The semiconductor industry is a multibillion-dollar endeavor situated at the heart of the modern world. During the development and production of semiconductor devices things can go wrong, and the source of the problems must be analyzed. This is the impetus for another large industry aimed at providing methods for failure analysis (FA) of semiconductor devices. FA systems make use of a variety of sophisticated techniques that provide a wealth of information at very high spatial resolutions. However, current techniques have limitations, especially regarding the non-destructive measurement of the internal structure and material composition of the devices.
Electron spin resonance (ESR) is a well-established technique for the characterization and imaging of paramagnetic materials. In semiconductors, ESR has been used extensively to study paramagnetic properties such as crystal impurities, defects, and dopants that greatly affect the devices’ performance. However, conventional ESR has very limited sensitivity and image resolution, making it unattractive for the study of the sub-micron-scale devices that are in use today. Our current ERC project has recently resulted in a method that improves ESR sensitivity by about 3 orders of magnitude compared to commercial systems. We also enhanced the spatial resolution of ESR-based images to the 100-nm length scale. Such capabilities heighten ESR’s attractiveness as a potential semiconductor-inspection tool. The purpose of the proposed project is to develop a demonstration prototype of a one-sided ESR imaging probe based upon this new methodology. This probe will examine semiconductor wafers and analyze their properties, composition, and structure in a non-destructive manner. We will also provide market research on the field of FA and examine how our new method fits into it. Special attention will be paid to strengthening IP protection for our idea and to finding means to attract more funds to commercialize this technology.
Fields of science
Call for proposal
See other projects for this call