a good approximation, equal to 1ꢀF , be-
F
[25]
[4,21,30]
TDCPPSO H to log K >4 for TDCPPSNHheptyl,
cause of the internal heavy-atom effect.
The triplet en-
3
OW
which emphasizes the wide range of water solubilities of the
photosensitizers selected for this study.
ergies of the halogenated molecules presented in Table 2
were calculated using F =1ꢀF and Equation (2).
T
F
The presence of chlorine atoms in the ortho-positions of
the phenyl rings severely restricts the rotation of the single
bond at the meso-position on 5,10,15,20-tetrakisphenyl por-
phyrins or bacteriochlorins. When the phenyl rings are un-
symmetrical, geometric isomers result from the different po-
sition of the ortho and/or meta substituents relative to the
porphyrin plane. Such atropisomers have been isolated for a
series of 5,10,15,20-tetrakis(2,6-dichloro-3-sulfamoylphenyl)
porphyrins, and shown to have substantially different spec-
troscopic properties. Atropisomers are also present in the
unsymmetrical molecules of Table 1, and the values reported
here correspond to a weighted average of the properties of
the individual atropisomers in the fractions collected after
chromatography. However, this will not affect the rates and
mechanisms discussed further below in this work.
Table 2. Triplet-state properties of sulfonyl and sulfamoyl porphyrins and
bacteriochlorins in ethanol, and of related compounds in toluene.
E
T
t
T
(N
2
)
t
T
(air)
k
q
ꢀ1
ꢀ1 ꢀ1
[
kcalmol
]
[ms]
[ns]
A[ m
C
H
T
U
N
G
T
R
E
N
N
U
N
G
s
]
[a]
[b]
9
TPP
TDCPP
TDCPPSO
33.0
43
17
40
38
41
32
33
38
28
349
641
855
710
758
257
226
265
295
1.4ꢄ10
8.6ꢄ10
5.4ꢄ10
6.6ꢄ10
6.2ꢄ10
2.1ꢄ10
2.1ꢄ10
1.8ꢄ10
1.6ꢄ10
[c]
8
8
8
8
9
9
9
9
3
H
TDCPPSNHethyl
TDCPPSNHheptyl
TDCPB
TDCPBSO
TDCPBSNHethyl
TDCPBSNHheptyl
33.0
36.1
30.4
[
27]
[
c]
3
H
25.7
27.4
[a] Reference [26]. [b] Reference [31]. [c] Reference [4].
Under our experimental conditions, no phosphorescence
emission could be detected unambiguously from our mole-
cules, as is generally the case for free-base porphyrins and
Triplet-state lifetimes in aerated and N -saturated solu-
2
tions were measured by laser flash photolysis, and Figure 2
shows representative decays. We were unable to detect the
presence of long-lived intermediates other than triplets. Ad-
ditionally, isosbestic points were observed for the ground-
state recovery and triplet decay, indicating that at least 90%
[28]
related macrocycles.
free-base bacteriochlorins must be regarded with caution,
as we also have been misled by the phosphorescence of im-
Reports of phosphorescence from
[29]
[
21,30]
purities in free-base chlorin samples.
proach to measuring the triplet-state energies of these mole-
An alternative ap-
[21]
cules is to use photoacoustic calorimetry (PAC).
This
technique measures the heat deposited in the medium as the
excited molecule decays by radiationless processes. The
energy balance of Scheme 2 in de-aerated solutions gives,
for the prompt release of heat, Equation (2), in which f is
1
the fraction of energy release as prompt heat (less than 8 ns
under our experimental conditions, associated lifetime t <
1
Scheme 2. Fractions of heat released (f
pulsed laser irradiation of a sensitizer, S. In the presence of oxygen, the
decay of the triplet state is also observed.
1
, f
2
) measured by PAC following
Figure 2. Transient absorption spectrum of TDCPBSNHheptyl in ethanol,
measured by laser flash photolysis. The inset shows the decay at 400 nm,
which has the same lifetime as the ground-state bleaching recovery.
9276
www.chemeurj.org
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 9273 – 9286

Porphyrins and Bacteriochlorins
FULL PAPER
of the triplet state returns to the ground state without kinet-
ically relevant intermediates. The presence of oxygen sub-
stantially reduces the lifetimes of the triplet states, as shown
in Table 2. The quenching rate constant was estimated from
Equation (3), using the oxygen concentration in the appro-
emission, trisetime, must correspond to the decay rate of the
triplet-state absorption in aerated solutions, tT ACHUTNGRNENGU( air). The fast
rise time of the weak emission of singlet oxygen could only
be observed using picosecond excitation of the bacterio-
chlorins (Figure 3). Nanosecond laser pulses produced too
ꢀ
3
ꢀ3
priate solvent ([O ]=2.1ꢄ10 m and 1.8ꢄ10 m in ethanol
2
[32]
and toluene at 208C, respectively).
The data in Table 2
were collected at room temperature (approximately 208C).
1
=tðairÞ ¼ 1=tðN Þ þ k ½O ꢂ
ð3Þ
2
q
2
Singlet oxygen in ethanol or toluene has a relatively long
[33]
[4]
lifetime: 15.2
and 33 ms, respectively. This favours the
detection of the singlet-oxygen emission at 1270 nm. The
singlet-oxygen lifetimes measured in this work are in very
good agreement with the literature, as seen in Table 3.
Table 3. Singlet-oxygen rise time, decay, and quantum yield generated by
sulfonyl and sulfamoyl porphyrins and bacteriochlorins in ethanol, or by
related compounds in toluene.
[
a]
t
risetime
t
decay
F
D
k
ACHTUNGENTRNUNG[ m ]
1
/k
n
1
CT CT
F E
ꢀ
[ns]
[ms]
[
b]
phenalenone
TPP
TDCPP
0.95
0.71
0.98
1.00
0.85
0.85
0.60
0.85
0.66
0.63
0.02ꢃ0.01
[
c]
[c]
d]
31.2
[
TDCPPSO
3
H
810
743
844
14.6
14.7
14.5
33
15.4
13.7
14.3
0.10ꢃ0.05
0.10ꢃ0.05
0.10ꢃ0.05
TDCPPSNHethyl
TDCPPSNHheptyl
ꢄ0
ꢄ0
Figure 3. Rise time of singlet-oxygen emission following picosecond laser
excitation of TDCPBSNHheptyl in ethanol.
[
e]
TDCPB
TDCPBSO
10.4
3
H
223
1.0ꢃ0.2
1.0ꢃ0.2
1.0ꢃ0.2
TDCPBSNHethyl
TDCPBSNHheptyl
8.1
10.1
330
much interference with the singlet-oxygen rise time for such
photosensitizers. A detailed kinetic analysis of the energy-
transfer mechanism provides further insight into the ob-
served rise times and decays, but Tables 2 and 3 show that
ꢀ
1
[
a] Maximum error of ꢃ2 kcalmol . [b] In reference [34]. [c] In refer-
ence [4], 0.78ꢃ0.04 was reported from a third degree polynomial fit to
the singlet-oxygen decay in benzene in reference [35], and 0.67ꢃ0.14 was
reported from PAC studies in toluene in reference [21]. [d] From PAC
studies in toluene in reference [26]. [e] From reference [4] in toluene.
trisetime and t AHCUTNGRNENUG( air) have similar values for each system.
T
Figure 4 presents EPR spectra in the presence of
TDCPBSO H measured in a phosphate buffer water solu-
3
The intensity of the singlet-oxygen phosphorescence
tion (PBS, pH 7.4) with air-saturated solutions in the dark
(Figure 4A), and an air-saturated solution irradiated for
12 min with the diode laser (Figure 4B). The computer sim-
ulation of this spectrum is shown for comparison in Fig-
ure 4C. The hydroxyl adducts spectrum was simulated on
the basis of two conformers (see Figure 4D and E) undergo-
ing chemical exchange. The line shape and the hyperfine
(hf) splitting of the signal are typical of BMPOꢀOH radicals
0
decay extrapolated to time zero (I ) can be used to deter-
D
mine the quantum yield of singlet-oxygen generation (F )
D
[
36]
by our photosensitizers using phenalenone as a reference.
