Trans. Nonferrous Met. Soc. China 27(2017) 2015-2021
Effects of solidification rate and excessive Fe on phase formation and magnetoclaoric properties of LaFe11.6xSi1.4
Xiang CHEN1, Yun-gui CHEN2, Yong-bai TANG2, Ding-quan XIAO2
1. College of Physics and Electronic Information Engineering, Neijiang Normal University, Neijiang 641002, China;
2. School of Materials Science and Engineering, Sichuan University, Chengdu 610065, China
Received 23 May 2016; accepted 31 December 2016
Abstract: The effects of solidification rate and excessive Fe on phase formation and magnetocaloric properties of LaFe11.6xSi1.4 (x=1.1, 1.2) were investigated by XRD, SEM and VSM measurements. The XRD results show that the amount of LaFeSi phase in the as-cast melt-spun ribbons prepared by a copper wheel at a speed of 10 m/s is less than that in the as-cast arc melting buttons with the same x values. The annealed melt-spun ribbons contain smaller amount of La(Fe,Si)13 (1:13) phase than the corresponding annealed arc melting buttons. Although the melt-spun sample has finer crystalline grains of α-Fe, as indicated by SEM analysis, its crystalline size has not reached nano-scale. Therefore, the magnetic exchange-coupling between 1:13 phase and α-Fe phase has not been observed in melt-spun ribbons. Further, the maximum negative magnetic entropy change (-SMax) and relative cooling power (RCP) of annealed melt-spun ribbons under a field change of 0-2 T are weaker than those of the corresponding annealed arc melting buttons.
Key words: LaFe11.6xSi1.4 alloy; solidification rate; microstructure; magnetocaloric properties
1 Introduction
Due to the coupling between field-induced paramagnetic (PM)-ferromagnetic (FM) transition and the discontinuous change of lattice constant, LaFe13-xSix alloys with cubic LaCo13-type structure (hereinafter called 1:13 phase) exhibit some interesting physical properties, such as giant magnetocaloric effect (MCE). To maximize their MCE, researchers have tried their best to eliminate the impurities and obtain LaFe13-xSix with single 1:13 phase. However, due to the intrinsic characteristics of peritectic reaction Fe+LaFeSi→ La(Fe,Si)13, α-Fe and LaFeSi phases co-exist propor- tionally in most annealed LaFe13-xSix samples [1-5]. It is well known that LaFeSi phase is corroded easily in some heat-conducting media [6,7], so its existence will deteriorate corrosion resistance of LaFe13-xSix alloys as magnetic refrigeration working materials. But α-Fe is a kind of soft magnetic materials and has remarkably high magnetic susceptibility; maybe a certain amount of α-Fe phase in LaFe13-xSix alloys can affect the magnetic behaviors of 1:13 phase, such as low field response characteristics and MCE. In fact, the low field response effect has been reported in Gd5Si2Ge2/Fe composite magnetocaloric alloy [8]. Our previous work showed that excessive Fe in LaFe11.6xSi1.4 buttons with x>1 prepared by arc melting results in reducing content of LaFeSi phase [9-11]. At the same time, the magnetocaloric properties of LaFe11.6×1.1Si1.4 are better than those of normal stoichiometric sample, which is an intriguing phenomenon. In addition, we found that the Curie temperatures (TC) of 1:13 phase were all about 190 K in LaFe11.6xSi1.4 alloys, and the magnetic coupling between 1:13 phase and α-Fe phase has not been observed. One possible reason is that the grain size of α-Fe phase is too large to generate strong exchange-coupling. If the grain size of α-Fe phase is decreased to a certain value, perhaps there exists a stronger exchange-coupling between magnetic atoms in 1:13 phase and α-Fe phase, which can affect TC and magnetic properties.
