Spin-space Groups: Theory and Applications

• Spin-space Groups (SSGs) are symmetry groups that decouple spin and lattice operations, generalizing both ordinary and magnetic space groups.
• They classify periodic magnetic structures in weak-spin-orbit-coupling systems, providing insights into band degeneracies, tensor responses, and topological phenomena.
• SSG enumeration and representation theory enable systematic identification of magnetic orders and band nodes, influencing materials design and superconductivity research.

Searching arXiv for papers on spin-space groups and related representation theory. Spin-space groups (SSGs) are the symmetry groups appropriate to periodic magnetic structures when spin-space operations and real-space operations must be treated jointly but not rigidly locked to one another. They generalize ordinary space groups and magnetic space groups by allowing a symmetry operation to combine an independent spin-space orthogonal transformation with a crystallographic operation, and they are the natural symmetry language for magnetic materials in the weak-spin-orbit-coupling regime, including collinear, coplanar, noncoplanar, commensurate, and incommensurate spiral orders (Song et al., 2024, Xiao et al., 2023). In this framework, magnetic order is described as a periodic vector or pseudovector field, and the resulting symmetry theory governs not only magnetic structure itself but also band degeneracies, magnons, tensor responses, superconducting order parameters, and the symmetry reduction induced by spin-orbit coupling (Jiang et al., 2023, Etxebarria et al., 6 Feb 2025, Feng et al., 2024).

1. Definition and group-theoretic setting

The defining feature of an SSG is the decoupling of spin and lattice operations. In the notation adopted for three-dimensional magnetic crystals, an SSG is written as

$$\mathcal G^{(S)}=\bigcup_{i=1}^{n} \{U_i || R_i|\bm{\tau}_i\} \bm{T},$$

where $U_i\in O(3)$ acts on spin space, $R_i$ is a lattice point operation, $\bm{\tau}_i$ is a fractional translation, and $\bm T$ is the pure translation group (Song et al., 2024). The action on a magnetic moment field $\mathbf M(\mathbf r)$ is

$$\{U||R|\boldsymbol\tau\}\mathbf M(\mathbf r)=U\,\mathbf M(\{R|\boldsymbol\tau\}^{-1}\mathbf r),$$

which differs from the locked axial-vector action of magnetic space groups by allowing $U$ to be independent of $R$ (Jiang et al., 2023).

Time reversal is incorporated through improper spin-space operations. In the common convention,

$$g=\left\{ \begin{aligned} &\{U || R_i|\bm{\tau}_i\}, ~\text{if }\det(U)=+1\ &\{-U || R_i|\bm{\tau}_i\}\mathcal{T}, ~\text{if }\det(U)=-1. \end{aligned} \right.$$

Thus antiunitary structure is encoded in the spin-space part rather than appended purely as a real-space decoration (Song et al., 2024). This is the point at which SSGs become strictly broader than MSGs: MSGs are recovered only in the restricted case where the spin and lattice parts are locked, while ordinary space groups arise for nonmagnetic systems after quotienting by the full spin-only symmetry (Song et al., 2024).

A central subgroup is the spin-only group

$U_i\in O(3)$0

and the decomposition

$U_i\in O(3)$1

makes explicit how magnetic geometry is organized by residual pure-spin symmetry (Song et al., 2024). Closely related point-group objects are the spin point groups $U_i\in O(3)$2, which govern tensor constraints and co-representation theory at the point-group level (Etxebarria et al., 6 Feb 2025, Schiff et al., 2023).

