Aging E ﬀ ect on Functionalized Silver-Based Nanocoating Braided Coronary Stents

: A previously developed ﬁber-based polyester (PES) stent, with mechanical properties comparable to commercial nitinol stents, was coated with metallic silver (Ag 0 ) and silver oxides (Ag x O) thin ﬁlms through direct current (DC) magnetron sputtering. Ag 0 and Ag x O coatings provide antimicrobial properties to the stents to minimize the occurrence of coronary stent infections. Nevertheless, the stent interacts with the atmosphere and then with the biological ﬂuids and may lead to the generation of silver species with diminished antimicrobial e ﬃ ciency and / or prone to induce cytotoxicity. Therefore, stent coating nanostructures aged 3 months were thoroughly analyzed by X-ray photoelectron spectroscopy (XPS) and their antimicrobial and cytotoxicity properties were assessed. Aging led to the presence of silver carbonate and bicarbonate as well as chemisorbed oxygen species in Ag 0 and Ag x O coatings. Bactericidal e ﬃ cacy was tested against an important nosocomial bacterium, particularly associated to indwelling devices: Staphylococcus epidermidis . Aged Ag 0 and Ag x O coating presented a Log reduction of 1 and 2 at their surface; respectively. However, aged stents were able to induce a Log reduction of 2 (Ag 0 ) and 4 (Ag x O) on the surrounding medium. Only aged Ag x O stent was able to provide a mild reduction of the bacterium at its surface and a clear antimicrobial e ﬀ ect (Log reduction > 3) within its vicinity. More importantly, both aged Ag 0 and Ag x O stents were shown to be compatible with ﬁbroblasts cells indicating that they can be safely used as indwelling devices, despite the aging e ﬀ ect.


Introduction
Cardiovascular diseases (CVD) represent nearly half of mortality in Europe, resulting each year in a loss of over €200 billion [1]. Atherosclerosis of the coronary arteries is the principal cause of CVD. It occurs when blood flow is obstructed due to cholesterol build-up within the arterial walls, causing the incidence of clot formation and severe circulatory problems [2]. Atherosclerosis treatment is mandatory and includes oral medication comprising vasodilators, anticoagulants or antiplatelet agents, as well as surgical procedures such as coronary artery bypass grafting, angioplasties and stent application [3,4]. Although angioplasties are effective in the immediate treatment of atherosclerosis, it often results in restenosis and reocclusion of arteries [5]. Instead, stenting techniques have been regarded the most

Stent Production and Functionalization
Braided fibrous polyester (PES) stents comprising monofilaments of PES 0.27 mm in diameter and a mandrel 6 mm in diameter with 35 • of braiding angle were produced according to the methodology described by Rebelo and co-workers [6]. Both stent and Ag impurities were removed in a petroleum ether ultrasonic bath for 1 h, followed by an argon (Ar) plasma etching during 5 min. The cleaning etching was performed using a pulsed DC power supply with the following specifications: 0.4 A, 200 kHz, 1536 ns of reverse time, a current density of 0.5 mA cm −2 within an 80 sccm Ar atmosphere. Ag 0 and Ag x O coatings were deposited by non-reactive and reactive pulsed DC magnetron sputtering, respectively. The deposition was performed as described by Rebelo and co-workers [21]. In brief, coatings were deposited from Ag target (200 × 100 mm 2 -purity of 99.99%), varying the oxygen flow. Stents were placed in a substrate holder 70 mm away from the target with a constant rotation velocity of 7 rpm. Samples of silicon and stainless steel were placed along the stents, in order to easily characterize the coating, namely by X-ray diffraction (XRD) analysis. The presence of oxygen was mandatory for the deposition of reactive mode to obtain Ag x O coatings, and its optimization is described elsewhere [20]. The stents were aged for a period of 3 months at room temperature in a desiccator, protected from light. To further analyze the nanostructure of the coated stent, an Ar + ionic erosion was performed according to the conditions described by Altstetter and Tortorelli, using a beam of 15 keV of Ar + [22].

Coatings Characterization
To determine the coating binding status, coated silicon samples were analyzed by X-ray photoelectron spectroscopy (XPS) (Kratos Analytical, Manchester, UK). The Kratos AXIS Ultra HSA equipment was used with a monochromatic Al Kα X-ray source (15 keV, 90 W) in fixed analyzer transmission mode (FAT) mode, encompassing a pass energy of 40 eV and a pass of 0.1 eV for regions of interest (ROI). A pressure lower than 1 × 10 −6 Pa and a charge neutralization system were applied. Experimental data were collected and analyzed using VISION software (Vision Manager software v. 2.2.10 revision 5) and CasaXPS software (Version 2.3.15), respectively. Carbon peak reference (285 eV) was used to correct the effect of electric charge.
The morphology of stents coated with Ag 0 and Ag x O, were evaluated by scanning electron microscopy (SEM) analysis with EDAX (FEI Company, Hillsboro, OR, USA). Nova nano-SEM200 equipment, prior to and after mechanical tests. XRD analysis, in a PANalytical X'Pert PRO MPD system using Cu Kα radiation (45 kV and 40 mA) with a parallel beam configuration (Malvern Panalytical Ltd., Malvern, UK), was undertaken with the purpose of understanding the structure and phase distribution of the coatings. The analysis was performed in grazing incidence mode with an angle of incidence of 5 • , and the grain size was determined by the Scherrer equation (Equation (1)).
where, T corresponds to the mean size of the ordered (crystalline) domains (nm), K to dimensionless shape factor, λ to the X-ray wavelength (nm), β to the half the maximum intensity (FWHM) (rad), and θ to the Bragg angle ( • ). The modified Auger parameter was estimated using Equation (2) [23].
where, BE(Ag 3d 5/2 ) corresponds to the maximum binding energy (BE) of Ag 3d 5/2 peak in the Ag 3d spectrum and KE(AgM 4 N 45 N 45 ) to the maximum kinetic energy (KE) of the AgM 4 N 45 N 45 peak in the AgMNN spectrum (strongest transitions between electrons whose orbitals are closest together).

