First off I should give some background on what ATLAS is, and what neutral B mesons are. ATLAS is one of the big multi-purpose experiments of the Large Hadron Collider at CERN, the machine that discovered the Higgs boson in 2012 and which is poised to search for new physics for the next two decades, studying proton-proton collisions at 13 TeV in the center of mass.
As for B mesons, these are particles composed of a bottom quark and a light antiquark (or vice-versa, a anti-bottom quark and a light quark). The neutral B meson, B0, has a down quark and an anti-bottom quark in its interior. The B0 meson is copiously produced in the LHC proton-proton collisions, and can be studied by detecting its decays into particles that we can track and measure in the complex particle detectors of which the LHC is equipped.
The property I am talking about in the first paragraph above is the difference in lifetime between the B0 and the anti-B0 mesons. Usually, when you read the prefix "anti-" in front of a particle name you are supposed to understand it is the antimatter copy of the original particle. In spite of that, the B0 and anti-B0 mesons have a tiny difference in their lifetime owing to the fact that they are observed as slightly different quantum-mechanical admixtures of two distinct states, the so-called "mass eigenstates" B_L and B_H. The admixtures are due to the fact that neutral B mesons exhibit oscillatory transitions between particle and antiparticle, a fantastic phenomenon whereby the weak interaction changes simutaneously a bottom quark into an antibottom quark, and an anti-down quark into a down quark.
The topic is too complex to explain given the limited amount of CPU I can put into finding good analogies and trivializations. I will try to do this another time, and will leave it aside for now; what matters, though, is that the Standard Model predicts this difference in lifetime to be of the order of 0.4% between the two states we measure; new physical processes have been hypothesized which could change this parameter while leaving most of the remaining phenomenology of B hadrons unaffected.
0.4% of the B0 lifetime is a small time interval in absolute terms! Given that the lifetime of neutral B mesons is in the picosecond ballpark, so already not really something you can measure with grandpa's chronograph, in order to assess the lifetime difference of B0 mesons we need to be sensitive to time scales of the order of the femtosecond (a billionth of a millionth of a second!) in our measurement if we are to test the model predictions. And yet we can do it!
To carry out its measurement, ATLAS used all the Run 1 statistics of neutral B mesons they could put together, about 25 inverse femtobarns of collisions. B mesons are easy to identify with small backgrounds when they decay to a J/ψ meson and a K or K*(892) meson. The J/? decays then to a muon pair, the K meson to a pair of charged pions, and the K* decays to a charged K-pion pair. Then the combined masses of J/ψ and K (or K*) are extracted, as shown in the graphs below. The large statistics in the Gaussian peaks, and the small backgrounds underneath, allow very precise measurements to be carried out with these events.
(Above, the B0 signals in the J/psi K_s and J/psi K* final states are shown on the left and on the right panel, respectively, from the 20.3 inverse femtobarns of 8 TeV data collected by ATLAS. The lower panels show the residuals of data with respect to the fit to signal plus background shown in blue in the upper panels).
Then a study is performed of the time evolution of the states decaying into J/ψ K_s and J/ψ K*. The ratio of these differential distributions is sensitive to the lifetime difference of the quantum-mechanical systems of B0 and anti-B0 mesons. However, a number of subtle details must be accounted for in order to be sensitive to a very small time asymmetry.
Among the subtleties one must reckon with is the fact that there are slight asymmetries in the production of B0 and anti-B0 particles in proton-proton collisions, owing to the fact that the proton contains one extra "valence" down quark, which can be transported by the collision into the final-state B0 meson -the B0 meson contains in fact a down-antibottom quark combination. For the anti-B0 meson this is not possible, as there is no anti-down "valence" quark counterpart to pick up in the proton. You would need an antiproton to do that.
After all effects are understood and corrections are accounted for, ATLAS is able to pull off a measurement of the lifetime asymmetry which reads as (-0.1 +-1.1 +-0.9)*10^-2, where the first quoted uncertainty is statistical and the second is systematic. This is of course insufficient to draw any conclusions on the existence of non-standard-model effects on the decay properties of B0 mesons, but it is definitely a step in the right direction.
For comparison, the previously determined world average of the measurements of the same quantity performed by LHCb, BELLE2 and BaBar totals (0.1+-1.0)*10^-2, a value entirely compatible with the ATLAS one and with a similarly sized uncertainty bar. The ATLAS result is the most precise single determination of the asymmetry.
Looking forward, I believe there is a possibility of reaching a sensitivity to the standard model value in the near future. This will require the combination of new results of the different experiments -and CMS might contribute as well (we know, in fact, that CMS has shown a better sensitivity than ATLAS to some important determinations of rare B decay properties; the obvious example is the branching fraction of neutral B decays to muon pairs). So, the message is always the same: stay tuned, and maybe one day we'll see new physics popping up in the B sector!