Factors Limiting Ammonia Conversion on Pt-Rh Gauze Catalysts

An industrial ammonia oxidation burner over a platinum-rhodium gauze pack is not a simple fixed-bed reactor; the catalytic surface itself reconstructs, volatilizes, and accumulates poisons while operating inside a narrow thermal and stoichiometric envelope. The principal oxidation reaction is 4 NH₃ + 5 O₂ → 4 NO + 6 H₂O, requiring 1.25 mol O₂ per mol NH₃, whereas the competing nitrogen and nitrous oxide routes consume 0.75 mol and 1.00 mol O₂ per mol NH₃ respectively. A production gauze pack may consist of 6–20 layers of woven or knitted Pt-Rh wire with wire diameter between 0.060 mm and 0.100 mm, open area fraction of 55–70%, and active face diameter from 1.2 m to 4.5 m. Fresh gauze exposed to 9.5–10.5 vol% ammonia in dry air at 850–950 °C can deliver ammonia conversion above 99%; however, the practical conversion limit in extended campaigns is set by surface poisoning, metal loss, maldistribution, and thermodynamic selectivity toward nitrogen rather than by the intrinsic activity of the clean alloy. Because the gauze operates at red heat under forced draft, external mass transfer through the stagnant boundary layer around each wire often becomes rate-controlling within hours after light-off, and the apparent conversion then depends more on gas velocity, wire diameter, and pack geometry than on the kinetic parameters of the uncorroded metal. Burner pressure ranges from near atmospheric to 15 bar absolute in high-pressure trains, and this alters both the equilibrium and the surface concentration of ammonia and oxygen. Platinum loss as volatile PtO₂ increases with temperature and oxygen partial pressure, while rhodium forms a stable Rh₂O₃ skin that changes surface composition from the as-manufactured alloy. These simultaneous processes make ammonia conversion a dynamic variable requiring continuous adjustment of air-to-ammonia ratio, filtration efficiency, and start-up heating rate.

What Limits Conversion When Air-to-Ammonia Ratio Falls Below the Stoichiometric Threshold?

The stoichiometric oxygen demand for NO formation is 1.25 mol O₂ per mol NH₃, but industrial burners are operated at 1.7–2.0 O₂/NH₃ because the side reactions leading to N₂ and N₂O compete more strongly when oxygen coverage is low. When the ratio falls below 1.4, ammonia conversion may remain high but selectivity to N₂ and N₂O increases markedly; if the ratio falls below 1.25, unconverted ammonia can appear in the outlet gas. This outcome is not solely an equilibrium effect. Under oxygen-lean conditions the Pt-Rh surface becomes covered with nitrogen-containing intermediates such as NHx and N adatoms, and the statistical probability of N-N recombination rises. In a low-pressure burner at 900 °C, reducing O₂/NH₃ from 1.8 to 1.3 can increase N₂ selectivity by several percentage points while ammonia slip remains below 0.5%; in high-pressure plants above 8 bar, the same reduction can produce N₂O concentrations in the tail gas that require downstream abatement. The air-to-ammonia ratio is maintained by an ammonia flow controller and an air flow controller, both compensated for temperature and pressure; a bias of more than 2% in either loop is sufficient to move the burner away from the optimum window. Because ammonia oxidation is exothermic, a reduction in air ratio also raises the adiabatic flame temperature and accelerates PtO₂ volatilization, even though oxygen partial pressure is lower. The observed conversion loss after an air compressor upset may therefore be caused by simultaneous surface reduction, thermal damage, and local ammonia breakthrough in cooler zones of the gauze pack.

Temperature-dependent surface reconstruction introduces a competing effect on ammonia conversion: higher temperature increases both ammonia adsorption/reaction rate and the vapor pressure of platinum oxide, while lower temperature preserves metal but increases N₂ and N₂O selectivity. On a clean Pt-Rh wire at 850–950 °C, surface oxidation produces a rough faceted morphology with exposed crystallographic planes that are more active for ammonia dissociation than the as-drawn smooth surface. The geometric surface area can increase by a factor of 10–100 during the first 48–120 h of operation, which explains the initial rise in conversion often observed after light-off. Simultaneously, platinum is lost as PtO₂ from the hottest regions, and the remaining surface becomes progressively enriched in rhodium as Rh₂O₃. Rhodium oxide is less volatile but also less active for ammonia oxidation on a per-site basis, so a heavily rhodium-enriched surface can show lower overall conversion after long exposure. This process is not uniform: hotter zones near the centre of the gauze show greater metal loss and surface coarsening than the cooler clamped edges, creating radial gradients in wire diameter and catalytic activity. Spent gauze inspections after 6–12 months of operation may show a loss of 5–20% of initial platinum mass depending on pressure and temperature, with the highest losses at pressures above 10 bar and temperatures above 900 °C. Published data for exact PtO₂ partial pressures over multi-layer packs under high pressure is limited, but the trend of accelerating platinum volatilization with increasing oxygen partial pressure is well documented in post-mortem gauze analyses.

Catalyst Poisoning by Sulfur, Iron, and Siloxanes Remains the Dominant Field Limitation

Feed-gas contaminants are the most common cause of premature ammonia conversion loss on production burners. Sulfur compounds, even at concentrations below 0.1 ppmv, adsorb strongly on platinum and rhodium sites, blocking ammonia oxidation and promoting N₂ formation. Iron particulates from corroded air piping or catalyst handling equipment deposit on the gauze surface as Fe₂O₃, which masks active sites and can react with platinum to form low-melting intermetallics; phosphorus compounds are particularly damaging because they create liquid phases that accelerate grain growth and embrittlement. Siloxanes from compressor oil and silicone gaskets decompose at gauze temperatures to silica, which is non-volatile and cannot be removed by acid washing; a silica film of only 50 nm can reduce ammonia conversion by several percentage points. Ammonia feedstock may also contain trace chlorides that form volatile PtCl₂ and RhCl₃ at burner conditions, causing direct metal loss instead of masking. The exact threshold for each poison depends on burner temperature, pressure, and the past history of the gauze. In practice, gas purification trains target total sulfur below 0.05 ppmv, total phosphorus below 0.01 ppmv, and oil aerosol below 0.1 mg m⁻³ at STP in the air feed. The compressed air quality standard ISO 8573-1:2010 is used to specify particulate, oil, and water classes; many nitric acid plants require at least class 2 for oil and class 2 for solid particulate upstream of the burner. The table below summarizes the main contaminant classes, typical sources, and limiting concentration bands.

When Feed-Gas Purification Is Inadequate, Conversion Drops Before Thermal Deactivation Is Observable

The best burner performance is obtained when the air and ammonia streams are treated as part of the catalyst system rather than as utilities. A compressed air system that meets ISO 8573-1:2010 class 2 for oil and particulate will still contain water vapour that must be removed by refrigerated or desiccant dryers before ammonia mixing, because liquid water droplets can hydrolyze salt deposits and create local cold spots. The air dew point should be maintained below -20 °C at operating pressure to prevent condensation in the ammonia-air mixer. If coalescing filters are not replaced on a pressure-drop schedule, oil carryover can reach 0.5 mg m⁻³, and the gauze will show a dark, dull surface within 24–48 h with corresponding ammonia conversion loss of 2–5%. A single episode of filter bypass during start-up can shorten gauze campaign life more than 1000 h of normal operation, because silica and iron oxides deposited early are not removed by later acid washing. The ammonia feedstock itself must be filtered through high-efficiency coalescers and, where sulfur is a concern, through a zinc oxide guard bed at 250–350 °C to reduce H₂S and COS to below 0.05 ppmv. Gas-phase analyzers for sulfur and oil are often validated under ASTM D5504-12 and ISO 8573-1:2010 respectively, but online analyzers cannot detect every contaminant, so regular grab sampling and spent-gauze metallography are required to confirm that purification equipment is functioning.

Gauze Pack Pressure Drop, Wire Diameter, and Flow Distribution Interactions

Pressure drop across a multi-layer gauze pack is not merely an energy penalty; it directly influences flow distribution and ammonia conversion. For a given burner diameter, a stack of 6–12 gauzes with wire diameter of 0.076 mm and open-area fraction near 60% operates with a pressure drop typically between 20 mbar and 100 mbar at superficial velocities of 0.5–2.0 m s⁻¹ at burner conditions. If one gauze experiences local fouling or metal loss, the flow redistributes to lower-resistance areas, which then operate at higher local oxygen availability and temperature, while fouled regions become ammonia-rich and cooler, creating a self-reinforcing maldistribution. This is especially relevant in burners above 3 m diameter, where the radial flow profile from the distributor is rarely perfectly uniform. Computational fluid dynamics of ammonia oxidation burners indicate that the inlet velocity coefficient of variation should remain below 5% to prevent localized ammonia slip and premature gauze failure. Wire diameter also determines mass-transfer area per unit volume: reducing wire diameter from 0.100 mm to 0.060 mm increases geometric surface area per gram and improves low-temperature conversion, but reduces tensile strength and resistance to creep. Manufacturer data for industrial gauzes indicate that the design trade-off is normally optimized at wire diameters between 0.076 mm and 0.090 mm for medium-pressure plants. Pressure drop is proportional to the square of mass velocity across the pack, so high-pressure plants with mass fluxes above 10 kg m⁻² s⁻¹ may require fewer gauze layers to avoid excessive backpressure, but this reduces residence time and can limit conversion if the feed is not perfectly mixed.

