Generation of highly enantioenriched crystalline products in reversible asymmetric reactions with racemic or achiral catalysts

Article abstract of DOI:10.1002/anie.200803877

(Chemical Equation Presented) Backwards is better: By combining forward and backward reaction steps of a reversible chemical process (see scheme) with the physical processes of crystallization, crystal crushing, and dissolution, asymmetric amplification u

Full text of DOI:10.1002/anie.200803877

Communications
DOI: 10.1002/anie.200803877
Asymmetric Amplification
Generation of Highly Enantioenriched Crystalline Products in
Reversible Asymmetric Reactions with Racemic or Achiral Catalysts**
Svetlana B. Tsogoeva,* Shengwei Wei, Matthias Freund, and Michael Mauksch*
The generation of enantiomerically pure compounds from
achiral precursors is the dream of any stereochemist. Except
in enzymatic reactions, such a high degree of stereoselectivity
is usually out of reach. Closely related is one of the greatest
unsolved problems of our time in chemistry: the origin of
homochirality in the biosphere; that is, the fact that l-amino
acids and d-sugars dominate in nature, while laboratory
experiments with stereoselective reactions under achiral
conditions only produce a racemic mixture.[1–3] Many scien-
tists assume that the choice of a single chirality occurred
before the advent of life on early earth.[4,5] Offering a
potential solution to this conundrum, in 1953 the crystallog-
rapher Charles Frank first introduced the concept of “sponta-
neous chiral symmetry breaking” in chemistry.[6] In such a
process, one enantiomer is preferentially formed in a kineti-
cally controlled autocatalytic process, even though the initial
conditions were achiral. The first such symmetry-breaking
transformation in chemistry was the homogenous Soai
reaction. This reaction, however, is irreversible and thus
couldnꢀt produce enantiomerically pure products.[7] Saito and
Hyuga first realized that a reaction network must be
reproductive to allow complete chiral symmetry breaking,
resulting in 100% enantiomeric excess.[8]
Another rather different symmetry-breaking mechanism
was reported by Kondepudi et al. for enantiomorphous
crystallization from supersaturated NaClO3 solutions on
cooling under the influence of stirring, which causes secon-
dary nucleation: one crystal handedness always randomly
dominated in the solid product with 99% ee.[9] In 1999,
Kondepudi et al. reported that even an intrinsically chiral,
atropisomeric 1,1’-binaphtyl species is able to crystallize from
the melt in high enantiomeric excess of 80% in a determin-
istic fashion (but with equal probability for the R and S crystal
forms),[10] extending on earlier work of Pincock et al., who
found a near Gaussian distribution of ee values centered
around the racemic outcome.[11] This result was explained by
secondary nucleation in conjunction with autocatalytic crystal
growth; this process was later kinetically modeled by Asakura
and co-workers.[12]
Viedma expanded on this work and obtained completely
monochiral states from previously racemic crystal mixtures of
NaClO3. This result was explained by a combination of
nonlinear autocatalysis and recycling through dissolution of
crystallites spawned off from the mother crystals by grind-
ing.[13–16] Very recently, Noorduin et al. demonstrated con-
vincingly that practically racemic or scalemic conglomerates
of another intrinsically chiral compound, an amino acid
derivative, can fully deracemize by solution-phase enantio-
merization in a base-catalyzed process.[17] Although such a
process offers fascinating prospects for the production of
enantiomerically pure compounds, not all chiral products of
interest might enantiomerize readily, or they could be
sensitive to the enantiomerization reaction conditions (e.g.
giving side products stemming from rearrangements or
elimination reactions). Moreover, products with more than
one stereocenter might deracemize in an undesired way,
because the synthesis of such complex chiral compounds
demands that the chirality at each stereocenter be controlled
individually.
Hence, understanding that the enantiomerization might
also occur through the reactant (that is, involve the reverse
reaction in a reversible stereoselective reaction),[18] we
realized that the product of an enantioselective reaction run
near equilibrium (under thermodynamic control) could read-
ily deracemize under mild conditions, for example, in the
presence of a suitable achiral or racemic catalyst or even
without a catalyst at all. Herein we demonstrate such a
deracemization process by example of a reversible Mannich
type reaction (Scheme 1), and we explain the observed
process by a kinetic rather than a thermodynamic model
that involves recycling by denucleation and nonlinear auto-
catalytic crystal growth.
