Optical Telescopes and Their Optics

An optical telescope is a photon bucket with a lens for a bottom. Everything it can do for astrophysics is fixed by one dominant number — the clear aperture diameter $D$ — expressed through two nearly independent figures of merit. The first is light-gathering power: the collected flux is the incident flux times the collecting area $\pi(D/2)^2$ , so doubling $D$ quadruples the photons per second and pushes the faint limit deeper. The second is resolving power: because light is a wave, a finite aperture cannot localise a direction more sharply than $\sim\lambda/D$ , and a point source is smeared into an Airy pattern whose central disk subtends $\theta_{\min}\approx1.22\,\lambda/D$ . Larger apertures therefore win twice — more light and finer detail — which is why $D$ , not magnification, is the honest headline of any research instrument. Around those two capabilities sits a small vocabulary of geometry. The focal length $f$ sets the physical scale of the image but not its quality; the focal ratio $f/\#=f/D$ governs how bright an extended image lands on the detector; and the plate scale $\mathrm{d}\theta/\mathrm{d}x=1/f$ converts millimetres at the focus into arcseconds on the sky. Reality then intrudes: from the ground, atmospheric turbulence — "seeing" — usually blurs images to $0.5$ – $1.5''$ regardless of $D$ , which drove the march from refractors to reflectors and the modern turn to adaptive optics and space. The quantities defined here are the entry point to all quantitative astrophysics: they turn the flux and magnitude of a distant source into a measurable signal feeding photometry, astrometry, and the entire cosmic distance ladder.

Key takeaways

A telescope does two jobs: it gathers light over an aperture of diameter $D$ (collected flux $\propto D^2$ ) and resolves detail down to a diffraction floor $\theta_{\min}=1.22\,\lambda/D$ set by the wave nature of light.
The $D^2$ area law becomes a depth law in magnitudes: a telescope reaches $\Delta m=5\log_{10}(D_1/D_2)$ magnitudes fainter than a smaller one — a 10 cm aperture beats the dark-adapted eye by $\approx5.8$ mag.
The focal ratio $f/\#=f/D$ fixes image surface brightness; the plate scale $\mathrm{d}\theta/\mathrm{d}x=1/f=206265''/f\,[\mathrm{mm}]$ links focal-plane millimetres to sky arcseconds and sets detector sampling.
From the ground, seeing ( $\sim1''$ ) usually beats diffraction for any $D\gtrsim0.14$ m at $\lambda=550$ nm, so extra aperture buys depth, not sharpness, unless adaptive optics or space removes the atmosphere.
Reflectors dominate above $\sim1$ m because mirrors are achromatic and back-supported, whereas lenses sag and disperse colour; the largest refractor ever built (Yerkes, 1.02 m) is a museum piece beside 8–39 m mirrors.
Key idea: A telescope is characterised by two nearly independent figures of merit — light-gathering power $\propto D^2$ (how faint you can go) and diffraction-limited resolution $\propto\lambda/D$ (how fine you can see) — and both improve with aperture, which is why $D$ is the single most important number describing a telescope.

The core insight

A telescope is characterised by two nearly independent figures of merit: light-gathering power $\propto D^2$ (how faint you can go) and diffraction-limited resolution $\propto\lambda/D$ (how fine you can see). Both improve with aperture, which is why $D$ is the single most important number describing a telescope.

Where this lives today

Aperture and diffraction still write the roadmap of observational astronomy. The frontier of light-gathering power is the 39 m Extremely Large Telescope, whose segmented primary collects roughly $15\times$ more light than a 10 m Keck mirror and, with built-in adaptive optics, delivers a diffraction limit near $0.005''$ in the near-infrared — sharp enough to image exoplanets and resolve stars in nearby galaxies. In space, JWST's 6.5 m primary reaches its $0.077''$ diffraction limit at $2\ \mu\mathrm{m}$ precisely because there is no seeing to fight. On the ground, the technology that reclaims the diffraction limit is adaptive optics with laser guide stars: a sodium-layer laser paints an artificial star, a wavefront sensor reads the turbulence hundreds of times a second, and a deformable mirror cancels it, letting 8–10 m telescopes rival Hubble's sharpness over a small field. At the survey extreme, the Vera C. Rubin Observatory pairs an 8.4 m primary with a fast $f/1.2$ optical train and a 3.2-gigapixel camera to image the entire southern sky every few nights — trading focal ratio and plate scale for etendue (aperture times field of view). The open problems are aperture problems in disguise: cost scales steeply with mirror size, so segmentation, active optics, and lightweight mirrors are what make the next factor of two in $D$ — and its factor of four in photons — affordable at all.

