Studies of the triplet state of the proton-transfer tautomer in salicylaldehydes

Article abstract of DOI:10.1016/S0009-2614(03)00165-9

The spectroscopy and dynamics of the low-lying triplet state of the proton-transfer tautomer in salicylaldehydes have been studied via internal heavy-atom effects coupled with a sensitive near-IR detecting system. For 3,5-diiodosalicylaldehyde a weak proton-transfer keto-tautomer phosphorescence was resolved with a maximum at 710 nm (τ_p ～ 1.8 μs, φ_obs ～ 5.23 × 10^-4) in a 77 K methylcyclohexane glass. The results, in combination with the time-resolved thermal lensing experiment, further deduced the triplet-state population yield and radiative decay rate to be 0.20 (298 K) and 3.12 × 10² s^-1, respectively. Consequently, the energetics and dynamics of the triplet states during a proton transfer cycle are discussed in detail.

Full text of DOI:10.1016/S0009-2614(03)00165-9

Chemical Physics Letters 370 (2003) 747–755
www.elsevier.com/locate/cplett
Studies of the triplet state of the proton-transfer tautomer
in salicylaldehydes
a,
b
b
b
Pi-Tai Chou *, Chau-Shuen Chiou , Wei-Shan Yu , Guo-Ray Wu ,
c
Tai-Huei Wei
a
Department of Chemistry, National Taiwan University, Taipei 621, Taiwan, ROC
b
Department of Chemistry, The National Chung-Cheng University, Chia Yi, Taiwan, ROC
Department of Physics, The National Chung-Cheng University, Chia Yi, Taiwan, ROC
c
Received 1August 2002
Abstract
The spectroscopy and dynamics of the low-lying triplet state of the proton-transfer tautomer in salicylaldehydes
have been studied via internal heavy-atom effects coupled with a sensitive near-IR detecting system. For 3,5-diiodo-
salicylaldehyde a weak proton-transfer keto-tautomer phosphorescence was resolved with a maximum at 710 nm
(sp ꢀ 1:8 ls, Uobs ꢀ 5:23 ꢁ 10ꢂ4) in a 77 K methylcyclohexane glass. The results, in combination with the time-resolved
thermal lensing experiment, further deduced the triplet-state population yield and radiative decay rate to be 0.20 (298
K) and 3:12 ꢁ 102 sꢂ1, respectively. Consequently, the energetics and dynamics of the triplet states during a proton
transfer cycle are discussed in detail.
Ó 2003 Elsevier Science B.V. All rights reserved.
1. Introduction
transfer tautomer possesses vast differences in
structure and electronic configuration from its
corresponding normal species. Accordingly, a
large Stokes shifted S01 ! S00 fluorescence (here-
after, the prime sign denotes the proton-transfer
tautomer) was observed. This unusual photo-
physical property has found many important ap-
plications. Prototypical examples are probes for
solvation dynamics [6] and biological environ-
ments [7], the development of laser dyes [8,9], ul-
traviolet stabilizers [10], metal ion sensors [11] and
radiation hard-scintillator counters [12], etc.
The fundamental approach to a proton transfer
process, which is crucial to many chemical and
biological reactions, has relied deeply on studies of
the excited state intramolecular proton transfer
(ESIPT) reaction [1–5]. The ESIPT reaction gen-
erally incorporates transfer of a hydroxyl (or
amino) proton to the carbonyl oxygen (or pyridi-
nic nitrogen) through a preexisting hydrogen
bonding (HB) configuration. The resulting proton-
In contrast to extensive efforts made on the
singlet manifold, studies on the spectroscopy and
dynamics of triplet states in ESIPT molecules are
* Corresponding author. Fax: +886-2-23695208.
E-mail address: chop@ccms.ntu.edu.tw (P.-T. Chou).
0009-2614/03/$ - see front matter Ó 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0009-2614(03)00165-9

748
P.-T. Chou et al. / Chemical Physics Letters 370 (2003) 747–755
relatively rare. Disregarding the first case of the
ESIPT reaction on methyl salicylate reported in as
early as 1956 [13], only recently has the investiga-
tion of T01 ! S00 phosphorescence become possible.
