Nanoscale copper and silver thin film systems display differences in antiviral and antibacterial properties

The current Coronavirus Disease 19 (COVID-19) pandemic has exemplified the need for simple and efficient prevention strategies that can be rapidly implemented to mitigate infection risks. Various surfaces have a long history of antimicrobial properties and are well described for the prevention of bacterial infections. However, their effect on many viruses has not been studied in depth. In the context of COVID-19, several surfaces, including copper (Cu) and silver (Ag) coatings have been described as efficient antiviral measures that can easily be implemented to slow viral transmission. In this study, we detected antiviral properties against Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2) on surfaces, which were coated with Cu by magnetron sputtering as thin Cu films or as Cu/Ag ultrathin bimetallic nanopatches. However, no effect of Ag on viral titers was observed, in clear contrast to its well-known antibacterial properties. Further enhancement of Ag ion release kinetics based on an electrochemical sacrificial anode mechanism did not increase antiviral activity. These results clearly demonstrate that Cu and Ag thin film systems display significant differences in antiviral and antibacterial properties which need to be considered upon implementation.

www.nature.com/scientificreports/ Pd, or Ir) within an electrolytic environment (such as biological fluids) the less noble Ag corrodes in favor of the more noble part (it is "sacrificed"). We have demonstrated that such sacrificial anode surfaces exert much higher antibacterial effects compared to pure Ag surfaces with much higher total Ag + due to the electrochemically driven enhanced dissolution of Ag. Even though the antibacterial activity of Ag is well understood very few studies describe the antiviral effects of Ag and most of them deal with Ag nanoparticles 13 . Recently, a virucidal effect of sputtered Ag nanocluster/silica composite on face masks was reported 14 .
In contrast to Ag surface materials the virucidal properties of Cu and Cu alloy surfaces have been studied more frequently 13 . The virucidal effects of Cu rely also on the release of ions and both Cu ionic species (Cu II and Cu I) contribute to the biocidal activity [15][16][17] .
Remarkably, there is an absence of comparative studies on Cu and Ag opposing their bactericidal and virucidal properties within same experimental setups using equal coating parameters and equal metal ion concentrations to avoid variations in testing conditions. The results of our study are relevant for the design of Cu and Ag containing surface materials with special emphasis on their virucidal activities. Thus, we aimed to analyze the antiviral properties of surfaces coated with Cu or Ag as well as several Ag-based sacrificial anode surfaces including combinations of Cu and Ag for possible synergistic effects and compare antiviral against antimicrobial performance.

Results
The results were achieved using different sputtered thin film surfaces (for details see Experimental Methods): (I) continuous and dense thin films of Ag and Cu (thickness 50 nm) and (II) nanostructured surfaces with high surface areas. The nanostructured surfaces were synthesized by (a) sequentially depositing Ag on Pt, Cu on Ag or (b) by co-sputtering Ag and Pt as well as Cu and Ag. We call the resulting surface structures "nanopatches", i.e., nanoislands which are formed by the two metals. A schematic workflow is depicted in Fig. 1. In case of sequential sputtering (Ag on Pt, Cu on Ag) the elements tend to be more separated as compared to co-sputtering which tends to mix elements on the atomic scale (Ag & Pt, Cu & Ag) and rather forms an alloy (forced solid solution) compared to the co-existence of elemental films (Ag on Pt, Cu on Ag). The latter is expected to have better sacrificial anode properties. We synthesized nanopatches of two different thicknesses, which offer different volumes of material (e.g., thin Ag/Pt vs. thick Ag/Pt) for releasing metal ions ( Table 1). Examples of scanning electron microscopy (SEM) and transmission electron microscopy (TEM) images of the nanoscale films are shown in Fig. 2. The 50 nm thick films are continuous and show different nanocrystalline surface microstructures: the Ag film is rougher and shows larger grains ( Fig. 2A) as compared to the Cu film (Fig. 2B). The SEM and TEM images ( Fig. 2C-F) of nanopatches (Ag, Ag & Pt, Ag & Cu) show that these nanostructures are discontinuous. They can be regarded as immobilized nanoparticles and thus offer a larger surface area than the continuous films. A detailed TEM investigation of the nanostructure of Ag-Pt nanopatches can be found in reference 12 .
