Novel chiral stationary phase for the enantioseparation of racemic drugs 2

Since most of the separated compounds showed the best chiral resolution results on ETHE-6PC, it is believed that the optimal spacer length for solute retention and selectivity was achiev

Chapter 5 Effects of Spacer Length

5.1 Introduction

In Chapter 4, we have studied the effects of substituents as chiral selectors and it appears that phenylcarbamate CSPs are suitable for separation of a wide range of racemic compounds Apart from this, we understand that CSPs are usually prepared by immobilizing chiral molecules such as modified CDs onto the silica supports via appropriate alkyl spacers.1-3 It appears that the type and length of the spacers in the CSPs are also important factors in affecting the enantioselectivity in HPLC.4-11 It was reported that the enantioseparations of the CD CSPs are partially attributed to the spacer which kept the cavity of the chiral molecules away from the support surface.12 The length of the attachment of the tether connecting the chiral selector to the chromatographic support can have a dramatic influence in the chiral separation because the spacer makes the solutes’ access easier for the chiral discrimination to take place.13

In this chapter, we investigate the effects of spacer length of the phenylcarbamate CSP based on the basis of the work done in the previous chapter In this work, three chemically bonded ether-linkage β-CD CSPs with different spacer lengths have been

synthesized as illustrated in Figure 5.1 The three CSPs ETHE-3PC, ETHE-6PC and ETHE-11PC (with 3-, 6- and 11-carbon spacer respectively) were derived from the

phenylcarbamoylated β-CD and immobilized onto the silica surface via a spacer arm with different lengths The spacer length is varied in the number of methylene group (-CH2-) and the CSPs were packed into HPLC microbore column [∅ 2.1 x 150 mm] for HPLC analysis

107

O OR)7-n

( OR )14

) (Si

n(

m

m = no of carbon

Figure 5.1 CSPs with different spacer lengths

Surface coverage for the three CSPs was calculated based on the following formula.6

αexp = (%C x 106) / (1200Nc - %C.MW) 340 [µmol/m2

] 5-1

where αexp is the surface concentration, %C is the percentage of carbon of the

β-CD CSP (from elemental analysis), MW is the relative molecular mass of the chiral ligand unit and Nc is the number of carbon in the total chiral ligand

Table 5.1 Elemental Analysis and Surface Concentration for the CSPs

CSP m Spacer length Elemental Analysis Carbon (%)

αexp,ave* (µmol/m2

From Table 5.1, it is evident that ETHE-3PC has the highest surface loading, followed by ETHE-6PC and the lowest for ETHE-11PC Shorter spacer arms such as

the 3- and 6-carbon spacer might contribute lesser hindrance in the immobilization process leading to higher surface coverage For the 11-carbon spacer, it is possible for the long and flexible chain to possess considerable degree of flexibility and rotational freedom that may reduce the immobilization efficiency and lead to a lower surface coverage The efficiency of each CSP was determined under normal phase using biphenyl

as standard The efficiency was calculated to be ~40000 plates per meter for the three CSP columns

Chiral recognition abilities of the three CSPs were evaluated under normal and reversed phase It was found that the length of the spacer contributes significantly to the retentivity, selectivity and resolution of the selected enantiomers Based on the chiral recognition trends in resolving series of chiral compounds, the dependence of the enantioselectivity on the alkyl chain length was proposed

5.2 Normal Phase

Substituted (1-aryl) but-3-en-1-ols are useful synthetic precursors in organic syntheses The alcohols were evaluated on the three CSPs under the same chromatographic conditions for comparison purposes as summarized in Table 5.2 It

appears that the largest k’ values and highest enantioselectivity were observed on 6PC for all the alcohols regardless of the position of substitution

ETHE-109

Table 5.2 Enantioseparations of substituted (1-aryl) but-3-en-1-ols under

Since most of the separated compounds showed the best chiral resolution results

on ETHE-6PC, it is believed that the optimal spacer length for solute retention and selectivity was achieved on ETHE-6PC, which seems to be closed to 6 methylene groups It appears that solutes were allowed to interact more sufficiently on ETHE-6PC

A longer spacer arm on ETHE-6PC causes the CD selector further away from the silica

surface and reduces the achiral H-bonding interaction between the solutes and the hydroxyl groups on the silica surface The diminished participation of achiral molecular structures is helpful for the improvement of the enantioselectivities In related studies done by the Chirosep group, a pentenyl moiety was determined to be the optimal spacer.14 Figure 5.2 gives the plot of separation factor (α) versus spacer arm length for the racemic samples

