Size modulated Griffiths phase and spin dynamics in double perovskite Sm1.5Ca0.5CoMnO6

Journal of Magnetism and Magnetic Materials 469 (2019) 161–170

Contents lists available at ScienceDirect

Journal of Magnetism and Magnetic Materials journal homepage: www.elsevier.com/locate/jmmm

Research articles

Size modulated Griffiths phase and spin dynamics in double perovskite Sm1.5Ca0.5CoMnO6 R.C. Sahooa, Sananda Dasa, S.K. Girib, D. Paladhia, T.K. Natha, a b

T

⁎

Department of Physics, Indian Institute of Technology Kharagpur, West Bengal 721302, India Department of Materials Science and Metallurgy, University of Cambridge, 27 Charles Babbage Road, Cambridge CB3 0FS, United Kingdom

A R T I C LE I N FO

A B S T R A C T

Keywords: Magnetic oxides Dynamic properties Griffiths phase Magnetic couplings

Griffiths phase like behavior has been investigated in details in size modulated Sm1.5Ca0.5CoMnO6 double perovskite, having ferromagnetic entities in the paramagnetic matrix. Temperature dependent real and imaginary part of ac susceptibility confirms that Sm1.5Ca0.5CoMnO6-1150 °C bulk-sample has spin-glass like ordering below Tg ∼ 34.88(3) K. This glassy behavior is attributed to the antisite disorders (∼7.12%) along with and magnetic frustration in the system. However, the ground state of Sm1.5Ca0.5CoMnO6-600 °C nanometric-sample is observed to be antiferromagnetic in nature due to the presence of large antisite disorder (∼53.58%). The observed unusual co-existence of antiferromagnetism with Griffiths-like phase in Sm1.5Ca0.5CoMnO6-600 °C sample is rare in nature. Interesting magnetic spin dynamics and aging effect are also observed in both the compounds suggesting different spin relaxation processes in low temperature regime. The observed magnetism in these compounds can be tuned by nearest neighbor Co-O-Mn and/or Co-O-Co/Mn-O-Mn exchange interactions as well as the next nearest neighbor Co-O-O-Co/Mn-O-O-Mn interactions.

1. Introduction In recent time one of the fascinating topics in the field of condensed matter physics is to explore various exciting features of complex magnetic oxides like Griffiths phase (GP), slow spin relaxation, aging and memory effects etc. [1–3]. Among them, the GP is one of the distinct magnetic state in which magnetization falls due to the completely random and competing magnetic interactions in the temperature range TCRand < T < TG, where TCRand is the critical temperature for random ferromagnetic (FM) entities and TG represents a temperature for the onset of completely random paramagnetic (PM) states [4,5]. In this intermediate temperature regime, the GP is microscopically distinguished by a FM cluster like system. This was first predicted theoretically in randomly diluted Ising ferromagnets by R. B. Griffiths in 1969 [6]. The presence of GP in magnetic materials is generally related to the competing magnetic interactions leading to FM entities. Moreover, the origin of GP varies from one material to other based on the synthesis conditions (size, dimension, shape, morphology, etc.), multiple crystallographic structure, quenched disorder and multi-magnetic phases. For example, de Teresa et al. have reported GP-like behavior in (La-Y/Tb)2/3Ca1/3MnO3 perovskite samples and they suggested competition between charge ordered antiferromagnet (AFM) and metallic FM appearing as nanoscale inhomogeneities in the PM regime [7]. The ⁎

Tb5Si2Ge2 compound showed the GP-like state (115 K < T < 200 K) in actual PM region and it was originated from the crystallographic structural disorder along with competing intralayer and interlayer magnetic exchange interactions [8]. Ouyang et al. have reported that short range FM correlations and dynamic FM clusters are responsible for GP-like phase (128 K < T < 240 K) in binary compound Gd5Ge4 [9]. The Ca3CoMnO6 spin chain system showed GP-like feature (13 K < T < 125 K) due to short-range FM correlations [10]. A similar GP-like ordering (6 K < T < 12 K) was noticed in another spin-chain compound Sr3CuRhO6 [11]. On the other hand, the GP singularities have been observed in a non-Fermi-liquid system due to the interplay between the Kondo-like and RKKY interactions where both the magnetic anisotropy and disorder are present [12]. Recently, GP singularity has been identified in a dilute magnetic semiconductor Fe1-xCoxS2 formation of which has been assigned to the clusters of localized ephemeral magnetic ordering [13]. The crystallographic disorder (including antisite defects (ASD)) in double perovskite compounds of general formula A′2-xA″xCoMnO6 [A′ = rare earth elements (Sm, La, Y, etc.) with oxidation state + 3 and A″ = alkaline-earth elements (Ca, Sr, Ba, etc.) with oxidation state + 2] can be introduced with the substitution of A′3+ by A″2+ with different ionic radii. This disorder induces distinct fascinating and complex properties like multi-glass behavior [14], phase separated

Corresponding author. E-mail address: [email protected] (T.K. Nath).

https://doi.org/10.1016/j.jmmm.2018.08.060 Received 18 June 2018; Received in revised form 10 August 2018; Accepted 22 August 2018 Available online 23 August 2018 0304-8853/ © 2018 Elsevier B.V. All rights reserved.

Journal of Magnetism and Magnetic Materials 469 (2019) 161–170

R.C. Sahoo et al.

Using Debye-Scherrer formula, the average crystallite size has been estimated to be ∼21.4 nm for SCCMO_N sample. All the XRD spectra have been examined through Rietveld refinement using FullProf Suite programme. The observed reflections for both the samples are assigned to the small B-site disordered monoclinic crystal symmetry with space group P21/n (14). The presence of ASD in the B-site sublattice has been accounted to achieve the best fit. The obtained refined structural parameters for both the samples are listed in Table 1. All the peaks in XRD pattern of SCCMO_N sample are indexed well using this space group. From Table 1, we conclude that the Mn-O-Co angles in both the samples are found to be quite different in contrast to that found in the ordered Sm2CoMnO6 compound [21]. The interatomic distances (Co/ Mn-O) for both the samples are not identical and more disorder in SCCMO_N, generally indicates the presence of oxygen vacancies or/and disorder in the crystal structure. Fig. 1(b) shows the FESEM micrograph of SCCMO_N sample demonstrating almost uniform grains separated by grain boundaries. The calculated average grain size is ∼20 nm (inset of Fig. 1(b)). In the Fig. 1(c), the ring pattern in the selective area electron diffraction (SAED) image confirms nanocrystalline nature of SCCMO_N samples. From this ring, we have calculated the average distance from two bright spots ∼7.01 Å and corresponding lattice spacing d ∼ 0.28 nm. Fig. 1(d) shows the HRTEM lattice image of SCCMO_N, revealing the good crystallinity of the sample and the value of lattice spacing (∼0.25 nm) corresponding to (1 1 2) plane is consistent with that estimated from SAED pattern. However, the estimated average grain size from FESEM micrograph of SCCMO_B is ∼1.5 μm (not shown here) which is also consistent with our XRD result and previously reported results [22]. When an element of different ionic radius (Ca2+, radii ∼ 1.0 Å) is substituted at Sm (Sm3+, radii ∼ 1.08 Å) site in order to relieve the internal chemical stress, CoO6 or MnO6 octahedra cooperatively rotates (unequal Co-O and/or Mn-O bonds) and/or tilts (unequal Co/Mn-O-Co/ Mn bond angles < 180°). Fig. 2(a) and (b) show the local environment of the SCCMO_B and SCCMO_N samples drawn using data from Table 1. The lattice parameters ‘b’ and ‘c’ are found to be smaller in SCCMO_N, while the parameter ‘a’ has increased as a result of enhancement of structural disorder. The inset of Fig. 2(a) shows crystal structure of SCCMO_B, where MnO6 and CoO6 octahedra are located at distinct preferential positions. The Co and Mn atoms are occupied along the three crystallographic directions at 2c (½, 0, 0) and 2d (0, ½, 0) positions, respectively. Here, the monoclinic structure consists of tilted and distorted CoO6 and MnO6 octahedra and are periodically arranged on the cubic edge. It makes the system consisting of reduced ASD. The inset of Fig. 2(b) shows crystal structure of SCCMO_N, where MnO6 and CoO6 octahedra are not located alternatively on the cubic edge and are assigned to large ASD (greater than50%). The observed structural information of both the samples suggests that the system may contain a disordered phase along with ASD.

