CEA-Leti has launched SubFAB Labs, an industry-funded research programme designed to accelerate the development and qualification of technologies for reducing emissions, chemical waste and resource consumption in semiconductor manufacturing.
The initiative has been established with International SubFAB Research Labs (ISRL) and will use CEA-Leti's existing Grenoble pilot-line infrastructure to test monitoring and treatment equipment under conditions representative of industrial semiconductor production.
Its first phase is already under way and will examine gaseous and aqueous emissions, material-use efficiency and process by-products associated with the principal families of integrated circuit manufacturing processes.
HORIBA, TOFWERK and Edwards Vacuum are among the first industrial participants. Their work will address the measurement of chemical waste streams, qualification of analytical equipment and development of techniques for treating emissions from semiconductor processes.
CEA-Leti introduced the programme on 10 November during SEMICON Europa 2026. The initiative is intended to provide an experimental environment in which technologies can be evaluated before manufacturers consider incorporating them into commercial fabrication facilities.
Semiconductor production involves multiple processing stages that deposit, remove, modify and clean materials on wafers. These operations require controlled chemical and physical conditions to create the structures needed for integrated circuits.
Process gases, liquid chemicals and cleaning materials enter manufacturing equipment according to the requirements of each operation. Substances leaving those processes may include unused materials, reaction products and other constituents requiring treatment or controlled disposal.
Much of the infrastructure responsible for handling these streams is located in the subfab, beneath or alongside the semiconductor cleanroom.
Subfab equipment includes vacuum pumps, exhaust management systems, emissions abatement equipment and installations for handling process chemicals and wastewater.
The performance of these systems depends on the manufacturing processes they support. Changes in gas composition, chemical consumption or operating conditions can alter the materials entering the waste handling infrastructure.
Deposition processes introduce substances intended to form layers on a wafer. Only part of the material supplied may become incorporated into the finished structure, leaving residual gases and reaction products that must be managed.
Etching operations use chemical or physical processes to remove selected material. Their exhaust streams can contain different substances according to the materials being processed and the chemistry used.
Cleaning and polishing processes create additional waste streams, including liquids containing chemicals, particles or dissolved material. The composition and treatment requirements of these streams differ from those associated with gas-based deposition or etching.
Consequently, an emissions treatment system must be selected according to the characteristics of the waste it receives. Equipment effective against one substance may not deliver the same result when concentrations or chemical compositions change.
The first phase of SubFAB Labs will run for up to four years, building a reference database of material consumption and by-products across semiconductor manufacturing processes.
Researchers will examine material-use efficiency and the composition of emissions, allowing different processes and waste streams to be characterised before suitable treatment technologies are evaluated.
The programme will also establish a reference monitoring system combining multiple sensor technologies. Its purpose is to measure the composition and concentration of substances as they move through different stages of waste handling.
Reliable measurement is necessary when evaluating treatment equipment. A system's effectiveness cannot be established without understanding the characteristics of the incoming stream and determining how those characteristics change during processing.
Measurements collected at several locations can help distinguish changes associated with individual treatment stages from variations introduced by the manufacturing process itself.
HORIBA will contribute analytical and measurement technologies, participating in the testing and qualification of existing and next-generation sensors.
The equipment will be assessed across different semiconductor processes and waste handling stages. These evaluations will examine whether the sensors can provide useful measurements under operating conditions representative of industrial applications.
TOFWERK, a Bruker company based in Switzerland, will use time-of-flight mass spectrometry to investigate process by-products during extended measurement campaigns.
Time-of-flight mass spectrometry separates ionised substances according to their mass-to-charge characteristics, allowing information about chemical composition to be obtained from their movement through the analyser.
The method can support detailed measurements of complex gas mixtures, although its effectiveness depends on instrument configuration, analytical procedures and the substances being examined.
Applying such equipment to semiconductor emissions presents challenges associated with changing concentrations, mixed chemical streams and the operating conditions found near manufacturing processes.
Extended testing will provide information about measurement stability and suitability for industrial use, including how instruments perform when exposed to different processes over longer periods.
Edwards Vacuum will examine process waste streams and investigate real-time emissions abatement strategies. Its work will focus on improving understanding of the substances entering treatment systems and how their characteristics influence abatement performance.
Abatement equipment is designed to treat selected substances before waste streams are discharged or transferred to subsequent handling systems. Its operating requirements depend on the chemicals involved, the flow rate and the treatment method.
The effectiveness of a treatment process must be considered alongside its wider consequences. A method that reduces the concentration of one substance may produce additional treatment products or affect the requirements of downstream equipment.
SubFAB Labs will therefore assess candidate technologies against environmental and economic criteria, including their effects on waste handling, safety and operating cost.
The experimental platform will operate in a mode representative of high-volume semiconductor manufacturing, allowing partners to conduct prolonged tests without interrupting commercial production lines.
Research carried out in an active manufacturing facility can be difficult where equipment trials risk affecting production availability or require changes to established process arrangements.
A dedicated pilot environment provides greater flexibility to modify monitoring equipment, test alternative treatment methods and collect information over extended operating periods.
The results must nevertheless be assessed before transfer to a commercial factory. Differences in equipment, production volume and waste handling infrastructure can influence the performance of a technology when it is deployed at a larger scale.
Modelling will be used to examine how findings from the Grenoble programme might translate into commercial semiconductor manufacturing operations.
Further research phases are planned to begin from 2027, with four additional phases addressing PFAS treatment, etching, deposition, cleaning and polishing.
PFAS compounds and their breakdown products are an area of particular concern because the regulatory framework surrounding these substances continues to develop.
Research into their treatment requires an understanding of the chemicals present, their behaviour during processing and the performance of technologies intended to reduce their release.
Other phases will examine the environmental effects of manufacturing operations and assess prototype solutions under representative process conditions.
The programme also intends to investigate opportunities to recover critical raw materials from waste streams. The feasibility of recovering a particular material depends on its concentration, chemical form and the resources needed for separation.
Recovery methods must be assessed against their energy, chemical and operating requirements. Obtaining a usable material from waste does not automatically establish that the process has a lower environmental impact than alternative supply arrangements.
CEA-Leti's Grenoble cleanroom operates continuously and contains more than 80 items of equipment supporting the development of semiconductor processes, components and circuits.
The existing pilot infrastructure provides access to a range of manufacturing technologies without requiring SubFAB Labs to reproduce every aspect of a commercial semiconductor factory.
Approximately 20 industrial organisations from the semiconductor value chain supported the concept during industry discussions in 2025, before the formal introduction of the research programme.
SubFAB Labs will complement the SEMI Climate Consortium and the European GENESIS project, which CEA-Leti coordinates. The new programme concentrates on validating technologies within the supporting infrastructure that handles manufacturing emissions and waste.
Initial work in Grenoble will establish the reference measurements needed to select and evaluate future treatment technologies. The first phase may continue for up to four years, while additional research phases are planned to begin from 2027, allowing several development activities to progress in parallel.
The programme's environmental benefits will depend on the results of these trials and the subsequent adoption of qualified technologies in commercial semiconductor production. No measured emissions reduction or material recovery improvement has yet been established for the newly launched initiative.
