Wen-Yuan Chuang

Properties of the poly(vinyl alcohol)/chitosan blend and its effect on the culture of fibroblast in vitro

In this work, the properties of poly(vinyl alcohol) (PVA) and PVA/chitosan blended membranes were investigated by scanning electron microscopy (SEM), differential scanning calorimetry (DSC) and electron spectroscopy for chemical analysis (ESCA). The SEM photographs show the PVA/chitosan blended membrane undergoes dramatic changes on the surface and bulk structure during the membrane formation. The DSC analysis shows that PVA and chitosan are not very compatible in the PVA/chitosan blended membrane, whereas the combination of two polymer chains of constitutionally different features is revealed. In addition, the surface of the PVA/chitosan blended membrane is enriched with nitrogen atoms at the ESCA analysis. These reflect the PVA membrane can be modified by blending with chitosan that in turn may affect the biocompatibility of the blended membrane. Therefore, adhesion and growth of fibroblasts on the PVA as well as PVA/chitosan blended membranes were investigated. Cell morphologies on the membranes were examined by SEM and cell viability was studied using MTT assay. It was observed that the PVA/chitosan blended membrane was more favorable for the cell culture than the pure PVA membrane. Cells cultured on the PVA/chitosan blended membrane had good spreading, cytoplasm webbing and flattening and were more compacting than on the pure PVA membrane. Consequently, the PVA/chitosan blended membrane may spatially mediate cellular response that can promote cell attachment and growth, indicating the PVA/chitosan blended membrane should be useful as a biomaterial for cell culture.

Mechanisms of PVDF membrane formation by immersion-precipitation in soft (1-octanol) and harsh (water) nonsolvents

Abstract Crystalline PVDF was precipitated, respectively, from water/DMF and 1-octanol/DMF solutions to produce membranes with asymmetric and uniform morphologies. The formation mechanisms of these specific structures were described both in the aspects of thermodynamics (equilibrium phase behavior) and the kinetics (diffusion trajectory). The phase diagrams of the investigated systems indicated the possibility of liquid–liquid demixing or crystallization or both during the immersion-precipitation process. The sequence of these phase separation events, which determined the ultimate membrane structure, was attributed to the kinetic factors. Into this context, a quantitative model describing the immersion-precipitation process was considered. The calculated diffusion trajectory and concentration distribution in the nascent membrane were found to be consistent with the experimental membrane morphologies. Moreover, the precipitation rate of the PVDF solution in water and 1-octanol were examined by the light transmission experiments. The latter results further confirmed the validity of the mass transfer calculations.

The effect of polymeric additives on the structure and permeability of poly(vinyl alcohol) asymmetric membranes

Abstract The manufacture and dextran-rejection properties of asymmetric poly(vinyl alcohol) (PVA) membranes have been described in this study. Membranes were prepared from a casting solution of PVA, water as solvent, and water-soluble polymeric additive by immersing them in Na 2 SO 4 /KOH/H 2 O as coagulant medium. Experiments showed that the dextran and poly(vinyl pyrrolidone) (PVP) additives exerted a different influence on the structure and permeability of membranes. Especially, the structure of skin layer strongly depended on the polymeric additives in the casting solution. The addition of dextran additives in the system could induce pores in the top layer. Conversely, the PVP additives effectively blocked the interstitial cavities within the top layer to generate a more compact structure. A mechanism describing that the affinity between additive and casting solution as well as between additive and coagulant medium was proposed to investigate the effect of dextran and PVP additives in the formation of PVA membranes. The results presented here offer a better understanding of relationships between the membrane formation mechanism and the skin structure when designing an asymmetric membrane by the addition of polymeric additives in the casting solution.

Morphology of crystalline Nylon-610 membranes prepared by the immersion-precipitation process : competition between crystallization and liquid-liquid phase separation

Abstract Nylon-610 membranes were prepared at 25°C by direct immersion of various dope solutions into either formic acid/water or 1-octanol bath. Depending on the dope and bath conditions, the precipitated membranes demonstrated characteristics derived from crystallization and/or liquid–liquid phase separation during the precipitation process. As a good dope solution was immersed in a harsh bath, e.g., water, precipitation occurred initiated by liquid–liquid phase separation. The formed membrane exhibited a cellular structure similar to that commonly observed in amorphous membranes. Alternatively, when a metastable dope (with respect to crystallization) was immersed in a soft bath containing a substantial amount of formic acid, crystallization dictated the precipitation process to yield bi-continuous, particulate membranes. Membranes with extensive macrovoids were observed, in the event that the dope contained a large amount of solvent. In the latter case, precipitation took place immediately after immersion, consistent with Strathmann and Smolder’s results for several membrane forming systems. In addition, skinless microporous membranes were prepared by precipitation of the dope solutions in a 1-octanol bath, in which precipitation occurred slowly and the formed membrane was composed of “sheaf-like” crystallites that interlocked into a homogeneous bi-continuous network .

