Performance of a new model of nasal filter when challenged against PM1 aerosols

ABSTRACT

Nasal filters are increasingly used in reducing personal haze exposures, but there is no retrievable evidence on its efficiency. In this study, a nasal filter testing system was designed. NaCl aerosols were generated to simulate PM1. Three different sizes of a new model of nasal filter were tested under two flow types (constant vs. cyclic) and three flow rates (15, 30, and 50 L/min). All PM1 penetration values were > 80%. Either enlarging filter size or lowering flow rate significantly reduced PM1 penetration, and PM1 penetration under cyclic flow was significantly higher compared with constant flow. To avoid mouth breathing, nasal filters could only be worn under light work-rates or rest conditions. The quality factor of nasal filters was only hundredths that of the N95 and P100 FFRs. This study reveals that there is a future need to design nasal filter materials with higher efficiency and lower flow resistance.

KEYWORDS:

• Haze
• air pollution
• pm1
• respiratory protection
• nasal filter

Introduction

In China, due to rapid urbanization and sharp increase of fossil energy consumption, more than 99% of the top 500 largest cities, at least 30% land areas, and nearly 800 million people are suffering from frequently (>100 days per year) severe haze pollution [1]. Various studies have shown that air pollution may increase respiratory diseases, cardiovascular and cerebrovascular cases, lung cancer deaths, and overall mortality [1–5]. Moreover, it was reported that around 1.2 million premature deaths per year were caused by air pollution in China, accounting for nearly 40% of this kind of global deaths [6]; correspondingly, the health-related economic losses caused by haze pollution were estimated to be as high as 87 billion dollars per year [6]. Air pollution or PM2.5 (particulate matters with an aerodynamic diameter ≤ 2.5 μm) related adverse health effects in other countries such as America were also not optimistic. It was estimated that air pollutants cause 200,000 early deaths per year in the U.S [7].

Faced with such a grim situation, global governments and scientists have devoted themselves in searching for the ‘real killer’ that threatens human health. Long-term epidemiological studies revealed that there is a close relationship between the mortality of urban residents and the size depended atmospheric particulate concentrations; the smaller the particle size the higher the mortality would be [8]. As a subpart of PM2.5, PM1 (particulate matters with an aerodynamic diameter ≤ 1 μm), with the range of size covers the wavelength of visible light (0.4–0.7 μm), played an important role in daylight extinction and may be the key factor in the formation of haze [9]. It was found that the PM1 mass concentration during haze days was several times higher than that of ‘non-haze’ days [10]; meantime, haze pollution was an indicator of the existence of high number concentration of PM1 [4,11]. Compared with PM2.5, PM1 features a larger surface area-to-mass ratio, thus carrying more toxic pollutants (such as heavy metals, PAHs, bacteria, etc.); with a smaller particle size, PM1 could penetrate deeper into human lungs [12]. It has been concluded that the PM1 constituted most of the ambient particulates (both mass and number concentrations) during haze days [10,11], and there is a statistical association between adverse health effects and PM1 [13]. Therefore, the high number concentration of PM1 may be the ‘real killer’ during the occurrence of haze pollution [4].

Since haze pollution affects all population, the concept of respiratory protection has been extended from industrial sectors and healthcare environment to the whole society, everyone needs to choose an effective respiratory protective equipment (RPE) during haze days’ commuting, especially those who work outdoors such as traffic policemen, construction workers, street cleaners, etc. Traditionally, face masks such as filtering facepiece respirators (FFRs) are widely used by people to reduce air pollutant exposure. However, there are several limitations associated with FFRs that may deter common people from wearing them: 1) FFRs are not comfortable to use [14]; 2) face masks may cause fogs on eyeglasses especially in winter days; 3) there is a muffling effect on voice and influence on cosmetic.

Since people breath 90% of air through their nose [15], nasal filters are developed to augment the filtering efficacy of nose by inserting two pieces of miniature filter into the nasal cavity. Compared with traditional face masks, nasal filters feature several advantages: 1) nasal filters are invisible (or less obvious) from outside which makes them more cosmetically appealing; 2) wearing nasal filters could avoid the problem of fogging on glasses; 3) the muffling effect on face-to-face communication can be lessened. In fact, it has been demonstrated that the foreign body sensation caused by nasal filters was mild/minimal; although some of the wearers may experience a nasal twang, the influence generally disappeared within 1–2 days [16]. Therefore, nasal filters may be a practical and comfortable alternative to traditional face masks to prevent against air pollution.

