M. I. Irshad*ab,
N. M. Mohamedab,
M. Z. Abdullahc,
M. S. M. Saheedab,
A. Mumtazab,
M. Yasarab,
A. Yarab,
M. A. Zeeshand and
J. Sorte
aCentre of Innovative Nanostructure & Nanodevices (COINN), Universiti Teknologi PETRONAS, 32610 Bandar Seri Iskandar, Malaysia. E-mail: imrancssp@gmail.com
bDepartment of Fundamental and Applied Sciences, Universiti Teknologi PETRONAS, 32610 Bandar Seri Iskandar, Malaysia
cDepartment of Mechanical Engineering, Universiti Teknologi PETRONAS, 32610 Bandar Seri Iskandar, Malaysia
dInstitute of Robotics & Intelligent Systems, Zurich, Switzerland
eInstitució Catalana de Recerca i Estudis Avançats (ICREA) and Departament de Física, Universitat Autònoma de Barcelona (UAB), 08193 Bellaterra, Spain
First published on 28th January 2016
Cobalt nanowires (NWs) have been synthesized by electrodeposition inside the pores of anodized aluminium oxide templates, at different values of applied deposition potential. The as-deposited NWs are parallel to one another and exhibit a high geometrical aspect ratio. The crystal structure of these NWs shows a strong dependence on the applied deposition potential during synthesis. X-ray diffraction indicates the predominance of hexagonal-closed-packed (HCP) phase with (002) texture at low applied deposition potentials, whereas a reorientation of the c-axis of the HCP structure is observed for high electrodeposition potentials. Moreover, for a given electrodeposition time, the length of the NWs also increases with the applied potential. As a result of these structural changes, a switch in the magnetic easy axis, from parallel to perpendicular to the NW axis, occurs depending on the applied potential. A simplified model is used to account for this reorientation of the effective magnetic anisotropy direction, which takes into account the interplay between shape anisotropy, magnetocrystalline anisotropy and interwire dipolar interactions.
The physico-chemical properties of NWs depend on their aspect ratio (i.e., diameter vs. length),7 surface roughness8 and the inherent crystallographic structure.9 Different crystallographic phases and textures have been reported in the literature for electrodeposited Co NWs.10–12 Some works have shown the presence of hexagonal-closed packed (HCP) as the main phase;11 others have obtained face-centered cubic (FCC) Co NWs.12 A thorough study is required to fully understand the transition between these two phases, eventual changes in the crystallographic orientation, and the impact of these structural features on the magnetic properties of the NWs.
The fabrication methods to prepare arrays of NWs have traditionally been physical vapour deposition, chemical vapour deposition, lithographic patterning and electrodeposition. The latter has attained a special interest because of its unique advantages, such as low production cost, high purity of the resulting materials, and possibility for mass production of NWs.13–18 Recently, Ivanov et al.19 have shown that the crystalline structure, and hence the magnetic properties, of Co NWs can be highly tuned by varying the pH value of the electrolytic bath. In turn, Xianghua et al.20 reported that the variation in temperature during the synthesis allowed controlling the uniaxial magnetocrystalline easy axis direction of HCP-Co NWs. Namely, by proper adjustment of bath temperature, it was possible to orient the c-axis of the HCP phase to be either parallel or perpendicular to the NW long axis. The ability to tune the magnetic easy axis in NWs is of practical importance in applications like magnetic MEMS/NEMS, magnetic sensors and magnetic recording media, among others. Furthermore, Wang et al.10 compared the growth of Co NWs in the HCP and FCC phases at two different applied deposition potentials. In spite of these previous works, a systematic and comprehensive study, as a function of applied deposition potential, of crystallographic reorientations of HCP phase, and correlation between crystalline structure and the effective orientation of the easy axis of magnetization is still lacking.
