Antibacterial activity of solid surfaces is critically dependent on relative humidity, inoculum volume, and organic soiling

Abstract Antimicrobial surface materials potentially prevent pathogen transfer from contaminated surfaces. Efficacy of such surfaces is assessed by standard methods using wet exposure conditions known to overestimate antimicrobial activity compared to dry exposure. Some dry test formats have been proposed but semi-dry exposure scenarios e.g. oral spray or water droplets exposed to ambient environment, are less studied. We aimed to determine the impact of environmental test conditions on antibacterial activity against the model species Escherichia coli and Staphylococcus aureus. Surfaces based on copper, silver, and quaternary ammonium with known or claimed antimicrobial properties were tested in conditions mimicking microdroplet spray or larger water droplets exposed to variable relative air humidity in the presence or absence of organic soiling. All the environmental parameters critically affected antibacterial activity of the tested surfaces from no effect in high-organic dry conditions to higher effect in low-organic humid conditions but not reaching the effect size demonstrated in the ISO 22169 wet format. Copper was the most efficient antibacterial surface followed by silver and quaternary ammonium based coating. Antimicrobial testing of surfaces using small droplet contamination in application-relevant conditions could therefore be considered as one of the worst-case exposure scenarios relevant to dry use surfaces.


Introduction
Surface transfer of microbes is one of the main routes of transmission of infectious diseases in critical applications including health-car e envir onment (Kr amer and Assadian 2014 ).Ther efor e, surfaces decr easing micr obial r esidence time and viability and thus, the likelihood of transfer of potential pathogens, may have a significant contribution to the decrease of infectious diseases and impr ov ement of public health (Muller et al. 2016 ).Ho w e v er, such efficacy may only be obtained if the surface elicits sufficient antimicrobial activity.According to legislative acts and their guidance documents in the European Union (ECHA 2018(ECHA , 2022 ) ) and in the United States (EPA 2023 ), a proof of at least of 3 log 10 (99.9%) reduction in microbial viability in up to 1-2 hours should be provided prior to making any commercial antimicrobial claims about surfaces with biocidal properties .US EPA guidance , originally developed for the assessment of copper surfaces, suggests antimicrobial activity assessment via application of a small volume of bacterial inoculum onto the test surface in the presence of organic soiling and subsequent dry exposure at ambient conditions (EPA 2022 ).Guidance documentation of the European Biocidal Product Regulation (ECHA 2018 ) suggests se v er al testing formats at concept le v el but r efers to onl y one established method, to the ISO 22196 (ISO 2011 ).The latter is an industrial standard that uses a low-organic liquid inoculum exposed to test surfaces as a thin lay er in w arm humid conditions resulting in high surface area to inoculum volume r atio.Rele v ance of the ISO 22196 in efficacy e v aluation of surfaces de v eloped for ambient/dry use conditions has been r epeatedl y questioned (Ojeil et al. 2013, Redfern et al. 2018, Dunne et al. 2020, Cunliffe et al. 2021 ) and a more relevant dry method is curr entl y under de v elopment by the ISO (ISO/DIS 7581) (ISO 2023 ).Experimental a ppr oac hes to sim ulated use methods can also be found in the liter atur e (Ojeil et al. 2013, Knobloch et al. 2017, McDonald et al. 2020 ) but those often r el y on r estrictiv e instrumentation or methodology such as aerosolization of the inoculum.To date, there is no consensus on the combined effect of drying and exposure medium contents to antimicrobial activity of solid surfaces in conditions resembling indoor ambient environment in different microbial contamination scenarios.
In terms of efficacy assessment, se v er al test formats have been described (Sjollema et al. 2018, Cunliffe et al. 2021 ).Ho w e v er, antimicrobial surfaces are more challenging than liquid formulations and it has been clearl y demonstr ated that a ppl ying differ ent test methods can result in substantial differences in the antimicrobial activity of the surfaces .T here are indications that methods involving dry exposure show lower antimicrobial activity than methods where bacterial exposure to surfaces is carried out in wet conditions.Methods also vary in the presence and amount of organic soiling as well as changes in incubation temperature during exposure.It has been demonstrated that a dry droplet and a surface transfer method results in markedly lo w er antibacterial activity than ISO 22196 standard test which requires exposure of bacteria to surfaces in a thin layer of liquid under a plastic film to avoid drying (Campos et al. 2016, van de Lagemaat et al. 2017 ).Similarly, it has been shown that antimicrobial effect of copper alloy surfaces was substantially lower in touch transfer assay than in ISO 22196 conditions (Knobloch et al. 2017 ).One possible explanation for the lo w er a ppar ent efficacy of antimicr obial surfaces in dry conditions may be the potential selection of desiccationresistant bacteria in dry environment and co-resistance of such bacteria to w ar ds antimicr obial compounds (Knobloc h et al. 2017 ).Also, high loss of viable bacteria in negative control has been proposed (Campos et al. 2016 ), although quality control of antimicrobial assay should eliminate this possibility.
To date there are no systematic studies on how the different test conditions affect antimicrobial activity of solid surfaces and most of the studies have compared one or two test formats with the current standard ISO 22196 or its relevant in-house modifications.One of the widest selections of varying test protocols has been used by van de Lagemaat et al. who compared the antibacterial efficacy of a quaternary ammonium compound (QAC) based polymer surface using five different test methods and demonstrated method-dependent differences in antibacterial activity (v an de La gemaat et al. 2017 ).Ho w e v er, the details of the used test methods varied in many aspects , e .g. in bacterial inoculum density, nutrient content and e v en exposur e time .T her efor e, the contribution of one specific test condition on antimicrobial efficacy of the tested QAC surface was very hard to r e v eal.Similarl y, Knobloc h et al .used in their study a specified touch transfer method where a bare finger or a glo ve-co vered fingertip that was pr e viousl y slightl y moistened and contained certain le v el of organic soiling was used to transfer bacteria to antimicrobial surface and sho w ed that in their assay, a series of antimicrobial surfaces were significantly less active than when tested with standar d w et method ISO 22196 (Knobloch et al. 2017 ).Although moisture and soiling of the environment appear as the most crucial factors in antimicrobial efficacy assessment, the effect of those factors was not studied in detail by Knobloch et al. and their contribution to the changes in antibacterial efficacy cannot be deduced from this study.Wiegand et al. studied a wider selection of environmental criteria affecting antibacterial efficacy but doing so only in wet test format analogous to ISO 22196 (Wiegand et al. 2018 ).This study sho w ed that the a ppar ent efficacy of polyamide surfaces with zinc ad diti ves was significantly affected b y the sour ce of bacterial inoculum, inoculum density, and expectedl y, or ganic content in the exposure medium.
Due to the lack of current understanding on the role of crucial variables of test environment during dry or semi-dry antimicrobial testing of surfaces, we carried out a systematic series of antibacterial experiments in a predefined range of environmental conditions to understand, to what extent antibacterial activity of solid surfaces is determined by drying and organic soiling during exposure.We selected five solid non-porous surfaces, which had pr e viousl y been included in antimicrobial studies or presented commercial antimicrobial claims.Antibacterial activity in varying test conditions was studied against two model bacteria, Gramnegativ e Esc heric hia coli and Gr am-positiv e Staph ylococcus aureus .The obtained data were expected to provide crucial information on the role of exposure conditions on antimicrobial efficacy assessment and provide insight into the possible discrepancy be-tween claimed and a pplication-r ele v ant efficacy of antimicrobial surfaces.

