Chemistry of photogenerated α-hydroxy-p-nitrobenzyl carbanions in aqueous solution: Protonation vs. disproportionation

Article abstract of DOI:10.1139/v03-045

The photochemistry of p-nitrobenzyl derivatives 6-10 has been studied in aqueous solution as a function of pH, using product analysis, UV-vis spectrophotometry, and laser flash photolysis (LFP). The compounds were chosen with the aim of further exploring the propensity of these systems to give rise to α-hydroxy-p-nitrobenzyl carbanions on photolysis, and to study their mechanisms of subsequent reaction, α-Hydroxy-substituted carbanions are anions that cannot be readily formed using thermal routes but which are believed to have some interesting chemistry. Three methods were employed for photogenerating these carbanions: (i) decarboxylation; (ii) retro-Aldol reaction; and (iii) carbon acid deprotonation. All three methods proved to be successful using the p-nitrobenzyl chromophore. Photogenerated α-hydroxy-p-nitrobenzyl carbanions react via disproportionation, giving rise to oxidized and reduced products; simple protonation of the anion was undetectable.

Full text of DOI:10.1139/v03-045

586
Chemistry of photogenerated ␣-hydroxy-p-
nitrobenzyl carbanions in aqueous solution:
protonation vs. disproportionation
James Morrison, Peter Wan, John E.T. Corrie, and V. Ranjit N. Munasinghe
Abstract: The photochemistry of p-nitrobenzyl derivatives 6–10 has been studied in aqueous solution as a function of
pH, using product analysis, UV–vis spectrophotometry, and laser flash photolysis (LFP). The compounds were chosen
with the aim of further exploring the propensity of these systems to give rise to α-hydroxy-p-nitrobenzyl carbanions on
photolysis, and to study their mechanisms of subsequent reaction. α-Hydroxy-substituted carbanions are anions that
cannot be readily formed using thermal routes but which are believed to have some interesting chemistry. Three meth-
ods were employed for photogenerating these carbanions: (i) decarboxylation; (ii) retro-Aldol reaction; and (iii) carbon
acid deprotonation. All three methods proved to be successful using the p-nitrobenzyl chromophore. Photogenerated α-
hydroxy-p-nitrobenzyl carbanions react via disproportionation, giving rise to oxidized and reduced products; simple
protonation of the anion was undetectable.
Key words: photodecarboxylation, nitrobenzyl carbanions, photoredox, nitroaromatic compounds, excited-state carbon
acids.
Résumé : Opérant en solutions aqueuses en fonction du pH et faisant appel à l’analyse des produits, à la spectrophoto-
métrie UV–vis et à la photolyse éclair au laser, on a étudié la photochimie des dérivés p-nitrobenzyles (6–10). On a
choisi les composés dans le but de pouvoir explorer plus à fond la propension de ces systèmes de donner naissance à
des carbanions α-hydroxy-p-nitrophényles par photolyse et d’étudier les mécanismes de leur réaction subséquente. Les
carbanions portant un substituant α-hydroxy sont des anions qui ne peuvent pas se former facilement par les voies ther-
miques, mais que l’on soupçonne d’avoir des propriétés chimiques intéressantes. On a utilisé trois méthodes pour la
photogénération de ces carbanions: (i) la décarboxylation; (ii) la réaction rétroaldolique et (iii) la déprotonation d’un
acide carbonique. Ces trois méthodes ont été couronnées de succès avec le chromophore p-nitrobenzyle. Les carbanions
α-hydroxy-p-nitrobenzyles photogénérés réagissent par une réaction de métathèse qui conduit à la formation de produits
oxydés et de produits réduits; on n’a pas pu mettre en évidence de protonation simple de l’anion.
Mots clés : photodécarboxylation, carbanions nitrobenzyles, photorédox, composés nitroaromatiques, état excité d’acides
carbonés.
[Traduit par la Rédaction] Morrison et al. 597
Introduction
an hydroxy group (an oxygen acid) is directly attached to
the carbanion, are an interesting type of carbanion whose
formation and stability has been the subject of theoretical in-
terest (2, 3). However, for obvious reasons, the breadth of
experimental knowledge of these carbanions is quite limited.
Nonetheless, α-hydroxy carbanions have been proposed as
intermediates in photochemical (4–6) and enzymatically cat-
alyzed (7, 8) reactions.
Carbanions are synthetically useful and fundamentally im-
portant reactive intermediates in organic chemistry (1). The
reactivity of carbanions as either nucleophiles, bases, or as
electron donors is well-known and substituents strongly in-
fluence their stability and reactivity. This is exemplified by
the enormous range of carbon acid pKa values (1) and these
data are useful in predicting the conditions necessary to gen-
erate a carbanion and its subsequent reactivity. Substituent
effects on the negatively charged intermediate may be mani-
fested through resonance, inductive effects, or by providing
a new reaction pathway. α-Hydroxy carbanions 1, in which
Received 3 January 2003. Published on the NRC Research Press Web site at http://canjchem.nrc.ca on 26 May 2003.
Dedicated to Professor Don Arnold for his contributions to chemistry.
J. Morrison and P. Wan.1 Department of Chemistry, Box 3065, University of Victoria, Victoria, BC V8W 3V6, Canada.
J.E.T. Corrie2 and V.R.N. Munasinghe. National Institute for Medical Research, The Ridgeway, Mill Hill, London, NW7 1AA, U.K.
1Corresponding author (e-mail: pwan@uvic.ca).
2Corresponding author (e-mail: jcorrie@nimr.mrc.ac.uk).
Can. J. Chem. 81: 586–597 (2003)
doi: 10.1139/V03-045
© 2003 NRC Canada