Grellmann and co-workers [14,15], first observed
the T01 ! S00 phosphorescence in 2-(20-hydroxy-
phenyl)benzoxazole. Subsequently, Chou et al. [16]
reported the weak 648 nm phosphorescence of 2-
(20-hydroxyphenyl)-benzothiazole (HBT) in a 77 K
hydrocarbon glass. Grabowska et al. [17] then
examined another case of proton-transfer phos-
phorescence in 2,20-bipyridyldiol systems. Re-
comes smaller [19]. Conversely, the rate of the
radiationless deactivation pathway is expected to
increase [20], in particular for ESIPT molecules
where the vibrational modes associated with in-
tramolecular hydrogen bonds usually act as good
quenchers for the emission [21]. One would thus
expect that in the condensed phase, the lowest-ly-
ing tautomer triplet state in NIR might give rise to
an extremely low yield of phosphorescence.
Very recently, we have reported the spectroscopy
and dynamics of the proton-transfer tautomer 735-
nm phosphorescence for 7,9-diiodo-10-hydroxy-
benzo[h]quinoline. The results, in combination with
the triplet-state sensitizing singlet oxygen experi-
ment, led us to deduce relative energetics in different
spin manifolds during a proton-transfer reaction
[22]. To further extend the approach on the spec-
troscopy and dynamics of triplet states in other
ESIPT systems, we have investigated salicylalde-
hyde (SA) and its halogenated derivative 3,5-di-
iodosalicylaldehyde (DISA) in the triplet manifold.
SA is structurally the simplest aromatic ESIPT
system, of which the physical properties in the sin-
glet-state manifold are readily accessible to both
spectroscopic and theoretical analyses. In the elec-
tronic ground-state SA exists dominantly in an enol
form (conformer E, see Fig. 1) possessing an intra-
molecular hydrogen bond between hydroxyl and
carbonyl groups [23–25]. Our approach to achieve
ꢀ
cently, Catalan and Diaz [18] applied a spin–orbit
perturbation technique, i.e., the intermolecular
and/or intramolecular Ôheavy-atomsÕ effects, to
achieve a feasible method of observing the desired
T01 ! S00 phosphorescence for halogenated meth-
ylsalicylates.
The T01–S00 energy gap in the above few cases is
usually restricted to the visible region (<700 nm).
For those ESIPT molecules in which the T01–S00
energy gap is expected to be in the far visible re-
gion extending to the near-infrared (NIR) 0region
(<15 000 cmꢂ1), the spectrally resolved T1 ! S00
phosphorescence was obscure. One major obstacle
for detecting the NIR phosphorescence is the
dominant T01 ! S00 radiationless deactivation pro-
cess. When the energy gap decreases, the forbidden
triplet ! singlet radiative decay rate constant be-
Fig. 1. Structures of SA and DISA and their corresponding methyl derivatives.

P.-T. Chou et al. / Chemical Physics Letters 370 (2003) 747–755
749
the feasibility of detecting the ultraweak phospho-
rescence is based on two key factors; namely the
intramolecular heavy-atom effect coupled with an
ultrasensitive NIR detection system. In this study,
we have further applied a time-resolved thermal
lensingtechnique to directly monitor the population
yield of the triplet state. The results lead to the es-
tablishment of energetics for DISA in different spin
manifolds during a proton-transfer cycle.
The time-resolved thermal lensing (TRTL)
technique [27] was applied to measure the radia-
tionless decay dynamics. Briefly, the TRTL setup
consists of an Nd:YAG laser (355 nm) of which
the output energy can be finely adjusted through a
dual-k=2-plates configuration. A He–Ne laser
(JDS Uniphase 1125; 632.8 nm, 5 mW) equipped
with a stabilized power supply was used as an
analyzing beam for TRTL signals. The He–Ne
beam was focused in front of the sample cell with
an f ¼ 30 mm lens, and was in a direction of 180°
with respect to the Nd:YAG excitation beam
(focused by a f ¼ 200 mm lens). The He–Ne laser
beam after the sample cell was passed through a
pinhole (Corion 2401; 300 lm diameter) and a set
of filters combination. The beam intensity was
then detected with a photomultiplier tube (Ham-
amatsu R943), of which the output signals were
converted into the voltage (Tektronix TDS 3054;
5GS/Sampling). The system response is limited by
the acoustic transient time defined by wp=vs, where
wp is the radius of the waist of the pump beam at
the focal point and was measured to be ꢀ 60 lm
in our current setup, vs denotes the speed of sound
in the solvent, which is 1:28 ꢁ 103 m/s in MCH.
