Observed (solid black line) and median expected (dashed black line) upper limits at 95% CL on the signal cross section times branching fraction $\sigma (\mathrm{pp} \to \mathrm{H})\,\mathcal{B}(\mathrm{H} \to \mathcal{A} \mathcal{A} \to 4 \gamma)$ at various $m_A$ mass points. The 68 and 95% confidence intervals around the median expected limit are shown as the green and yellow bands, respectively.

Relative cutflow fractions for representative signal samples, normalized to the number of events before any selection (All events = 1.0). The selections correspond to the trigger requirement, photon multiplicity and identification, kinematic requirements on the three-photon system, a selection on the angular separation (dR) of photons consistent with a merged topology, and a multivariate (MVA) discriminator and a charged-hadron isolation requirement applied to photon candidates used to suppress jet background. The final SR corresponds to the combination of the $m_H$ and $m_A$ signal regions used in the final fit.

The postfit pDNN distributions in the SR $\mu$ 3b3j assuming $m_{H^\pm} = 600$ GeV. Postfit signal for $m_{H^\pm} = 600$ GeV is also shown. Beneath plot the ratio of data to predictions is shown.

The postfit pDNN distributions in the SR e 2b2j assuming $m_{H^\pm} = 1$ TeV. The signal for $m_{H^\pm} = 1$ TeV is also shown before the fit, assuming a cross section of 1 pb.Beneath plot the ratio of data to predictions is shown.

The postfit pDNN distributions in the SR $\mu$ 2b2j assuming $m_{H^\pm} = 1$ TeV. The signal for $m_{H^\pm} = 1$ TeV is also shown before the fit, assuming a cross section of 1 pb.Beneath plot the ratio of data to predictions is shown.

The postfit pDNN distributions in the SR e 3b3j assuming $m_{H^\pm} = 1$ TeV. The signal for $m_{H^\pm} = 1$ TeV is also shown before the fit, assuming a cross section of 1 pb.Beneath plot the ratio of data to predictions is shown.

The postfit pDNN distributions in the SR $\mu$ 3b3j assuming $m_{H^\pm} = 1$ TeV. The signal for $m_{H^\pm} = 1$ TeV is also shown before the fit, assuming a cross section of 1 pb.Beneath plot the ratio of data to predictions is shown.

Observed and expected 95% CL limits on the cross section times branching ratio for different $H^{\pm}$ mass hypotheses with $ρ_{tc} = 0.4$, $ρ_{tt} = 0.6$ . The inner (green) band and the outer (yellow) band represent the regions containing 68% and 95%, respectively, of the distribution of limits expected under the background-only hypothesis. Theoretical prediction with different $ρ_{tc}$ and $ρ_{tt}$ couplingsare shown with the grey bands representing the corresponding uncertainties due to the factorization and renormalization scales and the PDFs.

Observed excluded phase-space regions, based on 95% CL limits, as a function of $m_{H^{\pm}}$ and $ρ_{tc}$ for various assumed $ρ_{tt}$ values represented with different colors. The limits are extracted from the pDNN distributions based on g2HDM assuming all the extra Yukawa couplings except $ρ_{tt}$ and $ρ_{tc}$ are zero. The results are obtained from the 95% CL limits on the cross section times branching fraction in Fig 6.

Expected excluded phase-space regions, based on 95% CL limits, as a function of $m_{H^{\pm}}$ and $ρ_{tc}$ for various assumed $ρ_{tt}$ values represented with different colors. The limits are extracted from the pDNN distributions based on g2HDM assuming all the extra Yukawa couplings except $ρ_{tt}$ and $ρ_{tc}$ are zero. The results are obtained from the 95% CL limits on the cross section times branching fraction in Fig 6.

Observed excluded phase-space regions, based on 95% CL limits, as a function of $m_{H^{\pm}}$ and $ρ_{tc}$ for various assumed $ρ_{tt}$ values represented with different colors. The limits are extracted from the pDNN distributions based on g2HDM assuming all the extra Yukawa couplings except $ρ_{tt}$ and $ρ_{tc}$ are zero. The results are obtained from the 95% CL limits on the cross section times branching fraction in Fig 6.

