nLab symmetry protected topological phase



Solid state physics

Topological physics

Representation theory



In solid state physics, by a phase of matter which is

or more generally

one means a topological phase of matter which is GG-equivariantly non-trivial, in that it cannot be adiabatically deformed to a trivial phase while respecting some GG-symmetry. In case of SPT one in addition requires that the underlying topological phase (forgetting the symmetry) is trivial, while in case of SET this constraint is not implied.

Since, if one forgets (theoretically) the GG-equivariance, a GG-equivariantly non-trivial SPT may be trivial as a plain topological phase, and two distinct SETs may be equivalent as plain topological phases, one may regard the GG-symmetry as “protecting” an SPT phase from decaying and as “enriching” one plain phase to several distinct SET phases – whence the terminology.

In other words:

  1. distinct SPT/SET phases with a given symmetry cannot be adiabatically deformed into each other, in particular not without going through a phase transition, if the whole deformation preserves the symmetry;

  2. but they may possibly be transformed into each other this way if the symmetry is broken during the deformation.

There are the following types of symmetry groups GG to which the concept of crystalline SPT/SET phases applies (as well as to any of their semidirect combinations):

(table from SS 22)
  1. GG may be an “external” spatial symmetry, namely a subgroup of the crystallographic point group of the underlying crystalline material, canonically acting on the Brillouin torus (BT) 𝕋 d\mathbb{T}^d;

    for example inversion II is the sign reversal involution on d𝕋 d\mathbb{R}^d \twoheadrightarrow \mathbb{T}^d.

    If a topological insulator-phase is “protected” by a crystallographic point group symmetry this way (possibly including time-reversal symmetry), then one speaks of a topological crystalline insulator.

  2. GG may be time reversal symmetry or charge reversal, which acts as II on the Brillouin torus, but in addition acts on the Hamiltonian by complex conjugation and, respectively, sign reversal;

  3. GG may be an internal symmetry which acts not on the position but on the “internal degrees of freedom” of (electrons in the) substance, which are, typically, located at each atomic site (whence also: “on-site symmetry”).

    The prime example of an internal symmetry are (finite) subgroups of the Spin(3)\simeqSU(2)-group which acts on the electron spins. This will be a symmetry to the extent that spin interactions (such as the spin-orbit coupling or the intrinsic Lorentz force due to an external magnetic field) are negligible.



  1. Beware that some authors insist on using the specific term “SPT” only for items further down in this list. For instance the claim below that SPT’s are “classified by group cohomology” applies really to the last item (see below).

    The first two items instead, for the case of free fermion systems at least, are expected to have a classification in twisted equivariant topological K-theory (see at K-theory classification of topological phases of matter).

    For the conceptual relation between these cases see further below.

  2. Beware that the first articles on the topic (Gu & Wen 09, PBTO 09) actually used the term “symmetry protected topological order”. This could be perceived as somewhat of a misnomer, since in examples the (symmetry protected) “topological order” is often trivial (in that the ground state is non-degenerate and/or the Berry connection is abelian) even though the underlying (symmetry protected) topological phase is non-trivial (ie. the valence bundle of a topological insulator has a non-trivial K-class).

    Due to this problem, the original authors argued that “SPT” could also stand for “symmetric protected trivial order” (X.-G. Wen, Sep 18, 2014). But then it seems more descriptive (and now fairly widely accepted) to speak of “symmetry protected topological phase”. (See also at red herring principle.)


Spatial symmetry

For more on spatial-SPT topological crystalline insulator-phases, see there for concrete examples.

Internal symmetry

More examples of internal-SPT: Ye & Wen 13, CLV 14, Yang-Liu 18



Spatial symmetries

Symmetry protected crystalline phases where the dynamics of the electrons may approximately be regarded free (but subject to the the atomic lattice Coulomb background field, see here) are thought to be classified by twisted equivariant topological K-theory of the Brillouin torus. This is the statement of the K-theory classification of topological phases of matter.

