Main lectures

PHENOMENOLOGICAL MODELS OF TRANSPORT AND SEPARATION IN MEMBRANES

BOHUMIL BERNAUER

Institute of Chemical Technology, Prague, Technicka 5, 166 28 Prague 6

e-mail: bohumil.bernauer/vscht.cz

1. Introduction

The significant variety of existing membrane operations is based on relatively simple, compact, and largely clarified fundamental transport mechanisms characterizing mass transfer in the dense or porous membrane layer and at the membrane - fluid interfaces. The understanding and prediction of transport phenomena in the membrane structure is now at least qualitatively possible, also theoretically through the newly available tools provided by molecular simulation, irreversible thermodynamics or relaible phenomenological models.

Prediction of the separation performance of membrane materials requires knowledge on multicomponent diffusion, adsorption parameters and structure parameters. These parameters should be determined by experimental methods and adequate modeling of transport phenomena.

2. Gas transport in simple membrane structures

The understanding and modeling of the diffusion process inside the micro, meso or macro pores of membranes poses many challenges for the experimentalist and theoretician alike. The proper description of diffusive transport within membranes is of considerable importance because of the considerable potential importance in chemical technology and new separation processes. A variety of models and techniques have been used to describe diffusion within membranes, e.g. molecular dynamics and Monte Carlo simulation, phenomenological models using Fick's law or Maxwell-Stefan formulation.

3. Gas permeation in macroporous and mesoporous systems

In the Wicke-Kallenbach experiments transport in the support layer can occur by molecular and Knudsen diffusion and by convective (viscous) flow mechanism. Knudsen diffusion can be neglected in the support layer if the Knudsen number (Kn) is significantly smaller than 1. Knudsen number is the ratio between the mean free path of the molecule () and the pore diameter ():

It is therefore assumed that permeation through the support layer can be described by dusty-gas model to compute concentration field in the support layer. Dusty gas model provides the constitutive flux equations in the following implicit form:

By eliminating the total pressure gradient from equation (3) using equation (4) we obtain the relation between the gradient of partial pressure of i-th compound and the fluxes:

The net flux density N_i through the support are the sum of molecular, Knudsen and viscous flux contributions

METHOD OF SOLUTION

System of ODE with boundary conditions converted to the system of algebraic nonlinear equation by

4. Experimental determination of geometric parameters of membranes

MODEL PARAMETERS

5. Diffusion within micropores, capillary condensation, multilayer diffusion

Generalised Maxwell-Stefan form

Transport in microporous crystal is depicted as a sequence of five steps, involving intracrystalline and interfacial processes. According to Barrer [] transport through porous crystal can be described by following consecutive steps:

• Adsorption on the external surface of crystal
• Transport from external surface into the pores
• Intracrystalline transport
• Transport out of the pores to the external surface
• Desorption from the external surface to the surrounding gas phase

These are all activated steps, which can be modeled assuming that molecule jumps between low-energy (low-potential) sites. The operating conditions (temperature, concentration of species) and the characteristics of the molecules and the microporous crystalline material determine which of these steps is rate determining. Steps A, B, D, and E are referred to as interfacial processes. Interfacial effects are important for thin membranes and depend on the difference between the activation energy for intracrystalline diffusion. Adsorption on the external surface is usually less strong than adsorption within the crystal, so it is unlikely that steps A or E are to be a rate limiting ones.

The generalized Maxwell-Stefan (GMS) constitutive relations have been successfully applied to many systems to describe diffusive transport phenomena in multicomponent mixtures. The GMS model is based on the momentum transfer between particles in motion. The driving force exerted on any molecule i is balanced by the friction this species experiences with all other species present in the mixture. The essential concept of the theory of diffusion in multicomponent mixture was already available more than a century ago following the pioneering works of James C. Maxwell and Joseph Stefan.

The Generalized Maxwell-Stefan equations provide an adequate basis for the description of multi-component mass transfer in microporous media. The basis of this theory is that the driving force for movement that is acting on a species if balanced by the "friction" with other molecules and vacant adsorption sites that is experienced by that species. We can use the dusty gas model approach to the description of diffusion in adsorbed phase by considering the vacant sites to be the (N+1)th pseudo-species in the adsorbed mixture:

where , the chemical potential gradient of the i-th adsorbed species, is the driving force exerting on this adsorbed molecule tending to move it the adsorbed phase. This driving force is balanced by interactions between i-th and j-th adsorbed molecules and the friction experienced by the species i from vacant adsorption sites. The q_i represents the surface occupancy of the adsorbed species i and q_N+1 represents the fraction of vacant adsorption sites.

Illustrative examples of permeation and separation with microporous membranes

Fig. 1 Computed dimensionless transient profile of the n-C₄H₁₀

A)	B)
1: Transient profiles of dimensionless concentrations of iso-C4H10 and H2 at the outlet of the tube (,) and the annulus(,) at different temperatures - A) 50 °C, B) 200 °C Mixture 5 vol.% of hydrogen and 5 vol.% of iso-butane in nitrogen, the inlet volumetric rates are 40 ml/min.

Catalytic membrane reactors

concepts, materials and applications

J-A. Dalmon

Institut de Recherches sur la Catalyse, CNRS - 2, av. A. Einstein 69626 -Villeurbanne France

Jean-Alain.Dalmon/catalyse.cnrs.fr

Introduction

Catalytic Membrane Reactors (CMRs) are reactors combining in a same unit a catalyst, providing conversion and a membrane that controls transfers to or from the catalyst. In a CMR, the presence of the membrane should improve the performance of the catalyst; in other words, CMRs are not only a sequence of two independent units.

For 20 years, numerous lab-scale studies have been developed to explore the capabilities of membrane and catalyst combinations and regular international meetings gathering specialists of membrane science, catalysis and chemical engineering are organised.

The membrane role, initially acting as a separation barrier, has been extended to other functions, such as reactant contactor. Up to now, most of the applications have considered heterogeneous catalysts and inorganic ceramic membranes. Promising results have been obtained at the laboratory level and few developments scaled up to the pilot level.

Some reviews, books and special issues can provide a complete literature survey of membrane catalysis [1-3].

In what follows, the various types of CMRs will be first described. In a second section, materials (both membrane and catalyst) used in CMR will be introduced. The third part will briefly present examples from our group for each CMR type. They will be detailed, together with other results from the literature, during the oral presentation.

A classification of membrane reactors

There are numerous ways of combining a catalyst and a membrane. Moreover, depending on the membrane role, the catalyst may be placed in a very different situation. All this makes a global analysis almost impossible. A classification of CMRs has been proposed according to the way of combining catalyst and membrane [2]. The main criterion is based on the presence in the CMR of a catalytic membrane (the same material acts as catalyst and membrane) or of an association of a conventional catalyst (packed or fluidized bed) and a membrane that just rules transfers.

Another classification is based on the membrane role [4]. As a matter of fact, as schematized in figure 1, the membrane can have 3 very different functions according to the CMR type:

(i) The membrane can be used to remove a reaction product from the reaction zone. This type of CMR, called extractor, is certainly the most studied. An extractor can be used to increase reaction yields. This is obtained either by improving the conversion in equilibrium-restricted reactions or, in consecutive reactions, by improving the selectivity towards a primary product via its selective extraction through the membrane.

In extractor CMR applications, the membrane permeation should be highly selective towards one product to get a performance improvement: any loss of another species involved in the reaction will be highly detrimental.

(ii) The membrane can control the introduction of one of the reactants in the reaction zone. This type of CMR, called distributor, is used to spread a reactant all along the catalytic zone, in which the other reactant is introduced as usually. In this way, when compared to conventional reactors, though the same (or even larger) amounts of the distributed reactant can be introduced, its concentration is kept at a low level in the entire reaction zone. This low concentration may increase the selectivity of reactions when the distributed reactant can undergo successive additions. Selective oxidations (or hydrogenations) can be improved in

Figure 1. Sketch of the different CMRs

distributor-type CMR if a membrane distributes oxygen (hydrogen). Another advantage of such reactor is related to flammable mixtures. Due to the low O₂ concentration in the catalyst bed, the local reactants composition can be kept outside the flammability region, though the total amount of reactants introduced corresponds to a ratio that is forbidden in conventional reactors.

In distributor-type CMR, the membrane acts as a diffusion barrier for one reactant. It should avoid back permeation of other species. More than the membrane pore characteristics, the operative conditions are crucial to control the transfers through the membrane.

