Effective Potential for Massive Particle Orbits

Geodesic motion of a massive particle around a spherical mass looks, at first, like an intractable problem: four coupled second-order equations in a curved four-dimensional spacetime. Two symmetries of the Schwarzschild geometry collapse it almost to triviality. Time-translation invariance makes $\partial_t$ a Killing vector and hands us a conserved specific energy $E$ ; axial symmetry makes $\partial_\phi$ a Killing vector and hands us a conserved specific angular momentum $L$ . Fix the motion to the equatorial plane $\theta=\pi/2$ — always allowed by spherical symmetry — impose the timelike normalization $g_{\mu\nu}u^\mu u^\nu=-1$ , and the radial coordinate obeys a single equation that is formally identical to nonrelativistic energy conservation for a unit-mass particle in a potential well. This is one of the cleanest reductions in all of physics: a hard problem in curved spacetime rewritten as a marble rolling in a one-dimensional landscape. The landscape is the effective potential $V_{\mathrm{eff}}(r)$ , and it is the Newtonian gravitational-plus-centrifugal potential with exactly one new term, $-ML^2/r^3$ , of purely relativistic origin. That single term rewrites the entire phenomenology. It falls off faster than the centrifugal barrier, so at small radius it always wins: the relativistic centrifugal wall is finite, not infinite, and a particle can plunge to the center even with angular momentum to spare — impossible in Kepler's universe. The competition between the three terms organizes everything downstream in this chapter: circular orbits and their stability, the innermost stable circular orbit at $r=6M$ , the perihelion precession of Mercury, gravitational capture, and the inspiral of compact binaries. Throughout we use geometrized units $G=c=1$ , so $M$ carries units of length and the Schwarzschild radius is $2M$ ; $E$ is dimensionless and $L$ has units of length, both taken per unit rest mass.

Key takeaways

The two Killing vectors $\partial_t$ and $\partial_\phi$ of Schwarzschild spacetime supply conserved energy $E$ and angular momentum $L$ per unit rest mass, reducing four-dimensional geodesic motion to a single radial equation.
That radial equation is Newtonian in form, $\tfrac12\dot r^2+V_{\mathrm{eff}}(r)=\tfrac12(E^2-1)$ , with $r$ a position, proper time $\tau$ the clock, and $\tfrac12(E^2-1)$ a total mechanical energy.
The potential $V_{\mathrm{eff}}=-\tfrac{M}{r}+\tfrac{L^2}{2r^2}-\tfrac{ML^2}{r^3}$ differs from Newton's by the single term $-ML^2/r^3$ , yet that term governs every distinctively relativistic orbit.
Extrema of $V_{\mathrm{eff}}$ are circular orbits; minima are stable, maxima unstable. The minimum vanishes when $L^2<12M^2$ , producing the innermost stable circular orbit at $r=6M$ .
The ISCO carries $L=2\sqrt3\,M$ , $E=\sqrt{8/9}\approx0.943$ , so its binding energy is $\approx5.7\%$ of rest mass — the maximum efficiency of thin-disk accretion onto a Schwarzschild black hole.
Key idea: two Killing vectors plus the geodesic normalization turn 4D geodesic motion into $\tfrac12\dot r^2+V_{\mathrm{eff}}(r)=\tfrac12(E^2-1)$ ; the potential equals the Newtonian one plus a single $-ML^2/r^3$ correction that governs all distinctively relativistic orbital behavior.

Where this lives today

The effective potential is no longer a textbook exercise — it is the backbone of how we weigh and study black holes. In X-ray binaries and active galactic nuclei, the inner edge of the accretion disk sits at the ISCO, and fitting the thermal continuum or the relativistically broadened iron K $\alpha$ line to the predicted disk profile measures the ISCO radius, which for a spinning (Kerr) hole shrinks from $6M$ toward $M$ and so returns the black hole's spin. The binding energy at the ISCO sets the $\approx6\%$ Schwarzschild — up to $\approx42\%$ for a maximally rotating hole — radiative efficiency that makes quasars the brightest steady beacons in the cosmos. In gravitational-wave astronomy, the LIGO–Virgo–KAGRA network watches compact binaries sweep down the sequence of quasi-circular orbits as they radiate; the transition from inspiral to plunge near the effective ISCO shapes the final chirp and appears in the waveform templates used to detect and characterize every merger. And the Event Horizon Telescope's images of M87 and Sgr A resolve emission from the very region between the photon sphere at $r=3M$ and the ISCO at $r=6M$ , so the bright ring and central shadow are, quite literally, a photograph of the landscape this article maps.

