O-Xylene to Phthalic Anhydride: Production Process and Applications

In commercial multi-tubular fixed-bed oxidation of ortho-xylene, the feedstock is vaporized and mixed with filtered, compressed air to deliver an inlet hydrocarbon concentration typically held between 60 g/Nm³ and 85 g/Nm³ at normal conditions. The reactor contains 12,000–25,000 vertical tubes of 21–25 mm internal diameter and 3.0–4.0 m length, packed with a vanadium pentoxide–anatase titania contact mass promoted with phosphorus, antimony, cesium, or niobium depending on the licensor formulation. The superficial gas velocity at normal conditions is maintained at 1.8–2.4 m/s, establishing turbulent flow through the catalyst bed and a pressure drop of 0.3–0.7 bar. The exothermic reaction between ortho-xylene and atmospheric oxygen proceeds through tolualdehyde, phthalide, and intermediate carboxylate species to phthalic anhydride, while parallel deep oxidation routes generate maleic anhydride, carbon monoxide, carbon dioxide, and water. Industrial selectivity to phthalic anhydride normally falls between 78 mol% and 82 mol% at ortho-xylene conversion above 99.0%. The heat of reaction is removed by a circulating molten salt mixture—typically a ternary nitrate-nitrite eutectic—maintained at 350–375 °C, and the product gas leaving the reactor at 360–390 °C is quenched in a waste-heat boiler to generate high-pressure steam before entering the recovery section. Experience on production-scale trains shows that small deviations in inlet hydrocarbon concentration or salt-bath temperature propagate rapidly into selectivity loss because the desired partial oxidation operates within a narrow window between incomplete conversion and runaway CO_x formation.

Because the oxidation chemistry is strongly exothermic and the reaction network contains both partial and total oxidation steps, the maximum catalyst temperature—commonly called the hot spot—must be controlled independently from the salt-bath set point. In a tube experiencing inadequate heat transfer, the hot spot can rise from a normal range of 420–450 °C to above 470 °C within minutes, at which point selectivity to phthalic anhydride collapses and the catalyst may undergo irreversible sintering of the anatase support. The temperature profile is monitored in selected tubes by axially inserted thermocouples, and the salt-bath circulation rate is designed to maintain a maximum radial temperature difference of 2–5 °C across the tube bundle. The apparent activation energy for the desired partial oxidation is lower than that of the competing total oxidation routes, so hot-spot suppression is a kinetic necessity rather than only a safety measure. Industrial operations therefore use catalyst pellets shaped as Raschig rings, trilobes, or miniliths with equivalent diameters of 3–6 mm and intra-particle void fractions high enough to reduce diffusion resistance without sacrificing mechanical crush strength. Fresh catalyst is loaded to a fill density of 0.9–1.1 kg/L, and the tubes are individually pressure-drop checked to maintain flow distribution within ±5% of the mean. Tube-to-tube flow maldistribution produces localized oxygen-to-hydrocarbon excursions, and the resulting temperature spikes are a common batch-to-batch variance source during start-up and after partial catalyst change-out.

Why Is Salt-Bath Temperature Control the Primary Constraint on Selectivity?

The salt bath functions as both heat sink and thermal reservoir, and its temperature uniformity determines whether the reactor operates within the narrow selectivity window. The molten salt system is circulated through the shell side by centrifugal pumps at a rate of 8,000–20,000 m³/h, depending on train capacity, and heat is removed by generating steam in external or internal waste-heat boilers. The salt composition is monitored for nitrate/nitrite ratio, carbonate concentration, and chloride content; chloride ingress from instrument leaks or feedstock contamination can increase corrosivity toward stainless steel tube sheets and initiates stress-corrosion cracking in sensitised weld zones. The salt-bath temperature is normally ramped at 5–10 °C/h during start-up to avoid thermal shock to the tube-to-tubesheet joints, and the reactor is brought up under air flow before ortho-xylene is introduced. Once the catalyst ages, the salt-bath temperature is increased gradually by 1–2 °C per year to compensate for declining activity, but this practice narrows the remaining operating margin because the hot-spot temperature rises nonlinearly with salt temperature. At feed loads above 85 g/Nm³, the hot spot may exceed 450 °C even with a salt bath at 365 °C, and the selectivity penalty is typically 2–4 mol% for every additional 10 °C of hot-spot temperature above the optimum. This threshold behaviour is the principal processing conflict in phthalic anhydride manufacture and forces operators to balance catalyst ageing, throughput, and selectivity on a daily basis.

