Activated Composite Membranes Containing the Chiral Carrier

Activated Composite Membranes Containing the Chiral Carrier...
0 downloads 0 Views 390KB Size
7696

Ind. Eng. Chem. Res. 2005, 44, 7696-7700

Activated Composite Membranes Containing the Chiral Carrier N-hexadecyl-L-hydroxyproline. Description of Morphology and Performance Ta` nia Gumı´,† Carles Torras,‡ Ricard Garcia-Valls,‡ and Cristina Palet*,† Centre Grup de Te` cniques de Separacio´ en Quı´mica, Unitat de Quı´mica Analı´tica, Departament de Quı´mica, Universitat Auto` noma de Barcelona, 08193-Bellaterra, Catalunya, Spain, and Departament d’Enginyeria Quı´mica, ETSEQ, Universitat Rovira i Virgili, Av. Paı¨sos Catalans 26, 43007-Tarragona, Catalunya, Spain

Activated composite membranes (ACMs) containing the chiral carrier N-hexadecyl-L-hydroxyproline have been characterized by scanning electron microscopy (SEM). SEM membrane images have been further treated with the Intrepretacio´ de Fotografies de Microsco`pia Electro`nica (IFME) program, and certain morphological properties of the ACM (such as mean internal pore size, internal irregularity, and internal asymmetry) were determined. In addition, ACMs have been tested for propranolol filtration, using a dead-end filtration membrane module working under 3.5 bar, and both the permeation rate and the enantioselectivity have been evaluated. Propranolol permeation through the three ACMs has been detected here; however, unfortunately, a clear enantioselective behavior has not been observed. The morphological structure of the ACM has been studied and related with their obtention procedure. Resulting data were finally compared with the results from other characterization techniques applied previously for such membranes. Introduction Membrane separation systems have been gaining importance recently as enatioresolution techniques, because they offer several advantages over traditional methods, such as low time cost, set-up simplicity, and the possibility to be used in continuous mode. In that sense, various membrane configurations have been already proposed for the separation of a broad number of chiral species, including amino acids and their derivatives, as well as drugs. Different liquid membrane types, based either on chiral liquids or on solutions of chiral molecules (namely, chiral selectors or carriers), were first proposed to resolve racemic mixtures.1 Crown ethers, polyamino acids, or cyclodextrines are some of the most widely used chiral molecules in membrane solutions.2 However, liquid membranes processes, when applied for the separation of enantiomers, experience a rapid decrease of selectivity, because of the free diffusion of analytes across the membranes, and, in addition, they show low stability and short lifetime when tested under industrial separation conditions.3 Therefore, solid polymeric membranes that are based primarily on ultrafiltration membranes4,5 or polymer impringting membranes6 have been developed. For the case of SR-propranolol, which is a β-blocking drug that is used to treat certain cardiovascular anomalies (with the S-enantiomers showing far more blocking activity than the R-enantiomer),7,8 different enantioselective carriers have been studied. On one hand, N-nalkyl-hydroxyprolines,9 and dialkyl tartrates10 were * Corresponding author. Tel: +34935813475. Fax: +34935812379. E-mail: [email protected]. † Centre Grup de Te`cniques de Separacio´ en Quı´mica, Unitat de Quı´mica Analı´tica, Departament de Quı´mica, Universitat Auto`noma de Barcelona. ‡ Departament d’Enginyeria Quı´mica, ETSEQ, Universitat Rovira i Virgili.

