Acc. Chem. Res.

1980,13, 45-50 45

Exciplex Intermediates in [2 + 21 Photocycloadditions

RICHARDA. CALDWELL"
Department of Chemistry, The University of Texas at Dallas, Richardson, Texas 75080

DAVIDCREED*
Department of Chemistry, The University of Southern Mississippi, Hattiesburg, Mississippi 39401
Received August 16, 1979

I n t h e subjects which we propose t o investigate, of association were ca. -20 eu, as expected for bimole-
our inquiries should be directed, not t o what cular association. Radiative lifetimes were remarkably
others have thought, nor t o what we ourselves structure independent, al1"'l being within an order of
conjecture, but t o what we can clearly and per- magnitude of 3 X s. The red shift of the struc-
spicuously behold and with certainty deduce; for tureless emission of the exciplex or excimer relative to
knowledge is not won in a n y other way. excited monomers was due both to the enthalpy of as-
sociation and to ground-state repulsion of the closely
Rene Descartes (1596-1650) held chromophores as sketched in Figure 1.
The polar nature of exciplexes was revealed by the
Photocycloadditions and cyclodimerizations have red shifts of their emission maxima in solvents of in-
been known since the early days of organic photo- creasing polarity and by the tendency of exciplex-
chemistry. Reactions such as the Paterno-Buchi re- forming systems to give radical ions or radical ion re-
action1 and photoannelation of cyclohexenones2with lated chemistry in highly polar solvents at the expense
simple olefins afford numerous examples of compounds of the emission.12J3 Weller's classic empirical analysis
accessible at best with difficulty by other routes. The of exciplex association6bin terms of excitation energies
formation of a-truxillic vs. 0-truxinic acid from different and redox properties of the partners, eq 1, provides a
crystal modifications of cinnamic acid is a landmark of good guide to the tendency of any photochemical sys-
our knowledge of the chemistry of the organic solid tem to undergo exciplex formation. In this model sta-
state.3 The dimerization of pyrimidine residues such bilization is quantitatively accounted for by one-electron
as thymine is one of the principal reactions responsible transfer from excited donor to acceptor (or from donor
for UV inactivation in vivo of genetic material.4 to excited acceptor). For this process in hydrocarbon
For a number of years, we have studied the mecha- solvents:
nisms of such reactions, both to model bimolecular
photochemical events generally and to understand the AHo = aEo,o- [El,20x(D)- El/zred(A) 0.13 V] (1) +
factors which affect reactivity in photocycloaddition. The Dilemma. In the fifties and sixties, organic
Exeiplexes. As the synthetic potential of photo- chemistry was deluged with a series of proposed new
cycloadditions was under study by organic chemists, an reaction intermediates, and in fairness one must admit
apparently unrelated development was occurring in the that the evidence was not always unambiguous. The
laboratories of several physical chemists. The work of most notorious was the nonclassical carbonium ion.
Forster: Weller: and Birks7was clearly establishing the Others included ?r complexes in electrophilic additions
photophysics of photoassociation. Long wavelength and charge-transfer complexes between Diels-Alder
emissions a t high concentrations of pyrene and other addends. Lively debates were stimulated regarding the
arenes were ascribed to excimers, and corresponding existence, nature, and relevance of the intermediates
exciplexes were reported as a result of diffusive in the reactions in question.
quenching of excited arenes by amines, especially aryl-
(1)D. R.Arnold, Adv. Photochem., 6,301(1968);see E.Paterno and
aminesa8 These species were shown to be formed re- G. Chieffi, Gazz. Chim. Ital., 39, 341 (1909).
versibly but with essentially diffusion-controlled rate (2)P. DeMayo, Acc. Chem. Res., 4,41 (1971).
constants, and, as evidenced by temperature-depen- (3)G.M. J. Schmidt, Pure Appl. Chem., 27,647(1971);see also A. W.
K. de Jong, Ber. Dtsch. Chem. Ges., 56,818 (1923).
dence studies,6*8were bound with heats of association (4)J. G.Burr, Adu. Photochem., 6, 193 (1968).
a few (5-10) kilocalories/mole exothermic. Entropies (5)T. Forster and K. Kasper, 2.Phys. Chem. (Frankfurt am Main),
1, 275 (1954);T. Forster, Angew. Chem., Int. Ed. Engl., 8, 33 (1969).
(6)(a) H.Leonhardt and A. Weller, Ber. Bunsenges. Phys. Chem., 67,
791 (1963); (b) H. Knibbe, D. Rehm, and A. Weller, ibid., 73,839(1969).
Richard A. Caldwell was born in 1940 at Huntington, West Virginia. He (7)J. B.Birks, D. J. Dyson, and I. H.Munro, Proc. R. Soc. London,
received his Bachelor's degree from M.I.T. and his Ph.D. from The University Ser. A , 275,575 (1963).
of California at Berkeley. Following 6 years on the Cornel1 faculty, he joined (8)See B.Stevens, Adv. Photochem., 8,161(1971),for a useful review.
the University of Texas at Dallas in 1971 where he is now Professor of (9)(a) N.Mataga, T. Okada, and N. Yamamota, Chem. Phys. Lett.,
Chemistry. His interests in organic photochemical mechanisms and photo- 1,119(1967);(b) Y. Shirota, I. Tsushi, and H.Mikawa, Bull. Chem. SOC.
physics have recently been broadened by a sabbatical year at the IBM San Jpn., 47,991 (1974).
Jose Research Laboratories and now Include photoelectrochemistry, some (10)K. H.Grellmann and U. Suckow, Chem. Phys. Lett., 32, 250
aspects of polymer photochemistry,and low-temperature photochemistry and (1975).
spectroscopy. (11)Other more recent values compare well; cf. D. V. O'Connor and
David Creed was born in Colchester, England, in 1943. He received the W. R. Ware, J. Am. Chem. SOC.,98,4706 (1976);M. H.Hui and W.R.
BSc., M.Sc., and Ph.D. degrees from the University of Manchester and alter Ware, ibid., 98,4718 (1976).
postdoctoral experiences too numerous to relate on both sides of the Atlantic, (12)Cf. A. Weller in "Fast Reactions and Primary Processes in
joined the faculty of the University of Southern Mississippi in 1977. His current Chemical Kinetics", S. Claesson, Ed.,Wiley-Interscience, New York, 1967,
research interests are in reaction mechanisms in photochemistry and the *D 413.
application of photochemical electron-transfer reactions to solar energy con- (13)H. Knibbe, K. Rollig, F. P. Schafer, and A. Weller, J. Phys.
version. Chem., 47,1184 (1967).

