A Green Bank Telescope search for narrowband technosignatures between 1.1-1.9 GHz during 12 Kepler planetary transits

Brief

Sheikh et al. (2022) used the GBT L-band receiver (1.1–1.9 GHz) on March 25, 2018 to observe 12 Kepler transiting exoplanet systems during their actual transit windows, the first SETI search designed around planetary transits as temporal Schelling points. The turboSETI pipeline, run with an unusually low S/N threshold of 5 and an unprecedented maximum drift rate of ±614.4 Hz/s, generated ~2.53 million hits, which 13 citizen scientists reduced to 338,473 events and ultimately 2 signals-of-interest. Both surviving signals occurred during their targets' transit midpoints but are non-narrowband and unconfirmed; if not redetected, the non-detection implies no transit-aligned transmitter in the sample exceeds 60× the former Arecibo radar's output.

Metadata

Category

Search

Venue

The Astronomical Journal

Type

Peer-reviewed

Year

2022

Authors

Sofia Z. Sheikh, Shubham Kanodia, Emily Lubar, William P. Bowman, Caleb I. Canas

DOI

10.3847/1538-3881/aca907

arXiv

2212.05137

Access

Open access

Length

2.5 M

Programs

Breakthrough Listen, Penn State Pulsar Search Collaboratory, Penn State Extraterrestrial Intelligence Center

Instruments

Robert C. Byrd Green Bank Telescope (GBT) L-band receiver (1.1–1.9 GHz), BL GPU spectrometer backend, turboSETI pipeline

Data sources

NASA Exoplanet Archive, ATNF Pulsar Database, Breakthrough Listen data archive (bldata.berkeley.edu/kepler-transits)

Tags

SETI, technosignature, radio SETI, narrowband, exoplanet transit, Schelling point, citizen science, astrobiology

Key points

• First SETI survey to pre-synchronize radio observations with exoplanet transit midpoints, treating the transit as a temporal Schelling point, a mutually derivable coordination epoch for transmitter and receiver.p.1
• turboSETI run at S/N threshold of 5 (vs. typical 10) and maximum drift rate ±614.4 Hz/s (vs. typical ±10 Hz/s), the first implementation of the Sheikh et al. (2019) ±200 nHz recommendation for exoplanet-hosted transmitters.p.5
• Pipeline generated ~2.53 million hits; after event-finding (off-on-off cadence across 52 scans), 338,473 unique events remained for vetting.p.6
• 13 citizen scientists reduced 338,473 events to ~1,600 flagged signals (~0.4%) via visual inspection of PDF waterfall plots, enabling sensitivity-preserving vetting at S/N 5 that a single researcher could not have performed.p.8
• Three GPS-linked frequency bands (1165–1185 MHz/L5, 1375–1385 MHz/L3, 1570–1580 MHz/L1) plus Iridium constellation signals account for 56% of all hits across only 5% of the usable band.p.7
• Final session-wide filtering reduced signals-of-interest to 2: a pulse in Kepler-1332b at 1749.4209 MHz and a broadband pulse in Kepler-842b spanning 1040–1438 MHz, both coinciding with their respective planets' transit midpoints.p.10
• Non-detection upper limit: if signals-of-interest are not redetected, the 12 targets produce no transit-aligned 1.1–1.9 GHz emission at transmitter powers exceeding 60× the former Arecibo radar.p.2
• Observing session covered March 25, 2018 (11:00 UT, 6 hours); each planet received 5-minute scans covering approximately 5% of its transit duration (range: 0.65–3.9 hours, median 1.8 hours).p.5

Verbatim

• “In this work, we used the L-band (1.1–1.9 GHz) receiver on the Robert C. Byrd Green Bank Telescope (GBT) to perform the first technosignature search pre-synchronized with exoplanet transits, covering 12 Kepler systems.”

  p.1

• “If the signals-of-interest are not re-detected in future work, it will imply that the 12 targets in the search are not producing transit-aligned signals from 1.1 – 1.9 GHz with transmitter powers > 60 times that of the former Arecibo radar.”

  p.2

• “This generated a total of ∼ 2.53 million hits.”

  p.6

• “5% of the band (excluding notch filter) accounts for 56% of the hits.”

  p.7

• “In the end, 13 citizen scientists (named in the Acknowledgements) flagged approximately 0.4% ( ∼ 1600 signals) of the dataset for further analysis, reducing the number of interesting events to a few thousand.”

  p.8

• “Interestingly, both the detection in Kepler-1332b (during scan 0031) and the detection in Kepler-842b (during scan 0056) were during their respective transit midpoints.”

  p.10

Most interesting

• The two surviving signals-of-interest both occurred at their targets' exact transit midpoints, the precise epoch the search was designed to prioritize, though neither has a narrowband morphology expected of an intentional ETI signal.
• Kepler-842b's signal-of-interest spanned 1040–1438 MHz (a ~400 MHz broadband pulse), far outside the narrowband morphology turboSETI was designed to detect; it was retained specifically to avoid excluding unexpected signal types.
• GPS satellites (L1, L3, L5 downlinks) and the Iridium constellation together saturated just 5% of the L-band but generated more than half of all 2.53 million detected hits, dominating the RFI environment.
• The ±614.4 Hz/s maximum drift rate, 60× higher than the SETI community norm of ~±10 Hz/s, was required because transiting exoplanets, viewed nearly edge-on, can produce the largest radial accelerations relative to Earth observers of any known planetary geometry.
• Lowering the S/N detection threshold from the standard 10 to 5 doubled sensitivity but inflated the initial hit count so severely that no single researcher could have visually inspected the output; citizen science was structurally necessary, not merely convenient.
• The GBT L-band notch filter blanks 1.20–1.34 GHz entirely, a 140 MHz gap in the water-hole region, because it is the most RFI-contaminated portion of the spectrum at the site.