In their 1997 analysis of semistable principal GG-bundles over elliptically fibered Calabi-Yau manifolds, Friedman, Morgan and Witten built the heterotic side of the heterotic/F-theory duality dictionary but left its abelian-variety sector incomplete: they could not describe the weighted projective bundle WE8\mathcal{W}_{E_8} uniformly near cuspidal fibers, and they posed but did not answer the question of when and how the Jacobian of the heterotic spectral data coincides with the corresponding factor of the intermediate Jacobian J(X)=H3(X,R)/H3(X,Z)J(X)=H^3(X,\mathbb{R})/H^3(X,\mathbb{Z}) of the F-theory fourfold. Later work resolved the neighboring parts of the original question: the twisting data for the classical groups and, over a curve, for E6,E7,E8E_6,E_7,E_8 (via Donagi's cameral covers and the 1998 companion paper), the deferred global proofs of the parabolic/spectral construction, and the matter-localization criteria. The continuous part of the Jacobian identification is now a theorem, in any base dimension and for any simply-laced group (Curio-Donagi 1998), but the discrete data is identified only up to finite index, conditional on the vanishing of two Weyl-group cohomology groups, and E8E_8 cuspidal fibers are excluded. The surviving open core is the unconditional and global completion of the dictionary: the Weyl-group cohomology groups controlling the discrete identification, a global model of WE8\mathcal{W}_{E_8} valid at cusps (or a rigorous impossibility proof and replacement), non-simply-laced groups with monodromy, and the explicit fourfold-level Jacobian/flux map beyond the stable degeneration limit.

Background

Heterotic/F-theory duality is a conjectured equivalence between two string compactifications: the heterotic string on an elliptically fibered Calabi-Yau threefold ZZ (a smooth projective variety with trivial canonical bundle carrying a surjective holomorphic map π:Z→B\pi:Z\to B whose general fiber is a genus-one curve, written in Weierstrass form y2=4x3−g2x−g3y^2=4x^3-g_2x-g_3 with g2,g3g_2,g_3 holomorphic functions on the base BB), and F-theory on an elliptically fibered Calabi-Yau fourfold XX over a base built from BB. On the heterotic side the gauge sector is a holomorphic principal GG-bundle V→ZV\to Z, with GG a compact Lie group (typically a subgroup of E8E_8); a bundle is called semistable when its coherent subsheaves have slope bounded by that of VV, and the physically relevant moduli are the equivalence classes of semistable bundles. Duality predicts that these moduli are matched with geometric data of XX.

Friedman, Morgan and Witten (1997) constructed the bundles fiber by fiber. On a single elliptic curve, a theorem of Looijenga identifies the moduli space of semistable GG-bundles with fixed fiberwise SS-equivalence class (fixed restriction to the general fiber, up to Jordan-Holder factors) with a weighted projective space, realized as the base of the miniversal unfolding (the universal deformation family) of the simple elliptic singularity G~\widetilde{G} attached to the affine Dynkin diagram of GG. Applied over the fibration, this yields a weighted projective bundle WG→B\mathcal{W}_G\to B together with spectral data. For G=SU(n)G=SU(n) the data is explicit: a spectral cover C⊂ZC\subset Z, an nn-sheeted cover of BB tracing the eigenvalues of a fiberwise Higgs field (a field valued in bundle endomorphisms that linearizes the construction), plus a twisting line bundle on CC controlling c2(V)c_2(V). For general GG, Donagi (1997) recast the construction in terms of cameral covers: W(G)W(G)-Galois covers of BB, where W(G)W(G) is the Weyl group of GG, carrying a distinguished Prym variety — an abelian variety obtained from the Jacobian J(C)=Pic0(C)J(C)=\mathrm{Pic}^0(C) of degree-zero line bundles on a cover by imposing W(G)W(G)-symmetric relations, known as a Prym-Tyurin variety.

