Effect of Various Surface Treatments on Wettability and Morphological Properties of Titanium Oxide Thin Films

The effect of three popular surface activation methods for a titanium oxide (titania) surface was thoroughly investigated to identify the most effective protocol for the enhancement of hydrophilicity. All the methods, namely H2O2 activation, UV irradiation and oxygen plasma treatment resulted in an enhanced hydrophilic titania surface, which was evidenced by the reduced contact angle values. To study in detail the chemical and morphological features responsible for the increased hydrophilicity, the treated surfaces were submitted to inspection with atomic force microscopy and X-ray photoelectron spectroscopy. The correlation between the treatment and titania surface hydroxylation as well as hydrophilic behavior have been discussed.


Introduction
Titania (titanium oxide, TiO 2 ) has been widely used in many areas of life (pigment, food coloring, cosmetics, toothpastes) and technology (medicine, energy and biosensing) [1,2]. After the discovery of the phenomenon of photocatalytic splitting of water on a titania electrode under ultraviolet irradiation, a huge growth of research activities has been seen on various TiO 2 materials, leading to many promising applications in areas of photovoltaics, photocatalysis and sensing [3,4]. Another interesting application of titanium dioxide is self-cleaning and self-sterilizing surfaces that can maintain their activity indefinitely as long as they are UV or even VL illuminated [5]. This is because when the titanium oxide film is illuminated with UV light, the contact angle is reduced to almost 0 • . In fact, TiO 2 is one of the few low-cost materials known to show this change in wettability under light illumination [6].
Applications of titania in biomedicine also gather significant attention owing to its excellent biocompatibility, unique photocatalytic properties, high chemical stability, and low toxicity [7]. Particularly, titania coatings are widely applied for improving the biocompatibility and bioactivity of implant materials or for the biosensing purposes, where a wide range of biomolecules can be immobilized to its surface and provide specific sensing functions [8]. In fact, the excellent biocompatibility of titanium itself is owed to a thin, several nm thick amorphous TiO 2 film naturally formed on its surface, protecting titania from various external environments, which could deteriorate the properties of the resulting surface [9].
In reality, the adsorption of ubiquitous water on TiO 2 surface always takes place [10], either as the molecular or dissociative form [11]. The latter plays an important role in further 2 of 11 application of the material as it leads to the formation of surface hydroxyl groups. As a consequence of dissociative adsorption of water resulting in the cleavage of some O-H bonds, on a TiO 2 surface, different kinds of hydroxyl groups are formed. The isolated and bridged hydroxyls are present on different sites depending on the specific titania crystal planes [12]. The fully hydroxylated TiO 2 surface presents two types of hydroxyl groups: (i) a terminal OH group, bound to a surface Ti 4+ site, which has five coordinates in respect to the lattice oxide ions, and (ii) a bridging OH group, bound to a Ti 4+ site, which has four coordinates with respect to the lattice oxide ions [13]. Moreover, the adsorbed water can be further bonded to surface hydroxyls through strong hydrogen-bonding interactions, significantly affecting the final properties of titania, e.g., wettability, chemical composition and surface morphology [14]. It ultimately drives the covalent and non-covalent interactions with polymers, proteins, and impacts the immobilization of organic molecules [15][16][17]. Therefore, it is extremally important to functionalize the surface with a significant amount of hydroxyls for further tuning of the interactions between titania and immobilized materials. In many cases, the content of surface hydroxyl groups has to be significantly increased to assure efficient attachment of the targeted molecules to the hydroxylated surface. Thus, pre-activation of the titania surface is of high importance for many applications, particularly in biomedicine where such surface can be used, i.e., as scaffold for tissue regeneration [18]. In addition to the chemical changes, the surface experiences a change in topography having an impact on a wide range of physical and chemical properties, morphology and wettability. All these together determine the hydrophilic properties as more effective for cellular interaction.
