X-ray technique for detecting defective computer chips
Computer chips consist of billions of interconnected transistors, making it challenging to manufacture them without defects. But do you determine if a chip is compromised?
Researchers from the Paul Scherer Institut (PSI) and the University of Southern California (USC) have now made it possible to validate the integrity of computer chips using X-rays, allowing companies and other organisations to non-destructively scan computer chips to ensure they haven’t been altered and have been manufactured to design specifications without error. Published in the journal Nature Electronics, the method should help improve hardware security, which is particularly crucial in the supply chain for advanced electronics.
Called ptychographic X-ray laminography, the technique utilises X-rays from a synchrotron to illuminate a small region of a rotating chip at an angle of 61° (with respect to the normal of the chip plane). The resulting diffraction patterns are measured with a photon-counting detector array. The data are then used to generate high-resolution slice images of the chip, from which 3D renderings are created.
Once the 3D image is generated, it can be compared with the original design as a type of forensics to help companies or organisations that are looking to ensure chips are manufactured correctly and meet design specifications.
The researchers indicate that chips have signature features, so it is possible to tell how and where they were manufactured. In addition, the process allows for reverse engineering of circuit designs without destroying the chip.
“The majority of a chip’s intelligence is how it is wired,” said study co-author Anthony FJ Levi, Department Chair of the Ming Hsieh Department of Electrical Engineering-Electrophysics at USC. “It is like the connectome of a brain.
“By viewing a chip in detail, you can non-destructively figure out what it does. With this technology, hiding intellectual property in a chip is over.”
With the researchers set to continue improving imaging speed, resolution and X-ray microscope performance, Levi imagines that the technology could one day contribute to a certification process to ensure the integrity of chips that are inserted into a computer or in communication hardware used by global businesses and governments.
As the use of automated face recognition for personal identification continues to grow, so too...
A high-throughput computational method developed at UC San Diego has generated 13 new material...
Seeking a solution for comfortable wearable electronics, Chinese researchers have developed a way...