In 1987, Nair, Shapere, Strominger, and Wilczek compactified the heterotic string on tori defined by even self-dual lattices of lorentzian signature (26-D, 10-D), combined with a freely acting Z2 twist that shifts the lattice by a half-vector in the left-moving bosonic sector while acting as (-1)^F (the fermion-parity operator) on the right-moving superstring sector — the string-theoretic realization of Scherk-Schwarz supersymmetry breaking (breaking by boundary conditions that depend on the compact coordinates), whose ten-dimensional member is the tachyon-free O(16)xO(16) heterotic string. They proved modular invariance at one loop, computed the cosmological constant as an explicit function of the lattice, located its extrema (including symmetric-lattice points, argued to be exact using an all-orders R <-> 1/R duality assumption), and analyzed decompactification. They left four issues open: higher-genus modular invariance and spin-statistics of the twist sum, the nonperturbative status and selection of these vacua, the backreaction of the nonzero one-loop cosmological constant (the rolling dilaton), and the generality of R <-> 1/R duality. The duality question is resolved in its well-defined forms, and the surviving open core is a single objective: whether these twisted compactifications define consistent, dynamically selected string vacua. That core comprises (a) all-genus modular invariance and unitarity of the diagonal spin-structure/twist-structure sum; (b) the endpoint and domain of validity of the resulting rolling-dilaton evolution; and (c) nonperturbative connectedness and vacuum selection. All three remain open in the literature.

Background

The heterotic string is a hybrid of two two-dimensional worldsheet theories: its left-moving (holomorphic) excitations are those of the 26-dimensional bosonic string, while its right-moving (antiholomorphic) excitations are those of the 10-dimensional superstring. Anomaly cancellation in ten dimensions selects gauge group E8xE8 or Spin(32)/Z2. Toroidal compactification to D dimensions is described by Narain lattices: lattices that are even (every lattice vector has even integer squared length) and self-dual (the lattice coincides with its own dual, the set of vectors having integer pairing with every lattice vector), carrying a lorentzian inner product of signature (26-D, 10-D), that package the left- and right-moving momenta, with the moduli space SO(26-D, 10-D)/(SO(26-D) x SO(10-D)) up to an arithmetic quotient. One-loop consistency — modular invariance of the genus-1 (torus) amplitude under the transformation of the worldsheet complex structure — follows from the self-duality of the lattice by Poisson resummation, and is what guarantees a consistent, anomaly-free spectrum.

Nair, Shapere, Strominger, and Wilczek studied such compactifications with an additional freely acting Z2 twist: a shift of the lattice by a half-vector in the left-moving sector combined with the fermion parity operator (-1)^F in the right-moving sector. Because the two sectors are treated differently, this is an asymmetric orbifold (an orbifold whose action is not a pure geometric symmetry of the target space); because the shift acts freely (without fixed points), it evades the fixed-point singularities of ordinary orbifolds. Physically it implements the Scherk-Schwarz mechanism in string theory: supersymmetry is broken by coordinate-dependent boundary conditions on the compact dimensions, at a scale set by the compactification volume, and the resulting spectra are non-supersymmetric but can be free of tachyons (modes of negative mass-squared that signal an instability). The ten-dimensional case of this construction is the tachyon-free O(16)xO(16) heterotic string. Within this family the authors computed the one-loop cosmological constant as an explicit function of the lattice, showed that its extrema include the symmetric (self-dual) lattice points, argued — assuming an R <-> 1/R duality (the exact equivalence, in string units, between a circle compactification of radius R and one of radius 1/R, with momentum and winding modes exchanged) valid to all loop orders, then unproven — that these extrema are exact in perturbation theory, and showed that a positive ten-dimensional cosmological constant drives decompactification (the growth without bound of the compactification volume).

