Temperature dependence of the photoisomerization of cis-1-(2-antrhyl)-2-phenylethene. Conformer-specificity, torsional energetics and mechanism

Article abstract of DOI:10.1021/ja972293a

Emission from cis-1-(2-anthryl)-2-phenylethene, c-APE*, in toluene is resolved into ¹t-APE(B)* and ¹c-APE* components at temperature ranging between 4.3 and 59.3°C. Decomposition of effective fluorescence quantum yields, φ(fc), into

Full text of DOI:10.1021/ja972293a

11202
J. Am. Chem. Soc. 1997, 119, 11202-11210
Temperature Dependence of the Photoisomerization of
cis-1-(2-Anthryl)-2-phenylethene. Conformer-Specificity,
Torsional Energetics and Mechanism
Jack Saltiel,* Yuxin Zhang, and Donald F. Sears, Jr.
Contribution from the Department of Chemistry, The Florida State UniVersity,
Tallahassee, Florida 32306-3006
ReceiVed July 10, 1997X
1
1
Abstract: Emission from cis-1-(2-anthryl)-2-phenylethene, c-APE*, in toluene is resolved into t-APEB* and c-
APE* components at temperatures ranging between 4.3 and 59.3 °C. Decomposition of effective fluorescence quantum
yields, φhfc, into pure component fluorescence quantum yields, φft-B and φfc, shows that φft-B increases 24% with
increasing temperature while φfc decreases more than 3-fold over this temperature range. On the basis of the fraction
1
of molecules that escape the c-APE* potential energy minimum, 1 - φfc, the efficiency of adiabatic formation of
1t-APEB* remains remarkably temperature independent at 50.5 ( 0.7%. These results, together with photoisomerization
quantum yields as a function of [c-APE] in degassed and air-saturated toluene, reveal a detailed photoisomerization
mechanism. At infinite dilution and in the absence of molecular oxygen, photoisomerization of c-APE occurs
predominantly via the adiabatic, conformer-specific 1c-APEB* f 1t-APEB* pathway. This torsional motion experiences
a 4.44 ( 0.14 kcal/mol barrier probably located at the perpendicular, 3p*, geometry. Since 12% of 1t-APEB* intersystem
3
cross to t-APEB*, the known triplet state quantum chain process enhances photoisomerization quantum yields at
1
higher [c-APE]. Triplets formed directly from c-APE* also contribute to this pathway. In air-saturated solutions,
3
oxygen eliminates the quantum chain process by reducing the lifetime of t-APE*. However, the quenching of
3
3
3
1c-APE* by O2 gives c-APE*, thus enhancing photoisomerization quantum yields via rapid c-APE* f t-APE*
adiabatic torsional displacement. No photoisomerization of 1c-APEA* need be postulated to account for our
observations. The enthalpy difference between ground state conformers, ∆HAB, favors c-APEB by 0.92 ( 0.02
kcal/mol.
Introduction
way photoisomerization of arylalkenes or of 1,2-diarylethylenes
with a 2-anthryl substituent on the double bond, as in the title
Olson proposed the first cis-trans photoisomerization mech-
anism of olefins in terms of potential energy curves.1 He
formulated the reaction as an adiabatic process in the excited
state.1a
3
3
compound. The c* f t* process is followed by excitation
transfer from 3t* to a cis ground state molecule and a quantum
chain process ensues.5 This mechanism applies when the triplet
energy of the aryl substituent of the arylalkene or of the
diarylethylene is low relative to the energy of the corresponding
perpendicular olefin triplet, 3p*. Extensive localization of triplet
excitation in the larger aryl substituent ensures that nearly planar
transoid and cisoid olefin geometries correspond to energy
minima in the triplet potential energy surface. The higher
molecule + hν f molecule′ + hν′
(1)
Mulliken’s expansion2 of Hu¨ckel’s description of the double
bond,3 published nearly simultaneously with Olson’s proposal,
provided a quantum mechanical basis for the idea of barrierless
rotation in the excited state. Lewis et al. reasoned that, if
Olson’s mechanism were correct, cis and trans olefins should
exhibit identical fluorescence.4 In a pioneering study of stilbene
photoisomerization they searched in vain for cis-stilbene
fluorescence. By establishing that the fluorescence intensity
from cis-stilbene, if present, was less that 1% of the fluorescence
of trans-stilbene, they showed that the contribution of an
adiabatic pathway to stilbene photoisomerization is, at best,
minor.4
3
3
energy of c*, due to steric hindrance, accounts for the c* f
3t* directionality of the process.
Adiabatic cis f trans photoisomerization in the lowest excited
singlet state was reported four years later by Sandros and Becker
for the 9-styrylanthracenes.6 The emission of solutions of the
cis isomer was shown to contain substantial contributions of
fluorescence from the trans isomer. Other examples followed
soon thereafter.7-9 Our work has shown that, as the aryl
substituent in cis-ArCH ) CHPh derivatives is changed from
(5) (a) Arai, T.; Karatsu, T.; Sakuragi, H.; Tokumaru, K. Tetrahedron
Lett. 1983, 24, 2873-2876. (b) Karatsu, T.; Arai, T.; Sakuragi, H.;
Tokumaru, K. Chem. Phys. Lett. 1985, 115, 9-15. (c) Arai, T.; Karatsu,
T.; Misawa, H.; Kuriyama, Y.; Okamoto, H.; Hiresaki, T.; Furuuchi, H.;
Zeng, H.; Sakuragi, H.; Tokumaru, K. Pure Appl. Chem. 1988, 60, 989-
998. (d) Tokumaru, K.; Arai, T. J. Photochem. Photobiol. A 1992, 65, 1-13.
(e) Arai, T.; Tokumaru, K. Chem. ReV. 1993, 93, 23-39. (f) Tokumaru,
K.; Arai, T. Bull. Chem. Soc. Jpn. 1995, 68, 1065-1087.
(6) (a) Sandros, K.; Becker, H.-D. J. Photochem. 1987, 39, 301-315.
(b) Go¨rner, H. J. Photochem. Photobiol. A 1988, 43, 263-288. (c) Sandros,
K.; Becker, H.-D. J. Photochem. Photobiol. A 1988, 43, 291-292.
(7) (a) Sandros, K.; Sundahl, M.; Wennerstro¨m, O.; Norinder, U. J. Am.
Chem. Soc. 1990, 112, 3082-3086. (b) Sandros, K.; Sundahl, M.;
Wennerstro¨m J. Phys. Chem. 1993, 97, 5291-5294. (c) Sandros, K.;
Sundahl, M. J. Phys. Chem. 1994, 98, 5705-5708.
Olson’s mechanism lay dormant for over 40 years. It was
resurrected in a series of papers starting in 1983 by Tokumaru
and co-workers.5 Adiabatic cis f trans photoisomerization in
the lowest triplet state was shown to be the key step in the one-
X Abstract published in AdVance ACS Abstracts, November 1, 1997.
(1) (a) Olson, A. R. Trans. Faraday Soc. 1931, 27, 69-76. (b) Olson,
A. R.; Hudson, F. L. J. Am. Chem. Soc. 1933, 55, 1410-1424. (c) Olson
A. R.; Maroney, W. J. Am. Chem. Soc. 1934, 56, 1320-1322.
(2) Mulliken, R. S. Phys. ReV. 1932, 41, 751.
