Efficiency to reconstruct an MS DV in the MS endcaps fiducial volume as a function of the longitudinal decay position of the LLP for scalar portal samples with $m_\varPhi=125$~\GeV\ for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling. The vertical lines show the relevant detector boundaries, where ``HCal end'' is the outer limit of the hadronic calorimeter, ``MDT 1/2'' represent the first/second stations of MDT chambers and ``L/S'' indicate whether they are in Large or Small sectors.

Expected and observed 95% CL limits on ($\sigma / \sigma_{\text{SM}}) \times B$ for $m_\phi=125$ GeV assuming 100% branching fraction of the long-lived scalar into fermion pairs.

Efficiency for the Muon RoI Cluster trigger in the MS barrel fiducial volume as a function of the transverse decay position of the LLP for non-SM Higgs benchmark samples for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling. The vertical lines show the relevant detector boundaries, where ``HCal end'' is the outer limit of the hadronic calorimeter, RPC 1/2 represent the first/second stations of RPC chambers, TGC 1 represents the first stations of TGC chambers and L/S indicate whether they are in the Large or Small sectors. The dependence on detector geometry can be seen by looking at the various detector boundaries depicted by dashed lines in the plots. In the barrel the trigger efficiency increases when the LLP decays are close to the end of the hadronic calorimeter (r ~ 4 m) and substantially decreases as the decay occurs closer to the middle station of the muon spectrometer (r ~ 7 m). For decays occurring close to the middle station the charged hadrons and photons (and their EM showers) are not spatially separated, resulting in low efficiency (the same reasoning is applicable to decays in the endcap regions).

Efficiency for the Muon RoI Cluster trigger in the MS endcaps fiducial volume as a function of the longitudinal decay position of the LLP for non-SM Higgs benchmark samples for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling. The vertical lines show the relevant detector boundaries, where ``HCal end'' is the outer limit of the hadronic calorimeter, RPC 1/2 represent the first/second stations of RPC chambers, TGC 1 represents the first stations of TGC chambers and L/S indicate whether they are in the Large or Small sectors. The dependence on detector geometry can be seen by looking at the various detector boundaries depicted by dashed lines in the plots. In the barrel the trigger efficiency increases when the LLP decays are close to the end of the hadronic calorimeter (r ~ 4 m) and substantially decreases as the decay occurs closer to the middle station of the muon spectrometer (r ~ 7 m). For decays occurring close to the middle station the charged hadrons and photons (and their EM showers) are not spatially separated, resulting in low efficiency (the same reasoning is applicable to decays in the endcap regions).

Efficiency for the Muon RoI Cluster trigger in the MS barrel fiducial volume as a function of the transverse decay position of the LLP for non-SM Higgs benchmark samples for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling. The vertical lines show the relevant detector boundaries, where ``HCal end'' is the outer limit of the hadronic calorimeter, RPC 1/2 represent the first/second stations of RPC chambers, TGC 1 represents the first stations of TGC chambers and L/S indicate whether they are in the Large or Small sectors. The dependence on detector geometry can be seen by looking at the various detector boundaries depicted by dashed lines in the plots. In the barrel the trigger efficiency increases when the LLP decays are close to the end of the hadronic calorimeter (r ~ 4 m) and substantially decreases as the decay occurs closer to the middle station of the muon spectrometer (r ~ 7 m). For decays occurring close to the middle station the charged hadrons and photons (and their EM showers) are not spatially separated, resulting in low efficiency (the same reasoning is applicable to decays in the endcap regions).

Efficiency for the Muon RoI Cluster trigger in the MS endcaps fiducial volume as a function of the longitudinal decay position of the LLP for non-SM Higgs benchmark samples for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling. The vertical lines show the relevant detector boundaries, where ``HCal end'' is the outer limit of the hadronic calorimeter, RPC 1/2 represent the first/second stations of RPC chambers, TGC 1 represents the first stations of TGC chambers and L/S indicate whether they are in the Large or Small sectors. The dependence on detector geometry can be seen by looking at the various detector boundaries depicted by dashed lines in the plots. In the barrel the trigger efficiency increases when the LLP decays are close to the end of the hadronic calorimeter (r ~ 4 m) and substantially decreases as the decay occurs closer to the middle station of the muon spectrometer (r ~ 7 m). For decays occurring close to the middle station the charged hadrons and photons (and their EM showers) are not spatially separated, resulting in low efficiency (the same reasoning is applicable to decays in the endcap regions).

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi = 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi = 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi = 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi = 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi = 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi = 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi = 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi = 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi \ne 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi \ne 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi \ne 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi \ne 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi \ne 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi \ne 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi \ne 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi \ne 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi \ne 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi \ne 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi \ne 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi \ne 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi \ne 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi \ne 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi \ne 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi \ne 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi \ne 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi \ne 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi \ne 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency of the Muon RoI Cluster trigger in the MS fiducial volume as a function of the LLP boost and decay position for scalar portal samples with $m_\varPhi \ne 125$ GeV for events passing the data quality requirements and having a reconstructed primary vertex. These efficiency distributions are based solely on MC simulation, without any corrections applied for mismodeling.

Efficiency to reconstruct an MS DV in the MS barrel fiducial volume as a function of transverse decay position of the LLP for non-SM Higgs benchmark samples for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling. The vertical lines show the relevant detector boundaries, where “HCal end” is the outer limit of the hadronic calorimeter, MDT 1/2 represent the first/second stations of MDT chambers and L/S indicate whether they are in Large or Small sectors.

Efficiency to reconstruct an MS DV in the MS endcaps fiducial volume as a function of longitudinal decay position of the LLP for non-SM Higgs benchmark samples for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling. The vertical lines show the relevant detector boundaries, where “HCal end” is the outer limit of the hadronic calorimeter, MDT 1/2 represent the first/second stations of MDT chambers and L/S indicate whether they are in Large or Small sectors.

Efficiency to reconstruct an MS DV in the MS barrel fiducial volume as a function of transverse decay position of the LLP for non-SM Higgs benchmark samples for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling. The vertical lines show the relevant detector boundaries, where “HCal end” is the outer limit of the hadronic calorimeter, MDT 1/2 represent the first/second stations of MDT chambers and L/S indicate whether they are in Large or Small sectors.

Efficiency to reconstruct an MS DV in the MS endcaps fiducial volume as a function of longitudinal decay position of the LLP for non-SM Higgs benchmark samples for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling. The vertical lines show the relevant detector boundaries, where “HCal end” is the outer limit of the hadronic calorimeter, MDT 1/2 represent the first/second stations of MDT chambers and L/S indicate whether they are in Large or Small sectors.

Efficiency to reconstruct an MS DV in the MS barrel fiducial volume as a function of the transverse decay position of the LLP for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling. The vertical lines show the relevant detector boundaries, where “HCal end” is the outer limit of the hadronic calorimeter, MDT 1/2 represent the first/second stations of MDT chambers and L/S indicate whether they are in Large or Small sectors.

Efficiency to reconstruct an MS DV in the MS endcaps fiducial volume as a function of the longitudinal decay position of the LLP for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling. The vertical lines show the relevant detector boundaries, where “HCal end” is the outer limit of the hadronic calorimeter, MDT 1/2 represent the first/second stations of MDT chambers and L/S indicate whether they are in Large or Small sectors.

Efficiency to reconstruct an MS DV in the MS barrel fiducial volume as a function of the transverse decay position of the LLP for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling. The vertical lines show the relevant detector boundaries, where “HCal end” is the outer limit of the hadronic calorimeter, MDT 1/2 represent the first/second stations of MDT chambers and L/S indicate whether they are in Large or Small sectors.

Efficiency to reconstruct an MS DV in the MS endcaps fiducial volume as a function of the longitudinal decay position of the LLP for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling. The vertical lines show the relevant detector boundaries, where “HCal end” is the outer limit of the hadronic calorimeter, MDT 1/2 represent the first/second stations of MDT chambers and L/S indicate whether they are in Large or Small sectors.

Efficiency to reconstruct an MS DV in the MS barrel fiducial volume as a function of the transverse decay position of the LLP for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling. The vertical lines show the relevant detector boundaries, where “HCal end” is the outer limit of the hadronic calorimeter, MDT 1/2 represent the first/second stations of MDT chambers and L/S indicate whether they are in Large or Small sectors.

Efficiency to reconstruct an MS DV in the MS endcaps fiducial volume as a function of the longitudinal decay position of the LLP for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling. The vertical lines show the relevant detector boundaries, where “HCal end” is the outer limit of the hadronic calorimeter, MDT 1/2 represent the first/second stations of MDT chambers and L/S indicate whether they are in Large or Small sectors.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi = 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi = 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi = 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi = 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi = 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi = 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi = 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi = 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi = 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi = 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi = 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi = 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi = 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi = 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi = 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi = 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Efficiency to reconstruct an MS DV in the MS fiducial volume as a function of the LLP boost and decay position for a scalar portal sample with $m_\varPhi \ne 125$ GeV for vertices that pass the baseline event selection (except for the trigger) and satisfy the vertex isolation criteria. The efficiency distributions are corrected for mismodeling.

Expected and observed 95% CL limits on ($\sigma / \sigma_{\text{SM}}) \times B$ for $m_\phi=125$ GeV assuming 100% branching fraction of the long-lived scalar into fermion pairs.

Expected and observed 95% CL limits on ($\sigma / \sigma_{\text{SM}}) \times B$ for $m_\phi=125$ GeV assuming 100% branching fraction of the long-lived scalar into fermion pairs.

Expected and observed 95% CL limits on ($\sigma / \sigma_{\text{SM}}) \times B$ for $m_\phi=125$ GeV assuming 100% branching fraction of the long-lived scalar into fermion pairs.

Expected and observed 95% CL limits on $\sigma \times B$ for $m_\phi=60$ GeV assuming 100% branching fraction of the long-lived scalar into fermion pairs.

Expected and observed 95% CL limits on $\sigma \times B$ for $m_\phi=60$ GeV assuming 100% branching fraction of the long-lived scalar into fermion pairs.

Expected and observed 95% CL limits on $\sigma \times B$ for $m_\phi=200$ GeV assuming 100% branching fraction of the long-lived scalar into fermion pairs.

Expected and observed 95% CL limits on $\sigma \times B$ for $m_\phi=400$ GeV assuming 100% branching fraction of the long-lived scalar into fermion pairs.

Expected and observed 95% CL limits on $\sigma \times B$ for $m_\phi=600$ GeV assuming 100% branching fraction of the long-lived scalar into fermion pairs.

Expected and observed 95% CL limits on $\sigma \times B$ for $m_\phi=600$ GeV assuming 100% branching fraction of the long-lived scalar into fermion pairs.

Expected and observed 95% CL limits on $\sigma \times B$ for $m_\phi=600$ GeV assuming 100% branching fraction of the long-lived scalar into fermion pairs.

Expected and observed 95% CL limits on $\sigma \times B$ for $m_\phi=1000$ GeV assuming 100% branching fraction of the long-lived scalar into fermion pairs.

Expected and observed 95% CL limits on $\sigma \times B$ for $m_\phi=1000$ GeV assuming 100% branching fraction of the long-lived scalar into fermion pairs.

Expected and observed 95% CL limits on $\sigma \times B$ for $m_\phi=1000$ GeV assuming 100% branching fraction of the long-lived scalar into fermion pairs.

Cumulative acceptance times efficiency as a function of resonance mass for each event selection step in the resolved channel for the spin-2 signal models. The local maximum at 251 GeV is a consequence of the near-threshold kinematics.

Corrected $m(HH)$ distribution in the resolved $4b$ validation region (dots), compared with the reweighted distribution in $2b$ validation region (teal histogram). The error bars on the $4b$ points represent the Poisson uncertainties corresponding to their event yields. The final bin includes overflow. The background uncertainty (gray band) is computed by adding all individual components in quadrature. The bottom panel shows the difference between the $4b$ and reweighted $2b$ distributions, relative to the $2b$ distribution.

Corrected $m(HH)$ distribution in the resolved $4b$ validation region (dots), compared with the reweighted distribution in $2b$ validation region (teal histogram). The error bars on the $4b$ points represent the Poisson uncertainties corresponding to their event yields. The final bin includes overflow. The background uncertainty (gray band) is computed by adding all individual components in quadrature. The bottom panel shows the difference between the $4b$ and reweighted $2b$ distributions, relative to the $2b$ distribution.

Corrected $m(HH)$ distribution in the resolved $4b$ signal region (dots), after the fit under the background-only hypothesis. The error bars on the $4b$ points represent the Poisson uncertainties corresponding to their event yields. The background model (teal histogram) is shown with its total post-fit uncertainty (gray band). The final bin includes overflow. Representative spin-0 signal hypotheses (dashed, dotted, and dashed-dotted lines) are overlaid, normalized to the overall expected limits on their cross-sections. The bottom panel shows the difference between the $4b$ distribution and the background model, relative to the background model. No significant excess of data relative to the SM background is observed.

Corrected $m(HH)$ distribution in the resolved $4b$ signal region (dots), after the fit under the background-only hypothesis. The error bars on the $4b$ points represent the Poisson uncertainties corresponding to their event yields. The background model (teal histogram) is shown with its total post-fit uncertainty (gray band). The final bin includes overflow. Representative spin-0 signal hypotheses (dashed, dotted, and dashed-dotted lines) are overlaid, normalized to the overall expected limits on their cross-sections. The bottom panel shows the difference between the $4b$ distribution and the background model, relative to the background model. No significant excess of data relative to the SM background is observed.

Corrected $m(HH)$ distribution in the resolved $4b$ signal region (dots), after the fit under the background-only hypothesis. The error bars on the $4b$ points represent the Poisson uncertainties corresponding to their event yields. The background model (teal histogram) is shown with its total post-fit uncertainty (gray band). The final bin includes overflow. Representative spin-2 signal hypotheses (dashed, dotted, and dashed-dotted lines) are overlaid, normalized to the overall expected limits on their cross-sections. The bottom panel shows the difference between the $4b$ distribution and the background model, relative to the background model. No significant excess of data relative to the SM background is observed.

Corrected $m(HH)$ distribution in the resolved $4b$ signal region (dots), after the fit under the background-only hypothesis. The error bars on the $4b$ points represent the Poisson uncertainties corresponding to their event yields. The background model (teal histogram) is shown with its total post-fit uncertainty (gray band). The final bin includes overflow. Representative spin-2 signal hypotheses (dashed, dotted, and dashed-dotted lines) are overlaid, normalized to the overall expected limits on their cross-sections. The bottom panel shows the difference between the $4b$ distribution and the background model, relative to the background model. No significant excess of data relative to the SM background is observed.