(iii) A membrane may be use to facilitate the contact between reactants and catalyst. This CMR type, called contactor, takes advantage of the very unique configuration of the membrane pore, which contrary to pores of conventional solids, presents two distinct ways of access, corresponding to the two sides of the membrane. In contactors, the membrane generally also acts as catalyst support (or is intrinsically active). A contactor can be used following two modes (figure 1). In the interfacial contactor mode, the reactants are separately introduced from each side of the membrane, and meet in the catalyst zone. Such configuration has been used in the case of non-miscible reactants, as in gas-liquid catalytic reactions. It has been pointed out that, contrary to what is generally observed in conventional reactors, the gaseous reactant is no more limiting. Dense polymeric membranes containing metal catalyst encaged in zeolites have also been used as interfacial contactors in the case of non-miscible aqueous and organic reactants. In this case, the polymeric membrane favours the transfer of the organic reactant to the active phase, which allows the contactor to perform better than conventional reactors. The other contactor mode is the flow-through contactor. In this second mode, the mixture of reactants is forced through the membrane, i.e. through the catalytic pores. Contact time and permeation regime in the active pore itself can be directly adjusted from the operative conditions and adapted to required values, which is hardly feasible in conventional reactors. Such reactors have been used for gas and gas-liquid reactions, and showed improvements in activity and selectivity.

CMR Materials

Membranes

Most of the first CMR studies were based on catalytic reactions requiring temperatures above those that polymeric membranes could withstand. This led to focus on inorganic membranes. Both dense and porous inorganic materials have been used. The membrane choice highly depends on the CMR type and application.

Dense inorganic membranes.

As dehydrogenation is one of frequently studied reactions, dense palladium membranes, selective for hydrogen permeation, have been used in extractor-CMR to improve the reaction yield by selective H₂ removal from the reactor. Most of these membranes are obtained by electroless plating techniques. The film is formed on top of a porous ceramic or metallic support. Pd is often combined with Ag to improve the resistance of the material at low temperature.

Beside metals, dense oxides have been also considered for membrane production, mainly for selective oxygen permeation in oxidation reactions. Such materials received an important support via several international consortia gathering leading industrial partners. Mixed conducting (ionic and electronic) oxide membranes, such as perovskites, were prepared using salts as precursors, by calcination at high temperature and sintering the powders under high pressures. Selective O₂ permetion occurs at high temperature (800°C). The main application was the separation of oxygen from air. The O₂ production is an economical key step in the syngas production from methane partial oxidation. As this reaction also occurs at high temperature membrane and catalyst are associated in a same unit. This is not stricto sensu a CMR, though it is has been proposed that the catalyst can take benefit of the presence of highly reactive O species at the membrane surface and that the absence of gaseous oxygen limits deep CH₄ oxidation and flammability problems.

Porous inorganic membranes.

If dense membranes are highly selective, they often suffer from low permeances and problems for large-scale production. Porous membranes are, in principle, less selective, but generally provide enough permeating fluxes to be combined with catalytic performances.

Depending on the application, microporous or mesoporous layers have been used. When high separation selectivity is required (extractor CMRs), microporous membranes may compete with dense membranes. Such microporous membranes can be obtained either by sol-gel techniques (as mesoporous membranes) or using zeolite application on a macroporous ceramic (or metallic) support. Microporous membranes can be used as molecular sieving systems, blocking permeation of species whose diameter is larger than the pore size. However, especially in the case of zeolite membranes, adsorption properties very often rule the permeation selectivity. In a mixture made of a small-size non-adsorbing molecule with a larger adsorbing species, the zeolite membrane separation can be in favour of the larger molecule that occupies all the microporous volume, thus blocking permeation of the smaller species.

Mesoporous inorganic membranes generally do not provide enough separation selectivity (based on Knudsen diffusion) for gas-phase catalytic reactions. They are mainly used in distributor or contactor CMR modes where separation selectivity is generally not the leading membrane effect.

Polymeric membranes

As above-mentioned, examples of CMRs using organic membranes are less numerous than those using inorganic materials. However, in principle, all the modes represented in Fig. 1 can be implemented with polymeric membranes. Extractor and contactor CMRs have been considered in low temperature catalytic reactions.

Catalysts

When compared to membranes, catalysts for CMR received much less attention. However, the specific conditions prevailing in membrane reactors may modify the catalyst state and properties in comparison to those in conventional systems [5]. In some cases, adapted catalysts have to be designed to get higher performance in CMRs.

CMR case studies

Extractor

Isobutanedehydrogenation [6]

The isobutane dehydrogenation was studied in two extractor-type CMRs. Combinations of a Pt-based catalyst with two different membranes, a Pd membrane obtained by electroless plating and a MFI zeolite membrane obtained by hydrothermal synthesis, were compared. Both CMRs give better results than conventional reactors. However, though the two membranes presented different separative properties, the two CMRs showed very similar yields in the isobutane dehydrogenation. This has been attributed to the limitation of both CMR performance by the catalyst lack of efficiency. A modeling approach, that combines catalysis kinetic law for and membrane gas transfer equations also contributes to the description of the CMRs performance.

Xylene isomerization [7]

A zeolite/alumina membrane was used to successfully separate xylene isomers. It was then applied, as a selective membrane, in an extractor type catalytic membrane reactor (CMR), used to enhance the xylene isomerization reaction selectivity towards para-xylene. The results of the CMR in different configurations (permeate-only and combined permeate-and-retentate mode) were compared to conventional fixed-bed reactor results. In both cases, the selectivity was significantly enhanced (up to 100% in permeate-only mode). In the combined mode, the CMR also provided a net increase in productivity over the conventional reactor.

Distributor

n-butane selective oxidation [8]

The n-butane selective oxidation has been studied in a membrane reactor, using high butane concentrations. Thanks to the oxygen distribution by the membrane, it is possible to keep the local composition outside the flammability zone. A MFI ceramic membrane was used to distribute oxygen (or part of it) in the catalyst bed, made of a VPO mixed oxide, either conventional or Co-doped. In a first step, the effect of the oxygen distribution has been studied, showing that, under standard reactant mixtures (O₂/butane=12, low butane concentration), the membrane reactor performed very close to the conventional one. Under high butane concentrations, the VPO system suffered from a drastic decrease of the selectivity towards maleic anhydride (MA). The addition of cobalt to the VPO catalyst allowed keeping the MA selectivity at a high level (75%). The combination of the CoVPO catalyst and the MFI membrane was used to explore the membrane reactor performance with high butane concentrations in the feed, corresponding to the flammability zone in a conventional reactor. For these conditions, the MA productivity was 3 times higher than that observed with the conventional reactor.

Contactor

Wet Air Oxidation process [9]

A new and innovative method for oxidation of dissolved compounds in water - the "Watercatox" process - has been developed in order to reduce the chemical oxygen demand and the total organic carbon in industrial wastewaters. This process is the result of a European Fifth Framework Program project. It can operate at much lower temperatures and pressures than conventional wet air oxidation or incineration, and it offers much smaller volume requirements than biological treatment plants. The operating principle of the Watercatox process is the oxidation of the dissolved molecules using oxygen from air within a catalytic membrane reactor in an interfacial contactor configuration. The catalytic contactor membranes, as well as the operating conditions, have been up-scaled from lab-scale to pilot unit. The technological efficiency was demonstrated by the results obtained using the pilot test unit on different industrial effluents from several origins.

Conclusion

Thanks to a large research effort at a laboratory level, important progresses have been made in CMRs on the basis of various catalyst/membrane combinations. Applications cover a large domain, from petrochemistry to fine chemicals or environmental issue remediation. However, in some cases, membrane costs could be a limiting factor and long duration experiments are still needed.

Though CMRs presented here are, for the time being, very far to reach the industrial development of membrane bioreactors, first CMR pilots have been recently developed.

References

[1] J-A. Dalmon, "Handbook of Heterogeneous Catalysis", VCH Publ. (1997), 1387

[2] J. Sanchez, T.T. Tsotsis "Catalytic Membranes & Catalytic Membrane Reactors", Wiley (2002).

[3] Catalysis Today, special issues on membrane catalysis, 25,3-4 (1995); 56,1-3 (2000); 67,1-3 (2001); 82,1-4 (2003); 104,2-4 (2005)

[4] S. Mota, S. Miachon, J.-C. Volta, and J-A. Dalmon, Catalysis Today, 67, (2001), 169.

[5] S. Miachon and J-A. Dalmon, Topics in Catalysis, 29 1-2 (2004) 59

[6] L. van Dyk, S. Miachon, L. Lorenzen, M. Torres, K. Fiaty, J-A. Dalmon, Cat. Tod, 82, (1-4) (2003), 167.

[7] L. van Dyk, L. Lorenzen, S. Miachon, J-A. Dalmon, Cat. Today, 104(2-4) (2005), 274.

[8] A. Cruz-López, N. Guilhaume, S. Miachon and J-A. Dalmon, Catalysis Today, 107-108, 949.

[9] E. E. Iojoiu, J. C. Walmsley, H. Raeder, S. Miachon, J-A. Dalmon, Cat. Tod., 104(2-4) (2005), 329.

