Introduction

Drinking water quality is one of the greatest factors affecting human health. In many countries, especially in developing countries, drinking water quality is poor, with consumption of poor-quality drinking water inducing many waterborne diseases (Li and Wu 2019). It has been estimated that > 800 children under the age of 5 years die daily from diarrhea caused by dirty water, and that, by 2030, 700 million people worldwide could be displaced by intense water scarcity (Oxfam 2018). One of the most important sustainable development goals developed by the United Nations is to, ‘ensure availability and sustainable management of water and sanitation for all’ (United Nations 2018).

Methods to supply adequate amounts of clean water to populations of underdeveloped countries remain inadequate. Water purification tablets containing calcium hypochlorite can sterilize water, eliminating disease-causing microorganisms (ScienceDaily 2013). Sterilized water, however, requires filtration (P and G 2021) to reduce turbidity and the number of floating particles (Kgabi et al. 2014). Much of the drinking water in underdeveloped countries is very turbid (Marobhe et al. 2007), indicating a need for additional filtration.

The COVID-19 pandemic has increased the use of masks and respirators for both health workers and non-health-related workers, resulting in the production of enormous amounts of plastic waste (Klemeš et al. 2020; Okoro et al. 2020). Anti-COVID-19 filtering facepiece respirators (FFRs) consist of an inner melt blown (MB) filter made of polypropylene, and an outer layer consisting mostly of synthetic fibers. These plastic materials are difficult to recycle and do not decompose naturally. Although studies have investigated the reuse of these FFRs (Su-Velez et al. 2020; Vernez et al. 2020), many parts of FFRs, including their outside/inside layers, ear bands, and plastic hooks, are difficult to recycle as plastic resources (Battegazzore et al. 2020). Therefore, a fresh point of view for recycling FFR is required. The present study, therefore, assessed whether FFRs could be recycled for use as preliminary water filters.

Materials and methods

Anti-COVID-19 Filtering Facepiece Respirators (FFRs)

N95 anti-COVID-19 FFRs is probably one of the most popular product that is used worldwide (Okoro et al. 2020) that have been reported to collect > 95% of particles under 0.3 μm in size present in air samples (United States National Institute for Occupational Safety and Health) (CDC 2019). The equivalent FFRs in other countries are called FFP2 in the European Union, KN95 in China, DL2 in Japan, and KF94 in South Korea (Liao et al. 2020). KF94 FFRs have been found to collect > 94% of dust particles under 0.4 μm (Park et al. 2020). The N95 and KF94 FFRs utilize the same principle, with MB (Melt Blown) filters in the middle layer filtering out dust and viruses. To avoid product differences between manufacturers, FFRs were purchased from the same company. N95 FFRs were purchased from Kimberly-Clark of the United States (Kimberly-Clark, Irving, Texas, USA), and KF94 FFRs were purchased from Yuhan-Kimberly of South Korea (Yuhan-Kimberly, Seoul, Korea). Dental mask type FFRs with MB filters in the middle layer were also purchased from Yuhan-Kimberly of South Korea.

Preparation of Soil–Water Mixtures

Soil was collected from a dried puddle at the roadside on Jeju Island, specifically from an area of bare soil without vegetation. After preliminary testing to determine the optimal soil:water ratio, the 2 g soil mixture, consisting of 1.4 g fine soil (sieved with 690 μm sieve) and 0.6 g coarse soil (sieved with 2 mm sieve), was mixed with 1 L water collected from a puddle (total area, 20 m2; maximum water depth, 30 cm) in a rock bed dry stream near Jeju University (Fig. S1 of Supplementary materials). Because people living in underdeveloped countries usually allow particles in collected water to settle out before drinking, the use of highly turbid water samples in this study would exaggerate the turbidity of the water collected in such underdeveloped countries. Thus, the maximum turbidity of the water samples was set at the margin where the bottom of the container became invisible.