The physical picture

Picture a marble rolling along a track whose height is $V_{\mathrm{eff}}(r)$ . The marble's energy is a horizontal line at height $\tfrac12(E^2-1)$ ; the marble can only be where that line lies above the track, and it turns around where the two meet. A dip in the track traps the marble in a bound, precessing orbit; the very bottom of the dip is a stable circular orbit, where the marble sits at rest in $r$ while circling in $\phi$ ; a hilltop is an unstable circular orbit, balanced and ready to roll either way. The whole theory of orbits reduces to reading this landscape — and the landscape has one feature no Newtonian potential can show. The Newtonian centrifugal barrier $L^2/2r^2$ is an infinitely high inner wall that keeps any spinning particle away from the center forever. General relativity drills straight through that wall with the $-ML^2/r^3$ term: at small enough $r$ the barrier turns over, the potential plunges toward $-\infty$ , and a particle with too little angular momentum or too much energy simply rolls over the crest and is swallowed.

Figure

Effective potential for L = 4M in Schwarzschild geometry. A stable circular orbit sits in the well minimum at r = 12M with E = 0.962, an unstable circular orbit crowns the barrier at r = 4M with E = 1, and because the barrier is finite the potential plunges toward r = 0 — any particle carried above the crest is captured.

Lowering the angular momentum flattens and lowers the whole landscape. Above a critical spin the track still has both a dip and a hill; at the critical value the dip and hill slide together and merge into a single flat inflection point; below it the track is a featureless downhill slide into the center and no circular orbit of any kind survives. That merger is the geometric origin of the innermost stable circular orbit — the last radius at which matter can circle a black hole without spiraling in — and it fixes the inner edge of every accretion disk. The second figure is the quantity an astrophysicist actually reaches for: the family of potentials that shows the ISCO being born.

Figure

The birth of the ISCO. Lowering the angular momentum flattens the well: for L² = 20M² a deep minimum and a barrier maximum coexist, at the critical L² = 12M² they merge into one inflection point at r = 6M, and below it (L² = 8M²) no circular orbit survives and every trajectory plunges.

The core equations

The two conserved quantities are the whole engine of the reduction. Each Killing vector contracts with the four-velocity to give a constant of the motion, and in Schwarzschild coordinates those constants are the specific energy and specific angular momentum.

E=\left(1-\frac{2M}{r}\right)\frac{\mathrm{d}t}{\mathrm{d}\tau},\qquad L=r^2\frac{\mathrm{d}\phi}{\mathrm{d}\tau} \qquad (1)

where: $E$ — specific energy (per unit rest mass), dimensionless, conserved because $\partial_t$ is Killing; $L$ — specific angular momentum (per unit rest mass), units of length, conserved because $\partial_\phi$ is Killing; $M$ — geometric mass of the source ( $G=c=1$ , so length); $\tau$ — proper time, the affine parameter of a timelike geodesic. Reading: the redshift factor $(1-2M/r)$ makes coordinate time $t$ run faster than proper time near the mass, which is why $E$ (not $\dot t$ ) is the constant. Eliminating $\dot t$ and $\dot\phi$ from the timelike normalization $g_{\mu\nu}u^\mu u^\nu=-1$ leaves a single equation in $r$ alone. It is algebraically identical to nonrelativistic energy conservation, which is the entire reason the effective-potential picture works.

\frac12\left(\frac{\mathrm{d}r}{\mathrm{d}\tau}\right)^2+V_{\mathrm{eff}}(r)=\frac12\left(E^2-1\right) \qquad (2)

where: $\mathrm{d}r/\mathrm{d}\tau$ — radial proper velocity, the "speed" of the fictitious 1D particle; $V_{\mathrm{eff}}(r)$ — effective potential, defined next; $\tfrac12(E^2-1)$ — the conserved total mechanical energy of the 1D problem, set by the real particle's $E$ . Reading: motion is allowed only where $\tfrac12(E^2-1)\ge V_{\mathrm{eff}}$ , and radial turning points are where equality holds. Written out, the potential is the Newtonian gravitational well plus the centrifugal barrier plus one extra term. That third term is the entire content of general relativity for this problem.

V_{\mathrm{eff}}(r)=-\frac{M}{r}+\frac{L^2}{2r^2}-\frac{ML^2}{r^3} \qquad (3)

where: $-M/r$ — Newtonian gravitational potential; $L^2/2r^2$ — centrifugal (angular-momentum) barrier, identical to Newton; $-ML^2/r^3$ — the relativistic correction, absent in Newtonian gravity and responsible for every effect below. Reading: the correction falls off one power of $r$ faster than the barrier, so it dominates at small $r$ and drags the potential to $-\infty$ . The same potential has a compact closed form that displays its geometric origin — a redshift factor multiplying a rest-mass-plus-centrifugal factor. The two forms are algebraically identical (Derivation 1).

