The TESS response model gives a better fit at TOI-2640 than at a bright neighbor in two 2021 sectors

Which star position best fits the 2021 dip near TOI-2640.01? We fit the archived TESS point response model to lost-light images from public 120-second data in Sectors 36 and 37. The two tested positions are the target, Gaia DR3 5404579488593432576, and the brighter Gaia DR3 5404579488593434240, about 15.97 arcseconds away. Each model has its own amplitude and constant image background. With archive coordinates and formal pixel weights, the neighbor has a larger weighted residual statistic by 118.01 and 89.66 in the two sectors. The sum is 207.67. The target retains the smaller residual with three dip durations, two image sizes, a background plane, and larger errors from cycle scatter. Resampling blocks of five observed cycles gives a central 95 percent range of 140.66 to 285.39 for the summed difference. These ranges are not null-test or host probabilities. Separate position shifts of up to 0.3 pixel also retain the preference; this allowance is a sensitivity test, not a measured error bound. Free position fits are offset from the target by 0.14 and 0.27 pixel. Some faint Gaia sources remain hard to distinguish, and response and coordinate errors are not calibrated. Thus, the result is a conditional comparison with one bright neighbor, not a unique host assignment or a planet confirmation. We also repeat the later depth change reported in prior SCI work. The package supplies source code, data provenance, original light curves, compact pixel statistics, response files, results, figures, and reproduction instructions.

TESS · TOI-2640.01 · TIC 135036669 · difference images · point response function · source position · transit vetting · Gaia DR3 · block bootstrap · reproducible astronomy