---
title: Four-Derivative Heterotic Supergravity
url: https://www.emergentmind.com/topics/four-derivative-heterotic-supergravity
type: topic
---

# Four-Derivative Heterotic Supergravity

Searching arXiv for recent and foundational papers on four-derivative heterotic supergravity to support a comprehensive article.
Four-derivative heterotic supergravity is the first higher-derivative extension of the ten-dimensional heterotic low-energy effective theory, retaining the leading $\mathcal O(\alpha')$ corrections beyond the two-derivative NS-NS and Yang-Mills sector. In the standard formulation, these corrections are encoded by curvature-squared terms, Chern-Simons modifications of the Kalb-Ramond three-form, and a corresponding deformation of supersymmetry. This sector is central to anomaly cancellation, torsionful geometry, duality-covariant reformulations, compactification, and corrected black-hole physics. Across recent work, it appears in several closely related guises: the Bergshoeff–de Roo torsionful completion in ten dimensions [2406.14600], T-duality-covariant heterotic Double Field Theory constructions [2104.09545; 2101.03376], compactification and consistent truncation analyses [2307.12420], explicit nonsupersymmetric Minkowski vacua with non-Ricci-flat internal spaces [2106.03625], six-dimensional higher-derivative descendants and dualizations [2304.06073; 2209.03981], and four-derivative corrected rotating heterotic black holes [2506.20077].

## 1. Ten-dimensional structure and the Bergshoeff–de Roo completion

In the ten-dimensional heterotic theory, the four-derivative sector is the leading correction to the bosonic Lagrangian and supersymmetry rules. One formulation keeps the Yang-Mills sector explicitly and uses
\[
\mathcal L = \sqrt{-g}\, e^{-2\phi} \left( R + 4(\partial\phi)^2 - \frac{1}{12}H_{MNP}H^{MNP} + \frac{\alpha'}{4} R_{MNPQ}R^{MNPQ} - \frac{\alpha'}{4}\operatorname{tr}(F_{MN}F^{MN}) \right),
\]
together with the heterotic Bianchi identity
\[
dH=\frac{\alpha'}{4}\left(\operatorname{tr}R\wedge R-\operatorname{tr}F\wedge F\right),
\]
which is the anomaly-cancellation condition used in flux compactifications [2106.03625].

A second, standard torsionful formulation truncates away the ten-dimensional gauge fields and writes the action in terms of the connection
\[
\Omega_+=\omega+\frac12\tilde{\mathcal H},
\qquad
\tilde H=H-\frac{\alpha'}{4}\omega_{3L}(\Omega_+),\qquad H=dB,
\]
with bosonic action
\[
e^{-1}\mathcal L_{\mathrm{het}}=e^{-2\phi}\Bigl(R+4(\partial\phi)^2-\frac{1}{12}\tilde H^2+\frac{\alpha'}{8}\bigl(R_{MNAB}(\Omega_+)\bigr)^2\Bigr).
\]
Supersymmetry then uses the opposite torsionful connection,
\[
\delta_\epsilon\psi_M=\nabla_M(\Omega_-)\epsilon,\qquad
\delta_\epsilon\lambda=\left(\Gamma^M\partial_M\phi-\frac{1}{12}\tilde H_{MNP}\Gamma^{MNP}\right)\epsilon,
\]
with $\Omega_-=\omega-\frac12\tilde{\mathcal H}$; the use of $\Omega_+$ in the action and $\Omega_-$ in the fermion variations is essential for supersymmetry [2406.14600].

