US6207936B1
Expired / Lapsed
US
⏱️ 0.0 years remaining
Model-based predictive control of thermal processing
Assignee: ASM America Inc
•
Filing: 1997-01-30
•
Granted: 2001-03-27
•
Expires: 2017-01-30
Abstract
Abstract A nonlinear model-based predictive temperature control system is described for use in thermal process reactors. A multivariable temperature response is predicted using a nonlinear parameterized model of a thermal process reactor. The nonlinear parameterized model is implemented using a neural network. Predictions are made in an auto-regressive moving average fashion with a receding prediction horizon. Model predictions are incorporated into a control law for estimating the optimum future control strategy. The high-speed, predictive nature of the controller renders it advantageous in multivariable rapid thermal processing reactors where fast response and high temperature uniformity are needed.
Inventors
- 👤 Henk de Waard
- 👤 James J. Donald
- 👤 Zhimin Lu
- 👤 Robin M. de Keyser
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-10-27 11:12
CPC Classifications
H; H05; H05B; H05B1/00; H05B1/02; G; G05; G05B; G05B13/00; G05B13/02; G05B13/04; G05B13/048; G05B13/0265; G05B13/027; G05D; G05D23/00; G05D23/19; G05D23/1917; G05D23/1927; G05D23/193; G05D23/1931
Claims Analysis
Total Claims: 55
Claim 1
Independent Claim
231 words
A temperature controlled thermal process reactor comprising: a reaction chamber enclosing an object to be heated; a plurality of sources of thermal energy which heat said object; a plurality of thermal sensors, each sensor configured to measure a sensor temperature, each sensor temperature related to an actual temperature of said object, where each thermal sensor provides an output signal representative of said sensor temperature, and where each of said sources of thermal energy affects each of said sensor temperatures; and a model-based predictive temperature controller comprising a nonlinear process model, said temperature controller configured to receive said output signals and control said sources of thermal energy in response to said output signals to produce a selected spatial and temporal distribution of heat energy to maintain a relatively uniform actual temperature on said object, said model-based predictive temperature controller comprising a multivariable thermal process model that relates multivariable process input thermal energy to multivariable process output temperature, a prediction calculator that uses said thermal process model to calculate a predicted nominal temperature output over a future time period, and a control calculator that uses said predicted nominal temperature output to calculate an optimum control strategy to control said sources of thermal energy, said prediction calculator configured to calculate said predicted nominal temperature output recursively over a predetermined future time period using a recursive approximation strategy that begins with an unoptimized initial estimate.
Claim 2
Dependent Claim
231 words
A temperature controlled thermal process reactor comprising: a reaction chamber enclosing an object to be heated; a plurality of sources of thermal energy which heat said object; a plurality of thermal sensors, each sensor configured to measure a sensor temperature, each sensor temperature related to an actual temperature of said object, where each thermal sensor provides an output signal representative of said sensor temperature, and where each of said sources of thermal energy affects each of said sensor temperatures; and a model-based predictive temperature controller comprising a nonlinear process model, said temperature controller configured to receive said output signals and control said sources of thermal energy in response to said output signals to produce a selected spatial and temporal distribution of heat energy to maintain a relatively uniform actual temperature on said object, said model-based predictive temperature controller comprising a multivariable thermal process model that relates multivariable process input thermal energy to multivariable process output temperature, a prediction calculator that uses said thermal process model to calculate a predicted nominal temperature output over a future time period, and a control calculator that uses said predicted nominal temperature output to calculate an optimum control strategy to control said sources of thermal energy, said prediction calculator configured to calculate said predicted nominal temperature output recursively over a predetermined future time period using a recursive approximation strategy that begins with an unoptimized initial estimate.
Claim 3
Dependent Claim
37 words
The temperature controlled thermal process reactor of claim 1 , wherein said prediction calculator calculates the predicted nominal temperature output using said nonlinear process model in a parallel recursion, said prediction calculator having a predetermined prediction horizon.
Claim 4
Dependent Claim
19 words
The temperature controlled thermal process reactor of claim 2 , wherein the prediction calculator assumes a future control strategy.
Claim 5
Dependent Claim
30 words
The temperature controlled thermal process reactor of claim 3 , wherein said thermal process model is adapted to substantially decouple the influence of system input variables from system input disturbance.
