Chiral purification of S-ibuprofen from ibuprofen enantiomers by stripping crystallization

chemical engineering research and design 1 1 7 ( 2 0 1 7 ) 301–308

Contents lists available at ScienceDirect

Chemical Engineering Research and Design journal homepage: www.elsevier.com/locate/cherd

Chiral purification of S-ibuprofen from ibuprofen enantiomers by stripping crystallization Lie-Ding Shiau a,b,∗ , Keng-Fu Liu a , Yu-Chao Hsu b a b

Department of Chemical and Materials Engineering, Chang Gung University, Taoyuan, Taiwan, ROC Department of Urology, Chang Gung Memorial Hospital, Linko, Taiwan, ROC

a r t i c l e

i n f o

a b s t r a c t

Article history:

A new technology, stripping crystallization (SC), is introduced in this work for chiral purifi-

Received 4 January 2016

cation of S-ibuprofen from ibuprofen enantiomers. Basically, SC combines distillation and

Received in revised form 20 June

crystallization operated at reduced temperature and pressure for the liquid mixture feed

2016

to produce pure S-ibuprofen crystals and mixture vapors by maintaining a series of the

Accepted 9 October 2016

three-phase equilibrium conditions based on the variations of the liquid composition. SC is

Available online 20 October 2016

continued until liquid is nearly eliminated. The final product only consists of S-ibuprofen

Keywords:

is developed to simulate the three-phase equilibrium during the SC operation and to direct

Crystallization

the batch SC experiments. The experiments show that, when SC is operated from 50 ◦ C

crystals as all the vapors produced are removed from the system. A thermodynamic model

Vaporization

and 290 Pa to 37 ◦ C and 148 Pa, seeding with ultrasound mixing or magnetic stirring can

Purification

be efficiently employed to purify S-ibuprofen from ibuprofen enantiomers. The experimen-

Thermodynamics process

tal results, including the final enantiomeric purity and recovery ratio of S-ibuprofen, are

Ibuprofen

consistent with the simulation results predicted by the model. © 2016 Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

1.

Introduction

Stripping crystallization (SC) is a new technology which combines distillation and crystallization to separate the mixtures with close boiling temperatures (Shiau et al., 2005, 2006, 2008; Shiau and Yu, 2009; Shiau and Liu, 2013). No solvent is added when SC is applied to produce

might cause some long-term side effects (Sheldon, 1993), ibuprofen is still sold as a racemic mixture in the market due to the complexity involved in the separation of enantiomers. Various separation methods to obtain enantiopure ibuprofen have been proposed in the literature, including chromatographic separation (Peper et al., 2002; Park et al., 2008), crystallization method (Zey et al., 1992; Trung et al., 2006), enzy-

crystals from the liquid mixture. As opposed to the solid–liquid equi-

matic separation (Huh et al., 2006; Kim et al., 2010), and membrane

librium involved in melt crystallization (Ulrich, 2003; Jiang et al., 2012;

separation (Cauwenberg et al., 1999; Long et al., 2005; Wang et al., 2007). The objective of this research is to study the feasibility of SC in

Micovic et al., 2013; Beierling et al., 2014), SC is operated at a triple-point condition, in which the liquid mixture is simultaneously vaporized and crystallized due to the three-phase equilibrium. By lowering tem-

purification of S-ibuprofen from ibuprofen enantiomers. The effect of seeding on the enantiomeric purity and recovery ratio of final Sibuprofen crystals is investigated. The results should provide important

perature and reducing pressure during the operation, SC results in the formation of pure crystals, and liquid phase and vapor phase of mixtures. In essence, SC is continued until the liquid phase is nearly

information in the pharmaceutical industry.

eliminated and only pure crystals remain in the feed.

1.1.

