This is the published version of the article: Eiler, Konrad; Fornell Beringues, Jordina; Navarro Senent, Cristina; [et al.]. «Tailoring magnetic and mechanical properties of mesoporous single-phase Ni- Pt films by electrodeposition». Nanoscale, Vol. 12, Issue 14 (April 2020), p. 7749-7758. DOI 10.1039/C9NR10757F This version is available at https://ddd.uab.cat/record/233361 under the terms of the license Jour nal Name Tailoring magnetic and mechanical properties of mesoporous single-phase Ni-Pt films by electrodeposition Konrad Eiler,†a Jordina Fornell,a Cristina Navarro-Senent,a Eva Pellicer,a and Jordi Sort‡a,b Homogeneous mesoporous Ni-rich Ni-Pt thin films with adjustable composition have been synthe- sised by one-step micelle-assisted electrodeposition. The films exhibit a face-centred cubic solid solution (single phase) and their magnetic and mechanical properties can be tuned by varying the alloy composition. In particular, the Curie temperature (TC ) is shown to decrease with the Pt content and thin films with a TC close to room temperature (i.e. Ni58 Pt42 ) and below can be produced. Hysteresis loops show a decrease of saturation magnetisation (Ms ) and coercivity (Hc ) with decreasing Ni content. A comparison of porous and dense films reveals significantly lower saturation magnetic field strength for porous films. Concerning mechanical properties, mainly two trends can be observed: a decrease of the Young’s modulus of the nanoporous films with respect to dense films by 10% in average and a progressive increase of Young’s modulus with the Ni content from 4.2 GPa to 5.7 GPa in both types of films. The tunability of properties and facility of synthesis make this alloy a promising material for microelectromechanical systems (MEMS). 1 Introduction ing heads for hard disks, where the main requirements include The ability to adjust the magnetic properties of a system with- corrosion resistance, low stress and thermal stability 4 . For many out drastically changing other physical properties—such as elec- applications, soft ferromagnetic materials with Curie temperature trical, mechanical or thermal properties—is beneficial for appli- (TC ) close to room temperature (RT) are of interest in order to re- cation in magnetic micro- and nanoelectromechanical systems duce energy consumption during storage and access of data. (MEMS/NEMS), where para- and ferromagnetic phases need to A TC close to RT also allows to study the magnetic behaviour be in electrical contact while ensuring mechanical compatibil- in this temperature region without applying heat, helping to ad- ity between different components 1 . If there is a low differ- vance in the theoretical models to understand the magnetic phase ence in composition between the paramagnetic and the ferro- transition effects. In general, TC can be tailored by varying the magnetic phase, there will be only little difference in electrical content of a ferromagnetic element in a solid solution comprising conductivity—thus avoiding local heating—, a low mismatch in ferromagnetic and paramagnetic elements. In this work, Ni-Pt is thermal expansion—avoiding excessive thermally induced inter- used to that purpose. While Pt is an excellent corrosion resistant facial stresses—and good mechanical integrity 2 . Single-phase al- material, Ni provides ferromagnetic properties. Moreover, Ni has loys have the potential to prevent these problems from occurring a high abundancy and provides mechanical stability and electri- intrinsically, and—in the case of full miscibility—do not undergo cal conductivity. This makes Ni-Pt a multifunctional alloy whose phase transition and avoid temperature-induced formation of sec- properties can be tuned by the composition, rendering it ideal for ondary phases 3 . Additionally, an electrical contact between two application in magnetic MEMS or NEMS. Ni and Pt promote the phases of similar composition is less susceptible to atmospheric formation of single-phase alloys due to their full miscibility, given corrosion caused by humidity. their identical lattice structure (fcc) and their comparable atomic Electrodeposited soft magnetic alloys are widely used as writ- radii 5 . Compared to coarse-grained materials, nanocrystalline met- als and alloys exhibit larger yield stress and hardness due to a Departament de Física, Universitat Autònoma de Barcelona, 08193 Bellaterra (Cer- enhanced dislocation pile-up at grain boundaries (Hall-Petch danyola del Vallès), Spain b strengthening) 6 . Using suitable conditions for electrodeposition, Institució Catalana de Recerca i Estudis Avançats (ICREA), 08010 Barcelona, Spain † Email: konrad.eiler@uab.cat nanocrystalline films can be readily obtained, thus improving the ‡ Email: jordi.sort@uab.cat mechanical properties without compromising any other proper- J our nal Name, [yea r][ ,vol . ],1–10 | 1 ties 7,8 . system. A dependence of TC , Ms and Hc on the composition Mesoporous materials, due to their high surface-to-volume ra- (Nix Pt1−x ) is observed, as well as