Kuo, Justin2018-10-232020-08-222018-08-30Kuo_cornellgrad_0058F_10912http://dissertations.umi.com/cornellgrad:10912bibid: 10489782https://hdl.handle.net/1813/59686This dissertation presents work on using the GHz sonar transducer, a new type of MEMS (microelectromechanical systems) bulk acoustic wave (BAW) technology, for communications and sensing applications. As these devices are fabricated with aluminum nitride (AlN), a CMOS-compatible piezoelectric thin film material, these devices can be integrated directly with CMOS circuits to allow for new circuit functionalities. The structure and fabrication details of GHz sonar transducers are introduced, followed by a discussion of how to model the devices. There are two primary effects that will be discussed – the electrical to acoustic conversion of the piezoelectric thin film transducers and how diffraction affects wave propagation from the transducers through the silicon substrate. In this work, three new applications enabled by the GHz sonar transducer will be discussed. The first is the GHz ultrasonic through-silicon via (UTSV), a new type of wireless 3D interconnect that enables chip-to-chip communication in multi-chip 3D integrated circuit (3DIC) stacks. The second application is the GHz sonic memory – a delay line memory that uses the UTSV as an ultrasonic delay line and stores digital bits as ultrasonic waves. The novelty of this memory is that it transforms the previously unused silicon substrate into 3D memory elements, as opposed to the traditional method of increasing memory density by stacking 2D memory chips. The third application is the GHz ultrasonic fingerprint sensor, a new CMOS compatible fingerprint sensor. The use of ultrasound allows for numerous advantages over current capacitive and optical fingerprint sensors, including higher penetration through glass and metal layers, as well as enhancing the spoof resistance of the fingerprint sensor by allowing the sensor to image elastic properties of tissue. The 1.3 GHz frequency of the sensor potentially allows for two orders of magnitude higher resolution over existing ultrasonic fingerprint sensors, which typically operate at biomedical ultrasound frequencies of tens of MHz.en-USFingerprint sensingGHz ultrasoundAcousticsMEMSElectrical engineeringaluminum nitridedissertation or thesishttps://doi.org/10.7298/X4PG1PZH