The water/isopropanol minimum-boiling binary azeotrope at 101.3 kPa is approximately 87.8 wt% 2-propanol with a boiling point of 80.3 °C, and this thermodynamic boundary defines the dehydration challenge for both direct catalytic hydration and indirect sulfate ester hydrolysis. In direct hydration, propylene is contacted with water over a fixed-bed acidic catalyst under pressure, producing a reactor effluent that is dilute in isopropanol and water-saturated after light ends are stripped. In indirect hydration, propylene is absorbed into sulfuric acid to form isopropyl hydrogen sulfate and diisopropyl sulfate, which are then hydrolysed with water to liberate isopropanol and regenerate an acid stream. Both routes generate an aqueous isopropanol mixture that approaches the azeotropic composition after primary distillation, and the achievable distillate composition is limited by the binary vapour-liquid equilibrium; simple fractional distillation cannot yield a bottoms product with water content below the azeotropic concentration. Azeotropic water removal therefore involves the deliberate addition of a third component that forms a minimum-boiling heteroazeotrope with water, or a pressure-swing adsorption unit that exploits the exclusion of water from zeolitic pores. The selection among direct hydration, indirect sulfation, heteroazeotropic distillation, and molecular-sieve dehydration is driven by feedstock cost, plant capacity, energy integration, corrosion resistance, and the regulatory status of the entrainer, because benzene-based azeotropic systems have been replaced in most regions by diisopropyl ether or cyclohexane systems due to REACH restrictions.

Thermal degradation pathways in supported phosphoric acid catalyst systems

Direct hydration of propylene over supported phosphoric acid operates within a narrow processing window because the catalyst activity requires a minimum reaction temperature of approximately 180 °C to overcome kinetic restraints, while temperatures above 250 °C accelerate acid loss and promote the formation of diisopropyl ether, acetone, and oligomeric by-products. Industrial fixed-bed reactors maintain pressures in the range 2.0–7.0 MPa to shift the equilibrium toward isopropanol; the pressure vessel and shell-and-tube reactor internals are typically designed to ASME BPVC Section VIII Division 1 or EN 13445-3 with post-weld heat treatment. Published kinetic data for supported phosphoric acid catalysts indicate apparent activation energies between 80 kJ mol⁻¹ and 120 kJ mol⁻¹, and rate expressions are generally first-order in propylene with inhibition by water at low water partial pressures. Single-pass propylene conversion is deliberately limited to roughly 5–12% because higher conversion increases reactor outlet water partial pressure and displaces phosphoric acid from the silica support. The resulting unreacted propylene is recycled through a high-pressure separator, a compressor knock-out drum, and a guard bed to remove water and acid aerosols before recompression. Field experience from multi-tubular reactors with tube lengths of 6–12 m and tube inner diameters of 25–50 mm shows that catalyst deactivation is rarely uniform: the upper tube sections lose acid first due to axial temperature gradients, and differential pressure increases by 0.1–0.5 bar over the catalyst life. The propylene feed must be limited to sulfur-poor quality because sulfur compounds neutralise the acidic sites; methanol in the feed may form dimethyl ether. A feed with methylacetylene or propadiene above 10 ppm can generate fouling polymers on the acid surface, so hydrotreated polymer-grade propylene is preferred, and the feed is analysed by ASTM D2163-14 before introduction to the hydration reactor.

The water-to-propylene mole ratio in direct hydration is maintained between 0.3:1 and 3:1 depending on the licensor configuration; lower ratios suppress acid elution but decrease conversion, while higher ratios improve conversion at the cost of increased water load in downstream distillation and greater acid entrainment. The reactor effluent is cooled in a feed-effluent heat exchanger with a shell-side temperature approach of 15–35 °C and then separated in a high-pressure drum designed for acid carryover; the aqueous phase is neutralised or evaporated, and the organic phase is distilled in a primary recovery column. Because the catalyst is sensitive to free water, the recycle propylene is dried over a molecular-sieve bed or a coalescing filter before recompression. The entire high-pressure loop must be constructed of corrosion-resistant materials, typically 316L stainless steel with alloy 20 or polytetrafluoroethylene-lined sections in the aqueous acid region, because trace phosphoric acid is entrained as an aerosol. The pressure relief system is sized for thermal expansion and for reactor runaway caused by loss of recycle gas quenching; relief valves are specified to ISO 4126-1 and discharge to a caustic scrubber rather than directly to atmosphere.