Cumulative signal acceptance times efficiency as a function of the resonance mass for various selection steps in the boosted channel. The steps up to the $b$-tag categorization are shown for the spin-0.

Cumulative signal acceptance times efficiency as a function of the resonance mass for various selection steps in the boosted channel. The steps up to the $b$-tag categorization are shown for the spin-0.

Cumulative signal acceptance times efficiency as a function of the resonance mass for various selection steps in the boosted channel. The steps up to the $b$-tag categorization are shown for the spin-2.

Cumulative signal acceptance times efficiency as a function of the resonance mass for various selection steps in the boosted channel. The steps up to the $b$-tag categorization are shown for the spin-2.

Cumulative signal acceptance times efficiency as a function of the resonance mass for various selection steps in the boosted channel. The efficiencies of the three b-tag categories are shown for the spin-0 scenario; this efficiency is obtained after the other selection steps including the SR definition. The signal efficiency in the 4b region has a maximum around 1.5 TeV. Above that value the track-jets start to merge together, and for the highest resonance masses the 2b category becomes the most efficient.

Cumulative signal acceptance times efficiency as a function of the resonance mass for various selection steps in the boosted channel. The efficiencies of the three b-tag categories are shown for the spin-0 scenario; this efficiency is obtained after the other selection steps including the SR definition. The signal efficiency in the 4b region has a maximum around 1.5 TeV. Above that value the track-jets start to merge together, and for the highest resonance masses the 2b category becomes the most efficient.

Cumulative signal acceptance times efficiency as a function of the resonance mass for various selection steps in the boosted channel. The efficiencies of the three b-tag categories are shown for the spin-2 scenario; this efficiency is obtained after the other selection steps including the SR definition. The signal efficiency in the 4b region has a maximum around 1.5 TeV. Above that value the track-jets start to merge together, and for the highest resonance masses the 2b category becomes the most efficient.

Cumulative signal acceptance times efficiency as a function of the resonance mass for various selection steps in the boosted channel. The efficiencies of the three b-tag categories are shown for the spin-2 scenario; this efficiency is obtained after the other selection steps including the SR definition. The signal efficiency in the 4b region has a maximum around 1.5 TeV. Above that value the track-jets start to merge together, and for the highest resonance masses the 2b category becomes the most efficient.

Comparison of the background model (stacked histograms) with data (dots) in the $2b$ validation region. The error bars on the data points represent the Poisson uncertainties corresponding to their event yields. The background uncertainty (gray band) is computed by adding all individual components in quadrature and is not allowed to extend below zero.

Comparison of the background model (stacked histograms) with data (dots) in the $2b$ validation region. The error bars on the data points represent the Poisson uncertainties corresponding to their event yields. The background uncertainty (gray band) is computed by adding all individual components in quadrature and is not allowed to extend below zero.

Comparison of the background model (stacked histograms) with data (dots) in the $3b$ validation region. The error bars on the data points represent the Poisson uncertainties corresponding to their event yields. The background uncertainty (gray band) is computed by adding all individual components in quadrature and is not allowed to extend below zero.

Comparison of the background model (stacked histograms) with data (dots) in the $3b$ validation region. The error bars on the data points represent the Poisson uncertainties corresponding to their event yields. The background uncertainty (gray band) is computed by adding all individual components in quadrature and is not allowed to extend below zero.

Comparison of the background model (stacked histograms) with data (dots) in the $4b$ validation region. The error bars on the data points represent the Poisson uncertainties corresponding to their event yields. The background uncertainty (gray band) is computed by adding all individual components in quadrature and is not allowed to extend below zero.

Comparison of the background model (stacked histograms) with data (dots) in the $4b$ validation region. The error bars on the data points represent the Poisson uncertainties corresponding to their event yields. The background uncertainty (gray band) is computed by adding all individual components in quadrature and is not allowed to extend below zero.

The $m(HH)$ distributions in the boosted $2b$ signal regions (dots), after the fit under the background-only hypothesis. The error bars on the data points represent the Poisson uncertainties corresponding to their event yields. The background model (stacked histogram) is shown with its total post-fit uncertainty (gray band). The uncertainty bands are defined using an ensemble of curves constructed by sampling a multivariate Gaussian probability density function built from the covariance matrix of the fit. Representative spin-0 signal hypotheses (dashed and dashed-dotted lines) are overlaid, normalized to the overall expected limits on their cross-sections. No significant excess of data relative to the SM background is observed.

The $m(HH)$ distributions in the boosted $2b$ signal regions (dots), after the fit under the background-only hypothesis. The error bars on the data points represent the Poisson uncertainties corresponding to their event yields. The background model (stacked histogram) is shown with its total post-fit uncertainty (gray band). The uncertainty bands are defined using an ensemble of curves constructed by sampling a multivariate Gaussian probability density function built from the covariance matrix of the fit. Representative spin-0 signal hypotheses (dashed and dashed-dotted lines) are overlaid, normalized to the overall expected limits on their cross-sections. No significant excess of data relative to the SM background is observed.

The $m(HH)$ distributions in the boosted $2b$ signal regions (dots), after the fit under the background-only hypothesis. The error bars on the data points represent the Poisson uncertainties corresponding to their event yields. The background model (stacked histogram) is shown with its total post-fit uncertainty (gray band). The uncertainty bands are defined using an ensemble of curves constructed by sampling a multivariate Gaussian probability density function built from the covariance matrix of the fit. Representative spin-2 signal hypotheses (dashed and dashed-dotted lines) are overlaid, normalized to the overall expected limits on their cross-sections. No significant excess of data relative to the SM background is observed.

The $m(HH)$ distributions in the boosted $2b$ signal regions (dots), after the fit under the background-only hypothesis. The error bars on the data points represent the Poisson uncertainties corresponding to their event yields. The background model (stacked histogram) is shown with its total post-fit uncertainty (gray band). The uncertainty bands are defined using an ensemble of curves constructed by sampling a multivariate Gaussian probability density function built from the covariance matrix of the fit. Representative spin-2 signal hypotheses (dashed and dashed-dotted lines) are overlaid, normalized to the overall expected limits on their cross-sections. No significant excess of data relative to the SM background is observed.

The $m(HH)$ distributions in the boosted $3b$ signal regions (dots), after the fit under the background-only hypothesis. The error bars on the data points represent the Poisson uncertainties corresponding to their event yields. The background model (stacked histogram) is shown with its total post-fit uncertainty (gray band). The uncertainty bands are defined using an ensemble of curves constructed by sampling a multivariate Gaussian probability density function built from the covariance matrix of the fit. Representative spin-0 signal hypotheses (dashed and dashed-dotted lines) are overlaid, normalized to the overall expected limits on their cross-sections. No significant excess of data relative to the SM background is observed.

The $m(HH)$ distributions in the boosted $3b$ signal regions (dots), after the fit under the background-only hypothesis. The error bars on the data points represent the Poisson uncertainties corresponding to their event yields. The background model (stacked histogram) is shown with its total post-fit uncertainty (gray band). The uncertainty bands are defined using an ensemble of curves constructed by sampling a multivariate Gaussian probability density function built from the covariance matrix of the fit. Representative spin-0 signal hypotheses (dashed and dashed-dotted lines) are overlaid, normalized to the overall expected limits on their cross-sections. No significant excess of data relative to the SM background is observed.

The $m(HH)$ distributions in the boosted $3b$ signal regions (dots), after the fit under the background-only hypothesis. The error bars on the data points represent the Poisson uncertainties corresponding to their event yields. The background model (stacked histogram) is shown with its total post-fit uncertainty (gray band). The uncertainty bands are defined using an ensemble of curves constructed by sampling a multivariate Gaussian probability density function built from the covariance matrix of the fit. Representative spin-2 signal hypotheses (dashed and dashed-dotted lines) are overlaid, normalized to the overall expected limits on their cross-sections. No significant excess of data relative to the SM background is observed.

The $m(HH)$ distributions in the boosted $3b$ signal regions (dots), after the fit under the background-only hypothesis. The error bars on the data points represent the Poisson uncertainties corresponding to their event yields. The background model (stacked histogram) is shown with its total post-fit uncertainty (gray band). The uncertainty bands are defined using an ensemble of curves constructed by sampling a multivariate Gaussian probability density function built from the covariance matrix of the fit. Representative spin-2 signal hypotheses (dashed and dashed-dotted lines) are overlaid, normalized to the overall expected limits on their cross-sections. No significant excess of data relative to the SM background is observed.

The $m(HH)$ distributions in the boosted $4b$ signal regions (dots), after the fit under the background-only hypothesis. The error bars on the data points represent the Poisson uncertainties corresponding to their event yields. The background model (stacked histogram) is shown with its total post-fit uncertainty (gray band). The uncertainty bands are defined using an ensemble of curves constructed by sampling a multivariate Gaussian probability density function built from the covariance matrix of the fit. Representative spin-0 signal hypotheses (dashed and dashed-dotted lines) are overlaid, normalized to the overall expected limits on their cross-sections. No significant excess of data relative to the SM background is observed.

The $m(HH)$ distributions in the boosted $4b$ signal regions (dots), after the fit under the background-only hypothesis. The error bars on the data points represent the Poisson uncertainties corresponding to their event yields. The background model (stacked histogram) is shown with its total post-fit uncertainty (gray band). The uncertainty bands are defined using an ensemble of curves constructed by sampling a multivariate Gaussian probability density function built from the covariance matrix of the fit. Representative spin-0 signal hypotheses (dashed and dashed-dotted lines) are overlaid, normalized to the overall expected limits on their cross-sections. No significant excess of data relative to the SM background is observed.

The $m(HH)$ distributions in the boosted $4b$ signal regions (dots), after the fit under the background-only hypothesis. The error bars on the data points represent the Poisson uncertainties corresponding to their event yields. The background model (stacked histogram) is shown with its total post-fit uncertainty (gray band). The uncertainty bands are defined using an ensemble of curves constructed by sampling a multivariate Gaussian probability density function built from the covariance matrix of the fit. Representative spin-2 signal hypotheses (dashed and dashed-dotted lines) are overlaid, normalized to the overall expected limits on their cross-sections. No significant excess of data relative to the SM background is observed.

The $m(HH)$ distributions in the boosted $4b$ signal regions (dots), after the fit under the background-only hypothesis. The error bars on the data points represent the Poisson uncertainties corresponding to their event yields. The background model (stacked histogram) is shown with its total post-fit uncertainty (gray band). The uncertainty bands are defined using an ensemble of curves constructed by sampling a multivariate Gaussian probability density function built from the covariance matrix of the fit. Representative spin-2 signal hypotheses (dashed and dashed-dotted lines) are overlaid, normalized to the overall expected limits on their cross-sections. No significant excess of data relative to the SM background is observed.

Expected (dashed black lines) and observed (solid black lines) 95% CL upper limits on the cross-section of resonant $HH$ production in the spin-0 signal models. The $\pm 1 \sigma$ and $\pm 2 \sigma$ uncertainty ranges for the expected limits (colored bands) are shown. Expected limits using each of the resolved and boosted channels individually (dashed colored lines) are shown. The nominal $H\rightarrow b\bar{b}$ branching ratio is taken as 0.582.

Expected (dashed black lines) and observed (solid black lines) 95% CL upper limits on the cross-section of resonant $HH$ production in the spin-0 signal models. The $\pm 1 \sigma$ and $\pm 2 \sigma$ uncertainty ranges for the expected limits (colored bands) are shown. Expected limits using each of the resolved and boosted channels individually (dashed colored lines) are shown. The nominal $H\rightarrow b\bar{b}$ branching ratio is taken as 0.582.

Expected (dashed black lines) and observed (solid black lines) 95% CL upper limits on the cross-section of resonant $HH$ production in the spin-2 signal models. The $\pm 1 \sigma$ and $\pm 2 \sigma$ uncertainty ranges for the expected limits (colored bands) are shown. Expected limits using each of the resolved and boosted channels individually (dashed colored lines) are shown. The theoretical prediction for the bulk RS model with $k/\bar{M}_{\text{Pl}} = 1$ (solid red line) is shown; the decrease below 350 GeV is due to a sharp reduction in the $G^{*}_{\text{KK}} \rightarrow HH$ branching ratio. The nominal $H\rightarrow b\bar{b}$ branching ratio is taken as 0.582.

Expected (dashed black lines) and observed (solid black lines) 95% CL upper limits on the cross-section of resonant $HH$ production in the spin-2 signal models. The $\pm 1 \sigma$ and $\pm 2 \sigma$ uncertainty ranges for the expected limits (colored bands) are shown. Expected limits using each of the resolved and boosted channels individually (dashed colored lines) are shown. The theoretical prediction for the bulk RS model with $k/\bar{M}_{\text{Pl}} = 1$ (solid red line) is shown; the decrease below 350 GeV is due to a sharp reduction in the $G^{*}_{\text{KK}} \rightarrow HH$ branching ratio. The nominal $H\rightarrow b\bar{b}$ branching ratio is taken as 0.582.

The BDT distribution of the di-Higgs ggF signal for two different values of $\kappa_{\lambda}$ and the main backgrounds in the low mass region ($m^{*}_{b\bar{b}\gamma\gamma} < 350$ GeV). Distributions are normalized to unit area. The dotted lines denote the category boundaries. Events with a BDT score below 0.881 in the low mass region are discarded.

The BDT distribution (with x-axis zoomed in) of the di-Higgs ggF signal for two different values of $\kappa_{\lambda}$ and the main backgrounds in the low mass region ($m^{*}_{b\bar{b}\gamma\gamma} < 350$ GeV). Distributions are normalized to unit area. The dotted lines denote the category boundaries. Events with a BDT score below 0.881 in the low mass region are discarded. The range of BDT scores is from 0.8 to 1.

The BDT distribution (with x-axis zoomed in) of the di-Higgs ggF signal for two different values of $\kappa_{\lambda}$ and the main backgrounds in the low mass region ($m^{*}_{b\bar{b}\gamma\gamma} < 350$ GeV). Distributions are normalized to unit area. The dotted lines denote the category boundaries. Events with a BDT score below 0.881 in the low mass region are discarded. The range of BDT scores is from 0.8 to 1.

The BDT distribution (with x-axis zoomed in) of the di-Higgs ggF signal for two different values of $\kappa_{\lambda}$ and the main backgrounds in the low mass region ($m^{*}_{b\bar{b}\gamma\gamma} < 350$ GeV). Distributions are normalized to unit area. The dotted lines denote the category boundaries. Events with a BDT score below 0.881 in the low mass region are discarded. The range of BDT scores is from 0.8 to 1.

