Researchers from Vrije Universiteit Amsterdam and research institute ARCNL have developed a new method to measure extremely small structures on chips faster and more accurately. The technique, published in the scientific journal Nature Communications, can contribute to the further development of more powerful and energy-efficient computer chips.
The production of modern chips requires measurement methods that can capture details just a few nanometers in size—thousands of times smaller than a human hair. As chip components become increasingly smaller, it also becomes more difficult to detect errors or deviations in time during production.
Broadband illumination
To this end, the researchers developed a new form of extreme ultraviolet scatterometry based on high harmonic generation. This involves using extreme ultraviolet light to analyze nanostructures without damaging the material. Unlike existing techniques, the new method works with broadband illumination and uses spectral information of the so-called zero-order diffraction. As a result, the quality of chip structures can be determined very accurately, without complex calibrations. Experiments show that the method can reconstruct the dimensions of nanostructures with an accuracy of a few nanometers. Critical dimensions of chip components can even be measured with an accuracy of up to approximately 10 nanometers. According to the researchers, this offers a fast, reliable, and non-destructive way to control production processes in the semiconductor industry.
Societal impact
Improved measurement methods make it possible to further miniaturize chips and produce them more efficiently. This is essential for applications such as artificial intelligence, medical technology, communication systems, and sustainable electronics. Moreover, more accurate process control can lead to less material waste and lower production costs in the chip industry.
The research was conducted by ARCNL, with a contribution from VU-group leader dr. Peter Kraus.
Article at ARCNL
Nature Communications