The abelian-variety sector is the crux of what remains open. On the F-theory side the relevant objects are the intermediate Jacobian J(X)=H3(X,R)/H3(X,Z)J(X)=H^3(X,\mathbb{R})/H^3(X,\mathbb{Z}), the complex torus of the fourfold on which the M-theory three-form field takes its periods, and the quantized four-form flux G4∈H2,2(X,Z)G_4\in H^{2,2}(X,\mathbb{Z}), subject to the shifted quantization condition of Witten (1997) and, for supersymmetry, the primitivity condition of Becker-Becker (1996). In the stable degeneration limit, where XX degenerates to a union W1∪ZW2W_1\cup_Z W_2 of two fibrations by del Pezzo surfaces (a del Pezzo surface is the projective plane blown up in at most eight points) glued along ZZ, Kanev (1989) had shown for pencils of del Pezzo surfaces over P1\mathbb{P}^1 that the intermediate Jacobian of the total space is exactly a Prym-Tyurin variety of the family — the structure the duality dictionary predicts. Friedman-Morgan-Witten then asked whether, and how, the Jacobian/Prym factors of the heterotic spectral data coincide with the appropriate factor of J(X)J(X), and recorded two obstructions: for G=E8G=E_8 the unfolding of the E~8\widetilde{E}_8 singularity admits no parametrization uniform in g2,g3g_2,g_3, so over bases of dimension at least two — where cuspidal fibers with g2=g3=0g_2=g_3=0 occur — their description of WE8\mathcal{W}_{E_8} is only local; and on the F-theory side the appropriate abelian subvariety of J(X)J(X) had not been identified.

The neighboring parts of the original question have since been settled. Cameral covers, together with the second Friedman-Morgan-Witten paper (1998), supply the twisting freedom for the classical groups and, over a one-dimensional base, for E6,E7,E8E_6,E_7,E_8 — the analog of twisting by a line bundle on the spectral cover — including an exact sequence expressing the twisting group through the relative Deligne cohomology (the cohomology theory whose cycles carry Abel-Jacobi equivalence, refining rational equivalence by real periods) of the cameral cover. The proofs deferred in the 1997 paper (its reference '[36]') were supplied by the 1998 companion, by 'Vector Bundles over Elliptic Fibrations' (1997), by Friedman-Morgan (2000) for the parabolic construction, and completed from the Fourier-Mukai side (a transform exchanging sheaves on ZZ and BB through the fibration) by Hernandez Ruiperez-Munoz Porras (2002). The matter-localization criterion — which cohomology groups of bundles in associated representations localize on which loci — was proved for arbitrary GG in the 1997 paper itself and refined by Diaconescu-Ionesei (1998) and by Hayashi-Tatar-Toda-Watari-Yamazaki (2009). For the Jacobian identification, Curio-Donagi (1998) proved, for any simply-laced group over a base of any dimension, that a connected component of the distinguished Prym is canonically isomorphic to the relative intermediate Jacobian of the associated family, which in the stable degeneration limit is the F-theory intermediate Jacobian: the continuous part of the dictionary. Two gaps remain in their theorem: the discrete data (components, torsion, component groups) is identified only up to finite index, conditional on the vanishing of two Weyl-group cohomology groups they did not compute, with the adjoint-type assumption (trivial center) introduced in their analysis only to simplify the discrete-part spectral sequence; and cuspidal fibers are excluded for G=E8G=E_8.

Problem Statement

Complete the heterotic/F-theory dictionary for bundle moduli as unconditional, global mathematics. The single scientific objective is to match the full — continuous and discrete — abelian-variety and flux data of the heterotic spectral covers with the intermediate-Jacobian and quantized-flux data of the F-theory compactification. Four interlocking results are requested, at the generality with which the sources pose them:

  1. Unconditional discrete identification. Compute the Weyl-group cohomology groups H2(W(G),T)H^2(W(G),T) and H2(W(G),Hom(Λ,Z[W(G)]))H^2(W(G),\mathrm{Hom}(\Lambda,\mathbb{Z}[W(G)])) for G=E6,E7,E8G=E_6,E_7,E_8 — where Λ\Lambda is the root lattice, TT the maximal-torus character module, and Z[W(G)]\mathbb{Z}[W(G)] the integral group ring — or rigorously account for their nonzero values, so as to remove the finite-index caveat in the Curio-Donagi identification; and settle the Friedman-Morgan-Witten Section 4.4 conjecture that components of the E8E_8 bundle moduli with fixed fiberwise restriction correspond bijectively to twists of the intermediate Jacobian J(W1)=H3(W1,R)/H3(W1,Z)J(W_1)=H^3(W_1,\mathbb{R})/H^3(W_1,\mathbb{Z}) of the del Pezzo fibration W1W_1 by primitive classes in H2,2(W1)H^{2,2}(W_1), i.e. to the quantized G4G_4-flux values in the stable degeneration X=W1∪ZW2X=W_1\cup_Z W_2.

  2. The E8E_8 cuspidal regime. Construct a global, uniform description of the weighted projective bundle WE8→B\mathcal{W}_{E_8}\to B, valid over bases BB of dimension at least two including the cuspidal fibers (g2=g3=0g_2=g_3=0) at which all known local parametrizations of the E~8\widetilde{E}_8 unfolding break down — or prove rigorously that no global uniform description exists in any form, and give the correct global replacement geometry (Friedman-Morgan-Witten stated the narrower belief that WE8\mathcal{W}_{E_8} is not the projectivization of a split sum of line bundles); in either case, extend the Prym/intermediate-Jacobian identification of the dictionary to this regime.

  3. Non-simply-laced groups. Extend the correspondence to F4F_4, G2G_2 and twisted forms, where F-theory monodromy acts on the elliptic fiber by outer automorphisms of the Dynkin diagram and the cameral-cover formalism requires a monodromy-equivariant version not yet available.

  4. Beyond the stable degeneration. Replace the stable-degeneration comparison by an explicit map on the smooth F-theory fourfold: from the spectral-cover Jacobians to the fourfold's intermediate-Jacobian and quantized four-form-flux data (with the quantization shifts and primitivity conditions above), valid also for degenerate and reducible spectral covers, i.e. including T-brane data (Higgs fields that cannot be diagonalized, so the spectral cover alone does not determine the bundle).

The four facets are jointly one objective: progress on any facet is meaningful, but the problem is solved only when the dictionary holds unconditionally and globally, with each facet at the generality above — no facet may be settled by restricting to generic covers, smooth bases, or the stable degeneration alone.

