Photolytic cleavage of 1-(2-nitrophenyl)ethyl ethers involves two parallel pathways and product release is rate-limited by decomposition of a common hemiacetal intermediate

Article abstract of DOI:10.1021/ja034354c

Time-resolved FTIR spectroscopic studies of the flash photolysis of several 1-(2-nitrophenyl)-ethyl ethers derived from aliphatic alcohols showed that a long-lived hemiacetal intermediate was formed during the reaction. Breakdown of this intermediate was

Full text of DOI:10.1021/ja034354c

Published on Web 06/18/2003
Photolytic Cleavage of 1-(2-Nitrophenyl)ethyl Ethers Involves
Two Parallel Pathways and Product Release Is Rate-Limited
by Decomposition of a Common Hemiacetal Intermediate
John E. T. Corrie,*,† Andreas Barth, V. Ranjit N. Munasinghe,
§,#
†
†
‡
David R. Trentham, and Michael C. Hutter
Contribution from the National Institute for Medical Research, The Ridgeway,
Mill Hill, London NW7 1AA, U.K., Institut f u¨ r Biophysik, Johann Wolfgang Goethe-UniVersit a¨ t,
Theodor Stern Kai 7, Haus 74, D-60590 Frankfurt, Germany, and Max-Planck Institut f u¨ r
Biophysik, Kennedyallee 70, D-60590 Frankfurt, Germany
Received January 27, 2003; E-mail: jcorrie@nimr.mrc.ac.uk
Abstract: Time-resolved FTIR spectroscopic studies of the flash photolysis of several 1-(2-nitrophenyl)-
ethyl ethers derived from aliphatic alcohols showed that a long-lived hemiacetal intermediate was formed
during the reaction. Breakdown of this intermediate was rate-limiting for product release. One of these
compounds (methyl 2-[1-(2-nitrophenyl)ethoxy]ethyl phosphate, 9) was studied in detail by a combination
of time-resolved FTIR and UV-vis spectroscopy. In addition, product studies confirmed clean photolytic
decomposition to the expected alcohol, 2-hydroxyethyl methyl phosphate, and the 2-nitrosoacetophenone
-
1
byproduct. At pH 7.0, 1 °C, the rate constant for product release was 0.11 s , very much slower than the
-
1
5
020 s rate constant for decay of the photochemically generated aci-nitro intermediate (pH 7.0, 2 °C).
Time-resolved UV-vis measurements showed that the hemiacetal intermediate is formed by two competing
pathways, with fast (∼80% of the reaction flux) and slow (∼20% of the flux) components. Only the minor,
slower path is responsible for the observed aci-nitro decay process. These competing reactions are
interpreted with the aid of semiempirical PM3 calculations of reaction barriers. Furthermore, AMSOL
a
calculations indicate that the pK of the nitronic acid isomer formed by photolysis is likely to determine
partition into the alternate paths. These unusual results appear to be general for 1-(2-nitrophenyl)ethyl
ethers and contrast with a related 2-nitrobenzyl ether that photolyzed without involvement of a long-lived
hemiacetal.
Introduction
Scheme 1
Flash photolytic release of bioactive molecular species by
rapid cleavage of photolabile protecting groups from inactive
precursors (colloquially termed caged compounds) is a valuable
1
technique for time-resolved studies of biological processes. The
most common photochemistry in the field utilizes the well-
known 2-nitrobenzyl rearrangement, in which a nitrobenzyl
group, or substituted version thereof, is converted to a 2-ni-
trosoarylcarbonyl compound, with release of the species to
which it was attached. In the archetypal example of caged ATP
1
, the dark reactions after the photolytic event have been
interpreted as in Scheme 1, with decay of the aci-nitro
intermediate 2 having been shown to be the rate-limiting step
†
National Institute for Medical Research.
Johann Wolfgang Goethe-Universit a¨ t.
Present address: Department of Biochemistry and Biophysics, Arrhe-
§
#
nius Laboratories for Natural Sciences, Stockholm University, SE-106 91
Stockholm, Sweden.
‡
Max-Planck Institut f u¨ r Biophysik.
(
1) (a) Corrie, J. E. T.; Trentham, D. R. In Bioorganic Photochemistry;
Morrison, H., Ed.; Wiley: New York, 1993; Vol. 2, pp 243-305. (b)
Adams, S. R.; Tsien, R. Y. Annu. ReV. Physiol. 1993, 55, 755. (c) Kaplan,
J. H. Annu. ReV. Physiol. 1990, 52, 897. (d) Marriott, G., Ed. Methods in
Enzymology; Academic Press: San Diego, 1998; Vol. 291. (e) Pelliccioli,
A. P.; Wirz, J. Photochem. Photobiol. Sci. 2002, 1, 441.
for product release.2,3 The reaction proceeds via rapid equilibra-
tion of the conjugate acid of 2 with the bicyclic species 3,
collapse of which generates the final products, namely, ATP
8546
9
J. AM. CHEM. SOC. 2003, 125, 8546-8554
10.1021/ja034354c CCC: $25.00 © 2003 American Chemical Society

Photolytic Cleavage of 1-(2-Nitrophenyl)ethyl Ethers
A R T I C L E S
Figure 1. IR difference spectra for photolysis of 9 at pH 7.0, 1 °C in 200 mM MOPS buffer. The traces represent spectra averaged in the following time
windows after the light flash: (s) 0.01-1.16 s, (- - -) 2.5-3.2 s, (‚‚‚) 3.2-11.0 s, (- ‚ -) 68.1-126.4 s. The inset shows a single-exponential fit to the
-
1
integrated area of the 1688 cm peak with respect to time.
and the nitrosoketone 4. The protonation state shown for 3 takes
account of recent studies, so differs from a previous proposal
aci-nitro intermediate formed on photolysis has been taken as
evidence for rapid release of choline. The result was proposed
to make 7 and its arsenic analogue suitable precursors for time-
4
2
in which 3 was suggested to be formed as its conjugate base,
i.e., by cyclization of the aci-nitro anion 2 without uptake of a
proton. In the case of caged ATP, an alternative collapse of 3
via a further hemiacetal species 5 was previously excluded, at
least insofar that the hemiacetal did not accumulate to a
6
resolved X-ray crystallographic studies.
2
measurable extent. In light of the results reported below, 5 may
2
in fact be in the pathway, but previous work implies that it
would break down immediately.
