Critical behavior and magnetocaloric effect in magnetic Weyl semimetal candidate Co2−xZrSn
Yu Tianlin, Yu Xiaoyun, Yang En, Sun Chang, Zhang Xiao, Lei Ming
State Key Laboratory of Information Photonics and Optical Communications & School of Science, Beijing University of Posts and Telecommunications, Beijing 100876, China

 

† Corresponding author. E-mail: zhangxiaobupt@bupt.edu.cn mlei@bupt.edu.cn

Abstract

We investigate the critical exponents and magnetocaloric effect of Co2−xZrSn polycrystal. The Co2−xZrSn undergoes a second-order ferromagnetism phase transition around the Curie temperature of Tc ∼280 K. The critical behavior in the vicinity of the magnetic phase transition has been investigated by using modified Arrott plot and Kouvel–Fisher methods. The obtained critical exponents, β, γ, and δ can be well described by the scaling theory. The determined exponents of Co2−xZrSn obey the mean-field model with a long-range magnetic interaction. In addition, the maximum magnetic entropy change of Co2−xZrSn is about and the relative cooling power (RCP) is at 50 kOe ( ).

1. Introduction

In the last decade, intensive research of the topology of electronic structures has deepened our understanding of the exotic physical properties of novel condensed matters and established a whole new classification. After the discovery of topological insulators,[1,2] topological semimetals (TSMs) have attracted considerable attention in recent years. The TSMs are characterized by the symmetry protected bulk band crossings near the Fermi energy EF.[36] According to the degeneracy of nodes, they can be classified into Dirac semimetals (DSMs), Weyl semimetals (WSMs), multi-degenerate semimetals, and so on.[7] Among these TSMs, DSMs have Dirac nodes that can be viewed as two Weyl nodes with opposite chirality degenerate at the same k-point. Such a degeneracy can be lifted by breaking either inversion symmetry (IS) or time-reversal symmetry (TRS), resulting in two types of WSMs; i.e., nonmagnetic noncentrosymmetric WSMs and magnetic WSMs.[7] The first class of WSMs has been theoretically predicted and experimentally confirmed.[810] Magnetic WSMs have been predicted in several materials, such as pyrochlore iridates, Heusler compounds, and the related materials such as Heusler alloys Co2TiX (X = Si, Ge, or Sn), Co2MnAl, and half-Heusler compound GdPtBi, etc.[4,1114]

Recently, Heusler compound Co2ZrSn has also been predicted to host Weyl fermions.[15] Moreover, it belongs to the family of 26 valence electron Heusler systems, which are known to exhibit half-metallic ferromagnetism (HMFM) with saturated magnetic moments of per formula unit.[16,17] According to the classification proposed by Coey,[18] Co2ZrSn is a type of IA HMFM, in which the spin-majority band is crossing EF (metallic) and there is a band gap for the spin-minority band. Because it is an HMFM compound with Curie temperature (Tc) near/higher than room temperature,[19] besides the unique feature of harboring Weyl fermions, it is vitally important to the spintronics applications due to the nearly 100% spin polarization of conduction electrons at EF.[20] Previous studies on Co2ZrSn single crystals show that there are some vacancies in the position of Co with the content of Co varying from 1.59 to 1.84. With higher Co content, the sample exhibits higher Tc and the Tc is above room temperature when the content of Co is larger than 1.70.[19] Meanwhile, when Cu is doped into the Co site, the Tc of Co2−yCuxZrSn decreases linearly and the low-temperature spin glass behavior appears when y is close to 1.[21] This trend can be explained by the weakening of half-metallicity of Co2ZrSn with the Cu substitutions, leading to the breakdown of the ferromagnetism.[21] To understand the itinerant ferromagnetism in Co2ZrSn further, a comprehensive investigation of phase transition and critical exponents is urgently needed, which will not only provide an insight into the range of ferromagnetic interaction but also the symmetry of the order parameter within a given universality class.[22,23] Among the various properties of ferromagnets, magnetocaloric effect (MCE) is another important feature that can be used for magnetic refrigeration, which is an environmentally friendly energy-conversion technology.

In this paper, the magnetic critical behavior and magnetocaloric effect of Co2−xZrSn in the vicinity of the paramagnetic–ferromagnetic (PM–FM) phase transition region are studied in detail. It is found that Co2−xZrSn undergoes a second-order phase transition with a long-range order in the nature of magnetic interaction. Moreover, the obtained critical exponents are found to follow the theoretically predicted values for mean-field model. Further analysis of the field-dependent magnetic entropy change indicates that the maximum magnetic entropy change of Co2−xZrSn is about at 50 kOe.

