DØ's Single Top Search in Run II

Search for Single Top Quark Production
Using Likelihood Discriminants at DØ in Run II

Plain english summary and previous DØ result

B. Clément, D. Gelé

E-mail the authors: Benoit Clément, Denis Gelé

E-mail the DØ single top group conveners: Arán García-Bellido, Ann Heinson, Reinhard Schwienhorst

Abstract

We present an improved search for single top quarks in two production modes, s-channel and t-channel. The search is performed in the electron+jets and muon+jets decay channels, with one or more b-tagged jets, on nearly 370 pb^-1 of DØ Run II data collected between August 2002 and October 2004. Impact-parameter based b-quark tagging is used to select signal-like events. We use a likelihood discriminant method to separate signals from backgrounds. The resulting expected/observed 95% confidence level Bayesian upper limits on the single top quark production cross sections are 3.3/5.0 pb (s-channel) and 4.3/4.4 pb (t-channel).


Please scroll down the page for: analysis description, plots for talks, tables for talks, cross section limits.

Analysis Description
Please read the 11 page conference note for a more detailed description of these Summer 2005 results
Analysis Flow Diagram	Background Measurement Methods

B-tagging efficiency plots

Clicking on a plot will give you the .eps version. Right click and "View Image" will get you the .gif version.

Jet LIfetime Probability (JLIP) b-tagging algorithm efficiency and mistag rate, including taggability for the electron channel analysis. The analysis use 2 different woking point for the b-tagger: Loose and Tight. The single tag sample ( = 1 tag) has exactly one Tight tagged jet and no Loose tagged jet. The double tag sample ( ≥ 2 tags) has at least one Tight and one Loose tagged jets.


Likelihood Discriminants

We use sixteen likelihood discriminant for signal-background separation. There are four signal-background pairs: s-channel/W+jets, s-channel/ttbar, t-channel/W+jets and t-channel/ttbar, two b-tagging scheme (single and double tags) and two lepton flavor (electron and muon). The variables used in each likelihood include tranverse momenta, invariant masses and angular variables combining the different reconstructed objects (lepton, neutrino, jets, b-jet) Data to backgroungd model comparisons and the expected signal for some of these variables are given below. Each plot combines single and double tag in both electron and muon channels.
	Centrality of the event : Σ_jetsp_T/Σ_jetsp where p is the jet momentum and p_T the jet transverse momentum. Used in: s-channel/ttbar and t-channel/ttbar filters.
	The minimum angular separation between 2 jets. ΔR_min(jj) with ΔR²=(Δη²+ Δφ²). Used in: s-channel/ttbar and t-channel/ttbar filters.
	The invariant mass of the system of all jets, i.e. the four-vector sum of all jets in the events. Used in: s-channel/Wbb , t-channel/Wbb, s-channel/ttbar and t-channel/ttbar filters.
	The tranverse momentum of the leading jet. Used in: s-channel/Wbb , t-channel/Wbb, s-channel/ttbar and t-channel/ttbar filters.
	The charge of the lepton, multiplied by the pseudorapidity of the leading untagged, jet Q x η : In the t-channel, the final state d produced with the top quark tends to go along the incoming proton direction. Similarly, the dbar quark produced with the anti-top quark goes along the anti-proton direction. We take this CP symmetry into account through multiplying the pseudorapidity distribution by the charge of the lepton, which reflects the charge of the top quark. Used in: t-channel/Wbb and t-channel/ttbar filters.
	The invariant mass of the system of the W boson (i.e. Lorentz vector of the neutrino and the lepton) and the leading tagged jet. It corresponds to the reconstructed top quark mass. Used in: s-channel/Wbb and t-channel/Wbb filters.

Outputs of the likelihood discriminant for single tag/t-channel and double tag/s-channel. The plots shows data to background model comparisons and the expected signal for electron and muon channel combined.

The 2-dimensional contours represent the ttbar vs. W+jets filter for the observed events, modelled background and expected signal. Single and double tags, as well as electron and muon are combined. The left plot is for the s-channel process, the right plot for the t-channel.

Tables for Talks

Likelihood discriminant input variables: (eps) (gif)
Signals and backgrounds yields: (eps) (gif).
Comparison with NN analysis (230 pb^-1): (eps) (gif).
MC systematics uncertainties: (eps) (gif).
Expected/observed cross section limits: (eps) (gif). Observed limits only: (eps) (gif)

More Plots for Talks


Cross Section Limits
Distributions of the Bayesian posterior probability density for the electron and muon channels combined and 95% CL limits: expected (left) and observed (right).

95% Confidence Level Expected/Measured Upper Limits in pb (after final selections, with systematics, using Bayesian statistics)
		s-channel	t-channel
Likelihood Discriminant	Electron	4.7/5.8	6.3/7.6
	Muon	4.7/6.4	6.2/5.0
	Combined	3.3/5.0	4.3/4.4

Last modified: July 19, 2005
Send comments to Benoit Clement