Study Of Oxygen Ionic Valence And Its Influence On The Magnetic Property Of The Perovskite Manganites La_1-xSr_xMnO₃

It is well known that there are both ionic and covalent bonds in oxides, and ionicity has been defined as the fraction of ionic bonds among all bonds, both ionic and covalent. The ionicity of oxides has been researched for decades, and the ionicity of materials were analyzed by various methods, including experimental and theoretical approaches. Considering the ionicity of oxides, a portion of oxygen ions in oxides are monovalent. There must be O2 p holes in the outer orbits of the monovalent oxygen ions. These O2 p holes are also expected to affect the magnetic ordering and electrical transport properties of the materials. However, few researchers have considered the influence of oxides ionicity on the magnetic properties of the materials.In traditional views, magnetic ordering of oxides was explained using magnetic superexchange(SE) and double exchange(DE) interaction models, in both which the valences of all oxygen ions were assumed as-2. However, there is no satisfactory explanation for some experimental phenomena. For example, it has not been found a quantitative explanation for the dependence of magnetic moments on the Sr-doping level x for the manganites La1-xSrxMnO3 with an ABO3 perovskite structure. For these questions, the main reason may be that the influence of oxide ionicity on the physical properties of the materials was neglected, that is, all oxygen ions were assumed to obtain 2 electrons, forming a 2s22p6 stable electron structure. Therefore, we explained the magnetic ordering of the manganites La1-xSrxMnO3 using the O2 p itinerant electron model proposed recently by our group, in which the ionicity was taken into account.This paper contains two main aspects. First, applying X-ray photoelectron spectra(XPS), we estimated the average valence(ValO) and the ionicity(fi) of the oxygen anions for several oxides. Second, taking into account the ionicity of oxides, we explained the magnetic ordering of the perovskite manganites La1-xSrxMnO3 using the O2 p itinerant electron model. The results are as follows:(1) O1 s photoelectron spectra of the powders samples including perovskite oxide BaTiO3, and the monoxides CaO, MnO, CoO, ZnO, NiO and CuO were investigated. The computer program XPSPEAK Version 4.1 was used to fit the narrow-scan spectra of the XPS peaks. The O1 s narrow-scan spectra of the XPS peaks were simulated using symmetrical Gaussian-Lorentzian product functions, and the satellite peak with about 2 e V of chemical shift ΔE(ΔE=E2-E1) from O2- peak was assumed to correspond to O1- ions. The average valence(ValO) of the oxygen anions was estimated using the fitting of O1 s spectrum. We found that the value of ValO(=-1.55) for BaTiO3 is close to the value(ValO=-1.63) calculated by Cohen using density functional theory(DFT) [Nature 358, 136(1992)]. Assuming that the value of the ionicity(fi) for the oxides, fi =|ValO|/2, the dependence of the fi for monoxides on the second ionization energy, V(M2+), of the cations is close to that reported by Phillips [Rev. Mod. Phys. 42, 317(1970)]. We therefore suggest that the analysis for O1 s photoelectron spectra should be accepted as a general experimental method for estimating the ionicity.(2) XPS of a single crystal sample of SrTiO3(SSTO) and a polycrystalline bulk sample of SrTiO3(PSTO) were studied, and 7 sets of photoelectron spectra were measured after the sample was etched with an argon ion beam for 0, 30, 60, 90, 120, 150 and 180 s. On the basis of the analysis of narrow-scan spectra of O, Ti and Sr ions for different etching time, two principle observations can be made:(i) the intensity ratio of the satellite peak with about 2eV of the chemical shift ΔE(ΔE=E2-E1) to the O2- peak decreased with etching time.(ii) The content of oxygen vacancies in the samples increased with etching time. These results show the assumption that the satellite peak with about 2eV of the chemical shift from the O2- peak is corresponds to the oxygen vacancies [Applied Physics Letters, 95, 203502 and 105, 152904], may be unsuitable. If it is assumed that the satellite peak corresponds to the monovalent oxygen(O1-) ions, these experimental observations can be readily explained.(3) The samples with nominal composition La1-xSrxMnO3(0.0≤x≤0.4) were prepared by sol-gel method, which crystal structure parameters and magnetic parameters were measured. We then examined the ionicities of the samples using XPS and found that there are no Mn4+ cations in La1-xSrxMnO3. Thus, the ferromagnetic ordering of Mn cations in the samplse cannot be explained using the DE model along the ion chains Mn3+-O2--Mn4+. Then, using the O2 p itinerant electron model, we explained the magnetic structure and fitted successfully the dependence on the Sr-doping level of the magnetic moments of the samples.(4) The three powder samples with the nominal composition La0.95Sr0.05MnO3 were prepared by different thermal-treatment conditions. The XRD patterns of the samples showed that all the samples have only a single ABO3 perovskite phase with space group R3 c, and that the crystal structure parameters are almost the same each other. Both Curie temperatures, TC, and magnetic moment, μexp, per formula at 10 K increase with increasing thermal-treatment procedure. Furthermore, we found that there are no Mn4+ cations in the samples using XPS analysis, and the average valence of Mn cations in the samples also increases with increasing thermal-treatment procedure. These phenomena have successfully been explained using the O2 p itinerant electron model. In addition, the variation tendency of the average valences of Mn cations calculated using the magnetic moments μexp, is in accordance with experimental results of XPS.