Photoacoustic Measurements of Porphyrin Triplet-State Quantum Yields
and Singlet-Oxygen Efficiencies
Marta Pineiro, Ana L. Carvalho, Mariette M. Pereira, A. M. dA. Rocha Gonsalves,
Luís G. Arnaut,* and Sebastia
Ä
o J. Formosinho
Abstract: Photoacoustic calorimetry was used to measure the quantum yields of
singlet molecular oxygen production by the triplet states of tetraphenylporphyrin
(TPP), Zn
ÿ
TPP and Cu
ÿ
TPP in toluene, yielding values of 0.67 0.14, 0.68 0.19
and 0.03 0.01, respectively. We show that a novel dichlorophenyl derivative of
Zn
ÿ
TPP is capable of singlet-oxygen production with a 0.90 0.07 quantum yield.
The synthesis and characterisation of a new photostable chlorin with high absorptivity
in the red that is capable of singlet-oxygen production with 0.54 0.06 quantum yield
is described. Our results suggest that chlorinated chlorins may be interesting new
sensitisers for photodynamic therapy.
Keywords: metalloporphyrins ´
photoacoustic calorimetry ´ por-
phyrinoids ´ singlet oxygen ´ sensi-
tizers
Introduction
The ubiquitous tetrapyrrolic macrocycles play highly diverse
roles in biological systems.
[1±5]
The natural roles of these
structures stimulated the search for new applications, exploit-
ing in particular the use of new synthetic porphyrins.
[6]
One of
the more recent and promising applications of porphyrin
chemistry in medicine is in the detection and cure of
tumours,
[7, 8]
referred to as photodynamic therapy (PDT).
The first reports of clinical trials of haematoporphyrin
derivatives (HPD) in PDT were followed by systematic
research for improved sensitisers over the last 20 years.
[9]
A good photosensitiser must be able to selectively photo-
damage the tumour tissue, while being irradiated with visible
or, preferably, near-infrared light. Two mechanisms are
possible. In one mechanism the excited photosensitiser reacts
directly with substrate molecules in the tissue by electron- or
hydrogen-transfer reactions (Type I process). In the other, it
transfers energy to the ground state of molecular oxygen,
generating singlet oxygen (
1
D
g
), which is the tissue-damaging
species (Type II process). Evidence favours the role of the
Type II photooxygenation process in cells.
[7, 8]
Adequate
sensitisers have specific biological and photochemical proper-
ties. The desired biological features of the sensitiser are:
1) little or no dark toxicity
2) selective accumulation and prolonged retention in tumour
tissues
3) controlled photofading to reduce the unwanted skin
photosensitivity side effects and increase light penetration
during therapy.
The chemical and photochemical requisites are:
1) stability, purity and long shelf-life
2) high absorption coefficient in the phototherapeutic win-
dow (600 ± 1000 nm)
3) high quantum yield for singlet molecular O
2
(
1
D
g
) sensi-
tisation.
The most important precursor of singlet oxygen is the
triplet state of the sensitiser, and a high singlet-oxygen
quantum yield requires at least three sensitiser triplet-state
properties: a near-unity quantum yield (F
T
1), an electronic
energy at least 20 kJ mol
ÿ1
above that of singlet oxygen (E
D
94 kJ mol
ÿ1
), and a long lifetime (t
T
> 5 ms).
The quantum yield of the triplet state of porphyrins and
related macrocycles is a critical quantity in determining their
efficiency in PDT. However, uncertainties persist concerning
the triplet quantum yields of basic members of this family of
compounds. For example, studies on the triplet quantum yield
of 5,10,15,20-tetraphenylporphyrin (TPP) reported values
ranging from as low as F
T
0.67 0.07
[10]
to as high as F
T
0.88 0.03.