COMBINING MEMBRANE SEPARATION AND FERMENTATION PROCESSES FOR IMPROVED PERFORMANCE

G. JONSSON

Department of Chemical Engineering, Technical University of Denmark, Building 229 DTU, DK-2800 Kgs. Lyngby, Denmark, (gj/kt.dtu.dk)

Introduction

Fermentation processes are often inhibited by either products or biproducts produced by the microorganisms thereby limiting the maximum product concentration which can be obtained during the fermentation. Also it will influence the maximum biomass concentration which can be obtained thereby decreasing the production rate. A typical batch fermentation as sketched in Fig. 1 consist of a lag phase where the microorganisms "wake up" followed by an exponential growth phase which goes into a stationary phase depending on the inhibiting effect of the different products. Removing these products directly from the fermentation by a membrane separation process makes it possible to operate at much higher biomass concentrations thereby increasing the production rate as well as the final product concentration which might have a beneficial influence on the further downstream processing. Depending on the type of products or inhibiting biproducts such membrane processes can be either ultra/microfiltration (size related separations), membrane distillation (volatile products like bioethanol) or electro membrane processes (charged products like organic acids or bases). Combining fermentation directly with membrane separation will further make it possible to operate a high cell density in even a continuous fermentation process which is today used on large scale in waste water treatment plants using membrane bioreactors (MBR) where the membrane plant can either be internal (submerged membranes) or external (recycle loop) to the biological waste water treatment plant.

Figure 1. Biomass growth rate in a typical batch fermentation.

The lecture will discuss the different advantages and problems related to using each of the three membrane separation processes: 1) ultra/microfiltration, 2) membrane distillation and 3) electro membrane processes directly coupled to fermentation processes. The main focus will however be on a newly developed electro membrane process named reverse electro-enhanced dialysis (REED):

REED Technology

The Reverse Electro-Enhanced Dialysis (REED) setup and design combines elements from electrodialysis reversal (EDR) and Donnan dialysis (DD) operations.

The REED unit improves fermenter productivity by continuously removing inhibiting organic acids and replacing them by alkaline hydroxide ions. This removes the acids, which would otherwise inhibit the growth of the microorganisms, as well as provides a pH regulation of the fermenter. Fermentation broth is continuously taken from the fermenter and pumped into the REED unit and then back to the fermenter.

The REED unit is a plate-and-frame module as demonstrated below. Ion-exchange membranes are placed in layers, separated by flow spacers, which facilitate the passage of the fermentation broth between each set of membranes. Inside the REED unit, the broth is diverted into every second flow spacer, where it streams along the surfaces of the two ion-exchange membranes on either side of the spacer. Inside the alternating flow spacers, an alkaline stream (the Dialysate) flows without direct contact to the broth. The ion-exchange membranes are positively charged, which allows the negatively charged ions to be exchanged through the membranes. Positively charged ions are excluded from passing through the membranes because they hold similar charges.

Figure 2. Set-up of the REED module.

The ion transport is carried by electrical current, which is added across the module from electrodes placed at each end of the stack. Since the concentration of organic acids is significantly higher than that of hydroxide ions in the broth, the current is mainly carried by the organics acids, which are transported to and through the ion-exchange membranes. In the alkaline solution, the concentration of alkaline hydroxide ions is much higher than that of the organic acids, which enters the solution through the membranes. Thus, in the alkaline solution the electrical current is mainly carried by hydroxide ions moving through the ion-exchange membranes and into the fermentation broth. This overall exchange of negative ions preserves electro-neutrality.

Bio-matter in the form of microorganisms, yeast extracts and other organic particular components as well as proteins and other high-molecular components are too bulky to pass through the dense ion-exchange membranes. But the bio-matter and proteins tend to foul the membrane surfaces, causing a build-up of organic matter on the membranes inside the flow spacers with fermentation broth. This membrane fouling causes a steady increase in the electrical resistance of the membrane stack. All the negative ions migrate in the same direction, towards the positive electrode (anode). Inside each flow spacer carrying fermentation broth, organic acids are leaving to one side while hydroxide ions are entering from the opposite site, as demonstrated in the process setup sketch above. The hydroxide ions entering the flow spacer through the ion exchange membranes destabilize any build-up of fouling on that side, which they enter. On the other side of the flow spacer, fouling builds up, only reduced by the shear stress from the bulk flow in the spacer along the membrane surface, which drags some of the top fouling layer off. The symmetrical setup of the REED system allows the direction of the electrical current to be completely reversed while the overall separation is continued. The principle is sketched below.

Figure 3. Sketch of the effect of the current reversal on membrane fouling.

In figure 3A, the negative ions migrate towards the positive electrode (anode), meaning that hydroxide ions enters the flow spacers with fermentation broth from the left, and fouling builds up on the right membrane surface. When the direction of the electrical current is reversed (figure 3B), hydroxide ions enter the flow spacer from the right, destabilizing and removing the fouling layer, while a new fouling layer builds up on the left membrane surface. The current is reversed at regular intervals of 5-10 minutes. The interval is determined by the build-up of fouling; when the build-up is severe, short periods are necessary. Each reversal reduces the overall separation efficiency due to the change in concentration gradients inside the membranes and the intervals should be kept as long as possible. The alkaline solution, which collects the organic acids in the form of acid salts (e.g. Sodium Lactate) is either discarded or recycled through the REED unit until sufficient amounts of lactate is collected in the solution. By continuously adding fresh alkaline solution to the REED system, the efficiency is high, but the amount of alkaline solution which must be discarded, is also high. Improving the utilization of the alkaline solution reduces the process efficiency and demands higher energy consumption and membrane area, so careful evaluation of these parameters must be taken into account when considering the system design.

Case study

The ability of the reverse electro-enhanced dialysis (REED) separation system to extract growth inhibiting lactic acid during live lactic acid bacteria fermentations has proven significant boosts in productivity and product yield for a Gram-positive expression system using Lactococcus lactis as a host organism for production of recombinant proteins for use in humans [1].

Lc. lactisproduces lactic acid from sugars as integral part of its primary energy metabolism, and if the cells are provided with sufficient glucose, growth as well as protein production will be limited by accumulation of lactic acid in the fermentation broth. Inhibition of growth and protein production can be delayed, but not prevented by titration of the lactic acid with base. Lactic acid and lactate also serve to induce production with the P170 system promoter; again the inducing concentration will be higher if pH is kept high with base addition. At a given pH value, the optimal inducing level of lactate will be severely but not absolutely growth inhibiting.

The patented electro-membrane separation process, REED, provides a technology for removing lactate from fermentation broth, while retaining cells, proteins, sugars and most other feed components. When the process is controlled to keep a constant pH in the fermenter, the amount of lactate in the broth will also be kept constant. This means that the level of growth inhibition and of expression system induction can be controlled during fermentation.

The same principles used for the P170 expression system in this abstract can be transferred to other fermentation and expression systems whether mammalian or bacterial.

REED fermentations

For REED fermentations, a fed-batch fermenter is fitted with an external loop, which passes through a membrane stack (Figure 4), where lactate is selectively extracted, before the acid-depleted ferment is recirculated to the fermenter.

Figure 4. REED fed-batch fermentation setup.

The fed-batch fermentation is operated in three steps. First, the fermentation runs through its lag phase and enters the exponential growth phase. When a certain non-inhibiting concentration of lactate is reached, the REED system takes over pH-control of the fermenter by continuously extracting newly produced lactate, replacing lactate ions with hydroxide ions, and thus maintaining constant fermenter pH. Through this method, it is possible to obtain significantly higher biomass owing to the uninhibited growth compared to standard fermentations. A concentrated solution of fresh substrate and other vital growth components are fed to the fermenter to supplement the growth.

When sufficient biomass has been produced, the fermenter's lactate concentration is increased through addition of sodium lactate to reach a new level, which triggers the expression system, but still remains below the inhibition level for lactate. At this third fermentation phase, the increased biomass expresses the pharmaceutical protein at a significantly higher rate compared to what can be obtained in similar batch fermentations. Since freshly produced lactate is continuously removed to maintain the desired lactate level in the fermenter, the production phase can be prolonged significantly resulting in yield improvement.

Results

Laboratory experiments were successfully carried out for different strains with similar results. When a standard batch would normally stop production due to inhibition, the cells in the REED fed-batch fermentation continued to grow and express the protein. Figure 5 shows the difference in biomass (OD₆₀₀) and produced recombinant protein between standard batch production and REED fed-batch fermentation. The time of induction, where lactate is added to the fermenter to promote protein expression is shown in figure 5.

Figure 5. Comparison between biomass output (OD) and produced protein for REED fed-batch fermentation versus standard batch fermentation.

Conclusion

The results show that the REED process is able to boost both productivity and product yield dramatically for the P170 expression system, and makes it feasible to produce even normally low-yield pharmaceutical components.

[1] L. Bredmose, S.M. Madsen, A. Vrang, P. Ravn, M.G. Johnson, J. Glenting, J. Arnau and H. Israelsen, "Development of a Heterologous Gene Expression System for use in Lactococcus lactis", O.-W. Merten, et al. (eds.), "Recombinant Protein Production with Prokaryotic and Eukaryotic Cells", p. 269-275, Kluwer Academic Publishers (Netherlands) 2001.

