Fitted event-level b-tagging selection efficiency as a function of the dijet invariant mass. The uncertainties are evaluated by propagating the uncertainties on the fit parameters to the fitted efficiency.

Expected (dashed black line) and observed (solid red line) 95% CL exclusion limits on the higgsino simplified model being considered. These are shown with $\pm 1\sigma_{\mathrm{exp}}$ (yellow band) from experimental systematic and statistical uncertainties, and with $\pm 1\sigma_{\mathrm{theory}}^{\mathrm{SUSY}}$ (red dotted lines) from signal cross-section uncertainties, respectively. The limits set by the latest ATLAS searches using the soft lepton and disappearing track signatures are illustrated by the blue and green regions, respectively, while the limit imposed by the LEP experiments is shown in gray. The dot-dashed gray line indicates the predicted mass-splitting for the pure higgsino scenario.

Expected (dashed black line) and observed (solid red line) 95% CL exclusion limits on the higgsino simplified model being considered. These are shown with $\pm 1\sigma_{\mathrm{exp}}$ (yellow band) from experimental systematic and statistical uncertainties, and with $\pm 1\sigma_{\mathrm{theory}}^{\mathrm{SUSY}}$ (red dotted lines) from signal cross-section uncertainties, respectively. The limits set by the latest ATLAS searches using the soft lepton and disappearing track signatures are illustrated by the blue and green regions, respectively, while the limit imposed by the LEP experiments is shown in gray. The dot-dashed gray line indicates the predicted mass-splitting for the pure higgsino scenario.

Expected (dashed black line) and observed (solid red line) 95% CL exclusion limits on the higgsino simplified model being considered. These are shown with $\pm 1\sigma_{\mathrm{exp}}$ (yellow band) from experimental systematic and statistical uncertainties, and with $\pm 1\sigma_{\mathrm{theory}}^{\mathrm{SUSY}}$ (red dotted lines) from signal cross-section uncertainties, respectively. The limits set by the latest ATLAS searches using the soft lepton and disappearing track signatures are illustrated by the blue and green regions, respectively, while the limit imposed by the LEP experiments is shown in gray. The dot-dashed gray line indicates the predicted mass-splitting for the pure higgsino scenario.

Expected (dashed black line) and observed (solid red line) 95% CL exclusion limits on the higgsino simplified model being considered. These are shown with $\pm 1\sigma_{\mathrm{exp}}$ (yellow band) from experimental systematic and statistical uncertainties, and with $\pm 1\sigma_{\mathrm{theory}}^{\mathrm{SUSY}}$ (red dotted lines) from signal cross-section uncertainties, respectively. The limits set by the latest ATLAS searches using the soft lepton and disappearing track signatures are illustrated by the blue and green regions, respectively, while the limit imposed by the LEP experiments is shown in gray. The dot-dashed gray line indicates the predicted mass-splitting for the pure higgsino scenario.

Expected and observed CLs values per signal point represented by the grey numbers. The expected (dashed) and observed (solid) 95% CL exclusion limits are overlaid along with $\pm 1\sigma_{\mathrm{exp}}$ (yellow band) from experimental systematic and statistical uncertainties, and with $\pm 1\sigma_{\mathrm{theory}}^{\mathrm{SUSY}}$ (red dotted lines) from signal cross-section uncertainties, respectively.

Expected and observed CLs values per signal point represented by the grey numbers. The expected (dashed) and observed (solid) 95% CL exclusion limits are overlaid along with $\pm 1\sigma_{\mathrm{exp}}$ (yellow band) from experimental systematic and statistical uncertainties, and with $\pm 1\sigma_{\mathrm{theory}}^{\mathrm{SUSY}}$ (red dotted lines) from signal cross-section uncertainties, respectively.

Expected and observed cross-section upper-limit per signal point represented by the grey numbers. The expected (dashed) and observed (solid) 95% CL exclusion limits are overlaid along with $\pm 1\sigma_{\mathrm{exp}}$ (yellow band) from experimental systematic and statistical uncertainties, and with $\pm 1\sigma_{\mathrm{theory}}^{\mathrm{SUSY}}$ (red dotted lines) from signal cross-section uncertainties, respectively.

Expected and observed cross-section upper-limit per signal point represented by the grey numbers. The expected (dashed) and observed (solid) 95% CL exclusion limits are overlaid along with $\pm 1\sigma_{\mathrm{exp}}$ (yellow band) from experimental systematic and statistical uncertainties, and with $\pm 1\sigma_{\mathrm{theory}}^{\mathrm{SUSY}}$ (red dotted lines) from signal cross-section uncertainties, respectively.

Truth-level signal acceptances for each production process ($\tilde{\chi}_1^\pm \tilde{\chi}_1^0$, $\tilde{\chi}_1^\pm \tilde{\chi}_2^0$, $\tilde{\chi}_1^+ \tilde{\chi}_1^-$, and $\tilde{\chi}_2^0 \tilde{\chi}_1^0$) in a SR with the $S(d_0)$ requirement removed. The acceptance is defined as the fraction of accepted events divided by the total number of events in the generator-level signal Monte Carlo simulation, where the signal candidate track is identified as the charged particle with the largest distance between the interaction vertex and the secondary vertex of the higgsino decays.

Truth-level signal acceptances for each production process ($\tilde{\chi}_1^\pm \tilde{\chi}_1^0$, $\tilde{\chi}_1^\pm \tilde{\chi}_2^0$, $\tilde{\chi}_1^+ \tilde{\chi}_1^-$, and $\tilde{\chi}_2^0 \tilde{\chi}_1^0$) in a SR with the $S(d_0)$ requirement removed. The acceptance is defined as the fraction of accepted events divided by the total number of events in the generator-level signal Monte Carlo simulation, where the signal candidate track is identified as the charged particle with the largest distance between the interaction vertex and the secondary vertex of the higgsino decays.

Truth-level signal acceptances for each production process ($\tilde{\chi}_1^\pm \tilde{\chi}_1^0$, $\tilde{\chi}_1^\pm \tilde{\chi}_2^0$, $\tilde{\chi}_1^+ \tilde{\chi}_1^-$, and $\tilde{\chi}_2^0 \tilde{\chi}_1^0$) in a SR with the $S(d_0)$ requirement removed. The acceptance is defined as the fraction of accepted events divided by the total number of events in the generator-level signal Monte Carlo simulation, where the signal candidate track is identified as the charged particle with the largest distance between the interaction vertex and the secondary vertex of the higgsino decays.

Truth-level signal acceptances for each production process ($\tilde{\chi}_1^\pm \tilde{\chi}_1^0$, $\tilde{\chi}_1^\pm \tilde{\chi}_2^0$, $\tilde{\chi}_1^+ \tilde{\chi}_1^-$, and $\tilde{\chi}_2^0 \tilde{\chi}_1^0$) in a SR with the $S(d_0)$ requirement removed. The acceptance is defined as the fraction of accepted events divided by the total number of events in the generator-level signal Monte Carlo simulation, where the signal candidate track is identified as the charged particle with the largest distance between the interaction vertex and the secondary vertex of the higgsino decays.

