Magnetocaloric response of binary Gd-Pd and ternary Gd-(Mn,Pd) alloys

Magnetocaloric response of binary Gd-Pd and ternary Gd-(Mn,Pd) alloys

Journal Pre-proofs Research articles Magnetocaloric response of binary Gd-Pd and ternary Gd-(Mn,Pd) alloys Piotr Gębara, Álvaro Díaz-García, Jia Yan L...

2MB Sizes 0 Downloads 22 Views

Journal Pre-proofs Research articles Magnetocaloric response of binary Gd-Pd and ternary Gd-(Mn,Pd) alloys Piotr Gębara, Álvaro Díaz-García, Jia Yan Law, Victorino Franco PII: DOI: Reference:

S0304-8853(19)32883-5 https://doi.org/10.1016/j.jmmm.2019.166175 MAGMA 166175

To appear in:

Journal of Magnetism and Magnetic Materials

Received Date: Revised Date: Accepted Date:

16 August 2019 18 November 2019 18 November 2019

Please cite this article as: P. Gębara, A. Díaz-García, J. Yan Law, V. Franco, Magnetocaloric response of binary Gd-Pd and ternary Gd-(Mn,Pd) alloys, Journal of Magnetism and Magnetic Materials (2019), doi: https://doi.org/ 10.1016/j.jmmm.2019.166175

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2019 Published by Elsevier B.V.

Magnetocaloric response of binary Gd-Pd and ternary Gd-(Mn,Pd) alloys Piotr Gębara1, Álvaro Díaz-García2, Jia Yan Law , 2, Victorino Franco2 1Institute

of Physics, Czestochowa University of Technology, Armii Krajowej 19, 42-200 Czestochowa, Poland

2Department

of Condensed Matter Physics, ICMS-CSIC, Universidad de Sevilla, P.O. Box 1065, 41080 Sevilla, Spain

Abstract: This work investigates the MCE of alloying Pd or (Mn,Pd) to Gd, which yields the formation of an extra Gd7Pd3 or Gd7(Mn,Pd)3 phase in addition to the Gd phase, forming a composite. The phase coexistence is observed from XRD and SEM/EDX results, whereby the phase fraction of secondary phase increases with the dopant content. The magnetocaloric behavior of the binary samples present two characteristic ΔSM peaks, attributed to the Curie transitions of the coexisting biphases (separated by ΔTC=45 K). Two minima are observed from the exponent n of field dependence of ΔSM, reinforcing the presence of the Curie transitions of the two phases. The largest phase proportion of Gd7Pd3 observed in Gd80Pd20 sample gives rise to the largest RC value, which is also 10% increase compared to single phase Gd. Keywords: XRD studies, magnetocaloric effect, Gd-based composites 1. Introduction The application of non-environmental friendly green-house refrigerant gases and the relatively low energy-efficiency of conventional refrigerators (based on the compression and expansion of gases) prompt for the search for alternate cooling methods near room temperature [1]. One well-known option is the magnetic cooling technology, which is based on the magnetocaloric effect (MCE) of solid magnetic refrigerants. This is achieved by the reversible adiabatic temperature that is induced when a varying magnetic field is applied to a magnetic material under adiabatic conditions. For that, magnetic cooling is more energetically efficient and also avoids the use of harmful refrigerant gases. The benchmark magnetocaloric material (MCM) is Gd, whose magnetic entropy change (ΔSM) is ~10 J (kg K)-1 for ~5T [2], despite that there are 

Corresponding author: [email protected]