The physical picture

Hold the two jobs side by side. As a bucket, the aperture catches a rain of photons; a wider mouth catches proportionally more per second, so a source trickling in below the sky noise can be built up above it. That is light-gathering power, and it grows as the area, hence as $D^2$ . As a wave-sampler, the same aperture can only pin down the arrival direction of a wavefront to about $\lambda/D$ radians — the Fourier trade-off that limits any finite instrument — so a wider aperture samples more of the wavefront and localises the source more sharply. The first figure fixes the geometry that turns a captured wavefront into a focused image and the two numbers a designer specifies.

Figure

How a telescope focuses starlight: parallel rays from a source at infinity fill the aperture D and converge at the focal length f, fixing the focal ratio f slash D and the plate scale 206265 arcsec over f — here D = 200 mm, f = 2400 mm give f slash 12 and 86 arcsec per mm

Focal length sets the scale of the picture, not its quality: a long $f$ spreads the same field angle over more millimetres at the focus (a small plate scale, a large image), while $D$ sets how much light and detail that image actually contains. The ratio $f/D$ then decides how concentrated the light of an extended object is per unit detector area — a fast $f/2$ system floods a bright, wide image onto the chip; a slow $f/15$ system gives a dim, finely sampled one. Mirrors beat lenses at large size for a plain mechanical reason: a lens can only be held at its rim and sags under its own weight, and it bends colours differently (chromatic aberration), whereas a mirror can be supported across its entire back and reflects every wavelength identically. The second figure is the one that decides observing strategy — it plots the diffraction floor against aperture and overlays the atmosphere.

Figure

Diffraction limit θ = 1.22 λ over D at λ = 550 nm falls as one over D, but ground-based seeing of 0.5 to 1.5 arcsec is a floor — beyond D ≈ 0.14 m the sky, not the optics, sets resolution unless adaptive optics or space intervenes; the eye reaches 20 arcsec, HST 0.058 arcsec, the ELT 0.0036 arcsec

The core equations

The first job of a telescope is light collection, and it is pure geometry: a source of fixed flux delivers photons in proportion to the unobstructed area the aperture presents to the sky. Because area scales as the square of the diameter, the collection rate does too, which is why aperture is quoted as a diameter and yet buys depth as a square.

P \propto A = \pi\left(\frac{D}{2}\right)^2 \qquad (1)

where: $P$ — light-gathering power (photon collection rate for a fixed source); $A$ — unobstructed collecting area; $D$ — aperture (clear) diameter. Reading: double $D$ , quadruple the area, gather four times the photons in the same exposure. Comparing two apertures, the constant of proportionality and the source flux cancel, leaving a clean ratio that is the workhorse of survey design. Fed through Pogson's magnitude definition it becomes a depth law: how many magnitudes fainter the bigger dish reaches.

\frac{P_1}{P_2}=\left(\frac{D_1}{D_2}\right)^2 \qquad\Longrightarrow\qquad \Delta m = 5\log_{10}\frac{D_1}{D_2} \qquad (2)

where: $P_1,P_2$ — light-gathering powers; $D_1,D_2$ — aperture diameters; $\Delta m$ — magnitudes of extra depth at fixed exposure. Reading: each factor $\approx1.58$ in diameter buys one magnitude, because $10^{0.4}=2.51$ in flux needs $\sqrt{2.51}=1.58$ in $D$ . The second job is resolution, and here the wave nature of light sets a hard floor. Even a flawless aperture spreads a point source into the Airy pattern, and two sources are declared "just resolved" when one Airy peak sits on the other's first dark ring — the Rayleigh criterion.

