Hunting for Higgses

Update: There’s a slightly expanded version of this post on the NOVA website, where I fill in some background on what the Higgs is and why we care.

Greetings from Geneva, where I’m visiting CERN to attend the much-anticipated Higgs update seminars on Wednesday, July 4. We’re all wondering whether they will say the magic words “We’ve discovered the Higgs,” but there’s more detailed information to watch out for. Hoping for some good book fodder, at the very least. (I personally am not hunting for Higgses, any more than someone who eats at a seafood restaurant has “gone fishing,” but you know what I mean.) Remember Higgs 101, and why we need it.

If at all possible, I’ll try to live-blog here at CV during the seminars. They will start at 9am Geneva time, a slot chosen to enable a simulcast in Melbourne for people attending the ICHEP Conference. For folks in the U.S., not so convenient: it’s 3am Eastern time, Midnight (July 3/4) Pacific time. Here is the seminar announcement, and of course CERN will have a live webcast. Or try to, anyway; last time something like this was arranged, back in December, the live feed collapsed pretty quickly under the load. I’m sure I won’t be the only one live-blogging: here’s Aidan Randle-Conde and Tommaso Dorigo.

So what are we looking for? You don’t find the Higgs by looking at an individual event in a detector and going “Yes! There it is!” For one thing, you don’t see the Higgs at all; it decays too quickly. For another, there’s nothing that the Higgs can decay into that can’t be produced in other ways as well. So what the experimentalists look for is a tiny bump — an enhancement of the rate of some process at some specific energy. That’s an indication that there’s a particle with a mass equal to that energy, whose decays have led to the extra events. Here’s the results from ATLAS last year — see the bump at 126 GeV?

We need statistics to tell whether any given feature is likely to be real or not, of course. We speak in “sigmas” — the number of standard deviations away from the expected value we’re looking at. In other words, how likely is it that the data we collected would be produced as a random fluctuation, rather than because we’re seeing a real effect? A three-sigma fluctuation happens less than 0.3% of the time, and that’s considered “evidence for” something — that’s what we got in December. A five-sigma result is less than one in a million; that qualifies as “discovery of” something, and that’s what everyone is waiting for.

It’s more complicated than that, of course. We have two experiments collecting data, CMS and ATLAS. Both are giant multi-purpose detectors; the LHC smashes protons together inside them, and they detect sprays of electrons, photons, muons, and all sorts of hadrons. If the Higgs is really there, both experiments should have comparable sensitivity to looking for it — so we’d love to see similar signals in both detectors. If both see a signal in exactly the same place with high statistical significance but not quite five sigma, it will be impossible to resist mentally combining them and concluding the total evidence for the Higgs is better than five sigma. (Although there are reasons to resist…)

Besides comparing different experiments, we want to compare different channels within each experiment, and we want to compare experiment to theory. In the Standard Model, once you fix the Higgs mass, there are no more free parameters to mess with; you make precise predictions for the rate of production of any particles the Higgs might decay into. Deviating from those predictions implies that you haven’t precisely found the Standard Model Higgs — but maybe a close relative thereof?

If the Higgs really is lurking at 125 GeV, Nature is giving us a very nice opportunity, because the Higgs should (if it’s the simple Standard-Model version) decay into a variety of different particles, and we can study each one separately. Every experimental possibility is a different “channel.” Here are the predictions for the simple Higgs at this mass. (Note that in some cases these particles quickly decay themselves.)

To make sure that what you’ve really found is the Higgs, you’d like to check that it decays into all the various channels with the right percentages. Not all final states are equally easy to find, however. When the Higgs decays into quarks or gluons, it releases a spray of jets that tend to get lost in all the background noise of other processes. The same is true if it decays into W’s or Z’s or taus and then those decay into quarks and gluons, which happens a lot. Our favorite processes, then, are when the Higgs decays all the way to leptons and photons. Indeed, a lot of the excitement from last December (including the ATLAS plot above) came from looking at two-photon events, even though those are expected to happen less than one percent of the time. Decays into four charged leptons (electrons or muons), which can happen when the Higgs goes to two Z’s and each Z decays into charged leptons, happen something like 0.01% of the time, but they’re so easy to spot that they’re also a favorite channel. With more data, they’re certainly hoping to do a better job in comparing with other channels as well.

So that’s what we’re looking for. Are there excess events at some particular energy, enough to declare a discovery? Are they consistent between the two experiments? Is the overall rate consistent with the predictions of the Standard Model? Are the rates in different channels compatible with each other?

The Higgs boson is not the end of a road; it’s a bridge from one world to another. It’s the last particle we need to make the Standard Model complete, but it also gives us a way to travel to what’s beyond, whether that might be dark matter, supersymmetry, extra dimensions, or what have you. Sadly we’re not in possession of a reliable map; we just have to cross the bridge and see where it takes us.