Two main contributions are presented in this thesis: the first evidence in the search for electroweak (EW) vector boson scattering (VBS) in semileptonic decays with $WVjj \rightarrow l\nu qqjj$ final state and the optimization of the CMS electromagnetic calorimeter (ECAL) trigger and local reconstruction for Run III of LHC. The VBS search uses a data set of proton-proton collisions at 13 TeV, corresponding to an integrated luminosity of 137 fb$^{-1}$ collected with the CMS detector during 2016-2018 (Run II). VBS is important because of its strong link with the EW symmetry breaking mechanism (the Higgs mechanism) in the Standard Model. Events are selected requiring one lepton (electron or muon), moderate missing transverse momentum, two jets with a large pseudorapidity separation and a large dijet invariant mass, and a signature consistent with the hadronic decay of a W/Z boson. Events are separated into two categories: either the hadronically decaying W or Z boson is reconstructed as one large-radius jet, or it is identified as a pair of jets with dijet invariant mass close to the bosons mass. The signal strength is measured by fitting the distributions of machine learning based multivariate discriminators, implemented to separate the signal from the backgrounds in each category. The observed EW signal strength is $\mew = 0.86 ^{+0.23}_{-0.21} = 0.86\pm0.12\,(\text{stat})^{+0.19}_{-0.17}\,(\text{syst})$, corresponding to a signal significance of 4.4 standard deviations (5.1 expected). The result reported in this thesis corresponds to the first evidence of vector boson scattering in a semileptonic channel at the LHC. The large total integrated luminosity delivered by LHC to the CMS detector and the increasing level of simultaneous interactions (pileup) has induced a dynamic and challenging environment for the performance of the ECAL detector during the Run II. Ageing effects due to radiation damage have affected both the crystal transparency and the signal pulse shape. The ECAL trigger and energy reconstruction algorithms need to adapt to the even more challenging conditions posed by the start of Run III in 2022. The ECAL trigger on-detector hardware employs fast digital algorithms to precisely measure the energy and timing information of ECAL energy deposits. This thesis describes the optimization of the digital filter weights which largely improves the robustness of the energy estimation against pileup. Moreover, a novel hardware configuration is explored for the first time to help reject anomalous signals or tag out-of-time energy deposits. The reconstruction of electrons and photons in CMS depends on topological clustering of the energy deposited by an incident particle in different crystals of the electromagnetic calorimeter (ECAL). The presence of upstream material causes electrons and photons to start showering before reaching the calorimeter. Due to the 3.8 T CMS magnetic field, the energy is spread in several clusters around the primary one. It is essential to recover the energy contained in these satellite clusters in order to achieve the best possible energy resolution for physics analyses. Historically, satellite clusters have been associated with the primary cluster using a purely topological algorithm that does not attempt to remove spurious energy deposits from additional pileup interaction. The performance of this algorithm is expected to degrade during LHC Run III because of the larger average pileup levels and the increasing levels of noise due to the ageing of the ECAL detector. New methods have been investigated in this thesis to exploit state-of-the-art deep learning architectures like Graph Neural Networks (GNN) and Transformer algorithms. These more sophisticated models, implemented for the first time in ECAL, improve the energy resolution up to 10% and are more resilient to pileup and noise, helping to preserve the electron and photon energy resolution achieved during LHC Runs I and II.

