SciELO - Scientific Electronic Library Online

vol.17 issue4-7A hybrid feedforward neural network model for the cephalosporin C production processDevelopment of a hydrodynamic model for air-lift reactors author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand




Related links


Brazilian Journal of Chemical Engineering

Print version ISSN 0104-6632On-line version ISSN 1678-4383

Braz. J. Chem. Eng. vol.17 n.4-7 São Paulo Dec. 2000 



C.C.Pereira, R.Nobrega and C.P.Borges
Federal University of Rio de Janeiro, COPPE/Chemical Engineering Program,
P.O. Box 68502, ZIP 21945-970, Phone: (021) 590 -2241, (021) 270-2189
Fax: (021) 590-7135,Rio de Janeiro - RJ, Brazil


(Received: September 9, 1999 ; Accepted: April 6, 2000)



Abstract - During hollow fiber spinning many variables are involved whose effects are still not completely clear. However, its understanding is of great interest because the control of these variables may originate membranes with the desired morphologies and physical properties. In this work, the phase inversion process induced by the immersion precipitation technique was applied to prepare hollow fibers membranes. It was verified that some of the variables involved, can promote a visco-elastic polymer solution expansion, called die-swell phenomenon, which is undesired since it may lead to low reproducibility of the permeation properties. The effects of the distance between spinneret and precipitation bath, the bore liquid composition, and the polymer solution composition were analyzed and discussed in order to avoid this phenomenon. According to the results, it was verified that the parameters investigated might promote a delay precipitation, which restrained the visco-elastic expansion.
Keywords: hollow fiber, die-swell, membrane formation




Membrane technology has a wide range of applications in several fields of industry. The membranes are highly selective and the separation membrane processes are characterized by low energy consumption, possibility of different module design and easy scale up. These advantages make these processes superior to many other established separation processes. Although some applications have already reached a great development, many aspects in the membrane formation are not completely clear. Several research studies (Pereira, 1999; Di Luccio, 1994; Pinnau, 1994) have focused the membrane formation in order to control the properties of the resulting membrane and, consequently, to optimize these applications, as well as other developing membrane processes, such as gas separation and pervaporation (Souza et al, 1998).

In these processes, the membrane geometry – flat or cylindrical, and consequently the module configuration, will be chosen according to the separation process, i.e., mixture characteristics, operational conditions, cleaning, investment and operation costs, and possibility of using compact systems, as well as, membrane replacement. In many cases, the hollow fiber geometry is applied, because it is self-supported and it has a high relation between permeation area and module volume, which means a reduction in the investment cost. However, the hollow fiber formation is much more complex than the flat geometry, since many other variables are involved. Besides that, the die-swell phenomenon, which is a visco-elastic expansion, leads to a deformation in the internal fiber diameter, and consequently, leads to membrane low reproducibility. In the literature some works are reported, in which this effect was minimized by adding a high solvent concentration to the bore liquid (Borges, 1993; Pereira et al, 1997, Van’t Hof, 1988, Roesink, 1989) In the present work, in order to avoid this phenomenon, besides the influence of the bore liquid composition, it was also investigated the distance between spinneret and the precipitation bath, as well as the polymer solution viscosity.



Usually, the membranes are prepared using the phase inversion technique by immersion precipitation. In this technique, a polymer solution is immersed into a precipitation bath, which is a non solvent to the polymer, or a mixture non-solvent/ solvent. After immersion, the solvent from the polymer solution diffuses into the precipitation bath, whereas the non-solvent diffuses into the polymer solution, i.e., mass transfer between bath and solution leads to changes in the composition and promotes liquid-liquid demixing process. Figure 1 represents schematically the immersion precipitation process.



Membranes with different morphologies may be obtained depending on the thermodynamics parameters involved, as well as, on the kinetics of precipitation of polymer solution. The mechanisms of nucleation and growth and the spinodal separation have been widely used to describe liquid-liquid separation evolution. According to Figure 2, which represents the free mixture energy gradient as a function of polymer composition, when a concentrated polymer solution (Fp >B2) changes its composition to the metastable region, where Fp varies in the range B2-C2, a nucleus is formed with polymer lean phase (Fp = A1). On the other hand, when a diluted polymer solution changes its composition to the metastable region, where Fp varies in the range B1-C1, a nucleus of polymer rich phase is formed (Fp = A2). These nuclei grow due to the mass transfer with the surrounding phase. This mechanism is known as nucleation and growth.



When the nuclei formation is not favoured, the polymer solution concentration can go into the instable region, where Fp varies in the range C1-C2. The liquid-liquid separation takes place, but the nuclei are not formed. This mechanism is called spinodal separation.

The separation takes place until the solidification of the polymer structure, by gelification, vitrification or crystallization phenomena.

