SCIENCE AND TECHNOLOGY OF ADVANCED MATERIALS 2020, VOL. 21, NO. 1, 424–434 https://doi.org/10.1080/14686996.2020.1780097 A comparative study of the influence of the deposition technique (electrodeposition versus sputtering) on the properties of nanostructured Fe70Pd30 films Matteo Cialone a,b, Monica Fernandez-Barcia c, Federica Celegato b, Marco Coisson b, Gabriele Barrera b, Margitta Uhlemann c, Annett Gebert c, Jordi Sort d,e, Eva Pellicer d, Paola Rizzi a and Paola Tiberto b a Chemistry Department and NIS, University of Torino, Torino, Italy; b Metrology of Innovative Materials and Life Science, INRiM, Torino, Italy; c Institute for Complex Materials, IFW Dresden, Dresden, Germany; d Department of Physics, Autonomous University of Barcelona, Cerdanyola Del Vallès, Spain; e Catalan Institution for Research and Advanced Studies, Barcelona, Spain ABSTRACT ARTICLE HISTORY Sputtering and electrodeposition are among the most widespread techniques for metallic thin Received 11 July 2019 film deposition. Since these techniques operate under different principles, the resulting films Revised 2 June 2020 typically show different microstructures even when the chemical composition is kept fixed. In Accepted 5 June 2020 this work, films of Fe70Pd30 were produced in a thickness range between 30 and 600 nm, using KEYWORDS both electrodeposition and sputtering. The electrodeposited films were deposited under FePd alloy; potentiostatic regime from an ammonia sulfosalicylic acid-based aqueous solution. electrodeposition; Meanwhile, the sputtered films were deposited from a composite target in radio frequency sputtering; thin films; regime. Both approaches were proven to yield high quality and homogenous films. However, magnetic properties; their crystallographic structure was different. Although all films were polycrystalline and Fe and mechanical properties; stripe Pd formed a solid solution with a body-centered cubic structure, a palladium hydride phase domains; perpendicular was additionally detected in the electrodeposited films. The occurrence of this phase induced magnetic anisotropy internal stress in the films, thereby influencing their magnetic properties. In particular, the CLASSIFICATION thickest electrodeposited Fe70Pd30 films showed out-of-plane magnetic anisotropy, whereas 105 Low-Dimension (1D/2D) the magnetization easy axis lied in the film plane for all the sputtered films. The domain pattern materials; 106 Metallic of the electrodeposited films was investigated by magnetic force microscopy. Finally, nanoin­ materials; 203 Magnetics / dentation studies highlighted the high quality of both the sputtered and electrodeposited Spintronics / films, the former exhibiting higher reduced Young’s modulus and Berkovich hardness values. Superconductors; 301 Chemical syntheses / processing; 303 Mechanical / Physical processing; 306 Thin film / Coatings; 503 TEM; STEM; SEM; 504 X-ray / Neutron diffraction and scattering 1. Introduction recording media [1–3], as microactuators and micro­ Iron-palladium alloys arise great attention from the sensors [4,5] or in spintronics [6]. In particular, the technological viewpoint as they show a unique and Fe70Pd30 (at.%) alloy shows the so-called ferromag­ very interesting combination of mechanical and mag­ netic shape memory effect, related to the occurrence netic properties. As a result, they find uses in magnetic of a martensitic-to-austenitic phase transition [7,8]. CONTACT Matteo Cialone matteo.cialone@unito.it Dipartimento di chimica, Università degli studi di Torino, Via Pietro Giuria, 7, 10125Torino (TO), Italy © 2020 The Author(s). Published by National Institute for Materials Science in partnership with Taylor & Francis Group. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Sci. Technol. Adv. Mater. 21 (2020) 425 M. CIALONE et al. This property is particularly appealing because, as available in the literature, a comparative study on the compared to standard non-magnetic shape memory impact of the deposition technique (in particular, elec­ alloys, it allows for wireless magnetic manipulation. trodeposition versus sputtering) is missing. In this The possibility to induce the re-orientation of the twin work, this is addressed for Fe70Pd30 composition boundaries in the martensite phase upon applying while trying to keep the rest of parameters involved a magnetic field [9,10] makes it an excellent candidate in the processes (e.g. substrate type, film thickness) the for wirelessly actuated micro- and nanoelectromecha­ same or as much comparable as possible. nical systems (MEMS/NEMS) or strain sensors [11]. Current trends towards miniaturization of materials and devices require the use of appropriate deposition 2. Experimental techniques. To date, Fe70Pd30 alloy films have been 2.1. Sputtering produced by both physical and (electro)chemical meth­ ods. Concerning physical methods, molecular beam Sputtered films were produced via radio frequency epitaxy [12], sputtering [13,14] and pulsed laser deposi­ (RF) sputtering from a composite target, where tiles tion [15], are among the most common approaches. of elemental palladium (purity 99.98%) are placed on Their major drawback is the need of ultra-high vacuum top of an elemental target of iron (purity 99.99%). This and high temperature for deposition. In comparison, configuration allows to change the exposed area of the electrodeposition represents a fast and cost-effective two elements, which in turn allows to tune the stoi­ alternative for the production of Fe70Pd30 thin films chiometry of the deposited film. Fe70Pd30 (at.