editor's blog
Subscribe Now

MEMS Laser Probing

I saw an announcement about a new MEMS diagnostic instrument, the M150 from Ardic. It uses a laser to measure the frequency response of a MEMS element. Sounds simple enough. Or it did until I started thinking more about it. After all, lasers require line of sight. And most MEMS elements are far out of sight. And if you bring them into sight, then you may have changed the environment (air instead of vacuum or controlled gas, microphone packaging and cavities, etc.) So how does that affect the measurement?

I had an interesting chat with founder and CEO Edward Chyau to get some perspective on the tool. First, the main use cases: this is typically used either to verify a design early on, for occasional production monitoring, or for diagnosing failed units. So it’s not a high-volume in-line production system. It’s set up to handle dice, although they have had requests for a unit that could handle 6” wafers. (In theory, on a wafer, neighboring dice could interact since they’re mechanically connected, but he said that the actual element movements are so small that by the time their vibrations reached the neighboring die, it would have attenuated to a negligible level.)

And yes, the moving element must be visible. An InvenSense gyro, for example, would need to have its capping ASIC removed to take a measurement. So there will be some difference between the elements behavior under the laser and its behavior in the wild. He says that the tests being run look for significant resonances within the region of interest, indicating that some mechanical element has a resonant mode, due to a design error or a production defect, that can interfere with normal operation. Those modes might shift frequency a little due to depackaging, but not a lot.

So, for example, if a microphone diaphragm exhibited a resonance at 1 MHz, then it wouldn’t be an issue. But if your window was up to 50 kHz and it had a 55 kHz mode, you’d want to pay attention even though it’s nominally outside the window.

It might be possible to anticipate any such shifts at design time. If design simulations were done assuming a vacuum, for example, then it would be possible to redo the simulations without a vacuum to establish a baseline expected behavior for laser probing. That could help set expectations for any change in behavior for correlation if that were a concern.

I also wondered if the laser itself would heat the element during the measurement, changing the performance as it heated. He said that the laser is low-power, like that on a CD-ROM drive, so no noticeable heating should occur during the 60-second scan.

That scan is part of what differentiates this unit from existing devices. Current technology uses laser Doppler interferometry, and it’s largely targeted at the academic environment with relatively high cost and a long learning curve. And a measurement can take 1-2 hours, with reasonably fussy setup.

By contrast, the Ardic system uses astigmatic detection, also used in their atomic-force microscope (AFM) tools. This gives them particularly good z-axis resolution, and they can scan from 1 Hz to 4.2 MHz in 60 seconds. Users often simply run that entire range even if they only need a subset just because it goes so fast.

I also asked whether this technique could apply down to the NEMS range, where quantum effects start to matter. And, in fact, here you’re dealing with features below the visible wavelength. So getting a laser focused that small and capturing the echo isn’t really feasible.

He said that you would likely need to take a more indirect approach. For example, you might have a MEMS-scale cantilever that acts as an atomic-force probe on the NEMS element. It doesn’t actually make contact (although I suppose that, on an atomic scale, where most of the atom is open space, “contact” is not such a clear concept); it just comes within Van der Waals force range. This allows the cantilever to track the motion of the NEMS element; the laser can then be used to measure the cantilever.

But, of course, the cantilever would have its own resonant modes, so you’d have to measure the cantilever first so that you could subtract its behavior out of the combined signal to isolate the NEMS element signal. This hasn’t been done yet, but he saw it as the kind of approach that would be required.

You can read more in their announcement.

Leave a Reply

featured blogs
Aug 18, 2018
Once upon a time, the Santa Clara Valley was called the Valley of Heart'€™s Delight; the main industry was growing prunes; and there were orchards filled with apricot and cherry trees all over the place. Then in 1955, a future Nobel Prize winner named William Shockley moved...
Aug 17, 2018
Samtec’s growing portfolio of high-performance Silicon-to-Silicon'„¢ Applications Solutions answer the design challenges of routing 56 Gbps signals through a system. However, finding the ideal solution in a single-click probably is an obstacle. Samtec last updated the...
Aug 17, 2018
If you read my post Who Put the Silicon in Silicon Valley? then you know my conclusion: Let's go with Shockley. He invented the transistor, came here, hired a bunch of young PhDs, and sent them out (by accident, not design) to create the companies, that created the compa...
Aug 16, 2018
All of the little details were squared up when the check-plots came out for "final" review. Those same preliminary files were shared with the fab and assembly units and, of course, the vendors have c...
Jul 30, 2018
As discussed in part 1 of this blog post, each instance of an Achronix Speedcore eFPGA in your ASIC or SoC design must be configured after the system powers up because Speedcore eFPGAs employ nonvolatile SRAM technology to store its configuration bits. The time required to pr...