Effect of Yttrium on Microstructure and Magnetocaloric Properties in La_1−xY_xFe_11.5Si_1.5 Compounds

Abstract

The effects of Y on the phase composition and magnetocaloric effect in La_1−xY_xFe_11.5Si_1.5 compounds were studied. Y₂Fe₁₇-type phase and α-Fe phase appear in the annealed La_1−xY_xFe_11.5Si_1.5 compounds when x ≥ 0.07 due to small solid solubility of Y in NaZn₁₃ phase. Y₂Fe₁₇ phase obstructs the formation of the 1:13 phase, leading to the decrease of magnetic entropy changes. But for x < 0.1, La_1−xY_xFe_11.5Si_1.5 compounds exhibit high magnetic entropy changes and low hysteresis loss compared with that of LaFe_11.5Si_1.5. Consequently, the La_1−xY_xFe_11.5Si_1.5 compounds (x < 0.1) are useful to realize large magnetocaloric effect with smaller hysteresis loss.

1. Introduction

LaFe_13−xSi_x compounds with a NaZn₁₃-type crystal structure (denoted as 1:13 phase) showed first-order magnetic phase transitions at a Curie temperature (T_C), which induced a giant magnetocaloric effect (MCE) when LaFe_13−xSi_x compounds underwent a change from a ferromagnetic state to a paramagnetic state at T_C [1]. Additionally, the peak value of the magnetic entropy change (ΔS_M) with temperature variation broadened to just above T_C for LaFe_13−xSi_x compounds because of the itinerant-electron magnetic transition [2]. These features make LaFe_13−xSi_x compounds ideal refrigerants for magnetic refrigeration technology [3,4]. However, LaFe_13−xSi_x compounds display large magnetic and thermal hysteresis due to the first-order nature of their transitions. The magnetic hysteresis decreases the actual efficiency of magnetic refrigeration [5]. Thermal hysteresis would also make the adiabatic temperature change (ΔT_ad) unstable as the refrigeration cycle runs [6]. Thus, the developments of the refrigerants with a large magnetic entropy change and small hysteresis losses are still the key for practical applications [7,8]. Partial substitution of Fe by Co in LaFe_13−xSi_x compounds could completely eliminate hysteresis and thermal hysteresis. However, the order of the magnetic phase transition changed from first to second-order type, and the magnetic entropy change decreased greatly near a Curie temperature by Co additions [9,10]. Grinding the compounds into micron scale particles could also reduce the hysteresis [11], but too small particles were easily oxidized because the LaFe_13−xSi_x compounds consisted of the rare earth La. In addition, the addition of carbon in LaFe_13−xSi_x systems effectively led to a reduction in magnetic hysteresis loss [12].

On the other hand, the influence of different element additions in LaFe_13−xSi_x compounds on phase structure and MCE have also been extensively studied to optimize the MCEs of LaFe_13−xSi_x compounds. Ni, Cr, and Nb additions had limited solubility in the 1:13 phase. Higher contents of the additions caused a large residual of impure phases after annealing, and a decrease of magnetic entropy change [13,14]. The substitution of rare earth Ce and Pr for La resulted in the enhancement of the magnetic entropy change. Ce and Pr additions had higher solubility in the 1:13 phase compared with the additions of Ni, Cr, and Mn [15,16]. Until now, little knowledge on the partial substitution of La by other rare earth elements with no orbital angular momentum contribution, like Y, in the LaFe_13−xSi_x compounds has been reported.

In the present work, the effect of a partial substitution of Y for La in LaFe_11.5Si_1.5 on the phase composition and magnetocaloric effect are investigated.

2. Materials and Methods

La_1−xY_xFe_11.5Si_1.5 (x = 0, 0.03, 0.07, 0.15, 0.2) ingots were prepared by arc melting of the stoichiometric mixtures of constituent elements under an argon atmosphere. The purity of starting materials was 99.4 wt % for La and Y (obtained from the Materials Preparation Center of Ames Laboratory), 99.9 wt % for Fe (purchased from Alfa Aesar Inc., Haverhill, MA, USA), and 99.999 wt % for Si (purchased from Alfa Aesar Inc.). The samples were melted four times to ensure homogeneity. The samples were wrapped in Ta foils, and sealed in helium-filled quartz tubesusing Mo Kα radiation at room temperature. The microstructural characterization was performed using JEOL 5910 scanning electron microscopy (SEM, JEOL USA, Inc., Ames, IA, USA) with an energy dispersive spectrometer (EDS).

