A truly decoupled pure four-dimensional $\mathcal{N}=1$ super Yang-Mills limit from M-theory on $G_2$-holonomy flop geometries

In 2001, Atiyah, Maldacena and Vafa proposed that the flop transition of a seven-dimensional manifold of $G_2$ holonomy is a large-$N$ duality of M-theory: on suitable $\mathbb{Z}_N$ quotients, the flop exchanges a phase containing $N$ D6-branes wrapped on a three-sphere for a phase with $N$ units of Ramond-Ramond flux, realizing in M-theory a duality of four-dimensional $\mathcal{N}=1$ super Yang-Mills proposed by Vafa. Their paper left four issues open: existence of deformed $G_2$ metrics with an asymptotically finite M-theory circle; removal of the $V=0$ singularity of the moduli space by membrane instantons; an eleven-dimensional supergravity solution with non-divergent effective string coupling; and the decoupling of a pure $\mathcal{N}=1$ super Yang-Mills field theory from the bulk. The first three issues have since been resolved in the literature: the required asymptotically locally conical $G_2$ metrics (the $B_7$ and $D_7$ families) were constructed by Brandhuber, Gomis, Gubser and Gukov and by Cvetič, Gibbons, Lü and Pope, and their existence and classification were proved by Bogoyavlenskaya and by Foscolo, Haskins and Nordström; and Atiyah and Witten argued that membrane instantons render the quantum moduli space a smooth curve. The fourth issue remains open: no limit of these constructions, nor of any known string or M-theory construction, yields pure $SU(N)$ $\mathcal{N}=1$ super Yang-Mills decoupled from gravity with parametric control, and the obstruction identified by the original authors — the chiral modulus $V_M$ and the coupled bulk — has not been overcome. The problem is to construct such a controlled decoupled limit, or to prove its impossibility.

Background

The holonomy group of a Riemannian manifold is the subgroup of rotations generated by parallel transport around closed loops; a seven-dimensional manifold whose holonomy is the exceptional Lie group $G_2$ admits covariantly constant spinors, and M-theory — the eleven-dimensional strong-coupling completion of type IIA string theory — compactified on such a manifold gives a four-dimensional theory with $\mathcal{N}=1$ supersymmetry, that is, one supersymmetry generator of each chirality. The basic complete example is the total space of the spinor bundle over $S^3$, namely $S(S^3)=\mathbb{R}^4\times S^3$ with the metric constructed by Bryant and Salamon: it is asymptotically conical (AC), meaning that at large distance it approaches a cone over $S^3\times S^3$, and its isometry group is essentially $SU(2)^3$ together with an interchange of factors. In M-theory on this space, the volume $V$ of the contractible $S^3$ and the period $C$ of the M-theory three-form gauge field (the $C$-field) over it combine into a complex chiral modulus $V_M=V+iC$, and the flop transition — in algebraic geometry, a birational operation that replaces one contracted submanifold by a partner of the same dimension — here sends $V_M\to -V_M$, exchanging the two three-spheres. Dividing by a $\mathbb{Z}_N$ subgroup of one $SU(2)$ isometry produces an $A_{N-1}$ singularity (a quotient singularity locally modeled on $\mathbb{C}^2/\mathbb{Z}_N$) transverse to a three-sphere in the quotient, supporting an $SU(N)$ gauge sector whose low-energy limit is four-dimensional $\mathcal{N}=1$ $SU(N)$ super Yang-Mills. Reducing along an appropriate circle fiber at weak coupling, one side of the flop is type IIA string theory on the deformed conifold (the smoothing of the conical threefold singularity in which a three-sphere retains finite size) with $N$ D6-branes (solitonic extended objects of type IIA string theory) wrapped on that three-sphere, and the other is type IIA on the resolved conifold (in which a two-sphere is blown up) with $N$ units of Ramond-Ramond two-form flux (flux of a two-form gauge field of type IIA string theory) through the two-sphere; Vafa conjectured the exact large-$N$ duality between these two pictures, with the complexified Kähler parameter $t$ of the two-sphere related to the three-sphere volume $V_A$ in the type IIA frame and to the string coupling $g_s$ by $(e^t-1)^N = a\,e^{-V_A/g_s}$, with $a$ a constant. Atiyah-Maldacena-Vafa rederived this duality purely geometrically in M-theory, where both sides lift to $G_2$-holonomy geometry and the duality becomes the flop.

The derivation was, however, conditional on several expectations. First, the known metric is asymptotically conical, so the M-theory circle decompactifies at infinity; the authors expected a deformation breaking one $SU(2)$ isometry to $U(1)$ so that the circle has finite asymptotic radius — a metric that is asymptotically locally conical (ALC), the seven-dimensional analogue of the Taub-NUT gravitational instanton — with such deformations existing for both signs of $V_M$ (issue i). Second, they expected but could not rigorously argue that Euclidean M2-brane instantons (membranes wrapped on supersymmetric three-cycles, which correct the $C$-field period relations nonperturbatively) remove the would-be singularity of the moduli space at $V=0$ (issue ii). Third, an eleven-dimensional supergravity solution interpolating to finite asymptotic string coupling was still to be constructed (issue iii). Fourth, they stated that they did not expect any limit of the construction to give a geometry truly decoupled from the bulk, since the four-dimensional theory carries the modulus $V_M$, which is not a parameter of pure $\mathcal{N}=1$ super Yang-Mills (issue iv).

