METHOD FOR A BIN RATIO FORECAST AT NEW TAPE OUT STAGE

• US 20110010215A1
• Filed: 07/08/2009
• Published: 01/13/2011
• Est. Priority Date: 07/08/2009
• Status: Active Grant

• Current Assignee: Taiwan Semiconductor Manufacturing Company Limited
• Original Assignee: Taiwan Semiconductor Manufacturing Company Limited
• Inventors: Mou, Jong-I, Wu, Sunny, Lin, Chun-Hsien, Huang, Chia-Hung, Tsen, Andy, Chen, Jui-Long
• Granted Patent: US 8,082,055 B2
• Time in Patent Office: Days
• US Class Current: 705/8
• CPC Class Codes: G06Q 10/06 Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
• CPC Class Codes: G06Q 30/0202 Market predictions or forecasting for commercial activities

Abstract

A method for providing a bin ratio forecast at an early stage of integrated circuit device manufacturing processes is disclosed. The method comprises collecting historical data from one or more processed wafer lots; collect measurement data from one or more skew wafer lots; generating an estimated baseline distribution from the collected historical data and collected measurement data; generating an estimated performance distribution based on one or more specified parameters and the generated estimated baseline distribution; determining a bin ratio forecast by applying a bin definition and a yield degradation factor estimation to the generated estimated performance distribution; determining one or more production targets based on the bin ratio forecast; and processing one or more wafers based on the one or more determined production targets.

Claims

1. A method comprising:

collecting historical data from a plurality of processed wafers lots;

collecting measurement data from one or more skew wafer lots;

generating an estimated baseline distribution from the collected historical data and collected measurement data;

generating an estimated performance distribution from the estimated baseline distribution based on one or more specified parameters and;

determining a bin ratio forecast by applying a bin definition and a yield degradation factor estimation to the estimated performance distribution;

determining one or more production targets based on the bin ratio forecast;

providing fabrication resources according to the one or more production target; and

processing a plurality of wafers based on the one or more determined production targets.

2. The method of claim 1 wherein collecting historical data from the plurality of wafer lots comprises collecting at least one of wafer acceptance testing, chip product yield, or final testing data.

3. The method of claim 1 wherein collecting the measurement data from one or more skew lots comprises collecting at least one of wafer acceptance testing, chip product yield, or final testing data.

4. The method of claim 1 wherein generating an estimated baseline distribution from the historical and measurement data comprises:

generating a historical distribution by graphing the historical data as one or more historical data points;

generating a skew lot distribution by graphing the measurement data as one or more measurement data points; and

mapping each historical data point to a closest measurement data point, wherein the closest measurement data point is marked as an estimated baseline data point.

5. The method of claim 4 wherein mapping each historical data point to a closest measurement data point comprises performing a minimal distance search.

6. The method of claim 1 wherein the one or more production target comprises a wafer start time.

7. The method of claim 1 further comprising adjusting a fab process according to the bin ratio forecast.

8. The method of claim 1 wherein generating an estimated performance distribution based on one or more specified parameters comprises specifying I_DDQ and P-delay parameters, wherein an I_DDQ and P-delay distribution map is generated.

9. The method of claim 1 wherein determining the bin ratio forecast comprises determining a bin ratio forecast prior to a mass production stage.

10. The method of claim 9 wherein determining the bin ratio forecast prior to a mass production stage comprises determining the bin ratio forecast at a new tape out stage.

11. The method of claim 1 wherein determining the bin ratio forecast by applying the bin definition and the yield degradation factor estimation to the estimated performance distribution comprises:

specifying a bin definition for a customer order, wherein the bin definition provides more than one bin categories;

applying the bin definition to the estimated performance distribution;

determining a bin count for each bin category based on the bin definition;

determining a yield degradation factor estimation for each bin category; and

applying to the bin count the corresponding yield degradation factor estimation.

12. The method of claim 1 wherein determining one or more production targets based on the bin ratio forecast and processing the plurality of wafers based on the one or more determined production targets comprises determining a number of wafers to begin processing.

13. A method for providing a bin ratio forecast prior to a mass production stage of a plurality of integrated circuit devices, the method comprising:

generating a simulated wafer acceptance test (WAT) baseline distribution;

generating a simulated circuit performance distribution from the generated simulated WAT distribution;

portioning the generated simulated circuit performance distribution into at least two bin categories based on a bin definition;

determining a bin ratio forecast by applying a yield degradation factor to each bin category;

determining a wafer quantity and device targets to begin processing based on the determined bin ratio forecast; and

processing the determined wafer quantity according to the determined device targets.

14. The method of claim 13 further comprising adjusting a semiconductor manufacturing process according to the bin ratio forecast before processing the determined wafer quantity.

15. The method of claim 13 comprising providing fabrication resources according to the device targets.

16. The method of claim 13 wherein generating a simulated WAT baseline distribution comprises:

generating a historical lot distribution of historical data points extracted from WAT testing data from one or more processed wafer lots;

generating a skew lot distribution of skew lot data points extracted from WAT testing data from one or more skew lots; and

performing a minimal distance search for each historical data point to the skew lot distribution.

17. The method of claim 16 wherein performing the minimal distance search comprises utilizing a Mahalonobis minimal distance search.

18. The method of claim 16 wherein applying the yield degradation factor to each bin category comprises obtaining a yield degradation factor for each bin category based on the simulated circuit performance distribution and the skew lot distribution.

19. The method of claim 13 wherein determining a bin ratio forecast by applying a yield degradation factor to each bin category comprises determining a number of IC devices associated with each bin category based on the simulated circuit performance distribution.

20. A system for providing a bin ratio forecast prior to a mass production stage in a semiconductor manufacturing environment, wherein a plurality of processes are performed on a plurality of wafer lots, the system comprising:

a virtual fabrication system coupled to a network;

a manufacturing execution system (MES) coupled to the network, wherein the MES comprises a bin-based control module configured to;

collect historical data from one or more processed wafer lots;

collect measurement data from one or more skew wafer lots;

generate an estimated baseline distribution based on the collected historical data and collected measurement data;

generate an estimated performance distribution based on one or more specified parameters and the generated estimated baseline distribution;

determine a bin ratio forecast by applying a bin definition and a yield degradation factor estimation to the generated estimated performance distribution;

determine one or more production targets based on the bin ratio forecast; and

process one or more wafers based on the one or more determined production targets.