Development of Particle Filters for Portable Air Purifiers by Combining Melt-Blown and Polytetrafluoroethylene to Improve Durability and Performance

,


Introduction
East Asian countries, including China, face immense challenges due to the recent increase in submicron particle (0.02-0.113 μm) concentrations associated with rapid industrialization [1,2].These submicron particles are mainly attributable to the burning of gasoline and diesel fuel, black carbon, tobacco, photochemical reactions, and oil particles released by cooking.The indoor submicron particle concentration is also influenced by the infiltration of submicron particles of an outdoor origin into the indoor environment [3].High concentrations of submicron particles are a health risk factor and a threat to the environment [4][5][6][7].In addition, harmful microbes, including bacteria, fungi, and viruses such as SARS-CoV-2, MERS-CoV, and SARS-CoV-1, have been reported as submicron particles [4,[8][9][10].
To curb indoor submicron particle concentration effects, the use of air purifiers has gained popularity in both residential and commercial spaces [4].Several studies have demonstrated that air purifiers are an efficient method for improving indoor air quality.A portable high-efficiency filter can be used as a treatment for patients with asthma [11].In addition, previous studies have reported significant improvement in some allergy symptoms through the use of a high-performance air purifier [12,13].However, several inefficient air purification systems have penetrated the market, highlighting the need to evaluate the quality of air purifier systems.In addition, understanding of air purifier performance is affected by a lack of real-use experiments.Currently, the performance of air purifiers is not measured in end-user settings.Each air purifier is tested for its clean air delivery rate (CADR) in testing chambers in each country according to its own standards, such as ANSI/AHAM AC-1-2006 [14], GB/T18801-2015 [15], and SPS-KACA002-132 [16], and the size and range of particles are defined differently in each country.
The performance of an air purifier is primarily determined by the type and quality of the filter.Filters fall into two popular categories: fiber filters and membrane filters [5].Air purifiers with high-efficiency particulate air (HEPA) filters have gained popularity because of their simplicity and high efficiency [5,7,17,18].However, many commercial air purifiers that work in the same way as HEPA units but are so-called HEPA-type electret fiber filters (e.g., charged melt-blown (MB) filters), do not comply with the same standards [4].Although HEPA-type filters have a high CADR, their collection efficiency decreases sharply as particles are loaded into the filter fibers [19][20][21].This may be attributable to fiber charge neutralization, filter charge screening, and chemical interactions between the aerosols and the filter material [19,21].In contrast, membrane filters generally have a relatively high solid fraction, which provides high efficiency [5,[22][23][24].However, particles larger than a regular void space may cause a drastic pressure drop, leading to a greater dependence on surface filtration than depth filtration [5,25].Although filters are a key determinant of air purifier efficiency and provide an opportunity for performance improvement, studies on the efficiency of air purifiers with different filter types are limited.
This study is aimed at developing a filter with an optimized combination medium to achieve low differential pressure and an extended lifespan.Such efforts would help overcome the problem of rapid performance deterioration experienced with conventional charged MB filters and extend the replacement cycle of filters, thereby mitigating environmental pollution caused by filters.To assess the performance of the filter we developed, we conducted a comparative evaluation with existing filters, including charged MB, glass fiber, and polytetrafluoroethylene (PTFE) filters.Additionally, we aimed to evaluate any performance differences of the new filter before and after particle loading, comparing it with existing filters manufactured in this study.These filters were subsequently installed in identical commercial air cleaners to investigate their impact on the air cleaner's CADR, emphasizing the critical necessity for efficient air purification systems in the current market.These results are expected to serve as a foundation for providing filters that implement space-specific optimized performance in existing air purifiers, thereby laying the groundwork for providing occupants with safe and comfortable air, countering the indiscriminate use of existing air purifiers.