\theta_{\min}=1.22\,\frac{\lambda}{D} \qquad (3)

where: $\theta_{\min}$ — smallest resolvable angular separation (radians); $\lambda$ — observing wavelength; $D$ — aperture diameter; the factor $1.22=3.8317/\pi$ is set by the first zero of the Bessel function $J_1$ . Reading: resolution sharpens as $1/D$ and coarsens as $\lambda$ grows, so infrared instruments need bigger mirrors for the same sharpness. Image geometry is governed by the focal length through two derived quantities. The focal ratio compresses "how fast" the system is into one number, and the plate scale converts focal-plane position into sky angle.

f/\# = \frac{f}{D}, \qquad \frac{\mathrm{d}\theta}{\mathrm{d}x}=\frac{1}{f}=\frac{206265''}{f\,[\mathrm{mm}]} \qquad (4)

where: $f$ — effective focal length; $f/\#$ — focal ratio ("f-number"); $\mathrm{d}\theta/\mathrm{d}x$ — plate scale (angle per unit length at the focus); $206265$ — arcseconds in one radian. Reading: long $f$ means a small plate scale — more millimetres per arcsecond — spreading detail across more pixels; short $f/\#$ concentrates an extended image into a brighter, smaller footprint. Finally, tying depth to aperture directly, the limiting magnitude — the faintest detectable star at fixed exposure and detection threshold — scales exactly as the depth law of equation (2).

m_{\lim}\approx m_0 + 5\log_{10}\!\left(\frac{D}{D_0}\right) \qquad (5)

where: $m_{\lim}$ — faintest detectable magnitude; $m_0,D_0$ — a reference limiting magnitude and aperture; $D$ — aperture diameter. Reading: this is why "how deep?" and "how big is the mirror?" are almost the same question — until sky background and read noise enter.

Derivation AThe diffraction limit from a circular aperture

The goal is to show that a perfect telescope, with flawless optics and no atmosphere, still cannot resolve arbitrarily fine detail, and to pin the exact coefficient $1.22$ in $\theta_{\min}=1.22\,\lambda/D$ . A plane wave from a point source at infinity fills a circular aperture of diameter $D$ ; by Fraunhofer (far-field) diffraction the intensity in the focal plane is the squared modulus of the Fourier transform of the aperture, and for a circle that transform is an Airy function. We locate its first dark ring and translate that angle into the Rayleigh two-source criterion.

Step 1. Write the far-field intensity of a uniformly illuminated circular aperture as the Airy pattern, $I(\theta)=I_0\,[\,2J_1(x)/x\,]^2$ with $x=\pi D\sin\theta/\lambda$ , where $J_1$ is the first-order Bessel function. Why: the circular aperture's Fourier transform is $2J_1(x)/x$ , so the point image is not a point but a bright central disk ringed by faint diffraction rings — the fundamental blur of any lens or mirror.

Step 2. Find the first dark ring by setting $J_1(x)=0$ at its first nontrivial root, $x_1=3.8317$ . Why: the edge of the central Airy disk — the natural "size" of the blurred point — is the first zero of the intensity, and Bessel-function tables fix that root once and for all.

Step 3. Solve $x_1=\pi D\sin\theta/\lambda$ for the angle: $\sin\theta=(3.8317/\pi)\,\lambda/D=1.2197\,\lambda/D$ , and for the small angles of astronomy $\sin\theta\approx\theta$ , giving $\theta\approx1.22\,\lambda/D$ to the radius of the first dark ring. Why: this is where the pure number $1.22$ comes from — it is $3.8317/\pi$ , nothing more mysterious than a Bessel root divided by $\pi$ .

Step 4. Apply the Rayleigh criterion: two equal point sources are "just resolved" when the peak of one Airy disk falls on the first minimum of the other, i.e. when their separation equals $\theta_{\min}=1.22\,\lambda/D$ . Why: the criterion turns the single-aperture blur into an operational resolution — closer than $\theta_{\min}$ the two disks merge into one, farther apart they show a dip between them. *As a check: $3.8317/\pi=1.2197$ ; for $\lambda=550$ nm and $D=1$ m, $\theta_{\min}=6.71\times10^{-7}$ rad $=0.138''$ , confirming the plotted line $\theta=0.1384/D$ arcsec.*

Go deeper · Derivation B · 1 minMagnitude gain from aperture

Derivation BMagnitude gain from aperture

The goal is to convert the geometric $D^2$ area law into the language observers actually use — magnitudes — and so answer "how much fainter can the bigger telescope see?" quantitatively. For a fixed source, exposure time, and detector, the number of collected photons is proportional to collecting area, and the magnitude scale is logarithmic in flux by construction (Pogson). Threading the area law through Pogson's relation gives a depth law that depends only on the ratio of diameters, independent of the source.