High-pressure operation above 4 bar changes the limiting factor for ammonia conversion from available active surface to contact time and product desorption. In atmospheric burners, a contact time of 10⁻³–10⁻² s through the gauze pack is normally sufficient to achieve ammonia conversion above 99% on a fresh, clean gauze, whereas in high-pressure trains the same volumetric throughput gives a much higher mass flux and a shorter residence time unless the burner diameter is increased or gauze layers are added. Raising pressure also increases the equilibrium partial pressure of N₂O and N₂ relative to NO for the competing ammonia oxidation reactions, so selectivity to NO typically falls from 96–98% at 1 bar to 93–95% at 10 bar, while N₂O formation may double. This selectivity loss is sometimes misinterpreted as ammonia conversion loss because nitrogen compounds in the tail gas are less valuable; online analyzers must distinguish unconverted ammonia from N₂ and N₂O using chemiluminescence or FTIR methods. ISO 7996:1985 or ASTM D6348-12 can be used for NOx and infrared-active compounds respectively, but ammonia slip analyzers require separate laser-based or colorimetric detection. The pressure drop across the gauze also increases with system pressure, and the supported catalyst baskets must withstand differential pressures up to 250 mbar during emergency shutdowns without sagging or tearing. In some plants, a downstream gold or palladium catchment gauze is used to recover platinum volatilized from the hotter upstream gauzes; this does not alter ammonia conversion directly but changes the economics of running at higher temperature and pressure.

Why Does Start-Up and Shutdown Thermal Cycling Produce Morphological Changes Without Incremental Catalyst Weight Loss?

During start-up, the gauze is heated from ambient to 850–950 °C in less than 30 min in many plants, which imposes severe thermal stress on the fine wire network. The coefficient of thermal expansion of Pt-Rh alloys is approximately 9 × 10⁻⁶ K⁻¹, and the wire intersections are constrained by the weave, so rapid heating produces plastic deformation at contact points. Repeated cycling causes grain growth, facet coarsening, and loss of the fine surface roughness that developed during burn-in. The result is a reduction in geometric surface area and a measurable drop in low-temperature ammonia conversion during the next campaign, even though the total gauze weight may be unchanged. This occurs because the active sites on a coarsened surface are fewer and less reactive than the fine crystallites formed during initial oxidative etching. Thermal cycling also exacerbates PtO₂ volatilization during the high-temperature part of each cycle, because oxygen partial pressure is often high during light-off before ammonia is fully admitted. If the burner is started with air only, the gauze surface is exposed to high oxygen potential at temperatures above 800 °C without the reducing ammonia species that would otherwise consume adsorbed oxygen, and this condition accelerates oxide volatilization. Production-scale guidelines therefore specify that ammonia should be introduced only after the gauze reaches the light-off temperature, and that air flow should be reduced during shutdown before the ammonia is fully purged. Published data for the exact number of cycles required to produce a given conversion loss are limited because burner history, contaminant load, and local hot-spot severity dominate; however, metallographic examination of spent gauzes after frequent start-stop operation consistently shows larger elongated grains and fewer etch pits than gauzes from continuous runs of equivalent total hours.

Ammonia conversion measurement on a production burner is complicated by the presence of water, nitric acid mist, and changing gas density due to temperature and pressure fluctuations. The standard control strategy uses an ammonia slip analyzer downstream of the gauze, an oxygen analyzer, and a pressure-differential transmitter across the pack. A slip value below 50 ppmv is generally acceptable, but excursions above 200 ppmv indicate either severe maldistribution, poisoning, or insufficient oxygen. The burner outlet gas is also monitored for NO and NO₂ by chemiluminescence under ISO 7996:1985, and the measured NOx concentration is used to compute yield. Because ammonia slip is often a more direct indicator of conversion than NOx yield, laser-based near-infrared analyzers are installed in many high-pressure plants to provide a response time below 1 s. Differential pressure across the gauze pack is an indirect measure of wire thinning and fouling; a gradual increase from 50 mbar to 80 mbar at constant throughput often indicates iron or silica deposition, while a gradual decrease may indicate metal loss and open-area increase. The pressure measurement is meaningful only if the mass flow is compensated for temperature and pressure, and if the tap points are located sufficiently far from the gauze to avoid wake effects. Control of the air-to-ammonia ratio is normally set by a ratio controller with a fixed bias based on online ammonia concentration; the bias may be adjusted when the catalyst ages or when feed gas composition changes. Because the response of ammonia conversion to pressure, temperature, and oxygen ratio is nonlinear, a multivariable control strategy with burner temperature constraints and ammonia slip limits is preferable to single-loop control in high-pressure plants.