Shoemaker-Levy 9

The discovery — a smudged comet on a Palomar plate

On the night of 1993 March 23–24, Carolyn Shoemaker, her husband Eugene Shoemaker, and their long-time collaborator David Levy were running their Palomar Planet-Crossing Asteroid Survey on the 0.46-m Schmidt telescope at Palomar Observatory. Conditions were poor; a stray photograph that should have been discarded was developed anyway because the film was already used. On the plate, near Jupiter, sat an elongated smudge — at first sight a bar-shaped comet, the kind nobody had ever cataloged. Carolyn Shoemaker described it as looking like a "squashed comet."

Follow-up imaging by Jim Scotti at Spacewatch the next night resolved the smudge into a row of discrete cometary nuclei strung along a line, each with its own coma and dust tail. The object was reported to the Central Bureau for Astronomical Telegrams as the ninth periodic comet co-discovered by the Shoemaker-Levy team — hence the name Shoemaker-Levy 9, with the formal designation D/1993 F2 (the D prefix indicating a "destroyed" or "disappeared" comet). Within days, orbital fits demonstrated something even more remarkable than the comet's appearance: it was orbiting Jupiter, not the Sun.

Captured by a giant — the previous twenty-five years

SL9 was the first comet humans had ever caught in orbit around a planet. Numerical integration of the discovery orbit backwards in time, by Don Yeomans and Paul Chodas at JPL, placed the original Jupiter capture in the window 1966–1970. Before that, SL9 was probably a member of the Jupiter-family comet population — short-period comets with aphelia near Jupiter's orbit and origins in the Kuiper belt and scattered disc.

Capture into a bound orbit around a single planet, without a third-body assist, requires the comet to lose orbital energy with respect to the planet. The most plausible mechanism is a sequence of close approaches: the first encounter dropped SL9 onto a marginally-bound Jupiter orbit; subsequent approaches over the next few decades dissipated additional energy via the three-body Sun-Jupiter-comet dynamics. By the early 1990s SL9 was looping Jupiter on a highly elliptical orbit with a period of roughly 2 years, with apojove out beyond the orbit of Callisto.

The 1992 disruption — inside the Roche limit

The pivotal close approach occurred on 1992 July 7, when SL9 dipped to a perijove of approximately 1.3 Jovian radii — about 30,000 km above Jupiter's cloudtops, well inside the Roche limit for a porous, self-gravitating body. The Roche limit is the distance inside which the tidal stretch across a small body exceeds its own self-gravity. For a fluid Roche limit it is

d ≈ 2.44 R_planet (ρ_planet / ρ_body)^(1/3)

For Jupiter (ρ ≈ 1.3 g/cm³) and a low-density rubble-pile cometary nucleus (ρ ≈ 0.5–0.6 g/cm³) the limit lies near 1.7 R_J. SL9 passed at roughly 1.3 R_J, deeply inside the limit. The differential acceleration across the ~2 km nucleus exceeded its self-gravity. With no significant cohesive strength to hold it together — comet nuclei being weakly-bound aggregates of ice and dust — the nucleus simply pulled apart along its long axis into 21 discrete fragments.

Each fragment was given a letter, A through W (with I and O skipped to avoid confusion with the numerals 1 and 0). Initially the fragments were close together, but each emerged from the encounter with a slightly different orbital energy and so drifted apart along the shared orbit at different rates. By the time of the 1993 discovery, the 21 nuclei stretched in a line roughly 1.1 million kilometres long — a "string of pearls" each trailing its own dust tail.

The collision prediction — and the question of survival

Once SL9's bound-Jupiter orbit was established in March 1993, orbital integration forward in time delivered an unprecedented prediction: the next perijove, in July 1994, would not be a flyby but an impact. Every one of the 21 fragments would strike Jupiter within a six-day window, hitting the planet just over the limb from Earth's point of view but rotating into Earth-visible longitude within minutes of each strike.