This torsionful BdR structure reappears in several later developments. The torus-reduction analysis of higher-derivative heterotic supergravity adopts
\[
e^{-1}\mathcal L = e^{-2\phi}\left[ R+4(\partial_M\phi)^2-\frac{1}{12}\tilde H_{MNP}^2 +\frac{\alpha'}{8}\big(R_{MNAB}(\Omega_+)\big)^2 \right] +\mathcal O(\alpha'^3),
\]
with
\[
\tilde H = H-\frac{\alpha'}{4}\,\omega_{3L}(\Omega_+),
\qquad
d\tilde H=-\frac{\alpha'}{4}\Tr\!\big[R(\Omega_+)\wedge R(\Omega_+)\big],
\]
and corresponding bosonic equations of motion for the dilaton, metric, and three-form [2307.12420]. The same general structure underlies the four-dimensional black-hole analysis, where the four-derivative term is the torsionful curvature-squared invariant accompanied by the Lorentz Chern-Simons modification of $H$ [2506.20077].

A recurrent misconception is that the four-derivative sector is exhausted by an ordinary $R^2$ deformation. The ten-dimensional heterotic completion is instead organized by torsionful curvature, Chern-Simons modifications, and deformed supersymmetry, so the curvature-squared term is inseparable from the Green-Schwarz structure in the standard heterotic scheme [2406.14600; 2104.09545].

## 2. Green-Schwarz mechanism, torsionful geometry, and supersymmetry

The Green-Schwarz mechanism is the structural principle that ties anomaly cancellation to the higher-derivative completion. In ordinary supergravity variables, the corrected three-form is written as
\[
\hat H = db - \Omega_{\rm L}(\omega_-) + \Omega_{\rm YM}(A),
\]
or, in the notation of the T-duality-covariant construction,
\[
H_{\mu\nu\rho} =3\partial_{[\mu}b_{\nu\rho]} - \Omega_{\mu\nu\rho}^{\rm YM} + \Omega_{\mu\nu\rho}^{\rm L} + \text{fermionic terms}.
\]
The two-form therefore acquires Lorentz and gauge Green-Schwarz shifts, and its supersymmetry variation is correspondingly deformed [2104.09545].

The T-duality-covariant construction based on exact ten-dimensional ${\cal N}=1$ supergravity coupled to Yang-Mills makes this mechanism manifest in a generalized-geometry framework with global duality group
\[
O(10,10+n_g)
\]
and local double Lorentz symmetry
\[
O(9,1)_L \times O(1,9+n_g)_R.
\]
Its central result is a supersymmetric and T-duality-covariant generalization of the Green-Schwarz mechanism that fixes the first-order $\alpha'$ corrections to both the transformation rules and the invariant action [2104.09545].

In heterotic Double Field Theory, the same logic appears as a generalized Green-Schwarz deformation of the frame,
\[
\delta^{(1)} E_M{}^{A} = - E_M{}^{B}\,\partial_B \Lambda^{CD}\,F_{CD}{}^{A},
\]
which the GKSA analysis identifies as the DFT counterpart of the heterotic anomalous $B$-field transformation [2101.03376]. This DFT presentation is not merely a reformulation; it organizes the four-derivative sector directly in duality-covariant variables and shows how higher-curvature data arise geometrically from generalized fluxes [2104.09545; 2101.03376].

Supersymmetry algebra closure is likewise deformed. The commutator of two supersymmetries still closes into generalized diffeomorphisms, local Lorentz transformations, and supersymmetry, but with field-dependent $\alpha'$ corrections to the parameters,
\[
[\delta_{\epsilon_1},\delta_{\epsilon_2}] = \delta_{\xi_{12}} + \delta_{\Lambda_{12}} + \delta_{\epsilon_{12}},
\]
where $\xi_{12}$, $\Lambda_{12}$, and $\epsilon_{12}$ each receive higher-derivative corrections built from generalized fluxes and fermionic curvatures [2104.09545]. This suggests that four-derivative heterotic supergravity is not merely a corrected action but a deformed local symmetry system.

