Temperature selective diastereo-recognition (TSD): Enantiomeric ibuprofen via environmentally benign selective crystallization

Article abstract of DOI:10.1021/op030203i

Selective crystallization of ibuprofen/lysinate from 1 mol of (R,S)-(racemic) ibuprofen and ≤0.5 mol of (S)-lysine in aqueous ethanol affords either (S)-(+)-ibuprofen/(S)-lysinate or (R)-ibuprofen/(S)-lysinate (in preponderance) depending on the crystallization conditions. The previously unreported temperature selective diastereo-recognition (TSD) provides simple and efficient means to prepare either enantiomer of ibuprofen from (R,S)-ibuprofen utilizing the same commercially available inexpensive resolving agent, (S)-lysine. The unwanted enantiomeric ibuprofen could be recovered from the mother liquor and racemized by a simple, relatively waste-free thermal process. This racemization method when utilized in conjunction with the selective crystallization technology provides a simple, efficient, and eco-friendly means to prepare (S)-(+)-ibuprofen lysinate in an overall essentially quantitative yield. This technology also incorporates the fundamental principle of atom economy (via direct production of the preferred pharmaceutical salt of (S)-lysine).

Full text of DOI:10.1021/op030203i

Organic Process Research & Development 2003, 7, 717−722
Temperature Selective Diastereo-Recognition (TSD): Enantiomeric Ibuprofen via
Environmentally Benign Selective Crystallization
Apurba Bhattacharya*
Department of Chemistry, Texas A&M UniVersity, KingsVille, KingsVille, Texas 78363, U.S.A.
David Murphy
Celanese Chemical Co., 1901 Clarkwood Road, Corpus Christi, Texas 78409, U.S.A.
Abstract:
to avert an undesirable load on metabolism, (S)-(+)-ibuprofen
in the form of its lysinate salt has been introduced as an
alternative by Merck.4 Asymmetric synthesis of (S)-(+)-
ibuprofen via an elegant chiral protonation of aryl ketenes
(Merck), asymmetric hydrogenation of vinyl arenes mediated
by chiral catalysts, as well as chiral auxiliary has been
reported.5 Ibis developed an enzymatic resolution of the
methyl ester.6 In light of such continuing interest in the area
of (S)-(+)-ibuprofen coupled with the recent demand for the
enantiomerically pure drug in chemotherapy, an efficient
preferential resolution of racemic (R,S)-ibuprofen is highly
desirable. Although already demonstrated, the existing
methods for resolving (R,S)-ibuprofen via fractional crystal-
lization of diastereomeric salts with chiral amines (e.g.,
R-methybenzylamine, (S)-lysine) suffers from the following
disadvantages: (a) Classical resolution is always accompa-
nied by the production of at least 1 mol of salt (waste) with
a maximum theoretical yield of only 50% (based on the chiral
auxiliary). (b) Although synthetic resolving agents (e.g.,
R-methybenzylamine) are available in both enantiomeric
forms, naturally occurring agents (e.g., (S)-lysine) are often
only available in one enantiomeric form, thereby limiting
the scope of such diastereomeric separations.2a,7 (c) And, the
success of resolution methods depends on an efficient
Selective crystallization of ibuprofen/lysinate from 1 mol of
(R,S)-(racemic) ibuprofen and e0.5 mol of (S)-lysine in aqueous
ethanol affords either (S)-(+)-ibuprofen/(S)-lysinate or (R)-
ibuprofen/(S)-lysinate (in preponderance) depending on the
crystallization conditions. The previously unreported temper-
ature selective diastereo-recognition (TSD) provides simple and
efficient means to prepare either enantiomer of ibuprofen from
(R,S)-ibuprofen utilizing the same commercially available
inexpensive resolving agent, (S)-lysine. The unwanted enantio-
meric ibuprofen could be recovered from the mother liquor
and racemized by a simple, relatively waste-free thermal
process. This racemization method when utilized in conjunction
with the selective crystallization technology provides a simple,
efficient, and eco-friendly means to prepare (S)-(+)-ibuprofen
lysinate in an overall essentially quantitative yield. This technol-
ogy also incorporates the fundamental principle of atom
economy (via direct production of the preferred pharmaceutical
salt of (S)-lysine).
Introduction
(R,S)-Ibuprofen belongs to a class of nonsteroidal anti-
inflammatory agents that has remained an area of intense
study.1 (S)-(+)-Ibuprofen is the pharmacologically active
component of (R,S)-ibuprofen. The (R)-(-) isomer is either
inactive or weakly active in vitro although the difference in
activity is markedly decreased in vivo due to metabolic
inversion of the (R)-(-) to the active (S)-(+) enantiomer.2
On the other hand, (R)-(-)-ibuprofen, the therapeutically
inactive isomer, is expected to give less gastrointestinal side
effects than the racemate but still retains its anti-inflammatory
activity via metabolic inversion to the active (S)-(+) isomer.
Also the (R) enantiomer can be potentially toxic due to its
storage in fatty tissue as the glycerol ester, whose long-term
effects are unknown.3 To realize enhanced specificity and
(4) (a) Tung, H. H.; Waterson, S.; Reynolds, S.; Paul, E. AIChE Symp. Ser.
1991, 87 (84), 64. (b) Tung, H. H. U.S. Patent 4,994,604, 1991.
(5) (a) Ohta, T.; Takaya, M.; Kitamura, K.; Nagai, K.; Noyori, R. J. Org. Chem.
1987, 52, 3174. (b) Chan, A. S. C. 199th American Chemical Society
Meeting, Boston, April, 1990. (c) Alessandra, A.; Amoroso, R.; Bettoni,
G.; De Filippis, B.; Giampietro, L.; Marco; Triccak, M. L. Tetrahedron
Lett. 2002, 43 (24), 4325. (d) Fan, Q. H.; Deng, G.-J.; Chen, X.-M.; Xie,
W.-C.; Jiang, D.-Z.; Liu, D. S.; Chan, A. C. S. J. Mol. Cat A: Chem. 2000,
159 (1), 37. (e) Chen, A.; Ren, Li.; Crudden, C. M. J. Org. Chem. 1999, 64
(26), 9704. (f) Larsen, R. D.; Corely, E. G.; Davis, P.; Reider, P.; Grabowski,
E. J. J. J. Am. Chem. Soc. 1989, 111, 7650. (g) Senanayake, C. H.; Bill, T.
J.; Larsen, R. D.; Liu, T. J.; Corley, E. G.; Reider, P. J. Synlett 1994, (3),
1999 and references sited therein.
(6) Mutsaers, J. H. G. M.; Kooreman, H. J. Recl. TraV. Chim. Pays-Bas. 1991,
110, 185.
(7) (a) Manimaran, T. U.S. Patent 5,015,764, 1991. (b) Chikusa, Y.; Fujimoto,
T.; Ikunaka, M.; Inoue, T.; Kamiyama, S.; Maruo, K.; Matsumoto, J.;
Matsuyama, K.; Moriwaki, M.; Nohira, H.; Saijo, S.; Yamanishi, M.;
Yoshida, K. Org. Process Res. DeV. 2002. 6, 291. (c) Jacques, J.; Collet,
A.; Wilen, S. H. Enantiomers, Racemates and Resolutions; Wiley: New
York, 251, 1981. (d) The cost of naturally occurring (S)-lysine is $2/kg,
whereas the enantiomeric (R)-lysine (which is required for the production
of (R)-(-)-ibuprofen via traditional classical resolution) costs $8000/kg.
(e) The Anastas, P. T.; Warner, J. C. Twelve Principles of Green Chemistry.
Green Chemistry Theory and Practice; Oxford University Press: New York,
1998.
(1) (a) Rieu, J.-P.; Boucherle, A.; Cousse, H.; Mouzin, G. Tetrahedron. 1986,
42, 4095. (b) Sonwane, H. R.; Bellur, N. S.; Ahuja, J. R.; Kulkarni, D. J.
Tetrahedron: Asymmetry. 1992, 3 (2), 163.
(2) (a) Avgerinos, A.; Hutt, A. J. Chirality 1990, 2, 249. (b) Kaiser, D. G.;
Vangiessen, G. J.; Reischer, R. J.; Wechter, W. J. J. Pharm. Sci. 1976, 65,
269.
(3) (a) Hutt, A. J.; Caldwell. J. Pharm. Pharmacol. 1983, 35, 693. (b) Bertola,
M. A. U.S. Patent 5,108,917, 1992. (c) Williams, K.; Day, R.; Kinhinicki,
R.; Duffield, A. Biochem. Pharmacol. 1986, 35, 3403.
10.1021/op030203i CCC: $25.00 © 2003 American Chemical Society
Published on Web 08/02/2003
Vol. 7, No. 5, 2003 / Organic Process Research & Development
•
717

Scheme 1
racemization and recycling of the unwanted (R)-isomer.
Existing technologies to racemize ibuprofen are often non-
trivial and waste producing. Examples include the following:8
(a) Conversion of the acid functionality to the thioester, base-
catalyzed racemization of the corresponding thioester, fol-
lowed by hydrolysis to regenerate the acid and thiol is
effective but cumbersome. (b) Refluxing in octane/Et3N for
18 h followed by acidification generates salt waste. (c)
Refluxing in concentrated HCl for 72 h requires expensive
chloride-tolerant equipment (e.g., glass-lined) and generates
acid waste. (d) Refluxing in excess 2-propanol/excess NaOH
for 16 h followed by acidification is waste producing. (e)
Refluxing in isopropyl acetate in the presence of acetic
anhydride (0.1 mol) and sodium acetate (0.1 mol) generates
salt/organic waste. (f) Heating ibuprofen acid chloride with
sodium ibuprofenate requires expensive chloride-tolerant
equipment (e.g., glass-lined) and generates salt waste. (g)
Refluxing (S)-(+)-ibuprofen in acetic anhydride and pyridine
is waste producing.
This report describes a selective crystallization of ibu-
profen/lysinate from one mole of (R,S)-ibuprofen and e0.5
mol of (S)-lysine. An unprecedented temperature selective
diastereo-recognition (TSD) leads to the preparation of either
enantiomer of ibuprofen (as well as the preferred lysinate
salt) utilizing the inexpensive, naturally occurring, and readily
available (S)-lysine7c as the chiral resolving agent and
appropriate choice of resolution conditions. In addition, we
also report a convenient, waste-free, thermal racemization
of (S)-(+)-ibuprofen that does not require any external
reagent, catalyst, and/or solvent, thus rendering alternate
racemization technologies less attractive. This racemization
method, when utilized in conjunction with the selective
crystallization technology, provides an efficient and envi-
ronmentally benign technology to prepare (S)-(+)-ibuprofen
lysinate in an overall yield which is nearly quantitative.7e
studies using 0.5 mol of (S)-lysine per 1 mol of (R,S)-
ibuprofen and seeding provided (S)-(+)-ibuprofen/(S)-lysi-
nate with modest de’s, up to 30%. Significant enhancement
of the diastereomeric purity of the product was achieved by
carrying out the crystallization under kinetic conditions (0-5
°C) with seeding using pure (S)-(+)-ibuprofen/(S)-lysinate
crystals. The crystallization was optimized by manipulating
the four key independent variables: (R,S)-ibuprofen:lysine
ratio; ethanol:(R,S)-ibuprofen ratio; temperature; and ethanol:
water ratio. The optimized crystallization afforded our best
results, a 95:5 diastereomer ratio (90% de) with an 80%
isolated yield of the product. The optimum conditions
were: (a) (R,S)-ibuprofen:lysine 2.5 mole:1mole; (b) 3.4 mL
ethanol/1 g of (R,S)-ibuprofen; (c) ethanol:water 95:5 (v/v);
and (d) 0 °C for 4 h.9a
The product of the first crystallization can be recrystallized
(see experimental below) once from aqueous ethanol (90 wt
%) to afford (S)-(+)-ibuprofen/(S)-lysinate monohydrate in
99% de.9b Maintaining the water content to at least the 10
wt % level or higher in the aqueous ethanol during the recrys-
tallization is imperative for obtaining the pharmacologically
desired monohydrate form in a reproducible manner.
The combined mother liquor and the wash liquor from
the first crystallization (enriched in (R)-(-)-ibuprofen) was
evaporated and partitioned between hexane and aqueous HCl.
The hexane phase was evaporated to provide a residue of
lysine-free ibuprofen that underwent >99% racemization on
heating at 235 °C for 4 h.
Results and Discussions
Selective Crystallization Studies of Ibuprofen Lysinate.
Classical resolution of (R,S)-ibuprofen via conventional
techniques utilizing equimolar quantities of (R,S)-ibuprofen
and (S)-lysine (as the resolving agent) in aqueous ethanol
afforded the diastereomeric (S)-(+)-ibuprofen/(S)-lysinate salt
in >99% optical purity after two crystallizations. Conceptu-
ally, we envisaged the possibility of selective crystallization
of (S)-(+)-ibuprofen/(S)-lysinate utilizing only 0.5 mol of
(S)-lysine per 1 mol of (R,S)-ibuprofen in the crystallization
medium, provided a significant differential rate of crystal
growth exists between the two diastereomeric salts,7b thereby
promoting the crystallization of only one diastereomer with
the mother liquor being enriched in the other enantiomer as
shown in Scheme 1. Crystallization studies were conducted
by adding an aqueous solution of (S)-lysine to an ethanolic
solution of (R,S)-ibuprofen, followed by seeding. The dia-
stereomeric excesses (de) of the crystals (sampled at 1 h
intervals) were monitored by chiral HPLC analysis. Initial
Further investigation revealed the following facts that
were crucial to the success of the selective crystallization
technology.
Seeding. The crystals of (S)-(+)-ibuprofen/(S)-lysinate
obtained by this method did not carry any water of crystal-
lization. From an experimental standpoint, it is important
that the crystallization process be seeded with the same
anhydrous form of crystals that is desired. Seeding with the
monohydrate form of the diastereomeric salt was shown to
inhibit the crystallization.10 The initial anhydrous seed
crystals were independently prepared by crystallization of
(S)-(+)-ibuprofen/(S)-lysinate from aqueous ethanol (97
wt %).
Crystallization Conditions (Time and Temperature).
Particularly noteworthy and remarkable is the fact that a
(9) (a) D-optimal design with a quadratic model was utilized for the optimization
studies. (b) Monohydrate is the form desired by Merck.
(10) External seeding is not an absolute necessity for the crystallization. Thus,
a crystallization mixture left at 0 °C in the refrigerator for 48 h showed
slow crystal growth of (S)-(+)-ibuprofen/lysinate without external seeding.
(8) (a) Chen, C. Y.; Chang, S. Y.; Lin, S. A.; Wen, H. I.; Cheng, Y. C.; Tsai,
S. W. J. Org. Chem. 2002, 67 (10), 3323. (b) Manimaran, T. U.S. Patent
5,015,764, 1991. (c) Larsen, R. D. U.S. Patent 4,946,997, 1990. (d) Ruchardt,
C.; Gartner, H.; Salz, Ulrich. Angew. Chem., Int. Ed. Engl. 1984, 23, 162.
718
•
Vol. 7, No. 5, 2003 / Organic Process Research & Development

Scheme 2
Scheme 3
Scheme 4
complete reversal of diastereoselectivity was observed when
the crystallization was conducted at 22 °C for 24 h under
otherwise identical conditions, thereby producing the (R)-
(-)-ibuprofen/(S)-lysinate in 80:20 diastereomer ratio (60%
de). The same diastereo-reversal was also observed when
the heterogeneous crystallization mixture obtained after 4 h
at 0 °C (where the isolated crystals were diastereomerically
enriched in (S)-(+)-ibuprofen/(S)-lysinate) was allowed to
stir at ambient temperature for an additional 24 h, as depicted
in Scheme 2. Selective precipitation of the (R)-(-)-ibuprofen/
(S)-lysinate also gives rise to an interesting possibility where
the mother liquor is enriched with the (S)-(+)-ibuprofen (as
the free acid and not as the lysinate salt) and direct recovery
of (S)-(+)-ibuprofen from the mother liquor is a distinct and
lucrative probability.
The ability to produce either diastereomer utilizing the
same chiral auxiliary ((S)-lysine) is previously unreported.7c
These results can be explained by assuming that under a
given set of conditions the crystallization process is under-
going two competing pathways to produce either diastere-
omer. Scheme 3 depicts a free-energy profile for such a
crystallization process in which (R)-(-)-ibuprofen/(S)-lysi-
nate (B) is the thermodynamic sink but (S)-(+)-ibuprofen/
(S)-lysinate (C) is formed faster due to lower Eact. If the
crystallization is reversible (which is usually the case when
the mother liquor is not supersaturated with the solute) and
the process is terminated well before the equilibrium has been
established (0 °C, 4-6 h.), the crystallization will be
kinetically controlled producing the (S)-salt. However, if the
crystallization is permitted to approach equilibrium (22 °C
24 h), the process will be thermodynamically controlled,
producing the (R)-salt.11 Therefore, when (S)-(+)-ibuprofen/
(S)-lysinate is the desired product, maintaining kinetic
conditions throughout the whole process is crucial. A
crystallization mixture which is preferentially enriched with
the (R)-(-)-ibuprofen/(S)-lysinate crystals can be reworked
by the following process: (a) redissolving the whole mixture
(60 °C) in the mother liquor and (b) recrystallizing under
kinetic conditions (0°, 4-6 h) to produce the kinetically
preferred (S)-salt.
Conceivably, the mechanism for the conversion from the
S-S diastereomeric salt to the corresponding R-S diaster-
eomeric salt does not necessarily exclude a temperature-
dependent equilibrium pathway. Further investigation indi-
cated that this was not the case. A crystallization mixture
which had achieved equilibrium conditions at 22 °C (80:20
R:S ratio) after 24 h, was cooled to 0 °C and kept at 0 °C
for 12 h. No change in the diastereomeric ratio of the isolated
crystals was observed. The above explanation is hypothetical
at this juncture, and further detailed kinetic/mechanistic
studies are needed.
Purity of the Ingredients. Purity of the ingredients plays
an important role in the success of the crystallization. For
example, the cycloamide, 3-amino-azepan-2-one, a common
impurity resulting from intramolecular dehydration of lysine,
(Scheme 4) tends to hinder the crystallization and produce
pasty, hard-to-filter material with poor de’s (c.a. 50%).
