---
title: Centre-Normalised Outgoing Null-Geodesic Gauge
url: https://www.emergentmind.com/topics/centre-normalised-outgoing-null-geodesic-gauge
type: topic
---

# Centre-Normalised Outgoing Null-Geodesic Gauge

Centre-normalised outgoing null-geodesic gauge is a centre-normalised version of the classical Newman–Unti gauge in which one fixes a timelike geodesic \(\Gamma\), foliates spacetime by outgoing null cones from \(\Gamma\), and uses an affine parameter along each outgoing null generator with its origin at the centre. In the formulation used for the global stability of Minkowski spacetime, the coordinates are \((u,r,\theta^1,\theta^2)\), the null cones are \(H_u\), and the metric is written
\[
g = - (d u \otimes d r+d r\otimes d u) - f d u \otimes d u +\gamma_{AB}(d\theta^A-b^A d u)\otimes (d\theta^B - b^B d u).
\]
The associated outgoing frame vector is \(e_4=\partial_r\), which is null and geodesic, so the gauge is simultaneously centre-normalised, outgoing, and null-geodesic-adapted [2606.31090].

## 1. Definition as a centre-normalised Newman–Unti gauge

In the relevant recent usage, the Newman–Unti gauge is defined by coordinates
\[
(u,r,\theta^1,\theta^2),
\]
with \(u\) a null coordinate and \(r\) chosen so that \(\partial_r\) is null and geodesic. The inverse metric is
\[
\begin{split}
g^{-1} = f \frac{\partial}{\partial r}\otimes \frac{\partial}{\partial r} - \left (\frac{\partial}{\partial u} + b^A\frac{\partial}{\partial\theta^A}\right )\otimes \frac{\partial}{\partial r}- \frac{\partial}{\partial r} \otimes \left(\frac{\partial}{\partial u} + b^A\frac{\partial}{\partial\theta^A}\right)+\gamma^{-1},
\end{split}
\]
and the associated null frame is
\[
e_4:= \frac{\partial}{\partial r}, \quad e_3:= 2 \left (\frac{\partial}{\partial u}  +b^A\frac{\partial}{\partial\theta^A}\right ) - f \frac{\partial}{\partial r}, \quad e_A:= \frac{\partial}{\partial \theta^A}.
\]
These satisfy
\[
g(e_4,e_4)=0,\qquad g(e_3,e_3)=0,\qquad g(e_3,e_4)=-2,\qquad g(e_4,e_A)=g(e_3,e_A)=0,
\]
and
\[
D_{e_4}e_4=0.
\]
Geometrically, the level sets \(H_u\) are outgoing null hypersurfaces, and \(e_4=\partial_r\) is tangent to their outgoing generators [2606.31090].

The centre-normalised version fixes a timelike geodesic \(\Gamma\), called the centre or spacetime axis, and declares that for each \(u\), \(H_u\) is the null cone generated by all future outgoing null geodesics from the point \(\Gamma(u)\). For fixed \((u,\theta)\), the curve
\[
r\mapsto (u,r,\theta)
\]
is the corresponding outgoing null generator, and \(r\) is chosen as an affine parameter on that generator. The normalization is that each cone starts at the centre and
\[
r=0
\]
at the centre. In Minkowski space this reduces to the standard retarded coordinates
\[
u=t-r,\qquad \Gamma=\{r=0\},
\]
with
\[
f=1,\qquad b^A=0,\qquad \gamma=\mathring{\gamma},
\]
where \(\mathring{\gamma}\) is the round metric of radius \(r\) [2606.31090].

A basic structural distinction from Bondi gauge is explicit: in Newman–Unti gauge, \(r\) is an affine parameter along outgoing null generators, whereas in Bondi/BMS gauge, \(r\) is an areal radius. Accordingly, in the centre-normalised outgoing null-geodesic gauge, \(r\) is not areal radius [2606.31090].

