Synthesis of magnetic microtubes decorated with nanowires and cells

The fascinating properties of hollow micro/nanomaterials are attracting great attention nowadays. Their large surface area, low density, and high loading capacity, open the possibility of a large variety of applications, such as micro nanoreactors, catalysis, energy storage, biomedicine, sensors, and environmental remediation.¹ Among hollow materials, magnetic micro/nanotubes have attracted great consideration not only in material science, but also in technological applications due to their magnetic property counterpart. As far as this point is concerned, magnetically driven systems encompass magnetic-bio micro-robots, water treatment, drug delivery, localized therapy via wireless intervention due to external magnetic field actuation.² All these applications involve transport phenomena at micro/nanoscale, which can be implemented by using biocompatible and magnetic actuated agents. The chemical and physical properties of nanomaterials are strongly influenced by their morphology and crystal symmetry, whereas technological applications are brought about by their properties. This virtuous cycle has pushed scientists to find simple synthesis methods to tune in morphology leading to hollow and/or hierarchically structured materials.

On the other hand, over the last years, significant effort has been focused on the synthesis of hematite (Fe₂O₃) nanowires and magnetite (Fe₃O₄) nanoparticles in order to understand the growth mechanism and possible applications in several fields. Their higher stability under ambient condition, corrosion, resistance and low toxicity^3,4 enabled them special candidates to sensors and dispositives. These properties place this two iron oxide phases as materials with great technological potential for a wide range of applications in fields such as catalysis, gas sensors, magnetic recording media, spintronic, sensor devices, and nanobiotechnology.^5,6 Different routes such as thermal oxidation,⁷ catalyzed oxidation,⁸ microwave-assisted reflux method,⁹ and pyrolysis methods¹⁰ can be used to synthetize nanostructures of these two compounds.

Among traditional methods to fabricate hollow structures are the use of templating, coating, chemical etching, and the Kirkendall effect.¹¹ The literature concerning hollow magnetic structure is scarce. To the best of our knowledge reports on hematite and magnetite hollow microtube fabrication are absent on literature, although hematite/graphene¹² and hematite/carbon nanotubes¹³ works have been performed. In the present work, we report the synthesis of hematite, magnetite, and combined hematite/magnetite microtubes and hematite nanowires on the surface of the microtube forming multifunctional architectures. We have shown that this hierarchical hollow microstructure combining antiferromagnetic hematite, ferrimagnetic magnetite, and nanowires on its surface can be produced in a low cost, fast, reproducible way by using thermal oxidation, controlling both electrical current and annealing time. The growth mechanism, which is assisted by in situ electrical resistivity measurements has been qualitatively discussed. Most important is that our proposed methodology enables one to tune in the amount of magnetite regarding to hematite going from a ferrimagnetic (high magnetic moment) to nonmagnetic-like (very low magnetic moment due to antiferromagnetic alignment of hematite) microtubes. Fibroblasts cell culture has been done to study the adherence, migration, and proliferation of cells on the surface of the microtubes, which indicated the absence cytotoxicity for this material. The microtubes behavior of simulated torque and driving magnetic force under external magnetic field gradient indicated that they are robust and promising candidates for magnetic bio microrobots applications and biotechnology.

Iron oxide microtubes have been synthesized by using thermal oxidation process along with the passage of electrical current. A home-made apparatus was built to measure in situ electrical resistivity during all the oxidation process. A high purity metallic iron microwire (99.995%) was connected to a four-wire resistivity measurement probe mounted inside of a tubular furnace. The furnace was heated up to a set point temperature (900 °C) under different heating rates and annealing times (0 min and 30 min). The microwires used in all experiments have a diameter of 75 μm and the experiment was made in air without flux of any gas. The humidity level was around 45-60%.

The “as prepared” samples were characterized by using scanning electron microscopy (SEM, JEOL S-5500), and X-ray diffractometer (STADI-P) operating in transmission mode with radiation of Mo. Magnetic properties were measured with high resolution VSM magnetometry (MPMS-3 system, Quantum Design). The Raman measurements were performed by using a triple spectrometer (T64000, HORIBA Jobin-Yvon) with a thermoelectric cooled CCD detector (Synapse, HORIBA Jobin-Yvon). The 532 nm line of an optically pumped semiconductor laser (Verdi G5, Coherent) was used as the excitation source. The laser power at the sample was maintained below 5 mW on a spot of ∼10 microns.