0
The standard procedure employs values of I measured at
D
different laser intensities for solutions of samples and the
reference with the same absorbance at the excitation wave-
length. Special care must be exercised with systems that ex-
hibit different degrees of non-linearity with the laser intensi-
(BMPO=5-tert-butoxycarbonyl 5-methyl-1-pyrroline N-
[
35,37]
ty.
The correct interpretation of the laser energy de-
oxide). The EPR parameters of BMPO adducts are present-
[38]
pendence of the singlet-oxygen emission intensity plots
relies on the solution of the kinetic equations of Wilkinsonꢅs
mechanism. A solution that retains the essential features of
ed in Table 4. Following earlier work, we assign the spec-
trum in Figure 4B to a mixture of diastereomers of the
BMPOꢀOH adduct.
this mechanism is discussed further below, and it is shown
0
D
that it supports the use of a linear fit to I , using the lowest
laser energy measurements to obtain F . The data obtained
D
with this procedure are presented in Table 3, and the linear
fits can be found in the Supporting Information.
In direct energy transfer from the photosensitizer triplet
state to molecular oxygen, the rise time of singlet-oxygen
Chem. Eur. J. 2010, 16, 9273 – 9286
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9277

L. G. Arnaut et al.
Figure 5. EPR spectrum of the DMPOꢀOOH adduct observed during
3
the illumination of TDCPBSO H (50 mm) in DMSO in the presence of
DMPO (100 mm): A) before illumination, B) during illumination of an
air-saturated sample, and C) simulation.
the contributions described by Equation (4), in which the
ꢀ
1
singlet-oxygen energy is E =22.5 kcalmol , E F is known
D
T
T
Figure 4. EPR spectrum of the BMPOꢀOH adduct observed during the
illumination of TDCPBSO H (50 mm) in PBS and in the presence of
BMPO (40 mm): A) before illumination, B) during illumination of an air-
saturated sample, C) simulation, D,E) simulation of two conformers of
the BMPOꢀOH adduct.
from the prompt release of energy, [Eq. (2)], and is consis-
3
tent with F ꢄ1 for all 5,10,15,20-tetrakis(2,6-dichlorophen-
T
yl) porphyrins and bacteriochlorins.
ꢀ2
Ehv ¼ F ðE ꢀ E Þ þ F ðE ꢀ E Þ þ ðF ꢀ F ꢀ FCTÞET
D
T
D
CT
T
CT
T
D
ð4Þ
Table 4. EPR parameters of BMPOꢀOH and DMPOꢀOOH adducts.
Spin adduct
Diastereo-
mers [%]
Hyperfine coupling constant [G]
Figure 6 shows representative PAC waves measured in
aerated ethanol solutions of the photosensitizers used in this
study. The energy released in the formation of the charge-
a
N
a
Hß
a
HU
7
2
–
3
7
14.1
14.2
13.3
12.8
15.8
10.6
0.68
0.77
1.34
1
4
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
[ N]BMPOꢀOH
1
4
separated state, FCTE , was calculated from Equation (4)
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
[ N]DMPOꢀOOH
CT
with the value of F in Table 3.
D
Similar experiments were conducted with DMPO (5,5-di-
methylpyrroline-N-oxide) in DMSO. We detected the COOH
Discussion
adduct in the presence of air and light; the EPR spectrum
of this adduct is presented in Figure 5B. However, in the ab-
sence of light, or in a nitrogen-saturated solution, or in the
presence of superoxide dismutase, this adduct was not ob-
served. Figure 5C shows the simulation of the DMPOꢀ
COOH adduct. The line shape and the hyperfine (hf) split-
ting of the signal are typical of DMPOꢀOOH radicals gen-
The discussion of the data is divided into three parts: a brief
overview based on a simplified mechanism that retains the
essential features of Wilkinsonꢅs mechanism, a detailed anal-
ysis of the mechanism of singlet-oxygen generation, and a
mechanistic interpretation of the evidence for the formation
ꢀ
of other reactive oxygen species (O C , H O , and OHC).
2
2
2
ꢀ
erated by the reaction of DMPO and O C (a =13.3 G,
The most immediate facts that emerge from this work are
as follows. In de-oxygenated ethanol solution, halogenated
tetraphenylporphyrins and bacteriochlorins have long-lived
triplet states, formed with near unit quantum yields, but the
quenching of such triplet states by molecular oxygen is dis-
tinctly different. Porphyrin triplets are quenched with a rate
2
N
aHß =10.6 G, aHU =1.34 G).
Singlet-oxygen and superoxide-ion generation in ethanol
was further investigated with PAC. In aerated solutions, the
decay of the triplet state occurs in the time window of the
experiment, and is measured as a second fraction of heat
decay, f , with a lifetime t that corresponds to the triplet
8
ꢀ1 ꢀ1
constant k =6ꢄ10 m s , which is virtually identical to
2
2
q
1
9
ꢀ1 ꢀ1
[32]
lifetime (Scheme 2). The heat decay in this time window has
/ kdiff, in which kdiff =5.4ꢄ10 m
s
in ethanol at 208C.
9
9278
www.chemeurj.org
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 9273 – 9286

Porphyrins and Bacteriochlorins
FULL PAPER
anism and in its kinetic analysis are presented in detail in
the Supporting Information. Two features of this mechanism
must be emphasized: 1) the simplified mechanism explicitly
considers the formation of the superoxide ion from the
charge-separated complex, which may be energetically fa-
vourable in polar solvents; 2) the simplified mechanism does
1
1
not consider the decay of O , and consequently [ O ] is
2
2
1
0
proportional to I , which is a measure of the total amount
D
of singlet oxygen produced by the photosensitizer.
For low-to-moderate light intensities, using the steady-
3
3
ꢀ
d+
dꢀ
state approximation for ( S*···O S ) and (S ···O2 ) in the
2
g
simplified mechanism, and subject to the condition that
[
S*]![O ], the concentration of singlet oxygen when t!1
2
is given by Equation (5), in which k =k +k +
n
ꢀ1
2
k3kCT
/
A
H
U
G
R
N
U
G
A
H
U
G
R
N
U
G
A
H
U
G
R
N
U
G
+
sep
3
ꢀCT
1
3
ꢀCT
3
3
3
ꢀ
g
d+
dꢀ
k
ꢀ
, and it is assumed that [ S*···O S ]+[S ···O2 ]!
et
[
O ].
2
ꢀ
ꢁ
kn
k2
k1CT
1
½
O2ꢂ1
¼
ln 1 þ
½S*ꢂ0
ð5Þ
Figure 6. Photoacoustic calorimetry waves of reference (short dashed
line) and TDCPBSNHethyl (long dashed line) in aerated ethanol solu-
tions following 355 nm laser excitation. The calculated wave was obtained
by reconvolution of the decay parameters with the reference wave, and
its difference from the TDCPBSNHethyl wave, called the residue, was
scaled by a factor of 10 to make it more visible in the plot.
k1CT k þ k k =ðk þ kꢀCTÞ
kn
2
3
CT
3
d+
dꢀ
+
ꢀ
Below, we will represent [S ···O2 ] by [SC ···O C ], be-
2
cause for the cases where charge transfer is most relevant, a
full charge transfer takes place.
For the absorbances employed in this work, [S*] is nearly
0
identical to the number of photons absorbed, Ia =
ꢀ
A
On the other hand, bacteriochlorin triplets are quenched
I
A
H
U
G
R
N
U
G
0
9
ꢀ1 ꢀ1
with a rate constant k =2ꢄ10 m s , which approaches
dependences of [ O ] versus [S*] and I versus Ehn to be
2 0
D
q
4
/9
kdiff. The singlet-oxygen phosphorescence quantum yields
identical. Figure 7 presents fits of expressions with the form
0
D
of such porphyrins and bacteriochlorins are also different:
whereas porphyrin triplets generate singlet oxygen with an
efficiency near unity, the corresponding value for bacterio-
chlorins is 40–15% lower. It must be emphasized that the
triplet state of TPP also generates singlet oxygen with unit
efficiency, because its triplet quantum yield in toluene is
of Equation (5) to the values of I measured for a wide
range of laser energies E . The fitting is remarkably good in
hn
view of the approximations involved and of the wide range
of laser energies employed. It is important to emphasise
that the singlet-oxygen quantum yields presented in Table 3
were not obtained from this fit, but from the linear fit to the
lower laser energies. Such linear fits are shown in the Sup-
porting Information.
