Platinum Metals Rev., 1993, 37, (2), 76
Knitted Platinum Alloy Gauzes
Catalyst Development and Industrial Application
The introduction of knitted catalyst gauzes into the nitric acid industry has had a major impact on both that industry and on the catalyst production procedures employed by Johnson Matthey. It is now two years since this technological advance was first reported and, because of this time interval, a summary of previous work is included here to serve as a background to this significant technical progress and the commercial acceptance that has taken place over the past few years.
The production of nitric acid is achieved by passing an ammonia-air gas mixture over a catalyst and absorbing the resultant gas in water. Most industrial plants use a platinum group metal catalyst, generally either rhodium-platinum or rhodium-palladium-platinum, to promote this reaction. The process is carried out at temperatures in the range of 800 to 940°C and at pressures of between 1 and 14 atmospheres. In order to realise optimum performance the catalyst is usually in the form of gauze, woven from wire. This structure was developed in 1909 (1), and until very recentiy little change in design had taken place.
Platinum behaves well as a catalyst, and early catalyst development involved the use of pure platinum (2) a relatively weak material. Among the first alloys considered for ammonia oxidation was platinum-iridium (2) and platinum containing 10 to 20 per cent palladium. Economic appraisal of these catalysts showed that none of these alloys was suitable for use in nitric acid plants, but platinum-rhodium alloys containing 5 to 10 per cent rhodium were subsequently found to have the best combination of properties in terms of ammonia conversion efficiency and economic performance; the rhodium increasing the relatively low tensile strength of pure platinum. That was 60 years ago.
Now platinum-rhodium alloys are at the heart of the nitric acid production process and as such have been the subject of much discussion. The sharp increase in the price of rhodium during the period 1989–1991, see Figure 1, led many nitric acid producers to consider reducing the rhodium content of their catalysts. The use of 10 per cent rhodium-platinum woven 1024 mesh gauze was well established (3), however, and alloys with lower rhodium content had reduced strength and a potential for increased metal loss.
The introduction of knitted catalyst gauzes had a major impact on catalyst producers and users. The production of knitted gauze and its performance, compared with that of traditional woven gauze, is discussed here.
Rotary Knitting Machine Development
After successful commercial trials of knitting gauzes, a full-scale production system was developed (3).
The circular knitting frame on the first production machine was 76.2 cm in diameter, with approximately 1990 needles around the circumference, see Figure 2, the size being determined by the need to reduce wastage and minimise seam fabrication. Having already determined the optimum density of the cloth on die earlier development machine, the needle spacing and stitch length were maintained to achieve continuity.
Although a single feed system could probably cope with the anticipated demand, it was decided that a 4-feed system would give greater flexibility. Indeed, all Johnson Matthey machines are capable of accepting 8 feeds although a requirement for such production rates is unlikely.
The cloth is produced by co-knitting the platinum group metal wire with a multistrand polyester yarn carrier which almost completely covers the wire and provides support during the knitting process (4), see Figure 3. The fabric is knitted in the form of a tube then split to give a cloth 2.4 metres wide which is cut to size. The polyester carrier is next removed and the catalyst is flamed with hydrogen to activate the surface. The carrier is not removed until this late stage of the process in order to give protection during storage; last minute cleaning also minimises the potential for contamination by dust. Any contaminants that are present during flaming could diffuse into the surface of the catalyst and subsequently cause problems during use.
During the carrier removal process no residues are left behind, but its composition is such that even if the gauze was installed without its removal the catalyst would still function efficiently.
Scale-up of the development process highlighted the problem of supplying the correct volume of material, at identical rates, to each of the four knitting heads, in order to obtain a cloth of uniform tension and with homogeneous density, see Figure 4. The introduction of computer control completely solved this problem. The reliability and efficiency of this knitting machine (which no longer resembles the original textile machine), together with the interest shown by customers has encouraged Johnson Matthey to purchase more machines for the production of knitted catalyst gauzes, as well as the catchment gauzes used in nitric acid production and in other similar chemical processes.
