Supercritical CO2 sterilization: An effective treatment to reprocess FFP3 face masks and to reduce waste during COVID-19 pandemic

The outbreak of COVID-19 pandemic unveiled an unprecedented scarcity of personal protective equipment (PPE) available in sanitary premises and for the population worldwide. This situation fostered the development of new strategies to reuse PPE that would ensure sterility and, simultaneously, preserve the filtering properties of the materials. In addition, the reuse of PPEs by reprocessing could reduce the environmental impact of the massive single-use and disposal of these materials. Conventional sterilization techniques such as steam or dry heat, ethylene oxide, and gamma irradiation may alter the functional properties of the PPEs and/or leave toxic residues. Supercritical CO2 (scCO2)-based sterilization is herein proposed as a safe, sustainable, and rapid sterilization method for contaminated face masks while preserving their performance. The functional (bacterial filtration efficiency, breathability, splash resistance, straps elasticity) properties of the processed FFP3 face masks were evaluated after 1 and 10 cycles of sterilization. Log-6 sterilization reduction levels were obtained for face masks contaminated with Bacillus pumilus endospores at mild operating conditions (CO2 at 39 °C and 100 bar for 30 min) and with low contents of H2O2 (150 ppm). Physicochemical properties of the FFP3 face masks remained unchanged after reprocessing and differences in efficacy were not observed neither in the filtration tests, following UNE-EN 14683, nor in the integrity of FFP3 filtration after the sterilization process. The herein presented method based on scCO2 technology is the first reported protocol achieving the reprocessing of FFP3 masks up to 10 cycles while preserving their functional properties.


H I G H L I G H T S
• Shortages of PPE could be reduced by a safe and effective sterilization method. • A 30-min sterilization with scCO 2 is effective and preserves functional integrity of PPE. • This process can be potentially used in hospital environments in a context of urgency. • Performance of PPEs was not remarkably compromised even after 10 sterilization cycles. • Reuse of PPE could be achieved using scCO 2 resulting in a reduction of waste.

