Fe nanowires

Fe is one of the most common elements on Earth because of its massive production by star nuclei as a consequence of fusion reactions. The abundance of this element has resulted in its extensive use in many applications since ancient times. In principle, Fe has been used in metallurgy; though pure Fe can be softer and more ductile than aluminum, the presence of different impurities can increase the hardness of metallic Fe by a factor of 1000, as in the case of carbon added to produce steel. However, the main disadvantage of Fe lies in its oxidation in the presence of oxygen or water, forming a flaky oxide layer that is easily peeled off to expose the Fe to the environment. Then Fe and Fe alloys are seriously damaged by corrosion if their surfaces are not passivated or coated with a protective layer. The role of Fe in biological processes where the Fe complexes with oxygen to form metalloproteins such as hemoglobin and myoglobin or redox reactions in plants and animals is also worth mentioning. Finally, from a magnetic point of view, Fe is widely used in magnetic devices because of its high saturation magnetization and magnetic susceptibility. One of the main magnetic applications of Fe over the years has been the production of electromagnet cores.

Industrial production of Fe starts with the extraction of Fe ores generally based on iron oxides. Such ores are processed in a blast furnace to obtain pig iron, which contains an elevated concentration of several impurities such as carbon, sulfur, and silicon. Then, this metal is submitted to several thermal treatments to produce different kind of Fe, steels, and compounds. The evolution of new technologies and the interest in small-dimension nanomaterials, however, have promoted the fabrication of Fe materials by means of physical vapor deposition, chemical vapor deposition, or electrochemical methods, among others. The possibility of electroplating Fe together with template-assisted methods has attracted attention for the fabrication of one-dimensional, Fe-based nanowires. However, the electrochemical conditions strongly affect the morphologic, structural, and magnetic properties of electroplated Fe (Izaki, 2010).

The most extended baths used in Fe electrodeposition are based on Fe chlorides, sulfates, or a mixture of two salts (Table 2.1). Although in those baths Fe(II) (ferrous) and Fe(III) (ferric) states are present, the latter causes a reduction of the cathode efficiency, which results in stressed and pitted deposits. In general, Fe(III) formation or precipitation can be controlled by the addition of different organic acids such as citric, tartaric, or ascorbic acid, or by adjusting the pH of the bath. The electrodeposition from sulfate-based baths results in smooth and homogeneous deposits but at slow deposition rates. In addition, such electrodeposition can be done at room temperature, maintaining the good properties of the deposit. By contrast, chloride-based electrolytes produce, in general, highly stressed and rough deposits that are reduced as the temperature of the electrodeposition process is increased.

A critical parameter that has to be controlled is the stability of the electrolyte, especially with large concentrations of Fe(III) that tend to form Fe(III) hydroxides precipitating at pH values around 3.5. Because changes in pH have a strong impact on the electrolyte, buffer salts such as boric acid (H₃BO₃) are commonly used to maintain a stable pH. Moreover, Fe oxidation can be avoided by bubbling inert gases such as nitrogen or argon to push away the oxygen ions present in the electrolyte.

From the crystalline point of view, the most stable phase at room temperature is the BCC, which is able to transit from a BCC to an FCC at high temperatures (> 900 °C). Such crystallization in a cubic crystal has a strong impact on the magnetic properties since it entails a reduction in the magnetocrystalline anisotropy of the material. This fact indicates that the magnetic properties of the material would be dominated by the shape anisotropy, which also increases as the saturation magnetization increases. In the case of nanomaterial science, one of the aims of reducing sizes is the achievement of high aspect ratios and, particularly in nanomagnetism, high and uniaxial-shaped magnetic anisotropies, which, in the case of Fe, are also enhanced by their high saturation magnetization.

Several studies of the magnetic properties of Fe nanowire arrays in porous anodic alumina templates have been performed over the past three decades (Yang et al., 2000; Zhang et al., 2003; Borissov et al., 2009). Because the shape anisotropy is supposed to be dominant, those studies focused on the effects of the geometric properties of nanowires (e.g., diameter, length) and nanowire arrays (e.g., porosity, homogeneity) over the magnetic properties (Peng et al., 2000; Sorop et al., 2004; Ou et al., 2009). These works cover nanowire diameters from 5 to 70 nm with a controlled nanowire length, ensuring in all cases a high aspect ratio. High uniaxial magnetic anisotropy with pore diameters smaller than 35 nm has been found, with the easy axis lying along the parallel direction with respect to the nanowire axis, as is predicted by shape anisotropy. Moreover, such nanowire arrays display a coercive field between 2 and 2.6 kOe, as can be seen in Figure 2.2, where the data of several nanowire arrays is collected (Sellmyer et al., 2001 and references therein).

Qin et al. (2012) reported a reduction in the coercive field and an apparent reduction in the uniaxial character of the magnetic anisotropy as the nanowire diameter is increased above 30 nm. Although the easy magnetization axis still lies parallel to the nanowires, the magnetostatic interactions strongly affect the shape of the hysteresis loops (HLs) made on nanowire arrays with 50- and 70-nm nanowire diameters (Figure 2.3). Since such interaction strongly depends on the distance between nanowires and on their saturation magnetization, the effect of the magnetostatic interaction among them has to be taken into account, especially with the high saturation magnetization value of Fe. Moreover, this interaction is negative because of the tendency to align the nanowire’s magnetization in an antiparallel configuration, which is able to reduce the coercivity and reduced remanence if it is strong enough.

Finally, although the BCC crystalline structure typical of Fe has a low magnetocrystalline anisotropy, the magnetic properties of Fe nanowires can be tuned by controlling the textured growth of nanowires (Hu et al., 2005). By means of dynamically changing the pH of the solution during electrodeposition, the preferred orientation [110] along the nanowire can be substituted by a [200] orientation, which enhances the easy magnetization axis parallel to the nanowires.

