Paczyński–Wiita Potential

The Paczyński–Wiita Potential is an empirical pseudo-Newtonian gravitational potential introduced by the Polish astrophysicist Bohdan Paczyński and the American astrophysicist Paul J. Wiita in 1980 to model radiation-pressure-supported thick accretion disks around non-rotating black holes without solving the full nonlinear equations of general relativity. By replacing the classical Newtonian point-mass denominator r with the shifted radial coordinate (r − R_S), where R_S = 2*GM* / c² is the Schwarzschild radius, the potential captures the steepening of strong-field relativistic gravity near the event horizon. Crucially, when inserted into standard Newtonian Euler or magnetohydrodynamic (MHD) fluid equations, the Paczyński–Wiita Potential reproduces the precise locations of both the general-relativistic innermost stable circular orbit (r_ISCO = 3R_S = 6*GM* / c²) and the marginally bound circular orbit (r_mb = 2R_S = 4*GM* / c²).

For a spherically symmetric, non-spinning compact object of mass M with gravitational radius r_g = GM / c² and Schwarzschild event-horizon radius R_S = 2r_g = 2*GM* / c², the Paczyński–Wiita Potential Φ_PW(r) is defined for all radial coordinates r > R_S as:

Φ_PW(r) = −GM / (r − R_S) = −GM / (r − 2*GM* / c²)

At large distances (r ≫ R_S), expanding Φ_PW(r) in powers of (R_S / r) recovers the Newtonian potential plus leading-order post-Newtonian corrections:

Φ_PW(r) = −(GM / r) × [ 1 + (R_S / r) + (R_S / r)² + … ]

Differentiating Φ_PW(r) with respect to r yields the inward radial gravitational acceleration g_PW(r):

g_PW(r) = −dΦ_PW / dr = −GM / (r − R_S)²

which diverges at the Schwarzschild horizon r = R_S, mimicking the infinite proper acceleration required for a stationary observer to hover just outside a black hole event horizon.

Keplerian Angular Velocity, Specific Angular Momentum, and Binding Energy

For a test particle or fluid ring moving in an equatorial circular orbit of radius r > R_S, balancing radial centrifugal acceleration rΩ_K²(r) against g_PW(r) determines the pseudo-Newtonian Keplerian angular velocity Ω_K(r), specific angular momentum ℓ_K(r) = r²Ω_K(r), and specific mechanical energy E_K(r) = ½r²Ω_K²(r) + Φ_PW(r):

1. Keplerian Angular Frequency: Ω_K(r) = √(GM / r) / (r − R_S)
2. Keplerian Circular Speed: v_φ(r) = rΩ_K(r) = √(GMr) / (r − R_S)
3. Specific Orbital Angular Momentum: ℓ_K(r) = √(GMr³) / (r − R_S)
4. Specific Orbital Mechanical Energy: E_K(r) = −(GM / 2) × (r − 2R_S) / (r − R_S)²

In classical Newtonian gravity (Φ_N = −GM / r), specific angular momentum ℓ_N(r) = √(GMr) increases monotonically down to r = 0, so circular orbits are dynamically stable at all radii and can extract infinite binding energy as r → 0. In general relativity, by contrast, strong-field spacetime curvature creates a minimum in ℓ_K(r) at a finite radius inside which circular orbits become radially unstable and plunge directly into the black hole without further viscous torque.

Derivation of r_ISCO = 3R_S and Rayleigh Epicyclic Frequency

By Lord Rayleigh's centrifugal stability criterion, an axisymmetric circular orbit is stable against small radial perturbations if and only if specific angular momentum increases outward with radius (dℓ_K² / dr > 0). Differentiating ℓ_K(r) = √(GMr³) / (r − R_S) with respect to r gives:

dℓ_K / dr = √(GMr) × (r − 3R_S) / [ 2(r − R_S)² ]

The radial epicyclic frequency κ(r), which governs radial oscillations of slightly eccentric orbits in accretion disks and quasi-periodic oscillations (QPOs) in X-ray binaries, is given by:

κ²(r) = (2Ω_K / r) (dℓ_K / dr) = GM(r − 3R_S) / [ r(r − R_S)³ ]

Consequently:

• For r > 3R_S, dℓ_K / dr > 0 and κ² > 0: circular orbits are stable, and displaced fluid elements undergo harmonic radial epicyclic oscillations.
• At r = r_ISCO = 3R_S = 6*GM* / c², dℓ_K / dr = 0 and κ = 0: the orbit is marginally stable, matching the exact Schwarzschild geodesic result.
• For R_S < r < 3R_S, dℓ_K / dr < 0 and κ² < 0: circular orbits are dynamically unstable, driving laminar transonic plunging inflow across the inner edge of the accretion disk.

