Articles
Articles
Articles
In fixed-bed vapor-phase production of 1,1,1,2-tetrafluoroethane from trichloroethylene and anhydrous hydrogen fluoride, the overall stoichiometry consumes 4 mol HF per 1 mol TCE and releases 3 mol HCl. The process is operated in multiple heat-transfer-controlled multitubular reactor stages, with the first stage favoring conversion of trichloroethylene to 1-chloro-2,2,2-trifluoroethane and the second stage favoring conversion of HCFC-133a to HFC-134a. Industrial fixed beds are charged with chromium oxide or prefluorinated chromium oxyfluoride extrudates of 3.2 mm to 5.0 mm diameter, having a mean crush strength above 2.5 MPa and a fresh surface area between 120 m²/g and 250 m²/g. The catalyst is loaded into tubes with internal diameters of 25 mm to 50 mm and tube lengths of 6 m to 12 m, using a moderate length-to-diameter ratio to control radial temperature gradients. Molten salt or a high-boiling heat transfer fluid is circulated on the shell side at a set point that is progressively increased as the aging bed loses activity, but the metal wall temperature must remain below the limits for hydrogen fluoride service in Alloy 20, UNS N06625, or UNS N10276 components per ASME B31.3 and the relevant corrosion allowances. Run time limits are therefore a function of multiple simultaneous constraints: the maximum allowable shell-side temperature, the hot spot temperature in the upper bed, the differential pressure across the catalyst bed, the concentration of unconverted HCFC-133a that can be tolerated in downstream distillation, and the carbon content accumulated on the catalyst. A fresh bed may exhibit a trichloroethylene conversion above 95% at a salt bath temperature near 260°C, whereas the same bed near the end of a production campaign may require a salt bath temperature above 320°C to maintain similar conversion, at which point the reactor metal temperature approaches the service ceiling and the run must be terminated. Published technical literature indicates that campaign lengths between regeneration events commonly fall between 3,000 h and 8,000 h, depending on feed purity, operating severity, and the regeneration protocol.
The first fixed-bed stage is exposed to a mixed feed of vaporized trichloroethylene, hydrogen fluoride, recycled HCFC-133a, and heavy by-products. The gas hourly space velocity, based on total vapor flow at standard conditions and the settled bed volume, is normally held between 250 h⁻¹ and 800 h⁻¹. The heat release in the first bed is concentrated in the upper 20% of the catalyst mass, where olefin addition and chlorine displacement reactions occur. Axial thermowells inserted at intervals of 0.5 m to 1.0 m monitor the traveling reaction zone. Because the active chromium oxyfluoride phase is formed during an initial activation step that fluorinates the fresh chromia surface, the run time clock begins only after the bed has been conditioned with 5 mol% to 15 mol% hydrogen fluoride in nitrogen at 250°C to 350°C for 8 h to 24 h. This activation sequence replaces surface oxide and hydroxyl groups with fluoride and creates a layer of chromium oxyfluoride that exhibits strong Lewis-acid character. Run time limits in the first reactor are frequently controlled by coke precursors generated from chlorinated alkenes when local temperatures exceed 330°C, while the run time limits in the second reactor are more often controlled by the loss of strong Lewis-acid sites required for the slower substitution of the third chlorine atom. Therefore, a single total run time value cannot be assigned to the entire plant; the first and second beds often require staggered regeneration schedules based on different deactivation rates and different tolerance for intermediate recycle.
