WorksPhysics and Astronomy
Kalb-Ramond black bounces: throat-blind shadows and throat-sensitive neutrinos
Abstract
Regularizing a black hole removes its singularity, but which observations could detect the regularized core? We address this question for a Simpson-Visser black bounce built on the power-law hairy black hole that a Lorentz-violating Kalb-Ramond background produces. Measured in the areal radius, the bounce leaves the metric function untouched and changes only the radial measure. This single fact sorts the observables into two classes. The shadow, the innermost stable orbit, orbital frequencies and gravitational redshifts are blind to the throat. The instability of the photon ring, the brightness of its subrings, the oscillation phase of neutrinos and the energy deposited by neutrino pair annihilation respond to it through one universal factor. We derive an exact law for the subring demagnification that holds for any metric function. We also show that the throat enters the weak-field neutrino phase and the light deflection with different weights, so that combined photon and neutrino lensing can separate the mass from the throat. Ray-traced polarized images of a turbulent thin disk confirm that the direct image hides the throat while the first subring reveals it. For neutrinos emitted near a stellar-mass compact object the throat reshapes the flavor survival pattern and enhances the annihilation energy deposition. Two independent computations reproduce each headline number.
Introduction
The images of M87* and Sagittarius A* released by the Event Horizon Telescope (EHT) have turned the size of the black-hole critical curve into a measured quantity [1, 2], and the polarimetric maps of M87* have added the pattern of the electric vector position angle around the ring [3]. These data test the metric function that controls circular photon orbits. They are much less sensitive to the structure of the spacetime inside the photon sphere, and this is precisely where proposals for singularity resolution put their new physics. The Simpson-Visser construction is the cleanest example [4]: replacing the areal radius by turns the Schwarzschild singularity into a throat of radius while leaving the shadow of the black hole untouched. Lensing [5, 6, 7, 8, 9], thin-disk images [10] and geodesic completeness [11] of such black bounces have been studied in detail, and their charged version has a known Reissner-Nordström limit [12]. Their field sources in general relativity require a phantom scalar together with a nonlinear electromagnetic field [13].
A second line of work concerns spontaneous Lorentz symmetry breaking in the gravitational sector. The antisymmetric Kalb-Ramond field [14], with a nonzero vacuum expectation value coupled nonminimally to the Ricci tensor, is a natural candidate within the Standard-Model Extension [15]. Lessa and collaborators found that a radial pseudo-electric expectation value produces an exact power-law hair, , controlled by the product of the coupling and the norm of the background [16]. Other choices of the Lagrange multiplier lead to different families [17], and rotating and charged versions have been used for shadows and photon dynamics [18, 19]. Lorentz-violating gravity also admits regular and geodesically complete black holes of its own [20, 21].
Neutrinos are the second messenger that can reach the inner region. Their oscillation phase accumulated along a curved null path was derived in the Schwarzschild metric long ago [22, 23], generalized to arbitrary static backgrounds [24], and extended to spin precession [25]. When a lens creates two or more paths, the flavor probability acquires interference terms sensitive to the absolute neutrino masses and to the mass ordering [26], with decoherence effects analyzed in Ref. [27]. This program has since been carried through a long list of geometries [28, 29, 30, 31, 32], and the possibility of lensed astrophysical neutrinos has been raised observationally [33]. The pair-annihilation process above neutrinospheres and accretion disks offers a third handle, since its energy deposition is strongly amplified by gravity [34, 35, 36] and it is sensitive to modified gravity [37, 38, 39]. Most directly relevant here, Shi and Araújo Filho studied annihilation, oscillation phases and lensed flavor probabilities in the Kalb-Ramond black hole [40, 41], in bumblebee gravity [42, 43], and Lorentz-breaking corrections to the annihilation jets of gamma-ray bursts were discussed in Ref. [44]. The Simpson-Visser metric has been confronted with the deposition rate as well [45].
The two ingredients are complementary. The Kalb-Ramond hair is an infrared modification controlled by a background tensor that pervades the exterior, while the throat is an ultraviolet modification of the core that leaves the exterior metric function intact. A geometry carrying both is the natural testing ground for a question that neither ingredient raises alone: which observables respond to the core and which only to the exterior, and can a single observing campaign tell them apart? Photons answer part of this question through the photon ring, whose subrings are governed by the instability of the bound photon orbit [46, 47, 48] and whose polarization follows a pattern fixed by the near-ring geometry [49, 50, 51]. Neutrinos answer another part, because the oscillation phase is an integral over the whole path rather than a property of a single orbit.
These studies treat each observable one at a time. None of them separates the information carried by as a function of the areal radius from the information carried by the radial measure, yet for a Simpson-Visser regularization the first is exactly unchanged and the second carries all the dependence on the throat. Whether neutrino phases and annihilation can see a throat that the shadow cannot, and how that compares with what the photon subrings and polarization can do, has therefore not been tested.
This paper fills that gap for the Kalb-Ramond black bounce. Three technical elements make the comparison possible. We show that in any static spherically symmetric spacetime the polarization of a ray can be transported with two conserved components, without a Walker-Penrose constant, so that polarized images of the bounce are computed exactly. We weight neutrino paths by the ratio of emitted to received solid angle rather than by equal amplitudes, which matters once paths wind around the photon sphere. And we identify the proper volume element as the only route by which the throat enters the annihilation rate. Section II sets up the field theory, the regularized metric and its effective source. Section III proves the split between throat-blind and throat-sensitive observables and derives the weak-field coefficients, including the combination of photon deflection and neutrino phase that separates the mass from the throat. Section IV presents polarized ray-traced images and their subrings, Sec. V the neutrino phases and multipath flavor probabilities, and Sec. VI the annihilation deposition. Section VII translates the results into observational windows, and Sec. VIII concludes.
Kalb-Ramond hair and its Simpson-Visser regularization
We use the signature and units ; in all numerical work lengths are measured in units of the mass . The construction has two steps, the derivation of the hairy solution in the Kalb-Ramond theory and the regularization of its center, and we keep them apart because the second step is not a solution of the field equations of the first.
Field equations and the power-law hair
The Kalb-Ramond field is a two-form with field strength [14]. A self-interaction with a nonzero vacuum value breaks local Lorentz symmetry spontaneously [15]. With a nonminimal coupling to curvature the action reads [16]
with and couplings of dimension length squared. On the vacuum the term is a constant rescaling of and drops out. For a static spherically symmetric metric the pseudo-electric configuration
has constant norm and vanishing field strength. Varying Eq. (1) with respect to the metric gives , whose and components reduce to , where
For these give , and the angular equation becomes
The trial power yields , so the general solution is [16]
where has dimension length. For the hair has the Reissner-Nordström fall-off with in place of , and for it mimics a cosmological constant. We adopt , so , as the fiducial value: the hair then decays faster than any electromagnetic charge and the solution is not a reparametrization of a known charged geometry.
Regularization and horizon structure
The metric (5) has a curvature singularity at . Following Simpson and Visser [4], we replace the areal radius by everywhere,
with and a throat at where . Since , the metric function and all its derivatives remain bounded and every curvature invariant is finite on the throat. We call the line element (6) the SV-KR geometry. Noted that serves as a cross-check against the charged-like black bounce [12]. Changing the radial coordinate to gives the form that organizes the rest of the paper,
The metric function in terms of the areal radius is identical to the Kalb-Ramond one, and the bounce appears only through .
Horizons are the roots with . For the hair produces an inner horizon, and the two merge at the extremal point , which gives
that is, for and for . For and the horizons are and , because factorizes. The geometry is a black bounce when and a wormhole otherwise, and we restrict the analysis to the black-bounce branch. Figure 1a shows the metric function across the throat: for the hair makes diverge at the center, while any turns it into a smooth even function of .
Figure 1. (a) Metric function of the SV-KR geometry with across the throat for and increasing throat radius, together with the hairless bounce of radius ; the unregularized solution diverges at the center, the regularized ones stay finite. (b) Contours of the subring demagnification exponent in the plane of throat radius and hair, each labeled by its value. The hatched region is the wormhole branch, the solid curve separates black bounces from the wormhole branch and the dashed line marks the extremal hair . Circles mark the four geometries used for the images.
Effective source and energy conditions
The regularized metric is not a vacuum solution of the Kalb-Ramond equations, and it is useful to see exactly what it requires. Reading , and from Eq. (6), a direct computation gives
For these reduce to , the effective Kalb-Ramond source of the hair [16]. The combination that controls the null energy condition takes a simple form,
independent of the hair except through . The null energy condition is violated everywhere outside the horizon and its violation is set by the throat alone, exactly as in the hairless bounce. A field theory producing Eq. (6) must therefore combine the Kalb-Ramond sector with a phantom component of the type found for the Simpson-Visser geometry in general relativity [13]. In what follows we treat Eq. (6) as an effective geometry and study its test-particle and test-field content.
Two further properties of the regularized geometry are worth recording. First, since in the coordinate , the vacuum value of Eq. (2) keeps the constant magnitude along the whole bounce, so the background tensor is smooth through the throat, whereas in the areal coordinate its component grows as toward . Second, the curvature at the throat is finite. For the Kretschmann scalar at is
which for becomes and diverges only in the singular limit . The radial coordinate covers two asymptotically flat regions joined through the throat; in the black-bounce branch both sides are separated from the throat by horizons, and all observers considered below sit in the exterior.
Throat-blind and throat-sensitive observables
The form Eq. (7) implies a general statement that holds for any metric function , and the Kalb-Ramond hair is only one realization of it. For geodesics with conserved energy and angular momentum in the equatorial plane,
with for null and for timelike geodesics. The throat multiplies the radial equation by a positive function and does nothing else. Any condition stated as the vanishing of the bracket and of its derivatives is therefore unchanged, while any quantity that integrates along or depends on the curvature of the radial potential at a fixed point picks up .
Throat-blind quantities
The photon sphere solves , so it satisfies and the critical impact parameter is
Timelike circular orbits have
and the marginally stable orbit solves
Here prime denotes ordinary derivative w. r. to argument. Since multiplies the whole bracket of Eq. (14), the second derivative of the radial potential at a circular orbit only rescales, so the condition for the ISCO is also unchanged. The same holds for the redshift of a static emitter, , and for the redshift of a circular emitter, . The photon sphere, the shadow, the ISCO, the Keplerian frequency, the radiative efficiency and all local redshifts are throat-blind, provided the corresponding radius lies above . For the four fiducial geometries Table 1 lists and , which depend on the hair but are the same for and .
Throat-sensitive quantities
Expanding Eq. (14) around the photon sphere gives the Lyapunov exponent of the unstable orbit in coordinate time [48],
where is the value for . Successive photon subrings are demagnified by with and [46, 47]. Because is throat-blind, we obtain the exact law
valid for any metric function. For the hairless bounce it reads , and for it gives exactly. The radial epicyclic frequency is likewise rescaled, , while the vertical and azimuthal frequencies are untouched; this offers a second, independent probe of through quasiperiodic oscillations.
