iiinova — where your digital self keeps evolving

Hannah Wakeford @stellarplanet.bsky.social · 13h

@lili-alderson.bsky.social then took all of the measurements & modeled the spectrum with loads of methods to work out if there were trends in the property of the planet's atmosphere at 10 different angles around the limb 🤯 Previous methods have only ever done this for half a planet at a time!😱 8/n🔭🧪

The spectrum of the planet measured at 10 different angles around the limb compared to: top, forward models from the SANSAR RCTE grid which considers chemical equilibrium and a thermodynamically equilibrium planet; middle, models which use a single temperature but are in chemical equilibriums; and bottom, which is a fully free model where all values are tuned to the data. Each of these show a transition around the planet with the bottom models showing the clearest trend in changing temperature from angle to angle around the planet.

140

Hannah Wakeford @stellarplanet.bsky.social · 13h

We then implemented a statistical method used in Machine Learning & AI called k-fold cross-validation (statistical methods developed and implemented since the 1950's) on our shape models (our transmission strings) to robustly map any deviations from a circular planet (🌑 or 🪨). 7/n🔭🧪

Evidence of the k-fold cross-validation tests. A top plot shows the cross validation score as a function of wavelengths where the higher the score the better. For each score there is also a grey shaded box which shows the range of scores from different scaling parameters used for that wavelengths. A bottom plot along the same wavelength axis shows a color map of these scores relative to the maximum score and the scaling parameter needed to reach the best validation score. These show that the scaling is relatively consistent which correlates with how deviant from a circle the results are with the maximum score showing a range of values which correlate with the signal to noise of the wavelength bin.

example light curves from our measurements for each JWST instrument mode and spectral order with the residuals of the model fit to the data and the unfolded shape of the planet as a function of angle for all of the model draws we sampled in our analysis. This shows the light curves from NIRISS SOSS order 2 (top) and 1 (bottom).

140

Hannah Wakeford @stellarplanet.bsky.social · 13h

To ensure that our measurements of the shape of the planet were accurate we had to refine the planet 🪐 parameters for the period ⏱️, inclination of the system📐, and the distance of the planet from the star📏. This allowed us to model the transit as precisely as possible. 6/n🔭🧪

The measured centre of transit time from repeated event which is then used to measure an accurate period of the planet around the star. We combine past measurements of the system with our new JWST measurements to get the period of the planet's orbit around the star accurate to less than 0.05 seconds

Left is the measured distance the planet is from the star in units of stellar radii from each of our individual JWST measurements. We then combine these together to get an accurate weighted average of the parameter. Right shows the same but for the inclination of the system.

150

Hannah Wakeford @stellarplanet.bsky.social · 13h

Excitingly our new measurements, which show finer details around the atmosphere, match well to predictions for the planet as well as other studies our JWST-TST DREAMS program have produced such as, dayside maps🗺️ by @danvalentine.bsky.social, dayside 🌕 and nightside 🌑 temp, & some work in prep 5/n🔭🧪

illustration of WASP-17b's orbit around its star. The eclipse as it passes behind the star is zoomed in on the left showing the illuminated dayside of the planet, the shift in the hottest point of the planet from the sub-stellar point, the gradient in temperature across the dayside and the measured average temperature from two different studies. The right shows a zoom in on the transit which shows the limb of the atmosphere where the light shines through before reaching our telescope. The atmosphere is not circular but skewed asymmetrically to the left with temperatures measured around the rim of the planet from hot on the left to cooler on the right.

160

Hannah Wakeford @stellarplanet.bsky.social · 13h

The atmosphere of WASP-17b has a smooth transition in temperature, T, from hot 🔥 on the evening decreasing across the poles with the coldest temperature ☃️ on the morning. This is caused by vast winds🌬️ that blow air from the permanent day to the night where it cools before restarting the loop. 4/n🔭🧪

a series of histograms showing the changing temperature and chemical properties for each angle of the planet's atmosphere. There is a steady change in the peak of the temperature histogram with angle.

