In continuous fiber-grade poly(ethylene terephthalate) manufacture from purified terephthalic acid, the molar ratio of ethylene glycol to terephthalic acid delivered to the primary esterifier is the highest-gain manipulated variable governing both esterification conversion and the accumulation of diethylene glycol as an unintended comonomer. A modern continuous esterification train normally prepares a slurry containing 67–70 wt% PTA in ethylene glycol, equivalent to an EG/PTA molar ratio of approximately 1.15:1 to 1.30:1, before feeding the paste through high-pressure progressive cavity or gear pumps into a first stirred esterification reactor operating at 255–265 °C and 1.2–1.8 bar g. The esterification cascade is designed so that the first reactor achieves 85–92% conversion of the two carboxylic acid functions to ester groups, while the second reactor raises conversion to 94–98% before the oligomer is transferred to the pre-polymerizer. Within this narrow operating region, the free ethylene glycol concentration is deliberately held above the stoichiometric requirement because the reaction mixture must remain pumpable and because unreacted PTA carryover into the polycondensation section cannot be tolerated in fiber-grade resins. The unavoidable consequence is that acid-catalyzed dehydration of free ethylene glycol produces diethylene glycol through the reaction 2 HOCH₂CH₂OH → HOCH₂CH₂OCH₂CH₂OH + H₂O; the resulting DEG is subsequently incorporated into the PET chain as a non-crystallizing comonomer. Since DEG units disrupt chain regularity, lower the melting point, and alter dye uptake in spun yarn, fiber-grade PET producers must control DEG within a tight envelope while simultaneously pushing esterification conversion high enough to avoid excessive carboxyl end group concentrations in the melt-phase product. The acidity of the reacting carboxyl groups provides the proton activity for etherification; therefore the local free carboxyl end group concentration and the molality of free EG jointly determine the DEG formation rate. In a continuously stirred first esterifier, this means that feed ratio changes affect DEG through two competing pathways: a lower EG/PTA ratio reduces free EG molality and tends to suppress the bimolecular etherification route, but the same reduction also slows esterification, increases carboxyl end group survival, and increases slurry viscosity, which can lead to local stagnation, wall fouling, and hot-spot formation on heat exchanger surfaces. Because the esterification rate is mass-transfer-limited in the paste phase at low EG loadings, industrial lines avoid ratios below approximately 1.10:1 unless a specific high-acidity process design with a dedicated recrystallization control scheme is installed. Published data for operation below 1.08:1 on continuous PTA lines is limited, and most licensors confine the guaranteed operating range to 1.15:1–1.25:1 for fiber-grade throughput turndown ratios between 60% and 110% of nameplate capacity.

When EG/PTA Molar Ratio Falls Below 1.15: Process Risks and DEG Response

At an EG/PTA feed ratio below 1.15:1, the mass fraction of PTA in slurry rises above approximately 70 wt%, and the paste can become shear-thickening in the transfer lines between the day tank and the reactor feed pump. Progressive cavity pumps with eccentric screw geometry and closed-loop variable-frequency drives are normally able to maintain suction at ratios down to 1.10:1, but batch-to-batch variation in PTA particle size distribution—typically D50 70–130 µm—may shift the sedimentation rate and create zones of settled solids in the slurry recirculation loop. Under these conditions, the first esterifier can lose local conversion, with the reactor outlet oligomer showing carboxyl end group values that rise from a typical 120–180 meq/kg at a 1.20:1 ratio to above 250 meq/kg at a 1.10:1 ratio when residence time is held at the same value. The elevated carboxyl end group concentration then provides stronger proton activity for the etherification route, so the expected reduction in DEG formation from lower free EG is partly offset if conversion is not maintained. This trade-off explains why simple feed-ratio reduction is not a reliable standalone DEG-control strategy in continuous fiber-grade esterification. Instead, the feed ratio is paired with a fixed residence-time template and a vent rectification scheme that removes water while retaining ethylene glycol in the reactor. The practical lower boundary for sustained operation is 1.12:1 to 1.15:1 for most continuous PTA-based fiber lines; operation below that boundary requires higher temperature or longer residence time, both of which re-accelerate DEG formation. In one production-scale continuous esterifier configuration with a first reactor working volume of 45 m³ and a second overhead condenser returning subcooled EG to the reaction zone, the recorded DEG in the esterified oligomer increased from 1.1 wt% to 1.4 wt% when the temperature was raised by 8 °C to compensate for a feed ratio reduction from 1.20:1 to 1.10:1; this illustrates the process conflict between conversion recovery and DEG selectivity.

