Extending the Matrix Element Method beyond the Born approximation: calculating event weights at next-to-leading order accuracy

Till Martini,Peter Uwer

doi:10.1007/jhep09(2015)083

Abstract

In this article we illustrate how event weights for jet events can be calculated efficiently at next-to-leading order (NLO) accuracy in QCD. This is a crucial prerequisite for the application of the Matrix Element Method in NLO. We modify the recombination procedure used in jet algorithms, to allow a factorisation of the phase space for the real corrections into resolved and unresolved regions. Using an appropriate infrared regulator the latter can be integrated numerically. As illustration, we reproduce differential distributions at NLO for two sample processes. As further application and proof of concept, we apply the Matrix Element Method in NLO accuracy to the mass determination of top quarks produced in e+e- annihilation. This analysis is relevant for a future Linear Collider. We observe a significant shift in the extracted mass depending on whether the Matrix Element Method is used in leading or next-to-leading order.

Highlights

Since it allows a direct comparison of observed event samples with expectations within a specific theoretical model
As described in ref. [22] the phase space of n + 1 massless partons can be factorised in terms of a phase space of n massless momenta — which we identify with the jet momenta Ji — and the dipole phase space measure dRij,k related to the emission of an additional parton: dRn+1 = dRn dRij,k (3.2)
From the above results we may conclude that the Born matrix element evaluated for mt = 178 GeV gives a better approximation of the next-to-leading order (NLO) corrections evaluated for mt = 174 GeV than the Born approximation evaluated for 174 GeV

Summary

Introduction

Since it allows a direct comparison of observed event samples with expectations within a specific theoretical model. To simplify the application of the Matrix Element Method the automated calculation of the required event weights has been studied recently in ref. It is worth stressing that for established jet algorithms, which rely on a 2 → 1 clustering, the jet masses created in the perturbative calculation have very little to do (in case of massless partons) with the jet masses observed in the experimentally measured jets in case of ‘light’ jets.1 The latter are mostly related to non-perturbative effects. A method needs to be constructed allowing the efficient integration of the regions in the n + 1-parton phase space contributing to the considered n-jet configuration.

Objectives

Methods

Findings

Conclusion