The selectivity decline observed after roughly 4,000 h is not governed by a single deactivation mechanism. The primary effects include the conversion of labile surface oxyfluoride sites to less active hydroxyfluoride or oxide domains, the accumulation of high-molecular-weight chlorinated residues at pore mouths, and the redistribution of active chromium species from the pellet surface into the bulk lattice. In the first reactor, a modest loss of total conversion can still be compensated by raising the salt bath temperature, but the selectivity pattern changes because the desired product HCFC-133a can undergo secondary dehydrochlorination on strong acid sites, producing unsaturated intermediates that are potent coke precursors. In the second reactor, the selectivity decline is observed as an increase in HCFC-133a breakthrough and a corresponding loss of HFC-134a product. The average HCFC-133a concentration in the second-stage effluent may be below 0.2 mol% on a fresh catalyst bed but can rise to 0.5 mol% to 1.0 mol% after extended service, requiring larger recycle streams and higher reboiler duties in the distillation train. The apparent activation energy for the overall trichloroethylene conversion over fresh chromium oxyfluoride is often reported in the range of 60 kJ/mol to 90 kJ/mol, while the second-stage substitution of the final chlorine atom has a higher barrier, typically 80 kJ/mol to 120 kJ/mol. As the catalyst ages, the observed temperature sensitivity changes because pore diffusion restrictions increasingly mask the kinetic regime, making the reactor less responsive to temperature adjustments and more vulnerable to run-time termination based on product specification drift.
The formation of carbonaceous deposits on chromia-based fluorination catalysts follows a time-dependent accumulation pattern that can be described by a Voorhies-type relationship of the form C = a tn, with n often between 0.5 and 1.0 for gas-phase chlorinated feedstocks. Fresh catalyst may contain less than 0.5 wt% carbon after the first 500 h of operation. By mid-campaign at 2,000 h to 4,000 h, carbon loadings in the top bed can reach 1.0 wt% to 2.0 wt%, and by the end of a prolonged run local carbon concentrations in the hottest zone may exceed 3.0 wt%. These deposits do not simply block active sites; they also restrict pore diffusion, increase the bed pressure drop, and create local exotherms during regeneration if not burned under controlled oxygen concentration. Selectivity loss is further amplified by the presence of iron, nickel, or chromium species carried into the bed from reactor walls and upstream piping. These metal contaminants can act as dehydrogenation catalysts and increase the yield of tars and chlorinated oligomers. Supplier data for chromium-based fluorination catalysts generally recommend limiting iron in the feed to less than 1 ppm and limiting total metals to less than 5 ppm to avoid accelerated coke formation. The run time limit in this stage is thus a selectivity limit rather than a total conversion failure: the reactor may still convert the organic feed, but the product distribution can no longer meet the required intermediate purity without excessive purification costs. Product-grade HFC-134a is specified under ISO 817:2014 and AHRI 700-2019 with maximum acceptable concentrations of chlorinated impurities, and failure of the distillation system to remove rising HCFC-133a levels effectively terminates the fluorination campaign.
The most direct operational boundary for extending fixed-bed fluorination run time is the exclusion of moisture, oxygenated stabilizers, and high-boiling chlorinated impurities from the feed. Anhydrous hydrogen fluoride used in HFC-134a production should meet a water specification of less than 100 ppm w/w, because water reacts with both the catalyst and the construction materials to produce metal oxides and additional HCl. Trichloroethylene feedstock is more demanding. Vapor-degreasing-grade trichloroethylene containing ketone, ether, or epoxide stabilizers is not suitable for fluorination catalyst beds; the oxygenated compounds decompose on strong acid sites to form carbon monoxide, carbon dioxide, water, and non-volatile carbonaceous residues. Even a stabilizer concentration of 100 ppm oxygen equivalent can shorten the interval between regeneration cycles by an estimated 30% to 50% relative to an uninhibited feed with 20 ppm oxygen equivalent. Moisture in trichloroethylene above 50 ppm w/w hydrolyzes the aluminum fluoride support phase, weakens pellet crush strength, and promotes fines formation. High-boiling chlorinated impurities such as pentachloroethane or hexachloroethane accumulate in the lower bed and act as solvent sinks that degrade conversion, while acidity in the trichloroethylene feed above 10 ppm as hydrochloric acid accelerates corrosion product carryover into the catalyst bed. These limits are generally monitored by on-line analyzers and laboratory methods such as ISO 760 for water, ASTM D6806-20 for solvent purity by gas chromatography, and ion chromatography for halide and sulfate species.
