Robert Rautenbach

Separation potential of nanofiltration membranes

Abstract Nanofiltration has two interesting features: 1. - a fractionation capacity for different organic components in aqueous solutions, the ‘cut-off’ being in the range of 300 kg/kmol molecular weight 2. - the potential of realizing the Donnan-effect with respect to anions of different valency. Promising areas of application are cases where • - a high rejection for single-valent salts is not required or even unwanted • - a separation of anions of different valency must be achieved • - fractionation of high and low molecular weight organics is required Systematic experiments concerning the fractionation of different organics in aqueous solutions like amino acids or solvents are carried out by the Institut fur Verfahrenstechnik of Aachen Technical University. Hopefully these experiments will result in • - relationships describing material transport in nanofiltration membranes • - a broad knowledge of separation problems which can be solved economically by nanofiltration — alone or in combination with other separation processes.

Separation of organic binary mixtures by pervaporation

Pervaporation is a membrane separation process which has the inherent advantage of excellent selectivity for a number of mixtures that are otherwise difficult to separate (e.g. azeotropic mixtures or mixtures of liquids with only small differences in vapor pressures). Important parameters for pervaporation are the operating pressure on the permeate side and the temperature drop at the membrane interface caused by the phase change from liquid to vapor. This paper presents a mathematical model for the mass transport of binary mixtures within the membrane. The calculations are based on the transport equations of Shelden and co-workers [1]. These equations were solved without any simplifying assumptions and tested experimentally with benzene—cyclohexane mixtures. The limitations of the mathematical model and possible methods of improving it are discussed. The temperature drop at the membrane interface and its influence on the rate of permeation are also examined.

The separation potential of pervaporation

Abstract This paper discusses transport equations for pervaporation based on the sorption—diffusion model, which is widely accepted for other membrane processes like reverse osmosis. The advantages of these phenomenological-transport equations are — a sufficient degree of accuracy — a minimum number of free parameters (material properties) is needed — the material properties can be determined from relatively simple, steady-state experiments. Their validity has been verified by experiments with the benzene/cyclohexane system using polyethylene membranes. Furthermore, pervaporation is compared with reverse osmosis. The comparison reveals that the high separation potential of pervaporation can also be achieved by reverse osmosis — at least in theory. Practically, however, the separation potential of reverse osmosis is much lower because of membrane imperfections.

The separation potential of pervaporation : Part 2. Process design and economics

Abstract Pervaporation is a relatively complex process compared to other membrane processes like reverse osmosis for two reasons: — the process is sensitive to pressure losses at the permeate side — the evaporation enthalpy has to be transferred to the membrane surface (permeate side). Selectivity and flux can decrease markedly in case of hindered permeate flow. This is demonstrated by numerical design calculations of hollow-fiber modules. The calculations indicate that optimal fiber dimensions of hollow-fiber pervaporation modules should be much larger than those employed in RO and gas permeation modules. In principle, several alternatives exist for the supply of the evaporation enthalpy. The most economical solution seems to be to draw this energy from the liquid and to maintain the operating temperature level by a combination of modules and heat-exchangers in series. An alternative is the sweeping of the permeate by a partially condensing, and with respect to the permeate immiscible, carrier vapor. Since the latent heat of the carrier vapor can be utilized only partially, this concept will not be economically competitive. Pervaporation has a wide range of possible applicatio for this reasons it is impossible to discuss the economics of pervaporation in general. In this paper the separation of benzene/cyclohexane, an azeotropic system with similar vapor pressures of the components, has been chosen as an example. But even such a limited discussion reveals tendencies which seem to be generally valid: 1.Pervaporation processes consisting of several stages (cascade) cannot compete with conventional separation processes like extractive distillation. 2. Hybrid processes like a combination of extractive distillation, pervaporation are very promising, especially in cases where high product purities required

Treatment of landfill gas by gas permeation — pilot plant results and comparison to alternatives

Landfill gas is an interesting source of prime energy since it contains about 50–55% volume methane. Utilization, however, must take into account the traces of halogenated hydrocarbons and H2S, which are present in the gas. The paper reports of two years experience with a pilot plant for the production of a gas with natural gas quality. Essentially the process consists of two stages, an adsorption for the separation of the trace components and a membrane unit for the separation of CO2. The process is discussed in detail and compared with alternatives for the utilization of landfill gas such as power generation by gas engines/generator, eventually in combination with a waste heat boiler.

Design and optimization of combined pervaporation/distillation processes for the production of MTBE

The purification of product streams in the production of MTBE requires complex processes due to the thermodynamic behaviour of the mixtures (i.e. formation of azeotropes). Combined processes consisting of distillation and pervaporation/vapour permeation might offer economically attractive alternatives as they can simplify the process structure, reduce the energy consumption and avoid entrainers. The paper presents possible process configurations and design strategies for an integration of pervaporation and vapour permeation into the Huels process. For all process calculations, the commercial simulation software ASPEN PLUS was used in combination with a compatible FORTRAN routine allowing the design and simulation of complete pervaporation units in ASPEN PLUS.

