Silver nanoparticle-enriched diamond-like carbon implant modification as a mammalian cell compatible surface with antimicrobial properties

The implant-bone interface is the scene of competition between microorganisms and distinct types of tissue cells. In the past, various strategies have been followed to support bony integration and to prevent bacterial implant-associated infections. In the present study we investigated the biological properties of diamond-like carbon (DLC) surfaces containing silver nanoparticles. DLC is a promising material for the modification of medical implants providing high mechanical and chemical stability and a high degree of biocompatibility. DLC surface modifications with varying silver concentrations were generated on medical-grade titanium discs, using plasma immersion ion implantation-induced densification of silver nanoparticle-containing polyvinylpyrrolidone polymer solutions. Immersion of implants in aqueous liquids resulted in a rapid silver release reducing the growth of surface-bound and planktonic Staphylococcus aureus and Staphylococcus epidermidis. Due to the fast and transient release of silver ions from the modified implants, the surfaces became biocompatible, ensuring growth of mammalian cells. Human endothelial cells retained their cellular differentiation as indicated by the intracellular formation of Weibel-Palade bodies and a high responsiveness towards histamine. Our findings indicate that the integration of silver nanoparticles into DLC prevents bacterial colonization due to a fast initial release of silver ions, facilitating the growth of silver susceptible mammalian cells subsequently.


Surface preparation and characterization
For the purpose of preparing silver nanoparticle-containing DLC coatings, a hybrid synthesis method based on the ion-induced transformation of silver-polymer nanocomposites has been used 1 .
The preparation of the silver-containing DLC-coatings is basically divided into three steps which are described in detail below. Five sample sets were coated differently, including four concentrations of silver in the coating (sample sets A, B, C and D) and one silver-free DLCcoating (sample set R). In addition, a set of uncoated substrates was used as a reference. As substrate, corundum-blasted medical Ti6Al4V discs (AST; F 136, MedTitan, R a~5 µm) with 10 mm diameter and 2 mm thickness were used in order to simulate bone implant surfaces.
For material characterisation, non-blasted discs were used.

Colloidal Ag nanoparticle dispersion
As a first step, a colloidal dispersion of stabilized silver nanoparticles has been synthesised. In this step, Ag concentration and nanoparticle sizes in the final coating can be controlled through the concentration of the educts. Silver nanoparticles were synthesized by photochemical reduction of silver nitrate (AgNO 3 ) dissolved in a mixture of ethanol and acetone. The addition of benzoin, which splits into a benzyl radical and a benzoin radical under UV irradiation, leads to a more homogenous particle distribution compared to a reduction in an alcoholic solution 2,3 . The particles were stabilized by steric hindrance, using polyvinylpyrrolidone (PVP). As large sample sets were needed for biological characterization of the samples (each set consisted of 150 samples), long-time stable dispersions had to be prepared. Dispersions described in earlier publications 4 turned out not to be long-time stable, as their silver concentration started decreasing after only one day due to precipitation of silver particles. Lower molar ratios of benzoin to AgNO 3 (benzoin:AgNO 3 1:2) as well as constant concentrations of the basic solutions (AgNO 3 in ethanol 0,1 mol/l; PVP in ethanol 0,6 mol/l; benzoin in aceton 0,05 mol/l) proved to give better stability, with constant silver concentrations for more than five days. Particle sizes, however, were bigger when compared to dispersions of higher molar ratio, independent of the used silver concentration. Fig. 1 shows a TEM bright field image of the silver particles, which were observed on a polyvinylformal-coated copper grid that was dip-coated in the dispersion, with an average particle diameter of 9 nm and a maximum diameter of 30 nm. Four different silver concentrations were prepared (Ag:PVP molar ratio in the solution 1:1, 1:2, 1:10 and 1:20), along with one pure PVP solution without silver. The solutions were stirred and de-gassed in an ultrasonic bath under exclusion of light and afterwards exposed to UV light (Hg vapor lamp) for 6 hours under N 2 atmosphere to initiate the photochemical reduction. As the 1:1 solutions showed precipitations with inconstant and lower than the expected silver concentrations, they were not used throughout the whole study.

Dip-coating
PVP polymer films with the different Ag concentrations were deposited by dip-coating. Film thickness was controlled by retraction speed, resulting in a highly homogenous coating with significant deviations only at the sample edges. All sample sets were coated with a polymer film of 130 nm thickness.

Plasma immersion ion implantation
An experimental vacuum set-up equipped with a magnetron microwave generator (Muegge), an ECR magnet and a high voltage pulse generator (GBS Elektronik) was used for PIII.
Earlier experiments showed optimal polymer-to-DLC transformation parameters for Ne implantation with 20keV implantation energy and a fluence of 5·10 16 ions/cm 2 . Ne and CH 4 were used as processing gases with a ratio of 2.5:1. In this way, the hydrocarbon species deposited out of the plasma provide compensation for sputtering losses. Note that this ratio is highly equipment specific, as in another PIII set-up it was 1.5:1. The working gas pressure was adjusted to 0.5 Pa. 800W power were applied to the microwave generator and 6A current to the ECR magnet to generate the plasma. High voltage pulses of -20kV with 5 µs duration and 200 Hz repetition rate were applied in order to conduct PIII.

Material characterization
Sample characterization included film thickness measurement by profilometry, nano hardness testing, Rutherford backscattering spectroscopy (RBS), Raman spectroscopy and electron microscopy (SEM and TEM). As most characterization techniques cannot be conducted on samples with a high surface roughness, polished Ti6Al4V substrates had to be used instead of the corundum-blasted specimens. Sample composition and areal density were determined by RBS and ERDA, utilizing 1.8 MeV He + ions at an angle of incidence of 70° with a Tandetron accelerator. Mass density was calculated from film thickness and areal atomic density, taking into account only H and C atomic fractions, as the contribution of O and N can be neglected.
In order to obtain pure carbon matrix densities, Ag fractions were disregarded as well in case of silver-containing samples.
Micro Raman spectroscopy was used to investigate the carbon bonding structure of the samples. The applied coatings were excited with the 488 nm line of an argon ion laser of 2 mW beam power. The illumination spot size was about 2 µm, focusing through a 50x microscope objective. The spectra were recorded using a CCD based spectrometer (T64000, Instruments S.A.).
Nano hardness testing of the coatings was conducted on a nanoindentation system (UNAT, Asmec). Indention force was 1 mN, applied with a Berkovich tip.
The distribution of the silver particles over the surface was observed via energy dispersive xray spectroscopy (EDX) on a SEM (ESEM XL 30 FEG, FEI).
The morphology of the nanoparticles in the film was analyzed using cross-sectional TEM.
The images were taken on a 200 keV high-resolution field emission TEM (JEM2100F, JEOL) equipped with an imaging filter (Gatan).