Mercury is the only planet besides Earth that has both a global magnetic field arising from a core dynamo and a crustal field produced by magnetized rocks. Since NASA's MESSENGER spacecraft orbited the planet between 2011 and 2015, the working assumption has been that at least part of that crustal magnetism is remanent — a fossil record of a stronger early dynamo. A new study in PNAS rebalances the two contributions.

The team, led by Catherine Johnson of the University of British Columbia, used MESSENGER's low-altitude magnetic field observations to build a magnetization model of Mercury's crust. The core of the calculation: how strong should induced magnetization be — that is, the effect of today's weak core field on iron-bearing rocks — given assumptions about crustal thickness, magnetic mineralogy, iron content and the present-day core field, which is about a hundred times weaker than Earth's?

The answer: across more than 85 percent of the area north of 38°N, induced magnetization alone can account for the observed crustal fields, provided it extends through the crust to a depth of roughly 30 kilometres. Over that majority of the mapped region, no remanent magnetization is required. The strongest values appear in the Caloris region, as in earlier work, alongside localised signals elsewhere. The surprise is in the pattern: magnetization strength correlates with crustal thickness up to about 30 km, but not with the near-surface iron content that MESSENGER's spectrometers measured — suggesting the magnetized rocks sit in the upper crust.

Where induced magnetization falls short, other explanations remain on the table. Iron delivered by ancient asteroid impacts could locally raise induced magnetization; elsewhere a contribution from remanent magnetization may still be needed. The result does not settle the question, but it shifts part of the burden of proof away from the ancient dynamo.

The study's central weakness is methodological. On Earth, separating induced from remanent magnetization requires laboratory measurements of rock samples. For Mercury, no samples exist, so the magnetic properties of the crust must be inferred from satellite observations constrained by assumptions about composition and magnetic behaviour. Lon Hood of the University of Arizona, who has worked extensively on Mercury's crustal magnetism but was not involved in the new study, puts it plainly: the delineation depends on assumptions about magnetic mineralogy that cannot be verified without returned samples. Even the crust's low-field susceptibility rests on a limited number of laboratory measurements of meteorites and minerals believed to resemble Mercury's crust.

Help could come from the European-Japanese BepiColombo mission, launched in 2018 and expected to run until 2029: better core-field models and higher-resolution iron maps are the hope. Hood cautions that neither spacecraft currently plans to approach close enough to map more of the crustal field, though both will eventually descend below 100 kilometres, giving a first look at the southern hemisphere. Rock samples from Mercury's crust — the measurement that would definitively settle the origin of the magnetization — are not planned at all.

So the argument is not closed. A second PNAS study this year, by Isaac Narrett, Benjamin Weiss and colleagues, reads the same crustal record the other way, inferring a dipole field with an equatorial surface strength of at least about 2,000 nanotesla — evidence, in their reading, of a stronger ancient dynamo. Two papers, the same spacecraft data, opposite conclusions: how much Mercury tells us about its magnetic past currently depends on assumptions nobody can yet test against the real rock.

Sources