Maintenance of 3T3 NIH (fibroblast) cells in culture: cells were cultivated in DMEM (Dulbecco’s Modified Eagle’s Medium) culture medium with 2 mM L-glutamine, adjusted with sodium bicarbonate 2.2 g/L and 10% fetal bovine serum of volume. Cells were kept in incubator at 37 ^oC and 5% of CO₂. When the cells extended to approximately 80% of the culture flask area, they were detached by enzyme action using trypsin solution 0.05%/EDTA 0.02% and amplified at 1:5 ratio or used on sample cell adhesion test. Sample cell adhesion test: samples were sterilized by dry heat in oven, and placed on 12 well plate. Each well was seeded with 2x10⁴ cells. The plates were placed in incubator at 37 °C and 5% of CO₂ during 3 days. After the incubation period, these samples were fixed in 4% formaldehyde buffered solution for 30 minutes, washed with water and stained with toluidine blue 0.1 % (m/v) in distilled water for 30 minutes. Samples were dried at room temperature, covered from dust and evaluated with confocal laser scanning microscope (CLSM) using a Olympus microscope, model OLS 4100.

By carrying out in situ electrical resistivity measurements we were able to assist simultaneously the oxidation process, oxide phase formation and consequently the microtube growth. These measurements revealed very important pieces of information about the whole process. Figure 1 shows electrical resistivity measurements using an electrical current of 10 mA as a function of temperature for two metallic iron microwires. We have used different heating rates for each curve and measured from T = 30 °C up to T = 900 °C followed by cooling down the system to room temperature. At the beginning of the process, the resistivity value of the microwires is in agreement with the value expected for metallic iron. As the system is heated up, the resistivity increases as also expected for a metallic behavior. However, at T = 750 °C the electrical resistivity starts to increase abruptly changing from 70 μΩ.cm to 2 x10³ μΩ.cm. As we shall see, this result reveals the initial metallic microwire transformation into a semiconducting and magnetic microtube. Depending on the heating rate of each thermal process, this critical transformation shifts to higher temperatures. It is worth noting that there is a jump on the resistivity values during the cooling process for both measurements around T = 500 °C. Interestingly, this abrupt change can be related to the oxidation process evolving dynamical phase transformation from magnetite to hematite and intermediate oxide phases. This dynamical phase transformation will be discussed elsewhere. Electric resistivity of magnetite has been reported to be lower than hematite.^14,15 We have estimated the Joule effect and found an increasing of ΔT = 2 °C which may be neglected.

In order to identify the crystal phases and study structural properties after oxidation process of the metallic microwire, X-ray diffraction measurements were performed. Fig. 2(a) shows the measurements made on the sample obtained according curve 1 in Fig. 1. Both magnetite (Fe₃O₄) and hematite (α-Fe₂O₃) crystal phases belonging to the $Fd3¯m$ and $R3¯c$ space group symmetry, respectively, were observed. In this case volumetric fraction of magnetite is 70% and hematite is 30%. Magnetite has a spinel inverse crystalline structure, where tetrahedral A sites are occupied by Fe³⁺, whereas the twice octahedral B sites are randomly occupied by Fe²⁺ and Fe³⁺. Hematite has a rhombohedral structure.¹⁶ If we change the heating rate as in curve 2, the volume fraction of magnetite and hematite phases changes. For example, if we decrease the rate giving more reaction time, hematite phase is enhanced. Moreover, if an additional time of 30 min at T = 900 °C (herein named annealing time) is included in the cycle, hematite phase dominates the oxidation process. Indeed, X-ray diffraction measurements, as shown in Fig. 2(b), reveal the presence of only hematite phase. We have observed that by changing parameters such as temperature, electrical current and annealing time different fractions of crystal phases can be obtained during the oxidation process. Representative SEM images of the obtained microtubes are shown in Fig. 3. Microtubes of panel (a) correspond to those obtained at T = 900 °C using a heating rate of 6.5 °C/min, and without electrical current. Panel (b) shows a microtube obtained by using a heating rate of 10 °C/min, without annealing time. Panel (c) shows a microtube obtained with lower heating rate and annealing time of 30 min. Panel (d) shows microtube surface synthetized with the same conditions of (a) but with an electrical current of 10 mA. In panels (c) and (d), it is possible to observe nanostructures formation, which appears in form of sticks and wires homogeneously dispersed on the microtube surface with diameters ranging from 100 nm to 300 nm.