The BDT distribution (with x-axis zoomed in) of the di-Higgs ggF signal for two different values of $\kappa_{\lambda}$ and the main backgrounds in the low mass region ($m^{*}_{b\bar{b}\gamma\gamma} < 350$ GeV). Distributions are normalized to unit area. The dotted lines denote the category boundaries. Events with a BDT score below 0.881 in the low mass region are discarded. The range of BDT scores is from 0.8 to 1.

The BDT distribution of the di-Higgs ggF signal for two different values of $\kappa_{\lambda}$ and the main backgrounds in the high mass region ($m^{*}_{b\bar{b}\gamma\gamma} > 350$ GeV). Distributions are normalized to unit area. The dotted lines denote the category boundaries. Events with a BDT score below 0.857 in the high mass region are discarded.

The BDT distribution of the di-Higgs ggF signal for two different values of $\kappa_{\lambda}$ and the main backgrounds in the high mass region ($m^{*}_{b\bar{b}\gamma\gamma} > 350$ GeV). Distributions are normalized to unit area. The dotted lines denote the category boundaries. Events with a BDT score below 0.857 in the high mass region are discarded.

The BDT distribution of the di-Higgs ggF signal for two different values of $\kappa_{\lambda}$ and the main backgrounds in the high mass region ($m^{*}_{b\bar{b}\gamma\gamma} > 350$ GeV). Distributions are normalized to unit area. The dotted lines denote the category boundaries. Events with a BDT score below 0.857 in the high mass region are discarded.

The BDT distribution of the di-Higgs ggF signal for two different values of $\kappa_{\lambda}$ and the main backgrounds in the high mass region ($m^{*}_{b\bar{b}\gamma\gamma} > 350$ GeV). Distributions are normalized to unit area. The dotted lines denote the category boundaries. Events with a BDT score below 0.857 in the high mass region are discarded.

The BDT distribution (with x-axis zoomed in) of the di-Higgs ggF signal for two different values of $\kappa_{\lambda}$ and the main backgrounds in the high mass region ($m^{*}_{b\bar{b}\gamma\gamma} > 350$ GeV). Distributions are normalized to unit area. The dotted lines denote the category boundaries. Events with a BDT score below 0.857 in the high mass region are discarded. The range of BDT scores is from 0.8 to 1.

The BDT distribution (with x-axis zoomed in) of the di-Higgs ggF signal for two different values of $\kappa_{\lambda}$ and the main backgrounds in the high mass region ($m^{*}_{b\bar{b}\gamma\gamma} > 350$ GeV). Distributions are normalized to unit area. The dotted lines denote the category boundaries. Events with a BDT score below 0.857 in the high mass region are discarded. The range of BDT scores is from 0.8 to 1.

The BDT distribution (with x-axis zoomed in) of the di-Higgs ggF signal for two different values of $\kappa_{\lambda}$ and the main backgrounds in the high mass region ($m^{*}_{b\bar{b}\gamma\gamma} > 350$ GeV). Distributions are normalized to unit area. The dotted lines denote the category boundaries. Events with a BDT score below 0.857 in the high mass region are discarded. The range of BDT scores is from 0.8 to 1.

The BDT distribution (with x-axis zoomed in) of the di-Higgs ggF signal for two different values of $\kappa_{\lambda}$ and the main backgrounds in the high mass region ($m^{*}_{b\bar{b}\gamma\gamma} > 350$ GeV). Distributions are normalized to unit area. The dotted lines denote the category boundaries. Events with a BDT score below 0.857 in the high mass region are discarded. The range of BDT scores is from 0.8 to 1.

The $BDT_{tot}$ score for the benchmark signal $m_{X}$ = 300 GeV and for the main backgrounds. Distributions are normalized to unit area. The dotted line denotes the event selection threshold. Events with a $BDT_{tot}$ score below 0.85 for $m_{X}$ = 300 GeV are discarded.

The $BDT_{tot}$ score for the benchmark signal $m_{X}$ = 300 GeV and for the main backgrounds. Distributions are normalized to unit area. The dotted line denotes the event selection threshold. Events with a $BDT_{tot}$ score below 0.85 for $m_{X}$ = 300 GeV are discarded.

The $BDT_{tot}$ score for the benchmark signal $m_{X}$ = 300 GeV and for the main backgrounds. Distributions are normalized to unit area. The dotted line denotes the event selection threshold. Events with a $BDT_{tot}$ score below 0.85 for $m_{X}$ = 300 GeV are discarded.

The $BDT_{tot}$ score for the benchmark signal $m_{X}$ = 300 GeV and for the main backgrounds. Distributions are normalized to unit area. The dotted line denotes the event selection threshold. Events with a $BDT_{tot}$ score below 0.85 for $m_{X}$ = 300 GeV are discarded.

The $BDT_{tot}$ score for the benchmark signal $m_{X}$ = 500 GeV and for the main backgrounds. Distributions are normalized to unit area. The dotted line denotes the event selection threshold. Events with a $BDT_{tot}$ score below 0.75 for $m_{X}$ = 500 GeV are discarded.

The $BDT_{tot}$ score for the benchmark signal $m_{X}$ = 500 GeV and for the main backgrounds. Distributions are normalized to unit area. The dotted line denotes the event selection threshold. Events with a $BDT_{tot}$ score below 0.75 for $m_{X}$ = 500 GeV are discarded.

The $BDT_{tot}$ score for the benchmark signal $m_{X}$ = 500 GeV and for the main backgrounds. Distributions are normalized to unit area. The dotted line denotes the event selection threshold. Events with a $BDT_{tot}$ score below 0.75 for $m_{X}$ = 500 GeV are discarded.

The $BDT_{tot}$ score for the benchmark signal $m_{X}$ = 500 GeV and for the main backgrounds. Distributions are normalized to unit area. The dotted line denotes the event selection threshold. Events with a $BDT_{tot}$ score below 0.75 for $m_{X}$ = 500 GeV are discarded.

Distributions of $m_{\gamma\gamma}$ in high mass BDT tight category for the nonresonant $HH$ search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband.

Distributions of $m_{\gamma\gamma}$ in high mass BDT tight category for the nonresonant $HH$ search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband.

Distributions of $m_{\gamma\gamma}$ in high mass BDT tight category for the nonresonant $HH$ search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband.

Distributions of $m_{\gamma\gamma}$ in high mass BDT tight category for the nonresonant $HH$ search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband.

Distributions of $m_{\gamma\gamma}$ in high mass BDT loose category for the nonresonant $HH$ search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband.

Distributions of $m_{\gamma\gamma}$ in high mass BDT loose category for the nonresonant $HH$ search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband.

Distributions of $m_{\gamma\gamma}$ in high mass BDT loose category for the nonresonant $HH$ search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband.

Distributions of $m_{\gamma\gamma}$ in high mass BDT loose category for the nonresonant $HH$ search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband.

Distributions of $m_{\gamma\gamma}$ in low mass BDT tight category for the nonresonant $HH$ search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband.

Distributions of $m_{\gamma\gamma}$ in low mass BDT tight category for the nonresonant $HH$ search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband.

Distributions of $m_{\gamma\gamma}$ in low mass BDT tight category for the nonresonant $HH$ search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband.

Distributions of $m_{\gamma\gamma}$ in low mass BDT tight category for the nonresonant $HH$ search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband.

Distributions of $m_{\gamma\gamma}$ in low mass BDT loose category for the nonresonant $HH$ search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband.

Distributions of $m_{\gamma\gamma}$ in low mass BDT loose category for the nonresonant $HH$ search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband.

Distributions of $m_{\gamma\gamma}$ in low mass BDT loose category for the nonresonant $HH$ search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband.

Distributions of $m_{\gamma\gamma}$ in low mass BDT loose category for the nonresonant $HH$ search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband.

Distributions of $m_{\gamma\gamma}$ for the selections used for the resonance mass point $m_{X}$ = 300 GeV for the resonant search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband. The scalar resonance signal is scaled to a total production cross section $\sigma(pp \rightarrow X \rightarrow HH)$ = 370 fb.

Distributions of $m_{\gamma\gamma}$ for the selections used for the resonance mass point $m_{X}$ = 300 GeV for the resonant search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband. The scalar resonance signal is scaled to a total production cross section $\sigma(pp \rightarrow X \rightarrow HH)$ = 370 fb.

Distributions of $m_{\gamma\gamma}$ for the selections used for the resonance mass point $m_{X}$ = 300 GeV for the resonant search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband. The scalar resonance signal is scaled to a total production cross section $\sigma(pp \rightarrow X \rightarrow HH)$ = 370 fb.

Distributions of $m_{\gamma\gamma}$ for the selections used for the resonance mass point $m_{X}$ = 300 GeV for the resonant search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband. The scalar resonance signal is scaled to a total production cross section $\sigma(pp \rightarrow X \rightarrow HH)$ = 370 fb.

Distributions of $m_{\gamma\gamma}$ for the selections used for the resonance mass point $m_{X}$ = 500 GeV for the resonant search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband. The scalar resonance signal is scaled to a total production cross section $\sigma(pp \rightarrow X \rightarrow HH)$ = 67 fb.

Distributions of $m_{\gamma\gamma}$ for the selections used for the resonance mass point $m_{X}$ = 500 GeV for the resonant search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband. The scalar resonance signal is scaled to a total production cross section $\sigma(pp \rightarrow X \rightarrow HH)$ = 67 fb.

Distributions of $m_{\gamma\gamma}$ for the selections used for the resonance mass point $m_{X}$ = 500 GeV for the resonant search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband. The scalar resonance signal is scaled to a total production cross section $\sigma(pp \rightarrow X \rightarrow HH)$ = 67 fb.

Distributions of $m_{\gamma\gamma}$ for the selections used for the resonance mass point $m_{X}$ = 500 GeV for the resonant search. The data-derived fractions of nonresonant $\gamma\gamma$, $\gamma$-jet or jet-$\gamma$, and dijet background are applied and the total background is normalized to the data sideband. The scalar resonance signal is scaled to a total production cross section $\sigma(pp \rightarrow X \rightarrow HH)$ = 67 fb.

The number of data events observed in the 120 GeV < $m_{\gamma\gamma}$ < 130 GeV window, the number of $HH$ signal events expected for $\kappa_{\lambda}$ = 1 and for $\kappa_{\lambda}$ = 10, and events expected for single Higgs boson production (estimated using MC simulation), as well as for continuum background. For the single Higgs boson, Rest includes VBF, $WH$, $tHqb$, and $tHW$. The values are obtained from a fit of the Asimov data set generated under the SM signal-plus-background hypothesis, $\kappa_{\lambda}$ = 1. The continuum background component of the Asimov data set is obtained from the fit of the data sideband. The uncertainties in $HH$ signals and single Higgs boson background include the systematic uncertainties discussed in Section 6. The uncertainty in the continuum background is given by the sum in quadrature of the statistical uncertainty from the fit to the data and the spurious-signal uncertainty.

The number of data events observed in the 120 GeV < $m_{\gamma\gamma}$ < 130 GeV window, the number of $HH$ signal events expected for $\kappa_{\lambda}$ = 1 and for $\kappa_{\lambda}$ = 10, and events expected for single Higgs boson production (estimated using MC simulation), as well as for continuum background. For the single Higgs boson, Rest includes VBF, $WH$, $tHqb$, and $tHW$. The values are obtained from a fit of the Asimov data set generated under the SM signal-plus-background hypothesis, $\kappa_{\lambda}$ = 1. The continuum background component of the Asimov data set is obtained from the fit of the data sideband. The uncertainties in $HH$ signals and single Higgs boson background include the systematic uncertainties discussed in Section 6. The uncertainty in the continuum background is given by the sum in quadrature of the statistical uncertainty from the fit to the data and the spurious-signal uncertainty.

The number of data events observed in the 120 GeV < $m_{\gamma\gamma}$ < 130 GeV window, the number of $HH$ signal events expected for $\kappa_{\lambda}$ = 1 and for $\kappa_{\lambda}$ = 10, and events expected for single Higgs boson production (estimated using MC simulation), as well as for continuum background. For the single Higgs boson, Rest includes VBF, $WH$, $tHqb$, and $tHW$. The values are obtained from a fit of the Asimov data set generated under the SM signal-plus-background hypothesis, $\kappa_{\lambda}$ = 1. The continuum background component of the Asimov data set is obtained from the fit of the data sideband. The uncertainties in $HH$ signals and single Higgs boson background include the systematic uncertainties discussed in Section 6. The uncertainty in the continuum background is given by the sum in quadrature of the statistical uncertainty from the fit to the data and the spurious-signal uncertainty.

The number of data events observed in the 120 GeV < $m_{\gamma\gamma}$ < 130 GeV window, the number of $HH$ signal events expected for $\kappa_{\lambda}$ = 1 and for $\kappa_{\lambda}$ = 10, and events expected for single Higgs boson production (estimated using MC simulation), as well as for continuum background. For the single Higgs boson, Rest includes VBF, $WH$, $tHqb$, and $tHW$. The values are obtained from a fit of the Asimov data set generated under the SM signal-plus-background hypothesis, $\kappa_{\lambda}$ = 1. The continuum background component of the Asimov data set is obtained from the fit of the data sideband. The uncertainties in $HH$ signals and single Higgs boson background include the systematic uncertainties discussed in Section 6. The uncertainty in the continuum background is given by the sum in quadrature of the statistical uncertainty from the fit to the data and the spurious-signal uncertainty.

Observed and expected limits at 95% CL on the cross section of nonresonant Higgs boson pair production as a function of the Higgs boson self-coupling modifier $\kappa_{\lambda}= \lambda_{HHH}/\lambda^{\textrm{SM}}_{HHH}$. The expected constraints on $\kappa_{\lambda}$ are obtained with a background hypothesis excluding $pp \rightarrow HH$ production. The $\pm 1\sigma$ and $\pm 2\sigma$ variations about the expected limit due to statistical and systematic uncertainties are also shown. The theory prediction curve represents the scenario where all parameters and couplings are set to their SM values except for $\kappa_{\lambda}$. The uncertainty band of the theory prediction curve shows the cross-section uncertainty.

Observed and expected limits at 95% CL on the cross section of nonresonant Higgs boson pair production as a function of the Higgs boson self-coupling modifier $\kappa_{\lambda}= \lambda_{HHH}/\lambda^{\textrm{SM}}_{HHH}$. The expected constraints on $\kappa_{\lambda}$ are obtained with a background hypothesis excluding $pp \rightarrow HH$ production. The $\pm 1\sigma$ and $\pm 2\sigma$ variations about the expected limit due to statistical and systematic uncertainties are also shown. The theory prediction curve represents the scenario where all parameters and couplings are set to their SM values except for $\kappa_{\lambda}$. The uncertainty band of the theory prediction curve shows the cross-section uncertainty.

