Selective modification of Ti6Al4V surfaces for biomedical applications

The surface of a medical implant is required to interact favourably with ions, biomolecules and cells in vivo, commonly resulting in the formation of the extracellular matrix. Medical grade Ti6Al4V alloy is widely used in orthopaedic and dental applications for bone replacement due to its advantageous mechanical properties and biocompatibility, which enhances the adhesion between native tissue and the implanted material. In this study, chemical and thermal modification of a medical-grade Ti6Al4V alloy were performed to enhance electrostatic interactions at the alloy surface with a synthetic peptide, suitable for conferring drug release capabilities and antimicrobial properties. The modified surfaces exhibited a range of topographies and chemical compositions depending primarily on the treatment temperature. The surface wetting behaviour was found to be pH-dependent, as were the adhesive properties, evidenced by chemical force titration atomic force microscopy.


Introduction
Titanium and its alloys are excellent materials for orthopaedic and dental implants due to their mechanical properties, including those conferred by the surface such as resistance to corrosion and bioactivity. 1 Titanium surfaces present negative charges when in contact with aqueous solutions at physiological pH (around 7.4) attributed to the native titanium oxide (TiO x ). The composition of the native TiO x surface is primarily TiO 2 in amorphous phase. 2,3 It was shown that the crystal phases of anatase and rutile present on the surface offer better bioactivity than the amorphous phase [4][5][6] and therefore the presence of these phases are more desirable for cell proliferation. A method commonly used for the crystallisation of amorphous TiO 2 is thermal oxidation. During this method, the formation of crystal phases is highly dependent on the temperature. For example, the TiO 2 layer formed following thermal oxidation treatment at temperatures below 500 C is mainly amorphous with some anatase present. 7,8 Aer the treatment at a temperature of 550 C, both rutile and anatase crystal phases are present. 7,9,10 In contrast, oxidation treatment at temperatures over 800 C leads to the formation of the rutile phase only. 9,11 The surface topography and charge density of Ti and its alloys are properties that can be modied by chemical treatments. 7,[12][13][14] One approach to alter these surfaces chemically is to use H 2 O 2 in aqueous solution. This modication produces a titania gel layer (TGL) rich in OH À with nano and microporosity 8,15 which is desirable for the adsorption of ions, peptides, and proteins, as well as the promotion of cell adhesion. [16][17][18] Furthermore, the TGL on the surface can protect the supporting metal from corrosion in the biological environment. 7,18 This is particularly relevant to the use of titanium for implantation. This chemical treatment, combined with a thermal treatment, form a TiO x surface that can prevent the liberation of cytotoxic elements from the bulk titanium into the surrounding tissues. This is supported by Browne and Gregson, who found that the release of titanium and aluminium ions from the oxide lms on Ti6Al4V annealed at 400 C for 0.75 h was reduced from 0.14 to 0.03 mg cm À2 (Ti), and from 0.25 to 0.026 (Al) mg cm À2 , respectively, compared with the standard nitric-acid passivated surface. 19 Control over the chemical moieties, charge density, and topography of the surface can be utilised for a multitude of applications. For example, rough surfaces can lead to the formation and deposition of calcium phosphates. The latter intervenes as anchorage for collagenous bres produced by bone cells or osteoblasts, and in most of the cases helps to introduce biocompatibility. 20 A previous study reported that a Ti6Al4V surface treated by H 2 O 2 exhibited improved electrostatic adsorption of peptides; 21 however, the surface properties were not reported. This study aims to show how the thermal oxidation and chemical treatment with H 2 O 2 of a medical-grade Ti6Al4V surface can affect its chemical composition and topography in the view of the surface capability to interact with organic molecules such as peptides for drug delivery applications and antimicrobial properties.

