Heavy-atom effects on metalloporphyrins and polyhalogenated porphyrins

Article abstract of DOI:10.1016/S0301-0104(02)00485-8

The photophysics of halogenated and metallated tetrakisphenylporphyrins is investigated using single-photon counting, photoacoustic calorimetry and luminescence techniques. The radiationless transition rates in these and related molecules are interpreted with a quantum-mechanical tunnelling model modified to include the effect of spin-orbit coupling in the intersystem crossing rates. It is shown that tetrakisphenylporphyrins with halogens in the ortho positions have long-lived triplet states that are formed in high yields.

Full text of DOI:10.1016/S0301-0104(02)00485-8

Chemical Physics 280 (2002) 177–190
www.elsevier.com/locate/chemphys
Heavy-atom effects on metalloporphyrins and
polyhalogenated porphyrins
a
Emılia G. Azenha , Armenio C. Serra , Marta Pineiro , Mariette M. Pereira ,
b
a
a
ꢀ
ꢀ
a
a,
a,1
*
~
J. Seixas de Melo , Luis G. Arnaut , Sebastiao J. Formosinho ,
A.M.d’A. Rocha Gonsalvesa
a
Chemistry Department, University of Coimbra, 3049 Coimbra Codex, Portugal
b
Universidade Catoꢀlica Portuguesa, P-3080 Figueira da Foz, Portugal
Received 23 January 2002; in final form 18 March 2002
Abstract
The photophysics of halogenated and metallated tetrakisphenylporphyrins is investigated using single-photon
counting, photoacoustic calorimetry and luminescence techniques. The radiationless transition rates in these and related
molecules are interpreted with a quantum-mechanical tunnelling model modified to include the effect of spin–orbit
coupling in the intersystem crossing rates. It is shown that tetrakisphenylporphyrins with halogens in the ortho positions
have long-lived triplet states that are formed in high yields. Ó 2002 Elsevier Science B.V. All rights reserved.
Keywords: Porphyrins; Radiationless transitions; Tunnel-effect theory; Heavy-atom effect
1. Introduction
and in the photodynamic destruction of viruses
[11,12]. The present work was motivated by the
need for better PDT sensitisers based on porphy-
rins, chlorins and bacteriochlorins [13–15], that is,
sensitisers with high absortivity in the photother-
apeutic window (600–900 nm) that generate singlet
oxygen in high yields. The most important pre-
Porphyrin derivatives represent an important
class of pigments in the storage of energy in both
natural and artificial photosynthesis [1–4], on the
making of molecular energy devices [5,6], catalysts
[7], and more recently as sensitisers of specific
1
1
products (e.g., the singlet Dg state of molecular
oxygen) in photodynamic therapy (PDT) [8–10]
cursor of the O2ð DgÞ state is the triplet state T1 of
the sensitiser. Therefore, an ideal sensitiser must
have a long-lived triplet state ðsT > 5 lsÞ formed
in high yields ðUT ꢀ 1Þ, and a triplet energy at least
20 kJ molÀ1 above that of the O2ð DgÞ state
1
* Corresponding author. Tel.: +351-39-822826; fax: +351-39-
827703.
E-mail address: lgarnaut@ci.uc.pt (L.G. Arnaut).
ðET > 115 kJ molÀ1Þ.
In order to maximise the quantum yield for
triplet formation one usually resorts to an internal
heavy-atom effect [16]. A quite extensive research
1
^
Present address: Escola Superior de Ciencias e Tecnologia,
ꢀ
Universidade Catolica Portuguesa, 3500 Viseu (Portugal).
0301-0104/02/$ - see front matter Ó 2002 Elsevier Science B.V. All rights reserved.
PII: S0301-0104(02)00485-8

178
E.G. Azenha et al. / Chemical Physics 280 (2002) 177–190
has been undertaken on the study of such effects in
halogenated porphyrins and metalloporphyrins
[1,13,17–19]. In this work we have revisited some
old data and supplement it with studies on new
molecules. We present the energies and quantum
yields (U) of formation of S1 and T1 states of
halogenated tetrakisphenylporphyrins and zinc-
tetrakisphenylporphyrins, and other metallo-
tetrakisphenylporphyrins (MeTPP) (Scheme 1)
together with their lifetimes ðsÞ and intersystem-
crossing rate constants ðkiscÞ. This systematic study
of heavy-atom effects is rationalised through the
use of a model for radiationless transitions in large
molecules [20]. The application of this model re-
veals trends that may be useful in the design of
new molecules with pre-required properties.
analysis on EA1108-CHNS-0 Fisons Instruments.
Absorption and luminescence spectra were re-
corded with Shimadzu UV-2100 and SPEX flu-
oromax 322-2 spectrophotometers, respectively.
Home-made equipment was employed in single-
photon counting and photoacoustic calorimetry
measurements.
All the solutions were carefully de-aerated with
N2 saturated in toluene before the photophysical
measurements were made.
Fluorescence quantum yields ðUFÞ were deter-
mined as described previously [13], with excitation
at the maximum of the Soret band. Fluorescence
excitation spectra were obtained for all the por-
phyrins and metalloporphyrins and agreed well
with the corresponding absorption spectra. The
phosphorescence studies were carried out in tolu-
ene at liquid nitrogen temperature, with excitation
at the Soret band.
2. Experimental
Fluorescence lifetimes were measured with an
apparatus using an IBH 5000 coaxial flashlamp,
filled with N2; D2 or H2 (or mixtures of them), as
excitation source, Philips XP2020Q photomulti-
plier, with wavelength selected with Jobin-Ivon
H20 monochromator and Canberra instruments
time-to-amplitude converter and multichannel
analyser. Alternate measurements (1000 counts
per cycle at the maximum) of the pulse profile at
420 nm and the sample emission were performed
until 5 Â 103 counts at the maximum were reached.
The fluorescence decays were analysed using the
method of modulating functions of Striker, with
automatic correction for the photomultiplier
‘‘wavelength shift’’ [21].
All solvents and reagents were purified using
standard procedures.
2.1. Instrumentation and methods
The 1H NMR spectra were recorded with a 300
MHz Bruker-AMX spectrometer. Mass spectra
were obtained on VG autospec and elemental
Time-resolved photoacoustic calorimetry (PAC)
measurements were made with a front-face cell
using a dielectric mirror [22], that allows for a
minimum background and maximum sensitivity of
these measurements. Our PAC apparatus and the
procedure employed in the measurements was re-
cently described in detail [14,23]. In short, the
sample and reference solutions, and the solvent,
were flown separately with a 1 ml/min rate (SSI
chromatographic pump) through a 0.11 mm thick
cell. They were irradiated at 421 nm with an un-
focused PTI dye laser (model PL202), pumped by a
N2 laser (PL2300) working at a frequency of 2 Hz.
A small fraction of the laser beam was reflected to a
Scheme 1. Porphyrins and metalloporphyrins used in this work.