Expected excluded phase-space regions, based on 95% CL limits, as a function of $m_{H^{\pm}}$ and $ρ_{tc}$ for various assumed $ρ_{tt}$ values represented with different colors. The limits are extracted from the pDNN distributions based on g2HDM assuming all the extra Yukawa couplings except $ρ_{tt}$ and $ρ_{tc}$ are zero. The results are obtained from the 95% CL limits on the cross section times branching fraction in Fig 6.

Observed excluded phase-space regions, based on 95% CL limits, as a function of $m_{H^{\pm}}$ and $ρ_{tc}$ for various assumed $ρ_{tt}$ values represented with different colors. The limits are extracted from the pDNN distributions based on g2HDM assuming all the extra Yukawa couplings except $ρ_{tt}$ and $ρ_{tc}$ are zero. The results are obtained from the 95% CL limits on the cross section times branching fraction in Fig 6.

Expected excluded phase-space regions, based on 95% CL limits, as a function of $m_{H^{\pm}}$ and $ρ_{tc}$ for various assumed $ρ_{tt}$ values represented with different colors. The limits are extracted from the pDNN distributions based on g2HDM assuming all the extra Yukawa couplings except $ρ_{tt}$ and $ρ_{tc}$ are zero. The results are obtained from the 95% CL limits on the cross section times branching fraction in Fig 6.

Observed excluded phase-space regions, based on 95% CL limits, as a function of $m_{H^{\pm}}$ and $ρ_{tc}$ for various assumed $ρ_{tt}$ values represented with different colors. The limits are extracted from the pDNN distributions based on g2HDM assuming all the extra Yukawa couplings except $ρ_{tt}$ and $ρ_{tc}$ are zero. The results are obtained from the 95% CL limits on the cross section times branching fraction in Fig 6.

Expected excluded phase-space regions, based on 95% CL limits, as a function of $m_{H^{\pm}}$ and $ρ_{tc}$ for various assumed $ρ_{tt}$ values represented with different colors. The limits are extracted from the pDNN distributions based on g2HDM assuming all the extra Yukawa couplings except $ρ_{tt}$ and $ρ_{tc}$ are zero. The results are obtained from the 95% CL limits on the cross section times branching fraction in Fig 6.

Observed excluded phase-space regions, based on 95% CL limits, as a function of $m_{H^{\pm}}$ and $ρ_{tc}$ for various assumed $ρ_{tt}$ values represented with different colors. The limits are extracted from the pDNN distributions based on g2HDM assuming all the extra Yukawa couplings except $ρ_{tt}$ and $ρ_{tc}$ are zero. The results are obtained from the 95% CL limits on the cross section times branching fraction in Fig 6.

Expected excluded phase-space regions, based on 95% CL limits, as a function of $m_{H^{\pm}}$ and $ρ_{tc}$ for various assumed $ρ_{tt}$ values represented with different colors. The limits are extracted from the pDNN distributions based on g2HDM assuming all the extra Yukawa couplings except $ρ_{tt}$ and $ρ_{tc}$ are zero. The results are obtained from the 95% CL limits on the cross section times branching fraction in Fig 6.

Two times the difference of the negative log-likelihood (NLL) as a function of σ(pp →(q$^{light}$)H$^{\pm}$) B(H$^{\pm}$ → tb, t → blν) and σ(pp → bH$^{\pm}$)B(H$^{\pm}$ → tb, t → blν) with the best fit point extracted from the pDNN distribution for the $H^{\pm}$ mass $m_{H^{\pm}} = 600$ GeV. The solid and dashed contours represent the observed and expected limits, respectively. The points along the contours represent 68% and 95% CL regions, extracted from the 2∆NLL values of 2.30 and 5.99, respectively. The g2HDM predictions are also shown. The points along the g2HDM prediction line represent different $ρ_{tt}$, $ρ_{tc}$ coupling sets, with all other extra Yukawa couplings assumed to be zero.

Two times the difference of the negative log-likelihood (NLL) as a function of σ(pp →(q$^{light}$)H$^{\pm}$) B(H$^{\pm}$ → tb, t → blν) and σ(pp → bH$^{\pm}$)B(H$^{\pm}$ → tb, t → blν) with the best fit point extracted from the pDNN distribution for the $H^{\pm}$ mass $m_{H^{\pm}} = 600$ GeV. The solid and dashed contours represent the observed and expected limits, respectively. The points along the contours represent 68% and 95% CL regions, extracted from the 2∆NLL values of 2.30 and 5.99, respectively. The g2HDM predictions are also shown. The points along the g2HDM prediction line represent different $ρ_{tt}$, $ρ_{tc}$ coupling sets, with all other extra Yukawa couplings assumed to be zero.