Mechanical Properties
The mechanical properties before and after stent functionalization were evaluated. Cover factor indicates the amount of fiber deposited on the mandrel surface, depending on fabrication parameters, and it is a good indicator of stent structural uniformity. Porosity is also an important parameter and it should range between 70% and 80% in order to avoid stent collapse [24]. Cover factor and porosity can be calculated using Equations (3) and (4) [6].
where, W y is monofilament width (mm), N c corresponds to the number of bobbins, R to mandrel radius (mm), and α is braiding angle (rad). Stents were of 40 mm in length and 6 mm in radial diameter. Radial and longitudinal compression were tested in a universal testing machine. In all tests, for each kind of sample (PES stents, silver-coated stent and silver oxide-coated stent) 15 samples were used and the necessary force for the compression was measured [6,25]. Radial compression plays an important role in stent behavior, since insufficient radial force can lead to revision surgeries. For radial compression tests, stents were compressed until 25% of their initial diameter. Stents were compressed by 15% of their original diameter (6 mm) during longitudinal compression, and longitudinal force was estimated.
We bent 40 mm-length stents until their extremities were separated by 2 cm and diameter was measured at midpoint of the bending section, according to Kim and colleagues [26]. The percentage of unchanged diameter, in the bending section, was calculated through Equation (5) [6].
where D b is stent diameter in bent condition (mm) and D i the initial stent diameter (mm).

Antibacterial and Cytotoxicity Assessment
All aged stents were sterilized at 121 • C during 15 min in an autoclave, prior to biological assays. A clinical isolate S. epidermidis (IE 186) was used to assess the bactericidal properties of the aged stents, by performing an adaptation of Standard Japanese Industrial Standard (JIS) Z 2801 [27]. Bacterium was cultured in tryptic soy broth (TSB), during 24 h at 37 • C and 120 rpm. The initial bacterium concentration used was 1 × 10 6 colony forming units (CFU) mL −1 . Aged stents were immersed in 3 mL of TSB with bacterium suspension and incubated for 24 h at 37 • C and 120 rpm. Afterwards, the stents were transferred into 5 mL of a sodium chloride 0.85% (w/v) solution with 1% (v/v) of polysorbate 80. Samples were vortexed 5 times for 5 s each, to detach bacteria from the stent surface. Both detachment buffer and TSB bacterium suspensions were submitted to serial dilution and were plated in tryptic soy agar (TSA), and incubated at 37 • C for 24 h. Thereafter, the CFU were estimated. All assays were run in triplicate.
For SEM analysis, aged stents were incubated in the same conditions. Then, the bacterium at the stent's surface were fixated according to the protocol described by Padrão and co-workers [28]. Briefly, stents were carefully immersed in a solution of 2.5% (v/v) glutaraldehyde in phosphate-buffered saline (PBS) for 1 h. Afterwards, stents were mildly rinsed with distilled water and dehydrated using ethanol solutions with increasing concentration (55%, 70%, 80%, 90%, 95%, 100% (v/v)), for 30 min per solution. The remaining ethanol was left to evaporate, and the samples were gold sputtered.
Cytotoxicity evaluation assays were performed using an adaptation of International Organization for Standardization (ISO) 10993-5 and CellTiter 96 ® AQueous One Solution Cell Proliferation protocols [29]. Stainless steel was coated with Ag 0 and Ag x O, and aged in the same conditions as the coated stents. These samples were used in cytotoxicity tests due to their less complex geometry when compared to braided stents.
Fibroblast 3T3 (CCl-163) acquired from the American Type Cell Collection (ATCC) (Manassas, VA, USA) were cultured in Dulbecco Modified Eagle's Medium (DMEM) containing 1% (v/v) of penicillin streptomycin (PS) and 10% (v/v) of fetal bovine serum (FBS) until obtaining 80% of confluence. After cells detachment, 50 µL of cell suspension containing 1 × 10 5 cells mL −1 were added to each sample in a 96 well plate and incubated for 24 h, at 37 • C in a humidified atmosphere containing 5% (v/v) of carbon dioxide. 50 µL of the medium in contact with the samples were removed from each well and added to a 96-well plate with cells, incubated in the same conditions for 24 h. After the incubation, 20 µL of 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) was added, in the dark, to each well of the 96-well plate. The plate was incubated in the same conditions for 1 h and the absorbance of the resulting solution was measured at 490 nm.
Cellular viability was measure taking in account the ratio of the difference between cells grown in the control (samples without coating considered 100% viability) and coated samples over the control growth (Equation (6)). Assays were run in triplicate.

