Detecting defects in a semiconductor specimen

A method of detecting defects in run-time during fabrication of a specimen, the method comprising:

• obtaining, by a processing and memory circuitry (PMC), a runtime image of at least a part of a specimen, the runtime image acquired by an examination tool, and the runtime image comprising a plurality of pixels;
• receiving, by the PMC, at least one characteristic of a defect of interest and a location of interest associated therewith, wherein the defect of interest is indicative of a specimen abnormality formed on or within the specimen at the location of interest;
• modifying, by the PMC in accordance with the at least one characteristic, image data informative of at least one pixel from the plurality of pixels of the runtime image and corresponding to the location of interest, thereby planting the defect of interest in the image data at the location of interest and obtaining modified image data;
• processing, by the PMC, the modified image data to detect a set of locations of potential defects of the specimen in accordance with a detection recipe, wherein the detection recipe is generated prior to processing the modified image data;
• determining, by the PMC, whether the set of locations of potential defects includes the location of interest;
• and
• subject to determining that the location of interest is not included in the set of locations of potential defects, modifying, by the PMC, the detection recipe to enable detection of the defect of interest planted at the location of interest.

A non-transitory computer readable storage medium comprising instructions, which when executed by a processor in run-time during fabrication of a specimen, cause the processor to perform operations comprising:

• obtaining a runtime image of at least a part of a specimen, the runtime image acquired by an examination tool, and the runtime image comprising a plurality of pixels;
• receiving at least one characteristic of a defect of interest and a location of interest associated therewith, wherein the defect of interest is indicative of a specimen abnormality formed on or within the specimen at the location of interest;
• modifying, in accordance with the at least one characteristic, image data informative of at least one pixel from the plurality of pixels of the runtime image and corresponding to the location of interest, thereby planting the defect of interest in the image data at the location of interest and obtaining modified image data;
• processing the modified image data to detect a set of locations of potential defects of the specimen in accordance with a detection recipe, wherein the detection recipe is generated prior to processing the modified image data;
• determining whether the set of locations of potential defects includes the location of interest;
• and
• subject to determining that the location of interest is not included in the set of locations of potential defects, modifying the detection recipe to enable detection of the defect of interest planted at the location of interest.