Space telescopes

Putting a telescope in orbit costs perhaps a hundred times what an equivalent ground telescope costs, and ground apertures are far larger. The case for doing it anyway rests on three distinct advantages that apply in different measure to different kinds of astronomy.

Three reasons, not one

Transparency. The atmosphere is opaque across most of the spectrum. Gamma rays, X-rays, most ultraviolet and most infrared never reach the ground at all. For those wavelengths, orbit is not an improvement but the only option.

Seeing. Atmospheric turbulence blurs images to about one arcsecond at a good site, far worse than the diffraction limit of any large telescope. This was once the dominant argument for space, and adaptive optics has substantially eroded it: a modern ground telescope with AO achieves diffraction-limited performance over a narrow field in the infrared, and its aperture is far larger than anything orbited.

Background. A warm telescope glows in the infrared, and so does the atmosphere. For mid-infrared observation the instrument’s own thermal emission overwhelms the signal unless the optics are cryogenic, which is far easier in space where deep space is a 3 K heat sink.

The honest summary is that space is decisively better for wavelengths the atmosphere blocks and for infrared background, and increasingly contested for visible-light resolution. This is why the newest flagship observatories are infrared instruments rather than visible ones.

Diffraction and aperture

Resolution is bounded by the same physics as any optic:

The second and third rows make an important point. JWST has nearly three times Hubble’s aperture but comparable angular resolution, because it observes at four times the wavelength. JWST is not a sharper Hubble; it is a far more sensitive instrument at wavelengths Hubble cannot reach, and sensitivity comes from collecting area and from being cold.

Segmented mirrors

A 6.5 m monolithic mirror does not fit in any launch fairing. JWST’s primary is therefore built from 18 hexagonal beryllium segments that fold for launch and deploy in orbit, each with actuators for position and radius of curvature.

Aligning them is the hard part. The segments must be co-phased to a fraction of a wavelength, tens of nanometres, so that the array behaves as a single optical surface rather than 18 separate telescopes. That alignment was performed in orbit over months using the science instruments themselves as the wavefront sensor.

Beryllium was chosen for its stiffness-to-density ratio and its low thermal expansion at cryogenic temperature, which matters because the mirror operates below 50 K and must hold its figure there.

The alternative to folding is assembly in orbit, which would remove the deployment risk entirely at the cost of requiring an assembly capability that does not yet exist. See In-space manufacturing and assembly.

Passive cooling and orbit selection

Orbit choice follows from what the telescope needs.

The $L_2$ argument is worth stating precisely because it is the reason JWST works. At Sun-Earth $L_2$, the Sun, Earth and Moon all lie within a narrow cone in one direction. A single sunshield therefore blocks all three at once, and the telescope can radiate freely to deep space on the other side.

JWST’s shield is roughly 21 by 14 m in five layers, holding the sun-facing side near 350 K and the observatory side below 50 K with no active cooler at all for the near-infrared instruments. Only the mid-infrared instrument, which needs 6.7 K, carries a mechanical cryocooler. See Thermal control and Lagrange points.

The cost of $L_2$ is that servicing is impractical, which is why JWST’s 344 single-point failures during deployment mattered so much. Hubble, by contrast, was serviced five times, and the first of those corrected a manufacturing defect that would otherwise have crippled the mission. See Fault tolerance and reliability.

The major observatories

Gaia deserves particular note. It is not an imaging telescope in the usual sense but an astrometric instrument, measuring positions and parallaxes to microarcseconds. Its catalogue underpins essentially all modern Galactic astronomy, because knowing distances accurately is what converts an apparent brightness into a physical one. Its data has been used by more papers than almost any other astronomical dataset.

X-ray telescopes are optically unlike the others. X-rays penetrate rather than reflect at normal incidence, so the optics are nested grazing-incidence shells that deflect photons through very shallow angles. This makes X-ray optics heavy, expensive and small in collecting area relative to their mass.

Operations and time allocation

Two features of this process matter more than they appear to.

Dual-anonymous review, adopted by Hubble in 2018 and since more widely, hides both proposer and reviewer identities. It measurably changed the distribution of allocated time, particularly by career stage and gender, which is evidence that the earlier process was not selecting purely on scientific merit.

The archive frequently outproduces the original programme. Observations are proprietary for a period and then public, and the resulting archive is used for science nobody proposed. Hubble’s archival papers outnumber those from its original allocated programmes, which is a strong argument for treating the archive as a first-class deliverable rather than a byproduct.

References

• Lallo, M. D. “Experience with the Hubble Space Telescope: 20 years of an archetype”, Optical Engineering, 51:011011, 2012.
• Gardner, J. P. et al. “The James Webb Space Telescope Mission”, Publications of the Astronomical Society of the Pacific, 135:068001, 2023.
• Menzel, M. et al. “The Design, Verification and Performance of the James Webb Space Telescope”, PASP, 135:058002, 2023.
• Weisskopf, M. C. et al. “An Overview of the Performance and Scientific Results from the Chandra X-ray Observatory”, PASP, 114:1, 2002.
• Gaia Collaboration, “The Gaia mission”, Astronomy and Astrophysics, 595:A1, 2016.
• Strolger, L. et al. “The Effectiveness of Dual-Anonymous Review in the Hubble Space Telescope Cycle 26 Time Allocation”, PASP, 2019.