Surgical masks decontamination for reuse by members of the public: feasibility study and development of home-based methods

This study aims to propose decontamination methods that are suitable for use by members of the public to cope with the shortage of surgical masks during the current COVID-19 pandemic. 3-ply surgical masks were subjected to different decontamination protocols (heat, chemical, ultraviolet irradiation) to assess their abilities to achieve at least 4-log reduction of two common respiratory pathogens, H1N1 Inuenza A virus, a single-stranded RNA enveloped virus similar to SARS-CoV-2 and Staphylococcus aureus, a Gram-positive bacterium that is more resistant to decontamination than single stranded RNA enveloped virus. Decontaminated surgical masks were assessed for differences in breathability, particle ltration eciency and bacteria ltration eciency as compared with non-decontaminated masks. The number of decontamination cycles that the 3-ply surgical masks could undergo without signicant changes in breathability and ltration eciencies were also determined. It was found that surgical masks decontaminated by either soaking for 60 min in 0.5% (v/v) aqueous hydrogen peroxide solution, or 30 min of soaking in 0.05% - 0.5% (v/v) aqueous sodium hypochlorite diluted from household bleach or ultraviolet irradiation by a surface dose of 13.5 kJ/m 2 were able to achieve at least a 4-log reduction of both Staphylococcus aureus and H1N1 Inuenza A virus spiked on surgical mask test swatches. No signicant changes in breathability and ltration eciencies of the surgical masks were observed after ten decontamination cycles of hydrogen peroxide or diluted bleach treatment or 30 cycles of ultraviolet irradiation.


Introduction
During outbreaks of acute infectious respiratory diseases such as the current coronavirus disease 2019 , transmission could occur through inhalation of infectious droplets expelled by  patients before the onset of symptoms [1] as well as asymptomatic infected individuals. [2] Studies have shown that viral copies expelled into the environment during exhalation are reduced by wearing a mask. [3] This has led to a widespread public health recommendation to wear a mask in public to limit the spread of COVID-19. The competing need of masks by both medical and frontline workers, as well as the public, exacerbated by worldwide restrictions on the export of medical supplies, could result in temporal shortage in supply of surgical masks when it is most needed. [4] Developing methods to ensure safe reuse of surgical masks by the public would help to reserve dwindling stock of fresh surgical masks for healthcare workers during such critical moments.
To allow for safe reuse of surgical masks that were designed for single use, it is essential to develop e cient and facile methods to remove microbial contaminants persisting on used masks without damaging the ltration properties of the mask. In particular, studies have indicated that ~0.1% of the original inoculum of infectious coronavirus persists on the outer layer of a surgical mask seven day after contamination. [5] To date, while there are many studies reporting on e cacy of various decontamination methods on N95 surgical masks, [6]- [12] there is a paucity of decontamination studies to allow re-use of contaminated surgical masks by healthy individuals in public to abate the spread of COVID-19.
Expanding on the research performed on reusing N95 ltering facepiece respirators, the goal of our study was to identify and propose accessible methods readily adoptable by the general public for decontaminating used surgical masks to allow for its safe re-use (single user reuse).
The decontamination methods were selected from similar studies performed on N95 respirators, [6]- [12] as well as recommendations from the United States Centers for Disease Control and Prevention (CDC) [13] and World Health Organization (WHO) for inactivating SARS-CoV-2 virus. [14] Decontamination methods such as application of dry heat at 70 o C for ve minutes or heating at high humidity levels, [10] chemical methods such as 70% ethanol, [10] bleach [5] and vaporized hydrogen peroxide, [10] as well as UV-C irradiation [10] were reported to be effective for inactivation of SARS-CoV-2. Since the target group is the public, the decontamination methods selected in this study explored techniques that could be readily carried out using household commodities and appliances. The methods included in our studies were steaming, immersion in hot water, boiling, soaking in different household disinfectants such as diluted bleach and aqueous hydrogen peroxide solution, 70% (v/v) ethanol and liquid detergent. Irradiation using ultraviolet (UV-C) light was also investigated.