ferroelectricity [15], GP-like ordering [11], superconductivity [16], colossal magnetoresistance [17] and magnetic relaxation dynamics [14]. However, the magnetic properties of these materials are also influenced by the strong spin orbit coupling on the 3d transition metal ions and electron delocalization in the intrachain (Co-O-Mn) network. This also induces ASD in the double perovskite systems, i.e., mislocation of either CoO6 or MnO6 octahedron inherently and forms interchain (Co-O-Co or Mn-O-Mn) interaction instead of intrachain interaction along each axis. The ASD is a common feature in most of the double perovskite systems and it destroys structural periodicity but introduces AFM superexchange coupling or/and magnetic frustration in the FM compound [18,19]. It leads to magnetic domain formation in the PM regime and as a result formation of GP and suppression of magnetization take place. In this study, we report the size modulated structural, magnetic and spin dynamics of 25% Ca substituted Sm2CoMnO6 (Sm1.5Ca0.5CoMnO6) double perovskite system. Interestingly, both the systems, Sm1.5Ca0.5CoMnO6-1150 °C (say, SCCMO_B with a bulk average grain size of 1.5 μm) and Sm1.5Ca0.5CoMnO6-600 °C (say, SCCMO_N with a nanometric average grain size of 20 nm) show GP-like ordering in a temperature regime TCRand < T < TG. Moreover, only SCCMO_B system exhibits glassy feature in the low temperature and we have tried to find out the possible origin of it. We have also investigated the magnetic aging behavior in the light of simple power law and stretched exponential models for both the samples [20]. 2. Experimental details Polycrystalline double perovskite Sm1.5Ca0.5CoMnO6 (SCCMO) powders have been synthesized by chemical sol-gel route from the stoichiometric mixture of high purity Sm2O3, CaCO3, Co(NO3)2 6H2O, Mn(CH3COO)2 4H2O precursor materials. The stoichiometric amount of Sm2O3 is dissolved into the minimum amount of nitric acid (HNO3) to form Sm(NO3)3 and other precursors are dissolved in deionized water to get clear solutions. All the solutions are mixed together and then heated at 150 °C using a hot plate for 4 days. The obtained precursor black and fluffy powder is calcinated at 1150 °C for 10 h to get bulk powder of Sm1.5Ca0.5CoMnO6 (SCCMO_B). The resin powder is calcinated at 600 °C for 6 h to get bulk powder of Sm1.5Ca0.5CoMnO6 (SCCMO_N). The crystallographic structural details of the sample have been investigated by rigorous high resolution x-ray diffraction (Panalytical x’pert pro HRXRD-I, PW 3040/60) using Cu-Kα radiation (λ ∼ 1.542 Å) at 300 K, the field emission scanning electron microscopy (FESEM, SUPRA 40, Carl Zeiss SMT AG, Germany with a resolution of 1 nm) image and high resolution transmission electron microscopy (HRTEM, JEM2100 with a resolution of 1.9 Å). Electronic structure of the sample has been determined using X-ray photoelectron spectroscopy (XPS, PHI 5000 Versa Probe II Scanning). The magnetization measurements have been carried out by using SQUID magnetometry (Quantum Design, USA with resolution 10−8 emu) and Physical Property Measurement System (PPMS, Quantum Design, USA) in the temperature range of 5–325 K with a maximum field of ± 7 T. Before starting each magnetization measurement, the samples were heated above their respective Curie temperature to ensure the complete demagnetization state of the sample. The samples were also demagnetized by using oscillating field with an appropriate protocol.

3.2. Comparative energy dispersive X-ray spectroscopy (EDX) study EDX spectra analysis of SCCMO_B and SCCMO_N samples has been carried out to determine the chemical homogeneity of the polycrystalline samples as shown in the inset of Fig. 2(c) and (d), respectively. For both the samples, all the constituent elements are present without having any impurity element. Fig. 2(c) and (d) shows the normalized atomic and weight percentage plot of constituent elements for both the samples. The presence of percentage of elements matches quite well with the nominal stoichiometry of SCCMO samples. The calculated chemical formula of SCCMO_B is Sm1.5Ca0.505Co0.939Mn1.09O6-δB and that for SCCMO_N is Sm1.5Ca0.496Co0.898Mn1.11O6-δN. The ratio of 3d transition metal elements (Co:Mn) in both systems is not perfectly 1:1. The estimated oxygen content, as calculated from charge balance, is ∼5.404 for SCCMO_B and ∼5.698 for SCCMO_N system. This confirms that the change in particle size or sintering temperature does affect the

3. Results and discussion 3.1. Structural properties Room temperature HRXRD patterns of SCCMO_B and SCCMO_N samples display typical perovskite structure as shown in Fig. 1(a) and inset of Fig. 1(a), respectively. The XRD patterns show polycrystalline nature of the samples and broadening of the peaks with the decrease of calcinations temperature signifying the smaller average crystallite size. 162

Journal of Magnetism and Magnetic Materials 469 (2019) 161–170

R.C. Sahoo et al. (112)

(b)

(204) (133)

(114) (131)

(020)

SCCMO_N

(021) (120) (211) (022) (202) (113) (212) (220) (004) (221)

(110) (111)

Intensity (a.u.)

24

(a) 20

30

40 50 2 (deg)

60

No. of grains (a.u.)

Intensity (a.u.)

Experimental Simulated Bragg's positions Difference

70

SCCMO_B

16

8

0

20

30

40

50 60 70 2 (deg)

80

9

18 27 Grain size (nm)

36

90

(c) (7.01 /nm)

5 /nm Fig. 1. . (a) The HRXRD pattern of SCCMO_B sample; Inset shows HRXRD pattern of SCCMO_N sample at room temperature with Rietveld refinement. (b) FESEM micrograph image; inset shows a histogram of particles size, (c) SAED pattern, and (d) HRTEM lattice image of SCCMO_N nanoparticles.

stoichiometry and ASD in the samples. Thus, we conclude that both the systems have small oxygen vacancies and less ASD at B site cations in SCCMO_B than SCCMO_N system.

configurations with respect to the ordered system, similar to that of disordered La1.5Ca0.5CoMnO6 [3], Sm2CoMnO6 [23] and La2CoMnO6 [24] perovskites reported earlier. The area ratio of Co2+: Co3+ and Mn3+: Mn4+ peaks in SCCMO_B system is 55.45:44.55 and 33.82:66.18, respectively. On the other hand, for SCCMO_N system the area ratio of Co2+: Co3+ and Mn3+: Mn4+ peaks are 54.62:45.38 and 66.20:33.80, respectively. Moreover, greater than 30% of Co3+ and Mn3+ cations are present in these disordered systems instead of 100% Co2+ and Mn4+ cations. The intensity of Co2+ and Mn4+ ions are lesser in SCCMO_N system than SCCMO_B. The large percentage of Mn3+ (66.20%) and Mn3+- Mn3+ interactions leads to antiferromagnetism in SCCMO_N. The area ratio of Co2+/Co3+ and Mn3+/Mn4+ confirms that both the samples have no oxygen deficiency [25]. In this case, the existence of mixed valence state of both the systems plays an important role in the formation of GP and magnetic frustration. For these reasons, grain size dependent SCCMO system has a great advantage in applications towards manipulation of multi magnetic ordering and for advanced scientific research.

3.3. Comparative valence state study of Mn and Co The substitution of Sm3+ by Ca2+ and structural disorder will lead to the change of valence state of 3d transition metal (Co/Mn) ions due to the charge neutrality in SCCMO system [21,23]. To evaluate the oxidation states of Co and Mn ions in SCCMO_B and SCCMO_N samples, we have used room temperature XPS measurement. The XPS spectra of Co 2p3/2 and Mn 2p3/2 core level in SCCMO_B are shown in Fig. 3(a) and (b) and that for SCCMO_N are shown in Fig. 3(c) and (d), respectively. The 2p3/2 core level spectrum of both the samples shows Co 2p3/ 2 peak consisting of two peaks. On the other hand, the Mn 2p3/2 core level spectrum also consists of two peaks. The corresponding binding energy (BE) for Co 2p3/2 and Mn 2p3/2 are summarized in Table 2. We observe in the Fig. 3(a) and (c) that the BE of Co 2p3/2 appears in the range of 780.15–780.20 eV for Co2+ state and 779.01–779.03 eV for Co3+ state. The BE of Mn 2p3/2 appears in the range of 641.10–640.84 eV for Mn3+ state and 642.40–642.38 eV for Mn4+ state as shown in the Fig. 3(b) and (d). Therefore, XPS study illustrates the mixed valence state of Co and Mn in both the samples. So the valence state of Co2+ and Mn4+ cations is shifted to Co3+ and Mn3+

3.4. Comparative Raman study In order to understand the degree of ASD in the present study, we have recorded room temperature Raman spectra of SCCMO_B and SCCMO_N samples as illustrated in Fig. 4(a) and (b), respectively. The 163

Journal of Magnetism and Magnetic Materials 469 (2019) 161–170

R.C. Sahoo et al.

decreases with increasing grain size for A1g mode and it shows opposite behavior for B1g mode. The interesting point here is that the symmetry of the stretching mode and anti-stretching mode are different in ordered P21/n phase. As a comparison, the stretching mode in ordered P21/n (14) phase is observed at 645 cm−1 with A1g symmetry along xx and x′x′ configurations. But the anti-stretching and/or bending mode vibrations are observed at 495 cm−1 with B1g modes of symmetry [26]. An increase in stretching and anti-stretching mode frequency of our two samples is in contrast to that of ordered P21/n structure and can be described qualitatively by considering a decrease of the < Co/MnO > bond lengths and bond angles. The Raman mode shift and FWHM at both the peaks also indicate the presence of disorder in the both SCCMO_B and SCMO_N systems, mainly appearing due to the ASD and structural disorder induced by substitution of Sm3+ by Ca2+ cations. All these conclusions are consistent with our structural analysis.

Table 1 . Unit-cell, positional parameters, discrepancy factors, bond distances and selected angles after the Rietveld refinement of the crystal structure from experimental HRXRD data at room temperature of SCCMO_B and SCCMO_N (CoO6 and MnO6 are not located alternatively) double perovskites. Crystallite size is also included. Compound Space group Cell parameters a (Å) b (Å) c (Å) β (°) V (Å3) Atomic position Sm/Ca 4e (x, y, z) x y z B (Å2) Mn/Co B (Å2) O1 4e (x, y, z) x y z B (Å2) O2 4e (x, y, z) x y z B (Å2) O3 4e (x, y, z) x y z B (Å2) Bond distances Mn-O1 (Å) Mn-O2 (Å) Mn-O3 (Å) Co-O1 (Å) Co-O2 (Å) Co-O3 (Å) Angles Mn-O1-Co (°) Mn-O2-Co (°) Mn-O3-Co (°) Discrepancy factor Rp(%) Rwp(%) Rexp(%) χ2 Crystalline size D (nm)

SCCMO_B P21/n (14)

SCCMO_N P21/n (14)

5.349(4) 5.434(5) 7.580(7) 89.98(3) 221.76(8)

5.354(1) 5.361(12) 7.503(18) 90.04(5) 220.35(3)