Truth-level signal acceptances for each production process ($\tilde{\chi}_1^\pm \tilde{\chi}_1^0$, $\tilde{\chi}_1^\pm \tilde{\chi}_2^0$, $\tilde{\chi}_1^+ \tilde{\chi}_1^-$, and $\tilde{\chi}_2^0 \tilde{\chi}_1^0$) in a SR with the $S(d_0)$ requirement removed. The acceptance is defined as the fraction of accepted events divided by the total number of events in the generator-level signal Monte Carlo simulation, where the signal candidate track is identified as the charged particle with the largest distance between the interaction vertex and the secondary vertex of the higgsino decays.

Truth-level signal acceptances for each production process ($\tilde{\chi}_1^\pm \tilde{\chi}_1^0$, $\tilde{\chi}_1^\pm \tilde{\chi}_2^0$, $\tilde{\chi}_1^+ \tilde{\chi}_1^-$, and $\tilde{\chi}_2^0 \tilde{\chi}_1^0$) in a SR with the $S(d_0)$ requirement removed. The acceptance is defined as the fraction of accepted events divided by the total number of events in the generator-level signal Monte Carlo simulation, where the signal candidate track is identified as the charged particle with the largest distance between the interaction vertex and the secondary vertex of the higgsino decays.

Signal efficiencies in SR-Low for each production process ($\tilde{\chi}_1^\pm \tilde{\chi}_1^0$, $\tilde{\chi}_1^\pm \tilde{\chi}_2^0$, $\tilde{\chi}_1^+ \tilde{\chi}_1^-$, and $\tilde{\chi}_2^0 \tilde{\chi}_1^0$), defined by the number of events of reconstructed-level signal simulation divided by the number of events obtained at generator level, where the $S(d_0)$ selecton efficiency has the largest impact. The higgsino decay products from $\Delta \mathrm{m}(\tilde{\chi}_1^\pm,\tilde{\chi}_1^0) < 0.4$ GeV signal have $p_{\mathrm{T}}$ too low to be reconstructed as the signal candidate tracks, and therefore the identified signal candidate tracks are typically from pile-up collisions or underlying events similar to the QCD track background, causing a low $S(d_0)$ selection efficiency in these plots.

Signal efficiencies in SR-Low for each production process ($\tilde{\chi}_1^\pm \tilde{\chi}_1^0$, $\tilde{\chi}_1^\pm \tilde{\chi}_2^0$, $\tilde{\chi}_1^+ \tilde{\chi}_1^-$, and $\tilde{\chi}_2^0 \tilde{\chi}_1^0$), defined by the number of events of reconstructed-level signal simulation divided by the number of events obtained at generator level, where the $S(d_0)$ selecton efficiency has the largest impact. The higgsino decay products from $\Delta \mathrm{m}(\tilde{\chi}_1^\pm,\tilde{\chi}_1^0) < 0.4$ GeV signal have $p_{\mathrm{T}}$ too low to be reconstructed as the signal candidate tracks, and therefore the identified signal candidate tracks are typically from pile-up collisions or underlying events similar to the QCD track background, causing a low $S(d_0)$ selection efficiency in these plots.

Signal efficiencies in SR-Low for each production process ($\tilde{\chi}_1^\pm \tilde{\chi}_1^0$, $\tilde{\chi}_1^\pm \tilde{\chi}_2^0$, $\tilde{\chi}_1^+ \tilde{\chi}_1^-$, and $\tilde{\chi}_2^0 \tilde{\chi}_1^0$), defined by the number of events of reconstructed-level signal simulation divided by the number of events obtained at generator level, where the $S(d_0)$ selecton efficiency has the largest impact. The higgsino decay products from $\Delta \mathrm{m}(\tilde{\chi}_1^\pm,\tilde{\chi}_1^0) < 0.4$ GeV signal have $p_{\mathrm{T}}$ too low to be reconstructed as the signal candidate tracks, and therefore the identified signal candidate tracks are typically from pile-up collisions or underlying events similar to the QCD track background, causing a low $S(d_0)$ selection efficiency in these plots.

Signal efficiencies in SR-Low for each production process ($\tilde{\chi}_1^\pm \tilde{\chi}_1^0$, $\tilde{\chi}_1^\pm \tilde{\chi}_2^0$, $\tilde{\chi}_1^+ \tilde{\chi}_1^-$, and $\tilde{\chi}_2^0 \tilde{\chi}_1^0$), defined by the number of events of reconstructed-level signal simulation divided by the number of events obtained at generator level, where the $S(d_0)$ selecton efficiency has the largest impact. The higgsino decay products from $\Delta \mathrm{m}(\tilde{\chi}_1^\pm,\tilde{\chi}_1^0) < 0.4$ GeV signal have $p_{\mathrm{T}}$ too low to be reconstructed as the signal candidate tracks, and therefore the identified signal candidate tracks are typically from pile-up collisions or underlying events similar to the QCD track background, causing a low $S(d_0)$ selection efficiency in these plots.

Signal efficiencies in SR-Low for each production process ($\tilde{\chi}_1^\pm \tilde{\chi}_1^0$, $\tilde{\chi}_1^\pm \tilde{\chi}_2^0$, $\tilde{\chi}_1^+ \tilde{\chi}_1^-$, and $\tilde{\chi}_2^0 \tilde{\chi}_1^0$), defined by the number of events of reconstructed-level signal simulation divided by the number of events obtained at generator level, where the $S(d_0)$ selecton efficiency has the largest impact. The higgsino decay products from $\Delta \mathrm{m}(\tilde{\chi}_1^\pm,\tilde{\chi}_1^0) < 0.4$ GeV signal have $p_{\mathrm{T}}$ too low to be reconstructed as the signal candidate tracks, and therefore the identified signal candidate tracks are typically from pile-up collisions or underlying events similar to the QCD track background, causing a low $S(d_0)$ selection efficiency in these plots.

Signal efficiencies in SR-Low for each production process ($\tilde{\chi}_1^\pm \tilde{\chi}_1^0$, $\tilde{\chi}_1^\pm \tilde{\chi}_2^0$, $\tilde{\chi}_1^+ \tilde{\chi}_1^-$, and $\tilde{\chi}_2^0 \tilde{\chi}_1^0$), defined by the number of events of reconstructed-level signal simulation divided by the number of events obtained at generator level, where the $S(d_0)$ selecton efficiency has the largest impact. The higgsino decay products from $\Delta \mathrm{m}(\tilde{\chi}_1^\pm,\tilde{\chi}_1^0) < 0.4$ GeV signal have $p_{\mathrm{T}}$ too low to be reconstructed as the signal candidate tracks, and therefore the identified signal candidate tracks are typically from pile-up collisions or underlying events similar to the QCD track background, causing a low $S(d_0)$ selection efficiency in these plots.

Signal efficiencies in SR-High for each production process ($\tilde{\chi}_1^\pm \tilde{\chi}_1^0$, $\tilde{\chi}_1^\pm \tilde{\chi}_2^0$, $\tilde{\chi}_1^+ \tilde{\chi}_1^-$, and $\tilde{\chi}_2^0 \tilde{\chi}_1^0$), defined by the number of events of reconstructed-level signal simulation divided by the number of events obtained at generator level, where the $S(d_0)$ selecton efficiency has the largest impact. The higgsino decay products from $\Delta \mathrm{m}(\tilde{\chi}_1^\pm,\tilde{\chi}_1^0) < 0.4$ GeV signal have $p_{\mathrm{T}}$ too low to be reconstructed as the signal candidate tracks, and therefore the identified signal candidate tracks are typically from pile-up collisions or underlying events similar to the QCD track background, causing a low $S(d_0)$ selection efficiency in these plots.