1

upcoming ones which can exhibit giant MCE (GMCE), such as Gd5Si2Ge2 [3], La(Fe,Si)13 [4,5], Heusler alloys [6-8], etc. These GMCE materials typically undergo a first-order phase transition, giving rise to an abrupt MCE in a narrow temperature span, which is usually accompanied by hysteresis and rate-dependent limits. Unlike them, the other type of MCM undergo a second-order phase transition (like Gd), leading to a non-hysteretic and moderate MCE in a wide temperature span, in which the enhanced reversibility is more suitable for some applications. The refrigerant capacity of this class of MCM can be optimized by appropriate selection of coexisting phases in composites and their transition temperatures in order to lead to table-like MCE responses (constant values in the operational temperature range), which are desirable for Ericsson refrigeration cycle [5,9], provided that the magnetocaloric peak of the composite remains large enough. Caballero et al. report that the refrigerant capacity can be improved up to 83% with proper phase selection of the composite system: optimum ∆TC (difference in the Curie temperatures) of the coexisting phases and their phase fractions [10]. Hence, there have been several works reporting the enhancement of refrigerant capacity in Gdbased magnetocaloric composites (since Gd is the reference MCM material near room temperature) by adopting this strategy. Law et al. showed the possibility of one stage production of biphasic materials based on Gd with the inclusion of a second GdZn phase (∆TC= 26 K) leading to up to 45 % improvement in the refrigerant capacity when compared to that of single phases (Gd or GdZn) [9]. Mo et al. investigated the MCE in Gd55Co35Mn10 ribbons [11]. They identified a coexistence of two phases with different TC, which enables a table-like shape of the temperature dependence of magnetic entropy change. On the other hand, Jayaraman et al. reported successive studies of partial substitution of Gd by Mn [12]. Their XRD studies reveal an occurrence of only Gd phase, whose lattice parameters contract with increasing Mn content. Moreover, a monotonic decrease of the ΔSM is detected with increasing Mn additions. The MCE in binary GdPd alloys was reported to be two times larger than that of pure Gd by Huo et al. [13]. Moreover, as was shown in Ref. [14], that small addition of Pd rises significantly the MCE in half-Heusler MnCoGe alloy. Hence in this work, appropriate selection of Pd additions to Gd for fabricating a magnetocaloric composite with coexistence of Gd and Gd7Pd3 phases (for ∆TC ~ 40 K) is being explored and studied for their MCE. In addition, with the aim to tune the TC of the phases, Mn doping to GdPd is also explored.

2

2. Experimental Ingot samples of the binary Gd100-xPdx and ternary Gd100-x(Mn,Pd)x alloys (with nominal x=5, 10, 15, 20 at. %) were prepared by arc-melting using high purity constituent elements. The XRD studies were carried out using a Bruker D8 Advance diffractometer with Cu-Kα radiation and semiconductor detector LynxEye. The analysis of XRD data was also supported by Rietveld refinement using PowderCell 2.4 package [15]. The morphologies and compositions of the samples were studied using JEOL JSM 6610LV scanning electron (SE) microscope, equipped with an energy dispersive X-ray spectrometer (EDX). Magnetic measurements were carried out using a Lake Shore 7407 vibrating sample magnetometer (VSM) with applied magnetic fields up to 1.5 T. The MCE was indirectly determined from M vs. μ0H curves collected at different temperatures. The calculations of magnetic entropy change, ΔSM, were performed using Maxwell relation [16]: H

 M (T , H )  S M (T , H )   0    dH , T H 0

(1)

where T is temperature, μ0 is magnetic permeability of vacuum, H is magnetic field and M is magnetization. Magnetization curves were corrected for demagnetizing field effects before calculating the magnetocaloric response. The refrigerant capacity (RC) of the samples were calculated using the following equation [18]: RC (T , H MAX ) 

Thot

 S

M

(T , H MAX )dT

(2)

Tcold

where δT = Thot – Tcold corresponds to full width at half maximum of the magnetic entropy peak, ∆𝑆𝑝𝑘 𝑀 , and HMAX is the maximum value of the magnetic field. 3. Results and discussion The SEM/EDX results of the as-cast samples show the presence of two phases: Gd and a secondary phase of Gd7Pd3 or Gd7(Mn,Pd)3 (depending on the series). As an example of the SEM/EDX results, those of Gd95Pd5 and Gd95(Mn,Pd)5 samples are presented in Figs. 1 (a) and (b) respectively. “Islands” that are constituted purely by Gd are observed for both cases. In addition, these “islands” are surrounded by the secondary phases, whose measured compositions reveal Gd7Pd3 or Gd7(Mn,Pd)3 depending on the series. For other alloys, their 3

EDX results of the distinguishable areas of each sample, which correspond to the different phases, are summarized in Table 1.

Fig. 1 SEM/EDX results of: (a) Gd95Pd5 and (b) Gd95(Mn,Pd)5. Table 1. EDX results of the phases from the distinguishable areas of the studied alloys.

