Combined CDF and D0 Search for Standard Model Higgs Boson Production with up to 10.0 fb-1 of Data

TL;DR

The paper presents a comprehensive Tevatron Higgs boson search by combining CDF and D0 results with up to 10 fb^-1 of data, spanning ggH, VH, and VBF production and multiple decay channels including H→bb, H→WW, H→γγ, and more. It uses both Bayesian and CLs statistical frameworks to extract limits on sigma x B(H) relative to the SM, incorporating extensive systematic uncertainties and correlations. The combination excludes substantial Higgs mass ranges and observes a modest excess around 120 GeV/c^2, primarily in H→bb channels, with a global significance around 2.2σ when accounting for the look-elsewhere effect. The analysis demonstrates the power of multi-channel, cross-experiment synthesis to probe the SM Higgs at the Tevatron and sets the stage for future improvements in Higgs sensitivity.

Abstract

We combine results from CDF and D0 on direct searches for the standard model (SM) Higgs boson (H) in ppbar collisions at the Fermilab Tevatron at sqrt(s)=1.96 TeV. Compared to the previous Tevatron Higgs boson search combination more data have been added, additional channels have been incorporated, and some previously used channels have been reanalyzed to gain sensitivity. With up to 10 fb-1 of luminosity analyzed, the 95% C.L. median expected upper limits on Higgs boson production are factors of 0.94, 1.10, and 0.49 times the values of the SM cross section for Higgs bosons of mass m_H=115 GeV/c^2, 125 GeV/c^2,and 165 GeV/c^2, respectively. We exclude, at the 95% C.L., two regions: 100<m_H<106 GeV/c^2, and 147<m_H<179 GeV/c^2. We expect to exclude the regions 100<m_H<119 GeV/c^2 and 141<m_H<184 GeV/c^2. There is an excess of data events with respect to the background estimation in the mass range 115<m_H<135 GeV/c^2 which causes our limits to not be as stringent as expected. At m_H=120 GeV/c^2, the p-value for a background fluctuation to produce this excess is ~3.5 x 10^{-3}, corresponding to a local significance of 2.7 standard deviations. The global significance for such an excess anywhere in the full mass range is approximately 2.2 standard deviations. We also combine separately searches for H->bbbar and H->W+W-, and find that the excess is concentrated in the H->bbbar channel, although the results in the H->W+W- channel are also consistent with the possible presence of a low-mass Higgs boson.

Figures (21)

• Figure 1: Distributions of $\log_{10}(s/b)$, for the data from all contributing channels from CDF and D0, for Higgs boson masses of 115, 125, and 165 GeV/$c^2$. The data are shown with points, and the expected signal is shown stacked on top of the backgrounds. Underflows and overflows are collected into the leftmost and rightmost bins.
• Figure 2: Integrated distributions of $s/b$, starting at the high $s/b$ side, for Higgs boson masses of 115, 125, and 165 GeV/$c^2$. The total signal+background and background-only integrals are shown separately, along with the data sums. Data are only shown for bins that have data events in them.
• Figure 3: Background-subtracted data distributions for all channels, summed in bins of $s/b$, for Higgs boson masses of 115, 125, and 165 GeV/$c^2$. The background has been fit, within its systematic uncertainties and assuming no Higgs boson signal is present, to the data. The points with error bars indicate the background-subtracted data; the sizes of the error bars are the square roots of the predicted background in each bin. The unshaded (blue-outline) histogram shows the systematic uncertainty on the best-fit background model, and the shaded histogram shows the expected signal for a Standard Model Higgs boson.
• Figure 4: Distributions of the log-likelihood ratio (LLR) as a function of Higgs boson mass obtained with the ${\rm CL}_{\rm s}$ method for the combination of all CDF and D0 analyses. The green and yellow bands correspond to the regions enclosing 1 s.d. and 2 s.d. fluctuations around the median expected value assuming only background is present, respectively. The red dashed curve shows the median expected value assuming a Higgs boson signal is present, separately at each $m_H$.
• Figure 5: Observed and expected (median, for the background-only hypothesis) 95% C.L. upper limits on the ratios to the SM cross section, as functions of the Higgs boson mass for the combined CDF and D0 analyses. The limits are expressed as a multiple of the SM prediction for test masses (every 5 GeV/$c^2$) for which both experiments have performed dedicated searches in different channels. The points are joined by straight lines for better readability. The bands indicate the 68% and 95% probability regions where the limits can fluctuate, in the absence of signal. The limits displayed in this figure are obtained with the Bayesian calculation.