Detector Studies for HL-LHC CMS Upgrade

Introduction

The Standard Model of particle physics

Introduction

It describes three out of the four fundamental forces in nature: the strong interaction (𝑆𝑈(3)), electromagnetism (𝑈(1)) and the weak force (𝑆𝑈(2)). There are eight different types of gluons in nature, which are massless and mediate the strong force.

The Lagrangian of the Standard Model

Spontaneous Symmetry Breaking (SSB)

𝐵𝜇 𝜈𝐵𝜇 𝜈 (2.10), where the field strength tensors are defined as:. 2.14) We will now define photons (𝐴𝜇), 𝑊± and 𝑍 bosons in SM as linear combinations of 𝐵𝜇 and 𝑊𝑎. In terms of the interaction, the coupling is proportional to the square of the gauge boson mass and is known as the Higgs boson gauge coupling.

Figure 2.2: The Mexican hat shaped Higgs potential.

Limitations of the SM

2 and the second term gives the interaction strength between the electron and the Higgs boson, which is proportional to the 𝑚𝑒 and is known as a Yukawa coupling. The Λ𝑈𝑉 limiting value is expected to be close to the Planck scale, 𝑀𝑃 ∼ 𝑂(1019) GeV, while the mass of the Higgs has been experimentally measured to be ∼125 GeV.

Higgs boson physics at the LHC

Single Higgs production and decay modes at the LHC
Double Higgs (HH) production and decay modes at the LHC 16

Dimuon system pseudorapidity𝜂(𝜇 𝜇) : The dimuon system from the VBF signal process is mostly produced centrally in the detector, unlike other background processes. T > 25 GeV and |𝜂| <4.7 The signal simulation considered in the training of BDT are 𝑔𝑔 𝐻, VBF, V𝐻 and 𝑡 𝑡 𝐻 processes.

Figure 2.3: Different production modes for the Higgs boson at the LHC : (a) ggH, (b) VBF, (c) VH, and (d) ttH.

The Compact Muon Solenoid (CMS) Experiment at the LHC 21

The CMS Experiment

Tracker
ECAL
HCAL
Superconducting solenoid magnet
Muon chambers
Trigger and data acquisition
Event reconstruction

Appendix B discusses the commissioning tests performed on the two preliminary versions of the prototype BTL readout electronics, i.e., TOFPET and TOFHIR2A ASIC. The T of AK8 Jet 2 (the aircraft with the second highest 𝑇Xbb score) is shown for events in the hadronic control region 𝑡 𝑡 after applying the ramp model drag correction, for 2016 (left), 2017 (middle) and 2018 (right) data sets separately.

Figure 3.3: Recorded instantaneous luminosity and pileup at CMS during Run 1 and Run 2

A MIP Timing Detector (MTD) for CMS at HL-LHC

Introduction

The current LHC will undergo major design upgrades so that it can increase its instantaneous luminosity by a factor of ~3-4, above the LHC's design value. By the end of its first few years of operation at 13 TeV (at the end of 2018), the LHC had collected approx. The HL-LHC aims to collect about 3000 fb−1 of data by the end of its 10 year operating period.

The CMS detector will also upgrade many of its existing subdetectors and install a new time-of-flight (TOF) subdetector, known as the MIP Timing Detector (MTD) [92].

MIP Timing Detector (MTD)

Barrel Timing Layer (BTL)

Studying the HH production and measuring the Higgs trilinear self-coupling (𝜆𝐻 𝐻 𝐻) is one of the main physics goals of the HL-LHC. In the BTL, the SiPMs will operate above their breakdown voltage in Geiger mode with a gain of the order of 105. These parameters drove the optimization of the sensor layout (crystal and SiPM configuration), as discussed later in Sect.4.3.

The size of the DCR increases with integrated brightness due to radiation damage that creates defects in silicon, increasing the probability of generating thermal electrons [100].

Figure 4.2: A schematic view of the GEANT geometry of the MTD, comprising a barrel layer (grey cylinder), at the interface between the tracker and the ECAL, and two silicon endcap (orange and light violet discs) timing layers in front of the endcap calorim

Studies to optimize BTL design parameters

Characterization of the sensor properties and optimizing the
Time resolution measurements at the Fermilab Test Beam
Thermal properties of BTL modules

Figure 5.38 shows the output of the BDT discriminator in the hadronic (left) and leptonic (right) 𝑡 𝑡 𝐻 categories.

Figure 4.7: Impact point studies with a laser. (Left) A schematic of a 3x3 mm 2 SiPM coupled with a 12x12x4 mm 3 LYSO tile using optical grease

First evidence of a Higgs boson decay to a pair of muons

Introduction

However, the Higgs-Yukawa couplings with the first and second generation fermions have yet to be established experimentally. TheB (𝐻 → 𝑓 𝑓) is expected to be small for fermions belonging to the first and second generations, since the Yukawa couplings are proportional to the mass of the fermion. The most recent search for 𝐻 → 𝜇 𝜇 decays from the ATLAS collaboration corresponding to the full Run 2 dataset [119], saw an observed (expected) significance relative to the background-only hypothesis for a mass Higgs boson of 125.09 GeV at 2.0𝜎(1.7𝜎).