0001-4842/80/0113-0045$01.00/0 0 1980 American Chemical Society

46 Caldwell and Creed Accounts of Chemical Research

1
tones with simple alkenes revealed the presence of a
complex with chage-transfer character formed prior to
I De+ A a 1,4-biradical. The singlet-state photodimers of 9-
cyanophenar~threne,~~ acenaphthylenejZ4and coumarinz5
or D t A” all had the endo structure which was sterically disfa-
‘\
vored but which was what was expected from the
maximum overlap structure of an excimer intermediate.
T ‘!\‘ T A structure-reactivity relationship in the triplet-state
E ,\ hyTax , D - - -A ) * reaction of substituted phenanthrenes with dimethyl
fumarate, the nonstereospeci€icity of which required a
B,4-biradical, suggested yet again the formation of a
charge-transfer coniplex prior to the biradical.26 The
inefficiency of several photoreactions could be better
accounted for by partitioning of an intermediate exci-
plex (or congener) than by quite separate pathways for
reaction and decaysz7
Clearly, this array of evidence was innpossible to ig-
r--+ noreiZa yet it, failed to define what the species was
Figure 1. Potential energy vs. intermolecular separation for (were?) or to allow any very clear generalizations that
ground- a n d excited-state association of D a n d A. AHo is the
exciplex association enthalpy, hum, is t h e energy corresponding
were predictively useful. The maximum overlap dimers
t o t h e exciplex emission maximum, a n d ERis the ground-state from singlet reactions were often accompanied by
repulsion energy. minimum overlap exo d i m e r ~ from
~ ~ 3triplet
~ reactions
or else ex0 dimers23dwere formed in triplet reactions
In 1964, @orey14suggested the intermediacy of both in related systems. The sequence triplet reactant 4-
an “oriented H complex” and lP-biradical intermediates olefins -+ “‘CT complex” 4 triplet 1,d-biradical 4
in enone photoannelation, the former to rationalize photoproducts increasingly appeared,2,14~21~23,z6 yet the
regiochemistry and intermolecular selectivity, the latter intermediacy of CT complexes was obviously a t best an
to rationalize stereorandomization. Others followed imperfect description for the symmetric dimerizations.
suit, proposing exciplexes, or at least some sort of In 1972 Farid, Doty, and published an
complex of the excited state, to explain other diverse observation of an exciplex-like emission accompanying
results. For example, Hammond et al.,15 in an early the photoaddition of phenanthrene to dimethyl fuma-
study of arene fluorescence quenching by dienes and rate. Studying the same reaction we had independently
especialiy by quadricyclene, suggested exciplex forma- made the same Observation in 1971 but un€ortunately
tion as the quenching mechanism despite the failure to were too timid to publish it. We learned later that Pac
observe exciplex emission. Not surprisingly, photo- and Sakurai a t Osaka were also observing exciplex
physicists and spectroscopists were skeptical. An ex- emissions in parallel with bimolecular photoreactionsOm
ciplex intermediate which did not luminesce? How The knowledge that exciplex-like emissions sometimes
could one claim its presence? Each of the present au- accompanied cycloadditions fueled our interest further,
thors remembers meetings in the early seventies a t but an obvious dilemma existed. Innumerable photo-
which organic chemists’ suggestions of exciplex inter- cycloadditions were known but almost never were ex-
mediates met with responses for which derision seemed ciplex or excimer emissions observed. Conversely,
a more appropriate description than skepticism. systems where bona fide excimer or exciplex emission
Yet it was undeniable that something was needed. was observed were generally photochemically unreac-
Chapman and Lura16J7had demonstrated inverse tem- tive. Did excited-state complex formation occur on the
perature dependence in the photoaddition of stilbene way to product or was it just an energy-wasting but
to tetramethylethylene, showing that a bound species stereoscopically observable dead end?
was involved. The photodimerimtion of anthracene was
(22) R. A. Caldwell, G. W. Sovocool, and R. P. Gajewski, J . Am. Chem.
enhanced in the presence of diethylanilhela or dienes,lg Soc., 95, 2549 (1973).
both quenchers of anthracene singlet, and the former (23) (a) H. Bouas-Laurent, pi. Lapouyade, A. Castellan, A. Nourmam-
in particular was well-known to form an exciplex. ode, and E. A. Chandross, 2. Phys. Chem (Frankfurt am Main), 101,39
(1976); see ah0 (lo) R. Galante, R. Lapouyade, A. Castellan, J. P. Morand,
McCullough20saw evidence for interception of some and H. Bouas-Laurent, C. R. Hebd. Seances h a d . Sci., Ser. C., 277,837
dipolar bimolecular complex of naphthalene and (1973); however, see (e) 6. Sugowdz, P. F. Collin, and W. Pi. F.Sasse,
acrylonitrile in protic solvents. Our own isotope effect Tetrahedron Lett., 3483 (1969), for an exo adduct of another phenan-
threne.
studiesz1vZ2of the Paterno -Biichi reaction of diary1 ke- (24) D. 0. Cowan and R. L. E. Drislro, J . Am. C h e n , Scc., 9Z9 6281,
6286 (1970); I. M. Nartmann, W. Hartmann, and G. 0. Schenck, Chem.
(14) E. J. Corey, J. D. Bass, R. LeMahieu, and R. B. Mitra, J. Am. Ber., 100, 3146 (1967).
Chem. Soc., 86, 5570 (1964). (25) R. Hoffman, P. Wells, and H. Morrison, J . Org. Chem., 36, 102
(15) L. M. Stephenson, D. G. Whitten, G. F. Vesley, and 6 . S. Ham- (1971), and references therein.
mond, J. Am. Chem. Soc., 88, 3665 (1966). (26) R. A. Caldwell, J . Am. Chem. Soc., 95, 1690 (1973).
(16) 0. L. Chapman and R. D. Lura, J. Am. Chem. Soc., 92, 6352 (27) R, A. Caldwell and R. P. Gajewski, J. Am. Chem. Soc., 93, 532
(1
~971))
- -_,_
. (1971);3. C. Dalton, P. A. Wriede, and N. J. Turro, ibid., 92, 1318 (1970);
(17) J. Saltiel, J. T. D’Agostino, 0. L. Chapman, and R. D. Lura, J . D, Valentine, Jr., and G. S. Hammond, ibid., 94, 3440 (1972).
Am. Chem. Soc., 93, 2804 (1971). (28) An excellent review including much or all of the early evidence
(18) (a) N. C. Yang, D. Shold, unpublished; quoted in (b) J. Saltiel, is by R. S. Davidson in “Molecular Association”, R. Foster, Ed., Academic
D. E. Townsend, B. D. Watson, P. Shannon, and S. C. Finson, J . Am. Press, London, 1975, p 215.
Chem. SOC.,99, 884 (1977). (29) 8. Farid, J. C. Doty, and J. L. R. Williams, J . Chem. SOC.,Chen.
(19) J. Saltiel and D. E. Townsend, J.Am. Chem. Soc., 95, 6140 (1973). Commun., 711 (1972); see also S. Farid, 8. E. Hartman, J. C. Doty, and
(20) R. M. Bowman, T. R. Chamberlain, C. W. Huang, and J. J. J. L. R. Williams, J . Am. Chem. Soc., 97, 3697 (1975).
McCullough, J. Am. Chem. SOC.,96,692 (1974), and references therein. (30) T. Sugihara, C. Pac, and €3. Sakurai, Chem. Lett., 667 (1972); K.
(21) R. A. Caldwell, J . Am. Chem. Soc., 92, 1439 (1970). Mizuno, C. Pac, and H. Salturai, J. Am. Chem. Soc., 96, 2993 (1974).
Vol. 13, 1980 Exciplex Intermediates 47
Scheme I
P + j F - M
t X

C H3/
P + F ,C
Figure 2. A maximum orbital overlap “sandwich” model for the
SCP-t-An exciplex. 3P f F - 3(P***F) - biradical -
/

