Polybutylene terephthalate (PBT) enters post-industrial regrind streams as sprues, runners, short shots, and non-conforming connectors, sensor housings, and appliance components. The virgin polymer is a semicrystalline polyester with melt temperature 223–225 °C, solid density 1.31 g/cm³, tensile strength 50–60 MPa under ISO 527-2:2012, flexural modulus 2.3–2.5 GPa under ISO 178:2019, and notched Izod impact 3–5 kJ/m² under ISO 180:2019. Because PBT is hygroscopic and hydrolytically sensitive, regrind recovery must be managed as a moisture and heat-history problem rather than a simple dilution problem. The operating question for regrind in PBT is how many cycles before strength and colour give way. The answer is not a single number; it is a cycle window dictated by residual moisture, processing temperature, regrind addition level, stabiliser consumption, and the colour tolerance of the final moulded part.

What Property Losses Appear Before Colour Shift Becomes Obvious?

Hydrolytic chain scission dominates PBT degradation when moisture exceeds 0.02 wt% at melt temperatures of 240–260 °C. The terephthalate ester linkage opens to yield carboxylic acid and hydroxyl end-groups, and the carboxylic acid end-groups catalyse further hydrolysis, creating an autocatalytic decline in molecular weight. On production-scale injection equipment with a 25 mm, 20:1 L/D screw and 60–90 s cycle time, viscosity loss appears as an increase in melt volume-flow rate at 250 °C under 2.16 kg load per ISO 1133-1:2022. Tensile strength is generally the least sensitive mechanical indicator because retained crystallinity continues to carry load until extensive chain scission has occurred. Elongation at break and notched Izod impact are earlier indicators; they reflect localised chain scission and brittle crack initiation. Comparative industrial data for unfilled PBT at 30 wt% regrind addition show that after three passes, tensile strength retention may remain at 88–93 % while elongation at break drops to 60–75 % and notched Izod falls to 75–85 %. Published data for a single defined regrind sequence is limited; these ranges are synthesis from ISO 527-2:2012 and ISO 180:2019 comparisons, not universal release limits.

Colour shift does not track molecular weight loss linearly. Discoloration follows stabiliser and antioxidant depletion, oxygen ingress, and temperature overshoot in the plastication unit. A light-grey PBT moulding can exceed a visible mismatch threshold of ΔE 2.0 under ISO 11664-4 or ASTM D2244-21 on the second or third regrind pass, while tensile strength remains above 90 % of virgin control. Therefore, colour “give way” is frequently the release limit in white and light-grey applications, not tensile failure. The exact cycle count depends on whether the material was dried to below 0.02 wt% moisture before each pass, whether hot-runner temperatures were maintained below 260 °C, and whether antioxidant levels were replenished.

Batch-to-batch variation in regrind colour is rarely caused by one degradation route alone. Thermomechanical shear in the screw plastication zone generates local heat spikes, while oxygen ingress through the feed throat and vent ports consumes phenolic and phosphite stabilisers. Once the stabiliser system is depleted, oxidation of butanediol-derived segments forms yellow-brown conjugated carbonyl chromophores. Black specks can appear from degraded material that adhered to screw valleys, check-ring grooves, or hot-runner dead spots and carbonised over several cycles. On a 40 mm reciprocating screw running a 4-s screw-recovery time, a barrel temperature overshoot of 10–15 °C above setpoint can shift colour more than an additional stable regrind cycle at 250 °C. Consequently, closed-loop PBT regrind operations monitor melt cushion, screw recovery time, and injection pressure as real-time degradation surrogates, because optical rejection frequently precedes mechanical rejection in light-coloured moulded parts.

When 30 wt% Regrind Re-enters a 250 °C Melt Stream

At 30 wt% regrind addition, a well-dried unfilled PBT feedstock can usually be reprocessed for 1–3 cycles without catastrophic tensile loss, provided no more than 30 wt% of the shot is regrind and the virgin top-up compensates for stabiliser depletion. The property cliff-edge for many unfilled injection-moulding grades appears between the third and fifth passes. In that window, elongation at break and notched Izod impact fall faster than tensile strength because the molecular-weight distribution broadens and low-molecular-weight chains concentrate at weld lines and gate-freeze zones. The table below shows representative retention windows for unfilled PBT processed on a 120-tonne hydraulic injection moulding machine at 250 °C melt temperature with 40 °C mould temperature.

