- Email: [email protected]
Large exchange bias effect in the super spin glass state of Mn50Ni38Al12 alloy
Intermetallics 86 (2017) 116e120
Contents lists available at ScienceDirect
Intermetallics journal homepage: www.elsevier.com/locate/intermet
Large exchange bias effect in the super spin glass state of Mn50Ni38Al12 alloy H. Pan a, L. Ma a, *, G.K. Li a, L.Y. Jia a, C.M. Zhen a, D.L. Hou a, W.H. Wang b, E.K. Liu b, J.L. Chen b, G.H. Wu b a b
Department of Physics, Hebei Advanced Thin Films Laboratory, Hebei Normal University, Shijiazhuang 050024, China Beijing National Laboratory for Condensed Matter Physics, Institute of Physics, Chinese Academy of Sciences, Beijing 100190, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 December 2016 Received in revised form 19 January 2017 Accepted 4 March 2017
A large exchange bias field (HEB) of 5.3 kOe which so far, is the largest value found in MnNi-based alloys, was observed in Mn50Ni38Al12 ribbon. DC and AC magnetic measurements show super spin glass (SSG) behavior below 130 K and superparamagnetic behavior above this temperature. Different from other systems, the pinning and pinned phases are both SSG. Due to the high content of Mn and low content of Al there exists considerable disorder in Mn50Ni38Al12 which is responsible for the large HEB. © 2017 Published by Elsevier Ltd.
Keywords: Exchange bias Super spin glass Heusler alloys
1. Introduction
2. Experimental details
In recent years, exchange bias (EB) effect in the multifunctional Heusler alloys have attracted significant interests owing to its potential application in giant magnetoresistance, ultrahigh-density magnetic recording and spin valve devices [1,2]. For Heusler alloys, EB was first observed in Ni50Mn36Sn14 and Ni50Mn25þxSb25-x alloys, but the exchange bias field (HEB) was just ~250 Oe [3,4]. The small HEB restricts the application of these alloys. By doping Co into these systems, the HEB was improved to several hundred Oe [5,6]. Recently, by changing the composition a large HEB has been also observed in Mn50Ni40Sn10 [7] (HEB ¼ 1170 Oe) and Mn50Ni42Sn8 [8] (HEB ¼ 3520 Oe) alloys. It is worth mentioning that a giant HEB (HEB ¼ 33 kOe) in Mn2.4Pt0.6Ga has been observed by Felser et al. [9]. Taking into account that Pt (or Ga) and Ni (or Al) are in the same main group, a large HEB may be obtained in MnNiAl system. Here in this work, based on the above idea, a large HEB of 5.3 kOe was observed in Mn50Ni38Al12. Besides, it is found that the pinning and pinned phases are both super spin glass by DC and AC magnetic measurements. Due to high content of Mn and low content of Al, there are considerable antisite disorder in MnNiAl, which is responsible for the large HEB.
The precursor polycrystalline ingot of Mn50Ni38Al12 with nominal composition was prepared by arc melting the pure metals (>99.9%) under an argon atmosphere. Subsequently, the ingot was melted in a quartz tube in an argon atmosphere and the melt spin technique used to obtain ribbons. The copper wheel rotated with a surface velocity of 25 m/s. The ribbons are hereafter referred as Al12, indicating that the Al content is 12 at. %. The crystal structure was identified by X-ray diffraction (XRD) using Cu Ka radiation. Magnetic properties were investigated utilizing a physical property measurement system (PPMS-9, Quantum Design, Inc.).
* Corresponding author. E-mail addresses: [email protected], [email protected] (L. Ma). http://dx.doi.org/10.1016/j.intermet.2017.03.003 0966-9795/© 2017 Published by Elsevier Ltd.
3. Results and discussion Fig. 1 shows XRD patterns for Al12 at room temperature. It can be seen that the sample shows a martensitic tetragonal structure (L10) with a ¼ b ¼ 5.48 Å and c ¼ 6.68 Å. Based on the valenceelectron site occupation rule in Heusler alloys, Al12 in the austenitic phase has an off-stoichiometric Hg2CuTi-type structure with 4 face center cubic sublattices (denoted as A (0,0,0), B (1/4 1/4 1/4), C (1/2 1/2 1/2) and D (3/4 3/4 3/4) along the body diagonal line), as shown in the insert of Fig. 1(b). The atom occupation form should be written as (Mn12Ni13)A (Mn25)B(Ni25)C(Al12Mn13)D for the ordered case. The additional Ni13 atoms take up the MnA site and
(312)
117
(400)
(222)
(422)
(220) (200)
(111)
Intensity (a.u.)
(204)
L21
(220)
(b)
(004)
Exp
(112)
(a)
(212) (220) (200)
H. Pan et al. / Intermetallics 86 (2017) 116e120
L10
60
70
46
80
(312)
(204)
50
44
(220)
40
(004)
(200)
30
(112)
20
(110)
(002)
(212)
(c)
90
2θ (degree) Fig. 1. (a) Experimental XRD patterns of Mn50Ni38Al12 ribbon at room temperature. (b) Simulated XRD patterns of the ordered Mn50Ni38Al12 with the cubic L21 structure, the insert shows the crystal structure of the Mn50Ni38Al12 in the austenitic state and the atoms occupation. (c) Simulated XRD patterns of the ordered Mn50Ni38Al12 with the tetragonal L10 structure, the insert shows the magnification between the (112) and (200) peaks.