This gives a system response time of ꢀ50 ns in
MCH.
2. Experimental
2.1. Material
€
Salicylaldehyde (SA, Riedel-de Haen, 99%) was
purified by vacuum distillation in the grease-
free condition. 3,5-Diiodosalicylaldehyde (DISA,
Lancaster, 98%) was purified by column chroma-
tography (eluent CH2Cl2) until a single spot
showed on the TLC plate. The purity was checked
by 1H NMR and fluorescence excitation spectrum.
2-Methoxybenzaldehyde (MBA) was synthesized
by the methylation of SA in basified methylio-
dide solution. 3,5-Diiodo-2-methoxybenzaldehyde
(DIMBA) was then synthesized by adding a solu-
tion containing MBA (0.01mol) and glacial acetic
acid (4 ml) in chloroform (10 ml) to a solution of
iodine (0.02 mol in 20 ml methanol) for a period of
1h. The precipitate was collected and neutralized
with sodium carbonate to give DIMBA. Methyl-
cyclohexane (MCH, Aldrich) was of spectrograde
quality and used without further purification.
3. Results and discussion
3.1. Absorption and fluorescence properties
The room-temperature absorption and emission
spectra of DISA in MCH are shown in Fig. 2. The
S0 ! S1 ðppÃÞ absorption exhibits an onset at ꢀ400
2.2. Method
Detailed steady state, time-resolved and quan-
tum yield measurements of fluorescence and phos-
phorescence have been elaborated in previous
reports [22,26]. The nanosecond transient absorp-
tion was recorded by a laser flash photolysis system
(Edinburgh LP920), in which an Nd:YAG laser
(355 nm) pumped optical parametric oscillator and
a white-light square pulse were used as pump and
probe beams, respectively. These two pulses were
crossed at a 90° angle with an overlapping distance
of ꢀ10 mm. The temporal resolution was limited by
the excitation pulse duration of ꢀ8 ns.
nm with a band maximum at 360 nm (e360
ꢀ
3600 Mꢂ1 cmꢂ1). The fluorescence is characterized
by a single emission band maximum at 560 nm
(Uf ꢀ 2:3 ꢁ 10ꢂ3) with a lifetime measured to be
ꢀ52 ps. The normal Stokes shifted fluorescence
expected to be in the region of 350–450 nm was not
detected. The excitation spectrum of the 560-nm
band is independent of the monitored emission
wavelength and is effectively identical with the ab-
sorption profile. In comparison, DIMBA, which is
considered a non-proton-transfer model with a
similar electronic configuration to DISA, exhibited

750
P.-T. Chou et al. / Chemical Physics Letters 370 (2003) 747–755
Fig. 3. Temporal evolution of the transient absorption spectra
of DISA in degassed MCH (298 K) with a delay time of (a) 130
ns, (b) 350 ns, (c) 500 ns, (d) 750 ns and (e) 1.2 ls. Inset: the plot
of 420-nm absorption intensity versus various delay times in
degassed (ꢃ) and aerated (N) solution.
Fig. 2. Absorption and emission spectra of DISA in MCH at
298 (–) and 77 K (- - -). (a)–(f): The time dependent spectral
evolution of the phosphorescence for DISA (5:7 ꢁ 10ꢂ5 M) in a
77 K MCH glass at >600 nm with a delay time of (a) 200 ns, (b)
800 ns, (c) 1.6 ls, (d) 2.4 ls, (e) 4.7 ls and (f) 6.2 ls. Inset: the
appearance of the E0 phosphorescence after 355-nm (5 mJ/
pulse) irradiation for 5 min in a 77 K MCH glass.