used in liquid membrane configurations. Also, the N-nalkyl-hydroxyprolines have been used as chiral carriers in polymeric membranes that are based on either polysulfone or chiral-derivatized polysulfone.11 Norbornadiene also has been considered to be a chiral polymer.12 Certain activated composite membranes (ACMs) containing different carriers, including different types of chiral carriers, have been prepared previously,13 and some of them were applied for the enantioseparation of racemic propranolol under dialysis conditions. However, propranolol transport across ACMs was not encountered in those cases. It was believed that such behavior was probably a consequence of two different factors: a deficient or null incorporation of the carrier in the ACMs and/or the formation of a dense top polyamide (PA) layer, hindering species transport across the ACM. Therefore, an in-depth characterization of the membranes was performed. The data collected gave evidence of an appropriate incorporation of the chiral carriers into the ACM,13 suggesting that the absence of propranolol transport could be related to the dense membrane properties. In this work, ACMs that have been developed previously,14 and successfully applied for the separation of metal ions15 or amino acids,16 have been considered for the enantioresolution of SR-propranolol. Therefore, ACMs containing the chiral carrier N-hexadecyl-Lhydroxyproline have been obtained, properly characterized by scanning electron microscopy (SEM) and the Intrepretacio´ de Fotografies de Microsco`pia Electro`nica (IFME) program, and finally tested under ultrafiltration conditions. Experimental Section Reagents. R-propranolol hydrochloride, S-propranolol hydrochloride, and racemic propranolol hydrochloride, all of analytical-reagent (AR) grade, were supplied

10.1021/ie0580295 CCC: $30.25 © 2005 American Chemical Society Published on Web 08/20/2005

Ind. Eng. Chem. Res., Vol. 44, No. 20, 2005 7697

by Sigma-Aldrich (Germany). N-hexadecyl-L-hydroxyproline (HHP), isopropyl myristate (IPM), triethanolamine, and hydroxypropyl-β-cyclodextrin (HP-β-CD), all of AR grade, were also purchased from Sigma-Aldrich (Germany). All other reagents used (such as acids and inorganic salts) were of analytical grade. Milli-Q water (from Millipore, Bedford, MA) was used for all aqueous solutions. Membranes. Polysulfone (PSU) (from Basf, Spain) was dissolved in AR-grade N,N-dimethylformamide (DMF, Sigma-Aldrich) to a concentration of 15 wt %. The PSU solution was cast on the surface of a nonwoven fabric and, all together, immersed into a coagulation bath at ca. 4 °C, to precipitate a porous PSU layer. Afterward, a thin polyamide (PA) layer was formed on the top surface of the PSU membrane, by in situ interfacial polymerization of 1,3-phenylenediamine (Merck, Germany) dissolved in water, with trimesoyl chloride dissolved in hexane (both from Sigma-Aldrich, Germany). The membranes were chemically activated by the addition of the chiral carrier, either in the PSU casting solution or into the hexane solution used for the preparation of the top PA layer. The carrier HHP was first dissolved in IPM, prior to being added to these solutions. More details of the membrane preparation are given elsewhere.17 A blank membrane, without carrier, was prepared and tested for comparison. The membranes are labeled hereafter as ACM1 (blank membrane), ACM2 (the chiral carrier is introduced during the formation of the top PA layer), and ACM3 (the chiral carrier is introduced in the PSU casting solution). Apparatus and Procedure. The filtration experiments were performed using a dead-end filtration membrane module, operating in batch mode, over a membrane area of 11.3 cm2 at a constant pressure of 3.5 bar,13 provided by a corresponding N2 flux. All experiments were performed at 24 ( 1 °C and at least twice. The membrane module was filled with different volumes of feed solution (always containing 0.1 g/L of racemic propranolol and adjusted at pH 8 with a Borax buffer) and conveniently pressured. The feed solution was stirred during the entire experiment with an electromagnetic stirrer, to avoid the formation of a diffusion layer that would be attached to the membrane.17 Experiments started when the permeation was initiated. The permeation of each enantiomer, during the filtration process, was determined by monitoring their concentration in the permeate. For this purpose, samples were periodically withdrawn over the entire experiment. Membranes were characterized by SEM, to study their superficial and internal morphology. ACM SEM images were obtained using an SEM microscope (Hitachi model S-570, Hitachi Ltd. Tokyo, Japan) in the UAB Microscopy Service. Furthermore, SEM images were subjected to the IFME program, which was created by Torras and Garcia-Valls, to determine the mean internal pore size, as well as the internal membrane asymmetry and irregularity.18 SR-Propranolol Determination. A capillary electrophoresis (CE) system (P/ACE MDQ System, Beckman Coulter, Fullerton, CA) was used to analyze the concentration of both enantiomers in the collected samples. Determination was performed in uncoated fused-silica capillaries with an inner diameter (ID) of 50 µm and a length of 60 cm (50 cm to the detector). Before each set of analyses, the capillary was rinsed

