State-of-the-art electroweak Higgs boson pair production in association with two jets at the LHC in the Standard Model and beyond

TL;DR

The paper addresses accurate modeling of electroweak Higgs boson pair production in association with two jets via vector boson fusion, including NLO-QCD corrections and anomalous Higgs couplings. It compares two Monte Carlo tools, GoSam+Whizard (full EW $HHjj$) and POWHEG-BOX (VBF-approximate), across SM and HEFT parametrizations mapped through the kappa framework, at $\sqrt{s}=13$–14 TeV. Key findings show large differences between full and VBF-approximate predictions in inclusive selections, but near agreement under VBF cuts; parton shower effects mainly affect non-VBF observables, with hadronisation enhancing NLO+PS predictions. The results guide tool choice for HL-LHC analyses, illustrating the validity of the VBF approximation in typical selections and outlining paths to combine full-EW accuracy with parton-shower realism in future work.

Abstract

We present a systematic comparison of two state-of-the-art tools for the simulation of Higgs boson pair production via vector boson fusion (VBF) as implemented in the Monte-Carlo tools GoSam+Whizard and the POWHEG-BOX. Cross sections and distributions are provided within the Standard Model and beyond, within scenarios typical for experimental physics analyses, and for a range of energies of relevance to the LHC and its upcoming high luminosity phase. We further perform a detailed study of the so-called VBF approximation, in particular in the presence of anomalous Higgs boson couplings.

Figures (6)

• Figure 1: Examples of diagrams that are not included in the VBF approximation. The shaded blob in diagrams \ref{['fig:s-channel']} and \ref{['fig:non-VBF-NLO']} can represent any of the sub-diagrams shown in \ref{['fig:VBF_LO_sub']}.
• Figure 2: Distributions of the rapidity of the hardest tagging jet (top) and transverse momentum of the hardest Higgs boson (bottom) for the VBF $HH$ process within the inclusive cut setup (left) and the VBF cut setup (right) as obtained with the POWHEG-BOX (LO: purple, NLO: blue) and GoSam+Whizard (LO: green, NLO: orange), and their ratios to the POWHEG-BOX results at NLO (lower panels).
• Figure 3: Rapidity (top) and invariant mass (bottom) distributions of the di-Higgs system for the VBF $HH$ process within the VBF cut setup at LO as obtained with the POWHEG-BOX and GoSam+Whizard (upper panels), and the ratios of the full process (GoSam+Whizard) to the VBF approximation (POWHEG-BOX) for each set of anomalous couplings (lower panels). The values chosen for the anomalous Higgs couplings are quoted in the format $\left[c_\lambda,c_V,c_{2V}\right]$ for each curve in the legend. In addition to the SM results, denoted by $[1,1,1]$, predictions for the values $[-1,0.9,0.5]$ and $[1,0.9,1]$ are shown on the left, and for the values $[1,1,0.5]$ and $[4,0.95,0.5]$ on the right.
• Figure 4: Transverse-momentum (left) and rapidity distributions (right) of the softer Higgs boson for the VBF $HH$ process as described in the text within the VBF cut setup at NLO (purple) and at NLO+PS accuracy using HERWIG7 (orange) or PYTHIA8 with the dipole shower without (green) and with (red) hadronisation/underlying event effects, and the Vincia shower (blue). Their ratios to the respective NLO results are shown in the lower panels. Error bars indicate statistical uncertainties, uncertainty bands correspond to a 7-point variation around the central scale $\mu_0$ of each curve.
• Figure 5: Azimuthal angle separation distributions of the Higgs bosons (left) and the two taggings jets (right) for the VBF $HH$ process as described in the text within the VBF cut setup at NLO (purple) and at NLO+PS accuracy using HERWIG7 (orange) or PYTHIA8 with the dipole shower without (green) and with (red) hadronisation/underlying event effects, and the Vincia shower (blue). Their ratios to the respective NLO results are shown in the lower panels. Error bars indicate statistical uncertainties, uncertainty bands correspond to a 7-point variation around the central scale $\mu_0$ of each curve.