Enantioenrichment by Iterative Retro-Aldol/Aldol reaction catalyzed by an achiral or racemic base

Article abstract of DOI:10.1002/chem.200903497

Chemical Equation Presented Did nature use the same strategy? The combination of crystal growth and iterative retro-aldol/aldol reactions catalyzed by a simple achiral (or racemic) base allows improvement of the enantiomer ratio of a product obtained from a stereochemically imperfect aldol reaction (see scheme).

Full text of DOI:10.1002/chem.200903497

DOI: 10.1002/chem.200903497
Enantioenrichment by Iterative Retro-Aldol/Aldol Reaction
Catalyzed by an Achiral or Racemic Base
Angelika M. Flock, Christine M. M. Reucher, and Carsten Bolm*[a]
Direct resolution of racemic or scalemic conglomerates by
crystallization has recently emerged as an alternative to the
separation of enantiomers by means of an additional chiral
compound such as the crystallization as diastereomeric salt
or inhibition of one of the enantiomorph’s crystallisation by
addition of tailor-made additives.[1,2] In contrast to the latter
techniques, the combination of resolution by crystallization
and racemization in solution allows to obtain the enantio-
pure compound in up to quantitative yield.
By means of abrasive grinding Viedma achieved a com-
plete resolution of intrinsically achiral NaClO3 starting from
a racemic mixture of the two enantiomorphic solid.[3] In this
context, the term “chiral amnesia” was introduced in order
to describe the loss of the solid-phase chiral identity (as l or
d crystal) during dissolution.[4] The resulting racemization
and the continuous preferential crystallization of the enan-
tiomer that already dominates the solid phase was suggested
to be the driving force for establishing homochirality in the
solid phase as it enables the transformation of one solid
enantiomer into the other via the solution phase.[4,5]
Several models have been developed to explain the dera-
cemization process under grinding. Kondepundi proposed
that in supersaturated solutions or melts secondary nuclea-
tion in conjunction with autocatalytic crystal growth is the
driving force for the amplification process.[6] Ribꢀ stated
that the autocatalytic crystal growth is driven by chiral inter-
actions between chiral aggregates in solution and the solid
phase.[7] These interactions, which also direct the aggrega-
tion of two or more molecules, are favorable between homo-
chiral partners and destructive between partners of different
chirality. In this scenario the system is striving for the ther-
modynamically most stable state, which is homochirality.
Also in the absence of supersaturation the trend towards ho-
mochirality is thermodynamically controlled and caused by
two interplaying processes: continuous attrition of crystals
by the grinding and Ostwald ripening.[5,8,9] The latter states
that large crystals grow at the cost of smaller ones. Driving
force for this process is the higher solubility of small crystals
compared to larger ones due to their higher surface to
volume ratio (Gibbs–Thompson effect).[10] Continuous
grinding of the solid resupplies the number of thermody-
namically less stable small crystals and amplifies thus the
Ostwald ripening effect. A drawback of this theory is, how-
ever, that it cannot explain the exponential amplification of
the initial enantiomeric excess that has been observed in
Viedma’s experiment.[3,11] Uwaha and Saito proposed micro-
scopic models that describe non-linear kinetics for the auto-
catalytic chiral crystal growth.[11,12] Both models are based
on the assumption that subcritical chiral clusters are formed
by the grinding that can only be integrated into crystals of
the same chirality. In Saito’s model the transformation of
one enantiomer into the other occurs directly on the crystal
surface and is thus an autocatalytic process as well.
