Journal of Materials Chemistry C · Periodic arrays of compositionally graded Cu–Ni alloy nanopillars (100 or 200 nm in diameter and 450 nm ... Atom probe tomography of the Cu-rich

Journal ofMaterials Chemistry C

PAPER

Publ

ishe

d on

09

Sept

embe

r 20

13. D

ownl

oade

d by

Uni

vers

itat A

uton

oma

de B

arce

lona

on

17/1

0/20

13 1

5:25

:38.

View Article OnlineView Journal | View Issue

aDepartament de F́ısica, Facultat de CièncCampus UAB, E-08193 Bellaterra, Spain. E-bInstitute of Robotics and Intelligent System

SwitzerlandcElectron Microscopy ETH Zurich (EMEZ), ETdICN2 – Institut Catala de Nanociencia i N

08193, Barcelona, Spain, Institució Catalan

Barcelona, SpaineInstitució Catalana de Recerca i Estudis Ava

Universitat Autònoma de Barcelona, Campu

[email protected]

† Electronic supplementary information (and overplated Cu–Ni nanopillars (Fig. S1

Cite this: J. Mater. Chem. C, 2013, 1,7215

Received 9th July 2013Accepted 6th September 2013

DOI: 10.1039/c3tc31310g

www.rsc.org/MaterialsC

This journal is ª The Royal Society of

Ordered arrays of ferromagnetic, compositionallygraded Cu1�xNix alloy nanopillars prepared bytemplate-assisted electrodeposition†

Äıda Varea,a Salvador Pané,b Stephan Gerstl,c Muhammad A. Zeeshan,b

Berna Özkale,b Bradley J. Nelson,b Santiago Suri~nach,a Maria Dolors Baró,a

Josep Nogués,d Jordi Sort*e and Eva Pellicer*a

Periodic arrays of compositionally graded Cu–Ni alloy nanopillars (100 or 200 nm in diameter and 450 nm

in height) have been fabricated by means of potentiostatic electrodeposition into patterned Au/Ti/Si(111)

substrates using a single electrolytic solution. The pillars do not bend after the template removal but

remain straight and perfectly attached to the substrate. The average composition of the alloy

nanopillars can be tuned between 34 and 70 at% of Ni by varying the applied potential from a [Ni(II)]/

[Cu(II)] ¼ 95.2 electrolytic solution. The nanopillars are Ni-rich at the bottom with a compositionalgradient about �20–25% above/below the average composition. The magnetic characterization revealsthat all the nanopillars are ferromagnetic with coercivity values around 100–150 Oe (in-plane) and 200–

500 Oe (out-of-plane), which are larger than those for continuous films of similar average composition.

Atom probe tomography of the Cu-rich Cu–Ni nanopillars indicates the segregation of Ni, which would

explain their unexpected magnetic character. The graded-anisotropy ferromagnetic Cu–Ni pillars could

be potentially applied in several fields including micro/nano-electromechanical systems (MEMS/NEMS),

magnetic recording media or spintronics.

1 Introduction

Nanocrystalline (nc) metallic lms are known to benet fromenhanced physical properties (e.g. mechanical, magnetic andoptical) as compared to their coarse-grained counterparts.1 Oneof the key applications of nc-metallic lms is in the micro/nanoelectromechanical systems (MEMS/NEMS) eld.2 Tosuccessfully implement nc-metallic materials in MEMS/NEMSdevices, the miniaturization (i.e. nanostructuring) of a contin-uous nc thin lm is required. Such miniaturization is usuallyperformed by electron beam lithography (EBL) combined withsputtering, thermal evaporation or e-beam evaporation.3

However, when using EBL combined with li-off in narrow

ies, Universitat Autònoma de Barcelona,

mail: [email protected]

s (IRIS), ETH Zurich, CH-8092 Zurich,

H Zurich, CH-08093 Zurich, Switzerland

anotecnologia, Campus UAB, Bellaterra,

a de Recerca i Estudis Avançats (ICREA),

nçats (ICREA) and Departament de F́ısica,s UAB, E-08193 Bellaterra, Spain. E-mail:

ESI) available: SEM images of detached). See DOI: 10.1039/c3tc31310g

Chemistry 2013

pores, sputtering only allows thicknesses below a few tens ofnm.4 Another technique oen used to fabricate metallic MEMS/NEMS, is focused ion beam (FIB) lithography.5 However, thistechnique also suffers from some drawbacks: (i) it requires along time to manufacture individual samples, which signi-cantly reduces the output and (ii) Ga+ ions are easily incorpo-rated into the objects being fabricated, thus modifying theirchemical composition and structure and, in turn, their physicalproperties.5

