Measurement of the $|V_{cb}|$ element of the CKM matrix in $t\bar{t}$ decays with the ATLAS detector

Abstract

The first measurement of the magnitude of the CKM quark mixing matrix element $|V_{cb}|$ using on-shell $W$-boson decays is presented. The measurement is made with $t\bar{t}$ events from $pp$ collision data corresponding to an integrated luminosity of $140$ fb$^{-1}$ at a centre-of-mass energy of 13 TeV collected by the ATLAS experiment at the Large Hadron Collider. A value of: $$|V_{cb}| = (50^{+11}_{-14})\times10^{-3}$$ is measured, where the uncertainties have roughly equal contributions from the limited size of the data sample and from systematic effects. This is consistent with existing measurements made at much lower energy scales in $B$ hadron decays. While the sensitivity does not yet rival that of previous determinations, this novel approach probes a very different physical situation, namely the rare hadronic $t\rightarrow b\bar{b} c$ decay at high momentum-transfer, rather than semileptonic $B$ decays at low momentum-transfer.

Figures (7)

• Figure 1: Leading-order Feynman diagram describing the $t\bar{t} \rightarrow bW^{+} \, \bar{b}W^{-} \rightarrow b \bar{b}c \, \bar{b} \ell^{-} \bar{\nu}$ decay. Vertices are labelled with the appropriate CKM matrix elements.
• Figure 2: Distribution of the $D_b$ scores for jets from events after applying the event selection requirements detailed in the text, with the thresholds used to define $c$- and $b$-tagged jets shown. For visualisation purposes, the distributions for each flavour are normalised such that the plotted numbers of true light-, $c$-, and $b$-jets are equal. Since at least one $b$-tagged jet is required in the event selection, there is a step in the distribution at the $b$-tag $D_b$ threshold.
• Figure 3: Pre-fit distribution of the Neural Network (NN) output score in the $3b1c$ tagging category. The data are shown as closed circles with statistical uncertainties; the prediction is shown as stacked histograms, broken down into the signal $|V_{cb}|=1$ and background processes. The signal $|V_{cb}|=1$ process is also overlaid separately for visualisation. The hatched bands represent the total systematic uncertainties. The lower panels show the ratios of the data to the prediction. The arrows indicate the optimised bin boundaries that are used in the profile-likelihood fit. The "Others" category consists of $Z$+jets, diboson, + $H/W/Z$ and multijet events.
• Figure 4: Ranking of the neural network input variables by importance, evaluated with data. The metric is defined as the magnitude of the averaged partial derivative of the discrimination output score with respect to the input. Here, $j_{1,2,3,4}$ refer to the four selected jets in descending order of $p_{\text{T}}$. The reconstructed neutrino is referred to as $\nu$, whose $p_{x}$ and $p_{y}$ are defined based on $E_{\text{T}}^{\text{miss}}$, and whose $p_{z}$ is derived using the $W$ boson invariant mass constraint $m(\ell\nu)=80.4$$\text{Ge V}$. The hadronic $W$ boson daughter jets are referred to as $j_{W1}$ and $j_{W2}$, in descending order of $p_{\text{T}}$.
• Figure 5: Post-fit distributions of the data and predicted event yields across all bins used in the fit. Each panel in the grid represents a tagging category ($1b1c$-$3b1c$), showing the event yields in the three Neural Network (NN) output score bins (Low, Medium, High). The grid is organised such that regions in the same row contain the same total number of heavy-flavour ($b$ or $c$) jets and share the same $y$-axis scales, while regions in the same column contain the same number of $b$-jets. The data are shown as closed circles with statistical uncertainties; the prediction is shown as stacked histograms, broken down into the signal $|V_{cb}|=1$ and background processes. The signal $|V_{cb}|=1$ process is also separately overlaid for visualisation. The hatched bands represent the total systematic uncertainties. The lower panels show the ratios of the data to the predictions.