Development of Titanium Surfaces Oxidized by Electrolytic Plasma, for Biomedical Application

Local drug delivery systems used in dental implants need to have uniform pore size distribution, adequate wettability, chemical composition, and biocompatibility. In the present work, titanium was treated by pulsed plasma electrolytic oxidation (PPEO), using 0.025 M sodium dihydrogen phosphate and 0.25 M calcium acetate as electrolytes, aiming at use in drug delivery systems. Pulse with Ton/Toff (width pulse/repetition time) of 50 μs/100 μs or 100 μs/50 μs, duty cycle 0,33, and 0.67, respectively, were used. After treatment, Ca/P ratio, wettability, crystalline phase, pore size and distribution were determined. The average pore size ranged from 1.5 μm to 2.3 μm according to the increase of energy supplied to the system. Pore distributions with lower dispersion were verified for the Ton/Toff condition of 50 μs /100 μs, using a current density of 30 mA/dm 2 . On the other hand, the 100 μs/50 μs conditions produce larger pores, but with greater dispersion. In general, conditions with lower currents (30 mA/ dm 2 and 38 mA/dm 2 ) and Ton/Toff ratio = 50/100 were the most appropriate for use in drug delivery systems, due to their size and distribution of uniform pores, greater hydrophilicity, and Ca/P ratio close to desirable (1.67) was obtained.


Introduction
Due sua non-toxicity, high specific strength, and good biocompatibility titanium is one of the most used metals in biomedical application 1 .However, in applications such as dental implants, a surface that induces bone growth is also desirable 2 .This complex osteointegration process involves several biological processes which occur at the tissue-implant interface 3 .Since titanium is a bioinert material, surface modifications are desirable to overcome this deficiency.One of the strategies used is to coat the implant with hydroxyapatite to improve the osseointegration cellular response, due to the chemical and crystallographic similarity with the inorganic components of dental and bone tissues 4 .Also modifying the microstructure and/or surface roughness, and wettability, contributes to increased cell adhesion 5,6 .Currently, the development of local drug-delivery systems in order to decrease the risk of peri-implant infections has attracted attention 3,7,8 .It is important that these systems contain pores (micro or nanotubes) of controlled dimensions, morphology, crystallinity and distribution 9,10 .These numerous desirable properties for titanium implant surfaces can be acquired using the plasma electrolytic oxidation (PEO) technique.In this technique, the titanium surface is polarized anodically in an electrochemical cell, to produce a porous surface, primarily containing titanium oxide 11 .High voltages (250-750 V) between two electrodes result in short-term plasma microdischarges at the electrode/electrolyte interface.The heating and cooling cycles produced during the appearance and extinction of microdischarges, determine the thermodynamic conditions of this interface, decisively influencing the phase evolution, chemical composition, microstructure, and morphology of the coatings 12,13 .In the pulsed current PEO, it is possible to control the number and duration of that microdischarge by adjusting the pulse parameter (Ton and Toff) [14][15][16] .It is known that duration time of microdischarges (30 -300 µs) and the incubation period between an individual discharge and a cascade (100 -1000 µs) can be controlled by adjusting the Ton and Toff in these time bands 15 .Understanding the correlation between surface properties and current pulse parameters, we *e-mail: clodomiro.jr@ufersa.edu.br are proposing a pulsed plasma electrolytic oxidation system, using shorter square current pulses (frequency=6.7 kHz), in order to interrupt microdischarges and thus, the main events of the process.To illustrate the importance of this work, we hypothesized an application in dental implants with drug delivery systems.

Sample preparation
Before plasma oxidation, grade 2 titanium discs (3 mm diameter; 2 mm thick) were electrically connected to a conductive wire.The discs were embedded in resin so that only the surface to be oxidized was exposed.Next the samples were metallographically prepared with a final polishing of 0.25 µm silica, then washed with distilled water, cleaned with alcohol and cotton wool and finally dried for further treatment.

