Quantitative cavities and reactivity in stages of crystal lattices: Mechanistic and exploratory organic photochemistry

Article abstract of DOI:10.1021/ja0124562

In continuing our research on solid-state organic photochemistry, we have been investigating the phenomenon of reactivity in stages. In this study we present new examples where the photochemical reactivity changes discontinuously at some point in the conversion. In these instances, the reaction course of the solid differs from that in solution. One example is the reaction of 2-methyl-4,4-diphenylcyclohexenone, where an unusual reaction course was encountered in the solid state; and, of two possible mechanisms, one was established by isotopic labeling. A second case is that of 4,5,5-triphenylcyclohexenone. The solid-state reaction of this enone was found to give a new photochemical transformation, the Type C rearrangement, a process that involves a δ to α aryl migration. In the case of 3-tert-butyl-5,5-diphenylcyclohexenone the Type C rearrangement occurred even in solution. The stage behavior was investigated using X-ray analysis and Quantum Mechanics/Molecular Mechanics computations. This permitted us to determine the sources and details of the stage phenomenon. The analysis revealed how a product molecule as a neighbor affects reactivity. The computations were employed to follow the course of a solid-state reaction from reactant through the succeeding stages. Additionally, the Delta-Density Analysis was utilized to ascertain the electronic nature of molecular changes. Besides product composition changing with extent of conversion, the reaction quantum yield was found to change as one stage gave way to a succeeding one.

Full text of DOI:10.1021/ja0124562

Published on Web 02/19/2002
Quantitative Cavities and Reactivity in Stages of Crystal
Lattices: Mechanistic and Exploratory Organic
Photochemistry1,2
Howard E. Zimmerman* and Evgueni E. Nesterov
Contribution from the Chemistry Department of the UniVersity of Wisconsin,
Madison, Wisconsin, 53706
Received October 29, 2001. Revised Manuscript Received December 21, 2001
Abstract: In continuing our research on solid-state organic photochemistry, we have been investigating
the phenomenon of reactivity in stages. In this study we present new examples where the photochemical
reactivity changes discontinuously at some point in the conversion. In these instances, the reaction course
of the solid differs from that in solution. One example is the reaction of 2-methyl-4,4-diphenylcyclohexenone,
where an unusual reaction course was encountered in the solid state; and, of two possible mechanisms,
one was established by isotopic labeling. A second case is that of 4,5,5-triphenylcyclohexenone. The solid-
state reaction of this enone was found to give a new photochemical transformation, the Type C
rearrangement, a process that involves a δ to R aryl migration. In the case of 3-tert-butyl-5,5-
diphenylcyclohexenone the Type C rearrangement occurred even in solution. The stage behavior was
investigated using X-ray analysis and Quantum Mechanics/Molecular Mechanics computations. This
permitted us to determine the sources and details of the stage phenomenon. The analysis revealed how
a product molecule as a neighbor affects reactivity. The computations were employed to follow the course
of a solid-state reaction from reactant through the succeeding stages. Additionally, the Delta-Density Analysis
was utilized to ascertain the electronic nature of molecular changes. Besides product composition changing
with extent of conversion, the reaction quantum yield was found to change as one stage gave way to a
succeeding one.
Introduction
ing molecules. We defined a subset of the X-ray crystal lattice
as a “mini-crystal lattice”.
Our research in solid-state photochemistry began 15 years
ago with the philosophy of treating crystal-lattice photochemistry
quantitatively. One major objective dealt with the pioneering
research of Cohen and Schmidt, who noted that the reaction of
a molecule can be envisioned as taking place in a cavity
composed of surrounding molecules of the lattice.3 This
qualitative, but invaluable, concept has been invoked in the last
four decades in a myriad of solid-state photochemical studies.
Surprisingly, despite the availability of X-ray structures, the
cavity concept remained qualitative. Thus, in our initial crystal-
lattice studies we concluded that, with the coordinates of all
atoms surrounding a single lattice molecule known, the cavity
could be defined quantitatively.4 Computational extraction of
one molecule leaves the cavity, which then is quantitatively
defined by the coordinates of the nearest atoms of the surround-
Our initial efforts focused on the “fit” of a reacting species
in the cavity of the mini-crystal lattice. “Fit” was gauged by
the overlap of the reacting species with the surrounding lattice.4
These initial, somewhat primitive, efforts were followed by
estimation of the energy of the reacting molecule by molecular
mechanics.5,6 But with the realization that molecular mechanics
omitted the electron delocalization effect, in subsequent efforts,
we constructed the cavity with an inert gas shell obtained by
replacing the nearest neighboring atoms by heliums for hydro-
gens and neons for carbon, oxygen, and nitrogen. With this
irregularly shaped cavity shell held rigid, it was possible to treat
the imbedded reacting molecule quantum mechanically.7 Most
recently,8,9 we have made use of the combined QM/MM
(ONIOM) programming by Morokuma10 as imbedded in Gauss-
* Address correspondence to this author. E-mail: zimmerman@
chem.wisc.edu.
(1) This is Paper 197 of our photochemical series and Paper 264 of our general
series.
(2) (a) For Paper 262 see: Zimmerman, H. E.; Wang, P. HelV. Chim. Acta
2001, 84, 1342-1346. (b) For a preliminary Communication of a portion
of the present results see: Zimmerman, H. E.; Nesterov, E. E. Org. Lett.
2000, 2, 1169-1171.
(3) (a) Cohen, M. D.; Schmidt, G. M. J. J. Chem. Soc. 1964, 1996-2000. (b)
Cohen, M. D. Angew. Chem., Int. Ed. Engl. 1975, 14, 386-393.
(4) (a) Zimmerman, H. E.; Zuraw, M. J. J. Am. Chem. Soc. 1989, 111, 2358-
2361. (b) Zimmerman, H. E.; Zuraw, M. J. J. Am. Chem. Soc. 1989, 111,
7974-7989.
(5) (a) Zimmerman, H. E.; Zhu, Z. J. Am. Chem. Soc. 1994, 116, 9757-
9758. (b) Zimmerman, H. E.; Zhu, Z. J. Am. Chem. Soc. 1995, 117, 5245-
5262.
(6) Zimmerman, H. E.; Sebek, P. J. Am. Chem. Soc. 1997, 119, 3677-3690.
(7) Zimmerman, H. E.; Sebek, P.; Zhu, Z. J. Am. Chem. Soc. 1998, 120, 8549-
8550.
(8) Zimmerman, H. E.; Alabugin, I. V.; Smolenskaya, V. N. Tetrahedron 2000,
56, 6821-6831.
(9) Zimmerman, H. E.; Alabugin, I. V.; Chen, W.; Zhu, Z. J. Am. Chem. Soc.
1999, 121, 11930-11931.
(10) (a) Maseras, F.; Morokuma, K. J. Comput. Chem. 1995, 16, 1170-1179.
(b) Matsubara, T.; Sieber S.; Morokuma, K. Int. J. Quantum Chem. 1996,
60, 1101-1109.
9
2818 VOL. 124, NO. 11, 2002 J. AM. CHEM. SOC.
10.1021/ja0124562 CCC: $22.00 © 2002 American Chemical Society

Reactivity in Stages of Crystal Lattices
A R T I C L E S
Scheme 2. Synthesis of the C-13 Labeled 2-Methyl Enone 9
Scheme 1. Synthesis of tert-Butyl Enone 4
Scheme 3. Synthesis of Photoproduct 14
ian 98.11 This permitted us to treat the reacting molecule with
ab initio quantum mechanics while treating the surrounding
molecules with molecular mechanics.8,9 This also enabled
quantitative assessment of crystal relaxation effects.
With this methodology available we proceeded to investigate
a number of further solid-state reactions and encountered a
remarkable phenomenon.2b Hitherto, it has been assumed that
solid-state reactions sometimes could be brought to high
conversions, but that more often at some point the crystal
structure is lost and the reaction leads to a heterogeneous, useless
aggregate. Our finding was that crystal lattice reactions often
proceed in stages, and that in those subsequent stages different
and useful products tend to be formed. The present paper deals
with that phenomenon and also describes a number of new
reactions and their theoretical treatment.
in Scheme 3. The elimination step, involving conversion of 11
to 12, proved difficult, primarily due to opening of the three-
membered ring under most conditions. However, a simple
thionyl chloride reaction under the conditions shown afforded
practical yields along with an equal amount of 1,3,4-triphenyl-
cyclohexa-1,4-diene (15). Another problem was encountered in
this synthesis: most oxidative conditions failed for the conver-
sion of the triphenylbicyclohexanol 13 to the corresponding
ketone (Swern oxidation, PCC, and most other CrO3 derived
reagents). However, Dess-Martin oxidation was successful. The
initial product was the 4-endo-phenylbicyclic stereoisomer,
which was identified by NMR analysis and epimerized directly
to afford the desired 4-exo-diastereomer 14 as depicted in
Scheme 3.
The stereochemical assignments were from NMR analysis
and are given in the Experimental Section. The case of alcohol
13 was especially critical and an assignment of the phenyl group
at C4 as endo was based on a comparison of HC4-HC3
coupling constants with literature examples.12
(B) Solid-State Photolyses. (1) 2-Methyl-4,4-diphenylcy-
clohex-2-enone (9). Previously2b,5 we have noted that solid-
state photolysis of 2-methyl enone 9 affords not only the
common Type B enone photoproduct 1613 but also the ring-
contracted cyclopentenone 17. Additionally, there was evidence
that the product ratio was dependent on the extent of conversion.
Thus, we planned to study the process more closely. The
reaction proceeds in stages2b with products 16 and 17 being
formed in a 1:1 ratio up to 10% conversion. At this point the
exclusive photoproduct being formed becomes the bicyclic
Results
(A) Synthesis. The present study required a number of
photochemical reactants as well as the independent synthesis
of certain photoproducts. Thus three syntheses are noteworthy
and are given in Schemes 1-3. In the case of the synthesis of
tert-butyl enone 4, allylic oxidation of cyclohexene 2 did afford
the desired enone 3; however, this was a minor product (ca.
25%); and enone 1 resulted as the major product. Nevertheless,
the diphenylcyclohexenone 1 formed was just recycled, and thus
the synthesis proved to be practical.
A second synthesis was of 13C-labeled 2-methyl enone 13C-
9. This was required for mechanistic studies (vide infra). A
practical synthesis began with 13C-labeled methyl iodide and
thence via formation of the phosphonium salt, the Wittig
reaction, hydroboration, Swern oxidation, and Robinson anne-
lation led nicely to the required 2-methyl product uniquely
labeled at C3. This is outlined in Scheme 2.
A more difficult synthesis was required for preparation of
an unusual bicyclic photoproduct 14. This synthesis is outlined
(11) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann,
R. E.; Burant, J. C.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.; Strain,
M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi, R.; Mennucci,
B.; Pomelli, C.; Adamo, C.; Clifford, C.; Ochterski, J.; Petersson, G. A.;
Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.; Rabuck, A. D.;
Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Stefanov,
B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.;
Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.;
Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; Johnson,
B.; Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-Gordon,
M.; Replogle, E. S.; Pople, J. A. Gaussian 98, Revision A.9; Gaussian,
Inc.: Pittsburgh, PA, 1998.
(12) Rees, J. C.; Whittaker, D. Org. Magn. Reson. 1981, 15, 363-369.
(13) (a) Zimmerman, H. E.; Wilson, J. W. J. Am. Chem. Soc. 1964, 86, 4036-
4042. (b) Zimmerman, H. E.; Rieke, R. D.; Scheffer, J. R. J. Am. Chem.
Soc. 1967, 89, 2033-2047. (c) Zimmerman, H. E.; Kutateladze, A. G. J.
Org. Chem. 1995, 60, 6008-6009.
9
J. AM. CHEM. SOC. VOL. 124, NO. 11, 2002 2819

