HPF discovers a Jupiter transiting the cool star TOI-1899

HPF has confirmed the planetary nature of a single-transiting Jupiter-sized planet (1.15 Jupiter radii) orbiting a nearby low mass M dwarf star on an orbital period of ~29 days. The paper has been published in the Astronomical Journal and is available on arXiv.

TESS observes a single transit

NASA’s Transiting Exoplanet Survey Satellite (TESS) mission conducted a 2-year (2018-2020) all-sky survey with the goal of identifying exoplanets transiting bright, nearby stars. TESS is currently on an extended mission where it will continue its search for exoplanets.  The video below contains additional information on TESS and its survey strategy:

Due to the nature of TESS’s observing strategy and design, it will search for exoplanets in 85% of the entire sky. However, 74% of this total sky coverage will only be observed for almost 1 month. Consequently, TESS will find single events which may be caused by genuine transiting exoplanets that have no detectable period. An analyses of the TESS yield, predicts that a few hundred single transit events will be detected by TESS, with an estimated 80 of these signals originating from planets located in the habitable zone at distances from the host star where water, if present, could exist in liquid form. In addition to potential habitability, single transit events of long-period planets present an opportunity to investigate a number of questions related to planet formation and migration.

TOI-1899 is a prime example of a single transiting system. While the transit was observed by TESS, it was only detected as a candidate due to efforts from the citizen science project Planet Hunters TESS (PHT). The traditional TESS pipeline searches light curves for candidate transit events with a low background noise and multiple events associated with a given orbital period. TOI-1899 b only crossed its host star once while TESS was watching, which means it was never formally detected by the TESS pipeline. It was not given the designation of a planet candidate until PHT notified the community. Finding these types of single events is an example of how citizen science can improve the scientific output of TESS. You can learn more about citizen science on the PHT blog.


Figure 1. The single 5-hour transit event for TOI-1899 was observed by NASA’s TESS mission in August. A single event only reveals the size of planet, and a detailed characterization of TOI-1899 required data from the HPF. Image credit: Caleb Cañas.

For our work, we separately detected the event with a custom pipeline designed to process TESS photometry of cool stars known as M dwarfs. TOI-1899 is an M dwarf and it is smaller (0.6 Solar radii) and cooler than the Sun (3900 Kelvin). The large transit depth (4%) suggested a Jupiter-sized companion, but it was impossible to derive a precise period from a single event. Following the detection, we set out to confirm the planetary nature of TOI-1899 b. As we have previously discussed on blog posts for G 9-40 b and TOI-1728 b, transforming a planet candidate into a full-fledged, well-characterized exoplanet requires significant observing from the ground. This is even more true when we lack any kind of constraint on the planet’s period!

Validation with no period

Statistical Validation

Additional ground-based observations of the transit were not possible without a precise period. We instead decided to statistically validate this planet. Statistical validation uses the observed transit and knowledge of the period and nearby stellar neighbors as constraints in the analysis. It is a tool to ensure your candidate is likely a planet without exhausting limited photometric and spectroscopic resources on every candidate planet. As with the Kepler mission, false positives in TESS can originate from systems that lack planets such as smaller, dimmer stars orbiting larger, brighter companion stars. Statistical validation uses simulations to (i) compare the observed transit to other confirmed signatures of simulated exoplanets and impostors, and (ii) determine if the observed transit is plausible in the Milky Way and its statistical probability of being a planet.

False Positives

Figure 2. Above are some example false positive scenarios that could masquerade as an exoplanet. These scenarios include transiting low-mass stars and brown dwarfs (upper right), a background multi-star system that are on-sky neighbors to the candidate host star (lower left), or a grazing stellar binary (lower right). Statistical validation is a tool that can determine how likely a given signal is due to a genuine planet (upper left) or some impostor system. Image credit: NASA Ames / W. Stenzel

We observed TOI-1899 with adaptive optics imaging from ShaneAO and high-contrast imaging from NESSI to determine if there were any on-sky bright, stellar companions. To approximate the period, we used the transit duration and the stellar density to get an approximate period of 30 days. TOI-1899 b had a >99% chance of being a planet which gave us confidence that high precision spectroscopic observations would not be wasted observing a multiple star system.

Verifying the Host Star

Given the very high precision of the space-based photometry, the ability to track the center of the photometric aperture is one method of identifying background systems which can masquerade as a planetary transit. The dimming of any object through an event, such as a transit, will shift the measured on-sky center of the light source. The apparent change in the position of the target star due to contaminating systems is dependent on the separation of the stars, their relative brightness, and the transit depth. We used the Discovery and Vetting of Exoplanets (DAVE)  software package for our centroid analysis. During transit, the lack of a systematic shift away from the expect host star, TOI-1899, suggested it was indeed the host and gave us confidence that we identified the correct target when planning subsequent observations.