During a normal oxidation process, the simple and well-understood steady-state diffusion of atoms throughout the structure is governed by Fick’s law.¹⁷ This means that the diffusion of atom and vacancies are induced by the difference in atomic concentration due to chemical potential gradient. The presence of an electric field, which leads to a voltage gradient, can also contribute to the chemical potential.¹⁸ The oxidation chemical reaction involves a process where a thin oxide layer is formed first on the metal surface, followed by simultaneous outward diffusion of metal ions through the oxide layer and inward diffusion of oxygen from the atmosphere into the core. Metal ions often diffuse outward faster than oxygen diffuses inward, which is consistent with the smaller ionic radius of cations than anions. The oxidation conditions including diffusion rates can determine the morphology of both intermediate and final product depending on the formation of voids and how they are arranged, as first point out by Kirkendall.¹⁹ In order to rationalize the hollowed structure formation in our experiments, we have considered the so-called Kirkendall effect.²⁰ The Kirkendall experiment revealed that diffusion of substitutional lattice atoms creates vacancies that facilitate atomic mobility. Condensation of vacancies can therefore give rise to void formation during the chemical reaction and Fe ions diffusion. Indeed, voids formation deteriorates the bonding strength of interfaces,^21,22 but such unwanted effect has become an interesting synthesis route to hollow nanostructures of various materials.²³ Since then, the Kirkendall effect has received great attention for fabricating and designed hollow structures, but in nanometer scaled materials.²⁴ On the other hand, for micrometer scale other effects should be considered during the oxidation process.²⁵ At intermediate stages of oxidation, when the inward diffusion of oxidizing species is not much faster than outward diffusion of metal cations, the solid shape of the wire is preserved and core/shell (metal/oxide) structures are often observed. However, as we shall see, when certain conditions are fulfilled hollow structured materials can be obtained. It is well known that when metallic microwire iron is oxidized different crystal phases can be formed on the surface of the metal. Nasibulin et. al.²⁶ have observed the formation of a layered sequence of Fe/Fe₃O₄/Fe₂O₃ on a thin film, followed by hematite nanowire growth on the outer layer.

In order to further characterize the structure and morphology of magnetic microtubes, micro-Raman spectroscopy measurements have been performed (Fig. 2). Two selected regions where probed: the outer layer, and the nanowires on the surface. The characteristic hematite Raman bands were observed at 225, 245, 292, 409, 497, 610, and 1321 cm^-1, whereas magnetite ones are at 298 and 663 cm^-1.²⁷ Interesting, the observed Raman bands of the two different regions were all assigned as belonging to hematite as can be seen in Fig. 2(c), (d). A careful analysis in the inner layer of the microtube revels the coexistence of a minority magnetite phase. It is know that bivalent iron from magnetite is easily prone to oxidation.²⁸ Indeed, a local oxidation of magnetite can be brought about by the laser beam during the Raman experiment.²⁹ 

The Raman spectrum of the nanostructures (Fig. 2(d)) corresponds to the hematite phase. However, a blue-shift and broadening were observed compared to the microtube, Fig. 2(c). This is a typical phonon confining effect that occurs as crystal domain size becomes of order of the Brillouin zone dimension. Due to the breaking down in the quasi-momentum conservation, contributions off Γ-point induce a frequency shift and an asymmetrical broadening of the phonon modes.³⁰ Following the method described on Refs. 31–33, the Raman intensity data for the 213 cm^-1 mode was fitted to the confined phonon in nanowires. The average diameter of the nanowires found from the fitting was D = 78 (2) nm. It is very interesting that the formation of regular hematite nanostructures on the microtube surface is increased when the electrical current is included in the oxidation process. This result suggests that the Fe ions may diffuse more efficiently through the lattice (tetrahedral and octahedral sites), grain boundary and surface of the samples when an electrical current is applied. Nasibulin et al.³⁴ have proposed a growth mechanism for the iron oxide layers and subsequent formation of nanostructures on the top through resistive heating method – a method where only electrical current is used. They claim that the driving force determining the migration of ions is the potential difference appearing during the oxidation process due to electrical current. Additionally, theoretical studies showed that the electrostatic potential across the bulk oxide is limited to ∼k_BT/e, which means that the total electrostatic potential drop across the oxide layer should have contributions mostly from the electric field interfaces.³⁵ 