As part of our use of time-resolved IR spectroscopy initiated
by flash photolysis to elucidate dark reactions in a range of
3
2
-nitrobenzyl photocleavage processes, we studied 6, the 1-(2-
nitrophenyl)ethyl ether of glycolic acid, because we had
observed photodecarboxylation of several related compounds
Following the initial observations on 6, we studied the
photolysis of ethers 7-9 and found that all had similar behavior,
with decay of a putative hemiacetal intermediate being the rate-
determining step for release of the various alcohol products.
Of these compounds, 9 was investigated in greatest detail, as it
was a readily accessed water-soluble species that was expected
to avoid unwanted side reactions such as the minor decarboxyl-
ation seen for 6. The compounds 6-8 are discussed only where
they give rise to specific points, particularly in the rate of
breakdown of the hemiacetal intermediate. In contrast to these
results for 6-9, the related ether 10 that lacks the benzylic
methyl substituent photolyzed with no evidence of a long-lived
intermediate. These results and their implications are described
below, together with results of semiempirical molecular orbital
calculations that shed light on the mechanistic processes
involved.
(A. Barth and J. E. T. Corrie, unpublished data). In fact, time-
resolved IR difference spectra recorded upon photolysis of 6
showed that photodecarboxylation was only a very minor
process for this compound. Of much greater interest in the
spectrum were changes of several strong transient bands that
appeared to be consistent with the formation and subsequent
decay of a relatively long-lived hemiacetal intermediate. At pH
7
, breakdown of this intermediate to release the caged species
(in this case the glycolate anion) was 4 orders of magnitude
slower than decay of the upstream aci-nitro intermediate. This
observation was particularly significant, as aci-nitro decay rates
have often been used to estimate product release rates, an
assumption based on the case of caged ATP, where product
release has unequivocally been shown to be synchronous with
2
,3
aci-nitro decay. Whether an aci-nitro decay rate necessarily
provides a reliable measure of the rate of product release is an
issue of concern in this field.1 In the case of a caged choline
Results and Discussion
d,5
It was evident from time-resolved IR difference spectra for
flash photolysis of 9 at pH 7.0, 1 °C (Figure 1) that slow
processes were involved in the overall reaction. In these spectra,
negative bands represent absorption by species that disappear
from the solution as a result of photolysis, while positive bands
7
, an ether derivative of the same type as 6, rapid decay of the
(
2) Walker, J. W.; Reid, G. P.; McCray, J. A.; Trentham, D. R. J. Am. Chem.
Soc. 1988, 110, 7170.
(
3) Barth, A.; Corrie, J. E. T.; Gradwell, M. J.; Maeda, Y.; M a¨ ntele, W.; Meier,
T.; Trentham, D. R. J. Am. Chem. Soc. 1997, 119, 4149.
4) (a) Il’ichev, Y. V.; Wirz, J. J. Phys. Chem. A 2000, 104, 7856. (b) Schw o¨ rer,
M.; Wirz, J. HelV. Chim. Acta 2001, 84, 1441.
(
(
5) McCray, J. A.; Trentham, D. R. Annu. ReV. Biophys. Biophys. Chem. 1989,
(6) (a) Peng, L.; Goeldner, M. J. Org. Chem. 1996, 61, 185. (b) Peng, L.;
Nachon, F.; Wirz, J.; Goeldner, M. Angew. Chem., Int. Ed. 1998, 37, 2691.
1
8, 239.
J. AM. CHEM. SOC.
9
VOL. 125, NO. 28, 2003 8547

A R T I C L E S
Corrie et al.
arise from species produced in the reaction. Notably, the
spectrum recorded in the 0.01-1.16 s time window after
photolysis showed minimal carbonyl absorption but had an
Table 1. Rate Constants for Processes Initiated by Flash
Photolysis of 9
rate constant (s-1)
-
1
aci-nitro decaya
(406 nm)
carbonyl formationb
nitroso dimerizationb,c
(1497 cm-1
intense band at 1497 cm that was assigned to a nitroso
monomer.3 A carbonyl band gradually appeared at 1688 cm
,7
-1
pH
(1688 cm-1
)
)
and reached its maximum intensity over the following ∼100 s,
while the nitroso monomer band decayed to zero on the same
time scale and was replaced by bands characteristic of nitroso
6.0
n.d.
0.01
0.11
n.d.
0.79
n.d.
5
n.d.
>50
n.d.
0.02
0.12
n.d.
t1/2 1.9 s
n.d.
t1/2 1.0 s
n.d.
t1/2 1.6 s
n.d.
7
7
8
.0
.5
.0
5020
2600
1490
650
225
246
-
1
-1
dimers at 1424 and 1382 cm (cis-dimer) and 1267 cm
8.5
9.0
3
,7
(trans-dimer).
9.5
These spectroscopic features, particularly the strong nitroso
1
1
0.0
0.5
190
138
monomer band and the absence of a carbonyl band in spectra
taken soon after the flash, are most readily compatible with the
hemiacetal 11 being a relatively long-lived intermediate.
Breakdown of 11 would release as products the alcohol 12 and
nitrosoketone 4, the latter giving rise to the observed carbonyl
band. Decay rates of the aci-nitro intermediate in photolysis of
a
Measured at 2 °C in 20 mM MES, MOPS, EPPS, CHES, or CAPS
buffers plus 150 mM NaCl (see Experimental Section for details).
b
Measured at 1 °C in 200 mM MES, MOPS, EPPS, CHES, or CAPS buffers
(see Experimental Section for details). c Data at pH 8, 9, and 10 did not fit
to single exponentials and are expressed as half-times for the overall
amplitude change.
9
(and the related ethers 6-8) were always too fast in the pH
range 6-10 to allow this species to be seen in our FTIR spectra,
in which 11 was the first observable intermediate for these
compounds. This contrasts with previous results for caged ATP,
not undergo significant dimerization, presumably because of
hindrance by the bulky ortho-substituent; so dimer formation
proceeds only as the nitrosoketone 4 is released. However, the
kinetics of nitroso dimerization were not the main focus of this
work and were not studied in detail.
To substantiate the formation of 11 and establish an overall
reaction mechanism, time-resolved UV and visible recordings
initiated by flash photolysis of 9 were made. Recordings of the
aci-nitro decay at 406 nm over the pH range 7.0-10.5 and at
3
,8
where decay of the aci-nitro intermediate was much slower.
To the best of our knowledge, the only related report of a
verified “stable” hemiacetal-like intermediate from photolysis
of 2-nitrobenzyl compounds is from N-(2-nitrobenzyl)-1-naph-
thalenecarboxamide, where the aminol 13 was reported as an
isolable compound that underwent slow decomposition to
9
1
-naphthalenecarboxamide and 2-nitrosobenzaldehyde.