2. Spin-only structure, magnetic taxonomy, and related symmetry classes

The classification of SSGs is organized by the spin-only subgroup, which encodes whether the magnetic structure is nonmagnetic, collinear, coplanar, or noncoplanar. In the formulation used for magnetic materials with weak SOC, the four standard cases are (Song et al., 2024):

• nonmagnetic:

$U_i\in O(3)$3

• collinear:

$U_i\in O(3)$4

• coplanar:

$U_i\in O(3)$5

• noncoplanar:

$U_i\in O(3)$6

The same structural idea appears in the enumeration framework based on invariant subgroups of ordinary space groups. There the spin-only-free part of the SSG is classified via quotient groups $U_i\in O(3)$7, and the representation dimension required to realize the spin part reduces with magnetic dimensionality: 3D real representations for general noncoplanar order, 2D real representations for coplanar order, and 1D real representations $U_i\in O(3)$8 for collinear order (Jiang et al., 2023). This is why collinear and coplanar families admit simplifications absent in general noncoplanar cases.

For spin point groups, the intrinsic spin-only symmetry of collinear and coplanar order is stated in point-group language as $U_i\in O(3)$9 for collinear structures and $R_i$0 for coplanar structures, while noncoplanar structures have only the identity intrinsically (Etxebarria et al., 6 Feb 2025). This point-group formulation is especially useful for tensor properties because it implies that collinearity and coplanarity are symmetry-protected in the SOC-free limit.

Several related notions extend this taxonomy. The paper on crystallographic spin point groups distinguishes nontrivial spin point groups, spin-only groups, and total spin groups $R_i$1, and shows that the genuinely new co-representations appear in collinear total spin groups because of continuous rotational freedom (Schiff et al., 2023). The oriented-SSG framework further refines the classification by fixing the relative orientation between spin space and lattice space, thereby bridging the exchange-dominated SSG description and the SOC-locked MSG description (Yu et al., 23 Apr 2026, Nomoto et al., 22 Jan 2026). This suggests that SSG classification captures magnetic geometry, while oriented SSGs and MSGs resolve how that geometry is embedded relative to the crystal axes once SOC is active.

3. Enumeration, nomenclature, and symmetry identification

The systematic enumeration of SSGs proceeds from invariant subgroups of ordinary space groups. Given a space group $R_i$2, one chooses an invariant subgroup $R_i$3, computes the finite quotient group $R_i$4, and realizes its elements through faithful real orthogonal representations to construct

$R_i$5

This is the core constructive formula behind the large-scale SSG databases (Jiang et al., 2023).

Because the number of all SSGs is infinite in principle, practical enumerations impose a cutoff on magnetic supercell size. One enumeration based on invariant subgroups of space groups considered magnetic unit cells up to 12 times the crystal unit cell and obtained $R_i$6 inequivalent non-coplanar SSGs, $R_i$7 inequivalent coplanar non-collinear SSGs, and $R_i$8 inequivalent collinear SSGs (Jiang et al., 2023). A parallel classification based on $R_i$9 representations reported complete classifications of $\bm{\tau}_i$0, $\bm{\tau}_i$1, and $\bm{\tau}_i$2 distinct SSGs for collinear, coplanar, and non-coplanar magnetism, respectively (Xiao et al., 2023). A further enumeration and representation-theory study established an extensive collection of over 100000 SSGs under a four-index nomenclature and International notation, starting from the 230 crystallographic space groups and finite translation groups with a maximum order of 8 (Chen et al., 2023). These differences reflect distinct but related counting conventions and equivalence relations.

The nomenclature records the parent space-group number, the magnetic supercell index, the point-quotient index, and the representation label. In one scheme, an SSG is denoted by

$\bm{\tau}_i$3

with collinear and coplanar cases marked by suffixes $\bm{\tau}_i$4 and $\bm{\tau}_i$5 (Jiang et al., 2023). Another closely related notation is used in the later representation-theory and oriented-SSG literature (Chen et al., 2023, Nomoto et al., 22 Jan 2026). The multiplicity of conventions is one reason recent work has emphasized standardization and interoperable data formats, including the spin crystallographic information file and online platforms such as FINDSPINGROUP (Yu et al., 23 Apr 2026).