%) films [16]. Moreover, unlike physical vapour deposition were produced in the thickness range between 46 and (PVD) methods, high aspect ratio structures are avail­ 500 nm. The base pressure of the sputtering chamber able by electrodeposition. was Pbase = (2.0 • 10−7) mbar, while during deposition For a given material composition, comparative stu­ the argon pressure was kept at Pdep = (1.0 • 10−2) mbar. dies on the morphological and structural characteris­ The substrate was a (100) oriented silicon single crys­ tics of films derived from physical and (electro) tal (500 μm thick) covered with 400 nm amorphous chemical deposition methods are rather scarce in lit­ SiO2. Before film deposition, Si/SiO2 substrates were erature [17–19]. Although the pros and cons of the cleaned with acetone, isopropanol and finally de- different deposition methods are well known by the ionized water in an ultrasonic bath. scientific community, specific considerations may apply to the particular system under investigation, 2.2. Electrodeposition and such features cannot be straightforwardly general­ ized. This is particularly true when it comes to electro­ Since conductive substrates are required for electro­ deposition. Although electrodeposition is commonly deposition, a 100 nm thick layer of gold (Au) was referred to as a cost-effective, faster technique which deposited on top of the Si(100)/SiO2 substrate. In can coat larger substrate sizes compared to physical order to ensure good adhesion, a thin layer of chro­ and chemical vapour deposition techniques, many mium (2 nm thick) was deposited between the SiO2 film properties (roughness, crystal structure, type of and the Au layer. Prior to electrodeposition, the defects, etc.) largely depend on the specific parameters substrates were cleaned consecutively with acetone, applied to grow the material. On the other hand, thin isopropanol and de-ionized water in an ultrasonic metallic layers deposited on smooth substrates by co- bath. As a reference electrode, a saturated calomel sputtering (thickness generally in the submicron electrode (SCE) was used (VSEC = 224 mV vs. stan­ range) are characterised by an almost fully dense and dard hydrogen electrode (SHE) at T = 25ºC), while uniform structure. Since microstructure and physical a thin foil of platinum was used as a counter elec­ properties (e.g., magnetic and mechanical) are tightly trode. The working and counter electrodes were correlated, alloy films produced from different deposi­ mounted into a cylindrical electrochemical cell in tion techniques might show very different properties. a horizontal arrangement, as previously reported by The choice of the appropriate deposition technique Iselt et al. [23]. Due to the particular setup of the is thus of utmost importance at the material design electrochemical cell, a portion of the Au surface was stage. Optimal selection requires a profound knowl­ used for electrical connection and remained edge of the influence of the deposition techniques on uncoated after electrodeposition. Depositions were the material characteristics in order to guarantee or be performed potentiostatically at room temperature closer to the target properties. Although studies on the and without stirring the electrolyte in a HEKA influence of the substrate type and the deposition potentiostat galvanostat PG 310. The electrolyte conditions on the properties of Fe-Pd films, and of composition, taken from a previous work of Fe70Pd30 alloy films, in particular, obtained by either Konczak et al. [16], is 0.01 M Pd(NH3)4 Cl2, 0.06 M electrodeposition [16,20,21] or sputtering [12,22] are sulfosalicylic acid (SSA), 0.05 M Fe2 (SO4)3 • 7 H2O, Sci. Technol. Adv. Mater. 21 (2020) 426 M. CIALONE et al. 0.3 M (NH4)2 SO4 (pH = 5). Considering the deposi­ has a roughness of Ra = (0.2 ± 0.1) nm, while the tion rates for electrodeposition and sputtering thickest film (500 nm) has an Ra = (4.4 ± 0.2) nm. processes, 1.1 nm/s and 0.1 nm/s, respectively, The evolution of the roughness for both the sputtered deposition times were adjusted on-demand to yield and electrodeposited films as a function of the film’s deposits of comparable thicknesses. thickness is summarized in the graph of Figure 2. The GIXRD curves of 100 nm, 300 nm and 500 nm thick sputtered films reported in Figure 3 (a) provide 2.3. Characterizations information on their crystal structure. Although the Morphology and stoichiometry of the films were stu­ phase diagram predicts the occurrence of two phases, died using atomic force microscope (AFM), scanning namely α-Fe + FePd, for the Fe70Pd30 composition electron microscope (SEM, FEI Inspect F) and trans­ [26], the GIXRD curves only show a set of reflections mission electron microscope (TEM, Jeol JEM-3010). shifted to lower angles with respect to pure α-Fe. This The latter was equipped with an energy dispersive might be an indication that the sputtered films are X-ray spectrometer (EDS). The crystallographic struc­ made of a single supersaturated solid solution of α- ture of the films was investigated by grazing incidence (Fe,Pd) [27]. Moreover, the (110) peak in the curves of X-ray diffraction (GIXRD) on a Panalytical X’Pert Figure 3 (a) is broader (larger full width at half max­ PRO MPD using the Cu Kα radiation, at a grazing imum) which can be related to a non-uniform distri­ angle of 0.4º. The magnetic properties were investi­ bution of Pd into the supersaturated solid solution gated using avibrating sample magnetometer (VSM) [27]. Considering the relative intensity of the diffrac­ from Lakeshore, at room temperature, up to tion peaks, no preferential orientation is observed. a maximum field of 20 kOe. Magnetic force micro­ Note that unlabeled peaks belong to the substrate. In scopy (MFM) was used to image the magnetic domain the inset of Figure 3 (b) is sketched the geometry of the patterns. A Bruker Multimode V Nanoscope 8 micro­ GIXRD configuration used to acquire the diffractions scope equipped with a fully non-magnetic head and plots for the different films. This particular geometry, scanner, and a commercial Bruker MESP-HR10 can­ where the incoming X-ray beam is kept at a fixed small tilever coated with Co/Cr hard magnetic alloy were incidence angle (α) while the detector is rotating utilized. The mechanical properties of the films were around the sample (with an angle θ, with θ ≠ α), measured by nanoindentation using a pyramidal- implies that the scattering vector Q has a different shaped Berkovich-type diamond tip [24]. Indentation direction for each different position (e.g. for each experiments were performed in raster across the sam­ different θ) of the detector and it is not always per­ ple surface and the values for the reduced Young’s pendicular to the sample surface. Hence, the detected modulus, Er, and the Berkovich hardness, HB, were vector K has components along x, y and z-direction. In determined as the average from ≈ 300 indentations for this case, the Scherrer’s equation can be used to esti­ each film using the method of Oliver and Pharr [25]. mate the overall dimensions of the crystalline grains. A complete list of the samples synthesized both via In particular, the Scherrer’s formula was applied on electrodeposition and via sputtering and analyzed in the (110) peak’s width to have a rough estimation of this work is reported in Table 1. the crystal size, < D >, of the α-(Fe,Pd) phase (Table 2). < D > increased from 5 nm to 15 nm with film thick­ ness, proving the nanocrystalline character of the 3. Results and discussions sputtered films. Worth it to notice that the value of the < D >, obtained with the Scherrer’s formula, are 3.1. Morphology and structural properties generally underestimated. Indeed, in thin films the 3.1.1. Sputtered films estimated values of crystallite size should be taken Sputtered films with thicknesses ranging from 46 nm with caution. to 500 nm were produced. The on-top SEM and AFM images of the two extreme thicknesses, namely 46 nm 3.1.2. Electrodeposited films and 500 nm, are shown in Figure 1. From the AFM For comparison purposes, electrodeposited films with images, the thickness-dependent surface morphology thicknesses ranging from 30 nm to 600 nm were of the films is clear. Namely, the thinnest (46 nm) film produced. Importantly, complete coverage of the sub­ strate was attained even in the 30 nm thick coating. Table 1. Summary of the samples produced and their Notice that complete coverage of the substrate is often thicknesses. not possible by electrodeposition for nanoscale thick­ Electrodeposited films Sputtered films nesses depending on the growth mode exhibited by 30 nm 46 nm the metal or alloy to be plated. As for the sputtered 100 nm 100 nm 305 nm 300 nm films, the on-top SEM and AFM images of the two 340 nm – extreme thicknesses, 30 and 600 nm, are depicted in 600 nm 500 nm Figure 4. The roughness also increases with film Sci. Technol. Adv. Mater. 21 (2020) 427 M. CIALONE et al. Figure 1. (a) SEM micrograph and (b) and (c) AFM images of the surface of the sputtered film with 46 nm thickness. (d) SEM micrograph and (e) and (f) AFM images of the surface of the sputtered film with 500 nm thickness. Interestingly, no cracks or inhomogeneities are appar­ ent from the images. Note that voids are often encoun­ tered in metallic electrodeposits due to hydrogen co- evolution. Likewise, cracks might develop due to layer dehydration in the SEM vacuum chamber. None of these were observed in our electrodeposited films, confirming that, under optimized conditions, electro­ deposited films can compete with those produced by PVD techniques. In order to investigate the cross-section of the electro­ deposited films, lamellas of the 30 nm and 600 nm thick films were prepared for TEM analyses and the results are Figure 2. Evolution of the roughness for the sputtered and shown in Figure 5 (a) and (b), respectively. The lamellas electrodeposited films as a function of the film’s thickness. were prepared via the ion beam milling technique. Firstly, platinum is deposited on the surface of the film thickness. The thinnest film (30 nm) has a roughness as a protective layer over the region of interest. Then, the of Ra = (0.9 ± 0.2) nm, whereas Ra = (3.2 ± 0.2) for the region is deep-trenched all around, separated from the thickest film (600 nm). These values are relatively substrate, and via a nanomanipulator probe transferred small for electrodeposited films, which are often and bonded to the TEM sample holder. Finally, the rougher than films produced by physical means. lamella is thinned down to a thickness of a few Sci. Technol. Adv. Mater. 