[11]
The triplet quantum yield of zinc 5,10,15,20-
tetraphenylporphyrinate (Zn
ÿ
TPP) is also subject to some
scatter; values of F
T
0.86
[12]
and F
T
0.97
[13]
can be found in
[*] L. G. Arnaut, M. Pineiro, A. L. Carvalho,
M. M. Pereira, A. M. dA. Rocha Gonsalves, S. J. Formosinho
Chemistry Department, University of Coimbra
3049 Coimbra Codex (Portugal)
Fax : ( 351)39-27703
E-mail: lgarnaut@cygnus.ci.uc.pt
S. J. Formosinho
Escola Superior de Cie
Ã
ncias e Tecnologia
Universidade Cato
Â
lica Portuguesa
3500 Viseu (Portugal)
FULL PAPER
Chem. Eur. J. 1998, 4, No. 11 WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998 0947-6539/98/0411-2299 $ 17.50+.25/0
2299
FULL PAPER
L. G. Arnaut et al.
WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998 0947-6539/98/0411-2300 $ 17.50+.50/0 Chem. Eur. J. 1998, 4, No. 11
2300
the literature. The various methods available for measuring
the quantum yield of singlet-oxygen production have been
recently reviewed,
[14]
and the value of 0.62 was selected as the
standard value for F
D
of TPP in aerated CCl
4
. Other F
D
values relevant to this work are 0.68 ± 0.93 for Zn
ÿ
TPP in
aerated benzene or toluene and < 0.01 for Cu
ÿ
TPP in aerated
CCl
4
. The use of photothermal methods to measure F
D
,
pioneered by Braslavsky and co-workers,
[15, 16]
is particularly
relevant to this work.
This work reports the use of photoacoustic calorimetry
(PAC),
[17, 18]
to measure energy-transfer rates and singlet-
oxygen sensitisation quantum yields for a selected range of
porphyrins and a chlorin (Figure 1). Some of these species are
new and include basic structures that can be derivatized with
more polar groups, notably hydroxyl and sulfonamide,
[19]
which modulate the solubility of the photosensitisers and
their selective accumulation in specific tissues.
[20]
Results
Porphyrin synthesis: 5,10,15,20-Tetrakisarylporphyrins have a
simple basic structure, but their characteristics can be
modified by peripheral structural changes. The methodology
first described by Rothemund in 1935
[21]
and modified by
Adler in 1967
[22]
was recently improved by Rocha Gonsalves
et al.
[23]
This methodology is now useful for the preparation of
large amounts of a wide range of pure porphyrins, which were
previously very difficult or impossible to prepare. A source of
potentially useful new porphyrins for PDT applications
became available.
In our approach, the aldehyde and pyrrole are made to
react in acetic acid or propionic acid in the presence of
nitrobenzene at 120 8C to give, very often by direct crystal-
Abstract in Portuguese: A calorimetria fotoacu
Â
stica foi
utilizada para determinar os rendimentos qua
Ã
nticos de forma-
çaÄo de oxigØnio singuleto molecular pelos estados tripleto de
TPP, Zn
ÿ
TPP e Cu
ÿ
TPP em tolueno, onde TPP representa a
tetrafenilporfirina na sua forma protonada ou desprotonada,
tendo-se obtido 0.67 0.14, 0.68 0.19 e 0.031 0.01, respec-
tivamente. Um novo derivado da Zn
ÿ
TPP, com a
Â
tomos de
cloro nas posiçoÄes orto do grupo fenilo, produz oxigØnio
singuleto com um rendimento qua
Ã
ntico de 0.90 0.07. E
Â
descrita a síntese e feita a caracterizaçaÄo de uma nova clorina
foto-esta
Â
vel, capaz de produzir oxigØnio singuleto com um
rendimento qua
Ã
ntico de 0.54 0.06. Estes resultados sugerem
que clorinas cloradas podem ser sensibilizadores apropriados
para a terapia fotodina
Ã
mica.
Figure 1. Structural diagram of porphyrin and chlorin macrocycles. Abbreviations: TPP, 5,10,15,20-tetraphenylporphyrin; Zn
ÿ
TPP, Zn
II
5,10,15,20-
tetraphenylporphyrinate; Cu
ÿ
TPP, Cu
II
5,10,15,20-tetraphenylporphyrinate; Mn
ÿ
TPP, Mn
III
5,10,15,20-tetraphenylporphyrinate; Zn
ÿ
TDFPP, Zn
II
5,10,15,20-tetrakis(2,6-difluorophenyl)porphyrinate; Zn
ÿ
TDCPP, Zn
II
5,10,15,20-tetrakis(2,6-dichlorophenyl)porphyrinate; TNP, 5,10,15,20-tetranaphthyl-
porphyrin; Zn
ÿ
TNP, Zn
II
5,10,15,20-tetranaphthylporphyrinate; TNC, 5,10,15,20-tetranaphthylchlorin.