Observed and expected limits at 95% CL on the cross section of nonresonant Higgs boson pair production as a function of the Higgs boson self-coupling modifier $\kappa_{\lambda}= \lambda_{HHH}/\lambda^{\textrm{SM}}_{HHH}$. The expected constraints on $\kappa_{\lambda}$ are obtained with a background hypothesis excluding $pp \rightarrow HH$ production. The $\pm 1\sigma$ and $\pm 2\sigma$ variations about the expected limit due to statistical and systematic uncertainties are also shown. The theory prediction curve represents the scenario where all parameters and couplings are set to their SM values except for $\kappa_{\lambda}$. The uncertainty band of the theory prediction curve shows the cross-section uncertainty.

Observed and expected limits at 95% CL on the cross section of nonresonant Higgs boson pair production as a function of the Higgs boson self-coupling modifier $\kappa_{\lambda}= \lambda_{HHH}/\lambda^{\textrm{SM}}_{HHH}$. The expected constraints on $\kappa_{\lambda}$ are obtained with a background hypothesis excluding $pp \rightarrow HH$ production. The $\pm 1\sigma$ and $\pm 2\sigma$ variations about the expected limit due to statistical and systematic uncertainties are also shown. The theory prediction curve represents the scenario where all parameters and couplings are set to their SM values except for $\kappa_{\lambda}$. The uncertainty band of the theory prediction curve shows the cross-section uncertainty.

Values of the negative log-profile-likelihood ratio ($-2ln\Lambda$) as a function of $\kappa_{\lambda}$ evaluated for the combination of all the categories of the nonresonant search. The coupling of the Higgs boson to fermions and gauge bosons is set to SM values in the profile likelihood calculation. The Asimov data set is generated under the SM signal-plus-background hypothesis, $\kappa_{\lambda}$= 1. All systematic uncertainties, including the theoretical uncertainties in the di-Higgs boson production cross section, are included. The intersections of the solid curves and the horizontal dashed lines indicate the 1$\sigma$ and 2$\sigma$ confidence-level intervals.

Values of the negative log-profile-likelihood ratio ($-2ln\Lambda$) as a function of $\kappa_{\lambda}$ evaluated for the combination of all the categories of the nonresonant search. The coupling of the Higgs boson to fermions and gauge bosons is set to SM values in the profile likelihood calculation. The Asimov data set is generated under the SM signal-plus-background hypothesis, $\kappa_{\lambda}$= 1. All systematic uncertainties, including the theoretical uncertainties in the di-Higgs boson production cross section, are included. The intersections of the solid curves and the horizontal dashed lines indicate the 1$\sigma$ and 2$\sigma$ confidence-level intervals.

Values of the negative log-profile-likelihood ratio ($-2ln\Lambda$) as a function of $\kappa_{\lambda}$ evaluated for the combination of all the categories of the nonresonant search. The coupling of the Higgs boson to fermions and gauge bosons is set to SM values in the profile likelihood calculation. The Asimov data set is generated under the SM signal-plus-background hypothesis, $\kappa_{\lambda}$= 1. All systematic uncertainties, including the theoretical uncertainties in the di-Higgs boson production cross section, are included. The intersections of the solid curves and the horizontal dashed lines indicate the 1$\sigma$ and 2$\sigma$ confidence-level intervals.

Values of the negative log-profile-likelihood ratio ($-2ln\Lambda$) as a function of $\kappa_{\lambda}$ evaluated for the combination of all the categories of the nonresonant search. The coupling of the Higgs boson to fermions and gauge bosons is set to SM values in the profile likelihood calculation. The Asimov data set is generated under the SM signal-plus-background hypothesis, $\kappa_{\lambda}$= 1. All systematic uncertainties, including the theoretical uncertainties in the di-Higgs boson production cross section, are included. The intersections of the solid curves and the horizontal dashed lines indicate the 1$\sigma$ and 2$\sigma$ confidence-level intervals.

The number of events observed in the 120 < $m_{\gamma\gamma}$ < 130 GeV window in data, the number of events expected for scalar resonance signals of masses $m_{X}$ = 300 GeV and $m_{X}$ = 500 GeV assuming a total production cross section $\sigma(pp \rightarrow X \rightarrow HH)$ equal to the observed exclusion limits of Figure 15, and events expected for SM $HH$ and single Higgs boson production (estimated using MC simulation), as well as for continuum background. The values are obtained from a fit of the Asimov data set generated under the signal-plus-background hypothesis. The continuum background component of the Asimov data set is obtained from the fit of the data sideband. The uncertainties in the resonant signals and the SM $HH$ and single-Higgs-boson backgrounds include the systematic uncertainties discussed in Section 6. The uncertainty in the continuum background is given by the sum in quadrature of the statistical uncertainty from the fit to the data and the spurious-signal uncertainty.

The number of events observed in the 120 < $m_{\gamma\gamma}$ < 130 GeV window in data, the number of events expected for scalar resonance signals of masses $m_{X}$ = 300 GeV and $m_{X}$ = 500 GeV assuming a total production cross section $\sigma(pp \rightarrow X \rightarrow HH)$ equal to the observed exclusion limits of Figure 15, and events expected for SM $HH$ and single Higgs boson production (estimated using MC simulation), as well as for continuum background. The values are obtained from a fit of the Asimov data set generated under the signal-plus-background hypothesis. The continuum background component of the Asimov data set is obtained from the fit of the data sideband. The uncertainties in the resonant signals and the SM $HH$ and single-Higgs-boson backgrounds include the systematic uncertainties discussed in Section 6. The uncertainty in the continuum background is given by the sum in quadrature of the statistical uncertainty from the fit to the data and the spurious-signal uncertainty.

The number of events observed in the 120 < $m_{\gamma\gamma}$ < 130 GeV window in data, the number of events expected for scalar resonance signals of masses $m_{X}$ = 300 GeV and $m_{X}$ = 500 GeV assuming a total production cross section $\sigma(pp \rightarrow X \rightarrow HH)$ equal to the observed exclusion limits of Figure 15, and events expected for SM $HH$ and single Higgs boson production (estimated using MC simulation), as well as for continuum background. The values are obtained from a fit of the Asimov data set generated under the signal-plus-background hypothesis. The continuum background component of the Asimov data set is obtained from the fit of the data sideband. The uncertainties in the resonant signals and the SM $HH$ and single-Higgs-boson backgrounds include the systematic uncertainties discussed in Section 6. The uncertainty in the continuum background is given by the sum in quadrature of the statistical uncertainty from the fit to the data and the spurious-signal uncertainty.

The number of events observed in the 120 < $m_{\gamma\gamma}$ < 130 GeV window in data, the number of events expected for scalar resonance signals of masses $m_{X}$ = 300 GeV and $m_{X}$ = 500 GeV assuming a total production cross section $\sigma(pp \rightarrow X \rightarrow HH)$ equal to the observed exclusion limits of Figure 15, and events expected for SM $HH$ and single Higgs boson production (estimated using MC simulation), as well as for continuum background. The values are obtained from a fit of the Asimov data set generated under the signal-plus-background hypothesis. The continuum background component of the Asimov data set is obtained from the fit of the data sideband. The uncertainties in the resonant signals and the SM $HH$ and single-Higgs-boson backgrounds include the systematic uncertainties discussed in Section 6. The uncertainty in the continuum background is given by the sum in quadrature of the statistical uncertainty from the fit to the data and the spurious-signal uncertainty.

Observed and expected limits at 95% CL on the production cross section of a narrow-width scalar resonance $X$ as a function of the mass $m_{X}$ of the hypothetical scalar particle. The black solid line represents the observed upper limits. The dashed line represents the expected upper limits. The $\pm 1\sigma$ and $\pm 2\sigma$ variations about the expected limit due to statistical and systematic uncertainties are also shown.

Observed and expected limits at 95% CL on the production cross section of a narrow-width scalar resonance $X$ as a function of the mass $m_{X}$ of the hypothetical scalar particle. The black solid line represents the observed upper limits. The dashed line represents the expected upper limits. The $\pm 1\sigma$ and $\pm 2\sigma$ variations about the expected limit due to statistical and systematic uncertainties are also shown.

Observed and expected limits at 95% CL on the production cross section of a narrow-width scalar resonance $X$ as a function of the mass $m_{X}$ of the hypothetical scalar particle. The black solid line represents the observed upper limits. The dashed line represents the expected upper limits. The $\pm 1\sigma$ and $\pm 2\sigma$ variations about the expected limit due to statistical and systematic uncertainties are also shown.

Observed and expected limits at 95% CL on the production cross section of a narrow-width scalar resonance $X$ as a function of the mass $m_{X}$ of the hypothetical scalar particle. The black solid line represents the observed upper limits. The dashed line represents the expected upper limits. The $\pm 1\sigma$ and $\pm 2\sigma$ variations about the expected limit due to statistical and systematic uncertainties are also shown.

Breakdown of the dominant systematic uncertainties. The impact of the uncertainties is defined according to the statistical analysis described in Section 7. It corresponds to the relative variation of the expected upper limit on the cross section when re-evaluating the profile likelihood ratio after fixing the nuisance parameter in question to its best-fit value, while all remaining nuisance parameters remain free to float. The impact is shown in %. Only systematic uncertainties with an impact of at least 0.2% are shown. Uncertainties of the "Norm. + Shape" type affect both the normalization and the parameters of the functional form. The rest of the uncertainties affect only the yields.

Breakdown of the dominant systematic uncertainties. The impact of the uncertainties is defined according to the statistical analysis described in Section 7. It corresponds to the relative variation of the expected upper limit on the cross section when re-evaluating the profile likelihood ratio after fixing the nuisance parameter in question to its best-fit value, while all remaining nuisance parameters remain free to float. The impact is shown in %. Only systematic uncertainties with an impact of at least 0.2% are shown. Uncertainties of the "Norm. + Shape" type affect both the normalization and the parameters of the functional form. The rest of the uncertainties affect only the yields.

Breakdown of the dominant systematic uncertainties. The impact of the uncertainties is defined according to the statistical analysis described in Section 7. It corresponds to the relative variation of the expected upper limit on the cross section when re-evaluating the profile likelihood ratio after fixing the nuisance parameter in question to its best-fit value, while all remaining nuisance parameters remain free to float. The impact is shown in %. Only systematic uncertainties with an impact of at least 0.2% are shown. Uncertainties of the "Norm. + Shape" type affect both the normalization and the parameters of the functional form. The rest of the uncertainties affect only the yields.

Breakdown of the dominant systematic uncertainties. The impact of the uncertainties is defined according to the statistical analysis described in Section 7. It corresponds to the relative variation of the expected upper limit on the cross section when re-evaluating the profile likelihood ratio after fixing the nuisance parameter in question to its best-fit value, while all remaining nuisance parameters remain free to float. The impact is shown in %. Only systematic uncertainties with an impact of at least 0.2% are shown. Uncertainties of the "Norm. + Shape" type affect both the normalization and the parameters of the functional form. The rest of the uncertainties affect only the yields.

Cutflow for nonresonant di-Higgs ggF signal sample, yields are normalized to 139 $fb^{-1}$.

Cutflow for nonresonant di-Higgs ggF signal sample, yields are normalized to 139 $fb^{-1}$.

Cutflow for nonresonant di-Higgs ggF signal sample, yields are normalized to 139 $fb^{-1}$.

Cutflow for nonresonant di-Higgs ggF signal sample, yields are normalized to 139 $fb^{-1}$.

Cutflow for resonant signal sample, with $m_{X}$ = 300 GeV, yields are normalized to 139 $fb^{-1}$.

Cutflow for resonant signal sample, with $m_{X}$ = 300 GeV, yields are normalized to 139 $fb^{-1}$.

Cutflow for resonant signal sample, with $m_{X}$ = 300 GeV, yields are normalized to 139 $fb^{-1}$.

Cutflow for resonant signal sample, with $m_{X}$ = 300 GeV, yields are normalized to 139 $fb^{-1}$.

Cutflow for resonant signal sample, with $m_{X}$ = 500 GeV, yields are normalized to 139 $fb^{-1}$.

Cutflow for resonant signal sample, with $m_{X}$ = 500 GeV, yields are normalized to 139 $fb^{-1}$.

Cutflow for resonant signal sample, with $m_{X}$ = 500 GeV, yields are normalized to 139 $fb^{-1}$.

Cutflow for resonant signal sample, with $m_{X}$ = 500 GeV, yields are normalized to 139 $fb^{-1}$.

Comparison of $m_{b\bar{b}}$ distributions when applying the specific b-jet energy calibration and the nominal jet energy calibration. The distributions are fitted using a Bukin function, and the values of the peak position, resolution and the relative improvement are reported in the legend.

Comparison of $m_{b\bar{b}}$ distributions when applying the specific b-jet energy calibration and the nominal jet energy calibration. The distributions are fitted using a Bukin function, and the values of the peak position, resolution and the relative improvement are reported in the legend.

Comparison of $m_{b\bar{b}}$ distributions when applying the specific b-jet energy calibration and the nominal jet energy calibration. The distributions are fitted using a Bukin function, and the values of the peak position, resolution and the relative improvement are reported in the legend.

Comparison of $m_{b\bar{b}}$ distributions when applying the specific b-jet energy calibration and the nominal jet energy calibration. The distributions are fitted using a Bukin function, and the values of the peak position, resolution and the relative improvement are reported in the legend.

Fit results of $m_{b\bar{b}}$ distributions when applying the specific b-jet energy calibration and the nominal jet energy calibration. The distributions are fitted using a Bukin function, and the values of the peak position, resolution and the relative improvement are reported in the legend.

Fit results of $m_{b\bar{b}}$ distributions when applying the specific b-jet energy calibration and the nominal jet energy calibration. The distributions are fitted using a Bukin function, and the values of the peak position, resolution and the relative improvement are reported in the legend.

Fit results of $m_{b\bar{b}}$ distributions when applying the specific b-jet energy calibration and the nominal jet energy calibration. The distributions are fitted using a Bukin function, and the values of the peak position, resolution and the relative improvement are reported in the legend.

Fit results of $m_{b\bar{b}}$ distributions when applying the specific b-jet energy calibration and the nominal jet energy calibration. The distributions are fitted using a Bukin function, and the values of the peak position, resolution and the relative improvement are reported in the legend.

The relative amount (purity) of expected events from SM $HH$ and single Higgs boson production processes for each of the four categories of the nonresonant search. The Higgs boson pair production with $k_{\lambda} = 1$ is considered as signal in (a), while the case with $k_{\lambda} = 10$ is considered as signal in (b).