[2] A. Garde, Ph.D. Thesis: Production of lactic acid from renewable resources using electrodialysis for product recovery, 2002, ISBN 87-90142-84-5

MEMBRANES FOR FUEL CELL AND HYDROGEN TECHNOLOGY

SUZANA PEREIRA NUNES

GKSS Research Centre, Max-Planck-Str. 1, 21502 Geesthacht, Germany, e-mail nunes/gkss.de

The need for reducing the worldwide CO₂ emission and searching for independence of fossil fuels has increased the motivation to establish fuel cell and hydrogen technology in all sectors of energy. In this task there are important challenges for membranes. Examples are:

- H₂ separation in petrochemical streams

- CO₂ sequestration

- Membrane reactors involving H₂

- Fuel cells, electrolysers

Hydrogen is still being produced in large scale by the petrochemical industry. Although the hydrogen technology is expected to be very important in the future, the world will still be using fossil energy in the next decades. It is decisive to make all the processes involving fossil fuel as clean as possible. Here membranes can play a decisive role in the hydrogen separation and processing and CO₂ sequestration. Many of the tasks for hydrogen separation in the petrochemical industry take place at quite high temperature, making inorganic membranes in these cases the only choice. Among the most investigated inorganic membranes for hydrogen separation are silica membranes based on molecular sieve mechanism and palladium membranes based on the strong interaction of hydrogen and palladium, which makes them extremely selective. More recent a new class of mixed electron and proton conductive oxides has emerged allowing a very selective of hydrogen through the membranes. However for the temperature range below 250^oC inorganic membranes can hardly compete with polymer membranes, which are cheaper and easy to manufacture in large scale. Asymmetric polymeric membranes (Figure 1) for hydrogen separation based on Matrimid^®, a commercial polyimide, have been developed for manufactured in technical machines at GKSS. The H₂/CH₄ selectivity is ca. 100 and CO₂/CH₄ selectivity 39 with gas permeability of 1 m³(STP)/m² h bar [1].

Membranes for fuel cell

In the low temperature fuel cell a polyelectrolyte membrane is a core component to transport protons from the anode to the cathode generating electricity (Figure 2). The requirements for a good membrane are: high chemical stability, high proton conductivity, low fuel and oxygen permeabilities, high electronic resistance and preferentially low cost. The membrane surfaces are in direct contact to a finely dispersed catalyst mainly on Pt base and the electrodes. The main commercial membrane for fuel cells is based on Nafion^Ò, a fluorinated polymer with sulfonated side chains. However well known drawbacks are the high costs, and mainly two technical aspects:

(1) for portable application, when using methanol as primary fuel, Nafion^Ò is too permeable, allowing methanol to migrate to the cathode and decreasing the cell performance;

(2) for automotive application, for which the preferential operation is considered to be above 100^oC and at low humidity, the membrane looses conductivity.

To overcome these problems many alternative polymeric materials are under investigation. Most of them are sulfonated polymers. The preparation varies from post sulfonation of commercial polymers to grafting of sulfonic segments in a polymer membrane by gamma [3] or electron [4] radiation and the polymerization from sulfonated monomers. Fluorinated polymers are usually more chemically stable, however more expensive and more difficult to synthesize. They have been reviewed recently [5].

A large number of non-fluorinated polymers are now under investigation. Some of these activities have been reviewed in recent publications [2, 6]. Approaches include the sulfonation of commercial polymers and the polymerization of new functionalized polymers. One of the first polymers chosen for sulfonation was polysulfone. More recently the synthesis of stable polysulfones from sulfonated monomers has been explored.

The investigation of different variants of sulfonated polyetherketones has been widely described in the literature: (polyetherketone, poly(ether ether ketone) [7] , poly(ether ketone ketone) [8] and poly(ether ether ketone ketone) (sPEEKK) [9], poly(oxa-pphenylene-3,3-phtalido-p-phenylene-oxa-p-phenileneoxy-phenylene) (PEEK-WC) [10]. The comparison between different structures of polyetherketones prepared by polycondensation of sulfonated monomers has been recently reported [11].

Polyphosphazene sulfonic acids [12] offer a unique combination of inorganic backbone with high stability, high ionic density and structural diversity with cross-linking alternatives. Polyphosphazenes with sulphonamide functionalisation have been reported with high proton conductivity [13]

A big challenge for fuel cell membranes is a good performance at temperatures higher than 100^oC. Reviews on materials under investigation to overcome this problem have been published by Alberti and Casciola [14], Q. Li et al. [15] and Hogarth et al. [16].

Polybenzimidazole (PBI) is one of the few polymers under consideration for high temperature operation. For that the membrane was immersed in concentrated phosphoric acid to reach the needed proton conductivity. Operation up to 200^oC is reported [17]. A disadvantage of this class of membranes is the acid leaching out during operation, particularly problematic for cells directly fed with liquid fuels. Additionally the phosphoric acid may adsorb on platinum surface.

The combination of organic and inorganic materials to develop membranes for fuel cell has become a versatile approach. Early reports and patent applications of Stonehart and Watanabe [18] and Antonucci and Arico [19] claim the advantage of the introduction of small amounts of silica particles to Nafion^® to increase the retention of water and improve the membrane performance above 100^oC. The effect is believed to be a result of the water adsorption on the oxide surface. As a consequence the back-diffusion of the cathode-produced water is enhanced and the water electro-osmotic drag from anode to cathode is reduced.

Another important motivation for the development of organic-inorganic membranes for fuel cell is the reduction of methanol and water permeability, which are highly relevant aspects for the establishment of the DMFC technology. The inorganic phase can be included in different ways, starting from the simple dispersion of isotropic particles (SiO₂, ZrO₂, etc), introduction of fillers with high aspect like layered silicates and the in situ generation of an oxide phase in the polymeric matrix by dispersion of inorganic precursors followed by their hydrolysis and polycondensation [20, 21]. The reduction of methanol and water permeability of sulfonated poly (ether ether ketone membranes) with the generation of ZrO₂ and SiO₂ in the membrane casting solution was reported in different papers [22]. These fillers are isotropic. Inorganic fillers like layered silicates, with high aspect ratios, are expected to lead to a more effective permeability reduction in membranes.

Silica and silicates are usually rather passive fillers. Active fillers able to contribute also for the proton conductivity include zirconium [14, 23] and boron phosphates [24] and heteropolyacids [25].

References

1. S. Shishatskiy, C. Nistor, M. Popa, S. P. Nunes, K. V. Peinemann. Advanced Engineering Materials 8 (2006) 390-397.
2. S. P. Nunes and K. V. Peinemann. Membrane Technology in the Chemical Industry, Wiley-VCH, 2^nd ed., 2006.
3. B. Gupta, F. N. Buchi, G. G. Scherer, Solid State Ionics 61 (1993) 213-218.
4. S. D. Flint and C. T. Slade. Solid State Ionics 97 (1997) 299-307.
5. G. Kostova, B. Ameduri, B. Boutevin. Journal of Fluorine Chemistry 114 (2002) 171-176.
6. M. A. Hickner, H. Ghassemi, Y. S. Kim, B. R. Einsla, and J. E. McGrath, Chemical Reviews 104 (2004) 4587-4612.
7. S. Kaliaguine, S.D. Mikhailenko, K.P. Wang, P. Xing, G. Robertson, M. Guiver, Catalysis Today 82 (2003) 213-222.
8. S. Vetter, B. Ruffmann, I. Buder and S. P. Nunes. Journal of Membrane Science 260 (2005) 181-186.
9. K. D. Kreuer, Journal of Membrane Science 185 (2001) 29-39.
10. E. Drioli, A. Regina, M. Casciola, A. Oliveti, F. Trotta, T. Massari. Journal of Membrane Science 228 (2004) 139-148.
11. S. Vetter, V. Abetz, G. Goerigk, I. Buder and S. P. Nunes, Desalination 199 (2006) 289-290.
12. A. K. Andrianov, A. Marin, J. Chen, J. Sargent and N. Corbett, Macromolecules 37 (2004) 4075-4080
13. M. A. Hofmann, C. M. Ambler, A. E. Maher, E. Chalkova, X. Y. Zhou, S. N. Lvov, H. R. Allcock, Macromolecules 35 (2002) 6490
14. G. Alberti and M. Casciola, Annual Review Materials Research 33 (2003)129-54.
15. Q. Li, R. He, J. O. Jensen and N. J. Bjerrum, Chemistry Materials 15, (2003) 4896-4915
16. W.H.J. Hogarth, J.C. Diniz da Costa, G.Q.(Max) Lu, J. Power Sources 142 (2005) 223-237
17. Q. F. Li, R. H. He, J. O. Jensen, N. J. Bjerrum. Chemistry Materials 15 (2003) 4896-4915.
18. P. Stonehart and M. Watanabe, US Patent 5,523,181, Jun 4, 1996.
19. P. L. Antonucci, A. S. Arico, P. Creti, E. Rammunni, V. Antonucci. Solid States Ionics
20. C. S. Karthikeyan, S.P. Nunes, L.A.S.A. Prado, M.L. Ponce, H. Silva, B. Ruffmann, K. Schulte. Journal of Membrane Science 254 (2005) 139-146.
21. S. P. Nunes, B. Ruffmann, E. Rikowski, S. Vetter and K. Richau. Journal of Membrane Science 203 (2002) 215-225.
22. V.S. Silva, B. Ruffmann, H. Silva, Y.A. Gallego, A. Mendes, L.M. Madeira, S.P. Nunes, Journal of Power Sources 140 (2005) 34-40.
23. L. A. S. A. Prado, H. Wittich, K. Schulte, G. Goerigk, V. M. Garamus, R. Willumeit, S. Vetter, B. Ruffmann and S. P. Nunes, Journal of Polymer Science-Physics 42 (2003) 567-575.
24. S. M. J. Zaidi. Electrochimica Acta 50 (2005) 4771-4777.
25. M. L. Ponce, L. A. S. A. Prado, V. Silva, S. P. Nunes, Desalination 162 (2004) 383-391.

