1. Introduction
The scientific community is actively engaged in materializing energy-efficient and environmentally friendly technologies to mitigate increasing energy demands. Low power refrigeration or cooling has emerged as a potential alternative for future energy-saving [
1,
2,
3,
4]. Common cooling techniques include absorption/adsorption refrigerators, ejector refrigeration systems and, less commonly, thermoelectric cooling, thermoacoustic refrigerators, and magnetocaloric (MC) refrigeration [
5]. Magnetocaloric effect (MCE) units have excellent prospects for cooling technology, due to their cooling efficiency, environmental friendliness, reliability, portability, cost-effectiveness, and long cyclic operational stability.
The magnetocaloric effect (MCE) is a magneto–thermodynamic phenomenon in which the thermal state of magnetic materials changes in response to an applied magnetic field [
6,
7]. At the heart of MCE lies the coupling between the magnetic moments and external magnetic field, and, in some cases, the MCE involves structural transitions concomitant with magnetic transitions [
8]. The magnetocaloric material (MCM) should have a high magnetic entropy change, a wide range of operating temperatures, low hysteresis loss, low specific heat, high thermal conductivity, etc., to produce a large MCE response, which is challenging to meet. Recent advances promise improved MCE efficiency and address challenges like fabrication complexity, less resistance to corrosion, poor conductivity, etc. [
9]. Different metallic alloys (e.g., iron-based alloy Gd
5Ge
1.9Si
2Fe
0.1) [
10] and composites e.g., La(Fe,Si)
13/CuLa(Fe,Si)
13H
y/In, Refs. [
1,
11] showed potential to overcome deficiencies, henceforth, improving the performance of MCE-based magnetic refrigeration or cooling.
However, these metallic materials are expensive, difficult to synthesize, and easily corrode, limiting their use in various applications. Recently, strontium-doped lanthanum manganite, with the general formula La
1−xSr
xMnO
3,
x = 0.0 < 1.0 (LSMO), has attracted much attention for its potential magnetocaloric materials at near room temperature [
12,
13,
14,
15,
16]. Current literature is replete with studies on RE
1−xMxMnO
3 type MCM. The search for oxide-based MCM ensued because they are corrosion-resistant, stress-resistant, easy to synthesize, have long cyclic stability and their magnetic properties can easily be tuned. Recently, it was reported that in multiphase composite materials, the temperature dependence of entropy change, −ΔS
M, curve broadens with enhanced relative cooling power (RCP) values [
17,
18,
19]. According to theoretical studies, −Δ
SM and RCP of the composites are predicted to show promising results [
20,
21,
22].
Large RCP values can be achieved by MCM composite with suitable transition temperatures as it undergoes multiple magnetic phase transitions [
22,
23]. El Boukili et al. reported that the transition temperature remained unchanged but decrease in ∆S
max and increase in RCP were recorded on increasing the content of Mn
2O
3 in (1 −
x)La
0.6Ca
0.4MnO
3/
xMn
2O
3 composite synthesized via solid-state reaction [
24]. The drop in transition temperature and ∆S
max values, along with the increase in RCP value by introducing Co
3O
4 in La
0.7Sr
0.3MnO
3/Co
3O
4 and La
0.7Ca
0.3MnO
3/Co
3O
4 composites synthesized via solid state reaction, was reported by Anwar et al. The antiferromagnetic (AFM) nature of Co
3O
4 induces magnetic inhomogeneity in the ferromagnetic perovskite to make these changes [
25]. Nasi et al. synthesized La
0.6Sr
0.4MnO
3/
x(Sb
2O
3) composites using a solid-state reaction and found that the ∆
Smax and RCP values dropped, compared to pure perovskite, due to the decrease in delocalization of the electron that minimizes the double-exchange interaction. However, the transition temperature remained constant, implying that the Sb
2O
3 phase remained only at the grain boundaries and did not react with La
0.6Sr
0.4MnO
3 to be included within the perovskite structure [
26]. Gavrilova et al. synthesized (1 −
x)La
0.7Sr
0.3MnO
3/
xNaF composites using a solid-state reaction and found that the magnetic entropy and transition temperature dropped due to the introduction of the secondary NaF phase, which was the contribution of grain size reduction. However, the RCP values did not change significantly, due to magnetic inhomogeneity [
27]. Neupane et al. reported that the values of ∆S
max increased, while the value of RCP decreased, with increasing metal oxides in La
0.45Nd
0.25Sr
0.3MnO
3/MO (MO=CuO, CoO, and NiO) composites synthesized via the autocombustion method. The proximity of the perovskite’s phase transition temperature and those of the metal oxides and the size of grains influenced the values of ∆S
max and RCP [
28]. An exchange bias interfacial FM/AFM effect was reported in (LaFeO
3)
x-(La
0.7Ca
0.3MnO
3)
1−x composite thin films prepared via pulsed laser deposition [
29]. At the same time, the impact of NiO particles was investigated on the magnetoresistance properties of LACM-NiO composites and it was realized that the NiO nanoparticles limited the charge transport in the composites to increase resistivity [
30]. Thus, understanding the role of the secondary oxide phase in the manganite composite is essential in designing high-performance magnetocaloric and spintronic materials, and for their applications [
31,
32].
