The Rayleigh Criterion

Why the Rayleigh criterion matters

Angular resolution is the single most important number that separates a good telescope from a great one. It is the reason astronomers keep building bigger mirrors and stringing telescopes across continents. The Rayleigh criterion is the physics that tells you, before you spend a billion dollars, exactly how fine a detail a given instrument can ever hope to distinguish — not because of poor optics or shaky mounts, but because light is a wave that spreads when it passes through any finite opening.

• It sets a hard ceiling. No amount of magnification or exposure time beats diffraction. A perfect 1 m telescope simply cannot split two stars closer than ~0.14 arcsec in visible light.
• It justifies giant apertures. Since θ ∝ 1/D, the 39 m Extremely Large Telescope will resolve detail roughly 16 times finer than Hubble's 2.4 m mirror.
• It explains interferometry. Swap the aperture D for a baseline B between telescopes and you can synthesize resolution far beyond any single dish — how the Event Horizon Telescope imaged a black hole's shadow.
• It defines the seeing problem. On the ground, atmospheric turbulence usually blurs images far below the diffraction limit — motivating adaptive optics and space telescopes.
• It governs exoplanet imaging. Directly seeing a planet next to its star is fundamentally a resolution-and-contrast problem set by 1.22 λ/D.
• It applies everywhere. The same rule limits microscopes, cameras, radar, and your own eye — it is one of the most universal statements in optics.

How it works, step by step

Light passing through a circular aperture does not travel in perfectly straight rays. By Huygens' principle, every point in the opening re-radiates, and the waves interfere. The result on the focal plane is not a point but a bright central blob surrounded by concentric rings — the Airy pattern, named for George Biddell Airy, who derived it in 1835.

1. A star becomes an Airy disk. Even a perfect optic images an infinitely distant point source as a disk of finite width. About 84% of the energy lands in the central Airy disk; roughly 7% goes into the first bright ring.
2. The first dark ring sets the scale. The first zero of the pattern lies at an angle θ = 1.22 λ/D from the center. This is the natural "radius" of a star image.
3. Two stars, two overlapping disks. Point the telescope at a close double star and you get two Airy patterns side by side. When they are far apart, you see two clean peaks. As they approach, the peaks slide together.
4. The just-resolvable limit. Rayleigh's rule: the pair is just resolvable when the central peak of one disk sits exactly on the first dark ring of the other — a separation of 1.22 λ/D. At that spacing the combined profile still shows a shallow central dip of about 26% below the two peaks, enough for the eye to register two sources.
5. Below the limit, they merge. Bring them any closer and the dip fills in; the two stars blur into a single elongated blob no instrument of that aperture can split.
6. Beat it with a bigger D or a shorter λ. A larger aperture shrinks every Airy disk; a shorter wavelength (blue, UV, X-ray) does the same. Both push θ down and let you separate tighter pairs.

The key equation and its variables

The Rayleigh criterion for a circular aperture is written:

θ = 1.22 · λ / D

Symbol	Meaning	Units
θ	Smallest resolvable angular separation	radians (× 206,265 for arcsec)
λ	Observing wavelength	metres (550 nm = 5.5 × 10⁻⁷ m for green light)
D	Aperture (mirror or lens) diameter	metres
1.22	Airy coefficient = 3.8317 / π (first Bessel zero)	dimensionless

The 1.22 is not arbitrary. A circular aperture's diffracted intensity follows the Airy function I(x) = I₀ [2 J₁(x)/x]², where J₁ is the first-order Bessel function of the first kind. Its first zero occurs at x = 3.8317 = π × 1.2197, and converting that argument into an on-sky angle divides out the π, leaving θ = 1.22 λ/D. A one-dimensional slit, by contrast, has no Bessel function and gives the simpler θ = λ/D — so the factor 1.22 is the fingerprint of a round hole specifically.

To convert radians to arcseconds, multiply by 206,265. For green light (λ = 550 nm) this collapses to a memorable rule of thumb:

θ (arcsec) ≈ 0.14 / D_metres

Worked example: splitting a binary star

Suppose you want to resolve the two components of a binary star separated by 0.10 arcsec in visible light (λ = 550 nm). What aperture do you need?