Titanium alloy surface sample preparation prior to surface modications
Single-sided-polished Ti6Al4V plates (ELI ASTM F136-2002a) were supplied by Titanium Industries UK Ltd. The bulk material composition was reported by the manufacturer and was: wt% in the top O 0.13%, V 4.12%, C 0.02%, Ti >88.854%, N 0.02%, Al 6.42%, H 0.006%, Fe 0.03% and in the bottom O 0.13%, V 4.14%,C 0.02%, Ti 88.864%, N 0.02%, Al 6.38%, H 0.006%, Fe 0.04%. For both sides, each residual elements were <0.10% and the residual elements total were <0.40%. The method of characterization is described in ASTM F136-13 Standard. 22 The Ti6Al4V plates were cut using a guillotine into square plates of dimensions 10 mm Â 10 mm. The 1.6 mm thick plates were then polished until a mirror-like nish following the Struers polishing protocol. 23 The plates were then washed following 15 min of consequent ultrasonic baths with deionised water (dH 2 O), acetone, and nally, dH 2 O, followed by a drying process in air at room temperature. This is the initial surface procedure used prior to the chemical and thermal modications and is denoted by MPT.
2.1.1 Chemical treatment. A mixture of 90 v/v % of 9.78 M H 2 O 2 supplied by Sigma Aldrich (chemical grade), and 10 v/v % of 1 M HCl supplied by Merck Millipore (analytical grade) was prepared. This solution was used to etch the surface of the metallic samples. Cleaned and dried Ti6Al4V samples were reacted with 10 mL of the etching solution at 80 C in an oven (Pickstone, UK) for 30 min. The plates were then rinsed with dH 2 O and dried in air at room temperature.
2.1.2 Thermal oxidation in air. Untreated and chemically treated Ti6Al4V plates were thermally oxidised in an air furnace (Lenton, UK) at 500, 600, 700 and 800 C. Upon reaching the set point temperature, the samples were further heat-treated for 1 h. The samples were then removed from the oven and le to cool in the air at room temperature. The samples were stored under air in sealed containers. All analyses were performed within 90 days of the modication procedure. Samples description and abbreviations are in Table 1.

Characterization
2.2.1 Scanning electron microscopy analysis. SEM micrographs of the untreated and treated surfaces were obtained using a JSM-7000F SEM (JEOL, Japan), operating with an electron voltage of 20 kV.
2.2.2 X-ray diffraction analysis. X-ray diffractograms of untreated and treated surfaces were obtained using a D8 Advance XRD (Bruker, Germany) with Bragg Brentano mode and Cu Ka (l ¼ 1.64 A) in the 2q between 20-60 using a step size of 0.02 . The obtained data were analysed using Match soware (Crystal Impact, Germany).
2.2.3 Raman spectroscopy analysis. Raman spectroscopy of untreated and treated surfaces was performed using a LabRAM HR 800 Raman microscope (Horiba Jobin Yvon, Japan) with a laser source of wavelength 532 nm. The analysed areas were 60 mm Â 70 mm, and the data were acquired for 45 s, in three nonoverlapping locations. The data were analysed by LabSpec 5 soware (Horiba, Japan).
2.2.4 AFM imaging analysis. AFM images of untreated and treated surfaces were obtained using a NanoWizard II AFM (JPK Instruments, UK) operating in intermittent contact mode at 18 C and relative humidity of <40%. Rectangular pyramidaltipped Si cantilevers (PPP-NCL, Windsor Scientic, UK) were employed (nominal length and width were 225 mm and 38 mm respectively; their nominal tip diameter was <10 nm). The nominal frequency of the rst resonant mode was 190 kHz. A tip velocity of 10 mm s À1 was employed when scanning across the sample surface. Each image was composed of 512 Â 512 pixels and took approximately 20 min to acquire. Images were analysed using a JPK Data Processing soware (JPK Instruments, UK) and a Scanning Probe Image Processor soware (Image Metrology, Denmark).
2.2.5 Dynamic contact angle analysis. Dynamic contact angle (DCA) data of all surfaces were obtained using a Theta Table 1 Sample list description and abbreviation

