Supersymmetric $R^2$ actions in ten dimensions: completing the two-parameter Riemann-squared/Gauss-Bonnet extension beyond $H_{(7)}$-independent variations, and uniqueness of the Bergshoeff-de Roo action

In 1987 Bergshoeff, Salam and Sezgin constructed off-shell superconformal curvature-squared actions in six-dimensional $N=(1,0)$ supergravity, including the Lorentz Chern-Simons coupling of the two-form field required by supersymmetry, and began the analogous program in ten dimensions: a two-parameter family combining the Riemann-squared and Gauss-Bonnet curvature combinations, supersymmetric only up to variations involving the seven-form field strength $H_{(7)}$ of the dual six-form formulation. The six-dimensional part of this program has since been completed (the off-shell Gauss-Bonnet invariant was finally constructed in 2017-2019, and the failure of the 1987 partial construction was traced to the assumption that the two-form dependence is captured by a torsionful connection alone), the ten-dimensional Riemann-squared member was completed by de Roo in 1992, and the required infinite-order deformation in $\alpha'$ with string-fixed coefficients is understood. What remains open is the ten-dimensional core: the supersymmetric completion of the two-parameter family including all $H_{(7)}$-dependent variations, whether supersymmetry relates or fixes the two parameters, whether the Bergshoeff-de Roo action is the unique four-derivative extension of ten-dimensional $N=1$ supergravity up to field redefinitions, and the fully supersymmetric all-orders dualization between the two-form and six-form formulations.

Background

Setting. String theory predicts that the low-energy supergravity action of the heterotic string is corrected by higher-derivative terms organized in powers of the Regge slope parameter $\alpha'$ of the worldsheet theory; at leading order these are curvature-squared terms. In $d>4$, a generic curvature-squared action $c_1 R_{\mu\nu ab}R^{\mu\nu ab}+c_2 R_{\mu a}R^{\mu a}+c_3 R^2$ propagates, besides the massless graviton, a massive spin-2 ghost (its linearized equations are fourth order). The unique ghost-free curvature-squared combination is the Euler density, the Gauss-Bonnet (GB) combination $R_{\mu\nu ab}R^{\mu\nu ab}-4R_{\mu a}R^{\mu a}+R^2$, topological in $d=4$ but dynamical in $d>4$ (here $R_{\mu\nu}{}^{ab}$ is the Riemann tensor, $R_{\mu}{}^{a}=R_{\mu\nu}{}^{ab}e_b{}^{\nu}$ the Ricci tensor built with the vielbein $e_\mu{}^a$ - the local orthonormal frame field carrying the tangent-space indices $a,b$ - and $R$ the scalar curvature).

Lorentz Chern-Simons term and anomaly cancellation. Ten-dimensional heterotic supergravity contains a two-form gauge potential $B_{\mu\nu}$ whose three-form field strength is modified by Chern-Simons (CS) three-forms, schematically $H=dB+\alpha'\,X_3(\omega)-\beta'\,X_3(A)$, where $X_3(\Gamma)=\mathrm{Tr}(\Gamma\wedge d\Gamma+\tfrac{2}{3}\Gamma\wedge\Gamma\wedge\Gamma)$ is the CS form of a connection $\Gamma$ (Lorentz $\omega$ or Yang-Mills $A$). This Green-Schwarz modification cancels the one-loop gravitational and gauge anomalies of the heterotic string, and it ties the $R^2$ terms to the CS couplings: supersymmetrizing one forces the other.

The 1987 construction. Bergshoeff, Salam and Sezgin (BSS87) used the six-dimensional $N=(1,0)$ superconformal tensor calculus ($N=(1,0)$ denotes minimal chiral supersymmetry in six dimensions, with 8 supercharges) - an off-shell formalism in which the Weyl multiplet (vielbein, gravitino (the spin-3/2 superpartner of the graviton), an SU(2) gauge field $V_\mu^{ij}$ and a two-form $B_{\mu\nu}$) closes under supersymmetry without using equations of motion - to construct the first off-shell supersymmetric Riemann-squared action, together with its required Lorentz CS coupling $-\tfrac{i}{4}\varepsilon^{\mu\nu\rho\sigma\lambda\tau}R_{\mu\nu}{}^{ab}(\omega_-)R_{\rho\sigma}{}^{ab}(\omega_-)B_{\lambda\tau}$. Here $\omega_\pm=\omega\pm\tfrac{1}{2}H$ are torsionful spin connections (the torsion being supplied by $H_{\mu\nu\rho}=3\partial_{[\mu}B_{\nu\rho]}+\ldots$), and the torsion interpretation is the ansatz that all $B$-field dependence of the invariant enters through $\omega_\pm$. BSS87 also showed that any $R^2$ action can be made locally conformally invariant via a compensating dilaton $\phi$ and inert vielbein $\hat e_\mu{}^a=\phi e_\mu{}^a$, and they began the analogous construction in ten dimensions, where the conformal multiplet is naturally written with a six-form potential $B_{\mu_1\ldots\mu_6}$ carrying a seven-form field strength $H_{(7)}$ (the magnetic dual of the three-form $H$; the corresponding soliton is the heterotic five-brane).