Magnetic measurements were performed using an MMPS XL-7 superconducting quantum interference device magnetometer manufactured by Quantum Design, Inc. Initial magnetization, M(T), was measured at varying temperatures at a constant magnetic field of 0.01 T in order to identify the Curie temperatures (T_Cs). The magnetization as a function of applied magnetic field, M(H), was measured isothermally near the corresponding T_Cs.

The isothermal magnetic entropy change, ΔS_M(T, H), was estimated from isothermal magnetization curves using the following equation [17]:

$$\Delta S_M{\left(T_{\mathrm{av}}\right)}_{\Delta H}=\frac{\delta H}{2\delta T}\left(\delta M_1+2{\displaystyle \sum }_{k=2}^{n-1}\delta M_k+\delta M_n\right)$$

(1)

Here, δT = T_u − T_l is the temperature difference between the two isothermal magnetization curves, n is the number of points measured for each of the two isotherms with the magnetic field changing from H₁ = H_I to H_n = H_F at δH = ΔH/(n − 1), and δM_k = M(T_u)_k− M(T_l)_kis the difference in the magnetization at T_u and T_l for each magnetic-field step from 1 to n.

3. Results and Discussion

3.1. Crystal Structure and Morphologies

Figure 1a–f demonstrates the back scattered electron (BSE) images of the as-cast La_1−xY_xFe_11.5Si_1.5 (x = 0, 0.03, 0.07, 0.1, 0.15, 0.2) ingots. From Figure 1 and the EDS data, the black parts correspond to the α-Fe phase, the white parts represent the LaFeSi phase, and the gray parts correspond to the Y₂Fe₁₇ phase. The microstructures of the as-cast samples were composed of the LaFeSi phase and α-Fe phase from the beginning at x = 0 to x = 0.1. Y only existed in the LaFeSi phase. The Y₂Fe₁₇ phase was detected as x increased from 0.15 to 0.2, indicating that the solid solubility of Y in the LaFeSi phase was much smaller. Y element was found in both the Y₂Fe₁₇ phase and the LaFeSi phase. The Y₂Fe₁₇ phase in the ingots reduced the amount of the 1:13 phase generated from the peritectoid reaction of the LaFeSi phase and α-Fe phase during the annealing process, which will be discussed below.

Figure 2 shows the X-ray diffraction patterns of annealed La_1−xY_xFe_11.5Si_1.5 compounds. Most of the Bragg peaks observed in this figure match the NaZn₁₃-type structure angular positions. It indicates that the 1:13 phase is the main phase. For x ≥ 0.1, a certain amount of impurities related to the α-Fe phase and the Y₂Fe₁₇-type structure phase can also be observed in small sets of the Bragg peaks. The diffraction intensities from the α-Fe phase and Y₂Fe₁₇ phase increase with increasing x.

Figure 3a–f demonstrates the BSE images of La_1−xY_xFe_11.5Si_1.5 compounds annealed at 1353 K for 14 days. The white areas, grayish areas, dark gray areas, and black areas in Figure 3 correspond to the LaFeSi phase, 1:13 phase, Y₂Fe₁₇ phase, and α-Fe phase respectively. In order to find out which phase contained the Y element, an EDS analysis was performed. The results show that Y existed in the 1:13 phase and LaFeSi phase for La_0.7Y_0.3Fe_11.5Si_1.5 compounds. Also, a large amount of the Y₂Fe₁₇ phase appeared in the compounds with x greater than 0.07, as shown in Figure 3c–f. It indicates that the solid solubility of Y in the 1:13 phase was much smaller compared with Ce and Nd for the annealing compounds [16]. There was no difference for the phase compositions of x = 0 and 0.03, and very small amount of the LaFeSi phase (white area) can be observed, as shown in Figure 3a,b. The amount of the Y₂Fe₁₇ phase and α-Fe phase increased with increasing x, which agreed with the results of the XRD observations as discussed above. It’s interesting to note that no Y₂Fe₁₇ phase was observed in the as-cast samples until x > 0.1, as shown in Figure 1. It suggests that the heat treatment procedure had favored the formation of the Y₂Fe₁₇ phase when La was partly replaced by Y in La_1−xY_xFe_11.5Si_1.5 compounds. The Y₂Fe₁₇ phase would hinder the formation of the 1:13 phases during the annealing process.