Recently, Noorduin et al. demonstrated that also nearly
racemic or scalemic mixtures of intrinsically chiral com-
pounds can quantitatively be resolved by means of the
aforementioned grinding–racemization technique.[13,14] Since
in this case the molecules do not lose their chiral identity
upon dissolution, base-catalyzed racemization in solution via
an achiral intermediate was the key to success. The protocol
was shown to be effective even for epitaxial racemic con-
glomerates.[15] Using the example of naproxen it was shown
that the deracemization time could significantly be reduced
when the enantiomeric excess of the solid was kept high by
a gradual in situ feed of racemic material.[16]
In a third type of deracemization, the forward- and back-
reaction steps of a reversible organic reaction are key for
the solution-phase racemization.[17] By using pyrrolidine as
catalyst, Tsogoeva and co-workers generated reaction condi-
tions that allowed to transform one enantiomeric Mannich
product into the other via the reactants. Also in this system
both abrasive grinding and the physical processes of aggre-
gation and crystal growth provide the driving force for com-
plete deracemization of the solid phase. From the observa-
[a] A. M. Flock, C. M. M. Reucher, Prof. Dr. C. Bolm
Institut fꢁr Organische Chemie der RWTH Aachen University
Landoltweg 1, 52056 Aachen (Germany)
Fax : (+49)241-8092391
E-mail: carsten.bolm@oc.rwth-aachen.de
3918
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 3918 – 3921

COMMUNICATION
tion that the deracemization was faster than the racemiza-
tion in solution, Saito’s prediction[12] that the former process
took place at the crystal/solution interface was verified ex-
perimentally.
Inspired by Tsogoeva’s results and following our previous
studies on the aldol reaction under grinding conditions, we
report here an enantioenrichment of a scalemic aldol prod-
uct via iterative retro-aldol/aldol reactions in presence of an
achiral or racemic catalyst (Scheme 1).[18]
Experiments performed in the absence of the secondary
amine showed that, in contrast to Tsogoeva’s observations
in Mannich reactions, no enantioenrichment occurred and
that simple mixing of 2 and 3 in DMSO did not lead to
aldol product 1. Both results suggested that the base was es-
sential and that 1 did not catalyze its own retro-aldol reac-
tion.[19]
In all cases, aldol product anti-1 was accompanied by
small quantities of the corresponding aldol condensation
product, significant amounts of
the syn diastereomer (ca. 30%)
and traces of other stereoiso-
mers of 1. Use of piperidine in-
stead of pyrrolidine avoided the
formation of the condensation
product.[20] Assuming that the
syn product resulted from a low
diastereoselectivity of the pyr-
rolidine- or piperidine-cata-
lyzed aldol reactions, rac-pro-
line was applied as organocata-
lyst instead. Starting from an
Scheme 1. Catalytic enantioenrichment of aldol product anti-1 in presence of achiral or racemic secondary
amines.
For the proof-of-concept experiment aldol product anti-1,
which is known to form a conglomerate in the solid phase,
was selected as starting material. In solution it can racemize
in the presence of a catalytic secondary amine via the reac-
tants 4-tert-butylcyclohexanone (2) and 4-nitrobenzaldehyde
(3). Accordingly, solid-solution mixtures of scalemic anti-1
were magnetically stirred (800 rpm) in DMSO in the pres-
ence of ZrO2 beads as grinding medium. After 30 min, pyr-
rolidine (10 mol%) was added to provide racemization con-
ditions in solution. Samples of the slurry were taken out
over time, and the enantiomeric ratio (e.r.) of anti-1 was de-
termined by CSP-HPLC.
As we had hypothesized, the e.r. of anti-1 in the slurry
changed significantly over time (Figure 1). Starting from an
initial 85:15 ratio it increased to 96:4 after one day, followed
by 97.5:2.5 and 98:2 after two and eleven days, respectively.