Demand for knitted gauzes from the nitric acid industry has grown rapidly and is expanding so quickly that Johnson Matthey expects 80 per cent of its customers will have converted by 1994. It is anticipated that most catalysts which are now woven will be knitted in future. Other major catalyst manufacturers are already producing or planning to produce knitted catalysts either by rotary or flat bed knitting.
In a previous article the following list of potential benefits accruing from the use of knitted gauzes was given (2):
Increased conversion efficiency
Reduced rhodium oxide formation
Greater surface area for catalysis
Fabric flexibility providing greater resistance to damage by thermal shock
Choice of alloy
Lower stock levels resulting from quicker response times
Since then two other major benefits have been observed, namely:
These benefits will now be considered in greater detail to show why the adoption of knitted gauzes has been so rapid.
There are two main types of nitric acid production plants: a single pressure process performed at either medium pressure (4 to 6 atmospheres) or at high pressure (7 to 14 atmospheres); and a dual pressure process where ammonia oxidation is carried out at medium pressure, and absorption is completed at high pressure. The overall reactions are:
The ammonia oxidation stage uses a catalyst pack containing up to 36 platinum group alloy gauzes designed to optimise the efficiency of the reaction.
The whole process is fast, but a contact time of 10-4 seconds is needed between gas and catalyst to maintain process efficiency. The catalysts have typical service lives of 50 to 300 days, depending mainly upon the pressure of the system.
Increased Conversion Efficiency
Pilot plant trials indicated that a 4 per cent increase in ammonia conversion efficiency was possible (3). Most industrial nitric acid plants claim conversion efficiencies of 96 to 99 per cent, but if these figures are accurate then a 4 per cent increase in efficiency is clearly impossible. It is, however, important to appreciate that pilot plant trials only indicate potential.
The majority of nitric acid plants measure the conversion efficiency of the overall system, not of the burner. Following the installation of knitted gauze catalyst packs, detailed production information has not been made available but, without exception, all the plants have reported no drop in conversion efficiency, and some have recorded reduced ammonia usage, which indicates an improvement in efficiency. Indeed, an increase in conversion efficiency must occur because less rhodium oxide and lower impurity levels have been recorded.
Reduced Rhodium Oxide Formation
Rhodium enrichment of a catalyst gauze occurs as a consequence of the loss of platinum from the catalyst surface. Platinum oxidises and volatilises at high temperatures in oxidising atmospheres; rhodium also oxidises, but does not volatilise as readily. When a 10 per cent rhodium-platinum alloy is in use, platinum is lost at a rate approximately 19 times faster than rhodium, so resulting in a natural enrichment of rhodium. However, if the actual catalyst is PtO2 and not platinum, then the need for replenishment and oxidation of the platinum is important.
When the initial composition of the gauze is 10 per cent rhodium-platinum, the final analysis of the bulk material can be 13 per cent rhodium-87 per cent platinum. An important factor in the catalytic process is the amount of RhO2 present at the catalyst surface. Although rhodium enrichment is regarded as a natural phenomenon other factors also can influence its formation. For instance, if the platinum-rhodium gauzes are operated at low temperatures (< 800°C) then a decline in catalyst activity can occur. The formation temperature of RhO2 is 600°C and it dissociates over the temperature range 1050 to 1150°C. Poisoning causes a loss in activity which slows the kinetic processes relative to the rate of gas diffusion transfer; thus the gauze contaminated by poison, usually the lead gauze, only allows limited diffusion. Visual evidence that this is happening is the appearance of a dull patch on an otherwise bright gauze. This low operating temperature and the subsequent increase in the oxygen partial pressure stabilises the rhodium oxide, which then rapidly accumulates at the surface. Efforts to reduce the RhO2 by raising the temperature of the gauze have only a limited success since, as stated above, the temperatures required to achieve this dissociation are high.
The structures of woven and knitted catalyst are shown in Figure 5. The more open knitted structure achieves the same weight as the closer woven structure, due to its bulky three-dimensional nature. This open structure also offers less resistance to gas flow, which enables the reaction to take place further into the gauze and ensures a more even temperature profile across the catalyst.