progressive variation of their tio, are generally of interest in biomedicine, water remediation, mechanical properties. The nanocrystalline, single-phase mate- catalysis, and energy storage and conversion devices such as bat- rial is characterised for both mesoporous and dense morphology, teries and fuel cells 9–13 . In the case of nanoporous magnetic ma- revealing that the mesoporosity influences both the magnetic and terials, the coercivity can be reduced by applying voltage, due mechanical behaviour. to large electric charge accumulation effects on the surface (con- verse magnetoelectric effect) 14 . Again, this can reduce the energy 2 Experimental consumption during writing of information in magnetic storage The electrodeposition of mesoporous Ni-Pt films was carried out media and other magnetically actuated devices 15 . using an aqueous solution containing 200 mM NiCl2 , 3 mM Furthermore, the interest in the biomedical field and bio-MEMS Na2 PtCl6 · 6 H2 O, 200 mM H3 BO3 , 25 mM NH4 Cl, and 10 g/l is growing 4 . Here, mesoporous materials are appealing for appli- Pluronic P-123 (average molecular mass Mn = 5 800). HCl was cation in drug delivery since the pores may act as reservoirs for added to adjust the pH to 2.7. For the deposition of dense films, the drugs to be delivered 16 . the block copolymer P-123 was omitted while keeping the con- Nanoporous materials are often superhydrophobic (and centration of all other chemicals constant. oleophilic), a property that may be utilised in water remediation A three-electrode set-up with an Ag|AgCl 3 M KCl reference (i.e. oil decontamination). If the material is at the same time fer- electrode and a platinum wire as counter electrode was used romagnetic, it can easily be retrieved from the water and thereby while the temperature during deposition was kept at 30 °C. All recycled for multiple usage 17,18 . given potentials refer to the Ag|AgCl electrode. De-aeration with The formation of a single-phase alloy generally depends on nitrogen gas was performed before each deposition, and the de- the thermodynamics of the system; the phase diagram and the position was performed while stirring the electrolyte at 100 rpm. mixing enthalpy for the element pair being suitable indicators For electrodeposition, silicon substrates were metallised with as to whether a single-phase alloy can be obtained. The Ni-Pt a 10 nm Ti adhesion layer and a 200 nm Cu seed layer, both phase diagram shows full miscibility with the existence of ordered deposited by sputter deposition. The back of the substrates was phases, and an approximately linear dependence of TC on com- insulating due to the presence of SiO2 . Potentiostatic deposition, position, reaching 373 K (100 °C) at 26 at% Pt 19 . Ni-Pt exhibits using potentials between –1.3 V and –0.6 V and deposition times a negative mixing enthalpy 20 , favouring the formation of a single from 30 s to 180 s, were applied using an Autolab 302N poten- phase. However, especially by electrodeposition the stabilisation tiostat/galvanostat to deposit films with different Ni to Pt ratios. of metastable phases is possible. Since the deposition rate was higher for more negative potentials, Another requirement for alloy electrodeposition is that the de- the deposition time was adjusted for each deposition potential position potentials of the two elements need to be sufficiently in order to obtain similar film thickness for all compositions. In close together; complexing agents are commonly used to shift practice, this meant increasing the deposition time when moving the deposition potential for one of the elements in order to ap- towards more positive potentials (Figure 1). After deposition of proach the deposition potential of the other and thus achieve the porous films, the block copolymer was removed by ultrasonic their co-deposition 7 . The use of a polymeric surfactant to in- cleaning in ethanol for 10 min. troduce nanoporosity can interfere with the crystal growth of the metallic film 21 , eventually changing the microstructure and lead- ing to phase separation. A general issue in the electrodeposition of metallic films is also the incorporation of oxygen and hydrogen, leading to the formation of oxide/hydroxide phases 22,23 and hy- drogen embrittlement, to which Ni is rather susceptible 24 . Single-phase dense Ni-Pt films have already been electrode- posited from K2 PtCl4 containing solution at acidic pH for the purpose of oxygen reduction reaction in the full compositional range 25 . Here, the synthesis of Ni-Pt thin films is accomplished using the well-described mechanism of micelle-assisted electrodeposi- Fig. 1 Deposition rate and Faradaic efficiency of mesoporous Ni-Pt films tion to introduce porosity 26–28 . In this one-step deposition pro- as a function of the applied deposition potential. cess, a block copolymer forms micelles when its concentration in water is above the critical micellar concentration (cmc), and the In order to determine the Faradaic current efficiency, the the- metal ions