Rearranging θ = 1.22 λ/D for D, and using θ = 0.10 arcsec = 0.10 / 206,265 = 4.85 × 10⁻⁷ radians:

D = 1.22 λ / θ = 1.22 × (5.5 × 10⁻⁷ m) / (4.85 × 10⁻⁷ rad) ≈ 1.38 m

So a telescope of about 1.4 m aperture is diffraction-limited to ~0.10 arcsec — in principle. On the ground, atmospheric seeing of ~1 arcsec would blur the pair together regardless. This is why such a measurement requires either a space telescope, adaptive optics, or speckle interferometry to recover the diffraction limit. Below is the diffraction limit for a range of real instruments at 550 nm.

Instrument	Aperture D	Diffraction limit θ (550 nm)
Human eye (dark-adapted pupil)	~7 mm	~20 arcsec
Amateur refractor	0.1 m	~1.4 arcsec
Hubble Space Telescope	2.4 m	~0.05 arcsec
Keck / VLT unit telescope	~10 m	~0.013 arcsec
Extremely Large Telescope	39 m	~0.003 arcsec
VLTI (optical interferometer)	~130 m baseline	~1 milliarcsec
Event Horizon Telescope (1.3 mm radio)	~10,000 km baseline	~20 microarcsec

Note the EHT operates at radio wavelengths (1.3 mm), which is 2,000 times longer than green light — normally terrible for resolution — but its Earth-spanning baseline more than compensates, delivering the sharpest angular resolution humanity has ever achieved.

Seeing-limited versus diffraction-limited

The Rayleigh criterion describes the best case: what a telescope could do in a vacuum. From Earth's surface, that best case is almost never met in the optical, because turbulence in the atmosphere continually distorts the incoming wavefront.

• Diffraction-limited — resolution set only by θ = 1.22 λ/D. Achieved in space (Hubble, JWST) or on the ground with adaptive optics locking onto the turbulence in real time.
• Seeing-limited — resolution set by the atmosphere's Fried parameter r₀, typically 10–20 cm at a good site. Turbulence caps ground images at roughly 0.5–1.5 arcsec no matter how big the mirror is. A 10 m telescope with a 0.013 arcsec diffraction limit still delivers ~1 arcsec unless corrected.

The practical resolution is the worse (larger) of the two angles. This is the whole reason adaptive optics, space telescopes, and interferometry exist — each is a strategy to recover the diffraction limit that the atmosphere would otherwise steal.

A note on history

George Biddell Airy, Astronomer Royal, derived the circular-aperture diffraction pattern in 1835 while studying the images stars form in telescopes; the central maximum still bears his name. Lord Rayleigh — John William Strutt, later a Nobel laureate for discovering argon — formalized the resolution criterion in the 1870s and 1880s in the context of spectroscopy and telescopes, choosing the first-dark-ring overlap as a convenient, physically motivated threshold. Rayleigh himself was clear that the choice is a convention: nothing in physics forbids resolving slightly finer detail, and the tighter Sparrow limit (~0.95 λ/D, where the central dip just vanishes) or high-signal model-fitting can do somewhat better.

Common misconceptions

• More magnification means more detail. No — magnification only enlarges what's there. Past the diffraction limit you get "empty magnification": a bigger, blurrier blob with no new information.
• The Rayleigh criterion is an unbreakable wall. It is a useful convention. Super-resolution, deconvolution, and the Sparrow limit all reach slightly finer, given enough signal-to-noise and prior knowledge of the source.
• Bigger telescopes just collect more light. They do (light-grasp ∝ D²), but they also resolve finer detail (θ ∝ 1/D). Both matter and both improve with aperture.
• Ground telescopes always hit their diffraction limit. Almost never in the optical — atmospheric seeing usually dominates unless adaptive optics or interferometry is used.
• Interferometry makes telescopes "see farther." It sharpens angular resolution (replaces D with baseline B); it does not collect more light or reach fainter objects on its own.
• Longer wavelengths are always worse. For a fixed aperture, yes, θ grows with λ — but radio interferometers overcome this with enormous baselines, which is exactly how the EHT beats every optical telescope on angular scale.