Two astronomers from Flagstaff’s Lowell Observatory are among the co-authors of a forthcoming study in the Monthly Notices of the Royal Astronomical Society that will announce the discovery of six exoplanets and an additional 13 planet candidates.

However, Gerard van Belle and Catherine Clark are not just contributing authors. They designed and built the instrument that the study used for the followup observations needed to confirm the exoplanets’ existence.

QWSSI

van Belle and Clark’s latest contribution to the hunt for exoplanets is the Quad-camera Wavefront-sensing Six-channel Speckle Interferometer, which replaced the observatory’s earlier Differential Speckle Survey Instrument.

Speckle imaging is a technique used to reduce the blurring effects of atmospheric turbulence in astrophotography. It involves taking numerous short exposures of a target object, so short, at 100 milliseconds or less, that the atmosphere has minimal time to shift during the photo. These exposures are then processed to recreate a high-quality original image.

Speckle interferometry, as performed by QWSSI and other similar instruments, involves incorporating additional data from the diffraction patterns produced by combining the speckle images into the reconstructed image of the target object, using Fourier analysis to cancel out noise and correct graininess.

Photographing stars through the atmosphere “is a lot like trying to look at a quarter at the bottom of a swimming pool after somebody’s just jumped in,” van Belle explained. “This QWSSI instrument helps us pull this apart … it lets us use the telescope at its full resolution.”

QWSSI is fitted to the 170-inch Lowell Discovery Telescope, the fifth largest telescope in the mainland U.S., which is located southeast of Flagstaff at Happy Jack. It can simultaneously image on six wavelengths, four visible and two near-infrared, and also incorporates a wavefront sensor to provide additional data about incoming light that can be used in post-processing. Its predecessor, the DSSI, could only image on two wavelengths.

“The design philosophy was optimized for an inexpensive, rapid build,” Clark and van Belle wrote in a description of QWSSI in the Bulletin of the American Astronomical Society in 2021. “Virtually all parts were commercial-off-theshelf, and custom parts were 3D printed.”

According to van Belle, design work took about six to nine months, and was greatly facilitated by using off-the-shelf parts, since they were able to build a virtual prototype using CAD files supplied by the manufacturers. Assembly required another six months.

“We were able to actually make a schedule and keep it, which was nothing short of remarkable,” van Belle commented on the build time.

Equally noteworthy is QWSSI’s low price of about $26,000. Van Belle pointed out that the main factor in keeping costs down was their decision to recycle the CCD detectors from the DSSI, which saved about $100,000. Since they successfully demonstrated QWSSI’s capabilities, van Belle said, they have received three additional grants to replace those detectors, add an extra set of detectors and upgrade the filters with purpose-built optics. The latter will cost $120,000, in contrast to the cost of the instrument’s offthe-shelf parts.

The assembled QWSSI instrument fills an optics box that is about two feet square and six inches thick. Testing began in summer 2020.

“It pretty much worked right out of the box when we bolted it on the telescope,” van Belle said with pleasure.

Clark took the lead role in assembling and properly aligning the instrument, and is also one of the primary users of QWSSI data, which requires a particular kind of mathematical expertise.

van Belle is a staff astronomer at Lowell Observatory who specializes in astronomical interferometry and planetary detection. He is also the chief scientist of the Navy Precision Optical Interferometer. Asteroid 25155 van Belle is named for him.

Clark, formerly his research assistant, received her doctorate from Northern Arizona University with a dissertation based on the development of QWSSI and is now a postdoctoral fellow at NASA’s Jet Propulsion Laboratory, where her research focuses on characterizing planets in multi-star systems.

Kepler Corrections

Launched in 2009, the Kepler space telescope was designed to detect exoplanets by measuring the drop in stellar brightness caused by planets transiting their parent stars. It observed 530,506 stars and detected 2,662 planets over nine years.

In 2013, two of the reaction wheels on the telescope failed, disrupting its aiming functions and reducing its light-gathering ability by greater than an order of magnitude. The lower precision of the data produced by this second phase of the Kepler mission required improved methods of analysis and additional followup observations to identify the signals of possible exoplanets within it.