Results
Differential pressure (delta-P), Particulate ltration e ciency (PFE) and Bacteria ltration e ciency (BFE) after decontamination A preliminary assessment on the feasibility of reusing surgical masks was carried out by comparing the PFEs before and after wearing the masks without prior decontamination. Wearing the masks for 8 h or more did not result in any signi cant change in the PFE 0.1µm and the trend in the PFE 0.05-0.6µm remained unaffected. This indicated that the PFE was retained in the worn mask and thus mask reuse could be feasible if the inactivation protocol of microbes does not alter the mask ltration properties.
Masks were subjected to heat, chemical and UV-C decontamination as listed in Table 1. Visual assessments performed on the decontaminated masks revealed no visible signs of damage. The melt blown lter layer of the untreated and treated masks were found to have no differences when viewed under the SEM. No chlorine off-gassing was detected on masks treated with bleach (Details are described in the supplementary information). The masks attained a metallic smell after UV-C decontamination, but this smell dissipated after airing for 15 minutes.
Effective decontamination methods should preserve the breathability of the masks. The delta-P of masks treated by all the methods in Table 1 were below 5.0 mmH 2 O/cm 2 , which indicated acceptable levels of breathability.
Prolonged duration of heating was found to degrade the lter layer. The PFE of masks treated by 30 mins boiling was signi cantly different from untreated masks (P-value=.0004). The PFEs 0.1µm of surgical mask treated by steaming, hot water and boiling decreased as the duration of applied wet heat increased from as shown by a decrease in PFE observed across this range of entire aerosol sizes ( Figure 1). Even though it has been reported that 70 o C for 5 minutes is adequate for inactivation of SARS-CoV-2, 5 such precise temperature measurement and control is di cult to achieve in most households. Thus, heat inactivation of microbes may not be a practical repeated decontamination approach for households.
In addition, the surgical masks treated with 70% ethanol and detergent had noticeable change in the ltration e ciency. Although there was no signi cant difference in PFEs 0.1µm for masks treated by either 70% ethanol (P-value=.08) or dishwashing detergent (P-value=.04), the BFE of both detergent (P-value=.002) and 70% ethanol (P-value=.007) were signi cantly reduced from the value of untreated masks. Furthermore, PFEs 0.05-0.6µm for these methods were observed to reduce by more than 40% from the untreated masks. These two methods were thus not pursued in subsequent investigations.  Table 1 and Figure   1). These methods of soaking in diluted bleach or hydrogen peroxide solution or UV-C exposure were also found to have minimal effects on the breathing resistance of the surgical mask. Hence, these methods were selected for antimicrobial studies to assess their decontamination effects.

Antimicrobial e cacy test
Antimicrobial studies could not be done on SARS-CoV-2 virus because BSL3 laboratory was inaccessible to the team. Instead, In uenza A subtype H1N1 virus, which is also a lipid enveloped single-stranded RNA virus and bacterium S. aureus were used in our investigations as they are common pathogenic microbes that causes respiratory diseases in the community. In addition, S. aureus, which is a type of vegetative bacteria that has been reported to be more resistant to inactivation by chemical and UV-C decontamination as compared to lipid enveloped viruses. [15] By choosing these two microbes, the decontamination method would be applicable to the public for the current pandemic COVID-19.
All 3 decontamination methods, namely soaking in 0.5% (v/v) hydrogen peroxide for 60 minutes or soaking in 0.05% (v/v) sodium hypochlorite (1% v/v bleach) for 30 minutes or UV-C irradiation using surface dose of 13.5 kJ/m 2 of UV-C, were found to achieve at least 4-log reduction for S. aureus bacteria and H1N1 virus ( Table 2).
Determination of number of decontamination cycles possible for each method Repeated cycles of the three decontamination methods were carried out to ascertain the maximum number of decontaminations that the 3-ply surgical mask could be subjected to without signi cant loss of barrier function or increased breathability. Our results showed that the breathing resistance (delta-P), PFEs and BFE of the masks did not change after being subjected to any of these 3 decontamination methods, namely ten cycles of hydrogen peroxide (0.5% v/v; 60 min) decontamination, or ten cycles of bleach (0.05% v/v and 0.5% v/v sodium hypochlorite; 30 min) decontamination, or 30 cycles of 13.5 kJ/m 2 UV-C irradiation (Table 1).