\
T

In 1972 and 1973 we formulated the dilemma more 1 P + F

precisely and began to attack it systematically. P + M

+
(1)Are exciplexes present in [2 21 cycloaddition
systems, and if so how do their properties compare to
the properties of the ones observed by the photophy- L C 0 , M e L C 0 2 M e 0
sicists?
(2) If present, are exciplexes in fact on the reaction
coordinate for cycloaddition or are they diversions of C T X
the starting excited species? We returned to the reaction of phenanthrene (P) with
(3) More generally, what is the nature of the first dimethyl fumarate (F) because Farid’s original con-
excited singlet adiabatic hypersurface relevant to [2 +
21 photocycloaddition? tention2*of exciplex emission had been q ~ e s t i o n e d . ~ ~
Repeating our original work, we indeed found the cis-
Resolution of the Dilemma. Exciplex exo (C) and trans (T)cyclobutane adducts, the oxetane
Quenching (X) reported by Farid, and isomerization of F to di-
We had been studying31the photoadditions in benz- methyl maleate (MI. All were quenched by electron
ene of 9-cyanophenanthrene (9CP) to trans-P-methyl- donors, and these observations could be nicely corre-
styrene (PMS), the anethole isomers (t-An and c-An), lated with quenching of the exciplex emission.34i35 Only
and 1-anisyl-2-methylpropene (pBA). Reaction 2 ap- partial quenching of C and T formation was observed