drive commensurate Mn13 atoms to the D site. For this case, the simulated XRD patterns is obtained as shown in Fig. 1(b). Considering that the martensitic transition is a diffusionless transformation, the relative position between the atoms remains unchanged after the martensitic transition. Based on the above analysis, the simulated XRD patterns of the ordered Al12 in the martensitic state is obtained as shown in Fig. 1(c). It should be noted that the simulated XRD patterns of the ordered Al12 have two superlattice reflection peaks (002) and (110) which are absent in the experimental patterns indicating that there exists antisite disorder in Al12 ribbon. In addition, it is found that there are two additional minor peaks between (112) and (200) peaks in our experimental patterns. The first minor peak at 43.8 may correspond to the (220) peak of the cubic L21 structure, as a result of the martensitic transformation of Al12 near room temperature (~320 K). The second minor peak at 45.4 may correspond to the (212) peak of the tetragonal L10 structure as shown in the insert of Fig. 1(c), due to the preferential orientation of the ribbon sample. Fig. 2(a) shows the zero field cooling (ZFC) and field cooling (FC) magnetization curves of Al12 in a magnetic field of 100 Oe. It can be seen that with decreasing temperature the magnetization suddenly decreases at around 350 K, which corresponds to the martensitic transformation. Here we use TMs (TMs ¼ 350 K) and TMf (TMf ¼ 300 K) to denote the start and finish temperatures of the martensitic transformation, respectively. Thus at room temperature we observe the martensitic structure (L10) by XRD measurement. With a
Fig. 2. (a) Thermal magnetization curves for Mn50Ni38Al12 ribbon after zero field cooling and 100 Oe field cooling from 400 K to 5 K. (b) Temperature dependence of the real part of the ac susceptibility (c0 ) measured at different frequencies. The arrows indicate the direction of increasing frequencies. Inserts (c) and (d) show the fitting results according to power law and the Vogel-Fulcher law, respectively.
further decrease in the temperature an irreversibility is observed between the ZFC and FC curves that becomes more apparent with decreasing temperature, indicating the presence of magnetically inhomogeneous phases. It is worth noting that in the ZFC curve there have a peak at Tp ¼ 130 K which is usually considered to indicate the critical temperature of spin glass (SG), super spin glass (SSG) or superparamagnetic state (SPM) [10,11]. To confirm the existence of SG, SSG or SPM, we performed AC susceptibility measurements. Fig. 2(b) shows the temperature dependence of the real part of the AC susceptibility (c0 ) at different frequencies (f) values. It can be seen that each curve has a distinct peak. This peak shifts toward higher temperature and its magnitude decreases with increasing frequency. In general, one can distinguish between SG-like behavior and SPM using the value of P [13]: P ¼ DTP =ðTP Dlog10 f Þ. For SPM the interaction between the particles can nearly be neglected, and the value of P is ~0.1 [12] whereas it is ~0.01 for a SSG [13] and ~0.001 for a SG [14] system. The value of P for Al12 is 0.0168 close to the reported value of 0.016 for Ni2Mn1.4Ga0.6 [15] indicating the SSG nature of low temperature magnetism. The SSG nature has also been further confirmed by c0 (T) data with the power law [13]: fitting the * t ¼ 1=2pf ¼ t ðTP =Tg 1Þzn and the Vogel-Fulcher law [12]:
u ¼ u0 exp½Ea =kB ðTf T0 Þ. In the first equation t* is the relaxation time of individual particle moment, Tg is the finite static
118
H. Pan et al. / Intermetallics 86 (2017) 116e120
freezing temperature for u ¼ 2pf/0 and zn is the dynamic critical exponent. In the second expression u is the circular frequency which is used to measure the sample, u0 is the characteristic frequency, Ea represents the activation energy and Tf is the frozen temperature. The best fit to the data is shown by the red line in the insert of Fig. 2(b). And the fitted parameters t* ¼ 1010 s, zn ¼ 4.9, Tg ¼ 132.4 K, u0 ¼ 2.77 108 rad/s and Ea/kB ¼ 91.01 K are close to the values reported for SSG. Based on the above measurements and analysis, it is found that below 130 K there is SSG composed of magnetic clusters, and among these clusters possess competition between ferromagnetic (FM) and antiferromagnetic (AFM) exchange interactions. Exchange bias effect may occur in this sample due to the coexistence of FM and AFM exchange interaction. Fig. 3 shows the M-H loop for Al12 after 30 kOe field cooling (HFC ¼ 30 kOe) from 300 K to 5 K with a maximum measurement field of 90 kOe. Here the criterion proposed by Harres, A et al. was used to confirm that it is a major hysteresis loop [16]. It can be seen that the M-H loop shifts in the negative field direction, indicating that the EB effect appears in this sample. The value of HEB ¼ - (HR þ HL)/2 is 2405 Oe and the value of HC ¼ j HR þ HL j/2 is about 3330 Oe, where HR and HL are the right and left coercive fields, respectively. In addition, it is interesting to find that the saturation magnetization (MS) of Al12 is only 11.8 emu/g (0.44 mB/f.u.) which is similar to that of Mn50Ni42Sn8 [8] but smaller than that of Mn2Ni1.6Sn0.4 and Ni2-xMn1.4þxGa0.6 [7,11]. Based on the ordered structure of Al12 the Korringa-Kohn-Rostoker coherent-potential approximation and the generalized gradient approximation (KKR-CPA-GGA) methods was used to calculate the magnetic structures. And the molecular moment of Al12 (1.29 mB/ f.u.) obtained through the theoretical calculation is much larger than the experimental value. Therefore, the small moment may be due to the antisite disorder mentioned above in Al12. It is known that EB comes from the interfacial pinning effect between the pinning phase and the pinned phase, and the value of HEB macroscopically characterizes the strength of the interfacial interaction [2]. Accordingly, to understand the physical mechanism of the EB in Al12, we have to confirm whether or not there is only SSG phase at low temperature. Fig. 4 shows the M(H) curves of Al12 at different temperatures. It may be seen that all the M(H) curves show a sigmoid shape and that no hysteresis can be observed. Moreover, the magnetization is quite low and even under a 90 kOe magnetic field does not show saturation. All these features indicate that the sample may be in the
Fig. 3. M-H loop of Mn50Ni38Al12 at 5 K after 30 kOe field cooling. And the insert shows the enlargement between 20 kOe and 20 kOe.
Fig. 4. M (H) curves for Mn50Ni38Al12 ribbon at different temperatures between 150 K and 300 K. The solid symbols represent the experimental data and the solid lines are the fitting curves according to Eq. (1). The insert lists the values of fitted parameters.
