Surface Energy Changes Involved in Apatite Formation in Copper-Containing Bioactive Glasses

2022 The aim of this work was to use the Washburn capillary rise method to evaluate changes in surface energy promoted by the insertion of CuO in a 58S glass and its implication in bioactivity. The presence of CuO decreased the surface wettability by decreasing the Lewis acid-base component γ SAB and increasing the Lifshitz-van der Waals component γ SLW . The insertion of CuO also diminished the value of the electron donor parameter γ S- ; i.e. , the number of negative groups. This was accompanied by a nonuniform apatite formation on the surface, likely because the presence of CuO-rich regions with lower electron donor values impaired the adsorption of Ca 2+ , preventing uniform apatite precipitation. Therefore, surface energy was extremely sensitive to small physical and chemical changes in the glass structure and the behavior of the electron donor parameter could be used as an adequate probe to predict glass


Introduction
Immediately upon implantation into the body, a biomaterial surface will undergo important transformations promoted by direct contact with biological fluids. These transformations are associated with dissolution, precipitation, ion exchange, and the adsorption of ions, proteins, and several other organic molecules present in the extracellular medium 1,2 . Living cells will reach this surface when several of these transformations have already begun, notably the protein adsorption and dissolution/precipitation processes. Therefore, surface transformations will ultimately define the bioactivity of the material and its biological efficiency 3 .
Among all ceramic biomaterials, bioactive glasses are the most bioactive materials, and their development has allowed great advances in the production of implants and grafts aimed at bone tissue regeneration [4][5][6] . When in contact with biological fluids, the siloxane groups (Si-O-Si) on the glass surface undergo hydrolysis, forming negatively charged silanol groups (Si-OH) at alkaline pH. These negative groups attract Ca 2+ ions, leading to the precipitation of a positively charged Ca 2+ -rich layer. This positive layer will further attract PO 4 3ions, forming amorphous calcium phosphate that further crystallizes into apatite on the bioglass surface [6][7][8] . Part of the ions involved in this process derives from the biological fluid, and another part is due the gradual dissolution of Ca-and P-containing bioactive glass. The apatite layer formed on the glass surface is responsible for the high biocompatibility associated with bioactive glasses 3 .
Along with the chemical composition, reactivity, and other parameters, surface energy is one of the driving forces involved in these surface transformations.
Several studies in the literature determine the surface energy of biomaterials, especially dense calcium phosphates and biocompatible metal alloys [9][10][11][12] . However, despite several authors citing the importance of surface energy to the bioactivity of glasses, little information is available about how surface energy components vary during a bioactivity assay or the exact relationship between the structure and surface energy of the glass. The morphology of the apatite coatings obtained from typical bioactivity assays prevents the use of the conventional sessile drop method for surface energy determination. The liquid drops are quickly absorbed by the porous coating, impairing contact angle measurements 13,14 . Conversely, the Washburn capillary rise method is specifically indicated for the measurement of contact angles in porous materials [13][14][15][16] . Therefore, the aim of this research was to employ the Washburn capillary rise method to measure the changes in the surface energy of CuO-containing 58S bioactive glasses (58% SiO 2 -(33% -x) CaO -x% CuO -9% P 2 O 5 , wt. %) during a bioactivity assay. This technique may be used to provide a quick method for predicting apatite formation on bioactive glasses, without effectively performing SBF assays, which normally require more time (e.g., an additional week or more).
Copper oxide (CuO) was used to partially replace calcium oxide (CaO) in the original 58S glass composition and to intentionally promote significant differences in surface energy. Copper is an element known to effectively modify the hydrophobicity of ceramic systems [17][18][19] . In addition, Cu 2+ is also known for its angiogenic action (stimulating vascular system formation) 20,21 and bactericidal action 20,22,23 in association with biomaterials

Bioactive glass synthesis
The sol-gel process was used to obtain 58S bioactive glass -60% SiO 2 • 36% CaO • 4% P 2 O 5 (mol %) 24 . Initially, 22.7 mL of TEOS was diluted in 48.7 mL of 0.1 mol/L nitric acid (HNO 3 ) at room temperature. This solution was stirred for 1 h to ensure the complete hydrolysis and polycondensation of TEOS. Then, 4.5 mL of TEP and 14.31 g of calcium nitrate were added with a 45 min interval between each addition. The solution was stirred for 1 h and then stored in sealed plates at room temperature for 10 days to allow the gel to form.
The resulting gel was dried at 60°C and 120°C for 48 and 72 h, respectively, for the gradual removal of water and other byproducts formed during gelation. After drying, the xerogel was ground in a planetary mill for 1 h at 250 rpm and then sieved (80 mesh). The powder was thermally treated at 700°C for 1 h at a heating rate of 10°C/min and stored in a dry environment for further characterization. To ensure reproducibility, three independent syntheses (n = 3) were performed in this work.