V_{\mathrm{eff}}(r)=\frac12\left[\left(1-\frac{2M}{r}\right)\left(1+\frac{L^2}{r^2}\right)-1\right] \qquad (4)

where: $(1-2M/r)$ — the metric redshift factor $-g_{tt}$ ; $(1+L^2/r^2)$ — combines the rest-mass energy and the centrifugal contribution. Reading: expanding the product and cancelling the constant reproduces equation (3) exactly. Circular orbits sit at extrema of $V_{\mathrm{eff}}$ . Setting the derivative to zero and clearing $r^4$ gives a quadratic whose two roots are the two circular-orbit radii for a given $L$ .

\frac{\mathrm{d}V_{\mathrm{eff}}}{\mathrm{d}r}=0\;\Rightarrow\; Mr^2-L^2 r+3ML^2=0\;\Rightarrow\; r_\pm=\frac{L^2\pm\sqrt{L^4-12M^2L^2}}{2M} \qquad (5)

where: $r_+$ — the larger root, a minimum of $V_{\mathrm{eff}}$ and hence a stable circular orbit; $r_-$ — the smaller root, a maximum and hence an unstable circular orbit. Reading: real distinct roots require the discriminant $L^4-12M^2L^2>0$ , i.e. $L^2>12M^2$ . When the discriminant vanishes the two roots coalesce into a single degenerate extremum: the marginally stable orbit, or ISCO. Its radius, angular momentum, and energy are pure numbers times powers of $M$ .

r_{\mathrm{ISCO}}=6M,\qquad L_{\mathrm{ISCO}}=2\sqrt3\,M,\qquad E_{\mathrm{ISCO}}=\sqrt{\tfrac89}\approx0.9428 \qquad (6)

where: $r_{\mathrm{ISCO}}$ — innermost stable circular orbit radius; $L_{\mathrm{ISCO}}$ — its specific angular momentum, from $L^2=12M^2$ ; $E_{\mathrm{ISCO}}$ — its specific energy. Reading: the fractional binding energy is $1-E_{\mathrm{ISCO}}\approx5.7\%$ , the ceiling on Schwarzschild thin-disk accretion efficiency. For any circular orbit — not just the ISCO — the energy and angular momentum follow from the two extremum conditions. These closed forms feed directly into the worked examples.

E^2=\frac{(1-2M/r)^2}{1-3M/r},\qquad L^2=\frac{Mr^2}{r-3M} \qquad (7)

where: $r$ — radius of the circular orbit; the denominators show both quantities diverge as $r\to3M^+$ . Reading: no timelike circular orbit exists inside $r=3M$ (the photon sphere), where $L$ and $E$ blow up; the marginally bound orbit ( $E=1$ ) sits at $r=4M$ .

Derivation AReduction of geodesic motion to a one-dimensional energy problem

The goal is to show, step by step, how the two Killing vectors of Schwarzschild spacetime and the timelike normalization of the four-velocity collapse the full set of geodesic equations into a single first-order equation for $r(\tau)$ that reads exactly like Newtonian energy conservation, and to identify from it the effective potential in both its expanded and closed forms. Every distinctively relativistic orbital effect in this chapter is a corollary of this one reduction, so it is worth doing carefully and checking that the expanded and closed forms agree.

Step 1. The Schwarzschild metric admits $\partial_t$ and $\partial_\phi$ as Killing vectors, and for any geodesic the contraction of a Killing vector with the four-velocity $u^\mu=\mathrm{d}x^\mu/\mathrm{d}\tau$ is conserved. This gives $E=-(\partial_t)_\mu u^\mu=(1-2M/r)\dot t$ and $L=(\partial_\phi)_\mu u^\mu=r^2\dot\phi$ in the equatorial plane $\theta=\pi/2$ , with an overdot meaning $\mathrm{d}/\mathrm{d}\tau$ . Why: symmetries are the only handle we have on a curved-spacetime problem, and each one turns a differential equation into an algebraic constant, removing $t$ and $\phi$ from the problem.

Step 2. Write the timelike normalization $g_{\mu\nu}u^\mu u^\nu=-1$ explicitly in the equatorial plane: $-(1-2M/r)\dot t^2+(1-2M/r)^{-1}\dot r^2+r^2\dot\phi^2=-1$ . This is the statement that proper time is the arc length of a timelike geodesic, and it is the fourth equation we need to close the system. Why: the three Killing-type relations alone underdetermine the motion; the normalization supplies the constraint that ties $\dot r$ to $\dot t$ and $\dot\phi$ .

Step 3. Substitute $\dot t=E/(1-2M/r)$ and $\dot\phi=L/r^2$ from Step 1 to eliminate the ignorable coordinates, giving $-E^2/(1-2M/r)+\dot r^2/(1-2M/r)+L^2/r^2=-1$ . Now only $r$ and its derivative remain. Why: replacing $\dot t,\dot\phi$ by the constants $E,L$ is exactly where the four-dimensional problem becomes one-dimensional — the angular and time motion are slaved to $r$ .

