What Are MEMS Devices?

Microelectromechanical systems (MEMS) are micrometer-sized systems that are a cross between an electronic and a mechanical device. In MEMS devices, an electrical signal is inputted into the device and the output is a mechanical response, and vice versa (a mechanical input will have an electrical output). However, MEMS always need to have a mechanical function, even if the mechanical structures don’t physically move. For this reason, even though they contain advanced electronics, they are often referred to as a mechanical system.

MEMS contain many miniaturized electronic elements and mechanical structures, including actuators, microsensors, cantilevers, micromirrors, membranes, small channels, switches, cavities, and microelectronic integrated circuits (ICs) that act as the “brain” and control center of the MEMS. Typically, a silicon substrate will be used to form the IC, and the other microsystem components will be added on top.

MEMS technology has been around for a number of years now, and as modern-day technology continues to get smaller, they are seen as the future of electronics. This is because MEMS fabrication is based on existing semiconductor microfabrication techniques used today ― such as surface micromachining, photolithography, and dry etching.

While it’s now a seasoned technology, MEMS didn’t see much commercial use until a MEMS-based accelerometer was used by Nintendo in 2006 in their Wii remotes. Since then, the popularity and the use of MEMS devices has expanded across many applications and industry sectors. With this expansion and market penetration, there are now many types of MEMS devices, driven by the ability to integrate and combine many small parts and components ― with different electro- and mechanical properties ― to build unique micron-sized systems with a high performance.

Many MEMS devices are used in a sensing, actuating, or resonating capacity and take advantage of advanced semiconductor fabrication techniques to build highly accurate, small, and low-weight devices that have a low power consumption. 

A lot of MEMS devices fall into sensor and actuator applications. The main difference between the two is that sensors convert a non-electrical signal (such as a mechanical signal) into an electrical output, whereas actuators take an electrical signal and convert it into a mechanical motion.

Many MEMS components can be fitted onto a silicon wafer, and engineers now have access to micron-sized devices where sensors can be co-located with other electronic signal conditioning electronics to build systems that are more akin to a transducer, rather than just a “MEMS sensor.”

MEMS devices typically fall into four main categories: capacitive, gyroscopic, piezoelectric, and laser-based MEMS. Many MEMS devices often fall into one or more of these categories, making it difficult to isolate devices into a single category. However, the main categories are:

1. Capacitive: Capacitive MEMS are used in conductive applications, and the internal elements inside the MEMS detect a change in capacitance. 
2. Gyroscopic: Gyroscopic MEMS measure the angular rate of a system by comparing the inertial force on an object against a reference. 
3. Piezoelectric: These MEMS produce an electrical current using the piezoelectric effect (redistribution of charges in material lattice) when the device undergoes mechanical deformation. 
4. Laser-based: Laser-based MEMS are used to tune lasers by varying their output wavelength to a desired size/region of the electromagnetic spectrum. Can be used to tune different types of laser for different applications, ranging from acousto-optic filters to optical communications and automotive lighting.

A number of MEMS sensors fall into the inertial measurement unit (IMU) category where the mechanical response is translated into an electrical output. IMUs include gyroscopes ― that are used in airbag deployment, virtual reality headsets, drone navigation, and mapping systems ― and accelerometers for video game consoles, camera, and aircraft attitude control systems applications.

Some common actuators include digital light processing (DLP) chips, speakers, micropumps, rotary micromotors, tweezers, printers, microgears, microvalves, micromirrors and switches, Switches is a key actuator area and requires understanding the “pull-in” voltage and the hysteresis between the pull in and the release voltage to optimize the design of very small switches.

Another MEMS-based sensor is the haptic sensor, which contains electroactive tapes that bubble up and send out an electrical signal when pressed, or through the use of magnetic effects and electro-active fluids ― and include applications such as touchscreens and fingerprint sensors. Other MEMS sensors include gas sensors and strain sensors.

MEMS oscillators are another key device architecture. MEMS oscillators contain a resonator that uses an analog driver to generate a piezoelectric excitation. MEMS oscillators produce stable frequencies ranging from 1 hertz (Hz) to hundreds of megahertz (MHz).

Radio frequency (RF) filters are another fundamental MEMS device and are currently one of the largest markets for MEMS technology. In this case, the mechanical output creates a filter that has a small size, low cost, and the ability to perform many filtering functions ― including wideband, narrowband, low-pass, and high-pass filtering. In the RF filter space, MEMS can be used to build both surface acoustic wave (SAW) and bulk acoustic wave (BAW) filters.