Morrison et al.
587
Scheme 1.
To study “free” carbanions in solution it is advantageous
to generate appreciable amounts of these intermediates
quickly. A well-known route is via photochemical reaction
(9), preferably one that is fast and efficient. The simplest
conceivable photochemical route to a carbanion is via a
heterolytic C—C or C—H bond cleavage, for example via
photodecarboxylation or photodeprotonation, respectively.
Nitrobenzyl compounds, with their electron-withdrawing ni-
tro group, are particularly adept at inducing these photo-
cleavage reactions, generally via the triplet excited state (10,
11). While it is well-known that o-nitrobenzyl systems are
highly photoreactive (generally via initial transfer of the
benzylic hydrogen to the nitro group), the m- and p-isomers
also display a range of photochemical reactivity not often re-
ported for the o-isomer (10, 11). Both m- and p-nitrobenzyl
compounds are known to undergo a variety of photofrag-
mentation reactions, such as photodecarboxylation (12),
photodeprotonation (4), and photoretro-aldol reaction (13),
all of which have been proposed to generate m- and p-
nitrobenzyl carbanions as intermediates. However, prior to
our studies, only the parent p-nitrobenzyl carbanion was ob-
served by transient methods (vide infra).
Margerum and Petrussis (12) were the first to investigate
the chemistry of photogenerated p-nitrobenzyl carbanion (2),
which they prepared by photodecarboxylation of p-nitro-
phenylacetic acid (3) (Scheme 1). The observed chemistry of
photogenerated 2, to give p-nitrotoluene (simple proto-
nation) and p,p′-dinitrobibenzyl (4) (disproportionation prod-
uct), may be described as “conventional” p-nitrobenzyl
carbanion chemistry since thermally generated 2 (formed by
deprotonation of p-nitrotoluene in tert-butyl alcohol – KOH)
has also been observed to give 4 (14–16). Carbanion 2 is
strongly stabilized by the p-nitrophenyl ring and is thus
readily observable by UV–vis spectrophotometry as well as
by µs and faster laser flash photolysis (LFP) (λmax 358 nm;
τ = 53 s at pH 13) (12). The complex decay kinetics of car-
banion 2 (mixed first- and second-order) were studied in
some detail by Craig et al. (17, 18), who explained the data
as arising via two competing pathways of reaction: (i) a sim-
ple protonation pathway to give p-nitrotoluene; and (ii) a bi-
molecular coupling pathway (of two molecules of 2),
followed by loss of two electrons (the final fate of which
was not determined) to form 4. Wan and Muralidharan (13)
discovered a second method for generating carbanion 2, by
base-catalyzed photoretro-aldol reaction of 5, with subse-
quent chemistry identical to that reported above.
solutions), via an apparent disproportionation of the photo-
generated carbanion. Disproportionation chemistry has also
been observed for photogenerated α-hydroxy-p-(p-nitro-
phenyl)benzyl carbanion in neutral solution (6). The obser-
vation of disproportionation and redox pathways is perhaps
not unexpected considering that even the parent p-
nitrobenzyl carbanion (2) displays this kind of reactivity.
Moreover, theoretical calculations (3) have shown that sim-
ple anions (in the gas phase) such as CH2OH and CH2NH2
are unstable with respect to their radicals. This suggests that
carbanions of this kind have an intrinsic propensity for
disproportionation chemistry (for example, via initial elec-
tron transfer). Some of the products described above (4, 6)
are consistent with this pathway.
–
–
The chemistry of α-hydroxy-nitrobenzyl carbanions is not
as well understood. Wan and Yates (4) first proposed water
and hydroxide-mediated photodeprotonation of the benzylic
proton of m- and p-nitrobenzyl alcohols, to give the corre-
sponding α-hydroxy nitrobenzyl carbanions (eq. [1]). In ret-
rospect, the proposal of C—H bond deprotonation in the
excited (triplet) state as a primary step was unusual, as there
was no precedent for such a pathway at the time. In addition,
reprotonation of these carbanions apparently does not occur,
as shown by the absence of deuterium exchange, but instead
redox chemistry was observed (4). p-Nitrobenzyl alcohol
was found to give a simple intramolecular redox product in
base (p-nitrosobenzaldehyde), whereas m-nitrobenzyl alco-
hol gave a mixture of m-nitrobenzaldehyde (60–70%) and
m-azoxybenzaldehyde (40–30%) (in neutral, acid, and basic
The aim of this work is to study in more detail the chem-
istry of a variety of photogenerated α-hydroxy-p-nitrobenzyl
carbanions (1, R′ = p-nitrophenyl) and related species, using
product studies and LFP. For this purpose p-nitromandelic
acid (6) and its O-acetyl derivative 7 are obvious initial sub-
strates since it is anticipated that these compounds will
undergo efficient photodecarboxylation, to give the corre-
sponding α-hydroxy and α-acetoxy-p-nitrobenzyl carban-
ions. p-Nitrophenylethylene glycol (8) would be expected to
undergo base-catalyzed photoretro-aldol reaction to give the
parent α-hydroxy-p-nitrobenzyl carbanion in basic medium.
Finally, p-nitrobenzhydrol (9), and the corresponding methyl
© 2003 NRC Canada

588
Can. J. Chem. Vol. 81, 2003
Scheme 2. Synthesis of 6 and 7. Reagents and conditions: (a) SOCl2; (b) Br2; (c) Hg(OAc)2, HOAc; (d) Cl3C(NH)O-t-Bu, CH2Cl2;
(e) CsCO3, MeOH; (f) TFA.
ether 10 are expected to give α-hydroxy (or methoxy) p-
nitrobenzyl carbanions via photodeprotonation.
Therefore, the 369 nm transient is assigned to α-acetoxy-p-
nitrobenzyl carbanion (14).
Preparative photolysis of 6 and 7 (1 × 10–3 M, 1:1 H2O–
CH3CN, Rayonet photoreactor, 300 nm) confirmed the ob-
servations from UV–vis spectrophotometry that indicated
different reaction pathways for these compounds. Low con-
version runs (<30%) of 6 at pH 7 (Ar purged) followed by
product isolation gave p-nitrobenzaldehyde (15) and p,p′-
azoxybisbenzaldehyde (16) (Scheme 3). The relative yields
of 15 and 16 (60–70% and 40–30%, respectively) were inde-
pendent of conversion (up to ~40%), consistent with both
being primary photoproducts. These are the expected
disproportionation products of an α-hydroxy-p-nitrobenzyl
carbanion intermediate, analogous to the reported products
of the photochemistry of m-nitrobenzyl alcohol (4) and p-(p-
nitrophenyl)benzyl alcohol (6). At high conversion (>50%)
insoluble orange precipitates and some soluble (but unidenti-
fied) secondary photoproducts were observed. Photolysis of
6 in oxygen-saturated solution gave mostly 15 (>90%) and
some minor unidentified products. At pH 13 (Ar purged),
low conversion (~15%) runs gave 15 and 16 (60–70% and
40–30%, respectively, as at pH 7). Photolysis at pH 3 gave
the intramolecular redox product p-nitrosobenzaldehyde (17)
in addition to 15 and 16 (relative yields 28:50:22). The for-
mation of 17 in acidic solution is reminiscent of the intra-
molecular photoredox chemistry reported for p-nitrobiphenyl
systems (6). We were not able to enhance the relative yield
of 17 since the photodecarboxylation quantum yield be-
comes negligible in solutions below pH 3, because ioniza-
tion of the carboxylic acid is suppressed. For example,
irradiation of 6 at pH 1 resulted in quantitative recovery of
substrate. These observations are consistent with a mecha-
nism in which 17 must arise via initial decarboxylation of
the ionized form of 6. The requirement of an acidic solution
for formation of 17 will be discussed later.
Results and discussion
Materials
p-Nitromandelic acid (6) is only sporadically reported in
the literature and its synthesis, unlike that of the correspond-
ing o-nitro isomer (19), is surprisingly difficult. In our hands
the complex procedure described by Carrara et al. (20),
based on formation of p-nitromandelonitrile and subsequent
acidic hydrolysis, was not reproducible. We, therefore,
adopted the route shown in Scheme 2 which yielded first the
O-acetyl derivative 7. Because of concerns about the stabil-
ity of 6 under basic conditions and possible difficulties of
isolation, we chose to prepare it via the tert-butyl ester 12.
Subsequent mild base treatment (CsCO3 in methanol) gave
the alcohol 13, from which 6 was obtained by TFA-mediated
cleavage of the tert-butyl group.
Glycol 8 was available from a previous study (21). p-
Nitrobenzhydrol (9) and its methyl ether derivative 10 were
made by standard methods (see Experimental).
Product studies
UV–vis spectrophotometry was used as a quick and infor-
mative tool to monitor the photochemistry of substrates 6–10
under various conditions and preparative photolyses were
used to identify products formed. Photolysis of p-nitro-
mandelic acid (6) as its carboxylate salt (~100% H2O, pH 7,
Ar purged, Rayonet photoreactor, 300 nm) in a UV cuvette
resulted in formation of new (permanent) absorption bands
at 289 and 308 nm (Fig. 1). No transient species was ob-
served on this time scale. Strikingly different spectral
changes that imply different photochemistry were observed
upon photolysis of the O-acetyl derivative 7 (Fig. 2). A
strong absorption band at 369 nm that was observed imme-
diately after brief irradiation (Ar purged) decayed via com-
plex kinetics over the course of several minutes (Fig. 3; the
decay cannot be satisfactorily fitted to either first- or second-
order kinetics; it is assumed to be a mixture of these compo-
nents). The overall decay was slower in D2O, consistent with
a component of the processes involving proton (deuteron)
transfer from solvent. The 369 nm absorption band was not
observed using this technique upon photolysis of 7 in oxy-
gen-saturated solution (τ < 10 s). These observations are
very similar to those reported for the 358 nm absorption
band of the parent p-nitrobenzyl carbanion (2) (12, 17, 18).
Preparative photolysis of 7 (pH 7, Ar purged) gave the
product of simple decarboxylation, i.e., p-nitrobenzyl acetate
(18) and the expected disproportionation product, α,α′-
bisacetoxy-p,p′-dinitrobibenzyl (19) (eq. [2]). The relative
yield of 18 to 19 was ~ 1:1 in H2O–CH3CN but this ratio
was dramatically altered in 1:1 D2O–CH3CN (~5% α-
deuterated 18 and ~95% 19). Thus, when the protonation
pathway (14 → 18) is retarded in D2O by a primary solvent
isotope effect, the disproportionation pathway that gives 19
becomes dominant. Photolysis of 7 in oxygen-saturated solu-
tion gave 15 (60%) and p-nitrobenzoic acid (40%); neither
18 nor 19 was observed. This indicates that interception of
© 2003 NRC Canada

