Biodegradable Ceramics Consisting of Hydroxyapatite for Orthopaedic Implants

: This study aims to analyze hydroxyapatite (HAP) coatings enriched with Mg and Ti prepared by a magnetron sputtering technique on Ti6Al4V substrate. For preparation of the coatings, three magnetron targets (HAP, MgO and TiO 2 ) were simultaneously co-worked. The concentration of Mg added was varied by modifying the power applied to the MgO target. In all coatings, the Ti concentration was maintained constant by keeping the same cathode power fed during the whole deposition. The inﬂuence of different Mg dopant contents on the formation of phase, microstructure and morphology of the obtained Ti-doped HAP coatings were characterized by Fourier transform infrared spectroscopy (FTIR) and scanning electron microscopy (SEM). Moreover, the effects of Mg addition upon corrosion, mechanical and biological properties were also investigated. Mg- and Ti-doped HAP coating obtained at low radio-frequency (RF) power fed to the MgO target provided material with high corrosion resistance compared to other coatings and bare alloy. A slight decrease in hardness of the coatings was found after the Mg addition, from 8.8 to 5.7 GPa. Also, the values of elastic modulus were decreased from 87 to 53 GPa, this being an advantage for biomedical applications. The coatings with low Mg concentration proved to have good deformation to yielding and higher plastic properties. Biological test results showed that the novel surfaces exhibited excellent properties for the adhesion and growth of bone cells. Moreover, early adherent vital cell numbers were signiﬁcantly higher on both coatings compared to Ti6Al4V, suggesting that Mg ions may accelerate initial osteoblast adhesion and proliferation.


Introduction
Calcium phosphate (CaP) coatings promote bone growth on the interface between a bone and an implant [1]. Therefore, CaPs are commonly used to coat orthopedic and dental implants [2][3][4]. CaP-based coatings, such as hydroxyapatite (HAP) and tricalcium phosphate (TCP), are known to have good bioactive ability and, consequently, to promote the regeneration of bones [5]. HAP has

Elemental Composition, Chemical Binding, Morphology and Mechanical Properties
Evidence of elemental composition was provided through energy-dispersive X-ray spectroscopy (EDS-SEM-TM3030Plus, Bruker, Berlin, Germany). Fourier transform infrared spectroscopy (FTIR) was performed using a Jasco FTIR 6300 spectrometer (Tokyo, Japan) at a resolution of 4 cm −1 and 150 scans, using the transmission configuration for the identification of the chemical groups. Topography was evaluated by an atomic force microscope (AFM) (INNOVA, Veeco, Berlin, Germany) running in the tapping mode, from 3 × 3 µm 2 area scans. Based on AFM images, the following roughness parameters were calculated: root-mean-square roughness (rms), skewness (S k ) and kurtosis (K).
A surface profiler (Dektak 150, Bruker, Berlin, Germany) was used for determining the roughness parameters. The testing conditions were: investigated area 800 µm × 800 µm, a low-inertia stylus sensor with 12.5 mm radius, 5 mg contact force, and 20 µm/s scan speed. The R a (average roughness), rms and S k parameters were calculated on each coating before and after corrosion measurements.
Nanoindentation tests were performed with a NanoTest device (Fischer-Cripps Lab. Pty. Limited, Sydney, Australia), using a Berkovich indenter (face angle of 65.3 • ). For calculating penetration depth (h), elastic modulus (E), and hardness (H), the load-unload curves were applied from 1.0 to 5.0 mN. The position of each measure was randomly selected. The holding time at the maximum load was 5 s in order to reduce the creep effect, which could affect the shape of the unload curve and, at the end, the values of the elastic modulus and hardness of the HAP coatings. The values of E and H were averaged over five measurements. The values of E and H were extracted using the method reported by Oliver and Pharr [23].