Methods
2.1.Flat Media and Pleated Filter Specifications.The specifications of the flat filter media, that is, the charged meltblown (MB), glass fiber (GF), PTFE membrane (PT), and PTFE-MB combined media (PM), are listed in Table S1.
The GF, MB, and PT filters are commercially available.The modified composite media was coupled to a spun bond (30 g/m 2 ) material on both sides of the PTFE membrane, and a further charged MB (20 g/m 2 ) layer was designed to bind to one side of the outer surface.The PTFE manufacturing process and the lamination process are shown in Figure 1.The flat media test was conducted according to the modified US 42 CFR Part 84 [26].US 42 CFR Part 84 is one of the regulations provided by the Centers for Disease Control and Prevention (CDC) in the United States, encompassing aerosol filtration efficiency, pressure drop, and user comfort for respiratory masks made of fibrous filter materials [27,28].At a face velocity of 5 cm/s, the particle removal efficiency determined using an automated filter tester (model 8130, TSI, Inc., Shoreview, MN, USA) of all the media used in this experiment was more than 99.95%.The percentage of particle penetration and pressure drop were measured at a flow rate of 32 L/min using a NaCl aerosol.Five samples of each flat filter medium were tested for particle penetration.For all types of media (MB, GF, PT, and PM), we produced a pleated filter with a size of 277 mm × 361 mm × 35 mm and area of 1.84 m 2 , as shown in Figure 2.

Measurement of Initial Filter Penetration and
Differential Pressure.Filter penetration and pressure drop were measured using the cabin air filter test system PAF-111 (Topas GmbH, Dresden, Germany).The experimental system and procedures were in accordance with ISO 11155-1, which is the standard for automotive passenger seat filters [29].The experimental apparatus was a modular system.The system was divided into a particle-generator section and a test section.In the particle-generator section, air passes through an EU13 HEPA filter, and then the particles exit the atomizer aerosol generators (model ATM 210, Topas GmbH, Dresden, Germany).The testing was performed using fluid diethylhexyl sebacate (DEHS) particles.In the test section, the upstream and downstream particle removal rates were calculated as the particles passed through the test filter using an optical particle counter (LAP 322, Topas GmbH).The differential pressure drop (ΔP) test measures the ease of air passage through the filter using the upstream-todownstream pressure drop at a flow rate of 1 m/s.The experiments were repeated three times for each filter.

Filter Loading
Test.The test particles were loaded using the user-defined dust loading test mode of the PAF-111 (Topas GmbH) to determine how the performance changed when particles were loaded into the filter.Many particles found in the environment, such as those produced from cooking, black carbon, oil, and dust, were considered when selecting DEHS (diethylhexyl sebacate) (<1 μm) and ISO 12103-1 A2 dust (0.97-176 μm) as representative test particles for evaluating particle loading [30,31].The reason for this selection is that ISO 12103-1 A2 represents coarse particles, while DEHS serves as a surrogate for oil particles.The reason for using standardized particles is that the amount of test dust must be uniform to ensure a constant value of dust concentration.Therefore, it is known that standardized experimental particles should be used in filter

2
Indoor Air loading tests where reproducibility is required for these reasons [32].The A2 fine dust was dispersed by dust dispersers (SAG410, Topas GmbH) at 20 g for 30 min, and the solid particles were neutralized using a bipolar aerosol neutralizer unit (EAN 581, Topas GmbH).DEHS particles were obtained from the atomizer aerosol generators (ATM 210, Topas GmbH) and were loaded at a speed of 2 g/30 min.
For the particle loading, solid particles were loaded onto the filter, and then liquid particles were loaded.One cycle lasted for 60 min.While the particles were being loaded, the change in the collection efficiency was measured using an optical particle counter (LAP 322, Topas GmbH) upstream and downstream of the test filter.Sensors were used to measure the change in the pressure drop across the filter.The loading experiment was implemented at a flow rate of 200 m 3 /h, and the filter performance was evaluated at 0.2, 0.5, and 1.0 m/s.To obtain representative and reliable results, the experiments were repeated three times for each of the four filter media (MB, GF, PT, and PM).The performance and loading test process are shown in Figure 3.

QF Calculation.
To evaluate the change in performance over time based on each filter's features, the quality factor (QF), or figure of merit, an established effective comparison tool for various filter types, was used to compare the performance before and after filter use [33][34][35].The QF equation was determined as follows: Here, P is the particle penetration (fraction of particles passing through a filter) and ΔP is the pressure drop (Pa) of the entire filter.The QF has a higher value when the filter's pressure drop is lower, and vice versa.