What governs entrainer selection and liquid-liquid split in azeotropic water removal?

A heteroazeotropic dehydration column for isopropanol-water mixtures is controlled less by the boiling point of the entrainer than by the shape of the ternary liquid-liquid equilibrium envelope and the mass fraction of water in the vapour phase that leaves the top tray. Candidate entrainers include diisopropyl ether, cyclohexane, and benzene; diisopropyl ether is often preferred because it can be generated on-site by the acid-catalysed intermolecular dehydration of isopropanol, whereas benzene is classified under REACH Annex XVII and cannot be used in new European installations without strict emission controls. The entrainer forms a minimum-boiling heteroazeotrope with water, and the overhead vapour condenses into two liquid phases: an organic-rich phase that is recirculated as reflux and an aqueous-rich phase that is withdrawn to waste or recovery. The composition of the aqueous phase governs whether the extractive or heteroazeotropic mechanism is stable; if the aqueous-phase IPA content is too high, the decanter can become single-phase and the column loses its liquid-liquid split. The design decanter temperature is typically held between 15 °C and 35 °C to reduce mutual solubility and increase the water-rich phase yield. The overhead drum must provide at least 20–40 min of residence time between the light organic layer and the heavy aqueous layer, and the interface is controlled by a conductivity probe or differential-pressure displacer. Table 1 summarises selected binary heteroazeotrope data for three entrainers at atmospheric pressure.

The selection of diisopropyl ether introduces a secondary process constraint because it forms a ternary minimum-boiling mixture with water and isopropanol; the overhead vapour temperature can be lower than the binary entrainer-water boiling point, and the decanter composition is strongly sensitive to the IPA concentration in the rectifying section. Published liquid-liquid equilibrium data for the DIPE-water-isopropanol system at 25 °C and 101.3 kPa are available but are often extrapolated from limited tie-line measurements; for design purposes, the organic phase is saturated with water and the aqueous phase is saturated with DIPE, and the mutual solubility increases substantially above 35 °C. The column must therefore be operated with a top pressure that allows condensation against cooling water at or below 35 °C; in hot climates a chilled water supply at 10–15 °C may be required to maintain decanter stability. Because DIPE is less dense than water and less stable thermodynamically than isopropanol at high temperature, the column reboiler is operated with a maximum skin temperature below 150 °C to minimise ether decomposition and acid-catalysed peroxide formation. The DIPE-water heteroazeotrope contains only 4.5 wt% water; therefore the volumetric entrainer circulation rate is high relative to the aqueous distillate flow, and the column diameter in the rectifying section is usually governed by entrainer recycle rather than product capacity.

Column internal selection for a heteroazeotropic dehydration system is constrained by the low interfacial tension and the high organic-phase loading in the rectifying section. Structured packing with a specific surface area of 250 m² m⁻³ to 350 m² m⁻³ is commonly used because the HETP for the ternary IPA-water-entrainer system falls between 0.25 m and 0.45 m under normal operating conditions, and the pressure drop is low enough to avoid excessive reboiler temperatures. Random packing of 25 mm to 50 mm diameter, such as stainless pall rings or ceramic saddles, is an alternative for fouling services or when capital cost is limited, but the hydraulic capacity is lower. Trayed columns with valve trays or sieve trays suffer from reduced efficiency in the high-entrainment region because the liquid phase is partially immiscible, and the froth density can collapse at high organic-water ratios. The liquid distributor above the structured packing must achieve a drip-point density of at least 100–200 points m⁻² and be levelled to a tolerance of 3 mm m⁻¹, because uneven distribution initiates local entrainer starvation and raises the water content of the overhead vapour. Below the feed, the stripping section is less dependent on liquid-liquid behaviour because the entrainer is stripped from the aqueous phase; the stripping section is designed with additional residence time to hydrolyse or remove residual sulfate esters in indirect-hydration feed streams. Process field data from columns with decanter undersizing show that entrainer carry-over into the aqueous draw is the most common failure mode, followed by interface controller hunting caused by incompatible density signals in the two-phase zone. The boot or horizontal decanter must be sized so the organic-phase superficial velocity is below 0.003 m s⁻¹, and the water draw leg must include a U-seal with a liquid height equivalent to at least 1.5 times the maximum continuous pressure differential.

