Design of chiral separation media using monodisperse functionalized macroporous beads: effects of polymer matrix, tether, and linkage chemistry

Article abstract of DOI:10.1021/ac971196x

Porous size monodisperse methacrylate beads containing amino and hydroxyl groups, which can be used as a platform for the production of chiral separation media, have been prepared using the staged templated suspension polymerization process. The monomers 2-hydroxyethyl methacrylate and several tert-butoxycarbonyl-protected aminoalkyl methacrylates were used for the preparation of both hydroxyl- and amino-functionalized beads. Attachment of a chiral selector based on L-valine-3,5-dimethylanilide through a carbamate and a urea linkage, respectively, provides chiral stationary phases with excellent enantioselectivities in the separation of the enantiomers of 3,5-dinitrobenzamido derivatives of α-amino acids. For comparison purposes, a separation medium was also prepared by the direct polymerization of a chiral monomer analogous to the hydroxyethyl methacrylate-based stationary phase. The chiral stationary phases prepared with the N-methyl-2-aminoethyl methacrylate platform exhibit the best selectivity, and separation factors as high as 15 were achieved under normal-phase HPLC conditions.

Full text of DOI:10.1021/ac971196x

Anal. Chem. 1998, 70, 1629-1638
The Design of Chiral Separation Media Using
Monodisperse Functionalized Macroporous Beads:
Effects of Polymer Matrix, Tether, and Linkage
Chemistry
Kevin Lewandowski, Peter Murer, Frantisek Svec, and Jean M. J. Fre´ chet*
Department of Chemistry, University of California, Berkeley, California 94720-1460
mercially available.3-5 Most commercial CSPs contain chiral
selectors supported by porous silica beads. Silica-based chro-
matographic supports have numerous qualities, such as high
mechanical stability, resistance to swelling, and excellent ef-
ficiency. However, residual silanol groups on the surface of silica
may contribute to nonspecific interactions with analyte enanti-
omers and decrease the overall selectivity of the separation
medium.4,6 A partial solution is to end cap the residual silanol
groups after the attachment of a chiral selector, which does lead
to improved enantioselectivities.7-10
P orous size monodisperse methacrylate beads containing
amino and hydroxyl groups, which can be used as a
platform for the production of chiral separation media,
have been prepared using the staged templated suspen-
sion polymerization process. The monomers 2 -hydroxy-
ethyl methacrylate and several ter t-butoxycarbonyl-pro-
tected am inoalkyl m ethacrylates were used for the
preparation of both hydroxyl- and amino-functionalized
beads. Attachment of a chiral selector based on L-valine-
3 ,5 -dimethylanilide through a carbamate and a urea
linkage, respectively, provides chiral stationary phases
with excellent enantioselectivities in the separation of the
enantiomers of 3,5-dinitrobenzamido derivatives of r-ami-
no acids. For comparison purposes, a separation me-
dium was also prepared by the direct polymerization of a
chiral monomer analogous to the hydroxyethyl methacry-
late-based stationary phase. The chiral stationary phases
prepared with the N-methyl-2 -aminoethyl methacrylate
platform exhibit the best selectivity, and separation factors
as high as 1 5 were achieved under normal-phase HP LC
conditions.
Synthetic polymer beads can also be used as a platform for
the preparation of chiral stationary phases. In contrast to porous
silica beads typically used for the preparation of chiral stationary
phases, synthetic polymers allow for more accurate control of both
the functionality and the porous structure of the support, leading
to separation media with greatly enhanced properties. Their
stability over the entire range of pH, in addition to the variety of
possible surface chemistries, makes them well suited for chiral
HPLC. Macroporous polymer beads have been used successfully
for a wide variety of chromatographic separations of both large
and small molecules.11 Despite the many advantages of polymeric
(1) Ahuja, S., Ed. Chiral Separations: Applications and Technology; American
Chemical Society: Washington, DC, 1997. Jinno, K. Chromatographic
Separations Based on Molecular Recognition; Wiley-VCH: New York, 1997.
(2) Francotte, E. In The Impact of Stereochemistry on Drug Development and
Use; Aboul-Enein, H. Y., Wainer, I. W., Eds.; Wiley-VCH: New York, 1997;
p 633.
(3) Allenmark, S. Chromatographic Enantioseparation: Methods and Applications,
2nd ed.; Ellis Horwood: New York, 1991.
(4) Ahuja, S., Ed. Chiral Separations by Liquid Chromatography; ACS Symposium
Series 471; American Chemical Society: Washington, DC, 1991.
(5) Siret, L.; Bargmann-Leyder, N.; Tabute, A.; Caude, M. Analusis 1 9 9 2 , 20,
427-435.
(6) Pirkle, W. H.; Welch, C. J. J. Chromatogr. 1 9 9 2 , 589, 45-51.
(7) Pirkle, W. H.; Liu, Y. J. Org. Chem. 1 9 9 4 , 59, 6911-6916.
(8) Guillaume, M.; Sebille, B. Eur. Polym. J. 1 9 9 6 , 32, 19-26.
(9) Schleimer, M.; Pirkle, W. H.; Schurig, V. J. Chromatogr. 1 9 9 4 , 679, 23-
24.
(10) Pirkle, W. H.; Readnour, R. S. Chromatographia 1 9 9 1 , 31, 129-132.
(11) Svec, F.; Fre´chet, J. M. J. Science 1 9 9 6 , 273, 205-211. Neue, U. D. HPLC
Columns. Theory, Technology, and Practice; Wiley-VCH: New York, 1997.
Lloyd, L. L. J. Chromatogr. 1 9 9 1 , 544, 201-217. Mikes, O.; Coupek, J. In
HPLC of Biological Macromolecules: Methods and Applications; Gooding, K.
M., Regnier, F. E., Eds.; Dekker: New York, 1990; pp 25-46. Unger, K. K.,
Ed. Packings and Stationary Phases in Chromatographic Techniques; Dek-
ker: New York, 1990. Tanaka, N.; Araki, M. Adv. Chromatogr. 1 9 8 9 , 30,
81-122.
Many chiral drugs and agrochemicals are increasingly sold
as single enantiomers as a result of different activities and toxicity
profiles according to their absolute configuration. This stimulates
the rapid development of methods leading to enantiomerically pure
intermediates and of analytical assays to evaluate the enantiomeric
purity of chiral compounds. For the purposes of biological
evaluation of both enantiomers, it may be desirable to synthesize
a racemic mixture and to resolve it. Alternatively, asymmetric
syntheses result in products with an excess of one enantiomer;
however, the product may not be enantiopure. In both cases, the
final isolation of individual enantiomers from their mixtures is
usually required. At the present time, liquid chromatographic
resolution has emerged as the method of choice for both analytical
and preparative separations because of the high efficiency that
can be achieved and the ease of operation in any scale.1,2
A large number of chiral stationary phases (CSPs) have been
developed during the past decade to meet the challenge of
enantiomer separations, and numerous CSPs are now com-
S0003-2700(97)01196-7 CCC: $15.00 © 1998 American Chemical Society
Published on Web 03/18/1998
Analytical Chemistry, Vol. 70, No. 8, April 15, 1998 1629

supports, there are only few examples of polymer-based “brush-
type” chiral stationary phases.12-14
br s), 4.17 (2H, t, J ) 5.4 Hz), 3.35-3.47 (2H, m), 1.90-1.92 (3H,
m), 1.41 (9H, s). Anal. Calcd for C11H19NO4 (229.27): C, 57.63;
H, 8.35; N. 6.11. Found: C, 57.76; H, 8.47; N, 6.22.
