Shielding surfaces from viruses and bacteria with a multifunctional coating

The spread of viral and bacterial pathogens mediated by contact with surfaces is a leading cause of infection worldwide. COVID-19 as well as the continuous rise of deaths associated with antibiotic-resistant bacteria highlights the need to impede surface-mediated transmission. We report a sprayable coating with an intrinsic ability to resist the uptake of bacteria and viruses from surfaces and droplets, such as those generated by sneezing or coughing. Our coating also provides an effective microbicidal functionality against bacteria, providing a dual barrier against pathogen uptake and transmission. This antimicrobial functionality is fully preserved following scratching and other induced damage to its surface or 9 days of submersion in a highly concentrated suspension of bacteria.


Introduction
COVID-19 has highlighted the global risk associated with the transmission of pathogens, and its rapid and severe impact on human health and the global economy [1][2][3] . There is broad evidence that surfaces facilitate pathogen dissemination, leading to widespread transmission throughout the population 4 .
Surface contamination also plays a major role in the evolution of antibiotic-resistant bacterial strains [5][6][7] .
Alternatives to traditional surface disinfection approaches are necessary as these are presenting declining e cacy and are often laborious and ineffective as recontamination is triggered episodically 8,9 .
Alternative approaches to prevent bacterial contamination of surfaces relies on the use of bactericidal coating materials including, silver, copper, zinc and TiO 2 along with other inorganic nanoparticles and quantum dots 10,11 . Among these, silver nanoparticles are the most widely used bactericidal agent.
Recent reports highlight their limitations, including the signi cant variations in potency between microbial species and harmful side effects due to the accumulation of high concentrations of silver ions within tissue 12,13 . Furthermore, bactericidal agents fail at protecting surfaces from contamination by viral pathogens.
Recently, the use of water repellent coatings that can be deployed on high touch surfaces, such as hospital infrastructure, has shown promising e cacy in preventing bacterial contamination 14 . Bacterial contamination of surfaces starts with the adhesion of few planktonic cells 15 . Engineering of water repellent surfaces can prevent this initial step by the formation of an air layer between the contaminated liquid and solid surface. This provides a steep mechanistic barrier to bacterial adhesion and subsequent bio lm formation [16][17][18][19] .
Inspired by these properties, recent studies (Table S1) have begun to investigate bacterial colonization of water repellent surfaces 20 . While an initial reduction of bacterial growth was observed on such surfaces, to date there is a lack of direct evidence on the mechanism preventing bacterial adhesion and growth. Furthermore, the observed bacteria inhibition is short-lived with bacteria colony densities converging toward those on non-water repellent surfaces within 4-6 h of immersion in a bacterial suspension. Water repellent coatings also often provide insu cient robustness for real-world applications, due to the intrinsic fragility of the nanoscale roughness required for attaining a Cassie-Baxter water repellent state 21 .
Here, we report a multifunctional coating able to shield surfaces from both viral and bacterial pathogens. We gain the rst direct insights on the ability of water repellent surfaces to prevent bacterial adhesion and on the eventual failure of this shielding mechanism, amid prolonged immersion in contaminated liquids or due to physical damage to the coatings. We use these learnings to engineer a coating that provides enhanced and prolonged resistance to pathogen contamination via a dual surface shielding mechanism.
Our coating features both an enhanced and robust water repellency, based on a multi-scale roughness, and a highly effective localized microbicidal action triggered by a stimuli-responsive ion release mechanism. This coating structure can provide 99.94% and 99.85% initial reduction in, gram-positive and gram-negative, bacterial and an immediate 11-fold decreased in viral pathogen contamination. Notably, it also retains its antimicrobial performance following either signi cant coating damage or extended submersion in a bacterial suspension for more than 9 days. This is more than 22 times longer than the best performing water repellent coating reported. Our coating shielding mechanism is effective in both dry and wet environments, providing means to inhibit contamination of hospital infrastructure, air and water lters and others surfaces at high risk of mediating pathogen transmission.

