Characterization and investigation of biological properties of silver nanoparticle-doped hydroxyapatite-based surfaces on zirconium

The infections leading to failed implants can be controlled mainly by metal and metal oxide-based nanoparticles. In this work, the randomly distributed AgNPs-doped onto hydroxyapatite-based surfaces were produced on zirconium by micro arc oxidation (MAO) and electrochemical deposition processes. The surfaces were characterized by XRD, SEM, EDX mapping and EDX area and contact angle goniometer. AgNPs-doped MAO surfaces, which is beneficial for bone tissue growth exhibited hydrophilic behaviors. The bioactivity of the AgNPs-doped MAO surfaces is improved compared to bare Zr substrate under SBF conditions. Importantly, the AgNPs-doped MAO surfaces exhibited antimicrobial activity for E. coli and S. aureus compared to control samples.

www.nature.com/scientificreports/ preventing infection is to prevent the initial adherence of bacteria to the implant surface 24 . Therefore, implant surfaces should be coated with antibacterial structures such as Ag, Cu and Zn 25 . Silver (Ag) is an important antibacterial agent that prevents microbial colonization. Furthermore, it exhibits biocompatibility and non-toxicity to human cells at low concentrations [26][27][28] . Ag may continue to show antibacterial properties after antibiotics are depleted and may increase the antifungal activity of antibiotics 29 . There are also studies using systemic antibiotic therapy and local delivery of Ag nanoparticles (AgNP) in vitro and in vivo. The results show that AgNPs increase the antibacterial efficacy compared to antibiotics and reduce the usage of antibiotics 30,31 .
The exact mechanisms of nanoparticle (NPs) toxicity to various bacteria are unclear. NPs can adhere to the bacterial membrane by electrostatic interaction and disrupt the integrity of the bacterial membrane. Nanotoxicity is usually triggered by the induction of oxidative stress with the formation of free radicals, i.e., reactive oxygen species, after nanoparticle application. Most importantly, compared to antibiotics, nanoparticles can effectively prevent microbial drug resistance in certain situations. Widespread use of antibiotics has created numerous public health hazards, such as super-drugs that do not respond to any available drug, and epidemics where the drug is not advocated. The search for new and effective bactericidal materials is important in combating drug resistance. Thus, NPs have been identified as a promising approach to overcome this problem. NPs are an effective therapeutic method in the fight against microbial resistance and multidrug-resistant mutants. In addition, it has gained importance as an anti-bacterial agent in recent years because it overcomes antibiotic resistance mechanisms and fights microbes using multiple mechanisms [32][33][34] .
In recent years, there have been some studies on ZrO 2 and/or HA coatings coated with MAO on zirconium and its alloys 8,10,12,17,35 . To improve the antibacterial ability of bioactive and biocompatible MAO coatings on zirconium, there are very limited researches on the preparation of MAO coatings with Ag ions/Ag layer/AgNPs 36-38 . Fidan et al. 37 fabricated Ag ions-doped ZrO 2 layer on zirconium by one-step MAO and investigated the antimicrobial activity against MRSA suspension. Durdu et al. 36 produced antibacterial Ag-nanolayers a thickness of 20 nm from bioceramic coatings formed by combined MAO and PVD techniques on zirconium and investigated the in vitro properties. Oleshko et al. synthesized AgNPs-decorated ZrO 2 coatings on ZrNb alloys by adding AgNPs to the electrolyte in a one-step MAO and investigated the antimicrobial activity for S. aureus 38 . However, antibacterial NPs structure must be formed on the outer layer because bacteria first contact the outer layer of the surfaces. Therefore, the effect of AgNPs on the MAO coatings should be studied in detail.
The MAO process performed in an alkaline electrolyte is an electrochemical process. Ceramic-like coatings with porous, homogeneous, hard, wear-resistant, corrosion-resistant, heat-resistant, electrically insulating and decorative ostentatious multifunctionality are produced by the MAO process 39 . The ED technique is based on the collection and deposition of ionic substances in solutions on metallic or non-metallic substrates through electrostatic interactions. This method offers a great advantage in producing desired shapes and covering large areas 40 .
