Across multitubular fixed-bed reactors used for ethane oxidative dehydrogenation over MoVTeNbO catalysts, axial temperature profiles in a 25 mm inner diameter tube typically display a hot spot located between 10% and 30% of the bed length from the feed inlet. The hot spot is driven by the exothermic oxidation of ethane and ethylene, with an adiabatic temperature rise estimated at 120–180°C per mole of O₂ consumed under typical feed compositions of 6–8 vol% ethane in air. Molten salt coolant systems, usually a eutectic mixture of NaNO₂ and KNO₃, are operated with a circulation rate sufficient to limit shell-side temperature variation to ±2°C, while the tube-side skin temperature at the hot spot can exceed the salt bath temperature by 30–60°C. The magnitude of this local temperature excess is monitored with a multi-point thermowell containing 4–6 type K thermocouples positioned at axial spacings of 150–300 mm. The control objective is not a single bed temperature set point but the suppression of the hot spot below the catalyst-specific deactivation threshold: for MoVTeNbO M1 phase, the threshold lies near 460°C because TeO_x species begin to volatilize irreversibly above that temperature. In the first 20% of the bed, dilution of the catalyst with inert α-Al₂O₃ spheres of identical particle size in a 1:1 volume ratio reduces local reaction density and shifts the hot spot downstream. This dilution practice is supported by comparative fixed-bed screening in a 10 mm internal diameter quartz reactor under gas hourly space velocities of 500–1000 h⁻¹, although published data for the exact influence of dilution ratio on hot spot magnitude remains limited. Tube pressure drop across the diluted zone must be measured before and after each regeneration cycle; an increase greater than 0.1 bar at a total flow of 1500 NL/h indicates catalyst fragmentation, partial bed collapse, or fines accumulation in the outlet frit. Temperature control additionally involves a feed-effluent heat exchanger that preheats the ethane-air mixture to 280–300°C before it contacts the first catalyst layer, a range selected to avoid homogeneous gas-phase radical reactions that would otherwise initiate upstream of the catalytic bed and reduce ethylene selectivity.

Why Does Propylene Yield Over VO_x/γ-Al₂O₃ Collapse When Bed Temperature Exceeds 480°C?

In propane oxidative dehydrogenation over supported vanadia, the propylene yield passes through a maximum because the apparent activation energy for propane conversion is lower than that for propylene combustion. The selectivity to propylene declines sharply once the catalyst surface temperature exceeds 480°C because the rate of secondary propylene oxidation to acrolein and CO_x increases. Laboratory fixed-bed data for 5 wt% V₂O₅ on γ-Al₂O₃ at a C₃H₈:O₂:He molar ratio of 1:1:8 and total gas hourly space velocity 1200 h⁻¹ show propane conversion of 18–24% with propylene selectivity of 62–68% at 450°C, decreasing to 38–45% at 500°C. Published data for this specific formulation and feed ratio are limited to laboratory-scale reproducibility studies. The thermal collapse is associated with increased vanadia reducibility at higher temperature, which accelerates oxygen insertion into adsorbed propylene intermediates and increases carbon oxide production. Temperature control for this system therefore uses a narrow set point band of 440–460°C, with an inter-bed quench stream of nitrogen or recycled off-gas at 25°C injected through a sparger between catalyst layers. The quench flow is trimmed by a control valve with a maximum opening of 40% under normal operation; exceeding this position indicates either excessive bed exotherm or degraded catalyst activity that forces higher feed temperature compensation. Catalyst deactivation in this system is dominated by coking at oxygen-lean local conditions and by slow sintering of the supported vanadia phase. A used catalyst sample after 120 h on stream typically shows a loss of BET surface area from 210 m²/g to 170–180 m²/g when measured by ASTM D3663, together with a shift in Raman spectroscopy from monomeric vanadyl species to polymeric vanadia domains. The combination of surface area loss and active phase restructuring reduces the number of accessible redox sites and lowers propylene yield even when the bed temperature is returned to the original set point.

