performance.

Two different composite polymeric membranes have been tested for this purpose. It has been tested for fluxes of pure carbon dioxide. The effect of pressure difference and the effect of feed pressure have been established. A polyethersulfone (PES) ultrafiltration membrane is used as a base membrane and two different selective polymeric top layers have been applied. This top layer consists of polyvinyl alcohol or polyamide type of polymers. The carbon dioxide fluxes are measured for both types of membranes at 313.4 K, pressures ranging from 10 to 200 bar and at 0 to 4 bar pressure difference across the membrane. Also the membranes have been tested for their physical stability. The effect of cross-linking of the top layer polymer has also been studied to improve the stability of the membranes. During the measurements, both membranes seem to have a typical relation between flux and the feed pressure.

Gas separation by selective separation through polymeric membranes is one of the fastest growing interests of separation technology. Membrane separation techniques present a large application potential to processes involving dense CO 2 in the phase of extract fractionation or solute recovery. During these steps, high depressurizing of the solvent is usually required, and the coupling of both processes can be quite useful in reducing pressure losses and recompression costs. Some of the studies have shown that few of the polymeric membranes used for carbon dioxide separation can be successfully used. The use of organo-mineral nanofiltration membranes were tested for the fractionation of triglycerides and purification of b- carotene in supercritical CO 2.

The polymeric membranes with polyethersulfone (PES) as base membrane and different polymeric top layer were obtained from TNO, The Netherlands. PES possesses very good chemical and thermal stability as indicated by its T g value 513 K. There are two different types of top layers have been developed, polyvinyl alcohol (PVA) and copolymer polyamide (IPC). Two different techniques were used to apply a thin top layer on a support, solution coating and interfacial polymerisation. Selective top layer of PVA (0.8 µm thick) was developed by solution coating method and the polyamide top layer is prepared by interfacial polymerization process. By interfacial polymerisation two very reactive bifunctional monomers (e.g. a di-acid chloride and a

diamine) or trifunctional monomers (e.g. trimesoylchloride) are allowed to react with each other at water/organic solvent interface and a typical network structure is obtained.

To start a flux measurement the on-off valve separating permeate and feed side is closed so that the membrane separates the feed and permeation section. Then the needle valve at the permeate side of the membrane was opened carefully to create a desired pressure difference across the membrane. Permeation experiments were performed with a 0.5 to 5 bar as pressure difference across the membrane. The flow rate of carbon dioxide is measured with the aid of flow meters.

Figure 6: Permeation of carbon dioxide through IPC (a) and PVA (b) membrane at

DP=3 bar

This typical trend of permeation at various feed pressure can be well explained by Hagen-Poiseuille’s law for viscous flows.

( 1 )

where Q = permeation, e = porosity , r = average pore radius, l = top layer thickness or effective pore length, t = tortuosity, Mco 2 = Molecular weight of carbon dioxide, r m= mean density, h m = mean viscosity. If all molecules move through pores at the same average velocity and in a laminar flow regime, the transport for viscous flows can be described by Hagen-Poiseuille’s law. For such small pore one can expect that the Knudson type of transport mechanism would dominate the overall flux behavior. However, from our calculations it comes out to be negligible. From figure 7 it can be seen that the experimental results are in good agreement with the Hagen-Poiseuille’s model given by equation (1).

Figure 7: Hagen-Poiseuille’s model fitting to the permeation data for IPC membrane

Membrane stability in supercritical carbon dioxide is important in order to get constant rejection performance. The well-known effect of plasticization and swelling of polymer due to carbon dioxide is certainly detrimental to the membrane performance. So to check these effects the stability of the membrane was experimentally determined. The flux through both the membranes was measured at 150 bar feed pressure and constant DP for a few hours, say up to four hours. PVA and IPC membrane showed some variations in the flux during the long run measurements. This indicates that these two membranes are not completely stable for use in supercritical carbon dioxide.

There are few techniques that have already been used to stabilize polymeric membranes. One of the techniques is adding the cross-linking in the top layers of the membrane. After inducing cross-linking (with 0.1wt% N 4) in top layer the permeation of carbon dioxide is found to be fairly constant over a time range of at least six hours. In figure 8 the permeation through cross-linked IPC membrane is given which is measured at 150 bar feed pressure and DP = 3 bar. But the cross-linking the top layer results in a less permeable membrane.

This is because the cross-linking makes the effective diameter of the pores reduced in size. The stability is acquired because entanglement of polymer chains. Cross-linked top layer reduces the freedom and flexibility of the polymer chains eventually reducing the permeation.

Figure 8: Stability test for IPC membrane

IPC membrane with more degree of cross-linking (0.3 wt% N 4) was also tested for its fluxes but there were no fluxes observed through the membrane. As the degree of cross-linking increases the carbon dioxide is unable to go through the membrane even up to 200 bar feed pressure and 4 bar DP. So it can be concluded that more cross-linking after a certain limiting value is not advantageous as it makes the effective pore size so small that CO 2 molecules can not enter into the polymer matrix to show any fluxes.

To conclude, different polymeric membranes have been compared in terms of flux and their stability towards carbon dioxide permeation. For PVA and IPC membranes, which are not cross-linked, have similar permeation behavior. Both membranes show a maximum in the permeation as a function of feed pressure of CO 2. In terms of stability, permeation of CO 2 through both membranes is stable for three to four hours. After cross-linking the IPC membrane the permeation of CO 2 decreases and stable permeation behavior for more than six hours has been observed. These results open the possibility for regeneration of supercritical carbon dioxide using polymeric membranes.