The relative amount (purity) of expected events from SM $HH$ and single Higgs boson production processes for each of the four categories of the nonresonant search. The Higgs boson pair production with $k_{\lambda} = 1$ is considered as signal in (a), while the case with $k_{\lambda} = 10$ is considered as signal in (b).

The relative amount (purity) of expected events from SM $HH$ and single Higgs boson production processes for each of the four categories of the nonresonant search. The Higgs boson pair production with $k_{\lambda} = 1$ is considered as signal in (a), while the case with $k_{\lambda} = 10$ is considered as signal in (b).

The relative amount (purity) of expected events from SM $HH$ and single Higgs boson production processes for each of the four categories of the nonresonant search. The Higgs boson pair production with $k_{\lambda} = 1$ is considered as signal in (a), while the case with $k_{\lambda} = 10$ is considered as signal in (b).

The expected significance in each of the four categories of the nonresonant search. The Higgs boson pair production with $k_{\lambda} = 1$ is considered as signal in (a), while the case with $k_{\lambda} = 10$ is considered as signal in (b). The single Higgs boson processes and the di-photon continuum spectrum are considered as background.

The expected significance in each of the four categories of the nonresonant search. The Higgs boson pair production with $k_{\lambda} = 1$ is considered as signal in (a), while the case with $k_{\lambda} = 10$ is considered as signal in (b). The single Higgs boson processes and the di-photon continuum spectrum are considered as background.

The expected significance in each of the four categories of the nonresonant search. The Higgs boson pair production with $k_{\lambda} = 1$ is considered as signal in (a), while the case with $k_{\lambda} = 10$ is considered as signal in (b). The single Higgs boson processes and the di-photon continuum spectrum are considered as background.

The expected significance in each of the four categories of the nonresonant search. The Higgs boson pair production with $k_{\lambda} = 1$ is considered as signal in (a), while the case with $k_{\lambda} = 10$ is considered as signal in (b). The single Higgs boson processes and the di-photon continuum spectrum are considered as background.

Spurious signal result for the exponential pdf for the various ggF nonresonant di-Higgs categories. In each category, the spurious signal value ($n_{sp}$) and its ratio to the expected statistical error ($Z_{spur}$) from data are shown.

Spurious signal result for the exponential pdf for the various ggF nonresonant di-Higgs categories. In each category, the spurious signal value ($n_{sp}$) and its ratio to the expected statistical error ($Z_{spur}$) from data are shown.

Spurious signal result for the exponential pdf for the various ggF nonresonant di-Higgs categories. In each category, the spurious signal value ($n_{sp}$) and its ratio to the expected statistical error ($Z_{spur}$) from data are shown.

Spurious signal result for the exponential pdf for the various ggF nonresonant di-Higgs categories. In each category, the spurious signal value ($n_{sp}$) and its ratio to the expected statistical error ($Z_{spur}$) from data are shown.

Spurious signal result for the exponential pdf as function of the resonant di-Higgs signal mass. The spurious signal value and its ratio to the expected statistical error from data are shown.

Spurious signal result for the exponential pdf as function of the resonant di-Higgs signal mass. The spurious signal value and its ratio to the expected statistical error from data are shown.

Spurious signal result for the exponential pdf as function of the resonant di-Higgs signal mass. The spurious signal value and its ratio to the expected statistical error from data are shown.

Spurious signal result for the exponential pdf as function of the resonant di-Higgs signal mass. The spurious signal value and its ratio to the expected statistical error from data are shown.

Distributions of the signal efficiency as a function of $\kappa_{\lambda}$, for the di-Higgs boson ggF nonresonant production mode. The range of $\kappa_{\lambda}$ in the table is from -10 to 10.

Distributions of the signal efficiency as a function of $\kappa_{\lambda}$, for the di-Higgs boson ggF nonresonant production mode. The range of $\kappa_{\lambda}$ in the table is from -10 to 10.

Distributions of the signal efficiency as a function of $\kappa_{\lambda}$, for the di-Higgs boson ggF nonresonant production mode. The range of $\kappa_{\lambda}$ in the table is from -10 to 10.

Distributions of the signal efficiency as a function of $\kappa_{\lambda}$, for the di-Higgs boson ggF nonresonant production mode. The range of $\kappa_{\lambda}$ in the table is from -10 to 10.

Distributions of the signal efficiency as a function of $\kappa_{\lambda}$, for the di-Higgs boson VBF nonresonant production mode. The range of $\kappa_{\lambda}$ in the table is from -10 to 10.

Distributions of the signal efficiency as a function of $\kappa_{\lambda}$, for the di-Higgs boson VBF nonresonant production mode. The range of $\kappa_{\lambda}$ in the table is from -10 to 10.

Distributions of the signal efficiency as a function of $\kappa_{\lambda}$, for the di-Higgs boson VBF nonresonant production mode. The range of $\kappa_{\lambda}$ in the table is from -10 to 10.

Distributions of the signal efficiency as a function of $\kappa_{\lambda}$, for the di-Higgs boson VBF nonresonant production mode. The range of $\kappa_{\lambda}$ in the table is from -10 to 10.

Values of the negative log-profile-likelihood ($-2ln\Lambda$) as a function of $\kappa_{\lambda}$ evaluated for the combination of all the categories of the nonresonant search. The coupling of the Higgs boson to fermions and gauge bosons is set to SM values in the profile likelihood calculation. The Asimov data set is generated under the SM signal-plus-background hypothesis, $\kappa_{\lambda}$= 1. All systematic uncertainties, including the theoretical uncertainties on the di-Higgs boson production cross section, are included. The intersections of the solid curves and the horizontal dashed lines indicate the 1$\sigma$ and 2$\sigma$ confidence level intervals. The best fit value corresponds to $\kappa_{\lambda}$ = $2.8^{+2.0}_{-2.2}(^{+3.8}_{-4.3})$ for the 1$\sigma$(2$\sigma$) confidence interval. The expected value corresponds to $\kappa_{\lambda}$ = $1.0^{+5.5}_{-2.4}(^{+7.3}_{-4.2})$ for the 1$\sigma$(2$\sigma$) confidence interval. The dashed curves represent values of the negative log-profile-likelihood where the Higgs boson branching fractions and the cross section of the production modes are varied as a function of $\kappa_{\lambda}$. In this case,the best fit value corresponds to $\kappa_{\lambda}$ = $2.7^{+2.0}_{-2.2}(^{+3.8}_{-4.3})$ and the expected value corresponds to $\kappa_{\lambda}$ = $1.0^{+5.4}_{-2.5}(^{+7.3}_{-4.3})$ for the 1$\sigma$(2$\sigma$) confidence interval.

Values of the negative log-profile-likelihood ($-2ln\Lambda$) as a function of $\kappa_{\lambda}$ evaluated for the combination of all the categories of the nonresonant search. The coupling of the Higgs boson to fermions and gauge bosons is set to SM values in the profile likelihood calculation. The Asimov data set is generated under the SM signal-plus-background hypothesis, $\kappa_{\lambda}$= 1. All systematic uncertainties, including the theoretical uncertainties on the di-Higgs boson production cross section, are included. The intersections of the solid curves and the horizontal dashed lines indicate the 1$\sigma$ and 2$\sigma$ confidence level intervals. The best fit value corresponds to $\kappa_{\lambda}$ = $2.8^{+2.0}_{-2.2}(^{+3.8}_{-4.3})$ for the 1$\sigma$(2$\sigma$) confidence interval. The expected value corresponds to $\kappa_{\lambda}$ = $1.0^{+5.5}_{-2.4}(^{+7.3}_{-4.2})$ for the 1$\sigma$(2$\sigma$) confidence interval. The dashed curves represent values of the negative log-profile-likelihood where the Higgs boson branching fractions and the cross section of the production modes are varied as a function of $\kappa_{\lambda}$. In this case,the best fit value corresponds to $\kappa_{\lambda}$ = $2.7^{+2.0}_{-2.2}(^{+3.8}_{-4.3})$ and the expected value corresponds to $\kappa_{\lambda}$ = $1.0^{+5.4}_{-2.5}(^{+7.3}_{-4.3})$ for the 1$\sigma$(2$\sigma$) confidence interval.

Values of the negative log-profile-likelihood ($-2ln\Lambda$) as a function of $\kappa_{\lambda}$ evaluated for the combination of all the categories of the nonresonant search. The coupling of the Higgs boson to fermions and gauge bosons is set to SM values in the profile likelihood calculation. The Asimov data set is generated under the SM signal-plus-background hypothesis, $\kappa_{\lambda}$= 1. All systematic uncertainties, including the theoretical uncertainties on the di-Higgs boson production cross section, are included. The intersections of the solid curves and the horizontal dashed lines indicate the 1$\sigma$ and 2$\sigma$ confidence level intervals. The best fit value corresponds to $\kappa_{\lambda}$ = $2.8^{+2.0}_{-2.2}(^{+3.8}_{-4.3})$ for the 1$\sigma$(2$\sigma$) confidence interval. The expected value corresponds to $\kappa_{\lambda}$ = $1.0^{+5.5}_{-2.4}(^{+7.3}_{-4.2})$ for the 1$\sigma$(2$\sigma$) confidence interval. The dashed curves represent values of the negative log-profile-likelihood where the Higgs boson branching fractions and the cross section of the production modes are varied as a function of $\kappa_{\lambda}$. In this case,the best fit value corresponds to $\kappa_{\lambda}$ = $2.7^{+2.0}_{-2.2}(^{+3.8}_{-4.3})$ and the expected value corresponds to $\kappa_{\lambda}$ = $1.0^{+5.4}_{-2.5}(^{+7.3}_{-4.3})$ for the 1$\sigma$(2$\sigma$) confidence interval.

Values of the negative log-profile-likelihood ($-2ln\Lambda$) as a function of $\kappa_{\lambda}$ evaluated for the combination of all the categories of the nonresonant search. The coupling of the Higgs boson to fermions and gauge bosons is set to SM values in the profile likelihood calculation. The Asimov data set is generated under the SM signal-plus-background hypothesis, $\kappa_{\lambda}$= 1. All systematic uncertainties, including the theoretical uncertainties on the di-Higgs boson production cross section, are included. The intersections of the solid curves and the horizontal dashed lines indicate the 1$\sigma$ and 2$\sigma$ confidence level intervals. The best fit value corresponds to $\kappa_{\lambda}$ = $2.8^{+2.0}_{-2.2}(^{+3.8}_{-4.3})$ for the 1$\sigma$(2$\sigma$) confidence interval. The expected value corresponds to $\kappa_{\lambda}$ = $1.0^{+5.5}_{-2.4}(^{+7.3}_{-4.2})$ for the 1$\sigma$(2$\sigma$) confidence interval. The dashed curves represent values of the negative log-profile-likelihood where the Higgs boson branching fractions and the cross section of the production modes are varied as a function of $\kappa_{\lambda}$. In this case,the best fit value corresponds to $\kappa_{\lambda}$ = $2.7^{+2.0}_{-2.2}(^{+3.8}_{-4.3})$ and the expected value corresponds to $\kappa_{\lambda}$ = $1.0^{+5.4}_{-2.5}(^{+7.3}_{-4.3})$ for the 1$\sigma$(2$\sigma$) confidence interval.

Minimum BDT value of the events passing the selection criteria of the resonant search. The combined BDT score is formed using as coefficients $C_{1}$ = 0.65 and $C_{2}$ = 1 − $C_{1}$. The selection efficiency for the resonant $X \rightarrow HH$ signal is also shown.

Minimum BDT value of the events passing the selection criteria of the resonant search. The combined BDT score is formed using as coefficients $C_{1}$ = 0.65 and $C_{2}$ = 1 − $C_{1}$. The selection efficiency for the resonant $X \rightarrow HH$ signal is also shown.

Minimum BDT value of the events passing the selection criteria of the resonant search. The combined BDT score is formed using as coefficients $C_{1}$ = 0.65 and $C_{2}$ = 1 − $C_{1}$. The selection efficiency for the resonant $X \rightarrow HH$ signal is also shown.

Minimum BDT value of the events passing the selection criteria of the resonant search. The combined BDT score is formed using as coefficients $C_{1}$ = 0.65 and $C_{2}$ = 1 − $C_{1}$. The selection efficiency for the resonant $X \rightarrow HH$ signal is also shown.

The efficiency for simulated VBF events to pass each analysis cut.

The efficiency for simulated VH events to pass each analysis cut.

The efficiency for simulated ttH events to pass each analysis cut.

The fractional contribution of each production mode to a given analysis region around the Higgs boson peak, defined as 105 < mJ < 140 GeV. The fraction is given with respect to the total signal yield in the analysis region in question.

Signal acceptance times efficiency within the fiducial volume used in the fiducial region.

Signal acceptance times efficiency for the STXS volumes in the differential measurement. Along with the pTH requirements shown, |yH | < 2 is required. For events with pTH < 300 GeV, the acceptance times efficiency is less than 0.1 × 10−2.

Event yields and associated uncertainties after the global likelihood fit in the pT(H) > 450 GeV fiducial region. The total background yield can differ from the sum of the individual components due to rounding.

Event yields and associated uncertainties after the global likelihood fit of the differential regions. The total background yield can differ from the sum of the individual components due to rounding. The five STXS volumes are labeled pT0–pT4 corresponding to pTH < 300 GeV, 300 – 450 GeV, 450 – 650 GeV, 650 – 1000 GeV, and > 1000 GeV, respectively. The pT0 event yield is constrained to its SM value within the theoretical and experimental uncertainties and free parameters act independently on the remaining four volumes.

The JER for R = 0.2 jets in Pb+Pb collisions as a function of $2|\Psi_2-\phi^{reco}|$ for centrality selections of 0-5%, 5-10%, 10-20%, 20-40% and 40-60%.

The systematic uncertainties in v2 for 20-40% centrality Pb+Pb collisions as a function of pT.

The systematic uncertainties in v2 for 5-10% centrality Pb+Pb collisions as a function of pT.

The systematic uncertainties in v3 for 20-40% centrality Pb+Pb collisions as a function of pT.

The systematic uncertainties in v3 for 5-10% centrality Pb+Pb collisions as a function of pT.

The systematic uncertainties in v4 for 20-40% centrality Pb+Pb collisions as a function of pT.

The systematic uncertainties in v4 for 5-10% centrality Pb+Pb collisions as a function of pT.

The systematic uncertainties in v2 for pT = 71--398 GeV jets as a function of centrality.

The systematic uncertainties in v3 for pT = 71--398 GeV jets as a function of centrality.