Questo lavoro di tesi si e' svolto seguendo due diverse linee di ricerca. La prima riguarda la ricerca del processo di vector boson scattering elettrodebole (EW VBS) con decadimenti semileptonici nello stato finale $WVjj \rightarrow l\nu qqjj$. La seconda riguarda l'ottimizzazione del sistema di trigger e degli algoritmi di ricostruzione del calorimetro elettromagnetico (ECAL) dell'esperimento CMS in vista del Run III di LHC. La ricerca del processo di VBS si basa sugli eventi registrati nelle collisioni protone-protone a 13 TeV, raccolti dall'esperimento CMS negli anni 2016-2018 (Run II), corrispondenti ad una luminosita' integrata di 137 fb$^{-1}$. Il VBS e' un processo importante perche' strettamente legato alla rottura spontanea della simmetria elettrodebole (EW) nel Modello Standard. La selezione degli eventi si basa su diversi requisiti: la presenza di un leptone (elettrone o muone), un moderato momento trasverso mancante, la presenza di due jet con una separazione in pseudorapidita' e una massa invariante elevate ed una segnatura consistente con il decadimento adronico di un bosone W/Z. Gli eventi sono separati in due categorie: il bosone W/Z con decadimento adronico viene ricostruito con un jet a largo raggio oppure identificato in una coppia di jet con massa invariante compatibile con la massa del bosone. Il segnale viene misurato con un fit sulle distribuzioni di discriminatori basati su tecniche di machine learning. L'intensita' del segnale EW osservato e' $\mu=0.86^{+0.23}_{-0.21}=0.86\pm0.12\,(stat)^{+0.19}_{-0.17}\,(syst)$, equivalente ad una significativita' di 4.4 deviazioni standard (attese 5.1). Tali risultati, riportati nel dettaglio in questa tesi, rappresentano la prima evidenza di VBS in un canale semileptonico misurata ad LHC. Le prestazioni di ECAL sono state significativamente influenzate dall'aumento della luminosita' integrata fornita da LHC e delle interazioni simultanee (pileup o PU) nel rivelatore durante il Run II. Gli effetti di ageing dovuti alla dose accumulata riguardano principalmente la trasparenza dei cristalli e la forma dell'impulso di segnale. Per stare al passo con l'aumento in luminosita' e pileup previsti in Run III (2022), e' necessario adattare il sistema di trigger e gli algoritmi di ricostruzione. Il sistema di trigger di ECAL si basa su algoritmi implementati nell'hardware on-detector per misurare con precisione l'energia depositata nel calorimetro. In questa tesi viene presentata l'ottimizzazione dei pesi del filtro digitale, che ha migliorato considerevolmente la robustezza della stima dell'energia rispetto al pileup. Inoltre, viene presentata un'innovativa configurazione dell'hardware, progettata per migliorare la rimozione di segnali anomali o l'identificazione di energia out-of-time. La ricostruzione di elettroni e fotoni in CMS si basa sul clustering dell'energia depositata dalle particelle incidenti nei cristalli di ECAL. Tuttavia, la presenza di materiale di fronte ai cristalli e del campo magnetico di 3.8 T, implica che elettroni e fotoni producano emissioni secondarie che distribuiscono l'energia in vari cluster attorno a quello primario. Fino ad ora, i cluster secondari sono stati associati a quello primario, per ottenere la migliore stima dell'energia totale, sfruttando un algoritmo geometrico che non tenta di eliminare i depositi spuri provenienti dalle interazioni di PU. L'aumento del noise e del PU nei Run futuri causera' inevitabilmente una degradazione della performance di tale algoritmo. In questo lavoro di tesi sono stati indagati nuovi modelli basati su moderne architetture di deep-learning come le graph neural networks e i Transformer. Questi modelli sofisticati, implementati per la prima volta in ECAL, hanno mostrato un potenziale miglioramento del 10% nella risoluzione energetica e una maggiore robustezza nei confronti di rumore e PU, contribuendo a preservare la qualita' delle performance di ECAL raggiunte durante i precedenti Run di LHC

(2022). First evidence of VBS in semileptonic decays with $WVjj \rightarrow l\nu qqjj$ final state and optimization of the CMS electromagnetic calorimeter for Run III. (Tesi di dottorato, Università degli Studi di Milano-Bicocca, 2022).

