Black hole solutions of the Einstein-Dirac-Maxwell system: eliminating the horizon power-law assumption, settling the extremal case, and extending to stationary axisymmetry

The Einstein-Dirac-Maxwell (EDM) system couples general relativity to massive charged spin-1/2 fermions and electromagnetism. In 1999 Finster, Smoller, and Yau proved that, under three regularity assumptions on the event horizon — most notably a power-law degeneracy $A(r)\sim c(r-\bar r)^s$ of the metric function — every static, spherically symmetric black hole solution of the EDM system is either a Reissner-Nordström solution with vanishing spinors or has an extremal-type ($s=2$) horizon for which numerical evidence shows the spinors cannot be normalized. Their 2000 survey explicitly listed as unresolved: eliminating the inessential power-law assumption; replacing the numerical exclusion of the extremal case by a rigorous proof; and settling the fully coupled stationary axisymmetric system. These three components remain open: Dafermos's alternative method generalized the nonexistence to time-periodic solutions but under different strong causal-structure and regularity assumptions; extensions to other gauge groups inherit comparable horizon assumptions; fixed-background analyses (including recent work on stationary Dirac states around Kerr) leave the coupled problem untouched; and unpublished claims of self-consistent Dirac solutions on Kerr-Newman have not been accepted. Two companion open questions raised in the same survey — whether multi-black-hole configurations can be stabilized by angular momentum in 3+1 dimensions, and the construction of quasilocal mass and angular momentum behaving well under time evolution — also remain open; they are separable program-level questions, while the core of the problem is the EDM black-hole problem.

Background

An isolated stationary black hole in general relativity is expected, according to the no-hair conjecture, to be characterized at infinity by only a few parameters (mass, angular momentum, charge), with no independent matter degrees of freedom persisting outside the event horizon — the surface through which no signal can escape to infinity. Rigorous no-hair statements depend delicately on the matter model. For gauge fields, counterexamples to naive expectations such as particlelike solutions and static 'colored' black holes of the Einstein-Yang-Mills equations (gravity coupled to a non-Abelian gauge field) are well known, but fermionic matter — fields of spin 1/2 governed by the Dirac equation — long resisted comparable analysis, both because a single spinor cannot be invariant under spatial rotations and because the Dirac probability current (a divergence-free, future-directed timelike vector field built from the wave function) enters the gravitational field equations in an essential way.

The system of interest is the Einstein-Dirac-Maxwell (EDM) system: Einstein's equations $R_{ij}-\frac{1}{2}R\,\delta_{ij}=-8\pi T_{ij}$ for the spacetime metric, coupled to the Maxwell field and to massive, electrically charged Dirac spinors $\Psi_a$ of mass $m$ and charge $e$ satisfying $(\mathcal{G}-m)\Psi_a=0$, where $\mathcal{G}$ is the Dirac operator (the spin-1/2 wave operator, depending on both the metric and the electromagnetic potential); the energy-momentum tensor $T_{ij}$ sums the Dirac and Maxwell contributions.

Finster, Smoller, and Yau (1998-2000) analyzed static, spherically symmetric configurations. Because individual spinors cannot be spherically symmetric, they used a 'filled shell' ansatz: $(2j+1)$ fermions occupying all states of angular momentum $j$, with radial amplitudes $\alpha(r)$, $\beta(r)$ and a common frequency $\omega$, arranged so that the total stress-energy is static and spherically symmetric. With the metric written as

$$ds^{2}=\frac{dt^{2}}{T(r)^{2}}-\frac{dr^{2}}{A(r)}-r^{2}(d\vartheta^{2}+\sin^{2}\vartheta\,d\varphi^{2})$$

and a purely electric Coulomb potential $\phi(r)$, the field equations reduce to a nonlinear system of ordinary differential equations for $\alpha$, $\beta$, $A$, $T$, and $\phi$. Two classes of solutions were considered: particlelike solutions, smooth for all $r\ge 0$, which were constructed numerically (with binding-energy curves forming spirals reminiscent of boson stars — localized gravitating scalar-field configurations — and a stability analysis via Conley index theory, a topological method for classifying invariant sets of flows); and black hole solutions, defined on $r>\bar r>0$ with $A(\bar r)=0$ and $A>0$ outside, the hypersurface $r=\bar r$ being the event horizon.

For black hole solutions one requires asymptotic flatness ($A,T\to 1$ as $r\to\infty$), finite ADM mass $\rho=\lim_{r\to\infty}\frac{r}{2}(1-A(r))$ (the total gravitational mass measured at spatial infinity), and normalization of the spinors,

$$0<\int_{r_0}^{\infty}(\alpha^{2}+\beta^{2})\,\frac{T}{\sqrt{A}}\,dr<\infty\qquad\text{for all }r_0>\bar r,$$

so that the Dirac wave functions define physically meaningful probability amplitudes. The Reissner-Nordström (RN) solution — the unique static, spherically symmetric electrovacuum black hole, electrovacuum meaning that gravity is coupled to nothing but an electromagnetic field — is the trivial solution of the EDM system with vanishing spinors; its non-extreme case has two horizons, while the extreme case (charge equal in magnitude to mass) has a single, doubly degenerate horizon.

