Search for diphoton resonances in the 66 to 110 GeV mass range using $pp$ collisions at $\sqrt{s}=13$ TeV with the ATLAS detector

A search is performed for light, spin-0 bosons decaying into two photons in the 66 to 110 GeV mass range, using 140 fb$^{-1}$ of proton-proton collisions at $\sqrt{s}=13$ TeV produced by the Large Hadron Collider and collected by the ATLAS detector. Multivariate analysis techniques are used to define event categories that improve the sensitivity to new resonances beyond the Standard Model. A model-independent search for a generic spin-0 particle and a model-dependent search for an additional low-mass Higgs boson are performed in the diphoton invariant mass spectrum. No significant excess is observed in either search. Mass-dependent upper limits at the $95\%$ confidence level are set in the model-independent scenario on the fiducial cross-section times branching ratio into two photons in the range of 8 fb to 53 fb. Similarly, in the model-dependent scenario upper limits are set on the total cross-section times branching ratio into two photons as a function of the Higgs boson mass in the range of 19 fb to 102 fb.

10 July 2024

Contact: Higgs conveners internal

JHEP 01 (2025) 053

e-print arXiv:2407.07546 - pdf from arXiv

Inspire record

Figures | Tables | Auxiliary Material

Figures

Figure 01a

(a) Distribution of electron--photon ambiguity BDT scores constructed by taking the minimum score of the two photon candidates in simulated ggF m_X = 100 GeV signal events (solid line), Z/γ^* → ee events (dashed line), and the continuum γγ background events (dotted line). The distributions are scaled independently for illustrative purposes. (b) Efficiency versus minimum requirement on the ambiguity BDT score, shown for the same samples. Only events in which both photon candidates are converted are shown.

png (48kB) pdf (11kB)

Figure 01b

Figure 02

Distributions of the category BDT scores for the merged SM-like Higgs boson considering all production modes (ggF, VBF, ttH, WH, ZH), the diphoton (γγ) and reducible backgrounds (γj, jj) continuum, and the simulated Z→ ee background prediction. The reducible background components are derived from dedicated data control regions where the photon identification and isolation requirements are inverted. The merged signal contains signals generated for m_H = 60, 80, 100, and 120 GeV and each MC sample is weighted according to the SM-like Higgs boson cross-section. Photons from all three conversion categories are used in the BDT training. The merged signal and backgrounds are separately normalised to unity. The vertical lines and arrows at category BDT scores of -0.2 and 0 define the categorisation used in this analysis. Events with category BDT scores below -0.2 are in BDT 1, events with category BDT scores between -0.2 and 0 are in BDT 2, and events with category BDT scores above 0 are in BDT 3.

png (47kB) pdf (9kB)

Figure 03a

Simulated diphoton invariant mass distribution of a narrow-width signal particle H of mass 80 GeV (points) in the (a) UU3 and (b) CC1 categories, overlaid with the DSCB function resulting from the signal model parameterisation (line). The error bars on the simulated data points indicate the statistical uncertainties. An arbitrary normalisation is used for illustration purposes.

png (130kB) pdf (17kB)

Figure 03b

Figure 04a

Background-only fit to the data (black markers) as a function of the diphoton invariant mass m_γγ for the model-independent conversion categories: (a) UU, (b) UC, and (c) CC. The solid lines show the sum of the Drell-Yan and the continuum background components and the dashed lines show only the continuum background components. The difference between the data and the total background component divided by the statistical uncertainty of the data, σ_data, is shown at the bottom panel separately for each category. The green (yellow) bands denote the total uncertainty in the background model at one (two) standard deviation.

png (136kB) pdf (41kB)

Figure 04b

Figure 04c

Figure 05a

Background-only fit to the data (black markers) as a function of the diphoton invariant mass m_γγ for each of the model-dependent BDT 3 categories: (a) UU3, (b) UC3, and (c) CC3. The solid lines show the sum of the Drell-Yan and the continuum background components and the dashed lines show only the continuum background components. The difference between the data and the total background component divided by the uncertainty, with σ_data denoting only the statistical error of the data, is shown at the bottom panel separately for each category. The green (yellow) bands denote the total uncertainty in the background model at one (two) standard deviation.

png (138kB) pdf (44kB)