The analysis described here targets the Higgs boson production via gluon fusion (ggH), vector boson fusion (VBF), and in association with a vector boson (VH, V = Z,W±) or with a top-antitop pair (ttH) .

Data sets and simulation

Data sets and triggers
Simulation overview

In the case of processes simulated with the MadGraph5_amc@nlo generator in leading order (next-to-leading order), jets from the matrix element calculations are matched to the parton shower produced by pythia following the MLM (FxFx) prescription [125, 126]. On the other hand, the MadGraph5_amc@nlo samples are used in the final signal extraction. Therefore, a Drell-Yan MC sample was generated via MadGraph5_amc@nlo with NLO precision in QCD and with up to 2 jets in the final state, applying a dimuon mass cut at the matrix element level in the range between 105 and 160 GeV.

For such studies, a Drell-Yan MC sample with a dimuon mass cut of 50 GeV at the matrix element level is used, again produced using MadGraph5_amc@nlo at NLO in QCD with up to 2 planes in the final state.

Table 5.1 provides the cross sections for each of the five main Higgs boson production modes at the LHC for a 125.0 GeV SM Higgs boson, along with the respective theoretical uncertainties, as recommended by the LHC Higgs Cross Section Working Group [40]

Physics objects

Primary vertex
Muons
Jets
Missing transverse momentum
Electrons
B-tagged jets
Track jets (additional soft hadronic activity)

To recover this loss in resolution, a procedure was developed to search for FSR photons inside the muon isolation cone and can be summarized as follows. T values are primarily calculated using the measured radius of curvature (𝑅) of the reconstructed muon track from hits in the inner tracker. Tmeasurement can be improved by including the position of the interaction point as an additional muon trace hit.

Trace jets consist only of charged traces originating from the primary interaction of the event.

Figure 5.2: Resolution, as a function of 𝑝

Corrections to data and simulation

Pileup re-weighting
L1 EGamma pre-firing corrections
Muon efficiency and trigger scale factors
Correction to Higgs boson transverse momentum

This led to an inefficiency in the T1 trigger decision; the trigger primitives generated by the ECAL deposits, belonging to the (𝑡 − 1)𝑡 ℎ cluster junction, resulted in inaccurate triggering of the 𝑡𝑡 ℎ cluster junction. Figure 5.4 shows the central maps of prefire probabilities (Pprefire(𝑗)) as a function of the pseudo-velocity and the electromagnetic component of the jet transverse momentum. The effectiveness of the muon identification and isolation requirements used in this analysis has been measured centrally by CMS, using a tag-and-probe technique on 𝑍+beam events in data and simulation.

T for the Higgs boson is corrected to the prediction from the NNLOPS generator, which is the highest order parton shower matched ggH simulation available that includes the finite top quark mass effects.

Figure 5.4: Centralized pre-firing probability provided per-jet as function of its electromagnetic transverse momentum and 𝜂 for 2016 (left) and 2017 (right).

Analysis strategy

Therefore, the identification and isolation scale factors are applied as a product of all selected muons. We also apply efficiency improvements for a muon that meets basic identification and isolation requirements to fire the single muon trigger. The Higgs boson signal produced by gluon fusion is simulated by Monte Carlo generators at NLO in QCD and then tuned to the parton shower.

Events that do not meet the requirements of the ttH or VH categories, but contain at least two hadron jets with 𝑝.

Figure 5.5: Graphical summary of the logical definition of each event category from the baseline dimuon selection

VBF category

VBF specific kinematic variables
Picking a discriminator – BDT vs DNN
The final Deep Neural Network for the VBF category
Systematic uncertainties
Fitting strategy
Results

The invariant mass of the two selected muons𝑚(𝜇 𝜇): this tends to reach close to the Higgs mass for the signal, while background processes have a more smoothly falling𝑚(𝜇 𝜇) spectrum. The minimum of the apparent speed difference between the selected jets (individually) and the dimuon system. To perform a data-driven fit to the 𝑚𝜇 𝜇 distribution (unlike the template-based fit to the DNN score output selected for . the VBF category), one must develop a 𝑚𝜇 𝜇 independent discriminator, to prevent the distribution of the backgrounds sculpting.

The signal strength is extracted by performing a built-in maximum likelihood fitting of the DNN distribution in the signal range.

Figure 5.6: Variable comparisons between VBF Higgs signal (blue) and Drell-Yan + Electro-weak Z backgrounds (orange)

ggH production
VH production
ttH production

The one (green) and two (yellow) standard deviation bands include the uncertainty in the background component of the fit. An overview of the selection criteria used in the W𝐻and𝑍 𝐻production categories is reported in Table 5.5. Finally, Table 5.6 reports the signal composition in the W𝐻 and 𝑍 𝐻 categories together with the hwhm of the expected signal shape.