Although it is claimed by the manufactures that nasal filters can effectively protect against air pollutants (filtration efficiency>90%), till now, except for several investigations regarding the effectiveness of nasal filters in reducing pollen allergy and improving life quality of wearers [17,18]; there is no traceable evidence on the efficacy of the nasal filters in protecting against PM1, which is assumed to be the ‘real killer’ during haze days. Therefore, utilizing a self-designed nasal filter testing system, this study aimed at assessing the performance of a new model of nasal filter when challenged with PM1 aerosols.

Materials and methods

Tested nasal filters

One new model of nasal filter designed to protect against air pollutants was tested in this study. As presented in Figure 1, the nasal filter consists of twin flexible filter holders that are bridged by a thin translucent plastic bar (connector). There is a retainer ring in each frame to secure the filter. The twin filter membranes are made of electret polypropylene fibers to enhance the filtration efficiency. The nasal filter membrane in shape is similar with human nostril, with a thickness of 1 mm. The two filter holders are soft and comfortable that can be inserted into both nostrils. The connector that holds the twin filters in place could prevent the filters from being pushed too far into the nasal vestibule, which also provides a means of insertion and removal of the nasal filter. Since the connector is translucent and located at the bottom of the nose, it is almost virtually invisible. Above stated filter material and design was quite similar for the tested and many other brands of nasal filters. In this study, to accommodate various sizes of nostrils, three different sizes of the tested nasal filter (one pair each: small, 88 mm² × 2; medium, 110 mm² × 2; large, 134 mm² × 2) were purchased from amazon.com, and under each combination of experimental condition, a brand-new nasal filter was tested. Therefore, six specimens of each size and totally 18 nasal filters were tested in this study.

Figure 1. Schematic diagram of the experimental set-up [modified from 23]

Challenge aerosols

Sodium chloride (NaCl) aerosols (0.02–1 μm) were used as the challenge surrogate for PM1. The submicron NaCl aerosols were generated by a constant output particle generator (Model 8026, TSI Inc., Shoreview, Minn., USA) using 2% NaCl solution and fed into a nasal filter testing chamber (see Figure 1). The testing chamber is 1.2 × 1.2 × 1.2 m³ and constructed by synthetic resin. Four fans were placed at the four corners of the chamber to circle the air to assure the generated aerosols were uniformly distributed. During each test, the particle generator was operated continuously at an output flow of 0.5 L/min to maintain a stable concentration level of 30,000 ± 300 particles/cm³, representing the average PM1 number concentration in the urban area of Beijing [11].

Experimental set-up

The performance of the tested nasal filters was assessed in two aspects: 1) the filtration efficiency, which was determined by the PM1 penetration, and 2) the comfort of the filter, which was represented by the breathing resistance (R_f). National Institute for Occupational Safety and Health (NIOSH)-approved FFRs are certified under 85 L/min constant flow [19]. However, actual human breathing patterns more closely resemble a sinusoidal waveform (cyclic flow) [20]. Furthermore, several studies reported that at the same test flow rates, particle penetration under cyclic flow may be higher than that of constant flow [21–23], which suggests that the cyclic flow may be a more stringent test flow regime. For a FFR, it is typically associated with a higher R_f, leading to the well-known unwillingness of the users to wear [14]. A good model of nasal filter should combine high efficiency with low resistance; however, the resistance and filter efficiency often conflict with each other. Comprehensively considering these two, the overall performance of a filter can be assessed by the quality factor (q_f) [24]. Higher q_f value corresponds to better overall performance of nasal filter.

Based on the analyses above, a nasal filter testing system was modified from 23, as shown in Figure 1. The tested nasal filters were inserted into two thin copper tubes (simulated as different size human nostrils). The interface between the frame of the filter and the wall of the tubes was fully sealed by silicone. The ends of the two thin copper tubes were connected to a large diameter copper pipe (simulating the upper human respiratory tract). The pipe was then connected to either a vacuum system or a breathing simulator (Series 1101, Hans Rudolph, Inc., Shawnee, KS, USA) to introduce constant or cyclic test flow. Aerosol samples (0.02–1 μm) from outside and inside the tested nasal filter were taken by a P-Trak Ultrafine Particle Counter (Model 8525, TSI Inc., Shoreview, Minn., USA) for PM1 number concentration analysis. Detail specifications for related equipment in this study were presented in Table 1.