Copper-containing bioactive glass synthesis
Copper-containing bioactive glass was obtained by following the same procedure described in section 2.2, with 1 mol % calcium nitrate replaced with copper nitrate. The samples were named 0% CuO and 1% CuO according to the copper content.

Elemental analysis
The composition of the obtained glass powders was determined by wavelength-dispersive X-ray fluorescence (WDXRF) with a Bruker S8-Tiger 4 kW spectrometer equipped with LiF 200, PET, XS-5 and XS-C crystals. The analyses were performed in a He atmosphere using the powders obtained after thermal treatment at 700°C for 1 h. The relative concentration of the elements was calculated considering their respective oxides.

Thermal analysis
The thermal behavior of the glasses was studied by a thermogravimetric analysis and differential thermal analysis (TGA/DTA) using STA 449 F3 Jupiter equipment. Analyses were performed with approximately 60 mg of the sample, obtained immediately after the gel was completely dried, in a N 2 atmosphere with a flow rate of 50 mL/min. The samples were heated from room temperature to 950°C in platinum crucibles at a heating rate of 10°C/min.

X-ray diffraction
The glasses obtained after thermal treatment at 700°C were analyzed by X-ray diffraction (XRD) to confirm the glassy state of the samples. The diffraction patterns were obtained with a Shimadzu diffractometer (XRD 6000) using a CuK α source (λ = 1.5405 Å) operated at 30 kV and 30 mA. The data were collected from 2θ angles of 15° to 45° in fixed time scanning mode (3 s/point) with a step size of 0.02°. The patterns were analyzed using HighScore Plus software. The standard structure used was hydroxyapatite Ca 10 (PO 4 ) 6 (OH) 2 (ICSD-151414).

Raman spectroscopy
The glasses obtained after thermal treatment were analyzed by Raman spectroscopy. Spectra were recorded by using a Horiba Raman microscope (LabRAM HR Evolution). A laser emitting a wavelength of 633 nm was employed, and the spectra were collected with an accumulation time of 30 s. The behavior of the bridging and nonbridging oxygens (NBOs) from silicate tetrahedra (SiO 4 4-) was estimated by the relative amount of Q n species obtained by deconvoluting the Raman spectra 8 . The curves were fit by using a multipeak Gaussian function-based approach via OriginPro software (2016). The peak positions assumed for each Q n specimen were based on the work of Sharma et al. 25 .

Specific surface area
The surface area was obtained by the Brunauer-Emmett-Teller (BET) method 26 , and the pore volume was obtained by the Barrett-Joyner-Halenda (BJH) method 27 . Nitrogen adsorption analyses were carried out using Quantachrome equipment (NOVA-1200e) at 77 K. Approximately 0.5 g of bioactive glass powder was degassed at 150°C under vacuum for 2 h. N 2 adsorption isotherms were obtained at P/Po range, varying from 0.05 to 0.99.

Bioactivity assay
The bioactivity assay 3,28 was carried out by immersing the glass powders in simulated body fluid (SBF) at a ratio of 10 mg/mL over a period of 1 and 7 days at 36.5 ± 0.5°C. After each time point, the suspensions were vacuum filtered, and the bioactive glass powders were dried at 120°C for 24 h.
The surface transformations that occurred during immersion in SBF were followed by X-ray diffraction (under the same conditions as described in section 2.6). Morphological changes were accompanied by scanning electron microscopy (SEM) performed with a JEOL microscope (JEM-5700). Images were acquired at 5 kV after the sample was carbon-coated with a DentonVacuum coating system (Desk V).

