In this paper
AbstractIntroductionKBR spacetime and ray tracingKBR geometry and null geodesicsObserver’s sky and backward ray tracingEquilibrium torus and radiative transferCircular flow and the effective potentialPolytropic closure and local radiative weightsFrequency shift and gray transferNumerical protocol and comparison gridNumerical results and image analysisExternal-field strength and torus fillingInclination, frequency shifts, and self-absorptionEquilibrium structure and rotational signaturesDiscussion and conclusionsAcknowledgementsAuxiliary vacuum and circular-source lens mapsRay classification and optical-depth mapsObserver-radius sensitivityReferencesWorksPhysics and Astronomy
Finite-Distance Imaging of Relativistic Equilibrium Tori around Genuine Kerr–Bertotti–Robinson Black Holes
Abstract
We calculate finite-distance images of stationary, pressure-supported tori around genuine, electrically neutral Kerr–Bertotti–Robinson black holes using general relativistic ray tracing with gray emission and absorption. The emitting volume is bounded by a constant-specific-angular-momentum equilibrium free surface, and the observer is placed at a radius of . Along the investigated external-field sequence, integrated brightness and apparent size both increase strongly as the field modifies the equilibrium source and photon propagation. Higher torus filling has a different effect: in a representative sequence, brightness grows by a factor of while the intensity-weighted image size increases by only . Increasing viewing inclination enhances lateral brightness contrast. Stronger self-absorption can coexist with higher absolute brightness because emission and absorption scale together in the adopted gray prescription. At common pressure center and filling, increasing the dimensionless spin from to dims the retrograde sequence by a factor of about as its equilibrium contracts, whereas prograde brightness changes by only . These contrasting responses show how the equilibrium boundary, fluid motion, and optical depth jointly shape the brightness, scale, and asymmetry of finite-distance torus images.
Keywords: Kerr–Bertotti–Robinson spacetime; equilibrium torus; black-hole imaging; ray tracing; radiative transfer
Introduction
Black-hole images resolve radiation from matter whose distribution, motion, and opacity shape the observed brightness. Event Horizon Telescope images and polarization measurements make this connection central to the interpretation of horizon-scale observations [39, 40, 41, 42]. The vacuum capture boundary supplies a geometric reference, while direct and strongly lensed emission depend on the source and viewing geometry [21, 22, 48, 47, 46, 49, 50, 52]. Fluid motion and radiative physics also affect intensity and polarization asymmetries [53, 54, 55]. For a pressure-supported source of finite thickness, its free surface determines the emitting volume, and foreground matter can attenuate radiation from deeper layers. Relating this equilibrium structure to the image therefore requires volume radiative transfer as well as null-ray propagation.
An external electromagnetic background adds a scale to the black-hole environment, as illustrated by the test-field and exact magnetic-universe constructions [30, 31, 32]. External-field effects on optical appearance have been studied through lensing, shadow, and emission calculations [33, 34, 35, 51, 36, 37, 38]. The Kerr–Bertotti–Robinson (KBR) solution describes a rotating black hole in a Bertotti–Robinson electromagnetic background [1, 24]. We consider the electrically neutral solution, called the genuine KBR solution in the literature, which is distinct from the earlier special-charge configuration [2, 3, 62]. Hereafter, KBR denotes this neutral solution. The geometry and thermodynamics of the wider Kerr–Newman–BR family have been examined in Refs. [25, 26, 27, 28]. It recovers Kerr at zero external field but is not asymptotically flat at nonzero field. In this work, we calculate images for a finite-distance observer and impose a radiative boundary condition at the outer edge of the ray-tracing domain. In this setting, varying the field changes both photon propagation and the potential supporting the emitting fluid.
Existing KBR and Kerr–Newman–BR optical calculations describe vacuum capture boundaries, photon rings, and images of thin emitting sources [4, 5, 6, 7, 8]. They establish propagation and surface-emission signatures, but do not determine attenuation through a three-dimensional pressure-supported torus. Equilibrium fluid models in other backgrounds [61, 16, 13] and thick-source imaging calculations [64, 15, 60] provide a basis for treating finite thickness. The neutral KBR geometry poses the coupled question of how its equilibrium free surface, fluid motion, and volume absorption affect the brightness and extent of the image received by a finite observer.
We address this question with a stationary, neutral, polytropic torus of constant specific angular momentum, following the relativistic equilibrium construction [9, 63, 10, 11, 17]. The BR field enters through the metric; the source has no internal magnetic pressure. Gray emission and absorption are integrated through the equilibrium volume using invariant radiative transfer [57, 12, 18]. The comparisons determine how brightness and image size respond to field strength and filling, how inclination and optical depth redistribute the received radiation, and how spin and rotation sense alter the source structure and its image. The finite-distance construction fixes the observer and radiative boundary for each comparison rather than using asymptotic screen coordinates.
Section II specifies the spacetime and ray tracing, Sec. III gives the source and transfer model, and Sec. IV presents the images and their brightness, size, and asymmetry. Section V summarizes the physical results and their limitations.
KBR spacetime and ray tracing
KBR geometry and null geodesics
In Boyer–Lindquist-type coordinates , the electrically neutral Kerr–Bertotti–Robinson metric takes the form [2, 3]
where the metric functions are
Here is the mass parameter, is the rotation parameter, and characterizes the external Bertotti–Robinson electromagnetic field. We use geometrized units ; and have dimensions of length, while has dimensions of inverse length. The coordinates and parameters follow the physical parametrization of Ref. [3], with coordinate tildes omitted. The regular-axis azimuth has period . We impose and , the latter ensuring for all polar angles. The outer horizon is at
The familiar Kerr limit is recovered at . At nonzero the geometry has BR asymptotics, and the optical construction therefore specifies both observer and source at finite radius.
The nonzero covariant metric components are
The metric components satisfy
The Hamilton–Jacobi action for a null ray separates, following the Carter construction and the quadratic-integral description of null motion [43, 45], applied to this neutral KBR branch as in Ref. [3]. Let and be the conserved Killing energy and axial angular momentum, and let be the separation constant. We introduce the energy-normalized constants
In the Kerr limit, is the usual Carter constant. The first-order equations in the Mino-like parameter are
The radial and polar potentials appearing in these equations are
The conformal factor cancels from the unparameterized null path but remains in its affine normalization: , where is an affine parameter. The variable gives a regular form at the polar axis. After normalizing , with , the polar potential and regularized equations are
Here primes on the trajectory variables denote differentiation with respect to , and a comma denotes a coordinate derivative. This polynomial system crosses polar turning points smoothly because it contains no division by .
Observer’s sky and backward ray tracing
We place a zero-angular-momentum observer at with the baseline radius and inclination , measured from the rotation axis. The most extended equilibrium in the pressure-center survey has , so this choice places the observer outside every emitting volume. Appendix § 8 quantifies the remaining dependence on . The local observer and orthonormal spatial triad follow the standard finite-distance construction [44, 22, 56]:
where and are evaluated at the observer. This local frame is required because the BR exterior does not provide an asymptotically inertial screen.
Let be gnomonic screen coordinates and . We plot and , with positive pointing toward decreasing Boyer–Lindquist polar angle. The backward spatial tracing direction has tetrad components
For unit local frequency before the Killing-energy rescaling, define
The constants used by the separated equations are then normalized to :
Equations (20) and (22) map each screen pixel to . The initial Mino derivatives follow from the same tetrad projection,
Thus every image pixel defines one backward ray without an intermediate asymptotic-impact-parameter approximation.
For spherical photon orbits, solving for the potentials in Eq. (17) gives, at ,
We retain unstable orbits that are radially accessible from the observer and project their constants through Eq. (22). The resulting vacuum capture boundaries appear as dashed curves in the intensity images. Backward rays terminate at the horizon or at their first outward recrossing of ; the latter carries zero incident intensity. These conditions specify the transfer boundary values on the finite imaging domain.
Equilibrium torus and radiative transfer
Circular flow and the effective potential
We model a stationary, axisymmetric, circular neutral test fluid, . Its specific angular momentum and angular velocity are
For a barotropic fluid the meridional Euler equation has the standard form [9, 63, 10, 11]
Defining
the constant- solution has the standard Polish-doughnut potential [9, 63, 10, 23]
The circular four-velocity is therefore fixed algebraically,
The physical torus is restricted to the connected region containing the chosen pressure center where the circular velocity is timelike and , .
We prescribe an equatorial center and set equal to the specific angular momentum of the corresponding circular geodesic. Explicitly, the standard circular-orbit relations give [9, 17, 29]
The subscript denotes the Keplerian (circular-geodesic) angular velocity or specific angular momentum. All metric derivatives in these relations are evaluated on the equator; for the positive spins used here, the plus and minus branches correspond to prograde and retrograde motion. The pressure center is a local minimum of on the equatorial plane, with potential value . The neighboring saddle of defines a cusp. We denote by the lowest potential level at which the connected lobe around the center opens; for the fiducial sequence, this is the potential at the inner cusp. For a closed equilibrium without cusp overflow, we choose a surface potential between and :
The zero-pressure free surface is the connected equipotential enclosing ; the fluid occupies its interior, where . The filling parameter specifies the fraction of the available potential depth occupied by the torus. As increases from zero toward unity, the closed surface expands from the pressure center toward the critical opening surface.
For the spatial plots, we use the auxiliary coordinates , , and . The meridional density sections are plotted in , where . These are spatial plotting coordinates, not proper distances or the observer’s screen coordinates . On the equatorial plane , ; the surface intersections in the upper panels of Figure 1 therefore match the equatorial torus edges in the lower panels, which use the same horizontal range.
Figure 1. Construction of the fiducial equilibria at , , and . Columns use the same field sequence as the image survey. The upper row shows , the selected surface (dashed), and the critical barrier (dotted); vertical marks are the equatorial surface intersections. The lower row shows the density obtained from the same potential. The light semicircle is the coordinate horizon and the cross marks . On the equatorial plane , , so the upper-row intersections coincide with the inner and outer equatorial edges of the corresponding lower-row torus.
At fixed and , the external field changes the potential depth and both equatorial surface intersections (Figure 1). The resulting density profiles and photon propagation therefore both vary along the field sequence.
Polytropic closure and local radiative weights
We adopt and , where is rest-mass density, is total rest-frame energy density, and is the polytropic constant. The specific enthalpy is [10, 17]
Integrating Eq. (26) for constant and imposing on gives
inside the connected lobe, with and outside. For the normalization , the polytropic constant is . We set throughout.
For the dimensionless gray coefficients, we define
we take and . Here is the rest-frame absorption coefficient, is the gray source function, is the emission coefficient, and is the common central source-function normalization. The opacity therefore tracks the equilibrium density, while the source function is more centrally concentrated. Figure 2 displays both weights on the equator and shows directly how changing or changes the emitting interval. At , the solid curves in its right panel are the equatorial cuts of the lower-row density maps in Figure 1; the dashed curves are the source-function weights constructed from the same equilibrium variable . No independent radial emissivity profile is imposed.
Figure 2. Equatorial profiles that enter the transfer calculation. Solid curves show and dashed curves show . The left panel varies the filling level at fixed field; the right varies the external field at fixed filling. Every curve terminates at the two free-surface intersections determined by Eq. (33).
The torus is neutral and non-self-gravitating. The BR field therefore enters the equilibrium through the metric, not through a Lorentz force or internal magnetic pressure.