[
21]
F =0.73ꢃ0.10.
T
According to Wilkinsonꢅs mechanism,
porphyrins follow the nCT channel, but bacteriochlorins
predominantly follow the CT channel. Additionally, our
data show that both energy and electron transfers take place
from the bacteriochlorin triplet states to molecular oxygen,
leading to singlet oxygen and the superoxide ion, respective-
ly.
The limit of Equation (5) for low laser energies (low
[S*] ), is the expression in Equation (6), which supports the
0
linear fit employed to obtain the value of FD presented in
Table 3.
The molecular understanding of the phenomena summar-
ized above can be built on the simplified mechanism shown
in Scheme 3, which retains the essential features of Wilkin-
sonꢅs mechanism, but is more amenable to an analytical so-
lution. The approximations involved in the simplified mech-
k
2
1
½
O2ꢂ1
¼
½S*ꢂ0
ð6Þ
k þ k k =ðk þ kꢀCTÞ
2
3
CT
3
In principle, it is possible to deduce FD from the fits in
Figure 7. In practice, the term k /k can only be precisely de-
1
n
termined when the curvature of the phenalenone-sensitized
1
O emission is appreciable, and this only occurs for very
2
high laser energies, for which the approximations involved
in deriving Equation (5) are no longer valid. Figure 7 indi-
cates the values of k /k that give acceptable fits to Equa-
1
n
tion (5). The procedure that we recommend to obtain F
D
0
consists in using the linear portion of the I versus Ehn de-
D
pendence, and, on the basis of Equation (6), calculating FD
from the ratio of the slopes of that dependence for a sample
and a reference with known FD.
Scheme 3. Simplified mechanism of photosensitizer quenching, together
with singlet-oxygen and superoxide-ion generation.
Chem. Eur. J. 2010, 16, 9273 – 9286
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9279

L. G. Arnaut et al.
Figure 7. Extended laser energy dependence of singlet-oxygen emission in two independent experiments. ^ =phenalenone. Left:
=TDCPBSO
TDCPPSNHethyl, * =TDCPPSNHheptyl, ~ =TDCPBSNHethyl. Right: H, ~ =TDCPBSNHethyl, * =TDCPBSNHheptyl. The curves
were fitted to an equation of the form of Equation (5) with k1CT/k =0.02, 0.1, or 1.0 for phenalenone, porphyrins, and bacteriochlorins, respectively. The
pre-logarithmic factors in the left graph were 0.031, 0.0083ꢃ0.0006, and 0.00124 for phenalenone, porphyrins, and bacteriochlorins, respectively
&
=TDCPPSO
3
H, ~ =
&
3
n
1
3
1
The resolution of the simplified mechanism kinetics also
relates the observed triplet lifetimes with the various rate
constants of that mechanism. The concentrations of the sen-
mote intersystem crossing to { S*···O D } and increase k .
2 g q
The fact that k of porphyrins does not increase with halo-
q
genation suggests that the triplet encounter complex has a
ꢀ
5
9
ꢀ1
sitizers under our experimental conditions are in the 10
m
short lifetime, (k +k )>10 s . On the other hand, fast in-
ꢀ1 2
range, and [S*]@[O ]. The quenching of the photosensitizer
tersystem crossing must occur within the charge-separated
2
1
,3
+
ꢀ
in the simplified mechanism can be effectively approximated
as a pseudo-first-order reaction, and the decay of the photo-
sensitizer triplet state and the generation of singlet oxygen
states {SC ···O C }, as suggested by Wilkinsonꢅs mechanism.
2
The kinetic analysis of the simplified mechanism does not
need to identify the multiplicity of these states, provided
’
1
1
are given by Equation (7), in which k =k [O ].
that we employ k = / k for the nCT channel and k =
/9
1
2
1
9
diff
1
4
kdiff for the CT channel.
ꢂ
ꢀ
ꢁ ꢃ
k
kn
ꢀ
1
The CT channel is expected to become relevant when
charge-transfer complexes become as stable as the corre-
sponding encounter complexes. Ethanol is a polar solvent
½
S*ꢂ ¼ ½S*ꢂ exp ꢀk 1 ꢀ
t
0
1
ꢄ
ꢂ
ꢀ
ꢁ ꢃꢅ
k2
k
ꢀ1
1
0
½
O ꢂ ¼
2
½S*ꢂ0 1 ꢀ exp ꢀk 1 ꢀ
t
1
(
e=24.55), and the CT complexes are best regarded as radi-
k þ k k =ðk þ kꢀCTÞ
k
n
2
3
CT
3
[39]
cal–ion pairs with energies given by Equation (8).
ð7Þ
2
0
e
As expected, the triplet-state decay and the singlet
oxygen rise time have the same lifetimes (Figures 2 and 3,
and Tables 2 and 3).
e
rip
ox
red
e
DG ¼ ðED ꢀ EA Þ ꢀ
ð8Þ
erDA
The last term in Equation (8) is negligible in ethanol. The
red
A
Singlet-oxygen generation: In the following, we combine
our kinetic and thermochemical data with the data available
in the literature to determine the rate constants of all the el-
ementary steps identified in the simplified mechanism of
Scheme 3, as well as to estimate the energies and quantum
yields of the species invoked by that mechanism.
half-wave reduction potential for oxygen (E =ꢀ0.78 V vs.
[40]
[18]
SCE in DMSO,
also found in other sources ) and the
half-wave oxidation potential for tetraphenylbacteriochlorin,
TPB (E =0.40 V vs. SCE in CH Cl ), give a radical–ion
pair energy of 27 kcalmol , similar to the E of bacterio-
chlorins. A different situation is obtained with the half-wave
oxidation potential for TPP (E =1.13 V vs. SCE in
which gives DG =44 kcalmol , much higher
than the E of porphyrins. Although different solvents were
ox
D
[41]
2
2
ꢀ
1
T
1
ox
The observation of k = / k in the quenching of triplet
q
9
diff
D
[42]
e
rip
ꢀ1
porphyrins has another implication for the triplet pathway.
DMF),
3
3
3
2
ꢀ
If the triplet encounter complex { S*···O S } had a nano-
g
T
second lifetime, like the singlet state of TPP and related
non-halogenated porphyrins, the increased spin-orbit cou-
pling introduced by the presence of Cl atoms would pro-
employed in these measurements and a quantitative compar-
ꢀ
1
ison is not possible, the 10 kcalmol endothermicity expect-
+
ꢀ
3
ed for the formation of {TPPC ···O C } from { TPP···O } is in
2
2
9280
www.chemeurj.org
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 9273 – 9286

Porphyrins and Bacteriochlorins
FULL PAPER
sharp contrast with the possible exothermicity of the forma-
+
ꢀ
3
tion of {TPBC ···O C } from { TPB···O }. These estimates lend
2
2
further support to the nCT channel in the quenching of trip-
let porphyrins by molecular oxygen, and to the early inter-
ruption of the triplet pathway.
Energy transfer in the encounter complex formed in the
singlet pathway has been interpreted as an internal conver-
1
3
3
ꢀ
1
1
sion of { S*···O S } into {S···O D }, following an early
2
g
2
g
[43]
suggestion by Kearns and co-workers.
This requires the
encounter complex decay rate constant to follow the
energy-gap law, and become slower as the triplet energy of
the sensitizer increases, DE=E ꢀE . Considering that the
D
T
nCT channel has F ꢄ1 and k can be neglected, then k ꢄ
D
CT
n
k
+k , and the observed decay constant, k ’
A
H
U
G
R
N
U
G
ꢀ
1
2
1
ꢀ1
n
Equation (7), can be rearranged to k ’k / ACHTUNGTRNEGNU( k +k ). Using the
1
2 ꢀ1 2
usual assumption that in the formation of encounter com-
[
20]
plexes with molecular oxygen, k diff/kdiff ꢄ1m,
we expect
ꢀ
that k /k ꢄ1m. The observed decay constant takes the
ꢀ
1
1
form k k / ACHTUNGTRENNUNG( k +k )[O ], and the energy-gap law should be ob-
1
2 1 2 2
Figure 9. Rate constants (left axis, *) and singlet-oxygen quantum yields
served when k <0.1k . Figure 8 compares the energy-trans-
2
ꢀ1
(
right axis, ) for charge-transfer assisted quenching of substituted naph-
&
fer rates from the triplet states of our porphyrins with those
thalenes (open symbols) and bacteriochlorins (filled symbols).