Step 4. Multiply through by $(1-2M/r)$ and rearrange to isolate $\dot r^2=E^2-(1-2M/r)(1+L^2/r^2)$ . The right side is manifestly a function of $r$ alone with $E$ as a parameter, so the radial motion is decoupled and first order. Why: a first-order equation of the form $\dot r^2=(\text{function of }r)$ is precisely an energy-conservation statement in disguise, which is the whole point of the reduction.

Step 5. Halve and regroup: $\tfrac12\dot r^2+\tfrac12[(1-2M/r)(1+L^2/r^2)-1]=\tfrac12(E^2-1)$ . Reading off the bracket as $V_{\mathrm{eff}}$ gives the closed form (4); expanding the product, $(1-2M/r)(1+L^2/r^2)=1+L^2/r^2-2M/r-2ML^2/r^3$ , and cancelling the $\pm1$ and halving reproduces the expanded form $V_{\mathrm{eff}}=-M/r+L^2/2r^2-ML^2/r^3$ of (3). Why: matching the equation to the template $\tfrac12\dot r^2+V=\text{const}$ is what lets us borrow the entire intuition of nonrelativistic mechanics, and confirming both forms of $V_{\mathrm{eff}}$ agree guards against an algebra slip.

Go deeper · Derivation B · 2 minThe innermost stable circular orbit at six geometric radii

Derivation BThe innermost stable circular orbit at six geometric radii

The goal is to locate the last stable circular orbit by demanding that the effective potential have a degenerate extremum — a point where the first and second derivatives both vanish, so that the stable minimum and the unstable maximum have merged into a single inflection point — and then to read off its radius, angular momentum, energy, and binding energy. This is the calculation that fixes the inner edge of an accretion disk and the transition from inspiral to plunge in a compact binary, so the numbers matter and must be checked against the standard references.

Step 1. A circular orbit requires $V'_{\mathrm{eff}}=0$ . Differentiating the expanded potential gives $V'_{\mathrm{eff}}=M/r^2-L^2/r^3+3ML^2/r^4$ , and multiplying by $r^4$ yields the circular-orbit condition $Mr^2-L^2r+3ML^2=0$ . Why: extrema of $V_{\mathrm{eff}}$ are exactly the radii where the radial force vanishes, so a particle placed there with the matching $L$ circles at constant $r$ .

Step 2. Marginal stability additionally requires $V''_{\mathrm{eff}}=0$ , the condition that the minimum and maximum coincide. Differentiating again gives $V''_{\mathrm{eff}}=-2M/r^3+3L^2/r^4-12ML^2/r^5$ , and multiplying by $r^5$ yields $-2Mr^2+3L^2r-12ML^2=0$ . Why: a stable orbit needs $V''_{\mathrm{eff}}>0$ and an unstable one $V''_{\mathrm{eff}}<0$ ; the boundary between them, where stability is marginal, is precisely $V''_{\mathrm{eff}}=0$ .

Step 3. Solve the circular-orbit condition of Step 1 for the angular momentum, $L^2=Mr^2/(r-3M)$ , and substitute into the Step 2 equation. After clearing the common denominator the two curvature terms combine and the algebra collapses to the single linear condition $r-6M=0$ . Why: eliminating $L^2$ between the two conditions leaves an equation in $r$ alone, and only its physical root ( $r>3M$ ) corresponds to a genuine timelike orbit.

Step 4. Hence $r_{\mathrm{ISCO}}=6M$ . Back-substituting into $L^2=Mr^2/(r-3M)$ gives $L^2=M(6M)^2/(3M)=12M^2$ , so $L_{\mathrm{ISCO}}=2\sqrt3\,M$ , and into $E^2=(1-2M/r)^2/(1-3M/r)$ gives $E^2=(2/3)^2/(1/2)=8/9$ , so $E_{\mathrm{ISCO}}=\sqrt{8/9}\approx0.9428$ . Why: the same two extremum conditions that fixed $r$ also fix the orbit's conserved quantities, turning a geometric statement into concrete, measurable numbers.

Step 5. The binding energy per unit rest mass is $1-E_{\mathrm{ISCO}}=1-\sqrt{8/9}\approx0.0572$ , i.e. about $5.7\%$ . This is the energy that must be radiated for matter to spiral quasi-statically from rest at infinity down to the ISCO. Why: for a thin accretion disk the matter drifts inward through circular orbits and plunges from the ISCO, so this binding energy is the maximum fraction of rest mass a Schwarzschild black hole can convert to radiation.

