Publications

The companion electron-channel analysis on DarkSide-50's 12 ton-day argon exposure. Ionisation signals from sub-GeV/c² dark matter scattering on atomic electrons constrain four distinct scenarios in one paper: DM-electron elastic scattering, axion-like-particle absorption, dark-photon kinetic mixing — and, for the first time in direct detection, the mixing angle |U_e4|² for keV-scale sterile neutrinos.

The Migdal effect — ionization shaken loose when a nucleus recoils — gives a sub-GeV WIMP a detectable electronic signature even when the bare nuclear recoil is well below threshold. This analysis, on DarkSide-50's full 12 306 kg·d low-radioactivity argon exposure using the ionization signal alone, extends the search down to 40 MeV/c² and sets the world-leading spin-independent limit on WIMP-nucleon scattering for masses below 3.6 GeV/c² at the time of publication. Establishes Migdal as a routine sub-GeV channel for argon.

Concept design study for DarkSide-LowMass: a ionization-only, electron-counting argon TPC built on the DarkSide-50 legacy and the DarkSide-20k programme, optimised for sub-10 GeV/c² dark matter. The projections show GeV-scale reach down to the solar neutrino fog and — through the Migdal effect or DM-electron scattering — significant sensitivity down to 10 MeV/c². Sets out the threshold, background and spurious-electron control targets a purpose-built low-mass argon experiment would need to hit.

Dark matter need not couple equally to protons and neutrons; relaxing that symmetry can move direct-detection limits by orders of magnitude. Earlier studies of isospin-violating dark matter worked in the zero-momentum limit; we extend the treatment to the full non-relativistic effective-field-theory (NREFT) basis and apply it to XENON1T, DarkSide-50, DEAP-3600, LZ and DarkSide-20k. The headline surprise: under the Standard-Model-gauge-invariant formulation of the canonical spin-dependent interaction, argon — a zero-spin nucleus — becomes competitive with xenon, illustrating how strongly target complementarity matters for interpretation.

A community roadmap for quarkonium physics at the High-Luminosity LHC, covering pp, pA and AA collisions in both collider and fixed-target modes. Five threads run through it: associated production of J/ψ and Υ with other hard particles; low-pT χ_{c,b} and η_{c,b} spectroscopy; quarkonia as probes of gluon PDFs, nuclear PDFs and the full transverse-momentum-dependent / generalised parton distribution family; the gluon Sivers effect; and quarkonia as a thermometer for the quark–gluon plasma in heavy-ion collisions.

An auto-regressive Gaussian-mixture density model designed for problems where the observables have hard physical boundaries and depend conditionally on external (theory) parameters — two features that defeat most off-the-shelf generative models. Projecting onto a configurable latent space gives the network enough expressive power to (a) constrain physical parameters from data, (b) sample as a Monte-Carlo surrogate, and (c) estimate likelihood ratios for event classification, all from one trained model. Demonstrated on anomalous electroweak Z+jj production at the LHC.

A full Run 2 ATLAS search (139 fb⁻¹) in the b-jets + missing transverse momentum final state, simultaneously interpreted in three well-motivated BSM scenarios: pair-produced bottom squarks decaying to b + neutralino; pair-produced dark-matter particles via a generic scalar or pseudoscalar mediator radiated off b-quarks; and pair-produced third-generation down-type scalar leptoquarks. No excess seen — the headline limit excludes bottom squarks below 1270 GeV at 95% CL for a massless neutralino, with the dark-matter and leptoquark reinterpretations setting the strongest LHC bounds in this channel at publication.

The flagship report of the LHC Reinterpretation Forum, audited against Run 2. Covers four dimensions of making LHC results reusable against future theories: how direct searches publish their results, how unfolded measurements are made available, what technical implementations and Open Data exist, and which software frameworks (Rivet, Contur, MadAnalysis, CheckMATE, SModelS, …) consume them. Closes with concrete recommendations on what ATLAS, CMS and LHCb should publish per analysis — recommendations that have since shaped the experiments' release practice.

Builds high-capacity density models for high-energy physics by designing a custom latent space tailored to the structure of physical observables, improving on generic deep generative architectures for simulation and inference.

CP-violation in the Higgs sector is one of the most direct routes to a Standard-Model explanation of the baryon asymmetry of the universe. ATLAS has reported two CP-sensitive differential measurements of the signed azimuthal angle between forward jets in h+2-jet events, yielding a combined asymmetry of 0.3 ± 0.2 — too coarse alone to identify the underlying physics. This paper interprets that pattern in SMEFT, exposes the blind directions left by inclusive φ_jj alone, and shows that additional CP-sensitive observables combined with phase-space-dependent effects can resolve them. The result is a concrete roadmap for a global CP-aware Higgs coupling analysis reachable with O(100 fb⁻¹) of LHC data.

