In this work, solid solutions of La_0.7Sr_0.3Mn_0.95Fe_0.05O_3-δ with different oxygen content were obtained by the solid-phase reactions technique. Based on the investigation of the dynamics of changes in the oxygen index (3 – δ) during heating of the samples, the formation of a stressed state in their grains as a result of annealing was established. This results in a decrease in the mobility of oxygen vacancies during the reduction of cations according to the Mn⁴⁺ + e– → Mn³⁺ scheme and explains the decrease of released oxygen amount with an increase of δ as well as the heating rate of the samples. When studying the magnetic properties of the obtained samples, it was found that the temperature dependence of the magnetization obeys the Curie–Weiss law and as the oxygen defficiency increases, the Curie temperature for solid solutions decreases. It was found that the particles are in a frozen ferromagnetic state when measured in the low-temperature region of the М (Т) dependence in “zero-field mode” at Т ˂ Т_В. The presence of ferromagnetism at Т ˃ Т_В leads to a magnetically ordered state, in which the resulting magnetic moment of the magnetic particle is influenced by thermal fluctuations. When considering the temperature values of the magnetization of lanthanum-strontium manganite samples, it was found that with an increase of temperature in the low-temperature region, magnetic ordering is disturbed due to the excitation of magnons with a quadratic dependence of the energy from the wave vector, the number of which increases in proportion to T^3/2. This results in a decrease in the manganite magnetization. The observed temperature dependence of the magnetization measured in the “field-cooling mode” was approximated taking into account the quadratic and non-quadratic dispersion laws of the magnon spectrum.

doped manganites, oxygen nonstoichiometry, temperature dependence of magnetization, Curie temperature, Bloch constant, exchange interaction constant.

Noticeable interest to doped manganites with general formula La_xR_1-xMnO_3-δ (where R is a rare earth element), as strongly correlated electronic systems, can be associated with the existence of competing electron-electron and electron-phonon interactions that contribute to formation of spatially separated ferromagnetic and antiferromagnetic regions [1–3]. The presence of orbital and charge ordering in such systems stimulates the appearance of giant magnetoresistance, spin-polarized electric transport, and other practically important characteristics [4–9]. The most promising is the partially substituted lanthanum strontium manganite of the composition La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ, which is characterized by the maximum values of the magnetoresistive effect near the “ferro-paramagnetic” phase transition. With the partial substitution of Fe³⁺ cations to Mn³⁺, competition is observed between the exchange interactions Fe³⁺—Mn³⁺ (antiferromagnetic) and Mn³⁺—Mn⁴⁺ (ferromagnetic), which is caused by double Zener exchange [10–13]. Such manganite has a unique relationship between electrical, magnetic and other properties. It is also characterized by oxygen nonstoichiometry and is promising for use as a cathode material [10–14].

In early studies [10–15] it was found that for La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ, there is a transition from a paramagnet to a ferromagnet at the Curie temperature T_С. At the same temperature range, the high-temperature “insulating” phase undergoes a transition to the “metallic” phase. To explain this behavior, Zener double exchange mechanism was introduced [10–15], which later turned out to be insufficient to explain the large increase in resistance in the high-temperature region and its sharp drop in the low-temperature region. It was suggested that other effects resulting from Jahn–Teller distortions of the Mn³⁺ ion (electron-phonon, electron-magnon, and electron-polaron scattering of free charge carriers) may contribute to the electrical resistance [10–15]. In this case, the transformation of Mn³⁺—O—Mn⁴⁺ chain, in which a double exchange is realized, can also occur due to the replacement of trivalent manganese by trivalent iron [5, 10–15]. It is also interesting because Fe doping is possible to a large extent, and Fe cation has uncompensated spins [10–15]. With the partial substitution of Fe for Mn in La_0.7Sr_0.3Mn_1–xFe_xO₃, a significant change in the density of states near the Fermi level and a decrease in T_C values are observed due to both the antiferromagnetic interaction between Fe³⁺ and Mn³⁺ ions and the effect of cationic disordering in the Mn site.

