Organic reactions in the solid state: Reactions of enclathrated 3,4- epoxycyclopentanone (= 6-oxobicyclo[3.1.0]hexan-3-one) in Tri-othymotide and absolute configuration of 4-hydroxy- and 4-chlorocyclopent-2-en-1-one

Article abstract of DOI:10.1002/(SICI)1522-2675(19990310)82:3<418::AID-HLCA418>3.0.CO;2-D

Several aspects of the heterogeneous actions of aqueous and gaseous HCl on the chemical behavior of 3,4-epoxycyclopentanone (=6- oxablcyclo[3.1.0]hexan-3-one; 1) included in the asymmetric cages of tri-o- thymotide (TOT) clathrates belonging to space groups P3₁21 are described, showing specific features strikingly at variance with those observed in liquid solutions. In a first step, the substrate underwent an acid-promoted allylic isomerization, as already observed in our previous investigations, to give optically active 4-hydroxycydopent-2-en-1-one (2). In a Consecutive step, a displacement of the OH group was accomplished by the Cl^- anion to afford the corresponding chloro compound 3. Polymorphism was encountered in the preparation of TOT/1 clathrates. Recrystallization of TOT in the pure guest 1 yielded micro-twinned crystals belonging to the P3₁ space group (host/guest ratio 1: 1), whereas the expected P3₁21 lattice grew from a mixture of TOT, 1 and MeOH. The structural determination of TOT/1 was carried out by X-ray diffraction (Fig. 1). Kinetic measurements were achieved that shed light on some striking features of this type of heterogeneous reactions for solid-liquid and solid-gas systems. Several reactions of pure clathrate antipodes (+)-TOT/1 with gaseous HCl were carried out under various conditions; concentration and enantiomer-excess(ee) determinations of the products 2 and 3 allowed to establish a larger ee for 3, thus demonstrating the influence of the host-guest diastereomeric association on the progression of the reaction. The correlation of optical activities of the host and products for the global reaction disclosed the sequence (+)-(M)-TOT/1 → (- )-2 → (-)-3. A new way for the preparation of 2 was devised. It was further demonstrated that the X-ray structure analysis of the chiral clathrate (M)- TOT/(+)-2 (Fig. 4) associated with chitoptical measurements was an efficient and straightforward method to determine the absolute (+)-(R)-configuration of the guest. The enantioselectivity of the TOT clathrate for 2 was established by two different methods which allowed the appraisal of an accurate revised value of the specific rotation of 2. The enclathration of 3 occurred exclusively in the orthorhombic centrosymmetric host lattice Pbca, thus prohibiting the X-ray structural determination of the guest absolute configuration. The problem of finding a pathway to the intended enantiomer enrichment of 3 was worked out through the action of aqueous HCI on microcrystalline (+)-TOT/(-)-(S)-2 that gave an optically active mixture of unreacted 2 with 3 as sole product. The pure optically active 3 was isolated by Subsequent TLC. The resolution of 3 was achieved by GC over a chiral column and its (unknown) specific rotation measured. The absolute configuration of 3 was established by the measurement of the enantiomer purity of the optically active mixture 3 obtained after the total conversion of (-)-(S)-2 in the presence of thionyl chloride in Et₂O, dioxane, and benzene. It was deduced that the (-)-3 enantiomer had the (S)-configuration.

Full text of DOI:10.1002/(SICI)1522-2675(19990310)82:3<418::AID-HLCA418>3.0.CO;2-D

418
Helvetica Chimica Acta ± Vol. 82 (1999)
Organic Reactions in the Solid State: Reactions of Enclathrated 3,4-
Epoxycyclopentanone (ˆ 6-Oxobicyclo[3.1.0]hexan-3-one) in Tri-o-
thymotide and Absolute Configuration of 4-Hydroxy- and 4-Chlorocyclopent-
2-en-1-one
1
by Raymond Gerdil*, Huiyou Liu ), and GeÂrald Bernardinelli
Department of Organic Chemistry and Laboratory of X-Ray Crystallography, University of Geneva, CH-1211
Geneva 4
Several aspects of the heterogeneous actions of aqueous and gaseous HCl on the chemical behavior of 3,4-
epoxycyclopentanone (ˆ 6-oxabicyclo[3.1.0]hexan-3-one; 1) included in the asymmetric cages of tri-o-
thymotide (TOT) clathrates belonging to space groups P3121 are described, showing specific features strikingly
at variance with those observed in liquid solutions. In a first step, the substrate underwent an acid-promoted
allylic isomerization, as already observed in our previous investigations, to give optically active 4-
hydroxycyclopent-2-en-1-one (2). In a consecutive step, a displacement of the OH group was accomplished
by the Cl anion to afford the corresponding chloro compound 3. Polymorphism was encountered in the
preparation of TOT/1 clathrates. Recrystallization of TOT in the pure guest 1 yielded micro-twinned crystals
belonging to the P31 space group (host/guest ratio 1 : 1), whereas the expected P3121 lattice grew from a mixture
of TOT, 1, and MeOH. The structural determination of TOT/1 was carried out by X-ray diffraction (Fig. 1).
Kinetic measurements were achieved that shed light on some striking features of this type of heterogeneous
reactions for solid-liquid and solid-gas systems. Several reactions of pure clathrate antipodes (‡)-TOT/1 with
gaseous HCl were carried out under various conditions; concentration and enantiomer-excess(ee) determi-
nations of the products 2 and 3 allowed to establish a larger ee for 3, thus demonstrating the influence of the
host-guest diastereomeric association on the progression of the reaction. The correlation of optical activities of
the host and products for the global reaction disclosed the sequence (‡)-(M)-TOT/1 ! ( )-2 ! ( )-3. A new
way for the preparation of 2 was devised. It was further demonstrated that the X-ray structure analysis of the
chiral clathrate (M)-TOT/(‡)-2 (Fig. 4) associated with chiroptical measurements was an efficient and
straightforward method to determine the absolute (‡)-(R)-configuration of the guest. The enantioselectivity of
the TOT clathrate for 2 was established by two different methods which allowed the appraisal of an accurate
revised value of the specific rotation of 2. The enclathration of 3 occurred exclusively in the orthorhombic
centrosymmetric host lattice Pbca, thus prohibiting the X-ray structural determination of the guest absolute
configuration. The problem of finding a pathway to the intended enantiomer enrichment of 3 was worked out
through the action of aqueous HCl on microcrystalline (‡)-TOT/( )-(S)-2 that gave an optically active mixture
of unreacted 2 with 3 as sole product. The pure optically active 3 was isolated by subsequent TLC. The resolution
of 3 was achieved by GC over a chiral column and its (unknown) specific rotation measured. The absolute
configuration of 3 was established by the measurement of the enantiomer purity of the optically active mixture
3 obtained after the total conversion of ( )-(S)-2 in the presence of thionyl chloride in Et2O, dioxane, and
benzene. It was deduced that the ( )-3 enantiomer had the (S)-configuration.
1. Introduction. ± Tri-o-thymotide (ˆ 1,7,13-trimethyl-4,10,16-tris(1-methylethyl)-
6H,12H,18H-tribenzo[b,f,j][1,5,9]trioxacyclododecin-6,12,18-trione; TOT; Fig. 1) has
become known as a versatile clathrate host crystallizing in 12 different space groups.
1
Á Â
The results presented hereafter are parts of the Ph.D. Thesis of H.L. (These No. 2702, Universite de
)
Á
Geneve).

