System and method for monitoring process quality control

• US 6,766,283 B1
• Filed: 10/13/2000
• Issued: 07/20/2004
• Est. Priority Date: 10/13/2000
• Status: Expired due to Fees

• Current Assignee: ADA ANALYTICS ISRAEL LTD (Advanced Emissions Solutions, Inc.)
• Original Assignee: Insyst, Ltd. (Advanced Emissions Solutions, Inc.)
• Inventors: Goldman, Arnold J., Hartman, Jehuda, Fisher, Joseph, Sarel, Shlomo
• Primary Examiner(s): PALADINI, ALBERT WILLIAM
• Application Number:
  US09/689,884
• Time in Patent Office: 1,376 Days
• Field of Search: 703/2, 700/108, 700/109, 700/110, 700/121, 706/16
• US Class Current: 703/2
• CPC Class Codes: G05B 19/41875 characterised by quality surveillance of production
• CPC Class Codes: G05B 2219/32184 Compare time, quality, state of operators with threshold value
• CPC Class Codes: G05B 2219/32187 Correlation between controlling parameters for influence on quality parameters
• CPC Class Codes: Y02P 90/02 Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]

Abstract

A method of modeling a monitorable stage in a process is provided. The method including the steps of: (a) measuring at least one input value of a parameter of the monitorable stage of the process; (a) measuring at least one output value of the parameter of the monitorable stage of the process; and (c) utilizing the at least one input value and the at least one output value for constructing a process output empirical model for uncovering a functional relationship between the at least one input value and at least one output value, the step of constructing the process output empirical modeler being effected by: (i) dividing at least one interval of the parameter into a plurality of sub intervals, such that each of the at least one interval is divided into at least two of the sub intervals; (ii) classifying the at least one output value according to the plurality of sub intervals, thereby presenting the at least one output value as a plurality of discrete variables defining the at least one output value; and (iii) using the plurality of discrete variables defining the at least one output value for defining the functional relationship between the at least one input value and the at least one output value, thereby modeling the monitorable stage of the process.

Claims

1. A method of monitoring a process having a plurality of input parameters, each parameter having respective variation range and taking a value within said respective variation range, and said process having at least one output parameter, the method comprising:

dividing said variation ranges of said input parameters into sub-ranges,

obtaining a series of input values for each of said input parameters, said input values respectively falling into respective ones of said sub-ranges,

representing each input value by its corresponding sub-range,

obtaining a corresponding series of output values for said at least one output parameter,

for each one of input states of said system, each input state comprising corresponding values for each of said input parameters, recording an association of a combination of sub-ranges with values of said at least one output parameter corresponding to said input state,

from said associations constructing a statistical model relating said sub-ranges of said plurality of input values of said input state with said output values and

using said statistical model to predict output behaviour of, said process in terms of said input states.

2. A method according to claim 1, wherein said plurality of input parameters define an input space, wherein said relating said sub-ranges of said plurality of input values with said output values comprises:

building said combinations of subranges as a plurality of tuples, being one sub-range for each input parameter, thereby defining the input range by said tuples,

associating measured outputs with a tuple describing corresponding inputs,

said tuples thereby defining said inputs states.

3. A method according to claim 2, wherein said constructing a statistical model comprises associating each tuple with a single value being a statistically processed result of respective corresponding values of said output parameter.

4. A method according to claim 2, wherein at least one tuple is associated with a probability of occurrence.

5. A method according to claim 4, wherein any tuple associated with a low probability of occurrence is further associated with an alarm.

6. A method according to claim 2, wherein said states are grouped into steady states of said process and states requiring corrective action to said process.

7. A method according to claim 2, wherein said process is a part of a larger process.

8. A method according to claim 2, wherein said process is at least part of a semiconductor wafer manufacture process.

9. A method according to claim 1, comprising the further steps of

identifying the statistical distribution of output values corresponding to at least some of said sub-ranges,

modifying at least one of the boundaries of a sub-range,

reassigning said input values to said modified sub-ranges in accordance with said modified sub-range boundaries,

re-identifying the statistical distribution of output values corresponding to at least some of said sub-ranges,

and selecting the subranges giving a better statistical distribution according to a predefined distribution criterion.

10. A method according to claim 9, wherein said steps of modifying, reassigning, re-identifying and selecting are repeated until a predefined finishing criterion is met.

11. A method according to claim 9, wherein said predefined distribution criterion is a low mean square distribution.

12. A method of modeling a relationship between a plurality of input parameters each having an variation range and an output parameter, said relationship having a plurality of possible states, the method comprising:

discretizing said variation ranges into a plurality of sub-range vectors such that there is a plurality of sub-range vectors for each input parameter,

building tuples of combinations of said sub-range vectors from respective ones of said input parameters,

determining a statistical association between each tuple and values of said output parameter, and

using said statistical association to model said relationship.

13. A method according to claim 12, wherein said step of determining a statistical relationship comprises associating said tuple with a single value being a statistically processed result of said corresponding values of said output parameter.

14. A method according to claim 12, comprising the modifying a current set of said sub-range vectors, said modifying comprising the further steps of

identifying the statistical distribution of output values corresponding to at least some of said subrange vectors,

modifying at least some of the boundaries of subranges used in at least some of said subrange vectors,

reassigning respective input values to said sub-range vectors in accordance with said modified sub-range boundaries,

recalculating the statistical distribution of output values corresponding to said modified subrange vectors,

and selecting the modified set of subrange vectors if the modification gives a better statistical distribution according to a predefined distribution criterion.

15. A method according to claim 14, wherein said steps modifying a current set of subrange vectors is repeated until a predefined finishing criterion is met.

16. A method according to claim 14, wherein said predefined distribution criterion is a low mean square distribution.

17. A method according to claim 12, wherein at least some of said states indicate actions to be taken.

18. A system for monitoring a process having a plurality of input parameters, each parameter taking values within an input variation ranges, at least one output parameter taking values within an output variation range, the system having a plurality of possible operational states each associated with different parts of said output range, the system comprising:

an input value recorder for recording a series of values of said input parameters,

an output value recorder for recording a corresponding series of values of said at least one output parameter,

a range divider for dividing said variation ranges of said input parameters into sub-range vectors,

a tuple builder for building combinations of sub-range vectors from respective ones of said input parameters,

a first associator for associating each tuple with statistically corresponding values of said at least one output parameter, and

a second associator for associating each tuple with one of said possible operational states of said process, thereby to monitor said process in terms of said states.

19. A system according to claim 18, further comprising a statistical analyzer, associated with said first associator, for producing a single value being a statistically processed result of said corresponding values of said at least one output parameter.

20. A system according to claim 18, further comprising an intelligent decision maker operable to use said tuples to provide numerical values for nodes in a decision tree of said process, and to make decisions based on desired outputs and on said values.

21. A system according to claim 18, further comprising a vector annealer for annealing sets of said sub-range vectors based on the statistical range of said associated corresponding values of a corresponding input parameter.

22. A system according to claim 21, operable to distinguish between inputs which are effective in governing said process and inputs which are ineffective.

23. A system according to claim 18, wherein said states are indications of a quality level of said process.

24. A system according to claim 18, wherein at least one of said states is indicative of corrective action to be taken in said process.

25. A system according to claim 18, wherein said process is a part of a larger process.

26. A system according to claim 18, wherein said process is at least a part of a semiconductor wafer manufacturing process.