There has been no study on the production of AgNPs-doped hydroxyapatite-based coatings on Zr by combined MAO and ED processes so far. Thus, this work aims to produce antibacterial AgNPs-doped hydroxyapatitebased bioceramic surfaces on zirconium by using MAO and ED techniques for implant applications. The phase structure, morphology, elemental amount, binding energy and wettability of all surfaces were analyzed by XRD, SEM, EDX-mapping, XPS and contact angle goniometer, respectively. In vitro bioactivity was investigated by immersion test in SBF for 28 days. Furthermore, bacterial tests were carried out for S. aureus and E. coli bacteria.

Experimental details
Sample preparation. Pure zirconium (Zr) plates were polished up to 2000 # sandpapers. Afterwards, all prepared plates were cleaned in an ultrasonic bath and were dried. MAO process. The MAO equipment with an AC power supply contains a stainless steel container, cooling and stirring systems as shown in Fig. 1a. The Zr substrate and stainless steel container served as the anode and the cathode, respectively. The prepared electrolyte contained (CH 3 COO) 2 Ca (calcium acetate, Alfa Aesar), β-C 3 H 5 (OH 2 )PO 4 Ca (β-calcium glycerophosphate, Alfa Aesar) and deionized water. The MAO was carried out at 0.379 A/cm 2 for 15, 30 and 45 min below 30 °C. After the MAO, all samples were cleaned in an ultrasonic bath and dried. ED process. Antibacterial AgNPs on the MAO surfaces were deposited at a constant potential value of -1 V for 0.5 min by using a potentiostat/galvanostat device (Metrohm Autolab). All MAO surfaces were coated in AgNO 3 (Merck) based aqueous electrolyte. The AgNPs structure was randomly accumulated on the MAO surfaces by the ED process. After the ED process, all samples were cleaned in an ultrasonic bath and dried. Schematic representation of experimental set up of the ED system is given Fig. 1b. Characterization of the surfaces. Phase structure on all surfaces was detected by XRD (Bruker D8 Advance) using Cu-Kα between 20° and 80° with a step size of 0.02°/min. The average thickness of the MAO coatings was measured at 40 different points by an eddy current device (Fischer Dualscope MP40). Average roughness of the MAO coatings was evaluated by surface profilometer (Dektak 8). The morphology, elemental distribution and elemental amount of the surfaces were investigated by SEM (Philips XL20S FEG) and EDXmapping and -area, respectively. The binding energies and surface chemistry of the surfaces were evaluated with Al-Kα radiation (1486.61 eV) by XPS (SPECS GmbH PHOIBOS 150). The average contact angles of the surfaces were investigated with a sessile constant drop technique by a Dataphysics OCA-15EC contact angle goniometer. All contact angles were measured within 1 min after water droplets with 1 µL were contacted the surfaces.

Microbial adhesion test. Microbial adhesion experiments were carried out with Staphylococcus aureus
and Escherichia coli. First, all samples were sterilized in an autoclave. Then, AgNPs-coated MAO and bare Zr samples were treated with test microorganisms adjusted according to 0.5 McFarland scale. For this process, the samples of sizes with 10 mm × 10 mm × 1 mm were immersed in 5 mL of MHB medium. After incubation at 37 °C at 125 rpm for 24 h on an orbital shaker, the samples were removed from the medium and washed with 15 mL of water to remove non-adherent organisms. This process was repeated for 3 times. Then, each sample was taken into a clean tube and 2 mL of 150 mM NaCl was added and vortexed for 2 min to collect the bacteria attached to its surface. Serial dilutions of the obtained bacterial solution were made and 100 µL were taken from the dilutions and applied to MHA medium by spreading method. At the end of 48 h of incubation at 37 °C, colony count was made and % inhibition was calculated. All experiments were performed in triplicate.

Statistical analysis.