This improved distribution of the catalytic activity reduces the metal loss per gauze, with a subsequent drop in the formation rate of rhodium oxide. The results given in Table I indicate significant reductions in RhO2 formation over the entire range of plant pressures, when using knitted gauzes. This is particularly significant since the major cause of campaign cessation in Example 3 (a high pressure plant) was the high levels of RhO2 formation.
Another major factor contributing to the formation of rhodium oxide is the presence of contaminants, the most common of which is iron. High surface levels of iron reduce catalytic activity, promote the breakdown of nitric oxide and increase the surface concentrations of RhO2. It is generally believed that the campaign life may be extended by increasing the number of gauzes in a pack, so that contamination of the leading gauzes results in the transfer of the reaction zone to a greater depth, without causing any loss of efficiency. This does not happen, however, since the ammonia is less selectively oxidised to nitric oxide; instead more is oxidised to nitrogen as the level of contamination increases.
The levels of iron impurity on the gauzes are listed in Table I. The improved gas flow through the knitted gauzes results in less solid particles being trapped and hence less RhO2 is formed due to iron contamination. Example 1 shows that there has been no drop in iron impurity levels (0.56 per cent during a woven campaign and 0.54 per cent during a knitted campaign) but the significant factor is the greatly reduced amount of impurity present on the surface of the knitted catalysts. This phenomenon is seen in all campaigns using knitted gauzes.
When woven catalyst gauzes are removed from an oxidation rig they are often in a fragile state, once the clamping supports and weights are removed. This does not occur with knitted gauzes, indeed many plant managers have commented on the catalyst integrity at the end of a campaign.
The strength of the catalyst is important, not only for high pressure plants but also for medium and low pressure ones. Catalysts are designed to be supported horizontally, but in high pressure units the support basket often deforms to such an extent that the edge of the burner wall cuts into the catalyst, sometimes causing it to tear. This also happens in low and medium pressure units which usually employ rashing ring support. During operation the rings may migrate to the centre, causing the catalyst to “dome”. This leaves the catalyst unsupported around its edge, which tends to result in tearing.
Tensile tests were carried out to determine the strength of knitted and woven catalyst gauzes, the results being given in Figure 6. Significantly more energy can be absorbed by the knitted structure than by woven gauze, as shown by the areas under the two curves. This is further illustrated in Figure 7; Figure 8 shows almost instant tearing of the herringbone – the strongest of all the conventional weaves – which results in rapid failure. A knitted sample stretches considerably before fracture, which usually occurs at the clamp interface and not within the gauze length. Flaws of different types, such as holes and cuts, were deliberately introduced into the knitted gauze with the aim of making the fracture occur within the gauze length, but without success. A simple analogy is to try tearing a knitted jumper compared to tearing a woven shirt.
Greater Surface Area for Catalysis
The knitted gauzes have more surface area available for catalysis because of their bulky three-dimensional structure. In woven gauzes the potential area for catalysis is masked, when compared to the knitted structure. Mathematical modelling was used to estimate the surface area available for catalysis for both types of gauze, and values of 83 per cent (17 per cent masked) for woven and 93 per cent (7 per cent masked) for knitted gauzes were obtained. In knitted gauzes the gas circulates around the wire, which provides more active sites for catalysis. This increases the potential for oxidation and reduces localised metal loss (wire thinning), so the wire retains its strength, which results in a longer life.
The flexibility of the knitted structure allows the catalyst to remain essentially fluid when subjected to start-up and shut-down procedures. When in use the catalyst gauze is held rigidly in a burner, but some distortion occurs with temperature change. The catalyst will contract or expand, which may cause deformation and/or tearing at the edges; if this does not happen the gauzes may slip from under the clamp allowing ammonia to by-pass the gauze. The flexible nature of the knitted gauze enables the metal to expand and contract readily without affecting its physical and catalytic properties.
Choice of Alloy
Historically the conventional alloy for ammonia oxidation has been 10 per cent rhodium-platinum woven gauze. Before weaving can commence on a loom between 50 and 100 kg of wire must be available, whereas a knitting machine only needs 5 kg of wire. This permits a more flexible approach to alloy composition and also allows a faster response to market changes. There are now four standard alloys in production: 5 per cent rhodium-platinum, 8 per cent rhodium-platinum, 10 per cent rhodium-platinum and 5 per cent rhodium-5 per cent palladium-platinum.
