YORP Effect

Sunlight torques an irregular rock

Every rock in the inner solar system is bathed in solar photons. Most discussions of solar radiation stop at radiation pressure — the small but persistent outward push that famously shapes comet tails and dust streamers. But absorbed light does not stay light; it warms the rock and is re-radiated, on average a fraction of an hour later, as thermal infrared photons. That re-emission carries away momentum too, and if the body is irregularly shaped or rotates with any obliquity, the outgoing photons do not balance.

The non-balance can produce two distinct second-order effects. If the lobe of preferentially emitted photons is offset from the asteroid's centre of mass along the orbit direction, you get a net translational force — the Yarkovsky effect — that slowly drifts the orbit's semi-major axis. If the emission asymmetry is about the spin axis instead — wedges, jagged equators, anything that breaks symmetry of the rotating profile — you get a net torque. That torque is the YORP effect. Yarkovsky changes where an asteroid goes; YORP changes how fast it spins and which way the pole points.

Naming and history

The acronym packs four contributors, none of whom worked on the same paper. Ivan Yarkovsky was a Russian-Polish civil engineer who in 1901 proposed the orbital-drift effect that bears his name (the YORP namesake by virtue of the close mechanism). John O'Keefe, a Goddard astronomer, sketched the rotational version informally in the 1970s. Vsevolod Radzievskii published Soviet work in the 1950s on spin-axis effects of solar radiation. Stephen Paddack worked out the modern thermal-recoil torque equations for irregular shapes in 1969 as a NASA technical report. David Rubincam consolidated and named the effect "YORP" in 2000, after which it rapidly became the dominant explanation for many small-body puzzles.

The first direct empirical confirmation came in 2007: Stephen Lowry's team and Patrick Taylor's team independently measured the spin acceleration of asteroid (54509) 2000 PH5 — now officially renamed (54509) YORP — by combining radar ranging with optical lightcurves over four years. The body spins up by 2.0 × 10⁻⁴ degrees per day², in agreement with thermal-recoil models from its measured shape.

The physics: a torque from photons

The instantaneous force on a surface facet of area dA, normal n, temperature T, and emissivity ε is

dF = − (2/3) (ε σ T⁴ / c) n dA

where σ is Stefan-Boltzmann, c is the speed of light, and the 2/3 factor comes from Lambertian re-emission integrated over the hemisphere. Integrate this over the body and you get the net force (Yarkovsky); take the moment about the centre of mass and you get the torque (YORP):

τ_YORP = ∫ (r × dF)

For a spherical asteroid the integral vanishes by symmetry. For any irregular shape it generally does not. The component of τ_YORP along the spin axis changes the spin rate; the perpendicular components reorient the pole. Both happen simultaneously, and both matter.

The strength of the effect scales steeply with size and distance:

dω/dt  ∝  F☉ / (ρ R² a²)         (Rubincam 2000)
        ∝  R⁻²   at fixed heliocentric distance a

where F☉ is the solar constant, ρ is bulk density, R is the body's radius, and a its semi-major axis. The R⁻² scaling is decisive: a 1 km body and a 10 km body at the same distance feel YORP angular accelerations that differ by a factor of 100. This is why YORP matters for small near-Earth asteroids but is negligible for anything above ~10 km.

Spin-up or spin-down — and pole reorientation

The sign of dω/dt is set entirely by the body's shape. There is no preferred direction; theoretical and numerical surveys of irregular shapes find roughly equal numbers of spin-up and spin-down cases. (Statzer and Vokrouhlický 2003 showed that for Gaussian-random asteroid shapes the median |dω/dt| is set by the size, but the sign is essentially random.) Observational confirmation comes from the spin-rate distribution of small asteroids, which is bimodal — bodies near the rotation-disruption limit and bodies near zero — exactly as one expects from a YORP-driven random walk in spin space with an absorbing boundary at the cohesion limit and a sticky boundary near zero.

The same torque also wanders the spin pole. Vokrouhlický and Čapek (2002) showed that YORP drives the obliquity toward attractor states: in many shape families, the pole evolves to either 0° or 180° (spin axis perpendicular to the orbital plane) on timescales similar to the spin-rate-change timescale. This explains why the observed obliquities of small main-belt asteroids cluster at the poles rather than being uniformly distributed.

The 2.2-hour cohesion limit

Set the centrifugal acceleration at the equator equal to the surface gravity of a uniform sphere of density ρ:

Ω² R  =  G M / R²  =  (4π/3) G ρ R
→  Ω_crit  =  √[(4π/3) G ρ]
→  P_crit  =  2π / Ω_crit  ≈  3.3 / √ρ   hours    (ρ in g/cm³)

For a typical asteroid density of 2 g/cm³, P_crit ≈ 2.33 hours. The observational distribution of asteroid rotation periods has an extremely sharp wall at 2.2-2.3 hours for objects larger than about 200 metres — virtually nothing rotates faster. Below 200 m the wall vanishes: dozens of "monolithic" small asteroids spin in tens of minutes or even minutes, held together by their own internal strength rather than gravity alone (Pravec et al. 2002; Hatch and Wiegert 2015). The 2.2-hour barrier is therefore the gravity-only stability boundary; YORP drives bodies into it, and what happens at the wall depends on whether the body is cohesive or a true rubble pile.

