Features

Searching for flavourful violations of the Standard Model

Katarina Anthony — Mon, 27 Oct 2025 13:05:27 +0000

Searching for flavourful violations of the Standard Model

Top HIghlight

False

Katarina Anthony Mon, 27/10/2025 - 14:05

Highlight

False

Feature

Lidia Dell'Asta

Jacob Julian Kempster

Daniele Zanzi

lepton flavour violation

lepton flavour universality

Every aspect of our world – from atomic nuclei and molecular chemistry to biology and astrophysics – emerges from the interactions of particles so small they are considered point-like. While some particles, like the electron, were discovered over a century ago, most remained hidden until the advent of modern particle accelerators. Particle physicists are dedicated to investigating the laws of nature that dictate how fundamental particles interact and combine to shape our everyday lives. By colliding particles at unprecedented energies inside massive detectors, physicists are able to create powerful magnifying lenses.

Accelerators led to the discovery of W and Z bosons, carriers of the weak fundamental force, in 1983 at CERN, as well as to the discovery of the top quark, the heaviest known particle, a decade later at Fermilab (USA). These discoveries supported the experimental foundation of the Standard Model (SM) of particle physics, the theory that describes how all known fundamental particles interact. The greatest triumph of the SM came in 2012 when its last missing piece – the Higgs boson – was discovered by the ATLAS and CMS Collaborations at CERN.

The SM is one of the most precisely tested scientific theories, with thousands of experimental results confirming its predictions to remarkable precision. Despite this success, the SM is not the “ultimate theory” of nature. It cannot explain experimental observations such as dark matter, an enigmatic form of matter detectable only through gravitational effects; the origin of neutrino masses, extremely light particles involved in processes like nuclear radioactive decay; nor the striking imbalance between matter and antimatter in the Universe.

These unanswered questions have driven researchers to develop theoretical frameworks that go beyond the Standard Model, with many predicting the existence of additional particles and interactions. Despite intensive searches, these hypothetical particles have yet to be observed. This could mean they are too heavy (requiring collision energies beyond the reach of the LHC), that their associated new interactions are too rare to be detected in the current amount of recorded data, or that the theory is simply not borne out by nature.

The focus of experimental particle physicists is, therefore, to study higher energies and collect more data, searching for any discrepancies between experimental measurements and SM predictions that could lead towards an ultimate theory of nature.

Nothing is truly forbidden in the quantum world. Even "forbidden" transitions are possible so long as they follow several intermediate, rule-following steps.

The Standard Model, its flavours and the search for new physics

One powerful way to search for new physics phenomena is to look for processes that are expressly forbidden by the SM, occurring only via new interactions. For example, physicists often search for processes that change the properties of a fundamental particle in an unexpected way. In the SM, quarks and leptons (known as fermions) come in different “flavours”. For example, electrons and muons are both charged leptons, but they differ in mass and are therefore considered different flavours. The SM sets strict limits on how fermions can change flavour and which flavours they are allowed to transition to. While the top quark may transition into a bottom, down or strange quark through a weak interaction using the electromagnetically-charged W boson (to conserve charge overall), it is forbidden from transitioning into an up or charm quark, even through an interaction with a neutral boson such as a Z boson, gluon, photon or Higgs boson. Equivalently, while the tau lepton may produce a muon or electron through a W boson interaction (including with the associated neutrinos), it cannot produce a muon or electron through a neutral boson interaction. These rules apply to all quarks and leptons alike.

Despite these rules, nothing is truly forbidden in the quantum world. Many of these transitions are physically possible so long as they follow several intermediate steps that adhere to the rules. But each additional step has a cost, reducing the overall probability of the process occurring. Thus, while these “forbidden” flavour-changing processes could occur according to the SM, they would be so rare as to never be detected by the ATLAS experiment, occurring in fewer than one in every 10¹⁴ interactions.

But what if ATLAS did detect them? Any observation of these processes would be indisputable evidence of new physics phenomena, indicating that something unknown has drastically increased the probability of these interactions. Particle physicists have even given these forbidden processes a name: those involving quarks are flavour-changing neutral currents (FCNCs), and those involving leptons are lepton flavour violation (LFV) (see Figure 1).

Figure 1: The production of a top-quark pair with a subsequent FCNC decay (left) and the production of a Z boson with a subsequent LFV decay (right, where the lepton ‘l’ is an electron or a muon, but not another tau). The forbidden interactions are highlighted with red circles. The pairs of gluons (g) and quarks (q) in the initial state come from the colliding protons. (Image: ATLAS Collaboration/CERN)

The search for FCNC interactions in top-quark processes

The LHC is the most powerful particle accelerator ever built, colliding bunches of protons every 25 nanoseconds at energies up to 13.6 TeV. Thanks to this unprecedented energy and collision rate, the LHC is often considered a particle “factory”, producing millions of top quarks and billions of Z bosons for study.

That’s great news for physicists looking beyond the Standard Model, where the heavier the particle involved in the interaction, the better. Since new particles and interactions likely require large amounts of energy to manifest, they may favour a connection with the heaviest and thus most inherently energetic known particle – the top quark. Searching for FCNC processes involving the top quark is thus considered a powerful technique for uncovering new physics phenomena.

In 2023, ATLAS physicists searched for FCNC interactions between a top quark and a Z boson, considering both modified production and decay mechanisms for the top quark (Figure 2). Both would generate a Z boson, a W boson and a bottom quark, with the modified decay also producing an up or charm quark. Physicists focused their search on leptonic decays of the W and Z bosons, with the W decaying to a charged lepton and a neutrino and the Z to a pair of oppositely-charged leptons of the same flavour. In practical terms, this meant looking for collision events with three charged leptons, one neutrino, one bottom quark and, in some cases, one additional up or charm quark.

Figure 2: The LHC production of a single top quark via a weak interaction utilising an FCNC vertex (left), and the production of a pair of top quarks via the strong interaction, including a subsequent FCNC decay vertex (right). The FCNC vertices are highlighted with the shaded circles. (Image: ATLAS Collaboration/CERN)

While the ATLAS detector is very good at measuring leptons, neutrinos pass through it completely undetected. Their presence can only be inferred from the missing transverse momentum in the collision. Quarks, meanwhile, produce showers of particles in the detector’s calorimeter called jets. Top quarks almost always decay into a bottom quark, which is extremely useful for physicists. Bottom quarks combine briefly with other quarks when they are formed, and so travel farther from the collision point before they decay. The resulting b-jets they produce have unique characteristics. To identify these b-jets (a process called b-tagging) physicists use machine-learning algorithms trained to spot these characteristics. Each jet is scored based on its likelihood of originating from a b-quark. High b-tagging efficiency is crucial for finding top-quark events.

However, all algorithms carry an inherent risk of misidentifying jets initiated by other quarks as b-jets. In their 2023 measurement, physicists calibrated their b-taggers to reach an efficiency of 77% for true b-jets, with a rejection factor of 5 for charm-quark-initiated jets (c-jets) and of 170 for lighter-quark-initiated jets. This means that one out of every five c-jets will be misidentified as coming from a bottom quark, and only one out of every 170 light-quark jets.

If ATLAS were to observe these flavour-changing processes, it would be indisputable evidence of new physics phenomena.

Figure 3: Distributions of recorded events (black points) and of simulated events (filled histograms) as a function of the Boosted Decision Tree output value. The distributions of the expected signal events, considering an FCNC coupling in single top quark production (dashed lines) or an FCNC process in the top quark decay (solid lines) are shown. The signals with an FCNC coupling between a top quark, a Z boson and an up quark (charm quark) are displayed in pink (pale blue). The lower panels show the ratio in each bin of event yields in data to the event yields from simulated background processes. A value of unity indicates that the events in data are in agreement with the predicted backgrounds. (Image: ATLAS Collaboration/CERN)

Physicists then defined an event selection tool to search for FCNC signatures, resulting in two signal regions (SR1 and SR2) that would be enriched in these processes. Figure 3 shows the results for the FCNC pair-produced top quarks and FCNC single-produced top quarks respectively, separated only by the number of jets required. The SM background processes resulting in similar final states are the production of a top-quark pair in association with a Z boson (see the light blue histogram in Figure 3) and the production of a pair of W/Z bosons in association with jets (see the orange histograms in Figure 3). Within each signal region, Boosted Decision Trees (BDT, a machine-learning technique) were employed to further separate the FCNC signatures from these SM backgrounds. The output of these BDTs determine the final shape of the distributions, with the aim being to push the FCNC signatures to one side of the distribution and the SM signatures to the other. This procedure was performed for FCNC processes between a top quark and an up quark (pink lines in Figure 3) and, separately, for FCNC processes between a top quark and a charm quark (blue lines in Figure 3).

Two interesting features can be observed in the distributions in Figure 3. First, the signal of the FCNC process in top-quark-pair decays is very well separated from backgrounds in the region targeting its signal (SR1), but not in the region targeting the FCNC production of a single top quark (SR2), and vice versa. Second, the sensitivity to the FCNC production of single top quarks is very dependent on whether an up quark or a charm quark initiates the process. As the colliding protons are themselves composed of up quarks and down quarks, the ingredients for an up-quark-initiated process are readily available. This results in good experimental sensitivity. By contrast, the charm-quark-initiated process arises from the constantly fluctuating mixture of “sea” quarks and gluons surrounding each proton, which carry a smaller proportion of the proton’s mass and energy. This reduces the probability for the charm-initiated process occurring, and subsequently reduces the experimental sensitivity.

After applying the event selection, physicists compared the data to the SM background prediction, searching for an excess of events that could be an FCNC signal. A statistical analysis was performed to determine whether the number of data events in each bin was larger and statistically incompatible with the predicted number of SM events. No excess was found and limits on the probability of an FCNC top-quark decay into a Z boson and a light quark were obtained. The most stringent limits were set for the top-quark decay into a Z boson and an up quark, which is excluded at probabilities greater than approximately 6 in 100,000 top-quark decays.

Figure 4: Summary of the current 95% confidence level observed limits on the branching ratios of the top quark decays via flavour-changing neutral currents (FCNC) to a quark and a neutral boson t → Xq (X = g, Z, ɣ or H; q = u or c) by the ATLAS and CMS Collaborations compared to several new physics models. Each limit assumes that all other FCNC processes vanish. The limits are expressed as FCNC top decay branching ratios, but several are obtained considering both FCNC top quark decay and FCNC top quark production vertices. (Image: ATLAS Collaboration/CERN)

The ATLAS Collaboration has performed many further FCNC searches for top quarks in association with gluons, photons and Higgs bosons, and for FCNC interactions of other particles such as the bottom quark. Figure 4 summarises the limits on the top-quark FCNC decay probabilities (branching ratios) obtained by the ATLAS (in blue) and the CMS (in red) Collaborations for various forbidden processes.

The search for lepton flavour violation in Z-boson decays

Searches for new physics phenomena have already led to the discovery of LFV in neutrinos, through the observation of neutrino oscillations. Neutrino oscillations could also induce LFV in electromagnetically-charged leptons. However, as the probability of this occurring is less than 1 in 10⁵⁴, any observation would indicate an exciting additional discovery.

A variety of experiments are currently searching for LFV processes with different lepton flavours and neutral bosons. This diversity of approaches is important, as new physics phenomena may interact with different leptons and bosons at different rates, and no single experiment can efficiently detect all LFV processes.

At high-energy colliders, the most promising opportunities to study LFV processes involve the production of tau leptons alongside massive neutral bosons, such as the Z boson. The LHC, in its role as a Z-boson “factory,” provides fantastic opportunities to search for flavour-violating decays involving the Z boson, such as Z→e𝜏 or Z→μ𝜏, where a tau lepton is produced alongside a charged lepton of a different flavour. While searches for LFV involving lighter leptons and massless photons are also viable at the LHC, they can be searched for with greater accuracy in low-energy experiments, i.e. in searches for muon decays into an electron and a photon (μ → eɣ).

The ATLAS Collaboration carried out a detailed search for flavour-violating Z→e𝜏 and Z→μ𝜏 processes. It was extremely challenging for several reasons, starting with the complicated task of spotting tau leptons. While electrons and muons can be directly detected by ATLAS, tau leptons have a lifetime of just 10^–13 seconds and decay before they reach the experiment. The most common tau decay modes include leptonic decays to electrons or muons and their associated neutrinos, or to jets of quarks as described above. Thus, the signature of a flavour-violating Z→e𝜏 or Z→μ𝜏 process may well mimic other perfectly viable SM signatures, such as Z→ee.

Figure 5: Diagrams of a LFV Z-boson decay (left) and of the two main SM processes which can produce final states experimentally similar to the LFV decays: a Z-boson decay into a pair of tau leptons (middle) and a W-boson decay into a light lepton and neutrino in association with an additional jet (right). The green arrows represent electrons or muons (l), the blue triangles are the jets produced by either the hadronic tau decay (𝜏 jet) or by a quark or a gluon (q/g jet), and the dashed blue lines indicate undetected neutrinos. (Image: ATLAS Collaboration/CERN)

Figure 5 (left) illustrates a flavour-violating Z-boson decay on a plane perpendicular to the proton beam direction. The unique feature of the LFV process is the alignment of the undetected neutrino (dashed line) and the tau lepton. This configuration, without any additional detected particles, cannot be produced by SM processes. Nevertheless, due to the random nature of quantum processes in the Universe and imperfections in experimental measurements, many SM processes can still look similar enough to cause a problem!

The main SM lookalike is a Z-boson decay into a pair of tau leptons, where one tau lepton subsequently decays into a light lepton and two neutrinos and the other into a tau-jet and a neutrino. While the number of detected particles in this SM process matches that of the LFV decay, there’s a greater number of undetected neutrinos (see Figure 5, middle). The second most important SM lookalike involves a W boson produced in association with a jet, where the W-boson decays into a light lepton and a neutrino, and the jet originates from a quark or gluon misidentified as a hadronically decaying tau. In this case, the detected final state is also the same as in the LFV decay, but the neutrino is typically emitted in a direction close to the light lepton instead of the jet (see Figure 5, right).

These kinematic differences are the key tool for isolating the LFV signature. Physicists designed a deep neural network (NN) to “learn” the kinematic properties of the signal events and how they differ on average from background events. NNs are extremely effective in this task as they can learn complex correlations among multiple features of an event. Figure 6 shows a histogram of events binned in a NN output (after applying the initial event selections). As with the BDTs in the FCNC searches, the events with high NN scores are more signal-like and the events with low values are more background-like. The red dashed line shows the expected distribution of signal events under an assumption that LFV decays occur as frequently as five in every 10,000 Z-boson decays.

As with all previous measurements, no excess above the SM prediction was observed in the data. The ATLAS Collaboration set limits on LFV Z-boson decays into a tau and a light lepton that are stronger than those set by experiments on the Large Electron Positron (LEP) collider. At the 95% confidence level, the ATLAS result established that if these LFV decays do occur in nature, their probability must be less than about five in every one million Z-boson decays. Using similar analysis methods, physicists also set even more stringent limits on LFV Z-boson decays into an electron and a muon.

The ATLAS Collaboration’s state-of-the-art searches for FCNC and LFV interactions have produced powerful probes of new physics phenomena.

Figure 6: Distribution of recorded events (black points) and of simulated events (filled histograms) as a function of the NN output value. The yellow and blue histograms are the events from the W and Z boson decays shown in Figure 5, respectively. The distribution of the expected signal events is shown with the red dashed line. The lower panel shows the ratio in each bin of event yields in data to the event yields from simulated background processes. A value of unity indicates that the events in data are in agreement with the predicted backgrounds. (Image: ATLAS Collaboration/CERN)

Outlook

The ATLAS Collaboration’s state-of-the-art searches for FCNC and LFV interactions have produced powerful probes of new physics phenomena. Dedicated FCNC studies set limits on the FCNC top-quark decay branching ratio as tight as 10^–5 (Figure 4), while further LFV searches focusing on Higgs-boson decays were also performed. The upper limits on the LFV decay probabilities of the Higgs boson are two orders of magnitude weaker than those for Z bosons, due to the much smaller number of Higgs bosons produced in LHC collisions and the more challenging task of distinguishing events with a Higgs boson from background processes. This sensitivity is expected to improve significantly with the collection of more data. ATLAS physicists also performed a search for a simultaneous FCNC and LFV process using Effective Field Theory in top-quark interactions with light quarks, muons and taus. Although this analysis is also statistically limited, it represents an exciting new direction for future studies.

Since 2022, the LHC experiments have been collecting data from proton-proton collisions at a centre-of-mass energy of 13.6 TeV. This data-taking period is expected to provide approximately 1.5 times more Z bosons, Higgs bosons and top-quark pairs than in all previous runs, thanks to both the increased collision energy and the larger dataset ATLAS is recording.

But larger datasets aren’t the only factor boosting searches for LFV and FCNC processes! Ongoing advancements in experimental methodologies and the ATLAS Collaboration’s deepening understanding of its ever-evolving detector will further improve sensitivity, driving searches for even rarer interactions. One continuous area of development is ATLAS’ in-house algorithms for identifying charm-quark-initiated jets, analogous to the existing and well-honed b-tagging capabilities. As mentioned above, limits on couplings involving the top and charm quarks are less stringent than those involving the up quark, and these algorithms will boost ATLAS’ charm-quark sensitivity.

Looking ahead, further searches will be conducted following the completion of the High-Luminosity LHC (HL-LHC), for which the ATLAS experiment will undergo major upgrades to cope with the much higher luminosity. At the current LHC energy of 13.6 TeV, every time two bunches of protons cross each other an average of 50 overlapping collisions occur. This already makes event reconstruction very challenging. At the HL-LHC, this number will increase to approximately 200 simultaneous collisions. To address this complexity, a new silicon tracking detector, the Inner Tracker (ITk), will be installed in the ATLAS experiment.

The ITk features new technologies that provide higher detection granularity, improved radiation hardness, faster readout speeds and a reduced material budget compared to the current inner detector – all of which will enable similar event reconstruction efficiencies to current data-taking, despite the much harsher future conditions. Moreover, the ITk’s extended coverage will enable ATLAS to reconstruct events in regions of the detector that are not currently equipped with a tracker, further extending and improving the data collection. The integrated luminosity that will be collected by the ATLAS experiment during the HL-LHC is expected to be 10 times greater than the total recorded at the LHC. This means that searches for very rare or forbidden processes, such as LFV and FCNC, will remain powerful, exciting and ever improving probes for physics beyond the Standard Model.

About the Authors

Lidia Dell'Asta joined the ATLAS Collaboration in 2007 while at the University of Milano. She then worked as a postdoctoral fellow at Boston University and as a research fellow at the University of Roma2. She is currently an associate professor at the University of Milano. In addition to her work on the detector trigger system, where she coordinated the muon trigger group, she has been active in data analysis. She has worked on Standard Model measurements as well as the measurement of the coupling of the Higgs boson to tau leptons. Over the past ten years, her research has focused on rare production processes of single top quarks, and she has served as a convener of the Single Top group. She has contributed to the measurement of single top quark production in association with a Z boson and to the search for top–Z FCNC couplings. Jacob Julian Kempster joined the ATLAS Collaboration in 2011 while at Royal Holloway, University of London. He subsequently worked as a Research Fellow at the University of Birmingham and is currently a Senior Research Fellow at the University of Sussex. His research focuses on using the top quark as a tool to search for new physics beyond the Standard Model. He performed the first ATLAS search for lepton flavor violating (LFV) couplings of the top quark to muons and tau leptons, and previously served as a subgroup convener for the Top+X working group. His other primary research area is Effective Field Theory (EFT), where he leads efforts on global EFT fits and recently completed his term as Chair of the LHC EFT Working Group. Daniele Zanzi was an active member of the ATLAS Collaboration until 2024. As researcher at the University of Melbourne, at CERN, and at the University of Freiburg, he has focussed on searching for LFV interactions. He has also contributed to the development and operation of the ATLAS trigger system, to measurements of the Higgs boson properties and to searches for dark matter and supersymmetric particles.

About the banner image: Visualisation of a H →μτ candidate event, with the μτ_e (top) and μτ_had (bottom) channel. An electron track is shown in green, a red line indicates a muon. A τ_had-vis candidate is displayed in purple, the ET_miss is shown by a white dashed line. (Image: ATLAS Collaboration)

Scientific articles

Search for flavor-changing neutral-current couplings between the top quark and the 𝑍 boson with proton-proton collisions at 13 TeV with the ATLAS detector (Phys. Rev. D 108 (2023) 032019, arXiv:2301.11605, see figures)
Search for flavour-changing neutral-current interactions of a top quark and a gluon in proton-proton collisions at 13 TeV with the ATLAS detector (Eur. Phys. J. C 82 (2022) 334, arXiv:2112.01302, see figures)
Search for flavour-changing neutral-current couplings between the top quark and the photon with the ATLAS detector at 13 TeV (Phys. Lett. B 842 (2023) 137379, arXiv:2205.02537, see figures)
Search for flavour-changing neutral-current couplings between the top quark and the Higgs boson in multi-lepton final states in 13 TeV proton-proton collisions with the ATLAS detector (Eur. Phys. J. C 84 (2024) 757, arXiv:2404.02123, see figures)
Top Quarks + X Summary Plots April 2024 (ATL-PHYS-PUB-2024-005)
Search for charged-lepton-flavour violation in Z-boson decays with the ATLAS detector (Nature Phys. 17 (2021) 819, arXiv:2010.02566, see figures)
Search for the charged-lepton-flavor-violating decay 𝑍 →𝑒⁢𝜇 in proton-proton collisions at 13 TeV with the ATLAS detector (Phys. Rev. D 108 (2023) 032015, arXiv:2204.10783, see figures)
Searches for lepton-flavour-violating decays of the Higgs boson into eτ and μτ in 13 TeV proton-proton collisions with the ATLAS detector (JHEP 07 (2023) 166, arXiv:2302.05225, see figures)
Search for charged-lepton-flavor violating 𝜇⁢𝜏⁢𝑞⁢𝑡 interactions in top-quark production and decay in proton-proton collisions at 13 TeV with the ATLAS detector at the LHC (Phys. Rev. D 110 (2024) 012014, arXiv:2403.06742, see figures)
The ATLAS ITk detector system for the Phase-II LHC upgrade (Nucl. Instrum. Methods Phys. Res., Sect. A 1045 (2023) 167597)
Search for flavour-changing neutral tqH interactions with H→γγ in proton-proton collisions at 13 TeV using the ATLAS detector (JHEP 12 (2023) 195, arXiv:2309.12817, see figures)
Search for flavour-changing neutral current interactions of the top quark and the Higgs boson in events with a pair of τ-leptons in proton-proton collisions at 13 TeV with the ATLAS detector (JHEP 06 (2023) 155, arXiv:2208.11415, see figures)

A transformative leap in physics: ATLAS results from LHC Run 2

Katarina Anthony — Mon, 07 Apr 2025 06:25:39 +0000

A transformative leap in physics: ATLAS results from LHC Run 2

Top HIghlight

False

Katarina Anthony Mon, 07/04/2025 - 08:25

Highlight

False

The onset of high-energy operation of the Large Hadron Collider (LHC) at CERN in 2010 heralded a new and transformative era in physics research. A powerful and versatile hadron collider paired with sophisticated particle detectors and worldwide grid computing, all performing beyond expectations, allowed the LHC experiments to chart new frontiers and advance particle physics. While the discovery of the Higgs boson by the ATLAS and CMS Collaborations in 2012 stands out as the LHC’s magnum opus, the wealth of results includes the observation of countless new processes and states, precision measurements matched with continuously refined theoretical predictions, and a broad and deep exploration of the new physics landscape. The non-observation of physics beyond the Standard Model so far, despite strong theoretical motivation for new TeV-scale particles, is another paramount outcome of the LHC. How can one fathom the puzzling tailoring or fine-tuning that seems to haunt the scalar sector? This mystery has fuelled many theoretical developments. It has prompted phenomenologists and model builders to propose new ways to preserve naturalness, and sparked renewed interest in more speculative anthropic concepts, stimulated by the string landscape and eternal inflation.

Figure 1: Measured modifiers of the Higgs boson coupling strength to elementary particles, along with their uncertainties, as a function of particle masses. The ordinate is linear for fermions and square-root for weak bosons, reflecting their mass dependence and enabling a linear trend. The diagonal red line represents the BEH mechanism’s prediction of non-universal, mass-dependent interactions. (Image: ATLAS Collaboration/CERN)

The discovery of the Higgs boson, a manifestation of the condensed Brout-Englert-Higgs (BEH) field, has enabled ATLAS and CMS physicists for the first time to directly study electroweak symmetry breaking and the process of mass generation. The first run of the LHC (Run 1), comprising the data-taking periods between 2010 and 2012 with proton collisions at 7 and 8 TeV centre-of-mass energies, marked the discovery of the Higgs boson via its decays to boson pairs, the measurement of its mass, a critical parameter in the energy evolution of the scalar field potential, and the confirmation of its basic quantum numbers and coupling properties. The second run of the LHC (Run 2), delivering seven times more data during the years 2015–2018 at 13 TeV centre-of-mass energy, allowed ATLAS and CMS to establish Higgs boson interactions with fermions and observe the expected non-universal, mass-dependent interaction strengths (Figure 1). This confirmed the BEH mechanism of spontaneous symmetry breaking of the electroweak vacuum through a scalar doublet field with an energy potential whose ground state is non-zero. Masses arise from particles interacting with this ever-present background field that permeates the universe. However, it remains unknown why some particles interact more strongly with the field, acquiring greater mass, than others. This is the flavour problem, and perhaps the most striking feature is not the order-one coupling strength of the top quark, but rather the exceptionally small coupling of the electron.

