Ultrafast time-resolved study of photophysical processes involved in the photodeprotection of p-hydroxyphenacyl caged phototrigger compounds

Article abstract of DOI:10.1021/ja0458524

A combined femtosecond Kerr gated time-resolved fluorescence (fs-KTRF) and picosecond Kerr gated time-resolved resonance Raman (ps-KTR³) study is reported for two p-hydroxyphenacyl (pHP) caged phototriggers, HPDP and HPA, in neat acetonitrile a

Full text of DOI:10.1021/ja0458524

Published on Web 01/12/2005
Ultrafast Time-Resolved Study of Photophysical Processes
Involved in the Photodeprotection of p-Hydroxyphenacyl
Caged Phototrigger Compounds
Chensheng Ma, Wai Ming Kwok, Wing Sum Chan, Peng Zuo, Jovi Tze Wai Kan,
Patrick H. Toy, and David Lee Phillips*
Contribution from the Department of Chemistry, The UniVersity of Hong Kong, Pokfulam Road,
Hong Kong S.A.R., P. R. China
Received July 10, 2004; E-mail: phillips@hkucc.hku.hk
Abstract: A combined femtosecond Kerr gated time-resolved fluorescence (fs-KTRF) and picosecond Kerr
gated time-resolved resonance Raman (ps-KTR3) study is reported for two p-hydroxyphenacyl (pHP) caged
phototriggers, HPDP and HPA, in neat acetonitrile and water/acetonitrile (1:1 by volume) solvents. Fs-
KTRF spectroscopy was employed to characterize the spectral properties and dynamics of the singlet
excited states, and the ps-KTR3 was used to monitor the formation and subsequent reaction of triplet state.
These results provide important evidence for elucidation of the initial steps for the pHP deprotection
mechanism. An improved fs-KTRF setup was developed to extend its detectable spectral range down to
the 270 nm UV region while still covering the visible region up to 600 nm. This combined with the advantage
of KTRF in directly monitoring the temporal evolution of the overall fluorescence profile enables the first
time-resolved observation of dual fluorescence for pHP phototriggers upon 267 nm excitation. The two
emitting components were assigned to originate from the 1ππ* (S3) and 1nπ* (S1) states, respectively. This
was based on the lifetime, the spectral location, and how these varied with the type of solvent. By correlating
the dynamics of the singlet decay with the triplet formation, a direct 1nπ* f 3ππ* ISC mechanism was
found for these compounds with the ISC rate estimated to be ∼5 × 1011 s-1 in both solvent systems.
These photophysical processes were found to be little affected by the kind of leaving group indicating the
common local pHP chromophore is largely responsible for the fluorescence and relevant deactivation
processes. The triplet lifetime was found to be ∼420 and 2130 ps for HPDP and HPA, respectively, in the
mixed solvent compared to 150 and 137 ns, respectively, in neat MeCN. The solvent and leaving group
dependent quenching of the triplet is believed to be associated with the pHP deprotection photochemistry
and indicates that the triplet is the reactive precursor for pHP photorelease reactions for the compounds
examined in this study.
Introduction
containing media. However, in pure organic solvents such as
acetonitrile (MeCN), photoexcitation leads solely to photo-
Rapid and localized liberation of biological stimulants is
needed for real-time monitoring of physiological responses in
biological systems.1-7 The p-hydroxyphenacyl (pHP) group was
recently found to be such an efficient “cage” for the photorelease
of various biological functional groups.8-10 Photodeprotection
of pHP phototriggers occur exclusively in aqueous or aqueous
physical processes resulting in the recovery of the corresponding
ground state molecule. The products of the deprotection are well
established, but much uncertainty exists about the events that
lead to their formation. An example of the products produced
from the deprotection of two pHP compounds are shown below:
(1) Givens, R. S.; Kueper, L. W. Chem. ReV. 1993, 93, 55-66 and references
therein.
(2) Givens, R. S.; Athey, P. S.; Matuszewski, B.; Kueper, L. W., III.; Xue, J.
Y.; Fister, T. J. Am. Chem. Soc. 1993, 115, 6001-6012 and references
therein.
(3) (a) Givens, R. S.; Athey, P. S.; Kueper, L. W., III; Matuszewski, B.; Xue,
J.-Y. J. Am. Chem. Soc. 1992, 114, 8708-8710. (b) Gee, K. R.; Kueper,
L. W., III; Barnes, J.; Dudley, G.; Givens, R. S. J. Org. Chem. 1996, 61,
1228-1233.
(4) (a) Il’ichev, Y. V.; Schworer, M. A.; Wirz, J. J. Am. Chem. Soc. 2004,
126, 4581-4595. (b)Rajesh, C. S.; Givens, R. S.; Wirz, J. J. Am. Chem.
Soc. 2000, 122, 611-618. (c) Hangarter, M.-A.; Ho¨rmann, A.; Kamdzhilov,
Y.; Wirz, J. Photochem. Photobiol. Sci. 2003, 2, 524-535.
(5) Rock, R. S.; Chan, S. I. J. Am. Chem. Soc. 1998, 120, 10766-10767.
(6) Namiki, S.; Arai, T.; Rujimori, K. J. Am. Chem. Soc. 1997, 119, 3840-
3841.
(7) (a) Lee, K.; Falvey, D. E. J. Am. Chem. Soc. 2000, 122, 9361-9366. (b)
Banerjee, A.; Lee, K.; Yu, Q.; Fan, A. G.; Falvey, D. E. Tetrahedron lett.
1998, 39, 4635-4638. (c) Banerjee, A.; Falvey, D. E. J. Org. Chem. 1997,
62, 6245-6251.
Previous mechanistic studies indicate a series of fast steps
are involved in the pHP deprotection.11-13 However, there are
(8) Givens, R. S.; Weber, J. F. W.; Conrad, P. G., II; Orosz, G.; Donahue, S.
L.; Thayer, S. A. J. Am. Chem. Soc. 2000, 122, 2687-2697 and references
therein.
(9) Conrad, P. G., II; Givens, R. S.; Weber, J. F. W.; Kandler, K. Org. Lett.
2000, 2, 1545-1547.
9
10.1021/ja0458524 CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 1463-1472
1463

A R T I C L E S
Ma et al.
conflicting interpretations of these processes and the actual
identity of the putative intermediates involved in the deprotection
reaction. The different interpretations are mainly focused on
the issue of the multiplicity for the reactive precursor and the
mechanism for the cleavage step in the overall reaction. Givens
et al. identified a short-lived triplet as the precursor for pHP
diethyl phosphate (HPDP) and favored a direct triplet cleavage
pathway for the phosphate release.8-11 Wan et al. reported a
singlet mechanism for pHP acetate (HPA) and proposed that
the primary event for the photochemistry is the singlet excited-
state intramolecular proton transfer (ESIPT) mediated by water
solvent molecules.12 Information regarding the intersystem
crossing (ISC) process and its solvent dependence (neat MeCN
vs H2O/MeCN mixed solvent) is crucial to resolving the
controversy about the multiplicity of the reactive precursor
intermediate. In addition, knowledge of the dynamics for event-
(s) that take place due to the presence of H2O in the solvent
environment and the possible influence on the dynamics by
different leaving groups is essential to understand the cleavage
mechanism for the deprotection reactions. Time-resolved mea-
surements are needed to provide direct experimental evidence
for the photophysical processes and reactive intermediates
involved in the deprotection reactions. To help address this,
transient absorption experiments with nanosecond and picosec-
ond time-resolution have been reported for HPA12 and HPDP,11
respectively. Although valuable information has been derived
from these studies, there are still several important gaps that
need to be addressed such as the initial photophysical events
and unambiguous evidence for the multiplicity of the reactive
precursor intermediate that reacts with water.
In this report, femtosecond Kerr gated time-resolved fluo-
rescence (fs-KTRF) combined with picosecond Kerr gated time-
resolved resonance Raman (ps-KTR3) spectroscopy have been
employed for both HPA and HPDP in neat MeCN and H2O/
MeCN mixed solvent systems. The spectral properties and rapid
decay of the singlet excited states were characterized by the
fs-KTRF measurements, and the generation and fate of the triplet
state were monitored by the ps-KTR3 experiments.
HPA and HPDP belong to aromatic carbonyl compound that
have been long known to have very small fluorescence yields
predominately due to ISC conversion that quickly deactivates
the singlet state(s).14-20 Because of experimental difficulties
associated with the extremely short singlet lifetime, weak
fluorescence, and the location of the fluorescence profile in the
UV region, there has been no solution phase time-resolved
fluorescence work reported for this class of compounds to our
knowledge. To deal with the problems associated with obtaining
ultrafast fluorescence spectra in the UV region, we have recently
developed an improved fs-KTRF system to extend the detectable
spectral range down to 270 nm in the UV region and still cover
the visible region up to 600 nm. With the advantage of the
KTRF being able to directly monitor the temporal evolution of
the overall fluorescence profile,21,22 dual fluorescence has been
observed for the first time for the pHP class of compounds.
