The Newman-Penrose spin-coefficient formalism

Intuition Beginner

In ordinary geometry you describe a curved surface by laying down little arrows that point along directions you care about and then asking how those arrows turn as you slide them around. In relativity the directions that matter most are the paths light can take. So instead of building a frame out of three space arrows and one time arrow, you build it out of light rays. Two of the frame's legs point straight along two crossing beams of light, and the other two are sideways combinations that sweep out the directions across the beams.

This light-ray frame is the heart of the Newman-Penrose method. Once you choose it, every fact about how spacetime bends gets repackaged as a list of numbers measuring how the frame twists, spreads, and shears as you carry it along the beams. There are twelve of these turning-rates, and physicists named each one with a Greek letter.

The payoff is that hard questions about gravity become bookkeeping. Whether a bundle of light rays is spreading apart, rotating, or being squashed into an ellipse is read straight off two of the twelve numbers. The gravitational field of a black hole, written out in this frame, collapses to a single non-zero number where the full curvature tensor would have had dozens of pieces.

Visual Beginner

Picture two flashlight beams crossing at a point. One frame leg rides forward along the first beam, a second leg rides forward along the second beam. The remaining two legs are sideways, and because the sideways plane is best handled with one complex direction and its mirror image, we draw them as a single arrow and its reflection. Now drop a tiny ring of nearby light rays into the first beam and watch the ring change shape as the rays travel: it can grow (expansion), spin (twist), and stretch into an ellipse (shear). Those three changes are exactly what two of the spin coefficients record.

Worked example Beginner

Take flat spacetime and the simplest outgoing light: spheres of light expanding from a point, like ripples from a stone but in three dimensions and moving at light speed. Put the first frame leg along the outgoing rays.

At radius $r=2$ the sphere of light has surface area proportional to $r$ squared, so area $\propto 4$. One unit of travel later, at $r=3$, the area is $\propto 9$. The rays are spreading: a small patch of the sphere grew from $4$ to $9$ in area, a factor of $2.25$.

The spin coefficient that measures this spreading turns out to be exactly $-1/r$ for these expanding spheres. At $r=2$ that number is $-1/2=-0.5$; at $r=3$ it is $-1/3\approx -0.333$. The size of the number shrinks as $r$ grows, matching the everyday fact that far from the source the ripples look almost flat and spread more gently.

The rays here only spread; they do not rotate and they do not get squashed. So the rotation rate and the squashing rate are both $0$. What this tells us: a single one of the twelve coefficients, namely $-1/r$, already captures the entire focusing behaviour of expanding light in empty space, and its two partners being $0$ certifies that the light is twist-free and shear-free.

Check your understanding Beginner

Formal definition Intermediate+

Work on a four-dimensional Lorentzian spacetime $(M,g)$ of signature $(+,-,-,-)$ equipped with its Levi-Civita connection $\nabla$ and, locally, a spin structure. A null tetrad is a frame $({\ell }^a,n^a,m^a,{\bar{m}}^a)$ in which $\ell$ and $n$ are real future-pointing null vectors and $m$ is complex with conjugate $\bar{m}$, subject to the normalisation

$${\ell }_a{n}^a=1,m_a{\bar{m}}^a=-1,$$

with all other contractions vanishing. The metric is recovered as $g_{ab}={\ell }_a{n}_b+n_a{\ell }_b-m_a{\bar{m}}_b-{\bar{m}}_a{m}_b$. Such a tetrad is the frame component of a spin dyad $(o^A,{\iota }^A)$ — a basis of the two-spinor space normalised by $o_A{\iota }^A=1$ — through the soldering form, so that ${\ell }^a=o^A{\bar{o}}^{A^{\prime }}$, $n^a={\iota }^A{\bar{\iota }}^{A^{\prime }}$, $m^a=o^A{\bar{\iota }}^{A^{\prime }}$, ${\bar{m}}^a={\iota }^A{\bar{o}}^{A^{\prime }}$.

The four legs act as directional derivative operators:

$$D={\ell }^a{\nabla }_a,\Delta =n^a{\nabla }_a,\delta =m^a{\nabla }_a,\bar{\delta }={\bar{m}}^a{\nabla }_a.$$

The twelve spin coefficients are the independent complex components of the connection in this frame, equivalently the dyad components ${\Gamma }_{ABC{D}^{\prime }}={\gamma }_{ABC{C}^{\prime }}$ of the spinor connection. They are named and grouped as

$$\kappa ,\sigma ,\rho ,\tau (\text{derivatives of }\ell \text{ in the four directions}),$$ $$\nu ,\lambda ,\mu ,\pi (\text{derivatives of }n),$$ $$ϵ,\gamma ,\beta ,\alpha (\text{the residual rotation of the dyad}),$$

with the standard definitions $\kappa =-m^a{\ell }^b{\nabla }_b{\ell }_a$, $\sigma =-m^a{m}^b{\nabla }_b{\ell }_a$, $\rho =-m^a{\bar{m}}^b{\nabla }_b{\ell }_a$, $\tau =-m^a{n}^b{\nabla }_b{\ell }_a$, and the analogous primed-frame and rotation expressions. Each is a genuine connection symbol: it measures the rate at which one tetrad leg turns into another along a chosen direction.