2
°C (Table 1) showed clean, single-exponential processes in
all cases (e.g., Figure 2c), and the amplitude of the signal was
constant across this pH range. The rate constant at pH 7.0, 2
-
1
-1
°
7
C (5020 s ) is consistent with the value of 6300 s at pH
6
a
.2, 20 °C reported for 7 by Peng and Goeldner (see below
for comment on the buffer dependence of this rate constant).
1a,2
As with other 1-(2-nitrophenyl)ethyl derivatives, the aci-nitro
decay became slower as the pH increased.
Further measurements to monitor the kinetics of nitroso
monomer formation (at 740 nm) and at 577 nm in the presence
of a pH indicator (Cresol Purple) to follow proton transients
were made at pH 8.5, where the rate constants were in a
convenient range. At 740 nm, we had expected to observe an
exponential increase in absorbance at a rate that mirrored decay
of the aci-nitro absorbance at 406 nm, as reported for photolysis
Time-resolved FTIR measurements for 9 over the pH range
-10 gave spectra qualitatively similar to those recorded at pH
(Figure 1), except that the rates of change of various bands
6
7
were pH-dependent. Rate constants for the formation of ketone
, and by inference for release of the alcohol product, were
2
of caged ATP 1. In practice, the 740 nm data for 9 upon flash
4
photolysis (Figure 2a) showed a large step increase in absorption
determined by integration of the area of the carbonyl band as a
function of time (Table 1). Formation of the carbonyl band
became faster at higher pH values; that is, breakdown of the
hemiacetal intermediate 11 was base-catalyzed (see below).
Similar integration of the nitroso monomer band gave informa-
tion on its rate of disappearance as a consequence of dimer-
ization (Table 1). This process at pH 6 and 7 had the same rate
constant (within experimental error) as that for carbonyl
formation; that is, the dimerization was rate-limited by decom-
position of the hemiacetal 11. The kinetics became more
complex at higher pH values and lagged behind the carbonyl
formation. The data suggest that the nitroso group of 11 does
5
-1
(
i.e., a rate constant >10 s , which was the resolution limit
of our apparatus) followed by an exponential increase with rate
-1
constant 650 s , which matched the aci-nitro decay at this pH
av ma l pu el it (u Td ae b. le 1). The slow phase comprised ∼20% of the total
Measurements of proton transients following flash photolysis
were made in weakly buffered solutions for 9 and caged ATP
2,10
at pH 8.5, 2 °C (Figure 2b) by a method previously described.
5
As expected at this pH, caged ATP showed only a rapid (>10
-1
s ) decrease in pH from the proton released during ionization
of the initially formed nitronic acid (Scheme 1). In contrast, 9
showed rapid proton release, with the amplitude of the initial
acidification step being ∼15% of that for caged ATP under the
same conditions, followed by an exponential return to the initial
(
(
(
7) Colthup, N. B.; Daly, L. H.; Wiberley, S. E. Introduction to Infrared and
Raman Spectroscopy, 3rd ed.; Academic Press: San Diego, 1990.
8) Barth, A.; Hauser, K.; M a¨ ntele, W.; Corrie, J. E. T.; Trentham, D. R. J.
Am. Chem. Soc. 1995, 117, 10311.
9) Peyser, J. R.; Fletchner, T. W. J. Org. Chem. 1987, 52, 4645.
(10) Papageorgiou, G.; Corrie, J. E. T. Tetrahedron 1997, 53, 3917.
8548 J. AM. CHEM. SOC.
9
VOL. 125, NO. 28, 2003

Photolytic Cleavage of 1-(2-Nitrophenyl)ethyl Ethers
A R T I C L E S
Figure 3. IR difference spectra for photolysis of 10 at pH 8.5, 1 °C in 200
mM Bicine buffer. The traces represent spectra averaged in the following
time windows after the light flash: (s) 0.01-0.185 s, (- - -) 0.196-0.51 s,
(‚‚‚) 3.2-29.0 s.
buffer (Bicine, pH 8.5) that showed no change of buffer
protonation state during or after the photolytic reaction (data
not shown). In contrast, similar FTIR spectra recorded for
compounds that undergo slower transient or permanent proton
release upon photolysis show large difference bands at 1621
-1
and 1572 cm for Bicine protonation (for examples of transient
and permanent changes, see Figure 3 and ref 11, respectively).
The proton transient recorded for 9 in the indicator study above
is too rapid to observe in our FTIR spectra. The absence of a
net pH change upon photolysis of 9 is expected, since the
reaction stoichiometry predicts that the proton released upon
ionization of the photolytically formed nitronic acid ultimately
returns to the product alcohol 12. However, the small amplitude
of the pH transient observed during the photolysis in comparison
with that for caged ATP 1 (Figure 2b) is noteworthy, since the
two compounds have similar quantum yields (0.63 for caged
2
ATP and 0.54 measured for 9; see Supporting Information)
and an identical near-UV chromophore. Despite this similarity
of quantum yields, the amplitude of the 406 nm aci-nitro
transient for 9 was ∼18% of that for caged ATP (Figure 2c),
close to the relative amplitudes of the proton transients from
the two compounds (Figure 2b). These 406 nm measurements
were made at different pH values for the two compounds to
bring their aci-nitro decay processes into a similar time frame
Figure 2. Transients for flash photolysis of caged ATP 1 and the ether 9.
(a) Nitroso monomer formation: measured at 740 nm for a 3 mM solution
of 9 in 20 mM EPPS, 150 mM NaCl, pH 8.5 at 2 °C. (b) Proton transients:
measured at 577 nm in the presence of Cresol Purple indicator for 0.5 mM
solutions of 1 or 9 in 1 mM EPPS, 150 mM NaCl, pH 8.5, 2 °C. (c) aci-
Nitro transients: measured at 406 nm, 2 °C for 0.5 mM solutions of 1 (226
-
1
-1
s
in 25 mM MES, pH 6.5) or 9 (688 s in 25 mM EPPS, pH 8.5). The
arrows on each panel indicate the time of the laser flash. Perturbations in
the records that start ∼0.8 ms prior to the laser flash arise from the flash
lamp that initiates the laser pulse. The clear lines superimposed on the traces
in each panel represent single-exponential fits to relevant sections of the
data.
(see caption to Figure 2c).