ELECTROCHEMICAL METHODS FOR CHARACTERISATION OF THE ION SELECTIVE MEMBRANES WITH A FOCUS ON THEIRS UTILISATION IN THE PEM TYPE FUEL CELLS

K. BOUZEK, S. MORAVCOVÁ, M. PAIDAR

Department of Inorganic Technology, Institute of Chemical Technology in Prague, Technická 5, CZ‑166 28 Prague 6, Czech Republic (bouzekk/vscht.cz, http://www.vscht.cz)

The ion selective membranes represent modern materials with a broad range of applications based on theirs specific properties. The ion selective membrane is typically applied as a so‑called solid polymer electrolyte. In this role membrane may fulfil two tasks. The more classical one consists in the separation of the anode and cathode compartments avoiding thus mixing of the electrolytes. This is either to hinder degradation of the electrochemical synthesis product on the counter electrode or to allow selective transport of the ions from the individual electrolyte streams. A typical example of such type of large‑scale industrial application represents the brine electrolysis to produce chlorine and caustic. In some applications the membrane replaces liquid electrolyte almost completely. Typical example of this arrangement represent ion selective electrodes, acidic water electrolysis and during the few last decades also PEM (proton exchange membrane) type fuel cells. There are more reasons for the replacement of the liquid electrolyte by the polymer membrane. In the case of the ion selective electrode it is typically selectivity of the membrane to particular ion or group of ions. In the electrolytic cells and fuel cells the reason consists in a substantial simplification of the cell construction and minimisation of danger of the electrolyte leakage and corrosion of the cell construction parts. At the same time reduction of the ohmic drop may be attained. This is due to the extremely low thickness of the modern ion selective membranes and theirs high conductivity.

The role of the membrane in the electrochemical systems is identical to those of the classical electrolyte. It means the electrical insulation of the electrodes and at the same time sufficiently conductive ionic contact between them allowing to close the cell electrical circuit with minimal ohmic potential loss. The demands on the membrane correspond to its task in the system. It is obvious, that in the electrochemical system sufficient ionic conductivity and mechanical and chemical stability represent the primary requirements. Beside them, various additional parameters have to be fulfilled according to the particular application, e.g. permeability and/or selectivity. Additional parameters are determined in order to allow more detail characterisation and understanding of the membrane behaviour. Following main representatives of the last group may be introduced: ion exchange capacity, ion exchange kinetics and degree of swelling. Since the aim of this contribution is to discuss in more detail electrochemical characterisation of the membrane materials with focus on theirs utilisation in the fuel cell technology, we skip majority of the parameters mentioned and focus just on the few of them relevant to this assignment.

The ionic conductivity of the membrane is one of the parameters being determined for each ion selective membrane prepared. This was already in history one of the main characteristics and it remains true until the recent days. Just the instrumentation and techniques have developed during the last decades rapidly. The methods used recently can be divided into the two main groups according to the electrical current signal used: direct current techniques and alternating current techniques.

Direct current techniques represent the original approach, which has been used historically and which is up to now well established especially in the industrial laboratories. The main advantage of this approach consists in the similarity to the membrane industrial application. It is based on the load of the membrane by the defined current density and determining the ohmic potential drop. Considering, that the membrane resistance can be approximated by the linear electrical resistance element, its value can be simply evaluated from these data using the known membrane thickness. The main danger of this approach consists in the polarisation of the membrane‑solution interface. Polarisation by the direct current causes ionic flux across the membrane. It results in the depletion of the boundary layer on one side and enrichment on the opposite one. This has two consequences. Increase in the concentration difference on the membrane results in a building up the liquid junction potential. This phenomenon represents an important artefact bringing uncertainty into the measurement of the ohmic potential drop. If the current density is to high or the electrolyte concentration to low, the concentration of the electroactive specie at the interface may fall down to zero. In such a case the potential drop on the interface between the membrane and electrolyte solution increases rapidly in order to provide sufficient amount of ions. Theirs source represents in the classical systems the water decomposition. Potential drop higher than 1.2 V on the membrane‑solution interface is necessary to decompose this molecule.

FIG. 1: Schematic sketches of the cells for the membrane conductivity measurements. A - two electrodes arrangement for the electrolyte solution environment, B - four electrode arrangement for the electrolyte solution environment, C - four electrode arrangement for the gaseous environment.

Alternating current techniques have an important advantage over the direct current ones. The periodic alternation of the current load results in keeping almost constant conditions on the both membrane surfaces. Thus, effect of the liquid junction potential and surface polarisation is eliminated. Application of the alternating current implicates appearance of the impedance, i.e. complex resistance, instead of the linear electrical resistivity. It is necessary to keep this fact on the mind when evaluating the data quantitatively. The most accurate method is based on the determining impedance spectra in the sufficiently broad perturbation signal frequency range. From such spectra it is possible by proper mathematical treatment to separate sample resistivity unencumbered with serious experimental errors. In principle two different electrical circuits can be used. The first, simpler one uses two electrodes arrangement. The electrodes used to impose perturbation signal on the system act at the same time as a sensing electrodes. Electrode reaction kinetics is thus included in the impedance spectra obtained. The alternative four electrodes arrangement uses sensing electrode separated from the electrodes imposing the perturbation signal. They are thus free of any electrochemical polarisation processes. Another problem represents the elimination of the electrolyte solution resistance. Electrolyte resistivity is similar to the membrane and it is thus difficult to distinguish between theirs contributions. The instrumental arrangement has to be used, which allows to eliminate into the necessary extend influence of these negative aspects.

The typical experimental arrangements used for membrane conductivity measurements are summarised in Fig. 1. Whereas Fig. 1A shows typical two electrode arrangement used for the measurements in the electrolyte solution, Figs. 1B and 1C summarize the four electrode arrangement cells. First of them is used for the electrolyte solution environment, whereas construction of the second one allows to work in the gaseous atmosphere.

We have to keep on the mind the fact that PEM type fuel cell works in the gaseous atmosphere and thus relevant conductivity experiments have to be performed at similar conditions. In an optimal case membrane sensitivity to its degree of swelling should be investigated and understood. This is not possible in majority of the two electrodes arrangements, one exception being application of the mercury electrodes. This arrangement is, in principle, identical to those shown in Fig. 1A. In this particular case the membrane squeezed between two electrode compartments is equilibrated with a gaseous atmosphere of defined relative humidity. Afterwards compartments are filled by the mercury. Mercury ensures membrane sample to keep its degree of swelling and at the same time precise contact of the electrode with the surface of the membrane. Since mercury is used, hydrogen and oxygen evolution reaction is strongly hindered. When the experiment is performed under the open circuit potential, spectra equivalent to the two electric capacities corresponding to the mercury - membrane interfaces with an electrical resistivity corresponding to the membrane resistivity is obtained. Resistivity is connected in series with the capacities and it is located between them. In the case of the four electrode arrangement the situation is easier, because sample can be exposed to the required gaseous atmosphere, i.e. conductivity at the required degree of swelling can be determined. Since no polarisation take place on the sensing electrodes and since no reaction occurs at these electrodes, theirs impedance spectra correspond to the simple resistivity and as such are they evaluated.