The separation factor R, or enantioselectivity observed in the
chromatographic resolution of a racemic mixture, is defined as
the ratio of the capacity factors of the eluted enantiomers and is
a measure of relative peak separation. This is affected not only
by the intrinsic capability of the immobilized chiral selector to
differentiate between the two enantiomers but also by a number
of other factors. These include the way by which the selector is
tethered to the support, the length and chemistry of the tethering
arm, the nature of the underlying support, and the presence of
undesired reactive sites within the support where nonspecific
interactions between the stationary phase and the analytes may
occur. All of these factors affect the chiral recognition process
and, therefore, the overall performance of the chiral separation
media.
Recently, we designed a “brush-type” chiral stationary phase
with improved performance based on our advanced size mono-
disperse porous polymer beads. We demonstrated that polymeric
separation media provide greatly enhanced enantioselectivities and
reduced retention times when compared to analogous silica-based
chiral stationary phases for the separation of enantiomers.14 In
contrast to silica, the use of a polymeric support gives us a unique
opportunity to probe the variables that may affect the chiral
recognition process since the type and number of reactive sites
on the beads can be easily controlled. For example, it is possible
to prepare a large variety of CSPs and to systematically determine
which types of functionality, surface coverage, length of the spacer
arm, and tethering groups lead to separation media with improved
enantioselectivities.
N-Methyl-N-(ter t-butoxycarbonyl)aminoethyl Methacry-
late (4 a). 3 a (10.0 g, 43.6 mmol) and methyl iodide (8.15 mL,
131 mmol) were dissolved in dry tetrahydrofuran (THF) (200 mL)
under argon, and sodium hydride (60% dispersion, 1.22 g, 30.6
mmol) was added cautiously with gentle stirring to the solution
at 0 °C. After being stirred for 2 h at 0 °C and 4 h at room
temperature, the white suspension was quenched with water (80
mL) at 0 °C. The aqueous phase was extracted with Et2O,
saturated with NaCl, and again extracted with Et2O. The
combined organic extracts were dried over MgSO4 and evaporated
under vacuum. The crude product was purified by flash chro-
matography on silica gel using ethyl acetate/ hexane (1:2) to yield
4 a as a colorless oil (9.1 g, 86%): 1H NMR (300 MHz, CDCl3) δ
6.08-6.12 (1H, br s), 5.53-5.60 (1H, br s), 4.19-4.28 (2H, br s),
3.44-3.56 (2H, br s), 2.85-2.95 (3H, 2 br s), 1.91-1.97 (3H, m),
1.44 (9H, s). Anal. Calcd for C12H21NO4 (243.30): C, 59.24; H,
8.70; N, 5.76. Found: C, 58.82; H, 8.88; N, 5.73.
N-(ter t-Butoxycarbonyl)aminopentanol (2 b). To an ice-
cold solution of 5-amino-1-pentanol (8.25 g, 80.0 mmol) and
triethylamine (13.4 mL, 96.0 mmol) in CH2Cl2 (130 mL) was slowly
added di-tert-butyl dicarbonate (21.0 g, 96.0 mmol) in 5-g portions
with an additional amount of CH2Cl2 (30 mL). The reaction
mixture was stirred at 0 °C for 30 min and at room temperature
for 20 h. The solvent was removed under reduced pressure, and
Et2O (400 mL) was added to the residue. The organic phase was
washed with 1 mol/ L HCl, saturated NaHCO3 solution, and brine
and then dried over Na2SO4. Evaporation of the solvent gave the
crude product which was purified by flash chromatography on
silica gel using ethyl acetate to yield 2 b as a colorless oil (14.6 g,
90%): 1H NMR (300 MHz, CDCl3) δ 4.56 (1H, s), 3.57-3.63 (2H,
m), 3.02-3.13 (2H, m), 1.69 (1H, br s), 1.40 (9H, s), 1.30-1.61
(6H, m). Anal. Calcd for C10H21NO3 (203.28): C, 59.09; H, 10.41;
N, 6.89. Found: C, 59.24; H, 10.19; N, 6.88.
In this report, we describe the use of different types of
macroporous monodisperse beads as a new platform for the
preparation of highly selective chiral separation media and
describe the effect of surface functionality and mode of selector
linkage on their performance.
EXPERIMENTAL SECTION
N-(ter t-Butoxycarbonyl)aminopentyl Methacrylate (3 b).
Prepared as described for 3a, starting with 2b. The crude product
was triturated in ice-cold hexane and further purified by flash
chromatography on silica gel using ethyl acetate/ hexane (1:2) to
yield 3 b as a light yellow viscous oil (77%): 1H NMR (300 MHz,
CDCl3) δ 6.05-6.09 (1H, m), 5.50-5.54 (1H, m), 4.47-4.59 (1H,
br s), 4.12 (2H, t, J ) 6.6 Hz), 3.05-3.16 (2H, m), 1.90-1.93 (3H,
m) 1.42 (9H, s), 1.30-1.73 (6H, m). Anal. Calcd for C14H25NO4
(271.35): C, 61.97; H, 9.29; N, 5.16. Found: C, 61.85; H, 9.46; N,
5.25.
N-(ter t-Butoxycarbonyl)aminoethyl Methacrylate (3 a).
An ice-cold solution of methacrylic acid (1 ; 9.84 mL, 116 mmol)
and 4-(dimethylamino)pyridine (1.42 g, 11.6 mmol) in CH2Cl2 (350
mL) was treated under argon with an ice-cold solution of N-(tert-
butoxycarbonyl)ethanolamine (2 a; 20.2 g, 125 mmol) in CH2Cl2
(100 mL). 1,3-Dicyclohexylcarbodiimide (24.8 g, 120 mmol) was
added in 5-g portions and with an additional amount of CH2Cl2
(50 mL). The resulting mixture was stirred at 0 °C for 2 h and
then at room temperature for 19 h. The precipitated urea was
filtered off and washed with CH2Cl2. The filtrate was extracted
with 1 mol/ L HCl, saturated NaHCO3 solution, and brine. The
organic phase was dried over MgSO4 and concentrated under
reduced pressure. An additional amount of urea was filtered off,
and the organic solution was concentrated until crystallization was
observed. After cooling, the white crystals 3 a (15.2 g, 57%) were
collected by filtration. A second crystallization of the filtrate gave
a second crop of 3 a (8.7 g, 32%): mp 84.0-85.5 °C; 1H NMR (300
MHz, CDCl3) δ 6.08-6.11 (1H, m), 5.54-5.58 (1H, m), 4.77 (1H,
N-Methyl-N-(ter t-butoxycarbonyl)aminopentyl Methacry-
late (4 b). Prepared as described for 4 a, starting with 3 b. After
stirring for 30 min at 0 °C and 23 h at room temperature, workup
and flash chromatography on silica gel using ethyl acetate/ hexane
(1:4) yielded 4 b as a light yellow oil (69%): 1H NMR (300 MHz,
CDCl3) δ 6.06-6.10 (1H, m), 5.52-5.56 (1H, m), 4.13 (2H, t, J )
6.6 Hz), 3.12-3.26 (2H, br t, J ) 6.9 Hz), 2.82 (3H, br s), 1.90-
1.94 (3H, m), 1.44 (9H, s), 1.29-1.74 (6H, m). Anal. Calcd for
C15H27NO4 (285.38): C, 63.13; H, 9.54; N, 4.91. Found: C, 62.94;
H, 9.80; N, 4.85.
