New actuators can traverse both long distances and subnanometer increments with equal ease.
No question the semiconductor industry now puts a premium on high-position resolution and speed: Wafers are getting larger, but chip features are ever smaller and closer together. Storage media is getting more dense. Equipment that fabs such devices must be speedy but precise. Semiconductor process technologies that utilize e-beam, ion beam, DUV, and X-ray all face such pressures. Stages must traverse a 300-mm wafer at high speeds, narrow in on a specific region, and perform ultrahigh-resolution work with high position stability.
Traditionally, the approach to synthesizing precise but speedy moves has been to combine two axes in tandem as a way of getting both a high traverse rate and ultrahigh resolution. The bottom axis would give coarse positioning and was typically driven by a mechanical screw or linear servomotor. The top axis provided fine positioning and was driven by traditional piezoactuators able to resolve distances in the tens of nanometers.
Combining parallel axes can not only be costly, but also compounds potential positioning errors. There are often issues related to position stability of one axis with respect to the other. And the machine footprint necessitated by two axes can be appreciable.
But recently a new technology has emerged that can combine the performance of a high-speed coarse drive with a high-resolution stage, eliminating the need for tandem axes. The technology is based on the principles of piezoceramic material, generating both coarse and fine positioning.
To synthesize coarse motion, piezoceramic elements made of proprietary material are used to generate an " ultrasonic standing wave." In this mode of operation the device functions as a "ceramic servomotor." It mounts to either a linear or rotary stage, coupled through a normal force perpendicular to the direction of travel. The ultrasonic standing wave functions as a closed-loop servomotor, driving a linear stage with appreciable travel, at velocities from 1m sec to 300 mm/sec.
The effect is one of the motor "wiggling" at 39.6 kHz. Amplitude oscillations at the end of a "fingertip" attached to the piezoceramic element trace out an ellipse. The fingertip can be pressed against any surface. In this mode of operation, the fingertip first pushes against a stage or other mechanical structure to advance it. Then the tip retracts and returns to the starting position. It again pushes on the stage to repeat the cycle.
The action is analogous to putting a piece of paper on a desk and using your fingertips to inch it toward you, allowing your fingers to make several cycles as the page moves.
In this coarse-motion or servomode, an amplifier driving the motor gets an ordinary ±10-V signal from a servocontroller. The amplifier excites the motor's piezoceramic element in a transverse and longitudinal manner as just described. The simultaneous excitation of the two modes is what creates the small elliptical path at the tip of the motor. This action provides exceptionally smooth motion. The resolution of the resulting moves is generally that of the system's position encoder.
The fine positioning mode is called the dc mode. In this mode, the range of travel is limited to ±300 nm with subnanometer resolution. The synthesis of fine positioning comes about from the piezoceramic material functioning as a piezoactuator when excited with a dc voltage. When driven with dc, the same drive motor able to move a stage at 300 mm/sec can make a 1-nm move.
The motor amplifier uses the same ±10-V input signal from the servocontroller to drive the motor in the dc mode. In this mode, for example, an input of 100 mV from the controller may produce about 0.8 nm of movement from the motor. Each motor and drive has a lookup table of motor displacement versus input voltage. The resulting motor/drive systems can operate open-loop or closed-loop with a high-resolution encoder or laser interferometer.
Clearly there are a host of mechanical bearing and encoder issues that also impact the ability to position equipment at the nanometer or subnanometer level. For example, the stiffness and smoothness of the bearings used have as much impact as the drive technology. So careful design is still a must.
The areas where such coarse and fine-positioning capabilities are useful include wafer fabrication and inspection, mask correction, disk-drive certification, and spin stands. In some instances, devices are configured with three parallel axes to get the levels of motion and position stability such systems demand.
Ceramic servomotors are also vacuum compatible, have no external or intrinsic magnetic field, and stack up well against brushless dc-linear motors and other precision drive technologies.
Finally, it is interesting to compare nonpiezo drives, such as leadscrews and ball screws, with the piezo approach. Conventional drives are, generally speaking, too compliant to manage tasks that combine speeds of 200 mm/sec or faster with manipulation at the nanometer level. They transmit linear motion through a rotary motor, flexible coupling, and series of shaft bearings. Moreover, it is exceptionally challenging to make all these components vacuum compatible.
Similarly, brushless linear motors have been widely used in such applications because they can hit high speeds, can deliver appreciable force, and exhibit high dynamic stiffness. But their position stability is not the best. These noncontact drives tend to experience servo dither at the level of ±1 encoder count. Linear motors also introduce high magnetic fields that require extensive and expensive shielding when used with e-beam and ion beam processes.