Experimental evidence for lamellar magnetism in hemo-ilmenite by polarized neutron scattering

Large local anomalies in the Earth's magnetic field have been observed in Norway, Sweden, and Canada. These anomalies have been attributed to the unusual magnetic properties of naturally occurring hemo-ilmenite, consisting of a paramagnetic ilmenite host ($\alpha$-Fe${}_2$O${}_3$-bearing FeTiO${}_3$) with exsolution lamellae ($\approx$$3\mu$m thick) of canted antiferromagnetic hematite (FeTiO${}_3$-bearing $\alpha$-Fe${}_2$O${}_3$) and the mutual exsolutions of the same phases on the micron to nanometer scale. The origin of stable natural remanent magnetization (NRM) in this system has been proposed to be uncompensated magnetic moments in the contact layers between the exsolution lamellae. This lamellar magnetism hypothesis is tested here by using polarized neutron diffraction to measure the orientation of hematite spins as a function of an applied magnetic field in a natural single crystal of hemo-ilmenite from South Rogaland, Norway. Polarized neutron diffraction clearly shows that the ilmenite spins do not contribute to the NRM and that hematite spins account for the full magnetization at ambient temperature. Hematite sublattice spins are shown to adopt an average angle of 56${}^{\circ }$ with respect to a saturating magnetic field, which is intermediate between the angle of 90${}^{\circ }$ predicted for a pure canted moment and the angle of 0${}^{\circ }$ predicted for a pure lamellar moment. The observed NRM is consistent with the vector sum of lamellar magnetism and canted antiferromagnetic contributions. The relative importance of the two contributions varies with the length scale of the microstructure, with the lamellar contribution increasing when exsolution occurs predominantly at the nanometer rather than the micrometer scale.

Figure 1

Antiferromagnetic structure of hematite and ilmenite. Left: Magnetic structure of hematite above the Morin transition. Right: Magnetic structure of ilmenite. The oxygen atoms are left out of the drawings, as is the small canting of the Fe${}^{3+}$ moments in hematite.

Figure 2

Sketch of the in-plane hematite spin directions and their response to a magnetic field applied applied in the plane. The net moment is the vector sum of the CAF moment, which is almost perpendicular to the AFM sublattice, and the lamellar moment which is parallel to the AFM sublattice. The two “domains” represent two of the possible six antiferromagnetic domains. In zero field the spin orientation depends on the remanent magnetization of the sample. If all six domains are equally represented, or if the spins are randomly oriented, the average spin angle with respect to $\mathbf{B}$ will be ${45}^{\circ }$. At a saturating field (2.5 T at room temperature) the net moment is aligned with the field and the average spin orientation is no longer random, but makes an angle with $\mathbf{B}$ that depends on the proportion of canted and lamellar moments.

Figure 3

Mapping of the $\left(003\right)$ peaks measured at MORPHEUS. 2$\theta$ is the scattering angle and $\omega$ is the azimuthal rotation angle of the sample. The most intense peak at $2\theta \approx {60}^{\circ }$ is the structural ilmenite reflection and the less intense peak at $2\theta \approx {62}^{\circ }$ is the magnetic hematite reflection. Both peaks have a shoulder, indicating that the sample consists of two distinct crystallites oriented at an angle of approximately $0.6^{\circ }$ with respect to each other.

Figure 4

Geometry in the IN12 experiment. ${\mathbf{P}}_i$ is the incoming polarization vector, $\mathbf{q}$ is the scattering vector, and ${\mathbf{M}}_{\perp }$ is the magnetic sublattice magnetization perpendicular to $\mathbf{q}$. The magnetic field is applied in the $z$ direction (parallel to ${\mathbf{P}}_i$). The angle between ${\mathbf{M}}_{\perp }$ and ${\mathbf{P}}_i$ is called $\theta$.

Figure 5

NSF and SF measurement of the $\left(003\right)$ structural ilmenite peak and the $\left(003\right)$ magnetic hematite peak. This measurement was in an applied field of 0.25 T and at a temperature of 65 K. This data have not been corrected for imperfect polarization as can be seen from the nonzero SF intensity at the position of the structural ilmenite peak.

Figure 6

NSF and SF measurement of the $\left(003\right)$ structural ilmenite peak and the $\left(003\right)$ magnetic hematite peak. This measurement was in an applied field of 0.25 T and at a temperature of 65 K. This data has been corrected for imperfect polarization with a flipping ratio of $R=43$. The ilmenite reflection is only present in the NSF signal, whereas the hematite peak is present in both the NSF and the SF signal.

Figure 7

Temperature variation of the $\left(10\overline{\frac{1}{2}}\right)$ magnetic ilmenite peak, showing the second order phase transition. Both black and hollow points are peak amplitude measurements, however, the fit is to the black points only, since the power law behavior is only valid within approximately this range.

Figure 8

Spin orientation as a function of applied field for different temperatures. The error bars were obtained from Monte Carlo simulations based on the errors on the Voigtian fits.

Figure 9

Unpolarized nuclear diffraction measurement at 2 and 150 K of the (101) peak, for applied fields between 0 and 11 T.