Negative-energy field disperses instead of exposing a singularity

A hypothetical field with negative energy might seem a promising way to challenge cosmic censorship—the idea that singularities produced by gravitational collapse are generally hidden behind black-hole horizons. But numerical experiments by Giovanni Caridi, Fabrizio Corelli and Paolo Pani found a different outcome: in every configuration successfully evolved, the field dispersed rather than forming an exposed singularity. Their study is available as an arXiv preprint, with the November 6 revised version listed as accepted for publication in Physical Review D.

Penrose’s weak cosmic censorship conjecture concerns the predictability of spacetime outside a collapsing object. An event horizon prevents a singularity from causally influencing distant observers. A naked singularity, lacking that protection, would threaten the predictability of classical evolution. The conjecture remains unproven in full generality, and its usual formulation assumes regular starting conditions and restrictions on matter’s energy properties.

The researchers deliberately abandoned one of those restrictions, the dominant energy condition. They studied a massless phantom scalar field, whose kinetic term has the opposite sign from that of an ordinary scalar field, giving it negative energy density. This was a theoretical experiment, not an observation of such matter. In principle, a negative-mass Schwarzschild geometry has a singularity without an event horizon. The question was whether a smooth field could dynamically produce that geometry, or another exotic outcome, rather than whether such a mathematical solution exists.

To investigate, the team evolved the coupled equations governing the scalar field and spacetime geometry in spherical symmetry. Their starting configurations were smooth, localized wave packets in an asymptotically flat spacetime—one approaching flat geometry far away. They tested both Gaussian pulses and broader, step-like profiles with smoothed edges, varying amplitudes and examining different widths. The initial waves were predominantly inward-moving. Fourth-order finite differences handled spatial derivatives, while a fourth-order Runge–Kutta method advanced the evolution in time.

Several checks helped distinguish physical behavior from computational errors. The researchers monitored curvature through the Kretschmann scalar, a coordinate-independent measure, and tracked the Misner–Sharp mass, which measures gravitational mass within a sphere. They also examined outgoing radial light propagation for signs of trapped regions. Tests of the gravitational constraint equations showed the expected fourth-order convergence, including near maximum compression in a high-amplitude run. As a benchmark, an ordinary positive-energy scalar field reproduced the expected threshold separating dispersion from black-hole formation.

The phantom field behaved differently. As inward-moving pulses compressed, curvature increased temporarily but then declined as the field bounced or spread outward. Stronger pulses produced more intricate sequences of outgoing curvature features. In the largest-amplitude example discussed in detail, part of the packet was prevented from reaching the center and moved outward instead. The authors interpret this behavior as the repulsive gravitational response associated with negative energy counteracting the initial focusing. They found no evidence of trapped surfaces, naked singularities or stationary exotic remnants in the configurations studied.

An important complication initially resembled a more dramatic result. At sufficiently high amplitudes, some simulations failed and showed rapidly growing curvature. However, the amplitude at which failure occurred shifted upward when the time step was reduced. Strong changes in the geometry increased the coordinate speeds of the simulated waves, making previously adequate time steps unstable. These are speeds expressed in the chosen coordinates, not evidence of light exceeding its local physical speed. In one example, reducing the time step by a factor of four changed an apparent curvature blow-up into a regular evolution followed by dispersion.

The result has clear boundaries. The simulations concern a massless scalar field, spherical symmetry, two families of initial profiles and a finite explored parameter range. They do not prove cosmic censorship, even for negative-energy matter generally. The chosen Schwarzschild-like coordinates also made strong-field evolution computationally demanding; an independent calculation using horizon-penetrating coordinates remains future work. Within those limits, the study finds that violating a standard energy assumption did not dynamically expose a singularity: dispersion prevailed.