Photolysis of thymine oxetanes produces triplet excited carbonyl compounds with high efficiency

Full text of DOI:10.1021/ja0040363

J. Am. Chem. Soc. 2001, 123, 3145-3146
3145
Synthesis of 1-3 is described elsewhere.15 These oxetanes are
thermally stable. For example, heating 1 at 100 °C for 8 h results
in no noticeable decomposition. The UV-vis absorption spectra
of these compounds show either a shoulder (1 and 2) or a peak
(3) in the region of 220-230 nm and then an asymptotic decay
to wavelengths >310 nm with no clear or distinct absorption
onset. No fluorescence could be detected from the oxetanes (CH3-
CN solutions, N2-purged).
Photolysis of Thymine Oxetanes Produces Triplet
Excited Carbonyl Compounds with High Efficiency
Arul Joseph and Daniel E. Falvey*
Department of Chemistry and Biochemistry
UniVersity of Maryland, College Park, Maryland 20742-2021
Photolysis of the oxetanes results in the clean and efficient
formation of DMT and the corresponding carbonyl compound.
Solutions of 1-3 in N2-purged CH3CN were irradiated with the
unfocused 266 nm output from a pulsed Nd:YAG laser and the
photolysis mixtures were analyzed by HPLC. Average yields of
the photoproducts from triplicate determinations range from 78%
to 96% (Table 1). No side products could be detected by HPLC.
The quantum yields for photolysis (ΦK) were determined in a
similar way. The amount of oxetane remaining following pho-
tolysis was determined by HPLC and the number of absorbed
photons was determined by measuring the average laser power
before and after photolysis with a pyroelectric joulemeter.16 In
all cases the conversion of the oxetane was kept below 10%. As
with the chemical yields, the ΦK values in Table 1 represent
averages of three independent determinations where each mea-
surement fell within 0.1 of the stated average.
Shown in Figure 1 is the transient spectrum resulting from LFP
(266 nm, 6 ns, 2-5 mJ/pulse) of oxetane 1 in CH3CN. This
spectrum shows excellent agreement with the well-characterized
triplet-state spectrum of benzophenone (λmax ) 525 nm).17 The
remaining oxetanes also give spectra of the corresponding al-
dehydes: 2 giving triplet tolualdehyde (λmax 350 nm) and 3 giving
triplet anisaldehyde (λmax ) 390 nm).18 For comparison purposes
we generated the same triplets through direct irradiation of the
carbonyl compounds, again using our 266 nm excitation source,
and again, excellent agreement was observed between the oxetane-
derived spectra and those of the authentic triplet carbonyl
compounds.
ReceiVed NoVember 21, 2000
ReVised Manuscript ReceiVed February 2, 2001
Most photochemical reactions produce ground-state products.
This is generally attributed to efficient electronic state mixing at
conical intersections (i.e., geometries where the excited-state
potential energy surface approaches the energy of the ground-
state surface).1,2 Reactions where stable products are produced
in their electronically excited state are termed “adiabatic photo-
chemical reactions”.3-5 Aside from proton transfers6 and formation
of various open-shell species, adiabatic photochemical processes
are extremely rare and limited to only a few structural classes.
Photochemical ring-opening of cyclobutenes7 (including dewar
benzenes)8,9 often yields electronically excited products. Photo-
induced cycloreversion of [2+2]10 and [4+4]3,11 dimers of arenes
can lead to the excited states of the final products. Photolysis of
1,1,2,2-tetramethyldioxetane (formally the [2+2] cycloadduct of
acetone) produces one triplet-state acetone and one ground-state
acetone.4,12,13
Pursuant to our longstanding interest in DNA photodamage
and photoenzymatic repair, we have prepared several oxetane
adducts of 1,3-dimethylthymine (DMT) with benzophenone (1),
tolualdehyde (2), and anisaldehyde (3).14,15 We studied their
behavior under photoinduced electron transfer conditions designed
to model the recently discovered photoenzymatic repair of the
DNA (6-4) photoproduct. In the course of these investigations
we had occasion to examine the direct photolysis of oxetanes
1-3. As reported previously such photolyses effect cyclorever-
sion, cleanly producing thymine and the carbonyl compound.
Herein are described laser flash photolysis (LFP) experiments
which show the photolysis of 1-3 results in a rare, adiabatic,
photochemical reaction producing ground-state 1,3-dimethylthy-
mine (DMT) and the excited triplet state of the carbonyl (eq 1).
The excited triplet carbonyl compounds are formed very
efficiently from photolysis of the oxetanes. This was ascertained
by comparing the initial intensity of transient signal from the
oxetanes with that from an optically matched solution of the
carbonyl compound. (Benzophenone as well as the aromatic
aldehydes all form their excited triplet states rapidly and with
unit quantum efficiency.17) These ratios, which correspond to the
quantum yields of adiabatic triplet formation from oxetane
photolysis (ΦT), range from 0.3 to 0.6 (Table 1).
Because adiabatic photochemical reactions are rare, it is
necessary to consider several alternative ways in which LFP of
the oxetanes could lead to formation of excited-state products.
The most obvious concern would be that the observed signals
originate from carbonyl impurities in the oxetane sample. HPLC
analysis of the oxetane sample both prior to and following the
LFP experiments showed that there was <1% of the carbonyl
compound at any point in the experiment.
(1) Fo¨rster, T. Pure Appl. Chem. 1973, 34, 225-234.
(2) Klessinger, M.; Michl, J. Excited States and Photochemistry of Organic
Molecules; VCH: New York, 1995.
(10) Noh, T.; Gan, H.; Halfon, S.; Hrnjez, B.; Yang, N.-c. C. J. Am. Chem.
Soc. 1997, 119, 7470-7482.
(11) Mori, Y.; Maeda, K. Bull. Chem. Soc. Jpn. 1997, 70, 869-875.
(12) Turro, N. J.; Devaquet, A. J. Am. Chem. Soc. 1975, 97, 3878-3860.
(13) Smith, K. K.; Koo, J.-Y.; Schuster, G. B.; Kaufmann, K. J. Chem.
Phys. Lett. 1977, 48, 267-2270.
(3) Fo¨rster, T. Pure Appl. Chem. 1970, 24, 443-449.
(4) Turro, N. J.; Lechtken, P.; Lyons, A.; Huatala, R. P.; Carnahan, E.;
Katz, T. J. J. Am. Chem. Soc. 1973, 95, 2035-2037.
(5) Turro, N. J.; McVey, J.; Ramamurthy, V.; Lechken, P. Angew. Chem.,
Int. Ed. Engl. 1979, 18, 572-586.
(14) Prakash, G.; Falvey, D. E. J. Am. Chem. Soc. 1995, 117, 11375-
11376.
(15) Joseph, A.; Prakash, G.; Falvey, D. E. J. Am. Chem. Soc. 2000, 122,
11219-11225.
(6) Solntsev, K. M.; Huppert, D.; Agmon, N.; Tolbert, L. J. Phys. Chem.
A 2000, 104, 4658-4669.
(7) Leigh, W. J.; Postigo, J. A.; Zheng, K. C. Can. J. Chem. 1996, 74,
951-964.
(16) This procedure slightly underestimates the actual number of photons
absorbed because there are some losses due to reflections from the walls of
the cuvette. However, we estimate these to be <10%.
(17) Murov, S. L.; Carmichael, I.; Hug, G. L. Handbook of Photochemistry,
2nd ed.; Marcel Dekker: New York, 1993; p 420.
(8) Yang, N. C.; Carr, R. V.; Li, E.; McVey, J. K.; Rice, S. A. J. Am.
Chem. Soc. 1974, 96, 2297-2298.
(9) Carr, R. V.; Kim, B.; McVey, J. K.; Yang, N. C.; Gerhartz, W.; Michl,
J. Chem. Phys. Lett. 1976, 39, 57-60.
(18) Go¨rner, H.; Kuhn, H. J. J. Phys. Chem. 1986, 90, 5946-5955.
10.1021/ja0040363 CCC: $20.00 © 2001 American Chemical Society
Published on Web 03/07/2001

3146 J. Am. Chem. Soc., Vol. 123, No. 13, 2001
Communications to the Editor
Table 1. Chemical Yields, Quantum Yields, and Enthalpy Changes
for the Splitting of Oxetane Adducts 1-3
yielda
ΦK
ΦT
∆Hsplitb (kcal/mol)
ETc (kcal/mol)
1
2
3
78
96
92
0.5
0.8
0.7
0.4
0.3
0.6
-22
-16
-16
72
72
69
a Yield based on carbonyl compound. b From AM1 calculation for
ground states. ∆Hsplit) ∆Hf(carbonyl compound) + ∆Hf(DMT) -
∆Hf(oxetane). c Triplet state energy of carbonyl compound, ref 17.
Figure 2. Potential energy diagram for the adiabatic photochemical
cycloreversion of thymine oxetane adduct 1.
lifetime to allow for a fraction of the excited species to cross to
the ground-state surface and from there partition between ground-
state products and reactants.
A potential energy diagram for 1 based on these considerations,
provided as Figure 2, is consistent with our observations.