It is known that in La_0.7Sr_0.3MnO_3–δ – manganites the value of δ cannot be higher than 3, since at δ > 3 the perovskite structure is unstable and attempts to obtain compositions with δ > 3 lead to the appearance of vacancies in the manganese and rare-earth sublattices [16–26]. In this case, the physico-chemical properties of manganite with the composition La_0.6Sr_0.4MnO_3–δ largely depend on the oxygen nonstoichiometry, which affects the oxidation state of manganese (Mn⁴⁺ and Mn³⁺ with electronic configurations t³_2ge_g¹ (S = 2) and t³_2ge_g⁰ (S = 3/2), respectively) and electronic exchange between Mn³⁺ and Mn⁴⁺ [16–26]. Crystal lattice distortions caused by defectiveness in the anion sublattice affect the bonds and the spatial arrangement of the Mn⁴⁺—O—Mn³⁺ chains, changing the magnitude of exchange interactions that depend both on the overlap of electron orbitals and on the bond angle between them. In this case, with a change in the oxygen deficiency δ and Mn cations, the sign of the exchange constant J_n,_n+p, which is included in the Heisenberg-type Hamiltonian, changes:

where the spins of S_n and S_n+p cations n and n + p are located at the nearest neighboring nodes. In this case, the value of the constant J_n,_n+p is determined by the superexchange interaction through the p_σ and p_π states of the О^2– anions [16–26].

To purposefully optimize the production of La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ with specified magnetic properties, it is important to establish the relationship between the macroscopic and microscopic parameters of the material, at which the saturation magnetization, and the general course of the temperature dependence are determined by microscopic parameters: the Bloch constant (В_T) and exchange interaction (A). The main point in calculating the above parameters, which are the key characteristics that determine the applied properties of a magnet, is the assumption that the easy magnetization axes in La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ – grains are oriented randomly. In addition, in the ensemble of polydisperse particles there is a scatter in size, which affects the uncontrollability and irreproducibility of their properties. Despite the fact that the magnetic properties of La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ have been widely studied by various authors, nevertheless, to date, there is no complete clarity in their understanding. Thus, it has not yet been established how the oxygen nonstoichiometry of the La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ – samples affects their magnetic characteristics. In this regard, the study of oxygen desorption and magnetic properties will make it possible to control the conditions for optimal saturation and uniform oxygen distribution in the anionic sublattice and obtain reproducible magnetic characteristics necessary to increase the service life of sensor devices based on them.

When preparing La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ – solid solutions, La₂O₃, Mn₂O₃, Fe₂O₃ metal oxides and strontium carbonate SrCO₃ of the “high-purity” grade have been used. Heat treatment of the samples was carried out in resistive thermal units, the temperature in which was maintained using a high-precision temperature controller RIF-101 and controlled by a Pt–Pt/Rh (10%) thermocouple with an accuracy of ±0.5 K. To remove crystallization moisture, the initial chemical compounds were kept in a thermal installation for 10 h at a temperature of 1120 K. A stoichiometric mixture of the starting metal oxides and strontium carbonate was stirred in the ethyl alcohol and dried at 370 K until the alcohol was completely removed. Preliminary annealing was carried out in air at 1170 K for 18 h. Secondary grinding was used to increase the homogenization of the charge. Then the powder was pressed into tablets with a diameter of 10 mm and a thickness of 4–5 mm under a force of 1500 kg. The synthesis of La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ – samples was carried out in air at 1770 K with holding for 3 h, followed by cooling in the switched-off thermal installation mode.

The phase composition and crystal lattice parameters were determined using the ICSD – PDF2 database (Release 2000) and the PowderCell software [27], FullProf [28] by the Rietveld method based on X-ray diffraction data. Diffraction patterns were recorded at room temperature at a rate of 60°/h in the range of angles θ = 10–90° using the DRON-3 setup in CuK_α-radiation. According to the data of X-ray phase analysis, the single-phase composition of the La_0.7Sr_0.3Mn_0.95Fe_0.05O_2.98 – samples was established (Fig. 1).

Figure 1.

XRD pattern of the La_0.7Sr_0.3Mn_0.95Fe_0.05O_2.98 sample with a rhombohedral structure (R3c).

