---
title: Hellings–Nordtvedt Gravity Theory
url: https://www.emergentmind.com/topics/hellings-nordtvedt-theory
type: topic
---

# Hellings–Nordtvedt Gravity Theory

Searching arXiv for the target paper and closely related Hellings–Nordtvedt references.
Hellings–Nordtvedt theory is a vector–tensor theory of gravity in which a vector field $A_\mu$ is nonminimally coupled to curvature through two independent interactions, $A^2{\cal R}$ and $A^\mu A^\nu{\cal R}_{\mu\nu}$. In the massive version studied in "Black holes and neutron stars in massive Hellings-Nordtvedt theory" [2605.14711], the original theory is supplemented by a bumblebee-type potential whose zero-energy minimum occurs at a nonzero value of the vector invariant $X\equiv A^2$. The resulting framework combines spontaneous Lorentz symmetry breaking with diffeomorphism invariance, and it supports distinct compact-object sectors depending on which nonminimal coupling is retained. A central result is that an asymptotic vacuum with nonzero vector vacuum expectation value does not generically permit both curvature couplings simultaneously; instead, it selects two disjoint single-coupling sectors with sharply different asymptotic and astrophysical behavior.

## 1. Lagrangian structure and field content

In geometric units $G=c=1$, the massive Hellings–Nordtvedt theory considered in [2605.14711] is defined by
$$
S \;=\; \frac{1}{16\pi}\int d^4x\,\sqrt{-g}\,L \;+\; S_{\rm m},
$$
with
$$
L \;=\; {\cal R} \;+\; \gamma_1\,X\,{\cal R} \;+\; \gamma_2\,A^\mu A^\nu {\cal R}_{\mu\nu} \;-\;\frac{1}{4}F^2 \;-\; V(X).
$$
Here the metric signature is $(-,+,+,+)$, the Einstein tensor is $G_{\mu\nu}={\cal R}_{\mu\nu}-\tfrac12{\cal R}g_{\mu\nu}$, the field strength is $F_{\mu\nu}=2\nabla_{[\mu}A_{\nu]}$, and $F^2\equiv F_{\mu\nu}F^{\mu\nu}$. The invariant $X$ is
$$
X\equiv A^2\equiv g^{\mu\nu}A_\mu A_\nu.
$$
The two dimensionless nonminimal couplings are $\gamma_1$ for $X{\cal R}$ and $\gamma_2$ for $A^\mu A^\nu{\cal R}_{\mu\nu}$. Matter is minimally coupled through $S_{\rm m}[g_{\mu\nu},\Psi]$.

The potential $V(X)$ is chosen so that its zero-energy minimum occurs at a nonzero value of $X$. This is the bumblebee-type ingredient that produces a vacuum with $X=b^2\neq 0$. In the terminology of the paper, the observable departures from general relativity are controlled by the combinations
$$
\ell_1=\gamma_1 b^2,\qquad \ell_2=\gamma_2 b^2.
$$
This split keeps track of which curvature contraction is responsible for a given deviation.

Varying the action yields the modified Einstein equations and the vector equation of motion,
$$
\begin{aligned}
E_{\mu\nu}\equiv\;& G_{\mu\nu}
+\gamma_1\Big[ X\,G_{\mu\nu} + {\cal R}\,A_\mu A_\nu + \big(g_{\mu\nu}\Box - \nabla_\mu\nabla_\nu\big)X\Big] \\
&+\gamma_2\Big[ 2{\cal R}^{\rho}{}_{(\mu}A_{\nu)}A_\rho -\frac{1}{2}g_{\mu\nu}{\cal R}_{\rho\sigma}A^\rho A^\sigma +\frac{1}{2}g_{\mu\nu}\nabla_\rho\nabla_\sigma\big(A^\rho A^\sigma\big) +\frac{1}{2}\Box\big(A_\mu A_\nu\big) -\nabla_\rho\nabla_{(\mu}\big(A_{\nu)}A^\rho\big)\Big] \\
&-\frac{1}{2}\left(F_{\mu\lambda}F_{\nu}{}^{\lambda} - \frac{1}{4}g_{\mu\nu}F^2\right)
+\frac{1}{2}g_{\mu\nu}V - V_X\,A_\mu A_\nu
\;=\; 8\pi\,T_{\mu\nu},
\end{aligned}
$$
and
$$
E_A{}^\nu \;\equiv\; \frac{1}{2}\,\nabla_\mu F^{\mu\nu} + \gamma_1\,{\cal R}\,A^\nu + \gamma_2\,{\cal R}^{\mu\nu}A_\mu - V_X\,A^\nu \;=\; 0,
$$
with
$$
V_X\equiv \frac{dV}{dX}.
$$
The vector equation is therefore a modified Maxwell equation with effective sources from curvature and from the potential gradient.