Results And Discussion
Engineering a dual functional coating for shielding surfaces from bacteria and viruses Our coating consists of two nanostructured components providing a pathogen repellent and microbicidal function, respectively (Fig 1, SI Note 1 and 2). Contamination by pathogen contained in liquid environments, such as respiratory droplets, is prevented by a water repellent superhydrophobic layer of uorinated silica nanoparticles, robustly anchored on an underlying interpenetrated polymer network inherently hydrophobic, contributing to the coating superhydrophobicity 23 . The spontaneous release of Zn 2+ ions from ZIF-8, due to its gradual decomposition in a broad range of pH values bears promise for its use as an antimicrobial agent 24,25 . Here, the presence of ZIF-8 is aimed at killing bacteria that may adhere to the coating due to damage to its structural integrity, and thus localized loss of superhydrophobicity, evaporation of their liquid shell environment before removal from the surface, or loss of the air layer due to prolonged immersion in a contaminated liquid (Fig 1b). For further discussion on the surface protection mechanism please see Note 5 in SI.

Protecting surfaces from viral contamination
We rst explored the performance of our water repellent surface structure in avoiding viral contamination. We used a steel rod with signi cant surface roughness ( Fig S15) and tested its ability to mediate transmission of viruses in its uncoated state, or coated with our water repellent coatings, with or without ZIF-8 nanoparticles (referred to as water repellent and dual functional coatings, respectively; SI Note 4, Fig  S15). For each test, the rod was dipped into a high titre solution of an EGFP-expressing lentiviral vector as a surrogate for infectious virus. The rod was dipped in a series of inert washing solutions (fresh Dulbecco's Modi ed Eagle's Medium (DMEM)) to determine the level of initial contamination and ease of cleaning (Fig 2a, Fig S15). Virus was quanti ed in the washing solutions by incubation on a monolayer of susceptible cells followed by uorescence microscopy to detect and count infected cells (Fig 2b).
The rst washing solution represents the capacity of the rod to resist viral contamination upon exposure to a high concentration lentiviral (i.e. human immunode ciency virus) solution (1×10 7 TU/mL). Analysis of the rst washing solution reveals a signi cant capability of the coatings to prevent viral contamination of the rod (p=0.05, Fig 2c). More than an 11-fold decrease in cell infection was observed from the bare rod (9.4% (± 0.7%)) to the ones coated with a water repellent surface (0.8% (± 0.3%)). The dual functional coating was also effective in preventing viral contamination, with a roughly 3-fold decrease in cell infection compared with bare steel. Remarkably, no virus was detected upon the rst wash on the rod coated with the water repellent coating and only a very small amount of virus was detected for the dual functional coating, that was eliminated after the 2 nd wash. The reduction in virus between the rst two steps (contaminated surface to wash 1) for the dual functional coating was ca. 11-fold, after which detection of virus became sporadic. By contrast, the bare steel carried virus through all four wash steps showing a linear decline, with infectivity reducing by around a half with each successive wash. More details of the speci c coating performance are provided in Note 4 in SI.

