Image source: Nanusens
Sensors are fundamental elements of the IoT and the smaller, cheaper, more rugged and more sensitive we make them, the more useful they become for applications such as smartphones, toys and wearables. However, at the microscopic level, the benefits of smaller size are countered by strange forces such as Van Der Waals and Casimir, which make Nanusens’ breakthrough in inertial sensing so interesting for developers of miniature IoT solutions.
MEMS, or microelectromechanical systems, are the technology of choice for sensing movement. They got their start in airbags, but they really began taking off with the introduction of the Nintendo Wii remote control, and then smartphones. MEMS are as they sound: miniature mechanical systems. They have electrical characteristics such that when two mechanical objects move relative to each other, an electrical characteristic, such as capacitance, changes and this change is sensed and its value helps determine the degree of mechanical movement.
The mechanical movement can be caused by linear acceleration, vibration, shock, tilt or rotation of the MEMS structure as the object in which it is placed moves.
Figure 1. MEMS sensors can be used to sense movement caused by a device as it accelerates, decelerates, vibrates or is dropped, tilted or rotated. (Image source: Analog Devices Inc.)
As useful as MEMS devices have become, there is always a desire to make them smaller and lower cost. The efforts to make them smaller have been hampered by stiction, caused by Casimir and Van Der Waals effects, as dimensions drop to the nanoscale.
Lower costs have been hampered by the need to manufacture them using separate processes to the CMOS processes used by semiconductors, thus preventing them from reaping the benefits of high-volume manufacturing and integration with the semiconductors upon which they depend to make use of their outputs.
Small Sensors, Big Opportunity
Nanusens has done some fundamental materials research and has come up with a way to reduce the effects of stiction while making electromechanical systems compatible with CMOS manufacturing processes.
Van der Waals and Casimir effects in MEMS are attractive forces that occur on microscopic levels that cause static friction, or stiction. In stiction, two microelectromechanical elements stick to each other when the moving mass moves out of its typical range of motion relative to the non-moving mass and they touch. This can happen if it strikes a surface too hard.
These stiction forces are surface-area dependent, not mass dependent. They can be countered by having stronger springs on the moving mass, but this reduces the sensitivity of the sensor. The sensitivity could be increased by increasing the mass, but this results in a greater surface area for the mass and so, unfortunately, more attractive forces.
To break this logjam, Nanusens reduced the sensor design size by an order of magnitude from MEMS, with linear feature sizes of 1-2 µm, to nanoelectromechanical systems (NEMS) where the features are 0.3 µm. This reduces the attractive forces significantly as the surface area reduction is in two dimensions, thereby giving almost two orders of magnitude reduction.
Reducing the proof mass (the mass that moves when the device accelerates in any direction) could result in decreased sensitivity, except this is offset by reducing the gap between the proof mass and the fixed electrode. The size reduction also means that the energy stored on the proof mass when it hits the surface in case of a shock, is much less and the traveling gap is also small. A shock with less energy is also easier to detach.
The new approach actually increases reliability to the point that sensitivity can also be increased, as there is less risk of stiction occurring. The kicker is that the nano-sensors can be made using standard CMOS processes and mask techniques, so the sensors can be directly integrated with active circuitry. In addition, the sensors can potentially have yields (percentage of devices within a preset quality and performance range) similar to CMOS devices.
To manufacture the sensors, Nanusens has partnered with Global Foundries (GF), which is using its 0.18-µm process for the project, along with its ability to scale quickly to high volume, assuming the technology takes off.
Right now, it’s very much in the development stages and may not be available for production-level IoT designs for some time. However, the technology looks promising. First silicon samples show strong resilience to stiction, going through 10,000 switching cycles, each equivalent to 1000 g shocks, with what it claims to be an order of magnitude increase in sensitivity (relative to MEMS).
For developers of IoT solutions, it’s best to continue with currently available devices to implement inertial sensing ideas and designs. Nanusens hasn’t made it clear when devices will be ready for prototyping and development. However, the key takeaways here are to be aware of the limitations of MEMS devices, while also being aware that those limitations are being overcome, opening up new applications with greater sensitivity, lower power consumption, smaller size and lower cost.