Photodecarboxylation of phenylglyoxylic acid: Influence of para-substituents on the triplet state properties

Article abstract of DOI:10.1039/a906567i

The photochemistry of several para-substituted derivatives of phenylglyoxylic acid (R-PA, R = OCH₃, CH₃, F, Cl, Br and CN) was studied in polar solvents. The triplet state was detected by laser flash photolysis and phosphorescence in acetonitrile, acetone or acetic acid at room temperature and by phosphorescence in polar glasses at -196°C. In fluid media the phosphorescence intensity and the triplet lifetime are both reduced on addition of propan-2-ol or water. The rate constant for H-atom abstraction by the triplet of the R-PA from propan-2-ol, 2 × 10⁴-8 × 10⁷ dm³ mol^-1 s^-1 in acetonitrile, increases with the electron-accepting power of the substituent. The intermediate is the corresponding α-carboxy-α-hydroxybenzyl radical. The main transient in neat aqueous solution is the triplet state of the corresponding benzaldehyde as secondary intermediate. Based on results of triplet quenching by water in mixtures with acetonitrile and of time-resolved conductivity measurements, heterolytic α-splitting of the triplet anion involving the benzoyl anion is suggested. Photodecarboxylation occurs in substantial yield for R = CH₃, F, Cl, Br and CN in acetonitrile and Φ(-CO₂) is enhanced in the presence of water, Φ(-CO₂) = 0.4-0.8 at 5-20% H₂O. The values are larger than in neat aqueous solution, where a decreasing dependence vs. pH was generally found. A reaction scheme is presented accounting for the observed intermediates and the dependence of Φ(-CO₂) on the water content and the pH.

Full text of DOI:10.1039/a906567i

View Article Online / Journal Homepage / Table of Contents for this issue
Photodecarboxylation of phenylglyoxylic acid: influence of para-
substituents on the triplet state properties
Helmut Görner* and Hans Jochen Kuhn
Max-Planck-Institut für Strahlenchemie, D-45413 Mülheim an der Ruhr, Germany.
E-mail: goerner@mpi-muelheim.mpg.de, kuhn@mpi-muelheim.mpg.de
Received (in Cambridge, UK) 12th August 1999, Accepted 27th September 1999
The photochemistry of several para-substituted derivatives of phenylglyoxylic acid (R–PA, R = OCH , CH , F, Cl, Br
3
3
and CN) was studied in polar solvents. The triplet state was detected by laser flash photolysis and phosphorescence in
acetonitrile, acetone or acetic acid at room temperature and by phosphorescence in polar glasses at Ϫ196 ЊC. In fluid
media the phosphorescence intensity and the triplet lifetime are both reduced on addition of propan-2-ol or water.
4
7
3
Ϫ1 Ϫ1
The rate constant for H-atom abstraction by the triplet of the R–PA from propan-2-ol, 2 × 10 – 8 × 10 dm mol
s
in acetonitrile, increases with the electron-accepting power of the substituent. The intermediate is the corresponding
α-carboxy-α-hydroxybenzyl radical. The main transient in neat aqueous solution is the triplet state of the
corresponding benzaldehyde as secondary intermediate. Based on results of triplet quenching by water in mixtures
with acetonitrile and of time-resolved conductivity measurements, heterolytic α-splitting of the triplet anion
involving the benzoyl anion is suggested. Photodecarboxylation occurs in substantial yield for R = CH , F, Cl, Br
3
and CN in acetonitrile and Φ(ϪCO ) is enhanced in the presence of water, Φ(ϪCO ) = 0.4–0.8 at 5–20% H O.
2
2
2
The values are larger than in neat aqueous solution, where a decreasing dependence vs. pH was generally found.
A reaction scheme is presented accounting for the observed intermediates and the dependence of Φ(ϪCO ) on the
2
water content and the pH.
2
1
Introduction
decarboxylation have been reviewed by Budac and Wan. The
role of water in the mixtures with inert polar solvents is
fourfold: the presence of water favours the heterolytic vs. a
homolytic mechanism and determines the acid–anion ground
state. Water interacts with the triplet state of the anion (thereby
forming the BA or naphthaldehyde precursor) and interacts
with the excited singlet state of the anion (thereby bypassing the
The role of water in the photodecarboxylation of phenylgly-
1,2
oxylic acid (PA) has been a long-standing puzzle. Of the
α-keto carboxylic acids mainly pyruvic acid, the simple ali-
phatic analogue of PA, has been investigated by photochemical
2–10
methods.
The photochemistry of PA, however, differs
significantly from that of pyruvic acid. PA undergoes photo-
reduction in the presence of H-donating solvents and photo-
decarboxylation in mixtures of water with inert polar solvents,
14–16
latter reaction towards the aromatic aldehydes).
This work provides a deeper insight into the mechanistic
details of the decarboxylation of R–PA. For this purpose we
have introduced several substituents (R = OCH , CH , F, Cl, Br
11–15
such as acetonitrile, acetone or acetic acid.
3
3
and CN) in the para position. The results in mixtures of water
with inert polar organic solvents and as a function of the pH
are presented. They support the proposed heterolytic splitting
reaction. Despite a large variation of the reactivity of water
with the triplet state of the anion, the para substituent
hν
(
1)
PA
BA ϩ CO2
The quantum yield for decarboxylation of PA in acetonitrile
in the presence of 10–20 mol dm water is Φ(ϪCO ) ≈ 0.8 and
the reaction is suitable as a chemical actinometer.
these conditions PA is present as anion and benzaldehyde (BA)
and CO are virtually the only photoproducts. On the basis of
phosphorescence and T-T absorption measurements at room
temperature we have recently concluded that the triplet anion
of PA is deactivated by a reaction with water. Moreover,
transient conductivity studies in neat aqueous solution and
mixtures with acetonitrile revealed that the formation of BA
is faster than (or as fast as) the neutralization reaction. Thus
we have suggested that reaction (1) occurs via heterolytic α-
splitting resulting in the benzoyl anion, C H CO , or its proto-
nated congruent. Free radicals play no role in the decarboxyla-
tion mechanism of PA in the presence of water.
Photolysis of 1- and 2-naphthylglyoxylic acid in mixtures of
acetonitrile with water yields CO and the corresponding
naphthaldehydes. A heterolytic C–C bond cleavage involving
carbanions has been demonstrated for other aromatic sys-
Ϫ3
2
1
2,13
(R = CH , F, Cl, Br and CN) has no significant effect on the
3
Under
Φ(ϪCO ) values with respect to the pH dependence in neat
2
Ϫ3
aqueous solution and in the mixtures of water (>5 mol dm )
2
with acetone or acetonitrile.
1
4
Results
Ground state and excited singlet state
The absorption spectra of the R–PA in acetonitrile have
maxima at 250–300 nm (see Experimental section) and, with
Ϫ
the exception of OCH –PA, a pronounced shoulder around
6
5
3
3
Ϫ1
Ϫ1
340–380 nm with an ε value of 50–90 dm mol cm . Addition
of water to acetonitrile leads to a blue-shift of the spectrum at
λ >280 nm. The absorption spectrum in neat water shows a
significant pH dependence, e.g. for CN–PA the absorbance at
250 nm increases with increasing pH between 0 and 2 and
remains almost constant thereafter (not shown). Similar curves
1
4,15
2
16
1
7,18
tems.
radicals.
Photolysis of alkyl phenylglyoxylates involves
The homolytic and heterolytic mechanisms for
were found for other R–PA, e.g. for A290 of OCH –PA and an
3
1
9,20
opposite dependence at longer wavelengths (A340). Three
J. Chem. Soc., Perkin Trans. 2, 1999, 2671–2680
This journal is © The Royal Society of Chemistry 1999
2671

View Article Online
a
Table 1 Phosphorescence maxima (in nm) in glassy and fluid media
Room temperature
Ϫ196 ЊC
Compound
Acetonitrile
Acetone
Ethanol
OCH –PA
470,505,550sh
465,500
470,505,550sh
472,520,555sh
470,508,550sh
470,508,555sh
486,524,560sh
463,501,544
466,506,550
466,505,545
465,505,550
470,510,558
468,508,555
480,520,566
3
CH –PA
472,513,554sh
472,511,555sh
475,515,557sh
477,515,560sh
472,514,560sh
3
b
PA
F–PA
Cl–PA
Br–PA
CN–PA
a
In air- and argon-saturated solutions at Ϫ196 and 25 ЊC, respectively,
λexc = 340 nm for acetone and 250–280 nm for the other media (sh:
b
shoulder). Values taken from ref. 14.
Fig. 2 (a) Phosphorescence spectra of OCH –PA in ethanol at
3
Ϫ196 ЊC (——) and in argon-saturated acetonitrile at 25 ЊC (– – –);
λexc = 340 nm and λ = 500 nm.
p
butyronitrile at Ϫ196 ЊC. The Φ values are in all cases between
p
0
.5 and 0.95, indicating that deactivating processes other than
those leading to the lowest triplet play only a minor role. The
phosphorescence lifetime is typically τ = 3 ms in both media
p
and shows only minor changes (smaller than ±0.6 ms) with
substitution. It should be emphasized that oxygen virtually
does not quench Φ nor τ in these glassy solvents. The first
p
p
maximum of the phosphorescence band of the R–PA in glassy
butyronitrile corresponds to a triplet energy of approximately
Ϫ1
2
59 kJ mol . Slightly smaller values were found at room
Ϫ1
temperature, the smallest with 242 kJ mol for CN–PA.
Phosphorescence in fluid media
Phosphorescence was also observed for the R–PA in argon-
saturated acetonitrile and acetone solutions at 24 ЊC (Table 1).
The spectra are less structured but the maximum is nearly the
same as in glassy ethanol or butyronitrile. The two major peaks
depend only little on the substituent and the medium. Examples
of the phosphorescence spectra in glassy and fluid media are
shown in Fig. 2. The quantum yield is reduced at least 100-fold
on going from Ϫ196 to 24 ЊC. This corresponds with the change
in lifetimes. We therefore suggest that the quantum yield of
Fig. 1 Absorption spectra of CN-, CH - and OCH –PA (from left to
3
3
right) in aqueous solution at pH 3.3 (——) and pH 0 (– – –).
examples below and above the pK value are shown in Fig. 1.
a
The pH for 50% change at the appropriate wavelengths is taken
14
as a measure of the pK value. For PA itself and the R–PA
a
pK = 1.1; only for CN–PA the pK value is slightly larger
a
a
(
≈1.2).
Apparently, the lifetime of the excited singlet state of R–PA
intersystem crossing (Φ ) is also close to unity for all of the
isc
is extremely short since only weak fluorescence could be
detected, either in fluid or in glassy media. The largest quantum
yield of fluorescence, ≈0.01 in ethanol at Ϫ196 ЊC, was found
R–PA in organic solvents at room temperature. The longest
phosphorescence lifetimes in the 10–30 µs range and first-order
decay were obtained with λexc = 308 nm at the lowest possible
pulse intensity (Table 2).
for OCH –PA.
3
The intensity at the maximum (I ) and τ are both quenched
p
p
o
by alcohols (Figs. 3a and 3b, respectively). Plots of I /I and
τ /τ vs. the propan-2-ol concentration are linear, the Stern–
p
p
Phosphorescence in glassy media
o
p
p
The emission spectrum of Br–PA in butyronitrile at Ϫ196 ЊC
has maxima at 464 and 502 nm and progressions at 545 and 595
nm. Very similar spectra with changes in λmax of less than 20 nm
were recorded for all R–PA in ethanol (Table 1). By comparison
with the parent molecule we conclude that the emission in the
Volmer constants (K2-p) were obtained from the slopes. K2-p and
the rate constant for phosphorescence quenching (k2-p) increase
both in the order CH –, F–, Cl–, Br– and CN–PA (Table 3).
3
For a given R–PA one would expect the same result from the
two methods, i.e. τ × k = K2-p. Indeed, this is fulfilled (within
p
2-p
2,14
case of the R–PA has to be attributed to phosphorescence.
the experimental error) in several cases, e.g., when τp was
For several R–PA (e.g., Br–PA) it was shown that the shape
obtained by using a low concentration and a small laser
of the spectrum and the maximum are independent of λ
intensity.
exc
Ϫ4
Ϫ2
Ϫ3
Ϫ1
(
230–380 nm for concentrations of 10 and 10 mol dm
,
From the linear part of Ip vs. [H O] (Figs. 3a and 3b) a
2
respectively). Similar phosphorescence spectra were recorded in
glassy ethylene glycol and water mixtures (2:1, v/v), where the
phosphorescence spectrum is less structured and the maximum
is slightly blue-shifted. Replacement of neutral by acid water
Stern–Volmer constant was estimated; K
to у70 dm mol (Table 4). As a second measure for the
ranges from 0.1
H
2
O
3
Ϫ1
phosphorescence quenching by water the slope (kH2O) of the
Ϫ1
linear part of τp vs. [H O] was taken. The reactivity against
2
(
pH р 1, R = CH , Br) has no marked effect on the spectrum.
water, as measured by KH2O and kH2O, increases for the sub-
stituent in the above order and is slightly larger for a given
R–PA in acetone than in acetonitrile and markedly smaller in
acetic acid. A further more accurate measure for the reactivity
of the triplet is obtained from T-T absorption (see below).
3
This indicates that in glasses the emission from both, acid and
anion, is similar.
The phosphorescence quantum yield (Φ ) and the corre-
p
sponding lifetime (τ ) were determined in ethanol and/or
p
2
672
J. Chem. Soc., Perkin Trans. 2, 1999, 2671–2680

View Article Online
a
Table 2 Phosphorescence and triplet lifetimes, absorption maximum and rate constants for quenching by oxygen and cyclohexa-1,3-diene
Ϫ9
3
1
0
k /dm
s
ox
Compound
Solvent
λexc/nm
τp b/µs
τT/µs
λTT c/nm
mol
Ϫ1 Ϫ1
d
OCH –PA
Acetonitrile
248
0.8
35
1
4
5
3
8
3
3
445
328,(410)
5 [9]
3 [6]
3
CH –PA
3
3
08
248
08
e
PA
322,(410)
2 [6]
3
20
16
F–PA
Cl–PA
1.5
2 [6]
248
08
248
08
320,(450)
332,(440)
320,(500)
3
17
Br–PA
3
2
1.5
3
14
<2
CN–PA
248
308
248
2 [8]
OCH –PA
Acetic Acid
2
8
440
320,(410)
318,(420)
3
CH –PA
20 (>10)
3
e
PA
6
3
08
22 (7)
30 (>8)
20
8
[3]
F–PA
Cl–PA
Br–PA
CN–PA
10
8
1
20 (>5)
325,(400)
a
Ϫ3
b
In argon-saturated solution at 25 ЊC; typical concentration: 0.2, 1.0 and 10 mmol dm for λexc = 248, 308 and 354 nm, respectively. For the
c
phosphorescence lifetime the concentration was reduced by a factor of 2–10; in parentheses: values in acetone, λ = 354 nm. Weak, second
maximum in parentheses. In square brackets: rate constant k for quenching by cyclohexa-1,3-diene. Taken from ref. 14.
exc
d
e
q
Table 3 Stern–Volmer constant and rate constants for quenching of
Absorption of the triplet state of phenylglyoxylic acid
a
phosphorescence and T-T absorption by propan-2-ol
Ϫ3
For most R–PA (0.1–0.5 mmol dm ) in acetonitrile solution
b
3
Ϫ6
c
3
Ϫ6
d
3
at room temperature two transients, a major short-lived (TPA
)
and a minor longer-lived (TKЈ) were observed by laser flash
photolysis (λexc = 248 nm). Examples are shown in Figs. 4a
K2-p /dm
mol
10
k
/dm
10
k
/dm
2-p
Ϫ1 Ϫ1
2-p
Ϫ1 Ϫ1
Ϫ1
Compound
mol
s
mol
s
OCH –PA
р0.03
0.82
1.5
2
2.6
>3
>30
р0.02
0.8 (1.4)
1.5 (1.7)
1.6 (1.8)
2.5 (3.0)
3 (2.8)
and 4c for CN- and OCH
comitantly with the pulse and decays by first-order kinetics
3
–PA, respectively. TPA is formed con-
3
e
CH –PA
10
24
40
3
f
PA
kobs
T
= 1/τ ). The triplet lifetime is in the 1–10 µs range (Table
(
F–PA
2
). TPA is quenched by oxygen, a typical value of the rate
Cl–PA
Br–PA
CN–PA
у40
>20
у100
9 3 Ϫ1 Ϫ1
constant is kox = 2 × 10 dm mol
s . TPA is also quenched
by cyclohexa-1,3-diene and rate constants (k ) close to the
q
diffusion-controlled limit were obtained from linear depend-
a
In argon-saturated acetonitrile at 25 ЊC using λexc = 248 nm, unless
Ϫ1
b
ences of τ
T
on [quencher].
specified otherwise. Stern–Volmer constant, λ = 250–280 nm in
acetonitrile, 330–350 nm in acetone. Rate constant for phosphor-
exc
c
Transient TPA, observed with the R–PA in the three inert
polar solvents at room temperature, is assigned to the lowest
triplet state, in analogy to the parent molecule. The absorption
maximum (λTT) changes only little on variation of the sub-
d
escence quenching. Rate constant for quenching of T-T absorption;
e
14
λ
= 354 nm in acetone. Values in parentheses refer to acetone.
exc
f
Taken from ref. 14.
stituent, e.g. from 320 nm (CN–PA) to 332 nm (CH –PA)
3
and a second weak band emerges in the visible range (Table 2).
The only exception is OCH –PA; here λ is around 450 nm
3
TT
and the spectrum extends to the red spectral range (Fig. 4c).
Absorption of the triplet state of benzaldehyde
Ϫ3
In aqueous solution at pH = 1 the R–PA (0.1–0.5 mmol dm
,
λexc = 248 nm) exhibit essentially a short-lived transient (TBA) in
the UV which decays by first-order kinetics and a longer-lived
one with a low ∆A. The lifetime of the major transient TBA
ranges from 30 ns (F–PA) to 1 µs (OCH –PA) and λ is red-
3
TT
shifted in the order PA (342 nm), Cl–PA and Br–PA (403 nm).
For CH –PA it was verified that T is quenched by oxygen
3
BA
9
3
Ϫ1
Ϫ1
(
rate constant: у10 dm mol
s ). The ratio of absor-
bances of the longer- and shorter-lived transients is substantial
only for OCH –, CH – and CN–PA with no transient at pH у 4
3
3
for the latter case. ∆A(TBA) increased on repeated flashing
Ϫ3
when the excited volume (ca. 0.01 cm ) was not exchanged,
indicating that TBA originates from an excited photoproduct
rather than from the given R–PA. This transient of PA itself has
14
already been assigned to the triplet state of BA.
On increasing the pH (up to pH = 5), ∆A (measured at λTT of
Ϫ1
Fig. 3 Plots of (a) I Њ/I and (b) τ vs. the quencher concentration for
p
p
p
TBA) is approximately constant for OCH –PA, but decreases in
3
Br–PA (squares) and CH –PA (triangles) in argon-saturated acetonitrile
at 25 ЊC; quencher: propan-2-ol (full symbols) and water (open sym-
bols); λexc = 340 and 308 nm, for (a) and (b), respectively.
3
the other cases. For PA and CN–PA the change in ∆A is so large
that practically no transient could be detected at pH > 4
J. Chem. Soc., Perkin Trans. 2, 1999, 2671–2680
2673

View Article Online
a
Table 4 Stern–Volmer constant and rate constants for phosphorescence and triplet quenching by water
b
3
Ϫ6
c
3
Ϫ6
d
3
KH2O /dm
mol
10
k
/dm
10
k
/dm
O
Ϫ1 Ϫ1
H
2
O
H
2
Ϫ1
Ϫ1 Ϫ1
Compound
Solvent
mol
s
mol
s
e
OCH –PA
Acetonitrile
0.1e
0.2
0.8
>4
>4
>3
>4
у40
р0.05
0.9
6
3.5
8.5
11
≈100
3
CH –PA
0.8
3
f
e
PA
2
1
e
F–PA
Cl–PA
Br–PA
CN–PA
7
6
e
>70
OCH –PA
Acetone
0.01
1.8
7
3
CH –PA
1.5
18
2
у5
3
f
PA
Cl–PA
Br–PA
10e
14
15
6
10
e
e
CH –PA
PA
F–PA
Cl–PA
Br–PA
CN–PA
Acetic Acid
0.2
1.6
0.4
<0.3
0.5
0.6
>0.5
0.2
0.6
3
f
0.5e
0.6
0.5
0.8
8
a
b
In argon-saturated solution at 25 ЊC using λexc = 308 nm, unless specified otherwise. Stern–Volmer constant, λexc = 250–280 nm for acetonitrile,
c
d
3
30–350 nm in the other solvents. Rate constant for phosphorescence quenching, taken from plots of 1/τ vs. the H O concentration. Rate constant
p
2
e
Ϫ1
f
for quenching of T-T absorption, data from Table 5. Ip vs. [H O] is upward curved, K taken as 1/[H O]1/2. Taken from ref. 14.