References

  1. Robert Friedman, John Morgan, Edward Witten, 'Vector Bundles And F Theory', Communications in Mathematical Physics 187 (1997) 679-743. DOI: 10.1007/s002200050154, arXiv:hep-th/9701162, https://arxiv.org/abs/hep-th/9701162
  2. Robert Friedman, John W. Morgan, Edward Witten, 'Principal G-bundles over elliptic curves', Mathematical Research Letters 5 (1998) 97-118. DOI: 10.4310/mrl.1998.v5.n1.a8, arXiv:alg-geom/9707004, https://arxiv.org/abs/alg-geom/9707004
  3. Robert Friedman, John W. Morgan, Edward Witten, 'Vector Bundles over Elliptic Fibrations', arXiv:alg-geom/9709029 (1997), https://arxiv.org/abs/alg-geom/9709029
  4. Ron Y. Donagi, 'Principal bundles on elliptic fibrations', Asian Journal of Mathematics 1 (1997) 214-223. arXiv:alg-geom/9702002, https://arxiv.org/abs/alg-geom/9702002
  5. Gottfried Curio, Ron Y. Donagi, 'Moduli in N=1 heterotic/F-theory duality', Nuclear Physics B 518 (1998) 603-631. DOI: 10.1016/S0550-3213(98)00185-0, arXiv:hep-th/9801057, https://arxiv.org/abs/hep-th/9801057
  6. Bjorn Andreas, Gottfried Curio, 'On discrete Twist and Four-Flux in N=1 heterotic/F-theory compactifications', Advances in Theoretical and Mathematical Physics 3 (1999) 1325-1413. DOI: 10.4310/ATMP.1999.v3.n5.a4, arXiv:hep-th/9908193, https://arxiv.org/abs/hep-th/9908193
  7. Vassil Kanev, 'Intermediate Jacobians and Chow groups of threefolds with a pencil of del Pezzo surfaces', Annali di Matematica Pura ed Applicata 154 (1989) 13-48. DOI: 10.1007/BF01790341, https://doi.org/10.1007/BF01790341
  8. D. Hernández Ruipérez, J. M. Muñoz Porras, 'Stable sheaves on elliptic fibrations', Journal of Geometry and Physics 43 (2002) 163-183. DOI: 10.1016/S0393-0440(02)00019-0, arXiv:math/9809019, https://arxiv.org/abs/math/9809019
  9. Duiliu-Emanuel Diaconescu, Gheorghe Ionesei, 'Spectral Covers, Charged Matter and Bundle Cohomology', Journal of High Energy Physics 9812 (1998) 001. DOI: 10.1088/1126-6708/1998/12/001, arXiv:hep-th/9811129, https://arxiv.org/abs/hep-th/9811129
  10. Hirotaka Hayashi, Radu Tatar, Yukinobu Toda, Taizan Watari, Masahito Yamazaki, 'New Aspects of Heterotic-F Theory Duality', Nuclear Physics B 806 (2009) 224-299. DOI: 10.1016/j.nuclphysb.2008.07.031, arXiv:0805.1057, https://arxiv.org/abs/0805.1057
  11. Robert Friedman, John Morgan, 'Holomorphic Principal Bundles Over Elliptic Curves II: The Parabolic Construction', Journal of Differential Geometry 56 (2000) 301-379. DOI: 10.4310/jdg/1090347646, arXiv:math/0006174, https://arxiv.org/abs/math/0006174
  12. Lara B. Anderson, He Feng, Xin Gao, Mohsen Karkheiran, 'Heterotic/Heterotic and Heterotic/F-theory Duality', Physical Review D 100 (2019) 126014. DOI: 10.1103/PhysRevD.100.126014, arXiv:1907.04395, https://arxiv.org/abs/1907.04395
  13. Mirjam Cvetič, Antonella Grassi, Denis Klevers, Maximilian Poretschkin, Peng Song, 'Origin of Abelian Gauge Symmetries in Heterotic/F-theory Duality', Journal of High Energy Physics 2016 (2016) 041. DOI: 10.1007/JHEP04(2016)041, arXiv:1511.08208, https://arxiv.org/abs/1511.08208
  14. Lara B. Anderson, Xin Gao, Mohsen Karkheiran, 'Extending the Geometry of Heterotic Spectral Cover Constructions', Nuclear Physics B 956 (2020) 115003. DOI: 10.1016/j.nuclphysb.2020.115003, arXiv:1912.00971, https://arxiv.org/abs/1912.00971
  15. Patrick Jefferson, Manki Kim, 'On the intermediate Jacobian of M5-branes', Journal of High Energy Physics 2024 (2024) 180. DOI: 10.1007/JHEP05(2024)180, arXiv:2211.00210, https://arxiv.org/abs/2211.00210
  16. Edward Witten, 'On Flux Quantization in M-Theory and the Effective Action', Journal of Geometry and Physics 22 (1997) 1-13. DOI: 10.1016/S0393-0440(96)00042-3, arXiv:hep-th/9609122, https://arxiv.org/abs/hep-th/9609122
  17. Katrin Becker, Melanie Becker, 'M-Theory on Eight-Manifolds', Nuclear Physics B 477 (1996) 155-167. DOI: 10.1016/0550-3213(96)00367-7, arXiv:hep-th/9605053, https://arxiv.org/abs/hep-th/9605053