Magnetocaloric effect and corrosion resistance of La(Fe, Si)₁₃ composite plates bonded by different fraction of phenolic resin

Nowadays, the magnetic refrigeration technology based on magnetocaloric effect (MCE) exhibits a brighter future than that of existing gas compression-expansion refrigeration method due to its clean and efficient energy conversion process.^1–3 Among various magnetocaloric materials, La(Fe, Si)₁₃-based compounds are considered to be the most promising materials for their giant MCEs and tunable Curie temperatures (T_C).^4–7 Hydrogenation of La(Fe, Si)₁₃-based compounds is usually used to tune their T_C to around room temperature.^8,9 Recently, La(Fe, Si)₁₃-based compounds are used in some experiments for the prototype of magnetic refrigerator.^2,3 However, the intrinsic brittleness of the hydrides hinders the formability, leading to the difficulties in practical application for magnetic refrigerator. The polymer binder is used to overcome the obstacles.^10,11 It is worth noticing that the heat exchange media in the magnetic refrigeration technology is water and the magnetocaloric materials are immersed into water for refrigeration.^12,13 Thus, the corrosion of magnetocaloric materials in water is directly related to their performance and lifetime. The investigation of corrosion behavior for polymer bonded La(Fe, Si)₁₃H_x composites in water is of great significance. Though the corrosion mechanism of La(Fe, Si)₁₃-based compounds have been elucidated,^14,15 few literatures reported the corrosion behavior of polymer bonded La(Fe, Si)₁₃-based composite material in water. In this work, phenolic resin was chosen as the binder and the influence of binder fraction on the corrosion behaviors and MCEs of La_0.7Ce_0.3Fe_11.495Mn_0.105Si_1.4H_y composite plates were investigated.

The ingot having the nominal composition of La_0.7Ce_0.3Fe_11.495Mn_0.105Si_1.4 was synthesized by induction melting furnace. La_0.7Ce_0.3Fe_11.495Mn_0.105Si_1.4 thin strip was prepared by strip-casting process from the ingot. In order to form a matrix NaZn₁₃-type 1:13 phase in the strips, the thin strips were annealed at 1393 K for 1 day in an argon atmosphere. The structure of the strips was identified by a X-ray powder diffractometer. The diffraction spectrum shows that the main phase of the strips is 1:13 phase. Then the strips were crushed into particles (≤ 0.18 mm). The particles were annealed in a high purity hydrogen atmosphere to make the hydrogen atoms into the 1:13 phase lattice interval. The hydrogen content of the particles was measured to be 1.48 atom per formula unit. The La_0.7Ce_0.3Fe_11.495Mn_0.105Si_1.4H_1.48 powder was mixed with different weight fractions of phenolic resin. The mixed powder was pressed into plates (φ 10 × 0.5 mm) under a pressure of 900 MPa at 423 K for 15 minutes. The microscopic morphology of sample was observed by a LEO-1450 scanning electron microscopy (SEM). A VersaSTAT MC workstation was used to conduct the electrochemical experiment. Magnetic measurements were performed by employing a VersaLab VSM system from Quantum Design Inc. All measurements were repeated for at least three times to ensure reliability of the presented results.

XRD pattern of La_0.7Ce_0.3Fe_11.495Mn_0.105Si_1.4H_1.48 powder and SEM images of bonded La_0.7Ce_0.3Fe_11.495Mn_0.105Si_1.4H_1.48 plates with 3 wt.%, 5 wt.% and 8 wt.% phenolic resin were shown in Fig. 1, respectively. The X-ray diffraction indicates that the sample consist of cubic 1:13 main phase and a small amounts of α-Fe phase in Fig. 1(a). By stereological quantitative analysis, the volume fraction of 1:13 phase was estimated about to be 81%. And the result was determined from the average of 10 SEM images to ensure reliability. It is found that 3 wt.% resin is too little to fill the pores and boundaries (black areas) between adjacent particles in Fig. 1(b). With the resin content increased from 3 wt.% to 8 wt.%, the pores and boundaries were more fully filled by more resin and behave as light areas in Fig. 1(f). The surface of plate containing 8 wt.% resin was more compact than that containing 3 wt.% and 5 wt.% resin.

To estimate the corrosion rates, the plates were immersed in static distilled water for 6, 9, 12 and 15 hours, respectively. It is found that after a 6-hour-long immersion, the average corrosion rates of plates decrease monotonically as the resin content increases from 3 wt.% to 8 wt.%, shown in Fig. 2(a). As time goes, the corrosion rates decrease due to the formation of corrosion products attaching to the surface of plates, which could prevent the plate from further corrosion. Furthermore, after being immersed in distilled water for 15 h, the corrosion rates of plates with 5 wt.% and 8 wt.% resin were 35.06% and 51.54% lower than that of plates with 3 wt.% resin, respectively. The higher the resin content, the better corrosion resistance the plate will have.