Ibuprofen, also called 2-(isobutylphenyl)-propionic acid, is a nonsteroidal anti-inflammatory drug widely used to treat headaches and minor pains (Adams et al., 1976). Although it is known that S-ibuprofen is responsible for the anti-inflammatory effects while R-ibuprofen

Principle of SC

The basic principles of the SC process can be explained by referring to the phase diagrams. In Fig. 1(a), the upper part illustrates the ideal vapor–liquid equilibrium (VLE) phase diagram for R-ibuprofen

∗ Corresponding author at: Department of Chemical and Materials Engineering, Chang Gung University, Taoyuan, Taiwan, ROC. Fax: +886 3 2118700. E-mail address: [email protected] (L.-D. Shiau). http://dx.doi.org/10.1016/j.cherd.2016.10.019 0263-8762/© 2016 Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

302

chemical engineering research and design 1 1 7 ( 2 0 1 7 ) 301–308

Nomenclature Hm,j Hm,R HV,j Ln P Pj sat R RB,E RB,S Sn T Tb,j Tm,j Tm,R t Vn W0 Wf Xj XB,0 XB,f XB,S

Yj

Heat of melting for j-component (>0), J/mol Heat of melting for rac-buprofen (>0), J/mol Heat of vaporization for j-component (>0), J/mol Mass of the liquid phase out of stage n, g Pressure, Pa Saturated pressure of the liquid of jcomponent, Pa Ideal gas constant, 8.314 J/mol-K Experimental recovery ratio, dimensionless Simulated recovery ratio, dimensionless Mass of the solid phase out of stage n, g Temperature, K Boiling temperature of j-component, K Melting temperature of j-component, K Melting temperature of rac-ibuprofen, K Time, min Mass of the vapor phase out of stage, g Initial weight of the liquid mixture feed, g Final weight of the liquid mixture feed, g Mole fraction of j-component in liquid phase, dimensionless Mole fraction of B-component in the liquid mixture feed, dimensionless Experimental purity of B-component in the final product, dimensionless Simulated purity of B-component in the final product predicted by simulation, dimensionless Mole fraction of j-component in vapor phase, dimensionless

Greek letters [˛]20 Specific optical rotation at 20 ◦ C for the sodium D light (589 nm) j Activity coefficient of j-component in liquid phase, dimensionless Subscript 0 f j n

pure S-ibuprofen or pure R-ibuprofen. It shows the existence of two three-phase states at T = 47.4 ◦ C. One is a three-phase state (points b and b ) having pure S-ibuprofen crystal (point b), and liquid phase and vapor phase of mixtures at XB = 0.90 (point b ) on the right-hand side of the figure. Note that the liquid composition is the same as the vapor composition at point b . The other is a three-phase state (points a and a ) having pure R-ibuprofen crystal (point a), and liquid phase and vapor phase of mixtures at XB = 0.12 (point a ) on the left-hand side of the figure. Note that the liquid composition is the same as the vapor composition at point a . Thus, SC provides a potential method not only to purify S-ibuprofen in the range 0.82 < XB < 1 but also to purify Ribuprofen in the range 0 < XB < 0.18. Note that cocrystallization occurs at Teu . 1.2.

Simulation of SC process

The SC process is simulated in a series of stage operations shown in Fig. 2, which can also be an abstract representation of a batch process in a single vessel. Each stage corresponds to a three-phase equilibrium at a given time, tn . During the temperature-decreasing operation in a batch process, the temperature of each stage is chosen to meet Tn−1 − Tn = T for n = 1, 2, ..., N. Note that T0 is the triple-point temperature of the mixture feed. The vapor formed in each stage is condensed to the liquid and removed while the solid and the liquid formed in each stage enter the next stage. The whole process starts from the liquid mixture feed and is usually stopped at Teu when no vaporization occurs. When SC is applied to purify S-ibuprofen from the liquid mixture feed in the range 0.82 < XB < 1, each stage is maintained at a threephase equilibrium state having pure S-ibuprofen crystals, and liquid phase and vapor phase of mixtures. Due to the formation of S-ibuprofen crystals in each stage, the liquid composition of S-ibuprofen decreases during the batch process. The corresponding three-phase equilibrium condition in each stage can be determined as follows. The SLE of S-ibuprofen in stage n is generally described by the Schroder–Van Laar equation as (Jacques et al., 1981) ln (XB )n =

Hm,B R



1 Tm,B

1 Tn

−

 (1)

