Application of preferential crystallization for racemic compound integrating thermodynamics, kinetics and optimization 5

Furthermore, a cyclic crystallization process, which provided alternating the pure mandelic acid enantiomer and the racemic compound, was proposed Lorenz et al., 2006b.. Fig 6.3 Chromato

Chapter 6 Systematic preferential crystallization process of

mandelic acid

6.1 Introduction

According to the introduction in Chapter 2, the applicability of preferential crystallization to racemic compounds would significantly widen the potential of usually cheap crystallization based techniques for enantioseparation (Jacques et al., 1981; Brock and Dunitz, 1994; Kinbara et at., 2001; Lorenz et al., 2001, 2006; Profir

et al., 2002) During the crystallization process, it is crucial to keep the freedom of supersaturation of the racemate in its metastable zone to avoid its spontaneous nucleation and to control the supersaturation of the target enantiomer as the spontaneous nucleation of the target enantiomer may easily initiate the spontaneous nucleation of its racemate

The model compound in this chapter is mandelic acid, which is a widely used reagent in classical resolution It has attracted some efforts to study the crystallization process and relevant data measurement due to its relatively cheap price and favorable phase diagram characteristics for preferential crystallization For example, the coupling process of liquid chromatography and preferential crystallization was suggested for efficient enantioseparation of the enantiomers of mandelic acid in aqueous solution by Lorenz et al (2001) The principle based on the thermodynamic phase diagram for the crystallization was introduced in this work The binary and ternary phase diagrams of mandelic acid enantiomer in water system were constructed and DSC was applied successfully for the solubility determination ((Lorenz and Seidel-Morgenstern, 2002; Mohan et al., 2002) Profir and Rasmuson (2004) reported that metastable conglomerate mandelic acid crystals can be formed upon primary nucleation in water and acetic acid After a time-lag, the conglomerate was transformed into the stable form racemic compound The time-lag range depended on

the operation conditions and decreased at increasing concentration or temperature and

in the presence of micrometer-size particles (Profir et al., 2002) They also studied the influence of solvent and the operation conditions, such as filtration, cooling rate and stirring rate, on the crystallization of mandelic acid (Profir and Rasmuson, 2004)

In addition, for the isothermal batch crystallization of the mandelic acid in water, different analytical techniques were evaluated to determine the solute concentration in the liquid phase and the metastable zone width measurement were presented, obtaining the basis for growth kinetics investigation (Perlberg et al., 2005) Preferential crystallizations of enantiomers were also discussed for the two enantiomeric systems, conglomerate threonine and racemic compound mandelic acid

by Lorenz et al (2006a) From this work, the 99.3% purity of (S)-MA was obtained

As well, the effects of the presence of the counter-enantiomer on the growth rate and crystal shape were illustrated in this work Furthermore, a cyclic crystallization process, which provided alternating the pure mandelic acid enantiomer and the racemic compound, was proposed (Lorenz et al., 2006b) During this process, the online polarimetry and online density measurement were used in application of preferential crystallization for mandelic acid Recently, the potential crystallization inhibitors were used on the chiral separation for mandelic acid (Mughal et al., 2007) The results showed that while none of the additives dramatically inhibit the

crystallization of (S)-MA, they significantly inhibit the crystallization of racemate

MA This may lead a new way to a crystallization process for the chiral separation

metastable zone, kinetics and supersaturation control profile to obtain crystal product with good quality

In this chapter, the certain enantiomerically enriched mandelic acid was obtained from a racemic composition by using a HPLC with a semi-preparative chiral column Based on the solubility and metastable zone data derived in the chapter 4 and kinetic data obtained in chapter 5, we demonstrated a systematic approach, in which a modified strategy and optimal operation profile were proposed for a racemic compound system Three different cooling profiles were applied to the subsequent crystallization process The final product’s optical purity, yield and crystal size distribution were examined and compared

6.2 Experiment

6.2.1 Semi-preparative HPLC separation of mandelic acid

Semi-preparative HPLC separation of mandelic acid was performed using

Chiralcel AD-H semi-preparative HPLC column (dimension 250mm L x 10mm I.D) The mobile phase is Heptane/TFA/IPA (95/0.1/5 v/v) at 25°C column temperature, with flow rate of 4.5ml/min and UV-Vis detection at 210nm