The literature shows that composite formation causes decrease in ∆S
max but increase in the RCP by increasing the full-width-at-half-maximum (FWHM) temperature range of ∆
S vs.
T curve, enhancing the operating temperature range for the magnetocaloric effect. Furthermore, the composite formation can help tune the paramagnetic (PM)-ferromagnetic (FM) transition temperature. Moreover, the literature abundantly shows the use of solid-state reactions to synthesize manganite and composites. The autocombustion route for synthesizing manganites and composites has not, however, been widely reported. The sol–gel autocombustion method is based on the contribution of both sol–gel and combustion processes [
33]. The autocombustion process involves self-sustained redox exothermic reactions between hydrated metal nitrates and fuel(s) mixed on the molecular level [
34,
35,
36]. The main advantages of autocombustion are fast and straightforward processing (the formation results in an oxide mixture within a few seconds), with relatively low-cost precursors (usually nitrate salts), no need for special equipment, low energy consumption, atomic-level diffusion, and a resulting powder that is ultrafine and homogeneous [
37,
38,
39].
In the present study, a single calcium-doped Lanthanum manganate La
0.7Ca
0.3MnO
3 (LCMO), with the composition chosen as the perovskite phase. La
0.7Ca
0.3MnO
3 is reported to have high ∆
Smax, and a paramagnetic (PM)-ferromagnetic (FM) transition temperature around room temperature [
40]. We chose three antiferromagnetic (AFM) metal oxides, Mos, having Neel temperature range of 230 to 525 K (CoO:
TN~291 K, NiO:
TN~525 K, CuO:
TN~230 K) [
41,
42]. The magnetic inhomogeneity introduced by AFM phases was observed to be beneficial in tuning the phase transition temperature and RCP value of the composite [
28,
43]. The present work reports enhanced RCP value for LCMO-NiO/CoO/CuO composites.
4. Results and Discussion
The room temperature XRD scans for LCMO–MO composites with different compositions are shown in
Figure 2. The observed peaks of only LCMO and MO indicated the absence of any impurity formed in the composite. Since the metal oxide was only present in a small amount, as compared to the perovskite LCMO, the perovskite peak dominated the XRD spectrum, as shown in
Figure 2a,c,e. In
Figure 2b,d,f, long scans were performed to detect MO phases in the composites. MO peaks showed an increase in intensity with an increase in their Wt.% content in the composite samples. The XRD spectrum was refined using TOPAS Bruker, Billerica, MA, USA) software to determine the lattice parameters and cell volume of the LCMO–MO composites. given in
Table 2. The X-ray diffraction analysis also showed that the LCMO perovskite existed in the orthorhombic phase (space group: Pnma, ICDD #: 00-063-0088). The metal oxides NiO (ICDD #: 01-085-5454) and CoO (ICDD #: 01-071-1178) existed in a cubic phase (space group:
), whereas CuO (ICDD #: 01-086-8837) existed in the monoclinic phase (space group: C12/c1). The minimal change in the lattice volume of LCMO with MO addition in the composite was noted. However, the possibility of M ion inclusion in the perovskite structure could not be ignored.