p+ g,,
since a substantial part of cyclobutane formation arises
from intersystem crossing of lP prior to exciplex for-
0 CN ,\\$\\@Me mation. Nevertheless, Scheme I, the complicated ki-
MeP O
R M e hV 0 Me‘“R
(2 I netic scheme, a synthesis of our and Farid’s earlier
0 R=HorMe
studies, was vindicated.
Reactive and Unreactive Exciplexes, The role of
peared to afford only the more hindered endo-phenyl exciplexes was now established in the phenanthrene-
adducts and to give only retention in the olefin even olefin singlet-state cycloadditions. Furthermore, in-
with c-An for which endo addition required two endo creasing numbers of photoaddition reactions were being
substituents. We were gratified to note that all the reported*& in which concomitant excimer or exciplex
reactions except the one with 0-MS exhibited exci- emissions were observed. The original contentions of
plex-like emissions (that reaction, we later32found out, exciplex intermediacy in organic photoreactions which
shows exciplex emission very weakly at -40 “C), and we showed no exciplex emission seemed much more likely
could therefore tackle the question of intermediacy of to be correct. We therefore next turned our attention
the exciplex in the reaction. to the relationship between the “reactive” exciplexes
If a compound could be found which selectively we were observing and the “unreactive” exciplexes
quenched the exciplex, without substantially quenching well-known to photophysicists. The introduction de-
the initial excited state, the question of intermediacy scribes several properties of “unreactive” exciplexes,
could be resolved. This proved to be the case. We and we set out to measure as many of these properties
thought that an exciplex stabilized by CT interactions as we could for our “reactive” exciplexes.
ought to be susceptible to further CT perturbation. We We first o b ~ e r v e d that
~ l ~ emission
~~ from our phen-
initially showed that the phenanthrene-fumaronitrile anthrene-styrene exciplexes showed the same sort of
exciplex could be quenched by 2,3-dimethyl-2-butene,
(33)G. Kaupp, Angew. Chem., Int. Ed. En& 12, 765 (1973).
a donor which did not significantly quench phenan- (34)D. Creed and R. A. Caldwell, J.Am. Chem. SOC.,96,7369(1974).
threne fluorescence. Donors were quite ineffective in (35)D. Creed, R.A. Caldwell, and M. M. Ulrich, J. Am. Chem. SOC.,
the case of 9-cyanophenanthrene-anethole; however, 100,5831 (1978).
(36)N. C. Yang, D. M. Shold, J. K. McVey, J. Am. Chem. SOC.,97,
good electron acceptors quenched the exciplex 5004 (1975).
fluorescence without significantly quenching the (37)N. C. Yang, K. Srinivasachar, B. Kim, and J. Libman, J. Am.
fluorescence of ‘9CP itself. In the 9CP-t-An system, Chem. Soc., 97,5006 (1975).
-. D. M. Shold. and B. Kim. J. Am. Chem. SOC..
(38)N.C. Yann. . 98.6587
,
quenching of photocycloaddition quantitatively paral- (igisj.
leled exciplex fluorescence quenching, rigorously dem- (39)E.A. Chandross and A. H. Scheibel, J. Am. Chem. SOC.,95,611,
onstrating the intermediacy of the exciple^.^^ The 1671 (1973).
(40)F.D. Lewis and R. H. Hirsch, J. Am. Chem. SOC.,98,5914(1976).
product structures suggested that they arose from ste- (41)F. D. Lewis, R. H. Hirsch, P. M. Roach, and D. E. Johnson, J.Am.
reospecific collapse of the maximum overlap, Chem. SOC.,98,8438 (1976).
(42)F. D. Lewis and C. E. Hoyle, J. Am. Chem. SOC.,99,3779(1977).
“sandwich” exciplex of Figure 2. (43)J. J. McCullough, R. C. Miller, D. Fang, and W.-S. Wu, J. Am.
Chem. Soc., 97,5942 (1975).
(31)R.A.Caldwell and L. Smith, J. Am. Chem. SOC., 96,2994(1974). (44)F. C. DeSchryver et al., Pure Appl. Chem., 49,237 (1977).
(32)R.A. Caldwell and T. S. Maw, unpublished results; T. S. Maw, (45)R. A. Caldwell, N. I. Ghali, C.-K. Chien, D. DeMarco, and L.
M.S. Thesis, The University of Texas at Dallas, Dec 1977. Smith, J. Am. Chem. SOC.,100,2857 (1978).
48 Caldwell and Creed Accounts of Chemical Research
solvent dependence as did that from arene-amine ex- In the Iiniit of low reversibility, attainable either by very
ciplexes. The emission increased sharply in wavelength high [A] or by structural or temperature variations, the
as solvent polarity increased and was undetectable in lifetime rC of the exciplex will be given by (4). If k F
T C =z (kF 4- h,,, -I- k R + kD)-'
solvents of higher polarity than acetone. We also
(4
noted46 that, in methanol, the photoreaction of 9-
cyanophenanthrene and trans-anethole afforded only
small amounts of cycloadduct as part of a complex
- 156--15'79-l and k,,, -
P07-108 s-l, as seems to be
usual, it is easy to see that k~ 5 IO6 will lead to an
mixture of electron-transfer-derived products. These apparently "Unreactive" exciplex and kR 2 IO9s-l will