This work tackles the influence of the electrodeposition potential on the crystallographic structure and magnetic properties of Co NWs. For this reason, all other electrodeposition parameters (i.e., composition of the electrolytic bath composition, temperature, pH, deposition time, etc.) are kept constant and only the applied potential is varied, from 0.6 V to 1.8 V. By doing so, the shape of the hysteresis loops is found to drastically change as a function of the applied potential. Namely, for low potentials, the easy axis of magnetization is oriented parallel to the NWs long axis, while for high potentials the easy axis of magnetization is oriented perpendicular to the NWs long axis. To some extent, this is a counterintuitive result since the length of the NWs (and hence the shape anisotropy, which favours orientation of the easy axis along the NW axis) increases with the applied deposition potential. After careful examination of the microstructure of the NWs, this magnetic reorientation can be explained from a crystallographic reorientation the c-axis direction of the HCP phase (which is the magnetocrystalline easy axis in HCP-Co). The effect can be interpreted in more detail using magnetic energy considerations that take into account all magnetic anisotropy contributions (shape, magnetocrystalline) as well as the interwire dipolar interactions between the NWs.
Sample no. | 1 | 2 | 3 | 4 | 5 |
Applied potential (V) | 0.6 | 0.9 | 1.2 | 1.5 | 1.8 |
Average nanowires length (μm) | 2.5 | 4 | 10 | 17 | 25 |
A schematic diagram that depicts the features of the AAO template from side and top view having a pore diameter around 100 nm and total pore length of 60 μm is shown in Fig. 1.
The pink-colored cylinders inside the pores of AAO templates represent Co NWs which are finally extracted from these templates by template dissolution in a basic solution.
All the samples were analysed using a Carl Zeiss Supra55 VP field-emission scanning electron microscope (FESEM) – to determine the morphology and length – assisted with energy dispersive X-ray spectroscopy (EDS) analysis for chemical composition. A Zeiss Libra 200 high resolution transmission electron microscope (HRTEM) was employed to assess the diameter and internal structure of the NWs. A Panalytical Empyrean X-ray diffraction (XRD) machine using characteristic Cu Kα X-rays having wavelength of 1.54 Å, was employed to study the structural phases in Co NWs embedded in the AAO template. The magnetic properties of the NWs were measured using a Lakeshore 7400 series vibrating sample magnetometer (VSM), with an applied magnetic field in parallel and perpendicular configurations with respect to the NWs long axis.
Fig. 2(d) shows the EDS pattern acquired for the Co NWs when still embedded inside the AAO templates. The spectrum indicates the presence of a high Co content. The peaks of Al and O appear due to the AAO template. The EDS pattern does not show any remarkable impurity except for phosphorous which comes from AAO template as they are synthesized in the presence of an electrolyte containing phosphoric acid (H3PO4).
The XRD patterns of Co NWs as a function of applied deposition potential are shown in Fig. 3. The XRD patterns reveal that at lower applied deposition potentials, the HCP phase of Co, matching the JCPDS reference number 89-4308 (a = 2.5070 Å, c = 4.0690 Å), is predominant, with a preferred orientation along the (002)H direction, similar to other works from the literature.2,21–24 Thus, in this case, the NWs grow with the (002)H planes oriented along their long axis. However, as the applied deposition potential increases from 1.2 to 1.5 V, the preferred HCP orientation shifts from (002)H to (010)H and (110)H directions, suggesting that the c-axis of the HCP phase becomes oriented perpendicular to the NWs long axis. Finally, the XRD peaks are rather sharp, indicating that the crystallite size is relatively large. In fact, application of the Scherrer's formula gives an average crystallite size larger than 50 nm for all the investigated NWs.
A more detailed investigation of the NW structure was carried out using HRTEM. For TEM observations, the template was dissolved in 1 M NaOH solution at 60 °C. The NaOH solution was removed by pouring de-ionized water slowly onto the suspension and taking it out with a suction pump. The detailed dissolution procedure of AAO template is described in our earlier research work.17 The NWs were then placed onto a Cu grid for analysis.
