| Summary: | In photolithography scanning tools, the functional patterns of integrated circuit layers are defined with critical dependence on the actuation of reticle and wafer stages along precisely synchronized trajectories. Patterning throughput of such tools is limited based on the velocity and acceleration at which the stages are actuated. Modern tools require sub-nanometer accuracy of stages along these trajectories during constant-velocity scan exposure to create feature sizes on the order of nanometers. At the ends of the constant velocity scans, high acceleration trajectories are used to reverse the scan velocity in minimal time. The next-generation of photolithography tools will require more aggressive trajectories along with the development of energy-efficient actuation solutions with higher force and precision capabilities to implement these demanding motion profiles.
In this thesis, we propose actuation concepts which may enable 100g reticle stage turnaround accelerations, and explore two such concepts in depth. The first concept is an array of piezoelectric stack actuators attached to the long-stroke stage which mechanically contact the short-stroke stage only during turnaround. In this context, we perform a scaled two degree-of-freedom experiment in which we attempt to control the contact of a 840 g payload moving at a velocity of 80 mm/s using a 50 𝜇m stroke piezo stack actuator which is driven open-loop. We are able to use the piezo current signal to detect mechanical contact with an estimated delay of 6-16 𝜇s. We are unable to control the dynamics of the contact during which the measured peak contact force of 150 N exceeds the planned amount by 80% and results in the payload bouncing off the actuator.
The second actuation concept we consider in theory is the use of dual-chamber pneumatic springs as energy storage devices to create turnaround forces for the long-stroke stage acceleration. We examine the use of such pneumatic springs in parallel with a conventional long-stroke linear motor to create a stage topology in which reactive power is stored and returned into kinetic energy. We study thermal aspects of the spring behavior first under an adiabatic assumption and then using a one-dimensional thermal model for heat flow through the piston chamber walls. The proposed design shows promise to reduce the motor power dissipation by 90% and the motor amplifier electrical power by 70%, showing promise for further study. Such energy savings can contribute to significant reduction in the energy consumption of lithography tools.
|
|---|