In continuous direct esterification of purified terephthalic acid (PTA) with 1,4-butanediol (BDO) for the production of polybutylene terephthalate (PBT), the measured loss of BDO as tetrahydrofuran (THF) is the dominant non-esterification sink in the molar balance and must be treated as an active stoichiometric variable rather than a waste-stream correction. The primary reaction sequence converts PTA and BDO first to 4-hydroxybutyl terephthalate and then to bis(4-hydroxybutyl) terephthalate, releasing water of esterification; in parallel, the acid-catalyzed intramolecular dehydration of BDO—HO(CH₂)₄OH → C₄H₈O + H₂O—consumes exactly 1 mol of BDO for every mole of THF produced and adds one additional mole of water that cannot be distinguished from esterification water unless the THF content of the overhead stream is measured directly. Because the theoretical BDO demand for one metric ton of PBT repeat unit is approximately 409 kg, a representative THF loss of 20 kg/t consumes 25 kg BDO per metric ton, equivalent to 6.1% of theoretical BDO demand; the exact fraction varies with temperature, residence time, and esterifier acidity. The cyclization reaction is first order with respect to protonated BDO concentration, with activation energies reported in publicly available kinetic studies between 85 kJ/mol and 120 kJ/mol; as a result, a 10 °C increase in esterifier temperature within the commercial range of 225–245 °C can increase the THF formation rate by a factor of approximately 2.0–2.5. This exponential thermal sensitivity conflicts directly with the kinetic requirement to operate at high temperature for acceptable PTA dissolution and esterification rate. The esterification reactor is a heterogeneous paste-fed system in which PTA particle size, paste uniformity, and the concentration of dissolved carboxyl groups govern the apparent esterification rate; therefore, molar balance control requires simultaneous reconciliation of esterification, transesterification, polycondensation, and BDO cyclization rates. The loss of BDO to THF alters the effective BDO-to-PTA stoichiometry in the melt phase, and if not measured at the overhead condenser and vacuum traps, it produces systematic errors in calculated conversion and end-group balance.

The esterification product contains 4-hydroxybutyl ester end groups that can undergo intramolecular backbiting to release THF and generate a free carboxyl end group without consuming a terephthalate unit. This reaction path decouples the apparent hydroxyl functionality from the converted BDO inventory and increases the carboxyl end-group concentration in the oligomer, so plant laboratories observe a carboxyl end-group drift that cannot be explained by incomplete PTA conversion alone. In a continuous train targeting a final PBT carboxyl end-group concentration below 30 mmol/kg when measured by titration methods referenced to ISO 1628-5:1998, an unaccounted THF loss of 1 kg/h corresponds to approximately 0.014 kmol/h of BDO consumption; for a line producing 10 t/h of PBT, that loss shifts the end-group balance by several millimoles per kilogram and may move the final polymer outside the viscosity and hydrolytic stability specification. The atmospheric esterifier overhead typically contains 2–8 wt% THF and 0.5–3 wt% free BDO when a partial condenser and reflux split are used to limit BDO carryover; however, published data for specific overhead compositions from commercial PBT esterifiers remains limited because the values depend heavily on reactor pressure, reflux ratio, column internals, and the local concentration of dissolved acid species. Direct measurement of THF in the esterifier overhead condensate is therefore mandatory for molar balance closure.

What Are the Primary Control Variables That Determine THF Formation Rate?

Among the controllable process variables, temperature exerts the largest exponential influence on THF formation. Given the activation energy range of 85–120 kJ/mol, a temperature reduction of 5 °C inside the esterification window of 225–245 °C lowers the THF rate constant by roughly 15–25%, while the corresponding decrease in esterification rate can be compensated by a modest increase in residence time or by improved paste homogenization before the reactor. The acid source is not limited to undissolved PTA; the carboxylic acid end groups of the oligomer continuum maintain catalytic acidity after PTA dissolution is complete, so THF formation continues during the later stages of esterification and into prepolycondensation. Water removal drives both esterification and BDO dehydration; the same vacuum or gas-stripping action that removes esterification water also shifts the BDO-to-THF equilibrium toward THF, creating an inherent process conflict that cannot be resolved by temperature alone. Residence time in the esterification cascade is the second variable: continuous units with two or three stirred reactors in series typically operate with total average residence times in the esterification stage of 2–5 h, and higher residence time improves PTA conversion but increases cumulative THF loss. The molar BDO-to-PTA feed ratio is the third variable. Higher BDO ratio increases the concentration of unreacted BDO and the first-order driving force for cyclization, whereas lower ratio reduces THF formation but increases oligomer melt viscosity and mass-transfer resistance. Commercial direct esterification designs therefore operate between 1.40 and 1.70 mol BDO per mol PTA, with the lower portion of the range typical of high-torque disc-ring reactors and the upper portion common in paste-fed stirred tank cascades. The onboard control system uses these three variables in a constraint-based scheme: the temperature is set by the minimum PTA conversion, the residence time by the allowable esterifier inventory, and the BDO ratio by the maximum acceptable melt viscosity. Table 1 summarizes the directional responses used in plant control schemes.