Assignment of the emitting states has been made based on their
temporal evolution and the dependence of their spectral proper-
ties on solvent.
Picosecond-KTR3 technique has been used for the triplet
Raman measurements. By correlating the singlet decay times
from the fs-KTRF spectra with the triplet formation time from
the ps-KTR3 spectra, the ISC mechanism has been determined
and the ISC rate was estimated. To evaluate the crucial role of
water for the deprotection reaction to take place, we compared
the femtosecond to nanosecond dynamics for HPA and HPDP
in neat MeCN vs H2O/MeCN mixed solvents. These results
provide strong evidence for the multiplicity of the intermediate
that is predominantly produced and further reacts with water in
the deprotection reaction. The present results also shed additional
light on the initial photophysical steps involved in the depro-
tection mechanism of the pHP class of phototrigger compounds.
Experimental Methods
Preparation of HPA and HPDP Samples. HPA was synthesized
following the methods given in refs 12, 23, and 24, and the synthesis
of HPDP is described in the Supporting Information. The identity and
purity of the samples were confirmed by analysis of MS, NMR, and
UV absorption spectroscopy. The details of the synthesis and charac-
terization of HPA and HPDP are given in the Supporting Information.
Spectroscopic grade acetonitrile (MeCN), methanol (MeOH), and
tetrahydrofuran (THF) as well as deionized water were used as solvents
for the experiments.
Femtosecond Kerr Gated Time-Resolved Fluorescence (fs-
KTRF) Experiments. Third harmonic (267 nm) and fundamental laser
pulses (800 nm) from a commercial Ti:Sapphire regenerative amplifier
laser system with a 150 fs pulse duration and 1 kHz repetition rate
were used as excitation and gating pulses, respectively, for the fs-KTRF
experiments. The excitation pulses (∼0.5 µJ) were focused (∼100 µm)
into a 0.5 mm thickness jet stream of sample placed at one focus of an
elliptical mirror. The emission from the sample was collected by the
elliptical mirror and passed through a film polarizer then focused into
the Kerr medium (a 1 mm thickness fused silica plate) placed at the
other focus point of the ellipse. The Kerr medium was placed between
a crossed polarizers pair with an extinction ratio of ∼104. The gating
beam was polarized at 45° and focused to the Kerr medium with
adjusted intensity to create, in effect, a half-waveplate that rotates the
polarization of the light from the sample allowing it to be transmitted
through a Glan Taylor polarizer for the duration of the induced
anisotropy created by the femtosecond gating pulse. The emission that
passed through the second polarizer was focused into a monochromator
and detected by a liquid nitrogen cooled CCD detector. The fluorescence
measurement was performed at the magic angle to eliminate the effect
of sample reorientation. The Kerr gate was opened to sample the
(10) (a) Gives, R. S.; Park, C.-H. Tetrahedron Lett. 1996, 37, 6259-6262. (b)
Park, C.-H.; Givens, R. S. J. Am. Chem. Soc. 1997, 119, 2453-2463. (c)
Givens, R. S.; Jung, A.; Park, C.-H.; Weber, J.; Bartlett, W. J. Am. Chem.
Soc. 1997, 119, 8369-8370.
(11) Conrad, P. G., II; Givens, R. S.; Hellrung, B.; Rajesh, C. S.; Ramseier,
M.; Wirz, J. J. Am. Chem. Soc. 2000, 122, 9346-9347.
(12) (a) Zhang, K.; Corrie, J. E. T.; Munasinghe, V. R. N.; Wan, P. J. Am.
Chem. Soc. 1999, 121, 5625-5632. (b) Brousmiche, D. W.; Wan, P. J.
Photochem. Photobiol., A 2000, 130, 113-118.
(13) Banerjee, A.; Falvey, D. E. J. Am. Chem. Soc. 1998, 120, 2965-2966.
(14) Ohshima, Y.; Fujii, T.; Fujita, T.; Inaba, D.; Baba, M. J. Phys. Chem. A
2003, 107, 8851-8855.
(15) Baronavski, A. P.; Owrutsky, J. C. Chem. Phys. Lett. 2001, 333, 36-40.
(16) (a) Ohmori, N.; Suzuki, T.; Ito, M. J. Phys. Chem. 1988, 92, 1086-1093.
(b) Kamei, S.; Okuyama, K.; Abe, H.; Mikami, N.; Ito, M. J. Phys. Chem.
1986, 90, 93-100.
(21) Browne, W. R.; Coates, C. G..; Brady, C.; Matousek, P.; Towrie, M.;
Botchway, S. W.; Parker, A. W.; Vos, J. G.; McGarvey, J. J. J. Am. Chem.
Soc. 2003, 125, 1706-1707.
(17) Greene, B. I.; Hochstrasser, R. M.; Weisman, R. B. J. Chem. Phys. 1979,
70, 1247-1259.
(22) Kwok, W. M.; Ma, C.; Matousek, P.; Parker, A. W.; Phillips, D.; Towrie,
M. J. Phys. Chem. A 2000, 104, 4188-4197 and reference therein.
(23) Kihara, M.; Ikeuchi, I.; Kobayashi, Y.; Nagao, Y.; Hashizume, M.; Moritoki,
H. Drug Des. DiscoVery 1994, 11, 175-183.
(18) Damschen, D. E.; Merritt, C. D.; Perry, D. L.; Scott, G. W.; Talley, L. D.
J. Phys. Chem. 1978, 82, 2268-2272.
(19) Rentzepis, P. M. Science 1970, 169, 239-247.
(20) Li. R.; Lim, E. C. J. Chem. Phys. 1972, 57, 605-612.
(24) Banerjee, A.; Lee, K.; Falvey, D. E. Tetrahedron 1999, 55, 12699-12710.
9
1464 J. AM. CHEM. SOC. VOL. 127, NO. 5, 2005

Ultrafast Study of p-Hydroxphenacyl Ester Phototriggers
A R T I C L E S
fluorescence spectrum at various time delays following the excitation
pulse by employing a computer controlled optical delay line. All KTRF
spectra were obtained by subtracting the negative time delay signal
from the positive time delay signal and corrected for the detection
system transmission/sensitivity variation with wavelength and the
wavelength-dependent time shift due to the group velocity dispersion
(GVD). The wavelength sensitivity variation correction curve was
obtained by comparing the 20 ps time delay KTRF spectrum of trans-
stilbene (this covers the spectral range of interest in the present study
and shows very little change on the shape of the KTRF spectra at
difference time delays) with its sensitivity corrected steady-state
spectrum from a commercial fluorescence spectrometer. The GVD of
the system (which was about 400 fs from 280 to 500 nm) was corrected
by equation δt ) 0.0033λ2 - 4.5117λ + 1253.8, where δt is time shift
(fs) and λ is wavelength (nm). The constants are estimated by peak
fitting of the KTRF spectra of white light continuum created by intense
pump pulse. The instrument response functions (IRF), which are
wavelength dependent and vary from 200 to 350 fs for wavelength
500 to 280 nm, were measured from the duration of the white light
continuum and also the fluorescence rise time from trans-stilbene.
Picosecond Kerr Gated Time-Resolved Resonance Raman (ps-
KTR3) Experiments. For the ps-KTR3 measurements, the Ti:Sapphire
regenerative amplifier laser system was operated in picosecond mode
with a 800 nm, ∼1 ps, and 1 kHz output. Samples were pumped by
267 nm pulse and probed with 400 and 342 nm pulses. The 267 nm
pump and 400 nm probe pulses were the third and second harmonic of
the regenerative amplifier fundamental, respectively. The 342 nm probe
pulse was generated by mixing of the 800 nm pulse and the 532 nm
output from a home-built OPA system pumped by the 400 nm laser
pulse. The pump and probe pulse durations were ∼1.5 ps; the pulse
energies at the sample were 2-3 µJ. The sampling and signal collecting
systems were the same as those in the fs-KTRF experiments except
that the film polarizer was replaced by a Glan Taylor polarizer and the
Kerr medium was changed to a 2 mm UV cell with benzene in it to
improve the extinction ratio and efficiency of the gate. The polarization
of the pump and probe beams were set at the magic angle and focused
to a spot size of approximately 200 µm at the sample. Details of the
ps-KTR3 experimental apparatus our laser system is based on are given
in refs 25 and 26.
