Structural aspects of nucleation inhibitors for diastereomeric resolutions and the relationship to Dutch resolution

Article abstract of DOI:10.1002/anie.200704021

You say you want a resolution: Nucleation inhibitors for the resolution of diastereomers by selective crystallization have been designed and tested. The resolution of racemic 3-methoxyphenylethylamine was optimized with (S)-mandelic acid as the resolving agent (see scheme). Multifunctional inhibitors are particularly effective. (Chemical Equation Presented).

Full text of DOI:10.1002/anie.200704021

Angewandte
Chemie
DOI: 10.1002/anie.200704021
Resolution of Diastereomers
Structural Aspects of Nucleation Inhibitors for Diastereomeric
Resolutions and the Relationship to Dutch Resolution**
Michel Leeman, Giuseppina Brasile, Edith Gelens, Ton Vries, Bernard Kaptein, and
Richard Kellogg*
The resolution of enantiomers[1a,b] by Dutch Resolution[2]
involves the use of an equimolar mixture of two or three
structurally similar and homochiral[3] resolving agents.Two
major factors contribute to the success of the method.First,
the structurally similar resolving agents often form a non-
stoichiometric solid solution of crystals.[4,5] Second, one of the
resolving agents can act as a nucleation inhibitor and retains
the more soluble diastereomeric salt in solution longer.[6a,b]
This kinetic effect effectively increases the solubility of the
more soluble diastereomer.For example, if one starts with a
solution of diastereomeric salts of composition S and an
eutectic composition E0, the phase rule requires that the solids
have composition A0 (Figure 1).If a nucleation inhibitor
moves the (kinetically determined) eutectic composition to
Ea, the solid composition changes from nearly racemic A0 (in
the example shown in Figure 1) to enantiopure Aa.Only a few
percent of the additive can lead to profound effects.[6a,b]
One should not expect the design of nucleation inhibitors
to be an easy task.Resolutions may be affected by many
factors.[1a] The effects of additives in crystallization processes
can be profound.[7] The separation of conglomerates by use of
either a pure enantiomer (entrainment)[1a] or “tailor-made”
additives has been well studied.[8–10] The number of candidates
can be limited if only compounds with some structural
resemblance to the resolving agent or racemate are inves-
tigated.The observation of Barton and Kirby that the
resolution of narwedine is aided by traces of optically pure
and structurally analogous galanthanine was seminal.[11a]
Our test system consisted of (S)-mandelic acid (MA) as
the resolving agent and (Æ )-3-methoxyphenylethyl amine
(3MeOPEA) as the racemate (Scheme 1).The ternary phase
Scheme 1. Resolution of (Æ)-3MeOPEA with (S)-MA.
diagram is shown in Figure 2, and the crystal structures of the
diastereomeric salts are given in the Supporting Information.
Under the conditions used (see the Supporting Information),
with the resolving agent but without an additive, the first salts
Figure 1. Ternary phase diagram for ideally behaving diastereomers
and the effects of nucleation inhibition thereon.
[*] M. Leeman, G. Brasile, Dr. E. Gelens, Dr. T. Vries, Prof. R. Kellogg
Syncom BV, Kadijk 3
Groningen 9747 AT (The Netherlands)
Fax: (+31)505-757-399
E-mail: r.m.kellogg@syncom.nl
Homepage: http://www.syncom.nl
Dr. B. Kaptein
DSM Pharmaceutical Products
Advanced Synthesis, Catalysis & Development
P.O. Box 18, 6160 MD Geleen (The Netherlands)
[**] We are grateful to the Samenwerkingsverband Noord- Nederland
(Cooperation Northern Netherlands) and the European Fund for
Regional Development (EFRO) for partial support of this work.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Figure 2. Phase diagram for mixtures of (S)-MA and (Æ)-3MeOPEA in
2-butanone at 208C; the scale is in wt%.
Angew. Chem. Int. Ed. 2008, 47, 1287 –1290
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1287

