Nonlinear Finite Element Calculations of Layered SAW Resonators.

In this work the nonlinear behavior of layered SAW resonators is studied with the help of Finite Element (FE) computations. The full calculations depend strongly on the availability of accurate tensor data. While there are accurate material data for linear computations, the complete sets of higher-order material constants, needed for nonlinear simulations, are still not available for relevant materials. To overcome this problem, scaling factors were used for each available nonlinear tensor. The approach here considers piezoelectricity, dielectricity, eletrostriction and elasticity constants up to fourth order. These factors act as a phenomenological estimate for incomplete tensor data. Since no set of fourth order material constants for LiTaO3 is available, an isotropic approximation for the fourth order elastic constants was applied. As a result, it was found that the fourth order elastic tensor is dominated by one fourth order Lamé constant. With the help of the FE model, derived in two different, but equivalent ways, we investigate the nonlinear behavior of a SAW resonator with a layered material stack. The focus was set to third order nonlinearity. Accordingly, the modeling approach is validated using measurements of third order effects in test resonators. In addition, the acoustic field distribution is analyzed.

A Low-Power RRAM Memory Block for Embedded, Multi-Level Weight and Bias Storage in Artificial Neural Networks.

Pattern recognition as a computing task is very well suited for machine learning algorithms utilizing artificial neural networks (ANNs). Computing systems using ANNs usually require some sort of data storage to store the weights and bias values for the processing elements of the individual neurons. This paper introduces a memory block using resistive memory cells (RRAM) to realize this weight and bias storage in an embedded and distributed way while also offering programming and multi-level ability. By implementing power gating, overall power consumption is decreased significantly without data loss by taking advantage of the non-volatility of the RRAM technology. Due to the versatility of the peripheral circuitry, the presented memory concept can be adapted to different applications and RRAM technologies.

Electrothermal Modeling of Surface Acoustic Wave Resonators and Filters.

In this article, an electrothermal modeling approach of surface acoustic wave (SAW) resonators and filters is presented. The starting point for the model is a preliminary design that has to be assessed for thermal aspects. Due to the high geometrical complexity of SAW components, simplifications are elaborated and qualified on resonator and filter levels to prepare the design for thermal simulation. A thermal model is created and simulated in a finite-element method environment. The simulated behavior is exported as a thermal impedance and implemented in a circuit model of a SAW filter. The layout's electromagnetic behavior is taken into account. Electrothermal models of the SAW resonators and the bus bars are developed. The interface to the thermal impedance is achieved by the use of electrothermal ports. The dynamic effect of the frequency shift is included. Verification is done by a comparison of the temperature increase of a resonator in a filter test structure to a corresponding simulation model. The filter is excited by a radio frequency large signal, and the temperature is detected by the use of a resistive temperature sensor. A simulation that shows the impact of mutual heating between the resonators in a filter environment is performed.

Design of an Integrated Subretinal Implant using Cellular Neural Networks for Binary Image Generation in a 130 nm BiCMOS Process.

Blindness caused by the eye diseases Retinitis-Pigmentosa and Age-Related-Macular-Degeneration leads to a degeneration of the photoreceptor layer while postsynaptic cells mostly stay intact. In this Paper a new concept for retinal implants is proposed. Instead of converting the incident light to a gray-scale picture with corresponding continuous-value stimulation levels, we here suggest to produce a binary image picture that only highlight edges in order to stimulate the retina solely at points which belong to an edge. An integrated test circuit is designed with a 130 nm BiCMOS process by using cellular neural networks for binary image generation. The circuit yields a simulated maximum rated power consumption of 2.61 mW for a 1000 information processing cells.

Investigation on Third-Order Intermodulation Distortions Due to Material Nonlinearities in TC-SAW Devices.

Nonlinearity can give rise to intermodulation distortions in surface acoustic wave (SAW) devices operating at high input power levels. To understand such undesired effects, a finite element method (FEM) simulation model in combination with a perturbation theory is applied to find out the role of different materials and higher order nonlinear tensor data for the nonlinearities in such acoustic devices. At high power, the SAW devices containing metal, piezoelectric substrate, and temperature compensating (TC) layers are subject to complicated geometrical, material, and other nonlinearities. In this paper, third-order nonlinearities in TC-SAW devices are investigated. The materials used are LiNbO3-rot128YX as the substrate and copper electrodes covered with a SiO2 film as the TC layer. An effective nonlinearity constant for a given system is determined by comparison of nonlinear P-matrix simulations to third-order intermodulation measurements of test filters in a first step. By employing these constants from different systems, i.e., different metallization ratios, in nonlinear periodic P-matrix simulations, a direct comparison to nonlinear periodic FEM-simulations yields scaling factors for the materials used. Thus, the contribution of the different materials to the nonlinear behavior of TC-SAW devices is obtained and the role of metal electrodes, substrate, and TC film are discussed in detail.

A Method for Accurate Modeling of BAW Filters at High Power Levels.

A novel approach for multiphysics modeling of bulk acoustic wave (BAW) filters is presented allowing accurate and at the same time efficient modeling of BAW filters at high power levels. The approach takes the different types of losses and their spatial distribution into account in order to provide the required input for thermal simulation. The temperature distribution determined by thermal simulation is used to modify the geometry and the layer stack of each single resonator of the filter. In this way, the required input for modeling of electromagnetic (EM) and acoustic behavior at high power level is generated. The high accuracy of the modeling approach is verified by the measurements of the S-parameters and the temperature distribution by infrared thermography during high-power loads. Moreover, the influence of the nonlinear behavior on the frequency shift of the resonance frequency is investigated. For this purpose, a parameterized nonlinear Mason model has been combined with a 3-D EM finite-element method and the required nonlinear material parameters were determined by fitting simulation results to the measured polyharmonic distortion model (X-parameters) of a BAW resonator.

Effect of spatial distribution of dissipated power on modeling of SMR BAW resonators at high power levels.

The modeling of bulk acoustic wave resonators at elevated power levels has been improved by taking the spatial distribution of the dominating loss mechanisms into account. The spatial distribution of the dissipated power enables more accurate modeling of the temperature increase caused by the applied power. Thus, it is also possible to more accurately model the frequency shifts of the resonators' impedance curves resulting from the temperature increase caused by the applied power. Simulation and measurement results for the temperatures and impedances of the resonators with different layerstacks at high power loads are presented. The simulation and measurement results are in good agreement, confirming the presented modeling approach. Furthermore, the de-embedding procedure used to obtain vectorial scattering parameters of the resonators during high power loads, the according measurement setup, and the procedure for measuring absolute temperatures by infrared thermography are discussed.