Specifications of the Air Purifiers.
A portable air purifier (AP-1018F, Coway Co., LTD, South Korea) (Table S2), which is available in both online and offline stores, was selected for the CADR performance tests of the MB, GF, PM, and PT filters.The filter system was composed of a prefilter, deodorant filter, and particle filter at the time of purchase.The product was certified with a Clean Air Certification from the Korea Air Cleaning Association, and the official CADR result was 4.28 m 3 /min.In this study, the prefilter and the experimental filters (MB, GF, PM, and PT) were 3 Indoor Air used for the tests.From among the four working modes, the CADR tests were conducted in the turbo mode.
2.6.Experimental Measurements.The CADR test was conducted according to the SPS-KACA 002-0132:2018 standard [16] (Figure S1).The performance of an air cleaner is evaluated by its clean air delivery rate (CADR), which is defined as the measure of its delivery of contaminant-free air in cubic meters per minute, with all particles of a given size distribution removed.The chamber size was 30 m 3 , following the Korea Air Cleaning Association (KACA) standard (4 0 m L × 3 0 m D × 2 5 m H ). The chamber temperature was 23 ± 5 °C, and the relative humidity was maintained at 55 ± 15%.Potassium chloride (KCl) was used as the test particle, and its initial particle concentration was 1 0-3 0 × 108 particles/m 3 .An aerosol spectrometer (Model 11-A, Grimm Aerosol Technik GmbH & Co. KG, Germany) was used to measure the particles, resulting in a 0.3 μm concentration decay.The sampling interval was one minute.Both natural decay and operational decay tests were conducted for 20 min each or until the concentration reached onetenth of its initial value.In the operational decay experiment, the first sampling point (t = 0) was taken after the air purifier had been running for three minutes.
We measured the submicron (0.02-0.113 μm) CADR (sCADR) to verify the difference in submicron particle removal performance among the filters.When an air purifier's sCADR was measured, submicron KCl particles were created using a nanoparticle generator (EP-NGS20, EcoPictures Co., Ltd., South Korea).Then, the submicron particles were measured using an electrostatic classifier (model 3082, TSI, Inc., Shoreview, MN, USA), an ultrafine condensation particle counter (model 3776, TSI, Inc.), and a differential mobility analyzer (model 3085, TSI, Inc.).The submicron KCl particles were sprayed for 10 min and sampled in 1 min intervals.The natural decay and operational decay were calculated based on SPS-KACA 002-0132:2018.Both CADR and sCADR tests were conducted.
The following equation was used to calculate the purifying capacity (Pc) of the CADR and sCADR: Here, V is the volume of the chamber, N is the number of experimental air purifiers, C i1 is the particle concentration at the beginning of the measurement (t = 0) when the capacity decreases naturally, C i2 is the particle concentration at the beginning of the measurement (t = 0) when the operation decreases, C t1 is the particle concentration for each measured t minute when the capacity decreases naturally, and C t2 is the particle concentration for each measured t minute when the operation decreases.
The operating electrical power was measured using a digital power analyzer (WR310, Yokogawa Electric Corp., Tokyo, Japan).Before the power measurements, the air purifier was turned to the maximum setting and allowed to operate for 5 min.The power indicator of the digital power analyzer was adjusted to 220 V (60 Hz), and the watt readings were recorded for 10 min at 1 min intervals.

Statistical Analysis. Regression analysis was performed
to determine the correlation between the CADR and energy consumption efficiency, and the correlation coefficient (r) was obtained using SPSS version 22.0 (SPSS Inc., Chicago, IL, USA), where a p value of less than 0.05 was considered statistically significant.

Filter Performance according to Particle Loading by Filter
Type.The filter performance and pressure drop varied with the material in Figure 4, and Figure S2 shows the changes in penetration rate and pressure drop after particle exposure.After exposing the filters to DEHS and A2 particles for 30 min each, the pressure drop increased for all filters.The PTFE filter exhibited the highest pressure drop.The MB filter showed an increased penetration rate, but the GF, PM, and PT filters showed decreased rates.The MB filter had the highest initial penetration rate (0.0118) and the lowest pressure drop (68 Pa).The pressure drop increased by 88.2% after 6 cycles of loading at 1 m/s, with a 30% increase in penetration rate (Figure 4(a)).The GF filter had a high initial penetration rate (0.0005) but the highest pressure drop (384 Pa).After six cycles, the pressure drop increased by 26.6% at 1 m/s and the penetration rate decreased by 0.04% (Figure 4(b)).The PM filter had the same initial penetration rate (0.0005) as the GF filter, but it had a lower pressure drop (165 Pa).After six cycles, the pressure drop increased by 91.5% at 1 m/s (Figure 4(c)).The PT filter had an initial penetration rate of 0.001 and an average pressure drop of 210 Pa.After six cycles, the pressure drop increased by 124.2% at 1 m/s and the penetration rate decreased by 0.04%.As the pressure drop of the PT filter increased, the particle removal efficiency also increased (Figure 4(d)).Figure 5 illustrates the changes in the performance of each filter as a result of particle loading using the QF values (Table S3).All filters used in the experiment were investigated with the lowest QF values in the range of 0.201-0.245μm.MB had the highest initial QF value (Figure 5(a)).The QF values were in the same order as the initial measurement results, i.e., MB, PM, PT, and GF, respectively.However, there was a decrease in the particle size range of 0.835-1.17μm, which had a high QF value of 46.3% for MB, 39.5% for PT, 25.7%    Indoor Air for GF, and 22.0% for PM (Figure 5(b)).The average QF values were higher in the order of PM, MB, PT, and GF, with the lowest QF value of 0.0079 Pa −1 for the MB filter in the range of 0.201-0.245μm (Figure 5(c)).As a result, the QF value increased in the order of PM, PT, GF, and MB.The greatest change observed in the MB filter was 90.4% lower than the initial value in the range of 0.201-0.245μm (Figure 5(d)).