21 (2020) 428 M. CIALONE et al. Figure 3. GIXRD plots labeled by film thickness from: (a) sputtered and (b) electrodeposited films. Intensity values are reported in linear scale. Unlabeled peaks in (a) belong to the substrate. Inset of panel (b) shows the geometry of the GIXRD set-up used for the acquisition of the GIXRD plots. Table 2. Values for the angular position (2 %\theta%), full of the Fe70Pd30 deposits is not observed here. The width at half maximum (FWHM) and average crystal size < crystal size of the electrodeposited films, as calculated D > for the sputtered films, the latter calculated by applying by applying the Scherrer’s formula on the (110) peak’s the Scherrer’s formula on (110) peak width of the α -(Fe,Pd) phase. width of the α-(Fe,Pd) phase, is around 31–32 nm Thickness (nm) 2 θ (º) FWHM (º) < D > (nm) (Table 3), hence larger than their sputtered counter­ 100 43.08 ± 0.02 1.67 ± 0.02 5±1 parts. Nevertheless, the films are also nanostructured. 300 43.40 ± 0.02 0.87 ± 0.02 10 ± 2 Note that microstrains also contribute to peak broad­ 500 43.51 ± 0.01 0.55 ± 0.01 15 ± 1 ening and thus the Scherrer’s formula underestimates crystal size. nanometers, using a low-intensity ionic current beam impinging on the lamella with a grazing angle [28]. 3.2. Magnetic properties From the cross-sections of Figure 5 (a) and (b) the presence of gross voids and cracks can be excluded. The magnetic hysteresis loops of Fe70Pd30 films pro­ Nevertheless, given the magnification used, the presence duced by both electrodeposition and sputtering were of cracks and voids in the sub-nanometer range cannot measured at room temperature along the in-plane be excluded. Furthermore, the composition of the films direction of the field. Different thicknesses were con­ was determined along their cross-sections using spot- sidered for both types of films and the results are EDS analyses in the TEM. A homogeneous composition shown in Figure 6 (a) and (b). The hysteresis loops matching a Fe70Pd30 stoichiometry was observed for of the sputtered films (Figure 6 (a)) are characterized both thicknesses. Note that a platinum layer was depos­ by low values of coercivity, HC, and a sharp irreversi­ ited on top of the Fe70Pd30 films to protect the surface ble jump of the magnetization. from oxidation during lamella fabrication by ion beam This behaviour is typical of soft magnetic materials. milling [28]. Moreover, the sputtered films are characterized by an The GIXRD curves of electrodeposited films with in-plane anisotropy (i.e. the easy axis for the magneti­ varying thicknesses (30 nm, 100 nm and 305 nm) are zation lies in the film plane) induced by the strong shown in Figure 3 (b). Unlike the 100 nm and 300 nm shape anisotropy due to the reduced film thicknesses. thick sputtered films, the plots show reflections origi­ Different values of coercivity ranging from (3.0 ± 0.1) nating from two different crystalline phases, namely, to (15.0 ± 0.3) Oe are observed, which do not vary α-(Fe,Pd) and tetragonal Pd1.5H2 phases. Interestingly, linearly with film thickness. Dissimilar coercive field the presence of the tetragonal Pd1.5H2 phase is inde­ values can be understood considering that the dimen­ pendent of film thickness, as it is detected even for the sions and shapes of the crystalline grains change thinnest coating. Its relative amount, however, posi­ depending on film thickness [32,33]. Indeed, the mag­ tively scales with film thickness. The formation of netic properties of nanocrystalline films depend on the palladium hydrides is attributed to the high affinity interplay between local magnetic anisotropy and between palladium and hydrogen [29,30]. This affinity exchange interactions, both of which are related to is underpinned by the absorption of hydrogen atoms the size of the crystalline grains. For large grains, the in the palladium crystal lattice. It was claimed that magnetization tends to align along the magnetic easy H:Pd ratios above 0.57 causes severe cracking of Pd axis direction of each grain. For large enough crystals, deposits during or after electrodeposition because of multiple domains can be generated [34]. Grain bound­ significant lattice distortions [31]. However, cracking aries indeed represent an obstacle for the motion of Sci. Technol. Adv. Mater. 21 (2020) 429 M. CIALONE et al. Figure 4. (a) SEM micrograph and (b) and (c) AFM images of the surface of the electrodeposited film with 30 nm thickness. (d) SEM micrograph and (e) and (f) AFM images of the surface of the electrodeposited film with 600 nm thickness. Figure 5. TEM images of the cross-section of the (a) 30 nm and (b) 600 nm thick electrodeposited films. domain walls, thereby hindering magnetization rever­ [37,38]. Indeed, according to the random anisotropy sal and increasing coercivity [35]. In general, as the model (originally proposed by Alben et al. [39]), for grain size is reduced, the coercive field first increases small grains, the exchange interaction forces the mag­ until it achieves a maximum and then it decreases netic moments of different grains to align parallel to [35,36]. The decrease in the value of coercivity for each other. As a consequence, the effective magnetic very small grain sizes is due to the enhanced exchange anisotropy is decreased (due to the random orienta­ interactions between neighbouring crystalline grains tion of exchange coupled grains) and coercivity Sci. Technol. Adv. Mater. 