Efficient Generation of Plasma‐Activated Aerosols with High Concentrations of Reactive Species via Silicon Dioxide Coated Surface Acoustic Wave Devices

An effective method to deliver plasma‐activated water onto contaminated surfaces is with a surface acoustic wave (SAW) nebulizer that can efficiently produce plasma‐activated aerosols. While higher reactive species concentrations in the plasma‐activated aerosols promote more efficient bacterial inactivation and hence reduce the amount of aerosols needed to completely eradicate the bacteria colonies, higher concentrations of reactive species themselves are found to lead to lower nebulization rates due to the increased electrical conductivity of the solution. To circumvent this problem, it is shown that coating the SAW substrate with a thin insulating layer of silicon dioxide (SiO2) can significantly mitigate the nebulization rate reduction. Specifically, when the electrical conductivity increases from 0.25 to 4.0 mS cm−1, the 62% drop in the nebulization rate is observed with the uncoated SAW devices is substantially reduced to 19% simply with the SiO2 coating. This can be attributed to the synergistic effects of increased wettability and the suppression of the electrical body force acting on the liquid film. The concomitant increase in the bacterial colony reduction from 17% to 76% demonstrates this solution to be a facile yet effective way of enhancing the efficiency for spray‐based bacterial inactivation on contaminated surfaces.


Introduction
[3] UV disinfection, for instance, requires extended treatment durations and is ineffective oxidative stresses, [13][14][15] thereby damaging microbial cell membranes and effectively disinfecting them. [16,17][20][21][22][23] Recently, spray-based nebulizers with integrated atmospheric pressure plasma generators to produce plasma-activated aerosols for surface disinfection have been demonstrated, allowing for enhanced portability, quicker drying times, and versatility for use on diverse surfaces. [24]ost nebulizers employ nozzles to generate fine droplets by directing liquid through an orifice.Nozzles are however susceptible to clogging due to condensates or trapped gas in the orifice, leading to vapor locks. [25,26]Foreign impurities and particles, even an order of magnitude smaller than the nozzle diameter, have been shown to induce aggregation and clogging. [27,28]To address the inefficiencies arising from the need for constant cleaning and maintenance interruptions, nozzleless nebulizers, such as those that employ surface acoustic waves (SAWs) to generate a fine mist of micron-dimension aerosol droplets by destabilizing and hence breaking up a liquid's free surface through its incredibly large ((10 8 ms −2 )) surface acceleration, [29,30] have been proposed.
SAW nebulizers, however, typically have lower nebulization rates compared to nozzle-based nebulizers. [31,32]Various methods have thus been explored in the past to enhance the nebulization rate of SAW nebulizers.Besides changing the SAW frequency, [33] the SAW nebulization rate can also be enhanced through the use of different modulation schemes: 1) pulse-width modulation, i.e., the introduction of on and off cycles to the carrier signal to increase the maximum peak-to-peak voltage; [34,35] 2) amplitude modulation, which alters not only the carrier signal peak-to-peak voltage based on a modulation index but also the modulation frequency; [36,37] ; and 3) hybrid modulation, which is a combination of both the pulse-width and amplitude modulation schemes. [38]hen SAW devices were used to nebulize plasma-activated activated water, an approximate 76% increase in the nebulization rate was reported with the use of hybrid modulation as compared to amplitude modulation.Nevertheless, a specific limitation is the significant drop in the nebulization rate that is observed as the concentration of the reactive species in the plasma-activated water, as measured by its electrical conductivity, is increased; this reduction in the nebulization rate can be attributed to the increase in electric force exerted on the liquid film due to the increased electrical conductivity, which, in turn, increases the wettability and hence reduces the loading of the liquid film on the SAW substrate.For instance, approximately 68% reduction in the nebulization rate was reported for deionized water and plasmaactivated water (≈0.3 mS cm −1 ) with the SAW nebulizer. [39]hile it was shown that a thin aluminium layer atop the SAW substrate led to an increase of approximately 36% in nebulization rate, [40] this was found to be less useful in practice due to the tendency of the aluminium coating to lift off with extended operation.In this work, we show that it is possible to increase the nebulization efficiency of plasma-activated water with high concentrations of reactive species by utilizing a thin silicon dioxide (SiO 2 ) coating.Besides being a much stronger and hence significantly more durable bonding to the SAW substrate, SiO 2 coatings on SAW substrates have been used to lower the temperature coefficient, [41] and to increase the sensor sensitivity. [42]More im-portantly, previous studies have shown that thin SiO 2 coatings protect the transducer under the high excitation powers required for nebulization, during which the temperature of the SAW device can increase significantly, making them susceptible to acoustomigration damage and ultimately device failure. [43,44]Additionally, we note that SiO 2 is chemically compatible with a wide range of fluids and biological applications. [45,46]he rest of the paper is organized as follows.In Section 2.1, we discuss the effect of increasing reactive species concentrations in plasma-activated water that changes its physical properties, which, in turn, alters the transmission of the acoustic wave from the piezoelectric substrate into the liquid, and therefore causes a reduction in the nebulization rate.The experimental setup for the production of plasma-activated water via the atmospheric pressure plasma is shown in Figure 1a and the computational domains to simulate the transmission of the acoustic wave is shown in Figure 2. In Section 2.2, we discuss the method to minimize such reductions in the nebulization rate (due to the higher concentration of active species in plasma-activated water) after sandwiching a thin SiO 2 electrically-insulating layer in between the piezoelectric substrate and the plasma-activated water to suppress the electric force on the liquid film that drives its spreading and hence thinning.The experimental setup for the production of plasma-activated aerosols via the uncoated/SiO 2 coated SAW devices is shown in Figure 1b.We then demonstrate in Section 2.3 that this allows for a considerably higher concentration of reactive species to be nebulized for surface disinfection: in contrast to previous studies, [39,40,47] wherein the electrical conductivity of the plasma-activated water feedstock is below  e < 1 mS cm −1 , considerably higher concentrations ( e ⩽ 4 mS cm −1 ) can be nebulized at sufficiently high rates, leading to significantly lower volumes of plasma-activated aerosols required to achieve complete inactivation of bacterial colonies on the surface.

