Filtering efficiency measurement of respirators by laser-based particle counting method

Graphical abstract


Introduction
The ambient air is full of different biological and non-biological aerosol particles, like viruses, bacterias, fungal spores, pollens, and simple fragments. The size of these aerosol particles varies between 0.1 and 10 µm, so without personal protection, they can easily penetrate the human respiratory system and may cause health issues, especially infectious diseases [1,2]. The latest worldwide public health problem is caused by the 2019 novel coronavirus (COVID-19), which has a particle size between 0.06 and 0.14 µm [3]. COVID-19 infection is mostly transmitted by direct mucosal contact to droplet-borne viruses originating from the mouth or nose of an infected person [4]. Therefore, some of the latest researches revealed that surgical masks and respirators could prevent or at least slow down the transmission of COVID-19 between human individuals [5][6][7][8]. Most of the governments suggest the wear of personal protective equipment in closed and crowded spaces like public buildings and transportations. In the case of a lack of personal protective equipment, it is mandatory to wear face coverings or so-called community masks. Such devices have low to no filtration efficiency, and they are perceived more as a hygienic measure that minimizes the projection of user's respiratory droplets saliva, sputum, or respiratory secretions when talking, coughing, or sneezing [9,10].
The respirators are classified according to their filtering efficiency. The European Respirator Standard (EN 149:2001) distinguishes three classes: FFP1 (Filtering FacePiece), FFP2, and FFP3 with corresponding minimum filtration efficiencies of 80%, 94%, and 99%, with a test agent, which has 0.4 and 0.6 µm median mass aerodynamic diameter [11]. The US versions are the N95 and N100 respirators (having 95% and 100% filtering efficiency) [12], according to the NIOSH-42C FR84 standard. The Chinese version is the KN95 (having 95% filtering efficiency), according to the GB2626-20 06 standard. Although the standards slightly differ in the determination methods of the filtering efficiency, the FFP2, N95, and KN95 respirators can suppose to be equivalent [13]. Generally, respirators are built up from two outer textile layers, which do not have any filtering role, and 1-2 inner filter layers, which can block liquid and solid particles as well (valid for EU tested respirators) [14]. (In the US, there is a distinction between the respirators that can be used to protect against both groups of contaminants.) The most essential and complicated layer is always the particle filter layer, which was usually made from melt-blown polypropylene microfibers [15].
The application of microfibers has limitations in filtering efficiency. Although the filtering efficiency could be improved with the increase of the thickness of the filter material, it decreases the breathability of the respirator [16]. Polypropylene electret melt-blown nonwovens are also applied, which could have increased filtration efficiency by electrostatic forces without sacrificing the air resistance [17]. In the case of FFP3 respirators, usually, exhalation valves needed to be used to avoid the lack of O 2 and the accumulation of CO 2 . The application exhalation valves have high health risks in a pandemic, since it lets out the exhaled air to the environment, so the protection of the mask is only one way. Therefore, some of the EU states already banned the marketing of respirators with exhalation valves. Recent researches showed that the electrospun nanofiber webs (mostly from polyacrylonitrile) are alternative candidates for filter materials due to their small pore size, small diameter, and large specific surface area [18,19].
The fast-spreading of COVID-19 caused a worldwide shortage of respirators, melt-blown non-woven fabrics, which are the raw material of the mask filters, and the respirator testing possibilities. The lack of respirators initiated to search for emergency solutions. The most obvious is the possible reuse of respirators with some decontamination methods. Boskoski et al. [20] suggested using highly energetic, shortwave, ultraviolet germicidal irradiation at 254 nm to decontaminate N95 respirators from viral agents. Grossman et al. [21] extended the wear life of N95 respirators using vaporized hydrogen peroxide (VHP) disinfection. Juand and Tsai [22] suggested heating to 70 • C or boiling the N95 respirators, which processes do not change the filtering efficiency, and they are against the use of soap or alcohol for disinfection. However, most of the studies agree that the reuse of respirators should only be practiced as a crisis capacity strategy [20][21][22][23]. In addition, Suen et al. [24] reported a significant fit factor and filtering efficiency drop of N95 respirators after nursing procedures. Ho et al. [25] found that the simple homemade cotton masks could be a temporary solution since they can block the respiratory droplets effectively. Different custom-made mask frames with changeable filters are also appeared on the market [26].
The lack of respirators also caused the partially controlled inflow of different chinses respirators into the EU; most of them were marked as KN95. Some EU member states reported quality problems directly after the arrival of them. Therefore, re-qualifying these respirators are essential before using them. In 2020, only eleven laboratories had the accreditation for EN 149:2001 standard in the EU. They are overloaded, so there is a strong need for capacity building. The EN 149:2001 standard measures the filtering efficiency with a photoelectric flame photometer. The flame photometer can determine the concentration of certain metal ions, mainly from groups 1 and 2, which have low excitation energies, like sodium, potassium, lithium, and calcium. It is a controlled flame test, the investigated material is vapourised, and the intensity of the flame color depends on the energy that had been absorbed [27]. This method is well used for decades. However, it is relatively expensive, can analyze only metals, and cannot give any information about the analyzed particle sizes, which would be important in the case of infection blocking of respirators.
Particle counting methods are also already used in laboratory practice. Rengasamy et al. [28] compared the penetrations of N95 respirators by photometric and by ultrafine condensation particle counting (UCPC) methods. They found that the penetration values obtained using the UCPC are more accurate in the case of the photometric method, which lacks sensitivity for small particles (<100 nm). Kim et al. [29] also used the UCPC method successfully to detect nanoparticles (3-20 nm) penetration through commercial filter media. However, in the case of UCPC methods, usually low volume flow rates are used. The standard operating flow rate is 0.30 to 0.5 l/min, which is far from the volume flow rate during the use of the respirators [30]. Balazy et al. [1] built a combined system, from a differential mobility diameter, from a condensation particle counter, and from a laser particle spectrometer. The combined machine cowered the measurable particle range between 10 and 1000 nm, but this solution does not suggest an easy and commercial use. Our aim was to develop a fast and easy filtering efficiency measurement method based on particle counting, which could help the respirator's classification during COVID-19 pandemic.