Frequency shift and gray transfer
The observer and emitter measure the photon frequency through (with a common frequency-unit convention), yielding [57, 12, 18]
Equation (38) also explains the screen normalization : after setting , . Along positive backward Mino distance the proper path element is . The formal gray bolometric solution with zero incident intensity at the outer boundary is [57, 12, 18]
Here is measured from the observer to the emitting segment along the backward ray. We denote the total optical depth through all traversed fluid by ; it is distinct from the coefficient normalization . The factor is the bolometric counterpart of the familiar specific-intensity transformation. The common fixes the relative normalization across models, but it does not provide a calibrated spectrum or polarization prediction.
Each numerical segment contributes , using midpoint fluid variables and the exact cell absorption factor. We also accumulate and report the contribution-weighted frequency shift .
The dimensionless integrated brightness and intensity-weighted image moments are defined on the observer’s local sky by
The corresponding moments are defined in the same way. The lateral brightness contrast is measured by , where the two integrals cover and , respectively, with the same solid-angle measure. uses the same emissivity and rays with the foreground attenuation factor removed. Thus is a dimensionless local-sky integral, and the plotted rms size is conditional on the screen radius used to define and .
Numerical protocol and comparison grid
The regularized equations are advanced with fourth-order Runge–Kutta steps limited by the local radial rate and . Every ray records its terminal state, torus intersection, accumulated optical depth, and the maximum normalized violations of and . The main images use pixels and the common field of view listed in Table 1; a separate common-center sequence is used only for the rotation-sense comparison, with pixels at and at .
| Quantity | Fiducial value | Values varied in figures |
| fixed | ||
| fixed | ||
| fixed | ||
| in Appendix § 8 | ||
| Image grid | displayed matrices | |
| Scalar-scan grid | for the Kerr field scan |
Table 1. Parameters of the fiducial comparison grid.
The three displayed filling levels sample low, intermediate, and nearly critical equilibria. The scalar scans resolve the trends with the denser samples
They are evaluated at pixels, with for the angularly compact Kerr control in the field scan. Every point is obtained from an independent ray-tracing calculation without interpolation.
For the fiducial model, halving the maximum integration step leaves the terminal and torus-hit labels unchanged and changes by in relative terms. For the high-spin common-center models, increasing the image grid from to changes by (prograde) and (retrograde), and rms size by less than . The retrograde horizontal and vertical cut maxima change by and , respectively. These maxima are more resolution sensitive than the integrated intensity and rms size, with residual percent-level resolution dependence in the finite-pixel profile amplitudes.
Numerical results and image analysis
The images probe three related aspects of the source–propagation problem: the response to external field and filling, the viewing and transfer effects for a fixed equilibrium, and the coupled structural and kinematic changes associated with spin and rotation sense. The field–filling and inclination/opacity sequences use , , and . The pressure-center and rotation-sense surveys use separate comparison families specified in § 4.3; their absolute brightnesses should not be compared across families as single-parameter variations. Within each image matrix, common axes and color normalization permit direct comparison.
External-field strength and torus filling
Figure 3 shows the equilibrium density across the field–filling grid. Increasing widens and thickens the connected lobe. Increasing the field at fixed and also changes the outer low-density envelope because the potential itself has changed. The density vanishes outside the closed free surface, including the gap between the horizon and the inner torus edge.
Figure 3. Normalized meridional density for the matched equilibrium grid. Rows give and columns give ; , , and . All panels share the same coordinate window and logarithmic color scale.
The corresponding intensity sequence is shown in Figure 4. Its rows and columns correspond to the equilibria in Figure 3. Filling mainly changes the source extent, whereas the BR field changes both that extent and the finite-distance lens map. The increasing image scale with field is therefore not a pure shadow-size effect.
Figure 4. Observed gray intensity for the same grid as Figure 3, at , , and . A single absolute logarithmic scale is used throughout. Dashed curves are the vacuum critical curves and are references, not emitting surfaces.
The prescribed circular-ring and celestial-source maps in Appendix § 6 show the field dependence of photon propagation without torus absorption.
Inclination, frequency shifts, and self-absorption
At fixed geometry, Figure 5 isolates viewing direction. Low-inclination images are nearly annular. Increasing produces a stronger left–right contrast and separates near- and far-side contributions, while larger increases the projected area and the range of emitting depths.
Figure 5. Observed intensity at fixed , , , , and . Rows give and columns give . The intensity normalization is common to every panel.
The contribution-weighted frequency-shift maps in Figure 6 link the intensity asymmetry in Figure 5 to the fluid motion on the same filling–inclination grid. The red–blue division rotates into a pronounced lateral contrast as the line of sight approaches the equatorial plane, linking the bright side in Figure 5 to the fluid motion. Increasing broadens the region over which the same kinematic pattern is sampled; it does not replace inclination as the main driver of the red–blue split. The corresponding ray-class and optical-depth matrices are given in Appendix § 7.
Figure 6. Contribution-weighted frequency shift for the same filling–inclination grid as Figure 5. Rows give and columns give at . All panels share a diverging scale centered on ; blue and red denote net redshift and blueshift, and non-emitting directions are black.
Figure 7 then varies while holding the equilibrium fixed. At small the source is faint on the common absolute scale; at larger the foreground approaches its local source function and deeper contributions are suppressed, most visibly at high inclination. Because , this changes emission and absorption together and should be read as an optical-depth normalization survey, not as pure absorption at fixed emissivity.
Figure 7. Observed intensity for and , at , , , , and . One absolute scale is used for all panels.