Ti6Al4V sample treatment Abbreviation
Polished surfaces MPT Thermally oxidised TO Chemically treated TGL Thermally treated TO n (n ¼ 500, 600, 700 and/or 800 C) Chemically and thermally treated TGL-TO n (n ¼ 500, 600, 700 and/or 800 C) Fig. 1 Description of measured forces in the force-distance curves obtained by AFM. In red, the approach curve and in blue, the retract curve. F rep is the repulsion force, F JT is the jump to force, and F PO is the pull-off force.
OneAttension optical tensiometer (Biolin Scientic, Sweden). Advancing contact angle (q a ) data were obtained in triplicate, on three different samples, with acidic and alkaline solutions at pH 3, 5, 7 and 9. The drop size was 2 mL, and the rate of volume change was 0.5 mL s À1 . Data were analysed with a Young-Laplace curve tting, using OneAttension soware (Biolin Scien-tic, Sweden). 2.2.6 AFM adhesion analysis. Acquisition of adhesion data was performed using a NanoWizard II AFM (JPK Instruments, UK) employing a CellHesion module (JPK Instruments, UK), operating in force spectroscopy mode at 18 C. The sample and cantilever were immersed in acidic and alkaline aqueous solutions, contained within a clean glass Petri dish. The solutions were manufactured using dH 2 O and pH adjusted with NaOH 0.01 M or HCl 0.01 M, to achieve solutions of pH 3, 5, 7 and 9. The samples were immobilised using a double-sided Shintron adhesive tape (Agar Scientic, UK). The vertical deection and z-axis displacement data were recorded at a frequency of 10 kHz. A grid of 100 forcedisplacement curves was acquired for each sample/liquid combination, equally spaced over an area of 100 mm Â 100 mm. The force-displacement data were analysed using JPK Data Processing soware (JPK Instruments, UK).
A schematic of the measured forces is shown in Fig. 1. The out-of-contact repulsion force (F rep ) was measured in the rst peak from right to le of the approaching curve, the height of the triangle that was formed in the peak from the origin to the highest part of the peak until the signal was stable in the approach curve was measured as the force. The jump to force (F JT ) was measured in the minimum point of the approach curve, and the pull-off force (F PO ) was measured in the minimum point of the retract curve.
2.2.7 X-ray photoelectron spectroscopy analysis. The surface composition of MPT, TO, TGL and TGL-TO were investigated using X-ray photoelectron spectroscopy (XPS). The analysis was performed using a Thermo Scientic™ K-Alpha™ X-ray Photoelectron Spectrometer with a monochromatic Al K a (1486 eV) X-ray source using charge compensation and a 400 mm spot size. Survey spectra were recorded with a pass energy of 200 eV and a step size of 0.4 eV, while high-resolution spectra were recorded with a pass energy of 40 eV and a step size of 0.1 eV. Three non-overlapping regions were measured for each sample. The data was processed in CasaXPS (V 2.3.16).
Peaks in the survey spectra were assigned using the Scoeld library, and a linear background was used for all peaks. Deconvolution of the Ti 2p peaks obtained from the highresolution scans was performed according to the literature 24,25 applying two adjacent regions ranging from 451.79 eV to 461.5 eV and 461.5 to 468.0, respectively. For both regions, a Shirley background with an average width of 2 was used, and the position at 461.5 eV was offset by a factor of 18. 8 components relating to the four oxidation states of Ti (0, II, III and IV) and the doublets for the 2p 1/2 and 2p 3/4 states expected for each of these four peaks were tted in all spectra. For the untreated Ti6Al4V surfaces, the positions of all peaks except the two prominent Ti(IV) and the Ti(0) 2p 3/2 peaks were constrained according to literature values. 24,25 The Ti(0) 2p 1/2 peak was restrained to +6.1 eV from the Ti(0) 2p 1/2 peak. For all other samples, only the two Ti(IV) peaks were not constrained, all other position constraints were kept constant. ESI Table 1 shows a summary of all constraints used for the treated surfaces. †

Surface topography
The topography of the polished Ti6Al4V surfaces (MPT), thermally oxidised Ti6Al4V surfaces (TO 500 C , TO 600 C , TO 700 C , TO 800 C ), H 2 O 2 /HCl chemically treated Ti6Al4V (TGL), and both chemically treated and thermally oxidised Ti6Al4V (TGL-TO 500 C , TGL-TO 600 C , TGL-TO 700 C , TGL-TO 800 C ) are shown in Fig. 2. From the SEM micrographs, it is apparent that the surface topography increased with increasing TO temperature. The roughness values of these surfaces measured by AFM are Table 2 Roughness analysis of surface topography measurements for MPT, TO, TGL and TGL-TO Ti6Al4V modified surfaces by AFM shown in ESI Fig. 1. In the first column is the sample name, in the second column, the arithmetical average roughness (S a ) and in the third column, the root mean square average roughness (S q ). Measurements were performed using a scan window of dimensions 10 Â 10 mm at a resolution of 512 Â 512 pixels    listed in Table 2. It was observed that the surface roughness increased with increasing TO temperature. In addition, TGL-TO Ti6Al4V surfaces presented a larger roughness compared to the TGL Ti6Al4V surface. An SEM micrograph of the TGL Ti6Al4V surface showing its topography is also presented in Fig. 3. The average roughness (S a ) of the modied Ti6Al4V surfaces ranged between 0.7 nm (MPT) to 121 nm (TGL-TO 800 C ).