Filtering System Development and Filtration

Three filtering systems were proposed (Fig. S2 of Supplementary materials). The first, a syringe-based filtering system, was composed of a 250 mL syringe, a rubber cap, and a hose clamp. The area used for filtering was about 20 cm2 (diameter, 5 cm), with an additional 1 cm margin required to avoid leaking. The second, a foot pump-based filtering system, was composed of a foot pump, a hose, two 2 L plastic bottles, and one 500 mL plastic bottle. The area used for filtering was about 50 cm2 (diameter, 8 cm) with additional margins of ~ 1–2 cm. The third, a plastic bottle-based filtering system, was composed of a soft (squeezable) plastic bottle and a rubber band. The area used for filtering was about 7.5 cm2 (diameter, 3.1 cm), with an additional 2 cm margin required to avoid leaking. Of the three methods, the syringe filtering system had the least leakage.

After adding 2 g soil to 1 L collected water, the samples were mixed thoroughly and filtered with the FFR filtering system. Most samples were filtered with the syringe-based filtering system, although the plastic bottle-based filtering system using a 1 L milk bottle and a 2 L mineral water bottle was also used.

Water and Soil Analysis

The filtered water was filtered again through Whatman filter paper no. 2 (Whatman plc, Little Chalfont, United Kingdom), and the weight of the soil removed by the filter paper was measured by determining the increase in weight of the dried filter. Carbon (C) and nitrogen (N) contents of the soil were determined using an elemental analyzer (Flash EA 1112; Thermo Fisher Scientific, Waltham, MA, USA).

The electroconductivity (EC) and pH of the water samples were measured with an Orion Star A329 portable multiparameter instrument (Thermo Fisher Scientific, MA, USA). To determine total phosphorus (TP), magnesium (Mg), potassium (K), and iron (Fe) contents of water samples, 10 mL aliquots of water were digested with 5 mL of 70% perchloric acid with shaking at 120 rpm for 24 h. These solutions were filtered through filter paper, and the contents of the filtrate were analyzed using an Inductively Coupled Plasma Atomic Emission Spectrometry (ICP-AES) (ICPS-1000IV; Shimadzu, Japan). To assess drinking water quality, total viable bacterial cell counts (TBCC), total viable Escherichia coli cell counts (TECC), biochemical oxygen demand (BOD), turbidity, total organic carbon (TOC), total nitrogen (TN), and SiO2 were analyzed by the Water Quality Analysis Center of the National Instrumentation Center for Environmental Management (NICEM), Korea.

Statistical Analysis

Differences between two groups were evaluated using the Wilcoxon two-sample test when normality assumptions were violated. Comparisons among three or more groups were assessed by one-way analysis of variance (ANOVA), followed by post hoc Tukey’s studentized range (HSD) test. All statistical analyses were performed using SAS 9.1 software (SAS Institute, Inc., Cary, NC, USA). Differences were considered statistically significant at p < 0.05.

Results

All types of FFRs and filtering systems tested were found to significantly remove soil from 1000 mL turbid water samples (Table 1). All FFRs showed > 90% filtration efficiency, with syringe-type filtration showing 95% efficiency. Dental masks showed significantly higher amounts of unfiltered soil than N95 and KF94 FFRs, but dental masks were able to filter > 50% of soil from water. Altering the volume filtered to 500 mL or 2000 mL had no effect on filtering efficiency.

Table 1 Weight of unfiltered soil remaining in water after filtering
Full size table

The soil used for filtering experiments had low concentrations of N (0.03%) and C (1.24%) (Table 2). Jeju Island is a volcanic island with soil composed of volcanic ash, which has lower nutrient and organic matter content than yellow loess (Nanzyo et al. 1993). The ECs of water before and after filtration did not differ significantly, as expected for soils with low nutrient contents (Table 3). While filtering with N95 FFRs increased the pH of water, filtering with KF94 FFRs reduced the pH (Table 3). The concentrations of P, Mg, K, and Fe were significantly lower after than before filtering with both N95 and KF94 FFRs (Table 4). TBCC and TECC values showed large deviations among replicates following filtration, with no significant differences between N95 and KF94 FFRs (Table 5). Similarly, BOD values did not differ significantly between filters. Water turbidity was significantly reduced, by > 85%, by filtration with both N95 and KF94 FFRs (Table 6). Furthermore, TOC values were also significantly reduced by filtration with both N95 and KF94 FFRs (Table 6). However, filtration by FFRs did not significantly alter TN and SiO2 values for these water samples (Table 6).

