Entropy vs enthalpy control of 1,5-biradical cyclization in the photochemistry of α-(o-alkylphenyl)acetophenones

Full text of DOI:10.1021/ja962683h

12856
J. Am. Chem. Soc. 1996, 118, 12856-12857
Scheme 1
Entropy Ws Enthalpy Control of 1,5-Biradical
Cyclization in the Photochemistry of
r-(o-Alkylphenyl)acetophenones
Peter J. Wagner,* Ali Zand, and Bong-Ser Park
Chemistry Department, Michigan State UniVersity
East Lansing, Michigan 48824
ReceiVed August 1, 1996
Table 1. Medium and Temperature Effects on Cyclization
Stereoselectivity
Over the past decade our group has explored δ-hydrogen
abstraction in various phenyl ketones in order to assess how
structural factors influence the formation and reactions of triplet
1,5-biradicals.1 We have noted the importance of benzylic
conjugation in inducing regio-2,3 and stereoselectivity4,5 and have
stressed the need to distinguish between pre-existing confor-
mational preferences and nonbonded interactions that are created
only during actual cyclization as factors determining product
ratios.5a During the same period the idea that triplet f singlet
intersystem crossing (isc) determines biradical lifetimes has
become widely accepted,6 although product ratios typically are
explained in terms of biradical geometries and product energies,
despite the possibility that varying rates of isc for different
biradical geometries could in fact dominate product ratios.7,8
We have attempted to resolve this inconsistency by suggesting
that isc occurs dynamically during cyclization rather than
statically beforehand.1,5 We report here temperature and
medium effects on the stereoselectivity of 1,5-biradical cycliza-
tion that, together with semiempirical computations of biradical
structure, reveal geometric effects on apparent entropies of
cyclization that may represent the first experimental evidence
for geometry-dependent biradical isc rates.
conditions
Z/E-2
11.5/1
14.6/1 (0.75)
16.5/1
26/1
2/1
100/0
0.69
Z/E-4
16/1
21/1 (0.52)
24/1
31/1
4/1
100/0
0.55
8
Z/E-6
1.5/1
1.2/1 (0.40)
1.0/1
0.7/1
1/1
1/1
-0.66
3.5
toluene, 110°
benzene, 24° a
toluene, 0°
toluene, -72°
methanol, 24°
crystal, 24°
EE - EZ, kcal/m
AZ/AE
4.6
a Product quantum yields in parentheses.
Scheme 1 compares the indanol yields obtained from the
ambient temperature photocyclization of R-(o-benzylphenyl)-
acetophenone (5) to those from the previously reported R-(o-
ethylphenyl)acetophenones 1 and 3.5 Z/E ratios for 1 and 3
are large in benzene but, as usually observed in such cycliza-
tions,9 smaller in methanol. In stark contrast, 5 shows negligible
selectivity.
Figure 1. Z/E indanol ratios for 9 3, 2 1, and b 5.
Table 1 lists product ratios in solution as a function of
temperature for these ketones10 plus results for crystals. The
corresponding Arrhenius plots are shown in Figure 1, with the
activation parameter differences listed in Table 1. These results
add to our earlier findings in two important ways: (1) changing
the hydrogen donor from an ethyl to a benzyl group eliminates
the stereoselectivity; (2) nonenthalpic factors appear to be as
important as enthalpy differences at determining diastereose-
lectivity. Our original interpretation of the room temperature
product ratios for 1 and 3 invoked cyclization from two biradical
conformational minima, one pre-Z and one pre-E, that differed
in energy by 1.7-2.0 kcal/mol.5a The measured ∆Ea values
for 1 and 3 indicate that enthalpy differences between biradical
conformers may be much smaller than we originally surmised,
with much of the Z/E preference contained in the A factor for
cyclization. Therefore, we have reexamined the geometric
aspects of all the steps in this reaction.
Equation 1 describes product ratios when conformational
equilibrium is established among various biradical geometries
before cyclization.11,12 The subscripts z and e refer to biradical
z′
Z/E ) øBRzkzcycl/øBRekceycl + ø
k
/ke′
(1)
BRx cycl cycl
conformers that can cyclize with minimal bond rotations to Z
or E indanol; x refers to any other conformer that must rotate
into a pre-Z or pre-E geometry before cyclizing. The ø values
(Σøi ) 1) describe the Boltzman distribution of conformers.
