Freestanding and flexible composites of magnetocaloric Gd5(Si,Ge)4 microparticles embedded in thermoplastic poly(methyl methacrylate) matrix
Vivian M. Andrade
’Gleb Wataghin’ Physics Institute, State University of Campinas (IFGW-UNICAMP), Rua Sergio Buarque de Holanda, 777, 13083-859 Campinas - SP, Brazil
IFIMUP and IN-Institute of Nanoscience and Nanotechnology, Departamento de Física da Faculdade de Ciências, Universidade do Porto, Rua do Campo Alegre, 687, 4169-007 Porto, Portugal
Nathalie B. Barroca
Ana L. Pires
João H. Belo
João P. Araújo
André M. Pereira
IFIMUP and IN-Institute of Nanoscience and Nanotechnology, Departamento de Física da Faculdade de Ciências, Universidade do Porto, Rua do Campo Alegre, 687, 4169-007 Porto, Portugal
Abstract
The implementation of processed magnetic materials onto thermoplastics can be an approach for practical application of brittle intermetallic materials with the advantage of enlarging the range of applications. In the present work, we evaluate the effect of blending magnetocaloric Gd5Si2.4Ge1.6 micrometric particles with 3.4 μm in different weight fractions onto a flexible, transparent and non-magnetic poly(methyl methacrylate) (PMMA). A close to homogeneous grain distribution along the polymer surface were achiever by using a simple solvent casting method for evaluation of their magnetocaloric properties. From XRD analysis, it was found a relative unit cell volume reduction of ∼2.5×103 ppm for the composite with 70 wt.% of powder as a result of interfacial interactions between the components. Although PMMA does not influence the magnetic nature of microparticles main phase, a reduction on the amount of secondary monoclinic phase occurs for all produced composite samples. As a consequence, a weakening on the effect of secondary phases on the micropowder magnetocaloric response is observed as a result of hydrostatic pressure from the difference between thermal expansions of matrix and filler.
Magnetic composite, magnetocaloric effect, flexible composite, PMMA
I Introduction
Magnetocaloric materials for heating/cooling applications at room temperature have been a field of interest for several decades for substitution of hazardous gases used in traditional vapour compression refrigerators. The discovery of a Giant Magnetocaloric Effect (GMCE) on the Gd5Si2Ge2 compoundPecharsky and Gschneidner (1997) was a milestone due to the potential for magnetic refrigeration at room temperature. Since then, several strategies have been followed in order to evaluate/increase the MCE properties of the Gd5(Si,Ge)4 family compounds aiming new applications, such as micro-cooling devices and medical treatmentsBoekelheide et al. (2017); Kitanovski et al. (2015); El-Gendy et al. (2017). An intensive research on this family led to the observation of impressive effects that raise from their strong magnetostructural coupling, such as: large magnetoresistance and colossal magnetostrictionPecharsky and Gschneidner Jr (2001); Pires et al. (2014). The crystallographic arrangement on these compounds is composed by a stack of pseudo-cubes connected by Si/Ge dimmers positioned at the cube edges, the so-called interslab positions, that are extremely sensitive to internal and external parameters. The number of Si/Ge bonds will determine the crystallographic structure and, consequently, the magnetic ordering: orthorhombic-II [O(II)] Gd5Ge4-type does not present any bond with an associated antiferromagnetic state, monoclinic (M) Gd5Si2Ge2-type where half of this bond is formed and orthorhombic-I [O(I)] with all Si/Ge bonds with both presenting a ferromagnetic orderingPecharsky and Gschneidner (1997). For this reason, thermal variations, applied magnetic fields and hydrostatic pressure, for example, can lead to the rupture or formation of Si/Ge dimmers and, thus, changing the structureGschneidnerJr, Pecharsky, and Tsokol (2005). However, most of the published research is performed with bulk samples that are brittle and fragileHarstad et al. (2017); Ozaydin and Liang (2014), which represents a significant drawback in shape design for technological applications. Besides that, when implemented into devices, the material is submitted to several thermal cycles that can lead to its degradationBooth and Majetich (2012); Pires et al. (2015). An approach to improve the material mechanical and chemical stability can be achieved by incorporating the inorganic magnetocaloric materials (MCM) into organic polymers with suitable properties. Particularly, flexibility on magnetic composites allows implementation on sensors and transducersAkdogan, Allahverdi, and Safari (2005), dielectric wave-guideXiang, Wang, and Yao (2007), microwave devicesBora et al. (2019) and bioengineeringLaurencin et al. (1999).