Table 1. Detail specifications for related equipment in this study

No.	Equipment	Manufacturer	Model	Measuring Range	Accuracy
1	Particle generator	TSI Inc.	8026	N/A	N/A
2	Particle counter	TSI Inc.	8525	0 ~ 5 × 10^5 particles/cm^3	±1 Particle
3	Manometer	TSI Inc.	PVM100	−3500 ~ 3500 Pa	±1 Pa
4	Breathing simulator	Hans Rudolph, Inc.	1101	−200 ~ 200 L/min	±2%
5	Flowmeter	TSI Inc.	4045	0 ~ 300 L/min	±2%

As expressed by Equation (1)(1) $D_{pipe}=D_{sample}\times \sqrt{\frac{Q_{test}}{Q_{sample}}}$(1) [24], to achieve isokinetic sampling results, based on the sampling flow of the P-Trak (Q_sample, 0.1 L/min) and the inner diameter of the sampling tube (D_sample, 6 mm), different inner diameters of copper pipes (D_pipe) 73, 104 and 134 mm were designed and applied to match different test flow rates (Q_test) 15, 30, and 50 L/min.

(1) $D_{pipe}=D_{sample}\times \sqrt{\frac{Q_{test}}{Q_{sample}}}$(1)

Test conditions

All three sizes of the nasal filter were tested under the fully sealed condition. To better simulate various human breathing efforts, sinusoidal cyclic flows with mean inspiratory flow (MIF) rates of 15, 30, and 50 L/min (breathing frequency = 15, 20, and 25 breaths/min, respectively) were applied [25,26]. To investigate the effect of the flow type on the PM1 penetration, the above-specified three flow rates were also performed at constant flow regime. A summary of the test conditions was shown in Table 2. For each tested nasal filter, the two flow types, three flow rates, and four replicates were fully randomized to minimize the experimental error. The PM1 penetration, R_f, and q_f were calculated by the following Equations (2)(2) $P\mathrm{\%}=\frac{C_{in}}{C_{out}}\times 100\mathrm{\%}$(2) -(4(4) $q_f=-ln(P\mathrm{\%})/R_f$(4) ) to assess the protection, comfort, and overall performance of the nasal filters, respectively.

Table 2. Summary of the experimental conditions

Variable	Levels
Nasal Filter	Small (88 mm^2); Medium (110 mm^2); Large (134 mm^2)
Flow Type & Rate	Constant flow (15, 30, 50 L/min); Cyclic flow (15, 30, 50 L/min)
Challenge Aerosol	0.02–1 μm NaCl particles
Replicates	4
Total Runs	3 filter sizes × 2 flow types × 3 flow rates × 4 replicates = 72

PM1 penetration (P%)

The PM1 penetration (P%) was determined as the ratio of inside (C_in) to outside PM1 concentration (C_out) [27–29]:

(2) $P\mathrm{\%}=\frac{C_{in}}{C_{out}}\times 100\mathrm{\%}$(2)

where C_in and C_out were the PM1 number concentration inside and outside of the tested nasal filters, particles/cm³.

Breathing resistance (R_f)

R_f was determined as the difference between the pressure outside (P_out) and inside (P_in) of the nasal filter when constant flow was applied [27–29]. The R_f was obtained by a manometer (AirflowTM PVM100, TSI Inc., Shoreview, Minn., USA).

(3) $R_f=P_{out}-P_{in}$(3)

where P_in and P_out were the air pressure inside and outside of the tested nasal filters, Pa.

Quality factor (q_f)

Comprehensively considering PM1 penetration and R_f, the overall performance of a nasal filter was assessed by the q_f [24,29]. The higher the q_f value, the better the overall performance of a nasal filter:

(4) $q_f=-ln(P\mathrm{\%})/R_f$(4)

where q_f was the quality factor of the tested nasal filters, Pa⁻¹.