In order to qualitatively compare antimicrobial and antiviral properties of sputtered Ag and Cu surfaces, we first evaluated the antibacterial properties of thin Ag/Pt and Ag/Cu sacrificial anode nanopatches. Bacterial  Pure Ti thin films as well as thin nanopatches of pure Pt, Ag, and Cu served as controls and exhibited no significant antibacterial activity against S. aureus (Fig. 3A). Similarly, bacterial growth was not affected by codeposited thin Ag/Pt nanopatches indicating the absence of a sacrificial anode effect (Fig. 3B). In contrast, sequentially-deposited thin Ag/Pt nanopatches as well as thin Ag/Cu nanopatches both co-and sequentiallydeposited, effectively prevented bacterial growth after 24 h of incubation (Fig. 3B).
As it is generally accepted that the antibacterial activity of Ag und Cu is strongly related to the release of ions and their interaction with cellular components and processes, solutions of silver acetate (AgAc) and copper sulfate (CuSO 4 ) were used as ionic controls for the antibacterial activity of Ag and Cu towards the gram-positive bacteria S. aureus 13,14 . Significant antibacterial effects were detected for AgAc at concentrations ≥ 1.0 µg/mL, whereas CuSO 4 provoked significant effects starting at concentrations ≥ 5.0 µg/mL (Fig. 4).
These results indicate that the absence of antibacterial effects of pure Ag and Cu nanopatches are due to an insufficient ion release from these structures, while the combination of Ag and Pt as well as Ag and Cu leads to enhanced antibacterial activity based on electrochemically driven enhanced dissolution of Ag and Cu, respectively (Fig. 3). Previously, we demonstrated such sacrificial anode effects for nanoparticular and nanostructured Ag/Pt systems 11,12,15 . Regarding the Ag/Cu system, in addition to a possible sacrificial anode effect of Ag on Cu, a combination effect of the two antibacterial metals might be considered 10,12 .
After demonstrating antibacterial effects of sputtered Ag and Cu surfaces as well as ionic Ag and Cu solutions, in accordance with previous results, we aimed in a next step to analyze potential antiviral effects of these surfaces against SARS-CoV-2. Viral contamination was mimicked upon inoculation of SARS-CoV-2 onto surfaces for either 1 h or 24 h, before viral titers were determined as TCID 50 /mL (Fig. 5). A pronounced antiviral efficacy was observed for thin Cu films sputtered on Si/SiO 2 pieces after 1 h and 24 h of incubation reducing viral titers for 3 log 10 and 4.5 log 10 , respectively, whereas Ag films did not reduce viral titers (Fig. 5A). In contrast, sputtered Ag surfaces only marginally and not significantly affected viral titers upon 24 h exposure. ICP-MS analysis suggested that Cu ions in the range of 10,000-14,200 µmol/L are released from these films during 24 h, indicating that this amount is sufficient for efficient inactivation of SARS-CoV-2 ( Supplementary Fig. 1). While Ag nanopatches did not affect viral infectivity, nanopatches with a thick layer of Cu reduced viral titers by 1 log 10 when incubated together with SARS-CoV-2 for 24 h (Fig. 5B), resulting in an ion release of around 720 µmol/L ( Supplementary  Fig. 1). Interestingly, Cu & Ag and Cu on Ag nanopatches enhanced the antiviral properties of solely Cu, with a significant antiviral effect after 24 h, despite similar amounts of Cu being released (870 µmol/L, Supplementary  Fig. 1). This could be related to the improved electrochemical Cu ion release mechanism in the sacrificial anode structure, where Cu can be easily oxidized and released to the medium as Cu ions, due to the presence of Ag, support the reduction process. Thin layers reduced viral titers by 2.2 log 10 , whereas thick layers increased virus inactivation resulting in reduction factors of 3.9 log 10 (Fig. 5C). There was no difference in the antiviral effect detected between co-and sequentially deposited nanopatches. This demonstrates that pure Cu films offer the highest antiviral effect (Fig. 5A), which correlates to the highest release of ions (13,000 µmol/L, Supplementary  Fig. 1). In contrast, co-deposited and sequentially deposited Ag-Pt nanopatches did not reduce viral infectivity within 1 h incubation. A mild, but not significant antiviral effect (1 log 10 reduction of viral titers) was observed after 24 h incubation with thin Ag & Pt and thin Ag on Pt nanopatches (Fig. 5D). In order to further examine the observation that Ag-coated surfaces do not affect viral infectivity, we used silver acetate solutions (AgAc) as a reference to test antiviral properties of Ag at higher ion concentrations as compared to what can be released from a thin film surface or a nanopatch structure. We included concentrations ranging from 1 µg/mL up to 50 µg/mL Ag and inoculated the solution with SARS-CoV-2 containing supernatant for 1 h or 24 h before determining viral infectivity as TCID 50 /mL. Only concentrations equal to or higher than 25 µg/mL displayed antiviral properties www.nature.com/scientificreports/ and completely abolished infectivity of SARS-CoV-2, however, only upon prolonged incubation of 24 h (Fig. 5E). In contrast, virus exposure towards CuSO 4 solution in a similar experimental setup had no effect on viral titers (Fig. 5F). In conclusion, we demonstrate a clear antiviral effect of Cu-coated surfaces against SARS-CoV-2 within 1 h exposure, whereas Ag-coated surfaces did not influence viral infectivity.