Despite its success, the newly discovered scalar sector raises profound questions about its naturalness (the Higgs boson is the only known elementary particle with a bare mass term — how come it is so light?) and the stability of the electroweak vacuum at high energy, in addition to the enigmatic flavour pattern. It may also relate to the matter-antimatter asymmetry in the universe, potentially through additional scalar fields, and to cosmological dark matter if it is composed of massive elementary particles. The remarkable similarity between the BEH mechanism and the Bardeen–Cooper–Schrieffer theory of superconductivity, where the Bose condensation of a Cooper pair of two electrons plays the role of the BEH field, raises the question whether the Higgs boson might not be elementary after all.

Figure 2: Installation of the new Insertable B-Layer (IBL) into the ATLAS Pixel detector in May 2014. (Image:H. Pernegger/ATLAS Collaboration)

Figure 3: Cumulative delivered and recorded luminosity during the four years of Run-2 operation (2015–2018). The good-for-physics data quality efficiency during Run 2 was 95.2%. (Image: ATLAS Collaboration/CERN)

ATLAS data-taking during Run 2 benefited from several detector upgrades implemented during the LHC’s long shutdown 1 (2013–2015), most prominently the addition of the insertable B-layer (IBL), a pixel layer with radiation hard sensors mounted at a radius of only 33 mm from the beam (see Figure 2). The IBL significantly improved the vertex resolution of the ATLAS Inner Detector and thus the heavy-flavour jet tagging performance, among other benefits. The ATLAS experiment achieved record data-taking and data quality efficiency during Run 2 (Figure 3). Of the 156 fb^–1 integrated proton–proton luminosity delivered by the LHC, 140 fb^–1 (90%) is used for physics analysis. Additional low-luminosity runs were conducted to enable precision measurements of W-boson properties. The Run-2 dataset further comprises 2.2 nb^–1 of lead–lead collisions taken in the years 2015 and 2018 at 5.02 TeV nucleon–nucleon centre-of-mass energy, 179 nb^–1 of proton–lead collisions taken in 2016 at 5.02 and 8.16 TeV, a small dataset of xenon–xenon collisions taken in 2017, and reference proton–proton collision datasets at corresponding energies. The luminosity measurement, calibrated using LHC beam-separation scans, reached the tremendous precision of 0.8% for the full ATLAS Run-2 proton–proton dataset, never before attained at a hadron collider. Online event selection employing low-threshold single and di-object triggers as mainstays, and a new first-level topological trigger for object combination, ensured high efficiency across the ATLAS physics programme.

Figure 4: Suite of improvements in the performance of the b-jet tagging algorithms versus year. DL1 denotes machine learning algorithms based on deep learning neural networks, and GN stands for graph neural networks. (Image: ATLAS Collaboration/CERN)

The worldwide LHC computing grid performed superbly. Heterogeneous, pledged and opportunistic computing resources were efficiently integrated, allowing ATLAS to promptly reconstruct the recorded data and produce substantial amounts of simulated data. The compute performance of the Geant-4 based full detector simulation was continuously improved, and a new fast calorimeter simulation, enhanced with generative machine-learning, was brought to completion during Run 2. New developments in the reconstruction and identification of charged leptons, flavoured and unflavoured jets, missing transverse momentum, and highly boosted objects benefited from ever more sophisticated machine learning algorithms, where in particular recent advances using graph neural networks stand out in performance (Figure 4). The excellent performance of the ATLAS detector and reconstruction software, together with its meticulous calibration, and the accuracy of the physics modelling and detector simulation, has made the Run-2 dataset the largest, most comprehensive, and highest quality collection of high-energy physics data at the time.

Figure 5: Visualisation of an ATLAS event from 2018, consistent with the simultaneous production of three W bosons, which decay into a muon, two electrons, and neutrinos. While neutrinos remain undetected, their presence is inferred from the momentum imbalance of all visible particles transverse to the beam direction. (Image: ATLAS Collaboration/CERN)

Figure 6: Visualisation of an ATLAS event from 2018, consistent with the production of a W-boson pair via photon–photon interaction. The W bosons decay into a muon and an electron (both detected) and neutrinos (undetected, but producing missing energy). As shown in the top left corner, no additional particles emerge from the same vertex, as expected for a photon fusion event. (Image: ATLAS Collaboration/CERN)

Run 2 has seen a surge in the breadth and depth of ATLAS’ physics results. The Collaboration has released 415 papers with the full Run-2 dataset to date (and counting), and a similar number of results using partial Run-2 datasets. Numerous rare processes have been observed for the first time. Each iteration of an analysis improved the results beyond the statistical yield, owing to better performance and analysis methods, notably by widely embracing machine learning algorithms.

ATLAS Run-2 results have set new benchmarks in multi-boson production. Forty years after the discovery of the W boson at CERN’s proton–antiproton collider, ATLAS observed for the first time the simultaneous production of three W bosons (Figure 5) as well as that of a W-boson pair through the interaction of two photons (Figure 6). Numerous vector-boson scattering processes were newly observed and measured. Angular analyses were performed to detect, again for the first time, the joint longitudinal polarisation of weak bosons in such events. These studies pave the way to testing another key property of the BEH mechanism: the regulation of longitudinal vector boson scattering at high momentum transfer.

Figure 7: Visualisation of a striking four-top candidate event. The event features seven jets, represented by cones, with four identified as b-tagged (blue cones). Three of the top quarks decay leptonically, producing two muons (red) and one electron (blue), while the fourth top quark decays hadronically into three jets. (Image: ATLAS Collaboration/CERN)

The production of almost 300 million top-quarks during Run 2, recorded with high trigger efficiency, has prompted a plethora of top-quark production and property measurements. It allowed ATLAS to measure rare high-mass processes such as the associated production of a top-quark pair with a Z boson, directly probing the top–Z coupling, and to observe four-top-quark production for the first time (Figure 7), a process four thousand times rarer than Higgs boson production. The first observation of quantum entanglement of top-quark pairs by ATLAS marked the beginning of high-energy quantum information research at the LHC. Top quarks also serve as a clean source of W bosons. When produced in pairs, one W boson can be used for identification while the other is studied in detail. This approach has enabled highly sensitive tests of lepton universality in the weak interaction and resolved a long-standing discrepancy from LEP.

The Run 2 data allowed ATLAS to observe and measure all major Higgs boson production modes at the LHC and to probe the non-universal Higgs couplings to fermions, thereby confirming the Brout-Englert-Higgs mechanism.

Figure 8: The first observation of the Higgs boson decay into two bottom quarks in Higgs plus W or Z boson production (left, reference), and the observation of top–antitop production in association with a Higgs boson decaying into two photons (right, reference), both achieved with a partial Run-2 dataset. (Image: ATLAS Collaboration/CERN)

Higgs physics underwent a transformation during Run 2. All major Higgs boson production modes at the LHC were observed and measured. Previously thought to be impossible or highly challenging, the decay of the Higgs boson into a bottom quark pair — which accounts for more than half of all Higgs boson decays — was observed for the first time by ATLAS in summer 2018 using a partial Run-2 dataset (Figure 8, left). The associated production of the Higgs boson with top-quark pairs, a rare process representing less than one percent of all Higgs boson production modes, was unambiguously observed by ATLAS in the same year through its decay into two photons, confirming the large predicted top–Higgs coupling (Figure 8, right). Evidence for exceedingly rare Higgs-boson decay channels was also seen. The small Higgs boson decay width, predicted to be 4.1 MeV in the Standard Model, has been constrained to better than 60% precision — well beyond expectations — by measuring off-shell production at high mass and comparing the measured rate to that observed on the mass shell.

The full suite of Run-2 analyses and their combination allowed ATLAS to draw a detailed map of the Higgs boson and its interactions with unprecedented depth and precision (Figure 9), a decade after its discovery. The possibility to directly constrain the quartic field coefficient in the BEH energy potential through the measurement of Higgs-pair production is a fascinating opportunity that has already been explored by ATLAS during Run 2, despite its minuscule cross section, three orders of magnitude smaller than that of single Higgs boson production. Spectacular improvements with each iteration of the analyses have sharpened the prospects for observing and measuring this channel in the upcoming High-Luminosity LHC era.

Figure 9: The ratio of observed rate to predicted Standard-Model event rate for different combinations of Higgs boson production and decay processes measured by ATLAS with the full Run-2 dataset (reference). (Image: ATLAS Collaboration/CERN)

While the LHC was primarily built for the direct production and study of new particles and phenomena, progress in the field also relies on precision measurements that refine our understanding of Standard Model parameters and probe new physics through quantum loops.

Figure 10: High-precision electron and photon energy calibration (left, reference) allowing to measure the Higgs boson mass to unprecedented precision in this channel, and in the combination with the four-lepton decay channel (right, reference). (Image: ATLAS Collaboration/CERN)

While the LHC was primarily built for the direct production and study of new particles, processes, and phenomena, progress in the field also relies on precision measurements that refine our understanding of Standard Model parameters and probe new physics through quantum loops. ATLAS has played a key role in this endeavour, delivering the first measurements of the W boson mass and width at the LHC, both found in agreement with the Standard Model predictions.

Figure 11: Harvest of ATLAS cross section measurements at different proton–proton collision energies for processes spanning 14 orders of magnitude — from milibarns to femtobarns. These unique measurements probe new forms of particle interactions and extremely rare processes, and (so far) confirm the predictive power of the Standard Model. Theoretical developments and computations have been key to this progress. (Image: ATLAS Collaboration/CERN)

ATLAS further released the most precise measurement of the Higgs boson mass to date (Figure 10), a precise preliminary measurement of the weak mixing angle, the most precise measurement yet of the top-quark mass in combination with CMS, the world’s most precise measurement of the strong coupling constant (these latter two results still based on the Run-1 dataset), the most precise result to date of the total proton–proton cross section by measuring forward elastic scattering, comprehensive constraints on parton distribution functions, as well as numerous measurements of total, fiducial, and differential cross-sections (Figure 11), angular properties and correlations. Sensitive measurements of rare B-meson decays and CP violation, along with the most precise measurement to date of the B-meson lifetime, complement these results. Effective field theoretical extensions of the Standard Model Lagrangian provide a general framework for quantifying the reach of direct and indirect searches for new physics in terms of higher dimensional field operators, allowing ATLAS to consistently combine measurements of different processes and thus enhance their impact.

ATLAS studies of heavy-ion collisions collected during Run 2 have deepened our understanding of the quark-gluon plasma through detailed measurements of its properties using soft and hard probes.

Figure 12: Left: per-event normalised dijet yields as a function of the dijet transverse momentum ratio x_J for leading jets between 126 and 141 GeV and shown for selections of event centrality in lead–lead collisions and in proton-proton (pp) collisions (reference). A larger jet asymmetry is seen in central collisions. Right: ratio of the observed yield in lead–lead collisions to the expectation from an equivalent number of nucleon–nucleon (NN) collisions in photon-jet events versus the photon momentum (reference). Both figures exhibit the phenomenon of “jet quenching” in the quark-gluon plasma created in central heavy-ion collisions. (Image: ATLAS Collaboration/CERN)

Collisions among ionised lead nuclei, each consisting of 208 nucleons (82 protons and 126 neutrons), generate temperatures several hundred thousand times hotter than the core of the Sun — hot enough to melt nucleons within the nuclei, forming a plasma of deconfined quarks and gluons, a near-perfect fluid with almost no viscosity. Due to its ephemerality (lifetime of about 10⁻²³ s) and microscopic size (10⁻¹⁴ m), the quark-gluon plasma cannot be observed directly, but only through the particles it emits.

Figure 13: Visualisation of an ATLAS event from 2018, displaying energy deposits from two photons in the electromagnetic calorimeter (green) on opposite sides, with no additional detector activity. This clear signature corresponds to “light-by-light” scattering of two incoming photons into two outgoing photons in an ultraperipheral collision of two lead ions. (Image: ATLAS Collaboration/CERN)

ATLAS studies of heavy-ion collisions collected during Run 2 have deepened our understanding of this new state of matter by employing soft probes via measurements of multi-particle correlations and flow harmonics, alongside hard probes to investigate energy-loss (“quenching”) effects in a variety of final states. Further insights have been gained by measuring the production of quarkonia and electroweak bosons as a function of the collision centrality. Jet quenching, a process where partons lose energy through collisions and gluon radiation in the quark-gluon plasma, was first observed at the LHC by ATLAS in 2010. Since then, the study of this striking phenomenon has evolved significantly with measurements of transverse momentum asymmetries in dijet, photon-jet, and Z-jet events (Figure 12), and by investigating varying jet energies and shapes.

Ultraperipheral collisions have enabled ATLAS to directly observe, for the first time, light-by-light scattering (Figure 13) and tau-lepton pair production. These processes are made possible by the extreme electric fields — up to seven orders of magnitude beyond the Schwinger limit of 10¹⁸ Volt per metre and among the highest observed in nature — generated by the quasi-real photons emitted by the colliding lead ions. Similarly, the enormous magnetic fields, reaching up to 10¹⁶ Tesla, in these processes allowed ATLAS to perform highly sensitive searches for hypothetical magnetic monopoles.

The non-observation of physics beyond the Standard Model so far, despite strong theoretical motivation for new TeV-scale particles, is another paramount outcome of the LHC.

Figure 14: Lower limits on the masses of supersymmetric particles — gluinos, top squarks, and charginos or neutralinos (shown on the horizontal axis, reference) — are plotted against the mass of the lightest supersymmetric neutralino (vertical axis), a potential dark matter candidate. Naturalness suggests that these particles should have masses at or below the TeV scale. (Image: ATLAS Collaboration/CERN)

The vast field of new physics searches featured prominently in the ATLAS physics programme during Run 2. The higher proton–proton collision energy and increased data sample provided a substantial gain in sensitivity. Searches for supersymmetry have excluded gluino masses up to 2.4 TeV, top-squark masses up to 1.2 TeV, and chargino and neutralino masses up to 1 TeV, depending on the scenario (Figure 14). These results set strong constraints on natural supersymmetry. Sensitivity gaps could be addressed through specialised searches.

Supersymmetry and other new physics models posit an extended scalar sector, explored through a variety of channels closing in on additional Higgs bosons. Heavy, strongly produced resonances such as excited quarks could be excluded up to almost 7 TeV (Figure 15, left), and heavy Z-like bosons up to about 5 TeV. Analyses looking for heavy particles decaying into boosted boson pairs exploited jet substructure techniques to increase their sensitivity far beyond the TeV mass scale. An apparent diphoton excess near 750 GeV in the 2015 data was not confirmed with larger datasets. The direct production of dark matter particles in proton–proton collisions was explored in a broad set of recoil modes involving jets (Figure 15, right), photons, weak bosons, Higgs boson, and top quarks, as well as in searches for supersymmetry.

Figure 15: Visualisations of a high-mass di-jet event (left), which is sensitive to potential new physics such as quark substructure, and a dark matter candidate event (right), characterised by a significant transverse energy imbalance (marked by the red dashed line). Both events were recorded by the ATLAS detector in 2017. (Image: ATLAS Collaboration/CERN)

Figure 16: Upper limits on the cross section for elastic scattering between a weakly interacting massive particle (WIMP) and a nucleon as a function of WIMP mass, derived from ATLAS searches for invisible Higgs boson decays for different mediator models. The limits are compared to results from dedicated underground experiments. Also shown is the expected neutrino fog from coherent elastic neutrino–nucleus scattering for a germanium target. (Image: ATLAS Collaboration/CERN)

Dark matter was also searched for via invisible decays of the Higgs boson produced through weak boson fusion or in association with other particles. These processes provide strong constraints on dark matter originating from weakly interacting massive particles, complementary to direct recoil-based searches underground and indirect searches for annihilation signals in space (Figure 16). Long-lived heavy particles may commonly occur in decay chains involving new mass-degenerate particles, weak couplings, or quantum loops with high virtuality (Figure 17). Searches for such phenomena have seen a rise of interest in ATLAS and produced numerous innovative analyses, which also benefited from advances in reconstruction techniques such as large impact parameter tracking. While the Standard Model is subjected to increasingly stringent tests and the ATLAS searches become ever deeper and more inventive, none have yet revealed compelling evidence of new physics.

ATLAS published a collection of six articles in Physics Reports to highlight the key findings and transformative implications of the landmark results obtained from its Run-2 proton–proton dataset. The articles cover Higgs boson physics, electroweak, QCD and flavour physics, top-quark physics, as well as searches for new phenomena covering an extended Higgs sector, supersymmetry, and other scenarios. A summary of heavy-ion results will be released at a later stage. As Run-3 data-taking progresses and the Collaboration prepares for major detector upgrades ahead of the High-Luminosity LHC — the next major milestone for the global community — the Run-2 results stand as the state-of-the-art in particle physics at the high-energy frontier.

Figure 17: Visualisation of an event satisfying selection criteria in a search for long-lived, massive particles produced in association with a muon (reference). Two of the reconstructed displaced vertices exhibit high mass and are consistent with potential decays of heavy long-lived particles. (Image: ATLAS Collaboration/CERN)

About the banner image: Visualisation of a Higgs candidate decay to four muons and the measured Higgs coupling strengths to elementary particles. (Edit: K. Anthony/ATLAS Collaboration)

A version of this text was released as a foreword to the six ATLAS summary articles on Run-2 physics results published early 2025 in Physics Reports.

Learning by machines, for machines: Artificial Intelligence in the world's largest particle detector

Katarina Anthony — Wed, 05 Jun 2024 12:57:58 +0000

Learning by machines, for machines: Artificial Intelligence in the world's largest particle detector

Top HIghlight

False

Katarina Anthony Wed, 05/06/2024 - 14:57

Highlight

False

In today’s age, you can't do much without interfacing with artificial intelligence and machine learning (AI/ML). This technology lets you unlock your phone via face recognition, helps curate your social media feed and powers internet search. In the future, it promises to automate tasks as mundane as driving a car and as cerebral as scientific outreach. Its clear transformative capability has captured our collective attention, sparking dialogue across scientific communities, governments and the general public, alike. But long before ChatGPT or DALL-E, the basic statistical principles that underpin the world's most sophisticated ML tools were hard at work in the field of high-energy collider physics. Today, they are enabling unprecedented progress in understanding the nature of our fundamental universe.

High-energy physics (HEP) can trace its relationship with ML back many decades, with the earliest neural networks coming into play in the 1990s. ML algorithms improved Higgs-boson searches at CERN’s LEP collider, powered CP-violation measurements at the B factories at KEK and SLAC, and enabled the observation of single top-quark production at Fermilab's Tevatron collider. They were also key for the discovery and study of the Higgs boson as well as the observation of the ultra rare two-muon decay of the Bs meson at the LHC.

But it wasn't until the 2010s that modern computational power and methodological innovation enabled deep learning, and let AI-based research methods like ML really shine. In many ways, the relationship between particle physics and ML is a natural and symbiotic one. High energy particle collisions offer a means to study fundamental interactions under conditions similar to the early universe, allowing a window into potential particles or processes that are “frozen out” in the current universe. In this way, finding optimal and intelligent ways to sift through the trove of data from experiments at CERN’s Large Hadron Collider (LHC) is crucial, as it enables researchers to precisely characterise the Standard Model (SM) and understand mysteries like dark matter and matter-antimatter asymmetry that motivate new physics beyond the Standard Model (BSM).

Collider-based experiments led to some of the original “big data” problems. Experiments like ATLAS at the LHC operate with staggering data rates, producing over 60 terabytes per second, yet only some of these proton collision events may contain processes of interest. What's more, these experiments offer a dataset that is unique in its complexity and statistical power, in which new ML architectures or problems such as systematic biases or hardware optimization can be studied.

High-energy physics can trace its relationship with machine learning back many decades, with the earliest neural networks coming into play in the 1990s.

The task of operating the ATLAS experiment and extracting results from its vast datasets is a computational labyrinth. From its inception in the 1990s, the ATLAS experiment was designed to process every proton collision digitally. Within seconds of a collision the data from millions of sensors have been filtered through a web of custom electronics and analysed on a computing farm with tens of thousands of CPUs. Collisions of interest are recorded and reanalysed countless times by physicists looking to better understand the nature of the universe. In all cases the goal of this analysis is to discover precisely what happened at the interaction point, where two protons travelling at 99.999999% the speed of light collide in the accelerator, producing a plethora of new particles.

Unfortunately, the physics at the interaction point can be elusive. Many of the most interesting SM or BSM particles produced in the proton collision will decay in less than one trillionth of one trillionth of a second! Physicists can only pick up on hints of SM or BSM physics by looking for the decay products of the most interesting particles, which may themselves decay before reaching any sensors in ATLAS. From a single collision the ATLAS experiment may record thousands of individual particles, and to make matters worse, it typically has to deal with dozens of simultaneous collisions. To understand what happened at the interaction point, physicists must carefully reconstruct, identify and measure each of these particles. These are then used to reconstruct the entire collision event, which are scoured for processes of interest that may lead to a better understanding of known particles, or shed light on the existence of never-before-observed ones.

ML methods are designed to harness large amounts of data in order to infer new information, making them naturally suited to various data processing tasks in ATLAS, from the moment a particle hits the detector all the way to the final published results. The examples that follow give a representative idea of how extensively this technology has pervaded the experiment, but merely scratch the surface of the full picture and potential of ML in HEP.

Machine learning at work in ATLAS

During regular operation of the ATLAS experiment, the first challenge is what to do with the data that is created. With multiple subsystems, an LHC collision frequency of 40 million per second and millions of individual channels of data to read out, the ATLAS experiment produces a data rate far greater than can possibly be written to disk. A complex trigger system implements algorithms that rapidly evaluate incoming data events to determine if they are interesting enough to keep, rejecting the overwhelming majority of events produced.

Figure 1: A single cell of a recurrent architecture used to determine particle transverse energy from ionisation in the ATLAS electromagnetic calorimeter cells, incorporating knowledge of the sequence of proton bunch crossings. Here ADC refers to the information inputted to the network from the on-detector analog-digital converters, which provide one data sample about the calorimeter activity every 25 ns. This input is given to a recurrent neural network (RNN) cell, which feeds the ADC information and the output of the previous RNN cell into a final dense operation to estimate the energy. Five bunch crossings in total are used to predict a given energy measurement. (Image: ATLAS Collaboration/CERN)

This task requires sophisticated inference that can be done very quickly, introducing the use of “fast ML” to accelerate ML algorithms traditionally run in software for use in hardware such as field-programmable gate arrays (FPGAs). This process allows for greater intelligence closer to the source of the data, leading to more accurate reconstruction and better trigger decisions. For example, the energy and timing of signals in the ATLAS electromagnetic calorimeter subsystem can now be estimated by convolutional and recurrent ML architectures in real time of LHC operation, outperforming existing signal filters (Figure 1). This capability of ML to perform fast and accurate regression of key physical quantities can also be used for more accurate calibrations of detector signals.

AI/ML also comes into play in the reconstruction algorithms that turn detector signals into physics objects. Well before ATLAS recorded its first data, physicists had developed hundreds of algorithms to reconstruct specific particle types based on the signatures they leave in the different ATLAS sub-detectors. Some particles, like b-hadrons, will decay before reaching any of the ATLAS sub-detectors, and are discerned by triangulating the trajectories of the decay products back to a displaced vertex that is separated from the proton collision point by just a few millimetres. ML has proved essential in identifying this distinctive signature. The latest tools to identify b-hadrons in the detector make use of cutting-edge architectures, such as transformers with attention mechanisms that carefully study simulated b-hadron decays and learn to reject vertices from regular light quark processes at the best rate achieved in ATLAS to date. Transformers have also been used to learn the complex signature of a particle decaying to two b-hadrons when the decay is collimated and the b-hadron tracks are overlapping.

Machine-learning methods are designed to harness large amounts of data in order to infer new information, making them naturally suited to various data processing tasks in ATLAS.

Once the data have been recorded and the events are reconstructed, it's time to study the underlying physics mechanism that produced the event. ATLAS physics analyses are predicated on effective solutions to classic signal-to-noise problems. Many processes of interest are incredibly rare and can be challenging to distinguish among the billions of ordinary proton collisions. Here is where ML can shine: its broad ability to exploit subtle features within a complex and high-statistics dataset make it a primary workhorse for isolating interesting signal processes.

Figure 2: The output BDT score distribution in the H→bb observation analysis. Here the black points correspond to ATLAS data, the red VH line represents the signal process of associated production of the Higgs boson with a vector boson scaled up by a factor of 50, and all other histograms represent Standard Model background processes. The BDT provides excellent separation of the signal VH process from background, enabling observation of the signal despite its very rare occurrence. (Image: ATLAS Collaboration/CERN)

No particle has held more interest in the past decade than the Higgs boson, discovered in 2012 by the ATLAS and CMS experiments and met with great fanfare and excitement. Understanding and characterising the Higgs boson and the underlying mechanism of mass generation remains an essential goal of high-energy physics today, and ML is in use throughout Higgs-boson analyses. The 2018 observation of the Higgs boson in its most common, but trickiest, decay channel, H→bb, made use of a classic boosted decision tree (BDT) architecture to classify the Higgs-boson signal from the overwhelming background of multijet processes to make observation possible (Figure 2).

ML has also enabled unprecedented study of the top quark, the heaviest known particle and one with a particularly interesting connection to the Higgs boson. In 2023, researchers adapted a graph neural network to model collisions in a geometrical way using the particles produced during the collision and their relationship to one another in the detector space. Training this model to separate the rare four-top-quark-production process from SM backgrounds allowed ATLAS to make its first statistically confident observation of such events, along with a measurement of its production rate and constraints on key possible extensions of the SM.

