Qualifying POM Regrind: Tests to Run Before Adding It Back

Polyoxymethylene (POM) regrind recovered from sprues, runners, and rejected moulded parts cannot be assumed to behave as virgin resin merely because the polymer is chemically identical. The acetal backbone is labile under combined thermal, oxidative, and shear stress, and each processing history shifts molecular weight distribution, consumes stabilisers, and generates formaldehyde or formic acid species that can accelerate further degradation. A qualification programme must therefore generate comparative data against the virgin lot using recognised test methods before regrind is reintroduced into production feedstock. On production-scale injection moulding machines with clamping forces between 800 kN and 25,000 kN, hot-runner sprue and cold-runner scrap typically carry one or two thermal histories; subsequent pelletising, conveying, and drying add further residence time and surface contamination risk.

Thermo-oxidative Chain Scission and Melt Flow Drift

The first measurement is melt volume-flow rate or melt mass-flow rate according to ISO 1133-1:2022 or ASTM D1238. POM homopolymer is conventionally measured at 190 °C under a 2.16 kg load using a 2.095 mm die. Injection-moulding grades commonly range from 8 cm³/10 min to 40 cm³/10 min, although the exact value depends on comonomer content and nominal molecular weight. Regrind that has undergone thermo-oxidative chain scission will generally show an MVR increase relative to the virgin reference; a decrease can indicate crosslinking, contamination, or high-fines packing in the melt indexer. A practical qualification gate is an MVR shift of no more than ±10 % from the virgin lot, but this tolerance must be validated against the moulding machine’s pressure-volume-temperature response and part dimensional stability. For thin-walled parts or tight mass limits, the shift may need to be tightened to ±5 %.

Melt flow ratio measured at a high load of 21.6 kg and at 2.16 kg gives additional sensitivity to molecular weight distribution. A narrowed ratio suggests loss of the high-molecular-weight tail; a broadened ratio suggests chain extension or gel-like contamination. Capillary rheometry using a 1 mm capillary die at 190 °C provides shear viscosity at processing-relevant shear rates from 100 s⁻¹ to 10,000 s⁻¹. If the shear viscosity at 1,000 s⁻¹ falls more than 15 % below the virgin reference, the blend may exhibit sink marks, flash, or screw recovery drift in production. High-temperature gel permeation chromatography in hexafluoroisopropanol can quantify number-average molecular weight; a reduction below 70 % of the virgin reference is generally associated with a measurable drop in tensile elongation, though published data for specific grade configurations may require internal correlation.

Moisture and volatile content must be measured before reintroduction because POM is not hygroscopic in the manner of polyamide, but surface condensation and contamination from storage can introduce water that hydrolyses the acetal linkage at melt temperatures above 210 °C. Karl Fischer titration according to ISO 15512:2019 method A or a suitable moisture analyser calibrated against that method should be run on the regrind after pelletising and before silo transfer. If moisture exceeds 0.10 %, pre-drying in a dehumidifying hopper dryer at 80 °C for 3–4 h is typical for many injection-moulding grades, but the dryer dew point must remain at or below −30 °C. Drying above 100 °C is not recommended for stabilised POM unless the grade-specific datasheet permits it; prolonged high-temperature exposure can deplete the formaldehyde scavenger and shift the melt flow rate before a single moulding cycle is recorded.

Residual volatiles include formaldehyde, trioxane, and low-molecular-weight oligomers. Thermogravimetric analysis at 10 °C/min under nitrogen from 30 °C to 350 °C quantifies low-temperature mass loss. A regrind lot that loses more than 0.5 % mass before 250 °C may indicate degraded additive packages or surface contamination; this material should be isolated and tested by purge-and-trap gas chromatography–mass spectrometry to identify the volatile species before adding it back. The TGA residue at 600 °C further supports inorganic contamination detection. A gravimetric moisture balance with 0.001 g resolution is insufficient for water content below 0.05 %; coulometric Karl Fischer titration with an oven autosampler set at 160 °C gives more reliable data for compounded POM containing carbon black or similar pigments that interfere with direct titration.

Which Mechanical and Impact Tests Detect Recyclate Embrittlement?

Melt flow data alone do not reveal the solid-state consequences of molecular weight loss; tensile and impact tests must be performed on specimens moulded from the candidate regrind-virgin blend using a consistent injection temperature profile, screw speed, and mould temperature. Tensile properties are determined according to ISO 527-2:2012 using Type 1A specimens conditioned at 23 °C and 50 % relative humidity for at least 40 h. The key acceptance indicator is not tensile stress at yield alone, which may remain above 60 MPa for many POM grades even after moderate degradation, but elongation at break. A regrind fraction with hydrolytic or thermo-oxidative damage typically loses elongation at break before yield stress; retention below 70 % of the virgin reference indicates that the material should be either down-converted to less demanding applications or blended at a lower ratio.

Notched impact testing according to ISO 179-1/1eA at 23 °C and, if service conditions demand, at −30 °C provides a stronger discriminator for contamination. POM notch sensitivity means that small particles such as aluminium flakes from regrind handling, glass fibre fragments from cross-contamination, or carbonised polymer from hot-runner degradation create local stress concentrations that lower notched Charpy values. A reduction in notched impact strength below 80 % of virgin should trigger microscopy and ash analysis. If the regrind contains glass-filled POM or flame-retarded POM, mechanical acceptance must be normalised against the correct virgin subfamily; substituting a homopolymer reference for a copolymer grade may produce false failures.

