key: cord-1015462-vj8p2mzr authors: Aydin, Onur; Emon, Bashar; Cheng, Shyuan; Hong, Liu; Chamorro, Leonardo P.; Saif, M. Taher A. title: Performance of fabrics for home-made masks against the spread of COVID-19 through droplets: A quantitative mechanistic study date: 2020-08-11 journal: Extreme Mech Lett DOI: 10.1016/j.eml.2020.100924 sha: f16b7363c30e7a495a82506a4dca2458347ca8c9 doc_id: 1015462 cord_uid: vj8p2mzr Coronavirus Disease 2019 (COVID-19) may spread through respiratory droplets released by infected individuals during coughing, sneezing, or speaking. Given the limited supply of professional respirators and face masks, the U.S. Centers for Disease Control and Prevention (CDC) has recommended home-made cloth face coverings for use by the general public. While there have been several studies on aerosol filtration performance of household fabrics, their effectiveness at blocking larger droplets has not been investigated. Here, we ascertained the performance of 11 common household fabrics at blocking large, high-velocity droplets, using a commercial medical mask as a benchmark. We also assessed the breathability (air permeability), texture, fiber composition, and water absorption properties of the fabrics. We found that most fabrics have substantial blocking efficiency (median values >70%). In particular, two layers of highly permeable fabric, such as T-shirt cloth, blocks droplets with an efficiency (>94%) similar to that of medical masks, while being approximately twice as breathable. The first layer allows about 17% of the droplet volume to transmit, but it significantly reduces their velocity. This allows the second layer to trap the transmitted droplets resulting in high blocking efficacy. Overall, our study suggests that cloth face coverings, especially with multiple layers, may help reduce droplet transmission of respiratory infections. Furthermore, face coverings made from materials such as cotton fabrics allow washing and reusing, and can help reduce the adverse environmental effects of widespread use of commercial disposable and non-biodegradable facemasks. professional respirators and face masks, the U.S. Centers for Disease Control and Prevention 18 (CDC) has recommended home-made cloth face coverings for use by the general public. While 19 there have been several studies on aerosol filtration performance of household fabrics, their 20 effectiveness at blocking larger droplets has not been investigated. Here, we ascertained the 21 performance of 11 common household fabrics at blocking large, high-velocity droplets, using a 22 commercial medical mask as a benchmark. We also assessed the breathability (air permeability), 23 texture, fiber composition, and water absorption properties of the fabrics. We found that most 24 fabrics have substantial blocking efficiency (median values >70%). In particular, two layers of 25 highly permeable fabric, such as T-shirt cloth, blocks droplets with an efficiency (>94%) similar to 26 that of medical masks, while being approximately twice as breathable. The first layer allows about 27 17% of the droplet volume to transmit, but it significantly reduces their velocity. This allows the 28 second layer to trap the transmitted droplets resulting in high blocking efficacy. Overall, our study 29 suggests that cloth face coverings, especially with multiple layers, may help reduce droplet 30 transmission of respiratory infections. Furthermore, face coverings made from materials such as 31 cotton fabrics allow washing and reusing, and can help reduce the adverse environmental effects 32 of widespread use of commercial disposable and non-biodegradable facemasks. 33 34 35 36 Keywords: cloth face covering, face mask, SARS-CoV-2, COVID-19, respiratory droplets, droplet 37 blocking, breathability 38 Respiratory infections caused by novel pathogenic agents (e.g., a novel virus) can lead to 39 epidemics or pandemics. Existing knowledge from respiratory infections such as influenza, SARS- 40 1, and MERS indicates three major routes of transmission, namely, droplets, aerosols, and 41 contact. [1] [2] [3] Although the mechanism of spread of the current novel coronavirus (SARS-CoV-2) is 42 not clearly understood, it is thought that spread can occur through respiratory droplets 43 containing virus particles that are released by infected persons when they sneeze, cough, or 44 speak. 4, 5 Larger droplets tend to fall nearby by gravity, and the sufficiently smaller ones can stay 45 in the air and travel longer. 1, 6, 7 Droplets containing viruses may be shed by symptomatic as well 46 as pre-symptomatic and asymptomatic individuals. [8] [9] [10] [11] If inhaled by a healthy individual, droplets 47 allow the virus to enter the respiratory system and cause infection. Face coverings offer a physical 48 barrier against virus transmission, and can be especially useful as a method of source control. For 49 example, medical masks have recently been shown to be highly effective in reducing the 50 dissemination of droplets of all sizes from COVID patients. 12 51 52 During the COVID-19 pandemic, the supply of commercially manufactured respirators and 53 facemasks has not been able to meet the demand. The U.S. CDC has therefore provided guidance 54 for the public to use alternatives such as cloth face coverings to slow the spread of COVID-19. 13 55 However, it is not yet clear what kind of fabric would be the most efficient material or how many 56 layers of cloth would protect against both spreading and contracting the virus. Existing literature 57 on cloth masks have mostly focused on the filtration efficiency of household materials against 58 dry or liquid aerosols with particle/droplet sizes within the range of ~10 nm to ~5-10 µm. 14-19 59 Such studies are based on the well-established theoretical and experimental foundations of 60 aerosol filtration by fibrous media. However, when an individual coughs, sneezes, or speaks, the 61 droplets that are released typically have a size distribution that includes larger droplets (up to ~1 62 mm in size) in addition to the aerosolized fraction (<5-10 µm size). [20] [21] [22] [23] Established knowledge 63 from aerosol science may not be immediately applicable to determine the efficiency of mask 64 materials at blocking larger droplets carrying virus particles. How effective cloth face coverings 65 can be at reducing transmission via large droplets therefore remains elusive. 66 67 To address this issue in the current context of COVID-19, we developed a method of quantifying 68 the effectiveness of fabrics at blocking large droplets containing 100 nm-diameter nanoparticles 69 which serve as a mimic for viruses in terms of size. We considered a wide range of regular 70 household fabrics using a commercial medical mask as a benchmark. Our work is meant to 71 complement existing information on the aerosol filtration efficiency of household fabrics [14] [15] [16] [17] [18] [19] and 72 offer insight into the possible mechanisms of how fabrics may block virus transmission via large 73 respiratory droplets. 74 75 In this study, we did not consider methods for producing a facemask, e.g., how they should be 76 stitched, how their boundaries should be designed, how to attach them to the face, and how 77 they should be used or decontaminated. For detailed information on how to make, use, and 78 decontaminate cloth face coverings, we refer the reader to guidelines provided by the U.S. CDC 79 and World Health Organization. 