Introduction
The pandemic caused by SARS-CoV-2 virus unveiled the limited contingency of the healthcare systems towards outbreak situations worldwide. Namely, the demand for personal protective equipment (PPE) dramatically surpassed the global production capacity of manufacturers (Rowan and Laffey, 2020). Subsequently, the general scarcity of different PPEs, particularly face masks, resulted in healthcare professionals being at the frontline of the COVID-19 outbreak without the suitable protective "shields" to mitigate the risk of becoming infected by the virus (Rowan and Laffey, 2021;Heinzerling et al., 2020). In the advent of the COVID-19 outbreak situation, the number of clinicians infected by SARS-CoV-2 represented ca. 15-20% of the total number of cases in several developed countries like Spain (Torrente et al., 2021). A dramatic increase in PPE production capacity by the manufacturer lines has momentarily alleviated the unmet demand, but the quest for providing tools in hospitals and other sanitary premises to prepare healthcare workers and individuals for future PPE scarcity episodes is still unsolved.
Different disinfection and sterilization methods have been tested with PPEs as reprocessing attempts to reuse them, including technologies already available at the hospital environment. These methods can be broadly categorized into heat (dry and moist), chemical (ethylene oxide) or radiation (gamma, X-ray and UV) treatments (Cadnum et al., 2020;Bennet et al., 2021;Wilson and Nayak, 2019). The limited number of possible sterilization cycles (e.g., the use of vaporized hydrogen peroxide sterilization causes changes in masks after the fourth cycle (RIVM, n.d.); alterations in masks are observed after the tenth cycle using UV-irradiation (Liao, 2020)), the modifications of the filtration efficacy after reprocessing (e.g., UV-exposure led to a certain increase in particle penetration (Lindsley et al., 2015); the use of steam altered the filtration efficiency after 5 cycles (Liao, 2020); gamma-ray caused significant reduction in the filtering capacity (de Man et al., 2020)) and the problems of adjustment of the respirators to the face (irradiation, heat (Viscusi et al., 2011), or hydrogen peroxide vapor (Commissioner, 2021)) are among the severe obstacles of implementing a PPE reprocessing strategy. Moreover, some technologies (e.g., UV-light) can only be applied to the surface of materials due to their low penetration capacity that hinders an efficient disinfection activity within the multi-layered filters of the filtering facepiece respirators (N95, FFP2 and FFP3) and the medical or surgical face masks (RIVM, n.d.;Rubio-Romero et al., 2020;Farke, 2020;Lore et al., 2012). Overall, a suitable sterilization technique to establish standardized protocols for PPE reprocessing is required with emergency not only from the healthcare but also environmental perspective. Due to the COVID-19 pandemic, the general and forced wear of PPE led to a high risk of improper disposal and it contributed to the ca. 8 million tons of pandemic-related plastics reported as of August 2021 (Peng et al., 2021). This dramatic increase in wastes becomes a new threat to the global ecosystem (Saliu et al., 2021;Hartanto and Mayasari, 2021). In general terms, the effective reuse by reprocessing of masks could reduce the overall consumption and therefore limit the contribution of PPEs to the world's plastic pollution (Kutralam-Muniasamy et al., 2022;De-la-Torre et al., 2022).
Sterilization with supercritical CO 2 (scCO 2 ) has emerged as a green alternative to conventional sterilization processes, being particularly effective for the processing of sensitive materials, such as novel polymeric scaffolds or synthetic tissues (Ribeiro et al., 2019). Namely, scCO 2 sterilization operates at mild conditions and has excellent penetration capacity in complex porous materials for biomedical applications (Ribeiro et al., 2019;Santos-Rosales et al., 2021). Moreover, it does not leave toxic residues that could cause skin irritation, so its use has been proposed in the disinfection of medical textiles and fabrics (Aslanidou et al., 2016;Cinquemani et al., 2007;Calvo and Casas, 2018). In this context, scCO 2 sterilization has been shown useful for the sterilization of nuclear, biological, and chemical suits for military purposes without altering the properties of the materials or their functionality (Calvo and Casas, 2018). In a recent investigation (Bennet et al., 2021), scCO 2 was used to inactivate spores and viruses in different environmental conditions on N95 and cloth masks with encouraging results. ScCO 2 combines several sterilization mechanisms for cell inactivation operating in tandem, being the cytoplasm and extracellular acidifications with CO 2 the most relevant ones (Spilimbergo and Bertucco, 2003;García-González et al., 2015;Matthew et al., 2013). The inactivation process mainly consists of three steps: 1) Acidification of the extracellular space due to the reaction between CO 2 and H 2 O forming bicarbonate and hydrogen ions. 2) Modification of the cell membrane composition due to the extraction of lipids from the membrane with scCO 2 (Lin et al., 1992). Additionally, the cell membrane structure is damaged by the acidification, increasing the permeability and further penetration of CO 2 (Kamihira et al., 1987;Garcia-Gonzalez et al., 2007). 3) Intracellular modifications due to cytoplasmatic pH lowering (Garcia-Gonzalez et al., 2007), which alter the activity of key enzymes and the cell metabolism pathways (Ishikawa et al., 1995;Hutkins and Nannen, 1993). The disbalance of the pH also modifies electrolytic concentrations (Ho-mu et al., 1993), resulting in the removal of vital constituents from the cell.
In this work, a supercritical sterilization protocol was developed for the sterilization of FFP3 face masks with the vision of being implemented at hospital environments. To achieve this, two main principles were set: the guarantee that the materials would preserve their safety and protection capacity after the sterilization treatment, and the development of a suitable sterilization protocol transferable to hospital sterilization units that effectively removes bioburden from the masks. Firstly, an effective sterilization protocol was developed against B. pumilus spores, a highly resistant microorganism towards scCO 2 sterilization methods (Santos-Rosales et al., 2021). A trade-off between mild pressure and temperature conditions, short processing times, and low amounts of chemical additives was established. Then, this protocol was applied to the FFP3 masks to evaluate the functional integrity (physicochemical properties, filtration efficacy, and mechanical resistance of straps) of the PPEs after 1 and 10 supercritical sterilization cycles.
2.2. Supercritical CO 2 sterilization procedures 2.2.1. Sterilization of bacterial endospores and contaminated FFP3 masks Sterilization tests of bacterial endospores were performed in a 600-mL NovaGenesis equipment provided with agitation (NovaSterilis Inc., Ithaca, NY, USA) using two different scenarios. On one hand, the sterilization of B. pumilus spore suspensions was performed in HPLC vials containing 100 μL (1.5-1.9 × 10 6 ) spores. Samples were placed in the autoclave and subjected to the sterilization conditions. On the other hand, FFP3 masks were contaminated with B. pumilus spores to recreate a more realistic and complex scenario. Briefly, FFP3 masks were cut into 4 cm 2 pieces and 100 μL of spore suspension (1.5-1.9 × 10 6 spores) were inoculated in the melt-blown polypropylene (PP) filter (Fig. 1). The cuts were confined in thermo-sealed pouches and put inside the autoclave. Varying volumes of a 30% (v/v) H 2 O 2 solution were dosed at the bottom of the autoclave achieving the desired additive contents (100 to 600 ppm of H 2 O 2 ), avoiding direct contact of the liquid with the samples. The high-pressure autoclave was conditioned to the sterilization parameters and finally the system was manually depressurized at a constant rate of 14 bar/min. A summary of all the tested parameters can be found in Table 1 to the manufacturer protocol for the evaluation of the membrane integrity. Control samples (untreated spores) and sterilized samples of 1.9 × 10 6 spores of B. pumilus were incubated for 30 min in the dark with a 1:1 mixture of Syto™ 9 (green dye) and propidium iodine (PI, red dye) to stain nucleic acids of live and dead spores. Stained spores were observed under confocal microscopy (LEICA TCS-SP2, LEICA Microsystems Heidelberg GmbH, Mannheim, Germany). In addition to the Live/Dead staining tests, the sterilization efficacy of the scCO 2 process was evaluated by turbidity measurements as follows. After processing, the spore suspensions from the HPLC vials were inoculated in 10 mL of tryptic soy broth (TSB) medium. After 24 h of incubation at 37°C, tubes were visually examined for bacterial growth. Finally, 1 mL of the culture was seeded in tryptic soy agar (TSA) plates to verify the absence of growth. The sterilization of FFP3 mask pieces was evaluated in the same way.