SPM state. In order to confirm this assumption we have analyzed the M(H) curves within the framework of the modified Langevin model [11,17] which characterizes the field response of SPM clusters in the form
MðHÞ ¼ N mLðxÞ þ cH;
(1)
where M is the mass magnetization, H is the applied magnetic field, N is the mass density of SPM clusters, m is the average magnetic moment of the SPM clusters, and L(x) ¼ coth(x) 1/x is the Langevin function, where x ¼ m0 mH=kB T, kB is the Boltzmann constant, and c is the mass susceptibility. As shown in Fig. 4 all the M(H) curves can be well fitted according to Eq. (1). The fitted values of the parameters are listed in the insert of Fig. 4. It is found that both the number of the magnetic clusters per unit mass (N) and the average magnetic moment of a magnetic cluster (m) decrease with increasing temperature due to the increase in the thermal activation energy. This means that the thermal activation energy not only reduces the number of magnetic clusters, but also reduces the size of the magnetic clusters. Based on the above measurements and analysis, it is confirmed that for 130 K < T < 300 K there is no long range ordered FM phase, but only an SPM state composed of magnetic clusters without exchange interactions between the clusters. Thus there would be no FM phase for T < 130 K. This is also confirmed by the small value of MS in Fig. 2. This means that there is only SSG for T < 130 K. Therefore, the pinning and pinned phases in Al12 are both SSG. We now turn our attention to the EB effect. Fig. 5(a) and (b) show the temperature dependence of HEB and HC, respectively, after 2000 Oe field cooling (The maximum value of HEB in Al12 has been obtained after this cooling field). It is worth noting that the maximum value of HEB is as high as 5.3 kOe which is 1.5 times as much as that of Mn50Ni42Sn8 (HEB ¼ 3520 Oe) which so far, is the largest value found for MnNi based alloys [8]. From this figure it may also be seen that with increasing temperature the value of HEB decreases dramatically and becomes zero at about 25 K (defined as the blocking temperature, TB), while the value of HC initially increases and reaches a maximum at 15 K then decreases to zero at 130 K. It is interesting to note that the blocking temperature (TB ¼ 25 K) is much lower than the SSG freezing temperature (Tf ¼ 130 K). We therefore speculate that the EB mechanism of Al12
H. Pan et al. / Intermetallics 86 (2017) 116e120
119
MS can be understood as being due to the competition between the anisotropy of the pinned phase and the thermal activation energy. For T < 25 K with increasing temperature the thermal activation energy (kBT) increases and the small magnetic clusters with large anisotropy gradually begin to rotate with applied field thus more and more magnetic clusters transform into the pinned phase, the anisotropy increases and exceeds the influence of the thermal activation energy, so MS increases. For T > 25 K, the pinned phase remains unchanged while the thermal activation energy increases, so MS decreases. Therefore, the abnormal change in MS provides experimental evidence to confirm the correctness of our speculation regarding the mechanism of the EB seen in the MnNiAl system. Let's now discuss the physical mechanism of the large HEB in Al12, based on the above experimental and theoretical results. In theory, it has been obtained that for the ordered atom occupation Mn50Ni38Al12 should be written as (Mn12Ni13)A (Mn25)B(Ni25) C(Al12Mn13)D. According to this atom occupation, the two superlattice reflection peaks (002) and (110) should appear in XRD patterns and the value of the molecular moment of Al12 should be as large as 1.29 mB/f.u. However, the experimental results are as follows: the superlattice reflection peaks are absent and the molecular moment of Al12 is only 0.44 mB/f.u. Therefore, there exists considerable antisite disorder in Al12. That means Mn atoms will occupy the A B and D sites randomly. Considering that the type and strength of the Mn-Mn exchange interaction strongly depends on the distance between Mn atoms, complicated Mn atom occupation due to antisite disorder will cause the competition between AFM and FM interactions and the formation of magnetic clusters and SSG in the martensitic phase. Thus the antisite disorder in Mn50Ni38Al12 should be responsible for the large HEB. 4. Conclusion Fig. 5. Temperature dependence of HEB, HC and MS for Mn50Ni38Al12 after 2000 Oe field cooling are shown in the Fig. 5(a) (b) and (c) respectively. At the top of the figure, from left to right, are the simplified schematic diagrams of the magnetic clusters in SSG with increasing temperature. Hm, HFC, and the black arrows ‘/’ represent the measurement field, cooling field and the direction of the field, respectively.
should be different from that for other Heusler alloys where TB is equal to Tf [8,18]. Furthermore, when the EB effect occurs for other Heusler alloys, there possess at least two magnetic phases [3,19] while only the SSG state exists for Al12. In order to illustrate clearly the mechanism behind the EB in Al12, we show simplified schematic diagrams at the top of Fig. 5. In this diagram balls of different sizes represent magnetic clusters of different sizes in the SSG, with a smaller magnetic cluster showing stronger anisotropy and vice versa [20,21]. Fig. 5(i) shows that magnetic clusters with large anisotropy are the pining phase and magnetic clusters with small anisotropy are the pinned phase. Fig. 5(ii) shows that with increasing temperature the relatively large clusters in the pinning phase transform into the new pinned phase because of the increase in the thermal activation energy, and this leads to the decrease in HEB and increase in HC. With further increases in the temperature as shown in Fig. 5(iii), at about 25 K all the magnetic clusters transform into the pinned phase and the pinning phase disappears, so HEB becomes zero. Therefore, the key feature of Al12, in distinction to other Heusler alloys, is that the pinned phase can be produced from one part of the pinning phase during the process of the increasing temperature. When the temperature is high enough, all of the magnetic clusters transform into the pinned phase. In brief, in Al12 there possess a temperatureinduced dynamic transformation between the pinning and pinned phases. Fig. 5(c) shows that MS first abnormally increases for T < 25 K, and then begins to decrease for T > 25 K. This abnormal change in