Generally, there has been an increase on research activities in polymer-based composites on the last 20 years focused on implementation in several industries such as transport, military, aerospaceKoniuszewska and Kaczmar (2016), biomedicalRamakrishna et al. (2001) and textileGereke et al. (2013). There are few reports dedicated to the studies of Gd-based materials in polymeric matrices. For instance, Ozaydin et al. have shown that the incorporation of low amounts (< 10%) of Gd5Si2Ge2 micropowder in polyvinylidene fluoride (PVDF) by spin coating revealed an enhancement on converting magnetic energy into electrical powerOzaydin and Liang (2014). More recently, Bora and co-workers demonstrated the potential of applying Gd5Si4 milled powder blended with the elastomer polydimethylsiloxane (PDMS) on microwave absorption in the Ku-bandBora et al. (2019). Regarding the MCE studies, Imamura et. al. have evaluated the effect of compacting and sintering Gd5.09Si2.03Ge1.88 powder with low amounts (∼ 15%) on a camphorsulfonic acid doped polyaniline (PANI-CSA) conductive polymerImamura et al. (2017). The authors observed a ∼ 22% reduction on the refrigerant cooling power (RCP) values and the gain on mechanical properties had no influence the magnetic nature of the powder.
In this work, we have choose a more flexible poly(methyl methacrylate) thermoplastic to reinforce powders of Gd5Si2.4Ge1.6 powder with 3.4 μm. The evaluation on the structural, magnetic and magnetocaloric effect of blending different weight fractions of the micropowder into a flexible, transparent and resistant PMMA using a simple and reproducible chemical technique. The chosen stoichiometry displays a broader working range temperature for the MCE at room temperature and with absence of hysteresis lossesHadimani et al. (2015); Misra and Miller (2006). The PMMA is suitable for making flexible magnetocaloric composites due to its large durability, flexibility, high resistance to scratches and very low water absorbency (∼ 3%)Ali, Karim, and Buang (2015); Kim et al. (2010); Park and Jana (2003). Considered as the hardest thermoplastic, PMMA presents a thermal stability in a wide range of temperature going from 200 K to 500 K which is of great importance for room temperature applications.
II Experimental techniques
II.1 Bulk and powder synthesis
The magnetocaloric filler for the composite solution were prepared through the Tri-arc technique under Ar atmosphereBelo et al. (2012). For this, each constituent elements (with purities higher than 99.99%) were first accurately weighted in order to obtain the Gd5Si2.4Ge1.6 composition. Since the weight loss, after arc-melting, was less than 2%, the final product stoichiometry was assumed unchanged. No heat treatment was performed on the as-cast ingot that was manually grounded and sieved through several filters with hole sizes from 50 μm to 5 μm in order to obtain a thinner powder and to guarantee a homogeneous dispersion on the composite.
II.2 Composites synthesis
The polymer composite samples were prepared via solvent casting technique using Gd5Si2.4Ge1.6 (GSG) 3.4 μ m particles obtained as described aboveMarycz et al. (2016). The composite solutions were prepared by dissolving PMMA in dichloromethane (DCM) (Sigma, 270997-1L) at 40 0C until complete dissolution was achieve in order to acquire a 10wt.% PMMA blend. Afterwards, composite solutions with GSG weight fractions of 10, 30, 50 and 70% were obtained by simply dispersing the GSG microparticles in the PMMA solution. The resultant solutions were then solvent casted in order to obtain the freestanding and flexible magnetocaloric films. Different moulds shapes were used to illustrate the possibility of producing flexible magnetocaloric composite materials with variable designs. Two of the products obtained through this technique are depicted in Fig. 1(a) and (b) where is possible to notice that the composites surface becomes visually tougher and darker as the GSG weight fraction in PMMA increases.
II.3 Characterization techniques
Crystallographic characterization was performed via X-ray diffraction (XRD) data obtained at room temperature using a Rigaku SmartLab diffractometer with Cu-Kα radiation (1.540593 Å), 45 kV and 200 mA at IFIMUP. The diffraction patterns were collected from 20o≤2θ≤60o range in a Bragg Brentano geometry, with 0.02 o steps. Analysis were performed through Rietveld refinement considering a Pseudo-Voigt function of the XRD patterns using the FullProf software Rodríguez-Carvajal (1993). Morphological properties of the composites were evaluated by Scanning Electron Microscopy (SEM) using a Philips-FEI/Quanta 400. Cross-section imaging were performed on freeze fractured samples using Nitrogen. Superconducting Quantum Interference Device (SQUID) was used for the magnetic characterization and evaluation of the MCE on [5,350] K temperature range.
III Experimental results
III.1 Crystallographic and morphological characterization
The XRD patterns for bulk, powder and composites samples are presented in Fig. 1(c). At a first glance, one highlight is the broadening of the diffracted peaks on the used powder in comparison the bulk counterpart. This is a direct consequence on the reduction of particle size after sievingAndrade et al. (2016a); Pires et al. (2015); Hunagund et al. (2018). From Rietveld refinements it was found the formation of expected O(I) structure, M-phase and eutectic R5M3-phase (that will be simply denoted as 5:3)Belo et al. (2012), with the results summarized for all samples on Table 1. Misra et al. have shown that at the Si-rich region of Gd5(Si,Ge)4 family compounds, Ge preferentially occupies the interslab positions at the O(I) structure and these values were considered for the calculationsMisra and Miller (2006). As for the M-phase, the Si/Ge sites were initially considered to be fully occupied and, after the refinement, it was found the same preferential occupation of Ge atoms positioned at the shortest distance between the dimers, with the atomic positions also displayed in SID Table S3. As highlighted by vertical lines in Fig. 1(c), the corresponding peaks intensities for M phase increase for the micropowder, indicating an increment on this phase amount with decreasing particle size. Such observation is confirmed by Rietveld refinement results which give the presence of more than 20% of M-phase. Similarly, the amount of detected eutectic phase was found to increase for Gd5Si4 compound when the particle size reduces from 700 to 80 nm through ball millingHunagund et al. (2018). Previously, it was shown an enlargement of secondary R5M4 phases, such behavior was observed for milled Gd5Si1.3Ge2.7 and Tb5Si2Ge2 followed by a reduction on the unit cell volume of the main phasePires et al. (2015). In our case, by sieving the GSG powders, a slight increase on the unit cell volume of the main O(I) phase from 865 Å3 for the bulk to 869 Å3 for the 3.4 μm powder was noted, while M and 5:3 unit cells remain unchanged, as summarized on Table 1. With the selection of smaller particles by the strainers, the defects at the grain boundaries become more evident and can be leading to the detection of larger amounts of the distorted monoclinic phaseToraya and Tsusaka (1995). In another words, there is a self-segregation of the crystallographic structures on the Gd5Si2.4Ge1.6 compound since there is no thermal effects involved on the chosen technique for grain separationPecharsky and Gschneidner (1997); Belo et al. (2012).