Sample

Phase

Gd95Pd5

Gd Gd7Pd3

Gd90Pd10

Gd Gd7Pd3

Gd85Pd15

Gd Gd7Pd3

Gd80Pd20

Gd Gd7Pd3

Weight fraction (wt. %) Gd – 100 Gd – 82.66 Pd – 17.34 Gd – 100 Gd – 79.19 Pd – 20.81 Gd – 100 Gd – 74.15 Pd – 25.85 Gd – 100 Gd – 77.89

Atomic fraction (at.%) Gd – 100 Gd – 76.33 Pd –23.67 Gd – 100 Gd – 72.02 Pd – 27.98 Gd – 100 Gd – 65.99 Pd – 34.01 Gd – 100 Gd – 70.44 4

Gd95(Mn,Pd)5

Gd Gd7(Mn,Pd)3

Gd90(Mn,Pd)10

Gd Gd7(Mn,Pd)3

Gd85(Mn,Pd)15

Gd Gd7(Mn,Pd)3

Gd80(Mn,Pd)20

Gd Gd7(Mn,Pd)3

Pd – 22.17 Gd – 100 Gd – 86.75 Mn – 2.70 Pd – 10.55 Gd – 100 Gd – 91.47 Mn – 1.33 Pd – 7.20 Gd – 100 Gd – 85.77 Mn – 2.72 Pd – 11.50 Gd – 100 Gd – 73.98 Mn – 3.68 Pd – 22.35

Pd – 29.56 Gd – 100 Gd – 78.81 Mn – 7.02 Pd – 14.17 Gd – 100 Gd – 86.39 Mn – 3.59 Pd – 10.05 Gd – 100 Gd – 77.57 Mn – 7.05 Pd – 15.38 Gd – 100 Gd – 62.95 Mn – 8.95 Pd – 28.10

The observed biphasic is also found in the XRD results (Fig. 2), which show diffraction peaks corresponding to the coexistence of hexagonal Gd (P63/mmc[194]) and hexagonal Gd7Pd3 (Th7Fe3-type, P63mc[186]) phases (while the observed cubic GdO1.5 (Fm3m) phase could be attributed to surface oxidation). For the analyses on the phase proportions and lattice parameters, the XRD data are further investigated by Rietveld refinement and the results are shown in Fig. 3 and Table 2. One example of the XRD refinement results for each series is also shown in Figs. 3 (a) and (b) while the compositional dependence of the recognized phases are presented in Figs. 3 (c) and (d). For the binary alloy series, it is clearly seen in Fig. 3 (c) that the volume fraction of Gd7Pd3 phase increases with Pd content. This is also observed in Fig. 3 (d) for the ternary compositions, in which the amount of Gd7(Mn,Pd)3 phase increases with (Mn,Pd) dopant concentration (but of a smaller proportion compared to that of Gd7Pd3 for Gd100-xPdx series). For the lattice parameters, no significant changes are observed.

Fig. 2 The XRD patterns of: (a) Gd100-xPdx and (b) Gd100-x(Mn,Pd)x alloys.

5

Fig. 3 Example of Rietveld refinement for (a) Gd95Pd5 and (b) Gd80(Mn,Pd)20. Phase proportions obtained from Rietveld analysis as a function of nominal dopant content for: (c) Gd100-xPdx and (d) Gd100-x(Mn,Pd)x alloys. Table 2. Results of the Rietveld analysis.

Sample Gd95Pd5

vol. % Lattice parameters (Å)

Gd 63 a = 3.639 c = 5.781

Gd90Pd10

vol. % Lattice parameters (Å) vol. % Lattice parameters (Å) vol. % Lattice parameters (Å) vol. % Lattice parameters (Å) vol. % Lattice parameters (Å) vol. % Lattice parameters (Å) vol. % Lattice parameters (Å)

59 a = 3.629 c = 5.752 49 a = 3.617 c = 5.740 45 a = 3.630 c = 5.745 78 a = 3.614 c = 5.768 77 a = 3.635 c = 5.778 71 a = 3.583 c = 5.683 65 a = 3.624 c = 5.739