An ATLAS measurement of the electroweak production of a W boson with two forward, high-mass dijets — the VBS-like topology that probes the triple-gauge-boson vertex (WWZ/WWγ) directly. Combines 4.7 fb⁻¹ of 7 TeV data with 20.2 fb⁻¹ at 8 TeV, providing both fiducial and differential cross-sections of the EW component (separated from strong Wjj production) and using these to set limits on anomalous triple-gauge couplings.

The first measurement of electroweak Zjj production at 13 TeV — 3.2 fb⁻¹ of early Run 2 data. The EW component is extracted in a fiducial region designed to suppress the dominant QCD Drell-Yan Zjj background, which is itself constrained by a data-driven method. Reports σ^Zjj_EW = 119 ± 16(stat) ± 20(syst) fb above m_jj = 250 GeV and 34 ± 6 ± 6 fb above 1 TeV — both consistent with Standard Model expectations. A first-look 13 TeV benchmark for EW Zjj that later vector-boson- scattering analyses build on.

An early ATLAS Run 2 (3.2 fb⁻¹) measurement built around an experimentalist-friendly idea: report the ratio R_miss = σ(jets + MET) / σ(jets + ℓ⁺ℓ⁻), in which most systematic uncertainties cancel. The ratio is measured differentially in two phase-space regions — an inclusive single-jet region and a VBF-enhanced dijet region — and the unfolded data are designed to be reinterpreted against any future BSM hypothesis. The accompanying analysis constrains simplified DM models, EFT operators and invisible Higgs decays. The first ATLAS measurement of its kind, and a template for reusable, model-agnostic BSM searches at colliders.

Comprehensive double-differential cross-sections (in pT and rapidity) for prompt and non-prompt J/ψ and ψ(2S), using 2.1 fb⁻¹ of 7 TeV and 11.4 fb⁻¹ of 8 TeV ATLAS data via the dimuon channel. Reports the ψ(2S)-to-J/ψ production ratio and the non-prompt fraction across kinematic phase space, separating the QCD-driven prompt component from the b-hadron decay contribution. The benchmark reference for quarkonium production theory at LHC Run 1 energies.

First observation of associated Z + J/ψ production at >5σ in both the prompt and non-prompt channels, using ATLAS's full 20.3 fb⁻¹ of 8 TeV data. The cross-section ratios relative to inclusive Z are reported as σ(Z + prompt J/ψ)/σ(Z) ~ 4 × 10⁻⁶ and σ(Z + non-prompt J/ψ)/σ(Z) ~ 7 × 10⁻⁶, and the double-parton-scattering component is separated to give a lower bound on σ_eff > 5.3 mb at 68% CL — a direct experimental handle on multi-parton interactions in the proton.

First observation of W + prompt J/ψ associated production at a hadron collider (5.1σ, 27 ± 7 events) in 4.5 fb⁻¹ of ATLAS 7 TeV data. The cross-section is reported as a ratio to inclusive W production, and a double-parton-scattering contribution is explicitly extracted — providing a clean handle on the σ_eff governing multi-parton interactions in the proton and a sharp test of charmonium production in the presence of an electroweak boson.

Differential measurements of χ_c1 and χ_c2 production at the LHC in 4.5 fb⁻¹ of 7 TeV ATLAS data. The χ_c states are reconstructed via the radiative decay χ_c → J/ψ γ using photons from conversions in the inner detector — a technique that gives the resolution needed to separate the two closely-spaced states. Both prompt and non- prompt cross-sections are extracted, the σ(χ_c2)/σ(χ_c1) ratio is measured as a function of J/ψ pT, and the contribution from χ_c feed-down to prompt J/ψ is derived. Key inputs to NRQCD tests of charmonium production.

The ATLAS muon-reconstruction performance paper for Run 1 (2011-12, 7 and 8 TeV), calibrated on large J/ψ → μμ, Z → μμ and Υ → μμ reference samples. Reports muon-reconstruction efficiency above 99% with per-mille precision across |η| < 2.7 and 5 < pT < 100 GeV, and momentum resolution from 1.7% (central, pT ≈ 10 GeV) to 4% (large rapidity, pT ≈ 100 GeV). Provides the simulation corrections used across the entire Run 1 ATLAS physics programme, underpinning B-physics, quarkonium and a broad range of new-physics analyses.