(3) Hu¨ckel, E. Z. Phys. 1930, 60, 423.
(4) Lewis, G. N.; Magel, T. T.; Lipkin, D. J. Am. Chem. Soc. 1940, 62,
2973-2980.
S0002-7863(97)02293-2 CCC: $14.00 © 1997 American Chemical Society

Photoisomerization of cis-1-(2-Anthryl)-2-phenylethene
J. Am. Chem. Soc., Vol. 119, No. 46, 1997 11203
Scheme 1
Photoisomerization Quantum Yields. A super-high-pressure
mercury lamp (Bausch & Lomb SP-200) was used as the excitation
source. The excitation wavelength, λexc ) 366 nm, was selected with
the use of a high-intensity grating mononochromator (Bausch & Lomb,
1200 grooves/mm). Incident light was beamed through an entrance
slit into a dewar that contained a miniature aluminum merry-go-round
apparatus in which five degassed Pyrex ampules (13 mm o.d.)
containing different c-APE concentrations in toluene and a sixth ampule
containing an actinometer solution were irradiated simultaneously.
Samples were degassed by using 5 freeze-pump-thaw cycles to ∼3
× 10-5 Torr and sealed at a constriction. Rotation of the merry-go-
round was achieved by use of a GT21 motor (Gerald K. Heller Co.).
The merry-go-round and samples were immersed in a water bath whose
temperature was maintained at 19.3 ( 0.1 °C. Temperature was
monitored by use of an Omega Engineering Model 199P2 RTD digital
thermometer equipped with a HYP-4 RTD hypodermic probe. Irradia-
tion periods were measured with a digital counter (Lab-CHRon).
Different irradiation times were achieved by temporarily blocking the
incident light beam.
Gas Chromatography. Actinometer solutions containing fluo-
renone and the stilbenes in benzene were analyzed by GLC (Varian
3300 gas chromatograph, equipped with a J&W Scientific DX-4
capillary column). The column temperature was programmed to
increase from 160 to 180 °C at a rate of 1 °C/min. Injector and detector
temperatures were maintained at 250 °C. The carrier gas (He) pressure
was 12 psi. Integration of peak areas was achieved by use of a Varian
4290 integrator.
phenyl (Ph) to 2-naphthyl (2-Np) and then to 2-anthryl (2-An),
the efficiency of the adiabatic 1c* f 1t* pathway increases from
0.2% to >2% and to >44%, respectively.8 The greater
localization of the electronic excitation in the largest aryl group
is reflected in a 1000-fold increase in the fluorescence lifetime
of cis-1-(2-anthryl)-2-phenylethene (c-APE) relative to cis-1-
(2-naphthyl)-2-phenylethene (c-NPE).8d,e Moreover, the adia-
batic photoisomerization pathway in these two molecules is
highly conformer specific. As illustrated in Scheme 1 for
c-APE, in each case, only the more extended conformers,
c-NPEB and c-APEB, exhibit the reaction.
In this paper we report the temperature dependence of the
1
1
adiabatic c-APEB* f t-APEB* process and the dependence
of the photoisomerization quantum yield on [c-APE] under
degassed and air-saturated conditions. These results provide
insights concerning the energetics of 1c* f 1t* torsional motion
in c-APEB and allow quantitative evaluation of the relative
contributions of singlet and triplet excited states in cis-trans
photoisomerization.
Experimental Section
Materials. Toluene, c-APE, and quinine sulfate were previously
described.8e Fumaronitrile (Aldrich) was recrystallized from benzene
prior to use. Fluorenone (Aldrich, 98%) was recrystallized from
ethanol, chromatographed on alumina with benzene as eluent, and
recrystallized twice from 95% ethanol (99.98% purity, GLC). trans-
Stilbene (Aldrich, 96%) was recrystallized from ethanol, chromato-
graphed on alumina with n-hexane as eluent, and sublimed under
reduced pressure (99.99% purity, GLC). Benzene (Aldrich, Spectral
grade) was passed through alumina prior to use.
Data Analysis. Principal component analysis (PCA) calculations
were performed on a Dell 80486/87 (25 MHz) microcomputer.
Results
Spectral Sets. Fluorescence spectra of c-APE in toluene
were measured in a flow-cell system as a function of λexc and
fumaronitrile, FN, concentration. Ar-saturated solutions were
employed and measurements were recorded at 4.3, 39.3, and
59.3 °C. At each temperature toluene background spectra were
recorded and then c-APE solution was added to a total volume
of 200 mL. The concentration of FN was varied by successive
additions of portions of a FN stock solution in toluene containing
the identical c-APE concentration present in the circulating
reservoir. Fresh solutions were employed for each temperature,
and actual working concentrations of c-APE and FN were
determined from UV absorption spectra by comparison with
standard solutions.11 Concentrations, corrected for solvent
density variation with temperature, were as follows: at 4.3 °C,
[c-APE] ) 2.89 × 10-5 M, [FN] ) 0.00, 1.43 × 10-3, 3.84 ×
10-3, 7.91 × 10-3, and 1.63 × 10-2 M; at 39.3 °C, [c-APE] )
Absorption Spectra. Absorption spectra were measured with a
Perkin-Elmer Lambda-5 spectrophotometer interfaced with a 80486 DX
(66 MHz) microcomputer.
Fluorescence Spectra. Fluorescence spectra were measured with
an extensively modified Hitachi/Perkin-Elmer MPF-2A spectropho-
tometer as previously described.10 All measurements were done under
flow cell conditions with Ar-outgassed solutions.8
(8) (a) Saltiel, J.; Waller, A.; Sun, Y.-P.; Sears, D. F., Jr. J. Am. Chem.
Soc. 1990, 112, 4580-4581. (b) Saltiel, J.; Waller, A. S.; Sears, D. F., Jr.
J. Photochem. Photobiol. A 1992, 65, 29-40. (c) Saltiel, J.; Waller, A. S.;
Sears, D. F., Jr. J. Am. Chem. Soc. 1993, 115, 2453-2465. (d) Saltiel, J.;
Tarkalanov, N.; Sears, D. F., Jr. J. Am. Chem. Soc. 1995, 117, 5586-5587.
(e) Saltiel, J.; Zhang, Y.; Sears, D. F., Jr. J. Am. Chem. Soc. 1996, 118,
2811-2817.
(9) (a) Spalletti, A.; Bartocci, G.; Mazzucato, U.; Galiazzo, G. Chem.
Phys. Lett. 1991, 186, 297-302. (b) Mazzucato, U.; Spalletti, A.; Bartocci,
G. Coord. Chem. ReV. 1993, 125, 251-260. (c) Kikuchi, Y.; Okamoto, H.;
Arai, T.; Tokumaru, K. Chem. Phys. Lett. 1994, 229, 564-570.
(10) Saltiel, J.; Sears, D. F., Jr.; Choi, J.-O.; Sun, Y.-P.; Eaker, D. W. J.
Phys. Chem. 1994, 98, 35-46, 8260.
