Process prediction system and process prediction method

Abstract

A process prediction system includes memory and a processor. The memory stores a plurality of instructions and an artificial intelligence algorithm. The processor is coupled to the memory. The processor executes the following steps according to the plurality of instructions in the memory: setting a test terminal for each of a plurality of dies in a wafer; measuring a plurality of process data of the plurality of dies using the test terminals of each of the plurality of dies; training and updating the artificial intelligence algorithm based on the plurality of process data of the plurality of dies; determining a measurement program data corresponding to one of the plurality of dies based on the artificial intelligence algorithm; and outputting measurement result data when the measurement program data corresponding to one of the plurality of dies is determined by the artificial intelligence algorithm.

Claims

1. A process prediction system, comprising: a memory for storing a plurality of instructions and an artificial intelligence algorithm; a processor coupled to the memory and performing the following steps according to the plurality of instructions in the memory: setting a test terminal for each of a plurality of dies in a wafer; measuring a plurality of process data of the plurality of dies through the test terminal of each of the plurality of dies; training and updating the artificial intelligence algorithm based on the plurality of process data of the plurality of dies; determining a measurement program data corresponding to one of the plurality of dies according to the artificial intelligence algorithm; when the measurement program data corresponding to the one of the plurality of dies is determined by the artificial intelligence algorithm, outputting measurement result data; wherein the processor outputs an optimal wafer probe test program based on the artificial intelligence algorithm and the plurality of process data.

2. The process prediction system as described in claim 1, wherein the processor acquires the plurality of process data for the plurality of dies at a plurality of process stages; wherein the plurality of process stages include a wafer acceptance testing stage, a wafer probe measurement stage, a package measurement stage, a final test measurement stage, and a reliability measurement stage; wherein the plurality of process data is related to the wafer acceptance testing stage.

3. The process prediction system as described in claim 2, wherein the processor outputs a warning signal based on the artificial intelligence algorithm and the plurality of process data.

4. The process prediction system as described in claim 3, wherein the optimal test program for the wafer probe is related to a plurality of wafer probe measurement data during the wafer probe measurement phase.

5. The process prediction system as described in claim 4, wherein the processor obtains a first mapping relationship based on the plurality of process data and the plurality of wafer probe measurement data; wherein the processor updates the artificial intelligence algorithm based on the first mapping relationship data.

6. The process prediction system as described in claim 5, wherein the processor outputs a final test optimal test program based on the artificial intelligence algorithm and the plurality of process data; wherein the final test optimal test program is related to the plurality of final test measurement data of the final test measurement stage.

7. The process prediction system as described in claim 6, wherein the processor obtains a second mapping relationship based on the plurality of process data and the plurality of final test measurement data; wherein the processor updates the artificial intelligence algorithm based on the second mapping relationship data.

8. The process prediction system as claimed in claim 7, wherein the processor obtains a third mapping relationship based on the plurality of process data and the plurality of reliability measurement data of the reliability measurement stage; wherein the processor updates the artificial intelligence algorithm based on the third mapping relationship data; wherein the first mapping relationship data, the second mapping relationship data and the third mapping relationship data are different from each other.

9. The process prediction system as claimed in claim 8, wherein the processor outputs at least one of a wafer probe measurement prediction model, a final test measurement prediction model, a reliability measurement prediction model, and an accurate circuit simulation model based on the artificial intelligence algorithm, the first mapping data, the second mapping data, and the third mapping data.

10. A process prediction method, comprising: in a design phase, setting a test terminal for each of a plurality of dies in a wafer; in a data collection phase, measuring a plurality of process data of the plurality of dies using the test terminal of each of the plurality of dies; training and updating an artificial intelligence algorithm based on the plurality of process data of the plurality of dies; determining a measurement program data corresponding to one of the plurality of dies based on the artificial intelligence algorithm; when the measurement program data corresponding to the one of the plurality of dies is determined by the artificial intelligence algorithm, outputting measurement result data; outputting at least one of a wafer probe measurement prediction model, a final test measurement prediction model, a reliability measurement prediction model, and an accurate circuit simulation model based on the artificial intelligence algorithm and the plurality of process data; wherein the measurement result data is related to one of the wafer probe measurement prediction model, the final test measurement prediction model, the reliability measurement prediction model, and the accurate circuit simulation model.

Images