E.G. Azenha et al. / Chemical Physics 280 (2002) 177–190
179
photodiode, used to trigger the transient recorder
(Tektronix DSA 601, 1 Gs/s). The photoacoustic
waves, detected with a 2.25 MHz Panametrics
transducer (model 5676) and captured by the
transient recorder, were transferred to a PC for
data analysis. In a typical PAC experience 100
waves of the sample, reference and pure solvent are
recorded and averaged in the same experimental
conditions. Four sets of averaged sample, reference
and solvent waves are used for the data analysis at
a given laser intensity, and four laser intensities are
employed in each experiment. The different laser
intensities are obtained by interposing neutral
density filters with transmissions between 25% and
100%. All the measurements were made in toluene
using either Mn(III)-5,10,15,20-tetrakisphenylpor-
phyrinate acetate or all trans-b-carotene as
photoacoustic reference.
following general procedure. Arylaldehyde (0.4
mmol) was dissolved in a mixture of acetic acid
(140 ml, 2.45 mol) and nitrobenzene (70 ml, 0.68
mol), and the temperature was raised to 120 °C.
Pyrrole (2.8 ml, 0.4 mmol) was then added. The
reaction was carefully followed by UV–Vis until
the Soret band intensity reaches its maximum. At
this point the solution was cooled to room tem-
perature to give porphyrin crystals, which were
filtered off and washed with methanol. The por-
phyrin was recrystallised with dichloromethane/
methanol. The data for H2TPP, H2TPPF8;
H2TPPCl4 and H2TPPCl8 characterisation are in
close agreement with our previous results [13].
Zinc(II) 5,10,15,20-tetrakis(2-bromo-5-methoxy-
phenyl)porphyrinate and zinc(II) 5,10,15,20-tetra-
kis(2,4-dibromo-5-methoxyphenyl)porphyrinate
were prepared according to
a
previous
described method [27]. To a solution of 120 mg
(0.11 mmoles) of the corresponding porphyrin in
30 ml of CHCl3, 100 mg (0.46 mmoles) of
ZnðOAcÞ2Á 2H2O in 4 ml of methanol was added.
The mixture was refluxed for 2 h to complete
metallation. The mixture was diluted with dichlo-
romethane, washed with water, dried over Na2SO4
and evaporated to dryness. The residue was
chromatographed on silica gel (CH2Cl2/petroleum
ether 40–60, 8:2) to give the corresponding zinc
complex. Recrystallisation in CH2Cl2/methanol
gave an analytical sample.
2.2. Synthesis
The synthesis of 2-bromo-5-methoxybenzalde-
hyde, 2,4-dibromo-5-methoxybenzaldehyde and
2,4,6-tribromo-5-methoxybenzaldehyde followed
literature procedures [24]; After recrystallisation in
water/ethanol we obtained 2-bromo-5-methoxy-
benzaldehyde, 2,4-dibromo-5-methoxybenzalde-
hyde or 2,4,6-tribromo-5-methoxybenzaldehyde
with 83%, 93% and 91% yields, respectively.
Characterisations are presented below.
5,10,15,20-Tetrakis(2-bromo-5-methoxyphenyl)-
porphyrin and 5,10,15,20-tetrakis (2,4-dibromo-
5-methoxyphenyl)porphyrin were synthesised
according to Lindsey procedure [25]. A solution of
1.8 mmol of the corresponding aldehyde and 0.13 ml
(1.8 mmol) of pyrrole in 300 ml of distilled CH2Cl2
was purged with N2 for 15 min, then 0.3 ml of
BF3 Á OðEtÞ2 was added at room temperature. The
solution was left during 3 h then neutralised with
triethylamine. Then DDQ, 0.8 g (3.5 mmol) in 40 ml
of toluene was added and the mixture is left over-
night. The solvent was removed and the residue was
chromatographed in alumina (CH2Cl2/petroleum
ether 40–60) giving the porphyrin as a red band.
Recrystallisation in CH2Cl2/methanol gave an an-
alytical sample.
Zinc(II)
5,10,15,20-tetrakis(2,4,6-tribromo-
5-methoxyphenyl)porphyrinate was synthesised
from a solution of 1.75 g (4.9 mmoles) of 2,4,6-
tribromo-5-methoxybenzaldehyde and 0.35 ml (4.9
mmoles) of pyrrol in 600 ml of distilled CH2Cl2
previously purged with N2. After 20 min, 0.2 ml of
BF3 Á OðEtÞ2 was added at room temperature. The
solution was left overnight, neutralised with tri-
ethylamine and concentrated in vacuum. The res-
idue was dissolved in 20 ml of CH2Cl2 and 100 ml
of acetic acid and 1.0 g of zinc acetate was added.
The mixture was stirred 30 min at room temper-
ature and then 1 ml of H2O2 (50%) was added.
After 2 h the solvent was evaporated in vacuum.
Then, 100 ml of CH2Cl2 were added to the residue
and the mixture was washed successively with
water, hydrogen carbonate (10%) and water. The
solvent was removed in vacuum and the residue
The other porphyrins were prepared according
to a previously described method [26], with the