Physical Properties and Acoustic Wave Transmission
Increasing the plasma treatment duration can increase the concentration of reactive species in the treated solution, leading to an increase in the electrical conductivity  e of the solution.As can be seen in Figure 3, this is accompanied by an increase in the solution density  f and viscosity  f , and a decrease in its surface tension  f .Within the lower electrical conductivity range ( e ≲ 1 mS cm −1 ), these physical properties (i.e.,  f ,  f ,and  f ) have almost a linear relationship with  e , suggesting that the increase in reactive species in plasma-activated water is almost linear as  e increases.Additionally, with the increase in density and decrease in surface tension, we expect the sound speed in plasma-activated water to decrease as the electrical conductivity increases, since c f ≈ ( f / f ) 2/3 . [48,49]We quantified this by adding microparticles to the solutions: with a device with two SAW transducers and at low excitation powers (P e ≪ 1 W) prior to the onset of acoustic streaming, the microparticles aggregate at pressure nodes, from which it is possible to approximate the acoustic wavelength.In particular, we observe that the acoustic waves are transmitted through the plasma-activated water with increasingly shorter wavelengths as its reactive species concentration and hence electrical conductivity is increased: a reduction in the acoustic wavelength  f with The ground electrode in (a) is a circular aluminium plate with diameter and thickness of approximately 52 and 2 mm, respectively.b) Schematic (not to scale) illustrating the experimental setups for the production of plasma-activated aerosols with uncoated/coated SAW nebulizers.The inset in (b-i) shows an uncoated SAW device, which comprises a chip-scale piezoelectric (lithium niobate, LN) substrate.The coated SAW device is identical to the uncoated device shown in (b-i), other than for the optically-transparent 500 nm thick SiO 2 coating atop it.As shown in the inset (b-ii), the SAW device is inverted to facilitate the deposition of the aerosols onto a beaker that was placed atop a weighing scale to quantify the nebulization rate.To evaluate the effectiveness of the plasma-activated aerosols for bacterial inactivation, the beaker is replaced with bacteria-plated agar plates.A paper strip (0.3 mm thick and 8 mm wide) is used as a conduit for the DI/plasma-activated water between the reservoir and the SAW device; upon excitation of the SAW, a thin film (length L film and height H film ) emanates from the paper strip and destabilizes to produce aerosols.The images in (c) shows the resultant static contact angles  c of the sessile droplets that form when 8 μL DI ( e = 0.003 mS cm −1 ) and plasma-activated water ( e = 4.0 mS cm −1 ) droplets were dispensed onto the (c-i) uncoated, and, (c-ii) SiO 2 coated SAW devices; the measured values for the contact angles are reported in Table 1.
respect to that for DI water of approximately 7.7%, 11.5%, 13.4% and 16.2% is seen for  e = 1, 2, 3, and 4 mS cm −1 , respectively.Given that the excitation frequency f SAW is fixed, this is due to the decrease in the speed of sound in the plasma-activated water c f = f SAW  f with increasing  e .
Using these estimated physical properties for plasma-activated water, we then simulate the transmission of acoustic waves from the LN substrate into DI and plasma-activated water ( e = 4 mS cm −1 ) (Figure 4a,b, respectively).It can be seen that the acoustic field transmission into both solutions have similar characteristics, although the acoustic wave can be seen to travel slightly faster in DI water compared its plasma-activated counterpart (Figure 4a-i, b-i).Additionally, due to the slightly lower acoustic impedance ( f c f ) of plasma-activated water, we observe that the attenuation length of the SAW along the solid-liquid interface  SAW to be slightly longer for the case of plasma-activated water, i.e., the u x and u y acoustic particle velocity components decay at a slower rate along the solid-0liquid interface.This is consistent with that reported in previous studies, since  SAW ≈  SAW ( SAW c SAW )/( f c f ), where  SAW ,  SAW , and c SAW are the wavelength, density and sound speed of the SAW on the LN substrate, which should remain constant given that the SAW frequency is fixed.As such, a decrease in  f c f as the liquid physical properties are altered leads to an increase in  SAW .
When a thin layer of SiO 2 is coated on the LN substrate, we observe a slight increase in the x-component velocity (u f, x ) as Sketch showing the three computational domains used in the numerical simulations: a piezoelectric solid domain (LN substrate), a solid domain (SiO 2 coating), and a liquid domain (water).To minimize wave reflection from the boundaries, a perfectly matched layer (absorber) is adopted.The red box indicates the approximate location from where the contour plots reported in Figure 4 are obtained.
compared to the y-component (u f, y ) within the solid domain, in addition to a slight increase in the presence of bulk waves, as shown in Figure 4c-ii.Since the SiO 2 layer is relatively thin compared to the SAW wavelength (H e = 500 nm ≪ SAW ≈ 131 μm), we do not expect appreciable changes in  SAW and c SAW when the SiO 2 is present.