Step 1. Write the flux ratio as the area ratio, $P_1/P_2=A_1/A_2=(D_1/D_2)^2$ , since both apertures view the same source and area scales as diameter squared. Why: the source flux and every constant cancel in the ratio, leaving pure geometry — a clean, source-independent statement.

Step 2. Insert into Pogson's definition of magnitude difference, $m_1-m_2=-2.5\log_{10}(P_1/P_2)$ , which encodes the historical five-magnitudes-per-factor-100 convention. Why: magnitudes are defined logarithmically so that equal ratios of flux map to equal steps of magnitude — the natural currency for "how much deeper".

Step 3. Combine: $\Delta m=-2.5\log_{10}(D_1/D_2)^2=-5\log_{10}(D_1/D_2)$ , so the larger aperture reaches a more negative $\Delta m$ , i.e. detects larger (fainter) magnitudes. Why: the factor $2$ from squaring the diameter turns Pogson's $2.5$ into a clean $5$ — the origin of the ubiquitous "five log D" in survey depth.

Step 4. Evaluate for the dark-adapted eye ( $D_2=7$ mm) versus a modest 10 cm telescope: $\Delta m=5\log_{10}(0.10/0.007)=5\log_{10}(14.3)=5.8$ magnitudes. Why: this is the everyday result that a small telescope carries the naked-eye limit of $m\approx6$ down to $m\approx12$ , revealing thousands of stars invisible to the eye. *As a check: $\log_{10}(0.10/0.007)=\log_{10}(14.29)=1.155$ , times $5$ gives $\boldsymbol{5.77}$ magnitudes.*

Go deeper · Derivation C · 1 minPlate scale and the f-number from paraxial imaging

Derivation CPlate scale and the f-number from paraxial imaging

The goal is to derive the two geometric numbers that decide how an image lands on a detector — the plate scale and the focal ratio — from a single paraxial fact, and to justify the magic constant $206265$ . A ray from a source at field angle $\theta$ to the optical axis is brought by the optic to a focal-plane height $x$ ; in the paraxial (small-angle) regime this mapping is linear, and its slope is the focal length. Differentiating and inverting gives the plate scale, and expressing the aperture as a fraction of that focal length gives the f-number.

Step 1. Write the paraxial image height, $x=f\tan\theta\approx f\,\theta$ , for a source at field angle $\theta$ . Why: to first order the optic maps sky angle linearly onto focal-plane position, with the focal length as the conversion factor — the whole basis of astrometry on a detector.

Step 2. Differentiate to get $\mathrm{d}x/\mathrm{d}\theta=f$ , hence the plate scale $\mathrm{d}\theta/\mathrm{d}x=1/f$ : a longer focal length spreads a given angle over more millimetres, sampling detail across more pixels. Why: the plate scale is literally the inverse focal length, which is why observers lengthen $f$ (e.g. with a Cassegrain secondary) when they need finer sampling.

Step 3. Convert radians to arcseconds using $1\ \mathrm{rad}=206265''$ , giving the practical form $\mathrm{d}\theta/\mathrm{d}x=206265/f\,[\mathrm{mm}]$ in arcsec per mm. Why: $206265=180\times3600/\pi$ is simply the number of arcseconds in a radian, the bridge between the radian-native optics and the arcsecond-native sky.

Step 4. Define the focal ratio $f/\#=f/D$ as the aperture measured in units of its own focal length; it fixes the converging cone angle and hence the surface brightness of an extended image. Why: two systems of the same $f/\#$ give equally bright extended images regardless of size, so $f/\#$ , not $D$ alone, is quoted for imaging speed. *As a check: for $f=2400$ mm, $206265/2400=85.9''/\mathrm{mm}$ and $f/\#=2400/200=12$ ; for $f=10$ m, $206265/10000=20.6''/\mathrm{mm}$ , so $1''$ maps to $48.5\ \mu\mathrm{m}$ .*

Limiting cases

The behaviour that matters observationally lives in three regimes — where diffraction wins, where the atmosphere wins, and where you are simply photon-starved.

