Back to results
US 10,948,903 B2Granted
System and method that consider tool interaction effects for identifying root causes of yield loss
AssigneeNational Cheng Kung University
InventorsChin-Yi Lin, Yu-Ming Hsieh, Fan-Tien Cheng
Life of this patent
Estimating the term…
Abstract
Embodiments of the present disclosure provide a two-phase process for searching root causes of a yield loss in a production line. In a first phase, an interaction between two process tools, that between two parameters, or that between one process tool and one parameter that is likely to cause the yield loss is identified. In a second phase, a threshold of the parameter that is likely to cause the yield loss and is obtained from the first phase is identified. In each phase, two different algorithms can be used to generate a reliance index (RII) for gauging the reliance levels of their search results.
Claims15
A method for identifying root causes of a yield loss, the method comprising:
- providing a production line comprising a plurality of process stages, each of the process stages comprising a plurality of process tools, each of the process tools comprising at least one process device, each of the at least one process device having a plurality of process parameters configured for processing one of a plurality of workpieces;
- identifying a plurality of inter-tool parameters between the process tools, wherein each of the inter-tool parameters represents an interaction between two of the process tools;
- processing the workpieces according to a plurality of production routes respectively, each of the production routes indicating a predetermined device of the process devices of the process tools in the process stages for processing one of the workpieces;
- performing metrology on each of the workpieces after being processed by its corresponding predetermined device, thereby obtaining a plurality of sets of in-line metrology values corresponding to the workpieces;
- performing at least one yield test on each of the workpieces after passing through the production line, thereby obtaining a plurality of sets of final inspection values corresponding to the workpieces respectively;
- determining if a yield test failure is encountered according to the sets of final inspection values;
- performing a first phase when the yield test failure is encountered, the first phase comprising:
- preparing a first searching model based on a first algorithm, wherein the first algorithm is a triple phase orthogonal greedy algorithm (TPOGA), a least absolute shrinkage and selection operator (LASSO) algorithm or a regression tree algorithm;
- obtaining a set of interaction data indicating whether the workpieces pass through or are produced by using both of any two elements of a first assembly consisting of the process devices, the process parameters and the inter-tool parameters, according to the production routes, the process parameters and the inter-tool parameters;
- and
- identifying at least one first key process device, and/or at least one first key process parameter, and/or at least one first key inter-tool parameter, and/or at least one first key interaction that are likely to cause the yield test failure by feeding the sets of final inspection values of the workpieces, the sets of in-line metrology values of the workpieces, the production routes of the workpieces, the process parameters, the inter-tool parameters, the set of interaction data into the first searching model;
- and
- when the result of identifying is good, sending at least one updated configuration to the at least one first key process device or devices related to the at least one first key process parameter, the at least one first key inter-tool parameter or the at least one first key interaction for continuous improvements of yield of the production line.
10 more elements
A method for identifying root causes of a yield loss occurring in a production line comprising a plurality of process stages, each of the process stages comprising at least one process tool, each of the at least one process tool comprising at least one process device, each of the at least one process device having a plurality of process parameters configured for processing one of a plurality of workpieces, the method comprising:
- identifying a plurality of inter-tool parameters between the process tools, wherein each of the inter-tool parameters represents an interaction between two of the process tools;
- obtaining a plurality of production routes each of which indicates a predetermined device of the process devices of the process tools in the process stages for processing one of the workpieces;
- receiving a plurality of sets of in-line metrology values of the workpieces, wherein the sets of in-line metrology values are obtained by performing metrology on the workpieces after being processed by their predetermined devices;
- receiving a plurality of sets of final inspection values corresponding to the workpieces, wherein the sets of final inspection values are obtained by performing at least one yield test on each of the workpieces after passing through the production line;
- determining if a yield test failure is encountered according to the final inspection values;
- performing a first phase when the yield test failure is encountered, the first phase comprising:
- preparing a first searching model based on a first algorithm, wherein the first algorithm is a triple phase orthogonal greedy algorithm (TPOGA), a least absolute shrinkage and selection operator (LASSO) algorithm or a regression tree algorithm;
- obtaining a set of interaction data indicating whether the workpieces pass through or are produced by using both of any two elements of a first assembly consisting of the process devices, the process parameters and the inter-tool parameters, according to the production routes, the process parameters and the inter-tool parameters;
- and
- identifying at least one first key process device, and/or at least one first key process parameter, and/or at least one first key inter-tool parameter, and/or at least one first key interaction that are likely to cause the yield test failure by feeding the sets of final inspection values of the workpieces, the sets of in-line metrology values of the workpieces, the production routes of the workpieces, the process parameters, the inter-tool parameters, the set of interaction data into the first searching model;
- and
- when the result of identifying is good, sending at least one updated configuration to the at least one first key process device or devices related to the at least one first key process parameter, the at least one first key inter-tool parameter or the at least one first key interaction for continuous improvements of yield of the production line.
9 more elements
A system for identifying root causes of a yield loss occurring in a production line comprising a plurality of process stages, each of the process stages comprising at least one process tool, each of the at least one process tool comprising at least one process device, each of the at least one process device having a plurality of process parameters configured for processing one of a plurality of workpieces, wherein there are a plurality of inter-tool parameters between the process tools, each of the inter-tool parameters representing an interaction between two of the process tools, the system comprising:
- a memory storing production information and a plurality of sets of final inspection values corresponding to the workpieces, the production information including a plurality of production routes, values of the process parameters, values of the inter-tool parameters and a plurality of sets of in-line metrology values of the workpieces, wherein each of the production routes indicates a predetermined device of the process devices of the process tools in the process stages for processing one of the workpieces, the sets of in-line metrology values are obtained by performing metrology on the workpieces promptly after being processed by their predetermined devices, and the sets of final inspection values are obtained by performing at least one yield test on each of the workpieces after passing through the production line;
- a processor configured to:
- determine if a yield test failure is encountered according to the final inspection values;
- and
- perform a first phase when the yield test failure is encountered, the first phase comprising:
- preparing a first searching model based on a first algorithm, wherein the first algorithm is a triple phase orthogonal greedy algorithm (TPOGA), a least absolute shrinkage and selection operator (LASSO) algorithm or a regression tree algorithm;
- obtaining a set of interaction data indicating whether the workpieces pass through or are produced by using both of any two elements of a first assembly consisting of the process devices, the process parameters and the inter-tool parameters, according to the production routes, the process parameters and the inter-tool parameters;
- and
- identifying at least one first key process device, and/or at least one first key process parameter, and/or at least one first key inter-tool parameter, and/or at least one first key interaction that are likely to cause the yield test failure by feeding the sets of final inspection values of the workpieces, the sets of in-line metrology values of the workpieces, the production routes of the workpieces, the process parameters, the inter-tool parameters, the set of interaction data into the first searching model;
- and
- when the result of identifying is good, send at least one updated configuration to the at least one first key process device or devices related to the at least one first key process parameter, the at least one first key inter-tool parameter or the at least one first key interaction for continuous improvements of yield of the production line.
8 more elements
Dependencies are inferred from the claim text.