Data analysis

For each combination of the experimental conditions, the mean value and standard deviation of the PM1 penetration, R_f, and q_f were calculated from the four replicates. Normal distribution test was performed. Three-way Analysis of Variance (ANOVA) was applied to investigate the effects of filter size, flow type, and flow rate on the PM1 penetration. Two-way ANOVA was used to study the significance of filter size and flow rate for the R_f and q_f. Pairwise comparisons were performed by Tukey’s range test. All statistical data analysis was conducted using SAS version 9.3 (SAS Institute Inc., Cary, NC). p-values < 0.05 were considered significant.

Results

PM1 penetration

The mean PM1 penetration values found for all nasal filters under different testing airflows were presented in Table 3. Three-way ANOVA revealed that filter size, flow type, and flow rate all had a significant effect on the PM1 penetration (p < 0.0001). As can be seen in Table 3, all PM1 penetration values were >80%; for the large size nasal filter (performed the best), the minimum penetration under the 15 L/min constant flow was around 83%, while the corresponding value at 15 L/min cyclic flow was nearly 90%. The reported data suggested that the tested nasal filters had a limited protection efficiency against PM1 aerosols.

Table 3. Mean PM1 penetration for three sizes nasal filters under different testing airflows

Flow Type	Flow Rate (L/min)	Mean Penetration, % (Mean ± SD)
		Small	Medium	Large
Constant	15	87.04 ± 1.68	85.84 ± 0.80	83.32 ± 0.45
	30	93.58 ± 1.13	87.31 ± 0.68	85.59 ± 0.83
	50	96.47 ± 1.18	92.84 ± 1.45	87.61 ± 0.51
Cyclic	15	96.31 ± 0.08	92.25 ± 0.49	89.51 ± 0.22
	30	97.78 ± 0.54	94.59 ± 2.06	92.45 ± 0.90
	50	99.27 ± 0.77	96.35 ± 0.63	93.28 ± 1.37

The pairwise comparison results for three sizes nasal filters, two flow types, and three flow rates were presented in Table 4. As expected, the PM1 penetration values obtained from the three filter sizes were all significantly different from each other (p < 0.05), the larger the filter size the lower the PM1 penetration value, indicating that filters with a bigger surface area provided a lower face velocity, and therefore a higher protection level. In addition, the PM1 penetration values under cyclic flows were found to be significantly higher than that of constant flows (p < 0.05), which was consistent with previous studies [21–23]. This finding suggests that the sinusoidal waveform, which better reflects a real human breathing pattern, is a more stringent and realistic flow regime to test the efficacy of a nasal filter. Meantime, it was found that under either constant or cyclic flow condition, increasing the test flow rate (or face velocity) would result in a significant increase of the PM1 penetration (p < 0.05). A similar conclusion has been drawn by many other scholars when testing the 95 and P100 FFRs or other types of respirators [22,23,27–32].

Table 4. Pairwise multiple comparisons: mean PM1 penetration among three sizes nasal filters under different testing airflows (ANOVA with Tukey’s Range Test)

Tukey Grouping*	Mean Penetration ±SD (%)	Levels	Variable
A	95.07 ± 5.36	Small	Filter Size
B	91.53 ± 3.54	Medium
C	88.63 ± 3.69	Large
A	94.64 ± 3.07	Cyclic	Flow Type
B	88.84 ± 4.29	Constant
A	94.30 ± 5.15	50	Flow Rate
B	91.88 ± 4.88	30
C	89.04 ± 3.62	15

* Within each group of the variable (filter size, flow type, and flow rate), mean penetration values with the same letter are not significantly different (p-value > 0.05).

Flow resistance (R_f)

R_f is an important parameter to characterize the performance of various types of respirators [27–29]. This study was the first of its kind to quantitively evaluate the R_f of a nasal filter. Unlike traditional respirators, for a nasal filter to function properly, wearers must breathe in through the nose only. However, higher R_f may lead the wearers to breathing through their mouth without any respiratory protection. The R_f values of the nasal filters under different constant flow rates are presented in Table 5. Two-way ANOVA showed that both filter size and flow rate were statistically significant (p < 0.0001). Pair-wise multiple-comparison analysis revealed that for each specific filter size, as the breathing flow increased, the R_f across the nasal filter increased significantly (p < 0.05). On the other hand, increasing the filter size resulted in a decrease in the face velocity and therefore a significant reduce in the average R_f values for all testing flow rates (p < 0.05).

