BEST publisher Vic Giles spoke to Glenn Tracey, portfolio director – Industrial Innovation and Circular Economy, Smiths Detection and Alexis Thoelen, general manager, Xbat.ai to learn more about battery identification – how they identify and separate end-of-life batteries for recycling.
As battery volumes surge across Europe, the ability to identify and separate end-of-life batteries safely and at scale has become a critical challenge for the recycling sector. Smiths Detection, working closely with Research Institute Vito and Belgian battery sorter Sortbat through the Xbat.ai joint venture, is applying multi-sensor, AI-driven technology to this fast-evolving space.
Smiths Detection is best known globally for aviation security, where its X-ray screening technologies are deployed at airports worldwide. The company is now transferring that proven detection and imaging expertise into the circular economy, applying industrialisation capability and global support infrastructure to battery recycling.
At the heart of this work is Smiths Detection’s Circular Economy and Industrial Innovation venture, focused on scaling solutions that can be deployed globally and supported throughout their operational lifecycle. That ability to industrialise technology, from manufacturing through service and support, is a key differentiator as recycling infrastructure expands.
The collaboration that led to the current battery sorting system began through Vito, which had already been working alongside Sortbat on practical, facility-based approaches. For Smiths Detection, the close involvement of an experienced end user was essential. The technology was designed inside an operational sorting plant, reflecting real-world constraints, safety requirements and throughput demand, rather than purely theoretical design assumptions.
Safety remains central when handling end-of-life batteries. While portable batteries are not actively discharged or deactivated, risk is mitigated through plant-level precautions and system design. Larger and higher-risk batteries are separated early, with dedicated handling and storage protocols intended to reduce the likelihood of incidents.
The system is designed to handle all portable batteries as defined under EU legislation below 5kg, including battery packs. Installed initially at Sortbat’s Belgian facility, it has undergone commissioning and optimisation to align performance with real operating conditions. Rejects are analysed continuously, enabling ongoing model refinement and consistent improvement over time.
From the outset, sorting accuracy exceeds 98%, and in normal operation typically surpasses 99%. Crucially, the system is engineered to manage risk. Where uncertainty arises, it makes conservative decisions to prevent hazardous materials, such as cadmium, entering the wrong recycling stream. Badly damaged batteries are rejected entirely, reflecting Sortbat’s operational experience and safety protocols.
Currently, lithium batteries are sorted into primary lithium and lithium rechargeable fractions. Further separation of rechargeable chemistries, including distinguishing LFP from other lithium-ion types, is technically possible and planned as the market evolves. While LFP remains relatively rare in portable batteries today, it is expected to become increasingly important. The partners are also exploring more granular chemistry separation to support direct recycling routes, responding to growing demand from downstream recyclers.
The Sortbat installation serves as a flagship deployment, with capacity exceeding 10,000 tonnes per year. The system is modular and scalable, allowing customers to start at lower throughput and expand over time. Using the same core characterisation module, installations can process between 1,000 and 10,000 tonnes annually, while increasing the number of chemistries sorted as requirements change. Seven battery classes are currently handled, with flexibility for future additions.
Before automated characterisation begins, waste, oversized items, and button cells are removed to protect downstream equipment and improve safety, a step that Sortbat estimates reduces overall risk by around 80%. Remaining batteries are then fed into the multi-sensor platform, which combines X-ray, laser and optical technologies. AI fuses these data streams to identify battery chemistry before compressed air ejects each item into the correct fraction.
Unlike systems that rely heavily on labelling, this approach is designed to perform even when labels are missing, damaged or incorrect – a common issue with end-of-life batteries. While improved labelling will matter more for larger batteries such as EV packs, portable batteries cannot be reliably identified by labels alone. By analysing hundreds of data points through multi-sensor AI, the system delivers far greater certainty than visual inspection.
This capability is particularly valuable when dealing with forged, mislabelled or heavily degraded batteries. Even cells that appear intact externally may show internal damage or structural changes under X-ray analysis. Swollen laptop batteries and other compromised cells can be identified with confidence, without relying on subjective judgement or extensive operator training.
By combining industrial-scale deployment, advanced sensing and AI-driven decision-making, Smiths Detection and Xbat.ai are helping to set a new standard for battery identification and sorting, supporting safer operations now while enabling more efficient and circular recycling pathways over time.