[44]
from other sensitizers. The rate constants for the porphy-
rin triplets have reached the maximum possible value of the
1
nCT channel, k = / kdiff, meaning that the Franck–Condon
are such that a noticeable acceleration is observed, but the
efficiency of singlet-oxygen generation is not appreciably
1
9
factors have been optimized.
Charge-transfer-induced quenching of triplet states is
clearly observed in the quenching of substituted naphtha-
lene triplet states by molecular oxygen in acetonitrile
eroded. When kCT is large, and consequently k
is small,
ꢀCT
k ꢄk and the triplet decay constant in Equation (7) ap-
n
CT
proaches k ’. Figure 9 shows that the limiting value is k =
1
1
[18]
4
(
Figure 9). The quenching of triplet bacteriochlorins also
/9kdiff, which can be assigned to efficient intersystem cross-
follows the inverse correlation between the rates and effi-
ciencies of singlet-oxygen generation observed when
quenching proceeds through the CT channel. However, the
energetics of the CT states in the bacteriochlorin systems
ing between singlet and triplet charge-transfer complexes
with sufficiently long lifetimes to reach equilibrium.
The semi-quantitative arguments presented above have
quantitative expression in Equations (5)–(7). From the ratio
k1CT/k in Equation (5), and for the nCT channel (negligible
n
kCT), we obtain Equation (9).
ꢀ
ꢁ
k1CT
kn
k1
¼ k
ð9Þ
þ k2
nCTchannel
1
ꢀ1
9
ꢀ1 ꢀ1
1
9
ꢀ1
With k = / 5.4ꢄ10 m s , k = / 5.4ꢄ10 s , and the
1
9
ꢀ1
9
ratio fitted in Figure 7, k /k ꢄ0.1m for porphyrins, we
1
CT
n
9
ꢀ1
obtain k =5.4ꢄ10 s . This energy-transfer rate constant
2
within the excited-state complex is also valid for the CT
channel, because they differ only in the presence of a
charge-transfer state of lower energy. In a finer analysis, it is
arguable that k refers to an internal conversion subject to
2
the energy-gap law, and may be slightly higher for bacterio-
chlorins. However, porphyrin and bacteriochlorin triplet en-
ꢀ
1
ergies differ by only 5 kcalmol , and it is reasonable to
9
ꢀ1
employ k =5.4ꢄ10 s also for our bacteriochlorins.
2
Another useful relation can be obtained by rearranging
the time constant in Equation (7) to the form in Equa-
tion (10).
Figure 8. Energy gap law for triplets of TPP, ketones, quinones, and other
k
kn
1
t k
T
ꢀ
1
aromatic molecules measured in CCl
4
&
( ), and for the porphyrins mea-
¼
1 ꢀ
ð10Þ
0
sured in this work in ethanol (*).
1
Chem. Eur. J. 2010, 16, 9273 – 9286
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9281

L. G. Arnaut et al.
ꢀ
4
From the ratio fitted in Figure 7 for bacteriochlorins
The values fitted in Figure 7 give 8.5ꢄ10 for the por-
ꢀ
3
which use the CT channel, we have k1CT/k ꢄ1m, and k can
phyrins and 1.2ꢄ10 for the bacteriochlorins. Only the
ratio between these values is meaningful because it cancels
instrumental constants. Although the ratio given by the ex-
perimental data has a large uncertainty, 1.4ꢃ0.76, it is reas-
suring that the same ratio calculated with the data in
Table 4, 0.71, is within the experimental limits.
n
n
be replaced by k1CT in Equation (10). After some algebraic
manipulation and using k /k ꢄ1m, we obtain Equa-
ꢀ
1
1
tions (11) and (12), in which t is the triplet lifetime of the
T
bacteriochlorin in the presence of oxygen.
0
k3 ¼ kꢀ1 þ ð1 ꢀ k t Þk
ð11Þ
ð12Þ
1
T
2
Superoxide-ion generation: The spectrum of the superoxide
spin adduct of DMPO (DMPO OOH) shown in Figure 5 is
prima facie evidence for the formation of the superoxide ion
in DMSO. However, the EPR spectra obtained after illumi-
nation of TDCPBSO H aqueous solutions in the presence of
BMPO show the formation of a spin adduct between this
kCT
1
ꢀ
¼
0
k3 þ k
ꢀCT
k t ꢀ 1
T
1
4
9
ꢀ1 ꢀ1
ꢀ3
With k = / 5.4ꢄ10 m s , [O ]=2.1ꢄ10 m, t =260 ns,
1
9
2
T
3
9
ꢀ1
8
ꢀ1
and k =5.4ꢄ10 s , we calculate k =7.2ꢄ10 s . The value
2
3
of k and Equation (12) can be used to estimate the rate
spin trap and the hydroxyl radical (BMPOꢀOH). In this sec-
3
constants kCT and k , because these are related by the free
tion we discuss the generation and decay of the superoxide
ion, including the mechanisms that can account for the ob-
servation of the hydroxyl radical in aqueous solutions.
We were unable to detect the BMPOꢀOOH spin adduct
ꢀCT
+
ꢀ
3
3
2
ꢀ
energy of formation of {SC ···O C } from { S*···O S }, as
2
g
shown in Equation (13), and this is given by the PAC data.
ꢀ
ꢁ
0
CT
ꢀ
kCT
DG
resulting from direct O C trapping by BMPO under diode
2
¼
CT
exp ꢀ
ð13Þ
k
ꢀ
RT
laser irradiation in PBS. Instead, we detected the spin
adduct with the hydroxyl radical. There are two possible ex-
ꢀ
Making FCT =0.33 and 0.38 for TDCPBSNHethyl and
planations of this observation: 1) all O
2
C
may act as a pre-
ꢀ
TDCPBSNHheptyl, respectively, we obtain E =24.7 and
cursor of OHC; 2) the BMPO
OOH adduct may be formed,
CT
ꢀ
1
0
ꢀ1
2
6.4 kcalmol , which correspond to DG =ꢀ1 kcalmol
but cannot be observed due to its lifetime being too short.
The presence of catalase inhibits the formation of the
BMPOꢀOH radical adduct. Catalase splits hydrogen perox-
CT
in both cases. This is consistent with Equation (8) and leads
0
CT
to F +F =1.00ꢃ0.01. This value of DG , together with
D
CT
8
ꢀ1
9
ꢀ1
Equation (12) and k =7.2ꢄ10 s , give k =5.7ꢄ10 s
ide into water and molecular oxygen, and the observed in-
hibition of OHC formation proves that hydroxyl radicals are
3
CT
9
ꢀ1
and k =1.1ꢄ10 s . Table 5 lists the rates calculated for
ꢀCT
singlet-oxygen generation by porphyrins and bacteriochlor-
ins.
not formed directly from the photosensitizer and molecular
oxygen, but rather as a secondary thermal product. The ad-
dition of superoxide dismutase, a known scavenger of the
superoxide ion, also inhibits the formation of the BMPOꢀ
ꢀ
3
Table 5. Typical rate constants in ethanol, [O
parameters relevant for kinetic modelling.
2
]=2.1ꢄ10 m, and other
OH radical adduct, corroborating the hypothesis that the
hydroxyl radical is formed consecutively to other species
such as the superoxide ion and hydrogen peroxide. The pho-
togeneration of superoxide and its subsequent reactions to
produce hydrogen peroxide can be described by the mecha-
nism in Equations (15)–(19).