13,24 80 Problem Definition 83 84 Aerosol Filtration versus droplet blocking: We first illustrate the key distinctions between filtering 85 aerosol particles and blocking larger droplets by fabrics ( Figure 1 ). There is a well-established 86 theoretical and experimental body of literature on aerosol filtration by fibrous materials. [25] [26] [27] [28] [29] [30] [31] For 87 aerosols, particle sizes considered are often within the ~10 nm to ~5-10 µm range, smaller than 88 or comparable to the fiber diameter and inter-fiber spacing of the fibrous filter. Typically, the 89 particles are treated as solids. This holds naturally for dry aerosols, but it is also a reasonable 90 assumption for liquid aerosols since, at the small scale, surface tension dominates and droplets 91 behave as solid particles. Such particles can be captured by fibers of the filter via mechanisms 92 such as direct impaction, interception, and diffusion ( Figure 1A ). Some particles can pass through 93 the inter-fiber spaces (i.e. pores) as projectiles or be carried across by bulk fluid flow. Particles 94 that are larger than the pores are simply blocked by straining or settling/caking. This is where the 95 key distinction between established aerosol filtration models and blocking of large droplets 96 emerges: While large solid particles will simply be blocked, a large droplet with sufficient 97 momentum can squeeze through the pores of the fabric against shear stress and surface tension 98 barriers ( Figure 1B ). This is a complex phenomenon involving non-equilibrium processes, 99 interface energies, and short time-scale events. Existing models of aerosol filtration may 100 therefore not be sufficient in predicting outcomes. This reveals a gap in the understanding of the 101 potential effectiveness of cloth face coverings in blocking virus particles carried by large droplets. 102 Our goal here is to close this gap, at least partially, through experimental studies with 11 different 103 household fabrics and commercial medical mask. We first identify the essential parameters for 104 droplet blocking outlined below. 105 106 Breathability and droplet blocking efficiency -the two key parameters for face coverings: Any 107 mask material must offer sufficient breathability (i.e., air permeability) and yet efficiently block 108 virus particles carried by droplets. In contrast to fit-tested respirators, medical masks or cloth 109 face coverings typically cannot ensure tight sealing against the contours of the face. As a result, 110 a significant portion of the air released during breathing, sneezing, and coughing may escape 111 through the gaps, potentially carrying some of the respiratory droplets with virus particles with 112 it. 32 A mask material with low breathability (high resistance to air flow across the mask) will result 113 in relatively large leakage, defeating the purpose of the mask, and providing a false sense of 114 protection -even if the mask material itself is highly efficient at filtering respiratory droplets. 115 Higher breathability can lead to less leakage as more air will pass through the mask material 116 which can block some of the droplets. However, higher air permeability may also correspond to 117 lower blocking efficiency. The problem of finding an appropriate material for a home-made mask 118 therefore involves a trade-off between breathability, , and efficiency, , of blocking virus 119 particles carried by droplets. Hence, we consider and as the two critical parameters for mask 120 materials. Throughout the rest of the manuscript, we refer to as "droplet blocking efficiency" 121 for short. 122 123 Here, we considered a diverse set of 11 common household fabrics, and used a medical mask as 124 our benchmark ( Figure 2 ) . The household fabrics were selected from new and used garments, 125 quilt cloths, bed sheet, and dishcloth, and characterized in terms of their fabric construction 126 (woven, knit, or napped), fiber content (cotton, polyester, polyamide, silk), weight, thread count, 127 porosity, and water absorption rate (see Methods). Sample descriptions and parameter values 128 are listed in Table 1 . We measured the droplet blocking efficiency and breathability of the fabrics 129 in a laboratory setting and empirically assessed the relationship between these two critical 130 parameters. 131 132 Droplet Blocking Efficiency 133 134 We developed a method to challenge the fabrics with large droplets and to quantify their droplet 135 blocking efficiency. A schematic of our experimental method is shown in Figure 3A . To generate 136 large droplets with high initial velocity, we repurposed a metered-dose inhaler. Such inhalers, 137 when pressed, generate sprays with consistent ejection pressure and duration. 33, 34 We loaded 138 the nozzle of the inhaler with a suspension of 100 nm-diameter fluorescent nanoparticles (beads) 139 in distilled water. The fluorescent beads mimic SARS-CoV-2 virus (70-100 nm-diameter) 35, 36 in 140 terms of size, and allow to quantify the efficiency of the fabric samples at reducing the 141 transmission of 100 nm-size particles carried by water. When the inhaler is pressed, the internal 142 pressure of the inhaler ejects the bead suspension out of the nozzle, creating high-speed droplets 143 (Video S1). The droplets then hit the fabric sample that is placed in front of the inhaler ( Figure 144 3B). 145 146 We recorded high-speed videos of the ejected droplets and performed image analysis to estimate 147 droplet size and velocity (see Methods for details). We detected droplets with diameters in the 148 ~0.1 mm to ~1 mm range within the spray ejected by the inhaler ( Figure S1 ). Droplets released 149 by sneezing, coughing, and speaking typically have size distributions within the range of ~1 µm 150 to ~1 mm. [20] [21] [22] [23] Our experimental platform thus offers a way to test blocking efficiency against 151 large droplets which fall within the size range of droplets released by respiratory events, thereby 152 complementing previous studies on aerosol filtration efficiency which considered particles 153 smaller than 10 µm in size. [14] [15] [16] [17] [18] [19] Analysis of high-speed videos also revealed that droplets exit the 154 inhaler with a median velocity of 17.1 m/s, measured within 25 mm (~1 inch) from the nozzle, 155 and gradually slow down as they travel through the air. Median droplet velocity reduces to 2.7 156 m/s after they travel 300 mm (~12 inch) from the inhaler nozzle ( Figure 3C ). Droplets released by 157 coughing and sneezing typically have velocities within the ranges of ~10 to ~20 m/s, 23 High-speed video recordings verify that high-momentum droplets can penetrate the medical 167 mask and T-shirt fabric (Fabric 6 in Table 1 ) with 1 or 2 layers ( Figure 3D , Videos S2-S4). Droplets 168 penetrating a single layer of T-shirt fabric had a median exit velocity of 9.6 m/s. In contrast, 169 droplets penetrating a medical mask and 2 layers of T-shirt fabric had median exit velocities of 170 2.2 m/s and 3.0 m/s, respectively ( Figure 3C ). In all three of these cases, exit velocities of the 171 droplets that penetrated the fabric barriers were significantly lower than the incident velocity of 172 17.1 m/s (p < 0.001 for all, two-sample t-test). 