The curvature splits into Weyl and Ricci parts, each repackaged into dyad scalars. The five Weyl scalars are the totally symmetric Weyl spinor's components,

$${\Psi }_0=C_{abcd}{\ell }^a{m}^b{\ell }^c{m}^d,{\Psi }_1=C_{abcd}{\ell }^a{n}^b{\ell }^c{m}^d,\dots ,{\Psi }_4=C_{abcd}{n}^a{\bar{m}}^b{n}^c{\bar{m}}^d,$$

equivalently ${\Psi }_k={\Psi }_{ABCD}{\iota }^k{o}^{4-k}$ in dyad components. The trace-free Ricci tensor becomes the Hermitian array of Ricci scalars ${\Phi }_{ij}$ ($i,j\in \{0,1,2\}$, ${\Phi }_{ij}={\bar{\Phi }}_{ji}$), and the scalar curvature contributes $\Lambda =R/24$.

Key theorem with proof Intermediate+

The content of the formalism is that the first structure equation of Cartan, written in a null tetrad, becomes a closed system relating directional derivatives of the spin coefficients to quadratic spin-coefficient terms and the curvature scalars. The cleanest such statement isolates the two coefficients $\rho$ and $\sigma$ and connects them to the geometry of a null congruence. We follow Chandrasekhar ^{[Chandrasekhar 1983]} and Penrose-Rindler ^{[Penrose 1984]}.

Theorem (Sachs optical equations / the NP $\rho$-equation). Let ${\ell }^a$ be tangent to a null geodesic congruence affinely parametrised, so $\kappa =0$ and $ϵ+\overline{ϵ}=0$. Then the expansion, twist, and shear of the congruence are encoded in $\rho$ and $\sigma$ by

$$\mathrm{Re}\rho =-\theta ,\mathrm{Im}\rho =\omega ,|\sigma |=|\hat{\sigma }|,$$

where $\theta$ is the expansion (fractional area growth rate), $\omega$ the twist, and $\hat{\sigma }$ the complex shear, and these satisfy the propagation equations

$$D\rho ={\rho }^2+\sigma \bar{\sigma }+{\Phi }_{00},D\sigma =(\rho +\bar{\rho })\sigma +{\Psi }_0.$$

Proof. Start from the definitions $\rho =-m^a{\bar{m}}^b{\nabla }_b{\ell }_a$ and $\sigma =-m^a{m}^b{\nabla }_b{\ell }_a$. Decompose the covariant gradient of $\ell$ restricted to the screen space spanned by $m,\bar{m}$. Since $\ell$ is null and geodesic with $\kappa =0$, the gradient ${\nabla }_b{\ell }_a$ projected onto the screen is a $2\times 2$ matrix whose trace is $-(\rho +\bar{\rho })$, whose antisymmetric part is $-(\rho -\bar{\rho })$, and whose symmetric trace-free part has magnitude $|\sigma |$. Identifying the trace with the fractional rate of change of a transverse area element gives $\rho +\bar{\rho }=-2\theta$, hence $\mathrm{Re}\rho =-\theta$; the antisymmetric part is the rotation of the screen, giving $\mathrm{Im}\rho =\omega$; the trace-free part is the area-preserving distortion, giving $|\sigma |$ as the shear magnitude.

For the propagation equations, apply the Ricci identity ${\nabla }_b{\nabla }_c{\ell }_a-{\nabla }_c{\nabla }_b{\ell }_a=R_{abcd}{\ell }^d$ and contract with the tetrad legs ${\ell }^c{m}^a{\bar{m}}^b$ for the first equation and ${\ell }^c{m}^a{m}^b$ for the second. On the left the commutator of $D$ with the screen derivatives, expanded through the connection coefficients, produces $D\rho$ together with the quadratic terms ${\rho }^2+\sigma \bar{\sigma }$ for the ${\bar{m}}^b$ contraction, and $(\rho +\bar{\rho })\sigma$ for the $m^b$ contraction. On the right the curvature contraction splits: the Ricci part of $R_{abcd}{\ell }^a{\ell }^d$ contracted into $m\bar{m}$ yields exactly ${\Phi }_{00}=-\frac{1}{2}{R}_{ab}{\ell }^a{\ell }^b$, while the Weyl part contracted into $m^a{\ell }^b{\ell }^c{m}^d$ yields ${\Psi }_0$. Collecting the two contractions gives the stated pair. $■$