Because the etherification side reaction is acid-catalyzed, controlled neutralization of a portion of the free carboxyl end groups is used on many fiber-grade lines to suppress DEG without requiring a large sacrifice in esterification rate. Buffering agents such as sodium acetate, lithium acetate, or small additions of alkali are metered into the slurry or into the first esterifier at concentrations equivalent to a few millimoles per kilogram of PTA; the target is not full neutralization, because some acid activity is required for esterification. The measured response at constant feed ratio is a reduction in DEG of 0.1–0.3 wt% relative to unbuffered operation when the alkali addition is held within a narrow window. Exceeding the window causes incomplete PTA dissolution, elevated oligomer haze, and higher filter pressure in the subsequent melt filter, while underdosing provides no measurable DEG benefit. The buffering strategy is sensitive to the PTA acid value and to the trace sodium already present in the recovered ethylene glycol stream; therefore analytical support for recovered EG aeration, ultraviolet spectroscopy for carbonyl compounds, and ion chromatography for alkali metals is normally integrated into the feed-forward control loop. Strong nitrogen-containing bases are generally avoided because residual amines can accelerate aminolysis and shift melt thermal stability beyond the narrow fiber-spinning tolerance.

What Limits the DEG Suppression Window in Continuous Fiber Grade Reactors?

The DEG suppression window is bounded on one side by the minimum esterification temperature required to keep the first-pass conversion above approximately 88%, and on the other side by the maximum temperature beyond which the etherification side reaction accelerates faster than the esterification main reaction. In continuous PTA-based esterification, this window is often no wider than ±5 °C around the design temperature when the EG/PTA ratio is fixed at 1.20:1. If the temperature is reduced from 260 °C to 252 °C to slow DEG formation, the esterification conversion at the end of the first reactor may fall by 5–10 percentage points, causing an increase in oligomer carboxyl ends that must be compensated in the second esterifier or the pre-polymerizer. The subsequent increase in residence time under vacuum reintroduces the DEG formation pathway, because two hydroxyethyl ester end groups can condense to a diethylene glycol unit during melt-phase polycondensation at temperatures above 270 °C. This second pathway is especially relevant in the finisher, where polymer melt temperature is raised to 275–285 °C under a vacuum of 1–3 mbar absolute to elevate intrinsic viscosity. Thus the control of fiber-grade DEG cannot be localized to the esterification section alone; it must be treated as a coupled esterification-to-finishing problem in which every increase in finishing temperature to correct IV also increases the probability of generating DEG and acetaldehyde. The esterification reactor pressure is available as a secondary manipulated variable, but reducing pressure strips ethylene glycol from the reaction mass and must be compensated by a higher overhead glycol reflux; otherwise the free-EG deficit will depress esterification conversion and expand the carboxyl-end tail of the oligomer.

Vapour-Liquid Partitioning and Recovered Glycol Impurity Loops

The esterification vent stream contains water, ethylene glycol, diethylene glycol, low-molecular-weight aldehydes, and trace terephthalic acid derivatives. The rectification column above the first esterifier is usually operated with a reflux ratio that keeps the water-rich overhead distillate below a specified ethylene glycol concentration—often 0.5 wt% EG in the water discharge—while sending the glycol-rich bottoms back to the reaction system. Because diethylene glycol has a normal boiling point approximately 47 °C higher than ethylene glycol, it tends to remain in the glycol-rich bottoms if the column bottom temperature is not allowed to drift upward. A recovered glycol stream containing 0.3 wt% DEG can contribute 0.02–0.05 wt% DEG to the final polymer when recycled at a typical recovered-EG-to-fresh-EG ratio; this contribution may be small in absolute terms but is large relative to the 0.1–0.3 wt% total DEG suppression margin available to the operator. Therefore the recovered EG train must use on-line refractive-index and gas-chromatographic monitoring for water, EG, and DEG composition. Batch discharge of the column bottoms is usually avoided because it would re-inject accumulated DEG slugs to the paste tank and disturb the steady-state fiber-grade melting point. The column pressure profile is also constrained: increasing column top pressure to gain capacity raises the bottom temperature and can promote additional DEG formation from the glycol present in the column, so the pressure tower is not treated as an independent variable.

