The ATLAS Experiment at the Large Hadron Collider

ATLAS [1] is the largest volume detector ever con- structed for a particle collider and one of the four major experiments at the Large Hadron Collider (LHC) at CERN. It is a general-purpose particle physics experiment run by an international collaboration and is de- signed to exploit the full discovery potential of physics opportunities that the LHC provides. ATLAS has the di- mensions of a cylinder, 46m long, 25m in diameter, and sits in a cavern 100m below ground. The four major components of the ATLAS detector are the Inner Detector, the Calorimeter, the Muon Spectrometer and the Mag- net System. The Trigger and Data Acquisition System selects in real time physics events with distinguishing characteristics and ﬁnally the Computing System allows to store, process and analyse vast amounts of collision Figure 2: Cumulative luminosity versus day delivered to AT- data. A view of the ATLAS detector is shown in Fig. 1. LAS during stable beams and for high energy p-p collisions.

undergo a significant upgrade. The ATLAS group at the Sapienza Universitàdi Roma and at INFN Sezione di Roma (the ATLAS Rome Group) is engaged in a large number of data analyses. These include detector performances assessment, measurement of properties of the Higgs boson, precision measurements of Standard Model physics, and searches for exotic processes predicted by theories beyond the Standard Model. Currently the group is also captivated by the detector upgrades and physics studies needed to preserve the AT- LAS ability to successfully perform in the challenging Figure 1: A view of the ATLAS Experiment [1]. environment expected during the High Luminosity program over the coming decade of the LHC machine. The The aim of the ATLAS experiment is to push the fron- group is deeply involved in the construction of the pre- tiers of knowledge by seeking answers to essential ques- cision detectors of the muon spectrometer, in the design tions related to the fundamental forces of nature, the and realization of the muon trigger, the high level trig- composition of the basic building blocks of matter, the gers and the data acquisition system, and finally in the problem of Dark Matter and the underlying symmetries development of the computing system. of our universe. In the years 2015-2018 during the so called Run-II, References LHC has delivered proton-proton collisions at a center [1] ATLAS Collaboration, JINST 3, S08003 (2008). of mass energy of 13 TeV with an increasing luminosity 34 2 1 up to a value of 2 10 cm s a factor of 2 larger than ⇥ Authors: F. Anulli⇤,P.Bagnaia,M.Bauce,C.Bini, the project value. The ATLAS experiment has collected N. Buscino, A. Chomont⇤, M. Corradi⇤, S. De Cecco, 1 in these years an integrated luminosity of 160 fb .The D. De Pedis⇤, A. De Salvo⇤, C. Dionisi, A. Di Domenico, luminosities collected in the di↵erent years are shown in S. Falciano⇤, S. Francescato, G. Frattari, P. Gauzzi, S. Gen- Figure 2. Moreover LHC has delivered lead-lead colli- tile, S. Giagu, V. Ippolito, M. Kado, F. Lacava, I. Longarini, sions and proton-lead collisions at the maximum center C. Luci, L. Luminari⇤, L. Maiani, F. Marzano⇤,A.Nisati⇤, of mass energies ever reached. E. Pasqualucci⇤, E. Petrolo⇤,A.Policicchio,M.Rescigno⇤, In parallel to the data taking and the wide analy- S. Rosati⇤, L. Sabetta, F. Safai Tehrani⇤, C.D. Sebastiani, R. Vari⇤, D. Vannicola, S. Veneziano⇤ sis program, the ATLAS collaboration is carrying out ⇤Istituto Nazionale di Fisica Nucleare, Sezione di Roma. a program of detector upgrades in order to be prepared for the further increase in total energy and luminosity http://www.roma1.infn.it/exp/atlas/ in the forthcoming Run-III and Run-IV. In 2019 we are entering the so called Long Shutdown 2 (LS2). In this Subject Area two years period, both LHC and the experiments will Particle Physics

