Chiral effective theory and low-energy constants

For two light flavors, package the three pion fields into $U=\exp(i\tau^a\pi^a/f)$, where the Pauli matrices have trace normalization two. The leading derivative term has coefficient f squared divided by four; expanding U gives canonically normalized pion kinetic terms. The constant f is the decay constant in the chiral limit in this convention. Different generator or charged-current normalizations can move square roots of two between a quoted decay constant and the Lagrangian.

Quark masses enter as a spurion, a source assigned a transformation law so that symmetry breaking can be tracked systematically. With real isospin-symmetric light masses, write chi equal to twice B times the quark mass matrix. The leading pion mass squared is twice B times the common quark mass. B and f are not fixed numerically by chiral symmetry alone; they require input from QCD calculations or data. The quark masses and B also have compatible renormalization conventions so that their product gives a physical leading mass scale.

In ordinary purely mesonic chiral perturbation theory, each derivative counts as one power of a soft momentum, while a light quark mass counts as two. This assignment follows the leading pion-mass relation. A connected amplitude in four spacetime dimensions then has chiral order D determined by its number of loops and the chiral orders of its vertices. The counting assumes all relevant soft momenta and pion masses are small relative to the appropriate breakdown scale.

The counting formula shows why one loop made entirely from leading order-two vertices contributes at order four. A tree diagram with an order-four vertex contributes at the same order and must be included alongside it. Canonical operator dimension alone would miss that relationship because U contains arbitrarily many pion fields. The familiar loop ratio involving momentum divided by four pi f is a useful estimate, but nearby omitted resonances and process-dependent enhancements can lower the practical breakdown scale.

Pion loops generate nonanalytic momentum and mass dependence, including chiral logarithms. Their ultraviolet divergences are absorbed by higher-order symmetry-allowed interactions. The associated renormalized low-energy constants depend on the subtraction scale, and that dependence cancels the loop scale dependence in observables. Keeping a logarithm while omitting all same-order analytic terms is not generally a complete prediction unless an additional hierarchy or approximation has been explicitly justified.

A constant fitted in one process can be used in another only with the same normalization, scale, and EFT order, including correlated uncertainties. Lattice QCD can determine constants or vary quark masses, but extrapolating those data still needs a convergence test. Extending from two to three light flavors introduces heavier pseudo-Goldstone masses and can worsen the expansion. The success of the two-flavor soft-pion organization does not imply uniform precision for kaon processes or for energies close to a hadronic resonance.