In conclusion, a large exchange bias field of 5.3 kOe has been observed in Mn50Ni38Al12 polycrystalline ribbon. This exchange bias field is almost the largest value found for MnNi-based Heusler alloys. The existence of super spin glass (SSG) below 130 K and superparamagnetic (SPM) state above 130 K in the martensite phase have been confirmed by DC and AC magnetic measurements and analysis. In distinction from other Heusler alloys, the pinning and pinned phases are both SSG for Mn50Ni38Al12. Due to this particularity, there possesses a temperature-induced dynamic transformation between the pinning and pinned phases, which makes the exchange bias blocking temperature (TB ¼ 25 K) much lower than the SSG freezing temperature (Tf ¼ 130 K), and makes the saturation magnetization increase abnormally with temperature for T < 25 K. By comparing the experimental and theoretical results, it is found that there exists considerable antisite disorder in Mn50Ni38Al12 which is responsible for the large HEB. Acknowledgements This work was supported by the National Natural Science Foundation of China (Grant Nos. 11504247), the Hebei Natural Science Foundation (Nos. E2016205268), and the Department of science and technology of Hebei province scientific and technological research project (Grant Nos. 13211032 and 15211036). References s, J. Sort, V. Langlais, V. Skumryev, S. Surin ~ ach, J.S. Mun ~ oz, M.D. Baro , [1] J. Nogue Exchange bias in nanostructures, Phys. Rep. 422 (2005) 65e117, http:// dx.doi.org/10.1016/j.physrep.2005.08.004. [2] JNeaIK Schuller, Exchange bias, J. Magn. Magn. Mater. 192 (1999) 203e232, http://dx.doi.org/10.1016/S0304-8853(98)00266-2. [3] Z. Li, C. Jing, J. Chen, S. Yuan, S. Cao, J. Zhang, Observation of exchange bias in
120
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
H. Pan et al. / Intermetallics 86 (2017) 116e120 the martensitic state of Ni50Mn36Sn14 Heusler alloy, Appl. Phys. Lett. 91 (2007) 112505, http://dx.doi.org/10.1063/1.2784958. M. Khan, I. Dubenko, S. Stadler, N. Ali, Exchange bias behavior in NieMneSb Heusler alloys, Appl. Phys. Lett. 91 (2007) 072510, http://dx.doi.org/10.1063/ 1.2772233. A.K. Nayak, K.G. Suresh, A.K. Nigam, Observation of enhanced exchange bias behaviour in NiCoMnSb Heusler alloys, J. Phys. D. Appl. Phys. 42 (2009) 115004. http://iopscience.iop.org/0022-3727/42/11/115004. J. Sharma, K.G. Suresh, Investigation of multifunctional properties of Mn50Ni40xCoxSn10 (x¼0e6) Heusler alloys, J. Alloys Compd. 620 (2015) 329e336, http://dx.doi.org/10.1016/j.jallcom.2014.09.141. L. Ma, W.H. Wang, J.B. Lu, J.Q. Li, C.M. Zhen, D.L. Hou, G.H. Wu, Coexistence of reentrant-spin-glass and ferromagnetic martensitic phases in the Mn2Ni1.6Sn0.4 Heusler alloy, Appl. Phys. Lett. 99 (2011) 182507, http:// dx.doi.org/10.1063/1.3651767. J. Sharma, K.G. Suresh, Observation of giant exchange bias in bulk Mn50Ni42Sn8 Heusler alloy, Appl. Phys. Lett. 106 (2015) 072405, http:// dx.doi.org/10.1063/1.4913268. A.K. Nayak, M. Nicklas, S. Chadov, P. Khuntia, C. Shekhar, A. Kalache, M. Baenitz, Y. Skourski, V.K. Guduru, A. Puri, U. Zeitler, J.M. Coey, C. Felser, Design of compensated ferrimagnetic Heusler alloys for giant tunable exchange bias, Nat. Mater. 14 (2015) 679e684, http://dx.doi.org/10.1038/ nmat4248. S. Chatterjee, S. Giri, S.K. De, S. Majumdar, Reentrant-spin-glass state in Ni2Mn1.36Sn0.64 shape-memory alloy, Phys. Rev. B 092410 (2009) 79, http:// dx.doi.org/10.1103/PhysRevB.79.235401. €tschke, C. Hürrich, L. Schultz, SuperD.Y. Cong, S. Roth, J. Liu, Q. Luo, M. Po paramagnetic and superspin glass behaviors in the martensitic state of Ni43.5Co6.5Mn39Sn11 magnetic shape memory alloy, Appl. Phys. Lett. 96 (2010) 112504, http://dx.doi.org/10.1063/1.3367721.
[12] J.A. Mydosh, Spin Glasses - an Experimental Introduction, Taylor & Francis, London, 1993. [13] A. Malinowski, V.L. Bezusyy, R. Minikayev, P. Dziawa, Y. Syryanyy, M. Sawicki, Spin-glass behavior in Ni-doped La1.85Sr0.15CuO4, Phys. Rev. B 84 (2011) 024409, http://dx.doi.org/10.1103/PhysRevB.84.115141. [14] C.A.M. Mulder, A.J. van Duyneveldt, J.A. Mydosh, Frequency and field dependence of the ac susceptibility of the AuMn spin-glass, Phys. Rev. B 25 (1982) 515(R), http://dx.doi.org/10.1103/PhysRevB.25.515. [15] Z.D. Han, B. Qian, D.H. Wang, P. Zhang, X.F. Jiang, C.L. Zhang, Y.W. Du, Magnetic phase separation and exchange bias in off-stoichiometric Ni-Mn-Ga alloys, Appl. Phys. Lett. 103 (2013) 172403, http://dx.doi.org/10.1063/ 1.4826654. [16] A. Harres, M. Mikhov, V. Skumryev, AMHd Andrade, J.E. Schmidt, J. Geshev, Criteria for saturated magnetization loop, J. Magn. Magn. Mater. 402 (2016) 76e82, http://dx.doi.org/10.1016/j.jmmm.2015.11.046. [17] A.C. Hsiao, L.H. Lewis, K. Kang, A.R. Moodenbaugh, Magnetism and metastability of melt-spun Pd40(Fe,Ni)40P20 metallic glass, J. Appl. Phys. 99 (2006) 08F117, http://dx.doi.org/10.1063/1.2170979. [18] A.K. Nayak, C. Shekhar, Winterlik Jr., A. Gupta, C. Felser, Mn2PtIn: a tetragonal Heusler compound with exchange bias behavior, Appl. Phys. Lett. 100 (2012) 152404, http://dx.doi.org/10.1063/1.4704160. [19] M. Khan, R. Chapai, Exchange bias phenomena in Ni2xMn1.4þxGa0.6, J. Alloys Compd. 647 (2015) 935e940, http://dx.doi.org/10.1016/j.jallcom.2015.06.124. [20] A.K. Nayak, R. Sahoo, K.G. Suresh, A.K. Nigam, X. Chen, R.V. Ramanujan, Anisotropy induced large exchange bias behavior in ball milled NieCoeMneSb alloys, Appl. Phys. Lett. 98 (2011) 232502, http://dx.doi.org/ 10.1063/1.3597305. [21] K. Younsi, V. Russier, L. Bessais, Structure and magnetic properties of nanocrystalline PrCo3, J. Appl. Phys. 107 (2010) 083916, http://dx.doi.org/10.1063/ 1.3388364.