Claim 6
Dependent Claim
43 words
The temperature controlled thermal process reactor of claim 1 , wherein the control calculator is adapted to compare said predicted nominal temperature output to a desired future temperature output and uses said comparison in a recursive algorithm to compute said optimum control strategy.
Claim 7
Dependent Claim
21 words
The temperature controlled thermal process reactor of claim 1 , wherein said thermal process model is based on a neural network.
Claim 8
Dependent Claim
86 words
The temperature controlled rapid thermal process reactor of claim 1 , wherein the nonlinear model-based predictive temperature controller comprises: a nonlinear multivariable thermal process model which relates multivariable process input thermal energy to multivariable process output temperature; a prediction calculator which uses said thermal process model, to calculate a predicted nominal temp re output over a predetermined future time period; and a control calculator which uses said predicted nominal temperature output to calculate an optimum control strategy by which to control said source of thermal energy.
Claim 9
Dependent Claim
24 words
The temperature controlled thermal process reactor of claim 7 , wherein said prediction calculator calculates the predicted nominal temperature output using a neural network.
Claim 10
Dependent Claim
19 words
The temperature controlled thermal process reactor of claim 8 , wherein the prediction calculator assumes a future control strategy.
Claim 11
Dependent Claim
19 words
The temperature controlled thermal process reactor of claim 9 , wherein said neural network is a feed forward network.
Claim 12
Dependent Claim
20 words
The temperature controlled thermal process reactor of claim 10 , wherein said neural network comprises a hidden layer of neurons.
Claim 13
Dependent Claim
20 words
The temperature controlled thermal process reactor of claim 4 , wherein said hidden layer of neurons comprises nonlinear sigmoid-type neurons.
Claim 14
Dependent Claim
22 words
The temperature controlled thermal process reactor of claim 8 , wherein said neural network is trained using a pseudo least squares method.
Claim 15
Dependent Claim
32 words
The temperature controlled thermal process reactor of claim 7 , wherein the control calculator compares said predicted nominal temperature output to a desired future temperature output to derive said optimum control strategy.
Claim 16
Dependent Claim
15 words
The temperature controlled thermal process reactor of claim 1 , further comprising a softsensor model.
Claim 17
Dependent Claim
25 words
The temperature controlled thermal process reactor of claim 15 , wherein said softsensor model is created from a dataset generated by using an instrumented wafer.
Claim 18
Dependent Claim
40 words
The temperature controlled thermal process reactor of claim 1 , further comprising a setpoint generator, said setpoint generator automatically generating a correction to said recipe inputs into said thermal process reactor, said correction facilitating control of actual wafer surface temperatures.
Claim 19
Dependent Claim
26 words
The temperature controlled thermal process reactor of claim 17 , said correction facilitating improved control of actual wafer surface temperatures based on measurement of susceptor temperatures.
Claim 20
Dependent Claim
204 words
A temperature control system for controlling a thermal process comprising: a controllable source of thermal energy which comprises multiple independently controllable thermal energy generators to heat an object; a plurality of temperature sensors, each of which measures a temperature related to an actual temperature of said object and which generates an output signal responsive to said temperature; and a model-based predictive temperature controller which receives said output signals and which controls said source of thermal energy in response to said output signals, said controller comprising: a nonlinear thermal process model which relates a process input thermal energy to a process output temperature; a prediction calculator which uses said nonlinear thermal process model to calculate a predicted nominal temperature output over a predetermined future time period; and a control calculator which uses said predicted nominal temperature output to calculate an optimum strategy by which to control said source of thermal energy to minimize spatial thermal gradients across said object, said controller generating output signals to said source of thermal energy in response to said optimum strategy, said control calculator configured to compare a predicted nominal temperature output to a desired future temperature output and use said comparison in an algorithm to compute said optimum strategy.
Claim 21
Dependent Claim
204 words
A temperature control system for controlling a thermal process comprising: a controllable source of thermal energy which comprises multiple independently controllable thermal energy generators to heat an object; a plurality of temperature sensors, each of which measures a temperature related to an actual temperature of said object and which generates an output signal responsive to said temperature; and a model-based predictive temperature controller which receives said output signals and which controls said source of thermal energy in response to said output signals, said controller comprising: a nonlinear thermal process model which relates a process input thermal energy to a process output temperature; a prediction calculator which uses said nonlinear thermal process model to calculate a predicted nominal temperature output over a predetermined future time period; and a control calculator which uses said predicted nominal temperature output to calculate an optimum strategy by which to control said source of thermal energy to minimize spatial thermal gradients across said object, said controller generating output signals to said source of thermal energy in response to said optimum strategy, said control calculator configured to compare a predicted nominal temperature output to a desired future temperature output and use said comparison in an algorithm to compute said optimum strategy.