(11) Saltiel, J.; Zhang, Y.; Sears, D. F., Jr. J. Phys. Chem. 1997, 101A,
7053-7060.

11204 J. Am. Chem. Soc., Vol. 119, No. 46, 1997
Saltiel et al.
φf0
) (1 + KSVc-B[FN])(1 + KSVt-B[FN])
(2)
( )
φf
t-B
c-B
t-B
1
1
where KSV
and KSV
are the SV constants for c- and t-
APEB*, respectively. Except for the substitution of quencher,
eq 2 is identical with the equation employed for O2 quenching.8e
Proceeding as in the oxygen case, we assign the just determined
c
1
KSV ) 62 M-1 value to c-APEB* and use the known value11
of KSVt-B ) 266 ( 6 M-1 to define the expected FN quenching
of 1t-APEB fluorescence based on eq 2. Location of the
combination coefficients of c-APE on the normalization line is
then based on reproducing the expected FN quenching curve,
Figure 3. The resulting resolved c-APE fluorescence spectrum
is shown in Figure 4 together with the spectra obtained at the
other temperatures by following the same procedure. Stern-
Volmer constants employed in the analyses and obtained for
c-APE are shown in Table 1.
Figure 1. Eigenvectors of the c-APE/FN fluorescence matrix at 4.3
°C. The four largest eigenvalues are 0.137, 0.196 × 10-3, 0.233 ×
10-4, and 0.912 × 10-5
.
Fluorescence Quantum Yields. Fluorescence quantum
yields for c-APE were determined under flow cell conditions
with use of quinine bisulfate in 0.1 N H2SO4 as the fluorescence
standard, φf ) 0.546.12 Ar-saturated toluene solutions of c-APE,
thermostated at the desired temperature, were employed and
absorbances of standard and unknown were matched closely at
λexc
. All spectra were corrected for background and for
nonlinearity in instrumental response. Fluorescence quantum
yields, φhfc, corrected for differences in index of refraction, are
shown in Table 2. Bars over symbols indicate that measured
quantities reflect composite behavior of a mixture of species.
Estimated uncertainties in the quantum yields are (3%.
Photoisomerization Quantum Yields. The fluorenone-
sensitized photoisomerization of trans-stilbene in benzene was
used for actinometry, φtc ) 0.47,13,14 as previously described.15
Conversions in actinometer solutions were corrected for back
reaction.16 Concentrations used for actinometry were [t-S] )
5.06 × 10-3 M and [Fl] ) 2.87 × 10-2 M. Corrected
actinometer conversions after 40.0 min of irradiation were 3.34
( 0.03 and 3.65 ( 0.02% cis-stilbene for degassed and air-
saturated cis-APE samples, respectively. Degassed and air-
saturated c-APE solutions were irradiated for 2.0 and 8.0 min,
respectively. Following irradiation, each c-APE solution was
diluted and its UV absorption spectrum was recorded. Isomer
compositions were determined on the basis of the entire
spectrum by performing PCA on a matrix containing the
absorption spectra of all irradiated solutions together with the
spectra of pure c- and t-APE. Since the formation of t-APE is
essentially irreversible, no back reaction correction is required.
Observed conversions and corresponding photoisomerization
quantum yields, φhct, are given in Table 3.
Figure 2. Normalization line based on the eigenvectors in Figure 1.
The points are combination coefficients for experimental spectra and
for pure component spectra (see arrows).
1.68 × 10-5 M, [FN] ) 0.00, 1.93 × 10-3, 4.04 × 10-3, 5.90
× 10-3, and 8.13 × 10-3 M; at 59.3 °C, [c-APE] ) 2.92 ×
10-5 M, [FN] ) 0.00, 1.47 × 10-3, 4.54 × 10-3, 7.82 × 10-3
,
and 1.08 × 10-2 M. Emission spectra were recorded in the
360-580-nm range in 0.5-nm increments with λexc ) 350, 360,
380, 390, and 400 nm. A single spectrum was measured for
each λexc and each [FN] yielding a (25 × 221) spectral matrix
at each temperature. Comparison of absorption spectra before
and after the fluorescence measurements showed no discernible
changes in [c-APE].
Spectral Resolutions. Fluorescence spectra were corrected
for background and self-absorption as previously described.8e
PCA was applied separately to the 25 spectra at each temper-
ature. Eigenvalues and eigenvectors from each matrix are
consistent with two-component systems, as illustrated in Figure
1 for the 4.3 °C matrix. Spectral resolutions were achieved
following the procedure described for O2 quenching at 19.3 °C.8e
The process relies on the previously determined t-APEB spectra
and Stern-Volmer (SV) constants for t-APEB fluorescence by
FN at each temperature.11 The analysis at 4.3 °C is typical and
is presented below in some detail. The combination coefficients
of the known t-APEB spectrum are defined by target analysis
as dot products of this spectrum with the eigenvectors in Figure
1. They define a point that falls on the normalization line and
corresponds to KSVc ) 62 ( 2 M-1 as the SV constant for FN
quenching of c-APE fluorescence, Figure 2. On the basis of
Discussion
PCA Resolution of c-APE Fluorescence. As described in
the Results section, our resolution of c-APE fluorescence spectra
into 1c-APE* and 1t-APE* components at different temperatures
follows faithfully the procedure employed earlier for measure-
ments at 19.3 °C.8e Substitution of FN for O2 as the quencher
avoids possible uncertainties in O2 concentrations at different
temperatures under our flow cell conditions. Earlier reports had
shown that FN is a highly efficient quencher of the fluorescence
(12) Meech, S. R.; Phillips, D. J. Photochem. 1983, 23, 193-217.
(13) Caldwell, R. A.; Gajewski, R. P. J. Am. Chem. Soc. 1971, 93, 532.
(14) Valentine, D., Jr.; Hammond, G. S. J. Am. Chem. Soc. 1972, 94,
3449.
(15) Saltiel, F.; Dabestani, R.; Schanze, K. S.; Trojan, D.; Townsend,
D. E.; Goedken, V. L. J. Am. Chem. Soc. 1986, 108, 2674-2687.
(16) Lamola, A. A.; Hammond, G. S. J. Chem. Phys. 1965, 43, 2129.
1
Scheme 1, the dependence of the fluorescence intensity of t-
APEB* on [FN] should follow the quadratic relationship

Photoisomerization of cis-1-(2-Anthryl)-2-phenylethene
J. Am. Chem. Soc., Vol. 119, No. 46, 1997 11205
Table 3. c-APE to t-APE Photoisomerization Quantum Yieldsa
(Toluene, 19.3 °C)
degassed
104[c-APE], M xtb
air-saturated
104[c-APE] xtb
c
φhct
φhct
1.06
1.50
2.00
2.54
2.77
0.174 1.02(4)
0.157 1.29(4)
0.143 1.59(4)
0.132 1.86(3)
0.130 1.99(3)
1.05
1.45
1.99
2.25
2.67
0.341 0.55(2)
0.282 0.62(2)
0.200 0.57(2)
0.182 0.58(2)
0.151 0.56(2)
a Values in parentheses are estimated uncertainties in the last
significant figure. They are based primarily on the uncertainty of the
GLC analyses of actinometer solutions. b Observed fractional conver-
sions to t-APE. c Photoisomerization quantum yields are corrected for
light filtering by the product t-APEB by dividing effective quantum
yields by (1 - 0.5xt), see text.