Two times the difference of the negative log-likelihood (NLL) as a function of σ(pp →(q$^{light}$)H$^{\pm}$) B(H$^{\pm}$ → tb, t → blν) and σ(pp → bH$^{\pm}$)B(H$^{\pm}$ → tb, t → blν) with the best fit point extracted from the pDNN distribution for the $H^{\pm}$ mass $m_{H^{\pm}} = 1000$ GeV. The solid and dashed contours represent the observed and expected limits, respectively. The points along the contours represent 68% and 95% CL regions, extracted from the 2∆NLL values of 2.30 and 5.99, respectively. The g2HDM predictions are also shown. The points along the g2HDM prediction line represent different $ρ_{tt}$, $ρ_{tc}$ coupling sets, with all other extra Yukawa couplings assumed to be zero.

Two times the difference of the negative log-likelihood (NLL) as a function of σ(pp →(q$^{light}$)H$^{\pm}$) B(H$^{\pm}$ → tb, t → blν) and σ(pp → bH$^{\pm}$)B(H$^{\pm}$ → tb, t → blν) with the best fit point extracted from the pDNN distribution for the $H^{\pm}$ mass $m_{H^{\pm}} = 1000$ GeV. The solid and dashed contours represent the observed and expected limits, respectively. The points along the contours represent 68% and 95% CL regions, extracted from the 2∆NLL values of 2.30 and 5.99, respectively. The g2HDM predictions are also shown. The points along the g2HDM prediction line represent different $ρ_{tt}$, $ρ_{tc}$ coupling sets, with all other extra Yukawa couplings assumed to be zero.

Covariance for spin correlation coefficients in beam basis for $p_{T}(t)$ vs. $|cos(\theta)|$ bins

Results for Eigenstate decomposition for $|\Phi^{+} \rangle$ in Helicity basis in $m(t\bar{t})$ vs. $|cos(\theta)|$ bins

Results for Eigenstate decomposition for $|\Phi^{-} \rangle$ in Helicity basis in $m(t\bar{t})$ vs. $|cos(\theta)|$ bins

Results for Eigenstate decomposition for $|\Psi^{+} \rangle$ in Helicity basis in $m(t\bar{t})$ vs. $|cos(\theta)|$ bins

Results for Eigenstate decomposition for $|\Psi^{-} \rangle$ in Helicity basis in $m(t\bar{t})$ vs. $|cos(\theta)|$ bins

Results for Eigenstate decomposition for $|\uparrow \uparrow \rangle$ in Beam basis in $m(t\bar{t})$ vs. $|cos(\theta)|$ bins

Results for Eigenstate decomposition for $|\downarrow \downarrow \rangle$ in Beam basis in $m(t\bar{t})$ vs. $|cos(\theta)|$ bins

Results for Eigenstate decomposition for $|\Psi^{+} \rangle$ in Beam basis in $m(t\bar{t})$ vs. $|cos(\theta)|$ bins

Results for Eigenstate decomposition for $|\Psi^{-} \rangle$ in Beam basis in $m(t\bar{t})$ vs. $|cos(\theta)|$ bins

Results for Purity in Helicity basis in $m(t\bar{t})$ vs. $|cos(\theta)|$ bins

Results for Purity in Beam basis in $m(t\bar{t})$ vs. $|cos(\theta)|$ bins

Results for Entropy in Helicity basis in $m(t\bar{t})$ vs. $|cos(\theta)|$ bins

Results for Entropy in Beam basis in $m(t\bar{t})$ vs. $|cos(\theta)|$ bins

Results for Entanglement in Helicity basis in $m(t\bar{t})$ vs. $|cos(\theta)|$ bins

Results for Entanglement in Beam basis in $m(t\bar{t})$ vs. $|cos(\theta)|$ bins

Results for Eigenstate decomposition for $|\Phi^{+} \rangle$ in Helicity basis in $p_{T}(t)$ vs. $|cos(\theta)|$ bins

