Paper II of this series described the chemical and microstructural evolution of ferri-ilmenite solid solutions during high-T quench and short-term annealing. Here we explore consequences of these Fe–Ti ordering-induced microstructures and show how they provide an explanation for both self-reversed thermoremanent magnetization and room-T magnetic exchange bias. The dominant antiferromagnetic interactions between (001) cation layers cause the net magnetic moments of ferrimagnetic ordered phases to be opposed across chemical antiphase domain boundaries. Magnetic consequences of these interactions are explored in conceptual models of four stages of microstructure evolution, all having in common that A-ordered and B-anti-ordered domains achieve different sizes, with smaller domains having higher Fe-content, lesser Fe–Ti order, and slightly higher Curie T than larger domains. Stage 1 contains small Fe-rich domains and larger Ti-rich domains separated by volumes of the disordered antiferromagnetic phase. Magnetic linkages in this conceptual model pass through disordered host, but self-reversed TRM could occur. In stage 2, ordered domains begin to impinge, but some disorder remains, creating complex magnetic interactions. In stages 3 and 4, all disordered phase is eliminated, with progressive shrinkage of Fe-rich domains, and growth of Ti-rich domains. Ordered and anti-ordered phases meet at chemical antiphase and synphase boundaries. Strong coupling across abundant antiphase boundaries provides the probable configuration for self-reversed thermoremanent magnetization. Taking the self-reversed state into strong positive fields provides a probable mechanism for room-temperature magnetic exchange bias.
Hemo‐ilmenite ores from Allard Lake, Quebec, were first studied over 50 years ago. Interest was renewed in these coarsely exsolved oxides, based on the theory of lamellar magnetism as an explanation for the high and stable natural remanent magnetizations (NRMs), 32 to 120 A/m, reported here. To understand the magnetism and evolution of the exsolution lamellae, the microstructures and nanostructures were studied using scanning electron microscopy and transmission electron microscopy (TEM), phase chemistry, and relations between mineral chemistry and the hematite‐ilmenite phase diagram. Cycles of exsolution during slow cooling resulted in lamellae down to 1–2 nm thick. Combined electron microprobe, TEM, and X‐ray diffraction (XRD) results indicate that hematite hosts reached a composition approximately ilmenite (Ilm) 14.4, and ilmenite hosts ∼Ilm 98. The bulk of the very stable NRM, which shows thermal unblocking ∼595–620°C, was acquired during final exsolution in the two‐phase region canted antiferromagnetic R c hematite + R ilmenite. Hysteresis measurements show a very strong anisotropy, with a stronger coercivity normal to, than parallel to, the basal plane orientation of the lamellae. Magnetic saturation (M s ) values are up to 914 A/m, compared to 564 A/m predicted for a modally equivalent spin‐canted hematite corrected for ∼15% R 2+ TiO 3 substitution. Low‐temperature hysteresis, AC‐susceptibility measurements, and Mössbauer results indicate a Néel temperature (T N ) of the geikielite‐substituted ilmenite at ∼43 K. The low‐temperature hysteresis and AC‐susceptibility measurements also show a cluster‐spin‐glass‐like transition near 20 K. Below T N of ilmenite an exchange bias occurs with a 40 mT shift at 10 K.
SUMMARY Lamellar magnetism is a source of remanent magnetization in natural rocks different from common bulk magnetic moments in ferrimagnetic minerals. It has been found to be a source for a wide class of magnetic anomalies with extremely high Koenigsberger ratio. Its physical origin are uncompensated interface moments in contact layers of nanoscale ilmenite lamellae inside an hematite host, which also generate unusual low-temperature (low-T) magnetic properties, such as shifted low-T hysteresis loops due to exchange bias. The atomic-magnetic basis for the exchange bias discovered in the hematite-ilmenite system is explored in a series of papers. In this third article of the series, simple models are developed for lamellae interactions of different structures when samples are either cooled in zero-field, or field-cooled in 5 T to temperatures below the ordering temperature of ilmenite. These models are built on the low-temperature measurements described earlier in Paper II. The important observations include: (i) the effects of lamellar shapes on magnetic coupling, (ii) the high-T acquisition of lamellar magnetism and low-T acquisition of magnetization of ilmenite lamellae, (iii) the intensity of lamellar magnetism and the consequent ilmenite magnetism in populations of randomly oriented crystals, (iv) lattice-preferred orientation of the titanohematite host crystal populations and (v) the effects of magnetic domain walls in the host on hysteresis properties. Based on exemplary growth models of lamellae with different geometries and surface couplings we here provide simple models to assess and explain the different observations listed above. Already the simplified models show that the shapes of the edges of ilmenite lamellae against their hematite hosts can control the degree of low-T coupling between ilmenite, and the lamellar magnetic moments. The models also explain certain features of the low-T exchange bias in the natural samples and emphasize the role of lattice-preferred orientation upon the intensity of remanent magnetization. The inverse link between ilmenite remanence and exchange-bias shift in bimodal low-T ilmenite lamellae is related to different densities of hematite domain walls induced by the clusters of ilmenite lamellae.