Beware that this case has mostly been discussed for CPT-symmetries such as time-reversal symmetry (see at topological insulator) and for crystallographic symmetries (see at topological crystalline insulator), not so much for internal (“on site”) symmetries (but see Wen 12).

Internal symmetries

In contrast, a widely cited claim CGLW 11 asserts (motivation is offered in CLW 11, Sec. V) that “bosonic” SPT orders for an internal symmetry group G intG_{int} are given by group cohomology H d+1(G;U(1))H^{d+1}\big(G;\, U(1)\big) of G intG_{int} with coefficients in the circle group and in degree d+1d + 1, for dd the effective dimension of the given material (in practice: d{0,1,2,3}d \in \{0,1,2,3\}). This claim was generalized to fermionic SPT orders via a kind of group-supercohomology (Gu & Wen14).

From X.-G. Wen (2013, in rev 1):

So the group (super-)cohomology theory may allow us to classify all SPT orders even for interacting systems, which include interacting topological insulator/superconductor.

But conceptual problems have remained with this proposal:

From Wang & Senthil 14, p. 1

this classification is now known not to be complete

From Xiong 18:

Despite tremendous progress, a complete classification of SPT phases for arbitrary symmetries in arbitrary dimensions remains elusive. A number of classification proposals have been made in the general case: the Borel group cohomology proposal [[CGLW11]], the oriented cobordism proposal, the Freed-Hopkins proposal, and Kitaev’s proposal in the bosonic case; and the group supercohomology proposal [[GW14]], the spin cobordism proposal, the Freed-Hopkins proposal, and Kitaev’s proposal in the fermionic case. These proposals give differing predictions in certain dimensions for certain symmetry groups that cannot be attributed to differences in definitions or conventions. While more careful analysis has uncovered previously overlooked phases and input from topological field theories has brought us closer than ever to our destination, we believe that we can do much more.

From BBCW 19, p. 3:

Although a remarkable amount of progress has been made on these deeply interrelated topics, a completely general understanding is lacking, and many questions remain. For example, although there are many partial results, the current understanding of fractionalization of quantum numbers, along with the classification and characterization of SETs is incomplete.

Moreover, while there have been many results towards understanding the properties of extrinsic defects in topological phases, there has been no general systematic understanding and, in particular, no concrete method of computing all the rich topological properties of the defects for an arbitrary topological phase. The study of topological phase transitions between different topological phases is also missing a general theory. In this paper, we develop a general systematic framework to understand these problems.

The proposal of BBCW 19 (may not yet have a physics “proof” either, but) is conceptually transparent: The authors assume, as often done, that a topological order with anyonic defects is characterized by a unitary fusion category 𝒞\mathcal{C}, and then propose that a G intG_{int}-SPT-phase Φ\Phi with this underlying topological order is what we may equivalently recognize as:

  1. the equivalence class [Φ][\Phi] of an ∞-action of G intG_{int} on 𝒞\mathcal{C}, namely

  2. the cohomology-class of a 2-group-homomorphism from G intG_{int} to the automorphism 2-group of 𝒞\mathcal{C}:

    (1)ΦHom(G int,Aut(𝒞)), \Phi \;\in\; Hom \big( G_{int} ,\, Aut(\mathcal{C}) \big) \,,

    (this is essentially BBCW 19, (1) – where it says “group action”, but later from (81) on (p. 14) it transpires that the correct 2-group-action is indeed meant),


  3. the pseudonatural transformation-class of a weak 2-functor

    ΦMaps(BG int,BAut(𝒞)), \Phi \;\in\; Maps \big( \mathbf{B}G_{int} ,\, \mathbf{B}Aut(\mathcal{C}) \big) \,,

    from the delooping groupoid of G intG_{int} to that delooping 2-groupoid of the automorphism 2-group of 𝒞\mathcal{C}.