Instrumented falling-weight impact or puncture tests are supplementary but not always required. When a moulded part is used in snap-fit applications, a notched Charpy result may overstate failure risk because snap-fit failures propagate from tensile stress concentrations in the hinge region. In these cases, a tensile test on a thin specimen or a flexural test according to ISO 178:2019 is more useful. The qualification report should record injection screw recovery time, cushion position, and peak injection pressure during specimen moulding because these process parameters can indicate viscosity drift before mechanical results are available.

Contamination and additive depletion require a separate analytical sequence because these variables do not always appear in melt flow or mechanical data at low concentrations. A lot of POM regrind stored in open boxes near a powder-moulding area may contain fine particulates that nucleate crystal growth and reduce toughness without changing MVR significantly. Ash content according to ISO 3451-1:2019 method A at 650 °C should be run on a 5 g sample and compared with the virgin lot. An increase of more than 0.05 % absolute is sufficient to warrant scanning electron microscopy or energy-dispersive X-ray spectroscopy on the residue. Magnetic separation before pelletising is not a substitute for ash testing because non-ferrous contamination from regrind blades and aluminium trim scrap is common in post-industrial POM streams.

Stabiliser depletion is more difficult to quantify directly. POM homopolymer and copolymer grades contain formaldehyde scavengers, acid scavengers, and antioxidants that are consumed with each heat history. A progressive increase in formaldehyde emission from a melt sample is an indirect measure of scavenger depletion. Melt-outgassing tests can be performed by heating a dried sample in a closed vial or headspace sampler at 160 °C for 30 min and analysing the headspace for formaldehyde using a calibrated detector tube or high-performance liquid chromatography after derivatisation. If formaldehyde concentration from regrind is more than double the virgin lot, the material may generate odour, plate-out, or mould corrosion during subsequent processing. In granulate form, formaldehyde emission can also be screened by climate chamber testing according to VDA 275 or equivalent automotive interior methods, but the measured values are highly dependent on pellet size, equilibrium time, and temperature.

Practical processing audits also matter. A single lot of POM regrind should be processed on a production injection moulding machine for at least 50 cycles with a constant cushion, and the nozzle melt temperature should be recorded using a handheld thermocouple or infrared pyrometer calibrated against a reference melt-temperature sensor. If the melt temperature at the nozzle deviates more than 5 °C from the virgin reference at identical barrel settings, frictional heating in the screw may have changed due to the regrind particle size distribution, or the material may be degrading in the compression zone. A twin-screw extruder with a 40:1 L/D ratio and vacuum degassing at −0.08 MPa or lower can sometimes recover mildly volatile-loaded regrind, but degassing alone cannot restore lost molecular weight.

When Regrind Exceeds 20 wt% in a Twin-Screw Feedstock Blend

When the addition ratio exceeds 20 wt%, the qualification logic changes from simple batch release to continuous process capability. The blend of virgin and regrind should be prepared in a loss-in-weight feeder with an accuracy of ±0.5 % at the target throughput, and the pellet feed should be homogenised before entering the extruder hopper. A single-screw extruder with a 24:1 L/D may be adequate for melt blending, but a co-rotating twin-screw extruder with a 40:1 L/D and a side vacuum vent is preferred because it can disperse regrind fines, remove low-level volatiles, and reduce unmelted particle passage. Barrel temperatures should follow the virgin POM processing range, typically 180 °C to 220 °C with the die at 190 °C to 210 °C, depending on grade. Screw speed should remain within the resin supplier’s recommended range, often 100–300 rpm for medium-size compounding lines, but the specific shear rate must be validated by measuring melt temperature at the die.

The final compounded pellets should be re-tested using the same qualification matrix as the initial regrind lot. Process capability indices can be applied to MVR and tensile elongation; a Cpk below 1.33 for MVR over 30 consecutive batches indicates that the blend ratio, feeder accuracy, or regrind variability is too high for dimensionally stable production. A deviation in notched impact strength greater than 10 % across the same batches often correlates with inconsistent regrind particle size. Sieve analysis of regrind with a 6.3 mm screen and a 2 mm screen should be recorded; POM regrind with a high proportion of fines below 1 mm can feed inconsistently and create melt blockages in the hopper throat.

Published data for specific POM copolymer/homopolymer regrind blends at ratios above 30 wt% is limited, especially for flame-retarded or impact-modified grades. In-house design of experiments should therefore treat regrind content, screw speed, and moisture content as independent variables, with MVR, tensile elongation, and notched impact as responses. A 2³ factorial design with centre points is often sufficient to establish a local operating window. The operational boundary for re-additisation is also grade-specific: adding a stabiliser masterbatch to reclaimed POM can improve melt stability, but it must be matched to the base resin because amine-based stabilisers can react with residual formaldehyde or form coloured condensation products. If the regrind lot originates from a different supplier or contains an unknown comonomer structure, infrared spectroscopy or differential scanning calorimetry to identify the melting peak may be required before compounding; blending homopolymer and copolymer POM is industrially possible but shifts crystallisation temperature and may alter mould shrinkage.