The physical properties of S-ibuprofen and R-ibuprofen needed in the simulation are taken from Table 1. However, as there are great deviations between the calculated SLE data from Eq. (1) and the experimental SLE data obtained by Dwivedi et al. (1992), the SLE of S-ibuprofen in stage n is described by the following experimentally fitted equation as

In the feed In the final product Component j (j = A or B) In stage n

2

Tn = −150.2(XB )n + 302.2(XB )n − 102.3

(2)

Note that the operable range for crystallization is from 50 ◦ C to 44 ◦ C for 0.82 < XB < 1 in Fig. 1(a). As SC is operated at low pressures (<0.101 MPa), the VLE in stage n can be described by (Smith et al., 2001; Sandler, 2006)

(A-component) and S-ibuprofen (B-component) while the lower part illustrates the experimental solid–liquid equilibrium (SLE) phase diagram obtained by Dwivedi et al. (1992). Thus, ibuprofen can form a racemic compound, which is characterized by a crystal form in which the two enantiomers coexist in the same unit cell (Jacques et al., 1981). In VLE phase diagram, the equilibrium liquid line coincides with the equilibrium vapor line due to the same saturated vapor pressure for Ribuprofen and S-ibuprofen. In SLE phase diagram, two eutectic points exist at T = 44.3 ◦ C, one at XB = 0.18 and the other at XB = 0.82.

(YA )n Pn = (XA )n (A )n (PA sat )n

(3)

(YB )n Pn = (XB )n (B )n (PB sat )n

(4)

As the temperature-dependent saturated vapor pressure data for S-ibuprofen and R-ibuprofen are not available, the Clausius–Clapeyron equation is adopted to describe the dependence of the saturated vapor pressure on temperature (Smith et al., 2001; Sandler, 2006), Thus,

As pressure is reduced, SLE usually remains almost the same while VLE will be moved downward. Fig. 1(b) illustrates the solid–liquid–vapor equilibrium (SLVE) phase diagram at P = 289 Pa, which is the triple-point pressure of pure S-ibuprofen or pure R-ibuprofen. It shows the existence of two three-phase states at T = 50 ◦ C. One is a three-phase state of pure S-ibuprofen (point b) on the right-hand side of the figure. The other is a three-phase state of pure R-ibuprofen (point a) on the left-hand side

 ln

Pj sat (T) Pj

sat

(Tb,j )

 =

HV,j R



1 1 − Tb,j T

 (j = A or B)

(5)

As Pj sat (Tb,j ) = 0.101 MPa (j = A or B), Pj sat (T) can be subsequently determined. Due to the structure similarity between S-ibuprofen and

of the figure.

R-ibuprofen, it is assumed that A = 1 and B = 1. Besides, we have

As pressure is further reduced, Fig. 1(c) illustrates the SLVE phase diagram at P = 252 Pa, which is lower than the triple-point pressure of

(XA )n + (XB )n = 1

(6)

303

chemical engineering research and design 1 1 7 ( 2 0 1 7 ) 301–308

Fig. 1 – (a) The experimental SLE phase diagram obtained by Dwivedi et al. (1992) and the ideal VLE phase diagram at P = 0.101 MPa; (b) the predicted SLVE phase diagram at P = 289 Pa; (c) the predicted SLVE phase diagram at P = 252 Pa.

Fig. 2 – Simulated SC operation where each stage is operated at a three-phase equilibrium state. Table 1 – Some physical properties of R-ibuprofen, S-ibuprofen and rac-ibuprofen. Property

R-ibuprofen

S-ibuprofen

Molecular structure

Molecular weight Boiling point, ◦ C (Lerdkanchanaporn and Dollimore, 1997) Melting point, ◦ C (Dwivedi et al., 1992) Triple-point pressure,a Pa (N/m2 ) Heat of melting, J/mol (Dwivedi et al., 1992) Heat of vaporization, J/mol (Lerdkanchanaporn and Dollimore, 1997) a

Estimated by Clausius–Clapeyron equation.