Gd85Pd15 Gd80Pd20 Gd95(Mn,Pd)5 Gd90(Mn,Pd)10 Gd85(Mn,Pd)15 Gd80(Mn,Pd)20

Recognized phases Gd7Pd3 GdO1.5 31 6 a = 9.975 a = 5.305 c = 6.275 34 a = 9.961 c = 6.286 44 a = 9.893 c = 6.362 46 a = 9.974 c = 6.281 14 a = 9.927 c = 6.267 16 a = 9.974 c = 6.277 19 a = 9.899 c = 6.282 24 a = 9.925 c = 6.298

7 a = 5.294 7 a = 5.312 9 a = 5.294 8 a = 5.391 7 a = 5.309 10 a = 5.371 11 a = 5.296

6

The M(T) curves for Gd100-xPdx as presented in Fig. 4 (a) show a gradual two-step decrease of M with T, which indicates the presence of two Curie transitions arising from non-interacting magnetic phases. Their first derivative dM/dT shown in Fig. 4 (b) reveals two minima at ~287 K and ~337 K, which agrees with the TC of Gd and Gd7Pd3 phases, respectively (a temperature increment of 5 K was used for the measurements). Furthermore, for higher Pd content, it can be seen that the M(T) slope at higher temperatures gets more prominent (Fig. 4 (a)) and the minimum at ~337 K becomes deeper (Fig. 4 (b)), whereby both indicate the increase in the Gd7Pd3 phase proportion. These results agree with those earlier observed from SEM/EDX and XRD refinement. On the other hand, for the ternary Gd100-x(Mn,Pd)x series, two-step M(T) and two minima are not observed in Figs. 4 (c) and (d) despite that the secondary Gd7(Mn,Pd)3 phase is detected by SEM/EDX and XRD. This could arise from the low content of the secondary phase in Gd100-x(Mn,Pd)x, which is observed from the Rietveld refinement.

Fig. 4 (a) M(T) and corresponding (b) dM/dT curves for Gd100-xPdx; (c) M(T) and corresponding dM/dT curves for Gd100-x(Mn,Pd)x for field of 0.1 T.

7

The temperature dependence of ΔSM for Gd100-xPdx and Gd100-x(Mn,Pd)x alloys is shown in Figs. 5 (a) and (b), respectively. Two minima in the vicinity of 287 and 332 K, which correspond to TC of Gd and Gd7Pd3 phase [17], are clearly visible for Pd content of x=10, 15 and 20 at. %. In the case of Gd95Pd5, a deep minimum occurs at ~287 K while a small hump is slightly visible at 327 K. This indicates the detection of two non-interacting Gd and Gd7Pd3 phases. In the studied series, the largest ΔSM peak obtained (2.76 J kg-1 K-1, which corresponds to 0.43 J mol-1 K-1) is observed for Gd95Pd5. Moreover, a successive decrease of the ΔSM peak corresponding to Gd phase (∆𝑆𝑝𝑘 𝑀, 𝐺𝑑) with increasing Pd content is observed, while that corresponding to Curie transition of Gd7Pd3 (∆𝑆𝑝𝑘 𝑀, 𝐺𝑑7𝑃𝑑3) increases. This shows the increasing fraction of the latter phase with Pd content, which is in agreement with SEM/EDX and XRD results. For Gd100-x(Mn,Pd)x, only one ∆𝑆𝑝𝑘 𝑀 is observed for all the compositions as shown in Fig. 5 (b) and it decreases with increasing (Mn,Pd) dopant concentration. A second peak has not been found for the ternary series, which may be due to different reasons: a) the fact that the concentration of the Gd(Mn,Pd) phase is very low, b) the transition of this phase might occur at lower temperatures than the measurement range or c) because the phase is paramagnetic.

Fig. 5 Magnetic entropy change as a function of temperature for (a) Gd100-xPdx and (b) Gd100-x(Mn,Pd)x alloys for 1 T.