Presumably, the free -NH2 group of the cycloamide is
capable of forming a different pair of diastereomeric salts
with (R,S)-ibuprofen, thereby complicating the selective
crystallization.12 Residual hydrocarbons (e.g., cyclohexane
(11) (a) Klumpp. ReactiVity in Organic Chemistry; Wiley: New York, 1982;
pp 36-89. (b) Even at 0 °C, the reversal of diastereomer in the crystallization
to produce the thermodynamically favored (R)-(-)-ibuprofen/(S)-lysinate
was observed by carrying out the crystallization over a longer period of
time. Thus, conducting the crystallization at 0 °C over an 18 h period under
otherwise identical conditions produced the diastereomeric (R)-salt in 33%
diastereomeric excess although the same crystallization when sampled after
4, 5, and 6 h showed the normal 93:7 (S over R) selectivity.
(12) Significant cycloamide was detected in samples of lysine, which has been
kept on the shelf for an extended period of time.
Vol. 7, No. 5, 2003 / Organic Process Research & Development
•
719

Scheme 5
and hexane) introduced as trace (1% based on ethanol)
impurities also produced slow-filtering, poor quality (60%
de or less) material. No crystallization was observed in the
presence of heptane. Particularly noteworthy is the fact that
absolute ethanol obtained from two different sources (Spec-
trum and Midwest) showed significant variability in their
performance. Reason for this is not entirely clear, but it may
be linked to the fact that hydrocarbon-mediated azeotropic
drying is often utilized to produce absolute ethanol.13
Choice of Solvent to Wash the Product Cake. Since
the preferential crystallization involves separation of (S)-(+)-
ibuprofen/(S)-lysinate salt from the enantiomeric R-(-)-
ibuprofen, the major impurity in the product cake is the (R)-
(-)-ibuprofen (in free acid form and not as the lysinate salt).
An optimum result (86-90% de) was obtained when the
product cake was washed with cold (0 °C), dry ethanol which
solubilizes the R-(-)-ibuprofen and not the salt. Increasing
the water content in the cake-wash solvent (e.g., 90:10
ethanol:H2O) decreased the diastereomeric purity of the
product by preferentially solubilizing the (S)-salt over (R)-
(-)-ibuprofen.
Racemization Studies of Enantiomeric Ibuprofen.
Racemization of a carboxylic acid with a chiral center at
the R-position via the corresponding ketene formed by the
treatment of acid chloride with a base (e.g., triethylamine)
was reported as early as 1919.14 We envisaged the possibility
of an efficient, waste-free racemization of ibuprofen by (a)
enolization of enantiomeric ibuprofen at an elevated tem-
perature, followed by (b) enantiorandom reprotonation of the
enol moiety to produce racemic ibuprofen. This surmise was
confirmed by experiment (Scheme 5).
Figure 1. Racemization of Ibuprofen.
Table 1. Racemization kinetics of (S)-(-)-ibuprofen
S
time temp
So measured (h) (K) (cal/mol) (h-1
xa
ya
k
S
)
‘(h-1
)
calculated
0.997 0.996
0.997 0.994
0.997 0.991
0.997 0.991
2 393 0.001 29 -7.5939 0.000 14 0.997
24 393 0.001 29 -8.9782 0.000 14 0.994
72 393 0.001 29 -9.3806 0.000 14 0.987
96 393 0.001 29 -9.6683 0.000 14 0.984
0.997 0.985 120 393 0.001 29 -9.1921 0.000 14 0.981
0.92
0.92
0.92
0.97
0.97
0.97
0.8
1
2
3
1
2
3
1
2
3
473 0.001 07 -1.7824 0.116 09 0.833
473 0.001 07 -1.6179 0.116 09 0.764
473 0.001 07 -1.2860 0.116 09 0.709
498 0.001 01 -0.6243 0.609 99 0.639
498 0.001 01 -0.6804 0.609 99 0.541
498 0.001 01 -0.7420 0.609 99 0.512
508 0.000 99 -0.0907 1.131 55 0.552
508 0.000 99 -0.4622 1.131 55 0.505
0.69
0.58
0.661
0.562
0.527
0.997 0.580
0.997 0.540
0.997 0.500
508 0.000 99
-
1.131 55 0.501
a x ) (1/RT), y ) ln[(1/2t) ln((2So - 1)/(2S - 1))]. Ea(calculated) ) 30.95
K cal/mole; Arrhenius frequency factor ) 2.6 × 1013 hr-1
.
A. Kinetic Studies. The kinetics of the racemization
process were measured by conducting a series of thermal
racemization experiments at different temperatures. Batch
temperatures throughout the experiments were maintained
constant with the aid of a digital Therm-O-Watch, and the
enantiomeric excesses as a function of time were monitored
by chiral HPLC analysis. The degree of racemization of (S)-
ibuprofen as a function of time at different temperatures is
depicted in Figure 1, and the data are given in Table 1. In
Figure 1, the range of rates is so great that secondary axes
are used to present data for the lowest temperature, 120 °C;
that is, the scales for 120 °C data are along the top and right
side of the figure.
