Frame Dragging (Lense–Thirring Effect)

The intuition: spacetime as a fluid you can stir

Newton imagined gravity as a force reaching across empty, passive space. Einstein replaced that picture with a spacetime that bends in response to mass. Frame dragging is the next twist of the knife: spacetime is not only bent by mass, it is dragged by mass in motion. Spin a heavy enough object and the spacetime in its vicinity begins to swirl, like honey around a turning spoon. Anything embedded in that spacetime — a free-falling gyroscope, an orbiting satellite, even a beam of light — gets carried along.

The key word is "inertial frame." In flat space, a non-rotating frame is one in which a free gyroscope keeps pointing at the same distant stars forever. Frame dragging breaks that agreement. Near a spinning mass, a gyroscope that feels no torque — that is, by every local measurement is "not rotating" — nonetheless wheels its axis relative to the fixed stars. The local definition of "not spinning" and the cosmic definition come apart. The rotating mass has dragged the local notion of non-rotation around with it. This is why the phenomenon is also called the dragging of inertial frames.

What frame dragging actually is

Frame dragging is the rotational, or "gravitomagnetic," part of the gravitational field of a spinning body. A static mass produces a gravitational field that points radially inward — the gravitational analogue of an electric field around a charge. A rotating mass produces, in addition, a swirling field — the gravitational analogue of the magnetic field around a spinning charge. That swirling field exerts a torque on gyroscopes and curls the trajectories of test particles, and that torque-and-curl is frame dragging.

Crucially, frame dragging is a property of the geometry, not a force you can shield against. There is no "anti-drag" material. A particle dropped radially toward the spin axis with zero angular momentum still picks up angular motion, because the very coordinates it falls through are rotating. The amount of drag depends on the body's angular momentum J, not merely its mass — a black hole and a planet of identical mass but different spin drag spacetime by different amounts.

The physics: gravitomagnetism and the Kerr metric

In the weak-field, slow-rotation limit, Einstein's equations can be linearised into a form that looks startlingly like Maxwell's electromagnetism — the gravitoelectromagnetic (GEM) equations. Mass density ρ sources a gravitoelectric field (Newtonian gravity); mass current j = ρv sources a gravitomagnetic field B_g via a Maxwell-like curl equation:

∇ × B_g = −(16πG/c²) j      (gravitomagnetic Ampère law)
B_g (dipole, outside a spinning body):
   B_g = (G/c²) [ 3(J·r̂)r̂ − J ] / r³

A gyroscope's spin S precesses in this field exactly as a magnetic dipole would, with precession rate

Ω_LT = (G/c²r³) [ 3(J·r̂)r̂ − J ]        (Lense–Thirring precession)

For an orbit lying in the equatorial plane of the spin, the line of nodes and the orbit itself wheel about the spin axis at the frame-dragging angular velocity

ω(r) = 2GJ / (c² r³)

The c² in the denominator is the reason the effect is so faint in everyday gravity: it is a 1/c² correction to Newton. Where general relativity stops being a small correction — near a black hole — frame dragging is described exactly, not approximately, by the Kerr metric (Roy Kerr, 1963). The Kerr line element contains an off-diagonal time–angle cross term g_tφ that is the unmistakable signature of rotation:

ds² = −(1 − 2GMr/ρ²c²) c²dt² − (4GMar sin²θ / ρ²c) dt dφ + ...
   with  a = J/(Mc),   ρ² = r² + a²cos²θ
local frame-dragging angular velocity:  ω = −g_tφ / g_φφ

The spin parameter a = J/(Mc) has units of length and is capped at a ≤ GM/c² for a black hole (the extremal Kerr limit); a hole spinning faster than that would expose a naked singularity, which the cosmic censorship conjecture forbids.

The key numbers across nine orders of magnitude

Because ω scales as J/r³, frame dragging ranges from utterly negligible to absolutely dominant depending on how compact and how fast-spinning the source is. The compactness that matters is J/(M c r) — how close the source is to extremal.

