Assessment of mask efficiency for preventing transmission of airborne illness through aerosols and water vapor

Study design

The study was performed at the Seattle Children’s Research Institute, Seattle, WA from April 2020- Oct. 2020. Studies in vitro were carefully designed to evaluate protective attributes based on the efficiency and permeability between simple cloth face masks, commonly worn in the community, and those applied in hospital settings. We also measured efficiency with the addition of a filter to the cloth mask.

Initially, we evaluated efficiency on prevention of inhaled ‘fine’ particles (0.03–0.05 µm) by assessing mask permeability of water vapor generated by a humidifier. These water vapor particles, also capable of transmitting airborne viruses, are generated in the exhaled breath of infected individuals while talking and present a significant risk for infection. We then characterized albuterol particle size generated by a nebulizer with a particle impactor. Lastly, we applied the nebulizer to a spontaneously breathing 3D printed airway affixed to an adult lung model and quantified particle deposition to the masks and filters (when applicable) and at the level of the face, airway model, and lung model filter. For this study, we considered nebulized aerosols as “coarse” particles within the range of particle sizes known to generate infectious airborne droplets generated through coughing or when performing aerosol generating procedures (AGPs) by caregivers in clinical settings.

Masks and filters

Figure 1 A–H show the different mask and filters used with each of the experimental conditions. The cotton or “cloth” facemasks, unisex fashion stretch lightweight cotton covering face and mouth (Getien Biotech, Inc; Nanjing, China), were chosen as they had elastic ear loops, a filter pocket and nose wire that allowed for fit adjustment (Figure 1 C, D, E). An N95 respirator (Figure 1A) (3M™; St. Paul, Minneapolis) and surgical mask (Figure 1B) (Wujin Economic Development Zone; Changzhou, Jiangsu) are commonly used in the hospital setting for respiratory precautions were also used.

Figure 1. Masks and filters tested.

The different masks used in all of the experimental conditions are shown, with N95 (A), surgical mask (B), cloth mask posterior (C), and anterior (D), with filter pocket (E). Filters used with cloth masks included PM2.5(F) MERV 15 (G) and Swiffer (H).

Prior to testing, masks and mask filters were preconditioned using a vacuum warmer (Labconco 7982010 CentriVap Aqueous System, 115V, 60Hz) at 40°C for 30 minutes. Three commercially available filter materials were used including PM2.5 filters (Getien Biotech, Inc; Nanjing, China) (Figure 1F), MERV 15 (Capital air filter; Raleigh, NC) (Figure 1G), and unscented Swiffer™ Sweeper Dry™ sweeping cloth (Procter & Gamble; Cincinnati, OH) (Figure 1H). Both the MERV 15 filter and the Swiffer™ dry cloth were cut to the same dimensions as the PM2.5 filters (12cm x 8cm) (see Figure 1F, G, H).

Conditions

Both phases of testing were conducted by evaluating 5 separate runs at each of the following conditions: 1) no mask; 2) cloth mask; 3) cloth mask with Swiffer™ filter; 4) cloth mask with MERV 15 filter; 4) cloth mask with PM2.5 filter and 5) surgical mask. Owing to a critical shortage of N95 masks in our region, at the time of testing, we were only able to test 3 each of the N95 respirator. For the fine particle testing we repeated testing on two of the masks after drying for a total of 5 runs (N=5). These same three masks were used for course particle testing for a total of 3 runs (n=3).

Fine particle permeability

The first phase of testing evaluated mask efficiency based on ‘permeability’ of fine water vapor particles. A 3D printed face and nasopharyngeal model was used. This airway was replicated from the Computed Tomography (CT) scans of a 17-year-old male’s head (printed with PLA black material on Crealty CR 10 printer with embedded nasa airway printed with FormLabs Clear V4 resin on the Form 2 printer). Spontaneous breathing was simulated using a Harvard Large Animal Ventilator (Model #613, Harvard Apparatus; Holliston, Massachusetts). The model simulated normal breathing through different mask conditions (Figure 2A).

Figure 2. Experimental set up.