Figure 3. Global Stern-Volmer plot for FN quenching of t-APEB
fluorescence at 4.3 °C. The curve is based on eq 2 with KSVc-B ) 62
The resolved c-APE fluorescence spectra in Figure 4 are
arbitrarily scaled to facilitate comparison. They are similar in
shape, showing little, if any, vibronic structure. The blue shift,
clearly evident at the onset of the 59.3 °C spectrum, may reflect
the temperature-induced decrease in medium polarizability as
discussed for the resolved conformer spectra of t-APE.11 By
analogy with t-APE,19 we expect the c-APEA UV absorption
spectrum to be somewhat red-shifted relative to the absorption
spectrum of c-APEB. This accounts for the sharp drop of φhfc
when excitation is at the onset of the absorption spectrum of
c-APE, from which an upper limit of φfc-A e 0.025 at 19.3 °C
was estimated.8e The effective fluorescence quantum yield of
the c-APE component is larger by a factor of g5 than this upper
limit in all spectra at 19.3 °C to which PCA was applied.8e
Accordingly, the resolved c-APE fluorescence spectrum consists
mainly of 1c-APEB* emission, at least for that temperature. The
M-1 and KSVt-B ) 266 M-1
.
1
preponderance of c-APEB* fluorescence in the c-APE com-
ponent at the other temperatures is indicated by the relative
insensitivity to λexc of the location of the combination coef-
ficients of the spectra on the normalization lines. The behavior
at 4.3 °C, illustrated in Figure 2, where points for spectra for
each [FN] appear as tight clusters on the normalization line, is
typical.
Figure 4. Arbitrarily scaled resolved c-APE fluorescence spectra
obtained at 59.3, 39.3, 19.3, and 4.3 °C in the order of decreasing
spectral area.
Table 1. Conformer-Specific SV Constants for FN Quenching of
APE Fluorescence in Toluenea
FN
SV
K
, M-1
1
SV Constants and the Lifetime of c-APEB*. Since the
b
b
T, °C
t-APEA
t-APEB
c-APEB
fluorescence of the c-APE component is associated mainly with
the B conformer, the SV constants in Table 1 for c-APE likewise
4.3
19.3
39.3
59.3
80.3(3.1)
116.0(6.0)
155.2(5.7)
217.4(12.8)
265.9(5.5)
347.8(3.3)
460.5(8.4)
547.5(21.5)
62(2)
57(4)c
50(3)
43(6)
1
reflect mainly the quenching of c-APEB* and are so labeled.
They decrease with increasing temperature, in sharp contrast
with the SV constants of t-APEA and t-APEB that change
strongly in the opposite direction. If FN quenching of c-APE
fluorescence is diffusion controlled, as has been shown for the
t-APE conformers,11 then the difference between the temperature
dependencies of the SV constants reveals the difference between
the temperature dependencies of the fluorescence lifetimes of
the t-APE conformers and of c-APEB. The strong increases in
the quenching rate constants kFqtN-A, kFN , and kFN with
a Values in parentheses are estimated uncertainties. b Reference 11.
c Estimated, based on SV constants for FN and O2 quenching of t-APEB
FN
and O2 quenching of c-APEB: KFN
) K
(KOX /KOX ).
SVc-B
SVt-B SVc-B SVt-B
Table 2. The Temperature Dependence of Effective Fluorescence
Quantum Yields from c-APE Solutionsa
xt-B(λexc
350 nm
)
φhfc(λexc)
qt-B
qc-B
T, °C
390 nm
350 nm
390 nm
FN
SV
increasing T are evident in the K
values of t-APEA and
4.3
19.3
39.3
59.3
0.585
0.718
0.805
0.853
0.580
0.723
0.808
0.833
0.569
0.531
0.505
0.484
0.488
0.456
0.423
0.397
t-APEB, whose fluorescence lifetimes are remarkably insensitive
FN
to temperature changes,20 but are countered in K
by an
SVc-B
(17) (a) Green, B. S.; Rejto¨, M.; Johnson, D. E.; Hoyle, C. E.; Simpson,
J. T.; Correa, P. E.; Ho, T.-I.; McCoy, F.; Lewis, F. D. J. Am. Chem. Soc.
1979, 101, 3325-3331. (b) Lewis, F. D.; Simpson, J. T. J. Phys. Chem.
1979, 83, 2015-2019. (c) Hub, W.; Klu¨ter, U.; Schneider, S.; Do¨rr, F.;
Oxman, J. D.; Lewis, F. D. J. Phys. Chem. 1984, 88, 2308-2315.
(18) Aloisi, G. G.; Elisei, F.; Mazzucato, U.; Prats, M. J. Photochem.
Photobiol. A Chem. 1991, 62, 217-228.
(19) Saltiel, J.; Zhang, Y.; Sears, D. F., Jr.; Choi, J.-O. Res. Chem.
Intermed. 1995, 21, 899-921.
(20) (a) Bartocci, G.; Masetti, F.; Mazzucato, U.; Spalletti, A.; Orlandi,
G.; Poggi, G. J. Chem. Soc., Faraday Trans 2 1988, 84, 385-399. (b)
Bartocci, G.; Masetti, F.; Mazzucato, U.; Baraldi, I.; Fischer, E. J. Mol.
Struct. 1989, 193, 173-183.
a Toluene solvent; entries at 19.3 °C are from ref 8e.
of trans-stilbene17 and of related 1,2-diarylethylenes,18 and our
work has established that FN and O2 quench t-APE fluorescence
at diffusion-controlled rates, showing no selectivity between
t-APEA and t-APEB.11,19 Furthermore, the absence of fluores-
cence from potential t-APE/FN or c-APE/FN exciplexes renders
the FN choice ideal for our resolutions as no additional
fluorescent components are introduced.

11206 J. Am. Chem. Soc., Vol. 119, No. 46, 1997
Saltiel et al.
1
even stronger decrease in the lifetime of c-APEB*. Direct
) 350 nm), and that this decrease is concomitant with an
1
measurement of the fluorescence lifetime of c-APE in toluene
at ambient temperature (∼27 °C) has yielded τfc ) 4.5 ( 0.5
ns,8e consistent with O2 fluorescence quenching observations.
increase in fct*, the efficiency of the adiabatic c-APEB* f
1t-APEB* process. Furthermore, independent of temperature,
fully 50.5 ( 0.5% of the initially excited c-APE molecules that
1
Lifetimes at the temperatures employed in this work were
do not decay radiatively reach t-APEB* on the excited state
FN
qc-B
and k
qt-A qt-B
surface. It follows that radiative decay from 1c-APEB* competes
with activated torsional motion to 1t-APEB*. We are left with
the question of accounting for the remaining ∼50% of excited
c-APE molecules that decay radiationlessly but do not reach
1t-APEB*. Three possibilities are the following: (i) intersystem
crossing to the triplet state from either 1c-APE* conformer (φisc
) 0.17 has been reported22), (ii) competing excitation of
relatively nonfluorescent c-APEA, and (iii) diabatic photo-
isomerization from twisted 1pA* and 1pB* intermediates. Path-
ways i and iii can potentially contribute to the cis f trans
photoisomerization process, and to the degree that this happens,
observed quantum yields should exceed values predicted by the
estimated from the SV constants by assuming that k
roughly equal to the average of the known11
values
is
FN
FN
k
FN
SVc-B
2K
τfc-B
)
(3)
FN
qt-A
FN
(k
+ k
)
qt-B
This approximation neglects differences in the small contribu-
tions of transient terms to the diffusion-controlled rate con-
stants.11,21 The resulting lifetimes are listed in Table 4. It is
reassuring that the estimated value at 19.3 °C, τfc-B ) 4.3 ns,
is well within experimental uncertainty of the measured value.
