For decades, the semiconductor industry has been built around silicon. Its material properties are well understood, and the measurement standards, calibration protocols, and reference materials that support manufacturing have been refined over many years. However, as demand grows for devices that can operate at higher voltages, higher power densities, and in harsher environments, the industry is increasingly turning to new materials.
Silicon carbide is one of the most important of these materials. It is already being deployed in power electronic devices, particularly in automotive applications, where it enables more efficient, more compact, and higher-voltage systems. Yet while silicon carbide devices are entering the market, the measurement infrastructure needed to support their validation and characterisation has not kept pace.
In particular, there is no reliable, standardised way to measure strain in silicon carbide. Strain can be introduced during wafer processing, device fabrication, and packaging, and it can have a significant impact on performance and reliability. At present, different laboratories and suppliers measure strain in different ways, often using techniques developed for silicon that do not translate well. The result is inconsistency, disagreement between measurements, and a lack of confidence across the supply chain.
Without a common, traceable approach to strain measurement, manufacturers struggle to compare materials from different suppliers, identify problems early in production, or communicate reliably about device performance. These issues increase qualification risk and slow the adoption of silicon carbide technologies at scale.
The National Physical Laboratory (NPL), as the UK’s National Metrology Institute, is well placed to address this challenge through its expertise in developing traceable measurement methods and standards. Rather than inventing an entirely new technique, this work focused on adapting and extending well-established measurement approaches so they could be applied reliably to silicon carbide.
Working with researchers at the University of Warwick, and supported by industry partners RENA, a global expert in wet-chemical surface processing, and Oxford Instruments, a UK company designing, manufacturing, and supporting high-technology tools and systems for research and industry, NPL developed a quantitative, traceable method for measuring strain in silicon carbide using Raman spectroscopy. Raman spectroscopy is a technique that uses light to probe how a material behaves at the atomic level.
In this approach, a laser is focused onto the material surface, and subtle changes in the scattered light reveal how the material responds to applied stress. By carefully applying force to silicon carbide samples and measuring the resulting shifts in their Raman spectra, the team established a calibration relationship between Raman peak changes and known levels of strain. This calibration allows other laboratories and manufacturers to interpret their own Raman measurements in a reliable and consistent way, moving strain measurement from estimation to traceable measurement.
NPL’s role was not only technical. It also led the effort to bring together expertise from across the UK metrology, equipment, and instrumentation landscape required to make this work possible, spanning sample preparation, mechanical loading, and spectroscopy. The resulting data were translated into usable measurement protocols and reference guidance that can now be deployed across industry.
The project delivered a practical calibration method that links changes in Raman measurements directly to the amount of strain in a specific type of silicon carbide, the 4H polytype. This creates a reference curve that manufacturers and researchers can use to interpret their own measurements, allowing strain to be quantified rather than estimated. In practice, this means engineers can detect and measure strain in silicon carbide earlier in the device fabrication process, when problems are faster and less costly to correct.
The work also establishes an important foundation for future measurement standards in advanced semiconductor materials. While the current method focuses on one type of strain in silicon carbide, it demonstrates how traceable calibration approaches can be developed for new materials where measurement methods are still emerging. Industry partners and researchers have shown strong interest in the results, recognising that reliable strain measurement is an essential step towards improving device yield, performance, and reliability across the silicon carbide supply chain.
If successful, this work will have a significant impact across the silicon carbide supply chain. Reliable strain measurement enables problems to be identified earlier in the manufacturing process, before devices are fully fabricated or assembled, when changes are far more costly to make. This leads to higher wafer yields, more predictable device performance, and reduced qualification risk.
For manufacturers and system integrators, it also creates a common language for discussing material quality. When suppliers, manufacturers, and customers can trust that strain measurements mean the same thing, no matter which laboratory they are quantified in, decisions can be made faster and with greater confidence, reducing duplication and delay.
More broadly, the work strengthens the UK’s position in an area of growing strategic importance. While large-scale fabrication of power devices may take place elsewhere, the UK can shape how emerging materials are measured, qualified, and trusted. By developing reference data and measurement protocols early, NPL and its industry partners are helping to ensure that silicon carbide innovation moves more quickly from research into reliable, deployable technology, supporting both national industry and the wider global transition to more efficient power electronics.
This project has successfully delivered a new calibration for the rapid, quantitative assessment of silicon carbide power electronics wafers at all stages of the supply chain. We expect to see real uptake of this new capability by industry, and for it to underpin evidence-based improvements in yield and reliability.
Dr. Ben Green, University of Warwick