rac-ibuprofen

–

206 231

206 231

206 231

50 289 2.0 × 104 4.4 × 104

50 289 2.0 × 104 4.4 × 104

77 1020 2.7 × 104 4.4 × 104

304

chemical engineering research and design 1 1 7 ( 2 0 1 7 ) 301–308

as displayed in Fig. 3. Note that pure S-ibuprofen solid is formed at the three-phase equilibrium. Once Tn , Pn , (XA )n , (XB )n , (YA )n and (YB )n are determined, Sn , Ln and Vn can be calculated as follows for a liquid feed of L0 with a known (XB )0 . The total material balance in stage n can be written as Sn−1 + Ln−1 = Sn + Ln + Vn

(9)

where S0 = 0 for a liquid feed. As pure S-ibuprofen is crystallized in the solid phase, R-ibuprofen only exists in the liquid and vapor phases. The material balance of R-ibuprofen in stage n can be written as Ln−1 (XA )n−1 = Ln (XA )n + Vn (YA )n

(10)

As the liquid mixture is simultaneously vaporized and crystallized due to the three-phase equilibrium in each stage, it is assumed that the latent heat released in forming crystals is removed by vaporizing portions of liquid mixtures in each stage. Thus, the energy balance in stage n is described by

Fig. 3 – The resulting P(T), XB (T) and YB (T) predicted by the model for the three-phase equilibrium. (YA )n + (YB )n = 1

(7)

When Tn is specified, Eqs. (2)–(4), (6) and (7) constitute a set of five equations that can be simultaneously solved for five unknowns—Pn , (XA )n , (XB )n , (YA )n and (YB )n for n = 1, 2, ..., N. It should be noted that Eq. (5) yields (PA sat )n = (PB sat )n due to Tb,A = Tb,B and HV,A = HV,B . Combining Eqs. (3) and (4) leads to Pn = (PA sat )n = (PB sat )n , (XA )n = (YA )n and (XB )n = (YB )n . Fig. 3 displays the variations of P, XB and YB with T by solving Eqs. (2)–(4), (6) and (7) simultaneously. Thus, P(T) in Fig. 3 represents the three-phase equilibrium pressure during the temperature-decreasing process. Fig. 3 shows that the three-phase equilibrium for pure S-ibuprofen lies at 50 ◦ C and 289 Pa. Then, both the three-phase equilibrium temperature and pressure decrease as XB decreases. By definition, the phase rule is given as (Smith et al., 2001; Sandler, 2006) F =C+2−

(8)

As there are two components at the three-phase equilibrium, Eq. (8) yields F = 1 due to C = 2 and  = 3. Thus, the degree of freedom equals 1. When Tn in each stage is specified at the three-phase equilibrium, Pn , liquid composition and vapor composition are implicitly determined

(Sn − Sn−1 )Hm,B = Vn [(YA )n HV,A + (YB )n HV,B ]

(11)

where Sn − Sn−1 represents the amount of liquid crystallized in stage n while Vn represents the amount of liquid vaporized in stage n. Thus, Sn , Ln and Vn can be obtained by solving Eqs. (9)–(11) simultaneously for n = 1, 2, ..., N. Note that Eq. (11) can be simplified as (Sn − Sn−1 )Hm,B = Vn HV,B due to HV,A = HV,B .

2.

Experimental

The experimental assembly consists of a 5-mL sample container in a 3-L stainless chamber as shown in Fig. 4. The whole chamber is fitted with a cooling jacket for temperature control. A mechanical vacuum pump is used to lower the pressure in the chamber. A temperature probe is positioned in the center of the liquid feed and a pressure gauge is connected to the chamber. S-ibuprofen (purity > 98.9%) and rac-ibuprofen (purity > 99.5%) are purchased from Tokyo Chemical Industry. The mixture is prepared by mixing S-ibuprofen and racibuprofen to obtain the desired feed concentration. As shown in Fig. 1, SC can be applied to purify S-ibuprofen from ibuprofen enantiomers in the range 0.82 < XB < 1 while cocrystallization occurs at XB = 0.82 and Teu0 = 44.3 ◦ C. At the beginning of the experiments, 1 g heated liquid mixture is

Fig. 4 – Schematic diagram of the experimental apparatus for the SC operation with the features: (1) mechanical pump, (2) constant temperature bath, (3) thermocouple, (4) sample container, (5) pressure gauge, (6) stirrer.