Spectroscopic analysis
The XRD diffractograms of the samples with and without chemical treatment aer TO with their respective controls (MPT and TGL) are shown in Fig. 4  The Raman spectra of anatase, rutile and samples with and without chemical treatment aer thermal oxidation are shown in Fig. 6 and 7, respectively. Active vibrational modes of anatase in the spectrum are assigned to E g at 140 cm À1 , E g at 196 cm À1 , B 1g at 395 cm À1 , A 1g at 515 cm À1 , and E g 639 cm À1 . 26 Active vibration modes of rutile are assigned to B 1g at 142 cm À1 , E g at 447 cm À1 , A 1g at 610 cm À1 , and E 2g 812 cm À1 . 27 The vibrational modes of Ti-O bonds were associated with the vibrational modes of anatase for the heat treatments at 500 and 600 C and associated with rutile for the heat treatments at 700 and 800 C for all Ti6Al4V modied surfaces.

Chemical composition
XPS spectra of MPT, TO, TGL, and TGL-TO Ti6Al4V modied surfaces are shown in ESI Fig. 2, † and the elemental composition obtained from these spectra is listed in Table 3. At the lower TO temperatures used, the elemental composition of the sample surface was close to that of the untreated surfaces, while higher TO temperatures altered the elemental composition notably for both the TO and TGL-TO Ti6Al4V modied surfaces. For both types of surface modication, Ti% and C% decreased when the TO temperature increased. In comparison, Al% and V% increased with increasing temperature, while the O% remained approximately constant.
The high-resolution Ti 2p spectra of the Ti6Al4V modied surfaces (Fig. 8) are used for the curve tting to determine the relative amounts of the four oxidation states of Ti present on the surfaces. The percentage contribution of each Ti oxidation state of the MPT Ti6Al4V modied surface is shown in Table 4. In all spectra, the predominant oxidation state was Ti(IV). Only the MPT Ti6Al4V modied surface displayed a notable second component for Ti(0). Ti(III) was also detectable in small amounts on all samples. Ti(II) and Ti(0) (with the exception of the MPT Ti6Al4V modied surface) were not present in any notable quantities.

Surface dynamic contact angle and force-distance curves
The results of DCA and AFM adhesion measurements for the MPT surface are shown in Fig. 9. This surface presented lower     roughness, S a $ 0.7 nm, compared to all other treated surfaces (TGL, TO, and TGL-TO). The q a for MPT is shown in Fig. 9a. The q a was in the range of 49-53 at pH 3, 7, and 9, whereas at pH 5 the q a increased to 62 . The force-distance data obtained by AFM performed against the MPT surface at pH 3, 5, 7 and 9 are shown in Fig. 9b-d. The F rep increased with increasing pH, the F JT was higher when pH was acidic compared to the other solutions pH, and the F PO decreased at pH 9 compared with pH 3, 5 and 7.
The q a of the TGL and TGL-TO 500 C surfaces are shown in Fig. 10, in which the TGL-TO 500 C surface exhibited a higher q a than the TGL surface. Furthermore, the TGL surface exhibited a lower q a at pH 5-9 (q a $ 15 ) than at pH 3 (q a $ 20 ), whereas the TGL-TO 500 C surface did not respond as strongly to pH change, exhibiting a minimum q a of 33 at pH 3, and a maximum q a of 38 at pH 7.

Surface topography
The thermally treated Ti6Al4V surfaces exhibited a range of topographies according to the treatment temperature. During TO at temperatures above 600 C, a crystalline TiO 2 layer was formed, as evidenced by XRD and Raman spectroscopy analyses. It has been reported that during the TO process, oxygen penetrates the substrate. 28 It has also been reported that aluminium diffusion, 29 as well as oxygen diffusion, 29 play a role in surface topography changes due to the formation of nodules, 29 small grains 30 and pores. 28 In our case, the roughness of modied Ti6Al4V surfaces increased with increasing TO temperature. Table 2 shows that the roughness of the treated surfaces increased with increasing TO temperature starting from S a $ 2 nm at 500 C to S a $ 100 nm at 800 C. Similar results were obtained by Kumar et al., who reported an increase in the roughness of Ti6Al4V surfaces with increasing TO temperature aer treatments between 500-800 C in air for 8-48 h. 9 A similar trend in surface roughness with increasing

Surface chemical composition
The crystal characterization of thermally oxidized samples by XRD and Raman spectroscopy are shown in Fig. 4. On the one hand, only rutile was possible to be identied by XRD on samples thermally treated at 800 C. It was observed that a-Ti was present for all the thermally treated samples except for samples treated at 80 C suggesting that the oxide layer formed due to TO at 800 C was thicker than that of the samples treated at 500-700 C. This observation about the increased oxide thickness was also reported by Kumar et al. 9 who observed that the TiO x thermally treated in air at 800 C for 8 h consisted mainly of rutile crystals and was thicker than the TiO x layer formed aer thermal treatment at 500 C and 650 C for 8 h.
On the other hand, Raman spectroscopy (Fig. 6) conrmed that samples treated below 800 C showed vibrational modes associated with both anatase and rutile crystal phases. In contrast, at 800 C, the vibrational modes were associated only with rutile. These results are in agreement with Chang and Huang 7 who studied the TiO 2 crystal transformation by Raman spectroscopy. They reported that amorphous TiO 2 and anatase were present in Ti6Al4V samples treated between 400 and 550 C. Rutile formation began at 550 C, whereas the metastable phase of anatase was transformed completely to rutile aer heat treatment at 700 C, suggesting that only rutile phase was formed at temperatures above 700 C.
In contrast, the crystal characterization for chemically treated Ti6Al4V that underwent further treatment by TO at low temperatures did not show the presence of TiO 2 crystal phases (Fig. 5). Kim et al. 32 reported that the TGL obtained aer treatment with NaOH and thermal treatment at 600 C was amorphous. In our case, the broad lines at 24, 28 and 48 in the X-ray diffractogram (Fig. 5), as well as the broad Raman peaks at the vibrational modes of Ti-O bonds in the Raman spectra (Fig. 7), indicated that the oxide layer formed was indeed amorphous.
Chemical analysis of MPT surfaces conrmed the presence of Al, Ti, O, C and small amounts of V and N in the passive oxide layer. This was in accordance with the XPS study of Ti6Al4V surfaces performed by Roessler et al. 2 and Milošev et al. 33 who reported the presence of Al in the passive oxide as well as C, associated with adsorbed hydrocarbons. In our case, small traces of other elements (Ca, P, S and Si) were also detected on the MPT Ti6Al4V surfaces and are believed to be associated with residual elements from manufacturing not specied by the supplier.
High-resolution spectra of the Ti 2p peak showed that Ti (IV) is the main Ti oxide present (above 90%; Fig. 8 and Table 3). Some Ti (III) was detected on all samples, whereas Ti(II) and Ti(0) were only present in trace amounts. This indicates that the temperatures used here had little impact on the composition of the oxidation states of the surfaces. An exception of this is the untreated polished MPT surface that contained 11% Ti(0) and 11.8% Ti(III) and therefore a slightly lower relative amount of Ti(IV) (74.6%). This result was in accordance with the study performed by Lee and Chang 34 on Ti6Al4V native surfaces who detected the same sub-oxides measured by XPS.
The Al and V content of the oxide layer was increased with increasing the treatment temperature, as is shown in Table 3. Du et al. 35 reported that Al diffused from the substrate to the outer surface as Al 2 O 3 . Al diffusion was more pronounced with increasing the treatment temperature. Kumar et al. 29 came to a similar conclusion, reporting that thermally treated Ti6Al4V surfaces were Al enriched compared with untreated Ti6Al4V. The V and Al enrichment were also reported by MacDonald et al. 35 in surfaces thermally treated at 600 C. In addition, C and N were observed to be present on the surface of thermally oxidised Ti6Al4V surfaces, as shown in Table 3. Still, their levels were considerably decreased when the temperature increased above 500 C.
The increased O/Ti ratio for the TO treated sample at 500 C (Table 3), with and without prior chemical treatment, compared with that of the untreated sample, was mainly due to O diffusion in the substrate and consequent formation of the oxide layer. The O/Ti ratio increase at treatment temperatures above 500 C was due to the formation of a more dense oxide layer, in agreement with Guleryuz and Cimenoglu. 28 They reported that dense oxide layers could be obtained at temperatures between 600-800 C.
In addition, the XPS analysis of TGL-TO Ti6Al4V surfaces, also shown in Table 3, reveals that the Al content increased with increasing the treatment temperature, whereas the change in the V was very small. Sun and Wang 36 reported that TGL formed on the Ti6Al4V surface might act as a diffusion barrier, preventing Al and V ions diffusing to the surface from the substrate. In our case, this might be true for V; however, it was obvious that the aluminium content increased consistently with increasing the treatment temperature for the TGL-TO Ti6Al4V.