assemble at the exterior (hydrophilic) part of the mi- oretical mass (mtheoretical ) of a Nix Pt(1−x) film was determined as- celles. The latter are thus co-deposited when the metal ions are suming that all current measured during the deposition process reduced at the working electrode. The polymer can later be easily corresponds to the reduction of the metal ions (Equation 1). dissolved, leaving a mesoporous metallic film behind. Since mostly magnetic properties of the Ni-Pt system are in- 1  M(Ni) M(Pt) Z t vestigated here, this work focuses on Ni-rich alloys of the Ni-Pt mtheoretical = x + (1 − x) i · dt (1) F z(Ni) z(Pt) 0 2| J our nal Name, [yea r][ ,vol . ], 1–10 where F is the Faraday constant, x is the atomic fraction of Ni in was applied with a harmonic displacement of 2 nm and a har- the deposit, M(Ni) and M(Pt) are the molar masses of Ni and Pt, monic frequency of 45 Hz. The experiments were performed in z(Ni)=2 and z(Pt)=4 are the numbers of electrons supplied for displacement control mode, with a strain rate of 0.05/s up to a the reduction of a Ni and a Pt ion, respectively. i is the measured maximum penetration depth of 100 nm. The Poisson’s ratio ν current and t is the deposition time. was assumed to be 0.3. Sixteen indents, separated by 10 µm Together with the real masses of the deposits determined by each, were performed on each sample from on top (on the films’ ICP-MS (see below for the procedure), which also allowed the surfaces). determination of the atomic fractions of Ni and Pt, the Faradaic Magnetic hysteresis loops were recorded at RT in-plane and efficiency is determined by the ratio of the real and the theoretical out-of-plane by vibrating sample magnetometry (VSM) on an mass mreal /mtheoretical . LOT-QuantumDesign MicroSense VSM up to 20 000 Oe. VSM For determination of the deposition rates, the real masses of was also used for measuring the temperature dependence of the the deposits were used to estimate their density, from which the magnetisation in the saturated regime at 1 000 Oe in order to de- total thickness was calculated. Thus the deposition rates reported termine TC , using nitrogen gas flow to control the temperature. show the total thickness divided by the deposition time. After first measurements of TC , the deposition potential was re- The synthesised Ni-Pt thin films were analysed by scanning fined in the range between –0.70 V and –0.60 V to obtain films electron microscopy (SEM) coupled with energy-dispersive X-ray with a TC in the desired range. All thin films were subsequently spectroscopy (EDX) on a Zeiss Merlin electron microscope. Imag- dissolved in aqua regia for chemical analysis by ICP-MS using an ing was done at an acceleration voltage of 1–2 kV using an InLens Agilent 7500ce spectrometer to obtain the exact compositions of detector to reveal the existence of mesoporosity. the Ni-Pt films, and to normalise the measured magnetic moment Transmission electron microscopy (TEM) was performed on a by the total mass of Ni and Pt for each sample. TC was determined Jeol JEM-2011 electron microscope with an acceleration voltage by the two-tangent method, applying tangents to the linear parts of 200 kV working in bright field and diffraction mode. Sample of the M-T (magnetisation over temperature) curve both below preparation for TEM was performed by making a cross-cut of a and above the magnetic transition and taking TC at the point of mesoporous sample with a diamond saw. By grinding and polish- their intersection 33 . ing with a final diamond polish of 1 µm particle size, the thickness was reduced to about 30 µm. Finally, Ar ion milling with an en- 3 Results and discussion ergy of 5 keV was performed at an angle of 8° for several hours to 3.1 Chemical composition reduce the thickness locally to a few nanometers. For the prepa- The composition of the synthesised thin films is determined ration of the cross-section of a dense Ni-Pt film, a lamella was cut by the electrodeposition parameters, mainly the deposition using a focused ion beam SEM (FIB-SEM). potential. Since Pt(IV) deposits at less negative potentials than Grazing incidence X-Ray diffraction (GIXRD) was conducted on Ni(II), the Ni content increases when the potential is more a Malvern-PANalytical Xpert Pro MRD diffractometer using Cu- negative (Table 1). The same trend was observed by Liu et Kα radiation for phase analysis of the films in a 2θ range from al. during the electrodeposition of dense Ni-Pt films from a 38° to 62°. Rietveld refinement was performed using the soft- different bath formulation 25 . Progressive enrichment in Ni as ware MAUD 29 in order to determine the lattice parameter and the applied potential was made more negative was accompanied crystallite size for each composition 30 . by an increase in the deposition rate up to –1.2 V (Figure 1). The mechanical properties of the films, i.e. hardness (H) and The compositions of the dense thin films are similar to their reduced Young’s modulus (Er ), were evaluated by nanoindenta- porous