What the impacts would actually look like was furiously debated through 1993–1994. Some models predicted Jupiter's deep atmosphere would simply absorb the energy with no visible signature — a "fizzle." Others predicted titanic plumes, sky-glow auroras, even a brief darkening of the planet visible to the naked eye. Coordinated observation campaigns were organised on essentially every operating telescope on Earth and several in space, and an aircraft-based observatory was scrambled to view from the southern hemisphere. The Galileo spacecraft, then en route to Jupiter, was repurposed to look directly at the impact face — the only platform with that geometry.

July 16 – 22, 1994 — the impacts

Fragment A struck Jupiter at 20:13 UT on 1994 July 16, at southern latitude near −44°. Within minutes a bright fireball was visible in IR through Earth-based telescopes as the impact site rotated over the limb. The expectation of a "fizzle" was demolished. Over the next six days, the remaining 20 fragments arrived in sequence, each striking near the same latitude band, leaving 21 distinct dark scars in the southern hemisphere.

The impact mechanics, reconstructed from Galileo direct imaging and HST follow-up, ran in three stages:

1. Entry — 0 to 2 seconds. Each fragment punched a kilometres-wide hole through Jupiter's upper atmosphere at 60 km/s, vapourising and shock-heating gas along the entry channel.
2. Fireball and plume — 2 seconds to ~10 minutes. The deposited energy expanded back through the entry channel as a hot, optically-thick plume. The plume rose ballistically to altitudes of ~3000 km above the cloudtops, plainly visible above the limb. Galileo imaging gave plume temperatures of 30,000–40,000 K at the brightest point.
3. Splash-back and scar formation — 10 minutes to ~1 hour. Plume material fell back onto the stratosphere at high speed, re-compressing and synthesising dark sulfur- and carbon-bearing aerosols that spread laterally on stratospheric winds. Each scar grew over an hour to roughly 6,000–12,000 km across — larger than Earth's diameter for the bigger impacts.

The Galileo SSI imaging team captured the brightest fireballs directly. HST's WFPC2 imaged the rising plumes from above the limb. Ground-based IR observatories (IRTF, ESO 2.2 m, AAT, Keck) tracked thermal emission from the plumes and scars. Detection of impact-generated water, ammonia, sulfur, and a host of organic species in Jupiter's stratosphere — by ground-based IR and HST UV — provided unprecedented insight into the composition of both cometary and Jovian material.

Impact energies — the headline numbers

Each fragment's diameter was inferred from its impact luminosity and from pre-impact HST imaging. Reconstructed values for the largest fragments cluster at 0.5–2 km. The impact speed v at the cloudtops is the planet's escape speed:

v_esc = √(2 GM_J / R_J) = 60 km/s

(Comets falling from interplanetary space arrive with at least this much, and any orbital velocity adds in quadrature — SL9 had been slow-moving relative to Jupiter, so v_esc is close to the actual impact speed.) The specific kinetic energy is therefore

KE / mass = (1/2) v² = (1/2)(6 × 10⁴ m/s)² ≈ 1.8 × 10⁹ J/kg

Adopting a cometary density of ~0.6 g/cm³ and the inferred fragment sizes yields the table below. Energies are quoted in joules and in equivalent megatons TNT (1 Mt = 4.184 × 10¹⁵ J).

Fragment	Est. diameter	Impact UT	Energy (J)	Equivalent
A	~0.6 km	Jul 16 20:13	~1 × 10²⁴	~240 Mt
G (largest)	~2 km	Jul 18 07:33	~6 × 10²⁵	~6 Gt · ~600× nuclear arsenal
H	~1.5 km	Jul 18 19:32	~2 × 10²⁵	~5 Gt
K	~1.0 km	Jul 19 10:24	~3 × 10²⁴	~700 Mt
L	~1.0 km	Jul 19 22:16	~5 × 10²⁴	~1.2 Gt
Q1	~1.2 km	Jul 20 20:11	~1 × 10²⁵	~2.4 Gt
W (last)	~0.3 km	Jul 22 08:06	~5 × 10²³	~120 Mt
21 fragments total	—	Jul 16 – 22	~3 × 10²⁶ – 1 × 10²⁷	~70 – 250 Gt