0.9886(15) 0.0419(6) 0.253(3) 0.051

0.9890(5) 0.0404(9) 0.254(3) 0.0031

0.082

0.0046

0.074(6) 0.514(6) 0.261(13) 0.358

−0.095(9) 0.535(8) 0.222(10) 0.0088

0.243(17) 0.286(14) 0.493(12) 0.229

0.327(13) 0.269(18) 0.422(7) 0.0086

0.828(8) 0.191(10) −0.088(6) 0.0135

0.706(12) 0.129(7) −0.010(13) 0.0033

1.8776 1.8106 1.9054 2.1042 2.0433 2.1159

2.1772 2.2376 1.7110 1.7691 1.8348 2.5670

150.990 140.673 172.594

147.662 139.566 159.787

8.1 6.7 16.7 0.267

15.9 12.0 22.19 1.294

1450

21.47

3.5. DC magnetization study: Size induced magnetism and formation of GP phase scenario Fig. 5(a) shows the temperature dependence of DC magnetization M (T) for SCCMO of different particle size in zero-field-cooled (ZFC) and field cooled (FC) conditions under a magnetic field of 100 Oe. The Co and Mn spin sublattices experience magnetic transitions at TC ∼ 120.5 K for SCCMO_B and at TC ∼ 94.3 K for SCCMO_N, while it is also strongly affected by the presence of any frustration in magnetic interactions. The magnetic transition temperature (TC) of different grain size is calculated from dMZFC/dT vs temperature curve (inset of Fig. 5(a)). The TC for both the samples has been assigned to positive magnetic exchange interaction (Co2+/Co3+–O2−-Mn4+/Mn3+). Since the ZFC-FC bifurcation is observed due to the magnetic anisotropy well below TC and/or formation of local spin clusters or AFM domain [27,28]. A clear ZFC-FC bifurcation is observed due to local clustering of the spins in the FM network below TC for SCCMO_B system and due to the growth of AFM interactions for the SCCMO_N system. In addition, the magnetic moment of bulk sample is higher in magnitude compared to that of the nanometric sample as a result of less disorder effect in bulk. The ZFC and FC magnetization curves measured in 100 Oe dc magnetic field bifurcate at the irreversibility temperature (Tirr) ∼ 116 K and ∼110 K for SCCMO_B and SCCMO_N, respectively. The irreversibility temperature decreases with decreasing particle size. The magnetic hysteresis loop, M(H), of SCCMO_B and SCCMO_N at 5 K presented in Fig. 5(b), exhibits a prominent soft FM like loop for SCCMO_B and AFM or mixture of FM/canted AFM like loop for SCCMO_N sample. It is found that the saturation magnetization of SCCMO_B at 5 K is 21.98 emu/gm. It is seen that the coercive field (HC) and remanent magnetization (MR) are strongly grain size dependent. According to our results, HC ∼ 9.18 kOe and MR ∼ 8.69 emu/gm for SCCMO_B and that for SCCMO_N are 8.15 kOe and 2.62 emu/gm, respectively. As discussed above, our samples have spin disorder due to Co3+/Mn3+ along with ASD. The percentage of ASD can be calculated using the formula,

spectra of SCCMO_B sample exhibits two notable peaks, one peak of A1g symmetry at 647.22 ± 0.37 cm−1 (full width of half maxima (FWHM): 47.14 ± 1.19 cm−1) corresponding to symmetric stretching vibrations of the (Co/Mn)O6 octahedron and the other at 518.36 ± 0.71 cm−1 (FWHM: 60.88 ± 2.32 cm−1) ascribing to the mixed nature of antistretching and/or bending mode vibrations with B1g symmetry. Similarly for SCCMO_N sample, the stretching mode of A1g symmetry is noticed at 622.16 ± 0.10 cm−1 (FWHM: 78.10 ± 0.36 cm−1) and the peak for the mixed nature of anti-stretching and/or bending mode vibrations with B1g symmetry is observed at 504.01 ± 0.47 cm−1 (FWHM: 110.63 ± 2.21 cm−1) of (Co/Mn)O6 octahedron. The grain size changes the intensity, position and shape of all peaks. The Raman shift and FWHM of A1g and B1g modes have been calculated using Lorentzian fit to the spectral data. Insets of Fig. 4(a) and (b) present the grain size variation of Raman shift of A1g and B1g modes, respectively. It is noticed that both A1g and B1g peaks shift to higher wavelengths as grain size increases. However, it is observed that the parameter FWHM

MS (ASD , x ) = (1−2ASD){MMn + MCo} + x (2ASD−1)

(1)

where MMn and MCo are the magnetization values of Mn and Co ions, respectively and the term x indicates the concentration of Ca2+ [29]. The first term in the right side of above equation represents the contribution of ASD to MS and the second term signifies a reduction in MS due to Ca2+ substitution. According to the above expression (Eq. (1)), we have calculated the percentage of ASD present of 7.12% and 53.58% in SCCMO_B and SCCMO_N systems, respectively. In SCCMO_N system, the M(H) loop is characteristic of a typical canted AFM like ordering [30]. Presence of large ASD in SCCMO_N system can cause additional AFM interactions (Co2+ –O2––Co2+ or Mn3+ –O2––Mn3+); this may result in reduction of MS, and such interactions get modulated by particle size. Interestingly, the isothermal ZFC magnetization curve exhibits small 164

Journal of Magnetism and Magnetic Materials 469 (2019) 161–170

R.C. Sahoo et al.

Sm/Ca Mn/Co O

O1 O1

O1 1.71 Å

2.11 Å

O2

Mn/Co Mn M n/C /Co Mn/Co

Sm/Ca Mn Co O

O2

(a) SCCMO_B

Mn/Co Mn/ Mn /Co Mn/Co

O3

1.90 Å

O3

O3

O2 O1

(b)

50

60

SCCMO_B

SCCMO_N

40

Weight (%)

Atom (%)

40

20

SCCMO_N SCCMO_B

(c)

0

SCCMO_N

Sm

Ca

(d)

30

SCCMO_N SCCMO_B

20 10

Co Mn Element

0

O

Sm

Ca

Co Mn Element

O

Fig. 2. . (a) Local environment and crystal structure (in inset of 2(a)) of SCCMO_B sample. (b) Local environment and crystal structure (in inset of 2(b)) of SCCMO_N sample. Inset of (c) and (d) shows EDAX spectra of SCCMO_B and SCCMO_N, respectively; comparison of EDAX data analysis showing the presence of elements (c) Atom (%) (d) Weight (%).

typical feature of GP singularity [4,5], which is a current concern of interest regarding the physics of disorder oxides. By definition, TG is a temperature at which the χ−1 (T) curve deviates from Curie-Weiss behavior [6]. We have obtained TG ∼ 133 K and 161 K for SCCMO_B and SCCMO_N, respectively. The shaded GP region in Fig. 6(a) and (b) are characterized by the short range FM domains in a PM matrix. It appears that in SCCMO_N, the strength of FM interactions in GP region is very strong than that of the SCCMO_B. Presence of large ASD and mixed valence states of 3d transition metal ions in SCCMO_N are most likely the origin of such behavior in GP region. The characteristic temperature dependence of inverse susceptibility in GP regime (TC < T < TG) can be described by the power law

hysteresis beyond TC (PM region) for both the samples. The upper inset and lower inset in Fig. 5(b) shows isothermal M(H) loops at 130 K (T > TC) of SCCMO_B and SCCMO_N sample, respectively. The M(H) loop shows a different shape from expected nature in a pure PM region. This demonstrates that the short range FM entities play an important role in the contribution of magnetization beyond TC (PM region). The high temperature magnetization data for both the samples follow a Curie-Weiss behavior as shown in Fig. 6(a) and (b). These are demonstrated by the linearity of χ−1 versus temperature plots in the high temperature regime. The values of the paramagnetic Curie temperature, ΘCW ∼ 91.86 K and −66.74 K are obtained from the CurieWeiss fitting for SCCMO_B and SCCMO_N, respectively. ΘCW is mostly determined by the strength of magnetic interaction between Co and Mn ions in the PM state. It is obvious that the ΘCW for SCCMO_B is smaller than TC but positive. In case of SCCMO_N, ΘCW is several times smaller (negative) than the actual FM ordering temperature. Such large negative ΘCW for SCCMO_N sample is a signature of dominating AFM interactions in the FM network. The derived effective paramagnetic moment (μeff) from Curie-Weiss fit in the PM regime of the samples are found to be 5.91 μB (SCCMO_B) and 16.34 μB (SCCMO_N). According to theoretical calculations [31], the effective PM moment for the different spin state of Co (+2 and +3) and Mn (+3 and +4) is ∼5.72 μB and 5.98 μB for SCCMO_B and SCCMO_N sample, respectively. Such difference of μeff can be explained well in terms of presence of ferromagnetically correlated short range clusters in PM state and such clusters are more in SCCMO_N system [23,31,32]. Upon cooling, a sharp downward deviation of χ−1 is observed for both the samples and it deviates from Curie-Weiss linear behavior. This downward deviation of χ−1 is a