1:5 ꢁ 106 sꢂ1 (s ¼ 0:67 ls), of which the relaxation
dynamics were significantly quenched by O2. As
shown in the inset of Fig. 3, the decay rate was
increased to ꢀ 7:0 ꢁ 106 sꢂ1 in the aerated MCH
(O2 ꢀ 2:3 ꢁ 10ꢂ3 M). On the basis of the Stern–
Volmer plot (not shown here), an O2 quenching
rate constant of 2:4 ꢁ 109 sꢂ1 was extracted, which
within experimental error is eqꢂu1al to 1/9 diffusion-
controlled rate (ꢀ 3:0 ꢁ 109 M sꢂ1 in MCH [28])
theoretically derived from the triplet (T)-O2 sen-
sitization. Thus, the assignment of the observed
transient absorbance in DISA to the triplet–triplet
transition is unambiguous. In a comparative
study, the transient absorption of DIMBA in de-
gassed MCH was also performed, of which both
a very weak, normal Stokes shifted fluorescence
maximized at ꢀ400 nm (see Table 1). Accordingly,
the ꢀ560 nm emission band with a Stokes shift of as
large as 9920 cmꢂ1 (peak-to-peak) measured in
DISA is unambiguously ascribed to the keto-tau-
tomer (K, see Fig. 1) fluorescence resulting from
ESIPT.
3.2. Photophysics of the triplet state
Fig. 3 depicts the room-temperature transient
absorption spectra of DISA in the degassed MCH.
The transient absorption maximized at 420 nm
undergoes a fast single exponential decay rate of
Table 1
Photophysical properties of SA, DISA and DIMBA in MCH at 298 and 77 K
Phosphorescencea
kmax (nm)
435b
Absorption
Fluorescence
kmax (nm)
kmax (nm)
Uobs
sf (ps)
Uobs
sp (s)
SA
330
510
505a
1:0 ꢁ 10ꢂ3
1:3 ꢁ 10ꢂ2;a
2:3 ꢁ 10ꢂ3
9:1 ꢁ 10ꢂ3;a
–
50
467a
52
–
1:3 ꢁ 10ꢂ3
DISA
360
352
560
550a
420b;c
ꢀ400c
710
455b
5:2 ꢁ 10ꢂ4
1:8 ꢁ 10ꢂ6
240a
–
3:6 ꢁ 10ꢂ4
–
DIMBA
a 77 K.
–
–
445
2:2 ꢁ 10ꢂ2
2:8 ꢁ 10ꢂ4
b The photolysis product at 77 K.
c Emission is too weak to perform dynamic and quantum yield measurements.

P.-T. Chou et al. / Chemical Physics Letters 370 (2003) 747–755
751
spectral feature (kmax ꢀ 470 nm) and dynamics of
decay (t1=2 ꢀ 20 ls) are considerably different from
those of DISA. Since DIMBA represents a non-
proton-transfer model, the assignment of triplet–
triplet transient absorption originating from the
enol conformer is obvious. Accordingly, the 420-
nm transient absorption in DISA is tentatively
ascribed to the triplet–triplet absorption of the
keto-tautomer (i.e., conformer K) species.
keto-tautomer phosphorescence of SA in a 77 K
MCH glass failed. In contrast, a long-lived emis-
sion maximum at 710 nm was apparently resolved
for DISA (Fig. 2). The relaxation dynamics can be
well fitted by single exponential decay kinetics with
a lifetime of 1.8 ls. Similar excitation spectra (not
shown here) between monitoring at tautomer flu-
orescence (i.e., the 550-nm emission) and the 710-
nm phosphorescence concluded that both emission
bands originate from a common ground-state
species, i.e., conformer E. For comparison, DIM-
BA in a 77 K MCH glass exhibits an enol phos-
phorescence maximum at ꢀ445 nm (sp ꢀ 0:28 ms,
see Table 1) of which both spectral feature and
relaxation dynamics are quite different from those
of DISA. Thus, the 710-nm emission in DISA can
be unambiguously ascribed to the T01 ! S00 phos-
phorescence of the keto-tautomer.
Caution has to be taken in performing phos-
phorescence measurements in the low-temperature
solid matrices. Nagaoka et al. [23] have reported
that except for the major channels of radiationless
deactivation (i.e., proton transfer reaction, internal
conversion, etc.), SA undergoes a minor photo-
isomerization channel, forming an open-form
conformer trapped in the cold, solid matrices.
Morgan et al. [25] further resolved the open con-
former by FTIR analyses of vibrational modes for
various deuterium isotope substitutions, and con-
cluded that in the open-form conformer both hy-
droxyl and aldehyde functional groups rotate by
180°. For DISA, the formation of a similar open-
form conformer (E0, see Fig. 1) was observed in the
77 K MCH glass matrices (vide infra). Such an
interference may hamper our goal of detecting the
T01 ! S00 phosphorescence. Fortunately, the open-
form product can be experimentally avoided by
taking a few laser shots during the data acquisition
accompanied by a low power of excitation
(ꢄ0:5 mJ=cm2), while spectra could still be ob-
tained with a good signal-to-noise ratio due to the
highly sensitive detecting system.
Under low-excitation energy, DISA exhibit
unique keto-tautomer fluorescence maximized at
550 nm (sf ꢀ 240 ps) in 77 K MCH glassy matrices
(Fig. 2). In order to obtain the presumably ultra-
weak phosphorescence attributed to the T01 ! S00
transition of the proton-transfer tautomer, the
gate of intensified charge coupled detector (ICCD
[22]) was opened at a delay time of ꢀ50 ns to
eliminate the keto-tautomer fluorescence. On the
other hand, due to its relatively much longer life-
time of ꢀ0.36 ms, the interfering phosphorescence
resulting from the open conformer (E0, vide infra),
if there is any, can be eliminated by gating ICCD
with a window as small as 200 ns. Under these
experimental conditions an attempt to detect the
Two close lying npà and ppà states (<3 kcal/
mol) have been reported for SA [25,29], in which
the photoinduced E ! E0 rotamerization origi-
nates from the lowest singlet excited state pos-
sessing a 1npà configuration [29]. For DISA,
induced by the diiodo-substitution, the first
1p ! pà absorption maximum of 360 nm is
ꢀ2500 cmꢂ1 red shifted relative to that of SA
(kmax ꢀ 330 nm, see Table 1). It is thus reasonable
1
1
1
to predict a reverse energetics between npà and
1ppà states in DISA. Accordingly, the rotamer-
ization efficiency should be significantly reduced
1
due to the competing fast 1npà ! ppà internal
conversion. Support of this viewpoint is given by
the lack of resolving any oxygen-independent
transient absorption associated with the E0 con-
former. Nevertheless, a non-negligible photoin-
duced rotamerization process still takes place, of
which the product (e.g., the E0 form) can be trap-
ped and accumulated for further analyses in the
77 K MCH glass matrices. Inset of Fig. 2 shows
the 455-nm phosphorescence spectrum of the
photolysis product of DISA under high laser in-
tensity (>5 mJ=cm2). Both spectral and dynamic
(s ꢀ 0:36 ms) features of the 455-nm phospho-
rescence are similar to that of the 445-nm phos-
phorescence observed in DIMBA. This, in
combination with the reversibility of the reaction
during a thawing/freezing cycle, led us to conclude
the 455-nm phosphorescence originate from a non-

752
P.-T. Chou et al. / Chemical Physics Letters 370 (2003) 747–755
hydrogen-bonded enol form, possibly the E0 con-
former. Because this issue is not the main focus of
this study, further quantitative elaboration has not
been pursued.
3.3. TRTL experiments
Previously, we made an attempt to quantify the
triplet-state yield through a triplet-state sensitizing
1O2 experiment [22]. This method is valid only if
1
the efficiency of O2ð DgÞ sensitized by the triplet
state is known. Unfortunately, this value varies
with sensitizers and is rather complex to deduce.
For simplicity of derivation the O2 sensitization
Fig. 4. The TRTL profile of DISA (1:1 ꢁ 10ꢂ4 M) in degassed
MCH. Inset: the plot of total signal intensity versus 355-nm
laser energies at the absorbance (355 nm) of 0.2 (ꢃ) and 0.5 (N).