Case	Result	Meaning
Diffraction limited (space or adaptive optics)	$\theta_{\min}\simeq1.22\,\lambda/D$	Ideal case: sharpness improves as $1/D$ . A 2.4 m telescope at $550$ nm reaches $0.058''$ ; a 39 m ELT reaches $0.0036''$ .
Seeing limited (typical ground)	$\theta\approx0.5\text{–}1.5''$	Atmospheric turbulence dominates and is nearly independent of $D$ , so beyond $D\approx0.14$ m extra aperture buys only light, not sharpness.
Light-gathering limit (faint sources)	$P\propto D^2$	Doubling $D$ quadruples collected flux, worth $2.5\log_{10}4\approx1.5$ magnitudes deeper at fixed exposure — the payoff that always survives.
Infrared vs optical	$\theta_{\min}\propto\lambda$	At $\lambda=2\ \mu\mathrm{m}$ the diffraction disk is $\sim3.6\times$ coarser than at $550$ nm, so JWST's 6.5 m gives $0.077''$ in the near-IR — still space-sharp because there is no seeing.

Common trapMagnification is a meaningful figure of merit for a research telescope.

It is not. Magnification $M=f_{\mathrm{obj}}/f_{\mathrm{eye}}$ is set by whichever eyepiece you screw in, and it cannot manufacture information the aperture never captured — push $M$ too high and you get a large, dim, blurry image ("empty magnification"), not a better one. What actually decides scientific reach is aperture (light and resolution) and the plate-scale/field-of-view match to the detector. A related trap: do not confuse the diffraction limit $1.22\,\lambda/D$ , an optical ideal that keeps improving with $D$ , with atmospheric seeing ( $\sim1''$ ), which from the ground usually dominates and is nearly independent of $D$ . A 10 m mirror on a mediocre site delivers $1''$ images unless it corrects the wavefront — its $0.014''$ diffraction limit is potential, not delivered.

Worked examples

1 · Fully specifying a telescope (all the numbers at once)

A Cassegrain has clear aperture $D=200$ mm and effective focal length $f=2400$ mm. Find (a) its focal ratio, (b) its diffraction limit at $\lambda=550$ nm, (c) its plate scale, (d) the pixel size that Nyquist-samples that diffraction disk, and (e) how many magnitudes deeper it goes than a $7$ mm dark-adapted eye.

Step 1. Focal ratio is the aperture in units of its focal length: $f/\#=f/D=2400/200=\mathbf{f/12}$ — a "slow", finely sampling system.

Step 2. Diffraction limit from equation (3): $\theta_{\min}=1.22\,\lambda/D=1.22\times(5.5\times10^{-7})/0.200=3.36\times10^{-6}$ rad. Convert with $206265''/\mathrm{rad}$ : $\theta_{\min}=\mathbf{0.69''}$ . (On any real ground site this is finer than the seeing, so the atmosphere, not the optics, sets the delivered image.)

Step 3. Plate scale from equation (4): $206265/f=206265/2400=\mathbf{86''/mm}$ , i.e. one arcsecond maps to $1/86=0.0116$ mm $=11.6\ \mu\mathrm{m}$ at the focus.

Step 4. Nyquist wants the $0.69''$ diffraction disk sampled by about two pixels, so one pixel should subtend $\approx0.35''$ . In millimetres that is $0.35/86=4.0\times10^{-3}$ mm $=\mathbf{4.0\ \mu m}$ per pixel — a fine but ordinary CMOS pitch.

Step 5. Depth versus the eye from equation (2): $\Delta m=5\log_{10}(200/7)=5\log_{10}(28.6)=5\times1.456=\mathbf{7.3}$ magnitudes, carrying a naked-eye limit of $m\approx6$ down to $m\approx13$ .

Result: $f/12$ , $\theta_{\min}=0.69''$ , plate scale $86''/\mathrm{mm}$ , Nyquist pixel $\approx4\ \mu\mathrm{m}$ , and $\approx7.3$ magnitudes of depth — a data sheet from two inputs.

2 · Space versus ground, and why aperture is not everything

Compare Hubble ( $D=2.4$ m, above the atmosphere) with an $8.2$ m ground telescope at a good site (seeing $0.8''$ ), at $\lambda=550$ nm. Which sees fainter, and which sees sharper?

Step 1. Depth (light-gathering): the $8.2$ m mirror wins by $\Delta m=5\log_{10}(8.2/2.4)=5\log_{10}(3.42)=\mathbf{2.67}$ magnitudes — it collects $(8.2/2.4)^2=11.7\times$ more photons, so at equal exposure it reaches far fainter.