Contents lists available at ScienceDirect
Intermetallics journal homepage: www.elsevier.com/locate/intermet
Large exchange bias effect in the super spin glass state of Mn50Ni38Al12 alloy H. Pan a, L. Ma a, *, G.K. Li a, L.Y. Jia a, C.M. Zhen a, D.L. Hou a, W.H. Wang b, E.K. Liu b, J.L. Chen b, G.H. Wu b a b
Department of Physics, Hebei Advanced Thin Films Laboratory, Hebei Normal University, Shijiazhuang 050024, China Beijing National Laboratory for Condensed Matter Physics, Institute of Physics, Chinese Academy of Sciences, Beijing 100190, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 December 2016 Received in revised form 19 January 2017 Accepted 4 March 2017
A large exchange bias field (HEB) of 5.3 kOe which so far, is the largest value found in MnNi-based alloys, was observed in Mn50Ni38Al12 ribbon. DC and AC magnetic measurements show super spin glass (SSG) behavior below 130 K and superparamagnetic behavior above this temperature. Different from other systems, the pinning and pinned phases are both SSG. Due to the high content of Mn and low content of Al there exists considerable disorder in Mn50Ni38Al12 which is responsible for the large HEB. © 2017 Published by Elsevier Ltd.
Keywords: Exchange bias Super spin glass Heusler alloys
1. Introduction
2. Experimental details
In recent years, exchange bias (EB) effect in the multifunctional Heusler alloys have attracted significant interests owing to its potential application in giant magnetoresistance, ultrahigh-density magnetic recording and spin valve devices [1,2]. For Heusler alloys, EB was first observed in Ni50Mn36Sn14 and Ni50Mn25þxSb25-x alloys, but the exchange bias field (HEB) was just ~250 Oe [3,4]. The small HEB restricts the application of these alloys. By doping Co into these systems, the HEB was improved to several hundred Oe [5,6]. Recently, by changing the composition a large HEB has been also observed in Mn50Ni40Sn10 [7] (HEB ¼ 1170 Oe) and Mn50Ni42Sn8 [8] (HEB ¼ 3520 Oe) alloys. It is worth mentioning that a giant HEB (HEB ¼ 33 kOe) in Mn2.4Pt0.6Ga has been observed by Felser et al. [9]. Taking into account that Pt (or Ga) and Ni (or Al) are in the same main group, a large HEB may be obtained in MnNiAl system. Here in this work, based on the above idea, a large HEB of 5.3 kOe was observed in Mn50Ni38Al12. Besides, it is found that the pinning and pinned phases are both super spin glass by DC and AC magnetic measurements. Due to high content of Mn and low content of Al, there are considerable antisite disorder in MnNiAl, which is responsible for the large HEB.
The precursor polycrystalline ingot of Mn50Ni38Al12 with nominal composition was prepared by arc melting the pure metals (>99.9%) under an argon atmosphere. Subsequently, the ingot was melted in a quartz tube in an argon atmosphere and the melt spin technique used to obtain ribbons. The copper wheel rotated with a surface velocity of 25 m/s. The ribbons are hereafter referred as Al12, indicating that the Al content is 12 at. %. The crystal structure was identified by X-ray diffraction (XRD) using Cu Ka radiation. Magnetic properties were investigated utilizing a physical property measurement system (PPMS-9, Quantum Design, Inc.).
* Corresponding author. E-mail addresses: [email protected], [email protected] (L. Ma). http://dx.doi.org/10.1016/j.intermet.2017.03.003 0966-9795/© 2017 Published by Elsevier Ltd.
3. Results and discussion Fig. 1 shows XRD patterns for Al12 at room temperature. It can be seen that the sample shows a martensitic tetragonal structure (L10) with a ¼ b ¼ 5.48 Å and c ¼ 6.68 Å. Based on the valenceelectron site occupation rule in Heusler alloys, Al12 in the austenitic phase has an off-stoichiometric Hg2CuTi-type structure with 4 face center cubic sublattices (denoted as A (0,0,0), B (1/4 1/4 1/4), C (1/2 1/2 1/2) and D (3/4 3/4 3/4) along the body diagonal line), as shown in the insert of Fig. 1(b). The atom occupation form should be written as (Mn12Ni13)A (Mn25)B(Ni25)C(Al12Mn13)D for the ordered case. The additional Ni13 atoms take up the MnA site and
(312)
117
(400)
(222)
(422)
(220) (200)
(111)
Intensity (a.u.)
(204)
L21
(220)
(b)
(004)
Exp
(112)
(a)
(212) (220) (200)
H. Pan et al. / Intermetallics 86 (2017) 116e120
L10
60
70
46
80
(312)
(204)
50
44
(220)
40
(004)
(200)
30
(112)
20
(110)
(002)
(212)
(c)
90
2θ (degree) Fig. 1. (a) Experimental XRD patterns of Mn50Ni38Al12 ribbon at room temperature. (b) Simulated XRD patterns of the ordered Mn50Ni38Al12 with the cubic L21 structure, the insert shows the crystal structure of the Mn50Ni38Al12 in the austenitic state and the atoms occupation. (c) Simulated XRD patterns of the ordered Mn50Ni38Al12 with the tetragonal L10 structure, the insert shows the magnification between the (112) and (200) peaks.
drive commensurate Mn13 atoms to the D site. For this case, the simulated XRD patterns is obtained as shown in Fig. 1(b). Considering that the martensitic transition is a diffusionless transformation, the relative position between the atoms remains unchanged after the martensitic transition. Based on the above analysis, the simulated XRD patterns of the ordered Al12 in the martensitic state is obtained as shown in Fig. 1(c). It should be noted that the simulated XRD patterns of the ordered Al12 have two superlattice reflection peaks (002) and (110) which are absent in the experimental patterns indicating that there exists antisite disorder in Al12 ribbon. In addition, it is found that there are two additional minor peaks between (112) and (200) peaks in our experimental patterns. The first minor peak at 43.8 may correspond to the (220) peak of the cubic L21 structure, as a result of the martensitic transformation of Al12 near room temperature (~320 K). The second minor peak at 45.4 may correspond to the (212) peak of the tetragonal L10 structure as shown in the insert of Fig. 1(c), due to the preferential orientation of the ribbon sample. Fig. 2(a) shows the zero field cooling (ZFC) and field cooling (FC) magnetization curves of Al12 in a magnetic field of 100 Oe. It can be seen that with decreasing temperature the magnetization suddenly decreases at around 350 K, which corresponds to the martensitic transformation. Here we use TMs (TMs ¼ 350 K) and TMf (TMf ¼ 300 K) to denote the start and finish temperatures of the martensitic transformation, respectively. Thus at room temperature we observe the martensitic structure (L10) by XRD measurement. With a
Fig. 2. (a) Thermal magnetization curves for Mn50Ni38Al12 ribbon after zero field cooling and 100 Oe field cooling from 400 K to 5 K. (b) Temperature dependence of the real part of the ac susceptibility (c0 ) measured at different frequencies. The arrows indicate the direction of increasing frequencies. Inserts (c) and (d) show the fitting results according to power law and the Vogel-Fulcher law, respectively.