305

chemical engineering research and design 1 1 7 ( 2 0 1 7 ) 301–308

Fig. 5 – Schematic diagram of the batch experiments: at t = 0, only liquid mixture feed in the container; at 0 < t < tf , three-phase equilibrium of pure crystals, liquid mixture, and vapor mixture; at tf , only pure crystals in the container. put in the sample container. Then, the sample temperature is lowered gradually from the melting point (50 ◦ C). The calculated P(T) in Fig. 3 is adopted to control the three-phase equilibrium conditions regardless of the feed concentration. Fig. 5 shows the schematic diagram of the batch experiments, in which the liquid mixture is simultaneously vaporized and crystallized due to the three-phase equilibrium. The vapor is condensed and removed from the sample container. The final sample remained in the vessel at the end of the experiments is weighed. The enantiomeric purity of the final sample is analyzed by Polarimeter (Horiba, model: SEPA-300). The polarimetry is measured by dissolving 0.1 g final product in 50 mL ethanol (purity > 99.5%). First, a plot of the measured specific optical rotation versus the known enantiomeric purity within the range XB,0 = 0.9 − 1.0 is fitted with a linear regression line with R2 = 0.995. Then, by measuring the specific optical rotation of the final sample the enantiomeric purity can be deter0 mined from the plot. Note that [˛]20 D = −59 for R-ibuprofen 20 0 and [˛]D = +59 for S-ibuprofen. The polarimetry of the mixture feed remains unchanged when the sample mixture is heated to 50 ◦ C at 1 atm for over 30 min, which implies that no reactions occur during the whole heating experiments. Thus, it is assumed that no reactions occur during the experiments as SC is operated at reduced temperatures and pressures. It is found that the experimental purity and recovery ratio of the resulting product are nearly independent of various cooling rates ranging from 0.10 ◦ C/min to 0.50 ◦ C/min. This implies that three-phase equilibrium can be achieved in each stage if the liquid mixture is cooled in the above-mentioned range. In the current work, the cooling rate is generally controlled at 0.20 ◦ C/min and a typical cooling curve during the batch experiment is plotted in Fig. 6.

3.

Results and discussion

Table 2 lists the simulated results for 1 g liquid feed with XB,0 = 0.90 using the proposed model for T = 0.73 ◦ C. In the simulation, N = 15 is adopted to simulate the temperaturedecreasing batch operation. As vaporization still occurs when the SC experiment is operated to Teu (44.3 ◦ C) in stage n = 5. Therefore, the SC experiment is operated to 37.0 ◦ C in stage n = 15 when no vaporization is observed. Table 2 indicates that the three-phase equilibrium for 1 g liquid feed with XB,0 = 0.90 occurs at T0 = 48 ◦ C and P0 = 261.7 Pa. As displayed in Fig. 3, once Tn is specified, Pn and (XB )n are determined by solving Eqs.

Fig. 6 – A typical cooling curve during the batch experiment for 1 g mixture feed with XB,0 = 0.90.

Table 2 – The simulation results for 1 g liquid feed with 90% S-ibuprofen and 10% R-ibuprofen (T = 0.73 ◦ C). n

T (◦ C)

P (Pa)

XB

L (g)

S (g)

V (g)

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

48.0 47.3 46.6 45.9 45.2 44.3 43.7 43.0 42.3 41.6 40.8 40.1 39.3 38.5 37.7 37.0

261.7 252.2 243.1 234.1 225.4 217.2 209.1 201.3 193.7 186.4 179.3 172.5 166.0 159.6 153.5 147.6

0.90 0.879 0.862 0.846 0.832 0.818 0.806 0.794 0.783 0.773 0.763 0.753 0.744 0.735 0.726 0.718

1 0.752 0.612 0.521 0.457 0.408 0.370 0.340 0.315 0.293 0.275 0.260 0.246 0.234 0.223 0.213