180
E.G. Azenha et al. / Chemical Physics 280 (2002) 177–190
1
chromatographed on alumina (CH2Cl2Þ to give
0.39 g of the zinc complex of 5,10,15,20-tetra-
kis(2,4,6-tribromo-5-methoxyphenyl)porphyrin as
a red band. Recrystallisation in CH2Cl2/MeOH
gave an analytical sample.
5.3%; H NMR (CDCl3): d (ppm) ¼ 8.9–8.7 (m,
8H, b-H), 7.9–7.3 (m, 12H, Ar), 4.2–3.9 (m, 12H,
Me), )2.6–2.7 (m, 2H, NH); Ms(FAB) m=z ¼ 1050
(molecular ion); V/UV (CH2Cl2): kmax (log e
(MÀ1 cmÀ1)) ¼ 419 (5.52), 453 (4.1), 513 (4.29), 546
(3.66), 588 (3.79), 644 (3.32).
Zn(II) 5,10,15,20-tetrakisphenylporphyrinate,
Zn(II) 5,10,15,20-tetrakis(2,6-halogenophenyl) por-
phyrinates, Cd(II) 5,10,15,20-tetrakisphenylpor-
phyrinate and Mg(II) 5,10,15,20-tetrakisphenyl-
porphyrinate were synthesised with slight
modifications of Adler’s method [28], with the
following general procedure. Porphyrin (0.33
mmol) was dissolved in 80 ml of dry dimetylfor-
mamide and the temperature was raised to 150 °C.
Then, acetate salt (3.3 mmol) of the desired metal
was added to the reaction mixture. The reaction
was monitored by UV–Vis spectroscopy until the
degeneracy of the Q bands was observed. After
work-up, the metal complexes were obtained by
several recrystallisations with dichloromethane/
methanol 1:5 giving the metal complexes. The
characterisations of ZnTPP, ZnTPPF8 and
ZnTPPCl8 were in agreement with our previous
results [13]. Yields and characterisation of CdTPP
and MgTPP are described below.
5,10,15,20-Tetrakis(2,4-dibromo-5-methoxyphe-
nyl)porphyrin. Yield: 18%; C48H30N4O4Br8Á H2O:
Found: C, 41.3; H, 2.8; N, 3.6. Calcd.: C, 41.7%;
1
H, 2.3%; N, 4.0%; H NMR (CDCl3): d (ppm) ¼
8.75 (s (large), 8H, b-H), 8.20–8.17 (m, 4H, Ar),
7.73–7.61 (m, 4H, Ar), 3.95–3.88 (m, 12H, Me),
)2.72 (s, 2H, NH); Ms(FAB) m=z ¼ 1366 (mo-
lecular ion); V/UV (CH2Cl2): kmax (log
e
(MÀ1 cmÀ1)) ¼ 419 (5.52), 513 (4.43), 545 (sh), 587
(4.11).
Zinc(II) 5,10,15,20-tetrakis(2-bromo-5-methoxy-
phenyl)porphyrinate (ZnTPPBr4Þ. Yield, 94%;
C48H32N4O4Br4Zn Found: C, 51.5; H, 3.0; N, 4.8.
Calcd.: C, 51.8%; H, 2.9%; N, 5.0%; 1H NMR (300
MHz; CDCl3) d (ppm), 8.82–8.77 (m, 8H, b-H),
7.85–7.59 (m, 8H, Ar), 7.21–7.15 (m, 4H, Ar),
3.88–3.77 (m, 12H, Me); V/UV (CH2Cl2): kmax (log
e (MÀ1 cmÀ1)) ¼ 420 (5.62), 548 (4.37); Ms(FAB):
m=z ¼ 1114 (molecular ion).
Zinc(II) 5,10,15,20-tetrakis(2,4-dibromo-5-meth-
oxyphenyl)porphyrinate (ZnTPPBr4ðBr4Þ). Yield:
96%; C48H28N4O4Br8Zn Á MeOH: C, 40.7; H, 2.9;
2.3. Characterisation
1
2-Bromo-5-methoxybenzaldehyde. Yield: 83%;
mp 74–75 °C (lit. mp 76 °C) [29]; 1H NMR
(CDCl3): d (ppm) ¼ 10.11 (s, 1H, CHO), 7.36 (d,
1H, J ¼ 8:8 Hz, ArH), 7.22 (d, 1H, J ¼ 3:2 Hz,
ArH), 6.88 (dd, 1H, J ¼ 8:8 Hz, J ¼ 3:2 Hz, ArH),
3.74 (s, 3H, OCH3).
N, 3.1. Calcd.: C, 40.3%; H, 2.2%; N, 3.8%; H
NMR (300 MHz, CDCl3): d (ppm) ¼ 8.84–8.79 (m,
8H, b-H), 8.20–8.14 (m, 4H, Ar), 7.75–7.59 (m,
4H, Ar), 3.94–3.86 (m, 12H, Me); Ms(FAB):
m=z ¼ 1430 (molecular ion); V/UV (CH2Cl2): kmax
(log e (MÀ1 cmÀ1)) ¼ 422 (5.61), 549 (4.32).
2,4-Dibromo-5-methoxybenzaldehyde. Yield: 93%;
1
Zinc(II) 5,10,15,20-tetrakis(2,4,6-tribromo-5-meth-
oxyphenyl)porphyrinate (ZnTPPBr8ðBr4Þ). Yield:
18%; C48H24N4O4Br12ZnÁMeOH: Found: C, 34.0;
H, 1.71; N, 3.29. Calcd.: C, 33.1%; H, 1.59%; N,
mp 111–112 °C (lit. mp 110 °C) [29]; H NMR
(CDCl3): d (ppm) ¼ 10.24 (s, 1H, CHO), 7.83 (s,
1H, ArH), 7.39 (s, 1H, ArH), 3.94 (s, 3H, OCH3).
2,4,6-Tribromo-5-methoxybenzaldehyde. Yield:
91%; mp 111–112 °C (lit. mp 113 °C) [29]; 1H
NMR (CDCl3): d (ppm) ¼ 10.15 (s, 1H, CHO),
7.89 (s, 1H, CH–Ar), 3.91 (s, 3H, Me); 13C d
(ppm): 190.4 (CO), 154.7 (COMe), 137.3 (CH),
133.5 (C–CO), 123.2 (CBr), 121.3 (CBr), 118.9
(CBr), 60.6 (Me).
1
3.15%; H NMR (300 MHz; CDCl3): d (ppm) ¼
8.72 (s, 8H, b-H), 8.25 (s, 4H, Ar), 4.11 (s, 12H,
Me); Ms(FAB) m=z ¼ 1743 (molecular ion); V/UV
(CH2Cl2): kmax (log e (MÀ1 cmÀ1)) ¼ 426 (5.52), 554
(4.49).
Cadmium(II) 5,10,15,20-tetrakisphenylporphy-
rinate (CdTPP). Yield: 90%; C44H28N4CdÁ
7CH3OH, Found: C, 65.48; H, 5.43; N. 5.91
Calcd.: C, 64.5%; H, 5.9%; N, 5.9%; 1H NMR (300
MHz; CDCl3): d (ppm) ¼ 8.84 (s, 8H, b-H), 8.23
5,10,15,20-Tetrakis(2-bromo-5-methoxyphenyl)
porphyrin. Yield: 22%; C48H34N4O4Br4, Found: C,
55.2; H, 3.4; N, 5.3. Calcd.: C, 54.9%; H, 3.3%; N,