In the TFAW survey, led by Daniel del Ser of the Royal Academy of Sciences and Arts of Barcelona, the Kepler light curve data is processed through a “series of pixel decorrelation and detrending algorithms” called EVEREST 2.0, then fed into “a novel waveletbased detrending and denoising algorithm” called TFAW that “delivers both better photometric precision and planet characterization than any detrending method applied to K2 light curves,” del Ser and his coauthors wrote in the MNRAS paper. The results from these algorithms are subjected to a further vetting process to eliminate false positives.

To ensure accurate analysis of the Kepler data by these methods, however, the research team found it necessary to update the background data on their target stars first to rule out stellar companions. This data was provided in part by QWSSI and the LDT, which supplied between “one thousand to several thousand speckle frames” per star, according to the study. Van Belle described this as doing “a lot of the cleanup work.”

In addition to the LDT in Flagstaff, the study made use of followup observations from the Pan-STARRS telescopes on Maui, the LAMOST telescope in northern China and the SOAR telescope in Chile.

Planetfall

The second phase of the TFAW survey, documented in the MNRAS paper, analyzed more than 300,000 Kepler light curves and identified 27 candidate planets in 24 star systems. The team statistically confirmed the existence of six planets and rejected eight candidates as false positives; confirming the existence of the remaining 13 candidates will require additional followup observations.

“Our sample of validated and candidate planets is comprised of three sub-Earth planets, seven Earths, four super-Earths, and four sub-Neptunes,” the authors summarized. Most of these planets have an orbital period in the range of three to 10 days, indicating that all are well inside their stars’ habitable zones. Eight of the validated and candidate planets have radii less than 1.5 times Earth’s radius, which the team suggests “points towards the improved detection of smaller planets by the combination of the TFAW corrected light curves and TLS.”

The confirmed planets are:

• EPIC 210768568.01: 2.34 Earth masses, 3.2-day orbital period, 965 lightyears from Earth
• EPIC 247422570.01: 5.58 Earth masses, 5.9-day orbital period, 2,181 lightyears from Earth
• EPIC 246078343.01 & EPIC 246078343.02: 825 light-years from Earth. The inner planet is 0.36 Earth masses and has a 0.8-day orbital period, making it one of the very few planets discovered with a period shorter than one Earth day. The outer planet is 1.83 Earth masses and has a 5.33-day orbital period.
• EPIC 246220667.01 & EPIC 246220667.02: 835 light-years from Earth. The inner planet is 1.22 Earth masses and has a 4.4-day orbital period. The outer planet is 5.03 Earth masses and has a 6.7-day orbital period.

The study’s unconfirmed candidate planets include EPIC 247560727.02, an outer planet of a multi-planet system 2,220 light-years distant, which has a mass of 6.78 Earth masses, orbits its star in 8.4 days and, given its density, may be a water world.

Looking to Tomorrow

In addition to providing data for the TFAW study, QWSSI has also been used to conduct a survey of the Solar System’s immediate neighborhood out to roughly 50 light-years to determine how many of the stars in this area are binaries. Van Belle stated that they examined approximately 1,200 stars during this reconnaissance and found about 35 new companion stars that were too faint to be detected by previous instruments. These new discoveries raised the stellar multiplicity rate by about 10%, which has significant implications for hunting exoplanets, as binary stars are unlikely to retain a planet in a stable orbit.

“We can discover stuff that nobody’s ever seen before,” van Belle said.

QWSSI is expected to continue the work of finding exoplanets and characterizing binary stars in tandem with the Yale Exoplanet Lab’s EXPRES high-resolution spectrometer, another of the LDT’s instruments. A spaceflight-rated version of QWSSI is also under development.

“It’s been a lot of fun,” van Belle said. “We were pretty successful in pulling this off.”

The preprint version of the MNRAS paper is available at arxiv.org/pdf/2210.10805.pdf