Regarding the composite samples patterns, also shown in Fig. 1(c), a shift on the peak positions towards higher 2θ angles in comparison with the micrometer powder pattern indicates a reduction on the unit cell volume. Besides that, there is a reduction on the intensity of diffracted peaks from secondary M-phase, suggesting that the presence of PMMA is affecting the amount of secondary phases. In order to confirm these, Rietveld calculations were performed by considering the resultant lattice parameters and atomic positions from the free powder diffractogram refinement as initial values. Since PMMA is amorphous, there is a large background contribution to the pattern leading to a decrease on Rietveld refinement quality; however, the results summarized on Table 1 are in fine agreement with previous reportsPecharsky and Gschneidner (1997); Misra and Miller (2006). The introduction of magnetocaloric powder on the polymeric matrix lead to a significant enhancement on the phase fraction of O(I)-phase from ∼76.2% for the free powder to ∼88.9% in average for the composite samples. Concomitantly, the detection of M-phase reduces from ∼22.0% (free powder) to ∼10.0% (composite). This observation for the reinforced system can be due to an applied pressure on the surface by PMMA. This effect can be attributed to PMMA solidification around the GSG grains that lead to a contraction at the powder boundaries. This is reflected on the unit cell volume reduction ΔV/V0 of these phases, where V0 correspond to the volume of the free powder, with the increase of GSG content, shown in Fig. 2(a). The normalized unit cell volume was obtained considering the O(I), M and 5:3 phase fractions from Table 1, that follows the same behavior as the individual phases. This is mainly due to the extreme sensitivity of M-phase to external parameters, where the PMMA surface in contact with the grains is working as a isotropic pressure cellPecharsky and Gschneidner Jr (2001); Carvalho et al. (2005). Mudryk et. al. have shown that an applied pressure of 2 GPa is required to induce a complete transition from the M to O(I) structure of Gd5Si2Ge2 polycrystalMudryk et al. (2005). The authors have obtained the isothermal compressibility (κT) of 3 TPa*-1* and 6 TPa*-1* for M and O(I) phases, respectively. Considering these values and using the experimental relative reduction on the unit cell volume, it is possible to estimate the isobaric pressure on each structural phase of the grains through the thermodynamic relation: κT=−(1/V)(dV/dP)T. The calculated values, presented in Fig. 2(b), show that the thermoplastic walls force along the M-phase at grain boundaries are estimated to be close to the critical value of 2 GPa for a polycrystalline sample. This might be the reason for an incomplete conversion of M into O(I) phase by the matrix. As a matter of fact, since the present composition have higher Si/Ge ratio than Gd5Si2Ge2 stoichiometry, these values can be overestimated and in situ measurements should be performed for accuracyMudryk et al. (2005).
SEM images were carried out in order to investigate the efficiency on separating microparticles through sieving the GSG bulk and to verify the particles distribution along the polymeric matrix volume. From the micrograph in Fig. 3(a) it was found that GSG thinner powder displays a log-normal particle size distribution averaging 3.4 μm, as shown on the inset. The presence of bigger grains is due to the fact that they do not present a uniform geometry. Due to gravity, the particles will agglomerate at the bottom, see Fig. 3(b), which led into stress on the PMMA surface and an increase on the curvature. The particle segregation at the bottom is the responsible for a non-regular surface of the films, as can be seen for 50 wt.% in Fig. 3(b). This evidence is another advantage for application in cooling systems since the thermal contact of the device can be selected at one side and being isolated by the polymer layer at the otherKitanovski and Egolf (2010). As can be noted also for 50wt.% composite in Fig. 3(b), the particle distribution along the film surface is quasi-homogeneous. It is known that particles with mean diameters ranging from 2-5 μm can not be perfectly dispersed onto polymeric matrices even with sonification, stirring and other conventional techniquesGuan, Dong, and Ou (2008).
Furthermore, Fig. 3(c)-(f) shows the cross-section of freeze-fractured composites, where it is possible to notice that the polymer structure grows around the particles, revealing the bonding between particle and PMMA interface. The amplifications on these images does not allow to infer the porosity level of the samples; however, it is possible to notice a few gaps around the grains that suggest higher porosity. These observations justify the intergrain pressure interpretation shown above. As already mentioned, the extreme sensitivity of powder M-phase on the grains boundaries lead to a reduction on the amount of this phase and also its unit cell volumePecharsky and Gschneidner Jr (2001). For comparison purposes, a pure PMMA sample was prepared by following the same procedure and a film with ∼ 15 μm of thickness was obtained. With the addition of 10 wt.% of GSG microparticles remarkably increases the thickness of the composite to ∼ 36 μm and reach the maximum of ∼ 42 μm for 70wt% of GSG. This initial characterization will be of great matter for the magnetic and magnetocaloric properties evaluation.