21 (2020) 430 M. CIALONE et al. Table 3. Values for the angular position, full width at half an indication of a lattice distortion that can be maximum (FWHM) and average crystal size < D > for the ascribed to the presence of internal stresses in the electrodeposited films, the latter calculated by applying the electrodeposited films. Even if the available data do Scherrer’s formula on the (110) peak width of the α -(Fe,Pd) phase. not allow elucidating the actual cause of the lattice Thickness (nm) 2 θ (º) FWHM (º) < D > (nm) distortion, it can be inferred that the presence of an 30 42.69 ± 0.05 0.26 ± 0.07 32 ± 4 additional phase in the electrodeposited films acts as 100 42.92 ± 0.03 0.87 ± 0.02 32 ± 4 a source of stresses in the deposit. Indeed, the PdH 305 43.10 ± 0.05 0.27 ± 0.06 31 ± 5 phase is embedded in the films, and because it has a lattice parameter (aPdH ≈ 2.99 Å) higher than the (which is proportional to the anisotropy) decreases as one of the α-(Fe,Pd) phase (aα-(Fe,Pd)≈ 2.87 Å), it acts well [34]. The exchange interaction between neigh­ as a source of internal stress. bouring crystalline grains increases with decreasing Considering the magnetic peculiarities of the elec­ crystal dimensions. The combination of these various trodeposited films, the configuration of their magnetic effects determines the modulation of the HC values domain patterns was investigated by MFM at room observed in the hysteresis loops of Figure 6 (a). temperature. The samples were previously saturated Regarding the hysteresis loops of the electrodepos­ with an in-plane magnetic field and subsequently mea­ ited films, a markedly different behaviour is observed sured at magnetic remanence. (Figure 6 (b)) as a function of film thickness. Both the The films displaying the transcritical hysteresis 30 nm and the 100 nm thick films show a soft mag­ loop (305 nm, 340 nm and 600 nm thick films) netic behaviour with in-plane anisotropy, similar to show the typical stripe domain pattern, as can be the sputtered films. These two films show different seen in Figure 7 (a), (b) and (c). The dark and bright coercivity values, from (20.0 ± 0.4) to (40.0 ± 0.8) Oe contrast is due to the magnetization canting up and which, as aforementioned, is ascribable to the differ­ down from the film plane [42]. Since MFM is only ent dimensions and hence, interactions between the sensitive to the out-of-plane magnetization compo­ crystalline grains. The shape of the hysteresis loops nent, the occurrence of a non-uniform contrast in drastically changes for the thickest films (305 nm, Figure 7 (a), (b) and (c) is a proof of the existence of 340 nm and 600 nm). These loops exhibit a sharp an out-of-plane component of the magnetization. jump of the magnetization at small fields around the Conversely, Figure 7 (d) shows a uniform contrast, coercive field. When the value of the field is further indicating that the magnetization is lying entirely in increased, the magnetization increases linearly before the film plane. reaching saturation, suggesting a rotation of the mag­ The domain pattern of the sputtered films was not netization from the perpendicular direction towards investigated by MFM since the hysteresis loops of the film plane. This particular shape of the loop is Figure 6 (a) show an in-plane anisotropy for all termed ‘transcritical loop’ [40,41], and indicates the films. Hence, the MFM images would have a constant existence of an out-of-plane anisotropy, hence, an contrast, providing no information on the actual out-of-plane component of the magnetization. The arrangement of the magnetic domains. origin of this component of the magnetic anisotropy perpendicular to the film plane can be correlated with 3.3. Mechanical properties the presence of strains inside the electrodeposited film. Indeed, for the electrodeposited film, a small As aforementioned, the Fe70Pd30 alloy shows a shift of the (110) peak of the α-(Fe,Pd) towards combination of shape memory effect with ferro­ lower angles have been observed. This shift is magnetic behaviour, which make thin films of this Figure 6. Hysteresis loops, measured at room temperature, for the in-plane orientation of the field corresponding to the (a) sputtered and (b) electrodeposited films with varying thickness. The insets show the evolution of the coercive field as a function of the film thickness for the sputtered and electrodeposited films. Sci. Technol. Adv. Mater. 