Figure 05b

Figure 05c

Figure 06a

(a) Compatibility of the data, in the model-independent search, in terms of local p-value (solid line), with the background-only hypothesis as a function of the assumed NWA signal mass m_X. The dotted-dashed lines correspond to the standard deviation quantification σ. (b) 95% CL upper limits on the fiducial cross-section times branching ratio B(X→γγ) as a function of NWA m_X, where the solid (dashed) line corresponds to the observed (expected) limit and the green (yellow) band corresponds to one (two) standard deviation from the expectation.

png (42kB) pdf (10kB)

Figure 06b

Figure 07

Compatibility of the data with the background-only hypothesis, using the local p₀ quantified in units of standard deviations, σ, as a function of the assumed signal mass m_X and of the relative width Γ_X/m_X for the model-independent search.

png (75kB) pdf (20kB)

Figure 08a

(a) Expected and (b) observed limits on the fiducial cross-section times branching ratio B(X→γγ) computed using the asymptotic approximation as a function of the assumed signal mass m_X and relative width Γ_X/m_X for the model-independent scalar resonance search.

png (77kB) pdf (11kB)

Figure 08b

Figure 09a

(a) Compatibility of the data with the background-only hypothesis as quantified by the local p-value (solid line) as a function of the assumed signal mass m_H, for the model-dependent search. The dotted-dashed lines correspond to the standard deviation quantification σ. (b) 95% CL upper limits on the total cross-section times branching ratio B(H→γγ) as a function of m_H, where the solid (dashed) line corresponds to the observed (expected) limit and the green (yellow) band corresponds to one (two) standard deviation from the expectation.

png (42kB) pdf (10kB)

Figure 09b

Tables

Table 01

The selection requirements and names of each category in the model-independent and model-dependent analyses.

png (20kB) pdf (49kB)

Table 02

The expected number of signal events, fractions of each Higgs boson production mode, and the number of background events per GeV at m_γγ = 90 GeV for each BDT category, and for all three photon conversion categories together. The per GeV binning corresponds to approximately 1σ of the mass resolution at m_γγ = 90 GeV. The background events are extracted from the background-only fit to the data and the "Total" category includes the number of Drell-Yan events (DY) that are also shown separately. The BDT categories are defined as follows: events with category BDT scores below -0.2 are in BDT category 1, events with category BDT scores between -0.2 and 0 are in BDT category 2, and events with category BDT scores above 0 are in BDT category 3.

png (21kB) pdf (66kB)

Table 03

The number of data events (N_data), the expected fraction of γγ, γj, jj events determined with the two dimensional side band method, and the fraction of Drell-Yan events in each category. The uncertainties in the fractions of γγ, γj, jj arise from the statistical uncertainty varying the identification requirements. The BDT categories are defined as follows: events with category BDT scores below -0.2 are in BDT Category 1, events with category BDT scores between -0.2 and 0 are in BDT Category 2, and events with category BDT scores above 0 are in BDT Category 3.

png (49kB) pdf (53kB)

Table 04

Summary of the systematic uncertainties considered in this analysis. In general, the values correspond to the uncertainties associated to the fit nuisance parameters. The DY uncertainty is the percent error on the nominal peak position, peak width, and normalisation. The spurious signal uncertainty is expressed as a number of events and relative to the expected statistical uncertainty (δS) of a fitted signal. The "Remarks" column indicates specific information about the systematic uncertainty, including whether or not the uncertainty varies as a function of resonance mass or analysis category.

png (75kB) pdf (95kB)

Auxiliary material

Figure 01a

Distribution of electron–photon ambiguity BDT scores for the leading (x-axis) and subleading (y-axis) photon candidates in simulated (a) m_X = 100 GeV ggF signal events, (b) QCD-initiated continuum diphoton background events, and (c) Z → ee events.

png (91kB) pdf (10kB)

Figure 01b

Figure 01c

Figure 02a

Drell-Yan invariant mass templates derived from fully simulated Z/γ^*→ ee events reconstructed as ee (dashed), reconstructed as γγ events (square points), and reconstructed as ee after applying the Smirnov transform (circle points) in the (a) UU and (b) CC categories. The error bars on the simulated data points indicate the statistical uncertainties of each event sample. The envelopes determined by shifting the Drell-Yan background model by the uncertainties of its peak position and width are added in quadrature and indicated by the grey band.