In this work, we have selected, applied to titania and compared three popular methods of surface activation, namely oxygen plasma, UV light, and H 2 O 2 activation. These methods were selected as we had previous experience (published and unpublished) of using plasma and H 2 O 2 on silica glass slides, and we observed increased wettability after those treatments (e.g., [19,20]). Plasma treatment is a physicochemical process where atoms of high energy can remove or significantly modify a surface. When it is applied for a short time and with low energy delivered to atoms, it allows for avoiding damaging the titania surface. Oxygen plasma improves the wettability of the surfaces, efficiently oxidizes and removes contaminants [21]. Implementing the oxygen/air plasma results in creation Ti 3+ and oxygen vacancies in TiO 2 films [22]. As a result, the oxygen plasma treatment promotes the hydrophilicity of the TiO 2 surface [23].
Additionally, UV light was chosen as it is known to increase the wettability of the TiO 2 surface [24,25]. Exposition of the TiO 2 surface to UV light results in the generation of highly reactive radicals that efficiently degrade nearby organic species [26]. UV activates only the light-available surface in contrast to plasma that can activate also UV light unavailable areas [27]. Of course, the efficacy of the plasma treatment is also pore size dependent, because the plasma cannot be generated in pores smaller than the mean free path [28]. The additional advantages of UV irradiation are photocatalytic sterilization and photo-induced superhydrophilicity (understood as complete wetting, i.e., water contact angles approaching 0 degrees) [29]. H 2 O 2 oxidation is a classical wet route of surface activation. Hydrogen peroxide is a powerful oxidant, as demonstrated by the E • value of 1.76 V measured for the half-reaction H 2 O 2 + 2 H + + 2 e − → 2 H 2 O [30]. Either TiO 2 or H 2 O 2 are not active when exposed to visible light, but the TiO 2 -H 2 O 2 complex shows visible light sensitivity because of the peroxy-complex formation on the TiO 2 surface [31]. H 2 O 2 is sometimes used with other co-activators aiming to provide a better activation efficiency [32]. In the current investigation, plasma, UV, and H 2 O 2 -exposed TiO 2 surfaces were compared to verify the impact of these methods on such surface properties as wettability, morphological and chemical features.

Materials
The following materials and reagents were used as received: titania-coated glass slides prepared according to the proprietary process of SDS Optic S.A., resulting in TiO 2 thickness

Activation Treatments
Based on the parameters used previously in the literature, the following three activation methods were applied to the TiO 2 -coated glasses: (i) Oxygen cold plasma activation-titania-coated glass slides were placed in a vacuum chamber of plasma cleaner (Pico, Diener Electronic, Ebhausen, Germany) and activated for one minute at 500 W power (using oxygen (pressure: 0.2 mbar, flow: 22 sccm)). (ii) UV activation-titania-coated glass slides were inserted into the UV chamber (KW-4AC UV Curer, Chemat) and subjected to irradiation with the use of four UV lamps with a total power of 16 W for 2 h. (iii) Titania-coated glass slide were immersed in 30 mL of 30% hydrogen peroxide. The Petri dish in which the activation was carried out was placed on a magnetic stirrer (RCT basic, IKA, Staufen im Breisgau, Germany) and gently stirred for 30 min.

Surface Characterization
X-ray photoelectron (XPS) spectra were collected using a multi-chamber UHV system (PREVAC, Rogów, Poland). Spectra were collected using a hemispherical R4000 electron analyzer, SAX-100 X-ray source (Al Kα, 1486.6 eV, 0.8 eV band) and XM 650 X-Ray Monochromator (0.2 eV band) with X-ray spot dimensions of 8 × 2 mm. The pass energy of the analyzer was set to 200 eV for survey spectra (with 500 meV step) and 50 eV for regions (with 100 meV step). The base pressure in the analysis chamber was 5 × 10 −9 mbar. The X-ray monochromator was placed at a magic angle relative to analyzer axis; thus, no angular correction was required. Spectra processing was completed using CasaXPS v. 2.3.23PR-1.0 software, and Scofield cross-sections included in the software library were used. Due to the insulating nature of the samples, we used a flood gun for charge neutralization. All the spectra were corrected by setting the aliphatic carbon peak binding energy at 284.7 eV. For spectra deconvolution, a Shirley background and Gaussian-Lorenzian peak shapes were used (30% of Gaussian, aka GL (30) in CasaXPS). In O 1s, the spectra peak widths were restricted to 1.35 eV [33].
