MEMS (Micro-Electromechanical Systems) is a field that is likely quite unknown to the general public but which has a relatively long history. According to one estimation, the first “MEMS” was built in 1965, less than twenty years after the transistor was invented, but pressure and stretch sensors have been made out of silicon even before that. MEMS became established as the common term only in the 1990s.
I ran into this subject in my twenties, in the late 1980s, when Tieteen Kuvalehti science magazine showed how microscopic, moving wheels and motors could be made out of silicon. When my semiconductor technology studies progressed to the thesis writing stage, I had the chance to design and process micromechanical sensors from polycrystalline silicon at the VTT Technical Research Centre of Finland’s semiconductor laboratory. After I completed my thesis, I continued my career with Murata (VTI Technologies at the time), which has been producing MEMS sensors for car movement measurements since 1991.
We do not talk about micromechanics anymore, as the current term is MEMS. The first letter, M, reveals that the basic idea is still the same: miniaturize, i.e., making various electronic, mechanical or three-dimensional small-scale structures on a wafer, applying the methods used for manufacturing integrated circuits or developed based on them. The most important methods are thin film growth, lithographic patterning, etching and wafer bonding. Silicon is, by far, the most important material. It is a tireless spring material up until its high ultimate strength point, and very accurate patterning methods have been developed for it. Even though many kinds of small-scale devices have been created through MEMS, not all concepts have led to industrial applications. With micromotors, for example, it was observed that the wheels wear down quickly, can easily become fused permanently with the wafer and the power they produce is so insignificant that transferring it from the wafer’s surface to be utilized is not sensible.
Even smaller sensors in the future?
Movement sensors, i.e., inertial sensors, are one of the success stories of MEMS, and Murata is one of the field’s forerunners. Inertial sensors are used to measure acceleration, tilting or rotational motion, and they are made of one or several masses made of silicon and equipped with silicon spring. These masses shift due to an external force targeted at the sensor, and this movement is then measured electronically. At first glance, it does not seem like an inertial sensor is a very popular object to miniaturize, as the power of the mass decreases when the mass grows smaller and the mass then moves less. And this is true, but when the dimensions become smaller, the accuracy of the movement measurements of the mass actually improves. In fact, the electric field between the electrodes changes more drastically the smaller the gap between them is. For example, the gap between the mass and the measurement electrodes in Murata’s acceleration sensor is less than one tenth of a hair width, and the mass movements can be measured with an accuracy of one thousandth of this. The oscillating rotary position sensor where mass is actively moved with electric power can measure movements on a scale even smaller than this. If the millimeters of a MEMS sensor were to be converted into meters, the sensor element would be the size of a dining room table and it would measure movements the size of a speck of dust on the table. The secret here is that the electronic and mechanic powers are more strongly connected to each other in small dimensions than in the macro world, and the two middle letters of MEMS – EM – reflect this: Electro-Mechanical.
But it is not quite that easy in practice. There are actually electric fields between all charged surfaces, and when the devices are miniaturized, these unwanted fields also grow, and easily more than those in the measurement electrodes. However, Murata’s technology offers a significant advantage regarding this problem. Murata has developed a method of combining silicon and glass insulation in a way that allows relatively thick and wide insulated gaps, which means that the unwanted electrical fields are smaller than in competitors’ devices. Thanks to this technology, Murata’s sensors are the most accurate devices on the market for demanding car applications. True to the basic idea of MEMS, their size has grown smaller over the years, even as an increasing number of smaller and more complex spring and mass structures have been added to them, allowing the same sensor to be used for measuring several different directions of movement. The improvement of the silicon wafers’ etching, patterning and bonding technologies has enabled this development.
But where is the limit for miniaturizing an inertial sensor? We have not likely seen it yet. Production technology for making the dimensions even smaller does exist already, as we know by the development of transistor linewidth. Of course, the production equipment, especially for patterning, are constantly growing more expensive, which means that miniaturization will not be financially profitable at some point. It can be assumed that the limit will be reached when other physical phenomena, such as surface powers caused by atoms or unwanted charges, start to disrupt the measurement too much and there are no longer ways to control them. But before this, I’m sure we will see new electronic and system-level solutions that help make electronic movement measurements even more accurate. The silicon structure in itself is not a functional sensor, after all, and it needs to be paired with an integrated electronic circuit and a carefully optimized casing for connecting them and attaching them to the integrated circuit, which minimizes the impact of unwanted external powers and disruptions. Hence, MEMS is not just about silicon and glass: it is a system combining several technologies the design and manufacturing of which requires comprehensive competence. S, the last letter of the term MEMS, reflects this: Systems.
Altti Torkkeli, Fellow, Research & Development