The 740 nm measurement described above gives direct
evidence of processes occurring on two very different time
scales, with a fast process accounting for the majority of the
reaction flux. The relative amplitudes of the comparative
measurements of proton stoichiometry and aci-nitro decay for
caged ATP and 9 agree well with those of the fast and slow
phases of the 740 nm data. In light of these congruent but
unexpected results, we were concerned that formation of
product(s) other than those of the “normal” 2-nitrobenzyl
photochemistry, as exemplified in Scheme 1, might be respon-
sible for the large-amplitude fast phase of the 740 nm transient
recorded for the photoreaction of 9. Evidence that this was not
the case came from two independent sources. First, clean
formation of the expected product (2-hydroxyethyl methyl
-
1
pH (rate constant 280 s ). Essentially the same rate constant
-1
(
250 s ) was measured for the aci-nitro decay of 9 under these
solution conditions but in the absence of the pH indicator (data
not shown). It should be noted that there are significant buffer
effects on the kinetics of this aci-nitro decay. Measurements at
pH 8.5, 2 °C in four solutions that were used in different parts
of this work (25 mM Na EPPS, 20 mM Na EPPS + 150 mM
NaCl, 1 mM Na EPPS + 150 mM NaCl, and 10 mM Na
pyrophosphate + 100 mM NaCl) gave rate constants of 688,
-
1
6
50, 250, and 1040 s , respectively. These differences appear
to be mediated more by the particular buffer conditions than
by nonspecific ionic strength effects (cf. ref 5 for an example
of the ionic strength dependence for caged ATP) but were not
investigated further.
Return of the pH to its initial value in the photolytic reaction
of 9 was consistent with FTIR spectra recorded in a carboxylate
1
phosphate 12) was confirmed by H NMR spectroscopy of a
partially photolyzed solution of 9 (∼40% conversion). This
NMR spectrum was compared with that of a solution that
contained a 1:1 mixture of unphotolyzed 9 and authentic 12.
(11) Barth, A.; Corrie, J. E. T. Biophys. J. 2002, 83, 2864.
J. AM. CHEM. SOC.
9
VOL. 125, NO. 28, 2003 8549

A R T I C L E S
Corrie et al.
Table 2. Rate Constants for Decomposition of the Hemiacetal
Intermediate 11 Formed upon Photolysis of 6-9
Scheme 2 . Proposed Two-Path Mechanism for the Dark
Reactions Subsequent to Flash Photolysis of 6-9a
rate constant (s-
1 a
)
pH
6
7
8
9
6
.0
.0
.0
.0
0.11
0.21
0.90
7.0
n.d.
1.1
n.d.
n.d.
0.26
0.35
0.45
0.70
0.012
0.11
0.79
5.0
7
8
9
a
Measured at 1 °C in 200 mM MES, MOPS, EPPS, or CHES buffers
-
1
by monitoring the formation of the carbonyl band at 1688 cm (see
Experimental Section for details).
The latter spectrum showed characteristic, clearly resolved
signals for each species, so was appropriate for this product
1
analysis. The H NMR spectrum of the partially photolyzed
sample of 9 essentially showed signals only for the starting
material and 12 together with the normal byproduct, 2-ni-
trosoacetophenone 4 (see Supporting Information, Figures S1
and S2). Signals for the latter were weak, probably because of
substantial precipitation of this water-insoluble compound
between the time of photolysis and recording the NMR
spectrum. Stronger evidence for predominant formation of 4
a
The arrows representing electron movements on structure 14 refer to
its direct route to 11, while those on 16 refer to the conversion into 17.
(in this case from compound 6, which had a biphasic 740 nm
transient upon photolysis very similar to that of 9) came from
FTIR spectroscopy at pH 6, 35 °C in the presence of dithio-
threitol. We have previously described the sequence of reactions
and consequent IR spectroscopic changes of 4 under these
conditions when generated by photolysis of caged ATP (i.e.,
initial rapid formation of 2-hydroxylaminoacetophenone, fol-
and its substituted analogues. The proposed partition for decay
of the nitronic acid 14, which is generally regarded as the
product of the photolytic step in 2-nitrobenzyl photochemistry,
is introduced to rationalize biphasic formation of the nitroso
monomer (Figure 2a) and the low amplitudes of the proton and
aci-nitro transients compared with corresponding signals from
caged ATP (Figures 2b and 2c, respectively). The minor
pathway for 9 proceeds via 15 and equilibration with the
isomeric nitronic acid 16, which further undergoes conversion
to the bicyclic species 17 and subsequent breakdown to the
hemiacetal 11. This path accounts for 15-20% of the reaction
flux for 9, as shown by the amplitudes of the aci-nitro and pH
transients in comparison with those of caged ATP and by the
amplitude of the slow phase of nitroso formation. This minor
path, at least up to 17, is analogous to the well-established
mechanism shown in Scheme 1 for caged ATP and represents
the exponential processes shown for 9 in Figure 2a-c. The
major pathway of Scheme 2 is suggested to proceed by an
electrocyclic rearrangement of 14 to 11 and is represented by
the sub-millisecond transient in Figure 2a, so comprises ∼80%
of the reaction flux. Because ionization of the initial nitronic
acid that leads into the “normal” pathway (i.e., as for caged
ATP) via 15 is rapid (within the time scale of the laser flash
3
lowed more slowly by dehydration to 3-methylanthranil).
Exactly the same sequence of spectra was observed upon
photolysis of 6 (data not shown). Thus the product studies
indicate that 9 undergoes photolysis to the expected products 4
and 12 with no evidence of additional products being formed
and indicate that both the fast and slow 740 nm transients
represent processes on pathways to 4 and 12.
Compounds 6-8 each gave FTIR spectra upon photolysis
with features similar to those in Figure 1, suggesting that the
unusual reactivity of 9 is common to 1-(2-nitrophenyl)ethyl
ethers of aliphatic alcohols. As mentioned above, 6 had the same
biphasic growth of absorption at 740 nm, indicating that a dual
reaction pathway also operates for that compound, and by
inference for 7 and 8 as well. The main difference detected
among the various compounds was in the rate constants for
release of the respective alcohol products. These data (Table 2)
show that the rates of breakdown of the hemiacetal intermediates
are modulated by the terminal substituent on the ethoxy group
and are compatible with base-catalyzed breakdown of a hemi-
acetal intermediate. The most notable variation is for 8, where
rates at pH 6 and 7 were similar to those for 6 and 9. However
by pH 8, where the terminal dibasic phosphate is fully ionized,
the rate began to slow and was less strongly catalyzed by
increasing alkalinity of the solution.