The nonexistence theorem of Finster-Smoller-Yau states: any static, spherically symmetric black hole solution of the EDM equations satisfying three regularity assumptions on the horizon — (I) the volume element $r^{2}A^{-1/2}T^{-1}$ is smooth and non-zero on the horizon; (II) the electromagnetic field strength $|\phi'(r)|$ is bounded near the horizon; (III) the metric function obeys a power law $A(r)=c\,(r-\bar r)^{s}+O((r-\bar r)^{s+1})$ with $c,s>0$ ($s=1$ is the Schwarzschild type, $s=2$ the extreme RN type) — is either a non-extreme RN solution with spinors vanishing identically, or has $s=2$, in which case the metric, electromagnetic field, and spinors are constrained to specific near-horizon expansions: the horizon is of extreme-RN type and the spinors decay like $(r-\bar r)^{\kappa}$ with $\kappa>\frac{1}{2}$. For the $s=2$ case, numerical integration shows that global solutions either develop singularities at some $r>\bar r$ or fail the normalization condition, so the coupled system is expected to admit no black hole solutions carrying fermions at all. Physically, a collapsing cloud of Dirac particles must disappear entirely through the horizon or escape to infinity. The authors note that (I) and (II) have physical content (a freely falling observer should feel no unbounded forces at the horizon), whereas (III) is a purely technical condition that 'does not seem to be essential' and could 'be weakened or even omitted completely' with more mathematical effort.

Two related results on fixed backgrounds (spinors not influencing the geometry) complete the picture: on a non-extreme RN background the Dirac equation admits no normalizable, time-periodic solutions — the probability integral over exterior hypersurfaces diverges, by a radial-flux argument using current conservation — and on an extreme RN background every normalizable time-periodic Dirac solution vanishes identically outside the horizon; the former result was extended to a general class of axisymmetric black hole geometries, including non-extreme Kerr-Newman (the unique stationary axisymmetric electrovacuum black hole, i.e. the charged, rotating generalization of Schwarzschild and RN). Dafermos later gave a different route to nonexistence for spherically symmetric systems, static or time-periodic, using null coordinates and the divergence-free Dirac current, under assumptions on the causal structure of the domain of outer communications (the region visible from infinity) that in effect exclude extreme-type horizons.

The same survey raised two companion program-level questions: whether multiple black holes can be held in stationary equilibrium by angular momentum (known explicit static double-black-hole solutions necessarily involve conical singularities — 'struts' — on the symmetry axis), and the lack of quasilocal (defined on a closed two-surface rather than at infinity) mass and angular momentum functionals that behave well under time evolution — the Hawking mass, a simple quasilocal mass functional, is positive for surfaces minimizing area under a volume constraint only when the scalar curvature is nonnegative (Christodoulou-Yau), and is not positive in general. These companion questions are separable from the EDM black-hole problem.

Problem Statement

Determine, with mathematical rigor, whether the Einstein-Dirac-Maxwell (EDM) system admits black hole solutions carrying nonvanishing, normalizable fermionic wave functions outside the event horizon — that is, complete the spinor no-hair theorem. Concretely, resolve the three linked components left open by the source survey:

(i) The static, spherically symmetric case without technical horizon assumptions. The known nonexistence theorem assumes, besides asymptotic flatness, finite ADM mass, and normalization of the spinors: (I) smoothness and non-vanishing of the volume element on the horizon, (II) boundedness of the electromagnetic field strength near the horizon, and (III) the power-law behavior $A(r)=c\,(r-\bar r)^{s}+O((r-\bar r)^{s+1})$. Prove that every static, spherically symmetric black hole solution of the EDM equations (in the filled-shell ansatz setting of the theorem, or a strictly more general symmetry class) satisfying asymptotic flatness, finite ADM mass, and normalization of the spinors is necessarily a Reissner-Nordström solution with vanishing spinors — equivalently, that no such solution carries nonvanishing normalizable spinors — under assumptions strictly weaker than (I)-(III), in particular without (III), covering horizons with arbitrary (including non-power-law) degeneracy of $A(r)$ at $r=\bar r$ (the horizon radius, where $A(\bar r)=0$); or exhibit an explicit counterexample — a black hole solution outside the assumption class (I)-(III) carrying nonvanishing normalizable spinors.