Cobalt nanowires

For regular Co at ambient pressure, the HCP crystal phase is the more stable one at temperatures below 422 °C, whereas the FCC crystal phase is the main stable phase at temperatures above 422 °C (Yoo et al., 2000). In electroplated films and nanowires, the FCC phase can be obtained from electroplating at ambient temperature and pressure using an electrolyte with a low pH (< 2.5). The low pH of the electrolyte enhances the evolution of hydrogen gas and the Co ion mobility during the electroplating process. The enhanced atomic mobility may be equivalent to a high temperature condition so that the FCC phase can be obtained (Cohen-Hyams et al., 2002). Nevertheless, the coexistence of both Co crystal phases has been reported in different works (Ye et al., 2009; Sánchez-Barriga et al., 2009). Also, one can expect a transition from an HCP to an FCC phase for nanowires with diameters ≥ 30 nm (Kitakami et al., 1997). Other investigations of Co nanoparticles show that the presence of an HCP phase or an FCC phase can be related to the particle diameter (Sánchez-Barriga et al., 2009). In this case, the transition from HCP to FCC was ascribed to a size effect caused by the lower surface energy of the FCC phase. However, electrodeposited Co (under standard conditions) usually presents the HCP crystallographic structure, although the FCC crystal structure can also be obtained by using particular deposition parameters, such as low pH and/or high temperature or deposition rate, as we have been discussing. Let us come back to the effect of the pH in the fabrication process through the discussion of the selected results in this section.

The nanowires described in this section were synthesized via direct current (DC) electroplating into pore channels of self-standing oxalic acid NAMs prepared by the two-step anodization process (Vivas et al., 2012b). Electrodeposition was carried out using an aqueous electrolyte containing 250 g L^− 1 cobalt(II) sulfate (CoSO₄) and 40 g L^− 1 H₃BO₃ at room temperature, under constant stirring, at a bias voltage of − 1 V (vs. Ag/AgCl). The prepared solution has a typical pH value of 3.5, which was gradually increased to 6.7 by the addition of diluted sodium hydroxide. The wires grew in the usual bottom-up fashion, starting from the gold electrode at the pore bottoms, and the average wire length (L) was controlled by monitoring the total charge that passed during deposition. Scanning electron microscopy (SEM; Figure 2.4a) allowed us to determine the pore diameter (50 nm), the interpore distance (105 nm), and the length of the wires (3 μm). Also, X-ray diffraction (XRD) patterns in Figure 2.4b reveal the HCP phase in all the cases, showing strong textures along the [100], [101], and [002] directions, depending on the pH value (3.5, 5.0, and 6.7, respectively; Vivas et al., 2012b). These textured structures at 2θ = 41.5°, 47.5°, and 44.5° are ascribed to hexagonal crystalline anisotropy with the c-axis nearly perpendicular, at ~ 45°, and parallel to the wires axis, respectively (Maaz et al., 2010).

Magnetic HLs for the three cases are shown in Figure 2.4c. A complex anisotropy distribution is suggested for the [100] texture (pH 3.5; Figure 2.4b), where there is a strong competition between magnetostatic (shape) and magnetocrystalline anisotropies to make the effective magnetic anisotropy to become parallel and/or perpendicular to the nanowires main axis. On the other hand, parallel magnetic anisotropy is deduced for the samples textured in the [101] direction (pH 5.0; Figure 2.4b), that is, the easy magnetic axis may be identified as the one parallel to the wires, in spite of the nonzero values of the coercivity (H_c) and reduced remanence (m_r = M_r/M_sat, where M_r is the remanence and M_sat is the saturation magnetization) when the magnetic field was applied perpendicular to the nanowires axis. Finally, the nanowires textured in the [002] crystallographic direction (pH 6.7; Figure 2.4b) are determined by a strong uniaxial magnetic anisotropy. These results clearly illustrate that the synthesis process can be used to tailor the crystalline structure to manipulate the magnetism of the Co nanowire arrays. In particular, by modifying the acidity of the electrolytic bath during the electrodeposition of Co nanowires, it is possible to induce different Co HCP phase textures, which result in different magnetic behaviors.

The measured coercivity and remanence values decrease with an increase in the nanowire aspect ratio in Co nanowires electrodeposited from a Co electrolithic bath with a typical pH of 4. Vivas et al. (2012c) provide more details of the fabrication process of these samples. This behavior contrasts with the magnetism expected from the shape anisotropy contribution. The experimental evolution of coercivity with nanowire length is interpreted considering micromagnetic simulations in the next section, taking into account the change in their crystalline structure.

Based on SEM characterization, the geometry of the nanowire arrays was determined as pore diameter D_p of 35 nm; L ranges from 120 to 1000 nm, and an interpore distance D_int of 105 nm in the hexagonal arrangement. XRD patterns for the samples with L = 120 and 1000 nm are presented in Figure 2.5a. The intense peak at 2θ = 44.7° arises from the [200] FCC phase of the underlying aluminum foil, which has a strong signal in the samples from which it is not removed, such as the ones under discussion. For shorter nanowires, only a broadening on the left side of the aluminum peak is visible. This suggests the presence of the diffraction peak from the [111] FCC crystallographic phase at 2θ = 44.2°. Nevertheless, this phase could overlap with the [002] HCP at 2θ = 44.5°. This overlapping complex diffraction pattern is observed up to L = 350 nm, which indicates that the electrodeposited Co nanowires could be composed of either single FCC or both FCC and HCP crystalline phases. In particular, this fact agrees with the observation reported by Pirota et al. (2011), who detected by structural and magnetic characterization an FCC phase at the bottom of Co electrodeposited nanowires up to a length of 270 nm. Whether there is any crystal texture presents in the samples, it can be estimated by using a standard analytical calculation; therefore, for longer wires, as in the present samples, there is a texture in the [100] direction of the HCP phase, which increases with the nanowires length.

Figure 2.5b and c shows the trends followed by the coercivity and the remanence at room temperature. In contrast to what is expected from the simple effect of modifying the shape anisotropy, one can see a decrease in these two magnetic properties as a function of the length, which can be attributed to the change in the magnetocrystalline anisotropy rather than the shape anisotropy (see the increase of the [100] HCP phase with L in Figure 2.5a; Vivas et al., 2011).