Surfaces
Five test surfaces with previously shown or claimed antimicrobial effect and two control surfaces were selected: metallic copper coupons, CuC (99% Cu, Metroprint OY, Estonia); copper-based Iniesta Covidsafe adhesive tape applied to stainless steel coupons, CuT (Clean Touch Medical, Finland); metallic silver coupons, AgC (99.95% Ag, Sur epur e Chemetals, USA); silv er-containing acrylic aqueous emulsion "TOUCH Antimicrobial coating" (silver paint) applied to stainless steel coupons; AgP (0.3-0.5 weight% silverbased or ganomodified bentonite, Br omoco International Ltd., UK); and a quaternary ammonium compound based Si-Quat coating applied to stainless steel coupons, SQ (activ e ingr edient dimethyloctadecyl (3-(trimethoxysil yl)pr opyl) ammonium c hloride (CAS 27668-52-6), Affix Labs , Finland).T hese test surfaces were selected to r epr esent materials with differ ent modes of antimicr obial action and potentially different efficiency in different humidity conditions.Two control surfaces were used: austenitic stainless steel, SS (AISI 304 SS, major elements in addition to Fe were 18% of Cr, 8% of Ni, 1.4% of Mn; 2B finish; Aperam-Stainless France) that has been fr equentl y suggested as a control surface in studies involving metal alloys as well as standard biocide testing methods , and borosilicate glass , NC (Corning Inc., USA) used as negativ e contr ol or an inert solid surface.Additionall y, pol ypr opylene plastic, PP (Etra OY, Finland) was used as an inert cover layer on glass control in modified ISO 22196 test.
Metallic copper , silver , and steel were obtained as 1-2 mm thick sheets, that were laser-cut to 20 mm diameter round coupons.Prior to use, the metal coupons were vigorously shaken in acetone follo w ed b y a w ash in deionized w ater and another shaking in 70% ethanol.Coupons were then air-dried in aseptic conditions and stored in sterile Petri dishes prior to use.Silver coupons were re-used during the experiment and therefore, those surfaces were always cleaned using the following pr ocedur e: a thr ee-step shaking (2 min), sonication (15 min) and shaking (2 min) pr ocedur e in 5% citric acid, 100% acetone, and 70% ethanol with 0.3 mm glass beads.Between e v ery wash pr ocedur e, the coupons wer e rinsed with deionized water and after the last step, were sterilized at least for 15 min by UVC in a biosafety cabinet.A slight increase in antibacterial activity was detected after 5 cleaning cycles ( Fig. S1 ).Based on this data, all silver coupons were subjected to 9 successive cleaning cycles before use.
CovidSafe surfaces were prepared by adhesion of ∅ 20 mm pieces of CovidSafe copper tape onto ∅ 20 mm steel coupons.Spincoating was chosen for application of TOUCH Antimicrobial silver paint and Si-Quat coating to ac hie v e r epr oducible uniform thin layer of coating on steel.For that, 100 μL of TOUCH Antimicrobial silver paint or Si-Quat coating was placed to the center of a steel coupon while spinning at 3000 rpm speed using an in-house built spin coater.Prior to use, these surfaces as well as glass and PP control surfaces were rinsed with 70% ethanol, air dried and treated for at least 15 min with UVC light in a biosafety cabinet.

Physico-c hemical c har acteriza tion of surfaces
To c har acterize the hydr ophobicity/hydr ophilicity contact angles of the surfaces wer e measur ed.The sessile dr op tec hnique (v an Oss 1993 ) and moving platform based on the Thorlab DT12 dovetail translation stage were used.Onto a dry surface, 2 μL drop of deionized water was pipetted in ambient conditions and after 10 seconds, w ater droplet w as photogr a phed using Canon EOS 650d camera equipped with an MP-E 65 mm f/2.8 1-5x macro focus lens.Using image analysis software (ImageJ 1.8.0 172 for Windows with a plugin for contact angle measur ement (Sc hneider et al. 2012 )), the contact angle formed by the liquid drop on the studied surface was calculated.Four water droplets per sample, each pipetted at a different area of the surface, were used.Photos of re presentati ve water droplets for contact angle measurement are presented on Fig. S2 .
X-r ay photoelectr on spectr oscopy (XPS) as an extr emel y surface-sensitiv e tec hnique was used to anal yze the elemental composition of the topmost layer of the surfaces.All XPS measurements were conducted using Scienta SES-100 electron energy analyzer and Thermo XR3E2 non-monochromatized twin-anode Xray source (Al K α 1486.6 eV, Mg K α 1253.6 eV) in ultrahigh vacuum (UHV) conditions at 10 −9 mbar.TOUCH Antimicrobial silver paint surface was pr epar ed for XPS by baking in the oven for 6 h at 150 • C. Al K α anode with po w er 100 W w as used for glass and TOUCH Antimicrobial surfaces, for other surfaces, po w er 400 W of Al K α anode was applied.For all the surfaces except Covidsafe copper tape, Ar ion sputtering (5 keV, current 10 mA during 1 h) under UHV conditions at 10 −6 mbar was performed to completel y r emov e the surface's topmost lay er.CasaXPS softw are (Fairley, N . CasaXPS;V ersion 2.3.23;Casa Software Ltd.: Teignmouth, UK, 2000) was used for XPS data processing.Monte Carlo method (Fairley and Carrick 2005 ) was used to calculate variation of XPS data.
Elemental analysis by ICP-MS was carried out with TOUCH Antimicr obial silv er paint and CovidSafe copper tape.For that, a kno wn w eight of CovidSafe copper tape and a kno wn v olume of liquid TOUCH Antimicrobial silver paint formulation was dissolved in 3:1 mixture of HNO 3 : HCl follo wing b y treatment in the Berghof Speedwave Xpert micro w a ve o ven.T he concentration of Ag in TOUCH Antimicrobial silver paint and Cu in CovidSafe copper tape was measured using Agilent 7700 ICP-MS from samples after their dilution up to 0.5% nitric acid.
Scanning electron microscope (SEM) (Tescan VEGA-II) with 5 kV and 20 kV electron acceleration voltages and secondary electron detector was used to visualize the top view of all the samples .Cross-sections of Co vidsafe copper tape , T OUCH Antimicr obial silv er paint and Si-Quat coating wer e pr epar ed by cutting the coupons with r espectiv e surface coatings and removing the overhangs by a scalpel.These samples were then placed dir ectl y on the SEM stub, a narrow strip of carbon tape was placed on top of it to impr ov e the conductivity between the surface and the SEM stub and the samples were imaged using FEI Nova NanoSEM 450 with 3 kV electron acceleration voltage.In all cases, non-conducting surfaces were cov er ed with 10 nm layer of gold.