Comparison of CADR for Different Particle Diameters.
The CADR was standardized to a particle size of 0.3 μm in accordance with SPS-KACA002-132 standards, and the sCADR (0.02-0.113 μm) was also measured for comparison.The most penetrating particle size (MPPS) for each filter was estimated through a CADR experiment using various particle sizes.The MB and PM filters recorded the lowest CADR when the particle size was 0.0914 μm, whereas the GF and PT filters recorded low values at 0.1018 μm.
Before particles (0 cycles), the CADR for the MB filter, based on a particle size of 0.3 μm, reached a maximum of 4.83 m 3 /min, and the sCADR was similar or lower.Under the same conditions, the GF filter had a CADR of 1.54 m 3 / min, the PM filter had a CADR of 3.62 m 3 /min, and the PT filter had a CADR of 3.22 m 3 /min (Figure 6(a)).After   Indoor Air particle loading for two cycles, the CADR of the MB filter decreased by approximately 3.6% in comparison to the initial value, whereas the GF, PM, and PT filters decreased by 18.1%, 9.5%, and 29.5%, respectively (Figure 6(b)).The sCADR decreased similarly.After particle loading for four cycles, the PT filter showed the greatest drop of 41.2% compared to the initial value.The PM filter value decreased by 23.0%, whereas the GF and MB filters decreased by 18.4% and 10.9%, respectively (Figure 6(c)).After particle loading for six cycles, the performance of the MB filter decreased by 21.2% in comparison to the initial state, whereas the performance of the GF, PM, and PT filters decreased by 25.3%, 31.4%, and 53.1%, respectively (Figure 6(d)).

QF and CADR Values.
The results of the regression analysis between QF and CADR are shown in Figure 7.The correlation was high for each filter type; however, the characteristics were different for each filter type.The slope of the linear equation for the MB filter was lower than that for the other filters.This is the case because the MB filter's QF decreased by 84%, and the air purifier's CADR decreased by 21.2% after particle loading.However, there was a high correlation coefficient (r = 0 878, p < 0 0001) between the QF and CADR for the MB filter.For the GF filter, the changes in the QF and CADR rates were 22.5% and 25.3%, respectively, and the correlation analysis showed a significant correlation (r = 0 767, p < 0 05) between QF and CADR.The slope of the PM filter was 48.82, a value between those of the MB and PT filters, and the correlation between the QF and CADR was high (r = 0 958, p < 0 0001).The QF change for the PT filter was approximately 55%, that for the CADR was 53.1%, and the correlation was high (r = 0 962, p < 0 0001).

CADR Performance after Particle Loading considering
the Energy Efficiency of each Filter.The CADRs after particle loading, considering the energy efficiency of the air purifier, were altered (Figure 8).The CADR per 1 watt for the MB filter remained relatively stable at 0.144 m 3 /min after six cycles of particle loading.However, the GF filter had the lowest initial CADR of 0.125 m 3 /min•W and experienced the greatest decline of 33.3% (0.083 m 3 /min•W) after six cycles of particle loading.The PM filter had the highest initial CADR of 0.151 m 3 /min•W, which decreased by 8.3% (0.138 m 3 /min•W) after particle loading.The PT filter had an initial CADR of 0.148 m 3 /min•W, which dropped by 28.8% after particle loading.