To overcome all these issues, new cost-effective methodsable to miniaturize high aspect ratio nc-metallic lms at rela-tively fast rates are being explored. Electrodeposition is a non-vacuum, simple and well-known method for depositingmicrometer-thick lms of metals and alloys onto conductivesubstrates. The fabrication of large aspect ratio metallic nano-wires in the cylindrical pores of membranes such as anodizedalumina, polycarbonate and polymethyl methacrylate (PMMA)is well documented in the literature.6–9 The diameter of thenanowires obtained ranges from hundreds down to tens ofnanometres depending on the pore size of the template. Theapplication of these 1D nanostructures in sensors, eld-emis-sion and magnetic devices oen requires the nanostructures tobe supported by a substrate (Si, glass, etc.) in order to ensure therobustness and mechanical stability of the device. For thisreason, the nanowires are kept embedded in the template to

J. Mater. Chem. C, 2013, 1, 7215–7221 | 7215

http://dx.doi.org/10.1039/c3tc31310ghttp://pubs.rsc.org/en/journals/journal/TChttp://pubs.rsc.org/en/journals/journal/TC?issueid=TC001043

Fig. 1 Schematic representation of the layered structure of the substrates usedfor nanopillar electrodeposition (left) and the final structure (right) showing thenanopillar array after resist stripping.

Journal of Materials Chemistry C Paper

Publ

ishe

d on

09

Sept

embe

r 20

13. D

ownl

oade

d by

Uni

vers

itat A

uton

oma

de B

arce

lona

on

17/1

0/20

13 1

5:25

:38.

View Article Online

serve as high-performance platforms. However, in certainapplications, the template needs to be removed because it caneither interfere with or hinder the nanowires' response. In suchcases, when the template is etched away, the nanowires oenbend and partially (or completely) collapse. Thus, for manyapplications where the active material needs to be exposed,nanopillars (with intermediate length, lling the gap betweendots and nanowires) are more adequate than nanowires (withhigh aspect-ratio) due to their superior mechanical stability.Self-organized metallic ferromagnetic nanopillars have beenproposed not only as MEMS/NEMS counterparts but also asbuilding blocks in spintronics devices.10 Interestingly, there isalso an increasing interest in magnetic oxide nanowires, bothferrimagnetic11,12 and diluted magnetic semiconductors.13,14

Although disordered or pseudo-ordered nanowire arrays ofmany different elements and alloys have been grown by elec-trodeposition in anodic aluminium oxide (AAO) templates,15

research on the growth of nanopillars is far scarcer.16,17 Up tonow, a few pure metals (e.g. Co, In, Sn, Au and Cu) have beenprepared using EBL combined with electrodeposition.18–22

Examples of alloy nanopillars grown using this approach areeven more limited (NiFe is almost the only example available),23

mostly due to the difficulty in achieving the co-deposition ofmetals in nano-sized cavities. Compared to AAO membranes,e-beam lithographed substrates offer the possibility to preparefully ordered arrays of diverse nanostructure geometries on pre-dened substrate areas, enabling easier integration in NEMSand spintronics devices, as well as their implementation inhigh-density magnetic storage systems.

Regarding magnetic recording media, interest has beendirected in recent years toward the development of the so-called“graded anisotropy materials” as possible future magneticrecording media. These materials exhibit a tunable and gradualvariation of their magnetic anisotropy leading to a controllabledecrease of coercivity without a signicant loss of the thermalstability.24,25 Although the rst graded anisotropy materialsconsisted of Co/Pd multilayers with variable Co thickness,26

recently, compositionally graded FePtCu lms have also beendemonstrated.27 Remarkably, in spite of the numerous works oncontinuous lms with graded magnetic anisotropy, reports onpatterned graded anisotropic structures remain rather scarce.28

Thus, development of novel types of lithographed pillars withgraded magnetic anisotropy holds promise for the improve-ment of current state-of-the-art in patterned recording media.

Cu–Ni is a binary isomorphous alloy that can be electro-deposited from aqueous solutions.29 Alloying Cu with Ni at thenanoscale has been demonstrated to be a suitable way tosynergistically combine the properties of the two pure metals,thereby expanding the range of technological applications. Thishas been proven not only for nanocrystalline thin lms, wheretuneable ferromagnetic and mechanical properties have beendemonstrated as a function of the Cu/Ni ratio,30–32 but also fordisordered/pseudo-ordered nanostructures electrodeposited inAAO (e.g. nanoparticles and nanowires).33–35 In all these cases,the composition is kept constant along the thickness/length ofthe objects. Interestingly, the possibility of producing fullyordered arrays of graded-anisotropy ferromagnetic Cu–Ni

7216 | J. Mater. Chem. C, 2013, 1, 7215–7221

nanopillars on pre-dened substrate areas that can be poten-tially useful for a variety of applications (magnetic MEMS/NEMS, spintronics or recording media) has not been attempted.

In this work, a straightforward approach to obtain compo-sitionally graded Cu1�xNix nanopillars by electrodepositiononto e-beam lithographed silicon-based substrates is presented.Template-free, ordered and straight arrays of compositionallygraded nanopillars have been obtained in all cases aerremoving the e-beam resist used for patterning. The structuralcharacterization demonstrates the compositional gradientalong the length of the nanopillars. The obtained pillars, withdiameters of 100 nm (300 nm pitch size) and 200 nm (600 nmpitch size), and aspect ratios of 4.5 and 2.25, respectively,exhibit a tuneable magnetic response as a function of theaverage Cu/Ni ratio.