Some models have been developed in order to describe the mass transfer before the liquid-liquid separation. According to Reuvers (1987), two mechanisms can be distinguished which depend on the kinetics of precipitation. The separation can take place immediately after the polymer solution immersion into the precipitation bath. This is called instantaneous precipitation. When it takes place after a certain time, it is assumed a delayed precipitation. Normally, the instantaneous precipitation occurs when there is a good affinity between solvent from the polymer solution and the non-solvent from the precipitation bath. On the reverse, the delayed precipitation may take place.



Hollow fiber membranes were prepared using poly(ether sulfone)- PES (VICTREX 5200) from ICI, and poly(ether imide) – PEI (ULTEM 1000) from BASF, as polymers. N-methyl-2-pyrrolidone -NMP from Merck, was used as solvent; and adipic acid – AD, propionic acid - PA, purchased from Rio Lab, and poly(vinyl pyrrolidone) – PVP-K90, acquired from Aldrich Chemie, were used as additives. Distillated water was used as the external precipitation bath, whereas as bore liquid, solutions of water/NMP were used. The compositions of the polymer solutions investigated, as well as the bore liquid composition and the distance spinneret-precipitation bath used are shown in Table 1.



During the hollow fiber spinning, the polymer solution is extruded through an annular orifice towards the external precipitation bath, whereas, to obtain the fiber lumen the bore liquid is pumped towards the spinneret and, then, it is co-extruded through an inner orifice. From the precipitation bath, the hollow fiber passes through a rinsing bath, and afterwards it is placed in a storage tank. The spinning apparatus is represented schematically in Figure 3.



After precipitation, the obtained membranes were dried by a liquid replacement technique in order to avoid collapsing the pores.

The membranes morphologies were observed by Scanning Electron Microscopy – SEM (JEOL JSM – 5300). The samples were prepared by immersion into liquid nitrogen in order to avoid deformation during fracture. Afterwards the samples were coated with gold by a sputtering device (JEOL JFC-1500).

Light transmittance measurements were carried out to characterize the polymer solution kinetics of precipitation. The polymer solutions were casted on a glass plate and immersed into a precipitation bath. Afterwards, a computer registered the light transmittance decay as a function of time. The light transmittance apparatus is represented in Figure 4.




The kinetics of precipitation of PEI/PVP/NMP was investigated, using water/NMP in different concentrations as precipitation bath. The results are shown in Figure 5.



Figure 5, shows that increasing the solvent concentration in the precipitation bath the precipitation kinetics of the polymer solution tends to decrease. It indicates that the mass transfer between the precipitation bath and the polymer solutions is reduced, i.e., the driving force to water inflow and solvent outflow decreases.

The effect of bore liquid composition was investigated using the system PEI/PVP/NMP. The morphologies obtained using the null distance spinneret-precipitation bath are shown in Figure 6.



At low solvent concentration in the bore liquid it was observed in Figure 6 (a) and (b) an irregular shape in the fiber internal perimeter. This effect occurs due to the polymer visco-elastic solution expansion, called die-swell phenomenon, and the high water concentration in the bore liquid effect which leads to instantaneous precipitation. In Figure 6(c) it cannot be observed any irregular shape by increasing the solvent concentration in the bore liquid, which decreases the water chemical potential. In this case, a delay precipitation is characterized, allowing enough time to accommodate the existing tensions before final polymer precipitation. When a delayed precipitation occurs in the internal surface of the membrane, the nuclei have more time to grow. Thus, the resulting membrane has a more opened pore structure in the cross section close to the internal surface, as can be observed in Figure 6.

Depending on the desired membrane properties, the use of a high concentration of water in the bore liquid may be more suitable. Besides that, due to the delayed precipitation in the region close to the fiber internal surface, the polymer solution may lose mechanical resistance for extrusion. In these conditions the addition of solvent, in order to avoid die-swell effect, becomes inadequate. Thus, in order to keep the water content in the bore liquid, other variables were investigated.

The influence of distance spinneret-precipitation bath was investigated. It allows the inflow of water from the atmosphere to the extruded polymer solution, which changes the polymer solution composition. In this work, the polymer solutions investigated were constituted of PES/PA/NMP and PES/AD/NMP. In these polymer solutions, one must consider that they are composed of acid:base Lewis complexes, where it is assumed that the acid is non-solvent, and the base is solvent to the polymer (Fritzsche, 1990; Pereira et al, 1998). It is assumed that the water inflow can dissociate the acid:base Lewis complex, and so increases the non-solvent (water and acid) concentration, which may lead to instantaneous precipitation and promote the die-swell effect. The photomicrographs of the cross sections obtained using PES/PA/NMP (30/19/51%w/w) are shown in Figure 7.



