This paper suggests a design for a Film Bulk Acoustic Resonator (FBAR) which utilizes a secondary piezoelectric layer for purposes of tuning the FBAR's resonant frequency. Currently, many ceramic resonators have difficulties in on-chip integration, power handling and electrode fabrication. FBARs are not only simple to fabricate and capable of full integration with CMOS/RF IC circuitry, but are also compact and can achieve high frequencies (GHz) with high quality factors. It is widely accepted that piezoelectric actuators encounter a significant change in mechanical stiffness between their open-circuit and closed-circuit states. In addition, it has been previously shown that the resonant frequency of a multi-layer FBAR is a function of the acoustic impedances and, correspondingly, the acoustic velocities, of its respective layers. Since the effective modulus term of the acoustic velocity of an FBAR layer is dependent on both the mechanical properties and electromechanical coupling of its piezoelectric element, and since electromechanical coupling can be altered by means a previously investigated shunt capacitor tuning concept, the stiffness of the piezoelectric tuning layer can be adjusted to vary the resonant frequency of the FBAR. Since difficulties have existed in matching FBAR resonant frequencies to specified values or making the frequencies stable during temperature variations, an active tuning capability for FBARs could offer many possible improvements. This work describes the application of the shunt capacitor tuning to a FBAR resonator and looks at the effects that varying different FBAR parameters have on the frequency range and degree of tunability of the device.
The silicon pressure sensor has been developed for over thirty years and widely used in automobiles, medical instruments, commercial electronics, etc. There are many different specifications of silicon pressure sensors that cover a very large sensing range, from less than 1 psi to as high as 1000 psi. The key elements of the silicon pressure sensor are a square membrane and the piezoresistive strain gages near the boundary of the membrane. The dimensions of the membrane determine the full sensing range and the sensitivity of the silicon sensor, including thickness and in-plane length. Unfortunately, in order to change the sensing range, the manufacturers need to order a customized epi wafer to get the desired thickness. All masks (usually six) have to be re-laid and re-fabricated for different membrane sizes. The existing technology requires at least three months to deliver the prototype for specific customer requests or the new application market. This research proposes a new approach to dramatically reduce the prototyping time from three months to one week. The concept is to tune the rigidity of the sensing membrane by modifying the boundary conditions without changing the plenary size. An extra mask is utilized to define the geometry and location of deep-RIE trenches and all other masks remain the same. Membranes with different depths and different patterns of trenches are designed for different full sensing ranges. The simulation results show that for a 17um thick and 750um wide membrane, the adjustable range by tuning trench depth is about 45% (from 5um to 10um), and can go to as high as 100% by tuning both the pattern and depth of the trenches. Based on an actual test in a product fabrication line, we verified that the total delivery time can be minimized to one week to make the prototyping very effective and cost-efficient.