Furthermore, the open circuit potential (OCP) and potent dynamic polarization curves of the plates were measured as shown in Fig. 2(b). The corrosion current densities and corrosion potential were calculated and shown in Table I. The results show that the corrosion potential increases slightly from −0.7821 V to −0.7437 V with the increase of phenolic resin content, while the corrosion current density decreases from 1.88 μA/cm² to 0.95 μA/cm². The corrosion potential of plate containing 8 wt.% resin is the maximum, while the corrosion current density is the smallest, which proves that the resin plays an important role in improving the corrosion resistance of the plates.

Microscopic corrosion morphologies of the bonded plates were displayed in Fig. 3. Fig. 3(a) and (b) were the backscattered electron images of plates containing 3 wt.% and 8 wt.% resin, after being immersed in distilled water for 1 h. The number of corroded spots on the plate with 8 wt.% resin is obviously less than that on the plate with 3 wt.%. It could be explained that more pores and the boundaries between the particles are filled by higher resin content. The low resin content could not fill the voids between adjacent particles, leading to a poor bonding quality. Corrosion is more prone to happen at those crevices that are not filled by resin between adjacent particles. Fig. 3(c) and (d) show a SEM image and corresponding backscattered electron image of the plate with 5 wt.% resin after being immersed in distilled water for 1 h. Fig. 3(e) and (f) show a SEM image and corresponding backscattered electron image of the plate with 5 wt.% resin after being immersed in distilled water for 1.5 h. The corroded spots first appear on the boundaries between adjacent resin and particle, then appear at the cracks inside the grain. Usually, the corroded spots of La(Fe, Si)₁₃ compounds occur between the main phase and the impurity phase, or on the boundary of the grain. The results for La(Fe, Si)₁₃-based composite plates can be explained by the gaps formed between adjacent resin and the particles, and the gaps width is closer to the most sensitive width of crevice corrosion(25 μm-15 μm), while the cracks width in grain is so narrow and is not prone to corrosion. Therefore, crevice corrosion preferentially forms at the crevices between the resin and the particles. Moreover, the elemental composition of corroded spots of plates containing 5 wt.% resin after being immersed in distilled water for 1 h was determined by an energy dispersive spectrometer (EDS), shown in Table II. The oxygen content of the corroded spots is as high as 17.57 at.%. It indicates 1:13 phase occurred oxygen-absorbing corrosion. It is observed that the corrosion mainly happened on the gray area (1:13 phase) from Fig. 3(d) and Table II.

Fig. 4(a) shows the temperature dependence of volumetric magnetic entropy change under a magnetic field change of 1 T. The density of all plates was measured using the Archimedes method. The ΔS_M decreases from 63.1 mJ/cm³ K to 59.8 mJ/cm³ K as the resin content increases from 3 wt.% to 8 wt.% under 1 T magnetic field change. The reduction of volumetric ΔS_M is due to more nonmagnetic phenolic resin with low density in the bonded plate with 8 wt.% resin. However, compared to other magnetic refrigeration materials, La(Fe, Si)₁₃-based composite plates still have a large MCE in the same temperature range. It indicates a good application prospect for the plates. In Fig. 4(a), the inset shows the temperature dependences of magnetization for bonded La_0.7Ce_0.3Fe_11.495Mn_0.105Si_1.4H_1.48 with different phenolic resin content under a magnetic field of 0.01 T. The results indicate that Curie temperatures are 302 K, 304 K and 302 K of plates with 3 wt.%, 5 wt.% and 8 wt.% resin, respectively, which is defined by the minimum of dM/dT vs. T curves. In Fig. 4(b) shows the hysteresis loops below Tc (292 K) for the plates, the plates exhibit a normal hysteresis loops. Fig. 4(c) shows the magnetization isotherms of bonded La_0.7Ce_0.3Fe_11.495Mn_0.105Si_1.4H_1.48 with 8 wt.% resin. It also can been seen that the coexistence of ferromagnetic and paramagnetic phase at the temperature of 312K. S-like variation signifies the magnetic field-induced ferromagnetic-paramagnetic phase transition. On the other hand, all of the materials show the performances of negligible hysteresis and coercivity. Which is considered to be candidate for magnetic refrigerator.

Bonded La_0.7Ce_0.3Fe_11.495Mn_0.105Si_1.4H_1.48 plates containing different content (3 wt.%, 5 wt.%, 8 wt.%) of phenolic resin were papered. The phenolic resin fills the voids and boundaries between the particles in varying degrees. When the plate is immersed in water, the corroded spots begin to appear between the resin and the particle gap and then at the cracks in the grain. The corrosion rates of the plates containing 5 wt.% and 8 wt.% resin are decreased by 35.06% and 51.54%, respectively, compared with that of 3 wt.%. However, the volumetric magnetic entropy change ΔS_M decreases slightly as the content of phenolic resin increases, The ΔS_M of the plates with 3 wt.%, 5 wt.% and 8 wt.% resin are 63.1, 61.2 and 59.8 mJ/cm³ K under a low magnetic field change of 1 T, respectively. Therefore, the addition of phenolic resin increases the corrosion resistance, but also has little influence on MCE for La(Fe, Si)₁₃-based composite plates.

This work is supported by the National Science Foundation of China (51371026), The National Key Research and Development Program of China (Grant No: 2017YFB0702704).