Helvetica Chimica Acta ± Vol. 82 (1999)
419
The TOT clathrates that have aroused most interest are chiral and of enantiomorphous
space group P3121, where any single crystal displays either only (P) or only (M)
helicity for the host molecule. These types of clathrate crystallize under spontaneous
resolution providing discrete C2 cavities that allow guest chiral discrimination. An
extensive review of this particular aspect of enclathration with respect to TOT, as well
of its applications, has been recently published [1].
Fig. 1. Tri-o-thymotide constitution and idealized view of the ( )-(M)-TOT configuration. The atom numbering
(arbitrary) refers to that used in the crystal-structure determinations.
Gerdil and Barchietto have reported that the heterogeneous interaction of a series
of enclathrated oxirane derivatives with gaseous hydrogen halides (HXg) and aqueous
hydrogen halide solutions (HXaq) exhibited regio- or stereospecific asymmetric
reactions of the guests [2].The hydrogen halide (X ˆ Cl, Br) can permeate the
crystalline phase and react with the included guest without interfering chemically with
the host lattice. Most of the observed processes consist of two consecutive trans-
formations. First, a specific allylic isomerization of the oxirane occurs, followed by
either one of two paths depending on the substrate: a nucleophilic substitution of the
allylic OH group leading to an a,b-unsaturated halide, or the addition of hydrogen
halide to the double bond to form the corresponding halohydrin. The allylic
isomerization in the solid-state differs strikingly from the corresponding ring-opening
occurring in the homogeneous acidic phase where no traces of allylic product or
unsaturated halide can be detected.
In contrast to the exhaustive study of TOT-enclathrated oxiranes with HXg and
HXaq, there had been no attempt to examine the cage-controlled reactivity of the
related epoxy ketone system. Epoxy ketones have aroused interest as synthetic
intermediates because of their facile preparation, often with substantial stereochemical
control, and their enhanced reactivity in acid- and base-catalyzed reactions as well as in
other transformations [3]. The present publication continues our preceding studies,
examining the cage-controlled chemical behavior of 3,4-epoxycyclopentanone (ˆ 6-
oxabicyclo[3.1.0]hexan-3-one; 1) with the same reactants under similar conditions. The
choice of 1 as an epoxy ketone was also dictated by its small size that must fit the cage
volume, the latter being a rather constant property of TOT P3121 clathrates.

420
Helvetica Chimica Acta ± Vol. 82 (1999)
2. Reactions of Enclathrated 3,4-Epoxycyclopentanone (1) with HCl. Kinetics and
Configuration Correlations. ± The epoxy ketone 1 was obtained by peroxidation of
cyclopent-3-en-1-one using trifluoroperacetic acid in CH2Cl2 [4]. Single and micro-
crystals of TOT/1 were prepared following the general procedure described in [2].
Micro-twinned clathrate crystals (space group P31, host/guest stoichiometric ratio 1 :1)
were obtained by recrystallization of TOT in the pure guest 1, whereas the expected
lattice P3121 grew from a mixture of TOT, 1, and MeOH2) and had an occupation ratio
of 0.52 ± 0.66, a below-average value for this type of clathrate. Optically active
microcrystalline samples were prepared by seeding and proved to be 99% enantiomeri-
cally pure. The X-ray structure of TOT/1 was determined with a single crystal grown
from a slowly cooled TOT solution in a mixture of 1 and MeOH (volume ratio 1 :2). As
depicted in Fig. 2, the guest five-membered ring is approximately parallel to the vertical
twofold axis (not shown), with the carbonyl group pointing towards the origin.
Fig. 2. Stereoview of the (M)-TOT/3,4-epoxycyclopentanone (1) clathrate structure (P3121) comprising the host
molecules bordering the cavity. For clarity, the top right TOT molecule has been removed. The crystallographic
twofold axis passing through the cage lies parallel with the vertical b axis (not shown). The guest is viewed
across the ꢀentry channelꢁ as discussed in Sect. 2. Only the H-atoms of the guest are shown. The solid filled O-
atoms are H-bonded as shown in TOT/2, on opening of the epoxy ring.
2
)
The molecular volume of MeOH is too small to allow facile inclusion of the molecule.

Helvetica Chimica Acta ± Vol. 82 (1999)
421
To acquire a view on the kinetic aspects of the intracrystalline conversion of 1 under
the action of HCl, a series of reactions was performed under various conditions. The
ensuing results may allow one to make some choice among plausible mechanisms
within the cage. Reactions of enclathrated 1 were investigated for heterogeneous solid-
gas and solid-liquid systems. In the first system, microcrystalline conglomerates were
reacted in a gas flow of anhydrous HCl (HClg); in the second, a suspension of
microcrystals was gently stirred in an aqueous solution of HCl (HClaq). It was found
that the reaction involves cage-specific isomerization of the guest 1 to the correspond-
ing allylic alcohol 4-hydroxy cyclopent-2-en-1-one (2) in a first step (Scheme 1, Step a),
followed by nucleophilic substitution of the OH group (Step b) leading to the a,b-
unsaturated halide 3 (for convenience, enclosed molecules are depicted in square
brackets).
Scheme 1
The molecular ratios of the starting material and products were determined by peak
integration of 1H-NMR spectra and expressed as dimensionless relative concentrations
in the calculations. Microcrystalline conglomerates were used in the measurements,
entailing the absence of discrimination between diastereoisomeric reaction paths. As a
result, the observed kinetic parameters must be considered as average values. All things
being equal, the average flux of Hd‡Cld into the crystal lattice is supposed to be
constant for a given weight of microcrystals. It is reminded that the kinetic parameters
depend implicitly on the average size of the microcrystals (i.e., on the mean solid-phase
area per unit weight of microcrystals) that has been shown to be reproducible
throughout treatment of kinetic data owing to the standardized method of preparation.
P
The rate constants were calculated by minimizing the sum of squares ( err2) of the
difference between the variations in observed relative concentrations and those calculated
on the basis of the integrated form of the assumed rate law [5]. Concentration-time
measurements were achieved in the solid-liquid system at different HClaq concen-
trations and fixed temperatures. Kinetic data were rationalized in terms of rate
expressions applied to liquid solutions, revealing remarkable analogies with the latter,
as also disclosed in [2]. A concentration-time profile is depicted in Fig. 3,a for the
reaction carried out at 408. The kinetic behavior agrees well with the occurrence of two
successive first-order (or pseudo-first-order) steps such as 1 ± k1 ! 2 ± k2 ! 3 (Sche-
me 1). The bimolecularity of the reaction is substantiated by the steric confinement
imposed by the cage upon the reactants. However, in both steps, the concentration of
HCl may assume a ꢀconstantꢁ unitary value on account of the cage enclosure, as
1
reflected in the order of the reaction. The rates (in h ) measured at 40 and 508 (HClaq
concentration 12.4 mol/l) gave the following values: k1 ˆ 0.003 and 0.007, respectively,

