ABSTRACT (original article): We consider Scherk-Schwarz compactifications of M-theory (toroidal compactifications with a non-trivial spin structure) in various dimensions and find isolated critical points of the potential on the moduli space. We demonstrate this by identifying the unbroken duality group and finding isolated points on the moduli spaces which are fixed by elements of the unbroken duality group. We work out concrete examples involving compactifications down to d=8, 6, 5 and 4 spacetime dimensions. We also conjecture a duality-covariant anomaly cancellation condition for M-theory on T^n orbifolded by discrete Z_N symmetries acting as phases on the charge lattice. This condition generalizes level matching for perturbative string orbifolds. CITATION (original article): Zihni Kaan Baykara, Shi Chen, and Cumrun Vafa, Isolated Critical Points for Scherk-Schwarz Compactifications of M-theory, arXiv:2605.16484. https://doi.org/10.48550/arXiv.2605.16484

Introduction

Scherk-Schwarz compactification breaks supersymmetry by imposing anti-periodic boundary conditions for fermions on a torus. At large radius the resulting potential can be studied perturbatively, but the Planckian interior of moduli space is strongly coupled. This notebook uses symmetry instead of an explicit formula for the potential.
A point is sufficient if a finite gauge symmetry fixes the background and acts without invariant directions on the linearized scalar moduli. Gauge invariance then forbids a term linear in any modulus, so the gradient of the potential vanishes there. The non-trivial spin structure is encoded by a mod-2 charge-lattice vector in the U-duality orbit of a geometric half-shift. The search therefore asks whether a candidate symmetric point fixes an admissible vector while leaving no neutral tangent direction.
This notebook presents the complete d=8 and d=7 searches and the partial d=6 search accompanying Isolated Critical Points for Scherk-Schwarz Compactifications of M-theory.

Computational strategy

Across the three dimensions, the calculation follows the same finite representation-theoretic test.
1
.
Choose an isolated symmetric point of the ordinary toroidal moduli space, represented by a symmetric lattice or Narain lattice.
2
.
Construct its finite point group and its actions on the charge lattice and on the linearized scalar moduli.
3
.
Reduce the phase-shift lattice modulo two and retain only vectors in the U-duality orbit of a geometric half-shift. The admissibility test is a rank condition in d=8, a Plucker condition in d=7 and a pure-spinor condition in d=6.
4
.
For each admissible orbit representative, compute its stabilizer and the stabilizer-invariant subspace of the moduli. A zero-dimensional fixed subspace identifies an isolated symmetric point.
5
.
When needed, construct an integral U-duality transformation that maps the generalized spin structure to a purely geometric one.

Results

◼
  • Eight dimensions. The complete scan finds exactly two isolated symmetric points in the spin moduli space: A1^3 x A1^2 and A3 x A1^2.
    • ◼
  • Seven dimensions. The complete scan of all seven lattice candidates finds none.
    • ◼
  • Six dimensions. The partial ADE and Z12-quasicrystal scans find seventeen inequivalent viable pure-spinor orbits among the tested Narain points. Possible non-cyclic quasicrystalline cases are outside the scan.
    • These symmetry arguments establish criticality but do not determine the value or sign of the potential at the critical point.

    8d: SL(3) x SL(2)

    This part performs the complete d=8 search. Candidate points are products of irreducible symmetric rank-three and rank-two lattices. The computation constructs their finite symmetry groups, acts on the mod-2 shift lattice (3,2) and on the seven linearized moduli, and tests whether a valid shift has a stabilizer with no invariant tangent direction.

    Implementation
    

    Results

    The benchmark reproduces the fully symmetric A1^3 x A1^2 point. The complete scan then finds exactly two isolated points in the spin moduli space, A1^3 x A1^2 and A3 x A1^2.

    Test: A1^3 x A1^2

    Run the orbit and stabilizer test on the simplest symmetric product lattice before starting the exhaustive scan.
    In[]:=
    testD8A13A12=AnalyzeD8Point["A1^3 x A1^2",d8PointData["A1^3 x A1^2"]];
    Analyzing A1^3 x A1^2
    SL3 point = A1^3
    SL2 point = A1^2
    |Aut3 det=1| = 24
    |Aut2 det=1| = 4
    product group size = 96
    pure phase shifts = 21
    pure-phase orbit count = 6
    fixed-moduli spectrum = 01,12,21,32
    min fixed moduli = 0
    successful representatives = 1
    possible = True