The measurement method and the system
Our method determines the filtering efficiency of the filtering material of the respirators, which is one of the most important parameters of a mask. It has to be noted that further important parameters of the masks have not been covered in this study, like inward leakage as a measure of fit to the wearer's face [12], breathing resistance, and efficient exchange of CO 2 under the facepiece. Our measurement method is based on particle counting. It is destructive from the point of ready respirators. A sample needs to be cut out from the filtering material of the investigated mask. The main parts of the measurement system are a particle counter (LASAIR III 310C) and a self-designed 3D printed sample holder that fixes the investigated filter material during the measurements. Fig. 1 shows the measurement system.
The surface size of the sample holder is 50 × 70 mm, and it has four main parts: bottom holder, upper holder, O ring, and fixing screws. The inner and outer diameter of the O ring is 20 and 26 mm, respectively. A 35 × 35 mm sample is positioned over the O ring. It keeps the sample during the measurement and blocks the leakage between the bottom and upper parts of the sample holder. So the ambient air is suctioned only through the sample. The O ring sits in a riffle, which sinks the O ring 1.5 mm into the bottom holder to ensure the optimal fixing of it. The fixing screws push the upper part to the bottom together with the sample and the O ring. The assembled sample holder is connected to the particle counter with the suction tube and suction pipe (Fig. 1).
The Lasair III 310C is a portable aerosol particle counter which is usually used to qualify clean-rooms in microelectronics or medicine production [31]. It works with a laser-based counting method. The accessing particle in the device passes through laser light. A photodetector detects the redirected light and the loss of the light, which determines the size of the obstructing particle. Finally, the particle counter distinguishes the counted particles according to the estimated sizes into channels. The used particle counter has six channels with the following particle size ranges: 0.3-0.5 µm, 0.5-1 µm, 1-5 µm, 5-10 µm, 10-25 µm, and 25 µm >, respectively. In a given particle size range, there is no information about the exact particle distribution.
The volume flow rate of the device is 30 l/min, which is approximately the average breathing flow of humans in a resting position. The measurement contained two steps. First, the Particle Number Concentration (PNC) [pcs./m 3 ] of the ambient air is measured for 1 min. Right after, the sample holder is installed on the particle counter, and the PNC behind the filter material is measured again for 1 min. The filtering efficiencies are determined respectively from the PNC differences.
A considerable advantage of our measurement method is simplicity, since it can be installed in a few days, availability of the particle counter device on the present market, and the low price (~15.000-18.000 USD) compared to the filtering efficiency measurement system according to the EN 149:2001 standard (~60.000-100.000 USD).