1 Precision Standard Model Measurements with ATLAS

Precise measurements of well-understood Standard Model (SM) processes at the LHC give the opportunity to test the internal consistency of the SM and to search for possible deviations originating from new phenomena. A prominent case is the production of lepton pairs, which can be calculated with small theoretical uncertainties thanks to the recent developments of perturbative QCD. Measurements of lepton-pair production have been per- formed by the ATLAS Collaboration in a wide range of di-lepton invariant masses and with data collected at dif- ferent center-of-mass energies [1–3]. To reach small sys- tematical uncertainties, these measurements rely on very accurate lepton reconstruction. The Rome group has an important role in the reconstriction of muons, with a leading role in the muon reconstruction software, in the calibration of the muon spectrometer, and in the muon Figure 1: ATLAS measurements of the mass of the W boa- trigger system. The ATLAS Collaboration has been us- son and of the top quark, compared to the result from elec- ing dilepton data to measure some of the foundamental troweak ﬁts to the Standard Model parameters [4]. parameters of the SM, such as the W -boson mass [4] and the electroweak mixing angle ✓W [5]. Figure 1 shows the ATLAS measurements of the W -boson mass, compared to indirect results from electroweak ﬁts. Further improvements are expected with the analysis of larger sets of data and from special runs dedicated to precision measurements. The ATLAS Rome group is currently working on an extension of the measurement of lepton-pair production towards small di-lepton invariant masses. Such a measurement will provide a unique test of the density of the constituents of the proton that carry a 4 very small fraction of the proton momentum (x 10 ). In this low-x regime, new QCD e↵ects are expected' beyond the standard Altarelli-Parisi evolution of the par- ton densities. These e↵ects may have a large impact in future collider at higher energy. Figure 2 shows a previous ATLAS measurement of the lepton-pair cross section at low mass [6] compared to QCD calculations. A large improvement in precision is expected from the ongoing analysis.

References Figure 2: Cross section for lepton pair production in pp [1] ATLAS Collaboration, Eur. Phys. J. C 77:367 (2017) . collisions at a center-of-mass energy of 7 TeV as a function [2] ATLAS Collaboration, JHEP 12 059 (2017). of the di-lepton invariant mass. Data are compared to QCD [3] ATLAS Collaboration, arXiv:1907.03567, submitted to calculations at the next-to-next-to-leading order (NNLO) [7]. Eur. Phys. J. C. [4] ATLAS Collaboration, Eur. Phys. J. C 78:110 (2018). [5] ATLAS Collaboration, “Measurement of the e↵ective Subject Area leptonic weak mixing angle using electron and muon pairs Particle Physics from Z-boson decay in the ATLAS experiment at ps =8 TeV”, ATLAS-CONF-2018-037. [7] ATLAS Collaboration, JHEP 06 112 (2014).

Authors The ATLAS Rome Group http://www.roma1.infn.it/exp/atlas

1 Studies of Higgs Boson Properties with the ATLAS detector

The discovery of the Higgs boson [1] is one of the most the LHC Run 2 data. It establishes that the Higgs boson prominent and landmark results of the first Run of the responsible for the electroweak symmetry breaking and LHC. The discovery was made through the three main the masses of the gauge bosons is also responsible for the decays of the Higgs boson to vector bosons (in the dipho- masses of the fermions. ton, the ZZ⇤ in the four leptons, and the WW⇤ in the The larger dataset and centre-of-mass energy delivered two leptons and two neutrinos final states). The dis- at the run 2 of the LHC, and the greatly improve pre- covery itself shed considerable light on the mechanism cision of theoretical predictions, have been key to reach responsible for the electroweak symmetry breaking. The a significantly improved precision on the measurement Rome group has significantly contributed to the discov- of the higgs bosons couplings in the combination of all ery and the first measurements of the properties of the Higgs boson measurement channels. These results are Higgs boson in the four-leptons channels [1]. illustrated in Figure [4]. In 2018, the LHC completed succesfully its second run 3 a higher centre-of-mass energy of 13 TeV, collecting a ATLAS Preliminary 1 σ interval 1 s = 13 TeV, 36.1 - 79.8 fb-1 dataset of approximately 150 fb . With the higher 2 σ interval m = 125.09 GeV, |y | < 2.5 H H 0.001 centre-of-mass energy and the larger dataset,8 a vast pro- 8 −1 −0.5 0 0.5 1 1.5 κ Z gram to measure with the highest possible7 precision the 7 κ W properties of the Higgs boson has been carried6 out. With 6 κ t first precision measurements of the coupling5 properties of 5 the Higgs boson, including studies from the four-leptons κ b 4 4 κ (ZZ⇤ channel) [2] and the direct observation of the cou- τ 3 3 pling of the higgs boson to third generation charged κ g 2 2 fermions [3]. κ γ ⏐κ ⏐ ≤ 1 1 1 B = 0 V BSM B 0 The Yukawa coupling of the Higgs boson to the top BBSM BSM ≥ quark is made through the pp ttH process [3]. Sub- −1 −0.5 0 0.5 1 1.5 −1 −0.5 0 0.5 1 1.5 stantial indirect evidence of the! coupling of the Higgs boson to top quarks is obtained from its main produc- Figure 1: ATLAS measurements of the Higgs boson cou- tion mode through the gluon fusion process (proceeding plings (to the Z and W bosons, the e↵ective coupling to predominantly through a loop of top quarks) and was photons and gluons, and the Yukawa couplings to the top, extensively studied through the discovery channels and b quark, and the tau lepton) under two assuptions: (i)-left that only Standard Model particles contribute to the total therefore also the ZZ four-elptons channel. ⇤ width of the Higgs boson, and (ii)-right that the coupling of The direct observation of the pp ttH production the Higgs boson to vector bosons cannot exceed 1. mode is very challenging, both due! its low cross section and the complex final state that is produced. The observation required the combination of various decay channels, including the diphoton, the b quark pair, the References tau pair, the W boson pair and the four-leptons. The Rome group has also significantly contributed to [1] ATLAS Collaboration, Observation of a new particle in the direct observation of this production mode and con- the search for the Standard Model Higgs boson with the AT- LAS detector at the LHC, Phys. Lett. B 716 (2012) 1, CMS sequently the direct evidence of the Yukawa coupling of Collaboration, Observation of a new boson at a mass of 125 the Higgs boson to top quarks in the Higgs boson decay GeV with the CMS experiment at the LHC, Phys. Lett. B mode with multiple leptons (electrons or muons) in the 716 (2012) 30. final state (originating from with the decays of W bosons [2] ATLAS Collaboration, Combined measurements of Higgs or tau leptons). The group is leading the analysis team boson production and decay, ATLAS-CONF-2019-005. in charge of one of the most sensitive channels in this [3] ATLAS Collaboration, Observation of Higgs boson pro- search: the final state topology where the higgs boson duction in association with a top quark pair, Phys. Lett. B decays to a pair of taus or W bosons with two same- 784, 173 (2018). sign electrons or muons and one reconstructed tau. This [4] ATLAS Collaboration, Measurement of the Higgs boson topology of events is extremely complex as in addition coupling properties in the 4` decay channel, JHEP 03 (2018) to the three aforementioned leptons, events are required 095. Authors to have four jets among which two are required to be The ATLAS Rome Group tagged as containing at least one B hadron. http://www.roma1.infn.it/exp/atlas The observation of the third generation Yukawa cou- Subject Area plings is a major and fundamental result obtained with Particle Physics