2. Experimental

Polycrystal Co2−xZrSn was synthesized using arc-melting method under an Ar atmosphere. High purity Zr particles and Co pieces were melted first to achieve a homogeneous mixture of the compound metal. Then, the obtained alloy was brought close to the Sn shots and remelted. The sample was flipped three times to avoid breaking up over a long time of arc-melting. Finally the arc-melted sample was cooling down naturally. The crystal structure and phase purity were identified by powder x-ray diffraction (pXRD) using the Cu–Kα radiation of Bruker D8. Rietveld refinements of the data were performed with the TOPAS package.[24] The magnetization measurements were conducted using a Quantum Design MPMS3.

3. Results and discussion

The XRD patterns of Co2−xZrSn can be readily indexed to the crystal structures of Heusler L21 phase (Fm3m) (Fig. 1(a)). No obvious characteristic peaks of other impurities are detected. The intense diffraction peaks suggest that the as-synthesized products are well crystallized. The fitted lattice parameter a is 6.21266(8) Å. When compared to previous single crystal results, it suggests that the value of x in Co2−xZrSn is about 0.33.[19] The temperature dependence of magnetization M(T) at H = 1 kOe with zero-field-cooling (ZFC) and field-cooling (FC) models are plotted in Fig. 1(b). The M(T) curves show a typical PM–FM phase transition. Inset of Fig. 1(b) shows the Tc is ∼280 K, which is defined as the minimum value of the . This is consistent with previous results for Co2−xZrSn single crystals.[19] The overlap of ZFC and FC M(T) curves suggests the absence of spin-glass behavior in Co2−xZrSn. Figure 1(c) shows the magnetization hysteresis loop of Co2−xZrSn measured at 2 K. The saturated moment Ms is about /f.u., in agreement with that of Co2−xZrSn single crystals.[19] Meanwhile, the magnetic coercivity (Hc) of Co2−xZrSn is smaller than 50 Oe, which indicates the Co2−xZrSn is a super soft magnet.

Fig. 1. (a) The powder XRD pattern and Rietveld refinement of Co2−xZrSn. Vertical tick marks is Bragg reflections in the Fm3m space group. (b) Temperature dependence of magnetization at H = 1 kOe with ZFC and FC modes. Inset: as a function of temperature near Tc. (c) Isothermal M(H) loop at 2 K. (d) Isothermal magnetization curves from 260 K to 300 K with 2-K temperature interval.

To further understand the fundamental feature of the FM phase transition in Co2−xZrSn, we measured the isothermal magnetization curves M(H) from 260 K to 300 K with temperature step of 2 K as shown in Fig. 1(d). According to the mean-field theory, the ferromagnet undergoes a second order phase transition with long range order interactions, the Landau free energy G(T,M) could be expressed as:[25]

where A and B are temperature-dependent parameters. When the free energy G(T,M) is minimal, ( ) = 0, the equation of state can be expressed as:

Based on this equation, we can plot a Arrott plot of M 2versus H/M, which should be a series of straight lines, and the positive intercepts of these straight lines on the H/M axis determine the second-order magnetic transition. But we should analyze the M(H) results for the short-range FM phase transition systems by using the Arrot–Noaks equation of state , where is the reduced temperature, and a and b are constants.[26] We constructed the Arrott plots, using 3D XY model (β = 0.345, γ = 1.316), 3D Heisenberg model (β = 0.365, γ = 1.386), 3D Ising model (β = 0.325, γ = 1.24), and mean field model (β = 0.5, γ = 1.0), as given in Fig. 2. All these models yield quasi-straight lines in the high-field region, the positive intercepts reveals a second-order PM–FM transition in Co2−xZrSn. However, none of the models yield a series of parallel straight lines in the high-field range and the line at Tc passes through the origin. Therefore, it is difficult to distinguish which model is the best for the determination of critical exponents.

Fig. 2. The MAP of (a) 3D XY model, (b) 3D Heisenberg model, (c) 3D Ising model, and (d) mean-field model.

