Electrochemical processes are interesting due to their general advantages:

• redox reactions are carried out directly with electrons at the electrodes,
  - no redox reagents are consumed,
  - no waste of redox reagents has to be disposed,
  
anodic and cathodic reactions are separated,

the reaction rate can be easily controlled by the electrical current,

reactions, which need much energy, are possible by high electrode potentials, e.g. production of chlorine, fluorine or lithium metal,

usually mild reaction conditions.

Our research is focused on ion exchange membranes, used as:

• permselective cell separators that are preferably permeable for
  - cations (cation exchange membranes CEM),
  - anions (anion exchange membranes AEM),
ion conductors (solid polymer electrolytes, SPE technology) which need no additional supporting electrolyte.
Of special interest is the combination with gas diffusion electrodes for energy saving.

Chlorine and caustic soda, amongst the largest products of chemical industry, are produced by electrolysis of common salt NaCl. The mod-ern membrane process (see Fig. 1) will replace in future the diaphragm and mercury processes because it reduces the energy consumption to about 70 % of the older processes and avoids environmental problems, caused by asbestos or mercury. The cation exchange membrane CEM separates the anodic and cathodic products due to its highly selective transport of Na+ cations and rejection of Cl- and OH- anions. Former research under the guidance of Prof. Dr. Karl Hans Simmrock refers to this process [publications 1-3,15, dissertation D1-D7].

A successful industrial project actually applies the membrane process to produce sodium hypochlorite solutions on-site for water disinfection. For such small devices the expensive brine system of the big plants ac-cording to Fig. 1 is unacceptable. Based on research results of our group the simplified process in Fig. 2 was developed. It runs reliably using directly cooking salt without supplementary purification [patent application P4]. An important property of the process is the formation of oxygen and hydrochloric acid additionally to chlorine at the anode, in order to protect the membrane against precipitation of calcium and magnesium compounds from impurities of the salt. Thus, the electrical energy consumption is somewhat increased, but this is without difficulty compensated by economized efforts in the brine system and the fail-safe operation.

The by-product hydrogen in industry frequently is not desired. Then, instead of the hydrogen evolving cathode in Fig. 1 an oxygen consum-ing cathode, known from fuel cell technology, according to Fig. 3 can be applied. Thus, 0.8 to 1.0 V of the cell voltage, equivalent to up to 30 % of the energy consumption, can be economized. First research in this area many years ago was carried out in Prof. Simmrock´s group [8,D9]. Actually, this technology is developed and realized in industry. The related work of our group is performed in collaboration with industry.

A large amount of the produced chlorine in industry is used in order to apply its high reactivity, but subsequent it is eliminated, forming chlorine free products and hydrogen chloride. One possibility to use this hydrogen chloride is electrolysis in order to recycle chlorine back into the production process. State of the art are diaphragm cells with graphite electrodes, but they need pure hydrochloric acid of high concentration. New industrial developments apply cation exchange membranes and additionally oxygen consuming cathodes, comparable to Fig. 3. Never-theless, a considerable part of the hydrogen chloride is produced as diluted hydrochloric acid which has to be neutralized. The resulting salts are an environmental problem in the waste water.

An alternative process is investigated in our group (see Dr. Vladimir Barmashenko, AiF project 12774 N, report of 2003 available). An anion exchange membrane transfers according to Fig. 4 chloride anions into the anode compartment. There, a high chloride concentration for opti-mal chlorine evolution with low oxygen content is assured by addition of a salt, preferably calcium chloride. Even low concentrated hydrochloric acid in the cathode compartment is sufficient to supply chloride anions. Usually, anion exchange membranes are destroyed immediately by oxidizing agents. But now a chlorine resistant anion exchange mem-brane, primarily developed by a Russian research institute, is available from the production of a German company (Fumatech, St. Ingbert).

A special problem of the process in Fig. 4 is the transfer of water into the anode chamber that necessitates partial discharging of anolyte or water separation. Thus, in the cell design of Fig. 5 an anode, which is directly connected to the membrane, is applied, and the anode chamber is only filled with gas. A wire mesh (platinum/iridium) is used as anodic current collector and the anode itself is made of a catalyst powder, e.g. of platinum black and/or ruthenium oxide, which is embedded in the surface of the membrane. This technique is known in principle from fuel cells (membrane electrode assembly MEA). First results are promising, further optimization is object of our investigations.

Recycling of salts by splitting into the corresponding acids and bases, e.g. of sodium sulfate into sulfuric acid and caustic soda solution, enables closed loops in chemical processes. Electrolysis and electro-dialysis, using cation and anion exchange membranes like in Fig. 6 and also bipolar membranes are investigated. Additionally an alternative, non-electrochemical process has been studied [20,D18].