The enantioenrichment appeared to be fast in the beginning
and then slowed down substantially. When the initial e.r. of
the aldol product was 80:20, it raised to 95:5 after only two
days and to 97:3 after seven days. In Figure 1 these data are
presented as ee versus time.
initial imbalanced e.r. of 85:15, aldol product anti-1 with an
e.r. of 97.5:2.5 was obtained after 42 days (Figure 1). As hy-
pothesized, less syn aldol product was formed, but apparent-
ly, the enantioenrichment was rather slow under those con-
ditions. Finally, column chromatography allowed to isolate
aldol product 1 with an anti/syn ratio of 80:20 and an e.r. of
>99:1 for the major diastereomer in 87% yield.[21]
In order to confirm that the enantioenrichment involved
the proposed retro-aldol reaction, a competition experiment
with cyclohexanone (4) as co-reactant was performed
(Scheme 2). Thus, in the presence of an excess of cyclohexa-
none (2.4 equiv) and with pyrrolidine as catalyst, 43% of
anti-1 (with an e.r. of 85:15) was converted into aldol prod-
uct anti-5 after six days. Interestingly, the e.r. of 1 was
higher than without cyclohexanone (99:1). In contrast, anti-5
was racemic, as expected.
That the current racemization conditions were favorable
for the formation of racemic aldol product anti-5 was also
shown in an experiment starting from a solid–solution mix-
ture of anti-5 having an e.r. of 97:3 (in absence of 1). As hy-
pothesized, the e.r. decreased to give anti-5 with a residual
e.r. of only 54.5:45.5.[22,23]
In conclusion, we demonstrated that an aldol product that
crystallizes as conglomerate can be enantiomerically en-
riched using a combination of crystal growth and iterative
retro-aldol/aldol reactions. Considering the importance of
aldol products in nature, these findings might have relevance
for the development of biological homochirality.
Experimental Section
Figure 1. Enantiomeric excess of anti-1 versus reaction time in the enan-
tioenrichment shown in Scheme 1. Catalyst/starting ee of anti-1 (for de-
Typical procedure: Solid–solution mixtures of scalemic anti-1 (120 mg;
prepared by the addition of an appropriate amount of racemic anti-1 to
enantiomerically enriched samples) were magnetically stirred (800 rpm)
&
*
tails see text): pyrrolidine 70% ( ); pyrrolidine 60% ( ); rac-proline
^
70% ( ).
Chem. Eur. J. 2010, 16, 3918 – 3921
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3919

C. Bolm et al.
[9] 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, 1675–1681.
[10] J. W. Gibbs, Thermodynamics,
Collected Works, Vol. 1, Yale Uni-
versity Press, New Haven, 1948.
[11] M. Uwaha, J. Phys. Soc. Jpn.
2008, 77, 083802/1–4.
[12] Y. Saito, H. Hyuga, J. Phys. Soc.
Jpn. 2008, 77, 113001/1–4.
[13] a) 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, 1158–1159;
Scheme 2. Enantioenrichment of aldol product anti-1 in presence cyclohexanone (4) as co-reactant.
in DMSO (50 mL) in presence of ZrO2 beads (670 mg) as grinding
b) W. L. Noorduin, H. Meekes, W. J. P. Van Enckevort, A. Millemag-
gi, M. Leeman, B. Kaptein, R. M. Kellogg, E. Vlieg, Angew. Chem.
2008, 120, 6545–6547; Angew. Chem. Int. Ed. 2008, 47, 6445–6447;
c) C. Viedma, J. E. Ortiz, T. de Torres, T. Izumi, D. G. Blackmond, J.
Am. Chem. Soc. 2008, 130, 15274–15275; d) see also M. W. van der
Meijden, M. Leeman, E. Gelens, W. L. Noorduin, H. Meekes,
W. J. P. van Enckevort, B. Kaptein, E. Vlieg, R. M. Kellog, Org. Pro-
cess Res. Dev. 2009, 13, 1195–1198.
medium. After establishing solid-solution equilibrium (30 min.), pyrroli-
dine (10 mol%) was added to provide racemization conditions in solu-
tion. Samples of the slurry were taken out over time and the e.r. of anti-1
was determined by CSP-HPLC.