## 3. Duality-covariant reduction and torus compactification

Compactification on a $d$-torus turns ten-dimensional heterotic supergravity without gauge fields into half-maximal supergravity in $10-d$ dimensions coupled to $d$ vector multiplets [2406.14600]. At two derivatives, the standard Kaluza-Klein ansatz is
\[
ds^2=g_{\mu\nu}dx^\mu dx^\nu+g_{ij}\eta^i\eta^j,\qquad
B=b_{\mu\nu}dx^\mu\wedge dx^\nu+B_{\mu i}dx^\mu\wedge\eta^i+\tfrac12 b_{ij}\eta^i\wedge\eta^j,
\]
with $\eta^i=dy^i+A_\mu^i dx^\mu$. The internal moduli reorganize into the $O(d,d)$ scalar matrix
\[
\mathcal H=\begin{pmatrix}g-bg^{-1}b & bg^{-1}\\ -g^{-1}b & g^{-1}\end{pmatrix},
\]
while the reduced scalar manifold is
\[
O(d,d;\mathbb R)/O(d)_-\times O(d)_+.
\]
A key result is that the continuous $O(d,d;\mathbb R)$ symmetry persists to all perturbative orders in the string $\alpha'$ expansion, and after suitable field redefinitions the reduced four-derivative action and supersymmetry variations become manifestly $O(d)_-\times O(d)_+$ invariant [2406.14600].

In the generalized-vielbein basis, vector fields split into $O(d)_-$ and $O(d)_+$ components, and the Maurer-Cartan form decomposes into a composite connection $Q$ and coset vielbein $P$,
\[
L=\begin{pmatrix}Q^{(--)}&0\\ 0&Q^{(++)}\end{pmatrix} +\begin{pmatrix}0&P^{(-+)}\\ P^{(+-)}&0\end{pmatrix}.
\]
In the conventions of the reduction, $O(d)_-$ is the supergravity $R$-symmetry and $O(d)_+$ is a flavor symmetry of the $d$ vector multiplets [2406.14600]. This asymmetry is specific to the heterotic theory and is tied to the Lorentz Chern-Simons term; a bosonic-string analogue would not distinguish the two $O(d)$ factors in the same way [2406.14600].

The first pass of the reduction produces many terms not manifestly covariant under $O(d)_-\times O(d)_+$, including cross terms such as
\[
R_{\gamma\delta}{}^{\alpha\beta}(\Omega_+)\supset -\frac14 F_{\alpha\beta}^{(-)\,c}F_{\gamma\delta}^{(+)\,c}.
\]
The crucial step is then a field redefinition of the internal metric/vielbein,
\[
\delta e_i^a=T^{ab}e_i^b,\qquad
\delta g_{ij}=2e_i^aT^{ab}e_j^b,
\]
with
\[
T^{ab}=\frac12\,F_{\mu\nu}^{(-)\,a}F_{\mu\nu}^{(-)\,b}+\frac14\,P_\mu^{(+-)\,ca}P_\mu^{(+-)\,cb},
\]
which removes the non-invariant pieces and makes the action and supersymmetry rules manifestly covariant [2406.14600].

After redefinition, the four-derivative action is organized as
\[
\mathcal L_{\partial^4}=\mathcal L_1+\mathcal L_2+\mathcal L_3,
\]
with $\mathcal L_1$ pure gravity-multiplet terms, $\mathcal L_2$ pure vector-multiplet terms, and $\mathcal L_3$ mixed couplings [2406.14600]. This decomposition is structurally important: it shows that higher derivatives do not destroy the lower-dimensional multiplet splitting, but they do induce controlled gravity/vector-multiplet mixing fixed by the coset geometry and composite connection.

## 4. Consistent truncations and lower-dimensional effective theories

Torus reduction generically produces extra vector multiplets, and a central question is whether they can be consistently removed once four-derivative corrections are included. The consistent truncation analysis shows that they can [2307.12420].

