Regrind in Polycarbonate Moulding: How Many Cycles Before Impact Numbers Fall?

Closed-loop regrind of bisphenol-A polycarbonate in injection moulding introduces a measurable shift in molecular weight distribution before bulk tensile or flexural properties show comparable change. The notched Izod impact value, determined according to ASTM D256-10 Method A or ISO 180:2000 /1A, is the most sensitive routine indicator of carbonate linkage cleavage because impact fracture energy depends on the high-molecular-weight tail of the distribution. Production moulders typically document a fall in notched Izod before the same lot exceeds an ASTM D1238-20 or ISO 1133-1:2022 melt flow rate shift of ±2 g/10 min or a tensile strength change above 5% at yield when tested to ASTM D638-14. The number of closed-loop cycles tolerated is therefore a function of moisture control, melt temperature, screw recovery residence time, and the proportion of regrind in the let-down.

Why Does Impact Loss Occur Earlier Than Tensile or Flexural Loss?

Bisphenol-A polycarbonate contains carbonate linkages susceptible to hydrolysis at melt-processing temperatures. Feedstock moisture above 0.020 wt% drives chain scission at the carbonate group, producing phenolic and hydroxyl end groups. At barrel temperatures from 280°C to 320°C, each scission event reduces weight-average molecular weight and creates a low-molecular-weight fraction that plasticizes the matrix but weakens craze stability. Notched Izod is governed by plane-strain deformation and craze fibrillation; a reduction in weight-average molecular weight from approximately 30,000 g/mol to 22,000 g/mol can reduce Izod by more than 30% while tensile strength at yield changes less than 5%. Published data for this specific configuration is limited, but the relative sensitivity is consistently observed in injection moulding quality-control laboratories using ASTM D256-10 and ISO 180:2000 /1A.

Chain scission also produces an increase in melt flow rate. ASTM D1238-20 testing at 300°C with a 1.2 kg load is often used as an indirect molecular weight indicator because it requires no solvent dissolution and can be performed at the press. A shift of 1 g/10 min can correspond to a measurable notched Izod change in a high-molecular-weight PC grade, particularly when the virgin lot has a melt flow rate near 6 g/10 min. Moulding trials that rely solely on melt flow rate without impact testing can therefore miss a low-energy impact failure mode.

Feedstock preparation for regrind begins with closed-loop drying in a desiccant dryer with a dew point of at most -40°C and an air temperature of 120°C for 4 h to reach a target moisture content below 0.020 wt%. Sprues, runners, and rejected parts should be granulated through a screen of 6 mm to 8 mm to limit fines below 500 µm, because fine particles melt earlier, absorb more shear heat, and hydrolyse faster in the feed throat. Hopper residence time of dried regrind should not exceed 30 min unless a closed-loop dry-air conveying system maintains a hopper dew point below -30°C. The use of virgin/regrind let-down ratios from 90:10 to 70:30 is common, but a fixed ratio without melt flow testing does not prevent lot-to-lot drift in impact response.

Moisture regain kinetics for amorphous PC at 23°C and 50% RH are rapid at the pellet surface. Dried pellets exposed to ambient air can surpass 0.020 wt% moisture in less than 20 min under humid conditions. A closed-loop vacuum or dry-air hopper is required when relative humidity exceeds 60%. Regrind with increased surface area from granulation absorbs moisture faster than virgin pellets. A moisture analyzer using weight loss at 150°C or Karl Fischer titration according to ISO 15512:2019 should be used at the hopper outlet, not only at the dryer inlet.

When Regrind Content Reaches 30 wt% in Hot-Runner Multi-Cavity Tools

In hot-runner tools with multiple gates, the additional residence time in manifold channels can exceed the barrel residence time by a factor of two to three. At melt temperatures above 310°C, polycarbonate chains undergo thermal rearrangement, Fries-type branching, and CO₂ formation, which shifts the molecular weight distribution and can introduce gel particles. Regrind containing thermally aged solidified melt from a hot runner should be segregated from cold-runner regrind because its impact retention is measurably worse. In a typical 20:1 L/D single-screw reciprocating unit with a 2.5:1 compression ratio and a 0.7 MPa back pressure, barrel residence time should be kept below 15 min during interruptions. A shot size below 25% of barrel capacity is associated with longer residence time, higher melt flow rate drift, and lower notched Izod; published data for this specific configuration is limited, but production records routinely show impact fallout when start-up purge is not discarded.

