Introduction

Antibacterial surfaces are of high need as current estimates suggest that from 4 to 10% of patients receiving dental implants develop postoperative infections1. Bacterial colonisation and biofilm formation harm the substratum, where bacterial attachment can lead to implant failure along with inflammations of peri-implant tissues2. To reduce this risk, several attempts have been made to develop Ag-containing antibacterial coatings on Ti-based substrates, considering the antibacterial nature of Ag3,4,5,6. It has been stated that the fraction and size (i.e. surface-area-to-volume ratio) of the Ag particles play an important role in the antibacterial efficiency of coatings. In general, larger surface-area-to-volume ratios allow the substantial release of Ag+ ions to interact with the lipopolysaccharide molecules forming the outer membrane of the bacteria, disrupting its integrity and penetrating deep into the cell, damaging its DNA7,8. In this way, Ag affects the functioning of bacteria cells, resulting in the elimination or significant reduction of their activity. However, excessive release of Ag+ ions may lead to a cytotoxic effect on the tissues. It was confirmed that blood concentrations of Ag should not exceed 300ppb to prevent eventual side effects9. It has been reported that allowable Ag concentration in the plasma sprayed coating should be in the range of 2wt.%, which is regarded as the cytotoxic limit for coatings applied to metallic biomedical alloys, according to the reports in the literature10.

As the most promising surface modification technique for titanium implants, the micro-arc oxidation (MAO) process has the potential for fabricating thick, adherent and micro-porous titanium oxide coatings containing Ag-NPs, whose antibacterial characteristics could be adjusted by altering the concentration of the antibacterial agents in the electrolyte. Recently, a superior combination of bioactivity and antibacterial efficiency against Staphylococcus aureus has been obtained from the MAO coatings containing 1.14 and 0.7 wt% Ag for commercially pure titanium (cp-Ti) and Ti-6Al-4V alloy, respectively11,12.

In the literature, Ag-NPs-incorporated antibacterial MAO coatings were fabricated by adding Ag-NPs or Ag-containing chemicals/salts (silver acetate and silver nitrate) into the electrolyte. While Ag-containing salt-added electrolytes favour the formation of very fine Ag-NPs (~ 1–3 nm) within the MAO coatings, Ag-NP-added electrolytes tend to produce MAO coatings containing relatively larger Ag-NPs (approximately tens of nm13). When the size of the Ag-NPs is of concern, the addition of Ag-containing salt-added electrolytes tends to favour the formation of MAO coatings exhibiting more efficient antibacterial properties. Aydogan et al.11 reported that even such small addition of the AgC2H3O2 to the base electrolyte as 0.001 mol/l forms MAO coating having ~ 99.98% antibacterial efficiency against S. aureus after 24hr of incubation. Furthermore, Ag concentration in simulated body fluid was kept at levels below the cytotoxic limits for the human body, that is, 1.5 ppm after 14 days of the immersion12. Although studies showed that antibacterial MAO coatings produced with the addition of Ag-containing salts are more effective against bacteria than standalone NPs, the reason behind this (participation of Ag in the MAO coating in the form of NPs) remains unknown. Moreover, too little attention has been regarded to detailed microstructure studies aimed at explaining the way of Ag is incorporated from the AgC2H3O2 into the MAO coating.

Thus, within this work a detailed microstructural analysis, with particular attention on high resolution electron microscopy, was carried out in order to find the explanation of the very high antibacterial effectivity of MAO coatings having in-situ formed Ag-NPs produced with AgC2H3O2 presented in the report ‘Optimisation of micro-arc oxidation electrolyte for fabrication of antibacterial coating on titanium’11. In this study, electron microscopy observations were supported by surface-sensitive X-ray photoemission spectroscopy (XPS) analysis, often applied in efficient study of Ag-NPs, also for the TiO2-Ag system14,15,16,17,18,19. The results of the present study allowed to propose mechanisms of Ag-NPs formation from used salts and provide new insights into the fabrication of antibacterial coatings by a simple and cost-effective approach.

Materials and methods

In this work, titanium of commercial purity (grade 4) was used as a substrate material. The initial material, purchased from Wolften company in the form of a rod having a diameter of 50mm, was subjected to a plastic deformation covering multi-pass hydrostatic extrusion reducing its diameter down to 5mm. Details of this processing were described elsewhere20. The rod after hydrostatic extrusion was cut into cylinders of 4mm in height, ground with sandpapers and ultrasonically cleaned in ethanol and distilled water to remove the surface contamination before the MAO processing, aimed at fabrication of the coating.

Micro-arc oxidation of titanium substrates was performed in a stainless steel vessel filled with the electrolyte containing 0.1 M of Na2HPO4 and 0.002 M of AgC2H3O2 inorganic salts, dissolved in distilled water. The samples were immersed in this electrolyte throughout the process and alternating pulses of the positive and negative electric potentials of 300 and 60V, respectively, were applied to them using a bipolar pulsed power supply for an overall time of 5min. Each electric pulse was maintained for 0.6ms, separated by 0.4ms of breaks between them. The temperature of the process was kept at ~ 22°C.

Phase analysis of the produced coatings in macroscale was performed using the grazing incidence X-ray diffraction (GI-XRD) technique at an angle of 3° with a Bruker D8 Discover diffractometer equipped with a Co anode. Phase composition was achieved by indexing of the recorded spectrum (with a 2θ = 0.01° angular step and an acquisition time of 0.25 s per step) with the help of the DIFFRACplus software and the PDF4 crystallographic database. The crystallite size was evaluated by the analysis of the broadening of the peaks on the XRD spectrum using the Williamson–Hall (W–H) method performed in the HighScore Plus computer programme.