To confirm using different approaches the important role that the crystalline structure plays in the magnetic properties of these kinds of systems, and looking for the actual magnetization reversal process, micromagnetic simulations are presented in this section. Because of the simulation time, this study is restricted to nanowires with lengths that range from 120 to 1000 nm, with typical wire diameters ranging from 35 to 75 nm. When the calculations are for nanowire arrays, the interwire distance was set to 105 nm, which is the typical distance for oxalic membranes (Schlörb et al., 2010; Vázquez and Vivas, 2011; Vivas et al., 2012b). The micromagnetic simulations presented here were performed using the MAGPAR package with finite element discretization (Scholz et al., 2003). The external magnetic field was applied in the direction of the nanowire axis. For the simulations, the typical magnetic parameters of bulk Co were considered (Ferré et al., 1997; Jamet et al., 2000).

Figure 2.6 shows part of the results from the demagnetization process simulated for seven individual nanowires with varying aspect ratios ordered in a hexagonal array. The simulations were for nanowires with either a single FCC or HCP phase. In the latter case, the c-axis was oriented perpendicular to the nanowire axis, considering that the Co samples obtained from the experiments exhibit an increase in the [100] HCP phase with increasing L. At the remanent state, the magnetic structure for co-FCC nanowires is a single domain with magnetization parallel to the applied field direction in the middle and two open vortex structures at both ends of the wires. Then, magnetization reversal occurs through vortex domain wall propagation (Figure 2.6a). In the case of nanowires with an HCP phase, the nucleation process also starts with the creation of vortices at the two ends with the same or opposite chiralities. The remanence is characterized by the nucleus magnetized along the nanowire length and a circularly magnetized shell. The corresponding remanence is, therefore, smaller than that of FCC nanowires. The coercivity is determined by the propagation of complex domain walls (Figure 2.6b). Thus, the modeling also shows the remanence decrease with the nanowire length (inset in Figure 2.6c) in qualitative agreement with the experimental results in Figure 2.5b and c. Finally, Figure 2.6d presents the coercivity values obtained from the experiments and the simulations. Both the FCC and HCP curves for the individual nanowires modeled follow a similar trend, but the second one has reduced values, as is expected from the strong perpendicular anisotropy effect. When magnetostatic interactions among nanowires are considered, the coercivity decreases considerably. However, the experimental behavior of H_c and m_r with L is just the opposite, namely, a reduction of both parameters is observed. The combined crystallinity/geometry effect can be understood by considering that short nanowires are predominantly FCC, where the shape anisotropy dominates the total effective anisotropy. On the other hand, for longer Co HCP nanowires, the transverse magnetocrystalline anisotropic easy axis competes with the longitudinal shape anisotropy. Furthermore, from both the experimental results and the calculations, the FCC phase may be the only one present up to approximately 350 nm. The mix curve in Figure 2.6d represents the results obtained from a model, where a percentage of randomly distributed HCP grains increases monotonically with the increase in L of the nanowires. This model provides coercive field values more in accordance with experimental ones. An additional remaining discrepancy between the experimental and modeled values may be attributed to other contributions such as the possible presence of an HCP [002] peak, which favors the partial parallel alignment of the c-axis with the nanowires and thus the reinforcement of the uniaxial anisotropy; to the presence of defects and fluctuations of wire diameters and lengths; and to additional dipolar interactions coming from an infinite array.

Another geometric parameter, which can be modified to investigate the evolution of the crystal structure, is the diameter of the nanowires. In a different study, where the magnetic properties of arrays of Co nanopillars with the same length (L = 120 nm) but different diameters (from 35 to 75 nm), both experiments and modelling showed crucial changes in HLs when D was increased (Vivas et al., 2013). Although this magnetic behavior could be related to changes in the shape anisotropy, the primary role of the magnetocrystalline anisotropy, as well as the importance of the magnetostatic interactions among nanopillars, was demonstrated. Vivas et al. also showed the change in the reversal mode with the increased diameter from vortex propagation to curling, when the field is applied parallel to the nanopillar axis, and from coherent rotation to curling, when it is applied perpendicular to this axis.

Even though this discussion has been focused in nanowires arrays, important results have been obtained for individual Co nanowires. Ivanov et al. (2013) reported the magnetic and structural properties of an individual 40-nm diameter Co nanowire prepared by electrodeposition onto anodic alumina templates. Using HR-TEM and XRD, the monocrystalline HCP structure along the whole 10-μm nanowire length with the c-axis almost perpendicular to the nanowire axis was revealed. This observation enabled the understanding of the magnetic properties of the nanowire. The magnetic state observed by magnetic force microscopy and modelled by micromagnetic simulations was interpreted as consisting of tilted vortices with alternating chiralities.

Through all these studies, it can be established that the magnetic properties of electrodeposited Co nanowire are strongly dependent on the crystalline structure present. Furthermore, through manipulation of the synthesis parameters and the geometry of the wires, the magnetocristalline anisotropy can be tailored, which could create desired characteristics for a huge range of applications.

Nickel nanowires

Ni is a transition metal with good resistance against corrosion, which makes it interesting for the production of Ni steels and coatings that offer protection against oxidation. Furthermore, it is an element that shows ferromagnetism at room temperature and has been used, in combination with other elements such as Alnico alloy, to create permanent magnets. Moreover, the abundance of Ni in the form of metal and soluble salts has increased the importance of this element in the electroplating industry. In particular, the low magnetocrystalline anisotropy exhibited by the FCC crystalline structure of Ni offers the possibility to control the magnetic properties of the material by tuning its shape. In the case of magnetic nanowire array growth by electrodeposition into the channels of porous templates, it is easy to modify the characteristics of the magnetization reversal process by changing either the diameter or length of the nanowires. Since these arrangements are closely packed, the magnetostatic interactions among nanowires become an important fact to take into account because they greatly affects the collective reversal magnetization process. The low saturation magnetization of Ni, in comparison with Co and Fe, reduces the strength of such interactions, which favors the application of such close-packed magnetic systems in magnetic storage media. Finally, Ni shows extensive and negative magnetostriction, which is of special interest in nanoactuator and sensor technologies.