The kinetic coupling between the desired partial oxidation and the undesired total oxidation becomes particularly severe when the catalyst is operated under oxygen-rich conditions and elevated temperature. The reaction rate of ortho-xylene oxidation follows a Langmuir–Hinshelwood-type rate expression that is first-order in hydrocarbon and fractional-order in oxygen, but the deep oxidation rate increases more rapidly with temperature. Selectivity is therefore maximised when the local catalyst temperature is kept below 440 °C, which requires an inlet oxygen-to-hydrocarbon ratio of approximately 20:1 by mass and a molten salt temperature of 352–368 °C for a fresh charge. On production-scale reactors, the hot spot is typically located at 30–50% of the tube length from the inlet, where the hydrocarbon concentration remains sufficient to generate high heat release but the local oxygen concentration has not yet decreased enough to suppress total oxidation. The use of 25 mm inner diameter tubes is a compromise between heat removal and catalyst loading; larger tubes reduce reactor capital cost but increase the radial temperature gradient above 10 °C, causing the centreline catalyst to operate under severe deep-oxidation conditions. The temperature control system therefore responds not only to salt-bath thermocouples but also to product-gas oxygen analysers, which detect oxygen breakthrough from maldistributed tubes. A rising tail-gas oxygen concentration above 2.5–3.5 vol% at constant feed load is an early diagnostic of partial catalyst deactivation or localised hot-spot instability.

When Inlet o-Xylene Loading Approaches the Upper Flammability Envelope

Ortho-xylene vapour forms flammable mixtures with air between a lower flammability limit of approximately 0.9 vol% and an upper limit of about 7.0 vol%, and the autoignition temperature is reported near 464 °C. At an inlet loading of 80 g/Nm³, the hydrocarbon concentration is roughly 1.8 vol%, placing the feed mixture inside the flammable envelope. The process cannot simply operate below the lower flammability limit because the resulting reactor productivity would be economically unattractive; instead, the mixing and reactor system is designed to exclude ignition sources, maintain gas velocities far above the flame propagation velocity, and withstand an internal deflagration. The ortho-xylene vaporiser is operated with hot oil at 160–190 °C, and the vaporised stream is diluted with air downstream of the vaporiser to avoid liquid-phase hot surfaces. The air supply is filtered to remove particles above 1 µm and dried to a dew point below -20 °C to protect the catalyst and prevent hydrate formation in instrument lines. The mixed feed is routed through a silo-type gas mixer containing static mixing elements, and the homogeneity of the hydrocarbon-air mixture is verified by multiple infrared analysers before the stream enters the reactor. Explosion protection includes rupture discs on the reactor inlet and outlet, flame arresters on the vaporiser vent, and continuous monitoring of oxygen and hydrocarbon concentrations. The operating margin between the lower flammability limit and the process set point is narrow, so control valves and trip logic are configured to isolate ortho-xylene feed automatically if the inlet concentration exceeds 110% of the design set point or if the reactor inlet temperature deviates by more than 15 °C.

Recovery of phthalic anhydride from the reactor effluent begins in gas coolers that reduce the stream from 360–390 °C to 160–180 °C, after which the stream enters a pair of air-cooled or water-cooled switch condensers. Desublimation occurs on finned tube banks at gas-phase temperatures between 55 °C and 70 °C, where phthalic anhydride crystallises directly from the vapour without passing through the liquid phase. The condenser is operated cyclically: during the loading phase, the tube banks accumulate solid phthalic anhydride, and then the unit is isolated and heated with hot oil at 180–200 °C to melt the crude product into a receiving tank. The switch condenser cycle time is typically 3–8 h, and the two vessels alternate automatically to maintain continuous operation. Incomplete desublimation leaves phthalic anhydride in the tail gas, requiring a thermal oxidiser or scrubber for emission control, while excessive cooling below 50 °C causes co-condensation of water and maleic anhydride, increasing the acidic impurity load. The crude molten phthalic anhydride is then held in agitated ageing tanks at 250–270 °C for 12–24 h, during which phthalide and colour-forming precursors are decomposed or condensed. The aged crude is distilled under vacuum at 50–100 mbar and 200–230 °C to produce refined phthalic anhydride with a purity above 99.8 wt%, maleic anhydride below 0.05 wt%, and colour below 20 Hazen after solidification. The purified product is flaked or pelletised under dry nitrogen and stored at temperatures below 30 °C and relative humidity below 60% to prevent hydrolytic formation of phthalic acid on the surface.