Results for Eigenstate decomposition for $|\Phi^{-} \rangle$ in Helicity basis in $p_{T}(t)$ vs. $|cos(\theta)|$ bins

Results for Eigenstate decomposition for $|\Psi^{+} \rangle$ in Helicity basis in $p_{T}(t)$ vs. $|cos(\theta)|$ bins

Results for Eigenstate decomposition for $|\Psi^{-} \rangle$ in Helicity basis in $p_{T}(t)$ vs. $|cos(\theta)|$ bins

Results for Eigenstate decomposition for $|\uparrow \uparrow \rangle$ in Beam basis in $p_{T}(t)$ vs. $|cos(\theta)|$ bins

Results for Eigenstate decomposition for $|\downarrow \downarrow \rangle$ in Beam basis in $p_{T}(t)$ vs. $|cos(\theta)|$ bins

Results for Eigenstate decomposition for $|\Psi^{+} \rangle$ in Beam basis in $p_{T}(t)$ vs. $|cos(\theta)|$ bins

Results for Eigenstate decomposition for $|\Psi^{-} \rangle$ in Beam basis in $p_{T}(t)$ vs. $|cos(\theta)|$ bins

Results for Purity in Helicity basis in $p_{T}(t)$ vs. $|cos(\theta)|$ bins

Results for Purity in Beam basis in $p_{T}(t)$ vs. $|cos(\theta)|$ bins

Results for Entropy in Helicity basis in $p_{T}(t)$ vs. $|cos(\theta)|$ bins

Results for Entropy in Beam basis in $p_{T}(t)$ vs. $|cos(\theta)|$ bins

Results for Entanglement in Helicity basis in $p_{T}(t)$ vs. $|cos(\theta)|$ bins

Results for Entanglement in Beam basis in $p_{T}(t)$ vs. $|cos(\theta)|$ bins

Measured and expected fiducial cross section for Z$\gamma$ production. Corresponding to Table 1 in the paper.

Normalized differential cross sections, compared with the SM predictions, as a function of the absolute value of pseudorapidity of the subleading jet in transverse momentum. The SM predictions are obtained using Madgraph5_aMC@NLO version 2.6.5 at NLO in QCD with PYTHIA version 8.240

Normalized differential cross sections, compared with the SM predictions, as a function of the invariant mass of the two jets. The SM predictions are obtained using Madgraph5_aMC@NLO version 2.6.5 at NLO in QCD with PYTHIA version 8.240

Normalized differential cross sections, compared with the SM predictions, as a function of the photon's transverse momentum. The SM predictions are obtained using Madgraph5_aMC@NLO version 2.6.5 at NLO in QCD with PYTHIA version 8.240

Normalized differential cross sections, compared with the SM predictions, as a function of the event centrality as defined in Ref.[13] of the paper. The SM predictions are obtained using Madgraph5_aMC@NLO version 2.6.5 at NLO in QCD with PYTHIA version 8.240

Normalized differential cross sections, compared with the SM predictions, as a function of the Zeppenfeld variable that is defined in Eq.(1) of the paper. The SM predictions are obtained using Madgraph5_aMC@NLO version 2.6.5 at NLO in QCD with PYTHIA version 8.240

Negative of twice in the difference in the log-likelihood as a function of $c_{\mathrm{W}}$, based on 138 $\mathrm{fb}^{-1}$ of CMS data at 13 TeV.

Negative of twice in the difference in the log-likelihood as a function of $c_{\mathrm{HWB}}$, based on 138 $\mathrm{fb}^{-1}$ of CMS data at 13 TeV.

Negative of twice in the difference in the log-likelihood as a function of $c_{\mathrm{HWB}}$ and $c_{\mathrm{W}}$, based on 138 $\mathrm{fb}^{-1}$ of CMS data at 13 TeV.

Predicted and measured EWK $\gamma\mathrm{jj}$ fiducial cross sections.

Data vs. the ABCD method background prediction for 2017 in $\mathrm{SR_{b}^{mm}}$. Events are divided by the bin width, hence fractional data counts. Error bars show statistical uncertainties of data. The blue band shows the propagated uncertainty of all individual fit variations in a given bin, which we consider to be uncorrelated. The lower panels show the ratio of the observed data to the background estimation.