Go deeper · Derivation C · 2 minEnergy and angular momentum along a circular orbit

Derivation CEnergy and angular momentum along a circular orbit

The goal is to derive the closed forms $E^2=(1-2M/r)^2/(1-3M/r)$ and $L^2=Mr^2/(r-3M)$ that give the conserved quantities of any circular orbit as functions of its radius, and to extract from them two landmark radii — the photon sphere at $r=3M$ where both diverge, and the marginally bound orbit at $r=4M$ where $E=1$ . These formulas are the workhorse of the worked examples and the bridge from the effective-potential picture to real astrophysical numbers.

Step 1. Start from the circular-orbit condition $Mr^2-L^2r+3ML^2=0$ of Derivation 2 and solve for $L^2$ : collecting the $L^2$ terms gives $L^2(3M-r)=-Mr^2$ , hence $L^2=Mr^2/(r-3M)$ . Why: expressing $L$ as a function of orbital radius lets us label every circular orbit by a single number, its radius, and immediately shows the orbit ceases to exist as $r\to3M$ .

Step 2. For a circular orbit $\dot r=0$ , so equation (2) reduces to $E^2=(1-2M/r)(1+L^2/r^2)$ . Substitute $L^2/r^2=M/(r-3M)$ from Step 1, so that $1+L^2/r^2=(r-3M+M)/(r-3M)=(r-2M)/(r-3M)$ . Why: setting $\dot r=0$ is the defining feature of a circular orbit, and it turns the radial energy equation into a direct expression for $E$ in terms of $r$ .

Step 3. Combine: $E^2=(1-2M/r)\cdot(r-2M)/(r-3M)=(r-2M)^2/[r(r-3M)]=(1-2M/r)^2/(1-3M/r)$ . Both $E$ and $L$ carry the factor $(r-3M)$ in a denominator and diverge as $r\to3M^+$ — the photon sphere, the innermost circular orbit of any kind, reached only in the null limit. Why: the common singularity at $r=3M$ is the deep statement that massive circular orbits must move ever closer to the speed of light as they approach the photon sphere, requiring unbounded energy.

Step 4. Set $E=1$ to find the marginally bound orbit — the unstable circular orbit whose energy just equals the rest-mass energy of a particle at infinity. Then $(1-2M/r)^2=1-3M/r$ , which expands to $4M^2/r^2=M/r$ , giving $r=4M$ . Any less bound unstable orbit ( $E>1$ ) lies between $4M$ and $3M$ and can be reached from infinity. Why: $E=1$ marks the boundary between capture and escape for an unstable orbit, and $r=4M$ is the radius at which a marginally bound particle from infinity can asymptotically circle — a key radius for gravitational capture.

Limiting cases

The effective potential interpolates between familiar and exotic regimes according to which of its three terms dominates.

Case	Result	Meaning
Newtonian limit $r\gg M$	$V_{\mathrm{eff}}\to-\dfrac{M}{r}+\dfrac{L^2}{2r^2}$	The $-ML^2/r^3$ term is negligible; orbits are closed Keplerian ellipses, the centrifugal barrier is infinite, and there is no ISCO and no capture.
Relativistic core $r\lesssim M$	$V_{\mathrm{eff}}\to-\dfrac{ML^2}{r^3}\to-\infty$	The relativistic term overwhelms the barrier; the potential plunges, so even a high- $L$ particle can reach $r=0$ . No Newtonian analogue.
Zero angular momentum $L=0$	$V_{\mathrm{eff}}=-\dfrac{M}{r}$	Purely radial infall; the potential is monotonic and identical in form to Newton, and the particle falls straight in with no barrier.
Photon-sphere approach $r\to3M^+$	$E,L\to\infty$	Circular timelike orbits require diverging energy and angular momentum; $r=3M$ is the innermost circular orbit of any kind, reached only by light.
Marginal stability $L^2=12M^2$	$V''_{\mathrm{eff}}(6M)=0,\ r_{\mathrm{ISCO}}=6M$	Minimum and maximum merge into an inflection point — the innermost stable circular orbit and the inner edge of an accretion disk.

Common trapThe relativistic term is a small correction, so it can be treated as a perturbation everywhere

It cannot. The term $-ML^2/r^3$ is indeed tiny for planetary orbits — it shifts Mercury's perihelion by a mere $43''$ per century — but it is a singular perturbation, not a uniform one. Near the center it does not merely correct the barrier; it inverts it. However large $L$ is, there is always a radius where $ML^2/r^3$ exceeds $L^2/2r^2$ (namely $r<2M$ ), so the infinite Newtonian wall is replaced by a finite hill of height $V_{\mathrm{eff}}(r_-)$ that a sufficiently energetic particle rolls straight over. Treating the term perturbatively throws away exactly the physics that has no Newtonian counterpart: the ISCO, gravitational capture, the plunge, and the very existence of a black hole's marginally stable orbit. Three companion traps: $\tfrac12(E^2-1)$ is the effective energy, not $E$ ; $E$ and $L$ are per unit rest mass; and $r$ is the areal Schwarzschild coordinate, not proper radial distance, while $\tau$ (not coordinate time $t$ ) is the clock.

