The Breakthrough Listen Search for Intelligent Life: Searching for Technosignatures in Observations of TESS Targets of Interest

Brief

Traas et al. (2021) observed 113 30-minute GBT cadences of 28 stars from the TESS Objects of Interest catalog across all four standard receiver bands (L, S, C, X; 1–11 GHz), processing 24 TB of 2.7 Hz-channelized data via the turboSETI Doppler-drift pipeline running on 20 Google Cloud Platform instances. The pipeline yielded 249,532 hits with S/N > 10; after ON/OFF cadence filtering, 651 candidates survived automated selection and were each rejected as RFI on visual inspection, leaving zero confirmed technosignature candidates. The study establishes a one-sided Poisson upper limit of 12.72% on the fraction of observed targets hosting narrowband transmitters at or above the 4.9 × 10^14 W EIRP threshold, over a wider simultaneous frequency band than any prior BL Doppler survey.

Metadata

Category

Search

Venue

The Astronomical Journal

Type

Peer-reviewed

Year

2021

Authors

Raffy Traas, Steve Croft, Vishal Gajjar, Howard Isaacson, Matt Lebofsky, David H. E. MacMahon, Karen Perez, Danny C. Price, Sofia Sheikh, Andrew P. V. Siemion, Shane Smith, Jamie Drew, S. Pete Worden

arXiv

2101.11137

Access

Open access

Length

3.8 M

Programs

Breakthrough Listen

Instruments

Green Bank Telescope (GBT), turboSETI pipeline

Data sources

TESS Input Catalog (TIC), TESS Objects of Interest (TOI) catalog, ExoFOP-TESS, Breakthrough Listen Open Data Archive

Tags

SETI, technosignature, radio SETI, exoplanets, narrowband search, Doppler drift

Key points

• 113 cadence observations (30 min each) of 28 TESS Objects of Interest were collected at L-, S-, C-, and X-band (1–11 GHz) with the 100-m GBT, totaling 58 hours on sky and 24 TB of fine-frequency-resolution data.p.1
• turboSETI at 2.7 Hz channelization identified 249,532 hits (S/N > 10, drift rate ±4 Hz/s); these collapsed to 43,275 events and 651 ON-source candidates, all rejected as RFI.p.4
• Zero technosignature candidates survived full analysis. All 20 single-source event groups were traceable to known RFI, including navigation satellite bands and the Iridium satellite band.p.7
• EIRP detection threshold is 4.9 × 10^14 W; the 12.72% transmitter rate limit is computed via one-sided Poisson confidence interval assuming a conservative 50% detection probability if a signal is present.p.7
• TESS transit geometry confers a theoretical search advantage: a detected transiting planet means Earth lies in that system's orbital ecliptic, raising geometric probability of intercepting both intentional beacons and leakage radiation directed along the ecliptic.p.3
• This search spans a wider simultaneous frequency band (1–11 GHz) than any prior Breakthrough Listen Doppler-drift survey.p.3
• L-band accounts for 76% of total hits but 97% of all 651 candidates, indicating dense RFI in the 1.1–1.9 GHz window; X-band produced zero candidates.p.4
• The turboSETI pipeline was validated against the Voyager 1 spacecraft signal as a blind-search proxy technosignature, confirming sensitivity to real extraterrestrial transmitters before applying it to the TESS sample.p.4

Verbatim

• “turboSETI successfully finds the Voyager 1 spacecraft (an "extraterrestrial" transmitter that is a good stand-in for a real technosignature) in a blind search”

  p.4

• “Our observations spanning 1 – 11 GHz were searched for narrowband signals exhibiting drift rates within ± 4 Hz s − 1 above a minimum S/N threshold of 10, and no candidate signals unattributable to RFI were found.”

  p.7

• “By visual inspection, it is easily verifiable that there are emissions of S/N < 10 present in 'OFF' observations–characteristic of RFI. In this way, each of the 651 candidate events were rejected.”

  p.12

Most interesting

• The TESS ecliptic geometry argument cuts both ways: any system where TESS detects a transit is one where the host civilization, if it exists, also sees Earth transiting its star, making reciprocal beacon transmission along the ecliptic the rational strategy for both parties.
• This was the first Breakthrough Listen Doppler search to run on Google Cloud Platform rather than UC Berkeley's own compute cluster, using 20 Docker-containerized turboSETI instances operating in parallel on a cloud storage bucket.
• The 28 target stars span distances from 37 pc (TIC359496368) to 880 pc (TIC294176967), a factor of roughly 24 in distance, implying a factor of ~576 difference in the EIRP sensitivity floor across the sample.
• L-band produced 97% of all 651 surviving candidates despite accounting for only 76% of total hits; the disproportionate candidate count reflects heavy navigation satellite contamination in the 1.1–1.9 GHz band, not increased astrophysical signal.
• GBT pointings incidentally captured background galaxies in the field of view ('bycatch'), which the authors note could in principle be used to constrain transmitter prevalence at cosmological distances beyond the 28 primary TESS targets.
• The ±4 Hz/s drift-rate search window is tuned to the maximum expected Doppler acceleration for a planet in an Earth-like orbit; signals with higher drift rates or those spread across multiple frequency channels fall below turboSETI's nominal detection efficiency.