The construction left four issues open, and each has a technical meaning. First, higher-genus consistency: a g-loop string amplitude is an integral over the moduli space of genus-g Riemann surfaces (surfaces with g handles), and must be invariant under the mapping class group, the group of large diffeomorphisms of that surface; at genus 1 the worldsheet fermions' spin structures (sign choices for transporting fermions around worldsheet cycles) are summed in a diagonal combination tied to the GSO projection (the Gliozzi-Scherk-Olive projection that removes inconsistent states from the superstring spectrum), and the question is whether the analogous diagonal pairing of right-moving spin structures with left-moving twist structures (the parallel sector choices determining which shifted boundary conditions the left-moving lattice bosons obey around each cycle) remains modular invariant — and unitary, i.e., free of negative-norm states and consistent with the spin-statistics connection — at every genus. Second, the backreaction of the nonzero one-loop vacuum energy: by the Fischler-Susskind mechanism, a one-loop vacuum energy is a tadpole (a linear source) for the dilaton, the field that sets the string coupling; resumming these tadpoles shifts the background at every loop order, and for a positive cosmological constant drives a time-dependent rolling dilaton rather than a static vacuum. Third, the nonperturbative status: whether these non-supersymmetric compactifications are connected to each other and to supersymmetric vacua by nonperturbative physics, and which (if any) is selected as a true vacuum. Fourth, the generality of R <-> 1/R duality beyond the toroidal setting.

The fourth issue has been resolved in every well-defined form — toroidal duality is proven at arbitrary genus, with the genus-g partition function satisfying Zg(1/R)=R2g−2Zg(R)Z_g(1/R)=R^{2g-2}Z_g(R) up to a dilaton shift; duality is exact to all orders in the string expansion for sigma models (two-dimensional theories of maps into a target space) with isometries (continuous target-space rotation symmetries); mirror symmetry — the equivalence, established in major classes, between string theory on a Calabi-Yau compactification space and on its mirror, exchanging the two worldsheet chiralities — realizes it for Calabi-Yau compactifications; and the T-duality groups of the twisted compactifications themselves have been identified as congruence subgroups (subgroups cut out by arithmetic conditions on the matrix entries) of the T-duality group O(d_L, d_R; Z), the integral transformations preserving the lorentzian pairing of the compactification momenta. The first three issues remain open and constitute the problem stated below.

Problem Statement

Determine whether the twisted heterotic string compactifications of Nair, Shapere, Strominger, and Wilczek — heterotic strings on even self-dual lorentzian lattices of signature (26-D, 10-D) with the freely acting Z2 twist that combines a left-moving lattice half-shift with right-moving fermion parity — define consistent, dynamically selected string vacua. The problem has three linked components, since the general statement subsumes partial cases:

(a) Higher-loop consistency. Prove, or refute, that at every genus g >= 2 the diagonal sum over right-moving spin structures paired with left-moving twist structures is modular invariant under the genus-g mapping class group and unitary — correct spin-statistics relations, no negative-norm or wrong-statistics states in any twisted sector. A refutation at any fixed genus resolves this component.

(b) Backreaction endpoint. Determine the fate of the background under the tadpole-induced rolling of the dilaton and the internal radii: whether the evolution ends in decompactification, in a controlled time-dependent background, at a singular boundary (as in the Dudas-Mourad-type codimension-one solutions, which depend nontrivially on a single spatial coordinate), or at strong coupling where the lattice description fails; and establish the domain of validity of the perturbative treatment, including the regime in which the one-loop analysis of the source paper can be trusted.

(c) Nonperturbative status and selection. Determine whether these non-supersymmetric compactifications are connected nonperturbatively to each other and to supersymmetric vacua, and which, if any, is selected as a true vacuum of string theory — or whether the construction is realized only as a subsector of a larger background.

The source paper's fourth open issue — the generality of R <-> 1/R duality beyond the toroidal setting — is excluded from the open core because it has since been established in all its well-defined forms (all-genus toroidal duality, all-order exactness for sigma models with isometries, and the identification of the T-duality groups of these compactifications; refs 9, 10, 16). No solution method is prescribed.