Because fluidized-bed operation disperses reaction heat through rapid particle circulation and high bed-to-surface heat transfer coefficients of 300–500 W/m²K, temperature control in n-butane oxidative dehydrogenation to butadiene over CrO_x/Al₂O₃ or supported vanadia catalysts differs from fixed-bed practice. The bed is typically maintained at 370–410°C, with a freeboard temperature 5–10°C lower than the dense bed due to post-reaction gas cooling. Differential pressure across the distributor plate is held at 20–35% of the dense bed pressure drop to achieve uniform gas distribution; a lower ratio permits bypassing and creates oxygen-rich stagnant zones that accelerate localized coking. Catalyst attrition is monitored by measuring the elutriation rate from the cyclone dipleg. An increase in fines carryover greater than 0.5 kg per tonne of catalyst inventory per day is a leading indicator of catalyst mechanical failure, requiring a reduction in superficial gas velocity from the normal 0.35–0.50 m/s to the lower end of the fluidization regime. Deactivation in this reactor type is more uniform than in fixed beds because the catalyst particles are continuously circulated through zones of different oxygen partial pressure. The high-temperature regeneration zone must be separated from the reaction zone by a loop seal with a minimum solids flux of 40 kg/m²s; insufficient solids circulation allows hot regenerated catalyst to return to the reactor with a surface temperature above 430°C, promoting nonselective oxidative cracking. Temperature control of the regenerator vessel uses a dilute-phase riser temperature of 650–700°C during coke combustion, with the regenerated catalyst cooled to 400°C in a shell-and-tube catalyst cooler before re-entry to the ODH reactor. The catalyst cooler tube-side outlet temperature is controlled by a cascaded loop that adjusts boiler feedwater flow; the water-side pressure is maintained at 40–50 bar to suppress boiling instability. This configuration is supported by operating data from circulating fluidized-bed pilot plants with a catalyst inventory of 25–50 kg, although published data for large-scale n-butane ODH units are limited.

Te Sublimation and M1 Phase Amorphization in Mixed-Oxide Beds

The distinctive deactivation pattern of MoVTeNbO mixed-oxide catalysts in ethane oxidative dehydrogenation is governed by the presence of the orthorhombic M1 phase, which contains tellurium in structural channels. At temperatures exceeding 440°C, the TeO_x component begins to sublime as Te(OH)₄ or TeO₂ depending on steam partial pressure. This loss is not uniform along the bed; it is fastest at the hot spot, where the local catalyst temperature can be 30–60°C above the nominal bath set point. The resulting axial profile of tellurium content, measured by X-ray photoelectron spectroscopy of spent catalyst sections, shows a depletion zone in the first 30% of bed length and a relative enrichment in downstream cooler sections. The loss of tellurium destabilizes the M1 framework, leading to amorphization and precipitation of MoO₃-rich domains detectable by X-ray diffraction and Raman imaging. As the M1 phase fraction declines below 40% of the total crystalline phases, ethane conversion per unit reactor volume falls, and selectivity to ethylene decreases because the remaining phases promote nonselective ethane combustion. Temperature control is therefore not simply a matter of maintaining average reactor inlet and outlet temperatures; the axial hot spot must be kept below 440°C to preserve the local tellurium inventory. In industrial practice, this is achieved by reducing the oxygen concentration in the feed gas to 5–7 vol%, increasing the steam-to-ethane ratio to 1.5–2.0 mol/mol, and increasing the salt circulation rate to reduce external heat-transfer resistance. The last measure is particularly effective when the salt Reynolds number is maintained above 10,000 in the baffled shell side, ensuring turbulent heat transfer and reducing the tube wall-to-salt temperature difference to 5–8°C. Without these measures, a sustained temperature excursion of 10°C above the threshold can reduce the catalyst Te inventory by 20–30% within 100 h, based on laboratory accelerated aging tests. Published data for complete industrial-scale Te mass balances remain limited because spent catalyst sampling from commercial multitubular reactors is not routinely reported.