0 0.171 0.267 0.329 0.374 0.407 0.433 0.454 0.471 0.486 0.498 0.509 0.518 0.527 0.534 0.541

0 0.078 0.044 0.028 0.020 0.015 0.012 0.0095 0.0079 0.0066 0.0057 0.0049 0.0043 0.0038 0.0034 0.0030

306

chemical engineering research and design 1 1 7 ( 2 0 1 7 ) 301–308

Fig. 7 – The variations of S, L and S + L with the operating temperature during the SC operation for 1 g mixture feed with XB,0 = 0.90. (2)–(4), (6) and (7) simultaneously for n = 1, 2, ..., N. Then, Sn , Ln and Vn are by solving Eqs. (9)–(11) simultaneously. As shown in Fig. 7, Sn decreases and Ln increases during the temperaturedecreasing operation. Note that Sn + Ln represents the total weight of the product, including the final crystals and the liquid, remained in the sample container as all the vapors are removed from the system. As indicated in Table 2, some liquid still remains along with the final crystals in stage n = 15. Thus, the simulated purity of the final product, including the final crystals and the remaining liquid, is defined as XB,S =

SN + LN (XB )N SN + LN

Fig. 8 – The simulated final purity of the product versus temperature in purification of S-ibuprofen ( represents initial purity, 䊉 represents simulated final purity, and the number in the parenthesis next to the ending point of each curve represents the simulated recovery ratio).

(12)

The simulated recovery ratio of S-ibuprofen is defined as RB,S =

SN + LN (XB )N W0 XB,0

(13)

The similar approach is applied for 1 g liquid feed with XB,0 = 0.93, 0.95 and 0.975, respectively. The calculated final purity of the product versus the final operating temperature, XB,S (T), is plotted in Fig. 8 as four curves, one for each feed. The starting point of each curve refers to the feed purity and the initial SC operating temperature, while the ending point of each curve refers to the simulated product purity and the final SC operating temperature. The number in the parenthesis next to the ending point of each curve represents RB,S . For example, the simulation predicts that 1 g liquid feed at XB,0 = 0.90 can be purified to XB,S = 0.92 with RB,S = 77% as SC is operated from and 48 ◦ C to 37 ◦ C for N = 15. Then, the SC experiment for 1 g liquid feed at XB,0 = 0.90 is performed based on the three-phase equilibrium condition in each stage for N = 15 as listed in Table 2. The experimental and simulation results are compared in Fig. 9. Although pure S-ibuprofen crystals should be formed based on the three-phase equilibrium, the resulting Sibuprofen crystals during the experiments might contain

Fig. 9 – Comparison between XB,f and XB,S for purification of S-ibuprofen (. . .. . .. . . represents XB,S ,  represents no represents no seeding with seeding and no stirring, ultrasound mixing,  represents no seeding with magnetic stirring,  represents seeding without magnetic stirring or ultrasound mixing, represents seeding with ultrasound mixing, 䊉 represents seeding with magnetic stirring). R-ibuprofen due to the following reasons: (1) incorporation of R-ibuprofen into the S-ibuprofen crystal lattice in crystal growth due to the structure similarity; (2) liquid inclusion in crystal growth; (3) incomplete elimination of liquid among crystals at the end of experiments. Fig. 9 compares the experimental and simulation results for various feeds (1 g liquid feed with XB,0 = 0.90, 0.93, 0.95 and 0.975). The dotted line represents XB,S calculated by Eq. (12),

chemical engineering research and design 1 1 7 ( 2 0 1 7 ) 301–308

which yields RB,S = 71% − 77% for XB,0 = 0.90 − 0.975 as shown in Fig. 8. Each data point represents the average XB,f for four repetitive experiments. The number in the parenthesis next to the data point represents the average RB,E . As XB,f represents the measured purity of S-ibuprofen in the final product, the experimental recovery ratio of S-ibuprofen is defined as

RB,E =

Wf XB,f W0 XB,0

(14)