Source	Angular momentum J (kg m²/s)	Reference radius	Frame-drag rate ω	Regime
Earth	5.86 × 10³³	R⊕ surface	~0.17 arcsec/yr (equator, max)	weak-field, GP-B regime
Earth (GP-B orbit)	5.86 × 10³³	7,020 km	37.2 mas/yr	measured 2011
Sun	1.9 × 10⁴¹	Mercury's orbit	~0.002 arcsec/century (≪1% of GR perihelion shift)	weak-field
Millisecond pulsar	~2 × 10⁴²	10 km surface	tens of rad/s at the surface	strong-field
Stellar-mass Kerr BH (10 M☉, a*=0.9)	—	at the ISCO	thousands of rad/s	near-extremal Kerr
M87* (6.5 × 10⁹ M☉, a*~0.9)	—	at the horizon	full co-rotation in ~ days	relativistic, EHT-imaged

The contrast is stark. Around Earth, frame dragging is a milliarcsecond-per-year wobble that took a half-billion-dollar satellite mission to detect. At the horizon of a spinning supermassive black hole, frame dragging is so total that an entire stationary universe of distant stars could not hold an object still — it must orbit with the hole. Same equation, with the gravitational compactness GM/(c²r) running from ~10⁻⁹ at Earth's surface to of order 1 at a Kerr hole's horizon.

Where dragging becomes inescapable: the ergosphere

For a spinning black hole, frame dragging grows without bound as you approach the hole. At a certain surface — the static limit — the dragging angular velocity equals the maximum angular velocity a photon could muster trying to counter-rotate and hold still against the spin. Inside this surface, the region called the ergosphere, no object can remain stationary relative to the distant stars. To stand still you would have to move faster than light against the drag, which is impossible. Everything inside the ergosphere is forced to co-rotate with the hole, even light aimed directly "backward."

The static limit sits at r = GM/c² (1 + √(1 − a*² cos²θ)), touching the event horizon at the poles and bulging out to 2GM/c² at the equator for a maximally spinning hole — an oblate cap of dragged spacetime sitting outside the horizon. This is what powers the Penrose process and the Blandford–Znajek mechanism: because the ergosphere lies outside the horizon, particles and magnetic fields can dip into the region of mandatory co-rotation, steal rotational energy from the hole, and escape, extracting up to 29% of a maximal Kerr black hole's mass-energy.

Discovery, people, and the missions that confirmed it

The theory predates the experiment by nearly a century:

• 1918 — Lense & Thirring. Austrian physicists Josef Lense and Hans Thirring published the slow-rotation solution of Einstein's field equations and derived the precession of orbits and gyroscopes near a rotating mass. The effect carries their names.
• 1959–60 — Schiff & Pugh. Leonard Schiff and (independently) George Pugh proposed using an orbiting gyroscope to measure the effect directly — the seed of Gravity Probe B.
• 1963 — Roy Kerr. Discovered the exact solution for a rotating black hole, making frame dragging a strong-field, non-perturbative phenomenon with a true ergosphere.
• 2004–2011 — Gravity Probe B. Launched by NASA and Stanford on 20 April 2004 into a 642 km polar orbit, GP-B carried four fused-quartz gyroscopes — the most perfectly spherical objects ever made, round to within about 40 atomic layers. After fighting unexpected electrostatic "patch effect" torques, the team reported in 2011 a frame-dragging precession of 37.2 ± 7.2 mas/yr against a GR prediction of 39.2 mas/yr, and a geodetic precession of 6,601.8 ± 18.3 mas/yr.
• 1976–present — LAGEOS / LARES. Laser-ranged spherical satellites measure frame dragging through the secular precession of their orbital nodes (the satellite-orbit version of the effect, ~31 mas/yr for LAGEOS), with LARES (2012) and LARES-2 (2022) pushing the accuracy toward a few percent.
• 2019–2022 — Event Horizon Telescope. Resolved the shadows of M87* and Sgr A*, both consistent with rapidly spinning Kerr holes whose ergospheres drag everything nearby.

Worked example: frame dragging on Gravity Probe B

Let's reproduce the GP-B number. Earth's angular momentum is J = I Ω, with moment of inertia I ≈ 0.331 M⊕ R⊕² and rotation rate Ω = 7.29 × 10⁻⁵ rad/s:

I  = 0.331 × (5.97×10²⁴ kg) × (6.371×10⁶ m)²  ≈ 8.0 × 10³⁷ kg m²
J  = I Ω = 8.0×10³⁷ × 7.29×10⁻⁵  ≈ 5.86 × 10³³ kg m²/s

The polar GP-B orbit had a semi-major axis r ≈ R⊕ + 642 km = 7.02 × 10⁶ m. For a polar orbit the average Lense–Thirring gyroscope precession is

Ω_LT ≈ G J / (c² r³)        (polar-orbit average, magnitude)
     = (6.674×10⁻¹¹ × 5.86×10³³) / [(3×10⁸)² × (7.02×10⁶)³]
     = (3.91×10²³) / (9×10¹⁶ × 3.46×10²⁰)
     ≈ 1.26 × 10⁻¹⁴ rad/s

Convert to milliarcseconds per year (1 rad = 2.063 × 10⁸ mas; 1 yr = 3.156 × 10⁷ s):