Worked examples

1 · The ISCO around a ten-solar-mass black hole (confidence builder)

A stellar-mass black hole has $M=10\,M_\odot$ . Find the radius of its innermost stable circular orbit in kilometers, the specific energy and angular momentum there, and the maximum efficiency with which it can radiate the rest mass of accreted gas.

Step 1. Convert the mass to geometric units. One solar mass corresponds to $GM_\odot/c^2=1.477\ \mathrm{km}$ , so the geometric mass is $M=10\times1.477=\mathbf{14.77\ km}$ . In these units the Schwarzschild radius is $2M=29.5$ km.

Step 2. The ISCO radius is $r_{\mathrm{ISCO}}=6M$ regardless of the mass, so $r_{\mathrm{ISCO}}=6\times14.77=\mathbf{88.6\ km}$ . Its angular momentum and energy are the universal numbers $L_{\mathrm{ISCO}}=2\sqrt3\,M=3.46\times14.77=51.2$ km and $E_{\mathrm{ISCO}}=\sqrt{8/9}=\mathbf{0.9428}$ .

Step 3. The maximum radiative efficiency is the binding energy of the ISCO, $\eta=1-E_{\mathrm{ISCO}}=1-0.9428=\mathbf{0.0572}$ , i.e. $5.7\%$ of the accreted rest mass. For comparison, hydrogen fusion converts only $\approx0.7\%$ ; a non-rotating black hole is already about eight times as efficient a power source, which is why accretion onto compact objects lights up the most luminous persistent sources in the universe. r_ISCO = 88.6 km, E = 0.943, and up to 5.7% of accreted rest mass is radiated.

2 · The two circular orbits for a given angular momentum (fully worked)

A massive particle in Schwarzschild geometry has specific angular momentum $L=4M$ . Find the radii of its two circular orbits, classify each as stable or unstable, and compute the specific energy of the stable one.

Step 1. Apply the circular-orbit roots of equation (5). With $L^2=16M^2$ the discriminant is $L^4-12M^2L^2=256M^4-192M^4=64M^4$ , whose square root is $8M^2$ . Then $r_\pm=(16M^2\pm8M^2)/2M$ , giving $r_+=24M^2/2M=\mathbf{12M}$ and $r_-=8M^2/2M=\mathbf{4M}$ .

Step 2. Classify by the second derivative, or equivalently by the rule that the larger root is the minimum. The outer orbit $r_+=12M$ is a minimum of $V_{\mathrm{eff}}$ , hence a stable circular orbit; the inner orbit $r_-=4M$ is a maximum, hence an unstable circular orbit poised to plunge or escape. Both lie outside the photon sphere $r=3M$ , as any timelike circular orbit must.

Step 3. Compute the energy of the stable orbit from equation (7): at $r=12M$ , $E^2=(1-2M/r)^2/(1-3M/r)=(1-1/6)^2/(1-1/4)=(5/6)^2/(3/4)=(25/36)(4/3)=25/27$ . So $E=\sqrt{25/27}=\mathbf{0.9623}$ , a binding energy of $3.8\%$ — less deeply bound than the ISCO, as expected for an orbit outside $6M$ .

Step 4. Check the unstable orbit for insight: at $r=4M$ , $E^2=(1-1/2)^2/(1-3/4)=(1/4)/(1/4)=1$ , so $E=1$ exactly. This inner orbit is marginally bound — its energy equals the rest-mass energy at infinity — so a particle nudged outward from it just barely escapes to infinity, while a particle nudged inward plunges. The orbits are r = 12M (stable, E = 0.962) and r = 4M (unstable, E = 1, marginally bound).

3 · Perihelion precession of Mercury

Use the leading relativistic precession formula $\Delta\phi=6\pi M/[a(1-e^2)]$ to compute the perihelion advance of Mercury per century, given semi-major axis $a=5.79\times10^{10}\ \mathrm{m}$ , eccentricity $e=0.2056$ , orbital period $0.2409\ \mathrm{yr}$ , and $M_\odot=1476.6\ \mathrm{m}$ in geometric units.

Step 1. Evaluate the semi-latus rectum $a(1-e^2)=5.79\times10^{10}\times(1-0.0423)=5.79\times10^{10}\times0.9577=5.545\times10^{10}\ \mathrm{m}$ . This is the length scale that sets the size of the relativistic correction.

Step 2. The advance per orbit is $\Delta\phi=6\pi M_\odot/[a(1-e^2)]=6\pi\times1476.6/(5.545\times10^{10})=2.784\times10^4/5.545\times10^{10}=\mathbf{5.02\times10^{-7}\ rad}$ . The $6\pi M$ numerator is a direct fingerprint of the $-ML^2/r^3$ term: expanding the radial oscillation about the $V_{\mathrm{eff}}$ minimum makes the radial period slightly exceed the azimuthal period, so the ellipse fails to close.

Step 3. Convert to arcseconds per century. Mercury completes $100/0.2409=415.1$ orbits per century, so the accumulated advance is $415.1\times5.02\times10^{-7}=2.084\times10^{-4}\ \mathrm{rad}$ . Multiplying by $206265''/\mathrm{rad}$ gives $\mathbf{43.0''}$ per century — the famous anomaly Le Verrier reported and Einstein explained, matching observation to within measurement error. The relativistic term precesses Mercury's perihelion by 43″ per century.

Physics of the portrait

The effective potential for massive orbits is inseparable from the man who first wrote down the geometry it lives in. Within weeks of Einstein's field equations appearing, an astronomer at the Russian front solved them exactly for a spherical mass — and everything in this article is a reading of the landscape hidden inside that solution.

Origins · 1 minKarl Schwarzschild and how this was discovered

Karl Schwarzschild

1873–1916 · the exact spherical solution of general relativity

Karl Schwarzschild was already a leading astrophysicist — director of the Potsdam Observatory, a pioneer of stellar photometry and radiative transfer — when the First World War broke out and he, past forty, volunteered for the German army. Serving on the Russian front and computing artillery trajectories, he read Einstein's newly published field equations of November 1915 and, within weeks, found their first exact solution: the metric of the spacetime outside a spherical mass. He communicated it to Einstein in a letter of December 1915, and Einstein, who had expected only approximate solutions, presented it to the Prussian Academy on Schwarzschild's behalf. A second paper gave the interior solution for a uniform-density sphere. It was this exterior metric — with its redshift factor $(1-2M/r)$ and its two Killing vectors — that turns geodesic motion into the effective-potential problem of this article, and from which the innermost stable circular orbit, the photon sphere, and Mercury's precession all follow. Schwarzschild contracted pemphigus, a rare and then-incurable autoimmune disease, at the front and died in May 1916, only months after his solution appeared and long before anyone understood that the radius $r=2M$ marks a black-hole horizon. The geometric mass $M$ , the Schwarzschild radius $2M$ , and the entire orbital phenomenology built on them carry his name.

History

1859

— Le Verrier reports the anomalous 43″ per century advance of Mercury's perihelion

1915

— Einstein derives Mercury's precession from the field equations of general relativity

1916

— Schwarzschild publishes the exact vacuum solution for a spherical mass

1916

— Schwarzschild dies of an illness contracted on the Eastern Front

1973

— Novikov, Thorne, and Bardeen place the inner edge of a thin accretion disk at the ISCO

Going further

Texts. Misner, Thorne & Wheeler, Gravitation (1973), §25.5 — the effective-potential treatment of Schwarzschild orbits · Wald, General Relativity (1984), §6.3 — circular orbits and the ISCO · Carroll, Spacetime and Geometry (2004), §5.4 — the reduction to $V_{\mathrm{eff}}$ and the $-ML^2/r^3$ term · Hartle, Gravity (2003), Ch. 9 — orbits, precession, and the ISCO with worked numbers · Schutz, A First Course in General Relativity, 2nd ed. (2009), §11.1 — geodesics of Schwarzschild · Chandrasekhar, The Mathematical Theory of Black Holes (1983), Ch. 3 — the definitive treatment of circular and radial geodesics. The key results of this article, all in geometrized units $G=c=1$ , are: the effective potential $V_{\mathrm{eff}}=-M/r+L^2/2r^2-ML^2/r^3$ and its closed form $\tfrac12[(1-2M/r)(1+L^2/r^2)-1]$ ; the circular-orbit condition $Mr^2-L^2r+3ML^2=0$ with roots $r_\pm=[L^2\pm\sqrt{L^4-12M^2L^2}]/2M$ ; the ISCO at $r=6M$ with $L=2\sqrt3\,M$ , $E=\sqrt{8/9}\approx0.9428$ , and $5.7\%$ binding energy; the circular-orbit relations $E^2=(1-2M/r)^2/(1-3M/r)$ and $L^2=Mr^2/(r-3M)$ , with the photon sphere at $r=3M$ and the marginally bound orbit at $r=4M$ ; and Mercury's perihelion precession of $43''$ per century from $\Delta\phi=6\pi M/[a(1-e^2)]$ .

Prerequisites. The Schwarzschild Metric · Killing Vectors and Conserved Quantities in Schwarzschild · Geodesic Equations in Schwarzschild Spacetime · Schwarzschild Radius and the Geometric Mass. Leads to. Circular Orbits and Their Stability · Innermost Stable Circular Orbit (ISCO) · Perihelion Precession in Schwarzschild Geometry · Photon Orbits and the Photon Sphere.