In the esterification of phthalic anhydride with C₈–C₁₀ oxo alcohols, the purified anhydride is charged to a jacketed stainless-steel or glass-lined reactor with excess alcohol and a tetraalkyl titanate or organotin catalyst. The alcohol-to-anhydride molar ratio is controlled between 2.2:1 and 2.8:1, and the reaction is run at 180–220 °C under reduced pressure to remove water continuously. The first esterification step opens the anhydride ring rapidly to form the monoester, while the second step is equilibrium-limited and requires removal of water to below 0.05 wt% in the reaction mass to achieve diester yields above 99%. The reaction is followed by neutralisation with dilute aqueous alkali, water washing, stripping of unreacted alcohol, and filtration with diatomaceous earth or activated carbon. The resulting phthalate esters—dioctyl phthalate, diisononyl phthalate, and diisodecyl phthalate—are used as primary plasticisers in flexible polyvinyl chloride compounds. In plastisol applications, the ester viscosity and solvation behaviour are controlled by the isomer distribution of the oxo alcohol; diisononyl phthalate typically exhibits a viscosity of 70–100 mPa·s at 20 °C, whereas dioctyl phthalate is about 80–85 mPa·s at the same temperature. The compatibility of the plasticiser with PVC is assessed through loop migration tests and through extraction resistance measured according to ASTM D1239-14, while plasticizer viscosity and acid number are controlled under ASTM D1045-19. On production-scale PVC compounding lines, phthalate plasticiser moisture above 0.05 wt% causes surface defects and reduced gelation in twin-screw extruders, and the use of a co-rotating twin-screw extruder with L/D 40:1 and vacuum venting at 60–80 mbar is standard for dry-blend processing. The vapour pressure and migration behaviour of low-molecular-weight phthalates have driven substitution toward higher-molecular-weight esters, but phthalic anhydride remains the central dicarboxylic acid precursor for this product class.

Does Residual Phthalide Content Control UPR Cure Stability?

Unsaturated polyester resins produced from phthalic anhydride, maleic anhydride, and propylene glycol are sensitive to the residual phthalide and colour-body content of the phthalic anhydride charge. In a typical two-stage polycondensation, phthalic anhydride is added in the first stage to build molecular weight, while maleic anhydride is added later to preserve the unsaturated sites required for styrene crosslinking. The reaction is run at 190–220 °C with an inert gas sparge, and water is removed through a partial condenser until the acid value falls to 20–35 mg KOH/g. The resin is then cooled and dissolved in styrene at 30–40 wt% to yield a curable liquid with a viscosity of 250–800 mPa·s at 25 °C. Residual phthalide in the phthalic anhydride does not directly consume styrene, but it contributes to an internal acid-catalysed side reaction during polycondensation that shifts the molecular weight distribution and produces a haze-forming fraction. When the phthalide content of refined phthalic anhydride exceeds 0.10 wt%, the cured UPR castings exhibit reduced Barcol hardness, lower tensile strength, and increased water absorption; the effect is measurable under ASTM D638-14 tensile testing and ASTM D2583-13 for hardness. In filled UPR systems such as cultured marble and solid-surface manufacturing, batch-to-batch variation in phthalic anhydride purity causes cure drift and surface porosity, particularly when the resin is cured with methyl ethyl ketone peroxide at 1.0–1.5 phr and cobalt naphthenate accelerator at 0.2–0.5 phr. The practical limit for phthalide is therefore not a simple specification value but a process control boundary: refiners maintain phthalide below 0.05–0.10 wt% because the UPR customer’s gel time and exotherm profile shift when the impurity concentration changes by as little as 0.03 wt%.

Alkyd resin formulations use phthalic anhydride as the aromatic dicarboxylic acid component to raise glass-transition temperature, hardness, and chemical resistance in oxidatively drying coatings. A medium-oil alkyd based on soybean or linseed oil contains phthalic anhydride at 25–35 wt% of the finished resin solids, while short-oil alkyds for industrial stoving enamels may contain 35–45 wt%. The resin is processed by monoglyceride alcoholysis followed by polycondensation with phthalic anhydride and polyols such as pentaerythritol or glycerol at 220–250 °C, with xylenes used as azeotropic solvent to remove water. The final acid value is typically below 10 mg KOH/g, and the resin solution viscosity is adjusted to 2–8 Pa·s at 25 °C. Coating performance is evaluated under ISO 1522 pendulum damping hardness, ISO 2813 specular gloss, and ISO 2409 cross-cut adhesion. When low-grade phthalic anhydride containing maleic anhydride above 0.05 wt% is used, the maleic functionality is incorporated into the alkyd backbone and changes gelation behaviour during processing, particularly in pentaerythritol-containing resins where the combination of tetrafunctional polyol and unsaturated diacid increases the risk of premature crosslinking. Production-scale alkyd reactors therefore specify phthalic anhydride colour below 20 Hazen and heat stability at 250 °C for 2 h with a colour increase of less than 10 Hazen, because colour development during processing is a direct indicator of iron contamination and oxidative degradation of the aromatic ring.