A R T I C L E S
Zimmerman and Nesterov
Figure 1. Stages and changes in reactivity with extent of conversion of reactant 9. The expanded graph (b) shows the area in graph a corresponding to low
conversions. Plot 1 corresponds to photoproduct 16, plot 2 corresponds to 17.
Scheme 4. Photochemistry of
2-Methyl-4,4-diphenylcyclohex-2-enone (9)
diphenyl bicyclic ketones, lacking the 2-methyl group, are
known to undergo cis-trans isomerization.14
(2) 4,5,5-Triphenylcyclohexenone (18). In previous studies15
we investigated the solution photochemistry of 4,5,5-triphenyl-
cyclohexenone (18). The corresponding behavior in the crystal
lattice proved quite different and particularly interesting. Both
reaction courses are illustrated in Scheme 6 along with details
of the stage behavior that was once again encountered. Hence,
the main products in solution were stereoisomeric cyclobu-
tanones 19 and the exo stereoisomer of bicyclic ketone 20. The
solid-state products differed completely. The bicyclic ketone
14 resulted from an unusual and novel rearrangement, termed
Type C. Also the endo stereoisomer 21, a product of the Type
B rearrangement, was formed. We note here the complete
reversal of the 6-exo-phenyl stereochemistry characteristic of
the solution photochemistry.
This example is particularly striking in giving a complete
change of products in the second reaction stage. Thus in stage
1 photoproduct 14 was formed exclusively while in stage 2
bicyclic ketone 21 resulted as the major product. Again, the
stage behavior had an early onset, at 12%. Note Figure 2. Ketone
21 is still another example of a product available synthetically
only in the second stage of a multistage reaction.
One additional observation is of importance. At 12% conver-
sion stage 1 ends and stage 2 begins. However, at close to one-
half of this conversion the quantum yield drops sharply. This
could be termed the existence of stage 1a and stage 1b, namely
a crystal lattice change within stage 1. This is depicted in Figure
3. We know that there is some change in the environment of
the reactant at the end of stage 1a to account for the drop in
quantum yield. The simplest assumption is the formation of one
product 14 as a neighbor at this point. This means, from the
quantum yield behavior in Figure 3, that at the end of stage 1,
there must be two neighboring molecules of 14.
Scheme 5. Two Alternative Labeling Results
ketone 16. Cyclopentenone 17 formed from stage 1 remained
except for a slow loss due to secondary conversion5 to bicyclic
16. The process finally leads to its total disappearance at the
end of stage 2. Scheme 4 gives these results, and the stage
behavior is shown in Figure 1 and summarized in the table in
Scheme 4 that gives incremental amounts of formation of the
products in the two stages. The zero value for 17 in stage 2
might well have been given a negative incremental value but
does clarify that this isomer is not being formed.
The unusual formation of the five-membered-ring product
17 had little precedent in the photochemistry of 4,4-diaryl-
substituted cyclohexenones. One can consider two, alternative
mechanisms. One involves direct ring-contraction and the other
involves a secondary reaction to be discussed (vide infra). By
use of the 13C-labeled reactant it was possible to differentiate
between the possibilities. The two labeling alternatives are
shown in Scheme 5 with the differing dispositions of the
originally labeled â-carbon in the products. Experimentally, the
NMR spectrum of the photoproduct 17 showed that the 13C
resulted at the benzhydryl carbon (i.e. label with an asterisk in
Scheme 5), thus precluding the simple ring contraction alterna-
tive (vide infra).
(3) 3-tert-Butyl-5,5-diphenylcyclohexenone (4). Interest-
ingly, tert-butyl enone 4 behaved quite differently than the
(14) Zimmerman, H. E.; Hancock, K. G.; Licke, G. J. Am. Chem. Soc. 1968,
90, 4892-4911.
One further interesting point is that bicyclic ketone 16 turned
out to be unreactive in solution (note eq 1), although 5,6-
(15) Zimmerman, H. E.; Solomon, R. D. J. Am. Chem. Soc. 1986, 108, 6276-
6289.
9
2820 J. AM. CHEM. SOC. VOL. 124, NO. 11, 2002

Reactivity in Stages of Crystal Lattices
A R T I C L E S
Scheme 7. Solution Photochemistry of tert-butyl Enone 4
Figure 2. Stages and changes in reactivity with the extent of conversion
of reactant 18. Plot 1 corresponds to photoproduct 14, plot 2 corresponds
to photoproduct 21.
with a quantum yield of 0.017. This chemistry is shown in
Scheme 7.
(C) Computations. Our theoretical approaches have been
based on our “mini-crystal lattice concept” which permits
quantitative description of the reaction cavity and of the behavior
of the reacting species within this cavity. Such a modeling of
the solid-state reaction is of importance since in many cases it
is the only possible method to get a reasonable answer to why
a reaction occurs in a given way. The mini-crystal lattice is
derived from the X-ray data for the reactant by using a special
computer program.5b It is composed of a central molecule
surrounded by a sufficient number of the neighboring molecules
building the reaction cavity. Then the central molecule is
computationally transformed to a molecule representing a
transition state of the reaction followed by molecular mechanics
and then ONIOM optimization of this transition structure inside
the reaction cavity. This produces the geometry and energy of
this species, which can be directly compared with those for the
alternative species and reaction pathways.
Figure 3. Conversion dependence of solid-state quantum yield for 4,4,5-
triphenylcyclohex-2-enone (18). The values of quantum yield are multiplied
by 104.
Scheme 6. Solution and Solid-State Photochemistry of
4,5,5-Triphenylcyclohex-2-enone (18)
The method allows ab initio optimization of reacting species
inside a two-layered minicrystal lattice that is treated by
molecular mechanics. The first layer is composed of the closest
surrounding molecules (5-25 molecules) and represents a
reaction cavity. All the other more distant molecules of the mini-
lattice are included in the second layer. The first layer is allowed
to optimize, while the second layer is kept frozen during the
entire optimization. An optimization of the reacting cavity
together with the reacting species allows removal of stress
imposed on the reacting central molecule by the reaction cavity
(i.e. crystal pressure).16 More details of the computational
method may be found in the Experimental Section.
ftriphenyl enone 18, which gave the Type C δ to R migration
only in the crystal lattice but afforded the Type B product in
solution. The solution photolysis of 4 led to three photoprod-
ucts: two isomeric bicyclic ketones, 22 and 23, and cyclohex-
enone 24. We note that all three photoproducts arose from a
Type C transformation.
(1) 2-Methyl-4,4-diphenylcyclohex-2-enone (9). In the case
of the unusual stage photochemistry of 2-methyl-4,4-diphenyl-
cyclohex-2-enone (9), we turned to computational modeling at
various points in the reaction course. A first problem was to
analyze why an originally formed bicyclic product 16 undergoes
such an unusual transformation to cyclopentenone 17. To this
end, a molecule of bicyclic triplet 16T was optimized inside a
mini-lattice consisting of ground-state neighbors. Then this
The solid-state photolysis of enone 4 also exhibited the Type
C reactivity, but in contrast to the solution behavior, selective
formation of a single Type C photoproduct 23 was encountered,
and this persisted to 67% conversion. Beyond this limiting
conversion, crystal melting precluded a search for a stage 2
reaction. Also the solid-state reaction proved quite efficient
triplet (16Tcr) was compared with the same triplet (16Tgas
)
optimized as a single molecule (i.e. gas phase). The ab initio
(16) McBride, J. M. Acc. Chem. Res. 1983, 16, 304-312.
9
J. AM. CHEM. SOC. VOL. 124, NO. 11, 2002 2821