Also given are the B ratios and the observation in the data within the hwhm of the signal peak.

Figure 5.33: The 𝑚 𝜇 𝜇 distribution for the weighted combination of VBF-SB and VBF-SR events

Results

p-values vs 𝑚 𝐻 scan
Limits on signal strength and 𝐻 → 𝜇 𝜇 branching ratio
Combination with CMS Run 1 results
Higgs couplings to muons

Table 5.8 gives the signal composition of each𝑡 𝑡 𝐻category, together with the hwhm of the expected signal shape. B ratios calculated within the hwhm of the signal peak, for each of the optimized event categories defined along the𝑡 𝑡 𝐻 hadronic and leptonic BDT outputs. The observed (expected for 𝜇=1) significance at 𝑚𝐻 =125.38 GeV of the incompatibility with the background-only hypothesis is 3.0 (2.5)𝜎. Fluctuations in the observed p-value of the VBF category. and for the combined fit is due to the nature of the signal extraction fit used in the VBF analysis.

The signal strength measured in the 𝐻 → 𝜇 𝜇 analysis cannot be directly translated into a measurement of the coupling of the Higgs boson to muons, because it is also sensitive to interactions between the Higgs boson and top quark/vector bosons.

Figure 5.39: Comparison between the observed data and the total background extracted from a signal-plus-background fit performed across 𝑡 𝑡 𝐻 hadronic (first row) and leptonic (second row) event categories

In Figure 6.10 we show the shape of Jet 2's regressive mass in the QCD background simulation as the BDT score requirements increase. The regressive mass distribution is shown in Figure 6.27 for different parts of the BDT score. 𝑇 𝐹𝑖(Jet 2𝑚 . reg) is a polynomial transfer factor of 𝑛-th order for the 𝑖-th passage area, as a function of the Jet 2𝑚.

The Jet 2 reduced the mass shape of the QCD multijet background in the different event categories (𝑝.

Figure 5.44: Higgs boson coupling modifiers. (Left) observed profile likelihood ratio as a function of 𝜅 𝜇 for 𝑚 𝐻 = 125

Nonresonant pair production of highly energetic Higgs bosons

Introduction

Higgs boson pair production (𝐻 𝐻) in SM provides unique sensitivity to explore the structure of the Higgs potential. The two leading diagrams of gluon fusion production (Fig.2.4) are known as the box (left) and triangle (right) diagrams. The box plot depends only on 𝑦𝑡 and therefore when the mass of the Higgs pair (𝑚𝐻 𝐻) exceeds twice the top mass threshold (2𝑚𝑡 ≈ 350 GeV), the 𝐻 𝐻 production probability increases from Fig. 6.1 same .

This chapter will be mainly focused on the 𝑔𝑔 𝐻 𝐻 analysis and will give a brief overview of the 𝑞 𝑞 𝐻 𝐻 analysis.

Figure 6.1: In the SM, the box (blue dashed line) and triangle diagram (red dashed line) for the 𝑔𝑔 𝐻 𝐻 process interfere destructively

Datasets and simulated samples

Datasets
Simulation overview

Therefore, with the requirement that the second 𝐻 be successfully tagged as a single large jet 𝐻 →𝑏 𝑏jet, about 103 background suppression is achieved without paying the price of percentage acceptance. Simulated samples for the ggF process are generated with next-order accuracy (NLO) using powheg 2.0, and corrected as a function of HH mass (𝑚𝐻 𝐻) based on Ref. Samples for other coupling combinations are constructed as linear combinations of the originally generated samples by applying appropriate event weights as described in Ref.

The differential 𝑡 𝑡 cross section as a function of the top quark 𝑝T has been corrected to NNLO QCD + NLO electroweak accuracy [214].

Physics objects

Muons
Electrons
AK4 jets
AK8 jets
Missing Transverse Energy (MET)

This technique provides a natural representation of the plane, unlike other algorithms that treat planes as images or ordered particles. We will use the mass distribution of one of the AK8 jets to derive the final signal strength in this analysis (explained later in Section 6.4.1), and therefore the sensitivity of this analysis is driven by the jet mass resolution. The performance of the smooth-fall measure algorithm is shown in blue and the regression measure algorithm is shown in black.

For 𝜏𝑁 ≫ 0, a large fraction of the beam energy is scattered away from the candidate subbeam, implying that the beam has at least N+1 subbeams.

Figure 6.5: The hadronization of a b quark results in a B meson, which travels 𝑂 (∼ 𝑚 𝑚 ) before decaying, resulting in a displaced secondary vertex (SV) inside the jet

The 𝑔𝑔 𝐻 𝐻 analysis

Event selection
Background estimation
Corrections to data and simulation
Results

System drag 𝑡 𝑡 is incorrectly modeled by the simulation in the relevant phase space for this analysis. System T As mentioned before, the variable used to derive the final signal is the regressed mass of the second plane.

If the F value for the data falls in the F distribution of the pseudo data set with a p value > 0.05, and the GOF test for .