Equilibrium structure and rotational signatures
Spin and rotation sense affect both the fluid motion and the equilibrium free surface. The resulting image signatures depend on the spatial extent and kinematics of each equilibrium. The pressure-center and common-center rotation-sense sequences sample these coupled structural and frequency-shift responses under different source constraints. The field–filling sequence above uses , whereas the pressure-center survey uses and , and the rotation-sense survey uses and . Each survey holds its own viewing and transfer parameters fixed.
Along the prograde branch, Figure 8 varies the pressure-center radius and spin at fixed and . Moving the pressure center outward substantially broadens the fluid region, whereas the change across the selected spin columns is smoother. All twelve equilibria lie on the same closure branch.
Figure 8. Normalized equilibrium density at fixed and . Rows give ; columns give . A common coordinate window and density scale are used.
The intensity matrix in Figure 9 shows the images of these equilibria at . Spin changes the near-side contrast and the position of the brightest arc, but the outward displacement of the pressure center produces the larger change in angular extent over the ranges shown.
Figure 9. Observed intensity for the equilibria in Figure 8, at , , and . Rows give and columns give . The field of view contains the complete images in this sequence.
For the rotation-sense comparison we fix and , with each signed set by its circular geodesic. At and , the selected prograde torus spans , while the retrograde torus spans only . Equal center and filling do not imply equal source volume. Figure 10 compares the density distributions and free surfaces of these equilibria.
Figure 10. Normalized density for the common-center sequence. Columns give and rows give prograde and retrograde rotation, at , , and . The shared logarithmic scale exposes the contraction of the retrograde lobe.
Figure 11 shows the corresponding high-inclination images. The two rotation branches have opposite Doppler-bright sides, while the smaller angular extent of the retrograde images also records their more compact equilibrium sources. The common and specify the equilibrium family; the two rotation senses have different free surfaces.
Figure 11. Images of the common-center sequence at , , , , , and . Columns give spin and rows give rotation sense. A common absolute intensity scale is used; dashed curves are the vacuum critical curves. The image grids are at and at .
The one-dimensional cuts in Figure 12 separate the broad primary emission from narrow lensed features more clearly than the 2D maps. On the common normalization, the screen-integrated intensity changes only slightly along the prograde sequence but falls sharply along the retrograde sequence as its lobe contracts. The centroid changes sign when the flow is reversed in the adopted screen convention. Specifically, changes from to along the prograde sequence and from to along the retrograde sequence as rises from to .
An auxiliary equal-width comparison illustrates the sensitivity to the chosen free surface. At , reducing the prograde filling to matches its equatorial width to the retrograde value at , while retaining and . On a common image grid, the prograde/retrograde integrated-intensity ratio then falls from for equal filling to for equal width. Matching this single coordinate width does not match the full volume or thermodynamic profiles, so it does not isolate a pure Doppler factor. The large equal-filling contrast is consequently sensitive to the equilibrium boundary.
Figure 12. Horizontal () and vertical () cuts through Figure 11. The upper panels retain the common absolute normalization; the lower panels normalize each cut to its own peak. Solid and dashed curves denote prograde and retrograde rotation. Each cut averages the two central pixel rows or columns, with the grids stated in Figure 11.
The integrated brightness, rms size, and lateral contrast distinguish the physical responses across these image families. Figure 13 quantifies the field–filling and viewing/transfer changes displayed in Figure 4–Figure 7. Its upper panels measure the integrated brightness and rms extent of the field–filling images in Figure 4. The lower-left panel measures the inclination-dependent brightness asymmetry of Figure 5, whose kinematic origin is illustrated by Figure 6. The lower-right panel measures how much foreground absorption reduces the intensity relative to the unabsorbed integral for the opacity sequence in Figure 7. These four panels summarize those sequences; the rotational signatures are quantified separately by Figure 11 and Figure 12. The denser parameter samples resolve their trends between the representative image panels; each marker is an independent ray-tracing calculation.
Figure 13. Quantitative summary of the field–filling, inclination, and optical-depth sequences in Figure 4–Figure 7. Upper left: screen-integrated intensity versus filling. Upper right: rms image size. Lower left: left/right flux ratio versus inclination. Lower right: attenuation relative to the optically thin integral versus . The scans use 8 filling values, 7 inclinations, and 7 approximately logarithmically spaced optical-depth values; every marker is a separately ray-traced calculation on a grid ( for the Kerr field scan) at . The line colors and legends identify the held or varied parameter in each panel.
The upper panels distinguish filling from the external-field response. At fixed , increasing from to raises from to , a factor of , while the rms size increases from to , only . The growing torus therefore contributes substantially more radiation without a comparable expansion of its bright image. The density envelope in Figure 3 and the intensity-weighted size measure different parts of the source: low-density outer layers carry little weight in the intensity moments. In contrast, at , increasing from to raises from to and the rms size from to . Thus a large increase in brightness can accompany either a modest size change along the filling sequence or a pronounced size change along the field sequence. These are responses of the specified equilibrium and finite-distance lens map, not a decomposition into independent source and lensing contributions.
The lower-left panel quantifies the inclination dependence at fixed field. For and , changing from to raises the left/right flux ratio from to . The frequency-shift maps in Figure 6 locate the enhanced side, while the ray and optical-depth maps in Appendix § 7 identify which paths cross foreground fluid. Inclination therefore changes the contrast much more strongly than can be inferred from image size alone.
The lower-right panel separates self-absorption from total brightness. At , increasing from to lowers from to , while rises from to . An increasingly self-obscured source can consequently become brighter in absolute terms in this model. The reason is that multiplies both the absorption coefficient and the emissivity ; the attenuation ratio measures the radiation lost relative to the corresponding unabsorbed source, not relative to a fixed-emissivity image.