Discussion
Studies by other researchers have demonstrated that a 5 minutes decontamination process with household bleach diluted up to 100 times was su cient to kill SARS-CoV-2 (concentration of sodium hypochlorite was not reported). 5 While we did not evaluated the anti-viral e ciency of diluted bleach for decontaminating surgical mask contaminated with SARS-CoV-2, the conditions evaluated in our bleach decontamination experiments (0.5% and 0.05% sodium hypochlorite, representing 10 and 100 times dilution of household bleach respectively) should be su cient to inactivate SARS-CoV-2 contaminants on the mask.
No study has been reported on the inactivation of SARS-CoV-2 using liquid hydrogen peroxide. A past study demonstrated that a concentration of 2.1% of hydrogen peroxide applied over 10 minutes (concentration x time = 21 %(v/v).min) could achieve a 4-log reduction of murine norovirus (a nonenveloped, single-stranded RNA virus) on surfaces. [17] As non-enveloped virus are reported to be more resistant to chemical disinfectants than lipid enveloped virus, [15] a concentration x time value of 30 % (v/v).min used in our hydrogen peroxide decontamination studies (pH=5, 25 to 35 °C) is expected to achieve similar 4-log reduction of SARS-CoV-2 single-stranded RNA virus.
For UV-C light decontamination, we observed that irradiation on the surface of the mask achieved 5.49log reduction for S. aureus and 4.26-log reduction for H1N1 by applying a UV-C surface radiance exposure dose of 13.5 kJ/m 2 . Attenuation of the light through the mask resulted in 9.2% of the UV-C light reaching the inner side of the masks to give a dose of 1.3 kJ/m 2 without ipping the mask during the irradiation process. To date, there is a paucity of peer-reviewed studies on the UV-C susceptibility of SARS-CoV-2 with the few reported studies pending peer review. [10], [19], [20] Comparing these initial UV-C disinfection studies against SARS-CoV-2 virus with peer-reviewed data on UV-C susceptibilities of SARS-CoV-1 and other enveloped viruses with single-stranded RNA, [6], [18], [21] a large variation in UV-C dose required for inactivation of coronaviruses (0.06 kJ/m 2 for the bovine coronavirus to 117.54 kJ/m 2 for SARS-CoV-1, Urbani strain was observed. [21] These extreme variations were reported to arise from variations in experimental conditions in these studies. [21] Excluding variations caused by experimental conditions, coronaviruses were reported to be very sensitive to UV-C light, with an upper dose limit of 0.1 kJ/m 2 required for 1-log inactivation. The corresponding dose required for 4-log inactivation of coronavirus was 0.42 kJ/m 2 , which is three times lower than the measured UV-C radiance exposure transmitted through a typical 3-ply surgical mask (~1.3 kJ/m 2 ). As SARS-CoV-2 virus does not differ structurally in any great extent from other known coronaviruses, they are expected to be similarly sensitive to the germicidal properties of UV-C radiation to produce 4-log reduction of these viruses trapped on the surface of a 3-ply surgical mask More investigation will be needed to ascertain the UV-C dose to inactivate SARS-CoV-2 trapped within 3ply surgical mask since factors such as material, attenuation of the mask, or presence of interference like mucin and saliva from the user has to be taken into considerations.

Materials
The surgical mask used was a 3-ply ear loop model with a melt-blown polypropylene middle layer (SM-3EP-50G, Arista Sensi, Indonesia The UV-C treatment was carried out in a self-built UV cabinet, tted with two 30 W low pressure germicidal lamp (G30T8, Sankyo Denki, Japan) with wavelength centred at 254 nm. A radiometer with a UV-C probe (Delta OHM H2101.2 and LP471UVC probe, Italy) was used for measuring the UVC power density between 220 to 280 nm within the cabinet. Viral ToxGlo™ Assay (Promega, USA) was used for the antiviral assays.

Inactivation methods
The treatment methods investigated include 1) heat treatment by boiling, immersing in hot water and steaming, 2) chemical treatment using hydrogen peroxide, bleach, ethanol and liquid detergent and 3) UV-C treatment. Unless or otherwise stated, the experiments were conducted with the full surgical mask.

Heat treatment
The procedure for heat treatment by hot water was as follows. Water was heated using an electric kettle. The water was then transferred to a heat-insulated bowl immediately. The mask was fully submerged into the hot water (80 to 90 o C), with weights to hold down the mask. The solution was covered to minimise cooling and left to stand for 10 and 30 min respectively. The temperature of water was found to be 40 to 45 o C after 30 min.

Page 7/15
The procedure for steaming was as follows. A metal steamer was placed inside a pot and water was lled to just below the bottom of the steamer. The water was brought to a boil. The mask was then placed on the metal steamer with the inner layer facing downwards. The masks were steamed for 10 and 30 min with the lid covered.
Boiling treatment was conducted as follows. Water was brought to a boil in a pot, after which, the mask was submerged inside the boiling water for 10 and 30 min, with the lid of the pot covered.
After each respective heat treatment, the mask was left to dry at room temperature.

Chemical treatment
Procedures for bleach treatment were as follows. Store-bought bleach was diluted to give 0.05% sodium hypochlorite (bleach: water=1:99) and the mask was submerged fully for 30 min. Similar procedure was carried out for 0.5% sodium hypochlorite solution (bleach: water = 1:9).
Steps for treatment by surfactant-based soap was as follows. Store-bought dishwashing detergent was diluted with water ( nal volume: 200 mL, nal concentration: 5 mg/mL) and the mask was submerged fully for 1 minute.
Steps for hydrogen peroxide treatment was as follows. Mask was submerged fully in 0.3% hydrogen peroxide solution for 30 min. Similar procedure was carried out with 0.5% hydrogen peroxide for 1 hour.
Steps for ethanol treatment was as follows. Mask was submerged fully in 70% ethanol (200 mL) for 10 min before removal.
After each respective chemical treatment, the mask was rinsed with tap water to remove the soaking chemicals and left to dry at room temperature.