For realistic magnetic structures, SSG identification is algorithmic. One workflow is: determine the parent crystal symmetry with spglib, identify admissible spin-space point-group operations from the set of moment directions, test all combinations that preserve the magnetic structure, extract the relevant subgroup and quotient data, and match the result against a database (Jiang et al., 2023). An exhaustive symmetry-search algorithm for commensurate spin arrangements formulates the recovery of the spin rotation part as an orthogonal Procrustes problem and is implemented in spinspg v0.1.1 (Shinohara et al., 2023). Database-scale applications have already identified SSGs for $\bm{\tau}_i$6 magnetic materials from MAGNDATA in one study (Jiang et al., 2023), and another study identified the SSG for each of the 1604 published magnetic structures in MAGNDATA, with body text discussing 1595 published experimental structures after excluding cases with non-integer occupancies (Xiao et al., 2023).

4. Little groups, projective irreducible representations, and band theory

At fixed crystal momentum $\bm{\tau}_i$7, the symmetry analysis is controlled by the little group

$\bm{\tau}_i$8

with $\bm{\tau}_i$9 a reciprocal lattice vector, and by the finite little co-group

$\bm T$0

Because quotienting out translations leaves residual phase factors and because spinful quasiparticles carry double-valued spin rotations, the relevant representations are generally projective irreducible representations rather than ordinary linear irreps (Song et al., 2024).

For antiunitary groups $\bm T$1, a projective representation satisfies

$\bm T$2

and the factor system obeys the antiunitary cocycle condition

$\bm T$3

Equivalent factor systems differ by a $\bm T$4 gauge transformation, so projective irreps are classified by cohomology classes of $\bm T$5 (Song et al., 2024). In SSGs, the total factor system has two sources: $\bm T$6 Here $\bm T$7 is the spin-$\bm T$8 factor from the double cover of $\bm T$9, while the nonsymmorphic translation factor is

$\mathbf M(\mathbf r)$0

For bosons such as magnons, the spinor factor is set to $\mathbf M(\mathbf r)$1, but $\mathbf M(\mathbf r)$2 remains operative (Song et al., 2024). This is why SSGs can be abstractly isomorphic to MSGs yet yield different degeneracy structures and different band consequences.

The paper “Constructions and Applications of Irreducible Representations of Spin-Space Groups” develops a practical route to all irreps by constructing the projective regular representation and decomposing it with a Hamiltonian method (Song et al., 2024). For a $\mathbf M(\mathbf r)$3-dimensional representation $\mathbf M(\mathbf r)$4, one defines

$\mathbf M(\mathbf r)$5

with $\mathbf M(\mathbf r)$6 a random matrix. Because $\mathbf M(\mathbf r)$7 commutes with all representation matrices, its eigenspaces resolve the reducible representation into irreducible components. This provides a uniform computational procedure for unitary and antiunitary projective groups (Song et al., 2024).

The same representation data determine $\mathbf M(\mathbf r)$8 Hamiltonians. If $\mathbf M(\mathbf r)$9 transforms by a projective representation $$\{U||R|\boldsymbol\tau\}\mathbf M(\mathbf r)=U\,\mathbf M(\{R|\boldsymbol\tau\}^{-1}\mathbf r),$$0, then the effective Hamiltonian obeys

$$\{U||R|\boldsymbol\tau\}\mathbf M(\mathbf r)=U\,\mathbf M(\{R|\boldsymbol\tau\}^{-1}\mathbf r),$$1

For linear terms, one obtains the antiunitary compatibility condition involving the dual vector representation, and a criterion for allowed linear dispersion is expressed through the indicator $$\{U||R|\boldsymbol\tau\}\mathbf M(\mathbf r)=U\,\mathbf M(\{R|\boldsymbol\tau\}^{-1}\mathbf r),$$2 (Song et al., 2024). A $$\{U||R|\boldsymbol\tau\}\mathbf M(\mathbf r)=U\,\mathbf M(\{R|\boldsymbol\tau\}^{-1}\mathbf r),$$3-dimensional irrep implies at least $$\{U||R|\boldsymbol\tau\}\mathbf M(\mathbf r)=U\,\mathbf M(\{R|\boldsymbol\tau\}^{-1}\mathbf r),$$4-fold degeneracy, and crossings between inequivalent irreps are symmetry-allowed. This is the representation-theoretic basis of SSG-protected band nodes, Dirac-like points, triple points, and nodal planes.

The worked example $$\{U||R|\boldsymbol\tau\}\mathbf M(\mathbf r)=U\,\mathbf M(\{R|\boldsymbol\tau\}^{-1}\mathbf r),$$5 illustrates these mechanisms. Its SSG is identified as

$$\{U||R|\boldsymbol\tau\}\mathbf M(\mathbf r)=U\,\mathbf M(\{R|\boldsymbol\tau\}^{-1}\mathbf r),$$6

where “P” denotes coplanar order (Song et al., 2024). Along $$\{U||R|\boldsymbol\tau\}\mathbf M(\mathbf r)=U\,\mathbf M(\{R|\boldsymbol\tau\}^{-1}\mathbf r),$$7-$$\{U||R|\boldsymbol\tau\}\mathbf M(\mathbf r)=U\,\mathbf M(\{R|\boldsymbol\tau\}^{-1}\mathbf r),$$8, the little co-group contains 1D irreps $$\{U||R|\boldsymbol\tau\}\mathbf M(\mathbf r)=U\,\mathbf M(\{R|\boldsymbol\tau\}^{-1}\mathbf r),$$9 and 2D irreps $U$0; a singly degenerate $U$1 band crosses a doubly degenerate $U$2 band, producing a threefold crossing. At the $U$3 point, a 4D irrep $U$4 enforces a fourfold Dirac-like degeneracy (Song et al., 2024). Under weak finite SOC the symmetry descends to an MSG with only 1D irreps along $U$5-$U$6, so the SSG-protected crossing is split or weakly avoided, and the line degeneracy is only approximately preserved (Song et al., 2024). This provides a concrete example of SSG symmetry as an exact nonrelativistic organizing principle and an approximate weak-SOC organizing principle.

5. Physical consequences beyond conventional magnetic-space-group analysis

SSGs affect electronic and bosonic spectra, momentum-space spin textures, and topological responses in ways that are not reducible to MSG symmetry alone. One study emphasizes that SSG electronic bands exhibit features never seen in MSGs, including nonsymmorphic SSG Brillouin zones, non-commuting translations behaving like an effective $U$7-flux, and unconventional spin-momentum locking determined purely by SSG symmetry (Xiao et al., 2023). In that framework, SSG Bloch states are labeled by an SSG momentum $U$8, and the transformed spin texture satisfies

$U$9

so the momentum-space spin pattern is symmetry-enforced rather than model-dependent (Xiao et al., 2023).

A unified SSG-based classification of magnetic orders exploits the fact that symmetry constrains real-space and reciprocal-space spin arrangements differently. The action on momentum-space spin textures contains an additional factor $R$0,

$R$1

which makes the relation between $R$2 and $R$3 richer than in conventional magnetic classifications (Song et al., 9 Dec 2025). Within this scheme, altermagnets, coplanar $R$4-wave magnets, and the predicted coplanar even-wave magnets are distinct SSG symmetry classes rather than unrelated phenomena (Song et al., 9 Dec 2025).

Topological diagnosis can likewise be upgraded from MSG to SSG. A recent symmetry-indicator program for collinear magnets treats SSGs as approximate symmetry groups neglecting SOC and MSGs as their SOC-reduced descendants, thereby exposing topological band structures invisible in MSG-only analysis (Huang et al., 8 Jan 2026). The paper classifies electronic topology across 484 experimentally synthesized collinear magnets and reports, for example, real triple points in ferromagnetic $R$5, Dirac nodal lines at generic $R$6-points in antiferromagnetic $R$7, and Weyl nodal lines in altermagnetic $R$8 (Huang et al., 8 Jan 2026). The case of $R$9 is especially notable because it is topologically trivial under MSG but hosts Dirac nodal lines within the SSG framework; when SOC gaps these lines, a sizable anomalous Hall conductivity appears (Huang et al., 8 Jan 2026). This suggests that nonrelativistic SSG topology can remain physically consequential after the exact symmetry has descended to an MSG.