The overhead condenser and vent system for DIPE-containing columns require specific attention to peroxide stability and low-flash-point handling. The condenser is typically a horizontal shell-and-tube unit with process fluid on the shell side and cooling water on the tube side, with a duty split between a primary condenser and a vent condenser sized for 3–5% of the total overhead flow. Non-condensables are vented from the top of the decanter through a nitrogen blanket, and the vent stream is sent to a thermal oxidiser because DIPE has a lower explosion limit of approximately 1.0 vol% in air. The reflux split to the column is controlled by the organic-phase overflow weir and not by a simple flow controller, because the organic phase must remain saturated with water to preserve the ternary azeotropic composition. In plants where the aqueous phase is returned to the primary distillation column, the aqueous draw is pumped through a plate-and-frame exchanger to recover heat and is then acidified or neutralised depending on the upstream route. The material of construction for the decanter and organic-reflux systems is typically 316L stainless steel or 304L stainless steel with post-fabrication pickling and passivation to reduce iron-catalysed peroxide initiation; copper alloys are avoided because copper ions accelerate DIPE peroxide decomposition and radical formation.

Sulfuric acid indirect hydration and hydrolysis unit corrosion boundaries

In the indirect route, propylene is absorbed into 70–85 wt% sulfuric acid at 20–35 °C in a packed absorber or stirred contactor to form a mixture of isopropyl hydrogen sulfate and diisopropyl sulfate; the absorption is exothermic and requires external cooling to prevent acid-catalysed polymerisation of propylene and ether formation. The ester mixture is then diluted with water to an acid strength of 30–45 wt% and hydrolysed at 70–100 °C to release isopropanol and reform sulfuric acid. The dilution and hydrolysis steps are kinetically complex: diisopropyl sulfate hydrolyses more slowly than isopropyl hydrogen sulfate, and the residence time in the hydrolysis vessel must be sufficient to reduce residual sulfate esters to below the detection limit of the subsequent neutralisation section. Materials of construction for the absorber and hydrolysis vessels include glass-bonded steel, tantalum-lined steel, or alloy 20 for the strong-acid zones; 316L stainless steel is generally unsuitable at acid concentrations above 80 wt% and temperatures above 40 °C. The acid reconcentration step uses a vacuum evaporator operated at 7–15 kPa absolute and 120–150 °C vapour temperature to return the sulfuric acid to the absorber strength, and the overhead water vapour is used to preheat the hydrolysis feed. Impurities such as propylene oligomers and diisopropyl ether are stripped from the hydrolysis product and sent to the primary distillation train.

Field failure data from indirect units identify localised corrosion at the dilution point as the dominant processing bottleneck, because mixing concentrated acid with water generates a region of intermediate acid strength where the corrosion rate of stainless steel is highest. The acid dilution tee and the hydrolysis feed quench are therefore specified with either tantalum inserts or PTFE-lined spools, and the mixing nozzle is designed to prevent backflow of dilute acid into the strong-acid header. The acid-catalysed formation of diisopropyl ether is suppressed by keeping the absorption temperature low and by minimising the residence time of the ester mixture before hydrolysis. The crude isopropanol stream from the hydrolysis column contains water, isopropyl alcohol, diisopropyl ether, and traces of acid; it is neutralised with caustic soda to a pH of 6.5–7.5 before entering the dehydration section to reduce ester hydrolysis by-product formation and protect downstream steel columns. A caustic scrubber is installed on the acid vent system, and the scrubber packing is checked at intervals defined by ISO 14001 environmental management requirements. The energy intensity of the indirect route is higher than direct hydration because the acid reconcentration step consumes low-pressure steam at a rate of 5–8 kg steam per kg of isopropanol product, depending on the initial acid strength and vacuum system efficiency; field measurements show that the reconcentration often contributes more than 40% of the total steam load. Consequently, the indirect route is economically attractive only when low-cost propylene is available and when the acid plant is integrated with a sulphur-burning or spent-acid regeneration unit.

