At the Large Hadron Collider (LHC), two counter-rotating beams of protons accelerated to high energy are made to collide head-on to result in the creation of myriad particles, known and unknown. From the debris of trillions of such collisions, scientists look for signals characteristic of processes involving the Higgs boson. Higgs is a very short-lived particle with a lifetime of only about ten-thousandth of a billionth of a second. Once created, the Higgs boson will immediately decay into several channels and experiments analyse the final products of such decays and see if these really came from the decay of a Higgs boson.
What makes this task really tough is to be able to pick the right events from a large background of other processes from known physics that mimic the decay of Higgs.
A signal for Higgs means that, in a plot of events observed, a bump sticks out above the large background from other mimicking processes. But such excess of events should be statistically significant to be ascribed to a new entity such as Higgs.
That is, the bump should not be explainable by statistical fluctuation in the background if there were no Higgs, and it is indeed due to processes involving Higgs. Statistical significance is measured in terms of what is called standard deviation (called sigma).
For any discovery in particle physics, the signal should be at least at ‘5 sigma level’ over the background, which is equivalent to one in 3.5 million chance of the bump being due to statistical fluctuation.