Discussion
This study evaluated the performance of a researcherdesigned PM filter in comparison to those of commonly utilized air purifier filters (MB, GF, and PT) by examining the changes in the initial capacity after particle loading.The objective was to determine the performance of the PM filter and its commercial viability.
Assessment of filter performance is generally based on two parameters: particle collection efficiency and pressure drop across the filter.The optimal filter offers the highest collection efficiency with the lowest pressure drop.Despite this, it is widely recognized that improving the collection efficiency of air filters leads to a corresponding increase in pressure drop [36,37].Consequently, the QF has become an invaluable tool for comparing the performance of different filters [38].In this study, we conducted an initial measurement of the QF for each filter, followed by an analysis of the evolution of the QF for each particle size distribution as the particles accumulated on the filter.
The results showed that the initial performance order of MB > PM > PT > GF changed to the order of PM > PT > GF > MB after six cycles of loading.In this study, the QF values of various particle size distributions were measured  Indoor Air for each filter.QF values between 0.201 and 0.245 μm were low for all filters, whereas QF values for the particle range of 0.835-1.17μm were high because of the higher particle collection efficiency compared to the pressure drop.The reason for the low particle capture efficiency in the range of 0.201-0.245μm is that this range is the MPPS range of the filter [36,39].The variation in QF value based on particle size distribution was consistent with previous research findings.
Additionally, as reported in earlier studies, pleated filters demonstrated greater fluctuations in QF value in response to particle size distribution than flat filters, thereby supporting our current findings [40].After particles were loaded onto the filters, QF values in the range of 0.201-0.245μm were reduced, with a reduction of 90.4% for the MB filter.These findings align with the results of previous studies that suggest that the static electricity safety of MB filters is 11 Indoor Air dependent on the filter material and aerosol, which can impact the static energy and a potential threat to human safety [41,42].The GF filter demonstrated a depth filtration structure, leading to a slow increase in pressure drop, which is consistent with prior research suggesting that the GF filter has a higher holding capacity than other filters [3].However, the low QF value of the GF filter can be attributed to its fiberglass composition of microfibers [43].The PT filter showed a minimal impact on the penetration rate after particle loading, but the pressure drop increased by 124%, reflecting its low holding capacity [3].In contrast, the PM filter exhibited both surface and depth filtration characteristics, effectively addressing the limitations of both the MB and PT filters.The analysis of the PM filter following particle loading indicated that there might be some benefits in terms of QF.
The performance of air filters is commonly evaluated using the QF value, but when assessing the purification ability of air purifiers, the CADR, which is expressed as the product of the airflow rate and particle removal efficiency, is the preferred metric [40,44,45].There has been a paucity of studies comparing the CADR performance of filters made from developed media for the purpose of evaluating filter lifespan.The present study evaluated the CADR and sCADR of each filter type, and the results showed that the order of MB > PM > PT > GF was not altered before and after particle loading.This finding is in contrast with the QF results.This contrast can be explained by the findings of previous studies, which showed that CADR measures the rate at which fine particles are removed from an enclosed space and that the airflow rate has the greatest impact on CADR performance [45].Therefore, the highest CADR results were obtained in the order of the lowest pressure drop for the filters.The MB filter obtained a higher CADR than the other filters, even after particle loading.This can be attributed to the electret filter method, which utilizes electrostatic attraction and has a lower initial pressure drop than other filters.Furthermore, the MB filter had a higher holding capacity because of the depth filtration method, resulting in a smaller increase in pressure drop even after loading [3,[46][47][48][49][50].We suggest that the reason for the inferior CADR of the PM filter compared to that of the MB filter is that the initial pressure drop was approximately 2.4-fold greater than that of the MB filter because of the combined PTFE and MB materials.Although the dust collection efficiency of the MB filter fabric was projected to deteriorate rapidly and impact the CADR after six cycles of particle loading, the primary factor that affected the CADR performance was the pressure drop rather than the dust collection efficiency.Therefore, further investigations are necessary to minimize the initial pressure drop to the level of the MB filter.
Effective management of indoor air quality necessitates a meticulous selection of the most suitable filters tailored to specific indoor environments.Establishing appropriate criteria that align with the indoor setting becomes imperative to determine whether to prioritize filters with high initial performance or opt for those with lower initial performance but extended lifespans.Notably, variations in clean air delivery rate (CADR) attributed to different filter materials further underscore the importance of selecting filters tailored to the prevailing situa-tion and environment.A pivotal representative criterion for filter selection involves cost-benefit analysis, considering both filter performance and expected lifespans.Additionally, in the context of optimizing filter selection, cost-reference particle filtering proves advantageous, leveraging sequential Monte Carlo (SMC) techniques to estimate the states of discretetime dynamic random systems [51,52].However, such an analysis was not conducted in this study.Therefore, further research is needed to derive an optimal method for selecting suitable filters, enabling the resolution of dynamic optimization problems using state models such as sequential Monte Carlo and other models to account for cost, performance, and lifespan considerations.
The sCADR results were similar to the CADR results because the materials used in the experiments, including PTFE, GF, and MB, have the capacity to eliminate submicron particles through most particle removal mechanisms.However, the lowest CADR in the range of 0.08-0.09μm was measured as a result of this experiment.Presently, the particle size range for CADR differs from one standard to another, but it typically is in the range of 0.15 to 0.3 μm, similar to the MPPS range of filters [36,39].This finding suggests that the MPPS range for the product may be distinct from that of the filter, necessitating further research into the new measurement range of CADR.
The correlation between the QF value and CADR was relatively high, but the slope varied depending on the filter's dust removal mechanism or material.Therefore, relying solely on the QF value may not accurately predict the air-purifying performance of an air purifier because the CADR result value, which is derived based on the filter material or filter specification, can have a different distribution than the QF value.
This study has some limitations.First, the media used were not optimized for the bending angle or filter media area based on its thickness or characteristics.Additionally, while manufacturing the filters, the performance of filters of different sizes was not compared when fixing the horizontal and vertical thickness standards of the filters.Furthermore, the experiment was limited to an air purifier with a CADR of 4.3CMM, so the performance change at high and low CADR could not be evaluated.During the QF evaluation, submicronsized particles were not evaluated; therefore, a comparative evaluation with sCADR was not possible.Moreover, the filter's durability characteristics against KCl, NaCl, and diesel particles, other than the DEHS and A2 particles used for filter loading, could not be analyzed.Finally, the study was limited because the flow path structure of the product and the type of fan was not optimized based on the filter pressure drop.

