This content has been downloaded from IOPscience. Please scroll down to see the full text.

Download details:

IP Address: 150.244.102.216

This content was downloaded on 15/12/2016 at 10:15

Please note that terms and conditions apply.

Spin configuration in isolated FeCoCu nanowires modulated in diameter

View the table of contents for this issue, or go to the journal homepage for more

2015 Nanotechnology 26 395702

(http://iopscience.iop.org/0957-4484/26/39/395702)

Home Search Collections Journals About Contact us My IOPscience

You may also be interested in:

Electrochemical synthesis and magnetic characterization of periodically modulated Co nanowires

I Minguez-Bacho, S Rodriguez-López, M Vázquez et al.

Vortex domain wall propagation in periodically modulated diameter FeCoCu nanowire as determined by

the magneto-optical Kerr effect

Ester M Palmero, Cristina Bran, Rafael P del Real et al.

Correlation between structure and magnetic properties in CoxFe100x nanowires: the roles of

composition and wire diameter

C Bran, E M Palmero, Zi-An Li et al.

Enhanced magnetocrystalline anisotropy of Fe30Co70 nanowires by Cu additives and annealing

Ester M Palmero, Ruslan Salikhov, Ulf Wiedwald et al.

Crystallographically driven magnetic behaviour of arrays of monocrystalline Co nanowires

Yu P Ivanov, D G Trabada, A Chuvilin et al.

Magnetic hardening of Fe30Co70 nanowires

Sara Liébana Viñas, Ruslan Salikhov, Cristina Bran et al.

Tuning the magnetic properties of multisegmented Ni/Cu electrodeposited nanowires with controllable

Ni lengths

M Susano, M P Proenca, S Moraes et al.

Magnetic reversal modes in cylindrical nanowires

Yu P Ivanov, M Vázquez and O Chubykalo-Fesenko

Spin configuration in isolated FeCoCunanowires modulated in diameter

Óscar Iglesias-Freire1,2, Cristina Bran1, Eider Berganza1,Ignacio Mínguez-Bacho1,3, César Magén4, Manuel Vázquez1 andAgustina Asenjo1

1 Instituto de Ciencia de Materiales de Madrid (ICMM-CSIC), 28049, Madrid, Spain2McGill University, H3A 2T8, Montreal, Canada3Nanyang Technological University, 637371, Singapore4 Instituto de Nanociencia de Aragón (INA), Universidad de Zaragoza 50018, Zaragoza, Spain

E-mail: oiglesias@physics.mcgill.ca

Received 27 May 2015, revised 2 August 2015Accepted for publication 13 August 2015Published 11 September 2015

AbstractCylindrical Fe28Co67Cu5 nanowires modulated in diameter between 22 and 35 nm aresynthesized by electroplating into the nanopores of alumina membranes. High-sensitivity MFMimaging (with a detection noise of 1 μNm−1) reveals the presence of single-domain structures inremanence with strong contrast at the ends of the nanowires, as well as at the transition regionswhere the diameter is modulated. Micromagnetic simulations suggest that curling of themagnetization takes place at these transition sites, extending over 10–20 nm and giving rise tostray fields measurable with our MFM. An additional weaker contrast is imaged, which isinterpreted to arise from inhomogeneities in the nanowire diameter.

S Online supplementary data available from stacks.iop.org/NANO/26/395702/mmedia

Keywords: nanowires, domain wall, magnetic force microscopy, magnetic materials, spintronics

1. Introduction

The determination of the spin configuration of nanoscalemagnetic systems and the control over their magnetizationreversal process is essential for progress in advanced informa-tion storage, logic systems, biomedical applications, cleanenergy sources and sensing devices [1–4]. Controlled dis-placement of magnetic domain walls (DWs) along ferromag-netic nanowires by means of electric currents [1, 5] or magneticfield pulses [6] has recently advanced to become one of themost intensively studied topics in magnetism [7]. In planarnanostrips fabricated by nanolithography, DWs are artificiallypinned by notches or local stray fields [8] but less attention waspaid to cylindrical nanowires (NWs) until the discovery of thenew opportunities they offer to control the shape and dynamicsof the DWs for spintronic application [9]. Cylindrical NWs arecurrently synthesized by electroplating into nanopores [10] withthe possibility to modulate their diameter [11], inducing localpinning centers that tailor the reversal mechanism [12]. On theother hand, concerns about resource scarcity and the search forrenewable and clean energy sources have led to novel families

of permanent magnets. In particular, CoFe-based NWs areamong the most optimal candidates because they exhibit highsaturation magnetization and good thermal stability [10].