further decrease in the temperature an irreversibility is observed between the ZFC and FC curves that becomes more apparent with decreasing temperature, indicating the presence of magnetically inhomogeneous phases. It is worth noting that in the ZFC curve there have a peak at Tp ¼ 130 K which is usually considered to indicate the critical temperature of spin glass (SG), super spin glass (SSG) or superparamagnetic state (SPM) [10,11]. To confirm the existence of SG, SSG or SPM, we performed AC susceptibility measurements. Fig. 2(b) shows the temperature dependence of the real part of the AC susceptibility (c0 ) at different frequencies (f) values. It can be seen that each curve has a distinct peak. This peak shifts toward higher temperature and its magnitude decreases with increasing frequency. In general, one can distinguish between SG-like behavior and SPM using the value of P [13]: P ¼ DTP =ðTP Dlog10 f Þ. For SPM the interaction between the particles can nearly be neglected, and the value of P is ~0.1 [12] whereas it is ~0.01 for a SSG [13] and ~0.001 for a SG [14] system. The value of P for Al12 is 0.0168 close to the reported value of 0.016 for Ni2Mn1.4Ga0.6 [15] indicating the SSG nature of low temperature magnetism. The SSG nature has also been further confirmed by c0 (T) data with the power law [13]: fitting the * t ¼ 1=2pf ¼ t ðTP =Tg 1Þzn and the Vogel-Fulcher law [12]:
u ¼ u0 exp½Ea =kB ðTf T0 Þ. In the first equation t* is the relaxation time of individual particle moment, Tg is the finite static
118
H. Pan et al. / Intermetallics 86 (2017) 116e120
freezing temperature for u ¼ 2pf/0 and zn is the dynamic critical exponent. In the second expression u is the circular frequency which is used to measure the sample, u0 is the characteristic frequency, Ea represents the activation energy and Tf is the frozen temperature. The best fit to the data is shown by the red line in the insert of Fig. 2(b). And the fitted parameters t* ¼ 1010 s, zn ¼ 4.9, Tg ¼ 132.4 K, u0 ¼ 2.77 108 rad/s and Ea/kB ¼ 91.01 K are close to the values reported for SSG. Based on the above measurements and analysis, it is found that below 130 K there is SSG composed of magnetic clusters, and among these clusters possess competition between ferromagnetic (FM) and antiferromagnetic (AFM) exchange interactions. Exchange bias effect may occur in this sample due to the coexistence of FM and AFM exchange interaction. Fig. 3 shows the M-H loop for Al12 after 30 kOe field cooling (HFC ¼ 30 kOe) from 300 K to 5 K with a maximum measurement field of 90 kOe. Here the criterion proposed by Harres, A et al. was used to confirm that it is a major hysteresis loop [16]. It can be seen that the M-H loop shifts in the negative field direction, indicating that the EB effect appears in this sample. The value of HEB ¼ - (HR þ HL)/2 is 2405 Oe and the value of HC ¼ j HR þ HL j/2 is about 3330 Oe, where HR and HL are the right and left coercive fields, respectively. In addition, it is interesting to find that the saturation magnetization (MS) of Al12 is only 11.8 emu/g (0.44 mB/f.u.) which is similar to that of Mn50Ni42Sn8 [8] but smaller than that of Mn2Ni1.6Sn0.4 and Ni2-xMn1.4þxGa0.6 [7,11]. Based on the ordered structure of Al12 the Korringa-Kohn-Rostoker coherent-potential approximation and the generalized gradient approximation (KKR-CPA-GGA) methods was used to calculate the magnetic structures. And the molecular moment of Al12 (1.29 mB/ f.u.) obtained through the theoretical calculation is much larger than the experimental value. Therefore, the small moment may be due to the antisite disorder mentioned above in Al12. It is known that EB comes from the interfacial pinning effect between the pinning phase and the pinned phase, and the value of HEB macroscopically characterizes the strength of the interfacial interaction [2]. Accordingly, to understand the physical mechanism of the EB in Al12, we have to confirm whether or not there is only SSG phase at low temperature. Fig. 4 shows the M(H) curves of Al12 at different temperatures. It may be seen that all the M(H) curves show a sigmoid shape and that no hysteresis can be observed. Moreover, the magnetization is quite low and even under a 90 kOe magnetic field does not show saturation. All these features indicate that the sample may be in the
Fig. 3. M-H loop of Mn50Ni38Al12 at 5 K after 30 kOe field cooling. And the insert shows the enlargement between 20 kOe and 20 kOe.
Fig. 4. M (H) curves for Mn50Ni38Al12 ribbon at different temperatures between 150 K and 300 K. The solid symbols represent the experimental data and the solid lines are the fitting curves according to Eq. (1). The insert lists the values of fitted parameters.
