A Review of MEMS Accelerometers Working Principles Sorush Salahshour Torshizi Institute for Microsensors, -actuators and –systems (IMSAS) Bremen, Germany [email protected] Abstract—This paper aims to provide a review of MEMS-Accelerometers (accs.) different operating principles. At first variety of acceleration sensing and their basic principles as well as a brief overview of their fabrication mechanism will be discussed and lastly the paper will be focused on most commercialized and well-known accelerometer technique, namely, capacitive. Moreover, a comparison table of their performance based on acceleration sensor characteristics such as dynamic range, sensitivity, resolution and linearity will be depicted. finally, an evaluation of the different sensing techniques of MEMS-Accelerometer as well as the conclusion wraps-up the paper. I. INTRODUCTION Acceleration sensors are playing a vital role in micromachined technology, moreover, the demand for new and high-performance accelerometers is increasing daily. The first industry which took the benefits of MEMS-Accelerometers was the automobile industry in 2000 by utilizing MEMS-Accs as for car suspension systems and controllability and in the same way for safety systems such as airbags system 1. Nowadays the application scope of accelerometers covered almost every aspect of engineering science. MEMS-Accs compare to conventional accelerometers have advantages of extremely small size and ability to be mass produced and importantly lower manufacturing costs 1. Consequently, the application spectrum of these acceleration sensors is not confined to car industry while they have opened up their way in multitude branch of science. For instance, nowadays in aviation and aerospace industry and after the emerging the modern technology autonomous unmanned aerial vehicles (UAVs) the demand for highly sensitive and low-cost accelerometers increased sharply 2. Moreover, MEMS accelerometers are now the crucial part of space crafts and rockets navigation systems. Moreover, they are now an inseparable part of smart devices navigation and tracking systems. Similarly, in Bio-engineering where the size of the sensor is highly under magnifier for researchers, MEMS accelerometers are used for health monitoring with help of implanting sensors inside the body 3. Based on aforementioned applications different technology and principle has been used up to now for their fabrication and operating method, the vast majority of application employed capacitive and piezoresistive accelerometer as their transduction mechanism and fabrication is easier to utilize, but there are more different working principles which will be discussed in the next section of this paper. II. OPERATING PRINCIPLES As for every accelerometer, the basic working principle is based on a fixed local inertial frame, beam, and of course the proof mass. When an external forces apply the mass will be displaced with respect to the local inertial frame, the source of this force could be constant gravity force which is called static force or it could be caused by shock or movement which can be named as dynamic forces 4. With reference to definition of sensor, acceleration sensor should convert mechanical motion which has deflected the proof mass, into readable computer signal, for this reason there are several transduction mechanisms which some of them are more relevant such as Capacitive or Piezoresistive accelerometers and also some other mechanisms like Optical, Piezoelectric, Thermal and Tunneling, Piezoelectric, Electromagnetic, Surface Acoustic Wave (SAW) accelerometers. Due to content restriction and less practical applications compare to other mechanisms, in this paper all of the above-mentioned principles except Electromagnetic and SAW will be discussed. A. Optical Accelerometers The working principle of optical accelerometers lies in characteristics of a beam of light. compare to wellknown capacitive based accelerometers, optical-Accs exhibit better sensitivity and resolution as well as higher thermal stability which make them applicable in hazardous environments. Optical accelerometers instead of measuring the displacement of proof mass measure the variation of light wave characteristics like measuring the stress distribution among the proof mass when it is deflected (Photoelastic effect) or determining the effect of different forces and mass displacement on optical signal phase (Phase modulation). Phase modulation is normally used when the higher dynamic range is required. The other methods are Intensity modulation which is simple for fabrication but highly dependent on high-quality light sources compare to Wavelength modulation which is completely independent of light source deviation and is highly accurate and sensitive. The outstanding advantage of Optical-Accs is their immunity against electromagnetic interference(EMI) 5. Figure. 