Data vs. the ABCD method background prediction for 2018 in $\mathrm{SR_{b}^{mm}}$. Events are divided by the bin width, hence fractional data counts. Error bars show statistical uncertainties of data. The blue band shows the propagated uncertainty of all individual fit variations in a given bin, which we consider to be uncorrelated. The lower panels show the ratio of the observed data to the background estimation.

Data vs. the ABCD method background prediction for 2016 in $\mathrm{SR_{b+\textrm{j}/b}^{mm}}$. Events are divided by the bin width, hence fractional data counts. Error bars show statistical uncertainties of data. The blue band shows the propagated uncertainty of all individual fit variations in a given bin, which we consider to be uncorrelated. The lower panels show the ratio of the observed data to the background estimation.

Data vs. the ABCD method background prediction for 2017 in $\mathrm{SR_{b+\textrm{j}/b}^{mm}}$. Events are divided by the bin width, hence fractional data counts. Error bars show statistical uncertainties of data. The blue band shows the propagated uncertainty of all individual fit variations in a given bin, which we consider to be uncorrelated. The lower panels show the ratio of the observed data to the background estimation.

Data vs. the ABCD method background prediction for 2018 in $\mathrm{SR_{b+\textrm{j}/b}^{mm}}$. Events are divided by the bin width, hence fractional data counts. Error bars show statistical uncertainties of data. The blue band shows the propagated uncertainty of all individual fit variations in a given bin, which we consider to be uncorrelated. The lower panels show the ratio of the observed data to the background estimation.

The asymptotic 95% CL limits on the product of cross section ($\sigma$), branching fraction into a pair of muons ($\mathcal{B}$) to dimuons, acceptance (A), and efficiency ($\epsilon$) in $\mathrm{SR_{b}^{mm}}$. A combination of both limits depends on the relative acceptance contributions in each region and is omitted here.

The asymptotic 95% CL limits on the product of cross section ($\sigma$), branching fraction into a pair of muons ($\mathcal{B}$) to dimuons, acceptance (A), and efficiency ($\epsilon$) in $\mathrm{SR_{b+\textrm{j}/b}^{mm}}$. A combination of both limits depends on the relative acceptance contributions in each region and is omitted here.

Cutflow for signal samples in 2016. Samples are divided by generated mass, signal region, and the b to s quark flavor changing coupling of the $Z^\prime$, dbs. For some mass values, multiple values of this coupling strength have been generated. It is described in this minimal lagrangian: $\mathcal{L} \supset Z^{\prime}_{\eta} \Big[ g_{\mu}\, \bar{\mu}\, \gamma^{\eta}\, \mu + \big( g_b\, \delta_{bs}\, \bar{s}\, \gamma^{\eta}\, P_{L}\, b + \text{h.c.} \big) \Big]$. The cutflow shows the selection into signal regions, and then the effects of specialized HTLT and RelMET selections, which grow with greater mass.

Cutflow for signal samples in 2017. Samples are divided by generated mass, signal region, and the b to s quark flavor changing coupling of the $Z^\prime$, dbs. For some mass values, multiple values of this coupling strength have been generated. It is described in this minimal lagrangian: $\mathcal{L} \supset Z^{\prime}_{\eta} \Big[ g_{\mu}\, \bar{\mu}\, \gamma^{\eta}\, \mu + \big( g_b\, \delta_{bs}\, \bar{s}\, \gamma^{\eta}\, P_{L}\, b + \text{h.c.} \big) \Big]$. The cutflow shows the selection into signal regions, and then the effects of specialized HTLT and RelMET selections, which grow with greater mass.

Cutflow for signal samples in 2018. Samples are divided by generated mass, signal region, and the b to s quark flavor changing coupling of the $Z^\prime$, dbs. For some mass values, multiple values of this coupling strength have been generated. It is described in this minimal lagrangian: $\mathcal{L} \supset Z^{\prime}_{\eta} \Big[ g_{\mu}\, \bar{\mu}\, \gamma^{\eta}\, \mu + \big( g_b\, \delta_{bs}\, \bar{s}\, \gamma^{\eta}\, P_{L}\, b + \text{h.c.} \big) \Big]$. The cutflow shows the selection into signal regions, and then the effects of specialized HTLT and RelMET selections, which grow with greater mass.

