Abstract
Rapid Thermal Processing (RTP) technology is a delicate field to the control engineer. Its compatibility to single-wafer processing is well suited for performing thermal steps in the state-of-the-art integrated circuit (IC) manufacturing. Control of the wafer temperature during the processing is essential. The main problem in the scalar (SISO) approach is due to the steady-state temperature uniformity. A solution to this problem is to vary the spatial energy flux distribution radiating to the wafer. To achieve this requirement, one approach is the use a multivariable (MIMO) control law to manipulate independently the different lamps banks. Thermal process are highly non linear and distributed in nature. Besides, these non-linearities implies process dynamics variations. In this work, after physically describing our process about a reference value of the power and temperature, we present an off-line identification procedure (in the aim of devising a linear multivariable model) using input/output data for different reference values from real experiences and multi-variable least square algorithm. Afterwards, particular attention is devoted to the structure parameter determination of the linear model. Based on the linear model, a multivariable PID controller is designed. The controller coupled with the least mean square identification algorithm is tested under real conditions. The
More>>
Citation Formats
Bardon, B.
Modelling and multiparameter control applied to a fast annealing furnace; Modelisation et Commande Multivariable Appliquee a un Four de Recuit Rapide.
France: N. p.,
1995.
Web.
Bardon, B.
Modelling and multiparameter control applied to a fast annealing furnace; Modelisation et Commande Multivariable Appliquee a un Four de Recuit Rapide.
France.
Bardon, B.
1995.
"Modelling and multiparameter control applied to a fast annealing furnace; Modelisation et Commande Multivariable Appliquee a un Four de Recuit Rapide."
France.
@misc{etde_462918,
title = {Modelling and multiparameter control applied to a fast annealing furnace; Modelisation et Commande Multivariable Appliquee a un Four de Recuit Rapide}
author = {Bardon, B}
abstractNote = {Rapid Thermal Processing (RTP) technology is a delicate field to the control engineer. Its compatibility to single-wafer processing is well suited for performing thermal steps in the state-of-the-art integrated circuit (IC) manufacturing. Control of the wafer temperature during the processing is essential. The main problem in the scalar (SISO) approach is due to the steady-state temperature uniformity. A solution to this problem is to vary the spatial energy flux distribution radiating to the wafer. To achieve this requirement, one approach is the use a multivariable (MIMO) control law to manipulate independently the different lamps banks. Thermal process are highly non linear and distributed in nature. Besides, these non-linearities implies process dynamics variations. In this work, after physically describing our process about a reference value of the power and temperature, we present an off-line identification procedure (in the aim of devising a linear multivariable model) using input/output data for different reference values from real experiences and multi-variable least square algorithm. Afterwards, particular attention is devoted to the structure parameter determination of the linear model. Based on the linear model, a multivariable PID controller is designed. The controller coupled with the least mean square identification algorithm is tested under real conditions. The performances of the MIMO adaptive controller is also evaluated in tracking as well as in regulation. (author) refs.}
place = {France}
year = {1995}
month = {Jan}
}
title = {Modelling and multiparameter control applied to a fast annealing furnace; Modelisation et Commande Multivariable Appliquee a un Four de Recuit Rapide}
author = {Bardon, B}
abstractNote = {Rapid Thermal Processing (RTP) technology is a delicate field to the control engineer. Its compatibility to single-wafer processing is well suited for performing thermal steps in the state-of-the-art integrated circuit (IC) manufacturing. Control of the wafer temperature during the processing is essential. The main problem in the scalar (SISO) approach is due to the steady-state temperature uniformity. A solution to this problem is to vary the spatial energy flux distribution radiating to the wafer. To achieve this requirement, one approach is the use a multivariable (MIMO) control law to manipulate independently the different lamps banks. Thermal process are highly non linear and distributed in nature. Besides, these non-linearities implies process dynamics variations. In this work, after physically describing our process about a reference value of the power and temperature, we present an off-line identification procedure (in the aim of devising a linear multivariable model) using input/output data for different reference values from real experiences and multi-variable least square algorithm. Afterwards, particular attention is devoted to the structure parameter determination of the linear model. Based on the linear model, a multivariable PID controller is designed. The controller coupled with the least mean square identification algorithm is tested under real conditions. The performances of the MIMO adaptive controller is also evaluated in tracking as well as in regulation. (author) refs.}
place = {France}
year = {1995}
month = {Jan}
}