Anomalies and bosonization (Q2200485)

A \textit{fermionic} quantum field theory \(Z_f\) is one which depends on (``couples to'') a background spin structure \(\eta\) on spacetime. Any fermionic quantum field theory may be \textit{bosonized} by summing over spin structures: \(Z_b := \sum_\eta Z_f(\eta)\), up to a normalizing factor. Spin structures are almost the same as principle \(\mathbb{Z}_2\)-bundles, and so bosonization is almost the same as gauging a \(\mathbb{Z}_2\)-symmetry. The upshot of this analogy is one of the basic observations of \textit{D. Gaiotto} and \textit{A. Kapustin} [in: Gribov-85 memorial volume: exploring quantum field theory. Proceedings of the memorial workshop devoted to the 85th birthday of V. N. Gribov, Landau Institute for Theoretical Physics, Chernogolovka, Russia, June 17--20, 2015. Hackensack, NJ: World Scientific. 464--508 (2017; Zbl 1365.81099)]: a \(D\)-spacetime-dimensional fermionic theory, when bosonized, will carry a distinguished (anomalous) \((D-2)\)-form \(\mathbb{Z}_2\)-symmetry, electromagnetic-dual to the spin structure, and the original fermionic theory \(Z_f\) can be recovered by ``ungauging'' this \((D-2)\)-form \(\mathbb{Z}_2\)-symmetry. Explicitly, \(Z_b\) has distinguished topological line operators which arise as Wilson lines for the (now dynamicalized) spin structure, and these are reinterpretted [\textit{D. Gaiotto} et al., J. High Energy Phys. 2015, No. 2, Paper No. 172, 62 p. (2015; Zbl 1388.83656)] as sourcing the \((D-2)\)-form \(\mathbb{Z}_2\)-symmetry. The present paper refers to \(Z_b\) by itself as the ``0th bosonization'' of \(Z_f\), and \(Z_b\) with its \((D-2)\)-form \(\mathbb{Z}_2\)-symmetry operators remembered as the ``1st bosonization''. Any time a quantum field theory has topological line operators, it also picks up topological surface and higher-dimensional operators known in the physics literature under the names ``Cheshire object'' [\textit{D. V. Else} and \textit{C. Nayak}, ``Cheshire charge in (3+1)-dimensional topological phases'', Phys. Rev. B 96, No. 4, Article ID 045136, 17 p. (2017; \url{doi:10.1103/PhysRevB.96.045136})] and ``condensation descendent'' [\textit{L. Kong} and \textit{X.-G. Wen}, ``Braided fusion categories, gravitational anomalies, and the mathematical framework for topological orders in any dimensions'', Preprint, \url{arXiv:1405.5858}]: these higher-dimensional operators can by built microscopically as networks of the line operators. The \(Z_b\) case is particularly elegant. There is a \(\mathbb{Z}_2\) worth of surface operators, the nontrivial element in which is called the \textit{Kitaev wire} (it measures the Arf-Brown-Kervaire invariant of the dynamical spin structure); and there is a \(\mathbb{Z }\)-worth of 3D membrane operator, generated by the \textit{\(p+ip\) superconductor}. The terms ``2nd bosonization'' and ``3rd bosonization'' are used in this paper to mean \(Z_b\) with these operators remembered. These higher-dimensional operators source higher-form symmetries. The real starting point of this paper is the observation that the Wilson lines, Kitaev wires, and \(p+ip\) superconductors correspond to the first few spin cobordism groups (Anderson-dual to the spin bordism groups) \(\Omega^1_{\mathrm{spin}}(*) = \mathbb{Z}_2\), \(\Omega^2_{\mathrm{spin}}(*) = \mathbb{Z}_2\), and \(\Omega^3_{\mathrm{spin}}(*) = \mathbb{Z}\). In other words, remembering these higher-dimensional operators in \(Z_b\) corresponds climbing the Postnikov tower of \(\Omega^1_{\mathrm{spin}}\). In other other words, one can analyze fermionic \(G\)-SPTs by decorating \(G\)-gauge defects by these operators, and this analysis is nothing but the Atiyah-Hirzebruch spectral sequence for \(\Omega^d_{\mathrm{spin}}(BG)\). The truncation in which only the 0th and 1st layers are included is the Gu-Wen-Freed cohomology of [\textit{X. Chen} et al., ``Symmetry protected topological orders and the group cohomology of their symmetry group'', Phys. Rev. B 87, No. 15, Article ID 155114, 48 p. (2013; \url{doi:10.1103/PhysRevB.87.155114})] and [\textit{D. S. Freed}, J. Differ. Geom. 80, No. 1, 45--77 (2008; Zbl 1146.81050)]. In this way, the present paper is able to translate the algebra of bosonization into the \(d_2\) differentials for this AHSS (up to spacetime dimension 3). In fact, this paper tells a more sophisticated story. In a fermionic theory with a \(G\)-symmetry, the group \(G\) may act nontrivially on the fermions themselves: this is the best way to understand time-reversal symmetries (and the choice \(T^2 = (\pm 1)^F\)) or theories with a \(\mathrm{Spin}^c\) or \(\mathrm{Spin}\)-\(\mathrm{SU}(2)\) connection. In this case, one needs to compute \textit{twisted} spin cobordism groups of \(BG\), and the present paper works out the \(d_2\) differentials for this generalization. Many other features of the story are analyzed (and told in a satisfying amount of physical detail), including the bosonization of boundary conditions and the relation between anomalous symmetries and SPTs in one higher dimension.