The investigation of the nature of oxygen desorption by lanthanum-strontium manganite, as well as the preparation of samples with the required oxygen content was carried out by means of the thermogravimetry (TGA) at a heating rate of 2.5 K/min in a flowing gas stream of 1% H₂ + Ar using a SETARAM SetSys 16/18 system. A sign of reaching equilibrium was the coincidence of the sample mass at the same temperatures during the increase and decrease of the temperature.

Scanning electron microscopy (SEM) investigations have been carried out by means of the Vega 3 Tescan set up, suitable for operation in both high –vacuum and low –vacuum modes. The set up is equipped with LaB6 filament with best resolution of 2 nm at 30 kV in high-vacuum mode and 2.5 nm at 30 kV in low-vacuum mode. Magnification from 4× to 1 000 000× was selected.

The magnetic characteristics of the samples were studied on a Cryogenic Limited universal set up. The temperature dependences of magnetization were measured in two modes, with preliminary cooling from 500 to 4.2 K in a magnetic field (FC – field cooling) or without it (ZFC – zero-field cooling), followed by heating to 500 K in a magnetic field of 0.86 T

In the work, the dynamics of changes in the oxygen index (3 – δ) was investigated during heating of the samples at a rate of 2.5 deg/min to a temperature of 1270 K, followed by holding until thermodynamic equilibrium with the gas phase was established (Fig. 2).

Figure 2.

Change in the value of the oxygen index (3 – δ) and its derivative d(3 – δ)/dt during thermal action on the sample La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ at heating rates of 2.5 deg/min (a) and 9 deg/min (b) in a gas flow of 1% H₂/Ar.

Upon heating at a rate of υ = 2.5 deg/min, the first minimum of the time derivative of the oxygen index T₁{min (d(3 – δ)/dt)_2.5deg/min} = 739 K was observed with a smooth transition to the second weak one, which stood out as an independent extremum at T₂{min (d(3 – δ)/dt)_2.5deg/min} = 889 K. With an increase in the heating rate from 2.5 deg/min to 9 deg/min, the temperature of the first minimum shifts towards higher temperatures and corresponds to T₁{min (d(3 – δ)/dt)_9deg/min} = 882 K. In this case, the presence of the second minimum was not detected, but only a slight bend has been observed at T₂{min (d(3 – δ)/dt)_9deg/min} = 987 K.

When considering the amount of desorbed oxygen during heating from 300 K to 1270 K and holding until the thermodynamic equilibrium of the sample with the gas phase was established, it was found that the values (3 – δ)_300→1270 increased with a decrease of the heating rate to υ = 2.5 deg/min, while at υ = 9 deg/min the rate of oxygen evolution decreased. This dependence (3 – δ)_300→1270 = f (υ) is most likely due to the appearance of additional kinetic difficulties in the diffusion of oxygen in La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ.

To substantiate the effect of the concentration of anionic vacancies on the oxygen mobility in iron-doped lanthanum-strontium manganite, the features of defect formation in it can be considered. Anionic vacancies (V) are formed in the crystal lattice of La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ when the oxygen index (3 – δ) is less than three, with a simultaneous charge redistribution between the Mn³⁺ and Mn⁴⁺ cations. Taking into account the observance of electroneutrality during the reduction of lanthanum-strontium manganite cations, the quasi-chemical reaction of defect formation can be written in the following form Eq. (1):

(1)

It can be seen from Eq. (1) that with the increase of δ the concentration [Mn⁴⁺] = 0.3 – 2δ decreases, and [Mn³⁺] = 0.65 + 2δ increases, leading to an increase in exchange interactions responsible for antiferromagnetic properties.

