Searches at the LHC and the discovery of the Higgs Boson Christos Leonidopoulos, University of Edinburgh March 18, 2014
Producing the Higgs boson at a hadron collider One of the most characteristic features of a hadron collider (such as the protonproton collider at the LHC) is the extremely high interaction rates. In the animation accompanying the lecture, you will see the collision of two individual protons inside the detector, producing a Higgs boson candidate 1. In reality, the two colliding protons are never in isolation. Protons at the LHC are grouped in bunches containing more than 100 billions protons each. Moving practically at the speed of light, the protons collide as the bunches transverse each other. 1 The reason that we talk about a Higgs boson candidate, and not simply a Higgs boson is that we can never know with 100% certainty if a particular Higgs-looking collision corresponds to the decay of a true Higgs particle or a background look-alike. Physicists carry out a statistical analysis of the full set of Higgs boson candidates and determine the probability that any given collision could have generated a Higgs boson or a Higgs-looking, background event. 2 / 20
Hadron colliders: a harsh experimental environment When we have two such beams of highly concentrated and highly energetic protons colliding against each other, there are several complications for the experimental environment in which we hope to detect the Higgs boson: We observe, on average, the staggering rate of approximately one billion proton-proton (pp) collisions per second. If we are lucky, maybe one of these collisions will produce a Higgs boson. A more accurate statement is that we expect on average one Higgs boson to be created for every four billion background collisions (i.e. collisions not producing a Higgs boson). The fragments from these collisions typically leave traces (e.g. ionising tracks or scintillation light) in dedicated subdetectors. This digitised activity inside the experimental apparatus captured within a time window of 25 ns 2 is defined as a collision event. More often than not, the detector records events that contain multiple collisions, with the average number of collisions per event depending on the running conditions. Obviously, the larger the number of multiple collisions per event, the harder the task of identifying a Higgs boson decay, as its decay will overlap with several other collisions fragments. At the end of the 2012 pp LHC run, the number of multiple collisions per event was greater than 20. When the LHC resumes the pp runs in 2015, this number will be closer to 50. 2 One ns is a billionth of a sec. 3 / 20
Physics production rates at hadron colliders One of the challenges here is that we are looking for evidence for a very rare event, the Higgs production and decay. Let s look at some concrete numbers. The table on the left shows the production rates of various processes at the LHC for nominal running conditions. The total production rate, representing the full activity inside the detector, is approximately 1 GHz (10 9 Hz). Production of W (Z) bosons that decay to leptons occurs at approximately 150 (20) Hz. The truly exciting physics is even more rare: New Physics that would manifest as new, exotic (i.e. hypothetical) particles (e.g. heavy cousins of the known Z boson and the gluino, notated as Z or ḡ), or a much heavier (hypothetical) Higgs boson at 500 GeV would only be created at a fraction of a Hz. 4 / 20
Hadron colliders and the Trigger The natural question to ask is how we manage to fish out the interesting physics out of the massive rate of less-exciting collisions. The answer is with the trigger : a very sophisticated hardware and software system which one can think of as a 100 megapixel digital camera that takes snapshots of the detector activity every 25 ns. Based on these detector signals, the trigger algorithms choose the collisions that look most interesting. This is a highly calculated selection of the most promising tiny slice of data for capturing a Higgs boson, or New Physics for that matter. The output of this filtering process is a few hundred events per second. The reason for keeping only a relatively small number of events is simply limited resources. Taking into account that the total collision rate is one billion Hz, we see that we are talking about a one-in-10-million selection mechanism. In other words, the trigger throws away 99.9999% of all physics interactions that we have managed to create with pp collisions inside the detector. The collaboration will only be able to analyse and search for the Higgs boson in the tiny fraction of collisions surviving the trigger. When a trigger decision is made to reject a particular collision, it is made for ever. If a mistake is made in this trigger filtering process, there is no undo button. The collision and the physics interactions that it contained are lost forever. 5 / 20
Hadron colliders and the Trigger This explains why these experiments, these highly sophisticated digital cameras, have to be so massive and complex, with tens of millions of channels, reading out signals from state-of-the-art detectors. Before we ever get a chance to look for the Higgs boson, the experiments have to make some very hard decisions about which collisions should be kept. These decisions have be to made extremely fast, leaving practically no room for errors. 6 / 20