Another highly important parameter characterising the ion selective materials represents ion exchange capacity. This parameter provides information on the concentration of the dissociated functional groups available. It is extremely important information not only from the point of view of the ion exchange processes, but also for the transport characteristics of the ion selective membrane. Whereas on one hand the increase of the ion exchange capacity provides increase in the charge carriers concentration inside the membrane, it causes at the same time increase in the volume of the solvatation water in the membrane interior and thus an increase in its void volume. Increase in the void volume is connected with the two phenomenons. The first one is represented by an increase in the free membrane cross‑section. It allows faster ions movement. As a result membrane conductivity increases. It causes, however, at the same time gradual increase in the distance between the individual polymer chains and thus deterioration of the membrane stability. It makes the membrane also more sensitive to the chemical attack. Therefore, it is necessary to have relevant information on the membrane properties allowing to predict its behaviour and possible critical values of the membrane functionalisation causing break down of its stability. Beside ion exchange capacity important complementary information can also be provided by the ion exchange kinetics. Both these information may be obtained during a single experiment if it is designed properly.

The theory of the ion exchange kinetics is based on the equations of the coupled diffusion in the semi‑infinitive plain represented by the ion selective membrane, Eq. (1).

(1)

Here D_AB is the diffusivity of the coupled diffusion in the membrane volume. It is defined by Eq. (2).

(2)

In Eqs. (1) and (2) t indicates time, c concentration, z charge number, D diffusivity and subscript i ion A or B.

An important complication, when compared to the rigid inorganic materials, represents the fact that values of diffusivities change with the composition of the pore fluid inside the membrane, i.e. during the ion exchange process itself. In the present study we consider linear dependence of these parameters values on the pore fluid composition given by Eqs. (3) and (4).

(3)

(4)

Where y_i represents a dimensionless concentration expressed by the Eq. (5).

(5)

In this equation c_x indicates concentration of the ion exchange sites in the membrane. It is assumed, that ion exchange process influences diffusivity of the individual ions on the similar way, i.e. following equations are supposed to be valid.

(6)

(7)

Superscript (1) indicates that the membrane is in the cycle of the respective ion indicated in the subscript. Superscript (0) corresponds to the opposite situation, i.e. membrane is in the cycle of the second ion. Variable g in Eqs. (3) and (4) is defined as follows.

(8)

Under these assumptions following relationship for the fractional attainment of the ion exchange process equilibrium F can be obtained.

(9)

Here t is the dimensionless time defined by Eq. (10) and x dimensionless coordinate defined by Eq. (11). Value of y_A(x,t) can be obtained by integrating Eq. (12).

(10)

(11)

(12)

Where L is equal to the half of the membrane thickness and a = g-1.

Under an assumption, that one value of diffusivity is known from an independent source, remaining three diffusivity values can be obtained by the analysis of the experimental data. So, the diffusivity value from an independent source remains the only open issue. Suitable option represents diffusivity evaluated from the conductivity value using the theory of the absolute ionic mobility inside the membrane. It is described by the following equation.

(13)

In Eq. (13) k indicates the Boltzman constant, T absolute temperature, r membrane resistivity (inverse value of the conductivity), F Faraday constant and R universal gas constant. The variable c indicates concentration of the charge carriers in the pore fluid. It can be evaluated according to the following relationship.

(14)

Here represents dry membrane density, C membrane ion exchange capacity, N number of water molecules absorbed by the membrane per an ion exchange site and than partial molar volume of water.

The problem of determining the membrane conductivity at the required conditions was already discussed. Let's turn the attention to the experimental evaluation of the ion exchange kinetics. In order to evaluate diffusivity of the ions participating in the ion exchange process the data from the course of the ion exchange have to be recorded. The concentration of the released ions is subsequently transformed to the F defined by Eq. (15). It is then plotted in dependence on the square root of the dimensionless time t. The linear part of the dependence corresponding to the diffusion in the semi-infinite plain is fitted by the straight line. Slope of this line is needed for the both directions of the ion exchange, i.e. cases of ion A or B being primary inside the membrane have to be evaluated.

(15)

The variable n indicates the molar amount of the exchanged ions released by the membrane into the solution.

In the classical case of the exchange of proton for the alkali metal ion the typical method of determining the amount of protons released by the membrane sample represents an acidobasic titration. It is, however, not a suitable option in the present case. This is because of its relatively low sensitivity at the early stages of the ion exchange process. Moreover, for some ions this method of quantitative analysis is rather difficult.

In an optimal case on‑line analysis is used. This is rather simple for the membrane in the proton cycle. In such a case proton released by the membrane is easily monitored by the glass pH electrode. An attention has to be paid to the stability and sensitivity of the electrode. An example of the proton concentration course in the solution during the ion exchange process recorded by the means of the glass electrode is shown in Fig.2A. Course of the evaluated dimensionless transition function is than shown in Fig. 2B.

In the case of the membrane being at the beginning of the experiment in the sodium cycle the situation is more complicated. On‑line analysis is hardly possible and samples have to be collected in the defined time intervals. The sodium ion content has to be analysed subsequently by the means of a suitable experimental technique. ICP, polarography or ion selective electrode represent typical examples of the analytical techniques applicable. The transport parameters of the ions inside the membrane have to be fitted according to the experimentally determined slopes of the transition functions. On this way the membrane transport properties are obtained. Parameters g and dprovide an additional qualitative information on the membrane properties. It concerns especially the role of the electroosmosis in the mass transfer. At the same time the information on the free membrane volume can be obtained. This allows deeper insight into the membrane structure in the swollen state and into the role of the individual mass transport mechanisms inside the membrane.

Selected membranes showing the most promising properties are subject of testing in the laboratory fuel cell in order to provide comparison of theirs performance with the standard material, typically prefluorinated sulphonated membrane, under the fuel cell conditions. In order to avoid unexpected side effects, several conditions have to be satisfied. The first of them represents application of the suitable gas diffusion electrode. If the fuel cell is working at the temperature below 100 °C, application of the commercial electrode ensures negligible distortion of the data by the irreproducible electrode performance. However, the situation is different in the case of the novel ionomers being tested with respect to theirs performance at the temperature above 100 °C. Perfluorinated sulphonated polymer used typically to impregnate the catalyst layer is loosing its performance at such a high temperature. This is because it looses the water. Therefore method of electrode impregnation has to be modified. In an optimal case it is impregnated by the ionomer identical with this constituting the membrane. Humidification of the gasses fed has to be carefully controlled to avoid any artefacts caused by the disturbance of the fuel cell water management. In order to provide the most reproducible cell assembly, single cell is typically used at this stage. Dimensions of the cell are subject of optimisation. To small cell doesn't allow precise process control. At the same time to large dimensions cause disturbances in the homogeneity of the gasses distribution along the electrodes surface. At the same time, using a single cell, the electronic load typically can't control process at the high current load due to the low cell voltages. Two types of characteristics are typically determined during the fuel cell testing. The first, more rapid one, is a load curve shown in Fig. 3A, i.e. current corresponding to the specific cell voltage. These characteristics can be divided into three different regions. In the first one the cell performance is controlled by the electrode reaction activation overpotential. The second one corresponds to the performance being controlled by the ohmic resistances in the system. The last one belongs to the domain of the mass transfer control. For the characterisation of the membrane suitability for this type of application is decisive the second region, because the membrane represents an electrolyte, i.e. ohmic element. Using cell voltage and corresponding current load, cell performance can be evaluated. It is shown in Fig. 3B. Using the results of this analysis, optimal conditions of the long term testing can be determined. Theoretically they correspond to the cell maximum power output. In practice, however, slightly lower current load is used. This is due to the possible changes of the fuel cell performance in time, more sensitive process control and reduced danger of the cell flooding.

The aim of this contribution is just to provide a brief survey of the selected electrochemical methods used for the characterisation of the ion selective membranes with a focus on theirs perspective application in the fuel cell technology as a solid polymer electrolyte. This is just a complementary text, which is, due to the space limitation, not able to provide detail information. This is the subject of the talk delivered during the summer school.

.

MEMBRANE PROCESSING USING NEOTERIC SOLVENTS

JOÃO G. CRESPO

Requimte / CQFB - Department of Chemistry, Faculty of Science and Technology, Universidade Nova de Lisboa, Campus da Caparica, 2829-516 Caparica, Portugal

email: jgc/dq.fct.unl.pt

In many situations the focus of innovation does not lie on the search for novel structures for active compounds, but on the optimization of production processes. The challenge today is also to concentrate engineering research in non-traditional reaction processes, using new synthetic methodologies such as microreactor technology, solventless reaction, solid-phase chemistry, and integrated reaction and product recovery. The same holds true for the design of new separation approaches: reactive distillation and crystallization, fluid extraction using new (neoteric) solvents with a greener character, and membrane processing, among others.

It is clear nowadays that many key areas concern the use of novel materials and solvents, and the integrated use of both. Advances in nanotechnology will enable the solution-oriented design of nanomaterials - stationary phases for chromatography, membranes with tuned selectivity, self-cleaning capabilities and extended life-time, nanoparticles, nanoassemblies and other custom-designed nanostructures.

Water is the most widespread solvent in nature and, without question, the most important for humankind. However, most reactions and chemical processes employ organic solvents that are usually volatile, flammable, toxic and hazardous, i.e. incompatible with the principles and aims of green chemistry [Anastas and Warner, 1998]. The search for replacement of classical organic solvents is therefore one of the most active fields of contemporary chemistry. Various fluids, including supercritical fluids, perfluorinated liquids and room temperature ionic liquids are being investigated as green alternatives to classical organic solvents.