Determination of the filtering efficiency
Contrary to the flame photometer based measurement methods, our particle counter measurement method can determine not only one but two different filtering efficiencies. The Particle Filtering Efficiency (PFE, %) can be measured on all channels of the particle counter. It is defined as: where PNC A is the particle number concentration of the ambient air [pcs./m 3 ], and PNC C is the particle number concentration behind the filter material [pcs./m 3 ]. The Concentration Filtering Efficiency (CFE, %) can be calculated on any particle size range of the measurements. It is defined as: where PC A is the particle concentration of the ambient air [mg/m 3 ], and PC C is the particle concentration behind the filter material [mg/m 3 ]. The particle counter can measure particle number concentrations (PNC) directly; the particle concentrations (PC) can be calculated from the PNCs and the particle sizes. (During the calculations, it was assumed that the specific density of the particles in the ambient air is near homogeneous).

Investigated respirators
The investigated respirators were distinguished into two groups: a reference group that contains three different FFP2-3 respirators with certificates according to the EN 149:2001 standard and an unknown group, which includes ten different KN95 respirators without a europian certificate. Table 1 shows the known information about the investigated respirators. Ten samples were tested from each mask type, and the measurements were half-blind.

Particle number concentration and size distribution of the ambient air
The measurements are carried out by the ambient air; therefore, it is important to know how the PNC and the particle size distribution changes from day to day. These parameters are influenced by a lot of factors, e.g. the weather or the traffic level around the measurement location. The investigations were carried out for two months in different environmental conditions. PNCs and particle size distributions were measured in rainy and dry weather, in cold and hot weather, and at low and high traffic levels. Altogether more than 300 measurement results were used. It was found that the PNC changed from between 1 × 10 7 (clear air) and 5 × 10 7 (polluted air) at our measurement location (Budapest). The particle size distribution in the function of PNC can be seen in Fig. 2 that uses a logarithmic scale due to the high differences.
The particle size distribution was almost constant in the range of the PNC between 2 × 10 7 and 5 × 10 7 . The majority of the particles (~98%) were smaller than 1 µm, the ratio of the particles between 1 and 5 µm was around ~1.5%, and only 0.5% was bigger than 5 µm. The EN 149:2001 standard defines CFE values in the case of aerosols with 0.4 and 0.6 µm MMAD (Median Mass Arodinamic Diameter). Our analysis of the MMAD parameter showed that in the 2 × 10 7 -5 × 10 7 PNC range, the MMAD was between 0.34 and 0.76 µm, if only the particles between 0.3 and 5 µm were counted. It is close to the requirements of EN 149:2001 standard. Therefore only the PNC between 2 × 10 7 and 5 × 10 7 were used during the measurements, and the particle concentrations (PC) were counted only from the 0.3-5 µm range during the CFE calculations. It has to be noted that we usually measured with a finer aerosol but with a lower flow rate (30 l/min) than the EN 149:2001 standard (95 l/min). These differences did not have a considerable effect on the measured results compared to the results of the standard EN149:2001 method (See in Section 3.2).