Communications
formed with only 10% de (Table 1, entry 1).Various additives
bifunctional compound (entry 7, Table 1, mixture of diaste-
reomers).
structurally related to MA were tested.The most effective are
shown in Table 1, entries 2a, 4a,b, and 5a.The diastereomeric
excess (de) of the first salts increases to nearly 100% although
the yields decrease.The additive used for entry 4a (Table 1) is
Nucleation inhibition should also apply to the racemate to
be resolved.We had previously observed that certain bifunc-
tional amines effectively blocked the activity of 1-phenylethyl
amine as a resolving agent.[6b] Some of these compounds have
been investigated as inhibitors in the resolution of (Æ )-
3MeOPEA by (S)-MA.Results with the best inhibitors are
shown in Table 2.The additives tested in entries 2, 4, and 5
Table 1: Resolutions of (Æ)-3MeOPEA with MA and nucleation inhib-
itors resembling mandelic acid.
Entry Additive[a]
(R)-
Yield [%][b] de [%][c] S-factor[d]
(S)-MA
Table 2: Resolutions of (Æ)-3MeOPEA with (S)-MA and nucleation
1
none
S
S
72
42
10
97
0.14
0.81
inhibitors resembling the racemate.
Additive[a]
Yield [%][b]
de [%][c]
S-factor[d]
0.14
2a
Entry
1
none
72
10
2b
3
R
S
70
31
12
99
0.17
0.62
2
32
96
0.61
3a
3b
32
31
97
96
0.62
0.60
[e]
–
4a
4b
4c
S
S
R
36
42
68
95
95
15
0.68
0.80
0.20
4
35
89
0.63
5
27
95
0.51
[a] In these experiments, 6mol% amine of additive was used relative to
(Æ)-3MeOPEA; the ratio of total acid to total amine was 1:1. [b] All
experiments were performed in duplicate. [c] The ee values determined
by HPLC on a chiral stationary phase are identical to de values. [d] S-
factor=yieldde2. [e] (R)-MA was used for the resolution.
5a
5b
6
S
R
S
R
40
63
26
16
96
11
94
93
0.77
0.15
0.48
0.30
(Table 2) are racemic and a mixture of diastereomers,
whereas that of entry 3 is enantiomerically pure.[14] This
inhibitor (see entries 3a and 3b, Table 2) is equally effective in
a resolution following the Marckwald principle (reciprocal
resolution).[15] We have no ready explanation for this.
A remarkable observation is that an achiral bisamine can
be an extremely effective nucleation inhibitor.In entry 4 of
Table 3 the results with 1,3-bis(2-amino-2-propyl)benzene
(1,3-BAPB) are given.Both the bifunctionality and the bulk
seem to be necessary for nucleation inhibition, as may be
concluded from entries 2 and 3 (Table 3).
7
[a] In these experiments, 6mol% of additive (based on the carboxylic
acid function) was used relative to enantiopure MA; the ratio of total acid
to total amine was 1:1. [b] All experiments were performed in duplicate.
[c] The ee values determined by HPLC on a chiral stationary phase are
identical to de values. [d] Resolution efficiency:[21] S-factor=yieldde2.
A larger scale resolution with 1,3-BAPB as the additive
was examined.Starting from 40 g of the ( Æ )-3MeOPEA/(S)-
MA salt, a resolution in 2-butanone with 0.5% and 1.0% of
the 1,3-BAPB/[(S)-MA]2 salt as an additive was performed
and compared to the results of resolution without an additive
(isolated salt 10% de).From the results the kinetic time frame
in which the additive is effective can be derived (Figure 3).
When 0.5% 1,3-BAPB was used, the first sample of isolated
solid after 1 h had > 90% de.However, a second sample
removed after another hour had only 10% de in the collected
solids, and this percentage remained constant over time.
Analysis by HPLC established that after 1 h the additive 1,3-
racemic.From the comparison of entries 2a and 2b, and
entries 5a and 5b (Table 1) it is clear that the enantiomer that
has the same chiral sense as the resolving agent (entry 4b) is
responsible for the inhibition.The additives could not be
detected by HPLC in the precipitated salts (sensitivity limit:
< 0.1%) except in the case of entry 4a (Table 1), in which 1%
incorporation was observed.[12] The stereochemical correla-
tions resemble the “rule of reversal” postulated by Lahav[8]
and others.[13] We also note the strong inhibitory action of a
1288
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 1287 –1290

