Search for single vector-like $B$ quark production in hadronic final states at the LHC

In this paper, we study the discovery potential of a Vector-Like $B$ quark (VLB) via the process $pp \to B(\to bZ)j\to b(Z \to \nu_l\bar{\nu_l})j$ at the Large Hadron Collider (LHC) with $\sqrt{s}=14$ TeV. In the framework of a simplified model, we perform a scan over its parameter space and test its viability following a Monte Carlo analysis developed to include all production and decay dynamics. We use cut-and-count combined with Extreme Gradient Boosting (XGBoost) methods to classify the signal and background events in order to improve the efficiency of signal identification and background rejection. We find that this approach can reduce background events significantly while the signal retention rate is much higher than that of traditional methods, thereby improving the VLB discovery potential. We then calculate the exclusion and discovery capabilities for VLBs and find that the advantages of the cut-and-count plus XGBoost method especially lie in the high-mass region, i.e., $m_B > 1500 \text{ GeV}$. We finally obtain the following LHC results in terms of the coupling and chiral structure of a singlet heavy VLB interactions: (i) for $g^{\ast}$=0.2 and $R_L=0$ with 3000 fb$^{-1}$, the $B$ quark mass can be be excluded (discovered) up to 3000 GeV (2500 GeV); (ii) for $g^{\ast}$=0.2 and $R_L=0.5$ with 3000 fb$^{-1}$, the exclusion (discovery) region can reach up to 4750 GeV (4250 GeV).