(A) The experimental set-up of the Harvard Apparatus use to emulate spontaneous breathing via 3D- printed realistic adult airway model. (B) Assessment of fine particle efficiency was characterized by measuring time to reach 95% humidity levels acquired from within the simulated trachea during spontaneous breathing. A valved, aerosol mask allowed egress of exhaled water vapor particles to atmosphere to prevent saturation of masks/filters and limit measurement to incoming water vapor particles. (C) Direct application of aerosol applied to the face with a cloth mask. Aerosol was eluted from upper airways, and lung model filter placed in filter housing following nebulization of 10 mg albuterol and quantified using spectrophotometry.

The adult test lung was configured with pre-set tidal volume of 700 mL, respiratory rate of 20 breaths/min, and inspiratory: expiratory (I:E) ratio of 1:1 with half sinusoidal inspiratory and expiratory flow profile to mimic normal spontaneous breathing. (Li et al., 2019) All values were monitored prior to testing with the TSI 5200 analyzer. A traceable hygrometer (Model #4185, Fisher Scientific, Pittsburg, Penn) was placed within the simulated trachea in series between the 3D printed airway and the lung model to measure relative tracheal humidity (RH). Increases in tracheal RH level beyond the ambient value was indicative of ultra-fine water vapor permeating beyond the different masks and filters. A pediatric Monaghan AeroChamber Plus Z STAT valved mask (Monaghan Medical; Plattsburgh, NY) incorporates a series of low resistance one-way valves that allows decoupling of inspiratory and expiratory flow paths. This prevented saturation of the hygrometer but also limited changes in RH and penetration of incoming fine particles to inhalation without rebreathing exhaled particles. Following preconditioning of the mask and filter as applicable, the valved, aerosol mask was secured on the 3D model and masks using rubber bands. A Fisher and Paykel (MR 850, Fisher & Paykel; Panmure, Auckland) heated humidifier was preset at 37°C in the invasive mode and applied to 20 L/min air via a heated-wire circuit (see Figure 2B). RH value of 99–100% was confirmed at the AeroChamber mask opening using a hygrometer. Following ambient RH measurement, humidified air was administered directly to the airway with each mask/filter. The masks efficiency in filtering fine airborne particles was determined based on time (seconds) to reach 95% relative humidity within the trachea. A new mask/filter was used for each of the 5 runs with the exception of the N95 respirators. Three of the respirators were pre-conditioned again in the dryer and re-tested for a total of 5 runs (n=5).

Course particle deposition

The first step for the aerosol testing was to characterize the course particles size emitted from a nebulizer. Coarse aerosol particles were generated with a vibrating mesh (VM) nebulizer (Aerogen Solo, Aerogen Ltd., Galway, Ireland), with a residual drug volume < 0.1 mL. (Li et al., 2019) A high-performance, multistage, next generation impactor (Model #NGI-0670, NGI, TSI Incorporated, Shoreview, MN) was used to quantify the distribution of coarse aerosol droplet size produced by the VM nebulizer and classify into respirable size fractions. The NGI uses eight different particle trays to collect aerosol droplets. The vacuum flow was adjusted with a frit resistor (S/N 511197-9, Cole Palmer, Vernon Hills, IL) at 15 L/min and confirmed with a TSI 5200 Flow and Pressure Analyzer (TSI Industries, Shoreview, MN). A standard leak test was performed per manufacturer’s specifications before each test. 10 mg of aqueous albuterol was aerosolized in continuous output mode. Following nebulization, aerosol droplets were eluted by instilling 10mL of hydrochloric acid in each plate and then withdrawing the solution from each individual stage. The mass was quantified using ultraviolet-visible spectrometry (UV-Vis) (Molecular Devices SpectraMax M3; San Jose, CA). The albuterol sulfate was measured by UV-Vis spectrometry at its local spectral maximum of approximately 276 nm. Using the Spectra Max M3 (Molecular Devices; San Jose, CA). Samples were pipetted into 96 well plates with UV transparent flat bottom, (Thermo Scientific, or equivalent). Albuterol sulfate powder (USP Reference Standard, Cat #1012633) was used to make standard concentrations for comparison of unknown samples. A calibration curve was created using 500 µ/ml, 200 µ/ml, 100 µ/ml, 50 µ/ml, 25 µ/ml, and 12.5 µ/ml known concentrations of albuterol. Three readings from each of these know concentration samples were plotted and analyzed by the Spectra Max M3 software. The standard curve was then be used to determine the concentrations of test samples from the absorptance measurements. This was repeated for a total of 5 runs (n=5). We then calculated median mass aerodynamic diameter (MMAD), geometric standard deviation (GSD) using a log-probit equation, fine particle fraction (FPF, %) based on proportional mass of respirable particles (MMAD <5.4 µm) and total drug mass delivered to the impactor stages.