In Vitro Corrosion Behaviour
The corrosion behaviour was observed by electrochemical tests in simulated body fluid (SBF) solution (composition presented in [24] pH = 7.2) at 37 ± 0.4 • C, using a PARSTAT 4000 system (Princeton Applied Research, Princeton, NJ, USA). Each sample (working electrode) was inserted in a Teflon holder and an area of 1 cm 2 was in contact with corrosive solution. The electrochemical tests were operated using a Pt foil as counter electrode and saturated calomel (SCE) as reference electrode. The scanning rate was of 1 mV/s, a value recommended by standard ASTM G5-94 [25] (reapproved 1999). Also, a magnetic microstirrer (VelpScientifica, Bohemia, NY, USA) with speed range 300 rpm

In Vitro Biological Assessment
Human osteogenic sarcoma cells (SaOS-2) were purchased from the American Type Tissue Collection (ATCC HTB 85, Manassas, VA, USA) and cultured in 85% McCoy's 5A medium, supplemented with 15% fetal bovine serum and 1% penicillin/streptomycin at 37 • C and 5% CO 2 . Cells at densities ranging from 10,000 (3 days) to 15,000 (1 day) cells/cm 2 were seeded onto coated or pure Ti6Al4V discs (15 mm diameter, 1 mm thick). The medium was changed every second day.
Cell viability was assessed by live/dead fluorescence staining using calcein-AM (Invitrogen, Waltham, MA, USA) and propidium iodide (PI, Sigma, Schnelldorf, Germany). On day 1 or 3 after seeding, cells were washed with HEPES buffered saline (HBSS) and stained with calcein-AM (2 µM) and PI (4 µM) in HBSS for 30 min at room temperature (RT) in the dark. Thereafter, staining solution was removed and samples were mounted upside down into coverslip bottom dishes (SPL, Gyeonggi-do, South Korea) using a drop of ProLong Diamond antifade solution (Molecular Probes, Eugene, OR, USA). Samples were immediately photographed and analyzed with Axiovision software (version 4.8) and an Axiovert 200 microscope (Zeiss, Göttingen, Germany) using appropriate filters. For each sample, at least 5 random microscopic visual fields were considered for cell counts.

Chemical Binding and Elemental Composition
Elemental composition was analyzed on coating deposited on Si substrate for eliminating the effect of the Ti content from the Ti6Al4V alloy substrate. The EDS results are illustrated in Table 2. The Ca/P ratio is not affected by Mg or Ti addition. The Mg content increases as RF power is fed to the MgO target, as was intended. The FTIR spectra of the coatings are displayed in Figure 1. For all coatings, the bands specific to HAP, originating from vibrations of PO 4 3− groups (1011-1040 cm −1 ), were observed. The FTIR spectrum of undoped HAP revealed sharp and well-defined bands placed at 1030 cm −1 and 1080 cm −1 , which can be attributed to (ν 3 ) symmetrical stretching vibration of the PO 4 3− group [26]. As can be seen from Figure 1, the prominent PO 4 3− band at 1030 cm −1 exhibited a gradual blue-shift up to 10 cm −1 , becoming broader when the Mg ion concentrations were increased in the Ti-doped HAP coating. This result confirms that some structural disorder had occurred during the Mg addition. The IR spectra corresponding to TM-0, TM-1 and TM-2 coatings showed the presence of a weak band at 1011 cm −1 , which is attributed to the ν 3− PO 4 3− group [27]. In the literature, it has been reported that the band found at 1030 cm −1 can be attributed to apatitic phosphate, while the band at 1010 cm −1 to non-apatitic phosphate [28,29]. We also found some characteristic peaks in all spectra at around 565 cm −1 and 717 cm −1 , which can be attributed to the ν 4 PO 4 3− and P 2 O 7 4− groups, respectively.
The presence of the PO 4 3− peaks in all the doped coatings demonstrates the apatite phase formation. Weak infrared (IR) bands corresponding to the CO3 2− group were visible in the spectral interval from 1650 to 1300 cm −1 . The presence of C-O bands is probably as a result of CO2 absorption from the atmosphere. The relative intensities of the bands corresponding to the OH − group were observed between 3600 and 3300 cm −1 , being more intense in the case of the TM-1 and TM-2 coatings. The reason is probably that the Ca 2+ is replaced by Mg 2+ and the increase in Mg concentration is compensated by OH − ions, or those of HPO4 2− . A weak band of P2O7 4− was found in all coatings, probably due to the conversion of HPO4 2− . In the case of doped coatings, it will be a competition between Mg 2+ and Ca 2+ to bond with anions, leading to a deficiency of Ca and a presence of water, and HPO4 2− is predicted. This effect was also found for the Zn-doped hydroxyapatite coatings [30]. Tank et al. [30] explained this effect as one in which stable interstitial protons form due to the attachment of oxygen of O-H or PO4 3− groups, leading to the formation of water or HPO4 2− . Han et al. reported that the OH − groups found between 3600 and 3300 cm −1 can be a sign of the presence of the HAP phase [31]. In the case of the TM-1 and TM-2 coatings, a weak band at ~640 cm −1 was seen, appearing probably due to the substitution of the OH − group by the CO3 2− group. This finding was also found by Antonakos et al. [32]. Figure 2 shows the AFM 3D images acquired on a 3 × 3 μm 2 area of uncoated and coated Ti6Al4V substrates. For comparison, Ti-doped HAP coating without Mg was also presented. Note that all of the coatings have smooth surfaces with uniform growth. Based on AFM images, the following roughness parameters were calculated: rms, Sk and K, being illustrated in Figure 3.  Weak infrared (IR) bands corresponding to the CO 3 2− group were visible in the spectral interval from 1650 to 1300 cm −1 . The presence of C-O bands is probably as a result of CO 2 absorption from the atmosphere. The relative intensities of the bands corresponding to the OH − group were observed between 3600 and 3300 cm −1 , being more intense in the case of the TM-1 and TM-2 coatings. The reason is probably that the Ca 2+ is replaced by Mg 2+ and the increase in Mg concentration is compensated by OH − ions, or those of HPO 4 2− . A weak band of P 2 O 7 4− was found in all coatings, probably due to the conversion of HPO 4 2− . In the case of doped coatings, it will be a competition between Mg 2+

