Modulation of lifetimes and diastereomeric discrimination in triplet- excited substituted butane-1,4-diones through intramolecular charge-transfer quenching

Article abstract of DOI:10.1021/ja9818708

Triplet lifetimes have been determined for the diastereomers of a broad set of butane-1,4-dione derivatives (1-3). A remarkable dependence of lifetimes on conformational preferences is revealed in that the lifetimes are shorter for the meso diastereomers of 1-3 than those for the racemic ones. The intramolecular β-phenyl quenching is promoted in the case of meso diastereomers by virtue of the gauche relationship between the excited carbonyl group and the β-aryl ring, while a distal arrangement in the lowest energy conformation (H-anti) in racemic diastereomers prevents such a deactivation. The involvement of charge transfer in the intramolecular β- phenyl quenching is suggested by the correlation of the triplet lifetimes of the meso diastereomers of compounds 2 with the nature of the substituent on the β-phenyl rings. In the case of racemic diastereomers, p-methoxy substitution on the β-phenyl ring (2-OCH₃, 3-OCH₃) also led to a decrease of the triplet lifetimes when compared to those of the nonsubstituted compounds (2-H, 3-H). This shortening is accounted for by the deactivation of a small proportion of the excited molecules through β-phenyl quenching. In addition to the above factors, the lifetimes in the case of meso diastereomers can further be controlled by increasing the energy spacing between the T₁ and T₂ states, since β-phenyl quenching occurs from the latter for compounds 2 and 3. Through a rational conformational control, a surprisingly long triplet lifetime (300 ns) has been measured for the first time for a purely n,π* triplet-excited β-phenylpropiophenone dimer (1- rac).

Full text of DOI:10.1021/ja9818708

J. Am. Chem. Soc. 1999, 121, 3093-3103
3093
Modulation of Lifetimes and Diastereomeric Discrimination in
Triplet-Excited Substituted Butane-1,4-diones through Intramolecular
Charge-Transfer Quenching
J. N. Moorthy,†,§ S. L. Monahan,† R. B. Sunoj,‡ J. Chandrasekhar,‡ and C. Bohne*,†
Contribution from the Department of Chemistry, UniVersity of Victoria, P.O. Box 3065,
Victoria, BC, Canada V8W 3V6, and Department of Organic Chemistry, Indian Institute
of Science, Bangalore 560012, India
ReceiVed May 29, 1998
Abstract: Triplet lifetimes have been determined for the diastereomers of a broad set of butane-1,4-dione
derivatives (1-3). A remarkable dependence of lifetimes on conformational preferences is revealed in that the
lifetimes are shorter for the meso diastereomers of 1-3 than those for the racemic ones. The intramolecular
â-phenyl quenching is promoted in the case of meso diastereomers by virtue of the gauche relationship between
the excited carbonyl group and the â-aryl ring, while a distal arrangement in the lowest energy conformation
(H-anti) in racemic diastereomers prevents such a deactivation. The involvement of charge transfer in the
intramolecular â-phenyl quenching is suggested by the correlation of the triplet lifetimes of the meso
diastereomers of compounds 2 with the nature of the substituent on the â-phenyl rings. In the case of racemic
diastereomers, p-methoxy substitution on the â-phenyl ring (2-OCH3, 3-OCH3) also led to a decrease of the
triplet lifetimes when compared to those of the nonsubstituted compounds (2-H, 3-H). This shortening is
accounted for by the deactivation of a small proportion of the excited molecules through â-phenyl quenching.
In addition to the above factors, the lifetimes in the case of meso diastereomers can further be controlled by
increasing the energy spacing between the T1 and T2 states, since â-phenyl quenching occurs from the latter
for compounds 2 and 3. Through a rational conformational control, a surprisingly long triplet lifetime (300 ns)
has been measured for the first time for a purely n,π* triplet-excited â-phenylpropiophenone dimer (1-rac).
Introduction
control of photophysical properties. Compounds for which
photophysical properties can be conformationally controlled
have the potential to probe how supramolecular structures affect
the intramolecular mobility of an incorporated molecule. In this
context, â-phenylpropiophenones are particularly interesting in
that their triplet lifetimes are limited by the conformation, in
which the triplet-excited carbonyl group and the â-phenyl ring
are in close proximity.12,13 Consequently, they have been
employed as excellent probes of supramolecular systems.14-21
The involvement of charge transfer in the intermolecular
quenching of triplet-excited ketones by arenes has been long
recognized.22-26 Detailed studies by Wagner and co-workers
Carbonyl compounds have traditionally played a key role in
the development of organic photochemistry as an independent,
vital branch of chemistry as a whole.1 The interest in ketone
photochemistry continues unabated in the quest for compre-
hensive mechanistic details of even such seemingly simple
reactions as photoreductions.2 One aspect of ketone photochem-
istry that has attracted a great deal of attention is the confor-
mational control of photochemical reactivity.1,3-5 There are
many reports in the literature to exemplify this aspect.1,3-11 In
noted contrast, little has been reported on the conformational
* Corresponding author. Phone: 250-721-7151. Fax: 250-721-7147.
E-mail: bohne@uvic.ca.
(12) Scaiano, J. C.; Perkins, M. J.; Sheppard, J. W.; Platz, M. S.; Barcus,
R. L. J. Photochem. 1983, 21, 137.
(13) Netto-Ferreira, J. C.; Leigh, W. J.; Scaiano, J. C. J. Am. Chem.
Soc. 1985, 107, 2617.
(14) Leigh, W. J. J. Am. Chem. Soc. 1985, 107, 6114.
(15) Workentin, M. S.; Fahie, B. J.; Leigh, W. J. Can. J. Chem. 1991,
69, 1435.
(16) Workentin, M.; Leigh, W. J. J. Phys. Chem. 1992, 96, 9666.
(17) Netto-Ferreira, J. C.; Scaiano, J. C. J. Photochem. Photobiol. A:
Chem. 1988, 45, 109.
(18) Bohne, C.; Barra, M.; Boch, R.; Abuin, E. B.; Scaiano, J. C. J.
Photochem. Photobiol. A: Chem. 1992, 65, 249.
(19) Casal, H. L.; Scaiano, J. C. Can. J. Chem. 1984, 62, 628.
(20) Scaiano, J. C.; Casal, H. L.; Netto-Ferreira, J. C. ACS Symp. Ser.
1985, 278, 211.
(21) de Lucas, N. C.; Netto-Ferreira, J. C. J. Photochem. Photobiol. A:
Chem. 1997, 103, 137.
(22) Guttenplan, J. B.; Cohen, S. G. J. Am. Chem. Soc. 1972, 94, 4040.
(23) Schuster, D. I.; Weil, T. M.; Halpern, A. M. J. Am. Chem. Soc.
1972, 94, 8248.
(24) Wolf, M. W.; Brown, R. E.; Singer, L. A. J. Am. Chem. Soc. 1977,
99, 526.
† University of Victoria.
‡ Indian Institute of Science.
§ Present address: Department of Chemistry, India Institute of Technol-
ogy, Kanpur 208016, India.
(1) Wagner, P. J. Top. Curr. Chem. 1976, 66, 1.
(2) Leigh, W. J.; Lathioor, E. C.; St. Pierre, M. J. J. Am. Chem. Soc.
1996, 118, 12339.
(3) Wagner, P. J. Acc. Chem. Res. 1983, 16, 461.
(4) Wagner, P. J. Acc. Chem. Res. 1989, 22, 83.
(5) Wagner, P. J.; Park, B.-S. In Organic Photochemistry; Padwa, A.,
Ed.; Marcel Dekker: New York, 1991; Vol. 11, p 227.
(6) Lewis, F. D.; Johnson, R. W.; Kory, D. R. J. Am. Chem. Soc. 1974,
96, 6100.
(7) Lewis, F. D.; Johnson, R. W.; Johnson, D. E. J. Am. Chem. Soc.
1974, 96, 6090.
(8) Wagner, P. J.; Chen, C.-P. J. Am. Chem. Soc. 1976, 98, 239.
(9) Wagner, P. J.; Lindstrom, M. J.; Sedon, J. H.; Ward, D. R. J. Am.
Chem. Soc. 1981, 103, 3842.
(10) Wagner, P. J.; Zhou, B.; Hasegawa, T.; Ward, D. L. J. Am. Chem.
Soc. 1991, 113, 9640.
(11) Zand, A.; Park, B.-S.; Wagner, P. J. J. Org. Chem. 1997, 62, 2326.