(12) Mosbach, K. Trends Biochem. Sci. 1 9 9 4 , 19, 9-14.
(13) Hosoya, K.; Yoshizaku, K.; Tanaka, N.; Kimata, K.; Araki, T.; Fre´chet, J. M.
J. J. Chromatogr. 1 9 9 4 , 666, 449-455.
P reparation of 2 -Hydroxyethyl Methacrylate (N-
L-Valine-
(14) Liu, Y.; Svec, F.; Fre´chet, J. M. J. Anal. Chem. 1 9 9 7 , 69, 61-65.
3 ,5 -dimethyl anilide) Carbamate (7 ). Triphosgene (0.38 g,
1630 Analytical Chemistry, Vol. 70, No. 8, April 15, 1998

1.28 mmol) was dissolved in CH2Cl2 (7 mL), and a solution of
2-hydroxyethyl methacrylate (HEMA, 5 ; 0.50 g, 3.84 mmol) in
pyridine (0.31 mL, 3.83 mmol) and CH2Cl2 (5 mL) was added
1.96 (2H, m), 1.30-1.49 (6H, m). 1 2 c: yield 58%; 1H NMR (300
MHz, DMSO-d6) δ 9.80 (1H, br s), 8.03-8.15 (1H, m), 7.21 (2H,
s), 6.45 (1H, s), 3.95-4.06 (1H, m), 3.03-3.08 (2H, m), 1.82-
2.89 (11H), 1.10-1.50 (16H) 0.78-0.90 (6H, m).
slowly over 5 min. After 1 h, a solution of L-valine-3,5-dimethy-
lanilide14 (6 ; 0.84 g, 3.81 mmol) in pyridine (0.21 mL, 2.60 mmol)
and CH2Cl2 (5 mL) was added dropwise at room temperature.
After 30 min, the organic phase was washed with 1 mol/ L HCl
and water and then dried over MgSO4. After evaporation of the
solvent under vacuum, the crude product was purified by flash
chromatography on silica gel using ethyl acetate/ CH2Cl2 (1:9) to
yield 7 as a white solid (1.10 g, 77%): mp 142-144 °C; 1H NMR
(300 MHz, CDCl3) δ 7.72 (1H, br s), 7.18 (2H, s), 6.77 (1H, s),
6.13 (1H, s), 5.58 (1H, s), 5.29 (1H, br d), 4.38-4.28 (4H, m),
4.00-4.10 (1H, m), 2.29 (6H, s), 1.94 (3H, s), 1.02 (3H, d, J ) 6.9
Hz), 0.98 (3H, d, J ) 6.9 Hz). Anal. Calcd for C20H28N2O5
(376.44): C, 63.81; H, 7.50; N, 7.44. Found: C, 64.00; H, 7.57; N,
7.41.
P reparation of Uniformly Sized Macroporous Beads. The
preparation of poly(2-hydroxyethyl methacrylate-co-ethylene
dimethacrylate) (1 3 ) beads was described previously.16 The
preparation of poly(3 a-co-methyl methacrylate-co-EDMA) (1 4 ),
poly(4a-co-methyl methacrylate-co-EDMA) (15), poly(4b-co-meth-
yl methacrylate-co-EDMA) (1 6 ), and poly(7 -co-EDMA) (1 7 )
beads using the staged templated suspension polymerization was
carried out as described in ref 16. The polymerizations were
performed in sealed, 500-mL Erlenmeyer flasks placed in an
orbiting shaker bath (Lab-Line) at 200 rotations/ min and 70 °C
for 17 h. The selector content of beads 1 7 is 1.22 mmol/ g based
on elemental analysis of nitrogen.
Deprotection of Amino-Functionalized Beads. A solution
of 5% trifluoroacetic acid in CH2Cl2 was added to 1 4 , 1 5 , or 1 6
(2.0 g), and the mixture was shaken slowly in an orbiting shaker
bath at room temperature. After 24 h, the beads 1 8 , 1 9 , or 2 0
were filtered, washed repeatedly with CH2Cl2, triethylamine/
CH2Cl2 (1:8), and Et2O, and dried under vacuum. The deprotec-
tion is accompanied by a disappearance of the amide II band at
1520 cm-1 of 1 4 and by the loss of the amide I band at 1700 cm-1
of 1 5 and 1 6 in the IR spectra.
Chiral Stationary P hases. To a slurry of amino-functional-
ized beads (1 8 , 1 9 , or 2 0 , 1.5 g) in THF (30 mL) and
triethylamine (0.63 mL, 4.5 mmol) at 0 °C was added 9 (1.7 g,
4.5 mmol) in two portions. After being stirred for 20 min at 0 °C
and 17 h at room temperature, the beads 2 1 -2 3 were filtered,
repeatedly washed with THF, water, 1 mol/ L KOH, CH3OH,
DMSO, and Et2O, and then dried under vacuum. The selector
content in the beads are 0.13 (2 1 ), 0.35 (2 2 ), and 0.26 mmol/ g
(2 3 ) based on elemental analysis of nitrogen.
Chiral stationary phases 2 4 -2 7 were prepared according to
a method described previously.14 To the 4-nitrophenyl carbonate-
activated beads 1 3 d (2.0 g) suspended in dry THF (25 mL) was
added 1 2 a-c (10 mmol) and triethylamine (0.5 mL). The
resulting slurry was stirred slowly at 60 °C for 17 h under nitrogen.
The beads were then filtered, washed with THF, water, CH3OH,
and Et2O, and then dried under vacuum. The selector content of
the beads is 1.10 (2 4 ), 0.80 (2 5 ), 0.70 (2 6 ), and 0.59 mmol/ g
(27) based on elemental analysis, assuming that all of the nitrogen
originates from the chiral selector functionalities.
4 -Nitrophenol (N-L-Valine-3 ,5 -dimethyl anilido) Carbam-
ate (9 ). To a solution of 6 14 (3.00 g, 13.6 mmol) in pyridine/
THF 1:2 (90 mL) was added 4-nitrophenyl chloroformate (8 ; 2.88
g, 14.3 mmol) at 0 °C. The resulting suspension was stirred for
10 min at 0 °C and for 30 min at room temperature. After the
reaction mixture was cooled to 0 °C, ethyl acetate (240 mL) was
added, followed by rapid extractions with ice-cold 1 mol/ L HCl,
water, and brine. The organic phase was dried over Na2SO4 and
evaporated under reduced pressure to yield 9 as a white solid.
(4.49 g, 86%) mp 172-173 °C; 1H NMR (300 MHz, CDCl3) δ 8.16-
8.21 (2H, m), 7.70 (1H, s), 7.26 (2H, s), 7.07 (2H, s), 6.78 (1H, s),
6.06 (1H, br d, J ) 8.8 Hz), 4.12-4.21 (1H, m), 2.25 (6H, s), 2.13-
2.34 (1H, m), 1.10 (3H, d, J ) 6.6 Hz), 1.09 (3H, d, J ) 6.7 Hz).