Validation of the measurement method
The validation steps of the measurement were done by the reference respirators (R1-R3). The repeatability of the measurements was investigated in the case of 4 × 10 7 PNC. It was found that the repeatability slightly depends on the particle sizes. Two ranges of particle sizes could be distinguished (0.3-5 µm and 5-25 µm), wherein the repeatability errors differ. However, they are very low in both ranges: lower than 0.05% in the range of 0.3-5 µm size particles and lower than 0.5% in the range of 5-25 µm size particles.
The PNC of the ambient air is hardly controllable, it is very important that the measurement cannot be too sensitive to this parameter. So it was also investigated whether the PNC of the ambient air affects the filtering efficiency of the filter materials. All the above-mentioned PNC values (1 × 10 7 -5 × 10 7 ) were investigated. Table 2. shows the filtering efficiencies of the R3 reference filtering material (FFP3 respirator) at different PNCs. The filtering efficiency did not show sensitivity to the PNC between 1 × 10 7 and 5 × 10 7 . Therefore the measurements were executed at the given particle number concentration on the given day (practically between 2 × 10 7 and 5 × 10 7 ).
Finally, the measurement method was tested with certified filtering materials. Table 3 shows the filtering efficiencies of the R1-R3 filtering materials. The measured CFE values were always slightly over the guaranteed CFE values by the manufacturer. It could be the overinsurance of the filtering materials due to the inward leakage of the respirator, or it could be due to the applied lower flow rate.
According to the validation results, it can be stated that our measurement method is a stable and capable method to investigate unknown respirators. Europe   Fig. 3 shows the CFE values of filtering material in the U1-10 "unknown" Chinese KN95 respirators. The respirators were distinguished into three categories. Two KN95 respirators reached the 95% CFE value (passed category, green in Fig. 3), and these two exceeded it considerably with ~98% (like in the case of the reference FFP2 respirators in Table 3). Four KN95 respirators exceeded the 90% CFE (almost passed category, yellow in Fig. 3). Four KN95 respiratory were under 90% CFE; one showed only 64% CFE (failed category, red in Fig. 3). In the failed category, the deviation of the filtering efficiencies was usually large, which also indicates the quality problems.

Analyses of the KN95 masks in
The particle filtering efficiencies (PFE) of the U1-10 "unknown" filtering materials were also investigated ( Table 4) to find the relation between CFE and PFE performance of the respirators. In the case of the passed respirators (green in Fig. 3), the PFE values were over 95% on each particle size range. In the case of the almost passed respirators (yellow in Fig. 3), the PFE of the 0.3-0.5 µm range never reached 95%. It was between 75 and 90%. Since the majority of the particles are in this size range, it caused the unlikely decrease of the CFE values below 95%. The PFE values of the larger particle size ranges (>0.5 µm) were over 95%. In the case of the failed masks (red in Fig. 3), the filter materials of

Conclusions
A fast and easy measurement method was developed to determine the filtering efficiency of respirators applied during COVID-19 pandemic. It is based on laser particle counting. The different validation tests proved that our measurement method could determine the concentration filtering efficiency (CFE) of the filter materials of the respirators similarly to the EN 149:2001 standard. Furthermore, our method can provide the particle filtering efficiency (PFE) on six particle size ranges, which is additional information compared to the flame photometry. The PFE values show the strengths and weaknesses of the respirators in given particle size ranges, so PFE could help to develop the respirators' filtering efficiency further. Considerable advantages of our measurement method are simplicity, availability, and the much lower price compared to the flame-photometer based methods. The analysis of the ten different types of chinse KN95 respirators showed that only two types could reach the theoretical 95% CFE value. Four types were over 90%, and the last four were under 90%. The failed respirators could not perform acceptable CFE in the 0.3-1 µm particle size range. These results suggest the more strict control of the KN95 respirators on the EU market. The work will be continued by the extension of our method with breathing resistance and inward leakage measurements and by supporting nano-fiber filtering medium researches with filtration measurements.

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.