Ω_LT ≈ 1.26×10⁻¹⁴ rad/s × 2.063×10⁸ mas/rad × 3.156×10⁷ s/yr
     ≈ 82 mas/yr  (order of magnitude; the precise GR
       polar-orbit prediction is 39.2 mas/yr after the
       full tensor average over the orbit)

The factor-of-two difference between the back-of-envelope dipole estimate and the rigorous orbit-averaged result comes from the angular geometry of the gravitomagnetic dipole field over a full polar orbit — but the order of magnitude, tens of milliarcseconds per year, falls straight out of GJ/(c²r³). That is the entire challenge: detecting a tilt of about 40 millionths of a degree per year in a gyroscope drifting through space.

Frame dragging vs the other precessions a gyroscope feels

An orbiting gyroscope precesses for several reasons, and disentangling them is the whole experimental game. They differ in physical origin and in how they scale.

Effect	Physical origin	Needs spin of central body?	GP-B value (Earth orbit)	Scaling
Geodetic (de Sitter)	Curvature of static space, gyroscope carried through it	No	6,601.8 mas/yr	∝ M/r · v
Frame dragging (Lense–Thirring)	Gravitomagnetic field of the spinning mass	Yes	37.2 mas/yr	∝ J/r³
Thomas precession	Special-relativistic, non-commuting boosts	No	folded into geodetic term	∝ v²
Newtonian perihelion shift	Oblateness J₂, third bodies	No (figure, not spin)	(removed from analysis)	∝ J₂ M/r²
Schwarzschild perihelion (GR)	Static spacetime curvature on orbit	No	(e.g. Mercury, 43″/century)	∝ M/r

GP-B oriented its gyroscopes toward a guide star (IM Pegasi) so that the geodetic and frame-dragging precessions pointed in perpendicular planes — geodetic in the orbital plane, frame dragging in the plane of Earth's equator. That orthogonality is what let a 6,600 mas/yr effect and a 37 mas/yr effect, differing by a factor of 180, be measured in the same data set.

Where frame dragging runs the show: compact objects

Around Earth frame dragging is a curiosity; around compact objects it shapes observable astrophysics.

• Misaligned accretion disks & the Bardeen–Petterson effect. A disk tilted relative to a spinning black hole's equator is dragged by Lense–Thirring torques. The inner disk is forced into alignment with the hole's spin while the outer disk stays tilted, producing a warp — the Bardeen–Petterson effect (1975) — that can torque and align relativistic jets.
• Quasi-periodic oscillations (QPOs). The Relativistic Precession Model attributes low-frequency QPOs in X-ray binaries to Lense–Thirring precession of a tilted inner flow. Observed low-frequency QPOs of ~0.1–30 Hz in systems like GRO J1655−40 and GRS 1915+105 are read as frame-dragging frequencies, turning the QPO into a probe of black-hole spin.
• Pulsar timing. In 2020, monitoring of the binary pulsar PSR J1141−6545 revealed a long-term drift in the orbit attributed to frame dragging by its rapidly spinning white-dwarf companion — a rare frame-dragging measurement outside the Solar System.
• Energy extraction. Inside the ergosphere the mandatory co-rotation enables the Penrose process and, with magnetic fields, the Blandford–Znajek mechanism — the leading model for how spinning black holes power relativistic jets in quasars and radio galaxies.

Common misconceptions and subtleties

• "Frame dragging is the spacetime being whirled by the surface friction of the spinning body." No friction is involved. The drag is a vacuum property of the geometry sourced by the body's angular momentum J; it exists in the empty space around the mass and is exactly captured by the vacuum Kerr solution.
• "Frame dragging is the same as the geodetic effect." They are distinct. The geodetic effect occurs even around a non-spinning mass; frame dragging requires J ≠ 0. GP-B measured them separately precisely because they are different terms in the precession equation pointing in perpendicular directions.
• "You could shield against it or hover motionless inside the ergosphere." Inside the ergosphere co-rotation is mandatory — there is no rocket thrust that lets you hold still relative to the distant stars, because doing so would require superluminal motion against the drag.
• "Lense–Thirring precession and frame dragging are two different effects." They are the same physics under two names: frame dragging is the phenomenon (inertial frames dragged by spin); Lense–Thirring precession is the observable consequence (the slow wheeling of a gyroscope or orbit). Lense and Thirring are simply the names attached to the 1918 derivation.
• "The effect falls off like gravity, ∝ 1/r²." It does not. Frame dragging is a dipole-like field that falls off as 1/r³, faster than Newtonian gravity, which is why it is comparatively even tinier far from the source and grows dramatically only close in.