| Stream | Impurity | Maximum practical limit | Observed run-time impact if exceeded | Reference method |
|---|---|---|---|---|
| Anhydrous hydrogen fluoride | Water | 100 ppm w/w | Hydrolysis of active oxyfluoride sites, increased HCl, metal fluoride scaling | ISO 760 |
| Trichloroethylene | Water | 50 ppm w/w | Support degradation, pellet softening, pressure drop increase | ISO 760 |
| Trichloroethylene | Oxygenated stabilizers | 20 ppm oxygen equivalent preferred | Coke deposition, early hot spot formation, selectivity loss | ASTM D6806-20 |
| Trichloroethylene | High-boiling chlorinated impurities | 200 ppm w/w | Lower bed fouling, distillation load increase | GC-FID |
| Trichloroethylene | Acidity as HCl | 10 ppm | Corrosion product metals deposit on catalyst, increased coke | ASTM D1613 |
| Anhydrous hydrogen fluoride | Sulfate | 50 ppm | Acid-site neutralization, lower fluorination activity | Ion chromatography |
Feed purification ahead of the fixed-bed reactors typically includes a desiccant dryer for trichloroethylene, an acid wash or coalescer for water removal from hydrogen fluoride, and a heavy-ends column for chlorinated high boilers. The upper temperature limit of the purification train is set by trichloroethylene's tendency to dehydrochlorinate; therefore, reboiler skin temperatures are normally kept below 130°C and residence times are minimized. Published data for the quantitative effect of specific stabilizer chemistries on chromium oxyfluoride deactivation remains limited, but the direction of the effect is confirmed by the difference between solvent-grade trichloroethylene and fluorination-grade trichloroethylene in commercial operation. Facilities that replace fluorination-grade trichloroethylene with vapor-degreasing-grade material frequently experience an increase in the rate of carbon accumulation and a reduction in the interval between regeneration from approximately 6,000 h to below 3,000 h. Since coke is concentrated in the top bed, increased feed purification capacity can extend the first-bed run time and decouple the two reactor stages.
Regeneration of a fixed-bed fluorination catalyst becomes necessary when either the differential pressure reaches the blower margin or the product selectivity drops below the acceptable intermediate purity for recycle. The regeneration sequence consists of purging with nitrogen, oxidizing the carbonaceous deposits with diluted air, and then refluorinating the catalyst surface with hydrogen fluoride before hydrocarbon feed is reintroduced. The oxidation step is the most thermally sensitive part of the run time cycle. Chromium oxide and chromium oxyfluoride catalysts are susceptible to irreversible sintering if the oxidation exotherm is permitted to exceed 420°C. Dilute air oxidation is therefore started with an oxygen concentration of 1 vol% to 2 vol% in nitrogen and gradually increased to 5 vol% only after the temperature front has passed through the bed. The superficial gas velocity during coke burn is typically maintained between 0.10 m/s and 0.30 m/s to remove the heat released from carbon combustion, which can be estimated at 32 MJ/kg to 35 MJ/kg of carbon. If the carbon loading is above 2.0 wt%, the temperature rise from the oxidation front can exceed 40°C at the front if oxygen is introduced too rapidly, and the hot spot may exceed the 420°C surface stability limit. At temperatures above this threshold, surface area loss of more than 30% can occur within a single regeneration cycle, and chromium oxyfluoride can volatilize as oxyfluoride species, permanently reducing the number of active sites.
After carbon removal is complete, the catalyst is refluorinated with a mixture of 5 mol% to 15 mol% hydrogen fluoride in nitrogen at 250°C to 350°C for 8 h to 24 h. This step restores the oxyfluoride surface but does not fully recover the original surface area or pore size distribution. Each regeneration cycle typically reduces catalytic activity by 5% to 15%, and the run time available before the next regeneration is often shorter than the first campaign. The cumulative effect of regeneration cycles is a staircase decline in the maximum achievable conversion at a given salt bath temperature. The practical end of catalyst life occurs when the catalyst can no longer achieve the required HCFC-133a conversion at the maximum allowable salt bath temperature, or when the regenerated bed pressure drop remains above the blower capability because fines have filled the void space. The regeneration gas should be vented through a scrubber and thermal oxidizer designed for HCl, HF, carbon monoxide, and trace chlorinated dioxins or furans; the outlet gas composition is monitored for CO and CO₂ to determine when carbon oxidation is complete. Total regeneration time, including purge, heat-up, oxidation, refluorination, and cool-down, normally requires 24 h to 72 h, during which the reactor is unavailable for production. Run time limits are therefore not solely a catalyst property; they are also a scheduling and maintenance limit that depends on whether the plant has spare reactor capacity or can tolerate a staged shutdown.