We recently observed an increase in product ee values
when a slurry of a racemic mixture of the preformed
[*] Prof. Dr. S. B. Tsogoeva, Dr. S.-W. Wei, M. Freund, Dr. M. Mauksch
Department of Chemistry and Pharmacy
Chair of Organic Chemistry I, University of Erlangen-Nuremberg
Henkestrasse 42, 91054 Erlangen (Germany)
Fax: (+49)9131-85-26865
E-mail: tsogoeva@chemie.uni-erlangen.de
Homepage: http://www.chemie.uni-erlangen.de/oc/tsogoeva/
[**] The authors gratefully acknowledge generous financial support from
the Deutsche Forschungsgemeinschaft.
Scheme 1. Deracemization experiments with a Mannich type reaction
carried out in toluene at room temperature with racemic or achiral
catalysts. 4-Br-Bz=4-bromobenzoyl.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200803877.
590
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 590 –594

Angewandte
Chemie
crystalline Mannich product 1[19] was stirred in its saturated
solution with a soluble enantiopure primary amine thiourea
catalyst (3). We assumed that the enantiomerization of the
product in solution could be involved and that it might occur
via the prochiral reactant 2. To test this hypothesis, and
because we recently became aware of the remarkable
observations of Noorduin et al.,[17] we set up a series of
experiments in which we, in effect, made the asymmetric
reaction that we studied earlier[19] run backwards: partly
dissolved mixtures of the crystalline product conglomerates
with varied but low initial ee values were vigorously stirred
together with achiral catalyst 4 or racemic catalyst 3
(Scheme 1), which was added in different amounts (see
Figures 1 and 2 and the Supporting Information). We carried
out experiments both with and without glass beads and at
constant stirring rates (1300 rpm). All the experiments show a
clear trend of asymmetric amplification with progressing
reaction time. All ee values given in Figures 1 and 2 refer to
samples taken from the product slurry.
We observed that the enantiomeric excess of the solution–
solid mixture increased slowly but clearly when the initial
enantioimbalance in the solid Mannich product 1 was already
significant (17.4% ee), for example to 79% ee after 24 days in
experiment 1 in the presence of 15 mol% of rac-3 and without
glass beads (Figure 1). An increase in catalyst loadings from
Figure 2. Enantiomeric excess value versus time determined from the
slurry for the Mannich product 1 in the deracemization process shown
in Scheme 1 (run 8 without glass beads; runs 5–7 and 9 with glass
beads). Runs 5 and 7–9 were carried out with (S)-1; run 6 was
performed with (R)-1.
19.3%): starting product absolute configuration was
conserved during the whole course of the reaction.
Experiment 5 (Figure 2) is essentially a repetition of
run 1 (Figure 1), with nearly identical initial conditions
(17.5 instead of 17.4% initial ee), but glass beads were
present in run 5. Accordingly, results differed between
the two runs, which also implies a probable temperature
influence, which was not controlled for during the
experiments: for example, run 1 gave product with 79%
enantiomeric purity after 24 days, while run 5 had
produced crystalline product with only 34.4% ee (from
slurry) after the same reaction time. We also compared
the results with catalyst 3 and glass-bead-assisted
attrition (i.e. with grinding, runs 2, 5, 6, and 7) with
those runs without glass beads in which the product
crystals were only stirred (1, 3, and 8). In some cases
(e.g. cf. experiments 7 and 8 with 17.4% initial product
ee and 50 mol% catalyst loading), the asymmetric
amplification was faster or more sustained without
glass beads than for experiments in which glass beads
were added to the stirred crystal conglomerate. After 30
days, run 8 gave chiral product in 76.8% enantiomeric
excess without glass beads, while 69.8% ee was obtained in
the concurrent run 7 with glass beads. Notably, the highest
catalyst loadings (runs 7 and 8) do not lead to improvement
over runs 2 and 3. In all experiments we observed formation
of reactant 2 (confirmed by NMR spectroscopy and mass
spectrometry), thus verifying the involvement of enantiomer-
ization via the reverse reaction.
Figure 1. Enantiomeric excess values versus time determined from the slurry
for the Mannich product 1 in the deracemization process shown in Scheme 1
(runs 1 and 3 without glass beads; runs 2 and 4 with glass beads).