Lower Stock Levels
Due to the different alloys which can be accommodated and the reduced amount of wire required for start-up, the production of large quantities of “stock” catalysts has not been required, despite the increase in the number of alloys being supplied. It also follows that a quicker response to a request for any knitted alloy is possible. Knitting can commence as soon as wire is produced. Certain non-standard alloys have been and can be produced to suit particular customer requirements.
Lower Metal Losses
The actual nature of the metal lost from catalysts has been the subject of speculation for a number of years. There appear to be two main types of metal loss: platinum volatilisation and mechanical removal. The platinum which is lost as PtO2 can be collected by installing palladium-based catchment systems, which can recover up to 95 per cent of the total platinum losses. Platinum is also lost by mechanical removal caused by vibration during plant operation, and this can be further compounded on start-ups and shut-downs. There may also be temperature differences in a catalyst bed which can result in increased losses.
The general conclusion from many studies is that volatilisation is the main contributor to platinum loss (3, 5–7). In order for platinum to be volatilised it must come into contact with oxygen molecules; therefore volatilisation is dependent upon the conversion of ammonia and hence on the production of acid. Assuming constant conversion efficiencies, the metal loss should be fixed and therefore the only possible reason for reduced metal loss must be smaller losses due to mechanical removal.
Metal losses from woven and knitted catalyst gauzes per tonne of acid produced are shown in Table II. The differences in metal loss can be partly explained by the restructuring that takes place on all gauzes, see Figure 9. The cauliflower-like formations that grow on the catalyst are very fragile, and become more so with increasing size; clearly, the more fragile the structure the greater its susceptibility to mechanical damage. With knitted gauzes there is widespread structural rearrangement throughout the catalyst pack and therefore less susceptibility to mechanical damage. This widely distributed restructuring also leads to a more even temperature distribution, and the increased structural flexibility enables it to withstand thermal shock more readily, which accounts for the reduced metal losses.
In some industrial plants, however, litde or no reduction in metal loss has been seen, and therefore it can be concluded that these plants do not experience any significant loss of metal by mechanical removal.
Extended Campaign Length
One of the main requirements in ammonia oxidation plants is long catalyst life which permits extended campaign lengths. There are three factors affecting the catalyst life: material strength, rhodium oxide formation and contamination. Improvements in all these area have been observed as a result of using knitted gauzes. Various types of plant are now running extended campaigns. Typical examples are: When ammonia oxidation plants are running they are reasonably cost effective and reliable, but shut-downs to change the catalyst are costly in terms of both money and manpower, and invariably cause other problems of a mechanical nature. For a medium pressure plant, it is estimated that having one less change in the yearly schedule can result in a saving of £60,000 on the metal costs alone.
Knitting is rapidly gaining acceptance as a gauze manufacturing technique, and material fabricated in this way is also being applied to areas other than nitric acid production. Samples of knitted platinum group metals have been supplied for various specialist applications and to industries that deal predominantly in base metals.
Many of the early claims made for knitted catalysts have now been substantiated and, indeed, many other beneficial and unexpected results have been achieved. Further developments in the knitted structure and in the alloys involved are continuing, and the future for knitted gauzes looks assured.
- 1 K. Kaiser, German Patent 271, 517 ; 1909
- 2 S. L. Handforth and J. N. Tilley, Ind. Eng. Ind., 1934 26, ( 12 ), 1287
- 3 B. T. Horner, Platinum Metals Rev., 1991, 35, ( 2 ), 58
- 4 Johnson Matthey PLC, European Patent 364, 153B
- 5 H. Connor, Platinum Metals Rev., 1967, 11, ( 2 ), 60
- 6 H. Holzmann, Chem.-Ing.-Tech., 1967, 39, ( 2 ), 89
- 7 H. Holzniann, Platinum Mertals Rev ., 1969, 13, ( 1 ), 2