Effect of the Electrical Conductivity of Plasma-Activated Water on the Nebulization Rate
We observe a significantly smaller reduction in the nebulization rate with the SiO 2 coated SAW devices when the electrical conductivity of plasma-activated water is increased, as shown in Figure 5.For example, it can be seen from Figure 5a that the nebulization rate at P e = 2 W decreases by approximately 62% and 19% for the uncoated and coated SAW device, respectively, when the electrical conductivity is increased from that of DI water ( e = 0.003 mS cm −1 ) to the highest value obtained for the plasmaactivated water with a high concentration of reactive species ( e = 4 mS cm −1 ).This is also true at higher SAW nebulization powers (P e = 4 W), wherein the decrease in nebulization rates are approximately 34% and 4% for the uncoated and coated SAW device, respectively (Figure 5b), therefore alluding to the potential for the SiO 2 coated SAW devices to efficiently nebulize plasmaactivated water with higher concentrations of reactive species.
The decrease in the nebulization rate of plasma-activated water with increasing electrical conductivity on the uncoated SAW device can be related to the corresponding decrease in the length L film and height H film of the thin meniscus film (Table 1) at the edge of the paper strip on the LN substrate where the nebulization occurs (Figure 1b).At equilibrium, the thin film, when subject to the SAW, is typically governed by three dominant forces, which sets its characteristic length L film : the acoustic streaming force (F , x ≈  f u f, x u f, y / v ) that draws the liquid film from the edge of the paper strip toward the transducer (in the positive x f -direction) counter to the SAW propagation direction, [50][51][52] an electric Maxwell body force (F e, x ≈  e E x ) arising as a consequence of the evanescent electric field in the liquid associated with the SAW [53] that repels the film away from the transducer (in the negative x f -direction), and the surface tension force (F c ≈  f cos  c ) that acts to prevent spreading of the film; is the viscous boundary layer thickness,  = 2f SAW is the angular frequency,  e is the charge density, E x is the electric field in the xdirection, and  c is the contact angle between the liquid film and the substrate.Assuming u f, x and u f, y to be constant under a fixed excitation power to first approximation, we expect F , x to decrease when  e increases, leading to a decrease in L film .In contrast, F e, x increases due to the increase in charge density with more conductive solutions, which also leads to shorter films, i.e., smaller L film .While the decreasing surface tension  f and contact angle  c that accompanies the increase in reactive species concentration leads to a decrease in F c that results in a longer film, this is negligible in comparison to the increase in F e, x and the decrease in F , x that drives a significant reduction in L film (Table 1).It is both this decrease in the film length, which reduces the available interfacial area over which nebulization occurs as a result of the breakup of the free surface, together with the corresponding decrease in the film height H film (Table 1) with which the nebulization rate has been found to correlate ( ṁ ≈ H 2 f ), [54] that is responsible for the decrease in the nebulization rate observed with increasing  e (Table 1).Concomitantly, the mean aerosol dimension is also expected to decrease with smaller film heights: [54,55] consistent with that observed in Table 1.
The suppression of the electric Maxwell body force F e, x in the presence of the insulating SiO 2 layer, on the other hand, suggests that the counteracting effects of decreasing F , x is balanced by the decrease in F c , leading to negligible changes in L film (Table 1).This, together with the statistically insignificant change in the film height H film (Table 1) results in only a very small reduction (≈4%) in the nebulization rate (Figure 5b), and consequently a negligible change in the mean aerosol diameter, at least within the limits of the measurement technique.Parenthetically, it can be seen that when  e ≲ 1 mS cm −1 , ṁ can be seen to decrease linearly with increases in  e , which can be attributed to the almost linear change in the physical properties within that range (0.03 mS cm −1 ⩽ e ⩽ 1 mS cm −1 ; see Figure 3).

Effect of the SiO 2 Coating on the Nebulization Rate
It can be seen by comparing the rate at which identical solutions are nebulized that the SiO 2 coated SAW device delivers superior nebulization rates compared to its uncoated counterpart.At P e = 2 W, we observe the nebulization rate to be up to 10% higher for DI water ( e = 0.003 mS cm −1 ) and 137% for plasma-activated water ( e = 4 mS cm −1 ), whereas at P e = 4 W, the increase is up to 9% for DI water ( e = 0.003 mS cm −1 ) and 59% for plasmaactivated water ( e = 4 mS cm −1 ) (Figure 5).This enhancement in nebulization rate can be attributed predominantly to the appreciable increase in wettability with the SiO 2 coating-as evidenced by the decrease in the static contact angle, especially at higher values of  e (Table 1)-that facilitates the spreading of the liquid over the substrate, leading to longer films and hence a larger  (c) surface tension  f -of deionized water (DI) and plasma-activated water with different electrical conductivities  e .These trends are consistent with that reported in an earlier study, [47] at least up to the highest value of  e = 1 mS/cm measured in that work.The electrical conductivity for the DI water was approximately  e = 0.003 mS/cm and that for plasma-activated water was controlled by varying the treatment time, i.e., the exposure duration of DI water to atmospheric pressure plasma.All measurements were conducted at room temperature.The solid lines represent the linear regression for the electrical conductivity within the range 0.03 mS/cm ⩽ e ⩽ 1 mS/cm: (a)  f = 3 e + 989 (R 2 = 0.961), (b)  f = 0.106 e + 0.898 (R 2 = 0.974), and, (c)  f = −3.2f + 72.3 (R 2 = 0.961).The dashed lines were added to aid visualization.
interfacial area that can be nebulized. [29,55]In all cases, the height of the liquid film remains greater than the sound wavelength in water, i.e., H film >  f , therefore permitting transmission of the sound wave through the liquid.We also note a slight increase in the u x -component acoustic particle velocity with the SiO 2 coated device (Figure 4c-ii) and hence F , x , which also leads to longer films that promote higher nebulization rates.