2
H
2
O
2
9
3
quenched by oxygen; the rate constant is typically 2 × 10 dm
Ϫ1 Ϫ1
mol
s . This transient, observed with the parent molecule
14
in several alcohols and with the R–PA in propan-2-ol, is
assigned to the corresponding α-carboxy-α-hydroxybenzyl
(
ketyl) radical. For OCH –PA and CH –PA in 100% propan-2-
3
3
ol the triplet appeared concomitantly to the laser pulse since the
rate constant (k2-p) for quenching by propan-2-ol is smaller than
5
3
Ϫ1 Ϫ1
1
0 dm mol
s . The k2-p values were obtained from the linear
dependence of kobs (of T-T absorption in acetonitrile solution)
Ϫ3
vs. the propan-2-ol concentration (0–1 mol dm ). They are
a direct measure for the reactivity of the triplet against H-
atom abstraction.
The remaining transient (T ) after triplet decay in aceto-
KЈ
nitrile is tentatively also assigned to the ketyl radical since the
spectrum and the dependence of λmax on the substituent are
similar to those of T . The ratio of absorbances of T and TPA
K
KЈ
(
0
∆A(KЈ)/∆A(PA)), measured at both maxima, ranges from
.2 to 0.5. Although ∆A(KЈ) is small the rather short triplet
lifetime in acetonitrile indicates a reaction competing with
phosphorescence and non-radiative intersystem crossing. This
reaction may be either H-atom abstraction from acetonitrile
or from the carboxyl group in the R–PA.
Fig. 4 Transient absorption spectra (a) of CN–PA (circles) and CH3–
PA (triangles) in argon-saturated acetonitrile, 100 ns (open symbols)
and 10 µs (᭹) after the pulse; (b) of CN–PA in argon-saturated propan-
Decay of the triplet in water–acetonitrile mixtures
2
-ol in the absence and presence (9:1) of water, pH 10 (circles and
As pointed out above, the triplet lifetime of most R–PA
becomes shorter on addition of water to acetonitrile. The plot
of k vs. the H O concentration (Fig. 5b), in analogy to that of
triangles, respectively) and (c) of OCH –PA in argon-saturated aceto-
3
nitrile, 0.1 µs (——) and 10 µs (– – –) after the pulse; λexc = 248 nm.
obs
2
Ϫ1
(
∆A(pH = 4)/∆A(pH = 1) <30). In other cases (OCH –, CH –
vs. [H O] (Fig. 3b), does not simply show initially a linear
3
3
τ
p
2
and Cl–PA) the absorption spectrum is similar to that of
TBA at pH = 1. We therefore propose that it is the same major
TBA transient observed in the accessible pH range. The reason
for its disappearance in the more basic range is the reduced
yield since the lifetime increases only slightly on increasing
the pH.
dependence on the H O concentration; k
increases only
obs
2
slightly in the 0–2% range and shows a more or less linear
dependence between 2 and 10% of water. Upon 248 nm excita-
tion and detection at ≈330 nm, the curve reaches a maximum
and decreases on going to neat water (results not shown). From
the above described behaviour in neat acetonitrile together with
that in neat aqueous solution it is clear that the maximum
originates from a change in formation of two different transi-
ents, the triplet states of the R–PA and the corresponding
photoproducts, TPA and TBA, respectively.
H-atom abstraction
In propan-2-ol a long-lived transient (T ) was observed for
K
the R–PA throughout (Fig. 4b). For R = OCH , CH , Cl, Br
In order to reduce the probability for excitation of R–BA
sufficiently, we used R–PA at 10–100 times higher concen-
trations, λexc = 308 nm. Then the initial dependence of kobs
vs. [H O] is the same as for λ = 248 nm (Table 5), but no
3
3
and CN the λmax values are 330, 320, 318, 324 and 320 nm,
respectively. In contrast to the triplet, TK decays by second-
order kinetics; half-lives (t1/2) range from 100 µs (Br–PA) to
about 1 ms (OCH – and CH –PA) under our conditions. T is
2
exc
3
3
K
maximum was found for the R–PA (Fig. 5b); kobs increases
2
674
J. Chem. Soc., Perkin Trans. 2, 1999, 2671–2680

View Article Online
Ϫ6
Ϫ1
a
Table 5 Triplet decay kinetics (10 kobs/s ) in the presence of water
Vol. H O (%)
2
Compound
Solvent
λexc/nm
1
5
10
15
5
6.6
у30
20
OCH –PA
Acetonitrile
248
0.4
1.3
1.1
5.5
3.4
<0.8
3.5
3.8
16
9
40
3
CH –PA
0.3
0.3
0.08
0.2
0.6
0.6
3
3
08
b
PA
F–PA
Cl–PA
Br–PA
CN–PA
248
354
12
22
>60
>50
>30
50
OCH –PA
Acetone
0.3
0.3
3.8
10
30
30
0.4
8
24
>50
>50
3
CH –PA
0.3
1.5
0.8
2
13
39
3
b
PA
Cl–PA
Br–PA
CH –PA
PA
Cl–PA
Br–PA
CN–PA
Acetic Acid
308
248
308
0.12
0.01
0.4
0.12
4
0.2
0.5
1
1.5
3.3
2.5
3
b
0.2
0.9
0.4
0.7
1.2
1.2
>50
a
b
In argon-saturated solution at 25 ЊC. Taken from ref. 14.
Fig. 6 Effect of pH on the rate constant for decay of ∆κ for CN– (᭺),
Br– (᭛), Cl– (᭝), CH – (᭞) and OCH –PA (ᮀ) in argon-saturated
3
3
aqueous solution; insets: conductivity change of OCH –PA at pH
3
9
.1 (lower) and pH 11.1 (upper); λexc = 248 nm.
pH = 10) the short-time dependence remains virtually un-
changed, i.e. ∆κ becomes positive and remains constant
between 1 and >100 µs. A small decrease of ∆κ within a few
Fig. 5 Plots vs. the H O concentration (pH = 7) of (a) ∆A(λmax) for F–
2
PA (᭿) and Cl–PA (᭡) and (b) kobs(TPA) for CN– (᭺), Br– (᭛), Cl– (᭝),
F– (ᮀ) and CH –PA (᭞) in argon-saturated acetonitrile; λ = 308 nm.
100 ns (in some cases, λexc
= 248 nm) is ascribed to photoioniza-
3
exc
14
tion and subsequent neutralization. This is supported by the
result that the relative contribution of this side reaction
decreases with decreasing laser intensity. On a long time scale
7
Ϫ1
monotonically up to the detection limit of about 5 × 10 s
.