6.2.2 Direct crystallization operation

The crystallization experiments were carried out in an automatic laboratory reactor system equipped with an l-L glass jacketed crystallizer as described in the Fig 4.2 Three types of cooling profiles, controlled cooling, linear cooling and forced

mandelic acid The start point of all experiments was the same solution which was the

80% mole percent (S)-MA saturated solution at 35 oC The crystallizer was kept 5 oC higher than that of the saturated solution in order to assure that no crystal exited in the solution prior to crystallization After 0.5h, the temperature of the crystallizer was reduced to the saturation temperature The batch cooling crystallization experiment

was started When the temperature decreased slightly, the pure (S)-MA seeds prepared

by fast cooling from the solution of pure (S)-MA were added into the solution Then

the temperature was cool down according to the different cooling profile with the range of 35-20 oC During these experiments, several samples were taken at definite time intervals to observe the optical purity of crystal products, solute concentration in liquid phase, and crystal size distribution

The optical purity of crystal products were measured by using DSC, Polarimeter, and HPLC In this work, the polarimeter was equipped with a sodium vapor lamp emitting light with a wavelength of 589.3 nm and a quartz cell of 50 mm path length HPLC was an Agilent 1100 series HPLC system with Chiralcel AD-H analytical column (dimension 250mm L x 4.6 mm I.D.) under separation conditions: Heptane /TFA/IPA (95/0.1/5 v/v) at 25°C column temperature, flow rate of 1.0ml/min and UV-Vis detection at 210nm The crystal size distribution was measured using a Malvern Mastersizer 2000 and the solute concentration was measured using Shimadzu

2450 UV-visible spectrophotometer

The loading capacity of Mandelic acid on Chiralcel AD-H was determined by injecting different amount of sample onto the column It was found that mandelic acid shows baseline separation at 8.07mg loading, while partial separation of mandelic acid are observed when sample loading increases to 16.15mg, shown in Fig 6.1

Fig 6.1 Partial separation of MA on Chiralcel AD-H semi-preparative HPLC column (dimension 250mm L x 10mm I.D) at different loadings 8.07mg and 16.15mg per injection using Heptane/TFA/IPA (95/0.1/5 v/v) as mobile phase, at 25°C column temperature, flow rate of 4.5ml/min and UV-Vis detection at 210nm

The sample loading was determined as 16.15 mg per analysis Through preparative chiral HPLC separation, we have successfully isolated the 80% mole

semi-percent S enantiomer and pure R enantiomer from its racemate, by collecting two

different fractions at two different retention time, t =27-37 mins and t =37-43 mins It

is presented in Fig 6.2

Fig 6.2 Fraction collection under semi-preparative HPLC separation of MA on

Chiralcel AD-H column (dimension 250mm L x 10.00 mm I.D.) under separation conditions: Heptane/TFA/IPA (95/0.1/5 v/v) at 25°C column temperature, flow rate of 4.5ml/min and UV-Vis detection at 210nm Fraction (a) collected at retention time 27-

37 minutes and fraction (b) is collected at 37-43 minutes

The optical purities of these two fractions were analyzed on analytical chiral column, Chiralcel AD-H using Heptane /TFA/IPA (95/0.1/5 v/v) as the mobile phase and a flow rate of 1.0ml/min The results are shown in Fig 6.3 These two fractions could be accumulated until a certain volume enough for the preferential crystallization process by using continuous HPLC separation with an antosampler and automated fraction collector

Fig 6.3 Chromatogram of two fractions (a) and (b) obtained through semi-preparative HPLC separation of mandelic acid on Chiralcel AD-H analytical column (dimension 250mm L x 4.6 mm I.D.) under separation conditions: Heptane/TFA/IPA (95/0.1/5 v/v) at 25°C column temperature, flow rate of 1.0 ml/min and UV-Vis detection at

210nm

6.3.2 Preferential crystallization operation for mandelic acid

6.3.2.1 The progression of preferential crystallization

A typical loading of preferential crystallization of mandelic acid was the

saturated solution with 80% mole percent (S)-MA at 35 oC According to the results of solubility and MSZWs of MA in chapter 4, this saturated solution was saturated in the eutectic composition of MA (Eu-MA) at 25.5 oC and saturated in the (RS)-MA at 23.3

o

C Six batches direct crystallization experiments were carried out starting from the same solution with different modes, which are (a) Exp_01: with seeding and final temperature at 28 oC; (b) Exp_02: with seeding and final temperature at 25.5 oC; (c) Exp_03: with seeding and final temperature at 24 oC; (c) Exp_04: with seeding and final temperature at 23.3 oC; (e) Exp_05: with seeding and final temperature at 22.5

o

C; (f) Exp_06: without seeding until nucleation occurred

As described in Fig 6.4, a saturation solutionwith 80% mole percent (S)-MA at

35 oC is repented by point A The saturated point of Eu-MA and (RS)-MA are

represented at point B and C respectively When the controlled cooling profile was

used, the solution seeded with (S)-MA was cooled in different runs to 28, 25.5, 24,