Bridge. This pair of equations builds toward the Goldberg-Sachs theorem and the optical classification of null congruences, and it appears again in the peeling theorem and in the Raychaudhuri focusing argument that drives the singularity theorems. The foundational reason the system closes is that a null tetrad turns the second-order Ricci identity into first-order transport along the tetrad legs: the central insight is that the connection in this frame is just the twelve scalars, so every curvature identity becomes a polynomial relation among directional derivatives of those scalars. The $D\sigma$ equation is exactly the statement that ${\Psi }_0$ sources shear, which generalises to the whole peeling hierarchy where each ${\Psi }_k$ governs a different falloff; and the vanishing locus $\sigma =0$ with ${\Psi }_0=0$ is dual to the shear-free condition that the next theorem promotes into an algebraic statement about the Weyl tensor. Putting these together, the optical scalars of 03.02.18 cease to be allusions and become computable functions of the affine parameter.

Exercises Intermediate+

Advanced results Master

The eighteen complex Ricci (NP) equations and the eight Bianchi equations form the full first-order system. Two structural refinements give the formalism its working power.

The first is the compacted (GHP) formalism of Geroch, Held, and Penrose ^{[Geroch 1973]}. The choice of dyad has a residual two-parameter freedom — a boost $o\mapsto \lambda o$, $\iota \mapsto {\lambda }^{-1}\iota$ and a spin $\lambda =e^{i\vartheta }$ — under which a quantity of boost weight $b$ and spin weight $s$ transforms by ${\lambda }^p{\bar{\lambda }}^q$ with $p=(b+s)/2$, $q=(b-s)/2$. Of the twelve spin coefficients, exactly the four optical ones $\kappa ,\sigma ,\rho ,\tau$ and their primed partners ${\kappa }^{\prime },{\sigma }^{\prime },{\rho }^{\prime },{\tau }^{\prime }$ are properly weighted; the remaining four, $ϵ,\beta ,\alpha ,\gamma$, are non-tensorial and are absorbed into weighted derivative operators $\text{thorn},{\text{thorn}}^{\prime },\text{edth},{\text{edth}}^{\prime }$ (thorn and eth). The field equations then collapse to a handful of weighted equations invariant under the discrete priming ($\ell \leftrightarrow n$) and starring ($m\leftrightarrow \bar{m}$) symmetries, halving the bookkeeping.

The second is the Goldberg-Sachs theorem in NP form: a vacuum spacetime (${\Phi }_{ij}=\Lambda =0$) is algebraically special if and only if it admits a null geodesic shear-free congruence, i.e. $\kappa =\sigma =0$ for a suitable $\ell$, and then $\ell$ is the repeated principal null direction with ${\Psi }_0={\Psi }_1=0$. This converts the algebraic speciality of 03.02.18 into a differential condition on the spin coefficients, and it is the engine by which Kerr's metric was found: imposing $\kappa =\sigma =0$ plus stationarity and axisymmetry reduces the vacuum equations to a tractable system.

The optical scalars feed the Raychaudhuri equation for null congruences: $D\theta =-{\theta }^2-|\hat{\sigma }{|}^2-\frac{1}{2}{R}_{ab}{\ell }^a{\ell }^b+{\omega }^2$, the real part of the $D\rho$ equation. With the null energy condition and zero twist this forces $\theta \to -\infty$ in finite affine time once it goes negative — the focusing theorem underlying Penrose's singularity argument.

Synthesis. The Newman-Penrose formalism is the foundational reason the curvature of a four-dimensional Lorentzian spacetime reduces to a finite list of scalars transported along light rays, and the optical scalars, the Weyl scalars, and the field equations are three faces of one structure: the spinor connection in a dyad. The central insight is that a null tetrad is a spin dyad in disguise, so the apparatus of 03.02.18 — the Weyl spinor ${\Psi }_{ABCD}$, its five components ${\Psi }_0\dots {\Psi }_4$, the principal null directions — is exactly the curvature half of the NP scalars, while the spin coefficients supply the connection half. This is exactly why Goldberg-Sachs reads as an identity between an optical condition ($\kappa =\sigma =0$) and an algebraic one (a repeated PND), and why the peeling theorem sorts the ${\Psi }_k$ radially: each governs a transport equation whose falloff rate is fixed by its dyad weight. The GHP refinement generalises the whole scheme by quotienting out the dyad freedom, so that the compacted equations are dual to the boost-weight grading; and the Raychaudhuri focusing law that drives the singularity theorems is nothing but the real part of the $D\rho$ equation. Putting these together, the bridge from local curvature to global causal structure runs entirely through twelve complex numbers and four derivative operators.