Calculation of the above values for S (mol fraction of
(S)-(+)-ibuprofen) follows a standard kinetic development.
It is assumed that the kinetics are first order in both
enantiomers and that the rate constants for both the forward
and reverse reactions are the same. The rate equation for
the change in (S)-(+) concentration is then
Racemization of (S)-(+)-ibuprofen was achieved by
heating molten (S)-(+)-ibuprofen at 235 °C under an inert
atmosphere for 4 h after which time complete racemization
(1:1 S:R) was confirmed by chiral HPLC analysis. The
optimum temperature range for conducting the racemization
was 230-235 °C, which enabled the racemization to be
complete within a reasonable time period (4-5 h) without
any significant decomposition (<2% by LC area percent)
of ibuprofen. Although these experiments were performed
in the batch mode, in a commercial process, these conditions
can be easily translated to a continuous process in a
commercial plug flow reactor.
(13) Selective crystallization performed under established conditions using
absolute ethanol obtained from “Spectrum” afforded product with the
expected diastereomeric ratio (93:7, S:R), whereas absolute ethanol obtained
from “Midwest Grain” afforded product with a lower diastereomeric ratio
(80:20, S:R). Traces of hydrocarbon have been detected by GC analysis of
the “Midwest Grain” ethanol.
d[S]
dt
) - k([S] - [R])
(1)
(14) McKenzie, A.; Christie, E. W. J. Chem. Soc. 1070, 1934. (b) Morrison, J.
D.; Mosher, H. S. Asymmetric Organic Reactions; American Chemical
Society: Washington, DC, 1980.
where [S] is the concentration of (S)-(+)-ibuprofen and [R]
is the concentration of (R)-(-)-ibuprofen. If the substitution
720
•
Vol. 7, No. 5, 2003 / Organic Process Research & Development

is made to replace concentration, [S], with mole fraction, S,
eq 1 becomes
enantiomeric species will exist, and if the conditions permit,
it will lead to selective crystallization. The TSD technology
introduces a new possibility in the preparation of either
separate enantiomer utilizing the same chiral resolving agent.
Application of this methodology to a host of R-arylpropionic
acids (profens) is under investigation. As chemical resolution
technology continues to play a dominant role despite
increasing competition from the enzymatic methods, the
potential offered by such selective crystallization technology
especially in the TSD of other R-arylpropionic acids should
not be overlooked.16
dS
dt
) - k(S - R) ) - k(S - (1 - S))
(2)
(3)
(4)
Separating variables and integrating gives
S
t
dS
1 - 2S
) - t)0 k dt
∫
∫
S)So
Substituting (1/2) d(2S) ) dS and integrating gives
1
2
2S - 1
2So - 1
-
ln
) kt
(
)
Experimental Section
Materials and Methods. Reactions were carried out
under an atmosphere of nitrogen and were stirred magnetic-
ally unless otherwise noted. All solvents were reagent grade.
1H NMR spectra were recorded on a Varian UNITY-400
spectrometer in CDCl3 solution with tetramethylsilane as an
internal standard. Analytical high performance liquid chro-
matography (HPLC) was carried out by using a Beckman
110B pump with a model 421-A gradient controller and a
Hitachi L 4000 variable wavelength detector. The chiral assay
was carried out with an Altex model 110A pump and a
Hitachi L 4000 variable wavelength detector.
Substituting the Arrhenius equation [in which R is the gas
constant, (not the concentration of the R enantiomer) and F
is a pre-exponential Arrhenius frequency factor]
- Ea
RT
k ) F exp
(5)
(
)
into the above eq 4, transposing, and taking the log of both
sides give
2So - 1
1
2t
1
ln ln
) -E
+ ln F
(6)
(
(
2S - 1 ))
aRT
Preparation of Seed Crystals of (S)-(+)-Ibuprofen/(S)-
Lysinate for the Selective Crystallization. A solution of
(S)-(+)-ibuprofen (5.26 g, 0.025 mol) dissolved in 25 mL
of absolute ethanol was added to an aqueous solution of (S)-
lysine (3.75 g, 0.025 mol) in 4.2 mL of water with stirring.
The resulting solution was filtered through a sintered glass
funnel to remove any insoluble material. To this clear
solution was added absolute ethanol (57.5 mL) with stirring
over 5 min at 22 °C. The turbid solution was stirred for 30
min, and then an additional 40 mL of ethanol was added to
this mixture over a period of 10 min with stirring. The
resulting suspension was cooled to 0 °C and stirred for an
additional 4 h. The crystals were filtered, washed with ice-
cold (2 °C) ethanol (10 mL), and dried under house vacuum
overnight to produce 7.4 g (84% yield) of (S)-(+)-ibuprofen/
(S)-lysinate. The product contained 0.2% water as determined
by a Karl Fischer titration. [Ethanol < 100 ppm by GC.]