Nowadays the most common bath used in the electrodeposition of Ni is called the Watts-type electrolyte, although there exist plenty of approaches, depending on the type of deposit required (Di Bari, 2010; Table 2.2). In general, this type of electrolyte gives rise to electrodepositions with high current density and good-quality homogeneous deposits. Although the crystalline structure of Ni is always FCC, single or polycrystalline Ni nanowires can be produced by controlling the electrochemical conditions (Pan et al., 2005). Moreover, single-crystal Ni nanowires are produced by high electrodeposition potentials and low temperatures, whereas polycrystalline structures are obtained in other ways. Single-crystal Ni nanowires have a larger coercive field and are more square and thus have a higher uniaxial anisotropy.

In general, the coercive field and reduced remanence of Ni nanowires are determined by its shape, which means that the easy magnetization axis commonly lies parallel to the nanowire. However, the spatial arrangement of nanowires affects the global magnetic behavior of all nanowires since they tend to be aligned in an antiparallel configuration because of the magnetostatic interactions among them. Such interactions increase when the diameter of the nanowires, and thus the saturation magnetization, increases despite the reduction of the distance between them. Avoiding any contribution from magnetocrystalline anisotropy, ensuring that each nanowire can be a single magnetic domain, makes this system of special interest when studying the magnetic interactions in nanowire arrays for their application in high-density magnetic storage technologies (Nielsch et al., 2001, 2002; Vázquez et al., 2004a,b, 2005; Samardak et al., 2012).

Co–Ni alloys can exhibit both either a soft or a hard magnetic behavior, depending on the amount of Co in the alloy (Talapatra et al., 2009; Vivas et al., 2012c). The room temperature uniaxial magnetocrystalline anisotropy constant of bulk Co–Ni alloys (K₁) varies from negative values (− 5 × 10⁴ erg cm^− 3) for pure Ni to positive ones (2 × 10⁴ erg cm^− 3) with the addition of a moderate amount of Co (8–22 at% of Co), whereas higher Co content leads to larger negative values (− 2 × 10⁵ erg cm^− 3; Kadowaki and Takahashi, 1975; Fan and Piron, 1996), thus allowing for the design of a material with tunable magnetic properties. This fact, together with the high solubility of Co atoms in the crystalline lattice of Ni (and vice versa) for a wide range of relative concentrations (Cheng and Huang, 2008), have drawn renewed interest in the fabrication and magnetic characterization of Co–Ni alloy nanowire arrays (Rheem et al., 2007; Cojocaru et al., 2010; Vivas et al., 2012b,c; Vega et al., 2012a,b,c; Ali and Maqbool, 2013; Wen et al., 2013a,b). In these nanostructured systems, the competition between shape and magnetocrystalline anisotropies, together with the magnetostatic dipolar interactions among nanowires, determine the effective magnetic anisotropy energy, which can favor a preferred magnetization direction (easy magnetization axis) of the system either along the longitudinal or perpendicular directions with respect to the nanowire axis (Vila et al., 2009; Pirota et al., 2011).

Co–Ni nanowire arrays with a diameter of 130 nm, a distance of 305 nm between adjacent nanowires, and a length of 15–30 μm, covering a wide range of compositions from pure Ni to Co₉₅Ni₅, were reported by Vega et al. (2012a,b,c). The Co–Ni nanowire arrays were electrodeposited into hard anodized alumina templates, which were previously functionalized with an SiO₂ shell by ALD. The electrodepositions were performed in Watts-type baths with changing amounts of Co and Ni ions in the electrolytes, which allowed the amount of Co in the alloy to be tuned by varying either the electrolyte composition or the electrodeposition potential, as indicated by the energy-dispersive X-ray (EDX) analysis of SEM cross-sectional images (Figure 2.7a). The crystalline structure, analyzed by XRD, reveals a change from an FCC to an HCP structure as the Co content of the alloyed nanowires is increased, as depicted in Figure 2.7b. At intermediate compositions, for example, Co₇₁Ni₂₉, the coexistence of both FCC and HCP crystalline phases was detected. Figure 2.8 (left column) displays room temperature HLs measured for Co–Ni nanowire arrays with different compositions along both the longitudinal () and radial (⊥) directions with respect to the nanowires axis, while the column on the right shows longitudinal HLs of single nanowires, measured by magneto-optic Kerr effect magnetometry. These results indicate that both Ni-rich single nanowires and nanowire arrays display a magnetic behavior dominated by a well-defined longitudinal shape anisotropy, showing typical bistable HLs for single nanowires that are tilted in the case of arrays of nanowires by magnetostatic interactions among adjacent wires (Zighem et al., 2011), thus decreasing the reduced remanence value. This tilting effect is more noticeable as the Co content of the alloy is increased. The increase of the saturation magnetization value of Co–Ni alloys with the addition of Co atoms, as shown in the Slater-Pauling curve (Bozorth, 2003) also causes an increase in the demagnetizing field due to magnetostatic interactions, which is responsible for diminishing the squareness value from almost 100% in single nanowires to around 60% for Ni nanowire arrays and down to 10% for Co₅₁Ni₄₉ nanowire arrays. However, Co-richer nanowires and nanowire arrays display HLs characterized by a progressive change in magnetization as the magnetic field is reversed, indicating a delocalized magnetization reversal associated with the polycrystalline structure of the nanowires and the high anisotropy constant of the HCP phase (K_hcp = 4.1 × 10⁶ erg cm^− 3), which competes with shape anisotropy (K_Sh ~ 6 × 10⁶ erg cm^− 3; Sánchez-Barriga et al., 2009). This competition between shape and magnetocrystalline anisotropies is also evident from the broad anisotropy field distribution obtained for Co₇₁Ni₂₉ nanowire arrays.

The dependence of the coercivity on the alloy Co content for both Co–Ni nanowire arrays and isolated nanowires, obtained from vibrating sample magnetometer and magneto-optic Kerr effect HLs, respectively, is depicted in Figure 2.9. In the composition range from pure Ni to Co₅₁Ni₄₉, the H_c value decreases as the Co content increases for both single nanowires (from 404 to 241 Oe) and nanowire arrays (from 391 to 180 Oe); the Co₅₁Ni₄₉ sample is optimal for exhibiting softer magnetic properties.