## 2. Centre normalisation, regularity at the axis, and fixing of affine freedom

The centre is the fixed timelike geodesic \(\Gamma\), parameterised by proper time \(u\). The coordinate construction starts from a marked point \(O\) on the initial Cauchy hypersurface and takes \(\Gamma\) to be the future timelike geodesic orthogonal to the initial surface at \(O\). For each \(u\), one chooses null initial directions \(V(u,\theta)\) along \(\Gamma\), obtained by parallel transport of the initial null directions \(N+\iota_*R(\theta)\), and exponentiates them. This yields the coordinate map and identifies the centre \(r=0\) as the blown-up boundary corresponding to the timelike axis [2606.31090].

The centre-normalisation imposes explicit regularity conditions at \(r=0\):
\[
f = 1 + O(r^2), \qquad b^A = O(r), \qquad \gamma_{AB} = \mathring{\gamma}_{AB} + O(r^4),
\]
and, earlier in the construction,
\[
g_{uu}=1+O(r^2),\qquad g_{uA}=O(r^3),\qquad g_{AB}=\mathring{\gamma}_{AB}+O(r^4).
\]
Thus \(f|_{r=0}=1\), \(b|_{r=0}=0\), and \(\gamma\) agrees to leading order with the round metric induced by the null cone structure from a regular point [2606.31090].

This regularity removes the residual coordinate freedom that would otherwise remain in Newman–Unti gauge. The paper makes the point explicitly: before centre normalisation, Newman–Unti gauge still permits residual changes preserving the null nature of \(u\), the null-geodesic character of \(\partial_r\), and the metric form. Centre normalisation removes this by fixing the centre geodesic \(\Gamma\), declaring \(H_u\) to be the cone from \(\Gamma(u)\), fixing \(r=0\) at the centre, making \(r\) affine along generators, and transporting angular coordinates from the centre [2606.31090].

The underlying affine ambiguity is consistent with the broader literature on null parametrization. Affine parameters for null geodesics are unique only up to
\[
\lambda \to a\lambda + b,
\]
so any construction that sets an affine origin and scale is imposing extra geometric data [2211.07835]. In the centre-normalised outgoing null-geodesic gauge, the translational freedom is removed by requiring \(r=0\) at the centre, and the multiplicative freedom is removed by affine normalisation from the centre itself [2606.31090].

A related but distinct normalization issue appears in the theory of null expansions. If null normals are rescaled by
\[
\ell^\mu \to \kappa(x)\,\ell^\mu,\qquad n^\mu \to \kappa(x)^{-1}\,n^\mu,
\]
then
\[
\theta_\ell \to \kappa\,\theta_\ell,\qquad \theta_n \to \kappa^{-1}\,\theta_n,
\]
while \(\theta_\ell\theta_n\) is invariant [2105.07521]. This suggests that centre normalisation should be understood as a fixing of otherwise available local null-normal scaling freedom, rather than as a normalization-independent invariant.