Testing of antibacterial effect of the surfaces
Antibacterial effect of surfaces was analyzed after inoculation of surfaces with large droplets (50 μL) or microdroplets (5 × 2 μL) and exposing those droplets to different environmental variables (Table 1 ).For r efer ence, surfaces wer e also anal yzed using a modified ISO 22196 protocol.
Bacterial suspensions of Esc heric hia coli DSM 1576 (ATCC 8739) and Staphylococcus aureus DSM 346 (ATCC 6538P), both acquired fr om DSMZ, wer e pr epar ed fr om fr esh ov ernight cultur es on LB solid medium (5 g/L yeast extract, 10 g/L tryptone, 5 g/L NaCl, 15 g/L agar).Bacterial biomass was suspended in sterile deion-ized water using a sterile inoculation loop, thor oughl y mixed by pipetting and vortexing.The suspension cell counts were photometrically adjusted to target suspension densities of 1.5 × 10 8 -5 × 10 8 or 7.5 × 10 8 -2.5 × 10 9 CFU/mL for large droplets and micr odr oplets, r espectiv el y.The pr epar ed bacterial suspension was mixed 1:1 with 2-fold concentrated organic soiling solution.The a pplied or ganic soiling was either the or ganic pr oportion of the EPA soil load ( SL ; final concentration: bovine serum albumin 2.5 g/L, yeast extract 3.5 g/L, mucin 0.8 g/L (EPA 2022 )) or the 500fold diluted nutrient broth of ISO 22196 ( NB ; final concentration of ingredients: 0.006 g/L meat extract, 0.02 g/L peptone, 0.01 g/L NaCl (ISO 2011 ) resulting in 3.75 × 10 6 -1.25 × 10 7 CFU/surface independent of inoculum droplet size .T he cell count per surface was intentionally selected higher than ISO 22196 r equir ement to allow assessment of at least 3 log 10 reduction in viability in conditions that also affect viability on control surfaces and to better resemble conditions used to e v aluate efficacy of surface disinfectants.Inoculum density from the European standard 13697 (CEN 2015 ) "dirty" conditions was used as guidance for lar ge dr oplets and the same cell count per surface was also used for micr odr oplet inoculation with higher inoculum density.(EPA 2022 ) Surfaces with large 50 μl droplets or 5 × 2 μl microdroplets on surfaces were incubated in open Petri dishes ( Fig. S2 ) in a climate chamber (Climacell EVO, Memmert, USA).Incubation was carried out at 22 • C and variable RH: 90% RH (high humidity), 50% RH (moderate humidity) or 20% RH (low humidity).After 0.5, 1, 2, and 6 h of incubation the surfaces submerged to 10 mL of toxicity neutralizing medium (SCDLP: 17 g/L casein peptone, 3 g/L soybean peptone, 5 g/L NaCl, 2.5 g/L Na 2 HPO 4 , 2.5 g/L glucose, 1.0 g/L lecithin, 7.0 g/L Tween80 (ISO 2011 )) in 50 mL conical centrifuge tubes .T he tubes w ere v ortexed for 30 seconds to detach bacteria, serially diluted in phosphate-buffered saline (10-fold stock diluted to: 8 g/L NaCl, 0.2 g/L KCl, 1.44 g/L Na 2 HPO 4 , 0.2 g/L KH 2 PO 4 ; pH 7.1).Dilutions wer e dr op-plated, incubated at 37 • C for optimal gr owth and plate counts r egister ed after 16-18 h for E. coli or 24 h for S. aureus.At least 3 biological and 2 technical repeats were analyzed.
For r efer ence and comparison, all the surfaces were tested using a modification of ISO 22196 format.For that, inoculums of E. coli and S. aureus wer e photometricall y adjusted to the tar get 3.14 × 10 6 CFU/mL in 500-fold diluted NB broth and 15 μL of this inoculum (ca 1.5 × 10 4 CFU/cm 2 ; 4.7 × 10 4 CFU/surface) was pipetted onto each surface .T he surfaces were then covered by 25 × 25 mm coverslips with an overhang but avoiding inoculum leakage and incubated at room temperature (22 • C) in a > 90% RH environment for 0.5, 1, 2, and 6 h.After exposure, the surfaces were placed into 10 mL of SCDLP neutr alizing medium, cells wer e r ecov er ed and plated for CFU counts as described abo ve .T hese experiments were performed in three biological replicates.

Sta tistical anal ysis
Statistical analysis of the data was performed with Gr a phP ad Prism 9.5.0 (Gr a phP ad Softwar e, San Diego, USA).Raw data used can be found in the Supplementary Raw Data file.Data from the 2 h time point was selected for a mor e thor ough statistical analysis as longest r ele v ant contact time in the context of current legislativ e fr ame work in the US and EU.Log 10 -tr ansformed data was used for the analysis and values below limit of quantification (LOQ) wer e br ought to LOQ prior to anal ysis.LOQ was consider ed as at least 3 colonies counted from of 20 or 500 μL undiluted surface wash-off plated from droplet inoculation or ISO format, respectiv el y, r esulting in LOQ of 3.18 or 1.48 log 10 CFU/surface.One-