Full proof set Master

Proposition (the four legs are a derivation basis dual to the connection). Let $(o^A,{\iota }^A)$ be a spin dyad with $o_A{\iota }^A=1$ and $(\ell ,n,m,\bar{m})$ the associated null tetrad. Then the twelve spin coefficients are precisely the independent components of the spinor connection ${\gamma }_{A{A}^{\prime }BC}$ in this dyad, and conversely the connection is reconstructed from them; equivalently ${\nabla }_a{o}^B=-(ϵo^B-\kappa {\iota }^B){\bar{o}}_{A^{\prime }}\cdots$ with each coefficient appearing exactly once.

Proof. The spinor connection is the $\mathfrak{sl}(2,\mathbb{C})$-valued one-form ${\gamma }_{A{A}^{\prime }BC}={\gamma }_{A{A}^{\prime }(BC)}$ defined by ${\nabla }_{A{A}^{\prime }}{\xi }_B={\partial }_{A{A}^{\prime }}{\xi }_B-{\gamma }_{A{A}^{\prime }B}{}^C{\xi }_C$; symmetry in $BC$ holds because the connection preserves the symplectic ${\epsilon }_{BC}$, so ${\gamma }_{A{A}^{\prime }BC}={\gamma }_{A{A}^{\prime }(BC)}$ has three independent symmetric $BC$-components for each of the four $A{A}^{\prime }$ choices, giving $3\times 4=12$ complex numbers. Expand each in the dyad: ${\gamma }_{A{A}^{\prime }BC}={\gamma }_{A{A}^{\prime }BC}(o,\iota )$-components. The four $A{A}^{\prime }$ projections are the four directional derivatives $D,\Delta ,\delta ,\bar{\delta }$ obtained by contracting $A$ with $o$ or $\iota$ and $A^{\prime }$ with $\bar{o}$ or $\bar{\iota }$. The three symmetric $BC$ components $oo$, $o\iota$, $\iota \iota$ pick out, respectively, the rate of rotation of $o$ into itself, the mixing of $o$ and $\iota$, and the rotation of $\iota$. Reading off the twelve entries and matching against the defining contractions of the previous section identifies them with $(\kappa ,\sigma ,\rho ,\tau ,\nu ,\lambda ,\mu ,\pi ,ϵ,\gamma ,\beta ,\alpha )$, each occurring once. Reconstruction is immediate: the connection one-form is the sum over the twelve dyad basis elements with these coefficients, so the data are equivalent. $■$

Proposition (reality and trace structure of the Ricci scalars). The Ricci spinor components ${\Phi }_{ij}$ defined by contracting the trace-free Ricci tensor into the tetrad satisfy the Hermiticity ${\Phi }_{ij}={\bar{\Phi }}_{ji}$, leaving nine real degrees of freedom, and $\Lambda =R/24$ is real, together accounting for the ten components of the Ricci tensor.

Proof. The trace-free Ricci tensor $S_{ab}=R_{ab}-\frac{1}{4}R{g}_{ab}$ is real and symmetric with nine independent components. In spinor form $S_{ab}=-2{\Phi }_{AB{A}^{\prime }{B}^{\prime }}$ with ${\Phi }_{AB{A}^{\prime }{B}^{\prime }}={\Phi }_{(AB)(A^{\prime }{B}^{\prime })}$ and the reality condition $\overline{{\Phi }_{AB{A}^{\prime }{B}^{\prime }}}={\Phi }_{AB{A}^{\prime }{B}^{\prime }}$ inherited from reality of $S_{ab}$. The dyad components ${\Phi }_{ij}$ ($i$ counting $\iota$-weight in the unprimed pair, $j$ in the primed pair) then satisfy ${\Phi }_{ij}={\bar{\Phi }}_{ji}$ by the reality of the spinor; the diagonal entries ${\Phi }_{00},{\Phi }_{11},{\Phi }_{22}$ are real and the off-diagonal pairs are complex conjugates, giving $3+2\cdot 3=9$ real numbers. The scalar curvature contributes the single real $\Lambda =R/24$, and $9+1=10$ matches the Ricci tensor count. The sign and factor conventions follow Penrose-Rindler ^{[Penrose 1984]}. $■$

Connections Master

The Weyl scalars ${\Psi }_0,\dots ,{\Psi }_4$ that this formalism transports are the dyad components of the Weyl spinor ${\Psi }_{ABCD}$ whose factorisation into principal spinors defines the Petrov types in 03.02.18; the present unit supplies the connection and field equations that the Petrov unit names but does not build, and Goldberg-Sachs is the hinge between them.

