A process engineer at the Dutch National Test Centre for Circular Plastics (NTCP), Dawn Manjaji, shares insights from a recent industrial Near-Infrared (NIR) study conducted for CEFLEX, which shows how different designs of flexible packaging affect sortability.

Why were the NIR sorting trials of flexible packaging conducted?

The CEFLEX NIR testing programme was essential because it has provided in-depth knowledge into how different flexible packaging materials are detected and classified by an industrial NIR sorting setup, offering a realistic view of the process at sorting facilities for packaging waste.

Ultimately, the trials have given us what we set out to achieve: a new in-depth understanding of the NIR technique and indications on how to potentially optimise sorting – especially for multi-layer multi-materials.

The samples we had gave us a realistic understanding of how the different materials are ‘seen’ by the NIR device and how effectively they are recognised when moving on the conveyor belt at a sorting facility. One of the myths in NIR reflection spectroscopy is that it only measures the surface of a material, whereas in reality, a penetration depth is typically hundreds of micrometres.

What set this testing apart and helped generate new insights?

A unique part of the testing was the wide portfolio of detailed samples that CEFLEX and its stakeholders were able to contribute to the programme. We were able to send around 240 samples in about ten different categories of flexible packaging through the NIR sorter as part of our contribution to the testing at NTCP. The testing process used an NIR sorter, identical to those employed by waste management and recycling companies, to collect data as the samples moved along the conveyor belt.

The team realised early on that the NIR struggled to recognise some clean, as produced, packaging films (samples were unused, pre-consumer materials), so we made adjustments to ensure the trials replicated real-world conditions, such as crumpling the samples. This effect is due to the lack of multiple scattering of the NIR light inside the material, reducing signal intensity. For thin films, scratches on the material enabled this to occur, aiding detection.

What were the main outcomes from the tests?

We could see whether a material combination was detectable and correctly classifiable or not by NIR. For example, the testing revealed challenges in recognising materials with metallisation, which is visible from the outside of the packaging due to being unprinted or unlaminated. This is mainly due to the overexposure of the detector at the NIR, causing high reflection which further results in error for classification. Also, any material underneath a metallisation layer cannot be detected by NIR. These issues were mitigated with printed metallised films where detectability had improved, for example with the presence of white pigment or ink.

What did the tests show regarding multi-layer multi-materials in NIR sorting?

Flexible packaging made from multiple materials are on the market, which can be an issue facing waste and recycling companies. We saw that NIR technology has the potential to help advance sorting and recycling practices and can identify complex material combinations for more effective NIR sorting.

Each material possesses a unique NIR spectrum, containing distinctive characteristics. NIR technology identifies materials by measuring these spectra, to reveal the composition of the material. Multi-layer structures can show a mixed spectrum of all penetrated material layers. For example, if any combination of PET, PE, PP or PA are present in the structures, individual polymers can be identified if the content is above approximately 20% thickness for that material.

Consult the full study and open-source results online: guidelines.ceflex.eu/testing-results

Dawn Manjaji, NTCP

Dawn Manjaji,