Jet Dipolarity: Top Tagging with Color Flow

Abstract

A new jet observable, dipolarity, is introduced that can distinguish whether a pair of subjets arises from a color singlet source. This observable is incorporated into the HEPTopTagger and is shown to improve discrimination between top jets and QCD jets for moderate to high pT.

Figures (6)

• Figure 1: Legoplot for a top jet reconstructed by the HEPTopTagger, with hard substructure identified by a combination of filtering and a fractional mass-drop criterion. The orange cells correspond to soft radiation associated with the $W^\pm$.
• Figure 2: Top: Eikonal radiation pattern $dp_T/d\eta d\phi$ for a color singlet with $\Delta R$=0.9, typical for a $W^\pm$ originating from a top with $p_T\!\sim$ 300 GeV. Bottom: As above with the partons instead color-connected to the beam (left/right-going parton connected to the left/right beam). Contours differ by powers of $e$. For the color singlet the radiation is mostly found in the region between the two subjets. For the background-like color configuration, the radiation is pulled towards the beam. Note that an absolute comparison cannot be made between the figures, since the collinear singularities in \ref{['eqn:sigpat']} and \ref{['eqn:bkgpat']} are not regulated.
• Figure 3: Schematic for a collection of QCD jets whose kinematics fake the top. The upper figures show various possibilities for quarks and gluons that undergo two branchings. The bottom figures show the corresponding large $N_c$ color diagrams, with dipole radiation patterns superimposed across color dipoles. Only the rightmost color configuration, which is suppressed by factors of $C_A/C_F$ with respect to the others, matches the radiation pattern of an actual top.
• Figure 4: Dipolarity distributions for $W^\pm$s reconstructed by the HEPTopTagger and passing default mass cuts with $m_{\text{filt}} \in$ [150 GeV, 210 GeV]. Thick solid lines indicate central values, whereas thin dashed lines correspond to values at 10% and 90%. Here and throughout the $p_T$ is that of the fat $R=1.5$ jet. For all $p_T$ the central value of the dipolarity for the background is ${\cal O}(50\%-100\%)$ larger than for the signal. This figure uses the HERWIG event samples; the PYTHIA event samples yield similar distributions. The dot-dash line roughly indicates where dipolarity cuts are made at the S=20% working point.
• Figure 5: Background mistag rates with and without dipolarity cuts at a fixed signal efficiency of 20% for HERWIG (top) and PYTHIA (bottom). The mistag rate at a given $p_{T0}$ is calculated from a $p_T$ window of 100 GeV centered at $p_{T0}$. Note that, as a consequence, each point is not statistically independent. Error bands are statistical.