Optimization of recycled material quality using continuous inline analysis

Plastics processors are under increasing pressure to meet recycling quotas. At the same time, the market does not provide a sufficient supply of quality-tested materials.

Many recycling companies do not have the capacity to carry out their own quality tests and send their materials to external laboratories. For this purpose, random samples of small quantities of the recycled batch produced in granulate form are usually analyzed using specialized offline laboratory methods. This is time-consuming, costly, and risky, because only a small amount of recycled granulate determines whether large batches of a material can be used for the intended purpose at all.

Not so in the ReDigital project. In the future, scientists at the IKK will be able to continuously monitor the extrusion of recycled material using highly sensitive digital real-time analysis. This will enable almost complete observation and AI-supported optimization of the quality and composition of recycled material during the ongoing extrusion process.

This process gives users greater confidence in the use of recycled plastics and gives reason to believe that it will increase both the acceptance of these materials and the demand for them.

Background

The current state of the art in monitoring the quality and chemical composition of recycled plastics involves taking only a small amount of the recycled batch produced in granulate form and analyzing it using random sampling, often with manual, costly, time-consuming, and specialized offline laboratory methods (e.g., melt flow index (MFI) measurement to characterize flow behavior, gas chromatography-mass spectrometry (GC-MS), Fourier transform infrared spectroscopy (FTIR), Raman spectroscopy, etc.). Depending on the method, sample quantities vary from approximately 50 g to the µg range. Recycling companies (often SMEs) cannot afford this internally and usually have random samples of their recyclates tested by an external laboratory for quality approval. As a result, batches are only released after days or weeks, and in the meantime, costs for temporary storage are incurred.

The measurement results of these extremely small sample quantities are used to approve several tens of tons of recycled material. This, in turn, carries the risk that the small sample quantities measured are not representative of the tons of approved plastic recyclates with a very heterogeneous composition of input streams. At the same time, if the specified values are not met in the random samples, the entire batch may not be used for its intended purpose. Due to these risks, many users categorically reject the use of recycled plastic. The approach pursued in the project therefore aims at continuous monitoring of recycling extrusion with highly sensitive digital real-time analysis and enables almost complete analysis and AI-supported optimization of recyclate quality and composition during the ongoing extrusion process.

From a technical perspective, four different inline/online analysis modules will be procured and installed on an existing semi-industrial recycling extrusion line at the IKK in order to determine the chemical composition of the plastic melt and the gas phase simultaneously using different complementary measurement principles. The data generated by the various modules will be collected and used in combination with innovative methods from the fields of artificial intelligence (AI), machine learning (ML), further regression models, and chemometrics for the direct analysis and optimization of the ongoing recycling extrusion process and the development of prediction models. 

Project expenditure: €1,170,000

Duration: January 1, 2026 – December 31, 2027

Contact

Dr Madina Shamsuyeva
Telefon: +49 511 762 18345
E-Mail: shamsuyeva@ikk.uni-hannover.de