Exothermic Spots in Molecular Sieve Bed: Runaway Adsorption Heating and Carbon Lay-Down Prevention

Engineering guide to exothermic spots in molecular sieve beds: runaway adsorption heating, multi-point thermocouple monitoring, TSA regeneration ramp rate control, carbon lay-down root cause analysis, and procurement specification for thermal-stable molecular sieve grades. Targets natural gas processing engineers, hydrogen PSA operators, refinery gas dryer technicians, and procurement teams specifying adsorbent beds for high-temperature service.

Why Exothermic Spots Are the Most Under-Diagnosed Failure Mode in Molecular Sieve Service

The standard molecular sieve adsorption cycle is inherently exothermic — water releases approximately 1,800 to 2,200 kJ per kg adsorbed onto a 4A or 13X zeolite at standard conditions, and CO2 releases another 1,500 kJ per kg. In a properly designed natural gas dryer or hydrogen PSA bed, this heat is carried out by the gas stream as the heat front moves down the bed at a steady velocity of 5 to 15 cm per hour, with a peak temperature rise of 15 to 35 deg C above inlet. Operators see this as a normal signature on the bed thermocouple tree and respond to it as part of the cycle timing. The dangerous failure mode is the runaway exotherm — a localized hot spot that reaches 150 to 350 deg C, drives coking of heavy hydrocarbons, sinters the zeolite framework, and permanently reduces bed capacity by 15 to 70% in a single event.

Across natural gas processing, hydrogen PSA, ethylene dehydration, and refinery gas dryer service, runaway exotherms account for an estimated 25 to 35% of unplanned bed change-outs in mature operating fleets. The cost of a single exotherm event is USD 100,000 to 500,000 in lost production, USD 50,000 to 200,000 in emergency sieve replacement, and in the worst documented cases (ethylene unit cracker gas dryers, LNG feed gas treaters) an extended plant shutdown that costs the operator USD 1 to 5 million per day. The single largest contributing factor in post-event audits is inadequate bed temperature monitoring — most failed beds had either too few thermocouples, or thermocouples that were not trended cycle-over-cycle, so the early warning signs of rising peak temperature were missed until the exotherm reached trip level. This article walks through the heat-of-adsorption mechanism, the thermocouple monitoring standard, the regeneration ramp rate that minimizes carbon lay-down, and the procurement specification that protects against thermal-stable grade variability.

The Heat-of-Adsorption Mechanism — Why Beds Run Hot

The differential heat of adsorption of water on 4A molecular sieve at low coverage (below 5 wt%) is approximately 80 to 90 kJ per mol of water, dropping to 50 to 60 kJ per mol at higher coverage as the most energetic sites fill first. The integrated heat release for a typical 4A bed taking a natural gas stream from 100 ppm H2O inlet to 1 ppm H2O outlet is 1,800 to 2,200 kJ per kg of water adsorbed. For a 5 MT bed processing 50 MMscfd of natural gas at 50 bar with 100 ppm inlet water, this translates to a heat release rate of 35 to 45 kW sustained over the 8 to 12 hour adsorption half-cycle, most of it concentrated in the mass-transfer zone at the leading edge of the heat front.

In a normal cycle, the heat front moves down the bed at a velocity governed by the gas superficial velocity and the mass-transfer zone width — typically 5 to 15 cm per hour for a 4A natural gas dryer at standard superficial velocity. The peak temperature in the heat front is 15 to 35 deg C above inlet gas temperature, and the trailing edge of the heat front is sharp because the zeolite is fully saturated and no more water is being adsorbed downstream of the front. The bed below the heat front remains at inlet temperature because it is dry (no adsorption) and the gas has cooled to inlet temperature by the time it passes through the saturated zone. This is the signature of a healthy cycle, and operators learn to read it as a normal sawtooth on the bed thermocouple trace.

How to Distinguish a Normal Heat Front from a Runaway Exotherm

The diagnostic distinction between a normal heat front and a runaway exotherm rests on three observable signatures. The first is temperature profile shape. A normal heat front is a moving peak that travels down the bed at steady velocity; a runaway exotherm is a stationary or upstream-moving peak. The second is peak temperature. A normal heat front peaks at 15 to 35 deg C above inlet; a runaway exotherm reaches 150 to 350 deg C in a 4A bed or 200 to 400 deg C in a 13X bed. The third is cycle-over-cycle trend. A normal heat front is reproducible cycle to cycle; a runaway exotherm trends upward cycle over cycle as coke accumulates and the bed becomes progressively less able to dissipate the heat.

The most informative diagnostic in field service is the cycle-over-cycle trend of peak bed temperature. Even when absolute peak temperature is within the normal range (under 50 deg C above inlet), a rising trend of 5 to 10 deg C per cycle over 20 to 50 cycles is the leading indicator of coke accumulation or other exothermic process. Operators who track this trend in their DCS historian catch the runaway in its early stages, when the bed can be taken offline, regenerated under controlled oxidative conditions, and returned to service with 70 to 90% capacity recovery. Operators who do not track the trend discover the runaway only when a single cycle exceeds the trip-level alarm at 200 deg C, by which point the bed damage is irreversible.