    Scan of all lattices

    Apply the same test to every pair of classified rank-three and rank-two symmetric lattice candidates.
    In[]:=
    d8Run=RunD8WithProgress[d8PointData];​​​​d8ResultsAll=d8Run["Results"];
    ==============================Starting A1^3 x A1^2 (1/6)==============================
    Analyzing A1^3 x A1^2
    SL3 point = A1^3
    SL2 point = A1^2
    |Aut3 det=1| = 24
    |Aut2 det=1| = 4
    product group size = 96
    pure phase shifts = 21
    pure-phase orbit count = 6
    fixed-moduli spectrum = 01,12,21,32
    min fixed moduli = 0
    successful representatives = 1
    possible = True
    Finished A1^3 x A1^2 in 0.06 seconds. Successful reps: 1.
    ==============================Starting A1^3 x A2 (2/6)==============================
    Analyzing A1^3 x A2
    SL3 point = A1^3
    SL2 point = A2
    |Aut3 det=1| = 24
    |Aut2 det=1| = 6
    product group size = 144
    pure phase shifts = 21
    pure-phase orbit count = 3
    fixed-moduli spectrum = 21,32
    min fixed moduli = 2
    successful representatives = 0
    possible = False
    Finished A1^3 x A2 in 0.04 seconds. Successful reps: 0.
    ==============================Starting A3 x A1^2 (3/6)==============================
    Analyzing A3 x A1^2
    SL3 point = A3
    SL2 point = A1^2
    |Aut3 det=1| = 24
    |Aut2 det=1| = 4
    product group size = 96
    pure phase shifts = 21
    pure-phase orbit count = 4
    fixed-moduli spectrum = 01,22,41
    min fixed moduli = 0
    successful representatives = 1
    possible = True
    Finished A3 x A1^2 in 0.07 seconds. Successful reps: 1.
    ==============================Starting A3 x A2 (4/6)==============================
    Analyzing A3 x A2
    SL3 point = A3
    SL2 point = A2
    |Aut3 det=1| = 24
    |Aut2 det=1| = 6
    product group size = 144
    pure phase shifts = 21
    pure-phase orbit count = 2
    fixed-moduli spectrum = 21,41
    min fixed moduli = 2
    successful representatives = 0
    possible = False
    Finished A3 x A2 in 0.04 seconds. Successful reps: 0.
    ==============================Starting A3* x A1^2 (5/6)==============================
    Analyzing A3* x A1^2
    SL3 point = A3*
    SL2 point = A1^2
    |Aut3 det=1| = 24
    |Aut2 det=1| = 4
    product group size = 96
    pure phase shifts = 21
    pure-phase orbit count = 4
    fixed-moduli spectrum = 12,32

    7d: SL(5)

    This part performs the complete d=7 search over the seven irreducible symmetric rank-five lattices. A generalized spin structure is represented by a decomposable two-form over F_2, while the fourteen moduli form the symmetric traceless representation. The scan finds no admissible stabilizer that removes every neutral modulus.

    Mathematical setup

    The following cells derive the representation-theoretic ingredients used in the implementation and state the fixed-moduli criterion explicitly.

    Moduli space and basic strategy

    In dimension five, the maximal finite irreducible positive definite lattice points are classified. The seven lattices considered in the scan are

    Lattice automorphisms

    Thus the automorphism group is

    Charge lattice and generalized spin structures

    Charge is written as a two-form
    So the phase shift lattice lives in the dual as

    Action on charges

    Equivalently, in components,

    Moduli representation

    This space has dimension

    Stabilizer criterion

    The point gives an isolated symmetric point in the spin moduli space precisely when this stabilizer fixes no scalar modulus. In other words, the only fixed scalar is the trace (which we subtract anyway)

    Results

    The A1^5 benchmark illustrates the obstruction, and the exhaustive scan over all seven lattices confirms the no-go: every candidate either lacks a valid fixed generalized spin structure or retains at least one neutral modulus.

    Test: A1^5

    Test the simplest rank-five symmetric lattice and display the surviving fixed directions or the failure of the spin-structure condition.

    Scan of all 7 lattices

    Run the same orbit, stabilizer and fixed-moduli test on the complete list of rank-five candidates.

    6d: SO(5,5)

    This part implements the partial d=6 search. Generalized spin structures are odd chiral spinors over F_2, and admissibility is the pure-spinor condition. For ADE Narain points and Z12-quasicrystalline extensions, the code enumerates pure-spinor orbits, tests their stabilizers on the twenty-five moduli and constructs dualities to geometric spin structures.

    Results

    The ADE and Z12-plus-ADE scans produce seventeen inequivalent viable pure-spinor orbits among the tested Narain points. This is a partial classification because possible non-cyclic quasicrystalline symmetry groups are not included.

    Test: A1^5

    Benchmark the full pure-spinor, stabilizer and geometrization pipeline on the most symmetric ADE Narain point.
    Here, M2 brane charges are represented by the dual indices, so 123 corresponds to M2 wrapped on 45.

    Scan of all ADE

    Apply the pipeline to all nine rank-five ADE lattice decompositions and count inequivalent viable spin structures at each point.

    Test: Q12

    Verify the lattice and symmetry data for the order-twelve quasicrystalline Narain block before forming direct sums.

    Scan of all Q12 + ADE

    Analyze the three Z12-quasicrystal plus rank-three ADE candidates and record their viable pure-spinor orbits.

    CITE THIS NOTEBOOK

    Isolated critical points for Scherk-Schwarz compactifications of M-theory​
    by Zihni Kaan Baykara, Shi Chen, and Cumrun Vafa
    Wolfram Community, STAFF PICKS, August 31, 2026
    ​https://jira.wolfram.com/jira/browse/COMMOR-2378