Development of the Noble Metal/Oxide Coated Titanium Electrode PART 111: COATED TITANIUM ANODES IN WIDELY RANGING OXYGEN EVOLVING SITUATIONS By €? C. S. Hayfield Bickenhill, West Midlands, England

The first two parts of this paper on titanium-based electrodes concentrated on the development of coatings comprising noble metals and their oxides suitable for use in chloride-bearing electrolytes. This concluding part deals with simi- lar types of electrodes which act as counter electrodes in commercial processes where the product is made on the cathode. Many electrolytes in this category involve sulfuric acid, which means that oqgen evolves at the anodes. Applications for which these electrodes are used range from those operating at very high current density, such as electrogalvanising, to others at much lower current densitr, such as impressed current cathodic protection of rebar in concrete.

When platinised titanium-type electrodes were type chlorine cell technology. first envisaged for commercial use, manufac- The electrolyte is usually strong sulfuric acid turers fixed their sights almost exclusively on (- 150 g 1.') at a temperature above ambient, using them as a possible replacement for graphite usually - 40°C, but at a significantly lower tem- in high current density chlor-alkali processes, perature than the electrolyte in chlor-alkali cells. in order to effect a lowering of the cell voltage The electrolyte is more likely to have an and therefore produce cost savings (26).In the organic content. author's experience, other applications, for There may be a high manganese level (10 instance replacing lead-based anodes in elec- to 15 g 1.'). trowinning, were not initially considered. Such Some chloride and fluoride may be present. applications, mostly involving sulfuric acid-based As the cathodes are periodically removed electrolytes, call for durability under oxygen to recover the metal, there are more disconti- evolving electrochemical conditions. Present nuities in operation than in chlor-alkali pro- day, large-scale applications for titanium-based duction. electrodes, including electrogalvanising and There may also be anodekathode shorting. impressed current cathodic protection of rebar All in all, this is rather a formidable environ- in concrete, did not then exist. ment in which to introduce viable titanium- based electrodes, and the incentives for change Electrowinning are therefore not straightforward. However, it In electrowinning, lead-based electrodes are can be advantageous to produce a cathodic prod- used to extract metal from a solution of its salts uct containing only low levels of lead, and while by electrolysis, with the electropositive metal there may be some savings in electrical cost, being deposited on the cathode. In this indus- high levels of manganese in solution usually try conditions differ from the chlor-alkali indus- make this difficult because of manganese diox- try in numerous ways: ide deposition on the surface of the anode. Current densities, in general, are about an The 'Bunker Hunt' affair in 1980, when the order of magnitude lower than in diaphragm- price of silver escalated, meant that some users

Platinum Metals Rev., 1998, 42, (3), 116-122 116 Fig. 12 A 20-cell plating section from a Gravitel electrogalvanisingplant in Usiminas, Brazil. The capacity is 360,000 tonneslannum. A 20 cell unit will contain around 400 electrodes, each of size - 0.5 ni x 1 m Reproduced by courtesy of Audritz AG Ruthner Surface Technologies switched to titanium-based electrodes to commercial importance, the characteristics of capitalise on the silver inventory in their lead- the electrocatalysts, iridium oxide and iridium silver anodes. In fact, the requirements of each oxidehantalum oxide mixtures, were studied in plant needed to be considered individually; but more detail. Usually, the favoured means of apart from a few plants which operated with depositing pure iridium oxide is by the paint/ chloride-based electrolytes and then switched thermal decomposition route using chlor-iridic to titanium-based electrodes, the majority of acid dissolved in an alcohol. If the decomposi- the world metal winning industry still uses lead- tion temperature is too low then water-soluble based electrodes. iridium chloride remains, but if the decompo- sition temperature is too high the iridium oxide Durable Electrocatalysts for Oxygen is converted to less catalytically-durableiridium Evolving Anodes metal. The window for the correct decompo- A good place to start looking at present-day sition temperature is small, only about 20°C. In oxygen electrode technology is the early de Nora the preferred iridium oxide form, the electro- patent assigned to Bianchi and colleagues, where catalyst has a tendency to change &om an attrac- the preferred electrocatalyst was IrO,/Ta,O, tively-low to an undesirably-high oxygen over- codeposit (27). It is said that electrodes with potential. Factors that cause passivation include: this coating were incorporated into a Sulfomat length of time under anodic polarisation cell installed in Japan by the U.S. company Ionics increasing current density and Inc., of Watertown, Massachusetts, but the lives decreasing acidity. of these electrodes were reputedly much shorter The inventors of the de Nora patent seem, like than those of anodes in chlor-alkali cells. Beer with his Ru02/Ti0, patents, to have found As oxygen electrodes became of greater the key parameters back in 1975. Subsequently,