422
Helvetica Chimica Acta ± Vol. 82 (1999)
and k2 ˆ 0.026 and 0.030, respectively, with root-mean-square deviations of the ob-
servations (rel. conc.) from the calculated data of 0.034 and 0.054, respectively. It
is pointed out that the molecule of H2O formed in the second step remains trapped in
the cage and only shows up as an immiscible layer on desolvation of the clathrate. This
again demonstrates the ability of the clathrate lattice to accommodate small structural
changes on variations of the cage contents. The chloride 3 appeared as major product at
both temperatures, whereas the concentration of 2 remained roughly constant, thus
pointing to a rate-determining step for the allylic conversion in accordance with the
1
rate constants and the activation energies DE1 75.3 and DE2 10.9 kJ ´ mol calculated
from the above k values measured at 40 and 508, respectively. It might be interesting to
emphasize that the magnitude of the rate constants for the allylic isomerization of the
enclathrated polymethyloxiranes [2] is from 60 to more than 100 times greater than that
of 1. For example, trimethyloxirane demonstrates a kinetic behavior very different
from that of 1 with k1 ˆ 0.870 h 1 at 228, followed by a rate-controlling HCl addition to the
1
allylic product with a constant k2 ˆ 0.032 h , in contrast to the Step-b transformation of
1. Reactions of TOT/1 with 48% aqueous HBr solution (HBraq) were carried out at 408
and shown to be very sluggish. The reaction mixture yielded 8% of 2 and no detectable
amount of 4-bromocyclopent-2-en-1-one after 96 h. The van der Waals volume of the
external reactant plays without doubt a role in slowing down the reaction.
As was already found for the oxirane clathrates, a slight change in the external
HClaq concentration influences drastically the intracystalline reaction rate. The
relationships k ˆ a ´ [HCl]n was postulated between k and the concentration of HClaq,
7
where a and n were optimized by a least-squares procedure to give k1 ˆ 3.01 ´ 10
´
[HCl]4 and k2 ˆ 1.69 ´ 10 5 ´ [HCl]3 (Fig. 3,b). By way of example, the k1 equation
Fig. 3. a) Selected concentration-time profile of the consecutive first-order reactions 1 ! 2 ! 3 at 408 (the
integrated rate-law expressions [5] (solid curves) are fitted to the data by a least-squares procedure;
concentration of HClaq 12.4 mol/l; k1 ˆ 0.0029, k2 ˆ 0.0262). b) Variations of k1 and k2, as functions of the
concentration of the external reactant HClaq at 508 ([HCl] in mol/l; k in h 1).

Helvetica Chimica Acta ± Vol. 82 (1999)
423
may be substituted for the rate constant in the first-order rate of change of 1 giving the
7
rate expression for the first step of the heterogeneous reaction: d[1]/dt ˆ 3.01 ´ 10
´
[1] ´ [HCl]4, where [HCl] is maintained constant. A reaction in solution is unlikely to
involve the fourth power of a concentration factor, and it is believed that this high
exponential dependency is due, in part, to the constitution of the outer diffusion layer at
the borderline liquid-crystal. The energy of desolvation of the hydrated acid might
contribute significantly to the barrier of crossing of the liquid-solid interface by a
ꢀnakedꢁ HCl molecule, the solvation shell of the ions being likely to increase with the
acid dilution.
The precise treatment of kinetic data for the solid-gas system is not easily feasible
owing to the experimental difficulty to express the ꢀconcentrationꢁ of HClg in the
reaction medium. The most critical part of the experimental set-up was the
maintenance of a constant and homogeneous HClg flow through the microcrystalline
conglomerate to ensure significant comparisons between the various samples. Consid-
erable desolvation of the solid phase was encountered during the experiments;
consequently, the reaction times were kept within the range from 60 to 180 h to limit the
decrease of the mol contents to a maximum of ca. 30%. Series of three to nine reactions
were carried out at three different temperatures, the kinetics agreeing with the
occurrence of two successive first-order steps as for the solid-liquid system. The
1
reaction rates (in h ) measured at 80, 100, and 1208 were k1 ´ 103 ˆ 17 Æ 2, 40 Æ 3, and
152 Æ 5, respectively, and k2 ´ 103 ˆ 15 Æ 6, 14 Æ 7, and 12 Æ 6, respectively. The reported
confidence limits, given at the 5% significance level, are those of the mean k values of
series of measurements at fixed temperatures. The allylic isomerization (k1) is
considerably accelerated with an increase in temperature, the Arrhenius plot giving a
1
rough estimation of the activation energy of 62 Æ 11 kJ ´ mol , a value close to that
observed for the solid-liquid system. The k2 values could only be estimated with
considerable inaccuracy in comparison with the k1ꢁs, thus showing the limitations of the
solid-gas investigations. It could only be concluded that the nucleophilic substitution
2 ± k2 ! 3 is practically temperature-independent within the range investigated.
A discussion of the possible mechanisms of the reaction of enclathrated 1 with HCl
might be useful and would benefit from a simplified but convenient description of the
clathrate cavity. The eight identical (M)-TOT molecules comprising the cavity (see
Figs. 2 and 4) might be considered as a model for a supermolecular chiral aggregate
with a C2 axis passing through the central cavity. In a recent comprehensive study of the
molecular packing of P3121 clathrates, the structural features of two small entry
channels leading into the central receptor were discussed and illustrated [1a][2]. They
are symmetry-equivalent and situated at either end of the longest axial dimension of
the cage (ca. 7 Š) viewed perpendicularly to Fig. 2. The presence of these ꢀchannelsꢁ
certainly imposes a preferred direction of attack of the incoming external reactant upon
the guest.
The epoxide 1 may be protonated along two directions of attack (Path A or B,
Scheme 2) to yield two different conjugate acids. The formal Cs symmetry of 1 entails
the presence of two pairs of enantiotopic protons in a-position to the ketone function
which can be abstracted by the basic Cl ion to produce the a,b-unsaturated alcohol
and a Hd‡Cld ion pair. Two pairs of six-membered enantiomeric transition states may
be envisaged for the concerted reaction, and the stereochemical outcome depends on

424
Helvetica Chimica Acta ± Vol. 82 (1999)
which proton is attacked. It has been shown [6] that on treatment with a strong base,
ring opening is prevalent consecutive to the attack on the enantiotopic proton Ha or Hb.
However, a limitation to the above symmetry considerations is worth mentioning.
Owing to the chiral environment and the distortion of the guest frame to fit the cage,
equivalence between the a-protons is cancelled resulting in the possible occurrence of
four distinct transition states. If the protonation occurs along Path B, the six-membered
transition state would be highly unfavorable due to the prohibitive Hc ´´´ Cld or
Hd ´´´ Cld distances and to the steric hindrance of the epoxide protons, therefore, the
ꢀsynꢁ mechanism along Path A should be favored.
Scheme 2
The conversion of enclathrated 1 gave optically active reaction mixtures of 2 and 3.
The presence of either one of the chiral products in the absence of the other was not
obtainable as follows from the kinetics and from adverse desolvation problems at long
reaction times (Sect. 5). A quantitative method for estimating the eeꢁs of the products
involves the relationship of Eqn. 1 where aobs is the optical activity of the reaction
mixture; the cꢀs and [a]ꢁs are the concentrations and specific rotations, respectively, of
the products 2 and 3. Both coefficients c and [a] are experimentally accessible for 2 and
3. Therefore, Eqn. 1 is linear in two unknowns (ee2 and ee3) that can be estimated by
least-squares analysis if we have a set of n such equations (n > 2).
aobs ˆ ee2 ´ c2 ´ [a]2 ‡ ee3 ´ c3 ´ [a]3
(1)
Five small microcrystalline samples of (‡)-TOT/1 (optical purity 99.8%) [2] were
reacted with HClg in a temperature range of 80 ± 1208 for 42 ± 96 h. Each sample was
then divided into three portions which were dissolved separately in CDCl3 and
1
analyzed by H-NMR spectroscopy for the concentrations of 2 and 3; the values 91.4
and 306.8 (see Sect. 4 and 5) were substituted in Eqn. 1 for [a]2 and [a]3, respectively.
Besides, chiroptical measurements showed levorotatory (aobs) reaction mixtures for
every sample. Thus, we had three sets of five independent linear Eqns. 1, each set

