Abstract
Surface characteristics of anodic films formed on electropolished and nonelectropolished titanium substrates have been evaluated using different sets of anodisation parameters at room temperature. Surfaces were analysed by light microscopy, Raman spectroscopy, scanning electron microscopy, and X-ray diffraction. The formation of TiO2 anatase phase was only detected on nonelectropolished substrates and there seems to be a larger amount of anatase as samples are anodised; consequently, the smallest crystals were obtained at the highest frequency of pulsed current. EIS results showed that there is no difference in the degree of compactness along the layer thickness.
1. Introduction
Titanium and its alloys have been widely used in biomedical devices, such as dental and orthopaedic implants. In order to improve the osseointegration process of orthopaedic implants, various chemical and topographic modifications have been suggested recently. Several authors have observed that surface modifications using methods such as polishing, electropolishing, etching, anodisation, and plasma electrolytic oxidation are able to improve the speed and quality of osseointegration processes, thus increasing bone deposition and shortening repair period [1–6].
Some modifications reveal nanostructured features that can enhance the growth and attachment of mesenchymal cells and osteoblasts due to increased bioactive area [7, 8]. Many methods have been used to improve interfacial properties and lifetime of implants. Moreover, anodisation at high voltages is rather interesting due to being a simple and replicable process that can influence the chemical composition, morphology, and thickness of anodic films. It can transform an amorphous oxide film into a crystalline oxide; thus the implant surface becomes nanostructured, thereby favouring osteoblastic cell growth and resulting in a more effective osseointegration process [8, 9].
According to Yao et al. [10], nanostructured features can better simulate cellular environments by providing greater surface area and reactive sites for initial protein interactions that can mediate osteoblast adhesion. Kim et al. [11] found that an anodic oxide layer which is relatively thicker and more uniform than those formed naturally enhances corrosion resistance and abrasion, thus significantly increasing alkaline phosphatase activity.
Shokuhfar et al. [2] analysed the interaction between osteoblasts and amorphous and crystalline titanium dioxide and concluded that the high wettability of their surfaces is due to crystallinity exerting greater influence on cell spreading. A hydrophilic surface also shows higher protein adsorption than a hydrophobic surface, which produces a positive effect on cell growth. Among crystalline phases of TiO2, anatase and rutile are less soluble in physiological fluids than amorphous films, thus forming a more corrosion-resistant passive oxide layer [12].
A crystallization mechanism of anodic films has been proposed by Chen and Zhou [13] by considering that amorphous, anatase, and rutile phases of TiO2 present a structure which is similar to reticular TiO6 octahedral units that can be coupled by sharing their edges (favouring the anatase structure) or their vertices (in the case of rutile). This process is initiated by dehydrating [Ti(OH)6]2−, which depends on the amount of hydroxyl ions and electrolyte temperature, thus affecting the solubility of Ti4+ ions in water.
Anatase can promote apatite nucleation at a faster growth rate than rutile due to its better lattice match with apatite crystals [14]. A progressive crystallization of TiO2 can be observed as samples are anodised. By altering anodisation parameters such as voltage and electrolyte composition, an anodic film can be formed on the titanium substrate [15–17].
A previous study [18] demonstrated that anodisation by pulsed current plays an important role in modifying implant surfaces, as it enhances cellular response, which leads to increased cellular population, brings benefits regarding medical and dental procedures, and results in a faster and more effective osseointegration process. In this context, the present work aimed to create a more biocompatible surface through anatase formation by pulsed current at different frequencies and room temperature for anodisation processes.
2. Material and Methods
Commercially pure titanium (Grade 2) sheets ( mm) were mechanically abraded by using 400, 600, and 1200 emery papers. Then, the sheets were separated into a couple of groups and submitted to two polishing methods: some of the samples were electropolished by automatic polishing and using an electrolytic etching Buehler Electromet 4, and another portion of samples was polished with an automatic rotating polisher (Allied Met Prep 3) by diamond suspension. The samples were anodised at 0.1 A cm−2 pulsed current and maximum voltage of 30 V by using a digital oscilloscope (MO2061 manufactured by Minipa) and a square wave pulsed rectifier (GI21P-10/30 by General Inverter). Each titanium sample was the working electrode, a copper plate was the counter-electrode, and 2.5 mol/L H3PO4 was used as electrolyte.