Effect of evaporation temperature on the formation of particulate membranes from crystalline polymers by dry-cast process

The membrane formation of crystalline poly(ethylene-co-vinyl alcohol) (EVAL), poly(vinylidene fluoride) (PVDF), and polyamide (Nylon-66) membranes prepared by dry-cast process was studied. Membrane morphologies from crystalline polymers were found to be strongly dependent on the evaporation temperature. At low temperatures, all the casting solution evaporated into a particulate morphology that was governed by the polymer crystallization mechanism. The rise in the evaporation temperature changed EVAL membrane structure from a particulate to a dense morphology. However, as the temperature increased PVDF and Nylon-66 membranes still exhibited particulate morphologies. The membrane structures obtained were discussed in terms of the characteristics of polymer crystallization in the casting solution theoretically. At elevated temperatures the crystallization was restricted for the EVAL membrane because the increase rate in the polymer concentration was fast relative to the time necessary for growth of nuclei. Nonetheless, the time available for PVDF and Nylon-66 with stronger crystalline properties was large enough to form the crystallization-controlled particulate structure that differed in particle size only. In addition, particles in the PVDF membrane were driven together to disappear the boundary, but those in the Nylon-66 membrane exhibited features of linear grain boundary. The difference in particle morphology was attributed to the Nylon-66 with the most strongly crystalline property. Therefore, the kinetic difference in the crystallization rate of the polymer solution play an important role in dominating the membrane structure by dry-cast process.

The formation mechanism of membranes prepared from the nonsolvent-solvent-crystalline polymer systems

Abstract Two different types of membranes were prepared by immersion–precipitation process in two nonsolvent–DMSO–poly(ethylene-co-vinyl alcohol) (EVAL) systems. The effect of nonsolvent on the resulting membrane structure was studied via scanning electron microscopy. On using 2-propanol as the nonsolvent, the membrane showed a homogeneous particulate morphology. If the nonsolvent was changed to water, the membrane structure was a typically asymmetric structure while the particulate morphology was suppressed. In order to understand the change of the obtained membrane structures, the phase diagrams of water–DMSO–EVAL and 2-propanol–DMSO–EVAL systems were contrasted at 25°C. Both crystallization-induced gelation and liquid–liquid demixing were observed and equilibrium crystallization lines are always positioned above the binodal boundaries. Moreover, the precipitation rate of the EVAL solution in 2-propanol and water were examined by the light transmission experiments and the local composition profiles of membrane solution during membrane formation were analyzed by using a ternary mass transfer model. Based on both thermodynamic behaviors and kinetic properties, the membrane structures obtained were discussed in terms of the sequence and mechanism of phase transformations during membrane formation.

Swelling behavior of hydrophobic polymers in water/ethanol mixtures

Swelling behavior of poly(ethylene-co-vinyl alcohol) (EVAL), polyurethane (PU) and poly(ethylene-co-vinyl acetate) (EVAc) in ethanol/ water mixtures was investigated. In the case of EVAL, the total swelling curve rises to a maximum that is above the swelling of either pure constituent. On the other hand, in the PU and EVAc systems, the total swelling is intermediate between the swelling of the two pure components. In the case of PU, the total swelling is close to the ideal system as the water-rich region is approached. However, in the case of EVAc, the deviation from ideal is significantly negative. Theoretical swelling values have been derived from Flory‐Huggins thermodynamics by using the equilibrium sorption data of the pure component. The theoretical results show good agreement with the swelling behavior of PU and EVAc in ethanol/water mixtures. However, the agreement between theoretical and experimental swelling values for EVAL is reasonably established when the Flory‐Huggins equation was modified by incorporating a ternary interaction parameterx T. This is probably due to the fact that the water/ethanol mixture produces the water‐ethanol complex that has a greater affinity for EVAL. In addition, the experimental results indicate that the shape of hydrophobic polymer swelling curve in the water/ethanol mixture is mainly controlled by the interactions between water and the polymer. Therefore, the correlation among the swelling equilibrium data, the polymer hydrophobicity and the structuring of water around the hydrophobic polymers is discussed. It is concluded that the hydrophobic interactions between the polymer and water are a major factor to influence the polymer swelling due to changes in the structuring of water around the hydrophobic polymers. These results might have important implications for the drug delivery and pervaporation, since these processes are influenced by the polymer swelling to a great degree. q 2000 Elsevier Science Ltd. All rights reserved.