Prompt and non-prompt ψ(2S) production at 7 TeV (2.1 fb⁻¹ ATLAS data), using the hadronic decay ψ(2S) → J/ψ(→μ⁺μ⁻)π⁺π⁻. Covers 10 < pT < 100 GeV and |y| < 2.0 — extending the prior dimuon-channel reach to higher transverse momentum and providing an independent cross-check of the colour-singlet, colour-octet and NRQCD descriptions of charmonium production.

Differential measurements of Υ(1S), Υ(2S) and Υ(3S) production at 7 TeV using 1.8 fb⁻¹ of ATLAS dimuon data, covering pT < 70 GeV and |y| < 2.25. Reports total cross-sections of 8.01, 2.05 and 0.92 nb respectively (assuming unpolarised production), with a quantified polarisation systematic that can shift integrated cross-sections by a factor of two under non-trivial azimuthal dependence. The headline finding: none of the contemporary quarkonium-production models reproduces the data across the full pT range — a clear signal that bottomonium production theory was not yet under control at LHC energies.

The legacy DØ Tevatron measurement of inclusive W(→eν) + n-jet (n=1-4) production on 3.7 fb⁻¹ of 1.96 TeV pp̄ data. Reports differential cross-sections across an unusually broad observable set — jet rapidities, lepton kinematics, H_T, leading-dijet pT and invariant mass, opening angles, azimuthal separations, and (the novel ingredient) rapidity separations between both pT-ordered and angular-ordered jet pairs — plus the mean jet multiplicity as a function of H_T. The angular-ordering observables exposed shortcomings in best-of-day predictions and informed parton-shower modelling that fed directly into later LHC W+jets analyses.

The first observation of a new particle at the Large Hadron Collider, made by ATLAS in 4.4 fb⁻¹ of 7 TeV data. A new structure at m = 10.530 ± 0.005(stat) ± 0.009(syst) GeV is seen in both the Υ(1S)γ and Υ(2S)γ radiative decay channels, identifying it as the χ_b(3P) bottomonium triplet. The result showcased the LHC's reach in heavy-quarkonium spectroscopy, was selected as a PRL Editors' Suggestion, and was confirmed shortly afterwards by DØ in re-analysed Tevatron data.

ATLAS's first B-physics publication. On just 2.3 pb⁻¹ of early 7 TeV data, the inclusive J/ψ → μ⁺μ⁻ cross-section is measured double-differentially in pT and rapidity over 1 < pT < 70 GeV and |y| < 2.4 — the widest reach of any J/ψ production measurement at the time. The prompt and non-prompt components are separated via the dimuon pseudo-proper time, and the results are compared against Colour Singlet NNLO*, the Colour Evaporation Model and FONLL. The work laid the groundwork for the LHC's quarkonium-production programme.

The DØ inclusive W(→eν) + n-jet measurement (n = 1-4) on 4.2 fb⁻¹ of 1.96 TeV pp̄ data, presenting cross-sections differentially in the nth-jet pT. Reaches considerably smaller uncertainties than previous W+jets measurements and is compared to NLO pQCD for the n = 1-3 bins and LO pQCD for n = 4. Generally agrees with the predictions but identifies phase-space regions where they fall short — informing later DØ Higgs searches and LHC W+jets analyses.

The first measurement of Υ(1S) production at the LHC — Υ(1S) → μ⁺μ⁻ with 1.13 pb⁻¹ of 7 TeV ATLAS data, presented differentially in pT and in two rapidity bands (|y|<1.2 and 1.2<|y|<2.4). The analysis is designed to minimise sensitivity to the poorly known polarisation by tight muon-pT requirements. The cross-section agrees with NRQCD predictions (colour singlet + colour octet) within a factor of two, but is off by up to an order of magnitude from the NLO colour-singlet-only prediction — a foundational bottomonium-production data point for the LHC programme.

The pre-data ATLAS detector + physics compendium — an 1852-page reference document that, before any 7 TeV collisions had been recorded, surveyed expected reconstruction performance (tracking, leptons, photons, MET, jets, b-tagging, trigger) and projected sensitivity across the planned physics programme. The chapters on low-pT muon reconstruction and quarkonium / heavy-flavour measurements (contributed to here) anchored the early ATLAS B-physics programme.

The detector paper for ATLAS — the canonical reference for the experiment's design, construction and performance, with contributions on electron bremsstrahlung-recovery algorithms.