Table 2 Properties of soil used in these experiments
Full size table
Table 3 Electro conductivity (EC) and pH of water after filtering
Full size table
Table 4 Concentrations of elements in water after filtering
Full size table
Table 5 Concentrations of biological indicators in water after filtering
Full size table
Table 6 Concentrations of physio-chemical indicators in water after filtering
Full size table

Discussion

Filtering Systems

Using syringe-based filtering system with attached FFRs in front of the syringe has several advantages such as tight sealing, rapid filtering, and easy construction. However, syringe-based filtering systems can only filter a maximum of 250 mL water for a time. After filtering 250 mL of water, rubber caps in front of the syringe must be removed and replaced with new pieces of FFRs. Thus, four separate filtering steps were required to filter 1 L water, making this a time-consuming process to pour water through the back of the syringe. However, because this method was the most accurate, with lowest leakage, it was utilized as the standard method for analysis. Filtering with a foot pump was fastest, but constructing the system was very difficult. Thus, this method was not used to analyze water quality. Filtering systems using a 1 L plastic milk bottle were easiest to construct but there was some leakage during filtering, and squeezing the bottle repeatedly made the bottle fragile. However, because this system was easiest to construct and apply, it was used to test the filtering abilities of N95 and KF94 FFRs. Both provided rapid filtering (link to video of filtering: https://youtu.be/-TRKcjBNDxo), and when compared water colors before the filtration and after the filtration were distinctive that filtered water showed much more clear color (Fig. S3 of Supplementary materials).

Water and Soil Analysis

The filtration efficiency of each FFR was > 90%, with syringe-type filtration having 95% filtration efficiency for removing soil particles (Table 1). After filtration, the water became much clear, indicating that the particles and soil samples suspended in the water had been removed (Fig. S4 of Supplementary materials). Most of these soil and other particles were removed by the MB filter (Fig. S5A of Supplementary materials), with only small amounts remaining after filtration. Most anti-COVID-19 FFRs consist of three or four layers; products made by Kimberly have four layers. The second layer, consisting of the MB filter, is a very dense meshed structure, which is very effective at keeping soil particles from passing through. Very little soil and other particles were present on the third layer, after the MB filter, as most of these particles had been removed by the MB filter (Fig. S5B of Supplementary materials). Dental masks (anti-droplet masks) filtered > 50% of soil from the water, indicating that the filtration efficiency of dental masks containing MB filters was much lower than those of N95 and KF94 FFRs. Because the values of the dental mask were high, the other values of N95 and KF94 FFRs did not differ significantly. N95 milk bottle and N95 water bottle was significantly highest group while KF 94 water bottle was middle group and others are lowest group (Table 1). These differences may be due to the filtering areas, as water bottle used much less area for filtering [half diameter (quarter area) compared to Syringe], the loading of soil and particles to filter can be the reason. However, FFRs cannot be completely sealed when using plastic water and milk bottles with rubber bands. Although tightly holding fingers were used to minimize leaking, leaking became inevitable (link to video of filtering: https://youtu.be/-TRKcjBNDxo). High values may, therefore, be due to leaking, not to the limits of filtration capacity. Because milk bottles with a greater filtering area yielded the same results as water bottles, and syringe filtering of 2 L water had no effect on the results, it seems the volume of water filtered was lower than the filtration capacity of the device. Soil and other particles, however, were markedly reduced after filtration, suggesting that the use of plastic bottles in developing countries lacking sufficient equipment can remove soil and other particles from water. It was difficult to purchase N95 FFRs in Korea, and we’ve lost most of prepared N95 FFRs during testing prototype and practice. Therefore, we could not use N95 FFRs for double volume (2000 ml) test. However, it is likely that the N95 FFRs we used in this study also would not reach their limit with respect to filtration capacity.

