Distribution of mT2 for events passing all the signal candidate selection requirements, except that on mT2 of the L100 selection, for DF events.

Distribution of mT2 for events passing all the signal candidate selection requirements, except that on mT2 of the L110 selection, for SF events.

Distribution of mT2 for events passing all the signal candidate selection requirements, except that on mT2 of the L110 selection, for DF events.

Observed 95% CL exclusion contour in the (STOP1, CHARGINO1) mass plane for a fixed value of m(NEUTRALINO1) = 1 GeV.

Expected 95% CL exclusion contour in the (STOP1, CHARGINO1) mass plane for a fixed value of m(NEUTRALINO1) = 1 GeV.

Observed 95% CL exclusion minus 1 sigma contour in the (STOP1, CHARGINO1) mass plane for a fixed value of m(NEUTRALINO1) = 1 GeV.

Observed 95% CL exclusion contour in the (STOP1, NEUTRALINO1) mass plane for a fixed value of m(STOP1) - m(CHARGINO1) = 10 GeV.

Expected 95% CL exclusion contour in the (STOP1, NEUTRALINO1) mass plane for a fixed value of m(STOP1) - m(CHARGINO1) = 10 GeV.

Observed 95% CL exclusion minus 1 sigma contour in the (STOP1, NEUTRALINO1) mass plane for a fixed value of m(STOP1) - m(CHARGINO1) = 10 GeV.

Observed 95% CL exclusion contour in the (CHARGINO1, NEUTRALINO1) mass plane for a fixed value of m(STOP1) = 300 GeV.

Expected 95% CL exclusion contour in the (CHARGINO1, NEUTRALINO1) mass plane for a fixed value of m(STOP1) = 300 GeV.

Observed 95% CL exclusion minus one sigma contour in the (CHARGINO1, NEUTRALINO1) mass plane for a fixed value of m(STOP1) = 300 GeV.

Observed 95% CL exclusion contour in the (STOP1, NEUTRALINO1) mass plane of m(CHARGINO1) = 2 m(NEUTRALINO1).

Expected 95% CL exclusion contour in the (STOP1, NEUTRALINO1) mass plane for m(CHARGINO1) = 2 m(NEUTRALINO1).

Observed 95% CL exclusion minus one sigma contour in the (STOP1, NEUTRALINO1) mass plane for m(CHARGINO1) = 2 m(NEUTRALINO1).

Observed 95% CL exclusion contour in the (STOP1, NEUTRALINO1) mass plane.

Expected 95% CL exclusion contour in the (STOP1, NEUTRALINO1) mass plane.

Observed 95% CL exclusion minus one sigma contour in the (STOP1, NEUTRALINO1) mass plane.

Observed 95% CL exclusion contour in the (STOP1, NEUTRALINO1) mass plane.

Expected 95% CL exclusion contour in the (STOP1, NEUTRALINO1) mass plane.

Observed 95% CL exclusion minus one sigma contour in the (STOP1, NEUTRALINO1) mass plane.

Observed 95% CL exclusion contour in the (STOP1, NEUTRALINO1) mass plane for a fixed value of m(CHARGINO1) = 106 GeV.

Expected 95% CL exclusion contour in the (STOP1, NEUTRALINO1) mass plane for a fixed value of m(CHARGINO1) = 106 GeV.

Observed 95% CL exclusion minus one sigma contour in the (STOP1, NEUTRALINO1) mass plane for a fixed value of m(CHARGINO1) = 106 GeV.

Expected CLs values for Fig. 14.

Observed CLs values for Fig. 14.

Observed cross-section limits for Fig. 14.

Expected CLs values for Fig. 15.

Observed CLs values for Fig. 15.

Observed cross-section limits for Fig. 15.

Expected CLs values for Fig. 16.

Observed CLs values for Fig. 16.

Observed cross-section limits for Fig. 14.

Expected CLs values for Fig. 17.

Observed CLs values for Fig. 17.

Observed cross-section limits for Fig. 17.

Expected CLs values for Fig. 19.

Observed CLs values for Fig. 19.

Observed cross-section limits for Fig. 19.

Number of generated events for Fig. 16.

Acceptance of S1 for Fig. 16.

Efficiency of S1 for Fig. 16.

Total systematic uncertainty IN PCT on signal yields in S1 for Fig. 16.

Acceptance of S2 for Fig. 16.

Efficiency of S2 for Fig. 16.

Total systematic uncertainty IN PCT on signal yields in S2 for Fig. 16.

Acceptance of S3 for Fig. 16.

Efficiency of S3 for Fig. 16.

Total systematic uncertainty IN PCT on signal yields in S3 for Fig. 16.

Acceptance of S4 for Fig. 16.

Efficiency of S4 for Fig. 16.