15 to 30 mol% (runs 2 and 3) resulted in faster asymmetric
amplification and higher attainable ee values. After final
workup (filtration of crystals and washing with toluene), the
solid-phase ee values for these two runs (2 and 3) were
determined to be 100 and 98.5%, respectively. This finding
demonstrated that even complete homochirality in the solid
state can be achieved by this method. Enantiomeric excess
values in the samples withdrawn from the slurry were about
5% smaller (95.5% ee (2) and 94.6% ee (3), Figure 1) than
those from the solid, indicating a considerably lower solution
ee. Notably, we always observed a deterministic outcome in
the experiments with significant initial product ee (13.9–
To better understand the role of the catalyst, we also
carried out a control experiment without any catalyst (run 9,
Figure 2). Interestingly, we nevertheless measured an increase
in the ee value to 32.6% after 22 days from an initial 19.3%
Angew. Chem. Int. Ed. 2009, 48, 590 –594
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
591

Communications
ee. We explain this finding by the fact that the Mannich
of the experiment, yields of at least 70% were achieved,
which means that the transformation of R-1 into S-1 must
indeed have occurred. We anticipate that the deracemization
might even become quantitative with larger batches and in
the absence of side reactions.
Meyerhofferꢀs double solubility rule for the solubilities of
racemic conglomerates with respect to monochiral crystals (in
analogy to vapor pressures in ideal gases),[22] which had only
recently been controversially discussed for the case of
racemizing conglomerates,[23] appears to hold well for our
system: the solubility of the enantiopure Mannich product 1 is
1.9 gLꢀ1, while it is 4.5 gLꢀ1 for rac-1. The rule predicts that a
racemic conglomerate, consisting of two different species,
should have exactly twice the solubility of a single enantio-
mer.[22] The observed deviation from the rule could be
attributed to the nonideality of the solution.[24]
reaction is already reversible and does not necessarily need
the further assistance of a catalyst. However, another control
experiment with a starting ee value of 0.2% (S) and without
added catalyst gave only 0.5% ee (S) after 32 days. It appears
that significant asymmetric amplification in reasonable reac-
tion times requires significant initial enantioimbalance.
Extended reaction times might even have adverse effects,
because side reactions could become significant, in particular
when the reaction runs for several weeks. The reverse of the
reaction depicted in Scheme 1, for example, gives both the E
and Z forms of hydrazone 2 (the Z/E isomerization is slow).[20]
However, only the E isomer can react in the forward reaction
step to give the Mannich product 1 with catalyst 3.
To see the effect of a different, simpler, and cheaper
catalyst, we also employed pyrrolidine (4). The deracemiza-
tion process is remarkably fast with 4 in the presence of glass
beads (an increase in Mannich product ee value from 13.9 to
60.6% was observed after only five days, see run 4, Figure 1).
While both rac-3 and 4 catalyze the deracemization process
by reversible formation of reactant 2, pyrrolidine might
additionally act as a base, abstracting a proton from the CH-
acidic position in 1 (run 4, Figure 1), which would also result
in deracemization, in analogy to the recently reported DBU-
catalyzed bimolecular enantiomerization of an amino acid
derivative (DBU = 1,8-diazabicyclo[5.4.0]undec-7-ene).[17,21]
Thus, the chiral product 1 might also enantiomerize directly
with pyrrolidine (i.e. without involvement of the reactant 2 in
the process S!R or R!S). To see whether pyrrolidine and/or
DBU could catalyze the Mannich type reaction itself, we
added them to separate reaction mixtures. Intriguingly, we
found that pyrrolidine is able to catalyze the reaction of 2 with
acetone, while DBU is not (no product formation was
detected). Instead, 4-bromo-N-benzoylhydrazine was
formed as a result of the fragmentation of reactant 2.
Notably, some enantioselective reactions that give a chiral
product with a CH-acidic position might even be more
tractable in the direct enantiomerization process than in
enantiomerization via the reactant—for example, when the
back reaction step is very slow (i.e. when practically
irreversible product formation is involved, as, for example,
in some asymmetric epoxidations) or when the desired chiral
product is formed under kinetic control.
We have also continuously monitored the enantiomeric
excess values in the solution and the solid state and compared
the latter with the results from samples taken directly from
the slurry before washing with toluene. We found that seven
days after the start of the experiment (with an initial ee value
of 17.5% and 30 mol% 3), the solid-state ee (or crystal
enantiomeric excess) value of 77.3% was dramatically higher
than the ee values in the sample from the slurry (25.7%). This
finding indicates that the actual deracemization is probably
much faster than the data depicted in Figures 1 and 2 implies.