A R T I C L E S
Zimmerman and Nesterov
Table 2. Ab Initio Energies for Triplet Intermediates 25Ta
Table 1. Comparison of Energies of Bicyclic Triplets 16Ta
ab initio energy, aub
∆E, kcal/molb
ab initio energy, aub
∆E, kcal/mol
16Tgas, optimized in gas phase
16Tcr, optimized in crystal
16Tcr′, optimized in crystalc
-804.7447991
-804.6830413
-804.6909315
0
38.8
33.8
25Tgas, optimized in gas phase
25Tcr, optimized in crystal
25Tcr′, optimized in crystalc
-804.7344249
-804.6929934
-804.7160792
0
26.0
11.5
a Geometries were optimized with HF/6-31G* for the gas phase and with
ONIOM (HF/6-31G*: Dreiding) for the enone crystal. b HF/6-31G*, note
energy differences of over 0.5 kcal/mol are significant. c Formed from the
second reacting molecule in the surrounding shell of enone molecules but
with the one adjacent enone converted to cyclopentenone 17.
a Geometries were optimized with HF/6-31G* for the gas phase and with
ONIOM (HF/6-31G*: Dreiding) for the crystal. b HF/6-31G*. c Formed
from the second reacting molecule in the surrounding shell.
Figure 4. Results of the Delta-Density analysis of the bicyclic triplet 16T
and suggested bond 1-6 fission to form biradical 32. Numbering in the
three-membered ring is shown in italics, ∆D changes are given in bold
print.
energies of both species are given in Table 1, where we note
the triplet 16Tcr to be 38.8 kcal/mol higher in energy than its
gas-phase counterpart. The importance of these relative energies
is considered in the Discussion section below.
Figure 5. Calculated (GIAO RHF/6-31G(df,pd)//B3LYP/6-31G(d,p))
versus experimental NMR chemical shifts for photoproduct 17.
Additionally, there is a striking difference in geometry relative
to 16Tgas. The C6 to ipso(phenyl) bond of 16Tcr is bent from
normal with a distortion angle 11° and seems responsible for
the much higher energy of this intermediate. Interestingly, almost
the same distortion was obtained in the preliminary molecular
mechanics optimization, indicating that the distortion mainly
results from intermolecular steric repulsion. The last species,
16Tcr′, in Table 1 is a neighbor to 16Tcr, which is converted to
cyclopentenone product.
9 onward to photoproduct 16. The energetic results of these
computations are given in Table 2.
A quantitative analysis of the distribution of bond distortions
was obtained from our Delta Density analysis.17 This method
compares the density matrices for two species as in eq 2. In the
present instance, the two species of interest are the relaxed, gas-
phase enone triplet 16Tgas and the triplet 16Tcr as formed in
the crystal lattice. The basis orbitals used are the Weinhold
NHO’s.18 Positive elements correspond to strengthened bonds
in 16Tcr relative to 16Tgas while negative elements correspond
to weakened bonds. This analysis revealed that the C1-C6 bond
of the bicyclic triplet 16Tcr is markedly weakened (by 0.3
electrons) as illustrated in Figure 4, which gives ∆D values for
bonds of interest. This accounts for the high reactivity of 16 in
stage 1 in proceeding onward to form 17.
(2) 4,5,5-Triphenylcyclohexenone (18). The conversion of
triphenyl enone 18 in the solid state to the bicyclic product 14
involves a novel transannular δ to R phenyl migration as noted
earlier. As a check on the structure assignment, it seemed
1
prudent to obtain the theoretical 13C and H NMR chemical
shifts for this photoproduct. This was done by using GIAO
computations, as available in Gaussian 98. The values are listed
in the Figure 5. The agreement with experiment is quite good.
A second item concerned the photochemistry. To understand
the reaction mechanism proceeding via an unusual δ to R phenyl
migration, which we term a Type C rearrangement, ab initio
computation of species along the mechanistic pathway proved
helpful. Similarly, we followed the pathway of the Type B
rearrangement of triphenyl enone 18; however, in this case, our
computations included only species 26T and 27T and the
transition structure between them. Energies of the intermediate
species of interest as obtained computationally with Hartree-
Fock UHF/6-31G* geometry optimization are summarized in
Table 3. These results led to the question of whether the two
reactions proceed via different reactant triplets of enone 18, with
these triplets differing not only in energy and electronic
configuration but also in geometry. Indeed this was the case.
Table 3 includes energies for triplets 26T (planar) and 29T
∆D ) Dcr - Dgas
(2)
To understand the product distribution in the reaction of enone
9 it seemed necessary to evaluate the structure and energy of
the half-migrated species 25T leading from the triplet of enone
(17) (a) Zimmerman, H. E.; Alabugin, I. V. J. Am. Chem. Soc. 2001, 123, 2265-
2270. (b) Zimmerman, H. E.; Alabugin, I. V. J. Am. Chem. Soc. 2000,
122, 952-953.
(18) (a) Foster, J. P.; Weinhold, F. J. Am. Chem. Soc. 1980, 102, 7211-7218.
(b) Reed, A.; Curtiss, L. A.; Weinhold, F. Chem. ReV. 1988, 88,
899-926.
9
2822 J. AM. CHEM. SOC. VOL. 124, NO. 11, 2002

Reactivity in Stages of Crystal Lattices
A R T I C L E S
Scheme 8. Mechanisms of Type B versus Type C Enone Rearrangements for Enone 18
Table 3. Energies of the Intermediate Species Included in the
Photochemistry of Enone 18
Table 6. Comparison of Energies of Reacting Intermediates for
Two Competing Pathways in Different Mediaa
Type B migration
(intermediate 27T)
Type C migration
(intermediate 30T)
ab initio
energy, aub
rel energy,
kcal/molb
reaction speciesa
ab initio
rel energy,
ab initio
rel energy,
A, S0 of enone 18, relaxed geometry
B, π-π* twisted triplet 29T
C, TS 29T f 30T
D, half-migrated triplet intermediate 30T
E, triplet biradical 31T
F, bicyclic product 14, S0
G, n-π* planar triplet 26T
H, TS 26T f 27T
-995.3351504
-995.2492053
-995.2005462
-995.2359950
-995.2804044
-995.3361895
-995.2684720
-995.2419138
-995.2581979
0.7
54.6
85.2
62.9
35.1
0
42.5
59.2
49.0
energy, aub
kcal/molb
energy, aub
kcal/molb
gas phase
crystal
-995.2581979
-995.1830517
0
47.2
31.3
-995.2359950
-995.2243063
-995.2050612
0
7.3
19.4
crystal, stage 2c -995.2083939
a Geometries were optimized with HF/6-31G* for the gas phase and with
ONIOM (HF/6-31G*: Dreiding) for the crystal. b HF/6-31G*. c Molecule
reacting in stage 2, optimized in the crystal together with two molecules of
product 14 (see the Discussion Section).
I, half-migrated triplet intermediate 27T
a Geometry was optimized with the HF/6-31G* level. b HF/6-31G*.
Table 7. Energies of Ground and Triplet States of tert-Butyl
Enone 4
Table 4. Delta Density Computation Results for the Two Excited
States of the Triphenyl Enone 18a,b
ab initio
rel energy,
kcal/molb
reaction speciesa
energy, aub
orbitals
planar triplet 26T
twisted triplet 29T
π CdC
π CdO
ny
-1174.9
-2391.8
-9146.0
-4483.8
-1420.9
-866.0
ground state, relaxed geometry
π-π* twisted triplet
n-π* planar triplet
-921.9353821
-921.8662925
-921.8450080
0
43.4
56.7
a UHF/6-31G*. b Versus ground-state S0 as a reference point.
a Geometry was optimized with the HF/6-31G* level. b HF/6-31G*.
Table 5. CASSCF/NBO Analysis of the Electronic Configuration of
the Triplets 26T and 29Ta,b
Computations were also carried out for the reaction inter-
mediates 27T and 30T in the mini-crystal lattice of triphenyl-
enone 18 obtained from the X-ray coordinates. This was done
with ONIOM. The results are given in Table 6.
(3) 3-tert-Butyl-5,5-diphenylcyclohexenone (4). With rela-
tively close energies for both the twisted and planar triplet states
of the triphenylenone 18, we proceeded to look at the 3-tert-
butyl-5,5-diphenylcyclohexenone (4) excited states. The energies
obtained are given in Table 7.
To search for a source of the high efficiency and selectivity
in the crystalline state reaction of this enone, an ONIOM
computation in the mini-crystal lattice was carried out for the
half-migrated triplet counterpart of 30T. Also computations were
carried out on the two alternative ground-state products 22 and
23. These results are shown in Table 8. The interpretation of
these results is given in the Discussion section.
c
electronic population
orbitals
planar triplet 26T
twisted triplet 29T
π2*
π1*
n
π2
π1
0.4198
1.0141
1.0073
1.6568
1.8929
0.2230
0.9235
1.9238
0.9853
1.8643
a Geometries were optimized with CASSCF(6,5)/6-31G* with the NBO
basis set. b CASSCF energies: for 26T -995.2822647 au, for 29T
-995.2796512 au. c In electrons.
(twisted). These were obtained as local energy minima by
starting the optimization, in one case (26T), with the ground-
state enone 18 and, the other (29T), from the search for the
transition structure between 26T and 30T. Both pathways are
outlined in Scheme 8. This is further considered in the
Discussion section.
Discussion
To establish the electronic structure of these triplets, we used
our Delta-Density analysis and also CASSCF/NBO computa-
tions.19 These computational results on triplets 26T and 29T,
planar and twisted, are given in Tables 4 and 5.
The Mechanism of Formation of Cyclopentenone 17. It
has been noted above and in Scheme 5 that two alternative
(19) Nemukhin, A. V.; Weinhold, F. J. Chem. Phys. 1992, 97, 1095-1108.
9
J. AM. CHEM. SOC. VOL. 124, NO. 11, 2002 2823