1. Optical wavelength modulation based Accelerometer The figure.1 shows the Wavelength modulation based Optical-Acc sensor by which the light goes through the photonic crystal (PhC) and then enter the photodetector for measuring the acceleration, when an external forces applied to proof mass it will move on its (y) axes which will cause a change of output wavelength. Consequently, the magnitude and direction of acceleration would be measured base on the wavelength difference occurred. B. Thermal Accelerometer Thermal accelerometers compare to other aforementioned techniques do not employ proof mass for sensing acceleration, they utilize thermal convection phenomenon. Thermal-Accs generally consist of silicon etched SNx heater with two temperature sensor on both sides of it, inside the thermal isolated encapsulated cavity. The heater reduces the density of its surrounded air(liquid) therefore when there is no acceleration two temperature sensor will sense the same temperature figure.2(A). By applying acceleration dense bubble will move within the direction of applied acceleration which will cause an asymmetric temperature profile for detectors figure.2(B), consequently, this temperature difference will be detected and amplified for converting into a digital signal by the principle of Wheatstone bridge. 6 Figure 2 Heat Accs,(a) rest mode (b) acceleration applied The fabrication process of this accelerometer is simple which means lower manufacturing cost compared to other mechanisms. Since there is no proof mass, the thermal accelerometer has extremely good shock resistance and compare to capacitive sensors it has more sensitivity, on the other hand, the dynamic range is confined and low-frequency range makes it not suitable for instant shocks measurements or fall sensing. 6 C. Tunnel Accelerometer Tunneling-Accs typically consist of metal tip connected to a proof mass which has few Nanometer distance to a counter electrode and the working principle lies in the quantum electron tunnelling. In order to activate the sensor small bias voltage (around 100mV) is needed to be applied, this voltage consequently create a small current between the metal coated tip and counter-electrode. 7 Figure 3 Simple Schematic of Tunnel Accelerometer When an acceleration applied the movement of proof mass will cause the sub-angstroms displacement of the tip which causes the change in tunnel current. The aim of this method is to keep the tunnel current (1nA) constant over the time, therefore, feedback forces have applied to bring the mass back to its rest position, as a result, the magnitude of acceleration could be measured by closed-loop detector circuit and with help of variation of deflection voltage. 7 The design and fabrication of Tunnel-Accs vary since the time they introduced, Cantilevered, Lateral and Bulk-micromachined are some of them. 7 Tunneling accelerometers have low drive voltage supported by wide frequency bandwidth as well as higher sensitivity compared to capacitive. On the other hand, with reference to the Nanoscale gap, they have complicated fabrication process and higher production costs. D. Piezoelectric Accelerometer These kinds of accs. take the benefit of the inherent piezoelectric effect of materials. A piezoelectric acc. as shown in the figure.4 Usually, consists of a piezoelectric material which is typically thin ZnO or PZT which is sandwiched by two electrodes and deposited over silicon cantilever beam. 8 Figure 4 Principle schematic of the piezoelectric The beam is fixed to frame on one side and on the other side there is proof mass. In the presence of acceleration, mass displacement cause deformation of the beam, in the same way, the piezo material experience compression or tensile. The acceleration then could be measured by calculating the potential difference occurred. PZT has higher piezoelectric constant and sensitivity but it could not be integrated or miniaturized, On the other hand, ZnO has lower sensitivity but easy to integrate, additionally, new fabrication technology compatibility and its sensitivity could be improved by miniaturization. Overall, piezoelectric-Accs. Has high sensitivity and compare to capacitive, lower power consumption and lower temperature dependence as well as higher bandwidth. 