Search for a Higgs boson in diphoton final states with the D0 detector in $9.6{\mathrm{fb}}^{-1}$ of $p\overline{p}$ collisions at $\sqrt{s}\mathbf{=}1.96\mathrm{TeV}$

We present a search for a Higgs boson decaying into a pair of photons based on $9.6{\mathrm{fb}}^{-1}$ of $p\overline{p}$ collisions at $\sqrt{s}=1.96\mathrm{TeV}$ collected with the D0 detector at the Fermilab Tevatron Collider. The search employs multivariate techniques to discriminate signal from the nonresonant background and is separately optimized for a standard model and a fermiophobic Higgs boson. No significant excess of data above the background prediction is observed and upper limits on the product of the cross section and branching fraction are derived at the 95% confidence level as a function of Higgs boson mass. For a standard model Higgs boson with mass of 125 GeV, the observed (expected) upper limits are a factor of 12.8 (8.7) above the standard model prediction. The existence of a fermiophobic Higgs boson with mass in the 100–113 GeV range is excluded at the 95% confidence level.

Figure 1

Comparison of the normalized ${\mathrm{O}}_{\mathrm{NN}}$ spectra for photons from DPP MC simulations and $Z\to {\ell }^{+}{\ell }^{-}\gamma$ data events (points with statistical error bars), and for misidentified jets from simulated dijet events.

Figure 2

Probability to select the correct hard-scatter PV as a function of $p_{\mathrm{T}}^{\gamma \gamma }$ as measured in $Z/{\gamma }^{*}\to e^{+}{e}^{-}$ events excluding the electron and positron tracks from consideration. The two different algorithms discussed in the text are compared.

Figure 3

Distribution of the reconstructed diphoton invariant mass distribution corresponding to a Higgs boson signal with $M_H=125\mathrm{GeV}$. The line shows the result of a fit to the distribution using the functional form described in Sec. 4d.

Figure 4

Distribution of $M_{\gamma \gamma }$ in (a) the photon-enriched sample and (b) the jet-enriched sample. The data (points with statistical error bars) are compared to the background prediction, broken down into its individual components. The expected distributions for a SM Higgs boson and a fermiophobic Higgs boson with $M_H=125\mathrm{GeV}$ are also shown scaled by a factor of 100.

Figure 5

Distribution of (a) $p_{\mathrm{T}}^{\gamma 1}$ in the photon-enriched sample and (b) $p_{\mathrm{T}}^{\gamma 2}$ in the jet-enriched sample. The data (points with statistical error bars) are compared to the background prediction, broken down into its individual components. The expected distributions for a SM Higgs boson and a fermiophobic Higgs boson with $M_H=125\mathrm{GeV}$ are also shown scaled by a factor of 1000. These two BDT input variables are used in both the photon-enriched and jet-enriched samples, but are displayed here for only one of the samples for illustrative purposes.

Figure 6

Distribution of (a) $p_{\mathrm{T}}^{\gamma \gamma }$ in the photon-enriched sample and (b) $\Delta {\varphi }_{\gamma \gamma }$ in the jet-enriched sample. The data (points with statistical error bars) are compared to the background prediction, broken down into its individual components. The expected distributions for a SM Higgs boson and a fermiophobic Higgs boson with $M_H=125\mathrm{GeV}$ are also shown scaled by a factor of 1000. These two BDT input variables are used in both the photon-enriched and jet-enriched samples, but are displayed here for only one of the samples for illustrative purposes.

Figure 7

Distribution of (a) ${\varphi }^{*}$ in the photon-enriched sample and (b) ${\mathrm{E̸}}_T$ in the jet-enriched sample. The data (points with statistical error bars) are compared to the background prediction, broken down into its individual components. The expected distributions for a SM Higgs boson and a fermiophobic Higgs boson with $M_H=125\mathrm{GeV}$ are also shown scaled by a factor of 1000. These two BDT input variables are used in both the photon-enriched and jet-enriched samples, but are displayed here for only one of the samples for illustrative purposes.

Figure 8

Distribution of (a) $O_{\mathrm{NN}}^{\gamma 1}$ and (b) $O_{\mathrm{NN}}^{\gamma 2}$ in the jet-enriched sample. The data (points with statistical error bars) are compared to the background prediction, broken down into its individual components. The expected distributions for a SM Higgs boson and a fermiophobic Higgs boson with $M_H=125\mathrm{GeV}$ are also shown scaled by a factor of 1000. These two BDT input variables are used as well in the photon-enriched sample, although their discrimination power is limited given the ${\mathrm{O}}_{\mathrm{NN}}>0.75$ requirement applied to both photons.

Figure 9

Distribution of the BDT output used in the SM Higgs boson search in (a) the photon-enriched sample and (b) the jet-enriched sample. The data (points with statistical error bars) are compared to the background prediction, broken down into its individual components. The expected distributions for a SM Higgs boson with $M_H=125\mathrm{GeV}$ are also shown scaled by a factor of 10.

Figure 10

Distribution of BDT output used in the fermiophobic Higgs boson search in (a) the photon-enriched sample and (b) the jet-enriched sample. The data (points with statistical error bars) are compared to the background prediction, broken down into its individual components. The expected distributions for a fermiophobic Higgs boson with $M_{H_{\mathrm{f}}}=125\mathrm{GeV}$ are also shown scaled by a factor of 10.

Figure 11

Distribution of the BDT output for data (points with statistical error bars) after subtraction of the fitted background (under the background-only hypothesis) in (a) the photon-enriched sample and (b) the jet-enriched sample, for $M_H=125\mathrm{GeV}$. The expected SM Higgs signal is normalized to the observed limit on $\sigma \times \mathcal{B}(H\to \gamma \gamma )$. The bands represent the 1 s.d. uncertainties on the background prediction resulting from the fit.

Figure 12

(a) Observed and expected 95% C.L. limits on the ratio of $\sigma \times \mathcal{B}(H\to \gamma \gamma )$ to the SM prediction as a function of $M_H$. The bands correspond to 1 and 2 s.d. around the median expected limit under the background-only hypothesis. (b) Observed LLR as a function of $M_H$ compared to the expected LLR under the background-only hypothesis (${\mathrm{LLR}}_{\mathrm{b}}$) and signal+background hypothesis (${\mathrm{LLR}}_{\mathrm{s}+\mathrm{b}}$). The bands correspond to the 1 s.d. and 2 s.d. around the expected median ${\mathrm{LLR}}_{\mathrm{b}}$.

Figure 13

(a) Observed and expected 95% C.L. limits on the ratio of $\sigma \times \mathcal{B}(H\to \gamma \gamma )$ to the fermiophobic Higgs model prediction as a function of $M_{H_{\mathrm{f}}}$. The bands correspond to 1 and 2 s.d. around the median expected limit under the background-only hypothesis. (b) Observed and expected 95% C.L. limits on $\mathcal{B}(H_{\mathrm{f}}\to \gamma \gamma )$ as a function of $M_{H_{\mathrm{f}}}$. The bands correspond to the 1 and 2 s.d. around the median expected limit under the background-only hypothesis. Also shown is the prediction for a fermiophobic Higgs boson.