Signal efficiencies in SR-High for each production process ($\tilde{\chi}_1^\pm \tilde{\chi}_1^0$, $\tilde{\chi}_1^\pm \tilde{\chi}_2^0$, $\tilde{\chi}_1^+ \tilde{\chi}_1^-$, and $\tilde{\chi}_2^0 \tilde{\chi}_1^0$), defined by the number of events of reconstructed-level signal simulation divided by the number of events obtained at generator level, where the $S(d_0)$ selecton efficiency has the largest impact. The higgsino decay products from $\Delta \mathrm{m}(\tilde{\chi}_1^\pm,\tilde{\chi}_1^0) < 0.4$ GeV signal have $p_{\mathrm{T}}$ too low to be reconstructed as the signal candidate tracks, and therefore the identified signal candidate tracks are typically from pile-up collisions or underlying events similar to the QCD track background, causing a low $S(d_0)$ selection efficiency in these plots.

Signal efficiencies in SR-High for each production process ($\tilde{\chi}_1^\pm \tilde{\chi}_1^0$, $\tilde{\chi}_1^\pm \tilde{\chi}_2^0$, $\tilde{\chi}_1^+ \tilde{\chi}_1^-$, and $\tilde{\chi}_2^0 \tilde{\chi}_1^0$), defined by the number of events of reconstructed-level signal simulation divided by the number of events obtained at generator level, where the $S(d_0)$ selecton efficiency has the largest impact. The higgsino decay products from $\Delta \mathrm{m}(\tilde{\chi}_1^\pm,\tilde{\chi}_1^0) < 0.4$ GeV signal have $p_{\mathrm{T}}$ too low to be reconstructed as the signal candidate tracks, and therefore the identified signal candidate tracks are typically from pile-up collisions or underlying events similar to the QCD track background, causing a low $S(d_0)$ selection efficiency in these plots.

Signal efficiencies in SR-High for each production process ($\tilde{\chi}_1^\pm \tilde{\chi}_1^0$, $\tilde{\chi}_1^\pm \tilde{\chi}_2^0$, $\tilde{\chi}_1^+ \tilde{\chi}_1^-$, and $\tilde{\chi}_2^0 \tilde{\chi}_1^0$), defined by the number of events of reconstructed-level signal simulation divided by the number of events obtained at generator level, where the $S(d_0)$ selecton efficiency has the largest impact. The higgsino decay products from $\Delta \mathrm{m}(\tilde{\chi}_1^\pm,\tilde{\chi}_1^0) < 0.4$ GeV signal have $p_{\mathrm{T}}$ too low to be reconstructed as the signal candidate tracks, and therefore the identified signal candidate tracks are typically from pile-up collisions or underlying events similar to the QCD track background, causing a low $S(d_0)$ selection efficiency in these plots.

Signal efficiencies in SR-High for each production process ($\tilde{\chi}_1^\pm \tilde{\chi}_1^0$, $\tilde{\chi}_1^\pm \tilde{\chi}_2^0$, $\tilde{\chi}_1^+ \tilde{\chi}_1^-$, and $\tilde{\chi}_2^0 \tilde{\chi}_1^0$), defined by the number of events of reconstructed-level signal simulation divided by the number of events obtained at generator level, where the $S(d_0)$ selecton efficiency has the largest impact. The higgsino decay products from $\Delta \mathrm{m}(\tilde{\chi}_1^\pm,\tilde{\chi}_1^0) < 0.4$ GeV signal have $p_{\mathrm{T}}$ too low to be reconstructed as the signal candidate tracks, and therefore the identified signal candidate tracks are typically from pile-up collisions or underlying events similar to the QCD track background, causing a low $S(d_0)$ selection efficiency in these plots.

Signal efficiencies in SR-High for each production process ($\tilde{\chi}_1^\pm \tilde{\chi}_1^0$, $\tilde{\chi}_1^\pm \tilde{\chi}_2^0$, $\tilde{\chi}_1^+ \tilde{\chi}_1^-$, and $\tilde{\chi}_2^0 \tilde{\chi}_1^0$), defined by the number of events of reconstructed-level signal simulation divided by the number of events obtained at generator level, where the $S(d_0)$ selecton efficiency has the largest impact. The higgsino decay products from $\Delta \mathrm{m}(\tilde{\chi}_1^\pm,\tilde{\chi}_1^0) < 0.4$ GeV signal have $p_{\mathrm{T}}$ too low to be reconstructed as the signal candidate tracks, and therefore the identified signal candidate tracks are typically from pile-up collisions or underlying events similar to the QCD track background, causing a low $S(d_0)$ selection efficiency in these plots.

Event selection cutflows for signal samples with $m(\tilde{\chi}_{1}^0)$ = 150 GeV and $\Delta m(\tilde{\chi}_{1}^\pm, \tilde{\chi}_{1}^0)$ = 1.5, 1.0, and 0.75 GeV, including all six production processes ($\tilde{\chi}_1^\pm \tilde{\chi}_1^0$, $\tilde{\chi}_1^\pm \tilde{\chi}_2^0$, $\tilde{\chi}_1^+ \tilde{\chi}_1^-$, and $\tilde{\chi}_2^0 \tilde{\chi}_1^0$). The cross-section used to obtain the initial number of events ($\sigma(\mathrm{n}_{\mathrm{jets}}) \geq 1$) refers to an emission of at least one gluon or quark with $p_{\mathrm{T}} > 50$ GeV at the parton level.

Event selection cutflows for signal samples with $m(\tilde{\chi}_{1}^0)$ = 150 GeV and $\Delta m(\tilde{\chi}_{1}^\pm, \tilde{\chi}_{1}^0)$ = 0.5, 0.35, and 0.25 GeV, including all six production processes ($\tilde{\chi}_1^\pm \tilde{\chi}_1^0$, $\tilde{\chi}_1^\pm \tilde{\chi}_2^0$, $\tilde{\chi}_1^+ \tilde{\chi}_1^-$, and $\tilde{\chi}_2^0 \tilde{\chi}_1^0$). The cross-section used to obtain the initial number of events ($\sigma(\mathrm{n}_{\mathrm{jets}}) \geq 1$) refers to an emission of at least one gluon or quark with $p_{\mathrm{T}} > 50$ GeV at the parton level.

The average tetrajet invariant mass distributions in data, along with the fitted background estimates for 0.16 < $\alpha$ < 0.18.

The average tetrajet invariant mass distributions in data, along with the fitted background estimates for 0.18 < $\alpha$ < 0.20.

The average tetrajet invariant mass distributions in data, along with the fitted background estimates for 0.20 < $\alpha$ < 0.22.

The average tetrajet invariant mass distributions in data, along with the fitted background estimates for 0.22 < $\alpha$ < 0.24.

The average tetrajet invariant mass distributions in data, along with the fitted background estimates for 0.24 < $\alpha$ < 0.26.

The average tetrajet invariant mass distributions in data, along with the fitted background estimates for 0.26 < $\alpha$ < 0.28.

