Tags:

ATLAS prepares for High-Luminosity LHC

2 April 2025 | By

In 2025, the global particle physics community turns its attention to Europe as it embarks on the third update of the European Strategy for Particle Physics. This follows the initial strategy process launched under the CERN Council in 2005–2006, and subsequent updates in 2012–2013 and 2018–2020. While discussions will focus on future projects, such as the next-generation high-energy collider, it is equally essential to reaffirm the critical importance of the upgrades of the LHC and its experiments for the High-Luminosity LHC (HL-LHC), approved to operate from 2030 to 2041. The HL-LHC is a large-scale effort to probe the structure of matter and its interactions more deeply than ever before. It is the highest mid-term priority for particle physics, as emphasised by the 2023 P5 report in the U.S. and the 2020 European Strategy update.

Simulated Events,Event Displays,Physics,ATLAS
Display of a simulated ttbar event with µ = 200 (1600 tracks) in the ATLAS ITk. (Image: ATLAS Collaboration/CERN)

To contribute to this update, the ATLAS Collaboration submitted four documents to the European Strategy Group (ESG), namely, an overview of the extensive ATLAS detector upgrade for the HL-LHC, a summary of the software and computing preparations to address the HL-LHC challenges, a joint update with the CMS Collaboration on HL-LHC physics expectations, and a report on perspectives in heavy flavour physics in collaboration with Belle II, CMS, and LHCb.

Through enhancements in beam intensity, emittance and final focus, the HL-LHC is expected to achieve a levelled instantaneous proton–proton luminosity of up to 7.5×10–34 cm–2s–1, leading up to 200 simultaneous inelastic collisions per bunch crossing (also known as "pileup") and a total delivered integrated luminosity of 3,000 fb–1 by 2041 – including about 520 fb–1 recorded during the LHC Runs 1–3. With this performance, the LHC and HL-LHC serve as a heavy-particle factory, producing in ATLAS alone 180 million Higgs bosons, 120 thousand Higgs boson pairs, 6.7 billion top quarks, 130 billion (12 billion) W (Z) bosons decaying to electrons or muons, and 360 million W bosons pairs. These quantities provide unparalleled physics potential. In many ways, they complement the cleaner but less abundant production of these particles (except for the Z boson) at a future e+e collider, where the initial state is precisely defined.

An ambitious ATLAS detector upgrade

The harsh conditions at the HL-LHC demand unprecedented detector and computing technologies, including radiation-hard components, high detection granularity and resolution, precise timing for improved pileup resilience, more selective triggers, higher-bandwidth data acquisition, deeply embedded machine learning, and high-performance heterogeneous software and computing.

To meet these challenges, ATLAS is pursuing an ambitious high-technology upgrade programme. It includes at its heart a new all-silicon based inner tracker (ITk) featuring the world’s largest silicon pixel detector with an active area of 13 m2, five billion readout channels, an innermost layer at a radius 33 mm from the beam, and extending forward coverage up to pseudorapidity of 4 (corresponding to a 2-degree polar angle). The pixel detector is surrounded by a 165 m2 silicon strip detector with active layers up to 1-metre in radius and 60 million readout channels, ten times that of the current ATLAS silicon strip detector.

ATLAS members visiting HL-ATLAS upgrade sites at CERN in October 2024. (Images: H.P. Beck & S. Guindon/ATLAS Collaboration, C. Krishna/CERN)

Additionally, a high-granularity timing detector based on newly developed 50 µm silicon sensor technology with fast signal rise is inserted behind the ITk into the forward region of ATLAS. It will provide 30–50 ps timing resolution per track. New resistive plate chambers (RPC) in the inner layer of the barrel muon spectrometer, partly accompanied by new small-diameter monitored drift tubes, improve the fast online muon acceptance. Comprehensive upgrades of the real-time event selection system (trigger) will enable ATLAS to maintain low-threshold detection and open up new physics opportunities. It operates with a first-level acceptance rate of 1 MHz, utilising global event selection through programmable hardware, followed by a second-level selection at 10 kHz output rate, implemented via algorithms running on a computing farm. The on-detector and off-detector electronics in the calorimeters and muon systems are being upgraded to accommodate the higher trigger rates and enhance overall performance. Smaller upgrade projects aim at improving the luminosity measurement and the detection of neutral particles along the beam line.

A new cooling system for the silicon detectors developed by CERN together with ATLAS and CMS, along with RPC leak repairs, consolidation, and an improved gas mixture, as well as the replacement of the transition radiation tracker will significantly reduce greenhouse gas emissions during HL-LHC operation.

The successful construction, installation, and commissioning of these upgrades engage thousands of ATLAS members worldwide and are the collaboration’s highest priority.

Rising to the computing challenge

ATLAS operates one of the world’s largest distributed scientific computing systems, with over a million computing cores at peak capacity and more than one exabyte of storage spread across more than 100 sites worldwide. This system must evolve to address the software and computing challenges of the HL-LHC era, focusing on scalability, efficient hardware utilization and adaptation, and the long-term sustainability of the workforce. Investments by the collaboration in software optimisation, heterogeneous computing and data management will not only benefit ATLAS but also shape future collider experiments.