Morphology
and Ca 2+ to bond with anions, leading to a deficiency of Ca and a presence of water, and HPO 4 2− is predicted. This effect was also found for the Zn-doped hydroxyapatite coatings [30]. Tank et al. [30] explained this effect as one in which stable interstitial protons form due to the attachment of oxygen of O-H or PO 4 3− groups, leading to the formation of water or HPO 4 2− . Han et al. reported that the OH − groups found between 3600 and 3300 cm −1 can be a sign of the presence of the HAP phase [31].
In the case of the TM-1 and TM-2 coatings, a weak band at~640 cm −1 was seen, appearing probably due to the substitution of the OH − group by the CO 3 2− group. This finding was also found by Antonakos et al. [32]. Figure 2 shows the AFM 3D images acquired on a 3 × 3 µm 2 area of uncoated and coated Ti6Al4V substrates. For comparison, Ti-doped HAP coating without Mg was also presented. Note that all of the coatings have smooth surfaces with uniform growth. Based on AFM images, the following roughness parameters were calculated: rms, S k and K, being illustrated in Figure 3.  Figure 2 shows the AFM 3D images acquired on a 3 × 3 μm 2 area of uncoated and coated Ti6Al4V substrates. For comparison, Ti-doped HAP coating without Mg was also presented. Note that all of the coatings have smooth surfaces with uniform growth. Based on AFM images, the following roughness parameters were calculated: rms, Sk and K, being illustrated in Figure 3.  The bare Ti6Al4V alloy showed rms roughness of approximately 12 nm. The roughness of coated sample decreased to 5.8 nm for Ti-doped HAP (TM-0). After a small Mg content (TM-1), a slight increase in roughness was observed (9.6 nm). The increment of RF power fed to the MgO target led to an increase in roughness to 18.9 nm (TM-2).