The subsequent literature settled the first three issues. Brandhuber, Gomis, Gubser and Gukov constructed the one-parameter ALC family now called $B_7$, describing the M-theory lift of $N$ D6-branes on the deformed conifold at any value of the coupling, with the M-theory circle of finite radius at infinity and the type IIA dilaton constant at infinity; Cvetič, Gibbons, Lü and Pope constructed the second branch $D_7$ and unified both families in one system of first-order equations. Bogoyavlenskaya and later Foscolo, Haskins and Nordström proved existence of these families, classified all complete simply connected $G_2$ manifolds with this symmetry, and constructed the conically singular ALC space through which the $B_7$-$D_7$ transition passes — the metric version of the $G_2$ flop. Atiyah and Witten argued that membrane instantons make the quantum moduli space of the cone over $S^3\times S^3$ a smooth genus-zero curve interpolating between the three classical branches, and that for the $\mathbb{Z}_N$ quotients this picture reproduces gluino condensation (a nonzero expectation value of the bilinear of the gluino, the supersymmetric partner of the gluon) and the breaking of the discrete chiral symmetry that yields the $N$ vacua of four-dimensional $SU(N)$ super Yang-Mills; Brandhuber, Gomis, Gubser and Gukov computed the instanton superpotential including multiple covers.

The fourth issue has resisted resolution. Pure four-dimensional $\mathcal{N}=1$ super Yang-Mills has only two parameters, the coupling $g_{YM}$ and the theta angle; the theory obtained from the $G_2$ construction always has the additional chiral modulus $V_M$ and remains coupled to the bulk of M-theory. Acharya proposed to realize pure $\mathcal{N}=1$ super Yang-Mills as the low-energy limit of M-theory near an ADE singularity (a complex quotient singularity $\mathbb{C}^2/\Gamma$ with $\Gamma$ a finite subgroup of $SU(2)$, of which $A_{N-1}$ is the case $\Gamma=\mathbb{Z}_N$) along a three-cycle in a $G_2$ manifold, with the flop branches as candidate duals, while conceding that gravity had been 'essentially ignored' in the argument; the review by Acharya and Gukov concluded that the evidence 'suggests strongly' that the low-energy limit near such singularities is super Yang-Mills theory — an infrared statement without parametric control. Maldacena and Núñez constructed a wrapped-brane supergravity dual 'towards' the large-$N$ limit of pure $\mathcal{N}=1$ super Yang-Mills, capturing confinement and chiral symmetry breaking qualitatively, but the Kaluza-Klein modes (modes of the fields on the compact internal directions) of the bulk do not decouple in the supergravity approximation. Girardello, Petrini, Porrati and Zaffaroni obtained a supergravity dual from a relevant deformation of the maximally supersymmetric $\mathcal{N}=4$ gauge theory, with a singular infrared endpoint. In the most recent refinement, Acharya, Foscolo, Najjar and Svanes generalized the flop using the modern existence and classification theorems, and the gauge sectors of their $G_2$ conifolds 'reduce to pure $\mathcal{N}=1$ super Yang-Mills theory at low enough energies' — the same uncontrolled infrared statement. Here 'parametric control' means a regime, specified by explicit limits of $N$, $V_M$ and the Planck scale, in which the corrections to the four-dimensional gauge description are quantifiably small, as in the large-$N$ limit of the AdS/CFT correspondence (the duality between conformal gauge theory and gravity on anti-de Sitter space); no such regime is known for pure $\mathcal{N}=1$ super Yang-Mills.

Problem Statement

The surviving open problem from Atiyah-Maldacena-Vafa is issue (iv). In the framework of M-theory on the $\mathbb{Z}_N$ quotients of the $G_2$-holonomy flop geometries — equivalently, the type IIA large-$N$ duality between $N$ D6-branes wrapped on the three-sphere of the deformed conifold and the resolved conifold carrying $N$ units of two-form flux — and in its modern rigorous refinements (the $B_7$ and $D_7$ asymptotically locally conical $G_2$ metrics and their conically singular limits and desingularisations), determine whether there exists a limit in which the resulting four-dimensional theory is exactly pure $\mathcal{N}=1$ $SU(N)$ super Yang-Mills, decoupled from bulk M-theory and type IIA gravity and from all non-gauge degrees of freedom, with parametric control: a regime specified by explicit limits of $N$, $V_M$ and the Planck scale in which corrections are quantifiably small, in the way the large-$N$ limit controls the AdS/CFT correspondence. The limit need not be a weakly coupled geometric background, since the original authors explicitly expected the decoupled regime to be strongly curved. The complexified volume $V_M$ is a chiral parameter of the geometric four-dimensional theory with no counterpart in pure $\mathcal{N}=1$ super Yang-Mills, whose only parameters are the coupling and the theta angle, so a decoupling limit must remove this modulus or show that its retention is incompatible with decoupling.