Preliminary evaluation of FFP3 masks under stress conditions
Prior to the sterilization tests, the processing window of the mask parts was evaluated at 200 bar of pressure and 45°C of temperature for a period of 2 h. The selected stress conditions of pressure and temperature correspond to more aggressive values than those defined as working conditions for the sterilization tests (Table 1). In addition, experiments were also performed under the previously mentioned processing parameters but in the presence of 1200 ppm of H 2 O 2 .
The tested FFP3 masks consisted of five sequential filters of different composition. Some of the filters of the mask were made from the same material with different fiber density. The following filter pattern was identified from the outer side to the internal layer in contact with the skin. (1) Filter a, made of spunbond (PP) fabric at a density of 50 g/m 2 ; (2) Filter b, cotton fabric at a density of 40 g/m 2 ; (3) Filter c, melt-blown PP at a density of 25 g/m 2 ; (4) Filter d, identical to filter b; (5) Filter e, spunbond PP at a density of 30 g/m 2 . A representative image of the FFP3 mask configuration is depicted in Fig. 1.
Small pieces of the FFP3 masks were put inside a high-pressure autoclave (Thar Process, Pittsburg, PA, USA) and the chamber preheated at 45°C was pressurized with CO 2 until 200 bar. Optionally, H 2 O 2 was incorporated in the autoclave before the pressurization step. Upon system depressurization, FFP3 masks were collected, their dimensions were measured and observed at naked eye, and CCD images taken with a 108 MP camera (Xiaomi Inc., Beijing, China) to evaluate major changes in their size and appearance, such as colour and shape.