Helvetica Chimica Acta ± Vol. 82 (1999)
425
yielding a best estimate for the eeꢁs on calculation by the least-squares method. For
each set, consistent results were only obtained with both coefficients [a] < 0, disclosing
unambiguously the following correlation:
(‡)-(P)-TOT/1 ! ( )-2 ! ( )-3
The eeꢁs were relatively poor and not very accurately estimated, with overall values
ee2 ˆ 9 Æ 3% and ee3 ˆ 22 Æ 2%. According to expectations, the correlation matrix
showed a large correlation between the eeꢁs. The ee of 3, larger than that of 2, might
reflect a leading influence of the structural differences between the diastereoisomeric
associations (P)-TOT/(S)-2 and (P)-TOT/(R)-2 on the displacement of OH by Cl.
Such an effect would require that (P)-TOT/(R)-2 favours partial inversion with
respect to the couple (P)-TOT/(S)-2. In the present instance, one can evaluate to ca.
11% the fraction of (S)-3 formed in the chlorination step by inversion of (R)-2. The
formation of an excess of (S)-2 is also consistent with its spontaneous inclusion as
major enantiomer by seeding a TOT solution in racemic 2 with (‡)-TOT clathrate
microcrystals.
In a subsequent experiment, microcrystals of ( )-TOT/1 of optical purity > 99%
were reacted with HClg to yield dextrorotatory reaction mixtures after racemization of
the host on solution in CHCl3. Besides, reaction products were recovered by
desolvation of clathrate samples and analyzed by GC over a chiral column, showing
the presence of unreacted 1 with (‡)-2 (ee 8%) and (‡)-3 (ee 7 ± 8%) as major
enantiomers. Chirality transfer from the cavity onto 2 is in agreement with that found
previously, whereas the lower value of the ee of 3 might be accounted for by a partial
inversion process on a molecule prone to racemization at higher temperature
(desolvation at 1808). The correlation of optical rotations ( )-TOT/1 ! (‡)-(2) !
(‡)-3 is complementary to that found in the preceding approach and reveals that
proton abstraction from position 2 is slightly favored with respect to position 1 in Fig. 2.
The low efficiency of the chirality transfer may be explained by the orientation of the
syn C H bonds, both facing the same gap of the entry channel. The proximal TOT
carbonyl O-atom, apt to form the observed H-bond with the OH group, is probably a
crucial orienting factor lowering the activation energy of the overall transition state
involving the nearby H-atom in a-position.
3. Crystal Structure of TOT/4-Hydroxycyclopent-2-en-1-one (2) and Absolute
Configuration of the Guest. ± The 4-chlorocyclopent-2-en-1-one (3) has already been
described [7], but its chiroptical properties remained unknown until now. The isolation
of optically active 3 as sole product from the reaction of TOT/1 was not possible as
pointed out above. The expected inclusion method was also made impossible on
account of the elusive P3121 TOT/3 clathrate (see Sect. 5). Consequently our attention
was aimed at obtaining pure 3 by reacting TOT/2 with HClaq. A large quantity of
racemic 2 was prepared by a new pathway (Scheme 3) differing from that used by
Noyori and coworkers [4]. Reaction of cyclopent-2-en-1-one with N-bromosuccinimide
in CCl4 afforded 4-bromocyclopent-2-en-1-one (4) in yields up to 73%. The conversion
of 4 to 2 was accomplished by hydrolysis of either 4 or 4-acetoxycyclopent-2-en-1-one
(5). Hydroxylation of 4 in aqueous solution (neutral, basic, or acidic) yielded the

426
Helvetica Chimica Acta ± Vol. 82 (1999)
cyclopentadiene dimer as main product in all cases (Scheme 3, Path a). Deacetylation
of 5 with K2CO3 aqueous MeOH solution [8] was inappropriate because 4-
methoxycyclopent-2-en-1-one was formed in 57% yield as the only identified product.
On the other hand, hydrolysis of 5 gave the expected pure product 2 in acidic media,
the most efficient condition (66% of 2) being the action of aq. 0.5n H2SO4 at 558
(Scheme 3, Path b).
Scheme 3
A challenging problem pertaining to X-ray crystal-structure analysis resides in the
determination of the absolute configuration of chiral crystal components. If the precise
recognition of the conformation of a chiral guest in P3121 clathrates is at stake
(unambiguous resolution of the guest atomic positions, the absolute configuration of
TOT being known), the inclusion of a large excess of one of the guest enantiomers is
peremptory. To avert this complication in case of low enantioselectivity of the clathrate,
a substrate presenting a large ee of one of its enantiomers can be employed in the first
recrystallization. Despite the fact that the enantioselectivity of TOT/2 seemed
reasonably ꢀhighꢁ (ee mean value 61.0 Æ 0.3%), optically active 2 was prepared
separately via asymmetric synthesis. Starting with (‡)-(R,R)-tartaric acid, ( )-(R,R)-
1,4-diiodo-2,3-(isopropylidenedioxy)butane (57%) was obtained in four steps [9]. The
diiodo derivative was allowed to react with the lithio derivative of methyl methyl-
thiomethyl sulfoxide [10], followed by acid hydrolysis (1n H2SO4/Et2O) affording ( )-
(S)-2 in 34% overall yield. The ee of ( )-2 amounted to 78% as shown by the 1H-NMR
spectrum of its Mosher derivative [11]. Though the absolute configuration of ( )-2 was
known from a previous work of Noyori and coworkers [12] and from the asymmetric
synthesis above, a knowledge of the structural features of TOT/2 was important to shed
light on the mechanism of the intracrystalline reaction of the guest with Hd‡Cld . The
correlation (‡)-TOT/( )-2 was demonstrated before hand by the usual polarimetric
approach [1b][2]. X-Ray crystal-structure analysis was carried out on a single crystal
grown from a TOT solution in ( )-2 the ee of which was 78.8%. Single crystals selected

Helvetica Chimica Acta ± Vol. 82 (1999)
427
from the same batch revealed an average ee of 92% for included ( )-2 owing to the
phenomenon of chiral amplification [1b], a proper value for a satisfactory crystallo-
graphic structural resolution of the guest.
A stereoview of the crystal structure of the enclathrated substrate ( )-2 is shown in
Fig. 4 3). The guest molecular plane is approximately parallel to the vertical crystallo-
graphic twofold axis (not shown) with the carbonyl C-atom located on the axis.
Interestingly, the OH group is H-bonded to one of the two TOT carbonyl O-atoms
protruding into the cage. This appears to be the first observed case of one of the TOT
carbonyl atoms acting as an active site in the cage and is at variance with all the other
known OH-bearing TOT guests, as e.g. butan-2-ol or acetic acid. Inspection of the host-
guest intermolecular contacts suggests that 2 fits snugly into the receptor as further
exemplified by a fair chiral recognition (ee 61%). The well-resolved guest conformer in
(‡)-(P)-TOT/2 had the S-configuration and the levorotatory residual optical activity
after racemization of the (‡)-TOT host of several clathrate crystals disclosed
unambiguously the correlation of configuration (‡)-(P)-TOT/( )-(S)-2, and conse-
quently the absolute configuration of 2.
Fig. 4. Stereoview of the cage of (M)-TOT/(R)-4-hydroxycyclopent-2-enone. For clarity, the top right TOT
molecule has been removed. One of the two TOT carbonyl O-atoms protruding into the cavity is H-bonded to
the OH group of 2, as depicted by the dotted line joining the solid O-atoms.
3
)
Standardized views of the P3121 cage comprising the host antipodes ( )-(M)-TOT have been consistently
presented in our studies of TOT clathrates, thus allowing an easy comparison between the guestsꢁ
orientations in different species (e.g., Figs. 2 and 4). Host-guest interactions are not affected by this
approach.

