Astronomy Featured Jun 16, 2026

Scientists Discovered a Planet Where It Rains Glass Sideways

HD 189733 b is a hot Jupiter whose blue colour, eastward winds and likely silicate haze are measured in part and inferred in part. We separate the data from the "glass rain" headline.

SC
ScienceTrace Science Desk
•
 8 min read
•
 1,547 words

Few exoplanets have been studied as intensively as HD 189733 b, and few have been described as loosely in popular coverage. The planet is not a recent find. It was identified in 2005 as a transiting hot Jupiter orbiting a bright, nearby K-type star about 64 light-years away in the constellation Vulpecula. What has changed over the past two decades is how much can be said about its atmosphere. The vivid phrase "it rains glass sideways" compresses several separate lines of evidence (a colour measurement, a temperature map, a wind measurement and a cloud model) into a single image. This article separates what has been measured from what has been inferred, because the difference matters for anyone using this planet as a benchmark.

A hot Jupiter that happens to be easy to observe

HD 189733 b has a mass a little above that of Jupiter and a radius somewhat larger, but it orbits at roughly 0.03 astronomical units from its star and completes an orbit in about 2.2 days. At that distance it is almost certainly tidally locked, like the Moon with respect to Earth, so one hemisphere permanently faces the star while the other faces away. Because its orbit is seen almost edge-on, the planet crosses the stellar disc on every orbit and then passes behind it half an orbit later. Both events are useful. And because the host star is bright by exoplanet standards, the number of photons collected per transit is high, which is why this system became a standard test case for atmospheric methods.

Two techniques account for most of what is known. In transmission spectroscopy, astronomers compare starlight before and during transit; the thin ring of atmosphere at the planet's day-night boundary (the terminator) absorbs and scatters a little of the light in a wavelength-dependent way. In eclipse and phase-curve measurements, the total light of star plus planet is tracked around the orbit; the dip when the planet disappears behind the star isolates the planet's own emission or reflected light. The first technique probes the terminator at high altitude; the second probes the dayside, and, through the phase curve, the whole planet in rough longitude bands.

The blue colour is a measurement; its cause is an inference

The claim that this world is blue comes from a specific observation. Using the Hubble Space Telescope's STIS spectrograph, Evans and colleagues (2013) measured how the combined light of the system fell when the planet slipped behind its star. The size of that drop at each wavelength gives the fraction of starlight the planet reflects. The planet reflected noticeably more light at the blue end of the measured range than at longer wavelengths, which is what "blue" means here. The result is an albedo spectrum, not a photograph; the deep-blue images that accompany articles are artistic renderings built on that measurement.

Why would a hot gas giant reflect blue light? The simplest explanation is scattering from very small particles, in the same family of physics (Rayleigh-type scattering) that makes the daytime sky on Earth blue. Gas that contains sodium, potassium and other absorbers tends to be dark, so the observed reflectivity implies that something bright and fine-grained sits high enough in the atmosphere to be seen. Independent transmission-spectrum work had already pointed in that direction: Pont and colleagues (2008) and Sing and colleagues (2011) found a featureless, sloping optical spectrum that suggested a high-altitude haze, and later work by Pont and colleagues (2013) combined Hubble and Spitzer data to argue that such particles are widespread in the atmosphere.

From haze to "glass": what the silicate hypothesis really says

Observations tell us that small particles are present; they do not uniquely tell us what the particles are made of. The leading proposal is that they are condensed magnesium silicates, minerals such as enstatite and forsterite. Silicates are the same family of materials as sand and glass, which is where the headline phrase comes from. But "glass" is a metaphor. The condensates in such models are generally treated as mineral grains, not molten droplets, and other candidates (sulfide clouds, or photochemical hazes made of complex carbon-bearing molecules) can produce similar scattering. The colour measurement constrains the particle size and altitude better than it constrains the chemistry.

The word "sideways" also deserves scrutiny. Tiny grains in a fast flow are carried along with the gas; they fall slowly relative to the wind, so the phrase says little more than that a windy atmosphere has suspended particles in it. A reader should not picture a horizontal downpour of shards. What can be defended is narrower and still interesting: a hot Jupiter whose upper atmosphere appears to be loaded with small scattering particles, and which is circulated by very strong winds.