Unlike a homogeneous catalytic reaction, fixed-bed fluorination run time is frequently terminated by a rise in differential pressure across the catalyst bed rather than by a failure of the reaction chemistry. The fresh bed pressure drop in a multitubular reactor is typically 0.05 MPa to 0.12 MPa at design gas flow. Over the course of a campaign, coke deposition, pellet fragmentation, and fouling from corrosion products increase the pressure drop to 0.25 MPa or higher, and the upper practical limit is generally set at 0.35 MPa to 0.45 MPa by the feed compressor or product compressor suction conditions. The pressure drop rise is not linear; a small increase in fines content below 1.0 mm can produce a disproportionately large increase in pressure drop because the fines migrate into the void spaces between intact extrudates. Catalyst pellets with a fresh mean crush strength above 2.5 MPa may lose 40% to 60% of their crush strength after repeated thermal cycles due to phase changes in the binder and hydrolysis of the support. When the pellet crush strength falls below 1.5 MPa, the rate of fines generation accelerates and the bed may compact by 5% to 15% of its original settled height. This compaction creates a hot zone near the top of the bed where flow distribution is poor and where coke deposition is most severe. To reduce particulate carryover into the second reactor, plants install guard beds or wire mesh screens above and below the catalyst bed; nevertheless, the pressure drop across these protective layers also rises with service time. The combined effect of pressure drop rise and bed compaction often forces a shutdown before the catalyst has lost all chemical activity.
| Operating variable | Fresh bed | Mid-run, 2,000–4,000 h | End-of-run threshold |
|---|---|---|---|
| Salt bath temperature set point | 250°C to 270°C | 280°C to 300°C | 320°C to 340°C termination |
| Top-bed hot spot | 280°C to 300°C | 310°C to 330°C | 350°C maximum before tar formation |
| Trichloroethylene conversion, first stage | 95% to 98% | 90% to 95% | 85% to 90% depending on recycle cost |
| HCFC-133a in second-stage effluent | 0.1 mol% to 0.2 mol% | 0.5 mol% to 1.0 mol% | 1.5 mol% termination |
| Differential pressure across bed | 0.05 MPa to 0.12 MPa | 0.15 MPa to 0.25 MPa | 0.35 MPa to 0.45 MPa termination |
| Carbon on catalyst, top bed | 0.1 wt% to 0.5 wt% | 1.0 wt% to 2.0 wt% | 2.5 wt% to 3.5 wt% regeneration required |
The table reflects a typical chromium-based catalyst campaign rather than a universal limit. The exact termination point depends on the installed compressor suction pressure, the regeneration capacity, and the product purity required by ISO 817:2014 and AHRI 700-2019. Operators often prolong a campaign by reducing throughput, increasing the hydrogen fluoride-to-organic ratio, or lowering the maximum bed temperature, but each of these adjustments has economic consequences. Reducing gas load lowers the differential pressure but also reduces plant output; increasing the hydrogen fluoride ratio improves heat removal and suppresses coke formation, but increases the energy consumption of the hydrogen fluoride recovery column. These trade-offs are made on a continuous basis by the production engineer, using on-line pressure transmitters and multi-point thermocouple arrays. The run time limit is therefore a moving boundary that reflects not only catalyst condition but also the downstream separation train's ability to remove recycled intermediate and by-products.
