Recycled HDPE Contamination: Testing for PVC, PP and Metal

Recycled HDPE feedstock derived from post-consumer rigid packaging, crates, and industrial scrap contains measurable polymer and metal contamination that shifts melt viscosity, shortens screen-pack service life, and changes oxidative stability. The contaminant classes must be separated by methods aligned to ISO 1183-1:2019, ASTM D5577-19, and ASTM D3418-21 because visual sorting cannot resolve PVC, PP, or dispersed metal particles below 5 mm in flake. Laboratory data are translated into production parameters such as barrel temperature, screw torque, and filtration pressure to prevent batch rejection at the pelletising stage. Sampling follows the gross reduction sequence of ASTM D5577-19 and requires homogeneous flake ground to 2 mm before analytical specimen preparation.

Incoming bales from material recovery facilities exhibit contamination heterogeneity that cannot be captured by hand sorting alone. Film labels, closure liners, and multi-layer bottles are sources of PP and PVC, while shredder wear and residual sensors add metal. The analytical hierarchy should begin with rapid spectroscopic screening, proceed through density and thermal confirmation, and terminate with melt filtration or ash-based quantification. Each step must use test standards because polymer contamination at 0.5 wt% can be process-relevant in high-speed extruders. A single test is insufficient; density separation alone cannot resolve metal fines, and XRF alone cannot distinguish PP from HDPE.

What Degradation Signature Does PVC Leave in Melt Processing?

PVC contamination in rHDPE is thermally hazardous because poly(vinyl chloride) begins dehydrochlorination near 140–160 °C, and autocatalytic HCl release accelerates above 170 °C. rHDPE is routinely compounded at 180–220 °C, meaning PVC domains degrade within the barrel residence time of 30–60 s and generate hydrochloric acid that attacks carbon steel surfaces, catalyzes HDPE chain scission, and forms black specks. In a production twin-screw extruder with 44:1 L/D, contaminated lots often exhibit torque excursions and pressure surges across a 120/80/40 mesh screen pack. Screw wear accelerates when HCl corrosion and abrasive fillers combine; the resulting iron wear debris adds another contaminant stream that appears as dark ash. The exact torque increase depends on PVC domain size and screw design, but the degradation path is consistent with rapid dehydrochlorination once the melt temperature exceeds the PVC stability limit.

Laboratory screening for PVC uses density separation where possible, then elemental confirmation by X-ray fluorescence. The chlorine Kα line at 2.62 keV provides a rapid pass/fail indicator, but the signal is not exclusively PVC because chlorinated flame retardants, cap liners, and salts contribute chlorine. ATR-FTIR confirmation is therefore required; PVC exhibits strong carbon-chlorine absorption bands between 600 cm⁻¹ and 800 cm⁻¹, with a characteristic 1427 cm⁻¹ methylene deformation band. Density separation under ASTM D5577-19 can be adapted from PVC-in-PET routine practice, but rHDPE flake with PVC labels may produce false positives if the label adhesive remains. Quantitative limits for handheld XRF vary with sample thickness, flake orientation, and measurement time; benchtop instruments with helium purge or vacuum improve light-element detection for chlorine but cannot match FTIR specificity.

Density-gradient columns operated according to ASTM D1505-18 separate HDPE, PP, and PVC because their densities differ sufficiently when the column is built from isopropanol–water or ethanol–water mixtures. HDPE flakes concentrate between 0.94 g/cm³ and 0.97 g/cm³, while PP floats at 0.89–0.91 g/cm³ and PVC sinks at 1.30–1.45 g/cm³. However, mixed-color flake, mineral-filled PP compounds, and multi-layer bottle structures create intermediate-density particles that escape single-point density cuts. In such cases, near-infrared sorting with spectral resolution below 10 nm separates PP from HDPE, but black flakes remain challenging because carbon black absorbs NIR. Consequently, differential scanning calorimetry provides the positive confirmation of PP contamination by resolving the PP melting endotherm near 160–170 °C from the HDPE main melt peak near 125–135 °C.

Thermal Profiling by Differential Scanning Calorimetry

DSC analysis under ASTM D3418-21 uses a heating rate of 10 °C/min in nitrogen at 50 mL/min, with first heating to erase thermal history, cooling at 10 °C/min, and second heating for interpretation. Contaminant PP in rHDPE is detected as a melt endotherm between 160 °C and 170 °C, while the HDPE melt endotherm appears between 125 °C and 135 °C. The integrated enthalpy ratio is compared against a calibration curve prepared from virgin HDPE/PP blends at 0.5 wt%, 1.0 wt%, 2.0 wt%, and 5.0 wt%. Calibration must use the same flake size distribution because coarse PP particles bias sampling and produce non-linear enthalpy responses. The method is insensitive to black flakes and does not require density differences. Its limitation is specimen mass: a 5–10 mg DSC specimen cannot represent a lot unless sample splitting follows ASTM D5577-19. Melt flow rate testing per ISO 1133-1:2022 at 190 °C and 2.16 kg tracks the viscosity shift from PP contamination; even 2 wt% PP can reduce MFR by 10–20 %, depending on PP melt flow state and phase dispersion. Injection molding trials with PP-contaminated rHDPE show visible flow lines and weld-line weaknesses when PP domains remain unmelted in the HDPE processing window; published data for this specific configuration is limited.