For RC comparison of the composites, the influence of magnetic phase proportions (data taken from Rietveld refinement results) on the RC values are presented in Fig. 6 ([GdO1.5] is neglected as it is not contributing to the studied MCE range). It is observed that Gd80Pd20 sample exhibits the largest RC, which is also 10 % larger than that of single phase Gd. It is the only sample exhibiting the highest content of the higher TC phase (which is the secondary Gd7Pd3 phase), which amounts to 53 wt. % of the composite. This is in agreement to the literature [10], wherein it reports that majority phase of the magnetocaloric composite for RC enhancement should 8

comprise of the higher temperature phase (i.e. for a composite of phases A and B and the latter exhibits larger TC,

[𝐵] [𝐴]

> 0.5 is needed for RC improvement). In addition, the RC calculations

of Gd80Pd20 sample also get contributions from ∆𝑆𝑝𝑘 𝑀, 𝐺𝑑7𝑃𝑑3 as the FWHM range expands. On the other hand, the RC of the ternary compositions decrease with Gd7(Mn,Pd)3 phase content as the phase optimization ratio is unfulfilled. For the binary alloys, RC decreases from the lower Pd dopant contents till that of Gd85Pd15 sample, in which it starts to increase due to the higher Gd7Pd3 phase content.

Fig. 6 RC of the composites (1 T) vs. phase proportion of the higher TC phase (neglecting [GdO1.5] as it is not contributing to the MCE). Solid symbols: Gd100-xPdx; open symbols: Gd100-x(Mn,Pd)x. Individual shapes of the symbols correspond to nominal dopant added.

Franco and coworkers in Refs. [19,20] reported that the ΔSM is strongly magnetic field dependent via the following power-law expression: Δ𝑆𝑀 ∝ 𝐻𝑛

(3)

where the exponent n can be locally determined as: 𝑛=

𝑑|ln Δ𝑆𝑀| 𝑑|ln H|

.

(4)

9

The exponent n is strongly dependent on the magnetic state of the material. When the sample follows the Curie-Weiss law, its value corresponds to 2 for paramagnetic state, while it is ~ 1 for ferromagnetic state. Based on critical scaling laws for MCE, it is possible to estimate n at the TC with the following expression: n  1

1  1  1    

,

(5)

where β and δ are critical exponents. In the case of a mean field theory, the critical exponents are β=0.5 and δ=3, which gives a value of n at TC equal to 2/3. Using equations (3) and (4), the temperature dependence of exponent n for studied alloys is presented in Figs. 7 (a) and (b) for binary Gd100-xPdx and ternary Gd100-x(Mn,Pd)x respectively. For the former, the value of n is slightly larger than 1 in the ferromagnetic region, which can be due to the paramagnetic contribution of the oxide, then proceeds to two distinct minima (one at ~ 288 K and another at ~ 338 K, which corresponds to the TC of Gd and Gd7Pd3 phases) before increasing up to 2 in the paramagnetic region. Moreover, the minimum corresponding to Gd7Pd3 phase in Gd95Pd5 sample can be observed, indicating the phase presence though it is not evident in its ΔSM(T) curve. Overall, the minimum corresponding to Gd7Pd3 phase increases with Pd additions, indicating the increasing amount of Gd7Pd3 phase with Pd doping. These are all in agreement with the results observed from ΔSM(T), M(T), SEM/EDX and XRD data. For Gd100-x(Mn,Pd)x series, similar n values were observed except that only a minor shoulder is observed at ~ 222-238 K, together with one minimum at ~ 288 K. The low content of the secondary phase could be the attributing factor for the absence of its corresponding minimum.

Fig. 7 Temperature dependence of exponent n for (a) Gd100-xPdx and (b) Gd100-x(Mn,Pd)x alloys for 1 T.