Statistical analysis was carried out by "IBM SPSS Statistics 22 SP" software. All data were reported as standard deviation (mean ± SD). Statistical significance between the means was decided by oneway ANOVA and Duncan's test, p < 0.05 was considered statistically important.  www.nature.com/scientificreports/

Results and discussion
The phase structures of all MAO and AgNPs-doped MAO coatings are illustrated in Fig. 2 Subsequently, ZrO 2 is syntheses structure through the MAO process 42 . The c-ZrO 2 is stable phase observed as major on the MAO surfaces since instantaneous temperature reached up to 10 6 K in discharge channels during the MAO 9 . With increasing MAO treatment time, the intensity of hydroxyapatite increases while the intensity of the metallic Zr relatively decreases. The formation mechanism of hydroxyapatite was discussed in detail in previous works 8,17,35,[43][44][45] . The partially stabilized Ca 0.15 Zr 0.85 O 1.85 was formed on the MAO surfaces by the corporation Ca 2+ and c-ZrO 2 . HAP structure easily occurs on zirconium oxide due to the catalytic effect of free radical Zr-OH groups on surface 16 . Cationic Ca 2+ and anionic PO 4 3− ionized from calcium acetate and β-calcium glycerophosphate-based electrolyte react with H 2 O molecules. Subsequently, they form bioactive and biocompatible hydroxyapatite during MAO process [46][47][48] . However, the elemental or compound phase structure of Ag could not be detected by XRD due to the existence of a trace amount of AgNPs on the MAO surfaces. Thus, XPS and EDX analyses were carried out to prove the existence of elemental and phase structure of AgNPs on the MAO surfaces.
The XPS analysis was performed to determine the binding energies and surface chemistry of the elements in the structure of AgNPs-doped MAO coating produced at 15 min. XPS survey and XPS spectra of Ag3p, Ag3d, Ca2p, O1s and Zr3d on AgNPs-doped MAO surface are shown in Fig. 3. According to the XPS survey spectrum, the AgNPs-doped MAO coatings contain Ca, P, O, Zr and Ag. These elements are also detected by EDX-mapping and EDX-area. The C1s spectrum is observed owing to the surface contamination during handling and cleaning. The Zr3d spectrum consists of two peaks at the binding energy of 185.9 eV for Zr3d 5/2 and at the binding www.nature.com/scientificreports/ energy of 188.3 eV for Zr3d 3/2 . Double Zr3d peaks in XPS spectra correspond to the presence of ZrO 2 49,50 . The O1s spectrum reveals a single peak at the binding energy of 534.9 eV. The Ca2p spectrum contains double peaks at the binding energies of 350.7 and 353.9 eV. Double Ca2p peaks in XPS spectra refer to the existence of hydroxyapatite 35,51 . The Ag3d spectrum contains double peaks at the binding energies of 371.1 eV and 377.2 eV. Moreover, The Ag3p spectrum reveals a single peak at the binding energy of 575.8 eV. This indicates the existence of AgNPs on the MAO surface as supported in the literature, SEM and EDX results 52,53 .
To distinguish surface structure, the surface morphology of the MAO coatings taken at low magnifications is given in Fig. 4. Typically, porous and rough MAO morphology occur on Zr substrates, which is due to the presence of micro sparks by MAO process. Following the breakdown of the dielectric film, micro sparks begin to form where the oxide film coating is weak. There are many large and small pores on the coating surface. This causes an increase in the micro-discharge channels formed during the process. With increasing time, larger sized discharge channels are formed. The reason for the expansion of the discharge channels is to reduce the pressure inside the discharge channels with the effect of increasing time during the MAO process. Large discharge channels eliminate preformed small discharge channels on the surface. As a result, with increasing time, there is a high anodic potential transition under galvanostatic conditions throughout the MAO process. Thus, while the discharge channels decrease in number, their size increases. As a result, a rough and porous surface is produced.  www.nature.com/scientificreports/ The micro pores on the surfaces are called as discharge channels (micro discharge channels). Usually, the size of discharge channels improves while the number of it decreases with increasing MAO treatment time. The surface morphology of AgNPs-doped MAO surfaces is seen in Fig. 5. The AgNPs structures are randomly and uniformly accumulated on the MAO surfaces. Furthermore, the AgNPs structures do not significantly change the morphology of MAO. The porous morphology on MAO is maintained by the ED process. The ED process is applied to all MAO surfaces by using identical experimental conditions such as voltage, electrolyte and time. However, the size and amount of AgNPs on the MAO surfaces increase from 15 to 45 min. The number of nucleation and growth sites for Ag structure increases on the MAO surfaces since the porosity and roughness of the surfaces enhance with increasing MAO treatment time. Also, anionic OHions on free radical Zr-OH groups may contribute to migrate cationic Ag + ions onto the surface. Compared to low treatment time, more Ag + ions can diffuse on negatively charged OHon the MAO surface since the intensity of c-ZrO 2 increases with increasing MAO treatment time, as seen in Fig. 2. Eventually, more AgNPs are accumulated onto the MAO surfaces with increasing MAO treatment time when the ED treatment time keeps constant.