1 Searches for Dark Matter and Invisible Higgs decays

During the Run 2 of LHC operation, at the collision [2]. energy in the center of mass of 13 TeV, the ATLAS ex- The searches mentioned above, as wall as other ones periment recorded a high-statistics dataset of proton- has been combined in a publication providing the best 1 proton collisions, corresponding to 139 fb .Thisisconstraints on di↵erent DM simpliﬁed model (Fig.1) [3]. currently scrutinized with a wide program of searches for evidence of new phenomena beyond those predicted 1.6 Dijet Dijet s = 13 TeV, 37.0 fb-1

[TeV] PRD 96, 052004 (2017) by the Standard Model (SM). Two particular topics in χ ATLAS 1.4 Dijet TLA s = 13 TeV, 29.3 fb-1 m PRL 121 (2018) 0818016 which the Rome group focused are the searches for ev- Dijet + ISR s = 13 TeV, 15.5 fb-1 idences of Beyond the Standard Model (BSM) provid- 1.2 Preliminary ATLAS-CONF-2016-070 Z' A = m 2 = 0.12 tt resonance χ m h -1 ing a candidate for Dark Matter (DM) and the search × Ω c s = 13 TeV, 36.1 fb 1 2 for anomalies of the SM in the Higgs boson sector, also EPJC 78 (2018) 565 Dibjet 0.8 Thermal Relictt resonance through its decay into invisibile particles. s = 13 TeV, 36.1 fb-1 PRD 98 (2018) 032016

0.6 miss Most of new phenomena initiated by proton collisions ET +X Dibjet miss -1 ET +γ s = 13 TeV, 36.1 fb Dijet Eur. Phys. J. C 77 (2017) 393 must couple any newly produced particle with the con- 0.4 miss -1 ET +jet s = 13 TeV, 36.1 fb JHEP 1801 (2018) 126 miss miss E +Z(ll) s = 13 TeV, 36.1 fb-1 stituent partons of the proton, and thus can produce par- E +X Axial-vector mediator, Dirac DM T 0.2 T PLB 776 (2017) 318 miss Perturbative Unitarity g = 0.25, g = 0, g = 1 E +V(had) s = 13 TeV, 36.1 fb-1 tons, which, after showering and hadronization, manifest q l χ T All limits at 95% CL JHEP 10 (2018) 180 as collimated jet of particles in the ﬁnal state. Signature 0 0.5 1 1.5 2 2.5 3 3.5