Step 2. Diffraction limits: HST at $1.22\times5.5\times10^{-7}/2.4=0.058''$ ; the 8.2 m at $0.017''$ . On paper the ground telescope is $\sim3.4\times$ sharper.

Step 3. Delivered sharpness: the ground images are smeared to the seeing, $0.8''$ , more than $10\times$ coarser than HST's diffraction-limited $\mathbf{0.058''}$ . Uncorrected, Hubble resolves structure the larger mirror cannot.

Step 4. The fix: laser-guide-star adaptive optics measures and cancels the turbulent wavefront in real time, letting the 8.2 m approach its $0.017''$ limit over a small field — recovering both more light and finer detail than HST.

Result: the bigger ground mirror always wins on depth ( $+2.7$ mag), but wins on sharpness only once adaptive optics defeats the $0.8''$ seeing floor — aperture and delivered resolution are different questions.

Physics of the portrait

The telescope is where a measured curiosity — that glass bends light and colours it — became an engine for astrophysics. The pivotal move was to stop bending light and start reflecting it, and that move belongs to the same man who explained the colour in the first place.

Origins · 2 minIsaac Newton and how this was discovered

Isaac Newton

1643–1727 · the reflecting telescope and the optics of light collection

Isaac Newton did not merely improve the telescope; he diagnosed why the refractors of his day were incurably blurred and then routed around the disease. His prism experiments of 1666 showed that white light is a mixture of colours and that glass refracts each colour by a different amount — chromatic aberration — so a single lens can never bring all colours to one focus, and the long, spindly refractors of the era smeared bright stars into coloured haloes. Newton's conclusion (too pessimistic, as it turned out) was that refractors were a dead end. His answer, built with his own hands in 1668 and presented to the Royal Society in 1671, was the first working reflecting telescope: a concave metal primary mirror gathered and focused the light, a small flat "Newtonian" secondary kicked the beam out to the side, and because reflection treats every wavelength identically, the image was achromatic. The instrument was barely 15 cm long yet outperformed refractors many times its length. That principle — collect light with a back-supported mirror, not a rim-supported lens — is why every research telescope above about a metre, from the 5 m Hale to the 10 m Keck twins to the 39 m ELT, is a reflector. The achromatic doublet (Chester Moore Hall, then John Dollond) later tamed colour in lenses well enough for refractors to survive in astrometry into the twentieth century, but for raw aperture the mirror won permanently. The two figures of merit this article is built on — light-gathering power and diffraction-limited resolution — are both maximised by the large, achromatic, back-supported mirror that Newton first put to the sky.

History

1608

— Hans Lipperhey files the earliest known telescope patent in the Netherlands

1609

— Galileo turns a refractor skyward and resolves Jupiter's four largest moons

1668

— Newton builds the first working reflecting telescope, sidestepping chromatic aberration

1758

— John Dollond patents the achromatic doublet, taming colour in refracting lenses

1879

— Rayleigh states the resolution criterion θ = 1.22 λ over D for a circular aperture

1990

— the 2.4 m Hubble Space Telescope reaches the diffraction limit from orbit

Going further

Texts. Carroll and Ostlie, An Introduction to Modern Astrophysics, 2nd ed., Ch. 6 (Telescopes) · Kitchin, Astrophysical Techniques, 6th ed., Ch. 1 · Schroeder, Astronomical Optics, 2nd ed. · Binney and Tremaine, Galactic Dynamics, 2nd ed. (observational appendices) · Hecht, Optics, 5th ed., §5.7 and §10.2 (diffraction and the Airy pattern). Key numerical anchors: the Airy/Rayleigh coefficient is $3.8317/\pi=1.2197\approx1.22$ ; $\theta_{\min}=1.22\,\lambda/D=0.138''$ for $\lambda=550$ nm, $D=1$ m (plotted line $\theta=0.1384/D$ arcsec); the depth law $\Delta m=5\log_{10}(D_1/D_2)$ follows from Pogson's relation with $A\propto D^2$ (eye example $5\log_{10}(0.10/0.007)=5.77$ ); the plate scale is $206265/f$ using $1\ \mathrm{rad}=206264.8''$ ; worked-example values are $f/12$ , $\theta_{\min}=0.69''$ , $86''/\mathrm{mm}$ . The Yerkes 1.02 m refractor, HST 2.4 m, VLT 8.2 m, Keck 10 m, JWST 6.5 m, and ELT 39 m apertures are standard catalogue values.