Selective Crystallization of (S)-(+)-Ibuprofen/(S)-Ly-
sinate. Racemic (R,S)-ibuprofen (41.25 g, 0.2 mol) was
dissolved in absolute ethanol (140 mL). The solution was
cooled to 0 °C and (S)-lysine (11.7 g, 0.08 mol dissolved in
7 mL water) was added to this solution while the solution
temperature was maintained 0 °C. The solution was seeded
with (S)-(+)-ibuprofen/(S)-lysinate (100 mg). The resulting
mixture was stirred at 0 °C for 4 h. The crystallization was
monitored by filtering a small sample through a disposable
centrifugal Teflon membrane microfilter. The crystals were
analyzed by chiral HPLC analysis for (S)- and (R)-ibuprofen
content, and the mother liquor was analyzed by achiral HPLC
for total contained ibuprofen. The crystals were filtered,
which is in the form y ) mx + b, since Ea and ln F are
constants and t, T, and S are variables with experimentally
known values.
Excel’s linear curve fit was used to solve for Ea and F.
The result is given in Table 1, where ln(1/2t ln(2So - 1/2S -
1)) is y and (1/RT) is x. The calculated Ea is 30.95 K cal/
mol, and the Arrhenius frequency factor is 2.6 × 1013 h-1.
These constants describe the rate over the entire range of
the experiments, a range of nearly 4 orders of magnitude.
Summary
We have demonstrated a simple, unprecedented, efficient,
and environmentally friendly temperature selective diastereo-
recognition (TSD) to convert (R,S)-ibuprofen to either
enantiomer utilizing the same commercially available inex-
pensive resolving agent, (S)-lysine.15 The process obviates
the shortcomings associated with both classical resolution
(involving expensive synthetic resolving agents) and enzyme
technology. The process also produces the enantiomeric
ibuprofen directly as the preferred lysinate salt, which is
desirable for human consumption thereby avoiding the
obligatory separation of ibuprofen from a chiral auxiliary
(atom economy).7e Diastereomeric enrichment of (S)-(+)-
ibuprofen lysinate to a 99% level via a single recrystallization
also demonstrates the superiority of lysine as a resolving
agent for ibuprofen. This TSD when utilized in conjunction
with the waste-free thermal racemization of enantiomeric
ibuprofen provides an efficient and environmentally friendly
process to prepare (S)-(+)-ibuprofen lysinate in an overall
essentially quantitative yield. Several resolving agents ob-
tained in nature are available only as a single enantiomer.
As long as the diastereomers are not formed by covalent
bonding, such equilibration between diasteroemeric and
(15) (S)-Lysine hydrochloride (feed grade, 98.5% chemical purity, 100% optical
purity) is available from Archer Daniel Midland Co. in bulk quantities at a
price of $2 per kg.
(16) Bhattacharya, A. U.S. Patent 5,332,834, 1994; U.S. Patent 5,380,867, 1994;
U.S. Patent 5,399,707, 1994.
Vol. 7, No. 5, 2003 / Organic Process Research & Development
•
721

washed with ice-cold (0 °C) ethanol (40 mL), and dried under
house vacuum overnight to give 22.6 g (80% yield based
on (S)-lysine) of (S)-(+)-ibuprofen/(S)-lysinate salt in 90%
diastereomeric purity as evidenced by chiral HPLC. The
mother liquor was concentrated in a rotary evaporator to
produce an amber oil which was partitioned between hexane
(200 mL) and water (50 mL containing 0.1 g of HCl) to
remove the residual lysine. Removal of hexane in a rotary
evaporator produced (R)-enriched ibuprofen (approximately
75:25, R:S) which was first racemized by heating at 230 °C
for 4 h under a nitrogen atmosphere and was then distilled
at about 220 °C, under vacuum (10 mM Hg), to produce
substantially pure racemized ibuprofen (26 g, 96% recovery)
that was recycled to the selective crystallization stream.
Analysis. A. Chiral HPLC Analysis. Column: CHIRAL-
HSA (astec) 100 × 4.0 mm2. Mobile phase: 5 mM octaonic
acid and 10% 2-propanol in 0.1 M phosphate buffer (pH )
7). Flow: 0.9 mL/min. Wavelength: 236 nm. Detector:
Hitachi L 4000, AUFS 0.02. Retention times: (S)-(+)-
ibuprofen, 7.93 min; (R)-(-)-ibuprofen, 12.92 min.
B. Achiral HPLC Analysis. 1. For (R,S)-Ibuprofen.
Column: ALTECH Hypersil ODS (C18), 25 cm × 4.6 mm.
Mobile phase: 40% H2O (0.1% H3PO4), 60% Acetonitrile.
Flow: 1.2 mL/min. Wavelength: 236 nm. Detector: Hitachi
L 4000, AUFS 0.02. Retention times: ibuprofen, 7.3 min.
