HIG-11-021

"Search for a Higgs boson decaying into two photons in the CMS detector"

This is a condensed description with plots for the analysis CMS-HIG-11-021

Abstract

This note describes the search for a Higgs boson decaying into diphotons in pp collisions at the LHC at a center-of-mass energy of 7~TeV. Data-driven methods are used to evaluate the reconstruction and selection performance for the signal, and the background is evaluated from a fit to the data. The analysis is performed on a dataset corresponding to 1.66 fb^-1 of data recorded in 2011 by the CMS experiment. Limits are set on the cross section of a Standard Model Higgs boson decaying to 2 photons, and on the cross section of a fermiophobic Higgs boson decaying to 2 photons. The expected exclusion limit at 95% CL is between 2.7 and 4.7 times the Standard Model cross section, and the observed limit fluctuates between about 1.3 and 8 times the Standard Model cross section. For the fermiophobic model, the expected exclusion limit at 95% CL covers the mass range between 110-116.5 GeV/c², while the data excludes only the mass range 110-112 GeV/c².

Main Results

A search has been made for a Higgs boson decaying into 2 photons in the CMS detector. The analysis uses reconstruction and selection techniques that take account of the multiple interactions occurring each bunch crossing in the current LHC operating conditions, which affect isolation and identification of the interaction vertex. The selected events are subdivided into classes according to indicators of mass resolution and signal to background ratio, and the results of a search in each class are combined. The expected exclusion limit at 95% CL is between 2.7 and 4.7 times the Standard Model cross section, and the observed limit fluctuates between about 1.3 and 8 times the Standard Model cross section. For the fermiophobic model, the expected exclusion limit at 95% CL covers the mass range between 110-116.5 GeV/c², while the data excludes only the mass range 110-112 GeV/c².

The small excess of events around 140 GeV/c², that causes the observed SM limit to fluctuate by more than 2 sigma above the expected limit, has been investigated in detail. The local p-value, measuring the probability that this excess is due to background, has been calculated to be 2.5 +- 0.3 x 10^-3. This calculation assumes that the relative signal strength between the event classes follows the Monte Carlo signal model for the Standard Model. The most probable signal strength is found to be 4.5 +1.9 -1.7 times the Standard Model cross section. Since background fluctuations can occur at any point in the searched mass range, a global evaluation of the p-value is performed and the global significance of a localized excess as the one observed found to be 1.6 sigma. This corresponds to a trials factor, or look-elsewhere effect, of about 20. It is further noted that almost no excess is seen in the event class which would be expected to contribute the most significance in the case of a signal.

Figures from CMS-HIG-11-021

PAS.figure	Image	File	Description
Fig. 1		pdf png	Fraction of Higgs boson vertices found within 10 mm of their true location, for a Monte Carlo signal sample (mH = 120 GeV/c2), as a function of the Higgs boson transverse momentum. The distribution of the number of interactions per bunch crossing (nPU) in the Monte Carlo is adjusted to be the same as in the data by weighting the events.
Fig. 2		pdf png	Diphoton mass distribution for data (data points), and Monte Carlo simulation of SM processes which constitute the background to the search (histograms). The uncertainty on the cross sections of the contributing processes is estimated to be about 15%.
Fig. 3	Background model fits to the pT(γγ) < 40 GeV/c event classes
		pdf png	Both photons in barrel, min(R9) > 0.94
		pdf png	Both photons in barrel, min(R9) < 0.94
		pdf png	One or more photon in endcap, min(R9) > 0.94
		pdf png	One or more photon in endcap, min(R9) < 0.94
Fig. 4	Background model fits to the pT(γγ) > 40 GeV/c event classes
		pdf png	Both photons in barrel, min(R9) > 0.94
		pdf png	Both photons in barrel, min(R9) < 0.94
		pdf png	One or more photon in endcap, min(R9) > 0.94
		pdf png	One or more photon in endcap, min(R9) < 0.94
Fig. 5		pdf png	Background model fit to the mγγ distribution for the combined data in all 8 event classes.
Fig. 6	Signal model fits to a simulated signal (mH=120 GeV/c2) for the pT(γγ) < 40 GeV/c event classes
		pdf png	Both photons in barrel, min(R9) > 0.94
		pdf png	Both photons in barrel, min(R9) < 0.94
		pdf png	One or more photon in endcap, min(R9) > 0.94
		pdf png	One or more photon in endcap, min(R9) < 0.94
Fig. 7	Signal model fits to a simulated signal (mH=120 GeV/c2) for the pT(γγ) > 40 GeV/c event classes
		pdf png	Both photons in barrel, min(R9) > 0.94
		pdf png	Both photons in barrel, min(R9) < 0.94
		pdf png	One or more photon in endcap, min(R9) > 0.94
		pdf png	One or more photon in endcap, min(R9) < 0.94
Fig. 8		pdf png	Simulated signal (mH=120 GeV/c2) for all 8 event classes combined. The line shows the weighted sum of fits to the signal models in the individual classes.
Fig. 9		pdf png	Exclusion limit on the cross section of a SM Higgs boson decaying into 2 photons as a function of the boson mass.
Fig. 10		pdf png	Exclusion limit on cross section of a SM Higgs boson decaying into 2 photons relative to the SM cross section, as a function of the boson mass.
Fig. 11		pdf png	Exclusion limit on the cross section of a fermiophobic Higgs boson decaying into 2 photons as a function of the boson mass.
Fig. 12		pdf png	Exclusion limit on the cross section of a Higgs boson decaying into 2 photons relative to the cross section predicted in the fermiophobic model, as a function of the boson mass.