To determine better parameters, we calculated their normalized slopes (NS) which are defined at the critical point as , where S(T) is the slope of versus at high field. Because the most appropriate model should be the one close to 1.0 in the vicinity of Tc,[27] the mean-field is the best model for Co2−xZrSn system, as shown in Fig. 3. However, the Arrott plot with the critical exponents β = 0.5 and γ = 1 could not lead to a set of parallel lines in the high magnetic field region. According to the scaling hypothesis, critical exponents β and γ can be briefly described by the following equation,[25,28]

where Ms(T) is the spontaneous magnetization below Tc, the inverse initial susceptibility above Tc, is the reduced temperature and M0, , and D are the critical amplitudes. Starting from the initial trial values of β = 0.5 and γ = 1, Ms(T) and obtained from the linear extrapolation are then fitted to Eqs. (2) and (4), respectively, yielding new values of β, γ and Tc. Repeat the iterative process until β and γ are stable. The final modified Arrott plot (MAP) generates the values β = 0.41(1) and γ = 1.08(9), and two sets of Ms(T) and with the fitting curves are shown in Fig. 4(a). The fixed Tc values from two curves are consistent with each other (280.1(8) K and 280.3(7) K). Thus, the Tc of Co2−xZrSn is 280 K, which is consistent with the values obtained from the M(T) curve. A further possibility to evaluate the values of β, γ, and Tc is given by the Kouvel–Fisher (KF) method,[29]

Fig. 3. Temperature dependence of the normalized slopes .
Fig. 4. (a) The MAP plot for Msversus T (left) and versus T (right). (b) The KF plot for versus T (left) and versus T. (c) The plot of the isotherm M versus H around T = 280 K. Insert: log(M) versus log(H). (d) Scaling plots of M versus H below and above Tc = 280 K. Isert: log (M) versus log (H).

Figure 4(b) shows that the critical exponents obtained from the KF method are β = 0.415(2) with Tc = 280.1(7) K and γ = 1.08(5) with Tc = 280.4(1) K. These are consistent with the values derived from the modified Arrott plots. The critical exponent δ can be determined by the critical isotherm analysis from M(H) at Tc following Eq. (5). The isothermal magnetization M(H) at Tc = 280 K is shown in fig 4(c), where the inset plots with the logarithmic scale. The inset shows that the log(M)–log(H) relation yields straight line at higher field range with the slope 1/δ, where δ = 3.53 is obtained. According to the statistical theory, the critical exponents should agree with the Widom scaling relation:[30]

Using the β and γ values determined from the modified Arrott plot and Kouvel–Fisher (KF) plot, we obtain δ = 3.66 and δ = 3.61, respectively, which close to the above value of δ. Furthermore, there is an important criterion for the critical regime based on the prediction of scaling hypothesis,[25]

where is used for and is used for . Then, the precise value of β and γ are used in fitting the plot of Fig. 4(d). the as a function of plots in the critical region should collapse into two universal curves: one above Tc and another below Tc. The emergence of two convergence curves in Fig. 4(d) indicate the critical exponents are correct and reliable.

The universality class of the magnetic phase transition depends on range of the exchange distance J(r). It decays as , where σ is the range of the interaction, r is the distance, d is the dimensionality, and b is the spatial scaling factor. The the parameter of σ can be calculated by the equation,[31] where n is the spin dimensionality, , . According to the renormalization group analysis, the spin interactions can be classified by the value of σ. The implies long-range spin interaction dominated in the system, while implies the short-range spin interaction. In the Co2−xZrSn system, the experimental value of γ is 1.26, for any combination of {d:n} the values of σ are always smaller than 2. This suggests that the long-range spin interaction plays an important role in the Co2−xZrSn system.

To gain more information on critical properties of Co2−xZrSn, we have studied the magnetic entropy change . According to thermal dynamics, the magnetic entropy change is defined as,[32]

For the magnetization measured at small discrete fields and temperature intervals, can be approximately expressed as,[33]

where Mi and are the magnetization at Ti and under the same magnetic field, respectively. The calculated results for have been shown in Fig. 5(a). It is obviously that the values of at different magnetic fields reach the maximum in the vicinity of Tc. The maximum magnetic entropy change is about ( ) for H = 50 kOe. Figure 5(b) depicts the values of vary with the magnetic fields. Generally, for the long-range spin interaction mean-filed model, obeys the power law of . Fitting of the gives that n = 0.696(3), confirming that the critical behavior of Co2−xZrSn can be depicted by mean-field theory.

Fig. 5. (a) The trend plots of with temperature variation at the different magnetic fields. (b) Field dependence of . The solid line represents a fit of to the data. (c) Field dependence of RCP.

The relative cooling power (RCP) is also an important indicator for characterizing the magnetothermal material, its mathematical definition is,[34]

where the RCP is the function of T with and the full width at half maximum . The values of calculating the RCP at different fields are shown in Fig. 5(c). With the increasing of magnetic field, the RCP curve shows a linear trend. The RCP at H = 50 kOe is about ( ). This value is smaller than those typical magnetocaloric materials[32] because of the relatively small . However, because the Tc of Co2−xZrSn is close to or above room temperature, it would be interesting to develop a method that can improve the magnetocaloric property of Co2−xZrSn in the future, such as the chemical doping method.

