Study of the spin of the Higgs-like boson in the two photon decay channel using 20.7 fb-1 of pp collisions collected at √s = 8 TeV with the ATLAS detector

ATLAS-CONF-2013-029

10 March 2013

These preliminary results are superseded by the following paper:

HIGG-2013-01
ATLAS recommends to use the results from the paper.

Content

Preview

Main document (CDS record) - internal

pdf from CDS

Figures Auxiliary Material

Abstract

Studies of the spin of the Higgs-like boson in the $H \rightarrow \gamma\gamma$ channel with the ATLAS detector at the LHC are presented. The dataset being used corresponds to 20.7 fb-1 of pp collisions collected at $\sqrt{s} = 8$ TeV. The diphoton invariant mass and the angular distribution of the photons in the resonance rest frame are used to separate signal and background, and to distinguish between the tested spin hypotheses. The Standard Model spin-0 hypothesis is confronted with a graviton-like spin-2 model with minimal couplings. Two analyses are reported and lead to compatible results. The data are in good agreement with the spin-0 hypothesis. The hypothesis of a spin-2 particle produced by gluon fusion is excluded at 99% CL, which corresponds to the expected exclusion. Results for various admixtures of $q\bar{q}$ and $gg$ production of the spin-2 resonance are also presented.

Figures

Figure 01:
Observed numbers of events (dots) in the signal region as a function of |cos theta*|, overlaid with the projection of the signal (blue/dark band) and background (yellow/light band) components of the pdf obtained from the inclusive fit of the data in the nominal analysis under the spin-0 hypothesis.

png (51kB) eps (15kB) pdf (6kB)

Figure 02a:
Distributions of background-subtracted data in the signal region as a function of |cos theta*| for the nominal analysis. The two sets of points correspond to the subtraction of the different profiled background shapes in the case of the conditional spin-0 and spin-2 fits (assuming the spin-0/spin-2 |cos theta*| shapes). The spin-0 and spin-2 (produced by gluon fusion) pdfs (normalized to the fitted number of signal events) are overlaid. The cyan bands around the horizontal line at zero show the systematic uncertainties on the background modelling before the fits which, for the nominal analysis, includes the statistical uncertainty on the data sidebands. The error bars on the points reflect only the data statistics.

png (60kB) eps (17kB) pdf (8kB)

Figure 02b:
Distributions of background-subtracted data as a function of |cos theta*| for the alternative analysis. The two sets of points correspond to the subtraction of the different profiled background shapes in the case of the conditional spin-0 and spin-2 fits (assuming the spin-0/spin-2 |cos theta*| shapes). The spin-0 and spin-2 (produced by gluon fusion) pdfs (normalized to the fitted number of signal events) are overlaid. The cyan bands around the horizontal line at zero show the systematic uncertainties on the background modelling before the fits. The error bars on the points reflect only the data statistics.

png (59kB) eps (16kB) pdf (8kB)

Figure 03a:
Expected distributions of the test statistics q for the spin-0 and spin-2 (produced by gluon fusion) hypotheses for the nominal analysis. The observed value is indicated by a vertical line. The coloured areas correspond to the integrals of the expected distributions used to compute the p-values for the rejection of each hypothesis.

png (64kB) eps (18kB) pdf (7kB)

Figure 03b:
Expected distributions of the test statistics q for the spin-0 and spin-2 (produced by gluon fusion) hypotheses for the alternative analyses. The observed value is indicated by a vertical line. The coloured areas correspond to the integrals of the expected distributions used to compute the p-values for the rejection of each hypothesis.

png (61kB) eps (18kB) pdf (7kB)

Figure 04:
Median expected values of the test statistics ln (L(0)/L(2)) for the spin-0 (dashed blue/upper line and corresponding 68% and 95% CL bands) and spin-2 (dashed red/lower line) hypotheses, as well as the observed values (solid line) as a function of the fraction of qqbar production of the spin-2 signal (fqq).

png (68kB) eps (12kB) pdf (6kB)

Auxiliary figures and tables

Figure 05:
Expected distributions of |cos theta*| for the spin-0 signal (blue line), the spin-2 signal produced by gg (solid red line) or qqbar (dashed red line) and the observed distribution from background events (black line) in the invariant mass sidebands (105 GeV < mgg < 122 GeV and 130 GeV < mgg < 160 GeV).

png (53kB) eps (11kB) pdf (5kB)