2 Results and discussion

Fig. 1 shows the layered structure of the substrates used forelectrodeposition (le image). The role of the upper 15 nm thickTi layer was to protect the Au seed-layer during the lithographyprocess and subsequent wafer slicing into single chips(substrates). This layer was selectively etched by a 0.1% HFsolution immediately before electrodeposition to ensure theadhesion of the nanopillars to the Au surface aer resistremoval; otherwise, the nanopillars become detached from thesubstrate due to the mismatch between the crystallographicstructure of Ti (hexagonal closed packed, P63/mmc spacialgroup) and that of Cu1�xNix structures (face-centred cubic,Fmm�3 spacial group) (Fig. S1a†). Nanopillars grown onto the Auseed layer show good adherence to the substrate since theircrystallographic structures are the same.

An electrolytic solution containing a [Ni(II)]/[Cu(II)] molarratio of 29 and a number of additives to control metal co-deposition and grain size were initially used for electrodeposi-tion. This bath was proven to be effective in achieving thedeposition of nanocrystalline, smooth, continuous Ni-rich Cu–Ni lms (up to 55 at% Cu).30 However, in contrast to the resultsin continuous lms, the composition of the nanopillars couldnot be tuned as a function of the applied potential/currentdensity using this electrolytic solution. Indeed, almost pureCu nanopillars were obtained regardless of the applied poten-tial/current density, likely because the potential inside the

This journal is ª The Royal Society of Chemistry 2013

http://dx.doi.org/10.1039/c3tc31310g

Table 1 Electrodeposition conditions for the deposition of pure Ni and Cu, andfor Cu–Ni nanopillars (concentration of Ni and Cu salts in solution, appliedpotential, time, stirring rate and temperature). For the Cu–Ni alloy, the Nipercentages correspond to the averaging of six EDX measures along the nano-pillar axis

Electrodeposition conditions

NiSO4(g L�1)

CuSO4(g L�1) E (V) t (s)

Pure Ni 630 0 �0.900 30 200 rpmT ¼ 30 �CPure Cu 624 6.24 �0.800 30

Cu–Ni alloy 34 at% Ni 624 6.24 �0.900 3041 at% Ni �1.000 3045 at% Ni �1.025 1563 at% Ni �1.100 770 at% Ni �1.150 5

Fig. 2 (a) Potentiostatic curves for different Cu1�xNix alloy nanopillar depositiononto patterned Au/Ti/Si substrates. (b) Averaged Ni percentage in the nanopillarsas a function of applied potential. The solid line is a guide to the eye.

Paper Journal of Materials Chemistry C

Publ

ishe

d on

09

Sept

embe

r 20

13. D

ownl

oade

d by

Uni

vers

itat A

uton

oma

de B

arce

lona

on

17/1

0/20

13 1

5:25

:38.

View Article Online

cylindrical cavities is not low enough to achieve the necessaryoverpotential for the reduction of Ni. To overcome this issue,the bath was re-formulated by increasing the [Ni(II)]/[Cu(II)]molar ratio in solution to 95.2, keeping the overall ionconcentration constant. The next step was to choose theappropriate window of deposition potentials so as to obtaindeposits with tuneable Cu/Ni ratios, and also to set the depo-sition time to avoid overplating (Fig. S1b†); otherwise, mush-room-like nanostructures are obtained upon stripping of theresin, as observed in other work36 (inset of Fig. S1b†). Note thatfor recording or magnetic sensor applications, overplating mustbe avoided because it can induce undesirable magnetic dipolarand exchange interactions between nanopillars.4 The optimizedparameters for the deposition of non-overplated structures(namely, the concentration of metal salts in solution, appliedpotential, deposition time, stirring rate and temperature) areshown in Table 1. The high [Ni(II)]/[Cu(II)] ratio was also selectedto induce signicant composition gradients along the alloynanopillar's axis.

The potentiostatic curves for Cu1�xNix deposition onto thelithographed substrates from the optimized bath are shown inFig. 2a. It can be observed that the absolute value of currentdensity increases as the applied potential is made increasinglynegative. The applied potentials (from E ¼ �0.900 V to E ¼�1.150 V) allowed growing nanopillars with varying composi-tion. The nickel percentage obtained from energy dispersive X-ray spectroscopy (EDXS) increased, on average, from 34 at% to70 at% as E was made more negative (Fig. 2b), since greateroverpotentials favour nickel discharge.

Pure Cu and Ni metals were also electrodeposited onto thelithographed substrates for comparison. Importantly, pure Cunanopillars could be obtained from the same bath formulationprovided that the applied potential was made more positive. Inparticular, for deposition potentials around �0.8 V, only Cu iselectrochemically reduced (Fig. 2b). To deposit pure Ni, though,the bath composition had to be modied since Ni deposition isalways accompanied by some Cu deposition even at high over-potentials. Therefore, the Cu sulphate in solution was replacedby Ni sulphate while keeping the overall ion concentrationconstant (Table 1).