The ATLAS Experiment The two experiments at the LHC that participated in the discovery of the Higgs boson are the ATLAS and CMS collaborations. The ATLAS detector is 45 metres long, 25 metres in diameter, and weighs about 7,000 tons. The collaboration consists of more than 3,000 physicists from all the over the world. 7 / 20
The CMS Experiment The CMS detector is more compact in size. It is 25 metres long, 15 metres in diameter, but weighs even more than the ATLAS detector, about 12,500 tons. Similarly, the CMS collaboration has more than 3,000 physicists from all over the world. 8 / 20
The ATLAS and CMS Experiments At a particle collider, it is not at all unusual to build two experiments looking for discoveries, instead of one. This co-existence of experiments, besides promoting a healthy (and mostly friendly) competition among physicists, it serves another very important purporse: the need for independent confirmation of any claim of a scientific discovery. It is difficult, but not entirely impossible for a collaboration to make a false discovery claim. It is much more unlikely to have two independent collaborations make the same mistake. This mechanism of independent confirmation or rejection of an experimental observation is the cornerstone of the scientific method. 9 / 20
The Higgs decay channels When the search for the experimental discovery of the Higgs boson started, physicists did not have much of a clue about its mass value. But the theoretical tools that we have allow us to calculate the probability that a Higgs will decay to a given set of particles as a function of the hypothetical mass. This is demonstrated in this plot 3. One can see that the probability that the Higgs decays to a b-quark and its antiparticle ( b), or two W bosons, or two Z bosons, or two τ leptons or two photons is very different for different hypothetical values of the Higgs mass. Since we don t know what the Higgs mass is, we have to look at several different channels to make sure that we will be able to observe the Higgs decay no matter what. 3 x-axis: hypothetical mass values for the Higgs boson in GeV (or GeV/c 2 ); y-axis: probability that the Higgs boson will decay to a partiular final state. 10 / 20
A Higgs boson candidate decaying into two photons The animation shown in the video lecture is an example of a Higgs boson candidate decaying into two photons. The two photons are represented by the two solid green towers. The very energetic photons get fully absorbed by a specialised detector: the calorimeter. The calorimeter measures the deposited energy of the photons. When combined, the energy of the twophoton system can give us the mass of the Higgs boson candidate. This animation depicts a single collision giving us a single measurement for the mass of a potential Higgs boson. We still do not know if this individual collision was a Higgs boson or not 4. We collect thousands upon thousands of similar collisions in order to better quantify the hypothesis that Higgs bosons are indeed created and observed inside the detector. The values for the reconstructed mass of all these Higgs candidates decaying into pairs of photons are then put into a single plot. 4 And we never will! 11 / 20
Di-photon mass distributions The diphoton mass distribution made by the ATLAS (CMS) collaboration can be seen in the plot on the left (right). A statistical analysis is needed to evaluate any potential deviation that stands out from the background that looks like a Higgs. The goal is to quantify the probability, in the event that there is no Higgs, that we could have a fluctuation caused exclusively by the background that looks so much like a Higgs peak; and, if there really is a Higgs boson, to determine its mass. This is typically at the place of the observed deviation. We can see very clear deviations in both plots, appearing around 125 GeV for both experiments. 12 / 20
A Higgs boson candidate decaying into two Z bosons The second animation included in the video lecture shows a Higgs boson candidate decaying to two Z bosons. One of the Z bosons decays to an electron and a positron. These are again depicted as solid green towers. They look very similar to photons, but these particles are charged, so they leave ionisation traces in a dedicated detector made of semiconductor material. The second Z boson decays to a muon and an anti-muon pair. Muons interact very weakly with the detector. They are able to travel very far without being absorbed by the calorimeter, like the electrons or the photons. They leave hits in another dedicated detector that can be used to reconstruct their tracks (pictured here as blue lines) and evaluate their momentum. When we combine the energy of the electron and the positron, and the momenta of the reconstructed muons, we can calculate the mass of the Higgs candidate. 13 / 20
Di-Z mass distribution by ATLAS This plot combines all the potential Higgs decays to two Z bosons for ATLAS. The plot looks very different than the diphoton distribution shown earlier. This is a cleaner channel (i.e. with smaller background), with fewer events and a very different background shape. The three coloured histograms (light blue, orange and grey) are hypothetical, simulated Higgs signals with different mass values. The black points correspond to the experimental collision data and they agree well with a simulated Higgs signal with a mass of 125 GeV. 14 / 20
