Search for the decays $B_{(s)}^0\to J/ψγ$ at LHCb

Abstract: A search for the rare decays $B_{(s)}^0\to J/ψγ$ is performed with proton-proton collision data collected by the LHCb experiment, corresponding to integrated luminosities of $3~\rm{fb}^{-1}$ at centre-of-mass energies of 7 and 8 TeV, and $6~\rm{fb}^{-1}$ at 13 TeV. Assuming no contribution from $B^0\to J/ψγ$ decay, an upper limit is set on the branching fraction $\mathcal{B}(B_{s}^0\to J/ψγ)<2.9\times10^{-6}$ at the 90% confidence level. If instead no contribution from $B_{s}^0\to J/ψγ$ decay is assumed, the limit is $\mathcal{B}(B^0\to J/ψγ)<2.5\times10^{-6}$ at the 90% confidence level. These results supersede the previous LHCb results, with the limit for $B_{s}^0\to J/ψγ$ improved by a factor of 2.5.

• The paper sets the most stringent upper limits on the branching fractions for B(s)⁰ → J/ψγ and B⁰ → J/ψγ decays.
• It employs full Run 1 and Run 2 datasets with advanced photon conversion reconstruction and multi-stage background suppression.
• The results exclude certain pQCD predictions at 99.7% CL, motivating refinements in nonperturbative QCD and new physics models.

Search for the Annihilation-Type Radiative Decays $B_{(s)}^0 \to J/\psi \gamma$ at LHCb

Theoretical Context and Motivation

The decays $B_s^0 \to J/\psi\gamma$ and $B^0 \to J/\psi\gamma$ are pure annihilation-type rare radiative processes in the Standard Model (SM), proceeding predominantly via $W$-exchange diagrams. These decay channels are highly suppressed in the SM due to CKM and color factors, with lowest-order branching fractions predicted to be $\mathcal{O}(10^{-7})$–$\mathcal{O}(10^{-6})$ for $B_s^0$ and an order of magnitude lower for $B^0$ decays depending on the adopted factorization scheme and the treatment of nonperturbative effects. Among the possible enhancements, scenarios including intrinsic charm in the $B$-meson wave function or contributions from non-SM right-handed currents stand out, making experimental investigation of these decays valuable probes for potential new physics.

Figure 1: Feynman diagram for the leading contribution to the $B^0_{(s)}\to J/\psi\gamma$ amplitudes, with photon emission from light quarks dominating; emission from heavy quarks is suppressed.

Experimental Approach and Analysis Strategy

The analysis utilizes the full Run 1 (3 fb$B_s^0 \to J/\psi\gamma$0 at 7, 8 TeV) and Run 2 (6 fb$B_s^0 \to J/\psi\gamma$1 at 13 TeV) $B_s^0 \to J/\psi\gamma$2 collision data sets from the LHCb detector, representing an integrated luminosity of 9 fb$B_s^0 \to J/\psi\gamma$3. The experimental signature targeted is the reconstruction of $B_s^0 \to J/\psi\gamma$4 combined with converted photons ($B_s^0 \to J/\psi\gamma$5) in the silicon tracking system. Photon conversions are subdivided into the $B_s^0 \to J/\psi\gamma$6 and $B_s^0 \to J/\psi\gamma$7 categories based on the conversion location, impacting the achievable mass resolution and selection efficiency.

To suppress backgrounds—dominated by partially reconstructed $B_s^0 \to J/\psi\gamma$8 hadron decays such as $B_s^0 \to J/\psi\gamma$9, $B^0 \to J/\psi\gamma$0, and combinatorial sources—a multi-stage selection was implemented including trigger requirements, muon and photon identification, vertex and kinematic selections, and BDT-based classification trained to suppress both combinatorial and specific partially reconstructed backgrounds.