Observed 95% CL upper limits on the $\tilde{t}$ production cross section, as functions of $m_{\tilde{t}}$ and $\Delta m$, for $\mathcal{B}(\tilde{t} \to bf\overline{f}'\tilde{\chi}^{0}_{1})$ of 50%. The observed (solid black) and expected (dashed red) exclusion curves are overlaid on the plots. The search excludes the region to the left of the exclusion curves.

Observed 95% CL upper limits on the $\tilde{t}$ production cross section, as functions of $m_{\tilde{t}}$ and $\Delta m$, for $\mathcal{B}(\tilde{t} \to bf\overline{f}'\tilde{\chi}^{0}_{1})$ of 50%. The observed (solid black) and expected (dashed red) exclusion curves are overlaid on the plots. The search excludes the region to the left of the exclusion curves.

Observed 95% CL upper limits on the $\tilde{t}$ production cross section, as functions of $m_{\tilde{t}}$ and $\Delta m$, for $\mathcal{B}(\tilde{t} \to bf\overline{f}'\tilde{\chi}^{0}_{1})$ of 100%. The observed (solid black) and expected (dashed red) exclusion curves are overlaid on the plots. The search excludes the region to the left of the exclusion curves.

Observed 95% CL upper limits on the $\tilde{t}$ production cross section, as functions of $m_{\tilde{t}}$ and $\Delta m$, for $\mathcal{B}(\tilde{t} \to bf\overline{f}'\tilde{\chi}^{0}_{1})$ of 100%. The observed (solid black) and expected (dashed red) exclusion curves are overlaid on the plots. The search excludes the region to the left of the exclusion curves.

Observed 95% CL upper limits on the production cross section for the bino-wino NLSP model, as functions of $m_{ ext{LLP}}$ and $c\tau$, for $\Delta m$ of 12 GeV. The observed (solid black) and expected (dashed red) exclusion curves are overlaid on the plots. The search excludes the region to the left of the exclusion curves.

Observed 95% CL upper limits on the production cross section for the bino-wino NLSP model, as functions of $m_{ ext{LLP}}$ and $c\tau$, for $\Delta m$ of 12 GeV. The observed (solid black) and expected (dashed red) exclusion curves are overlaid on the plots. The search excludes the region to the left of the exclusion curves.

Observed 95% CL upper limits on the production cross section for the bino-wino NLSP model, as functions of the $m_{ ext{LLP}}$ and $c\tau$, for $\Delta m$ of 15 GeV. The observed (solid black) and expected (dashed red) exclusion curves are overlaid on the plots. The search excludes the region to the left of the exclusion curves.

Observed 95% CL upper limits on the production cross section for the bino-wino NLSP model, as functions of the $m_{ ext{LLP}}$ and $c\tau$, for $\Delta m$ of 15 GeV. The observed (solid black) and expected (dashed red) exclusion curves are overlaid on the plots. The search excludes the region to the left of the exclusion curves.

Observed 95% CL upper limits on the production cross section for the bino-wino NLSP model, as functions of the $m_{ ext{LLP}}$ and $c\tau$, for $\Delta m$ of 20 GeV. The observed (solid black) and expected (dashed red) exclusion curves are overlaid on the plots. The search excludes the region to the left of the exclusion curves.

Observed 95% CL upper limits on the production cross section for the bino-wino NLSP model, as functions of the $m_{ ext{LLP}}$ and $c\tau$, for $\Delta m$ of 20 GeV. The observed (solid black) and expected (dashed red) exclusion curves are overlaid on the plots. The search excludes the region to the left of the exclusion curves.

Observed 95% CL upper limits on the production cross section for the bino-wino NLSP model, as functions of the $m_{ ext{LLP}}$ and $c\tau$, for $\Delta m$ of 25 GeV. The observed (solid black) and expected (dashed red) exclusion curves are overlaid on the plots. The search excludes the region to the left of the exclusion curves.

Observed 95% CL upper limits on the production cross section for the bino-wino NLSP model, as functions of the $m_{ ext{LLP}}$ and $c\tau$, for $\Delta m$ of 25 GeV. The observed (solid black) and expected (dashed red) exclusion curves are overlaid on the plots. The search excludes the region to the left of the exclusion curves.

