Newman–Janis Algorithm (NJA) Overview

• The Newman–Janis Algorithm (NJA) is a method that uses controlled complexification of static, spherically symmetric metrics to generate rotating black hole solutions.
• It employs a systematic transformation—from static seeds to null tetrad forms and Boyer–Lindquist coordinates—to include rotation and address coordinate ambiguities.
• The algorithm underpins canonical derivations such as Kerr and Kerr–Newman metrics and has been extended to incorporate gauge fields and cosmological terms.

Searching arXiv for recent and foundational papers on the Newman–Janis algorithm. The Newman–Janis algorithm (NJA), also presented in the broader Janis–Newman or Janis–Newman/Giampieri form, is a solution-generating prescription in classical gravity that starts from a static, spherically symmetric seed and produces a stationary, axisymmetric spacetime through a controlled complexification of coordinates and radial functions, usually in a Newman–Penrose null-tetrad framework or an equivalent coordinate-based formulation. Its canonical successes are the derivations of Kerr from Schwarzschild and Kerr–Newman from Reissner–Nordström, but the method is simultaneously a practical computational device, a locus of ambiguity in the complexification step, and, in recent work, an object of geometric reinterpretation rather than an ad hoc complex trick (Erbin, 2016, Gutierrez-Chavez et al., 2014, Kim, 2024).

1. Canonical algorithmic structure

In its standard four-dimensional form, NJA begins with a static seed metric such as

$ds^2=-f(r)\,dt^2+f(r)^{-1}\,dr^2+r^2 d\Omega^2,$

or, more generally,

$$ds^2 = g_{tt}\,dt^2 - g_{rr}\,dr^2 - g_{\theta\theta}\,d\theta^2 - g_{\phi\phi}\,d\phi^2.$$

The first step is a passage to null coordinates. In one common convention,

$du = dt - f^{-1}dr,$

while a generalized advanced Eddington–Finkelstein transformation is written as

$dv = dt + \sqrt{\frac{g_{rr}}{g_{tt}}}\,dr.$

The metric is then rewritten in Newman–Penrose form using a null tetrad $(l^\mu,n^\mu,m^\mu,\bar m^\mu)$ satisfying

$$g^{\mu\nu}= l^\mu n^\nu + n^\mu l^\nu - m^\mu \bar m^\nu - \bar m^\mu m^\nu$$

or, in another sign convention,

$$g^{\mu\nu}=-\ell^\mu n^\nu-\ell^\nu n^\mu+m^\mu\bar m^\nu+m^\nu\bar m^\mu.$$

A representative seed tetrad is

$$\ell^\mu=\delta_r^\mu,\qquad n^\mu=\delta_u^\mu-\frac{f}{2}\delta_r^\mu,\qquad m^\mu=\frac{1}{\sqrt2\,\bar r}\left(\delta_\theta^\mu+\frac{i}{\sin\theta}\delta_\phi^\mu\right)$$

(Gutierrez-Chavez et al., 2014, Erbin, 2014, Erbin, 2016).

The characteristic step is a complex coordinate transformation in the $(u,r)$- or $(v,r)$-plane. In one standard retarded-null convention,

$$ds^2 = g_{tt}\,dt^2 - g_{rr}\,dr^2 - g_{\theta\theta}\,d\theta^2 - g_{\phi\phi}\,d\phi^2.$$0

while equivalent sign choices appear in advanced-coordinate formulations. After transforming the tetrad, one reconstructs the inverse metric, inverts it to obtain the covariant metric, and then attempts a Boyer–Lindquist-type transformation to eliminate $$ds^2 = g_{tt}\,dt^2 - g_{rr}\,dr^2 - g_{\theta\theta}\,d\theta^2 - g_{\phi\phi}\,d\phi^2.$$1 and $$ds^2 = g_{tt}\,dt^2 - g_{rr}\,dr^2 - g_{\theta\theta}\,d\theta^2 - g_{\phi\phi}\,d\phi^2.$$2 terms. In this operational sense, NJA is intended to “include rotation into nonrotating solutions of the Einstein field equations with spherical symmetry or perturbed spherical symmetry” (Gutierrez-Chavez et al., 2014, Erbin, 2016).