The overhead system of a continuous PBT esterifier separates THF from water and unreacted BDO and determines whether the recovered BDO stream remains process-compatible or becomes a source of recycled THF that degrades reactor control. The THF-water minimum-boiling azeotrope at approximately 63.4 °C and 94.6 wt% THF under atmospheric pressure prevents complete separation in a single conventional distillation column unless a pressure-swing or extractive distillation step is added. The esterifier overhead is therefore processed in a column with a partial condenser and a reflux split that returns a BDO-rich bottoms stream to the reactor and withdraws an aqueous THF distillate for recovery or thermal oxidation. Structured packing with an HETP below 0.5 m for the THF-water-BDO system is commonly specified to limit pressure drop and residence time at elevated temperature; liquid BDO held in the column sump at temperatures above 150 °C can continue to dehydrate to THF, so sump inventory and reboiler temperature are controlled tightly. The reflux ratio is established from the online top-stage temperature and the measured THF concentration in the condensate; a reflux ratio below 1.0 can reduce stage efficiency and allow excessive BDO carryover into the THF-water distillate, while a reflux ratio above 2.5 increases steam demand without proportionally improving BDO recovery. The returned BDO stream must be analyzed for THF, water, and color-body precursors because recycled THF in the BDO feed inflates the apparent BDO input and confuses the molar balance unless the recycle composition is measured with an online gas chromatograph having a sample cycle time no longer than 10 min. The recovered THF-water distillate normally cannot be discharged directly to an industrial wastewater treatment plant without control of volatile organic emissions; the THF concentration in the overhead water is reduced by distillation or steam stripping to meet the facility’s permitted limits under EU Directive 2010/75/EU or equivalent local VOC regulations.

Catalyst Hydrolysis, Carboxyl End-Group Drift, and Commercial-Grade BDO Quality Boundaries

Titanium tetra-n-butoxide and other tetraalkyl titanates used as esterification and polycondensation catalysts are sensitive to water. When the water content of the BDO feed or the paste exceeds approximately 0.05 wt%, the hydrolysis of the titanate to titanium dioxide reduces active catalyst concentration and produces particulate titanium species that can affect filterability and polymer clarity. The catalytic activity loss raises the required esterifier residence time and can lead operators to increase temperature, which in turn accelerates THF formation and creates a negative control loop. Commercial BDO for PBT direct esterification is therefore specified with water below 0.05 wt%; additional limitations on carbonyl compounds and nitrogenous impurities are imposed to avoid color bodies and catalyst coordination losses. The carboxyl end-group concentration is a dynamic function of esterification conversion, THF backbiting, and melt residence time. In a direct esterification train, the carboxyl end-group concentration typically falls during the esterification stage as PTA dissolves and reacts, then can rise again if the melt is held at high temperature under insufficient vacuum because THF-forming backbiting regenerates carboxyl groups. Plant laboratories track this drift by end-group titration according to ISO 1628-5:1998, and the result is used to adjust the vacuum profile of the prepolycondensation stage. Final PBT grades for injection molding are usually characterized by melt volume-flow rate at 250 °C under a 2.16 kg load according to ISO 1133-1:2022, and by intrinsic viscosity according to ISO 1628-5:1998; both measurements are sensitive to the end-group ratio and molecular weight distribution. BDO containing amine-based additives or nitrogenous impurities should be excluded, because amines neutralize the acidic catalyst and can form colored imine or amide byproducts; strong coordination of amines to titanium also reduces catalyst activity. These incompatibilities set operational boundaries that are especially relevant when recycled or bio-based BDO is introduced into an existing PBT line.

During the transition from atmospheric esterification to prepolycondensation, the reactor pressure is reduced from near-atmospheric to 50–100 mbar, and the melt begins to release residual BDO, dissolved THF, and water of polycondensation. The vacuum overhead stream from the prepolycondensation reactor is frequently richer in free BDO than the atmospheric esterifier overhead, with BDO-to-water mass ratios that can exceed 1:1 depending on oligomer chain length and pressure ramp rate. Failure to measure this stream with a dedicated mass flow meter and composition analyzer is a common source of molar balance error exceeding 5%, because the BDO stripped in prepolycondensation is no longer available for transesterification but is often collected in a central vacuum condensate tank without a flow totalizer. The final polycondensation stage at pressures below 5 mbar strips additional BDO, low-molecular-weight oligomers, and cyclic species into surface condensers and cold traps; these low-volatile fractions may not be detected by an online gas chromatograph configured only for THF and water, so periodic liquid-phase analysis by Karl Fischer titration per ASTM D6869-17 and gas chromatography-mass spectrometry of the heavier condensate is required. A complete molar balance therefore includes the atmospheric esterifier overhead, the prepolycondensation vacuum condensate, the final polycondensation cold-trap contents, and vent losses from the THF recovery column. Without this multistream accounting, the apparent BDO consumption per ton of PBT will appear to increase during high-vacuum operation, and the recycle stream may be incorrectly adjusted.