The objective of this work is twofold. A first achievementis the fabrication of modulated NWs with smaller diametersthan previously reported (down to 22 nm). In addition, theresulting spin configuration in individual NWs has beenimaged and correlated to their morphologies. Note that fewexperimental techniques allow for direct imaging of spinconfigurations in this kind of nanostructure; among them,magnetic force microscopy (MFM) is currently the mostwidespread and accessible one for this purpose [13–15].

2. Methods

Fe28Co67Cu5 NWs were grown by electrodeposition onto thepores of anodic aluminum oxide (AAO) membranes [16],obtained after a two-step anodization process [17] (figure 1).The geometry of the nanochannels is controlled not only bythe anodization time, voltage or bath temperature but also by

Nanotechnology

Nanotechnology 26 (2015) 395702 (7pp) doi:10.1088/0957-4484/26/39/395702

0957-4484/15/395702+07$33.00 © 2015 IOP Publishing Ltd Printed in the UK1

the shape of the pulses in the second anodization step. Theresulting cylindrical modulated pores are formed by segmentswith diameters of around 22 and 35 nm, while the center-to-center inter-wire distance is kept constant at 65 nm [11]. NWswere subsequently grown into the modulated pores by elec-trodeposition from a sulfate-based electrolyte [17], with afinal length of around 7 μm.

Aluminum foils (99.999%; Goodfellow) are electro-polished in a mixture 1:4 of HClO4:C2H5OH at 20 V undervigorous stirring. Aluminum foils are anodized in sulfuricacid–based electrolyte with 3 wt% concentration. The ano-dization process is performed within a Teflon electro-chemical cell refrigerated at the base and the walls, so thatboth the aluminum foil and the electrolyte are directlycooled. This system provides high thermal stability at 0 °Cduring the processes. Prior to the pulsed anodization pro-cess, a nanopatterned aluminum foil is obtained afterremoval of the AAO resulting from 16 h of a conventionalfirst anodization at 25 V (figure 1). In order to improve themechanical stability and avoid the breakdown of the ano-dized film during the first hard anodization (HA) pulse, theduration of the first mild anodization (MA) pulse wasextended to 15 min, followed by alternating MA and HApulses. The sequence of the anodization pulses appliedduring the process is 150 s for MA segments and 200 ms forHA segments, repeated 10 times. The resulting length of thenanochannels is about 40 μm. The resulting cylindricalmodulated pores are formed by segments with diameters of22 and 35 nm for MA and HA respectively, while the center-to-center inter-pore distance (within the membrane) is keptconstant at 65 nm. After the anodization process, the alu-minum substrate is removed with an aqueous solution

composed of copper chloride and hydrochloric acid and thealumina barrier layer is etched with an aqueous solution ofH3PO4, 5 wt%. A gold layer is sputtered onto one side of thealumina template to ensure the good conductivity needed forelectrochemical deposition.

Cylindrical Fe28Co67Cu5 nanowires around 7 μm inlength were subsequently electroplated into the cylindricalmodulated pores by electrodeposition from a sulfate-basedelectrolyte containing CoSO4·7H2O (35 g l−1), CuSO4·5H2O(2 g l−1), FeSO4·7H2O (15 g l−1), H3BO3 (10 g l−1) andascorbic acid (10 g l−1). The resulting nanowires were thenperiodically modulated in diameter between 22 and 35 nm.The electrochemical deposition was carried out under poten-tiostatic conditions by applying −1.8 V to a Ag/AgClreference electrode, keeping the temperature of the electrolyteat 25 °C and the pH value at about 3.0. As a result, a mac-roscopic membrane enclosing a very large number of ferro-magnetic nanowires is obtained. The alloy composition ofthe electrodeposited samples was identified by means ofenergy dispersive x-ray spectroscopy (EDS). The membranewas then divided into different pieces so that bulk magne-tometry measurements were performed, on the one hand,whereas dissolution of the alumina and subsequent deposi-tion onto a Si substrate (figure 1) was followed up by TEMand MFM analysis of single Fe28Co67Cu5 elements, on theother hand.