Discussion
Rapid and effective prevention measures are urgently needed in order to combat microbial and viral diseases. Fomite transmission via contaminated surfaces was described for a variety of microbes, including several viruses and has been discussed in the context of COVID-19. Even though surfaces are believed to play a minor role for the spread of SARS-CoV-2, public health intervention strategies still rely on disinfectant procedures in order to reduce viral transmission. Various surface coatings exhibiting topographically and/or chemically induced antimicrobial activity have been suggested, including nanostructured materials as well as materials containing antimicrobial agents (e.g. antibiotics, antiviral drugs, nanoparticles) such as commercially available antibacterial/ www.nature.com/scientificreports/ antiviral foils, textiles, paints, and many more 19,20 . Although disinfection can be effective in the prevention of infection spread, the biologically active agents of many widely used surface and hand disinfectants might also be hazardous to humans and the environment, especially at prolonged application or misuse. In particular, skin and ocular irritation as well as chemical burns to the respiratory track might occur. Disruption of the normal skin flora, which normally represents a protection barrier to harmful agents, can even enhance the risk of infection 21 .
In line with van Doremalen et al., SARS-CoV-2 was less stable on sputtered Cu surfaces compared to all other thin films of this study 18 . A thicker film led to a more pronounced antiviral effect. Cu nanostructures in contrast reduced the antiviral effect, which can be attributed to the reduced amount of released Cu ions. Interestingly, combining Cu with Ag by either co-sputtering or sequentially depositing Cu on Ag, strengthened the antiviral www.nature.com/scientificreports/ effects for the nanoscale structures (Cu nanopatch in comparison to Ag-Cu nanopatch) especially after prolonged incubation of 24 h as monitored by limited dilution assays. In contrast, thin films of pure Ag or Ag-based sacrificial anode nanopatches (Ag in combination with Pt) displayed no significant capacity to inactivate SARS-CoV-2 ( Fig. 5B-D), despite efficient release of Ag ions. Ionic control experiments demonstrated that a minimum concentration of ≥ 25 µg/mL is necessary to significantly reduce viral titers when incubated for 24 h. Reducing the silver acetate concentration or incubation period completely abolished any antiviral effects (Fig. 5E). This means that only very high concentrations of Ag ions have an effect on SARS-CoV-2. In the past, most of nanoparticles (NPs) used for studies on antiviral activities were made of Ag (for review see 20,21 ) For instance, HIV-1 infectivity was inhibited at concentrations of 25 µg Ag/mL and above 19,21 . Recently, Jeremiah et al. reported that Ag NPs (10 nm) were already effective in lower concentrations inhibiting extracellular SARS-CoV-2 at concentrations between 1 and 10 µg/mL 22 , however, it is possible that nanoparticles exert an additional particle effect on viruses after binding such as steric inhibition. Also, a mixture of nanoparticles in solution containing Au-NP (1 µg/mL) and Ag-NP (5 µg/mL) was effective against SARS-CoV-2 and influenza viruses, however, this mixture also contained large amounts (60 µg/mL) of ZnO-NPs and additionally ClO 2 23 . In general, it must be considered that studies with nanoparticles are not directly comparable to studies on thin films (or thin bulk material) due to several facts such as enlarged surface, high ion release, uptake into cells, electrostatic interaction, steric interaction in case of the small particles, and other factors. Our results on thin film surfaces, nanopatches and ionic solutions clearly indicate differences between antiviral and antibacterial activities of Ag. For inactivation of SARS-CoV-2 about tenfold higher concentrations of Ag ions are required compared to efficient antibacterial concentrations. The reason may be related to the difference in the nature of viruses vs. bacteria. As living organisms, bacteria offer much more Ag-sensitive metabolic processes such as energy generation or cell proliferation compared to a virus particle. Furthermore, some bacteria such as S. aureus exhibit a significant inherent tolerance of growth at high Cu concentrations 24 . However, a copper sulfate solution up to 50 µg/mL did not inactivate SARS-CoV-2 (Fig. 5F). In our study the ionic Ag and Cu solutions had similar concentration up to 50 µg metal/mL to allow www.nature.com/scientificreports/ direct comparison. Whereas ionic Ag is antibacterial at concentrations of ≥ 1 µg/mL, ionic Cu is required at much higher concentrations to inhibit bacterial growth (e.g. 10 mM CuSO 4 ). Several groups observed similar differences in antibacterial efficiency between Ag and Cu ions [24][25][26] , however, at lower total ion concentrations due to the water test medium 26 . Thus, at equimolar concentrations Ag ions shows better antibacterial performance compared to Cu ions. Virus inhibition by soluble Cu ions might require higher concentrations above the used maximal values (50 µg/mL) in our experiments. In contrast to ionic species, antibacterial and antiviral activities of solid Ag and Cu sputtered surfaces have led to different results. Nanopatches of pure Ag neither exhibit antibacterial nor antiviral activities. Obviously, Ag ion release is insufficient under these experimental conditions. To overcome this limited Ag ion release, the more active ion-releasing sacrificial anode surfaces can be used as they exhibit much better antibacterial www.nature.com/scientificreports/ efficiency even at lower total Ag content 11,12 as was also observed here for Ag/Pt samples (Fig. 3). However, even these antibacterial Ag/Pt samples failed to induce any antiviral effect which indicates again the need of higher Ag ion concentrations to reach antiviral activity.