counterparts deposited under the same conditions, i.e. tion. The continuous stiffness measurement (CSM) technique was the composition is not influenced by the presence of the P-123 used since it allows the assessment of depth-dependent proper- triblock copolymer for a given potential. All percentages of Ni ties of materials in a single step. The method involves applying a and Pt in this work, ranging from 61 at% Ni to 99 at% Ni, are dynamic (oscillatory) load on top of a static load while the mate- given in atomic percentage. For the film compositions, oxygen rial is penetrated in order to determine the continuous stiffness, or other impurities are not taken into account, however those which is then further processed to calculate H and Er 31 . were monitored by EDX and the oxygen content never exceeded Er can be seen as a combined modulus of the tested sample and 10 at% (note that in EDX, it is not possible to discriminate the indenter (Equation 2) 32 . whether the oxygen originates from the film or the substrate 1 1 − νs2 1 − νi2 layers). Current efficiencies determined for the mesoporous Ni-Pt = + (2) films were between 60% and 80% (Figure 1). Er Es Ei where ν is the Poisson’s ratio and the indices r, s and i stand for reduced, sample and indenter, respectively. The nanoindentation tests were carried out with a Nanoinden- 3.2 Microstructure ter XP from MTS using a Berkovich-shaped diamond tip with All films deposited from the electrolyte containing P-123 exhibit Ei = 1140 and νi = 0.07. Due to the large Young’s modulus of a homogeneously distributed mesoporosity (Figure 2). The thin the indenter in comparison with most metals, the difference be- films containing 99% and 98% Ni have a clearly visible grain tween Er and Es is usually within a few per cent. The CSM mode structure with an appreciable roughness (Figure 2a and b). For J our nal Name, [yea r][ ,vol . ],1–10 | 3 Table 1 Composition of dense and mesoporous Ni-Pt films produced at a) b) given deposition potentials and determined by ICP-MS. Deposition Composition potential [V] porous dense -1.3 Ni99 Pt1 -1.2 Ni98 Pt2 Ni98 Pt2 -1.0 Ni95 Pt5 Ni93 Pt7 -0.9 Ni92 Pt8 Ni91 Pt9 -0.8 Ni84 Pt16 -0.7 Ni76 Pt24 Ni79 Pt21 -0.6 Ni61 Pt39 c) d) 95% Ni, a rough surface without perceptible grain structure is ob- served (Figure 2c), while the films ranging from 94% to 61% Ni are smoother, since no topographic contrast other than the one caused by the porosity is visible. Furthermore, those films do not show any apparent grain structure on the nanoscale, and they are highly mesoporous with a narrowly distributed pore size on the order of 10 nm. The microstructures of the dense films show the continuous sur- e) f) faces with an appreciable, low roughness (Figure 3). On a cross-section of a Ni92 Pt8 thin film, the porosity over the full film thickness is observed by TEM, together with the Cu seed layer (Figure 4a). Underneath the Cu layer, the Ti layer and Si substrate are vis- ible. The irregular shape of the film suggests that the film sur- face has been partially removed during ion polishing. SEM mi- crographs of this cross-section reveal the film’s homogeneity at surface level and the film thickness is around 280 nm (Figure 4b). At high resolution, crystal planes become visible and indicate the nanocrystalline structure of the thin film (Figure 5). A few g) h) nanometre thick layer covers the surface of the thin film, which may be a surface oxide, or a damaged layer caused by amorphi- sation of the metal due to the ion beam polishing during prepara- tion. In the cross-section of the dense Ni91 Pt9 film, a film thickness of ca. 250 nm is observed (Figure 6a). As for the mesoporous film (Figure 5), its nanocrystalline morphology is revealed in high resolution conditions (Figure 6b). 3.3 Phase analysis Fig. 2 SEM micrographs of mesoporous Ni-Pt thin films a) Ni99 Pt1 , b) The diffraction patterns obtained by GIXRD show an fcc single- Ni98 Pt2 , c) Ni95 Pt5 , d) Ni94 Pt6 , e) Ni92 Pt8 , f) Ni84 Pt16 , g) Ni76 Pt24 , h) phase solid solution with a cell parameter between those for fcc Ni61 Pt39 acquired by InLens detector. Ni and fcc Pt (as indicated by the discontinuous lines in Figure 7) for all compositions. The diffraction patterns of the Ni-Pt films are superposed on the reflections of the fcc Cu seed layer (Figure 7). a tendency towards cell parameters slightly higher than expected Remarkably, there is no apparent difference in the diffraction (Figure 8). This effect was also observed by Kumar et al 35 , and patterns between dense and porous films of identical composi- related to the size mismatch due to the different atomic radii of tion, i.e. the porosity has very little or no influence on the struc- Ni and Pt. ture (cf. Figure 7, by comparing curves a1 and a2 or d and e). Similar deviations, ascribed to the presence of tensile micros- Upon increasing Pt content in the alloy, the diffraction peaks train, have been documented for electrodeposited Ni films 36 . shift towards lower diffraction angles, indicating a