Table 5. Flow resistance (R_f) of three sizes nasal filters under different constant flow rates

Flow Rate (L/min)	Flow Resistance, cm H2O/L/s (Mean ± SD)
	Small	Medium	Large
15	8.1 ± 0.2	3.7 ± 0.0	3.3 ± 0.0
30	11.3 ± 0.1	5.0 ± 0.0	4.6 ± 0.0
50	14.3 ± 0.2	6.6 ± 0.1	6.2 ± 0.0

Previous studies [33,34] have demonstrated that when the nasal airway resistance was greater than 3.5 cm H₂O/L/s, people began to feel difficulties in nasal breathing; when resistance exceeded 4.5 cm H₂O/L/s, a significant number of subjects would inhale through their mouth. As listed in Table 5, even at the 15 L/min lowest test flow (corresponding to the minute air volume required by a medium sized person when performing light work activities, as defined in ISO/TS 16976–1:2015) [35], the R_f across the small size nasal filter was > 4.5 cm H₂O/L/s, suggesting that wearers may tend to execute mouth breathing. For the medium and large sizes nasal filters, users may feel nasal airflow restriction at the breathing flow of 15 L/min; when inhalation flow increased to 30 L/min (minute volume during moderate workload tasks, as listed in ISO/TS 16976–1:2015) [35], wearers may breathe through their mouth. These findings revealed that the tested nasal filters may not be recommended for moderate or heavier workload conditions.

Quality factor (q_f)

Based on the values of PM1 penetration and R_f reported in Tables 3 and 5, q_f was calculated to assess the overall performance of the tested nasal filters. The results are presented in Table 6. The higher q_f value corresponds to a better overall filter performance. Two-way ANOVA showed that both filter size and flow rate significantly affected the q_f (p < 0.0001). Specifically, for each pair of the tested nasal filter, the q_f value decreased rapidly with increasing the flow rates (p < 0.05), which was consistent with the finding of other scholars reported for FFRs [29–31]. The q_f values of the larger size nasal filters were significantly higher at all breathing flow rates (p < 0.05).

Table 6. Quality factor (q_f) of three sizes nasal filters and two models of N95 and P100 FFRs under different constant flow rates

Flow Rate (L/min)	Quality Factor (qf) × 10^−4 Pa^−1(Mean ± SD)
	Small	Medium	Large	N95-A	N95-B	P100-A	P100-B
15	7.0 ± 0.1	17.0 ± 0.2	22.0 ± 0.2	6270 ± 85	3210 ± 38	5920 ± 69	2220 ± 21
30	1.2 ± 0.0	5.5 ± 0.0	6.9 ± 0.1	850 ± 8	580 ± 5	1500 ± 13	520 ± 4
50	0.3 ± 0.0	1.4 ± 0.0	2.6 ± 0.0	N/A	N/A	N/A	N/A

Note: A and B represent FFRs from manufactures ‘A’ and ‘B’, respectively.

To better assess the performance of the tested nasal filters, the q_f values obtained in this study were compared to that of two models of N95 and P100 FFRs under the same testing conditions [data acquired from our previous study, please see detail testing information in reference 23]. As presented in Table 6, the q_f values for those four FFRs were hundredfold higher than those of the nasal filters, which means the overall performance of the tested nasal filters was far worse than N95 and P100 FFRs.

Discussion

China has been experiencing frequently severe haze pollution, which threatens the health of an 800 million population. PM1 aerosols constitute most of haze particles [10], which may be the ‘real killer’ during haze days. Assuming that in China, 800 million people wear single-use FFRs 100 days per year, one N95 respirator was one dollar, then the annual expense would be 0.8 billion dollars. The surface areas of FFRs were generally 100–300 cm² [36], while the sizes of nasal filters were 1–3 cm², which was only hundredths of the former. Therefore, the expense on filter material would be saved hundreds of times. Since human breathe 90% of the air through the nose [15], innovative nasal filters (worn inside the nose and worked as a miniature air filter) were designed to reduce the haze exposure. Currently, in several major cities of China (such as Beijing, Shanghai, Hangzhou, etc.), the tested nasal filters are increasingly used as a respiratory protection device by traffic policemen and street cleaners, though it is unknow how effective those filters are. The PM1 penetration results obtained in this study revealed that this new model of nasal filter, which was claimed to filter > 90% of PM2.5, failed to provide an adequate prevention against PM1 (penetration > 80%).

