New Science from Old Data: Finding Debris Disks in the Hubble Archive

Elodie Choquet,1, for the ALICE team at STScI,

Major breakthroughs in our knowledge of exoplanet populations have occurred over the last 10 years. With more than 2300 exoplanets detected by the Kepler mission, we now know that our solar system is not unique and that planets are ubiquitous in our galaxy. Two out of three sun-like stars have a planet the size of Neptune (or smaller) within 0.75 AU (Dong & Zhu 2013), while every dwarf M star—much more numerous than stars like our Sun—hosts at least two planets within similar orbits (Dressing & Charbonneau 2015). Yet several major questions about extra-solar planets remain open such as characterizing their structures and compositions as well as investigating if some harbor life. Clearly we must also learn more about their formation mechanisms. Understanding formation processes requires completing our planet census by detecting and characterizing massive exoplanets on wider orbits. As well, imaging the debris and dust left over from their formation can let us study their birthplaces and analyze their core compositions.

Large-orbit exoplanets are very difficult to find. They have orbital periods too long to be efficiently and unambiguously detected by indirect methods, such as transit or radial velocity measurements. They are also typically too dim to detect with direct imaging, that is, sufficiently distant from the parent star and bright enough to be identified. Such detections thus far have been limited to only very massive and young systems still warm enough to be self-luminous in infrared wavelengths.

Astronomers have learned the hard way that that new methods are necessary to push sensitivity limits, and improve the detectability of these systems. Dedicated instruments equipped with starlight-suppression systems, and exquisite wavefront controls, are capable of dimming the star by a factor of 1 million. Also, we require optimizing observing strategies to monitor the variability of the star due to instrumental instabilities, and developing post-processing algorithms to subtract the residual starlight from the images and finally reveal faint exoplanets. Through these techniques, and with instruments such as the Gemini Planet Imager (GPI), we are now able to detect giant planets a few times more massive than Jupiter on ∼10 AU orbits around young, 20 Myr-old stars such as 51 Eridani b (Macintosh et al. 2015; see Figure 1). We can also study the atmospheric structure and composition of such objects using new methods.

Figure 1: Image and spectrum of 51 Eri b discovered by the GPI instrument (from Macintosh et al. 2015). The planet spectrum shows evidences for methane and water content. The planet’s low brightness and young age is such that it can be explained by two different formation models.

The Hubble Space Telescope was designed well before the advent of these new technologies, and its “first-generation” coronagraphs were not optimized for finding such faint circumstellar objects. Recently developed post-processing algorithms might be applied to Hubble’s archival data to improve starlight subtraction methods to reveal faint circumstellar objects. This was the goal of the recently completed Archival Legacy Investigations of Circumstellar Environments (ALICE) project, led by R. Soummer. As part of this program, we consistently re-analyzed the entire archive of the NICMOS near-infrared data, spanning a decade from 1997 to 2008, achieving new starlight subtractions 100 times more precise than previously achieved.

NICMOS had been used to search for exoplanets around stars in the solar neighborhood. While no exoplanets were imaged, five debris disks were successfully detected with NICMOS, revealing dust left over from planet formation on large orbits around these stars, and similar to our solar system’s Kuiper belt.

Using advanced post-processing algorithms and building on the wealth and diversity of the NICMOS archive to monitor the instrument’s slightest instabilities enabled us to subtract the residual starlight 100 times more precisely from NICMOS images than was previously achieved. Three exoplanets around the 50 Myr-old HR 8799, previously discovered with ground-based telescopes, were newly uncovered in the NICMOS data with this method. By looking further back into the 10 year-old NICMOS dataset and measuring the motion of the planets, these detections enabled the measurement of the planets’ orbital parameters with unprecedented precision (Soummer et al. 2011).

Using the ALICE program, several additional planet-candidates were detected around other nearby stars, and the team is now working to confirm these detections. The program also has found 15 debris disks that were previously undetected in the NICMOS archives—11 of them never seen before by any instruments in this wavelength regime (Figure 2). Thanks to additional observations in complementary wavelengths, we will be able to study their structure and composition, as well as studying how they may be affected by undetected planets on nearby orbits by analyzing their morphology (shape, asymmetries, dust extent). Moreover, these new detections brought the number of debris disks imaged in near-infrared wavelengths to 31, which now enables interesting comparative studies of their physical properties and composition with respect to the characteristics of their host stars (age and mass).

Figure 2: Nine out of the 15 debris disks detected by the ALICE project by re-analyzing old archival Hubble-NICMOS data with advanced post-processing algorithms (from Soummer et al. 2014 and Choquet et al. 2016). These disks are formed by cold dust grains leftover from planet formation on large orbits around the stars, like our Kuiper belt.

Acknowledgements: This project was supported by NASA through grants HST-AR-12652.01 (PI: R. Soummer) and HST-GO-11136.09-A (PI: D. Golimowski), and by STScI Director’s Discretionary Research funds, and was made possible by the Mikulski Archive for Space Telescopes (MAST) at STScI, which is operated by AURA under NASA contract NAS5-26555.


Choquet, E., et al. 2016, ApJ, 817, L2
Dong, S., & Zhu, Z. 2013, ApJ, 778, 53
Dressing, D., & Charbonneau, D. 2015, ApJ, 807, 45
Macintosh, B., et al. 2015, Science, 350, 64
Soummer, R., et al. 2011, ApJ, 741, 55
Soummer, R., et al. 2014, ApJ, 786, L23

1 Hubble Fellow, Jet Propulsion Laboratory, California Institute of Technology.