Want to understand better what was shown at #CERN today on the search for the #Higgs boson? Non-specialists check:http://www.quantumdiaries.org/2012/07/04/new-baby-boson-is-born-weighing-in-at-about-126-gev/
To find this new particle, physicists sifted through billions of events, looking for more events having well defined characteristics than what is expected from other well-known processes described by the Standard Model. These we refer to as the background. An excess of events indicates something new is also present.
But particle physics follows statistical laws and what you get is never exactly what you expect, within a certain margin of error. Evaluating this margin is crucial to being able to make correct statements.
Imagine the following. In a large bag, mix a thousand blue marbles and a thousand red marbles. Then blindly draw ten marbles out of the bag, how many red ones will you get? Seven? Five? None? All these answers are probable, except 5 is more likely than 7 which is also more probable than none.
And if you draw 100 marbles, you are more likely to get closer to 50% red marbles. The same occurs in particle physics: the statistical fluctuations get smaller once you have a larger data sample. Hence, the error margin on the presence or not of the Higgs boson is smaller when adding the 6 inverse femtobarns (fb-1) of data collected in 2012 to the 5 fb-1 collected in 2011 (this is just how we measure the amount of data).
We cannot look for the Higgs directly since it decays into smaller, more stable particles. Hence, we find events potentially containing a Higgs boson by looking at its decay products.
The two decay channels that give the most precise measurements if a Higgs has a mass around 125 GeV are when a Higgs boson decays into two photons or when it decays into two Z bosons, each one breaking apart into two electrons or two muons. This is called the four-lepton channel, since both electrons and muons are leptons. As can be seen on the plot below, Higgs decays to a pair of b quarks or WW are more frequent but much less precise to find the exact mass and the background in these channels is also very large, making it harder to see anything.
The many different ways a Higgs boson can be produced and decay. At 125 GeV (shown by the vertical dashed line) many channels are possible. The vertical logarithmic scale shows which process occurs more often as a function of the Higgs boson mass (the horizontal axis). Any decay involving quarks (qq or bb), tau leptons τ or neutrinos ν is less precise since some of the fragments are totally or partially lost. Fewer events will happen where a Higgs boson decays to two photons γγ or four leptons l+l-l+l- but they are easier to reconstruct since nothing is lost.
If the debris we find really comes from a Higgs boson breaking apart, once you recombine them, they will all cluster at the same mass and we will see an excess of events above the background at this particular mass value.
In the two-photon channel, CMS sees an excess of events that corresponds to 4.0 times the error margin on the expected number of events coming from the background. This is what we call a 4 sigma variation. For ATLAS, the excess corresponds to a 4.5 sigma deviation at a mass of 126.5 GeV.
Here are the 59059 events selected by ATLAS in the two photon channel. The small bump at 126.5 GeV corresponds to the 170 events that could be coming from the Higgs boson. The bottom insert shows what is left after subtracting the background, estimated in the upper plot by the red curve, making the Small excès more visible.
Then both experiments also see an excess of events in the four-lepton channel. It ranks at 3.4 sigma for ATLAS and 2.5 sigma for CMS. The most probable mass is 125 GeV for ATLAS and 125.5 for CMS.
Here we can see under the peak drawn in red the excess of events attributed to Higgs bosons decaying into
The other channels, although less sensitive, offer good cross-checks. When each experiment adds up the probabilities for all channels and data analyzed so far, CMS obtains a total probability of XX while ATLAS sees 5.0 sigma. Taking into account that statistical fluctuations are always possible if you look at all the mass values in the range of 110-150 GeV, then the global significance is slightly reduced to 4.7 sigma for ATLAS and XX for CMS for the most probable mass being measured at 126.5 GeV in ATLAS and XX GeV in CMS.
The local probability that the observed excess of events found by ATLAS comes from a statistical fluctuation of the background is less than a chance in thirty million as shown by the solid curve. The dashed curve shows how strong the excess should be if it is produced by a Standard Model Higgs boson. Hence, the observed excess is slightly stronger than what is predicted by the Standard Model, but still within errors.
In particle physics, an experiment needs a 5 sigma excess to claim the evidence for a new particle. But since both experiments have it independently, it is clear that a new boson has just been discovered.
So welcome, little boson. But is it really the Higgs boson predicted by the Standard Model? That is too early to tell even though the chances are excellent. We must first check if this new particle is produced and decays exactly as the theory predicts for a Higgs boson in all possible channels.
Much work remains to be done and hopefully, other exciting discoveries will also come in due time as more data becomes available. Meanwhile, champagne is in order!
One event with two muons (tracks in red) and two electrons (tracks in green) found by CMS.
One of the four muon event selected by the ATLAS search. This could be coming from a Higgs boson decaying into four muons shown by the four red lines representing their tracks in the detector. There is no way to tell if this one in particular comes from a Higgs boson or some background event.
Pauline Gagnon