χ −1 ∝ Rand

1 ∝ (T −TCRand )1 − λ , M

(2)

where TC is the actual FM ordering temperature where susceptibility tend to diverge. The strength of the GP and the degree of deviation from Curie-Weiss law can be evaluated from the value of λ (0 ≤ λ ≤ 1) [1-5]. In the PM region, Eq. (2) modifies to the standard CurieWeiss equation for λ = 0. The inverse susceptibility in the GP regime obeys the modified Curie-Weiss behavior of Eq. (2) with λ≠0. To characterize the GP, it is very important to select TCRand for the correct estimation of λ (=0 in PM regime and ≠0 in GP regime). For GP in oxides reported so far, the Curie-Weiss temperature (ΘCW) and TCRand are found to be very close [10,33,34]. Therefore, we have chosen TCRand = ΘCW = 91.86 K and −66.74 K for SCCMO_B and SCCMO_N, respectively. To find out λ, we have plotted χ−1 as a function of (T − TCRand) in log10-log10 scale (shown in the inset of Fig. 6(a) and (b)) and 165

Journal of Magnetism and Magnetic Materials 469 (2019) 161–170

R.C. Sahoo et al.

Data Fit Co2+ Co3+

641.10 eV

(a)

780

778

776

Data Fit Co2+ Co3+ 779.03 eV

644

774

642

640

638 Data Fit Mn3+ Mn4+

Mn-2p3/2 640.84 eV

782

Co-2P3/2

(d)

(c)

SCMCO_N

784

(b)

SCMCO_N

642.38 eV

784

Data Fit Mn3+ Mn4+

642.40 eV

779.01 eV

Mn-2p3/2

SCMCO_B

780.2 eV

Intensity (a.u.)

780.15 eV

Co-2P3/2

SCMCO_N

782

780

778

776

774

644

642

640

638

B. E. (eV) Fig. 3. . Typical XPS spectra of Co 2p3/2 ((a) and (c)) and Mn 2p3/2 ((b) and (d)) core-levels at room temperature of both SCCMO_B and SCCMO_N samples, respectively.

the slope of fitted straight lines (using Eq. (2)) both in PM state and GP regime provides the values of λPM and λGP, respectively. The slope of the fitting lines in the two regimes provides λPM ∼ 0 and λGP ∼ 0.9 for SCCMO_B system and λPM ∼ 0 and λGP ∼ 0.63 for SCCMO_N system, respectively. The values of λGP are comparable to those values obtained for different oxides [11,33,34]. The obtained value of λGP for SCCMO_B is much larger than that for SCCMO_N. It is also noticed that the shaded GP region (95 K < T < 161 K) in SCCMO_N is much wider as shown in Fig. 6(b) in comparison with SCCMO_B sample (121 K < T < 133 K) as shown in Fig. 6(a). Both these behaviors are found to be in contrast to the conventional results mainly due to the presence of ASD and are not related to the crystal symmetry (space group) as grain size does not affect it [6,10]. The observed low value of λGP, large GP area and large ASD content indicate that GP singularity is sensibly strong in SCCMO_N. Bray [35] also extended these arguments that GP singularity can be noticed in the random magnetic system containing bond probability distribution that reduces the long-range ordering transition

temperature. The true PM state (except negligible structural anisotropy) of SCCMO_ B and SCCMO_N system has been observed above 133 K and 161 K, respectively.

3.6. Magnetic relaxation dynamics: Aging effect Now we have investigated the magnetic relaxation dynamics of the samples. The aging effect is clearly observed near the ground state and also well below the GP transition temperature of the cluster moments. To measure relaxation, the samples were cooled down to 30 K from above the TC in a FC mode with H = 100 Oe, and then the magnetic field was switched off instantly after a waiting time tw and M(t) was recorded. Various functional form has been introduced to describe the dependence of M(t) as a function of both waiting time and observation time. Finally, the M(t) curve of SCCMO_B is found to exhibit conclusively most popular stretched exponential dependence due to the glassy nature of the sample in the low temperature regime. On the other

Table 2. Binding energies (eV) and percentage contribution of core-level electrons of both the SCCMO_B and SCCMO_N samples from XPS studies. Samples

Co2p3/2

Mn2p3/2

Co2+

SCCMO_B SCCMO_N

Co3+

Mn3+

Mn4+

B.E. (eV)

(%)

B.E. (eV)

(%)

B.E. (eV)

(%)

B.E. (eV)

(%)

780.15 780.20

55.45 54.62

779.01 779.03

44.55 45.38

641.10 640.84

33.82 66.20

642.40 642.38

66.18 33.80

166

Journal of Magnetism and Magnetic Materials 469 (2019) 161–170

R.C. Sahoo et al.

A1g

A1g mode

SCCMO_B

70

FWHM

65

640

60

635

where M0 is an intrinsic FM component, Mr is the initial magnetization of the glassy component mostly contributing to the relaxation effects noticed, tr is the time constant, and n (0 < n < 1) is related to the relaxation rate [20,36]. The extracted fitting parameters are listed in Table 3. The non zero values of n and M0 clearly signifies the coexistence of FM/SG components in the relaxation process. The most significant component of this analysis, M0, decreases with increasing tW from 2.36(6) emu/gm to 1.378(1) emu/gm. The exponent, n, also gradually decreases with the increase of tW, indicating less time required to come back to an equilibrium state from a frustrated state below SG temperature in a magnetic field. Another interesting feature is that Mr/M0 decreases with the increase of tW, suggesting a change in phase fraction where SG phase reduces and the FM component enhances as a result of FC [37]. Fig. 7(b) displays M(t) curve for different waiting times (tw = 102, 103, and 5 × 103 sec) at 30 K (< < TC) for SCCMO_N sample and the solid (red) lines are power law fits of the form,

55

630

50

625

45

620 SCCMO_N

--

SCCMO_B

Intensity (a.u.)