3.2. Magnetic Properties

Figure 4 displays the temperature dependent magnetization measured under 0.01 T in heating and cooling processes for annealed La_1−xY_xFe_11.5Si_1.5 compounds (0 ≤ x ≤ 0.15). The T_Cs, determined from the point of the maximum of dM/dT in the heating process, were 188 K, 192 K, 188 K, 190 K, and 190 K for x = 0, 0.03, 0.07, 0.1, and 0.15 respectively. The measured T_Cs were not drastically modified with Y substitution. No regular changes were observed in T_Cs, probably due to formation of the Y₂Fe₁₇ phase and α-Fe.

The magnetizations in paramagnetic regions did not approach zero due to the existence of ferromagnetic phases, such as α-Fe and Y₂Fe₁₇ phases, in the compounds. In addition, the magnetizations above T_Cs increased with the increase of x because the amount of the ferromagnetic phases increased with x increasing.

Isothermal magnetization curves as a function of magnetic field up to 2 T for the annealed La_1−xY_xFe_11.5Si_1.5 (0 ≤ x ≤ 0.2) were measured under increasing and decreasing field strengths at various temperatures and are shown in Figure 5a–f. The magnetization curves at temperatures far above T_C show a curvature at low fields, which is caused by the α-Fe and Y₂Fe₁₇ phases as shown in the red rectangle in Figure 5c–f.

Figure 6 displays the magnetic hysteresis loss in the vicinity of T_Cs, derived from Figure 5, for the annealed La_1−xY_xFe_11.5Si_1.5 compounds. The solid line in Figure 6 is a guide for the eyes. The magnetic hysteresis loss of all Y-substituted compounds decreases compared with that of x = 0. Additionally, the hysteresis loss is obviously reduced with the increase of x. For x = 0.2, the hysteresis loss almost disappears, which is favorable for magnetic refrigeration.

Figure 7 shows the Arrott plots in the vicinity of T_C for La_1−xY_xFe_11.5Si_1.5 (0 ≤ x ≤ 0.2) compounds. The negative slopes decrease almost monotonically with the increase of the Y content from the beginning at x = 0 to x = 0.15, indicating that the feature of the first-order magnetic transition becomes weak after Y additions. A positive slope appears for x = 0.2, indicating that a second-order magnetic phase transition occurs.

Figure 8 shows the temperature dependence of ΔS_M(T, μ₀H) calculated by Equation (1) in a magnetic field change from 0 to 2 T for La_1−xY_xFe_11.5Si_1.5 compounds. The isothermal magnetic entropy change (ΔS_M) for x = 0, 0.03, 0.07, 0.1, and 0.15 are −24.1 J/kg·K, −24.3J/kg·K, −20.8 J/kg·K, −20.2 J/kg·K, and −16.9 J/kg·K respectively. No significant reduction was found in the magnetic entropy change ΔS_M by the partial substitution of Y for La in La_1−xY_xFe_11.5Si_1.5 for x ≤ 0.07. As to x > 0.07, the ΔS_M is significantly suppressed due to the formation of the α-Fe and Y₂Fe₁₇ phases.

4. Conclusions

The structural, magnetocaloric properties of La_1−xY_xFe_11.5Si_1.5 compounds were investigated. Y element existed only in the 1:13 phase, LaFeSi phase, and the Y₂Fe₁₇ phase in the La_1−xY_xFe_11.5Si_1.5 compounds. The solid solubility of Y in the NaZn₁₃ phase in the annealed La_1−xY_xFe_11.5Si_1.5 compounds was much smaller compared with Ce and Nd. The Y₂Fe₁₇ phase had a detrimental effect on the peritectoid reaction, resulting in a decrease in magnetic entropy change. The measured T_Cs were not drastically modified with Y substitution. The hysteresis loss decreased with increasing Y content in the compounds. High magnetic entropy changes and low hysteresis loss were obtained by minor amount of Y substitution for part of La in LaFe_11.5Si_1.5 compounds.