The elemental distribution of AgNPs-doped MAO surfaces is shown in Fig. 6. As expected, the Ca, P, O, Zr and Ag elements are observed on the surfaces. Moreover, all elements are homogenously distributed throughout the whole surface. The dark sites on the surfaces refer to discharge channels. The Ca, P, O and Zr elements are directly related with the phases m-ZrO 2 , c-ZrO 2 , Ca 0.15 Zr 0.85 O 1.85 and Ca 5 (PO 4 ) 3 (OH) phases coming from the MAO surface. Furthermore, the elemental Ag is uniformly distributed through the surfaces whereas the elemental or compound phase structure of Ag is not detected by XRD. The elemental amount of the surfaces is given in Table 1.
The wettability of bare Zr, bare MAO surfaces and AgNPs-doped MAO surfaces was investigated by contact angle goniometer as seen in Table 2 and Fig. 7. The contact angle of 65° is used to express the difference between www.nature.com/scientificreports/ hydrophilic and hydrophobic surface 54 . The bare Zr surface indicates a hydrophilic character since the average contact angle is lower than 65°. However, the average contact angles of the MAO surface are lower than one of bare Zr. The hydrophilic ability of the MAO surface is improved with increasing MAO treatment time. The contact angles of droplets on solid surfaces depend on some parameters such as the surface chemistry, topography and roughness 55,56 . Compared to bare Zr substrates, the water molecules on the MAO surfaces are easily adsorbed since the surface of MAO coatings is porous. Furthermore, one of the most important critical factors is the existence of OHgroup on the surface of coating. It is well known that a high amount of OHis associated  www.nature.com/scientificreports/ with improving hydrophilicity 57 . The intensity of c-ZrO 2 improves with increasing treatment time, as seen in Fig. 2. This refers to improve polarity of surface due to the existence of Zr-OH groups on the MAO surfaces. Polar surfaces improve wettability and exhibit low contact angles 58 . However, AgNPs-doped MAO surfaces indicate very low contact angles respect to the MAO surfaces. It is well known that a hydrophilic surface tends to enhance cell adhesion, cell differentiation, bone mineralization and biological activity 59 .