results point to the inherently polar nature of the bi- lead. to so great a dominance of photochemistry over
molecular interaction, as observed for "normal" exci- photophysics that emission will be essentially noncom-
plexes. petitive. For simultaneous emission and reaction to be
Through the use of nanosecond time-resolved spec- observed, the range of hR values is limited. We simply
trofluorimetry and with the collaboration of Professor conclude that it is common for organic bimolecular
L. A. Melton and Dr. P. H. Wine, we were able47to singlet photoreactions to proceed with nanosecond or
analyze the kinetics of formation and decay of several shorten. relaxation times once the exciplex is reached.
phenanthrene-styrene type reactive exciplexes as well The photochemically allowed nature (in the Wood-
as some unreactive model cases. We quickly learned mann sense), at least for [ 2 + 21 and [4 + 41
that exciplex formation was reversible; emission from ions and c y c ~ ~ d i m e r ~ z ~ renders
t i o ~ i s ~this
the phenanthrene persisted long after quenching by the conclusion perhaps unremarkable.
styrene was complete and tracked the exciplex
fluorescence decay. It proved possible to analyze the roeesses on the
results in considerable detail and to obtain equilibrium
constants for exciplex formation, exciplex lifetimes, and +
iplexes in [ 2 21 photocyclo-
(after fluorescence quantum yield measurements) ex- additions and the similarity of reactive and nonreactive
ciplex radiative lifetimes. Rates of formation were es- exciplexes clear, we have attempted more recently to
sentially diffusion controlled. Enthalpies of the exciplex understand the four possible exit channels from the
formation47were roughly in the range predicted by the excited [ 2 + 21 hypersurface via the exciplex, especially
Weller relationship,6beq 1; cf. -7.3 kcal/mol vs. -4.0 of leading to products.
kcal/mol calculated48for 1(9CP.-t-An) and -7.4 kcal/ ~~~~~~e~~The fluorescence and in-
mol vs. -6.3 kcal/mol calculated for l(SCP--pBA). For tersystem crossing processes appear to be rather
the latter, we measured the entropy of association and 5 t r ~ ~ ~ ~ t f oArw ar~~
fluorescence rate of ca. 3 X PO6 s-l
found a normal value of -13 eu. Clearly, the thermo- and an intersystem crossing rate of ca. 3 x IO7s-l are
dynamic behavior of our reactive exciplexes is almost pervasive in this s e r i e ~ . The
~ ~ ?rates
~ ~ are temperature
identical with that of the known, unreactive ones pre- independer~t,~~JO do not depend very much on substit-
viously examined.6b uent or even rather gross reactant variations, are similar

-
Even more importantly, the radiative lifetimes of the
reactive exciplexes (kF 2-4 X 106s-l in general)35-47,49
closely matched those that we and others6-11determined
to literature values7-11for unreactive exciplexes, and
thus have all the hallmarks of "normal" processes.
Little can be predicted about the magnitude of either.
for the unreactive ones. This observation points to the Since one-center integrals probably dominate spin-orbit
similarity of electronic structures in the two series. The coupling, locally excited configurations of the exciplex
conclusion seems inescapable that, in the phenan- (cf. a and b of (5)) may play an important In-
threne-styrene series a t least, and probably in other $ 1 ( ~ a ,= a$l~,*+ b$D,'A 3 CIC/D"A- -t d$D A' (5)
series of reactions in which exciplex emission and
chemical reaction occur together, the first step is for- creased polarity of the exciplex (high c or d, low a and
mation of an exciplex with all the properties expected b ) may inhibit intersystem crossing, and efficient isc in
from study of nonreactive models. one or the other component may enhance it. We note
Why are exciplex emissions still the exception in
organic photoreactions? We certainly must exclude
triplet photoprocesses from our expectations since fluid
that even very short-lived exciplexes can still undergo
intersystem crossing,35eg., '(P-eF), TC 1.2 ns with kk
= 1.2 X lo8 s-l. The lowest singlet of F is n,n*; perhaps
-
solution phosphorescence is seldom observed. Even for El Sayed's rules52are relevant. The 1(9CP-$3MS) ex-
singlet reactions, the following analysis shows that only ciplex, for which 7C is surely <1 ns,32still affords some
special cases will be expected to show simultaneous 30% intersystem crossing at room t e m p e r a t ~ r eand
~~
emission and reaction. Consider eq 3: thus kisc > 3 X IOR In this case, unusually efficient
SO coupling in one of the locally excited configurations
33 + A, 3 ( D * . * A ) , etc seems unlikely. Clearly, further study will be rewarding.
0th cycloadduct for-
mation (kR) and internal conversion (k,) show very
pronounced dependences on exciplex component