General P rocedure for the P reparation of Amino Alkyla-
mido- -valine-3 ,5 -dimethylanilides 1 1 a-c. A mixture of N-
L
(tert-butoxycarbonyl)alkanoic acid15 1 0 a-c (9.0 mmol), EEDQ
(2.3 g, 9.1 mmol), 6 (2.0 g, 9.1 mmol), and CH2Cl2 (20 mL) was
stirred at room temperature for 24 h. The product (1 1 a-c) was
collected by filtration. An additional amount of product can be
recovered from the mother liquor by removing the solvent under
vacuum, followed by crystallization from ethyl acetate. 11a: yield
80%; 1H NMR (300 MHz, DMSO-d6) δ 9.85 (1H, s), 7.96-8.02 (1H,
m), 7.21 (2H, s), 6.65-6.75 (2H, m), 4.22 (1H, t), 3.05-3.15 (2H,
m), 1.90-2.34 (9H), 1.39 (9H, s), 0.89 (6H, d, J ) 6.0 Hz). 1 1 b:
yield 74%; 1H NMR (300 MHz, DMSO-d6) δ 9.87 (1H, s), 7.95 (1H,
d, J ) 8.0 Hz), 7.21 (2H, s), 6.70-6.80 (1H, m), 6.68 (1H, s), 4.23
(1H, t), 2.87 (2H, q), 1.90-2.22 (9H), 1.18-1.55 (15H), 0.88 (6H,
1
d, J ) 5.5 Hz). 1 1 c: yield 60%; H NMR (300 MHz, CDCl3) δ
Capping of the Chiral Stationary P hases. To a slurry of
the CSP (2 1 , 2 2 , and 2 4 , 1.8 g) in CH2Cl2 (25 mL) was added
acetic anhydride (2.0 g, 20 mmol) and triethylamine (2.0 g, 20
mmol). The mixture was stirred slowly at room temperature for
20 h, the beads were then filtered and washed with THF, CH3OH,
water, THF, and CH2Cl2, and dried under vacuum. The capping
is accompanied by the appearance of the amide I stretch at 1685
and 1659 cm-1 of 2 1 and 2 2 , respectively, in the IR spectra.
Characterization of P orous P roperties. The specific surface
area of the beads was calculated from the BET isotherm of
nitrogen, and the pore size distribution in the dry state was
determined from mercury intrusion porosimetry using an auto-
8.89-8.94 (1H, m), 8.46 (1H, br s), 7.17 (2H, s), 6.73 (1H, s), 6.28-
6.38 (1H, m), 4.40-4.57 (1H, m), 3.02-3.12 (2H, m), 2.08-2.30
(9H), 1.58-1.70 (2H, m), 1.38-1.52 (11H), 1.10-1.50 (12H), 0.95-
1.06 (6H, m).
Cleavage of the tert-butoxycarbonyl groups of 1 1 a-c using
trifluoroacetic acid was performed according to previous meth-
ods.14 1 2 a: yield 72%; 1H NMR (300 MHz, DMSO-d6) δ 9.93 (1H,
s), 8.18-8.22 (1H, m), 7.21 (2H, s), 6.68 (1H, s), 4.21-4.32 (1H,
m), 2.68-2.79 (2H, m), 1.98-2.30 (11H), 0.82-0.97 (6H, m).
1 2 b: yield 69%; 1H NMR (300 MHz, DMSO-d6) δ 9.78 (1H, br s),
8.03-8.15 (1H, m), 7.78 (2H, s), 7.18 (1H, s), 4.68-4.82 (1H, m),
4.10-4.25 (2H, m), 2.87-3.05 (2H, m), 1.98-2.80 (13 H), 1.79-
(16) Lewandowski, K.; Svec, F.; Fre´chet, J. M. J. Chem. Mater. 1 9 9 8 , 10, 385-
(15) Houssin, R.; Bernier, J. L.; Henichart, J. P. Synthesis 1 9 8 8 , 259-261.
391.
Analytical Chemistry, Vol. 70, No. 8, April 15, 1998 1631

mated custom-made combined BET-sorptometer and mercury
porosimeter from Porous Materials Inc., Ithaca, NY.
Chromatography. The chiral stationary phases were slurry
packed at a constant pressure of 15.0 MPa into 150 × 4.6 mm i.d.
stainless steel columns. A Waters HPLC system consisting of two
510 HPLC pumps, a 717 plus autosampler, and a 486 UV detector,
and controlled by Millennium 2010 software, was used for all of
the chromatography.
Inverse size exclusion chromatography was carried out in THF
using benzene, toluene, amylbenzene, and polystyrene standards
with molecular weights ranging from 500 to 2 950 000 in order to
characterize the pore structure in a “wet” state. The peaks were
monitored at 254 nm and the distribution coefficients KD were
calculated according to the equation:
KD ) (VR - V0)/ VP
(1)
where VR is the retention volume of a standard, V is the void
0
Figure 1. Structures of the analytes used in this study.
volume of the packed column (elution volume of the largest
polystyrene standard), and VP is the total pore volume available
for the penetration of benzene.
rene particles as part of the porogenic system and results in fine
control of both the monodispersity and porous structure of the
beads.11,23 This polymerization technique is used to prepare
macroporous beads with a controlled size, density of functionality,
pore size distribution, and porous structure, which are important
variables in the design of new stationary phases for HPLC. In
the continued search for new polymeric separation media with
improved performance, we have prepared several new types of
monodisperse beads with a variety of surface functionalities.
Although monodisperse beads have previously been prepared
from monomers that do not dissolve readily in water, such as
styrene and its substituted derivatives24-27 or glycidyl methacry-
late,28 we recently developed a system that allows the direct
preparation of monodisperse beads even from water-soluble
hydrophilic monomers such as 2-hydroxyethyl methacrylate.16
The staged templated suspension polymerization, which pro-
vides monodisperse particles, is quite different from the “classical”
suspension polymerization.11 The first stage is the preparation
of small monodisperse spherical-shaped templates by an emulsion
or dispersion polymerization.29 A stabilized aqueous dispersion
of these micrometer-size uniform particles is swollen to the target
size of the final macroporous product with an emulsified mixture
of solvents, monomers, cross-linking agents, and free-radical
initiator which diffuse through the aqueous phase and enter the
templates. The swollen particles are then polymerized, resulting
in size monodisperse beads with the desired macroporous
structure.
Chiral separations were carried out using a hexane/ dichlo-
romethane mixture as a mobile phase. The selectivity values R
were calculated using the following equation:
R ) k′2/ k′1
(2)
where k′1 and k′2 are the retention factors of the enantiomers
defined as
k′i ) (tR - t0)/ t0
(3)
where tR and t0 represent the retention times of the compound
and 1,3,5-tri-tert-butylbenzene (void volume marker), respectively.
The racemic analytes N-(3,5-dinitrobenzoyl)-R-amino acid alkyla-
mides and 3,5-dinitroanilides were prepared by methods similar
to those reported elsewhere.17 1H NMR and IR spectra are in
agreement with the assigned structures shown in Figure 1.
RESULTS AND DISCUSSION
P reparation of the Supports. Macroporous beads are
typically prepared by a standard suspension polymerization which
leads to beads with a broad distribution of sizes.18,19 Therefore,
several techniques that afford size monodisperse beads for high-
end applications have been developed based on the seminal work
of Ugelstad, Vanderhoff, and El-Aasser.20-22 We have recently
developed the staged templated suspension polymerization pro-
cess, which involves the use of size monodisperse linear polysty-
(17) Pirkle, W. H.; Pochapsky, T. C. J. Chromatogr. 1 9 8 6 , 369, 175-177.