Bacterial Pathogen Shielding Performance and Mechanism
We used bacteria to evaluate the shielding performance against larger and heavier pathogens than virus, and to gain direct insights into the shielding kinetics of our coatings. Fig S16 shows a schematic representation of the initial interface between bacteria suspended in a liquid and the coating surface, and its time evolution. A continuous air lm, denoted as plastron layer, is initially formed at the solid-liquid interface preventing contact between bacteria and the coating surface. Over time parts of the plastron layer are removed via the formation and release of air bubbles, induced by the liquid pressure above the coating. These newly immersed coating surface becomes available for bacteria adhesion and eventually result in the formation of a bio lm. To investigate the speci c shielding kinetics of our multi-scale rough coating, a droplet of bacterial suspension, in its exponential growth phase, was placed and kept on the coating surface and on other non-coated materials (bare steel) used as a comparison (Materials and Methods in SI, Fig S 4(a)). suspension on a bare steel plate and a steel plate coated with the dual functional coating. On the bare steel surface, the bacteria immediately settle and start the surface adhesion process (Fig 3(b)). In stark contrast, on our coating the bacteria suspension keeps oating above the plastron layer (white horizontal line), without being able to reach the underlying surface 26 .
To gain further insights on the surface colonization mechanism, a droplet of bacterial suspension was kept on a bare and a coated steel surface and the cell behaviour was tracked with a confocal laser scanning microscope (CLSM) (Movie 1-2, Materials and Methods in SI, Fig S4(a)). Bacteria possess hairlike structures on their cell wall, denoted as agella, which aids them in exploring microscale topography and facilitates adhesion to a surface. On the bare steel surface, the bacteria displayed nanoscopic vibrations around an equilibrium position (Movie 1&2). This is characteristic of a surface-bound motility behaviour referred as swarming, which is usually observed upon bacterial adhesion to a surface 26 . More information regarding mobility and subsequent characterisation is described in Materials and Methods in SI.  (Fig S4(a), Movie 1-2) highlighting the dynamic response of the bacteria droplet suspensions on the bare steel and dual functional coating. Adhered bacteria were identi ed by lack of displacement and are false coloured in yellow. The average motion of the bacteria on both the surfaces were quanti ed by their average displacement between each time point (Fig 3(f, j)). Notably, bacteria cells on the bare steel surface had signi cantly (p<0.001) lower average velocity of 4.09 (± 1.42) µm/min than those on the water repellent surface (9.66 (± 1.43) µm/min).
Bio lm formation on an abiotic surface, such as the bare steel here, starts with a reversible primary adhesion stage followed by an irreversible secondary adhesion stage. The primary stage is greatly in uenced by hydrophobic interactions between the cell and the surface 27 . Our in-situ analysis (Movie 1) indicates that several bacterial cells on the bare steel surface have already entered the secondary stage, showing no translational movement, and rmly adhering to the surface. In contrast, on the water repellent surface, no bacteria cells lacking motion could be identi ed, indicating prevention of the primary adhesion stage.
To attain a quantitative assessment of the effectiveness of our surface structure in preventing bacterial adhesion, bare and coated steel substrates were immersed in a bacterial culture and stamped on agar plates (Fig S5), with or without a prior washing step in Phosphate Buffered Saline (PBS) . Fig 4(a,c) shows axio-observer microscopy of the agar plates stamped with the washed and the unwashed bare steel substrates. Notbly, both samples resulted in the formation of a square-shaped layer of green uorescing bacterial colonies, too densely packed to be counted. In stark contrast, the agar plates stamped with the un-washed water repellent coatings show only 12 (± 5) colonies (Fig 4(b)). Stamping the water repellent coatings after washing resulted in very few bacterial colonies, in the range of 2 (± 2) (Fig 4(d)). We attribute the small number of bacterial colonies, transferred from the un-washed coatings to the agar plate, to loosely adhered bacteria micro/nanodroplets. This is con rmed by the signi cant decrease in bacteria colonies upon a gentle washing step (Fig 4(d)), indicating the absence of bio lm formation (more discussion in Materials and Methods in SI).
To further quantify the e cacy in preventing bacterial adhesion, serial dilution of the residual PBS used to wash the samples after exposure to bacteria in their log growth phase was undertaken ( Fig S5). Fig   4(e) shows that the bare steel samples resulted in 2.1×10 5 (± 2.1×10 4 ) and 1.4×10 5 (± 3.1×10 4 ) CFUs/mL colonies for gram-negative and gram-positive bacteria strains, respectively. Our coating successfully prevented up to 99.85% and 99.94 % of the gram-negative and gram-positive bacteria strain growth (signi cantly different with a p-value of less than 0.01 in a one-way ANOVA test) with only 297 (± 56) and 89 (± 24) CFUs/mL, respectively. To validate the essential role of the water repellent texture, control samples consisting of the same polymer coating without the uorosilica layer were also investigated. The latter resulted in slightly higher CFUs/mL values than the bare steel surface, con rming that the polymer layer does not contribute to a decrease in bacterial adhesion.