Waste water treatment by a combination of bioreactor and nanofiltration

A new hybrid process has been developed and tested for the treatment of dumpsite leachate, a highly contaminated multicomponent waste water. The process consists of an activated sludge bioreactor, a nanofiltration stage and chemical oxidation or adsorption. The essential feature of the process is the recycling of the nanofiltration concentrate into the bioreactor. Due to the selectivity of nanofiltration membranes, the concentrations especially of recalcitrants in the loop are significantly increased, which in turn results in a higher rate of biodegradation without increasing the hydraulic residence time. Non-biodegradable components are removed from the system via nanofiltration permeate, excess sludge or chemical oxidation or adsorption on activated carbon. Compared to a straightforward process bioreactor/ chemical oxidation or bioreactor/adsorption, the specific consumption of oxidant or activated carbon of the new process is significantly lower. Pilot plant tests on four dumpsites, each over a period of several months, demonstrated the reliability of the process and confirmed our expectations with respect to specific consumptions. Observation of the nitrification, which is very sensitive to toxic components, confirmed that the biology was not affected by the substantially increased levels of rejected components of the nanofiltration. Compared to the straightforward process, the elimination rate of the bioreactor — ζ= 1 -(CODpermeate)/CODleachate) — was increased between 9% and 17%. We are convinced that this new process is not confined to the treatment of leachate but is applicable for a variety of industrial waste waters.

Investigation of mass transport in asymmetric pervaporation membranes

Abstract Selectivity and flux of pervaporation are determined by permeate pressure, feed temperature and feed composition. The influence of these parameters on process design will be discussed, using the separation of water-isopropanol by composite membranes as an example. Permeate flux increases exponentially with increasing temperature while the selectivity is almost independent of temperature. This result corroborates the sorption-diffusion model and explains why process temperature should be as high as possible. The most important parameter, however, is the local feed composition, especially in the region of technical importance (wF, H2O (0.2). Flux varies almost linearly with feed composition, decreasing to very low figures for high product purities (= low water concentration in the feed). Selectivity varies strongly in this region, and for this reason small errors in the measurement of the separation characteristics will lead to a totally erroneous calculation of the necessary membrane surface area. Two different model calculations will be presented and compared: (i) a module design based on the assumption of “ideal pervaporation”, and (ii) a module design based on the measurement of the separation characteristics and the flux-concentration relationship. In both cases, of course, the change of conditions (temperature, pressure, composition) along the module will be taken into account.

Some considerations on mass-transfer resistances in solution—diffusion-type membrane processes☆

Abstract The influence of feed-side concentration polarization and the resistance of the porous support on membrane processes where solution—diffusion-type composite membranes are involved is discussed in this paper. Reverse osmosis, nanofiltration, gas permeation, pervaporation and vapor permeation represent the complete variation of liquid and gaseous state at the feed and permeate side and cover cases where the driving force is realized by increasing the feed-side potential and cases where the permeateside potential is decreased. In the calculations, the major parameters, especially feed-side concentration of the preferentially permeating component is varied over a wide range - from “traces” to “fairly high”. The discussion of the results shows clearly the importance or unimportance of the resistances relative to each other for every process. Finally, the effect of transport resistances on module performance is discussed.

Ultrafiltration of macromolecular solutions and cross-flow microfiltration of colloidal suspensions. A contribution to permeate flux calculations

Abstract Ultrafiltration and cross-flow microfiltration are pressure-driven membrane processes which exhibit concentration polarisation in front of the membrane or gel layer, respectively. In ultrafiltration the concentration polarization is essentially determined by diffusion in the boundary layer. With the usual assumption of constant material properties, the film theory predicts fluxes quite accurately in the case of ultrafiltration of low molecular weight solutions. The calculated fluxes are, however, significantly lower than the observed figures in case of ultrafiltration of macromolecular solutions. It will be shown that with concentration-dependent material properties, especially a variable density (!) and a variable diffusion coefficient, the film theory can predict fluxes of ultrafiltration of macromolecular solutions very accurately. Such a theory fails, however, in case of cross-flow filtration of colloidal suspensions. Even with the assumption of a variable density in the boundary layer, the calculated fluxes are too low by a factor of about 10. For cross-flow microfiltration of colloidal suspensions, a model is presented considering basically the membrane-parallel drag forces acting on a particle in the vicinity of the membrane and a “friction” force as a consequence of the lateral drag forces. The model has been tested with different module systems and with different colloidal suspensions.