173 174 To collect the droplets that were transmitted through the fabrics, we placed a petri dish behind 175 the fabric samples ( Figure 3A ,B). Similarly, petri dish without a fabric barrier was used at the same 176 location to collect incident droplets for comparison. Figure 3E shows brightfield and fluorescence 177 images of droplets that were collected in the petri dish without a fabric barrier. Since fluorescent 178 beads are uniformly dispersed in the solution that is loaded into the inhaler nozzle, we can use 179 the number of fluorescent beads collected with and without a fabric barrier to measure the 180 droplet blocking efficiency. However, the droplet images in Figure 3E cannot be used to count 181 the beads because many beads are clustered together and cannot be counted separately. To 182 solve this problem, we developed a method to disperse and homogenize the beads collected in 183 the petri dish. We achieved homogenization by first collecting the droplets on petri dishes coated 184 with a layer of gelatin hydrogel, then melting the gelatin at 37°C, mixing the beads with liquid 185 gelatin, and re-gelling. We used a confocal microscope to image the gelatin hydrogels containing 186 fluorescent beads ( Figure 3F ). For each test, 100 to 150 images were taken from separate 187 locations of the hydrogel and the average number of beads per image, ̅, was computed by image 188 analysis. We carried out control measurements to validate our method and found that ̅ can be 189 used as an accurate predictor of the bead density in the gelatin mixture within the range of 1 < 190 ̅ < 1000 ( Figure S2 , see Methods for details of bead mixture homogenization, confocal imaging, 191 analysis, and validation). 192 193 To measure droplet blocking efficiency, we need to compute a ratio of the total number of beads 194 transmitted through the fabric and collected in the gelatin-coated dish, ℎ , to the total 195 number of beads collected without a fabric barrier, ℎ . To compute this ratio, we also 196 took into account the total volume of gelatin solution, , that the fluorescent beads were 197 dispersed in for each test. Since the average number of beads per image, ̅, correlates with bead 198 density (i.e., number of beads per unit volume), we use = ̅ · as a predictor of the total 199 number of beads collected in the petri dish. Droplet blocking efficiency, , of the fabric is thus 200 computed as: 201 202 203 We performed independent measurements of ̅ without any fabric barrier (15 separate 204 measurements, 100-150 images each), as well as with the medical mask, all 11 household fabrics 205 with a single layer, and fabric 2 and fabric 6 (see Table 1 ) with 2 and 3 layers (6 or 7 separate 206 sample measurements for each fabric, 100-150 images per measurement). We then computed 207 the droplet blocking efficiency, , using Equation 1, for all possible combinations of 208 ( ̅ ⋅ ) ℎ and ( ̅ ⋅ ) ℎ . Distributions of droplet blocking efficiency values for 209 all tests are shown in Figure 4 . The minimum, median, and maximum values are provided in Table 210 2. 211 212 The tests at 25 mm from the inhaler nozzle represent the case of mask users releasing droplets 213 with high-velocity onto the mask by sneezing or coughing. At this range, we found that most 214 fabrics with a single layer have relatively high droplet blocking efficiencies (median values > 70%) 215 ( Figure 4A , Table 2 ). The median droplet blocking efficiency was 98.5% (minimum 96.4%) for the 216 medical mask. For single layers of fabrics 2 and 6, median droplet blocking efficiency was 81.9% 217 (minimum 41.1%) and 83.1% (minimum 42.0%), respectively. The addition of second and third 218 layers increased the efficiency, with median values exceeding 94% (Table 2) . Note, however, that 219 for 3 layers of fabric 6, the average number of beads per image was below the detection limit 220 ( ̅ < 1). We assume that the droplet blocking efficiency of 3 layers of fabric will be at least as 221 high as the efficiency of 2 layers (median 98.1% for fabric 6) , and hence report the median 222 efficiency for 3 layers of fabric 6 as >98.1% (Table 2 ). For all other tests at 25 mm, ̅ was within 223 the accurate detection range (1 < ̅ < 1000). 224 225 Next, we measured droplet blocking efficiency for low-momentum droplets (300 mm away from 226 the inhaler nozzle), which represents the case of users releasing droplets with low velocity onto 227 the mask during speaking. Droplets released during respiratory events decrease in velocity and 228 size as they travel through air. 6, 7 Hence, our low-momentum droplet challenge can also represent 229 the case of a mask user receiving droplets released by a nearby individual. We measured for 230 the medical mask and 1 and 2 layers of fabric 6 at a distance of 300 mm from the inhaler nozzle 231 (6 separate sample measurements for medical mask, and fabric 6 with 1 and 2 layers). At this 232 distance, high-speed imaging of the droplets shows a median velocity of 2.7 m/s ( Figure 3C , Video 233 S5). The droplet sizes at this distance are also smaller ( Figure S1 ). Thus, the impact of the droplets 234 has decreased. At 300 mm, median values of for the medical mask and single layer of fabric 6 235 were 99.7% and 94.2%, respectively ( Figure 4B , Table 2 ), i.e., at 300 mm is higher compared to 236 that at 25mm. For 2 layers of fabric 6 at 300mm, the average number of beads per image was 237 below the detection limit ( ̅ < 1); Hence, as above, we report the median efficiency as >94.2% 238 (i.e., efficiency of 2 layers at least as high as the efficiency of a single layer). 239 240 Fabric Breathability 241 242 We define breathability, , of a fabric as, = ⁄ , where is a change in the flow rate of 243 air through unit area of fabric, and is the corresponding change in the pressure differential 244 across the sample that is required to induce . We used a plug flow tube 42, 43 to measure . 245 Here, the sample fabric seals the opening of a plug flow tube ( Figure 5A ,B). Pressurized air was 246 pumped through the tube. Pressure outside the tube is atmospheric and the pressure inside the 247 tube was measured with respect to the atmosphere. Thus, air was forced through the fabric by 248 the gauge pressure, . Since the area of the fabric sample subjected to air flow is the same as the 249 cross-sectional area of the plug flow tube, the change in flow rate through unit area of fabric, , 250 is equivalent to the change in average flow velocity, inside the tube. Hence, breathability can 251 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 J o u r n a l P r e -p r o o f Journal Pre-proof be written as = ⁄ . Note that this expression of breathability or air permeability is general 252 and allows linear as well as non-linear velocity-pressure relationships. Linearity is a special case 253 (Darcy's Law) at low Reynolds number flow regimes. However, deviations from linearity may still 254 arise while flow is laminar. 44 We did not assume linearity a priori and we measured as a function 255 of to assess the relationship. 