The historical importance of this construction lies in its repeated successful use in general relativity. The standard exemplars are Schwarzschild $$ds^2 = g_{tt}\,dt^2 - g_{rr}\,dr^2 - g_{\theta\theta}\,d\theta^2 - g_{\phi\phi}\,d\phi^2.$$3 Kerr and Reissner–Nordström $$ds^2 = g_{tt}\,dt^2 - g_{rr}\,dr^2 - g_{\theta\theta}\,d\theta^2 - g_{\phi\phi}\,d\phi^2.$$4 Kerr–Newman; related NJA-type constructions also produce Kerr–NUT- or Demiański–Newman-like metrics from suitable spherical seeds (Gutierrez-Chavez et al., 2014, Hansen et al., 2013).

2. Complexification rules, integrability, and algorithmic ambiguity

The central technical ambiguity of NJA is the complexification of radial functions. The standard empirical rules are

$$ds^2 = g_{tt}\,dt^2 - g_{rr}\,dr^2 - g_{\theta\theta}\,d\theta^2 - g_{\phi\phi}\,d\phi^2.$$5

For the Reissner–Nordström seed, this leads to a complexified lapse of the form

$$ds^2 = g_{tt}\,dt^2 - g_{rr}\,dr^2 - g_{\theta\theta}\,d\theta^2 - g_{\phi\phi}\,d\phi^2.$$6

which already exhibits the non-unique treatment of $$ds^2 = g_{tt}\,dt^2 - g_{rr}\,dr^2 - g_{\theta\theta}\,d\theta^2 - g_{\phi\phi}\,d\phi^2.$$7 and $$ds^2 = g_{tt}\,dt^2 - g_{rr}\,dr^2 - g_{\theta\theta}\,d\theta^2 - g_{\phi\phi}\,d\phi^2.$$8 terms. This arbitrariness is one reason several authors describe NJA as not a uniquely defined algorithm (Hansen et al., 2013, Erbin, 2014).

A major refinement is a unified prescription for powers of $$ds^2 = g_{tt}\,dt^2 - g_{rr}\,dr^2 - g_{\theta\theta}\,d\theta^2 - g_{\phi\phi}\,d\phi^2.$$9, proposed to subsume the familiar $du = dt - f^{-1}dr,$0 and $du = dt - f^{-1}dr,$1 replacements and extend them to $du = dt - f^{-1}dr,$2, which is needed for de Sitter and anti–de Sitter seeds. In that formulation, the usual replacements

$du = dt - f^{-1}dr,$3

are treated as special cases of a single rule, and $du = dt - f^{-1}dr,$4 is sent to $du = dt - f^{-1}dr,$5. This extension was used to produce Kerr-(A)dS and Kerr–Newman-(A)dS metrics within an NJA framework (Urreta et al., 2015).

A second technical restriction is integrability of the Boyer–Lindquist step. In several general NJA constructions, the coordinate shifts required to remove $du = dt - f^{-1}dr,$6 and $du = dt - f^{-1}dr,$7 terms are legitimate only if the relevant combinations depend on $du = dt - f^{-1}dr,$8 alone. In the notation of the general rotating metric literature, $du = dt - f^{-1}dr,$9 and $dv = dt + \sqrt{\frac{g_{rr}}{g_{tt}}}\,dr.$0 must be $dv = dt + \sqrt{\frac{g_{rr}}{g_{tt}}}\,dr.$1-only functions; otherwise there is no global Boyer–Lindquist-type chart of the desired form. This restriction is not formal bookkeeping: it determines whether the NJA output is a coherent stationary axisymmetric spacetime or only a partially transformed ansatz (Shaikh, 2019, Chen et al., 2019, Erbin, 2014).

The same ambiguity appears in symbolic implementations. A REDUCE program, Newman-Janis.red, automates the null-coordinate conversion, tetrad construction, metric reconstruction, and Boyer–Lindquist transformation, but explicitly leaves the complexification step to the user. Its applicability is restricted to metrics containing potentials of the form $dv = dt + \sqrt{\frac{g_{rr}}{g_{tt}}}\,dr.$2, and the authors warn against use for metrics with cosmological constant or without spherical symmetry (Gutierrez-Chavez et al., 2014).