8 E. Piezoresistive Accelerometer The first MEMS accelerometer was piezoresistive and was developed back in 19795. It took twenty years until the first MEMS accelerometer commercialized in the market by a car company for their safety systems. The backbone of this method is based on resistivity variation of a material under the stress. Early designs of piezoresistive-acc has a beam that holds the proof mass and supported by a fixed frame 9, moreover, piezoresistors were located on the special spot of the beam where the maximum deformation and stress happens (usually edges) and the readout circuit of them is based on Wheatstone bridge principle, figure.5(b). acceleration and displacement of proof mass will cause beam deformation and consequently, the resistivity of piezoresistors will change, resistance variation will end up changes in the output voltage. piezoresistive accelerometers are highly reliable and simple to fabricate but the integration is not simple. Figure 5 three axes piezoresistive accelerometer (a) model view (b) equivalent Wheatstone bridge model Up to know almost all of the papers are focused on improving the performance and sensitivity by modifying the geometric design and sensing mechanism or by utilizing different fabrication technologies. For instance, adding multiple beams instead of one flexure, figure.5(a) or using asymmetrically gaped cantilever or ion etching the resistors on beam instead of thermal diffusion. In some papers, the lateral movement of mass has also fabricated. Additionally, the length of flexure also matters, the longer the flexure will cause, the lower resonant frequency and therefore lower bandwidth. 9 Typically, in order to protect the sensor from high G or instant shock, the upper and lower part of the sensor is covered by Glass in almost most of the fabrications. F. Capacitive Accelerometer Capacitive-acc is the among most famous accelerometers in MEMS sensors, ADXL series is one the most successful accelerometers in MEMS market. 10 Their working principle is based on capacitance variation. The proof mass is located in a way that has a narrow gap with fix conductive electrodes, the displacement of the mass, therefore, will cause a change in distance between mass and electrodes, therefore, the capacitance will be varied. This variation then could be transferred to a digital signal with readout circuit. Capacitive-accs structures could be divided into lateral, vertical or see-saw. 1 Lateral accelerometers usually consist of surface micromachined fix fingers as well as a mass which shaped with sensing fingers, sensing inplane acceleration in x-y axis, vertical structures are usually bulk micromachined and has bigger mass which is located between two fix electrodes, thus they have better sensitivity and out of plane sensing in the z-axis. The see-saw accelerometers make use of torsional beams to suspend the mass and making one side of structure heavier, hence, same as verticals the have out of plane sensing mechanism. The outstanding advantages of capacitive-accs are high sensitivity and DC response, simple and easy to mass produce structure, high linearity and low power dissipation and easy to integrate. The only drawback of capacitive accelerometers is that they are at the mercy of electromagnetic interferences(EMI) which require special packaging. Figure.6 Is an integrated 3-axis accelerometer with two in-plane structure for x and z-axis and one out of plane structure for z-axis which they connected to each other by polysilicon connectors. The fabrication of both of them is depicted in figure.7. Figure 6 Three-axis single-chip micro-g accelerometer Figure 7 Fabrication process. (a) Boron doping; (b) DRIE trench; (c) oxide,nitride, poly deposition; (d) pattern oxide, nitride, poly; (e) electroplate metal;(f) anisotropic etching; (g) HF release. III. COMPARISION There are a variety of sensors in all of aforementioned transduction mechanisms in different dynamic range (from micro-g ranged to hundred kilos) and sensitivity as well as DC response or linearity. Therefore, it is not an easy task to compare this mechanism. For the sake of comparison six sensors which have close dynamic range has been selected and their performance has been written on the table1. Range Sensitivity Resolution Non.lin Optical + 22 3.1816 nm g?1 n/a Linear Thermal +10 375 mV/g 30 mg Linear Tunnel -20-+10 133 mV/g 22.8 mg 0.6% Piezo.res 50 3 mV/g 0.20 the mg <1% Piezo.e + 25 0.21 mV/g 0.22 mg <2% Capaci. + 27 0.5 mV/g n/a Linear Table 1 MEMS Accelerometers Comparison Table IV. CONCLUSION In this paper, six different transduction mechanism of the accelerometer, as well as their specifications with brief fabrication techniques descriptions, have been discussed. It worth to be noted that basically, each type of accelerometers has some pros and cons which make them applicable for the special and different application. For instance, EMI and temperature immunity of optical sensors for a special application which has electromagnetic fields or high sensitivity of Tunneling sensors for rocket navigation and good shock resistivity of Thermal sensors in high-g applications, simple fabrication of piezoresistive sensors for mass production and also low power consumption of piezoelectric sensors. The thing that makes the capacitive sensors different and put them in the first place in the market is that they have most of the above-mentioned advantages which make them applicable for wide variety spectrum of different applications.