3. Results and discussion

Hysteresis loops were obtained at RT with the NWs inside themembrane by means of vibrating sample magnetometry

Figure 1. The fabrication process. Schematics of the main steps followed during the fabrication of the FeCoCu nanowires and the preparationfor the subsequent MFM experiments.

2

Nanotechnology 26 (2015) 395702 Ó Iglesias-Freire et al

(VSM), with a maximum magnetic field of ±1.7 T appliedeither along the longitudinal or transverse directions. Analysisof the longitudinal loop (figure 2(a)) suggests the existence ofrelatively large magnetostatic interactions between neigh-boring elements, as expected from the small inter-pore dis-tance [18]. As a result, reduced values for the susceptibilityand remanent magnetization (about 41% of the saturationvalue) are obtained.

Upon dissolution of the AAO membrane, high-resolutiontransmission electron microscopy (HR-TEM) characterizationwas performed to evaluate the crystallinity and the structuralproperties of the NWs [19]. Figure 3(a) shows the modulationin diameter of a single NW, with magnified images of seg-ments about 35 nm and 22 nm in diameter presented infigures 2(b) and (c), respectively. The mean segment length ofconstant diameter was found to be 500–600 nm. Regardingcrystallinity, the nanostructures present a bcc crystallinestructure with the wider parts being highly textured along the[110] direction, whereas the narrow segments show weakertexture. A surface amorphous layer with a thickness of2–3 nm is observed throughout the NW length, in agreementwith previous studies [20] on uniform FeCoCu NWs, andoriginated by the mismatch between the NW crystallinestructure and the alumina during the growth process. A smallsurface roughness of a few nm is observed at both interfacesof the amorphous layer; its effect on the magnetic propertieswill be analyzed and recalled later.

In order to gain information about the magnetic behaviorof individual NWs, MFM imaging was performed with aNanotec Electrónica system making use of the amplitudemodulation (AM) and two-pass modes; in addition, a phase-locked loop (PLL) was enabled to track the resonancefrequency of the oscillating cantilever (k≈3 Nm−1,

f0≈75 kHz). The existence of eventual artifacts generated bythe overlap of electrostatic interactions [21] can be ruled out,as deduced from Kelvin probe force microscopy experiments(not shown); similarly, effects associated with irreversible tip-sample influence [22] were not relevant. MFM was imple-mented at low scan speed, allowing the lock-in amplifier tofilter out the majority of noise sources and yielding highresolution [23] and enhanced signal-to-noise ratio with afrequency shift noise level of 20 mHz. This value corre-sponds, assuming a point-mass model and the harmonicapproximation, to a force gradient detection limit of only1 μNm−1 in ambient conditions.

Figures 4(a) and (b) show the topography and its corre-sponding MFM image of several modulated NWs in theirremanent state. Similar MFM images showing the as-grownand demagnetized states can be found in the supplementaryinformation (SI.1). The interpretation of the MFM image israther difficult since several series of spots with relativelystrong contrast are found at different positions along thewires. At first sight, this suggests the existence of complexconfigurations involving several DWs as every NW displays anumber of attractively and repulsively interacting regions(dark and bright contrast, respectively). A deeper insight intothe underlying configurations can be obtained with the help ofmicromagnetic simulations. These were performed using anobject oriented micromagnetic framework (OOMMF) tosimulate the magnetic behavior of individual NWs. A(2×2×2) nm3 cubic cell size was considered, with asaturation magnetization MS=1.59·106 A m−1 andexchange stiffness A=10.7·10−12 J m−1. As previouslyreported [24], the weak magnetocrystalline anisotropy wasassumed not to play a significant role.

Figure 2. (a) Hysteresis loops of NWs embedded within the membrane. (b) Simulated loops of individual NWs with either constant(d=22 nm or 36 nm) or modulated diameters (dmin=22 nm and dmax=36 nm).

3

Nanotechnology 26 (2015) 395702 Ó Iglesias-Freire et al

Simulated hysteresis loops are shown in figure 2(b) forthe case of a longitudinal field applied to three differentNWs (2 μm in length) simulated independently. Two ofthem were designed with a uniform diameter of 22 nm(narrow) and 36 nm (wide), the third one having a modu-lated diameter with two narrow and two wide segments(500 nm long; see figure 5(c)). For each simulation, theinitial distribution of the magnetization was set to beoriented along the main axis of the NW (initial saturated

state) and its evolution with an external field was studied inthe range (1 T, −1 T), including a step at zero field(remanent state). Additionally (refer to the supplementaryinformation (SI) available at stacks.iop.org/NANO/26/395702/mmedia), the demagnetized state was simulated bystarting from a randomly oriented distribution of spins andstudying its evolution without the presence of any externalfield (SI.1), revealing the formation of transversal domainwalls during the process (shown in SI.2).