## 2. Asymptotic vacuum and sector selection

For isolated compact objects, the theory is studied in the static, spherically symmetric sector with
$$
ds^2 = -h(r)\,dt^2 + \frac{dr^2}{f(r)} + r^2 d\Omega^2,\qquad A_{(1)} = b\,\phi(r)\,dr,\qquad X=b^2\,\phi(r)^2\,f(r),
$$
with constant $b\neq 0$ [2605.14711]. The vacuum equations are then expanded near spatial infinity as
$$
h(r)=\sum_{i=0}^\infty \hat h_i\,r^{-i},\quad f(r)=\sum_{i=0}^\infty \hat f_i\,r^{-i},\quad \phi(r)=\sum_{i=0}^\infty \hat\phi_i\,r^{-i}.
$$

At leading order, the asymptotic value $\hat X_0=b^2\hat\phi_0^2\hat f_0$ must satisfy
$$
V(\hat X_0)=0,\qquad V_X(\hat X_0)=0.
$$
The paper focuses on the branch for which the vanishing minimum occurs at finite nonzero $X=b^2$,
$$
X\big|_{r\to\infty} = b^2,\qquad V|_{X=b^2}=0,\qquad V_X|_{X=b^2}=0,
$$
which enforces
$$
\phi(r)\;\xrightarrow{r\to\infty}\; f(r)^{-1/2}.
$$

The key asymptotic result is that this nonzero-vector-vacuum branch is not compatible with generic nonzero values of both $\gamma_1$ and $\gamma_2$. Instead, the field equations select two single-coupling sectors:

| Sector | Coupling choice | Asymptotics |
|---|---|---|
| Case I | $\gamma_2=0$, $\ell_1=\gamma_1 b^2$ | $h=f=1-\frac{m}{r}$, $\phi=\left(1-\frac{m}{r}\right)^{-1/2}$ |
| Case II | $\gamma_1=0$, $\ell_2=\gamma_2 b^2$ | $h=1-\frac{m}{r}$, $f=\frac{1}{1+\ell_2}\left(1-\frac{m}{r}\right)$, $\phi=\left[\frac{1+\ell_2}{1-\frac{m}{r}}\right]^{1/2}$ |

In Case I, the asymptotics are asymptotically flat. In Case II, the solid angle is rescaled by $1/(1+\ell_2)$, giving the monopole-like, deficit-angle asymptotic associated with a global-monopole geometry. The paper therefore shows that a nonzero vector vacuum does not, by itself, imply monopole-like asymptotics; that structure is tied specifically to the Ricci-tensor coupling sector. A direct corollary is that the Einstein-tensor combination $A^\mu A^\nu G_{\mu\nu}$ is excluded on this branch, because it is a nontrivial linear combination of the two elementary couplings.

## 3. Black-hole solutions and conserved mass

When the vacuum branch condition $X=b^2$ is enforced everywhere outside the horizon, the asymptotic forms become exact black-hole geometries [2605.14711]. In Case I, the metric is exactly Schwarzschild in terms of the integration constant $m$, while the radial vector one-form remains nontrivial. This is a stealth configuration: the metric solves general relativity, but the vector field does not vanish. In Case II, the metric is Schwarzschild-like with a global-monopole-type solid-angle deficit, and this reproduces the monopole-like compact-object sector studied previously in the restricted $A^\mu A^\nu{\cal R}_{\mu\nu}$ theory.