160

Hannah Wakeford @stellarplanet.bsky.social · 13h

This is the 1st application of this new mapping technique to real data & only combined with the exceptional precision of JWST 🛰️ can we extract the needed info. We map 🗺️, for the 1st time, the the changing shape of the 🪐 with radius & track the changes in shape relating to T, chemistry, & ☁️ 3/n🔭🧪

There are two semi-circles in the center with rectangular plots coming out radially at five angles from the top semi-circle. In each plot the radius of the planet as a function of wavelength is shown for that angle around the edge of the planet. On the left it goes from 180 degrees across the top pole of the plant at 90 degrees to 0 degrees on the right side of the planet. The semi-circle in the middle shows the measured amount of different gases found in the atmosphere at each angle. The bottom semi-circle shows the changing temperature of the atmosphere around the planet from hot on the left to cold on the right.

140

Hannah Wakeford @stellarplanet.bsky.social · 13h

In this paper we measure the shape of WASP-17b's atmosphere, a gas giant exoplanet orbiting in just 3.7 days around a star 1300 light years from Earth. We use a method called transmission strings 🧵 which traces the shape of the atmosphere as an angle 📐 around the limb for each color of light🌈 2/n🔭🧪

Two deformed circles on the left, the top more exaggerated than the bottom. Lines extend radially from the center to the edges (blue in the top red in the bottom). Middle shows the length of the lines against the angle the line extends to the edge of the deformed circle. On the right the red and blue points are plotted on the same graph showing that the two shapes are slightly different from each other as the length of the lines at a certain angle are different. In the bottom plot this is then flipped so that the x-axis shows blue vs red joined together. This represents the change in radius as a function of wavelength, or the spectrum of the planet.

160

Hannah Wakeford @stellarplanet.bsky.social · 13h

Planets are 3D! Shocking I know but when we measure an exoplanet we collapse the info into an average while assuming the planet is a perfect circle. UNTIL NOW! @lili-alderson.bsky.social presents: A Spatial Map of the Atmospheric Limb of a Hot Jupiter arxiv.org/abs/2610.18806 1/n🔭🧪

Artist impression of the hot Jupiter WASP-17b and its asymmetric atmosphere. The planet has a thick banded structure around the equator and the light is shining through from the back to show an extended left side of the atmosphere representing the evening terminator (limb) and a more compressed right side representing the morning terminator (limb).

Artist impression of the hot Jupiter WASP-17b and its asymmetric atmosphere as it transits in front of its host star. The planet has a thick banded structure around the equator and the light is shining through from the back to show an extended left side of the atmosphere representing the evening terminator (limb) and a more compressed right side representing the morning terminator (limb).

27024

Hannah Wakeford @stellarplanet.bsky.social · 04/09/2026

Excited to be at the @ioppublishing.bsky.social house in Bristol for a science writing workshop led by @andyextance.bsky.social for the Association of British Science Writers student group. #SciComm

Screen in an open plan room with the title. Science writing and interviews.

041

Hannah Wakeford @stellarplanet.bsky.social · 16/07/2026

Had a great time @bristoluni.bsky.social graduation today seeing the 2026 class of physicists graduate. My first time as staff at one of my students graduations (sorry Lili & Charlotte) & seeing Dr Matt Lodge = #ProudAdvisor. Also, lovely to wear my @exeter.ac.uk gown for the first time in a decade!

Me in my external gown. It has blue lines down down the front and silver and red hood

Me and Dr Matt Lodge he is in Bristol red and purple gown and me in my Exeter robe

A hot air balloon for Bristol Uni in the top of the great hall entrance.

080

Hannah Wakeford @stellarplanet.bsky.social · 21/06/2026

Got my nails done for all my trans friends and for all those feel unseen or too seen right now. Love you all, just be you! #pridemonth #transally

Nails painted light pink with bright blue French tips.

0150

Hannah Wakeford @stellarplanet.bsky.social · 02/06/2026

Finally, this work could not have been done without a big amazing team especially Katy, David, @offallingstars.bsky.social & Natasha. So, I wanted to make sure that the paper showed the life journey we've been on, so I introduce to you the HD209458b cloud babies in our acknowledgements! 🔭🧪 🧵15/n

Text reads HD209458b cloud babies: Ariah, Finn, Forest Moon.