We have measured magnetic properties on the microtubes obtained according to the curves 1 and 2 of the Fig. 1 where both magnetite and hematite phases coexist, but in different volume fractions. A third batch of microtubes with an additional heat treatment of 30 min at T = 900 °C was also measured. In the latter, only the hematite phase is observed, as shown in Fig. 2(b). Therefore, Fig. 4 displays magnetic moment as a function of temperature and magnetic fields (inset) for these three sets of samples. In Fig. 4(a), one can see the well-known Verwey transition at T = 128 K, characteristic of Fe₃O₄ compound.³⁶ The saturation magnetization reaches 41.5 emu/g at low field (see inset). Fig. 4(b) shows the same measurements for the second sample where one can see not only the Verwey transition of magnetite, but also the Morin phase transition at T = 260 K belonging to hematite. The saturation magnetization is lower, reaching 9.3 emu/g. The Morin transition is a spin-flip transition where antiferromagnetic (AF) structure of Fe₂O₃ changes to weakly ferromagnetic order.³⁷ Fig. 4(c) shows the presence of only Morin transition; the saturation magnetization is very low ∼ 0.5 emu/g even in magnetic field as large as 20 kOe confirming hematite microtubes. Corroborating the previous results, temperature, heating rate, and annealing time play a fundamental role turning magnetite into hematite phase. It is very interesting that by using this method one can tune in the amount of magnetite going from a ferrimagnetic (high magnetic moment) to a nonmagnetic-like (very low magnetic moment due to antiferromagnetic alignment of hematite) microtubes. To the best of our knowledge, the synthesis and control of magnetic phase volume fraction in a single microtube has not been reported yet. In the right panel on Fig. 4, we show a drawing illustrating the microtube along with magnetic lines and magnetic induced field produced by each one when they are saturated.

Controlling the magnetic phases of microtubes may have several applications in devices driven by external magnetic actuation. Indeed, the wirelessly rotational and translational device movement can be driven by applying magnetic field gradients generated by external electromagnets. As said before, such systems can be incorporated in microdevices for wireless magnetic manipulation with applications as local drug delivery or assistance in surgical operations. Recently, ex vivo testing in porcine eyes and in vivo in rabbit eyes, magnetic microtubes were successfully manipulated using external magnetic fields.³⁸ Rotational and translational movements of the microtubes require a magnetic torque and force acting on the magnetic structure. Magnetic torque, T_m = VM x B, and magnetic force, F_m = V ∇(M. B), can be calculated using the magnetic field (B) and magnetic field gradient (∇ B): V is the volume of the sample in m³, B is the applied magnetic field in A/m and M is the magnetization of the sample in Tesla.

We have calculated the magnetic torque and magnetic force for the microtubes measured in Fig. 4. Before, we have estimated the volume and saturation magnetic moment of microtubes which were found to be 8.2x10^-11 m³ and 334.8 kA/m, 8.84x10^-11 m³ and 3.84 kA/m, for microtubes shown in Fig. 4(a) and (c), respectively. Fig. 5 shows both torque and driving magnetic force as a function of external magnetic field gradient from 1 mT/m up 10 mT/m. Also, we have used a gradient field supposing translation movement of the microtubes along the x-axis direction. By selecting a set of possible angles one can also estimate the magnetic torque. Fig. 5 reveals robust rotation and translation response to an external magnetic field indicating promising technological applications of these hollow microstructures in magnetic actuation - for instance, the magnetic force reaches up to 300 nN. Here, it is important to note that we simulated using low external magnetic fields (1-10 mT) and even considering these conditions both torque and magnetic force are the highest reported in literature. If we assume higher values of external magnetic field (100-800 mT m-1), as used in other works,³⁹ both torque and magnetic force values for three microtubes increases considerably; for example, with a gradient field of 800 mT m^-1 ferrimagnetic microtubes would have values of T_m = 2.7x10^-4 N.m and F_m = 8.83 x 10^-5 N, while antiferromagnetic hematite microtubes T_m = 17.2x10^-9 N.m and F_m = 17.2x10^-9 N. In order to get further insights into wireless applications, we have simulated a magnetic field gradient produced by a finite solenoid (current = 1 A, length = 50 cm, radius = 10 cm) in Fig. 6. The solenoid axis is along the z-direction. The solid lines show the magnetic field gradient. After inserting the microtube in the bottom with the N-S axis along the solenoid radius direction, at initial times it rotates counterclockwise under the effect of the magnetic torque. The field gradient also exerts a force producing a linear movement along the z-axis. Controlling the current value, it is possible, for example, to equilibrate the weight of nano/microtube while keeping it levitating inside the solenoid. The acceleration acquired by a single microtube is also shown. At solenoid borders it could reach ∼ 30 g, being a noticeable value. This could increase the terminal velocity of falling spherical particles in a viscous fluid by a 30 factor. The translational motion dynamics will be published elsewhere.