A distinctive feature of the theory is that the physical mass is not simply $m/2$. Using the Wald covariant phase-space formalism, the Hamiltonian variation is
$$
\delta {\cal H} \;=\; \frac{1}{16\pi}\int_{\Omega_2}\big(\delta\boldsymbol{Q}_\xi - i_\xi \boldsymbol{\Theta}\big),
$$
and for the massive Hellings–Nordtvedt Lagrangian the computation gives
$$
\delta \boldsymbol{Q}_\xi - i_\xi \boldsymbol{\Theta} =
-\,4\pi\,r\,\sqrt{h}\,\sqrt{f}\,(1+\ell_1+\ell_2)\,\sin\theta\,\delta f\, d\theta\wedge d\varphi.
$$
The resulting Noether masses are
$$
M_1 = \frac{1}{2}(1+\ell_1)\,m
\qquad\text{in Case I},
$$
and
$$
M_2 = \frac{1}{2}\sqrt{1+\ell_2}\,m
\qquad\text{in Case II}.
$$
Hence the physical mass differs from the general-relativistic parameter $m/2$ by a coupling-dependent factor in both sectors.

Expressed in terms of the physical mass $M_1$, the Case I black hole takes the form
$$
\begin{aligned}
ds^2 &= -\Bigg[1-\frac{2M_1}{r(1+\ell_1)}\Bigg]dt^2
+ \Bigg[1-\frac{2M_1}{r(1+\ell_1)}\Bigg]^{-1}dr^2 + r^2 d\Omega^2,\\
A_{(1)} &= b\Bigg[1-\frac{2M_1}{r(1+\ell_1)}\Bigg]^{-1/2}dr,
\qquad \ell_1=\gamma_1 b^2,\quad \gamma_2=0.
\end{aligned}
$$
This reparametrization is essential in the weak-field regime, because observables must be compared using the physical mass rather than the integration constant.

## 4. Weak-field limit and Solar-System constraints

The weak-field interpretation of the $A^2{\cal R}$ sector differs from the appearance of the metric in terms of $m$ [2605.14711]. Although the line element is Schwarzschild-like when written with $m$, the physical mass is $M_1=\tfrac12(1+\ell_1)m$. The effective mass entering weak-field observables is therefore
$$
M_{\rm eff} \;=\; \frac{M_1}{1+\ell_1}.
$$
The paper emphasizes that this Noether-mass reparametrization is essential: Solar-System observables must be computed in terms of $M_{\rm eff}$ and compared to the physical mass $M_1$.

For the perihelion advance of Mercury, the relativistic correction scales as $M_{\rm eff}^2$, yielding
$$
\Delta\Phi_{\rm HN1} = \Delta\Phi_{\rm GR}\,(1+\ell_1)^2.
$$
Using $\Delta\Phi_{\rm GR}=42.9814''$ and $\Delta\Phi_{\rm obs}=(42.9794\pm 0.0030)''$ per century gives
$$
-1.2\times 10^{-5} \;\le\; \ell_1 \;\le\; 5.8\times 10^{-5}.
$$

For light deflection at the solar limb, the deflection angle scales as $M_{\rm eff}$,
$$
\Delta\Psi_{\rm HN1} = \Delta\Psi_{\rm GR}\,(1+\ell_1),
$$
and the VLBI constraint $\gamma_{\rm PPN}=0.99992\pm 0.00012$ yields
$$
-2.0\times 10^{-5} \;\le\; \ell_1 \;\le\; 1.0\times 10^{-4}.
$$

For the Shapiro time delay measured by Cassini, the excess time delay again scales as $M_{\rm eff}$,
$$
\Delta T_{\rm HN1} = \Delta T_{\rm GR}\,(1+\ell_1),
$$
and with $\gamma_{\rm PPN}=1+(2.1\pm 2.3)\times 10^{-5}$ one finds
$$
-2.2\times 10^{-5} \;\le\; \ell_1 \;\le\; 1.0\times 10^{-6}.
$$

Combining the three tests gives the representative bound
$$
-10^{-5}\;\lesssim\; \ell_1 \;\lesssim\; 10^{-6}.
$$
By contrast, in the Ricci-tensor sector the solid-angle deficit produces much tighter bounds, typically $|\ell_2|\lesssim 10^{-10}$, because the asymptotics are non-Euclidean. This sharp disparity is one of the main reasons the paper identifies the $A^2{\cal R}$ sector as the more useful branch for strong-field phenomenology compatible with current weak-field tests.