428
Helvetica Chimica Acta ± Vol. 82 (1999)
4. Revised Determination of the Specific Rotation of 4-Hydroxycyclopent-2-en-
one. ± The preceding results corroborate through a different pathway the (R)-
configuration assigned previously to (‡)-2 by Noyori and coworkers associated with a
specific rotation [a]D of ‡ 90.1 at 228 (c ˆ 0.43, MeOH) [12], no standard error being
given. The enantiomer purity of 2 was determined by 1H-NMR analysis of its Mosher
derivatives ((‡)-(R)-MTPA). The measurements were performed with samples
obtained along three different pathways (Fig. 5): a) by desolvation of the micro-
crystalline clathrate ( )-TOT/(‡)-2 prepared from racemic 2 (ee 61.2%), b) by
asymmetric synthesis from (R,R)-tartaric acid i.e. with ( )-2 (ee 78.8%, Sect. 3), and c)
by crystallization of the clathrate (‡)-TOT/( )-2 on seeding with the preceding
enantiomerically enriched probe (see b) (final ee 91.6% after desolvation). Subsequent
polarimetric measurements gave access to a value for the specific rotation of 2
(Table 1, standard deviations at the 5% level of significance) that differ from the result
of Noyori and coworkers. The value of the specific rotation of 2 was complemented by
additional measurements in which the ee of the sample of Fig. 5,c was determined by
GC analysis on a chiral column in association with polarimetric measurements. By the
same token, this allowed configurational peak assignments in the chromatograms. Both
20
methods delivered consistent [a]2D0 and [a] values in MeOH and CHCl3.
546
1
Fig. 5. Methylene H-NMR signals of the Mosher derivatives of 2: a) sample obtained after desolvation of the
clathrate ( )-TOT/(‡)-2 prepared from racemic 2; b) sample obtained from ( )-(S)-2 prepared from l-tartaric
acid; c) sample released from the clathrate (‡)-TOT/( )-2 prepared from enantiomer-enriched ( )-(S)-2 (ee
78.8%).
Table 1. Specific Rotation of 4-Hydroxycyclopent-2-en-1-one (2)
20
ee [%]
[a]2D0
[a]
Solvent
546
(‡)-(R)-2
61.2a)
78.8a)
91.6b)
‡ 97.2
97.2
96.9
Æ 97.1 Æ 0.4
‡ 100.3
100.8
100.3
Æ 100.5 Æ 0.7
MeOH
MeOH
MeOH
(
(
)-(S)-2
)-(S)-2
Mean [a]20 c) (c ˆ 2 ± 6, MeOH)
(‡)-(R)-2
61.2a)
78.8a)
91.6b)
‡ 91.3
‡ 92.8
CHCl3
CHCl3
CHCl3
(
(
)-(S)-2
)-(S)-2
91.5
93.9
91.4
Æ 91.4 Æ 0.3
93.2
Æ 93.2 Æ 1.4
Mean [a]20 c) (c ˆ 2 ± 3, CHCl3)
a) Based on Mosherꢁs derivative. b) Based on Mosherꢁs derivative and GC over a chiral column (Lipodex E).
c) R.m.s. errors calculated at the 5% significance level.

Helvetica Chimica Acta ± Vol. 82 (1999)
429
5. Enantiomer Enrichment of 4-Chlorocyclopent-2-en-1-one in the Solid State and
Determination of its Absolute Configuration. ± A knowledge of the configuration of 3
was primarily expected from the X-ray structure analysis of its TOT clathrate. Small
TOT/3 single crystals were grown in a TOT solution in the pure substrate or with ( )-a-
pinene as cosolvent which afforded larger crystals. The unit cell belonged to the
orthorhombic centrosymmetric space group Pbca (a ˆ 13.184, b ˆ 23.086, c ˆ 23.977 Š;
stoichiometric ratio TOT/3 1 : 1). No other crystallization conditions were found that
triggered the formation of the required P3121 space group, ruling out the straightfor-
ward method used for the analysis of the configuration of 2. A different approach was
used to determine the absolute configuration of 3. Enantiomer-enriched 3 was
expected from the reaction of enclathrated optically active 2 with HCl. However, ( )-2,
ee 78% in Et2O solution, was completely converted to racemic 3 under the action of dry
gaseous HCl for 1 h at room temperature, whereas no reaction took place overnight
with 20% aqueous HCl solution at the same temperature in CHCl3. Heterogeneous
reactions were preliminarily carried out making use of TOT/2 conglomerates in the
search for the proper conditions. The conversion from 2 to 3 was not complete under
the most favorable reaction conditions (HClaq (38%), 538, 110 h; molar ratio 2/3
40 : 60) and exhibited to a significant extent adverse desolvation of the guests caused by
the long reaction time and the higher temperature necessary to produce sizeable
amounts of 3. The slow reaction rate may be due, in part, to the host-guest H-bond
(Fig. 4), the cleavage of which would increase the activation energy. Desolvation of the
reacted clathrate (‡)-TOT/( )-2 (clathrate optical purity 99%, ee of 2 91.2%) with
HClaq (38%) afforded a mixture of unreacted 2 with 3 as sole product (molar ratio 2/3
1 : 3). The separation of the components of two independently reacted and desolvated
crystal batches (I and II) was achieved by preparative TLC (Et2O). This method did
not cause the racemization of 3, as demonstrated by comparing the enantiomer purity
of 3 before and after separation. The ee of the samples of 3 obtained by the above
treatments was determined by GC over a chiral stationary phase (Table 2, Batches I
and II). The ee of ( )-3 was 35 and 33% for the Batches I and II, respectively. The high
rotatory power of 3 was determined by subsequent polarimetric measurements in
CDCl3 and MeOH.
A chemical method was employed as a last resort to specify the absolute
configuration of 3. The absolute configuration of 2 being known, the related
configuration of 3 was expected from the conversion of optically active 2 to the
chloride 3 under the action of thionyl chloride in the presence of an amine. This
reaction is well known to induce predominantly either retention or inversion of
Table 2. Specific Rotation of 4-Chlorocyclopent-2-en-1-one (3) Prepared by Heterogeneous Reactions of
a
Enclathrated ( )-(S)-2 with HClaq
)
Batchb)
ee [%]c)
[a]2D0
[a]
Solvent
20
546
I
II
II
35
33
33
306.6 Æ 1.5
368.9 Æ 1.8
368.2 Æ 1.8
342.1 Æ 1.7
CDCl3
CDCl3
MeOH
306.9 Æ 1.5
285.4 Æ 1.4
a) R.m.s. errors estimated at the 5% significance level. b) Chemical purity of Batches I and II after desolvation:
97.3 and 98.9%, resp. c) Measured by GC over a chiral column (Lipodex E).