Rate constants
Porphyrins
Bacteriochlorins
ꢀ
1
ꢀ1
9
9
k
k
k
diff [m
s
]
5.4ꢄ10
5.4ꢄ10
ꢀ
1
ꢀ1
1
4
1
1
[m
’=k
s
]
/
/
9
k
[O
diff
/
/
9
k
diff
ꢀ
1
1
4
1
ꢀ
[O
2
] [s
]
9
2
] kdiff
9
2
[O ] kdiff
1
K
ꢀ1 [s
]
)
k
1
/1m
1
k /1m
ꢀ
1
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(k1CT/k
n
exp [m
]
0.10ꢃ0.05
1.0ꢃ0.2
ꢀ
1
9
9
k
2
[s
]
5.4ꢄ10
5.4ꢄ10
3
3
þ
ꢀ
2
TDCPBSO H þ O ! TDCPBSO H C þ O C
ð15Þ
ð16Þ
ð17Þ
ð18Þ
ð19Þ
3
2
3
t
T
[ns]
800
260
7.2ꢄ10
ꢀ
1
8
k
3
[s
]
þ
ꢀ
TDCPBSO H C þ O C ! TDCPBSO C þ HO C
3
2
3
2
F
D
0.93ꢃ0.08
0.7ꢃ0.1
0.3ꢃ0.1
ꢀ1.0ꢃ1
F
CT
0
ꢀ
þ
0
ꢀ1
2 O C þ 2 H ! O þ H O
DG CT [kcalmol
]
10
2
2
2
2
ꢀ
1
9
k
k
CT [s
ꢀCT [s
]
negligible
5.7ꢄ10
ꢀ
1
9
]
1.1ꢄ10
HO C þ HO C ! O þ H O
2
2
2
2
2
ꢀ
þ
HO C þ O C þ H ! O þ H O
2
2
2
2
2
An independent verification of the values in Table 5 is
given by the product of the pre-logarithmic factor in Equa-
The rate of superoxide disproportionation [Eq. (17)] at
pH 7.4 is reported as 2.4ꢄ10 m s . The disproportiona-
5
ꢀ1 ꢀ1 [45]
tion (4) by the factor multiplying [S*] , which yields Equa-
tion (14).
tion rate of the perhydroxyl radical [Eq. (18)] is 8.1ꢄ
10 m s . The oxidation of superoxide by the perhydrox-
0
5
ꢀ1 ꢀ1 [46]
yl radical reflects the fact that HO C is the more potent oxi-
2
dant. The subsequent formation of the hydroxyl radical may
proceed directly by Equations (20) and (21), but their rate
k2
ð14Þ
ꢀ
1
ꢀ1
[46]
k þ k k =ðk þ kꢀCTÞ
2
3
CT
3
constants, 16 and 3.7m
s
in water,
are much slower
9282
www.chemeurj.org
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 9273 – 9286

Porphyrins and Bacteriochlorins
FULL PAPER
ꢀ
2
than the relatively rapid dismutation of the HO C/O C radi-
cals.
2
[47]
ꢀ
ꢀ
O C þ H O ! O þ OHC þ OH
ð20Þ
ð21Þ
2
2
2
2
HO C þ H O ! O þ OHC þ H O
2
2
2
2
2
The Haber–Weiss reaction [Eq. (20)] was recently ob-
[48]
served in the gas phase,
but remains controversial in
water. Alternatively, we propose that bacteriochlorins may
also lead to the hydroxyl radical through photocatalysis,
[
Eq. (22)], in view of the facile electron-transfer from
3
TDCPBSO H discussed above, although hydrogen peroxide
3
is a worse electron acceptor than molecular oxygen.
3
TDCPBSO H þ H O !
3
2
2
þ
ꢀ
TDCPBSO H C þ H O C ! TDCPBSO C þ OHC þ H O
3
2
2
3
2
ð22Þ
ꢀ
It has been suggested that the transient species H O is
Figure 10. Decays of triplet TDCPBSO H absorption at 790 nm in N -sa-
2
2
3
2
accessed in a Franck–Condon transition in a dissociative
turated PBS solutions with the following concentrations of H
and 3 mm. The inset shows the H O concentration dependence of the re-
2 2
ciprocal of the triplet lifetimes.
2
O
2
: 0, 1
ꢀ
[49]
electron attachment to H O that produces OHC and HO .
2
2
Given the very weakly exothermic electron transfer between
3
TDCPBSO H and O , we expect electron transfer to H O
2
3
2
2
to be endothermic and slow. Nevertheless, the relatively
The OHC radical is a highly reactive oxidant and much
3
long lifetime of TDCPBSO H in water gives us the oppor-
more cytotoxic than singlet oxygen or the superoxide ion,
and it may contribute significantly to the efficacy of PDT.
However, hydroxyl-radical generation through photocataly-
sis is intrinsically limited by the amount of excited photosen-
sitizer. On the other hand, it is very well known that OHC is
3
tunity to investigate the photoinduced electron transfer to
H O .
2
2
Before addressing the excited-state reaction with H O , it
2
2
is important to note that aqueous solutions of TDCPBSO H
3
are remarkably stable in the dark, even in the presence of
H O . The half-life of this bacteriochlorin in aqueous solu-
very efficiently produced in the Haber–Weiss/Fenton reac-
ꢀ
tion, which consists of an iron reduction step by O C and an
2
2
2
[51]
tions with [H O ]=5 mm is approximately one week. We
OHC generation step through the Fenton reaction,
as
2
2
3
measured the lifetimes of TDCPBSO H in N -saturated
given in Equations (23) and (24).
3
2
PBS solutions (pH 7.4) with various amounts of H O , at
2
2
3
þ
ꢀ
2
! Fe2þ þ O2
Fe þ O C
ð23Þ
ð24Þ
208C. Figure 10 shows the reciprocals of the triplet lifetimes
as a function of [H O ], from which we estimate the rate
2
2
Fe þ H O ! Fe þ OHC þ OHꢀ
2
þ
3þ
7
ꢀ1 ꢀ1
constant for electron transfer to H O , k =3ꢄ10 m s .
H2O2
2
2
2
2
This is much higher than the rate constants of reactions
shown in Equations (20) and (21), and makes photocatalysis
the preferred mechanism for hydroxyl-radical generation by
bacteriochlorins. This rate should correspond to an outer-
sphere electron-transfer reaction that is endothermic by
[
52]
These reactions have been shown to take place in vivo.
Hence, on the basis of these reaction mechanisms, we argue
that co-locating the ferrous ion with a bacteriochlorin photo-
sensitizer may increase the efficacy of PDT.
ꢀ
1 [50]
5
kcalmol , as expected from the low electron affinity of
The EPR spectra of TDCPBSO H in the presence of
3
H O . It must be emphasized that in aerated solutions, elec-
DMPO in aerated DMSO under diode laser irradiation
show that the superoxide ion is formed under these condi-
tions. Since DMSO is an aprotic solvent, it prolongs the life-
2
2
tron transfer to H O competes with quenching by O , which
2
2
2
4
occurs with a rate constant of / kdiff. The production of the
9
ꢀ
hydroxyl radical only becomes effective with the depletion
time of O2C and the stability of the DMPOꢀOOH adduct,
of O from the solution and the concomitant accumulation
favouring the observation of this adduct. In summary, these
observations are consistent with the hypothesis that the su-
peroxide radical is produced from halogenated bacterio-
chlorins, and that the superoxide radicals in turn form hy-
droxyl radicals in aqueous solutions.
2
of H O .
2
2
Additional evidence for photocatalysis was found in the
pH dependence of the photodegradation of TDCPBSO H.
3
The bleaching is accelerated at lower pH and slowed at
higher pH with respect to PBS solutions, under laser irradia-
tion. Acidic media favour the formation of H O , but only
ꢀ
The generation of O2C in the simplified mechanism is
competitive with the charge recombination in the charge-
transfer complex [Eqs. (15) and (16) above]. This generation
occurs by separation of the radical–ion pair, which is nearly
2
2
in the presence of light does this species lead to efficient
degradation of TDCPBSO H.
3
Chem. Eur. J. 2010, 16, 9273 – 9286
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9283

L. G. Arnaut et al.
ꢀ
isoenergetic with the free ion in ethanol. A similar separa-
tion in acetonitrile takes place with a rate ksep =8ꢄ10 s .
efficient for the O /O C pair unless favoured by spin conser-
2 2
8
ꢀ1 [53]
vation. This means that only the species in the singlet chan-
nel are likely to evolve along this channel.
