MRI machine disables every iOS device in facility

Pretty fascinating story and conclusion for a nerd like me.

A medical facility was having a new MRI machine installed. Shortly after, dozens of iPhone/Watch owners reported that their devices froze. Other devices were fine, as were the humans.

It wasn’t an electrical problem, like you might expect. Instead, what happened was that a fairly small amount of helium was released during the MRI cooldown. MRI machines use liquid helium for cooling their superconductors.

Most of this will be vented outdoors, but this is an imperfect process. There’s not much need to do better since helium is completely inert and nobody will notice a percent or two of extra helium in the atmosphere.

However, newer Apple devices use what’s called a MEMS oscillator. MEMS are micromachines: very small devices with moving components. Apple uses an oscillator (a device meant to keep time by pulsing at a well-defined rate) rather like the balance wheel on a watch–a mass on a spring that swings back and forth at some rate–just much smaller.

As one might imagine, the vibration frequency depends on the ambient atmosphere. Pressure and atmospheric composition both play a role. And so MEMS oscillators are hermetically sealed in order to keep atmospheric variation in check.

However, helium is a very crafty atom. It is tiny, even as atoms go, and monatomic as well, being a noble gas. And so it is virtually impossible to build a perfect seal against helium. It will invade solid metal simply by squeezing between the metal atoms. A thin metal shell like on the MEMS device is not much of a barrier to helium.

What happened is that the very low helium concentration infiltrated the MEMS oscillators and changed their timing ever so slightly. iOS did not like this change in timing for whatever reason, and the devices hung. Airing them out and letting the battery discharge was sufficient to revive them.

Wow! That’s a new one.

Helium balloons for kids’ fun are still in common use where I live. If one were discharged next to an iPhone would that disable the phone temporarily?

I’ll bet someone has already tried it.

A friend used to work at Xerox and when introducing a new printer they had a photo shoot of the hardware operating. To make the photos more interesting, they opened up or removed some of the covers so you see the internals. Every time they took a certain shot the printer OS crashed. Turns out they exposed some EPROMs and the camera flash was erasing it.

Just popping a balloon nearby a phone wouldn’t do it–it needs a sustained high-helium atmosphere. One person made a video demonstrating the effect, but even then it took several minutes at nearly 100% concentration to affect the device. It’s probably not a linear effect but I’d suppose it would still need tens of minutes at, say, a ~1% concentration. The helium from a kid’s balloon would dissipate before it could have an effect unless you could trap it for a period of time.

In the actual incident, there was a fairly large volume of helium released over a large volume. Gases diffuse rapidly (not to mention forced ventilation), but that would still take a while to fully dissipate outdoors. Anything within the high-concentration bubble would have ample time to be affected.

Even just ordinary silicon chips can be affected this way. The Raspberry Pi 2 also crashed when you took a photo of it. It has a power supply chip with bare, exposed, silicon, which glitches when exposed to a pulse of light. The power is interrupted for a short period and the computer reboots.

I can see this as part of a spy movie in the near future.

“We’ll release helium into the building and that will cause the iPhone based security system to shut down”

A crafty atom that malevolently disables my iPhone can hardly be considered noble. :smiley:

Seriously, though, I truly find the science behind this to be fascinating.

“And the guards will begin talking in funny high-pitched voices. Win-win!”


Pretty interesting followup video by a guy with his own electron microscope.

TL;DW: Somewhere around 1-2% partial pressure of helium is enough to kill the oscillator. I’m surprised at this: I thought there would be just a small frequency shift, and the iPhones died because they detected the shift somehow. But in fact the oscillator just flat-out dies with just a small He concentration. Hydrogen did nothing.

A worse outcome would be all the iOS devices in the hospital suddenly disabling the MRI machines, CT scanners and ventilators. :eek: