Tunable magnetic nanowires for biomedical and harsh environment applications

We have synthesized nanowires with an iron core and an iron oxide (magnetite) shell by a facile low-cost fabrication process. The magnetic properties of the nanowires can be tuned by changing shell thicknesses to yield remarkable new properties and multi-functionality. A multi-domain state at remanence can be obtained, which is an attractive feature for biomedical applications, where a low remanence is desirable. The nanowires can also be encoded with different remanence values. Notably, the oxidation process of single-crystal iron nanowires halts at a shell thickness of 10 nm. The oxide shell of these nanowires acts as a passivation layer, retaining the magnetic properties of the iron core even during high-temperature operations. This property renders these core-shell nanowires attractive materials for application to harsh environments. A cell viability study reveals a high degree of biocompatibility of the core-shell nanowires.

oxidation of the 200-nm diameter Fe wires embedded inside an aluminum oxide membrane. To date, however, the magnetic properties of individual magnetic core-magnetite shell NWs have no yet been reported.
In this work, we propose a process to synthesize Fe-Fe 3 O 4 core-shell NWs with controllable geometric (core radius, shell thickness, and length) and crystalline (single or poly crystalline core) parameters. We introduce the magnetic properties of individual core-shell NWs to reveal new multifunctional and tunable properties.

Results
The facile fabrication process consists of Fe NW electrodeposition onto aluminum oxide templates (AOT) 35 , the release of the NWs from the AOT and their subsequent oxidization in an oven (see Fig. 1). By controlling the electrodeposition conditions, we were able to synthesize either 15-μm long polycrystalline or 1-μm single-crystalline Fe NWs with a 50-nm diameter (see Methods). NW oxidation was performed in an oven at 150 °C between 10 minutes and 72 hours in an ambient atmosphere. The Fe 3 O 4 shell was formed by heat-assisted oxygen diffusion between the grain boundaries to act as a passivation layer that prevents further oxidation of the bulk Fe core in regular conditions (e.g., room temperature).
We used transmission electron microscopy (TEM) to evaluate the evolution of the core-shell NW structure for different oxidation times. Figure 2a shows the TEM image of several polycrystalline Fe NWs ranging in grain size from several nanometers to tens of nanometers. The rings found in the selected area electron diffraction (SAED) pattern, shown in the inset, identify the body-centered cubic (bcc) phase of Fe. An increase of the oxidation time results in a decrease in the diameter of the Fe core and an increase in shell thickness. Figure 2b shows the results after 20 minutes of Fe 3 O 4 in the face-centered cubic (fcc) phase and Fe in the bcc phase. After 24 hours, oxidation is complete and only a polycrystalline Fe 3 O 4 structure is observed. Figure 3a shows a TEM image and SAED pattern of single-crystal Fe NWs. The crystallinity is pronouncedly different from those of the ones of the polycrystalline NWs shown in Fig. 2a. Single micrometer bcc-structured crystals are clearly visible and the single-crystal phase is confirmed by SAED. Note that due to the specific growth conditions of Fe in the AOT, the first section of the NWs, which is attached to the Au electrode, exhibits a polycrystalline structure. This polycrystalline area is typically a few hundred nanometers long and corresponds to the penetration depth of the Au into the pores of the AOT during sputtering of the Au electrode. After annealing, rings from the fcc Fe 3 O 4 shell appear in the SAED pattern (Fig. 3b). In contrast to the polycrystalline NWs, the single-crystalline NWs do not fully oxidize. Shell thickness was 3-5 nm after 10 minutes of annealing, 5-7 nm after 1 hour of annealing and 10-12 nm after 24 hours, which was the maximum shell thickness we could obtain (no further thickness was obtained even after 72 hours) (Fig. 3c).
The thicknesses of the cores and shells after different oxidation times were measured by energy-filtered TEM (EF-TEM). Figure 4 shows typical EF-TEM images of polycrystalline and single-crystalline Fe NWs, which were obtained by combining the Fe L 23 edge and O K edge. In core-shell NWs, the O clearly concentrates at the edges (Fig. 4a,c,d), while in fully oxidized NWs, O is distributed across the entire volume (Fig. 4b).
Shell thicknesses following annealing times are summarized in Fig. 5. Shell thickness of polycrystalline NWs increases linearly with annealing time until pure Fe 3 O 4 NWs are obtained after 24 hours (Fig. 4b). Single-crystal NWs oxidize slower such that shell thickness saturates at around 12 nm (Fig. 4d). This can be attributed to the absence of grain boundaries that are required for heat-assisted oxygen diffusion. The ability of single-crystal core-shell NWs to resist complete oxidation is a great advantage for applications that require high-temperature operations, while they maintain the magnetic properties of the Fe core. Figure 6a shows a high-resolution TEM (HRTEM) cross-sectional image of a single-crystal core-shell Fe nanowire after 72 hours of annealing. Agreement of the SAED patterns with those of the reference Fe 3 O 4 does not necessarily indicate that the phase of the oxide shell corresponds to magnetite. This is because the electron diffraction ring patterns of Fe 3 O 4 and γ -Fe 2 O 3 are very similar 36 . Thus, the oxide shell can be either Fe 3 O 4 or γ -Fe 2 O 3 or a mixture of both. Spectra obtained by electron energy loss spectroscopy (EELS) for a specific atomic species are influenced by both the coordination chemistry and the valence state of the atomic species being measured 37 . Figure 6b shows the EELS map of the cross-section of a core-shell NW (single-crystalline Fe NW after 24 hours of annealing) prepared by focused ion beam protocol. Two distinguished regions are clearly visible namely, the Fe core and the surrounding Fe-O shell. In contrast, the EELS map of the cross-section of a polycrystalline section (at the end of the NW, where the growth started from) shown in Fig. 6c denotes a completely oxidized Fe-O state. The blue circle around the NW arises from the Cr edge, caused by remnants of the chrome solution used to release the NWs from the AOT (see the Methods section).
Depending on the phase and the valence of Fe in the Fe-O structure, several Fe-O configurations are possible (see the Supplementary Information) 36 . The EELS fine structure of both the O K-edge and the Fe L 23 -edges imprints these structural differences and therefore can be used to identify a specific Fe-oxide phase 36,37 . The valence state of Fe can be determined from the chemical shift (dependence of the edge position with respect to the valence), fine structural features (splitting of the peaks), and the white-line ratios of the Fe L 2 and Fe L 3 spectra 38 .   Figure 6d shows the EELS spectra recorded from the Fe core and Fe-O shell. The chemical shift of the Fe edge is clearly observable from the inset. To determine the valence state of the shell, we calculated the white-line ratios of the Fe L 2 and Fe L 3 spectra and compared them with those measured from the reference samples (for details see Supplementary Materials). Figure 6e illustrates that the value of the white-line ratio from the NW shell is in the region that corresponds to the magnetite oxidation level 37,39,40 .
Magnetization curves were obtained using a vibrating sample magnetometer (VSM). The results of Fe NWs inside AOTs (Fig. 7a) confirm a longitudinal magnetic anisotropy (with magnetization easy axis parallel to the long axes of NWs), which can be related to the strong shape anisotropy of the high aspect ratio NWs 22 . A higher remanent magnetization is observed for single-crystal Fe NWs than for polycrystalline NWs, which is ascribed to a higher term of magnetocrystalline anisotropy 22 .
The magnetization curves of polycrystalline Fe NWs released from the AOTs and dispersed on a silicon substrate show a coercivity of around 1400 Oe (Fig. 7b), which is nearly double that of the NWs inside the AOT, which reach just 675 Oe. This can be at least partially attributed to the effect of the misalignment of the NWs with respect to the field. Inside the AOT, the NWs are parallel to the direction of the external magnetic field. NWs dispersed on the substrate have a random orientation compared to the NWs in AOTs, and the switching field increases with increasing angle between the axis of the NWs and the direction of the applied field 24 .
The hysteresis loops of core-shell NWs (Fig. 7b) differ considerably from those of Fe NWs, namely, they showed a lower saturation magnetization, a lower remanence, and a smaller coercive field. Reduced saturation magnetization is caused by the smaller magnetic moment of Fe 3 O 4 compared to Fe 41,42 . This also contributes to    a slight reduction of coercivity (see inset of Fig. 7b) that strongly depends on the shape anisotropy value of the NWs, which is proportional to the NW magnetization 22 .
To investigate the magnetic properties of individual core-shell NWs, magnetic force microscopy (MFM) studies under variable fields have been conducted. A NW was first saturated under a field applied parallel to its axis and then scanned at remanence. Afterwards, the applied field was reversed in steps of 20 Oe. After each step, the field was removed to avoid any effects caused by the stray field from the magnetic tip before conducting a scan. The coercivity was defined as the field at which the MFM contrast of the NW was reversed. At remanence, the black and white points at the ends of the NW were caused by strong magnetic stray fields, confirming the single-domain magnetic state with the magnetization along the axis of the NW for both polycrystalline (Fig. 8a-c) and single-crystal ( Fig. 8d-f) NWs. The contrasts found in MFM along the polycrystalline NW are due to stray fields originating from local defects in the shape of the NW. After saturating the NWs in the opposite direction, the contrast in MFM at remanence appears reversed, indicating a change in the direction of magnetization. Magnetization reversal from one saturation state to the opposite state occurs at the coercive fields of approximately 695 Oe for polycrystalline and 1250 Oe for single-crystalline NWs, which is slightly higher than the coercivity of the array (Fig. 7a). This magnetization reversal mechanism is in agreement with previously reported micromagnetic simulations, showing a reversal process by switching the direction of magnetization at the coercive field via propagation of the vortex domain wall 22 .
The core-shell geometry significantly modifies the magnetic behavior of individual NWs. Figures 9b and 9c shows MFM images at the remanent states of a core-shell NW (polycrystalline Fe NW annealed for 20 min) after magnetic saturation with a field parallel to the long axis of the NW as indicated in the schematics view. Similar to Fe NWs without a shell, single-domain magnetic states are observed over a wide field range. Magnetization switching occurs between 870 and 1100 Oe, where the magnetization reversal process includes multi-domain states (Fig. 9d). It is reasonable to assume that this is caused by strong pinning sites introduced by the core-shell structure. The higher coercivity compared to Fe NWs without a shell can be explained by the significant change in diameter (of the Fe core) due to oxidation 22 . The multi-domain structure is maintained after removing the field, resulting in different remanence values. This unique feature allows "programming" of different magnetization values into NWs, which could prove very useful for various applications.
In contrast to core-shell NWs produced by annealing polycrystalline Fe NWs, those produced from single-crystal Fe NWs did not form multi-domain states. Both single-crystal Fe NWs and core-shell NWs were in a single-domain state and magnetization switching occurred at the coercive field of 1250 Oe for Fe NWs and 540 Oe for core-shell NWs. This reduction in coercivity can be attributed to the change in the length of the Fe core (from 1.5 μm to 1.1 μm), due to the oxidation of the polycrystalline portion at the end of the NW, which affects the shape anisotropy more significantly than the change of the core's diameter. These findings were confirmed by finite element micromagnetic simulations (Fig. 8g). The biocompatibility of the Fe and Fe-Fe 3 O 4 core-shell NWs was assessed using the MTT assay, which measures the metabolic activity of the cell and is therefore a good indicator of cell health and viability (see Methods). Figure 10 shows the cell viability of HCT 116 cells incubated with three concentrations of Fe and Fe-Fe 3 O 4 NWs for 24, 48, and 72 hours. Core-shell NWs appear to be more biocompatible than Fe NWs. It is possible that this effect is due to the prevention of intracellular dissolution of the NWs, due to the presence of the passivation Fe 3 O 4 shell, as has been suggested elsewhere 43,44 . There was no statistically significant decrease in cell viability for most of the conditions tested for the two NW materials. We only observed a significant drop in cell viability at the highest concentrations and incubation times of Fe NWs. In general, these results show that both Fe and the core-shell NWs possess a good biocompatibility and are not cytotoxic to the model cell line used.

Discussion
We have demonstrated a simple and effective preparation method for Fe-Fe 3 O 4 core-shell NWs in which the thickness of the shell and the diameter of the core are tunable. This enables tailoring of the magnetic properties, in particular the values of saturation and remanence magnetizations, within a wide range. TEM studies of the shell  thicknesses and morphology/structure of the materials confirmed that these core-shell NWs can be synthesized with either polycrystalline or single-crystal Fe cores and distinctly different characteristics. In all cases, a magnetite shell was formed. The oxide shell of polycrystalline Fe NWs grows continuously with annealing time until the disappearance of the core. However, the thickness of the oxide shell of single-crystal Fe NWs does not increase beyond about 10 nm, and the Fe core is maintained. This is specifically attractive for application to harsh environments, where such core-shell NWs can provide the magnetic properties of the Fe core at high temperatures. Magnetization measurements revealed that the saturation and remanent magnetizations depend on the oxide shell thickness, i.e. they decrease with increasing shell thickness. Meanwhile, the coercive field value reflects the competition between the effect of the magnetite shell (it reduces coercivity) and the size of the iron core's diameter (a smaller diameter increases coercivity) and changes only slightly. Our method presents a simple technique for tailoring the magnetic properties of the NWs. On one hand, NWs with a thicker shell have a smaller remanence, which minimizes magnetostatic interactions between NWs and contributes to their dispersion in solutions. On the other hand, NWs with a thinner shell maintain a higher remanence and can be used as nano-sized permanent magnets. In-between, the NWs offer many opportunities for applications in cell separation, corrosion monitoring etc. In addition, polycrystalline core-shell NWs showed several remanent magnetization states, a unique feature, which could be exploited in the future to program different magnetization values into the same NWs.

Methods
NW fabrication. Fe NWs were fabricated by a well-known process using electrodeposition into self-ordered AOT synthesized by a two-step anodization process using oxalic acid (see Supplementary Information for details). Characterization. Electron microscopy studies were performed on a Cs corrected (scanning) TEM Titan 60-300 (FEI, Netherlands), operated at 300 kV. To study the crystal structure of individual NWs, the AOT membranes were dissolved and the NWs were dispersed in ethanol. Cross sections of core-shell NWs dispersed on a Si substrate were prepared by the conventional focus ion beam protocol. The EELS and EF-TEM experiments were performed with a post-column high-resolution Gatan energy filtering spectrometer. The optical conditions of the microscope for EELS imaging and spectroscopy were defined to obtain a probe-size of 0.2 nm, with a convergence semi-angle of 10 mrad, and a 12 mrad collection semi-angle.

Formation of the Fe-Fe
The magnetic properties of the NW arrays were studied using a VSM. The magnetization curves were measured under magnetic fields up to 10 kOe with the field applied parallel and perpendicular to the axes of the NWs.
MFM images were recorded in lift-off mode (100-nm distance) with an MFP-3D-Bio from an Asylum Research scanning probe microscope using standard atomic force microscopy nano-sensor probes with a magnetic coating. A single NW was selected using scanning electron microscopy and its position was marked with a focused ion beam. MFM measurements were done at the remanent state after applying magnetic fields parallel to the axes of the NWs using a variable magnetic field module (± 5kOe).
The demagnetization process of NWs was simulated by the MagPar package with finite element discretization 45 . The average finite element discretization size was chosen to be 2 nm. The magnetic parameters of Fe were taken from Ref. 22. The magnetization of magnetite was extracted from electron holography measurements − 0.28 T. Cytotoxicity assessment. The MTT (3-(4, 5-dimethylthiazol-2yl)-2, 5-diphenyl tetrazolium bromide) assay was used to assess the cytotoxicity of the Fe and core-shell NWs. HCT 116 colon carcinoma epithelial cells (ATCC ® CCL247TM) were cultured in McCoy's 5A modified medium (Gibco ® ), supplemented with 10% fetal bovine serum (Gibco ® ) and L-glutamine and grown in a 37 °C humidified incubator with 5% CO 2 . Upon reaching 80% confluence, cells were detached from the culture flasks with 0.25% trypsin-EDTA and counted using trypan blue staining. The cells were cultured in 96-well plates, and after 24 hours of stabilization, they were treated with NWs at different NW-to-cell ratios (number of NWs per cell in the culture at the time of treatment). After the desired incubation times, the cell medium was discarded and replaced with 10% MTT solution-5 mg/mL in phosphate buffered saline-in McCoy's medium. The cells were incubated for two hours, and then the medium was discarded and replaced with 90% dimethyl sulfoxide-10% sodium dodecyl sulfate lysis buffer to dissolve the MTT reduction products. Cell viability was evaluated through optical density with a microplate reader (XMarkTM, Bio-Rad) using a wavelength of 570 nm and a background wavelength of 630 nm.