Orbital angular momentum mode-mixing for Laguerre–Gaussian beams scattered by Kerr black holes

My new computational study maps the OAM scattering channel of Kerr in painstaking detail. Here is what it found, why it matters, and where it falls short.

Orbital angular momentum mode-mixing for Laguerre–Gaussian beams scattered by Kerr black holes

https://zenodo.org/records/19830674

Presentation

https://www.youtube.com/watch?v=uxgYcBmSQVs

So here is something that might surprise you if you have not thought about it before: light can be twisted.

Not in the polarization sense, where the electric field rotates as the wave moves. Something deeper. The phase fronts of a beam (the surfaces of constant phase, the things that look like rippling sheets in a textbook diagram) can wind around the beam's propagation axis like a corkscrew. The number of times they wind is an integer, it is conserved as the beam propagates through empty space, and it is called the topological charge. Physicists also call it orbital angular momentum, or OAM. A pulse of OAM light is sometimes called a vortex photon.

Allen and collaborators worked this out in a famous 1992 paper, and OAM has since become a real workhorse in optics labs. People use it for high-bandwidth communication, quantum information protocols, and, more recently, astronomy.

Which brings us to the question. What happens to a twisted beam of light when it scatters off a spinning black hole? Could it be possible that the missing information from the black hole information paradox might be encoded on OAM modes or light?

That is what this new preprint submitted to Physical Review D. sets out to approach in quantitative detail. And the answer turns out to be more interesting than either "nothing" or "the black hole eats the twist."

The seed for this whole line of inquiry is a 2011 Nature Physics paper by Tamburini, Thidé, Molina-Terriza, and Anzolin. They argued that a Kerr (rotating) black hole should imprint OAM on light passing through its near-zone, because of frame dragging. Spacetime around a spinning black hole literally rotates with the hole, and that rotation should leave a signature on the phase structure of any light that gets close.

If true, this opens up a tantalizing observational possibility. The Event Horizon Telescope and its planned successor, the next-generation EHT, are imaging the photon rings of supermassive black holes like M87* and Sgr A*. If those photon rings carry distinctive OAM signatures tied to the central black hole's spin, an OAM-resolved measurement could become a new probe of the geometry, complementary to standard imaging.

But the 2011 prediction was qualitative. Geometric optics. Order-of-magnitude.

The paper does the boring, hard, necessary thing: it computes the actual scattering, in the full-wave Teukolsky regime, on a 5,250-point grid covering most of the physically relevant parameter space (spin, frequency, input topological charge). And then it asks four very specific questions about what falls out.

Question one: does the twist survive?

The first thing I do is check whether the topological charge of the incident beam matches the topological charge of the outgoing scattered field.

Across all 4,985 grid points where the calculation converges, the deviation between input and output topological charge is at the floating-point machine precision floor. Median deviation: zero. Maximum deviation: about 1.78×10−151.78 \times 10^{-15} 1.78×10−15, which is just numerical noise from double-precision arithmetic.

In other words, twisted light stays twisted. If you send in a beam with topological charge 5, you get back a scattered field with topological charge 5. The black hole does not break the twist.

This is not actually a discovery so much as a certification. The Kerr metric has an axial Killing vector (it is symmetric under rotations around the spin axis), and the scattering equations decompose into modes labeled by a discrete azimuthal quantum number. So azimuthal phase winding is preserved analytically by the structure of the problem. What the calculation shows is that no convention error or implementation bug has crept in to break this property numerically. It is a sanity check on the numerical pipeline, not a physical finding. Azimuthal symmetry guarantees this analytically, so a deviation would have indicated a convention or implementation bug rather than new physics.

Some prior speculative work has suggested vorticity-shifting effects in strong-field gravitational interactions. This calculation is silent on those proposals, since they require breaking axial symmetry or going beyond classical single-frequency scattering. Within the regime the paper studies (axisymmetric, single-frequency, classical electromagnetic scattering), no such shift exists.

Question two: does the energy stay in one mode?

Here is where it gets more interesting. The topological charge is preserved, but that is just the integer that labels the helical phase. It does not say anything about how the energy of the beam is distributed across the spheroidal multipole tower.

I track the centroid of the OAM spectrum, written ⟨l⟩\langle l \rangle ⟨l⟩, which is the first moment of the mode-mixing distribution. What he finds is that ⟨l⟩\langle l \rangle ⟨l⟩ is always greater than or equal to the input linl_\text{in} lin​. The shift is strictly upward. Frame dragging plus finite-frequency wave-optics effects pump power from the input multipole into higher members of the tower, never lower.