It is reasonable to assume that the separation is slightly
8
ꢀ1
slower in ethanol, and we should have ksep ꢄ5ꢄ10 s . We
defined before k =k +k and determined the value of
The presence of a charge-transfer complex exacerbates
the laser energy dependence of singlet-oxygen generation by
triplet bacteriochlorins. Although bacteriochlorins generate
singlet oxygen efficiently at low laser intensities, the amount
of singlet oxygen that they generate at high laser intensities
is much less than that of porphyrin sensitizers. Current pro-
tocols for PDT dosimetry are simply based on the product
3
sep
ꢀet
8
ꢀ1
k . Now we can estimate k =2ꢄ10 s for reactions with
3
ꢀet
0
ꢀ
ꢀ1
exothermicities DG ꢄꢀ25 kcalmol . Charge recombina-
et
0
ꢀ
[50,54]
tion reactions are usually much faster at this DG
.
et
However, for radical–ion pairs formed from triplet states,
charge recombination rates of this magnitude have been
[55]
ꢀ2
measured. Thus, in polar solvents and for triplet donors
with low oxidation potentials, it is possible to produce the
superoxide ion.
between the fluence rate of light (in units of Wcm or
ꢀ
2
ꢀ1
[56]
Jcm s ) and exposure time. However, the full potential
of bacteriochlorins can only be exploited if lower light inten-
sities and longer irradiation times are employed in the ther-
apy, which does not change the dosimetry, but changes the
dosimetry mode. Repeating with bacteriochlorins the same
protocols as with porphyrins is not efficient. It was recently
reported that the laser fluence rate was a principal factor in
PDT efficacy. Empirically, it was found that compensating
a decrease in laser power with an increase in photosensitizer
dosage, which formally yields the same amount of singlet
oxygen, leads to dramatically different PDT efficacies. The
combination of lower laser power with higher photosensitiz-
er dosage is the most effective. This observation cannot
be accommodated by the usual argument of molecular
oxygen depletion at the irradiation site, but finds support in
our mechanism of energy and electron transfer. Interesting-
ly, molecular oxygen depletion and hydrogen peroxide accu-
mulation in photosensitization by bacteriochlorins favours
the photocatalytic pathway for the generation of the hydrox-
yl ion, which is one of the most potent reactive oxygen spe-
cies. The generation of this ROS may be further enhanced
by using ferrous iron as an adjuvant to PDT.
The mechanisms and kinetics of singlet-oxygen and super-
oxide-ion generation have been analyzed thoroughly. Inter-
system crossing between triplet and singlet pathways occurs
efficiently only in the charge-separated states. Singlet
oxygen is formed preferentially from the encounter com-
Conclusion
The bacteriochlorins employed in this work can generate
both singlet oxygen and superoxide ions. The quantum
yields of singlet-oxygen generation are high, and the energy-
transfer rates are accelerated by the presence of charge-
transfer interactions. These bacteriochlorins are particularly
interesting for photodynamic therapy (PDT) because of
their strong absorption in the phototherapeutic window,
photochemical stability, and wide range of 1-octanol/water
solubilities. The detailed analysis of the mechanism of
quenching of their triplet states by oxygen provides an im-
portant insight into this ubiquitous photochemical reaction.
For the first time it has been possible to develop a quantita-
tive understanding of the processes involved in the genera-
tion of singlet oxygen and superoxide ions, and to measure
the elementary rates involved. Scheme 4 summarizes the in-
formation obtained on these systems. This scheme considers
the possibility of singlet-oxygen generation from a CT com-
plex. This can be regarded as a weakly exothermic charge
[57]
[57]
0
ꢀ1
recombination, DG
ꢄꢀ4 kcalmol , which must compete
ꢀ
CT
with the very exothermic charge recombination to the
0
ꢀ
ꢀ1
ground state, DG
ꢄꢀ25 kcalmol , and is likely to be in-
CT
ꢀ
1
Scheme 4. Rates (in s ) and energies in the photosensitization of singlet oxygen and superoxide ion by porphyrins or bacteriochlorins.
9284
www.chemeurj.org
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 9273 – 9286

Porphyrins and Bacteriochlorins
FULL PAPER
plex, and the radical-ion pair is the precursor of the super-
oxide ion. The observed triplet decay constants are not a
direct measure of the internal conversion in the encounter
complex, to generate singlet oxygen, in the nCT channel.
Nor are such decay constants a measure of the charge sepa-
ration rates in the CT channel. In both cases it is necessary
to refer to Equation (7) to obtain the actual internal conver-
sion or charge separation rates. The correct interpretation of
the observed rates may contribute to establishing more fun-
damental relations between the molecular structure of pho-
tosensitizers and the kinetics and thermodynamics of the re-
action they initiate.
liquid nitrogen chamber (Products for Research, model PC176TSCE005).
Excitation was achieved with the third harmonic of Nd:YAG lasers
(
2
Spectra-Physics Quanta Ray GCR 130, 5–6 ns FWHM, or EKSPLA PL
143 A, 30 ps pulse width).
The modification of the spectrometer for time-resolved singlet-oxygen
phosphorescence measurements involved the interposition of a Melles
Griot cold mirror (03MCS005), which reflects more than 99% of the inci-
dent light in the 400–700 nm range, and of a Scotch RG665 filter. A 600
line diffraction grating was mounted in place of a standard one. This
equipment allows spectral identification of the singlet-oxygen phosphor-
escence and measurement of the singlet-oxygen lifetime in the nanosec-
ond and microsecond ranges. The filters employed are essential for elimi-
nating from the infrared signal all harmonic contributions of the sensitiz-
er emission in the 400–900 nm range. Singlet oxygen quantum yields in
[
36]
ethanol were obtained with a procedure described elsewhere,
using
[
34]
phenalenone as reference, for which F
D
=0.95ꢃ0.02 in ethanol.
Photoacoustic calorimetry (PAC) was performed with the front-face cell
design. The sample, reference and solvent solutions were flowed sepa-
Experimental Section
[
60]
ꢀ
1
rately with a 1 mLmin rate (SSI chromatographic pump) through a cell
of thickness 0.2 mm. They were irradiated at 355 nm with a Spectra-Phys-
ics Quanta Ray GCR 130 at a frequency of 10 Hz. A small fraction of
the laser beam was reflected to a photodiode, used to trigger the transi-
ent recorder (Tektronix DSA 601, 1GSa/s). The photoacoustic waves, de-
tected with a 2.25 MHz Panametrics transducer (model 5676) and cap-
tured by the transient recorder, were transferred to a PC for data analy-
sis. In a typical PAC experiment, 200 waves of the sample, reference, and
pure solvent were recorded and averaged in the same experimental con-
ditions. Four sets of averaged sample, reference, and solvent waves were
used for the data analysis at a given laser intensity, and four laser intensi-
ties were employed in each experiment. The different laser intensities are
obtained by interposing neutral density filters with transmissions between
Synthesis: The syntheses of 5,10,15,20-tetrakis(2,6-dichlorophenyl) por-
[
21,58]
phyrin (TDCPP), followed the methods described in the literature.
ꢀ
3
In the chlorosulfonation of TDCPP, TDCPP (100 mg, 0.11ꢄ10 mol)
and chlorosulfonic acid (6 mL) were stirred at 1008C for 3 h. After cool-
ing, chloroform (200 mL) was added, and the excess of chorosulfonic
acid was washed out with water in a continuous process. The organic
layer was dried, and after workup, the solvent was removed. The
5
,10,15,20-tetrakis(2,6-dichloro-3-chlorosulfonylphenyl) porphyrin was
obtained in rather pure form by this process, with a 90% yield, but, if
necessary, the crude product can be further purified by silica gel column
chromatography using chloroform as eluent. MS (FAB), m/z: 1285
+
1
[
M+H] ; H NMR (300 MHz, CDCl
Ph-H), 8.11–8.05 (m, 4H, Ph-H), ꢀ2.51 ppm (s, 2H, NH); elemental anal-
ysis calcd (%) for C44 : C 41.15, H 1.41, N 4.36; found: C
0.70, H 1.60, N 4.30.
The desired porphyrin derivatives TDCPPSO
3
): d=8.66–8.62 (m, 12H, b-H and
2
5
5
and 100%. The measurements were made using manganese
18 4 8 4
H N O Cl12S
,10,15,20-tetraphenylporphyrin (MnTPP) as photoacoustic reference
4
[
21]
(
excitation at 355 nm).
Electron paramagnetic resonance (EPR) measurements: Reactive
oxygen species produced by irradiation of TDCPBSO H in phosphate
3
H, TDCPPSNHethyl, and
TDCPPSNHheptyl were obtained after reaction of 5,10,15,20-tetra-
kis(2,6-dichloro-3-chlorosulfonylphenyl) porphyrin with water, ethyl-
3
buffer (PBS) or dimethylsulfoxide (DMSO) solutions, namely the hy-
droxyl radical and the superoxide ion, form adducts with various spin
traps. Two spin traps were employed: 5,5-dimethylpyrroline-N-oxide
ACHTUNGTRENNUNGa mine, or heptylamine. The physical properties of the compounds are in
[
25]
good agreement with previously described data.
