Mask inspection of a semiconductor specimen
There is provided a mask inspection system and a method of mask inspection. The method comprises: detecting, by the inspection tool, a runtime defect at a defect location on a mask of a semiconductor specimen during runtime scan of the mask, and acquiring, by the inspection tool after runtime and based on the defect location, a plurality sets of aerial images of the runtime defect corresponding to a plurality of focus states throughout a focus process window, each set of aerial images acquired at a respective focus state. The method further comprises for each set of aerial images, calculating a statistic-based EPD value of the runtime defect, thereby giving rise to a plurality of statistic-based EPD values each corresponding to a respective focus state, and determining whether the runtime defect is a true defect based on the plurality of statistic-based EPD values.
1. A mask inspection system, the system comprising:
an inspection tool configured to:
detect a runtime defect at a defect location on a mask of a semiconductor specimen during runtime scan of the mask; and
acquire, based on the defect location after runtime, a plurality of sets of aerial images of the runtime defect corresponding to a plurality of focus states throughout a focus process window, each set of aerial images acquired at a respective focus state; and
a processing and memory circuitry (PMC) operatively connected to the inspection tool and configured to:
for each set of aerial images, calculate a statistic-based Edge Positioning Displacement (EPD) value of the runtime defect, thereby giving rise to a plurality of statistic-based EPD values each corresponding to a respective focus state; and
determine whether the runtime defect is a true defect based on the plurality of statistic-based EPD values.
2. The mask inspection system according to claim 1 , wherein the PMC is configured to determine whether the runtime defect is a true defect by filtering out the runtime defect as a false defect when all values in the plurality of statistic-based EPD values are below a predefined EPD threshold.
3. The mask inspection system according to claim 1 , wherein the PMC is configured to, for each set of aerial images, calculate the statistic-based EPD value of the runtime defect using a Print Threshold (PT) characterizing the mask, wherein the PT is applied to each aerial image in the set of aerial images to calculate a respective EPD value of the runtime defect therein, and the statistic-based EPD value is an averaged EPD value obtained by averaging respective EPD values calculated for each aerial image in the set of aerial images.
4. The mask inspection system according to claim 3 , wherein the PMC is further configured to calibrate the PT prior to using the PT, comprising:
using a value of a wafer resist threshold to convert an aerial image of a representative pattern with width and length corresponding to a design intent in corresponding design data into a binary image;
calculating a grey level (GL) threshold for all pixels along the length of the representative pattern; and
calculating the calibrated PT as an averaged GL threshold.
5. The mask inspection system according to claim 1 , wherein each set of aerial images acquired at a respective focus state are overlapped aerial images captured for the runtime defect.
6. The mask inspection system according to claim 1 , wherein the inspection tool is an aerial mask inspection tool configured to capture aerial images, or a non-aerial mask inspection tool configured to capture non-aerial images where the aerial images are simulated from the non-aerial images.
7. The mask inspection system according to claim 1 , wherein the plurality of focus states throughout a focus process window comprises a negative defocus, a best focus and a positive defocus.
8. The mask inspection system according to claim 1 , wherein the runtime defect is from a plurality of runtime defects detected in a region of interest (ROI), and the acquiring, calculating and determining are performed for the plurality of runtime defects.
9. The mask inspection system according to claim 1 , wherein the inspection tool is further configured to acquire a plurality of sets of reference images corresponding to the plurality of focus states, the plurality of sets of reference images acquired from one or more corresponding reference locations of the defect location in one or more reference dies.
10. The mask inspection system according to claim 9 , wherein the PMC is further configured to, for each focus state, process the set of reference images acquired at the focus state to obtain a multi-reference; wherein the PMC is configured to use the multi-reference to calculate an EPD value for each image of the plurality of sets of aerial images at the focus state to yield a plurality of EPD values, and calculate the statistic-based EPD value as a statistical derivative of the plurality of EPD values.
11. A mask inspection method by a mask inspection system comprising an inspection tool and a processing and memory circuitry (PMC), the method comprising:
detecting, by the inspection tool, a runtime defect at a defect location on a mask of a semiconductor specimen during runtime scan of the mask;
acquiring, by the inspection tool after runtime and based on the defect location, a plurality of sets of aerial images of the runtime defect corresponding to a plurality of focus states throughout a focus process window, each set of aerial images acquired at a respective focus state;
for each set of aerial images, calculating, by the PMC, a statistic-based EPD value of the runtime defect, thereby giving rise to a plurality of statistic-based EPD values each corresponding to a respective focus state; and
determining, by the PMC, whether the runtime defect is a true defect based on the plurality of statistic-based EPD values.
12. The mask inspection method according to claim 11 , wherein said determining comprises filtering out the runtime defect as a false defect when all values in the plurality of statistic-based EPD values are below a predefined EPD threshold.
13. The mask inspection method according to claim 11 , wherein for each set of aerial images, the statistic-based EPD value of the runtime defect is calculated using a Print Threshold (PT) characterizing the mask, wherein the PT is applied to each aerial image in the set of aerial images to calculate a respective EPD value of the runtime defect therein, and the statistic-based EPD value is an averaged EPD value obtained by averaging respective EPD values calculated for each aerial image in the set of aerial images.
14. The mask inspection method according to claim 13 , further comprising calibrating the PT prior to using the PT, the calibrating comprising:
using a value of a wafer resist threshold to convert an aerial image of a representative pattern with width and length corresponding to a design intent in corresponding design data into a binary image;
calculating a grey level (GL) threshold for all pixels along the length of the representative pattern; and
calculating the calibrated PT as an averaged GL threshold.
15. The mask inspection method according to claim 11 , wherein each set of aerial images acquired at a respective focus state are overlapped aerial images captured for the runtime defect.
16. The mask inspection method according to claim 11 , wherein the plurality of focus states throughout a focus process window comprises a negative defocus, a best focus and a positive defocus.
17. The mask inspection method according to claim 11 , wherein the runtime defect is from a plurality of runtime defects detected in a region of interest (ROI), and the acquiring, calculating and determining are performed for the plurality of runtime defects.
18. The mask inspection method according to claim 11 , further comprising acquiring, by the inspection tool, a plurality of sets of reference images corresponding to the plurality of focus states, the plurality of sets of reference images acquired from one or more corresponding reference locations of the defect location in one or more reference dies.
19. The mask inspection method according to claim 18 , further comprising, for each focus state, processing the set of reference images acquired at the focus state to obtain a multi-reference, wherein said calculating comprises using the multi-reference to calculate an EPD value for each image of the plurality of sets of aerial images at the focus state to yield a plurality of EPD values, and calculating the statistic-based EPD value as a statistical derivative of the plurality of EPD values.
20. A non-transitory computer readable storage medium tangibly embodying a program of instructions that, when executed by a computer, cause the computer to perform a method of mask inspection, the method comprising:
detecting a runtime defect at a defect location on a mask of a semiconductor specimen during runtime scan of the mask;
acquiring, after runtime and based on the defect location, a plurality of sets of aerial images of the runtime defect corresponding to a plurality of focus states throughout a focus process window, each set of aerial images acquired at a respective focus state;
for each set of aerial images, calculating a statistic-based EPD value of the runtime defect, thereby giving rise to a plurality of statistic-based EPD values each corresponding to a respective focus state; and
determining whether the runtime defect is a true defect based on the plurality of statistic-based EPD values.