Platinum Metals Rev., 1998, 42, (3) 117 Comninellis and coworkers have clarified much precipitation, which caused such thick film of the detail (28,29). Perhaps the most impor- formation. At more elevated temperature, 70°C tant finding is that mixing tantalum oxide with and above, structural corrosion may still occur the iridium oxide acts, in some way, to slow but the continuing localised titanium dissolu- down the reduction of the iridium oxide to irid- tion no longer precipitates to form film. ium metal. Thus, while iridium oxide by itself An early warning of dire events that could hap- requires a decomposition temperature of pen to titanium-based electrodes was implicit 380 to 4OO0C, a codeposit of Ir02/Ta205can be in a paper by Antler and Butler who reported safely put into a furnace at 450 or even 5OO0C, degradation mechanisms in platinum- and and this also results in the expulsion of more rhodium-plated titanium electrodes at accel- residual chloride ion. erated high current densities during high tem- In retrospect, the platinum in a 70/30 Pt/Ir perature testing in simulated chlorate liquor coating appears to slow down the reduction of (31). Thick titanium oxide was found to have iridium oxide. Thus, with paints comprising formed underneath the noble metal coating. chlor-platinic and chlor-iridic acids in alcohol, How could this occur if the local anodising decomposition temperatures of 500°C can safely potential on such specimens was never more be used, producing coatings comprising fine than a few volts? As no similar electrode failures scale mixtures of Pt + IrO,. were found in sodium chlorate plants, the pos- Some evidence for the less powerful nature of sibility of thick film formation under noble iridium oxide as an oxygen electrocatalyst, com- metal/oxide coatings was put to one side for pared with ruthenium oxide as a chlorine elec- many years. trocatalyst, can be gauged from the finding that optimised IrO,/Ta,O, codeposits for oxygen evo- Electrogalvanising: The Andritz lution are always high in iridium oxide. A paint Ruthner Experience with a high iridium oxide content, such as 95/5 The next episode in this story came when the or 90/ 10, tends to favour the suppression of chlo- Austrian company, Andritz Ruthner of Vienna, rine evolution in electrolytes containing trace came up with a novel method of coating zinc chloride, while an optimised 'work horse' com- onto one side of continuous coils of steel strip, position is nearer 50/50 dwtper cent. principally for the automobile industry, but also With hindsight, an improved understanding for domestic appliances, see Figures 12 and 13. of the titanium substrate was needed to explain This Gravitel process also finds application for accumulating experiences with titanium-based coating both sides of steel strip to different thick- oxygen electrodes. A prerequisite for any com- nesses, see Figure 14. mercial coated titanium electrode must be an The Gravitel process involved zinc deposi- adherent and durable coating. In most chloride- tion at high current density (equivalent to mer- bearing electrolytes, even when quite acidic, the cury-type chlorine cells at 10 to 20 kA m-') from titanium is adequately durable to general cor- made-up electrolyte compositions. Initial eval- rosion, provided the design gives freedom from uation seemed to show that a pure iridium oxide anodic breakdown corrosion. Highly concen- electrocatalyst on a titanium substrate with an trated sulfuric acid is generally more aggressive overlay of tantalum oxide would sufEce to uprate towards titanium. In engineering terms anodic corrosion resistance in hot sulfuric acid. polarisation (protection) allows practical usage Accordingly, a massive quantity of electrodes for many years. In one reported situation (30), was made for - almost simultaneous - use in however, titanium, polarised at 12 volts in 10 three plants. However, their operation showed per cent sulfuric acid at 35"C, was found to be that the nominally identical anodes behaved quite covered with film of thickness 30 to 40 p.Later differently in each plant. In one plant, after a few examination showed that there was a slow months' operation, the electrodes suffered pre- but continuing pitting dissolution followed by mature degradation due to 'Antler-Butler-type'

Platinum Metals Rev., 1998, 42, (3) 118 Fig. 13 A smaller electrogalvanisingaddition to an existing coil coating plant Reproduced by courtesy of Andritz AG Ruthner Surface Teehnologiea thick film formation, which physically disrupted mining attack. In the third plant there was the electrocatalyst coating. The phenomenon neither thick film growth under the coating, nor was initially and erroneously thought to be some high wear rate, and the electrodes clearly behaved form of attack from a deposit. In the second plant as expected. How could these nominally simi- there was no evidence of thick film formation lar electrodes behave so differently in the three under the coatings, but there was a high rate plants? The explanation lay in the characteris- of intrinsic wear and loss of adhesion by under- tics of the titanium substrate described earlier.