The systematic uncertainties in v4 for pT = 71--398 GeV jets as a function of centrality.

Angular distribution of jets as a function of the observed psi2 plane for jets with 71 < pT < 79 GeV in the 10-20% centrality bin.

Angular distribution of jets as a function of the observed psi3 plane for jets with 71 < pT < 79 GeV in the 10-20% centrality bin.

Angular distribution of jets as a function of the observed psi4 plane for jets with 71 < pT < 79 GeV in the 10-20% centrality bin.

The v2 values for R = 0.2 jets as a function of centrality for jets in several pT ranges.

The v2 values for R = 0.2 jets as a function of pT for 0-5%, 5-10%, and 20-40% centrality collisions.

The v2, v3, and v4 as a function of centrality for jets with pT = 71-398 GeV.

The v3 values for R = 0.2 jets as a function of centrality for jets in several pT ranges.

The v4 values for R = 0.2 jets as a function of centrality for jets in several pT ranges.

The v2 as a function of pT for jets in 10-20% centrality collisions.

The v3 as a function of pT for jets in 10-20% centrality collisions.

The v2 as a function of pT for jets in 20-40% centrality collisions.

The v3 as a function of pT for jets in 20-40% centrality collisions.

The v2 for jets in 10-20% centrality collisions.

The v3 for jets in 10-20% centrality collisions.

R2max as a function of pT (filled circles). Also shown is 1 - 4v2/(1+2v2).

R3max as a function of pT (filled circles). Also shown is 1 - 4v3/(1+2v3).

The systematic uncertainties in v2 for 40-60% centrality Pb+Pb collisions as a function of pT.

The systematic uncertainties in v3 for 40-60% centrality Pb+Pb collisions as a function of pT.

The systematic uncertainties in v4 for 40-60% centrality Pb+Pb collisions as a function of pT.

The systematic uncertainties in v2 for 10-20% centrality Pb+Pb collisions as a function of pT.

The systematic uncertainties in v3 for 10-20% centrality Pb+Pb collisions as a function of pT.

The systematic uncertainties in v4 for 10-20% centrality Pb+Pb collisions as a function of pT.

The systematic uncertainties in v2 for 0-5% centrality Pb+Pb collisions as a function of pT.

The systematic uncertainties in v3 for 0-5% centrality Pb+Pb collisions as a function of pT.

The systematic uncertainties in v4 for 0-5% centrality Pb+Pb collisions as a function of pT.

The v2 as a function of pT for jets in 40-60% centrality collisions.

The v2 as a function of pT for jets in 5-10% centrality collisions.

The v2 as a function of pT for jets in 0-5% centrality collisions.

The v3 as a function of pT for jets in 40-60% centrality collisions.

The v3 as a function of pT for jets in 5-10% centrality collisions.

The v3 as a function of pT for jets in 0-5% centrality collisions.

Correlation coefficients $\rho_{i,j}$ for the statistical and systematic uncertainties between the $i$-th and $j$-th bin of the differential $A_E$ measurement as a function of the jet scattering angle $\theta_j$

The effects on the energy asymmetry of $1\sigma$ variations in its influencing nuisance parameters for the three $\theta_j$ bins. These are extracted from the samples of the posterior distribution with $\sigma_i^{(j)} = c_{ij}/\sqrt{c_{jj}}$ being the estimated shift of bin $i$ in conjunction with a shift $\Delta\theta_j$ of nuisance parameter $j$. The data statistical (Data stat.) uncertainty is obtained from running the unfolding with all nuisance parameters being fixed to their post-marginalised values, the MC statistical uncertainty on the response matrix ($t\bar{t}$ response MC stat.) is evaluated using a bootstrapping method from the covariance matrix of the ensemble of repeated unfolding results with varied response matrices. The $\gamma$ variations denote the statistical uncertainties of the background predictions in the corresponding bin of the $\Delta E$ vs $\theta_{j}$ distribution. The numbers appended to the $W$+jets PDF variations denote the corresponding NNPDF3.0 PDF sets.

The effects on the energy asymmetry of $1\sigma$ variations in its influencing nuisance parameters for the three $\theta_j$ bins. These are extracted from the samples of the posterior distribution with $\sigma_i^{(j)} = c_{ij}/\sqrt{c_{jj}}$ being the estimated shift of bin $i$ in conjunction with a shift $\Delta\theta_j$ of nuisance parameter $j$. The data statistical (Data stat.) uncertainty is obtained from running the unfolding with all nuisance parameters being fixed to their post-marginalised values, the MC statistical uncertainty on the response matrix ($t\bar{t}$ response MC stat.) is evaluated using a bootstrapping method from the covariance matrix of the ensemble of repeated unfolding results with varied response matrices. The $\gamma$ variations denote the statistical uncertainties of the background predictions in the corresponding bin of the $\Delta E$ vs $\theta_{j}$ distribution. The numbers appended to the $W$+jets PDF variations denote the corresponding NNPDF3.0 PDF sets.

Covariances $c_{ij}$ for the highest ranked systematic uncertainties in the energy asymmetry measurement differential in $\theta_j$.

Covariances $c_{ij}$ for the highest ranked systematic uncertainties in the energy asymmetry measurement differential in $\theta_j$.

Definition of the fiducial phase space with lepton candidate ($\ell$), hadronic top candidate ($j_h$), leptonic top candidate ($j_l$) and associated jet candidate ($j_a$).

Definition of the fiducial phase space with lepton candidate ($\ell$), hadronic top candidate ($j_h$), leptonic top candidate ($j_l$) and associated jet candidate ($j_a$).

Bounds on individual Wilson coefficients $C($TeV$/\Lambda)^2$ from the energy asymmetry, obtained from a combined fit to its values in all three $\theta_j$ bins.

Bounds on individual Wilson coefficients $C($TeV$/\Lambda)^2$ from the energy asymmetry, obtained from a combined fit to its values in all three $\theta_j$ bins.

Upper limits on $\mathcal{B}(H\rightarrow aa\rightarrow bb\mu\mu)$ at 95% CL including the BDT cut as a function of the signal mass.

Upper limits on $\mathcal{B}(H\rightarrow aa\rightarrow bb\mu\mu)$ at 95% CL without the BDT cut as a function of the signal mass.

The number of events when applying each cut in the analysis for $m_a$ = 16 GeV. The first row shows the approximate total number of events scaled to the theoretical signal cross-section and luminosity of the full Run 2 dataset, assuming $ \mathcal{B}(H\rightarrow aa\rightarrow bb\mu\mu)$ = 0.16%.

The number of events when applying each cut in the analysis for $m_a$ = 20 GeV. The first row shows the approximate total number of events scaled to the theoretical signal cross-section and luminosity of the full Run 2 dataset, assuming $ \mathcal{B}(H\rightarrow aa\rightarrow bb\mu\mu)$ = 0.16%.

The number of events when applying each cut in the analysis for $m_a$ = 30 GeV. The first row shows the approximate total number of events scaled to the theoretical signal cross-section and luminosity of the full Run 2 dataset, assuming $ \mathcal{B}(H\rightarrow aa\rightarrow bb\mu\mu)$ = 0.16%.

The number of events when applying each cut in the analysis for $m_a$ = 40 GeV. The first row shows the approximate total number of events scaled to the theoretical signal cross-section and luminosity of the full Run 2 dataset, assuming $ \mathcal{B}(H\rightarrow aa\rightarrow bb\mu\mu)$ = 0.16%.

The number of events when applying each cut in the analysis for $m_a$ = 50 GeV. The first row shows the approximate total number of events scaled to the theoretical signal cross-section and luminosity of the full Run 2 dataset, assuming $ \mathcal{B}(H\rightarrow aa\rightarrow bb\mu\mu)$ = 0.16%.

The number of events when applying each cut in the analysis for $m_a$ = 60 GeV. The first row shows the approximate total number of events scaled to the theoretical signal cross-section and luminosity of the full Run 2 dataset, assuming $ \mathcal{B}(H\rightarrow aa\rightarrow bb\mu\mu)$ = 0.16%.

Post-fit $m_\text{T}(\gamma, E_\text{T}^\text{miss})$ distribution in the inclusive signal region for the dark-photon search with the 125 GeV mass $\mathcal{B}(H \to \gamma \gamma_\text{d})$ signal normalization set to zero. A $H \to \gamma \gamma_\text{d}$ decay signal is shown for two different mass hypotheses, 125 GeV and 500 GeV, and scaled to a $\mathcal{B}(H \to \gamma \gamma_\text{d})$ of 2% and 1%, respectively. Events with $m_\text{T}(\gamma, E_\text{T}^\text{miss})$ larger than the rightmost bin boundary are added to that bin.

The 95% CL upper limit on the Higgs boson production cross-section times branching ratio to $\gamma \gamma_\text{d}$ is shown for different VBF-produced scalar-mediator-mass hypotheses in the NWA. The theoretically predicted cross-section of a Higgs boson produced via VBF and with the $\mathcal{B}(H \to \gamma \gamma_\text{d}) =$ 5% is superimposed on the $\pm 1\sigma$ and $\pm 2\sigma$ NNLO QCD + NLO EW uncertainty band of the expected production cross-section limit.

Post-fit $m_\text{jj}$ distribution in the inclusive signal region. The Higgs boson invisible decay signal is scaled to a $\mathcal{B}_\text{inv}$ of 37%. Events with $m_\text{jj}$ larger than the rightmost bin boundary are added to that bin.

Post-fit $m_\text{jj}$ distribution in the one-lepton control region $W_{\ell \nu}^\gamma$ CR. Events with $m_\text{jj}$ larger than the rightmost bin boundary are added to that bin.

Post-fit $m_\text{T}$ distribution in the one lepton control region. Events with $m_\text{T}$ larger than the rightmost bin boundary are added to that bin.

Post-fit photon centrality distribution in the zero lepton signal plus control region with the $\mathcal{B}_\text{inv}$ signal normalization set to zero in the fit.

Post-fit photon $E_\text{T}$ distribution in the zero lepton signal region with the $\mathcal{B}_\text{inv}$ signal normalization set to zero in the fit.

Post-fit photon centrality distribution in the zero lepton signal plus control region resulting from the fit to the $m_\text{jj}$ distribution for EW $Z \gamma + \text{jets}$. The post-fit uncertainties include statistical, experimental, and theory contributions.

Post-fit photon $E_\text{T}$ distribution in the zero lepton signal region resulting from the fit to the $m_\text{jj}$ distribution for EW $Z \gamma + \text{jets}$. The post-fit uncertainties include statistical, experimental, and theory contributions.

Post-fit DNN output score distribution in the one lepton control region.

Yields for the EW $Z \gamma + \text{jets}$ process are shown after each selection along with relative and absolute signal acceptance efficiencies.

Yields for the 125 GeV Higgs boson with $\mathcal{B}_\text{inv.} =$ 1 signal produced by the vector boson fusion process in association with a final state photon are shown after each selection along with relative and absolute signal acceptance efficiencies.

Yields for the 125 GeV Higgs boson with $\mathcal{B}(H \to \gamma \gamma_\text{d}) =$ 1 signal produced by the vector boson fusion process are shown after each selection along with relative and absolute signal acceptance efficiencies.

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.

Distributions of $s_{\text{T}}$ in the single-tau one-bin SR. The stacked histograms show the various SM background contributions. The hatched band indicates the total statistical and systematic uncertainty of the SM background. The $t\bar{t}$ (2 real $\tau$) and $t\bar{t}$ (1 real $\tau$) as well as the single-top background contributions are scaled with the normalization factors obtained from the background-only fit. Minor backgrounds are grouped together and denoted as 'Other'. This includes $t\bar{t}$-fake, single top, and other top (di-tau channel) or $t\bar{t}$-fake, $t\bar{t}+H$, multiboson, and other top (single-tau channel). The overlaid dotted lines show the additional contributions for signal scenarios close to the expected exclusion contour with the particle type and the mass and $\beta$ parameters for the simplified models indicated in the legend. For the leptoquark signal model the shapes of the distributions for $\text{LQ}_{3}^{\text{d}}$ and $\text{LQ}_{3}^{\text{v}}$ (not shown) are similar to that of $\text{LQ}_{3}^{\text{u}}$. The rightmost bin includes the overflow.

Distributions of $m_{\text{T}}(\tau)$ in the single-tau one-bin SR. The stacked histograms show the various SM background contributions. The hatched band indicates the total statistical and systematic uncertainty of the SM background. The $t\bar{t}$ (2 real $\tau$) and $t\bar{t}$ (1 real $\tau$) as well as the single-top background contributions are scaled with the normalization factors obtained from the background-only fit. Minor backgrounds are grouped together and denoted as 'Other'. This includes $t\bar{t}$-fake, single top, and other top (di-tau channel) or $t\bar{t}$-fake, $t\bar{t}+H$, multiboson, and other top (single-tau channel). The overlaid dotted lines show the additional contributions for signal scenarios close to the expected exclusion contour with the particle type and the mass and $\beta$ parameters for the simplified models indicated in the legend. For the leptoquark signal model the shapes of the distributions for $\text{LQ}_{3}^{\text{d}}$ and $\text{LQ}_{3}^{\text{v}}$ (not shown) are similar to that of $\text{LQ}_{3}^{\text{u}}$. The rightmost bin includes the overflow.

Distributions of $\Sigma m_{\text{T}}(b_{1,2})$ in the single-tau $p_{\text{T}}(\tau)$-binned SR. The stacked histograms show the various SM background contributions. The hatched band indicates the total statistical and systematic uncertainty of the SM background. The $t\bar{t}$ (2 real $\tau$) and $t\bar{t}$ (1 real $\tau$) as well as the single-top background contributions are scaled with the normalization factors obtained from the background-only fit. Minor backgrounds are grouped together and denoted as 'Other'. This includes $t\bar{t}$-fake, single top, and other top (di-tau channel) or $t\bar{t}$-fake, $t\bar{t}+H$, multiboson, and other top (single-tau channel). The overlaid dotted lines show the additional contributions for signal scenarios close to the expected exclusion contour with the particle type and the mass and $\beta$ parameters for the simplified models indicated in the legend. For the leptoquark signal model the shapes of the distributions for $\text{LQ}_{3}^{\text{d}}$ and $\text{LQ}_{3}^{\text{v}}$ (not shown) are similar to that of $\text{LQ}_{3}^{\text{u}}$. The rightmost bin includes the overflow.

