Efficient photoionization of a norrish II diradical

Article abstract of DOI:10.1002/anie.200504056

(Chemical Equation Presented) Extremely little is known about the factors governing the efficiency of photoionizations. The comparison of propiophenone and γ-methylvalerophenone highlights that a spontaneous intramolecular transformation accompanied by small structural and electronic changes - the conversion of a triplet moiety into a radical (anion) moiety by a Norrish II reaction - can have a pronounced effect on the electron yield.

Full text of DOI:10.1002/anie.200504056

Angewandte
Chemie
Photochemistry
DOI: 10.1002/anie.200504056
Efficient Photoionization of a NorrishII Diradical
Martin Goez* and Valentin Zubarev
The photochemistry of reactive intermediates in solution[1] is
a fascinating subject because of its promise of new mecha-
nisms and applications.[2] Photoionizations of excited singlet
and triplet states are common; examples of photoionizations
of radicals[3] and radical anions[4] are known. Here, we present
the first report on the photoionization of a diradical and show
that it is considerably more efficient than that of its precursor
triplet.
Figure 1. Electron concentrations relative to the substrate concentra-
tion c0 as functions of the laser intensity I at 308 nm for propiophe-
none (open symbols and dashed line; c0 =1.66”10ꢀ4 m) and g-
methylvalerophenone (filled symbols and solid line; c0 =7.15”10ꢀ5 m)
at pH 13.54. The formulas of the substrates are shown at the curves.
The fit functions are given by Equation (1). Constant parameter kdt: 0
(propiophenone) and 0.41 (g-methylvalerophenone); best-fit parame-
ters kexc and kion: 1.02”10ꢀ3 cm2 mJꢀ1 and 9.19”10ꢀ4 cm2 mJꢀ1 (propio-
phenone); 1.21”10ꢀ3 cm2 mJꢀ1 and 1.43”10ꢀ2 cm2 mJꢀ1 (g-
We performed laser flash photolysis at very high light
intensities, where mechanisms involving more than one
photon reveal themselves much more clearly than under the
usual conditions. Our setup[4c] allows two-flash two-color
experiments (355 or 266 nm, pulse width 6 ns; 308 nm, pulse
width 30 ns; freely adjustable interpulse delay) and homoge-
neously illuminates a well-defined reaction volume, thus
permitting quantitative determinations of absolute concen-
methylvalerophenone). The ratio of the kexc was fixed at the ratio of the
extinction coefficients of the two ketones at 308 nm (1.19).
and propiophenone as the longest-chain alkyl phenyl ketone
that is not yet capable of undergoing this reaction. It is evident
that the NorrishII substrate produces substantially more
electrons.
The upward curvature of the intensity dependence bears
out that the ionization of the NorrishII substrate is still
biphotonic. With the mechanism of Scheme 1, the electron
yield is determined by the (intensity-dependent) rates of both
steps. As expected, the extinction coefficients of the two
ground-state ketones, and thus the rates of production of *M
for given laser intensity, are very similar. Hence, this step
cannot explain the observed effect, and the different yields
have to be ascribed to different properties of the photo-
ionizable intermediate *M. It is natural to assume that they
reflect the differences between the triplet (for propiophe-
none) and the NorrishII diradical (for g-methylvalerophe-
none) because the latter is formed within less than 5 ns,[8] in
other words, much faster than the duration of the laser pulse.
This is corroborated by the results of pH variation. As
Figure 2 shows, the electron yield is independent of pH in the
case of propiophenone, whereas a titration curve with a
turning point at pH 11.5 is found in the case of g-methyl-
trations of transients. As the primary observable, the concen-
Cꢀ
tration of the hydrated electron e directly at the end of a
aq
laser pulse was measured through its optical absorbance.[5]
Alkyl phenyl ketones M can be ionized with 308 nm
light.[4d,6] The usual mechanism is a sequential two-photon
process as displayed in Scheme 1. Absorption of the first
Scheme 1. Sequential two-photon ionization of a ketone, M.
photon leads to an intermediate *M, normally the triplet
state, which is then ionized by the second photon. Even at our
high light intensities, the rate of the first step is limited by the
excitation of M because of the small extinction coefficient at
308 nm; the subsequent dark reactions to give *M, such as
intersystem crossing, are much faster.