Compared with arc melting LaFe13-xSix buttons, normal stoichiometric melt-spun LaFe13-xSix ribbons have lower content of 1:13 phases and MCE. The high solidification rate of melt-spun process results in small grain microstructure. For obtaining nonstoichiometric LaFe11.6xSi1.4 samples with different grain sizes of α-Fe phases, we used arc melting and melt-spun methods to prepare samples in this work. Up to now, the magnetic coupling between α-Fe phase and 1:13 phase in LaFe13-xSix alloys has seldom been studied. It is very difficult to precisely explain the magnetic behavior of this complex multiphase system. Thus, we explored the effects of solidification rate and excessive α-Fe on phase formation and magnetic properties of nonstoichiometric LaFe11.6xSi1.4 samples with x=1.1 and 1.2 only by comparing phase, microstructure and MCE.
2 Experimental
Approximately 15 g polycrystalline LaFe11.6xSi1.4 buttons were prepared by using high-purity starting elements (99.4% La, 99.9% Fe, and 99.9999% Si, mass fraction). The components were arc-melted on a water- cooled copper hearth under high-purity argon atmosphere. The buttons were turned over and re-melted five times to achieve a homogeneous composition. Each button was cut into two parts on average. An half of each button was re-melted and then spun into ribbons by a single-roller melt-spinner with a copper wheel at a speed of 10 m/s under a purified argon atmosphere. As the temperature of 1523 K was close to up limit temperature of peritectic reaction, button and ribbon samples were annealed at 1523 K for 5 h in a molybdenum wire furnace with a vacuum of 3×10-3 Pa, followed by furnace cooling down to room temperature. The crystallographic structures were determined by powder X-ray diffraction (XRD) analysis using Cu Kα radiation at room temperature. The microstructure and phase composition observations were carried out by scanning electron microscopy (SEM) and energy disperse spectroscopy (EDS) (Hitachi-S-3400N). Magnetic measurements were performed in vibrating- sample magnetometer (VSM, Lakeshore 7410) under a magnetic field up to 2 T.
3 Results and discussion
3.1 XRD analysis
Figure 1 displays the XRD patterns of the as-cast LaFe11.6xSi1.4 samples with x=1.1 and 1.2 prepared by arc melting and melt-spinning, respectively. The main phase is α-Fe phase and the second phase is LaFeSi phase. The differences in solidification rates between melt-spun and arc melting processes affect phase relations and microstructure. The relative diffraction intensities of LaFeSi phase in the as-cast melt-spun LaFe11.6xSi1.4 ribbons are weaker than those in arc melting LaFe11.6xSi1.4 alloys with the same composition, which indicates that the content of LaFeSi phase is lower in ribbons. However, the diffraction peaks of 1:13 phase, which appear in the as-cast arc melting buttons, almost cannot be observed in the as-cast melt-spun ribbons. This result is not consistent with the general points that the high solidification rates of melt-spun preparation enhance the formation of 1:13 phase and increase its content. According to our previous work [12], the temperatures of 1407.6 and 1530.5 K are melting point of LaFeSi phase and temperature of peritectic reaction, respectively. When annealing temperature is between 1407 and 1530 K, peritectic reaction will occur very fast, and the optimum annealing temperature will be about 1523 K [4]. The processing time is very short at 1407-1530 K in cooling process of melt-spun preparation; there is not enough time to complete the peritectic transformation. This behavior results in the fact that the content of 1:13 phase is lower in as-cast melt-spun LaFe11.6xSi1.4 ribbons than in arc melting LaFe11.6xSi1.4 buttons. In addition, the buttons need to be re-melted in melt-spun preparation, which leads to some loss of LaFeSi phase. Thus, the content of LaFeSi phase is lower in the as-cast LaFe11.6xSi1.4 ribbons.