III.2 Magnetic characterization
The normalized M-T curves for the Gd5Si2.4Ge1.6 in the form of bulk, powder and flexible composite obtained at cooling and heating between 5 and 350 K with an applied magnetic field of 0.1 T are shown in Fig. 4(a). The bulk presents a second order ferromagnetic to paramagnetic transition around 308 K are in agreement with previous reportsPecharsky and Gschneidner (1997); Misra and Miller (2006), as obtained through the dM/dT curves presented on the inset of Fig. 4(a). Notice that Gd5(Si,Ge)4 compositions that crystallize in a M structure will present a first order magnetic transition (FOMT), simultaneously changing to an O(I) structure at lower temperaturesPecharsky and Gschneidner (1997). As can be noted in the magnetization profile curves, with the selection of particles with 3.4 μm mean size, there is a thermal hysteresis ranging from 200 to 300 K that is a consequence on the increase of M-phase content. Such observation is confirmed through the appearance of a bump highlighted at the dM/dT curvesBelo et al. (2012). It is worth to point out that the reduction of particle size seems to be affecting only the secondary M-phase since there is no shift on TC for the main O(I) phase, as usually observed in magnetic materials at the micro/nanoscalePires et al. (2015); Andrade et al. (2016b); Checca et al. (2017). Such effect can be rising from the higher sensitivity of the M structure over O(I) to applied magnetic field, hydrostatic pressure and particle sizeCarvalho et al. (2005); Miller (2006). When the micropowder is immersed into the PMMA matrix, there is a slight lift on the magnetization curves at the thermal hysteresis temperature range that lead to a reduction on the bump at the derivative curves on the inset of Fig. 4(a). As for the O(I) phase temperature transition, there is no shift on TC due to the presence of non-magnetic PMMA matrix.
The strong spin-lattice coupling on the Gd5(Si,Ge)4 family compounds allows the use of its magnetic results to infer the crystallographic phases fractionsBelo et al. (2012). Above the ferro-paramagnetic temperature transition, the magnetic susceptibility is described by the Curie-Weiss law where we should consider the amount of each phase x on the sample as followsAndrade et al. :
[TABLE]
where C=μeff2/3kb represents the Curie constant and ΘP, the Curie paramagnetic temperature. In particular, for 10 wt.% composite, a correction on the magnetization curves due to the diamagnetic contribution from the matrix was performed, also performed in similar compositesTawansi et al. (2002). The best curves are shown in Fig. 5(a) for all samples with the returned values summarized on Table 2. For a better visualization, the amount of each phase is given in Fig. 5(b) obtained from χ−1 method and XRD results for comparison purposes. As can be noted, there is no significant change on the values obtained through the different methods, since they are within the error. Most notably, the magnetic analysis seems to detect a slightly higher content of M-phase than XRD calculations which might be related to the higher sensitivity of χ−1 method. These findings confirm the assumption that PMMA polymeric matrix works as a pressure cell on the grains surface and, thus, weakens the effects of secondary M-phase. Indeed, further evaluation on the MCE results will also reveal the strong influence of the non-magnetic thermoplastic on the magnetocaloric powder. Furthermore, the μeff value for this phase is below the expected 7.94 μB which has direct consequences on the MCE features of GSG magnetic materialCarvalho et al. (2005). Furthermore, θP and μeff are in fine agreement with the obtained for pure powder and the theoretically expectedBelo et al. (2012); Roger et al. (2006). Although there is a reduction on μeff values, they are still close to theoretically expected 7.94 μB for Gd3+ ionsBelo et al. (2012). Therewith, the presence of PMMA does not affect the intrinsic magnetic features of the 3.4 μm powder. In another words, the polymeric matrix is acting as an external agent on the grains in form of hydrostatic pressure that have been shown to have great influence on the M-phaseMiller (2006); Pecharsky, Gschneidner Jr, and Pecharsky (2003); Mudryk et al. (2005).