Additional approved plots and information for public talks

Image	File.link	Description
	linpdf linpng logpdf logpng	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 reweighting on the number of simulated in-time pile-up events (filled histogram). Both linear and logarithmic scales are available.
	pdf png	Signal efficiency for mH=120 GeV/c2 as a function of as a function of pT, for each of the four photon categories.
	pdf png	Signal efficiency for mH=120 GeV/c2 as a function of as a function of pseudorapidity, for each of the four photon categories.
	pdfa pnga pdfb pngb	Invariant mass distribution of Z \to ee events with a) both electrons in the barrel and R9>0.94, b) one in the barrel and one in the endcap, and both with R9<0.94. Points are data, the red histogram is MC, while the black empty histogram is after energy smearing has been applied to the MC.
	pdf png	Mass distribution for a SM Higgs boson (mH=120 GeV/c2) decaying to 2 photons reconstructed using the nominal interaction point (black histogram), the vertex whose associated tracks have the highest summed pT2 (red histogram), and the vertex chosen by the selection algorithm used in the analysis (blue histogram). (EPS approved version of the plot)
	pdf png	Efficiency x Acceptance as a function of mass. The band indicates 1 sigma systematic error. The variation of the Efficiency x Acceptance is due mainly to the fixed pT thresholds on the photons. The increase of the photon selection efficiency with pT also contributes.
	pdf png	The observed p-value as a function of mass (black line) for the hypothesis of a Standard Model Higgs boson decaying to 2 photons. The blue dashed line shows the median expected p-value for a SM Higgs boson at each tested mass. The look-elsewhere effect, which reduces the probability of the minimum near 140 GeV/c2 by a factor of about 20, is not included in the p-value.
	pdf png	The observed p-value as a function of mass (black line) for the hypothesis of a fermiophobic Higgs boson decaying to 2 photons. The blue dashed line shows the median expected p-value for a fermiophobic Higgs boson at each tested mass. The look-elsewhere effect, which reduces the probability of the minimum near 140 GeV/c2 by a factor of about 20, is not included in the p-value.

Background model fits to the pT(γγ) < 40 GeV/c event classes. Version with 120 GeV/c2 signal *5 overlayed
	pdf png	Both photons in barrel, min(R9) > 0.94
	pdf png	Both photons in barrel, min(R9) < 0.94
	pdf png	One or more photon in endcap, min(R9) > 0.94
	pdf png	One or more photon in endcap, min(R9) < 0.94
Background model fits to the pT(γγ) > 40 GeV/c event classes. Version with 120 GeV/c2 signal *5 overlayed
	pdf png	Both photons in barrel, min(R9) > 0.94
	pdf png	Both photons in barrel, min(R9) < 0.94
	pdf png	One or more photon in endcap, min(R9) > 0.94
	pdf png	One or more photon in endcap, min(R9) < 0.94
	pdf png	Background model fit to the mγγ distribution for the combined data in all 8 event classes. Version with 120 GeV/c2 signal *5 overlayed.

• Using the profile likelihood approximation the effect of different changes on the expected limit for both a SM Higgs boson and a fermiophobic Higgs boson, decaying to 2 photons have been investigated (m_H=120 GeV/c²). It can been seen that:
  • the use of event classes brings a 20% gain in sensitivity (SM)
  • removing the endcaps reduces the sensitivity by 5% (SM)
  • the use of the diphoton p_T classification adds 60% sensitivity for the fermiophobic model
  • the expected limit will improve by more than 15%, when the ECAL transparency corrections and intercalibration is mastered to the level achieved during the period in which the first 200/pb of data were taken in 2011 (i.e. no need for MC resolution smearing)

• The median expected limit set for each event class on it's own using the profile likelihood approximation (Standard Model):
  

• The median expected limit set for each event class on it's own using the profile likelihood approximation (fermiophobic model):