Quantum sensing compared to a standard failure analysis tool

Last month, exciting work using diamond-based quantum sensors done at Sandia National Laboratories was published in Physical Review Applied! The authors compared quantum diamond microscopy (QDM) with the industry-standard thermal-induced voltage alteration (TIVA) imaging technique for locating faults in the same die. The TIVA device tested was a commercial laser-scanning microscope (Checkpoint InfraScan 400 TDM)  using 140 mW of 1342 nm laser and a 20 minute measurement duration. The QDM apparatus also had a measurement time of 20 minutes. Their results concluded that magnetic imaging using quantum sensors in diamond (NV centers) can pinpoint the same faults. The magnetic image had the same grid patterns as layers 4 and 5, and disturbances indicate short circuits in the current traveling between them.

In the TIVA image, laser heating at defect locations changes the resistance, which shows in a circular pattern against a more or less homogeneous background. The signal-to-noise (SNR) for Fault B in the magnetic image was calculated to be 5.1, while for the TIVA image to be 0.15. This indicates that the SNR for the magnetic image is 35 times more. That means it is 1200 times faster in getting the same SNR! Furthermore, the SNR of TIVA measurements is strongly influenced by temperature-influencing factors such as heat absorbance from nearby metal but not QDM. 

L: a) Magnetic image with a 50 mA test current, three faults labeled A-C are circled.  b) TIVA image with 100 mA test current, showing faults in the same locations. R: a) Magnetic image shows weaker faults D-F b) D is barely visible, E and F are not visible in the TIVA image

In a different field of view, weaker faults were found. Faults E-F with magnetic signal 20-40x weaker than Fault D (which is already barely visible in the TIVA image) were detected by the QDM. Due to this incredible sensitivity, QDM is capable of measuring magnetic fields even from layers 50 - 100 µm deep. The authors acknowledge that however this would not be completely taking advantage of the low distance from the sensor to the source of the magnetic field which allows for the high resolution of the QDM and images would be blurred.

‍However the key point is that faults could still be visible deep within devices using backside power delivery for example and other devices with limited optical access to the deep layers! This is not possible for TIVA.

Citation:

Photo credit: Photothermal Spectroscopy Corp

For ease of reading, I have added the pre-print link here.

https://arxiv.org/abs/2302.01309

‍