SPM state. In order to confirm this assumption we have analyzed the M(H) curves within the framework of the modified Langevin model [11,17] which characterizes the field response of SPM clusters in the form
MðHÞ ¼ N mLðxÞ þ cH;
(1)
where M is the mass magnetization, H is the applied magnetic field, N is the mass density of SPM clusters, m is the average magnetic moment of the SPM clusters, and L(x) ¼ coth(x) 1/x is the Langevin function, where x ¼ m0 mH=kB T, kB is the Boltzmann constant, and c is the mass susceptibility. As shown in Fig. 4 all the M(H) curves can be well fitted according to Eq. (1). The fitted values of the parameters are listed in the insert of Fig. 4. It is found that both the number of the magnetic clusters per unit mass (N) and the average magnetic moment of a magnetic cluster (m) decrease with increasing temperature due to the increase in the thermal activation energy. This means that the thermal activation energy not only reduces the number of magnetic clusters, but also reduces the size of the magnetic clusters. Based on the above measurements and analysis, it is confirmed that for 130 K < T < 300 K there is no long range ordered FM phase, but only an SPM state composed of magnetic clusters without exchange interactions between the clusters. Thus there would be no FM phase for T < 130 K. This is also confirmed by the small value of MS in Fig. 2. This means that there is only SSG for T < 130 K. Therefore, the pinning and pinned phases in Al12 are both SSG. We now turn our attention to the EB effect. Fig. 5(a) and (b) show the temperature dependence of HEB and HC, respectively, after 2000 Oe field cooling (The maximum value of HEB in Al12 has been obtained after this cooling field). It is worth noting that the maximum value of HEB is as high as 5.3 kOe which is 1.5 times as much as that of Mn50Ni42Sn8 (HEB ¼ 3520 Oe) which so far, is the largest value found for MnNi based alloys [8]. From this figure it may also be seen that with increasing temperature the value of HEB decreases dramatically and becomes zero at about 25 K (defined as the blocking temperature, TB), while the value of HC initially increases and reaches a maximum at 15 K then decreases to zero at 130 K. It is interesting to note that the blocking temperature (TB ¼ 25 K) is much lower than the SSG freezing temperature (Tf ¼ 130 K). We therefore speculate that the EB mechanism of Al12
H. Pan et al. / Intermetallics 86 (2017) 116e120
119
MS can be understood as being due to the competition between the anisotropy of the pinned phase and the thermal activation energy. For T < 25 K with increasing temperature the thermal activation energy (kBT) increases and the small magnetic clusters with large anisotropy gradually begin to rotate with applied field thus more and more magnetic clusters transform into the pinned phase, the anisotropy increases and exceeds the influence of the thermal activation energy, so MS increases. For T > 25 K, the pinned phase remains unchanged while the thermal activation energy increases, so MS decreases. Therefore, the abnormal change in MS provides experimental evidence to confirm the correctness of our speculation regarding the mechanism of the EB seen in the MnNiAl system. Let's now discuss the physical mechanism of the large HEB in Al12, based on the above experimental and theoretical results. In theory, it has been obtained that for the ordered atom occupation Mn50Ni38Al12 should be written as (Mn12Ni13)A (Mn25)B(Ni25) C(Al12Mn13)D. According to this atom occupation, the two superlattice reflection peaks (002) and (110) should appear in XRD patterns and the value of the molecular moment of Al12 should be as large as 1.29 mB/f.u. However, the experimental results are as follows: the superlattice reflection peaks are absent and the molecular moment of Al12 is only 0.44 mB/f.u. Therefore, there exists considerable antisite disorder in Al12. That means Mn atoms will occupy the A B and D sites randomly. Considering that the type and strength of the Mn-Mn exchange interaction strongly depends on the distance between Mn atoms, complicated Mn atom occupation due to antisite disorder will cause the competition between AFM and FM interactions and the formation of magnetic clusters and SSG in the martensitic phase. Thus the antisite disorder in Mn50Ni38Al12 should be responsible for the large HEB. 4. Conclusion Fig. 5. Temperature dependence of HEB, HC and MS for Mn50Ni38Al12 after 2000 Oe field cooling are shown in the Fig. 5(a) (b) and (c) respectively. At the top of the figure, from left to right, are the simplified schematic diagrams of the magnetic clusters in SSG with increasing temperature. Hm, HFC, and the black arrows ‘/’ represent the measurement field, cooling field and the direction of the field, respectively.
should be different from that for other Heusler alloys where TB is equal to Tf [8,18]. Furthermore, when the EB effect occurs for other Heusler alloys, there possess at least two magnetic phases [3,19] while only the SSG state exists for Al12. In order to illustrate clearly the mechanism behind the EB in Al12, we show simplified schematic diagrams at the top of Fig. 5. In this diagram balls of different sizes represent magnetic clusters of different sizes in the SSG, with a smaller magnetic cluster showing stronger anisotropy and vice versa [20,21]. Fig. 5(i) shows that magnetic clusters with large anisotropy are the pining phase and magnetic clusters with small anisotropy are the pinned phase. Fig. 5(ii) shows that with increasing temperature the relatively large clusters in the pinning phase transform into the new pinned phase because of the increase in the thermal activation energy, and this leads to the decrease in HEB and increase in HC. With further increases in the temperature as shown in Fig. 5(iii), at about 25 K all the magnetic clusters transform into the pinned phase and the pinning phase disappears, so HEB becomes zero. Therefore, the key feature of Al12, in distinction to other Heusler alloys, is that the pinned phase can be produced from one part of the pinning phase during the process of the increasing temperature. When the temperature is high enough, all of the magnetic clusters transform into the pinned phase. In brief, in Al12 there possess a temperatureinduced dynamic transformation between the pinning and pinned phases. Fig. 5(c) shows that MS first abnormally increases for T < 25 K, and then begins to decrease for T > 25 K. This abnormal change in
