Meeting new inspection demands in compound semiconductors
Compound semiconductors are the backbone of high-power electronics, high-frequency electronics, and opto-electronics. As global demand for compound semiconductors grows, the notoriously low yield rates of materials such as silicon carbide (SiC) gallium nitride (GaN) become impossible to ignore. By developing better optical inspection tools, we seek to help manufacturers identify yield-killing defects in epitaxy layers long before the material is used in device production.
Over the last few years MONSTR Sense Technologies has won an NSF Phase II SBIR and poured substantial resources into applying ultrafast microscopy for defect inspection in compound semiconductors. Through this R&D we have demonstrated applications for SiC, GaN, and gallium arsenide (GaAs). You can use or try out ultrafast imaging for any of these applications using our Compound Semiconductor Inspection Service or by purchasing a KRAKEN.
When nonlinear optical measurements are better
A typical first step in bare wafer semiconductor inspection is dark-field imaging by scanning a tiny bright spot over the surface of a spinning wafer. This method is great for detecting every tiny surface defect that scatters light. For inspecting pristine silicon wafers, this inspection technique is great! However, the method does not work so well for epitaxy layers of compound semiconductors for two reasons:
- Scattering sites do not necessarily correlate with killer defect sites. Nuisance events, where an anomaly is detected that is not a defect, are common in these epi layers. -> Nonlinear imaging provides better identifiable information about defects. In fact, by exploiting the long decay time of defects in SiC, our nonlinear imaging modality measures these defects background free.
- Subsurface defects that impact device performance can go unnoticed. The insensitivity of conventional inspection methods to killer defects that are under the surface leads to reduced yield that is not realized until electrical test measurements of devices are performed. -> With nonlinear imaging we can penetrate through full wafers, and we are only sensitive to material at the focus of our microscope. Nonlinear measurements are therefore capable of selectively measuring deep semiconductor layers and volumetric imaging.
Nonlinear measurements, particularly four-wave mixing, are sensitive to fundamental defects in the crystalline structure of the semiconductor that impact the carriers. This means that defects, even subsurface defects, will strongly impact the four-wave-mixing image if the defect has a strong impact on the electrical properties of the semiconductor.
In the figures below we show how four-wave-mixing images are not only able to measure defects and inhomogeneity with high sensitivity, but we are able to record 3D volumetric images of those feature inside compound semiconducting materials and devices.
Automated defect detection incorporating Machine Learning
We use a combination of analytical computer vision and machine learning to find and classify defects across SiC. MONSTR Sense has developed and continues to advance data fusion algorithms for improving detection performance using multiple optical inspection channels such as nonlinear (NL), resonant linear reflectance (RLR), and photoluminescence (PL) imaging. After analytically finding defects using thresholding, laplacian contrast, and contouring, we use a Faster R-CNN model (Detectron 2) to classify defects.
Measuring electrical properties without probes
Ultrafast spectroscopy techniques measure charge transfer and energy transfer in materials. Ultrafast imaging microscopy measurements therefore identify regions where electrical energy is lost. For finding defects in semiconductors that impede device performance, this is everything. Regions of a semiconductor with uncharacteristically long optical decay constants will have correspondingly high nonradiative decay. For LEDs these regions will have reduced efficiency, and for detectors and solar cells these regions will have weak absorption.
Emission and absorption efficiency is typically determined with electrical measurements, requiring construction of an entire device around the semiconductor for testing. Our measurements can perform these tests 100% non-contact saving manufacturers time and process steps.
Testers, advisors, feature recommendations, and more needed
If you have an interest or are an expert in semiconductor manufacturing, please reach out to us. We are actively working to improve defect inspection in compound semiconductors, and we can use your help. Let’s start with a call!
Related links
Product page:
Compound Semiconductor Inspection Service (CSIS)
Characterize your full compound semiconductor wafers or test out the efficacy for your needs of our inspection equipment with our analytical testing service.
Product page:
KRAKEN Turnkey Ultrafast Microscope
Automated microscope acquires rapid hyperspectral, hypertemporal, and volumetric images. The KRAKEN is turnkey – no hassle dealing with your laser or aligning it into your setup on a regular basis.
Product page:
NESSIE Laser Scanning Microscope
First-of-its-kind laser-scanning microscope that integrates laser-based measurements with high resolution imaging. NESSIE features a large sample area for cryostats and is designed to work with our BIGFOOT spectrometer.