Non-perturbative foundations of the string duality web: exactness of type IIB $\mathrm{SL}(2,\mathbb{Z})$ self-duality, $U$-duality groups $G(\mathbb{Z})$, and the heterotic dualities

Witten (1995) organized the strong-coupling dynamics of the five ten-dimensional superstring theories into a web of conjectured non-perturbative dualities among the strings and with eleven-dimensional M-theory, while explicitly acknowledging that the load-bearing inputs remained unproven. The open problem: prove, or refute, within a controlled non-perturbative definition of the theories involved, the exactness of (a) the type IIB $\mathrm{SL}(2,\mathbb{Z})$ self-duality, (b) the $U$-duality groups $G(\mathbb{Z})$ of toroidal compactifications, (c) the six-dimensional heterotic-type IIA duality (the heterotic string on a four-torus versus type IIA on a K3 surface), and (d) the ten-dimensional $\mathrm{SO}(32)$ heterotic-type I duality, including at the degenerate limits of the K3 moduli spaces and in regimes admitting no parametrically valid description.

Background

Perturbative string theory provides five distinct, apparently unrelated consistent supersymmetric string theories in ten spacetime dimensions: type I (open, unoriented superstrings with $\mathrm{SO}(32)$ gauge symmetry), type IIA and type IIB (closed, oriented superstrings differing in the chirality of their supercharges), and two heterotic theories (closed strings whose left-moving modes are bosonic and right-moving modes supersymmetric, with gauge groups $\mathrm{SO}(32)$ and $E_8\times E_8$). Here supersymmetry is the quantum symmetry pairing bosons and fermions, and supergravity - the low-energy effective field theory of each string, containing gravity coupled to supersymmetric matter - controls the dynamics at energies far below the string scale. Each theory is characterized by a dimensionless string coupling $g_s$, in which its perturbative expansion is organized as a genus expansion (an expansion in the number of handles of the string world-sheet, the two-dimensional surface that a propagating string sweeps out); that expansion is reliable only for $g_s\ll 1$, and the strong-coupling regime of all five theories was, before 1995, essentially inaccessible.

String theory does possess one exactly established equivalence: T-duality, provable in perturbation theory, under which a string compactified on a circle of radius $R$ is equivalent to the T-dual theory on a circle of radius $\alpha'/R$, where $\alpha'$ is the inverse string tension. S-duality is the far stronger conjecture that a string theory at coupling $g_s$ is exactly equivalent to a (possibly different) string theory at coupling $1/g_s$.

Witten (1995), building on the $U$-duality conjecture of Hull and Townsend (1995), argued that the strong-coupling dynamics of all five strings in dimensions $d\ge 4$ is governed by a web of conjectured non-perturbative dualities. Its main elements are: (1) the strong-coupling limit of type IIA is eleven-dimensional supergravity, pointing to a single eleven-dimensional parent theory, soon called M-theory, whose existence is inferred but which has no independent definition; (2) ten-dimensional type IIB has an exact $\mathrm{SL}(2,\mathbb{Z})$ self-duality acting on the complexified coupling $\tau$ that combines the Ramond-Ramond scalar (a massless field from the RR sector of the closed-string spectrum) with the dilaton (the massless scalar whose expectation value sets the string coupling), under which the fundamental string is exchanged with the D-string (a solitonic one-dimensional extended object carrying Ramond-Ramond charge); (3) the $\mathrm{SO}(32)$ heterotic string is equivalent to type I; (4) the heterotic string toroidally compactified on a four-torus $T^4$ (the product of four circles, which preserves all sixteen of its supercharges) is equivalent, in six dimensions, to type IIA compactified on a K3 surface - the essentially unique compact four-dimensional Calabi-Yau manifold, that is, a complex manifold of vanishing Ricci curvature, which preserves half of the type II supercharges; and (5) upon toroidal compactification to $d$ dimensions, the continuous non-compact symmetry group $G$ of the corresponding maximal supergravity (for the maximally supersymmetric cases, split real forms - the maximally non-compact real versions of the exceptional Lie groups - such as $E_{6(6)}(\mathbb{R})$ in five dimensions) is conjectured to survive quantization as a discrete integral form $G(\mathbb{Z})$, the $U$-duality group, combining T- and S-dualities and acting on the full quantum spectrum.

Because a duality relates a weakly coupled theory to a strongly coupled one, no fixed order of perturbation theory can establish it. The 1995 argument was therefore necessarily indirect: identify the correct low-energy supergravity including its solitonic states, match the spectra of BPS states - states saturating a Bogomol'nyi-Prasad-Sommerfield bound between mass and charge, annihilated by part of the supersymmetry algebra, whose masses are fixed exactly by their charges and which therefore persist at strong coupling - and interpolate between parametrically controlled regimes, for instance where a compactification radius is large so that semiclassical eleven-dimensional supergravity is valid.

