Complete deracemization by attrition-enhanced Ostwald ripening elucidated

Article abstract of DOI:10.1002/anie.200801846

(Chemical Equation Presented) Shaken and stirred: The emergence of single chirality in the solid state during grinding of a slurry has been kinetically studied for a phenylglycine amide (see scheme). The insight obtained into the underlying process of attrition-enhanced Ostwald ripening enables the definition of suitable conditions to increase the deracemization rate drastically.

Full text of DOI:10.1002/anie.200801846

Angewandte
Chemie
DOI: 10.1002/anie.200801846
Homochirality
Complete Deracemization by Attrition-Enhanced Ostwald Ripening
Elucidated**
Wim L. Noorduin, Hugo Meekes, Willem J. P. van Enckevort, Alessia Millemaggi,
Michel Leeman, Bernard Kaptein,*Richard M. Kellogg,*and Elias Vlieg*
Direct resolution by crystallization of racemic mixtures,
induced by seeding or by adding tailor-made additives,
constitutes an attractive alternative to separation using
diastereomeric salts.[1–4] Unfortunately, the yield in a single
resolution step is low and the overall yield limited to 50% of
eachenantiomer. Combination of a direct resolution with
racemization in the solution permits a total enantiomeric
transformation.[5–14] Complete resolution in a single operation
would be of great practical use but examples have been
limited. Recently, Viedma has demonstrated the total reso-
lution of an intrinsically achiral molecule, NaClO3, by means
of abrasive grinding.[15,16]
Inspired by Viedma,[15] proof of principle was given witha
Scheme 1. Equilibria involved in the attrition-enhanced crystallization/
dissolution for compound 1.
solid phase of a chiral amino acid derivative in contact witha
solution in which racemization occurs; this evolved smoothly
to a single chiral solid end state on abrasive grinding.[17] This
was demonstrated for the conglomerate N-(2-methylbenzyli-
dene)-phenylglycine amide (1) (Scheme 1). Starting the
process with a mixture slightly enriched in one enantiomer
in the solid phase, the solid phase is completely converted into
the major enantiomer through the solution-phase racemiza-
tion reaction. Within a few weeks either solid enantiomer can
be obtained quantitatively with ee > 99.9%.
process as derived from theory by varying the experimental
conditions. The insight obtained also allows definition of
suitable conditions to increase the deracemization rate
drastically.
The Monte Carlo simulation model is expressed in terms
of the experimental parameters. In the model three param-
eters are crucial. First, the rate at which molecules racemize in
solution is described by the racemization efficiency c. Second,
the continuous ablation of crystals is determined by an
attrition probability z, and, third, the growth of larger crystals
at the cost of smaller proceeds according to the Ostwald
ripening probability k. The simulation results confirmed the
experimentally observed exponential behavior typical for
asymmetric autocatalysis of the time evolution of the
enantiomeric excess ee in the solid state according to
Equation (1)[20] for not too small values of the initial
enantiomeric excess ee(0).[21]
Monte Carlo simulations led to a model that showed that
two processes are responsible for the deracemization: con-
tinuous attrition of crystals and Ostwald ripening, which leads
[18–20]
to growthof large crystals at the cost of smaller ones.
Here we explore the parameters of this deracemization
[*] W. L. Noorduin, Dr. H. Meekes, Dr. W. J. P. van Enckevort,
Prof. Dr. E. Vlieg
IMM Solid State Chemistry, Radboud University Nijmegen
Heyendaalseweg 135, 6525 AJ Nijmegen (The Netherlands)
Fax: (+31)24-365-3067
eeðtÞ ¼ eeð0Þ expðktÞ
ð1Þ
E-mail: bernard.kaptein@dsm.com
A. Millemaggi, Dr. B. Kaptein
Innovative Synthesis & Catalysis, DSM Pharmaceutical Products
PO Box 18, 6160 MD Geleen (The Netherlands)
Fax: (+31)46-476-7604
The deracemization rate constant k (see Supporting
Information) can be expressed in terms of the three param-
eters [Eq. (2)], where the proportionality parameter n0 is an
event frequency and bsolid the amount of solid material.[22]
E-mail: r.m.kellogg@syncom.nl
M. Leeman, Prof. Dr. R. M. Kellogg
Syncom BV, Kadijk 3, 9747 AT Groningen (The Netherlands)
Fax: (+31)50-575-399
n0
k ¼
ckz
ð2Þ
bsolid
E-mail: e.vlieg@science.ru.nl
[**] We acknowledge Prof. D. G. Blackmond for the stimulating
discussions. The SNN agency (Cooperation Northern Netherlands)
and the European Fund for Regional Development (EFRO) are
acknowledged for financial support.