Because water in puddles used as a water source where undergoing limited water supply usually has many soil particles and characteristics of these soil particles are usually poor and bare, not organic-rich (WorldVision 2016, 2018). Therefore, in this study, we tried to use poor soils without many organic matters for our experiments. As those limited water source areas are usually arid places, with limited vegetation growth. Therefore, we collected soil from dried puddles without vegetation. As a result, the soil samples collected had low contents of N (0.03%) and C (1.2%) (Table 2). Because the soil samples used in the present experiments had relatively low concentrations of N and C, we expected that the concentrations of other elements, such as P and K, would also be low. The EC was not altered by the treatments (Table 3), indicating an absence of ionic substances from these soil samples, even though the water color was altered by filtration (Fig. S4 of Supplementary materials). Filtration of the water samples with N95 FFRs increased the pH, whereas filtration with KF94 FFRs reduced the pH, although both were around pH 7. KF94 FFRs are whiter in color than N95 FFRs, suggesting that the former probably contains a fluorescence brightening agent and that some chemical altered the pH of water. However, because the materials and processes used in manufacturing are proprietary information, we could not determine the reason for the difference in pH between waters filtered by the two FFRs. Still, the pH range of the filtered water was around pH 7, close to neutral range.

P, Mg, K, and Fe concentrations were found to be significantly lower after than before filtration (Table 4). Phosphorus triggers eutrophication, which leads to the degradation of drinking water (Wang et al. 2018). Most treatments reduced TP concentration to one-eighth of that in prefiltered water, whereas treatment of 1000 mL water with N95 FFRs reduced the TP concentration to one-fourth of that prior to filtration. Mg in drinking water is a risk factor for cardiovascular diseases (Rylander et al. 1991). Filtration with FFRs reduced Mg concentrations in water to less than one-half that prior to filtration. In nature, K concentrations are low in water, but contamination of water by salts and potassium chloride (detergent) can expose people to high concentrations of K, which cause many side effects (Organization 2009). All the filtration treatments reduced K concentrations in the water by > 20%. Fe is considered an important indicator of water quality (Saha et al. 2018), with FFR filtration reducing the concentrations of Fe to 10–20% of those observed in prefiltered water. The significant reductions in the concentrations of elements analyzed by ICP emission spectrometry could be due in part to the acid digestion. This digestion dissociated elements attached to the soil particles, allowing their detection by ICP. Because elements in unfiltered water can enter the human gastrointestinal tract, their removal by filtration with FFRs may be an important step in the purification of drinking water.

Water quality analysis of the microorganisms in these soil-in-water samples showed that different methods of filtration did not result in significant differences in TBCC and TECC values (Table 5). AS TBCC and TECC values had large deviations among replicates, no significant differences were detected. However, the absolute values of TBCC following filtration with N95 and KF94 FFRs were 25% and 60%, respectively, and TECC values following filtration were 20% of their prefiltered values. BOD values did not differ significantly, as mixed soil has low nutrient contents, thus having no effect on the concentrations of dissolved matter in water. The BOD value of prefiltered water was < 3 mg/L, satisfying the requirements for residential water (World Health Organization 2017), and indicating that the water collected from these puddles was not so contaminated.

The most important factor degrading the quality of drinking water in developing countries is turbidity, with children drinking muddy water being especially at risk of various diseases (WorldVision 2016, 2018). FFRs significantly reduced the turbidity of water by > 85%. Despite the importance of turbidity to the quality of water, few studies to date have assessed the turbidity of drinking water in developing countries. One study (Marobhe et al. 2007) found that the average turbidity of drinking water in a developing country was around 200 nephelometry turbidity units (NTUs), despite these waters being collected from lakes and reservoirs, not puddles. Because the turbidity of water is very high in places where people must walk for hours to collect water from pools (puddles), filtering this water through FFRs would be a very effective method for managing drinking water. FFRs also significantly reduced the TOC values in our study. Because TOC is positively associated with an increase in pathogens (Williams et al. 2015), FFR-associated reductions in TOC would likely enhance the safety of stored water. In contrast to TOC values, FFRs did not significantly affect TN and SiO2 values. Because the Water Quality Analysis Center of the NICEM does not use a digestion method to analyze water samples, the results obtained for turbid and clear water was not different.

