Measurement of the forward-backward asymmetry in top quark-antiquark production in $p\overline{p}$ collisions using the $\mathrm{lepton}\mathbf{+}\mathrm{jets}$ channel

We present a measurement of the forward–backward asymmetry in top quark–antiquark production using the full Tevatron Run II data set collected by the D0 experiment at Fermilab. The measurement is performed in $\text{lepton}+\mathrm{jets}$ final states using a new kinematic fitting algorithm for events with four or more jets and a new partial reconstruction algorithm for events with only three jets. Corrected for detector acceptance and resolution effects, the asymmetry is evaluated to be $A_{\mathrm{FB}}=(10.6\pm 3.0)\%$. Results are consistent with the standard model predictions which range from 5.0% to 8.8%. We also present the dependence of the asymmetry on the invariant mass of the top quark–antiquark system and the difference in rapidities of the top quark and antiquark.

Figure 1

The reconstructed $m_p$ for $l+3\text{jet}$ events with (a) one and (b) two or more $b$ tags. Sample composition determination is discussed in Sec. 6. The ratio between the data counts and the model expectation is shown in the lower panel, with the hashed area representing the systematic uncertainties.

Figure 2

The probability to correctly reconstruct the sign of $\mathrm{\Delta }y$ as a function of the production-level $|\mathrm{\Delta }y|$ for the algorithm of Ref. [27] used to measure the $A_{\mathrm{FB}}$ in Ref. [4] and the algorithms used to reconstruct $l+\ge 4\mathrm{jet}$ events and to partially reconstruct $l+3\text{jet}$ events in this paper.

Figure 3

The combined discriminant $D_c$. The region $D_c<1$ is not used to determine the signal $A_{\mathrm{FB}}$. The ratio between the data counts and the model expectation is shown in the lower panel, with the hashed area representing the systematic uncertainties. Figure is from Ref. [10].

Figure 4

Reconstructed difference between the rapidities of the top and antitop quarks, $\mathrm{\Delta }y$. The left column (a,c,e) shows $l+3\text{jet}$ events, and the right column (b,d,f) shows $l+\ge 4\mathrm{jet}$ events. Rows from top to bottom display events with 0 (a,b), 1 (c,d), and $\ge 2$ $b$ (e,f) tags. Overflows are included in the edge bins. The ratio between the data counts and the model expectation is shown in the lower panels, with the hashed area representing the systematic uncertainties.

Figure 5

Reconstructed invariant mass of the top quark–antiquark pair, $m_{t\overline{t}}$. The left column (a,c,e) shows $l+3\text{jet}$ events, and the right column (b,d,f) shows $l+\ge 4\mathrm{jet}$ events. Rows from top to bottom display events with 0 (a,b), 1 (c,d), and $\ge 2$ $b$ (e,f) tags. The ratio between the data counts and the model expectation is shown in the lower panels, with the hashed area representing the systematic uncertainties.

Figure 6

Production-level bins for the 2D measurement in the $(m_{t\overline{t}},\mathrm{\Delta }y)$ plane, overlaid on the distribution in these variables predicted from mc@nlo. The shading reflects the predicted event density in arbitrary units. The solid and dashed lines denote the production-level bins. The solid lines show bins that are used for the final result.

Figure 7

The bias as a function of the input $A_{\mathrm{FB}}$. Axigluon scenarios are indicated in the legend by the mass of the axigluon, $m_a$. The toy models are labeled by parameters $w$, $a$, and $\delta$ of Eq. (8). Unless stated otherwise $w=1$, $a=0$, and $\delta =1$. For each set of $a$, $\delta$, and $w$, the value of $\mu$ is varied to produce different input asymmetries. The dashed line indicates the calibration applied to the inclusive measurement and the dotted lines indicate the assigned calibration uncertainties. The point significantly outside of the dotted lines corresponds to an axigluon mass of 0.4 TeV and is discussed in the text.

Figure 8

The production-level $\mathrm{\Delta }y$ distribution. The D0 data points are shown with their statistical uncertainty indicated by the black rectangles and their total uncertainty, based on the diagonal elements of the covariance matrix, indicated by the hashed areas. The histogram shows the mc@nlo prediction [18]. The $x$ coordinate of each data point is the observed average of the $\mathrm{\Delta }y$ distribution in the corresponding bin.

Figure 9

The dependence of the forward-backward asymmetry on $|\mathrm{\Delta }y|$. The D0 data points are shown with the total error bars indicating the total uncertainty, based on the diagonal elements of the covariance matrix, while the statistical uncertainties are indicated by the inner error bars. The dashed line shows the fit to the data with the dotted lines indicating the fit uncertainty. The horizontal lines show the mc@nlo prediction for the asymmetry in each $m_{t\overline{t}}$ bin [18]. The last bin has no upper boundary. The $x$ coordinate of each data point is the observed average of $|\mathrm{\Delta }y|$ in the corresponding bin.

Figure 10

The dependence of the forward-backward asymmetry on the invariant mass of the $t\overline{t}$ system. The D0 data points are shown with the total error bars indicating the total uncertainty, based on the diagonal elements of the covariance matrix, while the statistical uncertainties are indicated by the inner error bars. The dashed line shows a linear fit to the data with the dotted curves indicating the fit uncertainty. The horizontal lines correspond to the mc@nlo prediction for the asymmetry in each $m_{t\overline{t}}$ bin [18]. Shaded boxes correspond to the prediction of Refs. [33, 35]. The last bin has no upper boundary. The $x$ coordinate of each data point is the simulated average of the $m_{t\overline{t}}$ distribution in the corresponding bin.