Statistical Analysis
One-way analysis of variance (ANOVA) with Tukey's multiple comparisons post-test was used (GraphPad Prism 6). All tests were performed with a confidence level of 95%.  (Figure 1e). The thin film of the coatings was generated by physical vapor deposition using direct current (DC) magnetron sputtering. It is a fast and user-friendly deposition method that does not require high temperature, essential to maintain the integrity of the PES stents. Furthermore, it allows a controllable film thickness and is environmentally friendly [30]. No red coloration of the uncoated PES fibers can be observed in both coated stents (Ag 0 and Ag x O). The aged Ag 0 stent is darker and more yellowish due to the metal silver oxidation and the presence or carbonates. The aged Ag x O stent exhibits a lighter brown color, probably, also due to the formation of silver carbonate that by contrast with the black coloration of silver oxide is known to have a white to yellow coloration. Fibroblast 3T3 (CCl-163) acquired from the American Type Cell Collection (ATCC) (Manassas, VA, USA) were cultured in Dulbecco Modified Eagle's Medium (DMEM) containing 1% (v/v) of penicillin streptomycin (PS) and 10% (v/v) of fetal bovine serum (FBS) until obtaining 80% of confluence. After cells detachment, 50 µL of cell suspension containing 1 × 10 5 cells mL −1 were added to each sample in a 96 well plate and incubated for 24 h, at 37 °C in a humidified atmosphere containing 5% (v/v) of carbon dioxide. 50 µL of the medium in contact with the samples were removed from each well and added to a 96-well plate with cells, incubated in the same conditions for 24 h. After the incubation, 20 µL of 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2Htetrazolium (MTS) was added, in the dark, to each well of the 96-well plate. The plate was incubated in the same conditions for 1 h and the absorbance of the resulting solution was measured at 490 nm.

Physicochemical Characterization
Cellular viability was measure taking in account the ratio of the difference between cells grown in the control (samples without coating considered 100% viability) and coated samples over the control growth (Equation (6)). Assays were run in triplicate.

Statistical Analysis
One-way analysis of variance (ANOVA) with Tukey's multiple comparisons post-test was used (GraphPad Prism 6). All tests were performed with a confidence level of 95%. The thin film of the coatings was generated by physical vapor deposition using direct current (DC) magnetron sputtering. It is a fast and user-friendly deposition method that does not require high temperature, essential to maintain the integrity of the PES stents. Furthermore, it allows a controllable film thickness and is environmentally friendly [30]. No red coloration of the uncoated PES fibers can be observed in both coated stents (Ag 0 and AgxO). The aged Ag 0 stent is darker and more yellowish due to the metal silver oxidation and the presence or carbonates. The aged AgxO stent exhibits a lighter brown color, probably, also due to the formation of silver carbonate that by contrast with the black coloration of silver oxide is known to have a white to yellow coloration.   XPS analysis of Ag and Ag x O coatings was carried out before (aged) and after a surface Ar + ionic erosion. The high-resolution spectra of Ag core-level, Auger peak, the C 1s peak, and the O 1s peak were examined with the intention of evaluating the chemical state and energy of the coating surface as previously performed [21]. Prior to erosion, both O 1s and C 1s envelopes contained shoulders that were deconvoluted indicating that multiple chemical states of carbon and oxygen were present on the coating surfaces. The core level photoemission spectra of C 1s obtained for Ag 0 coated surface was best fitted (full width at half maximum (FWHM) 1.2 eV) by means of four components ( Figure 2). The carbon peak at the lowest BE centered at 285 eV is assigned to adventitious carbon (C-C and/or C-H), while the second, located at 286.2 eV, is associated with C-OH bonding. The peak located at 288 eV is mainly ascribed either to silver carbonates or bicarbonates due to the interactions of Ag 0 with the outer atmosphere. Previous XPS studies of Ag 0 , AgO and Ag 2 O have shown that CO 2 present in the air could be responsible for the formation of Ag 2 CO 3 layers on these surfaces [31][32][33]. The higher BE peak at 289.2 corresponds to O-C=O bonding of the carbonate group of Ag 2 CO 3 [34]. High-resolution O 1s spectra depicted in Figure 2 corroborates the presence of Ag carbonate or bicarbonate (531.2 eV) [35]. The presence of stable chemically absorbed molecular oxygen (532.5 eV) and water (534.3 eV) further verifies the interactions of Ag 0 with the outer atmosphere [32,36]. A previous work confirmed through Raman analysis the existence of surface carbonate and surface oxygen species (superoxide oxygen and hydroxyl species) in Ag 0 displaying a similar peak in O 1s [37]. As expected, despite atmospheric interaction, no characteristic AgO and Ag 2 O (≈529 eV) are observable, since Ag 0 is not oxidized at ambient pressure by atmospheric oxygen [38]. After Ar + erosion, the Ag 0 surface was etched, removing all surface-contaminating species, revealing the coating original status obtained immediately after sputtering without any traces of C or O species.