Morphology
Usually, flat surfaces have skewness parameters with values close to zero. If the skewness has positive values, the surfaces consist of a few valleys and lots of peaks. In our case, the coating without Mg addition (TM-0) had a negative skewness, indicating a lot of valleys, probably due to the polishing substrates. Both coatings with Mg addition exhibited a positive skewness, being close to zero for the TM-2 coating, indicating that this coating is uniformly distributed on the entire substrate and most of the substrate valleys were covered.
The uncoated substrate has a high positive kurtosis, proving that the surface has high peaks and low valleys, which appeared during the polishing process. Gadelmawla et al. reported that the surfaces with few high peaks and low valleys presented kurtosis <3, and >3 for surfaces with many high peaks and low valleys [33]. In this context, the bare alloy exhibited a kurtosis value of about 8.1, which is higher than 3. The Ti-and Mg-doped coating (TM-1), prepared at low RF power fed to the MgO cathode, revealed a positive kurtosis value of less than 3, approximately 1, indicating quite a small number of high peaks and low valleys, meaning that the peaks are bumpy. The Mg-doped  The bare Ti6Al4V alloy showed rms roughness of approximately 12 nm. The roughness of coated sample decreased to 5.8 nm for Ti-doped HAP (TM-0). After a small Mg content (TM-1), a slight increase in roughness was observed (9.6 nm). The increment of RF power fed to the MgO target led to an increase in roughness to 18.9 nm (TM-2).
Usually, flat surfaces have skewness parameters with values close to zero. If the skewness has positive values, the surfaces consist of a few valleys and lots of peaks. In our case, the coating without Mg addition (TM-0) had a negative skewness, indicating a lot of valleys, probably due to the polishing substrates. Both coatings with Mg addition exhibited a positive skewness, being close to zero for the TM-2 coating, indicating that this coating is uniformly distributed on the entire substrate and most of the substrate valleys were covered.
The uncoated substrate has a high positive kurtosis, proving that the surface has high peaks and low valleys, which appeared during the polishing process. Gadelmawla et al. reported that the surfaces with few high peaks and low valleys presented kurtosis <3, and >3 for surfaces with many high peaks and low valleys [33]. In this context, the bare alloy exhibited a kurtosis value of about 8.1, which is higher than 3. The Ti-and Mg-doped coating (TM-1), prepared at low RF power fed to the MgO cathode, revealed a positive kurtosis value of less than 3, approximately 1, indicating quite a small number of high peaks and low valleys, meaning that the peaks are bumpy. The Mg-doped The bare Ti6Al4V alloy showed rms roughness of approximately 12 nm. The roughness of coated sample decreased to 5.8 nm for Ti-doped HAP (TM-0). After a small Mg content (TM-1), a slight increase in roughness was observed (9.6 nm). The increment of RF power fed to the MgO target led to an increase in roughness to 18.9 nm (TM-2).
Usually, flat surfaces have skewness parameters with values close to zero. If the skewness has positive values, the surfaces consist of a few valleys and lots of peaks. In our case, the coating without Mg addition (TM-0) had a negative skewness, indicating a lot of valleys, probably due to the polishing substrates. Both coatings with Mg addition exhibited a positive skewness, being close to zero for the TM-2 coating, indicating that this coating is uniformly distributed on the entire substrate and most of the substrate valleys were covered. The uncoated substrate has a high positive kurtosis, proving that the surface has high peaks and low valleys, which appeared during the polishing process. Gadelmawla et al. reported that the surfaces with few high peaks and low valleys presented kurtosis <3, and >3 for surfaces with many high peaks and low valleys [33]. In this context, the bare alloy exhibited a kurtosis value of about 8.1, which is higher than 3. The Ti-and Mg-doped coating (TM-1), prepared at low RF power fed to the MgO cathode, revealed a positive kurtosis value of less than 3, approximately 1, indicating quite a small number of high peaks and low valleys, meaning that the peaks are bumpy. The Mg-doped coating (TM-2), prepared at high RF power fed to the MgO cathode, showed a negative kurtosis very close to 0, signifying that the surface is formed of few high peaks and low valleys, meaning that many sharp peaks were found on the surface.
In the literature, it is stated that skewness and kurtosis could be used to estimate the corrosion resistance and friction performance of a surface [34,35]. Sedlacek et al. reported that a more negative skewness and a high kurtosis imply a low friction performance under dry sliding, due to the deep valleys, which act as wear particle traps [34]. Furthermore, Sedlacek et al. obtained the same results also under lubricated sliding, where the deep valleys play the role of micro reservoirs, intensifying the lubrication process [34]. The coatings under investigation in this paper exhibited kurtosis of less than 1, indicating that the surface consists of a few high peaks and low valleys, which can imply a good friction performance. Good corrosion resistance was found when the skewness had more positive values, while negative skewness indicated pitting corrosion [35]. In our case, both Mg-doped HAP coatings proved to have positive skewness, and were expected to have superior corrosion resistance than others.
Based on the results, one may conclude that the TM-1 coating could be a possible candidate for increasing the corrosion resistance and friction performance of the HAP coatings.