E.G. Azenha et al. / Chemical Physics 280 (2002) 177–190
181
(m, 8H, Ar), 7.75 (m, 12H, Ar); Ms (FAB):
m=z ¼ 725 (molecular ion).
cellent quality and we believe that it is a better
description of the lifetime of this species. We ex-
clude the possibility of quenching by an impurity
because this would lead to a decrease in UF.
Magnesium(II) 5,10,15,20-tetrakisphenylpor-
phyrinate (MgTPP). Yield: 80%; C44H28N4 MgÁ
3H2O Found: C, 76.84; H, 4.85; N, 9.76 Calcd.: C,
76.52%; H, 4.92%; N, 8.12%; 1H NMR (300 MHz;
CDCl3): d (ppm) ¼ 8.83 (s, 8H, b-H), 8.20 (d, 8H,
Ar), 7.70 (m, 12H, Ar); Ms (FAB): m=z ¼ 636
(molecular ion); V/UV (toluene): kmax (log e
(MÀ1 cmÀ1)) ¼ 426 (5.52), 564 (4.17), 604 (3.94).
4. Discussion
4.1. Experimental quantum yields
The discrepancy between the UF value of
MgTPP measured in this work and the values
found in the literature deserves a detailed discus-
sion, because MgTPP is a popular compound and
has been used as fluorescence [1]. We synthesised
two independent batches of MgTPP and for both of
them obtained crystals with excellent elemental
analysis (considering the presence of hydration
water molecules) and NMR. Two co-authors of
this work independently obtained UF ¼ 0:29 Æ 0:02
using fresh N2-saturated toluene solutions and
taking TPP as the fluorescence standard. In order
to explain the literature value, UF ¼ 0:15, we re-
peated the quantum yield measurements of the
same MgTPP solution in toluene at different times
after its preparation. The quantum yields dropped
rapidly with time, as shown in Fig. 2, reaching the
literature value in less than 24 h. The aged MgTPP
solutions have a new emission at 717 nm. The
fluorescence excitation spectrum of this emission
has a maximum at 418 nm, that is, at the same place
as the Soret band of TPP. The fluorescence exci-
tation of the fresh MgTPP solution does not show
this maximum, on the contrary, its maximum is at
426 nm, just like its absorption spectrum. We
conclude that MgTPP is prone to loose its metal in
solution and yields a mixture of free-base and
metallated TPP. It should not be used as fluores-
cence standard. We emphasise that we were par-
ticularly careful in neutralising our solvent after its
purification and did not find evidence for the
characteristic absorption and emission spectra of
protonated TPP in solution. Thus, de-metallation
may occur even in the absence of acids.
3. Results
Some photophysical properties of porphyrins
addressed in this work have already been mea-
sured in other laboratories [1,17,30,31]. We did
not measure the fluorescence lifetimes of the
species with UF 6 0:005 due to sensitivity limita-
tions at the wavelengths where the porphyrins
emit. Lifetimes shorter than 200 ps are beyond
the time resolution of our apparatus [32]. The
lifetimes of the brominated porphyrins were
ꢀ
measured at Centro de Quımica–Fısica Molecular
(Instituto Superior Tecnico, Lisboa). Thus, Table
ꢀ
ꢀ
1 collects the fluorescence and triplet yields, sin-
glet S1 and triplet T1 lifetimes and energies, and
S1 ! T1 and T1 ! S0 intersystem crossing rate
constants measured in our laboratory for all the
porphyrins synthesised in this work, except for
UF and sF of CdTPP, H2TPPCl8 and ZnTPPCl8.
For the reasons mentioned above, the singlet
lifetimes of these porphyrins were taken from the
literature.
With the exception of MgTPP and H2TPPF8,
our measurements are in good agreement with the
available data. We found that our sample of
MgTPP has a significantly larger fluorescence
quantum yield than the literature value,
UF ¼ 0:29 Æ 0:02 vs 0.15 [18] or 0.13 [33], but a
comparable lifetime, 7.6 vs 8.9 ns [1]. The reasons
for this discrepancy are discussed below. The
fluorescence lifetime of our H2TPPF8 sample is
appreciable smaller than that reported by Yang et
al. [1] 8.5 vs 15.1, although we agree on the value
of UF. Fig. 1 shows the fluorescence decay of
H2TPPF8 obtained with our single photon count-
ing apparatus. The monoexponential fit is of ex-
Triplet quantum yields were determined from
the triplet energy ðETÞ given by phosphorescence,
and from fraction of energy released in the for-
mation of the triplet state given by PAC ð/1Þ [13]

Table 1
Photophysical properties of porphyrins. Single state properties at room temperature in benzene, and triplet state at 77 K toluene matrix
a
a
n1 ðcmÀ1
Þ
Rn2 ðcmÀ1
Þ
Es
UF
sS ðnsÞ
ET
UT
kisc ðS1–T1Þ
sT ðT1–S0Þ
ðkJ molÀ1
Þ
ðkJ molÀ1
Þ
(ms)
H2TPP
MgTPP
ZnTPP
CdTPP
0.24
40.5
183.9
197.5
202.1
195.0
0.10b
0.29
0.033b
0.0004d
10.6
7.6
138.0c
150.6
153.7
146.9
0.73b
0.47
0.86b
6.9e7
6.2e7
4.3e8
1.4e10
115
30
390
1140
2.0
0.065d
0.90
2.5
H2TPPF8
H2TPPCl4
H2TPPCl8
2152
2348
4696
182.6
182.7
181.3
0.069
0.023
0.005
8.5
1.4
0.22f
0.84e
0.94e
0.98e
9.9ev
6.7e8
4.5e9
ZnTPPF8
ZnTPPCl8
390
390
390
390
390
2152
4696
211.5
202.7
194.0
194.8
193.2
0.026
0.003
0.032
0.0058
0.0050
2.0
163.2
159.5
159.7
159.5
159.5
0.99
1.02
0.90
1.02
0.97
4.1e8
2.9e9
45
19
0.34f
0.048
0.048
0.009
ZnTPPBr4
ZnTPPBr4ðBr4Þ
ZnTPPBr8ðBr4Þ
9840
9840
1.9e10
2.1e10
1.1e11
4.0
4.3
1.2
19 680
a Ref. [34].
b Ref. [13].
c Ref. [17].
d Ref. [30].
e The triplet quantum yields were assumed to be identical to the singlet oxygen quantum yields.
f Ref. [1].

E.G. Azenha et al. / Chemical Physics 280 (2002) 177–190
183
Fig. 1. Fluorescence decay and pulse instrumental response, with kexc ¼ 420 nm and kem ¼ 650 nm, for H2TPPF8 in toluene. The
quality of the fit can be judged from the chi-square autocorrelation function, and weighted residuals.
Fig. 2. Decrease of the fluorescence quantum yield ðUFÞ of MgTPP with time in toluene solution. (a) Fresh solution, UF ¼ 0:27; (b) 1 h
later, UF ¼ 0:22; (c) 2 h later, UF ¼ 0:20; (d) 4 h later, UF ¼ 0:19; (e) 24 h later, UF ¼ 0:15. The solutions were kept at room tem-
perature and exposed to air and ambient light between measurements. Before each fluorescence measurement, the sample used for
analysis was saturated with N2. TPP was used as fluorescence standard.

184
E.G. Azenha et al. / Chemical Physics 280 (2002) 177–190
ð1 À /1ÞEhm À UFEhm max
Within the Born–Oppenheimer approximation,
UT ¼
;
ð1Þ
the calculation of the radiationless transition rate
between the initial state a and the final state b of
large molecules is usually expressed in terms of the
product between an electronic and a nuclear factor
[37]
ET
where Ehm is the energy of the incident light and
Ehm max is the energy at the maximum fluorescence
intensity. We were unable to detect the phospho-
rescence of free-base porphyrins. Given their rel-
atively long lifetimes, we assumed that their UT
were identical to the corresponding singlet oxygen
quantum yields in aerated solution, given by [14]
2p
2
rateða ! bÞ ¼
jHabj FCWD;
ð3Þ
ꢁh
where Hab is the matrix element of the quantum-
mechanical Hamiltonian which governs the rate of
the transition, and FCWD is the Franck–Condon
factor, that is the number of substrates qualifying
as b per unit interval of energy. The interpretation
of internal heavy-atom effects in a series of com-
pounds usually focus on the expected dependence
of Hab on n, and employs constant or empirical
values for FCWD [38,39]. This approach is re-
stricted to families of closely related molecules
where only one parameter is changing while the
others remain constant. Systems involving multi-
ple substitutions at different sites of the molecules
concomitant with variable energy gaps require a
model that can simultaneously account for all
these factors.
Nearly 30 years ago, Formosinho showed that
the rates of radiationless transitions could be in-
terpreted as quantum mechanical tunnelling from
the bottom of the higher energy potential energy
surface to an isothermic position on the lower
energy surface, Fig. 3. Under the WKB approxi-
mation [20,40], the Tunnel Effect Theory (TET)
expresses the tunnelling rate as
UTET À Ehm/2
UD ¼
;
ð2Þ
ED
where ED is the singlet oxygen energy (94.1 kJ
molÀ1), /2 is the fraction of energy released in the
formation of singlet oxygen and UTET was ob-
tained from the first fraction of energy released,
according to Eq. (1).
4.2. Radiationless transitions
The main goal of this work is to rationalise the
intersystems crossing rates of porphyrins because
they control triplet quantum yields and lifetimes.
In the set of diamagnetic porphyrins selected for
this study, we can expect to correlate the variations
observed in kisc and sT with the presence of heavy
atoms in the structure of the porphyrins.
Heavy-atom effects in the transitions between
singlet and triplet states are usually interpreted in
terms of the coupling of the relativistic magnetic
field of the motion of a nuclear charge around an
electron considered to be fixed, and the magnetic
field of the electron–spin motion (spin–orbit cou-
pling). Those effects are proportional to the
(square of the) spin–orbit coupling factor, n, of an
atom with a high atomic number, Z, that out-
weighs all the other atoms (n ꢀ Z4Þ [16]. Addi-
tionally, just like other radiationless transitions in
the weak-coupling limit [35], intersystems crossing
rates present an inverse relationship with the dif-
ference in energy, DES–T, of the singlet and triplet
states and are proportional to the relative number
of vibrational modes, g, involved in the transition
[36]. Thus, even if the nature of the excited states is
not affected by the heavy-atom substitution and
symmetry characteristics do not reduce heavy-
atom/porphyrin orbital overlap, intersystems
crossing rates are a function of n2; DES–T and g.
ꢀ
ꢁ
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2p
Dx
2g
k ¼ vm exp
À
2lðD À EmÞ
;
ð4Þ
h
where v is the electronic factor, m is the frequency of
the vibrational mode ðmCH ¼ 3000 cmÀ1 and
mCC ¼ 1050 cmÀ1Þ, l its reduced mass ðlCH ¼ 0:923
and ðlCC ¼ 6Þ, D À Ev is the barrier height, Dx is
the barrier width defined in Fig. 3, and g is the
relative number of atoms. The exponential part of
the equation represents the Franck–Condon factor
between a vibrational wavefunction in the non-
adiabatic region of the initial electronic state and a
vibrational function in the final state, taken as a d-
function at the classical turning point of vibration.
Jortner and Ulstrup [41] have shown that the

E.G. Azenha et al. / Chemical Physics 280 (2002) 177–190
185
ꢂ
0.009 A upon electronic excitation from the S0 to
the S1 state [44–46]. This contraction of the CH
bonds directs tunnelling to the repulsive side of the
potential energy curves.
A simple way to describe the dependence of the
energy on the nuclear co-ordinates is to employ
Morse curves
ꢂ
ꢃ
ꢄ
ꢅꢆꢇ
2
V ðlÞ ¼ De 1 À exp À b l À leq
;
ð7Þ
where De is the depth of the potential well, leq is
the equilibrium bond length and the Morse pa-
rameter b is related to the anharmonicity of the
oscillator. The Morse curve of a CH bond in
benzene is given by De;CH ¼ 478:2 kJ molÀ1, bCH
¼
ꢂÀ1
ꢂ
1:972 A and leq;CH ¼ 1:080 A [47]. That of a
Cðsp2Þ–Cðsp3Þ bond can be built from its dissoci-
ation energy (D0 ¼ 368 kJ molÀ1), vibrational
frequency (mCC ¼ 1050 cmÀ1) [47], bond length,
Fig. 3. Anharmonic potential energy curve of the generalised
CH vibrational modes; Dx is the tunnelling barrier width on the
attractive side.
ꢂ
leq;CC ¼ 1:505 A and harmonic force constant,
ꢂÀ2
fCC ¼ 2:89 Â 103 kJ molÀ1
A
[48], using the re-
lation
Franck–Condon vibrational overlap factors of
some radiationless transitions can be recast is a
form that is practically identical with Eq. (4). The
importance of CH and CC stretching modes on
radiationless conversions depends on the number
of hydrogen (nH), carbon ðnCÞ and other atoms
(nX). The role of the relative number of atoms
nH
f ¼ 2Deb2:
ð8Þ
¼
We obtain De;CC ¼ 374:3 kJ molÀ1 and bCC
ꢂÀ1
1:965 A for the Morse curve of a Cðsp2Þ–Cðsp3Þ
bond.
The Morse curves obtained with these parame-
ters describe the energy variation associated with
the nuclear distortion of one oscillator, but the
transition involves all the promoting modes of a
certain type, CH or CC. It is well recognised that
the role played by CH stretching vibrations in ra-
diationless transitions of aromatic hydrocarbons is
better described in terms of local rather than nor-
mal modes [36]. A treatment of local modes as in-
dependent oscillators showed that a generalised CH
gH ¼
gC ¼
;
nH þ nC þ nX
ð5Þ
nC
nH þ nC þ nX
has been shown empirically by Siebrand [42].
This method of calculation of the Franck–
Condon factor requires the knowledge of the nu-
clear displacement between the initial and final
electronic states, and a function to describe the
dependence of the energy on the nuclear co-ordi-
nates. McCoy and Ross [43] showed that the
electronic transitions in polycyclic aromatic hy-
drocarbons (PAH) can be expressed in terms of a
bond-length displacement co-ordinate R,
vibration could be expressed by a force constant
p
ðnHÞfCH [20,40,49]. Given the relation expressed
by Eq. (8), we chose to express the Morse param-
eters of generalised CH and CC vibrations by
pffiffiffiffiffi
4
bðCHÞ ¼ nHbCH
;
ð9Þ
pffiffiffiffiffi
4
!
1=2
bðCCÞ ¼ nCbCC
:
X
R ¼
Drj2
;
ð6Þ
The contraction of the CH modes plays a
dominant role in the radiationless transitions of
benzene given their low reduced mass and thin
barrier. For example, the barrier width Dx for the
j
where Drj denotes the change in the length of the
jth bond. Each CH bond of benzene contracts by