Physicochemical Characterization
Coatings 2019, 9, x FOR PEER REVIEW 6 of 17 XPS analysis of Ag and AgxO coatings was carried out before (aged) and after a surface Ar + ionic erosion. The high-resolution spectra of Ag core-level, Auger peak, the C 1s peak, and the O 1s peak were examined with the intention of evaluating the chemical state and energy of the coating surface as previously performed [21]. Prior to erosion, both O 1s and C 1s envelopes contained shoulders that were deconvoluted indicating that multiple chemical states of carbon and oxygen were present on the coating surfaces. The core level photoemission spectra of C 1s obtained for Ag 0 coated surface was best fitted (full width at half maximum (FWHM) 1.2 eV) by means of four components ( Figure  2). The carbon peak at the lowest BE centered at 285 eV is assigned to adventitious carbon (C-C and/or C-H), while the second, located at 286.2 eV, is associated with C-OH bonding. The peak located at 288 eV is mainly ascribed either to silver carbonates or bicarbonates due to the interactions of Ag 0 with the outer atmosphere. Previous XPS studies of Ag 0 , AgO and Ag2O have shown that CO2 present in the air could be responsible for the formation of Ag2CO3 layers on these surfaces [31][32][33]. The higher BE peak at 289.2 corresponds to O-C=O bonding of the carbonate group of Ag2CO3 [34]. High-resolution O 1s spectra depicted in Figure 2 corroborates the presence of Ag carbonate or bicarbonate (531.2 eV) [35]. The presence of stable chemically absorbed molecular oxygen (532.5 eV) and water (534.3 eV) further verifies the interactions of Ag 0 with the outer atmosphere [36],35 [32]. A previous work confirmed through Raman analysis the existence of surface carbonate and surface oxygen species (superoxide oxygen and hydroxyl species) in Ag 0 displaying a similar peak in O 1s [37]. As expected, despite atmospheric interaction, no characteristic AgO and Ag2O (≈529 eV) are observable, since Ag 0 is not oxidized at ambient pressure by atmospheric oxygen [38]. After Ar + erosion, the Ag 0 surface was etched, removing all surface-contaminating species, revealing the coating original status obtained immediately after sputtering without any traces of C or O species.
(a) (b) Figure 2. High-resolution X-ray photoelectron spectroscopy (XPS) C 1s (a) and O 1s (b) spectra for Ag 0 coatings before and after Ar + ionic erosion. Figure 3 displays the XPS analysis of AgxO coatings. The highest BE peak of C 1s spectra is attributed to the functional C-C component (285 eV), followed by Ag carbonate or bicarbonate (288.5 eV), and its carboxyl group (286.5 eV) [32]. The AgxO O 1s spectra displays the presence of chemisorbed oxygen (533 eV) and a mixture of Ag2O and AgO (529.3 eV). The main peak is attributed  chemisorbed oxygen (533 eV) and a mixture of Ag 2 O and AgO (529.3 eV). The main peak is attributed to Ag carbonate/bicarbonate species (531.4 eV) [32]. After Ar + ionic erosion, Ag carbonate/bicarbonate is still present even if in a small quantity (531.3 eV) and it is denoted as a shift from 529.3 to 529.7 eV of the peak attributed to the Ag x O species. This minor shift can be ascribed to a difference in the ratio of the Ag 2 O/AgO species. Since the BE is lower in the aged coating, it may be attributed to higher concentration of AgO in the mixture [39,40]. to Ag carbonate/bicarbonate species (531.4 eV) [32]. After Ar + ionic erosion, Ag carbonate/bicarbonate is still present even if in a small quantity (531.3 eV) and it is denoted as a shift from 529.3 to 529.7 eV of the peak attributed to the AgxO species. This minor shift can be ascribed to a difference in the ratio of the Ag2O/AgO species. Since the BE is lower in the aged coating, it may be attributed to higher concentration of AgO in the mixture [39,40].
(a) (b) XPS analysis of silver and silver oxides is quite complex, due to the proximity of binding energies that are reported with a low dispersion of values or even superposition, and the Auger parameter (AP) was also considered. AP is used to identify the BE for oxidation, once AP does not depend on the sample charge shift, avoiding charge correction. Furthermore, the peaks of AP present larger chemical shifts than photoelectron peaks [40]. The main disadvantage of AP is that it cannot distinguish Ag(I) form others oxidation states [40]. Figure 4 shows the spectra of BE for Ag 3d. As expected, Ag 3d spectra display spin orbit components separated by 6 eV [40]. Ag 3d3/2 and Ag 3d5/2 binding energies prior and after Ar + ionic erosion are centered approximately at 374.9 ± 0.3 eV and 368.7 ± 0.3 eV, respectively. This is in good agreement with the values described for Ag 0 [33,37]. After Ar + ionic erosion, a minor shift in the energy binding and kinetic energy can be observed. Ag 3 d3/2 and Ag 3d5/2 peaks in the Ag 3d spectra display a negative shift of −0.4 eV, and a peak narrowing after Ar + ionic erosion. AgMNN (letters correspond to the principal energy level, or electron shells, associated with the Auger process) spectra also show narrowing of AgM4N45N45 and AgM5N45N45 peaks, and a slight negative shift of AgM4N45N45 after Ar + ionic erosion (−0.5 eV). These slight differences observed in all spectra may be associated with the elimination of the silver carbonate and bicarbonate species. The Ag 3d5/2 peak is located at 368.7 ± 0.3 eV. This fact, along with AP (726.2 ± 0.1 eV) ( Table 1), is in good agreement with the reported values for Ag 0 [41,42]. XPS analysis of silver and silver oxides is quite complex, due to the proximity of binding energies that are reported with a low dispersion of values or even superposition, and the Auger parameter (AP) was also considered. AP is used to identify the BE for oxidation, once AP does not depend on the sample charge shift, avoiding charge correction. Furthermore, the peaks of AP present larger chemical shifts than photoelectron peaks [40]. The main disadvantage of AP is that it cannot distinguish Ag(I) form others oxidation states [40]. Figure 4 shows the spectra of BE for Ag 3d. As expected, Ag 3d spectra display spin orbit components separated by 6 eV [40]. Ag 3d 3/2 and Ag 3d 5/2 binding energies prior and after Ar + ionic erosion are centered approximately at 374.9 ± 0.3 eV and 368.7 ± 0.3 eV, respectively. This is in good agreement with the values described for Ag 0 [33,37]. After Ar + ionic erosion, a minor shift in the energy binding and kinetic energy can be observed. Ag 3 d 3/2 and Ag 3d 5/2 peaks in the Ag 3d spectra display a negative shift of −0.4 eV, and a peak narrowing after Ar + ionic erosion. AgMNN (letters correspond to the principal energy level, or electron shells, associated with the Auger process) spectra also show narrowing of AgM 4 N 45 N 45 and AgM 5 N 45 N 45 peaks, and a slight negative shift of AgM 4 N 45 N 45 after Ar + ionic erosion (−0.5 eV). These slight differences observed in all spectra may be associated with the elimination of the silver carbonate and bicarbonate species. The Ag 3d 5/2 peak is located at 368.7 ± 0.3 eV. This fact, along with AP (726.2 ± 0.1 eV) (Table 1), is in good agreement with the reported values for Ag 0 [41,42]. High-resolution Ag 3d spectra of the AgxO coatings spectra do not denote any shift; however, both Ag 3d3/2 and Ag 3d5/2 became narrower after Ar + ionic erosion ( Figure 5a). As observed previously in Ag 0 Ag 3d spectra (Figure 4), peak narrowing may be due to the etching of silver carbonate/bicarbonate. However, the AgMNN spectra ( Figure 5b) exhibit a significant shift after ionic Ar + erosion (−1.1 eV). This shift is in agreement with that observed in Figure 3, regarding the BE shift of −0.4 eV. These shifts along with the AP value denote a distinct silver oxygen species ratio prior to and after Ar + ionic erosion coatings [43].   High-resolution Ag 3d spectra of the Ag x O coatings spectra do not denote any shift; however, both Ag 3d 3/2 and Ag 3d 5/2 became narrower after Ar + ionic erosion ( Figure 5a). As observed previously in Ag 0 Ag 3d spectra (Figure 4), peak narrowing may be due to the etching of silver carbonate/bicarbonate. However, the AgMNN spectra ( Figure 5b) exhibit a significant shift after ionic Ar + erosion (−1.1 eV). This shift is in agreement with that observed in Figure 3, regarding the BE shift of −0.4 eV. These shifts along with the AP value denote a distinct silver oxygen species ratio prior to and after Ar + ionic erosion coatings [43].