PEO treatment
The PEO treatment was carried out in an equipament (Figure 1) composed of a glass beaker containing 400 ml of electrolyte solution 0.25M of Ca(CH3CO2)2.H2O and 0.025M of NaH2PO4 .2H2O, with the sample anodically polarized (Figure 1).Opposite to anode a SAE 304 stainless steel square plate was placed (cathode, 50 mm length; 2.0 mm thick).
Square unipolar pulsed current were combined for two different duty cycles (0.33 and 0.67), using current of 30 A/ dm 2 , 38 A/dm 2 , and 45 A/dm 2 , resulting at six experimental conditions (Table 1).The duration of all treatments was fixed at 10 minutes.The solution temperature varied between 25 o C and 30 o C, depending on the current density used.

Samples characterization
Surface micrographs of the coating were obtained using a high resolution scanning electron microscope (SEM) JSM 6010LA, Jeol.For chemical analysis by energy dispersive spectroscopy (EDS), a Dry SD Hyper X-ray detector (EX-94410T1L11) was used, coupled to the SEM.The number, size and distribution of pores were analyzed from electron micrographs using ImageJ software.The wettability test was performed by measuring the static contact angle using the sessile drop method.Images of the accommodation of a drop of 10 μl of distilled water on the treated surfaces were recorded.The crystallographic phases were analyzed with the aid of an X-ray diffractometer, XRD-6000, Shimadzu, with Cu Kα radiation, operating at 30 kV, in the angular range of 20 ° to 80 °, with a scan speed of 2 degrees per minute.and angular pitch of 0.02°.The phases were identified with the help of the High Score Plus program.

Microstructural characterization
Changes in morphology and size were dependent on the process condition (Figure 2).In general, samples treated with lowest current values (30T5/10 and 30T10/5), pores are smaller, more homogeneous, and circular.The other samples present, in addition to the pore structure, a continuous film interconnecting the pores, called pancake's structure, which is a consequence of the strong micro-arcs that occur in the localized melting channels, covering pores and sometimes generating small cracks (Figure 2c).Mainly in Figure 2a, 2b and 2d, some pores protrude beyond the surface, suggesting columnar growth of the oxide, giving rise to nano or microtubes.
These porous structures are particularly important for dental implants, especially when considering local drug delivery systems.They must have a high specific surface area, rough and/or porous topography with the appropriate number, size, and distribution of pores that adequately meet the time required for release of specific drugs 17,18 .Pores with an average size between 0.4 and 3 µm are those that best accommodate the cells and thus have a better biological response 19,20 .In the present work, the mean diameters and dispersion of pore values were calculated using the imageJ software from the electron micrographs of the samples.It was verified that the values varied proportionally with the electrical current and pulse width.For the lowest values of current and pulse width (30T5/10), the average diameter  was 0.70 ± 0.30 µm, while for the highest values (45T10/5), the average diameter was 1.50 ± 1.00 µm.According to the literature, these values are within the desirable range for a good biological response.Regarding the thickness of the coating, an average value of 10 µm was obtained from its measurement by profile electron micrograph was found for all experimental conditions, except only for sample 30T5/10, which had a thickness of 25 µm.This difference can probably be attributed to the detachment of the layer that occurred during the metallographic preparation.A growth rate of 1.0 μm min −1 is in agreement with the literature for this type of treatment.Considering these pore dimensions, the volume/pore available for drug storage can be calculated by assuming the approximate shape of a hollow cylinder for the pore.These values varied between 3,8 µm 3 (30T5/10) and 17.7 µm 3 (45T10/5).Despite the lower volume/pore of samples produced with lower electrical current and pulse width, these are more interesting for drug delivery systems due to their homogeneity.It is also important to verify that pores with smaller diameters have a greater number of pores per unit area, being 0.14 pores/ µm 2 and 0.05 pores/ µm 2 , for the smallest and largest pore diameters, respectively.

