per unit cell.12 The racemic compound crystallizes in the
space group P21/c with four molecules per unit cell and no
solvent of crystallization.13 The absolute configuration of
(+)-(2R,4S)-1 was assigned from the known configuration
of (+)-(R)-3b, which is determined by the enantioselectivity
of the enzymatic hydrolysis.9
As illustrated in Figure 2, the structure of ketone (2R,4S)-1
is very similar in the two enantiomorphous (i.e., enantio-
merically pure and racemic) crystal forms.7 It is characterized
by a conformation where the 2-methyl and 4-cyano substit-
uents are approximately eclipsed with the ketone carbonyl
and the phenyl groups adopt a staggered orientation. The
main difference between the two molecular structures is the
orientation of the carboxylate ester groups. The ester carbonyl
points toward the cyano substituent in the racemic crystals
and in the opposite direction in the optically pure modifica-
tion.14
Assuming that double inversion in the crystal lattice is
extremely unlikely, the optical purity of (+)-(2R,3R)-2 is
determined by the optical purity of crystalline (+)-(2R,4S)-
1, which is expected to be 100% in its enantiomorphous
phase. An [R]D ) 128° (c 1, EtOAc) determined for
(+)-(2R,3R)-2 is of the same sign as those of the ketone
precursor (+)-(2R,4S)-1 and the acid ester (+)-3b. A >95%
de determined by GLC for optically active and racemic
samples (Figure 1, top trace) indicates a very small amount
of inversion at either one of the two stereogenic centers. A
small stereochemical loss may be due to reaction occurring
at the crystal surface or at defect sites. X-ray structural
elucidation of (+)-(2R,3R)-2 (mp ) 78-79 °C)15 and
(()-(2RS,3RS)-2 (mp ) 99-101 °C)16 confirmed the ex-
pected stereospecificity of the bond-forming reaction. The
simple visualization of the reaction illustrated in Scheme 1
is borne out from the X-ray structures shown in Figure 2.
The spatial orientation of the substituents at C2 and C4 in
ketone 1 is maintained at stereocenters C2 and C3 in product
2. Notably, the orientation of the ester group in crystals of
(+)-(2R,4S)-1 is different than that observed in crystals of
(+)-(2R,3R)-2 (Figure 2, top). However, for reactions in
crystals, it is well-known that changes in size and shape must
be relatively small.17 Therefore, it is likely that (+)-(2R,3R)-2
is originally formed with the ester carbonyl in a conformation
similar to that of the starting material and that it only changes
after the reacted sample is dissolved and recrystallized. Very
small conformational differences can be appreciated for the
structure of compound 2 in its two crystal forms.
Regarding the reaction mechanism, it is well-known18 that
decarbonylation proceeds in a stepwise manner by sequential
R-cleavage and decarbonylation.2 With experimental1,2a,3 and
computational2b evidence gathered in our group, we have
suggested some general guidelines to “engineer reactions in
crystals”.1-3 Our observations indicate that photodecarbo-
nylation in the solid state is enabled by substituents that lower
the bond dissociation energy of the two R-bonds by more
than ca. 12 kcal/mol as compared to that of acetone (BDE-
[Me-COMe] ) 81.1 kcal/mol).19 This suggestion arises from
a known relation between reaction efficiencies and heats of
reaction first proposed by Fisher and Paul.20 R-Cleavage and
decarbonylation in ketone 1 are enabled by phenyl, car-
bomethoxy, and cyano groups. On the basis of extensive
literature data,19 one may expect that phenyl, ester, and
methyl groups at C2 will lower the R-bond dissociation
energy by ca. 25 kcal/mol with respect to that of acetone.
Similarly, phenyl and cyano substituents at C4 are expected
to weaken the other R-bond by ca. 23 kcal/mol, which is
also well in excess of the suggested value. Not surprisingly,
compound 1 reacts in crystals with apparent efficiencies that
compete with those observed in solution reactions.
Knowing that accumulation of the product depresses the
melting point of the pure starting material, it is advantageous
to document the nature of the (ideally) two-component
system so that a suitable photolysis temperature is selected.21
Differential scanning calorimetric (DSC) analyses of
(()-(2RS,4SR)-1 before reaction and after ca. 45% conver-
sion indicate that a solid phase is preserved below ca. 30-
40 °C and that reaction proceeds by a phase separation
mechanism (Figure 3).21 The melting transition of pure
(()-(2RS,4SR)-1 at 97 °C broadens and shifts to lower
temperatures as the product accumulates in crystals of the
reactant phase. That phase separation occurs in this reaction
is deduced by the observation of a broad peak at about 60
°C, which corresponds to the formation of a eutectic mixture.
