Measurement of mixing-induced CP violation in the decay $B^0 \to \pi^0 \pi^0$

Technical Summary: Measurement of Mixing-Induced CP Violation in $B^0 \to \pi^0\pi^0$

Problem Statement
The determination of the Cabibbo–Kobayashi–Maskawa (CKM) quark-mixing matrix phase $\phi_2$ (also denoted as $\alpha$) is a critical test of the Standard Model and a probe for physics beyond it. While $\phi_2$ can be extracted from $B \to \pi\pi$ decays using isospin relations, the precision of this determination has been historically limited by discrete ambiguities. These ambiguities arise primarily from the lack of a measurement of the mixing-induced CP-violating coefficient, $S_{\pi^0\pi^0}$, in the decay $B^0 \to \pi^0\pi^0$.

Conventional measurements of time-dependent CP violation rely on reconstructing the decay-time difference ($\Delta t$) between the signal $B$ meson and a flavor-tagging partner $B$ meson. This requires reconstructing the decay vertices of both mesons. However, the $B^0 \to \pi^0\pi^0$ signal is reconstructed from four photons (via $\pi^0 \to \gamma\gamma$), which do not leave charged tracks and thus provide no signal vertex. Consequently, conventional $\Delta t$-based analyses of this channel are infeasible with current data samples, requiring datasets an order of magnitude larger than currently available.

Methodology
This work reports the first measurement of $S_{\pi^0\pi^0}$ using a novel approach that exploits the quantum entanglement of the $B^0\bar{B}^0$ pair produced in $\Upsilon(4S) \to B^0\bar{B}^0$ decays. Instead of measuring the decay-time difference $\Delta t$, the analysis utilizes the decay time of the flavor-tagging meson ($t_{\text{tag}}$) relative to the production point.

• Data Sample: The analysis uses 190 million $\Upsilon(4S) \to B^0\bar{B}^0$ events collected by the Belle II experiment at the SuperKEKB collider, corresponding to an integrated luminosity of 365 fb$^{-1}$.
• Time Evolution Model: Due to quantum entanglement, the probability density of the tag meson decaying at time $t_{\text{tag}}$ with flavor $q$ retains information about the time-dependent CP violation in the signal decay. The asymmetry is parameterized by coefficients $S_{\pi^0\pi^0}$ and $C_{\pi^0\pi^0}$ according to:
  $$P(t_{\text{tag}}, q) \approx \frac{e^{-t_{\text{tag}}/\tau}}{2} \left[ 1 - q D \left( S_{\pi^0\pi^0} \sin[\Delta m(t_{\text{tag}} - \hat{t})] + C_{\pi^0\pi^0} \cos[\Delta m(t_{\text{tag}} - \hat{t})] \right) \right]$$
  where $\tau$ is the $B^0$ lifetime, $\Delta m$ is the mixing frequency, and $D$ is a damping factor arising from integrating over the unmeasured signal decay time.
• Vertexing and Resolution: The $t_{\text{tag}}$ is derived from the displacement between the interaction point (inferred from the beam luminous region) and the reconstructed decay vertex of the tag $B$ meson. The SuperKEKB's micron-scale luminous region and the Belle II pixel detector's vertex resolution enable a per-event $t_{\text{tag}}$ resolution of approximately 1.5 ps.
• Signal Selection and Background: The $B^0 \to \pi^0\pi^0$ signal is reconstructed using only calorimeter information. A boosted-decision-tree classifier is trained on 27 observables to suppress the dominant continuum background ($e^+e^- \to q\bar{q}$), which exceeds the signal by three orders of magnitude. The sample also includes $B\bar{B}$ background, dominated by $B^+ \to \rho^+\pi^0$ decays.
• Calibration and Validation:
  • The $t_{\text{tag}}$ resolution function is calibrated using $36,000$ $B^0 \to D^-\pi^+$ decays.
  • The method is validated by measuring time-dependent CP violation in the benchmark channel $B^0 \to J/\psi K_S^0$ without reconstructing the signal vertex, reproducing known values for $S_{J/\psi K_S^0}$ and $C_{J/\psi K_S^0}$.
• Fit Strategy: An extended maximum-likelihood fit is performed on the unbinned distributions of seven observables ($\Delta E$, $M_{bc}$, continuum suppression output, flavor tag, wrong-tag probability, $t_{\text{tag}}$, and its uncertainty).

Key Results
The analysis yields the following measurements for the CP-violating coefficients:

• Mixing-induced CP violation: $S_{\pi^0\pi^0} = 0.61^{+0.75}_{-0.79} (\text{stat}) \pm 0.11 (\text{syst})$
• Direct CP violation: $C_{\pi^0\pi^0} = 0.05 \pm 0.28 (\text{stat}) \pm 0.07 (\text{syst})$

The result for $S_{\pi^0\pi^0}$ represents the first ever determination of mixing-induced CP violation in this decay mode. The precision achieved is comparable to what would require a dataset twenty times larger using conventional $\Delta t$-based analysis. The $C_{\pi^0\pi^0}$ measurement supersedes the previous time-integrated Belle II result.

Significance and Impact
Incorporating these measurements into the isospin analysis of $B \to \pi\pi$ decays significantly improves the constraints on the CP-violating phase $\phi_2$:

1. Reduction of Ambiguities: The inclusion of $S_{\pi^0\pi^0}$ reduces the number of discrete solutions for $\phi_2$ from eight to six.
2. Precision Improvement: The allowed 68.3% confidence level interval for $\phi_2$ is shortened by 40%.
3. Preferred Solution: The solution preferred by the Standard Model constraints is determined to be $(82.0^{+4.4}_{-3.5})^\circ$, which is in agreement with and comparable in precision to determinations from the $B \to \rho\rho$ system.

The paper concludes that the $t_{\text{tag}}$-based approach makes $B \to \pi\pi$ decays competitive for determining $\phi_2$ much earlier than previously anticipated. This method provides a viable path for more precise tests of the CKM framework in decay channels where the signal vertex cannot be reconstructed, accelerating progress in the search for new physics.