A representative eld-emission scanning electron micros-copy (FESEM) planar view of non-overplated Cu1�xNix alloynanopillars is shown in Fig. 3a. 70� tilted (side-view) typicalFESEM images of 200 nm and 100 nm-diameter nanopillars aredisplayed in Fig. 3b and c, respectively. Both images show welldened, periodically ordered nanopillars with a constant heightof 450 nm. The diameter at the top is slightly smaller than at thebottom (near the substrate), so the structures are in fact hill-shaped. Interestingly, the Cu–Ni nanopillars are composition-ally graded along their long axis (Fig. 4). The EDX mappingsshow, within the lateral resolution (roughly around 50 nm), thatall the different alloyed nanopillars are Ni-rich at the bottomwith the Cu/Ni ratio progressively increasing towards the top,i.e., a compositional gradient (see Fig. 4). The degree of gradientdepends on the average Cu1�xNix composition. For example, forCu0.55Ni0.45, the gradient is moderate and the compositionranges from Cu0.49Ni0.51 at the bottom to Cu0.68Ni0.32 at the top,i.e., the Ni content increases by 19% between the top and thebottom (see Fig. 4b). On the other hand, for pillars with higherNi contents, e.g., Cu0.37Ni0.63, the gradient becomes consider-ably higher and can be as much as 30% between the top and thebottom (Fig. 4c). Considering that the molar concentration ofNi(II) is almost 100 times as that of Cu(II), the observed grada-tion indicates that the codeposition is strongly dependent on

J. Mater. Chem. C, 2013, 1, 7215–7221 | 7217


Fig. 3 FESEM images of Cu0.66Ni0.34 nanopillars: (a) planar view of 200 nmdiameter nanopillars, and (b and c) 70� tilted images for the 200 and 100 nmdiameter nanopillars, respectively.

Fig. 4 (a) FESEM image of two Cu0.55Ni0.45 200 nm-diameter nanopillars, (b) thecorresponding EDX mapping analysis along their length, where green colordenotes Cu and red color Ni. (c) FESEM image of a Cu0.37Ni0.63 200 nm-diameternanopillar, for which the Ni percentage at different locations along its length isindicated.

7218 | J. Mater. Chem. C, 2013, 1, 7215–7221


Publ

ishe

d on

09

Sept

embe

r 20

13. D

ownl

oade

d by

Uni

vers

itat A

uton

oma

de B

arce

lona

on

17/1

0/20

13 1

5:25

:38.

View Article Online

the mass transport conditions and the current distributioninside the cylindrical holes. Two factors may be the reason forfavoured deposition of nickel at the bottom of the pillar: (a)copper deposition is controlled by diffusion; (b) the conne-ment of the deposition event at the cavities of the template canlead to local increases of current density over the depositionarea. Once the growing alloy approaches the bulk solution, thedeposition of copper then becomes more favoured.

Magneto-optical Kerr effect (MOKE) measurements indicatethat the Cu1�xNix nanopillars (diameter of 100 nm) are ferro-magnetic at room temperature in the whole compositional

Fig. 5 Room-temperature MOKE hysteresis loops for: (a) Cu–Ni nanopillar arrayswith a diameter of 100 nm, measured in longitudinal configuration, (b) Cu–Ninanopillar arrays with a diameter of 100 nm, measured in polar configuration and (c)continuous (unpatterned) films, measured in longitudinal configuration. The aver-aged Cu–Ni compositions are indicated for the nanopillars. The behaviour of pure Ni isalso shown for the aim of comparison.



Fig. 6 (a) Atom map (Ni and Cu atoms are green and orange dots, respectively)of a local-electrode atom-probe (LEAP) tomography reconstruction of aCu0.55Ni0.45 nanopillar with volume dimensions in nm (actual dimensions are:38 nm � 38 nm � 26 nm). The isoconcentration surfaces, displaying 49 at% Cu(orange) and 50 at% Ni (green), reveal where local concentration gradients exist.The clustered regions are higher (>50 at%) in Ni. Concentrations within one suchNi rich region is shown in (c) along the 8 nm diameter cylinder depicted in (b),using 0.5 nm bin sizes along the length of the cylinder.


Publ

ishe

d on

09

Sept

embe

r 20

13. D

ownl

oade

d by

Uni

vers

itat A

uton

oma

de B

arce

lona

on

17/1

0/20

13 1

5:25

:38.

View Article Online

range. The hysteresis loops measured in longitudinal congu-ration (with the magnetic eld applied along the in-planedirection, i.e., perpendicular to the nanopillar axis) are rathersquare (except for the average Cu0.37Ni0.63 composition) andexhibit coercivity values, HC, around 100–150 Oe (Fig. 5a).Similar results were recorded for the nanopillars with a diam-eter of 200 nm. The occurrence of ferromagnetic behaviour inall the investigated Cu1�xNix nanopillars, particularly for theCu-rich ones, (e.g., Cu0.55Ni0.45 which is paramagnetic in bulkand thin lm form)37,38 is consistent with the presence of acompositional gradient.