Morrison et al.
589
Fig. 1. UV–vis traces observed on photolysis of 6 in ~100%
H2O (6.6 × 10–5 M, 0.01 M pH 7 phosphate buffer, Ar purged).
Each trace represents 15 s irradiation at 300 nm (4 lamps). Ar-
rows indicate changes in absorbance.
Fig. 3. Decay in absorbance of 14 photogenerated from 7 in
~100% H2O (complete within 5 min) or ~100% D2O (complete
within 15 min) measured at 369 nm immediately following 15 s
photolysis (6.7 × 10–5 M, 300 nm, 16 lamps, pH(D) 7, Ar
purged). Neither trace could be fitted satisfactorily to a first- or
second-order kinetic equation.
Fig. 2. UV–vis traces observed before (dashed line) and immedi-
ately following 15 s photolysis (300 nm, 16 lamps) of 7 in
~100% H2O (solid lines) (6.7 × 10–5 M, 0.01 M pH 7 phosphate
buffer, Ar purged). Each trace of the dark reaction is separated
by 30 s. Arrows indicate changes in absorbance during the dark
reaction.
strong transient absorption band at λmax 403 nm that
decayed within 1 h (Fig. 4) with complex kinetics (Fig. 4,
inset). The initial decay appears to obey first-order kinetics;
however, after approximately 20 min the decay becomes
(reproducibly) more complex, suggestive of contributions
from competing reactions. This transient was observed only
in basic solution (Fig. 5); no spectral changes were observed
in neutral or acidic solution (below pH 11), even on pro-
longed irradiation. The transient was also not observed in
oxygen-saturated basic solution.
Preparative photolysis of 8 (1 × 10–3 M, pH 13, Ar) gave
15 and 16 (60–70% and 40–30%, respectively), as observed
for 6 in neutral or basic solution. In keeping with the cuvette
results above, no reaction was observed at pH 7. Base catal-
ysis of the reaction of 8 is in line with observations made for
5 (13), although the latter is not known to give an observable
403 nm transient. In a previous study of 8 in our laboratory
(21), a long-lived radical anion species was photogenerated
and assigned from its ESR spectrum as p-nitrobenzaldehyde
radical anion (20). Its lifetime in the ESR spectrum was sim-
ilar to that observed for the transient 403 nm absorption
band. Radical anion 20 has previously been reported to have
an absorption maximum of ~400 nm (22) and the present
data are in good agreement with this.
14 by dissolved oxygen is much more efficient than either
protonation or disproportionation.
Irradiation of p-nitrobenzhydrol (9) at pH 13 (~100%
H2O, Ar purged, 300 nm) in a UV cuvette resulted in initial
spectral changes that were consistent with formation of 21.
However, other components were evidently present, as the
initial spectrum was not stable over time and underwent
complex changes over the course of 10–30 min in the ab-
sence of light, indicating that relatively slow dark processes
were occurring. No spectral changes were observed at pH 7,
consistent with base catalysis of photoreaction. Preparative
p-Nitrophenylethylene glycol (8) was found to be photo-
reactive only in basic solution. UV–vis traces of the reaction
(~100% H2O, pH 13, Ar purged) allowed observation of a
© 2003 NRC Canada

590
Can. J. Chem. Vol. 81, 2003
Scheme 3.
omatic signals). This is in agreement with previous studies
of m- and p-nitrobenzyl alcohols (4), in which the corre-
sponding photogenerated carbanions were also found not to
undergo deuterium incorporation at the benzylic position. p-
Nitrobenzhydryl methyl ether (10) might be expected to dis-
play “conventional” p-nitrobenzyl carbanion chemistry upon
photodeprotonation, which would lead to deuterium incorpo-
ration. However, prolonged irradiation of 10 (pD 13, 10 h)
resulted in complete recovery of substrate, with no evidence
of deuterium exchange.
Fig. 4. UV–vis traces observed before (dashed line) and immedi-
ately following 30 s photolysis (300 nm, 16 lamps) of 8 in
~100% pH 13 H2O (7.4 × 10–5 M, Ar purged). Each trace is
separated by 5 min (solid lines). Arrows indicate changes in
absorbance during the dark reaction. Inset: observed decay of the
403 nm absorption band.
photolysis of 9 (1 × 10–3 M, 1:1 H2O–CH3CN, pH 13) gave
two isolable disproportionation products, p-nitrobenzo-
phenone (21) and p,p′-azoxybisbenzophenone (22) (60–70
and 40–30% relative yield, respectively), and a third minor
(<10%, relative to 21) product that has not been fully char-
acterized but is believed to be derived from further reduction
of 22 (see Experimental) (eq. [3]). Observation of base-
catalyzed products from 9, which is reminiscent of the
photochemistry of m- and p-nitrobenzyl alcohol (4), suggests
that the reaction pathway involves initial deprotonation from
the benzylic position of 9. However, attempts to trap the α-
hydroxybenzyl carbanion intermediate from 9 with D2O (at
pD 13) were unsuccessful. Thus, extended photolysis of 9 in
1:1 D2O–CH3CN (pD 13) to >80% conversion followed by
recovery of unreacted substrate showed complete absence of
deuterium incorporation at the methine positon (as moni-
tored by 1H NMR integration of this signal relative to the ar-
The quantum yield for reaction of 6 was estimated using
p-nitrophenylacetic acid (3) (Φ = 0.6) (12, 17) as secondary
actinometer. Comparative photolysis of 3 and 6 to low con-
version (<20%) gave a quantum yield for the reaction of 6 in
100% H2O (pH 6) of 0.4 ± 0.1. Quantum yields for reaction
of 7 (pH 7), 8 (pH 14), and 9 (pH 13) were also estimated
using product analysis and were 0.4, 0.5, and 0.02, respec-
tively.
© 2003 NRC Canada