Following the characterization of aerosols emitted from the nebulizers, we evaluated filtration efficiency of coarse aerosols between the different mask and no mask conditions using a similar experimental set-up as the previous study and is shown in Figure 2A. A low dead-space filter housing containing a low resistance medication filter (PARI Filter; Starnberg, Germany), was placed between the lung model and simulated trachea to collect inspired aerosol just prior to entering the lung model. 10 mg of aqueous albuterol ulfate (5mg/mL) was administered through the VM nebulizer with an operational flow of compressed at 3 L/min. The nebulizer was attached directly to the 3D printed face model or overprotective masks using a vented aerosol mask (see Figure 2C).

Following nebulization, the lung filter was removed and the 3D printed face was wiped with a PARI filter. The 3D printed model airway was eluted with 20mL of hydrochloric acid (0.1 N). PARI filters, mask and mask filters (when applicable) were placed in a sealing plastic bag and eluted with 20 mL of HCL. They were manually agitated for 3 minutes. All eluents were filtered through a 5-micron filter needle to remove debris. Recovered samples were then quantified for absolute albuterol mass (µg) using UV-Vis spectrophotometry (Molecular Devices SpectraMax M3; San Jose, CA). The albuterol sulfate was measured by UV-vis spectrophotometry at its local spectral maximum of approximately 276 nm. Using the Spectra Max M3 (Molecular Devices; San Jose, CA). Samples were be pipetted into 96 well plates with UV transparent flat bottom, (Thermo Scientific, or equivalent). Albuterol sulfate powder (USP Reference Standard, Cat #1012633) was used to make standard concentrations for comparison of unknown samples. A calibration curve was created using 500 µ/ml, 200 µ/ml, 100 µ/ml, 50 µ/ml, 25 µ/ml, and 12.5 µ/ml known concentrations of albuterol. Three readings from each of these know concentration samples were plotted and analyzed by the Spectra Max M3 software. The standard curve was then be used to determine the concentrations of test samples from the absorptance measurements. A new mask (with or without filter) was used for a total of 5 runs (n=5). Each of the 3 N95 respirators were tested once for a total of 3 runs (n=3).

For these testing purposes, we reported coarse aerosol as the cumulative deposited aerosol dose (%) because patients are most likely to become infected by touching the face or when infectious droplets are inhaled directly into the naso-oral airway airways, bronchial tubes or lungs. As such, cumulative deposited aerosol dose (%) included the sum of albuterol mass deposited on face, airway, and lung as referenced to the nominal dose placed into the nebulizer. We also included descriptive measurements for mask and filter deposition.

Analysis

Fine particle permeability based on time to 95% humidity was calculated in seconds and assessed for normality using histograms and QQ-plots and summarized using means and standard deviations for each mask/filter type. ANOVAs were used to compare the cumulative deposited dose (%) as well as the mean amount of time, in seconds, until the mask/filter reached 95% humidity between each mask/filter. The sum of cumulative albuterol particles deposited and recovered from face, airway, and lung was calculated and assessed for normality using histograms and QQ-plots. The cumulative mean value was referenced to the nominal dose placed into the nebulizer (10 mg) and at each individual location (face, airway, lung, filter, and mask) and expressed as means and standard deviations for cumulative deposited aerosol dose (%) for each condition. A Tukey’s post-hoc analysis was used to assess specific differences between each mask/filter type. A p-value < 0.05 was considered statistically significant. SAS 9.4 (Cary, NC) was used for all analyses. R is an open access alternative that could be used for the analysis. The deposited dose (%) at each individual location (face, airway, lung, filter, and mask) was not tested for differences but were included as descriptive data.

Sample size

Estimates from previous pilot runs were used to estimate a standard deviation of 92 for total drug deposition. With 7 conditions of 5 runs each, this estimates a standard deviation of means of 64. An overall difference of 183 between mask/filter types was able to be detected using an alpha of 0.05 and 80% power. Post-hoc analyses could detect a difference of 232 between each mask/filter type with 80% power using a Bonferroni correction and an alpha of 0.007.