Porphyrins 2299 ±2307
Chem. Eur. J. 1998, 4, No. 11 WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998 0947-6539/98/0411-2301 $ 17.50+.25/0
2301
lisation, the corresponding porphyrin free of any chlorin
contamination. Yields of porphyrins are presented in Figure 1,
and their full characterisation is described in the Experimen-
tal Section. The effect of nitrobenzene as oxidant and
aromatizing agent was described in the synthesis of other
aromatic compounds
[24, 25]
and also in the oxidation of
porphyrinogen to porphyrins or chlorins.
[26]
The Zn and Cu complexes were easily prepared by
refluxing the porphyrins in the presence of the desired acetate
salt with dimethylformamide as solvent.
[27]
Yields are presented in the Experimental
Section. Our interest in obtaining compounds
with high absorption coefficients prompted us
to synthesise 5,10,15,20-tetranaphthylchlorin
(TNC) by refluxing the corresponding por-
phyrin in g-picoline in the presence of p-
toluenesulfonylhydrazine and sodium carbo-
nate for about 6 hours.
[28]
After preparative
TLC purification, the new 5,10,15,20-tetra-
naphthylchlorin was isolated; its full charac-
terisation is presented in the Experimental
Section.
Porphyrin photophysics and photochemistry:
The relevant data from the absorption and
luminescence spectra of the porphyrins and
chlorin are summarised in Table 1. The con-
centrations used in our spectroscopic and PAC
studies were in the 10
ÿ7
± 10
ÿ5
m range. The
Beer ± Lambert Law was always obeyed in this
concentration range, and we found no evi-
dence for aggregation. The present results are
in good agreement with the available literature data for TPP
and the corresponding porphyrinates.
[29±31]
The spectroscopic
singlet-state energies (E
S
) were obtained from the intersec-
tion of the normalised absorption and fluorescence spectra. In
Figure 2 we show the absorption, fluorescence excitation and
fluorescence emission spectra of TNC. The Stokes shifts of the
free bases are very small, and the spectroscopic energies are
nearly identical to the relaxed energies of the singlet states.
The same is probably true for the triplet states
[32]
, and the
Table 1a. Absorption and luminescence data of the free bases in deaerated toluene solutions.
Absorption l
max
[nm] (e [m
ÿ1
cm
ÿ1
]) Fluorescence
l
max
[nm] (RT)
Phosphorescence
l
max
[nm] (77 K)
E
S
[kJ mol
ÿ1
]
F
F
E
T
[kJ mol
ÿ1
]
Q
x
(0 ± 0) Q
x
(1 ± 0) Q
y
(0 ± 0) Q
y
(1 ± 0) B(0 ± 0) Q(0 ± 0) Q(0 ± 1) T(0 ± 0) T(0 ± 1)
TPP 649.8 592.0 548.0 514.6 418.8 652 719 183.9 0.3 0.10 0.01 138.0
[a]
(9.6 10
3
) (1.0 10
4
) (1.16 10
4
) (1.8 10
4
) (2.67 10
5
)
TNP 654.0 589.5 548.5 514.0 423.0 654 715 766 778 183.8 0.3 0.16 0.02 153.9 1.9
(5.16 10
3
) (8.61 10
3
) (7.67 10
3
) (2.4 10
4
) (3.81 10
5
)
TNC 652.0 601.8 542.8 517.0 423.4 657 721 775 182.9 0.7 0.36 0.02 154.5 0.7
(3.7 10
4
) (1.9 10
3
) (4.5 10
3
) (1.0 10
4
) (1.57 10
5
)
[a] Ref [27].
Table 1b. Absorption and luminescence data of the free bases in deaerated toluene solutions.