Morrison et al.
591
Fig. 7. LFP traces for 6 in ~100% H2O, pH 6.0 (4.6 × 10–5 M,
Fig. 5. Initial change in 403 nm absorbance (∆OD = OD imme-
diately after irradiation minus OD before irradiation) upon irradi-
ation of 8 in ~100% H2O, pH 1–14 (300 nm, 16 lamps, 15 s
photolysis time, 7.4 × 10–5 M, Ar purged).
0.01 M phosphate buffer, N2 purged, λex 266 nm). Diamonds rep-
resent the spectrum sampled 700 ns after the laser pulse, which
includes α-hydroxy carbanion 25 at 360 nm. Squares and triangles
represent the decay of the spectrum at 2.6 and 6.9 µs after the la-
ser pulse, respectively. Final spectrum after the decay (sampled at
36 µs) of 25 to the product at 340 nm is represented by circles.
Top inset: decay of carbanion 25 (kobs = 2.6 × 105 s–1, monitored
at 360 nm). Bottom inset: growth of product (kobs = 3.1 × 105 s–1,
monitored at 290 nm).
Fig. 6. LFP traces for 7 (4.7 × 10–5 M) with 1 × 10–4 M IrCl26−
in ~100% H2O (pH 7, N2 purged, λex 308 nm). Diamonds repre-
sent the spectrum sampled 1.8 µs after the laser pulse, which in-
cludes carbanion 14 at 369 nm. Squares and triangles (sampled
at 9.2 and 34 µs) represent the decay of the 369 nm band simul-
taneous with the bleaching of the IrCl26− absorption bands at 430
and 490 nm to the final spectrum (at 130 µs) represented by cir-
cles. Top inset: decay of carbanion 14 (kobs = 6.2 × 104 s–1,
monitored at 360 nm). Bottom inset: bleaching of IrCl26− (kobs
=
Fig. 8. LFP traces for 8 in ~100% H2O, pH 13.0 (5.7 × 10–5 M,
N2 purged, λex 266 nm) which includes α-hydroxy carbanion 25
at 360 nm. Squares and triangles represent decay of the spectrum
at 900 µs and 2.8 ms, respectively. After the decay of 25 was
complete, the spectrum of radical anion 20 was observed; repre-
sented by circles, sampled at 14 ms after the laser pulse. Inset:
Decay of 25 (kobs = 6.7 × 10–2 s–1, monitored at 360 nm) is si-
multaneous with the growth of radical anion 20 (kobs = 4.2 ×
10–2 s–1, monitored at 410 nm).
6.2 × 104 s–1, monitored at 490 nm).
Laser flash photolysis
p-Nitrobenzyl carbanion 2 (which absorbs strongly at λmax
358 nm) has been extensively studied by transient methods
(12, 17). Its lifetime is sensitive to pH but not to oxygen and
is shortened in the presence of IrCl26− that acts as a one-
electron acceptor (23–25). When the π system is further ex-
tended with another phenyl ring, as is the case for p-(p′-
nitrophenyl)benzyl carbanion (23), the absorption maximum
is at much longer wavelength (λmax 470 nm) (6). Further
substitution of the carbanion with an α-hydroxy group does
not shift the absorption maximum, as α-hydroxy-p-(p′-nitro-
© 2003 NRC Canada

592
Can. J. Chem. Vol. 81, 2003
Fig. 9. LFP traces observed for 9 in ~100% H2O, pH 13.0
(5.3 × 10–5 M, N2 purged, λex 266 nm). Diamonds represent the
spectrum sampled 1.7 µs after the laser pulse, which includes
carbanion 27 at 460 nm. Squares and triangles represent decay
of the spectrum, sampled at 8.6 and 28 µs, respectively. Circles
represent spectrum after the decay of 27 at 77 µs. Inset: decay of
27 (kobs = 5.2 × 104 s–1, observed at 460 nm).
Fig. 11. LFP traces observed for 6 in ~100% H2O, pH 13.0
(4.6 × 10–5 M, N2 purged, λex 266 nm). Diamonds represent the
spectrum sampled 5.1 µs after the laser pulse, which includes
carbanion 28 (420 nm) formed by photodeprotonation of 6, and
carbanion 25 (360 nm) formed by photodecarboxylation of 6.
Squares and triangles represent partial decay of the spectrum
(decay of 28 and not 25), sampled at 26 and 69 µs, respectively.
Circles represent the spectrum after 28 has completely decayed,
leaving 25 and possible products from 28. Top inset: decay of 28
(kobs = 2.1 × 104 s–1, monitored at 420 nm). Bottom inset:
growth of product from 28 (kobs = 2.2 × 104 s–1, monitored at
350 nm).
Fig. 10. Plot of the initial change in absorbance (∆ODinit) at
460 nm upon LFP of 9 at pH 10–14 (5.3 × 10–5 M, ~100%
H2O, N2 purged, λex 266 nm). Each point represents the average
initial ∆OD observed for 10 shots, assigned to carbanion 27
formed by photodeprotonation of 9.
Fig. 12. Plot of the initial change in absorbance (∆ODinit) at
420 nm upon LFP of 6 at pH 3–13 (4.6 × 10–5 M, ~100% H2O,
N2 purged, λex 266 nm). Each point represents the average initial
∆OD observed for 10 shots. Above pH 7 ∆ODinit is mostly attrib-
uted to carbanion 28 formed by direct carbon acid deprotonation.
At pH 7 and below ∆ODinit is mostly attributed to 25 formed by
photodecarboxylation which has weak absorption at 420 nm (as-
signment based on lifetimes).
phenyl)benzyl carbanion (24) was also observed at 470 nm
(6). Thus, it was anticipated that the p-nitrobenzyl carban-
ions generated from 6–8 should also have absorption max-
ima near 360 nm. This indeed was found to be the case for
carbanion 14 formed from 7 (λmax 369 nm, vide supra),
which was sufficiently long-lived to allow detection by con-
ventional UV–vis spectrophotometry. LFP of 7 (266 nm,
pH 7, ~100% H2O, N2 purged) also gave 14 (λmax 369 nm)
which did not decay within the 150 ms (longest possible)
time window of our LFP system, consistent with it being a
very long-lived species (vide supra). However, LFP of 7 in
the presence of IrCl26− (308 nm, N2 purged) showed an accel-
erated decay of the 369 nm transient that was simultaneous
with bleaching of IrCl26− bands at 430 and 490 nm (Fig. 6).
LFP of 6 (266 nm) at pH 6 gave a transient at ~360 nm
with a similar absorption profile as that of carbanion 14
(Fig. 7). It decayed via first-order kinetics (τ = 3.8 µs,
0.01 M phosphate buffer, pH 6) to a weaker but permanent
absorption in the range 300–340 nm, assignable as a mixture
© 2003 NRC Canada