Namely, even if the average Ni content is below the range forroom temperature magnetism in these alloys, the bottom of thenanopillars has sufficiently high Ni contents to make the pillarsferromagnetic. However, two ferromagnetic contributions seemto be present in the hysteresis loops corresponding to theCu0.30Ni0.70 and Cu0.37Ni0.63 nanopillars. In principle, if thecompositional gradient was continuous, exchange couplingshould render smooth hysteresis loops.27 Thus, the two-compo-nent loopsmay indicate in some alloys the segregation of variousnon-exchange coupled Ni-rich regions with different Ni contents(i.e., separated by non-magnetic regions). To elucidate the originof such an unexpected ferromagnetic response, some of theelectroplated nanopillars were prepared for atom probe tomog-raphy (APT) analysis in order to locally probe the distribution ofCu and Ni elements. APT is a very powerful technique althoughthe spatial resolution of atom probe tomography is slightlyanisotropic, meaning its resolution along the analysis axis (typi-cally the Z-axis) can be sub-nanometer and laterally (X and Ydirections) ca. 1 nm. As shown in Fig. 6a, the distribution of Cuand Ni elements at the top of Cu0.55Ni0.45 nanopillars is nothomogeneous, indeed, showing the presence of Ni-rich regionsembedded in a Cu-rich matrix. Moreover, the concentrationprole across the cylinder depicted in Fig. 6b shows that the Nicontent at the centre of it is around 57 at% (Fig. 6c), close to thethreshold for ferromagnetism. Since the Ni content progressivelyincreases toward the bottom of the pillar, one can straightfor-wardly assume that the probability of nding clusters with higherNi contents increases toward the bottom of the pillar, thuscausing the ferromagnetic response. Note that the occurrence ofNi-rich clusters could be understood bearing in mind that theenthalpy of mixing between Cu and Ni is slightly positive.39

Remarkably, square hysteresis loops are also obtained usinga polar MOKE (i.e., applying the magnetic eld along thenanopillars axis) for pure Ni nanopillars with a diameter of100 nm (Fig. 5b). In fact, the coercivity in this case (HC� 500 Oe)is much larger than for the longitudinal hysteresis loops. Suchan increase of HC is probably due to the role of shape anisot-ropy, which is particularly pronounced for the nanopillars witha smaller diameter (i.e., with a larger aspect ratio). Again, twoferromagnetic contributions are observed using a polar MOKEin the Cu0.37Ni0.63 and Cu0.30Ni0.70 nanopillars (Fig. 5b). Thecoercivity is larger for pure Ni than for Cu–Ni nanopillarsbecause the addition of Cu tends to progressively reduce themagnetocrystalline anisotropy.40

The coercivity values obtained from the nanopillars arelarger than those corresponding to Cu–Ni thin lms with the


same Cu/Ni ratio (Fig. 5c). The increase in HC can be ascribed tosize effects, i.e., either to the formation of single domain statesor the hindrances imposed by the limited lateral size of thenanopillars on the propagation of domain walls. Note that, asexpected, continuous lms with the Cu0.55Ni0.45 compositionare not ferromagnetic. While for continuous Cu–Ni lms thecoercivity tends to progressively decrease as the Cu contentincreases (basically due to the decrease of the magnetocrystal-line anisotropy), the evolution of coercivity with the Cupercentage in the Cu–Ni pillars is more complex. This can beascribed to the interplay between different effects. First, theformation of Ni-rich and Cu-rich clusters can give rise todifferent magnetic contributions to the hysteresis loops, oenexchange coupled to each other. The size of these clusters (andthe concomitant surface anisotropy) and the exchange, dipolarand Ruderman–Kittel–Kasuya–Yosida (RKKY) interactionsbetween them (similar to Cu/Ni multilayer nanowires41) caninuence the magnetic state of the cluster and their magneti-zation reversal, thus strongly affecting HC. For example, thelarge coercivity observed for Cu0.55Ni0.45 in Fig. 5a could be dueto the isolated ferromagnetic regions surrounded by a non-magnetic Cu-rich matrix, resulting in even larger coercivity thanfor pure Ni pillars. In addition, in the nanopillars, the magne-tocrystalline anisotropy competes with the shape anisotropy.Shape anisotropy would promote larger coercivity along thepillar axis (i.e., polar conguration). However, the formation ofnon-magnetic inclusions can have a negative effect on the shapeanisotropy, eventually resulting in small changes in coercivitywhen comparing polar with longitudinal hysteresis loops (e.g.,Cu0.37Ni0.63).

J. Mater. Chem. C, 2013, 1, 7215–7221 | 7219



Publ

ishe

d on

09

Sept

embe

r 20

13. D

ownl

oade

d by

Uni

vers

itat A

uton

oma

de B

arce

lona

on

17/1

0/20

13 1

5:25

:38.