Mid‐Proterozoic granulites in SW Sweden, having opaque minerals hematiteilmenite with minor magnetite, and occurring in an area with negative aeromagnetic anomalies, have strong and stable reversed natural remanent magnetization ∼9.2 A/m, with 100% remaining after demagnetization to 100 mT. Samples were characterized by optical microscopy, electron microprobe (EMP), transmission electron microscopy (TEM), and rock‐magnetic measurements. Earliest oxide equilibrium was between grains of titanohematite and ferri‐ilmenite at 650°–600°C. Initial contacts were modified by many exsolution cycles. Hematite and ilmenite (Ilm) hosts and lamellae by EMP are Ilm 24–25, ILm 88–93, like titanohematite, and ilmenite above 520°C on Burton's diagram [1991]. Finer hosts and lamellae by TEM are Ilm16 ±3 and Ilm 88±4, like coexisting antiferromagnetically ordered (AF) hematite and ilmenite below 520°C on Burton's diagram. This may be the first example of analytical identification, in one sample, of former hematite, now finely exsolved, and AF hematite. TEM microstructures consist of gently curving semicoherent ilmenite lamellae within hematite, flanked by precipitate‐free zones and abundant ilmenite disks down to unit cell scale (1–2 nm). Strain contrast of disks suggests full coherence with the host, and probable formation at the reaction titanohematite ‐‐‐> AF hematite + ilmenite at 520°C. Magnetic properties are a consequence of chemical and magnetic evolution of hematite and ilmenite with bulk compositions ilmenite‐richer than Ilm 28, that apparently exsolved without becoming magnetized, down to 520°C where hematite broke down to AF hematite plus ilmenite, producing abundant AF hematite below its Néel temperature. Intensity of magnetization is greater than possible with hematite alone, and TEM work suggests that ultrafine ilmenite disks in AF hematite are associated with a ferrimagnetic moment due to local imbalance of up and down spins at coherent interfaces.
Chemical and microstructural evolution during quench and short-term annealing of a sample XFeTiO3= 0.61 is explored in the light of observations in Paper I. Ordering proceeds by (1) random appearance of ordered and anti-ordered domains within a disordered host, (2) coarsening of ordered and anti-ordered domains until they impinge along antiphase domain boundaries, (3) growth of regions where one ordered or anti-ordered phase becomes dominant over the other, with progressive reduction in surface area of antiphase domain boundaries and (4) dynamic development where antiphase boundaries migrate during annealing, leading to Fe enrichment of shrinking domains and Fe depletion of growing domains. These conclusions are supported by 2-D Monte Carlo simulations illustrating that Ti-Ti avoidance is a powerful driving force for Fe enrichment along antiphase boundaries, and by bond-valence calculations demonstrating that local charge balance is improved when antiphase domain boundaries contain a combination of Fe-rich contact layers and disordered boundary layers along (001). Chemical phase separation during quenching is driven by the disorder/order transition at temperatures above the tricritical point and by spinodal decomposition at temperatures below the tricritical point. The former explains microtextures and chemical features in samples quenched from high temperature; the latter produces textural and chemical evolution during subsequent annealing. All these features provide the atomic basis for self-reversed thermoremanent magnetization and room-temperature magnetic exchange bias as will be described in Paper III.