Contact angle and adhesion surface properties
The surface contact angle evaluated at pH 3, 5, 7 and 9 (Fig. 9a) showed variations that can be attributed to the change of the interfacial tension. These are directly related with the intermolecular forces bounded by the solid and the contacting liquid. 37 It has been reported that the liquid surface tension of unbuffered water does not change signicantly with the addition of salt concentrations up to 1 M. 38 Consequently, changes in the contact angle titration behaviour could be attributed to the interaction of surface chemical moieties with the ionic content of the droplet through protonation and deprotonation. 38,39 In Fig. 9a, the q a of the MPT surface was highest at pH 5. This maximum could be associated with the isoelectric point (IEP) of the native titanium oxide (nTiO x ), as explained by Chau and Porter. 37 They reported that in the absence of specic absorbed ions, the maximum angle in the contact angle titration must be related to the IEP. This was also conrmed by McCafferty and Wightman 40 who determined the IEP of Al 2 O 3 and Cr 2 O 3 as the maximum value obtained by contact angle titration. According to the literature, the IEP of TiO x occurs between pH 4-6. 41 It has also been reported that the IEP of TiO x is inuenced by the density of hydroxyl groups present in crystal phases, contamination, and inclusion of ions. 2 For example, the IEP of anatase can vary between pH 5.9-6.6 and of rutile between pH 4.7-6.0. 2 Moreover, Roessler et al. 2 found that the surface IEP of native oxide on Ti6Al4V was around pH 4.4. Based on the information above, the use of contact angle titration is a suitable method for accurately determining the IEP of MPT surfaces only, considering that the high roughness of thermally and/or chemically treated samples interfered in the measurement and the data obtained was more suited to a qualitative assessment of surface charge.
The AFM adhesion study of MPT surfaces reinforced the argument that the IEP is around pH 5. The AFM adhesion analysis was performed with a Si cantilever that typically has a native SiO 2 layer with an IEP reported in some studies between pH 1 and 4 (ref. [42][43][44] and in some other studies at pH 3.9. 45,46 Raiteri et al. 43 reported the dependence of SiO 2 charge as a function of pH. The lowest F rep of MPT surface shown in Fig. 9b, and the highest attraction force shown in Fig. 9c, were observed at pH 3; this suggests that the cantilever was negatively charged and the MPT surface was positively charged. For solutions with pH > 5 it was expected that the negative charges of both SiO 2 and the native TiO x layer formed on the MPT surface increased in magnitude. This was veried by the calculation of F rep, which increased with increasing pH (from pH 5) reaching the highest value at pH 9. Butt et al.. 47 also reported the F rep and attractive forces of a mica surface using an alumina probe. F rep was only detected when the pH was above 8.1, and attractive forces occurred below pH 8.1. This was because the mica surface was negatively charged at the pH measured by Butt (between pH 3.1 and 10.4) and the alumina probe IEP was 8.1. 47 The F PO of the MPT surface, shown in Fig. 9a-d, exhibited their highest values at pH 5 and 7, yet a lower value at pH 9. These results can be explained considering that one of the surfaces was more negatively charged than the other, suggesting that the effect of the diffuse double-layer was involved in the interactions between the cantilever and the surface. 48 Hence, the reduced negatively charged MPT surface (nTiO x ) attracted water molecules than the negatively charged cantilever surface (SiO 2 ). This allowed the forces in the aqueous interface to interact, generating larger adhesion forces at pH 5 and 7 compared with those at pH 9, where both surfaces were negatively charged. The reduction of the magnitude of the surface potential when the pH was closer to the IEP was reported by Walsh et al. 48 They experimented on surfaces and a cantilever that were coated with TiO 2 using atomic layer deposition (IEP at pH 5.1) and observed that the surfaces at pH 6.5 were negatively charged. Still, the charge was not as large as at pH 10, where the diffuse double-layer was large enough to create repulsion between both surfaces.
The trend of the advancing contact angle with the increase of the thermal treatment is to decrease due to roughness and topography, as evidenced in Fig. 3. However, the contact angle of TGL-TO 500 C surfaces was higher than that measured for TGL surfaces at all pH (Fig. 10), even when the thermally treated surface presented a higher average roughness (S a ) than TGL surfaces (Table 2). This was an unexpected result. It has been reported that chemical treatment with peroxide can introduce Ti-complexes such as Ti-superoxide, TiO 2 À , Ti-peroxide and TiO 2 . 49-51 However, Osaka et al. 52 reported that further heat treatment in the air would result in the elimination of these Ticomplexes, leading to reduced hydrophilicity; these chemical reactions have been described by Osaka et al. 52 and Tengvall and Lundström. 6 The Ti-complexes slowly form polynuclear cations and anions leading to the precipitation of peroxotitanium hydrates (Ti(IV)O 3 (H 2 O) x with x ¼ 1, 2 or TiO 3(aq) ). The peroxotitanium hydrate species was suggested to form m-bridges (Ti-O-Ti) and consist of Ti 2 O 5 . 6,52 Osaka et al. 52 detected the presence of -O-O-(related with the peroxotitanium hydrate species) on pure Ti treated with H 2 O 2 samples by FT-IR at 890 cm À1 . Aer the thermal treatment at 400 C, the absence of the peak at 890 cm À1 in the FTIR spectra suggested that the additional thermal treatment eliminated the -O-Ogroups. It is likely therefore that such an elimination of -O-Ogroups might happen in the case of the TGL-TO 500 C surface, and for that reason, the contact angle was larger than for the TGL samples that were not thermally treated.

