Mass effect: Maybe Higgs, maybe not

Contributed by
Dec 13, 2011
<?xml encoding="utf-8" ?>

Today, scientists at CERN in Geneva announced their results for their search for the Higgs boson, a subatomic particle that, if it exists, is thought to be responsible for giving other particles mass. It's no exaggeration to call it a keystone in quantum mechanics, and finding it for sure will be a huge accomplishment for particle physicists.

So, did they find it?

Maybe. Then again, maybe not.

Um, what? OK, this'll take a wee bit of explaining.

Last things first

I said Higgs, Magnum. HIGGS.
First, the conclusion, so at least you have that in mind as you read the rest. There are two experiments running at CERN looking for the Higgs particle. They don't smash particles together, look around with magnifying glasses and tweezers, and then yell "AHA!" when they find one. Instead, they build up a picture of it after doing gazillions of particle collisions. After a year of runs, both experiments see something that might be Higgs, but they're not 100% sure. One sees something at about the 94% confidence level, the other at 98%. That's pretty good, but it's not enough to be completely sure. It seems likely they've found something, but it's like a fuzzy picture: it looks like Higgs, but it still might be something else.

So why can't they be sure one way or another?

CSI: Geneva

Basically, what the Large Hadron Collider at CERN does is whip protons around at nearly the speed of light, then smashes them into each other. At that speed they have huge energies, and when they collide that energy gets converted into matter: other particles. Like shrapnel, these new particles explode away from the collision site. Many of these new particles aren't stable; they decay into yet lower energy particles after incredibly short time intervals. For example, electron and protons are almost certainly stable over long times (like the lifetime of the Universe), but neutrons decay after only a few minutes, turning into a proton, and electron, and a particle called an antineutrino.

So these daughter particles from the proton collisions in LHC decay, and they have daughter particles, and some of those decay, and so on. At the LHC there are two ginormous detectors called ATLAS and CMS. Both of these, in essence, measure the energy of the particles that hit them; like forensics team, they look at the aftermath of the collision and try to work backwards to figure out what happened.

We know to some extent how much energy is expected from these collisions due to all the particles that are currently known, so those can be accounted for. But if there's some excess of energy, that could very well indicate a new particle. And we have theories as to how much energy the Higgs particle should have. So the energies are measured, calibrated for known particles, and the excesses are examined.

What both experiments found is an excess of energy -- a bump in the graph -- indicating a particle that has an energy* about 125 times that of a proton -- right in the expected range for the Higgs particle. That's exciting! But what they're doing is counting up things statistically, so they can't be 100% sure. The bump in the graph is still fuzzy.

Rolling loaded dice

An article on Ars technica gave a great analogy, which I'll paraphrase: imagine you have a pair of dice. One of them is not normal: instead of the numbers 1 - 6, it replaces the 6 with a 5. If you roll the dice once, you might get 2-3, or 1-5, or 6-4. It's random. But because one of the die is missing a 6 and has an extra 5, if you roll them enough times that starts to become apparent. After a bunch of rolls, you see too many 5s and not enough 6s. If you roll them, say, three times you might not see anything, but roll them 10,000 times and you'll definitely know something is up. The more you roll, the more confident you'll be.

It's the same thing at CERN. Every run in the collider is a roll of the dice. Do it once and you might see something, but your confidence is low. Do it again and you get better statistics. Do it thousands or millions or billions of times, and you get more confident. In fact, you can calculate your confidence level. For one experiment, the bump at 125 times the energy of the proton has a confidence level of about 94%, the other experiment sees it at about 98%.

Here's the plot from the ATLAS experiment at LHC:

The horizontal axis is energy -- it's measured in a weird unit called GeV, but a proton has an energy of very close to 1 GeV, so think of the axis as being in proton units -- and the vertical axis is a measure of how certain they are the measurement is real, which depends on how many times they saw a particle at that energy. The units are called sigma, and are calculated statistically. 1 sigma means your confidence level is about 68%, 2 sigma is about 95%, and so on. The bigger sigma is, the more confident you can be.

The dashed line is what they would expect to see if Higgs doesn't exist, and the solid line connecting the dots are what they actually saw. You can see the bump around 125 GeV; it's way above what you'd expect if Higgs doesn't exist, and that's what has everyone so hot and sweaty. The bump tops out at about 2.4 sigma, which is where the 98% confidence comes from. The results from the CMS experiment are a bit noisier, and top out at 1.9 sigma, or 94%.

Exuding confidence

That sounds good, and it is, but in physics we want even better confidence than that. After all, a 94% confidence means there's a 6% chance of being wrong (as the CERN press release notes, there's a 3% chance of rolling two sixes with a pair of normal dice on the first try, so you have to be careful here). So while these confidence levels are good, they're not great. Physicists start getting excited around the 99.99% (4 sigma) confidence level, and start celebrating at 99.9999% (5 sigma). Seriously. After all, at that point the odds of being wrong are literally one in a million!

Some good news too is that this works the other way as well: looking at higher energies, they don't see any evidence for the Higgs particle. So while they can't be sure they see it at 125 GeV, they are in fact very confident they don't see it at energies higher than 125 GeV or so. That's good: it's always nice to eliminate possibilities when you're looking for something.

OK, so what does this mean?

So the grand conclusion here is:

Scientists at CERN cannot claim with enough confidence they have found the Higgs particle, but neither can they rule it out. There's a good chance they have have found something, and it very well may be real, but they cannot say with complete confidence that it's the Higgs.

They will continue to run more experiments and try to bump up that confidence level a few more notches. In other words, they have to keep rolling those dice, building up the numbers, and get better statistics. As they do, those confidence levels will change, and hopefully move into the "5-sigma-we-have-a-winner" stage. But that takes time, and it'll be 2012 at least before we know more one way or another.

If you want to read more about this, I suggest Dennis Overbye's article at the NYT, and this nice overview by Jon Butterworth at The Guardian.

* In quantum/particle physics, energy and mass are two sides of the same coin. All these experiments measure energy, but that's pretty much the same as the rest mass of the particle. So if energy makes you uncomfortable, think of it as mass and you'll be fine. So in this case, what they seem to have found is something that has a mass about 125 times that of a proton.

Related posts:

- LHC smacks some protons! (Includes a video I made when I toured the LHC a few years ago)
- Brian Cox calls ‘em like he sees ‘em
- Breaking: LHC still will not destroy the Earth
- Get your mass handed to you