Figure 1. (a) Kerr gated time-resolved fluorescence contour of HPA (p-
hydroxyphenacyl acetate) obtained with 267 nm excitation in MeCN. (b)
Normalized fluorescence decay at 330 nm (circles in blue) and 440 nm
(circles in red) for HPA in MeCN. The solid lines show two exponential
fittings to the experimental data; the dotted line is the instrumental response
function (see text for details). (c-f) Steady-state absorption spectrum (in
black) and typical fluorescence profile of the blue (in blue) and red (in
red) fluorescence for HPA in MeCN (c), THF (d), MeOH (e), and 90%
H2O/10% MeCN mixed solvent (f). The absorption spectra were normalized
to the blue fluorescence spectra. Sharp features indicated by “#” and “*”
are the solvent Raman band and the second harmonic generation of the
800 nm gating pulse from the Kerr medium, respectively.
The fs-KTRF and ps-KTR3 spectra were calibrated by using Hg lamp
and solvent Raman peaks, respectively. Sample solutions for these
experiments were typically at ∼1 mM concentration and renewed every
2 hours during the experiments, and within this period of time no
degradation was observed by UV absorption spectroscopy.
Results and Discussion
A. Femtosecond Kerr Gated Time-Resolved Fluorescence
(fs-KTRF) Spectroscopy. Figure 1a shows the 3D contour of
the KTRF spectra for HPA obtained with 267 nm excitation in
MeCN with a time delay up to 4 ps after excitation. Clearly,
rather than a simple decay of one band, the fluorescence profile
shows two components in both the time and wavelength regimes.
An intense emission band with a maximum near 340 nm
predominates at very early times and disappears at ∼0.5 ps.
The relatively weak emission band that has a maximum at ∼420
nm persists up to ∼3 ps. Fluorescence decay kinetics at 330
and 440 nm together with the IRF are displayed in Figure 1b.
The IRF is represented by a Gaussian band shape with a 290 fs
at 330 nm.
sum of two exponential components with time constants of 0.08
and 1.78 ps, respectively, but employing different proportionality
factors. The 330 nm decay is predominantly by the faster
component, while the longer-lived component contributes mostly
to the decay of the 440 nm emission band. Decay kinetics at
other wavelengths can be fitted in the same way. This appears
to indicate that the fluorescence is composed of two components
originating from different electronic excited states. Provided the
different lifetimes, we tentatively take the spectrum recorded
at -0.2 and 0.9 ps (shown in Figure 1c) as representative of
the net fluorescence spectrum for the short- and long-lived
emissive states, respectively. According to the spectral location,
they are denoted temporarily to “blue” and “red” fluorescence,
respectively. We found that, by using a log-normal function27
It is also shown in Figure 1b that, by convolution with the
IRF, both the decay curves can be simulated satisfactorily by a
(25) Matousek, P.; Towrie, M.; Ma, C.; Kwok, W. M.; Phillips, D.; Toner, W.
T.; Parker, A. W. J. Raman Spectrosc. 2001, 32, 983-988.
(26) Matousek, P.; Towrie, M.; Stanley, A.; Parker, A. W. Appl. Spectrosc. 1999,
53, 1485-1489.
9
J. AM. CHEM. SOC. VOL. 127, NO. 5, 2005 1465

A R T I C L E S
Ma et al.
Table 1. Kerr Gated Time-Resolved Fluorescence and Steady-State Absorption Spectral Parameters for HPA and HPDP in Neat and 90%
H2O/10% MeCN (v/v) Mixed Solvents
HPA
HPDP
MeCN
90%H2O
MeOH
THF
MeCN
90%H2O
τ1 (ps)
0.08 ( 0.03
1.78 ( 0.05
334
427
272
1.45
4.4
5.0
0.08 ( 0.03
0.08 ( 0.03
1.25 ( 0.05
341
420
279
1.35
6.9
5.5
0.11 ( 0.03
1.90 ( 0.05
334
433
273
1.42
5.4
7.3
0.08 ( 0.03
1.98 ( 0.05
331
433
273
1.28
6.1
5.5
0.07 ( 0.03
τ2 (ps)
0.89 ( 0.05
0.76 ( 0.05
λ1fluo (nm)
λ2fluo (nm)
356
392
280
1.23
16
354
401
280
0.86
15
λ
abs (nm)
ꢀ (× 104 L mol-1 cm-1
)
I2/I1 (× 10-2
)
Φ (× 10-5
)
6.5
4.0
(see Supporting Information for details) and nonlinear least-
squares regression, the fluorescence profile obtained at each time
delay can be reproduced by overlap of the blue and red
components with different weighted factors. The weaker
intensity of the red band compared to the blue band may reflect
the weaker oscillator strength of the former than latter transition.
The resemblance of the steady-state absorption spectra of
HPA and HPDP to that of the “cage” p-hydroxyacetophenone
(HA) (see Supporting Information for the comparative absorp-
tion spectra) indicates that, for both of the compounds, the pHP
moiety is the chromophore responsible for the photoexcitation.
The 267 nm photoexcitation of these compounds leads to the
prompt population of the strongly allowed ππ* (La type) S3
state associated with the lowest strong absorption band (shown
in Figure 1c). The two lower energy ππ* S2 (Lb type) and nπ*
S1 bands are buried under the intense S3 band due to their
closeness (the S2 state) to the S3 absorption and their weaker
oscillator strength (both the S1 and S2 states).28-30 The S1
absorption was found to be at ∼346 nm for HA.29 Extensive
investigations on aromatic carbonyl compounds reveal that the
energy levels of the excited ππ* and nπ* states are affected
significantly by the solvent H-bonding strength and the solvent
polarity: increasing the H-bonding strength or polarity stabilizes
the ππ* states, whereas these effects destabilize the nπ* state
at the same time.31-39 This solvent-dependent modification of
these energy levels has been well rationalized and ascribed to
account for the characteristic blue shift of the nπ* and red shift
of the ππ* spectral bands in solution.31-33,40,41 It is therefore
necessary to perform the KTRF measurement in solvents of
different polarity and H-bonding ability to help assign the
observed fluorescence and to justify the above dual fluorescence
assignment. In addition to neat MeCN that is an aprotic solvent
with high polarity (ꢀ ) 33.8) and 90% H2O/10% MeCN (v/v)
mixed solvent that is a strong H-bonding and very polar solvent,
we chose two additional solvents methanol (MeOH) and
tetrahydrofuran (THF) for KTRF experiments on HPA. MeOH
has a midrange H-bonding strength and a polarity (ꢀ ) 33.1)
similar to that of MeCN, while THF is an aprotic solvent of
weak polarity (ꢀ ) 7.6).27,41
The KTRF results obtained in the H2O/MeCN mixed solvent,
MeOH, and THF are generally similar to that in pure MeCN
but with distinct variations in the spectral location and decay
rate in the different solvents. Table 1 lists the time constants
obtained from fitting the 330 and 440 nm decay kinetics in the
four different solvents. The maximum band positions of the
typical blue and red fluorescence bands and the corresponding
steady-state S3 absorption band are also summarized in Table
1. The band positions were acquired by log-normal fitting of
the respective spectral profile (see Supporting Information). In
Figure 1d-f, the typical fluorescence and absorption spectra in
THF, MeOH, and H2O/MeCN mixed solvent are shown
respectively to compare with those recorded in pure MeCN.
From Table 1 and Figure 1c-f, it is clear that, from MeCN
to MeOH to H2O/MeCN mixed solvent, the blue fluorescence
downshifts to the red and the red fluorescence upshifts to the
blue. These changes in the spectral location are accompanied
by a lifetime shortening of the red band and an increase in the
intensity ratio of the red to blue band manifested by I2/I1 in
Table 1 (see Supporting Information for how these were
estimated). The THF data are rather close to those of MeCN
but with a slight red shift and longer lifetime for the THF red
band compared to that of MeCN. Obviously, the solvent
H-bonding and polarity dependent shifts coincide with the
characteristic solvent-dependent spectral shift expected for these
aromatic compounds. We consider this as convincing evidence
for the coexistence of two fluorescence components; one is the
ππ* character state emitting the blue fluorescence, and the other
is the nπ* state emitting the red fluorescence. We note that
processes such as excited-state intramolecular vibrational re-
distribution (IVR), vibrational relaxation, and solvent reorienta-
tion can lead to spectral shifts and bandwidth changes for TRF
profiles and may sometimes occur in similar time scales.27,42-44
Although some of these processes could possibly occur for
(27) Horng, M. L.; Gardecki, J. A.; Papazyan, A.; Maroncelli, M. J. Phys. Chem.
1995, 99, 17311-17337.
(28) Dym, S.; Hochstrasser, R. M. J. Chem. Phys. 1969, 51, 2458-2468.
(29) (a) Kearns, D. R.; Case, W. A. J. Am. Chem. Soc. 1966, 88, 5087-5097.
(b) Case, W. A.; Kearns, D. R. J. Chem. Phys. 1970, 52, 2175-2191.
(30) Chan, W. S.; Ma, C.; Kwok, W. M.; Zuo, P.; Phillips, D. L. J. Phys. Chem.
A 2004, 108, 4047-4058.
(31) (a) Bhasikuttan, A. C.; Singh, A. K.; Palit, D. K.; Sapre, A. V.; Mittal, J.
P. J. Phys. Chem. A. 1998, 102, 3470-3480. (b) Singh, A. K.; Bhasikuttan,
A. C.; Palit, D. K.; Mittal, J. P. J. Phys. Chem. A 2000, 104, 7002-7009.
(32) (a) Mitsui, M.; Ohshima, Y. J. Phys. Chem. A 2000, 104, 8638-8648. (b)
Mitsui, M.; Ohshima, Y.; Ishiuchi, S.-i.; Sakai, M.; Fujii, M. J. Phys. Chem.
A 2000, 104, 8649-8659. (c) Mitsui, M.; Ohshima, Y.; Kajimoto, O. J.
Phys. Chem. A 2000, 104, 8660-8670.
(33) (a) Morlet-Savary, F.; Ley, C.; Jacques, P.; Wieder, F.; Fouassier, J. P. J.
Photochem. Photobiol. A. 1999, 126, 7-14. (b) Ley, C.; Morlet-Savary,
F.; Jacques, P.; Fouassier, J. P. Chem. Phys. 2000, 255, 335-346.
(34) Cavaleri, J. J.; Prater, K.; Bowman, R. M. Chem. Phys. Lett. 1996, 259,
495-502.
(35) van der Burgt, M. J.; Jansen, L. M. G.; Huizer, A. H.; Varma, C. A. G. O.
Chem. Phys 1995, 201, 525-538.
(36) Hamanoue, K.; Nakayama, T.; Yamaguchi, T.; Ushida, K. J. Phys. Chem.
1989, 93, 3814-3818.
(37) Scaiano, J. C. J. Am. Chem. Soc. 1980, 102, 7747-7753.
(38) Dalton, J. C.; Montgomery, F. C. J. Am. Chem. Soc. 1974, 96, 6230-
6232.
(39) Rusakowicz, R.; Byers, G. W.; Leermakers, P. A. J. Am. Chem. Soc. 1971,
93, 3263-3266.
(40) Turro, N. J. Modern Molecular Photochemistry; University Science Book:
California, 1991.
(41) Reichardt, C. SolVent and SolVent Effect in Organic Chemistry; VCH:
Verlagsgesellschaft mbH, D-6940 Weinheim, 1988.
(42) Macpherson, A. N.; Gillbro, T. J. Phys. Chem. A 1998, 102, 5049-5058.
9
1466 J. AM. CHEM. SOC. VOL. 127, NO. 5, 2005

Ultrafast Study of p-Hydroxphenacyl Ester Phototriggers
A R T I C L E S
photoexited HPA, the specific solvent-dependent lifetime and
spectral shifts indicate that none of them are likely to be the
major factor accounting for the KTRF observations.
observation that the extent of the fluorescence shifts (for both
1ππ* and 1nπ*) correlates with the solvent H-bonding strength
indicates the importance of excited state solvent-solute H-
bonding interaction in affecting the π* f π and π * f n
emission processes. From Table 1, it can be seen that, due to
the H-bonding interaction, the energy separation between the
emitting 1ππ* and 1nπ* states reduces substantially from ∼6500
cm-1 in neat MeCN to ∼2600 cm-1 in H2O/MeCN mixed
solvent for HPA and a similar amount for HPDP. This
consequently results in stronger vibronic coupling between the
two states in H2O/MeCN mixed solvent than in neat MeCN
solvent. Out-of-plane vibrations such as -C(O)CH2- torsion
and -(O)C-CH2- rotation could be the most probable modes
to effectively couple these two states.49 This vibronic coupling
between the close-lying ππ* and nπ* states, the so-termed
“proximity effect”, influences substantially the radiative and
nonradiative dynamics of the excited states and can account
for the lifetime and additional spectral differences observed here
in aprotic vs protic solvents (see below).35,50 We note, however,
that the order of the ππ* vs nπ* energy level remains the same
in all the solvents investigated. This implies that deactivation
processes of the two singlet states will not be dramatically
different in the two kinds of solvent, and this is indeed confirmed
by the ISC dynamics revealed in the following ps-TR3 experi-
ments.
Consistent with the ππ* nature of the S3 state, the S3
absorption band shifts to the red from MeCN/THF to MeOH
to H2O/MeCN mixed solvent (Table 1), and this behavior
parallels the observed shifts of blue fluorescence band. This
combined with the allowed nature of the relatively intense blue
fluorescence band and the roughly mirror image between it and
the S3 ππ* absorption band leads us to assign the S3 state as
the source emitting the blue fluorescence. The weakness of the
red fluorescence is consistent with its origin from the weakly
allowed S1 nπ* state. According to these assignments and using
the 1ππ* (S3) extinction coefficient (the ꢀ values listed in Table
1) deduced from the absorption spectra,40,45 an overall fluores-
cence yield of ∼6 × 10-5 (this does not much depend on the
solvent system) can be estimated based on the integration of
the measured KTRF spectra and the fluorescence lifetimes
(Table 1; for details see Supporting Information).46 This is
consistent with previous observations that the fluorescence of
HA is below the normal steady-state detection limit and indicates
a very high efficiency of the singlet deactivation processes.20
We note the possibility of interference from the Lb type ππ*
S2 state to the blue fluorescence. The S0 to S2 transition is
symmetrically forbidden but the Lb state can mix with the nearby
La ππ* (S3) state by vibrational coupling and borrow intensity
from the optically allowed S3 state.28-30,47 However, based on
the result that the KTRF decay kinetics can be fitted reasonably
well by only two exponential components, we consider the blue
fluorescence is predominantly from the S3 state in this prelimi-
nary account and the S2 state will not be considered further
here. To avoid any ambiguity in the assignments, we refer
hereafter to the emitting states to be 1ππ* and 1nπ*, rather than
S3 and S1, as the origin of the blue and red fluorescence bands,
respectively.
1
1
The dual fluorescence phenomenon was first reported for
4-dimethylaminobenzonitrile (DMABN) and its derivatives more
than forty years ago.22,51 The solvent polarity has been deduced
as the main factor to influence the DMABN fluorescence profile
rather than the H-bonding strength.51-53 In nonpolar solvents,
the DMABN fluorescence is “normal” and originates from a
locally excited (LE) state. An additional deeply red shifted band
appears in polar solvents, and this new band has been found to
be from an intramolecular charge transfer (ICT) state produced
from the LE state on a picosecond time scale.22,52-54 Obviously,
the dual fluorescence observed here for HPA and HPDP is
different from that of the DMABN family of compounds due
to fundamental differences in the nature of the emitting states
as well as the behavior of the solvent dependence. Here the
1nπ* state is formed by internal conversion (IC) from the
The KTRF results for HPDP in MeCN and 90% H2O/10%
MeCN mixed solvent are similar to those of HPA in the
corresponding solvents and the data obtained is also listed in
Table 1. In addition, for both HPA and HPDP, the KTRF spectra
in 50% H2O/50% MeCN (v:v) mixed solvent were found to be
almost identical to the corresponding spectra recorded in 90%
H2O/10% MeCN mixed solvent. The similarity between the
KTRF spectra of HPA and HPDP implies the properties of the
emitting states are little affected by the leaving group consistent
with the pHP subgroup being the common chromophore for
the fluorescence processes.
1
initially populated ππ*. The IC dynamics cannot be resolved
in this work due to spectral overlap between the 1ππ* and 1nπ*
fluorescence profile and the ∼250 fs time resolution but is
1
expected to correspond to the ππ* decay that has a ∼80 fs
time constant as shown in Table 1. We note that, by using an
ultrafast TRF technique, fluorescence from singlet excited states
higher than S1 has been reported for several systems46,55-56
The moderate solvent polarity dependence and small Stokes
1
shift between the ππ* absorption (S3) and the fluorescence
bands imply the 1ππ* state has little charge transfer (CT)
character consistent with the previously reported La 1ππ*
character for aromatic carbonyl compounds with hydroxy
substitution at the para-position.31,48 On the other hand, the
(49) (a) Lim, E. C.; Li, Y. H.; Li, R. J. Chem. Phys. 1970, 53, 2443-2448. (b)
Li, Y. H.; Lim, E. C. Chem. Phys. Lett. 1970, 7, 15-18.
(50) (a) Lim, E. C. J. Phys. Chem. 1986, 90, 6770-6777. (b) Lai, T.-i.; Lim, E.
C. Chem. Phys. Lett. 1980, 73, 244-248.
(51) Grabowski, Z. R.; Rotkiewicz, K.; Rettig, W. Chem. ReV. 2003, 103, 3899-
4032.
(52) Kwok, W. M.; George, M. W.; Grills, D. C.; Ma, C.; Matousek, P.; Parker,
A. W.; Phillips, D.; Toner, W. T.; Towrie, M. Angew Chem., Int. Ed. 2003,
42, 1826-1830.