Scanning Electron Microscopy (SEM) images of the bare titania surface were collected with a Quanta 3DFEG (Thermo Fisher Scientific/FEI, Waltham, MA, USA) microscope with the accelerating voltage of 20 keV.
Surface topology experiments were carried out with Atomic Force Microscopy (AFM) on a NanoScope V instrument (Veeco, Plainview, NY, USA) equipped with a piezoscanner of a maximum scan range of 150 µm × 150 µm. A ScanAsyst-AIR-HR cantilever/tip (Bruker, Billerica, MA, USA) with a spring constant of 0.4 N/m, tip radius of 2-12 nm and resonance frequency of 100-160 kHz was used. The topography and amplitude/deflection images were obtained simultaneously. Taping mode was used for imaging the topography of the external surface of the TiO 2 layers. NanoScope Analysis ver. 1.40 software was applied to AFM data processing and statistical analysis. Three images were collected per sample, each from a different part of the surface. Parameters S q , S a and S diff (discussed later in this paper) were averaged from all three images for each sample.
Water contact angles (CA) were measured in triplicate (for each of three replicates, the CA measurement was performed on both sides of the droplet) using the GBX contact angle meter controlled by Win Drop++ software. Then, 6 µL water droplets were gently settled on the surface using an automatic deposition system and immediately submitted to contact angle measurements.

Results
The SEM analysis of the bare titania surface (sample R0) under various magnifications was performed to examine the properties of the surface. As shown in Figure 1, the glass slide is covered with a continuous and tight titania film; however, at higher magnification, random particles or agglomerates of these particles with sizes up to 100 nm are visible. The analysis of the AFM images for the R0 sample also confirms the presence of particles/agglomerates present on the bare TiO 2 surface-an exemplary photo of this sample is shown in Figure 2a.
to contact angle measurements.

Results
The SEM analysis of the bare titania surface (sample R0) under magnifications was performed to examine the properties of the surface. As sh Figure 1, the glass slide is covered with a continuous and tight titania film; how higher magnification, random particles or agglomerates of these particles with size 100 nm are visible. The analysis of the AFM images for the R0 sample also confir presence of particles/agglomerates present on the bare TiO2 surface-an exemplary of this sample is shown in Figure 2a. In order to monitor the morphological changes resulting from the applied act protocols, the AFM analysis was performed on the slides directly after activation p i.e., for the oxygen plasma (RP), UV (RU), and H2O2 (RH) activated sample. Selec projected AFM photos are shown in Figure 2. The surface topography of th untreated slide is similar to projection captured on the SEM image. However, the topography changes significantly after plasma activation (Figure 2b), i.e., the su smoother, and only a small number of agglomerates/adherent particles can be ob suggesting removal of the residual particles upon the surface treatment with plasma. In turn, the surface topography of the UV-activated sample RU ( The discussed visual observations of morphological changes are confirmed by statistical analysis of the AFM roughness data (Table 1), which also reveals impor differences. An arithmetical mean height (Sa) and root mean square height (Sq) are commonly accepted measures used to evaluate the surface roughness [34]. Sa expr the averaged difference in the height of each point compared to the arithmetical m the surface. Sq represents the root mean square value of ordinate values and is equivalent to the standard deviation of the heights. The third parameter, Sdiff, is th percentage difference between the observed three-dimensional surface area and projected two-dimensional image area (in this case, 1 μm 2 ). All three parameters have the highest mean values for the non-activated R0 surface (4.32 nm, 2.97 nm and 5.95%, respectively). Each treatment of the activation process results in a decrease in the mean value of all three parameters, and the level of the changes depends on the type of activation. For the surfaces RH and RP, the change of the mean values of the parameters is more pronounced when compared to the sample RU. The reason for this can be the fact that plasma and H2O2 activation are quite aggressive types of treatments, involving free radicals or active ions directly approaching the surface and provoking its modification. In contrast, during UV activation, no active chemical moieties capable of chemical reactivity with the surface are present directly. The active chemical forms occur in situ during the reaction of UV light with titania. This explains an effect of UV activation of the topmost layer available to UV light. The mechanism of these UVinitiated chemical changes was described elsewhere [35].   