Quenched ferri-ilmenite solid solutions X FeTiO3+ (1 –X) Fe2O3 with X≈ 0.60 contain chemical and magnetic structures important for understanding the unusual magnetic properties in this series, including self-reversal in igneous rocks and exchange bias. Here we study a composition X= 0.61, annealed at 1055 °C, above the Fe-Ti ordering temperature, then quenched. Presence of two interface-coupled phases is established by pot-bellied character of the room-temperature magnetic hysteresis loop, and large negative magnetic exchange bias below 30 K. Transmission electron microscopy (TEM) dark-field imaging with the 003 reflection shows dominant Fe-Ti disordered antiferromagnetic and lesser ordered ferrimagnetic phases, the latter in lenses ≤8 nm thick. Parts of the ordered phase are in antiphase relationship, shown by high-resolution TEM imaging of Fe-rich and Ti-rich layers. TEM-EDX analyses indicate chemical phase separation during quench, with dominant compositions X= 0.56–0.63, extremes 0.50 and 0.70. Thermomagnetic experiments indicate compositions X= 0.56–0.61 are antiferromagnetic, X= 0.61–0.64 are ferrimagnetic. A sample held ∼5 min at 1063 K, increased in order, demonstrated by twofold increase in induced moment at 1 T. This then acquired self-reversed thermoremanent magnetization between 490 and 440 K. Progressive annealings of another sample at 773 K, 973 K, 1023 K and 1063 K, followed by cooling in a 1 T field, produced positive room-temperature magnetic exchange bias, only for 1023 K and 1063 K runs. These properties suggest growth of ordered regions from disordered regions, and expansion of some ordered domains against others across antiphase boundaries, thus creating self-organized structures essential for magnetic self-reversal and magnetic exchange bias.
DNA structural analysis of the Qa region in two BALB/c mouse substrains with different Qa-2 phenotypes reveals that a deletion of DNA has occurred in BALB/cBy (Qa-2-) mice relative to BALB/c (Qa-2+) mice. We propose that this deletion arises from unequal crossing-over and recombination between adjacent BALB/c class I genes and results in the generation of a hybrid class I gene in BALB/cBy mice. Furthermore, we suggest that this is a direct cause of the change in Qa-2 phenotype. Further support for this model was obtained from transfection experiments in which cloned genes from the equivalent part of the Qa region in C57BL/10 mice were introduced into L cells. Four C57BL/10 genes, arranged in two almost identical pairs, encode polypeptides that are precipitated from lysates of transfectants with anti-Qa-2/3 antiserum. Although loss of one pair of these genes in BALB/c mice has no qualitative effect on Qa-2 phenotype, the loss of both pairs of genes via gene fusion leads to the loss of the Qa-2+ phenotype in BALB/cBy mice.
SUMMARY The Rogaland Igneous Complex (RIC) in southern Norway intruded into Sveconorwegian granulite crust beginning ∼930 Ma. Three massif anorthosite bodies, Egersund–Ogna, Helleren and Åna-Sira, were intruded some 10 Myr later by the Bjerkreim–Sokndal layered intrusion. The Garsaknatt leuconorite and the ilmenite-rich Tellnes norite, one of the youngest rock in the complex at ∼920 Ma, intrude the anorthosite or nearby country rock. Magnetic mineralogy and palaeomagnetic studies carried out on the Tellnes norite, the Garsaknatt leuconorite and the surrounding Åna-Sira anorthosite, indicate the magnetization of all three bodies are dominated by hemo-ilmenite carrying the remanence as a thermochemical remanent magnetization, although magnetite is present in some samples. The three bodies yield steep negative inclinations with northwesterly declinations (Tellnes, I = −71.9°, D = 305.0°, α95 = 10.6°; Garsaknatt, I = −73.1°, D = 312.7°, α95 = 4.7°; and Åna-Sira, I = −81.2°, D = 326.3°, α95 = 6.7°). When combined with data from other bodies in the RIC, the older anorthosites have steeper inclinations, and higher palaeolatitudes, while the younger units have less steep inclinations and shallower palaeolatitudes by nearly 10°, indicating northward plate motion during cooling of the intrusions. Age of the remanence is difficult to determine precisely, however, best estimates are ∼910 Ma for the older anorthosites and ∼900 Ma for the younger intrusions. Although these differences are significant, a unified pole position (35.6° N, 215.1° E), combining all the 111 sites from the RIC, strongly supports the assumed position of southern Baltica in Rodinia at ∼900 MA.
'%3e%3cpath fill='%23012169' d='M0 0v30h60V0z'/%3e%3cpath stroke='%23fff' stroke-width='6' d='m0 0 60 30m0-30L0 30'/%3e%3cpath stroke='%23C8102E' stroke-width='4' d='m0 0 60 30m0-30L0 30' clip-path='url(%23b)'/%3e%3cpath stroke='%23fff' stroke-width='10' d='M30 0v30M0 15h60'/%3e%3cpath stroke='%23C8102E' stroke-width='6' d='M30 0v30M0 15h60'/%3e%3c/g%3e%3c/svg%3e)
