TESS sectors 1-107 show no new transit of 659 RV planets in their RV-predicted windows; the same search finds 24 of 28 known transiting RV planets; the HD 44219 eclipses come from a neighbour star

We searched the TESS light curves (sectors 1 to 107; SPOC, TESS-SPOC and QLP) for transits of 795 planets that radial-velocity (RV) surveys found and that have no known transit. For each planet, 2000 Monte Carlo samples of the archive orbit give the transit windows. The search used only these windows, with one box-minus-flanks statistic and one set of vetting tests for all planets. A sign-inverted (null) search of the same windows set the detection thresholds (S* = 7.8 for single events, 7.0 for multiple events). 897 injected transits gave the detection efficiency. TESS data cover the predicted windows of 659 planets. The search found no transit candidate: the vetting tests rejected all 68 signals above the threshold. For each planet we give the detection probability P_det and the posterior transit probability. P_det is more than 0.9 for 164 planets, and for 181 planets the search detects a planet smaller than 2 Earth radii with 90 % efficiency. The search found 24 of 28 RV-discovered planets that are known to transit (positive controls), and the injection model predicts 20.6. One control (55 Cnc e) failed because its RV period is a one-day alias. The expected number of new detections is 11.4 ± 3.2 and the observed number is 0. This is a selection effect, because earlier surveys already found the transiting RV planets, and these planets are the controls. When we add the controls, the search finds 24 transits and the model expects 13.2 ± 3.5 (P = 0.003). Without the 9 controls in systems where a transit survey found a planet, it finds 18 and the model expects 12.7 (P = 0.09). The strongest signal that passed the first vetting step is the catalogued 8.17-d eclipsing binary at HD 44219. The centroid shifts and the difference images show that the eclipses come from the neighbour star Gaia DR3 3001428355964665728 (TIC 443088035, G = 12.1, 19.5 arcsec away), not from HD 44219 (the difference-image fits prefer the neighbour by Δχ² ≥ 490). The Gaia DR3 variability tables, VSX and Prša et al. (2022) do not list this neighbour as a variable star. Assumptions: the archive orbits and their errors are correct and independent, the radii follow the Chen & Kipping relation with m = m sin i, and the orbits have random orientation. The package contains the code, the archive query results (not the Gaia query), the per-planet results, the injections, the vetting tables and the figures. A script downloads the light curves again from MAST.

TESS · radial velocity · transit search · exoplanets · transit probability · injection-recovery · null search · positive controls · eclipsing binary · HD 44219 · centroid analysis · difference imaging · detection efficiency · false alarm · ephemeris propagation · NASA Exoplanet Archive · Gaia DR3 · Chen-Kipping mass-radius relation