How Would a Carrington Event Affect Modern Computers?

[jumpjack]

I just discovered about solar storm named "Carrington Event"; I'd like to know how such a solar storm could affect my PC today.

I found some publications:
http://www.bu.edu/csp/EMMREM/papers/carrington.pdf
http://engineering.dartmouth.edu/~Simon_G_Shepherd/research/Shielding/docs/Townsend_03.pdf

1859 solar storm:
http://www.solarstorms.org/SS1859.html

How can I determine if the "shielding" provided to my Pentium by my case would be enough to prevent it to burn up?!?

Is it possible to determine a "radiation curve" for Carrington event, in terms of rad(Si) vs g/cm^2 ?

 

[Norman]

I just discovered about solar storm named "Carrington Event"; I'd like to know how such a solar storm could affect my PC today.

I found some publications:
http://www.bu.edu/csp/EMMREM/papers/carrington.pdf
http://engineering.dartmouth.edu/~Simon_G_Shepherd/research/Shielding/docs/Townsend_03.pdf

1859 solar storm:
http://www.solarstorms.org/SS1859.html

How can I determine if the "shielding" provided to my Pentium by my case would be enough to prevent it to burn up?!?

Is it possible to determine a "radiation curve" for Carrington event, in terms of rad(Si) vs g/cm^2 ?

Hi Jumpjack,

I did my first postdoc under Larry Townsend (the first author on both of those papers) and have worked in space radiation shielding for just under a decade. Nice to see a question about this work!

What you need to do to is take the spectral form of the Carrington event and run it through 1000 g/cm^2 of Earth's atmosphere, make some approximation for your house and computer materials and depth. Then look at the response of the remaining radiation on the silicon. You do all this using a radiation transport code like HZETRN (NASA deterministic code designed specifically for space radiation), MCNPX (Los Alamos Monte Carlo code) or GEANT (CERN Monte Carlo code).

I have never run a Carrington like event, but I can give you some qualitative answers. If you look at the Table 5 of the ASR paper you linked to, you can see a table which gives Dose (rads in Si) vs Depth in aluminum. If I assume that your computer is here on Earth, you do not live near the poles, you would likely not see any effects as the dose even through only 50 g/cm^2 of Aluminum (shown in Table 5 of the ASR paper) is not of a real concern. I was not able to find any information on threshold dose for common microprocessors (Pentium lines) quickly, so this is all very hand waving.

A couple of comments, in the ASR paper, Townsend ran a code called BRYNTRN. There is no secondary pion production in BRYNTRN, so the dose estimates there are only for nucleons (protons + neutrons). This would increase the dose at large depths.

In addition, the spectral form of the event is very important. The latest thoughts on the Carrington Event were that it was a fairly soft spectrum (spectrum was peaked at low energy, not a lot of high energy particles). This is very important, because a change in the spectral shape can make a huge difference in the final dose estimates.

 

[jumpjack]

What was not clear is the definition of "spectrum", as I usually refer it to "frequency" context rather than power. After reading again those (and others) papers, now I get the point; indeed I didn't understand the table 5 is actually what I was looking for, as I read it refers to 1972 event; now I see it just refers to 1972 spectrum!

Your data about atmosphere being equivalent to 1000 g/cm^2 is also very interesting.

But now I have another question:
How can I calculate intensity of currents caused by asolar flare? I know in 1859 they were so strong they could power telegraph lines several kilometers long! And would a PC case shield against them, too?
But I don't know if this too is your subject.

 

[Norman]

But now I have another question:
How can I calculate intensity of currents caused by asolar flare? I know in 1859 they were so strong they could power telegraph lines several kilometers long! And would a PC case shield against them, too?
But I don't know if this too is your subject.

Dose is related to ionization rate. You could run one of the codes listed above and figure out the flux or dose in some defined cable. Then calculate the total energy deposited in the cable.

How strong a solar particle event (SPE) (like the Carrington Event) is a function of (very basically) two variables:
1. The total number of particles ejected.
2. The shape of the event spectrum- how many particles were emitted with a given energy.

If the total number of particles is not large, there is not total energy available to do damage. If the spectral form of the event is highly peaked at low energy, without a lot of particles in the high energy tail, then there are not enough particles with enough energy to penetrate.

The Carrington Event was a big deal because it had both a lot of particles released and a significant high energy component to the event. The Carrington Event exceeds all "modern" events in both aspects.

This is all only from the viewpoint of the event. One must also consider the target (your computer in this case). The density of transistors is always increasing. This higher density means lower energy is needed to cause errors in the chips. This is very crude, but true in general.

 

[jumpjack]

I thin I didn't properly explain my second question: apart from ionizing dose, solar storms are also "made" of strong magnetic field, if I understand correctly, which is able to "compress" Earth magnetosphere from 30 down to 6-7 km, and also to cause currents to flow into wires. I'm askin if it is possible to estimate how many Ampere would flow in my motherboard due to a Carrington Event.

 

[Norman]

I thin I didn't properly explain my second question: apart from ionizing dose, solar storms are also "made" of strong magnetic field, if I understand correctly, which is able to "compress" Earth magnetosphere from 30 down to 6-7 km, and also to cause currents to flow into wires. I'm askin if it is possible to estimate how many Ampere would flow in my motherboard due to a Carrington Event.

Very sorry, I misunderstood.

I don't know if you could estimate the change in the Earth's magnetic field simply from the ice core data that is driving the particle flux estimates which we were discussing previously. You might be able to make up some crude scaling law for order of magnitude changes.

You might want to consider looking up the term "Forbush decrease." It is used to describe the decrease in the Galactic cosmic rays due to the increasing magnetic field during solar event. You might gain some insight that might allow you to make some reasonable assumptions about how the magnetic field of the sun changes and affects the Earth.

In general, the interactions of Earth's magnetosphere are very complex, so I wouldn't hold much hope of attacking the problem from the point of view of the storm. That being said, you should be able to make some reasonable assumptions about likely currents in the wire, the wire dimensions and determine some value for the magnetic fields needed to create thos currents.