186
E.G. Azenha et al. / Chemical Physics 280 (2002) 177–190
S1 ! S0 internal conversion in benzene is the dif-
ference between the generalised bond contraction
of the S0 state at the S1-state energy ðES1 ¼ 459
kJ molÀ1Þ, and the bond-length displacement be-
It is likely that larger PAH have smaller chan-
ges in R than benzene but, following Eq. (9), this is
compensated by a larger b(CH). We avoid specu-
lating about the value of R and the role of b, and
employ the parameters of benzene for all the cal-
culations with PAH. Fig. 4 compares the results of
our calculations for a selection of PAH with the
experimental rates compiled by Birks [52]. The
agreement between calculated and experimental
radiationless transition rates is quite remarkable,
especially because our selection of data includes a
large range of rates (14 orders of magnitude), en-
ergy gaps (33–459 kJ molÀ1Þ and relative number
of atoms (0:333 < gH < 0:5). This selection com-
prises molecules such as benzene, naphthalene,
phenanthrene, anthracene, pyrene, coronene, tet-
racene and azulene.
The internal conversion rate of H2TPP, kic ¼
1:6 Â 107 sÀ1, can be estimated from the experi-
mental data in Table 1. The tunnelling rates were
evaluated with nH ¼ 10, nC ¼ 4 and nX ¼ 24, that
is, only the C atom of each phenyl group bonded
to the meso site was involved in a local mode and
included in the calculation of the CC tunnelling
rate. We calculate a rate of 1:4 Â 106 sÀ1 using the
ꢂ
ꢂ
tween the two curves (R ¼ 0:022 A), Dx ¼ 0:199 A.
The crossing between the Morse curves of the S0
and S1 state gives the barrier height D À Ev ¼
2:088 Â 103 kJ molÀ1. The curves represented in
Fig. 3 were calculated with the data pertaining to
benzene and give a Franck–Condon factor of
3:5 Â 10À9. The electronic factor can be calculated
as electron tunnelling through a triangular [50], or,
more exactly, a rectangular barrier [51] with width
Dx and height IP À ES1, where IP ¼ 9:25 eV is the
ionisation potential of benzene. In the case of a
triangular barrier we obtain v ¼ 0:7 and in the case
of a rectangular barrier we have v ¼ 0:5, this gives
an effective transition frequency mv ¼ 5 Â 1013 and
a nuclear tunnelling rate of kTET ¼ 1:7 Â 105 sÀ1
.
According to Birks [52], the experimental rate is
kexp ¼ 4 Â 104 sÀ1. We obtain the internal conver-
sion rate in perdeuterated benzene replacing lCH
by lCD and mCH by mCD. We estimate 1 Â 102 sÀ1
and Birks reports 3 Â 102 sÀ1. The calculation of
the S1 ! S0 internal conversion rate in benzene
does not involve the fitting of any parameters, al-
though we recognise that Eq. (9) involves some
empiricism.
The analogous S2 ! S1 rate requires the CH
bond length changes from the S1 to the S2 states.
This is not experimentally known, although it is
reasonable to expect that it should be smaller than
for the S1 ! S0 transition. We reproduce the ex-
perimental S2 ! S1 internal conversion rate in
benzene, k ¼ 4 Â 1010 sÀ1 [52], with Dr ¼ 0:005 A.
ꢂ
The T1 ! S0 intersystem crossing has a nonadia-
batic factor of v ¼ 10À6 [53]. With this v and
ꢂ
Dr ¼ 0:006 A we reproduce the experimental rate
of 0.03 sÀ1 [52]. Given that Dr ¼ 0:009 A for the
ꢂ
ꢂ
S1 ! S0 transition and Dr ¼ 0:006 A for the
ꢂ
T1 ! S0 transition, we must have Dr ¼ 0:003 A for
the S1 ! T1 transition. The experimental S1 ! T1
intersystem-crossing rate, k ¼ 8:8 Â 106 sÀ1 [52], is
reproduced with v ¼ 2 Â 10À3. The relative varia-
tions in Dr and in v among the four electronic
states meet our expectations and are a good
ground to explore radiationless transitions in other
PAH.
Fig. 4. Correlation between experimental and calculated radi-
ationless transition rates in PAH. The transitions are as follows:
T1 ! S0 represented as circles, S1 ! S0 as triangles, S1 ! T1 as
squares and S2 ! S2 as lozenges. The filled symbols refer to
halogenated naphthalenes.

E.G. Azenha et al. / Chemical Physics 280 (2002) 177–190
187
benzene curves and 1:2 Â 1010 sÀ1 using the nH
factor in R and b. The actual curves for H2TPP are
somewhat in between these limiting approxima-
tions. All the calculations presented below used the
benzene curves for the CH modes, for consistency
with the calculations on PAH. The contribution of
the CC modes was estimated assuming that
RCC ¼ 2RCH. Under this approximation, the CC
modes give a negligible contribution to radiation-
less transitions in tetraphenylphorphyrins. How-
ever, in octaalkylporphyrins, where nH ¼ 6, nC ¼ 8
and nX ¼ 24, the number of CC local modes is
sufficiently large to contribute to the nonradiative
rate. The contribution of the CC local modes to
the S1 ! S0 internal conversion in octa-
alkylporphyrin is 2:5 Â 105 sÀ1, whereas that of the
CH bonds is 6:2 Â 105 sÀ1. The experimental rate,
estimated from the data published by Bonnett et al.
[31], is 3:1 Â 107 sÀ1. The differences between ex-
perimental and calculated rates may be assigned to
the approximations involved in the estimate of
DrCC and DrCH in these molecules, but they are of
no consequence in the estimate of spin–orbit cou-
plings.
Fig. 5. Correlation between calculated and experimental radi-
ationless transition rates in metallated tetrakisphenylporphyrins
(triangles, including halogens in the phenyl rings) and halo-
genated porphyrins (other symbols). Diamonds represent por-
phyrins with halogen substituents in alkyl positions, circles
b-pyrrole positions and squares in meso positions. The crosses
refer to tetrakisphenylporphyrins with halogens in the phenyl
rings. The filled and open symbols represent S1 ! T1 and
T1 ! S0 transitions, respectively.
4.3. Spin–orbit coupling
figure includes the porphyrins presented in Table
1, metal tetrakisphenylporphyrins [18], free-base
octaalkylporphyrins having halogen substituents
in alkyl, b-pyrrole and in meso positions [31]. We
did not include multiple meso-fluorinated alkyl-
porphyrins because the very strong electron with-
drawing effect of this substitution pattern may
alter the electronic properties of the porphyrin [1].
The heavy atom effect at different sites of the
porphyrin ring can be interpreted in terms of the
coefficients ci, because they are fitted to the ex-
perimental dependence of the kisc on n. Table 2
collects the empirical parameters involved in these
fits.
The coefficients ci were chosen to yield a slope
close to unity in the correlation between the log-
arithms of calculated and experimental rates. They
determine the magnitude of the electronic factor
and their numerical value also depends on the
Franck–Condon factor, because changing an H
atom by a halogen atom also changes DES–T and
The description of spin–orbit coupling requires
additional empirical parameters, as many as the
number n of chemically different sites where heavy
atoms are introduced. In general, the nonadiabatic
factor can be written as
n
X
v ¼ v0
ð1 þ cin2Þ;
ð10Þ
i
i¼1
where v0 is the nonadiabatic factor in the absence
of significant spin–orbit coupling, and the coeffi-
cient ci measures the contribution of the heavy
atom in the position i to the total intersystem-
crossing rate. Fig. 4 also includes the internal
heavy-atom effect in S1 ! T1 and T1 ! S0 transi-
tions of halogenated naphthalenes, following the
analysis of Birks of the data of Ermolaev and
Svitashev [54]. The experimental data were fitted
with c1 ¼ 10À4
.
We are now in a position to calculate kisc for
halogenated and metallated porphyrins. Fig. 5
compares calculated and experimental rates. This

