Aromatic polyester polyol kettles fed with purified terephthalic acid (PTA) require control of the PTA particle size distribution before the first glycol charge reaches the reactor. The bulk solid properties of PTA—notably D10, D50, D90, span expressed as (D90 − D10)/D50, and sub-10 µm fines content—determine hopper discharge reliability, initial solid-liquid wetting behavior, and the temporal acid number trajectory established by ASTM D4662-20. Industrial PTA obtained from p-xylene oxidation and hydrogenation typically exhibits D50 values between 20 µm and 150 µm, with crystal habit and surface rugosity varying according to crystallization residence time, slurry temperature, purification train configuration, and dryer outlet conditions. A kettle process operating with a dual-flight helical ribbon impeller, a baffled 20,000 L reactor, and hot-oil coils is sensitive to PSD shifts because the reaction mass transitions from a solid-liquid suspension to a progressively homogeneous oligoester melt. Feed acceptance testing therefore combines laser diffraction per ISO 13320:2020, test sieving per ISO 2591-1:2008, and dynamic image analysis per ISO 13322-1:2014 when crystal aspect ratio or agglomerate structure must be quantified. Jenike shear testing per ASTM D6128-16 is used to establish the flow function for hopper and rotary valve design; flow function values below 4 indicate cohesive flow that can produce arching and erratic screw discharge. Fine PTA fractions below 40 µm create higher specific surface area, increased moisture adsorption at relative humidity above 60 %, and greater dust generation inside silos and charging chutes. The dust deflagration severity of fine PTA is evaluated where required by ASTM E1226 and minimum explosible concentration by ASTM E1515. In hopper and dense-phase transfer systems, narrow PSD spans below 1.2 are generally sought because coarse-plus-fine mixtures tend to segregate under center-fill conditions. Preheat and pre-drying of PTA, when silo storage has occurred at ambient humidity, is conducted at 110 °C to 120 °C until moisture by ASTM E203-16 is below 0.2 wt%. This prevents excessive foaming in the initial esterification stage and reduces the risk of premature organometallic catalyst hydrolysis.

At What D50 Does Solid–Liquid Mass Transfer Cease to Govern Kettle Cycle Time?

In batch esterification of PTA with diethylene glycol at initial molar ratios between 1.6 mol and 2.6 mol glycol per 1 mol PTA, the first portion of the temperature ramp operates with a heterogeneous solid phase. PTA has a density of approximately 1.51 g/cm³ and is sparingly soluble in cold glycol, so dissolution is a required precursor to oligomerization. The external surface area per unit mass scales inversely with median diameter; reducing D50 from 120 µm to 60 µm doubles the area-to-mass ratio by nonporous spherical geometry. The Stokes settling velocity scales with the square of particle diameter, so a 120 µm particle settles roughly 4 times faster than a 60 µm particle in laminar suspension. In a 20,000 L kettle with a low-speed helical ribbon impeller and no vertical baffles, just-suspended operation for coarse PTA may require a significantly higher impeller speed or a redesigned bottom clearance to sweep the dished head. Below 180 °C, the cycle time is commonly controlled by solid dissolution and the associated mass transfer boundary layer; between 180 °C and 210 °C, both dissolution and esterification kinetics influence the net disappearance of free acid; above 210 °C, the homogeneously dissolved PTA species and the oligoester acid ends undergo kinetic control. Quantitative kinetic-to-mass-transfer transition D50 values depend on impeller shear, solids loading, particle shape, and the solubility envelope of the specific glycol blend. Published data for this specific configuration is limited; the transition must be confirmed by hot slurry sampling through a 10 µm sintered stainless steel filter and by monitoring the acid number trajectory. If residual crystalline PTA persists beyond the expected dissolution window, the feed D50 or the pre-dispersion protocol is outside the operable envelope.

Interfacial mass transfer is further influenced by the low-solubility acid anion formation at the particle surface. Small particles can dissolve rapidly but can also form dense agglomerates when charged too quickly into hot glycol, especially if surface moisture exceeds 0.2 wt%. These agglomerates are not readily detected by dry PSD analysis but appear as oversized hydrated lumps in the reactor, causing transient torque spikes and local acid-rich zones. A wetting and dispersion protocol is therefore as important as the dry D50. Slurry charging via a high-shear inline disperser can mitigate this for fine PTA; however, the additional shear may raise the temperature and introduce dissolved oxygen if the transfer line is not nitrogen-blanketed. The process engineer must correlate the just-suspended impeller speed, the terminal settling velocity, and the hot-spot temperature profile to maintain the reaction within the narrow processing window. In many aromatic polyester polyol kettles, the optimal PTA D50 is not the smallest available but the coarsest material that still dissolves before the reaction exotherm and water removal reach their maximum rate.

Esterification Product Acid Number Trajectory Depends on Particle Surface Area

The analytical record of a kettle run is dominated by acid number measured by ASTM D4662-20, hydroxyl number by ASTM D4274-21, and rotational viscosity by ASTM D4878-15. When PTA D50 shifts from 90 µm to 120 µm without a change in crystallizer conditions, the initial acid number decline may lag because the total solid-liquid interface is smaller. The lag is most visible in the first 3 h of the heat-up phase, when kettle samples are heterogeneous and must be filtered hot before titration to avoid false acid contribution from suspended PTA. A coarse PTA charge with D90 above 180 µm may retain undissolved particles in low-shear zones near thermowells, bottom outlet flanges, and baffle attachments, producing local acidity that is not representative of the bulk. Fine PTA with D50 below 30 µm accelerates early acid number reduction but can create a fast initial water release that overloads the partial condenser if the heating rate is not reduced. The product acid number endpoint for aromatic polyester polyol production is often set at ≤2.0 mg KOH/g; this endpoint is linked to downstream polyurethane prepolymer stoichiometry. A batch that misses this target by 0.5 mg KOH/g due to coarse feed dissolution may require an extended hold at 220 °C to 230 °C, which increases the risk of color development measured by ASTM D1209. Thus PTA particle size is not an isolated powder property; it is a direct input to the kinetic control chart and to the final acid number capability index.

The following matrix summarizes typical operating bands for a 20,000 L coil-heated aromatic polyester polyol kettle. The bands are qualitative operational categories, not universal specifications, and plant-specific validation is required.

During the intermediate oligomerization stage the reaction mass becomes a viscous melt, and the torque on a close-clearance impeller becomes the most direct real-time process signal. Aromatic polyester polyols based on PTA and diethylene glycol typically show 25 °C viscosities between 1,500 and 8,000 mPa·s in manufacturer product bulletins, with hydroxyl numbers generally between 200 and 400 mg KOH/g. The reaction temperature is normally maintained between 210 and 235 °C, but the melt remains in the laminar regime inside the kettle, so the helical ribbon or anchor impeller consumes torque in proportion to apparent viscosity and speed. Particle size influences this stage indirectly through the amount of dissolution lag and through the distribution of oligomer chain lengths. Coarse PTA that has not fully dissolved acts as a suspended solid phase that can generate local regions of high acid-to-glycol ratio as it slowly releases PTA, resulting in higher molecular weight domains and subsequent viscosity heterogeneity. Fine PTA that dissolves rapidly tends to produce a more uniform acid-to-glycol concentration profile, but if the initial wetting step failed, the fines form acid-rich agglomerates that can deposit on thermowells and create a temperature signal bias. Thermowell fouling in a 20,000 L kettle has been observed as a lag between the coil oil temperature and the bulk thermocouple of 3 °C to 8 °C under high-fines feed. The processing window of ±5 °C around the optimum esterification setpoint requires that such temperature bias be minimized by selecting a particle size range that does not promote wall-scale formation. Water removal during this stage is driven by nitrogen sparge and vacuum; the water evolution profile is smoother with a controlled PTA dissolution rate because rapid water release can entrain glycol and fines overhead. The vacuum system response, not the polymerase kinetic curve, often becomes the limiting factor when fine PTA is charged into a hot kettle, and published data for this combined hydraulic and reaction configuration is limited; plant-specific validation is mandatory.

Sublimation Fouling and Vacuum System Carryover

PTA itself has a negligible vapor pressure at normal esterification temperatures, but fine PTA particles can be physically entrained during the water–glycol vapor flush that accompanies polycondensation. In a kettle with a partial condenser operated at 120 °C to 150 °C on the shell side and vacuum stripping at 5 kPa to 20 kPa absolute, the superficial vapor velocity in the riser determines whether a particle of a given terminal settling velocity can be lifted into the condenser. Sub-10 µm particles and low-molecular-weight oligomer droplets produce a fouling film on the tube side; the resulting pressure drop suppresses water removal and forces the reactor temperature upward to maintain the same distillation rate. This is a critical threshold risk because polyester color and decomposition rates accelerate sharply above 240 °C. The kettle processing window of ±5 °C is thus coupled to overhead cleanliness. Fouling precursors are controlled by adding PTA in a particle size band that minimizes dust carryover and by using a knockback condenser or demister pad with defined removal efficiency. When fine PTA is used to accelerate dissolution, a wet vent scrubber with a low-solids recycle stream may be required to prevent vacuum pump seal water from becoming acidic and laden with terephthalic acid fines. The acid number of the scrubber water is monitored by titration because absorbed PTA can depress pH and corrode carbon steel internals. The operational boundary is explicit: avoid pelletized or heavily dusted PTA feed extremes in a system designed only for a narrow D90 range, and do not initiate vacuum stripping until the overhead condensate is clear and the residual free acid concentration in the bulk has fallen below a pre-established limit.

When a Narrow PSD Feed is Not Achievable Due to Silo Segregation

Even if the supplier certificate of analysis reports an acceptable D50, silo segregation can deliver a widened PSD to the kettle. In a center-fill, center-discharge silo, fines tend to concentrate in the core and coarse particles at the periphery; discharge sequence then creates a time-varying D90/D10 ratio that is not captured by a single composite sample. A batch charged during the first silo heel may receive coarse material, while a batch charged after silo refill may receive a fines-rich core. The resulting acid number trajectories exhibit batch-to-batch scatter unrelated to catalyst activity. This scatter can be detected by control charting the hydroxyl number per ASTM D4274-21 and the Brookfield viscosity per ASTM D4878-15. The corrective action is not necessarily a tighter supplier PSD; it may involve mass-flow hopper retrofits, silo recirculation, or on-site screening. A vibratory screen with a 300 µm mesh removes only oversize agglomerates; a classifier or air separator is required to remove sub-20 µm fines. Dense-phase pneumatic conveying at 2 bar to 4 bar can also attrite PTA crystals, especially at bend elbows, and shift the D50 lower over transfer distance. The practice of pre-blending PTA with a small amount of glycol in a slurry tank before charging can reduce segregation impact, but the slurry must be continuously agitated to prevent fines flotation and coarse settling. If the span, defined as (D90 − D10)/D50, exceeds 2.0 at the kettle inlet, the process should not assume that dissolution time and acid number endpoint are governed by D50 alone. In that condition D90 and the fines fraction both drive the process in opposite directions: coarse particles extend the tail of solid dissolution, while fines overload the overhead system. The processing window of ±5 °C around the optimal esterification temperature is therefore insufficient to compensate for PSD segregation; feed engineering is required before the batch enters the kettle.

Termination of kettle processing is routinely based on acid number measured by ASTM D4662-20, water by ASTM E203-16, and viscosity by ASTM D4878-15; the clarity and filtration behavior of the discharged aromatic polyester polyol are equally sensitive to the upstream PTA particle size. Product specified with a final acid number of ≤2.0 mg KOH/g and a water content of ≤0.05 wt% can still fail downstream optical inspection if unreacted crystalline PTA or terephthalate oligomer particles remain. A final filtration step through a 10 µm bag or a 25 µm cartridge is typical for aromatic polyester polyols destined for rigid polyurethane or polyisocyanurate foam formulations. If the PTA D90 was above 200 µm, the filter differential pressure increases early in the discharge, and the batch may require multiple filter changes. This operational consequence is not captured by acid number alone because the solids may be localized in the heel or in dead zones. Hot filtration of a reactor sample through a 0.45 µm membrane is used to detect suspended particles before discharge. Haze measurements on the filtered liquid are often performed by visible transmittance at 500 nm or by the appropriate spectrophotometric method, and a failure typically triggers a rework hold at elevated temperature until the residual PTA dissolves. That rework hold can push the batch beyond the ±5 °C thermal window if the heat-transfer oil has limited capacity, increasing the APHA color value per ASTM D1209. It is therefore not sufficient to specify PTA particle size only by D50. The D90, fines fraction, moisture content, and discharge temperature profile must all be controlled because the final filterability of the polyol is a direct integration of particle dissolution history. The presence of amine-based neutralizers or tin catalysts should be sequenced carefully: adding free amines before residual acid number drops below 5 mg KOH/g may produce amine terephthalate salts with low solubility that blind final filters and increase turbidity. This incompatibility is an operational boundary, not a theoretical possibility.

Does PTA Particle Size Influence Catalytic Turnover in the Kettle?

Catalyst selection for aromatic polyester polyol kettle processing often includes organotitanates or organotin compounds. The active catalyst is sensitive to hydrolysis; trace water introduced by fine PTA fractions above 0.2 wt% can hydrolyze tetraalkyl titanate before the esterification reaches full temperature, forming titanium dioxide nuclei that are observed as haze and filterable solids. Particle size influences this behavior because fine PTA adsorbs atmospheric moisture more rapidly during storage and transfer, especially when the relative humidity exceeds 60 %. A coarse PTA with lower specific surface area carries less surface moisture under the same exposure conditions; however, the slower dissolution of coarse particles creates acid-rich pockets that can locally chelate titanium or tin active centers and delay the expected rate increase. This is one reason the use of fine PTA to shorten cycle time may not translate into a proportional improvement in catalytic turnover. The catalytic reaction is preceded by dissolution, and the dissolution state is influenced by both the solid-liquid interfacial area and the state of aggregation. If fine PTA is charged into a hot kettle too rapidly, local low-temperature zones can form at the feed point and cause the glycol to condense on the particles, producing a paste layer on the reactor wall. The paste layer may be visible only as a torque transient but can retain catalyst-bearing oligomer and delay attainment of the target acid number. The operational boundary is therefore to pre-mix or slowly charge PTA below 40 µm under nitrogen, to measure the moisture content by ASTM E203-16 before charge, and to avoid adding the full catalyst charge until the batch is fully dispersed and the free water overhead has dropped. Published data for explicit turnover frequency changes with PTA D50 in commercial polyol kettles is limited, so the catalyst-specific response must be confirmed on a laboratory reactor with the identical feed geometry. Once the acid number falls below 1.0 mg KOH/g, the particle size effect is no longer directly observable in the catalytic conversion; at that point, residual suspended solids and color are the remaining quality risks.