080

Hannah Wakeford @stellarplanet.bsky.social · 02/06/2026

Each detection of a cloud species gives us a better idea of the chemical inventory of the planet - especially the O as it is locked up in silicates. Without JWST MIRI data clouds are unconstrained & our knowledge on the planet is fundamentally limited! We need more data for more planets 🔭🧪 🧵14/n

Meme showing cartoon shouting person with the words "Give me all the data"

Meme of Data from Star Trek TNG with the words "You seem to require more data"

Meme of Christopher Walken from SNL skit with the cowbell with the words "I got a fever and the only prescription is more data"

131

Hannah Wakeford @stellarplanet.bsky.social · 02/06/2026

HD209458b joins the small but growing number of silicate cloud detections in hot Jupiter atmospheres. Silicate species provide insight into atmospheric chemistry & possibly formation pathways with the balance of Mg/Si in the star dictating the availability in the planet forming environment. 🔭🧪 🧵13/n

$Figure from Calamari et al. (2024, 2026) showing the Mg/Si bulk fraction versus the fraction of the oxygen inventory for the planet. The Mg/Si ratio comes from the star with HD209458b sitting in the regime where it would be expected that both Mg2SiO4 and MgSiO3 could be present at the same time while other hot Jupiters with silicate cloud detections tend to sit in the region most dominated by SiO2 cloud formation.$

131

Hannah Wakeford @stellarplanet.bsky.social · 02/06/2026

We tested our clouds against our 4-🍃 model to show that it is robust to the choice made in the rule-based tree structure for data reduction & that a Gaussian mixture model is a useful way to incorporate unknowns into our retrieved planetary information & can be applied in many ways. 🔭🧪 🧵12/n

Spectrum of HD209458b measured from 5 to 12 microns with JWST MIRI. The plot shows wavelength versus transit depth. The data are shown for two different ways on selecting the "best" reduction pathway we call the 1-leaf and one combining the information from multiple valid reduction pathways in this case 4 paths to make the 4-leaf spectrum of the planet. We then use the same model to determine if they convey the same information and find that both models equally well fit the data.

101

Hannah Wakeford @stellarplanet.bsky.social · 02/06/2026

Our forward models are then used to show if this can be physically replicated without looking at the data. This allows us to ground the detection in known physics & chemistry using more realistic atmospheric temperature profile. This confirms our magnesium silicate clouds to be most likely 🔭🧪 🧵11/n

121

Hannah Wakeford @stellarplanet.bsky.social · 02/06/2026

HD209458b was the first exoplanet to have its atmosphere measured & there is extensive data in the optical - near-IR with Hubble. Joining this information with NIRCam & our JWST measurements we can constrain the properties of the clouds finding particle size ~0.1 μm at a pressure ~1–10 mbar 🔭🧪 🧵10/n

a plot of wavelength versus transit depth showing the full spectrum of HD209458b combining Hubble data from 0.3 to 1.6 microns, JWST NIRCam data from ~2 to 5 microns, and out new JWST MIRI data from 5 to 12 microns. The spectrum shows prominent absorption from H2O, CO2, with scattering towards the blue optical indicative of small particle scattering from clouds. The MIRI data shows the broad absorption due to magnesium silicate clouds. The right side shows the probability density retrieved for the pressure of the cloud and the radius of particles. A broad unconstrained distribution is shown is pale when just using the MIRI data alone, by including the short wavelength data the location and particle size of the cloud can be constrained.

111

Hannah Wakeford @stellarplanet.bsky.social · 02/06/2026

Our retrieval is used to explore the potential of the data with the forward model grounding the physical feasibility of the retrieved atmosphere. Our models show the 1st direct measurement of cloud species in HD209458b to be Mg2SiO4(forsterite) or a combination of Mg2SiO4+MgSiO3(enstatite)! 🔭🧪 🧵9/n

Two plots showing wavelength (x) versus transit depth (y). Left: The measured planet spectrum with three models one for a cloud free atmosphere which is shown to deviate from the data structure significantly, two which include clouds of Mg2SiO4 or Mg2SiO4+MgSiO3 which both equally fit the data well. Right: the contribution of different species to the modelled atmosphere with a cloud of Mg2SiO4 where at short wavelengths the model is dominated by H2O absorption and at longer wavelengths the model is dominated by the cloud absorption.

111

Hannah Wakeford @stellarplanet.bsky.social · 02/06/2026

We test a ton of different cloud forming species looking at their crystalline and amorphous forms to determine what fits the data best. We also considered a combination of cloud species mixing in a single cloud deck or forming two stratified layers in the atmosphere. 🔭🧪 🧵8/n

Wavelength versus transit depth figure. The planet spectrum is shown as black points with vertical errorbars. Overlayed is the expected spectrum of different silicate absorption signatures which peak around 9 microns but have varying structures, peak amplitudes, and extents. The best fit species is in bold and matches to Mg2SiO4 sol-gel amorphous cloud particles.