The triplet yield of the R–PA, which is proportional to ∆A(PA),
depends also on the amount of water. It decreases with in-
creasing water content in acetonitrile as has been reported
∆κ decreases below the initial value prior to the pulse (Fig. 6,
insets) and the decay becomes significantly pH dependent. The
rate constant of this slow decay (k , see Discussion section)
14
7
for the parent molecule and shown in Fig. 5a for two cases.
Ϫ
increases linearly with the logarithm of the OH concentration
(
Fig. 6).
Transient conductivity
A positive ∆κ signal concomitant with the laser pulse was
A conductivity change concomitant with the laser pulse was
observed for the R–PA in argon-saturated aqueous solutions
in a broad pH range. At pH = 2 the signal (∆κ: positive for
increasing conductivity) decreases within 30 ns and remains con-
stant for seconds. The rate constant for the fast decrease
becomes measurable at pH = 3 and increases further to 0.2–1 µs
on going to pH = 4. The signal remains constant on a pro-
longed time scale (up to seconds). However, at pH = 6 the
conductivity signal increases within 1 µs and decreases, thereby
approaching the initial value within seconds.
observed for the R–PA in neat argon-saturated acetonitrile
solutions. The decay to the initial value (∆κ = 0) within 2–20 µs
indicates that no longer-lived charged species is formed. On
addition of water in amounts of 2–20%, the half-life of ∆κ is
reduced and the amplitude then becomes negative, i.e. smaller
than prior to the pulse. On a prolonged time scale the con-
ductivity remains constant between a few µs and about 1 s.
The rate constant for the decay of ∆κ shows also a non-
linear dependence on the H O concentration (Fig. 7) but the
2
absolute values are smaller than the kobs values for triplet decay
(Fig. 5b).
Ϫ
When the OH concentration is further increased (up to
J. Chem. Soc., Perkin Trans. 2, 1999, 2671–2680
2675

View Article Online
a
Table 6 Quantum yields of decarboxylation in mixtures with water
Vol. H O (%)
2
b
c
Compound
Solvent
Method
0
1
5
10
25
OCH –PA
Acetonitrile
A
A
A
B
A
B
A
B
A
B
A
0.34
0.31
0.86
0.66
0.6
0.25
0.79
0.6
0.8
0.50
0.69
3
CH –PA
0.17
3
d
PA
0.1
0.4
F–PA
0.54
<0.1
0.3
0.6
0.6
0.67
0.4
Cl–PA
Br–PA
CN–PA
0.2
0.11
0.70
0.65
0.47
0.75
0.65
0.8
0.53
0.71
0.25
0.53
CH –PA
Acetone
A
B
B
B
B
0.05
<0.05
<0.05
0.65
0.6
0.65
0.4
0.78
0.7
3
F–PA
Cl–PA
Br–PA
0.5
0.5
0.4
0.75
0.8
a
b
Ϫ3
In argon-saturated solution. A: λirr = 313 nm, GC analysis, B: λ = 366 nm, HPLC analysis, typical concentration: 2.5 and 50 mmol dm
,
irr
c
d
respectively. Vol % H O = 0.02 ± 0.01. Taken from ref. 14.
2
a
Table 7 Quantum yields of decarboxylation in aqueous solution at different pH
pH
b
c
Compound
Method
0
1
2
Nat.
5
OCH –PA
C
C
A
A
C
A
0.13
0.2
0.38
0.57
0.3
0.28
0.29
0.3
0.25
0.15
0.09
0.36
0.1
0.15
<0.07
0.03
3
CH –PA
0.2
0.3
0.4
0.4
0.35
0.31
0.35
0.3
0.15
0.21
0.46
0.1
0.10
0.26
0.12
0.13
0.05
3
PA
d
0.3
F–PA
Cl–PA
0.1
0.04
0.24
<0.05
0.04
d
A
C
C
C
Br–PA
CN–PA
0.3
0.12
0.08
0.02
0.18
a
b
Ϫ3
In argon-saturated solution. A: λirr = 313 nm (GC analysis), C: λ = 254 nm (UV absorption), typical concentration: 2.5 and 0.1 mmol dm
,
irr
c
Ϫ3
d
Ϫ3
respectively, unless indicated otherwise.
sensitized.
λ
= 366 nm for PA (concentration = 50 mmol dm ) at natural pH ≈2.6–2.8. Acetone (2 mol dm
)
irr
corresponding to the decomposition of the R–PA. The R–PA
exhibit a characteristic dependence of Φ(ϪCO ) on the water
2
concentration. Typically, Φ(ϪCO ) is low in neat acetonitrile,
2
reaches a maximum at 5–20% water and decreases further.
The photoreaction of OCH –PA forms a very special case.
3
Here, in neat acetonitrile or with low water content the products
are mainly anisic acid (λirr = 313 nm), some 4,4Ј-dimethoxy-
benzil, but only traces of anisaldehyde, as shown by HPLC
analysis. The quantum yield of disappearance (0.09 and 0.38
in acetonitrile and water, respectively) and that of anisalde-
hyde formation (0.002 and 0.35) increase considerably with
increasing water content, while those of anisic acid (0.08
and 0.03) and benzil (0.02 and <0.01) formation decrease.
Actually, only at >50% water, the aldehyde is the main product
besides some anisic acid (no benzil formed). The course of
the reaction at low water content is still not known. It was
established that in mixtures with D O, naturally, the formyl-
2
Fig. 7 Rate constant for decay of ∆κ vs. the H O concentration for
2
deuteriated aldehyde (BA) –OCH3 is formed which, how-
D
CN– (᭺), Br– (᭛), Cl– (᭝), CH – (᭞) and OCH –PA (ᮀ) in argon-
3
3
22
ever, cannot be obtained by direct photodeuteriation of
anisaldehyde.
saturated acetonitrile; λexc = 308 nm.
The effect of pH on Φ(ϪCO ) in aqueous solution is
2
Continuous irradiation
summarized in Table 7. The highest values were found in the
strongly acidic pH range. However, they are somewhat lower
than in the optimum water–acetonitrile mixtures. For F–,
Several R–PA (CH , Cl, F, Br and CN) readily undergo photo-
3
decarboxylation in acetonitrile–water yielding the correspon-
ding BA with high Φ(ϪCO ) (Table 6). The concentration of
Cl–, Br– and CN–PA Φ(ϪCO ) decreases with increasing pH,
2
2
both products increases as a function of the irradiation time,
showing a steep decrease around pH = 1 (Fig. 8). Photo-
2
676
J. Chem. Soc., Perkin Trans. 2, 1999, 2671–2680

View Article Online
As species involved in the transient conductivity of
R–PA in acetonitrile in the presence of small amounts of water
Ϫ3
ϩ
(
<1 mol dm ), we propose the strongly conducting H ion and
the triplet state of the anion according to reaction (4).
The recovery of the conductivity increase involves the triplet
state and is explained by efficient intersystem crossing and fast
recombination.