Distributions of $p_{\text{T}}(\tau)$ in the single-tau $p_{\text{T}}(\tau)$-binned SR. The stacked histograms show the various SM background contributions. The hatched band indicates the total statistical and systematic uncertainty of the SM background. The $t\bar{t}$ (2 real $\tau$) and $t\bar{t}$ (1 real $\tau$) as well as the single-top background contributions are scaled with the normalization factors obtained from the background-only fit. Minor backgrounds are grouped together and denoted as 'Other'. This includes $t\bar{t}$-fake, single top, and other top (di-tau channel) or $t\bar{t}$-fake, $t\bar{t}+H$, multiboson, and other top (single-tau channel). The overlaid dotted lines show the additional contributions for signal scenarios close to the expected exclusion contour with the particle type and the mass and $\beta$ parameters for the simplified models indicated in the legend. For the leptoquark signal model the shapes of the distributions for $\text{LQ}_{3}^{\text{d}}$ and $\text{LQ}_{3}^{\text{v}}$ (not shown) are similar to that of $\text{LQ}_{3}^{\text{u}}$. The rightmost bin includes the overflow.

Observed event yields in data ('Observed') and expected event yields for SM background processes obtained from the background-only fit ('Total bkg.' and rows below) in the signal regions of the di-tau and single-tau channels. The quoted uncertainties include both the statistical and systematic uncertainties and are truncated at zero yield. By construction, no $t\bar{t}$ (2 real $\tau$) events can pass the selections in the single-tau channel. As the individual uncertainties are correlated, they do not add in quadrature to equal the total background uncertainty.

From left to right: upper limits at the 95% confidence level (CL) on the visible cross section ($\sigma_\text{vis}$) and on the number of signal events ($S_{\text{obs}}^{95}$). The third column ($S_{\text{exp}}^{95}$) shows the upper limit at the 95% CL on the number of signal events, given the expected number (and $\pm 1\,\sigma$ excursions on the expectation) of background events. The last two columns indicate the confidence level observed for the background-only hypothesis ($\text{CL}_{b}$), the discovery $p$-value ($p(s=0)$) and the significance ($Z$). In the di-tau SR, where fewer events are observed than predicted by the fitted background estimate, the $p$-value is capped at 0.5.

Expected and observed exclusion contours at the 95% confidence level for the vector third-generation leptoquark signal model, as a function of the mass $m(\text{LQ}_{3}^{\text{v}})$ and the branching fraction $B(\text{LQ}_{3}^{\text{v}} \rightarrow b\tau)$ into a quark and a charged lepton. The plot shows the exclusion contour for the minimal-coupling scenario. The limits are derived from the binned single-tau signal region.

Expected and observed exclusion contours at the 95% confidence level for the vector third-generation leptoquark signal model, as a function of the mass $m(\text{LQ}_{3}^{\text{v}})$ and the branching fraction $B(\text{LQ}_{3}^{\text{v}} \rightarrow b\tau)$ into a quark and a charged lepton. The plot shows the exclusion contour for the minimal-coupling scenario. The limits are derived from the binned single-tau signal region.

Expected and observed exclusion contours at the 95% confidence level for the vector third-generation leptoquark signal model, as a function of the mass $m(\text{LQ}_{3}^{\text{v}})$ and the branching fraction $B(\text{LQ}_{3}^{\text{v}} \rightarrow b\tau)$ into a quark and a charged lepton. The plot shows the exclusion contour for vector leptoquarks with additional gauge couplings. The limits are derived from the binned single-tau signal region.

Expected and observed exclusion contours at the 95% confidence level for the vector third-generation leptoquark signal model, as a function of the mass $m(\text{LQ}_{3}^{\text{v}})$ and the branching fraction $B(\text{LQ}_{3}^{\text{v}} \rightarrow b\tau)$ into a quark and a charged lepton. The plot shows the exclusion contour for vector leptoquarks with additional gauge couplings. The limits are derived from the binned single-tau signal region.

Exclusion contours at the 95% confidence level for the stop-stau signal model as a function of the masses of the top squark $m(\tilde{t}_{1})$ and of the tau slepton $m(\tilde{\tau}_{1})$. Expected and observed limits are shown for the present search in comparison to observed limits from previous ATLAS analyses based on data from Run-1 of the LHC at $\sqrt{s} = 8$ TeV [Eur. Phys. J. C 76 (2016)] and on a partial dataset from Run 2 at $\sqrt{s} = 13$ TeV [Phys. Rev. D 98 (2018) 032008]. The green band indicates the limit on the mass of the tau slepton (for a massless LSP) from the LEP experiments.

Exclusion contours at the 95% confidence level for the stop-stau signal model as a function of the masses of the top squark $m(\tilde{t}_{1})$ and of the tau slepton $m(\tilde{\tau}_{1})$. Expected and observed limits are shown for the present search in comparison to observed limits from previous ATLAS analyses based on data from Run-1 of the LHC at $\sqrt{s} = 8$ TeV [Eur. Phys. J. C 76 (2016)] and on a partial dataset from Run 2 at $\sqrt{s} = 13$ TeV [Phys. Rev. D 98 (2018) 032008]. The green band indicates the limit on the mass of the tau slepton (for a massless LSP) from the LEP experiments.

Expected and observed exclusion contours at the 95% confidence level for the scalar third-generation leptoquark signal model, as a function of the mass $m(\text{LQ}_{3}^{\text{u}})$ and the branching fraction $B(\text{LQ}_{3}^{\text{u}} \rightarrow q\ell)$ into a quark and a charged lepton. The plot shows the exclusion contour for up-type leptoquarks $\text{LQ}_{3}^{\text{u}})$ with charge $+2/3e$. The limits are derived from the binned single-tau signal region. Shown in gray for comparison are the observed exclusion-limit contours from the previous ATLAS publication that targets the same leptoquark models but is based on a subset of the Run-2 data [JHEP 06 (2019) 144]. In this previous publication five different analyses are considered that target not only the final state studied here but also the final states that correspond to a branching fraction $B(\text{LQ}_{3}^{\text{u}} \rightarrow q\ell)$ of 0 or 1, leading to the concave shapes of the gray exclusion contours.

Expected and observed exclusion contours at the 95% confidence level for the scalar third-generation leptoquark signal model, as a function of the mass $m(\text{LQ}_{3}^{\text{u}})$ and the branching fraction $B(\text{LQ}_{3}^{\text{u}} \rightarrow q\ell)$ into a quark and a charged lepton. The plot shows the exclusion contour for up-type leptoquarks $\text{LQ}_{3}^{\text{u}})$ with charge $+2/3e$. The limits are derived from the binned single-tau signal region. Shown in gray for comparison are the observed exclusion-limit contours from the previous ATLAS publication that targets the same leptoquark models but is based on a subset of the Run-2 data [JHEP 06 (2019) 144]. In this previous publication five different analyses are considered that target not only the final state studied here but also the final states that correspond to a branching fraction $B(\text{LQ}_{3}^{\text{u}} \rightarrow q\ell)$ of 0 or 1, leading to the concave shapes of the gray exclusion contours.

Expected and observed exclusion contours at the 95% confidence level for the scalar third-generation leptoquark signal model, as a function of the mass $m(\text{LQ}_{3}^{\text{d}})$ and the branching fraction $B(\text{LQ}_{3}^{\text{d}} \rightarrow q\ell)$ into a quark and a charged lepton. The plot shows the exclusion contour for down-type leptoquarks $\text{LQ}_{3}^{\text{d}})$ with charge $-1/3e$. The limits are derived from the binned single-tau signal region. Shown in gray for comparison are the observed exclusion-limit contours from the previous ATLAS publication that targets the same leptoquark models but is based on a subset of the Run-2 data [JHEP 06 (2019) 144]. In this previous publication five different analyses are considered that target not only the final state studied here but also the final states that correspond to a branching fraction $B(\text{LQ}_{3}^{\text{d}} \rightarrow q\ell)$ of 0 or 1, leading to the concave shapes of the gray exclusion contours.

Expected and observed exclusion contours at the 95% confidence level for the scalar third-generation leptoquark signal model, as a function of the mass $m(\text{LQ}_{3}^{\text{d}})$ and the branching fraction $B(\text{LQ}_{3}^{\text{d}} \rightarrow q\ell)$ into a quark and a charged lepton. The plot shows the exclusion contour for down-type leptoquarks $\text{LQ}_{3}^{\text{d}})$ with charge $-1/3e$. The limits are derived from the binned single-tau signal region. Shown in gray for comparison are the observed exclusion-limit contours from the previous ATLAS publication that targets the same leptoquark models but is based on a subset of the Run-2 data [JHEP 06 (2019) 144]. In this previous publication five different analyses are considered that target not only the final state studied here but also the final states that correspond to a branching fraction $B(\text{LQ}_{3}^{\text{d}} \rightarrow q\ell)$ of 0 or 1, leading to the concave shapes of the gray exclusion contours.

Upper limits on the signal cross section at the 95 % confidence level for the stop-stau signal model.

Upper limits on the signal cross section at the 95 % confidence level for the scalar third-generation leptoquark signal model with up-type leptoquarks.

Upper limits on the signal cross section at the 95 % confidence level for the scalar third-generation leptoquark signal model with down-type leptoquarks.

Upper limits on the signal cross section at the 95 % confidence level for the vector third-generation leptoquark signal model with minimal coupling (MC).

Upper limits on the signal cross section at the 95 % confidence level for the vector third-generation leptoquark signal model with additional gauge couplings (YM).

Acceptance of the one-bin signal region of the single-tau channel for pair production of up-type leptoquarks $\text{LQ}_{3}^{\text{u}}$.

Efficiency of the one-bin signal region of the single-tau channel for pair production of up-type leptoquarks $\text{LQ}_{3}^{\text{u}}$. The plot does not show efficiencies for a branching fraction $B(\text{LQ}_{3}^{\text{u}} \rightarrow b\tau)$ of 0 or 1 because here the acceptance at generator level becomes zero and the efficiency is thus undefined.

Acceptance of the first bin of the multi-bin signal region (50 GeV $< p_{\text{T}}(\tau) <$ 100 GeV) of the single-tau channel for pair production of up-type leptoquarks $\text{LQ}_{3}^{\text{u}}$.

Efficiency of the first bin of the multi-bin signal region (50 GeV $< p_{\text{T}}(\tau) <$ 100 GeV) of the single-tau channel for pair production of up-type leptoquarks $\text{LQ}_{3}^{\text{u}}$. The plot does not show efficiencies for a branching fraction $B(\text{LQ}_{3}^{\text{u}} \rightarrow b\tau)$ of 0 or 1 because here the acceptance at generator level becomes zero and the efficiency is thus undefined.

Acceptance of the middle bin of the multi-bin signal region (100 GeV $< p_{\text{T}}(\tau) <$ 200 GeV) of the single-tau channel for pair production of up-type leptoquarks $\text{LQ}_{3}^{\text{u}}$.

Efficiency of the middle bin of the multi-bin signal region (100 GeV $< p_{\text{T}}(\tau) <$ 200 GeV) of the single-tau channel for pair production of up-type leptoquarks $\text{LQ}_{3}^{\text{u}}$. The plot does not show efficiencies for a branching fraction $B(\text{LQ}_{3}^{\text{u}} \rightarrow b\tau)$ of 0 or 1 because here the acceptance at generator level becomes zero and the efficiency is thus undefined.

Acceptance of the last bin of the multi-bin signal region (200 GeV $< p_{\text{T}}(\tau)$) of the single-tau channel for pair production of up-type leptoquarks $\text{LQ}_{3}^{\text{u}}$.

Efficiency of the last bin of the multi-bin signal region (200 GeV $< p_{\text{T}}(\tau)$) of the single-tau channel for pair production of up-type leptoquarks $\text{LQ}_{3}^{\text{u}}$. The plot does not show efficiencies for a branching fraction $B(\text{LQ}_{3}^{\text{u}} \rightarrow b\tau)$ of 0 or 1 because here the acceptance at generator level becomes zero and the efficiency is thus undefined.

Acceptance of the signal region of the di-tau channel for pair production of up-type leptoquarks $\text{LQ}_{3}^{\text{u}}$.

Efficiency of the signal region of the di-tau channel for pair production of up-type leptoquarks $\text{LQ}_{3}^{\text{u}}$. The plot does not show efficiencies for a branching fraction $B(\text{LQ}_{3}^{\text{u}} \rightarrow b\tau)$ of 0 because here the acceptance at generator level becomes zero and the efficiency is thus undefined.

Acceptance of the one-bin signal region of the single-tau channel for pair production of down-type leptoquarks $\text{LQ}_{3}^{\text{d}}$.

Efficiency of the one-bin signal region of the single-tau channel for pair production of down-type leptoquarks $\text{LQ}_{3}^{\text{d}}$. The plot does not show efficiencies for a branching fraction $B(\text{LQ}_{3}^{\text{d}} \rightarrow t\tau)$ of 0 or 1 because here the acceptance at generator level becomes zero and the efficiency is thus undefined.

Acceptance of the first bin of the multi-bin signal region (50 GeV $< p_{\text{T}}(\tau) <$ 100 GeV) of the single-tau channel for pair production of down-type leptoquarks $\text{LQ}_{3}^{\text{d}}$.

Efficiency of the first bin of the multi-bin signal region (50 GeV $< p_{\text{T}}(\tau) <$ 100 GeV) of the single-tau channel for pair production of down-type leptoquarks $\text{LQ}_{3}^{\text{d}}$. The plot does not show efficiencies for a branching fraction $B(\text{LQ}_{3}^{\text{d}} \rightarrow t\tau)$ of 0 or 1 because here the acceptance at generator level becomes zero and the efficiency is thus undefined.

Acceptance of the middle bin of the multi-bin signal region (100 GeV $< p_{\text{T}}(\tau) <$ 200 GeV) of the single-tau channel for pair production of down-type leptoquarks $\text{LQ}_{3}^{\text{d}}$.

Efficiency of the middle bin of the multi-bin signal region (100 GeV $< p_{\text{T}}(\tau) <$ 200 GeV) of the single-tau channel for pair production of down-type leptoquarks $\text{LQ}_{3}^{\text{d}}$. The plot does not show efficiencies for a branching fraction $B(\text{LQ}_{3}^{\text{d}} \rightarrow t\tau)$ of 0 or 1 because here the acceptance at generator level becomes zero and the efficiency is thus undefined.

Acceptance of the last bin of the multi-bin signal region (200 GeV $< p_{\text{T}}(\tau)$) of the single-tau channel for pair production of down-type leptoquarks $\text{LQ}_{3}^{\text{d}}$.

Efficiency of the last bin of the multi-bin signal region (200 GeV $< p_{\text{T}}(\tau)$) of the single-tau channel for pair production of down-type leptoquarks $\text{LQ}_{3}^{\text{d}}$. The plot does not show efficiencies for a branching fraction $B(\text{LQ}_{3}^{\text{d}} \rightarrow t\tau)$ of 0 or 1 because here the acceptance at generator level becomes zero and the efficiency is thus undefined.