Path integrals pick up the factor . For a null ray of impact parameter with turning point , the swept angle and the radial integral that enters the neutrino phase are
Expanding for we find for the deflection angle
and for the regularized radial integral of one leg, ,
with hair coefficients
For these give , the Reissner-Nordström deflection coefficient with , and for they give and . The term of Eq. (22) reproduces the Simpson-Visser result of Ref. [5]. The derivation of Eqs. (22)–(24) is given in Appendix A.
The comparison of Eqs. (22) and (23) is the first quantitative result of this paper. The throat enters both expansions at second order, but with different weights: in the deflection it shifts , in the two-leg neutrino phase it shifts . Denoting the two second-order coefficients by and , one obtains
so a photon measurement and a neutrino measurement of the same lens separate mass and throat at this order. For the hair enters both coefficients only at higher order, and the inversion is hair-independent. For the hair adds to and to , and Eq. (25) then returns and instead of and ; the charged black bounce therefore carries a degeneracy between charge and throat that the Kalb-Ramond hair with avoids at this order.
The structure of the problem can be stated in one line. Every observable considered in this paper is either a functional of the metric function alone or an integral of the form
with a kernel that again depends on alone, or a local quantity proportional to at a fixed radius, as in Eq. (18). The placement of each observable in one of these classes, and the factor itself, are universal: they hold for any metric function and any throat. The Kalb-Ramond hair is model dependent and enters only through , via the exponent and the amplitude . This is the sense in which the paper separates a universal statement from a model-dependent one.
For timelike orbits the same logic yields the radial epicyclic frequency in closed form. Perturbing Eq. (14) around a circular orbit and converting to coordinate time,
which vanishes on the ISCO of Eq. (17) and reduces to for Schwarzschild. The periastron precession per orbit, , is therefore throat-sensitive, and the ratio of vertical to radial frequencies is shifted by at every radius, while the orbital frequency itself is not. For the S2 star around Sagittarius A*, whose pericenter lies near , the factor differs from unity by about for , so stellar orbits are effectively throat-blind and the inner disk is where the epicyclic channel would operate.
Figure 2(a) summarizes the split as a function of the throat radius. The critical impact parameter, the ISCO and the Keplerian frequency stay at their values, the two second-order weak-field coefficients grow at different rates, the subring exponent decreases according to Eq. (19), and the annihilation rate, discussed in § 1.5, rises. Panel (b) shows the emission radius of the -th equatorial crossing for a face-on observer: the direct image barely moves, while the and lensing bands widen, as the inset makes clear near .
Figure 2. (a) Ratios to the values of throat-blind quantities (solid black, the same curve for , and ), of the second-order neutrino phase and deflection coefficients, of the subring exponent and of the annihilation rate at , for and . (b) Emission radius of the -th crossing of the equatorial plane versus impact parameter for a face-on observer in the Schwarzschild metric and in the hairless bounce with . The inset magnifies the and bands near the critical impact parameter, marked by the dotted vertical line.
Table I also tests Eq. (19) independently. The widths of the lensing bands were obtained by integrating the orbit equation in the variable and locating the impact parameters at which a ray reaches the -th crossing exactly at the horizon or exactly at infinity. Their ratio converges to as grows, and at it already agrees with the closed form to within about one part in seventy for every geometry. The Schwarzschild row reproduces , , and , and the row reproduces the charged black bounce with , for which and .
| 0 | 0 | 2.00000 | 5.19615 | 6.00000 | 3.14159 | 0.04274 | 0.04321 |
| 0.8 | 0 | 2.00000 | 5.19615 | 6.00000 | 3.02783 | 0.04785 | 0.04842 |
| 1.5 | 0 | 2.00000 | 5.19615 | 6.00000 | 2.72070 | 0.06488 | 0.06583 |
| 1.75 | 0 | 2.00000 | 5.19615 | 6.00000 | 2.55170 | 0.07680 | 0.07795 |
| 0 | 0.5 | 1.85464 | 5.03934 | 5.71764 | 2.96628 | 0.05084 | 0.05149 |
| 1.5 | 0.5 | 1.85464 | 5.03934 | 5.71764 | 2.52071 | 0.07915 | 0.08040 |
| 1.75 | 0.5 | 1.85464 | 5.03934 | 5.71764 | 2.33905 | 0.09540 | 0.09642 |
| \multicolumn{8}{c}{} | |||||||
| 1.5 | 0.5 | 1.70711 | 4.70960 | 5.16646 | 2.37959 | 0.09262 |
Table I. Geometric data of the fiducial SV-KR geometries with and of the charged-type case , in units of . The last two columns compare the measured ratio of the widths of the and lensing bands with from Eq. (19).
Polarized images and photon subrings
The images of an optically and geometrically thin equatorial disk provide the cleanest test of the split, since the direct emission comes from throat-blind circular orbits while the lensed emission travels along throat-sensitive paths. This section describes the ray-tracing scheme, the treatment of polarization and the resulting images.
Rays, emitters and polarization transport
Null geodesics in a spherically symmetric spacetime are planar. In the orbital plane we integrate
in its second-order form, which is regular through turning points and starts exactly at with . For an observer at inclination and a pixel at screen radius and polar angle , the swept angle at the -th crossing of the equator is with and . A table of on a grid refined logarithmically toward maps every pixel to its emission radii.
Outside the ISCO the gas follows Keplerian orbits, Eq. (16), and inside it plunges with the ISCO energy and angular momentum. In the static frame the plunging velocity components are
which again depend only on and , so the entire emitter model is throat-blind and any change in the image comes from the light paths. The redshift is , with the photon direction in the static frame. Following [49, 50], the observed intensity is
with spectral index , the path length capped at four, and primed quantities in the fluid frame. The local polarization is .