Nevertheless, magnetization data at 5 K with an applied magnetic field up to 5 T for the bulk, powder and composite samples were acquired for rating their M(5K,5T) value. Using this data it was also possible to extract their saturation magnetization (μsat) values through an extrapolation on the M versus 1/H, with the values are summarized in Table 2. As can be noted, there is a reduction on M(5K,5T) and μsat values from the bulk to the powder sample that is related to dimensionality reduction and from the 5:3-phase AFM ordering - the TN of this phase is report to be 75 KRoger et al. (2006); Belo et al. (2012); Hunagund et al. (2018). As for the composite samples, taking into account the contribution from all the system components, i.e., mGSG+mPMMA, we obtain the M(5K,5T) quantity that reduces as the amount of magnetic material decreases. This is a direct consequence on the dilution of magnetic powder along the non-magnetic PMMA matrix, as observed in previous magnetocaloric compositesZhang et al. (2015). The μsat values, however, were calculated by considering only the weight fraction of active magnetic material and the contribution from each crystallographic phase obtained through χ−1 method. In a similar way, there is a loss on the μsat values for lower GSG content composites that can be due to dilution effects or from possible evaporation losses during synthesisRamprasad et al. (2004). For this reason, the obtained values are within the error which corroborates with previous assumptions that the presence of a non magnetic polymer does not affect the intrinsic properties of magnetocaloric filler. Furthermore, the saturation is reached at lower intensities of magnetic field for the composite samples with lower filler contents and it starts to behave as the free powder with increasing the amount of magnetic material. Such behavior can raise from interparticle short range interactions with associated low level of deformation, as observed in soft magnetic elastomers, that can also contribute to this fast magnetization on the composite samplesStolbov, Raikher, and Balasoiu (2011); Stepanov et al. (2007).
IV Magnetocaloric Effect
The MCE is a well known phenomenon where the thermodynamic properties of a ferromagnetic material can be reversible varied through the application of a magnetic field. The thermal variations can be observed in two basic processes: i) adiabatic, that leads into changes in temperature (ΔT), and ii) isothermal, that is quantified by an entropy change (ΔS). For the present work, the MCE evaluation were performed through magnetization isothermal measurements obtained from 215 K to 350 K by increasing and decreasing the applied magnetic field up to 5 T, given on SID Fig. S3 and S4. It is worth to point out that, due to the magnetic irreversibility, the samples were warmed up to 330 K between each isothermal measurement. Therefore, through the magnetization map M(T,H), the magnetocaloric potential ΔS was calculated by using the integrated Maxwell relation ΔS(H,ΔT)=0∫H(∂T∂M(T,H))dHDe Oliveira and Von Ranke (2010). For the calculations, first it was considered only the contribution from the filler, i.e., the weight of magnetic active material, denoted as ΔSW. As can be noted, there is a reduction on the saturation magnetization as filler content decreases with the diamagnetic contribution from PMMA being more evident for 10 wt.% sample, as mentioned above. The obtained ΔSW(T) curves for powder and composite samples are depicted in Fig. 6(a) for Δμ0H=5T. The ΔSW curves for the composites follow a λ-shape, typical of SOMT, and its maximum values raise as the applied magnetic field increasesBelo et al. (2012). From these curves, the important MCE parameters were extracted as can be seen on Table 3.
Concerning application goals, the volumetric entropy change ΔSV is rather informative for device engineering than the mass oneZhang et al. (2015). The magnetic material density (ρ) was considered to be 7.45 g/cm3, in agreement with reported by Gschneidner and PecharskyGschneidner et al. (2000). As for the composite samples, the density was obtained through a carefully measurement on the film area and considering the thickness obtained by SEM cross-section imaging. It is worth to point out that the acquired density for the systems with 10 and 30 wt.% of magnetic material are below the expected and might be the reason for loss on ΔSMmax, shown in Fig. 6(b). As previously mentioned, this can also be a consequence of a non perfect homogeneous distribution along the film volume, that cannot be assured during castingJesson and Watts (2012), and from possible evaporation losses during synthesisRamprasad et al. (2004). A close composition Gd5Si2.5Ge1.5 have been reported to present a maximum entropy of 70.7 mJ/cm3K at 313 KGschneidner Jr and Pecharsky (2000). Taking into account the formation of secondary M-phase and eutectic 5:3 phase, the obtained value of ∼ 50 mJ/cm3K for 42 μm are close to the expectedGschneidnerJr, Pecharsky, and Tsokol (2005). For the smaller grains, it drastically decreases to ∼ 25 mJ/cm3K with the same profile as shown in Fig. 6(a). Although the effect of M-phase is blocked by the polymeric matrix, the presence of non-magnetic material diminish the system magnetocaloric response. However, the values of 5-10 mJ/cm3K are in the range for applications in micro-cooling devices for pump systemsKitanovski and Egolf (2010).