Multi-Point Thermocouple Monitoring Standard

Industry practice for natural gas dryers and hydrogen PSA beds is 5 to 7 thermocouples distributed along the bed height: one at the inlet (top) for inlet temperature reference, three to four at 25%, 50%, 75%, and 90% of bed depth to track the heat front position, and one at the outlet (bottom) for outlet temperature. For larger diameter beds above 2 meters, also add 2 to 3 radial thermocouples at mid-height because radial heat distribution can vary by 10 to 20 deg C across the cross-section, and an exothermic spot near the vessel wall can be missed by a centerline-only thermocouple tree.

Thermocouple type and installation matter as much as count. Type K (chromel-alumel) sheathed thermocouples with 316L stainless steel or Inconel 600 sheaths are the standard for molecular sieve service because they tolerate the 320 deg C regeneration ceiling and the reducing atmosphere. Sheath diameter of 3 to 6 mm provides adequate mechanical strength for installation in a packed bed; smaller diameters bend too easily during loading and larger diameters create excessive dead volume. Thermocouple wells should extend at least 50 mm into the bed cross-section to avoid measuring the wall boundary layer; a thermowell that terminates at the vessel wall reads wall temperature, not bed temperature, and will miss an exothermic spot at the centerline by 30 to 60 deg C. Aluminaworld publishes a recommended thermocouple tree drawing for each standard bed diameter from 0.5 m to 3.0 m.

Regeneration Ramp Rate — The Most Common Root Cause of Carbon Lay-Down

Carbon lay-down is the thermal cracking of heavy hydrocarbons (C6+) on the acidic sites of the zeolite at temperatures above 200 to 250 deg C, leaving a polymerized carbon residue that permanently reduces capacity. The most common root cause is regeneration ramp rate that is too high — a fast temperature ramp from the bed at adsorption-end temperature (often 50 to 80 deg C) to the regeneration setpoint (250 to 290 deg C) creates a localized hot zone near the regeneration gas inlet that briefly reaches 280 to 320 deg C, hot enough to crack any heavy hydrocarbon present on the sieve. The cracked hydrocarbon forms coke that cannot be removed by the standard regeneration gas, and the bed capacity drops cycle-over-cycle as more coke accumulates.

The mitigation is a controlled regeneration ramp rate. The standard ramp rate is below 30 deg C per minute, with 20 deg C per minute being a more conservative target for beds that have shown any sign of capacity decay. For beds with measured C6+ in the feed above 50 ppmv, the ramp rate should be limited to below 15 deg C per minute, and a regeneration gas pre-heater with a flow-rated capacity check at every quarterly inspection is required. Some operators add a 60-minute dwell at 150 deg C during the ramp to volatilize and purge heavy hydrocarbons before the temperature reaches the coking window — this dwell extends total regeneration time by 30 to 60 minutes but extends bed life by 30 to 60% in coking-prone service.

Carbon Lay-Down Chemistry and the Acid-Site Connection

The acid sites that catalyze hydrocarbon cracking during regeneration are the same sites that give 13X its high CO2 and water capacity — extra-framework aluminum atoms and silanol nests at the crystal surface. 4A has fewer acid sites than 13X because of its higher aluminum content and the resulting higher sodium loading, but even 4A has enough acid sites to crack a heavy hydrocarbon if the temperature is high enough. The acid-site density is controlled in part by the silica-to-alumina ratio of the zeolite — higher silica grades (SiO2/Al2O3 above 2.5 for 13X) have fewer acid sites and are more thermal-stable, at the cost of 10 to 15% lower equilibrium capacity. For hydrogen PSA service, demand controlled acid-site density by NH3-TPD, with total acid sites below 0.1 mmol/g to minimize hydrocarbon cracking activity.

Once carbon is laid down, capacity restoration requires controlled oxidative regeneration. The standard procedure is to heat the bed to 300 deg C in dry N2, then introduce 1 to 2% O2 in N2 at 350 to 400 deg C and hold for 12 to 24 hours while monitoring outlet CO2 and temperature. The exothermic combustion of coke can raise bed temperature by 30 to 60 deg C if O2 is introduced too quickly, so O2 concentration must be ramped from 0.5% to 3% over 4 to 6 hours. After oxidative regeneration, switch back to dry N2 and bring the bed to the normal regeneration temperature (290 deg C for 13X) for the final desorption step. Capacity recovery is typically 70 to 90% of fresh bed, with the residual 10 to 30% loss attributable to the most strongly bound carbon and to minor thermal damage from the elevated regeneration temperature.

Maximum Safe Regeneration Temperature and Alarm Setpoints

4A molecular sieve is stable to approximately 600 deg C in dry atmosphere; above 350 deg C the sodium cations begin to migrate and the pore structure partially collapses. The practical regeneration ceiling is 320 deg C to leave safety margin. 13X (sodium form) is stable to approximately 600 deg C but with more rapid capacity loss above 300 deg C than 4A. The practical regeneration ceiling for 13X is 290 deg C. Both grades should never exceed 350 deg C in service, and a high-temperature alarm at 320 deg C is standard practice for natural gas dryer and hydrogen PSA service.

