HIG-12-015

"Evidence for a new state decaying into two photons in the search for the standard model Higgs boson in pp collisions"

This is a condensed description with plots of the analysis HIG-12-015

Table of contents

Abstract

A search for a Higgs boson decaying into two photons is described. The analysis is performed using datasets recorded by the CMS experiment at the LHC from pp collisions at a centre-of-mass energies of 7 TeV and 8 TeV. The datasets correspond to an integrated luminosity of 5.1 plus 5.3 fb−1, respectively. Limits are set on the cross section of the standard model Higgs boson decaying to two photons. An excess of events above the expected standard model background is observed for a Higgs boson mass hypothesis of 125 GeV, with a local significance of 4.1σ. The global significance of observing an excess with a local significance ≥ 4.1σ anywhere in the search range 110–150 GeV is estimated to be 3.2σ. This result constitutes evidence for the existence of a new state that decays into two photons.

Further details

The search combines the results of the Multi-Variate Analysis described in: HIG-12-001 applied to the 2011 data sample after ECAL recalibration and some other minor changes with the results of a very similar analysis applied to 2012 data. Also the cut based analysis described in CMS-HIG-11-033 (published in Physics Letters B) has been repeated as a cross check.

In the MVA analysis, to maximise the sensitivity of the search, selected diphoton events are subdivided into classes according to the output value of a diphoton MVA which is designed to have the following properties:

• it classifies events with signal-like kinematic characteristics with a high score,
• it classifies good diphoton mass resolution events with a high score,
• it classifies events with a high score from the photon identification MVA with a high score,
• it should be mass independent, it should not select events according to the invariant mass.

Five(Six) mutually exclusive event classes are defined, four defined by the diphoton MVA output, and one(two) additional ones into which are put events containing a pair of jets passing selection requirements which are designed to select Higgs bosons produced by the vector boson fusion process. While in 2011 we had a single class for the dijet tag, in 2011 we defined two classes based on the transverse jet momenta and on the dijet mass in order to enanche the sensitivity and the separation between the VBF and gg-fusion contribution.

The background model is obtained by fitting polynomials to the observed diphoton mass distributions in each of the five event classes. A cross check analysis is performed using an alternative background model where the result is extracted from a fit to the output distribution of the MVA (mass-window MVA or sideband MVA) which has two inputs: the diphoton MVA output, and the mass.

The results of the cross check analysis and of the cut based analysis are consistent with those obtained from the mass fit.

Figures and tables from the PAS

Tables

Image	File links	Description
	png pdf	Table 2: Expected number of SM Higgs boson events (mH=125 GeV) and estimated background (at mγγ =125 GeV) for all event classes of the 7 and 8 TeV datasets. The composition of the SM Higgs boson signal in terms of the production processes and its mass resolution is also given. Further information can be found here.
	png pdf	Table 3: Separate sources of systematic uncertainties accounted for in the analysis of the 8 TeV data set. The magnitude of the variation of the source that has been applied to the signal model is shown.

Figures

Image	File links	Description
	png pdf	Figure 1 a: The result of signal plus background model fit to the mγγ distribution for the five event classes of the 7 TeV dataset.7 TeV mass spectrum in class 0 of MVA based analysis.
	png pdf	Figure 1 b: The result of signal plus background model fit to the mγγ distribution for the five event classes of the 7 TeV dataset.7 TeV mass spectrum in class 1 of MVA based analysis.
	png pdf	Figure 1 c: The result of signal plus background model fit to the mγγ distribution for the five event classes of the 7 TeV dataset.7 TeV mass spectrum in class 2 of MVA based analysis.
	png pdf	Figure 1 d: The result of signal plus background model fit to the mγγ distribution for the five event classes of the 7 TeV dataset.7 TeV mass spectrum in class 3 of MVA based analysis.
	png pdf	Figure 1 e: The result of signal plus background model fit to the mγγ distribution for the five event classes of the 7 TeV dataset.7 TeV mass spectrum in class 4 (dijet tag) of MVA based analysis.