188
E.G. Azenha et al. / Chemical Physics 280 (2002) 177–190
Table 2
Empirical parameters fitted in the calculation of intersystem crossing ratesa
S1 ! T1
T1 ! S0
v0
c1
c2
v0
c1
c2
Halogenated naphthalenes
Octaalkylporphyrins halogenated in
alkyl positions
2 Â 10À3
10À4
5 Â 10À7
10À6
5 Â 10À3
10À4
10À5
10À2
Octaalkylporphyrins halogenated in
alkyl pyrrole positions
Octaalkylporphyrins halogenated in
alkyl meso positions
10À2
10À2
2 Â 10À5
5 Â 10À3
5 Â 10À3
5 Â 10À4
10À4
10À3
Metal-TPP
Metal-TPPc
5 Â 10À3
2 Â 10À5
2 Â 10À5
10À2
10À2
2 Â 10À4
2 Â 10À4
10À4b
TPP with halogens in the ortho
positionsc
2 Â 10À6
2 Â 10À6
ZnTPP with halogens in the ortho
positionsc
10À2
10À4b
2 Â 10À5
5 Â 10À7
a Room temperature, except where noted; c2 refers to the effect of halogens in the ortho positions.
b T ¼ 77 K.
c This work.
gH. The good correlation obtained in Fig. 5 results
from a correct description of both electronic and
Franck–Condon factors.
Workstation using the Cerius2 (from Molecular
Simulations Incorporated-MSI) software. Geom-
etry optimisation was performed with the Austin
Model 1 with configuration interaction (AM1-CI)
[55], as included in the MOPAC6.0 version. Hy-
perchem software (release 6.01 for Windows) was
further used to obtain the frontier molecular
orbitals by means of the Zerner Intermediate
Neglect of Differential Overlap/1 with Configu-
ration interaction (ZINDO/1-CI) technique. Fig. 6
shows the contours of the HOMO and LUMO
calculated with the ZINDO method [56]. The
highest occupied molecular orbital (HOMO) and
the next-to-the highest occupied molecular orbital
(HOMO-1) are mainly localised on the porphyrin
moiety. The same applies to the lowest unoccu-
pied molecular orbital (LUMO) and to next-
to-the lowest occupied molecular orbital
(LUMO + 1). It is known from crystallographic
data that the angle of the phenyl group with the
porphyrin ring for the ZnTPP is about 60° [57].
The introduction of bulky atoms in the ortho
positions leads to a larger degree of rotation of
the phenyl groups with respect to the plane of the
porphyrin skeleton, facilitating the overlap of
halogen pz-orbital with orbitals of the porphyrin
moiety.
It is very interesting to remark that c1 is smaller
for S1 ! T1 than for T1 ! So transitions when the
heavy atoms are introduced in the alkyl, pyrrole
and meso positions of free-base octaalkylporphy-
rins. This means that such heavy atoms increase
more the triplet-state decay rate than its rate of
formation. Ultimately, these substitution patterns
will give rise to high triplet-state quantum yields,
but the triplets will have short lifetimes. On the
other hand, heavy metals or halogens in the ortho
positions of the phenyl rings tend to increase more
the triplet-state formation rate than its decay. This
favours high yields of triplet without compromis-
ing the triplet lifetime. The halogens in the ortho
positions are particularly remarkable, as they are
50 times more efficient in accelerating the forma-
tion of the triplet than its decay.
The efficiency of the halogen substituents at
the ortho positions should be related to a fa-
vourable overlap between the orbitals of the
halogen atoms and those of the porphyrin core.
We investigated the electronic structure of these
species performing quantum mechanical-molecu-
lar orbital calculations on a Silicon Graphics. O2

E.G. Azenha et al. / Chemical Physics 280 (2002) 177–190
189
5. Conclusion
Tetrakisphenylporphyrins with halogens in the
ortho positions are particularly efficient in pro-
moting S1 ! T1 intersystem crossing: kisc increases
by two orders of magnitude from H2TPP to
H2TPPCl8. On the other hand, the comparison
between the low-temperature phosphorescence
lifetimes of ZnTPP and ZnTPPCl8 reveals that the
introduction of 8 Cl atoms decreases the lifetime of
the triplet state by a very modest 37%. The Tun-
nel-Effect Theory, modified to account for spin–
orbit coupling, is a convenient toll to rationalise
and predict heavy atom effects in radiationless
transitions. The experimental data on halogenated
porphyrins and metalloporphyrins and their in-
terpretation with TET indicate that tetrakisphe-
nylporphyrins with halogens in the ortho positions
are very appropriate to generate long-lived triplet
states in high yields. Such compounds may find
application as singlet oxygen sensitisers, where
such properties and a strong absorption in the
visible and near-infrared are desirable.
Acknowledgements
We thank SAPIENS (European Union) and
FCT for financial support (Project No. POCTI/
QUI/42536/2001). M.P. received a scholarship
from Praxis XXI/BD/11398/97. We are grateful to
Prof. Gaspar Martinho (Instituto Superior
ꢀ
Tecnico, Lisboa) for the measurement of picosec-
ond lifetimes.
References
[1] S.I. Yang, J. Seth, J.-P. Strachan, S. Gentemann, D. Kim,
D. Holten, J.S. Lindsey, D.F. Bocian, J. Porphyrins
Phthalocyanines 3 (1999) 117.
[2] R. Bonnett, P. Charlesworth, B.D. Djelal, D.J. McGarvey,
T.G. Truscott, J. Chem. Soc., Perkin Trans. 2 (1999) 325.
[3] J.D. Spence, T.D. Lash, J. Org. Chem. 65 (2000) 1530.
[4] L.G. Arnaut, S.J. Formosinho, J. Photochem. Photobiol.
A 111 (1997) 111–139.
Fig. 6. LUMOs (top) and HOMOs (bottom) contours calcu-
lated with the ZINDO/1-CI method. The electron density
contours show that the porphyrin core carries nearly all the
electron density of these orbitals.
[5] A.J. Hudson, T. Richardson, J.P. Trirtle, R.A.W. John-
stone, A.J.F.N. Sobral, A.M.d’A. Rocha Gonsalves, Mol.
Cryst. 253 (1993) 103.