product(s)
-
structure and on temperat~re.~~2" Both are activated,
with Ald3 2-7 kcal/mol and with substantial negative
(50) R. A. Caldwell and D. Greed, 9.A m Chern. Soc , 101,6960 (1979).
(46) R. A. Caldwell and L. Smith, unpublished. (51) C. P. Keijzers, J. Duran, and D. €lamer, J Chem. Phys., 69,3563
(47) D.Creed, P. H. Wine, R. A. Caldwell, and L. A. Melton, J . Am. (iwn).
Chem. Soc., 98, 621 (1976). (52) M. A. EISayed, 9.Chem. Phys., 38, 2834 (1963); 36, 573 (1962);
(48) Redox parameters in acetonitrile: S.-M. Park and R. A. Caldwell, 41, 2464 (1964).
J. Electrochem. Soc., 124, 1859 (1977). (53) (a) By taking 2+5t..c for PMS as +5jx for the exciple^;^^^^ (b) R. A.
(49) D.Creed and R. A. Caldwell, J . Am. Chem. Soc., 100,2905 (1978). Caldwell, D. Creed, and T. S. Maw, J . Am. Chem. Soc., 101,1293 (1979).
Vol. 13, 1980 Exciplex Intermediates 49
Table I
Rate Constants and Activation Parameters for Cycloadduct Formation and Internal Conversion of ExciplexeP
AH,+, AS,+, AH, + , ASD +,
exciplexb kRC,s-’ kcalimol eu kDC, S-l kcal /mol eu
’( 3-Me0-10-CP. . .t-An) 2.8 x lo8 2.7 -11 2.1 x l o * 2.1 -13
‘(9CP. . .t-An) 4.2 x io7 2.2 -16 3.2 x 107 3.0 -14
I ( 3,10-CN2P. . $-An) 4.6 X lo6 4.2 -14 4 . 4 x 106 2.2 -20
‘(9CP. . ,pBA) 5.6 X 10’ 4.8 -16
‘(P. . ‘ F )-+ T 9 x 10’ 7.3 -7
7 x lo8 2.2 -1 1
1(P. . .Fj -+ x 2 . 2 x 10’ 4.4 -10
a From ref 49 and 50. M e 0 = methoxy, CN = cyano, P = phenanthrene; other symbols explained in text. 21-23 “C.