When the ketone bears a hydrogen atom in the g position,
its triplet is rapidly transformed into a 1,4-diradical by an
intramolecular hydrogen abstraction (the NorrishII reac-
tion).[7] We were interested in how this classic photoreaction
influences the photoionization. Figure 1 compares the elec-
tron yields, as functions of the laser intensity in single-flash
experiments at 308 nm, for two model compounds in degassed
aqueous base (nominal concentration of KOH, 0.4m), g-
methylvalerophenone as an efficient NorrishII substrate,[8]
Figure 2. Dependence of the relative electron concentrations [eCaꢀq]/c0 on
the pH (adjusted with KOH and phosphate buffers of variable
composition) at 308 nm and constant excitation intensity
[*] Prof. Dr. M. Goez, V. Zubarev
Fachbereich Chemie
Martin-Luther-Universität Halle-Wittenberg
Kurt-Mothes-Strasse 2, 06120 Halle/Saale (Germany)
Fax : (+49)345-552-7657
(392 mJcmꢀ2). Open symbols and dashed line: propiophenone,
c0 ꢁ2.7”10ꢀ4 m; filled symbols and solid line: g-methylvalerophenone,
c0 ꢁ7.9”10ꢀ5 m. The exact pH and substrate concentration c0 were
determined before each experiment in the solution used.
E-mail: goez@chemie.uni-halle.de
Angew. Chem. Int. Ed. 2006, 45, 2135 –2138
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2135

Communications
valerophenone. Although in neutral solution the NorrishII
substrate exhibits a reduced yield, it still produces more than
twice the amount of electrons than the simpler ketone does.
The absorption spectrum of M is not influenced by the pH
value. Nor can the extinction coefficient and photoionization
quantum yield of the triplet change with pH, as the absence of
an effect in the propiophenone experiment shows. The
difference between the two substrates that persists in neutral
solution finally excludes a transition from triplet ionization at
low pH to diradical ionization at high pH as the origin of the
pH dependence, which would also be difficult to reconcile
with the short lifetime of the g-methylvalerophenone triplet.
Instead, a deprotonation of the diradical in basic medium
provides a natural explanation.
Figure 3. Two-pulse experiments with the pulse sequence represented
in the inset; pulse intensities: 35 mJcmꢀ2 (266 nm) and 400 mJcmꢀ2
(308 nm). Shown are the electron yields scaled to maximum, [eCaꢀq]rel, as
functions of the interpulse delay Dt for propiophenone (open symbols
and dashed line; c0 =1.66”10ꢀ4 m) and g-methylvalerophenone (filled
symbols and solid line; c0 =8.68”10ꢀ5 m), both at pH 13.5.
Except for their intramolecular deactivation steps, Nor-
rishII diradicals are describable as separate alkyl and ketyl
radical moieties.[8] The alkyl radical moiety cannot be
deprotonated, so it could only cause a pH-independent
contribution to photoionization. Even that is thought very
unlikely in our case because of insufficient absorption (the
extinction coefficient of the tert-butyl radical at 308 nm is as
low as 200mꢀ1 cmꢀ1).[9] In contrast, both the ketyl radical and
its deprotonated form, the radical anion, have medium to
strong absorptions at 308 nm, and their photoionization is
much more favorable energetically than alkyl radical photo-
ionization because the by-product is a closed-shell moiety,
that is, the regenerated carbonyl group. Of importance for the
electron yield is the red shift of the absorption band caused by
the deprotonation,[10] which leads to a considerable increase
of the extinction coefficient at 308 nm.