2. For Lysine. Column: ALTECH Hypersil ODS (C18)
25 cm × 4.6 m. Solvent: 100% H2O (0.1% H3PO4). Flow:
0.6 mL/ min. Detector: Hitachi L 4000, AUFS 0.02.
Wavelength: 200 nm. Retention time: lysine, 5.42 min.
Recrystallization to Produce Diastereomerically Pure
(S)-(+)-Ibuprofen/(S)-Lysinate. (S)-(+)-Ibuprofen/(S)-ly-
sinate (2 g, 90% diastereomeric purity) was dissolved in a
mixture of ethanol (12 mL) and water (1.2 mL) at 60 °C.
The mixture was slowly cooled from 60 °C to 0 °C over a
period of 2 h and then stirred at 0 °C for an additional 2 h.
The crystals were collected by filtration, washed with cold
(0 °C) ethanol (6 mL), and dried under vacuum overnight
to give 1.72 g (90% yield of the available (S)-salt) of the
(S)-(+)-ibuprofen/(S)-lysinate in >99% diastereomeric pu-
rity. (The 30 °C/hour gradient during the cooling process is
important in obtaining the required diastereomeric purity.
Thus, when the crystallization mixture was cooled at a faster
rate (ca. 10 min), the diastereomeric purity of the product
was only 97%.)
6.6 mL of water) was added to this solution while maintain-
ing 12 °C temperature and then was followed by an
additional 107 mL of ethanol over a period of 1 h again
maintaining the 12 °C temperature. The reaction mixture was
stirred at 22 °C for 48 h. The crystals were collected by
filtration, washed with ethanol (40 mL), and dried under
house vacuum overnight to give 9.9 g (70% yield based on
(S)-lysine) of (R)-(-)-ibuprofen/(S)-lysinate salt in 60%
diastereomeric purity as evidenced by chiral HPLC. The salt
was recrystallized from 65.3 mL of ethanol/water (10:1 v:v)
as described above to produce 7.1 g (90% yield of the
available (R)-salt) of the (R)-(-)-ibuprofen/(S)-lysinate in
>99% diastereomeric purity.
Preparation of (S)-(+)-Ibuprofen from (S)-(+)-Ibu-
profen/(S)-Lysinate. (S)-(+)-Ibuprofen/(S)-lysinate, 35.2 g
(99% de), was dissolved in 300 mL of water at 22 °C. The
solution was adjusted to pH ) 1 by the addition of
concentrated aq HCl. The mixture was stirred for 1 h at 22
°C. The heterogeneous mixture was filtered, and the crystals
were washed with water (60 mL). The crystals were dried
at 50 °C under house vacuum to give 20.6 g of (S)-(+)-
ibuprofen (100% yield) of 99% enantiomeric purity as
evidenced by chiral HPLC. Mp 52-53 °C (lit7b 52-53 °C
for 99% pure (S)-isomer). 1H NMR δ (CDCl3) 7.21 (d, J )
8.0 Hz, 2H), 7.09 (d, J ) 8.0 Hz, 2H), 3.70 (q, J ) 7.2 Hz,
1H), 2.44 (d, J ) 6.8 Hz, 2H), 1.95-1.75 (m, 1H), 1.49 (d,
J ) 7.2 Hz, 3H), 0.89 (d, J ) 7.2 Hz, 6H).
Preparation of (R)-(-)-Ibuprofen from (R)-(-)-Ibu-
profen/(S)-Lysinate. (R)-(-)-Ibuprofen/(S)-lysinate, 35.2 g
(99% de), was subjected to the same acidification process
as described above to give 20.6 g of (R)-(-)-ibuprofen (100%
yield) in 99% enantiomeric purity as evidenced by chiral
HPLC.
Racemization of (S)-(+)-Ibuprofen. (S)-(+)-Ibuprofen
(10 g, >99% ee) was held at 230 °C for 4 h at the end of
which time complete racemization of (S)-(+)-ibuprofen was
observed as evidenced by chiral HPLC. Upon cooling to
room temperature, the racemic (R,S)-ibuprofen crystallized
on standing. HPLC analysis of the product showed no
significant (<1%) decomposition.
Acknowledgment
Partial financial support provided by the Welch Founda-
tion and Texas Excellence fund is greatfully acknowledged.
Selective Crystallization of (R)-(-)-Ibuprofen/(S)-Ly-
sinate. Racemic (R,S)-ibuprofen (20.6 g, 0.1 mol) was
dissolved in absolute ethanol (107 mL). The solution was
cooled to 12 °C. (S)-Lysine (5.84 g, 0.04 mol dissolved in
Received for review March 25, 2003.
OP030203I
722
•
Vol. 7, No. 5, 2003 / Organic Process Research & Development