Permeate-side Conditions

In pervaporation and vapor-permeation processes the partial vapor pressures of the components at the feed side are fixed by the nature of the components, composition, and temperature of the feed, whereas the total pressure is of no influence, as long as the liquid mixture can be regarded as incompressible. 

Only by increasing the temperature of the Hquid mixture the partial vapor pressure can be increased for a given feed mixture. Therefore the driving force for the transport of matter through the membrane is appHed and maintained by reducing the partial vapor pressure at the permeate side. 

The influence of the permeate side partial pressure can best be seen from a combination of Eq. (9) and Eq. (2). 

The argument under the logarithms in Eq. (24) is the ratio of the partial vapor pressure of the component in the feed over that on the permeate side. In most practical application %i,Peed is fairly small (the minor component has to be removed from the feed), but permeate will be close to unity for a high-selective 

Different means have been proposed in order to reduce the permeate-side partial vapor pressure (Fig. 3.4)  

---

The design of any real pervaporation and vapor-permeation installation has thus to be based on experimental data measured in the laboratory under conditions as similar as possible to those of the subsequent full-size plant. These conditions include the flow regime of the feed mixture, the temperature and the geometry of the feed side, the composition and nature of the feed mixture, the permeate side geometry and partial vapor pressure. From the experimental data the partial transmembrane fluxes of all components of a mixture and thus the selectivity can be determined as a function of composition, temperature and permeate-side conditions for the respective mixture and geometry. In practice the permeate-side conditions (total pressure, condensation temperature) are kept as close as possible to those expected in the final plant, thus changes of these parameters do not need to be considered. Figure 3.3 depicts the partial fluxes of EtOH and water measured for a PVA-membrane. 

For the determination of the four constants A, B, C, D of the flux equations (13) and (14), at least four experiments at different concentrations, but constant temperature, have to be performed. Another measurement, at an already fixed constant concentration, but a second temperature will yield the (then constant) activation enthalpy or activation temperature. When the activation temperature additionally depends on the feed concentration, two more measurements are necessary. In all these experiments other test parameters like feed velocity or permeate-side conditions should be as close as possible to those of the later real plant. By using Eqs. (13) to (15) the performance of the membrane can then be calculated with sufficient accuracy and even large industrial plants can be designed within the range of concentrations and temperatures of the experiments. 

The schematic diagram of the experimental setup is shown in Fig. 2 and the experimental conditions are shown in Table 2. Each gas was controlled its flow rate by a mass flow controller and supplied to the module at a pressure sli tly higher than the atmospheric pressure. Absorbent solution was suppUed to the module by a circulation pump. A small amount of absorbent solution, which did not permeate the membrane, overflowed and then it was introduced to the upper part of the permeate side. Permeation and returning liquid fell down to the reservoir and it was recycled to the feed side. The dry gas through condenser was discharged from the vacuum pump, and its flow rate was measured by a digital soap-film flow meter. The gas composition was determined by a gas chromatograph (Yanaco, GC-2800, column Porapak Q for CO2 and (N2+O2) analysis, and molecular sieve 5A for N2 and O2 analysis). The performance of the module was calculated by the same procedure reported in our previous paper [1]. 

Pervaporation. Pervaporation differs from the other membrane processes described so far in that the phase-state on one side of the membrane is different from that on the other side. The term pervaporation is a combination of the words permselective and evaporation. The feed to the membrane module is a mixture (e.g. ethanol-water mixture) at a pressure high enough to maintain it in the liquid phase. The liquid mixture is contacted with a dense membrane. The other side of the membrane is maintained at a pressure at or below the dew point of the permeate, thus maintaining it in the vapor phase. The permeate side is often held under vacuum conditions. Pervaporation is potentially useful when separating mixtures that form azeotropes (e.g. ethanol-water mixture). One of the ways to change the vapor-liquid equilibrium to overcome azeotropic behavior is to place a membrane between the vapor and liquid phases. Temperatures are restricted to below 100°C, and as with other liquid membrane processes, feed pretreatment and membrane cleaning are necessary. 