21 (2020) 431 M. CIALONE et al. Figure 7. Room temperature MFM images, acquired at magnetic remanence, for the electrodeposited films with a thickness of (a) 600 nm, (b) 340 nm, (c) 305 nm and (d) 30 nm. composition promising candidates for the design of using a standard Berkovich-type tip. Because surface wirelessly actuated MEMS/NEMS or strain sensors roughness was very low, indentation could be per­ [11]. For this reason, it was deemed convenient to formed from on top. The maximum indentation load investigate and compare the nanomechanical prop­ was 0.3–0.4 mN, which was chosen to keep the pene­ erties of the sputtered and electrodeposited films. tration depth approximately 23–25 nm from the surface Indeed, as pointed by [1], the mechanical proper­ (i.e., below one-tenth of the total film thickness). Under ties of Fe70Pd30 alloy films have been largely over­ these conditions, the mechanical properties of films looked. This is particularly true when it comes to thicker than 250 nm could be investigated. Specifically, electrodeposited samples. sputtered and electrodeposited films having comparable Similar to the magnetic properties, the mechanical thicknesses (300–305 nm and 500–600 nm) were sub­ response of the films can be correlated with the crystal­ ject to indentation. lographic phases and grain dimensions of the films [36]. The obtained results are summarized in Tables 4 Given the reduced thickness of the coatings, the and 5, while representative load–displacement curves mechanical properties were studied by nanoindentation are shown in Fig. 8 (a) and (b). Table 4. Berkovich nanohardness (HB), reduced Young’s modulus (Er), HB/Er, H3B /E2r and Ue/Utot for sputtered Fe70Pd30 films of varying thickness (t). t (nm) Er (GPa) HB (GPa) HB/Er H3B /E2r (GPa) Ue (nJ) Up (nJ) Ue/Utot 300 129 ± 3 6.4 ± 0.2 0.050 ± 0.002 0.015 ± 0.003 0.0015 ± 0.0003 0.0019 ± 0.0002 0.5 ± 0.1 500 162 ± 18 8±1 0.05 ± 0.01 0.02 ± 0.01 0.0028 ± 0.0006 0.0026 ± 0.0002 0.4 ± 0.1 Table 5. Berkovich nanohardness (HB), reduced Young’s modulus (Er), HB/Er, H3B /E2r and Ue/Utot for electrodeposited Fe70Pd30 films of varying thickness (t). t (nm) Er (GPa) HB (GPa) HB/Er H3B /E2r (GPa) Ue (nJ) Up (nJ) Ue/Utot 305 105 ± 2 5.9 ± 0.1 0.056 ± 0.002 0.018 ± 0.001 0.0020 ± 0.0004 0.0022 ± 0.0002 0.5 ± 0.2 340 112 ± 12 6.5 ± 0.9 0.05 ± 0.01 0.02 ± 0.01 0.0016 ± 0.0002 0.0024 ± 0.0001 0.6 ± 0.1 600 115 ± 12 6.8 ± 0.9 0.06 ± 0.01 0.02 ± 0.01 0.0014 ± 0.0008 0.0022 ± 0.0002 0.6 ± 0.4 Sci. Technol. Adv. Mater. 21 (2020) 432 M. CIALONE et al. Figure 8. Load-displacement nanoindentation curves of the: (a) 305 nm thick electrodeposited (circles) and 300 nm thick sputtered (squares) films and (b) 600 nm thick electrodeposited (circles) and 500 nm thick sputtered (squares) films. Overall, the sputtered films Fe70Pd30 show larger Er also list several meaningful ratios, namely, HB/Er, H3B /E2r values than their electrodeposited analogues (cf. Tables 4 and Ue /Utot . The former can be regarded as an indirect and 5). Although the latter looked very compact (see assessment of the wear resistance of a coating [44]. Figure 5), a slightly higher density can be claimed for the Results indicate that this parameter is similar for sputtered films. Since stiffness generally scales up with both sputtered and electrodeposited films and it is density, this would explain the higher Er values of the about 0.05–0.06. Similar hardness to reduced elastic physically deposited films. Similarly, HB values are also modulus ratios were found in nanostructured Co-Ni- higher. The Er and HB values for the sputtered films are Re-P alloys fabricated by electrodeposition with 18 at.% in agreement with the literature. For example, HB = 7.3 of rhenium as refractory metal [45]. The H3B /E2r ratio, GPa and E = 140 GPa were reported for Fe70Pd30 films which allows to estimate the material’s ability to dissi­ grown by PVD on Si/SiO2 substrate. Chiu et al. obtained pate energy at plastic deformation during loading, is HB = 12.5 GPa and E = 152 GPa for Fe50Pd50 films with about 0.02 (and a bit less for the thinnest films). a crystal size of 12 nm grown by magnetron sputtering Finally, the ratio between Ue and Utot (elastic recovery) [1]. An important observation is that for both sputtered was estimated from the areas enclosed between the and electrodeposited films, HB and Er values increase unloading indentation segment and the displacement with thickness. This seems contradictory especially for axis (Ue) and between the loading segment and the the sputtered films since the crystal size of the α-(Fe,Pd) displacement axis (Utot). This ratio is of particular phase increases with thickness (see Figure 3 and Table interest in impact loading applications since it indicates 2). According to the Hall–Petch relationship, hardness is how much energy is released from the material after expected to increase with a decrease of crystal size due to being loaded. Results indicate that the thickest electro­ dislocation pile-up at grain boundaries. Therefore, there deposited films (340 nm and 600 nm) are able to dis­ has to be another factor which overcomes the apparent sipate more energy than their sputtered counterparts increase of crystal size with thickness. The reason behind (500 nm). Indeed, Ue/Utot and HB/Er values are exactly this trend might be related to the relatively high hard­ the same in some cases, as it happens in elastic perfectly ness of the fabricated Fe70Pd30 films. Under conditions plastic materials. Overall, the mechanical properties for which the coating is stiffer than the substrate, the indicate that the electrodeposited Fe70Pd30 films are one-tenth rule might not necessarily apply and, there­ mechanically as good as or even superior to their sput­ fore, an influence from the Si/SiO2 (400 nm) substrate tered analogues. for the lowest Fe70Pd30 sputtered film thicknesses cannot be excluded [43]. A similar reasoning can be made for the electrodeposited films, for which the substrate 4. Conclusions is Si/SiO2 (400 nm)/Cr (2 nm)/Au (100 nm); the layer in In this work, the magnetic and mechanical properties direct contact with the electrodeposited film (Au) is of nanostructured Fe70Pd30 films produced via definitely softer than the here-fabricated Fe70Pd30 coat­ potentiostatic electrodeposition and radio frequency ings. There could be another factor contributing to the sputtering were investigated and correlated with their thickness-dependence of HB in the electrodeposited microstructure. Both approaches were proven to coatings. Since the relative amount of palladium produce homogeneous films, with roughness below hydrides increases with deposit thickness (see section 5 nm. The sputtered films, independently from their Electrodeposited films) and these compounds are thickness, were single phase with a body-centred known to induce strain, progressively higher HB values cubic structure belonging to a supersaturated solid would be expected as the film builds up. Tables 4 and 5 solution of α – (Fe,Pd). The electrodeposited films Sci. Technol. Adv. Mater. 21 (2020) 433 M. CIALONE et al. show, alongside with the α-(Fe,Pd) phase, the pre­ Gabriele Barrera http://orcid.org/0000-0002-3174-8092 sence of palladium hydrides. The Pd1.5H2 phase Margitta Uhlemann http://orcid.org/0000-0002-2764- causes internal stresses in the electrodeposited film 0329 Annett Gebert http://orcid.org/0000-0003-2748-3850 which, in turn, influences the magnetic properties Jordi Sort http://orcid.org/0000-0003-1213-3639 of the material. In particular, the stresses induce Eva Pellicer http://orcid.org/0000-0002-8901-0998 out-of-plane magnetic anisotropy which results in Paola Rizzi http://orcid.org/0000-0002-3977-2839 a transcritical hysteresis loop and a striped magnetic Paola Tiberto http://orcid.org/0000-0002-7409-2281 domain pattern in films thicker than 100 nm. The presence of an out-of-plane anisotropy is relevant for a number of applications such as magnetic storage References devices, sensors or spintronics. Regarding the [1] Chiu YJ, Shen CY, Chang HW, et al. Characteristics mechanical behaviour, the sputtered films show of iron-palladium alloy thin films deposited by mag­ high reduced Young’s modulus and Berkovich hard­ netron sputtering. Results Phys. 2018;9:17–22. [2] Sun S, Murray CB, Weller D, et al. Monodisperse fept ness values, which can be attributed to their compact­ nanoparticles and ferromagnetic fept nanocrystal ness. Yet, the electrodeposited films looked rather superlattices. Science. 2000;287(5460):1989–1992. dense by TEM and thereby a pronounced decrease [3] Weller D, Moser A, Folks L, et al. High ku materials of these parameters was not observed. The reduced approach to 100 gbits/in2. IEEE Trans Magn. 2000 Young’s modulus and Berkovich hardness were Jan;36(1):10–15. thickness dependent, indicating an influence from [4] Chin TS. Permanent magnet films for applications in microelectromechanical systems. J Magn Magn the underlying substrate for the lowest film thickness Mater. 2000;209(1):75–79. tested (300 nm). Interestingly, rather large HB/Er and [5] Sharma P, Waki J, Kaushik N, et al. High coercivity H3B /E2r ratios were obtained for both sets of samples, characteristics of feptb exchange- coupled nanocom­ suggesting good wear resistance and the ability posite thick film spring magnets produced by to dissipate energy during plastic deformation. sputtering. Acta Mater. 2007;55(12):4203–4212. [6] Morley SA, Riley ST, Porro JM, et al. Effect of FePd Fundamental understanding of the influence of the alloy composition on the dynamics of artificial spin deposition technique and film thickness on the phy­ ice. Sci Rep. 2018 February;8:4–11. sical properties of Fe70Pd30 films will help choose the [7] Pérez-Landazábal J, Lambri O, Bonifacich F, et al. right synthetic conditions in view of the end Influence of defects on the irreversible phase transi­ application. tion in fepd ferromagnetic shape memory alloys. Acta Materialia. 2015;86:110–117. [8] Cui J, Shield T, James R. Phase transformation and magnetic anisotropy of an iron-palladium ferromag­ Acknowledgments netic shape-memory alloy. Acta Materialia. 2004;52 Founding of this work by Marie Sklodowska Curie Action (1):35–47. (ITN/ETN, Grant number 642642). Partial financial support [9] Ullakko K, Huang JK, Kantner C, et al. Large mag­ from the Spanish Government (Project MAT2017-86357- netic field induced strains in Ni2MnGa single crystals. C3-1-R and associated FEDER) and the Generalitat de Appl Phys Lett. 1996;69(13):1966–1968. Catalunya (Project 2017-SGR-292) is also acknowledged. [10] Sozinov A, Likhachev AA, Lanska N, et al. Nanofacility Piemonte is also acknowledged. Giant magnetic-field-induced strain in NiMnGa seven-layered martensitic phase. Appl Phys Lett. 2002;80(10):1746–1748. Disclosure statement [11] Kauffmann-Weiss S, Hamann S, Gruner ME, et al. Understanding the magnetic shape-memory system There are no conflicts to declare. Fe-Pd-X by thin film experiments and first principle calculations. Adv Eng Mater. 2012;14(8):724–749. [12] Caro P, Cebollada A, Briones F, et al. Structure and Funding chemical order in sputtered epitaxial FePd(001) alloys. J Crystal Growth. 1998;187(3):426–434. This work was supported by the Generalitat de Catalunya [13] Tiberto P, Barrera G, Celegato F, et al. Microstructural [2017-SGR-292]; H2020 Marie Skłodowska-Curie Actions evolution and magnetic properties in Fe50Pd50 [ITN/ETN - 642642]; Spanish Government [MAT2017- sputtered thin films submitted to post-deposition 86357-C3-1-R, associated FEDER]. annealing. J Alloys Compd. 