The TEM images give more comprehensive details of internal structure of the NWs shown in Fig. 4(a and d) are TEM images of isolated Co NWs, deposited at 1.8 V and 0.6 V, placed onto a holey Cu grid. The average diameter of the NWs is approximately 100 nm. The high-resolution TEM image of the sample synthesized at 1.8 V reveals the occurrence of atomic fringes whose orientation is parallel to the long axis of NWs (Fig. 4(b)). The inter-planer spacing of these planes corresponds to the HCP (002) planes (see also Fig. 4(c)). However, for NWs deposited at deposition potential of 0.6 V, the HCP (002) planes are found to be roughly perpendicular to the NW long axis (see Fig. 4(e and f)). Thus, the Co NWs grow preferentially along the (002) HCP direction for low applied deposition potentials, whereas a reorientation of the c-axis of the HCP structure is observed for sufficiently high applied potentials. This is in agreement with the XRD results shown in Fig. 3.
The structural changes in the deposited Co NWs can be attributed to the hydrogen co-evolution during the synthesis process.25,26 This can be rationalized in terms of the lowering of the metal's surface energy and hydrogen adsorption phenomenon. Lowering of the surface-energy varies for different crystallographic planes because of their dissimilar adsorption abilities. It has been demonstrated, by means of quantum-mechanical calculations, that the activation energy for hydrogen adsorption is lower for greater distances between adjacent metal atoms. This means that the surface energy for loosely packed planes decreases more rapidly than for closely packed planes. The basal (002) planes have lattice spacing of 0.20 nm, which is lower in comparison to the (010) planes, whose lattice spacing is 0.22 nm. Thus formation of (010) planes at high voltages is more favourable due to the fact that activation energy is low for this plane for hydrogen adsorption. Krause et al.27 have also attributed the growth of HCP-Co phase to variation in surface free energies due to co-deposition of hydrogen.
Fig. 5 shows the combined hysteresis loops of Co NWs, measured while still embedded inside the AAO template, as a function of the applied deposition potential, both when the external magnetic field is applied parallel and perpendicular to the NWs axis. The specific saturation magnetization value of Co NWs is lower than for bulk Co since the total mass here includes not only the mass of the NWs but also the ones from the template and from the Cu thin layer.28,29
In an isolated NW, the magnetocrystalline and shape anisotropies are major factors that dictate the final orientation of the effective magnetic easy axis.30–32 For HCP Co, the magnetocrystalline anisotropy favours the magnetization to be preferably oriented along the (002)H direction (i.e., along the c-axis of the HCP phase). Hence, depending on the crystallographic texture, the magnetocrystalline anisotropy will promote the easy axis to be either parallel or perpendicular to the NW long axis. Conversely, the shape anisotropy always tends to align the magnetic moments along the long axis of the NWs.
It is a well-known fact that growth conditions of NWs have a tremendous effect upon their length. The NWs get longer as one increases the voltage during electrodeposition. This means that shape anisotropy (promoting easy magnetization axis along the wire axis) is enhanced by increasing the voltage. Nevertheless, one can observe that, contrary to the expectations, a reorientation of the easy axis (from parallel to perpendicular to the NWs long axis) occurs for sufficiently high voltages. This stems from the reorientation of the c-axis in the HCP phase. For high voltages, the magnetocrystalline anisotropy contribution is so important that the easy axis of magnetization becomes perpendicular to the NWs, even though they are longer.33
The interplay between magnetocrystalline and shape anisotropy for HCP-Co is depicted in Fig. 6. The shape anisotropy is represented by a dotted line that tends to orient the magnetic moments parallel to long axis of the NWs. The magneto-crystalline anisotropy is represented by a continuous line and, for high potentials, it tends to orient the magnetic moments in a direction perpendicular to the long axis of NWs, whereas for low potentials, the magnetocrystalline and shape anisotropies are parallel to each other. Hence, for high potentials, there is a competition between these two anisotropies and the final direction of easy axis of magnetization depends on which of them is stronger.34
For an array of NWs (still embedded in the AAO template) the effective magnetic easy axis will be also influenced by interwire dipolar interactions. Hence, the following three contributions will have an effect on the easy axis direction: (i) the shape anisotropy of the NWs (which depends on their aspect ratio), (ii) the magnetocrystalline anisotropy (whose direction depends on crystallographic texture) and (iii) the magnetostatic coupling among the NWs (which depends on the interwire dipolar interactions and tends to develop a magnetic easy axis perpendicular to the NW axis).