101

Hannah Wakeford @stellarplanet.bsky.social · 02/06/2026

Traditionally we would take a single 🍃 from our decisions but with our 4 🍃 we now wanted to combine the information content to show the marginalized result. We used a Gaussian mixture model* *In the appendix we show how the GMM works for an extreme case where the results are multi-modal 🔭🧪 🧵6/n

101

Hannah Wakeford @stellarplanet.bsky.social · 02/06/2026

These 432 results we called "leaves" 🍃representing the end point of a set of choices. We used the Weighted L2-norm, commonly used in stats & machine learning, to test branches against each other & tell us if & how it impacted the result: big diff, no diff, or 🤷‍♀️. We keep the 🤷‍♀️ leaving just 4🍃 🔭🧪 🧵5/n

the final rule-based tree-structure with uncle data at the top. The tree then splits in two with the number of groups to mask. Each branch then splits in two again with the amount of baseline to use. The four branch ends are related to the systematic model used such that long baselines require linear+exp fits while short baselines just require linear fits. This leaves four results, one at the end of each branch, that represent the potential spectra of the planet.

101

Hannah Wakeford @stellarplanet.bsky.social · 02/06/2026

When analyzing our data there are lots of choices we can make along the way, but there is no 1 way to do things. We decided to test these decisions by making a tree-structure where each branch represented rule-based choices made at that step in the analysis. Giving us 432 potential branches! 🔭🧪 🧵4/n

Diagram of the key steps considered in the MIRI/LRS Rule-based tree structure model for this dataset going from the uncal data to the final speci,j,k,.... At the nodes of the tree, branches are expanded for different reduction and analysis choices. Triple dots are used to indicate the continuation of the tree structure. The tree can be described as homogeneous or symmetric in that it can be thought of as a complete set of decisions with no dangling branches.

111

Hannah Wakeford @stellarplanet.bsky.social · 02/06/2026

We used JWST MIRI LRS instrument to get a spectrum of the canonical hot Jupiter HD209458b (Jupiter-sized gas giant on a close-in orbit of its star) from 5 - 12 microns. This ~1500K planet will not have water clouds like Earth but clouds made of rock🪨, most commonly expected to be silicates. 🔭🧪 🧵3/n

The high temperature cloud sequence. At around 700K high altitude clouds of hazes form with salts deeper and silicates deeper still. At 950K and above the hazes and salts evaporate and silicates become the top most cloud deck. At around 2000K the silicates evaporate and Aluminium oxides become the top most clouds in an atmosphere. From 950 - 2000K silicates dominate the cloud sequence.

111

Hannah Wakeford @stellarplanet.bsky.social · 02/06/2026

Clouds ☁️ exist in near-all planetary atmospheres but only with JWST mid-IR measurements can we identify what the they're made from & therefore how they impact the chemical & radiative properties of the planets. We measured the light absorbed by the cloud 🌤️⛅🌥️ to work out what it is made from 🔭🧪 🧵2/n

blue sky with high and low clouds. Thick yellow arrows coming from the top show the incoming solar rays. Thinner yellow arrows bounce off the cloud tops or from the bottom of the image depicting reflected starlight. Thick red arrows come from the bottom of the image showing IR radiation which is absorbed or reflected by the clouds. Lower clouds are more effective at reflecting back to the bottom more IR radiation and emitting more IR radiation to space.

131

Hannah Wakeford @stellarplanet.bsky.social · 23/04/2026

Mine repeats three letter but I appreciate how they used differ pictures for them.

160

Hannah Wakeford @stellarplanet.bsky.social · 18/12/2025

Got a wonderful gift from a former postdoc today! A terrarium based on The forest moon of Endor complete with its own little Ewok & R2D2 😁 The tank was previously used on BBC The Sky at Night to demonstrate optical depth in a planetary atmosphere. A happy repurposing & joyful addition to my office

Terrarium in a large rectangular fish tank. The moss and gravel is made to look like the forest moon of Endor from Star Wars and a small model of an Ewok and R2D2 are in the foreground

1291

Hannah Wakeford @stellarplanet.bsky.social · 16/12/2025

Doing some filming for a potential project this morning. I always get so nervous before filming things. The more prep for something there is the more nervous I become. Very much more comfortable with the hit and hope method. But nevertheless here we go!