While these examples of ML to isolate a specific signal demonstrate the depth of its effectiveness, another implementation of ML can reveal its breadth. A growing interest in the LHC community in anomaly detection has led to the proliferation of ML methods that can isolate unusual phenomena from a well-known background model. Such an approach lowers the need to rely on a specific signal model, making these search techniques very broad and sensitive to new physics that may have been missed by previous analysis approaches. In recent years, ATLAS published its first use cases of anomaly detection, implemented via ML algorithms without complete labelling information of training inputs, all in the context of searches for new heavy particles decaying to two-object final states (Figure 3). These analyses leveraged the power of data-driven ML training through a mix of conventional and novel architectures to perform model-independent searches for new particles with a variety of mass and decay hypotheses, providing an invaluable new approach for extracting the most from the ATLAS dataset.

Figure 3: An efficiency map of a neural net output, trained with the weakly supervised Classification WithOut LAbels (CWoLa) method in a dijet resonance search. The axes of the plot, m1 and m2, refer to the measured masses of each of the two jets in the final state. The neural net output has low efficiency (corresponding to signal-like events) at the location of the injected signal (marked with a green X), demonstrating enhanced sensitivity to beyond the Standard Model hypotheses. (Image: ATLAS Collaboration/CERN)

Subtleties and challenges

Despite these successes, there's no such thing as a flawless solution. While ML can offer incredible benefits to ATLAS throughout all stages of the analysis chain, its usage has to be closely coupled with continuous monitoring. Models can inherit unintended biases in the course of training, leading it to make spurious or, even worse, incorrect inferences. The risk of such biases is so significant that it has spawned a broader subfield of AI alignment and safety, and must be carefully considered when applying ML tools to produce physics results.

Luckily there are many ways ATLAS physicists can tackle this challenge. One potential source of such bias emerges from the use of simulated collision events to develop ML tools. While physicists have invested decades into generating accurate and fast simulations, there are still some known ways in which their predictive capabilities can break down. Furthermore, the development of a tool using a particular selection of data with limited statistical power can often require the intentional decorrelation of the model's learned conclusions from certain sensitive properties that should not be considered. To address these issues, physicists make use of dedicated de-biasing or decorrelation techniques from the HEP-ML research community, such as moment decomposition or distance correlation. The limitations of statistical power in simulation samples used for training can also be mitigated through the use of fast simulation methods, which use ML to circumvent the costly full Monte Carlo simulation chain by making fast estimations of key collision and detector properties.

While machine learning can offer incredible benefits to ATLAS throughout all stages of the analysis chain, its usage has to be closely coupled with continuous monitoring.

On top of it all, developing, training and running these advanced algorithms takes a staggering amount of power. To run and adequately cool the mainframes and supercomputers of the CERN Data Centre takes about 37 gigawatt-hours per year, about 3% of CERN's total annual electrical consumption when the LHC operates. While this computing covers all CERN operations, including many applications beyond AI/ML development, producing this quantity of electricity has a significant carbon footprint. The growing role of AI/ML, combined with the uptake of larger and larger models, means that associated power consumption will likely increase as well; for context, Open AI’s ChatGPT uses half a gigawatt-hour daily! Greener approaches are being investigated to continue these operations at CERN in an increasingly climate-focused society. Through dedicated sustainability initiatives, CERN is working with experts across areas of research to find environmentally friendly data management solutions and greener ways to run collider experiments.

A new AI era of scientific research

With this striking history of success, and expectations for computational power to continue its tremendous rise, the future of ML in high-energy physics is bright. ATLAS researchers are collaborative by nature, and much of the work described here wouldn't be possible without close ties to the computer science and AI/ML research communities. Maintaining and expanding these relationships means that physics experimentation will continue to benefit from the latest and greatest in ML algorithms and software capabilities. A recent push across CERN to provide more "open data" recorded by the experiments will further engage researchers outside of HEP who can benefit from the uniquely complex and high-statistics LHC datasets to design and optimise their tools.

Beyond the horizons of ATLAS, AI/ML techniques are similarly impacting the broader landscape in physics. Within theoretical physics, ML offers the promise of dramatically reducing computation cost/time of challenging calculations and simulations, among other things. Further, ML is being studied to perform comprehensive optimizations of future detector designs, which comes at an exciting time for the strategic planning of next-generation colliders.

The long-term future of AI will have an impact on our world that is exciting, transformative and yet unimaginable – and things are no different for particle physics. Through continued collaboration and thoughtful planning for the potential ethical and environmental consequences, researchers can properly harness AI/ML to usher in a new era of precision understanding (and potentially groundbreaking discoveries) in particle physics.

The author would like to thank Katarina Anthony, Dan Guest, Andreas Hoecker, Walter Hopkins, Michael Kagan, Zach Marshall, Benjamin Nachman, and Manuella Vincter for their input and feedback.

About the Author

Julia Gonski is a Panofsky Fellow (associate staff scientist) working on the energy frontier at SLAC National Accelerator Laboratory. Her research focuses on novel approaches to searching for beyond the Standard Model physics with the ATLAS experiment, particularly incorporating machine learning (ML) and anomaly detection. She also works on fast ML for electronics in advanced trigger and readout systems, and planning for next-generation global collider facilities.

About the banner image: Graphic representing a neural network transforming data in the ATLAS experiment. (K. Anthony/ATLAS Collaboration)

Scientific articles

Artificial Neural Networks on FPGAs for Real-Time Energy Reconstruction of the ATLAS LAr Calorimeters (ATL-LARG-PROC-2021-001 (2021))
The application of neural networks for the calibration of topological cell clusters in the ATLAS calorimeters (ATL-PHYS-PUB-2023-019 (2023))
Flavour tagging with graph neural networks with the ATLAS detector (ATL-PHYS-PROC-2023-017 (2023))
Transformer Neural Networks for Identifying Boosted Higgs Bosons decaying into bb and cc in ATLAS (ATL-PHYS-PUB-2023-021 (2023))
Observation of four-top-quark production in the multilepton final state with the ATLAS detector (Eur. Phys. J. C 83 (2023))
Dijet resonance search with weak supervision using 13 TeV pp collisions in the ATLAS detector (Phys. Rev. Lett. 125, 131801 (2020))
Anomaly detection search for new resonances decaying into a Higgs boson and a generic new particle X in hadronic final states using 13 TeV pp collisions with the ATLAS detector (Phys. Rev. D 108, 052009 (2023))
Search for new phenomena in two-body invariant mass distributions using unsupervised machine learning for anomaly detection at 13 TeV with the ATLAS detector (arXiv:2307.01612)
Deep Generative Models for Fast Photon Shower Simulation in ATLAS (Computing and Software for Big Science Vol 8 #7 (2024))

A new ATLAS for the high-luminosity era

Katarina Anthony — Wed, 18 Jan 2023 09:49:20 +0000

A new ATLAS for the high-luminosity era

Top HIghlight

False

Katarina Anthony Wed, 18/01/2023 - 10:49

Highlight

False

Stefan Guindon, Christian Ohm and Caterina Vernieri describe the major ‘Phase II’ upgrades taking place to prepare the ATLAS detector for the High-Luminosity LHC.

The discovery of the Higgs boson at the LHC in 2012 changed the landscape of high-energy physics forever. After just a few short years of data-taking by the ATLAS and CMS experiments, this last piece of the Standard Model (SM) was proven to exist. Since then, the Higgs sector has been studied using a rapidly growing dataset and, so far, all measurements agree with the SM predictions within the experimental uncertainties. In parallel, a comprehensive programme of searches for beyond-SM processes has been carried out, resulting in strong constraints on new physics. A harvest of precise measurements of a large variety of processes, confronted with state-of-the-art theoretical predictions, has further supported the SM. However, the theory lacks explanations for, among others, the nature of dark matter, the cosmological baryon asymmetry and neutrino masses. Importantly, the Higgs sector is related to “naturalness” problems that suggest the existence of new physics at the TeV scale, which the LHC can probe (CERN Courier July/August 2022 p47).

The high-luminosity phase of the LHC (HL-LHC) will provide an order of magnitude more data starting from 2029, allowing precision tests of the properties of the Higgs boson and improved sensitivity to a wealth of new-physics scenarios. The HL-LHC will deliver to each of the ATLAS and CMS experiments approximately 170 million Higgs bosons and 120,000 Higgs-boson pairs over a period of about 10 years. By extrapolating Run 2 results to the HL-LHC dataset, this will increase the precision of most Higgs-boson coupling measurements: 2–4% precision on the couplings to W, Z and third-generation fermions; and approximately 50% precision on the self-coupling by combining the ATLAS and CMS datasets. The larger dataset will also give improved sensitivity to rare vector-boson scattering processes that will offer further insights into the Higgs sector.

These precision measurements could reveal discrepancies with the SM predictions, which in turn could inform us about the energy scale of beyond-SM physics. In addition to improving SM measurements, the upgraded detectors and trigger systems being developed and constructed for the HL-LHC era will enable direct searches to better target new physics with challenging signatures. To achieve these goals, it will be essential to achieve a detailed understanding of the detector performance as well as to measure the integrated luminosity of the collected dataset to 1% precision.

Figure 1: The ATLAS ITk, comprising strip (outer layers) and pixel (inner layers) detectors with more than five billion readout channels. (Image: ATLAS Collaboration/CERN)

Rising to the challenge

To cope with the increased number of interactions when proton bunches collide at the HL-LHC, the ATLAS collaboration is working hard to upgrade its detectors with state-of-the-art instrumentation and technologies. These new detectors will need to cope with challenging radiation levels, higher data rates and an extreme high-occupancy environment with up to 200 proton–proton interactions per bunch crossing (see banner image). Upgrades will include changes to the trigger and data-acquisition systems, a completely new inner tracker, as well as a new silicon timing detector (see Figure 2).

The trigger and data-acquisition system will need to cope with a readout rate of 1 MHz, which is about 10 times higher than today. To achieve this, ATLAS will use a new architecture with a level-0 trigger (the first-level hardware trigger) based on the calorimeter and muon systems. Building on the upgrades for Run 3, which started in July 2022, the calorimeter will include capabilities for triggering at higher pseudorapidity, up to |η| = 4. During HL-LHC running, the global trigger system will be required to handle 50 Tb/s as input and to decide within 10 μs whether each event should be recorded or discarded, allowing for more sophisticated algorithms to be run online for particle identification. All the detectors will require substantial upgrades to handle the additional acceptance rates from the trigger.

To cope with the increased number of interactions when proton bunches collide at the HL-LHC, the ATLAS collaboration is working hard to upgrade its detectors with state-of-the-art instrumentation and technologies.

The readout electronics for the electromagnetic, forward and hadronic end-cap liquid-argon calorimeters, along with the hadronic tile calorimeter, will be replaced. The full calorimeter systems, segmented into 192,320 cells that are read out individually, will be read out for every bunch crossing at the full 40 MHz to provide full-granularity information to the trigger. This will require changes to both front-end electronics and off-detector components.

The muon system will also see significant upgrades to the on-detector electronics of the resistive plate chambers (RPCs) and thin-gap chambers (TGCs) responsible for triggering on muons, as well as the muon drift tubes (MDTs) responsible for measuring the curvature of the tracks precisely. The MDTs will also be used for the first time in the level-0 trigger decisions. These improvements will allow all data to be sent to the back-end at 40 MHz, removing the need for readout buffers on the detector itself. All hits in the detector will be used to perform trigger logic in hardware using field programmable gate-arrays. Additional improvements to increase the trigger acceptance for muons will come in the form of a new layer of RPCs to be installed in the inner barrel layer, along with new MDTs in the small sectors. The Muon New Small Wheel system was installed during Long Shutdown 2 (LS2) from 2019 to 2022 (CERN Courier November/December 2021 p27) and is located inside the end-cap toroid magnet containing both triggering and precision tracking chambers. Additional RPC upgrades were also made in the barrel leading up to Run 3, and the TGCs will be upgraded in the endcap region of the muon system during LS3.

Figure 2: Schematic of the major upgrades to the ATLAS detector for the HL-LHC era. (Image: ATLAS Collaboration/CERN)

State-of-the-art tracking

The success of the research programme at the HL-LHC will strongly rely on the tracking performance, which in turn determines the ability to efficiently identify hadrons containing b and c quarks, in addition to tau and other charged leptons. Reconstructing individual particles in the HL-LHC collision environment with thousands of charged particles being produced within a region of about 10 cm will be very challenging. The entire tracking system, presently consisting of pixel and strip detectors and the transition radiation tracker, will be replaced by a new all-silicon pixel and strip tracker – the ITk. This will feature higher granularity, increased radiation hardness and readout electronics that allow higher data rates and a longer trigger latency. The new pixel detector will also extend the pseudorapidity coverage in the forward region from |η| < 2.5 to |η| < 4, increasing the acceptance for important physics processes like vector-boson fusion (see Figure 1).

The ITk will comprise nine barrel layers, positioned at radii from 33 mm out to 1 m from the beam line, plus end-cap rings. It will be much more complex with respect to the present ATLAS tracker, featuring 10 times the number of strip channels and 60 times the number of pixel channels. The strip detectors will cover a total surface of 160 m²with 60 million readout channels, and the pixels an area of 13 m² with more than five billion readout channels. The innermost layer will be populated with radiation-hard 3D sensors, with pixel cells of 25 × 100 µm² in the barrel part and 50 × 50 µm² in the forward parts for improved tracking capabilities in the central and forward regions. Prototypes of the end-cap ring for the inner system and of the strip barrel stave are at an advanced stage (see Figure 3). A unique feature of the trackers at the HL-LHC is that they will be operated for the first time with a serial powering scheme, in which a chain of modules is powered by a constant current. If the modules were to be powered in parallel, the high total current would lead to either high power losses or a large mass of cables within the volume of the detector, which would impact the tracking performance.

Figure 3: Loaded prototypes for (left) a pixel end-cap ring of the inner system and (right) a strip barrel stave of the ITk detector. (Image: ATLAS Collaboration/CERN)

Given the challenging conditions posed by the HL-LHC, ATLAS will construct a novel precision-timing silicon detector, the High-Granularity Timing Detector (HGTD), which provides a time resolution of 30 to 50 ps for charged particles. The detector will cover a pseudorapidity range of 2.4 < |η| < 4 and will comprise two double-sided silicon layers on each side of ATLAS with a total active area of 6.4 m². The precise timing information will allow the collaboration to disentangle proton–proton interactions in the same bunch crossing in the time dimension, complementing the impressive spatial resolution of the ITk. Low-gain avalanche diodes (see Figure 4) provide timing information that can be associated with tracks in the forward regions, where they are more difficult to assign to individual interactions using spatial information. With a timing resolution six times smaller than the temporal spread of the beam spot, tracks emanating from collisions occurring very close in space but well-separated in time can be distinguished. This is particularly important in the forward region, where reduced longitudinal impact-parameter resolution limits the performance.

Building upon the insertable B-layer cooling system used since the start of Run 2, and to reduce the material budget, ATLAS will use a two-phase CO₂ cooling system for the entire silicon ITk and HGTD detectors. These will allow the detectors to be cooled to around –35 °C during the entire lifetime of the HL-LHC. The low temperature is required to protect the silicon sensors from the expected high radiation dose received during their lifetime. Two-phase CO₂ cooling is an environmentally friendly option compared to other suitable coolants. It provides a high heat transfer at reasonable flow parameters, a low viscosity (thus reducing the material used in the detector construction) and a well-suited temperature range for detector operations.

Luminous future

Precise knowledge of the luminosity is key for the ATLAS physics programme. To reach the goal of percent-level precision at the HL-LHC, ATLAS will upgrade the LUCID (Luminosity Cherenkov Integrating Detector) detector, a luminometer that is sensitive to charged particles produced at the interaction point. This is incredibly challenging given the number of interactions expected to be delivered by the machine, and the requirements on radiation hardness and long-term stability for the lifetime of the experiment. The HGTD will also provide online luminosity measurements on a bunch-by-bunch basis, and additional detector prototypes are being tested to provide the best possible precision for luminosity determination during HL-LHC running. Luminometers in ATLAS provide luminosity monitoring to the LHC every one to two seconds, which is required for efficient beam steering, machine optimisation and fast checking of running conditions. In the forward region, the zero-degree calorimeter, which is particularly important for determining the centrality in heavy-ion collisions, is also being redesigned for HL-LHC running.

The HL-LHC will deliver luminosities of up to 7.5 × 10³⁴ cm^–2s^–1, and ATLAS will record data at a rate 10 times higher than in Run 2. The ability to process and analyse these data depends heavily on R&D in software and computing, to make use of resource-efficient storage solutions and opportunities that paradigm-shifting improvements like heterogeneous computing, hardware accelerators and artificial intelligence can bring. This is needed to simulate and process the high-occupancy HL-LHC events, but also to provide a better theoretical description of the kinematics.

Figure 4: An 8-inch prototype wafer of low-gain avalanche diodes for the High Granularity Timing Detector. (Image: ATLAS Collaboration/CERN)

New era

The Phase-II upgrade projects described are only possible through collaborative efforts between universities and laboratories across the world. The research teams are currently working intensely to finalise the designs, establish the assembly and testing procedures, and in some cases start construction. They will all be installed and commissioned during LS3 in time for the start of Run 4, currently planned for 2029.

The HL-LHC will provide an order of magnitude more data recorded with a dramatically improved ATLAS detector. It will usher in a new era of precision tests of the SM, and of the Higgs sector in particular, while also enhancing sensitivity to rare processes and beyond-SM signatures. The HL-LHC physics programme relies on the successful and timely completion of the ambitious detector upgrade projects, pioneering full-scale systems with state-of-the-art detector technologies. If nature is harbouring physics beyond the SM at the TeV scale, then the HL-LHC will provide the chance to find it in the coming decades.

This feature first appeared in the CERN Courier Jan/Feb 2023 issue.

About the banner image: A simulated tt event in a proton–proton collision at 14 TeV at the HL-LHC, including approximately 200 pileup interactions in the same bunch crossing. No fewer than 88 primary vertices (blue balls) are reconstructed along the beam line, each producing many particles (orange tracks). In total, more than 2000 tracks are reconstructed in the proton–proton interaction. (Image: ATLAS Collaboration/CERN)

Looking inside trillion degree matter with ATLAS at the LHC

Katarina Anthony — Fri, 13 May 2022 16:06:35 +0000

Looking inside trillion degree matter with ATLAS at the LHC

Top HIghlight

False

Katarina Anthony Fri, 13/05/2022 - 18:06

Highlight

False

Stars, planets, animals, plants, you and me – everything that can be directly observed in our Universe is ordinary matter. But the Universe didn’t always look this way. This ATLAS feature describes the quark-gluon plasma, a unique state of matter that existed shortly after the Big Bang. The nature and properties of this matter are being revealed at CERN’s ATLAS Experiment.

Most of the mass of ordinary matter comes from protons and neutrons, which build up the atomic nucleus. Protons and neutrons belong to a family of particles called hadrons, which are themselves composite subatomic particles made of two or more quarks held together by the strong force. They are analogous to molecules that are held together by the electromagnetic interaction. In the Standard Model of particle physics, the theory that best describes our current understanding of elementary particles and their interactions, gluons are the carriers of the strong force. Gluons compose the majority of particles inside of hadrons and, despite being massless themselves, are the major contributors to a hadron’s mass due to the interaction of the quarks with the gluon field.

A particle’s ability to interact via the strong force is described by its colour. Quarks and gluons, collectively known as partons, are thus colour-charged particles. Because of a phenomenon called colour confinement, they cannot be isolated from hadrons and therefore cannot be directly observed in normal conditions. Thus, the total colour charge of hadrons must be zero (they are colourless). Two families of hadrons can be identified: mesons, which consist of a quark of one colour and an antiquark of the corresponding anticolour, and baryons, which are composed of three quarks of different colours but still carrying net-zero colour charge. Today’s Universe is baryon dominated, with quarks bound into ordinary matter. In fact, antibaryons almost do not exist in nature and can only be produced in relatively high abundances in the laboratory.

But what if conditions can be established in which individual partons can be deconfined?

Today’s Universe is dominated by quarks and gluons bound into ordinary matter. How would they behave if they were deconfined?

In the 1960s, German physicist Rolf Hagedorn from the Theory Division at CERN, while working on a statistical model for particle production, made an important prediction of particle yields at the highest accelerator energies available at the time (from the Proton Synchrotron at CERN). In his work, he introduced the concept of a “fireball”. In this approach, all the energy from a particle collision was regarded to be contained within a small space-time volume from which particles were radiated – like a burning fireball.

Hagedorn’s particle-production models turned out to be remarkably accurate at predicting yields for many different types of secondary particles produced from the primary high-energy collisions. He understood that while new particles are being produced, and more and more energy is poured into the system, the temperature does not increase. Instead it is the entropy that increases with the collision energy. If the number of particles of a given mass increases exponentially, the temperature becomes stuck at a limiting value called the Hagedorn temperature. It amounts to approximately 2 terakelvin (2 followed by 12 zeros), which is a million times higher than the temperature of the core of the Sun.

Later on, the Hagedorn temperature was interpreted as a limit where ordinary matter is no longer stable, and must either "evaporate" or be converted into quark matter; as such, it can be thought of as the "boiling point" of hadronic matter. In electronvolts, this temperature amounts to about 160 MeV, which is about 15% above the mass of the lightest hadron, the pion. Therefore, matter at Hagedorn temperature or above will spew out fireballs of new particles, and the ejected particles can then be detected by particle detectors. Thus, quarks and gluons cannot be separated from their parent hadron without producing new hadrons.

This brings us to a concept of the quark–gluon plasma (QGP), a state of matter in which the quarks and gluons that make up the hadrons of matter are freed of their strong attraction to one another under extremely high energy densities. In the QGP, quarks and gluons are deconfined. By producing the QGP in the laboratory, researchers can recreate and study the high energy density conditions that prevailed in the early Universe, shortly after the Big Bang, when matter was formed from free quarks and gluons. This corresponds to the time interval of 10⁻¹⁰–10⁻⁶ s after the Universe was born.

So far, the only way for physicists to produce the QGP is through the collision of two heavy atomic nuclei (called heavy ions as atoms are fully ionised). These nuclei are accelerated to energies of more than a hundred GeV, thus heating matter well above the Hagedorn temperature. Using the result of a head-on collision in a volume approximately equal to that of an atomic nucleus, it is possible to reproduce the conditions in the very first moments of the Universe.

The idea that particle production could help achieve a high enough particle density to allow deconfinement was first recognised around 1978. Physicists realised that matter made of hadrons would melt into a boiling QGP phase. In the following decade, two experimental facilities began to search for this new phase of matter: lead nuclei were collided at the Super Proton Synchrotron (SPS) at CERN, and gold nuclei were collided the Relativistic Heavy Ion Collider (RHIC) at the Brookhaven National Laboratory (BNL) near New York City. The nuclei accelerated by these facilities would travel almost to the speed of light before being directed towards each other – creating a "fireball" in the rare event of a collision.

But how can physicists recognise whether the QGP medium is created in a collision? Several probes have been proposed to search for signatures of deconfined matter. This feature article will be highlighting strangeness production and elliptic flow, with the phenomena of jet quenching reviewed in detail.

By producing the quark–gluon plasma in laboratories, researchers can study the conditions that prevailed in the early Universe, when matter was formed from free quarks and gluons.

Strangeness – that is, the production of strange quarks in relativistic heavy-ion collisions – is a signature and a diagnostic tool of the QGP formation. In 1982, strange quarks (and their antiquarks) were first proposed as signatures of the QGP. This relies on the fact that strange quarks and antiquarks are not present in ordinary matter since, when produced, they promptly undergo a "weak" interaction decay, akin to naturally radioactive isotopes. Moreover, the mass of strange quarks (and antiquarks) is below and close to the temperature at which protons, neutrons and other hadrons dissolve into quarks. Strange quarks are thus sensitive to the conditions, structure and dynamics of the deconfined matter phase and, if their number is large, it can be assumed that deconfinement conditions were reached. However, in the hot QGP environment, strange quarks can be produced abundantly and, due to their natural radioactivity, are relatively easy to observe. The ensuing production of matter particles comprising strangeness (especially multi-strange antibaryons) is the “gold standard” for the formation of the QGP.

In 2006, the NA57 Experiment at CERN found that hadrons made entirely from newly created quarks were produced up to 15-20 times more abundantly in heavy-ion reactions when compared to expected values from the reference proton-proton system. The pattern of enhancement follows the QGP prediction. There is no known explanation for these experimental results other than QGP formation.

Perhaps the most striking feature of heavy-ion collisions is that the geometry of the initial overlap between the two nuclei is reflected in the anisotropy of the final particle distributions. In order to understand this, it is first useful to classify collisions according to their impact parameter. Collisions are categorised into “centrality classes” depending on the magnitude of the impact parameter: central collision events with the largest overlap correspond to small impact parameter values, and peripheral collision events with the smallest overlap correspond to large impact parameter values. This is known as elliptic flow and it is illustrated in Figure 1.

