The antibacterial effect of the nanomaterials is due to different processes based on their surface biochemical functionality, partial dissolution of their components and mechanical antibacterial effect of their surface morphology. The utilisation of nanostructured materials as antimicrobial agents depends on their high surface to volume ratio. TiO₂ is a widely used material for scientific studies and different applications in clinical practice due to its biocompatibility and high stability at chemical treatments and environmental conditions. The modification of the TiO₂ band gap through doping with different chemical elements does not affect its ultraviolet light activity and increases the photocatalytic activity in sunlight. It is known that transition metals (W, V, Ag, Cu etc.) or non-metals (S, C, N₂) doped TiO₂ has a lower band gap than pure TiO₂ and can be used to increase the production of reactive oxygen species (ROS) under visible light irradiation in solutions containing undesired pollutants Fujishima et al. (2008), Huang et al. (2012). Doped with silver, TiO₂ nanoparticles (NPs) prepared by sol-gel technique were successfully applied for deactivation of different microorganisms (Escherichia coli, Pseudomonas aeruginosa, Bacillus subtilis and Staphylococcus aureus). A concentration dependence was established for the bacterial deactivation with Ag content in TiO₂ thin films with a quantity of 1, 3, 5 and 7 μg/ml Bahadur et al. (2016). Wong et al. (2015) reported increasing the visible-light response and durability of silver nanoparticles impregnated in TiO₂:(N) thin films against pathogens like E. coli, Streptococcus pyogenes, S. aureus and Acinetobacter baumanii.

TiO₂ porous coatings with different Ag concentration (0, 0.95, 1.36 and 1.93 wt%, respectively) prepared through combining magnetron sputtering with micro-arc oxidation were presented by Zhang et al. (2014). The interaction of the coatings with bacterial culture is characterised by different rates of Ag dissolution - initially high rate followed by a period of low rate. This process provided relatively long-term antibacterial activities and met the clinical requirements. No cytotoxic effect is found for the surfaces with Ag concentration lower than 0.95 wt%.

Heidenau et al. (2005) and Haenle et al. (2010) demonstrated that Cu induces bacterial cell toxicity without decreasing the cell biocompatibility. The osteoblastic adhesion, spreading, early proliferation and late differentiation on Cu-TiO₂ coatings were significantly enhanced due to the inhibitory effect of Cu and existence of a suitable porous nano-structured surface Han et al. (2008), Zhao et al. (2009). Excellent antibacterial properties of Cu-doped TiO₂ coatings were reported by Raffi et al. (2010). They accept that the reason for the antibacterial property of Cu NPs is their adhesion on to the bacterial surface due to their opposite charge with respect to bacteria. When they encounter each other, the reduction reaction occurs at the bacterial cell wall, resulting in formation of cavities or pits.

According to Zhu et al. (2013), Wu et al. (2014), Cu as an essential trace element participating in a variety of metabolic activities in living organisms, might be a more effective antimicrobial agent than Ag due to its low cytotoxicity. According to Norambuena et al. (2016), amongst metals with antibacterial properties, copper has shown superior in vitro antibacterial performance while maintaining an acceptable cytotoxicity profile. A thin film containing copper could prevent early formation of biofilm which limits periprosthetic infections. TiCuO containing 20%, 40% or 80% copper prepared through high-power impulse magnetron sputtering (HiPIMS) showed an increased biofilm and planktonic cell density reduction.

Zhang et al. (2016) used Micro-Arc Oxidation synthesised TiO₂ coatings doped with different doses of Cu²⁺ (0 - 1.93 wt%) on S. aureus. Compared to Ti, the adhesion and proliferation of fibroblast can be significantly enhanced on the TiO₂ coating with 0.67 wt% Cu, while these processes are greatly inhibited with 1.93 wt% Cu. Cu²⁺-doped TiO₂ coatings show good antibacterial properties, including contact-killing and release-killing of S. aureus.

Rtimi et al. (2014) compared the antibacterial effect of Ag and Cu surfaces as a functionalisation of polyester by very thin sputtered Ag and Cu layers inducing accelerated E. coli killing in the dark. In the case of Ag, the most suitable sample leading to a bacterial inactivation within 2h required a sputtering time of 75 s or 150 s. The deactivation kinetics of E. coli on Cu-polyester was observed in the sample which was sputtered for 30 s and 60 s. It was established that, at longer HIPIMS sputtering times, the inactivation time for E. coli becomes shorter. A local softening of the fibres at the contact sites with Cu may play a significant role in the inactivation time of E. coli. These films preclude formation of biofilm of all kinds of infectious bacteria in hospitals, schools and public places.

Santhosh and Natarajan (2015) reported excellent biocidal characteristics for the epoxy resin containing TiO₂ doped with Ag⁺. They found significant inhibition of the biofilm development by two pathogenic bacteria, Staphylococcus aureus and Escherichia coli K-12.

He et al. (2017) carried out a review of micro-arc oxidation assisted magnetron sputtering (MS) of Ag, Cu and ZnO to endow Ti-implant materials with antibacterial properties.