(18) Arshady, R. J. Chromatogr. 1 9 9 1 , 586, 181-197.
(19) Fre´chet, J. M. J.; Eichler, H.; Ito, H.; Wilson, G. Polymer 1 9 8 3 , 24, 995-
1000.
(20) Ugelstad, J.; Kaggerud, K. H.; Hansen, F. K.; Berge, A. Makromol. Chem.
1 9 7 9 , 180, 737-744. Ugelstad, J.; Stenstad, P.; Kilaas, L.; Prestvik, W. S.;
Rian, A.; Nustad, K.; Herje, R.; Berge, A. Macromol. Symp. 1 9 9 6 , 101, 491-
500.
(23) Wang, Q. C.; Hosoya, K.; Svec, F.; Fre´chet, J. M. J. Anal. Chem. 1 9 9 2 , 64,
1232-1238.
(24) Wang, Q. C.; Svec, F.; Fre´chet, J. M. J. Polym. Bull. 1 9 9 2 , 28, 569-576.
(25) Galia, M.; Svec, F.; Fre´chet, J. M. J. J. Polm. Sci., Polym. Chem. 1 9 9 4 , 32,
2169-2175.
(26) Lewandowski, K.; Svec, F.; Fre´chet, J. M. J. J. Liq. Chromatogr. 1 9 9 7 , 20,
227-243.
(27) Lewandowski, K.; Svec, F.; Fre´chet, J. M. J. J. Appl. Polym. Sci. 1 9 9 8 , 67,
597-607.
(28) Smigol, V.; Svec. F. J. Appl. Polym. Sci. 1 9 9 2 , 46, 1439-1448.
(29) Smigol, V.; Svec, F.; Hosoya, K.; Wang, Q.; Fre´chet, J. M. J. Angew.
Makromol. Chem. 1 9 9 2 , 195, 151-164.
(21) Cheng C. M.; Micale F. J.; Vanderhoff J. W.; El-Aasser M. S. J. Polym. Sci.,
Polym. Chem. 1 9 9 2 , 30, 235-244, 245-256.
(22) Omi, S.; Taguchi, T.; Nagai, M.; Ma, G. H. J. Appl. Polym. Sci. 1 9 9 7 , 63,
931-942.
1632 Analytical Chemistry, Vol. 70, No. 8, April 15, 1998

Scheme 1
Table 1. Characteristics of Macroporous Beads
Prepared from Functionalized Methacrylate
Monomersa
beads
1 3 a 1 3 b 1 3 c
1 7
1 8
1 9
2 0
Vp,SEC mL/ gb
Vp, mL/ gc
0.69 0.72
0.63 0.79 0.74 0.63
0.89 0.94 0.75 0.53 0.81 0.79 0.65
Dp, nmd
25
91
31
64
39
14
16
<5
18
116
1.4
22
105
1.2
19
6
Sg, m2/ ge
protected amino
groups, mmol/ g
1.1
f
a
Reaction conditions: porogen particles (shape templates, 0.10 mL
of 22% dispersion; dibutyl phthalate, 0.62 mL; 0.25% aqueous sodium
dodecyl sulfate solution, 23 mL); functional monomer, MMA, and
EDMA (5 mL); AIBN, 0.122 g; cyclohexanol, 4 mL and 1,2-dichloro-
ethane, 4 mL. b Pore volume determined by inverse size exclusion
chromatography. c Pore volume determined by mercury intrusion
porosimetry. d Median pore diameter determined from the pore
volume. e Specific surface area calculated from the BET isotherm of
nitrogen. f Determined by elemental analysis of nitrogen.
The thermodynamic groundwork for these processes has been
investigated by Ugelstad.30 The components of the emulsified
polymerization mixture must traverse the aqueous phase and swell
the templates as a result of their higher affinity toward the
nonpolar material of the template. The measure of the affinity is
the partition coefficient, which is defined as the ratio of concentra-
tions of the monomer in the aqueous and organic phases.
Transport of water-soluble monomers, characterized by a large
distribution coefficient, may represent a serious difficulty in this
process because the driving force can be diminished. To avoid
this, the hydrophilic functionality of the monomers must be
protected to increase their overall hydrophobicity. However, for
other monomers, the partition coefficient can be decreased by
using selected porogenic solvents that favor transport of the
monomer into the organic phase. Under these conditions, even
a hydrophilic monomer dissolves predominantly in the organic
phase, and if the resulting concentration of the monomer in the
aqueous phase is low, the staged polymerization process may
proceed smoothly.
The first approach involving functional group protection is
illustrated by our preparation of monodisperse macroporous beads
with pendant aliphatic amino groups using ethylene dimethacry-
late as the cross-linker. The protection approach was selected
since the solubility of ω-aminoalkyl methacrylates in water is too
high for the staged templated suspension polymerization to
proceed. Despite repeated attempts, the distribution coefficients
of the ω-aminoalkyl methacrylates could not be adjusted readily
using simple porogenic solvents. Therefore, we prepared a series
of substituted methacrylate monomers with tert-butoxycarbonyl-
protected primary and secondary amino groups (3 a,b and 4 a,b,
Scheme 1). Mixtures of 1,2-dichloroethane and cyclohexanol were
used as porogens to produce macroporous monodisperse 5-µm
beads. The proportion of protected amino groups incorporated
into the polymers 1 4 -1 7 was controlled by “diluting” the
functional monomer in the polymerization mixture with an “inert”
monomer such as methyl methacrylate. Beads containing up to
1.4 mmol/ g protected amino groups were prepared and their
composition determined from elemental analysis (Table 1). Once
the beads are obtained, their tert-butoxycarbonyl protecting groups
can easily be removed by treatment with a 5% solution of
trifluoroacetic acid in dichloromethane.
In addition to beads with aliphatic amino functionalities, size
monodisperse poly(2-hydroxyethyl methacrylate-co-ethylene
dimethacrylate) beads (13) rich in aliphatic hydroxyl groups were
also prepared. We found that a balanced mixture of cyclohexanol
and dodecanol provides an organic phase with enhanced affinity
for the hydrophilic 2-hydroxyethyl methacrylate monomer, which
facilitates the transport of monomers into the templates. This
porogenic mixture is well suited for the preparation of size
monodisperse poly(2-hydroxyethyl methacrylate-co-ethylene
dimethacrylate) beads with controlled properties containing up
to 80% of the hydrophilic monomer. A detailed study of this
polymerization process has been published elsewhere.16
Characterization of the Supports. Table 1 summarizes the
results of mercury intrusion porosimetry and BET nitrogen
adsorption measurements for the functionalized beads. Both the
porosity and the surface area of the amino-functionalized beads
do not change following deprotection of the tert-butoxycarbonyl
groups. The total pore volumes and surface areas measured in
the dry state for the hydroxyethyl methacrylate (1 3 ) and amino-
ethyl methacrylate (1 8 ) supports are similar. In contrast, the
median pore diameter decreases from 25 to 18 nm. Because the
pores of 1 8 are smaller, their overall number must be larger to
achieve a similar pore volume. Figure 2 shows the broad mercury
porosimetry distribution profiles for hydroxyl 1 3 and amino 1 8 -
2 0 functionalized beads. The steep increase in the total volume
of pores smaller than 30 nm for support 1 8 confirms the BET
data. Extension of the alkyl length in the amine monomer used
for the preparation of beads 2 0 from two to five methylene units
results in a large drop in both the pore volume and the surface
area of the support. This can be attributed to the decrease in
polarity of the monomer resulting from the introduction of a longer
alkyl chain spacer arm between the polar ester and amine
moieties. Since this polarity change is not offset by adjustments
in the composition of the porogenic solvent, the phase separation
process that is key to pore formation during polymerization is
also affected.