If Low-Molecular-Weight Color Bodies Exceed 0.05 wt%, Coating Gloss Fails

The colour bodies in refined phthalic anhydride are largely aromatic condensation products formed during the ageing and distillation steps, and they become visible in high-gloss coating films only when their concentration exceeds a threshold near 0.05 wt%. These impurities have molar extinction coefficients that are several orders of magnitude higher than phthalic anhydride itself, so even trace quantities impart a yellow or brown tint. The standard quality-control method for phthalic anhydride colour is the molten colour test, often expressed in Hazen units after heating the sample to 250 °C for 2 h, and many resin manufacturers require a value below 20 Hazen for premium applications. The corresponding method for maleic anhydride is based on polarographic or gas-chromatographic analysis with a detection limit of 0.01 wt%, and the specification for phthalic anhydride used in high-solids coatings is typically maleic anhydride below 0.05 wt%. In polyester powder coatings, phthalic anhydride is used as a chain extender and aromatic acid component, and the resin is melt-blended with a blocked isocyanate or triglycidyl isocyanurate curing agent in a twin-screw extruder at 80–120 °C. The extrudate is ground and classified to a median particle size of 30–50 µm, and the resulting powder is applied electrostatically and cured at 180–200 °C. Low-molecular-weight colour bodies in the phthalic anhydride charge survive the extrusion and curing steps and lower the reflectance and yellowness index of the finished coating, which is measured under ASTM D523-14 for gloss and ASTM D1925-70 or ISO 7724 for yellowness. The processing window for powder coatings is narrow because the resin’s glass-transition temperature must remain above 40 °C for storage stability, and impurities that plasticise the resin by as little as 1–2 °C can cause sintering in the powder bag.

The manufacture of anthraquinone dyes and pigments consumes a smaller fraction of phthalic anhydride than plasticizer and resin applications, but the purity requirements are stringent because the condensation chemistry is sensitive to acidic and oxidisable impurities. Phthalic anhydride reacts with benzene derivatives under Friedel-Crafts conditions to form anthraquinone intermediates, which are subsequently sulfonated, chlorinated, or aminated. The reaction is carried out in a solvent or in molten phthalic anhydride at 120–250 °C using aluminium chloride or acid catalysts, and the presence of maleic anhydride or phthalide leads to coloured by-products that are difficult to remove from the final pigment. The quality of phthalic anhydride for this application is controlled by molten colour, maleic anhydride content below 0.05 wt%, and iron content below 1 mg/kg, because iron promotes decomposition during the high-temperature condensation. The purity of the intermediate is verified by thin-layer chromatography and by UV-visible absorption spectroscopy. Published data for specific anthraquinone pigment yields as a function of phthalic anhydride purity is limited, but the industrial experience is that colour-body contamination increases the number of recrystallisation steps required to reach pigment-grade purity, thereby raising production cost and reducing yield. The same quality constraints apply to the production of phthalimide and phthalocyanine pigments, where phthalic anhydride is condensed with urea and a metal salt, and trace acidic impurities interfere with the cyclisation reaction.

The operational boundaries for phthalic anhydride storage and handling are determined by its hygroscopicity, reactivity with water, and solidification behaviour. The anhydride ring opens in the presence of moisture to form phthalic acid, increasing the acid number and reducing the reactive anhydride titre available for downstream esterification or polycondensation. Bulk storage tanks are therefore heated with hot-water jackets or steam tracing to maintain the product at 140–160 °C in the molten state, and the tank headspace is purged with dry nitrogen at a dew point below -20 °C. At temperatures above 220 °C, phthalic anhydride undergoes gradual thermal decomposition and colour development, so transfer lines and pumps are designed for turbulent flow with minimal dead legs, and the residence time in hot zones is kept below 4 h. The solidified material is classified as a skin and respiratory irritant, and exposure limits are enforced through workplace monitoring; the guidance for occupational exposure is generally aligned with a limit of 1–2 mg/m³ as respirable dust. Contact with strong bases releases exothermic neutralisation heat and can generate phthalate salts, while contamination with primary amines leads to imide formation and may accelerate corrosion of aluminium transfer equipment. The use of phthalic anhydride in food-contact polymers is constrained by the migration limits applicable to the derived phthalate esters under European Union Regulation (EU) No 10/2011 and by the restrictions on certain phthalate esters under REACH Annex XVII, which specifies a maximum concentration of 0.1 wt% of the individual restricted phthalate in plasticised toys and childcare articles. The refined product is not a single-use bulk chemical but a reactive intermediate, and its successful processing requires control of moisture, trace acids, temperature history, and downstream impurity interactions at every transfer point.