Surface energy -the Washburn capillary rise method
The contact angle of the glass surfaces, θ, was calculated by using the Washburn capillary rise method before and after immersion in SBF 13 . This method is based on measuring the penetration rate of liquids with different and known surface energies into a capillary containing a powder material. The Washburn equation 29 relates the rate of penetration with the contact angle θ, as follows ( where h is the height in which a liquid of a particular surface tension γ L and viscosity η causes the liquid in the capillary to rise in a certain time, t. This equation can also be written as a function of the gain of mass during the rise of the liquid in the capillary, assuming the following form 29 (Equation 2): where w is the mass of a liquid with surface tension γ L , density ρ, and viscosity η that causes it to rise in the capillary in a certain time, t. In our case, the contact angle, θ, was calculated by considering the gain of mass of three different liquids (water, formamide and n-hexane) of known Lifshitzvan der Waals component γ L LW , Lewis acid-base component γ L AB , total surface energy parameter γ L T ( ), electron acceptor parameter γ L + , and electron donor parameter γ L -( Table 1). The surface energy components of the bioactive glasses (solid) before and after the bioactivity assay in SBF were calculated by the van Oss-Chaudhury-Good method 30,31 according to the following Equations 3 and 4: This theory considers that two distinct components constitute the surface energy of a solid: the Lifshitz-van der Waals interactions γ S LW , and Lewis acid-base component γ S AB .
The first is associated with apolar or dispersive forces (van der Waals and other nonsite specific interactions between a solid surface and a liquid), and the second is associated with polar forces (hydrogen bonding, dipole-dipole, dipole-induced dipole and other site-specific interactions). In addition, this polar component is also associated with the acceptor parameter γ S + or donor parameter γ S of the chemical groups present on the solid surface.
The statistical significance of the obtained data was evaluated by analysis of variance (ANOVA) followed by Tukey's test. Differences at p ≥ 0.05 were considered statistically nonsignificant. The tests were performed in triplicate.