0 1 2 3

Under normal phase, the highest enantioseparation abilities were afforded by the

6-carbon spacer CSP ETHE-6PC for selected flavanones in Table 5.3 It appears that CSP ETHE-6PC displays greater enantioseparation abilities than the other two CSPs for

most of the test racemic compounds

Table 5.3 Enantioseparations of selected flavanones under normal phase

S/n Compound k’1 k’2 α Rs Spacer Length

The chiral recognition ability of ETHE-11PC appears to be less satisfactory and

afforded the poorest enantioseparations for all compounds in Table 5.2 and 5.3 The rotational freedom of the long spacer arm may have contributed to its poorer interaction with analytes There exists a possibility that the mutual interactions between the CD chiral selectors would lead to a possible decrease of CSP-solute interactions (Figure 5.4)

The poor solute retentivity on ETHE-11PC might also be attributed to its low surface

loading Often, stronger retention and enantioselectivity are observed for a CSP with higher surface loading

113

Figure 5.4 Possible mutual interactions between chiral selectors.

In summary, ETHE-6PC affords the highest solute retentivity and the best

enantioselectivity for almost all the test racemic compounds under normal phase Therefore, it is within reason to say the 6-carbon chain spacer appears to be the optimal spacer length for a phenylcarbamate CSP to achieve better enantioseparation under normal phase

5.3 Reversed Phase

The effects of spacer length were further investigated under reversed phase (Table 5.4) The dependence of the enantioselectivity on the spacer chain length under reversed

phase was found to be similar to normal phase In Table 5.4, ETHE-6PC affords the

strongest retention and highest selectivity factor values as expected under reversed phase

Table 5.4 Enantioselectivity of test racemic compounds under reversed phase

O N

0 2 4 6

Figure 5.5 Effects of spacer length on separation factor under reversed phase

It shows that the effects of spacer length are similar under normal and reversed phase for the three CSPs Mobile phase conditions and chiral discriminating mechanisms

have less influence on the alkyl chain length In both phases, the 11-carbon spacer CSP ETHE-11PC affords the poorest chiral recognition abilities Based on the results, the 6- carbon chain spacer on the ETHE-6PC appears to be the optimal spacer length under

normal and reversed phase

of a 6-carbon chain for achieving optimal enantioselectivities This is also demonstrated

by studies in the Chirosep group that a pentenyl spacer was found to be the optimal length At the same time, in our study, surface loading is likely to play a role in chiral discrimination by increasing the odds of interaction between analytes and the chiral selectors

117

References:

1 O.A Shpigun, I.A Ananieva, N.Y Budanova, E.N Shapovalova, Russ Chem Rev.,

2003, 72: 1035

2 X H Lai, S C Ng., Tetrahedron Lett., 2003, 44 (13): 2657

3 S C Ng, T T Ong, P Fu, J Chromatogr A, 2002, 968: 31

4 M Kato, T Fukushima, T Santa, K Nakashima, R Nishioka, K Imai, Analyst,

1998, 23: 2877

5 A Berthod, C.D Chang, D.W Armstrong, Talanta, 1993, 40: 1367

6 D Kontrec, A Abatangelo, V Vinkovic, V Sunjic, Chirality, 2001, 13: 294

7 Z Chen, T Fuyumuro, K Watabe, T Hobo, Anal Chim Acta., 2004, 518: 181

8 T Hargitai, Y Kaida, Y Okamoto, J Chromatogr., 1993, 628: 11

9 M.H Hyun, M.S Na, J.S Jin, J Chromatogr A, 1996, 752: 77

10 N Bargmann-Leyder, J Chromatogr A, 1994, 666: 27

11 C Wolf, W.H Pirkle, J Chromatogr A, 1998, 799: 177

12 K Cabrera, D Lubda, G Jung, 3 rd Int Symp On chiral discrimination (3’ISCD), Poster 16, 5-8 October, Tubingen, Germany

13 G Felix, C Cachau, A Thienpont, Chromatographia, 1996, 42(9/10): 583

14 R Duval, H Leveque, C Francois, Oral Communications, 1997

Chapter 6

Optimization of Enantioseparation Conditions and Chromatographic

Properties

6.1 Introduction

Selection of an appropriate chiral CSP in HPLC analysis is crucial for enantioseparation of racemic compounds Each type of CSP affords unique enantioseparation abilities based on its chiral selector, surface coverage, spacer length etc There is always no single CSP with universal abilities to resolve all kinds of analytes Enantioseparation is influenced not only by type of the CSP; it also covers a wide range

of chromatographic parameters that would help to optimize a HPLC separation For a particular CSP chosen, there exist many methods in resolving enantiomers with better chromatographic results; these include selection of more suitable mobile phase, flow rate, control of buffer pH and concentration under reversed phase and type of polar modifier used Investigation on these parameters was carried out to optimize the chromatographic conditions In addition, studies on the above-mentioned chromatographic properties can help to provide better understanding on the roles of each parameter plays in an enantioseparation Furthermore, loading capacity of a particular CSP was studied to investigate its potential in preparative scale analysis A simple thermodynamics study

was also carried out

Selected samples (flavanone, 5-methoxyflavanone, acebutolol and triazine) were used to demonstrate the influence of HPLC parameters in the optimization of

chromatographic conditions on the phenylcarbamate columns, ETHE-3PC-L and ETHE-3PC; the former was packed in a HPLC analytical column [∅ 4.6 x 250 mm] and the latter in a microbore column [∅ 2.1 x 150 mm]

