Structural a N D Magnetic Properties of Ferrihydrite

-57Fe MSssbauer spectra of two synthetic samples of ferrihydrite, recorded at 4.2 K in applied fields of up to 9 T, have been analysed by a mean-field model. The samples exhibit two and six X-ray diffraction peaks. It is shown that only one ferric ion site is present in the mineral, and that in this site the ions are octahedrally coordinated. The spectra show the presence of different magnetic states: ferrimagnetism in two-line ferrihydrite, and antiferromagnetism in six-line ferrihydrite. The ferrimagnetism in two-line ferrihydrite is analysed in terms of random fluctuations arising from the small numbers of ferric ions per particle, and it is shown that the different magnetic states may arise purely as a result of these fluctuations. Key Words--Ferrihydrite, Ferrimagnetism, Antiferromagnetism. I N T R O D U C T I O N Ferrihydrite is a poorly crystalline hydrated ferric oxyhydroxide which is the growth precursor to more crystalline iron minerals such as hematite. It has the approximate composit ion FesHOs-4H20. The poor crystallinity of ferrihydrite samples is evident in their X-ray diffraction patterns, which contain between two and six broad lines. Samples are described as, for example, two-line, four-line, or six-line ferrihydrites, as a means of indicating the degree of crystallinity. Both the structural and the magnetic properties of ferrihydrite are currently the subject of vigorous debate. It is generally accepted that ferrihydrite has a structure based on that of hematite, with a hexagonal unit cell (a = 5.08 /~, c = 9.4 /~) and the iron in octahedral coordination (Towe and Bradley, 1967; Chukhrov et aL, 1973). However, on the basis of X-ray diffraction, highresolution transmission electron microscopy, and X-ray absorption edge data, Eggleton and Fitzpatrick (1988) propose that a new structural model based on a trigonal unit cell is more appropriate. They conclude that 36% of the Fe is present in tetrahedral sites, with the remainder in octahedral sites. This proposal has been disputed by Manceau et al. (1990). Eggleton and Fitzpatrick (1990) contend that as yet none of the proposed models for the structure of ferrihydrite, ranging from all tetrahedral to all octahedral Fe, have been either proved or disproved. In support of this they refer to the room temperature MSssbauer spectrum of six-line ferrihydrite, which can be modelled equally well with paramagnetic doublets corresponding to either 0% or 25% tetrahedral Fe (Cardile, 1988). Further, Murad (1988) analysed both the paramagnetic and magnetic Mrssbauer spectra of a 'well-crystallised' ferrihydrite in terms of different distributions of hyperfine parameters, and inferred that two distributions were needed, and that they could be ascribed to the presence of both Copyright 9 1992, The Clay Minerals Society octahedral and tetrahedral Fe. However, subsequent work on this 'well-crystallised' material (Bigham et al., 1990) has shown that sulphur plays an important role, and that the material may in fact be an iron (III) oxyhydroxysulphate rather than ferrihydrite s e n s u stricto. The magnetic state of ferrihydrite is also the subject of controversy. It is supposed to be 'speromagnetic' , a state in which the atomic magnetic moments are individually oriented at random despite the presence of antiferromagnetic nearest-neighbour interactions. The prototype speromagnet was a natural ferric gel (Coey and Readman, 1973) now thought to have been ferrihydrite (Madsen et aL, 1986; Murad et al., 1988) and which showed applied-field M/Sssbauer spectra similar to those of ferrihydrite (Madsen et al., 1986). We have recently pointed out (Pollard and Pankhurst, 1992) that although the applied field MSssbauer spectra of the prototype speromagnet were said to demonstrate the new magnetic state (despite contrary indications from bulk magnetic measurements which suggested antiferromagnetism), the original M/Sssbauer spectra can in fact be explained most simply as resuiting from antiferromagnetic order in fine particles. Consequently, the nature of the magnetic order in ferrihydrite is thrown into some doubt. M/Sssbauer spectroscopy offers the prospect of detailed analysis of both the crystallographic and magnetic states offerrihydrite. The room temperature spectra are ambiguous because they comprise overlapping paramagnetic doublets; however, on cooling to temperatures below the ordering temperature, hyperfine magnetic splitting occurs which increases the resolution. Applying magnetic fields enhances the resolution further. In addition, the response of the material to the applied field indicates its magnetic state. High-field (up to 9 T) spectra of two-line and sixline synthetic samples at 4.2 K have been recorded (Pollard et aL, 1992), but have not been analysed with