XRD analysis was performed in order to obtain information about coating structure. Figure 6 shows XRD diffraction patterns for aged Ag 0 and Ag x O coatings. Crystalline phases were identified as Ag (International Centre for Diffraction Data (ICDD) card n • 00-004-0783) and AgO (ICDD card n • 00-043-1038). Regarding aged Ag 0 coating, XRD analysis, reveals that it crystallizes in a face-centered cubic (fcc-Ag) crystal structure with random oriented crystallites and a grain size of 78.5 nm, calculated for (111) peak, while aged Ag x O coating was identified as a monoclinic phase with a grain size of 35.9 nm for (−111) peak. These results are in agreement with the expected structures for Ag 0 and Ag 2 O/AgO mixture [44].
High-resolution Ag 3d spectra of the AgxO coatings spectra do not denote any shift; however, both Ag 3d3/2 and Ag 3d5/2 became narrower after Ar + ionic erosion ( Figure 5a). As observed previously in Ag 0 Ag 3d spectra (Figure 4), peak narrowing may be due to the etching of silver carbonate/bicarbonate. However, the AgMNN spectra ( Figure 5b) exhibit a significant shift after ionic Ar + erosion (−1.1 eV). This shift is in agreement with that observed in Figure 3, regarding the BE shift of −0.4 eV. These shifts along with the AP value denote a distinct silver oxygen species ratio prior to and after Ar + ionic erosion coatings [43].   XRD analysis was performed in order to obtain information about coating structure. Figure 6 shows XRD diffraction patterns for aged Ag 0 and AgxO coatings. Crystalline phases were identified as Ag (International Centre for Diffraction Data (ICDD) card n° 00-004-0783) and AgO (ICDD card n° 00-043-1038). Regarding aged Ag 0 coating, XRD analysis, reveals that it crystallizes in a facecentered cubic (fcc-Ag) crystal structure with random oriented crystallites and a grain size of 78.5 nm, calculated for (111) peak, while aged AgxO coating was identified as a monoclinic phase with a grain size of 35.9 nm for (−111) peak. These results are in agreement with the expected structures for Ag 0 and Ag2O/AgO mixture [44].   Figure 7, shows a homogenous coating in both samples with a more columnar and porous structure in Ag x O coating, when compared with Ag 0 , due to the introduction of oxygen, which minimizes the porosity and leads to the formation of more compact films, with lower column size, as evaluated in previous work [21]. The aged coated stents were subjected to mechanical forces and SEM was used to analyze coating integrity. From the analysis of Figure 7 it is possible to observe differences in the morphology of silver-coated stents after the mechanical tests. A rearrangement of silver columns is noticeable, due to the ductile properties of the coating that may accumulate the tensions to which it is subjected and deform without being disrupted. However, no other considerable changes, like delamination, were visible in silver and silver oxide coatings, after the mechanical tests. SEM analysis on coated stents, presented in Figure 7, shows a homogenous coating in both samples with a more columnar and porous structure in AgxO coating, when compared with Ag 0 , due to the introduction of oxygen, which minimizes the porosity and leads to the formation of more compact films, with lower column size, as evaluated in previous work [21]. The aged coated stents were subjected to mechanical forces and SEM was used to analyze coating integrity. From the analysis of Figure 7 it is possible to observe differences in the morphology of silver-coated stents after the mechanical tests. A rearrangement of silver columns is noticeable, due to the ductile properties of the coating that may accumulate the tensions to which it is subjected and deform without being disrupted. However, no other considerable changes, like delamination, were visible in silver and silver oxide coatings, after the mechanical tests.  8 also shows the mechanical behavior of uncoated and coated PES stents, upon longitudinal and radial compression. As previously observed in the mechanical characterization of these stents by our group, when the compressions were applied stents underwent a structural deformation modifying their diameter and braiding angle [6]. Mechanical analysis of coated surfaces is not linear, due to the complexity of atom interactions between the substrate and the thin-film coating [45]. All samples presented a linear behavior under compression, which indicates that for the required extension, stents remained in the elastic domain with no permanent alteration in the braided structure. When comparing polyester uncoated and coated stents in longitudinal compression, it is clear that the coating is the main determinant influencing the mechanical behavior. Stiffness of the uncoated stent 0.02 N mm −1 is similar to the aged Ag 0 stent and to the aged AgxO stent (0.05 and 0.03 N mm −1 , respectively). Higher stiffness implies the application of a higher force to perform the same elongation [46]. Regarding radial compression, the materials' influence is not so clear, as the braided structure seems to play a key role in the process, due to the fact that the contact area between the stent and the applied force is much higher. Aged Ag 0 -coated stent presented a lower radial stiffness (0.36 N mm −1 ) whereas the aged AgxO-coated stent displayed a higher radial stiffness (0.90 N mm −1 ), when compared to the uncoated PES stent (0.55 N mm −1 ). Ag coatings are considered lubricants, thus, and considering the structure possess the main contribution for the radial compression behavior, the incorporation of Ag coatings can lead to a slippage between the  8 also shows the mechanical behavior of uncoated and coated PES stents, upon longitudinal and radial compression. As previously observed in the mechanical characterization of these stents by our group, when the compressions were applied stents underwent a structural deformation modifying their diameter and braiding angle [6]. Mechanical analysis of coated surfaces is not linear, due to the complexity of atom interactions between the substrate and the thin-film coating [45]. All samples presented a linear behavior under compression, which indicates that for the required extension, stents remained in the elastic domain with no permanent alteration in the braided structure. When comparing polyester uncoated and coated stents in longitudinal compression, it is clear that the coating is the main determinant influencing the mechanical behavior. Stiffness of the uncoated stent 0.02 N mm −1 is similar to the aged Ag 0 stent and to the aged Ag x O stent (0.05 and 0.03 N mm −1 , respectively). Higher stiffness implies the application of a higher force to perform the same elongation [46]. Regarding radial compression, the materials' influence is not so clear, as the braided structure seems to play a key role in the process, due to the fact that the contact area between the stent and the applied force is much higher. Aged Ag 0 -coated stent presented a lower radial stiffness (0.36 N mm −1 ) whereas the aged Ag x O-coated stent displayed a higher radial stiffness (0.90 N mm −1 ), when compared to the uncoated PES stent (0.55 N mm −1 ). Ag coatings are considered lubricants, thus, and considering the structure possess the main contribution for the radial compression behavior, the incorporation of Ag coatings can lead to a slippage between the monofilaments which may explain the lower force observed for Ag 0 -coated stents [47,48]. It is harder to draw a hypothesis for the 1.6-fold increase of the radial compression exhibited by the Ag x O, due to the lack of information in the literature. However, there is a report of a hardness increase when Ag oxides are present, due to changes in the roughness of the material [49].