The paper explicitly acknowledges that its conclusions rest on unproven inputs: the $\mathrm{SL}(2,\mathbb{Z})$ exactness of type IIB, the integrality and action of the $U$-duality groups $G(\mathbb{Z})$, and the six-dimensional heterotic-type IIA (string-string) duality are stated as conjectures on which the arguments depend; the identification of the $\mathrm{SO}(32)$ heterotic string with type I is called an obvious speculation supported by scarcely strong evidence, with the specific puzzle that the heterotic spinor representations of $\mathrm{SO}(32)$ are absent from the perturbative type I spectrum and would have to arise as solitons; the RR-charged BPS states required by duality are assumed to exist without a microscopic construction, with the collective-coordinate quantization of the relevant instanton-solitons (localized, topologically non-trivial classical solutions) and the fate of instantons shrinking to zero size left open; the analysis of degenerations of the moduli space (the space of exactly marginal compactification parameters) was restricted to $d\ge 4$ and to large-volume classical geometry; and several steps rely on scaling arguments valid only when a compactification radius is parametrically large.

Two of these acknowledged gaps were closed within three years, at the standard physics level of evidence. Polchinski (1995) showed that D-branes - dynamical extended objects defined by mixed Dirichlet-Neumann boundary conditions on open strings - carry exactly a complete set of Ramond-Ramond charges with the charge quantum required by duality, supplying the microscopic origin of the RR-charged BPS states; Witten (1996) showed that instantons shrinking to zero size on the heterotic side are not singular but support an enhanced gauge symmetry $\mathrm{SU}(2)$, or $\mathrm{Sp}(k)$ for $k$ coincident instantons, in their core; and Sen (1998) constructed the missing $\mathrm{SO}(32)$ spinor states of type I as kinks of the tachyon field (a field of negative mass squared on an unstable brane-antibrane system) on a D-string-anti-D-string pair. The dualities themselves, however, remain unproven: every later framework that could host a proof - the BFSS matrix model of M-theory as a limit of the quantum mechanics of D0-branes (point-like D-branes), matrix string theory (a two-dimensional gauge-theory formulation of type IIA on a circle), the type IIB matrix model (a zero-dimensional reduced matrix model proposed as a constructive definition of type IIB), AdS/CFT holography (the conjectured equivalence between gravity theories on anti-de Sitter spaces, which are spacetimes of constant negative curvature, and conformal field theories, which are quantum field theories with symmetry under local angle-preserving transformations, on their boundary), and string field theory (a second-quantized formulation in which fields create whole strings rather than particle-like states) - is either itself defined by a conjectured equivalence or covers only restricted backgrounds and sectors, and standard reviews continue to present $U$-duality as a conjecture. The open problem is therefore the non-perturbative foundation of the duality web itself.

Problem Statement

The problem is to give the non-perturbative string duality web its foundation: to prove, or refute, from first principles - within an explicitly constructed and controlled non-perturbative definition of the theories involved - that

(a) ten-dimensional type IIB superstring theory possesses an exact $\mathrm{SL}(2,\mathbb{Z})$ self-duality acting on the complexified coupling $\tau$, under which the full quantum spectrum and observables are mapped onto themselves, with the fundamental string exchanged with the D-string;

(b) for every toroidal compactification to $d$ dimensions (the integrality conjecture is stated for all $d$, without the $d\ge 4$ restriction of the 1995 analysis), the continuous symmetry group $G$ of the maximal supergravity admits an integral form $G(\mathbb{Z})$ that is an exact symmetry of the quantum theory and acts on the spectrum as the $U$-duality conjecture predicts;

(c) the heterotic string toroidally compactified on a four-torus $T^4$ is exactly equivalent, in six dimensions, to type IIA compactified on a K3 surface; and

(d) the $\mathrm{SO}(32)$ heterotic string is exactly equivalent, in ten dimensions, to the type I superstring.

In every leg the claimed equivalence or symmetry is a statement about the full quantum theory - the complete spectrum and all observables - not only about the BPS-protected subsector or the low-energy supergravity limit. Because each of (a)-(d) relates a strongly coupled theory to a weakly coupled one, no fixed order of perturbation theory can settle any of them. Settling them therefore requires a controlled non-perturbative framework. These four legs are one scientific objective - the exactness of the single duality web organizing the strong-coupling dynamics of all five strings - not four independent questions. Because they form one objective, a refutation of any one of (a)-(d) shows that the web is not exact and by itself resolves the problem, whereas a proof must establish all four legs, or establish some of them together with an explicit demonstration that the remaining legs follow.

The same foundation is expected to settle two subsidiary questions acknowledged in the source that remain without microscopic control: the complete behavior at the degenerate limits of the K3 moduli spaces appearing in (c), beyond the large-volume classical-geometry regime available in 1995; and the dynamics in regimes where competing scales are comparable, so that neither perturbative string theory nor semiclassical eleven-dimensional supergravity is parametrically valid. The exactness claims of (a)-(d) bear on both regimes.