Photodecarboxylation involving the R–PA anion
It has been suggested that the initial increase of Φ(ϪCO2)
on increasing the H O concentration in acetonitrile (Table 6) is
2
14
due to formation of the triplet anion. The increasing amount
of water shifts equilibrium (2) towards the anion, which on
excitation yields the triplet of the anion in high yield. Once this
triplet is formed decarboxylation occurs via overall reaction (5).
Fig. 8 Relative quantum yield for decarboxylation in argon-saturated
aqueous solution as a function of pH for CN– (᭺), Br– (᭛), Cl– (᭝, ᭡),
F– (ᮀ) and CH –PA (᭞) (direct: open symbols, acetone sensitized: full);
3
λirr = 248 nm.
decarboxylation was also studied under sensitized irradiation
conditions. Larger Φ(ϪCO ) values were found for R–PA with
2
several sensitizers (e.g. acetone, acetophenone, 4,4Ј-dimethoxy-
benzophenone) than in their absence in the range pH = 2–6.
Here, the results with acetone are presented (full symbols in
Fig. 8).
This interpretation holds also for the conductivity results in
mixtures with acetonitrile. The initial increase and decrease in
∆
κ whose lifetime (<1 µs) and yield decrease with increasing
ϩ
water content (Fig. 7) arises from the release of H according to
reaction (4). The gradual increase in ∆κ and the subsequent
decrease at pH = 2–5 is ascribed to elimination of OH ,
followed by the neutralization reaction.
Discussion
Primary photochemical processes
Ϫ
The photochemistry of the R–PA is governed by the equi-
librium between the acid and the anion in the ground state.
Ϫ
ϩ
OH ϩ H
H O
2
(6)
Whether or not deprotonation takes place also from the
singlet state or only from the triplet state was previously an
open question. For 1- and 2-naphthylglyoxylic acids, where
fluorescence appears, in contrast to most R–PA, it has been
shown that the singlet state is non-reactive and that the triplet
pathway accounts for the Φ(ϪCO ) values of 0.2–0.4 in
mixtures of acetonitrile and water.
2
In organic solvents in the absence of water, the R–PA exist in
the acid form. Addition of water shifts the equilibrium towards
16
Reaction (5) can further be specified on the basis of conduc-
tivity measurements. The increase in ∆κ in aqueous solution at
the anion since the pK values of all the R–PA in neat aqueous
a
14,23
solution are around 1.2
In solvents of low polarity PA is
The triplet state of the R–PA in
Ϫ
pH = 8–11 (Fig. 6) is ascribed to OH elimination within the
4–26
no longer monomeric.
pulse width. The reason for the fast pH-dependent decrease
water-free solutions is well characterized by results from
phosphorescence and T-T absorption measurements. The
triplet nature of TPA is supported by (i) the absorption
spectra (Figs. 4a and 4c), (ii) lifetime measurements from
phosphorescence and T-T absorption (Table 2), (iii) the effect
of quenching of the decay of TPA by oxygen (Table 2) and
of κ in the acidic range is the larger equivalent conductivity
ϩ
2
Ϫ1
Ϫ1
Ϫ
2
of H (Λ = 350 cm Ω mol ) relative to OH (Λ = 190 cm
Ϫ1
Ϫ1
Ϫ
Ω
mol ). Thus CO and OH are suggested to be formed
2
Ϫ
within the laser pulse width throughout and OH remains as
conducting species in the basic range or is neutralized at lower
pH.
(
iv) the effects of other quenchers such as propan-2-ol (Figs. 3
The slow decrease of ∆κ in the 10 ms–10 s range at pH = 9–11
Fig. 6) is ascribed to formation of bicarbonate, reaction (7),
and Table 3). H-atom abstraction from propan-2-ol (SH),
28,29
(
27
yielding the α-carboxy-α-hydroxybenzyl (ketyl) radical
transient T ), occurs according to reaction (3).
which has2 aϪ1 markedly smaller equivalent conductivity
(
Ϫ1
Ϫ
K
(Λ р 50 cm Ω mol ) than that of OH .
Ϫ
Ϫ
CO ϩ OH
HCO3
(7)
2
3
Ϫ1 Ϫ1
A value of k = 7000 dm mol
s
is obtained from the slope
7
3
in Fig. 6 for several R–PA. A comparable value, k = 6900 dm
s , has been reported on the basis of a pulse radiolysis
study. The amplitude (at ca. 1 ms after the pulse) is a measure
7
Ϫ1 Ϫ1
mol
28
The k2-p values, obtained from the linear dependence of kobs
of the relative Φ(ϪCO ) value. This latter value is similar
2
vs. the propan-2-ol concentration (Table 3), are a direct measure
for the reactivity of the R–PA triplet against H-atom
abstraction.
for CH –, F– or Br–PG at pH 10 and somewhat larger for
3
OCH –PG, in agreement with the steady state values at pH = 5
3
(Table 5). Moreover, the yield was found to be roughly the same
J. Chem. Soc., Perkin Trans. 2, 1999, 2671–2680
2677

View Article Online
Scheme 1
for F–PG at pH = 3.5 and 10, using the amplitudes and the
equivalent conductivities of H and OH .
times larger in the presence than in the absence of the sensitizer
ϩ
Ϫ
(Table 7). This is in line with the low yield of triplet anion
formation due to the competing quenching reaction (8)
(Scheme 1) on direct excitation, but its efficient formation when
the triplet state of the anion is generated by energy transfer.
Effect of water
The non-linear dependence of kobs on the H O concentration
2
(
Fig. 5b) is suggested to result from a water assisted photo-
Photodecarboxylation involving the R–PA acid
process. Addition of water increases the amount of anion by
shifting the ground state equilibrium (2) to the right hand side.
In the absence of water the observed triplet of R–PA (except for
The enhancing effect on Φ(ϪCO ) in neat aqueous solution in
2
ϩ
the range pH = 0–3 with H concentration can be explained by
14
OCH ) in acetonitrile has probably n,π* character and the R–
postulating a decarboxylation step from the triplet of the acid.
3
PA are present as acids. Quenching of triplet R–PA by a single
water molecule is not efficient since a small amount of water
Ϫ3
(
<0.5 mol dm ) has only a small effect on the decay kinetics,
whereas kobs depends almost linearly on the H O concentration
2
Ϫ3
in the 0.5–10 mol dm range, indicating that the quenching
step involves more than one water molecule.
As the two triplet forms are probably not distinguishable by
T-T absorption spectroscopy, we cannot exclude that the π,π*
character of the triplet state increases on addition of water.
Hydrogen bonding to the carboxyl group may also play a role.
As quenching step we propose reaction (5) of the triplet state
of the anion with water, thereby forming the benzoyl anion or
its protonated isomer as short-lived intermediate (Scheme 1).
For all R–PA examined by nanosecond laser flash photolysis
in aqueous solution (e.g. at pH = 0) no transient could be
detected which may be assigned to the triplet state of the acid.
Excitation of the acid yields only the triplet state of the corre-
sponding BA (see Results section). This result, in contrast to
the features of the triplet of the acid in non-aqueous solution
Oxygen has only a small effect on Φ(ϪCO ) in mixtures with
2
water. This is readily explained by the proposed mechanism
since the reaction with water competes efficiently with triplet
quenching by oxygen.