Claim 22
Dependent Claim
25 words
The temperature control system of claim 19 , wherein said thermal process model substantially decouples the influence of system input variables from system input disturbances.
Claim 23
Dependent Claim
31 words
The temperature control system of claim 19 , wherein said prediction calculator is adapted to use a postulated future control strategy and a recursive algorithm to optimize said postulated future control.
Claim 24
Dependent Claim
28 words
The temperature control system of claim 19 , wherein said nonlinear thermal process model comprises a neutral network, said neural network comprising one or more nonlinear sigmoid-type neurons.
Claim 25
Dependent Claim
25 words
The temperature control system of claim 22 , wherein said thermal process model substantially decouples the influence of system input variables from system input disturbances.
Claim 26
Dependent Claim
16 words
The temperature control system of claim 22 , wherein said prediction calculator comprises a neural network.
Claim 27
Independent Claim
103 words
A method of controlling a thermal process comprising the steps of: measuring one or more process output temperatures; predicting a plurality of future process output temperatures by using a nonlinear thermal process model; using said one or more measured process output temperatures and said predicted future process temperatures to calculate an optimum process input control strategy by comparing one or more of said predicted future process output temperatures to a desired future temperature and using said comparison in an algorithm to compute said optimum process input control strategy; and controlling a process input thermal energy using the calculated optimum process input control strategy.
Claim 28
Independent Claim
103 words
A method of controlling a thermal process comprising the steps of: measuring one or more process output temperatures; predicting a plurality of future process output temperatures by using a nonlinear thermal process model; using said one or more measured process output temperatures and said predicted future process temperatures to calculate an optimum process input control strategy by comparing one or more of said predicted future process output temperatures to a desired future temperature and using said comparison in an algorithm to compute said optimum process input control strategy; and controlling a process input thermal energy using the calculated optimum process input control strategy.
Claim 29
Dependent Claim
57 words
The method of claim 25 , wherein the step of predicting a future process output temperate comprises: identifying a nonlinear thermal process model which relates process input thermal energy to process output temperature; and recursively predicting future process output temperatures using said nonlinear thermal process model, said process output temperature predicted over a predetermined future time period.
Claim 30
Dependent Claim
28 words
The method of claim 26 , wherein the step of predicting future process output temperatures further comprises periodically updating said predictions in accordance with a receding horizon calculation.
Claim 31
Dependent Claim
23 words
The method of claim 25 , wherein the step of predicting a future process output temperature comprises postulating a stationary future control strategy.
Claim 32
Dependent Claim
32 words
The method of claim 25 , wherein the step of calculating an optimum process input control strategy comprises comparing said predicted future process output temperatures to a desired future process output temperature.
Claim 33
Dependent Claim
60 words
The method of claim 25 , wherein the step of predicting a future process output temperature comprises: identifying a nonlinear thermal process model which relates process input thermal energy to process output temperature; and training a neural network to predict fixture process output temperatures using said thermal process model, said process output temperature predicted over a predetermined future time period.
Claim 34
Dependent Claim
28 words
The method of claim 30 , wherein the step of predicting future process output temperatures firer comprises periodically updating said predictions in accordance with a receding horizon calculation.
Claim 35
Dependent Claim
23 words
The method of claim 30 , wherein the step of predicting a future process output temperature comprises postulating a stationary future control strategy.
Claim 36
Dependent Claim
32 words
The method of claim 30 , wherein the step of calculating an optimum process input control strategy comprises comparing said predicted future process output temperatures to a desired fixture process output temperature.