190
E.G. Azenha et al. / Chemical Physics 280 (2002) 177–190
[30] A. Harriman, J. Chem. Soc., Faraday Trans. 2 (77) (1981)
[6] M.B. Grieve, T. Richardson, R.A.W. Johnstone, A.F.N.L.
Sobral, A.M.d’A. Rocha Gonsalves, Thin Solid Films 243
(1994) 581.
1281.
[31] R. Bonnett, A. Harriman, A.N. Kozyrev, J. Chem. Soc.,
Faraday Trans. 88 (1992) 763.
[7] A.M.d’A.R. Gonsalves, M.M. Pereira, J. Mol. Cat. A 113
(1996) 209.
[32] J. Seixas de Melo, P.F. Fernandes, J. Mol. Struct. 565–566
(2001) 69.
[8] T.J. Dougherty, Adv. Photochem. 17 (1992) 275.
[9] T.J. Dougherty, Photochem. Photobiol. 58 (1993) 895.
[10] R. Bonnett, Chem. Soc. Rev. 24 (1995) 19.
[11] S. Rywkin, E. Ben-Hur, Z. Malik, A.m. Prince, Y.-S. Li,
M.E. Kenney, N.L. Oleinik, B. Horowitz, Photochem.
Photobiol. 60 (1994) 165.
[33] P.G. Seybold, M. Gouterman, J. Mol. Spectrosc. 31 (1969)
1.
[34] S.L. Murov, I. Carmichael, G.L. Hug, Handbook of
Photochemistry, Marcel Dekker, New York, 1993.
[35] R. Englman, J. Jortner, Mol. Phys. 18 (1970) 145.
[36] O.S. Mortensen, W. Siebrand, A.W. Tarr, Chem. Phys. 125
(1988) 231.
[12] E. Ben-Hur, B. Horowitz, Photochem. Photobiol. 62
(1995) 383–388.
[13] M. Pineiro, A.L. Carvalho, M.M. Pereira, A.M.d.A.R.
Gonsalves, L.G. Arnaut, S.J. Formosinho, Chem. Eur. J. 4
(1998) 2299.
[37] D.L. Dexter, J. Chem. Phys. 21 (1953) 836.
[38] N. Nijegorodov, R. Mabbs, Spectrochim. Acta, Part A 57
(2001) 1449.
[14] M. Pineiro, M.M. Pereira, A.M.d’A. Rocha Gonsalves,
L.G. Arnaut, S.J. Formosinho, J. Photochem. Photobiol.
A 138 (2001) 147.
[39] N.V. Korol’kova, V.G. Klimenko, T.A. Kir’yanova, E.A.
Gastilovich, Chem. Phys. 248 (1999) 233.
[40] S.J. Formosinho, Mol. Photochem. 7 (1976) 41.
[41] J. Jortner, J. Ulstrup, Chem. Phys. Lett. 63 (1979) 236.
[42] W. Siebrand, in: A.B. Zahlan, et al. (Eds.), The Triplet
State, Cambridge University Press, Cambridge, 1967, p. 31.
[43] E.F. McCoy, I.G. Ross, Aust. J. Chem. 15 (1962) 573.
[44] A. Langseth, P.B. Stoicheff, Can. J. Chem. 31 (1956) 350.
[45] J.H. Callomon, T.M. Dunn, I.M. Mills, Philos. Trans. R.
Soc. London A 259 (1960) 499.
[15] M. Pineiro, A.M.d’A. Rocha Gonsalves, M.M. Pereira,
S.J. Formosinho, L.G. Arnaut, J. Phys. Chem. A 106
(2002) 3787.
[16] S.P. McGlynn, T. Azumi, M. Kinoshita, Molecular Spec-
troscopy of the Triplet State, Prentice-Hall, Englewood
Cliffs, NJ, 1969.
[17] A. Harriman, J. Chem. Soc., Faraday Trans. 1 (76) (1980)
1978.
[46] S. Shaik, A. Shurki, D. Danovich, P.C. Hiberty, J. Am.
Chem. Soc. 118 (1996) 666.
[18] A. Harriman, J. Chem. Soc., Faraday Trans. I 77 (1981)
369.
[47] A.A.C.C. Pais, L.G. Arnaut, S.J. Formosinho, J. Chem.
Soc., Perkin Trans. 2 (1998) 2577.
[19] J.R. Darwent, P. Douglas, A. Harriman, G. Porter, M.-C.
Richoux, Coord. Chem. Rev. 44 (1982) 83.
[20] S.J. Formosinho, J. Chem. Soc., Faraday Trans. 2 (70)
(1974) 605.
[48] A.J. Gordon, R.A. Ford, The Chemist’s Companion,
Wiley, New York, 1972, p. 107, 114.
[49] S.J. Formosinho, L.G. Arnaut, R. Fausto, Prog. Reaction
Kinet. 23 (1997) 1.
[21] G. Stricker, V. Subramaniam, C.A.M. Seidel, A. Volkmer,
J. Phys. Chem. B 103 (1999) 8612.
[50] S.J. Formosinho, Rev. Port. Quim. 14 (1972) 201.
[51] L.G. Arnaut, L.G. Formosinho, J. Photochem. Photobiol.
A 100 (1996) 15.
[22] L.G. Arnaut, R.A. Caldwell, J.E. Elbert, L.A. Melton,
Rev. Sci. Instrum. 63 (1992) 5381.
[23] C.S. Soares, L.G. Arnaut, J. Phys. Chem. 104 (2000)
11075.
[52] J.B. Birks, Photophysics of Aromatic Molecules, Wiley,
London, 1970, p. 185 and p. 187.
[24] M. Lock, Monatsh. Chem. 55 (1930) 307.
[25] J.S. Lindsey, I.C. Schreiman, H.C. Hsu, P.C. Kearny,
A.M. Marguerettaz, J. Org. Chem. 52 (1987) 827.
[53] M. Kasha, Discuss. Faraday Soc. 9 (1950) 14.
[54] V.L. Ermolaev, K.K. Svitashev, Opt. Spectrosc. 7 (1959)
399.
~
[26] A.M.d’A. Rocha Gonsalves, J.M.T.B. Varejao, M.M.
[55] M.J.S. Dewar, R.C. Bingham, D.H. Lo, J. Am. Chem. Soc.
97 (1975) 1285.
Pereira, J. Heterocyclic Chem. 28 (1991) 635.
[27] J.W. Buchler, in: D. Dolphin (Ed.), The Porphyrins, vol. 1,
Academic Press, New York, 1978, p. 328.
[56] M.C. Zerner, in: D.B. Boyd (Ed.), Reviews in Computa-
tional Chemistry, VCH Publishers, New York, 1991.
[57] M.D. Glick, G.H. Cohen, J.L. Hoard, J. Am. Chem. Soc.
89 (1967) 1996.
[28] A.D. Adler, F.R. Longo, F. Kampas, Inorg. Chem. 32
(1979) 2443.
[29] H.H. Hodgson, H.G. Beard, J. Chem. Soc. 127 (1925) 875.