Rotational fission and binary formation

For a cohesionless rubble pile, hitting the 2.2-hour barrier means equatorial material can no longer be held. The body sheds mass. Three outcomes are theoretically distinguished — and all three appear in the observed population.

Fission outcome	Primary spin	Mass shed	Observed type
Stable mass-shedding	Just above critical	~ a few %	Equatorial ridge / top shape (Bennu, Ryugu)
Satellite formation	Spun down by recoil	~ 1–10 % bound	Binary asteroid (1996 FG3, 1999 KW4)
Pair formation	Marginally bound system disrupts	~ 1–30 % escapes	Asteroid pair (3749 Balam, 6070 Rheinland)

The full sequence as currently understood: YORP spins a rubble pile up; the body undergoes deformation as equatorial regolith creeps toward the equator; a top shape develops; further spin-up causes mass to lift off; some of it re-accretes as a bound satellite. The satellite carries away angular momentum, dropping the primary's spin below critical. If the binary remains bound, you observe it today. If the binary's mutual orbit is unstable and dissolves, the two pieces drift apart at differential rates set by their masses and the YORP/Yarkovsky drift on each — and you observe an asteroid pair.

Walsh, Richardson and Michel (2008) ran N-body simulations of this entire process for cohesionless rubble piles and reproduced the top shape and the typical binary mass ratios (~5 percent). Their simulations are widely treated as the standard reference for the YORP-fission mechanism.

Worked example: Bennu and Ryugu

Both targets of recent sample-return missions exhibit the diagnostic top shape with a sharp equatorial ridge — and both have measurable YORP signatures.

Bennu (target of OSIRIS-REx, sample returned 2023): diameter ~ 490 m, density 1.19 g/cm³, rotation period 4.296 hours. OSIRIS-REx tracking showed the spin is currently accelerating: dP/dt ≈ −1.0 s/yr, corresponding to dω/dt ≈ +6 × 10⁻⁶ deg/day², consistent with thermal-recoil modelling. The cohesion limit at ρ = 1.19 g/cm³ is P_crit ≈ 3.0 hours, so Bennu is at ~70 percent of the disruption rate — and accelerating. Its top shape is the signature of YORP-driven equatorial migration earlier in its history. OSIRIS-REx also confirmed pebble-mass ejection events: tens-of-gram particles regularly lift off the equator and either escape, re-impact, or briefly orbit. This is direct in-situ confirmation of rotational mass shedding.

Ryugu (target of Hayabusa2, sample returned 2020): diameter ~ 900 m, density 1.19 g/cm³, rotation period 7.63 hours. Ryugu also displays the equatorial ridge but rotates more slowly than Bennu, possibly because YORP has spun it down since the equatorial ridge formed. Hayabusa2's high-resolution shape model showed that the ridge is composed of looser regolith than the polar regions — consistent with relatively recent (within Myr) equatorial slumping rather than primordial mass.

Itokawa (Hayabusa, 2005): the original "rubber duck" asteroid, with a barely-bound head-and-body structure that may itself be a contact binary produced by an earlier YORP-fission event. Spin measurement showed dω/dt consistent with YORP within the considerable uncertainties.

Asteroid pairs: (3749) Balam and others

Pravec et al. (2010) identified the first asteroid pairs: distinct main-belt asteroids on nearly identical orbits whose backward integration converges to a single source point and time. The list now includes dozens of pairs, with ages from 10⁴ to 10⁶ years — far younger than the parents themselves and inconsistent with a collision origin (which would be far too rare). The straightforward explanation is YORP-driven rotational fission: a parent spins up, sheds a fragment that fails to remain bound, and the two bodies drift apart.

(3749) Balam is the textbook case: it has both a bound satellite (~ 90 m) and a co-orbital unbound companion, plus a confirmed pair partner at ~ 1 km. All four bodies appear consistent with sequential YORP-driven fissions from a single rotating progenitor. Pair ages and the primary's measured fast rotation match the YORP-cascade hypothesis.