Banks of anodes set close to

between anodes and coating both

set close to strip Fig. 14 The principle features tor single side of the Gravitel method of zinc coating continuously coating steel Circulating zinc strip with zinc electrolvte

Platinum Metals Rev., 1998, 42, (3) 119 In the first plant, where thick film formed Gravitel plants formed the basis both for the under the electrocatalyst, the operating tem- construction of improved electrodes and meth- perature was about 65°C and the current den- ods of use. If the etching characteristicsof com- sity was high. Unexpectedly, even though the mercially pure titanium of differing residual surface electrode potential was controlled by impurity content are examined, etching rates the noble metal oxide electrocatalyst, there was increase with increasing impurity. Hence, high micropitting of the titanium and precipitation current density electrogalvanising anodes, see of film. In the second plant the electrolyte tem- Figure 15, benefit from: perature exceeded 65°C and the current den- being made from the purest titanium sity was particularly high in order to meet the having residual surface impurities leached special needs of the plant. Thus titanium dis- possibly having interlayers to help shield the solved without precipitation and the iridium electrolyte from the titanium, including oxide oxide suffered transpassivation dissolution. layers preformed by thermal oxidation These resulted in electrode replacement more having thick coatings of IrOz/TazO,to com- often than planned. The third plant, operated bat the effects due to the high current density. at lower electrolyte temperature and lower cur- While the first chlorate electrodes received rent density, did not experience the fault-caus- 70/30 Pt/Ir coatings with loadings of less than ing conditions of the other two plants, at least 5 g m-’, and most chlor-alkali electrodes use over protracted periods. ruthenium-based coatings with loadings up to The experience gained from these first three 20 g m-*, loadings on electrogalvanising anodes can be 50 g m-’ or higher, and even then the life- times are not much in excess of one year. Coatings for this application are still an on-going development area.

Corrosion Protection In addition to the use of oxygen evolving anodes in electrogalvanising, other opportuni- ties for using electrodes occurred in electrotin- ning, electroforming and the continuous ano- dising of aluminium strip for the printing industry. Another unexpected application then presented itself relating to reinforced concrete used in structural engineering. Over the years a small percentage, but huge in total, of rein- forced concrete had suffered serious internal corrosion of the reinforcing bars from the pres- ence of chloride ions. How could this most insid- ious and continuing corrosion situation be con- trolled? A consensus view in the 1980s was to use cathodic protection. Concrete is porous and contains electrolyte in pores in contact with the reinforcement. If the electrochemical poten- tial of the steel could be controlled electro- chemically, as in the protection of ships’ hulls (32), then corrosion could be delayed until the Fig. 15 Typical titanium electrodes used in electrogalvanising plants. The laboratory assistant damaging ions had migrated away shows the size of such electrodes towards an anode. Hence external anodes are

Platinum Metals Rev., 1998, 42, (3) 120 Fig. 16 Mesh positioned on a support beam prior to applica- tion of a cementitious overlay. This gives an impression of the scale of usage of the mesh. The beam is part of a wharf in the Middle East Reproduced by courtesy of Corrosion Control Services Ltd

necessary for this method of protection (33). compared with, say, electrogalvanising, only It has always been recommended that pla- very low loadings of electrocatalyst are neces- tinised titanium anodes, at least those operat- sary, often less than 1 g m-’. In fact, over the last ing at high current density, should be kept free 15 to 20 years, many hundreds of thousands from surface deposit (34). This is to prevent the of square metres of reinforced concrete - in formation of localised highly concentrated acid- roads, bridges and buildings have been treated, ity which could damage both the coating and with no reported failures of the titanium, which the titanium substrate. The application of coated is just as well as most are designed to have lives titanium in concrete is only really practical by of 25 or more years. virtue of operating at current densities some five The present owners of the patents on the mesh orders of magnitude lower, which limits the form of anode have a virtual world monopoly, amount of acidity formed. Acidity is further just as the owners of the Beer 2 patent held an neutralised by the proximity of highly alkaline almost world monopoly on codeposition meth- concrete. ods of applying coatings. As happened with the The commonest form of titanium-based anode Beer patent, numerous attempts have been made for application to concrete is an open mesh to compete with the titanium mesh form of not dissimilar to galvanised chicken-wire mesh, anode by various ingenious alternative devices. see Figure 16 (35-38). The holes in the pattern are around 5 cm x 5 cm, and the actual real sur- Discussion face area is about one tenth of the geometrical The titanium-based electrode business is now surface area it covers. To complete an electro- commercially huge, involving possibly billions chemical cell, the mesh is first applied to the of dollars. In this survey, it has not been possi- active surface and then attached to the struc- ble to cover the niceties of most applications, ture with a cementitious material, usually by such as anodes in chlorate, the production of spraying, see Figure 17. The electrolyte then persalts, hypochlorite generation or more recent makes contact between opposing electrodes via possibilities in organoelectrosynthesis, ground water in the pores of the concrete, which usu- remediation on local and large scale, controlled ally also helps to dissipate any anodically gen- ground water migration, effluent treatment, elec- erated gas. Such gas may include chlorine if troetching, and other uses. there is chloride in the concrete adjacent to With the increasing sophisticationof all aspects the anode, but it is predominantly oxygen. of modern technology, it is perhaps surprising Electrocatalysts chosen for titanium-based that the repetitive and laborious painthermal anodes in concrete include 70/30 Pt + LOz,IrOz decomposition method of building up the and IrOz/TazO,.Because the current density required electrocatalyst should continue to be of the application, at - 0.1 A m-*, is so low used. True, there are plants with a high degree