View Article Online

3 Conclusions

Ordered arrays of Cu1�xNix (0.34 # x # 0.70) nanopillars, withdiameters of 100 nm and 200 nm and length of 450 nm, havebeen obtained by e-beam lithography combined with electro-deposition onto Au/Ti/Si(111) substrates. The structures remainadhered to the substrate and do not collapse aer resinremoval. The nanopillars are compositionally graded alongtheir height, as demonstrated by EDX analyses, and exhibittuneable ferromagnetic behaviour depending on the averageCu/Ni ratio. Interestingly, their coercivity is larger compared tocontinuous lms of the same composition. Atom probetomography of Cu-rich nanopillars reveals the presence of Ni-rich clusters that account for the observed ferromagnetism inspite of the large Cu content. Owing to their graded-anisotropyferromagnetic properties, the synthesized Cu–Ni nanopillararrays could nd applications in a number of elds where theactive material needs to be fully exposed (template-free) andmechanical stability is a prerequisite. Among them, magneticNEMS actuation and recording media are envisioned.

4 Experimental details4.1 Substrates preparation

First, Ti (100 nm)/Au (400 nm)/Ti (15 nm) were evaporated ontoSi(111) substrates. An electron sensitive resist (ZEP-520A) ofabout 500 nm in thickness was then spin-coated onto thesubstrate and patterning was carried out by electron-beamlithography (EBL) with the CRESTEC CABL-9500C Nano-lithography equipment. 60 � 60 mm2 arrays of cylindrical holesof 100 nm and 200 nm in diameter and pitch sizes of 300 nmand 600 nm, respectively, were patterned. Following the litho-graphic procedure, the resist was developed in anisol.

Before electrodeposition, all samples were submerged in HFsolution for 1 minute to remove the 15 nm-thick Ti layer to allowthe deposition of the metals directly onto the Au layer. Finally,the substrates were dipped in diluted sulphuric acid to removeoxides and organic residues. Before deposition, the substrateswere immersed in the electrolyte for 10 min, which had beenpreviously deaerated with argon gas.

4.2 Electrodeposition

Cu, Ni and the Cu–Ni alloy were deposited by direct currentelectrodeposition in a one-compartment thermostatized three-electrode cell using a PGSTAT302N Autolab potentiostat/galva-nostat (Ecochemie). The analytical solutions were preparedfrom analytical grade reagents and Millipore Milli-Q water(18 MU cm).

Cu and the CuNi alloy were deposited from a bath containing6.24 g L�1 CuSO4$5H2O, 624 g L

�1 NiSO4$6H2O and 70 g L�1

Na3C6H5O7$2H20 (sodium citrate), 0.2 g L�1 NaC12H25SO4

(sodium dodecylsulfate) and 1.4 g L�1 C7H5NO3S (saccharine).The electrolyte volume was 50 mL. The pH was xed at 4.5 andthe temperature at 30 �C in all cases. Pure Ni was depositedfrom an electrolyte containing 630 g L�1 NiSO4$6H2O (nickelsulphate) and the same amount of the other chemicals (sodiumcitrate, sodium dodecylsulfate and saccharine).

7220 | J. Mater. Chem. C, 2013, 1, 7215–7221

The Si(111)/Ti (100 nm)/Au (400 nm)/ZEP-520A layers wereused as working electrodes, which were positioned verticallyinto the electrode. A double junction Ag|AgCl (E ¼ +0.210 V perSHE) reference electrode was used with a 3 M KCl inner solutionand a 1 M Na2SO4 outer solution. All the potentials are referredto this electrode. A Pt sheet was used as the counter electrode.

Deposition was conducted potentiostatically, by applying aconstant potential in the range of �0.800 to �1.150 V, undermild stirring (200 rpm) using a magnetic stirrer bar. Aer thedeposition, the samples were thoroughly rinsed in water andthe ZEP-520A resin was removed with dimethylacetamide(DMAC) at 35 �C for 1 min to get template-free nanopillars.

4.3 Characterization

FESEM images of the nanopillars were obtained on a MERLINFESEMmicroscope operated at 5 kV. The chemical compositionwas determined by energy dispersive X-ray (EDX) spectroscopyat 7 kV. Metal proportions are expressed in atomic percentage(at%). Mapping analysis was carried out with the aim to clarifyelement distribution through the nanopillar length. Aspect-ratio values and nanopillar element distribution images weretaken at a 70� tilt angle.

The magnetic properties of nanopillars were investigated byacquiring hysteresis loops using a magneto-optical Kerr effectmagnetometer (MOKE) in longitudinal and polar modes(perpendicular and parallel to the long axis of the structure,respectively) with a maximum applied eld of 825 Oe at roomtemperature.

Atomprobe tomography of a single pillar was achieved using aCameca LEAP 4000-X HR. Controlled eld evaporation from thespecimen cooled to 33 K was instigated by applying 150 pJ laserenergy pulses (of 355 nm wavelength) at a pulse rate of 160 kHz,resulting a Ni++/Ni+ charge-state-ratio of 0.2. The resultingreconstruction and analysis were completed with IVAS 3.6.4.