Morrison et al.
593
of 15 and 16. Assignment of the 360 nm transient to 25 is
based on its similarity to 14. Additionally, LFP of 6 in the
presence of IrCl26− showed an accelerated decay of the
360 nm band in synchrony with bleaching of the IrCl62−
bands at 430 and 490 nm.
shown to lead to reduction product 16 (21), while scaveng-
ing of the electron of 20 by residual oxygen or on work-up
would lead to 15. The formation of 17 at pH 3 suggests an
alternate pathway involving an overall intramolecular redox
reaction for 25. This new pathway is observed at a pH that
coincides with significant protonation of the nitro group of
25, to give 29 (similar aci-nitrobenzyl compounds have pKa
~ 3, (6, 17, 26)). Net dehydration of 29 gives 17.
LFP of 8 at pH 13 also gave carbanion 25 (τ = 1.5 ms) as
a transient (Fig. 8). No transients were observed at pH 7
consistent with its lack of photochemical reactivity in neu-
tral solution (vide supra). Decay of 25 (at pH 13) was con-
comitant with the growth of a band at 400 nm, which did not
decay within 150 ms (Fig. 8, inset). The long-lived 400 nm
transient (not observed in the presence of oxygen) is as-
signed to radical anion 20, which was also observed by UV–
vis spectrophotometry (Fig. 4). Closer examination of the ki-
netic trace at 410 nm (Fig. 8, inset) also showed evidence of
a shorter-lived species (τ = 63 µs) absorbing in this region,
the identity of which is not known with certainty but which
we have tentatively assigned to carbanion 26, arising via
benzylic deprotonation of 8.
LFP of 9 at pH 13 gave a broad absorption band at
460 nm (Fig. 9) that decayed by first-order kinetics to base-
line (Fig. 9, inset). The intensity of the 460 nm absorption
was strongest at pH 14 and was progressively weaker at
lower pH values; at pH 10 no signal was observable
(Fig. 10). This apparent base catalysis of transient formation
supports its assignment as carbanion 27. Its absorption maxi-
mum is red-shifted by 100 nm compared to carbanion 25
(λex 360 nm) which is compatible with the extended conju-
gation available from the additional phenyl ring.
At pH values > 8, photoionization of the benzylic proton
of 6 (to give 28) apparently competes with photodecarb-
oxylation, as shown in LFP studies, although to what extent
is not entirely clear with the data available. We have not iso-
lated any product at high pH that might be expected to arise
via 28, but only compounds 15 and 16 as seen at pH values
in the range ~4–7. The possibility of facile reprotonation of
28 to give back 6 (in the ground-state reaction) is a distinct
possibility. However, we have not attempted any deuterium
incorporation experiments at these high pHs because of the
obvious complication of thermal background exchange in
basic D2O. Thus, further understanding of the competing
benzylic C-H deprotonation of 6 requires additional studies.
The proposed reaction pathway for 8 (pH > 12)
(Scheme 5) is simpler. LFP observations at pH 13 indicate
that generation of 25 from 8 occurs much more cleanly, with
no significant photoionization at the benzylic position. Car-
banion 25 (from 8) gives radical anion 20 (via overall loss of
an electron and a proton), as shown by growth of the radical
anion absorption band (λmax 403 nm) concomitantly with de-
cay of the absorption band of 25 (λmax 360 nm) (Fig. 8). The
subsequent decay of 20 appears to be more complex, after a
phase with initial first-order kinetics (Fig. 4, inset), and is
suggestive of multistep pathways for subsequent formation
of 16.
The LFP spectrum of 6 at pH 13 was dominated by a
band at 420 nm (Fig. 11), which decayed (τ = 47 µs) to a
long-lived (τ > 150 ms) 360 nm band. The ∆OD of the
420 nm band as a function of pH is shown in Fig. 12, which
has an apparent titration plot, suggesting that the 420 nm
transient is carbanion 28 (with an excited stated pKa of ~8
for the benzylic proton of 6). This seems reasonable consid-
ering the analogous photodeprotonation observed for 9 at
high pH.
The α-acetoxy carbanion 14 generated via photodecarb-
oxylation of 7 behaves as a conventional p-nitrobenzyl car-
banion (mimicking 2, Scheme 1). The two observed
products, 18 and 19, arise via protonation and coupling path-
ways (giving a dianion, which subsequently loses two elec-
trons to give 19) (18). Given that the observed lifetime of
14 is very long compared to that of 25, both its protonation
and coupling processes must be slow. In contrast, the life-
time of 25 at pH 6 is more than 7 orders of magnitude
shorter (µs vs. min). Clearly the disproportionation pathway
available to 25 is sufficiently fast that protonation or cou-
pling cannot compete. Comparison of the pseudo-first-order
rate constants for reaction of these carbanions with IrCl62−
(25 reacts only about 2 times faster than 14) suggests that
the oxidation potentials of 14 and 25 are roughly equivalent.
Thus, the key structural feature that allows the efficient dis-
proportionation pathway observed for 25 is the α-hydroxyl
proton. This is consistent with gas-phase theoretical cal-
Mechanisms of reaction
Our study of the photochemistry of p-nitrobenzyl deriva-
tives 6–10, by a combination of product studies, UV–vis
spectrophotometry, and LFP, has provided additional insights
into their mechanisms of reaction. It seems clear that the
aqueous photochemistry for all of 6–9 involves p-nitrobenzyl
carbanions. Indeed, several α-hydroxy-p-nitrobenzyl carban-
ions, generated by C—H bond deprotonation or by decarb-
oxylation (or deformylation), were observed for the first
time. With the present data, an overall mechanistic picture of
the photochemistry of these compounds is now available.
Photodecarboxylation of 6 (assumed to be via T1) occurs
with appreciable quantum efficiency at pH 3–13 to generate
25 (Scheme 4). The lack of photoreactivity of 6 and absence
of an LFP signal attributable to 25 at pH values 1 and 2 indi-
cate that photodecarboxylation only occurs via the
carboxylate form of 6. The necessary steps to generate prod-
ucts 15 and 16 from 25 most likely involve overall loss of an
electron and a proton (sequentially or via one-step loss of a
hydrogen atom), to generate radical anion 20 (λmax 403 nm,
vide supra), which is known to be long-lived at high pH but
undetectable at lower pH (21). Radical ion 20 has been
–
culations (3) that have shown anions such as CH2OH and
–CH2NH2 to be unstable with respect to their radicals.
The reaction pathway of 9 is similar in essence to those of
6 and 8. It differs in that the carbanion 27 is formed by di-
rect C—H bond heterolysis, and is the only p-nitrobenzyl
carbanion that could form in this case. To our knowledge,
only one other example of LFP observation of a carbanion
generated by photoionization has been reported (6). The sub-
sequent product-forming steps from 27 are believed to be
similar to those proposed for 6 and 8; i.e., the carbanion for-
© 2003 NRC Canada