4. Conclusion

In summary, the magnetic properties of Co2−xZrSn polycrystal are studied in detail. It is found that the PM–FM transition in Co2−xZrSn polycrystal is the second-order phase transition. The critical exponents β, γ, and δ obtained from the MAP and KF methods are consistent each other. These parameters conform to the scaling rate very well. In addition, these critical exponents suggest that the magnetic transition of Co2−xZrSn polycrystal can be described by the mean-field model. Moreover, the and RCP are about and at H = 50 kOe, respectively.

Reference
[1] Hasan M Z Kane C L 2010 Rev. Mod. Phys. 82 3045
[2] Qi X L Zhang S C 2011 Rev. Mod. Phys. 83 1057
[3] Burkov A A Hook M D Balents L 2011 Phys. Rev. 84 235126
[4] Wan X Turner A M Vishwanath A Savrasov S Y 2011 Phys. Rev. 83 205101
[5] Xu G Weng H Wang Z Dai X Fang Z 2011 Phys. Rev. Lett. 107 186806
[6] Wang Z Sun Y Chen X Q Franchini C Xu G Weng H Dai X Fang Z 2012 Phys. Rev. 85 195320
[7] Weng H M Dai X Fang Z 2016 J. Phys.: Condens. Matter 28 303001
[8] Weng H M Fang C Fang Z Bernevig B A Dai X 2015 Phys. Rev. 5 011029
[9] Xu S Y Belopolski I Alidoust N et al 2015 Science 349 613
[10] Lv B Q Weng H M Fu B B Wang X P Miao H Ma J Richard P Huang X C Zhao L X Chen G F Fang Z Dai X Qian T Ding H 2015 Phys. Rev. 5 031013
[11] Chang G Xu S Y Zheng H Singh B Hsu C H Bian G Alidoust N Belopolski I Sanchez D S Zhang S Lin H Hasan M Z 2016 Sci. Rep. 6 38839
[12] Manna K Sun Y Muechler L Kübler J Felser C 2018 Nat. Rev. Mat. 3 244
[13] Kübler J Felser C 2016 Europhys. Lett. 114 47005
[14] Hirschberger M Kushwaha S Wang Z Gibson Q Liang S Belvin C A Bernevig B A Cava R J Ong N P 2016 Nat. Mater. 15 1161
[15] Wang Z Vergniory M G Kushwaha S Hirschberger M Chulkov E V Ernst A Ong N P Cava R J Bernevig B A 2016 Phys. Rev. Lett. 117 236401
[16] Fecher G H Kandpal H C Wurmehl S Felser C Schönhense G 2006 J. Appl. Phys. 99 08J106
[17] Kandpal H C Fecher G H Felser C 2007 J. Phys. D: Appl. Phys. 40 1507
[18] Coey J M D Venkatesan M 2002 J. Appl. Phys. 91 8345
[19] Kushwaha S K Stolze K Wang Z Hirschberger M Lin J Bernevig B A Ong N P Cava R J 2017 J. Phys.: Condens. Matter 29 225702
[20] De Groot R A Mueller F M Engen P G Buschow K H 1983 Phys. Rev. Lett. 50 2024
[21] Kushwaha S K Wang Z Kong T Cava R J 2018 J. Phys.: Condens. Matter 30 075701
[22] Sondhi S L Girvin S M Carini J P Shahar D 1997 Rev. Mod. Phys. 69 315
[23] Banerjee B K 1964 Phys. Lett. 12 16
[24] Bruker AXS. TOPAS Version 4.2 2009 Bruker AXS, Karlsruhe, Germany
[25] Stanley H E 1971 Introduction to Phase transitions and critical phenomena London Clarendon Press 7
[26] Arrott A Noakes J E 1967 Phys. Rev. Lett. 19 786
[27] Fan J Ling L Hong B Zhang L Pi L Zhang Y 2010 Phys. Rev. 81 144426
[28] Fisher M E 1967 Rep. Prog. Phys. 30 615
[29] Kouvel J S Fisher M E 1964 Phys. Rev. 136 A1626
[30] Widom B 1965 J. Chem. Phys. 43 3898
[31] Fischer S F Kaul S N Kronmuller H 2002 Phys. Rev. 65 064443
[32] Gschneidner K A Jr. Pecharsky V K Tsokol A O 2005 Rep. Prog. Phys. 68 1479
[33] Phan M H Yu S C 2007 J. Magn. Magn. Mater. 308 325
[34] Han H Menzel D Liua W Ling L Dua H Pi L Zhanga C Zhang L Zhang Y 2017 Mater. Res. Bull. 94 500