with a 0.1 M NaOH solution, Milli-Q water, and finally the separation buffer solution. The latter consisted of 100 mM phosphoric acid adjusted at pH 4.4 with triethanolamine, which contained 17.4 mM hydroxypropyl-β-cyclodextrin (HP-β-CD).19,20 The applied voltage was 23 kV, and UV detection was conducted at a wavelength of 210 nm. Samples were injected using the hydrodynamic mode for 5 s at 0.3 psi. The capillary was thermostated at 20 °C. Between consecutive determinations, the capillary was rinsed with Milli-Q water. At the end of the day, the capillary was washed with 0.1 M NaOH, Milli-Q water, and methanol (MeOH), which was used to remove organic material and water. Calculations Both the permeation rate of racemic propranolol through the chiral ACMs and its enatioselectivity were investigated. The permeation rate is expressed in terms of permeation percentage (P), which is calculated as the ratio of S- or R-propranolol concentration in the permeate at any time t (Cpt,E-) to the initial S- or Rpropranolol concentration in the feed phase (Cf0,E-):

PE- (%) )

( )

Cpt,E× 100 Cf0,E-

(1)

The enantioselectivity of the process is given in terms of alpha (R) values. Such R values were calculated by the following equation:11

RS-/R- )

(Cpt,S-/Cpt,R-) (Cf0,S-/Cf0,R-)

(2)

where Cpt,S- and Cpt,R- are the concentration, in the permeate, of S- and R-enantiomers of propranolol, respectively, at any time. Cf0,S- and Cf0,R- correspond to the initial feed concentration of the S- and Renantiomers, respectively. Results and Discussion ACMs that contain the chiral carrier HHP have been prepared here and tested for propranolol filtration purposes, working under pressure conditions. Characterization of ACM by SEM and IFME. The three obtained ACMs were characterized using SEM, to investigate their surface and cross-sectional morphological structure. Figure 1 shows the SEM images for the three ACMs. As may be observed, all of them, as expected, present the typical morphological structure of ultrafiltration membranes that have been prepared via a phase inversion technique with DMF/water as the solvent/nonsolvent pair, i.e.; asymmetric morphology with the presence of macrovoids,17 with an additional noteworthy dense top layer. However, in the case of ACM3, the asymmetric structure is not so evident and hardly any macrovoids may be observed. Such morphological features can be directly related with the composition of the PSU casting solution, which, in this case, contained the carrier HHP dissolved in IPM. The latest, IPM, acts as an extra nonsolvent, approaching the differences in polarity between DMF and water. Thus, it causes a decrease of the PSU precipitation rate in the coagulation bath, leading to a more-symmetric structure with a smaller number of macrovoids. From the SEM images, it can be also noted that the three ACMs

7698

Ind. Eng. Chem. Res., Vol. 44, No. 20, 2005

Figure 1. Cross-sectional images of (a) ACM1, (b) ACM2, and (c) ACM3 obtained by scanning electron microscopy (SEM).

Figure 2. Intrepretacio´ de Fotografies de Microsco`pia Electro`nica (IFME) results from (a) ACM1, (b) ACM2, and (c) ACM3 images treatment. All cases show, from left to right, the transformed photograph of cross-sectional SEM image of a piece of membrane (reproduced membrane surface is indicated by the x- and y-axes), and the variations of internal pore distribution along the vertical (y-axes) and horizontal (x-axes) length of the images. In the two graphics presented, of each membrane, pore distribution is plotted versus the piece of membrane analyzed.