594
Can. J. Chem. Vol. 81, 2003
Scheme 4.
Scheme 5.
mally loses a hydrogen atom to form the corresponding radi-
cal anion, which subsequently disproportionates to products
21 and 22. The complete absence of photochemistry from 10
indicates the sensitive nature of excited-state deprotonation
with respect to substituents. Further work will be needed to
understand the factors that influence the operation of these
substituent effects, which may be quite subtle.
Materials
p-Nitrobenzaldehyde (15) and p-nitrobenzophenone (21)
(Aldrich) were recrystallized (ethanol–H2O) prior to use. So-
dium hexachloroiridate(IV) hexahydrate 99.9+% (Strem)
was used without further purification. Product 18 was pre-
pared by substitution of p-nitrobenzyl bromide with sodium
acetate. An authentic sample of product 22 was prepared by
reduction of 21 with paraformaldehyde–KOH, as described
elsewhere (27).
Experimental
O-Acetyl p-nitromandelic acid (7)
General
A stirred suspension of 4-nitrophenylacetic acid (15.3 g,
84.5 mmol) in thionyl chloride (6.66 mL, 91.3 mmol) was
refluxed for 1 h. The solution was allowed to cool and bro-
mine (14.9 g, 93 mmol) was added in one portion. The mix-
ture was heated at 70°C for 5.5 h and stirred overnight at
room temperature. Ice water was added and the mixture was
extracted with Et2O. The organic phase was washed with
brine, dried and evaporated to a brown oil (22 g) that was
dissolved in a mixture of THF (230 mL) and water (100 mL)
and stirred for 6 h at room temperature. The solution was
concentrated under reduced pressure to remove most of the
THF and water (50 mL) was added. The mixture was ex-
tracted with Et2O and the ethereal extract was washed with
brine, dried and evaporated to give a solid (18.5 g) that con-
tained the starting nitrophenylacetic acid and the product 11
(typical ratio ~ 1:4). After two crystallizations from
1H NMR spectra were obtained on Bruker AC300 or
JEOL FX90Q instruments in CDCl3 unless otherwise speci-
fied. Low-resolution mass spectra were obtained on a
Finnigan 3300 (CI) and HR-MS were obtained on a Kratos
Concept H (EI) instrument. UV–vis spectra were recorded
on a Varian Cary 5 instrument. Acetonitrile (HPLC grade)
and distilled water were used for product studies and LFP
experiments. Reagent grade CH2Cl2 was redistilled before
use for work-up protocols. The following buffers (0.01 M)
were used for some product studies and LFP static and flow
cell experiments: potassium phosphate – phosphoric acid
(pH 3), acetate – acetic acid (pH 4 and 5), potassium phos-
phate (pH 6–8), sodium carbonate – sodium bicarbonate
(pH 9–11). Aqueous H2SO4 and NaOH were used for solu-
tions at other pH values.
© 2003 NRC Canada

Morrison et al.
595
benzene–hexanes, the contamination by starting material
was reduced to 8% and this material (11.7 g) was used with-
out further processing. Within the mixture, 11 had H NMR
δ: 8.22 (d, J = 8.8 Hz, 2H, Ar-H), 7.73 (d, J = 8.8 Hz, 2H,
Ar-H), 5.40 (s, 1H, CHBr).
2H, arom.), 8.17 (d, J = 8.8 Hz, 2H). HR-MS (EI) calcd. for
C13H11NO3: 229.0739; found: 229.0739.
1
p-Nitrobenzhydryl methyl ether (10)
A solution of 9 (300 mg) in methanol (50 mL) (in the
presence of 10% concd H2SO4) was refluxed for 24 h. Addi-
tion of ether and washing with H2O resulted in extraction of
A mixture of the impure bromo compound 11 (11.49 g,
41.65 mmol) and mercury(II) acetate (28.2 g, 88.4 mmol) in
glacial acetic acid (88 mL) was refluxed for 0.5 h. The solu-
tion was cooled, diluted with water, and extracted with Et2O.
The ether phase was washed with water and brine, dried and
evaporated under reduced pressure. The residue was crystal-
lized from water to give the acetate 7 (7.5 g, 37%), mp 153
1
a light yellow oil which by H NMR was found to contain
5% 9 and 95% 10. Preparative TLC (CH2Cl2 and silica gel)
allowed isolation of a colourless oil (Rf 0.9) p-nitrobenz-
hydryl methyl ether (10). IR (neat, cm–1): 1090 (s), 1190
(m), 1350 (s), 1450 (m), 1520 (s), 1600 (m). 1H NMR
(300 MHz) δ: 3.38 (s, 3H, methyl), 5.29 (s, 1H, methine),
7.30 (m, 5H, arom.), 7.51 (d, J = 8.8 Hz, 2H, arom), 8.16 (d,
J = 8.8 Hz, 2H, arom.). HR-MS (EI) calcd. for C14H13NO3:
243.0896; found: 243.0893.
1
to 154°C (lit. mp (28) 157–159°C). H NMR δ: 8.25 (d, J =
8.8 Hz, 2H, Ar-H), 7.69 (d, J = 8.8 Hz, 2H, Ar-H), 6.04 (s,
1H, CHOAc), 2.23 (s, 3H, Me).
p-Nitromandelic acid (6)
α,α′-Bisacetoxy-p,p′-dinitrobibenzyl (19)
tert-Butyl 2,2,2-trichloroacetimidate (4.0 g, 18.3 mmol) in
CH2Cl2 (32 mL) was added in one portion to a solution of 7
(2.73 g, 11.4 mmol) in CH2Cl2 (69 mL) and the solution was
stirred overnight at room temperature. The mixture was di-
luted with CH2Cl2, washed with saturated NaHCO3, water,
and brine, dried and evaporated. The residue was dissolved
in Et2O (~10 mL) and most of the trichloroacetamide was
precipitated by the addition of hexanes (~40 mL). The fil-
trate was evaporated and purified by flash chromatography
(EtOAc–hexanes, 3:17) to give tert-butyl acetoxy-(4-nitro-
phenyl)acetate (12) as a yellow solid, (2.56 g, 76%). A sam-
ple crystallized from hexanes had mp 80 to 81°C (lit. mp
Esterification of the corresponding diol (0.5 g), which was
available from a previous study (21), in refluxing (10 h) gla-
cial acetic acid (in the presence of 10% concd H2SO4) gave
89% yield of an approximate 1:1 mixture of cis- and trans-α,
α′-bisacetoxy-p,p′-dinitrobibenzyl (19). Recrystallization from
ethyl acetate – hexanes enriched one of the diastereomers,
1
thus allowing distinction between the H NMR signals of the
two diastereomers. IR (mixture of diastereomers, KBr
disc, cm–1): 1220 (s), 1350 (s), 1460 (w), 1530 (s), 1610 (m),
1
1750 (s). H NMR (300 MHz) diastereomer A (40%) δ: 2.07
(s, 6H, COCH3), 6.14 (s, 2H, methine), 7.34 (d, J = 8.8 Hz,
4H, arom.), 8.16 (d, J = 8.8 Hz, 4H, arom.). 1H NMR
(300 MHz) diastereomer B (60%) δ: 2.10 (s, 6H, COCH3),
6.11 (s, 2H, methine), 7.33 (d, J = 8.8 Hz, 4H, arom.), 8.13
(d, J = 8.8 Hz, 4H, arom.). MS (mixture of diastereomers, +
LSI-MS) 389 m/e (M+ + 1).
1
(28) 82°C). H NMR δ: 8.24 (d, J = 8.8 Hz, 2H, Ar-H), 7.66
(d, J = 8.8 Hz, 2H, Ar-H), 5.91 (s, 1H, CHOAc), 2.23 (s,
3H, Me), 1.41 (s, 9H, CMe3).
A solution of 12 (2.53 g, 8.6 mmol) in methanol (25 mL)
was treated with cesium carbonate (0.14 g, 0.43 mmol) in
methanol (5 mL) and stirred for 0.5 h at room temperature.
The solution was diluted with EtOAc, washed with satd.
NaHCO3 and brine, dried and evaporated under reduced
pressure. The residue was purified by flash chromatography
(EtOAc–hexanes, 3:17) to give tert-butyl hydroxy-(4-nitro-
phenyl)acetate (13) (1.51 g, 70%), mp 91.5–93°C (hexanes).
1H NMR δ: 8.21 (d, J = 8.8 Hz, 2H, Ar-H), 7.63 (d, J =
8.8 Hz, 2H, Ar-H), 5.16 (d, JH,OH = 4.5 Hz, 1H, CHOH),
3.72 (d, JH,OH = 4.5 Hz, 1H, OH), 1.42 (s, 9H, CMe3). Anal.
calcd. for C12H15NO5: C 56.91, H 5.97, N 5.53; found: C
56.75, H 5.97, N 5.38.
Product studies
UV–vis photolyses were carried out for 3 mL of the ap-
propriate solution in Suprasil quartz cuvettes. Immediately
before photolysis, each sample was Ar or O2 purged for
5 min, then quickly sealed with a Teflon stopper. During the
photolysis the cuvette sat on top of an inverted 1 L beaker
inside the Rayonet reactor (300 nm, 4–16 lamps) and was
cooled with an internal fan. Irradiations did not exceed
10 min (total photolysis time); no heating of the samples
was observed.
Preparative photolyses were carried out with the substrate
(20–100 mg) dissolved in the appropriate solvent (100 mL)
in a quartz tube. The solution was irradiated at 300 nm
(Rayonet reactor, 4–16 lamps) with continuous cooling (by a
cold finger) and purging by a stream of Ar for approxi-
mately 15 min before and continuously during irradiation
(via a long stainless steel needle). Photolysis times ranged
from 1 min to 20 h, depending on the conversion desired, the
efficiency of the reaction, and the amount of substrate used.
After photolysis, aqueous samples were acidified to pH 1 to
2 and extracted with CH2Cl2. Typical experiments are de-
scribed below.
The tert-butyl ester 13 (1.8 g, 7.1 mmol) was stirred with
trifluoroacetic acid (36 mL) for 1 h at room temperature,
then evaporated under reduced pressure to a brown gum
which solidified on trituration with Et2O. The solid was
crystallized twice from toluene to give 6 as yellow needles
(0.7 g, 48%), mp 128.5–130°C (lit. mp (29) 126 to 127°C).
1H NMR (CD3OD) δ: 8.25 (d, J = 8.8 Hz, 2H, Ar-H), 7.69
(d, J = 8.8 Hz, 2H, Ar-H), 5.26 (s, 1H, CHOH).
p-Nitrobenzhydrol (9)
Reduction of p-nitrobenzophenone (21) (Aldrich) with
NaBH4 gave a white solid, which upon recrystallization
from ethanol–H2O gave white crystals of p-nitrobenzhydrol
(9), mp 69 to 70°C. IR (KBr disc, cm–1): 1190 (m), 1340 (s),
1450 (m), 1510 (s), 1600 (m), 3300–3500 (s). 1H NMR
(300 MHz) δ: 2.37 (s, 1H, exchanges with D2O, OH), 5.90
(s, 1H, methine), 7.33 (m, 5H, arom.), 7.56 (d, J = 8.8 Hz,
Photolysis of p-nitromandelic acid (6) in 1:1 H2O–
CH3CN (pH 7)
A solution of 6 (20 mg) in 1:1 H2O–CH3CN (100 mL, pH
adjusted to 7 with NaOH, Ar purged) was photolyzed for
© 2003 NRC Canada