Figure 1: (left) ATLAS measurement of the angular distribution of particles with respect to the reaction plane for lead-lead collisions as a function of centrality [PLB 707 330 (2012)]. (middle) Illustration depicting the overlap of two lead nuclei colliding. In central collisions the overlap is nearly circular and in peripheral collisions the overlap region is more eccentric. (Image: ATLAS Collaboration/CERN)

Elliptic flow is a fundamental observable. It reflects the initial almond-shaped, azimuthally asymmetric region in the transverse plane of the nuclear overlap region, directly translated into the observed momentum distribution of particles. In very central collisions, elliptic flow is small, because the overlap of the two nuclei is nearly circular. As the overlap becomes more eccentric, the elliptic flow becomes larger. The reaction plane defines the minor axis of this eccentric shape. The high energy density in the overlap region creates pressure gradients which are largest in the reaction plane and particles are preferentially pushed out in that direction, similar to a ball rolling faster down a steep hill than a gentle slope. Elliptic flow is especially sensitive to the early stages of system evolution, when the spatial anisotropy is at its largest. In particular, the magnitude of the anisotropy is sensitive to the viscosity of the expanding fireball medium. The elliptic flow is measured to be nearly as strong as it could be, given the eccentricity of the overlap region. This means that the viscosity is nearly as small as it could be (see here). Elliptic flow is interpreted as strong evidence for the existence of the QGP.

Jet quenching is another unique phenomenon that can occur in relativistic heavy-ion collisions. Jets are produced in both proton-proton and heavy-ion collisions at the LHC. In both collision systems they are produced by two partons, one from each incoming projectile, colliding and exchanging a very large amount of momentum. These partons scatter at a large angle. As they leave the collision region, the scattered parton develops a shower. This is the process by which a colour-charged quark or gluon becomes an observable collection of colour-neutral particles. In proton-proton collisions, this showering process happens in the vacuum, but in nucleus-nucleus collisions, the shower develops inside the QGP (see Figure 2). Thus the showering process actually probes the short distance scale properties of the QGP itself. This makes jets one of the most powerful probes of QGP properties.

In these interactions, the energy of the partons is reduced through collisional energy loss and medium-induced gluon radiation, the latter being the dominant mechanism in the QGP. The effect of jet quenching in the QGP provides a main motivation for studying jets, as well as high-momentum particles and their correlations, in heavy-ion collisions.

Figure 2: Jet production in proton-proton collisions (left) and nucleus-nucleus collisions (right). In both cases two incoming quarks (labelled “q”) scatter off each other. The outgoing jets are shown as arrows (representing the particles in the jet). In nucleus-nucleus collisions the jet develops inside the QGP (orange region) and is modified by its interactions with it. (Image: M. Rybar/ATLAS Collaboration)

The heavy-ion programme at ATLAS was launched spectacularly with a paper on the discovery of jet quenching. In contrast to the back-to-back jets observed in proton-proton collisions, in lead-lead collisions, one jet was observed with the opposing jet either entirely missing or having a much smaller energy. Figure 3 shows an event display from one lead-lead collision with a single prominent jet and no clear corresponding opposing jet. Overall, pairs of jets (dijets) in lead-lead collisions were found to be much less balanced than in proton-proton collisions.

This observation was based on the first LHC lead-lead data from 2010. However, with the increased integrated luminosity of lead-lead collisions accumulated over the last several years at the LHC, more interesting questions can be asked and answered. The R_AA of jets as a function of their momentum in the transverse direction (perpendicular to the beam) was measured in 2019. R_AA is the ratio of the number of jets measured in lead-lead collisions compared to measurements from proton-proton collisions scaled by a factor to account for the thickness of the lead nuclei compared to protons; R_AAbeing less than one indicates jets are losing energy. R_AA of inclusive jets in the most head-on lead-lead collisions was found to be between 0.4 and 0.6 (depending on the momentum of the jet). This indicates substantial energy loss.

LHC physicists are now focused on measurements addressing how this jet suppression actually happens. Two questions of current interest are:

How does the suppression depend on the type of jet?
How does the energy loss depend on the path length the jet travelled through the QGP?

Figure 3: Three representations of an event display of a single lead-lead collision recorded in ATLAS. (left) Beams coming into and out of the page with a prominent jet in the upper left and a diffuse energy distribution toward the bottom right. Calorimeter energy (middle) and charged particle track (right) distributions for the same event. (Image: ATLAS Collaboration/CERN)

How does the suppression depend on the type of jet?

Jets can be classified in a number of ways. One common way is to classify them by the type of parton that gave rise to the jet – for example quarks and gluons. At the LHC, most jets are gluon jets. Separating quark jets from gluon jets is a difficult task on a jet-by-jet basis, but creating a sample of jets with an enhanced quark fraction can be done by looking for jets which are opposite to a photon rather than another jet. Those jets are dominantly produced in the process quark+gluon → photon+quark. The R_AA of these photon-associated jets is also shown in Figure 4. These jets have a larger R_AA and are thus less suppressed than inclusive jets. This is the expected effect, as gluons jets couple to the QGP more strongly because the colour charge of gluons is larger than that of quarks. It is also possible to sort quark jets by the different types of quarks. Bottom quarks are of particular interest due to their very large mass. The R_AA of jets arising from bottom quarks is found to be larger than that of inclusive jets. This is expected, at least in part, from a suppression of gluon radiation inside the QGP due to the quark mass.

Another way to classify jets is by their structure. The developing parton shower is what interacts with the QGP, therefore jets with different structures interact differently with the QGP. ATLAS researchers have classified jets according to the angular distance between two prongs of the parton shower (r_g). These prongs are constructed via the Soft-Drop procedure and the R_AA is calculated as a function of both transverse momentum (p_T) and r_g. Very collimated (small r_g) jets are shown to be suppressed less than inclusive jets, while wide (large r_g) jets are suppressed more than inclusive jets.

As the structure and partonic origin of the jet are expected to be related to each other, these two ways of classifying jets are not independent. What they both show is that the amount any jet is quenched by depends on the details of the jet itself – this is exactly what is needed to understand the QGP from its interactions with jets.

Figure 4: R_AA as a function of the jet transverse momentum (p_T) for various types of jets as shown in the legend. (Image: ATLAS Collaboration/CERN)

How does the energy loss depend on the path length the jet travelled through the QGP?

In order to investigate the path length dependence of energy loss, the angular distributions of jets with respect to the event planes are measured. In this way, one can directly change the average amount of QGP that the jet traverses through the QGP. The variation in the yield with respect to those planes provides information on the sensitivity to the path length. Like the flow measurements discussed earlier, these measurements are directly sensitive to the shape of the QGP. Figure 5 shows a pair of jets created in the QGP and shows the differing path lengths depending on the orientation of the jets with respect to the QGP.

To quantify this, ATLAS physicists have measured the yield of jets with respect to the angle of the impact parameter (Ψ). This has been quantified by measuring the distribution of jets with respect to Ψ. More jets are observed at Ψ = 0 and Ψ = π than at Ψ = π/2; this distribution is fit to a sinusoidal distribution and the amplitude of that variation is v₂. Figure 6 shows v₂ of jets as a function of centrality. In the most central collisions (smallest impact parameter), v₂ is very small; this is understood to be due to the fact that the path length differences are very short when the QGP is nearly circular. However, in less central collisions v₂ grows as the eccentricity of the QGP increases. This clearly shows that the number of jets varies with the path length through the QGP.

Figure 5: Illustration of a pair of jets (green arrows) and their path length (solid black lines) through the QGP (orange region). The jets were produced at the place where the two grey arrows meet. (Image: M. Rybar/ATLAS Collaboration)

Figure 6: v_2 as a function of centrality for jets in lead-lead collisions. (Image: ATLAS Collaboration/CERN)

Summary

Recent ATLAS results have shown that jet quenching is sensitive to both the structure of jets as they pass through the QGP and the path length of the jet travelling through. This demonstrates the usefulness of jet measurements for understanding the structure of the QGP. These experimental results are being incorporated into theoretical models which contain the microscopic quark and gluon interactions between the jets and the QGP. These models will have to describe these (and other) data to be possible interpretations of these interactions.

The ATLAS measurements presented here are part of a rich scientific programme seeking to understand the nature and properties of the QGP, and are but a small fraction of the exciting results to come out of ATLAS Run-2 data. Additional data collected in Runs 3 and 4 of the LHC will allow for more differential studies to further constrain the parton type, jet structure and path length dependence of jet quenching.

About the authors

Anne M. Sickles is an Associate Professor of Physics at the University of Illinois at Urbana-Champaign. She is a member of the ATLAS and sPHENIX collaborations and has worked on many experimental studies of the quark-gluon plasma created in ultra-relativistic collisions of nuclei. Her present research deals primarily with measurements of jets and their properties in heavy-ion collisions. Iwona Grabowska-Bold is a Full Professor of Physics at the AGH University of Science and Technology (Kraków, Poland) and member of the ATLAS Collaboration. She has been working on many topics including trigger preparations, data quality for heavy-ion data, as well as data analysis with a focus on weak bosons, heavy quarks and ultra-peripheral collisions (UPC). Her present research activity focuses on understanding strong and electroweak interactions using inelastic and UPC heavy-ion data.

Scientific articles

ATLAS Collaboration: Observation of a Centrality-Dependent Dijet Asymmetry in Lead-Lead Collisions at 2.76 TeV with the ATLAS Detector at the LHC (PRL 105 (2010) 252303)
ATLAS Collaboration: Measurement of the nuclear modification factor for inclusive jets in Pb+Pb collisions at 5.02 TeV with the ATLAS detector (PLB 790 (2019) 108)
ATLAS Collaboration: Comparison of inclusive and photon-tagged jet suppression in 5.02 TeV Pb+Pb collisions with ATLAS (ATLAS-CONF-2022-019)
ATLAS Collaboration: Measurement of the nuclear modification factor of b-jets in 5.02 TeV Pb+Pb collisions with the ATLAS detector (arXiv: 2204.13530)
ATLAS Collaboration: Measurement of substructure-dependent jet suppression in Pb+Pb collisions at 5.02 TeV with the ATLAS detector (ATLAS-CONF-2022-026)
ATLAS Collaboration: Measurements of azimuthal anisotropies of jet production in Pb+Pb collisions 5.02 TeV with the ATLAS detector (arXiv: 2111.06606)
E. V. Shuryak: Quantum chromodynamics and the theory of superdense matter (Phys. Rept. 61 (1980) 71-158)
J. Rafelski, B. Müller: Strangeness Production in the Quark-Gluon Plasma (Phys. Rev. Lett. 48 (1982) 1066)
NA57 Collaboration: Enhancement of hyperon production at central rapidity in 158 A GeV/c Pb–Pb collisions (J. Phys. G 32, 427 (2006))
J. E. Bernhard, et al.: Bayesian estimation of the specific shear and bulk viscosity of quark–gluon plasma (Nat. Phys 15 (2019) 11)
Y. L. Dokshitzera, D.E. Kharzeev: Heavy-quark colorimetry of QCD matter (PLB 519 (2001) 199)
A. J. Larkoski, et al.: Soft drop (JHEP 05 (2014) 146)

The Last Quark

Katarina Anthony — Fri, 23 Jul 2021 08:23:59 +0000

The Last Quark

Top HIghlight

False

Katarina Anthony Fri, 23/07/2021 - 10:23

Highlight

False

Feature

Richard Hawkings

top quark

One of the most successful paradigms in physics is to try to understand complex phenomena in terms of simpler building blocks. In the 19th century, Russian chemist Dimitri Mendeleev noticed that the properties of approximately 100 known chemical elements showed distinct patterns when arranged in a table according to their mass. With subsequent discoveries and insights from Antoine Becquerel, J.J. Thomson, Ernest Rutherford, Marie and Pierre Curie, James Chadwick and others, the modern ‘periodic table’ we know from school chemistry was born.

The disparate chemical elements are all made of the same building blocks: a compact ‘nucleus’ made of similar numbers of protons and neutrons, surrounded by a cloud of orbiting electrons. Each chemical element has a characteristic number of electrons, equal to the number of protons in its nucleus. The number and allowed arrangements of the electrons (governed by the laws of quantum mechanics) determine all the element’s chemical properties; whether it is stable or reactive, acid or alkaline, metal or non-metal etc. The rich structure of the periodic table can be explained by fundamental building blocks of electrons, protons and neutrons, and the ways in which they interact and fit together.

After the Second World War, increasingly powerful particle accelerators allowed protons or electrons to be boosted to high energies and smashed into dense material targets or, later, into other accelerated particles in head-on collisions. The debris from such collisions consisted not only of familiar protons, neutrons and electrons, but an ever-expanding ‘zoo’ of heavier and more exotic particles (collectively called ‘hadrons’) that would appear fleetingly and then rapidly decay into lighter, more familiar particles. As in Mendeleev’s periodic table, these particles could be arranged by their properties into intriguing patterns, which had their roots in a branch of mathematics called ‘group theory’.

In the 1960s, physicists proposed that hadrons were actually made up of combinations of two or three fundamental constituents, which they called ‘quarks’.

Murray Gell-Mann, one of the proposers of the quark model and recipient of the Nobel Prize in Physics in 1969, visiting the ATLAS experiment in 2012. (Image: M. Brice/CERN)

In the 1960s, physicists Murray Gell-Mann and George Zweig proposed that most of these particles were actually made up of combinations of two or three fundamental constituents, which they called ‘quarks’ (pronounced to rhyme either with ‘Mark’ or ‘York’). Most hadrons could be explained as combinations of just two quarks, dubbed ‘up’ and ‘down’, while a class of so-called ‘strange’ hadrons with anomalously long lifetimes could fit into the model by postulating a third ‘strange’ quark. Soon a fourth ’charm’ quark was proposed, providing an elegant solution to some problems in the description of weak interactions (the theory that underlies radioactive decay).

In parallel, experimentalists at the Stanford Linear Accelerator Center in California, USA, were colliding high-energy accelerated electrons with stationary protons, and measuring the angles at which the electrons came out from the collision. These showed that the proton was not a featureless solid object, but instead acted as if it contained three hard point-like constituents, exactly as would be expected in the quark model. Subsequent experiments have confirmed this picture: the proton is made of two up quarks and one down quark, the neutron is made of two down quarks and one up quark, and the other hadrons are all made of other combinations of quarks (and their antimatter partners, antiquarks). In the last few years, physicists have also found evidence for hadrons containing four or even five quarks.

But could there be more quarks? A fifth, dubbed the ‘bottom’ or ‘beauty’ (b) quark, was inferred in 1977 after the discovery of a new heavy hadron in proton–nucleus collisions at the Fermi National Accelerator Laboratory (Fermilab) near Chicago, USA. The quark model requires that quarks come in pairs, resulting in three ‘generations’ each with two quarks: up and down; strange and charm; and, finally, bottom and top – the sixth quark to complete the pair. Intriguingly, by this time the electron was also known to have two heavier cousins – the muon and tau – which, together with neutrinos, form three generations of leptons, another family of particles. The motivation for a sixth quark appeared compelling. But where might this quark be found?

Hunting the top quark

One of the fundamental characteristics of a particle is its mass, which determines not only how heavy it is (its weight under gravity), but how hard it is to accelerate. For example, a car is much harder to push by hand than a bicycle. Einstein’s famous equation E=mc² tells us that mass, m, and energy, E, are proportional (related by the speed of light, c, squared). This means that a heavy particle requires much more energy to create than a light one. Particle physicists use this relationship to measure particle masses in terms of ‘electron volts’, where 1 electron volt is the energy acquired by an electron when it is accelerated by an electric field of 1 volt. This is an extremely small unit, and the proton has a mass of about 1 giga electron volts, i.e. 1000 million electron volts or 1 GeV for short, equivalent to 1.8x10^-27 kg.

In these units, the up, down and strange quarks have masses of less than 0.1 GeV; the charm quark, 1.3 GeV; and the bottom quark, 4.2 GeV. So, it was natural to assume that the top quark fit this sequence – with a mass of perhaps 10 to 20 GeV. Surely, after the discovery of the bottom quark, the top quark would be ‘just around the corner’.

As each new and more powerful particle accelerator or collider began its work, physicists hoped it would have enough energy to discover the top quark. But no convincing hints were seen, and the first data from the CDF and D0 experiments at Fermilab’s Tevatron proton–antiproton collider in the early 1990s showed that if the top quark exists, its mass must be more than about 100 GeV. On the other side of the Atlantic Ocean, experimentalists at CERN’s Large Electron–Positron Collider (LEP) in Geneva, Switzerland, were probing the top quark indirectly through precise measurements of the decays of the Z boson (a fundamental particle connected to the electroweak interaction) into different types of quarks and antiquarks. Due to conservation of energy, the Z boson, with a mass of about 90 GeV, would not be heavy enough to decay into a top quark–antiquark pair if the top quark (and top antiquark) mass is greater than 45 GeV. Nevertheless, the relative proportions of Z boson decays into other types of quarks could be subtly influenced by even the possibility of decays into top quarks, and measurements at LEP suggested the top-quark mass should be somewhere between 150 and 200 GeV. But did it really exist?

As each new and more powerful particle accelerator or collider began its work, physicists hoped it would have enough energy to discover the top quark. But no convincing hints were seen.

In a particle collider, collisions between high-energy protons and antiprotons can be understood as collisions between two opposing ‘bags’ of quarks or antiquarks, the constituents of the (anti)protons. The total energy of the accelerated proton is shared among the three quarks, with a fraction also going to gluons, other particles in the proton that represent the force binding the three quarks together. Physicists expected that the most likely way to produce top quarks in Tevatron’s 1.8 TeV (1800 GeV) collisions was through a head-on collision of a quark from the proton and an antiquark from the antiproton, producing a top quark and corresponding top antiquark (a 'top-pair'). Again, due to conservation of energy, this process would require the initial quark and antiquark to have at least twice the energy equivalent of the top-quark mass – that’s more than their fair share of their parent proton’s energy. This is rather unlikely, making top-pair production a rare process that becomes even rarer if the top quark is very heavy.

In the early 1990s, the CDF and D0 experiments began to accumulate evidence for the production of top–antitop pairs in their data sample. They finally announced their joint discovery of the top quark in 1995, measuring its mass to be about 180 GeV. This was around 10 times larger than the original expectation, but in agreement with the indications from LEP. Over the next 16 years, tens of thousands of top-quark events were recorded and studied by the two experiments, allowing physicists to build a first portrait of this new particle. As far as they could see, it behaved just as a partner of the bottom quark would be expected to – but why was it so heavy?

Top quarks come to Europe

When CERN’s Large Hadron Collider (LHC) turned on at its initial 7 TeV collision energy in spring 2010, catching a first glimpse of the top quark was high on many ATLAS physicists’ wish lists. Due to the much larger collision energy, the production of top–antitop pairs at the LHC is dominated by the collision of gluons from the two colliding protons, with a rate much higher than at the Tevatron.

The top quark is not a stable particle. Rather than teaming up with other quarks to form hadrons, a process that takes several 10^-24 seconds, it decays within 10^-25 seconds to a W boson and a bottom quark. The W boson in turn decays into a quark–antiquark pair (usually up–down or charm–strange), and these quarks give rise to collimated sprays, or ‘jets’, of particles all heading roughly in the same direction in the detector. Alternatively, the W boson may decay to a lepton (e.g. an electron or muon) plus a neutrino. The neutrino passes through the ATLAS detector and escapes undetected, leaving an energy imbalance in the collision event once all the detected particles are accounted for. The b-quark also produces a particle jet, but one containing a hadron made up of a b-quark and a lighter antiquark, which travels a few millimetres in the detector before decaying. The charged particle tracks from this decay can be precisely reconstructed in ATLAS, and distinguished from other particles coming from the collision point. This allows the jet to be ‘b-tagged’ as likely to have been produced by a b-quark. With two top quarks producing two W bosons and two b-quark jets, the signature of top-pair production in the ATLAS detector is spectacular. It also provides an excellent ‘work out’ for the detector, and the processing, calibration and reconstruction of the recorded data.

The LHC accumulated data slowly at first, but by the summer of 2010 both the ATLAS and CMS collaborations were able to report the first observation of top-quark production in Europe. This milestone for the LHC programme set the stage for the discovery of the even more elusive Higgs boson just two short years later. One of the first top-pair events recorded in ATLAS, with the top quarks decaying to an electron, a muon, missing energy indicating neutrinos and a b-tagged jet, is shown in Figure 1.

Figure 1: One of the first candidate top-pair events recorded by ATLAS in July 2010. In the display of the inner detector to the right of the picture, the electron is shown by a short red track pointing to an energy cluster in the calorimeter (green blob). The muon is shown by the long red track going to the left, passing through three sets of muon chambers (blue boxes). The upper right inset shows a zoom on the collision point, with 3 blue tracks originating from a secondary vertex away from the main collision vertex, indicating the decay of a hadron containing a b-quark. (Image: ATLAS Collaboration/CERN)

The first observation of the top-quark at the LHC set the stage for the discovery of the even more elusive Higgs boson just two short years later.

As the LHC accumulated more data, and the collision energy was increased to 8 TeV in 2012 and 13 TeV in 2015, ever larger samples of top quarks were recorded by the ATLAS and CMS experiments, allowing it to be studied in greater detail. One of the primary measurements is that of the ‘cross section’, or rate at which top-quark pairs are produced for a given number of proton–proton collisions. Such measurements are conceptually simple: count the number of ‘top-like’ events seen in the detector; estimate and subtract off the number of events expected from other, non-top particle production processes giving rise to similar event signatures; and then correct for the estimated fraction of top-pair events that were missed.

In reality, years of painstaking work are required to understand the data and corrections well enough to make measurements with uncertainties of just a few percent. Calculating how many events we expect is also a formidable challenge, requiring knowledge of the exact mixture of quarks and gluons inside the proton (derived from many previous experiments), and the probabilities that quarks and gluons of particular energies will actually interact and produce a top–antitop pair. The result of this effort is shown in Figure 2: across the full range of proton–proton collision energies (denoted by √s) studied so far, top-antitop quark pairs are produced exactly as often as the theory says they should be, within uncertainties of a few percent. So far, this mysterious particle behaves just as expected for a very heavy partner of the b-quark.

Figure 2: ATLAS measurements of the top-pair production cross section at different LHC collision energies (√s) from 5 to 13 TeV, compared to the theoretical prediction (green band) , assuming a top-quark mass of 172.5 GeV. From the cross section at 13 TeV we can deduce that a total of about 130 million top-antitop pairs were produced in ATLAS during the years 2015–2018 of LHC operation at that collision energy. (Image: ATLAS Collaboration/CERN)

Weighing the top quark

The usual way to determine the mass of an unstable particle is to measure the energies and directions of the particles it decays into. As energy is conserved in the decay, the energy locked up in the mass of the particle reappears in the mass of the decay products. Any excess energy is converted into the motion of the decay products (i.e. their kinetic energy) as they fly apart from the initial decay point. This works well, for example, for the decay of a Z boson into two electrons or muons, whose energies and directions can be measured precisely in the ATLAS detector. The more Z-boson decays we record and analyse, the more precisely we can determine its mass.

However, the top quark decays into a W boson and a b-quark, neither of which is a stable particle. The b-quark produces a collimated jet of particles and, if the W boson decays into a quark–antiquark pair, two more such jets are produced. By measuring the energy and direction of these sprays of particles in the detector, the energy and direction of each of the original quarks can be inferred. By doing this for the quark and antiquark from the W boson, and for the original b-quark from the top decay, the mass of the top quark can be calculated.

But if the W boson decays to a lepton (e.g. an electron or muon) and a neutrino, things are a little more complicated. The lepton can be precisely measured, but the neutrino escapes the ATLAS detector without leaving any trace, so its energy and direction can only be guessed at from an ‘imbalance’ in the rest of the event. As there are two decaying top quarks in each event, it is also necessary to decide which particle jets and leptons belong to which top-quark decay – get this wrong, and a wrong value for the top mass, often far from the true value, is obtained.

The results of applying this technique to about 40,000 top-pair events recorded by ATLAS in 2012 are shown in Figure 3 (left). The peak of the distribution is around 160 GeV, but it is rather broad, with many events showing a top-quark mass below 150 GeV or above 170 GeV. This is mainly because the measurement of jets is an imprecise process – a quark with a true energy of 50 GeV may give rise to a jet that is incorrectly measured to have an energy of 40 or 60 GeV. But these effects average out over many events. The right plot in Figure 3 illustrates the reconstructed mass of the quarks assumed to come from the W-boson decay in each event; here the peak is close to 80 GeV and, since the W-boson mass is known from other measurements to be precisely 80.4 GeV, this provides an important ‘calibration’ measurement in the data analysis that helps reduce the uncertainty due to the jet measurement.

The final result from this analysis gives a top-quark mass value of 172.1 ± 0.9 GeV. This is about 12 GeV higher than the apparent peak position in Figure 3 (left), showing the importance of all the effects that distort the shape of the mass distribution, such as the imperfect jet resolution and the mis-assignment of jets to the wrong top-quark or W-boson decays. Physicists rely on theoretical models and computer simulations to predict these effects, and these models are limited by our lack of understanding of what actually happens when top quarks decay.

Figure 3: Reconstructed mass distributions of the jets assigned to a top-quark decay (left) and a W-boson decay (right) in top-pair events produced in proton–proton collisions recorded by the ATLAS detector at the LHC. (Image: ATLAS Collaboration/CERN)

Most quarks exist confined within hadrons, either in twos or threes. The force that keeps them bound together is the strong nuclear force, which has some similarities with the electromagnetic force that acts on charged particles (opposite electric charges attract, like charges repel). The analogous charge of the strong nuclear force is called ‘colour’, and comes in not one but three varieties: red, green and blue (although this has nothing to do with the everyday meaning of ‘colour’). Hadrons must be ‘colourless’, meaning that the colour charges of their constituent quarks must balance, e.g. red + anti-red, or red, green and blue.

