ETD system

Electronic theses and dissertations repository


Tesi etd-06042018-101256

Thesis type
Tesi di laurea magistrale
Search for H->mu mu in the VBF production channel with the CMS experiment at LHC
Corso di studi
relatore Azzurri, Paolo
correlatore Prof. Rizzi, Andrea
Parole chiave
  • higgs boson
  • CMS
  • Vector Boson Fusion
  • multivariate analysis
Data inizio appello
Riassunto analitico
Understanding the mechanism that breaks the electroweak symmetry and generates<br>the masses of the known elementary particles has been one of the fundamental<br>endeavors in particle physics. The breaking of the electroweak symmetry is allowed<br>if at least one new particle with well defined properties is added to the ensemble<br>of the elementary particles. Such a particle has long been know as the Higgs<br>boson. Its discovery at the Large Hadron Collider (LHC) at Cern in 2012 by the<br>ATLAS and CMS collaborations represents therefore a major achievement in<br>the field. Starting in 2012, the properties of the Higgs boson have been measured<br>in many of the accessible final states originating from its decay. The mass of the<br>Higgs boson has been determined to be 125.09 ± 0.21 (stat) ±0.11 (syst) GeV, from<br>a combination of the ATLAS and CMS measurements. Several results from<br>both experiments established that its measured properties, including its spin, CP<br>properties, and coupling strengths to fermions and bosons, are consistent with the<br>Standard Model (SM) expectations.<br>As new data is collected, the properties of the Higgs boson can be measured with<br>increasing precision and rarer decay modes become accesible. Such measurements<br>are interesting because any deviation from the prediction of the theory might be a<br>hint of new physics beyond the SM. Among the rare decay modes currently under<br>investigation, the Higgs boson decay into two muons (H → μμ) is the object of<br>study of this thesis.<br>For a Higgs boson with mass of approximately 125 GeV, the probability to decay<br>into a muon pair is expected to be B(H → μμ) = 2.2 × 10 −4, making it one of<br>the smallest accessible at the LHC. On the other hand, the H → μμ signature is one<br>of the cleanest to detect experimentally. Higgs boson decays in two muons are of<br>particular importance because they extend the study of its couplings from the third<br>generation to the second generation of fermions, where deviations from the SM<br>predictions, due to new physics are predicted to be larger.<br>The search for H → μμ presented in this work is performed selecting the vector-<br>boson fusion (VBF) production mode. The cross section is about 10% of the cross<br>section for the gluon-gluon fusion, which is the most important production mode.<br>However, the VBF process gives a cleaner experimental signature. In fact, in the<br>VBF process, a quark coming from each colliding proton radiates a W or Z bosons<br>vvi<br>that subsequently interacts. The two quarks therefore slightly deviated from their<br>original flight direction and typically fall inside the detector acceptance, while a<br>Higgs is emitted.<br>Restricting the scope of the search to the VBF production mode, makes the process<br>even rarer but the peculiar signature of the VBF production mode can be exploited<br>to effectively reduce the experimental backgrounds. The VBF quarks are revealed<br>as jets: two back to back high momentum narrow cones of hadrons and other<br>particles produced by the hadronization of a quark or gluon. Generally the two<br>VBF jets are expected to have high pseudorapidity and large invariant mass while<br>the Higgs decay products are expected to be in the central region of the detector.<br>Imposing the constraints to the invariant mass and the rapidity of the jets as addi-<br>tional cut one reaches an impressive improvement of the signal-to-background ratio.<br>The data used for this search were collected using proton-proton collision at<br>sqrt(s) = 13 TeV by the CMS experiment in 2016, corresponding to an integrated<br>luminosity of 35.9 fb −1 . Only 30 event are expected during the entire data taking<br>period. It is therefore essential to have a high signal efficiency, both in the online<br>and the offline selections, while greatly reducing the backgrounds. The dominant<br>sources of background in these studies are production of top quark pairs (tt) and<br>Drell-Yan leptons with associated jets (referred to as DY+jets). These have a good<br>probability to decay into muons, whose tracks risk to be misclassified as coming<br>from a Higgs decay. The DY+jets background is the hardest to discriminate because<br>it is characterized by two real prompt leptons from a virtual Z or γ boson in addition<br>to two jets, either from initial state radiation.<br>A multivariate approach is used to further discriminate signal from background.<br>As background processes are many orders of magnitude larger than the signal, a<br>Machine Learning (ML) classifier with an extremely good signal acceptance versus<br>background rejection performance is required. For this purpose two different ma-<br>chine learning techniques are used: Boosted Decision Trees (BDTs) and Deep Neural<br>Networks (DNNs). Such systems &#34;learn&#34; (i.e. progressively improve performance<br>on tasks) by considering examples, generally without task-specific programming.<br>The toolkits used in this thesis to implement the multivariate classifier algorithm<br>are TMVA [6] for the BDT method and the Keras library, running on top of Theano,<br>for the NN one. Both are integrated into the ROOT analysis framework.<br>My personal contribution has been the development of these dedicated multivariate<br>techniques, including the search and selection of the most discriminant variables.<br>In order to improve the suppression of the background sources and to obtain the<br>maximum sensitivity a particular attention was given to the choice of the variables<br>starting with the definition of an extensive set of kinematic observables. Several<br>tests were made to search the best discriminant variables checking also the cor-<br>relation between all the features. Seven variables are considered as the inputs of<br>the BDT. The same input variables are sent to the NN with the addition of other<br>five. After several network configurations the best one results using a pretraining<br>step without the muon invariant mass mll (that is the most discriminant variable) and then a training with the previous weights with the complete features set. In this way is possible to exploit the discriminating power of all the selected variables. <br>The expected final goal is an improvement of the branching ratio upper limit of<br>the process. The current results are still preliminary but encouraging: for a Higgs<br>boson decaying to two muons, the upper limit on the decay rate at 95% confidence<br>level (CL) is expected to be approximately 2.5 times the SM value.