However, there are still several aspects of the mechanism that
remain unclear. For example, it is not known whether intersystem
crossing to the triplet surface occurs before bond scission, after
the products are formed, or at some intermediate geometry. Also
the nature of the intermediate state responsible for reversion to
the ground state is unclear. We suggest that this is a 1,4 diradical
derived from C6-C bond scission as indicated in Figure 2.
However, a diradical derived from C5-O bond scission or some
sort of exciplex cannot be excluded. Tetramethyldioxetane
photolysis is thought to occur through a dissociative excited state
that does not vibrationally equilibrate prior to bond dissociation.4
If this holds for 1-3, then the 266 nm excitation would provide
a larger than estimated driving force for adiabatic dissociation.
We are unaware of previous reports of adiabatic cycloreversion
of oxetanes (or indeed any previous investigations of oxetane
photolysis). Numerous [2+2] cycloadducts of pyrimidine bases
in DNA are found in nature. For example, pyrimidine-cyclobu-
tane dimers are formed through the direct photolysis of DNA.20
Likewise photolysis of psoralen derivatives in the presence of
DNA leads to [2+2] adducts between psoralen and the pyrimidine
bases.21 There is also evidence that urocanic acid forms [2+2]
cycloadducts with DNA bases.22,23 The possibility that these bio-
logically important base cycloadducts might photocleave in a sim-
ilar adiabatic fashion will be the subject of future investigations.
Figure 1. Transient absorption spectra from laser flash photolysis (355
nm. 2 mJ, 6 ns) of oxetane 1 in CH3CN taken at 0.3, 0.6, 1.0, 2.0, and
5.0 µs following the laser pulse. The inset shows the dependence of the
oxetane signal on the laser pulse energy, which was determined by the
intensity of the signal from LFP excitation of benzophenone.
If the triplet carbonyl signals observed in the photolysis of 1-3
are due to photoproducts that accumulate from repetitive laser
photolysis of the samples, then we would expect that their intensity
will increase with the number of laser pulses applied to a given
sample. Therefore, we examined the LFP signals for the carbonyl
triplets on fresh solutions of 1-3 which were each subjected to
a single laser pulse (typically data are accumulated by signal
averaging ca. 20 shots). In no case did the intensity of these signals
differ from those of signals obtained after 10 shots into the same
solution. This provides further evidence against the signals being
due to accumulated photoproducts.
Also considered was the possibility that the excited triplets
result from a biphotonic process where the first photon effects a
diabatic splitting of the oxetane ring and a second photon from
the same laser pulse excites the carbonyl compound product. This
situation would result in a quadratic dependence of the carbonyl
triplet signal on the laser pulse energy. The intensity of the triplet
signal from oxetanes 1-3 on the excitation pulse energy was
examined and in each case found to be linear. This is illustrated
for the case of 1 in the inset to Figure 1 and similar behavior
was observed for 2 and 3. Thus the biphotonic mechanism can
be ruled out.
Acknowledgment. This work was supported by the Chemistry
Division of the National Science Foundation.
Supporting Information Available: Transient absorption spectra
from photolysis of oxetanes 2 and 3 (PDF). This material is available
free of charge via the Internet at http://pubs.acs.org.
The adiabatic process is energetically feasible. The lowest
singlet-state energy of oxetanes 1-3 can be estimated at 90 kcal/
mol from their absorption onsets of ca. 310 nm. The triplet state
energies for the carbonyl fragments range from 69 (1) to 71 kcal/
mol. Semiempirical calculations (AM1) provide the enthalpy for
the ground-state-to-ground-state splitting (∆Hsplit) of -22 kcal/
mol for 1 and -16 kcal/mol for both 2 and 3.19 Thus the estimated
energy changes in going from the vibrationally equilibrated singlet
states of the oxetanes to the observed triplet states of the carbonyls
range from -40 to -34 kcal/mol. Because ΦK < 1 there must
be some state along the reaction coordinate with a sufficient
JA0040363
(19) Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P. J. Am.
Chem. Soc. 1985, 107, 3902-3909.
(20) Cadet, J.; Vigny, P. In Bioorganic Photochemistry; Morrison, H., Ed.;
Wiley: New York, 1990; pp 1-272.
(21) Shi, Y.-B.; Lipson, S. E.; Chi, D. Y.; Spielmann, H. P.; Montforte, J.
A.; Hearst, J. E. In Bioorganic Photochemisty: Volume 1: Photochemistry
and the Nucleic Acids; Morrison, H., Ed.; Wiley: New York, 1990; pp 341-
425.
(22) Morrison, H.; Bernasconi, C.; Pandey, G. Photochem. Photobiol. 1983,
38, 23-27.
(23) Mohammad, T.; Morrison, H.; HogenEsch, H. Photochem. Photobiol.
1999, 69, 115-135.