23.3 and 22.5 oC, shown in Fig 6.4 as well The optical purity of the final crystal products with different cooling degree was analyzed by HPLC and polarimeter and the results are listed in Table 6.1 The calibration curve of optical rotation with

concentration for pure (S)-MA will be shown in section 6.3.2.3

BC

Exp 1Exp 2Exp 3Exp 4

Exp 5Exp 6

Fig 6.4 Progression of direct crystallization of mandelic acid

Table 6.1 The optical purity of the final crystal products with different cooling degree

for mandelic acid

Experiment

Optical purity of

product (% S

enantiomer) from polarimeter

Optical purity of

product (% S

enantiomer) from HPLC

Cooling degree

The interesting results were found here The product crystals were almost pure

(S)-enantiomer when the end temperature was higher than 23.3 oC (Exp_01-04) In

this region, only (S)-MA is supersaturated This implies that the pure enantiomer

product can be obtained by well designed and controlled direct crystallization However, when the crystallization final temperature was lower than 23.3 oC, such as Exp_05 at 22.5 oC, (RS)-MA began to supersaturate and the product crystals were in the form of mixture of (RS)-MA and (S)-MA This may suggest that there is no

selectivity of crystal growth of the pure enantiomer and racemate for a racemic

by well designed and controlled procedure (Wang and Ching 2006) However, for the

racemic compound, it was found here that only excess S enantiomer can be obtained

by the preferential crystallization process

When the solution was cooled to 21.5 oC without seeding as Exp_06, the primary nucleation occurred by exceeding the metastable zone of initial composition

mandelic acid The products were always in the form of (RS)-MA and (S)-MA This

could be due to the crucial characteristic of a racemic compound: no selectivity of nucleation between the pure enantiomer and racemate for a racemic compound

So, the optical purities of these experiments results show that almost pure S

mandelic acid crystal product was obtained from the preferential crystallization with

seeding within the safe supersaturation critical limit, while both S and RS were

crystallized out when the system temperature exceeded the saturated temperature of racemate It indicated that the key factor to get the pure enantiomer product is its solubility characteristic for direct crystallization of racemic compound

It was also found that though the end temperature reached the Eu-MA saturated temperature 25.5 oC, the eutectic component crystal of mandelic acid did not come out

from the solution and the pure S enantiomer was still obtained as a product It may be

explained that the eutectic mandelic acid is still in the metastable zone area, though the end temperature already surpass its saturated temperature Or it could be due to that the eutectic mandelic acid can not come out as a stable compound from the solution So during the course of the whole direct crystallization process, the

competition could be only between (S)- and (RS)- mandelic acid crystal in the solution

6.3.2.2 Optimal operation profile

In order to realize the supersaturation control in preferential crystallization process, an optimal operation profile should be used In this work, the classical simplified equation derived by Nyvlt et al (1973) for optimal operation of batch

crystallization with seed was used with assuming constant crystal growth rate G and constant nucleation rate B (Eq 6-1) As discussed in Wang and Ching’s work (2006),

the thermodynamics and crystallization kinetics should be combined together to apply this equation to control the supersaturation

3 0

0 )/( ) ( / )(T −T T −T f = t t c (6-1)

From the thermodynamics point of view, the 80% mole percent (S)-MA

saturated solution at 35 oC was used as the starting point for the preferential crystallization operations According to the solubility data, this saturated solution was

saturated in the (RS)-MA at 23.3 oC That is to say that only S enantiomer is

supersaturated in the temperature range of 35 to 23.3 oC After continually cooling,

the spontaneous nucleation of the racemic RS may be carried out According to

MSZW data in chapter 4, the supersaturation should be kept within circa 5 oC to avoid spontaneous nucleation of both enantiomers Also, the applicable supersaturation for

RS racemate is 5.5-25 oC Therefore it is acceptable to set the final targeted