mZ' [TeV] based on the reconstruction of hadronic jets in the de- A tector can be quite sensitive to new physics evidence. Exotics particles produced in the collisions, can, on Figure 1: Constraints from di↵erent searches on an Axial- the other side, be weakly interacting with the ordinary Vector mediated DM simpliﬁed model, as a function of the matter, therefore resulting as invisible in the detector, main model’s parameters [1]. but whose existence can be inferred from an unbalance in the detector transverse plane energy distribution, called Dark matter particles, if suciently light, may be pro- Emiss. duced in decays of the Higgs boson, therefore searches T for the invisible decays of the SM Higgs boson are car- The search for DM at colliders relies on the Weakly ried out. Given the invisible decay of the Higgs boson, Interacting Massive Particle (WIMP) paradigma, which the detectable signature of such process must be char- assumes that in the simplest SM extensions the DM can- acterized by the presence of visible elements, therefore didate particle is produced via a massive mediator (such the Higgs boson production through vector boson fu- as a Z ) which couples to it and to SM particles. Two 0 sion or in association with a massive boson W/Z are complementary approaches can probe such BSM models: exploited. A combination of searches in the di↵erent the search for such mediator particle as a new resonance production modes has been published, providing con- in di-jet events or the search for the decay to DM candi- strains on the Higgs branching ratio to invisible particle date in events with a large amount of Emiss and a single T [3]. The data collected during the LHC run at the cen- hadronic jet radiated from the initial state (ISR) of the ter of mass energy of 7, 8, and 13 TeV, corresponding to collision. 1 a total amount of 61 fb excluded the invisible Higgs Searches for new resonances decaying to two jets have decays with a branching ratio of 0.26, or larger, at 95% been carried out in a wide range of mediator masses, conﬁdence level. considering also events where an additional ISR jet is present or where the two jets from the resonance are merged in a single, wider, object. Given no deviation with respect to the smooth SM background prediction, interpretations in terms of exclusion for di↵erent BSM References models, besides DM mediators are given: excited quarks, [1] M. Aaboud et al.,Phys.Rev.D 96 052004 (2017). W 0,Z0, excited bosons W ⇤, quantum black holes [1]. [2] M. Aaboud et al.,JHEP 01 126 (2018). Complementary searches focus on the decay of the [3] M. Aaboud et al.,Phys.Rev.Lett.122 231801 (2019). aforementioned mediators to undetected WIMPs in as- Authors sociation with a single visible object in so-called mono-X N. Template, E. Fermi, E. AmaldiX signatures. One of the most sensitive searches is considering events with one energetic jet and classifying them group.webpage.if.any.it miss based on their ET . Good agreement is observed between events in data and SM predictions, and thus ex- Subject Area clusion limits are placed on di↵erent DM-related models Particle Physics

1 Search for Dark Matter using Long Lived Particles with the ATLAS Experiment

A class of theories beyond the Standard Model, col- and this is well matched by the high-granularity mea- lectively called Dark Sector models, can lead to the pro- surement capability of the ATLAS muon spectrometer. duction of unusual signatures in detectors at the Large The ATLAS collaboration searched for lepton jets sig- Hadron Collider (LHC), that may include long-lived col- natures using data collected from proton-proton colli- limated jets of displaced leptons or hadrons. These sig- sions at LHC at ps = 13 TeV corresponding to an in- 1 natures allow to evade the current stringent constraints tegrated luminosity of 36.1fb [1], without ﬁnding so of Standard Model extensions based on more conven- far any excess over the expected background. Figure 1 tional decays, and at the same time can provide a viable shows the results interpreted in the context of the Vector solution to important unanswered questions in cosmol- portal model as exclusion contours in the kinetic mixing ogy and particle physics these days: the Dark Matter parameter ✏ vs d mass plane, where it is possible to ap- problem, the anomalous magnetic moment of the muon, preciate the complementarity between the displaced and the asymmetries in the electron and positron ﬂuxes mea- the prompt ATLAS searches. sured in satellite experiments, and the recently reported The enormous amount of data that will be collected 8 1 anomaly in Be nuclear decays. by ATLAS during the Run-3 (300fb ) and High- Dark Sector models hypothesize the existence of a hid- Luminosity (3000fb1) 14 TeV LHC phase, and the up- den sector that is weakly coupled to the visible one. De- dated ATLAS detector setup, will o↵er a unique oppor- pending on the structure of the hidden sector and its cou- tunity to probe unexplored regions of phase space in the pling to the Standard Model, some light unstable neutral context of such searches. The sensitivity prospects for hidden states called dark photons (d)maybeproduced Run-3 and HL-LHC [2] have been estimated, extrapolat- at colliders, for example via Higgs boson decays. The ing the results of the Run-2 search, and here presented dark photon mixes kinetically with the SM photon and in Figure 2. decay back into SM leptons and light quarks with long lifetime. For a small kinetic mixing (✏) value, the d has a long lifetime, so that it decays at a macroscopic distance from its production point . Due to the small mass, the dark-photons are typically produced with a large boost producing collimated jet-like structures containing pairs of electrons and/or muons and/or charged pions collectively called “lepton jets”. The lepton-jet sig-