In conclusion, a large exchange bias field of 5.3 kOe has been observed in Mn50Ni38Al12 polycrystalline ribbon. This exchange bias field is almost the largest value found for MnNi-based Heusler alloys. The existence of super spin glass (SSG) below 130 K and superparamagnetic (SPM) state above 130 K in the martensite phase have been confirmed by DC and AC magnetic measurements and analysis. In distinction from other Heusler alloys, the pinning and pinned phases are both SSG for Mn50Ni38Al12. Due to this particularity, there possesses a temperature-induced dynamic transformation between the pinning and pinned phases, which makes the exchange bias blocking temperature (TB ¼ 25 K) much lower than the SSG freezing temperature (Tf ¼ 130 K), and makes the saturation magnetization increase abnormally with temperature for T < 25 K. By comparing the experimental and theoretical results, it is found that there exists considerable antisite disorder in Mn50Ni38Al12 which is responsible for the large HEB. Acknowledgements This work was supported by the National Natural Science Foundation of China (Grant Nos. 11504247), the Hebei Natural Science Foundation (Nos. E2016205268), and the Department of science and technology of Hebei province scientific and technological research project (Grant Nos. 13211032 and 15211036). References s, J. Sort, V. Langlais, V. Skumryev, S. Surin ~ ach, J.S. Mun ~ oz, M.D. Baro , [1] J. Nogue Exchange bias in nanostructures, Phys. Rep. 422 (2005) 65e117, http:// dx.doi.org/10.1016/j.physrep.2005.08.004. [2] JNeaIK Schuller, Exchange bias, J. Magn. Magn. Mater. 192 (1999) 203e232, http://dx.doi.org/10.1016/S0304-8853(98)00266-2. [3] Z. Li, C. Jing, J. Chen, S. Yuan, S. Cao, J. Zhang, Observation of exchange bias in
120
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
H. Pan et al. / Intermetallics 86 (2017) 116e120 the martensitic state of Ni50Mn36Sn14 Heusler alloy, Appl. Phys. Lett. 91 (2007) 112505, http://dx.doi.org/10.1063/1.2784958. M. Khan, I. Dubenko, S. Stadler, N. Ali, Exchange bias behavior in NieMneSb Heusler alloys, Appl. Phys. Lett. 91 (2007) 072510, http://dx.doi.org/10.1063/ 1.2772233. A.K. Nayak, K.G. Suresh, A.K. Nigam, Observation of enhanced exchange bias behaviour in NiCoMnSb Heusler alloys, J. Phys. D. Appl. Phys. 42 (2009) 115004. http://iopscience.iop.org/0022-3727/42/11/115004. J. Sharma, K.G. Suresh, Investigation of multifunctional properties of Mn50Ni40xCoxSn10 (x¼0e6) Heusler alloys, J. Alloys Compd. 620 (2015) 329e336, http://dx.doi.org/10.1016/j.jallcom.2014.09.141. L. Ma, W.H. Wang, J.B. Lu, J.Q. Li, C.M. Zhen, D.L. Hou, G.H. Wu, Coexistence of reentrant-spin-glass and ferromagnetic martensitic phases in the Mn2Ni1.6Sn0.4 Heusler alloy, Appl. Phys. Lett. 99 (2011) 182507, http:// dx.doi.org/10.1063/1.3651767. J. Sharma, K.G. Suresh, Observation of giant exchange bias in bulk Mn50Ni42Sn8 Heusler alloy, Appl. Phys. Lett. 106 (2015) 072405, http:// dx.doi.org/10.1063/1.4913268. A.K. Nayak, M. Nicklas, S. Chadov, P. Khuntia, C. Shekhar, A. Kalache, M. Baenitz, Y. Skourski, V.K. Guduru, A. Puri, U. Zeitler, J.M. Coey, C. Felser, Design of compensated ferrimagnetic Heusler alloys for giant tunable exchange bias, Nat. Mater. 14 (2015) 679e684, http://dx.doi.org/10.1038/ nmat4248. S. Chatterjee, S. Giri, S.K. De, S. Majumdar, Reentrant-spin-glass state in Ni2Mn1.36Sn0.64 shape-memory alloy, Phys. Rev. B 092410 (2009) 79, http:// dx.doi.org/10.1103/PhysRevB.79.235401. €tschke, C. Hürrich, L. Schultz, SuperD.Y. Cong, S. Roth, J. Liu, Q. Luo, M. Po paramagnetic and superspin glass behaviors in the martensitic state of Ni43.5Co6.5Mn39Sn11 magnetic shape memory alloy, Appl. Phys. Lett. 96 (2010) 112504, http://dx.doi.org/10.1063/1.3367721.
[12] J.A. Mydosh, Spin Glasses - an Experimental Introduction, Taylor & Francis, London, 1993. [13] A. Malinowski, V.L. Bezusyy, R. Minikayev, P. Dziawa, Y. Syryanyy, M. Sawicki, Spin-glass behavior in Ni-doped La1.85Sr0.15CuO4, Phys. Rev. B 84 (2011) 024409, http://dx.doi.org/10.1103/PhysRevB.84.115141. [14] C.A.M. Mulder, A.J. van Duyneveldt, J.A. Mydosh, Frequency and field dependence of the ac susceptibility of the AuMn spin-glass, Phys. Rev. B 25 (1982) 515(R), http://dx.doi.org/10.1103/PhysRevB.25.515. [15] Z.D. Han, B. Qian, D.H. Wang, P. Zhang, X.F. Jiang, C.L. Zhang, Y.W. Du, Magnetic phase separation and exchange bias in off-stoichiometric Ni-Mn-Ga alloys, Appl. Phys. Lett. 103 (2013) 172403, http://dx.doi.org/10.1063/ 1.4826654. [16] A. Harres, M. Mikhov, V. Skumryev, AMHd Andrade, J.E. Schmidt, J. Geshev, Criteria for saturated magnetization loop, J. Magn. Magn. Mater. 402 (2016) 76e82, http://dx.doi.org/10.1016/j.jmmm.2015.11.046. [17] A.C. Hsiao, L.H. Lewis, K. Kang, A.R. Moodenbaugh, Magnetism and metastability of melt-spun Pd40(Fe,Ni)40P20 metallic glass, J. Appl. Phys. 99 (2006) 08F117, http://dx.doi.org/10.1063/1.2170979. [18] A.K. Nayak, C. Shekhar, Winterlik Jr., A. Gupta, C. Felser, Mn2PtIn: a tetragonal Heusler compound with exchange bias behavior, Appl. Phys. Lett. 100 (2012) 152404, http://dx.doi.org/10.1063/1.4704160. [19] M. Khan, R. Chapai, Exchange bias phenomena in Ni2xMn1.4þxGa0.6, J. Alloys Compd. 647 (2015) 935e940, http://dx.doi.org/10.1016/j.jallcom.2015.06.124. [20] A.K. Nayak, R. Sahoo, K.G. Suresh, A.K. Nigam, X. Chen, R.V. Ramanujan, Anisotropy induced large exchange bias behavior in ball milled NieCoeMneSb alloys, Appl. Phys. Lett. 98 (2011) 232502, http://dx.doi.org/ 10.1063/1.3597305. [21] K. Younsi, V. Russier, L. Bessais, Structure and magnetic properties of nanocrystalline PrCo3, J. Appl. Phys. 107 (2010) 083916, http://dx.doi.org/10.1063/ 1.3388364.