jump to navigation

Cosmic-ray radiography January 25, 2007

Posted by dorigo in computers, internet, physics, science.

When discussing with people who believe in astrology, I have a hard time maintaining that our life is not influenced by celestial bodies and cosmological effects. I can usually win the argument thanks to my superior knowledge of Physics and Astronomy, but even then I feel like a half-loser, because I do know our bodies are indeed influenced in unpredictable ways by extraterrestrial influences.

Wow, what a sentence – let me qualify that a bit. What I am talking about is the fact that every second our body is traversed by gazillions of particles of cosmological origin. Most of them are neutrinos, and sure enough, they usually do not interact with our flesh and bones, being able to travel through light-years of material totally unseen. But we are also showered with tens of muons every second, and muons do interact with our body, leaving trails of ionization behind.

Muons are produced when high-energy protons and alpha particles (so-called primary cosmic rays)  from cosmological sources impinge on the Earth’s upper atmosphere, undergoing a nuclear interaction with atoms of nitrogen or oxygen. This results in a shower of light hadrons – mostly pions, the lightest hadronic-interacting particles-, electrons, and photons. Charged pions decay in about ten billionths of a second, yielding muons that continue the trip and easily reach the surface of our planet. Sure, charged hadrons, electrons, and photons also reach us in small amounts, but they are absorbed by a few inches of lead, while muons will punch through it easily.

To count muons from cosmic rays, therefore, one just has to place two particle detectors (such as planes of scintillating material connected to a photomultiplier) on top of another under a lead shield, and adjust the electronic readout of the photomultipliers such that it counts only if both see a signal at the same time. 

What is happen-ing is that the muon traverses the centimeter-thick sheet of plastic scintillator, leaving a trail of ionized atoms. The ionized atoms in the material (the red dot in the sketch above) emit ultraviolet light (the cyan lines) which bounces on the sides of the scintillator until it reaches the photomultiplier window. There, ultraviolet photons hit an alkali material which releases an electron by photoelectric effect. The electron is accelerated in the high voltage field inside the photomultiplier, and is subjected to a avalanche multiplication, which result in a sizable, fast peaked signal: a short current pulse. The signal is processed by the electronics, and put in coincidence with the output of the other photomultiplier. At the end of this chain, if both scintillators were traversed by a muon, you can see a counter turn its rightmost wheel by an unit – click! 

I used the setup shown on the left with two fellow students during an undergraduate course in experimentation. It was a really nice experience, although we worked hard for almost a year. We had four layers of scintillators (in blue) sandwiching as many sheets of aluminum (in cyan), and we used the instrument to measure the lifetime of the muon. In fact, if a muon got trapped in the aluminum, we could observe a signal from its entrance in our device (left when the red line crosses the first three blue sheets in the diagram sketched above), no signal of its exit (in the lower sheet), and a delayed signal from its decay to an electron (the brown track). The time distribution of these delayed electron signals fit to an exponential curve with a lifetime of 2.2 microseconds… I learned a lot from that experiment.

Cosmic ray muons nowadays are mostly used by particle physics experiments during their commissioning phase, when parts of the apparatus have to be tested and there is no particle beam to test them with. What we do then is to just turn on our devices and expose them to the shower of muons, detecting their signals, reconstructing their trajectories, and thus obtaining maps of dead channels, alignment constants, hints at the (mal)functioning of the whole machinery. CMS is exactly in this phase right now, and a full 25% of its silicon tracker is presently taking cosmic ray data for precisely that purpose. And I am analyzing the data coming out.  

But cosmic ray muons have practical applications which go beyond the study of subnuclear physics. In recent years the interest of using them to detect the smuggling of heavy substances – such as bomb-grade uranium – has sharply risen.

Imagine a container within which is concealed some uranium or plutonium. You could try to detect the radioactive decay of these substances with particle detectors, but if the substance were shielded well enough you would stand no chance of finding it without opening the box and examining its contents.

Muons come to the rescue. You can place a large detector plane on top of the container, and one on the bottom of it. Take a minute of data, and a computer program will reconstruct the muon tracks that have traversed the container – of the order of a few hundred thousand. Now comes the nifty part. When a muon traverses high-Z material such as uranium or plutonium, it scatters from its original direction much more than it would if traversing lighter material. A sufficiently smart reconstruction algorithm can use the amount of scattering to obtain a three-dimensional view of the density of material inside the container. Whatchamacallit ? A smugglers-be-gone!

The technique has been demonstrated in a paper by Los Alamos physicists back in 2003: see for instance http://www.lanl.gov/quarterly/q_spring03/muon_text.shtml .



1. Bee - January 25, 2007

Hi Tommaso,

this is interesting. One minute of data would be enough? I’m surprised. I didn’t realize there are so many muons/cross-section is sufficiently large. From the 10.000 muons per minute how many would actually be detected? Best,


2. dorigo - January 25, 2007

Well, 10,000/m^2/min means roughly 50,000 in a minute on a typical container. And most of these (say 80%) will go through the whole thing (if it is not full of lead!). Muons are very easy to detect, and with a suitable detector you can easily reach 98% efficiency for crossing particles. Then, having two planes of detectors of side L, separated by a similarly sized height difference h=L, means that your coincidences will cover a third of the solid angle. However, this third will contain much more than a third of the muons, since it will be the third which is oriented in the top-bottom direction, where rates are way higher (think of a cube, and think of rain impinging on it: the face up will get most of the water).

All in all, yes, I think you can reconstruct maybe 10,000 trajectories per minute of exposure, give or take a factor of 2. That amount might be enough for coarse analysis, and in case you see something weird, well then you take x10 more data!


3. Guess Who - January 25, 2007

By the way, is the artsy cosmic ray detector in the CERN foyer still operative? It used to have an almost hypnotic effect on me, a long, long time ago…

4. dorigo - January 26, 2007

Hi GW,

I have been there a few times, but I missed it. Maybe somebody else out there knows ?


5. mis 3 quarks » Breves: 26/01/07 - January 26, 2007

[…] Radiografía de rayos cósmicos (en inglés)Sobre cómo se detectan muones y qué se puede hacer con ello. […]

Sorry comments are closed for this entry

%d bloggers like this: