OJ 287–MOTIVATED BLACK-HOLE / DISK ENCOUNTERS

Magnetic capture during a finite disk passage

45° passage · β = 10 and 100

Stan DeLaurentiis · September 9, 2026 · Preliminary results

Result. Stronger fields deliver more horizon flux. Neither run meets the joint flux-and-power jet criterion during the simulated interval; gas is still draining.

Question. How much gas and magnetic flux survive to affect the next encounter?

Full accretion, raw horizon flux and normalized horizon flux histories for beta10 and beta100
Figure 1. Measured response over the full geometric capture interval. Blue: β=10; orange: β=100.
Methods & interpretation

Shading marks delivery across the inclined capture column (t≈1505–6636); the dotted vertical line marks central crossing (t≈4071). Accretion peaks near t=3980, while raw flux peaks much later. Peak φG is 5.46 and 1.88, respectively. The dashed φG=15 line is a campaign reference, not a universal jet criterion. Histories are shown without smoothing; flux and accretion retain code normalization. Data · SVG.

1. Setup

The solver holds the Kerr black hole fixed while a hydrostatically supported Gaussian slab passes it. Gravity focuses incoming gas; some is captured and some is deflected. The magnetic field is carried with the slab. Lower β means stronger magnetic pressure relative to gas pressure.

Hydrostatic support is imposed by a translating body force. This is a controlled local encounter, not a self-consistent global disk orbiting the primary black hole. The wide-view motion is a coordinate translation, not a moving spacetime metric.

Black-hole spin
a = 0.9
Wind / inclination
0.1c / 45° from slab normal
Gaussian scale height
H = 5 rg,2
Density contrast
1000
Field
Coherent slab-carried vertical field
Mesh
Static refinement; Δx=1 in capture ribbon, 2 in buffer, 0.25 near BH
Units
rg,2=GM₂/c²; time=GM₂/c³
Six density slices: beta10 and beta100 at initial time, central crossing vicinity, and final time
Figure 2. From approach to post-crossing capture. Rows show β=10 and 100; columns show t=0, 4060 and 6640.
Methods & interpretation

All panels retain the original x–y axes and identical logarithmic density limits, 10⁻²–10² code units. Coordinates are in the slab-rest frame: X=x+0.1(t−4070.71). The white marker locates the BH and is enlarged; it is not the horizon radius. The dashed line is the analytic slab midplane. Density below the color minimum is clipped; dark purple does not mean vacuum. These are slices, not projected column density or telescope images. Full-resolution PNG.

2. Inspect the encounter

Download full transit · MP4 73 s · 4.7 MB · 1600 × 1460

Full saved interval, t=0–6640: both β cases, wide density, near-hole density and magnetization. Fixed scales and actual slice times; two saved-time steps per second, no fluid interpolation. Slice montage only; linked history plots remain interactive below.

Loading…

β = 10

β = 100

(a–b) Wide density · x–y · slab frame · ±512 rg,2

(c–d) Near-hole flow · x–z · BH frame

Loading original snapshots…

Figure 3. Interactive, synchronized slice and history explorer. Playback steps through saved outputs; no intermediate fluid structures are generated.
Methods & interpretation

Each panel gives its actual timestamp and offset from the requested time because the two output schedules differ. Near-hole panels cover ±100 rg,2 in a different plane, with z along the spin axis; they are not simply a zoom into the wide panel. Magnetization σ=b²/ρ uses fixed limits 10⁻⁴–10³. Select gas density for limits 10⁻⁴–10². The horizon is masked. Linked history values use their nearest saved samples. This visualization alone cannot establish escaping energy or a jet. Frame times.

Raw magnetic flux versus accretion rate, colored by simulation time
Figure 4. Magnetic loading is not the same as accretion-rate normalization. Each point is a history sample; color gives time and thin lines connect successive samples.
Methods & interpretation

The identical axes compare the two β cases directly. Since φG=√(4π)ΦHL/√Ṁ, its growth can reflect both flux accumulation and declining accretion. Plotting raw flux against Ṁ exposes that distinction. This trajectory is not evidence of a closed cycle or periodic hysteresis. SVG.

3. Interpretation

For this topology, thickness, speed and inclination, delivery produces substantial accretion without establishing the campaign’s joint flux-and-power jet criterion. At t=6640 the accretion rates are still 104 and 56 code units, well above their pre-focus medians of ≈4.3. A later change during draining remains possible.

Recorded histories are finite, and maximum normalized magnetic divergence is 1.25×10⁻¹² / 3.76×10⁻¹³. These are numerical health checks, not convergence evidence. The β comparison does not test resolution, field topology, spin, radiation feedback, or repeated re-impact of a global disk. Earlier underresolved apparent-MAD runs are excluded from this note.