To find out whether one enantiomer has actually been
transformed into its antipode, we isolated and purified the
crystalline phase only after complete enantiomeric purity was
achieved in a parallel run, to determine the total yield in the
transformed conglomerate. Taking into account the unavoid-
able loss of solid product owing to dissolution at the beginning
The dependence of the enantiomeric excess on the
reaction time in our system is similar to that reported recently
for the stirred slurry of an organic conglomerate, which,
however was achiral in the solution phase.[25] In a logarithmic
plot, the exponential increase of the ee values is clearly
apparent (see the Supporting Information, Figure S1). This
result is in accord with the predictions from two recent models
for the deracemization of racemic conglomerates involving
solution-phase enantiomerization,[17b,26] a transformation first
studied by Havinga in 1941.[27] He reasoned that because of
the faster “deposition” of those molecules from the stagnant
solution that already dominate in the crystal phase, the
transformation from the less abundant to the more abundant
form must be inevitable once the enantiomerization process is
fast enough, thus keeping the solution in its racemic state and
resupplying the enantiomer which is taken away faster from
the solution.[27b] Havinga employed supersaturation as the
driving force in the crystallization process. Recently, it was
proposed that recycling by thermodynamically controlled
dissolution in a vigorously stirred slurry together with kineti-
cally controlled nonlinear autocatalytic nucleation and crystal
growth could explain these phenomena in merely saturated
(not supersaturated) solutions.[28] The crystal growth is
autocatalytic, because the bulk crystal surface must be
involved in it. Crystal growth of the major isomer must be
faster than its dissolution; while the rate of dissolution is
proportional to the mole fraction of the respective crystal
enantiomorph, the crystallization rate should have a more
than linear dependence on the available crystal surface (or the
mole fraction) of that enantiomorph.
Stirring, especially in the presence of glass beads, causes
“secondary nucleation” by spawning off tiny microcrystals or
clusters[16,26,29] from the mother crystals through crushing or
shear forces. The rate of this secondary nucleation for a
specific crystal handedness is obviously proportional to the
grinding rate and to the mole fraction of the respective
enantiopure mother crystals.[30] The microcrystals have the
same handedness as their mother crystals and usually differ in
size among each other. The ensemble of crystals obtained
from vigorous stirring or grinding is permanently at non-
equilibrium. It is fuelled by the influx of mechanical energy, as
the thermodynamic energy content of the microcrystals
increases with decreasing size according to the Gibbs–
592
www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 590 –594

Angewandte
Chemie
Thomson rule.[31] Smaller crystals have a higher interfacial
energy, because their surface is more strongly curved.
Furthermore, the surface tension also has a higher contribu-
tion to the total energy of such a microcrystal because of the
larger surface-to-volume ratio with decreasing size. The
solubility of the smaller microcrystals is therefore enhanced
with respect to the larger ones. This denucleation process[32]
provides a simple and straightforward way of recycling the
crystalline product of the minor enantiomer, exposing it to
potential enantiomerization and recrystallization with the
dominant enantiomorph, thereby offering an explanation for
the observed total conversion of one enantiomer into the
other (run 2, Figure 1).
ee) is about 20 days in a homogenous near-saturated solution
in the presence of three equivalents rac-3, which is a realistic
concentration of the readily soluble catalyst in the deracem-
ization experiments (Figures 1 and 2). It appears, therefore,
that solution-phase enantiomerization cannot be made
responsible for the observed swift asymmetric depletion in
the solution under the influence of grinding or vigorous
stirring in our heterogeneous system. Hence, while the
stagnant solution should have approximately the same
composition as the conglomerate, we assume faster removal
of the crystal phaseꢀs major enantiomer from the solution
owing to nonlinear crystal growth driven by the non-
equilibrium distribution of crystal sizes.