ATLAS Simulation Preliminary

Figure 2: A two-dimensional exclusion plot in the dark-

photon mass md and the kinetic mixing ✏ parameter space, showing the expected excluded regions by ATLAS after Run- 3andHL-LHC[2].

References [1] ATLAS Collaboration, ”Search for light long-lived neutral particles produced in pp collisions at ps =13TeV and decaying into collimated leptons or light hadrons with the ATLAS detector”, EXOT-2017-28. Figure 1: A two-dimensional exclusion plot in the dark- [2] ATLAS Collaboration, ”Search prospects for dark-photons photon mass m and the kinetic mixing ✏ parameter space, d decaying to displaced collimated jets of muons at HL-LHC”, showing the regions excluded by ATLAS [1]. ATL-PHYS-PUB-2019-002. nature represent a challenge both for the trigger and for Authors the reconstruction capabilities of the LHC detectors. In The ATLAS Rome Group the absence of information from the inner tracking sys- http://www.roma1.infn.it/exp/atlas/ tem it is in fact necessary to use the muon spectrometer for the reconstruction of tracks which originate from a Subject Area secondary decay far from the primary interaction vertex, Particle Physics

1 Artiﬁcial Intelligence applications in the ATLAS Experiment

Artificial Intelligence and representation learning In Figure 1 an example of the ROC curve (QCD multi- emergence in the recent years allowed for machine learn- jet background rejections VS jet from dark photon de- ing tools which could adeptly handle higher-dimensional cays signal eciency) obtained by the ATLAS group and more complex problems than previously feasible. in Rome in developing Convolutional Neural Network Traditionally the main problem of the analysis of HEP (CNN) in FPGAs for the Phase-2 Level-0 Muon Trig- data, characterised by large volumes and high dimen- ger of the ATLAS detector. The upper figure shows the sionality, is approached by dimensionality-reduction of hits pattern of the signal released by a muon crossing an data based on a series of analysis steps that operate ATLAS Resistive Plate Chamber detector. The pattern both on individual collision events and on collections can be interpreted as an image and analysed by a CNN of events. Machine learning, and Deep learning in par- trained to identify the muon and to measure its parame- ticular, provides an extremely powerful method to con- ters. The bottom figure shows the eciency as a function dense the relevant information contained in the low-level, of the transverse momentum of the muon obtained with high-dimensional data into a higher-level and smaller- a ternary CNN that represents an optimal solution for dimensional space, and can provide the needed ingredi- FPGA synthesis, given its low precision weight structure ent to overcome the limits of the traditional approach. (weight can in fact assume only -1,0,1 values). The Rome group of the ATLAS experiment is actively involved in the design and development of both state of the art and novel deep learning models for the feature extraction, simulation and analysis of LHC data. This range from developing novel low-precision ternary and quantised deep neural network to run in real time on FPGAs for next generation fast triggers for HL-LHC, to the study of convolutional neural networks, genera- tive adversarial networks and variational auto-encoders for physics object reconstruction to improve the discovery sensitivity of the ATLAS experiment for new physics beyond the Standard Model. Figure 2: ROC curve obtained with a novel CNN in com- parison with a BDT [2].

In Figure 2 the result of the application of deep learning techniques for the o✏ine classiﬁcation of jets from decays of long lived particles (dark photons) with respect conventional jets from QCD production is shown. Exploiting a deep CNN trained to identify substructures and low level features in the energy deposits in the AT- LAS calorimeters, improvements by more than a factor two in the rejection of background have been achieved with respect to the Boosted Decision Trees algorithm used in the previous generation of analysis [2].

References [1] ATLAS Collaboration, ”Fast Deep Learning for Phase-II L0 Muon Barrel Trigger”, ACAT 2019 proceedings. [2] ATLAS Collaboration, ”Search for light long-lived neutral particles produced in pp collisions at ps =13TeV and decaying into collimated leptons or light hadrons with the ATLAS detector”, EXOT-2017-28.

