Fixed TESS models give lower residual sums at Gaia binary positions than at the assigned stars for two TOIs in four sectors

We tested the sources of the eclipses assigned to TOI-2998.01 and TOI-3143.01. Each target has a nearby Gaia DR3 binary candidate with a matching period and eclipse epoch. We measured event times and images of lost light in public two-minute TESS data. We used sectors 63 and 89 for TOI-2998.01, and sectors 99 and 101 for TOI-3143.01. With fixed response models and pointing terms, all four main image fits put the loss closer to the Gaia binary. The binary position also gives a lower weighted residual sum than the assigned-star position in each sector. The fitted distances from the binaries are 0.124, 0.256, 0.041, and 0.082 pixel, respectively. Event bootstrap limits hold the response model and coordinate shift fixed. These limits omit model error. TOI-2998.01 has a position offset in sector 89. Its sector 63 result changes when the pointing terms are removed. The tested controls are more stable for TOI-3143.01. These tests compare two specified stars in each field. Other close sources remain possible. They support a nearby-binary explanation but do not establish a final planet disposition. The package supplies lossless data extracts, source queries, hashes, code, fitted images, event times, bootstrap samples, and four figures. The binaries and the image method were known before this work.

TESS · Gaia DR3 · TOI-2998.01 · TOI-3143.01 · eclipsing binaries · source location · contamination · pixel response function · reproducible research