Characterization of bore pressure change effects on Matrimid® fiber performance in pervaporation of acetic acid and water mixtures

https://doi.org/10.1016/j.ces.2005.12.019Get rights and content

Abstract

A mathematical model was used to account for the bore pressure change effects observed for Matrimid® hollow fiber membranes to obtain the inherent water permeability and membrane selectivity in pervaporation of 20%wt acetic acid (HAc) and water mixtures. The modeled water flux was close to the experimental result for a large bore size fiber (bore size 350μm) with 20 cm length due to its small bore pressure change. This indicated that the large bore size fiber can effectively minimize the water partial pressure change inside the bore and thus maximize the water driving force across the membrane. In addition to the large bore size, a short length fiber can also decrease the bore pressure change and improve all of the penetrant fluxes significantly; however, the short module is a less desirable option from the standpoint of practical module formation due to the need to form more modules for a given application. Experimentally, an unexpectedly decreased HAc flux was observed for the large bore size fiber with 20 cm length. Our analysis suggested that a reduced swelling effect was generated by the low HAc pressure change inside the large bore fiber. Based upon the modeled selectivity and permeability that accounted for the bore pressure change, the water permeability was found to be weakly HAc-concentration dependent, while the HAc permeability was strongly HAc-concentration dependent for 20%wt HAc/H2O feed streams.

Introduction

Acetic acid (HAc) is an important industrial chemical that has been widely applied in the manufacture of acetate esters. For example, acetic acid can react with hydroxyl-group polymers such as cellulose to yield cellulose acetate, which is used to make films and textiles. Vinyl acetate, another ester of acetic acid, polymerizes with acetic acid to form poly(vinyl acetate), which is used in water-based latex paints and in glues for paper and wood. Acetic acid can also be used to synthesize pharmaceuticals such as aspirin and fungicides, and as a solvent for many organic compounds such as purified terephthalic acid. However, the removal of water from acetic acid solutions is always a challenging problem (Kirk and Othmer, 2000). Membrane-based separation, with low energy and operation cost, is a promising alternative to the current energy-intensive distillation column (Baker, 2000).

A defect-free Matrimid® hollow-fiber based pervaporation system has been developed in our previous work (Zhou and Koros, 2006). This work proved the feasibility of adapting defect-free gas separation Matrimid® hollow fiber membranes to pervaporation of a model 20%wt HAc/H2O mixture at 101.5 °C. Shell feed was used in this study to avoid excessive pressure drop associated with liquid flow inside the bore (Rautenbach and Albrecht, 1985). Due to high water flux associated with pervaporation through the bore side of the fiber, however, this relatively small diameter or long fiber may lead to complicated permeation behavior with shell feed. That is, whenever gas or vapor permeates through a hollow fiber, the pressure change associated with high penetrant flux will be generated in the bore side along the axial direction of the fiber. Therefore, the overall production rate is reduced, because an increase of the pressure in the bore side decreases the driving force across the membrane wall. In addition, the bore pressure change affects the highly permeable gas much more than the less permeable one. The membrane selectivity for the high permeable/low permeable gas pair will be underestimated if the pressure change inside the bore is not accounted for properly (Nader and Abbas, 1999, Lim et al., 2000). In this study, an additional more subtle effect is shown to arise when the bore pressure change is significant. Specifically acetic acid transport properties show complex concentration dependent effects under these condition that negatively impact membrane selectivity. By overcoming the bore pressure change limitations, unexpected improvements in selectivity were observed.

Several analytical and numerical solutions have appeared in the literature to describe the pressure drop in hollow fiber membranes (Lim et al., 2000; Pan and Habgood, 1978a, Pan and Habgood, 1978b; Chern et al., 1985, Goran and Milan, 2001, Mellis et al., 1993, He et al., 1993, Wang et al., 2002). Pan and Habgood, 1978a, Pan and Habgood, 1978b modeled multicomponent permeation systems with high flux in asymmetric hollow fiber membranes including the permeate pressure variation. Chern et al. (1985) reported that the pressure buildup in the fiber lumen was substantial for smaller fibers when the permeator was operated at high permeation rates for the shell feed mode. Thorman et al. (1975) studied the pressure drop and separation in a binary system through a bundle of silicone rubber capillary tubing. Lim and Li (2001) modified the Hagen–Poiseuille equation to account for fiber permeability and gas compressibility. Koops et al. (1994) calculated the pressure loss in the bore side of a polysulfone hollow fiber for separation of HAc/H2O mixtures based on the Bernoulli equation and concluded that negligible loss of the driving force existed due to the pressure loss. Coker et al. (1998) and Thundyil and Koros (1997) developed a model to simulate multicomponent gas separation with cocurrent, countercurrent, and crossflow patterns using hollow fiber contactors. Thundyila et al. (1999) analyzed and characterized the dependence of permeability on the permeate pressure for a glassy polyimide. Nonzero permeate pressures were shown to depress the selectivity of the carbon dioxide and methane system.

The Hagen–Poiseuille equation, coupled with the permeation equation, is used to establish a simple mathematical model to describe the flow behavior of each penetrant through the bore side of a hollow fiber in this paper. The bore pressure change mainly relies on outer and inner diameters of a hollow fiber, available separation fiber length, and penetrant flux.

Section snippets

Model development

A finite element method is used to simulate the pressure change in the bore side of a hollow fiber. In other words, a hollow fiber has to be divided into a number of differential elements and a mass balance should be built in each differential element. Fig. 1 illustrates the schematic diagram of this method for a “one-ended single-fiber” module. This concept can be easily extended to a double-ended one.

The key assumptions involved in the mathematical model are as follows:

  • (1)

    The behavior of the

Results and discussion

The model was used to investigate the bore pressure change effects on hollow fibers for pure gas permeation tests, pure deionized (DI) water feed tests, and 20%wt HAc concentration feed tests. Table 1 shows the spinning condition, bore size, and skin thickness of the hollow fibers that are used in the model. Clearly all of the hollow fibers are defect-free based on the ability to achieve intrinsic O2/N2 separation factors for Matrimid® (Zhou and Koros, 2006). However, the fiber permeances from

Conclusion

The bore pressure change of a hollow fiber can affect fluxes of all penetrants, but highly permeable penetrants are influenced much more seriously. The true membrane selectivity will be underestimated for high/low permeable gas pairs without the bore pressure change correction if the highly permeable penetrant is the end-product. Large bore size or short length hollow fibers have been proved to be very effective to minimize the bore pressure change in this model. In addition, previously

Acknowledgements

The authors sincerely thank BP-Amoco for their financial support.

References (21)

There are more references available in the full text version of this article.

Cited by (9)

  • Enhancing the property of composite pervaporation desalination membrane by fabricating a less resistance substrate with porous but skinless surface structure

    2022, Desalination
    Citation Excerpt :

    Therefore, the pure water fluxes of the PSF membranes (PA15-60) increase from 654.3 ± 18.2 kg/(m2·h·bar) to 1014.8 ± 27.6 kg/(m2·h·bar) benefiting from the enlarged surface pore sizes and skinless structures. In a PV desalination process, the water molecules desorb from the backside of the dense selective layer of PV composite membrane and then permeate through the porous substrate in a gas state [37,38]. Therefore, determining the gas transport property of the substrate can better understand the transport resistance of the support layer of the composite membranes.

  • Fabrication of high-performance composite membranes based on hierarchically structured electrospun nanofiber substrates for pervaporation desalination

    2021, Journal of Membrane Science
    Citation Excerpt :

    This distinct difference at PV performance can be explained by the pressure build-up at the interface between the selective layer and substrate. By applying the solution-diffusion model to a PV composite membrane [40–43], water molecule first dissolves in the dense selective layer, diffuse through the top selective layer, and finally desorbs as water vapor at the top surface of the porous substrate. However, the permeation of the water vapor will be restricted at uncoherent or dead-end hole regions of the support layer, and thus a pressure build-up of water vapor is produced at the interface between the bottom surface of dense selective layer and the top surface of the support layer as shown in Fig. 8a. Supposing there is no intrusion of the selective layer into the substrate, we can estimate the gas transport property by the resistance model proposed by Henis and Tripodi using Eq. (6) [44].

  • Polyimide hollow fiber membranes and their applications

    2021, Hollow Fiber Membranes: Fabrication and Applications
  • Dehydration of alcohols by pervaporation through polyimide Matrimid<sup>®</sup> asymmetric hollow fibers with various modifications

    2008, Chemical Engineering Science
    Citation Excerpt :

    Heat treatment has been considered as an effective method to improve membrane selectivity and stability since the birth of RO and Celgard type membranes. Its effectiveness in improving asymmetric polymeric hollow fiber performance has been proved by Liu et al. (2005) and Zhou and Koros (2006a,b). Experimental in Sections 3.2 and 3.3 have demonstrated that the proposed chemical modification has the ability to produce hollow fibers with certain degree of separation efficiency.

  • Membrane-Based Technologies for Environmental Pollution Control

    2020, Membrane-Based Technologies for Environmental Pollution Control
View all citing articles on Scopus
1

Currently work in Hercules Incorporated, Wilmington, DE 19808.

View full text