A R T I C L E S
Zimmerman and Nesterov
Table 8. Computational Results for the Enone 4 Type C Rearrangement Speciesa
ab initio energy, au
∆E, kcal/molb
half-migrated intermediate optimized in the gas phase
half-migrated intermediate optimized in the crystal
photoproduct 23 optimized in the crystal
-921.8353739
-921.8289199
-921.8465375
-921.8224607
0
4.1
0
photoproduct 22 optimized in the crystal
15.1
a Geometries were optimized with HF/6-31G* for the gas phase and with ONIOM (HF/6-31G*:Dreiding) for the crystal. b Comparison between lines
1-2 and 3-4, respectively.
Scheme 9 The Mechanism of Formation of 17 According to the
Labeling Experiment
proceeds to form a molecule of 17 that then also has an enone
neighbor (i.e. termed 17(9) or 9(17)) which, in turn, is strongly
perturbed and is triggered to a very fast reaction of this enone
molecule 9(17) leading quickly to another molecule of 16 with
a neighbor of 17. Hence the process uses two enone molecules
to form one molecule of 16 and one of 17, and the ratio remains
1:1 overall.
Thus 16(9) and 9(17) are postulated to be particularly reactive
while bicyclic molecules with a cyclopentenone product 17
neighbor (i.e. 16(17)) are postulated to be unreactive. This is a
unique requirement for 1:1 formation of 16 and 17 in stage 1
of the reaction.
The high reactivity of 16(9) in the crystal cavity for bond 1-6
fission compared with the lack of reactivity in solution is
understood from Delta-Density considerations. Thus, as noted
in the section on Computational Results, a Delta Density analysis
of the bicyclic ketone triplet 16Tcr, geometry optimized within
the mini-crystal lattice composed of enone reactant, and
compared with its optimized gas-phase counterpart 16Tgas
,
Scheme 10. Understanding of Constancy of the 16 to 17 Ratio in
revealed bond 1-6 to have a large negative matrix element and
thus to be selectively weakened. Note Figure 4. It is this bond
that is opened in the mechanism leading to photoproduct 17 as
outlined in Scheme 9. Thus, bond 1-6 weakening arises from
distortion of the phenyl group attached to C6 by the cavity.
The computational data in Tables 1 and 2 also are in accord
with the experimentally observed reactivity presented above.
Interestingly, even before use of quantum mechanical optimiza-
tion, molecular mechanics optimization revealed that the bicyclic
triplet 16Tcr formed inside the crystal cavity is of much higher
energy than its gas-phase analogue 16Tgas. An ab initio
difference of 38.8 kcal/mol was found. The energy of 16Tcr,
corresponding to 16(9) in Scheme 10, is higher than the energy
of 16Tcr′, which corresponds to 16(17) in this scheme (note Table
1). This suggests a higher reactivity for the initially formed 16
compared with the second molecule of 16 produced. Also, this
is consistent with the lack of reactivity of 16(17). Table 2 gives
information about the “half-migrated” triplet species 25T of the
two processes of interest. In this table 25Tcr, corresponding to
the bridged species leading from 9 to 16(9), is of considerably
higher energy than 25Tcr′, which is the bridged intermediate
formed from 9(17) affording 16(17). Thus 9(9) is much less reactive
compared with 9(17). The net theoretical conclusion is that the
initial enone reaction occurs somewhat more slowly than the
subsequent reaction of a second molecule of the reaction cluster.
A further point concerns stage 2. It is seen that the energies
of the half-migrated intermediates of the surrounding unreacted
enone molecules at the beginning of stage 2 are in the range of
25-27 kcal/mol as determined by MM3 computation. This
compares with that for the initial central molecule of stage 1
(i.e. 25Tcr) where the energy is 26.9 kcal/mol. Hence these
species do not react competitively with the two molecules of
stage 1, but react only slowly at stage 2.
Stage 1
mechanisms were considered for the formation of photoproduct
17 that was formed in the solid-state photolysis in stage 1; see
Scheme 1. The direct ring contraction mechanism is certainly
the simplest and was initially favored. However, this mechanism
would leave the 13C label at the â carbon of the cyclopentenone
ring as outlined in Scheme 5 and not at the benzhydryl site.
Clearly the product arises from the mechanism outlined in
Scheme 9, which depicts an initial Type B phenyl migration
followed by a reaction of the initially formed bicyclic photo-
product 16. Opening of the three-membered ring in this manner
as noted has precedent in cis-trans stereoisomerization of
bicyclo[3.1.0]hexanones.14 In this case triplet diradical 32
undergoes a phenyl migration with concomitant intersystem
crossing to afford the photoproduct 17 with the experimentally
observed labeling.
One further point is the contrast between two findings: (a)
the sequential mechanism of conversion of 9 to 16 to 17 as
required by the labeling evidence and (b) the constant 1:1 ratio
of 16 to 17 during stage 1 that forms 16 and 17. Our
interpretation is outlined in Scheme 10. Here molecule pairs
are represented as “MOLECULE(Neighbor)”. Every molecule of
16 formed in stage 1 (we term this 16(9)) has a closest neighbor
enone 9 molecule adjacent. Each of these molecules 16(9)
9
2824 J. AM. CHEM. SOC. VOL. 124, NO. 11, 2002

Reactivity in Stages of Crystal Lattices
A R T I C L E S
two enone triplets, 29T is of higher energy by 12.1 kcal/mol
(note Table 3). The Delta Density computations reveal that the
triplet 26T had lost almost 10 times the py orbital density
compared with the twisted triplet 29T. Also, it had lost much
more π-bond density from the carbonyl. This identifies triplet
26T as n-π*. In contrast, the triplet 29T had lost electron
density mainly from the R,â-carbon bond, which identifies it
as π-π*.
The CASSCF/NBO computations19 using NBO’s as a basis
and an active space of (6,5) aimed at confirming the assignment
of these triplets. The active space was taken to include only
those eigenfunctions heavily participating in the excitation
process. Thus, the active space was comprised of four butadiene-
like orbitals and an oxygen lone pair (i.e. n or py) orbital. This
permitted efficient computation, and the results are given in
Table 5. Triplet 26T is seen again to involve lone pair electron
excitation (n-π*) while for triplet 29T a π-π* configuration
is dominant.
Turning to the stage phenomenon, from the quantum yield
versus conversion measurements (note Figure 3), we can see
that stage 2 does not begin until two molecules of reactant have
undergone stage 1 reaction. Thus, the quantum yield drops
precipitously after every reactant molecule has one molecule
of photoproduct as a neighbor. The second ONIOM computation
was carried out with the two neighbors of the central reactant
molecule replaced by photoproduct 14 molecules. The choice
of the neighbors had been taken at the locus where the Type C
rearrangement resulted in maximum geometric change and thus
maximum cavity shape modification. The result of this com-
putation is the third entry in Table 6. Here it is seen that there
is a predicted preference for the Type B rearrangement in stage
2. Moreover, the energetic preference of 2.1 kcal/mol over the
Type C pathway corresponds approximately to the observed
selectivity in stage 2 of 6:1.
The Mechanism of the 3-tert-Butyl-5,5-diphenylcyclohex-
enone 4 Reaction. It was originally computationally predicted
that the Type C rearrangement originates from the π-π* triplet
rather than the n-π* triplet. The latter commonly affords the
Type B rearrangement. The Type C rearrangement of the tert-
butylcyclohexenone 4 was of particular interest because it
occurred even in solution. Here, even without a crystal environ-
ment, geometry optimization (gas phase) led to a twisted triplet
(note Table 7) that is lower in energy. The corresponding planar
triplet proved to be 13.3 kcal/mol higher in energy as a
consequence of the steric effect of the tert-butyl group in
facilitating π-bond twisting. Thus, the more stable twisted π-π*
triplet led to the solution Type C rearrangement in this case.
Two important points are worth discussing. The first is the
high efficiency of the solid-state photoreaction of enone 4.
Extrapolated to zero conversion, quantum yield turned out to
be 0.017, which is in the usual range of the quantum yields for
many solution photorearrangements (e.g. the solution quantum
yield for the Type B enone rearrangement was determined to
be 0.04321). To explain this ease of crystalline Type C
rearrangement, a half-migrated intermediate corresponding to
the δ-R phenyl migration for the enone 4 has been optimized
with ONIOM inside the mini-crystal lattice. The results are given
in Table 8. For comparison purposes an energy of the same
Figure 6. Schematic representation of the reaction path for the photochemi-
cal rearrangements of 4,5,5-triphenylcyclohex-2-enone (18). The dashed
line corresponds to S0 and the solid lines correspond to two different T1
states.
There is another point requiring comment concerning our use
of molecular mechanics for this computation. It is clear that
molecular mechanics does not afford good absolute accuracy
as electronic (e.g. delocalization) effects are neglected. But since
the main source of intermolecular interactions in these systems
comes from steric repulsions, for comparative purposes, mo-
lecular mechanics with the MM3 force field is reliable.
The Mechanism of the Reaction of 4,5,5-Triphenylcyclo-
hexenone (18). The first item requiring attention is the unusual
course of the solid-state rearrangement of the triphenylenone
18. As outlined in Scheme 6, the known solution reaction15 leads
to the usual Type B bicyclohexanone product 20 and, with ring-
contraction, to vinylcyclobutanones 19. In contrast, in stage 1
of the solid-state reaction the exclusive formation of triphenyl
bicyclic 14, termed a Type C rearrangement, involves a δ to R
migration.
The preference for the Type B rearrangement in solution,
rather than the Type C, can be understood from inspection of
the data in Table 3. Thus, the half-migrated intermediate 27T
of the Type B reaction is 13.9 kcal/mol lower in energy than
the corresponding 30T of the Type C process.
Moreover, a comparison of the alternative transition states
TS1 and TS2 (note Figure 6) shows a preference of 26.0 kcal/
mol for the Type B rearrangement. We also see that the
calculated activation energy of 16.7 kcal/mol obtained for the
triplet Type B rearrangement as the difference between entries
H and G of Table 3 (TS2 and 26T in Figure 6) is in good
agreement with the experimental value of 10-11 kcal/mol for
the reaction in solution.20
However, for the crystal reaction, ONIOM computations for
the same half-migrated species imbedded in the usual reactant
mini-crystal lattice make clear the preference for this unusual
reaction in stage 1. Thus, it is seen that the energy of the δ to
R (Type C) migration is preferred over the common Type B
process by 7.3 kcal/mol.
Additionally, as noted, the computations reveal that the Type
C rearrangement is characteristic of the π-π* twisted triplet.
Thus, two energy minima were obtained for the triplet of enone
18, these corresponding to structures 26T and 29T. The former,
26T, had a nearly planar enone moiety (i.e. O-CR-Câ-Cγ
dihedral angle of 0.5°). However, 29T had the π-bond strongly
twisted with a dihedral angle of 35°. We note in that, of the
(21) Zimmerman, H. E.; Hancock, K. G. J. Am. Chem. Soc. 1968, 90, 3749-
(20) Zimmerman, H. E.; Elser, W. R. J. Am. Chem. Soc. 1969, 91, 887-896.
3760.
9
J. AM. CHEM. SOC. VOL. 124, NO. 11, 2002 2825