Example description

Example description

Normalized differential distributions of the $\Delta R (\gamma,\mathrm{tt})$. The nominal MC prediction used to compare the experimental results to is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and Madgraph5 at LO in QCD for photons from the decay part of the process. The alternative prediction is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and a POWHEG+Pythia $\mathrm{tt}$ simulation at NLO in QCD for photons from the decay part of the process.

Absolute differential distributions of the min. $\Delta R (\gamma,\mathrm{t})$. The nominal MC prediction used to compare the experimental results to is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and Madgraph5 at LO in QCD for photons from the decay part of the process. The alternative prediction is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and a POWHEG+Pythia $\mathrm{tt}$ simulation at NLO in QCD for photons from the decay part of the process.

Normalized differential distributions of the min. $\Delta R (\gamma,\mathrm{t})$. The nominal MC prediction used to compare the experimental results to is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and Madgraph5 at LO in QCD for photons from the decay part of the process. The alternative prediction is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and a POWHEG+Pythia $\mathrm{tt}$ simulation at NLO in QCD for photons from the decay part of the process.

Absolute differential distributions of the $m(\mathrm{tt})$. The nominal MC prediction used to compare the experimental results to is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and Madgraph5 at LO in QCD for photons from the decay part of the process. The alternative prediction is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and a POWHEG+Pythia $\mathrm{tt}$ simulation at NLO in QCD for photons from the decay part of the process.

Normalized differential distributions of the $m(\mathrm{tt})$. The nominal MC prediction used to compare the experimental results to is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and Madgraph5 at LO in QCD for photons from the decay part of the process. The alternative prediction is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and a POWHEG+Pythia $\mathrm{tt}$ simulation at NLO in QCD for photons from the decay part of the process.

Absolute differential distributions of the leading lepton $p_{\mathrm{T}}$. The nominal MC prediction used to compare the experimental results to is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and Madgraph5 at LO in QCD for photons from the decay part of the process. The alternative prediction is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and a POWHEG+Pythia $\mathrm{tt}$ simulation at NLO in QCD for photons from the decay part of the process. The theory predictions are computed at fixed order, at NLO in QCD, as described in Section 9.2 of the paper.

Normalized differential distributions of the leading lepton $p_{\mathrm{T}}$. The nominal MC prediction used to compare the experimental results to is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and Madgraph5 at LO in QCD for photons from the decay part of the process. The alternative prediction is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and a POWHEG+Pythia $\mathrm{tt}$ simulation at NLO in QCD for photons from the decay part of the process. The theory predictions are computed at fixed order, at NLO in QCD, as described in Section 9.2 of the paper.

Absolute differential distributions of the photon $p_{\mathrm{T}}$. The nominal MC prediction used to compare the experimental results to is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and Madgraph5 at LO in QCD for photons from the decay part of the process. The alternative prediction is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and a POWHEG+Pythia $\mathrm{tt}$ simulation at NLO in QCD for photons from the decay part of the process. The theory predictions are computed at fixed order, at NLO in QCD, as described in Section 9.2 of the paper.

Normalized differential distributions of the photon $p_{\mathrm{T}}$. The nominal MC prediction used to compare the experimental results to is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and Madgraph5 at LO in QCD for photons from the decay part of the process. The alternative prediction is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and a POWHEG+Pythia $\mathrm{tt}$ simulation at NLO in QCD for photons from the decay part of the process. The theory predictions are computed at fixed order, at NLO in QCD, as described in Section 9.2 of the paper.

Absolute differential distributions of the $\Delta \phi (\ell,\ell)$. The nominal MC prediction used to compare the experimental results to is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and Madgraph5 at LO in QCD for photons from the decay part of the process. The alternative prediction is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and a POWHEG+Pythia $\mathrm{tt}$ simulation at NLO in QCD for photons from the decay part of the process. The theory predictions are computed at fixed order, at NLO in QCD, as described in Section 9.2 of the paper.

Normalized differential distributions of the $\Delta \phi (\ell,\ell)$. The nominal MC prediction used to compare the experimental results to is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and Madgraph5 at LO in QCD for photons from the decay part of the process. The alternative prediction is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and a POWHEG+Pythia $\mathrm{tt}$ simulation at NLO in QCD for photons from the decay part of the process. The theory predictions are computed at fixed order, at NLO in QCD, as described in Section 9.2 of the paper.