In addition to these applications, we demonstrated that nano/microstructured hematite, which is the outer layer of our microtubes, has biocompatibility and affinity with cells opening the possibility to culture them developing bio-micro-robots. Therefore, in vitro experiments involving adherence, migration, and proliferation of cell culture assay by using fibroblasts have been carried out. These results revealed the presence of stained cells (blue) on the surface of all tested samples as shown in Fig. 7. The cell adhesion and growth considering the 3-day-test evaluation period indicated the absence of immediate cytotoxicity for this material, otherwise cellular death should have taken place.⁴⁰ The highly rough surface of the samples due to the presence of nanowires allowed cells interaction and survival on the nano-engineered topography. The cell spreading on the surface of the material, anchoring itself along it in different directions, indicates good interaction reflected by the cells accommodation and their recognition of friendly surfaces.^41,42 Images on the surface of the microtubes before and after the cell culture assay were also obtained by confocal laser scanning microscope. After 3 days immersed in the culture medium and in contact with fibroblast cells, one can observe neither structural changes nor any indication of corrosion byproducts from the material in the culture medium. This inherent corrosion resistance of the microtubes is certainly due to the presence of hematite layer. Interesting, despite the different magnetic properties between samples, from higher (magnetite) to lower (hematite) magnetic moment, there was no difference on the cellular interaction when considering cell spreading and adhesion. This may also be related to the presence of hematite passivation layer which is in direct contact with cells. A viability test using colorimetric assays to check number of live cells could not be performed due to lack of proper positive and negative controls. Indeed, the cells numbers in each sample grew exponentially, but the small size of samples limited the approach of standardized protocols in which a specific area of the sample is needed to perform the test. The initial propose was to evaluate the behavior of the cells at the new surface and their attachment on the material in a short time, keeping the adequate morphology.

These results revealed that the magnetic microtubes decorated with nanowires may be used as a hierarchical biomaterial device. The magnetic properties of those microtubes allowed them to be driven by magnetic actuation in a fluid environment,⁴³ but these micro-devices may also be used in drug delivery system, biopsy, locally induced hyperthermia, etc.⁴⁴ Hard magnetic materials such as Co and FePt could have higher saturation magnetization, but are toxic, which prevents their use in most of biomedical applications.⁴⁵ Furthermore, the product of this work integrates two hierarchical morphologies (microtube/nanowires) into one single device that could rationally combine the advantages of different applications of magnetic microtubes and nanowires with cells. We consider that the results presented here will attract attention from materials science and biotechnology scientific communities and other insights may derive along this line.

In summary, magnetic biomicrotubes decorated with nanowires and cells have been obtained. The magnetic microtube synthesis process is accompanied by in situ electrical resistivity measurements. Microtubes of both antiferromagnetic hematite and ferrimagnetic magnetite with nanowires on the surface can be obtained by controlling the set point temperature, heating rate, annealing time, and electrical current. The hollow structure formation is brought about by the difference in the diffusion coefficient of outward Fe and inward oxygen ions migration which lead to the voids formation. As revealed by electrical resistivity measurements, the hollow formation takes place in a temperature window of T = 800 – 900 °C. This simple method to fabricate magnetic microtubes and the facility to tune in the magnetic moment magnitude going from an antiferromagnetic (nonmagnetic –like) to a ferromagnetic state may bring about interesting applications. In vitro experiments of fibroblasts cell culture on the surface of the microtubes revealed adherence, migration, and proliferation which indicated the absence of cytotoxicity for this material. We have estimated the magnetic torque and magnetic force which reveal robust rotation and translation response to an external magnetic field confirming promising technological applications of these hollow microstructures in magnetic actuation. We believe that interesting breakthrough and new applications may be achieved by using magnetic bio microtubes in several fields. Taking advantage of the hollow structure, it is possible to build a transport mechanism of ionic fluid through it by means of Lorentz forces. Besides cells on the surface, it is also possible to think in trap inside chemical or biological agents towards drug and targeted cell delivery.

This material is based upon work supported by the Brazilian agency CNPq under grants Nos. 2011/19924-2, 306431/2014-9, 455092/2014-1, 311146/2015-5, 309202/2014-0, 402289/2013-7 and 307764-2015-5 and FAPESP under grants Nos. 2013/16172-5 and 2016/09769-3. The authors are grateful to the Multiuser Central Facilities (UFABC) for the experimental support. TSG is grateful to PNPD/CAPES fellowship.