Let us consider various forms of oxygen arrangement in the La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ compound. The first and most reactive form is oxygen adsorbed by the grain surface, for desorption of which the samples were preliminarily annealed at 1170 K in an argon flow for 2 h. The remaining three forms: 1 – lattice oxygen, bound with trivalent and tetravalent manganese (moreover, Mn⁴⁺ cations cause two different forms of oxygen); 2 – excess superstoichiometric oxygen, partially compensating for the presence of Mn⁴⁺ cations and 3 – oxygen, restoring electroneutrality with a decrease in charge due to the introduction of Sr²⁺. The oxygen bound to the Fe³⁺ cation was not considered in frames of the current investigation because of its low concentration. Based on the above, it can be assumed that T₁ (d(3 – δ)/dt)_2.5deg/min is due to the release of superstoichiometric oxygen and a decrease in the [Mn⁴⁺] concentration. With an increase in temperature, a significantly lower min T₂ (d(3 – δ)/dt)_2.5deg/min is observed, at which the bonds of anions with an octahedron are broken, in the center of which Mn⁴⁺(6) is located. This is apparently due to the fact that the force of electrostatic repulsion between the anions is higher than in octahedra with Mn³⁺(6) due to the difference in cationic radii (r (Mn³⁺(6)) = 0.0645 nm, r (Mn⁴⁺(6)) = 0.0530 nm) [16–17]. The appearance of defects promotes the redistribution of the electron density, the reduction of the manganese cation Mn⁴⁺, and the formation of Mn³⁺ in the pentahedral environment of the ligands with r (Mn³⁺(5)) = 0.0580 nm. Upon reduction, the effective ionic radii of manganese cations, r (Mn⁴⁺(6)) = 0.0530 nm, r (Mn³⁺(6)) = 0.0645 nm and r (Mn³⁺(5)) = 0.0580 nm differ significantly; therefore, an increase in the [Mn³⁺] radius in the octahedral and pentahedral environment of the ligands leads to an increase in the molar volume of manganite within the existence of a structure with the RС symmetry, as indicated by the by the XRD phase analysis data. Since an increase in the molar volume of manganite is observed during oxygen desorption, a stressed state in grains is formed in La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ, during annealing, which leads to a decrease in the mobility of during the reduction of cations according to the Mn⁴⁺ + e– → Mn³⁺ scheme. In this case, the rate of oxygen desorption is determined by its diffusion in the stressed regions of the grains. This explains the decrease in the amount of released oxygen with an increase in δ and the heating rate of La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ samples (Fig. 2).

Following the presented above, the effect of oxygen nonstoichiometry on the magnetic properties of manganite can be considered. Figure 3 shows the temperature dependence of the magnetization М (Т) of the La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ system with different 3–δ oxygen content in a magnetic field Н = 8.6 kOe (Fig. 3).

It was found that the temperature dependence of the magnetization obeys the Curie–Weiss law and the values of the magnetization parameters are presented in Table 1.

Figure 3.

Temperature dependence of the magnetization of the compositions La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ: δ = 0.02 (a), δ = 0.08 (b), δ = 0.15 (c), δ = 0.18 (d), investigated in magnetic field 8.6 kOe.

It can be seen from Figs 3 and 4, that at low temperatures the magnetization weakly depends on T. However, with temperature increase, a decrease begins on the M (T) curves, which continues in a wide temperature range. Obviously, the concept of the Curie temperature T_C of samples in an insulating magnetically two-phase state, such as the studied compositions La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ, is rather arbitrary. It is also incorrect to determine T_C from the magnetization in weak magnetic fields, since this magnetization is mainly due to the demagnetizing factor. The demagnetizing factor of the ferromagnetic phase depends on its configuration, which changes with temperature. Therefore, as the Curie temperature, the temperature obtained by extrapolating the steepest part of the M (T) curve, measured at the magnetic field strength H = 8.6 kOe, before its intersection with the temperature axis. Т_C values are presented in the Table 1 for all studied samples. Table 1 shows that as the oxygen deficiency increases, the Curie temperature decreases. It was found that above Т_C there is a “tail” of magnetization, while the difference in magnetization values at 300 and 5 K increases with decreasing oxygen index (Figs 3 and 4). This is additional evidence of the magnetic inhomogeneity of the samples.

Figure 4.

Temperature dependence of magnetization М (Т) for samples La_0.7Sr_0.3Mn_0.95Fe_0.05O_2.98 (a), La_0.7Sr_0.3Mn_0.95Fe_0.05O_2.92 (b), La_0.7Sr_0.3Mn_0.95Fe_0.05O_2.85 (c), La_0.7Sr_0.3Mn_0.95Fe_0.05O_2.82 (d), measured in a magnetic field H = 8.6 kOe in ZFC and FC modes, where black circles are experimental data, red line is approximation of experimental data by function (3).

