Measurement of inclusive dijet cross-sections in proton-proton collisions at $\sqrt{s} = 13$ TeV with the ATLAS detector

TL;DR

This ATLAS study delivers a high-precision measurement of inclusive dijet cross sections at $\sqrt{s}=13~\text{TeV}$ using $140~\mathrm{fb}^{-1}$ of Run-2 data. Jets are reconstructed with the anti-$k_t$ algorithm, $R=0.4$, and cross sections are unfolded to particle level as functions of $(m_{jj}, y^{*})$ and $(m_{jj}, y_{boost})$, probing a broad kinematic range up to $m_{jj} \approx 9.5~\text{TeV}$. The results are compared to state-of-the-art NNLO full-color pQCD predictions with NP and EW corrections, across multiple PDF sets; while the data tend to lie below theory by about 15–20%, the $(m_{jj}, y_{boost})$ phase space shows relatively better agreement, and the ATLASpdf21T3 set minimizes the global chi-squared. These measurements provide stringent constraints on pQCD, NP effects, EW contributions, and PDFs in a regime sensitive to the running of the strong coupling and high-x parton dynamics. They thus motivate further refinements of PDFs and higher-order calculations to improve both normalization and shape in high-mass dijet events.

Abstract

Inclusive dijet cross-sections have been measured in proton-proton collisions at a centre-of-mass energy of 13 TeV using data with an integrated luminosity of 140 fb$^{-1}$, recorded by the ATLAS detector at the Large Hadron Collider during 2015-2018. Jets are identified using the anti-$k_{t}$ algorithm with a radius parameter of $R = 0.4$. The inclusive dijet double-differential cross-sections are measured first as a function of the invariant dijet mass and the half absolute rapidity separation between the two leading jets, $(m_{\mathrm{jj}}$, $y^{\ast})$, and second as a function of the invariant dijet mass and the total longitudinal boost of the dijet system, $(m_{\mathrm{jj}}$, $y_{\mathrm{boost}})$. The measured dijet system covers the invariant mass range from 240 GeV to almost 10 TeV, with dijet separation $y^{\ast} < 3.0$ and dijet boost $y_{\mathrm{boost}} < 3.0$. The results are unfolded to the particle level and compared with state-of-the-art next-to-next-to-leading-order full colour perturbative QCD calculations, corrected for non-perturbative and electroweak effects.

Figures (14)

• Figure 1: An illustration of the dijet event topology in the laboratory frame as a function of the $y^{\ast}\xspace{}=|y_1 - y_2|/2$ and $y_{\mathrm{boost}}\xspace{}=|y_1 + y_2|/2$ observables, with $y_1$ and $y_2$ being the rapidities of the leading and subleading jets. Projections of individual rows (columns) to the $y^{\ast}$ ($y_{\mathrm{boost}}$) axis correspond to the bin configurations of the measurement. In each subfigure, the horizontal axis corresponds to the beam direction, while the vertical axis represents the $x$--$y$ plane.
• Figure 2: Relative systematic uncertainty in the dijet cross-section as a function of $m_\mathrm{jj}$ in the first and last $y^{\ast}$ (a, b) and $y_{\mathrm{boost}}$ (c, d) analysis bins. The components of the uncertainty are displayed as colour bands---Jet energy scale, Jet energy resolution, and Other (combining the residual unfolding bias, the effect of disabled Tile calorimeter modules, and the luminosity uncertainty). The total systematic uncertainty is obtained as a quadrature sum of the components and is compared with the statistical component.
• Figure 3: Relative pQCD theoretical uncertainties for dijet cross-section predictions for the 18NNLO PDF set. Panels a, b (c, d) correspond to the first and last $y^{\ast}$ ($y_{\mathrm{boost}}$) analysis bins. The uncertainties due to the renormalisation and factorisation scales, the value of , the choice of particular PDF, the statistical uncertainty on the calculation and the total uncertainty are shown. The total uncertainty is calculated by adding the individual uncertainties in quadrature.
• Figure 4: Relative pQCD theoretical uncertainties for dijet cross-section predictions for the 4.0 PDF set. Panels a, b (c, d) correspond to the first and last $y^{\ast}$ ($y_{\mathrm{boost}}$) analysis bins. The uncertainties due to the renormalisation and factorisation scales, the value of , the choice of particular PDF, the statistical uncertainty on the calculation and the total uncertainty are shown. The total uncertainty is calculated by adding the individual uncertainties in quadrature.
• Figure 5: The non-perturbative correction factors in the first and last $y^{\ast}$ (a, b) and $y_{\mathrm{boost}}$ bins (c, d). The nominal non-perturbative correction factors are derived using [8.210], 2.3 PDF set, and A14 tuned parameters, while the envelope defines the associated uncertainty applied to the theory.