(43) Gustavsson, T.; Cassara, L.; Gulbinas, V.; Gurzadyan, G.; Mialocq, J.-C.;
Pommeret, S.; Sorgius, M.; van der Meulen, P. J. Phys. Chem. A 1998,
102, 4229-4245.
(53) Kwok, W. M.; Ma, C.; George, M. W.; Grills, D. C.; Matousek, P.; Parker,
A. W.; Phillips, D.; Toner, W. T.; Towrie, M. Phys. Chem. Chem. Phys.
2003, 5, 1043-1050.
(44) Akimoto, S.; Yamazaki, I.; Sakawa, T.; Mimuro, M. J. Phys. Chem. A
2002, 106, 2237-2243.
(45) Dalton J. C.; Turro, N. J. J. Am. Chem. Soc. 1971, 93, 3569-3570.
(46) Takeuchi, S.; Tahara, T. J. Phys. Chem. A 1997, 101, 3052-3060.
(47) Leopold, D. G.; Hemley, R. J.; Vaida, V J. Chem. Phys. 1981, 75, 4758-
4769.
(48) Bhasikuttan, A. C.; Singh, A. K.; Palit, D. K.; Sapre, A. V.; Mittal, J. P.
J. Phys. Chem. A 1999, 103, 4703-4711.
(54) Kwok, W. M.; Ma, C.; Matousek, P.; Parker, A. W.; Phillips, D.; Toner,
W. T.; Towrie, M.; Umapathy, S. J. Phys. Chem. A 2001, 105, 984-990.
(55) Bhasikuttan, A. C.; Sapre, A. V.; Okada, T. J. Phys. Chem. A 2004, 107,
3030-3035.
(56) Leupin, W.; Berens, S. J.; Magde, D. J. Phys. Chem. 1984, 88, 1376-
1379.
9
J. AM. CHEM. SOC. VOL. 127, NO. 5, 2005 1467

A R T I C L E S
Ma et al.
Figure 3. Temporal dependence of the triplet ∼1600 cm-1 band areas for
HPDP (a) and HPA (b) in 50% H2O/50% MeCN mixed solvent (circles)
and neat MeCN (squares) obtained in 400 nm probe ps-KTR3 spectra. Solid
lines show an exponential fitting of the experimental data.
Figure 2. Picosecond Kerr gated time-resolved resonance Raman spectra
of HPDP obtained with 267 nm pump and 400 nm probe wavelengths in
50% H2O/50% MeCN mixed solvent (left) and neat MeCN (right).
the common chromophore (pHP moiety) and a similar triplet
structure. For details of the vibrational spectra and assignments,
the reader is referred to refs 60 and 61.
including various carotenoids.57-59 The observation here of the
strong S3 fluorescence is consistent with the large extinction
coefficient of this 1ππ* state vs the 1nπ* S1 state. It is
worthwhile to point out that, to our knowledge, the fluorescence
results presented here on HPA and HPDP represent the first
ultrafast time-resolved fluorescence study on the pHP photo-
trigger compounds and also for related aromatic carbonyl
compounds.
The temporal change of the Lorentzain fitted ∼1600 cm-1
band area was found, and Figure 3a and b display a kinetic
3
comparison for the ππ* state in pure MeCN and H2O/MeCN
mixed solvent for HPDP and HPA, respectively. Clearly, the
early time 3ππ* development is similar for both the compounds
in the two solvents (with ∼7-12 ps time constant for the triplet
growth); it is striking, however, that the 3ππ* lifetime is much
shorter in H2O/MeCN mixed solvent than in pure MeCN
solvent. The ∼1600 cm-1 Raman band intensity decays were
fitted using a single-exponential function, and lifetimes of 420
( 10 ps and 2130 ( 10 ps were estimated for HPDP and HPA,
respectively. The ∼420 ps HPDP triplet decay time coincides
with the ∼400 ps time constant reported by Wirz and co-workers
for the same system using transient absorption spectroscopy.11
The triplet lifetime has been found to be ∼150 and 137 ns for
HPDP and HPA, respectively, in neat MeCN.60,61 This solvent
and leaving group dependent quenching of the triplet implies
further reaction takes place on the triplet manifold specifically
in the H2O containing solvent. It is straightforward that the
reaction is related to the pHP photocleavage process even if it
does not directly undergo cleavage. We therefore ascribe the
triplet decay time to the time scale occurring for the further
deprotection photochemistry. The observations here are hence
compelling evidence that the triplet state is the predominant
reactive precursor to the pHP deprotection reaction and rules
out the singlet mechanism12 for HPA and HPDP after excitation
with 267 nm light.
B. Picosecond Kerr Gated Time-Resolved Resonance
Raman (ps-KTR3) Spectroscopy. To establish the deactivation
channel of the photopopulated 1ππ* and 1nπ* states, especially
the ISC conversion to the triplet manifold, the dynamics for
the triplet formation are essential to determine. Using a 267
nm pump and 400 nm probe excitation wavelengths, our recent
TR3 measurements found early picosecond formation and
hundreds nanoseconds lifetime of the ππ* triplet (3ππ*) for
HPA and HPDP in pure MeCN.60,61 We performed the same
experiments for the two compounds in 50% H2O/50% MeCN
(v/v) mixed solvent. Figure 2 displays the comparison between
the spectra obtained in 50% H2O/50% MeCN (v/v) mixed
solvent to spectra obtained in neat MeCN for HPDP. Due to
the 400 nm probe wavelength being near the maximum position
of the T1 f Tn absorption band11 plus the similarity between
the spectra in the two solvents, the spectra in the H2O/MeCN
mixed solvent can be attributed confidently to the HPDP 3ππ*
state. The corresponding ps-KTR3 spectra for HPA are very
similar to those of HPDP and are shown in the Supporting
Information. As discussed in our earlier work,60,61 the similarity
of the triplet spectra of HPDP to those of HPA is indicative of
Since the ∼7-12 ps growth time of the triplet does not
exactly match the decay time of either the 1ππ* or 1nπ*
fluorescence, further evidence is required to better elucidate the
ISC mechanism. ps-KTR3 spectra of HPA and HPDP obtained
in pure MeCN and 50% H2O/50% MeCN mixed solvents with
267 nm pump and 342 nm probe excitation wavelengths provide
this needed information about the ISC mechanism. Figure 4a
presents early time ps-KTR3 spectra in the 1450-1800 cm-1
(57) Holt, N. E.; Kennis, J. T. M.; Dall’Osto, L.; Bassi, R.; Fleming, G. R.
Chem. Phys. Lett. 2003, 379, 305-313.
(58) Mimuro, M.; Nagashima, U.; Nagaoka, S.-i.; Nishimura, Y.; Takaichi, S.;
Katoh, T.; Yamazaki, I. Chem. Phys. Lett. 1992, 191, 219-224.
(59) Kandori, H.; Sasabe, H.; Mimuro, M. J. Am. Chem. Soc. 1994, 116, 2671-
2672.
(60) Ma, C.; Chan, W. S.; Kwok, W. M.; Zuo, P.; Phillips, D. L. J. Phys. Chem.
B 2004, 108, 9264-9276.
(61) Ma, C.; Zuo, P.; Kwok, W. M.; Chan, W. S.; Kan, J. T. W.; Toy, P. H.;
Phillips, D. L. J. Org. Chem. 2004, 69, 6641-6657.
9
1468 J. AM. CHEM. SOC. VOL. 127, NO. 5, 2005

Ultrafast Study of p-Hydroxphenacyl Ester Phototriggers
A R T I C L E S
family of molecules, and such information is not available for
the pHP compounds studied here to the best of our knowledge.
It would therefore be helpful at some point to acquire transient
absorption spectra with femtosecond resolution in the future to
directly observe the S1f Sn absorption and its spectral and
kinetics evolution for these compounds.
C. ISC Mechanism and Deactivation of the Excited States.
1
The dynamics and nπ* S1 assignment of the early species
observed in the TR3 spectra combined with the KTRF result
1
3
indicates a direct nπ* (S1) f ππ* (T1) ISC pathway for the
HPA and HPDP in neat and H2O/MeCN mixed solvent. This
may also suggest that such a mechanism is generally applicable
for these compounds in other solvents. This ISC pathway is
consistent with El-Sayed’s rule and follows a direct spin-orbital
coupling mechanism.40 Givens and co-workers estimated the
triplet energies of HPA and HPDP are 71.0 and 70.6 kcal/mol
in EPA, corresponding in wavelength units to ∼403 and 405
nm, respectively.8 These values are rather close to the locations
1
of the nπ* fluorescence for the two compounds (Table 1)
implying a small energy separation between the 1nπ* and 3ππ*
level and consequent strong coupling between the two states.
This provides an interpretation for the rapid ISC conversion
and accounts for the short singlet lifetime and large triplet yield.