BdR action. The Bergshoeff-de Roo (BdR) action is the fully supersymmetric $O(\alpha')$ four-derivative extension of ten-dimensional $N=1$ supergravity including the CS-modified $H$, constructed by Bergshoeff and de Roo (1989); supersymmetry forces $H$ to be defined recursively, each order in $\alpha'$ built from the CS form of the previous-order connection ($H^{(0)}=dB$ and $H^{(n)}=\alpha'\,X_3(\omega_-)^{(n-1)}$ for $n\ge 2$, with $\alpha'$ here the Lorentz-CS coupling), so the action and its supersymmetry transformations are corrected to all orders in $\alpha'$. The $O(\alpha')$ heterotic effective action obtained from demanding Weyl invariance of the string worldsheet sigma model order by order in $\alpha'$ (the beta-function calculations of Metsaev and Tseytlin) agrees with the BdR action. The BdR action settles the two-form-formulation extension of the $O(\alpha')$ effective action with the CS-modified $H$; the Problem Statement asks beyond it for the off-shell completion of the general two-parameter curvature-squared family in the ten-dimensional conformal multiplet - the six-form formulation, whose $H_{(7)}$-dependent variations the 1987 construction reached only partially - and for the relation of the two parameters.

Dualization. A duality transformation here means trading the two-form potential $B_2$ for its magnetic dual six-form $B_6$ (or vice versa) by introducing Lagrange multipliers that impose the CS-modified Bianchi identity and then eliminating the original field - the classical field-theory counterpart of string/five-brane duality. The source paper remarked that this procedure appears obstructed for off-shell $R^2$ actions because of their derivative and higher-power dependence on $H$. Chang, Sezgin and Tanii (2022) subsequently carried out this dualization of the ten-dimensional Bergshoeff-de Roo action at the level of the action and matched its bosonic part, up to field redefinitions, to the earlier superspace dual formulation of Saulina, Terentev and Zyablyuk (reproduced in Ozkan, Pang and Sezgin 2024), obtaining the supersymmetry transformation rules of the dual fields to cubic-fermion order and first order in $\alpha'$. The fermionic-sector action and the matching of the full construction to the two-form formulation up to field redefinitions, and the extension of the duality to all orders in $\alpha'$, remain open.

Problem Statement

Complete the ten-dimensional part of the supersymmetric $R^2$-actions program left unfinished by Bergshoeff, Salam and Sezgin (1987). The six-dimensional part is finished and is not re-asked: the six-dimensional off-shell Riemann-squared invariant was given in the source paper, and the six-dimensional off-shell Gauss-Bonnet invariant has since been constructed (Novak et al. 2017; Butter et al. 2019).

Specifically, in ten-dimensional $N=1$ supergravity with the Green-Schwarz/Lorentz-Chern-Simons-modified field strength, determine the full structure of the four-derivative supersymmetric extensions:

1. Construct the supersymmetric extension of the two-parameter bosonic family $c_R\,R_{\mu\nu ab}R^{\mu\nu ab}+c_{GB}\,(R_{\mu\nu ab}R^{\mu\nu ab}-4R_{\mu a}R^{\mu a}+R^2)$ - including all Lorentz-Chern-Simons-related couplings and all terms involving the seven-form field strength $H_{(7)}$ of the dual six-form formulation - that is invariant under all supersymmetry variations, not only those independent of $H_{(7)}$, taking into account the differential constraints satisfied by the fields of the ten-dimensional conformal multiplet; or prove that no such two-parameter completion exists in ten dimensions.

2. Decide whether supersymmetry relates the two parameters (for example, fixes their ratio to the value realized in the Bergshoeff-de Roo action, or forces the Gauss-Bonnet combination), or whether a genuine two-parameter family of four-derivative extensions exists in ten dimensions.

3. Prove or disprove that the Bergshoeff-de Roo action is, up to field redefinitions, the unique four-derivative supersymmetric extension of ten-dimensional $N=1$ supergravity.