Scope note: two sub-questions acknowledged in the source - the microscopic origin of the RR-charged BPS states and the mechanism by which type I reproduces the heterotic $\mathrm{SO}(32)$ spinor states - were resolved at the standard physics level of evidence by later work (Polchinski 1995; Witten 1996; Sen 1998) and are therefore not part of the open core.

References

1. Edward Witten, String Theory Dynamics In Various Dimensions, Nucl. Phys. B 443 (1995) 85-126. DOI: 10.1016/0550-3213(95)00158-O. arXiv:hep-th/9503124, https://arxiv.org/abs/hep-th/9503124
2. C. M. Hull and P. K. Townsend, Unity of Superstring Dualities, Nucl. Phys. B 438 (1995) 109-137. DOI: 10.1016/0550-3213(94)00559-W. arXiv:hep-th/9410167, https://arxiv.org/abs/hep-th/9410167
3. Joseph Polchinski, Dirichlet-Branes and Ramond-Ramond Charges, Phys. Rev. Lett. 75 (1995) 4724-4727. DOI: 10.1103/PhysRevLett.75.4724. arXiv:hep-th/9510017, https://arxiv.org/abs/hep-th/9510017
4. Joseph Polchinski and Edward Witten, Evidence for Heterotic - Type I String Duality, Nucl. Phys. B 460 (1996) 525-540. DOI: 10.1016/0550-3213(95)00614-1. arXiv:hep-th/9510169, https://arxiv.org/abs/hep-th/9510169
5. Edward Witten, Small Instantons in String Theory, Nucl. Phys. B 460 (1996) 541-559. DOI: 10.1016/0550-3213(95)00625-7. arXiv:hep-th/9511030, https://arxiv.org/abs/hep-th/9511030
6. Edward Witten, Five-branes And M-Theory On An Orbifold, Nucl. Phys. B 463 (1996) 383-397. DOI: 10.1016/0550-3213(96)00032-6. arXiv:hep-th/9512219, https://arxiv.org/abs/hep-th/9512219
7. John H. Schwarz, The Power of M Theory, Phys. Lett. B 367 (1996) 97-103. DOI: 10.1016/0370-2693(95)01429-2. arXiv:hep-th/9510086, https://arxiv.org/abs/hep-th/9510086
8. Ashoke Sen, Unification of String Dualities, Nucl. Phys. Proc. Suppl. 58 (1997) 5-19. DOI: 10.1016/S0920-5632(97)00409-X. arXiv:hep-th/9609176, https://arxiv.org/abs/hep-th/9609176
9. Cumrun Vafa, Evidence for F-Theory, Nucl. Phys. B 469 (1996) 403-418. DOI: 10.1016/0550-3213(96)00172-1. arXiv:hep-th/9602022, https://arxiv.org/abs/hep-th/9602022
10. T. Banks, W. Fischler, S. H. Shenker and L. Susskind, M Theory As A Matrix Model: A Conjecture, Phys. Rev. D 55 (1997) 5112-5128. DOI: 10.1103/PhysRevD.55.5112. arXiv:hep-th/9610043, https://arxiv.org/abs/hep-th/9610043
11. R. Dijkgraaf, E. Verlinde and H. Verlinde, Matrix String Theory, Nucl. Phys. B 500 (1997) 43-61. DOI: 10.1016/S0550-3213(97)00326-X. arXiv:hep-th/9703030, https://arxiv.org/abs/hep-th/9703030
12. N. Ishibashi, H. Kawai, Y. Kitazawa and A. Tsuchiya, A Large-N Reduced Model as Superstring, Nucl. Phys. B 498 (1997) 467-491. DOI: 10.1016/S0550-3213(97)00290-3. arXiv:hep-th/9612115, https://arxiv.org/abs/hep-th/9612115
13. Juan M. Maldacena, The Large N Limit of Superconformal Field Theories and Supergravity, Adv. Theor. Math. Phys. 2 (1998) 231-252. DOI: 10.4310/ATMP.1998.v2.n2.a1. arXiv:hep-th/9711200, https://arxiv.org/abs/hep-th/9711200, https://doi.org/10.4310/ATMP.1998.v2.n2.a1
14. Ashoke Sen, SO(32) Spinors of Type I and Other Solitons on Brane-Antibrane Pair, JHEP 09 (1998) 023. DOI: 10.1088/1126-6708/1998/09/023. arXiv:hep-th/9808141, https://arxiv.org/abs/hep-th/9808141
15. N. A. Obers and B. Pioline, U-duality and M-Theory, Phys. Rept. 318 (1999) 113-225. DOI: 10.1016/S0370-1573(99)00004-6. arXiv:hep-th/9809039, https://arxiv.org/abs/hep-th/9809039