Recycling Black Plastics: How Thermography and AI Are Transforming Sorting
Exclusively for K-Mag
A new approach from Fraunhofer IZFP makes it possible to reintroduce previously lost plastic streams into the circular economy
13.05.2026
Black plastics have a visibility problem – because they are not detected by sorting plants. The result: materials that would actually be recyclable end up in thermal recovery. Particularly in the context of the circular economy and rising recycling rates, it becomes clear that a significant stream of recyclable materials is being lost here. This is where a new process developed by Fraunhofer IZFP comes in, designed to make black plastics specifically identifiable.
At a glance
NIR sensors do not reliably detect black plastics, as carbon black absorbs the reflected light.
For the first time, thermography and AI enable identification via heat signatures rather than reflection.
This allows material streams previously lost to be returned to the recycling loop.
Key Takeaway:
Black plastics are not a recycling problem, but a detection problem that can increasingly be solved with new technologies.
Why Black Plastics Fall Through the Cracks in Recycling
The standard method for sorting plastics is near-infrared (NIR) technology. It relies on reflected light – and this is precisely where the problem lies.
"As things stand, black plastics are difficult to sort into pure fractions," explain Kevin Schmitz and Andreas Keller from the Fraunhofer Institute for Nondestructive Testing IZFP. "The go-to method for plastic sorting, hyperspectral imaging in the NIR range, requires sufficient light reflected from the material’s surface."
The main obstacle: carbon black, as a colour pigment, absorbs light almost completely. The consequences:
- hardly any reflection for NIR sensors
- no reliable material recognition
- failure to separate the materials
Alternatives exist, but they are expensive. Schmitz and Keller are well aware of this: "Good sorting accuracy can be achieved with MWIR hyperspectral cameras, but the cost of the camera system alone runs into six figures."
Thermography: Sorting Black Plastics Using Thermal Signatures
A new approach focuses not on reflection, but on emission. Researchers at Fraunhofer IZFP are using thermographic methods to distinguish materials based on their thermal radiation.
The principle:
- Plastic parts are heated in a targeted manner
- A thermal imaging camera captures the thermal signature
- AI analyses the patterns in real time
- The system automatically makes the sorting decision
"Our approach differs in two fundamental ways," say Schmitz and Keller. "The wavelength range under consideration is long-wave infrared, whereas other systems rely on near- or mid-wave infrared."
The decisive factor is the shift in physical perspective: "In the non-spectrally resolved long-wave infrared range, the radiation generated by emission is used to analyse the thermal properties of the materials and thereby distinguish between them."
Mini Glossary
- Near-infrared (NIR): Identification of plastics via reflected light
- Hyperspectral: Analysis of many wavelengths simultaneously
- Thermography: Analysis of thermal radiation
- Artificial intelligence (AI): Pattern recognition and automated classification
From Research to Practice
A working demonstrator already shows how the process can be implemented in practice: plastic parts are selectively heated, thermally detected and automatically sorted on a conveyor belt.
However, there are still challenges to be overcome for industrial application:
- varying particle geometries after shredding
- contaminants in the material flow
- Conveyor speed for high throughput
At the same time, the potential of this approach is becoming clear. For operators of sorting plants, a previously untapped material stream could open up.
"Existing plants can use a system based on this technology to further break down the black waste, which is currently mainly used as a substitute fuel," explain Kevin Schmitz and Andreas Keller. "This allows high-quality plastics such as PA6 to be filtered out and marketed as an enriched stream."
The potential benefits:
- higher recycling rates
- additional revenue streams
- better compliance with regulatory requirements
However, several development cycles will still be required before widespread industrial application is achieved.
What Does This Mean for the Industry?
- Integration of new sensor systems: Thermography-based methods could complement existing NIR-based lines
- Data literacy is becoming increasingly important: AI-supported analysis is becoming a key factor in sorting quality
- Changing plant design: Combining different detection principles is becoming more relevant
- Scaling as the key: speed, contamination and particle diversity remain decisive variables
Context: A Detection Problem, not a Material Problem
Developments show that black plastics do not pose a fundamental recycling problem, but rather a recognition problem.
New sensor technologies, combined with artificial intelligence, are pushing the boundaries of what was previously possible. The key factor now will be how quickly these approaches can be integrated into existing sorting processes and scaled up under real-world conditions.
If this step is successful, material streams that were previously lost could be systematically tapped into in future – with direct effects on recycling rates, resource efficiency and the cost-effectiveness of sorting plants.
Autorin: Elena Blume | K-Mag
Elena Blume schreibt seit acht Jahren für verschiedene Fachmagazine und ist seit 2021 Teil des K-Mag-Redaktionsteams. Besonders interessieren sie Themen rund um innovative Kunststofflösungen, die zur Nachhaltigkeit beitragen und aktuelle industrie- und gesellschaftsrelevante Herausforderungen adressieren.