Characteristics of the surfaces
The physical a ppear ance of the surfaces studied is shown in Figs 1 and 2 .Metallic copper and silver were relatively smooth at their pristine state.Differ entl y fr om copper and silv er, other surfaces demonstrated certain surface structures already in their pristine form.CovidSafe copper tape exhibited a clear surface structure (Fig. 1 , CuT) and from cross-section (Fig. 2 , CuT) its thickness was estimated to be 5-6 μm.Both TOUCH Antimicrobial silver paint and Si-Quat coating were coated onto stainless steel coupons .T he steel itself exhibited clear surface structures (Fig. 1 , SS) and as those structur es wer e well visible in case of Si-Quat coating (Fig. 1 , SQ), we concluded that Si-Quat coating was r elativ el y thin allowing secondary electrons from the stainless steel pass through the coating la yer.T his was confirmed also fr om the cr oss-section of Si-Quat coated samples, where the thickness of the coating was calculated to be 3-5 μm.Unlike Si-Quat, no steel surface structur es wer e visible with SEM in case of TOUCH Antimicrobial silver paint (Fig. 1 , AgP).According to cross-section (Fig. 2 , AgP), the thickness of TOUCH Antimicrobial silver paint on steel was reaching 10 μm and was ther efor e able to mask the surface structure of steel.Glass surface a ppear ed completel y smooth as expected (Fig. 1 , NC).According to XPS, metallic copper and silver presented the r espectiv e elements Cu and Ag on their very surface (Table 2 ) as expected.Ho w e v er, since XPS is an extr emel y surface sensitive method (surface depth up to 10 nm) significant amount of O and C wer e pr esent in XPS spectr a as surface contaminants.Such a situation is typical for XPS measur ements.To r e v eal the composition of copper and silver inside samples, Ar + ion sputtering was used to r emov e uppermost sample layer prior to XPS anal ysis, whic h r e v ealed that metallic copper and silv er wer e presenting 100% Cu and 93% Ag on their surfaces r espectiv el y (Table 2 ).
On stainless steel, almost all expected metallic components (F e , Cr, and Mn) were identified.Nickel was not detected with XPS but given that this is a surface sensitive method, we suggest that it (8% Ni content in AISI 304 stainless steel) may be not present on the very first surface layers of the 2B finish.
Although Ag was expected to be present on TOUCH Antimicrobial silver paint surface, XPS was not able to r e v eal an y elemental Ag on the surface layer of that sample (Table 2 ).To eliminate the possibility that Ag was masked by surface contaminants such as C or O, TOUCH Antimicrobial surface was cleaned using Ar + ion sputtering.Inter estingl y, no Ag was detected also on ion-sputtered samples (Table 2 ) suggesting that the concentration of Ag was below the limit of determination of the method ( < 1% Ag).To pr ov e the presence of Ag in the sample, ICP-MS analysis of the original liquid TOUCH Antimicrobial coating formulation was performed and 20 mg of Ag/kg of the formulation was identified (Table 2 ).
On Si-Quat surface, Si, Cl and N were detected, which corresponds to the c hemical form ula of the active compound in Si-Quat coating (C 26 H 58 NO 3 ClSi) (Table 2 ).Ar + ion sputtering of Si-Quat surface r emov ed most of the Si-Quat coating as in ion-sputtered samples no Si or Cl was detected and instead, components of the base material, stainless steel (F e , Cr and Mn) were revealed (Table 2 ).
Due to the outgassing from the adhesive, UHV conditions in the anal ysis c hamber wer e not ac hie v able and CovidSafe copper tape could not be safel y anal yzed with XPS without the risk of damage to the instrument.To pr ov e the presence of Cu on that surface, ICP-MS analysis was performed on the bulk CovidSafe material, whic h pr ov ed the pr esence of Cu in that material (Table 2 ).
Most of the surfaces were hydrophilic in nature with just two surfaces-CovidSafe and steel having slightly hydrophobic properties with contact angles > 90 • (Fig. 3 ).Similar behavior of bacterial inoculum droplets was observed on all surfaces ( Fig. S2 ) and thus , in antibacterial tests , differ ences in hydr ophilicity of surfaces was not considered important.Howe v er, une v en drying time of 2 μL micr odr oplets on the same surface was often observed, possibly explained not solely or mainly by the physical characteristics of the surfaces but also ventilation in the climate chamber, variation in pipetting or combination of the abo ve .