The null tetrad and the directional derivatives are the spin-frame realisation of the Cartan tetrad and spin connection of 13.02.03; the twelve spin coefficients are exactly the tetrad components of the connection one-form ${\omega }_{\mu ab}$ specialised to a null frame and a two-spinor basis.

The whole apparatus is written in the two-spinor calculus and the soldering form $V^a\leftrightarrow V^{A{A}^{\prime }}$ that are the subject of the co-produced 03.02.41; the dyad $(o^A,{\iota }^A)$, the normalisation $o_A{\iota }^A=1$, and the spinor connection ${\gamma }_{A{A}^{\prime }BC}$ all live there, and the present unit is the curved-frame application of that calculus.

The zero-rest-mass field equation ${\nabla }^{A{A}^{\prime }}{\varphi }_{AB\cdots L}=0$ of the co-produced 03.02.42 is written and solved in exactly this formalism: the Maxwell scalars ${\varphi }_0,{\varphi }_1,{\varphi }_2$ and their NP transport equations are the spin-1 instance of the same dyad-component machinery used here for the spin-2 Weyl field.

The optical scalars feed the Raychaudhuri focusing argument and the singularity theorems through the Jacobi-field and geodesic-deviation analysis of 03.02.19; expansion, shear, and twist are the null-congruence analogues of the conjugate-point machinery for timelike geodesics.

Historical & philosophical context Master

The formalism was introduced by Ezra Newman and Roger Penrose in 1962 ^{[Newman 1962]} as a method for studying gravitational radiation, building on Penrose's 1960 spinor reformulation of general relativity and on the null-tetrad ideas already present in the work of Sachs ^{[Sachs 1961]} on the optical scalars and the peeling property. The original paper writes out all twelve spin coefficients, the eighteen field equations, and the eight Bianchi equations explicitly, and a published erratum corrected sign errors in the field equations — a measure of how error-prone the hand computation was before the compacted formalism existed.

Chandrasekhar's 1983 monograph ^{[Chandrasekhar 1983]} made the formalism the standard computational engine of black-hole perturbation theory, using it to separate the Teukolsky equation and to prove the modal stability of the Kerr metric. The compacted refinement of Geroch, Held, and Penrose ^{[Geroch 1973]} removed the non-tensorial spin coefficients by passing to weighted scalars, exposing the boost-and-spin grading that the original component formalism obscured. The Kerr metric itself was discovered in 1963 by imposing the geodesic shear-free condition that Goldberg and Sachs had shown to be equivalent to algebraic speciality.

Bibliography Master

@article{NewmanPenrose1962,
  author  = {Newman, Ezra T. and Penrose, Roger},
  title   = {An Approach to Gravitational Radiation by a Method of Spin Coefficients},
  journal = {Journal of Mathematical Physics},
  volume  = {3},
  pages   = {566--578},
  year    = {1962},
  note    = {Erratum: J. Math. Phys. 4 (1963) 998}
}

@book{PenroseRindler1984,
  author    = {Penrose, Roger and Rindler, Wolfgang},
  title     = {Spinors and Space-Time, Volume 1: Two-Spinor Calculus and Relativistic Fields},
  publisher = {Cambridge University Press},
  year      = {1984}
}

@book{Chandrasekhar1983,
  author    = {Chandrasekhar, Subrahmanyan},
  title     = {The Mathematical Theory of Black Holes},
  publisher = {Oxford University Press},
  year      = {1983}
}

@article{GHP1973,
  author  = {Geroch, Robert and Held, Alan and Penrose, Roger},
  title   = {A space-time calculus based on pairs of null directions},
  journal = {Journal of Mathematical Physics},
  volume  = {14},
  pages   = {874--881},
  year    = {1973}
}

@article{Sachs1961,
  author  = {Sachs, Rainer K.},
  title   = {Gravitational waves in general relativity. VI. The outgoing radiation condition},
  journal = {Proceedings of the Royal Society A},
  volume  = {264},
  pages   = {309--338},
  year    = {1961}
}

@book{Stewart1991,
  author    = {Stewart, John},
  title     = {Advanced General Relativity},
  publisher = {Cambridge University Press},
  year      = {1991}
}