Non-transitory computer-readable medium storing program code generating wafer map based on generative adversarial networks and computing device including the same

Patent metadata

Field	Value
Publication number	US-11775840-B2
Application number	US-202016909132-A
Country	US
Kind code	B2
Filing date	Jun 23, 2020
Priority date	Nov 26, 2019
Publication date	Oct 3, 2023
Grant date	Oct 3, 2023

Abstract

A non-transitory computer-readable medium storing a program code including an image generation model, which when executed, causes a processor to input input data including sampling data of some of a plurality of semiconductor dies of a wafer to a generator network of the image generation model and output a wafer map indicating the plurality of semiconductor dies, and to input the wafer map output from the generator network to a discriminator network of the image generation model and discriminate the wafer map.

First claim

What is claimed is:

1. A non-transitory computer-readable medium storing a program code including an image generation model, which when executed, causes a processor to:
train a generator network of the image generation model to generate a wafer image by inputting input data including sampling data of some of a plurality of semiconductor dies of a wafer to the generator network of the image generation model;
output a wafer map indicating the plurality of semiconductor dies generated during the training of the generator network;
input the wafer map output from the generator network to a discriminator network of the image generation model such that the discriminator network discriminates the wafer map generated during the training; and
perform additional training of the generator network based on a discrimination result of the discriminator network correctly differentiating the generated wafer map from a real wafer map.

2. The non-transitory computer-readable medium of claim 1 , wherein the image generation model, which when executed, causes the processor to:
arrange the sampling data on the input data depending on locations of the some semiconductor dies of the wafer.

3. The non-transitory computer-readable medium of claim 1 , wherein the image generation model, which when executed, causes the processor to:
generate the input data including the sampling data and at least one of a first value determined in advance and a second value based on the sampling data.

4. The non-transitory computer-readable medium of claim 3 , wherein the image generation model, which when executed, causes the processor to:
dispose a sampling value of the sampling data on the input data depending on a location of a first semiconductor die of the some semiconductor dies; and
dispose the at least one of the first value and the second value on the input data depending on a location of a second semiconductor die of the remaining semiconductor dies of the plurality of semiconductor dies.

5. The non-transitory computer-readable medium of claim 1 , wherein the image generation model, which when executed, causes the processor to:
reshape a size of the input data to a size of the wafer map.

6. The non-transitory computer-readable medium of claim 1 , wherein the image generation model, which when executed, causes the processor to:
reshape a size of the input data to a one-dimensional vector.

7. The non-transitory computer-readable medium of claim 1 , wherein the image generation model, which when executed, causes the processor to:
generate the input data including the sampling data and coordinate values indicating locations of the some semiconductor dies of the wafer.

8. The non-transitory computer-readable medium of claim 7 , wherein the coordinate values are based on one of a Cartesian coordinate system and a polar coordinate system.

9. The non-transitory computer-readable medium of claim 1 , wherein the plurality of semiconductor dies are a plurality of first semiconductor dies and the wafer is a first wafer, and
wherein the image generation model, which when executed, causes the processor to:
execute the generator network learned based on the discrimination result and generate a wafer map of a second wafer including a plurality of second semiconductor dies.

10. The non-transitory computer-readable medium of claim 1 , wherein the image generation model, which when executed, causes the processor to:
generate the input data including the sampling data and a process parameter of the wafer.

11. A non-transitory computer-readable medium storing a program code including an image generation model, which when executed, causes a processor to:
perform learning of a discriminator network of the image generation model so that the discriminator network discriminates a first wafer map including a plurality of sampling values measured from a plurality of semiconductor dies of a wafer as real and discriminates a second wafer map output from a generator network of the image generation model, to which some of the plurality of sampling values are input, as fake; and
perform learning of the generator network so that the second wafer map output from the generator network is discriminated as real by the discriminator network.

12. The non-transitory computer-readable medium of claim 11 , wherein the image generation model, which when executed, causes the processor to:
input input data including the some sampling values and coordinate values of some semiconductor dies, from which the some sampling values are measured, from among the plurality of semiconductor dies to the generator network.

13. The non-transitory computer-readable medium of claim 11 , wherein the plurality of semiconductor dies are a plurality of first semiconductor dies, the wafer is a first wafer, and the plurality of sampling values are a plurality of first sampling values,
wherein the image generation model, which when executed, causes the processor to:
perform learning of the discriminator network so that the discriminator network discriminates a third wafer map including a plurality of second sampling values respectively measured from a plurality of second semiconductor dies as real and discriminates a fourth wafer map output from the generator network, to which some of the plurality of second sampling values are input, as fake; and
perform learning of the generator network so that the fourth wafer map output from the generator network is discriminated as real by the discriminator network.

14. The non-transitory computer-readable medium of claim 11 , wherein the image generation model, which when executed, causes the processor to:
input input data including the some sampling values and a process parameter of the wafer to the generator network.

15. The non-transitory computer-readable medium of claim 11 , wherein the generator network outputs the second wafer map from input data including the some sampling values without up-sampling.

16. A computing device comprising a processor configured to execute an image generation model stored in a memory, wherein the processor is configured to:
perform learning of a discriminator network of the image generation model so that the discriminator network discriminates a first wafer map of a first wafer including a plurality of first sampling values measured from a plurality of first semiconductor dies of the first wafer as real and discriminates a second wafer map of the first wafer output from a generator network of the image generation model, to which some of the plurality of first sampling values are input, as fake;
perform learning of the generator network so that the second wafer map output from the generator network is discriminated as real by the discriminator network; and
input some of second sampling values measured from some of a plurality of second semiconductor dies of a second wafer to the generator network and generate a third wafer map of the second wafer.

17. The computing device of claim 16 , wherein locations of the some second sampling values on the third wafer map respectively correspond to locations of the some first sampling values on the second wafer map.

18. The computing device of claim 16 , wherein the processor is configured to:
input input data including the some first sampling values and coordinate values of some semiconductor dies, from which the some first sampling values are measured, from among the plurality of first semiconductor dies to the generator network.

19. The computing device of claim 16 , wherein the process

Assignees

• Samsung Electronics Co Ltd

Inventors

• Jang Wonik
• Myung Sanghoon
• Jeong Changwook
• Lee Sunghee

Classifications

• G06T11/10

  Texturing; Colouring; Generation of textures or colours (retouching, inpainting or scratch removal G06T5/77) · CPC title

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/094

  Adversarial learning · CPC title

• G06N3/0475

  Generative networks · CPC title

• H10P74/203

  Structural properties, e.g. testing or measuring thicknesses, line widths, warpage, bond strengths or physical defects · CPC title