Role of micro-scale coating defects on pathogen adhesion
To determine the role of micro-scale coating defect sites on pathogen adhesion, the water repellent uorosilica layer was removed in a quasi-elliptic area of 62.8 µm 2 , exposing the underlying steel surface. A droplet of bacterial suspension was kept over the defective region of the coating and its intact water repellent surrounding (Fig S4(a)). The interface between the microscale defect and the water repellent region was continuously in-situ imaged by CLSM. Fig 5(a) shows an orthogonal projection of the dropletsubstrate region. Over the intact water repellent surface (region A), a plastron layer with an above oating cloud of bacteria is observed. The plastron layer disappears on the defective coating area. The latter area consists of two parts (B,C), corresponding to the border with the intact region and the bulk of the defected area, respectively.
Bacteria over regions A and B exhibit a swimming behaviour whereas in region C, bacteria encounter the surface, due to the lack of the plastron layer. Movies 3 & 4 show the time-lapse view from the image plane 1 and 2 in Fig 5(a). Fig 5(b,c) shows static top view images taken after 14 min on these areas. Analysis of the velocity pro les in regions A and B (Fig 5(d)) con rm that the bacteria oat over the intact surface and over the interface with the defected area with average velocities between 10.9 (± 1.6) and 11.5 (± 2.3) µm/min, respectively. Whereas C has much smaller average values of 4.0 (± 1.6), in line with the analysis of the bare metal and water repellent surfaces above (Fig 3(i, m)). We also performed SEM analysis of such a defected area challenged with a bacterial culture for 10 min (Fig 5(e-g)). Bacteria began to colonise the defective area of the sample (Fig 5(f)), whereas there was no evidence of bacteria adhered to the neighbouring non-defective regions of the sample (Fig 5(g)). These results show that any microscale defects can act as a colonization site for bacteria. This may explain the observed few isolated colonies of bacteria observed on the agar plates stamped with the unwashed water repellent surfaces (Fig 4(b)).

Failure kinetic of the water repellent shielding mechanism
In addition to the presence of defects on the coating, the long-term stability of the plastron layer is a key factor to prevent bacteria and other pathogen adhesion to the surface. Previous studies report that the plastron layer on super-hydrophobic surfaces is lost within 1-1.5 h of continuous immersion in water (Table S1). Here, to investigate the stability of the plastron layer on our surface structure, coated steel substrates were immersed in a liquid bacterial culture at a depth of 2.5 cm, and the water-surface interface was continuously imaged by CLSM over a period of 8 h (Fig S4). Fig 6(a-h) show images with green uorescent bacteria on agar plates stamped with the water repellent steel substrates as a function of their immersion time from 1 to 8 h. Notably, for up to 4 h of continuous immersion very few bacterial colonies are able to adhere to the substrate (Fig 6(a-b)). For exposures of 4 to 6 h, a rapidly increasing number of bacteria colonies was observed (Fig 6(c-e)). While longer time results in complete colonization of the bacteria-challenged area (Fig 6(f-h)). For the full characterization of the failure kinetics of the water repellent shielding mechanism please refer to Note 5 in SI.
To gain insights on the mechanism determining the longer stability of the plastron layer on our water repellent surfaces, in situ studies were conducted by immersing our coating in a column of deionized water coloured with rhodamine B (Fig 6(j-m), Fig S4(b), Movie 5). The collapsing of the plastron layer was captured by the formation of a series of sequential contact point with the surface structure. This process rapidly accelerates after 6 h where the formation of bubbles supported by the rough surface features is observed. Subsequently, the bubbles are released resulting in the wetting of the surface as highlighted by the dye accumulation and the correlated increase in uorescence intensity (Fig 6(l, m)). This lifetime of the plastron layer is attributed to the multiscale surface roughness of our coatings (see Note 5 in SI for more characterisation and discussion)

