RESUMEN
The actual role of transition metals like iron in the room temperature ferromagnetism (RTFM) of Fe doped ZnO nanoparticles is still an unsolved problem. While some studies concluded that the Fe ions participate in the magnetic interaction, others in contrast do not believe Fe to play a direct role in the magnetic exchange interaction. To contribute to the understanding of this issue, we have carefully investigated the structural, optical, vibrational and magnetic properties of sol-gel synthesized Zn1-xFexO (0 < x < 0.10) nanoparticles. No Fe(2+) was detected in any sample. We found that high spin Fe(3+) ions are substitutionally incorporated at the Zn(2+) in the tetrahedral-core sites and in pseudo-octahedral surface sites in ZnO. Superficial OH(-) was observed in all samples. For x ≤ 0.03, an increment in Fe doping concentration decreased a and c lattice parameters, average Zn-O bond length, average crystallite size and band gap; while it increased the degree of distortion and quadrupole splitting. Undoped ZnO nanoparticles exhibited very weak RTFM with a saturation magnetization (Ms) of â¼0.47 memu g(-1) and this value increased to â¼2.1 memu g(-1) for Zn0.99Fe0.01O. Very interestingly, the Ms for Zn0.99Fe0.01O and Zn0.97Fe0.03O increased by a factor of about â¼2.3 by increasing annealing for 1 h to 3 h. For x ≥ 0.05, ferrimagnetic disordered spinel ZnFe2O4 was formed and this phase was found to become more ordered with increasing annealing time. Fe does not contribute directly to the RTFM, but its presence promoted the formation of additional single charged oxygen vacancies, zinc vacancies, and more oxygen-ended polar terminations at the nanoparticle surface. These defects, which are mainly superficial, altered the electronic structure and are considered as the main sources of the observed ferromagnetism.
RESUMEN
We propose a new direct method for calculating simultaneously two recoilless f-factors of any iron-bearing compound relative to that of a reference material by collecting only a single-temperature Mössbauer spectrum. This methodology is comparatively much simpler than the usual one which requires taking Mössbauer spectra of the compound at several temperatures and subsequently fitting the temperature dependence of the subspectral area or the isomer shift data with a lattice vibrational model. We demonstrate the applicability of this new methodology in the case of three common iron-bearing compounds: magnetite, akaganeite and goethite, but of course this type of study can be extended to other materials. The two f-factors for each compound were related to iron ions located in sites of different origin: for magnetite, these were related to irons with two different oxidation states; for akaganeite to irons in two different crystallographic sites; and for goethite to irons in similar crystal sites but located in grains of different sizes. In the case of magnetite, we found that the f-factors for the Fe(3+) and Fe(2.5+) sites relative to that of metallic iron powder were of fFe (3+)∕fFe = 0.97 ± 0.05 and fFe (2.5+)∕fFe = 0.92 ± 0.05, respectively. Interestingly, the quotient of these two f-factors, i.e., fFe (2.5+)∕fFe (3+), is equal to 0.95 ± 0.05, which compares fairly well with a value reported in literature obtained using the complex methodology based on the temperature dependence of the absolute subspectral area and the Debye approximation. For akaganeite, the f-factors of the doublet 1, D1, and doublet 2, D2, sites relative to that of metallic iron powder were of fD 1∕fFe = 0.95 ± 0.08 and fD 2∕fFe = 0.98 ± 0.15, respectively. And for goethite we found that the f-factors of the sextet 1, G1, and sextet 2, G2, sites relative to that of metallic iron powder were of fG 1∕fFe = 0.80 ± 0.02 and fG 2∕fFe = 0.80 ± 0.02, respectively. The similarity of these last two factors is perhaps due to a sharp distribution of large grains.