Authors The ATLAS Rome Group Figure 1: (top) Hits pattern of a muon particle crossing an http://www.roma1.infn.it/exp/atlas/ ATLAS Resistive Plate Chamber detector. (bottom) E- ciency curve (with a threshold at 10 GeV) obtained with a Subject Area ternary CNN [1]. Particle Physics

1 The Level-1 Barrel Muon Trigger of the ATLAS experiment at LHC

The trigger system of any collider experiment is the es- to 2018) the group successfully commissioned the new sential component responsible of deciding whether or not additional trigger regions and continued to be in charge to keep a given bunch crossing interaction data for the of the maintenance and operation of the full system, in- o✏ine analysis studies. The presence of prompt muons in cluding RPC timing calibrations and eciency studies. the final state is a distinctive signature for many physics A stable and smooth data taking was guaranteed dur- processes in high energy proton collisions at the LHC, ing the full Run-2 operations. The current system will therefore a high-performance muon trigger is essential. be continued to be operated by our group also in Run-3 The current ATLAS trigger system is made of a first (2021-2024), when the LHC luminosity will be twice the 34 2 1 hardware based system, called Level-1, and a second nominal luminosity of 10 cm s . software based system, called High Level Trigger (HLT). Run-4 (from 2026, the so-called· High Luminosity The Level-1 provides 100 kHz data to the HLT, which LHC) will be characterised by an increase of luminosity then selects about about 1.5 kHz of events recorded for of about a factor of 5. These high demanding conditions the o✏ine analysis. About 20 kHz Level-1 data are allo- impose to completely replace the current Level-1 trigger cated for muon triggers, and about 150 Hz are the muon system. Figure 2 shows the scheme foreseen for the Run- events selected by the HLT. 4 new Level-0 barrel muon trigger. The on-detector Data Muons are identified at Level-1 in the barrel region Collector Transmitter boards (DCT) sample and time ( ⌘ < 1) by the spatial and temporal coincidence of hits tag the RPC hits and send the data to the o↵-detector in| the| Resistive Plate Chamber (RPC) detectors point- Sector Logic boards (SL) which perform the coincidence ing to the beam interaction point, as shown in Figure 1. based Level-0 trigger algorithm. The foreseen Level-0 The low-pT trigger requires a coincidence in the middle trigger rate is 1 MHz, 50 kHz of which will be allocated RPC layers while the high-pT trigger requires a further for muon events. The Roma group is responsible of the coincidence of hits in the outer RPC layer. The degree full upgrade of the Level-0 barrel muon trigger system, of deviation from the hit pattern expected for a muon and it is supposed to play a leading role in the whole with infinite momentum is used to estimate the pT of trigger system as it did in the past. the muon. ATLAS cavern USA15 Counting room

DCT MDT BO RPC RPC 3 high pT RPC 2 (pivot) Barrel MDT DCT Sector MUCTPI RPC 1 low p BM RPC Logic

T MDT

MDT BI RPC DCT MDT Trigger Processor tile

MDT calorimeter

0 5 10 15 m Figure 2: The Run-4 Level-0 muon trigger in the barrel.

Figure 1: The current and Run-3 ATLAS Level-1 muon trigger in the barrel region ( ⌘ < 1). References | | [1] F. Anulli et al.,”The Level-1 Trigger Muon Barrel System of the ATLAS Experiment at CERN”,JINSTVolume:4 The ATLAS-Roma group was responsible of the de- Article Number: P04010 (2009). sign, realisation and installation of the Level-1 barrel [2] The ATLAS Collaboration, ”ATLAS Trigger and trigger system and was in charge of its commissioning Data Acquisition Phase-II Upgrade Technical Design and data taking in the early years of LHC functioning Report”, CERN-LHCC-2017-020; ATLAS-TDR-029; (Run-1, 2011 to 2013) [1]. During the Long-Shutdown-1 https://cds.cern.ch/record/2285584. (LS1, from mid 2013 to mid 2015) the group successfully installed the completion of the barrel trigger system in Authors the lower part of the spectrometer, which makes use of ATLAS Rome Group: www.roma1.infn.it/exp/atlas new additional RPC chambers in the feet region and in the elevators region, to recover respectively about 3% Subject Area Particle Physics and 0.8% geometrical acceptance. During Run-2 (2015