256 257 In a short plug tube, pressure is uniform and flow velocity is approximately uniform across any 258 cross section. 43, 45 Also, pressure changes negligibly along the length of the short tube (small 259 pressure gradient). Most of the pressure drop occurs across the sample fabric. We measured 260 and at the same cross-section at mid length of the tube. Figure 5C shows velocity vs. pressure 261 measured at various pump speeds for a single and double layered T-shirt (fabric 6) and the 262 medical mask. These measurements were made for the medical mask, single layers of all 11 263 household fabrics, as well as 2 and 3 layers of fabric 2 and fabric 6 (3 independent samples tested 264 for each case). In all cases, velocity vs. pressure showed a highly linear relationship, i.e., following 265 Darcy's Law. Hence, we computed breathability, , as the slope of the least-squares fit line to the 266 velocity-pressure data. Results are shown Figure 5D and Table 2 . 267 268 We found that, for household fabrics, breathability depends strongly on porosity (correlation 269 coefficient, r = 0.929 for woven fabrics, r = 0.894 for knit fabrics). Although the correlation 270 between breathability and porosity was high for both woven and knit fabrics, the slopes of the 271 regression lines were different. For the same porosity, knit fabrics had higher breathability than 272 woven fabrics ( Figure 5E ). This is due to the porosity being measured under static conditions, 273 whereas breathability measurement involved applying air pressure across the fabrics. In the 274 latter case, fabrics can stretch due to air pressure. Knit fabrics are overall more stretchable than 275 woven fabrics, and therefore have higher breathability for a given static porosity. 276 277 Furthermore, in aerosol science, air permeability of a fibrous filter usually anti-correlates with 278 aerosol filtration efficiency, i.e., the higher the air permeability, the lower the filtration 279 efficiency. 46 To assess whether this relationship also holds for blocking of large droplets, we 280 plotted droplet blocking efficiency against breathability, and found that they are indeed anti-281 correlated (r = -0.820, Figure 5F ). For example, a single layer of T-shirt (fabric 6) had high 282 breathability (7.23 ±0.55 mm/Pa•s) and relatively low blocking efficiency (median 83.1%, 283 minimum 42.0%). The addition of a second layer increased droplet blocking efficiency (median 284 98.1%, minimum 94%) while reducing breathability (to 3.87 ±0.06 mm/Pa•s, see Table 2 ). 285 286 A corollary of the porosity-breathability and efficiency-breathability relationships shown in Figure 287 5E and 5F is that droplet blocking efficiency should also anti-correlate with porosity. Indeed, we 288 found this to be the case (efficiency vs. porosity: r = -0.796 for woven fabrics, r = -0.822 for knit 289 fabrics). Similar to the porosity-breathability relationship, the slopes of the regression lines in 290 efficiency vs. porosity plots were different for woven and knit fabrics ( Figure S3 ). For the same 291 porosity (measured in static conditions), knit fabrics tend to have lower droplet blocking 292 efficiency than woven fabrics. High-speed recordings show that fabric samples can deform and 293 stretch as a result of the impact from high-velocity droplets (see Video S3). Since knit fabrics are 294 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 The data presented so far demonstrates that most home fabrics with one layer can block both 300 high and low impact droplets reasonably well. With 2 or 3 layers, their blocking efficiency 301 becomes comparable to that of medical masks while still having similar or higher breathability. 302 However, the materials of the medical mask and that of the home fabrics are very different. How 303 do the home fabrics achieve their blocking efficiency? While porosity plays a role, as discussed 304 above, we also observed differences between the medical mask material and home fabrics in 305 terms of wetting and water soaking behavior. Commercially manufactured medical masks ( Figure 306 2A,B) use 3 layers of hydrophobic fabric (non-woven plastic material, e.g., polypropylene) with a 307 high contact angle ( Figure 6A , Video S6 To better understand the underlying mechanism of droplet blocking by hydrophilic home fabrics, 312 we recorded high-speed videos of the incident droplets from the inhaler and subsequent 313 transmission through the medical mask, as well as 1 and 2 layers of T-shirt (fabric 6) ( Figure 6B , 314 Videos S2-S4). In all cases, the samples were attached to a 40 mm-diameter wire ring which was 315 placed 25 mm from the inhaler nozzle. Image analysis reveals that the large droplets impact and 316 push on the fabric (medical mask or T-shirt fabric) with a median velocity of 17.1 m/s. They also 317 stretch the fabrics. A few droplets squeeze through small pores of the fabrics and split into 318 smaller droplets as they exit within few milliseconds ( Figure 6B , Videos S2-S4). Thus, the droplets 319 deform during transmission through the pores of the fabric against surface tension ( Figure 1B ). 320 The kinetic energy of the incident droplets is spent deforming the fabric (bending and stretching), 321 overcoming surface tension barriers and splitting into smaller droplets ( Figures 1B, 6A) . Energy 322 needed to overcome surface tension barriers will be higher for hydrophobic fabrics compared to 323 hydrophilic fabrics ( Figure 1B ). 324 325 Transmission through fabric reduces kinetic energy of the droplets. The median exit velocity of 326 droplets penetrating a single layer of T-shirt fabric was 9.6 m/s, significantly lower than the 327 incident velocity of 17.1 m/s ( Figure 3C , p < 0.001, two-sample t-test). These transmitted droplets 328 therefore have much lower momentum with which they would impact a second layer, if present. 329 Indeed, high-speed imaging reveals very few droplets exiting the 2-layered T-shirt fabric sample 330 ( Figure 6B, Video S4 ). This reduction of momentum of large droplets by first layer followed by 331 the trapping of the lower-momentum droplets by the second layer may explain the high blocking 332 efficiency of 2-layered T-shirt fabric sample (median 98.1%, minimum 94.0%). Thus, energy 333 dissipation appears to be a key mechanism that contributes to the high droplet blocking 334 efficiency of multi-layered home fabrics. After completion of the impact and partial transmission, 335 the T-shirt fabric may soak and retain the remaining droplet volume, whereas hydrophobic 336 medical mask does not absorb any liquid (Figures 1 and 6A) . Large droplets may just roll down 337 the medical mask by gravity. 338 The above scenarios involve high-speed impact and represent droplets released by users 340 coughing and sneezing onto the mask. Droplets released by speaking onto a mask, as well as 341 droplets received on the outside surface of the mask, have low momentum. Hence they may not 342 have enough energy to overcome surface tension barriers and squeeze through the fabric. This 343 may explain our measurements of higher blocking efficiency against low-momentum droplets 344 compared to those with high-momentum (for medical mask and 1-2 layers of T-shirt Fabric, see 345 Table 2 , Figure 4 ). 