This ISC mechanism associated with strong singlet-triplet
coupling may be applicable generally to a wide range of
aromatic carbonyls that have a lowest nπ* singlet and ππ*
triplet, such as derivatives of acetophenone.
Figure 4. Picosecond Kerr gated time-resolved resonance Raman spectra
obtained for HPDP in 50% H2O/50% MeCN mixed solvent with 267 nm
pump and 342 nm probe (a) and 400 nm probe (b) wavelengths, respectively.
region for HPDP in 50% H2O/50% MeCN mixed solvent using
400 and 342 nm probe wavelengths after 267 nm pump
excitation. It is clear that the band at ∼1610 cm-1 appearing at
2 ps and afterward at both probe wavelengths is from the same
3ππ* triplet that is also observed in spectra obtained in neat
MeCN solvent (see Figure 2). Inspection of Figure 4 shows
that an additional Raman band at about 1550 cm-1 appears only
at very early times (up to 4 ps), and this feature decays very
rapidly to form the 3ππ* triplet state. The absence of the early
time ∼1550 cm-1 Raman band in the 400 nm probe ps-TR3
spectra implies this 1550 cm-1 Raman band is from another
short-lived species that is not the 3ππ* triplet state. The lifetime
of the species associated with the ∼1550 cm-1 Raman band is
within or close to the IRF of the present TR3 system (∼1-2
ps) and thus cannot be well-resolved but estimated to be on the
same time-scale as the IRF. Similar spectra and dynamics were
also observed for HPDP in neat MeCN solvent and for HPA in
both solvent systems (these spectra are shown in the Supporting
Information).
It is important to point out that our data provides no indication
1
for the existence of an alternative ISC pathway ππ* (S3) f
3
3nπ* (T2) f ππ* (T1) although such a process is possible,
since the T2 state lies between the S3 and T1 and close to both
the states.29 Provided the ∼80 fs S3 lifetime (Table 1) and the
fast T2 to T1 internal conversion that is believed to occur also
3
in femtosecond time scale, the ππ* state produced following
this ISC channel can be expected to be detected promptly by
our TR3 measurements, which is not our experimental observa-
tion: as obviously indicated by the TR3 spectra shown in Figures
2 and 4, the triplet is not detectable clearly until ∼2 ps. It is
1
3
therefore rather unlikely that this ππ* (S3) f nπ* (T2) f
3ππ* (T1) ISC pathway can account for the generation of the
3ππ* studied here for HPA and HPDP.
1
Our result that the ISC occurs from the nπ* S1 state even
with S3 excitation has important implications in elucidating the
deactivation processes of the excited states. It implies that (i)
the major deactivation channel of the photopopulated S3 state
is internal conversion to S1; (ii) the ISC of pHP caged
compounds may be relatively independent of the excitation
wavelength. We note that influence of excitation wavelength
on ISC processes has been reported previously for some
aromatic carbonyl compounds, and it appears that the effect can
Significantly, the decay time of the early transient with the
∼1550 cm-1 Raman band correlates with the lifetime of the
red fluorescence attributed to the 1nπ* S1 state presented in the
preceding KTRF measurements. Thus, we assign the new early
1
time ∼1550 cm-1 Raman band to be due to the nπ* S1 state.
1
The appearance of the nπ* Raman band in the 342 nm ps-
be especially significant under gas-phase conditions.62-65
A
KTR3 spectra and its absence in the 400 nm ps-KTR3 spectra
implies the 1nπ* S1 f Sn transitions absorb appreciably at 342
nm but not at 400 nm. Previously reported transient absorption
spectra indicate the 342 nm probe wavelength is located in the
blue tail of the T1 f Tn absorption for both HPA and HPDP in
the two solvents.11,12 This is consistent with the weaker Raman
signal for the 3ππ* triplet Raman bands in the 342 nm spectra
compared to the 400 nm spectra. Probably due to the short S1
lifetime, there has been very little S1 f Sn transient absorption
spectroscopy reported in the literature for the aromatic carbonyl
kinetic model including ISC from both the nπ* and ππ* singlet
states has also been proposed for aromatic carbonyl such as
xanthone in solution phase.34 On the other hand, Neckers,
Rodgers, and co-workers reported very recently that the ISC of
(62) (a) Rennert, A. R.; Steel, C. Chem. Phys. Lett. 1981, 78, 36-39. (b) Berger,
M.; Steel, C. J. Am. Chem. Soc. 1975, 97, 4817-4821.
(63) Itoh, T.; Baba, H.; Takemura, T. Bull. Chem. Soc. Jpn. 1978, 51, 2841-
2846.
(64) Warren, J. A.; Bernstein, E. R. J. Chem. Phys. 1986, 85, 2365-2367.
(65) (a) Hirata, Y.; Lim, E. C. J. Chem. Phys. 1980, 72, 5505-5510. (b) Hirata,
Y.; Lim, E. C. J. Chem. Phys. 1980, 73, 3804-3809.
9
J. AM. CHEM. SOC. VOL. 127, NO. 5, 2005 1469

A R T I C L E S
Ma et al.
benzophenone in solution is insensitive to the excitation
wavelength.66 This is consistent with our present results for HPA
and HPDP in solution. Since the pHP deprotection reaction
occurs on the triplet manifold, the independence of ISC process
on excitation wavelength implies that the photorelease processes
could also be relatively independent of the excitation wave-
length. This may have some implication for practical applica-
tions of the pHP phototriggers in terms of selection of an
excitation source. However, other factors also need to be
considered. Compared with the S1 excitation, using higher
excitation energy (to S2 or S3) could possibly result in an
enhanced efficiency of internal conversion back to ground state
to deactivate the singlet excited state.50,65 This could cause the
triplet yield to be smaller with higher excitation than lower
excitation energy. In this sense, the S1 excitation with the
wavelength above 300 nm, as generally employed in previous
pHP deprotection studies,8-12 would favor in giving high
efficiency for liberation of the protected groups. Further work
is needed to better address the actual wavelength dependence
of the pHP deprotection reactions.
available implies that the deprotonation is not likely to account
for the S1 deactivation. However, as discussed below, depro-
tonation on the other hand could be involved in the deactivation
of the triplet in aqueous solution. We therefore suggest the ISC
rate as being ∼5 × 1011 s-1 (reciprocal of the 1nπ* S1 lifetime)
for HPA and HPDP in neat as well as H2O/MeCN mixed solvent
and take this as an upper limit value. This value is very close
to the result obtained using the triplet quenching method as
reported by Givens, Wirz, and co-workers.8,9,10b,11
The stronger 1nπ* and 1ππ* vibronic coupling in H2O/MeCN
mixed solvent than in neat MeCN could also account for the
1
1
increased intensity ratio of the nπ* to ππ* fluorescence in
the former to latter solvent system. Our KTRF result shows
that relative fluorescence integration of the 1nπ* to 1ππ* band
increases (by ∼3 times larger in H2O/MeCN mixed solvent
related to neat MeCN, Table 1) as the two bands shift close to
each other. Assuming a similar oscillator strength for the 1ππ*
fluorescence transition, this implies an enhancement of the 1nπ*
1
1
oscillator strength associated with the smaller nπ* to ππ*
energy gap. This is well within expectation considering that
electronic transition between the S0 and 1nπ* states is symmetric
forbidden for planar aromatic carbonyls but can derive some
finite allowed nature via vibronic coupling with the nearby 1ππ*
state(s) (most probably the S3 state here); the extent of the
allowed character depends on the strength of the vibronic
coupling that is substantially favored by the closer energy gap
between the two states in H2O/MeCN mixed than in neat MeCN
solvent.28-30,37,38,49,50 It is expected that this result here may spur
further solution phase experimental and theoretical work on the
singlet excited state properties for aromatic carbonyls especially
in H-bonding solvents, since, to our knowledge, there has been
little work reported on this aspect in the literature.
According to the 1nπ* (S1) f 3ππ* (T1) ISC pathway,
1
estimation of the ISC rate requires, besides the nπ* lifetime
1
(Table 1), information of the nπ* f S0 internal conversion
rate kIC, that is not yet available for the compounds studied here
to our knowledge. It has been well established in previous
investigations that the relative contributions of IC vs ISC as
nonradiative channels to deactivate the S1 excited state are
strongly influenced by the vibronic interaction between the
close-lying 1nπ* and 1ππ* states that is, in turn, determined by
the 1nπ* and 1ππ* energy gap: the smaller the energy gap, the
stronger the vibronic coupling, and consequently the more
efficient the IC process becomes to deactivate the S1 state.50,65
Based on this viewpoint, our result that the S1 lifetime is shorter
in H2O/MeCN mixed solvent (∼0.89 and 0.76 ps for HPA and
HPDP, respectively, Table 1) than that in neat MeCN (∼1.78
and 1.98 ps for HPA and HPDP, respectively, Table 1)
combined with the much smaller 1nπ* and 1ππ* energy gap in
the mixed than neat solvent might indicate an increased
contribution of the IC to depopulate the S1 state in the former
rather than latter solvent. In this sense, the IC may not be
negligible and competes with the ISC to deactivate the S1 state.