430
Helvetica Chimica Acta ± Vol. 82 (1999)
configuration of the product depending on the participation (Et2O, dioxane) or the
absence of participation (benzene) of the solvent. The stereochemical study of the
reaction of ( )-(S)-2 (ee 78.8%) with thionyl chloride was performed in three different
solvents following the procedure of Young and Caserio [13], formerly applied to the
conversion of ( )-(R)-but-2-en-2-ol, a substructure of 2. Contrary to the expected
solvent-induced retention in Et2O, the chloro derivative of the latter compound was
formed predominantly with inversion of configuration. The same effect, unexpectedly,
occurred during chlorination of 2 in Et2O. This phenomenon might be imputed in both
cases to the intramolecular delocalization of the positive charge developed in the
reactant upon expulsion of SO2 that minimizes the charge transfer onto the solvent
while granting more migration freedom to the neighboring Cl ion. The formation of a
ꢀlong-livedꢁ ionic intermediate in the conversion of ( )-(R)-but-3-en-2-ol is supported
by the formation of 1-chloro-but-2-ene beside (‡)-(S)-3-chloro-but-2-ene, in the molar
ratio 1:2 [13]. The resolution of the optically active samples of 3 was achieved by GC
over a chiral column. In benzene, where structural inversion is expected, ( )-(S)-2 gave
(‡)-3 with 77% inversion, from which the conclusion could be reached that (‡)-3
corresponds to the (R)-configuration. In Et2O, thionyl chloride gave (‡)-3 with 83%
inversion, in accordance with the noteworthy behaviour of ( )-(R)-but-3-en-2-ol. In
dioxane, the expected prevalent retention was observed. The results are reported in
Table 3. Hereby, percent inversion or retention of configuration expresses the fraction
of the total amount of pure enantiomer present in the starting material that underwent
either one of the processes, the difference from 100% representing the contribution of
the inverse process. This is at variance with the values given usually with the difference
from 100% understood to be racemization. The same conclusion regarding the
configuration of 3 is arrived at, when the octant rule is applied.
Table 3. Chemical Determination of the Absolute Configuration of 4-Chlorocyclopent-2-en-1-one (3) Using the
Reaction of ( )-(S)-2 with Thionyl Chloride
b
Solvent
ee [%]a)
[a]2D0
)
(CDCl3)
3
Stereochem. result [%]c)
Benzene
Et2O
Dioxane
42.3
52.6
3.9
‡ 313
‡ 312
306
(c ˆ 3.6)
(c ˆ 1.9)
(c ˆ 1.9)
(‡)-(R)
(‡)-(R)
77 (inversion)
83 (inversion)
52 (retention)
( )-(S)
a) % Enantiomer purity of the chloride measured by GC over a chiral column (Lipodex E). b) Estimated
specific rotation. c) Calculated relative to the amount of enantiomerically pure alcohol initially present as
reactant (see text).
The reaction of optically pure microcrystalline (‡)-TOT/( )-(S)-2 (ee of 2 ca.
92%) in contact with HClaq (38%) yielded ( )-(S)-3 with an ee of 35%, indicative of
70% retention during the conversion, as expressed by the correlation of configurations
(P)-TOT/(S)-2 ! (P)-TOT/(S)-3. Hereby, ꢀ70% retentionꢁ means that, out of 100
enclathrated (S)-2 guest molecules, the configuration of ca. 70 molecules remained
unchanged under the action of Hd‡Cld . On the basis of the known configurations of 2
and 3, possible mechanisms of the reaction are depicted in Scheme 4. A bimolecular
interaction is by nature occurring within the cage. The reaction site (C OH moiety) of
the substrate is sterically well positioned with respect to both entry channels (see
Fig. 4), restricting the approach of the incoming reactant towards the chiral center of

Helvetica Chimica Acta ± Vol. 82 (1999)
431
(S)-2 to two opposite directions: Path I and II (decrease of the entropy term). It will be
noted that, owing to the H-bond between the alcohol and one of the cage carbonyl O-
atoms, the facility of formation of the oxonium intermediate 2' differs from one
direction of attack to the other. Protonation along Path I, contrary to Path II, will be
favored by the proper orientation of the OH lone electron pairs directed towards the
incoming positive charge. The negative end of the Hd‡Cld ion pair is likely to be
positioned ꢀaboveꢁ the ring mean plane and sterically close to the C(4) ± C(5) bond. As
a subsequent reaction step, the heterolytic dissociation of the C O bond leading to a
highly reactive allylic carbocation (Step a) might be envisaged. The fate of the
carbocation is twofold depending on the side of the ring attacked by Cl . It is
reasonable to assume that the confines of the receptor prevent a facile shift of Cl from
one side of the ring to the other, thus favoring Step b (retention) at the expense of
Step c (inversion), as was experimentally observed. The same stereochemical
implications might also allow for a frontside nucleophilic attack of Cl on C(4) with
retention of configuration (Step d). A Sn2 substitution with clear inversion (Step e)
seems less likely in view of the confinement of the reactants as expressed above.
Scheme 4
Experimental Part
General. TOT was synthesized by cyclodehydration of o-thymotic acid with POCl3 [1]. All solvents and
reagents were purchased from high-grade commercial sources, and the solvents redistilled before use. MgSO4
and Na2SO4 were employed as drying agents. Cyclopentadiene was prepared by pyrolysis of commercial
dicyclopentadiene (Fluka; 85 ± 90% by GC) and stored in a tightly capped bottle at dry-ice temperature. Et2O
was distilled from sodium benzophenone ketal (ˆsodium diphenylketyl) prior to use. Gaseous HCl was
produced by the action of conc. sulfuric acid upon a conc. HCl soln. and dried through two traps of conc. sulfuric
acid and one filled with P2O5. Gaseous HBr was generated by the action of Br2 on tetrahydronaphthalene. The
precise concentrations of HClaq were titrated with 1.000n or 0.100n standard soln. of NaOH. TLC: Merck TLC
silica gel 60 F254 plates. Gas chromatography (GC): Vega-GC-6000 capillary column (SPB 20, 30 m  1.0 mm),
chiral column Lipodex E. Polarimetric measurements: Perkin-Elmer-241 polarimeter (accuracy: Æ 0.0028). 1H-
(400 MHz) and 13C-NMR (100.6 MHz) Spectra: Bruker-AMX-400 spectrometer; SiMe4 as internal standard in
CDCl3; chemical shifts d in ppm, J in Hz. IR Spectra: IR-Perkin-Elmer 681 using NaCl soln. cells; in cm 1. Mass
spectra: GV-7070 mass spectrometer with mass data system; m/z (rel. %).