Figure 06:
Difference between the number of observed events (nobs) in intervals of |cos theta*| × mgg = 0.1×1 GeV (excluding the signal region 122 < mgg < 130 GeV) and the number of expected events (nexp) divided by the associated uncertainty (sigma_exp) as a function of |cos theta*| and the invariant mass mgg, in GeV (left). The distribution of this quantity is shown on the right plot, overlaid with a normalised Gaussian distribution. The number of expected events in each interval is given by the product of the marginal pdfs of mgg and |cos theta*|, obtained by projecting the two-dimensional distribution of mgg × |cos theta*| in either dimension. The compatibility between the histogram and the curve, expressed in terms of the chi2 divided by the number of degrees of freedom (ndf) and the confidence level (CL), indicates that the data is consistent with the absence of correlations between |cos theta*| and mgg.

png (61kB) eps (35kB) pdf (11kB)

Figure 07:
Residual correlations between the invariant mass mgg and |cos theta*|, from the ratio of the background models of |cos theta*| in the alternative and the nominal analyses. The background model in the nominal analysis is given by the normalised distribution of |cos theta*|, built from all events belonging to the sidebands (105 GeV < mgg < 122 GeV and 130 GeV < mgg < 160 GeV). For the alternative analysis, the distribution corresponds to the estimated number of background events in the signal region (122 < mgg < 130 GeV) in each bin, normalised by their sum. The number of events in the signal region in each bin is estimated from a fit to the invariant mass sidebands, using exponentials of second degree polynomials for the first nine bins and a third degree polynomial for the last bin. The orange and blue bands correspond, respectively, to the statistical uncertainty from the sidebands and the total uncertainty on the background model in the nominal analysis. The total uncertainty includes, in addition to the statistical errors, systematic uncertainties from the residual correlations between the invariant mass mgg and |cos theta*| estimated from Monte Carlo simulations. The compatibility of the ratio with the line at one implies that no significant correlation between mgg and |cos theta*|, within the current precision, is observed in the data.

png (62kB) eps (17kB) pdf (7kB)

Figure 08a:
Diphoton invariant mass (mgg) distributions in each bin of |cos theta*|, overlaid with the sum of signal (modelled independently in each bin by the sum of a Crystal-Ball function and a wide Gaussian component) and background models of the invariant mass (exponentials of second degree polynomials for the first nine bins, and a third degree polynomial for the last bin) for the alternative analysis. The background shape and normalisation in each bin is determined by the fit to the data, as well as the overal number of signal events, while the signal yield per bin is obtained assuming the spin-0 hypothesis.

png (72kB) eps (26kB) pdf (11kB)

Figure 08b:
Diphoton invariant mass (mgg) distributions in each bin of |cos theta*|, overlaid with the sum of signal (modelled independently in each bin by the sum of a Crystal-Ball function and a wide Gaussian component) and background models of the invariant mass (exponentials of second degree polynomials for the first nine bins, and a third degree polynomial for the last bin) for the alternative analysis. The background shape and normalisation in each bin is determined by the fit to the data, as well as the overal number of signal events, while the signal yield per bin is obtained assuming the spin-0 hypothesis.

png (71kB) eps (26kB) pdf (11kB)

Figure 08c:
Diphoton invariant mass (mgg) distributions in each bin of |cos theta*|, overlaid with the sum of signal (modelled independently in each bin by the sum of a Crystal-Ball function and a wide Gaussian component) and background models of the invariant mass (exponentials of second degree polynomials for the first nine bins, and a third degree polynomial for the last bin) for the alternative analysis. The background shape and normalisation in each bin is determined by the fit to the data, as well as the overal number of signal events, while the signal yield per bin is obtained assuming the spin-0 hypothesis.

png (72kB) eps (26kB) pdf (11kB)

Figure 08d:
Diphoton invariant mass (mgg) distributions in each bin of |cos theta*|, overlaid with the sum of signal (modelled independently in each bin by the sum of a Crystal-Ball function and a wide Gaussian component) and background models of the invariant mass (exponentials of second degree polynomials for the first nine bins, and a third degree polynomial for the last bin) for the alternative analysis. The background shape and normalisation in each bin is determined by the fit to the data, as well as the overal number of signal events, while the signal yield per bin is obtained assuming the spin-0 hypothesis.

png (72kB) eps (26kB) pdf (11kB)

Figure 08e:
Diphoton invariant mass (mgg) distributions in each bin of |cos theta*|, overlaid with the sum of signal (modelled independently in each bin by the sum of a Crystal-Ball function and a wide Gaussian component) and background models of the invariant mass (exponentials of second degree polynomials for the first nine bins, and a third degree polynomial for the last bin) for the alternative analysis. The background shape and normalisation in each bin is determined by the fit to the data, as well as the overal number of signal events, while the signal yield per bin is obtained assuming the spin-0 hypothesis.

png (73kB) eps (26kB) pdf (11kB)