Mechanical Properties
Knowledge of the mechanical properties of HAP coatings is necessary in order to anticipate their performance when employed, and in particular to draw comparisons between their performance and that of natural bones [36]. HAP demonstrates unsatisfactory mechanical properties, principally in corrosive surroundings such as those found in the human body. Important progress has been reported into the microstructures of biocompatible ceramics, aimed at achieving implants with suitable mechanical properties used in hard tissue replacement [37]. Nowadays, studies are being focused on the modification of HAP by the addition of other phases or elements in order to obtain enhanced mechanical and corrosive properties [38]. To show the modifications that materials undergo, hardness is one critical parameter for evaluating the mechanical properties of a material [39]. In our previous study, it was reported that the addition of Ti to the HAP structure caused an increase in mechanical properties, as a result of the lattice deformation caused by the Ti atoms being added to the base HAP cell unit [13].
For this study, load-unload measurements from 1.0 to 5.0 mN, with a step size (dP) 0.5 mN, were performed. By growing load from 1.0 to 5.0 mN, the penetration depth increases. Thus, for preventing the influence of substrate properties, we established that the best load to investigate our coatings was 1.0 mN. For this load, the depth of the indenter is less than 10-20% of the coating thickness. The maximum penetration depth achieved at the end of the loading is less significant for the coating without Mg addition (TM-0), suggesting that the Mg addition does not have any favorable influence on the mechanical properties of Ti-doped HAP coatings. The E and H values obtained from load-unload measurements at 1.0 mN are summarized in Table 3. It can be observed that both the H and E of the coated samples decrease with Mg addition in the Ti-doped HAP structure. The H/E ratio is called "plasticity index" and defines the elastic strain of a material to its failure [40]. A high value of this ratio indicates good elastic properties and good adhesion of the coating to the substrate. By taking into consideration this ratio, one may see that the addition of Mg into Si-based HAP coatings improves the elastic ability of HAP coatings. Moreover, a high H/E ratio was found for TM-2, indicating good adhesion to the Ti6Al4V substrate. The H 3 /E 2 ratio is an indicator that a material resists plastic deformation [40]. A high value indicates a good elastic behaviour. The TM-1 coatings exhibited the highest H 3 /E 2 ratio, meaning the best elastic properties, followed by TM-0 and TM-2. The H 2 /2E ratio describes resistance to the effect of external forces, known as the modulus of resilience [40]. It can be seen that the TM-0 has the highest H 2 /2E ratio, indicating a good ability to recover after applying external forces. It can be observed that the TM-2 exhibited high H/E and H 3 /E 2 ratios, indicating good deformation to yielding, while TM-1 showed high H 3 /E 2 and H 2 /2E ratios, meaning higher plastic properties.