(ii) The extremal case. Under the original theorem, every solution that is not a Reissner-Nordström solution with vanishing spinors must have $s=2$ (an extreme-Reissner-Nordström-type horizon) with the specific near-horizon expansions derived in the theorem, and numerical evidence indicates all such solutions either develop singularities at some $r>\bar r$ or have non-normalizable spinors. Prove this exclusion rigorously — no global solution of the coupled system with an $s=2$ horizon and normalized spinors exists — or construct such a solution.

(iii) The stationary axisymmetric case. Beyond spherical symmetry, settle the fully coupled stationary, axisymmetric EDM (or Einstein-Dirac) system: determine whether it admits black hole solutions with spinor matter, either by construction (an exact solution or a computer-assisted certified solution) or by a nonexistence theorem. No ansatz, construction, or theorem is currently known for this case.

The common objective uniting the three components is to settle whether fermionic matter can persist outside a black hole horizon in equilibrium, across the full setting of the original theorem (all horizon types, and beyond spherical symmetry). Component (ii) is the $s=2$ special case of component (i): resolving (i) also resolves (ii), whereas resolving (ii) alone leaves the non-power-law horizon degeneracies of (i) open.

References

1. F. Finster, J. Smoller, S.-T. Yau, 'Some recent progress in classical general relativity', Journal of Mathematical Physics 41 (2000) 3943-3963, DOI 10.1063/1.533332, arXiv:gr-qc/0001064, https://arxiv.org/abs/gr-qc/0001064
2. F. Finster, J. Smoller, S.-T. Yau, 'Non-Existence of Black Hole Solutions for a Spherically Symmetric, Static Einstein-Dirac-Maxwell System', Communications in Mathematical Physics 205 (1999) 249-262, DOI 10.1007/s002200050675, arXiv:gr-qc/9810048, https://arxiv.org/abs/gr-qc/9810048
3. F. Finster, J. Smoller, S.-T. Yau, 'Particle-Like Solutions of the Einstein-Dirac Equations', Physical Review D 59 (1999) 104020, DOI 10.1103/PhysRevD.59.104020, arXiv:gr-qc/9801079, https://arxiv.org/abs/gr-qc/9801079
4. F. Finster, J. Smoller, S.-T. Yau, 'Particle-Like Solutions of the Einstein-Dirac-Maxwell Equations', Physics Letters A 259 (1999) 431-436, DOI 10.1016/S0375-9601(99)00457-0, arXiv:gr-qc/9802012, https://arxiv.org/abs/gr-qc/9802012
5. F. Finster, J. Smoller, S.-T. Yau, 'Non-Existence of Time-Periodic Solutions of the Dirac Equation in a Reissner-Nordstrom Black Hole Background', Journal of Mathematical Physics 41 (2000) 2173-2194, DOI 10.1063/1.533234, arXiv:gr-qc/9805050, https://arxiv.org/abs/gr-qc/9805050
6. F. Finster, N. Kamran, J. Smoller, S.-T. Yau, 'Non-Existence of Time-Periodic Solutions of the Dirac Equation in an Axisymmetric Black Hole Geometry', Communications on Pure and Applied Mathematics 53 (2000) 902-929, DOI 10.1002/(SICI)1097-0312(200007)53:7<902::AID-CPA4>3.0.CO;2-4, arXiv:gr-qc/9905047, https://arxiv.org/abs/gr-qc/9905047
7. F. Finster, J. Smoller, S.-T. Yau, 'The Interaction of Dirac Particles with Non-Abelian Gauge Fields and Gravity - Black Holes', Michigan Mathematical Journal 47 (2000) 199-208, DOI 10.1307/mmj/1030374678, arXiv:gr-qc/9910047, https://arxiv.org/abs/gr-qc/9910047
8. F. Finster, J. Smoller, S.-T. Yau, 'Absence of Static, Spherically Symmetric Black Hole Solutions for Einstein-Dirac/Yang-Mills Equations with Complete Fermion Shells', Advances in Theoretical and Mathematical Physics 4 (2000) 1231-1257, DOI 10.4310/ATMP.2000.V4.N6.A3, arXiv:gr-qc/0005028, https://arxiv.org/abs/gr-qc/0005028
9. F. Finster, J. Smoller, S.-T. Yau, 'Non-Existence of Black Hole Solutions to Static, Spherically Symmetric Einstein-Dirac Systems - a Critical Discussion' (unpublished note, 2002), arXiv:gr-qc/0211043, https://arxiv.org/abs/gr-qc/0211043
