In melt conversion of polyethylene terephthalate, residual moisture is not a surface contaminant that vacuum venting can remove after plastication; it is a stoichiometric chain scission reagent. Hydrolysis of the ester carbonyl group consumes one water molecule per scission event, reducing number-average molecular weight and solution intrinsic viscosity. The processing target known as PET Hydrolysis Control: Moisture, Molecular Weight and the 0.02% Rule formalizes this relationship. The 0.02% threshold refers to pellet moisture content on a wet-mass basis, measured before the resin enters the barrel. Above this value, the probability of measurable molecular weight loss during normal melt residence time increases sharply. The resulting degradation produces shorter chains, higher carboxylic acid end-group concentration, lower melt elasticity, and reduced mechanical strength. These effects are irreversible because the extruder cannot reconnect severed ester bonds once hydrolysis has occurred. The 0.02% rule is not a clause in ISO 15512 or ASTM D4603; it is a widely applied upper processing guideline derived from the stoichiometric relationship between water and ester linkages.

What Makes Moisture a Degradation Reactant Rather Than a Processing Nuisance?

Hydrolytic degradation of PET proceeds through ester bond scission. The reaction consumes water and produces one hydroxyl and one carboxylic acid end group. The carboxylic acid end groups subsequently catalyze further hydrolysis, so the degradation becomes autocatalytic. Melt-phase kinetic measurements reported in polymer science literature describe the reaction as pseudo-first-order with respect to ester concentration over limited conversion ranges. Temperature accelerates the rate constant; at barrel temperatures of 270–300 °C, hydrolysis in wet resin can dominate residence-time-dependent molecular weight loss. The consequence is measured by solution viscosity according to ASTM D4603 or ISO 1628-5. A decline in intrinsic viscosity from 0.80 dL/g to 0.72 dL/g is not a cosmetic shift; it corresponds to lower extensional viscosity and reduced melt strength. In production, this appears as edge weave in sheet extrusion, parison sag in stretch blow molding, or short shots in thin-wall injection molding. The moisture concentration in the pellet controls the number of available hydrolysis events before the water is consumed or pumped out of the melt.

The autocatalytic character of PET hydrolysis means that carboxylic acid end groups generated in the barrel reduce the apparent activation barrier for subsequent scission. Thermal degradation also produces acetaldehyde and carboxyl end groups, so wet feedstock and high melt temperature are synergistic variables. At 280 °C, a pellet moisture content of 0.02% may produce acceptable intrinsic viscosity retention for short barrel residence times; the same resin at 300 °C with longer residence time may lose additional intrinsic viscosity because the hydrolysis rate constant increases. The practical processing window therefore couples moisture, melt temperature, and screw residence time rather than treating moisture as a single variable.

Desiccant-Wheel Dryer Performance and Hopper Purge Boundaries

A closed-loop desiccant-wheel dryer supplying air at a dew point no higher than −40 °C establishes the moisture gradient needed for removal. Typical drying temperatures for PET range from 160 °C to 180 °C, with residence times of 4 h to 6 h. The drying air volume, dew point, and residence time are coupled: a dew point of −20 °C cannot dry PET to the 0.02% limit at standard temperature, regardless of residence time, because the equilibrium moisture content remains too high. Dew point measurement at the dryer outlet, not the ambient plant air, is the controlling variable. Hopper purge air must also be dry. Field observations on production-scale sheet extrusion lines show that pellets dried to 0.008% moisture can regain moisture within 30–60 min in an unsealed hopper when relative humidity exceeds 60%. This reabsorption time is shorter than the residence time in many hopper configurations, so the feed throat condition becomes part of hydrolysis control.

Desiccant wheel performance is not constant across regeneration cycles. In production, dew point spikes occur immediately after bed switchover if the regenerated bed has not cooled adequately. A dew point sensor with alarm at −30 °C is commonly installed at the process air outlet. If the alarm is ignored, pellets may be discharged at 0.025–0.035% moisture while the controller still indicates a completed drying cycle. Batch logs should include dew point, drying temperature, and feed-throat moisture rather than dryer timer status alone. Moisture analysis at the feed throat is performed by ISO 15512 or ASTM D6869 Karl Fischer methods; the result must be reported as percent wet basis because the 0.02% rule is a mass fraction. Loss-on-drying instruments may underreport tightly bound water in PET unless the oven temperature is sufficient to release it. Additive masterbatches, regrind, and fillers should be included in the moisture assessment because the rule applies to total system moisture, not only virgin pellets. A silo of virgin chip at 0.010% moisture combined with a regrind stream at 0.040% can push the blended feed above the threshold even though the primary resin was acceptable.

