Axion searches in rare Higgs decays at ATLAS

像希格斯玻色子这样的标准模型的发现是大型强子对撞机的一个重要里程碑。尽管如此，标准模型（SM）及其基础理论迄今仍受到理论和实验上出现的问题的困扰。这包括强CP-问题，这是一个理论上的论点，即smm可以在强相互作用中破坏CP对称。然而，没有实验观察到这样的证据。希格斯玻色子的观测质量约为125 GeV，这确实为新粒子的存在留下了可能性，这些新粒子不会违反标准模型的基本规范对称性，因此创造了寻找低质量粒子的潜力。有了这个动机，我的项目将有助于寻找这样的粒子。目的是开始分析寻找伪标量低质量粒子，我们称之为a，从罕见的希格斯衰变到Za。粒子a可以被确定为可能的类轴子粒子（ALP）或次最小超对称模型（NMSSM）候选者。轴子和ALP是由peccei - quinn机制产生的理论粒子，该机制通过引入动态场来解释CP违反项，描述了CP问题的可能解决方案。ALP可以耦合到所有的smp粒子，然而，根据它们的耦合和可能的质量范围，对ALP的搜索有很大的不同。Bauer（2017）的理论研究表明，存在可能的参数区域，LHC数据可以很好地用于在衰变通道hza中搜索质量在0.1-10 gev范围内的ALP，并且可以长期存在。在希格斯物理学中，ALP可以衰变为一对射流或光子，而z玻色子则衰变为两个轻子，具有清晰的信号。以前在大型强子对撞机上进行过对阿尔卑斯粒子的搜索，但到目前为止还没有成功。这项研究面临的一个挑战是，ALP光子彼此之间的距离R太小，以至于在光子识别（ID）过程中光子实际上被错误识别。该项目的主要任务是通过评估电磁量热计（ECAL）第一层光子簇的子结构来分析这些光子，以更好地区分那些被错误识别的光子，从而改进对该衰减通道的整体分析。机器学习方法，如深度神经网络和其他多变量分析技术，辅以通过作为利物浦数据密集型博士培训中心（LIV.DAT）的一部分工作获得的技能，将被详尽地尝试更好的光子重建。通过逐步改进与该衰减通道的基础分析相关的问题，可以对伪标量低质量粒子等进行彻底的研究

The discovery of a Standard Model like Higgs boson was a major milestone forthe LHC. Despite this, the Standard Model (SM) and the underlying theories sofar are plagued by problems that arise both theoretically and experimentally. Thisincludes the Strong CP-problem, which is the theoretical argument that the SMcould break CP symmetry in strong interactions. However, no evidence of suchhas been experimentally observed. The Higgs boson observation with a mass ofapproximately 125 GeV does leave open the possibility for new particles which willnot violate the elementary gauge symmetries of the Standard Model and as suchcreates the potential to search for these particles at low mass .With this motivation my project will contribute to the searches of such particles.The aim is to initiate the analysis for searching for a pseudo-scalar low mass particlewhich we will call a, from the rare Higgs decay to Za. Particle a could identifyas a possible Axion-like particle (ALP) or Next-to-Minimal Supersymmetry Model(NMSSM) candidate.Axions and ALP's are theoretical particles resulting from the Peccei-Quinnmechanism that describes a possible solution to the CP problem by introducinga dynamic field to explain the CP-violating term, ALP's can couple to all SMparticles, however depending on their coupling and possible mass ranges searchesfor ALP's differ significantly. Theoretical studies by Bauer (2017) show that thereare possible parameter regions present were LHC data could well be utilized tosearch for ALPs in the decay channel H Za for masses in the range 0.1-10 GeVand could be long lived In Higgs physics, ALP's could decay into a pair ofjets or photons alongside a Z-boson decaying into two leptons with a clean signal.Previous searches for ALPs have been conducted at the LHC but have so far notbeen successful.A challenge with this study is the fact that R, the distance the ALP's photonsare from each other, is so small that during photon identification (ID) photons areactually being misidentified. A major task of this project will focus on the analysisof these photons by evaluating the substructure of the photon shower in the 1stlayer of the electromagnetic calorimeter (ECAL) to better distinguish those thatare misidentified and as such improve the overall analysis of this decay channel.Machine learning approaches such as deep neural networks and other multivariateanalysis techniques, complemented by skills gained through working as part of theLiverpool data intensive center for doctoral training (LIV.DAT), will be exhaustivelyattempted for better photon reconstruction. By incrementally improvingissues related to the underlying analysis of this decay channel can pseudo-scalarlow mass particles and such be thoroughly investigated