For response tensors, the associated spin point group rather than the MSG is the correct symmetry object in the SOC-free limit. The tensor formalism developed for spin point groups shows that spin parts transform with $$g=\left\{ \begin{aligned} &\{U || R_i|\bm{\tau}_i\}, ~\text{if }\det(U)=+1\ &\{-U || R_i|\bm{\tau}_i\}\mathcal{T}, ~\text{if }\det(U)=-1. \end{aligned} \right.$$0-indices, orbital parts with $$g=\left\{ \begin{aligned} &\{U || R_i|\bm{\tau}_i\}, ~\text{if }\det(U)=+1\ &\{-U || R_i|\bm{\tau}_i\}\mathcal{T}, ~\text{if }\det(U)=-1. \end{aligned} \right.$$1-indices plus a time-reversal sign from $$g=\left\{ \begin{aligned} &\{U || R_i|\bm{\tau}_i\}, ~\text{if }\det(U)=+1\ &\{-U || R_i|\bm{\tau}_i\}\mathcal{T}, ~\text{if }\det(U)=-1. \end{aligned} \right.$$2, and that in collinear and coplanar structures the effective magnetic point group is always a gray group $$g=\left\{ \begin{aligned} &\{U || R_i|\bm{\tau}_i\}, ~\text{if }\det(U)=+1\ &\{-U || R_i|\bm{\tau}_i\}\mathcal{T}, ~\text{if }\det(U)=-1. \end{aligned} \right.$$3 (Etxebarria et al., 6 Feb 2025). A sharp consequence is that all orbital tensor contributions that are odd under time reversal vanish under spin-group symmetry for collinear and coplanar structures (Etxebarria et al., 6 Feb 2025). The paper further states that the nonrelativistic anomalous Hall effect requires non-coplanarity (Etxebarria et al., 6 Feb 2025), which aligns with SSG analyses of noncoplanar antiferromagnets such as $$g=\left\{ \begin{aligned} &\{U || R_i|\bm{\tau}_i\}, ~\text{if }\det(U)=+1\ &\{-U || R_i|\bm{\tau}_i\}\mathcal{T}, ~\text{if }\det(U)=-1. \end{aligned} \right.$$4 (Xiao et al., 2023).

6. Extensions, computational frameworks, and current scope

Recent work extends SSG theory in several directions while preserving its central nonrelativistic logic. One extension concerns finite SOC via Dzyaloshinskii-Moriya interaction. The usual view that SOC destroys SSG symmetry and leaves only MSG symmetry is shown to be too restrictive: in two coplanar settings there are spin-only operations that remain exact symmetries even with considerable DMI, and in collinear magnets the magnon Hamiltonian within linear spin-wave theory retains decoupled spin and spatial rotations regardless of DMI strength (Mu et al., 1 Feb 2025). In the coplanar case the spin-only group remains

$$g=\left\{ \begin{aligned} &\{U || R_i|\bm{\tau}_i\}, ~\text{if }\det(U)=+1\ &\{-U || R_i|\bm{\tau}_i\}\mathcal{T}, ~\text{if }\det(U)=-1. \end{aligned} \right.$$5

and the preserved antiunitary spin-only symmetry forces the magnon Berry curvature to satisfy

$$g=\left\{ \begin{aligned} &\{U || R_i|\bm{\tau}_i\}, ~\text{if }\det(U)=+1\ &\{-U || R_i|\bm{\tau}_i\}\mathcal{T}, ~\text{if }\det(U)=-1. \end{aligned} \right.$$6

forbidding magnon thermal Hall response in those cases (Mu et al., 1 Feb 2025). This extends SSG applicability beyond the negligible-SOC limit in a controlled subset of systems.