(a)

520

120 B1g mode FWHM

110 100

FWHM (cm -1 )

Raman shift (cm -1 )

516

SCCMO_N

(3)

-1

FWHM (cm )

-1

Raman shift (cm )

645

n

{ ( )}

M (t , tW ) = M0−Mr exp − t tr

75

650

B1g

512

90 80

508

70 504

60

SCCMO_N

--

SCCMO_B

(b)

M (t , tW ) = M0 t −β

300

(4)

where M0 and β are power law fitting constants [35]. The values of M0 and β obtained from these fits are shown in Table 3. The β value decreases from 0.0026(2) to 0.00038(1) with an increase of tW whereas M0 (∼1.396(3) emu/gm) remains constant with the tW. The value of the exponent is two/three orders magnitude larger in SG phase (SCCMO_B system) than that in AFM phase (SCCMO_N system).

450 600 750 900 -1 Raman shift (cm )

Fig. 4. . Room temperature Raman spectra of (a) SCCMO_B and (b) SCCMO_N samples; inset shows comparative Raman shift and FWHM variation of the samples for (a) A1g and (b) B1g mode; lines are a guide to the eye only.

3.7. Magnetic ac susceptibility: Real and imaginary part hand, the M(t) curve of SCCMO_N sample is found to display power law dependence due to the presence of strong AFM couplings near ground state. The SCCMO_B sample, below 120.5 K, enters FM phase from random GP phase. On further lowering of temperature below ∼60 K, it enters spin-glass phase (shown in Fig. 8). Near ground state region, SCCMO_B system has a mixed state of FM and spin-glass (SG) phases. Fig. 7 (a) displays M(t) curve for different waiting times (tw = 102, 103, and 5 × 103 sec) at 30 K (below glassy temperature) for SCCMO_B sample and the solid (red) lines are stretched exponential fits of the form,

(a)

0

80

160 T (K)

240

300

M (emu/gm)

150 225 T (K)

ZFC_SCCMO_B FC_SCCMO_B ZFC_SCCMO_N FC_SCCMO_N

2 0

75

0 -1

0

-15

-10

-5

0 5 H (kOe)

10

15

3

-20 -40

320

-80

SCCMO_N @ 130 K

2

(b)

SCCMO_B SCCMO_N ZFC @ 5 K

-40

M (emu/gm)

dM/dT

-0.21 0

SCCMO_B @ 130 K

-2

TC= 120.5 K

4

20

TC=94.3 K

-0.14

2 1

SCCMO_B SCCNO_N ZFC

-0.07

6 M (emu/gm)

40

0.00

M (emu/gm)

8

In order to characterize the nature of spin glass ordering in the SCCMO_B sample, temperature dependent real (χ′) and imaginary part (χ″) of ac susceptibility (ACS) (χ = χ′ + i χ″) measurements have been carried out at different frequencies (f) with an ac field of 10 Oe. Fig. 8(a) shows the χ′(T) component of ACS near low temperature, where peak temperatures (Tf) increases with f. To characterize the origin of Tf, primarily we have calculated the value of ΔTf/ Tf Δ(log (2πf)). The estimated value of ΔTf/ Tf Δ(log(2πf)) between two neighboring highest frequencies is 0.042, the value of which lies in the range of (0.03–0.08) similar to the values as generally observed for insulating

1 0 -1 -2 -3

-15

0 H (kOe)

-10

-5

0 5 H (kOe)

40

10

15

80

Fig. 5. . (a) Temperature dependence of magnetization for 100 Oe field under ZFC and FC mode of both the samples; inset shows dM/dT vs. T plot in ZFC mode. (b) Isothermal Field dependence of ZFC magnetization at 5 K for both the samples; upper inset and lower inset shows M(H) curve at 130 K (> TC) of SCCMO_B and SCCMO_N samples, respectively. 167

Journal of Magnetism and Magnetic Materials 469 (2019) 161–170

R.C. Sahoo et al.

PM

-0.5

5000

GP region

0.0

=0.9

GP

0.5 1.0 1.5 Rand Log (T-TC )

TC=121 K =91.86 K CW

2.0

SCCMO_B

-1

1.5

TG=133 K

75

150 225 T (K)

SCCMO_N

TG=161 K

3000 3.6

2000 =

45 .74 6 -6

CW

1000

TC=95 K

=0

PM

3.0

gion PM re

R C

T = -66.74 K

2.4 1.8

n GP regio

=0.63

(a)

0 0

(b)

4000

2.0

10000

-1

=0

)

2.5

T =91.86 K

5000

ion reg

(Oe-gm/emu)

) -1

15000

3.0

Log (

(Oe-gm/emu)

20000

3.5

PM

R C

-1

4.0

Log (

25000

1.2

0

300

0

GP

2.0

2.2 2.4 Rand Log (T-TC )

50 100 150 200 250 300 T (K)

Fig. 6. . Inverse susceptibility at 100 Oe field with CW analysis holds only at T > > TC; Inset displays log-log plot of the power law analysis of magnetic susceptibility, 1/χ(T)∝(T-TCRand)1-λ and the solid lines represent the fit to the experimental data following the power law for (a) SCCMO_B and (b) SCCMO_N double perovskites.

SG phase in low temperature regime. Origin of this typical SG phase behavior is attributed due to the presence of ASD along with both structural and magnetic frustration in this system. It is also noticed here that no such behavior has been observed in case of SCCMO_N system due to the presence of strong AFM interaction. The imaginary part of ac susceptibility (χ″) also agrees well with the SG behavior of SCCMO_B compound. Fig. 8(b) shows the temperature dependent χ″(T) at various frequencies. It is also clear that the peak temperature also significantly shifts to the higher temperature with increasing frequency. We have also analyzed the frequency dependent shift of the peak temperature by using the empirical Vogel-Fulcher (VF) law of the form

SG compounds [37]. For further confirmation of nature of SG phase of SCCMO_B sample, the f dependence of Tf in χ′(T) is analyzed by employing dynamic scaling law of the form, −zv

Tf τ = τ0 ⎛⎜ −1⎟⎞ ⎝ Tg ⎠

(5)

where τ (∼1/f) is the spin relaxation time corresponding to frequency f, τ0 is the spin flipping time, zv is the dynamic critical exponent, Tg is equivalent to the freezing temperature of SG ordering (as f → 0 and Hdc → 0) and Tf is the frequency dependent freezing temperature determined by the maximum in χ′(T). The best fit (as shown inset of Fig. 8(a)) produces the flipping (relaxation) time τ0 ∼ 4.285 × 10−13 sec, SG freezing temperature Tg ∼ 34.88(3) K and critical exponent zv ∼ 10.99(5). We note that τ0 lies in the range of 10−12 to 10−14 sec and zv lies in the range of 4–12 in the case of canonical SG systems [3,22]. The estimated very small flipping time and dynamic critical exponent suggest that our SCCMO_B system goes to a pure (canonical)

EA ⎫ ω = ω0exp ⎧− ⎨ ⎩ KB (Tf −TVF ) ⎬ ⎭

where ω0, TVF and EA are the measuring frequency, VF temperature and activation energy, respectively [3]. The best fit of Tf from χ″(T) using

2.70

1.395

2.76

1.390

2

M (emu/gm)

Fit

t = 10 sec W

3

t = 10 sec W

4

t = 10 sec W

3.6

M (emu/gm)

SCCMO_B

2.82

SCCMO_N

1.385

Fit

2

t = 10 sec W

3

t = 10 sec W

4

t = 10 sec W

1.380

4.0 4.4

(6)

(b)

(a)

0

2

4

6 8 10 12 14 16 2 t (10 sec)

1.375

0

2

4 2 6 t (10 sec)

8

10

Fig. 7. . Plot of the time dependence of isothermal (at 30 K) magnetization measured in the FC mode for different waiting times (tW) 102, 103, and 5 × 103 sec; the solid red lines are the best fit for (a) SCCMO_B and (b) SCCMO_N double perovskites. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) 168

Journal of Magnetism and Magnetic Materials 469 (2019) 161–170

R.C. Sahoo et al.

