**3. Results**

#### *3.1. Surface Characterization*

Typical connective porous morphology after MAO were observed from Ag- and Cu-incorporated TiO2 layers. This morphology on each specimen was maintained after incubation in saline during 28 days (Figure 1).

**Figure 1.** Scanning electron microscopy (SEM) images of (**A**) the Ag- and (**B**) the Cu-incorporated TiO2 layers before and after incubation in saline during 28 days and the cross-sectional views of the specimens before incubation.

The XRD spectra obtained from the control, the specimens before incubation, and the specimens after incubation are presented in Figure 2, respectively, from the bottom to the top in each figure. Peaks corresponding to α-Ti and anatase TiO2 were detected, and those of Ag were undetected in the Ag-incorporated specimens before and after incubation. Peaks corresponding to α-Ti, anatase TiO2, and rutile TiO2 were detected, while those of Cu were undetected. Furthermore, the chemical structures of Ag- and Cu-incorporated specimens did not change by the incubation in saline during 28 days.

**Figure 2.** X-ray diffraction (XRD) spectra obtained from the (**A**) Ag- and (**B**) the Cu-incorporated specimens before and after incubation in saline during 28 days. The spectra presented at the bottom in each figure was obtained from the control.

Figure 3 shows the XPS survey scan spectra obtained from Ag- and Cu-incorporated specimens before and after incubation in saline during 28 days. The peaks originating from C, O, P, Ca, Ti, and Ag or Cu were detected from the XPS spectra of the Ag- and Cu-incorporated specimens. In addition to these elements, the peak originating from Na was detected in the specimens after incubation in saline from 7 to 28 days. P existed as a phosphate species and calcium existed as Ca2+, because the binding energies of the corresponding peaks of P 2p and Ca 2p3/2 were 133.7–134.1 eV and 347.6–347.9 eV, respectively. The binding energy of the Ti 2p3/2 peak was 458.9–459.3 eV, indicating that Ti existed as TiO2. The binding energies of the Na 1s peaks were 1071.7–1072.4 eV, indicating that Na exists as Na<sup>+</sup>.

**Figure 3.** X-ray photoelectron spectroscopy (XPS) survey scan spectra obtained from the (**A**) Ag- and (**B**) the Cu-incorporated specimens before and after incubation in saline during 28 days.

Figure 4 depicts the Wagner plot of Ag and Cu obtained from this XPS characterization and previous studies [46–48]. The chemical state change of Ag and Cu is determined by the comparison of their binding energy with α values on the Wagner plot. According to the Wagner plot of Ag, the α value of the Ag-incorporated specimen before incubation (0 days) was 724.9 eV, indicating that Ag mainly exists as Ag2O. The α values of Ag increased with incubation time. Since the α value finally converged to 726.1 eV, it became clear that the chemical state of Ag incorporated in TiO2 by MAO approached that of metallic Ag with increasing incubation times. On the other hand, the α values of Cu from the specimens immersed in saline for 28 days were 1849.3–1849.7 eV, indicating that Cu exists as Cu2O. The chemical state of Cu in the oxide layer did not change during the incubation in saline.

Figure 5 shows the changes in the concentrations of Ti, P, Ca and Ag or Cu detected from Figure 5A's Ag- and Figure 5B's Cu-incorporated specimens with the incubation time. The concentrations of Ag and Cu were relatively small (around 2.5 atom%), even before the incubation. Furthermore, from the results of the EDS analysis of the cross-section shown in Figure 1, 0.1 atom% of Ag and 0.1 atom% of Cu were detected from the inside oxide layers, respectively. Thus, the amounts of Ag and Cu incorporated during the MAO treatment were small. The amount of Ag and Cu dramatically decreased to approximately Ag 0.4 atom% and Cu 0.8 atom% upon incubation in saline from 0 to 7 days, respectively. These concentrations remained constant until 28 days. Moreover, the concentrations of Ca and P decreased, and that of Ti increased with the incubation time.

**Figure 4.** Wagner plot of (**A**) the Ag and (**B**) the Cu in the oxide layer incubated in saline for 28 days based on the photoelectron peaks and the Auger peaks. Each parameter of the Ag and the Cu compounds is plotted according to the previous studies [46–48].

**Figure 5.** Changes in the atomic concentrations in (**A**) the Ag- and (**B**) the Cu-incorporated TiO2 layers with the incubation time.

#### *3.2. Evaluation of Antibacterial Activity*

The normalized bacterial number of *E. coli* on each specimen is shown in Figure 6. The vertical axis represents the bacterial number normalized by the initial concentration of *E. coli*. The normalized bacterial number smaller than 1 (shown as a dashed line in the figure) indicates that the tested specimens exhibited an antibacterial effect. *E. coli* grew on the untreated Ti and control specimen, because the number of *E. coli* on those specimens significantly increased compared with the initial bacterial number. In contrast, Ag- and Cu-incorporated specimens developed antibacterial effects against *E. coli*.

**Figure 6.** Comparison of the antibacterial effects of the untreated Ti control specimen (micro-arc oxidation (MAO)-treated Ti without antibacterial elements), the Ag- and the Cu-incorporated specimens. Data are shown as the mean ± standard deviation. \* Significant difference between specimens (*p* < 0.05).