We further verified that there is no significant change in the liquid film temperature during the nebulization process in both the uncoated and coated devices.At P e = 4 W, for example, the measured steady-state liquid film temperature on the uncoated SAW device is T f = 70.1°C± 0.3°C for DI water ( e = 0.003 mS cm −1 ) and T f = 68.5°C± 0.3°C for plasma-activated water ( e = 4.0 mS cm −1 ).On the SiO 2 coated device, the corresponding temperature is T f = 71.7°C± 0.2°C for DI water ( e = 0.003 mS cm −1 ) and 69.8°C ± 0.4°C for plasma-activated water ( e = 4.0 mS cm −1 ).The slightly higher liquid film temperature (ΔT < 2°C) can be attributed to the smaller thermal conductivity of SiO 2 thin film [56] (≪1.4 W(m K) −1 ) compared to that for LN (>4 W(m K) −1 ), which results in slightly less efficient heat dissipation from the device surface.We also note that plasma-activated water has a lower temperature as compared to DI water (ΔT < 2°C) irrespective of whether it is nebulized from the uncoated or coated device, which we attribute to the higher nebulization rate with plasmaactivated water that removes heat from the liquid film.

Effectiveness of Plasma-Activated Aerosols with High Concentrations of Reactive Species for Bacterial Inactivation
Having demonstrated the ability to efficiently generate plasmaactivated aerosols with higher reactive species concentrations with the SiO 2 coated SAW device, we now turn to show the effectiveness of these high reactive species concentration plasmaactivated aerosols for bacterial inactivation.By controlling the amount of plasma-activated aerosols that are sprayed using the SiO 2 coated SAW device at P e = 4 W onto the Escherichia coli (E.coli) plated agar plate, we observe that the percentage reduction in bacteria count on the plate increases with increasing electrical conductivity of the solutions, i.e., higher concentrations of reactive species lead to more effective bacterial inactivation, as shown in Figure 6.In particular, we observe an initially linear increase in the bacteria load reduction when the electrical conductivity of the solution is below  e ⩽ 2 mS cm −1 , i.e.,  e ≈  e , followed by a more gradual increase when  e ⩾ 2 mS cm −1 , i.e.,  e ≈ ln  e .Based on the plasma treatment system shown in Figure 1a, the treatment duration to attain a solution electrical conductivity of  e = 2 mS cm −1 is approximately t pt ≈ 9 h; to reach  e = 4 mS cm −1 , the treatment duration increases to approximately t pt ≈ 22 h.As such, we note that while plasma-activated aerosols with  e = 4 mS cm −1 can achieve the highest percentage reduction, i.e.,  e ≈ 76% compared to  e ≈ 60% for  e = 2 mS cm −1 , the treatment duration to reach such reactive species concentrations more than doubles.When these differences in treatment duration are considered, it would thus appear that plasma-activated water with electrical conductivity  e = 2 mS cm −1 is a more viable choice in practice, and is therefore used in the subsequent analysis to examine the effectiveness of spraying different quantities of plasma-activated aerosols onto the agar plates.
The effectiveness of this higher concentration of reactive species ( e = 2 mS cm −1 ) in inactivating bacteria can also clearly be seen by spraying different volumes of the plasma-activated aerosols onto the E. coli plated agar plates.With lower reactive species concentrations (e.g.,  e = 0.25 mS cm −1 ), we observe the  percentage reduction in bacteria load to increase almost linearly with the volume of aerosol deposited.Consistent with that reported in earlier work, [40] in which  e ≈ 69% reduction efficiency was obtained with the deposition of 0.6 mL of plasma-activated aerosols with  e ≈ 0.34 mS cm −1 onto the plate, we observe a similar bacteria load reduction here of  e ≈ 67% with 0.6 mL of plasma-activated aerosols with  e = 0.25 mS cm −1 .When the concentration of reactive species and hence the electrical conductivity of the solution is however increased to  e = 2 mS cm-1, the volume of plasma-activated aerosols to obtain the same percentage reduction is significantly reduced to only ≈0.23 mL.This reduction of volume is equivalent to a reduction in the spraying (treatment) time from approximately 75 s ( e = 0.25 mS cm −1 ) to just 29 s ( e = 2 mS cm −1 ) based on the measured nebulization rates in Figure 5.Alternatively, it can be seen that with the same volume (0.6 mL) of plasma-activated aerosols that are deposited onto the agar plate, a complete ( e = 100%) reduction in bacteria load can be obtained with the  e = 2 mS cm −1 solution (Figure 7), or  e > 99% with only 0.4 mL of plasma-activated aerosols.

Conclusion
We report the efficient production of plasma-activated aerosols with high reactive species concentrations with the use of SiO 2 coated SAW nebulization devices.Studies to date on the use of SAW nebulizers to generate plasma-activated aerosols have been limited to electrical conductivities below 1 mS cm −1 .With an insulating SiO 2 layer, we show the possibility of generating plasmaactivated aerosols with electrical conductivities up to 4 mS cm −1 , but with substantially lower penalties in the nebulization rate reduction (4% to 19% decrease in the nebulization rate when increasing the solution electrical conductivity from that of DI water to 4 mS cm −1 , as opposed to a 34% to 62% drop with uncoated devices).This can be attributed to the synergistic effects of increased wettability with the SiO 2 coating, in addition to its insulating properties that minimize the electrical body force associated with the evanescent electric field of the SAW on the liquid film from which the plasma-activated aerosols are generated during nebulization.Finally, we show the effectiveness of the plasmaactivated aerosols with higher reactive species concentration in inactivating bacteria: for the same volume of aerosols sprayed and deposited onto E. coli plated agar plates, the reduction in bacterial colony count increases more than fourfold from  e ≈ 17% to 76% when the electrical conductivity of the aerosols is increased from  e = 0.25 to 4.0 mS cm −1 .Such an ability not only reduces the spraying/treatment time required to achieve a specific bacteria load reduction, but also eliminates the need for excessive amount of aerosols that need to be generated and deposited onto contaminated surfaces (advantageous, for example, in hand sanitizing applications, where extended spraying durations can become inconvenient, while excessive amounts of plasma-activated water can result in discomfort), thus alluding to the practical potential of this surface disinfection technique.