Population-level evidence

• Binary NEO fraction. Pravec and Harris (2007) reported that ~16 percent of near-Earth asteroids larger than 300 m are binaries. The mass ratios cluster near 5 percent and the primaries spin near the critical limit, both consistent with YORP-spin-up fission. Main-belt binaries are rarer (~ 2 percent) because the main belt is colder, YORP is slower, and the binary lifetimes against further perturbations are shorter.
• Spin distribution. The cumulative rotation-period distribution of small asteroids shows the predicted bimodality: pile-up at the 2.2-hour limit, depletion just below it (fast bodies promptly fission and are removed), and a long tail at slow rotation. The shape matches a YORP random-walk model with absorbing-at-fast and reflecting-at-slow boundaries.
• Top-shaped body abundance. Among small NEOs visited by radar, ~ 30 percent have the diagnostic equatorial ridge / spinning-top profile (1999 KW4, Didymos, 2008 EV5, Bennu, Ryugu). The fraction is too high for a single-event origin and is naturally explained by long-term YORP-driven mass migration.
• Obliquity clustering. Small main-belt asteroids show preferential pole obliquities at 0° and 180° — predicted by YORP attractor dynamics, not by primordial randomness.
• Asteroid pairs. > 100 confirmed pairs in the main belt, with backwards-integrated ages from 10⁴ to 10⁶ years and primary rotation rates near the critical limit. Each pair represents one fission event in the YORP framework.

Variants and complications

• Normal YORP vs tangential YORP. "Normal YORP" comes from emission asymmetries normal to the surface; "tangential YORP" (Golubov & Krugly 2012) is a smaller contribution from anisotropic conduction in surface boulders. Tangential YORP is harder to model but may dominate the steady-state spin-up for some bodies.
• YORP self-stochasticity. A subtle issue: as the asteroid loses material or shifts regolith, its shape — and therefore its YORP coefficients — change. Statler (2009) argued that YORP is so shape-sensitive that the long-term behaviour is effectively a random walk in spin and obliquity even at fixed total composition. This may smear out clean predictions of pole attractors and complicates pair-age dating.
• BYORP. Binary YORP — the analogous radiation torque on a tidally locked secondary in a binary system — was identified by Ćuk and Burns (2005). BYORP can drive a binary asteroid's mutual orbit to either expand (separation) or contract (re-merger or chaotic disruption) on Myr timescales, and is the leading mechanism for binary lifetimes against fission-driven creation.
• Thermal conductivity and the lag. The thermal-emission lag — and therefore the YORP torque amplitude — depends on the surface thermal inertia. Very low conductivity (fluffy regolith) means rapid re-equilibration and small lag; rocky surfaces have higher inertia and larger lag. This is why YORP rates from theoretical models always come bracketed by ±a factor of a few.

Where YORP shows up in mission data

• (54509) YORP (formerly 2000 PH5). First and most precisely measured spin acceleration: dω/dt = +(2.0 ± 0.2) × 10⁻⁴ deg/day². 113 m equivalent diameter; spin period decreasing from 12.17 min by ~ 1 ms/yr. The textbook YORP detection (Lowry et al. 2007; Taylor et al. 2007).
• Itokawa. Hayabusa imagery (2005) plus archival lightcurves yielded dω/dt consistent with the YORP prediction at the 1-2 σ level. Itokawa's contact-binary shape may be the relict of a past YORP-driven fission event.
• Bennu. OSIRIS-REx tracking refined the spin acceleration to dP/dt ≈ −1.0 s/yr — a robust 5-σ detection (Nolan et al. 2019; Hergenrother et al. 2019). Mass-shedding events directly observed.
• Ryugu. Hayabusa2 imaging refined the shape model; the top-shape and surface regolith distribution are interpreted as YORP-driven equatorial migration.
• 1999 KW4 (Moshup) binary. Diameter 1.3 km primary with 0.36 km secondary, primary spin period 2.76 hours (close to critical), top-shaped primary with an equatorial bulge — the archetypal YORP-fission binary.
• (65803) Didymos / Dimorphos (DART target). Top-shaped primary, secondary in nearly tidally locked orbit, spin near critical — fits the YORP-fission picture. DART's deliberate impact on Dimorphos in 2022 was set against this baseline of YORP-driven binary architecture.

Common pitfalls

• Conflating YORP with Yarkovsky. They share a name and a mechanism but act on different motions. YORP = torque, changes spin and pole. Yarkovsky = force, changes orbit. Both occur simultaneously on an irregular body; do not use one's timescale to argue about the other.
• Treating YORP as deterministic. Statler-style stochastic YORP shows that the effect is so shape-sensitive that small surface changes can flip dω/dt sign. Population-level predictions are robust; individual-body forecasts over > 10⁵ years are not.
• Assuming all top-shaped asteroids spin near critical. A body can develop a top shape during a spin-up phase and then spin down again by YORP; today's rotation period is not always diagnostic of the formation epoch.
• Forgetting BYORP for binaries. Once a binary forms, its mutual orbit evolves under BYORP, not YORP. Models that ignore BYORP overestimate binary lifetimes and undercount the rate at which new binaries replace dissolved ones.
• Ignoring the cohesion floor. Bodies smaller than ~ 200 m often have internal strength — they are not pure rubble piles. The 2.2-hour limit does not apply, and YORP can spin them to minutes-period rates without disrupting them. The disruption mechanism for small monoliths is fragmentation under thermal stress, not centrifugal mass shedding.