The racemization efficiency via the solution, c, depends
on the amount of the racemization catalyst, brc. Assuming a
linear relationship we can write c = ac,solv brc, where the
proportionality constant ac,solv depends on the solvent used.
The Ostwald ripening probability k is determined by the
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200801846.
Angew. Chem. Int. Ed. 2008, 47, 6445 –6447
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6445

Communications
solute exchange frequency between the solution and the
parameters fixed. As expected, k is inversely proportional to
bsolid
crystallites, which in turn depends on the surface free energy
of the crystallites in the solution. Nielsen and Söhnel have
shown that a logarithmic relation exists between the surface
free energy and the solubility.[23] Using this relation, k can be
shown to be proportional to the solubility bsolute, according to
k = aOst bsolute (see Supporting Information). The attrition
probability z is not further specified in experimental terms,
as this depends on complicated parameters like the stirring
speed, the grinding agents used and the crystal hardness.[24]
Combination of all the relations leads to Equation (3) for the
deracemization rate constant k:
.
To study the effect of the solvent, toluene was used as a
low solubility solvent (0.4 wt%) and tetrahydrofuran (THF)
as a good solvent (6.8 wt%), as compared to acetonitrile
(MeCN) (2.2 wt%, all at 258C). In toluene, after 30 days the
enantiomeric excess in the solid phase was merely 10%
(Figure 3). As expected, for the solvents with a higher
solubility, MeCN and THF, the deracemization is faster and
complete chiral purity is reached after 21–24 days.
n0
k ¼
ac,solv brc aOst bsolute z
ð3Þ
bsolid
Using this equation we have studied the dependencies of
the deracemization rate experimentally (see Experimental
Section). Deracemization experiments were performed for
various molalities brc of the base DBU (1,5-diazabicyclo-
[5.4.0]undec-5-ene) using MeCN as a solvent (Figure 1a).[25]
Figure 3. Evolution of lnee in the solid phase for the deracemization
~
&
*
experiments in THF ( ), MeCN ( ), and toluene ( ).
The measured solubilities bsolute, the solvent dependent
racemization parameters ac,solv, the amount of solid material
bsolid, and the normalized deracemization rate constant ktheo
,
calculated using Equation (3), are given in Table 1; z is kept
Figure 1. a) Evolution of the solid-phase ee in MeCN for differeÀn1t
!
~
&
*
amounts of DBU: 0.42 ( ), 0.25 ( ), 0.13 ( ) and 0.04 molkg ( )
(lines are guide to the eye). b) Rate constant k obtained from the data
in (a) according to Equation (1).
Table 1: Comparison of the theoretical deracemization rate ktheo and the
experimentally observed deracemization rate kexp
.
[a,b]
[a,c]
[a,d]
bsolute
bsolid
brc
ac,solv
ktheo
kexp
MeCN
THF
Toluene
0.088
0.287
0.018
0.387
0.390
0.204
0.0573
0.0409
0.0177
1.00
0.23
0.31
1.00
0.53
0.04
1.00
0.77
0.16
Fitting Equation (1) to the data shows that the rate constant k
increases linearly withincreasing brc (Figure 1b).
Figure 2 shows the dependence of the deracemization rate
on the amount of solid material, bsolid, in MeCN withall other
[a] Normalized for MeCN. [b] The solution racemization rates were
determined in saturated solutions for the three different solvents without
any solid present (Supporting Information). [c] Calculated using Equa-
tion (3). [d] Determined from the data in Figure 3.
constant. The decreasing trend in the experimental rate
constants for the three solvents is also found for the
theoretical values. These results show that for practical
applications the efficiency increases on use of a good solvent
but that the solution racemization rate must be considered as
well. On comparison of MeCN and THF, the latter has the
higher solubility but the former has a higher racemization rate
constant (Supporting Information) resulting in the highest
deracemization rate in MeCN.
All experiments described thus far were performed at
constant stirring rates and amounts of glass beads. This results
in an approximately constant crystal size distribution (CSD)
Figure 2. The deracemization constant k as a function of the amount
of solid, bsolidÀ1, in MeCN (35 g) with 0.13 molkgÀ1 DBU, 10 g of glass
beads and stirring at 1250 rpm.
6446 www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 6445 –6447

Angewandte
Chemie
Keywords: amino acids · asymmetric amplification ·
.
chiral resolution · chirality · grinding
[1] H. Murakami, Top. Curr. Chem. 2007, 269, 273.
[2] J. Jacques, A. Collet, S. H. Wilen, Enantiomers, Racemates and
Resolution, Krieger, Florida, 1994.
[3] L. Addadi, Z. Berkovitch-Yellin, N. Domb, E. Gati, M. Lahav, L.
Leiserowitz, Nature 1982, 296, 21.
[4] L. Addadi, S. Weinstein, E. Gati, I. Weissbuch, M. Lahav, J. Am.
Chem. Soc. 1982, 104, 4610.
[5] a) E. Havinga, Chem. Weekbl. 1941, 38, 642; b) E. Havinga,
Biochim. Biophys. Acta 1954, 13, 171.
[6] D. H. R. Barton, G. W. Kirby, J. Chem. Soc. 1962, 806.
[7] R. E. Pincock, R. R. Perkins, A. S. Ma, K. R. Wilson, Science
1971, 174, 1018.
[8] J. Van Mil, L. Addadi, M. Lahav, Tetrahedron 1987, 43, 1281.
[9] S. N. Black, L. J. Williams, R. J. Davey, F. Moffatt, R. V. H. Jones,
D. M. McEwan, D. E. Sadler, Tetrahedron 1989, 45, 2677.
[10] D. K. Kondepudi, R. J. Kaufman, N. Singh, Science 1990, 250,
975.
Figure 4. a) Evolution of ee as a function of the amount of glass beads
~
~
&
*
*
in grams: 2.5 ( ), 5.0 ( ), 7.5 ( ), 10.0 (&), 15.0 ( ), 20.0 g ( ), for a
constant amount of DBU (0.25 molkgÀ1), 4 g of 1 in MeCN (35 g), and
stirring at 1250 rpm (left). Additionally the deracemization was per-
formed under these conditions with 15.0 g glass beads using an
!
ultrasonic bath ( ). b) Rate constant k as a function of the amount of
glass beads obtained from the results in (a) according to Equation (1)
for a constant amount of DBU (0.25 molkgÀ1) in MeCN. The lines are
a guide to the eye.
[11] B. Kaptein, T. R. Vries, J. W. Nieuwenhuijzen, R. M. Kellogg,
R. F. P. Grimbergen, Q. B. Broxterman, Handbook of Chiral
Chemicals, CRC, Taylor & Francis, New York, 2006, pp. 97 – 116.
[12] R. Yoshioka, Top. Curr. Chem. 2007, 269, 83.
[13] K. M. J. Brands, A. J. Davies, Chem. Rev. 2006, 106, 2711.
[14] D. A. Chaplin, N. B. Johnson, J. M. Paul, G. A. Potter, Tetrahe-
dron Lett. 1998, 39, 6777.
[15] C. Viedma, Phys. Rev. Lett. 2005, 94, 065504.
[16] See also P. S. M. Cheung, J. Gagnon, J. Surprenant, Y. Tao, H.
Xu, L. A. Cuccia, Chem. Commun. 2008, 987.