Fig. 1 XRD patterns of as-cast samples
Figure 2 shows the comparison of XRD patterns between arc melting and melt-spun LaFe11.6xSi1.4 samples annealed at 1523 K for 5 h. The main phase is 1:13 phase and impurity phase is α-Fe phase. The relative diffraction intensities of α-Fe phase gradually increase with the increase of x. The diffraction peaks of LaFeSi phase almost disappear in the above two kinds of annealed LaFe11.6xSi1.4 samples. By comparing with relative diffraction intensities of α-Fe phase, it can be found that the value in melt-spun LaFe11.6xSi1.4 ribbons is higher than that in arc melting LaFe11.6xSi1.4 alloys with the same x, which suggests that the annealed melt-spun ribbons contain larger amount of α-Fe phases. This can be further confirmed by Rietveld refinement result, which shows that the mass fractions of α-Fe phase in arc melting samples with x=1.1 and 1.2 are 6.5% and 10.7%, but the corresponding values in melt-spun samples are 16.0% and 18.9%, respectively. This can be attributed to the lower content of LaFeSi phase in as-cast ribbons.
Fig. 2 XRD patterns of different samples annealed at 1523 K for 5 h
3.2 Microstructures and compositions
Figure 3 shows the back-scattered SEM images of the above two kinds of LaFe11.6xSi1.4 samples. Grey matrix phase is 1:13 phase, black region is α-Fe phase and the white part is LaFeSi phase. The distribution of α-Fe phase is not homogeneous in all samples, but its grain size in melt-spun ribbons is much smaller than that in arc melting buttons. For example, the largest size in melt-spun LaFe11.6×1.1Si1.4 is only about 3 μm, the corresponding size in arc melting sample is about 20 μm. Thus, the melt-spun technology is useful for grain refinement. Unfortunately, the melt-spun preparation in this work cannot refine the grain size of α-Fe phase to the nano-scale. The desired magnetic coupling between 1:13 phase and α-Fe phase will also be affected. In addition, it is hard to determine the amount of α-Fe phase from the limit number of SEM images due to its fine grain size and anisotropic distribution in matrix phase.
Since MCE is caused by 1:13 phase in LaFe13-xSix alloys, the element ratios in 1:13 phase should have a great influence on magnetic performance. Table 1 gives mole fractions of elements in 1:13 phases of two kinds of annealed LaFe11.6xSi1.4 samples obtained by EDS analysis. It is clear that the mole fraction of Fe in those samples does not increase with the increase of x, indicating that excessive Fe does not enter into 1:13 phase but exists as α-Fe phase. All element contents are very stable in two arc melting buttons. On the contrary, the contents are not stable in melt-spun LaFe11.6×1.1Si1.4 ribbons. This shows the existence of chemical component segregation in melt-spun samples. In total, La content in melt-spun LaFe11.6×1.1Si1.4 ribbons is smaller than that in arc melting LaFe11.6xSi1.4 alloys, and Si content is just contrary. To the best of our knowledge, there are few reports about the effect of element content in 1:13 phase on magnetic properties probably due to the perceived difficulties associated with coupling of crystallographic and magnetic structures.
Fig. 3 SEM images
Table 1 Mole fractions of elements in 1:13 phases and magnetocaloric properties of LaFe11.6xSi1.4 samples
3.3 Magnetic phase transition
The temperature dependence of magnetization (M) for two kinds of annealed LaFe11.6xSi1.4 samples under zero-field-cooled warming (ZFC) and field-cooled cooling (FC) protocols in magnetic field of 0.02 T is shown in Fig. 4. The TC values of 1:13 phases are listed in Table 1, determined as a peak in dM/dT plots (not shown here). Although the grain size of α-Fe phase in annealed LaFe11.6xSi1.4 ribbons is very fine, the differences of TC for two kinds of LaFe11.6xSi1.4 samples with the same x are only 2-3 K. This indicates that the magnetic coupling between 1:13 phase and α-Fe phase is negligible in melt-spun ribbons. At the same time, thermal hysteresis phenomena exist in two LaFe11.6xSi1.4 samples, and do not decrease with the increase of α-Fe content. Thus, the excessive Fe will not change the nature of the first order magnetic phase transition in those annealed LaFe11.6xSi1.4 samples.