Absolute differential distributions of the ratio between the cross sections of $\mathrm{tt\gamma}$ and $\mathrm{tt}$, as a function of the leading top quark $p_{\mathrm{T}}$. The nominal MC prediction used to compare the experimental results to is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and Madgraph5 at LO in QCD for photons from the decay part of the process. The alternative prediction is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and a POWHEG+Pythia $\mathrm{tt}$ simulation at NLO in QCD for photons from the decay part of the process.

Absolute differential distributions of the ratio between the cross sections of $\mathrm{tt\gamma}$ and $\mathrm{tt}$, as a function of the leading lepton $p_{\mathrm{T}}$. The nominal MC prediction used to compare the experimental results to is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and Madgraph5 at LO in QCD for photons from the decay part of the process. The alternative prediction is obtained with Madgraph5 at NLO in QCD for photons from the production part of the process and a POWHEG+Pythia $\mathrm{tt}$ simulation at NLO in QCD for photons from the decay part of the process. The theory predictions are computed at fixed order, at NLO in QCD, as described in Section 9.2 of the paper.

Correlation matrix of the total uncertainty in the absolute differential cross section as a function of the leading top quark $p_{\mathrm{T}}$.

Correlation matrix of the statistical uncertainty in the absolute differential cross section as a function of the leading top quark $p_{\mathrm{T}}$.

Correlation matrix of the total uncertainty in the absolute differential cross section as a function of the $\Delta R (\gamma,\mathrm{tt})$.

Correlation matrix of the statistical uncertainty in the absolute differential cross section as a function of the $\Delta R (\gamma,\mathrm{tt})$.

Correlation matrix of the total uncertainty in the absolute differential cross section as a function of the min. $\Delta R (\gamma,\mathrm{t})$.

Correlation matrix of the statistical uncertainty in the absolute differential cross section as a function of the min. $\Delta R (\gamma,\mathrm{t})$.

Correlation matrix of the total uncertainty in the absolute differential cross section as a function of the $m(\mathrm{tt})$.

Correlation matrix of the statistical uncertainty in the absolute differential cross section as a function of the $m(\mathrm{tt})$.

Correlation matrix of the total uncertainty in the absolute differential cross section as a function of the leading lepton $p_{\mathrm{T}}$.

Correlation matrix of the statistical uncertainty in the absolute differential cross section as a function of the leading lepton $p_{\mathrm{T}}$.

Correlation matrix of the total uncertainty in the absolute differential cross section as a function of the photon $p_{\mathrm{T}}$.

Correlation matrix of the statistical uncertainty in the absolute differential cross section as a function of the photon $p_{\mathrm{T}}$.

Correlation matrix of the total uncertainty in the absolute differential cross section as a function of the $\Delta \phi (\ell,\ell)$.

Correlation matrix of the statistical uncertainty in the absolute differential cross section as a function of the $\Delta \phi (\ell,\ell)$.

Correlation matrix of the total uncertainty in the absolute differential $\mathrm{tt\gamma}$ / $\mathrm{tt}$ cross section ratio as a function of the leading top $p_{\mathrm{T}}$.

Correlation matrix of the statistical uncertainty in the absolute differential $\mathrm{tt\gamma}$ / $\mathrm{tt}$ cross section ratio as a function of the leading top $p_{\mathrm{T}}$.

Correlation matrix of the total uncertainty in the absolute differential $\mathrm{tt\gamma}$ / $\mathrm{tt}$ cross section ratio as a function of the leading lepton $p_{\mathrm{T}}$.

Correlation matrix of the statistical uncertainty in the absolute differential $\mathrm{tt\gamma}$ / $\mathrm{tt}$ cross section ratio as a function of the leading lepton $p_{\mathrm{T}}$.

Observed differential dijet spectrum using full Run-2 data.

The observed 95% CL upper limits on the product of the cross section, branching fraction, and acceptance for quark-quark type dijet resonances.

The observed 95% CL upper limits on the product of the cross section, branching fraction, and acceptance for quark-gluon type dijet resonances.

The observed 95% CL upper limits on the product of the cross section, branching fraction, and acceptance for gluon-gluon type dijet resonances.

The observed and expected 95% CL upper limits on the universal quark coupling $g_{q}'$ as a function of resonance mass for a leptophobic Z' resonance that only couples to quarks.