The raw charge asymmetry of B^± → K^±μ⁺μ⁻ decays

Overview

This project measures the raw charge asymmetry of B^± → K^±μ⁺μ⁻ decays. The non-resonant component is strongly suppressed in the Standard Model, making it sensitive to small additional contributions. I split the data into six intervals of the dimuon invariant mass squared, fitted the signal yield of B⁺ and B⁻ candidates in each, and compared the resulting asymmetries against a charmonium control channel, B⁺ → J/ψK⁺, where any genuine CP violation is known to be negligible.

The control channel gives a raw asymmetry consistent in magnitude with the published LHCb measurement of the same quantity. Across the six rare-mode intervals, the raw asymmetries range from −0.158 to +0.250 with statistical uncertainties of 0.07–0.20; two of the six sit within one standard deviation of the control value, and the largest departure is 1.6σ. At this precision, I find no compelling discrepancy with the control.

Methods

Candidates were selected using a gradient-boosted decision tree (XGBoost) score already stored in the reduced ntuples; rather than take its behaviour on trust, I re-evaluated the stored models directly and characterised how the selection retains candidates and responds to their charge before using it. Signal yields in each mass interval come from binned maximum-likelihood fits to the reconstructed B mass, and the fits' calibration was checked by generating and refitting pseudo-experiments under the fitted model rather than relying on an asymptotic formula. Uncertainties and the sensitivity of the result to specific modelling choices were quantified directly rather than assumed.

Interpretation

This is a raw asymmetry, not a calibrated CP asymmetry. The reduced data used here doesn't retain what a dedicated CP measurement needs: the fiducial acceptance requirement, peaking-background vetoes, efficiency corrections, or per-polarity fits. So the comparison against the control channel is a consistency check of scale and convention, not a test of the Standard Model, and this project makes no claim about new physics.