Assessment of processibility and properties of raw post-consumer waste polyethylene in the rotational moulding process

2.1 Raw materials

The recyclate material used in this study originated from a mixed plastic post-consumer waste stream which was received by a recycling plant. The material was sorted using a density separation process to separate the mixed stream into individual polymer types. While the aim of the process is to achieve batches consisting of a single type of polymer this is in practice very difficult and batch purity ranges from 70 to 98%. The recyclate material used in this study was a batch predominately consisting of various unknown grades of polyethylene, initially believed to be towards the higher end of the range of purity with around 95% polyethylene content. Testing, however, using differential scanning calorimetry, later confirmed purity to be in the range 57–68% polyethylene. The remaining content is mostly polypropylene but may also include impurities such as pigments, additives, inks, adhesives, and other residual contamination.

Two “baseline” materials were chosen for comparison, these included a natural virgin polyethylene rotational moulding grade (vPE), and reground green pigmented polyethylene which was produced from scrap tanks from a rotational moulding factory (reground green tanks, RGT). The RGT material could potentially consist of one or a mixture of three similar rotational moulding tank grades of polyethylene used in the factory at that time. The RGT material was used to give an indication of the effect of reprocessing, without the additional factors in the recyclate material such as potentially significant contamination, unknown number of grades of materials, a range of processes which have been used to produce the products that have been recycled and unknown use conditions.

2.2 Preparation of polymer powder

The material was supplied as raw chipped flake and therefore required a number of preparation steps to produce a powder suitable for the rotational moulding process. This included a pre-compounding step using a single-screw extruder to produce pellets which could then be ground to a powder and dry-blended with the additives. A further compounding step was carried out using a twin-screw extruder followed by pelletising, grinding and rotational moulding.

Single-screw extrusion was carried out using a Killion 150 KN, 38 mm compounding unit, fitted with a barrier screw to promote mixing of the recyclate material. A single pass through the extruder was used. The barrel temperatures of the single-screw extruder were: 195 °C (die), 195 °C (adaptor), 185 °C (clamp ring), 185 °C (zone4), 180 °C (zone3), 180 °C (zone2), 180 °C (zone1).

The blends were compounded using a Collins ZK 25 co-rotating twin-screw extruder. The die and five barrel heating zones were set to 200 °C with the feed zone set to 180 °C.

Cryogenic grinding (to avoid melting of the PP content in the recyclate) was carried out on a Wedco SE-12-TC pilot plant grinding machine using 12” milling plates with 480 teeth and a gap size of 500 microns. Approximately 2 kg of recyclate pellets were placed in a vacuum insulated stainless steel Dewar flask and liquid nitrogen poured onto the pellets until the liquid level was above the pellet. The pellets were left to soak for approximately 10 min before being manually fed into the grinding machine using a ceramic cup. The resulting powder was collected from the outlet pipe of the grinding head using a pillowcase which allowed the entrained air to escape whilst also capturing the ground powder.

In this study the effect of compounding typical rotational moulding additives as well as compatibiliser (an ionomer of ethylene acid copolymer with a density of 0.94 g/cm³ and an MFI of 1.8 g/10 min) into the material was studied. By way of example, a typical series of material preparation steps were as follows:

1. Chipped flake recyclate dried for 72 h in a Jenco, recirculating, hot air, desiccant dryer at 60 °C. The dew point of the dry air was −40 °C.

2. Dried flake passed through a single-screw extruder and pelletised.

3. Pellets cryogenically ground to produce a powder.

4. Ground powder separated into five batches as shown in Table 1.

5. Additives (typically incorporated into rotational moulding grades of polymer) were ground using a pestle and mortar and blended into batches of recyclate using a Prism Pilot three high-speed mixer.

6. Blends were melt compounded using a twin-screw extruder with the resulting pellets cryogenically ground into a powder.

2.3 Rheology

The rheological properties of the compounded blends and two baseline materials [virgin PE (vPE) and post-industrial reground green tank (RGT)] were determined to assess their likely behaviour during the melting/densification stage of the rotational moulding process. Melt flow index (MFI) testing according to BS EN ISO 1133-1:2011 was initially carried out followed by a more detailed study using oscillatory rheometry according to BS EN ISO 3219:1995, at a temperature of 190 °C. This allowed comparisons to be made between the blends at near zero shear rate which is typical of the rotational moulding process.

2.4 Differential scanning calorimetry

In order to assess the crystallinity of each blend, and to obtain an estimate of the relative percentages of polyethylene and polypropylene within each blend, differential scanning calorimetry (DSC) was carried out according to BS EN ISO 11357-3:2018 on a Perkin Elmer DSC machine. Between 7 and 9 mg of material was used for each sample, and sealed in an aluminium pan. The samples were heated and cooled from 30 to 300 °C, followed by one reheat cycle. The enthalpies of crystallisation of the PE and PP components (∆H) were determined on reheat, and percentage crystallinity of each component was determined by Eq. (1), with the theoretical enthalpies of crystallisation of polyethylene and polypropylene taken as 293 J/g and 209 J/g respectively.