References

1. M. Atiyah, J. Maldacena and C. Vafa, 'An M-theory flop as a large N duality', Journal of Mathematical Physics 42, 3209-3220 (2001). DOI: 10.1063/1.1376159; arXiv:hep-th/0011256; https://arxiv.org/abs/hep-th/0011256
2. C. Vafa, 'Superstrings and topological strings at large N', Journal of Mathematical Physics 42, 2798-2817 (2001). DOI: 10.1063/1.1376161; arXiv:hep-th/0008142; https://arxiv.org/abs/hep-th/0008142
3. R. L. Bryant and S. M. Salamon, 'On the construction of some complete metrics with exceptional holonomy', Duke Mathematical Journal 58 (1989). DOI: 10.1215/s0012-7094-89-05839-0; https://doi.org/10.1215/s0012-7094-89-05839-0
4. A. Brandhuber, J. Gomis, S. S. Gubser and S. Gukov, 'Gauge theory at large N and new $G_2$ holonomy metrics', Nuclear Physics B 611, 179-204 (2001). DOI: 10.1016/S0550-3213(01)00340-6; arXiv:hep-th/0106034; https://arxiv.org/abs/hep-th/0106034
5. M. Cvetič, G. W. Gibbons, H. Lü and C. N. Pope, 'M-theory conifolds', Physical Review Letters 88, 121602 (2002). DOI: 10.1103/PhysRevLett.88.121602; arXiv:hep-th/0112098; https://arxiv.org/abs/hep-th/0112098
6. M. Cvetič, G. W. Gibbons, H. Lü and C. N. Pope, 'A $G_2$ unification of the deformed and resolved conifolds', Physics Letters B 534, 172-180 (2002). DOI: 10.1016/S0370-2693(02)01654-4; arXiv:hep-th/0112138; https://arxiv.org/abs/hep-th/0112138
7. M. Cvetič, G. W. Gibbons, H. Lü and C. N. Pope, 'Cohomogeneity one manifolds of Spin(7) and $G_2$ holonomy', Physical Review D 65, 106004 (2002). DOI: 10.1103/PhysRevD.65.106004; arXiv:hep-th/0108245; https://arxiv.org/abs/hep-th/0108245
8. O. A. Bogoyavlenskaya, 'On a new family of complete $G_2$-holonomy Riemannian metrics on $S^3\times\mathbb{R}^4$', Siberian Mathematical Journal 54, 431-440 (2013). DOI: 10.1134/S0037446613030075; https://doi.org/10.1134/S0037446613030075
9. L. Foscolo, M. Haskins and J. Nordström, 'Infinitely many new families of complete cohomogeneity one $G_2$-manifolds: $G_2$ analogues of the Taub-NUT and Eguchi-Hanson spaces', Journal of the European Mathematical Society 23, 2153-2220 (2021). DOI: 10.4171/jems/1051; arXiv:1805.02612; https://arxiv.org/abs/1805.02612
10. M. Atiyah and E. Witten, 'M-theory dynamics on a manifold of $G_2$ holonomy', Advances in Theoretical and Mathematical Physics 6, 1-106 (2002). DOI: 10.4310/atmp.2002.v6.n1.a1; arXiv:hep-th/0107177; https://arxiv.org/abs/hep-th/0107177
11. B. S. Acharya, 'On realising $\mathcal{N}=1$ super Yang-Mills in M theory', arXiv:hep-th/0011089 (2000); https://arxiv.org/abs/hep-th/0011089
12. B. S. Acharya and S. Gukov, 'M theory and singularities of exceptional holonomy manifolds', Physics Reports 392, 121-189 (2004). DOI: 10.1016/j.physrep.2003.10.017; arXiv:hep-th/0409191; https://arxiv.org/abs/hep-th/0409191
13. J. M. Maldacena and C. Núñez, 'Towards the large $N$ limit of pure $\mathcal{N}=1$ super Yang-Mills theory', Physical Review Letters 86, 588-591 (2001). DOI: 10.1103/PhysRevLett.86.588; arXiv:hep-th/0008001; https://arxiv.org/abs/hep-th/0008001
14. L. Girardello, M. Petrini, M. Porrati and A. Zaffaroni, 'The supergravity dual of $\mathcal{N}=1$ super Yang-Mills theory', Nuclear Physics B 569, 451-469 (2000). DOI: 10.1016/S0550-3213(99)00764-6; arXiv:hep-th/9909047; https://arxiv.org/abs/hep-th/9909047
15. B. S. Acharya, L. Foscolo, M. Najjar and E. E. Svanes, 'New $G_2$-conifolds in M-theory and their field theory interpretation', Journal of High Energy Physics 05 (2021) 250. DOI: 10.1007/JHEP05(2021)250; arXiv:2011.06998; https://arxiv.org/abs/2011.06998