Effect of test format on antibacterial activity
Antibacterial activity of the five potentially antimicrobial surfaces compared to 2 control surfaces was evaluated in two test formats: an envir onmentall y exposed lar ge dr oplet or micr odr oplet inoculation in various semi-dry conditions and a modified ISO 22196 (ISO 2011 ) wet test as an industrial r efer ence.Exposur e in semi-dry conditions was selected to r esemble m ucin-containing or al/nasal spr ay and spr ay-contamination of surfaces or lar ger droplet contamination with low organic content e.g.near faucets or cooking surfaces, on surfaces in public space , non-in v asiv e surfaces in health-care settings or in the food industry.As a widely used industrial r efer ence, ISO 22196 format employing lar ge sur-  face area to inoculum volume ratio was used in otherwise similar conditions to test the maximum expected efficacy of the surfaces.500-fold diluted nutrient broth (0.026 g/L organic content) with r elativ el y low metal ion complexing pr operties was c hosen as a low-or ganic exposur e medium also used in ISO 22196.The or ganic pr oportion of m ucin-containing soil load described in US EPA test guidance, originally developed for the efficacy assessment of copper surfaces (EPA 2022 ) was chosen as a high-organic exposure medium (6.8 g/L organic content) that has metal ion complexing capacity and supports metabolism.PBS was omitted from the EPA SL formulation to avoid introducing additional osmotic variable into the comparison of environmental conditions.In addition to organic soiling, relative air humidity (RH) was varied between exposures, to represent dry (20%), normal (50%), or humid (90%) indoor conditions .T he test surfaces were generally chosen due to their proven antibacterial efficacy.Copper and silv er wer e included as materials with historicall y w ell-kno wn antimicr obial pr operties as well as continuing r esearc h inter est and commercial use (Rosenberg et al. 2019 ).Quaternary ammoniumbased Si-Quat coating was selected as a non-metallic surface with a different mode of action.Antibacterial activity of those surfaces was e v aluated in comparison with borosilicate glass as inert surface and stainless steel as a widely used non-antimicrobial metal surface.
Figures 4 and 5 present the log 10 reduction in viability of E. coli and S. aureus , r espectiv el y, on studied surfaces after 2 h exposure in all the test conditions used (complete data from 0.5 to 6 h time points: Figures S3 and S4 and Tables S1 and S2 ).Copperbased surfaces pr ov ed to be the most efficient antibacterial surfaces across the conditions tested, follo w ed b y silver and Si-Quat surfaces (Figs 4 and 5 ) with the latter showing potential specifically to w ar ds S. aureus (Fig. 5 ).The CovidSafe copper tape demonstrated lo w er effect than pur e copper and the TOUCH Antimicr obial silver paint was the least effective with no effect in any of the test conditions in 2 h (Figs 4 and 5 ) and only a minor effect against E. coli in a couple of conditions by the longest 6 h time point.
Large contact area between the bacterial inoculum and the test surface in humid conditions in the modified ISO 22196 method (Figs 4 A and 5 A) gener all y ov er estimated antibacterial activity of metal-based surfaces compared to droplet and microdroplet inoculation in the same low-organic exposure medium and at ambient temper atur e (panels B and D on Figs 4 and 5 ).The precise log 10 r eduction v alues ar e c hallenging to compar e due to substan- bars .An a v er a ge of 3-6 biological r eplicates with standard de viation is shown.Statisticall y significant differ ence ( P < 0.05) fr om the NC contr ol surface is marked with * * * * ( P ≤ 0.0001), * * * ( P ≤0 001), * * ( P ≤ 0.01), * ( P < 0.05).The target of 3 log 10 reduction is displayed as a grey dotted line.Antibacterial activity after 0.5, 1, 2, and 6 h exposure is presented on Figs S3 and S4 , log 10 -transformed viable counts in Tables S1 and S2 and raw data in Supplementary Raw Data file.
tially lo w er cell count per surface in the ISO format as that can affect the antibacterial effect size .T he ISO format also expectedl y guar anteed sufficient bacterial viability on control surfaces ( Tables S1 and S2 ; Fig. 7 ) which simplifies analysis in a standard test format but does not reflect application-relevant conditions where a substantial amount of bacteria might be killed or inactivated by environmental variables e.g.desiccation, light exposur e, or ganic and non-organic surface contamination, too m uc h or too little oxygen etc.It could be that combinations of envir onmental v ariables might decr ease bacterial viability synergistically and complicate efficacy assessment of antimicrobial surfaces .For example , we ha v e pr e viousl y r eported that low-intensity UVA phototoxicity to w ar ds both E. coli and S. aureus is higher in lo w er air humidity on inert control surfaces (Kisand et al. 2022 ).Although we did not specifically design the current study to quantify synergism between the selected environmental parameters, there is clear evidence of synergistic effects represented by substantial contribution of statistically significant 2-way and 3way interactions of environmental parameters to total variabil-ity of viable counts based on 3-way ANOVA of 2 h exposure data (Fig. 6 ).
Ther efor e, the ISO format r epr esents the "best-case scenario" with high bacterial viability on control surface and high antimicrobial activity on surfaces of interest.With this study we demonstrate that in more plausible "worst-case scenarios" viability of E. coli is decreased on control surfaces and antibacterial activity of the potential antimicrobial surfaces is gener all y lo w er than in the ISO format.Ho w e v er, as can be seen for E. coli in 50% and 90% RH and micr odr oplets (Fig. 4 B) or lar ge dr oplet in 20% RH (Fig. 4 D) in low-organic medium there can be exceptions, where antibacterial activity of copper after 2 h exposure does not significantly differ between the ISO format ( −3.5 log 10 ) and droplet inoculations ( −3.1, 3.1 and 3.45 log 10 , r espectiv el y; P > 0.05).
Although se v er al surfaces elicited > 3 log 10 r eduction in bacterial viability in different test conditions after 6 h exposure, the regulatory goal of at least 3 log 10 reduction in up to (1-)2 h was in most cases not met, except copper-based surfaces in some conditions specified below.Copper and silver surfaces were generally The maximum possible log 10 reduction detectable depended on bacterial viability on glass control surface in each test condition.Maximum possible log 10 reduction of S. aureus was not achieved by any of the surfaces in any conditions after 2 h exposure.An av er a ge of 3-6 biological replicates with standard deviation is shown.Statistically significant difference ( P < 0.05) from the NC control surface is marked with * * * * ( P ≤ 0.0001), * * * ( P ≤0 001), * * ( P ≤ 0.01), * ( P < 0.05).On panel D, tw o-w ay ANOVA + post-hoc r esults ar e shown for the whole dataset (black * - * * * ) as well as with the most variable SQ gr oup r emov ed (r ed * * * * ).The tar get of 3 log 10 r eduction is displayed as a gr e y dotted line.Antibacterial acti vity after 0.5, 1, 2, and 6 h exposure is presented on Figs S3 and S4 , log 10 -transformed viable counts in Tables S1 and S2 and raw data in Supplementary Raw Data file.
mor e effectiv e a gainst E. coli than S. aureus (Figs 4 and 5 ) while Si-Quat surfaces tended to pr efer entiall y act better on S. aureus (Fig. 5 B and D).Ho w e v er, despite substantial mean log 10 r eduction Si-Quat surfaces did not demonstrate a reliable effect due to extr emel y high v ariability among biological r eplicates.It m ust be noted that while the CovidSafe copper tape elicited > 3 log 10 reduction in viability to w ar ds both species after 2 h in the ISO format (Figs 4 A and 5 A), it had consider abl y lo w er to no activity in other formats , e .g. < 1 log 10 reduction of S. aureus regardless of exposed droplet test conditions .T he T OUCH Antimicrobial silver paint did not show any antibacterial activity in any condition tow ar ds either of the species after 2 h exposure including the ISO 22196 format.This lack of effect could be explained by the extr emel y low silver content in the TOUCH Antimicrobial initial formulation and no Ag detected by XPS on the coated surfaces (Table 2 ).After 2 h exposur e, onl y copper-based surfaces r eac hed the r equir ed 3 log 10 r eduction in viability.Both metallic copper and CovidSafe copper tape demonstrated > 3 log 10 reduction of viable counts of both bacterial species in the ISO format ( P < 0.0001 in all cases).Ho w e v er, onl y metallic copper caused > 3 log 10 reduction in exposed droplet format only to w ar ds E. coli and only in the following conditions: low-organic microdroplets in 50% and 90% RH and low-or ganic lar ge dr oplets in 20% RH ( P < 0.0001 in all cases).