4
b
for solutions containing buffer salts ), entry to the competing,
major pathway must necessarily be on the same time scale.
Remarkably, the unsubstituted 2-nitrobenzyl ether 10, i.e.,
the analogue of 6 that lacks the R-methyl group, had quite
different reactivity from that of 6-9. Its aci-nitro decay rate
-1
(
515 s at pH 7.0, 20 °C) was much slower than for its methyl-
Scheme 2 shows a branched mechanism for the dark reactions
after photolysis to interpret these time-resolved IR and UV-
vis data for 9 and related ethers. In its final step, it shows base-
catalyzed breakdown of the hemiacetal 11 in accord with the
data of Table 1. Base catalysis is in line with studies by
McClelland et al.12 on hydrolysis of hemiacetals of acetophenone
-1
substituted analogue 6 (5690 s at pH 7.5, 17 °C). Measurement
at 740 nm to monitor nitroso formation showed a single-
-1
exponential rise in absorption with a rate constant of 590 s
(pH 7, 20 °C; data not shown), comparable to that of the aci-
nitro decay process. This trace had no evidence of a sub-
millisecond component as seen for 9 (Figure 2a). The slower
decay kinetics in this case made it possible to observe the aci-
nitro intermediate in the time-resolved IR spectrum (Figure 3;
(
12) McClelland, R. A.; Engell, K. M.; Larsen, T. S.; Sørensen, P. S. J. Chem.
Soc., Perkin Trans. 2 1994, 2199.
8550 J. AM. CHEM. SOC.
9
VOL. 125, NO. 28, 2003

Photolytic Cleavage of 1-(2-Nitrophenyl)ethyl Ethers
A R T I C L E S
Table 3. PM3 Computed Reaction Barriers (kJ mol-1) for
Bicine, pH 8.5, 1 °C). Positive bands at 1464, 1319, and 1176
-
1
Processes in Scheme 2
cm are close to positions of bands at 1465, 1330, and 1179
-1
reaction step
cm that were assigned in the aci-nitro spectrum of caged ATP
8
compounda
17 f 11b
to vibrations involving the nitronate group, and a sharp band
14 f 11
16 f 17
-
1
at 1027 cm probably arises from the enol ether moiety. The
spectrum also showed strong difference bands for buffer
6
81.6
118.5
113.9
112.2
113.5
118.9
110.1
119.7
98.4
62.4
102.2
113.9
105.9
98.0
102.2
93.8
110.6
93.0
106.3
105.1
104.3
130.2
136.9
147.0
135.7
131.1
134.4
136.1
132.3
134.8
137.3
142.4
139.0
6
7
7
+ H+
-
1
-1
protonation (negative at 1572 cm and positive at 1621 cm )
that return to baseline as this intermediate decays to the final
products; that is, the reaction has no net proton release since,
as discussed above for the related compound 9, the proton
initially released from the benzylic position is eventually taken
up on the alcohol product.
-
+ Cl
+
8 + H
+
8
9
9
+ 2H
+ H+
10
+
1
2
0 + H
0
119.7
118.5
122.7
Two further points are relevant to the IR spectra observed
upon photolysis of 10. First, apart from the negative bands that
correspond to disappearance of the original nitro group, the
spectrum at the end of the dark processes is almost featureless
21
a
Compounds 6, 9, and 10 were treated as the monoanions or their
conjugate acids, compound 8 as its mono- or diprotonated form, and
b
compound 7 as the free cation or as an ion pair with a chloride ion. Via
(
9
Figure 3), in contrast to the corresponding final spectrum for
. The lack of significant IR absorptions in this case can be
water-mediated proton exchange. The reaction barrier is expected to be
considerably lower in water.
attributed to hydration of the 2-nitrosobenzaldehyde byproduct.
Electron-withdrawing substituents are known to promote hydra-
tion of aromatic aldehydes,13 and in the present case hydration
would additionally be favored by formation of the cyclic
tautomer 18. It therefore seems probable that 2-nitrosobenzal-
dehyde undergoes hydration as it is formed, so only the weakly
absorbing tautomer 18 is present. Clean release of glycolate on
photolysis of 10 was inferred from previous data for a
insufficient information is available to permit detailed com-
parisons with our work.
2
-nitrobenzyl ether [2-O-(2-nitrobenzyl)-D-glucose 19], where
stoichiometric release of D-glucose upon photolysis was con-
To obtain some insight into the unusual reactivity shown by
firmed by an enzymatic assay.14 Time-resolved IR spectra for
6
-9 in comparison with the photocleavage of 10, we performed
19 upon flash photolysis also showed an aci-nitro intermediate,
17
gas-phase semiempirical PM3 calculations for a number of
the steps shown in Scheme 2. The PM3 method was chosen, as
it yields the best results for organophosphorus and nitro
followed by reversion to a near-featureless spectrum (data not
shown), as observed for 10. The second issue in the photolysis
of 10 is the absence of any evidence for a hemiacetal
intermediate. Rates for breakdown of hemiacetals derived from
benzaldehydes and acetophenones are dominated by hydroxide
ion catalysis at alkaline pH values, and the second-order rate
constants for hydroxide ion catalysis are an order of magnitude
18
compounds among the MNDO-type Hamiltonians. The suit-
ability of the method for 6-10 was also checked where possible
by comparison of our results with previous high-level calcula-
tions on the isomerization of 2-nitrotoluene to 2-nitrosobenzyl
4a
alcohol [B3LYP/6-311+G(2d,p)] or on the same isomerization
lower for aldehyde-derived hemiacetals than those for those
19
and that of 2-nitrobenzyl methyl ether [B3PW91/6-31G(d)]
derived from ketones.1
2,15,16
Hence if a hemiacetal intermediate
(
see below).
analogous to 11 were formed in the photolysis of 10 at pH 8.5,
it should have had a lifetime of at least several seconds,
substantially longer than the observed lifetime of the aci-nitro
intermediate from 10 (∼0.4 s at 1 °C), and therefore should
have been observable in our FTIR spectra, inter alia by the
presence of a strong nitroso monomer band. The absence of
such a band thus excludes a hemiacetal intermediate and 10
appears exclusively to undergo photocleavage by a “normal”
pathway analogous to that shown in Scheme 1, i.e., direct
collapse of a bicyclic intermediate equivalent to 3 to yield the
end products. A referee has drawn our attention to unpublished
work briefly described in a review article,1 where it is indicated
that photolysis of 2-nitrobenzyl methyl ether in nonaqueous or
aqueous solvents gave evidence of a hemiacetal intermediate.
