Calculation of W b bbar Production via Double Parton Scattering at the LHC

TL;DR

The paper assesses the observability of double parton scattering (DPS) in $pp \to W b\bar{b} X \to \ell \nu b\bar{b} X$ at the LHC (7 TeV) by comparing DPS and single-parton scattering (SPS) contributions and dominant backgrounds. It uses NLO predictions via POWHEG BOX to generate DPS, SPS, and backgrounds, and introduces kinematic discriminants—notably $S'_{p_T}$ and angular observables—to separate DPS from SPS and backgrounds. A combination of an upper missing-$E_T$ cut and two-dimensional distributions greatly enhances DPS significance, with $S/\sqrt{B}$ reaching roughly 12–15 for $\sigma_{\rm eff} \sim 12$ mb at 10 fb$^{-1}$. The study advocates experimental analyses exploiting these discriminants to measure DPS and extract $\sigma_{\rm eff}$, while also noting DPS's potential as a background in other searches.

Abstract

We investigate the potential to observe double parton scattering at the Large Hadron Collider in p p -> W b bbar X -> l nu b bbar X at 7 TeV. Our analysis tests the efficacy of several kinematic variables in isolating the double parton process of interest from the single parton process and relevant backgrounds for the first 10 inverse fb of integrated luminosity. These variables are constructed to expose the independent nature of the two subprocesses in double parton scattering, pp -> l nu X and pp -> b bbar X. We use next-to-leading order perturbative predictions for the double parton and single parton scattering components of W b bbar and for the pertinent backgrounds. The next-to-leading order contributions are important for a proper description of some of the observables we compute. We find that the double parton process can be identified and measured with significance S/sqrt(B) ~ 10, provided the double parton scattering effective cross section sigma_{eff} ~ 12 mb.

Figures (9)

• Figure 1: Schematic depiction of single parton scattering (left) and double parton scattering (right). In single parton scattering, one parton from each hadron is active in the scattering and the partonic process is ${\cal{A}}(i j \to a b c d)$. Double parton scattering assumes two partons from each hadron are active in the hard scattering and the total partonic process consists of two independent subprocesses ${\cal{A}}(i j \to a b)$ and ${\cal{A}}(k \ell \to c d)$.
• Figure 2: The event rate as a function of $\not{\!{\rm E}}_{T}$ for DPS and SPS. On the left, all backgrounds are included while the plot on the right compares the DPS events to those from $t\bar{t}$ alone. While the DPS signal is concentrated in the $\not{\!{\rm E}}_{T} < 45$ GeV range, the majority of the $t\bar{t}$ background lies above this range. Therefore, imposing an maximum $\not{\!{\rm E}}_{T}$ cut of 45 GeV can greatly reduce the background coming from $t\bar{t}$ production.
• Figure 3: The event rate as a function of the transverse momentum of the leading object (either a jet or lepton). The SPS contribution has a harder $p_T$ spectrum.
• Figure 4: The event rate for $Wb\bar{b}$ production from DPS (left) and SPS (right) as a function of $S_{p_T}^\prime$. In the DPS case, the distribution is peaked toward $S_{p_T}^\prime \simeq 0$; SPS production of $Wb\bar{b}$ produces bottom quarks that are not back-to-back, resulting in a broad distribution and a peak near $S_{p_T}^\prime \simeq 1$.
• Figure 5: The $S_{p_T}^\prime$ distribution for DPS and SPS production of $Wb\bar{b}$ including all relevant backgrounds. On the left, only the minimal acceptance cuts are imposed, while, on the right, an additional maximum $\not{\!{\rm E}}_{T}$ cut is imposed ($\not{\!{\rm E}}_{T} < 45$ GeV). Imposing a maximum $\not{\!{\rm E}}_{T}$ cut greatly reduces the background and produces a sharp peak in an $S_{p_T}^\prime$ region where DPS is expected to dominate.