Our lab has a new paper out in collaboration with the Berfield group about a windowpane-shaped microstructure that has two stable shapes. Tom Lucas (who graduated with a Ph.D. from our group in 2014) and Jaz pointed a heat gun at the MEMS devices one day, trying to flatten them by thermal expansion. This led to the discovery that the structure would flip from one to the other by aiming compressed air at it. Thanks to these experiments and Dan Porter’s model, we know that larger devices flip at lower airflow velocities.
The best applications for these bistable microelectromechanical (MEMS) devices are situations where power is limited, because they don’t need power to hold their shape. For this reason, bistable devices are candidates for radiofrequency (RF) switches in mobile phone applications, where it’s important to conserve the battery. The difference between this windowpane device and most other bistable devices is that the windowpane curls out of the plane of the silicon wafer due to residual stress. Usually MEMS makers try to cancel out residual stress, which is the enemy of predictable in-plane motion. Also different: the device in this paper would probably still be bistable if completely removed from the wafer.
The windowpane-shaped device in the paper could be used as a passive marker for peak airflows. Our group has also switched bistable devices using electric currents. With electrical contacts, it could potentially make a low-power electronic switch using significantly less wafer area than a planar device.
We have been dealing with bistable structures across different size scales. A common question is, how can we detect their state electronically. This project uses machine-sewable conductive thread to add an electronic switch to bendable compliant beams in a cm-scale structure. The beam material is 0.125 mm thick plastic film: thick enough to have some “snap” (video here) yet thin enough that a sewing needle easily punches through. In figure (a), silver-plated nylon conductive thread is suspended across a U-shaped cutout that has already had a conductive thread sewn down the middle. When the beam is down, the cutout moves away from the suspended threads, opening the circuit, and when it’s up, the threads are in contact. We can detect the beam position by connecting the switch to a digital input as shown in figure (b), it’s very crisp. In figure (c), there are a few different length beams that snap at different angles (paper here). The embroidered pattern needed to be aligned to these laser-cut beams, and to do that a pair of “+” shaped alignment marks helped line up the needle. Those marks can be seen on the middle beam, except the left one is stitched over so it’s hidden. This flexible design is a big improvement over a previous iteration (video here) where we had soldered on switches that kept flying off.
The upper plot in this video shows the “basins of attraction” for a bistable compressed beam as you bend its support angle from flat (0 degrees) to about 18 degrees. This beam is about 4% too large to fit in its assigned area, so the center pops up or down. From our earlier work we know that beams prefer to pop in the same direction as the supporting substrate, and we have the potential energy function that describes this behavior more quantitatively. It’s the lower plot.
When a beam is dropped into “phase space” (a plot of velocity vs position of the beam center) it will coast to one of the two minimal energy positions. The red spiral shape is the region of phase space corresponding to the higher energy state, and it shrinks as the substrate bends. When the high-energy basin of attraction goes to zero area, the beam snaps to the low-energy state if it wasn’t there already. We are looking at the area of the spiral as a way to measure the curvature of the underlying substrate through the statistics of repeated experiments. This is a “dartboard” style experiment; a smaller target should receive fewer hits than a larger one.
Pulling a wax-filled silicone balloon out of a mold
We had a physics student from Berea College, Fidel Tewolde, in our lab during summer 2013. His project was to create inflatable silicone actuators to drive our bistable beams from one state to another. With a bistable skeleton, you might be able to save a considerable amount of power by shutting off the air pressure when the actuator has flipped! Thanks to a great tutorial on printing your own robot, Fidel was able to get going fast. See the rest of this post for more photos.
This video shows how two curved beams can interact to produce bistability. The green and blue beams are constrained so that their tips stay a fixed distance apart, shown by the red dotted lines. (In practice, the red line would be another beam made out of the same material.) The green and blue beams also have a preferred radius of curvature that minimizes their stored energy. We adjust the preferred radius of the blue beam by changing the temperature. However, the constraint makes it impossible for the green beam to take on the preferred radius of curvature at the same time. Instead, there are two mirror-image energy-minimizing states. In the video, the beams are drawn thicker when in the energy-minimizing state; at the same time, the slider on the energy plot is at a local minimum.