Real accretion flows are not smooth. Magnetized turbulence, sheared by differential rotation, organizes the emission into thin trailing filaments, which dominate both simulated and observed images. To represent this in the displayed images we multiply by a stochastic texture of unit mean,
normalized to unit variance, with 260 modes of random phase, between 14 and 130, within two units of for a pitch angle , amplitudes and contrast . Since all wavevectors are nearly parallel in the plane, the superposition forms long logarithmic-spiral filaments. The texture is a property of the emitter and is the same function of in every geometry, so it cannot mimic a throat effect. The polarimetric coefficients quoted below and in Table II refer to the smooth emissivity; with the texture, of the full image at shifts by about and by less than one hundredth, while the throat-induced shift between the Schwarzschild and SV-KR images stays at .
Transporting the polarization is simple in spherical symmetry. Let be the unit normal of the orbital plane and the in-plane transverse direction in the static frame. Reflection symmetry about the orbital plane keeps constant under parallel transport, the norm is constant, and both and are unchanged by the gauge shift . Hence is conserved as well, and the polarization on the screen is
where and are the radial and azimuthal unit vectors of the screen. This replaces the Walker-Penrose constant used for Kerr [49] and is exact for any static spherically symmetric metric. As a check, a static face-on disk threaded by a radial field gives a purely azimuthal electric pattern with , and a toroidal field gives . We quantify the pattern by the coefficient [51]
with measured east of north, and by the net fractional polarization .
Images
Figure 3 shows the four fiducial geometries at , appropriate for M87*, and at , for a fluid-frame field . The four columns look nearly identical, which is the visual form of the throat-blind statement: the critical curve is the same circle for the two hairless and for the two hairy geometries, and the direct image of the disk, which dominates the flux, comes from throat-blind emitters seen along paths that barely approach the throat. The polarization ticks spiral in the same sense in every panel. Near edge-on, the lensed image of the far side of the disk arches over the shadow and its underside appears below it, framed by the thin photon ring, while Doppler beaming makes the approaching side of the filaments far brighter than the receding side.
Figure 3. Polarized images of a thin, turbulent Keplerian disk at (top row) and (bottom row) for the Schwarzschild metric, the Kalb-Ramond hair with , the hairless bounce with and the SV-KR geometry with both. All panels of a row share one color scale, , with the 99.6th percentile of the Schwarzschild intensity, so that brightness differences between geometries are real. Ticks show the electric vector position angle with length proportional to the fractional polarization; outside the Schwarzschild column their color gives the rotation relative to the Schwarzschild image at the same pixel. The dotted white circle is the Schwarzschild critical curve and the dashed cyan circle the critical curve of the geometry shown. Insets show the azimuthally averaged intensity versus screen radius for Schwarzschild and, in the other columns, its ratio to Schwarzschild, with excess in red and deficit in blue; dotted and dashed vertical lines mark the two critical impact parameters. The hair moves the ring inward and piles intensity up just inside the Schwarzschild critical curve, while the throat raises the intensity between and through its wider lensing bands.
The similarity of the four columns hides real differences, which Figure 4 makes explicit by subtracting the Schwarzschild image pixel by pixel. The hair alone produces a residual concentrated on the critical curve, a positive and a negative ring that record the shrinking of from to . The throat alone leaves the critical curve in place but displaces the lensed filaments, since their paths are governed by Eq. (28): the residual appears as a sequence of interleaved red and dark arcs throughout the inner disk and in the lensed image of the far side. With both ingredients the two patterns add. Measured by the summed absolute residual normalized to the summed Schwarzschild intensity, the difference at is for the hair, for the throat and for both, and at it is , and , respectively. The fine structure of the disk is what makes the throat visible in these maps: a smooth emitter would hide most of the displacement, which is why flux-weighted quantities such as respond only weakly.
Figure 4. Residual maps of the images of Figure 3 relative to the Schwarzschild image, which is repeated in the first column, in a field of view of . Color shows with , a stretch that keeps weak residuals visible next to strong ones, at (top) and (bottom). Red marks an excess and graphite gray to black a deficit of intensity. The dotted circle is the Schwarzschild critical curve and the dashed blue circle that of the geometry shown; the residuals of the hair cluster on the critical curve, those of the throat spread over the lensed filaments.
Blurring with a circular Gaussian of full width , which corresponds to for M87*, removes the photon ring as a separate feature and erases these residuals. The central brightness depression survives in all four geometries and its size tracks the hair rather than the throat. The blurring lowers from to in the Schwarzschild case and from to in the SV-KR case, while the phase changes by and , respectively, so the blurred polarimetric pattern carries almost no trace of the throat.
The throat becomes visible when the image is decomposed by the number of equatorial crossings. Figure 5 separates the , and contributions for the Schwarzschild and the SV-KR geometries. The direct images differ little, but the band of the SV-KR case is broader on its inner side and both sub-rings are brighter. The integrated flux grows by a factor relative to Schwarzschild at and the flux by roughly a factor of two, consistent with the larger of Table I. Because the critical curve is fixed by the hair alone, a measured subring width at a given ring diameter isolates the throat through Eq. (19).
Figure 5. Decomposition of the image at into the direct emission , the first and second subrings and , and their sum for the Schwarzschild metric (top) and the SV-KR geometry with , (bottom). The emissivity carries the texture of Eq. (31), and each column uses the color scale of the Schwarzschild panel above it, , so the brighter SV-KR subrings are a physical effect. In the and panels the solid line traces the edge of the band of the geometry shown and the dashed line the band of the other geometry: the SV-KR band extends further inward. The bottom row shows the residual SV-KR minus Schwarzschild for each order with the stretch of Figure 4: the direct image shows displaced filaments, the residual an inner excess ring from the wider band, and the residual a bright pair of rings at the shifted critical curve. The subring panels use a narrower field of view; the ring, thinner than a pixel, is lightly smoothed and drawn without ticks.