Hot-runner valve-gated tools require frequent inspection of gate wear because degraded PC regrind can generate carbonaceous deposits that increase shear heating. In multi-cavity tools with unbalanced filling, the cavity farthest from the sprue can receive material with a higher thermal history, producing a lower notched Izod in that cavity. Pack/hold pressure of 40–60 MPa is used; overpacking near the gate can create frozen-in stress that reduces Izod. Sort-and-segregate regrind by cavity or by runner type is therefore used in moulding operations where impact retention must remain above 85% of virgin.

One closed-loop cycle is defined as the sequential moulding, granulation, drying, and re-moulding of the same material without blending with virgin resin. A 30 wt% regrind let-down does not constitute one full cycle for the entire mass; a fraction of the stream sees repeated passes while virgin material dilutes it. The average number of heat histories is lower than the number of closed-loop cycles in the regrind fraction. When a moulder reports five cycles before impact loss, the reported cycle count must specify whether the material was 100% regrind or diluted at each pass. Even in a reciprocating screw, material is not ideal plug flow. A fraction of regrind can reside in dead spots behind the check ring, screw root, or hot-runner manifold for many cycles. This stagnant fraction can degrade severely and release into melt intermittently, causing batch-to-batch Izod scatter. Screw and barrel should be inspected for wear; check ring clearance greater than 0.10 mm can increase backflow and broaden residence time distribution.

Instrumented Puncture and Ductile-to-Brittle Transition Thresholds

Notched Izod values can miss a ductile-to-brittle transition at high deformation rates or at low temperatures. Instrumented puncture testing according to ASTM D3763-18 records force-deflection curves and total energy to failure, which are more sensitive to the high-molecular-weight fraction than standard Izod. Closed-loop regrind that retains 90% of virgin notched Izod at 23°C can show a larger loss at -20°C because the ductile-to-brittle transition shifts upward as molecular weight declines. Moulders supplying industrial guards or glazing applications should therefore use ASTM D3763-18 total energy and ductile failure percentage as additional pass/fail indicators.

A low-temperature notched Izod procedure such as ISO 180:2000 /1A at -30°C can reveal a brittle failure population that is invisible in room-temperature data. The test specimens should be milled from the same gate-to-flow-end distance because regrind with degraded skin can show anisotropic impact behaviour. Higher mould temperatures within 80–120°C improve impact retention by reducing frozen-in orientation, but extend cycle time. Regrind with reduced molecular weight may require a higher mould temperature to maintain ductile failure; the cooling time extension should be assessed against the impact target.

The number of closed-loop cycles before notched Izod values fall by more than 10% is not a single material constant. It depends on the initial weight-average molecular weight of the virgin resin, the end-cap design of the lot, the intensity of drying, the barrel temperature profile, the screw recovery speed, and the hot-runner system. Polycarbonate grades with a virgin melt flow rate of 10 g/10 min at 300°C/1.2 kg may tolerate five cycles of dry, low-residence-time moulding with less than 10% impact loss, while the same grade moulded with 0.040 wt% moisture can show a larger loss after the first regrind cycle. Published data for this specific configuration is limited; the statement should be treated as a process capability boundary rather than a guaranteed cycle count.

Closed-loop production lots should be sampled at the start, middle, and end of each production run. Notched Izod specimens should be conditioned at 23°C ±2°C and 50% ±10% RH for 40 h before testing according to ASTM D618-13. The sample should be tested along the flow direction and transverse to the flow direction because regrind with degraded skin layers may show directionally different impact failure. A single-number lot average without a standard deviation can conceal a low-energy failure population; notched Izod data should be reported with a minimum of 10 specimens and a coefficient of variation below 10%.

Granulated sprues and runners have bulk density typically 0.55–0.65 g/cm³ compared with 0.75–0.80 g/cm³ for cylindrical virgin pellets. Gravimetric dosing is mandatory because volumetric dosing can underfeed regrind by 10–20%. Fines fraction below 500 µm should be limited to <5 wt%; high fines levels create non-uniform melting, gas splay, and local hydrolysis. A dedusting step using an air elutriator or vibratory classifier is used in closed-loop systems above 30% regrind.

For Flame-Retardant and Glass-Filled Grades, Impact Retention Follows a Different Curve

Flame-retardant polycarbonate containing phosphate or sulfonate additives requires additional regrind segregation because the additive package can degrade under repeated thermal cycles and shift both impact and UL 94 performance. Moulding trials that use the same regrind ratio but do not monitor phosphorus content or additive distribution can produce lots that pass impact testing but fail flame-retardant classification. Glass-filled polycarbonate grades show a different damage mechanism: fibre-length reduction during granulation and repeated screw recovery reduces impact energy even when the matrix molecular weight is unchanged. The notched Izod loss in glass-filled regrind may appear after the first cycle, but the failure mode is dominated by fibre attrition rather than hydrolysis; melt flow and matrix molecular weight measurements may not predict the drop.