6.2 Mobile Phase Composition

For better resolution and selectivity, one can change the composition of the mobile phase.1-3 High polarity or strong eluting power solvents tend to give shorter elution time but generally poorer resolution due to insufficient interaction of analytes with the CSP On the other hand, reducing the organic content in the mobile phase will result in significant peak-tailing and broaden peaks due to long retention time There exists an optimal mobile phase composition for each separation under normal and reversed phase.4-6

6.2.1 Normal Phase

In both normal and reversed phase, solute retentivity is influenced by the composition of the mobile phase Table 6.1 lists the separation results of flavanone on

ETHE-3PC-L at different hexane/IPA composition under normal phase

Table 6.1 Separation of flavanone under normal phase on ETHE- 3PC-L

(v/v)

Chromatographic Results

Figure 6.1 Separation of flavanone at different hexane/IPA composition on

ETHE-3PC-L (a=separation factor) Hexane/IPA (v/v): A: 80/20; B: 70/30; C:

60/40

We observed a slight decrease in separation factor and resolution with an increase

in IPA content in Figure 6.1 It appears that higher IPA content gives shorter solute retention as expected

6.2.2 Reversed Phase

Under reversed phase, buffer and organic modifier content can be altered in order

to reach an optimal separation condition for a particular sample Enantioseparation of

acebutolol was evaluated on ETHE-3PC at different buffer composition in Table 6.2 It

shows that at higher MeOH content smaller k’ values were observed (Figure 6.2) This is because MeOH when compared to TEAA buffer, is less polar and will tend to compete with the analytes for the CD cavity sites Hence, inclusion complexation of the analytes will not be favored and analytes tend to be eluted faster.7 This resulted in weaker retention and faster elution of the solutes

Table 6.2 Separation of acebutolol under reversed phase

Condition: 1% TEAA pH 5.45/MeOH, flow rate at 0.2 ml/min

Stronger solute retention does not always give higher resolution At higher TEAA content, the peaks were found to have poorer resolution and peak-tailing was more significant due to mass diffusion under Condition C Optimal resolution was observed at Condition B In summary, changes in mobile phase composition greatly affect the solute retention Changes in selectivity are less significantly observed

Figure 6.2 Chromatograms of solute elution at Condition A, B and C for acebutolol

123

6.3 Effects of pH under Reversed Phase

Control of pH helps to determine the degree of ionization for analytes bearing weak acidic or basic functional groups It can help to eliminate or increase potential H-bonding interaction sites and change the interactions involved between analytes and the CSP Alteration in the amount of H-bond donors and/or acceptors often has significant effects on chiral recognition.6, 8-10 Neutral and ionized solutes perform differently as the

pH conditions are varied

A lower pH value usually results in shorter retention time It is believed that a lower pH buffer helps to ionize the analyte, making it less hydrophobic It can also help

to suppress the silanol activity on the column to decrease the unwanted interactions between basic molecules with the acidic silanols The former is important because hydrophobicity of an analyte contributes to inclusion complexation under reversed phase For most of the analytes, the presence of a weakly acidic and/or basic site on a solute requirs an optimal pH value for separation Under different pH values, the separation will change for a particular sample The greatest pH effect can be observed near the pKa value

of the compound Generally pH values accepted for CD CSPs range from 3.0 to 8.0 At higher pH values, the silica gel would be destroyed whereas at a pH lower than 3.0, cleavage of chiral ligands could be observed

In Table 6.3, a shorter retention time was observed at lower pH for acebutolol This might due to the formation of an acetate salt upon protonation of the secondary amine on the β-blocker (Figure 6.3)

Table 6.3 Separation of acebutolol on ETHE-3PC at different pH values

Condition: 1% TEAA/MeOH=60/40 and flow rate at 0.2 ml/min

The protonated, cationic species failed to form a tight inclusion complex with the

CD cavity due to its reduced hydrophobicity Ionization of the solute prevented bonding from taking place Hence, the analyte tends to be eluted faster due to poorer interaction with the CSP For acidic analytes, a low pH will suppress the ionization of the acids and make them more apt to be retained in a reversed phase system