SEM analysis on coated stents, presented in
Coatings 2019, 9, x FOR PEER REVIEW 11 of 17 monofilaments which may explain the lower force observed for Ag 0 -coated stents [47,48]. It is harder to draw a hypothesis for the 1.6-fold increase of the radial compression exhibited by the AgxO, due to the lack of information in the literature. However, there is a report of a hardness increase when Ag oxides are present, due to changes in the roughness of the material [49].

Biological Performance
S. epidermidis is one of the most ubiquitous human skin colonizer bacteria [50], described as an important nosocomial pathogen, responsible for numerous infections of immunocompromised patients carrying medical devices [51], particularly due to its biofilm forming capabilities [52]. Figure 8 comprises the antimicrobial efficacy of the aged coated stent's surface and within its vicinity. Figure 8a shows the bacterium cell concentration at the tested PES stent's surface. It denotes a Log 1 and a Log 2 reduction at the surface of aged Ag 0 and aged AgxO, respectively (Figure 9a). These results are corroborated by SEM images, which confirm S. epidermidis has an extensive biofilm formation on the stent's surface. The aged Ag 0 -coated stent exhibited an apparent lower bacterium concentration than the PES stent, and the aged AgxO stent shows not only a visual clear reduction in the bacterium concentration but also displays the presence of unviable bacteria (Figure 9b-d). Ag 0 displayed approximately a Log 2 reduction and AgxO exhibited an effective antimicrobial action (Log 4 reduction) (Figure 9e). This indicates a more potent vicinity action than at the stent surface. The higher susceptibility of several planktonic bacteria to the action of silver has already been reported in the literature. Planktonic Pseudomonas aeruginosa, Staphylococcus aureus and Enterobacter cloacae were promptly and extensively killed by silver bactericidal action whereas their biofilm counterparts