Figure 4. (a) Topographic and (b) MFM images (25×25) μm2 in size of modulated NWs in remanence deposited onto a Si substrate.

Figure 3. HR-TEM images of a modulated NW, where (a) a modulation in the diameter is clearly observed. (b) Area highlighted with a redsquare in (a), showing a bcc high textured region along the [110] direction. (c) Area framed in blue in (a), revealing a weaker texture.

4

Nanotechnology 26 (2015) 395702 Ó Iglesias-Freire et al

For each of the three simulated NWs, square hysteresisloops were obtained with remanent magnetizations up to 98%of the saturation value (inset in figure 2(b)), as expected fromsingle-domain wires [24]. The reversal process in uniformwires is mediated by two vortex walls nucleating at both endsand propagating inwards [26], this propagation taking place atlower field values in the case of wide NWs. In the samemanner, reversal in the modulated NW is triggered by thevortex wall nucleated in the wider segment, which explainsthe similarities of the hysteresis loops between the modulatedand 36 nm uniform wires (figure 2(b)). The slightly lowervalue of the magnetization observed in the modulated struc-ture right before the Barkhausen jump is associated withmagnetization curling taking place at the discontinuities [27].For these small transverse dimensions, Bloch-point DWs arenot expected to be energetically favorable [9, 28].

Comparison with the experiment can be made by con-sidering an individual NW as the one shown in figure 5. In the

MFM image (figure 5(b)), two types of magnetic contrast canbe identified: (i) relatively intense spots not only present at theends of the wires but also at specific positions along thenanostructures, and (ii) a weaker alternating contrast observedall along the element.

The topographic profile of the NW has been highlightedby white dashed lines in figures 5(a)–(b) to help correlate themagnetic contrast with the morphology. From there, one canreadily identify the strong contrast to be originated at dis-continuities in the diameter, in excellent agreement with thesimulated MFM image depicted in figure 5(c). Zooming into atransition region between segments (figure 5(d)) reveals thatthe orientation of the magnetization in remanence is essen-tially oriented along the longitudinal axis of the NWs.Therefore, the origin of the strong divergence is purely shape-induced at the modulation sites with the aim to minimize themagnetostatic energy. As a result, magnetization curling takesplace (figure 5(e)) in narrow regions that extend over

Figure 5. (a) Topography and (b) MFM images (2.7×0.4 μm) of a single NW in remanence. (c) Simulated configuration for a modulatedNW equivalent to the experimental case. A single domain state is obtained with large divergence of the magnetization at the discontinuities.The black–white scale stands for a negative–positive divergence of the magnetization, analogous to the MFM case. (d) Zoom-in on the regionframed in blue in (c). (e) Cross section view of the area enclosed by the dashed square line in (d).

5

Nanotechnology 26 (2015) 395702 Ó Iglesias-Freire et al

10–20 nm into the wider segments of the NW. These transi-tion regions might eventually be used to modify the speed of apropagating DWs in this kind of nanostructure [29, 30]. Moreexamples of single NWs showing more complex configura-tions in demagnetized states, including DWs determined to betransverse-like by simulations, are analyzed in SI.2.

In addition, weak alternating spots not predicted bysimulations are present along the NW (figure 5(b)). This con-trast becomes noticeable due to the enhanced resolution andsensitivity of the MFM data shown in this work. Analogouslyto the former case, a general correlation can be identifiedbetween subtle topographic protuberances and the weak con-trast in the MFM signal. As was previously mentioned in theHR-TEM analysis (figure 3), the outermost layers of the NWspresent a corrugation on the order of few nm; therefore, weattribute this quasi-periodic weak contrast to morphologicaldeviations from a smooth cylindrical surface. Further infor-mation on this can be found in SI.3, where MFM data onstraight NWs of the same composition but constant diameterare presented, displaying single-domain behavior characterizedby strong contrast at the ends and a similar series of weakspots. These results should encourage MFM users to be par-ticularly careful with the interpretation of data in this kind ofsample as similar effects should be observed in narrow nano-strips or NWs if substantial corrugation and enough sensitivityis reached during the measurement [31, 32].