Further, the crystallite size of the LCMO phase was calculated using Scherrer’s method [
44]. The diffraction peaks with Miller indices (
hkl) values (020), (200), (220), (040), (042), and (242), corresponding to peak positions (2θ) 23.1°, 32.8°, 40.4°, 47.1°, 58.5°, and 68.7°, were used to calculate the crystallite size of the LCMO phase. The obtained crystallite size for the LCMO phase is also presented in
Table 2. The LCMO lattice parameters and cell volume were similar to those reported by Shankar et al. [
45]. The crystallite size of LCMO decreased as the amount of metal oxides increased. The observed grain refinement of the LCMO phase in the composite might result from the following: (a) the increased micro-strain due to the size difference between LCMO and MO [
46,
47]; (b) MO diffusion to the boundaries, which could restrain the grain growth [
48]; (c) the reduction in the unit cell volume accompanied by shortening the diffusion path between nearby grains, which could, consequently, result in smaller grains during calcination. Similar grain refinement was observed in other ferrites and composites [
48,
49]. To check for unit cell distortion, the unit cell anisotropy, δ, defined as [(b/(a
2+c
2)
0.5) − 1], was calculated [
45]. The δ parameter is listed in
Table 2. The δ was zero for cubic cells. Thus, variation in δ parameter showed some evidence of deviation from the LCMO unit cell, which could result from the crystallite size reduction for these composites and/or M ion diffusion in the LCMO phase, which remained undetected in XRD. The irregular change in a lattice pattern with MO phase in
Table 2 could be due to different amounts of diffusion of metal ions in the main phase. This diffusion was dependent on reaction kinetics. Some metal ions dissociated more from MO than others, thus, causing the observed difference in lattice parameters.
5. Magnetic and Magnetocaloric Study
The temperature dependence of magnetization of the composites measured in a field-cooled (FC) process during warming-up in a constant external applied field of 1 kOe is shown in
Figure 3. The samples with low MO content (<5%) showed a sharp magnetic phase transition from a low-temperature ferromagnetic to a high-temperature paramagnetic phase at Curie temperature (
TC). It was noticed that the value of
TC decreased with an increasing amount of metal oxide in the composite. The curve became broader and appeared to show a second-order phase transition. The Curie temperature
TC was deduced from the minimum of
vs.
T curves, as shown in insets of
Figure 3. The peak of this curve represented the phase transition temperature, and the broadness of the peak suggested whether the transition was immediate or occurred over a wide range of temperatures. The obtained
TC values are listed in
Table 3. The transition temperature of 258 K for pure LCMO was in close agreement with the 267 K recorded in the literature [
50,
51]. The transition temperature dropped on all composites with increasing content of metal oxides. The drop in transition temperature was very similar for LCMO–NiO and LCMO–CuO samples, but there was a somewhat higher drop observed for LCMO–CoO samples. A significant drop in temperature from 258 K for pristine LCMO to 167 K was observed in the LCMO–5%CoO sample. The peaks also broadened as the content of metal oxide increased. This drop-in phase transition temperature was due to the weakening of the double exchange magnetic, Mn–O–Mn interaction, possibly due to the diffusion of M ions into the LCMO matrix. The 3
d ions brought in an unexpected variation of Mn valence states and oxygen vacancies [
25,
52,
53]. Further, the intensity of the rate of change of magnetization, with respect to temperature, decreased with an increase in metal oxide content; thus, the temperature range of the PM–FM transition became broader in the case of composite samples, which indicated the influence of increased magnetic inhomogeneity in the sample. Thus, in composites, the phase transition ranged over a more extensive temperature range than the pristine LCMO.
The value of magnetization at low temperature, 5 K, was observed to decrease with the increase in MO content for LCMO–NiO and LCMO–CoO, while LCMO–CuO showed an increase at 2.5% CuO content. The magnetic interaction in LCMO was mainly governed by Mn
3+–O
2−–Mn
4+ interactions and the diffusion of Cu
2+ into LCMO matrix substitutes for Mn
3+ and induced more Mn
4+ which made the Mn
3+–O
2−–Mn
4+ ferromagnetic chain increase, and, therefore, magnetization was enhanced [
52]. On the other hand, superfluous Cu
2+ resulting from an increase in CuO content in the composite occupied B sites between Mn
3+ and Mn
4+, and would induce disorder and change the bond angle and bond length of Mn–O–Mn interaction, and, thus, weaken the double-exchange interaction. Accordingly,
TC and
Ms decreased with the gradual increase in doping level [
54,
55]. A similar mechanism could be responsible for the observed decrease in magnetization at 5 K in the LCMO–MO composite.
Figure 4 shows the magnetic hysteresis of LCMO–MO composites at 5 K. The
M–H curves indicated that the magnetization of the samples increased rapidly at low fields and then tended to saturate at high fields, reflecting the ferromagnetic behavior of LCMO–MO composites. The inset in
Figure 4b shows the composites’ saturation magnetization (Ms) with their weight percentage of the MO phase. The decrease in the magnetization of the composite with increased Wt.% of the MO phase could be the combined effect of the following: (1) increased Wt.% of antiferromagnetic phase; (2) decrease in particle size, as reported by He et al. [
56]; (3) diffusion of metal ions into LCMO grains and grain boundaries, as explained above.