The size of this shift is set by the dimensionless combination aωa\omega aω, where aa a is the spin parameter and ω\omega ω is the frequency of the wave. When aω∼O(1)a\omega \sim O(1) aω∼O(1), the centroid can shift by 5 or more units of ll l. When you are far from that regime, the shift is suppressed below 0.5.

There is also a saturation. For input topological charge lin≥7l_\text{in} \geq 7 lin​≥7, the centroid shift caps out at around 1.5 to 2, almost regardless of where you are in the spin-frequency plane. The reason is the centrifugal barrier. The angular momentum potential goes like l(l+1)/r2l(l+1)/r^2 l(l+1)/r2, and at high ll l this gets steep enough that mixing into modes far above linl_\text{in} lin​ is suppressed by tunneling factors that do not really care about Kerr parameters. The light gets bumped up by one or two units, and that is about it.

Question three: what does the dominant mode do?

If you watch the dominant multipole lpeakl_\text{peak} lpeak​ (the single multipole carrying the most power in the outgoing field) as you sweep across the spin-frequency plane, it does not vary smoothly. It moves in a staircase. For lin=1l_\text{in} = 1 lin​=1, the dominant mode jumps from l=1l=1 l=1 to l=2l=2 l=2 to l=3l=3 l=3 and so on up to about l=8l=8 l=8, in sharp bands as you crank up the spin and frequency.

The bands have to be sharp because lpeakl_\text{peak} lpeak​ is by definition an integer. But the underlying physics here is that Kerr scattering pumps power through the multipole tower in unit-of-ll l increments. The spin and frequency together determine how many steps the dominant power has climbed.

This is the kind of structure you do not get from a smooth geometric-optics calculation. It is a wave-optics signature, full stop, and it lives in a regime where the eikonal description fails.

Question four: is the calculation actually right?

The total mode-mixing norm (the unnormalized power in the outgoing field, summed over modes) shows a sharp ridge in parameter space where the amplitude rises by one to two orders of magnitude.

The ridge tracks a specific curve: ω≈mΩH\omega \approx m \Omega_H ω≈mΩH​, where ΩH\Omega_H ΩH​ is the angular velocity of the black hole horizon and m=linm = l_\text{in} m=lin​ is the azimuthal quantum number. This is the superradiant threshold, the boundary across which a corotating mode gets amplified at the expense of horizon angular momentum.

The fact that the ridge sits exactly there, and migrates correctly as linl_\text{in} lin​ changes, is a non-trivial cross-check that the calculation is wired up correctly. Spin-1 superradiance is a known phenomenon (the maximum amplification factor is about 4.4% for electromagnetic perturbations near extremality), and this pipeline reproduces its location in parameter space without it being put in by hand.

So what does this all mean?

Here is where I want to be careful, because this paper is doing something important but limited, and the limits matter.

What it gives us is a quantitative reference dataset. If you want to know what the full-wave OAM scattering matrix of a Kerr black hole looks like in the parameter range relevant to photon rings, you can now go look it up. Geometric-optics calculations can be checked against it. Future observations can be compared to it. Lab analog experiments (acoustic superradiance from rotating absorbers, for instance, like the Cromb et al. 2020 result) can be benchmarked against scaled versions of these predictions.

What it does not give us is a direct path to observing Kerr OAM signatures with the EHT. At the relevant frequencies (230 GHz, 345 GHz) and for a black hole as massive as M87*, the dimensionless frequency Mω is around 10¹⁶, which is far above the parameter range the paper explores meaning EHT operates in the deep geometric-optics regime (Mω ≫ 1), not the long-wavelength wave-optics regime where these mode-mixing effects are pronounced. Direct detection would need either coherent integration across very broad frequency ranges or new OAM-resolved interferometric techniques that do not yet exist.

It also does not, by itself, tell us anything about the black hole information paradox or about whether OAM channels carry quantum information through Hawking radiation. The classical OAM channel exists, and it is well-behaved, and that is a necessary condition for any soft-hair information-channel proposal to work. But necessary is not sufficient. The quantum question stays open.

Twisted light bouncing off a spinning black hole comes out twisted in the same way it went in, but with its energy reshuffled along a discrete staircase whose steps you can predict. That is a clean, solid result. The follow-up questions (geometric-optics comparisons, gravitational waves, quantum information) are now sitting on top of an actual computational foundation rather than a pile of order-of-magnitude estimates.

Reference: T. Nestor, "Orbital angular momentum mode-mixing for Laguerre-Gaussian beams scattered by Kerr black holes," April 2026.