## 3. Null frame, connection coefficients, and the transport hierarchy

The null geometry in this gauge is encoded by the standard Ricci coefficients
\[
\begin{aligned}
\chi_{AB}=&\:g(D_Ae_4,e_B), \quad \quad \chib_{AB}=g(D_Ae_3,e_B),\\
\omega=&\:\frac{1}{4} g(D_4 e_4,e_3),\quad \quad  \omegab=\frac{1}{4} g(D_3 e_3,e_4),\\
\eta_A=&\:\frac{1}{2}g(D_3e_4,e_A),  \quad \quad \etab_A=\frac{1}{2}g(D_4e_3,e_A),\\
\xi_A=&\:\frac{1}{2}g(D_4e_4,e_A), \quad \quad  \xib_A=\frac{1}{2}g(D_3e_3,e_A), \\
\zeta_A=&\: \frac{1}{2}g(D_Ae_4,e_3),
\end{aligned}
\]
together with
\[
\chih_{AB}:=\chi_{AB}-\frac{1}{2}\gamma_{AB}\chi,\qquad
\chibh_{AB}:=\chib_{AB}-\frac{1}{2}\gamma_{AB}\chib.
\]
In this gauge they are tied to the metric coefficients by especially simple identities:
\[
\omegab=\frac12 \frac{\partial f }{\partial r},\qquad   \etab^A-\eta^A = \frac{\partial b^A}{\partial r},
\]
\[
\omega= 0, \qquad \xi_A = 0,\ \zeta_A=\eta_A=-\etab_A,\qquad \xib_A = \frac{ \partial f}{\partial\theta^A},
\]
\[
2\chi_{AB} = \frac{\partial \gamma_{AB}}{\partial r},
\]
\[
2\chib_{AB} = \left(2\frac{\partial }{\partial u} + 2 b^A\frac{\partial }{\partial \theta^A} - f\frac{\partial }{\partial r}\right)\gamma_{AB} + 2 \frac{\partial b^C}{\partial \theta^A}\gamma_{CB} + 2 \frac{\partial b^C}{\partial \theta^B}\gamma_{AC}.
\]
Thus \(\chi\) is literally the \(r\)-derivative of \(\gamma\), and \(\omegab\) is the \(r\)-derivative of \(f\) [2606.31090].

The Weyl curvature components are
\[
\alpha_{AB}= W_{A4B4}, \qquad \beta_A=\frac{1}{2}W_{A434}, \qquad \rho = \frac{1}{4}W_{4343},
\]
\[
\alphab_{AB}=W_{A3B3},\qquad \betab_A=\frac{1}{2}W_{A334}, \qquad \sigma=\frac{1}{4}(^*W)_{4343}.
\]
A central structural fact is that \(\alpha\) satisfies a decoupled tensorial wave equation, whereas the remaining curvature components and Ricci coefficients are then recovered by transport along \(e_4\), supplemented by elliptic sphere equations. Representative transport equations include
\[
\nab_4 \chih+  \chi \chih=\: - \alpha,
\]
\[
\nab_4\beta+2 \chi \beta- \alpha=\: \eta \cdot \alpha,
\]
\[
\nab_4(-\rho,\sigma)+\frac{3}{2} \chi (-\rho,\sigma)+( \beta,- \beta)=\cdots,
\]
\[
\nab_4 \omegab = 3|\eta|^2+\rho.
\]
The resulting hierarchy is
\[
\alpha \to (\chi,\hat\chi,\beta) \to (\gamma,\eta,\rho,\sigma,K)\to(\chib,\omegab,\betab,b)\to(f,\xib).
\]
The regular centre supplies canonical initial conditions for this hierarchy:
\[
f|_{r=0}=1,\qquad b|_{r=0}=0,
\]
and correspondingly
\[
\gamma_{AB}=\mathring{\gamma}_{AB}+O(r^4),\qquad g_{uA}=O(r^3),\qquad g_{uu}=1+O(r^2).
\]
All the \(e_4\)-transport equations are integrated from \(r=0\) [2606.31090].

## 4. Teukolsky analysis and the stability of Minkowski spacetime

The gauge becomes analytically powerful because the Einstein vacuum equations acquire a rigid hierarchy centered on the outgoing Weyl component \(\alpha\). The key wave equation is the tensorial Teukolsky equation
\[
\nab_3(r \nab_4(r^2\alpha)) + 2 \nab_4 (r^2 \alpha) + 3 \nab_3 (r^2 \alpha) + \frac{4}{r} (r^2 \alpha) -  r \slashed\Delta (r^2 \alpha) = \mathcal{E},
\]
with nonlinear perturbative error \(\mathcal E\). The analysis uses a Dafermos–Rodnianski \(r^p\)-multiplier adapted to this operator, essentially
\[
w(r,u)\,(r\nab_4\phi+3\phi),
\qquad
w(u, r) = \frac{(r+u)^{2p_{0}+2p_{\infty}}}{r^{3+2p_{0}}},
\]
interpolating between a near-centre weight and an asymptotic weight [2606.31090].