Claim 37
Dependent Claim
227 words
A temperature controlled thermal process reactor comprising: a reaction chamber enclosing an object to be headed; a plurality of sources of thermal energy which heat said object; a plurality of thermal sensors, each sensor configured to measure a sensor temperature, each sensor temperature related to an actual temperature of said object, where each thermal sensor provides an output signal representative of said sensor temperature, and wherein each of said sources of thermal energy affects each of said sensor temperatures; and a model-based predictive temperature controller which receive said output signals and which controls said sources of thermal energy in response to said output signals to produce a selected spatial and temporal distribution of heat energy to maintain a relatively uniform actual temperature on said object said model-based predictive temperature controller comprising a multivariable thermal process model that relates multivariable process input thermal energy to multivariable process output temperature, a prediction calculator that uses said thermal process model to calculate a predicted nominal temperature output over a predetermined future time period and a control calculator that uses said predicted nominal temperature output to calculate an optimum control strategy by which to control said sources of thermal energy said control calculator adapted to compare a predicted nominal temperature output to a desired future temperature output and use said comparison in a recursive algorithm to compute an optimum contort strategy.
Claim 38
Dependent Claim
227 words
A temperature controlled thermal process reactor comprising: a reaction chamber enclosing an object to be headed; a plurality of sources of thermal energy which heat said object; a plurality of thermal sensors, each sensor configured to measure a sensor temperature, each sensor temperature related to an actual temperature of said object, where each thermal sensor provides an output signal representative of said sensor temperature, and wherein each of said sources of thermal energy affects each of said sensor temperatures; and a model-based predictive temperature controller which receive said output signals and which controls said sources of thermal energy in response to said output signals to produce a selected spatial and temporal distribution of heat energy to maintain a relatively uniform actual temperature on said object said model-based predictive temperature controller comprising a multivariable thermal process model that relates multivariable process input thermal energy to multivariable process output temperature, a prediction calculator that uses said thermal process model to calculate a predicted nominal temperature output over a predetermined future time period and a control calculator that uses said predicted nominal temperature output to calculate an optimum control strategy by which to control said sources of thermal energy said control calculator adapted to compare a predicted nominal temperature output to a desired future temperature output and use said comparison in a recursive algorithm to compute an optimum contort strategy.
Claim 39
Dependent Claim
20 words
The temperature controlled thermal process reactor of claim 34 , wherein the model-based predictive temperature controller comprises multivariable temperature control.
Claim 40
Dependent Claim
29 words
The temperature controlled thermal process reactor of claim 34 wherein said prediction calculator calculates the predicted nominal temperature output using an auto-regressive moving average using a predetermined prediction horizon.
Claim 41
Dependent Claim
24 words
The temperature controlled thermal process reactor of claim 36 , wherein the prediction calculator calculates an unoptimized initial estimate for a future control strategy.
Claim 42
Dependent Claim
46 words
The temperature controlled thermal process reactor of claim 37 , wherein said prediction calculator is further adapted to calculate said predicted nominal temperature output recursively over a predetermined future time period using a recursive approximation strategy, said recursive approximation strategy beginning with said unoptimized initial estimate.
Claim 43
Dependent Claim
30 words
The temperature controlled thermal process reactor of claim 38 , wherein said thermal process model is adapted to substantially decouple the influence of system input variables from system input disturbances.
Claim 44
Dependent Claim
199 words
A temperature control system for controlling a thermal process comprising: a controllable source of thermal energy which comprises multiple independently controllable thermal energy generators to heat an object; a plurality of temperature sensors, each of which measures a temperature related to an actual temperature of said object and which generates an output signal responsive to said temperature; and a model-based predictive controller which receives said output signals and which controls said source of thermal energy in response to said output signals, said controller comprising; a thermal process model that relates process input thermal energy to process output temperature; a prediction calculator that uses said thermal process model to calculate a predicted nominal temperature output over a predetermined future time period; and a control calculator that uses said predicted nominal temperature output to calculate an optimum strategy by which to control said source of thermal energy to minimize spatial thermal gradients across said object, said controller generating output signals to said source of thermal energy in response to said optimum strategy said control calculator configured to compare a predicted nominal temperature output to a desired future temperature output and use said comparison in an algorithm to compute said optimum strategy.