2014;615:S236– S241. SI: ISMANAM 2013. [14] Bechtold C, Lotnyk A, Erkartal B, et al. Structural ORCID characterization of sputtered Fe70Pd30 thin films dur­ ing ex situ and in situ TEM heating. Adv Eng Mater. Matteo Cialone http://orcid.org/0000-0002-3018-2787 2012;14(8):716–723. Monica Fernandez-Barcia http://orcid.org/0000-0001- [15] Edler T, Buschbeck J, Mickel C, et al. Mechanisms of 6808-6267 stress generation and relaxation during pulsed laser Federica Celegato http://orcid.org/0000-0002-7076-6234 deposition of epitaxial Fe-Pd magnetic shape memory Marco Coisson http://orcid.org/0000-0002-2978-7615 alloy films on MgO. New J Phys. 2008 jun;10(6):063007. Sci. Technol. Adv. Mater. 21 (2020) 434 M. CIALONE et al. [16] Konczak C, Haehnel V, Schultz L, et al. Adjusting the [31] Baumgrtner M, Gabe D. Palladium-Iron alloy electro­ phase structure of electrodeposited Fe-Pd films. deposition. part I single metal systems. Trans IMF. Mater Chem Phys. 2016;174:150–155. 2000;78(1):11–16. [17] Wellings JS, Samantilleke AP, Warren P, et al. [32] Kaiser N. Review of the fundamentals of thin-film Comparison of electrodeposited and sputtered intrin­ growth. Appl Opt. 2002 Jun;41(16):3053–3060. sic and aluminium-doped zinc oxide thin films. [33] Thornton JA. Influence of substrate temperature and Semicond Sci Technol. 2008 sep;23(12):125003. deposition rate on structure of thick sputtered cu [18] Giroire B, Ahmad MA, Aubert G, et al. A comparative coatings. J Vac Sci Tech. 1975;12(4):830–835. study of copper thin films deposited using magnetron [34] Herzer G. Grain size dependence of coercivity and sputtering and supercritical fluid deposition techniques. permeability in nanocrystalline ferromagnets. IEEE Thin Solid Films. 2017;643:53–59. Functional Oxides Trans Magn. 1990 Sep;26(5):1397–1402. Synthesis, structure, properties and applications [35] Cullity B, Graham C. Introduction to magnetic mate­ (Symposium Z EMRS Fall Meeting 2016). rials. 1st ed. Piscataway (NJ): Wiley; 2009. ISBN:978- [19] Fernndez-Barcia M, Kurdi S, Terzopoulou A, et al. 0471477419 Comparative study of the sustainable preparation of [36] Arzt E. Size effects in materials due to microstructural femn thin films via electrodeposition and magnetron and dimensional constraints: a comparative review. co-sputtering. Surf Coat Technol. 2019;375:182–196. Acta Materialia. 1998;46(16):5611–5626. [20] Rezaei M, Ghorbani M, Dolati A. Electrochemical [37] Skomski R. Nanomagnetics. J Phys. 2003 may;15(20): investigation of electrodeposited Fe-Pd alloy thin R841–R896. films. Electrochim Acta. 2010;56(1):483–490. [38] Fukunaga H, Inoue H. Effect of intergrain exchange [21] Milad R, Davoud H, Mohammad G, et al. interaction on magnetic properties in isotropic nd-fe- Electrochemical behavior of nanostructured Fe-Pd b magnets. Jpn J Appl Phys. 1992 may;31(Part 1, No. alloy during electrodeposition on different substrates. 5A):1347–1352. J Electrochem Sci Technol. 2018;9(3):202–211. [39] Alben R, Becker JJ, Chi MC. Random anisotropy in [22] Inoue S, Inoue K, Fujita S, et al. Fe-Pd ferromagnetic amorphous ferromagnets. J Appl Phys. 1978;49 shape memory alloy thin films made by dual source (3):1653–1658. dc magnetron sputtering. Mater Trans. 2003;44 [40] Álvarez-Prado L, Alameda J. Micromagnetism of (2):298–304. nanowires with low out-of plane-anisotropy. [23] Iselt D, Gaitzsch U, Oswald S, et al. Electrodeposition Phys B Condens Matter. 2004;343(1): 241–246. and characterization of Fe80Ga20 alloy films. Proceedings of the Fourth Intional Conference Electrochim Acta. 2011;56(14):5178–5183. on Hysteresis and Micromagnetic Modeling. [24] Fischer-Cripps AC. Nanoindentation. 2nd ed. [41] Svalov AV, Aseguinolaza IR, Garcia-Arribas A, et al. New York: Springer-Verlag; 2004. Structure and magnetic properties of thin permalloy [25] Pharr G, Oliver W. Measurement of thin film films near the transcritical state. IEEE Trans Magn. mechanical properties using nanoindentation. MRS 2010 Feb;46(2):333–336. Bull. 1992;17(7):2833. [42] Yu J, Chang C, Karns D, et al. Thermal annealing effect [26] von Goldbeck O. Fepd ironpalladium. In: Ironbinary on FeCoB soft underlayer for perpendicular magnetic phase diagrams, 1st. 1982; Berlin, Heidelbersssg: recording. J Appl Phys. 2002;91(10):8357–8359. Springer. p. 88–91. [43] Chen J, Bull S. On the factors affecting the critical [27] Cialone M, Celegato F, Coïsson M, et al. Tailoring indenter penetration for measurement of coating magnetic properties of multicomponent layered hardness. Vacuum. 2009;83(6):911–920. structure via current annealing in FePd thin films. [44] Leyland A, Matthews A. On the significance of the h/e Sci Rep. 2017 dec;7(1):16691. ratio in wear control: a nanocomposite coating [28] Giannuzzi L, Stevie F. Introduction to focused ion approach to optimised tribological behaviour. Wear. beams. New York (NY): Springer; 2005. ISBN: 978- 2000;246(1):1–11. 0-387-23313-0 [45] Pan S, Özkale B, Sivaraman K, et al. Tailoring the [29] Flanagan TB, Oates WA. The palladium-hydrogen physical properties of electrode-posited CoNiReP system. Ann Rev Mater Sci. 1991;21(1):269–304. alloys with large re content by direct, pulse, and [30] DPBroom. Hydrogen storage materials. 1st ed. reverse pulse current techniques. Electrochim Acta. New York: Springer-Verlag; 2011. 2013;96:43–50.