Figure 5 shows magnetization-magnetic field (M-H) curves of two kinds of annealed LaFe11.6xSi1.4 samples in a field change of 0-2 T. Before measurement, the samples were heated to paramagnetic state and cooled to the measuring temperature in zero field in order to ensure a demagnetized sample. Magnetic hysteresis and field-induced metamagnetic transition from PM to FM above TC of 1:13 phase are clearly observed, which are manifestations of the first order magnetic transition (FOMT). As shown in Fig. 6, the Arrott plots at and above TC show “S” shape, further confirming that the metamagnetic PM-FM transitions in those samples are of FOMT [13]. In addition, the curvature of Arrott plot can reflect the intensity of FOMT characteristic [14]. For arc melting LaFe11.6×1.1Si1.4 alloy, the Arrott plot’s curvature at TC is higher than that of other three alloys. Therefore, this alloy has the strongest first order characteristic.
Fig. 4 Temperature dependence of ZFC and FC magnetizations measured in magnetic field of 0.02 T
3.4 Magnetocaloric properties
In Fig. 5, M-H curves have bending phenomena in the low field range when temperatures are higher than TC of 1:13 phase. It is from the contribution of soft magnetic α-Fe phase, which is in ferromagnetic state in the measurement temperature range due to its high TC of ~1023 K. By linearly extrapolating between the first large susceptibility stage and the small susceptibility stage, the M of intersection is the magnetization of α-Fe phase at the measured temperature. For different samples, with the increase of α-Fe phase content, the magnetization contribution from α-Fe phase also increases. However, α-Fe soft magnetic phase has high TC, and its magnetization contribution remains almost unchanged in the measured temperature range for the same sample. We can obtain the magnetizations of 1:13 phase at different temperatures in a 2 T applied field via eliminating the magnetization contribution of α-Fe phase. For example, the magnetizations of arc melting LaFe11.6×1.1Si1.4 and LaFe11.6×1.2Si1.4 alloys are 180 and 156 A·m2/kg at 166 K, the magnetization contributions of α-Fe phases are 31 and 44 A·m2/kg, respectively. Thus, the magnetizations of 1:13 phases under 2 T are 149 and 112 A·m2/kg at 166 K for arc melting alloys with x=1.1 and 1.2, respectively. Similarly, the magnetization of 1:13 phase in melt-spun LaFe11.6×1.1Si1.4 and LaFe11.6×1.2Si1.4 ribbons can be obtained, and the values are only 107 A·m2/kg at 168 K and 96 A·m2/kg at 164 K, respectively. The magnetization of 1:13 phase in melt-spun LaFe11.6xSi1.4 ribbons is lower than that in the corresponding arc melting buttons. According to Clausius-Clapeyron △SMax=-(△M·△HC)/△T [15], the negative isothermal magnetic entropy change (-△SMax) is proportional to the magnetization change of 1:13 phase. This predicts that -△SMax of annealed arc melting LaFe11.6xSi1.4 buttons should be higher than that of the corresponding melt-spun ribbons.
Fig. 5 Isothermal magnetization curves measured in applied magnetic fields from 0 to 2 T
Fig. 6 Arrott plots derived from isothermal magnetization data
In fact, the isothermal magnetic entropy change can be calculated by Maxwell equation [16]. The curves of -△SMax versus temperature in a magnetic field change of 0-2 T are shown in Fig. 7. The peak values of -△SMax for melt-spun LaFe11.6xSi1.4 ribbons with x=1.1 and 1.2 are 17.22 and 13.2 J/(kg·K), respectively, which are obvious smaller than those of the corresponding arc melting LaFe11.6xSi1.4 alloys, as shown in Table 1, especially at x=1.1. The relative cooling power (RCP) is calculated by -△SMax·δT [17], where δT is the temperature full width at half of -△SMax. The RCPs of arc melting LaFe11.6xSi1.4 buttons with x=1.1 and 1.2 are 198.6 and 144.0 J/kg in a field change of 0-2 T, respectively. For melt-spun LaFe11.6xSi1.4 ribbons, the corresponding values are only 146.2 and 105.6 J/kg, respectively. These results show that the fine grain sized α-Fe phase does not enhance the magnetocaloric properties in melt-spun LaFe11.6×1.1Si1.4 ribbons. On the contrary, due to the reduced 1:13 phase fraction and increased density of grain boundary, the magnetocaloric effect becomes weak.
