« PreviousContinue »
United States Patent 
Shah et al.
US005517594A [li] Patent Number: 5,517,594  Date of Patent: May 14, 1996
 THERMAL REACTOR OPTIMIZATION
 Inventors: Sunil C. Shah, Mountain View;
Pradeep Pandey, San Jose, both of
 Assignee: Relman, Inc., Mountain View, Calif.
 Appl. No.: 324,416
 Filed: Oct. 17,1994
 Int. CI.6 H05B 1/02; C23C 14/54
 U.S. CI 392/416; 219/494; 118/50.1;
 Field of Search 392/416, 418;
219/405, 411, 494; 118/50.1, 724, 725;
 References Cited
U.S. PATENT DOCUMENTS
4,276,603 6/1981 Beck et al 364/477
4,688,180 8/1987 Motomiya 364/477
4,711,989 12/1987 Yu 219/390
4,761,538 8/1988 Chiba et al 219/497
4,937,434 6/1990 Nakao 219/390
5,099,442 3/1992 Furutaetal 364/557
5,258,601 11/1993 Takano 219/486
5,280,422 1/1994 Moe et al 364/140
5,291,514 3/1994 Heitmann et al 373/135
G. F. Franklin, J. Powell, and M. Workman, "Digital Control of Dynamic Systems", pp. 349-483, Addison-Wesley, 1990. L. Ljung, "System Identification, Theory for the User", pp. 115-127, Prentice Hall, 1987.
B. D. O. Anderson and J. B. Moore, "Optimal Control: Linear Quadratic Methods", pp. 262-289, Prentice Hall, 1989.
A. Bryson and Y. C. Ho, "Applied Optimal Control", pp. 42-127, Halsted Press, 1975.
Y. M. Cho and T. Kailath, "Model Identification in Rapid Thermal Processing Systems", IEEE Semiconductor Manufacturing, vol. 6, No. 3, pp. 216-229, Aug. 1993.
C. M. Schaper, M. Moslehi, K. Saraswat, and T. Kailath, "Modeling, Identification, and Control of Rapid Thermal Processing Systems", pp. 1-29, Tables 1-5, and FIGS. 1-33, submitted to Journal of the Electromechanical Society, Sep. 1993.
S. Hirasawa, et al., "Temperature Distribution in Semiconductor Wafers Heated in a Vertical Diffusion Furnace", IEEE Trans. Semiconductor Manufacturing, vol. 6, No. 3, pp. 226-232, Aug. 1993.
S. A. Norman, "Optimization of Transient Temperature Uniformity in RTP System", IEEE Trans. Electron Devices, vol. 39, No. 1, pp. 205-207, Jan. 1992. C. M. Schaper, M. Moslehi, K. Saraswat, T. Kailath, "Control of MMST RTP: Repeatability, Uniformity, and Integration for Flexible Manufacturing", pp. 1-24, Table 1, and FIGS. 1-28, submitted to IEEE Trans. Semiconductor Manufacturing, Sep. 1993.
H. De Waard, and W. L. De Konig, "Adaptive Temperature Control of Industrial Diffusion/LPCVD Reactors", pp. 289-294, IFAC Intelligent Tuning and Adaptive Control, Singapore, 1991.
H. De Waard, and W. L. De Konig, "Optimal Control of the Wafer Temperatures in Diffusion/LPCVD Reactors", Automatica, vol. 28, No. 2, pp. 243-253, 1992.
Primary Examiner—Teresa J. Walberg
Assistant Examiner—J. Pelham
Attorney, Agent, or Firm—Blakely, Sokoloff, Taylor & Zafman
A system for controlling a thermal reactor is disclosed that characterizes the thermal reactor with a reactor model that indicates behavior of the thermal reactor and of a load contained in the thermal reactor and that accounts for interaction among a set of heating zones of the thermal reactor. An online reactor model is then determined that estimates the thermal behavior of the load based upon an online input power to the thermal reactor and upon an online temperature indication from the thermal reactor. A time varying temperature and reactant flow recipe is determined that minimizes end of run parameters on the load. A multivariable controller is employed to minimize temperature deviations of the load from a predetermined temperature recipe or time varying trajectory.
45 Claims, 7 Drawing Sheets
Determine a Reactor Model That Reflects the High
Bandwidth Dynamics of the Wafer Load and That Reflects
the Lower Bandwidth Dynamics of Thermal Reactor and
Wafer Load Drift
Determine an Online Reactor Model That Includes Wafer
Temperature Estimators for Differing Reactor
Temperature Ranges and That Includes Wafer
Temperature Tracking Between Estimators
Determine a Multi-Variable Controller From the Online
Reactor Model That Minimizes Wafer Temperature
Deviation From Constant or Time Varying Set Points
Determine a Reactor Time-Varying Trajectory That
Optimizes End of Run Process/Wafer Parameters
Refine the Reactor Time-Varying Trajectory by
Correlating End of Run Test Wafer Parameters to
Wafer Temperatures From the Online Reactor Model