The k values contain only relative efficiencies of isc (as an
entropy term) if isc occurs independently and if subsequent
biradical cyclization is faster than any conformational inter-
conversion.7 If singlet biradicals interconvert before cyclizing
or if isc occurs during cyclization, then k values also contain
differences in activation enthalpies for cyclization of different
conformers.
(1) Wagner, P. J. Acc. Chem. Res. 1989, 22, 83.
(2) Wagner, P. J.; Meador, M. A.; Scaiano, J. C. J. Am. Chem. Soc.
1984, 106, 7988.
(3) Wagner, P. J.; Giri, B. P.; Pabon, R.; Singh, S. B. J. Am. Chem. Soc.
1987, 109, 8104.
(4) Wagner, P. J.; Subrahmanyam, D.; Park, B.-S. J. Am. Chem. Soc.
1991, 113, 709.
(5) Wagner, P. J.; Park, B.-S. Tetrahedron Lett. 1991, 32, 165. Wagner,
P. J.; Meador, M. A.; Zhou, B.; Park, B.-S. J. Am. Chem. Soc. 1991, 113,
9630.
(6) (a) Scaiano, J. C. Acc. Chem. Res. 1982, 15, 252. (b) Caldwell, R.
A. Pure Appl. Chem. 1984, 56, 1167. (c) Doubleday, C., Jr.; Turro, N. J.;
Wang, J.-F. Acc. Chem. Res. 1989, 22, 199.
(7) Scaiano, J. C. Tetrahedron 1982, 38, 819.
Scheme 2 depicts for 1 and 3 the likely topology for biradical
formation and reaction. The geometric requirements for
(8) Griesbeck, A. G.; Stadtmuller, S. J. Am. Chem. Soc. 1991, 113, 6923.
(9) Wagner, P. J. J. Am. Chem. Soc. 1967, 89, 5898. Wagner, P. J.; Kelso,
P. A.; Kemppainen, A. E.; McGrath, J. M.; Schott, H. N.; Zepp, R. G. J.
Am. Chem. Soc. 1972, 94, 7506. Wagner, P. J.; Chiu, C. J. Am. Chem. Soc.
1979, 101, 7134. Wagner, P. J. Acc. Chem. Res. 1989, 22, 83.
(10) GC analysis of product ratios has provided slightly lower values
than originally estimated5 from NMR integration alone.
(11) Wagner, P. J. Acc. Chem. Res. 1983, 16, 461.
(12) Equation 1 is a modified form of the Curtin-Hammett principle in
that it includes conformers that cannot react directly but which must be
considered if static isc is the rate-determining step for biradical decay.
S0002-7863(96)02683-2 CCC: $12.00 © 1996 American Chemical Society

Communications to the Editor
J. Am. Chem. Soc., Vol. 118, No. 50, 1996 12857
Scheme 2
product ratios reflect equilibrium biradical conformational
populations, as predicted by eq 1. The A factor ratio indicates
a lower entropy of cyclization for BRe, which is twisted around
bonds a and b such that the two singly occupied p orbitals are
almost orthogonal but not pointed at each other, as they are in
BRz. The nonoverlapping orbital orientations in BRe and the
twist of the R-hydroxy radical site would decrease both through
space and through bond spin-orbit coupling.17 This effect would
produce less singlet-triplet mixing and thus less efficient isc
in BRe than in BRz if isc is static, occurring at conformational
minima before cyclization, or if isc is dynamic, occurring
concurrently with the bond rotation that swings the two p orbitals
toward each other. In this regard, the behavior of BRx, which
represents some 42% of the triplet biradical population, is very
important. Does it undergo static isc to form a singlet biradical
that must undergo rotation around bond b in order to cyclize?
If so, nearly half the product ratio could be determined by the
relative rates for rotation of BRx to BRz or BRe. As far as
our calculations can determine, there is a 0.7 kcal/mol difference
in the two rotational barriers, which indicates at most a 3:1
preference for formation of singlet BRz. Assuming the same
static isc rate for BRe and BRx, which have similar bond angles,
a 10-fold greater rate of isc for BRz is required to fit the
measurements to the calculations. That large of a difference
seems unlikely; we are currently analyzing the behavior of
several more such biradicals.