In order to monitor the morphological changes resulting from the applied activation protocols, the AFM analysis was performed on the slides directly after activation process, i.e., for the oxygen plasma (RP), UV (RU), and H 2 O 2 (RH) activated sample. Selected 3Dprojected AFM photos are shown in Figure 2. The surface topography of the bare untreated slide is similar to projection captured on the SEM image. However, the surface topography changes significantly after plasma activation (Figure 2b), i.e., the surface is smoother, and only a small number of agglomerates/adherent particles can be observed, suggesting removal of the residual particles upon the surface treatment with oxygen plasma. In turn, the surface topography of the UV-activated sample RU (Figure 2c) looks quite like that of the initial non-activated sample R0 (Figure 2a). Interestingly, the RH surface activated with hydrogen peroxide (Figure 2d) shows topography with the intermediate pattern between the RP plasma-activated surface (Figure 2b) and the UV-activated surface (Figure 2c).

AFM Roughness Data
The discussed visual observations of morphological changes are confirmed by statistical analysis of the AFM roughness data (Table 1), which also reveals important differences. An arithmetical mean height (S a ) and root mean square height (S q ) are commonly accepted measures used to evaluate the surface roughness [34]. S a expresses the averaged difference in the height of each point compared to the arithmetical mean of the surface. S q represents the root mean square value of ordinate values and is equivalent to the standard deviation of the heights. The third parameter, S diff , is the percentage difference between the observed three-dimensional surface area and projected two-dimensional image area (in this case, 1 µm 2 ). All three parameters have the highest mean values for the non-activated R0 surface (4.32 nm, 2.97 nm and 5.95%, respectively). Each treatment of the activation process results in a decrease in the mean value of all three parameters, and the level of the changes depends on the type of activation. For the surfaces RH and RP, the change of the mean values of the parameters is more pronounced when compared to the sample RU. The reason for this can be the fact that plasma and H 2 O 2 activation are quite aggressive types of treatments, involving free radicals or active ions directly approaching the surface and provoking its modification. In contrast, during UV activation, no active chemical moieties capable of chemical reactivity with the surface are present directly. The active chemical forms occur in situ during the reaction of UV light with titania. This explains an effect of UV activation of the topmost layer available to UV light. The mechanism of these UV-initiated chemical changes was described elsewhere [35]. On a macroscopic scale, the changes resulting from the conducted activation processes were monitored by means of contact angle measurements. These measurements taken immediately after the activation were performed just after the processing (0 h) and after 24 h post-activation (24 h). The reason was to determine the wettability change after the activation process as well as the persistence of the increased hydrophilicity of the surface due to activation. The obtained results are summarized in Table 2. Looking at the mean values of the contact angle measured immediately after activation (0 h), it is clear that for each of the activation methods used, the contact angle decreased, although to a different extent (Figure 3). The least pronounced change was observed for the surface activated with H 2 O 2 , for which the contact angle decreased from 64.1 • to 58.3 • . However, for the other two activation methods, the change in the contact angle is more pronounced: for the RP surface, the observed contact angle is 10.1 • , and for the RU surface, it reaches 9.4 • . This means that the obtained RU and RP surfaces are completely wettable [36,37] (Figure 3b,c). Interestingly, the origin of superhydrophilicity of the surfaces RP and RU can be slightly different, which is indicated later in the manuscript when discussing XPS data.
