Polyethylene terephthalate (PET) recovered from post-consumer bottles and thermoforms enters secondary melt processing with a heterogeneous moisture, carboxyl end-group and color history. Processing recycled PET: drying, IV loss and yellowing control are interrelated because residual water drives hydrolytic chain scission, chain scission reduces intrinsic viscosity, and oxidative degradation products contribute to yellowing. For bottle-grade rPET, intrinsic viscosity measured according to ASTM D4603-21 typically falls between 0.72 dL/g and 0.82 dL/g; mixed bales and thermoform-rich streams may enter below 0.65 dL/g. The processing window narrows as melt temperature, residence time and moisture interact, and secondary processors therefore impose separate constraints on feedstock drying, melt-phase history and pellet color.

What Moisture Threshold Prevents Hydrolytic Chain Scission in Recycled PET Flake?

Hydrolysis of the PET ester linkage is a random chain-scission reaction that consumes water and generates one carboxylic acid and one hydroxyl end group per scission. The reaction is not linear with water concentration because carboxyl end groups autocatalyze further hydrolysis. In practice, rPET flake dried to residual moisture above 100 ppm can lose 0.03-0.06 dL/g during a single screw plasticating cycle. Drying to ≤50 ppm is required for general extrusion, while bottle-to-bottle injection lines frequently specify ≤30 ppm because even 0.003 wt% water retains the potential to cleave a measurable fraction of chains at 280 °C. Moisture content is determined by coulometric Karl Fischer titration according to ASTM D7404-19, using sealed-vial sampling to avoid atmospheric regain. Intrinsic viscosity is derived from solution viscosity in a 60/40 phenol/1,1,2,2-tetrachloroethane solvent at 30 °C, as specified by ASTM D4603-21, and the Mark-Houwink relation converts viscosity-averaged molecular weight to solution viscosity.

Carboxyl end group concentration is a parallel quality indicator that quantifies hydrolytic history. Titration according to ASTM D7409-15 typically yields 20-30 meq/kg for bottle-grade rPET but can exceed 50 meq/kg after repeated hot-wet processing. High carboxyl content accelerates subsequent hydrolysis and lowers solid-state polymerization rate. Processors commonly reject feed with carboxyl end groups above 35 meq/kg for carbonated-bottle preforms because two batches with identical 0.76 dL/g can have different end-group distributions and therefore different melt stability.

On a production-scale continuous drying line, amorphous rPET flake cannot be heated directly to 160 °C because the glass transition occurs near 70 °C and unheated particles will stick, bridge and block hopper discharge. Precrystallization is therefore performed in an agitated vessel or crystallizer at 120-140 °C with residence time sufficient to raise flake crystallinity above 30%. This stabilizes the particles for subsequent drying at 160-175 °C in a desiccant-wheel hopper with air dew point of −40 °C or lower and residence time of 4-6 h. Processors observe batch-to-batch moisture variation when hopper design permits channeling or when aftercooling condensation occurs in conveying lines. A common failure mode is rPET flake dried to an average moisture below 50 ppm but fed from hopper periphery with local moisture above 120 ppm, producing sporadic IV loss in molded parts. Vacuum drying offers an alternative for small batch operations: a vacuum tumbler dryer operating at 140-150 °C and ≤30 mbar for 8-12 h reduces residual moisture below 50 ppm without high airflow, but energy cost and batch cycle time limit high-throughput use. Infrared rotary dryers can heat flake to 160-180 °C in 2-3 h, yet localized overheating can initiate thermo-oxidative yellowing before extrusion if temperature control is not uniform. Table 1 presents typical industrial ranges for clean bottle flake; published data for specific configurations is limited and exact values vary with feedstock contamination and screw design.