A limited source-function test replaces by , with , in the fiducial model on a grid. At , the attenuation ratio ranges from to , and the absolute intensity still rises with for each exponent. The absolute intensities at that optical depth differ by a factor of across the three prescriptions. The coexistence of increasing brightness and stronger self-absorption thus persists in this test, while its numerical amplitude depends on the assumed concentration of the source function.
The different brightness, size, and asymmetry responses across these families reflect the combined effects of equilibrium structure, fluid motion, and radiative transfer.
Discussion and conclusions
We have calculated gray images of stationary equilibrium tori in the neutral KBR spacetime, integrating emission and absorption through their finite volumes to an observer at . The field and filling sequences produce different relations between brightness and image size. Increasing from to at raises integrated intensity by a factor of but increases rms size by only . At , the rms size instead grows by a factor of between and . The equilibrium envelope and the bright-image extent therefore need not vary in proportion. The field response combines changes in the fluid potential and the lens map, whereas filling changes the free surface within a fixed spacetime.
Inclination enhances lateral brightness contrast, while self-absorption changes the emitting depths visible to the observer. At , increasing the optical-depth normalization from to lowers the received fraction of the unabsorbed intensity from to , even as absolute brightness rises. This coexistence follows from : emission and absorption scale together in the gray prescription. The attenuation ratio and absolute intensity consequently measure different aspects of the source. Optically thin KBR disk models [7, 8] omit this attenuation through the emitting volume.
Rotation sense affects the source structure as well as the bright side. At common pressure center and filling, increasing from to dims the retrograde sequence by a factor of about , whereas prograde brightness changes by only . The retrograde dimming accompanies contraction of its equilibrium lobe. Motion-dependent photon escape and frequency shifts are quantified for Kerr geodesic emitters in Ref. [65]; the torus comparison also changes the spatial distribution of emission. Its rotational brightness contrast thus contains both kinematic and equilibrium-structure effects, rather than a pure Doppler factor.
These results apply to stationary, neutral test fluids with the specified gray source function and finite observer. Internal magnetic stresses [14, 59], plasma emission prescriptions [64, 15, 19], and polarized transfer [20, 66, 58] require independent physical assumptions for observational predictions. Appendix § 8 quantifies the remaining observer-radius dependence. Within the investigated families, brightness, size, and asymmetry exhibit different responses to the source and transfer parameters. Their joint interpretation must account for the equilibrium emitting region as well as the propagation of null rays.
Acknowledgements
This work was supported by the National Natural Science Foundation of China under Grant No. 12305070, and the Basic Research Program of Shanxi Province under Grant Nos. 202303021222018 and 202303021221033.
Auxiliary vacuum and circular-source lens maps
Prescribed circular emitters and a celestial sphere isolate null propagation without torus absorption. Figure fig:ringmatrix shows the first three equatorial crossings; its source radii and field samples differ from the torus grid in Figure 3 and Figure 4.
Figure 14. Lensed celestial sphere at , using the spacetime and observer grid of Fig. fig:ringmatrix and a field of view. Colors identify source-texture quadrants in the tangent frame about the antipodal optical direction; black denotes capture and the light curves are mapped source-coordinate grid lines.
At the same observer radius, the finite celestial-source map displays the field- and spin-dependent distortion of a sphere at . Grid folding close to the capture boundary marks strongly bent paths. The representative trajectories then illustrate that bending at and ; their meridional projections use common axes.
Figure 15. Four representative outward-returning null rays selected just outside the capture boundary for , , , and . Rows give and ; the left column shows three-dimensional coordinate paths and the right column their meridional projections. Common, cropped coordinate windows emphasize the strongly bent near region; the integrations themselves continue to the source sphere at . Colors identify the four source-sphere directions, and the black star marks the observer.
Ray classification and optical-depth maps
Figure 16 separates horizon capture from return to the finite outer boundary and records whether each ray crosses the torus. Repetition along the filling axis is physical: capture or return is source-independent, while the torus-intersection label changes with the source boundary.
Figure 16. Ray classification for the models shown in Figure 5. Rows give and columns give . The shared legend distinguishes horizon capture from return to and records whether a ray intersects the torus.
Figure 17 gives the integrated optical-depth maps for the same filling–inclination sequence. They locate the columns most susceptible to foreground attenuation. Larger filling and higher inclination enlarge the projected high-depth regions, although the emergent intensity also depends on the source function and frequency shift.
Figure 17. Integrated optical depth for the same rays as Figure 5. Rows give , columns give increasing inclination, and all panels use one logarithmic scale.
Observer-radius sensitivity
We quantify the observer-radius dependence of the baseline images by repeating the ray tracing at while holding the local gnomonic field of view, source model, inclination, opacity normalization, and numerical step control fixed. Figure 18 compares the fiducial torus, an extended high-spin prograde torus, and the largest equilibrium in the pressure-center sequence. The coordinates and are local angular screen coordinates, so the changing outline is not hidden by rescaling with .
Figure 18. Finite-observer-radius comparison at . The upper row is the fiducial model ; the middle row is the extended prograde model with ; the lower row is the outer-center model with . Intensities are shown on one logarithmic scale within each row, dashed curves are the vacuum critical curves, and the inset numbers give the screen-integrated intensity relative to the calculation.
We measure the angular-size proxy , using the intensity-weighted gnomonic variances and , and the median critical-curve angular radius ; these avoid the explicit scaling of and . For the fiducial torus, moving the observer from to changes the angular rms size by only , the median critical-curve radius by , and the left/right flux ratio by . For the extended torus, the same displacement changes those quantities by , , and , respectively. In contrast, the screen-integrated intensity changes by and in the two models. For the largest pressure-center model, with , the – changes are in angular size, in critical-curve radius, in left/right ratio, and in integrated intensity. The observer is only at when , compared with at . The extended-source survey therefore retains appreciable observer-radius dependence in both size and intensity. Across all three cases, the asymmetry is substantially more stable than the absolute intensity. The comparison is therefore a specified finite-distance experiment, with the residual dependence reported here.
