A MEMS Thin Film AlN Supercritical Carbon Dioxide Valve
- Author(s): Chen, Ya-Mei
- Advisor(s): Pisano, Albert P.
- et al.
In this thesis, a new piezoelectric valve system with bi-chevron aluminum nitride (AlN) actuator is described. The intended application of the new piezoelectric valve is for the advanced printing technology with supercritical carbon dioxide as the solvent. With supercritical carbon dioxide as the solvent, the ink dissolved will start to nucleate with a micronozzle and generate extremely small and uniform ink particles due to rapid expansion of supercritical solution (RESS). Therefore, the resolution of the printing can be improved and induce a much better printing quality.
To successfully operate this new printing technology, the operation pressure of this valve system should be as high as 30 MPa. This operation pressure is much higher than what the current piezoelectric MEMS valve can offer. Aluminum nitride is chosen as the piezoelectric material over lead zirconate titanate (PZT) because the depolarization of PZT due to compressive stresses limits the operating pressure to less than 5 MPa. In order to meet high pressure requirements, thin-film AlN is selected because it doesn't experience compressive stress depolarization and is IC compatible. The piezoelectric valve system is designed with bi-chevron shape not only to amplify the generated stroke but also to reduce undesired out-of-plane motion. This amplification mechanism is achieved by the cantilever beam structure without increasing the size of the valve system significantly. In the bi-chevron, the use of piezoelectrics with matched sets of actuator arms enables a push-pull actuation in both directions and also reduces out-of-plane buckling.
To verify the working function of the bi-chevron actuator, a pre-prototype device is introduced and fabricated. And the result from the static testing of the pre-prototype device is compared with the finite element simulation and the theoretical analysis. The result shows a good fitting between finite element simulation and the pre-prototype device measurement (maximum error is about 17%). However, the error between the theoretical analysis and finite element simulation is significant due to the particle top/bottom electrode coverage. This is because partial-coverage results in a nonuniform electrical field along the width of the AlN beam and simultaneously reduces the generated stroke and the generated force. For the dynamic performance, an alternating current (AC) is used to actuate the pre-prototype device rather than direct current (DC). It is because AC actuating voltage gives the dynamic response of the pre-prototype device and then indicates the resonant frequency of the pre-prototype devices which corresponds to the operation speed while DC actuating voltage has no effect on the dynamic performance at all. The result of the dynamic performance shows that the prototype device will give 1.5 micron in-plane generated stroke with acceptable out-of-plane generated stroke when the device is actuated in 60 kHz with 10 V actuating voltage. The pre-prototype device has 1100 micron long, 10 micron wide, and 2° angle AlN beams.
In addition to the pre-prototype device, the prototype device is fabricated for the supercritical carbon dioxide valve system. The difference between prototype devices and supercritical carbon dioxide devices is that supercritical carbon dioxide devices have been sealed with a cap. That means the prototype devices can become goal devices, supercritical carbon dioxide valves, after sealing the device with a cap. This prototype device uses a SOI wafer with bi-chevron AlN actuator to control the flow of the supercritical carbon dioxide valve. This is prototype device is also evaluated and the result verify that this fabrication process is correct.