png (127kB) pdf (16kB)

Figure 02b

Figure 03a

Background-only fit to the data (black markers) as a function of the diphoton invariant mass m_γγ for each of the model-dependent BDT 1 categories: (a) UU1, (b) UC1, and (c) CC1. The solid lines show the sum of the Drell-Yan and the continuum background components and the dashed lines show only the continuum background components. The difference between the data and the total background component divided by the uncertainty, with σ_data denoting only the statistical error of the data, is shown at the bottom panel separately for each category. The green (yellow) bands denote the total uncertainty in the background model at one (two) standard deviation.

png (134kB) pdf (52kB)

Figure 03b

Figure 03c

Figure 04a

Background-only fit to the data (black markers) as a function of the diphoton invariant mass m_γγ for each of the model-dependent BDT 2 categories: (a) UU2, (b) UC2, and (c) CC2. The solid lines show the sum of the Drell-Yan and the continuum background components and the dashed lines show only the continuum background components. The difference between the data and the total background component divided by the uncertainty, with σ_data denoting only the statistical error of the data, is shown at the bottom panel separately for each category. The green (yellow) bands denote the total uncertainty in the background model at one (two) standard deviation.

png (137kB) pdf (54kB)

Figure 04b

Figure 04c

Figure 05a

(a) The acceptance A_X in the fiducial volume as a function of the assumed signal mass m_X for different production modes. The dashed line in (a) represents the parameterisation of the acceptance for the nominal ggF production mode assumed in the model independent analysis. In (b) to (d), the correction factors are split between the different conversion categories used in the model dependent analysis. The parameterisation of the correction factor as a function of signal mass is shown for the (b) UU, (c) UC, and (d) CC conversion categories. The dashed lines in figures (b) to (d) correspond to the parameterisation of the nominal ggF production mode and the dashed-dotted lines correspond to the parameterisation of the maximal deviation from the nominal values. The error bars on the data points indicate the statistical uncertainties.

png (77kB) pdf (10kB)

Figure 05b

Figure 05c

Figure 05d

Figure 06a

The acceptance times correction factors used in the model dependent analysis are given for (a) UU BDT category 1, (b) UU BDT category 2, and (c) UU BDT category 3 for several Higgs boson production modes as a function of the assumed signal mass m_H. The dashed line represents the parameterisation of the acceptance times correction factor for the nominal SM-like Higgs boson signal. The error bars on the data points indicate the statistical uncertainties.

png (99kB) pdf (12kB)

Figure 06b

Figure 06c

Figure 07a

The acceptance times correction factors used in the model dependent analysis are given for (a) UC BDT category 1, (b) UC BDT category 2, and (c) UC BDT category 3 for several Higgs boson production modes as a function of the assumed signal mass m_H. The dashed line represents the parameterisation of the acceptance times correction factor for the nominal SM-like Higgs boson signal. The error bars on the data points indicate the statistical uncertainties.

png (108kB) pdf (12kB)

Figure 07b

Figure 07c

Figure 08a

The acceptance times correction factors used in the model dependent analysis are given for (a) CC BDT category 1, (b) CC BDT category 2, and (c) CC BDT category 3 for several Higgs boson production modes as a function of the assumed signal mass m_H. The dashed line represents the parameterisation of the acceptance times correction factor for the nominal SM-like Higgs boson signal. The error bars on the data points indicate the statistical uncertainties.

png (100kB) pdf (11kB)

Figure 08b

Figure 08c

Figure 09a

The acceptance times correction factors used in the model dependent analysis are given for (a) UU BDT category 1, (b) UU BDT category 2, and (c) UU BDT category 3 for the VBF+VH, ggF+ttH, and all Higgs boson production modes as a function of the assumed signal mass m_H. The dashed line represents the parameterisation of the acceptance times correction factor for the nominal SM-like Higgs boson signal. The error bars on the data points indicate the statistical uncertainties.

png (45kB) pdf (11kB)

Figure 09b

Figure 09c

Figure 10a

The acceptance times correction factors used in the model dependent analysis are given for (a) UC BDT category 1, (b) UC BDT category 2, and (c) UC BDT category 3 for the VBF+VH, ggF+ttH, and all Higgs boson production modes as a function of the assumed signal mass m_H. The dashed line represents the parameterisation of the acceptance times correction factor for the nominal SM-like Higgs boson signal. The error bars on the data points indicate the statistical uncertainties.