Remarkably, our study shows that solid-state Cu either as a dense film or as nanopatches is able to induce antiviral activity, but not solid-state Ag. The antiviral effects are dependent on the total Cu amount (thickness of the sputtered Cu and ion release) and on the time of sample exposure.
It was reported that solid-state cuprous compounds exhibit efficient antiviral activities, whereas those of solidstate Ag are markedly lower 27 . In particular, the inactivation of influenza virus HA and NA surface proteins are affected by the exposure to Ag and Cu 28 . Solid-state Ag is less susceptible than solid-state Cu to surface oxidation under the experimental conditions to release ionic species 29 . It is known that several metabolic products of Cu such as cuprous oxide (Cu 2 O), sulfide (Cu 2 S), or chloride (CuCl) exhibit high antiviral activities and Cu surfaces retain their anti-infective properties even after oxide formation 27,30 .
Although there are numerous reports on antibacterial and antiviral effects of Ag or Cu and their related compounds, a direct comparison of Ag and Cu as performed in our study, including both bacterial and viral data is rarely found and a high variability of the reported study methods makes such direct comparison difficult 31 .
Recently, the rapid inhibition of SARS-CoV-2 on copper-silver (Cu-Ag) nanohybrid surfaces was reported and the authors attributed the role of primary SARS-CoV-2 inhibition to the Cu content 32 . Our results showed that the combination of Cu and Ag (Cu & Ag nanopatches) exerted significantly more antiviral activity than similar nanopatches consisting of Cu or Ag alone. This effect can be attributed to the enhanced Cu ion release of Cu & Ag nanopatches concurrent with a profound decrease in Ag-ion release compared to Ag nanopatches. Apparently, the presence of silver is nevertheless necessary to enable antiviral effects via nanopatches. These findings suggest that electrochemical processes such as a sacrificial anode effect may play important roles. However, the exact mechanisms for this still need to be clarified in further studies.
Taken together, biocidal surfaces could provide constant antiviral and antibacterial efficacy against reoccurring contamination, thus reducing the spread of certain pathogens, given that the surface stays clean and is not used up, whereas surface disinfection has to be reapplied with every contamination 33 . The antimicrobial activity of Cu-based materials and surfaces was demonstrated against different pathogens, including SARS-CoV-2, MRSA (meticillin-resistant S. aureus), VRE (vancomycin resistant enterococci), and other nosocomial pathogens, while techniques such as cold-spray coating or Cu-impregnation would circumvent the need to completely replace existing surfaces [34][35][36] . However, incubation periods greater than 1 h are not applicable for many administrations and prevention measures should therefore be critically evaluated with respect to the targeted pathogen.