To evaluate the effect of MO content on the magnetic entropy change in LCMO–MO composite, magnetic isotherm curves were collected. Magnetic isothermal curves were collected by measuring the sample’s magnetization with varying external magnetic fields at various temperatures around the phase transition temperature. The measurements were taken at an interval of 5 K to obtain more accurate data.
Figure 5 represents the
M vs.
H plot at different temperature values for all the samples. From
Figure 5, it is evident that the samples underwent a phase transition from PM–FM state with lowering temperatures. As expected, the
M–H curves revealed a strong variation of magnetization around the
TC. The lower portion of the graph represented the higher temperature measurements which increased gradually with the field. At lower temperatures, the graph changed nature and acquired more of a ferromagnetic ordering with the magnetization saturation in the presence of the external magnetic field. A jump in magnetization values for the same external field with a temperature decrease was noticed. The highest jump of magnetization corresponded to the phase transition temperature. This was because magnetization showed a sharp increase at the transition from the paramagnetic to the ferromagnetic phases. This indicated that a small variation in magnetic field value could significantly increase the sample’s spin-ordering and magnetization.
Figure 6 represents the Arrott plots for all the samples. The negative slope corresponds to the first-order phase transition, whereas the positive slope corresponds to the second-order phase transition [
57]. We observed a first-order phase transition in LCMO, LCMO–2.5%NiO, and LCMO–5%CuO, whereas a second-order phase transition was observed in LCMO–5%NiO, LCMO–2.5%CuO, LCMO–2.5%CoO, and LCMO–5%CoO. The change of order of magnetic phase transition was attributed to the Mn–O–Mn interaction with the diffusion of M ions into the LCMO matrix and particle size reduction [
58], as explained earlier.
From
M vs.
H curves in
Figure 5 at different temperatures, the change in magnetic entropy of the sample was evaluated. The difference in magnetization was proportional to the change in entropy. The magnetic entropy change (Δ
SM) was one of the parameters to evaluate the material’s magnetocaloric property. The isothermal magnetic entropy change was computed using the thermodynamic Maxwell relation [
59]:
Numerically calculated as:
where
Hi and
Hf are the initial and final external applied fields and
µo is the permeability of free space. The
−Δ
SM was calculated from the isothermal magnetization curve of
Figure 5. This negative change in entropy with the temperature at different applied external field values (1–5 T) was plotted in
Figure 7. The change in entropy was negative because heat was released when the magnetic field changed adiabatically. In other words, for an adiabatic process, a cooling effect during this phenomenon was observed. It was observed that this change in entropy reached a maximum at the phase transition temperature and increased with increase in the external field applied.
The maximum change in magnetic entropy (
) and the FWHM of the plots were evaluated from the entropy change graph,
Figure 7, and presented in
Table 4,
Table 5 and
Table 6 for LCMO–NiO, LCMO–CuO, and LCMO–CoO samples, respectively. It was evident that the
increased with the increase in strength of the applied field. The magnetic entropy change was at a maximum at 5 T for all the samples, because the large applied magnetic field induced higher ferromagnetic ordering in the sample. First, the maximum magnetic entropy change obtained for pristine LCMO was 4.75 JKg
−1K
−1 at 1 T, significantly higher than the recorded value of 1.2 JKg
−1K
−1 at 1.5 T [
51]. A consistent trend of decrease in the maximum magnetic entropy and peak broadening with the increase in the metal oxide content was observed in all samples. The broadening of the peak was further verified by the increase in FWHM of the plot as the content of respective metal oxide increased. The most significant magnetic entropy drop from the pure LCMO was seen in LCMO–5%NiO. The change in magnetic entropy dropped from 8.77 JKg
−1K
−1 in pure LCMO to 3.24 JKg
−1K
−1 in LCMO–5%NiO at 5T. The plot’s most significant increase in FWHM was for the LCMO–5%CoO sample. The FWHM increased from 24.63 K in pure LCMO to 64.09 K in LCMO–5%CoO. This defined the operational temperature for the samples, which increased significantly for LCMO–5%CoO. This increase in FWHM was observed in smaller amounts in other samples but was still significant, improving the practical application. The decrease in
value was primarily attributed to the antiferromagnetic insulation of the LCMO by metal oxides [
25]. The
values were independent on temperature, except for LCMO–5%CuO, where
shifted to higher temperature with the increased field. The magnetic inhomogeneity played a significant role in both
reduction and increase in the FWHM values of the composite. The volume fraction of the LCMO decreased with increased metal oxide content. This also reduced the fraction of ferromagnetic spins compared to the pure sample [
60].