Experimental Section
Production of Plasma-Activated Water: To produce plasma-activated water, an anode for the plasma system was constructed from an array consisting of 16 shortened needles fixed together side by side, as shown in Figure 1a.These needles were connected to the positive terminal of a high voltage DC power supply (SRS PS375; FuG Elektronik GmbH, Rosenheim, Germany) and were positioned above a petri dish filled with water in which Table 1.Estimated maximum liquid film length L film and height H film (Figure 1b), static contact angle  c , and mean aerosol diameter ϕ d for the liquid films and aerosols produced with the uncoated and SiO 2 coated (500 nm thick) SAW devices, respectively.The SAW excitation power and frequency were fixed at P e = 4 W and f SAW = 30.5MHz, respectively.Four different solutions were employed: DI water ( e = 0.003 mS cm −1 ) and plasma-activated water with electrical conductivities  e = 0.5, 2.0, and 4.0 mS cm −1 .Standard deviations in the measurements were independently calculated based on triplicate experiments.a round aluminium plate (ground electrode) was submerged.When a high electric field (≈10 kV) was applied between the needles and aluminium plate, a strong atmospheric pressure plasma was generated around the needles and directed toward the liquid-air interface.Plasma-activated water with different electrical conductivities ( e = 0.25, 0.5, 1, 1.5, 2, 2.5, 3, 3.5, and 4 mS cm −1 ) were produced by varying the treatment duration.
The electrical conductivity and viscosity of the plasma-activated water were measured using a conductivity meter (Lutron CD-4306; SIL Technology Sdn.Bhd., Subang Jaya, Selangor, Malaysia) and rheometer (MCR 102; Anton Paar, Graz, Austria), respectively, whereas its surface tension and

Generation of Plasma-Activated Aerosols via Surface Acoustic Wave Devices:
The surface acoustic wave (SAW) nebulization device consisted of a focusing elliptical single-phase unidirectional transducer (SPUDT; see Figure 1b-i) made from a 295 nm thick aluminium layer atop a 5 nm thick titanium layer patterned onto a 128°rotated Y-cut X-propagating singlecrystal lithium niobate (LN) piezoelectric substrate (Hangzhou Freqcontrol Electronic Technology Ltd., China) using a standard UV photolithography process.The SAW was generated by applying a sinusoidal electrical signal from a function generator (AFG1062; Tektronix Inc., Beaverton, Oregon, USA) and high-frequency amplifier (ZHL-5W-1; Mini-Circuits Inc., Brooklyn, NY, USA) to the SPUDT at a resonant frequency of f SAW = 30.5MHz, which was determined by the spacing and width of the SPUDT fingers.To investigate the effect of incorporating a thin electrical insulation layer between the LN substrate and the conductive liquid (i.e., plasmaactivated water), a thin (500 nm) coating of SiO 2 was deposited onto the SAW device via an RF sputtering process; [57] the SiO 2 coating was etched away using Ar ion etching at the electrical connection pads.
To increase the efficiency of the SAW devices, amplitude modulation with a modulation frequency of f m = 1 kHz was utilized to reduce the power required to maintain continuous nebulization.The total power of the electric signal to the transducer can be calculated from P e = P c (1 + m 2 /2), where P c = V c I c is the RMS power of the carrier signal and m = V m /V c is the modulation index, whereby V m = (V max − V min )/2 is the modulated signal voltage and I c = (I max + I min )/2 is the carrier signal current.V max is the maximum RMS voltage and V min is the minimum RMS voltage, whereas I max is the maximum RMS current and I min is the minimum RMS current, measured using an oscilloscope (TDS 2012C; Tektronix Inc., Beaverton, OR, USA) connected to voltage (TPP 0201; Tektronix Inc., Beaverton, Oregon, USA) and current (P6022; Tektronix Inc., Beaverton, OR, USA) probes.A paper strip (WIP-100DLE; Suorec Sdn.Bhd., Batu Berendam, Melaka, Malaysia) of length 80 mm, width 8 mm and thickness 300 μm was used as a conduit for the water from the reservoir to the SAW device: upon excitation of the SAW, the liquid was drawn from the reservoir through the paper strip and onto the SAW device due to the negative pressure it generated in the paper strip.The liquid emanating from the paper strip then forms a thin liquid film meniscus on the SAW device with height H f and width L f (see Figure 1b), whose interface subsequently destabilizes and breaks up under the SAW substrate surface acceleration to generate micron-dimension aerosols.
The plasma-activated aerosols that were generated through the nebulization process were subsequently collected in a beaker placed directly underneath the inverted SAW device over 5 min to quantify the nebulization rate.The weight of the beaker before and after the aerosol collection process was measured with a weighing scale (MS303S/01; Mettler Toledo, Greifensee, Switzerland).The diameters of the plasma-activated aerosols with different electrical conductivities  e = 0.5, 2 and 4 mS cm −1 , on the other hand, were estimated by examining the aerosols that were deposited onto a slide under a microscope (Eclipse Ci-E; Nikon Inc., Minato, Tokyo, Japan).Additionally, changes in the dimensions of the thin plasma-activated water meniscus film emanating from the paper strip atop the uncoated and coated SAW devices were also examined by using a high-speed camera (Phantom M310; Vision Research Inc., Wayne, NJ, USA) equipped with a long distance magnifying lens (1-50486; Navitar, Rochester, NY, USA), together with a high intensity illuminator (41723series; Cole Parmer, Vernon Hills, IL, USA).To study the temperature differences of the liquid film on the uncoated and SiO 2 coated SAW devices, a thermocouple was placed in contact with the liquid in the nebulisation zone and recorded the readings using a data logger (GL820, Graphtec America Inc., Irvine, CA, USA).