Absorption l
max
[nm] (e [m
ÿ1
cm
ÿ1
]) Fluorescence l
max
[nm] (RT) Phosphorescence
l
max
[nm] (77 K)
E
S
[kJ mol
ÿ1
] F
F
E
T
[kJ mol
ÿ1
]
Q(0 ± 0) Q(1 ± 0) B(0 ± 0) Q(0 ± 0) Q(0 ± 1) Q(0 ± 2) T(0 ± 0) T(0 ± 1)
Zn-TPP 588.4 549.6 423.4 600 648 715 757 ± 201.5 2.0 0.033 0.005 160.0 1.0
(2.8 10
3
) (1.21 10
4
) (4.52 10
5
)
Zn-TNP 583 549.2 426.2 595 644 770 ± 203.5 2.1 0.077 0.005 157.0 1.0
(8.28 10
2
) (1.56 10
4
) (2.73 10
5
)
Zn-TDFPP 546.0 420.8 588 644 720 737 ± 211.5 7.8 < 10
ÿ4
160.3 1.0
(1.45 10
4
) (3.15 10
5
)
Zn-TDCPP 586.6 550.4 423.8 595 652 725 754 ± 202.7 1.4 < 10
ÿ4
163.9 1.9
(3.6 10
3
) (1.22 10
4
) (2.42 10
5
)
Cu-TPP 539.8 418.4
[a] [a]
738 ± 205.3
[b] [a]
163.7 1.2
(1.65 10
4
) (2.4 10
5
)
[a] We observe a residual fluorescence atributable to TPP, suggesting that our sample of Cu-TPP is only 98% pure. [b] Ref. [29].
Figure 2. Absorption, fluorescence excitation and emission (normalised) spectra of TNC in
toluene at room temperature. The phosphorescence spectrum of TNC was obtained in
deaerated toluene at 77 K.
FULL PAPER
L. G. Arnaut et al.
WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998 0947-6539/98/0411-2302 $ 17.50+.50/0 Chem. Eur. J. 1998, 4, No. 11
2302
maximum of the highest energy phosphorescence band was
used to obtain the triplet energies (E
T
) presented in Table 1.
The Stokes shifts of the zinc porphyrins correspond to 3 ±
4 kJ mol
ÿ1
. Thus, we add 1.5 kJmol
ÿ1
to the energy corre-
sponding to the maximum of the first vibrational band in the
phosphorescence spectrum of Zn ± porphyrins to obtain the
E
T
energies presented in Table 1. The same procedure was
followed to obtain the triplet energy of Cu
ÿ
TPP from the the
phosphorescence maximum obtained in this work (l
max
738 nm), which is in good agreement with that reported by
Harriman.
[31]
Time-resolved PAC is based on the measurement of the
acoustic wave generated by the heat released in the non-
radiative processes following electronic excitation. The ex-
perimental wave (E-wave) of the sample studied is compared
with that of the pressure transducer (T-wave). The T-wave is
obtained with a calorimetric reference absorbing the same
fraction of light as the sample and releasing it as thermal
energy in a time much shorter than the transducer oscillation
frequency. The phase and amplitude differences between the
T- and E-wave allow for the simultaneous determination of
the thermal energy released by the transients and their
lifetimes. Typical background-corrected reference and sample
signals are shown in Figure 3.
Co
ÿ
TPP and Zn
ÿ
TPP have been used as PAC references.
[32]
Whereas the first of these compounds is radiationless, the use
of the second one in PAC requires a correction for its
fluorescence. The use of 5,10,15,20-tetrakis(p-sulfonylphe-
nyl)porphyrin as calorimetric reference in aqueous solu-
tions
[33]
has been shown to be inadvisable.