A R T I C L E S
Zimmerman and Nesterov
species optimized in the gas phase is included. One can see
that the in-crystal optimized intermediate is only 4.1 kcal/mol
higher than the corresponding gas-phase species. This reflects
a very close “fit” of the shape of the crystalline reaction cavity
to the reacting intermediate, as well as a spatial “compactness”
of the type C transition state. Both factors allow the reaction to
occur in the solid state with almost the same efficiency as in
the solution.
The second point deals with a very high selectivity of the
solid-state reaction that affords only one product 23 as compared
with the solution reaction where three isomeric type C photo-
products 22-24 form. Again, such a high selectivity of the
crystalline reaction may be qualitatively explained by the good
fit of the reaction cavity to the only product ketone 23. To obtain
a quantitative support for this statement, the ONIOM optimiza-
tions of two alternative isomers 22 and 23 inside the enone 4
mini-crystal lattice have been conducted. The results are
included in Table 8. An energetic preference of 15.1 kcal/mol
for product 23 versus 22 has been found that explains its
exclusive formation.
The Role of Cluster Size and Relaxation Effects. The last
subject for discussion is the role of the size of the “reaction
cluster”, i.e., the number of surrounding molecules, that affects
the reaction of a central molecule and whose reactivity is
affected by it. This cluster may be considered as a minimal
independent reacting unit of which the entire crystal is com-
posed. Clearly it will include all the closest neighbors which
form the first surrounding shell. Yet it is certainly smaller than
our mini-crystal lattice. As was shown in a previous paper,8
molecules more remote from the central molecule than the
nearest shell have a minor effect on the central molecule
reactivity. Thus, the “reacting cluster” practically coincides with
the first surrounding shell. In the present cases of 2-methyl enone
9 and triphenyl enone 18, this shell includes 14 molecules, so
one can anticipate the size of the “reacting cluster” to consist
of 15-20 molecules. This explains qualitatively why the switch
between two stages in both solid-state reactions occurs at 10-
12% conversion. This corresponds to the point where 2
molecules of 15-20 have already reacted, and only a slower
reaction of the others takes place.
packing potentials in which an empirical van der Waals function
is used to assess the energy of the molecular pair in the cavity.23
Warshel studied an excimer using an early QM/MM method.24
One elegant approach utilized the proximity of an individual
functional group to its nearest neighbor as in pretty studies by
Lahav and Scheffer.25 A different approach involved calculation
of cavity volumes in a pioneering study by Gavezzotti.26 Local
stress has been utilized in important studies by McBride.16
However, in essentially all of these studies, the methods relied
on reactant characteristics in the crystal to understand its
reactivity. Our approach has been to compare the reacting
species in the lattice with the original reactant. This is analogous
to ground-state solution chemistry where it is helpful to compare
a transition state with reactant rather than considering the
reactant alone.
In particular, our finding of the phenomenon of reaction stages
has provided an additional chance for obtaining photoproducts
which are otherwise inaccessible. While hitherto an early loss
of original crystal lattice ordering had been considered a
limitation, this now promises to reveal subsequent stages with
novel reactivity. Finally, the brilliant, but qualitative, concept
of Cohen and Schmidt in envisioning crystal cavities in which
reactions occur is now replaced by a “quantitative cavity” in
which the cavity is precisely defined with the consequence that
reactivity in the cavity can be computationally analyzed.
Experimental Section
General Procedures. All reactions were performed under an
atmosphere of dry nitrogen. Melting points were determined in open
capillaries and are uncorrected. Column chromatography was performed
on silica gel (Aldrich, 60 Å, 200-400 mesh, or Silica gel 60 Geduran
35-75 µm) mixed with Sylvania 2282 phosphor and slurry packed
into quartz columns to allow monitoring with a hand-held UV lamp.
All solvents were additionally purified and dried by standard techniques.
1H and 13C NMR spectra were recorded at 300 and 75 MHz respectively,
and are reported in ppm downfield from tetramethylsilane.
Solution Photolyses. Solution photolyses were carried out in an
immersion reactor with a water-cooled jacket using a 400 W medium-
pressure mercury lamp equipped with a 5 mm filter of a circulating
0.4 M solution of CuSO4 unless stated otherwise. All solutions were
purged with deoxygenated nitrogen for 1 h both prior to and during
photolysis.
General Procedure for Solid-State Photolysis. Irradiation of fine
crystalline material was performed between two quartz slides (5.0 ×
5.0 cm2) using a water-cooled medium-pressure 400 W mercury lamp
equipped with a 2 mm Pyrex filter. To prevent possible melting during
photolysis the samples were cooled in an ice-water bath. Each sample
contained 50-75 mg of reactant crystals. For time-dependent photolysis
purposes 5 mg portions were taken at fixed times and the product ratios
Relation to Previous Studies and Final Conclusion
Solid-state photochemistry has long exhibited remarkable
promise for synthetic purposes. When we began our efforts in
1985, solid-state organic chemistry already had a very large
number of examples where the crystal photochemistry differed
from that in solution.22 A restriction was that understanding was
limited and largely qualitative. Further, the reactions investigated
most often involved minimal molecular reorganization. Some
quantitative treatments focused on reactant geometry.
1
were determined using H NMR.
4,4-Diphenylcyclohexene (2) was prepared according to the pro-
cedure of Zimmerman et al.27 in 64% yield as colorless crystals, mp
61-63 °C (lit.27 mp 65-66.5 °C).
Most important, Cohen and Schmidt had proposed the concept
of the cavity in which molecules reacted.3 These researchers
also proposed the idea of least motion,3a and this has been
termed the Topochemical Principle. A quantitative postulate was
that for [2+2] cycloadditions a distance of 4.2 Å or less between
π bonds was required.22 Thomas and Ramamurthy advanced a
still more quantitative theory for [2+2] cycloadditions involving
5,5-Diphenylcyclohex-2-enone (3). To a stirred solution of 18.0 g
(76.9 mmol) of 4,4-diphenylcyclohexene (2) in 600 mL of benzene
(23) (a) Thomas, N. W.; Ramdas, S.; Thomas, J. M. Proc. R. Soc. London A
1985, 400, 219-227. (b) Murthy, G. S.; Arjunan, P.; Venkatesan, K.;
Ramamurthy, V. Tetrahedron 1987, 43, 1225-1240.
(24) Warshel, A.; Shakked, Z. J. Am. Chem. Soc. 1975, 97, 5679-5684.
(25) (a) Chang, H. C.; Popovitz-Biro, R.; Lahav, M.; Leiserowitz, L. J. Am.
Chem. Soc. 1987, 109, 3883-3893. (b) Ariel, S.; Askari, S. H.; Scheffer,
J. R.; Trotter, J.; Wireko, F. J. Am. Chem. Soc. 1987, 109, 4623-4626.
(26) Gavezzotti, A. J. Am. Chem. Soc. 1983, 105, 5220-5225.
(27) Zimmerman, H. E.; Epling, G. A. J. Am. Chem. Soc. 1972, 94, 7806-
7811.
(22) Schmidt, G. M. J. Solid State Photochemistry; Ginsberg, D., Ed.; Verlag
Chemie: New York, 1976.
9
2826 J. AM. CHEM. SOC. VOL. 124, NO. 11, 2002