Published data for a single defined PBT regrind sequence is limited; the tabulated ranges are representative windows synthesised from injection moulding evaluation of unfilled PBT, not absolute pass/fail limits. The practical criterion in production is usually not tensile strength alone but the combination of elongation, impact, and ΔE against the application release specification. Once notched Izod retention falls below 70 % or ΔE exceeds 3.0, the regrind is often diverted to black or non-load-bearing compounds.

On the same 120-tonne machine running a 4-cavity connector tool, nozzle pressure may fall by 10–15 % by the fifth pass as the melt becomes more fluid, while screw recovery time may shorten by 0.2–0.5 s. These signals can be captured by the machine controller and used as release criteria before laboratory tensile bars are tested. Melt viscosity loss under ISO 11443:2021 capillary rheometry is more sensitive than tensile data but is less common in production because it requires clean sample preparation.

Thermal Oxidation Is Not Linear With Cycle Number

Thermal oxidation in PBT regrind follows a stabiliser-depletion curve rather than a simple arithmetic addition of heat histories. Each pass consumes part of the antioxidant package; once the residual stabiliser concentration falls below a formulation-specific threshold, oxidation accelerates. The first two cycles may show only ΔE 0.5–1.5 because the stabiliser is still active. By the third to fifth cycle, the slope of the colour curve can steepen, producing ΔE values above 4.0. Colour can therefore appear stable for several cycles and then deteriorate rapidly in a single subsequent pass. The effect is magnified in hold-up zones: hot-runner manifolds that retain molten PBT for 10–15 min at 260 °C can oxidise material even before the first mechanical recycling cycle is completed. Measurement by ASTM D2244-21 or ISO 11664-4 should therefore be performed on moulded plaques or finished parts, not pellets, because pellet surface oxidation can exaggerate colour variability.

Process Stability Metrics for Closed-Loop PBT Regrind

Pre-drying is the first control point. A desiccant dryer with a dew point of -40 °C to -35 °C, 4–6 h at 120–130 °C, achieves residual moisture below 0.02 wt%; Karl Fischer titration per ISO 15512:2019 is the reference method. Regrind particle size should be 3–8 mm for uniform feeding and to minimise bridging in the hopper. The melt temperature during injection moulding is normally held at 240–260 °C; above 260 °C, degradation accelerates sharply, and below 240 °C, melt viscosity may be too high for thin-wall filling. Barrel residence time should be kept below 5 min; shot size should occupy 25–75 % of barrel capacity to avoid excessive hold-up. Back pressure of 0.5–1.5 MPa improves melt homogeneity but increases shear heating when regrind has lower viscosity than virgin.

Operational boundaries include immediate storage of regrind in sealed containers after drying; uncontrolled ambient humidity causes rapid moisture re-uptake and hydrolysis before the material reaches the feed throat. Regrind streams should be segregated from polycarbonate, PVC, and acetal contamination. Polycarbonate contamination above 2 wt% can generate transesterification, splay, and gloss changes; PVC contamination can release hydrogen chloride and accelerate acid-catalysed PBT hydrolysis. Flame-retarded PBT regrind requires separate handling because brominated and phosphorus-based additive systems shift thermal stability and colour response, and they may not be suitable for non-flame-retarded food-contact or electrical release requirements.

Glass-fibre-reinforced PBT enters regrind streams with a different degradation profile. Fibre attrition during each injection moulding pass shortens glass fibres from an initial length near 250–300 µm to below 150 µm after three cycles, reducing tensile strength and notched Izod impact even when the PBT matrix molecular weight remains acceptable. Ashing per ISO 1172:1996 and subsequent fibre-length distribution analysis are used to track this change; published data for a specific five-cycle glass-fibre length distribution in PBT is limited. In impact-critical automotive housings and structural brackets, 15–20 wt% regrind addition is typically the upper limit, and higher regrind ratios are diverted to non-structural components. Moulding trials on a 150-tonne injection machine with a 30 mm, 20:1 L/D screw indicate that the third pass of 30 wt% glass-filled PBT regrind can retain 85–90 % of tensile strength while notched Izod impact may fall below 70 % of virgin. When 20 wt% glass-filled PBT is re-pelletised on a vented twin-screw extruder with a 40:1 L/D and two atmospheric vents, fibre attrition dominates property loss rather than molecular weight reduction; matrix degradation is secondary but still measurable by rising melt volume-flow rate.