346 347 Home Test of Water Permeability 348 349 The techniques we developed and used in this study to measure breathability and droplet 350 blocking efficiency involved specialized laboratory equipment not available to the general public. 351 We therefore sought a simple method that could be used at home to help compare different 352 fabric choices. Here, we describe a simple test of water permeability and assess correlations 353 between its result and breathability and droplet blocking efficiency ( Figure S4 ). 354 355 We pre-wetted the fabrics with water (medical mask was not used here since it did not wet), 356 placed them onto the mouth of a bottle and tied in place with a rubber band. The bottle 357 contained a cup (~250 ml) of water and had a small hole punched at mid-height to equilibrate 358 pressure inside the bottle to atmospheric pressure. Next, we gently flipped the bottle to allow 359 water to drain through the wetted fabric ( Figure S4A ). We used a stopwatch to record the time 360 that it takes water to drain, , stopping when water stops streaming and begins to drip. 361 Water drains faster through fabrics with higher water permeability. Thus, the inverse of the 362 draining time, 1⁄ , offers a measure of water permeability. We found that 1⁄ 363 correlates strongly with breathability, which is a measure of air permeability (r = 0.943, Figure 364 S4B). This strong correlation implies that both measures (water and air permeability) dependent 365 on the same fabric properties, among which porosity is most likely a dominant one. Given that 366 breathability and droplet blocking efficiency were anti-correlated ( Figure 5F ), we expect that 367 water permeability should also anti-correlate with droplet blocking efficiency. Indeed, a 368 comparison of median droplet blocking efficiency and 1⁄ revealed this to be the case (r = 369 -0.828, Figure S4C ). 370 371 The simplicity and availability of this water draining test -requiring only a water bottle, rubber 372 band, and stopwatch -provides a means by which individuals at home can assess their choices 373 of household fabrics for masks. The result of this test, 1⁄ , correlates (or anti-correlates) 374 reasonably strongly with breathability and droplet blocking efficiency that were measured in a 375 laboratory setting ( Figure S4B,C) Regarding the practical use of home-made face coverings, the properties of home fabrics have 385 several important implications aside from breathability and droplet blocking efficiency. 386 Microscale and nanoscale properties, such as microstructure and surface chemistry, of the fabrics 387 contribute to their wetting behavior, which can be modified by engineering these properties. 47 Home fabrics can be washed and reused, as opposed to medical masks that are disposable. 389 Medical masks are typically made of non-biodegradable plastics such as polypropylene. Their 390 widespread use during a pandemic can therefore pose an environmental burden. The ability to 391 wash and reuse home-made face coverings offers an advantage in terms of reducing waste and 392 pollution. Washing home-made face covering is also necessary for decontamination. As we have 393 discussed above, most home fabrics are hydrophilic (water soaking). In contrast to the highly 394 hydrophobic medical mask material, home fabrics can soak and hold the droplets. While this 395 holding ability of hydrophilic fabrics may accrue possible benefits in terms of droplet blocking 396 performance of home-made masks, it also means that the fabrics can retain the viruses that were 397 in the soaked droplets. Home-made masks must therefore be washed regularly to 398 decontaminate, either by laundry machine at warmest temperature setting, or by hand using 399 water and household disinfectants such as bleach. 13 400 401 The results of droplet blocking efficiency we have presented here complement the existing 402 knowledge from studies on the aerosol filtration efficiency of household fabrics. [14] [15] [16] [17] [18] [19] The method 403 commonly employed in these studies, as well as in the standard testing of commercial facemasks 404 and respirators, involves challenging the sample by a pressure-driven airflow containing aerosols 405 with particle sizes usually within the ~10 nm to ~10 µm range, and using a particle counter to 406 measure and compare the number of particles upstream and downstream of the fabric. 48 A 407 recent study on the performance of common household fabrics against aerosols with particle 408 sizes within 10 nm to 6 µm range reported several findings that align with ours: Fabrics with lower 409 porosity (tighter weave) performed better than those with higher porosity. Multiple layers of 410 fabric were highly effective (> 90% filtration efficiency in most cases). 19 In our experiments, the 411 fabrics were challenged with large droplets (up to ~1 mm in size), significantly larger than the 412 particles used in aerosol filtration studies. As our results indicate, household fabrics can also be 413 effective at blocking such large droplets, especially with multiple layers. 414 415 Blocking of both small aerosol particles and large droplets by fabrics can be considered within 416 the general theoretical framework of advective-diffusive transport through a porous barrier. The 417 performance of the barrier can depend on which mode of transport dominates. In our 418 experiments, the motion of droplets towards the fabric samples can be described as projectile 419 motion with relatively high velocity. Similarly, in the previous aerosol filtration studies, aerosol 420 particles travel towards the fabric with momentum imparted upon them by pressure-driven bulk 421 airflow. Thus, in both cases, advective transport plays a dominant role, as opposed to diffuse 422 transport which may occur, for instance, due to osmotic pressure. In most practical situations of 423 facemask use, advective transport is dominant. Droplets or aerosols released by a mask wearer 424 during sneezing, coughing, or speaking will impact the mask with high velocity. In situations 425 where a facemask is intended for blocking incoming droplets or aerosol particles, advective 426 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 J o u r n a l P r e -p r o o f Journal Pre-proof motion is also expected due to ambient airflow or flow generated by inhalation. We therefore 427 expect our results, and the findings of previous aerosol filtration studies, to be applicable in these 428 scenarios. Situations where diffusive transport dominates may arise in confined spaces with high 429 concentration of airborne particles but with minimal ambient airflow, thereby creating osmotic 430 pressure across the mask. Future studies are required to assess the performance of facemasks in 431 such situations involving predominantly diffusive transport. 