Quantum sensors in semiconductor manufacturing
Enabling Moore's law to continue reduction in chip dimensions and improving processing speed and reduce energy consumption.
Inefficient quality control process exacerbating chip shortages
Semiconductor manufacturing is high-end - being both capital-heavy and technologically intensive. They are essential in devices needing computing power, having high demand but with very limited foundries producing them. A human, using optical scanning (optical beam induced resistance change) or imaging technology (lock-in thermography), reviews each circuit or the wafer as a whole to check for issues. Both techniques are slow and depend on the chip being infrared transparent. The flawed product can then be removed, and then the stamps or design can be reviewed for faults. Manufacturers are having a difficult time meeting quality assurance standards. This results in a costly and inefficient manufacturing process with up to 20% scrapping rate costing a single plant 87.5 to 140 million dollars a year in losses. Moreover, these losses exacerbate supply shortages.
Quantum diamond microscope
Current through an integrated circuit (IC) or a print circuit board generates a small vector magnetic field. For diamond based quantum sensing, there are two state-of-the-art methods which are relevant to the semiconductor industry. One is the scanning tip method, where a diamond tip is moved over the sample. This enables a very high spatial resolution (25 - 50 nm) in several hours. The other is wide-field imaging which uses a camera, simultaneously measuring and spatially correlating the signal recorded by each pixel in seconds.
Magnetic field sensing
Minimal field strength: 0.1 nT (AC), 1 nT (DC)
Frequency range: 1 Hz - 10 GHz (AC), <1Hz (DC)
Sensitivity: down to 10 pT/√Hz
Vectorial sensing possible
Magnetic field imaging
X/Y-resolution: 250 nm currently, could be down to 100 nm
Field of View: up to 16 mm²
The trillion dollar industry of the future
The semiconductor industry is expected to be a trillion dollar industry by 2029. ASML after spending 30 years in developing the chips that are used today is already making the ones for 30 years later - going down to nanometer and even smaller sizes. As they become more and more miniaturized, identifying defects in chips such as current leakages early in the manufacturing process becomes increasingly challenging and increasingly require the quantum sensing solution.