When pressure-swing adsorption replaces heteroazeotropic distillation for anhydrous IPA finishing

When the target water content is below 0.05 wt% or when entrainer emissions are tightly controlled, a pressure-swing adsorption unit using zeolite 3A molecular sieve can be operated downstream of a conventional distillation column that approaches the azeotropic composition. The molecular sieve has a pore diameter of approximately 0.3 nm, which excludes isopropanol with a kinetic diameter of 0.47 nm and permits water with a kinetic diameter of 0.265 nm to enter the cages. The equilibrium water capacity of zeolite 3A is approximately 20–22 wt% at 25 °C and high relative humidity, and the dynamic capacity in a typical two-bed adsorption cycle is reduced to 8–12 wt% due to mass-transfer limitations and residual water left after regeneration. Adsorption is conducted at 0.5–1.5 MPa and 25–50 °C in carbon steel columns lined with a high-temperature inorganic coating to prevent iron-catalysed IPA dehydration to propylene; regeneration is performed with heated dry nitrogen or superheated isopropanol vapour at 200–250 °C and 0.1–0.3 MPa. The adsorption beds are sized for a superficial velocity of 0.05–0.20 m s⁻¹ and a pressure drop below 0.35 bar across the bed, with a length-to-diameter ratio of 2:1 to 4:1 to reduce channelling. Field data from industrial PSA units indicate that the main operational risk is a gradual loss of water capacity caused by non-volatile impurities that block the crystal surface; therefore the feed to the PSA is filtered through a 5 µm particulate filter and pre-treated to remove traces of acid and peroxide. The product water content is verified continuously by an online NIR analyser or at-line Karl Fischer titration according to ASTM E203-16. Unlike distillation, the PSA unit does not produce a liquid entrainer recycle stream, but it does generate a regeneration condensate that is recycled to the primary distillation feed. The regeneration gas must be dried before being reheated, and the cooling step is performed in a closed loop to avoid thermal shock to the adsorbent; heating and cooling rates are limited to 30–50 °C h⁻¹ in vendor procedures. In installations where the feed is saturated with water at the azeotropic concentration, the PSA unit reduces water content to ≤0.05 wt% only if the adsorption cycle time is less than 45–60 min; longer cycles allow the mass-transfer zone to break through into the product line. The pressure-swing unit is therefore unsuitable for direct treatment of dilute aqueous IPA, and pre-distillation must bring the feed to at least 85 wt% isopropanol to avoid excessive bed size and rapid breakthrough.

The regeneration loop for zeolite 3A includes a high-temperature blower, an electric or steam-heated regeneration gas heater, and a water-cooled condenser that separates desorbed water from the gas stream. The heater is controlled to a maximum outlet temperature of 280 °C to avoid structural damage to the zeolite binder; the bed outlet regeneration gas temperature must remain above 200 °C for at least 1–2 h to achieve adequate water removal. Condensed regeneration water is analysed for isopropanol by gas chromatography, and if the IPA content exceeds 0.5 wt%, the regeneration temperature is reduced because excessive local heating can dehydrate isopropanol to propylene or diisopropyl ether. The switching valves between adsorption and regeneration are specified to API 607 or ISO 10497 fire-safe requirements and must be actuated with a maximum switching time of 2–5 s to prevent pressure surge and bed fluidisation. Process safety interlocks include a high-temperature shutdown on the regeneration heater, a low-flow bypass on the product line, and a hydrocarbon detector at the regeneration condenser vent. Because the molecular sieve is hygroscopic and degrades in the presence of liquid water, the bed is protected against liquid carryover by a high-level alarm in the feed knockout drum and by a demister pad with a separation efficiency of 99% for droplets above 10 µm. In comparison with heteroazeotropic distillation, the pressure-swing route eliminates the need for a hazardous entrainer but increases the solid-handling and regeneration-energy burden; published data for the specific configuration of azeotropic IPA feed and zeolite 3A are more limited than for ethanol dehydration, and vendor pilot tests are generally required before a guaranteed performance specification is accepted.