432 433 While fabric porosity appears to play a role in the blocking of both aerosols and larger droplets, 434 the mechanisms of penetration may be different. Aerosols with particles in the nanometer to 435 micrometer range may pass through the fabric pores without interacting with the fabric (Figure 436 1A). Thus, lower porosity (as a fraction of the total fabric area) will result in higher filtration 437 efficiency. With multiple layers of fabric, the net effective porosity is likely to decrease 438 significantly due to misalignment of the pores of the separate layers. This can significantly 439 increase filtration efficiency against small particles. Droplets, on the other hand, can squeeze and 440 flow through the pores. Lower porosity can still be advantageous since squeezing through smaller 441 pores is more energy consuming due to higher shear stresses and surface tension barrier 442 originating from the interfacial energies of air, water and the fabric material ( Figure 1B ). As we 443 have discussed above, household fabrics can vary in their degree of hydrophilicity and water 444 absorption properties. A more hydrophilic fabric may impose less activation barrier against 445 droplets squeezing through the pores compared to that of a hydrophobic fabric. A multi-layered 446 fabric barrier will likely misalign the pores of the separate layers, decreasing the effective 447 porosity, and significantly increasing the blocking efficiency. 448 449 The droplet blocking efficiency of the fabrics reported here is based on the count of 100 nm-450 diameter nanoparticles transported through the fabric, ℎ , compared to the total 451 number of incident particles, ℎ , both carried by water droplets. Thus, the efficiency 452 is really a measure of the ability of the fabrics to block the nanoparticles carried by water 453 droplets. On the other hand, SARS-CoV-2 (approximately 100 nm in size 35, 36 ) are carried by saliva 454 droplets released during respiratory events. 49, 50 Saliva typically has much higher viscosity than 455 water, even under high shear rates. 51, 52 Transmission of large droplets of saliva would therefore 456 require higher energy to squeeze through fabric pores due to higher shear stresses. Hence, the 457 blocking efficiency values that we measured, using water droplets, provide a conservative 458 estimate of the efficiency of fabrics to block transmission of viruses carried by saliva droplets. 459 460 We would like to stress that the efficiency values we report here are indicative of the ability of 461 the fabric materials to block 100 nm-sized particles carried by droplets, and are not meant to 462 reflect the net performance of a cloth facemask. The performance of a facemask would also 463 depend on how it is worn, and how much air leaks through the gaps between the mask and face 464 contours. Net efficiency of the facemask can be significantly lower than the efficiency of the fabric 465 material itself, due to leakage through gaps. 32 Cloth face coverings could therefore be made more 466 effective by ensuring proper sealing against the face contour. For example, a recent study 467 reported that adding a layer of nylon stocking over the masks minimized the air flow around the 468 edges of the masks and improved particle filtration efficiency for both home-made and 469 commercial masks. 53 As an alternative measure to ensure a better fit, an elastomeric net or half-470 mask can be used over the mask. 54 471 472 Although the relative contributions of aerosols and larger droplets to the spread of COVID-19 are 473 not fully understood, our results, taken together with those of aerosol filtration studies, suggest 474 that home-made face coverings may have considerable efficacy in blocking both. Furthermore, a 475 recent modeling study suggests that extensive adoption of even relatively ineffective face masks 476 may meaningfully reduce community transmission of COVID-19 and decrease peak 477 hospitalizations and deaths. 55 Mask use reduces transmission rate in a nearly linear proportion 478 to the product of mask effectiveness (as a fraction of potentially infectious contacts blocked) and 479 coverage rate (as a fraction of the general population), while the impact on epidemiologic 480 outcomes (death, hospitalizations) is highly nonlinear. 55 As a result, it is anticipated that face 481 coverings made from household fabrics can play a vital role as a mitigating strategy. 482 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 In this study, we asked whether face coverings made from home fabrics can be effective against 485 the dissemination of droplets carrying 100 nm size infectious viruses, such as SARS-CoV-2, and if 486 so, will their droplet blocking efficiency be comparable to that of a commercial medical mask. 487 We studied a diverse set of 11 common household fabrics with varying fiber types and 488 constructions. We quantified their breathability, and their ability to block 100 nm-diameter 489 nanoparticles carried by high-velocity droplets similar to those that may be released by sneezing 490 or coughing. We found that all of these fabrics have considerable efficiency at blocking high-491 velocity droplets, even as a single layer. With 2 or 3 layers, even highly permeable fabrics, such 492 as T-shirt cloth, achieve droplet blocking efficiency that is similar to that of a medical mask, while 493 still maintaining comparable breathability. 494 495 For low-velocity droplets, which mimic droplets released during speaking, we found that blocking 496 efficiency of T-shirt fabric is much higher compared to that for high-velocity droplets. A scenario 497 involving low-velocity droplets may also arise when a mask user receives droplets released by an 498 infected individual nearby. It thus follows that a 2 or 3-layered home-made mask with most 499 common fabrics may help prevent the dissemination of droplets by infected individuals, and 500 protect healthy individuals from inhaling droplets, with efficiencies similar to that of commercial 501 medical masks. 502 503 Considering our results within the context of recent literature on aerosol filtration efficiency of 504 home fabrics and epidemiological modeling of the potential impacts of mask use, we conclude 505 that during pandemics and mask shortages, home-made face coverings with multiple layers can 506 be effective against transmission of respiratory infection through droplets. Mask wearing by all 507 individuals, supported by proper education and training of mask making and appropriate usage, 508 can be an effective strategy in conjunction with social distancing, testing and contact tracing, and 509 other interventions to reduce disease transmission. 510 Fabric Characterization 513 To quantify fabric weight (mass per unit area), rectangular samples of known area were cut from 514 each fabric and medical mask using a paper trimmer and their masses were measured using a 515 high precision lab scale. Fabric construction (woven, knit, or napped) was determined by visual 516 inspection. Fiber contents were noted from the cloth labels. To measure thread count, close-up 517 views of the fabric samples (see Figure 2 ) were imaged using a digital camera (MS100 USB 518 Microscope, Teslong), and the thread count was calculated as the sum of the number of threads 519 per unit length in weft and warp directions. Thread count measurement was not applicable to 520 the medical mask and dishcloth due to their non-woven and napped fabric construction, 521 respectively. To measure fabric porosity, a digital image analysis method was used (see Figure 522 S5), adapted from previous studies. 