At leading order, the bosonic supersymmetric truncation is
\[
g_{ij}=\delta_{ij},\qquad b_{ij}=0,\qquad A_\mu^i=A_\mu^{(-)\,i},\qquad B_{\mu i}=-A_\mu^{(-)\,i},
\]
equivalently
\[
F^{(+)\,i}=0,\qquad F^{(-)\,i}\neq 0.
\]
The resulting two-derivative Lagrangian is
\[
e^{-1}\mathcal L = e^{-2\varphi}\left[ R+4(\partial\varphi)^2-\frac{1}{12}h^2-\frac18(F_{\mu\nu}^{(-)\,i})^2 \right],
\]
with
\[
h=db+\frac14 F^{(-)\,i}\wedge A^{(-)\,i}, \qquad dh=\frac14 F^{(-)\,i}\wedge F^{(-)\,i}.
\]
This retains the lower-dimensional gravity multiplet plus graviphotons [2307.12420].

At $\mathcal O(\alpha')$, the truncation requires corrections:
\[
g_{ij}=\delta_{ij}+\alpha'\delta g_{ij},\qquad b_{ij}=\alpha'\delta b_{ij},
\]
with the key choice
\[
\delta g_{ij}=\frac{1}{16}F_{\mu\nu}^{(-)\,i}F_{\mu\nu}^{(-)\,j}, \qquad \delta b_{ij}=0,
\]
together with
\[
\delta F_{\mu\nu}^j+\delta G_{\mu\nu j}=0.
\]
These shifts cancel the four-derivative sources in the vector-multiplet equations of motion [2307.12420]. The reduced truncated action then acquires curvature, $RFF$, $(\nabla F)^2$, and quartic $F^4$ terms, while the three-form becomes
\[
\tilde h = db+\frac14 A^{(-)\,i}\wedge F^{(-)\,i} -\frac{\alpha'}{4}\omega_{3L}(\omega_+),
\]
with
\[
d\tilde h = \frac14 F^{(-)\,i}\wedge F^{(-)\,i} -\frac{\alpha'}{4}\Tr\big[R(\omega_+)\wedge R(\omega_+)\big].
\]
The fermionic consistency check is equally nontrivial: the gaugini require the redefinition
\[
\tilde\psi_i=\psi_i-\frac{\alpha'}{4}F_{\mu\nu}^{(-)\,i}\,\mathcal D_\mu\psi_\nu,
\]
after which
\[
\delta_\epsilon \tilde\psi_i=0,
\]
so the truncation is compatible with the Killing spinor equations [2307.12420].

A plausible implication is that four-derivative consistent truncation is highly scheme-sensitive: the lower-dimensional field content survives only after $\alpha'$-dependent redefinitions of both bosons and fermions.

## 5. Compactification beyond Ricci-flatness

One of the most concrete applications of the four-derivative heterotic action is the construction of four-dimensional Minkowski vacua with curved internal manifolds that are not Ricci-flat [2106.03625]. The starting point is ten-dimensional heterotic supergravity with constant dilaton and $H=0$, for which the remaining equations reduce to
\[
R+\frac{\alpha'}{4}R_{MNPQ}R^{MNPQ} -\frac{\alpha'}{4}\operatorname{tr}(F_{MN}F^{MN})=0,
\]
\[
R_{MN}+\frac{\alpha'}{4}R_{MPQR}R_N{}^{PQR} -\frac{\alpha'}{4}\operatorname{tr}(F_{MP}F_N{}^{P})=0,
\]
\[
D_MF^{MN}=0.
\]
An important point is that the trace of the Einstein equation implies
\[
R=0,
\]
but this does not imply Ricci flatness $R_{MN}=0$ [2106.03625]. Positive and negative curvature components can cancel in the scalar curvature, opening a route to Minkowski compactifications with curved internal spaces.

The compactification ansatz takes
\[
M_{10}=M_0\times M_1\times M_2\times M_3,
\]
where $M_0$ is four-dimensional Minkowski spacetime and each $M_i$ is a two-dimensional space of constant sectional curvature,
\[
R_{mnpq}^{(i)} = \lambda_i\left(g_{mp}g_{nq}-g_{mq}g_{np}\right).
\]
Depending on the sign of $\lambda_i$, the factor is $S^2$, $T^2$, or a compact hyperbolic surface $H^2/\Gamma$ [2106.03625]. Gauge flux is chosen block-diagonally with Freund-Rubin-like form
\[
F^{(i)}_{mn}=\sqrt{g_i}\, f_A\, \epsilon_{mn},
\]
which automatically solves the gauge equation and permits anomaly cancellation without the standard embedding [2106.03625].