4. Conclusions

This work presents the fabrication of cylindrical Fe28Co67Cu5NWs with modulated diameter segments as thin as 22 and35 nm. The effect those modulations have on the spin orien-tation has been unveiled by high-sensitivity MFM imagingand supported by micromagnetic simulations. The complexcontrast in the resulting images requires a careful non-trivialinterpretation that suggests the single-domain state is favor-able in remanence, with a curling of the magnetization at thetransition regions between segments of different diameter. Inaddition, a weaker contrast was correlated to a surface cor-rugation of few nm caused by the mismatch of the crystallinestructure of the NWs and the AAO nanopores.

Acknowledgments

The authors acknowledge support from the Spanish MINECOunder projects CSD2010-00024 and MAT2013-48054-C2-1-R and from the EU commission under project ‘REFREE-PERMAG’ FP7NMP.2011.2.2.

Supporting information

- MFM images of Fe28Co67Cu5 NWs in the as grown anddemagnetized states.

- Domain walls in demagnetized individual NWs.- MFM images of non-modulated Fe28Co67Cu5 NWs.

References

[1] Parkin S S P, Hayashi M and Thomas L 2008 Fabrication ofconstricted compositionally-modulated Nix Fe1−xnanowires Science 320 190–4

[2] Allwood D A, Xiong G, Faulkner C C, Atkinson D, Petit D andCowburn R P 2005 Magnetic domain-wall logic Science 3091688–92

[3] Safi M, Yan M, Guedeau-Boudeville M A, Conjeaud H,Garnier-Thibaud V, Boggetto N, Baeza-Squiban A,Niedergang F, Averbeck D and Berret J F 2011 Interactionsbetween magnetic nanowires and living cells: uptake,toxicity, and degradation ACS Nano 5 5354–64

[4] Gutfleisch O, Willard M A, Brück E, Chen C H,Sankar S G and Liu J P 2011 Magnetic materials and devicesfor the 21st century: stronger, lighter, and more energyefficient Adv. Mater. 23 821–42

[5] Saitoh E, Miyajima H, Yamaoka T and Tatara G 2004 Nature4 203

[6] Kim J-S, Mawass M-A, Bising A, Krüger B, Reeve R M andSchulz T 2014 Synchronous precessional motion of multipledomain walls in a ferromagnetic nanowire by perpendicularfield pulses Nat. Commun. 5 3429

[7] Stamps R L et al 2014 The 2014 magnetism roadmap J. Phys.D: Appl. Phys. 47 333001

[8] Franken J H, van der Heijden M A J, Lavrijsen T H, Ellis R,Daniels C, McGrouther D, Swagten H J M and Koopmans B2014 Beam-induced Fe nanopillars as tunable domain-wallpinning sites Adv. Funct. Mater. 24 3508–14

[9] Piao H-G, Shim J-H, Djuhana D and Dong-Hyun K D-H2013 Intrinsic pinning behavior and propagation onset ofthree-dimensional Bloch-point domain wall in acylindrical ferromagnetic nanowire Appl. Phys. Lett. 102112405

[10] Nielsch K, Mueller F, Li A P and Goesele U 2000 Uniformnickel deposition into ordered alumina pores by pulsedelectrodeposition Adv. Mater. 12 582

[11] Minguez-Bacho I, Rodriguez-López S, Vázquez M,Hernández-Vélez M and Nielsch K 2014 Electrochemicalsynthesis and magnetic characterization of periodicallymodulated Co nanowires Nanotechnology 25 145301

[12] Allende S, Altbir D and Nielsch K 2009 Magnetic cylindricalnanowires with single modulated diameter Phys. Rev. B 80174402

[13] Pitzschel K, Bachmann J, Martens S, Montero-Moreno J M,Kimling J, Meier G, Escrig J, Nielsch K and Görlitz D 2011Magnetic reversal of cylindrical nickel nanowires withmodulated diameters J. Appl. Phys. 109 033907

[14] Shinjo T, Okuno T, Hassdorf R, Shigeto K and Ono T 2000Magnetic vortex core observation in circular dots ofpermalloy Science 289 930