Essential for the salt splitting processes is the transport behavior of the ion exchange membranes. Our results show that it is either dependent on the conditions at the anode or at the cathode side of the mem-brane, but never it is simultaneously influenced by the conditions on both membrane sides. This can be elucidated by the model of an either "alkaline" or "acidic" state of the membrane [7,15,D8,D19]. Based on this model, the cascade connection of cells for salt splitting can be calculated and has been experimentally verified, in order to realize better current efficiencies and/or higher product concentrations [7,19, D21, AiF project 10491 N, report of 1998 available].

The reality of the "acidic membrane state" has been confirmed by a successful long time electrolysis experiment with unpurified sodium sulfate [D21] (and also by the fail-safe hypochlorite production, mentioned above, see Fig. 2). On the other hand, under conditions which result in the "alkaline membrane state", the membrane is immediately and irreversibly destroyed in presence of salt impurities.

A further transport model ions has been developed to elucidate the undesired selectivity deterioration of anion exchange membranes by multiple charged anions like sulfate or phosphate. On the other hand the desired co-transport of H+ ions within hydrogensulfate or hydrogen-phosphate ions increases the effective membrane selectivity, possibly up to 100 % for acid concentrations on the anode side below a mem-brane specific limit [29,D22,D24].

Unfortunately, the selectivity of membranes that are available for the production of strong acids, like sulfuric acid, and strong bases, like caustic soda solution, according to Fig. 6 until now is unsatisfying (the highly selective membranes for chlor-alkali electrolysis, see Fig. 1, are not suitable here). At present, the concentration of sulfuric acid should not exceed 10 wt-% and of caustic soda solution 20 wt-% if the current efficiency should not decrease below 80 %.

Energy saving by gas diffusion electrodes is here of special interest. Oxygen consuming cathodes, like in Fig. 3, or hydrogen consuming anodes can be applied, using the gas that is produced at the opposite electrode. Hydrogen consuming anodes are preferable, because they obtain the lowest energy consumption (small overvoltage) and avoid undesired oxidation reactions due to their low potential [8,11,D17,D22].

Ion exchange membranes, working as solid polymer electrolytes (SPE technology), enable electrochemical reactions like electro-organic syn-theses without addition of any supporting electrolyte. This principle is very actual for fuel cells like the direct methanol fuel cell (DMFC), but in case of electro-organic syntheses some special conditions have to be considered [28,30]. Numerous reactions have been successfully tested using various types of cation and anion exchange membranes in com-bination with different electrode materials in aqueous and non-aqueous media by experiments up to 8 months [10,14,15,27,28,30].

A specific property of ion exchange membranes is the electro-osmotic stream (EOS). Fig. 8 shows the transport of solvated ions at the surface of one particle of a porous anode, e.g. at a single fiber of graphite felt (not to scale). This is surrounded by a laminar diffusion layer where convection in the bulk phase of the liquid electrolyte is not present. In a conventional electrolyte (left of Fig. 8) cations and anions are moving in opposite directions and no effective mass transfer is generated by their solvation shells. Therefore, uncharged reactants and products can be transferred by diffusion alone. Limited diffusion rates for reactants and/ or products of the desired anode reaction can enable side reactions.

In SPE technology the cell liquid is free from ions, which are present only in the membrane (right of Fig. 8). The fixed ions (here, in a cation exchange membrane, anions) are bound to the polymer and thus they are immobile, including their solvation shells which are formed due to the swelling of the membrane. The charge transfer is effected only by counter ions (here H+ cations) and their solvation shells generate the EOS. It includes unselectively all uncharged compounds with the con-centrations that are present at the working electrode.

In consequence, a convective mass transfer in the membrane is caused which is typical of SPE technology. The H+ ions are formed during reaction at the anode surface (see Fig. 7) and therefore they have to be solvated at this location for further transport through the membrane. This requires molecules - not only products but also reactants and solvents - which are flowing from the cell liquid to the anode surface. Thus, in contrast to a conventional electrolyte with a diffusion layer at the electrodes, in SPE technology convective mass transfer occurs directly at the electrode surfaces. Because the EOS and consequently the enhancement of mass transfer increase proportional with increasing current density relative high current densities are obtainable in the SPE technology. Of course, mass transfer by diffusion takes place addition-ally to the EOS.

The direction of the EOS can be decided by choosing a cation or an anion exchange membrane. The quantity of the EOS was observed from 2 up to 20 and more molecules per charge carrier ion. It is very dependent on the composition of the solution in the cell and on the type and the preparation of the membrane. Especially for Nafion® cation ex-change membranes (perfluorosulfonic acid) the EOS can be enhanced by swelling of the membrane at elevated temperature in amides like N,N-dimethylformamide (DMF). The EOS increases proportional with the increasing area of the membrane (about factor two for reactions like in Fig. 7) [13,30,D13,D20,D23].