Regarding the ΔSW and ΔSV curves features, the bump from 220 to 300 K is not observed for the GSG powder when is blended with PMMA thermoplastic. As already illustrated, the particles are confined on the polymeric matrix that can be acting as a pressure cell on the grain boundaries. It is known that Gd5(Si,Ge)4 is highly sensitive to internal and external stimuliPecharsky and Gschneidner Jr (2001). The thermal expansion of the filler, monoclinic Gd5Si2Ge2 stoichiometry presents an anomalous behavior on it volumetric thermal expansion (γ) that reaches a maximum ∼1.46×10*-2* ppm/K for a single-crystalHan et al. (2002), being the mechanism responsible for the GMCE on this material. Carvalho et. al. have already shown that by applying a hydrostatic pressure of 0.1 GPa reduces the ΔSmax of Gd5Si2Ge2 compound in ∼23% and eventually vanishes after 0.6 GPa due to a suppression of the FOMTCarvalho et al. (2005). Given this, we can assume that during heating, the thermal expansion of the M-phase changing to an O(I) by the sliding of the pseudo-blocks on the crystal structure is limited by PMMA surface around the grains - since the polymer present a lower thermal expansionPorter and Blum (2000). Besides that, it would be interesting to perform studies on temperature cycling of the composite samples to observe if the polymer deformation is reversible, which could allow the structural change on the magnetic material. Since there is particles with a broad particle size distribution and PMMA thermal expansion is anisotropic, there is no simple solution to describe the mechanism behind this effect to estimate the applied pressure from the matrix to the grains edges. For this reason, we can only assume that is above the 0.6 GPa observed on M Gd5Si2Ge2 single crystals. Nonetheless, these findings reveal the interplay between mechanical and magnetocaloric properties that can be used to tune the best material features to produce multifunctional devices.
V Conclusions
The results here presented shows that solvent casting is a suitable technique for the implementation of 10, 30, 50 and 70 weight fraction of 3.4μm Gd5Si2.4Ge1.6 particles in non-magnetic PMMA. The micropowder was obtained through sieving and the reduction of particle size have intensified the effect of deformity on the grain boundaries, leading to a detection of ∼23% of deformed M-phase. Although there are no changes on the magnetic nature of the microparticles when blended with the thermoplastic, a slight reduction on the amount of the secondary M-phase is observed. The saturation magnetization at 5 K reveal that the composites magnetic response are ruled by the particle density, being bigger for larger amounts of magnetic filler. This is reflected on the MCE results where the interface GSG/PMMA interaction seems to weaken the contribution from M-phase on the ΔS curves with the absence of the bump observed for the free powder. Further investigations on thermal and mechanical properties are required for a fully understand on the mechanism of GSG/PMMA composites during thermal cycles. Furthermore, these observations indicates that the system here present can allow several applications such as energy harvesting, microfluidic system and magnetic refrigerationHamann and Dahlberg (2017); Harstad et al. (2017); Zhang et al. (2015); Crossley, Mathur, and Moya (2015). In addition, the findings presented in this work open new avenues on the next generation of the magnetocaloric effectKitanovski and Egolf (2010).
acknowledgement
This work is funded by FEDER funds through the COMPETE 2020 Programme and National Funds throught FCT - Portuguese Foundation for Science and Technology under the project UID/NAN/50024/2013 and by NECL with the project NORTE-01-0145-FEDER-022096. This work was also supported by the European Union’s Horizon -2020 research and innovation program under the Marie Sklodowska-Curie Grant Agreement No. 734801. VMA thanks the CNPq for the Grant No. 203180/2014-3. J.H. Belo would like to thank CICECO-Aveiro Institute of Materials, POCI-01-0145-FEDER-007679 (FCT Ref. UID /CTM /50011/2013), financed by national funds through the FCT/MEC and when appropriate co-financed by FEDER under the PT2020 Partnership Agreement.