References

  1. V. P. Nair, A. Shapere, A. Strominger, F. Wilczek, 'Compactification of the twisted heterotic string', Nucl. Phys. B 287 (1987) 402-418. DOI: 10.1016/0550-3213(87)90112-X. https://doi.org/10.1016/0550-3213(87)90112-X — The source paper of this task.
  2. L. Alvarez-Gaumé, P. H. Ginsparg, G. W. Moore, C. Vafa, 'An O(16) x O(16) heterotic string', Phys. Lett. B 171 (1986) 155-162. DOI: 10.1016/0370-2693(86)91524-8. https://doi.org/10.1016/0370-2693(86)91524-8 — The ten-dimensional tachyon-free non-supersymmetric heterotic string, the D=10 member of the twisted-compactification family.
  3. K. S. Narain, M. H. Sarmadi, C. Vafa, 'Asymmetric orbifolds', Nucl. Phys. B 288 (1987) 551. DOI: 10.1016/0550-3213(87)90228-8. https://doi.org/10.1016/0550-3213(87)90228-8 — The asymmetric-orbifold framework within which the twist construction is formulated.
  4. A. Parkes, 'Two loop modular invariance and spin statistics', Phys. Lett. B 184 (1987) 19. DOI: 10.1016/0370-2693(87)90481-3. https://doi.org/10.1016/0370-2693(87)90481-3 — Genus-2 modular invariance and spin-statistics constraints for closed strings with general boundary conditions.
  5. Y. Kikuchi, C. Marzban, 'Two-loop modular invariance and proper spin-statistics: projection for general boundary conditions', Phys. Rev. D 36 (1987) 2583. DOI: 10.1103/PhysRevD.36.2583. https://doi.org/10.1103/PhysRevD.36.2583 — Parallel genus-2 analysis.
  6. E. Gava, R. Iengo, G. Sotkov, 'Modular invariance and the two loop vanishing of the cosmological constant', Phys. Lett. B 207 (1988) 283. DOI: 10.1016/0370-2693(88)90576-X. https://doi.org/10.1016/0370-2693(88)90576-X — Genus-2 modular structure of heterotic amplitudes with Narain lattice sums (untwisted case).
  7. J. Lauer, R. Zucchini, 'Covariance properties of higher genus twist correlators under R <-> 1/R duality in orbifold conformal field theory', Phys. Lett. B 235 (1990) 268-274. DOI: 10.1016/0370-2693(90)91962-B. https://doi.org/10.1016/0370-2693(90)91962-B — Higher-genus behavior of twist correlators under duality.
  8. W. Fischler, L. Susskind, 'Dilaton tadpoles, string condensates and scale invariance', Phys. Lett. B 171 (1986) 383-389. DOI: 10.1016/0370-2693(86)91425-5. https://doi.org/10.1016/0370-2693(86)91425-5; and 'Dilaton tadpoles, string condensates and scale invariance II', Phys. Lett. B 173 (1986) 262-264. DOI: 10.1016/0370-2693(86)90514-9. https://doi.org/10.1016/0370-2693(86)90514-9 — The tadpole-resummation mechanism underlying the backreaction thread.
  9. A. Giveon, M. Porrati, E. Rabinovici, 'Target space duality in string theory', Phys. Rept. 244 (1994) 77-202. DOI: 10.1016/0370-1573(94)90070-1. https://doi.org/10.1016/0370-1573(94)90070-1 — Review with proofs; section 2.3: all-genus toroidal duality Zg(1/R)=R2g−2Zg(R)Z_g(1/R)=R^{2g-2}Z_g(R) with a dilaton shift, exactness for sigma models with isometries, and mirror symmetry as the Calabi-Yau realization.
  10. T. H. Buscher, 'A symmetry of the string background field equations', Phys. Lett. B 194 (1987) 59-62. DOI: 10.1016/0370-2693(87)90769-6. https://doi.org/10.1016/0370-2693(87)90769-6; and 'Path integral derivation of quantum duality in nonlinear sigma models', Phys. Lett. B 201 (1988) 466-472. DOI: 10.1016/0370-2693(88)90602-8. https://doi.org/10.1016/0370-2693(88)90602-8 — Buscher's construction: the worldsheet path-integral derivation of R <-> 1/R duality, which establishes exactness to all orders in the string expansion for sigma models with isometries.