[34]
The same is
probably true for any compounds yielding long-lived tran-
sients. For this reason, we selected Mn
ÿ
TPP as the calori-
metric reference for our PAC studies in the visible and tested
it against reliable calorimetric references in the UV and
visible. It is known than Mn
ÿ
TPP is very weakly lumines-
cent,
[35, 36]
is soluble in a wide range of solvents and absorbs
strongly in the 350 ± 550 nm region. The absorption spectrum
of Mn
ÿ
TPP did not change as a result of prolonged irradiation
with the N
2
laser. We compared the acoustic waves of
Mn
ÿ
TPP and 2-hydroxybenzophenone (HBP), an established
PAC reference for UV irradiation,
[17]
using the N
2
laser. The
linearity of the photoacoustic response of Mn
ÿ
TPP with the
fraction of laser energy absorbed is not distinguishable from
that of HBP in toluene solutions. We also tested the photo-
acoustic response of Mn
ÿ
TPP with irradiation at 337 nm in
ethanol/water (1:1 by volume) against that of K
2
CrO
4
. When
the fraction of energy absorbed is less than 50 % of the N
2
-
laser energy, the photoacoustic responses of Mn
ÿ
TPP and
K
2
CrO
4
solutions are linear (correlation coefficient better
than 0.990) and have indistinguishable slopes. Mn
ÿ
TPP and
K
2
CrO
4
give slightly larger waves than HBP in this solvent
mixture. It is known that the ground-state repopulation of
HPB in non-hydrogen-bonding solvents is 35 5 ps, but in
ethanol a fraction of the molecules populate the triplet state,
which has a 1.5 ns lifetime.
[37]
This advises against the use of
HPB as a PAC reference in hydrogen-bonding solvents.
Finally, we tested Mn
ÿ
TPP against trans-b-carotene (Aldrich)
in toluene at 421 nm, because trans-b-carotene has a singlet-
state lifetime of 8.4 0.6 ps
[38]
and a fluorescence quantum
yield of 6 10
ÿ5
.
[39]
The photoacoustic responses of Mn
ÿ
TPP
and trans-b-carotene are linear with the fraction of laser
energy absorbed and have identical slopes.
We interpret the waves of N
2
-saturated
samples with two sequential exponentials, the
first one for the formation of the triplet state of
the sensitiser and the second one for its decay
(Figure 4). The formation of the triplet state is
faster than the time resolution of our experi-
ments, and we arbitrarily set the lifetime of the
first exponential decay to t
1
1 ns; smaller
values of t
1
do not change the other parameters
in the deconvolution. This is not strictly true for
Cu
ÿ
TPP, because t
2
is small. For this system we
set t
1
to 0.1 ns. We interpreted the waves of air-
saturated samples with three sequential expo-
nentials, the second one representing the two
decay channels now available for the triplet
(energy transfer to oxygen or nonradiative
decay to the ground state), and the last one
associated with the decay of singlet oxygen.
Each decay step is described by two parame-
ters: the lifetime of the transient and the
fraction of thermal energy released in that
lifetime (Figure 4). The convolution of the
reference wave with parameters of the kinetic
model for the decay of transient species gives a
calculated E-wave. The appropriateness of the
kinetic model and its parameters to describe
the observed E-wave can be evaluated by the
difference between the amplitudes of observed
Figure 3. Typical sample photoacoustic wave, E-wave (obs) and reference wave (T-wave),
obtained in a PAC experiment. The E- and T-waves depicted were corrected for the
background signal and normalised. The normalisation factor is the reciprocal of the largest
absolute value of the T-wave. The sample (TNP), reference (Mn
ÿ
TPP) and solvent (toluene)
data were obtained under the following experimental conditions : irradiation at 517 nm of N
2
-
saturated solutions with a filter with 93 % transmittance; absorbance of 1.50 at 517 nm;
solution flow of 1 mL min
ÿ1
. The calculated wave, E-wave (calcd), was obtained with two
sequential exponential decays with lifetimes t
1
1 ns and t
2
5.5 ms and fractions of heat
released f
1
0.2995 and f
2
0.3052. Res E-wave(calc) ÿ E-wave(obs).
Porphyrins 2299 ±2307
Chem. Eur. J. 1998, 4, No. 11 WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998 0947-6539/98/0411-2303 $ 17.50+.25/0
2303
and calculated E-waves at each decay time. As shown in
Figure 3, that difference is typically 1/100 of the amplitude of
the sample wave. The decay parameters were obtained by
deconvolution of the background-corrected and normalised
E- and T-waves with the algorithm described by Melton and
co-workers.
[40]
The fractions of laser energy released by each system were
measured at four different laser intensities. In some systems,
the first fraction of energy released varied with the laser
intensity. We used the Students t test, at the 95 % confidence
level, to decide whether the values obtained at the different
laser energies were significantly different. When the differ-
ences were significant, we plotted the first fraction of energy
released as a function of the laser energy and obtained linear
correlation coefficients greater than 0.96. The difference was
assigned to transient ± transient absorption and was corrected
by extrapolating the fraction of energy released to zero laser
intensity.