[14] For a controversy on the question if the solubility behaviour of a
system of rapidly racemising enantiomers and achiral molecules that
form a chiral state is completely analogous or not, see refs. [3] and
[6b]. For a discussion of the mechanisms also see: J. M. McBride,
J. C. Tully, Nature 2008, 452, 161–162.
[15] B. Kaptein, W. L. Noorduin, H. Meekes, W. J. P. van Enckevort,
R. M. Kellogg, E. Vlieg, Angew. Chem. 2008, 120, 7336–7339;
Angew. Chem. Int. Ed. 2008, 47, 7226–7229.
[16] W. L. Noorduin, B. Kaptein, H. Meekes, W. J. P. van Enckevort,
R. M. Kellogg, E. Vlieg, Angew. Chem. 2009, 121, 4651–4653;
Angew. Chem. Int. Ed. 2009, 48, 4581–4583;.
[17] a) S. B. Tsogoeva, S. Wie, M. Freund, M. Mauksch, Angew. Chem.
2009, 121, 598–602; Angew. Chem. Int. Ed. 2009, 48, 590–594; b) S.
Wei, M. Mauksch, S. B. Tsogoeva, Chem. Eur. J. 2009, 15, 10255–
10262.
Acknowledgements
We are grateful to the Fonds der Chemischen Industrie and the Deutsche
Forschunggemeinschaft (DFG) within the SPP 1179 (“Organocatalysis”)
for financial support and thank Prof. Dr. R. M. Kellogg (Syncom) and
Dr. B. Kaptein (DSM) for stimulating comments.
Keywords: aldol reaction
· asymmetric amplification ·
enantioenrichment · grinding · reversible reactions
[18] a) B. Rodrꢃguez, T. Rantanen, C. Bolm, Angew. Chem. 2006, 118,
7078–7080; Angew. Chem. Int. Ed. 2006, 45, 6924–6926; b) B. Ro-
drꢃguez, A. Bruckmann, C. Bolm, Chem. Eur. J. 2007, 13, 4710–
4722; c) A. Bruckmann, B. Rodrꢃguez, C. Bolm, CrystEngComm
2009, 11, 404–407.
[19] For a study that suggests that the aldol reaction between acetone
and 3 proceeds in the absence of a catalyst and expresses a sponta-
neous asymmetry by an autocatalytic pathway, see: M. Mauksch,
S. B. Tsogoeva, S. Wei, I. M. Martynova, Chirality 2007, 19, 816–
825.
[1] a) J. Jacques, A. Collet, S. H. Wilen, Enantiomers, Racemates and
Resolution, Krieger, Florida, 1994; b) A. Collet, M.-J. Brienne, J.
Jacques, Chem. Rev. 1980, 80, 215–230; c) W. M. L. Wood, Chirality
in Industry II (Eds.: A. N. Collins, G. Sheldrake, J. Crosby), Wiley,
New York, 1997, pp. 119–156. See also: d) R. M. Kellogg, B. Kap-
tein, T. R. Vries, Top. Curr. Chem. 2006, 269, 159–197.
[2] a) L. Addadi, Z. Berkovitch-Yellin, N. Domb, E. Gati, M. Lahav, L.
Leiserowitz, Nature 1982, 296, 21–26; b) L. Addadi, S. Weinstein, E.
Gati, I. Weissbuch, M. Lahav, J. Am. Chem. Soc. 1982, 104, 4610–
4617; c) L. Addadi, Z. Berkovitch-Yellin, I. Weissbuch, J. van Mil,
L. J. W. Shimon, M. Lahav, L. Leiserowitz, Angew. Chem. 1985, 97,
476–496; Angew. Chem. Int. Ed. Engl. 1985, 24, 466–485.
[3] C. Viedma, Phys. Rev. Lett. 2005, 94, 065504.
[20] Use of piperidine instead of pyrrolidine raised the e.r. of anti-1 from
85:15 to 95:5:4.5 after 1 d. It reached an e.r. of 97.5:2.5 after 4 d.