The proof scheme is explicitly hierarchical. First one estimates \(\alpha\) by weighted \(r^p\) estimates for \(r^2\alpha\). Next one upgrades \(\beta\) using Bianchi equations and elliptic estimates on spheres, avoiding derivative loss. Then one recovers all remaining quantities by \(e_4\)-transport from the centre. In this way, the centre-normalised outgoing null-geodesic gauge turns the global nonlinear problem into a mixed wave–transport system with canonical centre data [2606.31090].

The main theorem proves small-data global stability of Minkowski spacetime in this gauge. Under small asymptotically flat initial data, the maximal future development is globally smooth, future complete, and covered by a single centre-normalised Newman–Unti chart away from the axis. The theorem is stated to work for initial data which decay only weakly to flat space, and under stronger asymptotic structure one obtains additional asymptotic control, including almost sharp Bondi–Sachs peeling [2606.31090].

The role of the centre is not merely coordinate-theoretic. Near the axis, the Newman–Unti chart is singular enough that \(\alpha\) itself is not smooth there, so the argument uses weighted control near \(r=0\), with near-centre weights taking
\[
p_0=1+\delta.
\]
This yields estimates sufficient to recover smoothness in a different chart and to propagate local regularity near the centre. A plausible implication is that the centre-normalised formulation is not only a gauge choice but also a regularity mechanism adapted to transport from a regular timelike axis [2606.31090].

## 5. Relation to Bondi, double-null, affine-null, and light-cone gauges

The immediate comparison is with Bondi/BMS gauge. The two gauges differ only by radial normalisation: in Newman–Unti gauge, \(r\) is an affine parameter along outgoing null generators; in Bondi/BMS gauge, \(r\) is an areal radius. The centre-normalised outgoing null-geodesic gauge therefore shares the Newman–Unti one-null-foliation structure and should not be identified with areal-radius gauges [2606.31090].

A closely related spherical characteristic formalism is the metric ansatz
\[
ds^2=-2G\,du(dx+B\,du)+R^2\,d\Omega^2,
\]
for which \(u=\) const are outgoing null hypersurfaces and the affinely parametrised generators have tangent vector
\[
U^a:=-\nabla^a u=G^{-1}(\partial_x)^a.
\]
The preferred affine parameter is
\[
\frac{d}{d\lambda}:=G^{-1}\partial_x,
\qquad
\lambda(u,x)=\int_0^x G(u,x')\,dx'.
\]
In the regular-centre setting the centre is fixed by
\[
R(u,0)=0,
\]
with \(u\) taken as proper time at the centre and the exact centre conditions
\[
G(u,0)=R_{,x}(u,0), \qquad
B(u,0)=\frac{1}{2R_{,x}(u,0)}, \qquad
\Xi R(u,0)=-\frac12, \qquad
{\cal H}(u,0)=0.
\]
This is not the same formulation as the centre-normalised Newman–Unti gauge, but it is a very close spherical analogue of a centre-normalised outgoing null-geodesic coordinate system [2511.14874].

By contrast, long-time double-null simulations in spherical symmetry typically employ
\[
ds^{2}=-e^{\sigma(u,v)}dudv+r(u,v)^{2}d\Omega^{2},
\]
with residual gauge freedom
\[
u\to u'(u),\qquad v\to v'(v),
\]
and the standard initial choice
\[
\sigma(u,v_{0})=0,\qquad \sigma(u_{0},v)=0.
\]
That standard choice is explicitly identified as an affine gauge for both \(u\) and \(v\). The later adaptive-gauge constructions in this framework show that affine normalization alone is not enough for numerical control near horizons; where and how the null coordinates are normalised is decisive [1510.05273].