In Vitro Electrochemical Behaviour
The human body is a corrosive environment for any artificial materials inserted into it, especially for metallic surfaces such as implants. For this reason, it is very important to investigate the behaviour of artificial material in contact with solutions that mimic human physiological solutions. Electrochemical tests are the most common form of analysis for evaluating a material's behaviour in corrosive solution.
The evolution of the open circuit potential (E OC ) after 6 h of immersion in SBF at 37 • C is shown in Figure 4. Tafel curves of the coatings and uncoated Ti6Al4V substrate are presented Figure 5. Electrochemical parameters, corrosion potential (E i=0 ) and corrosion current density (i corr ) were extracted from Tafel curves by the extrapolation technique, and the results are presented in Table 4. Taking into account the electrochemical parameters, the porosity (P) was calculated, using Elsener's empirical equation [41]. The protective efficiency (P e ) was also determined by a formula reported in [24]. Mansfeld [42] and Scully [43] described how the E OC potential is the parameter that is related to the nobility of a surface at contact with a corrosive medium. According to Figure 4, for all coatings the potential exhibited some oscillation at the beginning of the test, indicating higher film stability to breakdown. After 2 h of immersion, it can be noted that there are fewer oscillations, suggesting that all the investigated surfaces started to be covered by stable protective oxides which are resistant to chemical dissolution in SBF solution. Both Mg-doped coatings have a similar evolution of E OC potential after 2 h, having a more electropositive E OC value than of uncoated Ti6Al4V alloy. TM-0 exhibited more electronegative E OC value, demonstrating that this coating is more or less affected by the SBF solution. Table 4. Electrochemical parameters of the coatings tested in simulated body fluid (SBF) solution at 37 • C (E i=0 , corrosion potential; i corr , corrosion current density; R p , polarization resistance; P, porosity; P e , protection efficiency). breakdown. After 2 h of immersion, it can be noted that there are fewer oscillations, suggesting that all the investigated surfaces started to be covered by stable protective oxides which are resistant to chemical dissolution in SBF solution. Both Mg-doped coatings have a similar evolution of EOC potential after 2 h, having a more electropositive EOC value than of uncoated Ti6Al4V alloy. TM-0 exhibited more electronegative EOC value, demonstrating that this coating is more or less affected by the SBF solution.    the potential exhibited some oscillation at the beginning of the test, indicating higher film stability to breakdown. After 2 h of immersion, it can be noted that there are fewer oscillations, suggesting that all the investigated surfaces started to be covered by stable protective oxides which are resistant to chemical dissolution in SBF solution. Both Mg-doped coatings have a similar evolution of EOC potential after 2 h, having a more electropositive EOC value than of uncoated Ti6Al4V alloy. TM-0 exhibited more electronegative EOC value, demonstrating that this coating is more or less affected by the SBF solution.   Corrosion resistance plays an essential role in biomedical applications, because any material, especially a metallic one, starts to degrade in contact with body fluids, leading to the formation of corrosion products that will affect the tissue around the implant. It is very important to modify the surface of metallic implants with coatings based on calcium and phosphorus in order to delay the corrosive process [44]. It is generally recognized that the corrosion resistance of metallic surfaces can be estimated by several criteria, as an electropositive (E i=0 ) and low i corr and high R p values reveal good anti-corrosive properties [45]. For all coatings, it is possible to say that these significantly improve the corrosion behaviour of the Ti6Al4V alloy. According to the i corr parameter, all investigated coatings have lower i corr values than uncoated Ti6Al4V alloy. The addition of Mg into the HAP-Ti structure also leads to a decrease in i corr value over undoped Mg coating (TM-0), whatever the deposition conditions, signifying an improvement of the resistance to corrosion. An increment of RF power fed to the MgO target has a significant influence on the electrochemical characteristics of the Ti-doped HAP coatings, leading to the deterioration of corrosion performance.

Sample
Comparison of the R p values showed that the coating with added Mg revealed a higher polarization resistance in SBF than that of uncoated Ti6Al4V alloy and coating without Mg (TM-0). Regarding the porosity (P), it can be observed that the values of the Mg-doped coatings were not so different. Somehow, the increase of Mg content leads to a slight increase of porosity. The coating with Mg addition has the highest porosity value. This effect can be an advantage for cell growth, proliferation and differentiation. In regard to protective efficiency (P e ), the best value was found for the TM-1 sample (78.1%), a coating with low Mg content. It seems that the Mg increment does not enhance resistance to the corrosive effect of SBF.
The electrochemical results of Mg-doped HAP-Ti coatings revealed that the electrochemical parameters increase with the increase of surface roughness parameters (R a ). Therefore, it can be concluded that roughness can be regarded as an important parameter effect on corrosion resistance. Consequently, the specimens coated with Mg-doped coatings presented a superior barrier effect to preserve the uncoated substrate and also to the coating without Mg addition. Corrosion involves the penetration of Cl − ions and H 2 O through the depth found on coated surfaces and accelerates the electrochemical reactions that take place at the substrate-coating interface [46]. A surface profiler (Dektak 150, Bruker, Berlin, Germany) was used to evaluate the roughness of the corroded surfaces on a large area of 800 µm × 800 µm. The R a , rms and S k parameters were calculated for each investigated sample ( Table 5). Roughness of the samples before electrochemical tests was also measured in order to compare the effect of the corrosion tests. According to the roughness parameters, the TM-0 sample is more affected by the SBF solution than the TM-1 or TM-2, indicating that Mg addition improves the corrosion resistance of Ti-doped HAP in SBF. Note that the TM-1 exhibited a skewness value close to zero, signifying that the hills found before corrosion were almost leveled, and leading to a flat surface. TM-1 and TM-2 samples revealed small differences in roughness parameters before and after corrosion compared to those of uncoated Ti6Al4V and TM-0, showing that the introduction of Mg can reduce the dissolution rate of the HAP coating in the SBF solution. As an example, Figure 6 shows 3D mapping of the TM-1 coating, which proved to have the best corrosion resistance, before and after electrochemical tests in the SBF solution. One may observe that the surface of the TM-1 coating is not affected by the SBF attack. This effect corresponds with the electrochemical results.

