It will still take me a while to wrap my head around this, even if I had a head-start. I have been skeptical of the existence of the Higgs boson, but my skepticism is – almost – no longer justified. ATLAS and CMS announced simultaneously the discovery of a new particle with a mass of around 125 GeV. This particle is a pretty good candidate to be the Higgs boson predicted over 40 years ago, but there is still a lot of experimental work to be done to establish that it really is.
By the way, you can find a lot more thoughts and information about this on the Quantum Diaries.
Both experiments report a statistical significance of 5 sigma and above. This is a big deal. A 5 sigma significance is actually a ridiculously high standard to rule out statistical fluctuations. In fact, the only reason why particle physics adopted 5 sigma as its threshold for discovery is because it can afford it. It can afford it because billions and billions of collisions are analysed so that reducing the chances of statistical fluctuations down to 1 in 10 million is actually possible. I don’t envy medical and social sciences who probe correlations in samples made of no more than a few hundred people (very often less). Attracting volunteers for such studies is difficult and costly.
In particle physics, we also have the benefit of having fewer goals than medical and social sciences. This allows us to pool resources. ATLAS and CMS are international collaborations with more than 3000 scientists each, with funding coming from more than 30 countries. That is without counting what it took to build and operate the LHC. That should illustrate how hard it has been to find this new particle. Don’t be surprised it took so long.
I see this new discovery as another triumph of human collaboration. This is the quintessential proof of what this young species inhabiting a lonely planet can accomplish. We can learn how the Universe works. This is what we are doing at CERN.
Let’s look more closely at the results.
The ATLAS results were presented by Fabiola Gianotti, the spokesperson of the collaboration. She gave an outstanding presentation. Let’s get this out of the way first: please forgive Fabiola for using Comic Sans on her slides again. That kind of stuff really shouldn’t matter.
ATLAS managed to get only two Higgs decay channels analysed for today. The Higgs decaying to a pair of photons (commonly referred to as gamma-gamma), and the Higgs decaying to a pair of Z bosons (referred to as ZZ).
The gamma-gamma channel is blessed for having very little background. It is easy to recognize among the variety of events that we expect in the LHC collisions, but it is also pretty rare, as if to make sure things are not too easy. To find a gamma-gamma signal, you need to find collision events where two photons have been found, and reconstruct the mass of the particle that made them. When looking at all the data collected by ATLAS so far in 2011 and 2012, you obtain something like this:
And then you use statistical techniques to determine if that bump at 125 GeV is a fluctuation or not. Everything you see in this mass range are events that are expected according to the Standard Model theory (minus the Higgs boson), and the little bump represent an excess on top of that expectation. The vertical bars on each data point show you an approximation of how much the data points are expected to fluctuate up and down, would you gather an equal amount of new LHC collisions and made that plot again. The bump has to be bigger than these hypothetical fluctuations in order to be considered real. It is already pretty unlikely that this bump in the gamma-gamma mass spectrum isn’t real. But there is also the ZZ channel.
The ZZ analysis has a pretty staggering excess as well, although it is not as statistically significant. Look at this:
Same kind of reasoning here. As you can see, the excess appear as the same place as in gamma-gamma, so it is a reasonable assumption that the bump we are seeing is the same phenomenon. Assuming this, we can combine the statistical significance of both channel. This is how ATLAS obtains a 5 sigma excess in its data.
The CMS results were presented by Joe Incandela, the new CMS spokesperson who is a gifted speaker. He was obviously but very understandably nervous during the talk. I would be too if I had the pressure of announcing a discovery like this to the world.
The gamma-gamma result by CMS is quite fantastic and very convincing:
The ZZ result is also pretty interesting, but they have less events than ATLAS. But since they also seem to have less background, their significance in this channel is slightly higher than ATLAS. Here is the ZZ mass spectrum:
Combined, gamma-gamma and ZZ channels yield a 5 sigma statistical significance. But it didn’t end there. CMS had more analysis channels ready to show! Next, they presented the WW results. WW is a little bit different because you don’t expect to reconstruct the mass of the new particle very well, so that you don’t see a bump, just an overall excess. This channel is not sensitive enough yet, but there still seems to be a little something above background expectations.
Next, CMS talked about their bb analysis (a pair of b quarks) which is not sensitive yet but is still consistent with the existence of the Standard Model Higgs boson.
Finally, they show their tau-tau result. This one is particularly interesting to me because this is the channel I am working on. Note that ATLAS didn’t manage to show a result there. ATLAS is indeed working on tau-tau (in even more tau-tau decay modes than CMS), but couldn’t make it in time. We are getting there.
What is especially interesting in the tau-tau channel is that it is pretty sensitive, but nothing is seen there yet! It is too early to say that the tau-tau CMS result is inconsistent with the Standard Model, but I guess we will have to keep an eye on this channel. This is the only reason I say that I can still be skeptical of the existence of the Higgs boson. If the tau-tau decay is not observed, this is not the Higgs boson we just discovered, but a stranger beast. Just for the record: that would be tremendously exciting.
One thing worth noting is that CMS makes a much better use of multi-variate analyses, which makes me quite jealous. This is the kind of analysis I am working on, and it is rather difficult to motivate such analyses in ATLAS. They are simply not part of the culture. I suppose that after these beautiful results shown by CMS, ATLAS will feel the pressure to move on to more multi-variate analyses.
I still can’t believe we found something. I spent the first 25 years of my life being in a state of “we don’t know if the Higgs exist or not”, which has been overturned in the past two weeks. I can’t imagine what this must feel like for the people who have been hunting this particle from the start. I can’t imagine what it must be like for my elders and mentors who have been in the field looking for that particle most of their lives. I can’t imagine what this must feel like for Philip Warren Anderson, Peter Higgs, François Englert, Gerard Guralnik, Carl Richard Hagen and Tom Kibble, the co-discoverers of the Higgs mechanism who are still alive today. Robert Brout is unfortunately deceased.