entropies of activation, typically -10 to -17 eu. Table

I summarizes selected rate constants and activation
parameters. They are quite revealing about the nature
of both processes.
Most striking is the general similarity of the two in
the substituted cyanophenanthrene-styrene series. The s’ll + 1;
magnitudes of the rate constants kR and k~ vary by as Loxcipiex
much as 500-fold, yet their ratio is nearly constant! The
entropies of activation correspond to the conversion of
a rather loose exciplex (for which have evi-
dence) to a rather tight transition state, quite in line
with other values for unimolecular thermal reorgani-
zations. For these reasons, we feel that both k R and kD
are best represented as adiabatic, at least through a
transition state immediately following the exciplex. A
surface-hopping (diabatic) process could only fortui- ~ ~ )[2 + 21
Figure 3. Schematic energy surface (after M i ~ h 1 for
tously give a sensible AS*. cycloadditions.
Variation with substituent similarly can be reasonably process. Second, exciplex formation may bias the sys-
interpreted by familiar ground-state polar and steric tem toward cycloaddition and away from competing
substituent effects and judicious consideration of the chemical processes. In the present case, for example,
energy surface suggested by M i ~ h for 1 ~[2~ + 21 photo- dimerization of the phenanthrene can be a serious
cycloaddition. The exciplex is a polar “pause point” on competing process when the phenanthrene-olefin ex-
the way to products, and must pass over a transition ciplex is relatively unstable, Le., highly reversible.
state a t the crossing of the polar S surface and the The Pericyclic Minimum. Another Entry. The
relatively nonpolar D (doubly excited) surface of Figure next frontier for those interested in photocycloadditions
3. Since D is nonpolar, the transition state is necessarily is clearly the pericyclic minimum in SI of Figure 3.
less polar than the exciplex, and the decrease in rate Michl’s proposed surfaces59 depict this as a second
of cycloadduct formation with increasing ability of the “pause point” and the one from which diabatic con-
exciplex to stabilize charge separation (Table I) thus version to the ground surface results. The nearly 50:50
also necessarily follows. A study of solvent effects by partitioning between k~ and kD we observe fits very well
Lewis for other exciplexes has afforded the same con- with the idea of a vertical drop to a flat top of the
clusion.60 The increased steric hindrance in the ground surface in the vicinity of the pericyclic mini-
l(9CP-pBA) case due to the endo-methyl must also be mum. However, there is no independent experimental
a retarding effect on kR. evidence for a pericyclic minimum of finite lifetime in
Since exciplexes appear now to be common (perhaps S1at present. We have attempted to obtain evidence,
universal) photocycloaddition intermediates, the de- if not for the nature and lifetime of the species a t the
crease in reaction rate constant with increase in exciplex pericyclic minimum, at least for common intermediates
stability is an apparent paradox which requires more in cycloadditions and cycloreversions consistent with
careful consideration of the importance of the exciplex such species.
to the overall reaction. We feel that exciplexes can exert Our work with the 1(9CP.-+An) and l(P.-F) exci-
a favorable effect on cycloaddition reactions in two plexes, especially determination of k R and kD, provided
ways. Most important is the prior association of the an understanding of the forward process. We noted
reacting chromophores with consequent overcoming of that both the ‘(SCP-.t-An) adduct and T, the trans
a substantial fraction of the AS* of the bimolecular adduct from l(P-.F), afforded efficient and nearly ste-
reospecific fragmentation upon irradiation. Only weak
(54) G. N. Taylor, E. A. Chandross, and A. H. Scheibel,J. Am. Chem. fluorescence ascribable to the adduct was observed for
Soc., 96, 2693 (1974).
(55) E. A. Chandross and H. T. Thomas, Chem. Phys. Lett., 9, 393 the latter, and none for the former. Consequently, we
~ - --,.
11971).
. suspected that, whatever the next “pause point” in
(56) T. Okada, T. Fujita, M. Kubota, S. Masaki, N. Mataga, R. Ide, cycloreversions was, it was reached essentially quanti-
Y. Sakata, and S.Misumi, Chem. Phys. Lett., 14, 563 (1972).
(57) R. S. Davidson, J. Chem. SOC.Chem. Commun., 827 (1976). tatively. A test of whether the pause points were t,he
(58) J. J. McCullough, private communication. same in both directions was then immediately apparent,
(59) W. Gerhartz, R. D. Poshusta, and J. Michl, J. Am. Chem. SOC.,
98, 6427 (1976); J. Michl, Photochem. Photobiol., 25, 141 (1977). since the fraction kR/(kR + kD) of exciplex leading to
(60) F. D. Lewis and C. E. Hoyle, J. Am. Chem. SOC.,99,3779 (1977). adduct and the fraction kD/(kR + kD) of adduct frag-
50 Caldwell and Creed Accounts of Chemical Research
Table I1 outcome in untried cases. High singlet energy in the
Comparison of A d d u c t P h o t o f r a g m e n t a t i o n Q u a n t u m excited reactant, low triplet energies for both, and high
Yields w i t h Photocycloaddition Q u a n t u m Yields
frontier orbital densities appear to presage high re-
R
'
___ activity. Tests of these predictions are now under way,
adduct @addend? '
R -t 'D sum The second critical feature, partitioning at the pericyclic
minimum, seems less amenable to simple prediction, yet
may also significantly affect quantum yields of photo-
0.38,c 0.3gd 0.57 0.96 reaction.
The studies we have made of triplet-state photo-
cycloadditions22126s53b suffer from our inability to observe
triplet reactive exciplexes by emission. Our work, and