4. Construct the dual (six-form, five-brane) formulation of the resulting four-derivative theory (or, if part 1 is answered in the negative, of the Bergshoeff-de Roo action) completely - including the fermionic sector and the supersymmetry transformations of the dual fields, matching the two-form formulation up to field redefinitions - and either extend the duality to all orders in $\alpha'$ (as the recursive definition of $H$ requires) or prove that an obstruction exists.

The negative branches overlap: a proof of uniqueness (part 3) settles both the no-inequivalent-family branch of part 2 and the no-two-parameter-completion branch of part 1, so a single uniqueness result discharges those branches. In every part the treatment of the $H_{(7)}$-dependent variations must be explicit, since the source construction was invariant only up to precisely those terms.

References

1. E. Bergshoeff, A. Salam, E. Sezgin, "Supersymmetric $R^2$ actions, conformal invariance and the Lorentz Chern-Simons term in 6 and 10 dimensions", Nucl. Phys. B 279 (1987) 659-683. https://doi.org/10.1016/0550-3213(87)90015-0
2. E. Bergshoeff, A. Salam, E. Sezgin, "A supersymmetric $R^2$-action in six dimensions and torsion", Phys. Lett. B 173 (1986) 73-76. https://doi.org/10.1016/0370-2693(86)91233-5
3. M. de Roo, "The $R^2$-action in $d=10$ conformal supergravity", Nucl. Phys. B 372 (1992) 243-269. https://doi.org/10.1016/0550-3213(92)90319-7
4. E. Bergshoeff, M. de Roo, "Supersymmetric Chern-Simons terms in ten dimensions", Phys. Lett. B 218 (1989) 210-215. https://doi.org/10.1016/0370-2693(89)91420-2
5. E. A. Bergshoeff, M. de Roo, "The quartic effective action of the heterotic string and supersymmetry", Nucl. Phys. B 328 (1989) 439-468. https://doi.org/10.1016/0550-3213(89)90336-2
6. J. Novak, M. Ozkan, Y. Pang, G. Tartaglino-Mazzucchelli, "Gauss-Bonnet supergravity in six dimensions", Phys. Rev. Lett. 119 (2017) 111602. arXiv:1706.09330, https://doi.org/10.1103/PhysRevLett.119.111602
7. D. Butter, J. Novak, M. Ozkan, Y. Pang, G. Tartaglino-Mazzucchelli, "Curvature squared invariants in six-dimensional $N=(1,0)$ supergravity", JHEP 04 (2019) 013. arXiv:1808.00459, https://doi.org/10.1007/JHEP04(2019)013
8. D. Butter, S. M. Kuzenko, J. Novak, S. Theisen, "Invariants for minimal conformal supergravity in six dimensions", JHEP 12 (2016) 072. arXiv:1606.02921, https://doi.org/10.1007/JHEP12(2016)072
9. H.-Y. Chang, E. Sezgin, Y. Tanii, "Dualization of higher derivative heterotic supergravities in 6D and 10D", JHEP 10 (2022) 062. arXiv:2209.03981, https://doi.org/10.1007/JHEP10(2022)062
10. E. A. Bergshoeff, M. de Roo, "The string effective action in the dual formulation of $D=10$ supergravity", Phys. Lett. B 247 (1990) 530-536. https://doi.org/10.1016/0370-2693(90)91896-J
11. E. A. Bergshoeff, M. de Roo, "Duality transformations of string effective actions", Phys. Lett. B 249 (1990) 27-34. https://doi.org/10.1016/0370-2693(90)90522-8
12. M. de Roo, H. Suelmann, A. Wiedemann, "The supersymmetric effective action of the heterotic string in ten dimensions", Nucl. Phys. B 405 (1993) 326-366. arXiv:hep-th/9210099, https://doi.org/10.1016/0550-3213(93)90550-9
13. M. Ozkan, Y. Pang, E. Sezgin, "Higher derivative supergravities in diverse dimensions", Phys. Rep. 1086 (2024) 1-95. arXiv:2401.08945, https://doi.org/10.1016/j.physrep.2024.07.002
14. H. Johansson, J. Nohle, "Conformal gravity from gauge theory", arXiv:1707.02965 [hep-th]. https://arxiv.org/abs/1707.02965
15. R. R. Metsaev, A. A. Tseytlin, "Order $\alpha'$ (two loop) equivalence of the string equations of motion and the sigma model Weyl invariance conditions: dependence on the dilaton and the antisymmetric tensor", Nucl. Phys. B 293 (1987) 385-419. https://doi.org/10.1016/0550-3213(87)90077-0
16. J. T. Liu, R. Minasian, "Higher-derivative couplings in string theory: dualities and the $B$-field", Nucl. Phys. B 874 (2013) 413-470. arXiv:1304.3137, https://doi.org/10.1016/j.nuclphysb.2013.06.002