In vitro immersion test in SBF is typically carried out to estimate the predicting bioactivity of a biomaterial surface. To analyze the bioactivity of bare Zr and AgNPs-doped MAO surfaces, in vitro immersion test was carried out at 36.5 °C under SBF for 10 days. The surface morphology and phase composition of bare Zr and AgNPs-doped MAO surfaces after immersion test are shown in Figs. 8, 9, respectively. Red arrows on Fig. 8 indicate newly formed secondary particles on bare and AgNPs-doped MAO surfaces at post-immersion in SBF. As shown in Fig. 9, bioceramic composite phases such as Zr (zirconium, JCPDS # 005-0665), m-ZrO 2 (monoclinic ZrO 2 , JCPDS # 037-1484), c-ZrO 2 (cubic ZrO 2 , JCPDS # 049-1642), Ca 0.15 Zr 0.85 O 1.85 (calcium zirconium oxide, JCPDS # 026-0341) and Ca 5 (PO 4 ) 3 (OH) (hydroxyapatite, JCPDS # 009-0432) are observed on the MAO and AgNPs-doped MAO coatings at post-immersion in SBF. The spherically shaped secondary particles, which are randomly distributed, are observed on the surfaces. These secondary particles refer to the existence of a hydroxyapatite (secondary apatite) structure on the surface. The low amount of secondary apatite structure is locally accumulated on bare Zr surface due to bioinert nature of it. However, more homogenously secondary apatite deposition on AgNPs-doped MAO surfaces reveals with respect to bare Zr substrate owing to the presence of hydroxyapatite, ZrO 2 and Ag contents. The hydroxyapatite, ZrO 2 and Ag contents on AgNPs-doped MAO surfaces trigger the diffusion of cationic Ca 2+ and anionic PO 4 3− ions in SBF with the electrostatic interactions 60 . The ZrO 2 provide effective epitaxial nucleation sites for secondary apatite structures under SBF 61 . Therefore, it is believed that the enhanced secondary apatite-forming ability of the ZrO 2 is related with the abundant Zr-OH groups on the MAO surfaces. Especially, Zr-OH groups trigger the absorption of anionic PO 4 3− by the cationic Ca 2+ in SBF, which contributes to the electrostatic potential interaction between secondary apatite nuclei and outer surface. Moreover, it is assumed that some of Ca 2+ sites in the apatite lattice were substituted with Ag + ions during SBF. Thus, this accelerates the dissolution rate of secondary apatite in SBF solution 62 . The faster dissolution of soluble ions to the SBF escalates the apatite formation and precipitation on AgNPs-doped MAO surfaces. Eventually, this results on the formation of more secondary apatite on AgNPs-doped MAO surfaces with respect to bare Zr surface.
For antimicrobial analysis of bare Zr and AgNPs-doped MAO surfaces, S. aureus and E. coli bacteria adhering to the surface treated with microorganisms were collected and re-cultured. After incubation, the reduction in bacterial viability is calculated as percent inhibition, as given in Fig. 10 and Table 3. S. aureus that is the most    Fig. 5. The amount of Ag on the MAO produced at 30 min can be slightly higher than 45 min since the AgNPs structures are not detected by EDX in dark regions (discharge channels). However, according to the SEM images, the distribution of AgNPs on the MAO surface produced at 45 min is more than ones at 15 min and 30 min. Thus, antimicrobial efficiency of the AgNPs-doped MAO produced at 45 min is the best obtained within all samples. Ag indicating broad spectrum antimicrobial effect at very low concentrations possesses many advantages such as biocompatibility and good antibacterial ability 66 . Ag passes through the microbial cell wall and can bind to DNA. It can interfere with the replication process. In the literature, the focus in this area is the development of Ag-substituted HA coatings to minimize the adhesion of bacteria 67 . Thus, Ag structure that exhibits a broad spectrum of antimicrobial activity is often preferred for antibacterial purposes in the medical field such as burn creams, vascular grafts and implants 27,37,68 . Oleshko et al. 38 obtained excellent antimicrobial effect the

Conclusion
In this work, the randomly distributed AgNPs-doped hydroxyapatite-based bioceramic composite MAO surfaces were fabricated on Zr by combined MAO and ED processes. Bioactive and biocompatible hydroxyapatite-based bioceramic structures were detected on the surface. Furthermore, the existence of AgNPs on MAO was verified by XPS, SEM and EDX. The AgNPs-doped MAO surfaces, which are beneficial to bone structures at postimplantation, were porous and rough. Furthermore, the hydrophilicity of the AgNPs-doped MAO surfaces is significantly improved compared to bare Zr and MAO surfaces. The bioactivity of AgNPs-doped MAO surfaces is improved compared to bare Zr substrate under SBF conditions. Importantly, the AgNPs-doped MAO surfaces indicated antibacterial activity for S. aureus and E. coli. Eventually, it can be concluded that the AgNPs-deposited MAO surfaces have potential for long-term usage of implant applications.

Data availability
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.