3. Radial extension, principal-null structure, and recent geometric reinterpretation

One line of analysis emphasizes that NJA does more than “spin up” a seed metric. In a detailed re-examination of the Reissner–Nordström $dv = dt + \sqrt{\frac{g_{rr}}{g_{tt}}}\,dr.$3 Kerr–Newman derivation, the complexification

$dv = dt + \sqrt{\frac{g_{rr}}{g_{tt}}}\,dr.$4

was argued to promote the original nonnegative radial variable to a Cartesian-like coordinate

$dv = dt + \sqrt{\frac{g_{rr}}{g_{tt}}}\,dr.$5

thereby introducing the negative-$dv = dt + \sqrt{\frac{g_{rr}}{g_{tt}}}\,dr.$6 branch directly within the NJA construction. In Boyer–Lindquist form, the corresponding Kerr–Newman metric has

$dv = dt + \sqrt{\frac{g_{rr}}{g_{tt}}}\,dr.$7

Within that interpretation, the $dv = dt + \sqrt{\frac{g_{rr}}{g_{tt}}}\,dr.$8 and $dv = dt + \sqrt{\frac{g_{rr}}{g_{tt}}}\,dr.$9 sectors are two asymptotically flat regions glued across the disk bounded by the ring singularity; horizons and ergosurfaces occur only for $(l^\mu,n^\mu,m^\mu,\bar m^\mu)$0 (Brauer et al., 2014).

A distinct geometric development supplements the standard NJA by a null rotation after the complex coordinate transformation. The purpose is to align the transformed tetrad with the repeated principal null directions of Kerr–Newman. In that extended algorithm, the null rotation is not decorative: it allows the metric, Maxwell tensor, Ricci tensor, Weyl tensor, Carter Killing tensor, and conformal Killing tensor to be generated coherently from their Reissner–Nordström counterparts. The same analysis shows that Schwarzschild $(l^\mu,n^\mu,m^\mu,\bar m^\mu)$1 Kerr is the uncharged special case of this larger principal-null-structure transport (Keane, 2014).

A more radical reinterpretation was proposed in 2024. There, the Kerr geometry in Kerr–Schild form is read as the nonlinear superposition of a self-dual and an anti-self-dual Taub–NUT instanton, interpreted as gravitational dyons of opposite chiralities. The ring singularity is written as

$(l^\mu,n^\mu,m^\mu,\bar m^\mu)$2

and under Wick rotation becomes

$(l^\mu,n^\mu,m^\mu,\bar m^\mu)$3

which collapses to the two isolated points

$(l^\mu,n^\mu,m^\mu,\bar m^\mu)$4

These two points are identified with the SD and ASD Taub–NUT instantons, joined by a finite Misner string. In this picture, Schwarzschild corresponds to coincident chiral constituents, while Kerr arises from separating them; the “complex shift” is then interpreted as the holomorphic bookkeeping of the two source positions rather than as a coordinate trick. The same framework is extended there to Kerr–Newman and Kerr–Taub–NUT (Kim, 2024).

4. Generalizations to gauge fields, NUT charge, cosmological terms, and higher dimensions

NJA has repeatedly been generalized beyond the original Schwarzschild/Kerr setting. For gauge fields, a streamlined Giampieri-based formulation shows that the electromagnetic potential can be transformed systematically rather than inserted by ansatz. Starting from Reissner–Nordström in a gauge where the radial component has been removed, the complexification

$(l^\mu,n^\mu,m^\mu,\bar m^\mu)$5

and the differential replacement

$(l^\mu,n^\mu,m^\mu,\bar m^\mu)$6

yield

$(l^\mu,n^\mu,m^\mu,\bar m^\mu)$7

the standard Kerr–Newman potential (Erbin, 2014).