The average tetrajet invariant mass distributions in data, along with the fitted background estimates for 0.28 < $\alpha$ < 0.30.

The average tetrajet invariant mass distributions in data, along with the fitted background estimates for 0.30 < $\alpha$ < 0.32.

The average tetrajet invariant mass distributions in data, along with the fitted background estimates for 0.32 < $\alpha$ < 0.34.

The average dijet invariant mass distributions in data, along with the fitted background estimates for 0.10 < $\alpha$ < 0.12.

The average dijet invariant mass distributions in data, along with the fitted background estimates for 0.12 < $\alpha$ < 0.14.

The average dijet invariant mass distributions in data, along with the fitted background estimates for 0.14 < $\alpha$ < 0.16.

The average dijet invariant mass distributions in data, along with the fitted background estimates for 0.16 < $\alpha$ < 0.18.

The average dijet invariant mass distributions in data, along with the fitted background estimates for 0.18 < $\alpha$ < 0.20.

The average dijet invariant mass distributions in data, along with the fitted background estimates for 0.20 < $\alpha$ < 0.22.

The average dijet invariant mass distributions in data, along with the fitted background estimates for 0.22 < $\alpha$ < 0.24.

The average dijet invariant mass distributions in data, along with the fitted background estimates for 0.24 < $\alpha$ < 0.26.

The average dijet invariant mass distributions in data, along with the fitted background estimates for 0.26 < $\alpha$ < 0.28.

The average dijet invariant mass distributions in data, along with the fitted background estimates for 0.28 < $\alpha$ < 0.30.

The average dijet invariant mass distributions in data, along with the fitted background estimates for 0.30 < $\alpha$ < 0.32.

The average dijet invariant mass distributions in data, along with the fitted background estimates for 0.32 < $\alpha$ < 0.34.

The expected and observed limits on $m_{4j}$ for signal templates using a 4-parameter fit function for 0.10 < $\alpha$ < 0.12.

The expected and observed limits on $m_{4j}$ for signal templates using a 4-parameter fit function for 0.12 < $\alpha$ < 0.14.

The expected and observed limits on $m_{4j}$ for signal templates using a 4-parameter fit function for 0.14 < $\alpha$ < 0.16.

The expected and observed limits on $m_{4j}$ for signal templates using a 4-parameter fit function for 0.16 < $\alpha$ < 0.18.

The expected and observed limits on $m_{4j}$ for signal templates using a 4-parameter fit function for 0.18 < $\alpha$ < 0.20.

The expected and observed limits on $m_{4j}$ for signal templates using a 4-parameter fit function for 0.20 < $\alpha$ < 0.22.

The expected and observed limits on $m_{4j}$ for signal templates using a 4-parameter fit function for 0.22 < $\alpha$ < 0.24.

The expected and observed limits on $m_{4j}$ for signal templates using a 4-parameter fit function for 0.24 < $\alpha$ < 0.26.

The expected and observed limits on $m_{4j}$ for signal templates using a 4-parameter fit function for 0.26 < $\alpha$ < 0.28.

The expected and observed limits on $m_{4j}$ for signal templates using a 4-parameter fit function for 0.28 < $\alpha$ < 0.30.

The expected and observed limits on $m_{4j}$ for signal templates using a 4-parameter fit function for 0.30 < $\alpha$ < 0.32.

The expected and observed limits on $m_{4j}$ for signal templates using a 4-parameter fit function for 0.32 < $\alpha$ < 0.34.

The expected and observed limits on $m_{\langle 2j \rangle}$ for signal templates using a 4-parameter fit function for 0.10 < $\alpha$ < 0.12.

The expected and observed limits on $m_{\langle 2j \rangle}$ for signal templates using a 4-parameter fit function for 0.12 < $\alpha$ < 0.14.

The expected and observed limits on $m_{\langle 2j \rangle}$ for signal templates using a 4-parameter fit function for 0.14 < $\alpha$ < 0.16.

The expected and observed limits on $m_{\langle 2j \rangle}$ for signal templates using a 4-parameter fit function for 0.16 < $\alpha$ < 0.18.

The expected and observed limits on $m_{\langle 2j \rangle}$ for signal templates using a 4-parameter fit function for 0.18 < $\alpha$ < 0.20.

The expected and observed limits on $m_{\langle 2j \rangle}$ for signal templates using a 4-parameter fit function for 0.20 < $\alpha$ < 0.22.

The expected and observed limits on $m_{\langle 2j \rangle}$ for signal templates using a 4-parameter fit function for 0.22 < $\alpha$ < 0.24.

The expected and observed limits on $m_{\langle 2j \rangle}$ for signal templates using a 4-parameter fit function for 0.24 < $\alpha$ < 0.26.

The expected and observed limits on $m_{\langle 2j \rangle}$ for signal templates using a 4-parameter fit function for 0.26 < $\alpha$ < 0.28.

The expected and observed limits on $m_{\langle 2j \rangle}$ for signal templates using a 4-parameter fit function for 0.28 < $\alpha$ < 0.30.

The expected and observed limits on $m_{\langle 2j \rangle}$ for signal templates using a 4-parameter fit function for 0.30 < $\alpha$ < 0.32.

The expected and observed limits on $m_{\langle 2j \rangle}$ for signal templates using a 4-parameter fit function for 0.32 < $\alpha$ < 0.34.

The expected and observed limits on $m_{4j}$ for Gaussian signal templates using a 4-parameter fit function for 0.10 < $\alpha$ < 0.12.

The expected and observed limits on $m_{4j}$ for Gaussian signal templates using a 4-parameter fit function for 0.12 < $\alpha$ < 0.14.

The expected and observed limits on $m_{4j}$ for Gaussian signal templates using a 4-parameter fit function for 0.14 < $\alpha$ < 0.16.

The expected and observed limits on $m_{4j}$ for Gaussian signal templates using a 4-parameter fit function for 0.16 < $\alpha$ < 0.18.

The expected and observed limits on $m_{4j}$ for Gaussian signal templates using a 4-parameter fit function for 0.18 < $\alpha$ < 0.20.

The expected and observed limits on $m_{4j}$ for Gaussian signal templates using a 4-parameter fit function for 0.20 < $\alpha$ < 0.22.

The expected and observed limits on $m_{4j}$ for Gaussian signal templates using a 4-parameter fit function for 0.22 < $\alpha$ < 0.24.

The expected and observed limits on $m_{4j}$ for Gaussian signal templates using a 4-parameter fit function for 0.24 < $\alpha$ < 0.26.

The expected and observed limits on $m_{4j}$ for Gaussian signal templates using a 4-parameter fit function for 0.26 < $\alpha$ < 0.28.

The expected and observed limits on $m_{4j}$ for Gaussian signal templates using a 4-parameter fit function for 0.28 < $\alpha$ < 0.30.

The expected and observed limits on $m_{4j}$ for Gaussian signal templates using a 4-parameter fit function for 0.30 < $\alpha$ < 0.32.

The expected and observed limits on $m_{4j}$ for Gaussian signal templates using a 4-parameter fit function for 0.32 < $\alpha$ < 0.34.

The expected and observed limits on $m_{\langle 2j \rangle}$ for Gaussian signal templates using a 4-parameter fit function for 0.10 < $\alpha$ < 0.12.

