Waste4Future: Technology for the complete recycling of plastic waste

From the perspective of creating a closed raw materials cycle, only a fraction of plastics currently get recycled. Large percentages are incinerated for the purpose of energy recovery. However, this inevitably generates CO₂ emissions, which are something that must be avoided in the future because of their impact on the climate and the general need to move away from fossil-based energy production. 

Maintaining the carbon cycle

The aim here is to recycle the carbon contained within the plastic. The first way to achieve this is by optimizing the mechanical recycling process, e.g., by using an optimized AI-assisted sorting technology to specifically separate out those plastic fractions that are less severely affected by aging. In a downstream compounding step, these can then be converted into high-quality granulate for use in conventional plastics processing. Secondly, to deal with those plastic fractions that cannot be melted down to create new, high-quality products, innovative solutions need to be devised for the area known as chemical recycling (pyrolysis and gasification). In this way, these plastic fractions within the flow of recyclables are to be converted into raw materials for the chemical industry.

The overall aim described above can be broken down into the following three subaims:

• To develop an evaluation model for breaking the flow of recyclables down into the two subflows of mechanical and chemical recycling. This is no easy task. Finally, it is necessary to minimize the level of expenditure required — both in terms of energy and the financial costs — to implement a comprehensive sorting system and the subsequent steps (processing via compounding and conversion into chemical raw materials). The input data in the ongoing process consists of data from various sensors (e.g., optical spectroscopy, visual image data and terahertz spectra). On the basis of this data, the “quality” of the flow of recyclables must be estimated in real time. Quality is primarily defined in terms of the aging condition of the plastic parts. In the case of heavily aged parts, chemical recycling is the only viable option. In order to correlate the sensor data with the level of quality, AI-based algorithms have to be developed. Another associated task is to get measurement technology that has not yet been used for sorting (e.g., terahertz technology) fit for providing complementary sensor data.
• To further develop mechanical recycling. This involves turning solvent-based fractionation paths and compounding paths into high-quality plastic granulates.
• To provide and optimize chemical recycling processes. These consist of pyrolysis and gasification. These two thermal processes split the macromolecules of the plastic in the absence of air, converting them into low-molecular chemicals that can be used as starting materials for chemical synthesis, either directly or by means of a downstream reaction.

The Fraunhofer Institute for Structural Durability and System Reliability LBF has a wealth of expertise

In light of its extensive know-how in the area of plastics technology, Fraunhofer LBF is involved in two work packages. To enable the age condition to be correlated with the various items of sensor data, the scientists involved in the project are providing plastic fractions that have undergone defined levels of aging for the purpose of training the AI algorithms. The crucial factor here is to select brand-new plastic that is in widespread use (e.g., in packaging) and can also be found in various conditions of aging within the flow of recyclables.

Furthermore, a compounding line is to be developed at Fraunhofer LBF, where additives are to be subsequently added to plastic fractions in various conditions of aging in a targeted manner with an eye on future applications in order to create high-quality batches with minimal variations in quality. The relevant composition of the additive blends is to be determined using data from the various sorting sensors and using rheological data measured online on the compounder. This can only be achieved with the aid of AI-based algorithms, which must be implemented accordingly.

Other project partners

• Fraunhofer Institute for Optronics, System Technologies and Image Exploitation IOSB
• Fraunhofer Institute for Microstructure of Materials and Systems IMWS
• Fraunhofer Institute for Nondestructive Testing IZFP
• Fraunhofer Research Institution for Materials Recycling and Resource Strategies IWKS
• Fraunhofer Institute for High Frequency Physics and Radar Techniques FHR
• Fraunhofer Institute for Process Engineering and Packaging IVV