The Demiański–Janis–Newman framework extends the algorithm to NUT charge, an extra parameter $(l^\mu,n^\mu,m^\mu,\bar m^\mu)$8, topological horizons, and electric charge. In that formalism, the complex transformation is written as

$(l^\mu,n^\mu,m^\mu,\bar m^\mu)$9

with $$g^{\mu\nu}= l^\mu n^\nu + n^\mu l^\nu - m^\mu \bar m^\nu - \bar m^\mu m^\nu$$0 for spherical horizons and $$g^{\mu\nu}= l^\mu n^\nu + n^\mu l^\nu - m^\mu \bar m^\nu - \bar m^\mu m^\nu$$1 for hyperbolic ones. The transformed gauge field takes the form

$$g^{\mu\nu}= l^\mu n^\nu + n^\mu l^\nu - m^\mu \bar m^\nu - \bar m^\mu m^\nu$$2

and the resulting Einstein–Maxwell analysis identifies $$g^{\mu\nu}= l^\mu n^\nu + n^\mu l^\nu - m^\mu \bar m^\nu - \bar m^\mu m^\nu$$3 as the NUT charge. For $$g^{\mu\nu}= l^\mu n^\nu + n^\mu l^\nu - m^\mu \bar m^\nu - \bar m^\mu m^\nu$$4, one obtains a charged topological Demiański/Kerr–NUT family; for $$g^{\mu\nu}= l^\mu n^\nu + n^\mu l^\nu - m^\mu \bar m^\nu - \bar m^\mu m^\nu$$5, the field equations within that framework force $$g^{\mu\nu}= l^\mu n^\nu + n^\mu l^\nu - m^\mu \bar m^\nu - \bar m^\mu m^\nu$$6, so only NUT charge survives (Erbin, 2014).

This restriction contrasts with other NJA-type constructions. The review literature presents an extended algorithm generating configurations with metric, gauge fields, real and complex scalars, and five of the six Plebański–Demiański parameters—mass, electric charge, magnetic charge, NUT charge, and angular momentum—while also emphasizing that the missing acceleration parameter remains outside the method in its standard form. The same review treats complex scalar fields as whole complex objects under the transformation, which is essential for deriving rotating axion–dilaton solutions (Erbin, 2016).

Higher-dimensional generalizations are possible but uneven. A five-dimensional extension using the Giampieri prescription and sequential rotations in the two orthogonal rotation planes generates Myers–Perry with two angular momenta and, after an extremal limit, BMPV. In that setting,

$$g^{\mu\nu}= l^\mu n^\nu + n^\mu l^\nu - m^\mu \bar m^\nu - \bar m^\mu m^\nu$$7

emerges after two successive rotations. The same work states that dimensions higher than six are much more challenging because the complexification of the seed functions does not generalize cleanly (Erbin et al., 2014).

5. Heuristic status, failure modes, and on-shell or non-complexified reformulations

A persistent theme in the NJA literature is that the algorithm is powerful but heuristic. Several papers explicitly state that it acts on metrics or solutions rather than directly on field equations, so the final spacetime must be checked independently. This caution becomes decisive outside general relativity (Hansen et al., 2013, Li et al., 22 Jan 2025).

A strong no-go result was obtained in quadratic gravity. There, applying NJA to a weakly deformed spherical black-hole seed produces a rotated metric that differs from the independently known slowly rotating solution, develops an unwanted $$g^{\mu\nu}= l^\mu n^\nu + n^\mu l^\nu - m^\mu \bar m^\nu - \bar m^\mu m^\nu$$8 component, fails the modified field equations through

$$g^{\mu\nu}= l^\mu n^\nu + n^\mu l^\nu - m^\mu \bar m^\nu - \bar m^\mu m^\nu$$9

and introduces a naked curvature singularity at

$$g^{\mu\nu}=-\ell^\mu n^\nu-\ell^\nu n^\mu+m^\mu\bar m^\nu+m^\nu\bar m^\mu.$$0

for $$g^{\mu\nu}=-\ell^\mu n^\nu-\ell^\nu n^\mu+m^\mu\bar m^\nu+m^\nu\bar m^\mu.$$1 and $$g^{\mu\nu}=-\ell^\mu n^\nu-\ell^\nu n^\mu+m^\mu\bar m^\nu+m^\nu\bar m^\mu.$$2. The conclusion drawn in that work is that, in general, NJA should not be used to construct rotating black holes outside general relativity (Hansen et al., 2013).