Transfer of the oligomer from the second esterifier to the pre-polymerizer shifts the control objective from conversion and DEG selectivity to molecular weight build and residual glycol removal. The oligomer normally has a degree of polymerization of 4–8 and contains free ethylene glycol, water, and esterifiable carboxyl ends; the pre-polymerizer operates at 270–275 °C and 20–50 mbar absolute, where the vacuum strips free EG and water and the melt progresses to a degree of polymerization near 25–35. During this stage, residual free EG can still be transformed to DEG, especially if the melt contains high carboxyl end group concentrations from a low-ratio esterifier. In a typical horizontal finisher with a rotating disc or cage agitator, the specific surface renewal rate is set to remove EG and water while avoiding excessive shear heating; high shear zones at the disc edge can generate local temperatures above the bulk melt temperature, and those hot zones are measurable as an increase in DEG and acetaldehyde. The finish is normally controlled by adjusting vacuum pressure and wall temperature setpoints within the limits of the polymer melt viscosity, which rises from 10–50 Pa·s in the pre-polymerizer to 200–500 Pa·s in the finisher. A fiber-grade product exiting the finisher with an intrinsic viscosity of 0.62–0.68 dL/g per ASTM D4603-18 and a carboxyl end group concentration below 30 meq/kg is then filtered through a melt filter with a typical rating of 20–40 µm before pelletization or direct spinning. Any attempt to increase IV by raising finisher temperature above 285 °C instead of improving vacuum will increase DEG generation more than linear proportion, because the melt-phase etherification route is thermally activated.

Across the DMT Route to Fiber Grade: Contrasting DEG Baselines

Fiber-grade PET produced by the dimethyl terephthalate route has a naturally lower DEG baseline than PTA-based product because the transesterification step does not contain the same concentration of free carboxylic acid groups that catalyze ethylene glycol etherification. In DMT-based lines, the first reactor operates on a DMT/EG molar ratio near 1:2.1 to 1:2.3 with a manganese or zinc acetate catalyst, and the by-product methanol is removed as vapor. The reaction mass remains essentially neutral until the last stages of polycondensation, so the etherification side reaction proceeds at a reduced rate relative to the PTA esterifier. Commercial fiber-grade resins from DMT routes have been reported with DEG contents in the range 0.7–1.1 wt% under standard texturing and spinning conditions, while PTA-based fiber resins typically control at 1.2–1.5 wt%. However, the DMT route requires methanol recovery, transesterification catalyst deactivation, and additional volatile organic compound management, so its lower DEG advantage must be balanced against capital and operating cost in a continuous fiber-grade plant. When a PTA-based line is retrofitted to produce low-DEG fiber resin, the relevant benchmark is therefore not a theoretical zero-DEG polymer but the established PTA process envelope with its intrinsic acid-catalyzed side reaction.

Fiber-grade PET characterization laboratories typically release each resin lot against a certificate of analysis that includes DEG content by ASTM D5815, intrinsic viscosity by ASTM D4603-18, melt volume-flow rate by ISO 1133-1:2022, and carboxyl end group content by ASTM D7409. For polyester staple and filament yarns, a DEG content above 1.8 wt% begins to lower the crystalline melting onset below 248–250 °C, which reduces yarn tenacity and increases the risk of fiber fusion during heat-setting at 180–220 °C. Conversely, a DEG content below 0.8 wt% may reduce dye uptake under boil dyeing with disperse dyes, producing shadow dyeing across fabric panels. In continuous direct-spinning fiber lines, the molten polymer from the finisher is transferred through a polymer manifold with residence-time distribution that can add 0.02–0.06 wt% DEG if the melt temperature is held above 285 °C for more than 20 min. Therefore the DEG specification is not only a reactor outlet target; it is a total melt-residence budget that includes the finisher, transfer lines, and spin beam. Published data for specific direct-spinning configurations is limited because spinneret back-pressure and manifold residence times are proprietary, but production sites frequently tighten the esterifier feed ratio by 0.02–0.04 units and reduce finisher temperature by 2–3 °C when a downstream dyeing lot fails the melting-point or DEG limit.

The use of strong nitrogen-containing bases as DEG suppressors is generally avoided in fiber-grade PET because residual amines can accelerate aminolysis, increase yellowing, and reduce melt thermal stability under direct-spinning residence times. Pre-drying of pelletized fiber-grade resin is required where ambient humidity exceeds 60% relative humidity, because absorbed moisture lowers the IV during remelting and complicates the melt-residence budget for DEG control. The same restriction applies to recovered EG streams: wet EG with water content above 0.2 wt% introduces water into the esterifier, shifting the effective feed ratio and forcing a compensating increase in fresh EG, which may carry an additional DEG impurity load. These operational boundaries are routinely verified by feed-tank moisture measurements, recovered-glycol gas chromatography, and periodic certification of the polymer analytical methods against ASTM D5815 and ASTM D4603-18 reference materials.