## 5. Neutron stars and slow rotation in the $A^2{\cal R}$ sector

For slowly rotating stars, the metric and matter content are taken to be
$$
ds^2 = -e^{\lambda(r)}dt^2 + f(r)^{-1}dr^2 + r^2(d\theta^2 + \sin^2\theta d\varphi^2) - 2\,\epsilon(\Omega - w(r))\,r^2 \sin^2\theta\, dt\,d\varphi,
$$
with perfect-fluid stress–energy
$$
T^{\mu\nu} = \big(\rho+p\big)u^\mu u^\nu + p g^{\mu\nu},\qquad
u^\mu=(u^t,0,0,\epsilon\Omega u^t),\qquad
u^\mu u_\mu=-1,
$$
and barotropic equation of state $p=P(\rho)$. The vector one-form remains radial, $A_{(1)}=b\,\phi(r)\,dr$. In the slow-rotation expansion, only $w(r)$ is first order; the static functions $(\lambda,f,\phi,\rho)$ are zeroth order [2605.14711].

The zeroth-order equations reduce to
$$
E^t{}_t=-8\pi\rho,\qquad
E^r{}_r=8\pi P(\rho),\qquad
E^\theta{}_\theta=E^\varphi{}_\varphi=8\pi P(\rho),\qquad
E^r{}_A=0,
$$
and can be rearranged into the coupled ODE system
$$
\lambda'' = F_1(r,\lambda',f,\phi,\rho),\quad
f' = F_2(r,\lambda',f,\phi,\rho),\quad
\phi' = F_3(r,\lambda',f,\phi,\rho),\quad
\rho' = F_4(r,\lambda',\rho).
$$
The first-order rotational equation is
$$
w'' + F_5(r,\lambda',f,\phi,\rho)\,w' + F_6(r,\lambda',f,\phi,\rho)\,w = 0.
$$

Regularity at the center is imposed through power-series expansions,
$$
\lambda(r) = \sum_i \lambda_i r^i,\; f(r)=\sum_i f_i r^i,\; \phi(r)=\sum_i \phi_i r^i,\; \rho(r)=\sum_i \rho_i r^i,\; w(r)=\sum_i w_i r^i,
$$
with input central density $\rho_0$. The leading coefficients include
$$
\lambda_2 = \frac{16\pi \ell_1 (2\rho_0+3P(\rho_0)) - V(X_0)}{18\,(1+\ell_1\phi_0^2)} - \frac{b^2 V_X(X_0)}{18\ell_1},\quad
f_0=1,\quad
w_2=\frac{8\pi w_0(\rho_0+P(\rho_0))}{5\,(1+\ell_1\phi_0^2)},\quad
X_0=b^2\phi_0^2.
$$
The condition $f_0=1$ is explicitly contrasted with the Ricci-tensor sector, where the central expansion already encodes a solid-angle deficit.

The stellar surface is defined by
$$
p(R)=0,
$$
and continuity across the surface is imposed for the fields and their first derivatives. In the exterior vacuum,
$$
\lambda''=\hat F_1(r,\lambda',f,\phi),\quad
f'=\hat F_2(r,\lambda',f,\phi),\quad
\phi'=\hat F_3(r,\lambda',f,\phi),
$$
while the slow-rotation function satisfies
$$
w'' + \hat F_4(r,\lambda',f,\phi)\,w' = 0.
$$
The asymptotic conditions are
$$
\lambda(r)\to \log\Big[1-\frac{2M_1}{r(1+\ell_1)}\Big],\quad
f(r)\to 1-\frac{2M_1}{r(1+\ell_1)},\quad
\phi(r)\to \Big[1-\frac{2M_1}{r(1+\ell_1)}\Big]^{-1/2},
$$
and
$$
w(r)\to \Omega - \frac{2J}{r^3},\qquad I\equiv \frac{J}{\Omega}.
$$

For numerics, the potential is taken to be
$$
V(X) = \alpha\,(\gamma_1^2+\gamma_2^2)\,(X-b^2)^2,
$$
and in Case I the parameter $b$ can be absorbed once $\ell_1=\gamma_1 b^2$ is used. The equations are solved as a boundary-value problem using a shooting method with the SLy EOS.