By contrast, some modified-gravity studies argue that NJA remains usable in theory-specific settings. In de Rham–Gabadadze–Tolley massive gravity, a rotating dRGT-Maxwell black hole is constructed from a static seed using a modified NJA adapted to the cosmological-constant case. There the method is defended by a lemma relating nonunitary gauge with Minkowski reference metric to unitary gauge with curved reference metric, and by explicit verification that the resulting metric reduces to the seed as $$g^{\mu\nu}=-\ell^\mu n^\nu-\ell^\nu n^\mu+m^\mu\bar m^\nu+m^\nu\bar m^\mu.$$3 (Li et al., 22 Jan 2025).

A different response to the off-shell criticism is to make NJA explicitly on-shell. One 2024 construction introduces the most general complex transformation

$$g^{\mu\nu}=-\ell^\mu n^\nu-\ell^\nu n^\mu+m^\mu\bar m^\nu+m^\nu\bar m^\mu.$$4

derives the corresponding axisymmetric metric from Schwarzschild, and then imposes Ricci-flatness as the equation of motion. In that treatment, Kerr, Taub–NUT, and Kerr–Taub–NUT arise as special parameter choices, and additional Ricci-flat axisymmetric black holes are suggested by the same master transformation. The authors call the resulting family the “on-shell Newman–Janis class of Schwarzschild black holes” (Lan et al., 2024).

Non-complexification procedures pursue a related strategy. In the Azreg-Aïnou-type variant, one performs the rotation-inspired tetrad deformation but replaces the complexified functions by real functions $$g^{\mu\nu}=-\ell^\mu n^\nu-\ell^\nu n^\mu+m^\mu\bar m^\nu+m^\nu\bar m^\mu.$$5, $$g^{\mu\nu}=-\ell^\mu n^\nu-\ell^\nu n^\mu+m^\mu\bar m^\nu+m^\nu\bar m^\mu.$$6, and $$g^{\mu\nu}=-\ell^\mu n^\nu-\ell^\nu n^\mu+m^\mu\bar m^\nu+m^\nu\bar m^\mu.$$7, fixed by integrability and Einstein-equation constraints. This method was used to construct rotating Janis–Newman–Winicour and rotating polytropic spacetimes in proper Boyer–Lindquist form; in the JNW case, the paper contrasts the no-complexification metric, which satisfies the standard energy conditions, with the complexification-based one, which does not (Solanki et al., 2021, Contreras et al., 2019). Recent work on rotating black holes with anisotropic matter likewise treats NJA as a generator of a Kerr-like ansatz whose remaining freedom is fixed by conditions such as $$g^{\mu\nu}=-\ell^\mu n^\nu-\ell^\nu n^\mu+m^\mu\bar m^\nu+m^\nu\bar m^\mu.$$8, rather than as a fully deterministic map from seed to solution (Kim et al., 10 Mar 2025).

6. Hidden symmetries, separability, and phenomenological uses

Beyond metric generation, NJA-generated geometries are important because of the hidden-symmetry structures they may preserve. For the general stationary, axisymmetric spacetime obtained through a successful NJA, null geodesics separate completely provided the Boyer–Lindquist integrability conditions hold. With impact parameters

$$g^{\mu\nu}=-\ell^\mu n^\nu-\ell^\nu n^\mu+m^\mu\bar m^\nu+m^\nu\bar m^\mu.$$9

the radial potential takes the generic form

$$\ell^\mu=\delta_r^\mu,\qquad n^\mu=\delta_u^\mu-\frac{f}{2}\delta_r^\mu,\qquad m^\mu=\frac{1}{\sqrt2\,\bar r}\left(\delta_\theta^\mu+\frac{i}{\sin\theta}\delta_\phi^\mu\right)$$0

and the shadow boundary is determined by unstable circular photon orbits satisfying $$\ell^\mu=\delta_r^\mu,\qquad n^\mu=\delta_u^\mu-\frac{f}{2}\delta_r^\mu,\qquad m^\mu=\frac{1}{\sqrt2\,\bar r}\left(\delta_\theta^\mu+\frac{i}{\sin\theta}\delta_\phi^\mu\right)$$1 and $$\ell^\mu=\delta_r^\mu,\qquad n^\mu=\delta_u^\mu-\frac{f}{2}\delta_r^\mu,\qquad m^\mu=\frac{1}{\sqrt2\,\bar r}\left(\delta_\theta^\mu+\frac{i}{\sin\theta}\delta_\phi^\mu\right)$$2. This yields a general parametric shadow formula for arbitrary successful NJA spacetimes (Shaikh, 2019).