Conclusion
In conclusion, meticulous filter selection is crucial for effective indoor air quality management.The newly developed PM filter shows promise in extending the filter lifespan while maintaining air purification efficacy.This suggests that it may serve as a substitute for the MB filter, which is used commercially but exhibits a rapid decline in particle collection efficiency after particle loading.Thus, if employed as an air purification device filter in medical institutions, childcare 12 Indoor Air centers, elderly care facilities, and other multiuse facilities requiring long-term indoor air quality management, the filter has the potential to maintain indoor air quality over an extended duration.However, its introduction may lead to a lower CADR, necessitating further research for CADR enhancement strategies.Future research should focus on optimizing filter design and performance to effectively address dynamic indoor air quality management challenges, underscoring the need for ongoing investigation in this area.

Figure 1 :igure 2 :
Figure 1: The test bench diagram for development of new flat media.

Figure 3 :Figure 4 2 .
Figure 3: Test bench diagram for performance and loading test process.

Figure 4 :
Figure 4: Penetration and pressure drop of the filters loaded with DEHS and A2 as a function of time.(a) Melt-blown (MB) filter, (b) glass fiber (GF) filter, (c) PT + MB (PM) filter, and (d) PTFE membrane (PT).

Figure 5 :
Figure 5: The performance of QF evaluated using the ExpDec2 model, which examines the changes in performance between the initial value and particle loading for various particle sizes in the melt-blown (MB), glass fiber (GF), PT + MB (PM), and PTFE membrane (PT) materials.(a) The initial cycle (without loading), (b) the second cycle, (c) the fourth cycle, and (d) the sixth cycle.

Figure 6 :
Figure 6: CADR performance of each filter depending on particle size and loading time: (a) the initial cycle (without loading), (b) the second cycle, (c) the fourth cycle, and (d) the sixth cycle.Data points represent the average of three repetitions, and error bars indicate standard deviations.

Figure 7 :Figure 8 :
Figure 7: Performance correlation between QF values of filters and CADRs for the particle size of 0.3 μm: melt-blown (MB), glass fiber (GF), PTFE membrane (PT), and PT-MB combined media (PM).Data points represent the average of three repetitions, and error bars indicate standard deviations.