progressive With the exception of the film containing 98 at% Ni, the re- increase in the cell parameter. The cell parameter of the Pt-Ni flections of the Ni-Pt phase are significantly broadened due to the phase obtained by Rietveld refinement scales linearly with the al- small crystallite size. The Rietveld refinement returned crystallite loy composition, thus following Vegard’s law 34 . However, there is sizes on the order of 5 nm for Ni contents between 61% and 92%, 4| J our nal Name, [yea r][ ,vol . ], 1–10 a) b) a) b) Fig. 3 SEM micrographs of dense Ni-Pt thin films a) Ni98 Pt2 , b) Ni91 Pt9 Fig. 6 TEM micrographs of dense Ni91 Pt9 thin film a) over the full cross- acquired by InLens detector. section and b) under high-resolution conditions. a) b) Fig. 4 a) TEM and b) SEM micrograph of a cross-section of mesoporous Ni92 Pt8 thin film. Fig. 7 GIXRD diffraction patterns of Ni-Pt thin films showing the decon- volution of the Ni-Pt (green) and Cu (red curve) phases after Rietveld refinement for a1 ) Ni98 Pt2 , a2 ) dense Ni98 Pt2 , b) Ni92 Pt8 , c) Ni84 Pt16 , d) dense Ni79 Pt21 , e) Ni76 Pt24 , f) Ni61 Pt39 . With exception of a2 ) and d) all films are mesoporous. Fig. 8 Dependence of cell parameter a on composition of Ni-Pt thin films Fig. 5 High resolution TEM micrograph of a cross-section of mesoporous as determined in this work as well as by Kumar et al 35 . Values for pure Ni92 Pt8 thin film with selected crystal planes indicated. The correspond- Ni and Pt are included as a reference 37 . ing fast Fourier transform (FFT) image for the entire zone is shown as an inset on the top left. are no other structural differences apparent. and 50 nm in the case of Ni98 Pt2 . Similar values are obtained The nanocrystalline nature of the films is also demonstrated by for dense films, hence the nanocrystallinity resulted entirely from TEM under diffraction conditions, showing that crystals are ori- the electrodeposition parameters and electrolyte composition (Ta- ented in all possible directions within the cross-section (Figure 9). ble 2). The presence of the P-123 surfactant did not provoke or influence the formation of a nanocrystalline structure. It can be 3.4 Mechanical properties concluded that differences in behaviour between mesoporous and Representative curves obtained from nanoindentation tests dis- dense films are entirely related to effects of porosity, since there play the dependences of Er and H on the penetration depth into J our nal Name, [yea r][ ,vol . ],1–10 | 5 Table 2 Crystallite sizes of representative Ni-Pt thin films obtained by a) Rietveld refinement of their respective GIXRD patterns. Composition Morphology Crystallite size [nm] Ni98 Pt2 porous 41 Ni98 Pt2 dense 53 Ni92 Pt8 porous 7 Ni84 Pt16 porous 6 Ni79 Pt21 dense 5 Ni76 Pt24 porous 5 Ni61 Pt39 porous 6 b) Fig. 9 Selected area electron diffraction pattern taken at a cross-section of mesoporous Ni92 Pt8 with the Miller indices of the diffraction rings indi- cated. Fig. 10 Representative curves of a) the reduced Young’s modulus and b) hardness as a function of the nanoindenter’s penetration into the surface the surface for selected samples (Figure 10). Nanoindentation for selected dense and porous thin films. measurements may have some influence from the substrate. It is commonly accepted that if the maximum penetration depth is kept lower than one tenth of the overall film thickness, the contri- of the porous and dense counterparts is the following bution of the substrate can be disregarded 31 . Considering the ex-  3/2 perimentally determined film thickness of approximately 280 nm ρporous σporous = C2 σbulk (4) for the Ni-Pt films, the contribution of the substrate may not be ρbulk neglected for penetration depths higher than 30 nm. Neverthe- where C2 is equal to 0.3 41 . Therefore, porosity has a more dras- less, since the film thickness is similar for all compositions, the tic influence on the Young’s modulus than on hardness since, as influence of the substrate on the obtained mechanical properties shown in the equations above, Eporous /Ebulk is proportional to may be considered similar for all samples, and thus the observed (ρporous /ρbulk )2 (where ρporous /ρbulk is the relative density of the trends in Er and H are representative of the films’ properties. material), whereas hardness is proportional to (ρporous /ρbulk )3/2 . For a given composition, the dense films show a higher Young’s Comparing the results for mesoporous and dense Ni-Pt films, the modulus than their porous counterparts, as expected due to Young’s modulus shows a stronger influence of porosity than the their higher density (Figure 10a). The decrease of hardness and hardness (Figure 11). Indeed, the decrease in hardness due to Young’s modulus with porosity is a well-documented effect 38–40 . the occurrence of porosity is only apparent for the Ni98 Pt2 film. The relation between the Young’s moduli of the porous and bulk For the rest of compositions, the hardness of the porous films is materials is given by equal or slightly higher than for their dense counterparts.  