Image	File links	Description
	png pdf	Figure 2 a: The result of signal plus background model fit to the mγγ distribution for the six event classes of the 8 TeV dataset.8 TeV mass spectrum in class 0 of MVA based analysis.
	png pdf	Figure 2 b: The result of signal plus background model fit to the mγγ distribution for the six event classes of the 8 TeV dataset.8 TeV mass spectrum in class 1 of MVA based analysis.
	png pdf	Figure 2 c: The result of signal plus background model fit to the mγγ distribution for the six event classes of the 8 TeV dataset.8 TeV mass spectrum in class 2 of MVA based analysis.
	png pdf	Figure 2 d: The result of signal plus background model fit to the mγγ distribution for the six event classes of the 8 TeV dataset.8 TeV mass spectrum in class 3 of MVA based analysis.
	png pdf	Figure 2 e: The result of signal plus background model fit to the mγγ distribution for the six event classes of the 8 TeV dataset.8 TeV mass spectrum in class 4 (Tight dijet tag) of MVA based analysis.
	png pdf	Figure 2 f: The result of signal plus background model fit to the mγγ distribution for the six event classes of the 8 TeV dataset.8 TeV mass spectrum in class 5 (Loose dijet tag) of MVA based analysis.

Image	File links	Description
	png pdf	Figure 3 a: For the mass fit MVA, the exclusion limit on the cross section of a SM Higgs boson decaying into two photons as a function of the boson mass relative to the SM cross section, where the theoretical uncertainties on the cross section have been included in the limit setting. The limit is calculated using the modified frequentist CLS method. Result for the 7 TeV dataset.
	png pdf	Figure 3 b: For the mass fit MVA, the exclusion limit on the cross section of a SM Higgs boson decaying into two photons as a function of the boson mass relative to the SM cross section, where the theoretical uncertainties on the cross section have been included in the limit setting. The limit is calculated using the modified frequentist CLS method. Result for the 8 TeV dataset.

Image	File links	Description
	png pdf	Figure 4 a: 95% CL limits on the cross section of a Higgs boson decaying to two photons relative to the SM expectation for the combined 7 and 8 TeV datasets. Mass fit MVA analysis.
	png pdf	Figure 4 b: 95% CL limits on the cross section of a Higgs boson decaying to two photons relative to the SM expectation for the combined 7 and 8 TeV datasets. Cut based analysis.
	png pdf	Figure 4 c: 95% CL limits on the cross section of a Higgs boson decaying to two photons relative to the SM expectation for the combined 7 and 8 TeV datasets. Mass window MVA analysis.

Image	File links	Description
	png pdf	Figure 4 a: Observed local p-values as a function of mH obtained with the mass fit MVA analysis.
	png pdf	Figure 4 b: Observed local p-values as a function of mH obtained with the cut based analysis.
	png pdf	Figure 4 c: Observed local p-values as a function of mH obtained with the mass window MVA analysis.

Image	File links	Description
	png pdf	Figure 6 a: The best fit signal strength relative to the SM Higgs boson cross section. Scan as a function of the Higgs boson mass hypothesis.
	png pdf	Figure 6 b: The best fit signal strength, (σ/σSM), for the combined fit to the five classes (vertical line) and for the individual contributing classes (points) for the hypothesis of a SM Higgs boson mass of 125.0 GeV. The band corresponds to +-1 sigma uncertainties on the overall value. The horizontal bars indicate +-1 sigma uncertainties on the values for individual classes. The symbol σ/σSM denotes the production cross section times the relevant branching fractions, relative to the SM expectation. The actual values plotted are tabulated here.

Image	File links	Description
	png pdf	Figure 7 a: For the mass fit MVA and combined datasets, result of summing the data and the signal plus background fits weighted by the ratio of signal to background in each event class as shown in Table 2.
	png pdf	Figure 7 b: For the mass fit MVA and combined datasets, result of summing the data and the signal plus background fits weighted by the ratio of signal to background in each event class as shown in Table 2. In the background-subtracted plot, the uncertainties shown around the axis correspond to the sum in quadrature of the estimated uncertainties on the background and the weighted data.

Additional plots

Data sample

Image	File links	Description
	png pdf	Distribution of the number of reconstructed vertices (using the deterministic annealing algorithm) for events with a Z decaying to 2 muons in the data (points) and MC after having applied the reweighing on the number of simulated in-time pile-up events (filled histogram). The distributions are shown for the full 2011 running period at a CM energy of 7 TeV.
	png pdf	Distribution of the number of reconstructed vertices (using the deterministic annealing algorithm) for events with a Z decaying to 2 muons in the data (points) and MC after having applied the reweighing on the number of simulated in-time pile-up events (filled histogram). The distributions are shown for the full 2012 running period at a CM energy of 8 TeV.