In consequence of the electro-osmotic stream EOS, that has in cation or anion exchange membranes a different direction, four configurations of the SPE technology are possible. In b) and d) of Fig. 9, and also in Fig. 7, mass transfer in the cell is generated only by the EOS, the membrane works like a dosing pump. The cell fluid is exposed to anodic and cathodic reactions. This may be desired, see section 6. If undesired reactions occur, possibly they can be oppressed, e.g. by choosing of electrode materials (see beneath). In the configurations a) and c) the EOS runs in opposite direction to diffusion. This offers the chance to perform only cathodic reactions in configuration a) or only anodic reactions in configuration c), respectively (see sections 6 and 7).

Some detailed results of our investigations using the SPE technology:

• A good contact between the electrodes and the membrane is needed, but it is not necessary to built up a composite of the electrode and the membrane, like the membrane electrode as-semblies (MEA) of fuel cell technology. Even simple pressing of permeable electrodes on the membrane surfaces is successful. Thus, the range of possible electrode materials is extended: e.g. graphite felt (advantageous properties: soft and elastic) or coated titanium (expanded or sintered metal).
Metal oxides on titanium showed interesting electro-catalytic properties and in combination with the SPE technology a highly increased lifetime compared with conventional sulfuric acid electrolytes:

-electrochemically deposited lead dioxide [D12,D14],

-electrochemically deposited manganese dioxide [24,D26],

-thermally prepared layers (similar to the dimension stable anodes (DSA) for chlor-alkali electrolysis) of cobalt-manga-nese and chromium-antimony-titanium oxides [24,D26].

Reasons for the longer life time may be the absence

-of aggressive electrolytes that could dissolve the oxides,

-of any turbulence as origin of erosion.


In the inert cell fluid without supporting electrolyte chemical side reactions are avoided, as was demonstrated for the methoxy-lation of furan [6,D11].
A lot of reactions, especially methoxylations in non-aqueous media, see Fig. 7, have been found to run better using the SPE technology than using a conventional supporting electrolyte, like the toxic, expensive and difficult to separate liquid salt tetrabutyl-ammonium-tetrafluoroborate [9,13-15,D13,D23]. A plausible elucidation is given by the enhanced mass transfer at the anode (graphite felt) due to the EOS: only with prepared membranes, as mentioned above, the selectivity of the reaction in Fig. 7 can be increased nearly up to 100 %. Using configuration b) of Fig. 9 (equivalent to Fig. 7) side reactions are avoided at graphite felt.
Various methods have been established for operating the SPE technology continuously in non-aqueous media with a stable, non increasing cell voltage [22,26, patent applications P1-P3]. These developments are based on measurements of the different parts of the voltage in an SPE cell [17,23,27,D25].
As example for using an anion exchange membrane in the SPE process, according to configuration d) of Fig. 9, g-butyrolactone was oxidized. At a graphite felt cathode it is converted into g-hydroxybutyrate anions, that migrate through the AEM and are oxidized at a lead dioxide anode (on titanium carrier) to succinic acid. This oxidation is also possible using the SPE process with a CEM, but it runs much better with an AEM [9,14,15,D14].
The reaction of Fig. 7 was investigated also in a pilot-plant cell of 250 cm2 membrane area. The distributions of current, concentrations and temperature were measured, and mass and heat balances were calculated. The results show that a scale-up for industrial dimensions should be possible [18,21,D20].

Electrochemical detoxification of water is possible without any additive using the SPE technology (see Dipl.-Ing. Anna Heyl). 2-chlorophenol as an example almost completely can be decomposed, with extensive mineralization of the included chlorine as chloride ions. The electro-osmotic stream (EOS) can be used for enhancing the mass transfer at the electrodes. Thus, significant differences using cation or anion ex-change membranes in the configurations according to Fig. 9 are ob-served. Best results are obtained with configuration b) of Fig. 9, i.e. first anodic oxidation and the cathodic reduction, using Nafion® membranes. An intensive preparation of Nafion® membranes as mentioned above increases the EOS to more than 20 (up to 31) molecules water per H+ ion. Consequently, due to the strong pumping effect of the EOS the flow rate is increased and the consumption of electrical energy is decreased (about 150 kWh/m3 of water, comparable with commercial processes of electrochemical waste water treatment).

A serious, until now unsolved problem of direct methanol fuel cells (DMFC) is the undesired transfer of methanol to the cathode through the usually applied cation exchange membrane (Fig. 10, left). This is caused by diffusion and also by the electro-osmotic stream (EOS). In case of an anion exchange membrane as solid polymer electrolyte (Fig. 10, right) the EOS could be applied to reject the diffusion of methanol at higher current densities. In a simulated DMFC, using a hydrogen evolving cathode in configuration c) of Fig. 9, this principle has been verified: methanol transfer to the cathode side was nearly stopped at a current density of 60 mA/cm2 [25]. The precondition for realizing a DMFC according to Fig. 10, right, (complete scheme shown in Fig. 11) would be the availability of anion exchange membranes of sufficiently high conductivity and stability against OH- ions.