Discussion
The absorption and emission spectra of porphyrins and
chlorin exhibit the typical features of this class of compounds.
It is worth mentioning the large molar absorption coefficient
of the Q
x
(0,0) band of TNC (e
652
37 000 m
ÿ1
cm
ÿ1
).
The series of halogenated Zn
II
complexes show fluores-
cence quenching due to the heavy-atom effect in the
intersystem-crossing rate. This effect was also described by
Quimby and Longo
[30]
for halogen substituents positioned on
phenyl rings of the free base and Zn
II
5,10,15,20-tetrakis-
(2-chlorophenyl)porphyrinates in benzene. However, these
authors reported that the fluores-
cence yield of Zn
II
5,10,15,20-tetra-
kis(2-chlorophenyl)porphyrinate
was larger than that of Zn
ÿ
TPP.
This unexplained result has no
parallel in Zn
ÿ
TDCPP. Actually
we find that introducing two halo-
gens in the ortho positions of the
phenyl ring is as effective in
quenching fluorescence as halo-
genation in the meso positions of
the porphyrin.
[41]
Energy conservation in N
2
-satu-
rated samples requires that the
energy of the laser light absorbed
(E
hn
) be given by Equation (1),
where E
F
is the integrated radiative
E
hn
(N
2
) E
F
E
hn
f
1
E
hn
f
2
(1)
energy of the singlet state, f
1
and
f
2
are the fractions of laser energy
released as thermal energy in the
lifetimes t
1
and t
2
, respectively.
The spectroscopic energy of the
singlet (E
S
) is an upper limit of E
F
.
This is given by Equation (2),
where I
F
and I
0
are the intensities of emitted and absorbed
light, and accounts for the energy emitted by the singlet at
each frequency. For a Gaussian emission band, Equation (2)
can be approximated by Equation (3), where E
nÄ
max
is the
energy at the maximum fluorescence intensity.
E
F
E
F
(n) I
F
(n)/I
0
dn (2)
E
F
E
n
max
F
F
(3)
When two or more emission bands are observed, they can
be broken into a series of Gaussians, each centred around a
maximum. The thermal energy released in a time shorter than
the resolution of the 2.25 MHz transducer (t
1
< 10 ns) is due
to the formation of the relaxed singlet followed by the
formation of the triplet and by the internal conversion to the
ground state (internal conversion quantum yield F
IC
). The
ground-state species formed by fluorescence also relax in this
time window and contribute (DE
r
F
F
) to the thermal energy
dissipated in lifetime t
1
[Eq. (4)].
E
hn
f
1
(E
hn
ÿ E
S
) (E
S
ÿ E
T
)F
T
E
S
F
IC
DE
r
F
F
(4)
The thermal energy released in the longer decay is
associated with the triplet-state energy [Eq. (5)], but triplets
with lifetimes longer than t
2
> 10 ms are difficult to follow with
E
hn
f
2
E
T
F
T
(5)
the 2.25 MHz transducer. In such cases, F
T
can be obtained
from Equation (4) from the value of f
1
determined by PAC,
E
T
from phosphorescence measurements (Table 1) and Equa-
Figure 4. Photoinduced processes: a) in the absence and b) in the presence of molecular oxygen. Full lines:
radiative processes; dashed lines: radiationless processes. DE
r
represents the relaxation energy of the
ground-state species formed radiatively from the lowest excited singlet state. In our PAC measurements the
radiationless processes in a) are described by two sequential exponentials: formation of the triplet state and
internal conversion to the ground state, followed by decay of the triplet. In b) there are three sequential
exponentials: formation of the triplet state and internal conversion to the ground state, followed by decay of
the triplet by intersystem crossing or energy transfer to oxygen and, finally, the decay of singlet oxygen.
![Table 2. Quantum yields and triplet lifetimes in toluene measured by time-resolved photoacoustic calorimetry.[a]](/figures/table-2-quantum-yields-and-triplet-lifetimes-in-toluene-11zc3w6i.webp)