10. M. Dafermos, 'On time-periodic black-hole solutions to certain spherically symmetric Einstein-matter systems', Communications in Mathematical Physics 238 (2003) 411-427, DOI 10.1007/s00220-003-0870-0, arXiv:gr-qc/0210025, https://arxiv.org/abs/gr-qc/0210025
11. Y. Bernard, 'Non-existence of black-hole solutions for the electroweak Einstein-Dirac-Yang/Mills equations', Classical and Quantum Gravity 23 (2006) 4433-4451, DOI 10.1088/0264-9381/23/13/009, https://doi.org/10.1088/0264-9381/23/13/009
12. S. Bhattacharya, 'Note on black hole no hair theorems for massive forms and spin-1/2 fields', Physical Review D 88 (2013) 044053, DOI 10.1103/PhysRevD.88.044053, arXiv:1307.2213, https://arxiv.org/abs/1307.2213
13. D. Ranganathan, 'A self consistent solution to the Einstein Maxwell Dirac Equations' (unpublished, 2003), arXiv:gr-qc/0306090, https://arxiv.org/abs/gr-qc/0306090
14. V. I. Dokuchaev, Yu. N. Eroshenko, 'Stationary Solutions of the Dirac Equation in the Gravitational Field of a Charged Black Hole', Journal of Experimental and Theoretical Physics 117 (2013) 72, DOI 10.1134/S1063776113080049, arXiv:1308.5927, https://arxiv.org/abs/1308.5927
15. V. Dzhunushaliev, V. Folomeev, 'Condensation of a spinor field at the event horizon', arXiv:2505.22163 (2025), https://arxiv.org/abs/2505.22163
16. J. L. Blázquez-Salcedo, C. Knoll, E. Radu, 'Einstein-Dirac-Maxwell wormholes: ansatz, construction and properties of symmetric solutions', European Physical Journal C 82 (2022) 533, DOI 10.1140/epjc/s10052-022-10488-6, arXiv:2108.12187, https://arxiv.org/abs/2108.12187
17. P. G. LeFloch, Y. Ma, W. Zhang, 'The global nonlinear stability of Minkowski spacetime with self-gravitating massive Dirac fields', arXiv:2510.20626 (2025), https://arxiv.org/abs/2510.20626
18. S. Guo, P.-Y. Chen, Y.-H. Huang, X. Li, Y. Liang, K. Lin, L. Wen, 'Stationary Dirac condensates around Kerr black holes', arXiv:2608.25436 (2026), https://arxiv.org/abs/2608.25436
19. M. Hernández Márquez, M. Alcubierre, 'Charged Dirac stars', arXiv:2607.26374 (2026), https://arxiv.org/abs/2607.26374
20. M. Astorino, M. Torresan, 'Rotating and swirling binary black hole system balanced by its gravitational spin-spin interaction', Physical Review D 112 (2025) 064034, DOI 10.1103/3954-766n (the APS-registered DOI for this article), arXiv:2502.08706, https://arxiv.org/abs/2502.08706
21. V. S. Manko, E. Ruiz, 'Black hole-naked singularity dualism and the repulsion of two Kerr black holes due to spin-spin interaction', Physics Letters B 791 (2019) 26-29, DOI 10.1016/j.physletb.2019.02.025, arXiv:1803.03301, https://arxiv.org/abs/1803.03301
22. J. A. Rueda, V. S. Manko, E. Ruiz, J. D. Sanabria-Gómez, 'The double-Kerr equilibrium configurations involving one extreme object', Classical and Quantum Gravity 22 (2005) 4887-4894, DOI 10.1088/0264-9381/22/22/013, arXiv:gr-qc/0508101, https://arxiv.org/abs/gr-qc/0508101
23. H. Elvang, P. Figueras, 'Black Saturn', Journal of High Energy Physics 0705 (2007) 050, DOI 10.1088/1126-6708/2007/05/050, arXiv:hep-th/0701035, https://arxiv.org/abs/hep-th/0701035
24. D. Christodoulou, S.-T. Yau, 'Some remarks on the quasi-local mass', in Mathematics and General Relativity, Contemporary Mathematics 71, American Mathematical Society (1988), pp. 9-14, DOI 10.1090/conm/071/954405, MR 954405, https://doi.org/10.1090/conm/071/954405
25. M.-T. Wang, S.-T. Yau, 'Quasilocal mass in general relativity', Physical Review Letters 102 (2009) 021101, DOI 10.1103/PhysRevLett.102.021101, arXiv:0804.1174, https://arxiv.org/abs/0804.1174
26. P.-N. Chen, M.-T. Wang, Y.-K. Wang, S.-T. Yau, 'Conserved quantities in general relativity - the view from null infinity', arXiv:2204.04010 (2022), https://arxiv.org/abs/2204.04010
27. B. Zhao, S.-T. Yau, L. Andersson, 'A proposal of quasi-local mass for 2-surfaces of timelike mean curvature', arXiv:2407.00593 (2024), https://arxiv.org/abs/2407.00593