Higgs Search: ATLAS and CMS Summer Data Combined

Matt Strassler 11/18/11

Today Gigi Rolandi, of the CMS experiment at the Large Hadron Collider (LHC), speaking on behalf of both the ATLAS and CMS experiments, presented what many of us have been curious to see for quite a while: an update on the search for the Standard Model Higgs particle, obtained by combining all of the summer’s data from the Large Hadron Collider’s ATLAS and CMS experiments. This does not include any of the LHC data gathered since the summer — that data is still being carefully analyzed. What’s new is that the experts in statistics, who understand how to take data from different experiments and put them together to get sensible statistical information, have stirred all the data into one pot, and extracted as much information as they can get.

Before I get started, let me remind you of the crucial distinction between (a) the search for the Higgs particle and (b) the search for the Standard Model Higgs particle. The latter search is Phase 1 of the former search, which is more general; the Standard Model Higgs particle is the simplest possible Higgs particle, and it is also the easiest to look for, so the natural thing to do first at the LHC is to search for it, and either find it or rule it out for good. The search strategies that are used in this effort are also sensitive to certain other more exotic Higgs particles, so it is possible we might stumble across one along the way. But many types of exotic Higgs particles are much harder to find, requiring either more time or entirely different search strategies. And of course we have to remember that even though there is at least one Higgs field, there might be no Higgs particle, and instead there may be new particles and/or forces of other types to take its place, whose details we don’t know. To find them would require a broad-based search too. So if in Phase 1 the LHC experiments rule out the Standard Model Higgs particle, Phase 2 will be a lot more elaborate and complex than Phase 1, and it will take much longer, possibly the rest of the decade.

But back to Phase 1, which we’re still in the midst of. What did we know from the summer? We knew that the LEP collider experiments, from the 1990s, had ruled out a Standard Model Higgs below 114.4 GeV. We knew that the Tevatron experiments had ruled out a small range from about 155 to about 180 GeV. Also we knew that they were slowly closing in on the low-mass range, to the point that they might see little hints if the Higgs were there, but would not have enough data in their final data set to exclude that range. And we knew that the LHC experiments had pretty much surpassed the Tevatron experiments in all but the low mass range (below about 130 GeV.) In particular, CMS and ATLAS had each ruled out very large portions of the middle mass range, between about 150 GeV and 450 GeV, though each had gaps. However, if you lined up their data, you found that most of the gaps in what CMS had excluded were filled by ATLAS, and vice versa.  As of now, only one gap was common to both experiments: 288 to 296 GeV. But if you actually looked in some detail at the data from the two experiments, it didn’t seem (to me) particularly consistent with a Standard Model Higgs sitting in that gap. So we could pretty much guess already that when the two experiments were combined together, the whole range from about 145 to 450 would be pretty much ruled out, maybe with a very weak hint around 290, and a slightly stronger hint around 143 extending down to somewhat lower masses. The results presented today confirm that; there were no surprises. At the 95% exclusion level (and remember that still means a 5% chance that the Standard Model Higgs is there and statistics played a trick on us) the Higgs is now excluded from 141 GeV up to 476 GeV. If you want 99% exclusion, then it looks like you’d exclude about 145 to about 420, with small gaps around 240 and 290.

The real unknown from the summer was this: when one combined all the experiments carefully, how far below 145 would the 95% excluded region go? And would there still be a significant gap in the exclusion around 140-145 GeV? That’s the question whose answer we got today: the 95% excluded region moves down to 141, though at 99% the limit is still around 145 or so.

Without including any theoretical considerations from precision calculations, this leaves us with two windows left to close experimentally on the Standard Model Higgs:

• the lightweight window of 115 to 141 GeV or so, and
• the heavyweight window, which runs from about 475 GeV to about 800 GeV (the upper limit being something we can debate a bit, but can’t be too much higher without requiring modification of the Standard Model’s equations).

Of course this still leaves lots of room, even in the middle region, for certain types of exotic Higgs particles.

Conversely, many theorists will tell you that precision measurements already exclude the heavyweight window (i.e., if the Standard Model Higgs were there, certain precision measurements made in other experiments would have come out differently than they did).  And others will point out that the lightweight window is constrained too.  If the Standard Model Higgs is too light, they say, lighter than about 130 GeV, the vacuum of ordinary space would be unstable; so if the Higgs is lighter than that, the Standard Model is incomplete.  (But this last argument holds only when if you insist the world is described by the Standard Model’s equations up to very much higher energies than the LHC can reach… so we won’t be able to check this claim experimentally.)  Of course the experiments should search anyway, but we’ll already start to have discussions about the Standard Model’s demise if the mass limit on the Standard Model Higgs in the low mass range move down below 130 GeV sometime next year.

We can try to learn a little bit more from another point of view. Obviously, if something you’re looking for is there, and you try to rule it out, you’ll fail. But discoveries happen gradually as you accumulate more data: first there’s a hint, then evidence, and finally near-certainty. So if what you’re trying to rule out is actually there, your first hint that you’re going to discover it eventually is that you’ll fail to exclude it as effectively as you’d expected to if it were not there. In short, if the Standard Model Higgs were now on the verge of being discovered, we’d expect to see that the observed exclusion zone is weaker than the expected exclusion zone.  And it is.  The expected 95% exclusion zone is 124 to 520 GeV, and the observed zone is weaker, 141-476 GeV.  The reason for a reduced exclusion at the high end could be an accident caused by just one or two unusual collisions, so we musn’t overinterpret it.  The reason at the low end is due to the search for Higgs decaying to two W particles, which shows a small excess in both experiments, and was responsible for this summer’s hints that were widely reported.  However, as I emphasized this summer, that search strategy is not one you would automatically be confident in; it’s very hard indeed, for many reasons.  So what we all really want is to have enough data that the easy searches, those for a Higgs decaying to two photons and for a Higgs decaying to two lepton-antilepton pairs, can come into their own in this low-mass region.  The data collected this fall is enough that this should begin to happen soon.

That’s the main news for the moment.  Granted it isn’t big news, since we could have guessed most of this from eyeballing the summer’s CMS and ATLAS results from the  Grenoble and Mumbai conferences.  The real question is when we’ll hear news that includes this fall’s big data set from CMS or ATLAS.   I’m not getting consistent answers to that question right now, but it will surely be no later than March, and probably before that.