Regeneration of coked VO_x/Al₂O₃ and MoVTeNbO beds is conducted as a temperature-programmed oxidation in which the oxygen concentration in the regeneration gas is ramped according to the carbon burn rate rather than as a fixed air purge. The oxidized reactor is first purged with nitrogen at 3–5 times the reactor volume per hour until the explosive limit monitor at the outlet indicates a hydrocarbon concentration below 0.2 vol%. The bed inlet temperature is increased at 20–30°C/h from the standby temperature of 200°C to 320°C under a gas mixture containing 1.0–1.5 vol% O₂ in nitrogen. At this oxygen level, the coke combustion rate is controlled by oxygen mass transfer to the coke surface, and the temperature rise across the bed is limited to 10–15°C. If the outlet CO concentration exceeds 1000 ppm, the oxygen feed is temporarily reduced to 0.5 vol% to prevent a runaway combustion front. Once the CO concentration falls below 200 ppm for a continuous period of 30 min, the oxygen concentration is raised stepwise to 3–5 vol%, and the bed temperature is increased to 400–450°C for final removal of graphitic carbon. The total regeneration time for a catalyst bed with a carbon loading of 2–5 wt% as measured by thermogravimetric analysis under ASTM E1131 is typically 12–18 h. A dual thermocouple pair is inserted at the axial midpoint of the catalyst tube to measure the combustion front temperature. The difference between the two thermocouples, one in the bed center and one at the tube wall, must not exceed 25°C during regeneration; a larger difference indicates that the combustion front is channeling through the bed or that catalyst fines have accumulated in low-velocity regions. This procedure is repeated after every 100–200 h of on-stream operation for VO_x-based propane ODH catalysts and after 300–500 h for MoVTeNbO ethane ODH catalysts. The longer interval for MoVTeNbO reflects the lower coking tendency of the ethane feed under the higher steam-to-hydrocarbon ratio used to suppress Te volatilization. In all cases, regeneration gas analyzers are calibrated against certified calibration gases using a Fourier transform infrared analyzer with a lower detection limit of 10 ppm for CO and 50 ppm for CO₂.

When Oxygen Partial Pressure at Regeneration Inlet Exceeds 3 kPa

If the oxygen partial pressure at the regeneration inlet reaches 3 kPa, corresponding to 3 vol% O₂ at atmospheric pressure, while the catalyst bed still contains a residual carbon loading above 1.5 wt%, the combustion front can accelerate beyond the heat removal capacity of the molten salt coolant. The resulting temperature excursion is most severe in the upstream 20% of the bed, where carbon accumulation is highest because the feed oxygen concentration during ODH operation was lowest at the reactor inlet due to preferential consumption. In that zone, the local temperature can rise from the set point of 380°C to 650°C in less than 5 min, as observed in pilot-scale regeneration experiments with a 25 mm internal diameter tube. The rapid temperature rise causes thermal sintering of the support, as measured by a BET surface area decrease of 30–40% per ASTM D3663, and in MoVTeNbO catalysts it induces complete loss of tellurium from the affected section. The temperature excursion also produces a sharp pressure-drop increase as the catalyst pellets fracture due to thermal stress; the pressure drop across the reactor can rise from 0.4 bar to 1.2 bar within the same period. To prevent this condition, the regeneration control system must be configured with an oxygen analyzer located immediately downstream of the gas mixing point, with a response time <5 s and an alarm set point at 2.5 vol% O₂. The analyzer signal is interlocked with the oxygen supply shutoff valve via a safety instrumented system designed to IEC 61511-1:2016. In addition, the molten salt circulation pump is interlocked to continue running during regeneration; any loss of salt flow must automatically trip the regeneration gas heater and initiate a nitrogen purge at 200°C within 30 s. This interlock sequence is validated by a HAZOP action item that assigns a target safety integrity level of SIL 2 for the oxygen shutdown loop. The exact SIL rating depends on the facility risk assessment and the maximum allowable temperature for the specific catalyst formulation, but published guidance for exothermic fixed-bed regeneration in petrochemical service commonly recommends at least SIL 2. The regeneration procedure must also limit the oxygen concentration during the initial coke burn by a restrictor orifice in the air line sized for a maximum air flow of 10% of the normal operating flow. This mechanical limit prevents an operator error or control valve failure from delivering an unacceptably high oxygen flow into a carbon-loaded bed.

