Copper-doped lanthanum manganite La0.65Ce0.05Sr0.3Mn1−xCuxO3 influence on structural, magnetic and magnetocaloric effects

Bulk nanocrystalline samples of La0.65Ce0.05Sr0.3Mn1−xCuxO3 (0 ≤ x ≤ 0.15) manganites are prepared by the sol–gel based Pechini method. The effect of the substitution for Mn with Cu upon the structural and magnetic properties has been investigated by means of X-ray diffraction (XRD), Raman spectroscopy and dc magnetization measurements. The structural parameters obtained using Rietveld refinement of XRD data showed perovskite structures with rhombohedral (R3̄c) symmetry without any detectable impurity phase. Raman spectra at room temperature reveal a gradual change in phonon modes with increasing copper concentration. The analysis of the crystallographic data suggested a strong correlation between structure and magnetism, for instance a relationship between a distortion of the MnO6 octahedron and the reduction in the Curie temperature, Tc. A paramagnetic to ferromagnetic phase transition at TC is observed. The experimental results confirm that Mn-site substitution with Cu destroys the Mn3+–O2−–Mn4+ bridges and weakens the double exchange (DE) interaction between Mn3+ and Mn4+ ions, which shows an obvious suppression of the FM interaction in the La0.65Ce0.05Sr0.3Mn1−xCuxO3 matrix. The maximum magnetic entropy change −ΔSmaxM is found to decrease with increasing Cu content from 4.43 J kg−1 K−1 for x = 0 to 3.03 J kg−1 K−1 for x = 0.15 upon a 5 T applied field change.


Introduction
Recently, perovskite manganites of R 1Àx A x MnO 3 (where R and A are trivalent rare earth and divalent alkaline earth ions, respectively) have been the subject of intense research due to their interesting physical properties around the ferromagnetic (FM)-paramagnetic (PM) transition temperature (the Curie temperature, T C ), such as the colossal magnetoresistance (CMR), the magnetocaloric effects (MCE) (related to a large magnetic entropy change) and the strong correlation between structural and magnetic properties. Doped lanthanum based manganites have been used in many technological applications, including magnetic recording, high-density data storage, hard disks, magnetic sensors, spin-electronic devices, and magnetic refrigerants. [1][2][3][4] These materials offer a high degree of chemical exibility leading to complex interplay between structural, electrical and magnetic properties. The double exchange (DE) effect in which e g electrons transfer between adjacent Mn 3+ and Mn 4+ ions and the Jahn-Teller effect were used to understand FM-PM transition and CMR in manganites. 5,6 A prominent feature of most manganites is that they will undergo a ferromagnetic-paramagnetic (FM-PM) phase transition at the Curie temperature T C associated with an metal-insulator (M-I) transition at temperature T MI , which explains the fact that there exists a close relationship between the electrical and magnetic properties of manganites. 7,8 Average ionic radius, electronic conguration, valance state and the concentration of the doping element are important parameters for tuning the magnetic and electronic properties of these materials. 9 On the other hand, the synthesis technique greatly inuences the physical and chemical characteristics of the rare-earth perovskite materials. There are various methods to synthesize the manganites compounds among them the Pechini sol-gel method. This method has been used successfully to produce high-quality specimens due to these potential advantages such as better homogeneities, lower processing temperatures, short annealing times, high purity of materials and improved material properties.
In manganites, it is possible to dope at both R-site and Mnsite, much research has been done on the substitution at the Rsite with transition elements 10-12 and/or rare-earths Eu, 13 Ce, 14 Pr, 15 which can modify the Mn 3+ -O 2À -Mn 4+ network and in turn will intensively affect the intrinsic physical properties, such as ferromagnetism and (MCE). The substitution at the Mn site in perovskite oxides, with other transition metal ions, [16][17][18][19][20] is more important because it not only modies the Mn 3+ -O 2À -Mn 4+ network but also brings about many new exchange interactions between the Mn ion and the doped transition metal ions.
Amongst the doping at Mn sites with transition elements, Cu substitution has been particularly investigated because of the special nature of its variable valence. [21][22][23][24][25][26][27][28][29][30][31] In particular, Kim et al. 30 The stoichiometric amounts of precursors were dissolved in distilled water at 90 C and then a suitable amount of citric acid and ethylene glycol as coordinate agents were added. The resulting gel was pre-calcined (673 K for 3 h) to eliminate the organic material, ground and calcined again (973 K for 15 h) to eliminate the residual organic material. The obtained powder was then pressed into pellets (13 mm in diameter and 2-3 mm thick under a pressure of 5 ton cm À2 ). Aer that, the powder was sintered at 1173 K for 12 h in air.