Prerequisites. Luminosity and Radiant Flux · The Magnitude System · The Celestial Sphere. Leads to. Photometric Systems and Filters · Stellar Parallax and the Parsec · The Cosmic Distance Ladder · Standard Candles.

Ask the tutor: Ask me to size an aperture for a target magnitude, decide whether your telescope is seeing- or diffraction-limited, or match a camera's pixels to a focal length.

Why is aperture, not magnification, the number that defines a telescope?

Because aperture sets the two things you cannot fake — how many photons you catch and how fine an angle you can resolve — while magnification only reformats what the aperture already captured. Light-gathering power grows as $D^2$ , so a bigger dish reaches fainter sources; diffraction-limited resolution improves as $1/D$ , so a bigger dish sees finer detail. Magnification is just the ratio of objective to eyepiece focal length; crank it past the point where the aperture's information runs out and you get a large, dim, blurry image, not a better one. That is why telescopes are advertised by their diameter — "an 8-inch", "a 10-metre" — and serious observers barely mention magnification at all.

What is the difference between the diffraction limit and seeing?

The diffraction limit, $1.22\,\lambda/D$ , is an optical ideal: the sharpest image a perfect aperture of diameter $D$ can form, set only by the wave nature of light. It improves without bound as you build bigger — a 39 m mirror reaches $0.004''$ . Seeing is a separate, atmospheric limit: turbulence in the air scrambles the incoming wavefront and smears stars to $0.5$ – $1.5''$ at good sites, almost independently of $D$ . For any aperture larger than about $0.14$ m at visible wavelengths, seeing is the coarser of the two, so the atmosphere — not the optics — sets the delivered resolution. Adaptive optics or going to space removes the atmosphere and lets the diffraction limit take over again.

If a bigger mirror gives sharper images, why does a 10 m ground telescope still look blurry?

On paper its diffraction limit is $0.014''$ , but the atmosphere never lets that potential through unaided. Turbulent cells of air act like a shifting, lumpy lens in front of the aperture, and by the time the wavefront reaches the mirror it is corrugated on scales of $10$ – $20$ cm (the Fried parameter $r_0$ ). The mirror faithfully focuses that corrugated wavefront into a $\sim1''$ blur. The extra aperture still helps enormously — it collects far more light — but the sharpness gain is thrown away until adaptive optics measures the corrugation hundreds of times a second and a deformable mirror irons it out, or until you observe from orbit where there is no air at all.

Why did astronomy abandon giant refractors for reflectors?

Two incurable problems with large lenses. First, chromatic aberration: glass bends blue light more than red, so a single lens focuses colours at different points, ringing bright stars with colour; achromatic doublets reduce but never fully cure this. Second, mechanics: a lens can only be supported at its rim, so under its own weight it sags and deforms, and light must pass through it, so any bubble or striae in the glass matters. A mirror bends no colours (reflection is achromatic), can be supported across its entire back so it holds its figure, and only its front surface needs to be perfect. The largest refractor ever built, the 1.02 m Yerkes lens of 1897, was essentially the practical ceiling; every telescope larger than that is a reflector.

What does the focal ratio f-number actually control?

The focal ratio $f/\#=f/D$ sets how concentrated the light of an extended object is at the focus, and hence the surface brightness and exposure time for extended sources. A "fast" system like $f/2$ pours the light of a nebula into a small, bright patch — great for wide-field survey imaging with short exposures. A "slow" system like $f/15$ spreads the same object over a large, dim area — better for high-resolution work and planetary imaging where you want fine sampling. Crucially, for extended sources the exposure time depends on $f/\#$ , not on $D$ alone: two telescopes at the same $f/\#$ take the same exposure on a nebula regardless of size. For point sources like stars, though, it is total aperture $D$ that sets how many photons you catch.

How do you match a detector's pixels to a telescope?

You use the plate scale to make the pixels sample the smallest real detail — the diffraction disk in space, or the seeing disk on the ground — at roughly two pixels across (the Nyquist criterion). The plate scale is $206265/f$ arcseconds per millimetre, so dividing by the pixel pitch gives arcseconds per pixel. If the seeing is $1''$ you want about $0.5''$ per pixel, which fixes the focal length. Undersampling throws away resolution; oversampling spreads each star over many pixels, wasting field of view and piling up read noise. Good design tunes $f$ so the science-limiting blur lands on about two pixels.