0.09

8

Data Fit using power law

ln (f)

0.03

' (a.u.)

45

50

5 4

0.00

40

2001 Hz 1501 Hz 1001 Hz 501 Hz 101 Hz 11 Hz

6

46

47 48 Tf (K)

49

50

" (a.u.)

(sec)

0.06

2001 Hz 1001 Hz 501 Hz 101 Hz 11 Hz

Data Fit using VF law

7

3 2

0.060

0.063 0.066 -1 1/(Tf-TVF) (K )

0.069

SCMCO_B SCMCO_B

(b)

(a)

60 70 T (K)

30

80

40

50

60

T (K)

Fig. 8. . (a) Temperature dependence real part of χ′(T) for different frequencies; inset shows Power law fit (solid red line) using Eq. (5) to the spin freezing temperature, (b) Imaginary part of χ″(T) for different frequencies with an ac field of 3 Oe of SCCMO_B; inset shows Vogel-Fulcher law fit (solid red line) to the spin freezing temperature. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Table 3 . Fit parameters of the magnetic relaxations at 30 K (< < TC) for different waiting times described in Fig. 7(a) and (b) for both SCCMO_B and SCCMO_N samples. tW (sec)

102 103 5 × 103

SCCMO_B

Table 4 . Comparison table for the expected magnetic interactions (Co-O-Mn or Co-O-Co or Mn-O-Mn) in SCCMO_B and SCCMO_N samples.

SCCMO_N

M0 (emu/ gm)

Mr (emu/gm)

1.396(3) 1.396(3) 1.396(3)

0.453(6) 0.129(4) 0.0022(9)

tr (sec)

13,019 992 569

n

0.799 0.678 0.621

M0 (emu/ gm)

β

1.396(3) 1.396(3) 1.396(3)

0.0026(2) 0.00089(2) 0.00038(1)

Co2+ (S = 3/2) Co3+ (S = 2 or 1) Co3+ (S = 0) Mn3+ (S = 2) Mn4+ (S = 3/2)

VF law is shown in the inset of Fig. 8(b). The obtained fitting parameters are EA/KB ∼ 524.7(2) K, TVF ∼ 35 K and f = 2.2 × 1013 Hz. In this framework, TVF and τ ∼ 1/f ∼ 4.5 × 10−14 sec are consistent with the observed results from χ′(T) analysis. Also TVF < < EA/KB, indicates a weak coupling between the interacting entities (FM/SG) in SCCMO_B. All the above observations suggest that our SCCMO_B system has canonical SG like freezing below 34.88(3) K with very small flipping time.

Mn3+ (S = 2)

Mn4+ (S = 3/2)

Co2+ (S = 3/2)

Co3+ (S = 2 or 1)

Co3+ (S = 0)

AFM AFM

FM FM

AFM AFM

AFM AFM

FM FM

FM AFM FM

AFM FM AFM

FM AFM FM

AFM AFM FM

AFM FM AFM

O-Mn4+ magnetic interactions may contribute ferromagnetism. The AFM ground state of SCCMO_N compound is attributed to large Co2+O-Co2+ and Mn3+-O-Mn3+ exchange interactions. The origin of these improved exchange interactions is due to the existence of large ASD. In this compound, the observed GP can be assigned to the Co-Co/Mn-Mn ASD pairs which causes random dilution of the B-site ordered lattice (Co-Mn), thus favoring short-range FM ordering of the Co/Mn sublattice. However, the SG state of SCCMO_B emerges in the presence of both FM interactions and less AFM interactions of Co2+-O-Co2+ and Mn4+-O-Mn4 resulting from the existence of large amount of Co2+ and Mn4+ ions. These competing magnetic interactions (FM and AFM) make the system frustrated and break the long-range ferromagnetism. Also the variety of magnetism is influenced by structural disorder (distinct bond lengths and bond angles) in the compounds, thus hindering the formation of long-range ferromagnetism.

3.8. Role of 3d transition metals in the compounds The 3d transition metal ions (Co and Mn) play a crucial role in structural and magnetic properties of these double perovskite systems. One of the species Co is typically magnetic while Mn species is generally nonmagnetic in the compounds. According to our XPS study, Co has both divalent and trivalent states, showing an electronic configuration 3d7 (t2g5eg2 (S = 3/2) and 3d6 (t2g4eg2 (S = 2)/ t2g5eg1 (S = 1)/ t2g6eg0 (S = 0)), respectively. Moreover, Mn is a mixed valence ion with trivalent and tetravalent states, showing an electronic configuration 3d4 (t2g4 (S = 2)) and 3d3 (t2g3 (S = 3/2)), respectively. The presence of both Co2+/Co3+ and Mn3+/Mn4+ ions influence the different magnetic correlations by competing magnetic interactions. Table 4 shows the possible combinations of magnetic interactions due to the present spin states of Co and Mn ions in the compounds. All these combinations of magnetic interactions are both temperature and composition dependent [38,39]. The Co2+-O-Mn4+ and Co3+(S = 0)-O-Mn3+(S = 2) interactions give rise to FM superexchange interactions through O2– in both the compounds. Also, Co2+(S = 3/2)-O-Co3+(S = 0) and Mn3+-

4. Conclusions We have shown that Sm1.5Ca0.5CoMnO6 double perovskite is a multi-magnetic phase material, which is tuned by grain size effect. Both the values of effective paramagnetic Curie temperatures and magnetization curves indicate that the SCCMO_N system is AFM and SCCMO_B is a FM in nature. The interesting complex magnetism of both the systems has been attributed to the ASD effect, structural disorder and magnetic frustration. The observed GP region (95 K < T < 161 K) in 169

Journal of Magnetism and Magnetic Materials 469 (2019) 161–170

R.C. Sahoo et al.

SCCMO_N is much larger than that of SCCMO_B (121 K < T < 133 K) and it mainly results from the presence of ASD effect. In the GP regime, the short-range FM correlations and dynamic FM clusters may be attributed to the finite magnetic moment in PM state. We have observed an anomaly in both χ′(T) and χ″(T) data near Tf of SCCMO_B system only. This result confirms the canonical SG like ordering in this SCCMO_B system and it is attributed to the competing magnetic interactions (FM and AFM). The aging effect in SCCMO_B system also confirms its magnetic glassy state at the ground state. On the other hand, the aging effect in SCCMO_N system proves AFM like features at the ground state.