Energy resolution and scale in 2012

Image	File links	Description
	png pdf	The energy resolution in different regions of the detector is investigated by smearing the energies of electrons in MC simulated Z→ee and comparing the resulting Z mass distribution with what is found in data. For electrons with R9>0.94 in the central part of the barrel (η<1), the required additional smearing is very small. This is illustrated in the plot.

Vertex Id in 2012

Image	File links	Description
	png pdf	The performance of the vertex identification algorithm is measured with Z to mumu events for the diphotn events with unconverted photons. The muon tracks can be used to identify the hard interaction vertex and they are subsequently removed from the collection of tracks used in the vertex reconstruction algorithm to mimic the topology of a Higgs boson decaying into two photons. The comparison of the vertex identification efficiency between data and simulation as a function of the Z boson pT is shown.
	png pdf	MC comparison between the true vertex identification efficiency and the average estimated vertex probability as a function of the reconstructed diphoton pT. The plots are obtained on signal events with mH =120 GeV selected by the MVA analysis.

Additional plots for the MVA based analyses (mass fit and sideband)

Image	File links	Description
	png pdf	Flowchart of MVA based analysis

Photon Id MVA on the 2012 8 TeV sample

Image	File links	Description
	png pdf	Photon ID MVA output distribution for Z->ee events in data and Monte Carlo, for photon candidates in the ECAL barrel (left) and endcaps (right). The systematic uncertainty applied to the Monte Carlo shape, corresponding to a shift of ±0.01 in the value of the MVA output is shown by the hashed region.

Diphoton Id MVA for the 2012 8 TeV sample

Image	File links	Description
	png pdf eps	Diphoton ID MVA output distribution trained on the 2012 8 TeV sample for data and BG MC. The shaded region is not used in the analysis and the discrepancy between data and MC at low values of the diphoton BDT output can be explained by the higher rate of fake photons in data that is due to the generator level enrichment filters used to simulate the background. The vertical dashed lines indicate the class bounds.
	png pdf	Diphoton ID MVA output distribution trained on the 2012 8 TeV sample for signal and BG MC. The vertical dashed lines indicate the class bounds.

Relating the diphoton MVA output for MC simulated signal, and for data, to the cut based event classes

Image	File links	Description
	png pdf	Diphoton MVA output distribution for events in a simulated Higgs boson signal (mH = 125 GeV) where pT(γγ)>40 GeV. The events are shown classified according to the eta/R9 event classes used in the cut based analysis. The event class boundaries based on the diphoton MVA output are shown as vertical blue lines, and events in the blue shaded region are not used in the statistical treatment, limit setting etc. Note that not all of the events shown in the plot would have entered the cut based event classes since the event selection used for that analysis was tighter than the preselection used for the multivariate analysis.
	png pdf	Diphoton MVA output distribution for events in a simulated Higgs boson signal (mH = 125 GeV) where pT(γγ)<40 GeV. The events are shown classified according to the eta/R9 event classes used in the cut based analysis. The event class boundaries based on the diphoton MVA output are shown as vertical blue lines, and events in the blue shaded region are not used in the statistical treatment, limit setting etc. Note that not all of the events shown in the plot would have entered the cut based event classes since the event selection used for that analysis was tighter than the preselection used for the multivariate analysis.
	png pdf	Diphoton MVA output distribution for events in data (100 < mgg < 180 GeV). Both photons in the events have additionally been required to pass the cut based selection requirements. The events are shown classified according to the eta/R9 event classes used in the cut absed analysis. The event class boundaries based on the diphoton MVA output are shown as vertical blue lines, and events in the blue shaded region are not used in the statistical treatment, limit setting etc.

Efficiency x acceptance of the MVA analysis on the 2012 8 TeV sample

Image	File links	Description
	png pdf	Signal efficiency × acceptance for the MVA based analysis as a function of mass hypothesis (all event classes combined). The band indicates 1 sigma systematic error.