Several safety concerns are associated with the use of worn FFRs, including the presence of pathogens, the viability of microorganisms on the mask surface, and the transmissibility of fomites. However, the risk of infection or reactivation of pathogens on FFRs will be reduced by maritime transport, as the surface viability of COVID-19 and other coronaviruses on porous surfaces was reported to be about 72 h (Van Doremalen et al. 2020). The survivability of microorganisms after 48 h on the surface of FFRs is expected to be low (Cho et al. 2014), with maritime transport under dry conditions eliminating most of the risk. To ensure 100% safety, decontamination with microwave heat and high dry heat can also destroy the activities of viruses and other microorganisms. As these processes also result in the degradation of the respirator material (Su-Velez et al. 2020), it is not used. However, as we are not using airborne particle filtration abilities of FFRs, applying such heating methods will not be a problem. Also, treatment with hydrogen peroxide (H2O2) has been shown to successfully remove mold spores and other pathogens without affecting airflow resistance or the fit of filters, and to maintain an initial filter penetration of < 5% (Rempel et al. 2021). Also, ultraviolet germicidal irradiation can effectively decontaminate respiratory viral agents (Yang et al. 2020). Proper sterilization of FFR filters can eliminate contamination problems, allowing the reuse of FFRs. Because we suggest using FFRs as preliminary water filters, mostly to remove suspended particles and soil samples from water, the additional use of water purification tablets can ensure sterilization of the water.

Overall, filtering contaminated water with FFRs significantly reduced turbidity, TOC, and major pollutants and indicators, such as P, K, Mg, and Fe. Because the soil used in these experiments was of poor quality, with low C and N contents, results might be more dramatic if soils rich in organic matter and nutrients were used. The World Health Organization (WHO) has set the drinking water quality standard for Fe at 0.3 mg/L (World Health Organization 2017). Most of the treatmentss tested in this study resulted in water with < 0.3 mg/L Fe, although one treatment, of 1000 mL water with N95 FFR, yielded a marginal value. The WHO standard for Mg in drinking water is 0.3 mg/L (World Health Organization 2017). All of the FFR treatments in this study reduced Mg contents below the standard. Also water filtered with FFRs satisfied WHO standards for K (20 mg/L), EC (300 μs/cm), and pH (6.8–8.0) (World Health Organization 2017). The 2017 WHO standard did not set a maximum value for turbidity, but the 1971 WHO standard for turbiity of drinking water was 25 NTUs (Kumar and Puri 2012; Public Health Nigeria 2020). Although filtering water through FFRs reduced turbidity eight-fold, it remained over 25 NTUs. The WHO also set Escherichia coli cell counts at zero (World Health Organization 2017), but these bacteria were detected in filtered water. Therefore, a combination of filtering through FFRs and treatment with water purification tablets to eliminate disease-causing microorganisms could help to generate secure drinking water. Moreover, water purification tablets have some precipitation effects that can reduce turbidity (ScienceDaily 2013), making filtered water more suitable for drinking.

Conclusions

Utilizing FFRs for preliminary water filtering before treatment with water purification tablets is an effective method for providing clean drinking water to places where access to clean water is difficult. This method does not require complicated techniques or devices: a plastic bottle with a rubber band is sufficient for using FFRs as preliminary water filters. This simple method could save the lives and health of many people who experience intense water scarcity and could help to achieve one of the sustainable development goals of the United Nations, ensuring the availability and sustainable management of water and sanitation for all. In addition, using recycled FFRs could reduce plastic waste and conserve resources. As about USD 22,143 million is expected the face mask market in 2021 (markets and markets, 2020), recycling FFRs is also likely to be of great importance as a pioneering study in finding additional uses for huge economic costs. Thus, recycling anti-COVID-19 FFRs as preliminary water filters is an effective, economical and sustainable method for both providing safe drinking water and eliminating waste associated with FFR production.