[21] Although our results suggest that the enantioenrichment involves a
deracemization through crystallization, we cannot, at the present
stage, exclude that the e.r. is also affected by a selective destruction
of the racemic compound in solution and/or its conversion into
other stereoisomers. The latter hypothesis is supported by the results
of the following experiments: Aldol products anti-1 having 30 and
70% ee were heated in DMSO (to 608C) until complete dissolution.
Then, pyrrolidine was added to the solutions, and with intense stir-
ring the mixtures were allowed to slowly cool to room temperature.
Samples were taken from the resulting slurries and analyses by
[4] D. G. Blackmond, Chem. Eur. J. 2007, 13, 3290–3295.
[5] C. Viedma, Astrobiology 2007, 7, 312–319.
[6] a) D. K. Kondepudi, J. Laudadio, K. Asakura, J. Am. Chem. Soc.
1999, 121, 1448–1451; b) D. K. Kondepudi, K. L. Bullock, J. A.
Digits, J. K. Hall, J. M. Miller, J. Am. Chem. Soc. 1993, 115, 10211–
10216; c) D. K. Kondepudi, R. J. Kaufman, N. Singh, Science 1990,
250, 975–976.
[7] a) J. Crusats, S. Veintemillas-Verdaguer, J. M. Ribꢀ, Chem. Eur. J.
2006, 12, 7776–7781; b) J. Crusats, S. Veintemillas-Verdaguer, J. M.
Ribꢀ, Chem. Eur. J. 2007, 13, 10303–10305.
1
HPLC and H NMR showed that 74 and 83%, respectively, of anti-1
had been converted to the corresponding syn-aldol products. The
stereochemical assignments of the resulting syn-1 indicated un-
changed ee values of 30 and 70%, respectively, and suggested that
the major products had 2R,4S,1’R-configuration. These results can
be interpreted as indication for a selective solution phase epimeriza-
[8] a) W. Ostwald, Lehrbuch der Allgemeinen Chemie, Vol. 2, Part 1,
Engelmann, Leipzig, 1896; b) W. Ostwald, Z. Phys. Chem. (Leipzig)
1897, 22, 289–330; c) for a recent overview, see: W. L. Noorduin, E.
Vlieg, R. M. Kellogg, B. Kaptein, Angew. Chem. 2009, 121, 9778–
9784; Angew. Chem. Int. Ed. 2009, 48, 9600–9606.
3920
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 3918 – 3921

Retro-Aldol/Aldol Reactions
COMMUNICATION
tion of anti-1 at C2 to give syn-1. Subsequent investigations of this
interesting “destructive conversion” of stereoisomers in solution
shall reveal to what extend this phenomenon is responsible for the
observed e.r. change in the original experiments.
[23] The reaction was conducted under the conditions of the competition
experiment (with DMSO as solvent and 2.4 equiv of cyclohexa-
none). Without cyclohexanone, the ee was lower to give anti-5 with
a residual e.r. of 81.5:18.5.
[22] Tsogoeva proposed that the enantiomerization process involved the
formation of an enamine intermediate in solution and a chemisorp-
tion of the other reactant (in their case an imine) on the crystal sur-
face of the major enantiomer (ref. [17b]). As a result, the major
enantiomer is reproduced at the crystal/solution interface in an au-
tocatalytic fashion. A racemic compound could reproduce itself in
the same mode. It is not clear, however, how compound mixtures
will behave. The formation of racemic anti-5 in our competition ex-
periments seems to indicate that the crystal surface of anti-1 does
not affect the stereochemical path of the anti-5 formation.
[24] While this manuscript was under review, we learned about an orga-
nocatalytic asymmetric retro aldol reaction catalyzed by a chiral dia-
mine triflic acid adduct which has been utilized in efficient kinetic
resolutions (as reported by Professor Dr. S. Luo at Symposium on
Catalysis and Fine Chemicals (C&FC) 2009 in Seoul on December
15, 2009). We thank Professor Luo for sending us subsequently an
unpublished summary of his work.
Received: December 21, 2009
Published online: March 5, 2010
Chem. Eur. J. 2010, 16, 3918 – 3921
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3921