UV-C treatment
Pre-determined locations with receiving power density of 10-15 W/m 2 was marked in the UV cabinet. The masks were placed at these locations and irradiated for 15 min. The total mask was subjected to a surface dose of 9 to 13.5 kJ/m 2 during each cycle of irradiation. The amount of light that reached the bottom layer was approximately 9.2 ± 0.3 % of the total irradiated light, which is equivalent to a dose of 1.3 kJ/m 2 .
Off-gassing measurement of chlorine Measurement of possible off-gassing of chlorine were performed at two stages of the decontamination process: 1) immediately after rinsing of the surgical mask with tap water and 2) after drying of the surgical mask. A chlorine detector with a detection range of 0 to 1.5 ppm with increment units of 0.01 ppm was used (Riken Keiki, SC-8000, Japan). No chlorine was detected when the air sampling probe was placed beside the mask at both stages.
Particle Filtration E ciency (PFE) and Differential pressure (Delta-P) measurements The PFE was evaluated based on a modi ed test method stated in ASTM-F2299 standards. The aerosol ltration test rig system constructed using stainless steel was set up based on Scheme S1. The main deviation from the standard was using a smaller mask coupon of 9.6 cm 2 and ltration e ciency was determined by measuring the aerosol concentration before and after insertion of the mask coupon using a single downstream aerosol isokinetic sampling probe. Concentration of PSL was measured by a condensation particle counter (TSI Model 3775) with the Differential Mobility Analyser (TSI Model 3081) electromobility diameter set at 0.11 µm. The DEHS concentration was measured in the scanning mobility particle sizer mode from 0.05 to 0.6 µm.
Differential pressure across the mask coupons was measured using a micromanometer (TSI Model 5815).
Bacterial Filtration E ciency (BFE) BFE was conducted by an external vendor (TUV-SUV-PBS Pte Ltd). ASTM-F2101 test methods for evaluating BFE of medical face mask materials using biological aerosol of S. aureus was followed. The mean particle size of the challenge aerosol was 3 0.3 µm over a ow rate of 28.3 0.3 L/min over a test area of approximately 38.5 cm 2 . The results of the untreated and treated mask samples were compared against negative control ( ltered air) and positive control stream of S. aureus.
Bacterial and Viral culture S. aureus (ATCC 25923), a Gram-positive bacterium found commonly in the upper respiratory tract, was used. It was cultured in tryptic soy broth (TSB) at 37 o C, overnight with shaking. S. aureus was diluted to 1 x 10 9 colony-forming units CFU/mL using 1:19, TSB in 0.05% Triton X-100 as the inoculating medium.
In uenza A/Puerto Rico/8/1934 (H1N1) was grown in embryonated chicken egg. The virus was harvested 3 days post-infection from the allantoic uid.

Inoculation of biological agents on test swatch
For anti-bacterial studies, 2 x10 8 CFU of S. aureus was inoculated onto square swatches (16 cm 2 ) prepared from the surgical masks resulting. For anti-viral studies, 2 x 10 6 Plague Forming Unit (PFU) H1N1 in virus growth medium was inoculated onto square swatches (1 cm 2 ). The inoculated swatches were allowed to dry at room temperature in a biosafety cabinet for 60 min, before proceeding with the treatment studies.
Treatment of the test swatch for anti-bacterial and anti-viral assays Brie y, for bleach treatment, the inoculated swatches were submerged fully in sodium hypochlorite solution (0.05% and 0.5%) for 30 min, followed by rinsing with deionised water thrice.
For hydrogen peroxide treatment, the inoculated swatches were submerged fully in H 2 O 2 (0.3% and 0.5%) for 60 min, followed by rinsing with deionised water thrice. The volume to surface area ratio was kept consistent at 0.4 mL solution (hydrogen peroxide or bleach) per centimetre square of test swatch.
For UV-C treatment, the inoculated swatches were placed at the pre-determined locations in the UV cabinet with the outer layer facing the UV lamp described in the earlier section.

Anti-bacterial assays
After treatment, the swatches were transferred to centrifuge tubes, and 10 mL of extraction buffer (TSB) was added, followed by vortexing for 1 minute. The supernatant was serially diluted, and plated in duplicates on tryptic soy agar (TSA) plates. The TSA plates were incubated for 24 hours before the CFUs on the plates were enumerated to determine the bacterial Log10 reduction for each treatment method, compared to the positive control (bacterial count extracted from inoculated swatches that were not inactivated).
Anti-viral assays  Figure 1 Plot of PFEs for different decontamination methods with different DEHS particle sizes of 0.05 to 0.6 µm. Data shown for heat decontamination are for samples exposed to 30 min of heat decontamination. The blue region indicates the prediction interval of the ltration e ciency of the clean mask at 90% con dence level. H2O2 = hydrogen peroxide, NaOCl = sodium hypochlorite found in bleach

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.