596
Can. J. Chem. Vol. 81, 2003
1
2 min (300 nm, 16 lamps). After work-up, H NMR of the
photolysate allowed observation of two aldehyde products
(60% 15, 40% 16) and 6 (36% total conversion). Preparative
TLC (CH2Cl2 and silica gel) allowed isolation of the two
products, one of which was identified as 15 (Rf 0.5) by com-
parison with an authentic sample; the other (Rf 0.3) identi-
fied as p,p′-azoxybisbenzaldehyde (16), mp 192°C. IR (KBr
disc, cm–1): 1200 (m), 1320 (m), 1460 (m), 1600 (m), 1700
that did not dissipate for up to several hours. Upon
acidification to pH 2 with aq HCl, the color turned to yellow
(between pH 10 and 11), with formation of a yellow precipi-
1
tate. Extraction of the whole mixture with CH2Cl2 gave a H
NMR that showed ~15% conversion of 9 to 21 and 22 (70
and 30%, respectively), and a third, unidentified product. In
a separate experiment, the precipitate formed during acidifi-
cation was isolated by suction filtration and found to contain
22 and the unidentified product (in a ~1:1 mixture), while
the filtrate contained 9 and 21. Preparative TLC (CH2Cl2
and silica gel) failed to separate 22 from the unidentified
product. Based on analysis of the 1H NMR pattern of the un-
identified product (as a mixture with 22), we have tenta-
tively identified it as a secondary photoproduct of 22, i.e., a
hydroxyazobenzene derived from 22, consistent with the
known photo-Wallach rearrangement of aromatic azoxy
compounds (31).
1
(s). H NMR δ: 8.01 (d, J = 8.1 Hz, 2H, arom.), 8.05 (d, J =
8.1 Hz, 2H, arom.), 8.26 (d, J = 8.1 Hz, 2H, arom.), 8.49 (d,
J = 8.1 Hz, 2H, arom.), 10.05 (s, 1 H, CHO), 10.14 (s, 1 H,
CHO). HR-MS (EI) calcd. for C14H10N2O3: 254.0692;
found: 254.0689.
Photolysis of p-nitromandelic acid (6) in 1:1 H2O–
CH3CN (pH 3)
A solution of 6 (20 mg) in 1:1 H2O–CH3CN (100 mL,
pH 3, 0.01 M phosphate buffer, Ar purged) was photolyzed
1
for 5 min (300 nm, 16 lamps). Following work-up, H NMR
Quantum yields
allowed observation of the aldehyde peaks of 15 and 16, as
well as an aldehyde signal which had not been present in the
pH 7 product studies. Based on integration the relative yields
of these products were 50% (15), 22% (16) and 28% of the
new aldehyde product, and the total conversion from sub-
strate was 61%. This procedure was repeated until approxi-
mately 90 mg of photolysate had been collected. An initial
attempt to separate the new product via preparative TLC
(CH2Cl2 and silica gel) resulted in recovery of 15 and the
new product in one band (Rf 0.6). A second attempt (1:9
ethyl acetate:hexanes and silica gel) succeeded in separating
15 (Rf 0.2) from the new light yellow product (Rf 0.3), iden-
The quantum yield of 6 was measured by preparative
1
photolysis and H NMR using 3 as secondary actinometer
(Φ = 0.6) (12). Photolysis of 3 (20 mg, pH 7, 100 mL 100%
0.01 M phosphate buffer, 300 nm, 4 lamps, 2 min, Ar
purged) followed by assessment of the conversion to product
(10–15%) allowed calculation of the number of photons
emitted by the apparatus. In a following photolysis of 6
(100% pH 6 0.01 M phosphate buffer) using identical condi-
tions, assessment of conversion (5–10%) allowed calculation
of the quantum yield for the reaction. These photolyses were
repeated four times, and the average quantum yield was cal-
culated.
1
tified as p-nitrosobenzaldehyde (17) (based on identical H
NMR with that reported in the literature (30)). IR (KBr
disc, cm–1): 1200 (m), 1260 (s), 1390 (m), 1460 (w), 1595
(m), 1690 (s). H NMR δ: 8.03 (d, J = 8.8 Hz, 2H, arom.),
Laser flash photolysis (LFP)
1
LFP data was acquired using either a Nd:YAG laser
(Spectra-Physics Quanta-Ray GCR-11, λex 266 nm; this la-
ser was used for most LFP experiments) or a Lumonics
excimer laser (EX-510, λex 308 nm; this laser was used for
experiments using IrCl62−) with pulse widths ~20 ns and
power attenuated to ~20 mJ/pulse. Signals were digitized
with a Tektronix TDS 520 recorder. Solutions with OD = 0.3
at λex were used to ensure even penetration of the laser light.
Quartz flow cells were used for all studies except for the
measurement of ∆ODinit at various pH values (Figs. 10 and
12), where quartz static cells were used. Samples in flow
cells were continuously purged with a stream of N2 or O2 for
15 min before and during the experiment. Samples in static
cells were purged with N2 or O2 for 5 min and subsequently
sealed prior to the experiment. The contents of static cells
were mixed between each laser shot to avoid depletion of the
solution in the irradiated volume.
8.17 (d, J = 8.1 Hz, 2H, arom.), 10.20 (s, 1H, CHO). HR-
MS (EI) calcd. for C7H5NO2: 135.0321; found: 135.0323.
Photolysis of O-acetyl p-nitromandelic acid (7) in 1:1
H2O–CH3CN (pH 7)
A solution of 7 (20 mg) in 1:1 H2O–CH3CN (100 mL, pH
adjusted to 7, Ar purged) was photolyzed for 20 min
(300 nm, 4 lamps). Following work-up, 1H NMR of the
photolysate revealed 23% conversion to 18 and 19 (47 and
53%, respectively), each of which were identified by addi-
tion of an authentic sample (vide supra) which enhanced the
1H NMR signals.
Photolysis of p-nitrophenylethylene glycol (8) in 1:1
H2O–CH3CN (pH 13)
A solution of 8 (100 mg) in 1:1 H2O–CH3CN (100 mL,
pH 13, Ar purged) was photolyzed for 30 min (300 nm, 16
1
lamps). Following work-up H NMR of the photolysate re-
vealed 69% conversion to 15 and 16 (60 and 40%, respec-
tively), identified as above (vide supra).
Acknowledgments
Photolysis of p-nitrobenzhydrol (9) in 1:1 H2O–CH3CN
(pH 13)
A solution of 9 (20 mg) in 1:1 H2O–CH3CN (100 mL,
pH 13, Ar purged) was photolysed for 60 min (300 nm, 16
lamps). After photolysis the solution turned a deep red color
We thank the Natural Sciences and Engineering Research
Council of Canada (NSERC) for continued support of this
research, and Mr. Xigen Xu for the initial studies on com-
pound 8.
© 2003 NRC Canada