Di-Z mass distribution by CMS A similar distribution produced by CMS can be seen in this plot. CMS has chosen here to zoom in at the place of the observed deviation. The red-line histogram is again a hypothetical, simulated Higgs distribution that matches very well the experimental data (black points) around 126 GeV. The value of 126 GeV is compatible with the value reported by ATLAS. The two measurements are consistent with each other within the experimental resolution. 15 / 20
Deviations observed. Now what? By now, we have observed deviations from the background in different channels in the Higgs boson searches. But if there is a Higgs boson that decays to multiple channels, the individual deviations that we observe must give a coherent picture. This means two things: not only must they provide cumulative evidence of a Higgs boson, but they must also give mass measurements that are consistent with each other. Furthermore, we have two independent experiments with very similar sensitivity. If one experiment claims a discovery, and the discovery is real, it would be extremely unlikely that the other experiment does not observe anything. If both experiments claim a discovery, they also have to agree on the Higgs mass. Can we combine this information from the individual channels to quantify how unlikely it is to see these deviations if there were no Higgs? This is an extremely difficult problem that requires very complex calculations and tedious cross checks. Describing the method used for calculating these probabilities is beyond the scope of this course. However, we can take a closer look at the results. 16 / 20
Evaluation of deviations by CMS The results presented by CMS are shown in this plot. A channel with very high sensitivity is the one with the two Z bosons discussed earlier and appears with the red colour. The red curve shows that the probability to observe such a deviation without a real Higgs behind it is smaller than 1 in a 1000 (as determined by checking that the minimum of the red curve reaches below the value 10 3, as read off the y-axis on the left). Even more sensitive is the diphoton channel depicted with the green colour. The corresponding probability is almost 1 in 100,000 (or 10 5 ). The combination, with the black colour, shows that the probability that we would observe these deviations in all channels together if there were no Higgs is smaller than 1 in a million (or 10 6 ). What is important here is that the incorporation of multiple channels contributes constructively in the evidence for the existence of the Higgs, and that the individual deviations appear at the same mass value. 17 / 20
Evaluation of deviations by ATLAS This is, similarly, the combined plot for ATLAS extended over a wide mass range. When looking at the full mass spectrum, we can observe smaller or larger deviations for several mass values. The deviations between 200 and 300 GeV do not suggest that there is a Higgs, as they are perfectly consistent with a fluctuation caused exclusively by background. There is a very large deviation, however, around 125 GeV (the deep inverse spike on the left). The calculations show that, if there were no Higgs, the combined probability to observe all these deviations in the different channels is almost as low as one in 10 million (or 10 7 ). The experimental observation crosses a standard, conventional threshold for claiming the discovery of a new particle that is referred to as a 5-sigma deviation. 18 / 20
Announcement of the discovery of the Higgs boson The experimental evidence for the discovery of the Higgs boson was presented on July 4, 2012 in a joint seminar by ATLAS and CMS at CERN. The audience included François Englert, Peter Higgs and the other prominent theorists (Guralnik, Hagen and Kibble) who worked on the theory almost 50 years earlier 5. The announcement of the discovery was met with standing ovation, something extremely unusual in particle physics seminars. 5 With the exception of deceased Robert Brout. 19 / 20
After the discovery, what? The Higgs discovery was the last missing piece of the Standard Model, a system of sophisticated equations that can predict a very large range of phenomena with very high accuracy. The Standard Model is now considered a completed theory. But here s the great paradox. It turns out that the powerful equations that have been introduced in this course describe extremely well just 5% of the known Universe. We now know that dark matter and dark energy make up the vast majority of the cosmos, namely 95% of it. What exactly is dark matter and dark energy? We don t know! How is it possible that we understand 5% of the Universe and everything we see so extremely well, but know absolutely nothing about what makes up most of the Universe? We don t know! Are there yet more particles that we have not discovered yet? Is there a connection between the Higgs boson, the potentially new particles and the dark matter? These are some of the questions that we are trying to answer at the LHC. We think that measuring the properties of the Higgs particle could give some clues that would help us begin to address these questions. 20 / 20