Figure 19 gives the corresponding scalar radius dependence. The integrated-intensity and angular-size curves flatten gradually over the sampled interval. The larger changes between and are especially important for the extended torus: because at but at , the former position samples a visibly more local view. All integrals here refer to the fixed aperture; the extended sources reach its edges most visibly at . A grid is used for these supplementary calculations. Comparing the nine image models at against a grid changes the integrated intensities by at most and the angular rms sizes by at most . All rays reach a terminal state, and the largest first-integral residual over the six observer radii remains below .
Figure 19. Dependence of image properties on the observer radius for the three rows of Figure 18. From left to right and top to bottom: relative integrated intensity, angular rms size, median critical-curve radius, left/right flux ratio, attenuation relative to the optically thin integral, and torus-hit solid angle. Each sequence samples , whereas the image comparison retains the three representative radii , , and . Lines connect independently ray-traced calculations only as visual guides.
References
- [1]Podolsky, J., & Ovcharenko, H. (2025). Kerr black hole in a uniform Bertotti--Robinson magnetic field: An exact solution. Phys. Rev. Lett., 135, 181401. https://doi.org/10.1103/rfgv-ybz5
- [2]Ovcharenko, H., & Podolsky, J. (2026). Kerr--Newman and genuine Kerr black holes in the Bertotti--Robinson magnetic field. https://arxiv.org/abs/2608.21672
- [3]Kubiznak, D., Ovcharenko, H., & Podolsky, J. (2026). Thermodynamics of Genuine Kerr--Bertotti--Robinson Black Holes. https://arxiv.org/abs/2608.22292
- [4]Wang, X., Hou, Y., Wan, X., Guo, M., & Chen, B. (2026). Geodesics and shadows in the Kerr--Bertotti--Robinson black hole spacetime. JCAP, 02, 050. https://arxiv.org/abs/2507.22494
- [5]Zeng, X.-X., Yang, C.-Y., & Yu, H. (2025). Optical characteristics of the Kerr--Bertotti--Robinson black hole. Eur. Phys. J. C, 85, 1242. https://arxiv.org/abs/2508.03020
- [6]Wan, X., Zhang, Z., Wei, F.-S., Hou, Y., & Chen, B. (2026). Critical behavior of photon rings in Kerr--Bertotti--Robinson spacetime. Phys. Rev. D, 114, 044005. https://arxiv.org/abs/2603.25049
- [7]Wan, X., Wang, H., Zhang, Z., Zhang, Z., & Chen, B. (2026). Shadows and thin-disk images of Kerr--Newman black holes in a Bertotti--Robinson magnetic field. https://arxiv.org/abs/2609.06648
- [8]Zhang, Z.-Y., Li, X.-Q., Yan, H.-P., & Yue, X.-J. (2026). Optical appearance of the Kerr--Bertotti--Robinson black hole with a magnetically driven synchrotron emissivity model. Eur. Phys. J. C, 86, 945. https://doi.org/10.1140/epjc/s10052-026-16174-1
- [9]Fishbone, L. G., & Moncrief, V. (1976). Relativistic fluid disks in orbit around Kerr black holes. Astrophys. J., 207, 962. https://articles.adsabs.harvard.edu/pdf/1976ApJ...207..962Farticles.adsabs.harvard.edu/pdf/1976ApJ...207..962F
- [10]Abramowicz, M. A., Jaroszynski, M., & Sikora, M. (1978). Relativistic, accreting disks. Astron. Astrophys., 63, 221. https://articles.adsabs.harvard.edu/pdf/1978A%26A....63..221Aarticles.adsabs.harvard.edu/pdf/1978A%26A....63..221A
- [11]Abramowicz, M. A., & Fragile, P. C. (2013). Foundations of Black Hole Accretion Disk Theory. Living Rev. Relativ., 16, 1. https://doi.org/10.12942/lrr-2013-1
- [12]Younsi, Z., Wu, K., & Fuerst, S. V. (2012). General relativistic radiative transfer: formulation and emission from structured tori around black holes. Astron. Astrophys., 545, A13. https://doi.org/10.1051/0004-6361/201219599
- [13]Zhang, Z., Chen, S., & Jing, J. (2024). Images of Kerr-MOG black holes surrounded by geometrically thick magnetized equilibrium tori. JCAP, 09, 027. https://arxiv.org/abs/2404.12223
- [14]Komissarov, S. S. (2006). Magnetized tori around Kerr black holes: analytic solutions with a toroidal magnetic field. Mon. Not. R. Astron. Soc., 368, 993–1000. https://doi.org/10.1111/j.1365-2966.2006.10183.x
- [15]Vincent, F. H., Yan, W., Straub, O., Zdziarski, A. A., & Abramowicz, M. A. (2015). A magnetized torus for modeling Sgr A* millimeter images and spectra. Astron. Astrophys., 574, A48. https://doi.org/10.1051/0004-6361/201424306
- [16]Prada-Méndez, G. D., Lora-Clavijo, F. D., & Velásquez-Cadavid, J. M. (2023). Synchrotron emitting Komissarov torus around naked singularities. Class. Quantum Grav., 40, 195011. https://doi.org/10.1088/1361-6382/acf17e