(
Table 2), indicates that reaction (9) is faster than 20 ns.
The effect of pH on Φ(ϪCO ) for the R–PA is similar to that
2
The decrease of Φ(ϪCO ) on going to 100% water (cf. Tables
2
14
of the parent molecule. The general trend is a decrease of
6
and 7) is suggested to be due to quenching of the singlet
Φ(ϪCO ) with pH, e.g. on going from pH Ϫ1 to ϩ5 (Table 7),
14
2
state, reaction (8) in Scheme 1. This reaction with water
reduces the yield of the observable triplet, as is illustrated in
Fig. 5a for two cases. In the presence of more than about
which is suggested to be due to three effects, (i) the shift
of the ground state equilibrium (2) with pK = 1.1, (ii) the
a
reduction of the yield of triplet anion for pH > 1, reaction (8)
and (iii) the efficiency of the decarboxylation reaction (9) at
pH р pKa.
2
0% water, due to the above effect of water on kobs, the triplet is
not accessible with our methods. Involvement of hydrates in the
excited singlet state is hypothesized. Photodecarboxylation
under acetone-sensitized irradiation supports the proposed
Reactivity of the triplet and effects of substitution
mechanism since Φ(ϪCO ) is strongly enhanced compared
2
to otherwise the same conditions. For example, for CH – and
Sequence (9) is in line with the observed pH dependence of
3
Br–PA in aqueous solution at pH = 5 Φ(ϪCO ) is about ten
Φ(ϪCO ) for the R–PA (CH , F, Cl, Br, CN) taking into
2
2
3
2
678
J. Chem. Soc., Perkin Trans. 2, 1999, 2671–2680

View Article Online
a reaction of the triplet anion with water, yielding CO and
2
the corresponding BA. However, a reaction of water with the
excited singlet anion reduces Φ(ϪCO ) in neat water. The
2
increase of Φ(ϪCO ) with decreasing pH is ascribed to
2
the decarboxylation reaction involving the triplet state of the
acid form of the R–PA.
Experimental
The substituted PA were prepared by conventional
Ϫ1
1
1–15
3
procedures.
Molar absorption coefficients (in dm mol
Ϫ1
4
cm × 10 ) in acetonitrile and (in parentheses) aqueous
solution are, R: OCH ε = 1.42 (ε288 = 1.65); CH ε = 1.18
3
293
3
265
(
ε263 = 1.40); F ε251 = 0.94 (ε255 = 1.27); Cl ε265 = 1.26 (ε263 = 1.25);
Br ε267 = 1.48 (ε267 = 1.78); CN ε254 = 1.68 (ε255 = 1.92). Solvents
Merck, Darmstadt): were p.a. (acetic acid, ethylene glycol,
ethanol, propan-2-ol and the other alcohols), Uvasol quality
acetone, acetonitrile, heavy water) or distilled (butyronitrile,
(
(
Fig. 9 Semilogarithmic plots of k2-p (open symbols) for triplet quench-
ing by propan-2-ol and kH2O (filled symbols) for quenching by water vs.
the Hammett σ parameter in acetone (triangles), acetonitrile (circles)
and acetic acid (squares) using data from Tables 4 and 5.
Fluka). Water was purified by a Millipore (Milli Q) system.
The output at 248 or 308 nm from one of two excimer lasers
(
Lambda Physik, EMG 201 MSC and EMG 200, respectively)
was used for most transient spectroscopic and kinetic
measurements. Excitation at 248 nm has the advantage of the
extension of the spectra to the shortest wavelength possible,
whereas with λexc = 308 nm photoionization can be avoided. For
a few measurements also the third harmonic of a Nd-laser
account that the decarboxylation of the triplet anion is reduced
by the efficient quenching step (8). For R = F, Cl and Br
Φ(ϪCO ) is 0.3–0.5 at pH = 0.4–1.3, while it drops to approxi-
2
mately 0.02 at pH ӷ pK (Fig. 8 and Table 7).
a
(
λexc = 354 nm) was used. The laser flash photolysis apparatus
The dependence of log k2-p vs. the Hammett σ parameter
and the procedures used were essentially the same as in previous
(
Fig. 9) reveals an increase of k2-p on increasing the electron
14,16,30
work.
The fast (0.05–10 µs) and slow (5 µs–10 s) con-
accepting power of the substituent, i.e. in the order
ductivity signals were measured by DC and AC bridges
OCH , CH , H, F, Cl, Br and CN. This change is quite re-
3
3
using appropriate transient digitizers (Tektronix 7912AD
markable since k2-p increases by more than three orders of
14,31
and 390AD, respectively).
The hydrated electron could be
magnitude on going from OCH –PA to CN–PA. For a given R–
3
observed for most R–PA in aqueous solution at pH = 3–10 at
high laser intensities (λexc = 248 nm). It was identified by the
PA the k2-p value in acetone is similar to that in acetonitrile or
only slightly larger. Owing to the small k2-p value for OCH one
3
reaction with N O or oxygen and its yield decreases in the order
2
may expect that T should not be formed in acetonitrile, but the
K
OCH –, CH – and F–PA. The occurrence of photoionization
3
3
possibility of intermolecular or intramolecular H-atom
abstraction in this case cannot be excluded.
Generally, the results from triplet quenching by alcohols
are in agreement with the assumption that in inert solvents all
R–PA are present in the n,π* triplet state of the acid. This
via a biphotonic step is not surprising since the hydrated
electron was also observed with several substituted BA under
31
comparable conditions. The hydrated electron was not
observable using λexc = 308 nm, confirming the biphotonic side
process with 10 eV excitation energy for λexc = 248 nm.
is different for OCH –PA, where the triplet energy is virtually
3
Continuous irradiation was performed with the 366 and
the same as for other R–PA, but (i) the photoproducts, (ii) the
3
13 nm lines of a 1000 W Xe/Hg-lamp combined with a
pH dependence of the Φ(ϪCO ) values (Table 7) and (iii) the
2
12,13
monochromator
lamp (Gräntzel). The quantum yield Φ(ϪCO
or 366 nm was determined by means of an electronic
or at 254 nm using a low-pressure Hg
T-T absorption spectrum (Fig. 4c) are quite different. Thus
3
2
) with λirr = 313
the π,π* state is favoured for OCH –PA.
3
Addition of water to an organic solvent gives rise to the
non-Stern–Volmer behaviour in the quenching of both T-T
absorption and phosphorescence (Figs. 3a and 3b). Neverthe-
less, the rate constants kH2O, estimated from the linear part of
Fig. 5b, may be used as a measure for the reactivity of triplet
quenching by water. These rate constants show a trend to
increase in the order R: OCH , CH , F, H, Cl, Br and CN,
12,13
actinometer.
Actinometry at 254 nm was carried out using
32
uridine in air-saturated aqueous solution at pH = 6, Φ = 0.016.