To summarise, no detailed research has been reported for antimicrobial activity of TiO₂ thin films doped with Ag and Cu and deposited through r.f.magnetron co-sputtering. This is the reason to study the antibacterial effect of these films and compare their influence on Gracilicutes and Firmicutes bacteria.

Inhibitory effect in static conditions

The result from the diffusion assay test have shown that the inhibition zone was different for the test bacteria: 11 mm for E. coli and S. epidermidis, 12 mm for B. cereus and 13-14 mm for Pseudomonas sp. It can be concluded that E. coli and S. epidermidis are the most resistant to the tested films.

Disinfection time

The disinfection time of the studied thin films was determined for all tested bacteria during cultivation in the liquid medium.

The disinfection time of the coating to Gram-positive pathogen, B. cereus is presented in Fig. 1. It is evident that the bacterial growth in the presence of TiO₂:Ag:Cu thin films was strongly inhibited. Both methods used to obtain the growth curve demonstrate similar dependencies. The number of bacteria in a logarithmic scale sharply decreased at the 4^th hour and the optical density dropped at the 3^th hour from the beginning of the experiment, which determined the process of disinfection. There are no surviving cells till the 48^th hour, which also suggests the killing effect of the thin film on the spores of Bacillus cereus strain.

The effect of tested thin films on the other Firmicutes bacteria, like S. epidermidis can be seen in Fig. 2a and Fig. 2b. In this case, a slower bactericidal effect and no surviving cells after the 24^th hour were observed. The result could be due to the formation of a biofilm on the surface, typical for this kind of bacteria and lower growth rate (0.2 in comparison with the 0.4 for Bacillus). As is well known, the cells in a biofilm are more resistant to the environmental factors.

Fig. 3 demonstrates the bactericidal effect of the same thin films TiO₂:Ag:Cu on P. putida growth, determined by the cultivation method Fig. 3a. The results obtained by the optical density measurements at λ = 610 nm are shown in Fig. 3b. The bactericidal effect on Pseudomonas strain occurred faster and the result was visualised as a morphological destruction of bacterial cells by a Scanning Electron Microscope (JOEL JSM 5510). The damaged cells can be viewed in two optical magnifications in Fig. 4.

The influence of nanostructured thin films TiO₂:Ag:Cu on the growth of S. enterica is represented in Fig. 5a and Fig. 5b. The bactericidal effect on Gracilicutes bacteria (S. enterica and Pseudomonas sp.) is faster than the bactericidal effect on the Firmicutes, which supposes the higher sensitivity to the nanocomposite material. The effect is not the same on E. coli - the thin film has only a bacteriostatic effect on it. Strong retention in the bacterial growth was observed from the beginning until the 10^th hour of the experiment. The results for E. coli are presented inFig. 6.

Atomic Flame Absorption Spectrometry

The bactericidal concentration of Ag and Cu (nanoparticles and ions) in solution was measured for the representatives of two types of the tested bacteria - Firmicutes and Gracilicutes. The values for blank control (organic nutrient medium with a thin film without bacteria) were determined as 5.2 mg/l for Ag and as 4.2 mg/l for Cu. For the samples with bacteria (B. cereus) and thin films, these values were 4.8 mg/l for Ag in organic growth medium and 3.6 mg/l for Cu also in organic medium; the difference between the control and the sample is the amount engulfed by the Bacillus cells (10⁵ CFU/ml) - 0.6 mg/l Cu and 0.4 mg/l Ag.

In the synthetic culture medium ISO 10712, the content of dissolved Ag was 1.04 mg/land Cu was 2.4 mg/l in the sample with 10⁸ CFU P. putida per ml.

In this study, five bacterial strains were exposed to TiO₂:Ag:Cu thin films. The results indicated that the tested materials have a bactericidal effect on the bacterial cells in different time periods.

The bactericidal effects of TiO₂:Ag:Cu was evident after 30 min and 90 min of exposure of P. putida and S. enterica determined by the Koch's method. Only on E. coli, known as a highly adaptable bacterial species, the effect was bacteriostatic. However, a strong retention until the 10^th hour of the treatment was observed. For the tested Firmicutes bacteria, the bactericidal effect was observed at the 4^th and 8^th hours for B. cereus and S. epidermidis respectively. This effect is within the time limits which do not allow biofilm formation. As reported by Thuptimdang et al. (2015) for monitoring the changes in bacterial activity during the biofilm formation, it is necessary to measure the bacterial quantity every 3 h for the first 24 h and every 12 h between 24th and 72^th hours. In this case, the inactivation of the bacteria occurred much faster. Scanning electron microscope images reveal a rough cell envelope instead of a smooth Pseudomonas capsule and leakage of the cell content around the cells (dark zones) which proved the destruction of the cell envelope. Similar effects were expected for the other test bacteria. The damage of the cell envelope leads directly to the leakage of cell content materials causing cell death. This proves the mechanical demolition by nanoparticles and chemical destruction by ROS, as was established by other authors Sondi and Salopek-Sondi (2004), Zavilgelsky et al. (2011), Zhang et al. (2014).