Based on the insight that the only thermodynamically
stable final state should consist of a large monochiral single
crystal with minimal surface energy, it has been proposed
recently that “Ostwald ripening”, in which larger crystals
grow at the cost of smaller ones, could constitute a determin-
istic route towards a final state with single chirality in a
thermodynamically driven process under laboratory condi-
tions.[17,33] In this case, Ostwald ripening would play the role of
supersaturation in providing the driving force for crystalliza-
tion in the non-equilibrium system. However, the authors
themselves noted[17b] that without further assumptions, their
model does not correctly predict the outcome for compounds
which are achiral in solution (i.e. with an infinite rate of
enantiomerization), as, for example, in the original Viedma
experiment.[13,17b]
To provide a plausible model of the required nonlinear
growth, Uwaha proposed, on the basis of earlier expect-
ations,[12,14,29,34] that crystal growth is partly nurtured from
coalescence of chiral clusters with the corresponding bulk
crystals, resulting in a nonlinear autocatalytic feedback loop
with a kinetic instability of the racemic state and a bifurcation
in the crystal ee values, while the solution becomes and
remains exactly racemic for sufficiently high enantiomeri-
zation rates.[26] A central feature of the Uwaha model is the
temporary excess in the minor-enantiomer concentration in
solution during the symmetry-breaking transition when
enantiomerization in the solution is slow.[35] Experimentally,
we have indeed found that vigorously grinding a slurry of
scalemic conglomerate of S-1 with 17.5% initial solid–state ee
for 19 h without catalyst resulted in a crystal ee increase to
33.7% (after washing with toluene), while the solution-phase
ee value is a mere 0.6% R. After 60 h, the solid-state ee value
was essentially unchanged (32.5% ee). This experiment was
repeated several times and with different initial solid-state ee
values. In all cases, we observed rapid asymmetric depletion
in the solution, which is usually accompanied by a tiny but
significant excess of the solid phaseꢀs minor enantiomer.
In a subsequent deracemization experiment in the pres-
ence of catalyst 3, we again observed that the solution phase
remained nearly racemic most of the time, with an inter-
mittent sudden increase in the minor formꢀs enantiomeric
excess up to 21% ee (R), which might be due to a very slow
enantiomerization process.[26,27] Indeed, the observed rate of
racemization in the homogenous solution is surprisingly low
and is found to be only slightly increased by the presence of
the catalyst 3. The half-life of nearly enantiopure 1 (at 99%
The observed low racemization rate also made us wonder
whether a polymorphic transition (e.g. induced by grinding)
from a metastable racemic phase to a potentially more stable
enantiomerically pure phase could be involved in our
deracemization experiments.[36] However, in light of the
solid evidence for the involvement of 3 both in the forward
reaction[19] and in the whole transformation process (Figures 1
and 2 and the Supporting Information), and because of the
absence of evidence for the existence of a metastable racemic
phase in the solubility data, we consider a kinetically
controlled crystal growth process in combination with an
assisted enantiomerization involving the solution phase in the
heterogeneous system to be the most probable explanation
for our results. Furthermore, the observed first-order kinetics
of the asymmetric amplification appear to fit better into a
kinetic than a thermodynamic model. To reconcile the
observation of a very low enantiomerization rate in the
homogenous solution with the swiftness of the transformation
process under heterogeneous conditions, we assume that the
enantiomerization occurs much more rapidly at or near the
crystal–solution interface,[37] in contrast to the results
reported recently.[17]
In conclusion, the combination of stirring (with or without
grinding) of preformed crystalline conglomerates of chiral
products (with initially low enantiomeric excess values) of
asymmetric reactions (e.g. the reversible Mannich type
reaction studied herein) with enantiomerization through the
reverse reaction might be a very promising new methodology
to attain highly enantioenriched products (up to 100% ee).
For the first time it was shown that glass-bead-assisted
attrition is not a necessary ingredient. The enantiomerization
through the back reaction of a reversible asymmetric reaction
could be a complement to the process revealed by Noorduin
et al.,[17] thus extending the scope of exploitable chemical
processes from direct enantiomerizations to reversible asym-
metric reactions. Application of this concept to asymmetric
transformations that provide only poor product enantiomeric
excess values or conversion rates because of their reversibility
might prove fruitful by converting such processes into
valuable methods to give highly enantioenriched products.[38]
Further mechanistic investigations are necessary to elucidate
the exact nature of the transformation process reported
herein. The challenge remains to accelerate the transforma-
Angew. Chem. Int. Ed. 2009, 48, 590 –594
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
593

Communications
[19] D. A. Yalalov, S. B. Tsogoeva, T. E. Shubina, I. M. Martynova, T.
Clark, Angew. Chem. 2008, 120, 6726; Angew. Chem. Int. Ed.
2008, 47, 6624.
tion process to make it more suitable as a method of
deracemization for broad practical use.
