Timekeeping, or why helium can (temporarily) kill your iphone/ipad

On the day when the US switches clocks back to standard time, here is a post about timekeeping and its impact.  

Conventional computers need a clock, some source of a periodic voltage that tells the microprocessor when to execute logic operations, shift bits in registers, store information in or retrieve information from memory.  

Historically, clocks in computer systems have been based on quartz oscillators or similar devices.  Quartz is an example of a piezoelectric, a material that generates a voltage when strained (or, conversely, deforms when subjected to a properly applied voltage).  Because quartz is a nice material with a well-defined composition, its elastic properties are highly reproducible.  That means that it's possible to carve it into a mechanical resonator (like a tuning fork), and as long as you can control the dimensions well, you will always get very close to the same mechanical resonance frequency.  Pattern electrodes on there, making the quartz into a capacitor, and it's possible to set up an electrical circuit that takes the voltage produced when the quartz is resonantly deforming, amplifies that signal, and feeds it back onto the material, so that the quartz crystal resonator will ring at its natural frequency (just like a microphone pointed at a speaker can lead to a ringing).  Because quartz's elastic and electrical properties depend only weakly on temperature, this can act as a very stable clock, either for a computer like your desktop machine or tablet or smartphone, or in an electric wristwatch.  

In recent years, though, it's become attractive for companies to start replacing quartz clocks with microelectromechanical resonators.  While silicon is not piezoelectric, and so can't be used directly as a substitute for quartz, it does have extremely reproducible elastic properties.  Unlike piezoelectric resonators, though, MEMS resonators typically have to be packaged so that the actual paddle or cantilever or tuning fork is in vacuum.  Gas molecules can damp the resonator, lowering its quality factor and therefore hurting its frequency stability (or possibly damping its motion enough that it just can't function as part of a stable self-resonating circuit).  

The issue that's come up recently (see this neat article) is that too much helium gas in the surrounding air can kill (at least temporarily) iphones and such devices that use these MEMS clocks.  In a helium-rich environment like when filling up superconducting magnets, helium molecules can diffuse through the packaging into the resonator environment.  Whoops.  Assuming the device isn't permanently damaged (I could imagine feedback circuits doing weird things if the damping is way out of whack), the helium has to diffuse out again to resolve the problem.  Neat physics, and something for helium-users to keep in mind.