Bacterial Inactivation: To investigate the effectiveness of using the plasma-activated aerosols to inactivate bacteria, the aerosols were deposited on agar plates plated with Escherichia coli (E.coli).Briefly, Luria-Bertani (LB) broth was first prepared by mixing 0.25 g of LB broth powder (LB Broth Miller 1.10285.0500;Merck KGaA, Darmstadt, Germany) along with 10 mL of DI water, 10 μL of 25 mgmL −1 kanamycin (Thermo Scientific ™ kanamycin sulfate CAS 25389-94-0; Thermo Fisher Scientific Inc., Waltham, MA, USA) and 10 μL of 30 mgmL −1 chloramphenicol (Calbiochem Chloramphenicol CAS 56-75-7; Merck KGaA, Darmstadt, Germany).The mixture was then autoclaved (HICLAVE HV-110; Hirayama Manufacturing Corp., Toyono-chō, Osaka, Japan) for 15 min at 121 °C prior to use.200 μL of E. coli (BL21 and DE3 strains) glycerol stock was subsequently cultured in 10 mL LB broth, with the cultivation process lasting for 16 h at 37 °C and 200 rpm in an incubation shaker (Certomat IS; Sartorius Stedim Biotech GmbH, Goettingen, Germany).After this incubation period, 1 mL of the culture was extracted and subsequently added to 50 mL of LB broth, together with a 50 μL mixture that consisted of 25 mgmL −1 kanamycin and 30 mgmL −1 chloramphenicol.The suspensions were then further incubated for 5 h at 37 °C and 200 rpm until the optical density (O.D.) of the culture increased by 1.0 with a logarithmic growth phase of approximately 10 8 CFU/mL.The O.D. was quantified using an absorbance microplate reader (Infinite M Nano; Tecan Group Ltd., Männedorf, Switzerland).100 μL aliquots from the prepared culture were then centrifuged (Sorvall Legend Micro 21 Microcentrifuge; Thermo Fisher Scientific Inc., Waltham, MA, USA) for 5 min at 5000 rpm to obtain a pellet at the bottom of the vial.The supernatant was slowly removed by using a pipette and replaced with 1 mL of normal saline, which comprised 0.9% sodium chloride solution (Sodium chloride A.R/ACS; R&M Chemicals Sdn.Bhd., Petaling Jaya, Selangor, Malaysia).The vial was subsequently agitated using a vortex shaker (Ika Genius 3; IKA-Werke GmbH & Co. KG, Staufen im Breisgau, Baden-Württemberg, Germany) to dissolve the pellet, following which the suspension was serially diluted with normal saline until a concentration of 10 3 CFU/mL was obtained.
To prepare the LB agar plates for E. coli plating, 18.5 g of LB agar powder (LB Agar Miller 1.10283.0500;Merck KGaA, Darmstadt, Germany) was first mixed with 500 mL DI water, followed by constant stirring while boiling to ensure complete dissolution.The agar mixture was then autoclaved for 15 min at 121 °C and cooled to under 50 °C, after which 500 μL of 25 mgmL −1 kanamycin was added to the agar mixture.10 mL of the agar mixture was subsequently dispensed into sterile petri dishes at least 30 min prior to use to ensure the agar mixture was set.
200 g of plasma-activated aerosols with electrical conductivities  e = 0.25, 0.5, 1, 2, 3, and 4 mS cm −1 were then nebulized directly onto the bacteria-plated agar plates.Additionally, the effect of different weights (0.05-0.6 g) of plasma-activated aerosols deposited on the agar plates was also studied.Following their exposure to the plasma-activated aerosols, the bacteria colonies on the agar plates were incubated at 37 °C for 12 h, after which the percentage reduction of colony forming units (CFU) before and after the treatment was estimated from to assess the bacteria inactivation efficiency.Acoustic Field Simulation: A simple 2D numerical model was formulated to elucidate the transmission of the acoustic waves from the LN substrate into the SiO 2 layer and subsequently the liquid above it.The computational domain was delineated into three regimes (Figure 2), namely, the piezoelectric solid domain comprising the LN substrate, a solid domain constituted by the SiO 2 coating, and the liquid domain above the coating.In brief, to model the generation and propagation of the acoustic waves in the piezoelectric substrate, the time-domain constitutive equation governing the motion of a piezoelectric solid is employed: [58][59][60][61][62] D i t = e ijkl S kl t where D i = ϵ ik E k is the electric displacement, E k is the electric field, ϵ ik are the dielectric coefficients, t is the time, e ikl are the piezoelectric stress coefficients, T ij are the stress components, and c ijkl are the elastic stiffness coefficients; the superscripts S and E denote that these quantities are measured at constant strain and constant electric field, respectively.Here, the electric field was quasi-static and hence ∂D i /∂t ≈ 0was assumed.Together with the infinitesimal strain-displacement relationship, Equations ( 2) and ( 3) can be simplified to ) where  is the particle displacement and v is the velocity of the solid.Equations ( 5) and ( 6) are solved together with Newton's second law of motion simultaneously using a finite difference time-domain method; in the above,  s is the mass density of the substrate.For the computation of the acoustic wave propagation through the SiO 2 coating, the stress-strain relationship (i.e., Equation ( 6), but without the second term on the righthand-side of the equation) was solved together with Equation ( 7), assuming the SiO 2 coating to be isotropic.The density of SiO 2 was approximately 2200 kgm − 3 and the Poisson ratio was assumed to be approximately 0.17.