When Solid-State Polymerization Becomes Necessary to Offset Melt-Phase IV Loss

Melt-phase devolatilization cannot restore chain length once hydrolysis has occurred. A co-rotating twin-screw extruder with L/D 40:1 and a vacuum vent operated at 10-20 mbar can strip dissolved water, acetaldehyde and low-molecular-weight oligomers, but carboxyl end groups remain and can continue chain scission in later processing. Solid-state polymerization is the standard industrial route to raise intrinsic viscosity. In bottle-to-bottle processing, rPET pellets are crystallized, then heated under vacuum or nitrogen at 200-210 °C for 10-16 h, producing an IV increase from approximately 0.72 dL/g to 0.82 dL/g. Solid-state polymerization kinetics are diffusion-limited; larger pellets and high initial carboxyl content slow polycondensation. Residual moisture in the solid-state polymerization feed must remain below 50 ppm because water shifts the equilibrium toward depolymerization. The final IV is confirmed by ASTM D4603-21. Melt viscosity, typically measured as melt volume-flow rate under ISO 1133-1:2022 at 285 °C and 2.16 kg, declines as IV falls, with a change of 0.05 dL/g detectable as a shift in screw torque and injection pressure.

Where solid-state polymerization capacity is not available, reactive chain extenders such as pyromellitic dianhydride or epoxy-functionalized styrene-acrylic oligomers are used at 0.1-0.5 wt% to link carboxyl and hydroxyl end groups. Overaddition creates branching, gel formation and die-pressure fluctuations. The safe addition window is narrow, and dispersion is best accomplished in a side-fed twin-screw configuration with melt temperature controlled below 280 °C. Chain extenders are not a direct substitute for drying; they can mask some IV loss while leaving low-molecular-weight acetaldehyde and carboxyl end groups to affect downstream products.

Thermal and Thermo-Oxidative Yellowing Pathways in rPET Melt Processing

Yellowing in rPET is not governed solely by thermal history. Acetaldehyde generated from β-scission of vinyl ester end groups is the dominant volatile, but acetic acid from residual poly(vinyl acetate) label adhesive and hydrogen chloride from poly(vinyl chloride) contamination also act as chromophore precursors and catalyst poisons. Melt temperatures above 290 °C or hot spots at the screw tip accelerate thermo-oxidative degradation even when the barrel setpoint is moderate. Color is measured on dried pellets by ASTM D6290-19 as CIE L* a* b*, or by yellowness index according to ASTM E313-20. A pellet CIE b* value below 2.0 is commonly targeted for clear bottle flake, while colored or mixed streams may show b* above 8.0.

Because measured rPET pellet color is influenced by crystallinity and surface scattering, pellets must be measured in a standardized cell with constant depth and identical surface finish. Reduction of yellowing begins before extrusion: near-infrared sorting and density separation reduce PVC, PLA and silicone contamination, but small residues remain. Extruder melt filtration with screen packs from 40-120 mesh removes particles above 0.1 mm yet cannot eliminate soluble chromophores. Vacuum venting reduces acetaldehyde concentration by 80-90% at 10-20 mbar. Acetaldehyde specifications below 3 µg/g are common for carbonated beverage preforms, measured by ASTM F2013-10. Phosphite stabilizers at 0.05-0.20 wt% interrupt peroxy radical propagation, but their hydrolysis products can form acidic species if feedstock moisture is not controlled. All additive packages intended for food-contact rPET must comply with 21 CFR 177.1630 or EU Regulation 10/2011 for the intended use.

Table 2 lists the control limits used as acceptance criteria for bottle-grade rPET pellet. They are not universal specifications; individual converters adjust them to the preform design and beverage fill pressure.

Injection Molding Parameter Limits for IV Retention and Color Stability

Preform injection molding of rPET requires the hot runner and nozzle to be held between 270 °C and 280 °C when pellet IV is above 0.72 dL/g. Residence time in the barrel is limited to less than 5 min at maximum throughput, and screw speed is set to avoid shear overheating beyond 5 °C above the barrel setpoint. Published data for specific mold configurations is limited; shear heating in the gating system can dominate the melt temperature, and gate freeze must be balanced against sink and acetaldehyde generation. Drying hopper outlets should feed directly into closed conveying lines because dried rPET reabsorbs atmospheric moisture rapidly. Handling of reground preforms follows the same moisture specification as flake, but repeated heat histories lower IV by 0.02-0.04 dL/g per pass. Melt filters and hot-runner screens must be selected to remove char and degraded gel that otherwise accumulate at the gate and produce color streaks in the perform wall.