Results and Discussion
The gain of mass with respect to time for each capillary filled with the glass powder is plotted in Figure 1. As discussed in section 2.10, these curves allow us to calculate the contact angle for each liquid on each surface and the respective glass surface energy components.
The Lifshitz-van der Waals component γ S LW was higher than the Lewis acid-base component γ S AB in the glasses before the SBF assay, regardless of the presence of CuO (Figure 1). The Lifshitz-van der Waals component γ S LW represents active forces between molecules containing dipoles. Most chemical bonds in silica-based glasses have intense covalent character, which can explain these higher values. The partial replacement of CaO for CuO seemed to accentuate the covalent character since the Lifshitz-van der Waals component γ S LW increased. The Lewis acid-base component γ S AB reflects the attraction between highly polar molecules, many times involving hydrogen bonds. In the presence of water, the sum of acid and base contributions can describe wettability. The glass with CuO had a lower Lewis acid-base component γ S AB , suggesting a lower wettability. In fact, the water contact angles measured on the glasses with and without CuO were approximately 68° and 53.5°, respectively. This means that the insertion of CuO in the glass structure plays a role in decreasing the polarity by decreasing the number of positive and negative groups on the surface, clearly expressed by the lower values of the electron acceptor parameter γ S + and donor parameter γ S -. In fact, it has been shown that copper can increase the hydrophobicity of ceramic surfaces 17 . Polarity reduction has been pointed out as one of the ways to reduce bacterial adhesion 17,18 , which has great importance for developing antibacterial systems. However, there is a lack of information that explains how copper can effectively modify the surface tension of glasses. To answer this question, it is essential to understand how copper changes the glass structure.
The WDXRF results confirmed that approximately 1 mol % CuO was included in the glass structure as a substitution for CaO ( Table 2). Even this small amount of CuO was able to significantly change the thermal behavior of the bioactive glass ( Figure 2a) above 600°C. Below 600°C, the thermal events did not show significant differences. At approximately 120°C, an endothermic event can be attributed to the release of physically adsorbed water and ethanol (a byproduct of TEOS hydrolysis). At approximately 300°C, exothermic events are associated with the release of chemically adsorbed water, especially by the reaction between surface silanol groups (Si-OH) producing siloxanes (Si-O-Si) and by the release of nonreacted TEP 32 . At approximately 500°C, nitrate decomposition produces a strong endothermic peak 23,33 . Above 600°C, the mass loss becomes insignificant, and the thermal events can be directly associated with structural phenomena, such as the glass transition temperature (T g ). The insertion of CuO led to a T g reduction (from 673 to 648°C), indicating that CuO acted as a network modifier in the glass structure.
Copper can be incorporated into the vitreous network in two distinct oxidation states: Cu 2+ and Cu + . In both states, it is believed that its insertion leads to preferential breaking of P-O-P bonds instead of Si-O-Si bonds due to the greater affinity of copper for phosphate groups. The Δ i G of Cu 2+ -O-P (0.303) is larger than that of Cu 2+ -O-Si (0.189) and that of Cu + -O-P (0.449) relative to that of Cu + -O-Si (0.335) 34 . Therefore, the Cu-O bonds being more covalent than Ca-O bonds induces an increasing relaxation time for the glass chains while at the same time decreasing the energy required to move those chains 34 . This is directly reflected in a T g reduction.
The mobility of the glass network can be explained by structural modifications in the organization of SiO 4 4tetrahedral units (Q n ) forming silicate chains. In Q n structures, n refers to the number of bridging oxygen (BO) atoms per silicate tetrahedron, as depicted in Figure 2b. These arrangements can be observed by Raman spectroscopy. Bands between 800 and 1100 cm -1 were attributed to the asymmetric vibration of SiO 4 tetrahedra 36 . Within this region, the bands at 800, 950 and 1060 cm -1 are associated with Q 1 , Q 2 and Q 3 silicate units, respectively 37 . At 960 cm -1 , there is also the presence of the ẟ P-O mode of phosphate-linked species 38 , which in our case is masked by the predominant ν Si-O mode at 950 cm -1 from SiO 2 . CuO insertion led to an increase in the number of BOs in the glass structure since the relative area under the Q 1 and Q 2 curves diminished while Q 3 increased. The reduction of Q 1 and Q 2 silicate units associated with the formation of more covalent Cu-O-Si bonds can explain why the Lifshitz-van der Waals component γ S LW increased after the insertion of 1 mol % CuO into the glass structure. As the bonds are more covalent than ionic, the electrons in the oxygen atoms from siloxane bridges become more shielded, thus reducing the electron donor parameter γ S of the surface groups.
The XRD patterns of the samples exhibited an amorphous profile with a halo between 20° and 35° (Figure 2c). Low- Figure 1. Average capillary rise with respect to time for three different liquids in bioactive glass powders before and after 7 days of immersion in SBF solution (top). The surface energy parameters calculated from the bioactive glasses before and after immersion in SBF solution using the Washburn capillary rise method (bottom). * Significant difference (n = 3; p < 0.05) compared with the sample without CuO. intensity and broad peaks at approximately 26° and 32° corresponding to a low-order hydroxyapatite phase segregated from the glass matrix as nanocrystals 33,39,40 were observed in the CuO-containing sample. TEP tends to segregate along with the modifying cations in regions around amorphous silica clusters during the sol-gel process 39,41 . After heat treatment, these regions can be transformed into nanocrystalline phases, such as hydroxyapatite 39 . No significant changes were observed in the surface area and pore volume after the insertion of 1% CuO in the glass structure. The sample without CuO exhibited a surface area and pore volume of 92.46 ± 4.68 m 2 /g and 0.170 cm 3 /g, respectively, while the CuO-containing samples exhibited a surface area and pore volume of 87.30 ± 4.66 m 2 /g and 0.164 cm 3 /g, respectively. If the surface area and pore volume were not significantly altered by the presence of CuO, surface energy parameters are even more important to explain the observed changes in apatite precipitation.