With so many classes of MEMS devices, there are many applications and industries ― such as automotive, aerospace, defense and health care ― where MEMS are making an impact today. MEMS sensors, for example, are used to detect a wide range of stimuli in different industries, including acoustics, fluid flow, temperature, pressure, vacuum level for semiconductor fabrication machines, inertial effects, magnetic fields, chemicals, and radiation.

Some common examples of MEMS sensor devices include infrared detectors, magnetometers, temperature sensors, and pressure sensors. MEMS accelerometers, gyroscopes, and other inertial sensors are widely used in the aerospace sector where everything is moving at fast speeds and sensing operations require the utmost precision.

MEMS can also be used in small-scale energy harvesting applications for powering medical and health-monitoring wearables and implantable medical devices (IMDs) ― in a sub-area known as bioMEMS ― as well as for powering other small-scale portable electronics. In the portable and consumer electronics space, MEMS are used in smartphones as both an RF filter and as a haptic sensor for the touchscreen display. Other RF filters ― either SAW or BAW ― are currently used in Wi-Fi, Bluetooth, and long-term evolution (LTE) applications.

Beyond the more conventional applications, MEMS exist in many specialist areas, including in the sensors used in self-driving cars, airbag deployment, and automation applications; micromirror arrays for high-definition projectors; inkjet printer heads; micro heat exchangers; optical switches and photonic devices for low-loss communications; and microfluidic devices.

The design and manufacturing process for MEMS can involve numerous challenges due to their small scale and sensitivity — which makes them susceptible to any motion or shock, as it can cause a false signal. There are also thermal and off-access compensation that need to be added to the device and accounted for. The challenge with designing MEMS is that they are small, and the geometries are complex, but the movement of the mechanical parts is orders of magnitude smaller. So advanced simulation capabilities are required to look at both the structural and operational aspects of MEMS and ensure that the design is robust enough to cope with the natural variability that exists within manufacturing processes.

Everything in a MEMS device is driven by sensitivity and quality factor, which is a measure of the energy loss. However, MEMS devices can have very high frequencies that need to be accounted for: for inertial sensors it’s hundreds of kilohertz (KHz) to MHz, and it’s in the gigahertz (GHz) region for RF filters. Filters are a step function so the accuracy of the predicted coupled displacement and voltage fields is important, and the accuracy drives the slope of a filter curve, which goes from 0 to infinity. Filters need to have a steep response for them to be effective filters, so very precise tools are required to accurately evaluate that curve sharpness and its sensitivity to temperature variation.

For many MEMS devices, designing and optimizing the size and materials used in the mechanical components is one of the most important aspects of the design process. The optimal structure can be designed by looking at the inputs and seeing how the signal travels between two points in the device, and what the resulting output is. If the resulting output is not right for the inputs given, the design space is not optimal. These aspects can all be analyzed and solved by using advanced simulation software to design high-performance MEMS devices.

Simulation tools need to be able to build complex designs and require a high accuracy. The tools from Ansys have a picometer-level resolution, so not only can these simulation tools be used for MEMS, but they can also be used for their smaller nanotechnology counterparts, nanoelectromechanical systems (NEMS). At the end of day, simulating NEMS is just like zooming the design to fit a smaller scale, and the picometer resolution provides this capability.

For designing and simulating the performance of MEMS, Ansys uses two software packages, called Discovery and Mechanical. Ansys Discovery is used for pre-processing and then Ansys Mechanical is used for the simulation itself. Discovery can be used to lay out different geometries and process induced variations in MEMS, such as investigating the etching process and relating to critical dimensions that might be measured during manufacturing. The Discovery package can pre-process most all the different variations in the geometry prior to simulation in Mechanical, where those geometries and any other features can be scaled on a wafer-scale simulation. For more detailed and unique geometric variations, the nodes representing the geometry in Mechanical can be moved using various automated approaches.

Using dedicated pre-processing and wafer-level simulation methods with picometer resolutions help to not only speed up the design process, but they also ensure that the design is accurate and that it has the required specification to deliver a MEMS device with a high performance for its intended application. To find out more about how the simulation tools at Ansys can improve the design process of your MEMS devices, contact our technical team today.

Related Resources