Total systematic uncertainty IN PCT on signal yields in S4 for Fig. 16.

Acceptance of S5 for Fig. 16.

Efficiency of S5 for Fig. 16.

Total systematic uncertainty IN PCT on signal yields in S5 for Fig. 16.

Acceptance of S6 for Fig. 16.

Efficiency of S6 for Fig. 16.

Total systematic uncertainty IN PCT on signal yields in S6 for Fig. 16.

Acceptance of S7 for Fig. 16.

Efficiency of S7 for Fig. 16.

Total systematic uncertainty IN PCT on signal yields in S7 for Fig. 16.

Acceptance of H160 for Fig. 16.

Efficiency of H160 for Fig. 16.

Total systematic uncertainty IN PCT on signal yields in H160 for Fig. 16.

Observed cross-section limits for Fig. 20.

Observed cross-section limits for Fig. 20.

Expected cross-section limits for Fig. 20.

Total number of generated MC events for each point of the grid.

Observed upper limit on the signal cross-section, in pb, for each point of the grid in the different flavour channel.

Observed upper limit on the signal cross-section, in pb, for each point of the grid in the same flavour channel.

The best expected signal region chosen for each point of the grid in the different flavour channel.

The best expected signal region chosen for each point of the grid in the same flavour channel.

Observed CLs for each point of the grid in the different flavour channel.

Observed CLs for each point of the grid in the same flavour channel.

Expected CLs for each point of the grid in the different flavour channel.

Expected CLs for each point of the grid in the same flavour channel.

Signal acceptance for all the analysis cuts, except the BDTG cut, for each point of the grid in the different flavour channel.

Signal acceptance for all the analysis cuts, except the BDTG cut, for each point of the grid in the same flavour channel.

Signal efficiency, including the acceptance of the BDTG cut, for each point of the grid in the different flavour channel.

Signal efficiency, including the acceptance of the BDTG cut, for each point of the grid in the same flavour channel.

Total signal experimental systematic uncertainty for each point of the grid in the different flavour channel.

Signal efficiency, including the acceptance of the BDTG cut, for each point of the grid in the same flavour channel.

Number of simulated events passing various stages of the selection in the hadronic mT2 analysis for a signal sample with m(STOP1)=300 GeV, m(CHARGINO1) = 150 GeV and m(NEUTRALINO1) = 50 GeV, and with the top squark decaying as STOP1 --> CHARGINO1+ BOTTOM -> W(*)+NEUTRALINO1+ BOTTOM with unit probability. Event weights are applied to correct simulated events to data. "Isolation" includes the effects of tight ID for electrons and the isolation selection for both electrons and muons. "Cleaning cuts" refer to cuts applied to remove non-collision backgrounds and detector noise.

Number of simulated events passing various stages of the selection in the hadronic mT2 analysis for a signal sample with m(STOP1)=250 GeV, m(CHARGINO1) = 106 GeV and m(NEUTRALINO1) = 60 GeV, and with the top squark decaying as STOP1 --> CHARGINO1+ BOTTOM -> W(*)+NEUTRALINO1+ BOTTOM with unit probability. Event weights are applied to correct simulated events to data. "Isolation" includes the effects of tight ID for electrons and the isolation selection for both electrons and muons. "Cleaning cuts" refer to cuts applied to remove non-collision backgrounds and detector noise.

Number of simulated events passing various stages of the selection in the leptonic mT2 analysis for two signal samples with the top squark decaying as STOP1 --> CHARGINO1+ BOTTOM -> W(*)+NEUTRALINO1+BOTTOM with unit probability. Event weights are applied to correct simulated events to data. "Isolation" includes the effects of tight ID for electrons and the isolation selection for both electrons and muons. "Cleaning cuts" refer to cuts applied to remove non-collision backgrounds and detector noise.

Number of simulated events passing various stages of the selection in the leptonic mT2 analysis for a signal sample with m(STOP1)=180 GeV and m(NEUTRALINO1) = 60 GeV, and with the top squark decaying as STOP1 --> W + BOTTOM + NEUTRALINO1 with unit probability. Event weights are applied to correct simulated events to data. "Isolation" includes the effects of tight ID for electrons and the isolation selection for both electrons and muons. "Cleaning cuts" refer to cuts applied to remove non-collision backgrounds and detector noise.

Number of simulated events passing various stages of the selection in the in the MVA analysis for all signal samples used to train the BDTG and with the top squark decaying as STOP1 --> TOP + NEUTRALINO1 with unit probability. Event weights are applied to correct simulated events to data. "Isolation" includes the effects of tight ID for electrons and the isolation selection for both electrons and muons. "Cleaning cuts" refer to cuts applied to remove non-collision backgrounds and detector noise. The index i in category (Ci) is the one reported in table 3 for each signal point, SF and DF.

Lower mass requirement for signal regions.