(30) Ugelstad, J.; Mork, P. C.; Mfutakamba, H. R.; Soleimany, E.; Nordhuus, I.;
Schmid, R.; Berge, A.; Ellingsen, T.; Aune, O.; Nustad, K. In Science and
Technology of Polymer Colloids; Poehlein, G. W., Ottewill, R. H., Goodwin, J.
W., Eds.; Nato ASI Series 67; Martinus Nijhoff: Boston, 1983; p 51.
In addition to mercury porosimetry used to characterize the
beads in the dry state, inverse size exclusion chromatography
Analytical Chemistry, Vol. 70, No. 8, April 15, 1998 1633

swell appreciably in THF, whereas the HEMA beads do, and the
swollen chains occupy part of the pore volume.
P reparation of the Chiral Stationary P hases. The separa-
tion factor observed in the chromatographic separations of
racemates is affected by several factors. These include the
intrinsic capability of the immobilized chiral selector to differenti-
ate between the two enantiomers of the analyte, the way by which
the selector is tethered, the length of the tethering arm, and the
chemical nature of the underlying support. In this study, the
effects of the chemistry of the support and tether were investigated
by preparing several chiral stationary phases containing an
identical chiral selector 6 based on substituted
directly, or through a spacer, to the support. A somewhat
analogous but silica-based phase derived from -proline has proven
L-valine attached
L
to be very effective in the separation of π-acidic compounds.32 The
valine-based selector was coupled to the beads either through a
carbamate or through an urea linkage, depending on the type of
functionality present on the original support (Schemes 2 and 3).
To obtain a broad perspective of the effects of the nature of the
matrix, the resin-binding chemistry, and the tethering arm on the
separation process, the various chromatographic conditions were
kept constant for all the chromatographic runs. These conditions
included the composition of the mobile phase (20% hexane in
dichloromethane), as well as the flow rate (1 mL/ min) and the
temperature (25 °C)
Figure 2. Differential pore size distribution curves of poly(2-
hydroxyethyl methacrylate-co-ethylene dimethacrylate) (13) ([) and
poly(aminoalkyl methacrylate-co-methyl methacrylate-co-ethylene
dimethacrylate) beads 18 (9), 19 (b), and 20 (2) measured by
mercury porosimetry.
Loading. The type of beads used in the preparation of the
chiral stationary phases influences the amount of attached selector
on the chiral stationary phase. The CSPs 2 1 -2 3 derived from
the amino-functionalized beads have loadings in the range of 0.13-
0.35 mmol/ g, which are similar to those reported for typical
analogous silica-based supports (e0.30 mmol/ g).32 These load-
ings indicate that less than 25% of the amino groups originally
present in the beads have reacted with the chiral ligand. From
this it can be inferred that the chiral selector is only attached to
the most accessible sites at the surface of the pores while the
amino groups buried within the mass of the polymer remain
unaffected. To reduce possible interferences in the separation
process, the residual amino groups on the beads were capped by
reaction with acetic anhydride to afford amide-blocked end groups.
The capping reaction was monitored by IR spectroscopy, which
showed the appearance of a strong carbonyl stretch at about 1650
Figure 3. Size exclusion calibration curves of poly(2-hydroxyethyl
methacrylate-co-ethylene dimethacrylate) (9) and poly(2-aminoethyl
methacrylate-co-methyl methacrylate-co-ethylene dimethacrylate) (0)
beads with polystyrene standards and alkylbenzenes in tetrahydro-
furan. Conditions: columns 150 × 4.6 mm i.d.; flow rate 1 mL/min;
UV detection at 254 nm.
cm-1
.
Unlike the amino beads, the selector contents of the hydroxy-
ethyl methacrylate-based CSPs 2 4 can be much higher and the
loading is controlled within a broad range up to 1.3 mmol/ g by
reacting the beads with varying amounts of the activated selector.
As was done earlier, the residual hydroxyl groups on 2 4 were
capped by reaction with acetic anhydride to provide a CSP without
readily accessible residual hydroxyl groups. For comparison
purposes, macroporous beads of CSP 1 7 with a high loading of
1.2 mmol/ g were prepared directly by copolymerization of chiral
monomer 7 with ethylene dimethacrylate (Scheme 4), to obtain
an analogue of chiral stationary phase 2 4 that does not contain
any excess hydroxyl groups.
(ISEC) can be used to determine the porous properties of the
beads in a solvent environment that is more typical of their use.31
Figure 3 shows the ISEC calibration curve for beads 1 3 and 1 8
measured with polystyrene standards in THF. These curves
confirm the mercury porosimetry data and demonstrate that for
both types of beads about 80% of the pore volume is due to small
pores only accessible to smaller solutes such as polystyrene with
MW <10 000. A simple calculation reveals that the pore volume
in the swollen state is similar to that determined by mercury
porosimetry for the amino-functionalized beads and is smaller for
the HEMA beads. This indicates that the amino beads do not
The common perception is that inserting a spacer between the
matrix and the immobilized active moiety may improve acces-
sibility to the terminal functionalities, making them more available
(31) Tanaka, N.; Kimata, K.; Mikawa, Y.; Hosoya, K.; Araki, T.; Ohtsu, Y.;
Shiojima, Y.; Tsuboi, R.; Tsuchiya, H. J. Chromatogr. 1 9 9 0 , 535, 13-31.
(32) Pirkle, W. H.; Murray, P. G. J. Chromatogr. 1 9 9 3 , 641, 11-19.
1634 Analytical Chemistry, Vol. 70, No. 8, April 15, 1998

Scheme 2
Scheme 3
Scheme 4
Figure 4. Separation of 3,5-dinitrobenzamidoleucine-N,N-diallyla-
mide enantiomers on chiral stationary phase 24. Conditions: column,
150 × 4.6 mm i.d.; mobile phase, 20% hexane in dichloromethane;
flow rate, 1 mL/min.
and naproxen (V) (Figure 1). The separation results for both the
HEMA-based 1 7 and 2 4 and amino-based CSPs 2 1 -2 3 are
shown in Table 2. In general, excellent enantioselectivities with
separation factors R as high as 14 are achieved for most of the
separations attempted. As a result of the better control of the
reactions used for the preparation of the CSPs, these values are
even higher than those obtained for a similar chiral polymeric
separation medium we prepared earlier by a multistep process
from glycidyl methacrylate beads and from silica beads. For
example, CSPs supported by glycidyl methacrylate and silica beads
afford R values of 7.25 and 3.26 for analyte Ib (Figure 1),
respectively,14 while a separation factor of 13.20 is obtained with
HEMA-based CSP 2 4 . Figure 4 shows a typical separation of
3,5-dinitrobenzamidoleucine-N,N-diallylamide enantiomers carried
out with chiral stationary phase 2 4 .