10.1021/ja9818708 CCC: $18.00 © 1999 American Chemical Society
Published on Web 03/19/1999

3094 J. Am. Chem. Soc., Vol. 121, No. 13, 1999
Moorthy et al.
Chart 1
on the photoreduction of substituted acetophenones and ben-
zophenones with alkylbenzenes showed that both n,π* and π,π*
triplets react with alkylbenzenes predominantly by a rate-
determining charge-transfer complexation.25,26 However, the
reactivity shifts to direct hydrogen transfer when the triplet
ketone becomes harder to reduce.26 â-Phenylpropiophenone
constitutes an intramolecular case of quenching. Its lack of
photoreactivity, which was recognized more than 50 years ago,27
was explained as being due to efficient intramolecular charge-
transfer quenching of the n,π* triplet-excited ketone by the
phenyl ring at the â-position.12,28-30 From time-resolved studies,
a surprisingly short triplet lifetime of ca. 1 ns was revealed by
Wismontski-Knittel and Kilp.31 Subsequently, Scaiano and co-
workers demonstrated that substitution in the â-phenyl ring by
electron-donating groups has only a marginal effect on the triplet
lifetimes of â-phenylpropiophenones.13 This lack of dependence
led them to suggest that bond rotations, which bring the excited
carbonyl group and the â-phenyl ring into close proximity,
determine the overall triplet decay. From Arrhenius parameters
for the triplet decay, a tight transition state for quenching was
suggested.12,13 In contrast, for â-phenyl-p-methoxypropiophe-
nones, the lifetime increases by 2 orders of magnitude (3τ )
230 ns in methanol).13,32,33 This increase is attributed to the need
to thermally populate the high-energy n,π* state from the lowest
triplet state with π,π* character.13 In the case of â-aryl-p-
methoxypropiophenone, the triplet lifetimes depend on the
nature of the substituent on the â-phenyl ring.33 The triplet decay
rate constants correlate with the Hammett σ+ constants, indicat-
ing that charge-transfer interactions between acceptor carbonyl
and donor aryl groups are responsible for intramolecular
quenching. For this reason, it was suggested that the rate-limiting
step for the triplet decay is the â-phenyl quenching process and
not conformational motions to achieve the quenching geom-
etry.33 In addition to the substituent effects, conformational
restrictions that lead to increased proximity between the carbonyl
and the â-phenyl rings have been shown to influence the triplet
lifetimes.32,34
We have recently demonstrated a diastereomeric discrimina-
tion in the triplet lifetimes of the racemic and meso diastereo-
mers of a substituted 1,4-diketone (2-H, see Chart 1).35
Conformational control was suggested to be responsible for the
shorter triplet lifetime of 2-H-meso, where the efficient in-
tramolecular quenching was ascribed to the gauche relationship
between the â-phenyl and carbonyl moieties in the excited-triplet
conformer with lowest energy. In the case of 2-H-rac, its triplet
lifetime is mainly determined by the intrinsic decay of the triplet,
with little contribution from intramolecular quenching. This
initial study was expanded to compounds 1-3 (Chart 1) to
establish how the triplet lifetimes of both diastereomers can be
altered by substitution in the benzoyl and/or â-phenyl rings.
Furthermore, we show that, through conformational control, it
is possible to increase the lifetime to the nanosecond time
domain for a purely n,π* triplet-excited â-phenylpropiophenone
dimer (1-rac).
(25) Wagner, P. J.; Lam, M. H. J. Am. Chem. Soc. 1980, 102, 4166.
(26) Wagner, P. J.; Truman, R. J.; Puchalski, A. E.; Wake, R. J. Am.
Chem. Soc. 1986, 108, 7727.
(27) Bergman, F.; Hirshberg, Y. J. Am. Chem. Soc. 1943, 65, 1429.
(28) Whitten, D. G.; Punch, W. E. Mol. Photochem. 1970, 2, 77.
(29) Wagner, P. J.; Kelso, P. A.; Kemppainen, A. E.; Haug, A.; Graber,
D. R. Mol. Photochem. 1970, 2, 81.
(30) Stermitz, F. R.; Nicodem, D. E.; Muralidhadran, V. P.; O’Donnell,
C. M. Mol. Photochem. 1970, 2, 87.
(31) Wismontski-Knittel, T.; Kilp, T. J. Phys. Chem. 1984, 88, 110.
(32) Boch, R.; Bohne, C.; Netto-Ferreira, J. C.; Scaiano, J. C. Can. J.
Chem. 1991, 69, 2053.
(33) Leigh, W. J.; Banisch, J. H.; Workentin, M. S. J. Chem. Soc., Chem.
Commun. 1993, 988.
Results
Synthesis of Ketones. The precursor benzyl phenyl ketones
were synthesized from Friedel-Crafts reactions between R-aryl-
(34) Boch, R.; Bohne, C.; Scaiano, J. C. J. Org. Chem. 1996, 61, 1423.
(35) Moorthy, J. N.; Patterson, W. S.; Bohne, C. J. Am. Chem. Soc. 1997,
119, 11094.

Triplet Lifetimes of Butane-1,4-dione Diastereomers
J. Am. Chem. Soc., Vol. 121, No. 13, 1999 3095
acetyl chloride and anisole/1,3-dimethoxybenzene in the pres-
ence of AlCl3 using carbon disulfide as solvent (eq 1).36 The
1
racemic and meso diastereomers of diketone 1 were prepared
according to the reported procedure; benzyl phenyl ketone was
subjected to oxidative dimerization by KMnO4 in refluxing
acetone (path a, eq 2).37 Although this procedure afforded both
Figure 1. Triplet-triplet absorption spectra in methanol of 2-H-rac
([) and 3-H-rac (b) measured respectively at delays of 1170 and 470
ns after the laser pulse.
Table 1. Triplet Lifetimes (µs) and T-T Absorption Maxima (nm)
of Diketones 1-3
acetonitrile
methanol
diketone
1-rac
1-meso
3τ
λmax
3τ
λmax
0.25 (0.01)a
340
0.31 (0.04)a
340
<0.01c
<0.01c
2-H-rac
10 (1)b,d
0.23 (0.01)a
>4.5e
400
400
33 (8)b,d
400
410
410
410
400
400
400
2-H-meso
2-CH3-rac
2-CH3-meso
2-CF3-rac
2-CF3-meso
2-OCH3-rac
0.63 (0.04)a
400 >6.6e
0.06 (0.01)a
>5.3e
410
0.17 (0.02)b
380 >6.1e
390 >5.3e
>3.3e
3.5 (0.1)b,d
400
7.5 (0.2)b,d
2-OCH3-meso <0.01c
<0.01c
3-H-rac
3-H-meso
4.1 (0.3)a
430
430
430
10.2 (0.6)b,d
2.2 (0.1)a
2.3 (0.2)a
440
450
430
0.77 (0.05)b
1.2 (0.1)a
3-OCH3-rac
3-OCH3-meso
diastereomers of the diketone 2-CF3 in a rather low yield (9%),
only the meso diastereomers were isolated in the cases of 2-H
and 2-OCH3. Therefore, the synthetic procedure was modified,
and that involving CuCl2-mediated oxidative dimerization of
the respective lithium enolates (path b, eq 2) reported for the
preparation of analogous diketones was adopted.38 Thus, treat-
ment of the lithium enolates at -78 °C with CuCl2 afforded
the racemic and meso diastereomers of diketones 2-H, 2-CH3,
2-OCH3 and 3-H, 3-OCH3 in 10-30% yields. No attempt was
made to optimize the reaction conditions to improve the yields.
In general, the diketones were isolated in higher yields when
the R-aryl ring was substituted with an electron-donating group
at the para position; the reaction completely failed when the
para substituent was trifluoromethyl.
0.030 (0.002)b 440
0.060 (0.003)a 450
a From three or more measurements; the errors in parentheses refer
to standard deviations. b From two independent measurements; the error
refers to an average deviation. c Limiting higher value; no signal was
detected on the nanosecond time scale. d Lifetime as extrapolated to
zero ketone concentration. e Minimum value since efficient bimolecular
self-quenching occurs.
pionyl chloride. The diketone 1,4-bis(4-methoxyphenyl)butane-
1,4-dione (6) was synthesized according to the reported
procedure.40 A similar procedure was employed for the prepara-
tion of 1,4-bis(2,4-dimethoxyphenyl)butane-1,4-dione (7). Thus,
the Friedel-Crafts reaction between 1,3-dimethoxybenzene and
succinyl chloride in the presence of AlCl3 in refluxing carbon
disulfide furnished the diketone 7.
The separation of racemic and meso diastereomers of ketones
2 and 3 was achieved by fractional crystallization in methanol
and in diethyl ether-hexanes mixtures. Only in the case of
2-CF3 was it not possible to isolate the pure diastereomers
without contamination of each other; however, the extent of
contamination (<5%) did not present a serious problem in the
time-resolved work (see below). The stereochemical assignments
for meso and racemic diastereomers of diketones 2 and 3 were
Triplet-Triplet Absorption Spectra and Triplet Lifetimes
of the Ketones. Laser flash photolysis of all the ketones 2 and
3, with the exception of 2-OCH3-meso, led to the detection of
transient absorptions corresponding to their excited triplet states
(Figure 1). The assignment of the transient with absorption
maxima at 380-410 nm for compounds 2 and 430-450 nm
for compounds 3 (Table 1) to the triplet state is based on the
experiments described for 2-H35 and the similarity to the triplet-
triplet (T-T) absorption spectra previously reported.13,32 The
spectra for the meso and racemic diastereomers of each
compound were similar and were independent of the solvent
used. The decay of compounds 2 and 3 showed a monoexpo-
nential behavior throughout the T-T absorption spectrum, and
the same triplet lifetimes were measured at different wave-
lengths. A long-lived residual was observed after the decay of
the triplet state. This residual at the T-T absorption maxima
was smaller for the meso diastereomers (<5%) than for the
racemic ones (<10%). In the cases of 2-OCH3 and 3-OCH3,
the residues were larger than those observed for 2-H and 3-H.
1
made by comparison of the characteristic H NMR chemical
shifts of the benzylic protons with those of the diastereomers
of diketones 1. In the latter case, the stereochemical assignments
were established on the basis of the influence of chiral chemical
shift reagents on the benzylic resonances at δ ∼5.4 (racemic)
and ∼5.7 (meso).39
2,4-Dimethoxypropiophenone (5) was synthesized from
Friedel-Crafts acylation of 1,3-dimethoxybenzene with pro-
(36) Buck, J. S.; Ide, W. S. J. Am. Chem. Soc. 1932, 54, 3012.
(37) Suginome, H. J. Org. Chem. 1958, 23, 1044.
(38) Ito, Y.; Konoike, T.; Harada, T.; Saegusa, T. J. Am. Chem. Soc.
1977, 99, 1487.
(39) Guerri, F. A.; Sagredo, R. A.; Sanz, J. J. Photochem. Photobiol. A:
Chem. 1990, 55, 87.
(40) Buchta, E.; Scheffer, G. Justus Liebigs Ann. Chem. 1955, 597, 129.

3096 J. Am. Chem. Soc., Vol. 121, No. 13, 1999
Moorthy et al.
Figure 2. Transient absorption spectra in methanol of 1-rac at delays
of 85 ns (b) and 1.7 µs (O) after the laser pulse and for 1-meso at a
220 ns delay (9).
Scheme 1
Figure 3. Decay of 1-rac at 340 nm in methanol. The residuals (44A)
correspond to the difference between the experimental and calculated
values for the fit of the decay to a monoexponential function with an
offset of the baseline.
eomers of 1, which probably corresponds to the radicals formed
in the photodecomposition of this compound. Since the lowest
triplet of 1 has an n,π* configuration, we expect the photo-
cleavage to be competitive with other deactivation channels for
the excited triplet state. Photodecomposition studies of 1 at 254
nm in a Rayonet reactor showed that both diastereomers are
decomposed. However, the photodecomposition rate is at least
20 times faster for the racemic diastereomer than that for the
meso compound.42 The transient phenomena regarding the
formation of the triplet of 1 in acetonitrile are similar to those
observed in methanol. However, additional transients are formed
in the 400-500 nm region.43 In analogy to the results for 1 in
methanol, we assign the fast decay at 340 nm to the triplet of
1-rac (Table 1). No short-lived transient was observed for
1-meso in acetonitrile.
Residual absorptions after the decay of the triplet were previ-
ously reported for other â-phenylpropiophones,32 and in the
cases of the diketones 2 and 3 they are probably due to the
formation of radical products from R- and â-cleavage reactions
(Scheme 1).41 The fact that no absorption corresponding to the
triplet was observed for 2-OCH3-meso suggests that its lifetime
is significantly shorter than the 10 ns time resolution of the laser
flash photolysis system.
The transient phenomena for 1 were more complex since
photodecomposition of this compound was significant. At short
delays after the laser pulse, a transient with an absorption
maximum at 340 nm was observed for 1-rac in methanol but
was absent for 1-meso (Figure 2). The initial decay at 340 nm
of the transient observed for 1-rac, assigned to its triplet-excited
state, was monoexponential, followed by a significant long-lived
residual (15-20%) (Figure 3). The lack of the signal at 340
nm in the case of the 1-meso suggests that its triplet lifetime is
shorter than 10 ns. At long delays after the laser pulse, a
featureless spectrum (Figure 2) was observed for both diaster-
The values for the triplet lifetimes of 1-rac (Table 1) were
independent of the ketone ground-state concentration. These
lifetimes include the contribution from the photocleavage
reaction, which is responsible for the residual absorption
observed. Since the cleavage reaction is one of the decay
pathways of the triplet, offsetting the baseline to the value
corresponding to the residual absorption yields the correct triplet
lifetimes.44 The lifetimes for the meso diastereomers of 2-H
and 2-CH3 in acetonitrile and methanol (Table 1) were found
to be independent of the ketone concentrations, while a strong
dependence was found in the case of the racemic diastereomers
of all the diketones 2 and 2-CF3-meso. The shortening of the
lifetimes with increasing diketone concentration is due to a
bimolecular self-quenching process in which a ground-state
diketone quenches the excited triplet state of another molecule.
The bimolecular self-quenching rate constants for 2-H-rac and
2-OCH3-rac (ksq; kobs ) k0 + ksq[ketone]) were determined from
the quenching plots at different diketone concentrations (Figure
(41) The photodecomposition of both diastereomers of 2-H irradiated
in a Rayonet reactor with 254 nm lamps was followed by HPLC. The
photodecomposition of 2-H-rac is faster than that observed for 2-H-meso,
as expected from the longer lifetime of the racemic diastereomer. The
relative photodecomposition rate for 2-H-rac is much smaller than that for
1-rac and is similar to that observed for 1-meso. Furthermore, the residuals
observed in the laser flash photolysis experiment for the triplet decay of
â-phenyl-p-methoxypropiophenone and 2-H are similar, suggesting that
photodecomposition of 2-H is not a major component of its triplet
deactivation, as shown previously for â-phenyl-p-methoxypropiophenone.13
The mechanistic studies on the photocleavage reactions, which are not the
subject of this work, will be described separately.
(42) The photodecompositions of both diastereomers of 1 in methanol
and acetonitrile were followed by monitoring the decrease of the concentra-
tion of 1 by HPLC. The estimated decomposition rates were 180 and 160
nM/s, respectively, for 1-rac in methanol and acetonitrile and 5 and 7 nM/
s, respectively, for 1-meso in methanol and acetonitrile with matched
absorbances at 254 nm.
(43) The UV-vis absorption spectra of the photodecomposition of 1 in
acetonitrile are different from those observed in methanol.
(44) Encinas, M. V.; Wagner, P. J.; Scaiano, J. C. J. Am. Chem. Soc.
1980, 102, 1357.

Triplet Lifetimes of Butane-1,4-dione Diastereomers
J. Am. Chem. Soc., Vol. 121, No. 13, 1999 3097
Table 2. Triplet Lifetimes (µs)a and T-T Absorption Maxima
(nm) of Ketones 4-7
acetonitrile
3τ
methanol
ketone
λmax
3τ
λmax
4
5
6
7
5.6 (0.2)b
7 (0.4)
7 (1)
380
420
380
420
33 (4)
36 (6)
28 (1)
18 (2)
390
430
390
440
7 (2)
a Determined as extrapolated to zero ketone concentrations. The
values in parentheses represent average deviations from two experi-
ments. b No self-quenching was observed.
Figure 4. Concentration dependence for the triplet decay of 2-H-rac
in (a) methanol and (b) acetonitrile.
4).45 Thus, for diketones 2-H-rac and 2-OCH3-rac, the triplet
lifetimes were determined in methanol and acetonitrile as
extrapolated to zero ketone concentrations. The concentration
dependence was not examined for 2-CH3-rac and for both the
diastereomers of diketones 2-CF3. Therefore, the values in Table
1 constitute the limiting lower values for their triplet lifetimes.
The lifetimes for compounds 3 in methanol were found to be
concentration-dependent for 3-H-rac,46 but no such dependence
was revealed in acetonitrile. The diketones 3-H-meso, 3-OCH3-
rac, and 3-OCH3-meso displayed concentration-independent
triplet lifetimes in both acetonitrile and methanol.
Figure 5. Arrhenius plots for triplet decay of 3-OCH3-meso in
acetonitrile (9), 3-OCH3-meso in methanol (0), 2-H-meso in aceto-
nitrile (b), and 2-H-meso in methanol (O).
described above (Table 2).48 The lifetimes of 2-H-rac (Table
1) are similar to those of 4 and 6, whereas for 3-H-rac one
observes that its lifetime is shorter than those for 5 and 7.
A strong solvent effect on the lifetimes of all the ketones is
observed (Tables 1 and 2). A significant lengthening of triplet
lifetimes is observed in a polar protic medium such as methanol;
for example, the lifetime for 2-H-meso varies from ca. 600 ns
in methanol to 100 ns in apolar nonprotic solvent such as carbon
tetrachloride,35 while it is 220 ns in acetonitrile. A similar solvent
dependence for the triplet lifetimes is observed for the model
compounds 4-7. Such an effect of medium on the lifetimes
has long been recognized, and the effect of polar protic solvents
has been interpreted in terms of hydrogen-bonding effects13 and
stabilization of π,π* triplet-excited states.49 Regardless of the
mechanism involved, it is noteworthy that the lifetimes of both
racemic and meso diastereomers of diketones 2-3 follow a
similar trend in the different solvents.
Temperature Dependence of the Lifetimes. The dependence
of lifetimes on the temperature was examined for 1-rac and the
meso diastereomers of 2-H, 2-CH3, 3-H, and 3-OCH3 over a
range of 223-323 K. The triplet decays were measured by
monitoring the transient absorptions at 340, 400, and 440 nm
for diketones 1, 2, and 3, respectively. In all the cases, the decay
traces followed first-order kinetics, and the values of kobs were
obtained from single-exponential fits. A linear behavior was
observed for log kobs vs 1/T (Figure 5). The activation energies
and preexponential factors associated with triplet decays of the
ketones examined are given in Table 3.
The results in Table 1 show that the triplet lifetimes for all
the meso diastereomers of diketones 1-3 are remarkably shorter
than those for the corresponding racemic ones.47 A striking
dependence of the lifetimes of meso diastereomers on the nature
of the para substituent of the phenyl rings is observed. The
lifetimes progressively increase in the order methoxy < methyl
< hydrogen < trifluoromethyl (Table 1). A similar trend is
observed in the case of diketones 3. While a 4-fold difference
is noted in the lifetimes for 3-H-rac and 3-OCH3-rac, a 20-
40-fold difference is evident in the lifetimes of the meso
diastereomers.
Molecules 4 and 5 were studied to establish the triplet
lifetimes of simple analogues of the racemic diastereomers of
2 and 3. Ketones 6 and 7 were studied to establish the
importance of intramolecular self-quenching, i.e., the quenching
of one of the carbonyl chromophores by the other in the same
molecule in compounds 2 and 3. The T-T absorption spectra
for p-methoxypropiophenone (4) and 1,4-bis(p-methoxyphenyl)-
butane-1,4-dione (6) were found to be similar to those of
diketones 2, while the spectra for 2,4-dimethoxypropiophenone
(5) and 1,4-bis(2,4-dimethoxyphenyl)butane-1,4-dione (7) re-
sembled those for 3 (Table 2). The triplet lifetimes of all the
ketones 4-7, with exception of 4 in acetonitrile, exhibited strong
dependences on the ketone concentrations. Therefore, the triplet
lifetimes were determined at zero ketone concentrations as
MM PI Calculations. The differences in the lifetimes for
the meso and racemic diastereomers of each ketone are related
to the relative population of the triplet-excited conformers and
the rotational energy barriers for their interconversion. More
detailed calculations than those previously reported35 were
(48) The ksq’s for the ketones 4-7 in (a) methanol are (4 ( 1) × 107, (6
( 1) × 108, (2 ( 1) × 108, and (10 ( 1) × 108 M-1 s-1, and in (b)
acetonitrile are nonobservable, (5 ( 1) × 108, (7 ( 1) × 107, and (14 ( 5)
108 M-1 s-1, respectively.
(49) Wagner, P. J.; Kemppainen, A. E.; Schott, H. N. J. Am. Chem. Soc.
1973, 95, 5604.
(45) The self-quenching rate constants (ksq’s) for 2-H-rac and 2-OCH3-
rac in (a) methanol are (6 ( 2) × 108 and (4 ( 1) × 108 M-1 s-1 and in
(b) acetonitrile are (5 ( 1) × 108 and (3 ( 1) × 108 M-1 s-1, respectively.
(46) ksq ) (8 ( 1) × 108 M-1 s-1
.
(47) See also ref 35 for a figure showing the triplet decays for 2-H.

3098 J. Am. Chem. Soc., Vol. 121, No. 13, 1999
Moorthy et al.
Table 3. Activation Energies and Preexponential Factors for
Triplet Decay of the Diketones 1-3
acetonitrile
methanol
diketone
1-rac
2-H-meso
2-CH3-meso
3-H-meso
3-OCH3-meso
Ea (kcal/mol)
log A
Ea (kcal/mol)
log A
6.4 (0.1)a
6.4 (0.1)b
5.8 (0.1)a
6.0 (0.1)b
4.2 (0.2)a
11.5 (0.1)a
11.4 (0.1)b
11.6 (0.1)a
10.6 (0.1)b
10.7 (0.2)a
5.9 (0.3)a
6.6 (0.1)b
6.6 (0.2)a
7.0 (0.1)b
5.6 (0.1)a
10.8 (0.3)a
11.1 (0.1)b
11.7 (0.2)a
10.9 (0.1)b
11.4 (0.1)a
a From a single experiment; the error in parentheses represents
statistical deviations of the data from a linear fit. b From two indepen-
dent measurements; the error refers to an average deviation.
Table 4. Calculated Steric Energies and Relative Energies with
Respect to the Lowest Energy Conformation for Both Diastereomers
of 2-H and 3-H and Energy Barriers from the Lowest Energy
Conformation for the Racemic Conformers of 2-H and 3-H
min energy rel energy energy barrier
compound conformation (kcal/mol) (kcal/mol)
(kcal/mol)
Figure 6. Structures for the calculated lowest energy H-anti conforma-
tions of 3-H-meso (a) and 3-H-rac (b). The shortest contacts between
the carbonyl oxygen atoms and a carbon on the â-phenyl rings are
shown.
2-H-meso
H-anti
H-gauche
H-anti
Ph-anti
Bz-anti
H-anti
H-gauche
H-anti
53.3
56.4
51.1
54.6
57.5
56.8
61.9
55.3
58.5
60.6
0
3.1
0
3.5
6.4
0
5.1
0
2-H-rac
8.4
8.0
3-H-meso
3-H-rac
ordering of proximate n,π* and π,π* levels and cause the lowest
excited state to be π,π* in character.51,52 Thus, the lowest excited
triplet is purely n,π* in the case of 1, while it is π,π* for
diketones 2 and 3.
Ph-anti
Bz-anti
3.2
5.3
8.5
7.5
The significant results from Table 1 are the remarkable
diastereomeric discrimination and the variation in the triplet
lifetimes, which range from <10 ns to 30 µs. While the former
evidently pertains to conformational factors, the latter is
traceable to substituent effects. Also, one observes shorter triplet
lifetimes for 2-OCH3-rac versus 2-H-rac and for 3-OCH3-rac
versus 3-H-rac caused by remote p-methoxy substitution in the
â-phenyl ring. In the ensuing discussion, these observations are
rationalized by addressing first the diastereomeric discrimination
followed by substituent effects.
Diastereomeric Discrimination. A perusal of the results in
Table 1 reveals that the triplet lifetimes for all the meso
diastereomers, except for 2-CF3-meso, are significantly shorter
than those for the corresponding racemic ones. We have
established for diketones 2-H that the difference in the lifetimes
for the two isomers is, indeed, real and not due to the presence
of an adventitious impurity, which efficiently quenches the
triplet-excited state in one of the diastereomers.35 To reconcile
the different lifetimes for the racemic and meso diastereomers,
the effect of intramolecular self-quenching and intramolecular
triplet energy transfer on the lifetimes for the two diastereomers
should be considered. The importance of these two factors has
been examined in detail by Wagner and co-workers in analogous
diketones.53
performed for both diastereomers of 2-H and 3-H to establish
the difference in energies of their conformers. These calculations
include the effect of π delocalization by explicitly identifying
the phenyl groups, carbonyl moieties, and methoxy oxygen
atoms. It was confirmed for both diastereomers that the
conformers with lowest energy correspond to the H-anti
conformers (Table 4). The relative energies for the other
conformations are high enough to ensure that, under equilibrium
conditions, most molecules will be populated in the H-anti ones.
For the meso diastereomers, the H-anti conformation has the
â-phenyl rings and the carbonyl moieties in close proximity,
whereas in the H-anti conformation of the racemic compound,
the shortest contact between these groups is larger than 4.3 Å,
and the carbonyl oxygen is not directed toward the π cloud of
the phenyl ring (Figure 6). All other conformations, i.e.,
H-gauche for the meso and Ph-anti and Bz-anti for the racemic
diastereomers, have at least one contact between the carbonyl
and â-phenyl moieties of 3 Å.50 It is important to note that the
interconversion barriers for the racemic diastereomers were
calculated by varying the torsional angle about the central C-C
bonds and reoptimizing the structure at each new angle. This
procedure does not necessarily follow the lowest energy
pathway. For this reason, the computed energy barriers (Table
4) represent upper limits.
The intramolecular self-quenching which corresponds to the
deactivation of one of the triplet-excited carbonyls by the other
ground-state carbonyl in the same molecule can, in principle,
occur. The rate constant for such a process in analogous n,π*-
excited 1,2-dibenzoylethane has been shown to be 1.7 × 107
Discussion
Apart from establishing the conformational dependence of
triplet lifetimes, our rationale to study diketones 1-3 was
twofold: first, to establish the extent of charge transfer for the
intramolecular â-phenyl quenching for diketones 2; second, to
investigate the influence of the configuration of the lowest
triplet-excited state on the lifetimes of diketones 1, 2-H-meso,
and 3-H-meso. It has long been known that substitution with
electron-donating substituents in alkanophenones can invert the
s-1 53
. To investigate the importance of the intramolecular self-
quenching, the triplet lifetimes of diketones 6 and 7 were
measured. These lifetimes are comparable with those for the
monoketones 4 and 5, suggesting that intramolecular self-
(51) Turro, N. J. Modern Molecular Photochemistry; Benjamin/Cum-
mings Publishing Co.: Menlo Park, CA, 1978; p 628.
(52) Gilbert, A.; Baggott, J. Essentials of Molecular Photochemistry;
Blackwell Scientific Publications: Oxford, 1991; p 538.
(50) The calculated structures for the lowest energy conformers of 2-H
and 3-H are given in the Supporting Information.
(53) Wagner, P. J.; Freking, H. W., Jr. Can. J. Chem. 1995, 73, 2047.

Triplet Lifetimes of Butane-1,4-dione Diastereomers
J. Am. Chem. Soc., Vol. 121, No. 13, 1999 3099
quenching is not an important deactivation process.54 Moreover,
the lowest energy conformations of the racemic diastereomers
of 2 (H-anti, Table 4) have the two benzoyl moieties in a gauche
relationship, whereas for the H-anti conformation of the meso
compound, the benzoyl groups are in an anti relationship.
Consequently, the intramolecular self-quenching process should
be faster for the racemic diastereomers. Therefore, the lifetimes
for the racemic compounds would be expected to be shorter
when compared to those for the meso isomers if intramolecular
quenching was occurring. This is clearly not the case, since the
lifetimes of the racemic diastereomers are longer than those
measured for the respective meso compounds.
flash photolysis studies, where a significant amount of residual
absorption is apparent after the decay of the triplet state. In this
respect, it is important to establish if photodecomposition is the
factor responsible for the different triplet lifetimes observed for
1-meso and 1-rac, since the shorter lifetime for the former could
be due to an enhancement of its photodecomposition rate
constants. However, the photodecomposition yield for 1-meso
was much smaller than that for 1-rac. This result shows that
the shorter triplet lifetime for 1-meso is not due to an enhanced
photodecomposition rate constant. On the contrary, the larger
photodecomposition observed for 1-rac is due to the longer
triplet lifetime of this diastereomer, which makes it possible
for photocleavage to compete with triplet deactivation.
The lifetimes for â-phenylpropiophenones are shorter than 1
ns at 300 K, and they are not very dependent on the nature of
the substituents on the â-phenyl ring.13 These results were
explained by suggesting that molecular motions to achieve the
quenching conformation are rate limiting. In the case of
compound 1, the triplet lifetime is given by the sum of the
intramolecular â-phenyl quenching process (kâ) and the pho-
tocleavage rate constant (kcleavage):56
The intramolecular energy transfer between the two carbonyl
moieties in diketones has been shown to occur with a rate
constant of the order of 109 M-1 s-1 53,55
In the case of
.
compounds 1-3, this intramolecular energy-transfer process is
isoenergetic, which would cause the excitation energy to reside
on either of the carbonyl chromophores for half of the time.
This intramolecular energy transfer cannot affect the triplet
decays differently to account for the observed differences in
the triplet lifetimes of the diastereomers.
As mentioned at the outset, the deactivation of triplet states
by â-aryl rings occurs through an efficient intramolecular
quenching mechanism.13,28-33 The quenching process is believed
to involve a conformation in which the triplet-excited ketone
and the â-phenyl ring are sufficiently close (gauche geometry)
to permit efficient overlap between the electron-deficient
n-orbital on oxygen and the π-orbitals of the electron-rich
â-phenyl ring. In this respect, the intramolecular â-phenyl
quenching is expected to occur from the ketone excited states
with n,π* configuration. The quenching of the π,π* triplet-
excited ketone, which readily occurs for arenes as an inter-
molecular process, is not involved in the case of â-phenylpro-
piophenones. This is probably because the geometry with an
effective overlap between the â-phenyl ring and the π-system
does not occur in these ketones.26,31,33 Therefore, we will
consider throughout our discussion that â-phenyl quenching
occurs only from excited states with n,π* configuration. The
differences in the triplet lifetimes for the meso and racemic
diastereomers of 1-3 are a consequence of how efficiently
â-aryl quenching contributes to the triplet deactivation for each
diastereomer. Since the diastereomers differ only in terms of
spatial arrangement of atoms, their lowest energy conformations
are of immediate interest. We assume, based on the calculations
performed for 2-H and 3-H, that for all compounds the
conformations with lowest energy for the meso as well as the
racemic diastereomers are the H-anti conformers.
kd ) kâ + kcleavage
(3)
In the case of 1-meso, the lowest energy conformation (H-
anti) has the â-phenyl rings gauche to the benzoyl moieties,
and â-phenyl quenching after excitation occurs so readily as to
preclude observation of the triplet in the nanosecond time
domain. The small photodecomposition quantum yield when
compared to that observed for 1-rac also indicates that the
â-phenyl quenching rate constant is much larger than the
photocleavage rate constant. However, in the case of 1-rac, the
phenyl rings lie away from the benzoyl groups in the lowest
energy H-anti conformation. For this reason, the â-phenyl
quenching does not readily occur since it requires a rotation
from the H-anti to either the Bz-anti or Ph-anti conformers.
This rotation is slow, and photocleavage readily competes with
intramolecular â-phenyl quenching. Based on the cleavage rate
constant estimated for â-phenylpropiophenone of 106 s-1 13
it
,
is evident that kcleavage contributes considerably (g30%) to the
triplet decay rate constant of 1-rac. Therefore, the observed
triplet decay rate constants constitute an upper limit for the
â-phenyl quenching rate constant in 1-rac. Since â-phenyl
quenching for triplet 1-rac requires bond rotation, the rate
constants for these bond rotations are remarkably smaller (e2.8
× 106 and e 2.3 × 106 s-1 in acetonitrile and methanol,
respectively, assuming that φâ is smaller than 70%) than the
motion necessary for the intramolecular quenching in â-phe-
nylpropiophenone (1 × 109 s-1). This large difference suggests
that rotation is much more hindered in the case of 1 than that
for â-phenylpropiophenone. The activation energy for the
â-phenyl quenching process in â-phenylpropiophenone has not
been determined.13 However, this value can be estimated from
the activation energy for â-phenyl-p-methoxypropiophenone (4.3
kcal/mol),13 taking into account that its activation energy
incorporates an energy difference of ca. 2-3 kcal/mol between
the T1 state with π,π* configuration and the T2 state with n,π*
configuration.13,49,57 Thus, the activation energy for â-phenyl
quenching in â-phenylpropiophenone is ca. 1.3-2.3 kcal/mol.
From the temperature dependence on the triplet decay kinetics
of 1-rac, we obtained activation energies for the triplet decay
The lowest triplet-excited state (T1) in the case of diketones
1 is n,π* in character, as is the case with â-phenylpropiophe-
none. The intramolecular quenching in the latter occurs so
efficiently that its triplet lifetime is short (<1 ns), and the
R-cleavage reactions cannot compete with triplet deactivation.13
For this reason, â-phenylpropiophenone is relatively photostable
and has a very low photodecomposition quantum yield of ca.
10-3 13
In the case of 1-rac, the photodecomposition yield
.
increases when compared to that of â-phenylpropiophenone
because the 1-rac triplet lifetime is much longer, the carbon-
centered radical formed is more stable, and â-cleavage could
be an additional photodecomposition process. The large con-
tribution from the photocleavage process is observed in the laser
(54) Only in the case of 7 in methanol is a shortening of the lifetime
observed, which may be due to an inefficient self-quenching process (3 ×
104 s-1).
(55) Bays, J. P.; Encinas, M. V.; Small, J., R. D.; Scaiano, J. C. J. Am.
Chem. Soc. 1980, 102, 727.
(56) The photocleavage rate constant incorporates the rate constant for
R-cleavage as well as any contribution from a â-cleavage process.
(57) Encinas, M. V.; Lissi, E. A.; Lemp, E.; Zanocco, A.; Scaiano, J. C.
J. Am. Chem. Soc. 1983, 105, 1856.

3100 J. Am. Chem. Soc., Vol. 121, No. 13, 1999
Moorthy et al.
Scheme 2
described by the sum of the reactivities from the T1 and T2 states,
taking into account the relative populations in each state.
Provided that there exists a rapid equilibrium between the T1
and T2 states, i.e., k-1 > kCT and k1 > kint, the triplet decay can
be described by eq 4, where K is the equilibrium constant
between the two triplet states:
kd ) kint + KkCT + Kkcleavage
(4)
of 5.9-6.4 kcal/mol (Table 3). These energies obviously include
the activation barriers for photocleavage and â-phenyl quench-
ing. Nevertheless, the activation energy for â-phenyl quenching
in 1-rac is 4-5 kcal/mol higher than that for â-phenylpro-
piophenone, which can readily account for the decrease of ca.
3 orders of magnitude for the â-phenyl quenching rate constant
in 1-rac. The marginal difference observed in the lifetimes for
1-rac in the two different solvents, with moderate difference in
their viscosities (η(methanol) ) 0.55 and η(acetonitrile) ) 0.36
cP),58 is consistent with the rate-limiting step for the intra-
molecular quenching process being the bond rotation.
Thus, the overall triplet decay rate constant is given by the
sum of the intrinsic triplet decay, the product of the equilibrium
constant and the rate constant for â-phenyl quenching, and the
product of the equilibrium constant and the photocleavage rate
constant. Since the amount of transients formed due to the
photocleavage reaction for 2 and 3 is comparable to that
observed for â-phenyl-p-methoxypropiophenone and much
smaller than that for 1-rac, we will consider that only the first
two terms contribute significantly for the triplet decay of
diketones 2 and 3. The intrinsic decay rate constants in diketones
2 and 3 can be equated to those for the model compounds 6
and 7, respectively. From the lifetime for 1-meso, we have
established that kCT is >108 s-1. The value of K is directly
related to the energy spacing between T1 and T2. In this respect,
the value of K should be smaller for 3 than that for diketones
2, since the dimethoxy substitution in the benzoyl ring leads to
additional stabilization of the π,π* state. In addition, the
methoxy substitution of the benzoyl rings will lower the electron
affinity of the carbonyl group, which would lead to a decrease
of kCT. Both of these factors are expected to decrease the
efficiency of the â-phenyl quenching reaction in 3 when
compared to that in 2.
The introduction of p-methoxy substituents in aryl ketones,
such as diketones 2 and 3, leads to the stabilization of the triplet
with π,π* configuration relative to the n,π* triplet and to a
reduction of the photoreactivity of these compounds.51,52 There
are two models to explain this decrease in reactivity for ketones
where the lowest triplet state has a π,π* configuration. One
model assumes that the reactivity occurs from the lowest triplet
which has a mixed character, whereas the second model assumes
that the upper n,π* triplet state has to be thermally populated
for reactions to occur.49,59 The latter model was first proposed
to explain the difference observed for the intramolecular
hydrogen abstraction efficiency of phenyl ketones for which
3
3
the lowest triplet states had a n,π* or π,π* configuration.49
Steel and co-workers considered both models and showed that
they can be differentiated by measuring the temperature
dependence of the photoreaction rate constants.59 In the case
of the mixing model, the changes in reactivity would appear in
the preexponential factors, whereas the equilibration model
predicts higher activation energies as the energy gap between
the lowest triplet with π,π* character and the T2 state with n,π*
character increases. The reactivity differences for the intramo-
lecular hydrogen abstraction reactions involving triplet ac-
etophenone derivatives were mainly due to changes in the
activation energies, which is consistent with the equilibration
model.59 Additional support for the equilibration model is
provided by the fact that p-methoxy substitution on aryl alkyl
ketones led to the same increase in the activation energies for
reactions that have different mechanisms, such as the hydrogens
abstraction and R-cleavage processes.57 Furthermore, the slower
intramolecular deactivation of â-phenyl-p-methoxypropiophe-
none when compared to that of â-phenylpropiophenone has been
explained using the thermal equilibration model.13 For this
reason, we adopt this model to explain the differences for the
triplet lifetimes of the meso and racemic diastereomers of
diketones 2 and 3.
The significant differences observed for the triplet lifetimes
of the meso and racemic diastereomers of 2 and 3 can now be
rationalized on the basis of conformational preferences in the
T1 state and by considering the fact that the lifetime for T2 is
too short to permit bond rotations oVer torsional barriers. The
latter is easily established by comparing the T2 lifetime (<1
ns) with that to overcome the rotational barrier in the case of
1-rac, which is g250 ns. For this reason, the conformation in
the T2 state should be considered to be frozen in the geometry
defined by the T1 state from which thermal population occurs.
As a result, population of the T2 state in a conformation in which
the phenyl ring and the carbonyl group are in an anti relationship
does not lead to â-phenyl quenching, and the triplet decay will
be determined by the intrinsic decay (kint) from the T1 state.
On the other hand, when T2 is populated in a conformation in
which the â-phenyl is gauche to the excited carbonyl group,
intramolecular quenching (KkCT) can occur efficiently, so that
depopulation through this channel primarily dictates the triplet
decay.
In the case of racemic diastereomers, the â-phenyl ring and
the carbonyl are distal in the lowest energy conformations, and
when the T2 state is populated no deactivation by â-phenyl
quenching occurs. Hence, the triplet decay is determined by
kint, and the lifetimes are long. Indeed, the lifetimes in the case
of 2-H-rac correspond to those measured for 6, which lacks
the phenyl rings in the â-positions. In the case of the meso
diastereomers, both conformations have at least one carbonyl
group in a gauche relationship with a â-phenyl ring. Conse-
quently, irrespective of the conformation in which T2 is
populated, the triplet decay occurs primarily through intra-
molecular quenching (kâ), and hence the lifetimes are shorter
than those for the racemic diastereomer. In addition, no slow
decay component is expected for the H-gauche conformer of
The intramolecular â-phenyl quenching and the photocleavage
reactions occur only for the triplet state with n,π* character (see
above). For this reason, â-phenyl quenching for compounds 2
and 3 occurs from the upper triplet state (Scheme 2), and the
triplet lifetimes of both diastereomers of these compounds are
longer than that for 1. The overall triplet decay (kd) can be
(58) Scaiano, J. C. In Handbook of Organic Photochemistry; Scaiano,
J. C., Ed.; CRC Press: Boca Raton, FL, 1989; Vol. II, p 343.
(59) Berger, M.; McAlpine, E.; Steel, C. J. Am. Chem. Soc. 1978, 100,
5147.

Triplet Lifetimes of Butane-1,4-dione Diastereomers
J. Am. Chem. Soc., Vol. 121, No. 13, 1999 3101
the meso diastereomers, since intramolecular energy transfer
between the benzoyl moieties is fast.
of magnitude as the deactivation rate constants, one would
predict a decay which corresponds to the sum of two exponen-
tials (small and large), with each exponential factor representing
a combination of all rate constants.
The above simple explanation which accounts for the
discrimination in the triplet lifetimes of the meso and racemic
diastereomers does not explain why the lifetime of 3-H-rac is
shorter than that observed for the model compound 7 and why
p-methoxy substitution in the â-phenyl ring of the racemic
diastereomers of 2-OCH3 and 3-OCH3 also leads to a reduction
of the triplet lifetimes when compared to those of the model
compounds 6 and 7. There are two possibilities for the increase
in the overall triplet deactivation rate constant: either a
contribution from â-phenyl quenching (KkCT) or a contribution
from the photocleavage reactions (Kkcleavage) (eq 4).
The second limiting scenario occurs when the T1 lifetime is
much longer than the time required for conformational equili-
bration to be achieved. A single-exponential decay is expected,
which is given by the sum of the decays from each conformer
with fractional equilibrium populations taken into account. In
this respect, the decay from the Ph-anti and Bz-anti conformers
formed by the rotation from the H-anti conformation corre-
sponds to a leakage pathway because â-phenyl quenching from
the former conformers is much faster than the intrinsic decay
from the H-anti conformation. The shorter lifetime for the
diketones with p-methoxy substitution on the â-phenyl ring can
be understood on the basis of an enhancement of the â-phenyl
quenching rate constant (KkCT, see below), which would lead
to a more efficient leakage through the Ph-anti and Bz-anti
conformations. In contrast, the increased triplet decay rate
constant for 3-H-rac when compared to that of 2-H-rac is more
difficult to interpret. One explanation would be that the rotational
barriers between the conformers are lower for 3-H than those
for 2-H. Unfortunately, the calculations performed do not lead
to a conclusive answer to this question, since the calculated
barriers may not represent the lowest energy path between the
conformers. For this reason, subtle changes in the barriers which
would lead to a change in rate constants by a factor of ca. 2
(∼0.5 kcal/mol) are within the errors of these calculations.
Substituent Effect. In the case of the meso diastereomers of
2 and 3, where â-phenyl quenching is the rate-limiting factor
for the triplet decay, the triplet lifetime depends on the nature
of the para substituent on the â-phenyl ring (Table 1). While
the lifetime for 2-CF3-meso approaches that for 2-CF3-rac, the
lifetime for 2-OCH3-meso is so short as to be undetectable on
the nanosecond time scale. This dependence can be explained
by the electronic effect of the substituents on kCT, where
electron-donating groups enhance this rate constant, whereas
electron-withdrawing groups decrease the charge-transfer ef-
ficiency. Assuming a lifetime of 10 and 3 ns for 2-OCH3-meso
in methanol and acetonitrile, respectively,60 a linear correlation
is obtained for the triplet decay of meso diastereomers of
diketones 2 with Hammett σ+ constants of the substituents (F+
) -1.9 ( 0.2 and -2.1 ( 0.1 in methanol and acetonitrile,
respectively), which falls in excellent agreement with that
reported for a broad set of â-phenyl-p-methoxypropiophenone
derivatives (F+ ) -1.8 ( 0.2 in methanol).33 This clearly attests
to the fact that KkCT is the process that controls the decay in
the meso diastereomers of the diketones 2 and 3. In contrast,
only a weak correlation with σ+ was observed for n,π*-excited
â-phenylpropiophenones,13 where rotational barriers probably
constitutes the rate-limiting step.
The efficiency of the R- and â-cleavage reactions (Scheme
1) for compounds 3-H, 2-OCH3, and 3-OCH3 is enhanced when
compared to that for 2-H because substitution of electron-
donating groups on the benzoyl moiety will stabilize the benzoyl
radicals formed in the R-cleavage reactions, whereas p-methoxy
substitution on the â-phenyl ring stabilizes the radical formation
in both cleavage processes. If the increase in kd is solely due to
an enhancement in the photodecomposition yield, the cleavage
quantum yields can be equated to (kd - kint)/kd, where kint
corresponds to the lifetime of the model compounds 6 and 7.
This analysis leads to quantum yields in excess of 0.4.
Furthermore, in the laser flash photolysis studies, the residuals
observed after the triplet decay correspond to the radicals
formed. Only a slight increase was observed for the residuals
when 3-H-rac, 2-OCH3-rac, and 3-OCH3-rac are compared
to 2-H-rac. In the latter case, the photocleavage reaction is not
significant since the 2-H-rac triplet lifetime is the same as that
for the model compound 6. The high quantum yields calculated
above and the small increase in the residuals observed in the
kinetic studies are clearly inconsistent with the decrease of the
triplet lifetimes for 3-H-rac, 2-OCH3-rac, and 3-OCH3-rac
being solely due to an increase in the photocleavage process
(Kkcleavage).
The second possibility for the decrease in the triplet lifetimes
of the racemic diastereomers of 3-H, 2-OCH3, and 3-OCH3 is
a contribution of â-phenyl quenching to the excited triplet
deactivation. This intramolecular quenching process can occur
from the Ph-anti and Bz-anti conformers. Population of these
conformers can occur either directly upon excitation or by
rotation from the most stable H-anti conformation during the
triplet lifetime. There are two limiting scenarios for the
relationship between the time necessary for bond rotation and
the triplet lifetimes. If rotations are slow compared to the triplet
lifetime, excitation into T1 leads to the formation of three
conformers, each of which would be considered to be frozen in
their staggered conformation. Consequently, one would expect
a decay which corresponds to the sum of two exponentials in
which the fast component corresponds to the decay through
â-phenyl quenching of the Ph-anti and Bz-anti conformations
and the longer lifetime corresponds to the decay from the H-anti
conformation. No double-exponential decays were observed,
which suggests that the Ph-anti and Bz-anti conformations are
not directly excited to a significant extent. In addition, the
relative population of the triplet conformers in this scenario
corresponds to their equilibrium population in the ground state.
Based on the relative energies calculated for 2-H-rac and 3-H-
rac, only a very small amount of triplets (<1%) would be
excited in the Ph-anti and Bz-anti conformations. In this respect,
the calculations are in agreement with the fact that only a
monoexponential decay was observed. Likewise, for a scenario
in which the rotation rate constants would be of the same order
The activation energies measured for the triplet decay of the
meso diastereomers incorporate the energy difference between
the T2 and T1 states and the activation energy for the â-phenyl
quenching:
Ea ) E(T2 - T1) + Ea(â)
(5)
The preexponential factors in acetonitrile (Table 3) do not
depend on the nature of the substituent on the â-phenyl ring.
(60) A lifetime of 10 ns for 2-OCH3-meso in methanol was assumed
since the signal corresponding to the triplet was not detected within the
time resolution of the nanosecond laser flash photolysis setup. This value
was considered reasonable by comparison to an estimated lifetime of 8 ns
for â-(4-anisyl)-p-methoxypropiophenone.33 Since the lifetimes in aceto-
nitrile are approximately 3-fold shorter when compared those in methanol,
a lifetime of 3 ns was derived for 2-OCH3-meso in acetonitrile.

3102 J. Am. Chem. Soc., Vol. 121, No. 13, 1999
Moorthy et al.
For this reason, the data in acetonitrile will be employed to
analyze the effect of substituents on the activation energies. In
methanol, a more complex behavior is observed, probably due
to the effect of hydrogen bonding of the solvent to the diketone.
Indeed, the effect of hydrogen bonding on the rate of â-phenyl
quenching has been proposed before for the intramolecular
triplet quenching of â-phenylpropiophenone13 and R-guaiacoxy-
acetoveratrone.61
to a larger population of the Ph-anti and Bz-anti, which should
lead to shorter lifetimes or the observation of double-exponential
decays. In this respect, the complexation of the enantiomers of
the racemic diketones to supramolecular structures with chiral
environments could lead to the observation of different triplet
lifetimes for each enantiomer.