This result suggests that high reaction yields with maintained
selectivity may be attained by controlling the temperature
(12) (+)-(2R,4S)-1: C19H17NO3, MW ) 307.34, space group P65, a )
19.154(3) Å, Å, b ) 19.154(3) Å, c ) 8.393(2) Å, R ) â ) 90°, γ )
120°, V ) 2666.5(9) Å3, Z ) 6, Fcalcd ) 1.156 Mg/m3, F(000) ) 978, λ )
0.71073 Å, µ(Mo KR) ) 0.080 mm-1, T ) 100(2) K, crystal size ) 0.60
× 0.10 × 0.10 mm3. Of the 15 993 reflections collected (2.72° e θ e
28.29°), 4062 [R(int) ) 0.0297] were independent reflections; max/min
residual electron density 357 and -176 e nm-3, R1 ) 0.0369 [I > 2σ(I)]
and wR2 ) 0.0874 (all data).
(13) (()-(2R,4S)-1: C19H17NO3, MW ) 307.34, space group P21/c, a
) 12.238(3) Å, b ) 9.468(2) Å, c ) 13.810(4) Å, â ) 97.769(5)°, V )
1585.6(7) Å3, Z ) 4, Fcalcd ) 1.287 Mg/m3, F(000) ) 648, λ ) 0.71073
Å, µ(Mo KR) ) 0.087 mm-1, T ) 100(2) K, crystal size ) 0.4 × 0.3 ×
0.3 mm3. Of the 9819 reflections collected (1.68° e θ e 28.26°), 3760
[R(int) ) 0.0298] were independent reflections; max/min residual electron
density 327 and -189 e nm-3, R1 ) 0.0371 [I > 2σ(I)] and wR2 ) 0.1030
(all data).
(14) Calculated (AM1) dipole moments are 5.92 and 3.53 D for structures
in racemic and optically active crystals, respectively. However, only the
chiral crystals are macroscopically polar.
(15) (+)-(2R,4S)-2: C18H17NO2, MW ) 279.33, space group P21, a )
8.4096(6) Å, b ) 9.4818(7) Å, c ) 9.3272(6) Å, â ) 96.9610(10)°, V )
738.25(9) Å3, Z ) 2, Fcalcd ) 1.257 Mg/m3, F(000) ) 296, λ ) 0.71073 Å,
µ(Mo KR) ) 0.082 mm-1, T ) 100(2) K, crystal size ) 0.50 × 0.40 ×
0.33 mm3. Of the 6629 reflections collected (2.20° e θ e 28.27°), 3454
[R(int) ) 0.0132] were independent reflections; max/min residual electron
density 272 and -150 e nm-3, R1 ) 0.0284 [I > 2σ(I)] and wR2 ) 0.0680
(all data).
(16) (()-(2RS,3RS)-2: C18H17NO2, MW ) 279.33, space group P21/n,
a ) 11.1764(16) Å, b ) 9.0080(13) Å, c ) 15.168(2) Å, â ) 98.844(2),
V ) 1508.9(4) Å3, Z ) 4, Fcalcd ) 1.230 Mg/m3, F(000) ) 592, λ ) 0.71073
Å, µ(Mo KR) ) 0.080 mm-1, T ) 100(2) K, crystal size ) 0.40 × 0.35 ×
0.25 mm3. Of the 9649 reflections collected (2.12° e θ e 28.30°), 3568
[R(int) ) 0.0170] were independent reflections; max/min residual electron
density 360 and -190 e nm-3, R1 ) 0.0345 [I > 2σ(I)] and wR2 ) 0.0875
(all data).
(17) Cohen, M. D. Angew. Chem., Int. Ed. Engl. 1975, 14, 386-393.
(18) Weiss, D. Org. Photochem; Padwa, A., Ed.; Marcel Deker: New
York, 1981; Vol. 5, pp 347-420.
(19) Luo, Y.-R. Handbook of Bond Dissociation Energies in Organic
Compounds; CRC Press: Boca Raton, FL, 2003.
(20) Fisher, H.; Paul, H. Acc. Chem. Res. 1987, 20, 200-206.
(21) Keating, A. E.; Garcia-Garibay, M. A. Molecular and Supramo-
lecular Photochemistry; Ramamurthy, V., Schanze, K., Eds.; Marcel
Dekker: New York, 1998; Vol. 2, pp 195-248.
Org. Lett., Vol. 5, No. 14, 2003
2533