2.5 Powder quality analysis

In the rotational moulding process, there are several indicators of powder quality which have an influence on powder flow and packing in the initial stages of processing, bubble removal, final surface finish and mechanical properties [9], [10], [11], [12], [13], [14], [15]. The main properties commonly measured are the time taken for 100 g of powder to flow through a standard sized funnel (dry flow rate), the packing of the powder (bulk density) and particle size distribution. Dry flow and bulk density testing were carried out according to standard ASTM D 1895 and particle size distribution analysis according to ASTM D 1921.

2.6 Moulding trials

All recyclate and baseline batches were rotationally moulded on a Ferry Rotospeed machine. An aluminium cube test mould (300 × 300 × 300 mm) was used to manufacture parts from which test specimens could be cut. A shot weight of 1.8 kg resulted in a nominal 3 mm wall thickness. The oven set temperature was 300 °C and an arm:plate speed of 8 rpm:2 rpm was used. A datapaq temperature acquisition system was used to measure in real-time the air temperature inside the mould during processing. For each material, two mouldings were produced, one exiting the oven at an internal air temperature of 140 °C and the other exiting at 190 °C. This achieved peak internal air temperatures (PIATs) of approximately 180 °C and 220 °C, respectively.

2.7 Mechanical testing

Tensile testing was performed on a Lloyd LS5 machine according to BS EN ISO 527:2019 at a strain rate of 5 mm/min. Drop dart impact testing was carried out according to BS EN ISO 6603-2:2000 on a CEAST Fractovis falling dart impact testing machine, at both room temperature and −40 °C.

3.1 Rheological properties

The difference in melt flow rate (Table 2) amongst the blends ranged from 2.80 g/10 min for post-industrial waste (reground tank) to 4.14 g/10 min for recyclate without additives or compatibiliser. This was encouraging as these were therefore all within a range that would be considered appropriate for use in the rotational moulding process, where typical values range from 2–10 g/10 min [16]. Whilst melt flow rate is a commonly used metric within the industry, the rotational moulding process is characterised by very low shear rates and therefore further analysis of the rheological properties was carried out using oscillatory rheometry.

The results from oscillatory rheometry are shown in Figure 1 where it was observed that blends containing recyclate exhibited much higher values of zero shear viscosity compared to the two baseline rotational moulding grade powders. This is known to have a major influence on the processibility of the resins by making it more difficult for air molecules within bubbles to dissolve into the viscous polymer melt that surrounds each bubble. The polyethylene contained within post-consumer recyclate is likely to contain a high proportion of high density grades with low MFI and therefore it is not unexpected that a more viscous melt would be observed. The longer polymer chains associated with higher viscosity melts resist the diffusion of air molecules through the polymer and therefore reduce the rate at which air molecules can dissolve into the melt across the bubble/melt boundary. This can subsequently have a detrimental effect on mechanical properties, especially impact resistance, as the entrapped bubbles become voids on cooling beneath the melt temperature and act as sites where cracks can initiate and propagate more easily. These bubbles near the outside surface will also result in a lower quality outer surface finish known as ‘pin-holes’. A study by Greco et al. [9] also suggests that sintering in recycled HDPE which is the joining and fusing together of adjacent melting polymer particles will only occur if the viscosity is sufficiently low, which would suggest further that bubbles in HDPE are more difficult to remove from the melt and likely require higher temperatures than typical rotational moulding grades of polyethylene.

Figure 1:

Viscosity of blends at low shear rates.

3.2 Differential scanning calorimetry

Results from the DSC testing (Table 3) showed that there was a notable variability in the composition of the blends, with the percentage of polyethylene varying between approximately 58–68% compared to 32–42% of polypropylene. This is problematic when trying to develop processing recommendations for inherently non-homogeneous material batches. In addition, the purity of the recyclate batches was much lower than expected overall the total crystallinity of the recyclate blends slightly increased with a second extrusion pass, and again with the addition of additives of the polyethylene component in the recyclate blends as additives were introduced.

3.3 Powder quality analysis

Table 4 shows the results from dry flow and bulk density testing. The lowest dry flow time is observed for the virgin PE, increasing by 22% for the RGT powder and between 33 and 53% for the recyclate batches. This indicates that flow of the powder within the mould during the initial heating and lay-up stages may be somewhat reduced compared to virgin PE and could have implications for wall thickness distribution. Previous work [3] has indicated that a dry flow time between 20 and 30 s is preferable in rotational moulding with only the rPE/E + C blend being significantly above this at 34.7 s.