Long-term FFP3 masks reprocessing for several cycles
Face masks were subjected to 1 and 10 sterilization cycles at the optimized static conditions (Table 1, condition highlighted in bold) for the evaluation of the integrity of the materials. Entire masks were put inside the autoclave, and 150 ppm of H 2 O 2 were incorporated at the bottom of the autoclave by dosing a commercially available solution (30% v/v) at the beginning of each cycle. These tests were carried out at 39°C and 100 bar during 30 min. Finally, the system was depressurized at a rate of 14 bar/min and the material was subjected to the following cycle or collected for further evaluation.

Evaluation of reprocessed FFP3 masks
Entire face masks subjected to 1 and 10 scCO 2 sterilization cycles were analysed to assess whether the exposure to the reprocessing induced relevant changes on the polymeric components of the FFP3 masks. Special attention was paid to the structure of the FFP3 masks filters, which was analysed by scanning electron microscope (FESEM, ULTRA PLUS Zeiss, Oberkochen, Germany) running at a voltage of 5 kV. Prior to their imaging, samples were iridium-sputtered (5-10 nm thickness) depending on the nature of the samples and the required information. The characterization of the filter layers was performed in processed masks (1 and 10 cycles) and in control (unprocessed) masks. All post-sterilization and control samples were visually inspected and scrutinized for any visible sign of degradation, deterioration or changes in texture or colour.

Evaluation of the filtration efficiency
In order to assess the filtration efficacy of the masks, bacterial filtration efficacy, breathability, and splash tests were performed based on EN 14683 standard. All tests were carried out on the finished face mask product and 5 samples per test were used. All face masks were conditioned at 21 ± 5°C and 85 ± 5% relative humidity for a minimum of 4 h prior to each test.
2.4.4.1. Bacterial filtration efficiency. Staphylococcus aureus (ATCC 6538) were inoculated in 40 mL TSB in an Erlenmeyer flask and incubated with a shaking velocity of 60 rpm at 37°C for 16 h. The culture was then diluted in peptone water to reach a concentration of 5 × 10 5 CFU/mL. The bacterial challenge was maintained between 1.7 and 3.0 × 10 3 CFU per test.
To perform the test, the extremities of the mask were removed first, and the mask with the test area of minimum 49 cm 2 was clamped between a sixstage cascade impactor and an aerosol chamber. An aerosol of Staphylococcus aureus was introduced into the aerosol chamber and drawn through the mask material and the impactor under vacuum. The testing was performed with the inside of the medical face mask in contact with the bacterial challenge. For each sample, the bacterial filtration efficiency percentage B was calculated using the following equation (Eq. (1)): where C is the mean of the total plate counts for two positive control runs without a sample, and T is the total plate count for the face mask sample. b a c d e  2.4.4.2. Breathability. The differential pressure through a measured surface area of the face mask material at a constant air flow rate of 8 L/min was used to measure the air exchange pressure. From each mask, 5 different test areas of 25 mm diameter were analysed. In order to perform the test, a 4.9 cm 2 area of the face mask was placed between the top and bottom parts of the holder and was clamped using a mechanical clamp with sufficient pressure to avoid air leaks. The differential pressure was then measured using a differential pressure manometer. This procedure was repeated 5 times for each mask. The differential pressure per mask area (ΔP, in Pa/cm 2 ) was then calculated as follows (Eq. (2)): where X m1 is the pressure in Pa at the low-pressure side of the material, X m2 is the pressure in Pa at the high-pressure side of the material, and 4.9 is the test area in cm 2 .
2.4.4.3. Splash resistance. The resistance of the face mask to penetration of synthetic blood was tested in accordance with ISO 22609:2004. The conditioned face masks were mounted on the device at a distance of 30.5 cm and exposed to a 2 mL splash of synthetic blood on the outer side of the mask at a stream pressure of 16 kPa. The inner side of the mask was then evaluated in terms of the penetration of the synthetic blood within 10 s after performing the test.

