A Parity Signature Distinguishing $SO(5)_1$ from $Spin(4)_1$ Boundaries
Jeff Rohlfing
Abstract
The level-one Wess-Zumino-Witten models $SO(5)_1$ and $Spin(4)_1$ share the same total quantum dimension and therefore the same topological entanglement entropy, yet they differ in the parity of their Majorana channel count: five (odd) versus four (even). We show that this even/odd difference is a sharp, falsifiable boundary signature even though the standard entropic diagnostic is blind to it. Under non-abelian bosonization $SO(5)_1$ is five Majorana fermions, and the odd count leaves one mode unpaired and the spinor primary a single Ising-class sector of quantum dimension $\sqrt{2}$, where the even neighbor $Spin(4)_1$ pairs cleanly and its spinor splits into two abelian sectors. On an explicit coupled-Majorana-chain realization, a class-D $\mathbb{Z}_2$ index and a two-parameter phase scan show that a protected unpaired-Majorana parity persists across an entire gapped phase for the odd ($n=5$) case and is absent for the even ($n=4$) control. On the interacting bilinear-biquadratic chains at the Reshetikhin point, a charge-conserving entanglement-spectrum computation resolves the two cases directly: the $SO(5)_1$ tower carries the five-fold vector multiplet including the zero-weight state, the $Spin(4)_1$ tower a four-fold shell without it, and the topological entanglement entropy $\gamma=\log 2$ is identical for both and cannot separate them where the parity and channel count can. A direct diagonalization of the genuine twisted sector resolves the spinor multiplet, which appears as the tensor square $\mathbf{4}\otimes\mathbf{4}=\mathbf{1}\oplus\mathbf{5}\oplus\mathbf{10}$ rather than the bare chiral spinor, the expected consequence of the Nielsen-Ninomiya obstruction to a single chiral mode on a non-chiral lattice. We also propose a candidate mechanism in which the open non-abelian sector is completed non-locally through a three-dimensional Chern-Simons bulk; the demonstrated lattice results stand independently of this conjecture.