ρporous 2 Comparing the reduced Young’s moduli of porous and non- Eporous = C1 Ebulk (3) porous films with similar composition using Equation 3, the rel- ρbulk ative density is different for each composition. From the mi- where C1 is a geometry constant close to 1 41 . In turn, hardness is crostructures it can be assumed that the mesoporosity intro- directly related to the yield stress σ according to H = 3σ for met- duced by the surfactant is constant over all compositions, how- als, which has been confirmed to hold for a nanoporous, single- ever, Ni98 Pt2 may have some additional porosity due to voids phase metallic material 42 . The relation between the yield stress in-between the grains (cf. Figure 2). For Ni98 Pt2 , the relative 6| J our nal Name, [yea r][ ,vol . ], 1–10 was determined for nanocrsytalline Pt for grain sizes lower than 10 nm 44 . The Ni-Pt films possess hardness values lying between 4 GPa and 6 GPa. The generally high values are a result of the nanocrys- tallinity, however, internal stresses may also contribute to an in- creased hardness. In the compositional range between 75% and 95% Ni, where the grain size is constant, a trend of increasing hardness with increasing Pt content is observed, which may be related to a solid solution strengthening. For Ni98 Pt2 , this trend is not continued and its higher hardness may be the result of the difference in grain size (cf. Figure 2b and Table 2). It has been shown that below a grain size of 14 nm, the breakdown of the Hall-Petch relationship is reached for Ni 43 . Thus, apart from the Ni98 Pt2 films, all other Ni-Pt films are subject to an inverse Hall- Petch effect, resulting in lower hardness values for those films. An effect of the porosity on the hardness of the films in the compositional range between 75% and 95% Ni lies within the measurement uncertainty and cannot be observed (Figure 10b and 11a). Only for the composition Ni98 Pt2 , the hardness is sig- nificantly lower when porosity is present. 3.5 Magnetic properties In-plane hysteresis loops conducted by VSM show highest satu- ration magnetisation Ms and coercivity Hc for Ni-rich films, both decreasing with Pt content (Figure 12). A change of slope (the orientation of magnetic domains is increasingly hindered) occur- Fig. 11 Averaged reduced Young’s modulus and hardness of porous and dense Ni-Pt thin films (a) and the relative changes in Er and H introduced ring between the coercivity field and the saturation magnetisation by porosity, in percentage of the values determined for the dense films is observed for both mesoporous and dense films. This effect is (b), at a penetration depth of 40 nm. generally stronger for dense films—with the exception of Ni98 Pt2 which shows the same behaviour as its porous counterpart— indicating that the observed behaviour may be caused by internal density yields 61%, i.e. a porosity of 39%. Contrarily, for the stresses which are less pronounced in the porous films. In the pair Ni94 Pt6 (porous) and Ni93 Pt7 (dense), the determined rela- case of Ni98 Pt2 , it is assumed that internal stresses are minimised tive density is 92% (and thus corresponding to a porosity of 8%). due to the fact that the material is almost pure Ni and therefore The relative densities of the two remaining porous/non-porous the effect of stresses due to dissolution of Pt (with a higher atomic pairs (Ni92 Pt8 /Ni91 Pt9 and Ni76 Pt24 /Ni79 Pt21 ) are 82% and 83%, radius) into the Ni lattice is smaller. respectively. An interplay between porosity, Ni/Pt ratio and grain size may be appreciated in the Young’s modulus. As observed, E is al- ways larger in bulk solid films than in their porous counterparts. Also, the Young’s modulus tends to increase with the Ni content in agreement with the rule of mixture of two different elements where ENi = 200 GPa and EPt = 172 GPa 5 . Grain size might also have an influence on E but its effect should be negligible com- pared to the other two. Depending on the state of annealing, the hardness of a pure metal can vary significantly from a metallurgical viewpoint. Gen- erally, Ni bulk metal can reach a Vickers hardness of 1.7 GPa in an annealed state, and 6.3 GPa in a hardened state 5 . Concerning electrodeposited pure Ni, nanocrystalline films with grain sizes (d) between 12 nm and 22 nm exhibit hardness values between 3.9 (d =22 nm) and 6.4 GPa (d =14) nm 43 —in this case, the high hardness is a result of the nanocrystallinity. For Pt bulk ma- terial, the hardness values vary between 0.4 GPa (annealed) and 2.1 GPa (hardened) 5 . The inversion of the Hall-Petch relation- Fig. 12 In-plane magnetic hysteresis loops for porous and dense Ni-Pt ship due to a change in the mechanism of plastic deformation thin films with different compositions. J our nal Name, [yea r][ ,vol . ],1–10 | 7 The saturation magnetisation Ms follows a linear trend with the composition for all films. For a given composition, the mass- normalised Ms is identical for both dense and