Figure 08f:
Diphoton invariant mass (mgg) distributions in each bin of |cos theta*|, overlaid with the sum of signal (modelled independently in each bin by the sum of a Crystal-Ball function and a wide Gaussian component) and background models of the invariant mass (exponentials of second degree polynomials for the first nine bins, and a third degree polynomial for the last bin) for the alternative analysis. The background shape and normalisation in each bin is determined by the fit to the data, as well as the overal number of signal events, while the signal yield per bin is obtained assuming the spin-0 hypothesis.

png (73kB) eps (26kB) pdf (11kB)

Figure 08g:
Diphoton invariant mass (mgg) distributions in each bin of |cos theta*|, overlaid with the sum of signal (modelled independently in each bin by the sum of a Crystal-Ball function and a wide Gaussian component) and background models of the invariant mass (exponentials of second degree polynomials for the first nine bins, and a third degree polynomial for the last bin) for the alternative analysis. The background shape and normalisation in each bin is determined by the fit to the data, as well as the overal number of signal events, while the signal yield per bin is obtained assuming the spin-0 hypothesis.

png (74kB) eps (27kB) pdf (11kB)

Figure 08h:
Diphoton invariant mass (mgg) distributions in each bin of |cos theta*|, overlaid with the sum of signal (modelled independently in each bin by the sum of a Crystal-Ball function and a wide Gaussian component) and background models of the invariant mass (exponentials of second degree polynomials for the first nine bins, and a third degree polynomial for the last bin) for the alternative analysis. The background shape and normalisation in each bin is determined by the fit to the data, as well as the overal number of signal events, while the signal yield per bin is obtained assuming the spin-0 hypothesis.

png (73kB) eps (26kB) pdf (11kB)

Figure 08i:
Diphoton invariant mass (mgg) distributions in each bin of |cos theta*|, overlaid with the sum of signal (modelled independently in each bin by the sum of a Crystal-Ball function and a wide Gaussian component) and background models of the invariant mass (exponentials of second degree polynomials for the first nine bins, and a third degree polynomial for the last bin) for the alternative analysis. The background shape and normalisation in each bin is determined by the fit to the data, as well as the overal number of signal events, while the signal yield per bin is obtained assuming the spin-0 hypothesis.

png (73kB) eps (26kB) pdf (11kB)

Figure 08j:
Diphoton invariant mass (mgg) distributions in each bin of |cos theta*|, overlaid with the sum of signal (modelled independently in each bin by the sum of a Crystal-Ball function and a wide Gaussian component) and background models of the invariant mass (exponentials of second degree polynomials for the first nine bins, and a third degree polynomial for the last bin) for the alternative analysis. The background shape and normalisation in each bin is determined by the fit to the data, as well as the overal number of signal events, while the signal yield per bin is obtained assuming the spin-0 hypothesis.

png (75kB) eps (27kB) pdf (11kB)

Figure 09a:
Expected distribution of ln(L(0)/L(2)) – the logarithm of the ratio of profiled likelihoods under the spin-0 and spin-2 hypotheses – in the presence of spin-0 (blue distributions) or spin-2 (red distributions) signals for 0% gg + 100% qqbar production of the spin-2 signal. The coloured areas correspond to the integrals of the expected distributions up to the observed values and are used to compute the p-values for the rejection of each hypothesis.

png (58kB) eps (17kB) pdf (8kB)

Figure 09b:
Expected distribution of ln(L(0)/L(2)) – the logarithm of the ratio of profiled likelihoods under the spin-0 and spin-2 hypotheses – in the presence of spin-0 (blue distributions) or spin-2 (red distributions) signals for 25% gg + 75% qqbar production of the spin-2 signal. The coloured areas correspond to the integrals of the expected distributions up to the observed values and are used to compute the p-values for the rejection of each hypothesis.

png (55kB) eps (15kB) pdf (7kB)

Figure 09c:
Expected distribution of ln(L(0)/L(2)) – the logarithm of the ratio of profiled likelihoods under the spin-0 and spin-2 hypotheses – in the presence of spin-0 (blue distributions) or spin-2 (red distributions) signals for 50% gg + 50% qqbar production of the spin-2 signal. The coloured areas correspond to the integrals of the expected distributions up to the observed values and are used to compute the p-values for the rejection of each hypothesis.

png (59kB) eps (17kB) pdf (8kB)

Figure 09d:
Expected distribution of ln(L(0)/L(2)) – the logarithm of the ratio of profiled likelihoods under the spin-0 and spin-2 hypotheses – in the presence of spin-0 (blue distributions) or spin-2 (red distributions) signals for 25% gg + 75% qqbar production of the spin-2 signal. The coloured areas correspond to the integrals of the expected distributions up to the observed values and are used to compute the p-values for the rejection of each hypothesis.

png (60kB) eps (15kB) pdf (7kB)

© 2013 CERN for the benefit of the ATLAS Collaboration. Reproduction of the article, figures and tables on this page is allowed as specified in the CC-BY-4.0 license.
2024-04-18 01:10:41