By directly comparing the values of H_c for Co–Ni nanowire arrays with their respective counterparts for single isolated nanowires, it can be observed that the latter have larger values than the former for all the measured nanowires in the range of 0 ≤ x ≤ 62, in good agreement with previous works (Escrig et al., 2008), which, based on mcromagnetic simulations, demonstrated that the magnetostatic interactions among nanowires in the array are responsible for such a decrease in H_c. The perpendicular coercive field values (H_c^⊥) of nanowire arrays show a small decrease from 71 to 44 Oe with increasing Co content in the same compositional range, whereas for higher Co contents it increases up to values around 200 Oe, close to those of H_c, due to the competition between the shape and HCP phase magnetocrystalline anisotropies of Co-rich nanowires.

Vivas et al. (2012c) studied pure Co, Co₈₀Ni₂₀, and Co₅₀Ni₅₀ nanowires deposited into anodic alumina membranes with the following geometric parameters: interwire distance D_int ~ 105 nm, nanowire diameter D_p ~ 35 nm, and nanowire length L ~ 2.5 μm. In the XRD θ–2θ scans, displayed in Figure 2.10a, one can observe only the [100] peak from the Co HCP phase for Co nanowires, whereas the presence of mixed FCC-HCP is observed with the addition of Ni. Also, the presence of the Co HCP [100] peak decreases significantly and instead, two peaks are observed around 44.5°; these are more intense for Co₅₀Ni₅₀ (indicating a textured structure; Vivas et al., 2012c) than for Co₈₀Ni₂₀. This behavior suggests that the crystal structure of the nanowires has evolved from an HCP to an FCC phase.

Figure 2.10b and c shows the room temperature HLs of these samples for parallel () and perpendicular (⊥) applied field configurations. Based on a comparative analysis, a certain effective magnetic anisotropy in the nanowires axis can be assumed for Co nanowires, since both H_c and m_r have slightly larger values than the corresponding perpendicular ones. That reflects a balance between magnetic anisotropies. While the textured [100] HCP phase of Co leads to a magnetocrystalline anisotropy with a perpendicular magnetic easy axis (Vázquez and Vivas, 2011), the strong effect of the shape anisotropy results in the preferential orientation of the magnetization parallel to the nanowires. These observations are in agreement with other results reported for polycrystalline Co nanowires (Vivas et al., 2012b,c; Rheem et al., 2007; Vázquez and Vivas, 2011). With the addition of Ni, a reinforcement of the parallel easy magnetization axis is observed, reflected in the increase of H_c and m_r seen in Figure 2.10b. However, an overall increase in H_c^⊥ is also visible. In this case, the sample’s polycrystallinity and the presence of both HCP and FCC phases may lead to local crystal inhomogeneities, causing the increased irreversibility observed in the perpendicular HL.

Similar crystallographic results have been observed for Co–Ni nanowire arrays deposited in anodic alumina membranes prepared in sulfuric acid-based electrolytes, with D_int = 65 nm and D_p = ~ 35 nm. However, a reduction of both magnetic parameters, H_c and m_r, may occur when comparing the oxalic membrane samples (D_int = 105 nm). The ratio of D_p to D_int is higher in the sulfuric samples; hence a reduction in the effective magnetic anisotropy of the system is expected because of the reinforcement in the magnetostatic interactions among nanowires, which is in agreement with several reports from this field.

The study of the angular dependence of magnetic properties, in particular coercivity (H_c) and remanence (M_r), gives us a way to obtain information about the actual magnetization reversal mechanism. Escrig et al. (2008) calculated the angular dependence of the coercivity for three idealized modes of the magnetization reversal process: coherent rotation mode, in which all the spins rotate simultaneously; the curling wall mode, in which spins reverse progressively via propagation of a vortex-like (curling) domain wall; and the transverse wall mode, in which spins reverse progressively via propagation of a transverse domain wall. Although all three modes are possible in principle, the corresponding system will reverse its magnetization by whichever mode opens an energetically accessible route first, that is, by the mode that offers the lowest coercivity. Figure 2.11a illustrates the results obtained for a sample of Co₅₀Ni₅₀ nanowires. In this calculation, an anisotropy constant value of 4.89 ⋅ 10³ J m^− 3 was used to obtain good agreement with the experiment, reflecting an axial easy axis, as deduced from the experimental values. In this case, it was concluded that the magnetization reversal is driven by the nucleation and propagation of a transverse wall, independent of the angle between the magnetization and the applied magnetic field. Actually, the same result is obtained for the other samples in this study (data not shown), although for Co₅₀Ni₅₀ nanowires the curling and the transverse reversal modes showed a similar coercivity: ~ 60°. Regarding the lack of perfect fitting for angles close to 0° and 90°, it is important to remember the limitations of the model, as well as the fact that it does not consider magnetostatic interactions with surrounding wires. These interactions may be responsible for an additional antiferromagnetic or ferromagnetic contribution in fields applied at 0° and 90° to the wires axis, respectively. Figure 2.11b collects the values of the crystalline anisotropy constant used in the calculations to obtain the best agreement when fitting with experimental results. This figure shows that the easy magnetization axis evolves from perpendicular (in-plane of the membrane) in the case of Co nanowires to longitudinal (parallel to the wires axis) in the case of Ni-rich alloy samples.

The temperature dependence of m_r and H_c for several arrays of Co–Ni nanowires with selected compositions (D_p ~ 35 nm, D_int ~ 65 nm) is presented in the left and right panels of Figure 2.12, respectively. The evolution of these magnetic parameters is depicted for both a and ⊥ configuration of the applied field. In the case of Co₂₅Ni₇₅ nanowires (Figure 2.12a), corresponding to a Ni-rich sample, an effective longitudinal easy axis in the whole range of investigated temperatures is deduced, where both reduced remanence and coercivity take higher values in the parallel applied field configuration than in the perpendicular one. Increasing the Co content to Co₅₀Ni₅₀ (Figure 2.12b) shows that the parallel magnetic parameters still have the highest values through the changes of temperature, except at room temperature where m_r has about the same value for both and ⊥ applied magnetic field configurations. For Co-rich Co₇₅Ni₂₅ nanowires (Figure 2.12c), a crossover effect from parallel to perpendicular reduced remanence is observed. Also, the perpendicular coercivity values are larger than those corresponding to the Co₅₀Ni₅₀ case. However, lower values are observed for H_c when comparing again with this previous sample. Finally, for Co nanowires (Figure 2.12d), a behavior similar to that for the m_r parameter is observed, whereas a crossover effect in H_c is detected, decreasing the temperature. Below 100°K, the effective anisotropy of this last system clearly lies perpendicular to the nanowires long axis.