[20] “Fast Isomerisation About Double Bonds”: H.-O. Kalinowski,
H. Kessler, Topics of Stereochemistry, Vol. 7 (Eds.: N. L.
Allinger, E. L. Eliel), Wiley, New York, 1973.
[21] We have carried out a similar experiment with the Michael
product from the reaction of trans-b-nitrostyrene and acetone
with either added pyrrolidine or DBU. After 12 days we
observed asymmetric depletion in the product slurry with
DBU (from 17.3 to 11.0% ee) but asymmetric amplification
(from 17.3 to 20.4% ee) with pyrrolidine.
[22] W. Meyerhoffer, Ber. Dtsch. Chem. Ges. 1904, 37, 2604.
[23] a) D. G. Blackmond, Chem. Eur. J. 2007, 13, 10306; b) J. Crusats,
S. Veintemillas-Verdaguer, J. M. Ribo, Chem. Eur. J. 2006, 12,
7776.
[24] J. S. Chickos, D. G. Hesse, Struct. Chem. 1991, 2, 33.
[25] P. Shan Monica Cheung, J. Gagnon, J. Surprenant, Y. Tao, H. Xu,
L. A. Cuccia, Chem. Commun. 2008, 987.
Received: August 6, 2008
Revised: October 13, 2008
Published online: December 12, 2008
Keywords: asymmetric amplification · chiral resolution ·
.
chirality · crystal growth · nucleophilic addition reactions
[1] M. Calvin, Chemical Evolution, Oxford University Press, 1969,
Oxford.
[2] J. Podlech, Cell. Mol. Life Sci. 2001, 58, 44.
[3] a) N. Fujii, T. Saito, Chem. Rec. 2004, 4, 267; b) S. Pizzarello, Acc.
Chem. Res. 2006, 39, 231; c) J. Cohen, Science 1995, 267, 1265;
d) J. Bailey, Acta Astronaut. 2000, 46, 627; e) J. L. Bada, Nature
1995, 374, 594.
[4] W. A. Bonner, Origins Life Evol. Biosphere 1991, 21, 59.
[5] a) V. Avetisov, V. Goldanskii, Proc. Natl. Acad. Sci. USA 1996,
93, 11435; b) V. A. Avetisov, S. A. Anikin, V. I. Goldanskii, V. V.
Kuzmin, Dokl. Akad. Nauk SSSR 1985, 282, 115.
[6] F. C. Frank, Biochim. Biophys. Acta 1953, 11, 459.
[7] a) K. Soai, T. Shibata, H. Morioka, K. Choji, Nature 1995, 378,
767; b) K. Soai, I. Sato, T. Shibata, S. Komiya, M. Hayashi, Y.
Matsueda, H. Imamura, T. Hayase, H. Morioka, H. Tabira, J.
Yamamoto, Y. Kowata, Tetrahedron: Asymmetry 2003, 14, 185;
c) T. Kawasaki, K. Suzuki, M. Shimizu, K. Ishikawa, K. Soai,
Chirality 2006, 18, 479; d) D. A. Singleton, L. K. Vo, Org. Lett.
2003, 5, 4337.
[8] Y. Saito, H. Hyuga, J. Phys. Soc. Jpn. 2004, 73, 33.
[9] D. K. Kondepudi, R. Kaufman, N. Singh, Science 1990, 250, 975.
[10] D. K. Kondepudi, J. Laudadio, K. Asakura, J. Am. Chem. Soc.
1999, 121, 1448.
[11] R. E. Pincock, R. R. Perkins, A. S. Ma, K. R. Wilson, Science
1971, 174, 1018.
[26] M. Uwaha, J. Phys. Soc. Jpn. 2008, 77, 083802.
[27] a) E. Havinga, Chem. Weekbl. 1941, 38, 642; b) See also: E.
Havinga, Biochim. Biophys. Acta 1954, 13, 171.
[28] a) C. Viedma, Astrobiology 2007, 7, 312; b) D. G. Blackmond,
Chem. Eur. J. 2007, 13, 3290; c) M. Lahav, I. Weissbuch, E.
Shavit, C. Reiner, G. J. Nicholson, V. Schurig, Origins Life Evol.
Biosphere 2006, 36, 151.
[29] Y. Saito, H. Hyuga, J. Phys. Soc. Jpn. 2005, 74, 2601.
[30] B. Kaptein, W. L. Noorduin, H. Meekes, W. J. P. van Enckevort,
R. M. Kellogg, E. Vlieg, Angew. Chem. 2008, 120, 7336; Angew.
Chem. Int. Ed. 2008, 47, 7226.
