Another intriguing result about B-mesons

Today's update on the measurement of the Bs → μμ branching fraction from CDF makes your heart beat a little faster. The Tevatron collider produces huge numbers of neutral Bd and Bs mesons and they're being looking at from every angle in desperate attempts to spot any departures from the Standard Model predictions. One interesting process to look for is when the 2 quarks making a B-meson annihilate, inducing a decay of that meson to a μ+ μ- pair. This process is mediated by flavor changing neutral currents and therefore within the Standard Model it occurs only via loop processes (see the diagrams), as opposed to much more frequent tree-level charged current decays of the b-quarks. As a consequence, the Bx → μμ decays are suppressed by small loop and CKM factors and the branching fraction ends up being tiny, 3×10^-9 for Bs mesons and 10^-10 for Bd mesons, which is below the current sensitivity. At the same time, these decays has been searched for vigorously because it's fairly easy for new physics to mess them up. For example, additional Higgses in 2-Higgs-doublet models, Z-prime gauge bosons, or SUSY particles in R-parity violating models could mediate these decays pump up the branching fraction.

CDF just posted the latest update on that search based on 7fb-1 of data. They pick up pairs of opposite sign muons originating from the same displaced vertex and measure the dimuon invariant mass. If that mass falls into the window of the Bs or Bd meson mass then we have a Bx → μμ candidate. On top of that, several other properties of these events are cooked into a magic potion (called the neural network discriminant by those in the know) to better distinguish the signal from background. See the plot of the number of events in various bins of the NN discriminant as a function of the dimuon invariant mass. A tantalizing excess can be seen in the upper right window of the plot, with 4 observed vs. 0.9 expected dimuon events having a large likelihood of coming from Bs decays. You should not look in the 2nd left window in that row showing a large excess (16 observed, 8 expected) in the bin where they don't expect any signal ;-) Based on the 3 highest bins, CDF estimates the branching fraction of Bs → μμ is (1.8 ± 1) × 10−8, which is about 2 sigma above the expected Standard Model value. The middle row corresponds to events where one of the muons is detected in the forward region, in which case less signal is expected and no excess is seen. The lower row tells you there is no excess of Bd → μμ events.

So is it interesting or not? First of all, it's merely a 2 sigma excess. Secondly, the data do not trace very well the expected background outside the signal window which casts doubts whether CDF has everything under control. Nevertheless, the new CDF result is very exciting in the context of the D0 observation of the anomalous dimuon charge asymmetry. That anomaly is related to a different decay process where two B-mesons decay to *one* muon each. It is however plausible that both anomalies have a common origin, see for example this paper for quantitative estimates of the Bs → μμ branching fraction in concrete models addressing the D0 anomaly.

The best thing is that we'll learn more very soon. The LHCb experiment is well equipped to make the same measurement. At the moment they have over 400 pb-1 of data on tape. Their own estimates suggest they should be able to see a 3 sigma excess if the Bs → μμ branching fraction is equal to the CDF central value. Moreover, ATLAS and CMS may also try to stick a foot in the door. Hold your breath for just a bit longer; in case anyone sees something the rumor will soon be out on blogs ;-)

See also Tommaso's post. The Wine&Cheese seminar will take place this Friday 9pm Europe time.