First evidence of VBS in semileptonic decays with $WVjj \rightarrow l\nu qqjj$ final state and optimization of the CMS electromagnetic calorimeter for Run III

VALSECCHI, DAVIDE
2022

Abstract

Two main contributions are presented in this thesis: the first evidence in the search for electroweak (EW) vector boson scattering (VBS) in semileptonic decays with $WVjj \rightarrow l\nu qqjj$ final state and the optimization of the CMS electromagnetic calorimeter (ECAL) trigger and local reconstruction for Run III of LHC. The VBS search uses a data set of proton-proton collisions at 13 TeV, corresponding to an integrated luminosity of 137 fb$^{-1}$ collected with the CMS detector during 2016-2018 (Run II). VBS is important because of its strong link with the EW symmetry breaking mechanism (the Higgs mechanism) in the Standard Model. Events are selected requiring one lepton (electron or muon), moderate missing transverse momentum, two jets with a large pseudorapidity separation and a large dijet invariant mass, and a signature consistent with the hadronic decay of a W/Z boson. Events are separated into two categories: either the hadronically decaying W or Z boson is reconstructed as one large-radius jet, or it is identified as a pair of jets with dijet invariant mass close to the bosons mass. The signal strength is measured by fitting the distributions of machine learning based multivariate discriminators, implemented to separate the signal from the backgrounds in each category. The observed EW signal strength is $\mew = 0.86 ^{+0.23}_{-0.21} = 0.86\pm0.12\,(\text{stat})^{+0.19}_{-0.17}\,(\text{syst})$, corresponding to a signal significance of 4.4 standard deviations (5.1 expected). The result reported in this thesis corresponds to the first evidence of vector boson scattering in a semileptonic channel at the LHC. The large total integrated luminosity delivered by LHC to the CMS detector and the increasing level of simultaneous interactions (pileup) has induced a dynamic and challenging environment for the performance of the ECAL detector during the Run II. Ageing effects due to radiation damage have affected both the crystal transparency and the signal pulse shape. The ECAL trigger and energy reconstruction algorithms need to adapt to the even more challenging conditions posed by the start of Run III in 2022. The ECAL trigger on-detector hardware employs fast digital algorithms to precisely measure the energy and timing information of ECAL energy deposits. This thesis describes the optimization of the digital filter weights which largely improves the robustness of the energy estimation against pileup. Moreover, a novel hardware configuration is explored for the first time to help reject anomalous signals or tag out-of-time energy deposits. The reconstruction of electrons and photons in CMS depends on topological clustering of the energy deposited by an incident particle in different crystals of the electromagnetic calorimeter (ECAL). The presence of upstream material causes electrons and photons to start showering before reaching the calorimeter. Due to the 3.8 T CMS magnetic field, the energy is spread in several clusters around the primary one. It is essential to recover the energy contained in these satellite clusters in order to achieve the best possible energy resolution for physics analyses. Historically, satellite clusters have been associated with the primary cluster using a purely topological algorithm that does not attempt to remove spurious energy deposits from additional pileup interaction. The performance of this algorithm is expected to degrade during LHC Run III because of the larger average pileup levels and the increasing levels of noise due to the ageing of the ECAL detector. New methods have been investigated in this thesis to exploit state-of-the-art deep learning architectures like Graph Neural Networks (GNN) and Transformer algorithms. These more sophisticated models, implemented for the first time in ECAL, improve the energy resolution up to 10% and are more resilient to pileup and noise, helping to preserve the electron and photon energy resolution achieved during LHC Runs I and II.
PAGANONI, MARCO
GOVONI, PIETRO
MASSIRONI, ANDREA
CMS; VBS; calorimetria; trigger; ricostruzione
CMS; VBS; machine-learning; trigger; ricostruzione
FIS/04 - FISICA NUCLEARE E SUBNUCLEARE
English
19-gen-2022
FISICA E ASTRONOMIA
34
2020/2021
open
(2022). First evidence of VBS in semileptonic decays with $WVjj \rightarrow l\nu qqjj$ final state and optimization of the CMS electromagnetic calorimeter for Run III. (Tesi di dottorato, Università degli Studi di Milano-Bicocca, 2022).
File in questo prodotto:
File Dimensione Formato  
phd_unimib_785396.pdf

accesso aperto

Descrizione: PhD thesis - final version
Tipologia di allegato: Doctoral thesis
Dimensione 18.11 MB
Formato Adobe PDF
18.11 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/10281/364125
Citazioni
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
Social impact