Angewandte
Chemie
Table 3: Resolutions of (Æ)-3MeOPEA with (S)-MA and nucleation
cannot have such a chiral conformation, was foiled by the
total insolubility of this material.
inhibitors resembling 1,3-BAPB.
Additive[a]
Yield [%][b]
de [%][c]
10
S-factor[d]
0.14
A search for inhibitors to improve resolutions may lead to
success if homochiral structural analogues of either resolving
agent or racemate are examined, although in many cases
racemic additives may safely be tested.Along this line,
bifunctional analogues of resolving agent or racemate may be
particularly effective, and bifunctional achiral materials with
a structural resemblance to either racemate or resolving agent
may also be active.It has been noted that the use of slightly
contaminated racemates or resolving agents as a result of a
previous reaction step might improve the subsequent reso-
lution, because of the structural resemblance to the racemate
or the resolving agent.[20] A general point is that there is a
better chance of optimizing a resolution that has a significant
diastereomeric excess at the eutectic point and little tendency
to form end solid solutions.Determination of eutectic
compositions is time well spent.
Entry
1
2
none
72
59
13
0.16
3
69
43
14
95
0.20
0.82
4[e]
[a] In these experiments, 6% equiv wt amine of additive was used
relative to (Æ)-3MeOPEA; the ratio of total acid to total amine was 1:1.
[b] All experiments were performed in duplicate. [c] The ee values
determined by HPLC on a chiral stationary phase are identical to de
values. [d] S-factor=yieldde2 [e] The isolated solids contained 2%
additive.
Experimental Section
General procedure for the resolution of (Æ )-3MeOPEA with (R)- or
(S)-MA.: A 2.5-mL aliquot of a 0.13m stock solution of (Æ )-
3MeOPEA in 2-butanone was pipetted into a Kimble reactor tube
equipped with a magnetic stirrer (12 cm).Subsequently, a 25.-mL
aliquot of a 0.13m stock solution of (R)- or (S)-MA in 2-butanone was
added to the tube and spontaneous crystallization occurred within a
couple of minutes.In a typical experiment, 6% of either 3MeOPEA
or MA was replaced by an equal amount of additive in such a manner
that the ratio of the total amount of amine relative to the total amount
of acid was maintained at 1:1 and the total volume was 5 mL.In this
manner, direct comparison to the experiment without an additive
(entry 1, Table 1) can be made.Further details are given in the
Supporting Information.
Figure 3. Composition of the solids without washing; after washing
Received: August 31, 2007
^
(line ) the composition is 96% de in accord with Table 3, entry 3.
Published online: January 10, 2008
Keywords: chiral resolution · crystal growth · Dutch resolution ·
.
BAPB was for the major part incorporated in the precipitated
less soluble salt.The resulting low concentration of additive in
the mother liquor made the system unstable and in the next
hour the more soluble salt also precipitated with incorpo-
ration of the rest of the additive.However, when 10. % 1,3-
BAPB was used, the more soluble diastereomeric salt
remained dissolved for at least 5 days.[16] Analysis by HPLC
established that only 0.38% of the additive was still present in
the mother liquor, the rest was incorporated in the precipi-
tated less soluble salts.The solids were collected, washed, and
dried to give (S)-3MeOPEA/(S)-MA in 43% yield and
96% de (S-factor= 0.83). Apparently at least 0.5% additive
is consumed during crystallization.[17] The mother liquor
shows more enrichment: 79% de with additive versus
53% de without additive (see Figure 1).This finding is very
useful if also the other enantiomer is required.[18]
nucleation inhibitors · phase diagrams
[1] a) For the theory and practice of resolutions, see; J.Jacques, A.
Collet, S.H. Wilen, Enantiomers, Racemates and Resolutions,
Krieger, Malabar, FL, 1994; b) See for example “Novel Optical
Resolution Technologies”: Topics in Current Chemistry (Eds.: K.
Sakai, N.Hirayama, R.Tamura), Springer, Berlin, 2007.
[2] T.Vries, H.Wynberg, E.van Echten, J.Koek, W.ten Hoeve,
R.M.Kellogg, Q.B.Broxterman, A.Minnaard, B.Kaptein, S.
van der Sluis, L.Hulshof, J.Kooistra, Angew. Chem. 1998, 110,
2491; Angew. Chem. Int. Ed. 1998, 37, 2349.
[3] In the sense of same relative absolute stereochemistries.
[4] C.Gervais, R.F.P.Grimbergen, I.Markovits, G.J.A.Ariaans, B.
Kaptein, A.Bruggink, Q.B.Broxterman,
2004, 126, 655.
J. Am. Chem. Soc.
[5] This resembles Peachey–Pope resolutions, in which less than one
Under the conditions of the resolution 1,3-BAPB is
probably doubly protonated.The C2 conformation with the
ammonium groups located on opposite faces of the molecule
is chiral.Both steric factors and charge repulsion could favor
this conformation.Unfortunately an attempt to test this idea
with 1,4-BAPB/[(S)-MA]2 salt (structure not shown), which
equivalent resolving agent is used.[1b]
[6] a) J.W. Nieuwenhuijzen, R.F.P. Grimbergen, C. Koopman,
RM. .Kellogg, TR. .Vries, K.Pouwer, E.van Echten, B.
Kaptein, L.A. Hulshof, Q.B. Broxterman,
2002, 114, 4457; Angew. Chem. Int. Ed. 2002, 41, 4281; b) J.
Dalmolen, T.D.Tiemersma-Wegman, J.W.Nieuwenhuijzen, M.
Angew. Chem.
Angew. Chem. Int. Ed. 2008, 47, 1287 –1290
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1289