q.,#02cH3 0.98,c l . O O d 1.3 X 0.99
that of others generally, is suggestive but we believe not
definitive on the points we have here touched on for
singlets. While we suspect that there will be, mutatis
mutandis, a general similarity to singlet exciplexe~,"~
a Fkagmentation a t 295 nm. * F r o m Table I, f o r
confirmation must await observation in key examples.
a d d u c t f o r m a t i o n f r o m corresponding a d d e n d s . Phenan- Our work on exciplex quenching has convinced us
threne. Olefin. that an enormous body of termolecular photophysics
and photochemistry lies fallow. Stable, emissive excited
menting to addends could now be compared. They termolecular complexes, which we have named exter-
would be required to sum to unity if a common inter- plexes, have been detected by other^,^^^^^^^ generally
mediate existed and (as we knew) no other chemistry under either forcing conditions of high concentration,%
occurred. low temperature,1° or surpassing what must be a sub-
The results61 (Table 11) certainly support the idea of stantial termolecular entropy problem via incorporation
a common intermediate, which we believe (but cannot of two of the chromophores in a polymer chain67or via
yet prove) is the species a t the pericyclic minimum. their linking by alkyl chains.% Chemical reactions now
O t h e r P have noted similar outcomes in dimerization ascribable to termolecular processes include enhanced
of compounds related to anthracene (cf. especially the cycloaddition^,^^^^^ electron-transfer reactions,3Rand in
work of Jones62c),and indeed the [4 9 41 surface is some cases the formation of adducts that do not form
expected to behave in similar fashion. We cannot yet via bimolecular route^.^^,^^ While the separation of
rule out the possibility of separate one-bond processes charge in chloroplasts during photosynthesis presuma-
leading to the same singlet 1,4-biradical. bly follows a different mechanism in detail,70the ne-
Active pursuit of the pericyclic minimum and an cessity for polymolecularity is clear and is obviously at
understanding of the dynamics of systems passing least indirectly relevant to our termolecular CT
through it are now major challenges for experimentalists q ~ e n c h i n g . Future
~~ work may reveal a significant and
interested in excited-state cycloaddition hypersurfaces. fascinating set of useful and practical ramifications of
termolecular processes.
Challenges for the Future
T h i s work could not have been done without the efforts of the
The role of exciplexes in [2 + 21 photocycloaddition co-workers listed in the references, efforts we greatly appreciate.
systems is now established beyond doubt, and exciplex W e are grateful f o r financial support by t h e Robert A . W e l c h
quenching is generally well understood. However, de- Foundation (Grant AT-532) and t h e National Science Foun-
tailed understanding of factors leading to exciplex or dation. While our colleagues who have provided stimulation and
excimer collapse to products has hardly emerged. How encouragement are too numerous to list, Professor F. D. Lewis
to predict heights of energy barriers and determine the and Dr. C. E. Hoyle deserve our t h a n k s f o r a number of stimu-
properties of surfaces near critical points (maxima and lating discussions and f o r t h e sharing o f results prior t o p u b -
minima) are the real keys to the prediction of success lication.
or failure of any chemical reaction, including photo- (64) (a) H. Beens and A. Weller, Chem. Phys. Lett., 2, 140 (1968); (b)
cycloadditions. E. A. Chandross and H. T. Thomas, ibid., 9, 397 (1971); (c) G. S. Bed-
dard, s. E. Carlin, and C. Lewis, J. Chem. SOC.,Faraday Trans. 2, 71,
We have recentlf13 used the Michl model to correlate 1894 (1975).
known photocycloaddition and photodimerization re- (65) J. Saltiel, D. E. Townsend, B. D. Watson, and P. Shannon, J. Am.
activities. If the D-S energy gap a t infinite reactant Chem. Soc., 97, 5688 (1975).
(66) 1'. Mimura and M. Itoh, J. Am. Chem. SOC.,98,1095 (1976); see
separation and the behavior of each as the reactants also T. Mimura and M. Itoh, Bull. Chem. SOC.Jpn., 50,1739 (1977), and
approach are both known, the D-S crossing and thus T. Okamoto, M. Itoh, T. Ohta, and T. Oramto, ibid., 50,1665 (1977); R.
the cycloaddition transition state will be located. It A. Beechcroft, R. S. Davidson, and T. D. Whelan, J . Chem. SOC.,Chem.
Commun., 911 (1978).
proves possible to make reasonable zeroth-order esti- (67) C. E. Hoyle and J. E. Guillet, Mucromolecules, 11, 221 (1978).
mates of reactivity easily, and thus also to predict the (68) R. 0. Campbell and R. S. H. Liu, Chem. Commun., 1191 (1970).
(69) F. D. Lewis, Acc. Chem. Res., 12, 152 (1979); see also F. D. Lewis
and D. E. Johnson, J. Am. Chem. SOC.,100, 983 (1978).
(61) R. A. Caldwell and N. Giao, unpublished. (70) For leading references to the possible role of chlorophyll aggre-
(62) (a) G. Kaupp, Angew. Chem., Int. Ed. Engl., 11, 313 (1972) (an- gates in photosynthesis, see S. G. Boxer and R. R. Bucks, J . Am. Chem.
thracenophane); (b) W. J. Tomlinson, E. A. Chandross, R. L. Fork, C. SOC.,101, 1883 (1979); A. Warshel, ibid., 101, 744 (1979); N. Periasamy
A. Pryde, and A. A. Lamola, Appl. Opt., 11,533 (1972) (acridizinium ion); and H. Linschitz, ibid., 101, 1056 (1979); F. K. Fong, A. J. Hoff, and F.
(c) W. R. Bergmark, G. Jones, T. Reinhardt, and A. M. Halpern, J. Am. A. Brinkman, ibid., 100, 619 (1978).
Chem. Soc., 100, 6665 (1978). (71) R. A. Caldwell, D. Creed, D. C. DeMarco, L. A. Melton, H. Ohta,
(63) R. A. Caldwell, J . Am. Chem. SOC.,submitted for publication. and P. €1. Wine, J. Am. Chem. SOC.,in press.