56, 57 Samples of woven and knit fabrics were imaged while 523 being illuminated from the backside by diffused light. Light that passed through the fabric was 524 used to identify the pores. Backside illumination intensity was kept constant and 9 samples were 525 imaged for each fabric. Images were then analyzed in MATLAB as follows: Images were converted 526 to greyscale. A median filter was applied to reduce noise. Filtered images were converted to 527 binary by applying a threshold at 90% intensity; hence, pores were identified as pixels with 528 intensity values in the top 10 th percentile. Binarized images were analyzed to calculate porosity 529 as the ratio of the number of bright pixels to the total number of pixels in the image ( Figure S5 ). 530 For the medical mask and dishcloth, this method did not produce images with sufficient contrast; 531 therefore porosity measurement for these fabrics was omitted. To quantify the water absorption 532 behavior of fabrics ( Figure S4 ), a small droplet (100 µl) of water was dispensed on dry fabric 533 samples, and video of the soaking process was recorded using a digital camera (MS100 USB 534 Microscope, Teslong). Food coloring was added to the water to provide contrast and allow us to 535 visualize the soaked area. Images were converted to binary and analyzed to measure the soaked 536 fabric area as a function of time. Water soaking speed was then computed as the time derivative 537 of the soaked area. We note that there are a wide variety of standardized tests available to 538 characterize fabrics, listed as ISO 9073, for industrial applications. 539 540 Droplet Generation and Characterization 541 We used a metered-dose inhaler (HFA-propelled, 210 sprays, GlaxoSmithKline) and loaded its 542 nozzle with 10 µl of distilled water to generate droplets. To measure droplet velocity, we 543 recorded videos of the ejected droplets at 10,000 frames per second (fps) using a high-speed 544 camera (FASTCAM-ultima APX, Photron). Videos of 4 separate sprays were recorded and 545 analyzed in ImageJ. Droplet positions were tracked manually across consecutive frames and 546 velocity was calculated by dividing the distance travelled by the elapsed time. To estimate ejected 547 droplet size, we used a separate imaging setup (4MP 2560×1600 CMOS camera, Phantom) which 548 offered higher spatial resolution with a lower frame rate (400 fps). Droplets were illuminated 549 with a laser (50 mJ Terra PIV dual cavity YLF laser, Amplitude) for clear visualization. Images of 550 ejected droplets from 3 separate sprays were analyzed using ImageJ and MATLAB. Captured 551 images were converted to binary and droplets were identified as objects (8-connected 552 components) in the binary image. Droplet diameter was calculated as the 'equivalent diameter' 553 of a circle with the same area as that of the imaged droplet. Using this imaging and analysis 554 method, we were able to detect droplets with diameters greater than ~0.1 mm ( Figure S1 ). For 555 measurement of droplet size at 300 mm from the inhaler nozzle (low momentum droplets), the 556 droplets were collected on a polystyrene dish placed perpendicular to the spray direction. The 557 dish was then imaged immediately using brightfield microscopy. ImageJ and MATLAB were used 558 as before to convert images to binary and measure the equivalent diameter of the landed 559 droplets. Since water contact angle on polystyrene is about 87°, 58 we approximated the landed 560 droplets as half-spheres, and calculated the diameters of corresponding incoming droplets using 561 volume conservation ( Figure S1 ). 562 563 Droplet Challenge Tests 564 To challenge the fabric samples with droplets, we placed the inhaler at mid-height of an acrylic 565 channel open at both ends. The channel prevents air flow within the room from interfering with 566 the tests. Droplets were generated using a suspension of 100 nm-diameter red fluorescent beads 567 (ex/em 580/605 nm, Invitrogen, catalog #F8801) diluted in distilled water. For testing the fabric 568 samples, we first coated the bottom of a petri dish with 1 ml of warm gelatin solution (5% wt/v) 569 which was prepared by dissolving powdered gelatin (Sigma Aldrich, catalog #G9391) in distilled 570 water. Gelatin forms a hydrogel at and below room temperature and melts at higher 571 temperatures. 59 We let the gelatin solution gel inside the petri dishes at 4°C, then covered the 572 dishes by attaching the fabric cut-outs to the rim of the dish using double-sided tape. We placed 573 samples inside the acrylic channel, at mid-height, 25 mm or 300 mm away from the inhaler 574 nozzle, to challenge with high or low-velocity droplets, respectively. In each test, after the inhaler 575 was pressed and droplets containing fluorescent beads impacted the fabric samples, droplets 576 that penetrated the fabric samples were collected on the gelatin layer. Similarly, droplets were 577 collected on separate gelatin-covered petri dishes without any fabric barrier, as control. Next, we 578 warmed the samples to 37°C in an incubator to liquefy the gelatin layer and allow the beads to 579 dissolve into the gelatin solution. This mixture was transferred to a vial, vortexed for 20 s, and 580 then sonicated for at least 30 min to uniformly disperse the fluorescent beads in solution. This 581 homogenized bead-gelatin mixture was re-gelled at 4°C. The beads were thus frozen in place in 582 the hydrogel. 583 584 Fluorescent Bead Counting and Validation 585 We imaged the gels containing fluorescent beads on a confocal laser scanning microscope (LSM 586 710, Zeiss) using a 40X water-immersion lens (NA = 1.2). For each sample, we picked five random 587 fields of view and took z-stacks with 10 µm spacing (to ensure the same set of beads were not 588 imaged twice) and 20 to 30 slices, resulting in 100 to 150 images per sample. The bead 589 distribution was reasonably uniform in plane and across the gel thickness. Confocal microscopy 590 images were analyzed in MATLAB. Images were converted to binary by applying a threshold at 591 50% intensity. Beads were identified as objects (8-connected components) in binary images, and 592 the average number of beads per image, ̅, was calculated from the 100-150 images for each 593 separate test. To validate that ̅ can be used as an accurate predictor of the bead density in the 594 gelatin mixture, we prepared gelatin mixtures with known bead densities by serial dilution of the 595 original bead solution (known bead density provided by vendor). We then gelled these solutions, 596 performed confocal imaging, and computed ̅, averaged over 100 images for each density tested. 597 For samples with 1000 ≳ ̅ ≳ 1, the average number of beads per image, ̅, correlated very 598 strongly with the known bead density in the mixture. Bead densities corresponding to ̅ ≳ 1000 599 were not tested. For ̅ < 1, there was significant deviation from the regression model ( Figure 600 S2). Thus, for all droplet blocking efficiency measurements, we took ̅ = 1000 and ̅ = 1 as our 601 upper and lower detection limits, respectively. 