present a regular surface, which confirms the existence of the top dense polyamide layer. Figure 2 and Table 1 show IFME results from the SEM images of Figure 1.18 Both the internal asymmetry and the irregularity of membranes (which are defined as asymmetry along the y- or x-axis, respectively) were evaluated from a piece of the corresponding membrane cross-sectional SEM images. As may be seen in the figure, ACM1 (Figure 2a) and ACM2 (Figure 2b) exhibit

a huge variation of the internal pore distribution, especially along y-axes. This is a direct consequence of the presence of macrovoids. Therefore, relatively high asymmetry and irregularity values (collected in Table 1) were encountered. As may be seen, ACM2 exhibits lower asymmetry than ACM1, because of the location of the macrovoids inside the membrane structure. On the other hand, in the case of ACM3 (Figure 2c), the internal pore distribution variation encountered was

Ind. Eng. Chem. Res., Vol. 44, No. 20, 2005 7699

Figure 3. Initial and steady-state permeation flux of filtration experiments performed with the membranes under study. Error bars correspond to the standard deviation of the values. Table 1. Numerical Values of Mean Internal Pore Size, Asymmetry, and Irregularity of Membranes, as Calculated by the Intrepretacio´ de Fotografies de Micrsco` pia Electro` nica (IFME) Program

ACM1 ACM2 ACM3

mean internal pore size (µm)

asymmetry percentage

global irregularity

2.705 (0.238) 2.764 (0.303) 2.798 (0.383)

46 20 14

0.0021 0.0033 0.0004

smoother and more linear. Therefore, the lowest asymmetry and irregularity values are found for this membrane (see Table 1). As expected, ACM1 and ACM2 have similar irregularity values, because of the same PSU layer preparation step (not containing the chiral carrier HHP in IPM). The PSU layer occupies the largest portion of the membrane cross-section image. The top dense PA layers, obtained by interfacial polymerization, usually have a thickness of ca. 1 µm.13 Filtration of Propranolol through ACM. The propranolol permeation throughout this study was determined using a dead-end filtration membrane module at 3.5 bar, as mentioned previously. Figure 3 shows the initial and the steady-state permeation flux values of the three ACMs. The first thing to note is that the permeation flux values obtained here, for all the membranes, are much lower than permeation flux values commonly defined for filtration experiments performed at 3.5 bar, which are in the range of 170-1700 L m-2 h-1.17 This observation is undoubtedly due to the dense PA top layer of the ACM, which considerably increases the membrane resistance to the feed permeation. The working pressure used in this study corresponds to that of ultrafiltration experiments, which are performed, as a rule, with porous membranes, instead of composited membranes. The latter membranes generally operate under pressures of 10-60 bar. Nevertheless, the aim of this work was not the achievement of a rapid and efficient permeation flux (as is the case for other chemical separation systems), but rather a selective permeation, which usually occurs at low permeation rates.11 For this reason, dense membranes and lower working pressures were combined. Taking into consideration the huge range of permeation flux (PF) usually defined at 3.5 bar, it seems that significant differences are not encountered among the ACMs. Even so, if considering the slight differences within them, it is observed that the permeation flux follows the order ACM2 > ACM3 . ACM1. From this

Figure 4. Evolution of alpha (R) values versus time, for the different ACM composite membranes. Error bars correspond to the standard deviation of the values. Table 2. Mean Value of Permeation Percentages of Filtration Experiments Achieved for Both Propranolol Enantiomers PS

ACM1 ACM2 ACM3

PR

mean permeation percentage

error (%)

mean permeation percentage

error (%)

