Symmetry protected topological phases from decorated domain walls

TL;DR

This paper introduces a physical picture for understanding and constructing symmetry protected topological (SPT) phases in various dimensions by decorating domain walls with lower-dimensional SPT phases, providing intuitive wavefunctions and connections to cohomology.

Contribution

It presents a novel domain wall decoration approach to realize and analyze SPT phases across different dimensions, linking physical intuition with mathematical classification.

Findings

SPT phases can be understood via decorated domain walls.

Ground states match group cohomology and field theory results.

Gauging symmetries yields new insights into phase properties.

Abstract

Symmetry protected topological (SPT) phases with unusual edge excitations can emerge in strongly interacting bosonic systems and are classified in terms of the cohomology of their symmetry groups. Here we provide a physical picture that leads to an intuitive understanding and wavefunctions for several SPT phases in d=1,2,3 dimensions. We consider symmetries which include a Z_2 subgroup, that allows us to define domain walls. While the usual disordered phase is obtained by proliferating domain walls, we show that SPT phases are realized when these proliferated domain walls are `decorated', i.e. are themselves SPT phases in one lower dimension. For example a d=2 SPT phase with Z_2 and time reversal symmetry is realized when the domain walls that proliferate are themselves in a d=1 Haldane/AKLT state. Similarly, d=3 SPT phases with Z_2 * Z_2 symmetry emerges when domain walls in a d=2 SPT with Z_2 symmetry are proliferated. The resulting ground states are shown to be equivalent to that obtained from group cohomology and field theoretical techniques. The result of gauging the Z_2 symmetry in these phases is also discussed. An extension of this construction where time reversal plays the role of Z_2 symmetry allows for a discussion of several d=3 SPT phases. This construction also leads to a new perspective on some well known d=1 SPT phases, from which exactly soluble parent Hamiltonians may be derived.