Whether nasal filters or traditional FFRs are worn, human body requires the same volume of air during certain work activities, and consequently the face velocity of airflow passing through the nasal filters should be hundredfold higher compared with the FFRs, as presented in Table 7. It has been well documented that with the increasing face velocity, the filter efficiency significantly decreases [22,23,27–32], because high face velocity would greatly reduce the residence time for particles passing through the respirator filter, resulting in a lower probability of particles being captured by filtering mechanisms of diffusion and electrostatic attraction [23,27–29,31]. Therefore, to reduce the PM1 penetration, there may be two ways: one is to increase the surface area of the nasal filter through various techniques such as fold methods; another solution may be reengineering the filter membrane with higher filtration efficiency and lower flow resistance.

Table 7. Face velocity of three sizes nasal filters and two models of N95 and P100 FFRs under different constant flow rates

Flow Rate (L/min)	Face Velocity (m/s)
	Small	Medium	Large	N95-A	N95-B	P100-A	P100-B
15	1.420	1.136	0.933	0.013	0.017	0.008	0.010
30	2.841	2.273	1.866	0.025	0.033	0.017	0.020
50	4.735	3.788	3.109	N/A	N/A	N/A	N/A

Note: A and B represent FFRs from manufactures ‘A’ and ‘B’, respectively.

Since the development and usage of nasal filters have just begun in recent years, there is still a lack of filtration testing standards. Current NIOSH certification test flow for FFRs is 85 L/min constant flow [19], which was closely related to the minute volume of breathing under high workloads defined in ISO/TS 16976–1:2015 [35]. The prerequisite of using nasal filters is that the wearer must breathe only through the nose, while under moderate or heavier physical exertions, people may change to mouth breathing (the switching point from nasal to oronasal breathing was 105 W/m², corresponding to light-moderate work rates) [35,37,38]. In other words, the nasal filters may only be worn under light workloads or rest conditions. Obviously, the 85 L/min test flow exceeds the light work condition, and nasal filters should be tested under a lower airflow standard. As for the test flow type, this and many other studies have demonstrated that at the same flow rate, particle penetration under sinusoidal cyclic flows (close to real human breathing pattern) was higher when compared with the constant flow [21–23]. Therefore, a cyclic flow type may provide a more stringent testing condition for nasal filters.

A limitation of this study is that only one model of nasal filter was tested. In future, more models should be investigated to study if the results will differ from one model to another. In addition, the nasal filters in this study were fully sealed, while there may exist nasal seal leakage under the actual usage condition.

Conclusion

Although nasal filters may have good application prospects in reducing haze exposure. The PM1 penetration results in this study were all > 80% (significantly higher than the penetration <10% claimed by the manufacturer), suggesting that the tested nasal filters failed to offer a sufficient protection against haze particles. In addition, to avoid mouth breathing, the nasal filters may only be used under light work rates or rest conditions. Overall, the findings suggest that more efforts should be done by the manufactures to improve the filtration efficiency and guarantee the protection and comfortability of the wearer.

Disclosure statement

The authors declare that they have no competing interests.

Additional information

Funding

This work was supported by the [National Natural Science Foundation of China] under [Grant No. 51904291), the [Basic Research Program of Jiangsu Province] under [Grant No. BK20190638], the [Fundamental Research Funds for the Central Universities] under [Grant No. 2020XGYJ08], the [Project funded by China Postdoctoral Science Foundation] under [Grant No. 2020M681781], the [Jiangsu Planned Projects for Postdoctoral Research Funds] under [Grant No. 2020Z076], and the [State Key Laboratory of Strata Intelligent Control and Green Mining Co-founded by Shandong Province and the Ministry of Science and Technology] under [Grant No. SICGM202105];Basic Research Program of Jiangsu Province [BK20190638];National Natural Science Foundation of China [51904291];Project funded by China Postdoctoral Science Foundation [2020M681781];State Key Laboratory of Strata Intelligent Control and Green Mining Co-founded by Shandong Province and the Ministry of Science and Technology [SICGM202105];Jiangsu Planned Projects for Postdoctoral Research Funds [2020Z076];Fundamental Research Funds for the Central Universities [2020XGYJ08].