Reactivity in Stages of Crystal Lattices
A R T I C L E S
were added Celite (50.0 g) and PDC (44.0 g, 117.0 mmol). The resulting
mixture was cooled to 10 °C, and tert-butyl hydroperoxide (11.7 g of
90% solution in t-BuOH, 116.7 mmol) was added dropwise. The
reaction mixture was stirred for 26 h at room temperature. Then it was
filtered through a Celite pad, and the filter was thoroughly washed
with ether. The filtrate was successively washed with 5% HCl and water
and dried over Na2SO4. After rotary evaporating the solvents, an oily
residue was separated by column chromatography on silica gel (column
47.0 × 4.0 cm2, eluent ethyl acetate-hexane 1:5) to give the following
fractions: (a) 8.6 g of starting material as a colorless oil, crystallizing
upon standing, Rf 0.78; (b) 5.5 g (57% based on consumed starting
material) of 4,4-diphenylcyclohex-2-enone (1) as a colorless solid, Rf
0.47, which was recrystallized from an ether-hexane mixture to give
colorless prisms, mp 94-97 °C (lit.28 mp 96-97 °C); (c) 2.0 g (21%
based on consumed starting material) of 5,5-diphenylcyclohex-2-enone
(3) as a colorless solid, Rf 0.33, which was recrystallized from hexane-
CH2Cl2 mixture to give 1.9 g of colorless crystals, mp 104-105 °C
(lit.27 mp 110.5-112.5 °C). 1H NMR (CDCl3) δ 7.32-7.14 (m, 10H),
7.01 (dt, J1 ) 9.6 Hz, J2 ) 5.4 Hz, 1H), 6.04 (dt, J1 ) 9.6 Hz, J2 ) 1.8
Hz, 1H), 3.22 (s, 2H), 3.19 (dd, J1 ) 4.2 Hz, J2 ) 1.8 Hz, 2H).
3-tert-Butyl-5,5-diphenylcyclohex-2-enone (4). A solution of t-BuLi
(9.5 mL of 1.7 M solution in pentane, 16.2 mmol) was added to a
stirred suspension of 0.73 g (8.1 mmol) of CuCN in 35.0 mL of ether
at 0 °C. The resulting mixture was stirred for 10 min, and a solution of
1.0 g (4.0 mmol) of 5,5-diphenylcyclohex-2-enone (3) in 40.0 mL of
ether was added dropwise at 0 °C, followed by stirring at this
temperature for 3 h. A solution of PhSeBr29 was prepared by addition
of Br2 (1.42 g, 0.46 mL, 8.87 mmol) to a solution of 2.87 g (9.2 mmol)
of diphenyldiselenide in 20.0 mL of ether at intensive agitation. This
PhSeBr solution was added rapidly to the reaction mixture (this was
accompanied by immediate decolorization). The resulting mixture was
poured into 150 mL of 5% HCl and 150 mL of ether-pentane (1:1).
The organic fraction was separated, washed with saturated NaHCO3
solution and water, dried over Na2SO4, and concentrated in vacuo. The
oily residue was dissolved in 100 mL of methanol, and 0.77 g (9.2
mmol) of NaHCO3 in 15.0 mL of H2O and 7.88 g (36.8 mmol) of
NaIO4 in 35.0 mL of H2O were added. After the mixture was stirred at
room temperature for 3.5 h, the resulting solution was poured into a
mixture of 80 mL of saturated NaHCO3 solution and 100 mL of ether-
pentane (1:1). The organic layer was washed with water and concen-
trated NaCl solution, dried over Na2SO4, and concentrated in vacuo.
The crude product was purified by column chromatography on silica
gel (column 48.0 × 2.5 cm2, eluent ethyl acetate-hexane 1:5), and a
fraction with Rf 0.44 afforded 0.58 g of colorless oil, which was
crystallized from pentane to give 0.47 g (38%) of 30 as colorless prisms,
which were decanted, combined together, washed with water, and dried
over Na2SO4. After concentration in vacuo the residue was purified by
column chromatography (silica gel, column 21.0 × 2.5 cm2, eluent
1
pentane) to afford 0.84 g (76%) of the product as a colorless oil. H
NMR (CDCl3) δ 7.37-7.29 (m, 10H), 5.46 (d, J ) 158.4 Hz).
1-13C-2,2-Diphenylethanol (7). A solution of 0.67 g (3.7 mmol) of
2-13C-1,1-diphenylethylene (6) in 10.0 mL of THF was treated with a
solution of BH3-THF complex (7.4 mL of 1.0 M solution in THF, 7.4
mmol) at ice-bath temperature. The reaction mixture was stirred for 4
h at room temperature, cooled in a water-ice bath, and carefully treated
with 2.0 mL of a 10% solution of water in THF followed by addition
of 12.0 mL of a 3.0 M aqueous solution of NaOH (37.0 mmol) and
10.0 mL of 30% H2O2 (96.2 mmol). The resulting mixture was stirred
for 1.5 h at room temperature, poured into water, acidified with 10%
HCl, and extracted with ether. The organic fraction was washed with
water and dried over Na2SO4. Removal of solvents in vacuo afforded
a colorless oil, which was crystallized from hexane to give 0.57 g (74%)
of the product as colorless cotton-like crystals, mp 51-52 °C. 1H NMR
(CDCl3) δ 7.39-7.15 (m, 10H), 4.22 (dd, J1 ) 14.1 Hz, J2 ) 6.9 Hz,
1H), 4.18 (dt, J1 ) 143.7 Hz, J2 ) 7.2 Hz, 2H), 1.51-1.42 (m, 1H).
1-13C-2,2-Diphenylacetaldehyde (8). A solution of 0.23 g (1.8
mmol) of oxalyl chloride in 5.0 mL of methylene chloride cooled to
-70 °C was treated with a solution of 0.29 g (3.7 mmol) of DMSO in
1.0 mL of CH2Cl2.31 The reaction mixture was stirred for 2 min, and
a solution of 0.33 g (1.7 mmol) of 1-13C-2,2-diphenylethanol (7) in
1.0 mL of CH2Cl2 was added via syringe. The resulting mixture was
stirred for 20 min, and triethylamine (0.84 g, 8.3 mmol) was added
dropwise followed by the stirring at -70 °C for an additional 10 min.
Then the reaction mixture was allowed to warm slowly and stirred for
40 min at room temperature. It was diluted with CH2Cl2, washed
successively with 3% HCl, water, saturated aqueous NaHCO3, and
water, dried over Na2SO4, and concentrated in vacuo. The crude product
was purified by column chromatography (silica gel, column 35.0 ×
1.5 cm2, eluent ethyl acetate-hexane 1:5), the fraction with Rf 0.55
was collected, which afforded 0.25 g (76%) of the product as a colorless
1
oil. H NMR (CDCl3) δ 9.89 (dd, J1 ) 178.5 Hz, J2 ) 2.6 Hz, 1H),
7.42-7.19 (m, 10H), 4.89 (dd, J1 ) 7.8 Hz, J2 ) 2.6 Hz, 1H).
3-13C-2-Methyl-4,4-diphenylcyclohex-2-enone (13C-9). A solution
of KOH (0.15 mL of a 1.0 M solution in EtOH, 0.15 mmol) was added
dropwise at 0 °C to the stirred solution of 0.17 g (0.86 mmol) of 1-13C-
2,2-diphenylacetaldehyde (8) and 80 mg (0.95 mmol) of ethyl vinyl
ketone in 10.0 mL of anhydrous ether. The reaction mixture was stirred
at 0 °C for 2 h followed by stirring at room temperature for 12 h. It
was successively washed with 10% HCl, water, a concentrated solution
of NaHCO3, and water and dried over Na2SO4. After being concentrated
in vacuo the oily residue was dissolved in 5.0 mL of ethanol, 1.0 mL
of concentrated HCl was added, and the resulting mixture was stirred
for 30 min at 60 °C. Then it was allowed to cool to room temperature,
poured into water, extracted with ether, washed with concentrated
NaHCO3 and water, dried over Na2SO4, and concentrated in vacuo to
afford a yellowish solid. This was recrystallized from methanol to give
0.15 g (65%) of colorless crystals, mp 116-118 °C. 1H NMR (CDCl3)
δ 7.37 (m, 10.5 H), 6.80 (br. s, right part of d13CH doublet, 0.5H),
2.72-2.63 (m, 2H), 2.41 (dd, J1 ) 7.0 Hz, J2 ) 5.5 Hz, 2H), 1.92 (dd,
J1 ) 6.0 Hz, J2 ) 1.5 Hz, 3H). HRMS m/e 263.1383 (calcd for 13C
C18H18O 263.1391).
1
mp 102-104 °C. H NMR (CDCl3) δ 7.30-7.12 (m, 10H), 5.95 (t,
J ) 1.2 Hz, 1H), 3.19 (s, 2H), 3.09 (s, 2H), 1.05 (s, 9H). UV (benzene),
nm: 334 (ꢀ ) 45). HRMS m/e 304.1813 (calcd for C22H24O 304.1827).
2-13C-1,1-Diphenylethylene (6). The Wittig reaction was performed
following the general procedure of Corey et al.30 DMSO (10.0 mL)
was added to sodium hydride (0.27 g of 60% suspension in mineral
oil, 6.6 mmol) previously washed with pentane to remove mineral oil,
and the resulting mixture was stirred at 50 °C to give a clear solution
(ca. 1 h). This was cooled in an ice-water bath, and a solution of 2.62
g (6.47 mmol) of 13C-methyltriphenylphosphonium iodide in 15.0 mL
of DMSO was added dropwise. The resulting dark-yellow solution was
stirred at room temperature for 30 min followed by addition of a solution
of 1.18 g (6.47 mmol) of benzophenone in 5.0 mL of DMSO. After
the addition was complete, the reaction mixture was stirred for 24 h at
room temperature. Then it was poured into 100 mL of water, the mixture
of solid and liquid was shaken with 5 × 50 mL portions of pentane,
trans-5,6-Diphenylbicyclo[3.1.0]hexan-2-one (10). A solution of
4.51 g (18.2 mmol) of 4,4-diphenylcyclohex-2-enone28 (1) in 1200 mL
of benzene was degassed with nitrogen for 1 h and photolyzed in a
water-cooled photochemical reactor equipped with a 400 W medium-
pressure Hg lamp with the CuSO4 filter solution for 3 h. Concentration
in vacuo afforded yellowish oil that was chromatographed on silica
gel (column 45.0 × 4.0 cm2, ethyl acetate-hexane 1:5) to give the
(28) Zimmerman, H. E.; Samuelson, G. E. J. Am. Chem. Soc. 1969, 91, 5307-
5318.
(29) Reich, H. J.; Renga, J. M.; Reich, I. L. J. Am. Chem. Soc. 1975, 97, 5434-
5447.
(31) Oxidation was performed according to the procedure given in the
following: Mancuso, A. J.; Huang, S.-L.; Swern, D. J. Org. Chem. 1978,
43, 2480-2482.
(30) Greenwald, R.; Chaykovsky, M.; Corey, E. J. J. Org. Chem. 1963, 28,
1128-1129.
9
J. AM. CHEM. SOC. VOL. 124, NO. 11, 2002 2827

