Interdigitated-electrode-based mems-scale piezoelectric energy harvester modeling and optimization using finite element method

Abstract

This paper presents a novel optimization method for interdigitated electrode (IDE)-based, cantilever-type piezoelectric energy harvesters at microelectromechanical system (MEMS) scale. A new two-stage approach based on the finite element method is proposed to examine the performance of such devices. First, detailed electrostatic poling simulations are presented. The results of these poling orientation simulations are used while calculating electrical energy and conversion efficiency in response to a constant external force. The proposed approach is used to find the optimum piezoelectric material thickness and IDE geometry for a cantilever beam which is constructed on top of a 4-μm Si structural layer and a 1-μm SiO2 isolation layer. Cantilever and IDE lengths are fixed at 320 μm and 240 μm, respectively, whereas the lead zirconate titanate (PZT) thickness, IDE finger widths, and number of finger pairs are varied. Maximum output energy of 0.37 pJ for a 15-μN force is obtained at a PZT thickness of 0.6 μm and an IDE consisting of 12 finger pairs. This energy is reduced to 1.5 fJ for 5 μm PZT thickness with 2 electrode finger pairs, which shows that device geometry has a significant impact on device performance. The proposed method presents an accurate framework for the rapid design and performance prediction of novel piezoelectric energy harvester structures.

A combination of experimental and numerical method for optimizing the sensitivity of ultra-thin piezoelectric sensor with interdigitated electrodes

“…A similar non-linear relationship between the piezoelectric layer thickness, the electrode width and the poling orientation has been previously observed by e.g. Toprak and Tigli [9].…”

“…However, this may be challenging because of the time and computational power required in dividing the piezoelectric material model into small elements, assigning each of these with a local poling direction/magnitude and finally solving the complex model. Toprak and Tigli [9] used similar two step approach; however, they simplified the latter step by using an average poling orientation/magnitude on top of the electrode instead of element specific local orientation/magnitude.…”

“…perpendicular to the electrode plane. Furthermore, the magnitude of the poling varies significantly in the case of IDE, whereas for MIM it is homogenous throughout the material: electrostatic FE-modelling [6][7][8][9][10] has shown that the IDE-based devices have an inactive area close to the centerline of the electrode where the electric field magnitude, and therefore also the poling magnitude, is zero. Thus, to fully understand the factors affecting the sensitivity of the IDE based device, one would have to solve for the net effect generated by the various magnitudes and directions of poling throughout the material.…”

Bio-inspired directional sensor with piezoelectric microfiber and helical electrodes

“…As shown in Figure 3(a), the induced polarization field after poling is non-uniform and can only penetrate to a limited depth since the electrodes are on the surface. The poling effect with IDEs has been studied previously (Beckert and Kreher, 2003; Bowen et al, 2006; Hareesh et al, 2012; Kim et al, 2012; Toprak and Tigli, 2013). Poling will result in three different types of ferroelectric zones in the fiber: (1) the active zones between a pair of electrodes, where the electric field can be considered to be uniform and parallel to the axis of the fiber; (2) the transition zones below the corners of the electrodes; and (3) the dead zones that are below the electrode surfaces and where the electric field cannot penetrate.…”

“…Ideally, to simulate the piezoelectric response of the sensor, the initial polarization directions in the fiber should be exactly the same as the result in the electrostatic simulation (Beckert and Kreher, 2003; Toprak and Tigli, 2013). In this study, however, we manually divided the fiber into different subdomains with orthogonal polarization directions as shown in Figure 4(d), neglecting the bending of the polarization fields in fiber: zones ③ and ① are in the 6 x -direction; zones ④ and ② are in the 6 z -direction.…”