Biological Performance
S. epidermidis is one of the most ubiquitous human skin colonizer bacteria [50], described as an important nosocomial pathogen, responsible for numerous infections of immunocompromised patients carrying medical devices [51], particularly due to its biofilm forming capabilities [52]. Figure 8 comprises the antimicrobial efficacy of the aged coated stent's surface and within its vicinity. Figure 8a shows the bacterium cell concentration at the tested PES stent's surface. It denotes a Log 1 and a Log 2 reduction at the surface of aged Ag 0 and aged Ag x O, respectively (Figure 9a). These results are corroborated by SEM images, which confirm S. epidermidis has an extensive biofilm formation on the stent's surface. The aged Ag 0 -coated stent exhibited an apparent lower bacterium concentration than the PES stent, and the aged Ag x O stent shows not only a visual clear reduction in the bacterium concentration but also displays the presence of unviable bacteria (Figure 9b-d). Ag 0 displayed approximately a Log 2 reduction and Ag x O exhibited an effective antimicrobial action (Log 4 reduction) (Figure 9e). This indicates a more potent vicinity action than at the stent surface. The higher susceptibility of several planktonic bacteria to the action of silver has already been reported in the literature. Planktonic Pseudomonas aeruginosa, Staphylococcus aureus and Enterobacter cloacae were promptly and extensively killed by silver bactericidal action whereas their biofilm counterparts exhibited higher resistance, by exhibiting less cell death and by delaying the silver bactericidal activity (in approximately 70 h) [53], underscoring the higher virulence of biofilms.
Coatings 2019, 9, x FOR PEER REVIEW 12 of 17 exhibited higher resistance, by exhibiting less cell death and by delaying the silver bactericidal activity (in approximately 70 h) [53], underscoring the higher virulence of biofilms. Ag-based coatings using several deposition techniques has been extensively studied and good results in terms of antibacterial efficiency are commonly observed [54]. Despite its widespread use, the exact mechanism of actuation of the Ag as an antimicrobial agent is still not completely understood. Moreover, the bactericidal efficacy of each Ag species does not generate a clear consensus. In this work, the coatings Ag concentration is very similar (Ag 0 coating: 0.17 ± 0.01 mg L −1 and AgxO coating: 0.16 ± 0.01 mg L −1 ) [21]. Therefore the observed 2-fold difference may be associated with a suggested modus operandi of Ag antibacterial properties, the release of Ag + [55]. Ag 0 firstly needs to be oxidized to release Ag + , which is a slow process [56] whereas, Ag + elution from AgxO is a fast process (up to 0.07 ppm min −1 ) [57]. Ag + disrupts the activity of enzymes containing L-cysteine residue, and it was demonstrated that Ag + mediated the generation of reactive oxygen species (ROS), particularly of superoxide radical (O2 − ) [58]. ROS are known to induce cell death through several mechanisms, DNA being one of its critical targets [59]. Extensive mutagenesis irrevocably causes bacterial cell death [60], moreover ROS are soluble species, which may explain the observed higher Log reduction in the medium in comparison with the stent surface. Previous studies report that Ag 0 Ag-based coatings using several deposition techniques has been extensively studied and good results in terms of antibacterial efficiency are commonly observed [54]. Despite its widespread use, the exact mechanism of actuation of the Ag as an antimicrobial agent is still not completely understood. Moreover, the bactericidal efficacy of each Ag species does not generate a clear consensus. In this work, the coatings Ag concentration is very similar (Ag 0 coating: 0.17 ± 0.01 mg L −1 and Ag x O coating: 0.16 ± 0.01 mg L −1 ) [21]. Therefore the observed 2-fold difference may be associated with a suggested modus operandi of Ag antibacterial properties, the release of Ag + [55]. Ag 0 firstly needs to be oxidized to release Ag + , which is a slow process [56] whereas, Ag + elution from AgxO is a fast process (up to 0.07 ppm min −1 ) [57]. Ag + disrupts the activity of enzymes containing L-cysteine residue, and it was demonstrated that Ag + mediated the generation of reactive oxygen species (ROS), particularly of superoxide radical (O 2 − ) [58]. ROS are known to induce cell death through several mechanisms, DNA being one of its critical targets [59]. Extensive mutagenesis irrevocably causes bacterial cell death [60], moreover ROS are soluble species, which may explain the observed higher Log reduction in the medium in comparison with the stent surface. Previous studies report that Ag 0 thin films display less Ag ionization [20,21], thus aged Ag x O coating may generate a higher Ag + , which may explain the 2-fold higher antimicrobial activity. Interestingly, aging may have promoted the higher Log reduction in the medium, since both aged coatings exhibited the presence of Ag carbonate/bicarbonate on the stents' surface, as observed in XPS. Considering that, Buckley and colleagues observed a Log 3 reduction of S. aureus by AgCO 3 [61], a similar action may have occurred with the aged stents. Finally, the higher resistance of the biofilm cells present on the surface of the stents may be correlated to the mitigation of the ROS by the extracellular polymeric substances (EPS) [62]. Figure 10 exhibits a diagram of the hypothesized bactericidal activity of the coated stents.
Coatings 2019, 9, x FOR PEER REVIEW 13 of 17 thin films display less Ag ionization [20,21], thus aged AgxO coating may generate a higher Ag + , which may explain the 2-fold higher antimicrobial activity. Interestingly, aging may have promoted the higher Log reduction in the medium, since both aged coatings exhibited the presence of Ag carbonate/bicarbonate on the stents' surface, as observed in XPS. Considering that, Buckley and colleagues observed a Log 3 reduction of S. aureus by AgCO3 [61], a similar action may have occurred with the aged stents. Finally, the higher resistance of the biofilm cells present on the surface of the stents may be correlated to the mitigation of the ROS by the extracellular polymeric substances (EPS) [62]. Figure 10 exhibits a diagram of the hypothesized bactericidal activity of the coated stents. Cytotoxicity evaluation assays depict a cellular viability superior to 70% (Figure 11), determined as a threshold for cytotoxicity [28]. ROS usually display a lower toxicity for the animal cells in comparison to bacteria, since animal cells possess a plethora of ROS damage mitigation mechanisms [63]. Thus, according to cytotoxicity results, it is viable to insert coated stents in the human body and use antibacterial braided material. Cytotoxicity evaluation assays depict a cellular viability superior to 70% (Figure 11), determined as a threshold for cytotoxicity [28]. ROS usually display a lower toxicity for the animal cells in comparison to bacteria, since animal cells possess a plethora of ROS damage mitigation mechanisms [63]. Thus, according to cytotoxicity results, it is viable to insert coated stents in the human body and use antibacterial braided material. Coatings 2019, 9, x FOR PEER REVIEW 14 of 17 Figure 11. Fibroblast culture viability percentages.