Three methods for detection of photoproducts were used here;
Φ(ϪCO ) is derived from CO measurements using GC and
2
2
HPLC analysis, A and B, respectively and with method C
Φ(ϪCO ) is derived from changes in the absorption spectrum
2
3
3
at appropriate wavelengths (typically longer than λ ). The
irr
covering a range of more than three orders of magnitude (Table
). The dependence of log kH2O vs. σ (Fig. 9) is essentially linear
samples were purged by argon and measured at 25 ± 1 ЊC unless
4
otherwise indicated. GC and HPLC analyses were essentially
with the exception of the case OCH –PA. A similar dependence
3
14,16
the same as reported previously.
on the substituent was found in acetone with slightly larger
kH2O values than in acetonitrile. A corresponding dependence
of log kH2O vs. σ was also measured in acetic acid, where,
however, the kH2O values are typically tenfold smaller than in
acetonitrile.
Phosphorescence quenching was carried out on a spectro-
fluorimeter (Perkin Elmer LS-5) with small amounts of
additives (up to 5 vol %) during continuous purging with argon.
The quantum yield was determined on a Spex-Fluorolog
(
corrected spectra) by using benzophenone in butyronitrile at
Ϫ196 ЊC as reference and Φ = 1. The phosphorescence decay
p
Concluding remarks
of the R–PA in acetonitrile, observed after laser excitation
(λexc = 248, 308 and 354 nm), depends somewhat on the experi-
mental conditions. On increasing the laser intensity the half-life
decreases. Using 248 nm and a typical concentration of 0.1–0.5
The triplet state of six para-substituted PA in acetonitrile,
acetone or acetic acid is characterized by results from time-
resolved and steady-state phosphorescence and T-T absorption
measurements. The corresponding carboxyhydroxybenzyl
radical is generated by an H-atom abstraction reaction from
Ϫ3
mol dm , the phosphorescence lifetime is in the 1–3 µs range.
Compared to the intensity-induced effect of T-T annihilation,
the effects of inert solvents, λexc and the concentration are
smaller. The latter is demonstrated for several R–PA in acetone
propan-2-ol. The increase of Φ(ϪCO ) on addition of water to
acetonitrile at room temperature is suggested to be due to
2
J. Chem. Soc., Perkin Trans. 2, 1999, 2671–2680
2679

View Article Online
(
λexc = 354 nm); although much larger concentrations had to be
12 H. J. Kuhn and A. Defoin, EPA Newsletter, 1986, 26, 23.
1
3 A. Defoin, R. Defoin-Straatmann, K. Hildenbrand, E. Bittersmann,
used, τ is also in the 10–30 µs range (not shown). The intensity
at the maximum (I ) and τ are markedly reduced by oxygen
(
most R–PA). When water was added in small amounts to a R–
PA in acetonitrile, acetone or acetic acid, the phosphorescence
spectrum remained unchanged, but I and τp were reduced.
However, the plots of I /I and τ /τ vs. the H O concentration
are initially non-linear and nearly linear only above a certain
water concentration. Deviations from linear Stern–Volmer
dependences in the range I /I = 2–4 are indicated.
p
D. Kreft and H. J. Kuhn, J. Photochem., 1986, 33, 237.
4 H. J. Kuhn and H. Görner, J. Phys. Chem., 1988, 92, 6208.
5 H. J. Kuhn and H. Görner, XII IUPAC Symposium Photochem.,
p
p
1
1
e.g. down to <1/10 in air-saturated acetonitrile solutions of
1
988, Bologna, 736.
16 C. Laurich, H. Görner and H. J. Kuhn, J. Photochem. Photobiol. A:
Chem., 1998, 112, 29.
17 E. Krogh and P. Wan, J. Am. Chem. Soc., 1992, 114, 705.
p
o
o
p
p
p
p
2
1
1
2
8 D. Shulka and P. Wan, J. Photochem. Photobiol. A: Chem., 1998,
1
13, 53.
9 S. Hu and D. C. Neckers, J. Photochem. Photobiol. A: Chem., 1998,
114, 103.
o
p
p
0 S. Hu and D. C. Neckers, J. Photochem. Photobiol. A: Chem., 1998,
1
18, 75.
21 D. Budac and P. Wan, J. Photochem. Photobiol. A: Chem., 1992, 67,
35.
Acknowledgements
We thank Professor K. Schaffner for his support and Mr L. J.
Currell and Ch. Laurich for technical assistance.
1
2
2 A. Defoin, R. Defoin-Straatmann and H. J. Kuhn, Tetrahedron,
1984, 40, 2651.
2
2
3 M. S. Wheatley, Experientia, 1956, 12, 339.
4 G. Oehme, G. Fischer and A. Schellenberger, Chem. Ber., 1967, 100,
4
25.
References
2
5 A. Rindt, A. Schellenberger, G. Oehme and R. Pinske, Tetrahedron,
1973, 29, 3051.
26 G. Fischer, G. Oehme and A. Schellenberger, Tetrahedron, 1971, 27,
5683.
27 J. C. Scaiano, J. Photochem., 1973/74, 2, 81.
28 M. N. Schuchmann and C. von Sonntag, Z. Naturforsch., B: Anorg.
Chem., Org. Chem., 1982, 37b, 1184.
29 D. M. Kern, J. Chem. Educ., 1960, 37, 14.
30 H. Görner and H. J. Kuhn, J. Phys. Chem., 1986, 90, 5946.
31 H. Görner. L. J. Currell and H. J. Kuhn, J. Phys. Chem., 1991, 95,
5518.
1
2
L. Claisen, Ber. Dtsch. Chem. Ges., 1877, 10, 844.
P. A. Leermakers and G. F. Vesley, J. Am. Chem. Soc., 1963, 85,
3
776.
3
4
Y. Sawaki and Y. Ogata, J. Am. Chem. Soc., 1981, 103, 6455.
D. S. Kendall and P. A. Leermakers, J. Am. Chem. Soc., 1966, 88,
2
766.
5
T. Fujisawa, B. M. Monroe and G. S. Hammond, J. Am. Chem. Soc.,
970, 92, 542.
1
6
7
8
9
T. Fujisawa and G. S Hammond, J. Am. Chem. Soc., 1972, 94, 4175.
B. M. Monroe, Adv. Photochem., 1971, 8, 77.
G. L. Closs and R. J. Miller, J. Am. Chem. Soc., 1978, 100, 3483.
S. Emori, J. P. Colpa and J. K. S. Wan, Chem. Phys. Lett., 1983, 98,
32 G. J. Fisher and H. E. Johns, Photochemistry and Photobiology
of Nucleic Acids, ed. S. Y. Wang, Academic Press, New York, 1976,
p. 169.
1
42.
0 L. J. Mittal, J. P. Mittal and E. Hayon, J. Phys. Chem., 1973, 77,
482.
1
1
1
1 A. Defoin, R. Defoin-Straatmann and H. J. Kuhn, J. Labelled
Compd. Radiopharm., 1982, 19, 891.
Paper 9/06567I
2
680
J. Chem. Soc., Perkin Trans. 2, 1999, 2671–2680