The porphyrins TDCPPSO H, TDCPPSNHethyl, and TDCPPSNHheptyl
3
(
(
DMPO) and 5-tert-butoxycarbonyl 5-methyl-1-pyrroline N-oxide
BMPO). The adducts formed were identified by EPR. EPR measure-
were thoroughly mixed with p-toluenesulfonylhydrazide (1:35), placed in
a degassed reactor, and heated to 1508C for 10 min. This procedure
ments were performed at room temperature using a Bruker ESP 300
spectrometer (IBM Instruments Inc.). The EPR spectra were recorded
under in situ irradiation with a Hamamatsu diode laser (748 nm). Typical
instrument settings for superoxide detection were: microwave power
yielded
substantially
pure
bacteriochlorins
3
TDCPBSO H,
TDCPBSNHethyl, and TDCPBSNHheptyl. Purification and characteriza-
[59]
tion details are presented elsewhere.
Photochemistry: Absorption and luminescence spectra were recorded at
room temperature with a Shimadzu UV-2100 spectrophotometer and
SPEX Fluorolog 3.22 spectrophotometer, respectively. Most photochemi-
cal studies employed photosensitizer concentrations adjusted to produce
absorbances in the 0.2–0.25 range at the excitation wavelengths. We
found no evidence for sensitizer aggregation at these concentrations. The
absorptions of both reference and sample solutions in fluorescence quan-
tum yield measurements were matched at ꢄ0.2 at the excitation wave-
length of 515.5 nm, and then the solutions were diluted by a factor of 10
before collecting the fluorescence. The fluorescence quantum yields were
obtained from the ratio of the fluorescence bands of the samples vs. the
reference, multiplied by the fluorescence quantum yield of the reference,
after correction for the difference in refractive indexes between the
sample and reference solutions. The reference employed was 5,10,15,20-
1
0 mW, modulation amplitude 0.8 G, sweep width 60.0 G. Slightly differ-
ent settings were employed to register the BMPOꢀOH adduct: micro-
wave power 4 mW, modulation amplitude 0.2 G, narrow scan range 60 G,
and 20 scans were recorded for each spectrum. The Hamamatsu diode
laser, type LA0873, S/N M070301, delivered 100 mW at 748 nm. This
diode laser was controlled by a ThorLabs 500 mA ACC/APC Laser
Diode Controller. The laser energies of this and the other higher-energy
lasers employed in this work were regularly checked with an Ophir
model AN/2E laser power meter. The EPR spectra were simulated using
[
61]
the software EPRsim32.
The PBS employed in these measurements
was previously treated with chelating resin, Chelex 100, in order to
remove any contaminating metal ions that may catalyze the decomposi-
tion of peroxides. DMPO was first purified with activated charcoal/ben-
zene, and then a 1.0m stock concentration was determined spectrophoto-
metrically using e226 =7200m cm
The EPR experiments were performed in the presence of 30–80 mm of
TDCPBSO H and 40 mm BMPO in PBS, or with 50 mm of TDCPBSO H
3
and 100 mm DMPO in DMSO, under the following conditions: air-satu-
rated solutions in the dark, nitrogen-saturated solutions irradiated for 1–
15 min with the diode laser, air-saturated solutions irradiated for 1–
F
tetrakis(2,6-dichlorophenyl) bacteriochlorin (TDCPB), for which F =
ꢀ
1
ꢀ1
[
4]
.
0
.012 in toluene.
Transient triplet–triplet absorption was obtained with an Applied Photo-
physics LKS.60 flash photolysis spectrometer with the R928 photomulti-
plier from Hamamatsu for detection and HP Infinium (500 MHz,
3
ꢀ
1
ꢀ1
1
GSas ) or Tektronix DPO 7254 (2.5 GHz, 40 GSas ) oscilloscopes.
An adaptation of this spectrometer allowed the detection of singlet-
oxygen phosphorescence at room temperature. This emission was detect-
ed using a Hamamatsu R5509–42 photomultiplier, cooled to 193 K in a
ꢀ
1
15 min with the diode laser in the presence of catalase (30 mgmL ) or
superoxide dismutase (50 mgmL ).
ꢀ
1
Chem. Eur. J. 2010, 16, 9273 – 9286
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9285

L. G. Arnaut et al.
Acknowledgements
[27] A. S. M. Ressurreiżo, M. Pineiro, L. G. Arnaut, A. M. d. A. R.
Gonsalves, J. Porphyrins Phthalocyanines 2007, 11, 50–57.
[
[
28] D. J. Quimby, F. R. Longo, J. Am. Chem. Soc. 1975, 97, 5111–5117.
29] S. K. Pandey, A. L. Gryshuk, A. Graham, K. Ohkubo, S. Fukuzumi,
M. P. Dodhal, G. Zheng, Z. Ou, R. Zhan, K. M. Kadish, A. Oseroff,
S. Ramaprasad, R. K. Pandey, Tetrahedron 2003, 59, 10059–10073.
30] M. Pineiro, M. M. Pereira, A. M. d. A. R. Gonsalves, L. G. Arnaut,
S. J. Formosinho, J. Photochem. Photobiol. A 2001, 138, 147–157.
31] A. Harriman, J. Chem. Soc. Faraday Trans. 2 1981, 77, 1281–1291.
32] S. L. Murov, I. Carmichael, G. L. Hug, Handbook of Photochemistry,
Marcel Dekker, New York, 1993.
This work was funded by Fundażo para a CiÞncia e a Tecnologia (Portu-
gal) and FEDER (project no. ERA-CHEM/0002/2008). Support from
CCDRC and Cꢆmara Municipal de Coimbra is gratefully acknowledged.
C.J.P.M. and E.F.F.S. thank FCT for grants BD/37652/2007 and BD/
[
4
6658/2008. This work was also supported in part by the National Centre
for Research and Development (Era Chemistry grant nr. 60 303). The au-
thors thank Dr. M. Elas for providing us with the spin probe (mHCTPO)
and spin trap (BMPO), Prof. Z. Sojka for the EPRsim32 simulation pro-
gram, and Dr. K. Kruczała for his help with EPR spectra simulation.
[
[
[
[
33] M. A. J. Rodgers, J. Am. Chem. Soc. 1983, 105, 6201–6205.
34] R. Schmidt, C. Tanielian, R. Dunsbach, C. Wolff, J. Photochem.
Photobiol. A 1994, 79, 11–17.
[
1] P.-C. Lo, J.-D. Huang, D. Y. Y. Cheng, E. Y. M. Chan, W.-P. Fong,
W.-H. Ko, D. K. P. Ng, Chem. Eur. J. 2004, 10, 4831–4838.
2] J. Arnbjerg, A. Jimꢇnez-Banzo, M. J. Paterson, S. Nonell, J. I.
Borrel, O. Christiansen, P. R. Ogilby, J. Am. Chem. Soc. 2007, 129,
[
[
35] R. Schmidt, C. Tanielian, J. Phys. Chem. A 2000, 104, 3177–3180.
36] J. M. Dabrowski, M. M. Pereira, L. G. Arnaut, C. J. P. Monteiro,
A. F. Peixoto, A. Karocki, K. Urbanska, G. Stochel, Photochem.
Photobiol. 2007, 83, 897–903.
[
5
188–5199.
[
[
37] C. Tanielian, C. Wolff, J. Phys. Chem. A 1995, 99, 9825–9830.
38] H. Zhao, J. Joseph, H. Zhang, H. Karoui, B. Kalyanaraman, Free
Radical Biol. Med. 2001, 31, 599–606.
[
3] C. Liu, M. P. Dobhal, M. Ethirajan, J. R. Missert, R. K. Pandey, S.
Balasubramanian, D. K. Sukumaran, M. Zhang, K. M. Kadish, K.
Ohkubo, S. Fukuzumi, J. Am. Chem. Soc. 2008, 130, 14311–14323.