Figure 2.17 Flux of n-hexane through a rubbery membrane as a function of the hexane concentration difference in the membrane. Data taken from both reverse osmosis ( ) and pervaporation (O) experiments. Feed-side and permeate-side membrane concentrations, Ci0 m) and Cie m), calculated from the operating conditions through Equations (2.26), (2.36) and (2.76). Maximum flux is obtained at the maximum concentration difference, when the permeate-side membrane concentration cit(m)), equals zero [19]. Reprinted from Driving Force for Hydraulic and Pervaporation Transport in Homogeneous Membranes, D.R. Paul and D.J. Paciotti, J. Polym. Sci., Polym. Phys. Ed. 13, 1201 Copyright 1975. This material is used by permission of John Wiley Sons, Inc.

Pervaporation operates under constraints similar to those for low-pressure gas separation. Pressure drops on the permeate side of the membrane must be small, and many pervaporation membrane materials are rubbery, so both spiral-wound modules and plate-and-frame systems are in use. Plate-and-frame systems are competitive in this application despite their high cost, primarily because they can be operated at high temperatures with relatively aggressive feed solutions, conditions under which spiral-wound modules might fail. 

Equation (9.1) is the preferred method of describing membrane performance because it separates the two contributions to the membrane flux the membrane contribution, P /C and the driving force contribution, (pio — p,r). Normalizing membrane performance to a membrane permeability allows results obtained under different operating conditions to be compared with the effect of the operating condition removed. To calculate the membrane permeabilities using Equation (9.1), it is necessary to know the partial vapor pressure of the components on both sides of the membrane. The partial pressures on the permeate side of the membrane, p,e and pje, are easily obtained from the total permeate pressure and the permeate composition. However, the partial vapor pressures of components i and j in the feed liquid are less accessible. In the past, such data for common, simple mixtures would have to be found in published tables or calculated from an appropriate equation of state. Now, commercial computer process simulation programs calculate partial pressures automatically for even complex mixtures with reasonable reliability. This makes determination of the feed liquid partial pressures a trivial exercise. 

The calculations shown in Figure 11.18 assume that a hard vacuum is maintained on the permeate side of the membrane. The operating and capital costs of vacuum and compression equipment prohibit these conditions in practical systems. More realistically, a carrier facilitated process would be operated either with a compressed gas feed and atmospheric pressure on the permeate side of the membrane, or with an ambient-pressure feed gas and a vacuum of about 0.1 atm on the permeate side. By substitution of specific values for the feed and permeate pressures into Equation (11.19), the optimum values of the equilibrium constant can be calculated. A plot illustrating this calculation for compression and vacuum operation is shown in Figure 11.19. 

Pervaporation is a concentration-driven membrane process for liquid feeds. It is based on selective sorption of feed compounds into the membrane phase, as a result of differences in membrane-solvent compatibility, often referred to as solubility in the membrane matrix. The concentration difference (or, in fact, the difference in chemical potential) is obtained by applying a vacuum at the permeate side, so that transport through the membrane matrix occurs by diffusion in a transition from liquid to vapor conditions (Figure 3.1). Alternatively, a sweep gas can be used to obtain low vapor pressures at the permeate side with the same effect of a chemical potential gradient. 

A set of boundary and initial conditions (BCs and ICs) is necessary to solve the system of Equations 13.8,13.9 and 13.11. The specific contribution of the permeation is expressed by means of a BCrelated to the membrane surface for both reaction (13.8) and permeation side (13.9). It is equal to the permeating flux. 

Due to its complexity (conversion and separation in the same unit) and because this system has been most widely studied experimentally, CMRs for dehydrogenation (or more generally for equilibrium-restricted reactions) have been the subject of modeling approaches [6, 54-59]. The modeling of CMRs requires mass and energy balances in both feed and permeate sides of the reactor (plug-flow behavior is always assumed) and appropriate boundary conditions. Generally these models fit the experimental data well. 