The expected and observed limits on $m_{\langle 2j \rangle}$ for Gaussian signal templates using a 4-parameter fit function for 0.12 < $\alpha$ < 0.14.

The expected and observed limits on $m_{\langle 2j \rangle}$ for Gaussian signal templates using a 4-parameter fit function for 0.14 < $\alpha$ < 0.16.

The expected and observed limits on $m_{\langle 2j \rangle}$ for Gaussian signal templates using a 4-parameter fit function for 0.16 < $\alpha$ < 0.18.

The expected and observed limits on $m_{\langle 2j \rangle}$ for Gaussian signal templates using a 4-parameter fit function for 0.18 < $\alpha$ < 0.20.

The expected and observed limits on $m_{\langle 2j \rangle}$ for Gaussian signal templates using a 4-parameter fit function for 0.20 < $\alpha$ < 0.22.

The expected and observed limits on $m_{\langle 2j \rangle}$ for Gaussian signal templates using a 4-parameter fit function for 0.22 < $\alpha$ < 0.24.

The expected and observed limits on $m_{\langle 2j \rangle}$ for Gaussian signal templates using a 4-parameter fit function for 0.24 < $\alpha$ < 0.26.

The expected and observed limits on $m_{\langle 2j \rangle}$ for Gaussian signal templates using a 4-parameter fit function for 0.26 < $\alpha$ < 0.28.

The expected and observed limits on $m_{\langle 2j \rangle}$ for Gaussian signal templates using a 4-parameter fit function for 0.28 < $\alpha$ < 0.30.

The expected and observed limits on $m_{\langle 2j \rangle}$ for Gaussian signal templates using a 4-parameter fit function for 0.30 < $\alpha$ < 0.32.

The expected and observed limits on $m_{\langle 2j \rangle}$ for Gaussian signal templates using a 4-parameter fit function for 0.32 < $\alpha$ < 0.34.

The product of the analysis acceptance and selection efficiency for all analysis selection criteria is shown as a function of $m_Y$ and $m_{X}/m_{Y}$.

The cross-section of the signal samples is shown as a function of $m_Y$ and $m_{X}/m_{Y}$.

The post-fit distributions of $p_\textrm{T}^{\textrm{bal}}$ for the SR. Data are compared with background predictions, and six signal predictions covering a representative mediator mass and invisible fraction range are overlaid. The uncertainties include all systematic and statistical components. The last bin contains the overflow.

The post-fit distributions of |$\phi_{\textrm{max}} - \phi_{\textrm{min}}$| for the SR. Data are compared with background predictions, and six signal predictions covering a representative mediator mass and invisible fraction range are overlaid. The uncertainties include all systematic and statistical components. The last bin contains the overflow.

Observed 95% CL limits on the t-channel production cross-section for semi-visible jets for R$_{inv}$ = 0.1.

Observed 95% CL limits on the t-channel production cross-section for semi-visible jets for R$_{inv}$ = 0.2.

Observed 95% CL limits on the t-channel production cross-section for semi-visible jets for R$_{inv}$ = 0.4.

Observed 95% CL limits on the t-channel production cross-section for semi-visible jets for R$_{inv}$ = 0.6.

Observed 95% CL limits on the t-channel production cross-section for semi-visible jets for R$_{inv}$ = 0.8.

Observed 95% CL limits on the t-channel production cross-section for semi-visible jets for R$_{inv}$ = 0.9.

The LO theory prediction (for λ=1).

Cutflow table for 6 signals of different mediator masses and R$_{inv}$ fractions.

$v_{2,2}^{unsub}$ and $v_{2,2}$ as a function of $\Delta\eta$ calculated from the 2-D per-trigger yields in figure 4(a) and 4(b), respectively.

$v_{3,3}^{unsub}$ and $v_{3,3}$ as a function of $\Delta\eta$ calculated from the 2-D per-trigger yields in figure 4(a) and 4(b), respectively.

$v_{4,4}^{unsub}$ and $v_{4,4}$ as a function of $\Delta\eta$ calculated from the 2-D per-trigger yields in figure 4(a) and 4(b), respectively.

The per-trigger yield distributions $Y^{corr}(\Delta\phi)$ and $Y^{recoil}(\Delta\phi)$ for events with $N_{ch}^{rec} \geq$ 220 in the long-range region $|\Delta\eta| >$ 2.

The per-trigger yield distributions $Y^{corr}(\Delta\phi)$ and $Y^{recoil}(\Delta\phi)$ for events with $N_{ch}^{rec} \geq$ 220 in the long-range region $|\Delta\eta| >$ 2.

The per-trigger yield distributions $Y^{corr}(\Delta\phi)$ and $Y^{recoil}(\Delta\phi)$ for events with $N_{ch}^{rec} \geq$ 220 in the long-range region $|\Delta\eta| >$ 2.

The per-trigger yield distributions $Y^{corr}(\Delta\phi)$ and $Y^{recoil}(\Delta\phi)$ for events with $N_{ch}^{rec} \geq$ 220 in the long-range region $|\Delta\eta| >$ 2.

The per-trigger yield distributions $Y^{corr}(\Delta\phi)$ and $Y^{recoil}(\Delta\phi)$ for events with $N_{ch}^{rec} \geq$ 220 in the long-range region $|\Delta\eta| >$ 2.

The per-trigger yield distributions $Y^{corr}(\Delta\phi)$ and $Y^{recoil}(\Delta\phi)$ for events with $N_{ch}^{rec} \geq$ 220 in the long-range region $|\Delta\eta| >$ 2.

The per-trigger yield distributions $Y^{corr}(\Delta\phi)$ and $Y^{recoil}(\Delta\phi)$ for events with $N_{ch}^{rec} \geq$ 220 in the long-range region $|\Delta\eta| >$ 2.

Integrated per-trigger yield, $Y_{int}$, on the near-side as a function of $p_{T}^{a}$ for 1 $< p_{T}^{b} <$ 3 GeV.

Integrated per-trigger yield, $Y_{int}$, on the near-side as a function of $p_{T}^{a}$ for 1 $< p_{T}^{b} <$ 3 GeV.

Integrated per-trigger yield, $Y_{int}$, on the near-side as a function of $p_{T}^{a}$ for 1 $< p_{T}^{b} <$ 3 GeV.

Integrated per-trigger yield, $Y_{int}$, on the near-side as a function of $p_{T}^{a}$ for 1 $< p_{T}^{b} <$ 3 GeV.

Integrated per-trigger yield, $Y_{int}$, on the near-side as a function of $p_{T}^{a}$ for 1 $< p_{T}^{b} <$ 3 GeV.

Integrated per-trigger yield, $Y_{int}$, on the away-side as a function of $p_{T}^{a}$ for 1 $< p_{T}^{b} <$ 3 GeV.

Integrated per-trigger yield, $Y_{int}$, on the away-side as a function of $p_{T}^{a}$ for 1 $< p_{T}^{b} <$ 3 GeV.

Integrated per-trigger yield, $Y_{int}$, on the away-side as a function of $p_{T}^{a}$ for 1 $< p_{T}^{b} <$ 3 GeV.

Integrated per-trigger yield, $Y_{int}$, on the away-side as a function of $p_{T}^{a}$ for 1 $< p_{T}^{b} <$ 3 GeV.

