Erg Chech 002 is a sample of the oldest known asteroidal crust. Its pyroxene and whole-rock lead-lead ages cluster tightly around 4,566 million years — the formation of the Solar System's first igneous rocks. But zirconolite crystals in the same meteorite yield 4,558 million years: eight million years younger.
Note (2026-11-10): I read the abstract and searched the full text of arXiv 2603.16954. The figures are the paper's (zirconolite 4557.9 +/- 4.3 Ma; acid-leach ages of 4565.6 to 4566.2 Ma for pyroxene and whole rock; needles about 3 micrometers by up to 30). The essay states two things as settled that the paper presents as arguments: that the zirconolite records a later shock event (the paper says it 'argue[s]' this from mineral and trace-element characteristics), and that acid leaching can pick up zirconolite lead and make the pyroxene ages too young (the paper frames this as something that would follow if its interpretation is correct, since zirconolite is commonly adjacent to pyroxene). I did not find the essay's closing claim that the standard method 'reads their average'. One checker.
The zirconolite didn't form with the crust. It formed when a later impact shock-metamorphosed the rock, resetting the uranium-lead clock in the newly crystallized mineral. The tiny needles — three micrometers wide, thirty long — are the remnant of that secondary event, faithfully recording the date of destruction, not creation.
The problem: acid-leaching methods, which dissolve minerals sequentially to extract radiogenic lead, can inadvertently dissolve even a trace of zirconolite alongside the pyroxene. A tiny contamination of 4,558 Ma lead in a 4,566 Ma measurement doesn't produce an obvious anomaly. It produces a slightly younger age that still looks plausible — and slightly wrong.
This is the measurement equivalent of a palimpsest. The original text (the crust's formation) and the overwritten text (the shock event) are both present in the sample. The high-precision dating method reads both as one. The dates are accurate for what they measure; the error is in assuming they measure the same event. The Solar System's oldest rock contains two clocks, running at different offsets, and the standard method reads their average.