Acknowledgements

Financial support from the MAT-2010-20616-C02 and MAT-2011-27380-C02-01 research projects and from ICTS-11/07(Infraestructuras Cient́ıcas y Tecnológicas Singulares) of theSpanish MINECO, and the 2009SGR-1292 from the Generalitatde Catalunya are acknowledged. M.D.B. acknowledges partialnancial support from an ICREA-Academia Award.

Notes and references

1 M. Aliohazraei and N. Ali, Two-Dimensional Nanostructures,CRC Press, Taylor & Francis Group, Boca Raton, Florida,USA, 2012.

2 MEMS/NEMS Handbook Techniques and Applications, ed. C. T.Leondes, Springer, New York, USA, 2006, vol. 1, pp. 219–225.

3 Y. D. Park, K. B. Jung, M. Overberg, D. Temple, S. J. Peartonand P. H. Holloway, J. Vac. Sci. Technol., B: Microelectron.Nanometer Struct.–Process., Meas., Phenom., 2000, 18, 16.

4 J. I. Mart́ın, J. Nogués, K. Liu, J. L. Vicent and I. K. Schuller,J. Magn. Magn. Mater., 2003, 256, 449.




Publ

ishe

d on

09

Sept

embe

r 20

13. D

ownl

oade

d by

Uni

vers

itat A

uton

oma

de B

arce

lona

on

17/1

0/20

13 1

5:25

:38.

View Article Online

5 M. D. Uchic and D. M. Dimiduk, Mater. Sci. Eng., A, 2005,400–401, 268.

6 M. A. Zeeshan, K. Shou, S. Pané, E. Pellicer, J. Sort,K. M. Sivaraman, M. D. Baró and B. J. Nelson,Nanotechnology, 2011, 22, 275713.

7 J. Qin, J. Nogués, M. Mikhaylova, A. Roig, J. S. Mu~noz andM. Muhammed, Chem. Mater., 2005, 17, 1829.

8 M. Tian, J. Wang, J. Krutz, T. E. Mallouk and M. H. W. Chan,Nano Lett., 2003, 3, 919.

9 M. L. Curry, K. Crews, V. Warke, M. G. Bakker, K. Hong,J. Mays, P. Britt, X. Li and J. Wang, J. Vac. Sci. Technol., A,2001, 29, 031401.

10 I. Žutić, J. Fabian and S. D. Sarma, Rev. Mod. Phys., 2004, 76,323.

11 L. Suber, P. Imperatori, G. Ausanio, F. Fabbri andH. Hofmeister, J. Phys. Chem. B, 2005, 109, 7103.

12 K. S. Shankar, S. Kar, A. K. Raychaudhuri andG. N. Subbanna, Appl. Phys. Lett., 2004, 84, 993.

13 G. Z. Xing, J. B. Yi, J. G. Tao, T. Liu, L. M. Wong, Z. Zhang,G. P. Li, S. J. Wang, J. Ding, T. C. Sum, C. H. A. Huan andT. Wu, Adv. Mater., 2008, 20, 3521.

14 D. Wang, Q. Chen, G. Xing, J. Yi, S. R. Bakaul, J. Ding,J. Wang and T. Wu, Nano Lett., 2012, 12, 3994.

15 G. Kartopu and O. Yalçin, in Fabrication and Applications ofMetal Nanowire Arrays Electrodeposited in Ordered PorousTemplates, Electrodeposited Nanowires and theirApplications, ed. N. Lupu, InTech, Vukovar, Croatia, 2010.

16 A. Santos, L. Vojkuvka, J. Pallarés, J. Ferré-Borrull andL. F. Marsal, Nanoscale Res. Lett., 2009, 4, 1021.

17 L. G. Vivas, Y. P. Ivanov, D. G. Trabada, M. P. Proenca,O. Chubykalo-Fesenko and M. Vázquez, Nanotechnology,2013, 24, 105703.

18 S. Jin, M. J. Burek, R. D. Evans, Z. Jahed and M. C. Leung,Mater. Sci. Eng., A, 2012, 552, 104.

19 G. Lee, J.-Y. Kim, A. S. Budiman, N. Tamura, M. Kunz,K. Chen, M. J. Burek, J. R. Greer and T. Y. Tsui, ActaMater., 2010, 58, 1361.

20 M. J. Burek, A. S. Budiman, Z. Jahed, N. Tamura, M. Kunz,S. Jin, S. M. J. Han, G. Lee, C. Zamecnik and T. Y. Tsui,Mater. Sci. Eng., A, 2011, 528, 5822.

21 M. J. Burek and J. R. Greer, Nano Lett., 2010, 10, 69.22 A. T. Jennings and J. R. Greer, Philos. Mag., 2011, 91,

1108.