Process control architecture for oxidative dehydrogenation reactors is specified around the need to manage a distributed exotherm rather than a single well-mixed temperature. In fixed-bed multitubular configurations, the primary control variable is the molten salt bath temperature, which is adjusted by controlling the fuel gas flow to a fired heater in the salt circulation loop. The bath temperature set point is ramped from 340°C at start-of-run to 400°C at end-of-run for VO_x-based systems as catalyst activity declines. This ramp rate is limited to 2°C/day to avoid thermal damage to the tube sheet and to prevent sudden changes in conversion. The individual reactor tube skin temperatures are measured with thermocouples welded to the outside of selected tubes at three axial positions: 0.25 m, 1.0 m, and 2.0 m from the tube inlet. The number of instrumented tubes in a typical multitubular reactor is 10–20 out of 2,000–5,000 tubes, depending on the tube sheet diameter. Because the uninstrumented tubes can develop hot spots without direct measurement, the reactor inlet and outlet gas analyzers are used as indirect indicators of hot spot formation. A sudden increase in the outlet CO₂ concentration of more than 0.5 vol% relative to the inlet, at constant feed composition, indicates that nonselective oxidation has increased and that a hot spot may have formed. The control system then reduces the bath temperature set point by 5–10°C and increases the steam-to-hydrocarbon ratio by 0.2 mol/mol until the CO₂ concentration returns to the expected range. The temperature transmitters used for bath control are calibrated against a dry-block calibrator with a reference probe traceable to ISO/IEC 17025; the maximum permissible error is ±0.3°C over the operating range. Control valves on the salt bypass line are specified with a deadband of 0.2% of full stroke and a maximum stroking time of 10 s to ensure that the bath temperature remains within ±1°C of set point. The combustion air and fuel gas loops for the salt heater use cross-limiting control to prevent the fuel gas flow from exceeding the air flow needed for complete combustion; the oxygen set point in the heater flue gas is 2–3 vol%. This arrangement prevents the development of a reducing atmosphere in the heater, which would cause soot deposition on the heater tubes and lower heat transfer efficiency.

Thermowell Placement Standards and Thermocouple Drift in High-Temperature ODH Service

Thermowell placement in oxidative dehydrogenation reactors must account for vibration-induced fatigue, coke deposition on the thermowell surface, and the presence of highly reactive oxygenated intermediates. A thermowell fabricated from Inconel 600 with a 6 mm outer diameter and a bore diameter of 3 mm is inserted through the reactor nozzle at an angle of 30° to the flow direction to minimize flow-induced vibration. The insertion length is set to 40–60% of the tube internal diameter in pilot-plant units and 10–15 tube diameters in commercial multitubular reactors. According to ASME PTC 19.3 TW-2016, the thermowell must be evaluated for natural frequency and Strouhal frequency to avoid resonance; the acceptance criterion is that the wake frequency remains below 80% of the thermowell natural frequency. In ODH service, the additional factor of surface coke formation changes the effective mass and heat transfer of the thermowell, which can increase the response time from 1.5 s to 4–8 s after 100 h of operation. For this reason, the thermocouple signal is filtered with a first-order digital filter with a time constant of 0.5 s during normal operation, and the signal is compared with a redundant tube skin thermocouple welded to the outside of the tube. A deviation greater than 5°C between the immersed and skin thermocouples at the same axial position triggers a calibration check or replacement of the immersed thermocouple. Thermocouple drift in type K elements is governed by chromium migration in the positive leg at high temperature; in oxidizing atmospheres above 400°C, type K thermocouples can drift by 1–3°C per 100 h due to oxidation of chromium and formation of a depletion zone. This drift is especially problematic in ODH reactors where the thermocouple is exposed to steam and carbon oxides during normal operation and to low-oxygen regeneration gas during coke burn. To reduce the effect, facilities replace immersed thermocouples after every 500–1000 h of operation or when the difference between the immersed and skin thermocouples exceeds 5°C for more than 2 h. The replacement interval is verified against the observed drift rate from a dedicated calibration thermocouple installed in a bypass well. If the drift rate exceeds 0.5°C/100 h for three consecutive calibration intervals, the thermowell material is upgraded from Inconel 600 to Inconel 601 or to a mineral-insulated type N thermocouple with a protective sheath of Nicrosil/Nisil. The type N thermocouple is less susceptible to drift in oxidizing and steam-containing atmospheres and is rated for continuous service up to 1200°C per IEC 60584-1. The limitation of type N is its higher cost and slightly lower output voltage in the 300–600°C range, which may require recalibration of the temperature transmitter. The calibration of the entire thermocouple-transmitter loop is performed at 0°C, 250°C, and 450°C using a stirred liquid bath and a reference standard resistance temperature detector with an uncertainty of ±0.05°C. The loop calibration data are stored in the plant historian with a time-stamped record and are reviewed monthly to detect slow degradation of the temperature measurement system before it affects the cascade control loop.