  11. S. Kachru, E. Silverstein, 'On vanishing two loop cosmological constants in nonsupersymmetric strings', JHEP 01 (1999) 004. arXiv:hep-th/9810129. DOI: 10.1088/1126-6708/1999/01/004. https://arxiv.org/abs/hep-th/9810129 — Two-loop cosmological constants of KKS-type nonsupersymmetric models, including modular-structure subtleties.
  12. K. Aoki, E. D'Hoker, D. H. Phong, 'On the construction of asymmetric orbifold models', Nucl. Phys. B 695 (2004) 132. arXiv:hep-th/0402134. DOI: 10.1016/j.nuclphysb.2004.06.038. https://arxiv.org/abs/hep-th/0402134 — The definitive modern treatment of the asymmetric-orbifold construction problem (the framework of the source paper's twist): states that operator and Hilbert-space methods do not easily generalize beyond one loop, that full partition functions for asymmetric orbifolds remain much less well-understood, and that modular invariance cannot in general determine the chiral-block pairing matrix.
  13. K. Aoki, E. D'Hoker, D. H. Phong, 'Two loop superstrings on orbifold compactifications', Nucl. Phys. B 688 (2004) 3. arXiv:hep-th/0312181. DOI: 10.1016/j.nuclphysb.2004.04.001. https://arxiv.org/abs/hep-th/0312181 — Genus-2 chiral measures for Z2 reflection orbifolds; the two-loop cosmological constant of KKS-type models does not vanish pointwise on moduli space.
  14. E. Dudas, J. Mourad, 'Brane solutions in strings with broken supersymmetry and dilaton tadpoles', Phys. Lett. B 486 (2000) 172. arXiv:hep-th/0004165. DOI: 10.1016/S0370-2693(00)00734-6. https://arxiv.org/abs/hep-th/0004165 — Codimension-one time-dependent solutions for the ten-dimensional O(16) x O(16) string with the dilaton tadpole included; their boundaries are singular.
  15. S. Abel, R. J. Stewart, 'Exponential suppression of the cosmological constant in nonsupersymmetric string vacua at two loops and beyond', Phys. Rev. D 96 (2017) 106013. arXiv:1701.06629. DOI: 10.1103/PhysRevD.96.106013. https://arxiv.org/abs/1701.06629 — Two-loop cosmological constant for bosonic and heterotic Scherk-Schwarz compactifications with explicit Sp(4,Z) modular checks and a residual one-loop dilaton tadpole.
  16. H. Itoyama, Y. Koga, S. Nakajima, 'Target space duality of non-supersymmetric string theory', Nucl. Phys. B 975 (2022) 115667. arXiv:2110.09762. DOI: 10.1016/j.nuclphysb.2022.115667. https://arxiv.org/abs/2110.09762 — The T-duality groups of the twisted compactifications of the source paper identified as congruence subgroups of O(d_L, d_R; Z), with transitions among nonsupersymmetric models induced by T-duality.
  17. G. Leone, S. Raucci, 'Aspects of Strings without spacetime supersymmetry', Riv. Nuovo Cim. 49 (2026) 75. arXiv:2509.24703. DOI: 10.1007/s40766-025-00078-z. https://arxiv.org/abs/2509.24703 — Review; section 5.2.2: no consistent all-orders modular-invariant tadpole-subtraction scheme exists, the endpoint of the Fischler-Susskind shift is unknown, and the existence of non-supersymmetric vacua is raised as an open question; section 5.3: the uncertain fate of Dudas-Mourad-type vacua.
  18. Y. Honda, J. Kaidi, Y. Koga, Y. Liu, 'Maximal enhancements in eight-dimensional non-supersymmetric heterotic strings', arXiv:2608.19166 (2026). https://arxiv.org/abs/2608.19166 — Classification of eight-dimensional nonsupersymmetric heterotic strings at maximal gauge enhancement, with one-loop cosmological constants and Hessian stability criteria (1210 points, 71 tachyon-free).