Ask the tutor: Ask me to sketch $V_{\mathrm{eff}}(r)$ for any angular momentum you choose and read off its circular orbits, or to walk a particle down the sequence of orbits from infinity to the ISCO and show where it plunges.

Why is the radial equation written in proper time τ rather than coordinate time t?

Because the geodesic equation and the normalization $u^\mu u_\mu=-1$ are statements about an affine parameter, which for a timelike geodesic is proper time — the time read by a clock carried along the orbit. Coordinate time $t$ runs at a position-dependent rate, $\mathrm{d}t/\mathrm{d}\tau=E/(1-2M/r)$ , so using $t$ would smuggle gravitational time dilation into the "kinetic energy" term and destroy the clean analogy with a nonrelativistic particle. The whole power of the effective-potential picture rests on choosing the parameter that makes the radial equation look Newtonian, and that parameter is $\tau$ .

What exactly is the difference between E and the effective energy ½(E²−1)?

$E$ is the physical conserved quantity — the specific energy of the particle, equal to its energy per unit rest mass as measured at infinity, and dimensionless in geometrized units. The combination $\tfrac12(E^2-1)$ is a bookkeeping device: it is the constant that plays the role of "total mechanical energy" in the one-dimensional analogue problem $\tfrac12\dot r^2+V_{\mathrm{eff}}=\tfrac12(E^2-1)$ . A particle at rest at infinity has $E=1$ , so $\tfrac12(E^2-1)=0$ ; a bound particle has $E<1$ and a negative effective energy; an unbound one has $E>1$ . Confusing $E$ with $\tfrac12(E^2-1)$ is the single most common error in setting up these problems.

Is the effective potential the same for a photon as for a massive particle?

No, though the two are close cousins. For a photon one sets the normalization to $g_{\mu\nu}u^\mu u^\nu=0$ instead of $-1$ , which drops the rest-mass "1" and leaves $V_{\mathrm{eff}}^{\mathrm{null}}=\tfrac{L^2}{2r^2}(1-2M/r)$ . This null potential has a single maximum at $r=3M$ — the photon sphere, an unstable circular light orbit — and no minimum, so there are no stable circular photon orbits at all. The massive potential, with its rest-mass term, additionally develops the $-M/r$ well that permits bound, stable, and marginally stable (ISCO) orbits. The photon case is treated fully in the photon-sphere article.

Why does lowering the angular momentum destroy the stable orbit rather than shrink it?

Because the stable minimum and the unstable maximum are the two roots of the same quadratic, $Mr^2-L^2r+3ML^2=0$ , and lowering $L$ moves them toward each other. As $L^2$ drops toward $12M^2$ the minimum (outer root) moves inward and the maximum (inner root) moves outward until, at exactly $L^2=12M^2$ , they collide at $r=6M$ and annihilate, leaving an inflection point. Below that value the quadratic has no real roots at all — there is no extremum, the potential is monotonically downhill, and any particle simply plunges. So the stable orbit does not shrink to zero radius; it disappears by merging with its unstable partner at a finite radius, which is precisely why the ISCO exists at $6M$ and not at the horizon.

How does the effective potential produce perihelion precession?

For a bound orbit the particle oscillates radially between a perihelion and an aphelion — the two turning points where the energy line meets $V_{\mathrm{eff}}$ — while simultaneously advancing in $\phi$ . In pure Newtonian gravity the radial oscillation period exactly equals the azimuthal period, so the ellipse closes. The $-ML^2/r^3$ term slightly changes the curvature of $V_{\mathrm{eff}}$ near its minimum, making the radial period a touch longer than the azimuthal one; the orbit therefore does not quite close, and the perihelion advances by $\Delta\phi=6\pi M/[a(1-e^2)]$ each revolution. It is the same single relativistic term, read in the weak-field regime instead of the strong-field one, that gives the ISCO deep inside.

Does the 5.7% ISCO efficiency mean a black hole radiates 5.7% of everything that falls in?

Only for idealized thin-disk accretion onto a non-rotating hole. The figure $1-\sqrt{8/9}\approx5.7\%$ is the binding energy of the ISCO: matter that spirals inward quasi-statically through nearly circular orbits must shed this much of its rest-mass energy as radiation to reach $r=6M$ , after which it plunges in and carries its remaining energy across the horizon. Real accretion is messier — some energy is advected inward or lost to winds, and rapid rotation moves the ISCO inward and pushes the efficiency toward $\approx42\%$ for a maximal Kerr hole. But the $5.7\%$ is the clean benchmark: it is why accretion onto compact objects, not fusion, powers the most luminous sources in the universe.