Integrated per-trigger yield, $Y_{int}$, on the away-side as a function of $p_{T}^{a}$ for 1 $< p_{T}^{b} <$ 3 GeV.

The integrated per-trigger yield, Y_{int}, on the near-side, the away-side and their difference and Y_{int} from the recoil as a function of event activity. Errors are statistical.

The integrated per-trigger yield, Y_{int}, on the near-side, the away-side and their difference and Y_{int} from the recoil as a function of event activity. Errors are statistical.

The Fourier coefficients $v_{n}$ as a function of $p_{T}^{a}$ extracted from the correlation functions, before and after the subtraction of the recoil component.

The Fourier coefficients $v_{n}$ as a function of $p_{T}^{a}$ extracted from the correlation functions, before and after the subtraction of the recoil component.

The Fourier coefficients $v_{n}$ as a function of $p_{T}^{a}$ extracted from the correlation functions, before and after the subtraction of the recoil component.

$v_{2}$, $v_{3}$, $v_{4}$ and $v_{5}$ as a function of $p_T^a$ for 1 $< p_{T}^{b} <$ 3 GeV for different $N_{ch}^{rec}$ intervals.

$v_{2}$, $v_{3}$, $v_{4}$ and $v_{5}$ as a function of $p_T^a$ for 1 $< p_{T}^{b} <$ 3 GeV for different $N_{ch}^{rec}$ intervals.

$v_{2}$, $v_{3}$, $v_{4}$ and $v_{5}$ as a function of $p_T^a$ for 1 $< p_{T}^{b} <$ 3 GeV for different $N_{ch}^{rec}$ intervals.

$v_{2}$, $v_{3}$, $v_{4}$ and $v_{5}$ as a function of $p_T^a$ for 1 $< p_{T}^{b} <$ 3 GeV for different $N_{ch}^{rec}$ intervals.

$v_{2}$, $v_{3}$, $v_{4}$ and $v_{5}$ as a function of $p_T^a$ for 1 $< p_{T}^{b} <$ 3 GeV for different $N_{ch}^{rec}$ intervals.

$v_{2}$, $v_{3}$, $v_{4}$ and $v_{5}$ as a function of $p_T^a$ for 1 $< p_{T}^{b} <$ 3 GeV for different $N_{ch}^{rec}$ intervals.

The values of factorization variable $r_{2}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{2}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{2}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{2}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{2}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{2}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{2}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{2}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{2}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{2}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{2}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{2}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{2}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{2}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{2}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{2}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{3}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{3}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{3}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{3}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{3}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{3}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{3}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{3}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{3}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{3}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{3}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The values of factorization variable $r_{3}$ defined by Eq.(11) before and after the subtraction of the recoil component. Errors are total experimental uncertainties.

The centrality dependence of $v_{2}$ as a function of $N_{ch}^{rec}$. Values from before and after the recoil subtraction are included.

The centrality dependence of $v_{3}$ as a function of $N_{ch}^{rec}$. Values from before and after the recoil subtraction are included.

The centrality dependence of $v_{4}$ as a function of $N_{ch}^{rec}$. Values from before and after the recoil subtraction are included.

The centrality dependence of $v_{2}$ as a function of $E_{T}^{Pb}$. Values from before and after the recoil subtraction are included.

The centrality dependence of $v_{3}$ as a function of $E_{T}^{Pb}$. Values from before and after the recoil subtraction are included.

The centrality dependence of $v_{4}$ as a function of $E_{T}^{Pb}$. Values from before and after the recoil subtraction are included.

The $v_{2}$ as a function of $E_{T}^{Pb}$ obtained indirectly by mapping from the $N_{ch}^{rec}-dependence of $v_{2}$ using the correlation data shown in Fig. 2(b).

The $v_{3}$ as a function of $E_{T}^{Pb}$ obtained indirectly by mapping from the $N_{ch}^{rec}-dependence of $v_{3}$ using the correlation data shown in Fig. 2(b).

The first-order harmonic of 2PC before recoil subtraction, $v_{1,1}^{unsub}$, as a function of $p_T^a$ for different $p_T^b$ ranges for events with $N_{ch}^{rec} \geq$ 220.

The first-order harmonic of 2PC before recoil subtraction, $v_{1,1}^{unsub}$, as a function of $p_T^a$ for different $p_T^b$ ranges for events with $N_{ch}^{rec} \geq$ 220.

The first-order harmonic of 2PC before recoil subtraction, $v_{1,1}^{unsub}$, as a function of $p_T^a$ for different $p_T^b$ ranges for events with $N_{ch}^{rec} \geq$ 220.

The first-order harmonic of 2PC before recoil subtraction, $v_{1,1}^{unsub}$, as a function of $p_T^a$ for different $p_T^b$ ranges for events with $N_{ch}^{rec} \geq$ 220.

The first-order harmonic of 2PC before recoil subtraction, $v_{1,1}^{unsub}$, as a function of $p_T^a$ for different $p_T^b$ ranges for events with $N_{ch}^{rec} \geq$ 220.

The first-order harmonic of 2PC before recoil subtraction, $v_{1,1}^{unsub}$, as a function of $p_T^a$ for different $p_T^b$ ranges for events with $N_{ch}^{rec} \geq$ 220.

The first-order harmonic of 2PC after recoil subtraction, $v_{1,1}$, as a function of $p_T^a$ for different $p_T^b$ ranges for events with $N_{ch}^{rec} \geq$ 220.

The first-order harmonic of 2PC after recoil subtraction, $v_{1,1}$, as a function of $p_T^a$ for different $p_T^b$ ranges for events with $N_{ch}^{rec} \geq$ 220.

The first-order harmonic of 2PC after recoil subtraction, $v_{1,1}$, as a function of $p_T^a$ for different $p_T^b$ ranges for events with $N_{ch}^{rec} \geq$ 220.

The first-order harmonic of 2PC after recoil subtraction, $v_{1,1}$, as a function of $p_T^a$ for different $p_T^b$ ranges for events with $N_{ch}^{rec} \geq$ 220.

The first-order harmonic of 2PC after recoil subtraction, $v_{1,1}$, as a function of $p_T^a$ for different $p_T^b$ ranges for events with $N_{ch}^{rec} \geq$ 220.

The first-order harmonic of 2PC after recoil subtraction, $v_{1,1}$, as a function of $p_T^a$ for different $p_T^b$ ranges for events with $N_{ch}^{rec} \geq$ 220.

The first-order harmonic $v_1$ obtained using factorization from $v_{1,1}$, as a function of $p_T^a$ for different $p_T^b$ ranges for events with $N_{ch}^{rec} \geq$ 220.

The first-order harmonic $v_1$ obtained using factorization from $v_{1,1}$, as a function of $p_T^a$ for different $p_T^b$ ranges for events with $N_{ch}^{rec} \geq$ 220.

The first-order harmonic $v_1$ obtained using factorization from $v_{1,1}$, as a function of $p_T^a$ for different $p_T^b$ ranges for events with $N_{ch}^{rec} \geq$ 220.

$v_{2}$ for Pb+Pb collisions in 55-60% centrality interval obtained using an EP method.