Communications
van der Sluis, E.van Echten, TR. .Vries, B.Kaptein, QB. .
[17] In resolutions with 1% additive some of the additive is initially
incorporated into the salt, but over a period of hours it slowly
returns into solution.
[18] Bifunctional inhibitors work in other resolutions.Sulfonic acid 2
(absolute configuration not established) and derivatives thereof
are good resolving agents.[19] The resolution efficiency for
racemate 1 is improved from S = 0.14 without the bifunctional
additive 3 to S = 0.38 with 3% of additive 3.
Broxterman, R.M.Kellogg, Chem. Eur. J. 2005, 11, 5619.
[7] J.W. Mullins, Crystallization, 4th ed., Elsevier, Oxford, 2001,
pp.181 – 215.
[8] L.Addadi, S.Weinstein, E.Gati, I.Weissbuch, M.Lahav, J. Am.
Chem. Soc. 1982, 104, 4610.
[9] Polymeric additives: a) I.Weissbuch, D.Zbaida, L.Addadi, L.
Leiserowitz, M.Lahav, J. Am. Chem. Soc. 1987, 109, 1869; b) Y.
Mastai, M.Sedlꢀk, H.Cꢁlfen, M.Antonietti, Chem. Eur. J. 2002,
8, 2429.
[10] Effects of trace impurities on amino acids: M.Lahav, I.
Weissbuch, E.Shavit, C.Reiner, G.J.Nicholson, V.Schꢂrig,
Origins Life Evol. Biosphere 2006, 36, 151.
[11] a) D.H.R.Barton, G.W.Kirby,
J. Chem. Soc. 1962, 804; b) see
also: K.R.Wilson, R.E.Pincock, J. Am. Chem. Soc. 1975, 97,
1474.
[12] Incorporation of a small amount of the inhibitor, perhaps in a
polymorph formed at the beginning of precipitation of solid
material, cannot be excluded.
[13] See also: a) S.Kojo, E.Tanaka, Chem. Commun. 2001, 1980;
b) S.Kojo, H.Uchino, M.Yoshimura, K.Tanaka,
Chem.
Commun. 2004, 2146.
[19] R.M.Kellogg, J.W.Nieuwenhuijzen, K.Pouwer, T.R.Vries,
QB. .Broxterman, RF.P..Grimbergen, B.Kaptein, RM. .
[14] J.Dalmolen, M.van der Sluis, J.W.Nieuwenhuijzen, A.Meet-
sma, B.de Lange, B.Kaptein, R.M.Kellogg, Q.B.Broxterman,
Eur. J. Org. Chem. 2004, 1544.
La Crois, E.de Wever, K.Zwaagstra, A.C.van der Laan,
thesis 2003, 1626.
Syn-
[15] See reference [1a], p.306.
[20] D.W.Armstrong, J.T.Lee, L.W.Chang,
Tetrahedron: Asym-
[16] Similar observations in other systems have been made: J.S.C.
Loh, W.J.P. van Enckevort, E. Vlieg, C. Gervais, R.F.P.
Grimbergen, B.Kaptein, Cryst. Growth Des. 2006, 6, 861.
metry 1998, 9, 2043.
[21] E.Fogassy, A.Lopata, F.Faigl, F.Darvas, M.ꢃcs, L.Toke,
Tetrahedron Lett. 1980, 21, 647.
1290 www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 1287 –1290