Considering the low-temperature region of the М (Т) dependence, measured in the ZFC mode, it can be assumed that at Т < Т_В (where Т_В is the blocking temperature), the particles are in a frozen ferromagnetic state (Figs 3 and 4). The presence of ferromagnetism at Т ˃ Т_В leads to a magnetically ordered state, in which the resulting magnetic moment of the magnetic particle is influenced by thermal fluctuations. It is shown that with an increase in oxygen deficiency, the Т_В values decrease, which indicates an increase in the effect of magnetocrystalline anisotropy.

When considering the temperature values of La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ sample magnetization, it was found that with an increase of temperature in the low-temperature region, magnetic ordering is disturbed due to the excitation of magnons with a quadratic dependence of the energy on the wave vector E (k) ~ k², the number of which increases in proportion to T^3/2, leading to a decrease in the magnetization of the manganite (Fig. 3). In this case, the temperature dependence of the magnetization, according to Bloch’s law, should be presented in the following form Eq. (2):

M (T) = M_Т0 (1 − В_TT^3/2), (2)

where M_Т0 is the average magnetization of ferrimagnetic particles at 4.2 K, В_Т is an adjustable coefficient corresponding to the Bloch constant. It was found that the best approximation according to Bloch’s law of the dependence M (T) measured in an external magnetic field of 0.01 T was realized in the temperature range 4.2 < T < 100 K. An increase of temperature promotes the excitation of magnons with large values of the wave vector k, characterized by non-square dispersion law and interacting with each other. In this regard, it is necessary to use corrections that take into account such effects. Dyson [29] showed that deviation from Bloch’s law can be described in the model by including the term C_TT^5/2 according to equation Eq. (3):

M (T) = M_Т0(1 − B_TT^3/2 − C_TT^5/2), (3)

where all the coefficients are positive and the third term is related to the non-square dispersion law of the magnon spectrum.

The exchange interaction constant A in the La_0.7Sr_0.3Mn_0.95Fe_0.05O_3-δ samples was determined from the Eq. (4):

(4)

k _B is the Boltzmann constant, g = 2.02 is the Lande factor, μ_В is the Bohr magneton. The values obtained for magnets with different oxygen index contents are presented in Table 1. It was found that, according to Eq. (4), the exchange interaction constant A in the La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ compound decreases with increasing oxygen nonstoichiometry.

In frames of the presented work an increase in the molar volume of manganite is observed during oxygen desorption. It results in a stressed state formation in grains of La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ compound during annealing. This leads to a decrease in the mobility of oxygen vacancies during the reduction of cations according to the scheme Mn⁴⁺ + e– → Mn³⁺. In this case, the rate of oxygen desorption is determined by its diffusion in the stressed regions of the grains. This explains the decrease in the amount of released oxygen with an increase of δ and the heating rate of the La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ samples.

It was established that in the low-temperature region of the М (Т) dependence measured in the ZFC mode at Т < Т_В the particles are in a frozen ferromagnetic state. The presence of ferromagnetism at Т ˃ Т_В leads to a magnetically ordered state, in which the resulting magnetic moment of the magnetic particle is influenced by thermal fluctuations. It was found that with an increase in oxygen deficiency, the Т_В values decrease, which indicates an increase in the effect of magnetocrystalline anisotropy.

It was found that the observed temperature dependence of the magnetization measured in the FC mode was approximated taking into account the quadratic and non-quadratic dispersion laws of the magnon spectrum. The calculated values of the Bloch coefficients and the exchange interaction constant indicate their dependence on the composition of the magnetic.

It was demonstrated that the exchange interaction constant A in the La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ compound decreases with increasing oxygen nonstoichiometry. In this work, the magnetic characteristics of manganites La_0.7Sr_0.3Mn_0.95Fe_0.05O_3–δ, which are important for practical applications, are calculated, which can serve as a reference point in the development of new memory elements, magnetic field sensors with enhanced magnetic characteristics.

A support of the work in frames of the European Union project H2020-MSCA-RISE-2018-823942 – FUNCOAT and in frames of the project of the Belarusian Republican Foundation for Fundamental Research No. F21ISR-004 are gratefully acknowledged.