Definitive evidence that the NorrishII diradical is respon-
sible for the difference in the electron yields was obtained by
two-pulse experiments. They allow the separation of ground-
state excitation and the actual photoionization, which pro-
vides a twofold advantage. First, the weak absorption of the
ground-state ketone M at 308 nm can be overcome by the
choice of a different wavelength, 266 nm, for the first pulse;
when the second pulse is left at 308 nm, the small extinction
coefficient of M is turned into an asset because that pulse now
predominantly excites the intermediate. Second, the lifetime
of the photoionizable intermediate can be determined by
varying the interpulse delay.
With both ketones studied, the extinction coefficient of M
is about 16 times higher at 266 nm than at 308 nm. However,
this is not paralleled by a similar increase of the extinction
coefficients of the photoionizable intermediates: at 266 nm,
the electron yield is even slightly smaller than at 308 nm. With
a suitably attenuated 266-nm pulse, about 30% of the
substrate can be converted into *M with negligible electron
generation. By the second pulse, at 308 nm, the respective
intermediate is photoionized after a delay Dt.
In Figure 3 the electron yields relative to the maximum in
each system have been plotted as functions of Dt, allowing a
direct comparison of the lifetimes of the electron precursors.
With propiophenone one observes a slow decrease on a
timescale of a few microseconds, as expected for the photo-
ionization of the triplet state. However, with g-methylvalero-
phenone the decrease is much more rapid; a fit by a
monoexponential decay curve gave a time constant of 74 ns.
It is well known[8] that NorrishII diradicals cleave on that
timescale to give an alkene and an enol (or, in the case of a
deprotonated diradical, an enolate). These diamagnetic
products are not expected to be photoionized as efficiently
as the diradical, if at all, so the two-pulse experiment monitors
the diradical decay. For final confirmation, we omitted the
second pulse and observed the decay of the diradical directly,
using its characteristic absorption at 440 nm. The diradical
lifetime obtained in this way (76 ns) and the decay time
constant in the two-pulse experiment were identical within
experimental error. A very similar lifetime (84 ns) was found
in neutral solution.
We determined the pKa of the g-methylvalerophenone
ketyl radical to be 9.7 by generating hydrated electrons from
8-amino-2-naphthalenesulfonic acid with 355-nm light at
different pH values, scavenging them with our ketone,
which does not absorb at that wavelength, and observing
the ketyl radical or its anion at 308 nm. The obtained value
agrees very well with that for the acetophenone ketyl radical
(pKa = 10.0).[10] The significantly higher turning point of the
titration curve in Figure 2, at pH 11.5, presents a discrepancy
that might seem to contradict the paradigm of independent
ketyl and alkyl radical moieties in the NorrishII diradical.
An apparent pKa in methanol/water of 11.8 was also
reported for the diradical of g-phenylbutyrophenone, a very
close analogue to our diradical, and ascribed to a solvent
effect.[11] In our experiments, however, an apparent pKa that is
shifted to a higher value is most likely due to a rate effect:
Regardless of pH, direct formation of the diradical with the
deprotonated ketyl moiety is impossible, because it is the
product of an intramolecular hydrogen abstraction by the
carbonyl group. At pH values higher than the true pKa, the
ketyl moiety will subsequently be deprotonated, resulting in a
higher photoionization probability, but our experiment only
responds to those deprotonated diradicals that are formed
within the duration of the laser pulse. When the pH is
increased, the concentration of proton acceptors rises, the
deprotonation becomes faster, and the fraction of deproton-
ated diradicals available within the kinetic window increases.
Experiments with different buffer concentrations might
support this hypothesis, but the feasible range of variation is
rather limited at that high pH.
2136
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 2135 –2138

Angewandte
Chemie
Scheme 2 summarizes the mechanism of the photoioniza-
tion of g-methylvalerophenone as inferred from these experi-
ments. After excitation of the substrate by the first photon
and intersystem crossing on a subnanosecond timescale,[6] the
coefficients of the g-methylvalerophenone ketyl radical and
radical anion at 308 nm of 1900 and 15100mꢀ1 cmꢀ1, the ratio
of which closely parallels the ratio of kion in neutral and basic
solution. Hence, we conclude that deprotonation of the ketyl
moiety does not influence the photoionization quantum yield,
and that the different electron yield (Figure 2) is solely due to
the change of the extinction coefficient. The extinction
coefficient of the g-methylvalerophenone triplet should
practically be identical to that of the propiophenone triplet,
4400mꢀ1 cmꢀ1 at 308 nm.[4d] The same comparison thus reveals
that the quantum yield of photoionization of the diradical is
almost five times higher than that of the triplet, which is in
good agreement with the previously obtained[4d] ratio for the
propiophenone triplet and radical anion.
