Hardening by Design Against Neutron Rain

A local election result was altered by neutron rain interacting with an e-voting computer. How can we design against damage done by neutron rain, to electronic systems like this,...
25 November 2012

Interview with 

Dr. Chris Frost, ISIS


Ben -   ISIS can be used as a type of microscope to image the positions and dynamics of atoms, but it's also ideal for seeing how neutrons themselves interact with systems.     Neutrons can actually be a problem.  They can damage sensitive electronic circuits and interact with computer systems.  A new target station at ISIS allows real components to experience many years' worth of neutron impact in just a few seconds as Dr. Chris Frost explained...  

Chris -   The problem is that it's raining and it's not the rain that you and I normally see which is water.  It's neutrons and these neutrons come from cosmic rays that hit the top of the atmosphere and generate, through particle reactions and nuclear reactions, by the time it's down to ground level or human occupational level, what we see is a gentle rain of neutrons.    

Humans and other animals have evolved in this, so while it does damage to us, it damages our DNA, we have mechanisms that repair it.  The trouble is that the electronic systems that we now rely on for all our everyday lives haven't evolved in the neutron rain and so they get damaged, but they can't repair it.  So we have to learn how to do that repairing for them.  

Printed circuit boardBen -   What sort of damage actually occurs?  Are we talking about breaks in wires or are we talking about a more subtle, general corrosion?  

Chris -   It's much more subtle than that.  It's to do with a nuclear reaction, so the subatomic particle of the neutron collides with silicon atoms in these devices, and that causes charge to be dumped into the devices, charge is the thing that makes electronics work.  So, if you put charge in the wrong place through these nuclear reactions, then you get things happening that you don't want in your electronic devices.    

A nice example is in 2003 in Schaerbeek in Belgium when an electronic voting machine added 4,096 more votes to one of the candidates.  I quite like to say that neutrons can change governments but it was in fact a local election and so, they worked out that there was something wrong really because there were more votes than there were people.  They worked out in fact, it couldn't have happened and they traced this back to a neutron.  So a neutron effectively voted in a local election in Belgium through this neutron rain idea.  

Ben -   How widespread a problem really is it?  

Chris -   It is all over the world.  The neutrons are generated through cosmic rays and these cosmic rays are from the galactic source.  They're not the ones that come from the Sun.  They're pretty much constant.  They've been constant for millennia and it's constant day and night all over the world.  There is a slight variation due to the magnetic field of the Earth, but it's a problem worldwide and it's a problem associated with pretty much all of the electronic industry.  Of course, the biggest problems are in the critical electronic systems; examples of that are flight systems on aircrafts.  As you go up in the atmosphere, it's 300 times greater at flight altitudes than the ground level and if you think about that, that means is if you would see something go wrong once a year on the ground, you'd see it go wrong once a day at flight altitudes.  So, where this emerged first was in the aerospace industry.  But the sky is falling and what I mean by that is that electronic systems on the ground are increasingly seeing these effects as well, both in critical systems and in ones that we rely on every day. 

In the course of our lives, if you do banking, if you go on theIntegrated circuit chip internet, if you do almost anything in your life, you're interacting with electronic devices and if those electronic devices go wrong, it could be inconvenient.  In the worst possible case, it can cause injury.  

Ben -   So, what actually are you doing here at ISIS?  

Chris -   Well, what we're able to do is that we have a neutron source and that neutron source thankfully, can mimic the spectrum of neutrons. What I mean by it is that the neutrons don't come in at one energy, they come in at all sorts of energies, and they come in at different numbers across that energy spectrum.  What we are able to do here is to mimic what the spectrum looks like with neutrons on the ground so we can mimic the way they exist in nature.  The difference is that we can accelerate it somewhere between a million and 100 million times.  What that means in reality is if you come and put your electronic device in our beam line for say, an hour, then you can end up the equivalent of something like hundreds of years in the real environment.  So, we do a thing called accelerated testing and from that, we can work out how susceptible it is and from that, work out ways in which we can correct the faults in those electronic devices and make them safe.  

Ben -   So essentially, you'll see how often these devices fail and that can tell you how likely they are to go wrong in the real world.  You then must tally that up to how long their operating life would be anyway, and therefore, you can tell whoever your client is, that this thing will be safe to use in the wild for 10 years or whatever it may be.  

Chris -   Yes, it's exactly right.  Not only that, it can tell you whether you need to correct it, because some devices might not be actually susceptible, or their susceptibility means that for instance on a phone, it might fail once every 3 years, in which case, you don't care.  But if it was to fail every 10 minutes, you would care.  So in that case, you might decide that actually it's worthwhile in your development stage of your phone to put correction mechanisms for these events that goes on into your phone.  That will cost you money to do that because it takes time for software programmers and engineers to design it in.  So it tells you not only that it could fail, but it tells you how often it fails and it can also tell you how you're going to correct for it, or even if it's worth it.  So, it's a great deal of commercial value to people who produce things that we use every day.  

Ben -   What sorts of industries do you think are likely to knock on your door?  

Chris -   We've already had them knocking on our door and the kinds of industries are pretty much anybody that produces electronic devices where at the moment, failure will cost them either in safety issues, so that would be the aerospace industries; or in commercial issues and examples of that are people who build the servers and the connections that build up the internet.  So, if you have a business that relies on a server, the person who sells you that server wants to tell you that that's a very reliable server, so people will come along and test those servers or the devices in those servers to make sure they work correctly.  

Ben -   And how can you protect from this neutron rain?  

Chris -   Well, it's a thing called hardening by design.  It's a bit of a mouthful, but what it means is you're put in redundancy, so you build many circuits and all of them around at the same time and they can vote on which one's the best or as you put things that constantly check to see where the errors are occurring.  That increasingly becomes difficult as electronics gets more complicated, the kinds of errors we're seeing and the way that they propagate through the system is changing, and so, we have to stay ahead of the game, and test to make sure that it works properly.  

Ben -   So, you'll not just recommend that everybody keep their servers in a lead lined room in order to keep them safe.  

Chris -   Unfortunately, a lead lined room won't do anything.  The neutrons are so energetic they go straight through lead.  So, you can't protect by burying them in the ground unless you go a long way into the ground.  So, what you have to do is hardening by design.  You have to think about it carefully.  You have to understand what's going on and then you design within your circuitry, within your software, within all the firmware and the way that it works to pick up the errors and correct them.

Ben -   That was Dr. Chris Frost, the Project Leader of Chip Radiation Research at ISIS.


Add a comment