Mechanical tests
The straps of sterilized FFP3 masks were subjected to tensile tests in a TA.XT Plus Texture Analyzer (Stable Micro Systems, Ltd., UK) fitted with a 30 Kg load cell at a crosshead speed of 5 mm/s, according to ASTM standards ( D13 Committee,n.d.). In each test, straps of defined dimensions were fixed to the upper and lower clamps with a gap of 25 mm (5 cm specimens) or 75 mm (10 cm specimens), depending on the equipment configuration, and subjected to 5 cyclic loads. In each cycle the strap was stretched 25 mm in length (i.e., 100% strain for 25 mm gap and 33% strain for 75 mm gap). Tests were performed in duplicate at room temperature. The Young's modulus (E) was calculated from the stress-strain plots previous conversion to engineering stress and engineering strain.

Statistical analysis
All results were expressed as mean ± standard deviation. Statistical analyses of the breathability, the bacterial filtration efficacy results and of the mechanical performance of the straps (1-way ANOVA) were performed using GraphPad Software v9 (GraphPad Software., San Diego, CA, USA) followed by the post hoc Tukey HSD multiple comparison test.

Results and discussion
3.1. Definition of the operating window for the supercritical CO 2 sterilization of FFP3 masks The main operating parameters mastering the efficacy of the scCO 2 sterilization process are temperature, pressure, treatment time, presence of additives and depressurization rates (Ribeiro et al., 2019;Perrut, 2012). Thermal treatment can deactivate a wide variety of microorganisms, including resistance forms. A classical steam sterilization procedure requires a minimum operating temperature of 121°C, according to ISO standards (ISO 17665-1:2006(en), n.d.). The temperature effect on the scCO 2 sterilization efficacy relies on an increase of the solubility of CO 2 in the liquid cell medium, as well as the fluidity of the cell membrane. Both features enhance the acidification process and therefore the inactivation ability of scCO 2 . In comparison to steam/dry heat sterilization, only mild operating temperatures are required, but temperatures below 40°C usually start compromising the efficacy of the process (Zhang et al., 2006). Pressure is also essential for the efficacy of the process as higher pressures correlate with higher amounts of CO 2 dissolved in the intra-and extracellular media. However, this influence is easily observable at mild pressures (80-150 bar), and an increase in pressure above 200 bar does not significantly alter the sterilization efficacy (Perrut, 2012). Treatment time is also a critical parameter as it sets the contact time of bacterial cells with CO 2 . The addition of other additives, such as hydrogen peroxide, peracetic acid, or even water, boosts the bactericidal activity of the scCO 2 and they are often used in low contents (in ppm) as chemical additives to achieve the sterilization of the most resistant forms, such as bacterial spores (Bernhardt et al., 2015;Leggett et al., 2015). Finally, the sterilization efficacy of the process can be favoured by tuning the depressurization rate, inducing a mechanical burst of cells (Fraser, 1951;Foster et al., 1962).
A trade-off solution between the processing parameters and the material integrity was firstly set for the development of an efficient sterilization for the reprocessing of contaminated FFP3 masks. Since the main component of the mask is polypropylene (PP), the thermal stability of the polymer must be considered under a scCO 2 atmosphere. Due to the plasticizing effect of compressed CO 2 , PP may experiment swelling depending on the processing temperature and pressure (Hasan et al., 2010;Gong et al., 2017).
FFP3 face masks were subjected to 45°C and 200 bar of CO 2 during a 2hour period to confirm the resistance of the materials to a CO 2 atmosphere saturated with H 2 O 2 . After the sterilization process, the collected materials showed no relevant changes in the textiles of the mask layers (Fig. S1). Similar results were obtained for masks processed under scCO 2 atmosphere in absence of additives. Based on these encouraging results, the sterilization procedures were designed in a more conservative manner under milder operating conditions. For instance, the working parameters were maintained below these thresholds of temperature (<45°C), pressure (<200 bar), contact time (<120 min) and additive incorporation (<1200 ppm).