This is plausible (relative to the assumption that 𝒞\mathcal{C} characterizes the un-proteced topological order) since under an “internal symmetry” one wants to mean a global group action under which all other constructions have equivariant-structure, and (1) is exactly the data that equips anyon-species (the simple objects of 𝒞\mathcal{C}) and their fusion (the tensor product) and braiding (the braiding) with such G intG_{int}-equivariant structure.

But neither of these proposals connects recognizably to the twisted equivariant K-theory classification of topological phases (TE-K). While the latter may not apply to all cases of symmetry symmetry protected topological phases, it certainly applies to some of them, and it currently stands out among all other proposals on the classification of topological phases of matter as being the one with the most detailed support by theory and experiment. Therefore it would be reassuring to see how any other classification proposal plausibly connects to the TE-K proposal in appropriate special cases (notably for phases well-approximated by free fermion dynamics).

Conversely, inspection of the mathematical construction of twisted equivariant K-theory (as made explicit in SS 21, (4.1.28)) readily reveals the evident way in which it subsumes internal symmetries and how the resulting structure has an equivalent expression in terms of ∞-group cohomology. This is an immediate consequence of the mapping stack- \infty -adjunction, as shown the following diagram

(graphics from SS 22b)

Here the mapping stack adjunction…


Original articles

See also:


See also:

Review with focus on the case of topological insulators protected by crystallographic group-symmetry:

A proposal for classifying symmetry-protected topological order in terms of 2-groupoidal-actions on modular tensor categories:

See also:

Some of the above material is taken from:


Of free fermionic SPT phases
Of bosonic and interacting SPT phases

Claim of classification of SPT phases via group cohomology:

Classification for free fermion SPT phases in twisted equivariant K-theory:

For more on this see at K-theory classification of topological phases of matter

Proposal that the general classification involves 2-group-symmetries:

Proposal that the general classification involves equivariant Whitehead-generalized cohomology theory:


The 1d Haldane phase:

Examples with braid group effects:

and with loop braid group-effects:

  • Chao-Ming Jian, Xiao-Liang Qi, Layer Construction of 3D Topological States and String Braiding Statistics, Phys. Rev. X 4 (2014) 041043 [[doi:10.1103/PhysRevX.4.041043]]

and with both:

  • Zhen Bi, Yi-Zhuang You, Cenke Xu, Anyon and loop braiding statistics in field theories with a topological Θ\Theta term, Phys. Rev. B 90 (2014) 081110(R) (doi:10.1103/PhysRevB.90.081110)
  • Michael Levin, Zheng-Cheng Gu, Braiding statistics approach to symmetry-protected topological phases, Phys. Rev. B 86, 115109 (2012), arXiv:1202.3120.

  • Yuan-Ming Lu, Ashvin Vishwanath, Theory and classification of interacting ‘integer’ topological phases in two dimensions: A Chern-Simons approach, Phys. Rev. B 86, 125119 (2012), arXiv:1205.3156.

  • Davide Gaiotto, Theo Johnson-Freyd, Symmetry protected topological phases and generalized cohomology [[arxiv/1712.07950]]

  • Yizhi You, Trithep Devakul, F. J. Burnell, Titus Neupert, Higher order symmetry-protected topological states for interacting bosons and fermions, Phys. Rev. B 98 (2018) 235102 (arXiv:1807.09788v2, doi:10.1103/PhysRevB.98.235102)

  • Rongge Xu, Zhi-Hao Zhang, Categorical descriptions of 1-dimensional gapped phases with abelian onsite symmetries [[arXiv:2205.09656]]

Observation in experiment:

  • Zhihuang Luo, Wenzhao Zhang, Xinfang Nie, Dawei Lu, Observation of a symmetry-protected topological phase in external magnetic fields [arXiv2208.05357]

Conference and seminar cycles

Last revised on November 1, 2023 at 07:57:06. See the history of this page for a list of all contributions to it.