Magnetic Monopole Phenomenology at Future Hadron Colliders

Abstract In the grand tapestry of Physics, the magnetic monopole (MM) is a holy grail. Therefore, numerous efforts are underway in search of this hypothetical particle at CMS, ATLAS, and MoEDAL experiments of Large Hadron Collider (LHC) by employing different production mechanisms. The cornerstone of our comprehension of monopoles lies in Dirac’s theory which outlines their characteristics and dynamics. Within this theoretical framework, an effective U (1) gauge field theory, derived from conventional models, delineates the interaction between spin magnetically-charged fields and ordinary photons under electric–magnetic dualization. The focus of this paper is the production of MMs through Drell–Yan (DY) and the Photon-Fusion (PF) mechanisms to generate velocity-dependent scalar, fermionic, and vector monopoles of spin angular momentum <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"><mml:mn>0</mml:mn><mml:mo>,</mml:mo><mml:mstyle displaystyle="false"><mml:mfrac><mml:mrow><mml:mn>1</mml:mn></mml:mrow><mml:mrow><mml:mn>2</mml:mn></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>,</mml:mo><mml:mn>1</mml:mn></mml:math> respectively at LHC. A computational study compares the monopole pair-production cross-sections for both methods at various center-of-mass energies ( <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"><mml:msqrt><mml:mrow><mml:mi>s</mml:mi></mml:mrow></mml:msqrt></mml:math> ) with different magnetic dipole moments. The comparison of kinematic distributions of monopoles at Parton and reconstructed level are demonstrated for both DY and PF mechanisms. Extracted results showcase how modern machine-learning techniques can be used to study the production of MMs at the future proton-proton particle colliders at 100 TeV. We demonstrate the observability of MMs against the most relevant Standard Model background using multivariate methods such as Boosted Decision Trees, Likelihood, and Multilayer Perceptron. This study compares the performance of these classifiers with traditional cut-based and counting approaches, proving the superiority of our methods.