432
Helvetica Chimica Acta ± Vol. 82 (1999)
Microcrystalline P3121Clathrates. The general procedure for their preparation and desolvation were
described elsewhere, as well as the method for the determination of the optical purity of the crystalline samples
[1][2].
1. Synthesis of 1. 6-Oxabicyclo[3.1.0]hex-2-ene. Peracetic acid (40%; 126 g, 0.66 mol), previously treated
with AcONa (3 g, dried), was added drop by drop to an ice-cold stirred mixture of freshly cracked
cyclopentadiene (45 g, 0.68 mol) and powdered anh. Na2CO3 (290 g) in CH2Cl2 (750 ml). The temp. was
maintained at 208 during the addition (ca. 1 h). After stirring for an additional 1.5 h, a starch/NaI test proved
negative. The solid cake was filtered off and washed with CH2Cl2 (3 Â 100 ml). The solvent was removed by
distillation through a Vigreux column and the higher-boiling residue collected under reduced pressure in an ice-
cold receiver to give 6-oxobicyclo[3.1.0]hex-2-ene (24.4 g, 59%). B.p. 38 ± 408/45 Torr. 1H-NMR: 2.31 ± 2.68
(m, CH2); 3.79 ± 3.91 (m, 2 CH); 5.94 ± 6.15 (m, 2ˆCH).
Cyclopent-3-en-1-one [4]. The 6-oxabicyclo[3.1.0]hex-2-ene (27 g, 0.33 mol) was added slowly to a soln. of
[Pd(PPh3)4] (50 mg, 43 mmol) in CH2Cl2 (80 ml) cooled in an ice bath. The reaction was highly exothermic and
the temp. rose to ca. 58. After heat evolution had ceased (ca. 4 h), the mixture yielded cyclopent-3-en-1-one
(15.8 g, 95%), after flash distillation and redistillation through a Vigreux column. B.p. 57 ± 618/107 Torr. No trace
1
of cyclopent-2-en-1-one was detected. H-NMR: 2.83 (s, 2 CH2); 6.04 (s, 2ˆCH).
6-Oxabicyclo[3.1.0]hexan-3-one. (1) A mixture of cyclopent-3-en-1-one (8.2 g, 0.1 mol), NaHCO3 (30 g),
and EDTA disodium salt (ˆdisodium dihydrogen ethylenediaminetetraacetate; 33 mg) in CH2Cl2 (70 ml) was
treated with CF3CO3H (prepared from 92% H2O2 soln. (3.9 ml, 0.15 mol) and (CF3CO)2O (24 ml, 0.17 mol) in
CH2Cl2 (70 ml) [14]) for 15 min at 08, then for 20 h at r.t. The reaction was quenched with Na2S2O3 ´ 5H2O (30 g)
and the org. layer, together with the CH2Cl2 extracts (5 Â 40 ml), dried (MgSO4). The solvent was carefully
removed under partially reduced pressure, and further distillation at 32.5 ± 348/0.3 Torr gave oily, pale yellow 1
‡
(6.7 g, 68%). 1H-NMR: 2.49 (m, 2 CH2); 3.76 (m, 2 CH); MS: 98 (72, M ), 70 (100, C4H6O), 55(42).
2. A Novel synthesis of 2. 4-Bromocyclopent-2-en-1-one (4). A mixture of cyclopent-2-en-1-one (50 g,
0.61 mol), N-bromosuccinimide (120 g, 0.67 mol), and a,a'-azodiisobutyronitrile (ˆ 2,2'-azobis[2-methylpro-
panenitrile]; 2.5 g) in CCl4 (650 ml) was stirred at 77 ± 788 for 2.5 h. After cooling in an ice-water bath, the
mixture was filtered and the filter cake washed with cold CCl4 (4 Â 50 ml). The filtrate was washed with ice-
water (3 Â 300 ml), then with 0.1n Na2S2O3 (300 ml), and dried (MgSO4). After evaporation at 458, the residual
yellowish oil was distilled to yield 4 (70.3 g, 73%) which was stored at 08. B.p. 47 ± 488/0.1 Torr [15]. IR (film):
1719s, 1578m, 1341m, 1178m. 1H-NMR: 2.71 (dd, J ˆ 19.5, 1.4, 1 H, CH2); 3.03 (dd, J ˆ 19.5, 6.2, 1 H, CH2); 5.17
(m, CHBr); 6.27 (dd, J ˆ 5.6, 1.2, ˆCH); 7.68 (dd, J ˆ 5.6, 2.4, ˆCH). 13C-NMR: 42.6, 44.8, 134.6, 162.2, 204.4.
‡
MS: 161 (48, M ), 160(50), 133(10), 81(100), 53(54).
4-Acetoxycyclopent-2-en-1-one (5). A suspension of AcOAg (28.4 g, 0.17 mol) in a soln. of 4 (27 g,
0.168 mol) in AcOH (240 ml) was stirred under reflux for 5 h. AgBr (30 g) was filtered off and AcOH
evaporated. Vacuum distillation of the residue gave 5 (18.5 g, 78%). Colorless liquid. B.p. 58 ± 608/0.1 Torr [15].
1
IR (film): 1723s, 1589w, 1373m, 1351m, 1240s, 1023s, 795m. H-NMR: 2.11 (s, Me); 2.33 (dd, J ˆ 18.4, 2.4, 1 H,
CH2); 2.83 (dd, J ˆ 18.4, 6.4, 1 H, CH2); 5.86 (m, H C(4)); 6.34 (dd, J ˆ 5.6, 1.2, ˆCH); 7.58 (dd, J ˆ 5.6, 2.4,
‡
ˆCH). 13C-NMR: 20.7, 40.9, 71.8, 136.9, 158.9, 170.3, 204.7. MS: 140 (29, M ), 112(36), 98(100), 97(68),
80(37), 53(92).
4-Hydroxycyclopent-2-en-1-one (2). A mixture of 5 (48.2 g, 0.344 mol) in 0.5n H2SO4 (150 ml) was stirred
for 12 h at 558. After complete disappearance of 5 (1H-NMR monitoring), the hydrolysed mixture was
neutralized with Na2CO3 (ca. 12 g) and extracted thoroughly with Et2O (Soxhlet). The combined Et2O extracts
were dried (MgSO4) and evaporated. Distillation of the residue yielded 2 (22.5 g, 66%). Colorless oil. B.p. 62 ±
1
648/0.1 Torr. IR (film): 3392, 172, 1671, 1586, 1404, 1044. H-NMR: 2.27 (dd, J ˆ 2, 6.4, 1 H); 2.77 (dd, J ˆ 6.4,
‡
18.4, 1 H); 5.05 (m, 1 H); 6.22 (dd, J ˆ 1.6, 5.6, 1 H); 7.60 (dd, J ˆ 2, 5.6, 1 H). MS: 98 (100, M ), 84(86),
70(55).
1H-NMR signals of the Mosher derivatives of 2: (‡)-(R)-MTPA/(‡)-(R)-2: 2.31 (dd, J ˆ 18.8, 2.2, 1 H,
CH2); (‡)-(R)-MTPA/( )-(S)-2: 2.41 (dd, J ˆ 18.8, 2.2, 1 H, CH2).
3. Racemic 4-Chlorocyclopent-2-en-1-one (3). Anh. HCl was bubbled under vigorous stirring through a
soln. of 1 (2.65 g, 0.027 mol) in CH2Cl2 (10 ml) for ca. 2 h at r.t.; the exothermic reaction was controlled by a
cold-water bath, maintaining the temp. below 258. The soln. was refluxed until the starting material was
completely consumed (1H-NMR monitoring), then neutralized with a 7% NaHCO3 soln. The org. layer was
washed with H2O and the aq. soln. extracted with CH2Cl2. The combined org. layer and extracts were dried
(MgSO4) and evaporated. The residue was distilled under reduced pressure to yield 3 (2.25 g, 72%). Pale yellow
liquid, unstable at r.t. [7]. B.p. 40 ± 418/1.5 Torr. IR (film): 1721.4s, 1585.4m, 1400.7, 1181.1, 792.3. 1H-NMR: 2.60
(dd, J ˆ 2, 19.2, 1 H, CH2); 2.98 (dd, J ˆ 6.8, 19.2, 1 H, CH2); 5.10 (m, CHCl); 6.31 (dd, J ˆ 1.2, 6.4, ˆCH); 7.58