Diphoton mass spectrum data-MC comparison for the MVA analysis

Image	File links	Description
	png pdf	Diphoton mass distribution for data (data points), and Monte Carlo simulation of SM processes which constitute the background to the search (histograms) for the 2011 7 TeV data sample. A simulated signal for a Higgs boson with a mass of 120 GeV is shown by the red outline histogram. The magnitude of the signal is what would be expected if its cross section were 7 times the SM expectation.The shaded band represents the theoretical (k-factor) uncertainty on the the MC prediction. Note that in the analysis the background is evaluated from a fit to the data without reference to the MC simulation.
	png pdf	Diphoton mass distribution for data (data points), and Monte Carlo simulation of SM processes which constitute the background to the search (histograms) for the 2012 8 TeV data sample. A simulated signal for a Higgs boson with a mass of 120 GeV is shown by the red outline histogram. The magnitude of the signal is what would be expected if its cross section were 7 times the SM expectation.The shaded band represents the theoretical (k-factor) uncertainty on the the MC prediction. Note that in the analysis the background is evaluated from a fit to the data without reference to the MC simulation.

Higgs boson mass resolution 2011 7 TeV sample

Image	File links	Description
	png pdf	7 TeV mass resolution in class 0 of MVA based analysis
	png pdf	7 TeV mass resolution in class 1 of MVA based analysis
	png pdf	7 TeV mass resolution in class 2 of MVA based analysis
	png pdf	7 TeV mass resolution in class 3 of MVA based analysis
	png pdf	7 TeV mass resolution in class 4 (dijet tag) of MVA based analysis
	png pdf	Combined 7 TeV mass resolution of MVA based analysis

Higgs boson mass resolution 2012 8 TeV

Image	File links	Description
	png pdf	8 TeV mass resolution in class 0 of MVA based analysis
	png pdf	8 TeV mass resolution in class 1 of MVA based analysis
	png pdf	8 TeV mass resolution in class 2 of MVA based analysis
	png pdf	8 TeV mass resolution in class 3 of MVA based analysis
	png pdf	8 TeV mass resolution in class 4 (Tight dijet tag) of MVA based analysis
	png pdf	8 TeV mass resolution in class 5 (Loose dijet tag) of MVA based analysis
	png pdf	Combined 8 TeV mass resolution of MVA based analysis

Combined mass spectrum in 2011 7 TeV and 2012 8 TeV sample

Image	File links	Description
	png pdf	Combined mass spectrum in 2011 7 TeV and 20128 TeV sample of MVA based analysis

Additional plots and for the mass fit MVA based analysis

Fitted signal strength as function of mass

Image	File links	Description
	png pdf	The best fit signal strength, in terms of the standard model Higgs boson cross section, as function of mass, the green band indicates the error.
	png pdf	The best fit signal strength, in terms of the standard model Higgs boson cross section, as function of mass, the green band indicates the error.

Fitted signal strength in individual classes

	png pdf	The best fit signal strength, (σ/σSM), for the individual classes for the hypothesis of a SM Higgs boson mass of 125.0 GeV. Note that these are the values that appear in Fig. 6 b), and are the result of a fit where the signal strength in each event class is not constrained by the signal strength in any of the other class. The symbol σ/σSM denotes the production cross section times the relevant branching fractions, relative to the SM expectation.
	png pdf	The best fit signal strength, (σ/σSM), for the combined fit to the five classes (vertical line) and for the individual contributing classes (points) for the hypothesis of a SM Higgs boson mass of 136.0 GeV. The band corresponds to +-1 sigma uncertainties on the overall value. The horizontal bars indicate +-1 sigma uncertainties on the values for individual classes. The symbol σ/σSM denotes the production cross section times the relevant branching fractions, relative to the SM expectation.

P-values per class

Image	File links	Description
	png pdf	P-values for each class and combined for the mass fit MVA analysis on the 2011 7 TeV data sample.
	png pdf	P-values for each class and combined for the mass fit MVA analysis on the 2012 8 TeV data sample.

Additional plots for the mass sideband MVA based analysis

Image	File links	Description
	png pdf	Illustration of the definition of the signal region and sidebands for mass hypothesis mH=125 GeV for the 2012 8 TeV data sample.
	png pdf	Distribution of the data, background model and signal model across the optimized bins of the sideband MVA analysis for mH=125 GeV for the 2012 8 TeV data sample.
	png pdf	Distribution of the background subtracted data and signal model across the optimized bins of the sideband MVA analysis for mH=125 GeV for the 2012 8 TeV data sample. The error bar on the points also includes the uncertainty on the background.