Fine particle permeability

The results from the fine particle testing can be found in Figure 3. Based on time to 95% RH, permeability with the PM2.5, MERV 15, Swiffer™, and N95 was lower than no mask (Ringer, 2021). The cloth mask and surgical mask alone did not demonstrate significant (P<0.05) filtration of water vapor when compared to no mask condition. The cloth mask with Swiffer™ filter performed as well as the N95 respirator with filtration of the fine water vapor particles.

Figure 3. Permeability of fine water vapor particles.

Data shown represent the time to 95% relative humidity (seconds), based on tracheal relative humidity levels, following exposure of the 3D airway to humidification with and without different mask options. Data are mean ± SD (standard deviation) with n=5 runs with each configuration. Data were analyzed with one-way ANOVA (analysis of variance), with Tukey test post hoc comparisons between testing conditions. Means with the same letter are not significantly different, P<0.05.

Course particle size characterization

Nebulized coarse aerosol albuterol particle sizes ranged from 0.54 to 11.72 µm with MMAD of 4.87 ± 2.18 µm (Figure 4). The particle size fractions (%) were based on the total mass (µg) of albuterol recovered from the individual impactor stages with 43, 27, and 30% for particles >5.4, 3.2- 5.4 and <3.2 µm, respectively (n=5).

Figure 4. Particle size characterization of nebulized coarse aerosols.

The particles generated by the Aerogen. Aeroneb (n=5) were characterized using an aerosol impactor prior to testing with masks. Particle sizes ranged from 0.54 to 11.72 µm with median mass aerodynamic diameter (MMAD) 4.87±2.18 µm. The particle size fractions (%) were based on the total mass (µg) of albuterol recovered from the individual impactor stages with 43, 27, and 30% particle for particles >5.4, 3.2- 5.4 and <3.2 µm, respectively (n=5).

Course particle deposition

Data shown in Figure 5 are the sum of total cumulative mass of albuterol particles recovered from the face, airway, and lung with and without different masks. These data were then referenced to the nominal albuterol dose (10 mg) placed into the nebulizer and shown as delivered dose %. Data are mean ± SD with n=5 runs with each configuration and n=3 for N95. Data were analyzed with one-way ANOVA, with Tukey test post hoc comparisons between testing conditions. Testing conditions with the same letter on Figure 5 have means that are not significantly different, P<0.05.

Figure 5. Cumulative deposited dose of coarse aerosol particles.

Data shown are the sum of total cumulative mass of albuterol particles recovered from the face, airway, and lung with and without different masks. These data were then referenced to the nominal albuterol dose (10 mg) placed into the nebulizer and shown as delivered dose %. Data are mean ± SD (standard deviation) with n=5 runs with each configuration and n=3 for N95. Data were analyzed with one-way ANOVA (analysis of variance), with Tukey test post hoc comparisons between testing conditions. Means with the same letter are not significantly different, P<0.05.

There were differences in the cumulative deposited dose (%) between different testing conditions (P<0.05). The no mask condition was the least efficient at preventing coarse particles from depositing on the face, airways, and lungs when compared to all of the different mask testing conditions. Although the cloth mask alone showed greater efficiency with a nearly 5-fold reduction in coarse particle transmission than the no mask condition, the addition of a filter (Swiffer™, PM2.5, or MERV 15) demonstrated greater filtration efficiency based on cumulative deposited dose (%). Surgical masks showed lower cumulative deposited dose than did no mask, cloth only, and cloth with Swiffer™ filter. The N95, prevented particle transmission with lower deposition than all the testing conditions except surgical mask and cloth mask with MERV 15 filter where performance was not different. Figure 6 shows descriptive data on the deposition of aerosol droplets measured in filters, masks, face, airways and lungs. The drug mostly deposited in the mask/N95 respirator with negligible delivery to the face, airway, and lung filter.

Figure 6. Location of course particle deposition.

Data shown represent the mass of albuterol particles recovered from the face, airway, and lung with and without different masks. These data were then referenced to the nominal albuterol dose (5 mg) placed into the nebulizer as delivered dose %. Data are mean ± SD (standard deviation) with n=5 runs with each configuration and n=3 for N95. These descriptive data were not analyzed for statistical differences.

Underlying data

Figshare: Assessment of Mask Efficiency for Preventing Transmission of Airborne Illness Through Aerosols and Water Vapor. https://doi.org/10.6084/m9.figshare.14758497 (Ringer, 2021).

This project contains the following underlying data:

Data are available under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0).