The polarimetric counterpart is shown in Figure 6. Over the range the phase of of the full image moves by about and its modulus by about one part in seventy, both for the unblurred and the blurred images. The ring alone responds roughly five times more strongly: its phase moves by for and by for , and its modulus grows by three parts in a hundred. Table II extends the comparison to purely radial, vertical and toroidal fields. The toroidal field is nearly insensitive to the throat. The radial field shifts the phase by and the mixed field by at . The vertical field moves the phase of the full image by about , which is the largest response of the direct emission, because its polarization depends most strongly on the direction of the photon at the source. For the full-image phase at , is , close to the value, so the sensitivity of the polarization to the hair exponent is weak at fixed critical-curve size.
Figure 6. (a) Shift of the phase of relative to and (b) modulus of relative to , as functions of the throat radius at for the full image, the blurred image and the ring, for (circles, solid) and (squares, dashed).
| Field | Model | |||||
| mixed | S | 0.870 | −135.19 | 0.677 | −98.97 | 0.0423 |
| SV | 0.862 | −134.22 | 0.693 | −104.17 | 0.0414 | |
| SV-KR | 0.858 | −134.02 | 0.693 | −103.94 | 0.0412 | |
| radial | S | 0.961 | −163.63 | 0.914 | −141.83 | 0.0234 |
| SV | 0.958 | −162.96 | 0.909 | −144.46 | 0.0234 | |
| SV-KR | 0.956 | −162.82 | 0.906 | −144.44 | 0.0235 | |
| vertical | S | 0.673 | 89.72 | 0.644 | 143.81 | 0.1121 |
| SV | 0.670 | 92.47 | 0.609 | 142.30 | 0.1087 | |
| SV-KR | 0.666 | 92.80 | 0.601 | 141.64 | 0.1068 | |
| toroidal | S | 0.989 | 2.45 | 0.954 | 5.82 | 0.0200 |
| SV | 0.988 | 2.57 | 0.956 | 5.96 | 0.0208 | |
| SV-KR | 0.988 | 2.63 | 0.956 | 5.97 | 0.0207 |
Table 2. Polarimetric coefficient at for four fluid-frame field geometries, for the Schwarzschild metric (S), the hairless bounce with (SV) and the SV-KR geometry with , . Phases are in degrees; refers to the full image.
Neutrino phases and multipath flavor oscillations
Neutrinos produced near a compact object probe the radial integral of Eq. (21) directly, and they do so in a way that photons cannot, because the oscillation phase depends on the neutrino masses. This section derives the phase in the SV-KR geometry, checks its weak-field expansion and studies the interference of paths for a source close to the bounce.
Oscillation phase
A flavor state propagating from the source S to the detector D accumulates for each mass eigenstate the phase . Evaluated along the light ray and expanded to first order in , it becomes [22, 23, 24]
where is the energy at infinity, and the second form uses in the coordinate of Eq. (6). The radial measure of the phase is exactly the factor that enters the orbit equation. For radial propagation in the Schwarzschild metric the phase reduces to [22]. In the SV-KR geometry radial propagation gives in the coordinate , which corresponds to the proper areal path , so that a source at areal radius acquires the extra phase relative to the hairless and throatless case. Away from the radial direction the phase of one leg, regularized at infinity, is with of Eq. (23).
Figure 7(a) compares the exact quadrature of with the expansion of Eq. (23). The quantity approaches for the hairless geometries without throat and with throat, confirming the coefficient of . At the exact values are 2.3588 and 4.1267 against the asymptotic 2.3562 and 4.1233, and the residual falls as . The hair with enters only at order and is invisible on this scale, which is the weak-field version of the hair independence of Eq. (25). The quadrature itself was checked against an independent arbitrary-precision evaluation and agrees to ten digits.
Figure 7. (a) Exact second-order coefficient of the one-leg neutrino phase integral versus impact parameter; dotted lines show the asymptotic values and of Eq. (23), and the inset magnifies the range , where the hair lowers the coefficient of the throatless geometries. (b) Change of the radial-integral difference of the two primary paths relative to Schwarzschild for a source at , as a function of the detector direction. (c) Electron-neutrino survival probability for the normal ordering, a ten-solar-mass lens and MeV; the inset magnifies the region .
Multipath flavor probabilities
In the weak-lensing regime studied in Refs. [26, 29, 31], where the impact parameters are of the order of the Einstein radius, Eq. (23) shows that the throat changes the phase difference between paths by . This is of second order in and negligible for any realistic Einstein radius. The throat becomes accessible when the neutrinos are emitted close to the compact object, as in the inner region of a hyper-accreting flow or the neutrinosphere of a collapsar. We therefore place a point source at areal radius on the negative axis and a detector at infinity in the direction . A ray leaving the source with local emission angle , measured from the outward radial direction, has . It escapes either directly or after a turning point, provided in the second case, and it reaches the detector when its total swept angle equals modulo . For each such path we compute and the regularized radial integral , and assign the amplitude with
the ratio of emitted to received solid angle. Earlier studies assigned equal amplitudes to all paths, which overstates the contribution of the paths that wind around the photon sphere. The flavor transition probability is then
which is normalized by construction. We use the mixing angles , , , and the splittings , close to the global fit [52], a lightest mass of , a lens of ten solar masses () and .
Figure 7(b) shows the change of the path difference between the two primary paths, . The throat lengthens the difference monotonically as the detector moves toward the source side, reaching for near , while the hair shortens it by up to . For the ten-solar-mass lens the throat contribution is , which for at corresponds to a phase shift of . With both ingredients the shift is and the hair partly compensates the throat. At the four paths with have weights , , and in the Schwarzschild metric; in the SV-KR geometry the third and fourth paths are about twice and three times as bright, reflecting the smaller .
The resulting electron-neutrino survival probability is shown in Figure 7(c) for the normal ordering. The fast oscillation in is driven by the large path difference and by , and the slower envelope by . The throat shifts the fast pattern in phase, with differences of a few tenths in near , where the paths are most unequal. At the survival probability is 0.866 in the Schwarzschild metric and 0.918 in the SV-KR geometry for the normal ordering, and 0.879 and 0.904 for the inverted ordering. The ordering matters at the same level as the throat: at fixed geometry the two orderings differ by about one hundredth in at this angle, while the throat changes by five hundredths for the normal and by two and a half hundredths for the inverted ordering.
Pair-annihilation energy deposition
The third messenger is the energy deposited by above a neutrino-emitting surface. Here the throat enters through the proper volume, which makes this observable throat-sensitive at leading order rather than through path differences.
We model the emitter as an isothermal neutrinosphere of areal radius , above the photon sphere, with luminosity fixed at infinity. The local temperature and the angle subtended by the neutrinosphere depend only on and , and , and the local deposition rate is proportional to with [34, 37]. The rate measured at infinity integrates this over the proper volume weighted by the redshift, . Normalizing to the Newtonian rate at the same we obtain
with , and . The factor converts to the local temperature, and the only place where the throat enters is in the numerator. For small , the relative change of the rate is , with the average taken over the deposition profile.
The small- estimate is accurate. At in the hairless case the exact relative enhancement is for and for , against and from the small- estimate; at the corresponding pairs are against , and against . The deposition is concentrated near the emitter: about 0.99 of the total rate is deposited within for both radii, which explains why the throat effect follows the surface value of of the small- estimate because the deposition extends beyond the neutrinosphere. Panel (b) shows that this relative enhancement decays as , and that the hair with acts in the opposite direction, so that in the SV-KR geometry the two effects cancel near for and the throat wins farther out. At the SV-KR rate is 16.13 for and 16.95 for , against 14.40 for the hair alone. The independent evaluation of Eq. (37) with a different quadrature reproduces the SV-KR value at to fourteen digits.
Two remarks put these numbers in context. The enhancement over the Newtonian rate near the photon sphere is driven by and by the factor , which describe the gravitational blueshift of the neutrinos falling back toward the emitter; these factors are throat-blind and set the overall scale, as for the modified-gravity backgrounds of Refs. [37, 39, 45]. The throat adds a geometric factor on top of them, which is why it appears as a relative correction of order that is the same for any metric function. For a black-hole accretion torus rather than a spherical neutrinosphere the angular integration changes [35, 36], but the proper-volume factor, and hence the structure of the throat correction, does not.
Observational windows
The split of Sec. III dictates which data can constrain which parameter. The size of the critical curve constrains only the hair. Writing , the fiducial hair with gives for and at extremality, while negative hair increases the ring and reaches only at . An EHT window of one tenth in , of the size reported for Sagittarius A* [2], therefore leaves every positive hair up to extremality allowed. For the same window excludes . In both cases the throat radius is completely unconstrained by the shadow, as is evident from Eq. (15), and comparisons of black-hole mimickers with the Sagittarius A* image [53] cannot fix it either.
It is useful to quantify the size of the throat signals. For the hairless bounce the subring exponent decreases from to at , so the flux ratio of consecutive subrings rises from 0.043 to 0.066, a factor of 1.5; at fixed this is a large change. For the hairy bounce with the same throat moves from 0.051 to 0.080. In polarimetry, the phase shift of at is small compared with the spread between field geometries in Table II, which exceeds a hundred degrees, so a throat measurement through needs independent information on the field geometry; the full-image shift of about lies well below the uncertainties of present EHT polarization maps [3]. For neutrinos from a ten-solar-mass engine the throat changes the primary path difference by up to 8.8 km, while the corresponding arrival-time difference between paths, of order ms, sets the coherence time a wave packet would need to keep the interference terms of Eq. (36).
Three channels see the throat. The first is the ratio of successive subring fluxes or widths, which measures and, once the hair is known from the ring diameter, gives through Eq. (19). The universal interferometric signature of the photon ring [46] makes this an observable for long baselines, and Figure 1(b) shows that ranges from about 0.67 to 1.07 over the black-bounce domain. The second is the phase of restricted to the ring, which moves by about five degrees at , roughly five times more than the phase of the full image. The third is the neutrino channel, through either the multipath interference of § 1.4 for neutrinos from the inner accretion region, or the deposition enhancement of § 1.5, or, for a lens observed in both photons and neutrinos, the inversion (25).
Conclusions
We have shown that the Simpson-Visser regularization of the Kalb-Ramond power-law hair leaves the metric function, written in the areal radius, exactly as it was and moves all dependence on the throat into the radial measure . This single statement sorts the observables. The critical curve, the ISCO, Keplerian frequencies and local redshifts are throat-blind; the Lyapunov exponent, the subring demagnification, the radial epicyclic frequency, light deflection, the neutrino oscillation phase and the pair-annihilation deposition are throat-sensitive. We derived the exact law , the weak-field coefficients of deflection and phase, including the hair coefficients for any fall-off exponent, and the combination of photon and neutrino data that separates mass and throat. The effective source has a null-energy violation set by the throat alone, .
Figure 8. (a) Pair-annihilation energy deposition rate normalized to the Newtonian rate at fixed luminosity at infinity, as a function of the neutrinosphere radius; the inset magnifies the region near the photon sphere. (b) Relative change of the rate with respect to the Schwarzschild metric.
The polarized images confirm the split. The direct image and the full polarimetric pattern barely respond to the throat, while the ring widens and its phase moves by about to over the black-bounce range. For neutrinos emitted at from a stellar-mass lens, the throat changes the interfering path difference by up to and the flavor survival pattern by order unity, and near the photon sphere it raises the annihilation deposition, an effect that the positive hair partly offsets.
All benchmarks were recovered: the Schwarzschild values , , , , the Simpson-Visser deflection coefficient, the Reissner-Nordström deflection coefficient for , the charged black bounce for , the radial phase of Fornengo and collaborators, and the Newtonian limit of the deposition. The closed forms and an independent kernel evaluation agree to ten significant figures for the geometric quantities and the phase integral and to fourteen for the deposition, while the band-width ratios approach to two significant figures at . Our neutrinos are test particles described by plane waves along geometric-optics paths, coherence between paths with different arrival times is assumed rather than derived, the disk is optically and geometrically thin, and only the static black-bounce branch is treated, so none of the results bears on ringdown or gravitational-wave signals.
In the near future, we plan to extend the analysis to the rotating Kalb-Ramond bounce, where the critical curve itself is no longer circular and the split between throat-blind and throat-sensitive quantities must be reformulated with the Carter constant. A wave-packet treatment of the multipath neutrino amplitudes, which fixes how much interference survives the time delays between paths, and a computation of the radial epicyclic frequencies for quasiperiodic oscillations are the other two tasks on our agenda.
Acknowledgments
F.A. acknowledges the Inter University Centre for Astronomy and Astrophysics (IUCAA), Pune, India, for granting a visiting associateship. I.S. is grateful to Eastern Mediterranean University (EMU), the Scientific and Technological Research Council of Türkiye (TÜBİTAK), the Turkish Academic Network and Information Center (ANKOS) and the Sponsoring Consortium for Open Access Publishing in Particle Physics (SCOAP3) for their support. He also acknowledges the networking support of COST Actions CA18108, CA22113, CA21106, CA23130, CA23115, and CA21136.
Author contributions
F.A. conceived the study and carried out the geodesic, lensing and weak-field calculations. A.A. contributed to the analysis of the shadow, the deflection angle and the neutrino phases, and checked the calculations. İ.S. carried out the ray-traced imaging, polarimetry and pair-annihilation analysis and prepared the figures. All authors discussed the results, wrote the manuscript via Overleaf, and approved its final version.
Data availability
No observational data were generated in this study. The computational worksheets that produce the figures and tables are available from the corresponding author upon reasonable request.
Weak-field coefficients
We sketch the derivation of Eqs. (22)–(24). Writing and , the integrals (20) and (21) become
with . The subtraction in Eq. (39) removes the flat-space path , and the difference between and vanishes at infinity. Since vanishes linearly at , the substitution makes both integrands regular. The throat contribution follows at once: to first order in ,
which doubles to the term of in Eq. (22) and gives the term of Eq. (23). For the hair, a term in changes by to first order. Expanding the integrands and integrating with gives the coefficients of Eq. (24), and the ratio comes out of the same expansion. The mass terms are obtained in the same way to third order in , after re-expressing in terms of .
We verified each coefficient against direct quadrature with thirty-digit arithmetic. For the numerical hair coefficients at are , and , against the closed forms , and ; the coefficient of in the deflection is against ; and the Schwarzschild third-order coefficient approaches .
Numerical methods and convergence
The geometric quantities of Table I were obtained by root finding on Eqs. (15)–(17) with a tolerance of and checked against an independent symbolic and numerical kernel, which agreed to ten significant figures for , , , , and in the five geometries checked, including the case. The band widths were computed by integrating Eq. (28) with an eighth-order Runge–Kutta scheme at relative tolerance and bracketing the band edges to in .
The images use a table of with about 1400 impact parameters, 520 of them distributed logarithmically within of down to , and 3600 values of up to , so that crossings up to are resolved. Pixel grids of points were averaged in blocks of four to suppress aliasing of the thin rings. For the SV-KR geometry at , changes from to and for grids of , and pixels, and stays at to four digits, so the image coefficients quoted in the text are converged to about in phase and in modulus.
The neutrino integrals were evaluated with the substitution of Appendix A and adaptive quadrature. Against thirty-digit arithmetic the one-leg integral for the SV-KR geometry agrees to at and to at . For the multipath problem a fan of about 3000 emission angles, refined logarithmically toward the critical angle, was interpolated with monotone cubic splines in , which gives the emission angle, the radial integral and the solid-angle weight of each path. The deposition ratio was integrated to ; raising the cutoff to 4000 changes it by less than .
References
- [4]A. Simpson and M. Visser, JCAP 2019 (02), 042, 1812.07114.
- [5]J. R. Nascimento, A. Y. Petrov, P. J. Porfirio, and A. R. Soares, Phys. Rev. D 102, 044021 (2020).
- [6]N. Tsukamoto, Phys. Rev. D 103, 024033 (2021).
- [7]N. Tsukamoto, Phys. Rev. D 104, 064022 (2021).
- [8]S. U. Islam, J. Kumar, and S. G. Ghosh, JCAP 2021 (10), 013.
- [9]G. He, Y. Xie, C. Jiang, and W. Lin, Phys. Rev. D 110, 064008 (2024).
- [10]M. Guerrero, G. J. Olmo, D. Rubiera-Garcia, and D. S.-C. Gomez, JCAP 2021 (08), 036.