Characterization
The morphological properties of the samples were investigated by scanning electron microscopy (SEM) on a JSM-6400 apparatus working at 20 kV. The structure and phase purity were checked by powder X-ray diffraction (XRD) using a "Panalytical X pert Pro" diffractometer with Cu K a radiation (k ¼ 1.5406 A). Data for Rietveld renement were collected in the range of 2 h from 10 to 120 with a step size of 0.017 and a counting time of 18 s per step. Raman scattering data was collected in the frequency range 100-1000 cm À1 using a Raman spectrometer. Magnetic measurements versus temperature and magnetic applied eld were realized using a SQUID (Quantum Design) developed at Louis Neel Laboratory of Grenoble. The isothermals M versus H at various temperatures around T C have been measured in applied elds up to 5 T.  15. We rst discuss the structural parameters of the studied samples in present work. Using the Rietveld renement method, we noted that all samples are single-phase with a rhombohedral structure of the R 3c space group (no. 167), in which the (La, Ce, Sr) atoms are at 6a (0, 0, 1/4) positions, (Mn, Cu) at 6b (0, 0, 0) and O at 18e (x, 0, 1/4). These results are consistent with the values of the Goldschmidt tolerance factor t G :

Structural properties
where r A , r B and r O are respectively the average ionic radii of A and B perovskite sites and of the oxygen anions. The tolerance factor is an important structural parameter, which reects the local microscopic distortion from the ideal perovskite (ABO 3 ) structure (t ¼ 1), for which the B-O-B bond angle q is equal to 180 . The values of t G were estimated and listed in Table 1.
The structural parameters were rened by the standard Rietveld renement method using the FullProf program. 32 We utilize the pseudo-Voigt function in order to t parameters to the experimental data set. The parameters used are: a scale factor, a zero shiing factor, three cell parameters, ve shapes and width of the peak factors, one global thermal factor and two asymmetric factors, the background was rened by a linear interpolation between a set background points with renable heights. The weighted prole factor R wp , the goodness of t c 2 , and the difference between the calculated and observed proles were evaluated at each renement cycle to determine the renement quality. The nal renement analysis shows that the experimental spectra and the calculated values obtained by the Rietveld renement are in good agreement with each other, and all observed peaks have been suitably indexed. The calculated results are shown in Fig. 1 30,31 The crystal structure and lattice parameters were affected because of the mismatch of ionic radius between the dopant and Mn ions. The B-site ionic radius of Cu 2+ (0.73 A) is larger than Mn 3+ (Mn 4+ ) and the ionic radius of Cu 3+ (0.54 A) is close to that of Mn 4+ (0.53 A) and smaller than the radius of the high spin state of Mn 3+ (0.645 A). 33 Further, it is expected that Cu 2+ ions substitute Mn 3+ ions and Cu 3+ ions substitute Mn 4+ ions. Furthermore, substitution of Cu 2+ for Mn 3+ and Cu 3+ for Mn 4+ would lead to a proportionate conversion of Mn 3+ to Mn 4+ and Mn 4+ to Mn 3+ . But the ionic state of Cu 2+ is expected to be dominant in the samples, therefore, the overall copper doping effect leads to a change in the relative fraction of different valence Mn ions, which results an increase in the number of Mn 4+ ions in order to preserve charge neutrality and therefore the unit cell volume is found to decrease.  Fig. 2(a).
The surface morphology of samples examined by scanning electron microscopy (SEM) is illustrated in Fig. 2(c). The SEM images show that the particles have an almost homogeneous distribution. The average crystallite size of the samples are obtained by applying the following Rietveld renement formula where k is the X-ray wavelength and IG is the Gaussian size parameter given by Rietveld renement. In all samples a nanometric size for the crystallites is found, between 77 nm and 94 nm (AE2 nm), which is related to the moderate synthesis temperatures of these samples, obtained from very reactive precursors from sol-gel procedures. These values are close to those shown by SEM micrographs (the average particles size is $100 (AE10 nm).