Monitoring Ash Content After Filtration

Ash content is a coarse but production-relevant indicator of metal, glass, and mineral contamination. The resin is ashed in a muffle furnace according to ASTM D5630-13, with polyolefin samples heated to 600 °C in porcelain crucibles until constant mass. Residual ash is then examined by XRF for Fe, Al, Cu, and Ti. A typical clean rHDPE pellet exhibits ash below 0.05 wt%, while metal-contaminated lots may exceed 0.2 wt%. The method does not identify PVC or PP unless carbonate fillers or chlorine-bearing residues are present, so it is used after melt filtration rather than as a primary polymer contamination test. In a production wash line, a 120/80/40 mesh screen pack captures particles larger than 120 µm, but fine aluminum and oxidized iron can pass through and appear only in ash. Continuous screen changer pressure rise greater than 30 bar across a run indicates excessive particulate loading and warrants offline ash testing. The ash value is also used to detect abrasive contamination that causes screw and barrel wear; hard particles such as glass or sand increase ash without changing melt temperature.

Metal contamination enters rHDPE through residual caps, foil induction seals, and wear debris from shredder blades. Ferrous fragments as small as 1–3 mm are removed by magnetic grizzly or grate magnets in the feed chute, but stainless steel particles from plant equipment may exhibit weak magnetic susceptibility and pass through. Non-ferrous aluminum and copper are not captured by magnetic separators and require eddy-current separators with rotor speeds adjusted to the flake trajectory. XRF can confirm metal type by detecting Fe Kα at 6.40 keV, Cu Kα at 8.04 keV, and Al Kα at 1.49 keV. A plate magnet of 10,000–12,000 gauss is insufficient for work-hardened stainless; rare-earth magnets above 13,000 gauss are specified where stainless is suspected.

When Tramp Iron and Non-Ferrous Residues Require Staged Detection

Ferrous detection at the feed stream uses a three-stage arrangement: a plate magnet at 10,000–12,000 gauss captures low-susceptibility iron before granulation, a grate magnet after granulation removes finer ferrous particles, and a rare-earth tube magnet above 13,000 gauss catches work-hardened stainless. Non-ferrous residues require either eddy-current separators or induction sensors because aluminum closure liners and copper wire fragments are not magnetic. Eddy-current separation on flake depends on particle conductivity, particle size, and rotor frequency; flakes below 5 mm may be ejected with HDPE because the induced field is weak. XRF metal identification complements these devices by detecting Fe Kα at 6.40 keV, Cu Kα at 8.04 keV, and Al Kα at 1.49 keV. Where aluminum is suspected, a density float in water cannot differentiate aluminum from HDPE because aluminum density is 2.70 g/cm³, far above water, but hydrocyclone cut points can be tuned for heavy particles. In practice, production lines combine magnetic capture, eddy-current sorting, and melt filtration to reduce metal to below the detection limit of ash analysis.

Sampling error dominates recycled feedstock testing because contaminant distribution is heterogeneous. A lot sampled from a super-sack may pass XRF but fail after blending because PVC pockets were not represented. A sampling plan should follow ASTM D5577-19 for gross reduction and split fractions, with at least 1 kg of flake ground to 2 mm before analytical specimen preparation. The probability of missing a contaminant at 0.1 wt% in a 1 kg sample can be estimated from binomial statistics, but published data for this specific configuration is limited. Inline NIR and XRF sorters on wash lines serve as first-stage screens, and laboratory methods serve as confirmatory tests. A staged melt filtration trial on a 25 mm twin-screw extruder with 30:1 L/D can reveal process-relevant contamination, but it does not replace polymer-specific identification.

Operational boundaries apply to all test methods. Density separation fails when fillers or multilayer structures produce overlapping densities. DSC fails when PP contamination is below the enthalpy detection threshold or when the PP melt peak overlaps with low-density polyethylene fractions. XRF chlorine screening is not PVC-specific and cannot detect metal particles below the instrument detection limit. Melt filtration is affected by melt viscosity and screen pack condition, so filtration pressure rise alone cannot distinguish PVC degradation products from aluminum fines. Laboratories must therefore combine at least two independent detection principles for each contaminant class.