Winds measured through the Doppler effect

The winds are not purely a model prediction. Louden and Wheatley (2015) used high-resolution spectra of the sodium absorption line during transit and looked for small Doppler shifts, which would arise if gas on one side of the terminator is moving toward us and gas on the other is moving away. They reported a pattern consistent with an eastward flow combined with the planet's rotation, with speeds of the order of a couple of kilometres per second and an uncertainty that is a substantial fraction of the value. The widely repeated figure of about 8,700 km/h corresponds to roughly 2.4 km/s, so it sits inside that range, but it should be quoted as an estimate with error bars rather than a settled number.

"Supersonic" is also looser than it sounds. In hot hydrogen-dominated gas at a thousand kelvin the speed of sound is on the order of 2 to 3 km/s by my own estimate, so these winds are better described as comparable to the local sound speed. That is consistent with general circulation models, which predict a fast equatorial jet in the eastward direction for tidally locked hot Jupiters.

The temperature map supports the same picture. Knutson and colleagues (2007) used the Spitzer Space Telescope at 8 micrometres to chart brightness around the orbit and found that the hottest region lies east of the point directly beneath the star, and that the day-night contrast is only a few hundred kelvin (roughly 1,200 K versus about 1,000 K in brightness temperature). A planet with no heat transport would show a far larger contrast. The eastward offset is the signature of winds that carry heat downstream before it can be radiated away.

How hot is it, really?

Earlier versions of this article cited a temperature above 1,000 degrees Celsius. The infrared photosphere is better described as roughly 1,000 to 1,200 kelvin, or about 700 to 930 degrees Celsius, with deeper layers hotter. Nothing in the data requires dramatic language. The planet is a gas giant with no solid surface in the ordinary sense, and describing it as the most hostile environment in the galaxy is rhetoric, not a measurement; hotter and more extreme planets have since been found.

Limitations and open questions

  • Stellar contamination. The host is an active star with spots and faculae. Unocculted spots can mimic or hide atmospheric slopes in transmission spectra, and this is a recognised limitation of the haze interpretation, not a footnote.
  • Composition is unresolved. Silicates, sulfides and organic hazes can all scatter blue light. Distinguishing them needs spectral features, not just slopes.
  • Different probes, different places. Transmission measures the terminator at altitude; reflected light measures the dayside. Clouds need not be the same in both places, so a single composite picture can mislead.
  • Wind uncertainty. Doppler measurements are low in signal-to-noise and depend on the assumed line model and the planet's orbital velocity, so the exact jet speed remains imprecise.

What researchers should watch

The sharpest test of the silicate hypothesis is spectral. Small silicate grains are expected to produce broad vibrational features in the mid-infrared, in the neighbourhood of 10 micrometres, whereas sulfide or organic hazes would not. Instruments that cover that range with enough precision could in principle confirm or exclude the idea. A second productive direction is asymmetry: whether the morning and evening limbs of the terminator differ in cloud content, as circulation models predict, and whether reflected light varies in a way that tracks cloud patterns across the dayside. Finally, multi-epoch monitoring would show whether the haze and the activity of the host star vary together, which is the cleanest way to separate planetary signals from stellar ones.

References and further reading

  • Bouchy, F., et al. (2005). ELODIE metallicity-biased search for transiting hot Jupiters II: A very hot Jupiter transiting the bright K star HD 189733. Astronomy & Astrophysics.
  • Knutson, H. A., et al. (2007). A map of the day-night contrast of the extrasolar planet HD 189733b. Nature.
  • Pont, F., et al. (2008). Detection of atmospheric haze on an extrasolar planet: the 0.55-1.05 micron transmission spectrum of HD 189733b with the Hubble Space Telescope. Monthly Notices of the Royal Astronomical Society.
  • Sing, D. K., et al. (2011). Hubble Space Telescope transmission spectroscopy of the exoplanet HD 189733b: high-altitude atmospheric haze in the optical and near-ultraviolet with STIS. Monthly Notices of the Royal Astronomical Society.
  • Evans, T. M., et al. (2013). The deep blue color of HD 189733b: albedo measurements with HST/STIS at visible wavelengths. The Astrophysical Journal Letters.
  • Louden, T., and Wheatley, P. J. (2015). Spatially resolved eastward winds and rotation of HD 189733b. The Astrophysical Journal Letters.
  • Seager, S., and Deming, D. (2010). Exoplanet atmospheres. Annual Review of Astronomy and Astrophysics.
#HD 189733 b #exoplanet #glass rain #silicate particles #hot Jupiter #Hubble Space Telescope #exoplanet atmosphere #alien weather #space discovery #astronomy
All Scientific Breakthroughs