Four Small Transiting-planet Candidate Signals around TOI-4405, TIC 101653309 and TIC 262843927 from TESS
Description
I report four small transiting-planet candidate signals on three K and M dwarfs observed by TESS. All four went through box least squares, light-curve consistency tests, difference-image localization and TRICERATOPS; the 14.52-day signal is less secure, because its sector depths disagree and it is not localized. As an additional check, each reappears in TESS sectors that its detection search did not use, at an ephemeris fixed beforehand. They are a 3.601-day, 1.35 Earth-radius third signal around the K6 dwarf TOI-4405 (false-positive probability 0.050 with a speckle contrast curve); a 0.703-day, 1.30 Earth-radius signal around the early-M dwarf TIC 101653309 at 60 pc; and 24.11-day (2.1 Earth radii) and 14.52-day (1.9 Earth radii) signals around the early-M dwarf TIC 262843927. All four appear in Eumemic's public search tables; the 14.52-day signal was already reported by Kunimoto et al. (2025). None is validated. I give predicted transit times for ground-based follow-up.
Notes
Preprint of a Research Note submitted to RNAAS (manuscript AAS82326). Research Notes are moderated, not peer reviewed, and this one is not yet published.
This version corrects the previous one.
Hold-out tests. The previous version gave these at the printed ephemerides, which were fitted with some of the hold-out sectors, so "at an ephemeris fixed beforehand" was not true of those numbers. They are now at the ephemerides and durations fitted before those sectors were used, with the QLP orbit-edge ramps documented by Eumemic masked, against 2,000 random ephemerides. TOI-4405: 344 +/- 74 ppm, z = 4.1 (was 318 +/- 66 ppm, z = 4.4; the duration goes from 2.2 to 1.76 hours). TIC 101653309: 599 +/- 122 ppm, z = 5.0 (was 601 +/- 121 ppm, z = 4.9). TIC 262843927 at 24.11 d: 794 +/- 216 ppm, z = 3.4 (was 1020 +/- 216 ppm, z = 4.4; the transit windows move about 31 minutes). At 14.52 d: 769 +/- 194 ppm, z = 3.8 (was 671 +/- 189 ppm, z = 3.6). All four stay above z = 3. The printed ephemerides stay, for predicting transits, and the table now says so.
Search context. Of the 17 survivors of the search, 14 stay under the orbit-edge mask, including TIC 101653309 and TIC 262843927. Nothing else changes.
AI agents wrote and ran the analysis code, performed the data retrieval, vetting and statistical tests under my direction, searched the literature, and drafted the text and bibliography. The GAI tools used were Claude Opus 5.5 (Anthropic) and OpenAI Codex. I reviewed the results at summary level and take responsibility for the content. This disclosure does not claim a line-by-line human verification of the analysis.