For no mixing without seeding, no crystallization or vaporization is observed for various cooling rates ranging from 0.10 ◦ C/min to 0.50 ◦ C/min when SC is operated to the threephase equilibrium condition for each feed. This is attributed to the subcooling of the liquid mixture during the SC operation. Ultrasound mixing or magnetic stirring only slightly reduces subcooling. However, once crystallization is induced in the subcooled liquid mixture, crystals are formed abruptly. Due to the structure similarity between S-ibuprofen and Ribuprofen, R-ibuprofen might be more easily incorporated into the S-ibuprofen crystal lattice when S-ibuprofen crystals are formed abruptly. As shown in Fig. 9, it yields XB,f ∼ = XB,0 with RB,E ∼ = 1. Therefore, the ibuprofen mixture can not be further purified by SC without seeding. Subcooling can be eliminated by seeding (0.01 g seeds with size 15–60 ␮m) with ultrasound mixing or magnetic stirring during the SC operation. Subsequently, crystals are gradually formed preferentially on the seeds, leading to a product with a higher purity. Fig. 9 shows that seeding with ultrasound mixing or magnetic stirring is effective to further purify the ibuprofen mixture. As compared in Figs. 8 and 9, for seeding with ultrasound mixing or magnetic stirring, the resulting XB,f and RB,E is close to XB,S and RB,S , respectively. Thus, the experimental results are consistent with the simulation results predicted by the model. As magnetic stirring helps to uniformly disperse the seeds, seeding with magnetic stirring leads to a product with a higher purity compared to seeding with ultrasound mixing. For example, Fig. 9 shows that 1 g liquid feed at XB,0 = 0.93 can be purified to XB,f = 0.954 with RB,E = 74% for seeding with magnetic stirring, as compared to XB,f = 0.950 with RB,E = 77% for seeding with ultrasound mixing. For seeding without magnetic stirring or ultrasound mixing, subcooling of the liquid mixture is still observed. Subsequently, once crystallization is induced in the subcooled liquid mixture, crystals are formed abruptly, leading to a product with XB,f ∼ = XB,0 and RB,E ∼ = 1 as shown in Fig. 9.

4.

Conclusions

In application of SC, S-ibuprofen can not be purified from the ibuprofen enantiomers without seeding. On the other hand, seeding with ultrasound mixing or magnetic stirring can be employed to further purify the ibuprofen enantiomers. A thermodynamic model is developed to simulate the three-phase equilibrium conditions during the SC operation. The experimental results with seeding, including the final enantiomeric purity and recovery ratio of S-ibuprofen, are consistent with the simulation results predicted by the model. Thus, the developed model is adequate to describe the heat and mass transfer phenomena in the SC process. Compared to conventional crystallization, the unique feature of SC is that impurity is vaporized and no mother liquor is present at the end of SC.

307

Subsequently, filtration and crystal washing is not required for final obtained crystals.

Acknowledgments The financial support of this work by Chang Gung Memorial Hospital (CMRPD2B0141, CMRPD2B0142, CMRPD2B0143) and Ministry of Science and Technology of Taiwan (MOST103-2221E-182-067-MY3) is greatly appreciated.