When a high-energy quark (say a red one) transforms into a jet of colourless hadrons (with no overall colour), the original red must be ‘neutralised’ by another coloured quark or quarks elsewhere in the event. This ‘colour exchange’ process, which also implies exchange of momentum and energy, is not well understood theoretically and has implications for top-quark decays. Does the top quark exchange colour with other quarks in the event before it decays? Does it pass its colour to the b-quark produced in its decay, and where does that ‘excess’ colour go? After all this shuffling of colour, what happens to the reconstructed top-quark mass in the ATLAS detector? Estimates range from a negligible change, to shifts of about 1 GeV – larger than the precision with which we can now measure the mass experimentally.

One way to explore these questions is to measure the top-quark mass in other ways – ways that are less reliant on an analysis of the decay products, and easier to relate theoretically to a ‘clean’ definition of the quark’s mass. For example, the prediction of the top-pair production cross-section shown in Figure 2 depends on the top-quark mass – the heavier it is, the less frequently it is produced. Changing the top-quark mass assumed in the prediction by more than about 2 GeV would destroy the agreement between prediction and measurement.

The momentum distributions of the top quarks produced at the LHC, or the leptons in their decays, also depend on the mass. Several such indirect measurements exist, and they typically agree with the average of 172.7±0.5 GeV obtained from all the ATLAS measurements made so far using the top quark decay products. However, the uncertainties on the indirect measurements performed so far are at the level of 1 to 2 GeV, too large to clearly see possible effects of 1 GeV or less that might be present in the decay-product-based measurements.

Why does this matter? Since the discovery of the Higgs boson at the LHC in 2012 and subsequent precise measurements of its mass, powerful new ‘consistency tests’ of the theoretical underpinnings of particle physics (the Standard Model) became possible. In particular, the precise measurements of the top-quark and Higgs-boson masses can shed light on particle interactions at extremely high energy scales, far beyond what we can ever hope to probe with particle accelerators on Earth. If the top quark is too heavy relative to the Higgs boson, the Standard Model breaks down – it would even suggest that the entire Universe is in an unstable state, making it implausible that it has survived the 14 billion years since the Big Bang. Based on the current measurements of the top-quark mass, the Universe might be considered on a ‘knife edge’, balanced between stability and instability.

But since we are all still here, a heavy top quark probably implies that the Standard Model is incomplete – we have not yet understood all there is to learn about fundamental particles and their interactions. If the top quark is slightly lighter, perhaps the Standard Model is enough, though the question remains why we are so close to the edge. A more precise, and better understood, top-quark mass measurement could tell us whether we really are in this intriguing situation.

Tops with everything

Is the top quark special, or is it ‘just another quark’? Its enormous mass of around 173 GeV compared to the other quarks (all less than 5 GeV) also puts it above the masses of the electroweak W and Z bosons (80.4 GeV and 91.2 GeV, respectively), and the Higgs boson (125 GeV). If the top quark is just a quark, the Standard Model precisely predicts how it should interact with these bosons. If the top quark is something else – perhaps more closely connected to the electroweak or Higgs bosons – it may interact with them differently. The main signature of these interactions is the production of events with a top-pair and Higgs or Z boson, making for spectacular and complex signatures in the detector. The cross section of such events, as well as characteristics such as the momentum distribution of the produced bosons, depends on the details of the top–boson interaction (its ‘coupling’).

Using the large data sample collected during Run 2 of the LHC from 2015 to 2018, ATLAS has begun to make measurements of these processes. Since the W bosons from the top quarks can decay in a variety of ways (into quark–antiquark pairs, electrons, muons or taus), and the Z and Higgs (H) bosons can also decay in many different ways, there are a plethora of experimental signatures of such events. These typically feature two, three or four electrons or muons, and several jets from e.g. the b-quarks in the top decays. However, the production rate of top-pair + Z/H events is small, and other types of events can give rise to similar signatures. For example, top-pair + W events with the W boson decaying to an electron or muon, or events where some other particle is mistakenly identified as an electron or muon.

Figure 4: Measurements of the rate of top-pair + Higgs-boson production in events with two electrons or muons (leptons) with the same charge (2lSS), 3 or 4 leptons, or events with several leptons and one or two tau decays (𝜏had). The results are expressed as ratios μ of the measurement divided by the expected rate according to the Standard Model.(Image: ATLAS Collaboration/CERN)

Figure 4 shows one search for top-pair + Higgs production, analysing decays of Higgs bosons and top quarks to leptons. The results are expressed in terms of a ‘signal strength’ represented by μ, which is the ratio between the measured production rate and that expected in the Standard Model. A value of one indicates exact agreement between the measurement and the Standard Model prediction, whereas zero indicates no such events were seen. For this analysis, the final μ value averaged over all the explored decay signatures is μ=0.58 +0.36 -0.33. This is a ‘glass half full’ value – compatible with both top-pair + Higgs production at the Standard Model rate, or none at all. As a by-product, this analysis also measured the rate for top-pair + W production and found results about 40% higher than expected, consistent with earlier dedicated measurements on a smaller fraction of the ATLAS Run 2 dataset, and reflecting the difficulties in unambiguously separating events likely to come from top-pair + H and top-pair + W processes.

An alternative strategy is to exploit the rare but very striking decay of the Higgs boson to two photons (the massless boson of the electromagnetic interaction). In this analysis, events with two photons and additional particles consistent with the production of a top-pair were selected, and the mass distribution of the two photons was studied, as shown in Figure 5. The upper plot shows a bump where the mass of the two-photon system is close to 125 GeV, the mass of the Higgs boson. The contributions from other processes giving rise to the same final state should not have any such bump, and can be modelled with a smooth curve. After subtracting these background processes (lower plot), the contribution from top-pair + Higgs production is clear, albeit from just a few handfuls of events. The signal strength from this analysis is μ=0.92 + 0.27 -0.24, compatible with the Standard Model expectation of unity, and strongly suggesting the top quark and Higgs boson do indeed interact with about the expected strength.

Figure 5: Reconstructed mass distributions for the two-photon system in events which also have activity consistent with the production of top quarks. The upper plot shows the full distribution, with the red line showing a fit to the smoothly-varying background plus the bump from Higgs boson production. The lower-plot shows the distribution with the fitted background removed, leaving only the Higgs contribution, corresponding to the events with top-pair + Higgs boson production. (Image: ATLAS Collaboration/CERN)

The production of top-pair + Z events is somewhat easier to observe, thanks to the distinctive decay of the Z boson to two electrons or two muons, which can be efficiently reconstructed with very little possibility of confusion with other processes. With the full Run 2 dataset, ATLAS physicists have isolated enough top-pair + Z events to not only measure the rate, but to study the momentum distribution of the Z bosons, as shown in Figure 6. The largest fraction of events have Z-boson momentum values (p_T^Z) around 100 GeV, and both the overall rate of events and the shape of the p_T^Z distribution are well-reproduced by the theoretical predictions.

So, at least for top-pair + Z events, the Standard Model appears to be winning. But more data from Run 3 of the LHC, due to start in 2022, will be needed to complete the picture, especially for the top-pair + Higgs and top-pair + W boson processes.

As far as we can see, the top quark looks like a quark, it swims like a quark, and quacks like a quark. And yet – it does not fit the pattern.

While waiting for more data, physicists are also working to tease as much information as possible from the measurements that we have. The Standard Model provides one clear set of predictions as to how top quarks, and W, Z and Higgs bosons should behave at high energy – but there are lots of alternatives involving the effects from additional hypothesised particles or interactions. Rather than testing them all, an elegant mathematical approach is provided by Effective Field Theory. This technique considers all possible modifications to e.g. the couplings between the top quark and the bosons that are consistent with basic physical principles. The new couplings are made to vary in such a way that they have no effect at low energies (where the theory is equivalent to the Standard Model), but become important at higher energies, where deviations from the Standard Model predictions might be expected.

Figure 6: Measurement of the transverse momentum distribution of Z bosons produced in top-pair + Z events in the complete ATLAS Run 2 data sample at √s=13 TeV. The data measurements are shown by the black points, with uncertainties indicated by the error bars. The coloured lines indicate predictions from various theoretical calculations of the momentum distribution expected in the Standard Model. The lower plot shows the ratios of the predictions to the measured values, with the uncertainties indicated by the orange bands. (Image: ATLAS Collaboration/CERN)

The effects of each of these modifications can be calculated generically, without knowing the details of the new particles or interactions that might give rise to them, providing a library of deviations that can be compared to current measurements. Effective Field Theory also provides a way to treat data from different particle interactions (e.g. WW-scattering as discussed here) in a consistent way, allowing a 360° view of our current understanding of particle physics at the highest energies. Physicists are now applying these techniques to data involving top quarks, and the results should become even more interesting as the data become more precise in LHC Run 3.

The rarest process involving top quarks studied at the LHC so far is the simultaneous production of two top quarks and two top antiquarks, referred to as ‘four top production’. This is expected to be around 70,000 times rarer than top-pair production, yet ATLAS and CMS have recently seen first hints of such events using their full Run 2 dataset. The ATLAS analysis gives μ=2.0 +0.8 -0.6; this is about twice the expected rate, but still compatible with the Standard Model prediction within uncertainties. Many theories proposing new particles or interactions beyond the Standard Model could give rise to such an enhancement. This just maybe a first tantalising hint of excitement to come, but Run 3 data will be needed to know for sure.

Twenty-five years after the discovery of the top quark, we now know a great deal about this super-heavy fundamental particle. As far as we can see, it looks like a quark, it swims like a quark, and quacks like a quark. And yet – it does not fit the pattern. It is so much heavier than all the other quarks, and seems to sit more easily among the heavy bosons of the electroweak sector of the particle zoo. Is this a coincidence? Trying to understand patterns has been a powerful technique in our exploration of Nature thus far, and has brought us the periodic table, our understanding of hadrons, and much more. But when something does not fit the pattern, this is often a hint that something new is lurking around the corner. We have not found it yet – but we will keep looking!

About the author

Richard Hawkings did his PhD in Oxford, UK, and has worked as a CERN fellow and DESY fellow before becoming a staff research physicist at CERN. He has been a member of the OPAL collaboration at LEP and the ATLAS collaboration at LHC, where his work focuses on precision Standard Model physics, particularly involving top quarks. He has held various coordination roles on ATLAS including physics coordinator and top working group convener. He now co-leads the ATLAS luminosity measurement group.

Scientific papers

Measurement of the ttbar cross-section and lepton differential distributions in e𝜇 dilepton events from pp collisions at √s=13 TeV with the ATLAS detector (Eur. Phys. J. C 80 (2020) 528)
Measurement of the top quark mass in the tt→ lepton+jets channel from √s=8 TeV ATLAS data and combination with previous results (Eur. Phys. J. C79 (2019) 290)
Analysis of ttH and ttW production in multilepton final states with the ATLAS detector (ATLAS-CONF-2019-045)
Measurement of the properties of Higgs boson production at √s=13 TeV in the H→𝛾𝛾 channel using 139 fb^-1 of pp collision data with the ATLAS experiment (ATLAS-CONF-2020-026)
Measurement of the inclusive and differential production cross sections of a top-quark-antiquark pair in association with a Z boson at √s=13 TeV with the ATLAS detector (arXiv: 2103.12603, see figures)
Measurement of the tttt production cross-section in pp collisions at √s=13 TeV with the ATLAS detector (arXiv: 2106.11683, see figures)

Unraveling Nature's secrets: vector boson scattering at the LHC

Steven Goldfarb — Tue, 22 Sep 2020 08:52:00 +0000

Unraveling Nature's secrets: vector boson scattering at the LHC

Top HIghlight

False

Steven Goldfarb Tue, 22/09/2020 - 10:52

Highlight

False

Figure 1: "Wandering the immeasurable", a sculpture designed by Gayle Hermick welcomes the CERN visitors. From the Mesopotamians’ cuneiform script to the mathematical formalism behind the discovery of the Higgs boson, the sculpture narrates the story of how knowledge is passed through the generations and illustrates the aesthetic nature of the mathematics behind physics. (Image: J. Guillaume/CERN)

In 2017, the ATLAS and CMS Collaborations announced the detection of a process in high-energy proton–proton collisions that had not been observed before: the vector boson scattering. It results in the production of two W particles with the same electric charge as well as two collimated sprays of particles called “jets" (see Figure 2). The observation of vector boson scattering didn't receive as much attention from the media as the Higgs discovery in 2012, even though it was an important event for the particle physics community. Another missing piece of the big puzzle had been found – the puzzle that is the mathematical description of the microscopic world (see Figure 1).

The W⁺ and W^– bosons are unstable particles, which decay (transform) into a lepton and an antilepton or a quark and an antiquark with a mean lifetime of only a few 10^-25 seconds. They have integer spin (characteristic of bosons) and are carriers of the weak force. Though the weak force is not directly experienced in everyday life, it is nevertheless important as it is responsible for radioactive β decay, which plays a role in the fusion of hydrogen into helium that powers the Sun's thermonuclear process.

To appreciate the importance of this discovery, it is instructive to follow the history of how and why the W⁺ and W^– bosons were introduced; it illustrates nicely how the interplay between experimental information, theoretical models and mathematical principles drives progress in physics.

With the observation of vector boson scattering, another missing piece of the big puzzle had been found – the puzzle that is the mathematical description of the microscopic world.

Figure 2: Simplified view of a proton–proton collision event recorded with the ATLAS detector that was selected as a candidate for vector-boson-scattering production. The insert depicts a schematic view of the candidate physics process. Protons (p) from the LHC beam travel from left to right and from right to left in this view. They collide at approximately the centre of the detector. Within a very short period of time, too short to be resolved, two W bosons are emitted independently by the incoming quarks (q) from each of the LHC proton beams. These W bosons interact and each of the resulting W bosons decays to a muon (μ) and a neutrino (ν), where the neutrinos leave the ATLAS detector undetected. The outgoing quarks undergo a process called hadronisation and manifest as a spray of particles called a “jet”. (Image: ATLAS Collaboration/CERN)

Enrico Fermi originally formulated a mathematical description of the weak force in 1933 as a “contact interaction” between particles, occurring at a single point without a carrier particle propagating the force. This formulation successfully described the known experimental observations, including the radioactive β decay for which it was developed. However it was soon realised that its predictions at high energy, a regime not yet experimentally accessible at that time, were bound to fail.

Indeed, Fermi's theory predicts that the production rate of some processes caused by the weak force – such as the elastic scattering of neutrinos on electrons – increases linearly with the neutrino energy. This continuous growth, however, leads to the violation of a limit derived from the conservation of probability in a scattering process. In other words, predictions become unphysical at a high enough energy. To overcome this problem, physicists modified Fermi's theory by introducing “by hand" two massive spin-one (“vector”) charged particles propagating the interaction between neutrinos and electrons, dubbed “intermediate vector bosons". This development came well before the discovery of the W bosons decades later.

So even if the discovery of the long-awaited W^± bosons in 1983 – and, five months later, of a neutral companion, the Z boson – didn't come as a real surprise to physicists, it was certainly an epochal experimental achievement. Fermi’s theory remains an example of an effective theory valid only at low energy (well below the mass of the force carrier boson) – an approximation of a more general, universally valid theory.

Along this line, the search for a consistent description of the fundamental forces between the ultimate constituents of matter has led to the Standard Model of particle physics: a mathematical construction based on fundamental principles and experimental observations. The Standard Model provides a coherent, unified picture of three of the four fundamental interactions, namely the electromagnetic, weak and strong force. The fourth force, not included in the Standard Model, is gravity. So far, the Standard Model has been successful at describing a myriad of experimental measurements in the microscopic world. Its success is, by all means, mind blowing.

We do not know why natural phenomena are so well described by mathematical entities and relations but, experimentally, we know that it works. Just as Galileo said four hundred years ago, the big book of Nature is written in a mathematical language^[1] – the Standard Model and Einstein’s theory of gravity, for example, are additional chapters of this book.

So far, the Standard Model has been successful at describing a myriad of experimental measurements in the microscopic world. Its success is, by all means, mind blowing.

Particle physics makes use of a theoretical tool in which all particles are represented mathematically by quantum fields. These entities encode properties like spin and mass of a particle. In the Standard Model, the existence of the electromagnetic, weak and strong force carriers follows from the invariance of the behaviour of quantum fields under a “local gauge transformation". This is a transformation from one field configuration to another, which can be imagined as a rotation in an abstract mathematical space. The parameters of the transformation may vary from point to point in space-time, and thus the transformation is defined as “local”. Gauge invariance or gauge symmetry is the lack of changes in measurable quantities under gauge transformations, despite the quantum fields (which represent particles) being transformed.

Gauge invariance holds if spin-one particles are introduced, which interact in a well-defined manner with the elementary constituents of matter such as electrons and quarks (constituents of the proton and neutron, or, more generally, of hadrons). These spin-one particles are interpreted as “the carriers of the interaction” between the matter particles, with the photon the carrier for the electromagnetic force, the W^–, W⁺ and Z bosons for the weak force, and eight gluons for the strong force. These are the (intermediate) vector bosons introduced above.

In this way, the Standard Model forces (or interactions) emerge in a very elegant manner from one general principle, namely a fundamental symmetry of Nature. Interestingly enough, in the model, the electromagnetic and weak interactions manifest themselves at high energy as different aspects of a single “electroweak" force, while at low energy the weak interaction remains feebler than the electromagnetic interaction. As a consequence, the photon, Z boson and W^± bosons are collectively named “electroweak gauge bosons".

In the example above, Fermi’s followed a bottom-up approach: going from an observation to a mathematical description (a contact-interaction theory), which was modified “by hand" with few additions to obey the general principle of probability conservation (known as “unitarity” in physics). Starting from this premise, the work of many physicists consequently led to a more general theory. One in which the description of the fundamental forces follows the opposite path: predictions are obtained from fundamental principles (as gauge invariance) in a mathematically and physically coherent framework.^[2] The interplay between these two ways of developing knowledge had been common in physics since before Newton's time, and still valid today.

In both cases, a theory is successful not only when it describes the known experimental facts, but also when it has predictive power. The Standard Model possesses both virtues and examples of its predictive power include the discoveries of the Higgs boson and the neutral kind of weak interaction mediated by the Z boson.

As a matter of fact, the Standard Model tells us (much) more: the quantum fields representing the new spin-one particles will also transform under a local gauge transformation. To ensure that the measurable quantities describing their behaviour do not change (gauge invariance, mentioned above), interactions among the carriers of the weak force must also exist, as well as among the carriers of the strong force. These self-interactions may involve three or four gauge bosons. No self-interaction among photons is possible, except indirectly through virtual processes involving intermediate particles such as electrons, as observed in a dedicated ATLAS measurement.

The process first observed by the ATLAS and CMS Collaborations in 2017, characterised by the presence of two W bosons with the same electric charge and two jets, is a signature of the occurrence of an electroweak interaction. The dominant part of the process is due to the self-interaction among four weak gauge bosons; another central prediction of the Standard Model finally confirmed by the LHC experiments. This self-interaction manifests as a “vector boson scattering", where two incoming gauge bosons interact and produce two, potentially different, gauge bosons as final state particles. The production rate of this electroweak process is very low – lower than that of Higgs boson – which is why it was observed only recently. And just like the Higgs boson discovery, the observation of this process didn't come out of the blue.

At the Large Electron–Positron (LEP) collider, which operated at CERN between 1989 and 2000 in what is today the LHC tunnel, physicists had already observed the self-interaction among three gauge bosons. They measured the production of a pair of gauge bosons of opposite charge, a W⁺ and a W^– boson, in the collisions of beams of electrons and positrons, the antiparticle of the electron. According to the Standard Model, three main processes contribute to this production. They proceed via the exchange of either a photon, neutrino or Z boson between the electron and positron of the initial state and the W pair of the final state (Figure 3).

Figure 3: Diagrams representing three processes contributing to the e+e- → W+W- production. They illustrate the exchange between the initial e+e- and final W+W- of (from left to right), a neutrino (ν), photon (γ), and a Z boson, respectively. The symbol γ* is often used when a photon mediates the interaction. Following the rules of quantum mechanics, the production rate of a process is computed by the square of the sum of all possible diagrams contributing to it. The diagrams in the sum may have different relative signs, so they may cancel (destructive interference), in the same way that waves can cancel each other if they arrive out of synchronisation. In the case discussed here, each diagram is necessary to avoid an unphysical continuous increase of the production rate with the collision energy and to ensure the preservation of the gauge invariance of the theory. (Image: S. Pagan Griso, L. Di Ciaccio/ATLAS Collaboration)

The exchange of a photon or a Z boson occurs via the self-interaction of three weak gauge bosons: WWγ and WWZ, respectively. The main point here is that without considering all three processes, the calculated production rate would grow continuously with energy, leading to the already encountered unphysical behaviour. The observation of this process at LEP, with a production rate consistent with the Standard Model prediction, therefore confirmed the existence of a self-interaction among three bosons.

It is striking that the theory predicts the structure of each underlying process such that, even though each of them gives to the calculated production rate a contribution which at high energy becomes unphysical, violating unitarity, the unphysical behaviour cancels out when all of the processes are considered together.

By studying vector boson scattering, physicists can investigate the Higgs mechanism in the highest energy domain accessible, where there may be signs of new physics.

So far, so good – but there’s a catch. The W^± and Z bosons observed and identified by experiments as the carriers of the weak interaction are massive, yet gauge invariance is only preserved if the carriers are massless. Should physicists give up the principle of gauge invariance to reconcile the theory with experimental facts?

A solution to this puzzle was proposed in 1964, postulating the existence of a new spin-zero (“scalar”) field with a slightly more complex mathematical structure. While the basic laws of the forces remain exactly gauge symmetric, in the sense explained above, Nature has randomly chosen (among many possibilities) a particular lowest-energy state of this field, breaking with this choice the gauge symmetry in a limited way, called “spontaneous”.

The consequences are dramatic. Out of this new field, a new particle emerges – the scalar Higgs boson – and the W^± and Z bosons become massive. Physicists now believe that gauge symmetry was not always spontaneously broken. The universe transitioned from an “unbroken phase” with massless gauge bosons to our current “broken phase” during expansion and cool-down, a fraction of a second after the Big Bang.

The discovery of the Higgs boson in 2012 by the ATLAS and CMS Collaborations is a great success of the Standard Model theory, especially when considering that it was found to have the mass that indirect clues were pointing to. While the Higgs boson mass is not predicted by theory, the existence of the Higgs boson with a given mass leaves a delicate footprint in natural phenomena such that, if measured very precisely (as was done at LEP and at Tevatron, the smaller predecessor of the LHC at Fermilab, nearby Chicago, USA), physicists could derive constraints on its mass. The Higgs boson’s discovery was thus an experimental prowess as well as a consecration of the Standard Model. It emphasized the remarkable role of the precision measurements at LEP, even though the energy of that accelerator was not high enough to directly produce the Higgs boson.

Obviously, the story doesn't end here. Solid indications exist that the Standard Model is not complete and that it must be encompassed in a more general theory. This possibility is not surprising. As Fermi’s weak interaction theory exemplifies, history has shown that a theory’s validity is related to the energy range (or, equivalently, size of space) accessible by experiments.

More generally, classical mechanics is appropriate and predictive for the macroscopic world, when the speed of the objects is small with respect to the speed of light. To describe the microscopic world, however, quantum mechanics must be invoked, and the special theory of relativity must be applied to appropriately describe the behaviour of objects moving close to light speed.

The LHC is the perfect place to look for rare processes like vector boson scattering, as it collides protons with the highest energy and rate ever reached.

How can physicists find experimental signs that may help to formulate a more general theory than the Standard Model?

A valuable approach is to directly search collision events for particles not included in the Standard Model. However this is inherently limited: only particles with a mass at or below the collision energy can be directly produced, due to the fundamental principle of energy conservation and following the equivalence between mass and energy. Alternative avenues, which suffer less from this limitation but are indirect, include performing very precise measurements of fundamental parameters of the Standard Model or measuring rare processes to look for deviations with respect to theoretical predictions. Such measurements are able to explore a higher energy domain, as the LEP Higgs-boson example showed.

Vector boson scattering is one of these rare processes. It is special because closely related to the Higgs mechanism, and able to shed light on unexplored corners of Nature at the highest energy available in a laboratory. Similar to the LEP vector-boson study described above, vector boson scattering is expected to proceed via several processes, this time including the self-interaction of four gauge bosons as well as the exchange of a Higgs boson (see Figure 4). Without accounting for all of the processes, the calculated scattering rate grows indefinitely with energy, leading to the above-mentioned unphysical behaviour (violation of unitarity).

It could be argued that this question is already settled, since we know that the Higgs boson exists. The key issue is that the way in which the Higgs boson interacts with the gauge bosons in the Standard Model is exactly what is required to moderate the growth of the scattering rate at high energy; a minimal deviation of the Higgs mechanism from the Standard Model prediction could result in an apparent breakdown of unitarity.

Vector boson scattering would then occur at a rate different from what is predicted by the Standard Model, and unitarity would have to be recovered by a yet-unknown mechanism. The study of vector boson scattering thus allows physicists to investigate the Higgs mechanism in the highest energy domain accessible, where there may be signs of new physics.

Figure 4: Diagrams of some of the processes contributing to the W+W+ → W+W+ process. Analogous diagrams contribute to the W-W- → W-W- process. Similarly to the explanation given in Figure 3, in order to compute the production rate each contribution is first added before their sum is squared. The individual contributions may have different relative signs leading to cancellations. In this case each contribution is necessary to avoid an unphysical continuous increase of the production rate with (the square or fourth power of) the scattering energy. (Image: S. Pagan Griso, L. Di Ciaccio/ATLAS Collaboration)

The LHC is the perfect place to look for rare processes like vector boson scattering, as it collides protons with the highest energy and rate ever reached. Furthermore, the ATLAS and CMS experiments are designed to select and record these rare events.

