Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Combining Monte Carlo generators with next-to-next-to-leading order calculations: event reweighting for Higgs boson production at the LHC

Giovanna Davatz, Fabian Stoeckli, Charalampos Anastasiou, Guenther Dissertori, Michael Dittmar, Kirill Melnikov, Frank Petriello

TL;DR

This work proposes and tests a pragmatic method to incorporate NNLO QCD corrections into Monte Carlo event generators by bin-by-bin reweighting to NNLO Higgs production distributions in Higgs $p_T$ and rapidity. The approach is validated for $pp \to H+X$ using PYTHIA and MC@NLO, and extended to Higgs decays $H\to\gamma\gamma$ and $H\to W^+W^- \to l^+l^-\nu\bar{\nu}$, showing good agreement with NNLO predictions in regions where fixed-order or resummed calculations are reliable. Key findings include the necessity of NNLO-level reweighting to handle jet-veto observables, the robustness of lepton distributions against reweighting in the $WW$ channel, and the preservation of resummation features near kinematic boundaries. The study suggests that NNLO reweighting provides a practical and accurate way to include higher-order corrections in MC-based Higgs analyses at the LHC, with potential applicability to other processes.

Abstract

We study a phenomenological ansatz for merging next-to-next-to-leading order (NNLO) calculations with Monte Carlo event generators. We reweight them to match bin-integrated NNLO differential distributions. To test this procedure, we study the Higgs boson production cross-section at the LHC, for which a fully differential partonic NNLO calculation is available. We normalize PYTHIA and MC@NLO Monte Carlo events for Higgs production in the gluon fusion channel to reproduce the bin integrated NNLO double differential distribution in the transverse momentum and rapidity of the Higgs boson. These events are used to compute differential distributions for the photons in the pp \to H \to γγdecay channel, and are compared to predictions from fixed-order perturbation theory at NNLO. We find agreement between the reweighted generators and the NNLO result in kinematic regions where we expect a good description using fixed-order perturbation theory. Kinematic boundaries where resummation is required are also modeled correctly using this procedure. We then use these events to compute distributions in the pp \to H \to W^+W^- \to l^+l^- ν\barν channel, for which an accurate description is needed for measurements at the LHC. We find that the final state lepton distributions obtained from PYTHIA are not significantly changed by the reweighting procedure.

Combining Monte Carlo generators with next-to-next-to-leading order calculations: event reweighting for Higgs boson production at the LHC

TL;DR

This work proposes and tests a pragmatic method to incorporate NNLO QCD corrections into Monte Carlo event generators by bin-by-bin reweighting to NNLO Higgs production distributions in Higgs and rapidity. The approach is validated for using PYTHIA and MC@NLO, and extended to Higgs decays and , showing good agreement with NNLO predictions in regions where fixed-order or resummed calculations are reliable. Key findings include the necessity of NNLO-level reweighting to handle jet-veto observables, the robustness of lepton distributions against reweighting in the channel, and the preservation of resummation features near kinematic boundaries. The study suggests that NNLO reweighting provides a practical and accurate way to include higher-order corrections in MC-based Higgs analyses at the LHC, with potential applicability to other processes.

Abstract

We study a phenomenological ansatz for merging next-to-next-to-leading order (NNLO) calculations with Monte Carlo event generators. We reweight them to match bin-integrated NNLO differential distributions. To test this procedure, we study the Higgs boson production cross-section at the LHC, for which a fully differential partonic NNLO calculation is available. We normalize PYTHIA and MC@NLO Monte Carlo events for Higgs production in the gluon fusion channel to reproduce the bin integrated NNLO double differential distribution in the transverse momentum and rapidity of the Higgs boson. These events are used to compute differential distributions for the photons in the pp \to H \to γγdecay channel, and are compared to predictions from fixed-order perturbation theory at NNLO. We find agreement between the reweighted generators and the NNLO result in kinematic regions where we expect a good description using fixed-order perturbation theory. Kinematic boundaries where resummation is required are also modeled correctly using this procedure. We then use these events to compute distributions in the pp \to H \to W^+W^- \to l^+l^- ν\barν channel, for which an accurate description is needed for measurements at the LHC. We find that the final state lepton distributions obtained from PYTHIA are not significantly changed by the reweighting procedure.

Paper Structure

This paper contains 6 sections, 16 equations, 10 figures, 1 table.

Table of Contents

  1. Introduction
  2. The reweighting technique for $pp \to H +X$
  3. The reweighting procedure
  4. The di-photon channel
  5. The $W^+W^- \to l^+ l^- \nu \bar{\nu}$ channel
  6. Conclusions

Figures (10)

  • Figure 1: Distributions of the Higgs transverse momentum (left) and rapidity (right).
  • Figure 2: The reweighting factors integrated over rapidity for PYTHIA and MC@NLO as functions of the Higgs $p_\bot$ (left). The reweighting factors, integrated over $p_\bot$, as functions of rapidity (right). The inclusive $K$-factor for the total cross-section is also shown on both plots as a horizontal line.
  • Figure 3: The Higgs boson $p_{\perp}$ spectrum for NLO, MC@NLO, and PYTHIA reweighted (left panel); comparison of the Higgs boson $p_{\perp}$ at NLO, and with MC@NLO when only a single jet is observed (right panel).
  • Figure 4: The Higgs boson $p_{\perp}$ spectrum from a resummed calculation, MC@NLO reweighted and PYTHIA reweighted.
  • Figure 5: The Higgs boson rapidity distribution with a jet veto of 30 GeV applied for NNLO, MC@NLO reweighted and PYTHIA reweighted.
  • ...and 5 more figures