1 The New Small Wheel and the Micromegas chambers for ATLAS

In the High Luminosity LHC protons will collide at by a thin metal grid (micro-mesh) from a 128 µmthick a14 TeV center of mass energy at a luminosity in the multiplication gap. Electrons drift to the mesh and 34 2 1 range 5 7 10 cm s . Such scenario is particu- enter in the multiplication gap where in a strong 40 larly demanding÷ ⇥ for the detectors in the forward regions kV/cm field they produce avalanche collected on resis- of the ATLAS experiment where large fluxes of particles tive microstrips. The signals are readout by capacitively are expected. coupled metallic microstrips below an insulating layer At present in the pseudorapidity range 1.3 < ⌘ < 2.7, underlying the resistive microstrips. the first stations (‘Small Wheel’) for the muon detection| | at 7 m from the interaction point, are composed by The first New Small Wheel (the NSW-A) will be in- TGC± (Thin Gap Chamber) chambers for the trigger and stalled in ATLAS before the end of the Long Shutdown CSC (Cathode Strip Chamber) and MDT (Monitored 2 (LS2), that is within 2020, while the second Wheel Drift Tube) detectors for the position measurement (see will be installed in a forthcoming shutdown. Seven Figure 1). These detectors are not adequate for the new italian groups from the universities and Sezioni INFN of Cosenza, Lecce, Napoli, Pavia, Roma Sapienza, Ro- maTRE and Laboratori Nazionali di Frascati are com- mitted to build 32 large trapezoidal chambers about 2.2 m 1.3 m, that is one quarter of the MM NSW chambers⇥ and to design and realize the 32 Trigger Pad Logic Boards used to find the tracks for the trigger in the New Small Wheel. Within this project the Sapienza group has already delivered 96 drift panels that are being used for assembling the chambers. At the moment 10 out of 32 chambers have been delivered to CERN and are now under integration. Four of them are shown in Fig. 2 It is planned to complete the production of the 32 chambers by the summer of 2020.

Figure 1.2: AFigure z-y view 1: of 1/4The of location the ATLAS of detector. the Small The blueWheel boxes detectors indicate the in end-cap the Monitored DriftATLAS Tube Experimentchambers (MDT) (x-y and view the yellow of one box quarter in the Small of the Wheel detector). area the Cathode Strip Chambers (CSC). The green boxes are barrel MDT chambers. The trigger chambers, Resistive Plate chambers (RPC) and Thin Gap Chambers (TGC), are indicated by the outlined white andhigh the magenta rates of boxes. particles This is a cut-out expected on the during muon spectrometer HL-LHC. at the Fur- large sectors, hence thethermore names ‘End-cap a trigger Inner Large’ station (EIL), added ‘End-cap in Middle the position Large’ (EML) of and the ‘End-cap Outer Large’ (EOL). The detector regions of the Small Wheel and Big Wheel are also outlined. Small Wheel, and able to track particles with a O(mrad) resolution, will reduce the trigger rate due to fake muons. Both of these two issues represent a serious limitation on the ATLAS performance beyond design luminosity: reducedThe acceptance ATLAS of experiment good muon tracking, is presently and an unacceptable building a rate new of fake high pT Level-1 muondetector, triggers coming called from New the forward Small direction. Wheel (NSW) [1], with In order to‘small-strip solve the two problems Thin Gap together, Chambers’ ATLAS proposes as main to replace trigger the device present muon Small Wheels with the ‘New Small Wheels’ (NSW). The NSW is a set of precision tracking and trigger detectors ableand to work ‘Micromegas’ at high rates chambers with excellent (MM) real-time to spatial measure and timethe resolution.po- These detectors cansition provide of the the muon particles. Level-1 trigger To system achieve with a online resolution track segments in mo- of good angular resolution tomentum confirm that better muon tracks than originate 15% for from 1 the TeV IP. In muons, this way the the end-cap posi- fake triggers will be considerablytion of reduced. a track With before the proposed the Endcap NSW the Toroidal ATLAS muon magnets system will in maintain the full acceptancethe of forward its excellent regions, muon tracking has atto the be highest measured LHC luminosities with a res- expected. At the same time the Level-1 low pT (typically pT > 20 GeV) single muon trigger rate will beFigure kept at an 2: Four Micromegas chambers built from the italian acceptable level.olution of about 50 µm. The resolution in position collaboration The ⌘ coverageon the of the single proposed hit NSW in the (and detectors the existing has Small to Wheel) be of is the1.3 or-< ⌘ < 2.7. The | | remaining region of the inner station 1.0 < ⌘ < 1.3 is covered by the existing EIL4 detectorsReferences of der of 100 µm and a similar| | accuracy is demanded the current muonin the end-cap assembly system. of The the TGC components chamber in EIL4 of the will bechamber. used to provide[1] a ATLAS rough Collaboration - New Small Wheel - Technical confirmation that a particle has traversed the end-cap toroid zone, reducing the fakeDesign end-cap Report, CERN-LHCC-2013-006 / ATLAS-TDR-020. triggers in this region. The EIL4 chambers however only cover about 50% of the full azimuthal The ATLAS-Sapienza group is since several years angle, while the rest of the space taken by the barrel toroid coils. A plan for a small scaleAuthors upgrade is being studied to fill the uncovered region. involved in the realization of the Micromegas chambers. The ATLAS Rome Group Prior to theThis installation is a Micro of the NSW Pattern in 2018, Gas the Detectorexisting TGC built chambers with of the the Small Wheel and EIL4 will be integrated in the end-cap trigger system during LS1 to reduce fakehttp://www.roma1.infn.it/exp/atlas/ triggers within a limitedmodern acceptance photolithographic1.0 < ⌘ < 1.9. Due technology. to the small number In this of layers detector (2 layers/chamber) | | and coarse spatialthe charged segmentation particles of the existing crossing detector, the only detector a rough ionize hit position the canSubject be used Area gas in a conversion/drift region of a few mm, separated Particle Physics