The bulk density of the virgin PE and the RGT powders were identical, but recyclate blends showed a reduction in bulk density of 11–17%. This indicates less efficient packing in these powders which can have a direct effect on consolidation of the melt and bubble removal in the rotational moulding process [17].

The particle size distribution of the various blends (Figure 2) are similar for all recyclate batches however these differ significantly when compared with the two baseline materials. The vPE and RGT powders show a distribution that is more in line with those commonly used in the rotomoulding process, where powders are ground to below 35 mesh size (500 μm maximum particle size) [18]. In contrast, the recyclate powders contained 42.5–55.0% of particles which were greater than 600 μm. This is likely to have a strong influence on powder flow and layup, the behaviour of the melt, and some influence on the quality of the surface finish of the mouldings produced.

Figure 2:

Particle size distribution.

Microscopy of the powder samples (Figure 3) showed that the VPE and RGT powder samples were generally rounded in shape than the recyclate blend samples, which tended to consist of sharper and more irregularly shaped fragments. This is consistent with the results from powder quality tests.

Figure 3:

Microscopy images of powder. Top left to right: vPE, RGT, rPE. Bottom left to right: rPE/E, rPE/E+C, rPE/E+A+C.

The reduction in all three measures of powder quality for the recyclate batches are likely to largely be due to the cryogenic grinding process used which causes brittle fracture of pellet fragments. The vPE and RGT batches on the other hand were ground above room temperature where pellets undergo a series of high-speed, shearing, cuts rather than brittle fracture. Micrographs taken of the various powders displayed much more rounded, smooth and more consistently sized particles for the virgin PE in comparison to the recyclate powders which are of much more angular shape and a greater range of sizes.

3.4 Moulding trials

Visual inspection of the mouldings showed excellent surface finish for the vPE mouldings but very poor internal surface finish for all the recyclate batches. The internal surface was very rough and uneven, and displayed evidence of unconsolidated patches, potentially indicating that not all the material had become molten during the processing stage. The RGT mouldings showed some minor ‘orange-peel’ effect on the inner surface.

The poor surface finish of mouldings incorporating recyclate was consistent with the zero shear viscosity results along with the lower powder quality reported above. However, it is likely that there are other factors which have contributed to the extent of the poor surface finish. These could include the inherent incompatibility of any minor polymer components within the melt, as well as the effect of non-polymer contaminants in the blends e.g. inks and glues and non-ideal processing conditions for the particular polyethylene grades within the recyclate.

3.5 Mechanical testing

3.5.1 Tensile testing

The two baseline materials which did not contain any PP (vPE and RGT) had very similar tensile strength and tensile modulus (Figure 4) and this was not significantly influenced by internal air temperature at removal from the oven. Much lower values of tensile strength were recorded for all blends containing recyclate at approximately 30% of that seen for vPE. The wide range of grades of both PP and PE likely to be contained within the post-consumer recyclate used are thought to be major contributing factors to the reduction in tensile strength seen in these tests. It was also observed that both the additive package and compatibiliser used in this study did not appear to significantly improve the tensile strength. In contrast with tensile strength, the Young’s modulus of the recyclate mouldings did not show consistently lower values when compared with the vPE/RGT samples; in fact, several of the recyclate mouldings had notably higher modulus values. This is thought to be due to the recyclate containing a proportion of PPs and PEs with much lower (fractional) MFIs that are typical of bottles, tubs and pots. This is supported by the higher zero shear viscosities of recyclate blends observed using parallel plate rheology and reported in Figure 1. When comparing the effect of processing temperature for recyclate mouldings, statistically significant improvements in modulus were seen at an oven removal temperature of 190 °C for all recyclate blends except the rPE/E + C mouldings where no difference was observed (Table 5). Whilst higher processing temperatures in rotational moulding would typically be associated with more bubble removal and better mechanical properties as a result, for recyclate blends, it is possible that the polypropylene content within the recyclate samples, which would have a melting temperature range of approximately 160–170 °C, may not have enough time to fully consolidate in the samples processed with an oven removal temperature of 140 °C (reaching PIAT of approximately 180 °C). This may also help to explain the differences in performance between the samples removed from the oven at 140 °C compared to 190 °C.

Figure 4:

Tensile test results.

Table 5:

t-Test showing effect of processing conditions on Young’s modulus.

t-test comparing Young’s modulus values at two processing conditions of 140 °C and 190 °C oven exit temperature	p-Value	Significant?
rPE	<0.001	Yes
rPE/E	0.036	Yes
rPE/E + A	0.001	Yes
rPE/E + C	0.797	No
rPE/E + A + C	0.004	Yes