BS IEC 63068-3:2020 pdf free.Semiconductor devices — Non-destructive recognition criteria of defects in silicon carbide homoepitaxial wafer for power devices Test method for defects using photoluminescence.
A combination of light source and focusing optics should be optimized to achieve sufficientuniformity of the excitation light intensity on the wafer surface.The PL intensity at each pointon the epitaxial layer is adjusted in an appropriate range so that defects are clearly detected.Uniformity of excitation light intensity can be achieved using hardware and/or software.
The spectral and power distributions of the excitation light are maintained constant during thewhole measurement period.
Wafers shall be positioned in the plane of Cartesian coordinate system (X-Y) or cylindricalcoordinate system (R-6).The third axis (Z) is the optical axis of image capturing system.The Z-axis is perpendicular to the plane and its point of intersection with the plane shall bethe point of focus.The distance between the front-end portion of image-capturing optics andthe wafer surface shall be constant,independent of the thickness of the wafers, so that focusing and magnification are not mutually adversely affected.
The PL imaging system is typically composed of a light source,focusing optics, CCD image sensor as an optical digital sensor，lighting-geometry adjustment system，wafer stage and light-tight enclosure. A dark box or a rack housing is often used to prevent the interference by external illumination.The spatial resolution of the PL imaging system shall be high enough to capture distinct features of small size defects. The image information is digitized directly within the optical image sensor unit.
To ensure the repeatability and reproducibility of the image capturing procedure， parameter settings should be carried out at a regular interval.This can be performed using specified reference wafers, for example,silicon or silicon carbide wafers.
The image processing covers numerous features such as brightness,contrast, edge detection,shading correction,and inversion.
Different software solutions may employ different mathematical algorithms for similaroperations，and images processed by different image-processing algorithms will not beidentical. Parameter settings，e.g. using reference wafers， are performed to ensure thatresults are comparable.
Two different methods are used for image analysis: binary (black/white) analysis and grey-level analysis.To obtain a binary image from a grey-level image, threshold procedure is used.
An appropriate algorithm should be used for image analysis to detect successfully defects in test wafers.BS IEC 63068-3 pdf download.