Secondary surface shielding mechanism
Our coating provides protection against both virus and bacteria until failure of the plastron layer. However, despite providing a longer stability than previous reports on water-repellent surfaces (> 4 h of continuous submersion), we acknowledged that the eventual dissolution of the plastron layer limits the breath of applications, for instance, in marine environments, membrane systems, or bathrooms, where an antibiofouling surface that could tolerate extended submersion is required. To this end, we have exploited the antimicrobial and hydrophobic properties of ZIF-8 nanograins (Fig S11) to impart a secondary shielding mechanism to the water repellent coatings. For material property optimization and wetting performance of the dual functional coating with ZIF-8 please see Note 2 in SI.
The antimicrobial effect of the ZIF-8 became more accentuated with increasing immersion time with bacteria colonies further decreasing on the dual functional coatings after 9 days of continuous immersion. At such prolonged interaction between bacteria and coated surfaces, the impact of the ZIF-8 content became measurable, with the 15 wt% ZIF-8 containing coatings showing the smallest number of colonies of 3×10 1 (± 1.8×10 1 ) versus the 1.6×10 2 (± 6.9×10 1 ) of the 5 wt% ZIF-8 containing coatings and the 5.9×10 5 (± 2.8×10 5 ) of the bare surfaces.
We further evaluated the e cacy of this antimicrobial mechanism to prevent bacteria adhesion in case of damage to the coating and subsequent loss of its water repellent functionality. A scratch was imparted on the pure uorosilica water repellent coatings and the dual functional coatings with 15 wt% ZIF-8. A droplet containing bacteria culture in log phase was left on the defect (Fig S7), upon which the samples were washed with PBS and then incubated for 24 h. Subsequently the samples were stamped on agar plates, then incubated and analysed by microscopy (Fig 6(o-r)). The bare steel samples, used as controls, shows signi cant bacteria adhesion (Fig 6(o)). The non-scratched water repellent coatings had no bacterial adhesion (Fig 6(p)) in line with the plastron layer-based bacterial repulsion mechanism. The scratched water repellent coatings made of pure uorosilica particles resulted in bacteria adhesion in proximity of the scratched region (Fig 5(q)). Notably, the scratched dual functional coatings containing ZIF-8 nanoparticles resulted in no bacteria adhesion despite the loss of their water repellent functionality (Fig 6(r)).

Conclusions
Realizing surfaces that are inherently able to resist pathogen contamination is central to healthcare strategies aimed at preventing pandemics. We introduced a dual functional coating capable of both shielding the surface from virus and bacteria. Our dual functional coating decreased the adhesion of bacteria by 99.8%, even when continuously immersed for 9 days in highly concentrated bacteria solutions. Importantly, the water repellent mechanism was also functional against virus, decreasing uptake by 11 times. Complete removal of the virus only required a gentle, single wash, while it was not possible to completely remove the same viruses from an uncoated steel surface even after 4 washes. The ease of fabrication of our coating by a rapid spray process and its compatibility with most materials makes it suitable for preventing pathogen contamination of a broad range of high-touch surfaces in public settings.    minutes; (f) the defect is colonised by bacteria, whereas (g) the non-defective area has no bacteria attached to it.

Figure 6
Failure kinetic of water-repellent shielding mechanism and performance of the secondary shielding mechanism (a-h) Axio-observer images of the agar plate imprints corresponding to water repellent surfaces submersed in bacteria solutions for 1 hour to 8 h. Up to < 6 h immersion, a decrease in colonization with respect to the bare steel coating is observed.