Other than this, unintentional minor doping of metal ions on the B-site of the LCMO perovskite replacing the Mn ions could also reduce
[
57,
58]. The published literature shows that the Mn ions in the Ca
2+ substituted LaMnO
3 exhibit double-exchange interaction. Mn ions exist in two different oxidation states, +3 and +4, due to the presence of La
3+ and Ca
2+ cations in the A-site [
61,
62,
63,
64,
65,
66]. The magnetic Mn
3+ and Mn
4+ cations bonded through non-magnetic oxygen anion induce the hopping of electrons between the cations through oxygen to attain the least energy state. This is called double-exchange interaction. This phenomenon promotes ferromagnetic ordering. The replacement of such transition metal ions with M cations could minimize this delocalization of electrons and decrease the ferromagnetic ordering. Eventually, yielding a lower change in magnetic entropy [
52].
The magnetic entropy change ∆
SM is proportional to the strength of the magnetic field. The ∆
SM value may be affected in the composite nanoparticles due to individual elements’ properties. The effect of minority phases and demagnetization fields on the ∆
SM value in the composite can be studied by fitting the numerical formula at a fixed temperature:
where C is a constant, and N is the exponent obtained from experimental fitting that determines the sensitivity of magnetic entropy change to the applied field. The higher value of N signifies higher sensitivity of the entropy change to the applied field. The exponent N was derived by fitting the experimental data at respective temperatures.
Figure 8 shows the change in exponent N at different temperatures. From
Figure 8, it is seen that the exponent N was at a minimum at
TC and exhibited a minor increase at temperatures below
TC and a significant increase at temperatures above
Tc, which was expected [
17]. Below
TC, the material was in the ferromagnetic phase, so the magnetic entropy did not change significantly due to the externally applied field. In contrast, above
TC, the paramagnetic phase of the material was highly influenced by the externally applied field, which substantially increased the N value.
One of the most important parameters to quantify the magnetocaloric property of a material is RCP. It quantifies the heat energy transferred in a complete ideal magnetic refrigeration cycle. This accounts for the maximum change in entropy and the operational temperature of the given material. as given mathematically by the equations:
where
is the maximum value of entropy changes at the given applied field (peak of ∆S
M vs.
T plot), and δT
FWHM is the FWHM of the ∆
SM vs.
T plot.
RCP was calculated using the data from
Table 4,
Table 5 and
Table 6 and listed in
Table 7 for all the samples.
Figure 9 displays the RCPs at different applied magnetic fields from 1–5 T. We observed that the RCP values increased with the applied magnetic field. This was because the increase in change in entropy with the increasing magnetic field improved ferromagnetic ordering, and RCP was linearly related to the change in entropy, as given by the formula. The RCP of all composite samples increased, compared with the pristine LCMO except for LCMO–5%NiO. The highest RCP value of 266.89 JKg
−1 was recorded for LCMO–5%CuO samples, which was 23.4% higher than the 213.13 JKg
−1 observed for the pure LCMO at the 5T applied field. This was a substantial gain, considering the decrease in ∆
Smax values, which was again attributed to the magnetic inhomogeneity introduced by the antiferromagnetic nature of the metal oxides [
45,
47]. The loss in magnetic entropy change was compensated for more than enough by broadening the peak, which increased the RCP. The decrease in RCP for LCMO–5%NiO was because of the small increment in the FWHM of the curve. This could be credited to the difference in phase transition temperature of the metal oxides with respect to LCMO [
49]. The phase transition temperature for NiO was 525 K which was farther away from 258 K for LCMO compared to 230 K for CuO and 291 K for CoO. Due to this high difference in magnetic phase transition temperature, the RCP for a higher content of NiO in LCMO dropped to 152.71 JKg
−1. From this observation, it emerged that the addition of MO as a secondary phase in LCMO samples indeed enhanced their cooling efficiency. The field dependence of RCP could also be studied using the relation:
where C is a constant and R is the exponent obtained from experimental fitting that determines the sensitivity of RCP to the externally applied field. The obtained R-value is shown in
Table 8. A small drop in R-value with composite formation for all samples, except La
0.7Ca
0.3MnO
3–2.5%NiO, was observed. This drop in R-value was reasonably close to 1 for all samples, which signified that RCP had an almost linear dependence on the externally applied field.