Our results for other systems[4b–e] suggest that the con-
version of a triplet state into the corresponding radical or
radical anion, as can be achieved by the addition of a suitable
quencher, is a general method to increase the photoionization
quantum yield, presumably because this transformation
suppresses reverse intersystem crossing of the upper excited
state. The diradical studied here represents an intramolecular
application of that principle.
Scheme 2. Photoionization of g-methylvalerophenone via its diradical.
first species potentially capable of absorbing the second
photon is the triplet. However, the probability of its photo-
ionization is low because it is converted into a NorrishII
diradical within a few nanoseconds,[8] which is still fast
compared to the duration of the ionizing laser pulse. In
basic solution, the ketyl moiety of the diradical is deproton-
ated. Both the unmodified and the deprotonated ketyl moiety
can be photoionized, the latter yielding more electrons, and
this ionization competes with chemical deactivation of the
diradical.
An approximate kinetic treatment is possible for the
limiting cases of very high or very low pH, where the diradical
is present exclusively in either of its two forms (see Figure 2).
Photoionization of the short-lived precursor triplet and of the
ketone formed by diradical cleavage are neglected, the latter
because its concentration during the pulse is low and its
structure precludes another NorrishII reaction, so its ioniza-
tion must be less efficient than that of the parent compound.
The mechanism is thus that of Scheme 1 enhanced by an
additional light-independent deactivation pathway of *M
(with rate constant kd). With the assumption of a rectangular
laser pulse of duration t, diradical formation occurs with a
rate constant kexcI/t and photoionization with a rate constant
Received: November 15, 2005
Published online: February 24, 2006
Keywords: diradicals · ketones · laser chemistry ·
.
photoionization · photolysis
[1] a) J. C. Scaiano in Kinetics and Spectroscopy of Carbenes and
Biradicals (Ed.: M. S. Platz), Plenum, New York, 1990, pp. 353 –
368; b) L. J. Johnston, Chem. Rev. 1993, 93, 251 – 266; c) W. G.
McGimpsey in Molecular and Supramolecular Photochemistry,
Vol. 2 (Eds.: V. Ramamurthy, K. S. Schanze), Marcel Dekker,
New York, 1998, pp. 249 – 306; d) A. Natarajan, L. S. Kaanu-
malle, V. Ramamurthy in CRC Handbook of Organic Photo-
chemistry and Photobiology, 2nd ed. (Eds.: W. Horspool, F.
Lenci), CRC, Boca Raton, FL, 2004, pp. 107/1 – 107/24.
[2] a) I. A. Shkrob, M. C. Sauer, Jr., A. D. Liu, R. A. Crowell, A. D.
Trifunac, J. Phys. Chem. A 1998, 102, 4976 – 4989; b) C. R.
Lambert, I. E. Kochevar, R. W. Redmond, J. Phys. Chem. B
1999, 103, 3737 – 3741; c) V. V. Jarikov, A. V. Nikolaitchik, D. C.
Neckers, J. Phys. Chem. A 2000, 104, 5131 – 5140; d) M. Goez, V.
Zubarev, Angew. Chem. 2001, 113, 2948 – 2951; Angew. Chem.
Int. Ed. 2001, 40, 2867 – 2869; e) M. Goez, M. Schiewek,
M. H. O. Musa, Angew. Chem. 2002, 114, 1606 – 1609; Angew.
Chem. Int. Ed. 2002, 41, 1535 – 1538; f) M. Goez, B. H. M.
Hussein, Angew. Chem. 2003, 115, 1697 – 1700; Angew. Chem.