For perspective: the total deployed nuclear arsenal of all nations at the time held about 10 Gt of yield. Fragment G alone exceeded that total by ~600×. The 1908 Tunguska airburst released ~10–15 Mt — flattening 2,000 km² of Siberian forest. The 2013 Chelyabinsk airburst delivered ~500 kT, shattering windows across a city. SL9-G, scaled to Earth, would have ended industrial civilisation.

The scars — Earth-sized bruises on Jupiter

The dark scars left by each impact were the most photogenic part of the event and the longest-lasting record. The largest, from fragment G, was 12,000 km across — comfortably wider than Earth's 12,742 km diameter. The scars were visible through a backyard 4-inch telescope and dominated Jupiter's appearance for the next month.

The scars' composition was inferred from HST UV, ground-based mid-IR, and millimetre-wave spectroscopy. The dark material is dominated by sulfur-bearing molecules (S₂, CS₂, H₂S, OCS) and complex hydrocarbons, synthesised in the high-temperature plume re-entry and condensed at stratospheric temperatures. Carbon monosulfide (CS) — a transient radical — was detected at unprecedented abundance. The scars slowly faded over the following year as Jovian stratospheric winds dispersed and chemically converted the aerosols. Some sulfur enhancement remained detectable in Jupiter's stratosphere for nearly a decade.

Jupiter as the inner Solar System's vacuum cleaner

SL9 dramatised a long-standing dynamical claim: that Jupiter, by virtue of its mass and position, gravitationally shepherds away or accretes a significant fraction of the comets and asteroids that might otherwise reach the inner planets. George Wetherill's 1994 calculations (published essentially simultaneously with the impacts) estimated that Jupiter intercepts incoming long-period comets at roughly 1,000× the rate of Earth and reduces Earth's impact flux by about an order of magnitude compared with a Jupiter-less Solar System.

This "cosmic vacuum cleaner" framing is now nuanced — Jupiter also occasionally delivers comets to Earth-crossing orbits via gravitational scattering, so the net effect on Earth's impact rate is debated and likely smaller than Wetherill's first estimate. But SL9 remains the canonical demonstration that Jupiter does in fact catch comets, and that being a giant planet a few AU outside the habitable zone is, on net, helpful for the inner planets' long-term peace.

The planetary-defence wake-up call

Before 1994, the cosmic-impact threat lived in cult-classic films and Walter Alvarez's K-Pg iridium layer. After 1994, it lived on the front page. Within a year, the U.S. Congress directed NASA to identify 90% of near-Earth objects larger than 1 km diameter — the "Spaceguard goal." A cascade of dedicated surveys followed:

• LINEAR (Lincoln Near-Earth Asteroid Research) — operational 1996, MIT Lincoln Laboratory; first dedicated wide-field asteroid survey using military surveillance optics.
• Catalina Sky Survey — 1998 onward; University of Arizona; still one of the most productive NEO discoverers globally.
• NEAT (Near-Earth Asteroid Tracking) — 1995, JPL; pioneered automated detection software now standard across the field.
• Pan-STARRS — 2008 onward; Maui-based 1.8 m survey telescopes, currently a leading source of new NEO discoveries.
• ATLAS (Asteroid Terrestrial-impact Last Alert System) — 2015; rapid-cadence wide-field survey optimised to give days-to-weeks warning of small (~20 m) Earth impactors.
• NEOWISE — 2014 onward; the WISE infrared satellite repurposed for asteroid hunting; key for measuring asteroid sizes via thermal emission.
• NEO Surveyor — launching 2027; a dedicated infrared NEO-discovery satellite that should complete the 140 m-and-larger NEO catalog.