The equations reduce to algebraic conditions,
\[
\lambda_i+\frac{\alpha'}{2}\left(\lambda_i^2-\sum_A (f_A^{(i)})^2\right)=0,
\qquad
\sum_{i=1}^3 \lambda_i=0.
\]
Flux quantization and the Gauss-Bonnet theorem enforce discreteness,
\[
\int_{M_i} F = 2\pi n_i,\qquad n_i\in\mathbb Z,
\qquad
\int_{M_i} R = 4\pi \chi_i,
\]
implying
\[
f_A \propto \frac{n_A}{\chi_i}.
\]
The allowed solutions fall into exactly three classes [2106.03625].

| Internal space | Curvature balance | Noted feature |
|---|---|---|
| $S^2\times T^2\times H^2/\Gamma$ | $\lambda_{S^2}+\lambda_{H^2/\Gamma}=0$ | minimal nonzero flux uses two $U(1)$ fluxes |
| $S^2\times H^2/\Gamma\times H^2/\Gamma$ | $\lambda_{S^2}+\lambda_{H^2/\Gamma}^{(2)}+\lambda_{H^2/\Gamma}^{(3)}=0$ | no flat torus is required |
| $S^2\times S^2\times H^2/\Gamma$ | $\lambda_{S^2}^{(1)}+\lambda_{S^2}^{(2)}+\lambda_{H^2/\Gamma}=0$ | at least three flux components are needed |

Representative examples include $(\lambda_1,\lambda_3)=(1,-1)$ for $S^2\times T^2\times H^2/\Gamma$, and $(\lambda_1,\lambda_2,\lambda_3)=(4,4,-8)$ or $\lambda_1=\lambda_2=2,\lambda_3=-4$ for $S^2\times S^2\times H^2/\Gamma$ [2106.03625]. These are nonsupersymmetric, non-Ricci-flat Minkowski vacua whose existence relies essentially on the curvature-squared term. Without the $\alpha'R^2$ contribution, the sign structure preventing such positive/negative curvature balance would remain [2106.03625].

## 6. Six-dimensional descendants, dualization, and independent invariants

In six dimensions, the four-derivative heterotic sector exhibits a richer invariant structure than in ten dimensions. For $N=(1,0)$ supergravity coupled to a tensor multiplet, there are two known off-shell supersymmetric invariants at four derivatives: a Riemann-squared invariant $L_{\text{Riem}^2}$ and a Gauss-Bonnet invariant $L_{\text{GB}}$ [2209.03981]. The general two-parameter theory is
\[
L_{\alpha,\gamma}=L_{\text{EH}}-\frac18 \alpha\,L_{\text{Riem}^2}-\frac18 \gamma\,L_{\text{GB}}.
\]
A key result is that these two invariants are not related by field redefinitions. Although some $R_{\mu\nu}R^{\mu\nu}$ and $R^2$ terms can be shifted away, the difference still contains structures such as
\[
H^4,\qquad H^2(\partial\phi)^2,\qquad (\partial\phi)^4,
\]
so the invariants remain genuinely distinct [2209.03981].