Effect of Loading. The amount of selector attached to the poly-
(2-hydroxyethyl methacrylate-co-ethylene dimethacrylate) beads
can be controlled within a broad range by using beads with
different contents of HEMA. It can be assumed that the degree
of substitution will have an effect on the enantioselectivity of the
separations. The higher the number of interacting sites, the
higher the number of recognition steps for an analyte molecule
for interactions with an analyte. While this expectation has been
demonstrated in affinity chromatography,33 studies of the effects
of a tether on molecular recognition in chiral separations are
scarce.34,35 Therefore, we prepared chiral stationary phases 2 5 -
2 7 from a single batch of HEMA beads 1 3 using selectors
attached via an amide bond to ω-aminoalkyl carboxylic acid spacer
groups with different chain lengths (Scheme 5). As expected,36
as the alkyl length of the spacer group is increased from 2 to 5
and 10 methylenes, the loading of the selector decreases from
0.80 to 0.70 and 0.59 mmol/ g, respectively.
Chromatographic Evaluation. The capabilities of the chiral
stationary phases were evaluated in the separations of 3,5-
dinitrobenzamido derivatives of racemic R-amino acids (I-III) and
derivatized nonsteroidal antiinflammatory drugs, ibuprofen (IV)
(33) Turkova, J. Bioaffinity Chromatography; Elsevier: Amsterdam, 1993.
(34) Pirkle, W. H.; Hyun, M. H. J. Chromatogr. 1 9 8 5 , 328, 1.
(35) Bargmann-Leyder, N.; Truffert, J.-C.; Tambute´, A.; Caude, M. J. Chromatogr.,
A 1 9 9 4 , 666, 27.
(36) Svec, F.; Hrudkova´, H.; Horak, D., Kalal, J. Angew. Makromol. Chem. 1 9 7 7 ,
63, 23-36.
Analytical Chemistry, Vol. 70, No. 8, April 15, 1998 1635

Scheme 5
Table 2. Retention Factors k′ and Separation Factors r Obtained for Enantioselective Separations of Racemic
Compounds on Chiral Stationary Phasesa
CSP (loading)b
2 4 (1.10)c
1 7 (1.22)
2 1 (0.13)
2 2 (0.35)
2 3 (0.26)
analyted
k′2
R
k′2
R
k′2
R
k′2
R
k′2
R
Ia
Ib
2.75
2.87
1.04
0.85
1.93
2.02
3.10
3.07
12.11
13.20
4.24
3.64
9.05
9.48
1.00
1.00
7.82
8.50
2.26
2.97
5.06
4.81
2.48
2.51
10.48
11.22
3.41
3.65
7.76
7.50
1.00
1.00
17.23
16.79
4.27
4.62
8.60
8.05
2.10
3.13
9.84
10.29
2.68
3.01
7.15
7.34
1.13
1.12
11.92
12.49
3.52
4.00
7.80
7.44
1.59
2.20
13.81
14.24
3.18
8.10
7.89
2.08
2.33
4.51
4.25
2.26
3.41
9.94
10.74
2.69
3.11
7.65
7.81
1.12
1.11
IIa
IIb
IIIa
IIIb
IV
V
4.18
10.34
10.82
1.28
1.27
a
Chromatographic conditions: column 150 × 4.6 mm i.d., mobile phase, 20% hexane in dichloromethane; flow rate, 1 mL/ min; void marker,
1,3,5-tri-tert-butylbenzene; UV detection at 254 nm. b Selector content in mmol/ g. c This CSP was prepared from beads 1 3 b. d For structures of the
analytes, see Figure 1.
passing through the column. Indeed, as the loading of CSP 2 4
Table 3. Retention Factors k′ and Separation Factors r
increases, the separation factor also increases as a result of the
rapidly increasing retention of the more strongly interacting
enantiomer. For example, using CSP 2 4 prepared from beads
1 3 a containing 50% HEMA, the retention factor k′ for the less
retained enantiomer of leucine derivative Ia increases only by
about 20% from 0.31 to 0.37 upon an increase in loading of the
selector from 0.14 to 0.71 mmol/ g, while k′ for the other
enantiomer grows from 1.4 to 3.7 as a result of a much higher
number of recognition events. This almost 300% increase in k′
for the second peak results in a remarkable increase in the
selectivity factor R from 4.4 to 10.1. An even higher loading of
1.30 mmol/ g is achieved with beads 1 3 c containing 80% HEMA.
This CSP provides an excellent selectivity factor of 15.8 in the
separation of modified leucine enantiomers Ib.
Effect of Spacer. The CSPs prepared with spacer groups of
different lengths (2 5 -2 7 ) were also tested using the same
chromatographic conditions (Table 3). Unexpectedly, the reten-
tion times are shorter and the observed enantioselectivities are
significantly lower than those of CSPs without spacer groups.
Since the tether inserts 2-10 methylene groups between the
selector and the matrix, the surface polarity of the pores is lower
than is the case with CSP 2 4 . In the normal-phase chromato-
graphic mode, this leads to a continuous decrease in retention as
the length of the spacer increases. In addition, the presence of
Obtained for Enantioselective Separations of Racemic
Compounds on Chiral Stationary Phasesa
CSP (loading)b
2 5 (0.80)
2 6 (0.70)
2 7 (0.59)
analytec
k′2
R
k′2
R
k′2
R
Ia
Ib
1.87
1.90
0.66
0.75
1.36
1.37
3.02
3.03
6.82
7.70
2.26
2.60
5.64
6.13
1.00
1.00
1.76
1.75
0.56
0.61
1.07
1.06
3.28
3.28
6.36
7.16
1.99
2.29
5.06
5.46
1.00
1.00
1.45
1.45
0.46
0.51
0.82
0.82
3.09
3.09
5.38
5.84
1.75
1.95
4.25
4.34
1.00
1.00
IIa
IIb
IIIa
IIIb
IV
V
a
Chromatographic conditions: column 150 × 4.6 mm i.d., mobile
phase, 20% hexane in dichloromethane; flow rate, 1 mL/ min; void
marker, 1,3,5-tri-tert-butylbenzene; UV detection at 254 nm. b Selector
content in mmol/ g. c For structures of the analytes, see Figure 1.
an additional amide functionality near the stereogenic center of
the selector may lead to strong hydrogen bonding (Scheme 5).
This increases the level of nonspecific interactions and, conse-
quently, deteriorates the enantioselectivities of all of the CSPs.37
Therefore, the binding chemistry available for linear aliphatic
(37) Pirkle, W. H.; Hyun, M. H. J. Chromatogr. 1 9 8 5 , 322, 295.
1636 Analytical Chemistry, Vol. 70, No. 8, April 15, 1998

amino acids makes them unsuitable for use as spacers in the
design of efficient chiral stationary phases. Comparisons of the
separations achieved in columns packed with CSPs 2 5 -2 7
indicate that the overall effect of the length of the tether itself on
selectivity is rather small. The specific selectivity calculated as
the ratio of selectivity factor over loading remains almost constant
for all of the analytes separated on these three CSPs. For
example, the specific selectivity varies in the range 8.6-9.1 for
analyte Ia, 2.8-2.9 for IIa, and 7.1-7.2 for IIIa.