A R T I C L E S
Zimmerman and Nesterov
Table 9. Amounts of Photoproducts 16 and 17 vs Total
following fractions: (a) 2.4 g (60%) of trans-5,6-diphenylbicyclo[3.1.0]-
hexan-2-one (10) as colorless solid, Rf 0.46, which was used for the
next step without further purification (the analytical sample was
recrystallized from hexane-ether mixture to give colorless prisms, mp
Conversion in the Solid-State Reaction of Enone 9a
conversion
0.05
0.06 0.075 0.08 0.10 0.11 0.17 0.29 0.39
1
72-74 °C (lit.32 mp 73-74 °C). H NMR (CDCl3) δ 7.45-7.20 (m,
product 16 0.025 0.03 0.038 0.04 0.05 0.07 0.11 0.25 0.37
product 17 0.025 0.03 0.038 0.04 0.05 0.04 0.06 0.04 0.02
10H), 3.12 (d, J ) 9.6 Hz, 1H), 2.68 (d, J ) 9.6 Hz, 1H), 2.44-2.36
(m, 2H), 2.15-2.01 (m, 1H), 1.21-1.06 (m, 1H)); (b) 0.44 g of starting
enone as colorless oil solidifying upon standing, Rf 0.38; and (c) 1.0 g
(25%) of cis-5,6-diphenylbicyclo[3.1.0]hexan-2-one as colorless solid,
Rf 0.31, with all spectral data corresponding to the literature32 values.
2,trans-5,6-Triphenylbicyclo[3.1.0]hexan-2-ol (11). A solution of
phenylmagnesium bromide was prepared from 3.38 g (21.5 mmol) of
bromobenzene and 0.53 g (22.0 mmol) of Mg in 35.0 mL of ether. A
solution of 3.54 g (14.3 mmol) of trans-5,6-diphenylbicyclo[3.1.0]-
hexan-2-one (10) in 50.0 mL of ether was added dropwise to the
solution of the Grignard reagent at room temperature. After the addition
was complete, the reaction mixture was stirred for 12 h at room
temperature followed by quenching with saturated aqueous NH4Cl and
extracting with ether. The combined organic fractions were washed
with water and concentrated NaCl solution and dried over Na2SO4.
After concentration in vacuo the crude product was purified by column
chromatography on silica gel (column 45.0 × 4.0 cm2, ethyl acetate-
hexane 1:4) followed by crystallization from hexane to afford 3.5 g
(75%) of colorless crystals, mp 94-96 °C. 1H NMR (CDCl3) δ 7.67-
7.59 (m, 2H), 7.51-7.44 (m, 2H), 7.42-7.20 (m, 11H), 2.72 (d, J )
8.7 Hz, 1H), 2.51 (d, J ) 8.7 Hz, 1H), 2.50 (dd, J1 ) 13.2 Hz, J2 )
7.8 Hz, 1H), 2.35-2.18 (m, 1H), 2.15 (s, 1H), 1.83 (dd, J1 ) 13.2 Hz,
J2 ) 7.8 Hz, 1H), 1.18-1.05 (m, 1H). HRMS m/e for [M - H2O]
308.1565 (calcd for C24H20 308.1568).
conversion
0.53
0.63
0.70
0.79
0.91
1.00
product 16
product 17
0.50
0.03
0.61
0.02
0.69
0.01
0.79
0
0.91
0
1.00
0
a All the amounts are in fractions of unity.
(2.46 g, 61.6 mmol) and 16.5 mL (18.2 g, 0.16 mol) of 30% H2O2.
The resulting mixture was stirred for 1.5 h at room temperature, poured
into 10% aqueous HCl, and extracted with a sufficient amount of ether
(2 × 250 mL). The organic fraction was washed with water and brine
and dried over Na2SO4. Solvents were concentrated in vacuo to a
volume of approximately 5.0 mL to afford a white precipitate that was
filtered. The clear filtrate was concentrated in vacuo to give a yellow
oil, which was chromatographed on silica gel (column 50.0 × 2.5 cm2,
ethyl acetate-hexane 1:5). The fraction with Rf 0.28 was collected and
afforded 0.38 g (38%) of alcohol 13 as a colorless oil. 1H NMR (CDCl3)
δ 7.45-7.00 (m, 13H), 6.91-6.83 (m, 2H), 3.74 (dd, J1 ) 8.4 Hz,
J2 ) 4.4 Hz, 1H), 3.49-3.37 (m, 1H), 2.83 (dd, J1 ) 12.8 Hz,
J2 ) 6.9 Hz, 1H), 2.57 (dd, J1 ) 8.7 Hz, J2 ) 4.4 Hz, 1H), 2.50 (d,
J ) 8.7 Hz, 1H), 2.36 (dd, J1 ) 12.8, J2 ) 8.1 Hz, 1H), 1.55 (br. s,
1H). 13C NMR (CDCl3) δ 145.13, 141.12, 136.82, 130.21, 128.48,
128.06, 127.93, 127.34, 126.43, 126.37, 126.18, 125.93, 72.63, 51.66,
41.10, 33.91, 33.88, 32.18. IR (neat) cm-1: 3550-3150. The white
precipitate previously filtered was recrystallized from methanol-THF
mixture (2:1) to give 0.27 g (25%) of 2,3,5-triphenylcyclohexan-1,4-
diol (35) as colorless microcrystals, mp 259-261 °C. 1H NMR (CD3-
OD) δ 7.50-6.90 (m, 15H), 4.71 (t, J ) 11.1 Hz, 1H), 4.03 (narrow
m, 1H), 3.71 (dd, J1 ) 10.8 Hz, J2 ) 5.4 Hz, 1H), 3.45-3.20 (partially
overlaping with an OH signal from the solvent, m, 2H + possibly 2H
from OH), 2.46 (td, J1 ) 14.7 Hz, J2 ) 2.7 Hz, 1H), 1.99 (dm, J )
14.7 Hz, 1H). HRMS m/e for [M - 2H2O - H2] 306.1410 (calcd for
C24H18 306.1410).
1,exo-4,endo-6-Triphenylbicyclo[3.1.0]hexan-3-one (14). To a
solution of 0.27 g (0.72 mmol) of Dess-Martin periodinane33 in 25.0
mL of CH2Cl2 was added pyridine (0.047 g, 0.60 mmol), and the
resulting mixture was stirred for 10 min at room temperature. A solution
of 100 mg (0.31 mmol) of alcohol 23 in 5.0 mL of CH2Cl2 was added
dropwise, and the resulting mixture was stirred for 6 h at room
temperature. Then it was cooled to 0 °C, and a mixture of 20.0 mL of
a 5% aqueous solution of Na2S2O3 and 12.0 mL of a concentrated
aqueous solution of NaHCO3 was added followed by stirring for an
additional 45 min. The resulting solution was diluted with ether, the
organic layer separated, and the aqueous layer extracted with ether.
The combined organic fractions were washed with water and dried over
Na2SO4. Concentration in vacuo afforded a yellowish oil. This was
treated with sodium ethoxide solution (prepared from 0.077 g (3.35
mmol) of Na and 4.0 mL of EtOH) for 8 h at room temperature. The
reaction mixture was diluted with ether, washed with water, dried over
Na2SO4, and concentrated in vacuo to afford a dark oil, which was
purified by preparative plate TLC (silica gel, eluent ethyl acetate-
hexane 1:5). The Rf 0.39 band gave 0.06 g (60%) of 1,exo-4,endo-6-
triphenylbicyclo[3.1.0]hexan-3-one (14) as a colorless oil. All the
spectral characteristics were in agreement with those of the photoproduct
14 obtained in a solid-state preparative photolysis of enone 18.
Exploratory Solid-State Photolysis of Enone 9. Summarized results
of some separate runs are presented in Table 9.
trans-2,5,6-Triphenylbicyclo[3.1.0]hex-2-ene (12). A solution of
2.5 g (7.7 mmol) of 2,trans-5,6-triphenylbicyclo[3.1.0]hexan-2-ol (11)
in a mixture of 10.0 mL of pyridine and 30.0 mL of methylene chloride
was cooled in an ice-water bath, and a solution of 1.37 g (0.84 mL,
11.5 mmol) of SOCl2 in 4.0 mL of methylene chloride was added
dropwise. The reaction mixture was stirred at 0 °C for 3 h. The mixture
was then gradually allowed to rise to room temperature, and stirred
for an additional 12 h. Then it was diluted with methylene chloride,
washed with a 5% solution of HCl and water, and dried over Na2SO4.
After concentration in vacuo, the residue was subjected to column
chromatography on silica gel (column 45.0 × 4.0 cm2, ethyl acetate-
hexane 1:6), and the fraction with Rf 0.54 was collected, which afforded
a white solid. Recrystallization from hexane gave 1.1 g (47%) of
colorless cotton-like crystals, mp 103-105 °C. 1H NMR analysis
showed the product to be mixed crystals of trans-2,5,6-triphenylbicyclo-
[3.1.0]hex-2-ene (12) and 1,3,4-triphenylcyclohex-1,4-diene (15) in
1
approximate ratio 1:1. H NMR (CDCl3) δ 7.65-7.59 (m, 2H of 12),
7.46-7.05 (m, 13H of 12 + 15H of 15), 6.36 (tm, J ) 4.0 Hz, 1H of
15), 6.28 (dt, J1 ) 4.4 Hz, J2 ) 1.5 Hz, 1H of 15), 5.65 (t, J ) 2.7 Hz,
1H of 12), 4.74-4.63 (m, 1H of 15), 3.47-3.38 (m, 2H of 15), 3.05
(dd, J1 ) 7.8 Hz, J2 ) 2.7 Hz, 1H of 12), 2.96 (dd, J1 ) 19.0 Hz,
J2 ) 2.7 Hz, 1H of 12), 2.82 (d, J ) 7.8 Hz, 1H of 12), 2.64 (dt, J1 )
19.0 Hz, J2 ) 2.7 Hz, 1H of 12). 13C NMR (CDCl3) δ 145.43, 143.83,
141.91, 140.93, 140.54, 137.41, 136.48, 135.62, 131.48, 130.47, 128.52,
128.50. 128.46, 128.31, 128.26, 128.07, 127.85, 127.13 (2C), 127.09,
126.79, 126.76, 126.28, 126.23, 126.19, 125.94 (2C), 125.12, 123.92,
123.52, 45.88, 39.28, 37.97, 37.27, 35.17, 29.54.
1,endo-4,endo-6-Triphenylbicyclo[3.1.0]hexan-exo-3-ol (13). A
solution of 0.95 g (3.1 mmol) of mixed crystals obtained in the previous
step in 15.0 mL of anhydrous THF was treated with a solution of BH3-
THF complex (10.8 mL of 1.0 M solution in THF, 10.8 mmol) at ice-
bath temperature followed by stirring at this temperature for 2.5 h, and
for an additional 2.5 h at room temperature. The resulting mixture was
cooled to 0 °C and carefully treated with 3.0 mL of a 10% solution of
water in THF followed by addition of 20.5 mL of 3.0 M aqueous NaOH
(32) Zimmerman, H. E.; Albrecht, F. X.; Haire, M. J. J. Am. Chem. Soc. 1975,
97, 3726-3740.
(33) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4155-4156.
9
2828 J. AM. CHEM. SOC. VOL. 124, NO. 11, 2002