95% CL exclusion limits for sideband MVA analysis

Image	File links	Description
	png pdf	7 TeV 95% CL exclusion for MVA based sideband analysis
	png pdf	8 TeV 95% CL exclusion for MVA based sideband analysis

Additional plots for the cut based analysis

More information relative to the cut based analysis based on 2011 7 TeV data can be found at: https://twiki.cern.ch/twiki/bin/view/CMSPublic/Hig11033TWiki

Efficiency x acceptance of the cut based analysis on the 2012 8 TeV sample

Image	File links	Description
	png pdf	Signal efficiency × acceptance for the cut based analysis as a function of mass hypothesis (all event classes combined). The band indicates 1 sigma systematic error.

Mass spectra in event classes for the cut based analysis 2011 7 TeV sample

Image	File links	Description
	png pdf	7 TeV mass spectrum in class 0 of cut based analysis
	png pdf	7 TeV mass spectrum in class 1 of cut based analysis
	png pdf	7 TeV mass spectrum in class 2 of cut based analysis
	png pdf	7 TeV mass spectrum in class 3 of cut based analysis
	png pdf	7 TeV mass spectrum in class 4 (dijet tag) of cut based analysis
	png pdf	Combined 7 TeV mass spectrum of cut based analysis

Mass spectra in event classes for the cut based analysis 2012 8 TeV

Image	File links	Description
	png pdf	Mass spectrum in class 0 of the cut based analysis for the 2012 8 TeV data sample.
	png pdf	Mass spectrum in class 1 of the cut based analysis for the 2012 8 TeV data sample.
	png pdf	Mass spectrum in class 2 of the cut based analysis for the 2012 8 TeV data sample.
	png pdf	Mass spectrum in class 3 of the cut based analysis for the 2012 8 TeV data sample.
	png pdf	Mass spectrum in class 4 (tight dijet tag) of the cut based analysis for the 2012 8 TeV data sample.
	png pdf	Mass spectrum in class 5 (loose dijet tag) of the cut based analysis for the 2012 8 TeV data sample.

95% CL exclusion limits of the cut based analysis

Image	File links	Description
	png pdf	95% CL exclusion limit relative to the SM Higgs boson cross section calculated in the asymptotic approximation for the cut based analysis on the 2011 7 TeV data sample. The theoretical uncertainties on the cross section have been included in the limit setting.
	png pdf	95% CL exclusion limit relative to the SM Higgs boson cross section calculated in the asymptotic approximation for the cut based analysis on the 2012 8 TeV data sample. The theoretical uncertainties on the cross section have been included in the limit setting.

Fitted signal strength as function of mass

Image	File links	Description
	png pdf	The best fit signal strength, in terms of the standard model Higgs boson cross section, as function of mass, the green band indicates the error.
	png pdf	The best fit signal strength, in terms of the standard model Higgs boson cross section, as function of mass, the green band indicates the error.
	png pdf	The best fit signal strength, in terms of the standard model Higgs boson cross section, as function of mass, the green band indicates the error.

Fitted signal strength in individual classes

png pdf

The best fit signal strength, in terms of the standard model Higgs boson cross section, for the combined fit to the five classes (vertical line) and for the individual contributing classes (points) for the hypothesis of a SM Higgs boson mass of 125.0 GeV. The band corresponds to +-1 sigma uncertainties on the overall value. The horizontal bars indicate +-1 sigma uncertainties on the values for individual classes.

P-values per class

Image	File links	Description
	png pdf	P-values for each class and combined for the cut based analysis on the 2011 7 TeV data sample.
	png pdf	P-values for each class and combined for the cut based analysis on the 2012 8 TeV data sample.

Event displays

2012 Diphoton Event:
https://cdsweb.cern.ch/record/1459459

Event parameters:
mass=125.9 GeV
ptgam1=89.8 GeV
ptgam2=46.5 GeV
eta1=0.06
eta2=-0.81
sigmaM/M=0.89%
ptgammagamma=78.4 GeV

Image	File links	Description
	png pdf	Standard 3D view
	png pdf	Standard Rho-Phi view
	png pdf	Standard Rho-Z view

	png pdf	Standard 3D view
	png pdf	Standard Rho-Phi view
	png pdf	Standard Rho-Z view

-- MarcoPieri - 29-Jun-2012