Fig. 5 reveals that, for low applied potentials, the hysteresis loops measured along the NWs long axis are squarer in shape than those measured along the perpendicular configuration. However, with the increase of the applied potential (from 1.2 to 1.5 V), the squareness of the parallel loops is progressively reduced and the parallel and perpendicular loops almost overlap. At an applied deposition potential of 1.8 V, the perpendicular hysteresis loop is squarer than the one measured along the NW axis, just opposite to what is observed for low deposition potentials. The dependence of the squareness ratio (defined as the ratio between the remanent, Mr, and the saturation, Ms, magnetizations) on the applied deposition potential is shown in Fig. 7.
![]() | ||
Fig. 7 Squareness ratio of Co NWs as a function of applied deposition potential, for parallel and perpendicular configurations. |
For low potentials, the effective easy axis parallel to the NW axis is energetically favourable since, in this case, both the shape anisotropy and the magnetocrystalline tend to align the magnetization along the long axis of the NW, which corresponds to the (002)H direction, as well. Conversely, for high electrodeposition potentials (e.g., 1.8 V), the magnetic easy axis seems to be perpendicular to the NW axis. This could indicate a strong role of the magnetocrystalline anisotropy (in this case the c-axis of the HCP structure is perpendicular to the NW axis), at the expenses of the shape anisotropy.35 The eventual partial phase transformation of Co, from HCP (low deposition potential) to FCC (high deposition potential), would have a little effect in reducing the overall value of the magnetocrystalline anisotropy, since FCC is more isotropic (i.e., with various equivalent magnetic easy axes) than HCP (i.e., with a single easy axis along the (002)H direction). Furthermore, the change in length of the NWs as a function of applied potential has an effect on the strength of interwire dipolar interactions.
To take all these effects into account and make this reasoning more quantitative, let us assume a two-dimensional infinite array of cylindrical particles with radius r, length L, aspect ratio Ar = L/2r and interwire distance D. One can then consider that the total dipolar field generated when all NWs are oriented along the NW axis direction can be written as:36–38
![]() | (1) |
On the contrary, if all moments are aligned perpendicular to the wires direction, then the dipolar field would be:
![]() | (2) |
The self-demagnetizing field of one NW can be then expressed as:
Hdemag = (N⊥ − N‖)MS | (3) |
![]() | (4) |
![]() | (5) |
Hence, the total effective anisotropy field can be then expressed as:36
![]() | (6) |
This equation predicts that if HK,eff is positive the magnetic easy axis will be parallel to the wires axis; otherwise, for negative HK,eff, the magnetic easy axis will tend to be oriented along the perpendicular direction. Here Hmc represents the magnetocrystalline anisotropy contribution, which for single-crystalline HCP-Co is approximately equal to 5MS.38 If one uses approximate values of the NW length and diameter and the distance between NWs (pitch size), one can try to estimate the effective anisotropy field of a given array of NWs using eqn (6). For high applied potentials, HK,eff is clearly negative, irrespective of the value of Hmc, since the second factor in eqn (6) is much larger than the self-demagnetizing field and Hmc, which at most (i.e., for pure HCP phase) would be equal to 5MS. Conversely, for low voltages, HK,eff is actually slightly positive, assuming that Hmc is 5MS and contributing to orient the easy axis along the NW axis. Hence, this simple model is capable of describing, in a semi-quantitative manner, the reorientation observed in the magnetic easy axis of the hysteresis loops depicted in Fig. 5.
Additional microstructural effects can also be partially responsible for some of the observations made on the hysteresis loops. For example, a higher deposition potential enhances the nucleation rate of the Co grains, resulting into finer grain formation and, eventually, formation of non-compact grains. It also promotes larger hydrogen atoms adsorption during electrodeposition.40 These changes in structural properties can contribute to a further curvature in the shape of the hysteresis loops.
This journal is © The Royal Society of Chemistry 2016 |