Single top-quark production with the Matrix Element Method in next-to-leading order accuracy

TL;DR

This paper extends the Matrix Element Method to next-to-leading order accuracy for single top-quark production, enabling more precise parameter measurements and reducing biases in collider data analysis.

Contribution

It introduces an efficient NLO calculation method for the MEM and demonstrates its application to top-quark mass determination, the first of its kind for jet production at NLO.

Findings

NLO MEM reproduces the top-quark mass accurately in pseudo-data.

Using Born likelihoods on NLO events causes significant bias.

The method can be applied to new physics searches with likelihood ratios.

Abstract

Single top-quark production offers a unique laboratory for precision tests of the Standard Model and searches of possible extensions. Furthermore, assuming the Standard Model, single top-quark production can be used to determine top-quark related couplings. For precise determinations of parameters like the electroweak gauge couplings or the mass of the top quark, efficient, unbiased, and theoretically unambiguous analysis methods are needed. Within this context, the Matrix Element Method (MEM) has been established in hadron collider analyses due to its possibility to top out at utilising the information available in experimental data. However, so far it has mostly been applied in Born approximation only. We discuss the extension to next-to-leading order (NLO) accuracy. As a necessary prerequisite we introduce an efficient method to calculate NLO QCD weights for jet events. As proof of concept and representative example we use the MEM at NLO to reproduce the top-quark mass in a toy experiment where we treat single top-quark events generated at NLO accuracy as pseudo-data. This is the first application of the MEM at NLO accuracy to the hadronic production of jets originating from coloured final state partons. We observe that analysing NLO events with Born likelihoods can introduce a pronounced bias in the extracted mass which would require significant calibration with associated uncertainties. Although we focus on parameter determinations, the methods presented here can also be used to search for new physics using likelihood ratios.