4. Illustrative emission

Figure 5 is not a radiation prediction. Absolute luminosity, temperature and diffusion are prescribed. No observational data or detection thresholds are plotted.

β=10   β=100
Arbitrary start

Spectrum and observing bands

Idealized top-hat windows, not calibrated telescope responses. Selecting a band highlights its energy range; it does not change the emission model.

Figure 5. Sensitivity to the assumed emission prescription. The driver is the positive excess of horizon accretion above its pre-focus median, convolved with a causal exponential kernel.
Methods & interpretation

Each prescription imposes a β=10 peak luminosity and hot/reprocessed Planck components; β=100 retains its relative driver amplitude. The spectrum keeps a fixed shape within each prescription. The time slider changes the selected sample and spectral normalization, not the simulation. Scaling assumes M₂=1.5×10⁸ M☉, z=0.306 and dL=1588.6 Mpc; one simulation time unit becomes 0.0111682 observer days. This is neither a mass measurement nor a calendar forecast. Absorption, the primary jet, telescope response and radiation feedback are omitted. Log axes omit zero and clip sub-range values. Actual predictions require a 3-D emissivity, opacity and cooling audit. Assumptions and data.

Gas-derived light curves require electron temperatures, opacities and radiation transport. The emission models above are prescribed, not fitted to observations.

5. Memory between impacts

Prescribed first and second capture windows for three two-slab delay experiments
Figure 6. Current two-slab experimental design, not a completed result. The quiet delay is measured from full first-column clearance to second-column entry, not between central crossings.
Methods & interpretation

Two initially distinct same-polarity slabs test gas and magnetic flux retained by the secondary. They do not represent a second strike through the same evolving disk patch. No second-passage result is included here. For same-disk recovery, a separate experiment must retain shear, restoring gravity, thermal evolution and the disturbed wake. SVG.

Next: measure the second-hit response against a single-hit control while varying waiting time and wake offset. Track surface density, thickness, energy and magnetic stress; infer light only after validating transport.

For QPEs and EMRIs, the target is a history-dependent delay between crossing and flare. Extrapolation requires matching gravitational, disk and cooling scales; these runs do not evolve an inspiral.

Context and comparisons

  1. Komossa et al. (2023): missing predicted thermal outburst; contrast with Valtonen et al. (2023) on revised timing and solar-gap coverage. Our local runs do not settle this orbital interpretation.
  2. Pelle et al. (2026): radiative post-processing distinguishing secondary accretion from ejecta emission. The most direct methodology comparison for future emission work.
  3. Ressler et al. (2025) and Lam et al. (2025): global magnetized and local relativistic collision studies. Accretion after impact is not itself a novel claim.
  4. Garain, Zrake & Haiman (2026): global disk response and secondary pickup disks. Our local calculation cannot test the primary’s global response.
  5. Yao et al. (2026): stellar debris from prior encounters affects later disk interactions; stellar stripping is a different memory mechanism from our black-hole capture problem.
  6. Zhu et al. (2026): disk atmosphere and breakout geometry affect flare duration. The ambient atmosphere must be assessed physically, not removed for visual clarity.
  7. Chakraborty et al. (2025): inference from QPE timings motivates testing history-dependent emission delays.

Wake and deflected flow

Density and projected flow direction at six encounter times, beta 10 above beta 100
Figure 7. Downstream response through the encounter. Wind flows right to left; the hole is fixed at the origin. Both runs develop a downstream overdensity and a broader disturbed cloud.
Methods & data

Equatorial slices, z=0; identical logarithmic density limits. Arrows show normalized in-plane coordinate velocity, not speed or trajectories. Display samples native cells every 4 rg,2, without smoothing; inner 20 rg,2 masked. This is a morphological diagnostic, not a drag measurement or a telescope image. Full-resolution PNG.

Density enhancement and depletion relative to the analytic translating slab
Figure 8. Density perturbations relative to the prescribed slab. Red: enhancement; blue: depletion.
Methods & interpretation

The reference is an analytic translating Gaussian, not a no-hole control simulation. Large enhancement outside the slab partly reflects the low ambient density; compare Figure 7. These slices do not establish a steady BHL wake. The incoming capture ribbon has about five cells per slab scale height; near-hole spacing is 0.25 rg,2. Neither is evidence of resolution convergence. Full-resolution PNG.

Next tests: follow retained flux and outward energy transport after passage; test slab equilibrium across resolutions before a refined encounter comparison. Slab-carried flux is a finite supply, unlike continuous magnetized feeding.