[15] Jaafar M, Yanes R, Asenjo A, Chubykalo-Fesenko O,Vazquez M, Gonzalez E M and Vicent J L 2008 Fieldinduced vortex dynamics in magnetic Ni nanotrianglesNanotechnology 19 285717

[16] Martin C R 1994 Nanomaterials: a membrane-based syntheticapproach Science 266 1961–6

[17] Bran C, Ivanov Y P, Garcia J, del Real R P, Prida V M,Chubykalo-Fesenko O and Vazquez M 2013 Tuning themagnetization reversal process of FeCoCu nanowire arraysby thermal annealing J. Appl. Phys. 114 043908

[18] Lim J-H, Rotaru A, Min S-G, Malkinski L and Wiley J B 2010Synthesis of mild–hard AAO templates for studyingmagnetic interactions between metal nanowires J. Mater.Chem. 20 9246–52

[19] Ivanov Y P, Iglesias-Freire O, Pustovalov E V,Chubykalo-Fesenko O and Asenjo A 2013 Magnetic

6

Nanotechnology 26 (2015) 395702 Ó Iglesias-Freire et al

configurations of Co(111) nanostripes with competing shapeand crystalline anisotropies Phys. Rev. B 87 184410

[20] Rodríguez-González B, Bran C, Warnatz T, Rivas J andVazquez M 2014 Structural and magnetic characterization ofas-prepared and annealed FeCoCu nanowire arrays inordered anodic aluminum oxide templates J. Appl. Phys. 115133904

[21] Jaafar M, Iglesias-Freire O, Serrano-Ramón L, Ibarra M R,de Teresa J M and Asenjo A 2011 Distinguishing magneticand electrostatic interactions by a Kelvin probe forcemicroscopy–magnetic force microscopy combinationBeilstein J. Nanotechnol. 2 552–60

[22] Iglesias-Freire Ó, Bates J R, Miyahara Y, Asenjo A andGrütter P H 2013 Tip-induced artifacts in magnetic forcemicroscopy images Appl. Phys. Lett. 102 022417

[23] Moya C, Iglesias-Freire Ó, Pérez N, Batlle X, Labarta A andAsenjo A 2015 Direct imaging of the magnetic polarity andreversal mechanism in individual Fe3−xO4 nanoparticlesNanoscale 7 8110–4

[24] Bran C, Ivanov Y P, Trabada D G, Tomkowicz J, del Real R P,Chubykalo-Fesenko O and Vazquez M 2013 Structuraldependence of magnetic properties in Co-based nanowires:experiments and micromagnetic simulations IEEE Trans.Magn. 49 4491–7

[25] Jaafar M, Serrano-Ramón L, Iglesias-Freire O,Fernández-Pacheco A, Ibarra M R, de Teresa J M andAsenjo A 2011 Hysteresis loops of individual Co

nanostripes measured by magnetic force microscopyNanoscale Res. Lett. 6 407

[26] Vivas L G, Yanes R, Chubykalo-Fesenko O and Vázquez M2011 Coercivity of ordered arrays of magnetic Conanowires with controlled variable lengths Appl. Phys. Lett.98 232507

[27] Salem M S, Sergelius P, Corona R M, Escrig J, Görlitz D andNielsch K 2013 Magnetic properties of cylindrical diametermodulated Ni80Fe20 nanowires: interaction and coercivefields Nanoscale 5 3941–7

[28] Da Col S et al 2014 Observation of Bloch-point domain wallsin cylindrical magnetic nanowires Phys. Rev. B 89180405 (R)

[29] Yan M, Kákay A, Gliga S and Hertel R 2010 Beating theWalker limit with massless domain walls in cylindricalnanowires Phys. Rev. Lett. 104 057201

[30] Burn D M and Atkinson D 2013 Suppression of Walkerbreakdown in magnetic domain wall propagation throughstructural control of spin wave emission Appl. Phys. Lett.102 242414

[31] Iglesias-Freire Ó, Jaafar M, Pérez L, de Abril O,Vázquez M and Asenjo A 2014 Domain configuration andmagnetization switching in arrays of permalloy nanostripesJ. Magn. Magn. Mater. 355 152–7

[32] Dennis C L et al 2002 The defining length scales ofmesomagnetism: a review J. Phys.: Condens. Matter 14R1175

7

Nanotechnology 26 (2015) 395702 Ó Iglesias-Freire et al