Int. Ed. 2003, 42, 1659 – 1661; g) M. Goez, D. von Ramin-Marro,
M. H. O. Musa, M. Schiewek, J. Phys. Chem. A 2004, 108, 1090 –
1100; h) X. Cai, M. Sakamoto, M. Hara, S. Tojo, A. Ouchi, A.
Sugimoto, K. Kawai, M. Endo, M. Fujitsuka, T. Majima, J. Phys.
Chem. A 2005, 109, 3797 – 3802.
k
ionI/t, where each parameter k is proportional to the product
of the extinction coefficient of the species absorbing the
photon and the quantum yield of the resulting process.[4e] The
intensity dependence of the electron yield at the pulse end is
given by Equation (1). When kd is set to zero, Equation (1)
can also be applied to the photoionization of propiophenone.
ꢀ
k
ionI
k
excI
excIꢀðkionI þ kdtÞ
ionI þ kdt
excIꢀðkionI þ kdtÞ
½eCaꢀqŠ=c0 ¼
1ꢀ
exp½ꢀðkionI þ kdtފ
k
ionI þ kdt
k
ð1Þ
ꢁ
k
þ
exp½ꢀkexcIŠ
k
[3] a) R. W. Redmont, J. C. Scaiano, L. J. Johnston, J. Am. Chem.
Soc. 1990, 112, 398 – 402; b) J. L. Faria, S. Steenken, J. Am.
Chem. Soc. 1990, 112, 1277 – 1279; c) J. L. Faria, S. Steenken, J.
Phys. Chem. 1993, 97, 1924 – 1930; d) J. L. Faria, S. Steenken, J.
Chem. Soc. Perkin Trans. 2 1997, 1153 – 1159.
[4] a) P. Natarajan, R. W. Fessenden, J. Phys. Chem. 1989, 93, 6095 –
6100; b) V. Zubarev, M. Goez, Angew. Chem. 1997, 109, 2779 –
2781; Angew. Chem. Int. Ed. Engl. 1997, 36, 2664 – 2666; c) M.
Goez, V. Zubarev, G. Eckert, J. Am. Chem. Soc. 1998, 120, 5347 –
Fitting Equation (1) to the data (see the legend of
Figure 1) shows that for the deprotonated diradical the
constant kion is larger by a factor 15.6 compared to that of
the triplet; in neutral solution, in other words, for the initially
formed diradical, this factor is found to be only 2.1. The
scavenging experiment described above gave extinction
Angew. Chem. Int. Ed. 2006, 45, 2135 –2138
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
2137

Communications
5348; d) M. Goez, V. Zubarev, Chem. Phys. 2004, 307, 15 – 26;
e) M. Goez, B. H. M. Hussein, Phys. Chem. Chem. Phys. 2004, 6,
5490 – 5497.
Cꢀ
[5] For better sensitivity, e was not observed at its absorption
aq
maximum (720 nm) but at 829 nm; see, Ref. [4d], footnote 9. By
control experiments (saturation of the solutions with the
electron scavenger N2O) it was ascertained that no other species
Cꢀ
absorbs at 829 nm. The extinction coefficient of e at 829 nm,
aq
13750mꢀ1 cmꢀ1, was determined by comparison with the max-
imum extinction coefficient at 720 nm.[4e]
[6] R. Boch, M. K. Whittlesey, J. C. Scaiano, J. Phys. Chem. 1994, 98,
7854 – 7857.
[7] P. J. Wagner, Acc. Chem. Res. 1971, 4, 168 – 177.
[8] J. C. Scaiano, Acc. Chem. Res. 1982, 15, 252 – 258, and references
therein.
[9] C. Huggenberger, H. Fischer, Helv. Chim. Acta 1981, 64, 338 –
353.
[10] E. Hayon, T. Ibata, N. N. Lichtin, M. Simic, J. Phys. Chem. 1972,
76, 2072 – 2078.
[11] R. A. Caldwell, S. N. Dhawan, D. E. Moore, J. Am. Chem. Soc.
1985, 107, 5163 – 5166.
2138
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 2135 –2138