S-wave pi-pi I=0 and I=2 scattering at physical pion mass

Tianle Wang, Christopher KellyProceedings of 37th International Symposium on Lattice Field Theory — PoS(LATTICE2019) 2020 · DOI

Abstract

The $\pi\pi$ scattering phase shifts for both the s-wave $I=0$ and $I=2$ channels are determined from a lattice calculation performed on 741 gauge configurations obeying G-parity boundary conditions with a physical pion mass and lattice size $32^3\times 64$. The phase shifts are determined for both stationary and moving $\pi\pi$ systems, at four different center of mass momenta. We implement several interpolating operators including a sigma operator and $\pi\pi$ operator with the pions carrying relatively larger momentum. We use both correlated fitting and the generalized eigenvalue treatment and compare these two results with dispersive predictions. A detailed excited state contamination error analysis is included.

Introduction

Understanding CP violation in kaon decays ($K \to \pi\pi$), specifically the parameter $Re(\epsilon'/\epsilon)$, is a major objective in particle physics. Theoretical predictions from the Standard Model are notoriously difficult because they involve the non-perturbative regime of Quantum Chromodynamics (QCD). A critical component of this calculation is the interaction between the two final-state pions, described by scattering phase shifts.

Previous lattice calculations showed significant discrepancies (up to 7$\sigma$) compared to dispersive methods based on experimental data. This study addresses the likely culprit: excited state contamination, where higher-energy states pollute the measurement of the ground-state energy on the lattice. By isolating the ground state more effectively, the authors aim to reconcile lattice results with theoretical expectations.

Methodology

The research team utilized Lattice QCD on a $32^3 \times 64$ grid with 741 gauge configurations and a physical pion mass. Key technical innovations included:

  • G-parity Boundary Conditions (GPBC): Allowing pions to carry non-zero momentum in a finite box, essential for studying scattering at various energy levels.
  • Multiple Interpolating Operators: To isolate the ground state, the team used a basis of operators. For the noisy $I=0$ channel, they introduced a 2-quark sigma ($\sigma$) operator alongside standard 4-quark $\pi\pi$ operators.
  • Advanced Analysis: The Generalized Eigenvalue Problem (GEVP) and multi-state correlated fits were employed to extract energies, supplemented by a new "frozen fit" method to estimate systematic errors from unincluded states.

Results and Discussion

The results demonstrate that the $I=2$ channel is relatively clean and consistent with dispersive predictions. However, the $I=0$ channel requires more sophisticated treatment to handle excited state contamination.

Excited state contamination in a moving frame $\pi\pi$ system. The plot shows the dependence of the fitted ground-state energy on the minimum time $t_{min}$ used in the fit. Without the $\sigma$ operator, the results (black and red points) remain above the dispersive prediction (solid line).

As shown in the figure above, moving frame calculations using only $\pi\pi$ operators exhibit clear excited state contamination, as indicated by the downward trend with increasing $t_{min}$. In contrast, the inclusion of the $\sigma$ operator in stationary frame calculations significantly suppresses this contamination.

The critical role of the sigma ($\sigma$) operator in the $I=0$ channel. The inclusion of $\sigma$ (blue and magenta points) dramatically reduces the discrepancy with the dispersive prediction (solid line) compared to using only $\pi\pi$ operators (grey and red points).

With the $\sigma$ operator, the lattice results for the stationary $I=0$ channel at the kaon mass yield a phase shift of $32.4(2.3)^\circ$, which is highly consistent with dispersive predictions. This reconciliation helps bridge a long-standing gap in the field.

Conclusion and Future Work

This work confirms that once excited states are properly handled—primarily through the use of the $\sigma$ operator—Lattice QCD results for $\pi\pi$ scattering are consistent with dispersive theory. While stationary frame results are now robust, moving frame $I=0$ calculations still suffer from visible contamination. Future work will focus on incorporating the $\sigma$ operator into moving frame analyses to further reduce systematic uncertainties in $Re(\epsilon'/\epsilon)$ predictions.

Summary updated

Abstract

The pi-pi scattering phase shifts for both the s-wave I=0 and I=2 channels are determined from a lattice calculation performed on 741 gauge configurations obeying G-parity boundary conditions with a physical pion mass and lattice size $32^3\times 64$. The phase shifts are determined for both stationary and moving pi-pi systems, at four different center of mass momenta. We implement several interpolating operators including a sigma operator and $\pi\pi$ operator with the pions carrying relatively larger momentum. We use both correlated fitting and the generalized eigenvalue treatment and compare these two results with dispersive predictions. A detailed excited state contamination error analysis is included.
BibTeX
@inproceedings{wang2020s,
  author = {Tianle Wang and Christopher Kelly},
  title = {{S-wave pi-pi I=0 and I=2 scattering at physical pion mass}},
  booktitle = {Proceedings of 37th International Symposium on Lattice Field Theory — PoS(LATTICE2019)},
  year = {2020},
  doi = {10.22323/1.363.0142},
}
Figure