Confirmation and Characterization

Once we verified TOI-1899 was the most probable host of the transit and the companion was most likely a planet, we began observing TOI-1899 as part of our M dwarf TESS follow up program with HPF. The goal was to determine the orbital period and the planetary mass via the star’s Doppler shift. After a 2-month span of observations for TOI-1899, we used the Doppler measurements to determine TOI-1899 b is a planet that is two-thirds the mass of Jupiter, 1.15 times larger in radius, and has an orbital period of approximately 29 days.

HPF Data

Figure 3. The HPF radial velocity curve after a 57 day span of observations. While TESS could only observe one transit, the planet is large enough to induce a wobble on the star that can be detected with HPF. The periodicity of the above curve is almost 29 days and corresponds to the orbital period of TOI-1899. The best fit is from our modeling of the data is plotted as a dashed line. Image credit: Caleb Cañas.

From previous exoplanetary surveys, it was evident that close-in Jupiter-sized planets are rare: almost 1% of Sun-like stars host a close Jupiter-sized exoplanet and the occurrence rate falls off around cool stars like TOI-1899. Before the characterization of TOI-1899 b, there were only 4 other known Jupiter-sized objects transiting cool stars: Kepler-45 b, HATS-6 b, NGTS-1 b, and HATS-71 b, all with orbital periods <4 days. HPF data were crucial in identifying TOI-1899 b as the first warm Jupiter (WJ) transiting a cool star.

Potential for Additional Characterization

Atmospheric characterization of TOI-1899 would help us understand the properties of these cool gas giants. TOI-1899 also appears to be slightly inflated when compared to planetary models of gas giants with little heating from the host star. The larger radius in comparison to the known population and existing planetary models may be the result of a planetary composition low in heavy elements, or some atmospheric property could distort heat distribution and result in a larger size. Future missions, such as the James Webb Space Telescope, will have the precision and wavelength coverage to detect atmospheric signatures from TOI-1899 to begin probing its atmospheric properties and understand the apparent discrepancy in size.

Compared to WJs

Figure 4. TOI-1899 (star) is compared to other characterized WJs. It is a larger than other WJs of comparable equilibrium temperature. The planetary models include a core-free model, which would suggest a solid-poor composition, or a model assuming a gaseous envelope with solar metallicity. In either case, TOI-1899 is has a median radius that is larger than predicted. Image credit: Caleb Cañas.

TOI-1899 also presents an opportunity to probe the formation of WJs. There are a lot of questions and theories on the formation of both WJs and hot Jupiters (orbital periods < 10 days): Do these objects form exactly where we observe them? Have they migrated due to interactions with the planetary disk? Were they on elliptical orbits and migrated due to tidal interactions with the star? The measurement of the apparent obliquity, or the angle between the star’s equator and the planetary orbit, through the Rossiter-McLaughlin (RM) effect could provide insight as to how TOI-1899 b formed. A direct measurement of the obliquity would limit the physical processes involved during formation, as some mechanisms, such as disk migration, prohibit large obliquity.

We note both atmospheric characterization and a measurement of obliquity require observations during transit. The first step for any additional characterization of TOI-1899 is to determine a precise orbital period. We have continued our photometric and spectroscopic observations of TOI-1899 to obtain a more precise mid-transit time with additional HPF observations and, hopefully, another transit. This will allow us to plan more for subsequent observations during transit to further characterize this cool planet transiting a cool star.

Posted in Exoplanet Science, HPF Science | Leave a comment

HPF discovers a warm super Neptune – TOI-1728b

The Discovery

A component of HPF’s on-sky time is used to follow up on potential planets discovered by NASA’s Transiting Exoplanet Survey Satellite (TESS), launched in 2018. TESS is nominally a 2-year mission designed to conduct an all-sky survey to find transiting exoplanets. It spends one year in each celestial hemisphere, which is divided into 12-13 sectors that TESS monitors for 27 days each.  You can see more about TESS and its survey strategy in the video below.

One of the planet candidates discovered by TESS in its second year was TOI-1728b from Sector 20. Orbiting an “early M” star (it has a surface temperature of around 4,000 Kelvin; the Sun’s surface temperature is approximately 6,000 Kelvin), the depth of the transit represented a planet larger than Neptune, with a radius 5 times larger than Earth’s.  The planet completes an orbit around its star every 3.5 days.

As we have recently discussed on this blog, transforming a planet candidate into a full-fledged, well-characterized exoplanet takes a lot of follow-up observing from the ground.  Here, we’ll show you how we confirmed the discovery of TOI-1728b, and measured its mass with HPF.

Transit Follow up

We followed up this object with the 0.4 m PlaneWave telescope located on the roof of Davey Lab at University Park, Penn State in order to catch more transits of the planet.


PSU 0.4 m on Davey lab. Credit: Gudmundur Stefansson

This transit follow up helped us to confirm the TESS transit, and obtain an independent measurement unaffected by spurious background stars which can affect the TESS detection due to its larger pixel size. To insulate us from potential bad weather (as can often happen at State College), we also followed up on this target from the 17″ Perkin telescope at Hobart and William Smith Colleges in Geneva, NY, to obtain a similar result. Shown below is what the transit shape looks like from the three different instruments –


Transit light curves obtained for TOI-1728 from TESS, Davey 0.4 m, and 17″ Perkin telescope.