The SEM micro-scale observations of the coatings as well as the preparation of the lamellae for TEM studies were done using a ThermoFisher Scios 2 Dual Beam microscope, equipped with an EasyLift nanomanipulator for lift-out of the samples (FIB method). The TEM nanoscale microstructure observations were carried out with a ThermoFisher Themis G2 200 kV FEG microscope equipped with a ThermoFisher Ceta™ bottom camera, a Fischione HAADF/STEM detector and a ChemiSTEM™ energy dispersive X-ray spectrometer (EDS).

The chemical composition of the coatings was determined by the XPS method. X-rays from a water-cooled Al anode (providing 20 mA emission current using 15 keV excitation) irradiated the sample from an 18mm work distance and spectra were detected using a special retarding field cylindrical mirror analyser type DESA 150 (Staib Instruments Ltd, Germany). Spectra were recorded with a constant 1.5 eV energy resolution at 0.1 eV energy steps. The measurements were performed under an ultra-high vacuum of 1 × 10–9 mbar. The sample surface was cleaned to remove contamination before the XPS measurement through 30min of ion sputtering with an Ar+ ion beam using 1 keV energy, set to a glancing angle of 15°. To determine the peak intensities, the Gaussian/Lorentzian peak model function was fitted to the measured peaks involving the Shirley-type background subtraction. Additionally, concentration calculation took place assuming the homogeneous distribution model and using sensitivity factors from21.

Results and discussion

The SEM/BSE images presenting the MAO coatings with the addition of Ag are shown in Figure 1. They show that the whole surface is covered with the coating material with an average thickness of ~ 5 µm (Figure 1b). The MAO coating exhibits a high porosity with numerous micro-pores of diameters ranging from several tens of nm to several µm (Figure 1c). The images obtained at higher magnifications indicate the presence of a large number of fine spherical particles of high-atomic-number-elements that cover almost homogenously the whole coating surface (Figure 1d). It was observed that some of these particles protrude from the surface of the coating. The FIB lamella was cut out from the place shown in Figure 1c in a way to present the cross-section of the coating while avoiding larger porosity sections. SEM observations showed that the surface of the coating is free from larger agglomerates of particles, confirming that the MAO processing conditions were optimised efficiently. Additionally, indexing of the GI-XRD diffraction pattern revealed the presence of the anatase phase of the TiO2 with an average crystallite size of ~ 40 nm (Figure 1e). Furthermore, a weak signal from the substrate (α-Ti) was recorded. The lack of signal from the Ag in the GI-XRD pattern is explained by an insufficient signal-to-noise ratio, similar to the observation of Aydogan et al.11.

Figure 1

figure 1
Macroscale photo of the cp-Ti sample with MAO coating (a), cross-section SEM/BSE image (b), plan-view SEM/BSE image of MAO coatings with the addition of Ag (c,d) and GI–XRD pattern recorded from the top surface of the MAO coating (e).

The TEM/BF images obtained from the coating cross-section showed the microstructure features typical for the MAO-coated samples (Figure 2). Presence of a thin continuous amorphous layer (up to 50nm in thickness) can be noticed at the coating/substrate interface, similar to the observations in20 securing a good bonding between them (Figure 2a). Above this, a porous zone consisting of a mixture of crystalline and amorphous oxides was formed. An almost fully amorphous zone (with only small crystalline inclusions) was visible farther from the substrate. The formation of amorphous zones can be attributed to a high cooling rate, which is up to 108 K/s due to very short micro-arc durations (mostly less than 1ms). On the other hand, the observations carried out at higher magnifications revealed the presence of fine spherical particles with a size < 10nm dispersed in the amorphous matrix (Figure 2b,c). These particles are non-uniformly distributed within the amorphous matrix: a much higher amount was located close to the pores than for the other areas of the coating. Only a small number of NPs has been observed in the areas far from the pores. The EDS maps of chemical composition presented in Figure 2d confirmed that the investigated particles are characterised by the increased signal of Ag with a simultaneous decrease in the number of counts from the other elements: Ti, O, Na, and P (dashed circles). This is an indicator that the Ag-NPs almost entirely aggregate within the amorphous phase, as confirmed by the SAED pattern presented as an inset of Figure 2c. Moreover, Figure 2c also shows the typical halo for amorphous phases, together with numerous spots of intensity originating from the crystalline Ag. The previous observation coincides with the studies carried out by Mosquera et al.22 that showed the addition of Ag tends to inhibit the amorphous-to-anatase transformation for the pulsed cathodic arc-deposited Ag-TiO2 coatings. In addition, our recent microstructure studies of MAO-fabricated TiO2 coatings obtained with Na2HPO4 electrolyte proved the formation of mostly amorphous transformation with a small contribution of the anatase phase residing mainly in the areas near the interface20.

Figure 2
figure 2
TEM/STEM images of MAO coating with Ag addition: BF images presenting a general overview of the coating (a), Ag-NPs in the central part of the coating (b), and agglomerates near the pores (c) as well as STEM/HAADF image showing Ag-NPs and accompanying EDS maps presenting distribution of Ag, O, Ti, P, and Na distribution (d).