Raman spectroscopy
Raman spectroscopy is a powerful and sensitive tool for the non-destructive investigation and characterization of all kinds of materials. This technique is useful in understanding crystal symmetry, the local structural distortion and its dependence on doping. Our manganites samples shows rhombohedral crystal symmetry using the R 3c space group that assumes six equal distances of the Mn-O bonds of MnO 6 octahedra ( Fig. 2(b)). This structure can be described with respect to the ideal cubic structure by considering a rotation of MnO 6 octahedra about the [111] pseudo cubic diagonal. According to the group theory, for R 3c (D 3d 6 ) rhombohedral structure, thirty vibrational degrees of freedom at the G point are distributed among the irreducible representation as: The rhombohedral distortion gives rise to ve Raman active modes.
Room temperature Raman spectrum of as synthesized La 0.65 Ce 0.05 Sr 0.3 Mn 1Àx Cu x O 3 (x ¼ 0, 0.05, 0.10 and 0.15) samples in the frequency range of 200-900 cm À1 is shown in Fig. 3(a). Five vibration modes have been identied, one (A 1g ) and four (E g ). These broad bands are located at 162 (A 1g ), 302 (E g ), 424 (E g ), 460 (E g ) and 667-703 (E g ) cm À1 , which are associated with  rotational-, bending-, and stretching-like vibrations of the MnO 6 octahedra, respectively. 35,36 It has been noticed from the graph that with increasing Cu concentration, the Raman scattering intensity of the phonon modes are increasing. The frequencies of the experimental peaks are plotted with doping level for the high-frequency mode E g in Fig. 3(b). This mode shows a substantial shi toward lower frequencies (a downshi of about 30 cm À1 ) as a function of Cu concentration. These shis are related to the change in the average (Mn/Cu)-O distance. 37 Similar observations have been reported for some other perovskite compounds with distorted rhombohedral lattice. 35,38,39 In this work, we underline the E g mode allowed for the symmetric stretching vibration of oxygen in MnO 6 octahedra. It is reasonable to relate the changes on the E g mode frequency to the modications of the oxygen octahedral MnO 6 . Moreover, the introduction of substitutional defect in the B-site (like Cu) has a strong effect in the structural changes of the lattice. Since Cu substitution induces a strong local stress, it can be expected that (Mn/Cu)O 6 octahedra rotate, and the Mn-O bond lengths decrease under this compression. 40

Magnetic properties
The magnetization of La 0.65 Ce 0.05 Sr 0.3 Mn 1Àx Cu x O 3 (0 # x # 0.15) as a function of temperature from 5 K to 400 K under an applied eld of 100 Oe is shown in Fig. 4(a).
All samples exhibit a clear transition from paramagnetic to ferromagnetic state with decreasing temperature. The Curie temperature T C is the temperature at which the absolute value of dM/dT is maximum (inset Fig. 4(a)), are summarized in Table 2.
One can see that T C decreases monotonically with increasing Cu-doping content. This behavior is consistent with those reported previously on Mn-site substitution of copper. 22 3 , the magnetic susceptibility (c) could be tted to the Curie-Weiss law: c ¼ C/(T À q p ); where q p is the Curie-Weiss temperature (the temperature at which c À1 intercepts the temperature axis) and C is the Curie constant were determined by linear tting of the temperature dependent c À1 data in the high temperature paramagnetic region, as displayed in Fig. 4(b) and the values are given in Table 2.  From the estimated Curie constant (C), we have deduced the experimental effective moment (m exp eff ) using the following where N A ¼ 6.023 Â 10 23 mol À1 is the number of Avogadro, m B ¼ 9.274 Â 10 À21 emu is the Bohr magneton, M m is the molecular weight and k B ¼ 1.38016 Â 10 À16 erg K À1 is the Boltzmann constant.
The theoretical effective paramagnetic moment (m th eff ) was calculated using the calculated Mn 3+ /Mn 4+ contents under the assumption that all the Cu ions exist in either Cu 2+ or Cu 3+ state. The spin-only magnetic moments for free Mn 3+ , Mn 4+ , Cu 2+ and Cu 3+ are 4.89m B , 3.87m B , 1.73m B , 2.83m B , respectively. The obtained values of (m exp eff ) and (m th eff ) are listed in Table 2. The experimental (m exp eff ) value is little larger than the calculated value using the spin-only moment. Such a difference in (m eff ) value may be ascribed to the appearance of short-range FM interactions in the paramagnetic state. This result is commonly observed in manganites. 8,41,42 In order to investigate the magnetic behavior at low temperatures, we have performed magnetization measurements as a function of the applied magnetic eld m 0 H up to 5 T at various temperatures. We plot in Fig. 5 the magnetization evolution versus the applied magnetic eld obtained at different temperature (isothermal magnetization) for (a) x ¼ 0, and (b) x ¼ 0.15 samples. At a given lower elds, (M-H-T) curves show a rapid increase and get saturated at higher elds. For all the studied samples, the magnetization has been found to increase with decreasing temperature in the selected temperature range, where thermal uctuation of spins decreases with decreasing temperature.
To determine the nature of the FM-PM phase transition (rst or second order) for our samples, we plotted in Fig. 6(a and b) the Arrott plot 43 (m 0 H/M versus M 2 ) for x ¼ 0 and x ¼ 0.15. Table 2 Values of the Curie temperature T C , the Curie constant C, the Curie-Weiss temperature q P and the experimental and theoretical effective paramagnetic moment (m exp eff ) and (m th eff ) for La 0.65 Ce 0.05 Sr 0.3 Mn 1Àx Cu x O 3 (0 # x # 0.15)