As weak gauge bosons are extremely short-lived particles, experiments search for the scattering of vector bosons by looking for the production of two jets and two lepton–antilepton pairs in proton-proton collisions. Imagine this as two gauge bosons being emitted by the quarks from each of the incoming LHC proton beams. These gauge bosons subsequently scatter off each other and the bosons emerging from this interaction promptly decay (see Figure 2). The quarks are subsequently deflected and appear in the detector as jets of particles, typically emitted at a relatively small angle with respect to the beam direction. This is called an “electroweak” process as it is mediated by electroweak gauge bosons.

The experimental signature of vector boson scattering is therefore characterised by the presence of the decay particles of the two bosons, accompanied by two jets with large angular separation. The W and Z bosons predominantly decay into a quark and antiquark pair. Nevertheless, the search of these rare events preferentially exploits the decays into a lepton and an anti-lepton because a concurrent process, the multi-jet production, being mediated by the strong interaction has an overwhelming rate and obscures processes with a much smaller rate.

Still, the search for vector boson scattering is very challenging. This is not only because the rate of the process is low – accounting for only one in hundreds of trillions of proton–proton interactions – but also because, even making use of the leptonic decays, several “background” processes produce the same kinds of particles in the detector, mimicking the process’ signal.

Due to its high rate, a particularly challenging background process is one in which the jets accompanying the decay products of the gauge bosons arise as a result of the strong-force interaction. The impact of this background with respect to the signal depends on the kind of gauge bosons which scatter. When they are W bosons with the same electric charge, the production rate of the two processes (signal and background) is comparable.

For this reason, same-charge WW production is considered the golden channel for experimental measurements and was the first target for the ATLAS Collaboration to study vector-boson-scattering processes. ATLAS physicists reported for the first time strong hints of the process in a 2014 paper – a milestone in the LHC physics programme. However, it took three more years to arrive at an unambiguous observation, passing the five-sigma threshold that particle physicists use to define a discovery and corresponding to a probability of less than one in 3.5 million that a signal observation could be due to a mere upward statistical fluctuation of the number of background events. In the years between the first hint and discovery, the LHC was upgraded to increase its proton–proton collision energy – from 8 TeV to 13 TeV – as well as its collision rate – yielding about six times more collected data. These improvements made observation of vector boson scattering possible – the era of its study had at last begun.

The Standard Model only allows a specific set of combinations of four-gauge-boson self-interactions: WWWW, WWγγ, WWZγ and WWZZ, forbidding interactions among four neutral bosons.

However, not all electroweak bosons are equal. While the observation of two same-charge W bosons has allowed physicists to start testing the interaction of four W bosons (WWWW, Figure 2), the quest to test other self-interactions remained. The Standard Model only allows a specific set of combinations of four-gauge-boson self-interactions: WWWW, WWγγ, WWZγ and WWZZ, forbidding interactions among four neutral bosons.

Not all of these electroweak interactions are predicted to have the same strength and, because of this, probing them requires identifying processes that are less and less frequent. Similarly to the case of two same-charge W bosons, electroweak processes involving two jets and a WZ pair, a Zγ pair, or a ZZ pair are increasingly rare or have significantly larger backgrounds. Hunting for such processes among the billions of proton–proton collisions recorded by ATLAS requires physicists to look for subtle differences in order to distinguish a signal from very similar background processes occurring at much higher rates.

Table 1. List of processes presented in the text (first column) that are used to study vector boson scattering: WW with same charge, WZ, ZZ or Zɣ production in association with two jets (j), and photon-induced production of two W bosons. For each process a check indicates the four bosons involved in the self-interaction. Other measurements performed at the LHC also play a role to test these self-interactions but have been omitted in this table for simplicity. (Image: ATLAS Collaboration/CERN)

While such a task was commonly regarded as requiring a much larger amount of data than collected so far, the LHC experiments used artificial-intelligence algorithms to distinguish between the sought-after signal and the much larger background. Thanks to such innovations, in 2018 and 2019, ATLAS reported the observations of WZ and ZZ electroweak production, and saw a hint of the Zγ process. Suddenly, this brand-new field saw a surge in the number of processes that could be used to probe the self-interaction of gauge bosons.

The most recent addition is ATLAS’ observation of two W bosons produced by the interaction of two photons, each radiated by the LHC protons. This phenomenon occurs when the accelerated protons skim each other, producing extremely high electromagnetic fields, with photons mediating an electromagnetic interaction between them. Such an interaction is only possible when quantum mechanical effects of electromagnetism are taken into account.

This is a direct and clean probe of the γγWW gauge bosons interaction. A peculiarity of this process is that the protons participate as a whole and can remain intact after the interaction; this is very different from inelastic interactions where the quarks, the protons’ constituents, are the main actors (see Figure 2).

Table 1 summarises the processes that are used to study vector boson scattering at the LHC. It also shows the four bosons involved in the self-interaction. The study of each process provides a different test of the Standard Model, as modifications of the theory can differently alter the strength of the self-interactions.

An extensive upgrade of the LHC experiments is also ongoing, which will improve further the detection capabilities for the vector-boson-scattering processes.

Figure 5. Distribution of the photon energy in the search for events resulting from the electroweak production of two vector bosons (a Z and a γ) associated with two jets (Zγjj-EW). The black polymarkers represent the data, the full histograms with different colours represent the Standard Model predicted contributions for the signal (in brown) and the many background processes (in different colours). All expected contributions are stacked. The dotted blue line in the upper panel indicates the calculated signal distribution when a new term is added to the Standard Model theory. (Image: ATLAS Collaboration/CERN)

Now, ten years on from the first high-energy collisions took place in the LHC, the study of the vector boson scattering is a very active field – though still in its adolescence, both from the experimental and theoretical point of view. Experimentally, the size of the available signal sample is limited. The upcoming data-taking period (from 2022 to 2024) and the high-luminosity phase of the LHC (starting in 2027) will increase the amount of collected data by more than a factor two and by an additional factor of ten, respectively. An extensive upgrade of the LHC experiments is also ongoing, which will improve further the detection capabilities for the vector-boson-scattering processes.

In parallel, physicists will continue to improve their analysis methods, relying on more and more advanced artificial-intelligence algorithms to disentangle the rare signal processes from the abundant backgrounds. Physicists are also employing advanced calculation techniques to improve the precision of Standard Model predictions to match the increased measurement precision.

Furthermore, a bottom-up approach is being introduced which follows in the footsteps of Enrico Fermi. Physicists have developed a theoretical framework that allows new mathematical terms, respecting basic conservation rules and symmetries, to be added “by hand" to the Standard Model, without relying on a specific new physics model. These terms change the predictions in the high-energy regime where new physics could be expected (Figure 5). The simplest form of this approach is called Standard Model Effective Field Theory.

Even though we know that an effective theory cannot work at an arbitrary high energy scale, history has shown that, supplemented by measurements, it can provide useful guidance at lower energy. Different production-rate measurements – including those of the Higgs boson, boson self-interactions and the top quark – can be, separately or simultaneously, compared to predictions in the same effective theoretical framework.

It would be a sensation if more precise measurements indicated that such new terms are necessary to describe the data. It would be a sign of physics beyond the Standard Model and indication of the direction to take in order to develop a more complete theory, depending on which kinds of terms are needed. The interplay between experimental observations and models in the quest for a complete theory would continue.

Ultimately, all ongoing experimental collider and non-collider studies in particle physics will contribute to building knowledge – be they direct searches for new particles, precision measurements exploiting the power of quantum fluctuations or studies of rare processes. This experimental work is complemented by ever more precise theoretical calculations. In this task, the next generation of powerful particle accelerators now being planned are indispensable tools to find new phenomena that would help us understand the remaining mysteries of the microscopic world.

About the Authors

Lucia Di Ciaccio is a Professor of Physics at the University of Savoie Mont Blanc (France) and member of the ATLAS Collaboration. During her career she has worked on different topics, including lepton and hadron collider physics. Her present research activity deals with the search for signs of new physics phenomena in the multiple gauge boson sector. Simone Pagan Griso is a staff scientist at the Lawrence Berkeley National Laboratory and member of the ATLAS Collaboration. His research topics range from measurements of rare phenomena predicted by the Standard Model to direct searches of new particles that only exist in extension of the Standard Model theory, with emphasis on signatures that require the development of dedicated charged-particle reconstruction algorithms and innovative analysis techniques.

[1] Il Saggiatore (The Assayer) is a book published by Galileo Galilei in October 1623 and is considered to be one of the milestones of the scientific method, propagating the idea that Nature must described and understood using mathematical tools rather than scholastic philosophy, as generally was believed at the time.

[2] In some cases the postulated principles are inspired by experimental facts, like, for example, the measurement of the speed of light for the theory of special relativity.

Scientific results

Observation of electroweak production of a same-sign W boson pair in association with two jets in proton–proton collisions at 13 TeV with the ATLAS detector (Phys.Rev.Lett. 123 (2019) 16, 161801, arXiv: 1906.03203)
Observation of electroweak W^±Z boson pair production in association with two jets in proton–proton collisions at 13 TeV with the ATLAS detector (Phys. Lett. B 793 (2019) 469, arXiv:1812.09740)
Observation of electroweak production of two jets and a Z-boson pair with the ATLAS detector at the LHC (Submitted to NPHYS, arXiv: 2004.10612)
Evidence for electroweak production of two jets in association with a Zγ pair in proton–proton collisions at 13 TeV with the ATLAS detector (Phys. Lett. B 803 (2020) 135341, arXiv:1910.09503)
CMS Collaboration: Observation of Electroweak Production of Same-Sign W Boson Pairs in the Two Jet and Two Same-Sign Lepton Final State in Proton-Proton Collisions at 13 TeV (Phys. Rev. Lett. 120 (2018) 081801, arXiv: 1709.05822)

News articles

ATLAS observes W-boson pair production from light colliding with light, Physics Briefing, August 2020
ATLAS observes light scattering off light, Physics Briefing, March 2019
The Higgs boson: the hunt, the discovery, the study and some future perspectives, ATLAS Feature, July 2018
The rise of deep learning, CERN Courier, July 2018
ATLAS finds evidence for the rare electroweak W±W± production, Physics Briefing, September 2014
The LEP story, CERN press, October 2000

Searching for Dark Matter with the ATLAS detector

Steven Goldfarb — Tue, 05 Mar 2019 10:15:00 +0000

Searching for Dark Matter with the ATLAS detector

Top HIghlight

False

Steven Goldfarb Tue, 05/03/2019 - 11:15

Highlight

False

Figure 1: An event with a highly energetic jet of particles and no other significant visible energy (monojet) recorded in 2016 by the ATLAS detector. This is how invisible particles can be detected in ATLAS, as they recoil against the visible ones (in this case, the jet of particles). The direction of the invisible particle is indicated by the dashed line. (Image: ATLAS Collaboration/CERN)

When we look around us, at all the things we can touch and see – all of this is visible matter. And yet, this makes up less than 5% of the universe.

We now know that the vast majority of matter is dark. This dark matter does not emit or reflect light, nor have we yet observed any known particle interacting with it. It is through the gravitational effect of dark matter on other matter in space that astronomers inferred its existence.

The first evidence for the existence of dark matter came as early as the 1930s^[1]. Many astronomers had been observing the motion of galaxies, and found a discrepancy with respect to their expectation that only accounted for matter that was emitting light. This was corroborated in the 70s through observations of the rotational velocity of galaxies made by Vera Rubin and collaborators.

Figure 2: Percentage of ordinary matter, dark matter and dark energy in the universe, as measured by the Planck satellite. (Image: E. Ward/ATLAS Collaboration, Credit: ESA and the Planck Collaboration)

Because of gravitational lensing, an effect related to Einstein’s general theory of relativity, matter that stands between a light source and its observer can bend the light from the source so that the observed image is distorted. From comparing the known position of the source (e.g. obtained through direct emission of visible particles from the source) to its distorted image, one can reconstruct the distribution of the matter causing the distortion. Observations of gravitational lensing also pointed to additional matter with respect to what was visible.

More recently, supercomputer simulations of the structure of our universe show that only including visible matter will not reproduce the structures that are observed in the universe, while if dark matter is included then a closer agreement is obtained between observations and simulations.

The presence of dark matter and its amount in the universe can also be inferred from the variations of temperature in the early universe. This leftover amount of dark matter is called its “relic density”, and it amounts to about 27% of the matter-energy content of the universe.

However, none of the observations or simulations involving dark matter give a clear indication of what dark matter is made of. We only know that if dark matter is a particle^[2], then it must have mass, since it interacts with other matter through the force of gravity. We can hope to understand its nature by observing rare dark matter particles and their interactions from space (where we have already seen its effects), and by trying to produce them in controlled laboratory conditions.

We hope to further understand the nature of dark matter by producing it in controlled laboratory conditions.

How particle collisions can create dark matter in a lab

Experiments at particle accelerators have revealed much about the nature of visible (ordinary) matter, starting from the first prototypes that aided the discovery of the proton and the antiproton to the recent discovery of the Higgs boson. All of the particles observed so far are part of the Standard Model of Particle Physics, describing the fundamental components of matter and their non-gravitational interactions.

The most powerful accelerator ever built is the Large Hadron Collider (LHC) at CERN in Geneva, accelerating protons and colliding them with a total energy of 13 TeV. According to Einstein’s most famous equation, E=mc², the more energy (E) the more massive particles (with a mass m) one can create (13 TeV corresponds to roughly 14 thousand times the rest mass of a proton). The hope is that at the LHC we can create massive dark matter particles by colliding known particles, in the same way we create the Higgs boson in proton-proton collisions.

Particles are regularly accelerated to very high energies in the universe in "natural" particle accelerators, such as supernovae explosions, and then collide with other particles in our atmosphere. Cosmic rays, for example, are particles that are generated in outer space and make it to Earth. However, the advantage of laboratory particle accelerators such as the LHC is that there we know the initial conditions of the collisions – namely the type and energy of the particles being collided. We can also create a large (and known) number of collisions and observe them in a controlled environment. These are essential features for detecting dark matter particles at experiments like ATLAS.

Characteristics of dark matter and consequences for detector signatures

Since dark matter is dark, it will not interact significantly with instruments made of ordinary matter. For this reason, the underlying signature of dark matter production at the LHC, used by all ATLAS searches, is the presence of invisible particles in proton-proton collisions.

One might reasonably ask how invisible particles can be observed, since they are by definition undetectable! We solve this problem with a little ingenuity. Before each collision, the protons travel along the direction of the LHC beams, and not in directions perpendicular to the beams. This means that their momenta in these perpendicular directions – their "transverse momentum" – is zero. A fundamental principle of physics is that momentum is conserved and so, after the collision, the sum of the transverse momenta of the products of the collision should still be zero. Therefore, if we add up the transverse momenta of all the visible particles produced in the collision and find it not to be zero, then this could be because we have missed the momentum carried away by invisible particles.

Figure 3: Diagram showing how missing transverse momentum (ETmiss) is determined in the transverse cross-section of a LHC detector. The LHC beams are entering/exiting through the plane. (Image: C. Doglioni, L.T. Wang & E. Ward/ATLAS Collaboration)

This happens routinely in ATLAS, in the case of physics processes involving neutrinos. We refer to this missed transverse momentum as "ETmiss". LHC searches for dark matter look for collisions with large values of ETmiss, where the dark matter is produced in association with other, visible particles from the Standard Model, such as photons, quarks or gluons (forming "jets" of particles), or electrons, muons or tau leptons. While ETmiss can be difficult to measure because it relies on accurate measurements of all the other particles in the collision, it is a powerful tool for observing dark matter.

A further requirement for the identification of dark matter particles in collisions is that the invisible particles should not decay as they travel through the ATLAS detector. In order for an invisible particle to be a candidate for the "relic" dark matter produced in the Big Bang, it should have a lifetime of at least the age of the universe – of the order of 14 billion years. Particles created in LHC collisions take about 40 nanoseconds to cross the ATLAS detector, so requiring that their lifetime be longer than this is not enough, on its own, to prove they constitute the dark matter. Complementary information from astroparticle experiments searching for relic dark matter would be required. However, it is a very good start!

It is worth noting that other particles that are connected to dark matter might also be detected at the LHC, for example new short-lived particles that can decay both into dark matter and into known matter. Observing those would be an important complement to an observation of dark matter particles from space, as it would allow us to better understand the landscape of dark matter interactions.

Theoretical models of dark matter can tell us more about how the interaction of dark matter with ordinary matter may take place. From that, we can predict what to expect in our detectors if that model were realised in nature.

What could dark matter be? Theoretical hypotheses

Experimentally, there are very few indications of what dark matter might be. We can, however, make theoretical hypotheses on the nature of dark matter, which are useful to experimentalists. The theorist and experimentalist communities often collaborate, for example within the LHC Dark Matter Working Group^[3]. Theoretical models of dark matter can tell us more about how the interaction of dark matter with ordinary matter may take place. From that, we can predict what to expect in our detectors if that model were realised in nature. This is relevant for designing detectors sensitive to dark matter, and for deciding how to analyse the products of the collisions once they have been recorded. It is also useful to know what to look for, as we have to decide in real-time which collisions to save data from (this is done using the ATLAS trigger system). A solid theoretical framework for dark matter is also necessary to put LHC searches into context and to compare them with dark matter searches from other instruments.

Searches for dark matter at the LHC are commonly guided by theoretical models that would allow us to explain the relic density of dark matter with one or a few kinds of particles. A class of models that satisfies these requirements includes a dark matter particle that only interacts weakly with ordinary particles and has a mass within the energy range that can be probed at the LHC – a Weakly Interacting Massive Particle (WIMP).

Using WIMP models as our starting point for LHC searches doesn’t mean that we are bound to the idea that dark matter should be described with a single particle and a single interaction! This is especially important when you consider that the content of dark matter in the universe is five times the content of ordinary matter, and ordinary matter is described by a variety of different particles and interactions. At the LHC, we have begun our tour into possible theoretical models of dark matter^[4] hoping that the few most prominent components and interactions of dark matter will be detected first, just as the electron, proton and electromagnetic interaction were discovered before all other particles of the Standard Model.

Figure 4: Key particle discoveries from 1898 to today! (Image: E. Ward/ATLAS Collaboration)

The simplest models one can build in terms of particle content are those where the dark matter particle is added to the Standard Model. In these models, the interaction between visible and dark matter must proceed through existing particles, such as the Z or Higgs boson. This means that the Z or Higgs boson could decay into two dark matter particles^[5], in addition to their ordinary decay modes involving Standard Model particles.

These models are called “portal” models of dark matter, as known particles act as the portal between what we know (ordinary matter) and what we don’t know (dark matter). While models with a Z boson portal are fairly constrained by precision measurements, including those done at the LEP collider at CERN during the 1990s, now is the first time in the history of particles that we can study the properties of the Higgs boson in detail. We could discover whether one or more of those properties lead to a connection to dark matter.

In addition to dark matter, one can also conceive of another particle not included in the Standard Model that acts as a portal particle. These are called “mediator” particles, since they mediate a new interaction between ordinary matter and dark matter. In the simplest versions of these models, the mediator is an unstable heavy particle that is produced directly from the interaction of Standard Model particles, such as quarks at the LHC. Therefore, it must also be able to decay into those same particles, or into a pair of dark matter particles. If a model of this kind occurs in nature, we have a chance to directly discover this mediator particle at the LHC, as we would be able to detect its Standard Model decay products. Other simple models don’t have a mediator that can also decay to Standard Model particles, but instead foresee the production of dark matter particles in association with Standard Model particles that can aid the detection of the process over known backgrounds.

Figure 5: Measurement of the number of events in data for each range of missing transverse momentum, compared to the sum of different physics processes that produce this signature in the detector, from simulation. (Image: ATLAS Collaboration/CERN)

While these models are commonly used to interpret the results of many LHC searches in terms of dark matter, they are too simple to represent the full complexity of a dark matter theory. However, they are still useful as building blocks for more complete theories with more ingredients.

The most popular example of a more complete theory that includes a dark matter candidate is supersymmetry (SUSY). SUSY was one of the first dark matter models to be studied extensively at the LHC. An appealing feature of supersymmetry is that it also solves a stability problem of the relatively low mass of the Higgs boson and other electroweak particles of the Standard Model (around 100 GeV) compared to the Planck scale (10¹⁹ GeV), at which gravity is expected to become strong and the Standard Model must break down. Quantum field theories like the Standard Model naturally prevent such large differences in energy scale from developing, so a physical mechanism is required to generate them. SUSY models provide such a mechanism and, in many cases, predict the existence of a new stable, invisible particle - the lightest supersymmetric particle (LSP) - which has exactly the right properties to be a WIMP dark matter particle. The search for particles predicted by SUSY is a major focus of the ATLAS physics programme. If produced in LHC collisions, these particles could decay to produce a variety of Standard Model particles that can be observed in the ATLAS detector, together with two escaping LSP dark matter particles that generate the characteristic ETmiss signature discussed above.

Many other theories, of various degrees of completeness and complexity, contain dark matter particle candidates. Some of them predict new particles similar to the Higgs boson that can decay into dark matter, while others go beyond the WIMP paradigm and include mediators with extremely feeble interactions with known particles that only decay after traveling significant distances inside (or outside!) the detector, or more complex sectors of particles mirroring the Standard Model^[6]. It is important for LHC searches to cover all this ground, while also preparing for unexpected, not-yet-theorised discoveries. No stone must be left unturned!

Experimental techniques and results

ATLAS already measures many processes involving invisible particles, namely neutrinos from the Standard Model. Fig. 5 shows the results of the measurement of the number of Z bosons decaying into a pair of neutrinos (about one fifth of all Z boson decays). As shown in the diagram in Fig. 6, we use a visible object (in this case a photon) to detect the presence of invisible particles and measure their missing transverse energy, as explained in the previous section.

Figure 6: Diagram of a Z boson decaying into a neutrino-antineutrino pair where the Z boson is produced in association with a photon. (Image: ATLAS Collaboration/CERN)

Figure 7: Diagram of a new mediator particle decaying into a pair of dark matter particles, produced in association with a photon. (Image: ATLAS Collaboration/CERN)

A very similar technique can be used for detecting the presence of dark matter particles. If we take the process in Fig. 6, replace the neutrinos with dark matter particles, replace the Z boson with a generic mediator between ordinary matter and dark matter, then we have the diagram in Fig. 7.

The detector signature of the processes shown in Fig. 6 and Fig. 7 is identical (and is shown in the event display in Fig. 8). Since we cannot distinguish the processes on a collision-by-collision basis, we have to take a different approach. We start by collecting a large number of events that have a large amount of missing transverse momentum and a highly energetic object. Then, we estimate precisely the number of expected events from Standard Model processes (called “backgrounds”), and look for an excess of additional events that could be due to dark matter processes. This kind of search is called “ETmiss+X”, where X stands for what the dark matter recoils against^[7].

So far, we have not found any excess with respect to backgrounds in this kind of search,as shown in Fig. 12 where the data agrees with the Standard Model-only prediction. Still, the journey of ETmiss+X searches at the LHC is far from over. Adding data and improving the experimental precision of future searches will enable us to search for even weaker dark matter interactions yielding processes that are still rarer than those to which we are already sensitive.

Figure 8: A visualisation of a photon and ETmiss event recorded in 2016, is shown in the ATLAS detector. A photon with transverse momentum of 265 GeV (yellow bar) is balanced by a ETmiss of 268 GeV (red dashed line in the opposite side of the detector). (Image: ATLAS Collaboration/CERN)

The advantage of this kind of search is that it makes no specific assumption about the nature of the invisible particles, other than that they are produced in association with a Standard Model particle. It is therefore well-suited to cast a wide net on a variety of dark matter models, as long as the model’s signature includes invisible particles and includes dark matter–Standard Model interactions. Conversely, the very large Standard Model backgrounds in ETmiss+X searches can be reduced by giving up some of their generality, for example by requiring distinctive particles (e.g. top quarks, the Higgs boson or related particles) to be produced in association with the dark matter.

The mediator particle can also decay to visible particles, leading to a peak or "resonance" in the total mass of those particles. Searches for new particles using resonances in the total mass of visible particles have led to numerous discoveries at colliders, including, most recently, the Higgs boson at the LHC. Given that the LHC is the highest-energy laboratory particle collider, the most obvious goal is to search for extremely massive particles that could not have been produced before.

Still, dark matter mediators could also appear at lower masses, escaping detection because of very low couplings to protons. This is a region where it has been increasingly difficult to perform searches due to the overwhelming Standard Model backgrounds that exceed the experiment’s data capacity if recorded in their entirety. Since background events are indistinguishable from events coming from decays of dark matter mediators, there is a risk of discarding both. Being able to detect this kind of process has provided motivation for overcoming technical limitations. All the main LHC experiments now employ data-taking techniques that allow them to retain a smaller amount of information for some events, so that more events can be recorded^[8]. These searches have not yet yielded any new particles, but improvements to the data selection and data acquisition system may bring surprises for the next LHC run.

The results of searches for invisible and visible dark matter-mediator decays bring complementary information on different parameters of dark matter models. Together, they could help to characterise the nature of a discovery. We must keep in mind, though, that these searches are interpreted in terms of the processes shown in Fig. 7, which stem from a very simple theoretical model. In this model, the only two new particles are the dark matter and the mediator of the interaction, and that may not describe the full complexity of the unknown matter in the universe.