- [17]Penna, R. F., Kulkarni, A., & Narayan, R. (2013). A new equilibrium torus solution and GRMHD initial conditions. Astron. Astrophys., 559, A116. https://doi.org/10.1051/0004-6361/201219666
- [18]Davis, S. W., & Gammie, C. F. (2020). Covariant radiative transfer for black hole spacetimes. Astrophys. J., 888, 94. https://doi.org/10.3847/1538-4357/ab5950
- [19]Dexter, J. (2016). A public code for general relativistic, polarised radiative transfer around spinning black holes. Mon. Not. R. Astron. Soc., 462, 115–136. https://doi.org/10.1093/mnras/stw1526
- [20]Mościbrodzka, M., & Gammie, C. F. (2018). ipole---semi-analytic scheme for relativistic polarized radiative transport. Mon. Not. R. Astron. Soc., 475, 43–54. https://doi.org/10.1093/mnras/stx3162
- [21]Gralla, S. E., Holz, D. E., & Wald, R. M. (2019). Black Hole Shadows, Photon Rings, and Lensing Rings. Phys. Rev. D, 100, 024018. https://doi.org/10.1103/PhysRevD.100.024018
- [22]Perlick, V., & Tsupko, O. Y. (2022). Calculating black hole shadows: Review of analytical studies. Phys. Rept., 947, 1–39. https://doi.org/10.1016/j.physrep.2021.10.004
- [23]Tlemissov, A., & Kovář, J. (2024). Egg-shaped sections of fluid structures around black holes. Eur. Phys. J. C, 84, 1115. https://doi.org/10.1140/epjc/s10052-024-13496-w
- [24]Al-Badawi, A., & Halilsoy, M. (2004). Revisiting the rotating BR electromagnetic universe. Nuovo Cim. B, 119, 931. https://www.emu.edu.tr/stafffile/JOHXAGD.pdfemu.edu.tr/stafffile/JOHXAGD.pdf
- [25]Ovcharenko, H. (2026). Global structure of Kerr--Newman--BR spacetime. https://arxiv.org/abs/2608.28265
- [26]Zhou, Y.-S., Fu, W.-T., Cao, L.-M., & Cai, R.-G. (2026). On the extended Kerr--Newman--Bertotti--Robinson spacetime: two black holes and a naked singularity in Bertotti--Robinson universe. https://arxiv.org/abs/2609.20626
- [27]Zhang, Z., Zhang, Z., & Chen, B. (2026). Thermodynamics of Kerr--Newman--Bertotti--Robinson black holes. https://arxiv.org/abs/2609.09816
- [28]Zhou, Y.-S., Wu, L.-B., Ji, M.-F., Fu, W.-T., Cao, L.-M., & Cai, R.-G. (2026). Notes on Kerr--Bertotti--Robinson spacetime. https://arxiv.org/abs/2608.19136
- [29]Wang, T. (2025). Innermost stable circular orbit of Kerr--Bertotti--Robinson black holes and inspirals from it: exact solutions. https://arxiv.org/abs/2508.04684
- [30]Wald, R. M. (1974). Black hole in a uniform magnetic field. Phys. Rev. D, 10, 1680–1685. https://doi.org/10.1103/PhysRevD.10.1680
- [31]Ernst, F. J. (1976). Black holes in a magnetic universe. J. Math. Phys., 17, 54–56. https://doi.org/10.1063/1.522781
- [32]Ernst, F. J., & Wild, W. J. (1976). Kerr black holes in a magnetic universe. J. Math. Phys., 17, 182. https://doi.org/10.1063/1.522875
- [33]Stuchlík, Z., & Hledík, S. (1999). Photon capture cones and embedding diagrams of the Ernst spacetime. Class. Quantum Grav., 16, 1377–1387. https://arxiv.org/abs/0803.2536
- [34]Wang, M., Chen, S., & Jing, J. (2021). Kerr black hole shadows in Melvin magnetic field with stable photon orbits. Phys. Rev. D, 104, 084021. https://arxiv.org/abs/2104.12304
- [35]Junior, H. C. D. L., Cunha, P. V. P., Herdeiro, C. A. R., & Crispino, L. C. B. (2021). Shadows and lensing of black holes immersed in strong magnetic fields. Phys. Rev. D, 104, 044018. https://arxiv.org/abs/2104.09577
- [36]Zhu, H., & Guo, M. (2022). Polarized image of synchrotron radiation from hotspots in Schwarzschild--Melvin black hole spacetime. https://arxiv.org/abs/2205.04777
- [37]Hu, Z., Zhong, Z., Li, P.-C., Guo, M., & Chen, B. (2021). QED effect on a black hole shadow. Phys. Rev. D, 103, 044057. https://arxiv.org/abs/2012.07022
- [38]Zhong, Z., Hu, Z., Yan, H., Guo, M., & Chen, B. (2021). QED effects on Kerr black hole shadows immersed in uniform magnetic fields. Phys. Rev. D, 104, 104028. https://arxiv.org/abs/2108.06140
- [39]Event Horizon Telescope Collaboration, Akiyama, K., et al. (2019). First M87 Event Horizon Telescope Results. I. The shadow of the supermassive black hole. Astrophys. J. Lett., 875, L1. https://arxiv.org/abs/1906.11238
- [40]Event Horizon Telescope Collaboration, Akiyama, K., et al. (2022). First Sagittarius A* Event Horizon Telescope Results. I. The shadow of the supermassive black hole in the center of the Milky Way. Astrophys. J. Lett., 930, L12. https://arxiv.org/abs/2311.08680
- [41]Event Horizon Telescope Collaboration, Akiyama, K., et al. (2021). First M87 Event Horizon Telescope Results. VII. Polarization of the ring. Astrophys. J. Lett., 910, L12. https://arxiv.org/abs/2105.01169
- [42]Event Horizon Telescope Collaboration, Akiyama, K., et al. (2021). First M87 Event Horizon Telescope Results. VIII. Magnetic field structure near the event horizon. Astrophys. J. Lett., 910, L13. https://arxiv.org/abs/2105.01173