23 M. Zenger, W. Breuer, M. Zö, R. Pulwey, J. Raabe andD. Weiss, IEEE Trans. Magn., 2001, 37, 2094.

24 D. Suess, Appl. Phys. Lett., 2006, 89, 113105.25 R. Skomski, T. A. George and D. J. Sellmyer, J. Appl. Phys.,

2008, 103, 07F531.26 B. J. Kirby, J. E. Davies, K. Liu, S. M. Watson, G. T. Zimanyi,

R. D. Shull, P. A. Kienzle and J. A. Borchers, Phys. Rev. B:Condens. Matter Mater. Phys., 2010, 81, 100405.

27 C. L. Zha, R. K. Dumas, Y. Y. Fang, V. Bonanni, J. Nogués andJ. Åkerman, Appl. Phys. Lett., 2010, 97, 182504.

28 H. Wang, H. Zhao, T. Rahman, Y. Isowaki, Y. Kamata,T. Maeda, H. Hieda, A. Kikitsu and J.-P. Wang, IEEE Trans.Magn., 2013, 49, 707.

29 A. Brenner, Electrodeposition of alloys, Academic Press, NewYork/London, vol. 1, 1963.

30 E. Pellicer, A. Varea, S. Pané, B. J. Nelson, E. Menéndez,M. Estrader, S. Suri~nach, M. D. Baró, J. Nogués and J. Sort,Adv. Funct. Mater., 2010, 20, 983.

31 E. Pellicer, A. Varea, K. M. Sivaraman, S. Pané, S. Suri~nach,M. D. Baró, J. Nogués, B. J. Nelson and J. Sort, ACS Appl.Mater. Interfaces, 2011, 3, 2265.

32 A. Varea, E. Pellicer, S. Pané, B. J. Nelson, S. Suri~nach,M. D. Baró and J. Sort, Int. J. Electrochem. Sci., 2012, 7, 1288.

33 A. Foyet, A. Hauser and W. Schäfer, J. Solid StateElectrochem., 2008, 12, 47.

34 Z. Liu, G. Xia, F. Zhu, S. Kim, N. Markovic, C.-L. Chien andP. C. Searson, J. Appl. Phys., 2008, 103, 064313.

35 X.-K. Tian, X.-Y. Zhao, L.-D. Zhang, C. Yang, Z.-B. Pi andS.-X. Zhang, Nanotechnology, 2008, 19, 215711.

36 X. Jin, Y. Hu, Y. Wang, R. Shen, Y. Ye, L. Wu and S. Wang,Appl. Surf. Sci., 2012, 258, 2977.

37 J. C. Ododo and B. R. Coles, J. Phys. F: Met. Phys., 1977, 7,2393.

38 S. A. Ahern, M. J. C. Martin and W. Sucksmith, Proc. R. Soc.London, Ser. A, 1958, 248, 145.

39 I. A. Abrikosov and H. L. Skriver, Phys. Rev. B: Condens.Matter Mater. Phys., 1993, 47, 16532.

40 J. G. Booth, H. P. J. Wijn and G. Zibold, in Landolt-Börnstein,Numerical Data and Functional Relationships in Science andTechnology, ed. B Subvol and H. P. J. Wijn, Springer-Verlag,Berlin, Germany, vol. 19, 1993.

41 L. Sun, Y. Hao, C. L. Chien and P. C. Searson, IBM J. Res. Dev.,2005, 49, 79.

J. Mater. Chem. C, 2013, 1, 7215–7221 | 7221


Ordered arrays of ferromagnetic, compositionally graded Cu1tnqh_x2212xNix alloy nanopillars prepared by template-assisted electrodepositionElectronic...Ordered arrays of ferromagnetic, compositionally graded Cu1tnqh_x2212xNix alloy nanopillars prepared by template-assisted electrodepositionElectronic...Ordered arrays of ferromagnetic, compositionally graded Cu1tnqh_x2212xNix alloy nanopillars prepared by template-assisted electrodepositionElectronic...Ordered arrays of ferromagnetic, compositionally graded Cu1tnqh_x2212xNix alloy nanopillars prepared by template-assisted electrodepositionElectronic...Ordered arrays of ferromagnetic, compositionally graded Cu1tnqh_x2212xNix alloy nanopillars prepared by template-assisted electrodepositionElectronic...Ordered arrays of ferromagnetic, compositionally graded Cu1tnqh_x2212xNix alloy nanopillars prepared by template-assisted electrodepositionElectronic...Ordered arrays of ferromagnetic, compositionally graded Cu1tnqh_x2212xNix alloy nanopillars prepared by template-assisted electrodepositionElectronic...Ordered arrays of ferromagnetic, compositionally graded Cu1tnqh_x2212xNix alloy nanopillars prepared by template-assisted electrodepositionElectronic...

Ordered arrays of ferromagnetic, compositionally graded Cu1tnqh_x2212xNix alloy nanopillars prepared by template-assisted electrodepositionElectronic...

Documents

Journal of Materials Chemistry C · Periodic arrays of compositionally graded Cu–Ni alloy nanopillars (100 or 200 nm in diameter and 450 nm ... Atom probe tomography of the Cu-rich