The conditions are substantially more favorable for the microporous catalytic membrane reactor concept. In this case the membrane wall consists of catalyti-cally active, microporous material. If a simple reaction A -> B takes place and no permeate is withdrawn, the concentration profiles are identical to those in a catalyst slab (Fig. 29a). By purging the permeate side with an inert gas or by applying a small total pressure difference, a permeate with a composition similar to that in the center of the catalyst pellet can be obtained (Fig. 29b). In this case almost 100% conversion over a reaction length of only a few millimeters is possible. The advantages are even more pronounced, if a selectivity-limited reaction is considered. This is shown with the simple consecutive reaction A- B- C where B is the desired product. Pore diffusion reduces the yield of B since in a catalyst slab B has to diffuse backwards from the place where it was formed, thereby being partly converted to C (Fig. 29c). This is the reason why in practice rapid consecutive reactions like partial oxidations are often run in pellets composed of a thin shell of active catalyst on an inert support [30],... 

When only the feed side and permeate side mass balance equations are considered under the isothermal condition, the resulting equations arc a set of first-order ordinary differential equations. Furthermore, a co-current purge stream renders the set of equations an initial value problem and well established procedures such as the... 

Plants are cleaned, sanitized, and rinsed immediately after processing, and right before processing to ensure satisfactory initial process conditions from microbiological standpoint [3]. Because chlorine is freely permeable to most membranes that it is able to sanitize the permeate side of the system as well as the retentate side, using solutions of sodium hypochlorite containing 100-200 ppm of active chlorine is a common sanitation technique for many membranes, except cellulose acetate reverse osmosis membranes, which can only tolerate brief exposure to chlorine at 10-50 ppm level [3]. 

The PDMS and POMS membranes and process parameters were investigated using experimental studies for comparison (Sampranpiboon et al., 2000). The following operational conditions were applied for both membranes temperature, 303.15 K downstream pressure (permeate side), 0.3997 kPa membrane thickness, 10 pm. 

The gas is applied as a mixture to the retentate (high pressure) side of the membrane, the components of the mixture diffuse with different rates through the membrane under the action of a total pressure gradient and are removed at the permeate side by a sweep gas or by vacuum suction. Because the only segregative mechanisms in mesopores are Knudsen diffusion and surface diffusion/capillary condensation (see Table 9.1), viscous flow and continuum (bulk gas) diffusion should be absent in the separation layer. Only the transition state between Knudsen diffusion and continuum diffusion is allowed to some extent, but is not preferred because the selectivity is decreased. Nevertheless, continuum diffusion and viscous flow usually occur in the macroscopic pores of the support of the separation layer in asymmetric systems (see Fig. 9.2) and this can affect the separation factor. Furthermore the experimental set-up as shown in Fig. 9.11 can be used vmder isobaric conditions (only partial pressure differences are present) for the measurement of diffusivities in gas mixtures in so-called Wicke-Callenbach types of measurement. 

Hydrogen from the membrane reactor is converted in a gas turbine with a high efficiency. The process efficiency will increase when the hydrogen production (CO conversion) and recovery (on the permeate side) from the membrane reactor is raised. CO2 abatement increases with increasing recovery of carbon components on the retentate side of the membrane. The performance of the reactor can be measured in terms of these three parameters. The boundary conditions for the membrane reactor in the total system depends upon final... 

The value of K is determined from an RO experiment in which the permeate side of the membrane is exposed to a recirculation solution to produce a known c /cj ratio to simulate CCRO conditions. The measured permeate flux is substituted into Equation 15 to compute the value of K that would match the experimental value of J. ... 

RO tests were conducted at various pressures to measure ethanol rejection and permeate flux as functions of feed concentration. To mimic CCRO conditions, a solution equal in concentration to the feed was used for permeate-side recirculation, and the changes in flux was monitored as recirculation was switched on and off. 

---