Rob Hargraves and the External Force in Lamellar Magnetism

The strong remanence and extreme coercivity of slowly cooled rocks rich in hemo-ilmenite or ilmeno-hematite, poor in or lacking magnetite, was recognized and puzzled over by Rob for 44 years and highlighted in recent studies. Together these are properties neither of paramagnetic (PM) ilmenite nor spin-canted antiferromagnetic (CAF) hematite. The minerals contain fine exsolution lamellae, now shown by TEM to go down to unit-cell scale, suggesting lamellar interfaces as the key. Atomic simulations of PM ilmenite lamellae in CAF hematite show formation of "contact layers" on (001) coherent interfaces that have a hybrid composition between hematite Fe3+ layers and ilmenite Fe2+ layers. These reduce interface charge imbalance, and, more important, have a magnetic moment coupled anti-parallel to but weaker than adjacent hematite layers. Each ilmenite lamella has an odd number of non-magnetic layers plus two contact layers coupled to hematite. The hematite host has an odd number of layers so magnetic moments of all but one cancel. This, combined with two opposite moments of contact layers (2MC-1MH), gives the moment of one lamella, about 4 Bohr magnetons. The maximum moment per formula unit is the moment per lamella times number of lamellae divided by formula units. One key to achieving a high moment is abundant lamellae. Rob's discussion of etched Allard Lake samples brought attention to the 3-phase cooling reaction in the ilmenite-hematite system, where R3c PM titanohematite transforms to about 20% PM R3 ilmenite and 80% CAF hematite, producing instantaneously the required coupled contact layers and a CRM. Lamellar yields up to 33% are obtained by undercooling below the 3-phase reaction, with eventual very fine nucleation. Long-term heating indicates major loss of the high-coercivity component due to lamellar resorption. A second key to strong remanence is that lamellae be magnetically "in-phase". This is optimized in crystals with (001) parallel to the external force of the magnetizing field during lamellar nucleation, and in nature by having (001) lattice-preferred orientations parallel to the field. Rob demonstrated this, using AMS of Allard Lake samples, where angle from the mean (001) to the Proterozoic magnetic vector is inversely related to intensity.