In Vitro Biological Properties
SaOS-2 cells were seeded directly on standard titanium alloy (Ti6Al4V) or the coated surfaces in order to investigate bone cell adhesion and morphology. Figure 7 shows that attached cell numbers seemed to be higher on TM-1 and TM-2 than on Ti6Al4V, indicating preferable adhesion conditions on the novel coatings. The relative densities of living cells (Table 6) proved those differences to be significant after the first day of seeding. On day 3, however, cell numbers were similar on all three surfaces. SaOS-2 osteoblasts displayed their typical polygonal shape with prominent actin stress fibres visible on all tested surfaces.

In Vitro Biological Properties
SaOS-2 cells were seeded directly on standard titanium alloy (Ti6Al4V) or the coated surfaces in order to investigate bone cell adhesion and morphology. Figure 7 shows that attached cell numbers seemed to be higher on TM-1 and TM-2 than on Ti6Al4V, indicating preferable adhesion conditions on the novel coatings. The relative densities of living cells (Table 6) proved those differences to be significant after the first day of seeding. On day 3, however, cell numbers were similar on all three surfaces. SaOS-2 osteoblasts displayed their typical polygonal shape with prominent actin stress fibres visible on all tested surfaces. Figure 6. 3D mapping images of the TM-1 coating deposited on Ti6Al4V substrates: (a) Before electrochemical tests; (b) After electrochemical tests. The investigated area was 800 μm × 800 μm.