In contrast to CSPs 2 5 -2 7 in which the tethering was
achieved by chemical modification of the selector, the spacer in
CSP 2 3 is “built in” as a part of the reactive monomer. Compared
to CSP 2 2 , the length of the alkyl group that links the secondary
amine to the matrix in CSP 2 3 is increased from two to five
methylene groups. This leads to a decrease in the overall polarity
of the medium, and consequently, the retention in this normal
phase decreases as observed for CSPs 2 5 -2 7 based on HEMA
beads. However, this linker only extends the chain length
between the matrix and the selector but does not add any new
interacting functionality. Therefore, the decrease in enantiose-
lectivities of 2 3 compared to 2 2 can safely be attributed as least
in part to the smaller number of attached selector moieties found
in the former CSP (0.26 vs 0.35 mmol/ g).
the HEMA beads, these amide functionalities are more polar38
than the amino groups they replace. Therefore, the overall
polarity of the amidated supports increases and leads to longer
retention. Compared to the uncapped 2 1 , the enantioselectivities
on the capped CSP are generally 10% lower. For example, the k′
values for the first eluted enantiomer of Ia increase from 1.75 to
2.08 while the selectivity decreases from 9.8 to 9.0, respectively.
Although the amide groups are not located in close proximity to
the selector, they may still interact in a nonspecific way forming
H-bonds that would affect the recognition process. Capping CSP
2 2 does not induce any changes in retentions and enantioselec-
tivities of the CSP. The polarity of the tertiary amide groups in
capped 2 2 is similar to that of secondary amines, and therefore,
the overall polarity does not change significantly after capping.
The tertiary amide groups are also less prone to H-bond formation
and their presence does not appear to lead to detrimental
nonspecific interactions. These results suggest that capping of
free amino functionalities of the CSPs does not improve their
performance and, therefore, is essentially useless.
Effect of Support Chemistry. The surface chemistry of a support
itself may affect chiral separations in two ways. First, it may
contribute to the surface polarity, and second, the surface
functionalities may also interact with the analytes and, hence,
indirectly affect the separation process. For example, CSPs
derived from poly(hydroxyethyl methacrylate) beads generally
display shorter retention times when compared to those prepared
from the amino-functionalized supports. Specifically, the retention
factor for the second eluted enantiomer of analyte Ia is 2.75 for
CSP 2 4 and 17.23 for 2 1 (Table 2). This matches with the
increased polarity of the polymeric matrix that still contains a large
number of free amino groups. In a practical application, the
retention could be adjusted to achieve the separation within a
shorter period of time by changing the composition of the mobile
phase. This was not done in the context of this study because,
had conditions been changed, the direct comparisons of different
CSPs would have been impossible.
The other difference between these two CSPs is in the mode
of attachment of the chiral selector to the support. While
carbamate chemistry is used to attach the selector to surface
hydroxyl groups, a urea linkage results from attachment to the
amine matrix. In contrast to 2 4 , for which no selectivity was
observed with antiinflammatory drugs, this rather small change
in support chemistry is sufficient to afford some selectivity with
CSPs 2 1 -2 3 for the enantiomers of substituted ibuprofen and
naproxen (Table 2).
The change of the surface functionalities from primary to
secondary amino groups leads to a decrease in retentions. CSP
2 1 displays higher retentions for all of the analytes tested, most
likely as a result of the increased number of residual primary
amino groups compared to the smaller number of secondary
amino groups of 2 2 as determined from the loading data. Since
the porous properties of the amino-functionalized beads are
similar, the lower loading for 2 1 can be attributed to the less
reactive primary amino groups compared to the secondary amino
groups of 2 2 . The slightly lower enantioselectivity of CSP 2 1
can then be related to a much lower loading rather than to the
Effect of Surface Polarity. The retention of an analyte on a chiral
stationary phase depends on the overall polarity of the environ-
ment surrounding the selector. A comparison of chiral stationary
phases 1 7 and 2 4 prepared from chiral monomer 7 and beads
1 3 , respectively, may help to elucidate this effect. Since 1 7
contains no residual hydroxyl groups and, therefore, is less polar
than 2 4 , one would expect that its retention would be lower.
Surprisingly, CSP 1 7 exhibits much higher retentions and shows
a slightly decreased enantioselectivity when compared to the
analogous CSP based on HEMA beads with a similar selector
loading (Table 2). This suggests that the polarity factor is not
overly important for these two CSPs while the enhanced retention
demonstrates the overwhelming effect of differences in the porous
structures of the two chiral media. As shown in Table 1, CSP 1 7
has a much lower porosity, pore size, and surface area than its
counterpart 2 4 prepared by chemical modification of optimized
beads.
Another method that can reveal the effect of the overall polarity
of the separation medium is the capping of residual hydroxyl
groups in CSP 2 4 . Since the porous properties of the matrix
remain essentially unchanged by this procedure, any change in
selectivity should illustrate the net effect of surface polarity.
Indeed, capping leads to a less polar chiral stationary phase with
lower retentions for all of the analytes tested. For example, the
k′ value for the first eluted enantiomer of analyte IIa is 0.22 on
the uncapped CSP and 0.16 on the capped CSP, and the separation
factor increases from 13.2 to 16.9, respectively. These results
suggest that capping the residual hydroxyl groups not only
increases the overall hydrophobicity of the support but also
enhances its enantioselectivity.
Capping of the residual amino groups of 2 1 and 2 2 with acetic
anhydride results in CSPs with additional amide functionalities.
In contrast to the formation of less polar ester groups by capping
(38) CRC Handbook of Chemistry and Physics, 76th ed.; Lide, D. R., Ed.; CRC
Press Inc: Boca Raton, FL, 1995-96; pp 6-159.
Analytical Chemistry, Vol. 70, No. 8, April 15, 1998 1637

effect of surface chemistry because the specific selectivity calcu-
lated for this CSP is the highest we have ever observed and
reaches values up to 79 for analyte Ib. This is very promising
because an increase in the substitution of these beads may afford
CSPs with extremely high selectivities.
Our results have provided good insight into both the positive
and negative contributions of the individual variables to the
enantioselectivity of the whole system. These findings are
invaluable and will help in the design of extremely selective CSPs.
Our results also show that carefully functionalized uniformly sized
synthetic polymer beads can be used to prepare stationary phases
with vastly improved chromatographic properties for enantiosepa-
rations. Further improvements in the enantioselectivity are
expected to result from the extent and density of functionalization
of the amine beads with the selector moieties. In addition, we
anticipate that the reactive beads can be useful not only for the
preparation of very selective chiral separation media but also in
the design of polymeric reagents and supports for combinatorial
chemistry.
CONCLUSION
Polar monodisperse porous polymer beads containing reactive
hydroxyl or amino groups with controlled porous properties and
well-defined contents of functionalities are well suited as a platform
for the production of chiral separation media. While the selector
surface loading of 2-hydroxyethyl methacrylate beads can be much
higher, the N-methylaminoethyl methacrylate support affords a
chiral separation medium with a higher specific selectivity. The
use of a spacer group that only increases the distance between
the selector and the matrix does not affect the enantioselectivity
of the CSP, but it affects the polarity of the surface and thus the
retention of the separation medium. However, spacers based on
R,ω-aminoalkylcarboxylic acids, which add an extra amide linkage
close to the selector moiety, dramatically decrease the selectivity
and should be avoided. Spacer groups may well be useful for
the modification of the surface polarity of chiral stationary phases;
however, they should not be attached to the support using
functional groups that can interfere with the chiral recognition
process.
ACKNOWLEDGMENT
Support of this research by a grant of the National Institute of
General Medical Sciences, National Institutes of Health (GM-
44885), is gratefully acknowledged. P.M. thanks the Swiss National
Science Foundation for a postdoctoral fellowship.
Received for review October 28, 1997. Accepted January
28, 1998.
AC971196X
1638 Analytical Chemistry, Vol. 70, No. 8, April 15, 1998