Investigation of parallel-connected MEMS electrostatic energy harvesters for enhancing output power over a wide frequency range

Microelectromechanical-Systems (MEMS) electrostatic energy harvesters, converting vibration energy into electricity, is an active area of research and is envisioned to supply power to the myriad of sensors for Internet of Things applications. However, although a variety of approaches have been investigated to improve either power output or frequency range, a solution able to simultaneously address both is yet to emerge. This paper explores an energy harvesting system comprised of two parallel connected devices of different resonant frequencies, which is demonstrated to exhibit a broadened operating frequency range and enhanced power output as compared with an individual device. While it is straightforward that the frequency operating band is to be broadened due to the difference in natural frequencies, an increase in system's power output is not guaranteed due to intrinsic phase shift exhibited by the two dissimilar devices undergoing oscillation. However, it is shown that the power output by the two parallel connected devices may be significantly enhanced above that of a single device, an effect referred to as power-combining or constructive interference, which appears under operating conditions that enable partial synchronization of the oscillatory motion of each device. Such an approach is especially promising since it is possible to employ multiple parallel connected devices to increase the output power to levels matching those seen in applications. However, this technic is not yet explored theoretically or experimentally, a task undertaken here, where a case study consisting of a system with two parallel coupled devices is presented. System's dynamics is investigated theoretically and experimentally to understand conditions enabling power-combining. A modular theoretical model, which can be readily extended to any number of devices connected in parallel, is developed with co-dependent, simultaneous mechanical and electrical simulations carried out in VerilogA and Cadence Spectra.