4 Conclusions
1) The content of 1:13 phases in the annealed melt-spun ribbons than that in the annealed arc melting buttons with the same composition. The main reason is
that the contents of 1:13 and LaFeSi phases in the as-cast melt-spun LaFe11.6xSi1.4 ribbons are less than those in the corresponding arc melting LaFe11.6xSi1.4 buttons.
2) The expected magnetic atom coupling between 1:13 phase and α-Fe phase is not observed in annealed melt-spun LaFe11.6xSi1.4 ribbons. The grain size of excessive α-Fe does not reach the nano-scale, although rapid solidification technology is useful for gain refinement.
3) Magnetocaloric properties of annealed melt-spun LaFe11.6xSi1.4 ribbons are weaker than those of the corresponding arc melting LaFe11.6xSi1.4 buttons. At x=1.1, the maximum of -△SMax values for melt-spun ribbon and arc melting buttons are 17.2 and 33.1 J/(kg·K) in a magnetic field change of 0-2 T, respectively.
Fig. 7 Comparison of -△SMax-T curves for LaFe11.6xSi1.4 samples in magnetic field change of 0-2 T
References
[1] CHEN Xiang, CHEN Yun-gui, TANG Yong-bai, XIAO Ding-quan. Effect of Ce, Co, B on formation of LaCo13-structure phase in La(Fe,Si)13 alloys [J]. Transactions of Nonferrous Metals Society of China, 2014, 24: 705-711.
[2] MANDAL K, GUTFLEISCH O, YAN A, HANDSTEIN A, 
K H. Effect of reactive milling in hydrogen on the magnetic and magnetocaloric properties of LaFe11.57Si1.43 [J]. Journal of Magnetism and Magnetic Materials, 2005, 290-291: 673-675.
[3] HU Feng-xia, SHEN Bao-gen, SUN Ji-rong, CHEN Zhao-hua, RAO Guang-hui, ZHANG Xi-xiang. Influence of negative lattice expansion and metamagnetic transition on magnetic entropy change in the compound LaFe11.4Si1.6 [J]. Applied Physics Letters, 2001, 78: 3675-3677.
[4] HU Feng-xia, SHEN Bao-gen, SUN Ji-rong, WANG Guang-jun, CHENG Zhao-hua. Very large magnetic entropy change near room temperature in LaFe11.2Co0.7Si1.1 [J]. Applied Physics Letters, 2002, 80: 826-828.
[5] LIU X B, ALTOUNIAN Z, TU G H. The structure and large magnetocaloric effect in rapidly quenched LaFe11.4Si1.6 compound [J]. Journal of Physics: Condensed Matter, 2004, 16: 8043-8051.
[6] ZHANG Min, LONG Yi, YE Rong-chang, CHANG Yong-qin. Corrosion behavior of magnetic refrigeration material La-Fe-Co-Si in distilled water [J]. Journal of Alloys and Compounds, 2011, 509: 3627-3631.
[7] ZHANG En-yao, CHEN Yun-gui, TANG Yong-bo. Investigation on corrosion and galvanic corrosion in LaFe11.6Si1.4 alloy [J]. Materials Chemistry and Physics, 2011, 127: 1-6.
[8] LEWIS L H, YU M H, GAMBINO M H. Simple enhancement of the magnetocaloric effect in giant magnetocaloric materials [J]. Applied Physics Letters, 2003, 83: 515-517.
[9] CHEN Xiang, CHEN Yun-gui, TANG Yong-bai. The influence of Fe on phase and magnetic property in the LaFe11.6Si1.4 compound [J]. Journal of Rare Earths, 2011, 29: 354-358.