Helvetica Chimica Acta ± Vol. 82 (1999)
433
‡
(dd, J ˆ 2.4, 6.4, ˆCH). 13C-NMR: 44.5 (t); 54.2 (d); 135.3 (d); 161.3 (d); 204.5 (s). MS: 116 (42, M ), 81(100),
53(73).
4. Enantiomer Enrichment of 3. The heterogeneous reactions of (‡)-TOT/( )-2 with aq. HCl were
performed according to the method indicated above. Desolvation was carried out at 180 ± 1858/0.02 ± 0.05 Torr
with the volatile fraction collected at liq. N2 temp. The separation of optically active 3 from unreacted 2 was
performed by prep. TLC (Et2O). The enantiomer purity of 3 was measured by GC (Lipodex E; oven temp.
1008; injector temp. 2858; detector temp. 3008).
Absolute Configuration of 3. General procedure [13]: thionyl chloride (119 mg, 1 mmol) was slowly added
to a stirred soln. of ( )-(S)-2 (ee 78.8%; 98 mg, 1 mmol) and tributylamine (185 mg, 1 mmol) in the adequate
solvent (5 ml) as given below. The soln. was stirred at r.t., followed by evaporation. Fractionating the residue by
prep. TLC (Et2O) gave optically active 3 as a pale yellow oil. The eeꢁs were determined as above; the specific
rotations are given in Table 2. Diethyl ether: after 2 h stirring, 60 mg (52%) of (‡)-3; chemical purity 97.7% by
GC, ee 52.6%. Benzene: after 2 h stirring, 60 mg (52%) of (‡)-3; chemical purity 96.8%; ee 42.3%. Dioxane:
after 6 h stirring, 50 mg (45%) of ( )-3; chemical purity 93.8%; ee 3.9%.
4
Crystal-Structure Determination of TOT/1 ). (C33H36O6) ´ (C5H6O2)0.5, Mr 577.7, m ˆ 0.616 mm 1, F(000) ˆ
1848, dx ˆ 1.18 g ´ cm 3, trigonal, P3121, Z ˆ 6, a ˆ b ˆ 13.6602(4), c ˆ 30.304(2) Š, Vˆ 4897.2(4) Š3; from 24
reflections (348 < 2q < 468); colorless prism 0.30 Â 0.30 Â 0.28 mm mounted on a quartz fiber. Cell dimensions
and intensities were measured at r.t. on a Nonius-CAD4 diffractometer with graphite-monochromated CuKa
radiation (l 1.5418 Š), w-2q scans, scan width 1.28 ‡ 0.14 tg q, scan speed 0.02 ± 0.148/s. Two reflections
measured every 100 reflections showed variations less than 3.0 s(I); 0 < h < 12, 0 < k < 12, 0 < l < 31, and all
antireflections of these; 3948 unique reflections were measured of which 3375 were observable (j Fo j> 4s (Fo)).
Data were corrected for Lorentz and polarization effects. The host structure was refined starting from the
previously determined coordinates in the isomorphous TOT/but-3-en-2-ol (2 : 1) clathrate [16]. Calculations
were carried out using the XTAL program [17]. DF syntheses revealed the major peaks of the two symmetry-
equivalent guest enantiomers observable within the region of the cage after refinement of the overall model.
The TOT structure was refined anisotropically by full-matrix least-squares with unit weights for the j F j values;
the positions of the H-atoms were calculated and included in the model with overall isotropic displacement
parameters 0.060 Š2; guest atoms were refined isotropically and included with equal weight 0.30 in the structure
(overall occupation factor 60%). Final values: R ˆ wR ˆ 0.070 and S ˆ 1.98 for 378 variables and 3372
contributing reflections. The mean shift/error was 0.57 ´ 10 3. The final difference electron density map showed a
3
maximum of ‡ 0.46 and a minimum of 0.27 eŠ
.
4
Crystal Structure Determination of TOT/2 ). (C33H36O6) ´ (C5H6O2)0.5, Mr 577.7, m ˆ 0.616 mm 1, F(000) ˆ
1848, dx ˆ 1.17 g ´ cm 3, trigonal, P3121, Z ˆ 6, a ˆ b ˆ 13.7649(5), c ˆ 30.057(2) Š, Vˆ 4932.0(4) Š3; from 24
reflections (348 < 2q < 468); colorless prism 0.22 Â 0.25 Â 0.35 mm mounted on a quartz fiber. Cell dimension
and intensity measurements and data reduction were performed under conditions identical to those used for 1.
Of the 3971 unique reflections measured, 3082 were observable (j Fo j> 4s(Fo)). The approach to the crystal-
structure determination was the same as that for 1. A minimum-energy conformation was calculated for 2 by the
force-field method (MM2 program [18]). The (‡)-(R)-molecule was initially positioned on the sites revealed by
DF syntheses within the region of the cage after refinement of TOT alone. The TOT structure was refined
anisotropically by full-matrix least-squares with unit weights for the j F j values; the positions of the H-atoms
were calculated and included in the model with overall isotropic displacement parameters 0.060 Š2. The guest
atoms positions were refined isotropically applying restraints on bond lengths and angles and included with
equal occupation factor 0.30 in the structure (measured guest occupation factor in the clathrate 61%). Final
values; R ˆ wR ˆ 0.077 and S ˆ 2.38 for 381 variables and 3971 contributing reflections. The mean shift/error
was 0.31 ´ 10 2. The final difference electron density map showed a maximum of ‡ 0.56 and a minimum of
3
0.29 eŠ
.
4
)
Crystallographic data have been deposited with the Cambridge Crystallographic Data Center, University
Chemical Laboratory, 12 Union Road Cambridge CB21EZ, England. The P3121 TOT clathrates are
isomorphous, and the coordinates of the host atoms are very similar in all species studied. A typical
coordinates set is given in [16].

434
Helvetica Chimica Acta ± Vol. 82 (1999)
REFERENCES
[1] a) R. Gerdil, in ꢀComprehensive Supramolecular Chemistryꢁ, Vol. 6, ꢀSolid-State Supramolecular
Chemistry: Crystal Engineeringꢁ, Ed. D. D. MacNicol, F. Toda, and R. Bishop, Elsevier Science Ltd,
Oxford, 1996, p. 239 ± 280; b) R. Gerdil, in ꢀTopics in Current Chemistryꢁ, Ed. E. Weber, Springer-Verlag,
Berlin-Heidelberg, 1987, Vol. 140, p. 75 ± 105.
[2] R. Gerdil, G. Barchietto, Helv. Chim. Acta 1994, 77, 691.
Â
Â
[3] M. Bartok, K. L. Lang, in ꢀ Heterocyclic Compounds, Small Ring Heterocyclesꢁ, Part 3, ꢀOxiranesꢁ,
Ed. A. Hassner, Wiley, New York, 1986, Vol. 42, p. 1 ± 196.
[4] M. Suzuki, Y. Oda, R. Noyori, J. Am. Chem. Soc. 1979, 101, 1623.
[5] K. B. Wiberg, in ꢀPhysical Organic Chemistryꢁ, John Wiley & Sons, Inc., New York, 1964.
[6] R. P. Thummel, B. Rickborn, J. Am. Chem. Soc. 1970, 92, 2064.
[7] B. W. Ponder, D. R. Walker, J. Org. Chem. 1967, 32, 4136.
[8] D. Bianchi, P. Cesti, E. Battistel, J. Org. Chem. 1988, 53, 5531.
[9] K. Ogura, M. Yamashita, G. Tsuchihashi, Tetrahedron Lett. 1976, 759; T. Tanaka, S. Kurozumi, T. Toru,
S. Miura, M. Kobayashi, S. Ishimoto, Tetrahedron 1976, 32, 1713; M. Kitamura, K. Manabe, R. Noyori,
H. Takaya, Tetrahedron Lett. 1987, 28, 4719.
[10] K. Ogura, M. Yamashita, M. Suzuki, G. Tsuchihashi, Tetrahdron Lett. 1974, 3653; K. Ogura, M. Yamashita,
S. Furukawa, M. Suzuki, G. Tsuchihashi, ibid. 1975, 2767.
[11] J. A. Dale, D. L. Dull, H. S. Mosher, J. Org. Chem. 1969, 34, 2543.
[12] M. Suzuki, T. Kawagishi, A. Yanagissawa, T. Suzuki, N. Okamura, R. Noyori, Bull. Chem. Soc. Jpn. 1988,
61, 1299.
[13] W. G. Young, F. F. Caserio Jr., J. Org. Chem. 1961, 26, 245.
[14] W. D. Emmons, J. Am. Chem. Soc. 1954, 76, 3468.
[15] C. H. DePuy, M. Isaks, K. L. Eilers, G. F. Morris, J. Org. Chem. 1964, 29, 3503.
[16] J. Siripitayananon, R. Gerdil, Acta Crystallogr., Sect. C 1989, 45, 768.
[17] S. R. Hall, H. D. Flack, J. M. Steward, ꢀXTAL 3.2 Userꢁs Manualꢁ, Universities of Western Australia and
Maryland, 1992.
[18] N. L. Allinger, J. Am. Chem. Soc. 1977, 99, 8127.
Received December 23, 1998