$v_{2}$ for Pb+Pb collisions in 55-60% centrality interval obtained using an EP method, after the scaling.

$v_{3}$ for Pb+Pb collisions in 55-60% centrality interval obtained using an EP method.

$v_{3}$ for Pb+Pb collisions in 55-60% centrality interval obtained using an EP method, after the scaling.

$v_{4}$ for Pb+Pb collisions in 55-60% centrality interval obtained using an EP method.

$v_{4}$ for Pb+Pb collisions in 55-60% centrality interval obtained using an EP method, after the scaling.

Correlation between $E_{T}^{FCal}$ and $N_{ch}^{rec}$ for MB events (without weighting) and MB+HMT events (with weighting), errors are statistical.

Differential production yield per binary collision for $W^{-}$ bosons as a function of $|\eta_\ell|$.

The lepton charge asymmetry $A_{\ell}$ from $W^\pm$ bosons as a function of absolute pseudorapidity.

Events in bins of the $E_{miss}^T$ in the SR.

Measured normalised differential fiducial cross sections of $\gamma\gamma \rightarrow \gamma\gamma$ production in Pb+Pb collisions at $\sqrt{s_{NN}}$ = 5.02 TeV for diphoton $|cos(\theta*)|$ are shown as points with error bars giving the statistical uncertainty and grey bands indicating the size of the total uncertainty. The results are compared with the prediction from the SuperChic v3.0 MC generator (solid line).

Measured normalised differential fiducial cross sections of $\gamma\gamma \rightarrow \gamma\gamma$ production in Pb+Pb collisions at $\sqrt{s_{NN}}$ = 5.02 TeV for average photon transverse momentum are shown as points with error bars giving the statistical uncertainty and grey bands indicating the size of the total uncertainty. The results are compared with the prediction from the SuperChic v3.0 MC generator (solid line) with bands denoting the theoretical uncertainty.

Measured normalised differential fiducial cross sections of $\gamma\gamma \rightarrow \gamma\gamma$ production in Pb+Pb collisions at $\sqrt{s_{NN}}$ = 5.02 TeV for average photon transverse momentum are shown as points with error bars giving the statistical uncertainty and grey bands indicating the size of the total uncertainty. The results are compared with the prediction from the SuperChic v3.0 MC generator (solid line).

Measured differential fiducial cross sections of $\gamma\gamma \rightarrow \gamma\gamma$ production in Pb+Pb collisions at $\sqrt{s_{NN}}$ = 5.02 TeV for diphoton rapidity are shown as points with error bars giving the statistical uncertainty and grey bands indicating the size of the total uncertainty. The results are compared with the prediction from the SuperChic v3.0 MC generator (solid line) with bands denoting the theoretical uncertainty.

Measured normalised differential fiducial cross sections of $\gamma\gamma \rightarrow \gamma\gamma$ production in Pb+Pb collisions at $\sqrt{s_{NN}}$ = 5.02 TeV for diphoton rapidity are shown as points with error bars giving the statistical uncertainty and grey bands indicating the size of the total uncertainty. The results are compared with the prediction from the SuperChic v3.0 MC generator (solid line).

The 95% CL upper limit on the ALP cross section $\sigma_{\gamma\gamma\rightarrow a \rightarrow\gamma\gamma}$ for the $\gamma\gamma\rightarrow a \rightarrow \gamma\gamma$ process as a function of ALP mass m$_{a}$. The observed upper limit is shown as a solid black line and the expected upper limit is shown by the dashed black line with its $\pm1$ and $\pm2$ standard deviation bands. The discontinuity at $m_{a}=70 GeV$ is caused by the increase of the mass-bin width which brings an increase in signal acceptance.

The 95% CL upper limit on the ALP coupling $1/\Lambda_{a}$ for the $\gamma\gamma\rightarrow a \rightarrow \gamma\gamma$ process as a function of ALP mass m$_{a}$. The observed upper limit is shown as a solid black line and the expected upper limit is shown by the dashed black line with its $\pm1$ and $\pm2$ standard deviation bands. The discontinuity at $m_{a} = 70 GeV$ is caused by the increase of the mass-bin width which brings an increase in signal acceptance.

Fiducial cross section for light-by-light scattering

Distribution of the NN output in the $\geq$1fj$\,$SR in data and the expected contribution of the signal and background processes after the profile-likelihood fit. The "Total" column corresponds to the sum of the expected contributions from the signal and background processes. The uncertainty represents the sum of statistical and systematic uncertainties in the signal and background predictions considering the correlations of the uncertainties as obtained by the fit.

Distribution of the NN output in the $t\bar{t}\gamma\,$CR in data and the expected contribution of the signal and background processes after the profile-likelihood fit. The "Total" column corresponds to the sum of the expected contributions from the signal and background processes. The uncertainty represents the sum of statistical and systematic uncertainties in the signal and background predictions considering the correlations of the uncertainties as obtained by the fit.

Total event yield in the $W\gamma\,$CR in data and the expected contribution of the signal and background processes after the profile-likelihood fit. The "Total" column corresponds to the sum of the expected contributions from the signal and background processes. The uncertainty represents the sum of statistical and systematic uncertainties in the signal and background predictions considering the correlations of the uncertainties as obtained by the fit.

Distribution of the scalar sum of the jet transverse momenta in the 0fj$\,$SR in data and for the sum of all processes expectations before the profile-likelihood fit. The "Total" column corresponds to the sum of the expected contributions from the signal and background processes. The uncertainty represents the sum of statistical and systematic uncertainties in the signal and background predictions. The first and last bins include the underflow and overflow, respectively.

Distribution of the $\eta$ of the $b$-tagged jet in the 0fj$\,$SR in data and for the sum of all processes expectations before the profile-likelihood fit. The "Total" column corresponds to the sum of the expected contributions from the signal and background processes. The uncertainty represents the sum of statistical and systematic uncertainties in the signal and background predictions.

Distribution of the reconstructed top-quark mass in the 0fj$\,$SR in data and for the sum of all processes expectations before the profile-likelihood fit. The "Total" column corresponds to the sum of the expected contributions from the signal and background processes. The uncertainty represents the sum of statistical and systematic uncertainties in the signal and background predictions. The first and last bins include the underflow and overflow, respectively.

Distribution of the $p_{\mathrm{T}}$ of the top-quark+photon system in the 0fj$\,$SR in data and for the sum of all processes expectations before the profile-likelihood fit. The "Total" column corresponds to the sum of the expected contributions from the signal and background processes. The uncertainty represents the sum of statistical and systematic uncertainties in the signal and background predictions and the last bin includes the overflow.

Distribution of the photon $p_{\mathrm{T}}$ in the 0fj$\,$SR in data and for the sum of all processes expectations before the profile-likelihood fit. The "Total" column corresponds to the sum of the expected contributions from the signal and background processes. The uncertainty represents the sum of statistical and systematic uncertainties in the signal and background predictions and the last bin includes the overflow.

Distribution of the photon $\eta$ in the 0fj$\,$SR in data and for the sum of all processes expectations before the profile-likelihood fit. The "Total" column corresponds to the sum of the expected contributions from the signal and background processes. The uncertainty represents the sum of statistical and systematic uncertainties in the signal and background predictions.