4] M. Pineiro, A. M. d. A. Rocha Gonsalves, M. M. Pereira, S. J. For-
mosinho, L. G. Arnaut, J. Phys. Chem. A 2002, 106, 3787–3795.
5] M. M. Pereira, C. J. P. Monteiro, A. V. C. Simꢂes, S. M. A. Pinto,
L. G. Arnaut, G. F. F. Sꢈ, E. F. F. Silva, L. B. Rocha, S. Simꢂes, S. J.
Formosinho, J. Porphyrins Phthalocyanines 2009, 13, 567–573.
6] C. Hopper, Lancet Oncol. 2000, 1, 212–219.
[
[
39] A. Weller, Z. Phys. Chem. 1982, 133, 93–98.
40] C. Huang, M. Tian, Y. Yang, F. Guo, M. Wang, J. Electroanal. Chem.
[
[
1
989, 272, 179–184.
[
[
[
[
[
41] J. Fajer, D. C. Borg, A. Forman, R. H. Felton, D. Dolphin, L. Vegh,
Proc. Natl. Acad. Sci. USA 1974, 71, 994–998.
42] M. Tezuka, Y. Ohkatsu, T. Osa, Bull. Chem. Soc. Jpn. 1976, 49,
[
[
[
1
435–1436.
43] K. Kawaoka, A. U. Khan, D. R. Kearns, J. Chem. Phys. 1967, 46,
842–1853.
44] M. Bodesheim, M. Schꢉtz, R. Schmidt, Chem. Phys. Lett. 1994, 221,
–14.
45] C. Rizzi, A. Samouilov, V. K. Kutala, N. L. Parinandi, J. L. Zweier,
P. Kuppusamy, Free Radical Biol. Med. 2003, 35, 1608–1618.
46] B. G. Ershov, A. V. Gordeev, Radiat. Phys. Chem. 2008, 77, 928–935.
47] J. Weinstein, B. H. J. Bielski, J. Am. Chem. Soc. 1979, 101, 58–62.
48] S. J. Blanksby, V. M. Bierbaum, G. B. Allison, S. Kato, Angew.
Chem. 2007, 119, 5036–5038.
7] T. J. Dougherty, J. Clin. Laser Med. Sci. 2002, 20, 3–7.
8] C. M. Moore, D. Pendse, M. Emberton, Nat. Clin. Pract. Urol. 2009,
1
6
, 18–30.
[
9] F. H. van Duijnhoven, J. P. Rovers, K. Engelmann, Z. Krajina, S. F.
Purkiss, F. A. N. Zoetmulder, T. J. Vogl, O. T. Terpstra, Ann. Surg.
Oncol. 2005, 12, 808–816.
7
[
10] J. Trachtenberg, A. Bogaards, R. A. Weersink, M. A. Haider, A.
Evans, S. A. McCluskey, A. Scherz, M. R. Gertner, C. Yue, S. Appu,
A. Aprikian, J. Savard, B. C. Wilson, M. Elhilali, J. Urol. 2007, 178,
[
[
[
1
974–1979.
[
[
[
11] M. Hoebeke, H. J. Schuitmaker, L. E. Jannink, T. M. A. R. Dubbel-
[
[
[
49] D. Nandi, E. Krishnakumar, A. Rosa, W.-F. Schmidt, E. Illenberger,
Chem. Phys. Lett. 2003, 373, 454–459.
50] S. J. Formosinho, L. G. Arnaut, R. Fausto, Prog. React. Kinet. 1997,
man, A. Jakobs, A. Van de Vorst, Photochem. Photobiol. 1997, 66,
5
02–508.
12] Y. Vakrat-Haglili, L. Weiner, V. Brumfeld, A. Brandis, Y. Salomon,
B. McIlroy, B. C. Wilson, A. Pawlak, M. Rozanowska, T. Sarna, A.
Scherz, J. Am. Chem. Soc. 2005, 127, 6487–6497.
13] S. Fukuzumi, K. Ohkubo, X. Zheng, Y. Chen, R. K. Pandey, R.
Zhan, K. M. Kadish, J. Phys. Chem. B 2008, 112, 2738–2746.
14] L. W. Oberley, G. R. Buettner, Cancer Res. 1979, 39, 1141–1149.
15] C. Schweitzer, R. Schmidt, Chem. Rev. 2003, 103, 1685–1757.
16] F. Wilkinson, A. A. Abdel-Shafi, J. Phys. Chem. A 1997, 101, 5509–
516.
17] A. A. Abdel-Shafi, F. Wilkinson, J. Phys. Chem. A 2000, 104, 5747–
757.
18] A. A. Abdel-Shafi, F. Wilkinson, Phys. Chem. Chem. Phys. 2002, 4,
48–254.
2
3, 1–90.
51] T. Nunoshiba, F. Obata, A. C. Boss, S. Oikawa, T. Mori, S. Kawa-
nishi, K. Yamamoto, J. Biol. Chem. 1999, 274, 34832–34837.
[52] J. A. Imlay, S. M. Chin, S. Linn, Science 1988, 240, 640–642.
[53] I. R. Gould, S. Farid, J. Phys. Chem. 1992, 96, 7635–7640.
[54] C. Serpa, P. J. S. Gomes, L. G. Arnaut, S. J. Formosinho, J. Pina, J.
Seixas de Melo, Chem. Eur. J. 2006, 12, 5014–5023.
[
[
[
[
[
55] C. Serpa, L. G. Arnaut, J. Phys. Chem. A 2000, 104, 11075–11086.
56] General Medical Physics Committee Task Group #5, Photodynamic
Therapy Dosimetry, Vol. 88 (Ed.: AAPM), AAPM, College Park,
5
[
[
[
5
2
005.
[
57] M. Seshadri, D. A. Bellnier, L. A. Vaughan, J. A. Spernyak, R. Ma-
2
zurchik, T. H. Foster, B. W. Henderson, Clin. Cancer Res. 2008, 14,
19] R. Schmidt, F. Shafii, C. Schweitzer, A. A. Abdel-Shafi, F. Wilkin-
son, J. Phys. Chem. A 2001, 105, 1811–1817.
2
796–2805.
[
58] A. M. d. A. R. Gonsalves, J. M. T. B. Varej¼o, M. M. Pereira, J. Het-
erocycl. Chem. 1991, 28, 635–640.
[
[
20] R. Schmidt, Photochem. Photobiol. 2006, 82, 1161–1177.
21] M. Pineiro, A. L. Carvalho, M. M. Pereira, A. M. d. A. R. Gonsalves,
L. G. Arnaut, S. J. Formosinho, Chem. Eur. J. 1998, 4, 2299–2307.
22] B. Wang, P. R. Ogilby, J. Photochem. Photobiol. A 1995, 90, 85–89.
23] L. G. Arnaut, S. J. Formosinho, A. M. da Silva, J. Photochem. 1984,
[
59] M. M. Pereira, C. J. P. Monteiro, A. V. C. Simꢂes, S. M. A. Pinto,
G. F. F. Sa, E. F. F. Silva, L. Rocha, J. M. D a˛ browski, S. J. Formosin-
ho, S. Simꢂes, L. G. Arnaut, unpublished results.
[
[
[
60] L. G. Arnaut, R. A. Caldwell, J. E. Elbert, L. A. Melton, Rev. Sci.
Instrum. 1992, 63, 5381–5389.
2
7, 185–203.
[
[
[
24] E. G. Azenha, A. C. Serra, M. Pineiro, M. M. Pereira, J. Seixas de
Melo, L. G. Arnaut, S. J. Formosinho, A. M. d. A. R. Gonsalves,
Chem. Phys. 2002, 280, 177–190.
[
61] T. Spalek, P. Pietrzyk, Z. Sojka, J. Chem. Inf. Model. 2005, 45, 18–
2
9.
25] C. J. P. Monteiro, M. M. Pereira, S. M. A. Pinto, A. V. C. Simꢂes,
G. F. F. Sꢈ, L. G. Arnaut, S. J. Formosinho, S. Simꢂes, M. F. Wyatt,
Tetrahedron 2008, 64, 5132–5138.
Received: January 15, 2010
Please note: Minor changes have been made to this manuscript since its
publication in Chemistry—A European Journal Early View. The Editor.
Published online: June 22, 2010
26] M. Pineiro, PhD Thesis, University of Coimbra (Portugal), 2001.
9286
www.chemeurj.org
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 9273 – 9286