Wave-equation separability is more restrictive. In the general NJA metric with functions $$\ell^\mu=\delta_r^\mu,\qquad n^\mu=\delta_u^\mu-\frac{f}{2}\delta_r^\mu,\qquad m^\mu=\frac{1}{\sqrt2\,\bar r}\left(\delta_\theta^\mu+\frac{i}{\sin\theta}\delta_\phi^\mu\right)$$3, $$\ell^\mu=\delta_r^\mu,\qquad n^\mu=\delta_u^\mu-\frac{f}{2}\delta_r^\mu,\qquad m^\mu=\frac{1}{\sqrt2\,\bar r}\left(\delta_\theta^\mu+\frac{i}{\sin\theta}\delta_\phi^\mu\right)$$4, and $$\ell^\mu=\delta_r^\mu,\qquad n^\mu=\delta_u^\mu-\frac{f}{2}\delta_r^\mu,\qquad m^\mu=\frac{1}{\sqrt2\,\bar r}\left(\delta_\theta^\mu+\frac{i}{\sin\theta}\delta_\phi^\mu\right)$$5, the Hamilton–Jacobi equation is separable when $$\ell^\mu=\delta_r^\mu,\qquad n^\mu=\delta_u^\mu-\frac{f}{2}\delta_r^\mu,\qquad m^\mu=\frac{1}{\sqrt2\,\bar r}\left(\delta_\theta^\mu+\frac{i}{\sin\theta}\delta_\phi^\mu\right)$$6 is additively separable,

$$\ell^\mu=\delta_r^\mu,\qquad n^\mu=\delta_u^\mu-\frac{f}{2}\delta_r^\mu,\qquad m^\mu=\frac{1}{\sqrt2\,\bar r}\left(\delta_\theta^\mu+\frac{i}{\sin\theta}\delta_\phi^\mu\right)$$7

whereas the massive Klein–Gordon equation requires the stronger condition

$$\ell^\mu=\delta_r^\mu,\qquad n^\mu=\delta_u^\mu-\frac{f}{2}\delta_r^\mu,\qquad m^\mu=\frac{1}{\sqrt2\,\bar r}\left(\delta_\theta^\mu+\frac{i}{\sin\theta}\delta_\phi^\mu\right)$$8

equivalently $$\ell^\mu=\delta_r^\mu,\qquad n^\mu=\delta_u^\mu-\frac{f}{2}\delta_r^\mu,\qquad m^\mu=\frac{1}{\sqrt2\,\bar r}\left(\delta_\theta^\mu+\frac{i}{\sin\theta}\delta_\phi^\mu\right)$$9 after imposing asymptotic flatness. For the massless Klein–Gordon equation, it is sufficient that $(u,r)$0 depend only on $(u,r)$1 (Chen et al., 2019).

The tensorial extension of NJA also clarifies where some of these integrability properties come from. In the null-rotation-augmented construction, the Carter Killing tensor of Kerr–Newman is generated from the Reissner–Nordström angular-momentum tensor, and the conformal Killing tensor is transported analogously. This shows that, at least for the Kerr–Newman family, NJA is tied not only to the metric but also to the underlying principal-null and hidden-symmetry structures (Keane, 2014).

These mathematical properties underpin extensive phenomenology. NJA-generated rotating spacetimes have been used to analyze shadows, accretion disks, Hawking temperature, tunneling, and orbital precession in a wide range of Kerr-like or non-Kerr-like backgrounds, including rotating Sen, rotating JNW, rotating regular black holes with cosmic strings, and rotating polytropic metrics (Cai et al., 1 Sep 2025, Solanki et al., 2021, Ali et al., 2022, Contreras et al., 2019). Taken together, this body of work suggests a dual status for NJA: it is both a historically successful recipe for constructing rotating geometries and a diagnostic framework whose validity depends on integrability, field-equation compatibility, and the geometric meaning assigned to the complex transformation.