porous films. The same is true for the coercivity of the porous films (Figure 13). For Ni98 Pt2 , Ms reaches 46 emu/g, approaching the value for pure Ni of 54 emu/g 45 and thus confirming again that the oxygen content is negligible. Fig. 14 In-plane and out-of-plane magnetic hysteresis loops for two se- lected dense and porous Ni-Pt thin films with similar composition. Fig. 13 Dependence of saturation magnetisation Ms and coercivity Hc on composition of Ni-Pt thin films, taken from in-plane hysteresis mea- surements. The linear trends for the mesoporous films are displayed as dotted lines. Comparing the behaviour along in-plane and out-of-plane mea- suring directions, it is clearly observed that the magnetically hard axis lies out of plane, where the loops are strongly tilted with re- spect to the in-plane direction. Therefore, the hysteresis loops of dense and porous films are dominated by shape anisotropy. Here, the orientation of magnetic domains normal to the plane is en- ergetically unfavourable, and an increased energy input (applied field) is needed to fully magnetise the films normal to the plane (Figure 14). The out-of-plane coercivity is lower, i.e. it is easier to demagnetise the material in this mode compared to in-plane. Fig. 15 Dependence of magnetisation on temperature of mesoporous Contrarily to the in-plane mode, the out-of-plane hysteresis loops Ni-Pt thin films at 1 000 Oe. show a widening when approaching saturation. This effect is gen- erally observed in Ni-rich alloy films—permalloy in particular—as the so-called transcritical state 46 . When the film thickness is suf- ficiently high, magnetic stripe domains form by part of the mag- netisation orienting out of the plane (when applying an external magnetic field in plane) due to an anisotropy caused by in-plane internal stresses 46 . The temperature scans reveal a phase transition of the Ni-Pt al- loys from ferromagnetic to paramagnetic state (Figure 15). The fact that a single transition is observed consolidates that the struc- ture of the films is indeed single-phase. As a general trend, the magnetic transition and therefore the Fig. 16 Curie temperature TC as a function of composition of meso- Curie temperature shifts to lower temperatures for films with porous Ni-Pt thin films. TC for pure Ni is shown for reference . The gen- eral trend is indicated by a dotted line. higher Pt content. The magnetisation curve for Ni58 Pt42 shows only the final part of the phase transition, indicating that its TC is very close to RT (Figure 15, violet curve). The resulting values for TC show a linear relationship with the 4 Conclusions percentage of Ni (Figure 16). The TC was only determined for Mesoporous and dense Ni-Pt thin films have been synthesised by those films where the magnetic transition appeared sufficiently electrodeposition from aqueous media, achieving the mesoporos- above the start of the measurement range (298 K). ity through micelle-assisted deposition with a block copolymer. 8| J our nal Name, [yea r][ ,vol . ], 1–10 The composition of the thin films can be tuned in a large range 6 J. Weertman, Mat. Sci. Eng. A, 1993, 166, 161. between Ni98 Pt2 and Ni61 Pt39 by varying the deposition potential, 7 E. Pellicer, A. Varea, S. Pané, B. J. Nelson, E. Menéndez, resulting in a single-phase Ni-Pt solid solution in all cases. The M. Estrader, S. Suriñach, M. D. Baró, J. Nogués and J. Sort, porosity is homogeneously distributed for all compositions and Adv. Funct. Mater., 2010, 20, 983. the nanocrystallinity is not affected by the porosity. Crystallite 8 H. Li and F. Ebrahimi, Mat. Sci. Eng. A, 2003, 347, 93. sizes are around 50 nm for the most Ni-rich composition Ni98 Pt2 , 9 Z. Li, J. Barnes, A. Bosoy, J. Fraser Stoddart and J. Zink, Chem. and between 5–7 nm for the other compositions. The mechanical Soc. Rev., 2012, 41, 2590. properties are strongly dependent on both porosity and composi- 10 N. Linares, A. Silvestre-Albero, E. Serrano, J. Silvestre-Albero tion. The ratio of the reduced Young’s moduli suggests a porosity and J. García-Martínez, Chem. Soc. Rev., 2014, 43, 7681. of around 18% for Pt contents of 8 at% and more, whereas a 11 L. Zhang, L. Jin, B. Liu and J. He, Front. Chem., 2019, 7, 22. porosity of 39% is obtained for Ni98 Pt2 . Apart from the effect of 12 L. X. Ding, A. L. Wang, G. R. Li, Z. Liu, W. X. Zhao, C. Su and porosity, the reduced Young’s modulus shows a strong composi- Y. X. Tong, J. Am. Chem. Soc., 2012, 134, 5730. tional dependence, increasing with the Ni content. The measured 13 A. L. Wang, H. Xu, J. X. Feng, L. X. Ding, Y. X. Tong and G. R. hardness shows the effect of solid solution strengthening, and Li, J. Am. Chem. Soc., 2013, 135, 10703. an effect of the microstructure in the case of Ni98 Pt2 , due to the significantly different morphology and grain size at this 14 A. Quintana, J. Zhang, E. Isarain-Chávez, E. Menéndez, particular composition. The single-phase character also makes it R. Cuadrado, R. Robles, M. D. Baró, M. Guerrero, S. Pané, possible to tune the magnetic properties with the composition; B. J. Nelson, C. M. Müller, P. Ordejón, J. Nogués, E. Pellicer TC can be adjusted to a desired value reaching from the TC of Ni and J. Sort, Adv. Funct. Mater., 2017, 27, 1701904. (630 K) down to RT and below. Adjusting TC to RT may allow 15 C. Navarro-Senent, A. Quintana, E. Menéndez, E. Pellicer and to study the magnetic behaviour close to TC , and other possible J. Sort, APL Mater., 2019, 7, 030701. effects such as a voltage-dependency of TC . Apart from TC , the 16 B. G. Trewyn, S. Giri, I. I. Slowing and V. S.