Sterilization of B. pumilus spores and contaminated FFP3 masks
The resistance of B. pumilus towards scCO 2 sterilization was previously reported and the incorporation of additives is usually required to achieve the inactivation of resistance forms (Zhang et al., 2006;Shieh et al., 2009;Santos-Rosales et al., 2019. For instance, the spores of this microorganism were recently proposed as the biological indicator to control scCO 2based sterilization protocols (Santos-Rosales et al., 2021).
Temperature was fixed at 39°C since protocols achieving high sterilization levels of B. pumilus dry spores under this working temperature were previously reported (Santos-Rosales et al., 2021. Regarding the processing pressure, an upper limit of 100 bar was defined to avoid deformations in the PP mask filters. A controlled depressurization during ca. 7 min was performed to prevent morphological changes in the masks. Namely, scCO 2 treated PP experimented was reported to transition from colourless material towards a slight yellow colouration due to a quick depressurization (Shieh et al., 1996). This aspect could be particularly detrimental for the multiple reprocessing of FFP3 masks.
At the former pre-set processing parameters, an optimization of the efficacy of the sterilization was performed in terms of additive content and overall contact time (Table 1). Commercial suspensions of B. pumilus spores were employed, which are known to be more vulnerable to scCO 2 compared to dry forms (Kamihira et al., 1987;Nakamura et al., 1994;Kumagai et al., 1997;Dillow et al., 1999). However, the use of H 2 O 2 as additive was required to achieve complete inactivation of the initial bioburden (10 6 spores). Regarding the purpose of the developed method, the presence of residual H 2 O 2 in the sterilized FFP3 mask can induce an irritation effect on the skin during usage. Moreover, inhalation of highly concentrated solutions of H 2 O 2 can induce severe irritation and inflammation of the respiratory tract (Watt et al., 2004;Urban et al., 2019). Therefore, only low contents of H 2 O 2 (less than 600 ppm) were incorporated during scCO 2 treatment, avoiding direct contact of the liquid with the samples.
The inactivation of B. pumilus spores was firstly evaluated through a specific staining method. Accordingly, Live/Dead staining was shown to work not only with eubacteria (Boulos et al., 1999) but also with archaea (Leuko  (Zhang and Fang, 2004) or bacterial endospores (Laflamme et al., 2004). The selected stain only allows the differentiation between intact and damaged cytoplasmic membranes. However, it is often used to assess active and dead cells (Berney et al., 2007) since membrane-compromised microorganisms can be considered as dead (Nebe-von-Caron et al., 2000). Regarding the dyes, the green fluorescing (SYTO9) can enter all cells, therefore both control (untreated) and sterilized samples would be stained. In contrast, spores with damaged membranes get also stained with propidium iodine. In our case, untreated spores ( Fig. 2A, B) only had green fluorescence and sterilized spores (Fig. 2C, D) displayed the green fluorescence paired with an intense red colouration. In addition, treated spores were also cultured in TSB to confirm the absence of microbial growth (cf. Section 2.3.1). An optimized sterilization procedure was developed requiring only 30 min of exposure to scCO 2 environment and using 150 ppm of H 2 O 2 ( Table 1). Neither decreasing additive content further nor using shorter contact times were able to reduce the bacterial endospore load. Since the sterilization procedure was controlled with endospores, which are the more resistant forms, the inactivation of other microorganisms is implicit (Zhang et al., 2006). Thus, results in this section verify the development of a powerful tool under mild processing conditions not only for the inactivation of endospores but also vegetative bacteria, viruses, and even fungi in contaminated materials.
Sterilization of dry porous materials, including masks, can be challenging due to poor accessibility regions and absence of water. Therefore, all the tested parameters for sterilizing spore suspensions were assessed in the more complex and realistic scenario of contaminated FFP3 masks. Due to the deep penetration capacity of scCO 2 , the experimental settings developed for the inactivation of liquid spore suspensions rendered identical results in contaminated FFP3 masks.