This contrasts with the results described here for the 2-nitroben-
zyl ether 10 in purely aqueous conditions. Full publication of
these unpublished data will clearly be of interest, but at present
Reaction barriers were calculated for the steps 14 f 11, 16
f 17, and 17 f 11 for each of compounds 6-9, nitrobenzyloxy
compound 10, and model alcohols 20 and 21, which are
uncharged analogues of 6-10 (Table 3). For the ionic com-
pounds, the calculations were performed in different ionization
states, as shown in Table 3, to establish whether the charged
side chains significantly influenced the calculated reaction
barriers. Several points stand out among these data. With the
exception of the data for 6 in its anionic form, the barriers
calculated over the range of compounds did not vary greatly
within each process, and values for the uncharged compounds
e
2
0 and 21 were similar to those for the ionic species. Second,
the barriers calculated for the direct conversion of 14 to 11 were
consistently higher than for the step of the competing pathway
that involves cyclization of 16 to the benzisoxazoline 17, but
-
1
in no case by >19.3 kJ mol . The relatively close magnitude
of these calculated barriers leaves open a reasonable possibility
that the two pathways could operate in competition (see below),
(
(
(
13) McClelland, R. A.; Coe, M. J. Am. Chem. Soc. 1983, 105, 2718.
14) Corrie, J. E. T. J. Chem. Soc., Perkin Trans. 1 1993, 2161.
15) Cox, B. G.; Kresge, A. J.; Sørensen, P. E. Acta Chem. Scand. 1988, A42,
(17) Stewart, J. J. P. J. Comput. Chem. 1989, 10, 209.
(18) Stewart, J. J. P. J. Comput. Chem. 1989, 10, 221.
(19) Dunkin, I. R.; Gebicki, J.; Kiszka, M.; San ´ı n-Leira, D. J. Chem. Soc., Perkin
Trans. 2 2001, 1414.
2
02.
(
16) Engell, K. E.; McClelland, R. A.; Sørensen, P. E. Can. J. Chem. 1999, 77,
9
78.
J. AM. CHEM. SOC.
9
VOL. 125, NO. 28, 2003 8551

A R T I C L E S
Corrie et al.
especially as the photochemically generated 14 is probably
formed in a vibrationally excited state that would facilitate
crossing of activation barriers. No pathway with a barrier that
could reasonably be crossed was found for a direct route from
the initial nitronic acid 14 to the bicyclic intermediate 17. Such
a step would have provided an alternative product-directed exit
from 14, but its inaccessibility appears to leave only the one-
step 14 f 11 route available.
Figure 4. Arrhenius plot of the kinetics of aci-nitro decay following flash
photolysis of 9 at pH 8.5.
Our measurements of aci-nitro and pH transients have
confirmed the presence of 15 as an intermediate between 14
and 16, and direct isomerization of 14 to 16 by rotation about
the CdN bond has been excluded by previous workers, either
process (Table 3), which we calculated also to be exothermic
-1
by ∼105 kJ mol .
Finally, barriers calculated for ring opening of 17 to give
hemiacetal 11 were fairly consistent over the range of com-
pounds. The calculation was done for a water-mediated pathway,
and the barriers reported in Table 3 represent the rate-limiting
step of initial transfer of the N-OH proton to a water molecule.
For the subsequent shift of the proton to the endocyclic oxygen
of the benzisoxazoline, the computed barrier was only a few kJ
-
1
by comparison with experimental values of ∼210 kJ mol for
4
a
19
-1
oxime ethers or from a calculated barrier of 157 kJ mol .
-
1
Our calculations yielded a barrier >124 kJ mol for this
rotation in the species 14 derived from 6-9. For an alternative
conversion of 14 to 16 via direct proton shift between the two
oxygens of the aci-nitro group (without participation of a water
molecule), the activation barriers computed by density functional
calculations are in the range 85-95 kJ mol- for the unsubsti-
tuted nitronic acid derived from 2-nitrotoluene,19 while calcu-
-1
mol . As discussed above, the barrier for the initial proton
1
transfer would be substantially lower in bulk water than for the
gas-phase calculation, so the actual barrier in solution for the
process 17 f 11 is probably comparable to that for cyclization
of 16. Furthermore, base catalysis of this process is likely to
operate under the conditions studied here. It is relevant in this
context that benzisoxazoline intermediates similar to 17 have
not to our knowledge been observed to accumulate in photolytic
reactions of any 2-nitrobenzyl system in aqueous solution.
The extent of the reduction in the calculated activation barriers
when reaction takes place in bulk solvent was investigated for
decay of the aci-nitro intermediate 15 generated upon photolysis
of 9, i.e., the step presumed to operate via the pathway 15 h
-
1
lated barriers for our substituted molecules are ∼142 kJ mol .
This difference is not surprising, given the different basis sets
used. The reaction would involve strong localization of positive
charge on the shifted proton, so it can be expected that high-
level methods, which employ substantial basis sets for hydrogen,
will yield lower activation barriers than MNDO-type methods
that use only a minimal basis set. Furthermore, inclusion of a
single water molecule to mediate the shift of this proton is
computed in our PM3 calculations to lower the activation barrier
-
1
by at least 29 kJ mol compared to the direct shift. The
presence of additional water molecules in bulk solution that are
within hydrogen bonding distance of the transiently formed
16 f 17, which is considered from the experimental data above
to represent the minor pathway for overall product formation.
The temperature dependence of the process was measured over
the range 2.3-31.5 °C for a solution buffered at pH 8.5 by 10
mM sodium pyrophosphate. This salt was used, as phosphates
+
H3O should lower the activation barrier for water-mediated
proton exchange even further. A previous computational study
of such effects on proton transfer barriers showed that the barrier
for proton transfer from imidazole to water is lowered by 84 kJ
2
1
have a low heat of ionization. The Arrhenius plot for these
-
1
20
mol in the presence of two ligand water molecules. Taking
this solvation effect into account, water-mediated proton
exchange via the anion 15 can be assumed to be the most likely
path between the nitronic acid isomers 14 and 16, with an
-
1
data (Figure 4) gave an activation energy of 24.8 kJ mol ,
which is presumably for the highest barrier in the steps between
1
5 and 11. This implies that the rate-determining step in the
overall process is not opening of the bicyclic intermediate 17
to the hemiacetal 11, since the measured rate constants of the
slow phase of nitroso formation (that corresponds to 17 f 11)
and of aci-nitro decay (that corresponds to overall 15 f 17)
-
1
activation barrier of only a few kJ mol .