If the FN quenching processes were all fully diffusion controlled,
then these lifetimes would be slightly overestimated, as the
transient term contribution to kq is expected to be larger in
shorter-lived species. The results show that τfc-B decreases by
at least a factor of 3 as the temperature is increased from 4.3 to
59.3 °C.
1
1
adiabatic c-APEB* f t-APEB* process. Pathway (i) should
be reflected in [c-APE] dependent φct values since APE triplets
are known to photoisomerize via a quantum chain process in
the 3c* f 3t* direction.5 Such a concentration dependence was
reported by Tokumaru and co-workers for the direct excitation
photoisomerization of c-APE and was attributed entirely to path
(i).5c
Pure Component Fluorescence Quantum Yields and Rate
Constants. Fractional contributions of t-APEB and c-APE
fluorescence, xt-B + xc ) 1.00, to the spectra on which the
quantum yields in Table 2 are based were obtained by
determining the location of (R,â) values for each spectrum on
the normalization line for each temperature, e.g., Figure 2.
Resulting fractional contributions, xt-B, for 350- and 390-nm
excitation are nearly identical, Table 2, at each temperature
despite a 16-22% drop in φhfc at the longer wavelength. Since
excitation at 390 nm is expected to increase the relative
population of 1c-APEA*, this result is consistent with the
conclusion that this conformer has a much smaller fluorescence
quantum yield. On the basis of the fractional contributions,
the effective fluorescence quantum yields are decomposed into
pure component fluorescence quantum yields, Table 5. Ne-
glecting incident light absorbed by c-APEA gives the minimum
Photoisomerization Quantum Yields. Under degassed
conditions, all φhct values in Table 3 exceed unity. In addition,
they are strongly [c-APE] dependent over the modest concentra-
tion range employed in our experiments. This behavior reflects
the participation of triplets in the photoisomerization.5 Neglect-
ing, for the moment, the presence of two distinct conformers
with conformer-specific rate constants, the simplest mechanism
that accounts for the observations is given by the following
sequence:
1c-APE
9
hν8 1c-APE*
(6)
(7)
(8)
kfc
1c-APE* 98 1c-APE + hν
kSct
1c-APE* 98 1t-APEB*
1
1
fractions of c-APEB* that reach t-APEB*, fct*
,
kisc
1c-APE* 8 3c-APE*
(9)
(10)
(11)
(12)
(13)
(14)
φft-B
φ0ft-B
fct*
)
(4)
kft
1t-APE* 98 1t-APE + hν
where φ0ft-B ) 0.88 is the pure component fluorescence
quantum yield of t-APEB.8e,11 Similarly, the minimum fractions
of c-APEB* that escape radiationlessly the potential energy
kist
1t-APE*
9
8 3t-APE*
1
kTct
3c-APE* 98 3t-APE*
minimum at the cisoid geometry and adiabatically reach the
potential energy minimum at t-APEB*, φesc, are given by
1
ct
kdt
3t-APE* 98 1t-APE
φft-B
φ0ft-B(1 - φfc)
esc
ct*
f
)
(5)
ket
3t-APE* + c-APE 98 1t-APE + 3c-APE*
In view of the temperature independence of τft-B, Table 4,20
φ0ft-B is assumed to be similarly temperature independent.
Equations 4 and 5 give minimum values for the desired
Application of the steady state approximation on all excited
species gives
φhct - fct*φ0ft-B ) (φhisc + fct*φi0st-B)(1 + ketτT[c-APE]) (15)
quantities because the φft-B’s are not corrected for incident light
esc
ct*
absorbed by c-APEA. Fractions fct* and f , based on the
larger fluorescence quantum yields for λexc ) 350 nm, are also
listed in Table 5.
Table 5 reveals that φfc decreases sharply with increasing
temperature (a factor of 3.3 in the 4.3 to 59.3 °C range for λexc
where the subscripts c and t designate cis and trans isomers,
respectively, and all other symbols have their usual meanings.
The term fct*φ0ft-B ) φft-B in eq 15 is the effective fluorescence
quantum yield of t-APEB following 366-nm excitation of c-APE.
Values of φhfc for λexc ) 340, 350, 360, and 366 nm are in the
narrow range of 0.49-0.53, and the fluorescence spectra for
(21) (a) Weller, A. Z. Phys. Chem. NF 1957, 13, 335-352. (b) Hui,
M.-H.; Ware, W. R. J. Am. Chem. Soc. 1976, 98, 4718-4727.

Photoisomerization of cis-1-(2-Anthryl)-2-phenylethene
J. Am. Chem. Soc., Vol. 119, No. 46, 1997 11207
Table 4. Fluorescence Lifetimes of t-APE and c-APE (Toluene)
τf, ns
a
a
b
T, °C
t-APEA
t-APEB
c-APEB
4.3
19.3
39.3
59.3
7.7
8.2
8.2
8.2
26.0
27.8
27.8
27.8
6.0
4.3
2.8
1.9
a By interpolation from ref 20. b Estimated by using eq 2.
Table 5. Species-Specific Quantum Yields
φft-B(λexc
)
φfc(λexc)
esc
ct*
350 nm)
f
ct* (λexc
)
f
(λexc )
T, °C 350 nm 390 nm 350 390 nm 350 nm)
4.3
19.3
39.3
59.3
0.333
0.381
0.405
0.413
0.283
0.330
0.342
0.331
0.236
0.150
0.098
0.071
0.205
0.126
0.081
0.066
0.378
0.433
0.461
0.469
0.495
0.509
0.511
0.505
Figure 5. The dependence of φhct on [c-APE]avg at 19.3 °C in toluene
(see eq 16).
°C. It follows that roughly 56% of the triplets of APE, formed
following excitation of a c-APE solution, are populated from
1c-APE*, a precursor with a ∼4.4-ns lifetime, and the rest come
from 1t-APEB*, a precursor with a 28-ns lifetime. The kinetics
of the growth of triplet-triplet absorption starting from 1c-APE*
should reflect the decays of the two precursors, but have not
been reported to date. It has been claimed, however, that the
3c-APE* f 3t-APE* process is too fast to be detected by
transient triplet-triplet absorption measurements in the nano-
second time scale.23 This conclusion is in good agreement with
photoisomerization quantum yields in the presence of oxygen5f,23
(see below).
these λexc are composed of essentially identical contributions
1
1
of fluorescence from c-APE* and t-APEB*.8e The average
effective fluorescence quantum yield from adiabatically formed
1t-APEB* for these λexc’s, 0.383 ( 0.013,8e is in very good
agreement with 0.381, the value found in this work for λexc
)
350 nm at 19.3 °C, Table 5. This quantity represents the
contribution of the singlet excited state adiabatic pathway to
cis f trans photoisomerization that proceeds strictly according
to Olson’s eq 1. The right-hand side of eq 15 gives the total
triplet contribution to the photoisomerization quantum yields.