Lower mass requirement for signal regions.

Lower mass requirement for signal regions.

Expected and observed events in the 16 discovery regions along with the according control regions.

Expected and observed events in the 16 discovery regions along with the according control regions.

Expected signal yield and acceptance x efficiency, estimated background and observed number of events in data for the full range of simulated masses in the MS-agnostic R-hadron search.

Expected signal yield and acceptance x efficiency, estimated background and observed number of events in data for the full range of simulated masses in the MS-agnostic R-hadron search.

Expected signal yield and acceptance x efficiency, estimated background and observed number of events in data for the full range of simulated masses in the full-detector R-hadron search.

Expected signal yield and acceptance x efficiency, estimated background and observed number of events in data for the full range of simulated masses in the full-detector R-hadron search.

p0-values and model-independent upper limits on cross-section x acceptance x efficiency for the 16 discovery regions.

p0-values and model-independent upper limits on cross-section x acceptance x efficiency for the 16 discovery regions.

Expected signal yield and acceptance x efficiency, estimated background and observed number of events in data for the full range of simulated masses in the MS-agnostic search for metastable gluino R-hadrons.

Expected signal yield and acceptance x efficiency, estimated background and observed number of events in data for the full range of simulated masses in the MS-agnostic search for metastable gluino R-hadrons.

Expected signal yield and acceptance x efficiency, estimated background and observed number of events in data for the full range of simulated masses in the full-detector direct-stau search.

Expected signal yield and acceptance x efficiency, estimated background and observed number of events in data for the full range of simulated masses in the full-detector direct-stau search.

Expected signal yield and acceptance x efficiency, estimated background and observed number of events in data for the full range of simulated masses in the full-detector chargino search.

Expected signal yield and acceptance x efficiency, estimated background and observed number of events in data for the full range of simulated masses in the full-detector chargino search.

Upper cross-section limit in gluino R-hadron search.

Upper cross-section limit in gluino R-hadron search.

Upper cross-section limit in sbottom R-hadron search.

Upper cross-section limit in sbottom R-hadron search.

Upper cross-section limit in stop R-hadron search.

Upper cross-section limit in stop R-hadron search.

Upper cross-section limit in stau search.

Upper cross-section limit in stau search.

Upper cross-section limit in chargino search.

Upper cross-section limit in chargino search.

Lower mass limit as function of gluino lifetime.

Lower mass limit as function of gluino lifetime.

Acceptance x efficiency, acceptance and efficiency for the full range of simulated masses in the MS-agnostic R-hadron search.

Acceptance x efficiency, acceptance and efficiency for the full range of simulated masses in the MS-agnostic R-hadron search.

Upper cross-section limit in meta-stable gluino R-hadron search.

Upper cross-section limit in meta-stable gluino R-hadron search.

Flavor composition of 800 GeV stop R-hadrons simulated using the generic model as a function of radial distance from the interaction point.

Flavor composition of 800 GeV stop R-hadrons simulated using the generic model as a function of radial distance from the interaction point.

Flavor composition of 800 GeV anti-stop R-hadrons simulated using the generic model as a function of radial distance from the interaction point.

Flavor composition of 800 GeV anti-stop R-hadrons simulated using the generic model as a function of radial distance from the interaction point.

Flavor composition of 800 GeV stop R-hadrons simulated using the Regge model as a function of radial distance from the interaction point.

Flavor composition of 800 GeV stop R-hadrons simulated using the Regge model as a function of radial distance from the interaction point.

Flavor composition of 800 GeV anti-stop R-hadrons simulated using the Regge model as a function of radial distance from the interaction point.

Flavor composition of 800 GeV anti-stop R-hadrons simulated using the Regge model as a function of radial distance from the interaction point.

ETmiss trigger efficiency as function of true ETmiss (EtmissTurnOn).

ETmiss trigger efficiency as function of true ETmiss (EtmissTurnOn).

Single-muon trigger efficiency as function of $|\eta|$ and $\beta$ (SingleMuTurnOn).

Single-muon trigger efficiency as function of $|\eta|$ and $\beta$ (SingleMuTurnOn).

Candidate reconstruction efficiency for ID+Calo selection (IDCaloEff).

Candidate reconstruction efficiency for ID+Calo selection (IDCaloEff).

Candidate reconstruction efficiency for loose selection (LooseEff).

Candidate reconstruction efficiency for loose selection (LooseEff).

Efficiency for a loose candidate to be promoted to a tight candidate (TightPromotionEff).

Efficiency for a loose candidate to be promoted to a tight candidate (TightPromotionEff).

Resolution and average of reconstructed dE/dx mass for a given simulated mass for ID+calo candidates.

Resolution and average of reconstructed dE/dx mass for a given simulated mass for ID+calo candidates.