At the HL-LHC, data volumes and processing rates are expected to grow 3 to 5 times, requiring more efficient storage and data management solutions. To prepare for this, network stress tests are already underway, with full-scale evaluations planned by 2029. While high-performance computing centres offer significant computational power, they face storage limitations. To address this, ATLAS is collaborating with the broader HEP software community to develop resource-efficient data formats and compression techniques that could reduce disk usage by 10–50%. Alternative computing resources are also being explored, including commercial cloud services, which provide greater flexibility but at a higher cost, and the integration of trigger computing farms for offline data processing. To boost computing performance, hardware accelerators – especially GPUs in AI-driven computing centres – are being integrated.

ATLAS is modernising metadata and data management by transitioning to a more flexible and interoperable conditions data infrastructure, making database interactions more efficient.

Significant progress has been made in Geant4 based full simulation, which now runs twice as fast as in Run 2, and is complemented by high-fidelity fast simulation employing generative machine learning. Event generation has also been accelerated by up to a factor of 40. Machine learning techniques are enhancing charged particle tracking and calorimeter clustering, further optimising resource efficiency.

A new physics landscape

The 3,000 fb–1 integrated proton–proton luminosity of the HL-LHC offer an extraordinarily rich spectrum of physics opportunities, particularly for the exploration of rare processes. ATLAS and CMS have updated several of their HL-LHC physics projections for this year’s European Strategy update.

ATLAS
Projected ATLAS and CMS combined uncertainties of the Higgs to particle coupling strength modifiers (measured relative to predicted coupling strength). (Image: ATLAS Collaboration/CERN)

Higgs boson physics remains a central focus of the HL-LHC physics programme. Assuming Standard Model values and a realistic – though likely conservative – systematic uncertainty evolution, both rare Higgs decay channels into µµ and Z𝛄 will be measured with a combined ATLAS and CMS cross-section uncertainty of 6% and 14%, respectively.

Higgs boson couplings to Standard Model particles will be measured with remarkable experimental precision ranging from 0.9% to 1.2% (1.7% for the bottom quark) for the W and Z bosons, photon, gluon, tau lepton, and top quark, when fixing the Higgs boson width to its predicted value. These measurements will be limited by theoretical uncertainties, currently projected between 1.3% and 3.2%, highlighting the need for theoretical developments to match the experimental precision.

Not updated in this edition, earlier projections indicate that the Higgs boson width can be determined via off-shell production in the four-lepton channel with better than 20% precision, and that the branching ratio to invisible decays (such as those involving dark matter) can be constrained to 2.5%.

Di-Higgs production, with a predicted rate of less than a permille of single Higgs production, may be individually observed by both ATLAS and CMS, with their combined sensitivity exceeding 7σ. The measurement of the triple Higgs coupling or Higgs self interaction – a fundamental parameter of the Brout-Englert-Higgs (BEH) potential and predicted to be 0.13 in the Standard Model – is expected to achieve better than 30% precision. Any deviation greater than about 75% from the predicted central value would allow ATLAS and CMS to exclude the Standard Model prediction, offering a crucial and unprecedented test of the Higgs sector.

ATLAS
Constraints on the triple Higgs coupling modifier. (Image: ATLAS Collaboration/CERN)

ATLAS and CMS have also investigated their sensitivity to the nature of the electroweak phase transition during the early universe. A strong first-order, that is non-continuous, phase transition would enable the nucleation and expansion of “bubbles” containing the broken phase within the surrounding gauge-symmetric phase. New CP-violating interactions at the bubble walls could potentially generate baryogenesis, as baryon- and lepton-number-violating sphaleron processes would be prevented from reaching equilibrium due to the rapid bubble expansion. However, in the Standard Model, the relatively large Higgs boson mass leads to a continuous phase transition, ruling out this mechanism. Analyses of several well-motivated extensions of the BEH energy potential indicate that, with a dataset of 3,000 fb–1, ATLAS and CMS together may have the sensitivity to exclude all these scenarios. This conclusion remains robust even in the presence of an additional scalar singlet field.

ATLAS and CMS have also updated their projections for W and Z boson scattering, a crucial test of the BEH mechanism. The Higgs boson regulates the interaction of longitudinally polarized W and Z bosons at high energy — polarization states that exist only due to the BEH mechanism generating the W and Z masses — thereby ensuring unitarity. The new analysis projects joint longitudinal W polarization can be observed above 5σ, with a cross-section measured to better than 20% precision.

ATLAS and CMS have further explored their measurement potential for rare top-quark processes, including four-top production and top–antitop production in association with a photon or a Z boson. They examined constraints on new physics using an effective field theory approach, where new, heavy particles interact only through virtual exchanges, parameterised via an operator product expansion. The analysis shows that sensitivity to energy scales above 2 TeV can be achieved for couplings of order unity.

Only a subset of studies was updated for this year’s European Strategy update. Earlier projections, particularly for direct new physics searches, can be found in here and here.

A tremendous opportunity, and a tremendous challenge

The HL-LHC offers extraordinary physics opportunities over the next two decades, addressing crucial questions on electroweak symmetry breaking and providing unmatched potential for discovering new physics for decades to come.

The necessary accelerator and experiment upgrades are ambitious global projects now under construction. Their successful and timely completion is a conditio sine qua non for high-energy particle physics and any future large-scale project of the field. Achieving this goal requires the collaborations to remain fully committed and focused on this number one priority, and relies on the continuing support and allocation of the essential resources by our funding agencies.

We count on the European Strategy Group and the CERN Council to reaffirm this as a top priority for the years ahead.


Submissions to the European Strategy Group

Futher reading