Research on Multi-variable Closed-loop Control Strategy for Electrostatic Chuck High Voltage Power Supply in Semiconductor Front-end Process
Semiconductor front-end manufacturing processes require precise control of wafer conditions. Electrostatic chucks hold wafers securely during processing using electrostatic attraction. The high voltage power supply that biases the electrostatic chuck must maintain precise control under varying conditions. Multiple variables interact to determine the chucking performance. Closed-loop control strategies address the multi-variable nature of the problem. Understanding the control requirements enables development of effective electrostatic chuck power supplies.
Electrostatic chuck operation principles involve electrostatic attraction. The chuck electrode is biased at high voltage relative to the wafer. The electric field induces charge in the wafer and chuck dielectric. The attractive force holds the wafer against the chuck surface. The force must be sufficient to resist process forces. The force distribution must be uniform across the wafer.
Process requirements for electrostatic chucks are demanding. Plasma etching processes generate ion bombardment forces. Thermal cycles cause wafer expansion and contraction. Backside gas cooling requires sealed contact. Wafer handling requires controlled de-chucking. The chuck must perform reliably through many process cycles. The power supply must support all these requirements.
Control variables for electrostatic chuck systems include multiple parameters. The chuck voltage determines the attractive force. The wafer temperature affects the chucking characteristics. The backside gas pressure affects the thermal contact. The plasma conditions affect the wafer charging. The process time affects the chucking history. These variables interact in complex ways.
Chucking force control is the primary objective. The force must be sufficient for secure holding. Excessive force can cause wafer damage or sticking. The force must be uniform across the wafer. The force must respond to process variations. The force control must be stable and repeatable.
De-chucking control is critical for wafer handling. The attractive force must be removed for wafer release. Residual charge can cause sticking or wafer damage. Controlled voltage reduction enables safe release. The de-chucking sequence must be optimized. The release must be complete and repeatable.
Multi-variable interactions complicate the control problem. Temperature affects the dielectric properties. Backside gas affects the voltage distribution. Plasma affects the wafer charging. Process history affects the chuck surface condition. The control strategy must account for all interactions.
Closed-loop control strategies address the multi-variable nature. Multiple sensors provide feedback of relevant variables. The controller processes the sensor inputs. The control outputs adjust the power supply parameters. The control algorithm coordinates all variables. The closed-loop approach compensates for disturbances.
Sensor requirements for closed-loop control include several measurements. Voltage sensors measure the chuck bias. Current sensors measure the leakage current. Temperature sensors measure the wafer and chuck temperatures. Force sensors may measure the chucking force indirectly. The sensors must have adequate accuracy and response time.
Control algorithm design for multi-variable systems is complex. Proportional-integral-derivative control provides basic regulation. Model-based control uses process models for prediction. Adaptive control adjusts parameters for varying conditions. Model predictive control optimizes over a future horizon. The algorithm must be appropriate for the control requirements.
Decoupling strategies address variable interactions. The interactions cause coupling between control loops. Decoupling controllers compensate for the interactions. The decoupling improves the control performance. The decoupling design requires understanding of the interactions. The decoupling must be robust against model uncertainties.
Disturbance rejection is important for process control. Plasma conditions can change rapidly. Temperature variations occur during processing. Backside gas pressure may fluctuate. The control must reject these disturbances. The disturbance rejection bandwidth must be adequate. The control must maintain performance despite disturbances.
Setpoint tracking enables process optimization. The chucking force may need to vary during the process. The temperature profile may require specific patterns. The control must track the setpoint changes accurately. The tracking performance affects the process results. The tracking bandwidth must be appropriate for the process.
Implementation considerations affect the practical deployment. The control hardware must have adequate capability. The software must be reliable and maintainable. The tuning must be appropriate for each process. The calibration must be maintained over time. The implementation must support production requirements.
Validation of control performance requires comprehensive testing. Step response testing characterizes the dynamic behavior. Disturbance rejection testing verifies the robustness. Long-term testing verifies the stability. Process integration testing verifies the overall performance. The validation must cover all relevant operating conditions.