As a summary of the comparison between parallel and perpendicular remanence and coercivity for each composition, we deduce that (i) for high Co content, there is an evolution toward perpendicular anisotropy as temperature decreases and (ii) for high enough Ni content, the parallel anisotropy is well defined through the whole temperature range. Nevertheless, a crossing temperature is observed when there is the same ratio of Ni and Co contents.

For Co–Ni nanowires, the monotonous increase in longitudinal coercivity with Ni content is ascribed to the decrease of the HCP crystal phase, as was discussed earlier. It should be mentioned that for high Ni content, a significant contribution of magnetoelastic anisotropy may be expected because of the magnetoelastic coupling between a higher magnetostriction and the mechanical stress induced by the different thermal expansion coefficients of Ni and the NAM (for these experiments the aluminum foil had been removed; Schlörb et al., 2010). On the other hand, in the case of Co-rich nanowires, the change in the crystalline anisotropy constant with temperature is expected to be more relevant. In addition, the contribution from changes in the shape anisotropy caused by variation in saturation magnetization seems to be relatively modest since the effect is observed well below the Curie temperature.

Single Co–Ni alloy nanowires have also been studied in a broad compositional range for the characterization of their magnetothermal and magnetotransport properties, paying particular attention to the magnetothermoelectric effect and thermopower response exhibited by these kind of systems at the nanoscale, together with the correlation of these effects with the anisotropic magnetoresistance phenomenon in these alloys. Absolute and relative Seebeck coefficient values in magnetic fields applied perpendicularly to nanowire axes at room temperature are determined to be 1.5 μV K^− 1 for Co_0.24Ni_0.76 nanowires and 8.1% for Co_0.39Ni_0.61 nanowires. Power factors of 3.7 mW mK^− 2 have been achieved, which are competitive with common values for more typical thermoelectric materials such as semiconducting bismuth telluride nanowires. For Co–Ni nanowires containing up to 39% Co, a linear relationship between the magnetic field-dependent change in the Seebeck coefficient and the electrical conductivity has been found (Böhnert et al., 2013).

Fe–Ni nanowires

The most exploited Fe–Ni alloy corresponds with 20% Fe and 80% Ni content. This alloy is commonly known as permalloy (Py). The high magnetic permeability exhibited by this material makes it interesting for magnetic cores and magnetic shielding applications. In addition, the low coercivity, negligible magnetostriction, and large anisotropic magnetoresistance, together with the possibility of reducing its size at the nanoscale, favor a wide range of applications in electrical and electronics-based nanotechnologies (Elmen and Arnold, 1923; Jiles, 1998).

The fabrication of Py nanowire arrays by means of electrodeposition processes commonly makes use of Ni Watts-type electrolytes with the addition of FeSO₄ as a source of Fe ions, among other sulfate- and chloride-based electrolytes (Cho et al., 2010). Some typical baths and operational conditions are collected in Table 2.3. Potentiosatic or galvanostatic methods provide good compositional control, as well as high homogeneity of the deposits. The electrodeposition potential usually ranges between − 1.2 and − 2 V versus a Ag/AgCl reference electrode. In Ni–Fe electrodeposition, there exists an anomalous codeposition reaction, which tends to preferentially deposit the Fe (the less noble metal); this, together with the differences in their standard electrode potentials, offers the possibility of tuning composition by controlling the electrochemical conditions. Figure 2.13 displays the Fe content in Ni–Fe nanowire arrays as a function of the deposition potential (Ohgai et al., 2005). Moreover, Salem et al. (2012) also demonstrated accurate control of Ni–Fe nanowire composition by changing the off or resting time between pulses in a pulsed potentiostatic electrodeposition process.

A single-phase FCC crystalline structure has been found in electrodeposited Py nanowires, which points out a reduction in the magnetocrystalline anisotropy. One effect of Fe content in Ni–Fe alloys on its crystalline structure is to change the lattice constants in such way that those increase as the Fe content increases. This is of special interest in ferromagnetic transition metals (FTMs)/copper (Cu) multilayered or multisegmented systems, where the crystalline cell size of the FTM layer should fit well to the corresponding one of the Cu layer. As a consequence of the reduced magnetocrystalline anisotropy, the effective magnetic anisotropy is determined mainly by the shape anisotropy; Py nanowires with a high aspect ratio can exhibit coercive fields as high as 1300 Oe because of the miniaturization of the material at the nanometric scale (Salem et al., 2012; Figure 2.14).

As the Fe content increases, however, the average saturation magnetization value of each nanowire also increases, and the increase in the strength of the magnetostatic interactions among nanowires provokes an apparent reduction of the effective anisotropy along the nanowire axis, thus reducing both the longitudinal coercivity and reduced remanence (Pardavi-Horvath et al., 2008).

Fe–Co nanowires

Particular attention has been paid to Fe–Co alloys because of their high saturation magnetization and Curie temperature—parameters that cannot be equalized by any other alloy systems. Bulk Fe–Co alloys with Fe content in the range of 60–70 wt% have a saturation magnetization of about 2.4 T, which is much higher than Fe–Ni and Co–Ni alloys or Fe, Co, and Ni metals (Bozorth, 1993). Fe–Co alloys are used mainly as soft magnetic materials because of their low coercivity and low magnetocrystalline anisotropy (Turgut et al., 1998; Yu et al., 2000). The first-principles calculations predict that tetragonal Fe–Co alloys with a very specific structure and alloy concentration have much improved properties for future applications: a large saturation magnetic moment and a giant magnetocrystalline anisotropy (Burkert et al., 2004). Fe–Co nanowires, however, show high coercivity and squareness, properties that make them good candidates as hard magnetic materials. A number of studies have reported the evolution and correlation of magnetic properties with the microstructure for arrays of nanowires with different composition and geometric characteristics (e.g., nanowire length and diameter; Fodor et al., 2002; Qin et al., 2002, 2003; Ramazani et al., 2011, 2012). This section deals with the evolution of magnetic properties with the composition and diameter of Co_xFe_100–x nanowires and their correlation with the corresponding crystal symmetry.