Morrison et al.
597
15. G.A. Russell and E.G. Janzen. J. Am. Chem. Soc. 89, 300
(1967).
16. E. Buncel and B.C. Menon. J. Am. Chem. Soc. 102, 3499
(1980).
17. B.B. Craig, R.G. Weiss, and S.J. Atherton. J. Phys. Chem. 91,
5906 (1987).
References
1. D.J. Cram. Fundamentals of carbanion chemistry. Academic
Press Inc., New York. 1965. p. 289.
2. P.M. Mayer and L. Radom. J. Phys. Chem. A, 102, 4918
(1998).
3. K.M. Downard, J.C. Sheldon, J.H. Bowie, D.E. Lewis, and
R.N. Hayes. J. Am. Chem. Soc. 111, 8112 (1989).
4. P. Wan and K. Yates. Can. J. Chem. 64, 2076 (1986).
5. P. Wan and X. Xu. Tetrahedron Lett. 31, 2809 (1990).
6. J. Morrison, H. Osthoff, and P. Wan. Photochem. Photobiol.
Sci. 1, 384 (2002).
7. M. St. Maurice and S.L. Bearne. Biochemistry, 39, 13324
(2000).
8. M. Garcia-Viloca, A. Gonzalez-Lafont, and J.M. Lluch. J. Am.
Chem. Soc. 123, 709 (2001).
18. B.B. Craig and S.J. Atherton. J. Chem. Soc., Perkin Trans. 2,
1929 (1988).
19. A. McKenzie and P.A. Stewart. J. Chem. Soc. 104 (1935).
20. G. Carrara, F. Fava, C. Martinuzzi, and A. Vecchi. Gazz.
Chim. Ital. 82, 674, (1952).
21. S. Muralidharan and P. Wan. J. Photochem. Photobiol. A. 57,
191 (1991).
22. T. Shida. Electronic absorption spectra of radical ions.
Elsevier, New York. 1988.
23. S. Steenken and P. Neta. J. Am. Chem. Soc. 104, 1244 (1982).
24. J. Halpern, R.J. Legare, and R. Lumry. J. Am. Chem. Soc. 85,
680 (1963).
25. J.Y. Chen, H.C. Gardner, and J.K. Kochi. J. Am. Chem. Soc.
98, 6150 (1976).
9. P.K. Das. Chem. Rev. 93, 119 (1993).
10. T.-I. Ho and Y.L. Chow. In The chemistry of amino, nitroso
and related groups. Vol. 1. Edited by S. Patai. John Wiley and
Sons, Toronto. 1996. p. 747.
11. D. Döpp. In CRC handbook of organic photochemistry and
photobiology. Edited by W.M. Horspool and P.-S. Song. CRC
Press, Boca Raton. 1995. Chap. 81. p. 1019.
12. J.D. Margerum and C.T. Petrusis. J. Am. Chem. Soc. 91, 2467
(1969).
13. P. Wan and S. Muralidharan. J. Am. Chem. Soc. 110, 4336
(1988).
14. G.A. Russell, A.J. Moye, E.G. Janzen, S. Mak, and E.R.
Talaty. J. Org. Chem. 32, 137 (1967).
26. M. Schwörer and J. Wirz. Helv. Chim. Acta, 84, 1441 (2001).
27. R.B. Davis and J.D. Benigni. J. Org. Chem. 27, 1605 (1962).
28. K. Schank. Chem. Ber. 103, 3093 (1970).
29. G. Heller. Chem. Ber. 46, 280 (1913).
30. X. Creary, P.S. Engel, N. Kavaluskas, L. Pan, and A. Wolf. J.
Org. Chem. 64, 5634 (1999).
31. N.J. Bunce, J.P. Schoch, and M.C. Zerner. J. Am. Chem. Soc.
99, 7986 (1977).
© 2003 NRC Canada