The standard alarm setpoints for a 4A natural gas dryer operating with inlet gas at 25 to 40 deg C are: high-high temperature alarm at 200 deg C (trip-level, immediate shutdown); high temperature alarm at 150 deg C (operator notification, investigate within 1 hour); temperature differential alarm at 30 deg C above inlet (early warning of abnormal heat generation). For 13X service, add 20 to 30 deg C to each setpoint because of the higher normal adsorption temperature rise. The high-high alarm should trigger an automatic regeneration cycle with no feed flow, not an immediate bed cooldown — a fast cooldown on a hot bed can cause thermal shock and bead fracture, which produces fines that cause pressure drop problems downstream.

Procurement Specification for Thermal-Stable Service

For thermal-stable molecular sieve service, demand the following on the COA: (1) thermal stability test result (5 cycles between 25 deg C and 320 deg C in dry air, capacity loss less than 3%); (2) attrition index less than 0.2 wt% by ASTM D4058; (3) sodium content by ICP-OES for 4A grade (Na2O 16 to 18 wt% for the sodium form; deviation suggests binder contamination); (4) particle size 4x6 or 8x12 mesh with less than 1.0 wt% fines; (5) total carbon content less than 0.05 wt% (pre-coke test). For hydrogen PSA service also demand controlled acid-site density by NH3-TPD, with total acid sites below 0.1 mmol/g to minimize hydrocarbon cracking activity.

Beyond the standard COA, demand a thermal-shock test result. The sieve is cycled 5 times between 25 deg C and 320 deg C at 30 deg C per minute and analyzed for fines generation — a passing lot generates less than 0.1 wt% additional fines. Both tests are inexpensive and catch lot-to-lot variation in binder quality that the standard COA does not detect. Aluminaworld publishes lot-level thermal stability and thermal-shock data on every 13X and 4A shipment bound for natural gas and hydrogen service.

Operational Data Summary — Quick Reference

The table below summarizes the alarm setpoints and operating limits referenced in the previous sections. All values are industry-typical for natural gas dryer and hydrogen PSA service with 4A or 13X molecular sieve, and should be used as a starting point for operating procedure development, not as a substitute for site-specific hazard analysis.

Common Mistakes to Avoid

Six mistakes account for most runaway exotherm events in molecular sieve service. The first is installing fewer than 5 thermocouples in the bed — a 2-thermocouple configuration (inlet and outlet only) cannot distinguish a normal heat front from a runaway exotherm because the runaway peak is between the thermocouples. The second is not trending peak bed temperature cycle-over-cycle in the DCS historian — operators see the absolute temperature but miss the rising trend that is the leading indicator. The third is regeneration ramp rate above 30 deg C per minute — the localized hot zone near the regeneration gas inlet exceeds the coking threshold even though the bulk bed temperature is within the safe range. The fourth is omitting the upstream C6+ removal step when feed gas is refinery off-gas or natural gas with entrained compressor oil — a single upset that pushes C6+ above 500 ppmv will coke the bed within 5 to 10 cycles.

The fifth mistake is selecting standard 13X rather than thermal-stable 13X for high-temperature service above 280 deg C — the standard grade has higher acid-site density and cokes 2 to 3 times faster than the thermal-stable grade. The sixth mistake is failing to specify acid-site density on the procurement COA — the standard COA does not include NH3-TPD acid-site data, and the lot-to-lot variation in acid-site density between suppliers and between production batches of the same supplier can be a factor of 2 to 3, which is the difference between a bed that runs 5 years and a bed that runs 18 months before requiring oxidative regeneration.

Acid-Site Density by Molecular Sieve Grade — Procurement Reference

The acid-site density by NH3-TPD is the single most informative COA parameter for thermal-stable service but is not part of the standard sieve certificate. The table below summarizes typical acid-site density ranges across commercial molecular sieve grades. Demand lot-level data for any sieve bound for service above 250 deg C or for any hydrogen PSA service where trace hydrocarbon carryover is a risk.

Next Steps — Get a Sample or Talk to Our Engineers

For thermal-stable 4A or 13X molecular sieve samples sized for a specific natural gas dryer or hydrogen PSA installation, COAs with lot-level thermal stability and NH3-TPD acid-site data, or guidance on thermocouple tree design and alarm setpoint selection, reach out to the Aluminaworld technical team. We hold stock of thermal-stable 4A and 13X grades in our Zibo warehouse with same-week dispatch for pilot orders (minimum 25 kg), and we can also produce custom acid-site-controlled grades for high-temperature hydrogen PSA service.

Reach our engineering team directly via WhatsApp at +86 13325222240 (wa.me/8613325222240) or by email at barry@aluminaworld.com. Sample requests are normally dispatched within 48 hours of receipt of a complete RFQ; production orders follow 15 to 25 days after sample approval and receipt of purchase order.