What Is Exoplanet Direct Imaging? Seeing Planets Around Other Stars

Finding exoplanets indirectly — by watching a star dim slightly when a planet transits in front of it, or by measuring tiny radial velocity wobbles in a star's spectrum — is difficult enough. But astronomers dream of something more profound: actually seeing the planet itself. Resolving the photons from an alien world, studying its reflected starlight or its own thermal glow, reading the fingerprints of its atmosphere directly.

Direct imaging of exoplanets is arguably the most technically demanding observational challenge in astronomy. The obstacles are formidable: a planet is 10⁶–10¹⁰ times fainter than its host star (depending on wavelength and separation), yet separated from it by an angle that may be less than 1 arcsecond — smaller than the atmospheric seeing disk. Seeing the planet beside the star is like seeing a firefly next to a stadium floodlight from 100 km away.

Yet direct imaging has been achieved. Dozens of young, massive gas giant planets have been directly resolved orbiting young stars. Spectroscopy of these directly imaged planets reveals their temperatures, atmospheric compositions, and rotation rates. And the planned Habitable Worlds Observatory targets something even more ambitious: directly imaging Earth-like planets in the habitable zones of nearby stars.

Why direct imaging is so hard

The fundamental challenge is the contrast ratio between the star and the planet:

In the near-infrared (thermal emission):

• Young (~10 Myr) gas giants (~5 M_J) are relatively bright — they are still warm from formation, radiating thermally
• Contrast: ~10⁶ (star 1 million times brighter than planet) at 1–2 μm
• Favourable for ground-based adaptive optics imaging

In the optical (reflected starlight):

• An Earth at 1 AU from a Sun-like star reflects ~10⁻¹⁰ of the stellar flux
• Contrast: ~10⁹–10¹⁰ (star 10 billion times brighter than planet) at visible wavelengths
• Extremely challenging; requires space-based coronagraph

Angular separation: A planet at 1 AU orbiting a star 10 pc away subtends only 0.1 arcsecond — at the limit of what ground-based AO systems can resolve (and only possible with extreme adaptive optics). A Jupiter analogue at 5 AU from the same star would be 0.5 arcseconds — more accessible.

Key technologies

Coronagraphs

A coronagraph is an optical device that blocks the central star's light while allowing the planet's light to pass through. It was invented by Bernard Lyot in 1931 for studying the solar corona — and adapted for exoplanet imaging.

Types of coronagraphs:

• Lyot coronagraph: A focal plane mask blocks the star; a Lyot stop in the pupil plane removes diffracted light around the mask
• Band-limited coronagraph: Mask with transmission graded to suppress diffraction
• Phase-induced amplitude apodisation (PIAA): Reshapes the beam to reduce diffraction
• Vortex coronagraph: A spiral phase plate that diffracts stellar light away from the focal plane to the pupil edge, where it is blocked
• Shaped pupil coronagraph: The pupil is shaped (by a mask) to control the diffraction pattern, creating dark "discovery zones" away from the star

Modern coronagraphs achieve contrasts of ~10⁻⁸ to 10⁻⁹ in the lab (at small angular separations). Space-based coronagraphs without atmospheric distortion can achieve deeper suppression for longer — enabling reflected-light imaging of nearby super-Earths with the Roman Space Telescope and eventually the HWO.

Extreme adaptive optics (ExAO)

Ground-based direct imaging requires extreme AO (ExAO) — adaptive optics with much higher order correction than standard AO. Where standard AO uses ~250 actuators, ExAO systems use 2,000–10,000+ actuators, correcting smaller-scale atmospheric distortions and achieving higher Strehl ratios (80-90% in H-band for the best systems).

Leading ExAO systems:

• GPI (Gemini Planet Imager): 1,500 actuator DM, operating at Gemini South (Chile)
• SPHERE (Spectro-Polarimetric High-contrast Exoplanet Research): VLT instrument, 41×41 actuator DM, H-band Strehl >90%
• SCExAO (Subaru Coronagraphic Extreme AO): Subaru Telescope, 2,000 actuator DM, advanced coronagraph designs
• MagAO-X: Magellan Clay telescope, 2,000 actuator DM

Angular Differential Imaging (ADI) and post-processing

Even with ExAO + coronagraph, residual speckles from imperfect atmospheric correction and instrument optical imperfections can mimic planet signals. The key technique to remove these:

ADI (Angular Differential Imaging): The telescope pupil is not rotated to track field rotation as Earth turns. Background speckles (which are instrument-fixed) remain stationary, while real astrophysical sources (including planets) rotate with the field. By subtracting a median frame from each image, speckles are removed while planets leave a curved trail — which is then de-rotated and combined.

PCA (Principal Component Analysis) / KLIP: Advanced post-processing that models the speckle pattern using a library of reference frames and subtracts it. Combined with ADI, KLIP achieves close to the theoretical photon-noise limit.

Directly imaged planets: the gallery

Beta Pictoris b

Beta Pictoris — a young A-type star (~23 Myr old) at 19 pc — shows a prominent edge-on debris disk discovered by IRAS in 1983. The inner disk showed a warp — indirect evidence for a massive planet.