## 6. Phenomenology, comparison of sectors, and open issues

The neutron-star calculations in [2605.14711] use $\ell_1=10^{-6}$, motivated by the Solar-System bounds, together with two representative values of $\alpha$: $10^{-4}\alpha_\star$ and $10^{-2}\alpha_\star$. For central density $\rho_0=1.2\times 10^{15}\,{\rm g\,cm^{-3}}$ with the SLy EOS, the paper reports:
- $\alpha=10^{-4}\alpha_\star$: $(M,R,I)=(1.71\,M_\odot,\,8.12\,r_\star,\,47.87\,I_\star)$, with $R\simeq 11.9$ km;
- $\alpha=10^{-2}\alpha_\star$: $(M,R,I)=(1.66\,M_\odot,\,7.98\,r_\star,\,43.09\,I_\star)$, with $R\simeq 11.7$ km.

The characteristic scalings are given by
$$
M_\odot \sim r_\star \sim p_\star^{-1/2}\sim \rho_\star^{-1/2}\sim \alpha_\star^{-1/2}\sim I_\star^{1/3},
$$
with
$$
r_\star=\frac{GM_\odot}{c^2}=1.47\times 10^5{\rm cm},\;
\rho_\star=\frac{c^6}{G^3 M_\odot^2}=6.18\times 10^{17}{\rm g\,cm^{-3}},\;
p_\star=\frac{c^8}{G^3 M_\odot^2}=5.55\times 10^{38}{\rm g\,cm^{-1}\,s^{-2}},\;
I_\star=\frac{G^2 M_\odot^3}{c^4}=4.34\times 10^{43}{\rm g\,cm^2}.
$$

The mass–radius relations in the $A^2{\cal R}$ sector show noticeable departures from general relativity despite the small coupling. At low central densities, both mass and radius are reduced relative to general relativity; at high central densities, both become larger than in general relativity, and the maximum mass occurs at a lower $\rho_0$. The moment-of-inertia relations mirror this pattern: relative to general relativity, $I$ is smaller for low-mass stars and larger for high-mass stars. Compared with the Ricci-tensor sector at $\ell_2=10^{-10}$, deviations are similar at low densities but diverge at high densities, where mass grows faster with $\rho_0$ in the $A^2{\cal R}$ sector. The representative parameter choices are described as broadly consistent with the massive pulsar measurement $M=2.01\pm 0.04\,M_\odot$ and the GW170817 radius estimate $R_{1.4}=11.0^{+0.9}_{-0.6}$ km.

The theoretical interpretation of these results rests on three points. First, the nonzero vacuum $X=b^2$ breaks local Lorentz symmetry spontaneously while preserving diffeomorphism invariance. Second, the asymptotic vacuum requirement with $V=V_X=0$ at $X=b^2$ forbids having both $\gamma_1$ and $\gamma_2$ nonzero on this branch. Third, the paper does not perform a full perturbative stability analysis. It notes that related vector–tensor theories can exhibit both stable regions and pathologies, including ghosts, gradient instabilities, and loss of hyperbolicity, and it identifies stability of the compact-object solutions as an urgent direction for future work.

A plausible implication is that massive Hellings–Nordtvedt theory should be viewed less as a single phenomenological branch than as two asymptotically selected sectors with different observational logic. The $A^\mu A^\nu{\cal R}_{\mu\nu}$ sector retains the monopole-like asymptotics but is tightly constrained in the weak field. The $A^2{\cal R}$ sector, by contrast, admits asymptotically flat Schwarzschild geometries with a nontrivial radial vector and remains compatible with current weak-field tests after the Noether-mass redefinition. Within the scope of the paper, this identifies the $A^2{\cal R}$ sector as a viable framework for studying strong-field compact objects with a nonzero vector vacuum [2605.14711].

Source: https://www.emergentmind.com/topics/hellings-nordtvedt-theory