Effect of en vironmental v ariables on antibacterial activity
All the antimicrobial surfaces were generally less effective in dry conditions (20% RH) and/or in high-organic medium.No biologically ( > 1 log 10 reduction in viable counts compared to contr ols) and/or statisticall y significant antibacterial activity of any of the surfaces, in any of the conditions to w ar ds either of the bacterial species was observed after 2 h exposure in case of micr odr oplet inoculation and low r elativ e humidity (panels B-C on Figs 4 and 5 ).There were also no surfaces that could biologicall y and statisticall y significantl y r educe bacterial viability after 2 h in large droplets of exposure medium with high organic content (Figs 4 E and 5 E).Both findings highlight an urgent need Figure 6.Contribution of drying of the inoculum, surface type ( vs glass contr ol, NC), exposur e medium (high vs low organic content) and their interactions (indicated by variable x variable interaction in the legend) to the total variation in viable counts based on 3-way ANOVA analysis of 2 h exposure data of E. coli (left) and S. aureus (right).The parameters of relative humidity (RH) and inoculum droplet size proved to be interdependent in their effect on viability pr obabl y by affecting drying time and could thereby impair main effect and interaction interpretation in multifactorial ANOVA.Instead, a combined variable, drying, was used.As a categorical variable drying comprised of all 6 possible combinations of RH (20%, 50% and 90%) and inoculum droplet sizes (5 ×2 μL and 50 μL).Only the most effective surface type of each active agent and stainless steel as a metallic control surface are shown compared to glass (NC): CuC, metallic copper; AgC, metallic silver coupon; SQ, Si-Quat coating; SS, stainless steel.All values above 6% are statistically significant.Numerical data is presented in Table S3 .(E-H) in variable testing conditions.CuC, metallic copper; AgC, metallic silver; SS, stainless steel; NC borosilicate glass.Blue hues denote lo w and y ello w hues high organic content in exposure media.Lighter color indicates lo w er air humidity and dotted lines microdroplet inoculation.An average of 3-6 biological replicates with standard deviation is shown.Values at or above the detection limit of at least three colonies counted in all undiluted plated drops (dotted line at y = 3.18) were used for the statistical analysis of CFU counts.Lo w er v alues ar e shown to demonstrate that growth of viable bacteria was detected in at least some repetitions.Tabular presentation of mean values and standard deviations is presented in Tables S1 and S2 and raw data in Supplementary Raw Data file.
to use a pplication-r ele v ant methodology to adequatel y assess efficacy of antimicrobial surfaces used in environments exposed to air.
T he en vir onmental v ariables affected viability of both E. coli and S. aureus on control surfaces (Fig. 7 C, D, G and H).As a general rule, on surfaces with lo w er to no known antibacterial activity, viability of E. coli was more affected by drying analyzed as a combined effect of RH and inoculum droplet size (Fig. 6 , left) while S. aureus was more affected by the selection of exposure medium (Fig. 6 , right).Our results indicated that the effect of environmental variables depended also on the surface type (Fig. 6 ; based on 3-way ANOVA of 2 h exposure data).While metallic copper surfaces demonstrated the highest antibacterial activity, we can see that its effect was most dependent on exposure medium, surface type and their interaction jointly explaining 77% and 85% of total variation in viable counts for E. coli and S. aureus on copper, r espectiv el y (Fig. 6 ).

Effect of drying on antibacterial activity
Drying had only a minor contribution to ov er all v ariability in viable counts on copper (Fig. 6 ), thus not dr amaticall y affecting bacterial viability.Ho w e v er, drying, exposure medium and their interaction had a more even contribution to variability of viable counts on silver surface (Fig. 6 ).Based on literature, silver surfaces are expected to have a larger effect in wet conditions and copper also demonstrating substantial effect in dry conditions (Michels et al. 2009, Knobloch et al. 2017, Villapún et al. 2018 ).This was with some r eserv ation confirmed for E. coli , but not for S. aureus, partly because silver had overall very low antibacterial activity towards S. aureus .Inter estingl y, while metal-based surfaces seemed more efficient to w ar ds E. coli , Si-Quat coating pr efer entiall y worked against S. aureus in low-organic medium and more importantly, its activity was not significantly affected by RH as none of the drying-r elated causes statisticall y significantl y contributed to viable count variation (Fig. 6 , right).Viable counts retrieved from Si-Quat coating in exposed droplet format were also characterized by extr emel y high v ariability compar ed to other surfaces, possibly indicating the unstable nature of microbe-surface interaction parameters beyond our selected criteria.The Si-Quat coating is described as a contact-killing and not release-based surface by the producer.Based on our results, it seems that Si-Quat surfaces could be most effective in low-organic large droplets independent of RH conditions (Fig. 5 D).This in combination with our finding on higher antibacterial activity of Si-Quat coating to w ar ds S. aureus than to w ar ds E. coli suggests the possible crucial role of water envir onment and potentiall y electr ostatic inter actions with nonmotile S. aureus as opposed to motile E. coli .
Inter estingl y, in the ISO 22196 format, the antimicrobial effect, if pr esent, gener all y r eac hed its maxim um by 2 h and plateaued ther eafter wher eas substantial decr ease in bacterial viability between 2 and 6 h could be observed on several surfaces in case of envir onmentall y exposed droplets (Fig. 7 A ,B,E and F;Figs S3 and S4 ).In case of exposed dr oplets, suc h plateauing of the antibacterial effect was seen after complete drying of inoculum droplets (Table 3 ; Figs S3 and S4 ).There seems to be a balance between the effect size of antibacterial action and the duration of the physical process of drying which can cause higher antibacterial activity in lo w er RH if the droplets do not dry so fast as microdroplets in 20% RH, possibly due to longer wet contact time, longer active agent release and e v entuall y drying i.e. decreasing droplet volume while incr easing activ e a gent concentr ation (e .g .Fig. 7 A and B; blue lines , both drop sizes).Plateauing of the antibacterial effect after drying Table 3. Drying time of the inoculum droplets.Time point (h) by which the inoculum droplets had visibly completely dried on the studied surface, irr espectiv e of or ganic soiling or bacterial species used.In the case of 5 ×2 μl micr odr oplet inoculation wher e une v en drying time of individual droplets on the same surface, the timepoint after which all droplets had visibly dried was r egister ed.Repr esentativ e ima ges of the dr oplets ar e pr esented in Fig. S5 .

20% RH 50% RH 90% RH
Lar ge dr oplet (50 μL) 6 h 6 h > 6 h Micr odr oplets (5 ×2 μL) 0.5 h 0.5-1 h 6 h could indicate that at least some amount of free water is needed to elicit the antibacterial activity.In this light the normal indoor RH (50%) provides conditions that, at least in low-organic environment, could benefit most of the use antimicrobial surfaces.Howe v er, the pr ocess of drying is not as simple as just the e v a por ation of water and should be studied further in the context of antimicrobial surfaces.Not only RH but also the physical c har acteristics of surfaces , microbes , and liquids can impact the kinetics of drying and may potentially alter the antibacterial properties of surfaces.(Grinberg et al. 2019, Lin and Marr 2020, Majee et al. 2021, Deleplace et al. 2022, Cunliffe et al. 2023 ).