HPF follow up

As we worked to catch more transits, we also started observing this target as part of our M dwarf TESS follow up program with HPF, to measure the planet’s mass via the star’s Doppler shifts.  For a transiting planet, it is especially useful to use Doppler measurements to estimate its mass, as that information can be combined with the radius (which we got from the transit) to estimate the planet’s density.  The density is the first hint at a planet’s bulk composition: that is, is it rocky like Earth, or dominated by a gaseous atmosphere?  As an example, consider two planets of the Solar system.  Earth, which is composed mostly of rock and metal, has a bulk density of about 5.5 grams per cubic centimeter.   The gas giant Jupiter, on the other hand, has a bulk density of just 1.3 grams per cubic centimeter!  To compare the densities of exoplanets, astronomers typically use the mass-radius diagram, which allows us to examine how planets’ compositions vary with size, temperature, and other physical properties.

TOI-1728b falls in a region of Mass-Radius (M-R) space for M dwarf planets that is occupied by two other Neptune-sized planets:  GJ 3470b and GJ 436b. Using our HPF Doppler measurements, we measure the mass of the planet to be around 26 Earth masses, thereby being more massive than Neptune, which is only 17 times as massive as Earth.


Mass-Radius plane for M dwarf planets showing TOI-1728b. TOI-1728b is both more massive and larger than the other two Neptunes orbiting M dwarfs (GJ 3470b and GJ 436b).

Searching for signs of atmospheric escape

GJ 3470b was studied by an HPF team led by Joe Ninan, where we detected helium escaping from the planet’s atmosphere. Given the similarities between these exo-Neptunes, we decided to search for signs of atmospheric escape in TOI-1728b using the same technique.

We did not find any evidence for helium escape in the planetary atmosphere, and placed an upper limit of 1.1% for helium escape measured in the near-infrared atomic absorption line. For comparison, our detection in GJ 3470b was at 1.5%.  The lack of detection is interesting for the following reasons: as can be seen in the comparison chart below, TOI-1728b receives slightly more energy (insolation) from its star than the other two comparable planets, and at the same time, has a lower  bulk density. Both of these factors favour atmospheric escape, potentially making the detection easier. This lack of detection could be due to differences in the composition of the planet, as well as a relative lack of high-energy radiation from the star compared to GJ 3470b.

Comparing the planetary radius and insolation for M dwarf planets. TOI-1728b orbits an earlier type (hotter) host star, and therefore receives more insolation than the other sub-Jovian (smaller than Jupiter) M dwarf planets.

Conversely, TOI-1728b could be similar to GJ 436b, where the CARMENES team also report no detection of helium atmospheric escape using similar methods. GJ 436b does have a detection of hydrogen escape as seen from an ultraviolet absorption feature with the Hubble Space Telescope (HST).  Therefore, TOI-1728b represents a good opportunity for further follow up observations with HST to potentially detect hydrogen atmospheric escape. If HST also shows no hydrogen escape, that information would be very informative for understanding how stars dislodge atoms from their planets’ atmospheres.

Potential for transmission spectroscopy

While our HPF observations immediately place constraints on any gas escaping from TOI-1728b, they also set the stage for future measurements of the planet’s real atmosphere with HST and the upcoming NASA flagship James Webb Space Telescope (JWST).  TOI-1728 is bright enough to facilitate detailed observations with space telescopes, and now our HPF mass measurement indicates the planet’s density is low enough that plenty of starlight should be filtering through its atmosphere to create a detectable signal.  Hopefully, we are just starting to learn what this exciting planet is made of!

Posted in Exoplanet Science, HPF Science | Leave a comment

HPF Team Member Gudmundur Stefansson Runner-Up in IAU Dissertation Prize

We are thrilled to announce that HPF Team Member Dr. Guðmundur Stefánsson was recently recognized as one of two runners-up for the International Astronomical Union’s (IAU) Division B (Facilities, Technology, Data Science) PhD Prize to Recognize Excellence in Astrophysics.  This is an extremely competitive award, which is open to astronomers all over the world.

Stefansson Portrait

HPF Team Member Guðmundur Stefánsson

Dr. Stefánsson’s PhD dissertation represents a culmination of all the research he performed as a graduate student at Penn State, and includes his contributions to the HPF instrument’s development such as its vacuum chamber construction and outfitting, and its unprecedented environmental control.

Dr. Stefánsson graduated with a PhD in Astronomy & Astrophysics from Penn State in 2019 and is now a Henry Norris Russell postdoctoral fellow at Princeton University. The entire HPF team congratulates Dr. Stefánsson on this well-deserved recognition.

Posted in HPF Hardware, HPF Science | Leave a comment