Monthly Archives: June 2023

Four new planetary systems from TESS & CHEOPS

Today we announce four new planetary systems which were discovered via a combination of TESS and CHEOPS photometry through a project I started on the CHEOPS GTO (see our ESA and Swiss press releases).

The problem with TESS

TESS is surveying the whole sky to find new transiting planets around bright stars, and it has done a fantastic job in the past few years – the vast majority of characterisable planets around stars brighter than V=9 come from TESS. However, because its observing strategy involves observing much of the sky for only 27 days every two years, planets on periods longer than around 20 days are unlikely to produce consecutive transits. Instead, what we have seen with TESS, is that many planets produce single transits during TESS observations with large (often 700+ day) gaps between transits. This is hugely problematic as it means the orbital period is still poorly constrained and follow-up characterisation efforts – for example constraining planet mass through radial velocities or atmospheric composition through transmission spectroscopy – are impossible.

This is where ESA’s CHEOPS satellite can step in. When two transits are observed this constrains the orbital period to a range of periods, usually 30-40. By modelling those transits and extracting as much information as possible from the star, transit shape, and planetary system (using my MonoTools code), this can be further reduced to a handful of possible so-called period aliases – we know that one of them must be the true period, but not which.

CHEOPS to the rescue

So how can we recover these lost planets? By using ESA’s CHEOPS satellite. This 30cm ESA telescope is far more precise than TESS, but targets specific stars. But we can target these interesting stars with long-period planets during the times when we expect the transit of one of the possible periods. We chose a selection of candidate planets best-suited to CHEOPS – planets producing shallow transits impossible to spot in ground-based telescopes (i.e. planets smaller than Neptune) orbiting bright stars.

This began a game of hide and seek between the planets and CHEOPS – we monitor the stars and check the data hoping to see a transit which would confirm the planet and its period… but more often than not we would see a flat line, ruling out that orbital period, allowing CHEOPS to move onto another one.

The four systems:

HD 15906 c

This star, much cooler than the Sun, was seen to have a single transiting planet on a 10d orbit in TESS. However, looking at the data, I spotted the sign of a second transiting planet producing two transits. We hunted the true period, and revealed an outer planet on a 21d orbit. This paper was led by Amy Tuson and published in MNRAS.

HD 22946 d

This bright sunlike star was already confirmed to host two transiting short-period planets in a previous paper. They also identified a single transit of an outer planet, but once again we spotted a second transit in the data, and went after it with CHEOPS to find the true period. We were able to confirm it as 47d for the outer planet d, as well as improving the radii of the inner planets (which due to a bug had incorrect values in the previous paper). This paper was led by Zoltan Garai and published in A & A.

TOI-5678 b

In this case, we had been successful with previous systems and therefore decided to try the same technique with slightly larger planets but using less CHEOPS data – hoping to rely on only a fraction of the transit to confirm the planetary period. So we scheduled short observations at the aliases for the Neptune-sized planet TOI-5678, and managed to catch the ingress – the first half of the transit – confirming a 48 day period for this planet. Solene Ulmer Moll, who led the paper, was also able to observe the planet with high-resolution HARPS spectra, with the radial velocities providing a mass. The planet was published in A & A.

HIP 9618 c

For this bright star, TESS initially spotted two transits which were thought to be from an 11d planet, but those transits were actually from two different planets. The inner planet transited again three times in two later TESS sectors, revealing it to have a 21d period, but the outer (and slightly smaller) planet transited only once more. As for the other three systems, we quickly turned to CHEOPS to seek the true period, finding a beautiful transit on the fourth observation. This confirmed a 52.5d period. HARPS-N, SOPHIE and CAFE radial velocities also constrained the masses for these planets, showing them to be low-density mini-Neptunes about ten times the mass of Earth and 3-3.5 times the radius. The RVs also revealed that the system has an outer brown-dwarf companion in the outer reaches of the system. I lead this paper in MNRAS.

Uploading an overleaf paper to arXiv

It’s not always obvious exactly how to proceed, so here’s a quick guide.

1) Fix compilation problems on overleaf

Even though overleaf might still be able to compile a pdf with a few orange/red warnings, arxiv definitely will not. So click on “logs and output files” and go through the bugs & warnings one-by-one fixing them.

2) Download the source

You’ll need a .zip file for arxiv

3) Recompile the source on your computer

Unfortunately that .zip file isn’t usually enough (especially if, like most of us, you use bibtex). So you’ll have to unzip the and recompile the source. Of course, in order to compile the paper, you will need to have latex locally running on your computer, so if you dont (i.e. if running “pdflatex” in the command line give you nothing) then download latex from here

If you now have latex, get the command line up, cd into the (unzipped) source folder, and now run pdflatex [your .tex file], followed by the bibtex [your .bib file], and once again pdflatex [your .tex file]. Verify that the pdf that is generated looks good and has the correct references, and that a .bbl file has been created. If so, you need to re-zip the folder.

4) Upload to arXiv

Now you can start a submission to arxiv. You’ll need to choose the license – I haven’t read much into this, but I usually choose arxiv’s own license for this. Then upload the source and hope that nothing breaks. For the abstract you can copy it from overleaf (most peoples computers should compile any latex maths in the abstract) but make sure to remove any macros and comments.

For the comments, it’s typical to put the paper status (is it submitted, accepted, etc?), the number of tables and figures, and any other accompanying links/data/etc.

Some extra points:

  • Do not upload the proof-corrected version. The journal you submitted to has rights to that – you only have rights to the pre-corrected (i.e. preprint) version.
  • Go through the paper comments and remove anything… silly. Otherwise the overheard on astroph bot might find it.