93.6 117.3 112.5

10.8 3 2.6

97.6 119.1 111.1

12.3 3.4 1.7

order, it can be observed that the incorporation of the carrier HHP in the membrane (in the case of ACM2 and ACM3) facilitates the pathway for propranolol to migrate across the ACM. In addition, the highest PF value is reached when the carrier is added into the PA top layer (ACM2), because this is directly in contact with the feed solution. It is also important to note that, in all cases, the PF suffers from an initial decrease during the first hours of the experiment. This permeation flux decrease is inherent to filtration experiments, and it is usually related to the concentration polarization on the active membrane side.17 The mean permeation percentage of both enantiomers of propranolol, for each membrane type, are collected in Table 2. S- and R-propranolol permeation percentages were constant, along with the time of filtration experiments, and the permeation percentage values corresponded to the initial propranolol concentration in the feed solution. Therefore, these observations mean that propranolol is not retained in this membrane system. Such SR-propranolol permeation occurs, most probably, across the nanopores of the ACMs, which are usually present in PA dense layers.17 Figure 4 shows the R values calculated for the different experiments. Unfortunately, as can be observed, a clear enantioselective tendency was not detected, although R values of >1 were observed in certain cases. This phenomenon may be due to the lack of retention of propranolol in the membrane, under the studied conditions. The pressure applied in this system, as a driving force, even being low in comparison with that of common systems (as indicated), is too high to permit enantioselectivity for the system conditions investigated in this work.11,21 Similar membrane systems, working under chemical and physical conditions that lead to lower propranolol

7700

Ind. Eng. Chem. Res., Vol. 44, No. 20, 2005

transport/permeation rates, and serving from the same chiral carrier HHP, allowed certain enantioselectivity to be attained.11 Concerning morphological parameters determined by SEM and IFME, they do not show a direct relationship with membrane filtration properties. It may be due to the fact that morphological parameters are mainly determined for the PSU layer, whereas, most probably, the membrane filtration skills are governed by the top dense PA layer. Conclusions The three obtained activated composite membranes (ACM1, ACM2, and ACM3) were characterized by scanning electron microscopy (SEM) and the Intrepretacio´ de Fotografies de Micrsco`pia Electro`nica (IFME) program. Differences in morphological properties have been encountered between ACM3 and the other ACMs (ACM1 and ACM2). Whereas the former presents a very symmetric and regular internal structure, with a top dense layer, ACM1 and ACM2 show an asymmetric internal structure with macrovoids and a top dense layer. The composition of the polysulfone (PSU) layer casting solution (with a carrier, in the case of ACM3, and without it, for ACM1 and ACM2) is a determining factor for the internal morphological parameters, as expected. The permeation of propranolol throughout the three ACMs was attained. Even so, clear enantioselective permeation has not been detected in any case. Therefore, ACMs, under the working conditions assayed in this work, do not seem to be appropriate for use in enantioresolution membrane systems. Acknowledgment This work has been supported by CICYT (Ref. Nos. PPQ2002-04267-C03-01 and PPQ2002-024201-C02-01). T.G. acknowledges el Ministerio de Educacio´n, Cultura y Deporte, for the predoctoral fellowship. Literature Cited (1) Keurentjes, J. T. F.; Voermans, F. J. M. Membrane separations in the production of optically pure compounds. In Chirality in Industry II. Developments in the Commercial Manufacture and Applications of Optically Active Compounds; Collins, A. N., Sheldrake, G. N., Crosby, J., Eds.; Wiley: Chichester, U.K., 1997. (2) Brice, L. J.; Pirkle, W. H. Enantioselective Transport Through Liquid Membranes. In Chiral Separations. Applications and Technology; Ahuja, S., Ed.; American Chemical Society: Washington, 1997; p 309. (3) Kemperman, A. J. B.; Bargeman, D.; Van den Boomgaard, Th.; Strathmann, H. Stability of Supported Liquid Membranes: State of the Art. Sep. Sci. Technol. 1996, 31, 2733. (4) Masawaki, T.; Sasai, M.; Tone, S. Optical resolution of an amino acid by enantioselective ultrafiltration membrane. J. Chem. Eng. Jpn. 1992, 25, 33. (5) Higuchi, A.; Yomogita, H.; Yoon, B. O.; Kojima, T.; Hara, M.; Maniwa, S.; Sayito, M. Optical resolution of amino acids by ultrafiltration using recognition sites of DNA. J. Membr. Sci. 2002, 205, 203.

