UNFOLD
A FIELD GUIDE TO THE JAMES WEBB SPACE TELESCOPE
Ten chapters on a telescope too large to launch fully built — folded like origami into a rocket fairing, unpacked over two weeks a million and a half kilometres from any repair crew, and now rewriting how early the universe got interesting. Scroll to begin the mission.
BEGIN — CHAPTER 01 →A parking spot a million miles out
Hubble circles Earth a few hundred kilometres up and slips into Earth's shadow about every hour and a half [1]. Webb does not orbit Earth at all. It circles the Sun about 1.5 million km beyond Earth, looping around the second Sun–Earth Lagrange point, L2, once every six months or so on a halo orbit about as wide as the Moon's orbit around Earth [1]. L2 is a balance point, not a trap: left alone, a spacecraft drifts away, so Webb fires small station-keeping burns to stay [1][2]. Why the balance exists is the subject of EQUILIBRIUM.
From there the Sun, Earth and Moon always sit on the same side of the sky, so one sunshield can block all three at once and keep the telescope in permanent shadow, with no day and night to warm and cool it [1]. That matters for an instrument that must stay colder than almost anything else humans have built. The same spot lets the Deep Space Network's antennas in Australia, Spain and California stay in touch as Earth turns [1].
A 318 °C drop across five thin layers
The sunshield is 21.2 by 14.2 metres — about the size of a tennis court — and is made of five layers of Kapton film spaced apart so heat can radiate away out of the gaps between them [3]. The sun-facing layer is 50 microns thick and the other four are 25 microns, thinner than a human hair [4].
NASA's thermal models put the outermost layer at up to 383 K (110 °C) and the coldest part of layer 5 at 36 K (−237 °C); the cards above are NASA's typical hot- and cold-side figures, 85 °C and −233 °C, and their difference [3]. Behind the shield the telescope runs at roughly 40 K: in flight the primary-mirror segments settled at 35–55 K, warmest nearest the shield [2].
Eighteen hexagons, one mirror
At 6.5 metres across, the primary mirror was too large to launch in one piece, so it is built from 18 hexagonal segments of beryllium, each about 20 kg, coated with gold 100 nanometres thick — about 48 grams of gold for the whole mirror [3]. The segments fold for launch and were then aligned in space until they act as one mirror. In flight the wavefront error is about 60–80 nanometres, sharp enough to be diffraction-limited at wavelengths as short as 1.1 microns, better than the 2-micron requirement [2]. How 18 segments were phased into one is told in PARABOLA, chapter 5.
An observatory, not just a telescope
Four instruments, a mirror, a sunshield, and a spacecraft bus — every piece with no possibility of a repair mission. Select a component.
Two weeks of unfolding, no rehearsal in space
Before launch the project counted 344 single-point failures — parts with no backup whose failure could end the mission — and most of them were retired during the deployments of the first two weeks [5]. None of it could be rehearsed in real conditions: the sequence had to work the first time, a million and a half kilometres past any repair mission's reach. Step through the sequence.
Five images, six months in the making
Seeing through dust, and back through time
Infrared light passes through the dust clouds that block visible light, letting Webb see stars being born inside nebulae that would otherwise look like solid dark walls.
It also solves a problem unique to the edge of the observable universe. As light crosses expanding space its wavelength is stretched by a factor of 1 + z, the redshift. Ultraviolet light from the first galaxies arrives as infrared, beyond Hubble's reddest camera band at 1.7 microns [18] but well inside Webb's 0.6–28.5 micron range [3].
Infrared has a price: everything warm glows in it. Wien's law puts the peak of a body's glow at 2,898 microns divided by its temperature in kelvins [15]. A telescope at room temperature would glow brightest near 10 microns, right in the middle of the band it is trying to see — which is why Webb's mirror segments are kept at 35–55 K and MIRI's detectors at 6.4 K [2]. Try both effects below.
Peak wavelength λ = b / T with b = 2,897.77 µm·K [15]; the Sun's effective temperature 5,772 K [16]. Hydrogen's Lyman-alpha (121.57 nm) and H-alpha (656.47 nm) are computed from the Rydberg constant in vacuum [17] and multiplied by 1 + z. The age at each redshift comes from a flat ΛCDM model with Planck 2018 parameters (H₀ = 67.36, Ωm = 0.3153), which gives 13.80 billion years today [20]. Ranges: Hubble's WFC3 0.2–1.7 µm [18], Webb's NIRSpec 0.6–5.3 µm [19], the whole observatory 0.6–28.5 µm [3]. The glow curve is drawn scaled to its own peak: it shows shape, not brightness.
The early universe grew up faster than expected
Two eyes, two kinds of light, one target
Astronomers regularly point both telescopes at the same object, combining Hubble's ultraviolet, visible and near-infrared view with Webb's infrared one to build a far more complete picture than either could alone.
A bigger mirror buys two things. It gathers more light, and it sees finer detail: the smallest angle a telescope can separate, its diffraction limit, is about 1.22 λ / D. But the limit grows with wavelength, so Webb's 6.5-metre mirror working in the infrared is not automatically sharper than Hubble's 2.4-metre mirror working in visible light. Slide the wavelength and compare.
Rayleigh limit θ = 1.22 λ / D for a circular mirror, with D = 6.5 m for Webb [3] and 2.4 m for Hubble [30]; Webb's segmented, hexagonal mirror gives a slightly different pattern, so treat its figure as a guide. Webb is diffraction-limited from 1.1 µm upward; below that its images are a little softer than the formula [2]. Pale disc: the Rayleigh limit; dark core: the width at half brightness (FWHM ≈ 0.42 of it). Collecting areas: Webb 25.4 m², measured in flight with obstructions already excluded [2]; Hubble 4.5 m², the full 2.4-metre disc — subtracting Hubble's central obstruction would make the ratio somewhat larger.
Built for decades, not years
Webb's only consumable is propellant for station-keeping and for unloading its reaction wheels. It was required to carry enough for at least 10.5 years; after an accurate launch and frugal course corrections, the commissioning team concluded the remainder will last more than 20 years [2]. Its successors are already flying or on the drawing board.
Sources cited throughout this guide
Bracketed numbers in the text above correspond to the sources below.