[31] J. W. Gibbs, Thermodynamics, Collected works, Vol. 1, Yale
University Press, New Haven, 1948.
[32] J. H. E. Cartwright, O. Piro, I. Tuval, Phys. Rev. Lett. 2007, 98,
165501.
[33] a) W. Ostwald, Lehrbuch der Allgemeinen Chemie, Vol. 2, Part 1,
Engelmann, Leipzig, 1896; b) W. Ostwald, Z. Phys. Chem. 1900,
34, 495.
[12] K. Asakura, Y. Nagasaka, S. Osanai, D. K. Kondepudi, J. Phys.
Chem. B 2005, 109, 1586.
[34] Asakura, studying a system crystallizing from the unstirred
supercooled melt, assumed that “secondary nucleation” resulted
from the spontaneous formation of chiral clusters from the melt
in the vicinity of the bulk crystal surface, see reference [12] and
K. Asakura, Y. Nagasaka, M. Hidaka, M. Hayashi, S. Osanai,
D. K. Kondepudi, Chirality 2004, 16, 131.
[13] C. Viedma, Phys. Rev. Lett. 2005, 94, 065504.
[14] The Viedma experiment has been interpreted by Uwaha,
employing a kinetic model in computer simulations and involv-
ing “minimal chiral units”, see: M. Uwaha, J. Phys. Soc. Jpn.
2004, 73, 2601.
[15] a) In a recent theoretical study, we showed that linear autoca-
talysis in a closed microreversible recycling network can result in
fully chiral stationary-state solutions in open-flow and homoge-
nous systems, see: M. Mauksch, S. B. Tsogoeva, ChemPhysChem
2008, 9, 2359; b) for a similar mechanism, which employs an
irreversible recycling activation step fuelled by an external
energy source, see: R. Plasson, J. Phys. Chem. B 2008, DOI:
10.1021/jp803588z.
[35] For very low or even zero solution-phase enantiomerization
rates, phase disproportionation of enantiomers might become an
additional factor. Furthermore, factors that govern the solubility
of the conglomerate (like the nature of solvent and the
temperature, see for example references [17b] and [30]) could
also crucially affect the effectiveness of the transformation.
[36] The “preferential enrichment” of racemic compounds by a
polymorphic transition involves a near enantiopure solution, in
contrast to the observations for our system, see: R. Tamura, H.
Takahashi, U. Takanori, J. Chem. Soc. Jpn. 2001, 2, 71.
[37] While this manuscript was under review, a paper was submitted
which describes a mechanism for the generation of homochir-
ality by grinding and an autocatalytic enantiomerization at the
crystal surface Y.Saito, H. Hyuga, J. Phys. Soc. Jpn. 2008, DOI:
arXiv:0810.0910v1.
[16] J. H. E. Cartwright, J. M. Garcia-Ruiz, O. Piro, C. I. Sainz-Diaz,
I. Tuval, Phys. Rev. Lett. 2004, 93, 035502.
[17] a) W. L. Noorduin, T. Izumi, A. Millemaggi, M. Leeman, H.
Meekes, W. J. P. van Enckevort, R. M. Kellog, B. Kaptein, E.
Vlieg, D. G. Blackmond, J. Am. Chem. Soc. 2008, 130, 1158;
b) W. L. Noorduin, H. Meekes, A. A. C. Bode, W. J. P. van Enck-
evort, B. Kaptein, R. M. Kellog, E. Vlieg, Cryst. Growth Des.
2008, 8, 1675; c) W. L. Noorduin, H. Meekes, W. J. P. van Enck-
evort, A. Millemaggi, M. Leeman, B. Kaptein, R. M. Kellogg, E.
Vlieg, Angew. Chem. 2008, 120, 6545; Angew. Chem. Int. Ed.
2008, 47, 6445.
[38] The method described herein might be compared with the
recently demonstrated symmetry breaking in homogenous
reversible reactive systems closed to matter flow, see: M.
Mauksch, S. B. Tsogoeva, S.-W. Wei, I. M. Martynova, Chirality
2007, 19, 816.
[18] M. Mauksch, S. B. Tsogoeva, I. M. Martynova, S.-W. Wei,
Angew. Chem. 2007, 119, 397; Angew. Chem. Int. Ed. 2007, 46,
393.
594
www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 590 –594