of the surfaces RP and RU can be slightly different, which is indicate manuscript when discussing XPS data. It was noticed that after 24 h, the surface wettability decreased as increasing contact angles. Such behavior is expected, as an activated hi titania surface is very sensitive to ubiquitous volatile organic compounds few hours, the wettability of the surface might drastically change [38]. The induced re-relaxation of the titania substrate can also take place [39]. Theref variations in the activation protocol, handling, and storage procedures of surfaces might be responsible for the large deviations and final properties o TiO2 surface.  It was noticed that after 24 h, the surface wettability decreased as manifested by increasing contact angles. Such behavior is expected, as an activated highly energetic titania surface is very sensitive to ubiquitous volatile organic compounds, and even in a few hours, the wettability of the surface might drastically change [38]. Then, adsorptioninduced re-relaxation of the titania substrate can also take place [39]. Therefore, even small variations in the activation protocol, handling, and storage procedures of such activated surfaces might be responsible for the large deviations and final properties of the activated TiO 2 surface.
Interestingly, after 24 h, the wettability of the RH surface (67.6 • ) is close to the wettability of the non-activated surface (65.7 • ), which means that H 2 O 2 treatment provides the least stable surface activation. In the case of the other two treatments, the wettability after 24 h is much higher (39.9 • for RP and 37.9 • for RU) than in the case of the untreated surface (65.7 • ). When confronting wettability data with the AFM imaging data (with the greatest morphological changes on the RH surface), it is evident that the surface topography could not be alone responsible for the increase in surface wettability (hydrophilicity). Obviously, the chemical composition of the surface may also play a role, particularly the ability of each activation method to hydroxylate the surface.
Therefore, chemical compositions on the untreated and activated surfaces were examined using the XPS approach. The activation protocols were synchronized such that they were completed exactly at the same time and the freshly activated samples were immediately transferred to the XPS vacuum chamber for analysis upon reaching the appropriate low vacuum. The results of the XPS elemental analysis are presented in Table 3. Importantly, the studied samples coated with TiO 2 film contained a significant amount of carbon that seems an integral component of the studied samples. The deconvolution pattern of the C1s signal is somehow typical for organic species. The carbon content turned out to be a very convenient way of monitoring the effectiveness of activation. Its initial amount (33.3 at %) significantly decreased after the plasma activation (26.8 at %) and the process of UV irradiation (24.4 at %). On the other hand, the treatment with H 2 O 2 had no effect on the carbon amount (34.7 at %). The reason for the most effective carbon elimination by UV radiation lies in a well-known photoinduced self-cleaning capability of titania due to the surface reorganization mechanism as follows [24]: This mechanism is confirmed by changes in the content of other elements, i.e., oxygen and titanium. There was a negative correlation between both elements and the surface carbon amount-the greater the decrease in the amount of carbon, the greater the increase in the amount of oxygen and titanium. For example, after plasma treatment, the carbon content is reduced from 33.3 at % to 26.8 at % and can be considered as an indicator of the efficiency of the cleaning process. At the same time, the oxygen content increased from 44.3 at % to 48.1 at %, and the titanium content increased from 18.0 at % to 20.3 at %. It turns out that H 2 O 2 activation caused no loss of carbon from the surface as well as no significant increase in the amount of oxygen, unlike the other two activation methods (plasma and UV activation). These data are fully consistent with the surface wettability data obtained from the contact angle measurements and confirm the fact of the chemical changes in the surface that are mainly responsible for the increased hydrophilicity.
Very interesting conclusions can be drawn from thorough analysis of the deconvoluted O 1s signal (the most indicative and most frequently discussed signal in the literature in relation to titania materials [33,[40][41][42]). After Shirley background subtraction and de-convolution, the O 1s core energy level of the untreated R0 titania surface gives rise to four components, as shown in Figure 4a. The main component at 530.0 eV BE can be undoubtedly interpreted as originating from bulk O 2− oxygen atoms in titania lattice, which was always reported for TiO 2 samples. The two other components should correspond to the surface groups. Two different OH groups have been reported in the literature for hydrated/activated titania surfaces: acidic hydroxyl groups linked to the bridging oxygens and basic hydroxyl groups bonded to five-coordinated Ti 4+ cations. Indeed, the existence of the two next components shifted ≈1.6 eV and ≈2.9 eV compared to the main bulk peak is observed. These shifted components can be attributed to the two types of hydroxyl groups on the titania surface: (i) a two-fold bridging −OH group with acidic character (shifted 1.6 eV) and (ii) a single-fold ones "top" −OH group with basic character (shifted ≈2.9 eV) [43,44]. The weakest contribution at ≈533.6 eV can be tentatively attributed to the oxygen of molecularly physisorbed water in the upper hydration layer [33].