-Y. Lin, Chem. saturation magnetisation and—in the case of the mesoporous Commun., 2007, 31, 3236. films—also the coercivity scales with the composition, facilitating 17 R. Du, Q. Zhao, Z. Zheng, W. Hu and J. Zhang, Adv. Energy to predict the magnetic behaviour for a certain composition and Mater., 2016, 6, 1600473. to easily select a suitable composition in order to obtain the 18 O. Kharissova, H. V. Rasika Dias and B. I. Kharisov, RSC Adv., desired magnetic properties. 2015, 5, 6695. 19 P. Nash and M. F. Singleton, Bull. Alloy Phase Diagr., 1989, 10, 258p. Conflicts of interest 20 R. A. Walker and J. B. Darby Jr., Acta Metall., 1970, 18, 1261. There are no conflicts to declare. 21 A. Gomes and M. da Silva Pereira, Electrochim. Acta, 2006, Acknowledgements 51, 1342. 22 C. Navarro-Senent, J. Fornell, E. Isarain-Chávez, A. Quintana, This work has received funding from the European Union’s Hori- E. Menéndez, M. Foerster, L. Aballe, E. Weschke, J. Nogués, zon 2020 research and innovation programme under the Marie E. Pellicer and J. Sort, ACS Appl. Mater. Interfaces, 2018, 10, Skłodowska-Curie grant agreement no 764977. Further funding 44897. has been obtained from the Generalitat de Catalunya (2017-SGR- 23 E. Isarain-Chávez, M. D. Baró, E. Pellicer and J. Sort, 292 Project) and the Spanish Ministerio de Economía, Indústria y Nanoscale, 2017, 9, 18081. Competitividad (MAT2017-86357-C3-1-R and associated FEDER 24 B. Craig, ASM Handbook, Volume 13A: Corrosion: Fundamen- Project). The European Research Council (SPIN-PORICS 2014- tals, Testing, and Protection, ASM Intl., 2003, pp. 367–380. consolidator grant, agreement no 648454) is also acknowledged. J.F. is grateful to MINECO for the “Juan de la Cierva” (IJCI-2015- 25 Y. Liu, C. M. Hangarter, U. Bertocci and T. P. Moffat, J. Phys. 27030) contract. Chem. C, 2012, 116, 7848–7862. 26 V. Malgras, H. Ataee-Esfahani, H. Wang, B. Jiang, C. Li, K.-W. Notes and references Wu, J. Kim and Y. Yamauchi, Adv. Mater., 2015, 28, 993. 1 G. Schiavone, M. P. Y. Desmulliez and A. J. Walton, Microma- 27 H. Wang, L. Wang, T. Sato, Y. Sakamoto, S. Tominaka, chines, 2014, 5, 622. K. Miyasaka, N. Miyamoto, Y. Nemoto, O. Terasaki and Y. Ya- 2 MEMS: A practical guide to design, analysis, and applications, mauchi, Chem. Mater., 2012, 24, 1591. ed. J. Korvink and O. Paul, William Andrew, NY and Springer- 28 J. Zhang, A. Quintana, E. Menéndez, M. Coll, E. Pellicer and Verlag, Germany, 2006. J. Sort, ACS Appl. Mater. Interfaces, 2018, 10, 14877–14885. 3 R. Coutu, Jr, P. Kladitis, K. Leedy and R. Crane, J. Micromech. 29 Materials Analysis Using Diffraction, http://maud. Microeng., 2004, 14, 1157. radiographema.com/, (accessed Dec 2019). 4 D. Niarchos, Sensors and Actuators A, 2003, 106, 255. 30 L. Luterotti, Nucl. Instrum. Methods Phys. Res. B, 2010, 268, 5 F. Cardarelli, Materials Handbook, Springer Intl. Publishing 334. AG, Switzerland, 3rd edn, 2018. 31 A. C. Fischer-Cripps, Nanoindentation, Springer, NY, 1st edn, 2002. J our nal Name, [yea r][ ,vol . ],1–10 | 9 32 W. C. Oliver and G. M. Pharr, J. Mater. Res., 1992, 7, 1564. 40 J. Biener, A. M. Hodge, A. V. Hamza, L. M. Hsiung and J. H. 33 C. S. Grommé, T. L. Wright and D. L. Peck, J. Geophys. Res., Satcher Jr., J. Appl. Phys., 2005, 97, 024301. 1969, 74, 5277. 41 L. J. Gibson and M. F. Ashby, Cellular solids: structure and 34 L. Vegard, Z. Physik, 1921, 5, 17. properties, Cambridge University Press, UK, 2nd edn, 1997. 35 U. Kumar, K. G. Padmalekha, P. K. Mukhopadhyay, D. Paudyal 42 H. J. Jin, L. Kurmanaeva, J. Schmauch, H. Rösner, and A. Mookerjee, J. Magn. Magn. Mater., 2005, 292, Y. Ivanisenko and J. Weissmüller, Acta Mater., 2009, 57, 2665. 234–240. 43 C. Schuh, T. Nieh and T. Yamasaki, Scr. Mater., 2002, 46, 735. 36 S. Sengupta, A. Patra, S. Jena, K. Das and S. Das, Metall. 44 X. Shu, D. Kong, Y. Lu, H. Long, S. Sun, X. Sha, H. Zhou, Mater. Trans. A, 2018, 49A, 920–937. Y. Chen, S. Mao and Y. Liu, Sci. Rep., 2017, 7, 13264. 37 A. W. Hull, Phys. Rev., 1921, 17, 571–588. 45 B. Cullity and C. Graham, Introduction to magnetic materials, 38 F. Tancret and F. Osterstock, Philos. Mag., 2003, 83, 125. John Wiley & Sons, Inc., NJ, 2nd edn, 2009. 39 S. Cariou, F.-J. Ulm and L. Dormieux, J. Mech. Phys. Solids, 46 A. V. Svalov, I. R. Aseguinolaza, A. Garcia-Arribas, I. Orue, 2008, 56, 924. J. M. Barandiaran, J. Alonso, M. L. Fernández-Gubieda and G. V. Kurlyandskaya, IEEE Trans. Magn., 2010, 46, 333. 10 | J our nal Name, [yea r][ ,vol . ],1–10