The morphology of the glass particles before the bioactivity assay in SBF (0 days) was not changed by the presence of CuO, as shown in the SEM images ( Figure 3). However, after 1 day of immersion in SBF, the glass surfaces were coated by a layer of globular particles distributed across the surface. XRD analyses (Figure 4) showed the formation of important peaks at approximately 25.9° and 31.8° that are typical of the (0 0 2), (2 1 1), (1 1 2) and (3 0 0) hydroxyapatite planes 42,43 .
The presence of these peaks confirmed the formation of a hydroxyapatite phase typically observed in biomineralization assays using SBF 3,39 . The same peaks were found in the CuOcontaining sample at days 1 and 7. However, in this sample, the area under the amorphous halo observed at approximately 22° compared with the area under the crystalline peaks was larger than in the sample without CuO (Figure 4), especially after 7 days of immersion. This indicates that the coating was thinner or did not cover the entire glass surface completely. In fact, SEM images showed uncoated areas on the CuOcontaining sample (Figure 3, yellow arrows), explaining the higher amount of amorphous phase on these samples. This behavior substantiated the inhibitory effect of CuO on the glass surface reactivity and, consequently, on hydroxyapatite formation (bioactivity).
The inhibitory effect of CuO on apatite formation can be attributed to the changes in surface energy, directly caused by the structural rearrangements shown previously. One could assume that the decrease in the electron donor parameter γ S of the surface by the presence of CuO impaired the adsorption of Ca 2+ , further preventing the capture of PO 4 3-and formation of calcium phosphate clusters on the surface. This seems true because the CuO-containing glass surface becomes even more negative after immersion. Conversely, the sample without CuO became less negative after immersion, indicating that the negative charges were neutralized by adsorbing Ca 2+ from the SBF, with the consequent formation of a surface hydroxyapatite layer after the incorporation of PO 4 3groups. It has been demonstrated that an increase in the hydrophilicity (polar component) of treated titanium surfaces 44 promotes apatite precipitation 45 . Indeed, the polar component γ S AB measured on the samples without CuO was larger than that observed on the CuO-containing samples. Nevertheless, the inhibitory effect of copper ions on apatite formation must also be considered.
Several ions are known to inhibit apatite precipitation in aqueous media 43,46 . In general, these ions tend to avoid the nucleation of calcium clusters at the beginning of the precipitation process or further apatite crystallization. In any case, inhibition occurs along with significant morphological changes in the hydroxyapatite crystals 43 . The average crystallite sizes calculated for the [0 0 1] crystal direction, i.e., the (0 0 2) plane of hydroxyapatite, did not show significant differences between the samples with and without CuO ( Figure 4). This means that the amount of copper ions in the solution was not enough to promote substantial disorder in the lattice, thus decreasing the crystallite size, as expected for Cu-doped hydroxyapatite phases 47 .
The uncoated regions observed on the CuO-containing bioglass in the SEM images ( Figure 3) could reflect the heterogeneous distribution of copper and phosphate along the glass structure, as discussed before. The previous presence of CuO-rich regions across the glass structure might have generated regions with a very lower electron donor value. In these regions, hydroxyapatite precipitation during the SBF assay was avoided because of the small attraction of Ca 2+ ions to the surface, as depicted in Figure 5. In adjacent regions with higher electron donor values, calcium phosphate clusters could be formed.   Because of its modifying nature, copper is known to interfere not only with the dissolution of vitreous systems 22,23 but also with apatite precipitation 43,46 . Although there were no copper ions in the SBF, the inherent dissolution of the glass led to the release of these ions during the assay. The slight decrease in T g observed for the CuO-containing samples suggested that they experienced increased dissolution as well. Therefore, one could assume that a higher degree of dissolution interfered with the coating process and added to the copper inhibitory character promoted by these changes in surface energy.
These results suggest that the Washburn capillary rise method can be very useful to follow the surface energy changes in bioactive glass systems during bioactivity assays. The surface energy was extremely sensitive to small physical and chemical changes in the glass structure, and the behavior of the electron donor parameter could be directly associated with the degree of apatite formation on the surface. This specific parameter of surface energy seems to be an adequate probe to anticipate the bioactive glass behavior in SBF assays, i.e., bioactivity.

Conclusion
The insertion of CuO into the glass structure decreased the glass transition temperature and crystallization temperature. The Cu-O bonds being more covalent than the Ca-O bonds increased the amount of Q 2 and Q 3 silicate units and decreased the energy required to move silicate chains. The presence of CuO diminished the surface wettability by means of a considerable drop in surface polarity γ S AB and a rise in the Lifshitz-van der Waals component γ S LW . In general, the electron donor parameter of the surface energy γ S was higher than the electron acceptor parameter γ S + , confirming the predominance of negative groups on surfaces. However, the insertion of CuO into the glass structure remarkably diminished the number of negative groups on the surface. This behavior was accompanied by the presence of uncoated regions after the SBF assay. In this case, the presence of CuO-rich regions decreased the electron donor parameter of the surface, preventing the adsorption of Ca 2+ ions necessary to form the apatite layer in the SBF assay. Therefore, the electron donor parameter seems to be an adequate probe to predict the bioactivity behavior of bioactive glasses.