Platinum Metals Rev., 1998, 42, (3) 121 Fig. 17 Coated chicken-wire type titanium mesh positioned for use in the impressed current cathodic protection of rebar, just prior to application of a cementitious overlay Reproduced by courtesy of Corrosion Control Services Ltd

of inbuilt automation, but attempts (and there the irrepressible Joe Cotton, the quietly have been many) to build up coatings quickly, working Cliff Angell, but let no-one lose sight for example, by heavy increments using of the enormity of credit due to the charming, paintkhermal decomposition, have not been gentle, unassuming, Henri B. Beer. successful. In fact, computerised atmosphere- controlled plasma deposition is making inroads Acknowledgements for some applications, including cathodic pro- It is sincerelyhoped that organisations whose names have been included will not take issue over inclusion tection, but it seems likely that the delicate, quasi in this narrative. Appreciation is extended to the huge equilibrium, high sensitivity electrocatalysts will number of conmbutors to the technology. Finally the continue to be made by the paint/thermal writer thanks a number of technical colleagues for reading the manuscript, including B. H. Hanson, decomposition method for many years to come. D. A. Hughes, A. T. Kuhn and last, but by no means The Beer patents have now expired worldwide, least, my long-established friend M. A. Warne. allowing the number of electrode manufactur- ers to increase, although most electrodes con- References 26 P. C. S. Hayfield, PlatinumMetalsRev., 1998,42, tinue to be made by a few experienced large (l), 27; op. cir., 1998, 42, (2), 46 companies who have the financial resources to 27 0. de Nora, A. Nidola, G. Trisoglio and G. guarantee their products fully. Bianchi, British Patent 1,399,576; 1973 The extensive use of noble metals and oxides 28 Ch. Comninellis and G. P. Vercesi, 3. Appl. EZecmochem., 1991,21, (4), 335 as coatings on titanium has contributed to a bet- 29 Ch. Comninellis and E. Plattner, Symposium ter understanding of their electrochemistry. on ‘Performance of Electrodes in Industrial However, despite the growth of the noble Electrochemical Processes’, The Electrochemical Society, 1989, pp. 229-243 metal/oxide coated titanium anode, the total 30 P. C. S. Hayfield, “Titanium Science and consumption of noble metals in these industries Technology”, AIME Publication, ed. R. I. Jaffe has not attained levels that dictate world prices. and H. M. Burte, 1973,4,2405-2418 31 M. Antler and C. A. Butler, J- Elecmochem. SOC., Rather, the electrode market remains at the 1967,5, 126 mercy of the vagaries of other applications. Few 32 P. C. S. Hayfield, Platinum Metals Rev., 1983,27, in the industry over the last 40 years will for- (1)>2 get the marked price changes for iridium, the 33 P. C. S. Hayfield, Platinum Metals Rev., 1986,30, volatility in the price of silver, even the quite (413 158 34 J. B. Cotton, Platinum Metals Rev., 1958, 2, (2), recent fall in rhodium price, but throughout 45 the years the price of platinum has remained 35 P. C. S. Hayfield, British Patent 2,175,609; 1986 relatively stable. 36 J. E. Bennett, G. R. Pohto, T. A. Mitchell and C. H. Brown, European Parent 225,343; 1990 Finally, this article has been prepared as a 37 J. I. Bennett, G. R. Pohto and T. A. Mitchell, tribute to the few who started the commercial- European Parent 222,829; 1989 isation of titanium-based electrodes. There was 38 G. L. Mussinelli, U S. Parent 5,062,934; 1989

Platinum Metals Rev., 1998, 42, (3) 122