Evaluation of reprocessed FFP3 masks
Entire FFP3 face masks subjected to 1 and 10 sterilization cycles and using the optimized conditions (100 bar, 39°C, 150 ppm of H 2 O 2 , 30 min, depressurization at 14 bar/min) were observed under a scanning electron microscope (SEM) to confirm that the porosity, surface and diameter of the filters were not compromised when subjected to several sterilization cycles (Fig. 3). The fibrous structure of the filtering cloths was unaltered after the scCO 2 processing with no differences in comparison to the control samples, regardless of the number of sterilization cycles. Additional images with a variety of magnifications can be found as Supplementary material (Figs. S2, S3, S4, and S5). These results are in line with previous reports where masks subjected to 5 cycles of scCO 2 sterilization did not cause any visible damage on their surface properties (Bennet et al., 2021).

Bacterial filtration efficacy (BFE) and breathability
To investigate whether 1 and 10 sterilization cycles affected the integrity of the masks, the bacterial filtration efficacy (BFE) testing was performed according to UNE-EN 14683 standard concerning surgical mask performance (Standards, n.d.). Among the different categories of face masks, surgical masks are the only masks specifically designed and regulated to prevent bioaerosol dissemination from the wearer into the environment. Masks specified under the abovementioned standard are classified into two types according to their bacterial filtration efficiency (BFE) level: type I masks have a BFE ≥ 95%; type II masks have a BFE ≥ 98%. In this Fig. 2. Representative results of the membrane integrity assay of B. pumilus spores after scCO 2 sterilization (100 bar, 39°C, 30 min, 600 ppm H 2 O 2 ), stained with Live/Dead BacLight. Untreated spores (A, green filter; B, red filter) compared to sterilized samples (C, green filter; D, red filter). Images were not merged since green signal overlapped red signal. Complete spore inactivation was achieved under the abovementioned processing parameters. section, results showed that the scCO 2 sterilization treatment induced significant reduction in the BFE values (Fig. 4). Mask subjected to 1 and 10 cycles presented the same performance during the filtration test, displaying an identical 1.5% reduction in BFE value. Nevertheless, sterilized masks presented an overall BFE value above 98%, accomplishing the target values for type II masks, even though filtering masks (FFP3) are not considered medical devices. In a context of urgency, results in this section place great value on scCO 2 sterilization for the reprocessing of FFP3 mask while ensuring a suitable filtering performance according to UNE-EN 14683.
Resistance of a material towards the air flow can be measured by the pressure differential across the analysed area. From these values, a reliable parameter indicating the ease of breathing trough can be obtained (Kwong et al., 2021). Breathability is an essential criteria regarding costumer comfort during use and, if inadequate, might lead to an inappropriate PPE use (Pei et al., 2020). In this work, the breathability of FFP3 masks sterilized by scCO 2 was analysed and compared to untreated masks (Fig. 5). The sterilization treatment had a significant impact on the mask breathability properties. Namely, a downward trend was identified in the differential pressures as the sterilization cycles were performed. Masks subjected to one sterilization cycle experimented a decrease of 4% in the pressure drop values compared to the control masks. However, this reduction was not consistent with further reprocessing cycles and masks subjected to 10 sterilization cycles only suffered decreases close to 7%. Despite the scCO 2 treatment enhanced the FFP3 breathability, the filtration efficacy of the masks was not compromised to the point of avoiding use in a hospital setting.

Splash resistance
The evaluation of synthetic blood penetration on medical face masks is required from a regulatory perspective to ensure the suitable performance of PPEs (Standards, n.d.). Accordingly, the splash test was performed   4. Bacterial Filtration Efficacy of FFP3 masks before (control) and after (1 and 10) sterilization cycles. Asterisk denotes significant differences (1-way ANOVA; p < 0.05). The scCO 2 sterilization induced significant modifications in BFE results of the masks, although values were in the range set in the UNE-EN standards.
following ISO 22609:2004 and the results were determined in a qualitatively manner by visual inspections. ScCO 2 sterilization did not induce any modifications on the performance of the FFP3 masks in this regard (Fig. 6). Footprint of the splash blood can be easily observed in the outer part of the masks (A, C, E), but no trace of synthetic blood was seen on the inner sides of the masks (B, D, F) resulted intact regardless despite being subjected to multiple sterilization cycles. All the tested materials passed the splash test for a set stream blood pressure of 120 mmHg, which corresponds to the pressure threshold defined by ISO standards.

