ketones. The two substituents clearly have dramatically
opposite effects on stereoselectivity.
by minimizing OH H-bonding. The differences among the
Scheme 2 displays the computed minimum energy biradi-
cal conformers involved in this cyclization process, with BRt
able to undergo least motion radical coupling (rotations of
bonds a and b) easier than the higher energy BRc (rotation
of bond c also), as mentioned above. Our original study of
temperature effects on the diastereoselectivity of photocy-
clization by 25 revealed that the large trans/cis indanol ratio
is due to both enthalpic and entropic factors, which clearly
affect both 2-om and 2-cn as well. The fact that there are
both enthalpic and nonenthalpic cis/trans preferences is in
accord with there being at least two separate biradical
conformers that cyclize to diastereomers. We have suggested5
that the measured enthalpic differences most likely reflect
the thermodynamically controlled proportions of separate
biradical conformers as well as varied activation energies
for the bond rotations that allow cyclization of different
conformers to different isomers. The nonenthalpic differences
probably reflect different alignments of the two singly
occupied p orbitals in each of the biradical conformers that
presumably cyclize to different isomers. In the world of
ground states, intramolecular cyclization has a relatively fixed
negative activation entropy whether stereoisomers are formed
and regardless of reacting conformers with differing geom-
etries. But in the world of triplet biradicals, the nonenthalpic
spin-orbit induced conversion to singlet multiplicity required
for product formation does depend on biradical geometry.6
With two biradical conformers, rates of direct intersystem
crossing (isc) before cyclization would differ because of
different distances between the two radical sites and different
orientations of the singly occupied p orbitals. Rates of isc
coupled with radical-radical coupling would differ as the
two radical sites approach each other and as the p orbitals
realign for σ-bond formation.
Ketone 2 displays a 670 cal enthalpic preference and a
nearly 5:1 (28 eu) nonenthalpic (or entropic) preference for
trans cyclization.3 Both the enthalpic and nonenthalpic trans
preferences are slightly greater for 2-om, perhaps indicating
a small electron-donating effect by the methoxy group. The
cyano group in 2-cn has a larger electronic effect; the
enthalpic trans preference is 50% greater than it is in 2,
whereas the nonenthalpic trans preference is only 1/3 that
for 2. The enthalpic increase may reflect diminution of OH
H-bonding, which removes one impediment to the bond
rotation required for cyclization, while the nonenthalpic
decrease is likely due to partial electron withdrawal from
the benzylic carbon.
There are two ways that the para substituents’ electronic
effects can affect biradical behavior. The simpler one is that
the electron donor enhances the phenyl ring’s basicity, thus
favoring a large BRt/BRc ratio via OH H-bonding as
suggested above, while the strong electron withdrawer
reduces the ring’s basicity and thus lowers the BRt/BRc ratio
three reactants in quantum yields and stereoselectivity of
cyclization could well be due largely to this effect. However,
the substituents being para to the benzylic radical site adds
a factor to biradical behavior that has not received much
attention, namely the ionic effect of π conjugation. As Arnold
and co-workers have reported,7 both p-methoxy- and p-
cyanobenzyl radicals have delocalized spin, with the benzyl
carbon becoming either somewhat negative or positive,
respectively. With the spin density at the benzyl site of the
biradical lowered, spin-orbit coupling could be very dif-
ferent from that for a simple 1,5-pentanediyl biradical. This
could well explain why the nonenthalpic factor for 2-cn
displays the lowest trans/cis cyclization preference. Likewise,
the lack of any increase in indanol quantum yield that could
have been afforded by the methoxy group on 2-om may be
caused by the partial negative charge on each radical site
discouraging radical-radical coupling.
A possible cause of the low diastereoselectivity for 2-cn
and the low quantum yields for 1-cn and 2-cn may be the π
conjugation-induced opposite partial charges on the two
benzylic carbons of the biradical intermediates. Salem’s
original analysis of biradical multiplicity and spin-orbit
coupling6 suggested that opposite charges at the two radical
sites would favor primarily singlet multiplicity and thus
enhance the rate of cyclization by BRc and possibly enhance
the rate of disproportionation by the no-longer hydrogen-
bonding BRt. The low quantum yields may also result from
kH values being so low that significant CT quenching of the
triplet ketones by solvent benzene8 can compete and lower
the overall triplet lifetime, whereas triplet δ-hydrogen
abstraction by keones 1 and 2 is some three orders of
magitude faster than solvent quenching. In any event, these
results tend to verify our belief that internal hydrogen
bonding promotes both high quantum yields and diastereo-
selectivity in the photocyclization of R-aryl ketones via
δ-hydrogen abstraction from ortho substituents on the R-aryl
ring.
Finally, the significant effects that the two π-conjugating
substituents had on 1,5-biradical intermediates suggest that
additional studies, both experimental and computational, be
performed on how such substituents can affect triplet
biradical behavior.
Acknowledgment. This work was partially supported by
NSF Grant No. CHE98-11570.
Supporting Information Available: Synthesis of reactant
ketones, photochemical procedures, identification of indanol
photoproducts, and computational methods. This material is
(5) Wagner, P. J.; Zand, A.; Park, B.-S. J. Am. Chem. Soc. 1996, 118,
12856.
(6) Salem, L. and Rowland, C., Angew. Chem., Int. Ed. Engl. 1972, 11,
92; Michl, J. J. Am. Chem. Soc. 1996, 118, 3568.
(7) Nicholas, A.; de martin, P.; Arnold, D. R. Can. J. Chem. 1986, 64,
270.
(8) Wagner, P. J., Leavitt, R. A. J. Am. Chem. Soc. 1973, 95, 3669.
OL0528383
Org. Lett., Vol. 8, No. 4, 2006
647