This is why ATLAS searches target many other experimental signatures in addition to MET-X and resonance searches. For example, models including putative new Higgs bosons yield an assortment of detector signals that can be targeted by different searches. These results can be compared to see whether there are regions in the model parameter space where we haven’t yet looked and, in some cases, they can be combined to strengthen the discovery potential or constraints on dark matter models. A comprehensive summary of these kinds of searches for dark matter, as well as their connection to astrophysical searches (described in the next section), can be found in a new ATLAS paper published today (arXiv: 1903.01400).

Many supersymmetry models predict the existence of a new stable, invisible particle – the lightest supersymmetric particle (LSP) – which has exactly the right properties to be a candidate dark matter particle.

Compared with ETmiss+X searches, detector signatures from SUSY scenarios offer the possibility to make use of some additional tricks to identify a dark matter signal from the Standard Model background.

In many models, SUSY particles are produced in pairs due to a requirement to conserve a quantity called “R-parity”^[9] (sometimes also denoted “matter-parity”). Whenever a SUSY particle decays, the resulting decay products must include exactly one lighter SUSY particle. The decay chain ends when the lightest SUSY particle, which is a candidate dark matter particle, is produced.

In contrast to many non-SUSY dark matter models, SUSY particle decays can generate many visible Standard Model particles of high energy. Hence, events containing SUSY particles can be identified by requiring these particles as well as missing transverse momentum. A further trick is to make use of constraints on the momenta of the visible particles produced in the SUSY decays coming from the high masses of their SUSY particle parents. In particular, when two visible particles are produced from two identical decay chains in a SUSY event, we can measure properties of the event which can take on much larger values than those expected in Standard Model background events. An example is shown in Fig. 10.

Figure 9: Missing transverse momentum distribution in data after selecting events with an energetic photon and ETmiss, compared to the Standard Model predictions. The different background processes are shown in different colours. The expected spectra of an example WIMP dark matter scenario is illustrated with red dashed lines. (Image: ATLAS Collaboration/CERN)

Figure 10: Distribution in data of a quantity sensitive to the production of pairs of SUSY particles whose decays include dark matter particles, after selecting events with two electrons or muons and ETmiss, compared to the Standard Model predictions. The different background processes are shown in different colours. The expected spectra of example SUSY dark matter scenarios are illustrated with blue and green dashed lines. (Image: ATLAS Collaboration/CERN)

With the help of these tools, SUSY searches are able to set tight requirements for events with a given set of characteristics, targeting specific models. This makes them less general than ETmiss+X searches, but also less impacted by large numbers of background events.

ATLAS has not yet found evidence of SUSY LSPs, and has strongly constrained many of the models that would simultaneously solve the dark matter puzzle and provide an explanation for the low mass of the Higgs boson. Nevertheless, many SUSY variants remain interesting and the search isn’t over, as described in the dedicated feature article.

Many other searches for particles from more complex dark matter theories, e.g. those in footnote 7, are also performed in ATLAS even though we don’t cover them in detail in this article. Some of the characteristics of these particles make them behave very differently compared with the particles the LHC was built to observe. Therefore, searching for these (still well-motivated) variants of dark matter is generally more challenging and requires dedicated techniques to identify and reconstruct candidate particles that would hint at the presence of dark matter. These searches are now at the forefront of the ATLAS and LHC quest for dark matter, and have gathered at least as much interest as searches for WIMPs and their associated particles.

Connecting collider searches to astrophysical searches

Searches for dark matter at the LHC are typically searches for the production, rather than the interaction or annihilation, of potential dark matter particles. As such, data from ATLAS would not provide proof that a new particle constitutes the dark matter – the sensitivity to dark matter lifetimes is just too short (see above). Nevertheless, ATLAS data could establish consistency with the predictions of dark matter models, and within those models ATLAS can provide complementary information to the broad range of astroparticle searches for the interaction of relic dark matter particles being carried out around the world. This complementarity can be illustrated taking, for example, the simple dark matter-mediator model.

Figure 11. Diagram showing dark matter (DM) interactions and their corresponding experimental detection techniques, with time going from left to right. (a) shows DM annihilation to Standard Model (SM) particles, as sought by Indirect Detection (ID) experiments. (b) shows DM -> SM particle scattering, targeted by Direct Detection (DD) experiments. (c) shows the production of DM particles from the annihilation of SM particles at colliders. (d) again shows the pair production of DM at colliders, but in this case the interaction occurs through a mediator particle between DM and SM particles. (Image: C. Doglioni & A. Boveia/ATLAS Collaboration)

Within this model, in order for dark matter particles to be produced in pairs at the LHC, two strongly interacting quarks or gluons from the colliding protons must interact to produce the two dark matter particles (Fig. 11(b)). These same interactions could enable relic dark matter particles trapped in the Milky Way galaxy to scatter off atomic nuclei on Earth, generating the nuclear recoil signature exploited by "direct" astroparticle searches for dark matter such as XENON in Europe, LUX in North America and PANDA-X in China. Constraints from ATLAS searches can therefore be translated, albeit with assumptions on the mediator–proton and mediator–dark matter interaction, into constraints on the possible signals in those experiments (Fig. 12).

Furthermore, the same interactions also enable relic dark matter particles produced in the early universe to annihilate and create Standard Model particles (Fig. 11(a)). This leads to the signatures for dark matter sought by "indirect" dark matter search experiments – typically high-energy photons (observed by telescopes such as HESS, MAGIC and VERITAS), neutrinos (observed by neutrino telescopes such as IceCube) or anti-particles (detected by space experiments such as AMS on the International Space Station). Results from collider searches can therefore also be compared with results from those experiments.

Figure 12: A comparison of the inferred limits from ATLAS data, including those from both ETmiss+X and mediator resonance searches, to the constraints from direct detection experiments on the WIMP-proton scattering cross section in the context of a model with a new vector particle mediating the Standard Model-dark matter interaction, fixing the given mediator / quarks (gq) and mediator / dark matter (gDM) couplings to the value in the plot. (Image: ATLAS Collaboration/CERN)

The complementarity between recent ATLAS searches and astroparticle searches for dark matter is illustrated by Fig. 12, for the case of the simple dark matter-mediator model.

When interpreting and combining ATLAS results and those from astroparticle dark matter searches, we need to consider whether the dark matter model being tested is consistent with the observed density of relic dark matter particles. This has been measured with a precision better than 1% through observations of the cosmic microwave background by satellites like Planck. When considering a particular dark matter model, this only sets an upper limit on the amount of dark matter the model should produce. This is because, in principle, the dark matter could consist of multiple types of particles, with any one type only contributing a fraction of the amount measured by Planck.

The relic dark matter density constraint is particularly important for SUSY dark matter models, where the models can often predict more dark matter than the Planck satellite observed. Special characteristics of the model, such as closely-spaced SUSY particle masses or increased dark matter interactions, can reduce this density to values consistent with Planck observations, and searches for models with these characteristics are a high priority for ATLAS.

The upcoming LHC data-taking period is expected to more than double the current dataset, and the high-luminosity period will deliver at least another factor of 10 more data. With this data, LHC experiments will be able to probe dark matter processes that are rarer and more challenging to reconstruct than the ones studied today.

Outlook: where do we go from here?

ATLAS is searching for dark matter at the LHC in synergy with other experimental collaborations, such as CMS and LHCb. LHC experiments have not yet discovered dark matter candidates from Run 1/2 data, but there is a large number of proton-proton collisions ahead. The upcoming LHC data-taking period (2021-2023, known as Run 3) is expected to more than double the current dataset, and the high-luminosity period beginning 2026 will deliver at least another factor of 10 more data. The experiments will be able to probe dark matter processes that are rarer and more challenging to reconstruct than the ones studied today. In view of the upcoming data-taking, experiments are also making use of more advanced data-collection and data-analysis techniques, such as machine learning^[10].
Direct and indirect searches for signals of the existing dark matter in our galactic neighbourhood are important complementary strategies to LHC searches, since astrophysical experiments are able to detect relic dark matter and they are necessary to confirm that a new invisible particle discovered at the LHC could make up dark matter. We will continue the dialogue with these experiments, exchanging scientific results and perspectives, share theoretical models, and extend the discussion to the broader astrophysics community.
Other experiments can probe dark matter models to which the LHC experiments are not sensitive, for example models where the interactions between dark matter and ordinary matter are too feeble for dark matter to be produced in collisions of known particles. These experiments are being discussed in the Physics Beyond Colliders effort that recently started at CERN.
As one of the main outstanding questions in fundamental physics, the identification of the nature of dark matter is a key scientific driver for the future of particle physics. For this reason dark matter searches are a main focus of the discussions, including both experimentalists and theorists, which have taken place in recent initiatives to draw up roadmaps for the future of the field. While the nature of dark matter is currently still unknown, it is clear that the quest to better understand it will be a highlight of humanity’s study of the fundamental constituents of the universe for many years to come.

About the authors

Caterina Doglioni is associate senior lecturer at Lund University, Sweden, and member of the ATLAS collaboration. She searches for new phenomena in ATLAS data. One of the targets of her research are possible mediators of dark matter interactions, discoverable using non-standard data recording techniques. Doglioni has been one of the LHC Dark Matter Working Group organisers. Dan Tovey is Professor of Physics at the University of Sheffield and a renowned expert of supersymmetry, a theory which predicts – among other phenomena – dark matter particles. Tovey coordinated the ATLAS physics programme between 2016 and 2017.

Research by C. D. is part of a project that has received funding from the European Research Council under the European Union’s Horizon 2020 research and innovation program (grant agreement 679305) and from the Swedish Research Council. Research by D. T. is part of a project that has received funding from the European Research Council under the European Union’s Horizon 2020 research and innovation program (grant agreement 694202) and from the UK Science and Technology Facilities Council (STFC).

[1] For an exhaustive overview of the history of dark matter, with ideas on dark matter that date even further back in time, see Bertone and Hooper's “A History of Dark Matter” (arXiv: 1605.04909), or Bertone, de Swart and van Dongen's “How dark matter came to matter” (arXiv: 1703.00013).

[2] This piece will not discuss the possibility that scientists haven’t understood all of the details of the structure of space-time, including how gravity acts. That hypothesis is discussed in more detail in this article and its references: "Shaking the dark matter paradigm" (Symmetry magazine, 2017).

[3] For this reason, the community of theorists and experimentalists looking for dark matter at the LHC has joined forces, forming first the Dark Matter Forum and then the Dark Matter Working Group. The goal and results of those group are described here.

[4] This article does not contain an exhaustive list of models. For a graduate-level lecture series on models of dark matter see, for example, the TASI "Lectures on Dark Matter Physics" by M. Lisanti (arXiv: 1603.03797).

[5] If the dark matter mass is less than half of that of the Z or the Higgs boson.

[6] For an introduction to these kind of models see, for example, "If You Can’t Find Dark Matter, Look First for a Dark Force" (Nautilus article, 2017), "Hunting for Dark Matter’s ‘Hidden Valley’" (BNL feature story, 2016), "Voyage into the dark sector" (Symmetry magazine, 2018) and "Long-lived physics" (CERN article, 2018).

[7] For more information on the missing transverse momentum+jet search, see the 2017 ATLAS Physics Briefing "Chasing the Invisible".

[8] For more information on this kind of searches, see the 2018 ATLAS Physics Briefing "A new data-collection method for ATLAS aids in the hunt for new physics".

[9] R-parity ensures that in SUSY models protons, and hence all of the atoms in the universe, are unable to decay to other particles quickly by exchanging SUSY particles. In models without R-parity conservation, this can also be prevented. However, introducing R-parity is the simplest possibility.

[10] For more information on ongoing efforts on Machine Learning, see the DarkMachines research collective. For general perspectives on data acquisition and collection see the HEP Software Foundation.

The Higgs boson: the hunt, the discovery, the study and some future perspectives

Steven Goldfarb — Wed, 04 Jul 2018 17:17:00 +0000

The Higgs boson: the hunt, the discovery, the study and some future perspectives

Top HIghlight

False

Steven Goldfarb Wed, 04/07/2018 - 19:17

Highlight

False

Figure 1: A candidate Higgs to ZZ to four-lepton event as seen in the ATLAS detector. The four reconstructed muons are visualised as red lines. The green and blue boxes show where the muons passed through the muon detectors. (Image: ATLAS Collaboration/CERN)

The origins of the Higgs boson

Many questions in particle physics are related to the existence of particle mass. The “Higgs mechanism,” which consists of the Higgs field and its corresponding Higgs boson, is said to give mass to elementary particles. By “mass” we mean the inertial mass, which resists when we try to accelerate an object, rather than the gravitational mass, which is sensitive to gravity. In Einstein’s celebrated formula E = mc², the “m” is the inertial mass of the particle. In a sense, this mass is the essential quantity, which defines that at this place there is a particle rather than nothing.

In the early 1960s, physicists had a powerful theory of electromagnetic interactions and a descriptive model of the weak nuclear interaction – the force that is at play in many radioactive decays and in the reactions which make the Sun shine. They had identified deep similarities between the structure of these two interactions, but a unified theory at the deeper level seemed to require that particles be massless even though real particles in nature have mass.

In 1964, theorists proposed a solution to this puzzle. Independent efforts by Robert Brout and François Englert in Brussels, Peter Higgs at the University of Edinburgh, and others lead to a concrete model known as the Brout-Englert-Higgs (BEH) mechanism. The peculiarity of this mechanism is that it can give mass to elementary particles while retaining the nice structure of their original interactions. Importantly, this structure ensures that the theory remains predictive at very high energy. Particles that carry the weak interaction would acquire masses through their interaction with the Higgs field, as would all matter particles. The photon, which carries the electromagnetic interaction, would remain massless.

In the history of the universe, particles interacted with the Higgs field just 10^-12 seconds after the Big Bang. Before this phase transition, all particles were massless and travelled at the speed of light. After the universe expanded and cooled, particles interacted with the Higgs field and this interaction gave them mass. The BEH mechanism implies that the values of the elementary particle masses are linked to how strongly each particle couples to the Higgs field. These values are not predicted by current theories. However, once the mass of a particle is measured, its interaction with the Higgs boson can be determined.

The BEH mechanism had several implications: first, that the weak interaction was mediated by heavy particles, namely the W and Z bosons, which were discovered at CERN in 1983. Second, the new field itself would materialize in another particle. The mass of this particle was unknown, but researchers knew it should be lower than 1 TeV – a value well beyond the then conceivable limits of accelerators. This particle was later called the Higgs boson and would become the most sought-after particle in all of particle physics.

The Higgs boson would become the most sought-after particle in all of particle physics.

The accelerator, the experiments and the Higgs

The Large Electron-Positron collider (LEP), which operated at CERN from 1989 to 2000, was the first accelerator to have significant reach into the potential mass range of the Higgs boson. Though LEP did not find the Higgs boson, it made significant headway in the search, determining that the mass should be larger than 114 GeV.

In 1984, a few physicists and engineers at CERN were exploring the possibility of installing a proton-proton accelerator with a very high collision energy of 10-20 TeV in the same tunnel as LEP. This accelerator would probe the full possible mass range for the Higgs, provided that the luminosity^{^[1]} was very high. However, this high luminosity would mean that each interesting collision would be accompanied by tens of background collisions. Given the state of detector technology of the time, this seemed a formidable challenge. CERN wisely launched a strong R&D programme, which enabled fast progress on the detectors. This seeded the early collaborations, which would later become ATLAS, CMS and the other LHC experiments.

On the theory side, the 1990s saw much progress: physicists studied the production of the Higgs boson in proton-proton collisions and all its different decay modes. As each of these decay modes depends strongly on the unknown Higgs boson mass, future detectors would need to measure all possible kinds of particles to cover the wide mass range. Each decay mode was studied using intensive simulations and the important Higgs decay modes were amongst the benchmarks used to design the detector.

Meanwhile, at the Fermi National Accelerator Laboratory (Fermilab) outside of Chicago, Illinois, the Tevatron collider was beginning to have some discovery potential for a Higgs boson with mass around 160 GeV. Tevatron, the scientific predecessor of the LHC, collided protons with antiprotons from 1986 to 2011.

In 2008, after a long and intense period of construction, the LHC and its detectors were ready for the first beams. On 10 September 2008, the first injection of beams into the LHC was a big event at CERN, with the international press and authorities invited. The machine worked beautifully and we had very high hopes. Alas, ten days later, a problem in the superconducting magnets significantly damaged the LHC. A full year was necessary for repairs and to install a better protection system. The incident revealed a weakness in the magnets, which limited the collision energy to 7 TeV.

When restarting, we faced a difficult decision: should we take another year to repair the weaknesses all around the ring, enabling operation at 13 TeV? Or should we immediately start and operate the LHC at 7 TeV, even though a factor of three fewer Higgs bosons would be produced? Detailed simulations showed that there was a chance of discovering the Higgs boson at the reduced energy, in particular in the range where the competition of the Tevatron was the most pressing, so we decided that starting immediately at 7 TeV was worth the chance.

The LHC restarted in 2010 at 7 TeV with a modest luminosity – a luminosity that would increase in 2011. The ATLAS Collaboration had made good use of the forced stop of 2009 to better understand the detector and prepare the analyses. In 2010, Higgs experts from experiments and theory created the LHC Higgs Cross-Section^{^[2]} Working Group (LHCHXSWG), which proved invaluable as a forum to accompany the best calculations and to discuss the difficult aspects about Higgs production and decay. These results have since been regularly documented in the “LHCHXSWG Yellow Reports,” famous in the community.

The discovery of the Higgs boson

Figure 2: The invariant mass from pairs of photons selected in the Higgs to γγ analysis, as shown at the seminar at CERN on 4 July 2012. The excess of events over the background prediction around 125 GeV is consistent with predictions for the Standard Model Higgs boson. (Image: ATLAS Collaboration/CERN)

As Higgs bosons are extremely rare, sophisticated analysis techniques are required to spot the signal events within the large backgrounds from other processes. After signal-like events have been identified, powerful statistical methods are used to quantify how significant the signal is. As statistical fluctuations in the background can also look like signals, stringent statistical requirements are made before a new signal is claimed to have been discovered. The significance is typically quoted as σ, or a number of standard deviations of the normal distribution. In particle physics, a significance of 3σ is referred to as evidence, while 5σ is referred to as an observation, corresponding to the probability of a statistical fluctuation from the background of less than 1 in a million.

Eager physicists analysed the data as soon as it arrived. In the summer of 2011, there was a small excess in the Higgs decay to two W bosons for a mass around 140 GeV. Things got more interesting as an excess at a similar mass was also seen in the diphoton channel. However, as the dataset increased, the size of this excess first increased and then decreased.

By the end of 2011, ATLAS had collected and analysed 5 fb^-1 of data at a centre-of-mass energy of 7 TeV. After combining all the channels, it was found that the Standard Model Higgs boson could be excluded for all masses except for a small window around 125 GeV, where an excess with a significance of around 3σ was observed, largely driven by the diphoton and four lepton decay channels. The results were shown at a special seminar at CERN on 13 December 2011. Although neither experiment had strong enough results to claim observation, what was particularly telling was the fact that both ATLAS and CMS had excesses at the same mass.

In 2012, the energy of the LHC was increased from 7 to 8 TeV, which increased the cross-sections for Higgs boson production. The data arrived quickly: by the summer of 2012, ATLAS had collected 5 fb^-1 at 8 TeV, doubling the dataset. As quickly as the data arrived it was analysed and, sure enough, the significance of that small bump around 125 GeV increased further. Rumours were flying around CERN when a joint seminar between ATLAS and CMS was announced for 4 July 2012. Seats at the seminar were so highly sought after that only the people who queued all night were able to get into the room. The presence of François Englert and Peter Higgs at the seminar increased the excitement even further.

At the famous seminar, the spokespeople of the ATLAS and CMS Collaborations showed their results consecutively, each finding an excess around 5σ at a mass of 125 GeV. To conclude the session, CERN Director-General Rolf Heuer declared, “I think we have it.”

The ATLAS Collaboration celebrated the discovery with champagne and by giving each member of the collaboration a t-shirt with the famous plots. Incidentally, only once they were printed was it discovered that there was a typo in the plot. No matter, these t-shirts would go on to become collector’s items.

ATLAS and CMS published the results in Physics Letters B a few weeks later. The ATLAS paper titled “Observation of a New Particle in the Search for the Standard Model Higgs Boson with the ATLAS Detector at the LHC.” The Nobel Prize in Physics was awarded to Peter Higgs and François Englert in 2013.

Figure 3: A recent distribution of candidate Higgs events from the H to ZZ to 4 leptons analysis using 13 TeV data from the LHC. The excess of events around 125 GeV is consistent with Standard Model predictions for the Higgs boson. (Image: ATLAS Collaboration/CERN)

Figure 4: The measured interaction strength as a function of the mass of different particles in the Standard Model. (Image: ATLAS and CMS Collaborations/CERN)

What we have learned since discovery

After discovery, we began to study the properties of the newly-discovered particle to understand if it was the Standard Model Higgs boson or something else. In fact, we initially called it a Higgs-like boson as we did not want to claim it was the Higgs boson until we were certain. The mass, the final unknown parameter in the Standard Model, was one of the first parameters measured and found to be approximately 125 GeV (roughly 130 times larger than the mass of the proton). It turned out that we were very lucky – with this mass, the largest number of decay modes are possible.

In the Standard Model, the Higgs boson is unique: it has zero spin, no electric charge and no strong force interaction. The spin and parity were measured through angular correlations between the particles it decayed to. Sure enough, these properties were found to be as predicted. At this point, we began to call it “the Higgs boson.” Of course, it still remains to be seen if it is the only Higgs boson or one of many, such as those predicted by supersymmetry.

The discovery of the Higgs boson relied on measurements of its decay to vector bosons. In the Standard Model, different couplings determine its interactions to fermions and bosons, so new physics might impact them differently. Therefore, it is important to measure both. The first direct probe of fermionic couplings was to tau particles, which was observed in the combination of ATLAS and CMS results performed at the end of Run 1. During Run 2, the increase in the centre-of-mass energy to 13 TeV and the larger dataset allowed further channels to be probed. Over the past year, the evidence has been obtained for the Higgs decay to bottom quarks and the production of the Higgs boson together with top quarks has been observed.^{^[3]} This means that the interaction of the Higgs boson to fermions has been clearly established.

Perhaps one of the neatest ways to summarise what we currently know about the interaction of the Higgs boson with other Standard Model particles is to compare the interaction strength to the mass of each particle, as shown in Figure 4. This clearly shows that the interaction strength depends on the particle mass: the heavier the particle, the stronger its interaction with the Higgs field. This is one of the main predictions of the BEH mechanism in the Standard Model.

We don’t only do tests to verify that the properties of the Higgs boson agree with those predicted by the Standard Model – we specifically look for properties that would provide evidence for new physics. For example, constraining the rate that the Higgs boson decays to invisible or unobserved particles provides stringent limits on the existence of new particles with masses below that of the Higgs boson. We also look for decays to combinations of particles forbidden in the Standard Model. So far, none of these searches have found anything unexpected, but that doesn’t mean that we’re going to stop looking anytime soon!

Analysis of the large Run-2 dataset will not only be an opportunity to reach a new level of precision in previous measurements, but also to investigate new methods to probe Standard Model predictions and to test for the presence of new physics.

Outlook

2018 is the last year that ATLAS will take data as part of the LHC’s Run 2. During this run, 13 TeV proton-proton collisions have been producing approximately 30 times more Higgs bosons than those used in the 2012 Higgs boson discovery. As a result, more and more results have been obtained to study the Higgs boson in greater detail.

Over the next few years, analysis of the large Run-2 dataset will not only be an opportunity to reach a new level of precision in previous measurements, but also to investigate new methods to probe Standard Model predictions and to test for the presence of new physics in as model-independent a way as possible. This new level of precision will rely on obtaining a deeper level of understanding of the performance of the detector, as well as the simulations and algorithms used to identify particles passing through it. It also poses new challenges for theorists to keep up with the improving experimental precision.

In the longer term, another big step in performance will be brought by the High-Luminosity LHC (HL-LHC), planned to begin operation in 2024. The HL-LHC will increase the number of collisions by another factor of 10. Among other measurements, this will open the possibility to investigate a very peculiar property of the Higgs boson: that it couples to itself. Events produced via this coupling feature two Higgs bosons in the final state, but they are exceedingly rare. Thus, they can only be studied within a very large number of collisions and using sophisticated analysis techniques. To match the increased performance of the LHC, the ATLAS and CMS detectors will undergo comprehensive upgrades during the years before HL-LHC.

Looking more generally, the discovery of the Higgs boson with a mass of 125 GeV sets a new foundation for particle physics to build on. Many questions remain in the field, most of which have some relation to the Higgs sector. For example:

A popular theory beyond the Standard Model is “supersymmetry”, which presents attractive features for solving current issues, such as the nature of dark matter. The minimal version of supersymmetry predicts that the Higgs boson mass should be less than 120-130 GeV, depending on some other parameters. Is it a coincidence that the observed value sits exactly at this critical value, hence still marginally allowing for this supersymmetric model?
Several models have been recently proposed where the only link of Dark Matter with regular matter would be through the Higgs boson.
Stability of the universe: the value of 125 GeV is almost at the critical boundary between a stable universe and a meta-stable universe. A meta-stable system possesses another baseline state, into which it can decay anytime due to quantum tunnelling.^{^[4]} Is this also a coincidence?
The phase transition: the details of this transition may play a role in the process which led our universe to be entirely matter and not contain any anti-matter. Present calculations with the Standard Model Higgs boson alone are inconsistent with the observed matter-antimatter asymmetry. Is this a call for new physics or only incomplete calculations?
Are fermion masses all related to the Higgs boson field? If yes, why is there such a huge hierarchy between the fermion masses spanning from fractions of electron-volts for the mysterious neutrinos up to the very heavy top quark, with a mass on the order of hundreds of billions of electron-volts?