However, the generally high ISC yield as reported in previous
work for aromatic carbonyl compounds19,67 implies the pre-
dominance of the ISC as the deactivation channel. In addition,
we note that phenols become more acidic in the excited than in
the ground state; it is therefore necessary to consider the
possibility that deprotonation of the phenolic proton could act
as a competing process to deactivate the S1 state. Previous
work9,11,12 found that the deprotonation does occur after pho-
toexcitation of the pHP caged phototriggers but with a much
slower rate, such as ∼9 × 106 s-1 for HA in 50% H2O/50%
MeCN (v/v),11 than the decay of the S1 state observed here.
This combined with the fact that a similarly small fluorescence
quantum yield and short singlet lifetime have been observed
for several closely related aromatic compounds with the para-
substitution with moieties other than the hydroxy group19,67 (for
example the methoxy1,2,11,67 group) where the proton is not
1
Since the triplet formation time is associated with the nπ*
lifetime decay (that is, ∼2 and 1 ps for the two compounds in
neat and H2O/MeCN mixed solvents, respectively), there must
be an additional process responsible for the much longer time
scale (∼7-12 ps) observed for the full development of the triplet
TR3 spectrum. By checking the early time spectra carefully and
as shown in Figure 4, we found that temporal growth of the
triplet band intensity is accompanied by a slight frequency
upshift (by ∼40 cm-1) and bandwidth narrowing (by ∼40 cm-1
)
for HPA and HPDP in both H2O/MeCN mixed and neat MeCN
solvents (for details of these changes in spectral profile, the
reader is referred to refs 60 and 61). Such an early time spectral
change in Raman band profile is a characteristic indication to
an excess energy relaxation process.60,61,68-73 It has been shown
that a typical time scale for this kind of relaxation is around 10
ps that matches well with the ∼7-12 ps time constant obtained
3
here. We therefore attributed the observed early stage ππ*
spectral evolution to combined dynamics involving growth of
the triplet population and subsequent relaxation of excess energy
(∼11 000 cm-1) introduced by 267 nm excitation and rapid ISC
(68) Ma, C.; Kwok, W. M.; Matousek, P.; Parker, A. W.; Phillips, D.; Toner,
W. T.; Towrie, M. J. Raman Spectrosc. 2001, 32, 115-123.
(69) Matousek, P.; Parker, A. W.; Towire, M.; Toner, W. T. J. Chem. Phys,
1997, 107, 9807-9817.
(70) Iwata, K.; Hamaguchi, H. J. Phys. Chem. A 1997, 101, 632-637.
(71) Hester, R. E.; Matousek, P.; Moore, J. N.; Parker, A. W.; Toner, W. T.;
Towire, M. Chem. Phys. Lett. 1993, 208, 471-478.
(72) Weaver, W. L.; Huston, L. A.; Iwata, K.; Gustafson, T. L. J. Phys. Chem.
1992, 96, 8956-8961.
(66) Shah, B. K.; Rodgers, M. A. J.; Neckers, D. C. J. Phys. Chem. A 2004,
108, ASAP.
(67) Wagner, P. J.; Kemppainen, A. E.; Schott, H. N. J. Am. Chem. Soc. 1973,
95, 5604-5614.
(73) Matousek, P.; Parker, A. W.; Towrie, M. J. Chem. Phys. 1997, 107, 9807-
9817.
9
1470 J. AM. CHEM. SOC. VOL. 127, NO. 5, 2005

Ultrafast Study of p-Hydroxphenacyl Ester Phototriggers
A R T I C L E S
1
3
conversion. Given the closeness of the nπ* and ππ* energy
levels, this implies, on the other hand, that the fluorescence
observed in the KTRF spectra are from the vibronically hot 1ππ*
electronic and structural properties of the triplet for the pHP
phototriggers in terms of the p-hydroxyphenacyl chromophore.60,61
Our calculations also revealed some structural and bonding
changes associated with the leaving group on going from the
ground to the triplet state. There is a slight lengthening of the
C-O bond connecting the pHP cage and leaving groups and a
significant variation in the relative orientation between the two
parts. However, the extent of these changes were found to be
not very sensitive to the kind of leaving group and rather similar
for HPA and HPDP (for a detailed description of the triplet
structure, the reader is referred to refs 60 and 61). This, on the
other hand, implies that, upon the triplet formation, further
processes are required to induce a favorable structural modifica-
tion and charge redistribution to consequently trigger the
cleavage and accompanying relevant rearrangement. The fact
that pHP photorelease occurs only in H2O containing solvent
indicates that H2O may act not only as solvent but also as a
reactant and/or catalyst that could be the source of the driving
force for the cleavage process.
1
and nπ* states consistent with the short singlet lifetimes pre-
venting full relaxation of the emitting states. Due to the efficient
ISC deactivation of the singlet state, relaxation of excess energy
on the triplet manifold has been observed generally for various
aromatic carbonyls in the solution phase, and our results here
are consistent with these other observations.74-77
In summary, we suggest the following nonradiative pathway
for the 267 nm photoexcitation of HPA and HPDP in solvent:
This scheme provides a consistent interpretation to the
dynamics and related solvent dependence observed in the present
KTRF and ps-TR3 experiments. Our results show clearly that
photophysical and photochemical processes of pHP phototrig-
gers occur in well-separated time scales. The photophysical
associated processes including photoexcitation, IC between the
singlet excited states, ISC from the singlet to triplet, relaxation
of excess energy, etc. take place in the femtosecond to early
picosecond time regime and finish within tens of picoseconds.
The deprotection related photochemistry is much slower and
proceeds from the fully relaxed triplet on the several hundred
picoseconds or nanosecond time scale depending on the kind
of leaving group.
Except for the photochemistry step(s) occurring subsequently
to the triplet generation in H2O containing solvents, we found
that the deactivation pathway shown in the scheme also applies
to several closely related aromatic carbonyls. However, provided
the complex dependence of fluorescence and ISC behavior on
many factors, such as the solvent properties,31-39 the temper-
ature,78 the electronic nature of the ring substituted groups,
etc.,28,29,67 further work is under way to understand the general
nonradiative channels for a selected diverse range of aromatic
carbonyls compounds.
D. Brief Discussion of Implications for the Deprotection
Reaction. The observation that the decay rate of the 3ππ* triplet
is strongly leaving group dependent and is larger for HPDP than
for HPA in H2O/MeCN mixed solvent appears to lend support
to the direct triplet heterolytic deprotection pathway proposed
by Givens, Wirz, and co-workers.11 This is because, as scaled
by the pKa value to describe the anion stability and leaving group
ability, diethyl phosphate has been found to be a better leaving
group than acetate.79 However, our recent studies show that the
triplet ps-TR3 spectra of HPA and HPDP in neat MeCN closely
resemble those of the “cage” HA and are barely leaving group
dependent.61 This has also been corroborated by theoretical
calculations and has been interpreted to be due to the analogous
Our KTRF results described in the preceding section indicate
the important influence of intermolecular H-bonding on the
radiative and nonradiative dynamics for the singlet excited states.
It is thus reasonable to consider possible H-bonding effects on
the triplet and subsequent deprotection reaction. Although the
general character of the triplet TR3 spectrum is comparable in
neat MeCN and H2O/MeCN mixed solvents, a careful examina-
tion of the spectra displayed in Figure 2 shows some small but
clear differences (such as relative intensity distribution of Raman
bands) in the two solvents that are likely associated with the
H-bonding interaction.11,12 But the observation that the triplet
spectra in H2O/MeCN mixed solvent are little different for HPA
and HPDP indicates the H-bonding modification of the triplet
in terms of the pHP chromophorm is similar for pHP photo-
triggers with different leaving groups and thus is unlikely to
fully account for the leaving group dependent deprotection
processes.
Plausible processes accounting for the water and leaving
group dependent dynamics quenching the 3ππ* triplet could be
(i) water assisted and/or catalyzed direct heterolysis or homolysis
followed by rapid electron transfer cleavage of the leaving
group, (ii) deprotonation and/or protonation of the triplet state.
Regarding the first possibility, we note that solvent-assisted
heterolysis cleavage of the leaving group, including the phos-
phate and acetate studied here, was postulated previously to
account for photocleavage of (7-methoxycoumarin-4-yl) methyl
(MCM) caged compounds;79 the photoheterolysis cleavage
mechanism has also been suggested for several other system
bearing the phosphate as a leaving group.80-82 In addition, a
very recent combined experimental and theoretical study by
Phillips and co-workers demonstrates that water-catalyzed
solvation of a leaving group into ions is the driving force
responsible for photodecomposition of isopolyhalomethanes in
water solvated environments.83 With regards to the second
possibility, it is essential to emphasize the well-established fact
that phenols become stronger acids upon photoexcitation, while
(74) (a) Turek, A. M.; Krishnamoorthy, G.; Phipps, K.; Saltiel, J. J. Phys. Chem.