Resolution and average of reconstructed ToF mass for a given simulated mass for ID+calo candidates.

Resolution and average of reconstructed ToF mass for a given simulated mass for ID+calo candidates.

Resolution and average of reconstructed ToF mass for a given simulated mass for FullDet candidates.

Resolution and average of reconstructed ToF mass for a given simulated mass for FullDet candidates.

Distribution of $m_{\mathrm{JJ}}$ in the SR (black points) compared to the background-only expectation, where the shape is derived from the CR and the normalisation is fitted in the SR.

Observed and expected limits at 95% CL on the cross-section times branching ratio for the production of dark quarks through a $Z^\prime$ as a function of the $Z^\prime$ mass for model A (see the text for details). The darker and lighter shaded bands around the expected limits represent the $\pm 1 \sigma$ and $\pm 2 \sigma$ uncertainty range, respectively.

The predicted theory cross-section times branching ratio for the production of dark quarks through a $Z^\prime$ as a function of the $Z^\prime$ mass for model A (see the text for details). In the theory model, a universal coupling of the $Z^\prime$ to each SM quark $g_q$ is assumed and the dark quarks are all represented by one species with an effective coupling $g_{q_d}$ to the $Z^\prime$. For model A, the predicted cross-section is shown for $g_q=0.05$ and $g_{q_d} = 0.2$.

Observed and expected limits at 95% CL on the cross-section times branching ratio for the production of dark quarks through a $Z^\prime$ as a function of the $Z^\prime$ mass for model B (see the text for details). The darker and lighter shaded bands around the expected limits represent the $\pm 1 \sigma$ and $\pm 2 \sigma$ uncertainty range, respectively.

The predicted theory cross-section times branching ratio for the production of dark quarks through a $Z^\prime$ as a function of the $Z^\prime$ mass for model B (see the text for details). In the theory model, a universal coupling of the $Z^\prime$ to each SM quark $g_q$ is assumed and the dark quarks are all represented by one species with an effective coupling $g_{q_d}$ to the $Z^\prime$. For model B, the predicted cross-section is shown for $g_q=0.15$ and $g_{q_d} = 0.5$.

Observed and expected limits at 95% CL on the cross-section times branching ratio for the production of dark quarks through a $Z^\prime$ as a function of the $Z^\prime$ mass for model C (see the text for details). The darker and lighter shaded bands around the expected limits represent the $\pm 1 \sigma$ and $\pm 2 \sigma$ uncertainty range, respectively.

The predicted theory cross-section times branching ratio for the production of dark quarks through a $Z^\prime$ as a function of the $Z^\prime$ mass for model C (see the text for details). In the theory model, a universal coupling of the $Z^\prime$ to each SM quark $g_q$ is assumed and the dark quarks are all represented by one species with an effective coupling $g_{q_d}$ to the $Z^\prime$. For model C, the predicted cross-section is shown for $g_q=0.15$ and $g_{q_d}= 0.5$.

Observed and expected limits at 95% CL on the cross-section times branching ratio for the production of dark quarks through a $Z^\prime$ as a function of the $Z^\prime$ mass for model D (see the text for details). The darker and lighter shaded bands around the expected limits represent the $\pm 1 \sigma$ and $\pm 2 \sigma$ uncertainty range, respectively.

The predicted theory cross-section times branching ratio for the production of dark quarks through a $Z^\prime$ as a function of the $Z^\prime$ mass for model D (see the text for details). In the theory model, a universal coupling of the $Z^\prime$ to each SM quark $g_q$ is assumed and the dark quarks are all represented by one species with an effective coupling $g_{q_d}$ to the $Z^\prime$. For model D, the predicted cross-section is shown for $g_q=0.05$ and $g_{q_d} = 0.2$.

Relative efficiency (in %) of each stage of the analysis selection with respect to the previous one for models A through D and $m_{Z^\prime}=2.5$ TeV. Approximately 90000 events were generated per signal mass hypothesis.

Expected and observed 95% CL upper limits on the branching ratio of the Higgs boson decay to $Za$ times the branching ratio $a $->$ \gamma \gamma$ as a function of the $a$ particle mass in the merged ($m_a \le 2$ GeV) and the resolved ($m_a > 2$ GeV) categories. The figure uses $pp$ collision data at $\sqrt{s}=13$ TeV corresponding to 139 fb$^{-1}$.

ATLAS observed 95% CL exclusion contours limits in terms of the ALP mass and its effective coupling to photons, $|C_{\gamma\gamma}|/\Lambda$, for different values of the Higgs coupling to $Za$, $|C_{Zh}|/\Lambda$. The figure uses $pp$ collision data at $\sqrt{s}=13$ TeV corresponding to 139 fb$^{-1}$. The value 100 in the tables is presentational, denoting the upper edge of the plot.