Electrostatic MEMS energy harvesting for sensor powering

Thumbnail Image
Li, Jinglun
Issue Date
Electronic thesis
Mechanical engineering
Research Projects
Organizational Units
Journal Issue
Alternative Title
The growing need for small-scale wireless power sources is driving the development of the MEMS energy harvesters. Among various MEMS energy harvesters, the electrostatic energy harvester is attracting much attention due to its lower power consumption and compatibility with most CMOS technology and fabrication processes. Despite its usefulness and widespread applications, to date, electrostatic MEMS energy harvesters still have low output, insufficient to power even low duty cycle wireless sensors. The objective of this work is to investigate new approaches to address this persisting challenge. The focus is on electrostatic vibrational energy harvesters (e-VEHs) consisting of variable capacitors that convert mechanical vibration to electricity via periodic changes in the capacitance. The output power of such electrostatic harvesters depends on both: i) the ratio of maximum and minimum capacitance, and ii) the variable capacitors’ vibration frequency. This work explores three approaches to increase the power output of electrostatic energy harvesters, investigating i) a device topology responding to out-of-plane excitations, where capacitor’s electrodes can be fully misaligned, leading to almost zero minimum capacitance and increasing capacitance ratio; ii) power combining technology where several selected devices are connected in parallel to produce a higher power output; and iii) devices with impact-based frequency-up-conversion. The first approach was investigated theoretically. A one-dimensional model was developed for predicting dynamic behavior of a e-VEH structure employing an interdigitated variable capacitor vibrating out-of-plane. The variable capacitor is constructed with the help of a plate supported by suspension beams that has electrodes on the edges. The variable capacitor is formed between the first set of electrodes attached to the plate and a second set fixed to a rigid substrate in the same plane with the suspended plate. The main advantages of the out-of-plane e-VEH structure are 1) it carries electrodes on four sides of the plate which can double the output power as compared with a structure moving in plane that has electrodes on two sides only; 2) the capacitance variation can be maximized by completely misaligning the moving and fixed electrodes of the variable capacitors, producing a large increase in output power. Despite its usefulness and potential, the dynamic response of such a structure is hard to predict due to the multi-dimensional deformation of the suspended plate membrane. A simplified model was proposed to address this challenge. The out-of-plane device is modelled as a sectionalized beam compried of the original suspension beam and an equivalent beam representing the plate. The Euler Bernoulli beam equation was solved to determine the electrodes' static displacement, natural frequency, and dynamic displacement. The model results compares with the finite element modeling results. Finally, the steady-state response under dynamic loads was determined by using second order system’s classical theory. The proposed model can potentially be used for mixed domain design optimization for pressure sensors. The second approach investigates the operating condition of a parallel-connected MEMS system. Although it is obvious that parallel systems have a wider bandwidth as compared to a single device, the output voltage may decrease. In this project, a numerical model is developed for the parallel system using Verilog-A language in Cadence spectra environment, which is capable of providing fast and accurate solutions to parallel systems with thousands of dissimilar devices. The condition of power enhancement is also explored for parallel systems. The device selection criteria and the proper operation conditions are discussed in order to ensure a power enhancement within desired operating frequency region. The third approach explores impact-based frequency-up-conversion, a phenomenon previously reported for electrostatic MEMS harvesters, although not fully understood. In this work a gap-closing type e-VEH vibrating in-plane is studied theoretically. The in-plane e-VEH is composed of a central plate supported by the suspension beams on its two parallel sides, and the variable capacitor. The mechanical vibration is converted to electricity by the variable capacitor, which consists of the first set of electrodes attached to the plate and a second set fixed to a rigid substrate. Additionally, the device employs cantilever beams as soft stoppers to absorb a portion of the impact energy as the plate approaches the impact point, and a parylene thin film deposited on the electrode sidewalls to avoid electrical shorting. In such devices, frequency-up conversion is triggered by mechanical impact between the two sets of electrodes and results in a relatively large increase in generated power from low-frequency base vibrations. As mentioned, the origin of this phenomenon is not fully understood, and in the absence of a model design optimization is not possible. Thus, the task here was to develop a model able to predict observed experimental results. The basic model of electrostatic harvesters consists of a point mass representing the plate and the attached electrode, suspended by a spring in parallel to a dashpot (capturing both air and electrostatic damping). To describe the effect of the electrodes’ impact, which deform, and experience damped free vibration following the collision, this model is modified as follows. The mass of the electrodes is separated from that of the shuttle mass and is represented by a second mass attached to the first mass (shuttle) by a second spring. The dashpot is added to this second mass, while the first mass is supported by a spring attached to the frame undergoing vibration. Predictions from this modified lumped model are consistent with the experimental results, and clear frequency-up-conversion effects are observed with exponentially decaying motion signals as seen in experiments. Besides these, parametric studies have been carried out to determine the critical design parameters that control the frequency up-conversion effect based on the validated model. Although such devices employing a titled sidewall and multiple stoppers are shown to significantly improve the output frequency, these devices require a special etch process to make sloped sidewalls. This special process which creates nonuniform electrode width through etching is costly, unstable, and not compatible with the standard fabrication process of MEMS devices. To address this issue, a new design compatible with the standard fabrication process is proposed. In this new device the electrodes are designed to have a vertical wall, however they still exhibit a non-uniform gap by having a variable width. More precisely, the gap along between the electrode now varies along their length. In this way, the new device has trapezoidal prism shaped electrodes where the trapezoid can be seen from the top view. Parametric studies are carried out to understand design trade-offs and to aid in design optimization. A new set of design parameters is proposed following design rules of MEMSCAP to allow for future fabrication and validation of the frequency up-conversion in this novel geometry. Finally, a neural network optimization method is proposed and tested to aid in future optimization efforts.
May 2021
School of Engineering
Full Citation
Rensselaer Polytechnic Institute, Troy, NY
PubMed ID