[12] A.H. Castro Neto, G. Castilla, B.A. Jones, Phys. Rev. Lett. 81 (1998) 3531. [13] S. Guo, D.P. Young, R.T. Macaluso, D.A. Browne, N.L. Henderson, J.Y. Chan, L.L. Henry, J.F. DiTusa, Phys. Rev. Lett. 100 (2008) 017209. [14] D. Choudhury, P. Mandal, R. Mathieu, A. Hazarika, S. Rajan, A. Sundaresan, U.V. Waghmare, R. Knut, O. Karis, P. Nordblad, D.D. Sarma, Phys. Rev. Lett. 108 (2012) 127201. [15] H. Sakai, J. Fujioka, T. Fukuda, D. Okuyama, D. Hashizume, F. Kagawa, H. Nakao, Y. Murakami, T. Arima, A.Q.R. Baron, Y. Taguchi, Y. Tokura, Phys. Rev. Lett. 107 (2011) 137601. [16] M.V. Mostovoy, F.M. Marchetti, B.D. Simons, P.B. Littlewood, Phys. Rev. B 71 (2005) 224502. [17] M. Uehara, S. Mori, C.H. Chen, S.W. Cheong, Nature 399 (1999) 560. [18] V.N. Singh, P. Majumdar, Eur. Phys. Lett. 94 (2011) 47004. [19] D.D. Sarma, Sugata Ray, K. Tanaka, M. Kobayashi, A. Fujimori, P. Sanyal, H.R. Krishnamurthy, C. Dasgupta, Phys. Rev. Lett. 98 (2007) 157205. [20] M. Patra, S. Majumdar, S. Giri, J. Phys. Condens. Matter 22 (2010) 116001. [21] L. Wang, W. Zhou, D. Wang, Q. Cao, Q. Xu, Y. Du, J. Appl. Phys. 117 (2015) 17D914. [22] S.K. Giri, R.C. Sahoo, Papri Dasgupta, A. Poddar, T.K. Nath, J. Phys. D Appl. Phys. 49 (2016) 165002. [23] P.R. Mandal, R.C. Sahoo, T.K. Nath, Mater. Res. Express 1 (2014) 046108. [24] V.L. Joseph Joly, Y.B. Khollam, P.A. Joy, C.S. Gopinath, S.K. Date, J. Phys. Condens. Matter 13 (2001) 11001. [25] G.D. Khattak, M.A. Salim, A.S. Al-Harthi, D.J. Thompson, L.E. Wenger, J. Non-Cryst, Solids 212 (1997) 180. [26] J.K. Murthy, K.D. Chandrasekhar, H.C. Wu, H.D. Yang, J.Y. Lin, A. Venimadhav, J. Phys. Condens. Matter 28 (2016) 086003. [27] R. Havlicek, J. Poltierova, D. Vejpravova Bochenek, J. Phys. 200 (2010) 0120581. [28] E. Vincent, V. Dupuis, M. Alba, J. Hammann, J.-P. Bouchaud, Eur. Phys. Lett. 50 (2000) 674. [29] J. Navarro, C. Frontera, L.l. Balcells, B. Martinez, J. Fontcuberta, Phys. Rev. B 64 (2001) 092411. [30] M.-R. Li, J.P. Hodges, M. Retuerto, Z. Deng, P.W. Stephens, M.C. Croft, X. Deng, G. Kotliar, J. Sanchez-Benitez, D. Walker, M. Greenblatt, Chem. Mater. 28 (2016) 3148. [31] R.C. Sahoo, D. Paladhi, A. Papri Dasgupta, Ripandeep Singh Poddar, A. Das, T.K. Nath, J. Magn. Magn. Mater. 428 (2017) 86. [32] S.K. Giri, S.M. Yusuf, M.D. Mukadam, T.K. Nath, J. Alloys Comp. 591 (2014) 181. [33] W. Jiang, X. Zhou, G. Williams, Eur. Phys. Lett. 84 (2008) 47009. [34] A.K. Pramanik, A. Banerjee, Phys. Rev. B 81 (2010) 024431. [35] A.J. Bray, Phys. Rev. Lett. 59 (1987) 586. [36] R.V. Chamberlin, G. Mozurkewich, R. Orbach, Phys. Rev. Lett. 52 (1984) 867. [37] J.A. Mydosh, Spin Glasses: An Experimental Introduction, 1993, Taylor and Francis, London. [38] J.B. Goodenogh, A. World, R.J. Arngtt, N. Mknyuk, Phys. Rev. 124 (1961) 2. [39] P.A. Joy, Y.B. Khollam, S.K. Date, Phys. Rev. B 62 (2000) 8608.

Acknowledgments RCS acknowledges MHRD, India, for financial assistance and Central Research Facility (CRF), IIT Kharagpur, India for experimental measurement equipment supports. The authors gratefully acknowledge G. Gehring for fruitful discussion. References [1] L. Lundgren, P. Svedlindh, P. Nordblad, O. Beckman, Phys. Rev. Lett. 51 (1983) 911. [2] N. Khan, P. Mandal, D. Prabhakaran, Phys. Rev. B 90 (2014) 024421. [3] R.C. Sahoo, S.K. Giri, D. Paladhi, A. Das, T.K. Nath, J. Appl. Phys. 120 (2016) 033906. [4] R.C. Sahoo, Sananda Das, T.K. Nath, J. Appl. Phys. 123 (2018) 013902. [5] J. Deisenhofer, D. Braak, H.-A. Krug von Nidda, J. Hemberger, R.M. Eremina, V.A. Ivanshin, A.M. Balbashov, G. Jug, A. Loidl, T. Kimura, Y. Tokura, Phys. Rev. Lett. 95 (2005) 257202. [6] R.B. Griffiths, Phys. Rev. Lett. 23 (1969) 17. [7] J.M. De Teresa, M.R. Ibarra, P.A. Algarabel, C. Ritter, C. Marquina, J. Blasco, J. Garcia, A. Del Moral, Z. Arnold, Nature (London) 386 (1997) 256. [8] C. Magen, P.A. Algarabel, L. Morellon, J.P. Araujo, C. Ritter, M.R. Ibarra, A.M. Pereira, J.B. Sousa, Phys. Rev. Lett. 96 (2006) 167201. [9] Z.W. Ouyang, V.K. Pecharsky, K.A. Gschneidner Jr., D.L. Schlagel, T.A. Lograsso, Phys. Rev. B 74 (2006) 094404. [10] Z.W. Ouyang, N.M. Xia, Y.Y. Wu, S.S. Sheng, J. Chen, Z.C. Xia, L. Li, G.H. Rao, Phys. Rev. B 84 (2011) 054435. [11] E.V. Sampathkumaran, Niharika Mohapatra, Sudhindra Rayaprol, Kartik K. Iyer, Phys. Rev. B 75 (2007) 052412.

170