Later ZTF data show a second brightening of Gaia DR3 microlensing event 172 in 2019 and no repeat in 80 other testable events from 2018 to 2025

A single-lens microlensing event does not repeat. We tested the 363 Gaia DR3 microlensing events (2014-2017) for later brightening. We used ZTF DR24 light curves from 2018 to 2025 for 81 events, OGLE, KMTNet and Gaia Science Alerts event lists for all events, and Gaia epoch photometry. A fixed outburst detector flagged 2 of 198 event light curves, and same-magnitude neighbour stars predict 1.95. Both flags are low-altitude artefacts. The same detector found later outbursts in 25 of 31 dwarf novae that Gaia alerted in 2014-2017. No event has a second list event at its position, and the OGLE base rate predicts 1.0. The 95% upper limit on sources that repeat in 2018-2025 is 4.5% to 11.6%, depending on the outburst type. Event 172 (Gaia DR3 2046379503690071296) is an exception. In 2019, at least 707 days after the Gaia peak, it was 0.7 to 0.8 mag above its 2023 level. Two ZTF fields and the ZTF alert stream confirm this, and it stayed 0.2 to 0.35 mag above that level in 2018-2020. A single lens cannot make this, and the source type is not known. For event 257, which the catalogue paper did not decide, a point-lens model with annual parallax fits Gaia and ZTF data together (tE 612 +/- 135 d). The residuals show no quasar-like red noise. One contaminant in 81 events (1.2%; 95% interval 0.03-6.7%) agrees with the published 0.6-1.7% estimate. Limitations: only 22% of events have ZTF data, the detector does not see slow variability, and contaminants that repeat after 2025 are not constrained.

gravitational microlensing · Gaia DR3 · ZTF · catalogue contamination · OGLE · KMTNet · Gaia Science Alerts · microlensing parallax · variable stars · out-of-sample test