As foreshadowed in the Introduction, direct cyclization of
the aci-nitro anion 15 to the anion of 17 can be excluded.
4
a
Il’ichev and Wirz found this step in the case of the reaction
for 2-nitrotoluene to have an activation barrier in the region of
-1
are identical. The measured 24.8 kJ mol barrier for the overall
decay of the aci-nitro anion 15 is therefore in one of the steps
between 15 and 17, but in this kinetically complex process we
do not attempt to assign it to an individual step. The activation
-
1
1
67 kJ mol and to be strongly endothermic. In contrast, they
reported that cyclization of the neutral form (corresponding to
the process 16 f 17) was exothermic, with an activation barrier
-1
energy measured in solution is 63-126 kJ mol lower than
-
1
19
in the range 77-102 kJ mol . Dunkin et al. calculated a
the calculated gas-phase values in Table 3 for the steps 16 f
-1
barrier of 59 kJ mol for the corresponding step in the reaction
of 2-nitrobenzyl methyl ether. These calculations agree reason-
ably with our calculated barriers of ∼105 kJ mol- for the
-1
1
7 and 17 f 11 and is substantially below the 62 kJ mol
previously measured for the aci-nitro decay process of caged
1
(
21) Edsall, J. T.; Gutfreund, H. Biothermodynamics. The Study of Biochemical
(20) Lill, M. A.; Hutter, M. C.; Helms, V. J. Phys. Chem. A 2000, 104, 8283.
Processes at Equilibrium; Wiley: Chichester, 1983; pp 165-169.
8
552 J. AM. CHEM. SOC. VOL. 125, NO. 28, 2003
9

Photolytic Cleavage of 1-(2-Nitrophenyl)ethyl Ethers
A R T I C L E S
3
ATP. It is reasonable to assume that the activation barrier for
From these data it appears reasonable in the methyl-
substituted series, where the initial nitronic acid 14 is of
substantially weaker acidity than the corresponding species that
lacks the R-methyl group, that the rearrangement of 14 that leads
directly to the hemiacetal 11 could be in competition with the
alternative of ionization to the aci-nitro anion 15. Once
ionization has occurred, the aci-nitro anion will be kinetically
trapped, with its decay controlled by the second-order rate
constant for reprotonation. The protonation step is more likely
to take place on the less-hindered oxygen to give 16, followed
by cyclization to 17. For the nitronic acid of higher acidity
formed initially in the nonmethylated series, ionization is likely
to compete more effectively with the direct rearrangement path,
so the reaction proceeds entirely (subject to the limits of
detectability) via the “conventional” benzisoxazoline route.
The discussion above gives a broad rationale for the reactivity
of the compounds investigated here. Further detailed studies
beyond the capability of our present instrumentation will be
required to delineate details of individual steps. The present work
however does highlight the likely importance of differing
acidities of structurally diverse nitronic acids that has not
previously been considered.
the direct 14 f 11 path in solution is no higher than that
measured from the Arrhenius plot for the overall process 15 f
1, i.e., that both pathways have a low barrier in aqueous
solution.
To rationalize the occurrence of the two competing pathways
in the overall reaction of 6-9 (Scheme 2) and the contrary
position with 10, which appears to react only by a single
pathway (the non-methyl analogue of the route via 16 and 17),
it was necessary to consider the additional factor of flux through
1
4
b
different intermediates. The study by Schw o¨ rer and Wirz
explains the kinetics of acid-catalyzed decay of aci-nitro anions
of the type considered here, i.e., that cyclization to a benzisox-
azoline such as 17 requires prior protonation of the aci-nitro
anion and is therefore controlled by the pK of this species. The
decay of 15 is considerably faster than that of its nonmethylated
analogue, which indicates that 15 is the stronger base of the
two compounds. The underlying reasons for this difference in
basicity and its consequences are discussed below.
aci-Nitro anions of type 15 are expected to be stronger bases
than the corresponding species lacking the R-methyl substituent,
because of the combined actions of the inductive effect of the
methyl group and poorer solvation of the more substituted
system. A stereoelectronic factor probably also operates to
destabilize 15 relative to its nonmethylated analogue. Our gas-
phase PM3 calculations showed a dihedral angle between the
exocyclic C-C and C-N bonds in 15 of 42.6° for the
Z-configuration shown (calculated for the species derived from
Conclusions
The overall results of this study and the proposal outlined
above to explain the kinetic records of Figure 2 obtained upon
flash photolysis of 9 have general implications for the field of
caged compounds. A particularly significant point for their use
in time-resolved biological studies is that reliance solely on aci-
nitro decay rates can be misleading as an indicator of the rate
of product release. Our data demonstrate that the discrepancy
can in some cases be very large and reinforce the value of
complementary time-resolved infrared measurements to study
this photochemistry. The elegant work by Toscano and col-
6
with its carboxylate group in an ionized state; note that
calculations for the E-configuration gave a similar dihedral
angle). By contrast in the non-methyl analogue of 15, where
the alkoxy substituent has a preferred E-configuration that
minimizes steric crowding (i.e., structure 22), this dihedral angle
1
9
was calculated to be 2.5° (cf. calculations by Dunkin et al.
for different isomers of the nitronic acid derived from 2-ni-
trobenzyl methyl ether, which consistently show a planar
conformation for the E-configuration but twisting of ∼30° for
the Z-configuration). To test the effects of this twisting,
semiempirical calculations that include solvation effects were
24
leagues on photochemistry of diazeniumdiolates further shows
the value of the technique in unraveling mechanistic details of
dark reactions triggered by photolysis. Indeed, it is noteworthy
that they were also able to detect two competing mechanistic
pathways in their study.
The duality of mechanism uncovered in our work is interest-
ing, as it requires competition between very rapid alternative
pathways. From the discussion presented here, this partition
appears to be controlled by the pKa of the initially formed
nitronic acid, but it is of interest that the participation of dual
pathways has not previously been observed in o-nitrobenzyl
photocleavage reactions. In cases such as caged ATP, it seems
likely that the presence of an anionic side chain very near the
nitro group would facilitate ionization of the initial nitronic acid
22,23
carried out using the AMSOL program.