- [11]K. Pal, K. Pal, and T. Sarkar, Gen. Relativ. Gravit. 55, 121 (2023).
- [12]E. Franzin, S. Liberati, J. Mazza, A. Simpson, and M. Visser, JCAP 2021 (07), 036, 2104.11376.
- [13]K. A. Bronnikov and R. K. Walia, Phys. Rev. D 105, 044039 (2022), 2112.3198.
- [14]M. Kalb and P. Ramond, Phys. Rev. D 9, 2273 (1974).
- [15]B. Altschul, Q. G. Bailey, and V. A. Kostelecký, Phys. Rev. D 81, 065028 (2010), 0912.4852.
- [16]L. A. Lessa, J. E. G. Silva, R. V. Maluf, and C. A. S. Almeida, Eur. Phys. J. C 80, 335 (2020).
- [17]K. Yang, Y.-Z. Chen, Z.-Q. Duan, and J.-Y. Zhao, Phys. Rev. D 108, 124004 (2023), 2308.06613.
- [18]R. Kumaar, S. G. Ghosh, and A. Wang, Phys. Rev. D 101, 104001 (2020).
- [19]E. Sucu and I. Sakalli, Phys. Rev. D 111, 064049 (2025).
- [20]R. Carballo-Rubio, F. D. Filippo, S. Liberati, and M. Visser, JHEP 2022 (02), 122.
- [21]J. Mazza and S. Liberati, JHEP 2023 (03), 199.
- [22]N. Fornengo, C. Giunti, C. W. Kim, and J. Song, Phys. Rev. D 56, 1895 (1997).
- [23]C. Y. Cardall and G. M. Fuller, Phys. Rev. D 55, 7960 (1997).
- [24]L. Visinelli, Gen. Relativ. Gravit. 47, 62 (2015).
- [25]M. Bruggen, Phys. Rev. D 58, 083002 (1998).
- [26]H. Swami, K. Lochan, and K. M. Patel, Phys. Rev. D 102, 024043 (2020).
- [27]H. Swami, K. Lochan, and K. M. Patel, Phys. Rev. D 104, 095007 (2021).
- [28]S. Capozziello, G. Iovane, and G. Lambiase, Nuovo Cim. B 115, 545 (2000).
- [29]Y.-R. Wang, Z.-W. Li, and S.-J. Rong, Int. J. Mod. Phys. A 41, 2650133 (2020).
- [30]H. Chakrabarty, A. Chatrabhuti, D. Malafarina, B. Silasan, and T. Tangphati, JCAP 2023 (08), 018.
- [31]M. Alloulokov, H. Chakrabarty, D. Malafarina, B. Ahmedov, and A. Abdujabbbarov, JCAP 2025 (02), 070.
- [32]S. Mannobova, B. Narziloev, F. Atamurotov, A. Abdujabbarov, I. Hussain, and B. Ahmedov, Phys. Dark Universe 51, 102194 (2026).
- [33]S. Britzen, Z. Younsi, E. Kun, I. N. Paschenko, F. Jaron, A. Tursunov, and H. Olivares, Mon. Not. Roy. Astron. Soc. 549, stag658 (2026).
- [34]K. Asano and T. Fukuyama, Astrophys. J. 531, 949 (2000).
- [35]K. Asano and T. Fukuyama, Astrophys. J. 546, 1019 (2001).
- [36]R. Birkl, M. A. Aloy, H.-T. Janka, and E. Mueller, Astron. Astrophys. 463, 51 (2007).
- [37]G. Lambiase and L. Mastrototaro, Astrophys. J. 904, 19 (2020).
- [38]G. Lambiase and L. Mastrototaro, Eur. Phys. J. C 81, 932 (2021).
- [39]G. Lambiase and L. Mastrototaro, Astrophys. J. 934, 12 (2022).
- [40]Y. Shi and A. A. A. A. Filho, JHEP 2025 (08), 28.
- [41]Y. Shi, A. A. A. A. Filho, K. E. L. de Farias, V. B. Bezerra, and A. R. Queiroz, Eur. Phys. J. Plus 141, 1038 (2026).
- [42]Y. Shi and A. A. A. A. Filho, JCAP 2025 (11), 045.
- [43]Y. Shi and A. A. A. A. Filho, Ann. Phys. 490, 170496 (2026).
- [44]M. Khodadi, G. Lambiase, and L. Mastrototaro, Eur. Phys. J. C 83, 239 (2023).
- [45]A. A. A. Filho, Eur. Phys. J. C 84, 73 (2024).
- [46]M. D. Johnson et al., Sci. Adv. 6, eaaz1310 (2020), 1907.04329.
- [47]S. E. Gralla, D. E. Holz, and R. M. Wald, Phys. Rev. D 100, 024018 (2019), 1906.00873.
- [48]V. Cardoso, A. S. Miranda, E. Berti, H. Witek, and V. T. Zanchin, Phys. Rev. D 79, 064016 (2009), 0812.1806.
- [49]E. Himwich, M. D. Johnson, A. Lupascu, and A. Strominger, Phys. Rev. D 101, 084020 (2020), 2001.08750.
- [50]R. Narayan, D. C. M. Palumbo, M. D. Johnson, et al., Astrophys. J. 912, 35 (2021), 2105.01804.
- [51]D. C. M. Palumbo, G. N. Wong, and B. S. Prather, Astrophys. J. 894, 156 (2020), 2004.01751.
- [52]I. Esteban, M. C. Gonzalez-Garcia, M. Maltoni, T. Schwetz, and A. Zhou, JHEP 2020 (09), 178, 2007.14792.
- [53]R. Shaikh, Mon. Not. Roy. Astron. Soc. 523, 375 (2023).