Claim 45
Dependent Claim
199 words
A temperature control system for controlling a thermal process comprising: a controllable source of thermal energy which comprises multiple independently controllable thermal energy generators to heat an object; a plurality of temperature sensors, each of which measures a temperature related to an actual temperature of said object and which generates an output signal responsive to said temperature; and a model-based predictive controller which receives said output signals and which controls said source of thermal energy in response to said output signals, said controller comprising; a thermal process model that relates process input thermal energy to process output temperature; a prediction calculator that uses said thermal process model to calculate a predicted nominal temperature output over a predetermined future time period; and a control calculator that uses said predicted nominal temperature output to calculate an optimum strategy by which to control said source of thermal energy to minimize spatial thermal gradients across said object, said controller generating output signals to said source of thermal energy in response to said optimum strategy said control calculator configured to compare a predicted nominal temperature output to a desired future temperature output and use said comparison in an algorithm to compute said optimum strategy.
Claim 46
Dependent Claim
29 words
The temperature control system of claim 40 , wherein said thermal process model has parameters selected to substantially decouple the influence of system input variables from system input disturbances.
Claim 47
Dependent Claim
33 words
The temperature control system of claim 40 , wherein said prediction calculator is adapted to use a recursive algorithm to compute said a predicted nominal temperature output over a predetermined future time period.
Claim 48
Dependent Claim
36 words
The temperature control system of claim 40 , wherein said prediction calculator comprises means for calculating an auto-regressive moving average, said means for calculating an autoregressive moving average comprising means for computing a receding calculation horizon.
Claim 49
Dependent Claim
32 words
The temperature control system of claim 40 , wherein said prediction calculator is adapted to use a postulated future control strategy and a recursive algorithm to optimize said postulated future control strategy.
Claim 50
Independent Claim
98 words
A method of controlling a thermal process comprising: measuring a plurality of process output temperatures; predicting a plurality of future process output temperatures; using said measured process output temperatures and said predicted future process temperatures to calculate an optimum process input control strategy by comparing one or more of said predicted future process output temperatures to a desired future temperature and using said comparison in an algorithm to compute said optimum process input control strategy; and controlling a plurality of process input thermal energy sources using the calculated optimum process input control strategy to reduce spatial thermal gradients.
Claim 51
Independent Claim
98 words
A method of controlling a thermal process comprising: measuring a plurality of process output temperatures; predicting a plurality of future process output temperatures; using said measured process output temperatures and said predicted future process temperatures to calculate an optimum process input control strategy by comparing one or more of said predicted future process output temperatures to a desired future temperature and using said comparison in an algorithm to compute said optimum process input control strategy; and controlling a plurality of process input thermal energy sources using the calculated optimum process input control strategy to reduce spatial thermal gradients.
Claim 52
Dependent Claim
55 words
The method of claim 45 , when the step of predicting a future process output temperature comprises: identifying a thermal process model that relates process input thermal energy to process output temperature; and recursively predicting future process output temperatures using said thermal process model, said process output temperature predicted over a predetermined future time period.
Claim 53
Dependent Claim
31 words
The method of claim 46 , wherein the step of predicting future process output temperatures further comprises periodically updating said predictions in accordance with a receding horizon auto-regressive moving average calculation.
Claim 54
Dependent Claim
23 words
The method of claim 45 , wherein the step of predicting a future process output temperature comprises postulating a stationary future control strategy.
Claim 55
Dependent Claim
32 words
The method of claim 45 , wherein the step of calculating an optimum process input control strategy comprises comparing said predicted future process output temperatures to a desired future process output temperature.
References Cited (Backward Citations)
US3836751A
Cited
US4358822A
Cited
US4255133A
Cited
US4349869A
Cited
EP0037579A2
Cited
US4714988A
Cited
US4639853A
Cited
US4616308A
Cited
US4688180A
Cited
US4720807A
Cited
US4694390A
Cited
US4680451A
Cited
US4634946A
Cited
US4761538A
Cited
US4736316A
Cited
US4769766A
Cited
US4975561A
Cited
US4904912A
Cited
US4959767A
Cited
US5032977A
Cited
US5099442A
Cited
US5301101A
Cited
US5291514A
Cited
CA2054423A1
Cited
US5488561A
Cited
US5790750A
Cited
US5660472A
Cited
US5755511A
Cited
EP0829784A1
Cited
US5793022A
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 2004-03-26 | 2004-09-25 | 2005-03-27 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2008-03-26 | 2008-09-25 | 2009-03-27 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2012-03-26 | 2012-09-25 | 2013-03-27 | $7700.00 | Window Closed / Lapsed |