Solvation energies
were calculated for the deprotonation reaction of the respective
nitronic acid isomers derived from 6 and 10, respectively, to
their corresponding aci-nitro anions. These energy differences
calculated for the deprotonation reaction (14 f 15) showed that
the aci-nitro anion derived from 10 is 15 kJ mol better
stabilized than the anion from 6, which corresponds to a
difference of at least 2.6 pKa units between the two nitronic
acids; that is, the nitronic acid derived from 10 is significantly
more acidic than that derived from 6 (and by extension from
-
1
(analogous to 14), thereby channeling the reaction via the slower
7
-9).
aci-nitro path. Nevertheless, we recognize that our analysis
leaves questions open. For example, the fact that the partitioning
between fast and slow processes for photolysis of 6-9 is
independent of pH (as judged by the pH-invariant amplitude of
the 406 nm transients for 9) suggests that additional consider-
ations may be relevant, in particular the extent to which
partitioning of 14 is influenced by reaction from a vibrationally
excited state. The factors that promote formation of a hemiacetal
(
22) Hawkins, G. D.; Giesen, D. J.; Lynch, G. C.; Chambers, C. C.; Rossi, I.;
Storer, J. W.; Li, J.; Zhu, T.; Rinaldi, D.; Liotard, D. A.; Cramer, C. J.;
Truhlar, D. G. AMSOL version 6.5.3; University of Minnesota: Min-
neapolis, 1997.
(24) Srinivasan, A.; Kebede, N.; Saavedra, J. E.; Nikolaitchik, A. V.; Brady,
D. A.; Yourd, E.; Davies, K. M.; Keefer, L. K.; Toscano, J. P. J. Am.
Chem. Soc. 2001, 123, 5465.
(
23) Cramer, C. J.; Truhlar, D. G. J. Comput. Chem. 1992, 13, 1089.
J. AM. CHEM. SOC.
9
VOL. 125, NO. 28, 2003 8553

A R T I C L E S
Corrie et al.
intermediate in the photolysis of 6-9 but not for the nor-
analogue 10 also remain to be explored. We hope the data
presented here will stimulate further enquiry into this mecha-
nistically diverse area.
using the Tektronix Wavestar software. Buffer solutions were prepared
by pH adjustment with dilute NaOH of appropriate molarity solutions
of salts specified in figure captions. Records for measurements at 406
or 577 nm were processed by averaging 2-4 individual traces
(uncorrected for blanks) and fitted to single-exponential functions in
Experimental Section
Microsoft Excel. The 740 nm records were averaged from 16 individual
traces and were corrected by subtraction of an averaged equal number
of blank traces (obtained from a solution containing the same buffer
but without the relevant photolabile compound) to compensate for flash
artifacts before fitting the slow component of the trace.
General and Spectroscopic Details. Details of the preparation of
compounds 6-10 and 12 and of the product study and quantum yield
determination for photolysis of 9 are given as Supporting Information.
Note that both 2-methyl-2-(2-nitrophenyl)-1,3-dioxolane (the starting
material for preparation of 6-10) and 12 are compounds previously
described but were prepared here by modified, more satisfactory
procedures, as described in the Supporting Information. The time-
resolved IR spectrometer was as described previously,8 and IR
measurements were performed in buffers made by adjusting 200 mM
solutions of 2-(N-morpholino)ethanesulfonic acid (MES), 3-(N-mor-
pholino)propanesulfonic acid (MOPS), N-(2-hydroxyethyl)piperazine-
N′-3-propanesulfonic acid (EPPS), Bicine, 2-(N-cyclohexylamino)-
ethanesulfonic acid (CHES), or 3-(N-cyclohexylamino)propanesulfonic
acid (CAPS) with 3 M NaOH to pH values specified in the text. To
analyze the kinetics of carbonyl formation and nitroso monomer decay
for compounds 6-9, we averaged the kinetic difference spectra from
three separate samples of each compound and integrated the appropriate
bands. Integration was performed using Bruker Opus NT 3D 3.1
software with respect to a baseline drawn between two data points of
the spectrum at each side of the band. The two baseline points were
calculated by averaging data points in a specified spectral range. Spectral
ranges for calculating the two baseline points and for the integration
Computational Methods. Semiempirical PM3 calculations were
2
5
performed using the program package VAMP Version 6.5 with the
PM3 Hamiltonian.17 The molecules investigated were energetically
-
1
-1
optimized to a gradient norm below 0.4 kJ mol
Å
using the
Eigenvector Following (EF) algorithm.25 In the reaction path calcula-
tions a step width of 0.05 Å was used. Geometries of transition states
were taken from the points on the reaction path calculations that showed
a maximum in energy. Subsequent transition state searches showed
-
1
that the resulting energies are within 0.4 kJ mol of those taken from
the reaction path calculations and are thus well within the error limit
of the underlying method. Calculations that included solvation effects
22
were carried out with the AMSOL program using the default settings
along with the PM3-SM3 Hamiltonian.23
Acknowledgment. We are grateful to Professor Werner
M a¨ ntele for provision of facilities, to the Royal Society for an
ESEP grant that facilitated collaboration between the U.K. and
German laboratories, and to the MRC Biomedical NMR Centre
for access to NMR instrumentation. We also thank the EPSRC
National Mass Spectrometry Centre for negative ion mass
spectrometric measurements.
-
1
were 1665-1674 and 1706-1734 cm (baseline points) and 1679-
-
1
-1
1
694 cm (carbonyl band) and 1477-1489 and 1512-1523 cm
-
1
(baseline points) and 1494-1508 cm (nitroso band), respectively.
The data were analyzed as single-exponential functions by least-squares
fits in Microsoft Excel.
Supporting Information Available: Sources or synthetic
procedures and characterization for compounds 6-10 and 12,
quantum yield determination, and product study for photolysis
of 9 (PDF). This material is available free of charge via the
Internet at http://pubs.acs.org.
The time-resolved UV-visible spectrometer was as previously
2
described except that photolysis was triggered by 347 nm light from
a frequency-doubled ruby laser (Lumonics QSR2/6, Rugby, U.K.), pulse
width ∼30 ns. Separate photomultiplier tubes were used for measure-
ments at 740 nm and for measurements in the range 406-577 nm, as
JA034354C
2
previously described. The cuvette had 4 mm optical paths for both
(
25) Rauhut, G.; Alex, A.; Chandrasekhar, J.; Steinke, T.; Sauer, W.; Beck, B.;
Hutter, M.; Gedeck, P.; Clark, T. VAMP Version 6.5; Oxford Molecular:
Erlangen, 1997.
the laser pulse and the probe light. Data were collected on a TDS 210
oscilloscope (Tektronix Inc., Beverton, OR) and transferred to a PC
8554 J. AM. CHEM. SOC.
9
VOL. 125, NO. 28, 2003