The anodisation process occurred in two stages, in which the effect of anodisation at periods of 15, 30, and 60 min, 100 Hz frequency, and 40% duty cycle was performed in the first one, while 60 min of anodisation time was set for each duty cycle at varied frequencies (10, 1, and 0.33 Hz) in the second one.
Titanium oxide films were identified using a Horiba Scientific T64000 Raman spectrometer through laser radiation beam with excitation wavelength of 514 nm (10-fold 60 sec exposure to obtain each spectrum). The surface topography of anodised samples was analysed by using scanning electron microscopy (SEM, ZEISS DSM 940). The crystal structure of anodised surfaces was determined by using low-angle X-ray diffraction (D8 Advance, Bruker Co.). Cross-sectional layer thickness was evaluated by using an optical microscope Zeiss Axio Imager Z2m with circular differential interference contrast (C-DIC). Measurements of electrochemical impedance spectroscopy (EIS) were made in a 0.9% NaCl solution containing chloride ion concentration which is similar to that found in blood plasma [19]. Electrochemical measurements were made by a PGSTAT302 potentiostat (Eco. Chemie B. V., Utrecht, Netherlands). A thermostatic cell (37°C 1°C) was used with three electrodes, in which each anodised sample was the working electrode. A Pt wire was used as counter-electrode and all potentials were measured against a saturated Ag/AgCl/ electrode. EIS spectra were recorded at open circuit potential (OCP) by applying a 10 mV sinusoidal potential at frequencies ranging from 100 kHz to 10 mHz. Data fitting was carried out by FRA software (4.9 Frequency Response Analyser, Eco. Chemie B. V., Utrecht, Netherlands).
3. Results and Discussion
Figure 1 shows instantaneous pulsed current values (frequency pulses at 100 Hz) after 30 and 55 minutes of anodisation time, which enables a comparison of electropolished (Figure 1(a)) and nonelectropolished (Figure 1(b)) surfaces. All other experimental conditions for obtaining such values were identical in both cases. According to these graphs, a square wave was only observed while the polished surfaces were being anodised. An ideal square wave alternates regularly between two levels, maximum and minimum. For such a process, three parameters should be adjusted: pulsed current duration (), time between current pulses (), and current pulse magnitude. The relation between and determines duty cycle (dc) and frequency () of the process, which are defined as and [20].
(a)
(b)
Raman spectra (Figures 2(a) and 2(b)) showed weak bands that became stronger as a result of increased anodisation time as regards electropolished substrates. The lack of strong and well-defined peaks indicates that films are predominantly amorphous or nanocrystalline, but it can also be due to the fact that the film is very thin [21, 22]. After 60 min of anodisation, an intense peak at 147 cm−1 and three well-defined bands with respect to the polished sample were observed, which are consistent with literature on the anatase phase [23]. Greater crystallinity is a result of longer anodisation time, which is in agreement with observations made by Xing et al. [24]. XRD spectra (Figure 3) showed that the anodic film on a polished titanium surface consisted mainly of anatase. A higher intensity peak that is close to 25 degrees is another indicative of anatase.
(a)
(b)
(c)
(d)
Figure 4 describes SEM images of anodic titanium oxide films formed at 30 V and different current pulse frequencies of polished substrates. The film morphology has not significantly changed by decreasing frequency from 100 to 0.33 Hz, but an increase in grain size, number, and pore size was observed at lower frequencies. An increase in average size of crystalline grains can be similarly explained by the crystal growth in a precipitation process from a supersaturated solution. Precipitation consists of two distinct events: nucleation, or the initial formation of smaller stable crystals, and their subsequent growth. Larger crystals are formed when crystal growth rate exceeds nucleation rate.
(a)
(b)
(c)
(d)
A similar structure, such as a flower-like or nodule-like structure, has been suggested in literature [24], which is composed mainly of crystalline TiO2 grains and can be associated with the film dielectric breakdown features. Light microscopy in Figure 5 shows a cross-sectional anodic layer on one of the polished samples, which has revealed that the oxide film is about 20 μm thick and does not seem to have discontinuities on the titanium substrate.
The mechanism for amorphous oxide formation, as well as its transition to crystalline oxide, has been reported by several authors [15, 17, 25–28]. The formation of TiO2 via anodic oxidation starts from metal oxidation and dissociation of water at the metal/electrolyte interface. The applied electric field directs the Ti4+ ions towards the electrolyte, while O2− and OH− ions are towards the anode. Pores are formed by the migration and oxidation of O2− ions along the TiO2 matrix, which is accumulated in the form of O2 bubbles. The anatase phase is known to have better electrical conductivity than amorphous TiO2, thus favouring charge transfer processes during anodisation, whereas the cathode reaction reduces the number of H+ ions.