Use of a diffusion model for assessing the performance of poly(vinyl alcohol) bioartificial pancreases

Islets of Langerhans surrounded by a semipermeable membrane to prevent an immune response by the host immunosystem is a potential way of treating type I diabetes mellitus. In this study, poly(vinyl alcohol) (PVA) tubular membranes with added polyethylene glycol to create pores in the skin layer were prepared to improve their diffusion property. In a static incubation study, islets cultured in the PVA tubular membranes still demonstrated their function of secreting insulin after 30 days. When the tubular PVA bioartificial pancreas was perifused in a small chamber with RPMI-1640 medium containing glucose at concentrations of 5.6-16.6 mmol/L, insulin release began to increase without delay. Therefore, such a membrane is an alternative potential material for a bioartificial pancreas. In addition, a mathematical mass transfer model of insulin release was developed and compared with the perifusion data. It was shown that satisfactory kinetics could be achieved with a PVA membrane. However, the model showed that the insulin output of islets cultured in the PVA tubular membrane must be increased to improve the performance significantly. These findings suggest that a bioartificial pancreas using a PVA membrane is a promising material, but the technique for seeding islets in the chamber requires further modification.

The effect of acetic acid on the structure and filtration properties of poly(vinyl alcohol) membranes

Abstract The role played by acetic acid in the formation of poly(vinyl alcohol) (PVA) membranes was investigated. Membranes were prepared from a casting solution of PVA, water, and acetic acid by immersion in Na 2 SO 4 /KOH/H 2 O coagulation bath. Experimental results show that the acetic acid additive exerts an influence on the structure and filtration properties of membranes. Not only the surface morphology but also the structure of cross-section could be modulated by adding the acetic acid in the casting solution. Obviously, the increase of the amount of acetic acid in the casting solution decreased the thickness of skin layer. This could be attributed to the fact that the increase of the amount of acetic acid in the casting solution increases H 3 O + ion of the casting solution enhancing the influx rate of coagulant medium for acid-base equilibrium. A mechanism describing the affinity between the PVA solution and the coagulant medium is proposed to estimate the PVA membrane structure by adding the acid. The results presented here offer a better understanding of relationships between the membrane formation mechanism and the skin structure when designing an asymmetric membrane with acetic acid as an additive.

Assessment and modeling of poly(vinyl alcohol) bioartificial pancreas in vivo

Pancreatic islets surrounded by a semipermeable membrane to prevent an immune response by the host immunosystem is a potential way of treating type I diabetes mellitus. Our previous in vitro studies have demonstrated that poly(vinyl alcohol) (PVA) membranes satisfy the basic requirements for a bioartificial pancreas. This study was designed to evaluate the performance of PVA tubular membrane chambers containing islets in vivo. When the m-2 type of PVA chamber was implanted into streptozotocin-induced diabetic rats, nonfasting blood glucose levels dropped from 500 +/- 35 mg/dl to the lowest value 210 +/- 22 mg/dl. Furthermore, the performance of the bioartificial pancreas can be enhanced by the increased numbers of implanted chambers. If three m(-2) chambers were implanted, nonfasting blood glucose levels in the diabetic rats decreased to 130-160 mg/dl and such a low blood glucose value was maintained for 1 month. This indicates that implanting three m-2 chambers in the diabetic rats could provide improved permeability of insulin to normalize blood glucose levels and improved survival of islets from the immune system of the recipient. For improving the design of the bioartificial pancreas, a mathematical model was developed to account for the changes in blood glucose levels of the diabetic rats. We demonstrated such a mathematical analysis was helpful to understand the characteristics of islet inside an artificial environment.