When the 0.02% Ceiling Is Evaluated as a Stoichiometric Limit

If pellet moisture is held at 0.02%, hydrolytic chain scission may still be unacceptable for stretch blow molding grades with an intrinsic viscosity above 0.80 dL/g. The same moisture mass fraction represents a larger number of water molecules per chain in high-molecular-weight resin because fewer polymer chains are present per kilogram. For high-IV grades, industrial drying targets are typically 0.005% or lower. This is not a linear extrapolation of the 0.02% rule; it is a stoichiometric consequence of higher number-average molecular weight. Conversely, filled and reinforced compounds may tolerate 0.02% moisture but fail by other mechanisms because mineral or glass reinforcement introduces additional surface moisture and nucleating effects. The 0.02% threshold must therefore be interpreted as an upper bound for unfilled general-purpose grades, not a universal specification.

The stoichiometric argument can be made explicit. At 0.02% moisture, a 1 kg charge contains 0.2 g water, equivalent to 0.011 mol. If the starting resin has a number-average molecular weight of 20,000 g/mol, the charge contains 0.05 mol chains. Complete reaction of the available water would generate 0.011 mol additional chain ends, reducing number-average molecular weight to approximately 16,400 g/mol. Even partial consumption produces measurable intrinsic viscosity loss. This is why the 0.02% rule remains a ceiling, not a target.

Measuring Molecular Weight Retention with Intrinsic Viscosity and Inline Rheometry

Solution viscosity measurement by ASTM D4603 uses a 60/40 phenol/1,1,2,2-tetrachloroethane solvent at 30 °C. The measured intrinsic viscosity is converted to number-average molecular weight using published Mark-Houwink parameters for PET in that solvent. Intrinsic viscosity should be reported with solvent and temperature because values measured by ISO 1628-5 may differ from ASTM D4603 unless the same solvent system is used. For process verification, melt mass-flow rate by ISO 1133-1:2022 and capillary viscosity by ISO 11443 provide faster proxies, but they do not distinguish hydrolysis from thermal or oxidative degradation. The 0.02% moisture rule is most directly tied to solution viscosity because hydrolytic scission changes end-group concentration and chain length without necessarily generating crosslinked residue. On a 75 mm co-rotating twin-screw extruder with 36:1 L/D and vacuum venting, the vent removes volatile degradation products but does not reverse ester bond scission. Therefore, feed moisture must be controlled before the first barrel zone.

Residence time distribution in a single-screw extruder can span 1–3 times the mean residence time. The long tail means some melt elements see much longer hydrolytic exposure than the average, so a moisture content near 0.02% may pass average intrinsic viscosity retention but fail for the slow-moving fraction. Melt viscosity depends on weight-average molecular weight through a power law with exponent near 3.4 above the critical entanglement molecular weight. Hydrolysis therefore produces a magnified change in melt pressure for a relatively modest change in intrinsic viscosity. A decline of 0.02 dL/g in solution viscosity can shift melt pressure in a sheet die by a measurable amount, requiring downstream gauge adjustment. Inline melt viscosity sensors in a melt pipe after the extruder can detect hydrolysis-induced shifts before laboratory intrinsic viscosity results are available. A drop in melt pressure at constant throughput in a gear-pump-assisted sheet line often indicates a reduction in molecular weight. This production-floor feedback is used to adjust dryer residence time or barrel temperature. In thin-wall PET injection molding, hydrolysis-induced melt viscosity reduction may initially appear as easier fill but leads to lower part molecular weight and tensile strength. This is a false process gain because clamp force requirements decline only after load-bearing chain length is lost.

Thermal Drying Limits and the Cost of Overdrying

Drying residence time is not infinitely adjustable. Below 160 °C, water diffusion from semi-crystalline PET is slow and the required residence time may exceed 8 h. Above 190 °C, pellets can stick, yellow, or crystallize in the hopper. The optimum window for conventional virgin PET is therefore narrow, typically 160–180 °C, with a hopper residence time of 4–6 h. Desiccant wheel performance is confirmed by dew point transmitter at the dryer outlet and by moisture analysis on samples taken at the feed throat. Overdrying is not a neutral action; prolonged exposure to hot dry air can increase acetaldehyde formation during subsequent melt processing and raise energy consumption. The 0.02% rule serves as a release criterion: stop drying when pellet moisture is below the threshold, not when the dryer has simply met a fixed time setting.

Reground post-consumer PET introduces variable crystallinity, moisture history, and additive residues. The same dryer settings that remove water from virgin chip may not remove water from mixed flake with high surface area and trapped fines. Fines increase moisture carryover and can cause localized hydrolysis because their residence time in the screw is longer. Data for specific recyclate streams is configuration-dependent; published data for a single universal drying protocol is limited. The 0.02% rule remains a starting point, but recycled PET with intrinsic viscosity below 0.65 dL/g may require solid-state re-polymerization before melt conversion. Drying alone cannot restore molecular weight that has already been lost. In such streams, moisture control is necessary but insufficient.