1
It consists of two terms since triplets form directly from c-
1
APE* and, following its adiabatic formation, from t-APEB*.
The slope to intercept ratio of the plot in Figure 5 gives ketτT
) (5.0 ( 1.3) × 104 M-1, for toluene at 19.3 °C. This is
considerably larger than the value of ∼1 × 104 M-1 estimated
on the basis of the dependence of φhct on [c-APE] in benzene
under direct and biacetyl-sensitized excitation conditions.23
Some of the discrepancy is deceiving as it reflects primarily
the fact that the adiabatic photoisomerization pathway in S1 was
not taken into account in the earlier treatment of the direct
excitation results. In particular, our results are in reasonable
agreement with those obtained for the direct excitation of c-APE
in benzene by Arai et al.5e,f Their plot of φhct vs [c-APE] gave
The derivation of eq 15 assumes that the φhct values are based
on negligibly small conversions to t-APE. Experimentally, the
imposition of this condition would lead to highly inaccurate φhct
values. We have compromised by keeping conversions within
the 13-17% range, Table 3. The t-APE product interferes in
two ways. First, since 366 nm is very close to an isosbestic
point of the APE isomers,8e t-APE acts as an internal filter and
its presence leads to artificially low φhct values. Second, this
effect is alleviated somewhat because excitation of t-APE yields
3t-APE*, albeit inefficiently, whose participation in eq 15
sensitizes the reaction. We can correct roughly for these effects
by employing average concentrations: [c-APE]avg ) (1 - 0.5xt) ×
[c-APE]o and [t-APE]B ) 0.5xt[c-APE]o, where xt is the observed
fractional conversion to t-APE and the average fractional
contributions of c- and t-APE are given by xcavg ) (1 - 0.5xt)
and xtavg ) 0.5xt. Equation 15 is modified to
i ) 0.43 and s ) 4.2 × 103 M-1 23
.
Our data plotted in the
same way give i ) 0.437 ( 0.026 and s ) (5.6 ( 0.1) × 103
M-1. The difference in slope could reflect interference by
residual O2 in the Ar-outgassed solutions employed in the earlier
work. The effect of residual O2 should be negligible on the
adiabatic singlet photoisomerization pathway and should not
influence the intercept of the plot as is observed. Identical triplet
( φhct - xcavg fct*φf0t-B) ) [xcavg(φisc + fct*φi0st-B) + xtavgφhist] ×
(1 + ketτT[c-APE]) (16)
3
transient absorption spectra, assigned to t-APE*, have been
reported 20 and 90 µs following the excitation pulse starting
from either c-APE or t-APE.5b Under degassed conditions in
benzene, the triplet lifetime is τT ) 190 µs.5b On the basis of
this lifetime, our results give ket ) 2.6 × 108 M-1 s-1 for the
propagation step of the quantum chain process. We note that,
under Ar-outgassed conditions, incomplete O2 removal can lead
to significantly smaller triplet lifetimes. The longest reported
lifetime obtained for 3t-APE* in benzene with Ar outgassing is
117 µs,24 and Tokumaru and co-workers have used both 190
µs5c and 30 µs23 for τT in alternative interpretations of their
data.
The plot of (φhct - 0.38xcavg) vs the average concentration,
[c-APE]avg, is linear, Figure 5, r ) 0.9992, and gives values of
0.118 ( 0.026 and (5.88 ( 0.13) × 103 M-1 for the intercept
and slope, respectively. With the use of φhist ) 0.11, measured
by transient spectroscopy,22 the intercept in Figure 5 gives (φhist
+ fct*φ0ft-B) ) 0.12 ( 0.03 as the overall triplet yield following
excitation of c-APE. This value is probably well within
experimental uncertainty of the reported value of 0.17 that was
based on triplet-triplet absorption measurements in benzene.22
The use of φ0ist-B ) (1 - φ0ft-B) ) 0.12 in eqs 15 and 168e,11 is
justified because only 1t-APEB* is accessed adiabatically.
Together with fct* ) 0.43 from Table 5, it gives φhisc ) 0.07 (
0.03 as the effective intersystem crossing yield of c-APE at 19.3
Participation of 3t-APE as the chain carrier for photoisomer-
ization via the quantum chain process is eliminated in air-
saturated benzene or toluene solutions because O2 quenching
(22) Arai, T.; Karatsu, T.; Tsuchiya, M.; Sakuragi, H.; Tokumaru, K.
Chem Phys. Lett. 1988, 149, 161-166.
(23) Karatsu, T.; Tsuchiya, M.; Arai, T.; Sakuragi, H.; Tokumaru, K.
Bull. Chem. Soc. Jpn. 1994, 67, 3030-3039.

11208 J. Am. Chem. Soc., Vol. 119, No. 46, 1997
Saltiel et al.
Table 6. The Temperature Dependencies of Limiting Rate
diminishes τT by about 1000-fold.5b,24 This accounts for the
sharp decrease in φhct of c-APE that was observed at very low
[O2] in benzene.5b Starting from φhct ) 2.1 under O2-free
conditions, the quantum yield drops to ∼0.59 at [O2] = 4 ×
10-4 M and appears to roughly plateau there at higher [O2].5b
The fact that φhct remains as high as 0.50 at [O2] ) 7.7 × 10-3
M, the highest [O2] employed, was a puzzle to Tokumaru and
Arai5b because they were unaware of the singlet excited state
photoisomerization pathway. On the basis of our mechanism,
O2 may interfere with cis f trans photoisomerization only by
quenching c-APE excited states
1
Constants for Radiative and Radiationless Decay of c-APE* a
T, °C
10-8ksct+, s-1
10-8ksct-, s-1
10-7 fc-, s-1
k
4.3
19.3
39.3
59.3
1.27
1.99
3.20
4.99
0.63
1.01
1.64
2.52
3.9
3.5
3.5
3.8
a Quantum yields for λexc ) 350 nm were employed.
1t-APEB* pathway, 15% fluoresce, and ∼7% undergo intersys-
tem crossing. The radiationless decay pathways that account
for the remaining one-third of excited c-APE molecules do not
contribute to cis f trans photoisomerization and may involve
cyclization to dihydrophenanthrenes, particularly from 1c-
APEA*.25
kTqox
1c-APE* + 3O2
3c-APE* + 3O2
8 3c-APE* + 1O2* or 3O2
(17)
(18)
kSqox
8 1c-APE* + 1O2* or 3O2
1
The Energy Barrier for Torsional Motion in c-APEB*.
Semiempirical quantum mechanical calculations predict that the
Since O2 eliminates the quantum chain process, quenching of
1t-APEB* or 3t-APEB* has no effect on φhct, except that it
eliminates sensitization of the reaction by the product, t-APE.