Figure 6.3 Protonation of acebutolol at low pH

As the pH increases, effects on retention time become less significant Generally, greater resolution and separation factors were observed as the pH increases as illustrated

in Figure 6.4

125

1 1.1

pH

Separation Factor Resolution

Figure 6.4 Effects of pH on solute retentivity, selectivity and resolution

6.4 Flow Rate

Flow rate of the mobile phase plays an important role on the k’ and Rs values of a solute Generally larger k’ values were observed at slower flow rate A slower flow rate often gives better resolution because solutes are allowed to interact with the CSP sufficiently Control of the flow rate is therefore one of the easiest ways to optimize an enantioseparation though resolution of peaks within shorter retention time is always preferred Generally the flow rate of an analytical column is between 0.1 ml/min to 2.0 ml/min depending on the dimension of the column used

Table 6.4 Separation of 5-methoxyflavanone on ETHE-3PC at different flow rate

Separation of 5-methoxyflavanone on ETHE-3PC at different flow rates was

studied in Table 6.4 The highest selectivity was observed at a flow rate = 0.2 ml/min k’ values were greatly affected by the flow rate as depicted in Figure 6.5 It can be clearly seen from Figure 6.6 that a shorter retention time, sharper and closer peaks were observed

at higher flow rate; poorer resolution and broader peaks occurred at lower flow rates due

to mass diffusion

Figure 6.5 Influence of flow rate (0.1, 0.2 and 0.5 ml/min) on the enantioseparation of

5-methoxyflavanone on ETHE-3PC (a=separation factor)

Figure 6.6 Stack view of chromatograms at different flow rate

127

6.5 Organic Modifier under Reversed Phase

CD CSPs are usually used under reversed phase because inclusion complexation can be easily formed in aqueous or aqueous-organic solutions.11-14 The inclusion constant and binding strength of an inclusion complex decrease with the addition of a sufficient amount of organic modifier Two types of commonly used organic modifiers under reversed phase HPLC are MeOH and CH3CN (or ACN) It was found that CH3CN has a stronger displacement effect than MeOH due to its higher affinity with the CD cavity Because of its weaker displacement ability, MeOH is often used as the starting solvent

Separation of acebutolol in MeOH and CH3CN was evaluated on ETHE-3PC in

Figure 6.7, it appears that CH3CN reduces retention time to a great extent and affords a slight decrease in selectivity Baseline resolution of acebutolol was achieved at higher MeOH content under Conditions B and C (Table 6.5) The highest selectivity was observed at Condition C for CH3CN

Table 6.5 Separation of acebutolol at Condition B and C using MeOH or CH 3 CN as organic modifier

B

k’1=1.96 k’2=2.71 α=1.38

Rs=1.89

k’1=0.76 k’2=0.84 α=1.11

Rs=0.43

C

k’1=3.62 k’2=4.92 α=1.36

Rs=1.70

k’1=1.08 k’2=1.52 α=1.41

Rs=1.65

Condition: 1% TEAA buffer pH 4.10/organic modifier (v/v):

B: 70/30; C: 80/20

6.6 Concentration of TEAA Buffer under Reversed Phase

We observed that higher TEAA concentration usually gives shorter retention time It is believed that a higher TEAA concentration helps to mask residual silanol groups and gives weaker solute retentivity.6, 8

Table 6.6 Separation of acebutolol at different concentrations of TEAA

Faster elution of solutes was observed at higher TEAA concentration Selectivity

of a solute separation decreases slightly as the TEAA concentration increases Table 6.6

tabulates separation the results of acebutolol on ETHE-3PC under 40/60 (v/v)

MeOH/TEAA buffer at different concentrations From Figure 6.8, it appears that TEAA

% has significant influence on resolution

Figure 6.8 Effect of % TEAA (0.5, 1.5 and 2.0) on enantioselectivity for acebutolol

(a=separation factor).

6.7 Loading Capacity

Loading capacity is a critical factor in preparative HPLC separation It can be used to investigate a CSP’s potential for preparative scale separation The higher the loading capacity, the greater the potential it can be packed as a preparative column.15The CSPs should have a high sample capacity, i.e the number of enantio-differentiating sorption sites per gram of sample should be as high as possible A stationary phase composed mostly of silica gel with only a few chiral elements can be rapidly overloaded

The loading capacity is also related to the physical properties of the silica gel surface that

would change when immobilized with chiral ligands

Loading capacity was investigated by injecting the same volume (1µL) of

acebutolol at different sample concentrations (in terms of mg) The analysis was carried

out under reversed phase at 1% TEAA pH 5.45/MeOH = 60/40 and flow rate at 0.2

ml/min The separation results are tabulated in Table 6.7 The effects of sample

concentration on k’, selectivity and resolution are illustrated in Figure 6.9 Optimal

enantioselectivity was achieved on sample concentration of 0.15 mg for acebutolol under