References Adams, S.S., Bresloff, P., Mason, C.G., 1976. Pharmacological differences between the optical isomers of ibuprofen: evidence for metabolic inversion of the (−) isomer. J. Pharm. Pharmacol. 28, 256–257. Beierling, T., Micovic, J., Lutze, P., Sadowski, G., 2014. Using complex layer melt crystallization models for the optimization of hybrid distillation/melt crystallization processes. Chem. Eng. Process.: Process Intensif. 85, 10–23. Cauwenberg, V., Vergossen, P., Stankiewicz, A., Kierkels, H., 1999. Integration of reaction and separation in manufacturing of pharmaceuticals: membrane-mediated production of S-Ibuprofen. Chem. Eng. Sci. 54, 1473–1477. Dwivedi, S.K., Sattari, S., Jamali, F., Mitchell, A.G., 1992. Ibuprofen racemate and enantiomers: phase diagram, solubility and thermodynamic studies. Int. J. Pharm. 87, 95–104. Huh, Y.S., Jun, Y.S., Hong, Y.K., Kim, W.H., 2006. Microfluidic separation of (S)-ibuprofen using enzymatic reaction. J. Mol. Catal. B: Enzym. 3, 96–101. Jacques, J., Collet, A., Wilen, S.H., 1981. Enantiomers, Racemates, and Resolutions. John Wiley & Sons, New York. Jiang, X.B., Hou, B.H., He, G.H., Wang, J.K., 2012. Falling film melt crystallization (I). Model development, experimental validation of crystal layer growth and impurity distribution process. Chem. Eng. Sci. 84, 120–133. Kim, Y.S., Hyun, C.J., Kim, T.A., Gu, M.B., 2010. Isolation and characterization of enantioselective DNA aptamers for ibuprofen. Bioorg. Med. Chem. 18, 3467–3473. Lerdkanchanaporn, S., Dollimore, D., 1997. A thermal analysis study of ibuprofen. J. Therm. Anal. 49, 879–886. Long, W.S., Kamaruddin, A., Bhatia, S., 2005. Chiral resolution of racemic ibuprofen ester in an enzymatic membrane reactor. J. Membr. Sci. 247, 185–200. Micovic, J., Beierling, T., Lutze, P., Sadowski, G., Górak, A., 2013. Design of hybrid distillation/melt crystallisation processes for separation of close boiling mixtures. Chem. Eng. Process.: Process Intensif. 67, 16–24. Park, J.S., Kim, W.S., Kim, J.M., Kim, I.H., 2008. Ibuprofen racemate separation by simulated moving bed. J. Chem. Eng. Jpn. 41, 624–626. Peper, S., Lubbert, M., Johannsen, M., Brunner, G., 2002. Separation of ibuprofen enantiomers by supercritical fluid simulated moving bed chromatography. Sep. Sci. Technol. 37, 2545–2566. Sandler, S.I., 2006. Chemical, Biochemical, and Engineering Thermodynamics. John Wiley & Sons, Asia. Sheldon, R.A., 1993. Chiritechnology. Marcell Dekker, New York. Shiau, L.D., Liu, K.F., 2013. Investigations into the effects of the cooling rate on stripping crystallization. Ind. Eng. Chem. Res. 52, 1716–1722. Shiau, L.D., Yu, C.C., 2009. Separation of the benzene/cyclohexane mixture by stripping crystallization. Sep. Purif. Technol. 66, 422–464. Shiau, L.D., Wen, C.C., Lin, B.S., 2005. Separation and purification of p-xylene from the mixture of m-xylene and p-xylene by distillative freezing. Ind. Eng. Chem. Res. 44, 2258–2265. Shiau, L.D., Wen, C.C., Lin, B.S., 2006. Application of distillative freezing in the separation of o-xylene and p-xylene. AIChE J. 52, 1962–1967.

308

chemical engineering research and design 1 1 7 ( 2 0 1 7 ) 301–308

Shiau, L.D., Wen, C.C., Lin, B.S., 2008. Separation of p-xylene from the multicomponent xylene system by stripping crystallization. AIChE J. 54, 337–342. Smith, J.M., Van Ness, H.C., Abbott, M.M., 2001. Introduction to Chemical Engineering Thermodynamics. McGraw-Hill, Singapore. Trung, T.Q., Kim, J.M., Kim, K.H., 2006. Preparative method of R-(−)-ibuprofen by diastereomer crystallization. Arch. Pharm. Res. 29, 108–111.

Ulrich, J., 2003. Melt Crystallization: Fundamentals, Equipment and Applications. Shaker, Aachen. Wang, Y., Hu, Y., Xu, J., Luo, G., Dai, Y., 2007. Immobilization of lipase with a special microstructure in composite hydrophilic CA/hydrophobic PTFE membrane for the chiral separation of racemic ibuprofen. J. Membr. Sci. 293, 133–141. Zey, E.G., Shockley, T.H., Ryan, D.A., Moss, G.L., 1992. Method for purification of ibuprofen comprising mixtures. U.S. Patent No. 5 151 551.