1
perpendicular geometry, pA*, of the singlet excited state of
1
1
APEA is a transition state along the c-APEA* f t-APEA*
reaction coordinate 16 kcal/mol above the energy of 1c-
APEA*.20a This prediction is consistent with the absence of
photoisomerization via this reaction coordinate, including the
1
Quenching of c-APE* in toluene at 19.3 °C occurs with
OX
SVc
K
) 160 ( 15 M-1 8e
. Transient spectroscopic measure-
ments, using nanosecond time resolution, indicate that 3c-APE*
1
inaccessibility of pA* as a potential intermediate. No such
3
f t-APE* occurs too rapidly to be quenched by O2 in air-
calculations have been reported for APEB. However, on the
saturated solutions.5f For this limiting case, O2 interferes with
unimolecular c-APE photoisomerization only by quenching 1c-
APE* and the quantum yield is given by
1
basis of the rate constant for radiationless decay of c-APEB*
in toluene in 19.3 °C, the activation energy barrier along the
1c-APEB* f 1t-APEB* reaction coordinate was estimated as e7
kcal/mol.8e Quantitative evaluation of this barrier can now be
based on the temperature dependencies of the fluorescence
lifetime of c-APE, Table 4, and of the resolved fluorescence
quantum yields, Table 5. Sums of all nonradiative decay rate
constants, knrc, of c-APE are given by (1 - φhfc)/τfc. In addition
fct* + φisc + KSVc[O2]
φhct )
(19)
1 + KSVc[O2]
3
At the other extreme is the case for which O2 quenches all c-
APE*, effectively shutting down its adiabatic photoisomerization
by eq 14. For this case, the photoisomerization quantum yield
is given by
to ks , knrc includes radiationless decay via intersystem crossing
ct
as well as all other radiationless decay pathways of both
conformers. The knrc values set a maximum limit on ksct and
are listed in the ksct+ column in Table 6. Since the fraction of
incident light absorbed by c-APEA is unknown, minimum values
fct*
φhct )
(20)
of ks can be estimated from fct*/τfc. These values are listed in
1 + KSVc[O2]
ct
the ksct- column in Table 6. Also listed in Table 6 are
minimum values of kfc based on φhfc/τfc. It is gratifying that kfc
is temperature independent, within experimental uncertainty,
whereas ksct+ and kcst- both decrease by a factor of 4 upon
decreasing the temperature from 59.3 °C to 4.3 °C. The
ksct+/ksct- ratio is 2.0, independent of temperature. Both sets of
ksct values in Table 6 adhere nicely to Eyring’s transition state
equation
Both eqs 19 and 20 are consistent with our observation that φhct
is independent of [c-APE] in air-saturated toluene, Table 3. Since
all parameters in these equations are known, we can readily
3
calculate φhct ) 0.61 and 0.34 for none of c-APE* quenched
(eq 19) and for all of 3c-APE* quenched (eq 20), respectively.
The average experimental value for air-saturated toluene solu-
tions, φhct ) 0.58 ( 0.03, is in excellent agreement with the
3
value predicted for negligible c-APE* quenching by O2, eq
19, confirming the conclusion of Tokumaru and co-workers that
q
q
3
3
∆Sct
∆Hct
the c-APE* f t-APE* process is nearly barrierless in fluid
solutions.5f Our photoisomerization quantum yield in the
presence of air in toluene is also in excellent agreement with
Tokumaru’s observations in benzene.5f
κk
h
ln(ksct/T) ) ln
+
-
(21)
(
)
R
R
q
q
where ∆Hct is the enthalpy of activation, ∆Sct is the entropy
of activation, and κ is a transmission factor. Plots of ksct+ and
ksct- vs T based on eq 21 are shown in Figure 6. The two lines
We stress that the photoisomerization quantum yields in Table
3 are accounted for, nearly quantitatively, without the necessity
of postulating additional photoisomerization pathways involving
twisted singlet excited state intermediates from either conformer,
q
are astonishingly parallel, giving ∆Hct- ) 3.99 ( 0.12 and
s
s
q
∆Hct+ ) 3.93 ( 0.007 kcal/mol from kct- and the kct+ plots,
respectively. Since ksct+ includes substantial contributions
from all the other radiationless decay processes, including those
1
1pA* or pB*. All photoisomerization in c-APE singlet excited
states proceeds Via the conformer-specific adiabatic pathway
shown in eq 5. All other photoisomerization events involve
triplet states that are formed inefficiently by intersystem crossing
1
of c-APEA*, two possible explanations for this slope coinci-
dence are (i) all other radiationless decay pathways, coinciden-
1
1
from c-APE* and t-APEB*, φisc ) 0.07 and φist-B ) 0.12.
Thus following c-APE excitation at 366 nm in toluene at 19.3
°C fully 43% of 1c-APE* isomerize via the 1c-APEB* f
tally, have the same activation energies as does ks and (ii) fct*
ct
1
and φfc account for all c-APEB* decay.
(24) Wismontski-Knittel, T.; Das, P. K. J. Phys. Chem. 1984, 88, 1168-
1173.
(25) Laarhoven, W. H.; Cuppen, Th. J. H. M.; Nivard, R. J. F.
Tetrahedron 1970, 4865-4881.

Photoisomerization of cis-1-(2-Anthryl)-2-phenylethene
J. Am. Chem. Soc., Vol. 119, No. 46, 1997 11209
Figure 7. Transition state plot for ksct.
Figure 6. Transition state plots for ksct+ (9) and kcst- (b).
Table 7. The Temperature Dependencies of Quantum Yields and
1
Rate Constants for c-APEB* a
φct-B*(λexc
)
T, °C
350 nm
390 nm
10-8ksct, s-1
10-7kfc-B, s-1
4.3
19.3
39.3
59.3
0.616
0.743
0.825
0.877
0.611
0.748
0.828
0.851
1.03
1.74
2.93
4.72
6.4
6.0
6.2
6.6
a See text for assumptions; rate constants are based on λexc ) 350
nm data.
The second explanation appears more attractive. Normalized
fct* and φfc values were obtained by dividing each by the
corresponding (fct* + φfc) sum. This procedure neglects small
contributions of 1c-APEA* to φfc. The validity of the assumption
Figure 8. The temperature dependence of KAB, the c-APEB h c-APEA
equilibrium constant based on λexc ) 350 (b) and 390 nm (9).
