Assessment of dielectric charging in electrostatically driven MEMS devices: A comparison of available characterization techniques

Abstract

The present work investigates the results of different characterization methods for the dielectric charging phenomenon applicable to metal–insulator–metal (MIM) capacitors and electrostatically actuated micro-electro-mechanical-systems (MEMS). The discharge current transients (DCT), thermally stimulated depolarization current (TSDC) and Kelvin probe force microscopy (KPFM) assessment methods have been applied to either MIM capacitors or electrostatic capacitive MEMS switches or both. For the first time, the KPFM methodology has been used to create a link between the results obtained from the DCT and TSDC techniques applicable for MIM and the results from MEMS switches. The comparison shows that the application of KPFM method to MIM and MEMS leads to the same results on the electrical properties of the dielectric material. This provides a novel powerful tool for the assessment of dielectric charging for MEMS switches using MIM capacitors which have much simpler layer structure. On the other hand the TSDC method reveals a continuous distribution of relaxation time constants, which supports the dependence of relaxation time constant calculated for MEMS on the duration of the observation time window.

Introduction

The increasing demand for more functional and flexible, yet lightweight and low-power-consumption wireless systems, has generated the need for a technology that can dramatically reduce manufacturing cost, size, and weight, and improve performance. Since MEMS technology enables batch fabrication of miniature mechanical transducers integrated with complementary metal–oxide semiconductor (CMOS) circuits, MEMS for radio frequency applications (RF-MEMS) provides an opportunity to meet these requirements. The term “RF-MEMS” encompasses several distinct types of devices, including RF switches, resonators, varactors and tunable inductors. Compared to conventional RF components, RF-MEMS offer significant benefits, including lower power consumption, lower insertion loss, lower cost and smaller form factor. Due to ease of fabrication and integration issues, electrostatically actuated MEMS devices are of high interest. However, such devices suffer from reliability issues related to dielectric charging. This phenomenon is found to be more pronounced in the case of electrostatic capacitive MEMS switches and currently results in hindering the commercialization of these devices [1].

The charging/polarization of a dielectric material arises from dipolar and space charge polarizations that often coexist and the electric field and polarization must then be considered as averaged over the thickness of the sample [2], [3]. This implies that the resulting charge displacement and microscopic dipoles orientation will give rise to a macroscopic dipole moment, which after removing the stressing voltage will decay and lead to discharge current transients (DCT) [4], [5], [6] or thermally stimulated depolarization currents (TSDC) [7] measured in external circuit of metal–insulator–metal (MIM) capacitors. In capacitive MEMS switches although the charging process is similar to the one in MIM capacitors the relaxation differs significantly since the injected charges are collected only through the bottom electrode when the applied voltage is removed and the suspended electrode is in the pull-up state. In MEMS the dielectric charging has been monitored through the shift of pull-down and pull-up voltages [8], [9] as well as voltage at which the pull-up capacitance becomes minimum [1]. Recently, Kelvin probe force microscopy (KPFM) has proved to constitute an efficient method to simulate charging through asperities [10], [11], [12], [13], [14] and assess the discharge process in MEMS dielectric by monitoring the film surface potential [13], [15].

In spite of these efforts there has been no correlation between the abovementioned assessment methods. Thus, the methods associated to MIM capacitors have been considered as efficient tools to assess the bulk properties of the dielectric film, which provide limited information and therefore being not appropriate to simulate MEMS. The aim of the present work is to reveal the common features and main differences between these assessment methods. In order to succeed this, the decay of surface potential of both MEMS insulating film and MIM capacitor top electrode were monitored and compared. Moreover, KPFM surface potential evolution with time was compared with the decay of DCT and TSDC spectra of MIM capacitors. To our knowledge, this is the first time to introduce a characterization methodology for the dielectric charging phenomenon based on KPFM for MIM capacitors.

The paper is organized as follows. First the theoretical background for the abovementioned assessment methods is presented. This is followed by the experimental details for each applied method. Finally, the results for each method as well as the correlation between the results obtained from different techniques are presented and discussed.