Changes in the antibacterial effects of Ag- and Cu-incorporated specimens before and after incubation in saline for 28 days are shown in Figure 7. The antibacterial effect of the Ag-incorporated specimen after incubation was significantly weakened compared to that before incubation. This effect was at the same level as that of Cu-incorporated specimens. On the other hand, the antibacterial effects of Cu-incorporated specimens did not change upon incubation, and were maintained even after the 28-day incubation in saline.

**Figure 7.** Changes of the antibacterial effects of (**A**) the Ag- and (**B**) the Cu-incorporated specimens before and after incubation in saline. Data are shown as the mean ± SD. \* Significant difference between the specimens before and after incubation (*p* < 0.05).

## **4. Discussion**

The porous oxide layers formed by MAO did not change their surface morphology and crystal structure during the incubation in saline (Figures 1 and 2). The lack of change is highly beneficial for the development of antibacterial property on implant surfaces.

The chemical states of P, Ca, and Ti in the Ag- and Cu-incorporated oxide layers were phosphate, Ca2+, and TiO2, respectively. These chemical states did not change upon incubation in saline for up to 28 days (Figure 3). The porous oxide layer consisted of TiO2 as well as incorporated Ca, P, and antibacterial elements (Figures 3 and 5). The presence of Ca and P in the porous oxide layer makes the hard-tissue compatibility of Ti better [27]. The peak originating from Na<sup>+</sup> was detected from the specimens after incubation in saline for 7 to 28 days. It is conceivable that compounds related to Na<sup>+</sup> were generated on the specimen surfaces, owing to the interfacial reactions between oxide layer and saline.

Ag- and Cu-incorporated porous oxide layers were formed on Ti surfaces, and slight amounts of Ag and Cu were incorporated by MAO using the electrolyte containing Ag and Cu (Figures 1 and 5). These results indicate that the constituent elements in the electrolyte were incorporated into the porous oxide layer during the MAO treatment. The incorporations of Ag and Cu by MAO is beneficial for realizing the antibacterial property on the implant surface (Figure 6). In addition, our previous studies revealed that specimens with Cu and a suitable amount of Ag did not a ffect the cellular adhesion, proliferation, di fferentiation or the calcification of the osteoblast cells [27,28]. Therefore, MAO can be imparting dual-function to the Ti surface, namely antibacterial property and hard-tissue compatibility.

The concentration of Ag in the oxide layer was dramatically decreased, and the chemical state of Ag in the oxide layer was changed from Ag2O to metallic Ag during the incubation in saline (Figures 4 and 5). These results indicate that Ag2O was converted into chemically stable metallic Ag in saline, due to the release of Ag ions. A study that investigated the formation mechanism of Ag particles in sodium citrate solution described the reduction of Ag on Ag particles via the radical-to-particle electron transfer [49]. The α' values of Ag in the oxide layer increased with incubation time, indicating that the density of electrons increased. Therefore, like the reduction of Ag on Ag particles in the solution containing sodium citrate, the change of the chemical state of Ag in the oxide layer incubated in saline may be caused by radical-to-Ag electron transfer. Moreover, the antibacterial e ffects of Ag-incorporated specimens changed upon incubation in saline (Figure 7). This finding implicates changes of both the concentration and the chemical state of Ag in the oxide layer in the antibacterial effect. However, it can be considered that Ag changes its chemical state more drastically depending on components, such as sulfur, in the actual biological environment, since Ag changed its chemical state, even in the simple simulated body fluid. Therefore, changes in the chemical state of Ag in the actual biological environment could influence the antibacterial e ffect.

The concentration of Cu in the oxide layer was dramatically decreased, and the chemical state of Cu in the oxide layer did not change upon incubation in saline (Figures 4 and 5). The chemical state of Cu in the oxide layer was stabilized as Cu2O despite the di fference in incubation time. This result indicates that the Cu2O is a stable chemical state in the TiO2 layer. In addition, the antibacterial effect of the Cu-incorporated specimen was maintained even after the 28 days of incubation in saline. These results indicate that Cu2O in a stable chemical state has a more important role in the development of an antibacterial e ffect, compared with the change of surface concentration of Cu. Previous studies revealed a substantial di fference in antibacterial e ffects between CuO and Cu2O. CuO inhibited the development of an antibacterial e ffect compared to metallic Cu. In contrast, thermally generated Cu2O was as e ffective as metallic Cu [50,51]. In other words, the presence of Cu2O and the slight amount of Cu in the TiO2 layer are necessary to develop the antibacterial e ffect.

The results support the proposal that the concentrations of Ag and Cu in the oxide layer are easily and dramatically changed by the incubation in saline. In addition, the chemical state of Ag changed from Ag2O to metallic Ag, while that of Cu did not change. The antibacterial e ffect of an Ag-incorporated specimen for Gram-negative bacteria changed by the 28-day incubation in saline, while the activity of Cu was maintained. The collective findings indicate the importance of the time-transient e ffects of Ag and Cu. This knowledge will be useful in the design of antibacterial implants based on the surface changes of Ag and Cu in vivo. Further study will be necessary to reveal the long-term e ffects of Ag and Cu for Gram-positive bacteria.