To model the acoustic wave propagation in the liquid, on the other hand, a regular perturbation expansion of the liquid velocity u was undertaken, pressure p, and density  fields in the asymptotically small  ≡ U/c 0 limit (U being the local characteristic velocity scale of the fluid elements and c 0 the speed of sound) in the equations governing conservation of mass and momentum in the liquid, which gives rise to the following first order approximations: [63,64]  1 t +  0 (∇ ⋅ u 1 ) = 0 ( 8 ) and the first order approximation of the equation of state for an adiabatic process, The zeroth term denoted by the subscript '0' refers to the unperturbed equilibrium state whereas first order approximations were denoted by the subscript '1' and constitute the solution for the propagation of the sound wave in the liquid.
To close the equations, a stress-free boundary condition was assumed on the surfaces of the piezoelectric substrate/SiO 2 coating not in contact with the liquid.Split-field perfectly matched layers (PMLs) to minimize wave reflection from the boundaries, on the other hand, were adopted along the left, right and bottom boundaries of the piezoelectric solid domain (LN substrate), the left and right boundaries of the solid domain (SiO 2 coating), and the right and top boundaries of the liquid domain (DI/plasma-activated water); the wave amplitude decays quadratically within the PML. [65,66]o generate the SAW, a sinusoidal electric potential, ϕ = ϕ p-p sin (2x/ SAW )sin (t), was imposed on the surface of the piezoelectric substrate, wherein ϕ p-p is the peak-to-peak voltage and  = 2f SAW the angular frequency.The SAW frequency was set at f SAW = 30.5MHz.At the interfaces between the piezoelectric substrate and the SiO 2 coating, and between the SiO 2 coating and the liquid, the domains were coupled through continuity in the velocities and stresses

Figure 1 .
Figure1.a) Schematic (not to scale) illustrating the production of plasma-activated water with different electrical conductivities via atmospheric pressure plasma (image shown in the inset).The ground electrode in (a) is a circular aluminium plate with diameter and thickness of approximately 52 and 2 mm, respectively.b) Schematic (not to scale) illustrating the experimental setups for the production of plasma-activated aerosols with uncoated/coated SAW nebulizers.The inset in (b-i) shows an uncoated SAW device, which comprises a chip-scale piezoelectric (lithium niobate, LN) substrate.The coated SAW device is identical to the uncoated device shown in (b-i), other than for the optically-transparent 500 nm thick SiO 2 coating atop it.As shown in the inset (b-ii), the SAW device is inverted to facilitate the deposition of the aerosols onto a beaker that was placed atop a weighing scale to quantify the nebulization rate.To evaluate the effectiveness of the plasma-activated aerosols for bacterial inactivation, the beaker is replaced with bacteria-plated agar plates.A paper strip (0.3 mm thick and 8 mm wide) is used as a conduit for the DI/plasma-activated water between the reservoir and the SAW device; upon excitation of the SAW, a thin film (length L film and height H film ) emanates from the paper strip and destabilizes to produce aerosols.The images in (c) shows the resultant static contact angles  c of the sessile droplets that form when 8 μL DI ( e = 0.003 mS cm −1 ) and plasma-activated water ( e = 4.0 mS cm −1 ) droplets were dispensed onto the (c-i) uncoated, and, (c-ii) SiO 2 coated SAW devices; the measured values for the contact angles are reported in Table1.

Figure 3 .
Figure 3. Measured physical properties-(a) density  f , (b) viscosity  f and (c) surface tension  f -of deionized water (DI) and plasma-activated waterwith different electrical conductivities  e .These trends are consistent with that reported in an earlier study,[47] at least up to the highest value of  e = 1 mS/cm measured in that work.The electrical conductivity for the DI water was approximately  e = 0.003 mS/cm and that for plasma-activated water was controlled by varying the treatment time, i.e., the exposure duration of DI water to atmospheric pressure plasma.All measurements were conducted at room temperature.The solid lines represent the linear regression for the electrical conductivity within the range 0.03 mS/cm ⩽ e ⩽ 1 mS/cm: (a)  f = 3 e + 989 (R 2 = 0.961), (b)  f = 0.106 e + 0.898 (R 2 = 0.974), and, (c)  f = −3.2f + 72.3 (R 2 = 0.961).The dashed lines were added to aid visualization.
(a) LN and DI water (b) LN and plasma-activated water (c) SiO 2 coated LN and plasma-activated water Normalized acoustic particle velocity magnitude U/U max Normalized acoustic particle velocity u/U

Figure 4 .
Figure 4. Contour plots of the computed instantaneous magnitude of acoustic wave transmission (a) from the LN substrate into DI water, (b) from the LN substrate into plasma-activated water,and, (c) from the SiO 2 coated LN substrate into plasma-activated water.Note that for (c), the contour for the SiO 2 coating was not shown as its thickness (H c = 500 nm) is relatively small compared to the SAW wavelength  SAW ≈ 131 μm.Shown below are the corresponding instantaneous acoustic particle velocities in the (i) solid (u s, x , u s, y ), and, (ii) liquid (u f, x , u f, y ) domains along the location indicated by the gray line in the contour plots.For plasma-activated water ( e ≈ 4 mS cm −1 ), the measured properties shown in Figure3were used in the simulations.The excitation frequency is f SAW = 30.5MHz and the results are shown after ten sinusoidal cycles, each with a period of f −1 SAW , i.e., t = 10f −1 SAW .

Figure 6 .