In 2008, Beta Pictoris b was directly imaged by Anne-Marie Lagrange and colleagues using NAOS-CONICA (NaCo) on the VLT — a ~13 M_J gas giant at ~8-9 AU, consistent with causing the disk warp. The planet transited its star in 2017, briefly dimming it — allowing direct measurement of the planet's size. Spectroscopy with GRAVITY (VLTI) measured the planet's Doppler broadening from rotation: ~25 km/s equatorial velocity (similar to Jupiter's 12 km/s for a larger planet).

Beta Pictoris c: A second planet was detected in radial velocity data in 2019 and later confirmed by GRAVITY: ~8 M_J at ~2.7 AU. The two-planet system provides a natural laboratory for comparing the spectra of planets at different separations and evolutionary states.

HR 8799: the first multi-planet direct imaging system

HR 8799 — a young (~30 Myr) F-type star at 39 pc — is the most celebrated direct imaging system: four giant planets (b, c, d, e) resolved simultaneously.

• HR 8799 b (2008, Marois et al.): ~6 M_J, 68 AU separation — detectable because it is young and warm
• HR 8799 c and d (2008, same paper): ~7 and 7 M_J at 38 and 24 AU
• HR 8799 e (2010): ~7 M_J at 15 AU — closest, requires exceptional AO

The system's orbital periods (40-400 years) are long but the planets have been imaged over years — arc motions are clearly visible. The planets appear to be in a 1:2:4:8 mean motion resonance — constraining their formation history (likely formed in the disk and migrated into resonance).

Spectroscopy: GPI and SPHERE have obtained near-infrared spectra of HR 8799 b, c, d, e. All show non-equilibrium chemistry (CO present where CH₄ expected — vertical mixing), patchy cloud coverage, and temperatures of 900-1,200 K.

51 Eridani b

51 Eri b (2015, GPI): A ~2 M_J gas giant at 13 AU from a 20-Myr star at 29 pc — notable for being the most methane-rich directly imaged planet (young planets are warm and usually show no methane; 51 Eri b's methane detection implies a relatively lower temperature and possibly more massive planet with faster cooling). Its spectrum most closely resembles a T dwarf among all directly imaged planets.

AF Leporis b

AF Lep b (2023): A ~3 M_J planet at ~9 AU from a young (24 Myr) F star. Notable because its detection was aided by Hipparcos/Gaia proper motion anomaly (the star's proper motion differed between the two catalogues — implying a companion pulling on the star). The Hipparcos-Gaia astrometric acceleration is now a powerful tool for prioritising direct imaging targets.

Directly imaged planets around M dwarfs

VHS 1256 b: A low-mass star (M dwarf) binary system with a directly imaged ~16 M_J companion — on the border between planet and brown dwarf. JWST spectroscopy of VHS 1256 b obtained the most detailed atmosphere spectrum of any directly imaged companion — showing water, methane, CO, CO₂, silicate clouds, and potassium. A benchmark for atmosphere characterisation.

2M1207 b: The first directly imaged companion to a brown dwarf (itself ~25 M_J), at ~3-8 M_J and 46 AU separation. Imaged in 2004.

Spectroscopy of directly imaged planets

Direct imaging enables reflected or emitted light spectroscopy — without needing to transit the star. This provides:

• Temperature: Fit to thermal emission spectrum
• Surface gravity: From spectral pressure-broadening
• Atmospheric composition: Detection of H₂O, CH₄, CO, CO₂, NH₃, TiO/VO (for warmer objects)
• Cloud properties: Photometric variability and spectral slope
• Rotation: Doppler broadening of spectral lines (measured for Beta Pic b by GRAVITY)
• Radial velocity and astrometry: Position relative to star, orbital parameters

Molecular mapping: Cross-correlating a high-resolution spectrum with molecular templates can detect individual molecules even in noisy data. GRAVITY on VLTI has obtained high-resolution spectra of Beta Pic b and HR 8799 e by combining interferometric astrometry with spectroscopy.

The future: imaging Earth-like planets

Roman Space Telescope Coronagraph Instrument (CGI): The Nancy Grace Roman Space Telescope (launch ~2027) includes a technology demonstration coronagraph achieving contrast of ~10⁻⁸ at 0.15–0.45 arcseconds — potentially imaging super-Jupiters around nearby stars in reflected light for the first time. Not designed to image Earth analogues, but a technology demonstration.

ELT extreme AO + coronagraph: The ELT's EPICS/METIS instruments, with 40-m class collecting area and next-generation ExAO, will image young super-Earths and directly characterise Uranus/Neptune-sized planets around nearby stars in reflected light.

Habitable Worlds Observatory (HWO): The next major NASA flagship mission, recommended by the 2020 Decadal Survey — a ~6-metre space telescope with a high-performance coronagraph achieving contrast ~10⁻¹⁰ at angular separations of ~50-200 mas. This would enable:

• Direct detection of reflected light from Earth-like planets in the habitable zones of ~25 nearby Sun-like stars
• Spectroscopy for O₂, H₂O, CO₂, CH₄ — biosignature detection
• The first definitive search for life on exoplanets

HWO is the most ambitious astronomical goal of the coming decades — potentially answering, for the first time, whether Earth is unique or whether life is common. Launch target: ~2040s.

Related Reading

• What Are Exoplanets and Could Any of Them Host Life?
• How Do We Detect Exoplanet Atmospheres?
• What Is the Habitable Zone?
• What Are Brown Dwarfs?
• How Do Telescopes Work?