In conclusion, a remarkable diastereomeric discrimination is
demonstrated in the triplet lifetimes of racemic and meso
diastereomers of a broad set of 1,4-diketones. The differences
in the lifetimes for meso and racemic diastereomers are
reconciled from preferential operation of intramolecular â-phen-
yl quenching in the former. It is shown that, by making an
appropriate choice of the substitution in the â-phenyl ring and
by controlling the spacing between the n,π* and π,π* triplet
states through substitution in the benzoyl group, the triplet
lifetimes can be modulated. The detailed study on a broad set
of substrates has permitted an appreciably longer lifetime (230
ns) to be measured for a pure n,π* triplet-excited â-phenylpro-
piophenone derivative for the first time.
The value for the spacing between the T1 and T2 states may,
in principle, be determined from the phosphorescence spectra.
Unfortunately, the structureless broad spectral features did not
permit unequivocal identification of the maxima for n,π* and
π,π* transitions. Assuming that substitution on the R-phenyl
ring in the diketones does not affect the spacing between these
levels, one calculates that the difference in the activation
energies between 2-H-meso and 2-CH3-meso and between 3-H-
meso and 3-OCH3-meso is 0.6 and 1.8 kcal/mol, respectively.
The lower activation energies when a p-methyl or p-methoxy
substituent is introduced in the â-phenyl ring are in agreement
with the explanation that electron-donating substituents increase
the charge-transfer rate constant for â-phenyl quenching.
The lifetimes for the meso diastereomers of compound 3 are
much longer than those for compounds 2. For example, the
dimethoxy substitution causes a 3-fold increase in triplet lifetime
for 3-H-meso when compared to that for 2-H-meso. Also, the
triplet, which is not observable in the case of 2-OCH3-meso,
becomes sufficiently long-lived in 3-OCH3-meso to be moni-
tored in a nanosecond time scale. As expected, the dimethoxy
substitution increases the energy spacing between T2 and T1 by
0.4 kcal/mol. However, changes in the preexponential factor
also contribute to the increase in the lifetimes, suggesting that
the steric crowding created by the substituent at the ortho
position is important.
Tunability of the Triplet Lifetimes. The present investiga-
tion has shown that conformational preferences and electronic
factors can be exploited to modulate the triplet lifetimes over a
range which spans 4 orders of magnitude. This property can be
explored to probe organized environments. The tunability of
lifetimes is related to the configuration of the lowest excited
state and the conformational preferences, which determine the
competition between intrinsic decay and intramolecular â-phenyl
quenching. In the case of the meso diastereomers of 1-3, the
lifetimes are always determined by â-phenyl quenching because
both staggered conformations have at least one carbonyl moiety
and the â-phenyl group in a gauche relationship. The lifetimes
can be tuned by changing the configuration of the lowest triplet
state and, in the case of π,π* triplets, by changing the T1-T2
energy spacing. Further control can be achieved by changing
the charge-transfer efficiency with appropriate substituents on
the â-phenyl ring. In addition, the lifetimes of the meso
diastereomers are influenced by the nature of the solvent, and
this property can be explored to probe the local environments
of supramolecular structures.
Experimental Section
General Aspects and Synthesis. The detailed description of the
general instrumentation, synthesis of various compounds, and their
characterization is provided in the Supporting Information.
Laser Flash Photolysis. Nanosecond laser flash photolysis experi-
ments were carried out using the pulses from a Lumonics Excimer laser
(308 nm) or a Spectra Physics Nd:YAG laser (266 nm). The full details
of signal detection, monitoring, and data processing have been described
elsewhere.62
The transient absorption spectra were recorded using a flow system
to avoid interference of the photoproducts resulting from several laser
shots. For this purpose, a freshly prepared solution was pumped, under
nitrogen atmosphere, through a custom-designed Suprasil-cell (7 × 7
mm) at a rate of 1.5-2.0 mL/min. As a result, a fresh solution was
irradiated by each laser shot. For kinetic measurements, the samples
were contained in cuvettes made of Suprasil quartz cells (7 × 7 mm),
sealed with rubber septa, and deoxygenated with nitrogen for >15-
20 min. The absorptions of the solutions at the excitation wavelengths
were in the range 0.07-0.7. Approximate concentrations for an
absorbance of 0.3 in the Suprasil cells are as follow: 1, (2-3) × 10-5
M at 266 nm; 2, (0.5-1) × 10-4 M at 308 nm; and 3, (1-2) × 10-5
M at 308 nm.
For temperature-dependent studies, a Brucker B-VT-100 temperature
control system was employed to control the sample temperature. The
sample holder was surrounded by an in-house quartz jacket, which was
equipped with four windows for excitation and monitoring beams. The
liquid nitrogen, which was boiled off from the Dewar reservoir and
allowed to pass through a heating element, was let into the quartz jacket.
The degree of heating was set by the Bruker control unit. Prior to each
measurement, the sample was allowed to equilibrate at the defined
temperature for 15-20 min. The temperature of the sample holder was
registered by a thermocouple (Omega, model DP 2000), and the
information was transferred to a computer.
Photodecomposition Studies. The relative photodecomposition rates
for both diastereomers of diketones 1 and 2-H in acetonitrile and
methanol were followed by HPLC (Hewlett-Packard series 1100). The
samples were filtered through Millipore filters into quartz cells. After
N2 was bubbled for at least 25 min, the absorbances at 254 nm were
matched by addition of deoxygenated solvent. These solutions were
irradiated in a Rayonet reactor with four 254-nm lamps, and samples
were drawn at appropriate irradiation times. These samples were
analyzed by UV-vis absorption and HPLC (3:1 acetonitrile/water, 0.5
mL flow, 30 °C, Supelco Supelcosil LC-PAH column, 10 cm × 4.6
mm, 3 µm particle size, detection at 250 nm for 1 and 280 nm for 2).
Several photodecomposition products were observed by HPLC. The
purity of the peak for the diketone was checked by the HP series 1100
data analysis software by comparison of the absorbance spectra taken
The longer triplet lifetimes for the racemic diastereomers are
due to the anti relationship of the carbonyl and â-phenyl moieties
for the conformer with lowest energy, which precludes the
intramolecular â-phenyl quenching process. However, some
variability of the triplet lifetimes has been observed, probably
due to leakage through the Ph-anti and Bz-anti conformers from
which â-phenyl quenching is possible. Inclusion of the racemic
diketones in appropriate supramolecular structures could lead
(61) Schmidt, J. A.; Goldsmidt, E.; Heitner, C.; Scaiano, J. C.; Berinstain,
A. B.; Johnston, L. J. In Photochemistry of Lignocellulosic Materials;
Heitner, C., Scaiano, J. C., Eds.; American Chemical Society: Washington,
DC, 1993; Vol. 531, p 122.
(62) Liao, Y.; Bohne, C. J. Phys. Chem. 1996, 100, 734.

Triplet Lifetimes of Butane-1,4-dione Diastereomers
J. Am. Chem. Soc., Vol. 121, No. 13, 1999 3103
at regular intervals across the width of the peak. Repeated injections
of equal volumes of the same sample led to a reproducibility of 1% of
the absorption values measured. The rate of decomposition was
determined from the amount of decomposed diketone measured from
the decrease in the absorption of the HPLC peak corresponding to the
diketone. The photodecompositions of 1-meso, 2-H-rac, and 2-H-meso
were kept below 10%, whereas the decomposition for 1-rac was kept
below 30%.
Calculations. Molecular mechanics calculations on 2-H and 3-H
were carried out using the MMX force field in the PCMODEL program
(Serena Software, Bloomington, IN). The effect of conjugation was
explicitly included by treating the carbonyl group, phenyl rings, and
methoxy substituents as π atoms. Minima corresponding to H-anti and
H-gauche conformers were optimized in the meso compounds, while
H-anti, Ph-anti, and Bz-anti were considered for the racemic structures.
Since the molecules are sterically crowded and considerable confor-
mational flexibility is possible, several initial geometries (typically 10)
were used to obtain the lowest energy conformations. Barriers for
interconversion between the different minima were calculated by
varying the torsional angle about the central C-C bonds in 2-deg
intervals and reoptimizing the structure at each stage. This procedure
need not necessarily follow the lowest energy pathway. Hence, the
computed values represent upper limits for the barriers.
Acknowledgment. This research was supported by the
Natural Sciences and Engineering Research Council of Canada
(NSERC). R.B.S. thanks CSIR (New Delhi) for a fellowship.
The authors at the University of Victoria thank Luis Netter for
software development.
Supporting Information Available: Details of the general
instrumentation, syntheses of various compounds, and figures
showing the calculated structures of the staggered conformations
of 2-H and 3-H with the lowest calculated energies (PDF). This
material is available free of charge via the Internet at
http://pubs.acs.org.
JA9818708