The lack of stereoselectivity for 5 is consistent with our earlier
findings that phenyl substitution substantially lowers rotational
barriers at a benzyl radical site in a biradical, since both benzene
rings do not have to conjugate with the half-occupied p orbital.18
As shown below, biradical geometries with the phenyl tilted
both up and down are of comparable energy but have signifi-
cantly different orbital orientations, which could explain the AZ/
AE ratio, as for 1. Why Ea is higher for the Z isomer is not
immediately apparent. The situation when both radical sites
are conformationally mobile is obviously complicated and
requires additional computational study.
hydrogen transfer5b limit the R-hydroxy radical site to a single
initial orientation, with the OH facing the central benzene ring.
Both models and calculations indicate that BRz would be the
major initial conformer; any BRi also formed would have only
a ∼4 kcal/mol barrier to cross in order to form the more stable
BRz. Thus, we estimate that the δ-methyl group is anti to the
other radical center in >99% of the reacting 1,5-biradicals. The
fact that both 1 and 3 yield only Z indanols in the crystal
confirms that BRz is the predominant initial biradical geometry
and that rotations around the R-hydroxy radical site, which
presumably are forbidden in the crystal, are necessary for
formation of E indanols. This conclusion would demand that
the stereoselectivity for 1 and 3 depends on geometric varia-
tions at the R-hydroxy radical site, as Scheme 2 shows. The
large solvent effects on 1 and 3 strongly support such a
conclusion, as originally suggested.5 Moreover, as we discuss
below, the contrasting behavior of 5 further solidifies these
conclusions.
Calculations on the conformational distribution around the
R-hydroxy radical site of the biradical indicate only two other
minima within 5 kcal/mol of BRz, the global minimum: BRx
has the same energy as BRz but cannot cyclize directly; BRe
lies only 0.4 kcal/mol above BRz.13,14 Comparable results are
calculated for 3. The calculated energy of BRz varies by 1.5
kcal/mol depending on whether the OH hydrogen points toward
or away from the central benzene ring. The former, favored
geometry represents hydrogen bonding by the OH to the benzene
ring, a well-established phenomenon15,16 that apparently controls
selectiVity in these reactions. Solvation of the OH group lowers
øBRz by reducing this stabilization of BRz.
These findings and the spin-orbit calculations on biradicals17
of Michl provide an explanation for the activation parameters.
Given the 1 kcal/mole activation energy for biradical decay,5b
part of the measured ∆Ea could be generated during cyclization.
Nonetheless, the similar values for ∆Ea and the calculated
enthalpy difference between BRz and BRe do suggest that the
Acknowledgment. This work was supported by NSF grants CHE91-
20931 and CHE94-08308. The NMR spectrometers used were partially
funded by NSF grant CHE88-00770 and NIH grant RR047550. A.Z.
is grateful for a Harold Hart Fellowship; B-S.P. is grateful for a BASF
fellowship.
(13) Minimization of the lowest energy conformations found by an AM1
double-dihedral drive search.
(14) We thank Dr. Pablo Wessig of Humboldt Universitat in Berlin for
sharing with us his similar calculations.
(15) (a) Oki, M.; Iwamura, H. Tetrahedron 1968, 24, 1905 and references
therein. (b) Biali, S. E.; Rappoport, Z. J. Am. Chem. Soc. 1984, 106, 5641
and references therein.
(16) Wagner, P. J.; Park, B.-S.; Sobczak, M.; Frey, J.; Rappoport, Z. J.
Am. Chem. Soc. 1995, 117, 7619.
(17) Michl, J. J. Am. Chem. Soc. 1996, 118, 3568.
JA962683H
(18) Wagner, P. J.; Meador, M. A.; Scaiano, J. C. J. Am. Chem. Soc.
1984, 106, 7988. Wagner, P. J.; Giri, B. P.; Pabon, R.; Singh, S. B. J. Am.
Chem. Soc. 1987, 109, 8104.