Desk with spotlight lamp, microphone and camera set up for filming.

170

Hannah Wakeford @stellarplanet.bsky.social · 02/12/2025

At the wonderful venue of the Balmoral ship docked in Bristol Quay to preset to the Bristol Science Cafe. A great team of volunteers on the ship and the science cafes lead Bob who has been running this for over 2 decades bring science to Bristolian life. #exoplanets #scicomm

Large iron ship docked along Bristol Quay. The name Balmoral can be seen in yellow letters along the front

010

Hannah Wakeford @stellarplanet.bsky.social · 21/11/2025

Had a fabulous week at Ringberg for the “Signal in the Noise“ workshop hosted by two brilliant ECRs in the #exoplanet field @eahrer.bsky.social & @lili-alderson.bsky.social. Loads of ideas on how to get the most out of #jwst data & many new collaborations formed across an eager cohort of scientists

Ringberg castle in the snow with mountains in the background

Looking through the gateway arch of Ringberg castle at the main doors and a courtyard blanketed in a dusting of snow

0180

Hannah Wakeford @stellarplanet.bsky.social · 17/11/2025

We have snow in Ringberg! ❄️🏰❄️ the sky has fallen, wonderful rainout of the atmosphere followed by a clearing and lowered opacity.

snow falling over the forest viewed through a large two story double wide castle doorway

View over a forest and lake with the trees glistening with fresh snowfall

1180

Hannah Wakeford @stellarplanet.bsky.social · 16/11/2025

Beautiful lessons in opacity from normal versus slant geometries from the plane as I left Abu Dhabi, some beautiful cloud formations over the Iranian mountains, and beautiful sunset over Romania

Looking out the left side of an airplane with the jet engine and wing visible. The view to the ground is clear but the view to the near horizon steadily gets whiter and whiter due to the observation at slant geometry through thin scattering particles in the atmosphere

Epic clouds seen out a aeroplane window. A sun beam cuts from top to bottom. The clouds appear to form two cliffs with deep valley between them. Multiple layers of different cloud shapes can be seen. Formed due to tiny changes in the local conditions.

Sunset through the atmosphere viewed from an airplane. Deep blue at the top turns to lighter and lighter blue until almost white and then becomes a thin layer of orange blending into red. The sun is partially obscured by a dark cloud and the image fades to black at the bottom

0142

Hannah Wakeford @stellarplanet.bsky.social · 29/03/2025

Taught the next door neighbors kids about eclipses and they loved looking through the glasses! They were so excited ☀️😁 Space is the best!

050

Hannah Wakeford @stellarplanet.bsky.social · 04/03/2025

Hey bluesky there is beautiful blue sky above the fog today in Bristol 😁🌁

Looking out over Bristol skyline with low lying white fog and bright blue sky above

0191

Hannah Wakeford @stellarplanet.bsky.social · 05/02/2025

Feeling mildly like a traitor that my first time at Burlington House is to the Royal Society of Chemistry and not over the courtyard to Astrophysicists but found some planets 🪐 represented in here so all is good. Man I love old buildings with so much history 🤓

The view of the Royal Astronomical Society from across the courtyard in the Royal Society of Chemistry

Different colored glass orbs with apparent rings like the planet Saturn suspended on curly sticks in front of an old dark wood bookcase filled with chemistry texts

2160

Hannah Wakeford @stellarplanet.bsky.social · 14/12/2024

My little Pearl-z is in a cone of shame after getting an infected cat bit that resulted in a massive abscess and now hole in her leg last week. All was healing nicely until today she decided to use the cone to attack the wound and scrape off any scab that formed so we are back to square one!

Small black cat siting on a box. They have a small cone around their neck to prevent them licking any wounds

1160

Hannah Wakeford @stellarplanet.bsky.social · 14/03/2024

The main take away You cannot use IR data to easily predict the UV/optical spectrum of an #exoplanet predominantly due to the presence of clouds in the atmosphere which cause additional absorption that has less impact on long wavelength data. @NASAHubble is key!

040

Hannah Wakeford @stellarplanet.bsky.social · 14/03/2024

We looked at the results from our models as a population but find no clear linear trends in cloudiness of the atmosphere with temperature or amount of water vapour, however, more complex trends are likely. We do show that the model determined temperature is similar to GCMs!