Reactivity in Stages of Crystal Lattices
A R T I C L E S
Table 10. Amounts of Photoproducts 17 and 18 vs Total Conversion in the Solid-State Reaction of Enone 13a
conversion
0.01
0.03
0.04
0.065
0.07
0.11
0.145
0.21
0.24
0.30
0.31
0.40
0.42
product 14
product 21
0
0.01
0
0.03
0
0.04
0
0
0.07
0
0.11
0.043
0.102
0.09
0.12
0.10
0.14
0.15
0.15
0.16
0.15
0.23
0.17
0.24
0.18
0.065
conversion
0.47
0.51
0.57
0.62
0.66
0.66
0.77
product 14
product 21
0.29
0.18
0.32
0.19
0.38
0.19
0.42
0.20
0.47
0.19
0.46
0.20
0.58
0.19
a All the amounts are in fractions of unity.
Exploratory Solid-State Photolysis of 4,5,5-Triphenylcyclohex-
2-enone (18). Summarized results of some separate runs are presented
in Table 10.
19.2 Hz, 1H), 1.92 (dt, J1 ) 6.4 Hz, J2 ) 1.9 Hz, 1H), 1.11 (d, J ) 6.4
Hz, 1H), 0.67 (s, 9H); HRMS (FAB with Na+) m/e 327.1714 (calcd
for C22H24ONa 327.1725). H NMR for 23 (CDCl3) δ 7.60-7.00 (m,
1
Preparative Solid-State Photolysis of 4,5,5-Triphenylcyclohex-
2-enone (18). A portion of 1.0 g (3.09 mmol) of crystalline enone 18
was irradiated between two quartz plates (diameter 14 cm) for 100 h.
A part (0.7 g) of the reaction mixture was separated by column
chromatography on silica gel (column 50.0 × 2.5 cm2, eluent ethyl
acetate-hexane 1:8) to give the following fractions: (a) 25 mg of
unidentified yellow oil, Rf 0.60; (b) 95 mg (14%) of exo-2,trans-5,6-
10H), 3.62 (s, 1H), 3.39 (dm, J ) 18.3 Hz, 1H), 2.73 (d, J ) 18.3 Hz,
1H), 1.87 (dd, J1 ) 6.3 Hz, J2 ) 3.3 Hz, 1H), 0.62 (s, 9H), 0.41 (d, J
) 6.3 Hz, 1H); HRMS (FAB with Na+) m/e 327.1718 (calcd for C22H24-
ONa 327.1725); and (d) 15 mg (25%) of starting enone as a colorless
oil, Rf 0.33.
Exploratory Sensitized Photolysis of 3-tert-Butyl-5,5-diphenyl-
cyclohex-2-enone (4). A solution of 6 mg (0.02 mmol) of enone 4 and
30 mg (0.25 mmol) of acetophenone in 2.5 mL of C6D6 was irradiated
in a NMR tube with a Pyrex filter for 5 h. During the photolysis nitrogen
1
triphenylbicyclo[3.1.0]hexan-3-one (14) as colorless oil, Rf 0.42. H
NMR (CDCl3) δ 7.50-6.90 (m, 15H), 3.70 (s, 1H), 3.06 (d, J ) 18.7
Hz, 1H), 2.88 (d, J ) 8.6 Hz, 1H), 2.80 (d, J ) 18.7 Hz, 1H), 2.66 (d,
J ) 8.6 Hz, 1H). 13C NMR (CDCl3) δ 212.72, 143.29, 138.30, 133.35,
131.07, 128.96, 128.92, 128.84, 127.37, 127.29, 127.22, 126.50, 125.86,
55.59, 41.53, 35.41, 32.52, 31.96. IR (neat) cm-1: 1745.8. HRMS m/e
324.1525 (calcd for C24H20O 324.1514); (c) 100 mg (14%) of 4,4,endo-
6-triphenylbicyclo[3.1.0]hexan-2-one (21) as a colorless solid, Rf 0.33,
whcih was recrystallized from an ethyl acetate-hexane mixture to give
1
was passed through the solution. H NMR analysis showed the ratio
of two products 22 and 23 to be 1:1 at the conversion 70%.
Preparative Solid-State Photolysis of 3-tert-Butyl-5,5-diphenyl-
cyclohex-2-enone (4). Two portions (25 mg each) of crystalline 4 were
irradiated for 2 h at 0 °C. The combined portions were subjected to
column chromatography on silica gel (column 35.0 × 1.5 cm2, eluent
ethyl acetate-hexane 1:6) to give the following fractions: (a) 28 mg
(56%) of 1-tert-butyl-endo-2,5-diphenylbicyclo[3.1.0]hexan-3-one (22),
Rf 0.49; and (b) 18 mg (36%) of the starting material, Rf 0.33.
1
colorless prisms, mp 161-162 °C (lit.34 mp 161-162 °C). H NMR
(CDCl3) δ 7.40-7.10 (m, 10H), 7.6 (t, J ) 6.9 Hz, 1H), 6.95 (t, J )
6.9 Hz, 2H), 6.70 (d, J ) 6.9 Hz, 2H), 3.18 (dd, J1 ) 8.6 Hz, J2 ) 5.4
Hz, 1H), 3.02 (t, J ) 8.6 Hz, 1H), 2.73 (dd, J1 ) 8.6 Hz, J2 ) 5.4 Hz,
1H), 2.41 (d, J ) 17.5 Hz), 2.17 (d, J ) 17.5 Hz); and (d) 420 mg
(60%) of starting material as a colorless solid, Rf 0.28.
Preparative Solid-State Photolysis of 3-13C-2-Methyl-4,4-diphen-
ylcyclohex-2-enone (13C-9). A 50 mg (0.19 mmol) portion of crystalline
13C-labeled enone 13C-9 was irradiated for 29 h. The crystals after
irradiation were subjected to column chromatography on silica gel
(column 35.0 × 0.9 cm2, eluent ethyl acetate-hexane 1:5) to afford
the following fractions: (a) 45 mg of a mixture of starting material
and bicyclic product 13C-16; and (b) 3 mg (6%) of 2-methyl-3-diphenyl-
13C-methylcyclopent-2-enone (13C-17) as a colorless oil, Rf 0.28, which
was crystallized from pentane to afford colorless crystals, mp 64-67
Preparative Solution Photolysis of 3-tert-Butyl-5,5-diphenylcy-
clohex-2-enone (4). A solution of 60 mg (0.20 mmol) of enone 4 in
40.0 mL of benzene was irradiated in a Pyrex tube placed next to a
water-cooled medium-pressure mercury lamp equipped with 2 mm
Pyrex filter for 12 h. Nitrogen was passed through the solution during
the photolysis. After rotary evaporation of the solvent, the oily residue
was separated by column chromatography on silica gel (column
35.0 × 1.5 cm2, eluent ethyl acetate-hexane 1:6) to give the following
fractions: (a) 5 mg (8%) of unidentified product as a colorless oil, Rf
0.89; (b) 12 mg (20%) of 3-tert-butyl-2,5-diphenylcyclohex-2-enone
(24) as a colorless oil solidifying upon standing, Rf 0.62, which was
recrystallized from hexane to give colorless thin needles, mp 141-
1
°C. H NMR (CDCl3) δ 7.38-7.11 (m, 10H), 5.44 (d, J ) 126.3 Hz,
1H), 2.49-2.38 (m, 4H), 1.69 (s, 3H). HRMS m/e 263.1374 (calcd for
13CC18H18O 263.1391).
Fraction 1 was additionally separated by preparative TLC on a plate,
eluted with ethyl acetate-hexane 1:5 to afford the following frac-
tions: (a) 5 mg (10%) of 6-13C-1-methyl-trans-5,6-diphenylbicyclo-
[3.1.0]hexan-2-one (13C-16) as a colorless solid, Rf 0.54, which was
recrystallized from methanol to afford colorless crystals, mp 139-141
1
143 °C. H NMR (CDCl3) δ 7.49-7.19 (m, 9H), 7.01 (s, 1H), 3.88
(tdd, J1 ) 12.2 Hz, J2 ) 5.4 Hz, J3 ) 1.8 Hz, 1H), 3.25 (t, J ) 12.2
Hz, 1H), 2.91 (dd, J1 ) 12.2 Hz, J2 ) 5.4 Hz, 1H), 2.49 (d, J ) 12.2
Hz, 1H), 2.38 (t, J ) 12.2 Hz, 1H), 0.90 (s, 9H); HRMS m/e 304.1812
(calcd for C22H24O 304.1827); (c) 21 mg (35%) of a mixture of bicyclic
ketones 22 and 23 as a colorless oil, solidifying upon standing, Rf 0.49.
Fractional crystallization from methanol afforded 10 mg of 1-tert-butyl-
endo-2,5-diphenylbicyclo[3.1.0]hexan-3-one (22) which was addition-
ally recrystallized from hexane to give colorless prisms, mp 171-173
°C. A structure of the product 22 was confirmed by a single-crystal
X-ray analysis. Evaporation of the mother liquor gave 8 mg of 1-tert-
butyl-exo-2,5-diphenylbicyclo[3.1.0]hexan-3-one (23) as a colorless oil.
1H NMR for 22 (CDCl3) δ 7.50-7.44 (m, 2H), 7.39-7.21 (m, 8H),
4.34 (s, 1H), 3.15 (dt, J1 ) 19.2 Hz, J2 ) 1.9 Hz, 1H), 2.73 (d, J )
1
°C. H NMR (CDCl3) δ 7.43-7.21 (m, 10H), 3.00 (d, J ) 157.2 Hz,
1H), 2.34-2.22 (m, 2H), 2.06 (dt, J1 ) 19.2 Hz, J2 ) 5.1 Hz, 1H),
1.22 (d, J ) 3.6 Hz, 3H), 1.10 (dt, J1 ) 19.2 Hz, J2 ) 8.7 Hz, 1H).
HRMS m/e 263.1380 (calcd for 13CC18H18O 263.1391); and (b) 33 mg
of starting material as a colorless solid, Rf 0.46.
Conversion Dependence of Solid-State Quantum Yield. A method
used by Zimmerman et al.4b was used to determine solid-state quantum
yield. Ten to twenty milligram portions of fine crystalline powder of
enone were used in each run. The photoproducts ratio was analyzed
1
by H NMR. The data obtained are summarized in Table 11.
Computational Procedures. Ab initio, DFT, and combined QM/
MM computations were performed on an Intel Pentium III based multi-
processor cluster with Linux OS using the Gaussian 9811 computational
package. Molecular mechanics computations were performed with
(34) Zimmerman, H. E.; O’Brien, M. E. J. Org. Chem. 1994, 59, 1809-1816.
9
J. AM. CHEM. SOC. VOL. 124, NO. 11, 2002 2829

A R T I C L E S
Zimmerman and Nesterov
Table 11. Conversion Dependence of Solid-State Quantum Yield for Enone 18a
conversion
3.9%
1.5%
3.8%
5.2%
12.0%
Φ of product 14
Φ of product 21
Φ total
4.73 × 10-4
5.07 × 10-4
2.40 × 10-4
1.43 × 10-4
1.16 × 10-4
4.79 × 10-5
1.64 × 10-4
4.73 × 10-4
5.07 × 10-4
2.40 × 10-4
1.43 × 10-4
a Irradiation at 320 nm.
MacroModel 6.535 using the MM3* force field. Minicrystal lattices were
generated based on X-ray data for reactant using Smartpack program.5b
General Procedure for Computational Treatment of the Mini-
Crystal Lattice. A central molecule was chosen, and a first shell of
closest surrounding molecules was defined (all the other more remote
molecules were considered as a second shell). Conversion of the central
molecule to a reacting intermediate was performed by creating new
bonds to be formed and deleting bonds to be broken followed by MM
optimization in frozen shells. This was followed by MM optimization
of the reacting intermediate together with the first shell inside the frozen
second shell (a force constant of 30 kJ/Å was chosen for optimization
of the first shell). In some cases it was necessary to use additional
constraints to prevent unrealistically strong bending of migrating phenyl
group in the central reacting molecule. The optimized central molecule
and first shell were extracted and further optimized by ONIOM. In
this computation the central molecule was subjected to ab initio
optimization, and the first shell was treated by MM (using the Dreiding
force field) and kept frozen throughout the entire optimization.
Computational Details. Optimization of intermediates 16Tcr and
25Tcr: shell 1-14 molecules, shell 2-87 molecules.
in each case, i.e., a central part included 2 molecules to be optimized
(for ONIOM treatment one molecule (16Tcr′ or 25Tcr′) was optimized
by ab initio, and one molecule of 17 by MM), shell 1-20 molecules,
and shell 2-80 molecules.
Optimization of intermediates 27T and 30T: shell 1-14 molecules,
shell 2-60 molecules.
Optimization of intermediates 27T and 30T for a second stage of
the solid-state reaction was performed together with 2 neighboring
molecules of photoproduct 14, i.e., a central part included 3 molecules
to be optimized (in ONIOM treatment one molecule (27T or 30T) was
optimized by ab initio, and two molecules of 14 by MM), shell 1-26
molecules (20 molecules for ONIOM optimization), shell 2-84
molecules.
Acknowledgment. Support of this research by the National
Science Foundation is gratefully acknowledged with special
appreciation for its support of basic research.
Supporting Information Available: X-ray coordinates of
compounds 4 and 22 (PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
Optimization of intermediates 16Tcr′ and 25Tcr′ was performed
together with one molecule of neighboring cyclopentenone product 17
(35) Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Liskamp, R.; Lipton, M.;
Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J. Comput. Chem.
1990, 11, 440-467.
JA0124562
9
2830 J. AM. CHEM. SOC. VOL. 124, NO. 11, 2002