Magnetocaloric effect and second-order magnetic phase transition
MCE is an intrinsic property of magnetic materials. It is the response of the material toward the application or removal of a magnetic eld. This response is maximized when the material is near its magnetic ordering temperature. In an isothermal process, the magnetic entropy change of the materials can be derived from the Maxwell relation as shown below: 45 The magnetic entropy changes, DS M , of La 0.65 Ce 0.05 Sr 0.3 -Mn 1Àx Cu x O 3 (x ¼ 0, 0.05, 0.1 and 0.15) have been calculated using the Maxwell relation 46 and are plotted in Fig. 7 as a function of temperature and eld.
The maximum value of magnetic entropy change DS M is found to be around T C and it increases with increasing the magnetic applied eld due to the enhancement of FM interactions. As the Cu content increases the magnitude of DS M decreases under a given eld strength. Indeed, under the magnetic eld change from 0 to 5 T, the values of |DS max M | observed for x ¼ 0.00, 0.05, 0.1 and x ¼ 0.15 are found to be 4.43, 5.15, 3.37 and 3.03 J kg À1 K À1 , respectively. The value of |DS max M | for m 0 H ¼ 1 and 5 T are listed in Table 2 along with related compounds 47,48 for easy comparison.
Relative cooling power (RCP) is another important parameter to quantify the efficiency of the magnetocaloric material. It is a measure of how much heat can be transferred between the cold and the hot tanks in one ideal refrigeration cycle. It can be dened as where d FWHM ¼ DT is the full-width at half maximum peak and ÀDS max M is the maximum value of magnetic entropy change which is occurred at Curie temperature.
The  Table 3.
The maximum value of RCP is obtained for x ¼ 0.05, which is 43% of that of pure Gd, the prototype magnetic refrigerant material (RCP ¼ 410 J kg À1 ). 48 Our results indicate that this compound is promising for room temperature magnetic refrigeration.
For further clarication of the phase transition of the samples as an alternative to the Banerjee criterion 44 we used the phenomenological universal curve method, proposed by Franco et al., 49,50 which is a general approach to determine the order of the phase transition. In order to construct this phenomenological universal curve, an analogous procedure to that described in ref. 51 has been used. It consists in normalizing the DS M curves with respect to their maximum and rescaling the temperature axis as where T R1 and T R2 are the two reference temperatures corresponding to the half maximum of DS M (T R1 ) ¼ DS M (T R2 ) ¼ DS max M /2. For the materials undergoing second order magnetic phase transition, the rescaled magnetic entropy change curves should follow a universal behavior. While the scaled DS M curves do not collapse as a single curve, the materials undergo a rstorder phase transition. 42   normalized entropy change curves collapse into a single universal curve, which conrms that the PM-FM phase transition observed for our compounds is of a second-order. Hence, this result is consistent with the trends observed in the Arrott plots (Fig. 6). The solid line corresponds to the average of the universal scaling. This average curve, once the temperature axis is back transformed from the reduced temperature to the unnormalized one, allows making extrapolations to lower temperatures for the high eld data and obtaining a ner description of the peak for the low eld curves. 52,53

Conclusion
In summary, we have studied the effect of copper doping lanthanum manganite ions on structural, magnetic and magnetocaloric properties of La 0.65 Ce 0.05 Sr 0.3 Mn 1Àx Cu x O 3 (0 # x # 0.15) prepared using the Pechini sol-gel method. Rietveld renement of XRD patterns shows that all samples crystallized in a rhombohedral structure with R 3c space group. The Cudoping induces the suppression of the one-electron bandwidth W of e g electron due the variations of the bond length and bond angle, leading to destruction of the DE interaction. The Curie temperature and the maximum magnetic entropy change decrease with the increase in the Cu content. This is attributed to the structural distortion of MnO 6 octahedron and the changes in the valence states of the Cu and Mn ions upon Cu doping, weakening the ferromagnetic exchange interaction. A uniform phenomenological function that describes the magnetic entropy change is found for these materials, which provides good handle on designing of magnetocaloric materials for micro magnetic refrigerators.
The La 0.65 Ce 0.05 Sr 0.3 Mn 0.95 Cu 0.05 O 3 sample is found to have a comparable MCE around 330 K with a maximum DS M of 1.34 J kg À1 K À1 and a RCP of 44 J kg À1 under a magnetic eld change of 1 T, and can be considered as competitive candidate for magnetic refrigerant materials operating near room temperature.

Conflicts of interest
There are no conicts to declare.