Conclusions
Three popular methods of surface activation of the titania surface were compared in a comprehensive and detailed manner for the first time. Each of them resulted in a titania surface with different properties, both in terms of surface morphology and, above all, The activation protocols employed in this paper lead to significant changes in the intensity of the individual components of the O 1s core energy level. As discussed earlier, protocols for plasma and UV activation led to a significant increase in the oxygen content, but importantly, the relative amounts of individual surface oxygen groups also changed. For a plasma-activated sample, the relative shares of bulk O 2− oxygen atoms increased (from 33.8 at % to 38.2 at %), and the number of bridging hydroxyl groups and top hydroxyl groups are also changed (from 6.1 at.% to 5.5 at.% and from 3.3 at % to 4.0 at %, respectively).
Activation by UV irradiation leads to a slight reduction in the number of hydroxyl groups while significantly increasing the amount of bulk O 2− oxygen atoms (from 33.8 at % to 41.6 at %) due to the most effective surface cleaning (which is also confirmed by the most pronounced reduction in the amount of carbon on the surface). In the case of the H 2 O 2 -activated sample, the relative share of bulk O 2− oxygen atoms increased slightly (from 33.8 at % to 35.6 at %), while the content of both types of hydroxyl groups decreased significantly (from 6.1 at % to 3.8 at % and from 3.3 at % to 2.8 at %, for bridging and top hydroxyl groups, respectively). Comparing the above results, it is evident that plasma is most effective in generating stable surface hydroxyl groups of acidic (bridging −OH) and basic nature (top −OH). Moreover, it is worth noting the difference between plasma and UV radiation with regard to the hydrophilicity of the surface. In the case of plasma, the hydrophilicity increases rather due to the large number of hydroxyl groups (especially basic top −OH groups), while in the case of UV irradiation, the main reason is the surface cleaning effect. This effect is clearly confirmed by the greatest loss of carbon and the increase in oxygen and titanium on the surface under the influence of UV radiation. Finally, it should be noted that due to the fact that for some applications, the presence of either acidic or basic carboxyl groups may be required, the selection of the appropriate activation protocol can be of a huge importance. It was reported that the type of surface hydroxyls could significantly modulate the surface property affecting the adsorption mode of studied adsorbents and promote the photocatalytic efficiency [13,45,46]. Therefore, the possibility of influencing amounts of acid and basic hydroxyl groups on the TiO 2 surface makes the appropriate activation method an important tool for controlling the final properties of this material. For example, plasma activation can decrease the surface acidity by increasing the relative share of basic hydroxyl groups on the surface.

Conclusions
Three popular methods of surface activation of the titania surface were compared in a comprehensive and detailed manner for the first time. Each of them resulted in a titania surface with different properties, both in terms of surface morphology and, above all, hydrophilicity and surface chemistry. Oxygen plasma and UV activation were the most effective, which is clearly indicated by the complete wetting of the resulting surfaces (contact angles: 10.1 • and 9.4 • , respectively) and significant changes in the surface chemistry (much lower carbon amount, significantly higher oxygen and noticeably higher titanium content). On the other hand, H 2 O 2 activation was shown to be the least effective procedure-it led to only a slight reduction of wettability (from 64.1 • to 41.3 • ) and minor changes in the surface chemistry. As shown in this work, hydrophilicity appears to depend mainly on the resulting surface chemistry, but the effect of surface morphology must also be considered. Consequently, our results show that the choice of the activation method of the titania coated surface might play a key role in the subsequent applications of the activated surfaces/materials.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.