Co_xFe_100–x (0 ≤ x ≤ 100) alloy nanowires were grown into the nanopores of NAMs at room temperature using DC electrodeposition in a three-electrode electrochemical cell equipped with an silver/silver chloride (Ag/AgCl) reference electrode, a platinum (Pt) mesh counter electrode, and a gold-coated membrane acting as a working electrode. The wires were grown using sulfate-based electrolytes containing CoSO₄ · 7H₂O (0.02–0.16 M) + FeSO₄ · 7H₂O (0.02–0.16 M) + H₃BO₃ (0.16 M) + ascorbic acid (0.06 M). Fe and Co nanowires were prepared by electrodeposition using the electrolytes FeSO₄ · 7H₂O (0.22 M), H₃BO₃ (0.16 M), and ascorbic acid (0.06 M; Fe nanowires) and CoSO₄ · 7H₂O (0.36 M) and H₃BO₃ (0.16 M; Co nanowires). The pH value was kept constant at ~ 3.0. The geometric characteristics (length and diameter) of the pores of the templates, as well as of the final nanowires, were determined by SEM. The final lengths of the nanowires were around 3 and 4 μm for the samples with a 40-and 20-nm diameter, respectively.

Over a wide composition range, Fe_100–xCo_x alloys exist in the BCC (α) phase. The phase diagram for bulk Fe–Co is shown in Figure 2.15. In the range of 22 < x < 70, Fe–Co alloys undergo a phase transformation from a disordered BCC structure (α) to an ordered caesium chloride structure (α₂) below 730 °C (Bozorth, 1993). At high temperatures (up to 983 °C), equiatomic Fe–Co alloys present an α BCC structure.

For x > 50, the BCC phase intersects with a region with mixed FCC Co (γ) and BCC Fe (α) phases. For Co concentrations less than about 17%, the Curie temperature is reached before the to α-to-γ phase transition (Nishizawa and Ishida, 1990).

The XRD patterns of NAMs filled with Co_xFe_100–x nanowires with a 40-nm diameter are shown in Figure 2.16a. The results show cubic symmetry in the whole range of compositions that evolve from a BCC structure for pure Fe nanowires and those with small amount of Co toward FCC as Co content increases. For high Co content, including pure Co nanowires, the only stable phase is the γ phase, FCC structure with a [110] preferred orientation. A similar compositional evolution of the crystalline structure is found for Co_xFe_100–x nanowires with a 20-nm diameter (Figure 2.16b): (1) BCC cubic crystals with [110] texture for lesser Co content up to pure Fe nanowires, (2) FCC crystals with [110] texture for high (around 90%) Co content, and (3) HCP hexagonal symmetrical structure with [100] texture for pure Co nanowires. A cubic symmetry is observed in the whole range of compositions that evolve from a preferential FCC structure toward a BCC structure as Fe content increases. For x < 78, we obtain a BCC structure with a [110] preferred orientation. We should note that an α-to-γ phase transition in Co_xFe_100–x alloy occurs at x ≥ 78, in agreement with previous reports of bulk and thin-film alloys (Kakuno et al., 1997; Hunter et al., 2011). Fe₂₂Co₇₈ nanowires exhibit a mixed BCC and FCC structure with a [110] BCC preferred orientation. By further increasing the Co content, the only stable phase is the γ-phase FCC structure with a [110] preferred orientation. Similar data were reported for Co nanowires with diameters between 20 and 30 nm (Cattaneo et al., 2012; Kaur et al., 2013; Yuan et al., 2014). Based on a comparison of Figure 2.16a and b, we can deduce that a small amount of Fe content seems to be enough to promote FCC cubic symmetry (Bran et al., in preparation). It should be pointed out that, according to previous reports, Co wires can also crystallize into an FCC structure.

A number of studies have focused on the structural characterization of Fe–Co nanowires (Menon et al., 2001; Qin et al., 2002, 2003; Ramazani et al., 2011, 2012). They show similar phase transitions with various alloy compositions but few differences in phase boundaries. Zhan et al. (2002) showed that for Fe–Co nanowires with a 20-nm diameter the α-to-γ phase transition occurs at higher Co concentration (82 < x < 90), whereas the γ-to-ɛ transition (HCP phase) occurs at x > 92. Fodor et al. (2003) reported the structural and magnetic characterization of Fe–Co nanowires with a diameter of 40 nm. The increase in Co concentration in the alloy promoted only a transition from the BCC to the HCP phase, starting with x = 90.

The magnetic characterization of Fe_100–xCo_x nanowires presented by many authors (Menon et al., 2001; Zhan et al., 2002; Qin et al., 2002, 2003; Fodor et al., 2003; Ramazani et al., 2011, 2012) shows similar trends in the evolution of coercivity and remanence with an applied field with different values for coercivity, remanence, and saturation magnetization. The high aspect ratio of the nanowires results in a strong shape anisotropy, which is considerably larger than magnetocrystalline anisotropy, forcing the easy magnetization axis to lie along the nanowire axis. The magnetic characteristics of the nanowire arrays have been determined using vibrating sample magnetometry under a maximum field of ± 2 T applied parallel and perpendicular to the nanowire axis. The magnetization curves and their hysteresis parameters under a magnetic field applied parallel to nanowire axes have been determined as a function of nanowire composition and diameter. Figure 2.17 shows the evolution of coercivity (a) and squareness ratio (M_r/M_s) (b) under a magnetic field parallel to the nanowire axis for arrays of 40- and 20-nm diameter nanowires as a function of composition. By increasing the Co content, the coercivity increases, reaches a maximum at about 90% Co, and then decreases for pure Co nanowires. Values of coercivity > 0.27 T and a squareness ratio > 0.9 were obtained for the Co-rich samples with a diameter of 20 nm. As reported before, by increasing the wire diameter, the coercivity, and squareness decrease gradually because of the increased magnetostatic interactions between the nanowires as they become closer to each other (the diameter varies with a constant center-to-center distance). A maximum value for the saturation magnetization of about 2.35 T is obtained for the Co₄₀Fe₆₀ alloy nanowires.