1
efficiency is generally assumed to be unity. Ac-
cordingly, this indirect method may be subject to a
great deal of uncertainty. Alternatively, a TRTL
technique has been developed in this study to di-
rectly quantify the yield of triplet state for DISA.
For the case of DISA, the rates for both proton
transfer and keto-tautomer fluorescence are fast,
as indicated by the system response limited rise
(<15 ps) and ꢀ52 ps lifetime of the keto-tautomer
emission. Furthermore, supported by the identical
temporal evolution of transient absorption spectra
(vide supra), it is reasonable to conclude that ex-
cept for the triplet-state species, there are no other
resolvable, long-lived transient species existing
during a proton transfer cycle. Consequently, the
rate of K ! E ground-state reverse proton trans-
fer for DISA is concluded to be faster than the
system response of ꢀ8 ns. Since TRTL is limited
by the acoustic transient time which is ꢀ50 ns in
MCH in this study (see Section 2), the only slow,
resolvable transient signal should originate from
the triplet state. Other relaxation pathways during
the proton transfer cycle should reveal instant,
acoustic-transient-time limited rise kinetics.
Fig. 4 reveals the TRTL time profile of DISA in
MCH. The signal apparently consists of a system
irresolvable and a 0.69 ls rise component. The
former is ascribed to the sum of all fast decay
components, while the latter with its rise time
similar to the lifetime measured from the transient
absorption can thus be assigned to the decay time
of the keto-tautomer triplet state. As shown in the
inset of Fig. 4, the plot of the intensity of the total
signal, Ut, versus the applied laser energy (5–30 lJ)
is sufficiently linear, eliminating the possibility of
the multiphoton effect contributed to the TRTL
signal. Since the intensity of the TRTL signal is
proportional to the absolute energy dissipated, the
intensity (voltage) of the slow rising signal, Us
versus Ut, can be expressed as
Us UptUKiscEp
¼
;
ð1Þ
Ut
aEex
where Upt and UKisc denote the yields of ESIPT and
S0 ! T0 intersystem crossing, respectively, Eex is
1
the excitation energy, and Ep is the energy gap
of the keto phosphorescence. a denotes the heat
conversion efficiency and is equivalent to
ðEex ꢂ Uf EsÞ=Eex, where Uf and Es are the quantum
yield and 0–0 transition energy of the keto fluo-
rescence. Because the yield of fluorescence Uf is
as small as 2:3 ꢁ 10ꢂ3 (298 K) in DISA, a is
essentially equal to 1. Due to the lack of enol flu-
orescence and ultrafast rise dynamics of the keto-
tautomer emission, it is also reasonable to assume
Upt to be ꢀ1. Ep was deduced from its average
energy expressed as
R
IpðvÞvdv
R
hEpi ¼
;
ð2Þ
IpðvÞdv
where IpðvÞ denotes the relative emission intensity
as a function of wavenumber. Accordingly, a value
of 13 985 cmꢂ1 was obtained. The ratio of Us=Ut
(see Fig. 4) was calculated to be ꢀ0.10. Applying
Eex of 28 169 cmꢂ1 (355 nm) in this study, UKisc was
thus deduced to be 0.2. UKisc is equivalent to kisc=kf ,

P.-T. Chou et al. / Chemical Physics Letters 370 (2003) 747–755
753
Scheme 1. Relative energy levels (in cmꢂ1) and photophysical parameters for the enol and keto-tautomers of DISA estimated from
either experimental or theoretical approaches. See text for the detailed description.
where kisc and kf are rate constants of intersystem
crossing and the fluorescence decay, respectively,
for the keto-tautomer. Knowing kf to be 1:9 ꢁ
1010 sꢂ1 ðsf ꢀ 52 psÞ, kisc was thus deduced to be
3:8 ꢁ 109 sꢂ1 at 298 K.
respectively denote the radiative and measured
decay rate constants of the keto phosphorescence.