When Tetrahydrofuran Accumulates in BDO Recycle and Shifts the End-Group Balance

If the overhead recovery column fails to strip THF from the BDO-rich reflux and recycle stream, the recycled BDO will contain increasing THF concentrations that are returned to the esterifier. The presence of THF in the feed BDO does not simply dilute the diol; it adds a low-boiling component that flashes across the esterifier and suppresses the partial pressure of water, changing the temperature profile and the apparent reflux requirement. More importantly, recycled THF is measured as BDO-equivalent flow if the feed analyzer uses density or refractive index as a stand-alone signal, and the resulting overstatement of BDO feed can mask real BDO losses until the end-group balance drifts. The corrective control strategy uses the online GC measurement of THF in both the fresh BDO feed and the recycled BDO stream to compute the true BDO molar flow as n_BDO,true = n_BDO,measured − n_THF,recycle, with the THF concentration expressed as mole fraction. The esterifier overhead THF measurement is then compared against the true BDO feed to close the molar balance in real time. If the recycle THF concentration exceeds approximately 0.5 wt%, the reflux ratio of the recovery column should be increased and the reboiler steam flow adjusted; if the concentration exceeds 2 wt%, the recycle should be diverted to a storage tank and the recovery column reboiler inspected for fouling. These limits are operational setpoints derived from industrial DCS strategies and may vary with column pressure and feed composition, but they provide a practical boundary for maintaining molar balance control.

The distributed control system of a continuous PBT direct esterification line implements molar balance closure as a data reconciliation loop that uses Coriolis mass flow meters on the PTA paste, fresh BDO, recycled BDO, overhead condensate, and vacuum condensate streams; online gas chromatographs on the overhead and recycle streams; and loss-in-weight feeders for solid PTA. The Coriolis meters provide mass flow accuracy of approximately ±0.1–0.2% of rate, while the online GC measurements for THF and water typically have relative standard deviations of 1–2% under frequent calibration with certified gas and liquid standards. The calculated BDO accountability is updated every 5–15 min and compared with a moving average of the fed BDO and the measured BDO-equivalent losses. When the accountability falls outside the range 98–102%, the control system issues a deviation alarm and the operators are directed to inspect the vacuum trap inventory, the recovery column reflux split, and the BDO recycle analyzer. This reconciliation prevents the common error of attributing molar balance drift to a single leak or instrument offset when the underlying cause is a change in THF formation rate due to temperature or catalyst variability. In plants with THF recovery, the recovered THF flow is not used as a control variable because it is a derived stream whose flow depends on recovery column operation; instead, the esterifier overhead THF concentration before recovery is the primary measured disturbance.

Thermal Degradation of 1,4-Butanediol in High-Temperature Esterification Loops

BDO is subject not only to acid-catalyzed dehydration to THF but also to thermal degradation and oxidation side reactions when exposed to high-temperature wetted surfaces in the esterification loop and hot-oil reboiler. The dehydrogenation of BDO can yield small amounts of hydroxybutyraldehyde, acetals, and color-forming unsaturates that are concentrated in the recovery column bottoms and can accelerate yellowing of the final PBT. The formation of these byproducts is promoted by wall temperatures above 180 °C in the reboiler and by long hold-up times in the column sump; therefore, reboiler design and sump inventory are controlled to limit thermal degradation. In addition, oxygen ingress into the hot BDO system must be prevented because dissolved oxygen can initiate radical oxidation that increases the carbonyl content of the recovered BDO and reduces the effectiveness of the titanium catalyst. Published data for specific degradation byproduct yields in commercial PBT direct esterification is limited because the byproduct profile is strongly influenced by plant-specific piping geometry and hot-oil temperature uniformity. Nevertheless, the same online GC and periodic titration methods used for THF balance also track the main oxygen-containing impurities, and the recovered BDO is periodically purged or redistilled to maintain color and acidity below the limits required for fiber and engineering thermoplastic grades. The operating window is therefore bounded at the lower end by the temperature required for PTA conversion and at the upper end by the combined penalties of THF loss, degradation byproduct formation, and titanate catalyst deactivation.