Light-cone gauges in cosmology have a different orientation and normalization logic. In Geodesic Light-Cone variables
\[
x^\mu=(\tau,w,\tilde\theta^a),
\]
the metric is
\[
ds^2_{\rm GLC} =\Upsilon^2dw^2-2\Upsilon\,dw\,d\tau +\gamma_{ab}(d\tilde\theta^a-U^a dw)(d\tilde\theta^b-U^b dw),
\]
with \(w\) null, photons traveling at constant \(w,\tilde\theta^a\), and \(\tau\) equal to the proper time of geodesic observers. The applications are primarily to an observer’s past light cone rather than to future outgoing cones. The residual condition
\[
w_o=\frac{\xi^0_o}{a_o}
\]
is interpreted as requiring that the origin of the polar coordinates coincide with the observer’s position. This is a centre-normalisation analogue, but not an affine outgoing-null normalisation [2009.14134].

## 6. Scope, nearby analogues, and common misunderstandings

A recurring misunderstanding is to treat “centre-normalised” as if it referred to a universally invariant normalization of null directions. The broader null-expansion literature shows the opposite: with
\[
\ell^\mu \to \kappa(x)\,\ell^\mu,\qquad n^\mu \to \kappa(x)^{-1}\,n^\mu,
\]
the individual expansions rescale, while the product \(\theta_\ell\theta_n\) and the marginal condition \(\theta_\ell=0\) remain invariant [2105.07521]. The centre-normalised outgoing null-geodesic gauge is therefore a gauge fixing of null-normal freedom, not an elimination of normalization dependence from all null-geometric quantities.

A second misunderstanding is to identify the gauge with any null coordinate system. The recent stability work is asymmetric: it is a one-null-foliation gauge built around the outgoing generators \(e_4=\partial_r\), not a double-null gauge. Likewise it is not a Bondi areal-radius gauge, and it is not a GLC observer-past-cone gauge [2606.31090].

A useful local model is given by locally inertial null normal coordinates. There one chooses a point \(p\), a null basis \(\{l^a,k^a,e_A^a\}\) with
\[
k_a k^a = l_a l^a = k_a e_A^a = l_a e_A^a = 0,\qquad l_a k^a = -1,
\]
constructs a codimension-2 spacelike surface
\[
\Sigma:\quad u=v=0,
\]
and then extends away from \(\Sigma\) along the geodesics normal to it using null coefficients \(U,V\). The distinguished null geodesic \(\Gamma\) tangent to \(k^a\) is the coordinate curve \((U,x^A)=(0,0)\), and \(V\) serves as affine parameter along that generator. This is a local codimension-2 analogue of a centre-normalised outgoing null-geodesic gauge, though it treats the two null directions symmetrically [1201.0542].

Near horizons, related constructions often cease to be ordinary coordinate-gauge choices and become dynamical or microlocal descriptions of null-geodesic flow. In the \(b\)-geometric treatment of nondegenerate Killing horizons, compactification at future infinity and the use of \(b\)-geometry replace explicit horizon-penetrating coordinates by a boundary-adapted Hamiltonian framework, and the surface gravity \(\varkappa\) appears as the linearized exponential rate at which null bicharacteristics are attracted to or repelled from the radial sets over the horizon [1701.00555]. This is not a centre-normalised outgoing null-geodesic gauge, but it clarifies that null-geodesic adaptation can be organized either by explicit coordinate choice or by invariant phase-space dynamics.

Within general relativity as treated in current arXiv work, the phrase therefore denotes a specific and comparatively rigid construction: choose a regular timelike centre, emit future outgoing null cones from it, use affine parameter \(r\) on each generator with \(r=0\) at the centre, and write the metric in Newman–Unti form. Its analytical significance lies in the fact that once the outgoing Weyl component \(\alpha\) is controlled by a Teukolsky \(r^p\) estimate, the remaining geometry is recovered by transport from the regular axis [2606.31090].

Source: https://www.emergentmind.com/topics/centre-normalised-outgoing-null-geodesic-gauge