- [43]Carter, B. (1968). Global structure of the Kerr family of gravitational fields. Phys. Rev., 174, 1559–1571. https://doi.org/10.1103/PhysRev.174.1559
- [44]Bardeen, J. M. (1973). Timelike and null geodesics in the Kerr metric. In Black Holes (Les Astres Occlus) (pp. 215–239). Gordon and Breach. https://ui.adsabs.harvard.edu/abs/1973blho.conf..215B/abstractui.adsabs.harvard.edu/abs/1973blho.conf..215B/abstract
- [45]Walker, M., & Penrose, R. (1970). On quadratic first integrals of the geodesic equations for type [22] spacetimes. Commun. Math. Phys., 18, 265–274. https://doi.org/10.1007/BF01649445
- [46]Johannsen, T. (2013). Photon rings around Kerr and Kerr-like black holes. Astrophys. J., 777, 170. https://arxiv.org/abs/1501.02814
- [47]Cunha, P. V. P., & Herdeiro, C. A. R. (2018). Shadows and strong gravitational lensing: a brief review. Gen. Relativ. Gravit., 50, 42. https://arxiv.org/abs/1801.00860
- [48]Chen, S., Jing, J., Qian, W.-L., & Wang, B. (2023). Black hole images: A review. Sci. China Phys. Mech. Astron., 66, 260401. https://arxiv.org/abs/2301.00113
- [49]Gralla, S. E., & Lupsasca, A. (2020). Lensing by Kerr black holes. Phys. Rev. D, 101, 044031. https://arxiv.org/abs/1910.12873
- [50]Hou, Y., Liu, P., Guo, M., Yan, H., & Chen, B. (2022). Multi-level images around Kerr--Newman black holes. Class. Quantum Grav., 39, 194001. https://doi.org/10.1088/1361-6382/ac8860
- [51]Hou, Y., Zhang, Z., Yan, H., Guo, M., & Chen, B. (2022). Image of a Kerr--Melvin black hole with a thin accretion disk. Phys. Rev. D, 106, 064058. https://doi.org/10.1103/PhysRevD.106.064058
- [52]Chael, A., Johnson, M. D., & Lupsasca, A. (2021). Observing the inner shadow of a black hole: A direct view of the event horizon. Astrophys. J., 918, 6. https://arxiv.org/abs/2106.00683
- [53]Gates, D. E. A., Hadar, S., & Lupsasca, A. (2020). Maximum observable blueshift from circular equatorial Kerr orbiters. Phys. Rev. D, 102, 104041. https://arxiv.org/abs/2009.03310
- [54]Gelles, Z., Himwich, E., Palumbo, D. C. M., & Johnson, M. D. (2021). Polarized image of equatorial emission in the Kerr geometry. Phys. Rev. D, 104, 044060. https://arxiv.org/abs/2105.09440
- [55]Palumbo, D. C. M., Wong, G. N., & Prather, B. S. (2020). Discriminating accretion states via rotational symmetry in simulated polarimetric images of M87. Astrophys. J., 894, 156. https://arxiv.org/abs/2004.01751
- [56]Chang, Z., & Zhu, Q.-H. (2021). The observer-dependent shadow of the Kerr black hole. JCAP, 09, 003. https://arxiv.org/abs/2104.14221
- [57]Lindquist, R. W. (1966). Relativistic transport theory. Ann. Phys., 37, 487–518. https://doi.org/10.1016/0003-4916(66)90207-7
- [58]Hou, Y., Huang, J., Guo, M., Mizuno, Y., & Chen, B. (2025). Near-horizon polarization as a diagnostic of black hole spacetime. Astrophys. J. Lett., 988, L51. https://doi.org/10.3847/2041-8213/adee09
- [59]Hou, Y., Zhang, Z., Guo, M., & Chen, B. (2024). A new analytical model of magnetofluids surrounding rotating black holes. JCAP, 02, 030. https://arxiv.org/abs/2309.13304
- [60]Zhang, Z., Hou, Y., Guo, M., & Chen, B. (2024). Imaging thick accretion disks and jets surrounding black holes. JCAP, 05, 032. https://arxiv.org/abs/2401.14794
- [61]Cassing, M., & Rezzolla, L. (2023). Equilibrium non-selfgravitating tori around black holes in parameterised spherically symmetric spacetimes. Mon. Not. R. Astron. Soc., 522, 2415. https://arxiv.org/abs/2302.09135
- [62]Ma, L., & Lü, H. (2026). Ricci-flat Seeds, Trombone Rescaling and Magnetizing Kerr Black Holes. https://arxiv.org/abs/2610.30256
- [63]Kozlowski, M., Jaroszynski, M., & Abramowicz, M. A. (1978). The analytic theory of fluid disks orbiting the Kerr black hole. Astron. Astrophys., 63, 209–220. https://articles.adsabs.harvard.edu/pdf/1978A%26A....63..209Karticles.adsabs.harvard.edu/pdf/1978A%26A....63..209K
- [64]Straub, O., Vincent, F. H., Abramowicz, M. A., Gourgoulhon, E., & Paumard, T. (2012). Modelling the black hole silhouette in Sagittarius A* with ion tori. Astron. Astrophys., 543, A83. https://doi.org/10.1051/0004-6361/201219209
- [65]Su, Y., Guo, M., Yan, H., & Chen, B. (2024). Photon emissions from Kerr equatorial geodesic orbits. Eur. Phys. J. C, 84, 733. https://doi.org/10.1140/epjc/s10052-024-13113-w
- [66]Huang, J., Zheng, L., Guo, M., & Chen, B. (2024). Coport: A New Public Code for Polarized Radiative Transfer in a Covariant Framework. https://doi.org/10.48550/arXiv.2407.10431



