The same paper develops direct action-level dualization with a Lagrange multiplier,
\[
\Delta L(B,C)=\frac{1}{2\cdot 3!}\,\epsilon^{\mu\nu\rho\sigma\lambda\tau}H_{\mu\nu\rho}\,\partial_\sigma C_{\tau\lambda},
\]
treating $H=dB$ as independent and obtaining a dual two-form theory with field strength
\[
G_{\mu\nu\rho}=3\partial_{[\mu}C_{\nu\rho]} - 3\alpha\,\omega_{\mu\nu\rho}(\omega_-(H)) - 3\gamma\,\omega_{\mu\nu\rho}(\omega_+(H)).
\]
The special case $\gamma=\alpha$ gives the dual of the BdR-type combination; $\gamma=-\alpha$ yields a different higher-derivative extension whose curvature-squared sector can be removed by field redefinitions at the cost of more complicated supersymmetry rules [2209.03981].

In ten dimensions, by contrast, dualizing the BdR action does not produce an analogous clean split into two independent invariants. The dualized result contains terms such as
\[
R_{\mu\nu\rho\sigma}H^{\mu\nu,\rho\sigma},\qquad H^4,\qquad (DH)^2,\qquad H\phi\,DH,\qquad H_7\omega(\omega),
\]
and the paper identifies an obstacle to separating the result into a sum of two independent invariants because some terms do not lift from $6D$ to $10D$ [2209.03981]. This underpins the view that the ten-dimensional BdR completion is much closer to unique than its six-dimensional descendants.

A related six-dimensional development constructs the four-derivative supersymmetric extension of $(1,0)$ supergravity coupled to Yang-Mills and hypermultiplets on
\[
\mathbb{H}P(n)=\frac{Sp(n,1)}{Sp(n)\times Sp(1)_R},
\]
with a new independent coupling $\gamma$ governing higher-derivative hypermultiplet interactions [2304.06073]. The hyperscalar kinetic term
\[
-\frac12 e\,e^{2\phi} P_\mu{}^{aA}P^\mu{}_{aA}
\]
is not deformed, so the quaternionic Kähler geometry is preserved [2304.06073]. In the special $\mathbb H P(1)$ case, comparison with heterotic supergravity compactified on $T^4$ shows agreement for $\gamma=0$ or $\gamma=-2\alpha$, depending on how $Sp(1)\subset SO(4)$ is embedded [2304.06073]. This suggests that four-derivative heterotic descendants can preserve target-space geometry while still admitting new independent couplings not fixed solely by the BdR invariant.

## 7. Black holes, observables, and broader extensions

Four-derivative heterotic supergravity has recently been used to obtain explicit corrections to the Kerr-Sen black hole [2506.20077]. The starting point is the four-dimensional string-frame BdR action
\[
e^{-1}\mathcal{L} = e^{-2\phi}\left[ R(\Omega)+4(\partial\phi)^2-\frac{1}{12}\tilde H^2 +\frac{\alpha'}{8}R_{\hat\mu\hat\nu\hat\rho\hat\sigma}(\Omega_-)^2 \right],
\]
with
\[
\tilde H = H+\frac{\alpha'}{4}\omega_{3L}(\Omega_-),\qquad H=dB,
\qquad
d\tilde H=\frac{\alpha'}{4}\operatorname{Tr}\!\left[R(\Omega_-)\wedge R(\Omega_-)\right].
\]
After dualizing $B_{\hat\mu\hat\nu}$ to an axion $\varphi$, the Einstein-frame two-derivative sector becomes
\[
e^{-1}\widetilde{\mathcal L}^{(2)}_E = R-2(\partial\phi)^2-\frac12 e^{4\phi}(\partial\varphi)^2,
\]
while the $\alpha'$ correction is an explicit functional of $R_{\mu\nu\rho\sigma}$, $\phi$, and the axion current $C_\mu=e^{2\phi}\nabla_\mu\varphi$ [2506.20077].

The Kerr solution can be embedded with $\phi=\varphi=0$ at zeroth order, and an important structural feature is that in Einstein frame the Kerr metric itself is uncorrected at this order; only the scalars acquire hair [2506.20077]. The Kerr-Sen solution is then generated by an $O(2,1)$ boost, but only after field redefinitions that restore manifest higher-derivative hidden symmetry in the reduced action [2506.20077]. This requirement reflects the same general theme seen in torus compactification: duality symmetry survives at four derivatives, but only in an appropriate field basis.