At the end of the experiment, only 2.8% retention in the maximum growth rate at the 24^th hour for E. coli was observed. In contrast, all the other bacterial strains were completely deactivated. The maximum growth rate counted on the controls data obtained by Koch's method (test bacterial strains without nanomaterials) were 0.404, 0.279, 0.214, 1.472 and 0.283 for B. cereus, S. epidermidis, P. putida and S. enterica respectively.

The toxic effect was confirmed by an assessment of the bacterial dehydrogenase activity. As is well known, this activity is proof of normal cell metabolism and its inhibition is due to destruction of the bacterial cell wall and membrane. The stress response in the treated E. coli cells was the dehydrogenase acivity increase in the sample in comparison to the control. The other bacterial strains, in contact with the E. coli, have shown no enzyme activity when tested with INT and TTC. Only E. coli retains the dehydrogenase activity at the 24^th hour; the formazan quantity in the sample measured was OD₄₇₀ = 0.804 compared to the OD₄₇₀ = 0.671 in the control and the calculated enzyme activity in the sample was 0.253 FNU (Formazin Nephelometric Units), compared to the control (0.08 FNU). The bacteriostatic effect and the adaptation of E. coli to the tested nanomaterials could be explained by the ability of this bacterial species to activate special transport pumps for export of the metal ions out of the cell. This fact is the mechanism of stress response as has been reported by Stoyanov et al. (2003). They found that a special transport pump coded by CopA gene of E. coli can control cytoplasmic silver, in addition to copper content.

The tested material demonstrated bactericidal concentration on P. putida at 3.77 x 10^-7 mol/cell for Cu and 9.64 x 10^-11 mol/cell for Ag. For the Firmicutes representative B. cereus, the toxic effect was observed at 3.71 x 10^-8 mol/cell for Ag and 9.44 x 10^-6 mol/cell for Cu. These results confirmed that the Gracilicutes bacteria are more sensitive to dissolved Ag and Cu from the studied thin films.

The results from the atomic absorption spectrometry proved a partial dissolution of metal atoms and ions from the tested thin films. Probably as a result of the faster penetration of Ag and Cu atoms in the cell, they cause a quicker toxic effect on the Gracilicutes bacteria. Moreover, due to the established low concentration of Ag and Cu in our experiment, low cytotoxicity to human cells can also be expected. A concentration of 5 μg/ml Ag nanoparticles (NPs) affects the epithelial cell line, according to Comfort et al. (2011).

The neurotoxicity of Ag NPs with an average size of 15 nm at a concentration of 10 μg/ml for 24 h was investigated using the dopaminergic neuronal cell line PC12 by Wang et al. (2009). From the results of this study, it can be inferred that the silver is involved not only in the oxidative stress, but also in alteration of the enzymatic functions which play an important role in the dopamine depletion.

The effects of the tested thin films depend on the structure of the bacterial cell wall Allaker (2010), Wang et al. (2012). In these cases, the peptidoglycan thickness plays an important role in the interaction with the nanoparticles, but this effect might be modulated by the other features of the environment, such as pH, the metal structure or the carrier Allaker (2010), Seil and Websters (2012), Wang et al. (2012). A similar effect has been observed in the study of Spacciapoli et al. (2001), Sondi and Salopek-Sondi (2004), Taglietti et al. (2012). A ten times higher concentration was obtained for Ag for Bacillus cells in neutral pH=7.0 than that of Pseudomonas at slightly acidic pH=6.0 and a longer time for the bacterial contact killing - 4 h for Bacillus in contrast with Pseudomonas sp. - only 30 min for their inactivation. According to Foster et al. (2010), the mechanisims of action include a release of more ionic silver than Ag⁰ atoms.

The study of Santhosh and Natarajan (2015) indicates that the effective dispersion and the optimum release of biocidal agents was responsible for the anti-biofilm activity of the surface. Quiñones-Jurado et al. (2014) reported that the surface plasmon resonance of Ag NPs plays an important role in dependence on the NPs' size and increases the bactericidal effect of Ag/TiO₂ nanocomposite. They supposed that, under visible light, the reactivity of TiO₂ cannot be amplified when it supports Ag NPs which have an inactive photocathalytic surface. In this case, the reactivity of Ag and Cu metals is increased probably due to the synergistic effect of Ag and Cu in the TiO₂ matrix. Also, nanoparticles of Cu possess a high oxidative redox potential leading to a fast bacterial loss of viability. The Cu ions entering the bacterial cytoplasm bind S, N and COO- electron donor negative groups of the cell wall.

The antibacterial activity of TiO₂ thin films doped with Ag and Cu using two types of Gram-negative and Gram-positive test bacteria with clinical significance (Gracilicutes and Firmicutes bacteria) were tested. The antibacterial activity of the thin films was evaluated by the classical Koch's method and optical density and dehydrogenase activity measurements. The bactericidal effect was established at different time periods between 30 min - 90 min for Pseudomonas sp. and S. enterica. The Firmicutes bacteria B. cereus and S. epidermidis were killed at the 4^th and 8^th hours of the treatment respectively. The effect against E. coli was bacteriostatic until the 10^th hour. The studied TiO₂:Ag:Cu thin films have potential for application as antibacterial agents on clinically significant bacterial strains. For the future perspective, other technological variants with similar combinations will be tested.