Derivation of r_mb = 2R_S and the Roche Lobe Cusp

Examining the specific mechanical energy E_K(r) = −(GM / 2)(r − 2R_S) / (r − R_S)² shows that E_K(r) reaches its minimum binding energy E_K(3R_S) = −GM / (8R_S) = −c² / 16 at the ISCO (giving a radiative accretion efficiency η = 6.25%, close to the true Schwarzschild value of 1 − 2√2 / 3 ≈ 5.72%) and crosses zero binding energy (E_K = 0) at the marginally bound orbit:

r_mb = 2R_S = 4*GM* / c²

In thick supercritical accretion tori ("Polish doughnuts" with constant specific angular momentum ℓ₀ bounded between ℓ_ISCO and ℓ_mb), equipotential surfaces of the effective potential Φ_eff(r, z) = Φ_PW(√(r² + z²)) + ℓ₀² / (2r²) form a self-crossing inner Roche-lobe cusp located between r_mb = 2R_S and r_ISCO = 3R_S. Accreting gas overflows this relativistic cusp dynamically without requiring microscopic viscosity, just as in full general-relativistic hydrodynamics.

While the Paczyński–Wiita Potential excels for non-rotating black holes in hydrodynamical and radiation-MHD codes, it does not capture light bending, gravitational redshift, Lense–Thirring frame dragging around spinning Kerr black holes, or the true apsidal precession rate of eccentric orbits at large radii (predicting twice the Einstein perihelion advance as r / R_S → ∞). For spinning black holes or closer epicyclic frequency matching, subsequent pseudo-Newtonian potentials—such as the Nowak–Wagoner (1991), Artemova–Bjoernsson–Novikov (1996), and Kluźniak–Lee (2002) potentials—introduce spin-dependent radial exponents or velocity-dependent forces.

Why is the Paczyński–Wiita Potential still widely used when full General Relativistic MHD (GRMHD) codes exist?

Many astrophysical problems—such as three-dimensional radiation-hydrodynamics with multi-group frequency-dependent radiative transfer, nuclear reaction networks in collapsar supernovae, or long-term global simulations spanning millions of dynamical timescales—are computationally dominated by non-gravitational microphysics. Using Φ_PW(r) in existing Newtonian Eulerian or smoothed-particle hydrodynamics (SPH) codes captures the transonic inner plunge at r_ISCO = 3R_S with zero metric-tensor overhead.

Why does the Paczyński–Wiita Potential overestimate the Einstein perihelion precession at large distances?

In general relativity, spatial curvature (g_rr) and gravitational time dilation (g_tt) both contribute to perihelion advance, producing Δϖ = 3*πR*_S / a(1 − e²) per orbit. Because a scalar Newtonian potential modifies only the radial force law without curving three-dimensional space, matching the strong-field horizons at 2R_S and 3R_S causes the post-Newtonian r⁻² potential term to be twice as large as the weak-field limit requires, yielding Δϖ_PW = 6*πR*_S / a(1 − e²) at r ≫ R_S.

Can the Paczyński–Wiita Potential be used inside r = R_S?

No. Because Φ_PW(r) diverges to −∞ at r = R_S, numerical simulations place an absorbing inflow boundary condition slightly outside the Schwarzschild radius (typically at r_in = 1.1 R_S to 1.5 R_S), where the inflowing gas is already highly supersonic (Mach number M ≫ 1) after passing through the sonic point near r_ISCO = 3R_S, so no acoustic signals can propagate back outward into the disk.

1. Paczyński, B., & Wiita, P. J. (1980). Thick accretion disks and supercritical luminosities. Astronomy and Astrophysics, 88(1–2), 23–31. https://ui.adsabs.harvard.edu/abs/1980A%26A....88...23P
2. Abramowicz, M. A., Jaroszyński, M., & Sikora, M. (1978). Relativistic accreting disks. Astronomy and Astrophysics, 63(1–2), 221–224. https://ui.adsabs.harvard.edu/abs/1978A%26A....63..221A
3. Artemova, I. V., Bjoernsson, G., & Novikov, I. D. (1996). Modified Newtonian potentials for the description of relativistic effects in accretion disks around black holes. The Astrophysical Journal, 461, 565–571. https://doi.org/10.1086/177084
4. Tejeda, E., & Rosswog, S. (2013). An accurate Newtonian description of particle motion around a Schwarzschild black hole. Monthly Notices of the Royal Astronomical Society, 433(3), 1930–1940. https://doi.org/10.1093/mnras/stt853