[10] CHEN Xiang, CHEN Yun-gui, TANG Yong-bo, XIAO Ding-quan. Effects of the excess Fe on phase and magnetocaloric property of LaFe11.6*xSi1.4 alloys [J]. Journal of Rare Earths, 2015, 33: 1293-1297.
[11] CHEN Xiang, CHEN Yun-gui, TANG Yong-bai, XIAO Ding-quan. The study of phase, microstructure, and magnetocaloric properties in LaFe11.6xSi1.4B0.1 alloys [J]. Phase Transitions, 2015, 88:1045-1053.
[12] CHEN Xiang, CHEN Yun-gui, TANG Yong-bo. High-temperature phase transition and magnetic property of LaFe11.6Si1.4 compound [J]. Journal of Alloys and Compounds, 2011, 509: 8534-8541.
[13] FUJITA A, AKAMATSU Y, FUKAMICHI K. Itinerant electron metamagnetic transition in La(FexSi1-x)13 intermetallic compounds [J]. Journal of Applied Physics, 1999, 85: 4756-4758.
[14] CHEN Xiang, CHEN Yun-gui, TANG Yong-bo, XIAO Ding-quan. The system study of 1:13 phase formation, the magnetic transition adjustment, and magnetocaloric property in La(Fe,Co)13-xSix alloys [J]. Journal of Magnetism and Magnetic Materials, 2014, 368: 155-168.
[15] 
A, FOLDEAKI M, RAVI G B, CHAHINE R, BOSE T K, FRYDMAN A, BARCLAY J A. Direct measurement of the “giant” adiabatic temperature change in Gd5Si2Ge2 [J]. Physical Review Letters, 1999, 83: 2262-2265.
[16] FOLDEAKI M, CHAHINE R, BOSE T K. Magnetic measurements: A powerful tool in magnetic refrigerator design [J]. Journal of Applied Physics, 1995, 77: 3528-3537.
[17] PECHARSKY V K, GSCHNEIDNER K A Jr. Magnetocaloric effect from indirect measurements: Magnetization and heat capacity [J]. Journal of Applied Physics, 1999, 86: 565-575.
凝固速度和过量铁对LaFe11.6xSi1.4合金相形成和磁热性能的影响
陈 湘1，陈云贵2，唐永柏2，肖定全2
1. 内江师范学院 物理与电子信息工程学院，内江 641002；
2. 四川大学 材料科学与工程学院，成都 610065
摘 要：采用XRD、SEM和VSM等方法研究凝固速度和过量Fe对LaFe11.6xSi1.4(x=1.1, 1.2)试样相形成和磁热性能的影响。XRD研究结果表明，以10 m/s速度甩带制备的铸态LaFe11.6xSi1.4试样中LaFeSi相的含量低于采用电弧熔炼的相同配比铸态LaFe11.6xSi1.4纽扣试样中LaFeSi相的含量，且在对应热处理后试样中，甩带试样中的La(Fe, Si)13相(1:13相)的含量也更低。SEM结果表明，虽然甩带能使热处理后试样中α-Fe相晶粒组织细化，但并未达到纳米级，因此，未观察到1:13相和α-Fe相中磁性原子的磁耦合现象。磁热性能研究表明，经过相同热处理后甩带LaFe11.6xSi1.4试样在0~2 T磁场中的最大磁熵变和相对制冷能力也低于同成分LaFe11.6xSi1.4纽扣试样的最大磁熵和相对制冷能力。
关键词：LaFe11.6xSi1.4合金；凝固速率；显微组织；磁热性能
(Edited by Wei-ping CHEN)
Foundation item: Project (16ZB0301) supported by the Research Program of Sichuan Provincial Education Department, China
Corresponding author: Xiang CHEN; Tel: +86-832-2340497; E-mail: gxucx@163.com
DOI: 10.1016/S1003-6326(17)60226-7