Figure 6 illustrates EIS spectra in Nyquist format obtained for polished titanium after anodisation at different current pulse frequencies, through which distorted capacitive semicircles can be observed in Figure 6(a). One disadvantage of displaying experimental data through complex plane plots is the fact that the impedance dependence on the applied signal frequency is not directly shown. Therefore, Bode plots are shown in Figures 6(b) and 6(c). The modulus and phase shift dependence on the frequency of each experimental condition are very similar, which indicates that the morphologies of anodic films are similar as well. An overlap of graphs shows differences in diameters of capacitive arcs with small variations in the impedance modulus and phase angle at low frequencies, which may indicate that the morphology of films does not change, even by varying the current pulse frequencies during anodisation. These differences are due to the number and distribution of crystalline nuclei on anodised surfaces.
(a)
(b)
(c)
(a)
(b)
An interpretation of these graphs was obtained by a fitting procedure with equivalent electrical circuit software, and fitted data are shown overlaid on each respective curve as solid lines.
Equivalent circuit was initially based on a two-layer structure model, Figure 7(a), as proposed by Williamson et al. [12] and Roach et al. [14], in which is the uncompensated ohmic resistance, otherwise known as solution resistance. The parallel association ( and ) represents an outer porous layer and the electrical elements ( and ) symbolize an inner barrier layer which is bound to a Warburg diffusion element near the titanium substrate. Since real surfaces of solid electrodes are always inhomogeneous, a representation of the electrochemical behaviour to investigate protecting properties of films on metals using simple equivalent electrical circuits is not viable. Possible defects and characteristics of films that are detectable by using EIS may often be ascribed to an imperfect cell geometry, which results in a nonuniform current distribution, since the corrosion process takes place on defects in a nonhomogeneous surface. and are the constant phase elements used for fitting data instead of capacitance and are related to nonuniform current distributions due to surface roughness or inhomogeneity. The inclusion of a constant phase element () in the equivalent circuit provides parameter that corresponds to either spectral degree of distortion or surface roughness.
Tables 1 and 2 show the parameters fitted for experimental results using equivalent circuit models depicted in Figures 7(a) and 7(b). Low values, in the order of 10−3, that is, the sum of quadratic deviations between experimental and calculated data, were observed for both circuit models, suggesting that both of them may be suitable for explaining the impedance spectra.
Unlike Williamson et al. [12] and Roach et al. [14], the inner and outer layer resistances herein ( and , respectively) were very similar (Table 1). This result indicates that there is no difference in the degree of compactness along the layer thickness. Thus, a simpler equivalent electric circuit model was chosen, given that the film should be composed of a single layer, in which polarization resistance () can be considered as film resistance. values (Table 2) were more coherent with experimental data in Figure 6 for the present model. Thereby, the anodic film formed at 10 Hz current pulse is the most resistant to corrosion.
Another apparently contradictory result regarding the choice of the most complex equivalent circuit was the fact that values are lower than those obtained for (Table 1). In addition, their respective values were greater than 0.9 (typical of a homogeneous and compact surface) while those found for were around 0.7 (normally designed to represent a rough surface), once it is expected that by the fact that refers to the most porous layer.
In both circuit models, the best data fitting was obtained after inserting a Warburg diffusion element, thence indicating that a diffusion process occurs through the oxide layer. As stated by Roach et al. [14], a nanostructured pore network seems to interconnect the outer layer with the inner layer; thus the electrolyte reaches the titanium substrate.
4. Conclusions
This work aimed to study anatase formation on titanium by using low-voltage current pulses in an anodisation process at room temperature. It was found that the anatase film was only formed on nonelectropolished titanium surfaces. Anatase phase proportion seems to be greater with increased anodisation time, and the smallest crystals were obtained at the highest current pulse frequencies. An amorphous or nanocrystalline film was formed on the electropolished titanium surface in the same experimental conditions as those of nonelectropolished titanium surfaces. For anodic film thicknesses of about 20 μm, EIS results revealed that there is no difference in the degree of compactness along the oxide film thickness, hence suggesting single layer formation.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Acknowledgments
This work was supported by FAPESP. Thanks are due to the Multiuser Laboratory of Advanced Optical Spectroscopy, Unicamp, and to Dr. L. R. O. Hein for SEM images and DRX spectra.