[17] W. L. Noorduin, T. Izumi, A. Millemaggi, M. Leeman, H.
Meekes, W. J. P. van Enckevort, R. M. Kellogg, B. Kaptein, E.
Vlieg, D. G. Blackmond, J. Am. Chem. Soc. 2008, 130, 1158.
[18] C. Viedma, Astrobiology 2007, 7, 312.
[19] J. H. E. Cartwright, O. Piro, I. Tuval, Phys. Rev. Lett. 2007, 98,
165501.
[20] W. L. Noorduin, H. Meekes, A. A. C. Bode, W. J. P. van Enck-
evort, B. Kaptein, R. M. Kellogg, E. Vlieg, Cryst. Growth Des.
2008, 8, 1675.
[21] The simulations show that in the case of instantaneous race-
mization in the solution the deracemization process becomes
stochastic. Here, we limit the discussion to relatively low
racemization efficiencies (c ꢀ 0.1, Supporting Information) and
nonzero values for ee(0), for which the process becomes
deterministic.
during the process. The computer simulations, however,
predict that the deracemization process can be enhanced by
applying intense attrition conditions leading to a CSD of
smaller crystals.[20] Therefore, the influence of the attrition on
the deracemization time was explored by varying the amount
of glass beads. As expected, the deracemization time
decreases for increasing amounts of glass beads as the crystals
are ground to smaller sizes (Figure 4).[26] The shape of the
curve shows that the relation between the amount of glass
beads and k in Equation (3) is rather complicated. Note that
even in the absence of glass beads, the magnetic stirrer will
cause attrition of crystals. To intensify the grinding even more,
we repeated the experiment with 15 g glass beads (see
Figure 4), now using a thermostated standard ultrasonic
bathinstead of magnetic stirring. These experiments led to
an approximately five times higher rate with k = 3.9 dÀ1,
reducing the deracemization time to a single day (Figure 4).
As predicted from Equation (1) a further decrease in the
deracemization time can be realized by starting witha solid
state that is already enriched, instead of beginning with
almost racemic material.[27]
These results offer a detailed understanding of the
processes involved in attrition-enhanced Ostwald ripening
leading to an enantiomerically pure solid phase. Attrition-
enhanced Ostwald ripening is a valuable new technique to
achieve single step conversion rapidly with virtually 100%
yield of the solid phase.[28]
[22] Throughout the paper we express amounts in terms of molalities
(mol per kg solvent).
[23] A. E. Nielsen, O. Söhnel, J. Cryst. Growth 1971, 11, 233.
[24] H. Briesen, Powder Technol. 2007, 178, 87 – 98.
[25] The linearity of the relationship between the solution racemi-
zation time and the DBU concentration was also checked
(Supporting Information).
[26] The computer simulations show a minimum in the deracemiza-
tion time as a function of the attrition probability z, which was
found at high attrition rates, for which the CSD is dominated by
the minimal crystal size that can be obtained by attrition
(Supporting Information).
Experimental Section
Solution-solid mixtures of
1 (4 g) were magnetically stirred
[27] Starting with ee(t=0) ¼
6
0 also suppresses stochastic behavior at
(1250 rpm) at ambient temperature in a given solvent (36 g) in the
presence of 2.5 mm glass beads (10 g). After establishing solution–
solid equilibrium (24 h), solution-phase racemization was initiated by
adding 200 mg DBU (ca. 0.04 molkgÀ1). Samples of the solid were
collected over time and the enantiomeric purity was measured using
two independent chiral HPLC methods.[17]
low ee values.[20]
[28] As expected from Equation (3), scaling up of all parameters in
the experiment has no influence on the deracemization rate. This
was verified by performing experiments at a five times larger
scale (ca. 20 g (R,S)-1, 0.25 molkgÀ1 DBU, 50 g glass beads, 175 g
MeCN). The enantiopure end state was again reached in 7 days
(cf. Figure 1).
Received: April 21, 2008
Published online: July 11, 2008
Angew. Chem. Int. Ed. 2008, 47, 6445 –6447
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 6447