1
that fluorescence from c-APEA* contributes negligibly to φfc
for λexc e 390 nm is confirmed by the fact that the two sets of
between the fct* values in Table 5 and the φct-B values in Table
7 is given exactly by
normalized fct* values, derived independently from the λexc
)
350 and 390 nm data, are nearly identical, Table 7 (normalized
fct* values are labeled φct-B* in Table 7). The resulting
conformer specific rate constants are also shown in Table 7.
fct* ) fc-Bφct-B
(23)
q
The transition state plot for kct*, Figure 7, gives ∆Hct ) 4.44
Substitution of eq 22 into eq 23 and rearrangement gives
( 0.14 kcal/mol and, with κ ) 1, ∆Sctq ) -5.7 ( 0.5 eu. The
calculated kfc-B value, (6.3 ( 0.3) × 107 s-1, is temperature
independent as expected.26 However, because the overall
intersystem crossing yield of c-APE is low, the other extreme
φct-B
ꢀc-A
ln
- 1 ) ln
+ ln KAB )
(
)
fct*
ꢀc-B
1
model for which only c-APEB* gives triplets also leads to a
ꢀc-A ∆SAB ∆HAB
reasonable fit of the data.26
ln
+
-
(24)
ꢀc-B
R
RT
Ground State Conformer Equilibration in c-APE. The
difference in slopes between the plots in Figures 6 and 7 is
significant. Temperature changes affect both ksct and the
equilibrium fraction of c-APEB in the conformer mixture. The
raw quantum yields employed to generate ksct( and the plots in
Figure 6 reflect both of these changes. Since we have shown
that c-APEA fluorescence is negligible for λexc e 390 nm, all
fluorescence arises from the fraction of excited c-APEB
where KAB is the equilibrium constant for c-APEB h c-APEA
conformer interconversion and ∆SAB and ∆HAB are the ground
state entropy and enthalpy differences between c-APEA and
c-APEB. Plots for the λexc ) 350 and 390 nm data, based on
eq 24, Figure 8, are parallel and linear (|r| ) 0.9953 and 0.9997,
respectively), and their slopes give ∆HAB ) 0.91 ( 0.06 and
0.93 ( 0.02 kcal/mol, respectively. It follows that c-APEB is
0.92 ( 0.04 kcal/mol more stable than c-APEA, consistent with
the conclusion that the more extended B conformer is similarly
more stable in t-APE.19,20,27 Neglecting entropy differences,
this energy difference predicts a small variation in the percent
[c-APEB] in the temperature range employed (from 84.0% to
80.0% as the temperature is increased from 4.3 to 59.3 °C).
Potential Energy Surface. A self-consistent interpretation
of the data has emerged based on the assumption that essentially
all 1c-APEB* molecules that do not fluoresce undergo adiabatic
molecules, fc-B
,
ꢀc-B[c-APEB]
fc-B
)
(22)
ꢀc-A[c-APEA] + ꢀc-B[c-APEB]
where ꢀc-A and ꢀc-B are the molar absorptivities of c-APEA
and c-APEB at a specified λexc, respectively. The relationship
(26) At the other extreme is the model that assigns all intersystem
crossing to 1c-APEB* and leads to kisc-B ) 2.38 × 107 s-1 at 19.3 °C. With
the further assumption that kisc-B is temperature independent, the 350-nm
data give ∆Hctq ) 4.73 ( 0.16 kcal/mol and, with κ ) 1, ∆Sctq ) -5.0 (
0.5 eu. However, kfc-B is now predicted to increase systematically from
5.5 × 107 to 6.8 × 107 s-1 as the temperature is raised from 4.3 to 59. 3
°C.
1
photoisomerization to t-APEB* over a 4.4 kcal/mol torsional
(27) (a) Bartocci, G.; Mazzucato, U.; Spalletti, A.; Elisei, F. Spectrochim.
Acta, A 1990, 46, 413-418. (b) Spalletti, A.; Bartocci, G.; Masetti, F.;
Mazzucato, U.; Cruciani, G. Chem. Phys. 1992, 160, 131-144.

11210 J. Am. Chem. Soc., Vol. 119, No. 46, 1997
Saltiel et al.
the basis of our estimated kfc-B, Table 7, the radiative lifetime
1
of c-APEB* is smaller by at least a factor of 2. Thus, the cis
f trans process in S1 appears to involve passing from a more
allowed cis electronic state to a more forbidden trans electronic
state. This difference in the nature of the electronic states of
the two isomers may render κ < 1 in eq 21 and may account
for the negative activation entropy that results when κ ) 1 is
assumed.
Concluding Remarks. In the lowest excited singlet state
the cis-trans isomerization of c-APE in toluene involves
efficient adiabatic one-way cis f trans torsional displacement
over a 4.4 kcal/mol barrier in 1c-APEB* only. This conformer-
specific reaction is a striking example of Havinga’s nonequili-
bration of excited rotamers (NEER) principle.29 It is remarkable
in that it illustrates conformer-specific reactivity that arises not
from structurally imposed proximity of the nuclei involved in
bond-breaking or -making, as is the usual case, but from
profound conformer-controlled differences in electronic distribu-
tion. It is these differences in electronic distribution that also
account for the conformer-specific photocyclization of 1c-
APEA*.25 Structural considerations alone do not explain the
Figure 9. Proposed potential energy curve for the 1c-APEB* f
1t-APEB* reaction.
1
1
different propensities of c-APEA* and c-APEB* toward cis
f trans photoisomerization. On the contrary, the steric interac-
tions that militate against the planar geometry in cis-stilbenes
and provide the driving force for torsional relaxation in S1 of
barrier. The energy difference between the ground states of
t-APEB and c-APEB has been attributed to the steric interaction
of the ortho hydrogens in the cis isomer and is roughly assigned5
the value of 5 kcal/mol as in stilbene.28 The origins of t-APE
and c-APE S1 r S0 transitions had been assigned values of
70.2 and 71.3 kcal/mol, respectively, on the basis of unresolved
spectra.5f We assign values of 67.6 and 73.3 kcal/mol for
t-APEB and c-APEB, respectively, on the basis of our resolved
fluorescence and absorption spectra.8e,11,19 The resulting po-
1
such molecules appear to be more severe in c-APEA* than in
1c-APEB*. In addition to the usual ortho hydrogen interaction,
in 1c-APEA* we have the proximity of a meta hydrogen of the
phenyl group to the peri hydrogen at the 9-position of the
anthracene moiety. Whether conformer control of electronic
distribution plays a role in the reactivity of photobiological
systems remains to be established. It is intriguing that the
double bond of the cyclohexene moiety of the retinyl chro-
mophore is s-cis to the acyclic double bond in rhodopsin but
s-trans in bacteriorhodopsin.30 The efficient double bond
photoisomerization that is triggered by light in these two systems
is in different directions and exhibits different regiochemistry.
1
1
tential energy curve for the c-APEB* f t-APEB* torsional
coordinate is shown in Figure 9. It accounts nicely for the
absence of trans f cis photoisomerization in S1 of APEB as
this process faces an insurmountable 15 kcal/mol activation
barrier. In principle, a shallow minimum cannot be ruled out
1
at pB*. However, the data require that any residence time at
Acknowledgment. This research was supported by NSF,
most recently, by Grant CHE-9612316.
that geometry be insufficient to allow radiationless decay from
there to contribute to cis f trans photoisomerization.
The radiative lifetime of S1 of t-APEB is at least 32 ns, as
the usual τf /φf calculation neglects contributions to these
quantities by the more allowed S2 f S0 emission.11,20b,27 On
JA972293A
(29) For a review see: Jacobs, H. J.; Havinga, E. AdV. Photochem. 1979,
11, 305-373.
(28) Saltiel, J.; Ganapathy, S.; Werking, C. J. Phys. Chem. 1987, 91,
2755-2758.
(30) For a review see: Birge, R. R. Biochim. Biophys. Acta 1990, 1016,
293-327.