3+1d Boundaries with Gravitational Anomaly of 4+1d Invertible Topological Order for Branch-Independent Bosonic Systems

Abstract

We study bosonic systems on a spacetime lattice defined by path integrals of commuting fields. We introduce branch-independent bosonic (BIB) systems, whose path integral is independent of the branch structure of the spacetime simplicial complex, even for a spacetime with boundaries. In contrast, a generic lattice bosonic (GLB) system's path integral may depend on the branch structure. We find the invertible topological order characterized by the Stiefel-Whitney cocycle (e.g., 4+1d w${}_2$w${}_3$) to be nontrivial for BIB systems, but this topological order and a trivial gapped tensor product state belong to the same phase for GLB systems. The invertible topological orders in GLB systems are not classified by the oriented cobordism. The branch dependence on a lattice may be related to the orthonormal frame of smooth manifolds and the framing anomaly of continuum field theories. The branch structure on a discretized lattice may be related to a frame structure on a smooth manifold that trivializes any Stiefel-Whitney classes. We construct BIB systems to realize the w${}_2$w${}_3$ topological order, and its 3+1d gapped or gapless boundaries. A 3+1d ${\mathbb{Z}}_2$ gauge theory with (1) fermionic ${\mathbb{Z}}_2$ gauge charge particle trivializes w${}_2$ and (2) fermionic ${\mathbb{Z}}_2$ gauge flux line trivializes w${}_3$. In particular, if the flux loop's worldsheet is unorientable, then an orientation-reversal 1d worldline corresponds to a fermion worldline carrying no ${\mathbb{Z}}_2$ gauge charge. Spin and Spin${}^c$ structures trivialize the w${}_2$w${}_3$ global pure gravitational anomaly to zero (which helps to construct 3+1d ${\mathbb{Z}}_2$ and all-fermion U(1) gauge theories), but the Spin${}^h$ and Spin${\times }_{{\mathbb{Z}}_2}$Spin$(n\ge 3)$ structures modify the w${}_2$w${}_3$ into a global mixed gauge-gravitational anomaly, which helps to constrain Grand Unifications (e.g., $n=10,18$) or construct new models.

Comments: 34 pages two columns (many appendices on the toolkits of cochain, cocycle, cohomology, characteristic class, cobordism, anomalies, and higher dimensional bosonization with emergent fermions). This work includes not only unpublished work in 2018 after arXiv:1810.00844, but also many new results. v2: Phys. Rev. B accepted