A second extension is superconductivity. In the spin-space-group setting, superconducting order parameters transform under

$$g=\left\{ \begin{aligned} &\{U || R_i|\bm{\tau}_i\}, ~\text{if }\det(U)=+1\ &\{-U || R_i|\bm{\tau}_i\}\mathcal{T}, ~\text{if }\det(U)=-1. \end{aligned} \right.$$7

with the spin part rotating the triplet $$g=\left\{ \begin{aligned} &\{U || R_i|\bm{\tau}_i\}, ~\text{if }\det(U)=+1\ &\{-U || R_i|\bm{\tau}_i\}\mathcal{T}, ~\text{if }\det(U)=-1. \end{aligned} \right.$$8-vector and the spatial part acting on $$g=\left\{ \begin{aligned} &\{U || R_i|\bm{\tau}_i\}, ~\text{if }\det(U)=+1\ &\{-U || R_i|\bm{\tau}_i\}\mathcal{T}, ~\text{if }\det(U)=-1. \end{aligned} \right.$$9 independently (Feng et al., 2024). The resulting superconducting channels and basis functions can differ substantially from those obtained under magnetic point-group analysis, especially in collinear groups where pair-spin angular-momentum labels appear and nonunitary triplet states are naturally organized by SSG symmetry (Feng et al., 2024).

A third extension is the oriented-SSG program, which separates magnetic-structure generation into a nonrelativistic SSG stage and a later SOC-driven orientation stage. In this framework, magnetic structures are first generated as totally symmetric representations of an SSG, yielding “spin-symmetry-adapted” structures, and are then globally rotated so that they belong to a maximal induced MSG (Nomoto et al., 22 Jan 2026). In a benchmark over MAGNDATA, $U_i\in O(3)$00 of reported structures were reproducible at the SSG level and $U_i\in O(3)$01 of the SSG-matched set were fully reproduced within the oriented-SSG scheme; for 283 materials, the experimental magnetic structures were reproduced as energetically most stable in $U_i\in O(3)$02 of cases at the SSG level without SOC and in $U_i\in O(3)$03 of cases at the oriented-SSG level with SOC (Nomoto et al., 22 Jan 2026). The characteristic energy scale among oriented SSA structures was only $U_i\in O(3)$04 meV per magnetic atom, about 300 times smaller than that among distinct SSA structures (Nomoto et al., 22 Jan 2026). This makes explicit the symmetry hierarchy between exchange-scale magnetic geometry and SOC-scale spin-axis selection.

Low-dimensional descendants also require dedicated treatment. Spin layer groups are obtained as quotients of three-dimensional SSGs by a normal translation subgroup along the nonperiodic direction and are not, in general, a trivial restriction of the 3D theory (Zhang et al., 25 May 2026). For $U_i\in O(3)$05, their classification yields 448 collinear, 6083 coplanar, and 33556 noncoplanar spin layer groups, for a total of 40087 inequivalent SLGs (Zhang et al., 25 May 2026). This indicates that the SSG framework extends naturally to genuinely two-dimensional magnetic crystals but acquires new equivalence relations and corepresentation structures.

The present scope of SSG theory is therefore broad but not unlimited. The framework is designed for weak SOC or controlled SOC-descendant situations; when SOC is appreciable, SSG is generally no longer exact and one must descend to an MSG description, though SSG can remain an approximate organizing principle (Song et al., 2024, Etxebarria et al., 6 Feb 2025). Collinear groups require special treatment because their spin-only subgroup is continuous, and the practical representation theory often replaces continuous $U_i\in O(3)$06 by a discrete subgroup such as $U_i\in O(3)$07 to capture the relevant degeneracy structure (Song et al., 2024). A plausible implication is that SSGs now constitute a mature symmetry language for nonrelativistic magnetic structure, band theory, and response theory, while current work is refining how that language interfaces with SOC, dimensional reduction, and high-throughput materials prediction (Nomoto et al., 22 Jan 2026, Yu et al., 23 Apr 2026).