This lecture discusses the integration of neoteric solvents - such as supercritical fluids, fluorinated solvents and room temperature ionic liquids - with membranes, aiming the design of new processes combining the potential of these solvents with the unique properties of membrane materials.

In what concerns the use of supercritical fluids (SCF) this lecture will discuss new ideas recently proposed in the field of SCF [Osuna, Serbanovic and Nunes da Ponte, 2005] and membrane coupled processes [Sarrade, Guizard and Rios, 2002]. The first issue to be discussed concerns the recovery of CO₂ by membranes during a supercritical fluid extraction process. This procedure aims enhancing the extraction process by lowering the energy required for recycling CO₂. The use of supercritical CO₂ extraction coupled with membrane nanofiltration is also discussed, as well as its use for purification of low molecular weight compounds. This new process has been proposed as a solution allowing for the recovery of the product(s) of interest at high flow rates, with a good control of the whole extraction/separation process and reduced energy consumption. Cross-flow filtration applied to viscous liquids (namely, room temperature ionic liquids) fluidified with supercritical carbon dioxide constitutes another interesting example. The basic idea of SCF assisted membrane filtration is to reduce the liquid apparent viscosity at ordinary temperature by injecting a gas (CO₂) under supercritical conditions before filtration. Finally, it will be discussed the preparation of solvent-free membranes - extremely interesting for medical applications - using a CO₂ assisted phase inversion method. The effect of the solvent affinity and depressurization rate on the morphology and resulting performance of the membranes will be also discussed [Temtem, Casimiro and Aguiar-Ricardo, 2006].

The use of perfluorinated hydrocarbons, represented by perfluorohexanes (FC-72^TM), in synthesis as solvent had been quite limited until recently. Perfluorocarbons are generally immiscible with water and organic solvents at room temperature except for some low molecular weight solvents. However, there are several cases in which a mixed solvent system composed of perfluorinated solvents and other organic solvents becomes homogeneous on heating. This thermomorphic behaviour in a fluorous/organic mixed system was applied by Horváth and Rábai, 1994, which achieved facile separation of catalysts and products in Rh-catalyzed hydroformylation of alkenes [Horváth and Rábai, 1994]. The use of perfluorocarbon solvents in liquid membrane systems has been proposed and described, but full exploitation of this approach is still in its infancy.

Room Temperature Ionic Liquids (RTILs) are a special class of molten salts that are fluid at room temperature and possess relatively low viscosity (when compared with some of the traditional molten salts). Although the list of new RTILs grows very rapidly, most of them are based on cations of tetraalkyl ammonium and phosphonium salts and hetero-aromatics, typically associated with inorganic and organic anions such as BF₄, PF₆, N(CF₃SO₂)₂, CF₃SO₃, RCO₂, NO₃, ClO₄, etc. These liquids possess physical-chemical properties that qualify them as "tailor-made" green solvents for various reactions and processes [Dupont, 2005].

• They exhibit a negligible vapour pressure;
• They usually possess higher thermal, electrochemical and chemical stabilities compared with classical solvents;
• They dissolve a very broad spectrum of organic and inorganic compounds and polymeric materials, and their miscibility with these substances can be finely tuned by changing the nature of the cation and/or anion;
• They are typically non-coordinating solvents and their hydrophobicity can also be modulated by the proper choice of the cation and/or anion or by changing the temperature of the process;
• They are easily prepared from commercially available reagents through classical synthetic procedures and several of these liquids are now commercially available.

Based on these properties a number of new processes have been proposed integrating RTILs and membranes. This lecture will discuss some of these new approaches: 1 - development of RTILs supported liquid membranes and membrane contactors for liquid extraction [Schäfer et al., 2005] and gas/vapour permeation; 2 - multiphase catalysis employing RTILs, including biocatalytic processes, making use of the capabilities of the membrane as a permeselective barrier and/or as a catalytic compartment; 3 - development of new membrane materials, doped with RTILs, for use as new electrochemical devices and sensors. This lecture will also address the use of different techniques for characterization of the membrane interaction with RTILs, namely employing impedance spectroscopy, X-ray photoelectron spectroscopy and Raman confocal spectroscopy.

References

P. T. Anastas and J. C. Warner, Green Chemistry Theory and Practice, Oxford University Press, New York, 1998.

A. B. Osuna, A. Serbanovic and M. Nunes da Ponte, Supercritical Fluids, in: Green Separation Processes, Edited by C. A. M. Afonso and J. G. Crespo, Wiley-VCH, Weinheim, 2005.

S. Sarrade, C. Guizard and G. M. Rios, Desalination, 2002, 144, 137.

M. Temtem, T. Casimiro and A. Aguiar-Ricardo, Journal of Membrane Science, published electronically in 2006.

I. T. Horváth and J. Rábai, Science, 1994, 266, 72.

J. Dupont, Ionic Liquids: Structure, Properties and Major Applications in Extraction / Reaction Technology, in: Green Separation Processes, Edited by C. A. M. Afonso and J. G. Crespo, Wiley-VCH, Weinheim, 2005.

T. Schäfer, L. M. Branco, R. Fortunato, P. Izák, C. M. Rodrigues, C. A. M. Afonso and J. G. Crespo, Opportunities for Membrane Separation Processes Using Ionic Liquids, in: ACS Series on Ionic Liquids III: Fundamentals, Progress, Challenges, and Opportunities, Edited by. R. D. Rogers and K. R. Seddon, 2005.

NEW CONCEPTS TO PRODUCE INNOVATIVE POLYMERIC MEMBRANES FOR BROAD RANGE OF APPLICATIONS

A. FIGOLI

Research Institute on Membrane Technology, (ITM-CNR), c/o University of Calabria, Via P. Bucci 17/C

87030 Rende (CS), Italy (a.figoli/itm.cnr.it)

New concepts to produce innovative polymeric membranes for a broad range of industrial applications, such as cosmetic, pharmaceutical and chemical, are described in this work. In particular, two novel membranes, such as a) the multilayer membrane and b) the polymeric capsules prepared by phase inversion technique, were investigated.

The idea of this new multilayer film is presented in figure 1.

The multilayer film is made of three layers:

a) An outer dense layer to control the exchange rate of gases and vapour between the external and internal environment of the food packaging

b) An intermediate adhesive tie-layer which has also the function of reservoir of antimicrobials.

c) A porous third layer, made by phase inversion, which is able to control the release of antimicrobials to the food in time. The release of antimicrobials can be adjusted changing the morphology of the porous layer. The morphology can be controlled varying the phase inversion process conditions.

The multilayer films were prepared as shown in figure 2. All separate layers of the multilayer film could be cast subsequently on one another without removing the dense film from the glass substrate.

a) The dense (80/20) PEEK-WC/p-α-p layer, used as substrate for the multilayer film, was prepared by casting a solution of 10 wt% 80/20 PEEK-WC/p-α-p in chloroform with a hand casting knife fixed at 250 μm. A dense 80/20 PEEK-WC/p-α-p film remained on the clean glass substrate after the solvent was evaporated [3].

b) The second layer made of poly-α-pinene with and without oxalic acid (0,5, 10 and 25 wt%), was stirred at 70°C and cast at 0% RH and 70°C in a climate chamber (Angelo Antoni, Aquile 302 Challenge). The handcasting knife was fixed at 100 μm. Directly after casting the formed double layer film was removed from the climate chamber and allowed to cool at room temperature.

c) A porous PEEK-WC film was then cast on the double layer film previously prepared. The porous layer was prepared by dry-wet phase inversion. A casting solution was prepared with different concentrations of PEEK-WC in N,N diethylacetamide (DMA) (15, 19 and 23 wt%). The films were cast in a climate chamber at 50% RH and 20°C with the hand casting knife fixed at 350 μm. The porosity and morphology of each membrane were varied changing the time of exposure to air before precipitation (45s./240s.) and the water-bath temperature (0°C and 40°C) [4-7]. The multilayer film was removed 5 minutes after immersion from the coagulation bath and dried overnight at room conditions.

The preparation and characterization of the single layers which makes the multilayer, and the whole multilayer film, was deeply studied. The single layers and the multilayer film were characterized for better understanding the gas/water transport and mechanical properties through the film itself. The gas properties of the multilayer film are similar to the single dense layer. This means that the other two layers do not affect the gas transport through the multilayer film. It was also observed during the mechanical test that the multilayer film broke in two separate steps. The porous film breaks first while the dense remains operational after the break.

The cross sections of the final antimicrobial multilayer films was examined by SEM (figure 3). The individual layers of the film, indicated by the arrows, can clearly be distinguished from the reported images.