To obtain UKisc at 77 K, kisc is further assumed to be
independent of temperature. This assumption
holds true if the triplet-state population proceeds
only through the S01 ! T10 intersystem crossing and
the S01–T01 energy gap is ꢅ kT. For DISA, support
for the latter assumption is given by the >4000
cmꢂ1 gap measured from the difference in peak
3.4. Radiative lifetime of the keto-phosphorescence
The yield of keto-tautomer phosphorescence,
UPobs, was measured to be 5:23 ꢁ 10ꢂ4 in the 77 K
MCH glass, which can be further expressed as
maxima between keto fluorescence (ꢀ18 182 cmꢂ1
)
and phosphorescence (ꢀ14 084 cmꢂ1). Knowing
the decay rate of the keto-tautomer fluorescence to
be 4:2 ꢁ 109 sꢂ1, a value of 0.93 was estimated for
UKisc in a 77 K MCH glass. As a result, UKp was
deduced to be 5:62 ꢁ 10ꢂ4. With a known koPbs
value of 5:56 ꢁ 105 sꢂ1, the radiative decay rate of
keto-tautomer phosphorescence, krP, was then de-
UPobs ¼ UptUiKscUpK;
ð3Þ
where Upt denotes the yield of the ESIPT and is
assumed to be ꢀ100% efficiency at 77 K due to the
irresolvable rise of the keto-tautomer fluorescence.
UKp represents the intrinsic keto phosphorescence
duced to be 3:12 ꢁ 102 sꢂ1
.
yield and is equivalent to krP=koPbs where krP and koPbs
,

754
P.-T. Chou et al. / Chemical Physics Letters 370 (2003) 747–755
3.5. Energetics during a proton transfer cycle
is of fundamental importance. For DISA the study
of T01 ! T1 tautomerism is not feasible due to the
highly endergonic process. Further approach
aimed at designing ESIPT molecules so that both
dynamics and thermodynamics of the T01 ! T1
tautomerism are accessible is in progress.
With the above spectroscopic data, an attempt
has been made to construct an energy diagram for
DISA in a proton transfer cycle. Applying Eq. (2)
the average S01 ! S00 and T10 ! S00 gaps were de-
duced to be 18 034 and 13 985 cmꢂ1, respectively.
Conversely, the absorption maximum of 360 nm
(27 778 cmꢂ1) was adopted for the S1 state. The T1
state of the enol conformer (E) is not available due
to its negligible population efficiency. Alterna-
tively, the average phosphorescence frequency
(21357 cm ꢂ1) of the E0 conformer was taken to be
the normal T1 state. Finally, the relative energy
between conformers E and K in the ground-state is
experimentally inaccessible, and has to be obtained
through the theoretical approach. The large spin–
orbit-coupling factor limits precise ab initio cal-
culations at higher-level basis sets. Alternatively, a
semi-empirical PM3 method was performed, esti-
mating the S00 state to be higher in energy than S0
by 13.67 kcal/mol. Accordingly, relative energy
levels for an overall proton transfer cycle in DISA
References
[1] E.M. Kosower, D. Huppert, Annu. Rev. Phys. Chem. 37
(1986) 127.
[2] Special Issue (P.F. Barbara, H.D. Trommsdorff (Eds.),
Spectroscopy and Dynamics of Elementary Proton Trans-
fer in Polyatomic systems), Chem. Phys. 136 (1989)
153.
[3] S.F. Formosinho, L.G. Arnaut, J. Photochem. Photobiol.
A Chem. 75 (1993) 21.
[4] A. Douhal, F. Lahmani, A.H. Zewail, Chem. Phys. 207
(1996) 477.
[5] S. Scheiner, J. Phys. Chem. A 104 (2000) 5898.
[6] F. Parsapour, D.F. Kelley, J. Phys. Chem. 100 (1996) 2791.
[7] A. Sytnik, M. Kasha, Proc. Natl. Acad. Sci. USA 91(1994)
8627.
[8] P.T. Chou, D. McMorrow, T.J. Aartsma, M. Kasha,
J. Phys. Chem. 88 (1984) 4596.
are depicted in Scheme 1. The S1 ! S0 tautomer-
1
ꢀ
[9] A.V. Acuna, F. Amat-Guerri, J. Catalan, A. Costella,
J. Figuera, J. Munoz, Chem. Phys. Lett. 132 (1986) 576.
ism is thermally allowed by ꢀ14.2 kcal/mol. The
T1 state of the enol form is estimated to be ꢀ7.4
kcal/mol higher in energy than the T01 state of the
keto-tautomer. This value has to be treated as an
upper limit since T1 of the HB conformer E is in
general lower in energy than that of the non-HB
conformer E0, resulting in a reduction of the T1–T01
gap.