A straight comparison of experimental data indicates that (1) for both coercivity and the squareness ratio, a similar trend with composition is observed for 40- and 20-nm diameter nanowires: After a nearly continuous increase in Co content, maximum values are observed for 80–90% Co. (2) Overall smaller values of coercivity and squareness ratio are observed for 20-nm diameter nanowires for small and intermediate Co content, whereas a kind of overlapping/crossing effect occurs for nanowires with high Co content and pure Co nanowires. The magnetic behavior of nanowires, such as the coercivity, remanence, and shape of the hysteresis curve, depends strongly on the magnetization reversal process.

The increase in coercivity can be explained using micromagnetic simulations, where a polycrystalline structure is considered for both Co and Co–Fe nanowires with the corresponding FCC and BCC phase. As for the magnetic parameters for FCC Co, we used a saturation magnetization M_s = 1440 emu cm^− 3, exchange constant A = 13 × 10^− 7 erg cm^− 1, and magnetocrystalline anisotropy constant K = 2.7 × 10⁶ erg cm^− 3, whereas for BCC Co–Fe, we consider M_s = 1600 emu cm^− 3, A = 10.7 × 10^− 7 erg cm^− 1, and K = 10⁵ erg cm^− 3 (Vivas et al., 2011; Ferré et al., 1997). To account for the actual polycrystalline nature of nanowire structures, a meshed tetrahedral (mesh size is 4 nm) was randomly assigned to corresponding textured phases. The [111] direction for FCC Co and the [110] direction for BCC Fe–Co were considered parallel to the nanowires axis, whereas the in-plane easy axis components were considered random.

Figure 2.18 presents the micromagnetic simulation of magnetic loops for FCC Co and BCC Fe–Co nanowire arrays in a parallel field configuration (Bran et al., 2013a,b). Both loops are characterized by a high squareness ratio, which is reasonable considering that the shape overcomes the crystalline anisotropy term. The Co FCC phase has a significantly stronger magnetocrystalline anisotropy, with a net component along the axial direction that reinforces the shape anisotropy. As a result, FCC Co coercivity and remanence take higher values than for the arrays of BCC Co–Fe nanowires that present with a weaker cubic magnetocrystalline anisotropy. Micromagnetic simulations indicate that the magnetization reversal proceeds by vortex domain wall propagation in both cases.

Simulated loops in Figure 2.18 are for given anisotropy strength values for FCC and BCC structures and fit only to some extent the data in Figure 2.18 for a wide range of Co–Fe alloy compositions. Full interpretation of data in Figure 2.18 should involve essentially the consideration of the role played by crystalline anisotropy with relative Co/Fe content. One can deduce an overall agreement in the trend of increasing coercivity with Co content because the FCC phase plays a more dominant role. The lack of full agreement can be understood considering different arguments: (1) an increase in the crystallite size with Co content (Cooper et al., 2005); (2) the defects and inhomogeneities in the samples due to the rapid deposition of Fe^2 + and Co^2 + ions, which can enhance coercivity (Qin et al., 2003); (3) small fluctuations in the diameters from sample to sample; and (4) magnetostatic interactions in a large-scale array.

Many groups (Zhan et al., 2002; Su et al., 2005) interpret the magnetization reversal mechanism in Fe–Co nanowires using a model based on “chains of spheres” (Jacobs and Bean, 1955), where each nanowire is considered to be a chain of single domain spheres with uniaxial magnetic anisotropy connected with each other by a single contact point, which make them magnetically isolated. The calculated coercivity of the nanowires using the chains of spheres model is in good agreement with the experimental results presented by the authors.

FeCoNi nanowires

Ni–Fe (Py) is one of the softest materials used in the industry, as mentioned above. It has low saturation magnetization, however, which may limit its application in, for instance, magnetic recording media. Co–Fe alloys were proposed to be an appropriated substitute for the Py to enhance the saturation magnetization; however, this alloy comes together with a magnetic hardening of the material. Another step in creating a soft magnetic material with high saturation magnetization is the combination of both previous compositions in a FeCoNi ternary alloy for, among others, high-frequency magneto-impedance applications (Atalay et al., 2013). In addition, these alloys can exhibit magnetostriction, which is interesting for the development of several sensors and actuators.

Sulfate-based and chloride-based electrolytes are commonly used in the fabrication of magnetic thin films and nanowires by means of electrochemical deposition processes. Depending on the relative concentration of the different ion sources, the composition can be tuned under the same electrochemical conditions. The pH usually ranges from 2.6 to 3.6, which also influences on the morphology and structure of the deposit (Kockar et al., 2013). Sometimes calcium chloride and ascorbic acid are added to the electrolyte to increase its conductivity and avoid Fe oxidation, respectively. Potentiostatic, galvanostatic, and alternating current (AC) electrodeposition methods have proved suitable for the fabrication of FeCoNi nanowire arrays via the template-assisted method.

As Fe, Ni, and Co crystallize into BCC, FCC, and HCP structures, the crystalline structure shown by electrodeposited FeCoNi depends strongly on the composition, and in many cases a mixture of crystalline phases is observed (Myung and Nobe, 2001; Kim et al., 2003; Yoo et al., 2006; Liu et al., 2011; Rafailovic et al., 2012).

For low Co contents without considerable HCP-phase crystallization, the magnetic shape anisotropy is dominant, resulting in an easy magnetization axis along the nanowire for large enough aspect ratios. However, the increase of saturation magnetization with respect to Py nanowires results in an increase of the magnetostatic interaction between nanowires, reducing the longitudinal coercive field values of FeCoNi nanowire arrays bellow those estimated from shape anisotropy calculations (Sharma and Grimes, 2004).