Conclusions
Ti6Al4V surfaces were modied by thermal oxidation and/or chemical treatment using a mixture of 90 v/v % of H 2 O 2 9.78 M/10 v/v % of HCl 1 M solution and were characterised by several advanced surface and structural characterisation techniques including AFM, DCA, XRD, Raman spectroscopy and XPS. The oxide layer formed on the surface exhibited different characteristics in terms of composition, topography and roughness, leading to enhanced electrostatic interactions with antimicrobial peptides. This paper presents a systematic study on the effect of thermal and chemical modications on the quality of the oxide layer formed on a Ti6Al4V surface.
All modied Ti6Al4V surfaces showed an increase in surface roughness with increasing the thermal treatment temperature. In addition, TGL-TO samples treated at low temperatures (500-600 C) exhibited an etched surface originated from the H 2 O 2 treatment performed prior to the thermal treatment, whereas the TGL-TO samples treated at high temperatures presented a topography surface due to the thermal oxidation. Amorphous TiO 2 was predominantly present in the passive TiO x layer for all samples, but aer heat treatments and with increasing the temperature, this layer became crystalline and predominantly rutile conrmed by both XRD and Raman spectroscopy.
Al and V diffusion from the substrate occurred in almost all thermally treated samples with Al content increasing with increasing the treatment temperature.
DCA analysis showed that the IEP of the MPT sample surface was around pH 5. In addition, AFM suggested that the surface was positively charged in aqueous solutions below pH 5, and negatively charged above pH 5. The F PO values measured for the MPT surface showed, that when a strong negatively charged cantilever was used, a reduction in magnitude, as well as increased adhesion forces, were observed when the pH was close to the IEP.
Finally, the thermal treatment of TGL samples at 500 C inuenced the surface wettability. This suggested that the elimination of Ti-H 2 O 2 during the thermal treatment at 500 C for the latter resulted in different surface composition and topography.

Conflicts of interest
There are no conicts to declare.