ꢀ
[10] J. Catalan, J.C. del Valle, R.M. Claramunt, D. Sanz,
J. Dotor, J. Lumin. 68 (1996) 165.
[11] A.D. Roshal, A.V. Grigorovich, A.O. Doroshenko, V.G.
Pivovarenko, J. Phys. Chem. A 102 (1998) 5907.
[12] C.L. Renschler, L.A. Harrah, Nucl. Inst. Methods Phys.
Rev., A 235, Sept., U.S. Patent, 1985, 635.
[13] A.Z. Weller, Elektrochemistry 60 (1956) 1144.
[14] M.F.R. Prieto, B. Nickel, K.H. Grellmann, A. Mordzinski,
Chem. Phys. Lett. 146 (1988) 387.
In conclusion, we have applied internal heavy-
atom effects and an ultrasensitive near-IR detect-
ing system to study the energetics and dynamics of
the low-lying triplet states for ESIPT molecules.
TRTL was proven to be a direct, feasible method
to obtain the yield of the triplet state, and conse-
quently the radiative decay rate can be extracted.
The results lead to the establishment of relative
energy levels for an overall proton-transfer cycle of
DISA in different spin manifolds. While the oc-
currence of proton transfer has been widely ac-
cepted in the 1ppà state [1–5], it has been proposed
that hydrogen-atom rather than proton transfer
takes place in the triplet states [5,14,15,30]. Thus,
whether the triplet-state hydrogen-atom transfer is
operative and can be generalized in other systems
[15] K.H. Grellmann, A. Mordzinski, A. Heinrich, Chem. Phys.
136 (1989) 201.
[16] P.T. Chou, S.L. Studer, M.L. Martinez, Chem. Phys. Lett.
178 (1991) 393.
[17] A. Grabowska, P. Borowicz, D.O. Martire, S.E. Braslav-
sky, Chem. Phys. Lett. 185 (1991) 206.
ꢀ
[18] J. Catalan, C. Diaz, J. Phys. Chem. A 102 (1998) 323.
[19] S.P. McGlynn, T. Azumi, M. Kinoshita, in: Molecular
Spectroscopy of the Triplet State, Prentice Hall, Engle-
wood Cliffs, NJ, 1969, p. 18.
[20] W. Sieband, J. Chem. Phys. 47 (1967) 2411.
€
[21] W. Klopffer, Adv. Photochem. 10 (1977) 311.
[22] P.T. Chou, G.R. Wu, Y.I. Liu, W.S. Yu, C.S. Chiou,
J. Phys. Chem. A 106 (2002) 5967.
[23] S. Nagaoka, N. Hiorta, M. Sumitani, K. Yoshihara, J. Am.
Chem. Soc. 105 (1983) 4220.
[24] S. Nagaoka, U. Nagashima, N. Ohta, M. Fujita, T.
Takemura, J. Phys. Chem. 92 (1988) 166.

P.-T. Chou et al. / Chemical Physics Letters 370 (2003) 747–755
755
[25] M.A. Morgan, E. Orton, G.C. Pimentel, J. Phys. Chem. 94
(1990) 7927.
[28] This number is calculated from Stokes–Einstein equation,
see: J.B. Birks, in: Photophysics of Aromatic Molecules,
Wiley, New York, 1970, p. 313.
[26] P.T. Chou, Y.C. Chen, W.S. Yu, Y.H. Chou, C.Y. Wei,
Y.M. Chen, J. Phys. Chem. A 105 (2001) 1731.
[27] For example, see: S.E. Braslavsky, K. Heihoff, Handbook
of Organic Photochemistry, CRC Press, Boca Raton, FL,
1989.
ꢁ
[29] M. Cuma, S. Scheiner, T. Kar, J. Am. Chem. Soc. 120
(1998) 10497.
[30] For example, see: T. Suzuki, T. Omori, T. Ichimura,
J. Phys. Chem. A 101 (2000) 11671.