ATLAS reports first observation of single top-photon production

15 March 2022 | By

The ATLAS Collaboration announces the first observation of “tqγ production”: the associated production of a single top quark and a photon in proton-proton collisions at the LHC.

The top quark is special. It’s the heaviest known elementary particle, plays a special role in electroweak symmetry breaking and its interactions provide promising leads for searches for physics beyond the Standard Model. Though it has already been 27 years since the top quark's discovery, the relative difficulty of identifying it in experimental data means that many of its properties are less well understood than those of the lighter quarks. By taking accurate measurements of its properties with rare processes, physicists can explore the impact of new physics phenomena at the highest energies (i.e. constraining the parameters of the “Effective Field Theory”).

The ATLAS Collaboration recently announced the first observation of one of these rare processes: the production of a single top quark in association with a photon. In their new analysis, ATLAS researchers analysed the full LHC Run-2 dataset, recorded by the detector between 2015 and 2018. The result is the first unambiguous observation of the tqγ production. This achievement was far from straightforward, as the search was dominated by background processes, 10 times larger than the tqγ signal events, mimicking the same signature in the detector.

ATLAS physicists focused on collision events where the top quark decays to an electron or a muon, a neutrino, and a bottom quark. To further narrow their search, they also sought out a particular characteristic of tqγ events: a “forward jet”, which is a jet of hadrons that is commonly produced and travels at small angles to the beam pipe.

The result is the first unambiguous observation of the tqγ production, with a statistical significance of 9.1 standard deviations.

The distribution of neural network output for data and Monte Carlo predictions in the most sensitive signal region with at least one forward jet. The fitted yield for the tqγ signal is shown in orange and the data points are shown in black. (Image: ATLAS Collaboration/CERN)

The main background processes in this analysis were the production of two top quarks with a photon (ttγ), the production of a W boson with a photon (Wγ), and events with “fake photons”, which are either electrons or hadrons but are mis-identified as photons. One of the major challenges in this analysis is that Monte Carlo generators are not good at simulating background processes with fake photons. To overcome this, researchers developed new techniques to correct the rates of backgrounds with fake photons using data in regions with experimental signatures close to those of the signal.

To separate the tqγ events from the background events, researchers trained a neural network in each signal region. Distinct features of the signal, such as the kinematic properties of the photon and lepton, are used as the input variables for the training. The figure shows the neural network output for the most signal-rich selection.

The statistical significance of the measurement is 9.1 standard deviations – well above the 5 standard-deviation threshold to claim observation. The expected significance is 6.7 standard deviations, were the tqγ signal equal to the Standard Model prediction. The cross-section of tqγ production at parton level was measured to be 580 ±19 (stat.) ±63 (syst.) fb, which is consistent with the Standard Model within 2.5 standard deviations.

This exciting measurement also allows physicists to look for hints of new interactions that might exist beyond the current kinematic reach of the LHC. In particular, physicists can now use the tqγ production process to find new particles that could alter the top-photon interaction. Further studies with new analysis techniques and a significantly larger dataset from the upcoming Run 3 of the LHC provides an exciting road ahead!