The ability to maintain a stable axial temperature profile is a prerequisite for extending fixed-bed fluorination run time. In the first reactor, the reaction zone is deliberately kept near the inlet of each tube to minimize the residence time of hot trichloroethylene and unsaturated intermediates. The bed is equipped with multiple axial thermowells, typically placed at intervals of 0.5 m along the length, and the shell-side heat transfer fluid is divided into upper and lower zones to manage the temperature profile. A fresh catalyst bed usually shows a sharp temperature rise in the top 20 cm to 40 cm of the tube, with a hot spot that is 20°C to 30°C above the salt bath set point. As the catalyst ages, the hot spot moves downward, and the temperature difference between the hot spot and the salt bath narrows because the reaction front becomes more diffuse. If the hot spot exceeds 350°C, trichloroethylene dehydrochlorination produces dichloroacetylene and other highly reactive chlorinated alkynes that readily polymerize on acid sites, causing a rapid stepwise increase in coke and a corresponding loss of selectivity. The process control window for the top bed is often set at ±5°C around the desired hot spot target, because a low deviation below the target decreases conversion and a high deviation above the target accelerates coking. Distributed control systems execute cascade control loops in which the salt bath temperature and the organic feed flow are adjusted based on the measured hot spot, with the maximum temperature deviation from set point logged as a critical process alarm.
The migration of the hot spot down the tube is a diagnostic indicator of catalyst aging. In a new bed, the upper 30% of the catalyst inventory performs the majority of the fluorine addition. As the top bed deactivates, the reaction front moves into the middle section of the tube, and the downstream section then experiences the full heat and mass transfer burden. The axial temperature profile flattens, and the exit gas temperature may rise by 5°C to 15°C compared with the fresh bed profile. In some plant designs, the shell side is divided into three heat transfer zones so that the middle and lower zones can be operated with independent set points; this allows operators to preheat or cool the catalyst as the reaction zone moves. Even with such zoning, there is a practical limit: when the hot spot reaches a distance corresponding to 60% to 70% of the total bed length, the remaining catalyst volume is insufficient to complete the conversion, and the run must be terminated or the feed rate must be reduced substantially. The required processing window of ±5°C is difficult to hold in a bed with an aged, diffuse reaction front because the transient response time of the shell-side heat transfer fluid is often on the order of 5 min to 20 min, while the gas residence time in the catalyst bed is only 2 s to 10 s. This mismatch between the fast gas-phase response and the slow heat-removal loop means that small perturbations in feed composition or flow can produce short-term hot spot excursions that exceed the target. The run time limit is thus set not only by the steady-state temperature profile but also by the dynamic controllability of that profile as the reaction front approaches the lower bed sections.
Temperature diagnostics are supplemented by on-line analysis of the reactor effluent. An increase in carbon monoxide and carbon dioxide in the exit gas can indicate that oxygenated impurities are decomposing on the acid sites and generating water, which then participates in hydrolysis reactions. An increase in unsaturated chlorinated by-products, measured by gas chromatography, indicates that the hot spot has become too high or that the residence time in the high-temperature zone is too long. These analytical signals are integrated into the run time management system: if the concentration of carbon monoxide exceeds 50 ppm or if the concentration of unsaturated chlorinated intermediates exceeds 500 ppm in the first-stage effluent, the bed is typically evaluated for a reduction in temperature set point or an increase in the hydrogen fluoride ratio. Because the fluorination reactions are highly exothermic, a single tube can be driven into a thermal excursion by a localized loss of heat transfer, such as a salt bath flow channel blockage or a localized catalyst packing void. Thermocouples placed across the tube sheet and at multiple radial positions help identify maldistribution, but once a tube has experienced a thermal excursion above 400°C, the catalyst within that tube may be permanently deactivated by sintering or may be cemented by carbonized residues. The operator must then decide whether to plug the tube or continue the campaign with reduced total conversion. Multitubular reactor designs therefore include spare tube capacity, and the run time limit may be declared when the number of plugged or reduced-flow tubes reaches 5% to 10% of the total. This operational boundary is specific to fixed-bed equipment and is not captured by bench-scale catalyst aging studies, which is one reason that published data for the exact run time limits of specific commercial configurations remains limited.