Figure6.Estimated percentage reduction  e in the bacterial colony count as a measure of the effectiveness of the plasma-activated aerosols generated with the SiO 2 coated SAW device with different electrical conductivities  e = 0.25, 0.50, 1.00, 2.00, 3.00, 3.5, and 4.0 mS cm −1 for bacterial inactivation.The dashed line represents the linear regression (R 2 = 0.998) of the data over the range 0 mS cm −1 ⩽ e ⩽ 2 mS cm −1 , whereas the solid line represents the logarithmic fitting (R 2 = 0.995) of the data over the range 2 mS cm −1 ⩽ e ⩽ 4 mS cm −1 .The weight of plasma-activated aerosols deposited on the E. coli plated agar plate was fixed at 0.2 g.The SAW excitation frequency was f SAW = 30.5MHz and the excitation power was fixed at P e = 4 W. The error bars indicate the standard deviation of the data acquired across triplicate measurements.

Figure 7 .
Figure 7.Estimated percentage reduction  e in the bacterial colony count under different weights of plasma-activated aerosol generated with the SiO 2 coated SAW device with electrical conductivities  e = 0.25 mS cm −1 (+) and 2.0 mS cm −1 (×) deposited on the E. coli plated agar plate.The SAW excitation frequency was f SAW = 30.5MHz and the excitation power was fixed at P e = 4 W. Trendlines were added to aid visualization and the error bars indicate the standard deviation of the data acquired across triplicate measurements.density were measured using a force tensiometer (Sigma 702; DYNE Testing Ltd., Lichfield, UK).Generation of Plasma-Activated Aerosols via Surface Acoustic Wave Devices:The surface acoustic wave (SAW) nebulization device consisted of a focusing elliptical single-phase unidirectional transducer (SPUDT; see Figure1b-i) made from a 295 nm thick aluminium layer atop a 5 nm thick titanium layer patterned onto a 128°rotated Y-cut X-propagating singlecrystal lithium niobate (LN) piezoelectric substrate (Hangzhou Freqcontrol Electronic Technology Ltd., China) using a standard UV photolithography process.The SAW was generated by applying a sinusoidal electrical signal from a function generator (AFG1062; Tektronix Inc., Beaverton, Oregon, USA) and high-frequency amplifier (ZHL-5W-1; Mini-Circuits Inc., Brooklyn, NY, USA) to the SPUDT at a resonant frequency of f SAW = 30.5MHz, which was determined by the spacing and width of the SPUDT fingers.To investigate the effect of incorporating a thin electrical insulation layer between the LN substrate and the conductive liquid (i.e., plasmaactivated water), a thin (500 nm) coating of SiO 2 was deposited onto the SAW device via an RF sputtering process;[57] the SiO 2 coating was etched away using Ar ion etching at the electrical connection pads.To increase the efficiency of the SAW devices, amplitude modulation with a modulation frequency of f m = 1 kHz was utilized to reduce the power required to maintain continuous nebulization.The total power of the electric signal to the transducer can be calculated from P e = P c (1 + m 2 /2), where P c = V c I c is the RMS power of the carrier signal and m = V m /V c is the modulation index, whereby V m = (V max − V min )/2 is the modulated signal voltage and I c = (I max + I min )/2 is the carrier signal current.V max is the maximum RMS voltage and V min is the minimum RMS voltage, whereas I max is the maximum RMS current and I min is the minimum RMS current, measured using an oscilloscope (TDS 2012C; Tektronix Inc., Beaverton, OR, USA) connected to voltage (TPP 0201; Tektronix Inc., Beaverton, Oregon, USA) and current (P6022; Tektronix Inc., Beaverton, OR, USA) probes.A paper strip (WIP-100DLE; Suorec Sdn.Bhd., Batu Berendam, Melaka, Malaysia) of length 80 mm, width 8 mm and thickness 300 μm was used as a conduit for the water from the reservoir to the SAW device: upon excitation of the SAW, the liquid was drawn from the reservoir through the paper strip and onto the SAW device due to the negative pressure it generated in the paper strip.The liquid emanating from the paper strip then forms a thin liquid film meniscus on the SAW device with height H f and width L f (see Figure1b), whose interface subsequently destabilizes and breaks up under the SAW substrate surface acceleration to generate micron-dimension aerosols.The plasma-activated aerosols that were generated through the nebulization process were subsequently collected in a beaker placed directly underneath the inverted SAW device over 5 min to quantify the nebulization rate.The weight of the beaker before and after the aerosol collection process was measured with a weighing scale (MS303S/01; Mettler Toledo, Greifensee, Switzerland).The diameters of the plasma-activated aerosols with different electrical conductivities  e = 0.5, 2 and 4 mS cm −1 , on the other hand, were estimated by examining the aerosols that were deposited onto a slide under a microscope (Eclipse Ci-E; Nikon Inc., Minato, Tokyo, Japan).Additionally, changes in the dimensions of the thin plasma-activated water meniscus film emanating from the paper strip atop the uncoated and coated SAW devices were also examined by using a high-speed camera (Phantom M310; Vision Research Inc., Wayne, NJ, USA) equipped with a long distance magnifying lens (1-50486; Navitar, Rochester, NY, USA), together with a high intensity illuminator (41723series; Cole Parmer, Vernon Hills, IL, USA).To study the temperature differences of the liquid film on the uncoated and SiO 2 coated SAW devices, a thermocouple was placed in contact with the liquid in the nebulisation zone and recorded the readings using a data logger (GL820, Graphtec America Inc., Irvine, CA, USA).Bacterial Inactivation: To investigate the effectiveness of using the plasma-activated aerosols to inactivate bacteria, the aerosols were deposited on agar plates plated with Escherichia coli (E.coli).Briefly, Luria-Bertani (LB) broth was first prepared by mixing 0.25 g of LB broth powder (LB Broth Miller 1.10285.0500;Merck KGaA, Darmstadt, Germany) along with 10 mL of DI water, 10 μL of 25 mgmL −1 kanamycin (Thermo Scientific ™ kanamycin sulfate CAS 25389-94-0; Thermo Fisher Scientific