The signal extraction is done via simultaneous unbinned maximum-likelihood fits to the $B^0 \to J/\psi\gamma$1 spectra in both $B^0 \to J/\psi\gamma$2 and $B^0 \to J/\psi\gamma$3 categories. Signal shapes are modeled with double-sided Crystal Ball functions, with resolution parameters validated and adjusted using control channels such as $B^0 \to J/\psi\gamma$4. Backgrounds are modeled with a combination of ARGUS and polynomial functions, with normalization of certain physics backgrounds constrained using external measurements and simulation-derived acceptances.

Figure 2: Invariant-mass distributions for the $B^0 \to J/\psi\gamma$5 search in the $B^0 \to J/\psi\gamma$6 and $B^0 \to J/\psi\gamma$7 photon categories with fit results superimposed.

Statistical Treatment and Results

No significant excess is observed over the background expectations in either decay channel. The best fits yield central values of

$B^0 \to J/\psi\gamma$8

where the first uncertainty is statistical and the second systematic.

Figure 3: Invariant-mass distribution for $B^0 \to J/\psi\gamma$9 signal candidates in a narrow mass window around the $W$0 mass with fit result; no significant excess is seen.

Systematic uncertainties are dominated by imperfect knowledge of the photon energy resolution, background modeling, and selection efficiency mismodeling. The total systematic uncertainty remains subdominant relative to the statistical uncertainty; the analysis achieves a factor of 2.5 improvement in sensitivity over previous results for $W$1.

Given the absence of statistically significant signals, the $W$2 method is employed to set conservative upper limits. At 90% (95%) CL, the branching fraction upper limits are

$W$3

Notably, the $W$4 limit excludes the perturbative QCD prediction of $W$5 at the $W$6 CL, thus directly constraining a specific SM calculation.

Figure 4: Expected and observed $W$7 values as a function of assumed branching fraction for $W$8 and $W$9, including 1$\mathcal{O}(10^{-7})$0 and 2$\mathcal{O}(10^{-7})$1 uncertainty bands; vertical lines show the observed/expected upper limits.

Implications and Prospects

This analysis severely constrains annihilation-type radiative decay rates and associated long-distance contributions within the SM, removing significant phase space for large nonperturbative enhancements or new-physics scenarios with moderate branching fraction enhancements. The exclusion of the $\mathcal{O}(10^{-7})$2 pQCD model explicitly disfavors the upper edge of SM predictions, intensifying the focus on more precise nonperturbative approaches such as those predicting $\mathcal{O}(10^{-7})$3–$\mathcal{O}(10^{-7})$4.

From a phenomenological standpoint, these results validate the use of converted photons at hadron colliders as a competitive strategy for low-rate heavy flavor radiative decay searches, leveraging the high photon energy resolution to reject irreducible backgrounds from partially reconstructed decays. Furthermore, as improvements in photon reconstruction and data-driven efficiency calibration are realized, sub-$\mathcal{O}(10^{-7})$5 sensitivity for these and related decay modes becomes accessible.

On the theoretical side, further refinement of hadronic form factor calculations and more systematic studies of nonfactorizable effects in annihilation-type radiative $\mathcal{O}(10^{-7})$6 decays are motivated. Direct sensitivity to possible right-handed current effects or intrinsic charm contributions remains limited by the current statistical power, but is within reach with projected higher-luminosity datasets.

Conclusion

A comprehensive search for the rare decays $\mathcal{O}(10^{-7})$7 and $\mathcal{O}(10^{-7})$8 at LHCb using the full Run 1 and Run 2 data sets observes no significant signals and sets the most stringent upper limits to date: $\mathcal{O}(10^{-7})$9 and $\mathcal{O}(10^{-6})$0 at 90% CL (2604.02933). These constraints directly test and in some cases exclude specific SM predictions, and provide strong guidance for model building. Future enhancements in luminosity and reconstruction performance are anticipated to probe the remaining allowed SM parameter space, enabling not only branching fraction measurements but, following observation, studies of photon polarization and $\mathcal{O}(10^{-6})$1 violation observables in these rare channels.