The corrected multipole moments are a particularly sharp result. At two derivatives, Kerr-Sen, Kerr, and Kerr-Newman have coincident gravitational multipoles, but at four derivatives the heterotic solution develops distinct corrections such as
\[
\delta M_2 = -\frac{143}{120}M\chi_a^2\chi_Q^2\left(1-\frac{265}{2002}\chi_a^2+\cdots\right),
\]
\[
\delta\mathcal S_3 = -\frac{11}{10}M^2\chi_a^3\chi_Q^2\left(1-\frac{219}{1232}\chi_a^2+\cdots\right),
\]
\[
\delta\mathcal Q_2 = -\frac{1}{30}M\chi_a^2\chi_Q\left(1+\frac{25}{56}\chi_a^2+\cdots\right),
\]
\[
\delta\mathcal P_1 = \frac{9}{32}\chi_a\chi_Q\left(1-\frac{1}{6}\chi_a^2+\cdots\right),
\]
showing that the four-derivative heterotic corrections are distinguishable from both Kerr and Kerr-Newman, even allowing the most general choice of four-derivative Einstein-Maxwell corrections [2506.20077].

Beyond relativistic backgrounds, the four-derivative gravitational corrections have also been extended to the non-relativistic heterotic limit. There, the BdR identification is adapted to Newton-Cartan/string-Newton-Cartan geometry, and the gauge connection is identified with a torsionful $SO(8)$ spin connection,
\[
a_\mu{}^{a'b'}=\frac1{\sqrt2}\,S^{(-)}_\mu{}^{a'b'},
\qquad
F_{\mu\nu a'b'}\equiv -\frac1{\sqrt2}R^{(-)}_{\mu\nu a'b'}.
\]
A field redefinition
\[
\bar b_{\mu\nu} = b_{\mu\nu} + \frac{5}{12}\,S^{(-)}_{[\mu}{}^{a'b'}\,\tau_{\nu]}{}^+\, \tau_{\rho\sigma}^-\,e^\rho{}_{a'}e^\sigma{}_{b'}
\]
then exposes an $SO(8)$ Green-Schwarz mechanism in the non-relativistic theory [2508.09250]. This mechanism is trivializable by field redefinitions, unlike the relativistic Lorentz Green-Schwarz shift, and the resulting higher-curvature completion is finite and symmetry-consistent [2508.09250]. This suggests that the organizing role of BdR geometry and Green-Schwarz structure extends beyond relativistic supergravity.

Another direction is exact four-derivative couplings in lower-dimensional heterotic CHL orbifolds, where a three-dimensional scalar coupling
\[
F_{abcd}(\Phi)(\nabla\Phi)^4
\]
is proposed as an exact U-duality invariant modular integral over $\Gamma_0(N)\backslash\mathbb H$, reproducing tree-level, one-loop, and non-perturbative effects as well as exact four-dimensional $F^4$ couplings [1702.01926]. Although this is not a ten-dimensional supergravity construction, it shows how the four-derivative sector of heterotic theories is tightly constrained by supersymmetric Ward identities, automorphic structure, and BPS spectra.

Taken together, these developments define four-derivative heterotic supergravity as a broad but tightly structured framework. Its characteristic ingredients are torsionful curvature-squared invariants, Green-Schwarz-modified three-form geometry, deformed supersymmetry, and nontrivial field-redefinition dependence. Its applications range from explicit compactifications and consistent truncations to exact duality-covariant formulations and corrected black-hole observables. The current literature indicates that the ten-dimensional BdR completion remains the central organizing principle, while reduction, dualization, and symmetry-covariant reformulation reveal both its rigidity and its diverse lower-dimensional consequences [2406.14600; 2104.09545; 2307.12420; 2209.03981; 2506.20077].

Source: https://www.emergentmind.com/topics/four-derivative-heterotic-supergravity