png (50kB) pdf (11kB)

Figure 10b

Figure 10c

Figure 11a

The acceptance times correction factors used in the model dependent analysis are given for (a) CC BDT category 1, (b) CC BDT category 2, and (c) CC BDT category 3 for the VBF+VH, ggF+ttH, and all Higgs boson production modes as a function of the assumed signal mass m_H. The dashed line represents the parameterisation of the acceptance times correction factor for the nominal SM-like Higgs boson signal. The error bars on the data points indicate the statistical uncertainties.

png (46kB) pdf (11kB)

Figure 11b

Figure 11c

Figure 12

Compatibility of the data with the background-only hypothesis as quantified by the local p-value as a function of the assumed signal mass m_X for the model-independent analysis. The compatibility is computed for the combined analysis including all three categories (black solid line), only considering the UU category (solid blue line), only considering the UC category (solid red line), and only considering the CC category (solid green line). The dotted-dashed lines correspond to the standard deviation quantification σ.

png (124kB) pdf (15kB)

Figure 13a

Compatibility of the data with the background-only hypothesis as quantified by the local p-value as a function of the assumed signal mass m_H for the model-dependent analysis. The compatibility is computed by splitting the analysis categories into BDT Category 1 (dotted line), BDT Category 2 (dashed line) and BDT Category 3 (solid line) for the (a) UU diphoton events (b) UC diphoton events and (c) CC diphoton events. The dotted-dashed lines correspond to the standard deviation quantification σ.

png (106kB) pdf (14kB)

Figure 13b

Figure 13c

Figure 14

The density distribution of the category BDT scores for the merged SM-like Higgs boson, separated into the distinct production modes considered in the analysis (ggF, VBF, ttH, VH). The merged signal is comprised of signals generated for m_H = 60, 80, 100, and 120 GeV and each MC sample is weighted according to the SM-like Higgs boson cross-section. Each signal production mode is separately normalised to unity.

png (111kB) pdf (9kB)

Figure 15a

95% CL upper limits on the total cross-section times branching ratio B(H→γγ) as a function of m_H, assuming production of the new additional Higgs boson via (a) the ggF and ttH processes, (b) the VH and VBF processes, (c) 100% VBF process, and (d) 100% VH process. The solid (dashed) line corresponds to the observed (expected) limit and the green (yellow) band corresponds to one (two) standard deviation from the expectation. The most significant excess is seen using the 100% VBF production hypothesis at a mass m_H=77.2 GeV, corresponding to a local significance of 2.5σ.

png (63kB) pdf (13kB)

Figure 15b

Figure 15c

Figure 15d

Table 01

Description of the input variables for the ambiguity BDT to discriminate between electrons and photons.

png (88kB) pdf (52kB)

Table 02

Dependence of the DSCB parameters describing the narrow-width model-independent scalar mass resolution model as a function of m_X [GeV] for the unconverted-unconverted category.

png (13kB) pdf (62kB)

Table 03

Dependence of the DSCB parameters describing the narrow-width model-independent scalar mass resolution model as a function of m_X [GeV] for the unconverted-converted category.

png (11kB) pdf (61kB)

Table 04

Dependence of the DSCB parameters describing the narrow-width model-independent scalar mass resolution model as a function of m_X [GeV] for the converted-converted category.

png (11kB) pdf (62kB)

Table 05

Dependence of the DSCB parameters describing the narrow-width model-dependent Higgs boson mass resolution model as a function of m_X [GeV] for the three BDT categories in the unconverted-unconverted conversion category.

png (26kB) pdf (63kB)

Table 06

Dependence of the DSCB parameters describing the narrow-width model-dependent Higgs boson mass resolution model as a function of m_X [GeV] for the three BDT categories in the unconverted-converted conversion category.

png (27kB) pdf (62kB)

Table 07

Dependence of the DSCB parameters describing the narrow-width model-dependent Higgs boson mass resolution model as a function of m_X [GeV] for the three BDT categories in the converted-converted conversion category.

png (26kB) pdf (62kB)