From what we’ve learned about it so far, the Higgs boson seems to play a very special role in nature… Can it show us the way to answer further questions?

About the authors

Heather Gray is an experimental physicist at the Lawrence Berkeley National Lab, USA. She is a member of the ATLAS experiment at CERN’s Large Hadron Collider to which she has contributed in many ways, including the measurement of Higgs boson interactions with quarks. Bruno Mansoulié is scientist at CEA-IRFU, Saclay, France. He has worked both as a theoretical and experimental physicist, and is a founding member of ATLAS where he performed, among others, combined Higgs boson analyses and led the Higgs working group. Both enjoy communicating particle physics to non-specialists.

[1] The luminosity is the machine parameter which drives the number of events per second for a given physical process. The higher the luminosity, the more events per second

[2] The cross-section is a measure of the probability for that process to occur during any proton-proton collision. Processes with larger cross-sections occur more often than processes with small cross-sections.

[3] Since the publication of this feature, ATLAS has observed the Higgs boson decaying into a pair of bottom (b) quarks with a significance of 5.4 standard deviations. Read the press statement on this result here.

[4] Fortunately, even if we are a little on the meta-stable side, the lifetime for decay is extremely long compared to the present age of the universe…

Broken symmetry: searches for supersymmetry at the LHC

Steven Goldfarb — Fri, 08 Dec 2017 11:21:00 +0000

Broken symmetry: searches for supersymmetry at the LHC

Top HIghlight

False

Steven Goldfarb Fri, 08/12/2017 - 12:21

Highlight

False

Feature

A commentary by ATLAS physicists Paul de Jong and George Redlinger on the history, progress and future of the search for supersymmetry.

George Redlinger

Paul de Jong

supersymmetry

Broken Symmetry, a sculpture designed by Fermilab’s first director, Robert R. Wilson, straddles one of the entrances to the lab. The sculpture appears perfectly symmetric when observed directly from below, but asymmetric otherwise. (Image: Reidar Hahn/Fermilab)

The close connection between symmetry and the laws of physics is one of the greatest scientific insights to develop in the twentieth century. Noether’s Theorem – named after the German mathematician Emmy Noether, who proved the theorem in 1915 – drew the first connection between the symmetries of space-time and the conservation laws of physics. It posed, for example, that the symmetry of the laws of physics under time translation result in the conservation of energy. In other words, they are the same today as they were a billion years ago. Symmetry under spatial translation, meanwhile, results in the conservation of momentum.

Yet another symmetry, known as “gauge symmetry”, is also observed. Consider a bird sitting on a live wire. Why does it not die? Because absolute voltages are meaningless: the laws of physics are invariant as to what we arbitrarily call “ground”. Only differences in voltages between objects are important, as they lead to a current between those objects. As long as our bird does not experience voltage differences, it is safe.

A similar concept is operational in the quantum world. In particle physics, gauge symmetry posits an invariance of the laws of physics under the “rotation” of quantum fields in an abstract mathematical space. The implications are staggering. Gauge symmetry gives rise to a unifying mathematical structure, known as the “Standard Model”, that describes all matter and (so far) three of the four fundamental forces in the Universe.^[1] As a result, nuclear forces – which power the Sun, and give rise to radioactive decay and the promise of unlimited energy through nuclear fusion – can be described in the same mathematical framework as electromagnetic forces – underlying radio communications and most chemistry. The Standard Model clearly stands out as one of the crowning intellectual achievements of the twentieth century.

And yet, there is an underlying paradox. Our world is visibly not symmetric; it is clumpy and irregular. Even at the subatomic scale, where one might imagine things to be more uniform, there are profound asymmetries. For example, the weak nuclear force acts only over subatomic distances, while the effects of electromagnetism can be felt at macroscopic scales.

Here we face another of the most profound insights into the structure of physical law: the notion of broken symmetry. Condensed matter physicist Phillip Anderson is usually credited with coming up with the first ideas on how an asymmetric world might arise from underlying symmetric laws of physics. In a beautiful example of cross-disciplinary physics, these ideas were then taken to the subatomic realm by Yoichiro Nambu in the early 1960’s. Three independent efforts by Robert Brout and Francois Englert in Brussels; Gerald Guralnik, Carl Hagen and Tom Kibble at Imperial College in London; and Peter Higgs at the University of Edinburgh, would lead to a concrete model known as the Brout-Englert-Higgs Mechanism.^[2] It posits a Higgs field that fills the entire Universe, with localised excitations of this field, namely the Higgs boson.

The different interactions between the Higgs field (that fills empty space), and the carriers of the weak force (W/Z bosons) and electromagnetic force (photons), break the symmetry inherent in the unified description of these two vastly different forces. The W/Z bosons acquire mass through their interaction with the Higgs field, making them short-range force-carriers, while the photon remains massless, allowing the electromagnetic force to have infinite range. The energy scale of this unified description, where weak and electromagnetic forces have similar strength, defines the so-called electroweak scale. This practical application of symmetry breaking was put forward as a “toy model", but was not expected to be found in Nature in this simple form.

A fundamental problem (known today as the “hierarchy problem”) was first noticed by Kenneth Wilson in the early 1970’s. The Higgs boson, which “gives mass” to all fundamental particles, gives itself mass. However, this mass turns out to be theoretically unstable and, in this model, is around a factor of 10¹⁶larger than the electroweak scale. This difference was widely considered “unnatural”: a problem asking for a solution.

The Standard Model clearly stands out as one of the crowning intellectual achievements of the twentieth century. And yet, there is an underlying paradox. Our world is visibly not symmetric; it is clumpy and irregular.

The 1960’s through the mid 1980’s were truly a golden age for particle physics, with remarkably fruitful interactions between experimental and theoretical advances that led to the firm establishment of gauge symmetry (and its breaking) as fundamental, unifying principles. Yet, some important issues remained unresolved, most notably the experimental discovery of the exact mechanism for breaking electroweak symmetry, and the theoretical hierarchy problem.

The 1970’s saw the arrival of a radical new symmetry principle: “supersymmetry”, or SUSY for short. This remarkable new idea was simultaneously and independently developed on both sides of the Iron Curtain. How it went from a mathematical curiosity (largely outside the mainstream) to dominating theoretical (and to a large extent, experimental) particle physics is a fascinating story… one much too long to go into here!

SUSY extends the symmetries of space and time into the quantum domain, connecting the purely quantum mechanical property of “spin” to classical macroscopic quantities like space and time. By imposing certain conditions of “locality”, you can obtain General Relativity, Einstein’s theory of gravity. Is this starting to sound like quantum gravity? Then you’ll understand some of the excitement!

SUSY puts bosons (force carriers) and fermions (matter) on a symmetric footing, unifying the description of forces with that of matter. In practical terms for experimentalists, SUSY doubles the number of fundamental particles, with a SUSY “partner” particle for each Standard Model particle.^[3] Under supersymmetry, the strengths of the fundamental forces become the same (unified) at short distances. Supersymmetry naturally gives the Higgs boson the property that allows it to break the (electroweak) symmetry between the weak and electromagnetic forces. Supersymmetry naturally gives rise to a particle that fits all the characteristics of the dark matter in the Universe. And, as if that were not enough, supersymmetry provides a natural solution to the hierarchy problem: the unstable terms in the calculation of the Higgs boson mass, arising from interactions of the Higgs boson with Standard Model particles, are canceled by similar terms, of opposite sign, arising from the SUSY partners.

Understanding the breaking of electroweak symmetry (from the Higgs boson, or some other mechanism?) and the search for a solution to the hierarchy problem (supersymmetry?) have arguably been two of the primary drivers of research in particle physics since the 1980’s. At CERN, the Super Proton Synchrotron (SPS) and Large Electron-Positron (LEP) colliders elucidated the nature of the weak interaction, but did not find the Higgs boson nor SUSY. In the United States, plans were made for a Superconducting Super Collider (SSC) to find the Higgs boson. The cancellation of this project in the mid 90’s was a tremendous shock to the “big science” community. A smaller machine at Fermilab in Chicago, the Tevatron, that operated between 1983 and 2011 was very successful. It discovered the top quark and a significant number of other phenomena, but found neither the Higgs boson nor SUSY.

With these machines in mind, the completion of the Large Hadron Collider (LHC) at CERN was greeted with high hopes in 2007.

SUSY extends the symmetries of space and time into the quantum domain, connecting the purely quantum mechanical property of “spin” to classical macroscopic quantities like space and time.

The LHC is an enormous facility. It is a circular accelerator, 27 kilometres in circumference, straddling the Swiss-French border on the outskirts of Geneva. Four experiments are situated around the LHC ring, two of which (ATLAS and CMS) are general-purpose facilities. They study the collisions of protons, accelerated by the LHC to 99. 999999% of the speed of light, which, owing to special relativity, increases their rest mass by a factor of about 7000. These experiments can be considered giant cameras, taking 40 million “photos” per second, and generating a few petabytes of data per year.

The LHC has two advantages over its predecessors: first, it collides protons with the highest energy ever achieved; second, it records a much larger number of collisions. Furthermore, the ATLAS and CMS detectors are truly state-of-the-art.

Data-taking began in earnest in 2009, with the volume of data growing quickly as the detectors worked above expectation. Initial searches for supersymmetric particles focused on particles that could be discovered with relatively little data: the SUSY partners of quarks and gluons, known as “squarks” and “gluinos”. Their production mechanisms benefit from the high collision-energy and, for these particles in particular, the LHC opened up a completely new domain. The production rate of squarks and gluinos, encoded in a quantity called “cross-section”, is well calculated, with only their unknown masses as free parameters. Squarks and gluinos are expected to be unstable and decay into the quarks and gluons of the Standard Model. But they are accompanied, in the most popular SUSY model, by a stable and very weakly interacting particle, called the “Lightest Supersymmetric Particle” or LSP.

Due to their very weak interactions, LSPs will typically leave no trace in ATLAS and CMS detectors, just as the well-known neutrinos do. However, their presence can be deduced with some detective work: they would leave an imbalance in the distribution of the energy of the visible particles in the detector, in apparent contradiction to the laws of conservation of momentum. Of course, these laws are respected by Nature, and allow us to play the opposite game: to reconstruct any LSPs that may occur in the event. The signature of heavy squarks and gluinos would then be a surplus of collision events with a high number of very energetic jets (arising from the quarks and gluons in the decay of the heavy squarks or gluinos) and a significant amount of missing energy, carried away by the LSPs.

Before the LHC started, the concept of SUSY as a solution to the hierarchy problem gave rise to certain expectations on the masses of the SUSY particles. After all, it was only if the masses of the Standard Model particles and their SUSY partners do not differ too much that SUSY would provide a solution. However, what “do not differ too much” means exactly is hotly debated. Expectations were that the carriers of the weak force (the W and Z bosons) and their SUSY partners would have roughly the same mass. Partners of the top quark were thought to be perhaps as much as five times heavier than these weak-force partners. And the partner of the gluon (the carrier of the strong nuclear force) could be ten times heavier.

The hoped-for surplus of events described above was not observed in the initial run of the LHC, nor has it yet been seen in LHC Run 2 with increased collision energy. Given this absence, constraints have been set on the production cross-section of squarks and gluinos, and from these, limits on the squark and gluino masses have been derived. These constraints are now very tight: it is unlikely that gluinos exist with a mass below about 2000 GeV (more than 20 times the mass of the weak force carriers), for example.

In parallel to these first SUSY searches, ATLAS and CMS carried out the search for the Higgs boson in full force, with exciting results. The first hints of a signal appeared in late 2011; these were tentative, but having two independent experiments showing the same signs gave credence to the notion that discovery was at hand. Full confirmation came half a year later, leading the Director General of CERN to famously declare, “I think we have it” to a packed auditorium (and thousands watching online) which included Englert, Higgs, Hagen and Guralnik.

Since then, refined measurements seem to confirm with higher and higher accuracy the simple picture of the relatively light Higgs boson, something that was initially disparaged as a mere “toy model”. These results only enforced the uneasiness that particle physicists feel about the hierarchy problem. We came to realise, however, that not all squarks are equal in their importance to solve the hierarchy problem: the largest contributions to the unstable terms in the Higgs mass calculation arise from the top quark, and must be balanced by its SUSY partner, the scalar top quark or “stop”.

Example of an event selected in the analysis searching for gluinos. High-energy jets, such as those seen here in yellow, are accompanied by missing energy. However, also the Standard Model can give rise to such events, and the SUSY analysis must search for a surplus. (Image: ATLAS Collaboration/CERN)

As the decays of the stop differ in a number of respects from those of other squarks, generic searches for squarks and gluinos would not be very effective for stops. So with renewed vigour, dedicated searches for stop quarks were designed and applied to the data. These searches are complicated by the fact that the stop can decay in a number of different ways, and that the importance of each decay mode is strongly influenced by the stop mass, as well as by the unknown masses of other SUSY particles. Despite the large cross-section of a relatively light stop (with a mass close to the top quark mass), such particles are quite difficult to spot due to presence of very large Standard Model background signals.

The newly-designed dedicated searches also came out empty, allowing us to constrain the properties of the stop. If it exists, its mass should exceed approximately 1000 GeV, at least a factor of 10 above the masses of the weak force carriers.

But there’s a catch. The searches described above are most effective when the LSP is light, but not so much when the LSP is heavy. However, the LSP mass is also a free and unpredicted parameter. If the LSP is heavy, the amount of missing energy reconstructed in the detectors is rather small, and difficult to separate from the background. Recognising this, ATLAS and CMS designed dedicated searches for SUSY events accompanied by a very energetic jet or photon radiated by one of the protons just before collision. In such events, the SUSY particles will get a “kick” that increases their visible energy. Indeed, this improved the sensitivity of the searches involving heavy LSPs, but all limits on squarks – including stops – and gluinos are still rather weak when the LSP mass is high. It is important to realise this “fine print” whenever limits on SUSY are quoted.

The properties of the LSP also correspond very well with those predicted for dark matter particles. This is important motivation for the search for direct production of the class of SUSY particles of which the LSP is most likely a member: charginos and neutralinos, the supersymmetric partners of the gauge bosons of the weak force and of the Higgs boson. This is significantly more difficult than the search for squarks or gluinos. Charginos and neutralinos are produced with rather small cross-sections. Thus, any experimental search for them requires a lot of data. Such data has only recently become available, now that we have had a few years of LHC runs. Furthermore, the properties of charginos and neutralinos, including the production cross-sections and decay modes, are subject to much more uncertainty than those of squarks and gluinos. Searching for them is most effective in final states with low backgrounds, involving two or more charged leptons, and some missing energy.

Recent results have also shown no surplus of such events, setting lower limits on chargino and neutralino masses typically of a few hundred GeV. However, these searches cover far from all the options and are also limited by the size of the current LHC dataset. Therefore, they are still very much at the forefront of current SUSY research.

When Einstein was asked how he would react if his theory of General Relativity was not confirmed by experiment, he reportedly replied: “the Lord has missed a most marvellous opportunity”.

The beautiful theoretical ideas underlying supersymmetry have not been seen in Nature – at least, not in the simplest form we expected. When Einstein was asked how he would react if his theory of General Relativity was not confirmed by experiment, he reportedly replied: “the Lord has missed a most marvellous opportunity”. To many, supersymmetry is a similarly beautiful “opportunity” that still cannot be simply discarded. Supporters continue to hope that supersymmetry will eventually be discovered at the LHC, though perhaps in a different form than initially expected. It is a cruel irony that sometimes toy models turn out to be correct and beautiful ideas fall by the wayside. Ultimately, measurements of Nature decide what is fundamental.

There are still windows where supersymmetry (or some other solution to the hierarchy problem) might appear. These windows are narrowing, but if the experience of the Higgs boson is any guide, the last window is sometimes where things finally show up! In the history of particle physics, there are stories (perhaps apocryphal) of giving up too soon only to be scooped by a later experiment. Some think the window is in fact not that narrow; supersymmetry is also a broken symmetry, after all, and our belief in a narrowing window of discovery is based on certain prejudices about supersymmetry breaking.

However, it is clear where the search for supersymmetry will be headed next. Over the coming years, experiments will be in pursuit of SUSY particles with small cross-sections, such as charginos and neutralinos; SUSY particles that do not decay immediately after production, but instead have a finite lifetime leading to particular, and rather special, final states; final states with heavy LSPs; and final states without a LSP, with little or no missing energy.

Though the experimental search for supersymmetry has come up empty-handed so far, it has left a permanent mark on our field. Clever, sophisticated techniques and observables have been devised to optimally separate signal from background, and which have applications beyond the SUSY search. Furthermore, the interpretation of the results no longer takes place in a limited set of highly constrained models of supersymmetry, but rather in simplified models that are much more general and less theory driven. Combined with the publication of experimental information, such as efficiencies and backgrounds, this allows our published results to be reinterpreted in different models by theoretical physicists.

It is also important to continue studying the Higgs boson in ever more detail; it may provide a hint that leads to deeper understanding. It is a unique beast; there is no other particle like it. No field other than the Higgs field fills the entire Universe with a non-zero average value, and then gives the Universe a lower energy than if the Universe were empty. Some (mostly theorists) have punted on the hierarchy problem, claiming that it is “just so” that the Higgs boson has such an un-natural mass. This is not the first time that things turn out “just so”. In the 16^th century, elaborate models were developed to explain why the planets in our solar system have the spacing that they do. The recent solar eclipse in the United States revived the question of why the moon and Sun cover almost exactly the same area in the sky. We now understand that there is nothing special about these spacings – there are many planetary systems and ours is “just so”. Could it be that there are many universes and we happen to live in the one that is “just so”? This is an active area of research, but it strikes many as a dead-end because it seems impossible to verify experimentally.

Though the experimental search for supersymmetry has come up empty-handed so far, it has left a permanent mark on High-Energy Physics.

While the LHC continues to perform at unprecedented levels, both the ATLAS and CMS experiments are preparing to be upgraded with even more sophisticated technology. The promise of ever increasing datasets of higher and higher quality is a dream for any experimentalist. Yet there hangs over the community a certain shadow. The breaking of symmetries turns out to be central to realization of the laws of Nature as we know them. The breaking of electroweak symmetry has been confirmed by experiments to be described very well by the Standard Model. The expectation, however, of a somewhat (but not too strongly) broken supersymmetry, that could solve some of the unanswered questions left behind by the Standard Model, has not yet been confirmed.

There is still plenty of opportunity to find some of these answers in the upcoming data. But we may also need to come to terms with a different picture of Nature: one with the absence of such answers at the LHC scale. Perhaps it is our over-simplistic view that is broken, and not just SUSY. Perhaps Nature is telling us that we are looking at the problems of the Standard Model in the wrong way. History has shown that times such as these are precious, since they may well contain the seeds of true progress.

About the authors

Paul de Jong is professor in experimental particle physics at the University of Amsterdam and senior scientist at the Dutch National Institute for Subatomic Physics (Nikhef). George Redlinger is scientist at the Brookhaven National Laboratory, USA. Both are members of the ATLAS experiment at CERN’s Large Hadron Collider and have been mining ATLAS data for signs of supersymmetry for many years.

[1] The four forces are: gravity, electromagnetism and the strong and weak nuclear forces. The strong force holds atomic nuclei together while the weak force is responsible for radioactive decay. The latter three forces can be described in terms of a unified “gauge” theory, commonly known as the Standard Model. To date, only the gravitational force remains outside this unified framework. A quantum theory of gravity remains as the Holy Grail of fundamental physics.

[2] Some physicists have advocated using the (admittedly awkward) nomenclature, the ABEHGHK mechanism/field/boson, after the afore-mentioned seven physicists who played the most prominent role in developing this powerful idea. We follow common practice here and use the shortened form.

[3] This would not be the first time the cast of fundamental particles is doubled. The experimental discovery in 1922 of what was eventually to be called “spin” was perhaps the first doubling, with “spin up” and “spin down” particles. Spin arises naturally in relativistic quantum mechanics formulated in 1928 by Paul Dirac; the theory also predicted the existence of antiparticles. Experimental observation of the first antiparticle (the anti-electron, or positron) by Carl Anderson in 1932, the further development of quantum field theory and the observation of many other antiparticles confirmed this second doubling of the number of fundamental particles. The SUSY partners would carry the same quantum numbers as their known counterparts, except for their spin that differs by half a unit. Thus, the partners of matter particles have the spin of force carriers and vice versa.

Features

Searching for flavourful violations of the Standard Model

Nothing is truly forbidden in the quantum world. Even "forbidden" transitions are possible so long as they follow several intermediate, rule-following steps.

The Standard Model, its flavours and the search for new physics

The search for FCNC interactions in top-quark processes

If ATLAS were to observe these flavour-changing processes, it would be indisputable evidence of new physics phenomena.

The search for lepton flavour violation in Z-boson decays

The ATLAS Collaboration’s state-of-the-art searches for FCNC and LFV interactions have produced powerful probes of new physics phenomena.

Outlook

About the Authors

Further reading

Scientific articles

A transformative leap in physics: ATLAS results from LHC Run 2

The Run 2 data allowed ATLAS to observe and measure all major Higgs boson production modes at the LHC and to probe the non-universal Higgs couplings to fermions, thereby confirming the Brout-Englert-Higgs mechanism.

While the LHC was primarily built for the direct production and study of new particles and phenomena, progress in the field also relies on precision measurements that refine our understanding of Standard Model parameters and probe new physics through quantum loops.

ATLAS studies of heavy-ion collisions collected during Run 2 have deepened our understanding of the quark-gluon plasma through detailed measurements of its properties using soft and hard probes.

The non-observation of physics beyond the Standard Model so far, despite strong theoretical motivation for new TeV-scale particles, is another paramount outcome of the LHC.

Learning by machines, for machines: Artificial Intelligence in the world's largest particle detector

High-energy physics can trace its relationship with machine learning back many decades, with the earliest neural networks coming into play in the 1990s.

Machine learning at work in ATLAS

Machine-learning methods are designed to harness large amounts of data in order to infer new information, making them naturally suited to various data processing tasks in ATLAS.

Subtleties and challenges

While machine learning can offer incredible benefits to ATLAS throughout all stages of the analysis chain, its usage has to be closely coupled with continuous monitoring.

A new AI era of scientific research

About the Author

Further Reading

Scientific articles

A new ATLAS for the high-luminosity era

Rising to the challenge

To cope with the increased number of interactions when proton bunches collide at the HL-LHC, the ATLAS collaboration is working hard to upgrade its detectors with state-of-the-art instrumentation and technologies.

State-of-the-art tracking

Luminous future

New era

Further reading

Looking inside trillion degree matter with ATLAS at the LHC

Today’s Universe is dominated by quarks and gluons bound into ordinary matter. How would they behave if they were deconfined?

By producing the quark–gluon plasma in laboratories, researchers can study the conditions that prevailed in the early Universe, when matter was formed from free quarks and gluons.

How does the suppression depend on the type of jet?

How does the energy loss depend on the path length the jet travelled through the QGP?

Summary

About the authors

Further Reading

Scientific articles

The Last Quark

In the 1960s, physicists proposed that hadrons were actually made up of combinations of two or three fundamental constituents, which they called ‘quarks’.

Hunting the top quark

As each new and more powerful particle accelerator or collider began its work, physicists hoped it would have enough energy to discover the top quark. But no convincing hints were seen.

Top quarks come to Europe

The first observation of the top-quark at the LHC set the stage for the discovery of the even more elusive Higgs boson just two short years later.

Weighing the top quark

Tops with everything

As far as we can see, the top quark looks like a quark, it swims like a quark, and quacks like a quark. And yet – it does not fit the pattern.

About the author

Further reading

Scientific papers

Unraveling Nature's secrets: vector boson scattering at the LHC

With the observation of vector boson scattering, another missing piece of the big puzzle had been found – the puzzle that is the mathematical description of the microscopic world.

So far, the Standard Model has been successful at describing a myriad of experimental measurements in the microscopic world. Its success is, by all means, mind blowing.

By studying vector boson scattering, physicists can investigate the Higgs mechanism in the highest energy domain accessible, where there may be signs of new physics.

The LHC is the perfect place to look for rare processes like vector boson scattering, as it collides protons with the highest energy and rate ever reached.

The Standard Model only allows a specific set of combinations of four-gauge-boson self-interactions: WWWW, WWγγ, WWZγ and WWZZ, forbidding interactions among four neutral bosons.

An extensive upgrade of the LHC experiments is also ongoing, which will improve further the detection capabilities for the vector-boson-scattering processes.

About the Authors

Further reading

Scientific results

News articles

Searching for Dark Matter with the ATLAS detector

We hope to further understand the nature of dark matter by producing it in controlled laboratory conditions.

How particle collisions can create dark matter in a lab

Characteristics of dark matter and consequences for detector signatures

Theoretical models of dark matter can tell us more about how the interaction of dark matter with ordinary matter may take place. From that, we can predict what to expect in our detectors if that model were realised in nature.

What could dark matter be? Theoretical hypotheses

Experimental techniques and results

Many supersymmetry models predict the existence of a new stable, invisible particle – the lightest supersymmetric particle (LSP) – which has exactly the right properties to be a candidate dark matter particle.

Connecting collider searches to astrophysical searches

Outlook: where do we go from here?

About the authors

Further Reading

The Higgs boson: the hunt, the discovery, the study and some future perspectives

The origins of the Higgs boson