(6) Yoshikawa, M.; Yonetani, K. Molecularly imprinted polymeric membranes with oligopeptide tweezers for optical resolution. Desalination 2002, 149, 287. (7) Ahuja, S. Chiral Separation and Technology. An overview. In Chiral Separations. Applications and Technology; Ahuja, S., Ed.; American Chemical Society: Washington, 1997; p 1. (8) Coelhoso, I. M.; Cardoso, M. M.; Viegas, R. M. C.; Crespo, J. G. Modelling of Transport Mechanisms in Liquid Membranes. In Proceedings of Engineering with Membranes; Luque, S., Alvarez, J. R., Eds.; Universidad de Oviedo: Oviedo, Spain, 2001; p 425. (9) Gumı´, T.; Valiente, M.; Palet, C. Characterization of a Supported Liquid Membrane Based System for the Enantioseparation of SR-Propranolol by N-Hexadecyl-L-hydroxyproline. Sep. Sci. Technol. 2004, 39, 431. (10) Keurentjes, J. T. F.; Nabuurs, L. J. W. M. Vegter, E. A. Liquid membrane technology for the separation of racemic mixtures. J. Membr. Sci. 1996, 113, 351. (11) Gumı´, T. Membranes en la separacio´ enantiome`rica del fa`rmac propranolol. Desenvolupament i caracteritzacio´, Doctoral Thesis, Universitat Auto`noma de Barcelona, Barcelona, Spain, 2004. (12) Aoki, T.; Ohshima, M.; Shinohara, K.; Kaneko, T.; Oikawa, E. Enantioselective permeation of racemates through a solid (+)poly{2-[dimethyl(10-pinalyl)silyl]norbornadiene}membrane. Polymer 1997, 38, 235. (13) Gumı´, T.; Valiente, M.; Khulbe, K. C.; Palet, C.; Matsuura, T. Characterization of activated composite membranes by solute transport, contact angle measurement, AFM and ESR. J. Membr. Sci. 2003, 212, 123. (14) Oleinikova, M.; Garcia-Valls, R.; Valiente, M.; Mun˜oz, M. Procedimiento para la obtencio´n de membranas compuestas para el transporte de especies quı´micas. Spanish Patent No. P200000536. (15) Gumı´, T.; Oleinikova, M.; Palet, C.; Valiente, M.; Mun˜oz, M. Facilitated transport of lead(II) and cadmium(II) through novel activated composite membrane containing di-(2-ethyl-hexyl)phosphoric acid as carrier. Anal. Chim. Acta 2000, 408, 65. (16) Calzado, J. A.; Palet, C.; Valiente, M. Facilitated transport and separation of aromatic amino acids through activated composite membranes. Anal. Chim. Acta 2001, 431, 59. (17) M. Mulder, Basic Principles of Membrane Technology, Second Edition; Kluwer Academic Publishers: Dordrecht, The Netherlands, 2000. (18) Torras, C.; Garcia-Valls, R. Quantification of membrane morphology by interpretation of scanning electron microscopy images. J. Membr. Sci. 2004, 233, 119. (19) Pak, C.; Marriot, P. J.; Carpenter, P. D.; Amiet, R. G. Enantiomeric separation of propranolol and selected metabolites by using capillary electrophoresis with hydroxypropyl-β-cyclodextrine. J. Chromatogr., A 1998, 793, 357. (20) Fillet, M.; Bechet, I.; Chiap, P.; Hubert, Ph.; Crommen, J. Enantiomeric purity determination of propranolol by cyclodextrinmodified capillary electrophoresis. J. Chromatogr., A 1995, 717, 203. (21) Lee, S. B.; Mitchell, D. T.; Trofin, L.; Nevanen, T. K.; So¨derlund, H.; Martin, C. R. Antibody-Based Bio-Nanotube Membranes for Enantiomeric Drug Separations. Science 2002, 296, 2198.

Received for review March 31, 2005 Revised manuscript received June 28, 2005 Accepted July 8, 2005 IE0580295