7 1 The Computing System of the ATLAS Experiment at the LHC

The ATLAS Computing System[1] is responsible for the provision of the software framework and services, the data management system, user-support services, and the world-wide data access and job-submission system. The development of detector-specific algorithmic code for simulation, calibration, alignment, trigger and reconstruction is under the responsibility of the detector projects, but the Software and Computing Project plans and coordinates these activities across detector bound- aries. In particular, a significant e↵ort has been made to ensure that relevant parts of the o✏ine framework and event-reconstruction code can be used in the High Level Trigger. Similarly, close cooperation with Physics Coor- dination and the Combined Performance groups ensures the smooth development of global event-reconstruction code and of software tools for physics analysis. Italy provides facilities to the ATLAS collaboration. Figure 1: Number of parallel jobs running in the ATLAS The Tier-1, located at CNAF, Bologna, is the main cen- Italian Tier1 and Tier2’s. tre, also referred as regional centre. The Tier-2 centres are distributed in di↵erent areas of Italy, namely in Fras- cati, Napoli, Milano and Roma La Sapienza. infrastructure in 2000. The members of the group have been pioneers of many subsystem, including the creation The computing activities of the ATLAS collaboration of the ATLAS Virtual Organisation [2], the distribution have been constantly carried out in 2018 and later, in of the software and its evolutions to CVMFS [3,4], the order to finalize the analysis of the data of the Run- access to the condition database via the Frontier System 2, produce the Monte Carlo data needed for the 2018 [5] and the remote calibration of the Muon Detectors run and produce the data for the upgrade studies. In (Remote Calibration Centers) [6]. The ATLAS group this period, the Tier1 and the four Tier2’s, have been operates one of the most ecient Tier-2 infrastructures involved in all the computing operations of the collab- in the ATLAS Grid and hosts some of the Central Ser- oration: data reconstruction, Monte Carlo simulation, vices, like the Installation System services. Currently user and group analysis and data transfer among all the the group is starting a new project called Harvester to sites. Besides these activities, the Italian centers con- extend the optimisation of the ATLAS resources, by tributed to the upgrade of the Computing Model both creating a new resource-facing service, to be plugged in from the testing side and the development of specific the global production system called PANDA [7]. working groups. Several improvements in the Comput- ing Model has been achieved in 2018 and the first part of 2019, more precisely in the software domain and the in- References frastructure. The use of the grid in 2018/2019 has been [1] The ATLAS Computing Technical Design Report stable on 320k simultaneous jobs, with peaks around the ATLAS-TDR-017; CERN-LHCC-2005-022, June 2005. conferences periods above 500k, showing the reliability [2] I Bird et al., Computer, 42, 36-46, January 2009, doi:10.1109/MC.2009.28. and e↵ectiveness of the use of grid tools. [3] J. Blomer et al., The CernVM File System, Technical The contribution of the Italian sites to the computing Report; CERN. activities in terms of processed jobs and data recorded [4] A. De Salvo et al, Journal of Physics: Conference Series, has been of about 9%, corresponding to the order of the Volume 396, Part 3. resource pledged to the collaboration, with very good [5] D. Dykstra et al., J. Phys.: Conf. Ser. 219 072034. performance in term of availability, reliability and e- [6] T. Dai et al., Journal of Physics: Conference Series 219 ciency. All the sites are always in the top positions in (2010) 022028. the ranking of the collaboration sites. Figure 1 shows the [7] T. Maeno et al., 2008 J. Phys.: Conf. Ser. 119 062036. number of parallel jobs in the Italian Computing System Authors of ATLAS from 2018 to 2019. The ATLAS Rome Group The ATLAS group at the Sapienza Università di http://www.roma1.infn.it/exp/atlas/ Roma and at INFN Sezione di Roma has been heavily involved in barely all the Computing activities of the Subject Area ATLAS ecosystem, since the beginning of the Grid Particle Physics