Effect of exposure media on antibacterial activity
Organic soiling is generally expected to reduce antimicrobial activity of surfaces.In this study organic proportion of soil load form ulation fr om the EPA method was as used.The individual components of soil load are bovine serum albumin (BSA), yeast extract and m ucin, whic h hav e been individuall y used for v arious purposes.BSA or fetal bovine serum (FBS) have often been used to mimic "dirty" conditions in various standard protocols and literature sources (e.g.EN 13697 (CEN 2015 )) but published results are contradictory.For example, BSA has been demonstrated to protect bacteria a gainst light-activ ated antimicr obial surfaces (Lour enço et al. 2018 ) but also increase efficacy of copper surfaces exposed to BSA-containing bacterial aerosols (Ojeil et al. 2013 ).Yeast extract has been shown to decrease the toxicity of heavy metals such as cadmium (Mariano-da-Silva et al. 2009 ).T he hea vil y gl ycosylated mucins in respiratory droplets may offer some protection against inactivation by drying in case of enveloped viruses (Yang et al. 2012 ) that could be considered somewhat similar to Gr am-negativ e bacteria with their envir onmentall y exposed outer membrane.
T he beha vior of one and the same surface in different conditions is illustrated in Fig. 7 .Lines for high-organic conditions (bro wn/y ello w) are generally grouping together to the top of the gr a phs irr espectiv e of RH clearl y indicating that antibacterial activity of the surfaces is more affected by the choice of exposure medium than drying of the inoculum.None of the test surfaces elicited any substantial antibacterial activity to w ar ds either bacterial species in high-organic exposure medium in low to medium RH.Ho w e v er, these ar e the conditions often encounter ed in r eal applications of commercial antimicrobial surfaces when we think of surfaces contaminated by bodily fluids or the food industry for example.High organic content also demonstrated a protective effect against desiccation effects in microdroplet inoculation, especially in low RH conditions where otherwise a substantial decrease in viable counts was also observed on control surfaces (Fig. 7 C and D).
Although higher efficacy than presented on Fig. 7 (bro wn/y ello w on panels A, E) definitely can be achieved in Figure 8.Effect of exposure medium, its organic content and individual organic soiling components on antibacterial activity of copper surfaces ( CuC , gr een) a gainst E. coli after 1 h exposur e in micr odr oplets at 50% r elativ e humidity compared to stainless steel ( SS , black).Organic soiling scenarios tested: 500-fold diluted nutrient broth from ISO 22196 ( 0.002x NB ; 0.026 g/L organics), 100-fold diluted EPA soil load ( 0.01x SL ; 0.068 g/L organics), 10-fold diluted EPA soil load ( 0.1x SL ; 0.68 g/L organics), undiluted EPA soil load ( 1x SL; 6.8 g/L or ganics), m ucin component of undiluted soil load ( 1x MUC ; 0.8 g/L), BSA component of undiluted soil load ( 1x BSA ; 2.5 g/L), yeast extract component from soil load ( 1x YE ; 3.5 g/L).Black dotted line represents log 10 CFU/surface inoculated onto the surfaces.Mean value of 5-8 data points is shown.Values at or above the detection limit of at least three colonies counted in all undiluted plated drops (red dotted line) were used for the statistical analysis of CFU counts.Lo w er v alues ar e shown to demonstr ate that gr owth of viable bacteria was detected in at least some replicates.
high-organic EPA soil load for certain copper-based surfaces as suc h pr oducts hav e been a ppr ov ed for commercial use in the US employing the EPA method (EPA 2022 ) to back the antimicrobial claims, it has to be noted that the method instructs to smear the inoculum over a large area on the test surface.We used non-smear ed dr oplets to intentionall y mimic contamination via spray or semi-wet cross-contamination.Earlier studies have shown that copper-based surfaces elicit lower antibacterial activity when surfaces are inoculated as undisturbed microdroplets as opposed to smearing the inoculum (Dauv er gne et al. 2020 ).In our pr e vious studies, we have seen similar effect and noted that bacteria survived better in 50 μl or 75 μl droplets of LB on copper surface than in 25 μl (Rosenberg et al. 2018 ).The latter also resulted in higher copper release from the surface during the 1 h exposure.Indeed, also in the current study, copper surfaces tended to be most efficient in case of micr odr oplet inoculation and high RH (Figs 4 B and 5 B), especially in the longest time-point and highest RH where the microdroplets had not completely dried on the surfaces ( Figs S3 and S4 ).Due to the test format and long exposures in dry conditions, we could unfortunately not determine the amount of released copper.It is also possible that differences in viability can at least partly arise from different flow dynamics and surface tension in the droplets of different exposure media during drying (Majee et al. 2021 ).
As the content of the EPA soil load (EPA 2022 ) is complex, we further aimed to clarify whether it is the total organic content or one of the individual organic components of the soil load medium that protects bacteria from the antibacterial as well as from drying effects.For that, we studied the antibacterial effect of the most effective metallic surface (copper) in comparison with stainless steel in the presence and absence of soil load or its components.
Figure 8 reveals that in moderate drying conditions (50% RH) and micr odr oplet inoculation in case of whic h micr odr oplets dry by the end of 1-hour exposure (Table 3 ) viability of E. coli on cop-per was clearly dependent on organic content of the exposure medium.No decrease in viability on copper surface ( P > 0.05) was detected in the presence of soil load r epr esenting high-or ganic environment and its 10-fold dilution (Fig. 8 ).The same was true when the components of soil load were used separately, indicating that any of the organic components in soil load is more than sufficient in eliminating the antibacterial activity of copper surfaces in moder atel y dry conditions.Onl y 100-fold diluted soil load enabled copper to cause significant reduction in viable count (-3.6 log 10 ; P < 0.0001) that was r elativ el y similar to that induced by a low-organic exposure medium (-4.8 log 10 ; P < 0.0001).While high organic content generally reduces antimicrobial activity of metal-based agents, exceptions can be found where the opposite is claimed.Contradicting our results with 3 g/L BSA in microdroplets, copper-based surfaces have also been shown to present faster antibacterial activity when exposed to aerosols containing microbes and 2.5 g/L BSA (Ojeil et al. 2013 ).To put the organic content in the EPA soil load (6.8 g/L) into a context, the commonly used l ysogen y br oth (LB; [36]) often used in laboratory conditions as a growth-supporting exposure medium contains double the amount, i.e. 15 g/L of organic ingredients.
Minor drying-related decrease of viability was also seen on stainless steel, where the effect to w ar ds E. coli in low-organic medium ( −0.5 log 10 from 1x soil load, P < 0.0001) was not seen in any of the other exposure media with higher organic content indicating that organic content, but not 0.8 g/L mucin alone, protected bacteria from desiccation.The latter disproves the idea that m ucin could exclusiv el y pr otect bacteria a gainst drying-r elated inactiv ation similarl y to env eloped viruses (Yang et al. 2012 ).