Methods
Sputter deposition of thin films. Thin film samples were prepared by direct current magnetron sputtering in Ar atmosphere (0.5 Pa) at room temperature on thermally oxidized Si substrates (Si/SiO 2 , 4.4 mm × 4.4 mm), which were placed on a rotating substrate plate. Sputter targets, 2-inch diameter, of Cu (purity 99.99%, Evo-Chem), Ag (99.99%, EvoChem) and Pt (99.99%, ESG Edelmetall Services) were used. Data on all films are listed in Table 1. The nominal thickness of the films was calculated from the pre-determined sputter rates of the used elements and the indicated power levels. Examples of scanning electron microscopy (SEM) and transmission electron microscopy (TEM) images are shown in Fig. 1. Antibacterial tests. Bacterial tests were performed with Staphylococcus aureus (S. aureus, DSMZ 1104) obtained from the German Collection of Microorganisms and Cell Cultures (Braunschweig, Germany). S. aureus cultures were grown overnight in brain-heart infusion broth (BHI broth, bioMerieux, Lyon, France) at 37 °C using a shaking water bath (JULABO GmbH, Seelbach, Germany) and bacterial concentrations were determined by turbidity measurements (Densichek turbidity photometer, bioMerieux). The adhesion and proliferation of S. aureus on the different nanopatch samples were analyzed using a drop-based experimental setup as reported previously 10,11 . Briefly, 30 µL of a bacterial solution in BHI broth containing 10 4 cells per mL (CFU/ mL) were placed in the middle of each test sample followed by incubation for 24 h in a humid chamber (water saturated atmosphere) under cell culture conditions (37 °C, 5% CO 2 ). Subsequently, the drops were aspirated, serially diluted (1:10 4 ) and plated on Columbia blood agar plates (bioMerieux) for quantitative analysis of planktonic bacteria. Qualitative analysis of the adherent bacteria was performed by SYTO-9 staining (Molecular Probes, Invitrogen, Karlsruhe, Germany) and detected by fluorescence microscopy (BX61 microscope, Olympus, Hamburg, Germany).
Silver acetate (AgC 2 H 3 O 2 , AgAc) and copper sulfate (CuSO 4 ) solutions were used as ionic controls for the antimicrobial activity of Ag and Cu, respectively. Each solution was prepared in sterile ultrapure water and normalized to the total metal content (i.e., for example 100 µg/mL AgAc contains 100 µg/mL Ag). Different bacterial concentrations (10 3 , 10 4 , 10 5 CFU/mL) of S. aureus were incubated in BHI for 24 h with different concentrations of AgAc and CuSO 4 solutions (0.5, 1.0, 2.5, 5.0, 10, 25, 50 µg/mL) in 96-well microplates at a total sample volume of 200 µL under cell culture conditions. Subsequently, quantification of viable cells was performed by the AlamarBlue assay. Therefore, the bacterial suspensions were incubated with 20 µL of the AlamarBlue reagent (Invitrogen) until visible color change and the fluorescence intensity was analyzed at 590 nm by a microplate reader (FLUOstar Optima, BMG LABTECH GmbH, Ortenberg, Germany). The data of the treated cultures (mean ± SD) are given as percentage of the untreated controls (bacteria cultured without AgAc or CuSO 4 ). www.nature.com/scientificreports/ and 24 h at room temperature (22 ± 1 °C) and 32 ± 1% humidity. Virus was recovered by adding 225 µL of Dulbecco's modified Eagle's medium (DMEM, supplemented with 10% (v/v) fetal calf serum (FCS), 1% (v/v) nonessential amino acids, 100 IU/mL penicillin, 100 µg/mL streptomycin and 2 mM L-Glutamine). Subsequently, viral titers were determined by an endpoint dilution assay performed on VeroE6 cells (kindly provided by C. Drosten and M. Müller) seeded at 5 × 10 4 cells/mL in DMEM one day prior titration. The remaining TCID 50 was calculated according to Spearman and Kärber. According to the antimicrobial tests silver acetate and copper sulfate solutions were used as ionic controls. Therefore, SARS-CoV-2 was spiked with different concentrations of AgAc and CuSO 4 solutions (1.0, 2.5, 5.0, 10, 25, 50 µg/mL) and was incubated for 1 h and 24 h at room temperature. Remaining viral titers were again quantified by an endpoint dilution assay followed by TCID 50 calculation.

Antiviral tests.
Inductively coupled plasma-mass spectrometry (ICP-MS). ICP-MS was used to determine the amount of ions released into the medium containing the virus during contact with the surface of interest. Dilutions were adjusted to a calibration range between 1 and 100 ppb, which represents the higher concentration part of the linear calibration curve. In accordance with the calibration measurements, the samples were acidified with 2% HNO 3 . For allowing multiple measurements or to further adjust the dilution of the same sample, 10 mL sample solution were obtained by adding 300 µL of an ultra-pure 69% HNO 3 (Roth, Supra quality), thorough mixing and filling up to 10 mL to adjust an acid concentration of 2%. The samples were analyzed with an iCAP RQ system from Thermo Fisher Scientific in KED mode.
Statistics. Statistical analysis of antibacterial effects was performed by one-way ANOVA with Bonferroni post hoc test. Statistical analysis of antiviral effects was performed by two-way ANOVA in a mixed-effects analysis with Dunnet's multiple comparison. p values ≤ 0.05 were considered as statistically significant.

Data availability
All data generated or analysed during this study are included in this published article.