10

4. Conclusions In this paper, the phase composition and magnetocaloric properties of Gd100-xPdx and Gd100-x(Mn,Pd)x alloys were studied. The phase coexistence of hexagonal Gd and hexagonal Gd7Pd3 (or Gd7(Mn,Pd)3) is observed from XRD and SEM/EDX results. The phase fraction of secondary phase increases with the dopant content. The magnetic investigations also reveal the Curie temperatures of both phases coexisting in the binary composites. Their ΔSM(T) curves show two peak values (separated by 45 K), which arise from the Curie transitions of Gd and Gd7Pd3. The phase proportions of the latter phase (which exhibits larger TC) increases with Pd content, arising up to 53 wt.% for x=20, which shows the largest RC value (also 10% increase compared to single phase Gd). On the other hand, very low phase fraction of the secondary Gd7(Mn,Pd)3 phase is attained in the ternary alloys, which in turn, is unable to contribute to ΔSM(T) and RC. Acknowledgements This study was supported by AEI/FEDER-UE (project MAT-2016-77265-R), the PAI of the Regional Government of Andalucía and by the Rector of Częstochowa University of Technology – Prof. Norbert Sczygiol. 5. References [1] V. Franco, J. Blázquez, J. Ipus, J.Y. Law, L. Moreno-Ramírez, A. Conde, Magnetocaloric effect: From materials research to refrigeration devices, Prog Mater Sci 93 (2018) 112-232 [2] K. A. Gschneider, V. K. Pecharsky, “Magnetic refrigeration in Rare Earths”, Science Technology and Applications III ed R. G. Bautista, C. O. Bounds, T. W. Ellis, B. T. Killbourn (The Minerals, Metals and Material Society) p. 209 [3] V. K. Pecharsky, K. A. Gschneidner, Appl. Phys. Lett. 70 (1997) 3299 [4] X.B. Liu, Z. Altounian, J. Magn. Magn. Mater. 264 (2003) 209–213. [5] P. Gębara, P. Pawlik, J. Magn. Magn. Mater. 442, 145-151, (2017). [6] Z. Yang, D.Y. Cong, L. Huang, Z.H. Nie, X.M. Sun, Q.Z. Zhang, Y.D. Wang, Mater. Des. 92 (2016) 932–936. [7] Y.W. Li, H. Zhang, K. Tao, Y.X. Wang, M.L. Wu, Y. Long, Mater. Des. 114 (2017) 410– 415. [8] O. Tegus, E. Brück, K.H.J. Buschow, F.R. de Boer, Nature 415 (2002) 150–152. 11

[9] J.Y. Law, L.M. Moreno-Ramirez, J.S. Blazquez, V. Franco, A. Conde, J. Alloys Compd. 675 (2016) 244. [10] R. Caballero-Flores, V. Franco, A. Conde, K. E. Knipling, and M. A. Willard, Appl. Phys. Lett. 98 (2011) 102505. [11] H.Y. Mo, X.C. Zhong, D.L. Jiao, Z.W. Liu, H. Zhang, W.Q. Qiu, R.V. Ramanujan, Phys. Lett. A 382 (2018) 1679-1684. [12] T.V. Jayaraman, L. Boone, J.E. Shield, J. Magn. Magn. Mater. 345 (2013) 153-158. [13] J.J. Huo, Y.S. Du, G. Cheng, X.F. Wu, L. Ma, J. Wang, Z. Xia, G. Rao, J. Rare Earth 36 (2018) 1044-1049. [14] P. Gębara, Z. Śniadecki, J. All. Compd. 796 (2019) 153 – 159. [15] W. Kraus, G. Nolze, Powder Diffr. 13 (1998) 256. [16] V.K. Pecharsky, K.A. Gschneidner Jr., J. Magn. Magn. Mater. 200 (1999) 44. [17] F. Canepa, M. Napoletano, S. Cirafici, Intermetallics 10 (2002) 731-734. [18] M.E. Wood, W.H. Potter, Cryogenics 25 (12) (1985) 667-683. [19] V. Franco, J.S. Blazquez, A. Conde, J. Appl. Phys. 100 (2006) 064307. [20] V. Franco, C.F. Conde, A. Conde, L.F. Kiss, J. Appl. Phys. 101 (2007) 093903.

Highlights     

Gd-Pd magnetocaloric composites of Gd:Gd7Pd3 are obtained by adding Pd to Gd Their Curie transitions with ∆TC ~ 45 K are observed from magnetocaloric results Table-like magnetocaloric effect is obtained for the composites Refrigerant capacity increases when sufficient amount of Gd7Pd3 is obtained Field dependence exponent n explicitly shows the presence of the secondary phase

Piotr Gębara: conceptualization, alloy preparation, microstructural characterization. Álvaro DíazGarcía: VSM measurements, magnetocaloric analyses. Jia Yan Law, Victorino Franco: visualization, supervision. Piotr Gębara, Álvaro Díaz-García, Jia Yan Law, Victorino Franco: writing - reviewing and editing. Piotr Gębara, Victorino Franco: funding acquisition.

12

The authors declare that there is no conflict of interest.

13