Mechanical performance of the elastic straps of the FFP3 mask
The respirator integrity following disinfection is crucial to demonstrate the preservation of the functional properties of the masks. The strap behaviour of the sterilized FFP3 mask was investigated through tensile tests following ASTM standards ( D13 Committee,n.d.). Modifications in this regard could compromise not only the patient comfort levels but also the appropriate fit once used. In this context, the implementation of the scCO 2 sterilization for the recycling of FFP3 masks could be jeopardized if the mechanical properties of the elastic straps are remarkably affected.
In a recent study, peracetic acid was used to decontaminate N95 masks and the strap elasticity was not affected after 5 decontamination cycles (John et al., 2021). However, steam sterilization was reported to be incompatible for the reprocessing of N95 masks since critical losses of strap elasticity were observed after a single sterilization cycle (Grinshpun et al., 2020).  5. Breathability values of FFP3 masks before (control) and after several (1 and 10) sterilization cycles. Equal letter denotes statistically homogeneous groups (1-way ANOVA; p < 0.05). The scCO 2 sterilization induced a significant increase in the breathability values of the treated masks but did not have practical relevance. To simulate a realistic scenario, specimens of 10 cm were cut and subjected to cyclic tensile tests applying a 33% deformation per cycle. Under this experimental setup, the straps presented a perfectly elastic behaviour when subjected to stress, and the deformation-recovery profiles obtained after the 5 cyclic charges were symmetrical for all the tested specimens, including those sterilized with scCO 2 (Fig. S6). The straps maintained their initial dimensions after the consecutive deformation cycles. Overall, the scCO 2 did not induce any modification on the mechanical response of the straps, and the values of the Young's modulus of the material were statistically identical (1-way ANOVA, p < 0.05), even after 10 reprocessing cycles (Table 2).
In a more exhaustive approach, specimens of 5 cm were subjected to cyclic tensile tests and applying a 100% deformation per cycle. Similar results were found regarding the mechanical behaviour of the straps under this experimental setup (Fig. S7). Namely, the scCO 2 sterilization induced certain increase on the Young's modulus values compared to control straps. This effect was constant and dependent on the sterilization cycles (Table 3), although was not statistically significant (1-way ANOVA, p < 0.05). Overall, the developed sterilization method based on scCO 2 preserved the mechanical properties of the processed straps.

Conclusions
In this work, the successful sterilization of FFP3 masks while preserving their filtration properties was achieved using scCO 2 technology. High sterilization levels (logR-6) were achieved for spores of B. pumilus, a microorganism particularly resistant towards the scCO 2 sterilization. At 39°C and 100 bar, only 30 min were required to achieve complete inactivation of the initial bioburden at the realistic scenario of contaminated FFP3 masks. This sterilization procedure represents an efficient, fast, safe and sustainable method for the multiple reprocessing (at least for 10 cycles) of face masks, in the absence of an alternative procedure. Although scCO 2 sterilization had an impact on the filtering and breathability properties of the FFP3 masks, the general performance of the PPEs was not remarkably compromised, even for those subjected to 10 sterilization cycles. Overall, the developed sterilization conditions underline the scCO 2 sterilization as a promising tool for the decontamination of FFP3 face masks and to allow their reuse in a context of urgency, such as the COVID-19 pandemic with a crucial impact on restricting the pathogen spread. The extension of this technology to other medical and technical equipment than FFP3 masks is currently under evaluation. PPEs from several suppliers will be also reprocessed with the developed scCO 2 method to assess their universal performance after sterilization. In addition, scale-up tests are also being performed to validate the applicability of the technology to a pilot scale. The low logistic complexity of the scCO 2 sterilization encourages its implementation in hospital settings, fostering the incorporation of this technology to the healthcare arsenal.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.  Table 3 Tensile mechanical properties of FFP3 mask straps after being subjected to scCO 2 sterilization compared to untreated straps. Straps were subjected to 5 consecutive deformation cycles (100% deformation/cycle).