602 603 Breathability Measurement 604 Our plug flow apparatus ( Figure 5B ) consists of an acrylic tube with 100 mm inner diameter, a 605 pump on one end, and a set of long aluminum tubes aligned within the acrylic tube on the other 606 end which help stabilize the flow. 42, 43, 45 Flow through the tube can be controlled by varying the 607 pump speed with an analog dial. We measured the air flow velocity, , and the gauge pressure, 608 , at the same cross-section at mid-length of the tube ( was measured at the center of the cross-609 section). A pressure gauge (Magnehelic) and a hot-wire anemometer (Omega) were used to 610 measure pressure and velocity, respectively. A small hole in the acrylic tube was used to insert 611 the pressure probe and the anemometer and the hole was sealed with clay dough. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 J o u r n a l P r e -p r o o f 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 J o u r n a l P r e -p r o o f **For these tests, the average number of fluorescent beads per image was below the detection limit. Hence, the expected value based on the median efficiency of one fewer layer of the same fabric is reported . 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 J o u r n a l P r e -p r o o f Figure 1 . Distinction between aerosol filtration and large droplet blocking by fabrics. (A) Typical mechanisms of particle capture and transport during aerosol filtration: Particles 1, 2, and 3 are captured by the fiber via interception, impaction, and diffusion, respectively. Particle 4 is smaller than the interfiber spacing and is transmitted through the fabric, carried by air flow. Particle 5, being larger than the inter-fiber spacing, is captured by straining. Particle 6 is subsequently captured by settling/caking. (B) Blocking of nanoparticles carried by large droplets. Top and bottom rows represent transmission through hydrophilic and hydrophobic fabrics, respectively. Droplets impact the fabric with high velocity, squeeze through the pores, and part of the volume can transmit. This process involves energy costs associated with interfacial energies and shear stresses, which may be influenced by fabric porosity, fabric type, and viscosity of the droplet. Energy barriers for transmission increase with decreasing pore size, increasing droplet viscosity, as well as hydrophobicity of the fabric. For example, interfacial energy barrier for transmission through hydrophobic fabric is much higher than that for hydrophilic one. 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 at 300 mm, measurements were below detection limit, therefore, only a line corresponding to the estimated lower bound is presented for these samples. 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 Toward Understanding the Risk of Secondary Airborne Infection: Emission of Respirable Pathogens Routes of Influenza Transmission Review of Aerosol Transmission of Influenza A Virus The Airborne Lifetime of Small Speech Droplets and Their Potential Importance in SARS-CoV-2 Transmission US Centers for Disease Control and Prevention | How COVID-19 Spreads Violent Expiratory Events : On Coughing and Sneezing Human Cough as a Two-Stage Jet and Its Role in Particle Transport Clinical Characteristics of 24 Asymptomatic Infections with COVID-19 Screened among Close Contacts in Nanjing SARS-CoV-2 Viral Load in Upper Respiratory Specimens of Infected Patients Transmission of 2019-NCoV Infection from an Asymptomatic Contact in Germany Asymptomatic Cases in a Family Cluster with SARS-CoV-2 Infection Respiratory Virus Shedding in Exhaled Breath and Efficacy of Face Masks US Centers for Disease Control and Prevention | Use of Cloth Face Coverings to Help Slow the Spread of COVID-19 Use of Cloth Masks in the Practice of Infection Control -Evidence and Policy Gaps Professional and Home-Made Face Masks Reduce Exposure to Respiratory Infections among the General Population Evaluating the Efficacy of Cloth Facemasks in Reducing Particulate Matter Exposure Testing the Efficacy of Homemade Masks: Would They Protect in an Influenza Pandemic? Simple Respiratory Protection-Evaluation of the Filtration Performance of Cloth Masks and Common Fabric Materials Against Aerosol Filtration Efficiency of Common Fabrics Used in Respiratory Cloth Masks The Size and the Duration of Air-Carriage of Respiratory Droplets and Droplet-Nuclei Droplet Expulsion from the Respiratory Tract Characterizations of Particle Size Distribution of the Droplets Exhaled by Sneeze Characterization of Expiration Air Jets and Droplet Size Distributions Immediately at the Mouth Opening World Health Organization | Advice on the use of masks in the context of COVID-19 Filtration of Aerosols by Fibrous Media Theory of Aerosol Filtration Filtration of Aerosols Theoretical Study of Aerosol Filtration by Fibrous Filters Experimental Study of Aerosol Filtration by Fibrous Filters Aerosol Filtration by Fibrous Filters -I Aerosol Filtration by Fibrous Filters -II Performance of an N95 Filtering Facepiece Particulate Respirator and a Surgical Mask During Human Breathing: Two Pathways for Particle Penetration The Physics of Aerosol Droplet and Particle Generation from Inhalers Instant Velocity and Consistency of Emitted Cloud Change by the Different Levels of Canister Filling with Metered Dose Inhalers (MDIs), but Not with Soft Mist Inhalers (SMIs): A Bench Study SARS-CoV-2 (COVID-19) by the Numbers Identification of Coronavirus Isolated from a Patient in Korea with COVID-19 Study on the Initial Velocity Distribution of Exhaled Air from Coughing and Speaking Study on Transport Characteristics of Saliva Droplets Produced by Coughing in a Calm Indoor Environment Droplet Infection of Air: High-Speed Photography of Droplet Production by Sneezing Public Health Applications of High-Speed Photography Visualization of Sneeze Ejecta: Steps of Fluid Fragmentation Leading to Respiratory Droplets Optimal Control of Free-Stream Turbulence Intensity by Means of Honeycombs. Fluid Dyn Experiments on Management of Free-Stream Turbulence Transition from Creeping via Viscous-Inertial to Turbulent Flow in Fixed Beds Comparison of Experimental and Theoretical Turbulence Reduction from Screens, Honeycomb, and Honeycomb-Screen Combinations Relation of Structure, Properties and Performance of Fibrous Media for Gas Filtration Superhydrophobic Textiles: Review of Theoretical Definitions, Fabrication and Functional Evaluation A Comparison of Facemask and Respirator Filtration Test Methods Saliva -Friend and Foe in the COVID-19 Outbreak Saliva-Based Molecular Testing for SARS-CoV-2 That Bypasses RNA Extraction The Rheological Properties of Saliva A Study of the Viscosity of Saliva at Different Shear Rates in Dentate and Edentulous Patients Quantitative Method for Comparative Assessment of Particle Removal Efficiency of Fabric Masks as Alternatives to Standard Surgical Masks for PPE Effect of the Tight Fitting Net on Fit Performance in Single-Use Filtering Facepieces for Koreans To Mask or Not to Mask: Modeling the Potential for Face Mask Use by the General Public to Curtail the COVID-19 Pandemic Determination of Pore Size, Porosity and Pore Size Distribution of Woven Structures by Image Analysis Techniques Porosity Determination of Jersey Structure Contact Angle of Water on Polystyrene Thin Films: Effects of CO2 Environment and Film Thickness Effects of Concentration, Bloom Degree, and PH on Gelatin Melting and Gelling Temperatures Using Small Amplitude Oscillatory Rheology