Acceptance of the signal region of the di-tau channel for pair production of down-type leptoquarks $\text{LQ}_{3}^{\text{d}}$.

Efficiency of the signal region of the di-tau channel for pair production of down-type leptoquarks $\text{LQ}_{3}^{\text{d}}$. The plot does not show efficiencies for a branching fraction $B(\text{LQ}_{3}^{\text{d}} \rightarrow t\tau)$ of 0 because here the acceptance at generator level becomes zero and the efficiency is thus undefined.

Acceptance of the one-bin signal region of the single-tau channel for pair production of vector leptoquarks $\text{LQ}_{3}^{\text{v}}$ in the minimal-coupling scenario.

Efficiency of the one-bin signal region of the single-tau channel for pair production of vector leptoquarks $\text{LQ}_{3}^{\text{v}}$ in the minimal-coupling scenario. The plot does not show efficiencies for a branching fraction $B(\text{LQ}_{3}^{\text{v}} \rightarrow b\tau)$ of 0 or 1 because here the acceptance at generator level becomes zero and the efficiency is thus undefined.

Acceptance of the first bin of the multi-bin signal region (50 GeV $< p_{\text{T}}(\tau) <$ 100 GeV) of the single-tau channel for pair production of vector leptoquarks $\text{LQ}_{3}^{\text{v}}$ in the minimal-coupling scenario.

Efficiency of the first bin of the multi-bin signal region (50 GeV $< p_{\text{T}}(\tau) <$ 100 GeV) of the single-tau channel for pair production of vector leptoquarks $\text{LQ}_{3}^{\text{v}}$ in the minimal-coupling scenario. The plot does not show efficiencies for a branching fraction $B(\text{LQ}_{3}^{\text{v}} \rightarrow b\tau)$ of 0 or 1 because here the acceptance at generator level becomes zero and the efficiency is thus undefined.

Acceptance of the middle bin of the multi-bin signal region (100 GeV $< p_{\text{T}}(\tau) <$ 200 GeV) of the single-tau channel for pair production of vector leptoquarks $\text{LQ}_{3}^{\text{v}}$ in the minimal-coupling scenario.

Efficiency of the middle bin of the multi-bin signal region (100 GeV $< p_{\text{T}}(\tau) <$ 200 GeV) of the single-tau channel for pair production of vector leptoquarks $\text{LQ}_{3}^{\text{v}}$ in the minimal-coupling scenario. The plot does not show efficiencies for a branching fraction $B(\text{LQ}_{3}^{\text{v}} \rightarrow b\tau)$ of 0 or 1 because here the acceptance at generator level becomes zero and the efficiency is thus undefined.

Acceptance of the last bin of the multi-bin signal region (200 GeV $< p_{\text{T}}(\tau)$) of the single-tau channel for pair production of vector leptoquarks $\text{LQ}_{3}^{\text{v}}$ in the minimal-coupling scenario.

Efficiency of the last bin of the multi-bin signal region (200 GeV $< p_{\text{T}}(\tau)$) of the single-tau channel for pair production of vector leptoquarks $\text{LQ}_{3}^{\text{v}}$ in the minimal-coupling scenario. The plot does not show efficiencies for a branching fraction $B(\text{LQ}_{3}^{\text{v}} \rightarrow b\tau)$ of 0 or 1 because here the acceptance at generator level becomes zero and the efficiency is thus undefined.

Acceptance of the signal region of the di-tau channel for pair production of vector leptoquarks $\text{LQ}_{3}^{\text{v}}$ in the minimal-coupling scenario.

Efficiency of the signal region of the di-tau channel for pair production of vector leptoquarks $\text{LQ}_{3}^{\text{v}}$ in the minimal-coupling scenario. The plot does not show efficiencies for a branching fraction $B(\text{LQ}_{3}^{\text{v}} \rightarrow b\tau)$ of 0 because here the acceptance at generator level becomes zero and the efficiency is thus undefined.

Acceptance of the one-bin signal region of the single-tau channel for pair production of vector leptoquarks $\text{LQ}_{3}^{\text{v}}$ with additional gauge couplings.

Efficiency of the one-bin signal region of the single-tau channel for pair production of vector leptoquarks $\text{LQ}_{3}^{\text{v}}$ with additional gauge couplings. The plot does not show efficiencies for a branching fraction $B(\text{LQ}_{3}^{\text{v}} \rightarrow b\tau)$ of 0 or 1 because here the acceptance at generator level becomes zero and the efficiency is thus undefined.

Acceptance of the first bin of the multi-bin signal region (50 GeV $< p_{\text{T}}(\tau) <$ 100 GeV) of the single-tau channel for pair production of vector leptoquarks $\text{LQ}_{3}^{\text{v}}$ with additional gauge couplings.

Efficiency of the first bin of the multi-bin signal region (50 GeV $< p_{\text{T}}(\tau) <$ 100 GeV) of the single-tau channel for pair production of vector leptoquarks $\text{LQ}_{3}^{\text{v}}$ with additional gauge couplings. The plot does not show efficiencies for a branching fraction $B(\text{LQ}_{3}^{\text{v}} \rightarrow b\tau)$ of 0 or 1 because here the acceptance at generator level becomes zero and the efficiency is thus undefined.

Acceptance of the middle bin of the multi-bin signal region (100 GeV $< p_{\text{T}}(\tau) <$ 200 GeV) of the single-tau channel for pair production of vector leptoquarks $\text{LQ}_{3}^{\text{v}}$ with additional gauge couplings.

Efficiency of the middle bin of the multi-bin signal region (100 GeV $< p_{\text{T}}(\tau) <$ 200 GeV) of the single-tau channel for pair production of vector leptoquarks $\text{LQ}_{3}^{\text{v}}$ with additional gauge couplings. The plot does not show efficiencies for a branching fraction $B(\text{LQ}_{3}^{\text{v}} \rightarrow b\tau)$ of 0 or 1 because here the acceptance at generator level becomes zero and the efficiency is thus undefined.

Acceptance of the last bin of the multi-bin signal region (200 GeV $< p_{\text{T}}(\tau)$) of the single-tau channel for pair production of vector leptoquarks $\text{LQ}_{3}^{\text{v}}$ with additional gauge couplings.

Efficiency of the last bin of the multi-bin signal region (200 GeV $< p_{\text{T}}(\tau)$) of the single-tau channel for pair production of vector leptoquarks $\text{LQ}_{3}^{\text{v}}$ with additional gauge couplings. The plot does not show efficiencies for a branching fraction $B(\text{LQ}_{3}^{\text{v}} \rightarrow b\tau)$ of 0 or 1 because here the acceptance at generator level becomes zero and the efficiency is thus undefined.

Acceptance of the signal region of the di-tau channel for pair production of vector leptoquarks $\text{LQ}_{3}^{\text{v}}$ with additional gauge couplings.

Efficiency of the signal region of the di-tau channel for pair production of vector leptoquarks $\text{LQ}_{3}^{\text{v}}$ with additional gauge couplings. The plot does not show efficiencies for a branching fraction $B(\text{LQ}_{3}^{\text{v}} \rightarrow b\tau)$ of 0 because here the acceptance at generator level becomes zero and the efficiency is thus undefined.

Acceptance of the one-bin signal region of the single-tau channel for pair production of top squarks with decays via tau sleptons.

Efficiency of the one-bin signal region of the single-tau channel for pair production of top squarks with decays via tau sleptons.

Acceptance of the first bin of the multi-bin signal region (50 GeV $< p_{\text{T}}(\tau) <$ 100 GeV) of the single-tau channel for pair production of top squarks with decays via tau sleptons.

Efficiency of the first bin of the multi-bin signal region (50 GeV $< p_{\text{T}}(\tau) <$ 100 GeV) of the single-tau channel for pair production of top squarks with decays via tau sleptons.

Acceptance of the middle bin of the multi-bin signal region (100 GeV $< p_{\text{T}}(\tau) <$ 200 GeV) of the single-tau channel for pair production of top squarks with decays via tau sleptons.

Efficiency of the middle bin of the multi-bin signal region (100 GeV $< p_{\text{T}}(\tau) <$ 200 GeV) of the single-tau channel for pair production of top squarks with decays via tau sleptons.

Acceptance of the last bin of the multi-bin signal region (200 GeV $< p_{\text{T}}(\tau)$) of the single-tau channel for pair production of top squarks with decays via tau sleptons.

Efficiency of the last bin of the multi-bin signal region (200 GeV $< p_{\text{T}}(\tau)$) of the single-tau channel for pair production of top squarks with decays via tau sleptons.

Acceptance of the signal region of the di-tau channel for pair production of top squarks with decays via tau sleptons.

Efficiency of the signal region of the di-tau channel for pair production of top squarks with decays via tau sleptons.

Cutflow for the benchmark signal model $m(\tilde{t}_{1}) = 1350$ GeV, $m(\tilde{\tau}_{1}) = 1090$ GeV for the di-tau SR. The simulated sample contains 30,000 raw MC events. Weighted event yields are reported, normalized to an integrated luminosity of 139 fb$^{-1}$. 'Preselection' refers to the preselection for the di-tau channel.

Cutflow for the benchmark signal model $m(\tilde{t}_{1}) = 1350$ GeV, $m(\tilde{\tau}_{1}) = 1090$ GeV for the single-tau one-bin SR. The simulated sample contains 30,000 raw MC events. Weighted event yields are reported, normalized to an integrated luminosity of 139 fb$^{-1}$. 'Preselection' refers to the preselection for the single-tau channel.

Cutflow for the benchmark signal model $m(\tilde{t}_{1}) = 1350$ GeV, $m(\tilde{\tau}_{1}) = 1090$ GeV for the single-tau multi-bin SR. The simulated sample contains 30,000 raw MC events. Weighted event yields are reported, normalized to an integrated luminosity of 139 fb$^{-1}$. 'Preselection' refers to the preselection for the single-tau channel.

Cutflow for the benchmark signal model $m(\text{LQ}_{3}^{\text{u}}) = 1.2$ TeV, $\beta = 0.5$ for the di-tau SR. The simulated sample contains 210,000 raw MC events. Weighted event yields are reported, normalized to an integrated luminosity of 139 fb$^{-1}$. 'Preselection' refers to the preselection for the di-tau channel.

Cutflow for the benchmark signal model $m(\text{LQ}_{3}^{\text{u}}) = 1.2$ TeV, $\beta = 0.5$ for the single-tau one-bin SR. The simulated sample contains 210,000 raw MC events. Weighted event yields are reported, normalized to an integrated luminosity of 139 fb$^{-1}$. 'Preselection' refers to the preselection for the single-tau channel.

Cutflow for the benchmark signal model $m(\text{LQ}_{3}^{\text{u}}) = 1.2$ TeV, $\beta = 0.5$ for the single-tau multi-bin SR. The simulated sample contains 210,000 raw MC events. Weighted event yields are reported, normalized to an integrated luminosity of 139 fb$^{-1}$. 'Preselection' refers to the preselection for the single-tau channel.

Cutflow for the benchmark signal model $m(\text{LQ}_{3}^{\text{d}}) = 1.2$ TeV, $\beta = 0.5$ for the di-tau SR. The simulated sample contains 210,000 raw MC events. Weighted event yields are reported, normalized to an integrated luminosity of 139 fb$^{-1}$. 'Preselection' refers to the preselection for the di-tau channel.

Cutflow for the benchmark signal model $m(\text{LQ}_{3}^{\text{d}}) = 1.2$ TeV, $\beta = 0.5$ for the single-tau one-bin SR. The simulated sample contains 210,000 raw MC events. Weighted event yields are reported, normalized to an integrated luminosity of 139 fb$^{-1}$. 'Preselection' refers to the preselection for the single-tau channel.

Cutflow for the benchmark signal model $m(\text{LQ}_{3}^{\text{d}}) = 1.2$ TeV, $\beta = 0.5$ for the single-tau multi-bin SR. The simulated sample contains 210,000 raw MC events. Weighted event yields are reported, normalized to an integrated luminosity of 139 fb$^{-1}$. 'Preselection' refers to the preselection for the single-tau channel.

Cutflow for the benchmark signal model $m(\text{LQ}_{3}^{\text{v}}) = 1.4$ TeV, $\beta = 0.5$ in the minimal-coupling scenario for the di-tau SR. The simulated sample contains 50,000 raw MC events. Weighted event yields are reported, normalized to an integrated luminosity of 139 fb$^{-1}$. 'Preselection' refers to the preselection for the di-tau channel.

Cutflow for the benchmark signal model $m(\text{LQ}_{3}^{\text{v}}) = 1.4$ TeV, $\beta = 0.5$ in the minimal-coupling scenario for the single-tau one-bin SR. The simulated sample contains 50,000 raw MC events. Weighted event yields are reported, normalized to an integrated luminosity of 139 fb$^{-1}$. 'Preselection' refers to the preselection for the single-tau channel.

Cutflow for the benchmark signal model $m(\text{LQ}_{3}^{\text{v}}) = 1.4$ TeV, $\beta = 0.5$ in the minimal-coupling scenario for the single-tau multi-bin SR. The simulated sample contains 50,000 raw MC events. Weighted event yields are reported, normalized to an integrated luminosity of 139 fb$^{-1}$. 'Preselection' refers to the preselection for the single-tau channel.

Cutflow for the benchmark signal model $m(\text{LQ}_{3}^{\text{v}}) = 1.4$ TeV, $\beta = 0.5$ in the Yang--Mills scenario for the di-tau SR. The simulated sample contains 50,000 raw MC events. Weighted event yields are reported, normalized to an integrated luminosity of 139 fb$^{-1}$. 'Preselection' refers to the preselection for the di-tau channel.

Cutflow for the benchmark signal model $m(\text{LQ}_{3}^{\text{v}}) = 1.4$ TeV, $\beta = 0.5$ in the Yang--Mills scenario for the single-tau one-bin SR. The simulated sample contains 50,000 raw MC events. Weighted event yields are reported, normalized to an integrated luminosity of 139 fb$^{-1}$. 'Preselection' refers to the preselection for the single-tau channel.

Cutflow for the benchmark signal model $m(\text{LQ}_{3}^{\text{v}}) = 1.4$ TeV, $\beta = 0.5$ in the Yang--Mills scenario for the single-tau multi-bin SR. The simulated sample contains 50,000 raw MC events. Weighted event yields are reported, normalized to an integrated luminosity of 139 fb$^{-1}$. 'Preselection' refers to the preselection for the single-tau channel.