The photomicrograph in Figure 7 (a) shows that the die-swell phenomenon occurs again due to the presence of pure water in the bore liquid. Increasing the distance spinneret-precipitation bath to 30 cm, the polymer solution has more time to reach the water from the external precipitation bath. During this time, the inflow of water towards the region close to the external surface is reduced, since it occurs only due the sorption of water from the atmosphere. This provides a delayed precipitation close to the external surface, and the die swell phenomenon is minimized.

The polymer solution composed of PES/AD/NMP (30/18.9/51.1 %w/w) is characterized by high viscosity values (Pereira, 1999). The effect of polymer solution viscosity to die-swell phenomenon was also investigated in this work. Figure 8 shows the photomicrographs of the cross sections obtained.



Figure 8 shows that the die-swell phenomenon did not affect the hollow fiber morphologies. The high viscosity of this polymer solution inhibits the water inflow to the polymer solution. This promotes stability to the solution, and enough time to accommodate the existing tensions. For the distance spinneret-precipitation bath of 30 cm, the polymer solution is exposed to the atmosphere for 33 seconds. However, according to another part of this study to be published (Pereira, 1999), this solution starts its precipitation after approximately 20 minutes, when exposed to atmosphere. This means that the solution region close to the external surface remains stable, during the time of exposition investigated. Hence, the die-swell effect is not favoured.

Van’t Hof (1988) considered that the die-swell effect is characteristic of high viscous solutions. However, the results from this work show that one must consider that increasing the polymer solution viscosity may promote a delay in the kinetics of precipitation. This allows the accommodation of the mechanical tensions due to the visco-elastic expansion, avoiding the internal perimeter deformation.



The results of this work showed that the die-swell effect can be avoided by increasing the solvent concentration in bore liquid, as well as by increasing the distance spinneret-precipitation bath even with a high content of water in the bore liquid. It was verified also that a high polymer solution viscosity might minimize the die-swell effect. All these parameters lead to reduce the mass transfer between the polymer solution and the concentration bath. The delayed precipitation takes place, which provides time enough for accommodating the tensions due to visco-elastic expansion. Hence, the internal perimeter deformation is inhibited.



C.C. Pereira would like to thank CAPES/ CNPq for scholarship during the development of this work.



AD adipic acid
NMP n-methyl-2-pyrrolidone
PA propionic acid
PEI poly(ether imide)
PES poly(ether sulfone)
PVP poly(vynil pyrolidone)
SEM Scanning Electron Microscopy

Greek Symbols

DGm free energy of mixture gradient (FL/mol)
Fp polymer concentration



Borges, C.P., Fibras ocas compostas para a remoção de poluentes orgânicos de soluções aquosas pelo processo de pervaporação, Tese de D.Sc., COPPE/UFRJ, Rio de Janeiro, Brasil (1993).        [ Links ]

Di Luccio, M., Borges, C.P., Nobrega, R., Habert. A.C., Microporous membranes by phase inversion I. Polycarbonate/NMP/water system, Proceedings of II CITEM (1995).        [ Links ]

Fritzsche, A.K., Cruse, C.A., Kesting, R.E., Murphy, M.K., Polysulfone hollow fiber membranes spun from Lewis acid:base complexes. II. The Effect of Lewis acid to base ratio on membrane structure, Journal of Applied Polymer Science, v. 39, p. 1949 (1990).        [ Links ]

Pereira, C.C., Nobrega, R., Borges, C.P., Die-swell phenomena during the formation of hollow fiber membranes characterized by Scanning Electron Microscopy and Digital Image Processing, XVI Meeting of the Brazilian Society for Electron Microscopy, 266-267, Caxambu – MG, 1-5 set. 1997.        [ Links ]

Pereira, C.C., Desenvolvimento de membranas anisotrópicas para separação de gases, D.Sc. thesis (1999).        [ Links ]

Pinnau, I., Wind, J, Peinemann, K.V., Ultrathin multicomponent. Poly(ether sulfone) membranes for gas separation made by dry/wet phase inversion, Ind. Eng. Chem. Res., v.29, p.2028 (1990).        [ Links ]

Reuvers, A.J., Membrane formation. Diffusion induced demixing processes in ternary systems, Ph.D thesis, University of Twente (1987).        [ Links ]

Roesink, H.D.W., Microfiltration membrane development and module design, Ph.D. thesis, University of Twente, Enschende, The Netherlands (1989).        [ Links ]

Souza, J.N.M, Pereira, C.C., Borges, C.P., Fibras Ocas de Poli(éter sulfona ) por imersão e precipitação utilizando ácido acético como aditivo à solução polimérica, II Encontro Regional de Polímeros – ABPOL Regional RJ, Rio de Janeiro-RJ, p.26 (1998).        [ Links ]

Van’t Hof, J.A, Wet spinning of assymetric hollow fibre membranes for gas separation, Ph.D. dissertation, Twente University, Enschede, The Netherlands (1988).        [ Links ]

Creative Commons License All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License