Chemical analysis
EDS chemical analyzes were carried out throughout the scan region as well as at the points illustrated in Figure 3. Attempts were made to correlate the chemical composition of the film with the distance to the pore, but no consistency was achieved.What can be assumed from this inconsistency is that the process was controlled mainly by surface reaction than by thermal activation.If the oxidation process were thermally activated, it was expected that in the vicinity of the pores there would be a greater formation of TiO2, since this region has a higher temperature.The Ti/O ratio was less than 0.5 (stoichiometric ratio for TiO2) for the 38T5/10 and 45T5/10 condition (Table 2), indicating that there was no complete formation of TiO2, unlike the other conditions that presented similar values to 0.5 or greater.In the condition in which there was an excess of oxygen (45T10/5), other oxides such as CaO may have formed.
Regarding the Ca/P ratio, it was observed that the values were generally higher than 1.62 (stoichiometric ratio of hydroxyapatite).However, when these ratios are observed punctually, it is verified that in many cases the stoichiometric ratio was reached.This suggests the existence of dispersed hydroxyapatite particles, and in a heterogeneous way, on the titanium surface.From the diffractograms (Figure 4), it can be seen that the titanium peaks are more intense for surfaces produced with lower value of electric current.This means that X-ray radiation by oxide films on the titanium surface was lower.That is, although the layer modified by the electrolytic plasma has the same thickness for all conditions, the oxide films have different thicknesses.This means that the oxide films are thinner where the titanium phase peaks were more intense.Another interesting aspect of the diffractogram is that the peaks referring to the anatase phase vary inversely with the electric current, while for rutile the opposite happens.
It is observed that for the same duty cycle, the rutile peak area increases with the current, while the anatase peak area decreases.For T on =50 µs, the decrease in anatase is more pronounced.It is known from the literature that the anatase   to rutile transformation occurs at temperatures between 450 and 850 •C, respectively 21 .
In the case of PEO, it is estimated that this happens when the electrolyte-film system is around 6 o C 22 .The temperature and reaction time of the compounds that generate important elements for film formation, such as Ca, P and O, are directly associated with the duration of the pulse repetition.They can occur by the following reactions 23 When titanium is dissolved in the vicinity of the sample (Equation 5), oxygen will be produced and may either evolve as a gas (Equation 6) or dissolves in solution to produce TiO 2 (Equation 7): Ti O TiO Thus, depending on the reaction temperature, anatase or rutile oxides can be produced.Samples oxidized with Ton = 50 µs and Toff = 100 µs tend to produce more anatase due to both the short duration of the microdischarge, 50 µs, and the cooling time of the molten zone by the microdischarge (100 µs).

Wettability
There is a growing body of literature showing the hydrophilic-hydrophobic conversion of flat titanium dioxide layers induced by irradiating the dry sample with light of a specific wavelength 24,25 .The wettability of TiO2 can also vary with topography 26 , temperature 27 , and stoichiometric composition 28 .On TiO2 surfaces produced by PEO, the size and distribution of pores, in addition to the chemical composition and crystal structure can be controlled.Greater wettability ensures more effective body fluid contact, increasing the initial interaction of proteins with the surface.By controlling the values of polar and non-polar components, one can control the selectivity of fibrinogen and albumin proteins 29 .The average values of the contact angles observed in our samples (Figure 6) ranged from 55.4 o to 106.8 o , which means a wide range of character from hydrophilic to hydrophobic.A full discussion of changes in wettability and the influence on the osseointegration and/or drug release lies beyond the scope of this study.Our interest is to advance in this area, carrying out a systematic study of the variation in the wettability of TiO2 surfaces produced by PEO, when we vary the temperature and surface morphology, irradiating the surface with different UV-Vis wavelengths.

Conclusion
The objective of the present research was to examine the use of PEO with short current pulses (50 and 100 µs) on titanium surfaces, aiming at application in implants, including those with drug delivery systems.This study showed that it was possible to control pore size and distribution, crystalline phase, Ca/P ratio and, still preliminary, wettability.Current data highlights the importance of using smaller pulse widths (50 µs) to achieve greater pore size and uniformity.These results are important because they have pore diameters within the range indicated by the literature for the best cell adhesion response and their values are uniform, providing a homogeneous release of drugs.By reducing the pulse width, it was also possible to obtain a greater amount of anatase, which has better photocatalytic properties.These surfaces, when properly excited by optical radiation, can selectively release drugs.This work contributes to the existing knowledge of new surface modification technologies aimed mainly at biomedical applications.

Figure 3 .
Figure 3. Analyzed scan region, containing the 5 points for punctual analysis.

Figure 6 .
Figure 6.Result of the average values of the contact angles measured and calculated from three samples for each experimental condition.

Table 1 .
Experimental conditions for electrolytic plasma oxidation.

Table 2 .
Elemental analysis by EDS of oxidized samples and stoichiometric Ti/O and Ca/P ratios.