The 2013 Chelyabinsk airburst — a 20 m asteroid, undetected before entry, injuring ~1,500 people in Russia — sharpened priorities further, motivating the small-body component of ATLAS and NEO Surveyor's design. The 2022 DART mission, which successfully altered the orbit of asteroid moon Dimorphos by kinetic impact, demonstrated the first operational asteroid-deflection technique. Every one of these programmes traces a direct lineage to the SL9 wake-up call.

Scientific legacy

• First observed cometary tidal disruption. Before SL9, tidal disruption of comets was a theoretical hypothesis to explain split-comet observations. SL9 turned theory into a watched event.
• First direct probe of Jupiter's deep atmosphere chemistry. The impacts dredged up material from beneath the visible cloud deck, producing stratospheric ammonia, water, and sulfur signatures that constrained Jupiter's interior composition.
• First real-time observation of an impact's plume physics. Galileo and HST coverage gave plume heights, temperatures and timing that calibrate every subsequent impact simulation — directly relevant to dinosaur-killer K-Pg modelling.
• Trigger for the modern NEO survey programme. The post-1994 cadence of NEO discovery exploded: pre-1994 NEO discovery rate was ~10/year; today it exceeds 3,000/year.
• Confirmation of the rubble-pile cometary structure. SL9's clean tidal disruption is consistent with — and supports — the picture of cometary nuclei as low-cohesion aggregates rather than monolithic ice blocks. The 2014 Rosetta mission to 67P/Churyumov-Gerasimenko later confirmed this directly.

Comparing SL9 to other historical impacts

Event	Year	Object size	Energy	Notes
Chicxulub (K-Pg)	66 Mya	~10 km asteroid	~10²³ J · 10⁸ Mt	Earth — ended non-avian dinosaurs
SL9 fragment G	1994	~2 km comet fragment	~6 × 10²⁵ J · 14,000 Mt	Jupiter — Earth-sized scar
SL9 total	1994	21 fragments	~10²⁶ – 10²⁷ J	Jupiter — six-day cascade
Tunguska	1908	~60 m asteroid/comet	~5 × 10¹⁶ J · 10–15 Mt	Earth — Siberian forest flattened
Chelyabinsk	2013	~20 m asteroid	~2 × 10¹⁵ J · 500 kT	Earth — ~1,500 injured
2024 Jupiter impact	2024 Aug	~10 m bolide	~10¹⁴ J	Jupiter — amateur-detected flash

Note that SL9-G's single-fragment energy was roughly half the impact that ended the dinosaurs. The fact that we watched it through binoculars from Earth — peacefully, on a different world — is the entire point of the story.

Common pitfalls

• Confusing capture date with discovery date. SL9 was discovered in March 1993 but had been orbiting Jupiter since roughly 1966–1970. Its observation as a planet-orbiting comet started in 1993, but its existence as one is roughly thirty years older.
• Treating "comet" as monolithic. SL9's clean tidal break-up is direct evidence that its nucleus was a weakly-bound aggregate, not a solid ice ball. Modern cometary nuclei are best modelled as rubble piles with very low tensile strength (~10 Pa).
• Misquoting the energy comparison. Fragment G released ~6 × 10²⁵ J. The world's deployed nuclear arsenal in 1994 was roughly 10²² J — 10 Gt TNT. The comparison is "~600× the global nuclear arsenal," not "~600,000×."
• Assuming Jupiter "protects" Earth simply. Wetherill's vacuum-cleaner argument is correct on long-period comets, but Jupiter also scatters Kuiper-belt and main-belt objects into Earth-crossing orbits. The net effect on Earth's impact flux is a subject of active modelling, not a settled order-of-magnitude reduction.
• Forgetting the impacts hit the limb, not the visible disk. Each fragment struck just over Jupiter's far limb from Earth — direct fireballs were only visible to Galileo from its in-cruise vantage. From Earth, the scars rotated into view minutes after each impact.