Conclusions
Commercial application Ag 0 and AgxO thin-film PES stents are not expected to be implanted in an atherosclerosis patient immediately after their production. Therefore, an aging period is paramount to provide an insight on their performance in terms of surface properties, mechanical behavior, antimicrobial activity and biocompatibility. XPS showed that both aged Ag 0 and AgxO stent surfaces possessed Ag carbonate or bicarbonate, chemically absorbed molecular oxygen, and water, due their reactivity with the atmosphere. Aged Ag 0 stents exhibited lower antibacterial activity at their surface and in the vicinity than AgxO and no relevant cytotoxicity. More importantly, aged AgxO stents showed a mild bactericidal effect on the surface and an effective antibacterial activity in its vicinity, with tolerable cytotoxicity towards fibroblasts. These conclusions are important to highlight the bactericidal effectiveness of the surfaces of Ag 0 and AgxO thin films despite the presence of Ag carbonate or bicarbonate species. In fact, once the stents are implanted within the human body, the carbon dioxide present in the blood flow can promptly generate Ag carbonate or bicarbonate species at the surface of the coated stents. This work provides a first indication that the presence of these species will not hinder the bactericidal effectiveness nor increase the cytotoxicity, further corroborating the adequacy of these indwelling devices.

Conclusions
Commercial application Ag 0 and Ag x O thin-film PES stents are not expected to be implanted in an atherosclerosis patient immediately after their production. Therefore, an aging period is paramount to provide an insight on their performance in terms of surface properties, mechanical behavior, antimicrobial activity and biocompatibility. XPS showed that both aged Ag 0 and Ag x O stent surfaces possessed Ag carbonate or bicarbonate, chemically absorbed molecular oxygen, and water, due their reactivity with the atmosphere. Aged Ag 0 stents exhibited lower antibacterial activity at their surface and in the vicinity than Ag x O and no relevant cytotoxicity. More importantly, aged Ag x O stents showed a mild bactericidal effect on the surface and an effective antibacterial activity in its vicinity, with tolerable cytotoxicity towards fibroblasts. These conclusions are important to highlight the bactericidal effectiveness of the surfaces of Ag 0 and Ag x O thin films despite the presence of Ag carbonate or bicarbonate species. In fact, once the stents are implanted within the human body, the carbon dioxide present in the blood flow can promptly generate Ag carbonate or bicarbonate species at the surface of the coated stents. This work provides a first indication that the presence of these species will not hinder the bactericidal effectiveness nor increase the cytotoxicity, further corroborating the adequacy of these indwelling devices.