A 2002, 106, 6044-6052. (b) Sun, Y.-P.; Sears, D. F., Jr.; Saltiel, J. J.
Am. Chem. Soc. 1989, 111, 706-711.
(75) Wolf, M. W.; Legg, K. D.; Brown, R. E.; Singer, L. A.; Parks, J. H. J. Am.
Chem. Soc. 1975, 97, 4490-4497.
(76) Kiritani, M.; Yoshii, T.; Hirota, N.; Baba, M. J. Phys. Chem. 1994, 98,
11265-11268.
(80) Newcomb, M.; Horner, J. H.; Whitted, P. O.; Crich, D.; Huang, X.; Yao,
Q.; Zipse, H. J. Am. Chem. Soc. 1999, 121, 10685-10694.
(81) Whitted, P. O.; Horner, J. H.; Newcomb, M.; Huang, X.; Crich, D. Org.
Lett. 1999, 1, 153-156.
(82) Horner, J. H.; Taxil, E.; Newcomb, M. J. Am. Chem. Soc. 2002, 124, 5402-
5410.
(77) Sneh, O.; Cheshnovsky, O. J. Phys. Chem. 1991, 95, 7154-7164.
(78) Jones, P. F.; Calloway, A. R. J. Am. Chem. Soc. 1970, 92, 4997-4998.
(79) Schade, B.; Hagen, V.; Schmidt, R.; Herbrich, R.; Krause, E.; Eckardt, T.;
Bendig, J. J. Org. Chem. 1999, 64, 9109-9117.
9
J. AM. CHEM. SOC. VOL. 127, NO. 5, 2005 1471

A R T I C L E S
Ma et al.
the 3ππ* triplet of aromatic carbonyls become stronger
bases.11,12,84-86 It is therefore reasonable to deduce that, for pHP
compounds with the two functional groups (hydroxy and
carbonyl) on the same chromophore, protonation at the carbonyl
oxygen and deprotonation at the hydroxy site is highly likely
to occur after photoexcitation.12,86 Such processes could formally
manifest themselves as the so-called ESIPT suggested by Wan
and co-workers12 but on the triplet rather than the singlet
manifold because of the rapid ISC dynamics as demonstrated
in our present study. Indeed, the deprotonation related reaction
has been reported for the HA cage by Wirz and co-workers11
and proposed for HPA by Wan and co-workers.12 However,
we note that there remains some uncertainty about the explicit
correlation between the deprotonation and cleavage reaction. It
appears that the deprotonation process displays little leaving
group dependence and thus could function as a competing
process with the cleavage reaction to deactivate the triplet state.
It is therefore reasonable to suggest that the overall photochemi-
cal pathway and deprotection efficiency for a specific pHP
phototrigger could be mostly controlled by a dynamic competi-
tion between the heterolysis cleavage and deprotonation process.
In this sense, the leaving group ability and therefore the cleavage
rate and rate of the deprotection would be the major factors in
determining the pHP photochemistry. Further TR3 work on the
picosecond to nanosecond time scale is in progress to more fully
address the cleavage mechanism and relevant rearrangement
reaction.
rate of internal conversion relative to ISC to deactivate the 1nπ*.
This increased IC is likely caused by the substantially reduced
1
1nπ* and ππ* gap in the H2O/MeCN mixed solvent than in
neat MeCN solvent (the proximity effect). Spectral evolution
3
of the ππ* triplet Raman bands on the early tens picosecond
time scale was attributed to the combined dynamics of the
development of the triplet population and relaxation of the
excess energy induced by the S3 excitation and rapid ISC
conversion. In H2O/MeCN mixed solvent, the triplet lifetime
was found to be ∼420 and 2130 ps for HPDP and HPA,
respectively, substantially shorter than that in neat MeCN (150
and 137 ns for HPDP and HPA, respectively). This leaving
3
group and solvent-dependent dynamic quenching of the ππ*
triplet is a clear indication for occurrence of the deprotection
related photochemistry. The observations here provide direct
experimental evidence that the photochemical step(s) takes place
from the fully relaxed triplet and thus rules out the singlet
mechanism for HPA and HPDP after 267 nm excitation. The
results show clearly that the photophysics and photochemistry
of pHP caged compounds proceed on a well-separated time
scale. The photophysical steps occur on the femtosecond to tens
picosecond time regime and are independent of the existence
and kind of leaving group, whereas the deprotection related steps
happen on the several hundreds picosecond to nanosecond and
later time scales. The strong dependence of the deprotection
on the leaving group ability suggests a heterolysis cleavage
mechanism may be involved in the deprotection mechanism.
Acknowledgment. This research was done in the Ultrafast
Laser Facility at the University of Hong Kong and supported
by grants from the Research Grants Council of Hong Kong
(HKU/7108/02P) and (HKU 1/01C) to D.L.P. and (HKU 7112/
02P) and (HKU 7027/03P) to P.H.T.. W.M.K. thanks the
University of Hong Kong for the award of a Postdoctoral
Fellowship.
Conclusion
By using fs-KTRF spectroscopy to monitor the singlet excited
states and ps-KTR3 to detect the triplet state, the present
investigation reports a combined ultrafast time-resolved study
on photophysical processes of two pHP caged compounds HPDP
and HPA in neat MeCN and H2O/MeCN mixed solvent. These
results show that 267 nm excitation leads to instantaneous
Supporting Information Available: Details for the synthesis
and characterization of HPA and HDPD. Details for using log-
normal function to simulate the experimentally measured KTRF
spectra including (i) estimation of intensity ratio of the 1nπ* to
1ππ* fluorescence band and (ii) estimation of the overall
fluorescence quantum yield. Additional spectral parameters
obtained from the log-normal simulation of the fluorescence
and absorption spectra are listed in Table S1. Fluorescence
spectra of the 1nπ* and 1ππ* state calculated based on the KTRF
spectra for HPA in neat MeCN with 267 nm excitation is
displayed in Figure S1. Figure S2 shows a comparison of the
normalized absorption spectra for HA, HPA, and HPDP in neat
MeCN solvent. Figure S3 shows ps-KTR3 spectra of HPA in
neat MeCN and H2O/MeCN (1:1 by volume) mixed solvent
obtained with 267 nm excitation and 400 nm probe wavelength.
Figure S4 shows comparison of early time ps-KTR3 spectra of
HPA in H2O/MeCN (1:1 by volume) mixed solvent obtained
with 267 nm excitation, 400 and 342 nm probe wavelength.
This material is available free of charge via the Internet at
http://pubs.acs.org.
1
population of the ππ* S3 state that has a lifetime of ∼80 fs
1
and is deactivated by internal conversion to the nπ* S1 state.
The 1nπ* lifetime was found to be ∼1 and 2 ps in H2O/MeCM
mixed and neat MeCN, respectively. Spectral properties of the
1ππ* and 1nπ* fluorescence transitions are strongly influenced
by the solvent H-bonding strength indicating the importance of
intermolecular H-bonding interaction for the excited molecules.
Dynamic correlation between the singlet decay and triplet
formation indicates a direct 1nπ* f 3ππ* ISC mechanism with
the ISC rate estimated to be ∼5 × 1011 s-1 in both solvents.
The shorter 1nπ* lifetime in the H2O/MeCN mixed solvent than
in neat MeCN was interpreted as being due to the increased
(83) (a) Kwok, W. M.; Zhao, C.; Li, Y.-L.; Guan, X.; Phillips, D. L. J. Chem.
Phys. 2004, 120, 3323-3332. (b) Kwok, W. M.; Zhao, C.; Li, Y.-L.; Guan,
X.; Wang, D.; Phillips, D. L. J. Am. Chem. Soc. 2004, 126, 3119-3131.
(c) Kwok, W. M.; Zhao, C.; Guan, X.; Li, Y.-L.; Du, Y.; Phillips, D. L. J.
Chem. Phys. 2004, 120, 9017-9032.
(84) (a) Canonica, S.; Hellrung, B.; Wirz, J. J. Phys. Chem. A 2000, 104, 1226-
1232. (b) Ramseier, M.; Senn, P.; Wirz, J. J. Phys. Chem. A 2003, 107,
3305-3315.
(85) Leigh, W. J.; Lathioor, E. C.; St. Pierre, M. J. J. Am. Chem. Soc. 1996,
118, 12339-12348.
(86) Das, P. K.; Encinas, M. V.; Scaiano, J. C. J. Am. Chem. Soc. 1981, 103,
4154-4162.
JA0458524
9
1472 J. AM. CHEM. SOC. VOL. 127, NO. 5, 2005