In Vitro Biological Properties
SaOS-2 cells were seeded directly on standard titanium alloy (Ti6Al4V) or the coated surfaces in order to investigate bone cell adhesion and morphology. Figure 7 shows that attached cell numbers seemed to be higher on TM-1 and TM-2 than on Ti6Al4V, indicating preferable adhesion conditions on the novel coatings. The relative densities of living cells (Table 6) proved those differences to be significant after the first day of seeding. On day 3, however, cell numbers were similar on all three surfaces. SaOS-2 osteoblasts displayed their typical polygonal shape with prominent actin stress fibres visible on all tested surfaces.    Live/dead fluorescence staining was used to determine osteoblast viability after 1 or 3 days in direct contact with standard alloy or the coated surfaces ( Figure 8). Virtually all cells were vital (green), indicating no cytotoxic distress exerted by the novel coatings. Notably, after three days of contact this assay also showed more attached vital cells per microscopic view on TM-1 and TM-2 coatings compared to the standard titanium alloy, thus supporting the outcome of the adhesion test.  Live/dead fluorescence staining was used to determine osteoblast viability after 1 or 3 days in direct contact with standard alloy or the coated surfaces ( Figure 8). Virtually all cells were vital (green), indicating no cytotoxic distress exerted by the novel coatings. Notably, after three days of contact this assay also showed more attached vital cells per microscopic view on TM-1 and TM-2 coatings compared to the standard titanium alloy, thus supporting the outcome of the adhesion test. In vitro biological results demonstrated that all tested surfaces are biocompatible, accepting the adhesion and growth of bone cells. Whereas cell viability is excellent on an uncoated alloy and the novel surfaces, both novel coatings seem to accelerate cell adhesion compared to Ti6Al4V. Regarding the cell viability, both novel coatings had very similar properties.
Magnesium plays an important role in the control of biochemical reactions, therefore influencing 1 day 3 days In vitro biological results demonstrated that all tested surfaces are biocompatible, accepting the adhesion and growth of bone cells. Whereas cell viability is excellent on an uncoated alloy and the novel surfaces, both novel coatings seem to accelerate cell adhesion compared to Ti6Al4V. Regarding the cell viability, both novel coatings had very similar properties.
Magnesium plays an important role in the control of biochemical reactions, therefore influencing crucial physiological functions such as nucleic acid metabolism, protein synthesis and energy production. It was expected that the coatings containing Mg would show lower cell numbers than of uncoated Ti6Al4V. It has already been demonstrated that cell adhesion and proliferation of MG63 osteosarcoma cells is inhibited in direct contact with pure Mg surfaces or Mg-based coatings [47,48]. A possible explanation for this could be due to the Mg ions being released from the surface of the implant. Degradation of Mg in contact with human physiological solution involves hydrogen decomposition and an elevated pH at the implant surface. Both factors can strongly affect initial osteoblast adhesion and delay cell spreading and proliferation [47]. In this study, the TM-1 coating was less affected by the SBF corrosive attack, probably leading to lesser hydrogen decomposition. This result might explain the slightly better cell spreading and proliferation on the TM-1 coating compared to TM-2.
On the other hand, extracellular Mg 2+ and Ca 2+ ions are both essential for the adhesion and proliferation of human osteoblasts and cannot compensate each other [49]. Although the cell culture medium contained both ions, the specific Mg 2+ concentration may be higher at the immediate interface between bone cells and implant surface due to magnesium ions released from the coating. This might accelerate initial cell adhesion on both Mg-doped coatings compared to Ti6Al4V alloy, leading to significantly higher numbers of adherent vital cells 24 h after seeding. This early positive effect of Mg 2+ on osteoblast proliferation seems later to be neutralized by hydrogen evolution and/or alkalinisation of the implant surface, as cell numbers on both Mg coatings were statistically indiscernible from those on Ti6Al4V on day 3.
The free Mg 2+ concentration is the regulatory agent for cell proliferation, leading to a coordinate physiological response, which is an important step immediately after the insertion of medical implants [50]. Rubin [50] further found that there is an optimal Mg 2+ concentration for DNA and protein synthesis. A lower or higher magnesium level will lead to a decline in both processes and thus will inhibit cell proliferation. Furthermore, Leid et al. showed that the formation of a calcium phosphate mineral matrix is affected by a high free Mg 2+ level [51]. In bone formation and healing, the formation of this mineral matrix is an important phase, and thus it is crucial to have the optimal amount of free Mg 2+ in the coating.

Conclusions
The consequence of Mg and Ti addition on the mechanical characteristics, corrosion behavior and cell viability of a HAP system was studied. The following conclusions can be drawn from this research: • Surface roughness of the coatings increases with Mg incorporation into a Ti-doped HAP structure. Additionally, with increasing RF power fed to the MgO cathode there is intensification in the roughness of the coating surface.

•
The decrease in the surface roughness plays an important role in increasing the corrosion resistance.

•
Corrosion results show that coating obtained at low RF power fed to the MgO target led to a decrease in the current density compared with other coatings and bare alloy, suggesting a high corrosion performance. • Mg-and Ti-doped HAP coatings obtained at low RF power fed to the MgO target provided the material with high protection efficiency (78.1%) to the corrosive SBF attack, indicating a higher corrosion resistance.
• It can be concluded that the elastic modulus of the coated samples decreases with Mg addition in the HAP structure, especially in the case of the coatings with high Mg concentration (53 GPa). • All tested surfaces are non-cytotoxic, as they support the adhesion and growth of bone cells.

•
Mg ions seem to accelerate initial osteoblast adhesion and proliferation.
The results obtained from this study demonstrate that the novel coatings are promising candidates for orthopaedic applications. Further in vivo tests will be needed for validation of the in vitro results.