Distribution of the scalar sum of the jet transverse momenta in the $\geq$1fj$\,$SR in data and for the sum of all processes expectations before the profile-likelihood fit. The "Total" column corresponds to the sum of the expected contributions from the signal and background processes. The uncertainty represents the sum of statistical and systematic uncertainties in the signal and background predictions and the last bin includes the overflow.

Distribution of the invariant mass of the $b$-tagged jet and the highest-$p_{\mathrm{T}}$ forward jet in the $\geq$1fj$\,$SR in data and for the sum of all processes expectations before the profile-likelihood fit. The "Total" column corresponds to the sum of the expected contributions from the signal and background processes. The uncertainty represents the sum of statistical and systematic uncertainties in the signal and background predictions and the last bin includes the overflow.

Distribution of $p_{\mathrm{T}}$ of the highest-$p_{\mathrm{T}}$ forward jet in the $\geq$1fj$\,$SR in data and for the sum of all processes expectations before the profile-likelihood fit. The "Total" column corresponds to the sum of the expected contributions from the signal and background processes. The uncertainty represents the sum of statistical and systematic uncertainties in the signal and background predictions and the last bin includes the overflow.

Distribution of the difference in $\eta$ between the highest-$p_{\mathrm{T}}$ forward jet and the photon in the $\geq$1fj$\,$SR in data and for the sum of all processes expectations before the profile-likelihood fit. The "Total" column corresponds to the sum of the expected contributions from the signal and background processes. The uncertainty represents the sum of statistical and systematic uncertainties in the signal and background predictions and the last bin includes the overflow.

Distribution of the energy of the system formed by the highest-$p_{\mathrm{T}}$ forward jet and the photon in the $\geq$1fj$\,$SR in data and for the sum of all processes expectations before the profile-likelihood fit. The "Total" column corresponds to the sum of the expected contributions from the signal and background processes. The uncertainty represents the sum of statistical and systematic uncertainties in the signal and background predictions. The first and last bins include the underflow and overflow, respectively.

Distribution of the $\eta$ of the $b$-tagged jet in the $\geq$1fj$\,$SR in data and for the sum of all processes expectations before the profile-likelihood fit. The "Total" column corresponds to the sum of the expected contributions from the signal and background processes. The uncertainty represents the sum of statistical and systematic uncertainties in the signal and background predictions.

Distribution of the reconstructed top-quark mass in the $\geq$1fj$\,$SR in data and for the sum of all processes expectations before the profile-likelihood fit. The "Total" column corresponds to the sum of the expected contributions from the signal and background processes. The uncertainty represents the sum of statistical and systematic uncertainties in the signal and background predictions. The first and last bins include the underflow and overflow, respectively.

Distribution of the $p_{\mathrm{T}}$ of the top-quark and photon system in the $\geq$1fj$\,$SR in data and for the sum of all processes expectations before the profile-likelihood fit. The "Total" column corresponds to the sum of the expected contributions from the signal and background processes. The uncertainty represents the sum of statistical and systematic uncertainties in the signal and background predictions and the last bin includes the overflow.

Distribution of the photon $p_{\mathrm{T}}$ in the $\geq$1fj$\,$SR in data and for the sum of all processes expectations before the profile-likelihood fit. The "Total" column corresponds to the sum of the expected contributions from the signal and background processes. The uncertainty represents the sum of statistical and systematic uncertainties in the signal and background predictions and the last bin includes the overflow.

Distribution of the photon $\eta$ in the $\geq$1fj$\,$SR in data and for the sum of all processes expectations before the profile-likelihood fit. The "Total" column corresponds to the sum of the expected contributions from the signal and background processes. The uncertainty represents the sum of statistical and systematic uncertainties in the signal and background predictions.

Ordered list of the 30 systematic uncertainties with the largest impact on the measured signal normalisation in the fit to data in the parton-level measurement considered as nuisance parameters (NPs) in the profile-likelihood fit. The column "NP value, error" corresponds to the nominal best-fit values and the corresponding uncertainties. The impact of each NP, $\Delta\sigma$/$\sigma_{\mathrm{pred}}$, is computed by comparing the nominal best-fit value of the POI ($\sigma$/$\sigma_{\mathrm{pred}}$) with the result of the fits when fixing the considered NP to its best-fit value shifted by its pre-fit and post-fit uncertainties. The corresponding impacts are listed in the "POI impact prefit high/low" and "POI impact high/low" columns, respectively. The "MC stat." NPs represent the MC statistical uncertainty and they enter the likelihood with a Poisson term, while all the other NPs enter the likelihood via a Gaussian term.

Ordered list of the 30 systematic uncertainties with the largest impact on the measured signal normalisation in the fit to data in the particle-level measurement considered as nuisance parameters (NPs) in the profile-likelihood fit. The column "NP value, error" corresponds to the nominal best-fit values and the corresponding uncertainties. The impact of each NP, $\Delta\sigma$/$\sigma_{\mathrm{pred}}$, is computed by comparing the nominal best-fit value of the POI ($\sigma$/$\sigma_{\mathrm{pred}}$) with the result of the fits when fixing the considered NP to its best-fit value shifted by its pre-fit and post-fit uncertainties. The corresponding impacts are listed in the "POI impact prefit high/low" and "POI impact high/low" columns, respectively. The "MC stat." NPs represent the MC statistical uncertainty and they enter the likelihood with a Poisson term, while all the other NPs enter the likelihood via a Gaussian term.

This table lists the kinematic requirements on parton-level objects used to define of the fiducial phase space for the parton-level measurement. Frixione isolation ($\href{https://arxiv.org/abs/hep-ph/9801442}{\text{hep-ph/9801442}}$) with a chosen radius of $\Delta R = 0.2$ is applied to photons ($\gamma$). The measured fiducial parton-level cross section is $\sigma_{tq\gamma}\times\mathcal{B}(t\rightarrow l\nu b) = 688\pm 23(\text{stat.})^{+75}_{-71}(\text{syst.})\,$fb.

This table lists the kinematic requirements on particle-level objects used to define of the fiducial phase space for the particle-level measurement. The particle level objects are photons ($\gamma$) not from a hadron decay, neutrinos not from a hadron decay ($\nu$), prompt electrons and muons ($\ell$) "dressed" by adding close-by ($\Delta R < 0.1$) photons, and anti-$k_t$ $R = 0.4$ jets built from stable particles ($\tau > 30\,$ps) and tau leptons excluding neutrinos and prompt dressed muons. Jets are $b$-tagged ($b$-jet) using ghost-matched $b$-hadrons with $p_{\text{T}} > 5\,$GeV. Apart from the kinematic requirements, isolation and overlap removal criteria are applied. Jets within $\Delta R = 0.4$ of a photon are removed if the $p_{\text{T}}$ of charged particles within $\Delta R = 0.3$ of the photon is smaller than $10\,\%$ of its $p_{\text{T}}$. Jets within $\Delta R = 0.4$ of a lepton are removed. Events are removed where a photon is close ($\Delta R < 0.4$) to a lepton or a surviving jet. The measured fiducial particle-level cross section is $\sigma_{tq\gamma}\times\mathcal{B}(t\rightarrow l\nu b)+\sigma_{t(\rightarrow l\nu b\gamma)q} = 303\pm 9(\text{stat.})^{+33}_{-32}(\text{syst.})\,$fb.