Furthermore

It was proven that the release rate of antimicrobials from the multilayer films depended strongly on the phase inversion processing conditions of the third porous layer. The release of oxalic acid increases mainly with decreasing the coagulation bath temperatures. The release rate of antimicrobials can be modulate changing the morphology of the porous layer.

b) Micro-capsules using a membrane process combined with phase inversion technique was exploited [8]. This technique can be identified as an integration between the traditional chemical capsule techniques (coacervation or phase inversion) [4,5,7] and the mechanical capsule technique (pressure extrusion).

This technique permits the formation of monodispersed polymer (modified polyetheretherketone) macrocapsules with different morphologies. The capsule morphology, porosity, size and shell thickness is easily adjusted changing the ingredient parameters such as polymer concentration, solvent and non solvent involved phases in the process. In particular, the prepared capsules show a size diameter from 200 to 1200 micron.

Active compounds, such as antimicrobials, active carbon, are loaded in the capsules during its production, which are then employed for preparing a well-defined capsule containing film. Such film has been prepared dispersing the capsules in a polymeric solution, which has just the function of bounding the capsules, and then cast it, for example, on a traditional food packaging film such as PET.

The capsules, in contact with the food environment, release the antimicrobial agent in a controlled way thanks to the different morphology and shell porosity. Experiments of antimicrobials release are reported as function of the different membrane morphology. Antimicrobials agents such as oxalic or ascorbic acid have been employed.

The experimental capsule dimensions have been compared with the droplet diameter obtained by balance force analysis along the pore month. The force balance analysis, along the droplet contact line, has allowed to understand the droplet detachment mechanism in function of pore diameter, membrane hydrophobicity, polymeric solution and process parameters.

In Figure 4, the SEM pictures of the made PEEKWC capsules are presented. These capsules are obtained with pore membrane diameter of about 500 mm, using the same polymer solvent (DMF) and non-solvent (water).

Figure 4. SEM pictures of different prepared capsules using a film with the pore size of about 500 mm using as phase 1, PEEKWC/DMF 8 wt.%; phase 2, dodecane and phase 3, water.

The formation of mono-dispersed PEEKWC capsules with different morphologies has been carried out. The capsule morphology and dimension can be easily adjusted changing the involved phases (phase 1, 2 and 3) in the process and the membrane pore sizes respectively. These preliminary tests, using membrane pore diameter ranging from 500 mm to 1000 mm, show that droplet sizes are almost constant. Furthermore, computational studies on the controlled release of active chemical compounds in function of capsule morphology and dimensions are in progress. This study can be considered as the starting point toward the production of loaded nano-sized capsules.

References

[1] J.Jansen, Master Thesis work: "Preparation and Characterisation of a New Antimicrobial Polymer film for food Packaging Application" performed at Institute on Membrane Technology (ITM-CNR), Italy (A. Figoli, E. Drioli) in collaboration with the Membrane Technology Group, University of Twente, The Netherlands, (M.Wessling).

[2] A. Figoli, E. Drioli, J.Jansen, M.Wessling, Film antimicrobico per prolungare la shelf-life, Rassegna dell'Imballaggio, ISSN0033-9687, Nov. 2004, n.16, Year 25°.

[3] A. M. Torchia, G. Clarizia, A. Figoli, E Drioli, Potentiality of PEEWC as new material in food packaging, Italian Journal of Food Science, 16 (2004) 173-184.

[4] J.C.Jansen, M.G.Buonomenna, A.Figoli, E.Drioli, Asymmetric membranes of modified poly(ether ether ketone) with an ultra-thin skin for gas and vapour separations, Journal of Membrane Science, 272 (2006) 188-197.

[5] M.G. Buonomenna, A. Figoli, J.C. Jansen, M. Davoli, E. Drioli, Mat. Res. Soc. Symp. Proc., Membranes, Preparation, Properties and Applications, 752 (2003) 3-8.

[6] Han M-J, Effect of Propionic Acid in the Casting Solution on the Characteristics of Phase Inversion Polysulfone Membranes, Desalination, 121 (1999) 31-39.

[7] Buonomenna M.G. Figoli A. Jansen J.C. Drioli E., Preparation of Asymmetric PEEKWC Flat Membranes with Different Microstructures by Wet Phase Inversion, Journal of Applied Polymer Science, 92 (2003) 576-591.

[8] W. Bakker et al., Patent WO2004/096405, Akzo Nobel B.V., 2004.

SMART MEMBRANES RELY ON SMART ARCHITECTS

M. WESSLING

University of Twente, Membrane Technology Group, 7500 AE Enschede, The Netherlands

Progress in micro and nanotechnology enables the material scientist to design membrane architectures having very specific functions. Off course, not the membrane is smart, but the designer has used his smartness and creativity to prepare these new membranes. Often such new membranes and membrane functions evolve at the interface of different disciplines. The membrane architect gets impulses for new ideas when discussing with scientist from unknown research communities.This presentation will give four examples where membranes are not used in its classical applications, but at interfaces with other disciplines like microfluidics, emulsification, tissue engineering and biomimetic surfaces..

In this article we present a new versatile replication method to produce thin polymeric microfluidic devices with tunable porosity. This method is based on phase separation of a polymer solution on a microstructured mold. Compared to existing microfabrication techniques, such as etching and hot embossing, our technique offers four advantages: (a) simple and cheap process that can be performed at room temperature outside clean room facilities; (b) very broad range of applicable materials (including materials that could not be processed before); (c) ability to make thin flexible chips; (d) ability to introduce and tune porosity in the chip. By introducing porosity, the channel walls can be used for selective transport of gasses, liquids and solutes. A proof-of concept will be given, by showing fast CO2 transport through the channel walls of a porous polymer chip. Furthermore, it will be demonstrated that the gas permeation performance of chips can be enhanced dramatically by a decrease in chip thickness and incorporation of porosity. We expect that the development of porous chips can lead to the on-chip integration of multiple unit operations, such as reaction, separation, gas liquid contacting and membrane emulsification.

The same microfluidic chips described above can be used to prepare mono-dispers droplets. The continuous phase permeates from the porous chip into the channel. Droplet breakup in the channel occurs in a very controlled manner and the size of the channel determines the size of the droplet. To produce oil-in-water or water-in-oil droplets one must change the surface properties of the porous chip.

Scaffold design is a major theme in tissue engineering; an optimal design of a scaffold appears to be crucial. Two topics should be addressed profoundly; i.e., porosity and micro-patterning of the surface. Adequate porosity ensures sufficient nutrient diffusion to the cells and micro-patterning induces alignment of the cells and therefore organizates in vitro tissue to mimic in vivo tissue. Currently, fabrication of scaffolds does not associate porosity with micro-patterning. This paper reports a novel one-step method to fabricate porous micro-patterned 2-D scaffolds with tunable porosity and pore-morphology to acquire adequate nutrient diffusion and with incorporated a micro-pattern to align the cells. Scanning electron microscopy showed that the poly(l-lactid acid) scaffolds have high porosity and pore-interconnectivity and subsequently, that the various micro-pattern designs imprinted in the scaffold are of excellent quality. Diffusion of glucose through the scaffold was high, indicating the scaffolds are successful with respect to nutrient diffusion. HUVEC culturing proved the scaffolds are suitable for cell culturing. More extensive experiments with culturing mouse myoblasts, C2C12, showed tissue organization could be controlled; the micro-pattern design affects the extent of cell alignment. The 2-D scaffold design appears to be very effective and is very promising as basis for 3-D structures due to the combination of nutrient diffusion through the porous scaffold and induced cell alignment controlled by the micro-pattern.

Phase Separation Micro Molding is utilized to prepare surfaces of a hydrophobic material with a hierarchal roughness, mimicking the self-cleaning properties of the lotus plant leaves. First of all, the microstructure increases the roughness of the surface, and additionally the porosity of the microstructure superimposes a roughness on a smaller length scale. The hydrophobicity of Hyflon ® AD was optimized by variation of the patterns of the microstructure and the phase separation process, tuning the roughness on both levels. The optimized surfaces, in terms of hydrophobicity, display water contact angles well above 160 degrees. Due to the very low contact angle hysteresis, water droplet roll-off angles are as low as 0.5 degrees.

References
(1) J. de Jong, B. Ankone´, R. G. H. Lammertink* and M. Wessling, New replication technique for the fabrication of thin polymeric microfluidic devices with tunable porosity, Lab Chip, 2005, 5, 1240-1247

(2) Bernke J Papenburg; Laura Vogelaar, Dr; Lydia A Bolhuis-Versteeg; Dimitrios Stamatialis, Rob G Lammertink, Matthias Wessling, One-step Fabrication of Porous Micro-patterned Scaffolds to Control Cell Behavior, submitted to Biomaterials

(3) Laura Vogelaar, Rob G. H. Lammertink,* and Matthias Wessling, Superhydrophobic Surfaces Having Two-Fold Adjustable Roughness Prepared in a Single Step

Langmuir, 22 (7), 3125 -3130, 2006