Conclusions
Here we carried out a systematic series of experiments in variable experimental conditions to understand, how r elativ e air humidity of the test envir onment, dr oplet size of the bacterial inoculum and organic surface soiling affect antibacterial activity of solid surfaces.Antibacterial activity of fiv e copper, silv er and quaternary ammonium-based surfaces was compared against control surfaces of stainless steel and glass using modified ISO 22196 condition which requires bacterial exposure in a thin layer of liquid in nutrient-poor medium, and bacterial exposure in high-or lowor ganic dr oplets that wer e exposed to v ariable r elativ e air humidity.Across the conditions tested, copper-based surfaces proved to be the most efficient follo w ed b y silver and Si-Quat surfaces with the latter showing potential effect specifically towards S. aureus .Although se v er al surfaces elicited > 3 log 10 r eduction in bacterial viability in different test conditions after 6 h exposure, the regulatory goal of at least 3 log 10 reduction in up to 1-2 h was only met in case of copper-based surfaces in modified ISO protocol or in low-organic exposures in droplet exposures.
Our results from droplet exposures indicated that antibacterial activity of the tested surfaces against E. coli and S. aureus decreased with decreasing RH and increasing organic soiling.At low RH, none of the surfaces exhibited biologically or statistically significant antibacterial acti vity, exce pt for copper in lar ge dr oplets with low organic content.At high RH, copper surfaces were found to be most efficient in the case of micr odr oplet inoculation.Results on Si-Quat surfaces sho w ed antibacterial effect only in modified ISO 22196 format and in large low-organic droplets, but high variability of the results made it impossible to draw further conclusions.
The studied environmental variables: humidity and organic content of the medium affected the viability of both E. coli and S. aureus .As a general rule, on surfaces with lo w er to no antibacterial activity, viability of E. coli was more affected by drying analyzed as a combined effect of RH and inoculum droplet size while S. aureus was more affected by the selection of exposure medium.
To study the effect of exposure medium on antibacterial effect, we compared the antibacterial activity of copper as the most effective surface and stainless steel as the control surface, in high organics containing soil load and its components compared with low organics environment and sho w ed that the organics in exposur e medium efficientl y inactiv ated the antibacterial activity of metal-based surfaces.
There has been a lot of discussion along with a recent ISO initiativ e to de v elop a dry standard test method (ISO/DIS 7581) alongside the current wet ISO 22196 industrial standar d.Ho w e v er, pr actical challenges in testing as well as knowledge gaps regarding the effects of environmental variables, especially drying, remain.Our study highlights the need for an open discussion about the testing conditions with an a pplication-r ele v ant test protocol beyond the wet versus dry formats and include spray inoculation as one of the worst-case scenarios to accur atel y assess the antimicr obial activity of solid surfaces in real-life like conditions.

Figure 2 .
Figure 2. Cross sections of CuT (CovidSafe copper tape), AgP (TOUCH Antimicr obial silv er paint) and SQ (Si-Quat coating) on stainless steel under SEM.In case of Co vidSafe , both, copper-containing (darker) surface layer and adhesive polymer (lighter) base layer can be seen.

Figure 3 .
Figure 3. Contact angle measurements of the studied surfaces (CuC, metallic coupon; CuT, CovidSafe copper tape; AgC, metallic silver; AgP, TOUCH Antimicrobial silver paint; SQ, Si-Quat coating; SS, stainless steel; NC, borosilicate glass).The coupons were rinsed with acetone and ethanol the same way as prior to antibacterial testing.Statistically significant differences ( P < 0.05) from the NC and SS control surfaces are marked with * * * * ( P ≤ 0.0001), * * ( P ≤ 0.01).The 90 • boundary above which the values can be considered hydrophobic, is marked with a red dotted line.

Figure 4 .
Figure 4. Antibacterial activity of the surfaces towards E. coli after 2 h exposure in low organic (A, B, D) and in high organic (C, E) media exposed either as a liquid layer ISO 22196 format (A), micr odr oplets (5 ×2 μL; B, C) or large droplet (50 μL; D, E).Log 10 reduction is calculated from glass (NC).The maximum possible log 10 reduction detectable depended on bacterial viability on glass control surface in each test condition.Maximum log 10 reduction was achieved by copper coupon (CuC) on panel A, B (50% RH) and D (20% RH) indicated by high antibacterial activity values with no errorbars .An a v er a ge of 3-6 biological r eplicates with standard de viation is shown.Statisticall y significant differ ence ( P < 0.05) fr om the NC contr ol surface is marked with * * * * ( P ≤ 0.0001), * * * ( P ≤0 001), * * ( P ≤ 0.01), * ( P < 0.05).The target of 3 log 10 reduction is displayed as a grey dotted line.Antibacterial activity after 0.5, 1, 2, and 6 h exposure is presented on Figs S3 and S4 , log 10 -transformed viable counts in TablesS1 and S2and raw data in Supplementary Raw Data file.

Figure 5 .
Figure 5. Antibacterial activity of the surfaces towards S. aureus after 2 h exposure in low organic (A, B, D) and in high organic (C, E) media exposed either as a liquid layer in ISO 22196 format (A), micr odr oplets (5 ×2 μL; B, C) or large droplet (50 μL; D, E).Log 10 reduction is calculated from glass (NC).The maximum possible log 10 reduction detectable depended on bacterial viability on glass control surface in each test condition.Maximum possible log 10 reduction of S. aureus was not achieved by any of the surfaces in any conditions after 2 h exposure.An av er a ge of 3-6 biological replicates with standard deviation is shown.Statistically significant difference ( P < 0.05) from the NC control surface is marked with * * * * ( P ≤ 0.0001), * * * ( P ≤0 001), * * ( P ≤ 0.01), * ( P < 0.05).On panel D, tw o-w ay ANOVA + post-hoc r esults ar e shown for the whole dataset (black * - * * * ) as well as with the most variable SQ gr oup r emov ed (r ed * * * * ).The tar get of 3 log 10 r eduction is displayed as a gr e y dotted line.Antibacterial acti vity after 0.5, 1, 2, and 6 h exposure is presented on Figs S3 and S4 , log 10 -transformed viable counts in TablesS1 and S2and raw data in Supplementary Raw Data file.

Figure 7 .
Figure7.Comparison of antibacterial effect of the most effective surface types and control surfaces to w ar ds E. coli (A-D) and S. aureus (E-H) in variable testing conditions.CuC, metallic copper; AgC, metallic silver; SS, stainless steel; NC borosilicate glass.Blue hues denote lo w and y ello w hues high organic content in exposure media.Lighter color indicates lo w er air humidity and dotted lines microdroplet inoculation.An average of 3-6 biological replicates with standard deviation is shown.Values at or above the detection limit of at least three colonies counted in all undiluted plated drops (dotted line at y = 3.18) were used for the statistical analysis of CFU counts.Lo w er v alues ar e shown to demonstrate that growth of viable bacteria was detected in at least some repetitions.Tabular presentation of mean values and standard deviations is presented in TablesS1 and S2and raw data in Supplementary Raw Data file.

Table 1 .
Envir onmental v ariables used in exposed lar ge or micr odr oplet format antibacterial activity assessment.
CuC -metallic copper; CuT -CovidSafe copper tape; AgC -metallic silver; AgP -TOUCH Antimicrobial silver paint; SQ-Si-Quat coating a Large contact area method, modified from ISO 22196 was also used for the gener al e v aluation of antibacterial activity of the surfaces but due to lar ge differ ences in inoculum cell counts not in direct comparison with environmental variables affecting bacterial viability on test surfaces.way, tw o-w ay, and three-w ay ANOVA follo w ed b y Tuck e y or Dunnett post-hoc tests at α = 0.05 were used where appropriate.

Table 2 .
Characteristics of the surfaces used in the study.