Retrieved temperature of the planet's limb compared to GCMs from Kataria et al. (2016) showing that the limb does not equal the equilibrium temperature but is in line with predictions from complex 3D modelling.

120

Hannah Wakeford @stellarplanet.bsky.social · 14/03/2024

We show that the inclusion of data below 0.6 microns (@nasawebb-mirror.bsky.social lower limit) is important to constrain information on the clouds in the #exoplanets atmosphere and that near-IR data alone is a poor predictor of the UV/optical spectrum measurable by @NASAHubble - associated with☁️

Two exoplanet spectra with models fit by increasing the amount of data used by 0.1 micron increments. Demonstrating that near-IR data alone cannot predict the cloud properties accurately as the models do not fit the data.

110

Hannah Wakeford @stellarplanet.bsky.social · 14/03/2024

To assess changes to the measured parameters from one model fit to the next containing more data we compute the change in Information Content - a metric to define the increase in information you have with the new model. Higher number == more information gained with new model

Information content change for 13/14 planets in the study. Showing a gain in information when shorter wavelength data is included in the fit. The bottom panel shows the cloud parameters broken down into the four specific components.

110

Hannah Wakeford @stellarplanet.bsky.social · 14/03/2024

We took 14 #exoplanet spectra from the literature and used the open-source retrieval code POSEIDON (@distantworlds.space ) to ask the question: "what information do I gain as I add in more optical data?" to demonstrate the need for @NASAHubble UV/optical measurements

14 planetary transmission (absorption) spectra with three model fits. Orange - just considering data longword of 1.1 microns. Blue considering data long of 0.6 microns. Purple considering data long of 0.3 microns. With added data tighter constraints on the model can be seen.

130

Hannah Wakeford @stellarplanet.bsky.social · 16/10/2023

This work shows what an excellent team @HUBBLE_space and @ESA_Webb make for studying clouds in exoplanet atmospheres & the combination helped solidify the detection of quartz as our clouds in the atmosphere of this hot (1700K) giant thanks to the scattering & water Hubble found

120

Hannah Wakeford @stellarplanet.bsky.social · 16/10/2023

Another key aspect we were able to measure thanks to data from the Hubble Space Telescope (@lili_alderson), was that these quartz crystals are just a fraction of a micron across (0.01micron) - in fact 10,000x smaller than width of a hair!

110

Hannah Wakeford @stellarplanet.bsky.social · 16/10/2023

We didn't stop there! We also ran a suite of comprehensive retrievals (@MartianColonist, Elijah Mullens, Ana Glidden), which step through millions of potential combinations of models to match the spectrum. We found that clouds were vital to fit the spectrum by 3.5-sigma

110

Hannah Wakeford @stellarplanet.bsky.social · 16/10/2023

Here we show the breakdown of the PICASO+VIRGA model. Each color represents what the atmosphere would look like without a particular material. In yellow we show what we would have measured if there was no quartz clouds in WASP-17b's atmosphere

110

Hannah Wakeford @stellarplanet.bsky.social · 16/10/2023

We used models from PICASO+VIRGA (@NatashaBatalha) and ATMO (Jayesh Goyal) to fit the spectrum with & without quartz clouds. We tested thousands of models & found that those with quartz clouds were preferred over those without by 4.2 sigma (>3 is used for a significant detection)

110

Hannah Wakeford @stellarplanet.bsky.social · 16/10/2023

This result come from #JWST's Mid-InfraRed Instrument (MIRI) which measured the transit of WASP-17b over the course of nearly 10 hours. Two teams (led by David Grant and @NikoleKLewis) carried out a reduction of the data to get the spectrum in different chunks to check it

110

Hannah Wakeford @stellarplanet.bsky.social · 16/10/2023

These measurements were part of the JWST-TST DREAMS program where we are conducting a deep reconnaissance of exoplanet atmospheres using multi-instrument spectroscopy. In this program led by the amazing @NikoleKLewis we took a deep dive into WASP-17b with all #JWST can offer

130

Hannah Wakeford @stellarplanet.bsky.social · 16/10/2023

We expected clouds made of a mix of magnesium and silicates, like sand on our beaches, but this pure quartz may suggest they act as seeds for larger cloud particles to form onto. Clouds in the atm. of WASP-17b are the first in a cloud sequence we hope to examine with #JWST

140