Comparison of particles leakage with different types of masks using a coughing simulator in a negative pressure isolation room.

. Although the rate of wearing masks has increased to 97.9%, there is a possibility that a gap between mask and face may occur because it does not completely fit to face, or that the virus may spread due to the poor performance of the mask. Accordingly, many studies are being conducted to evaluate performance of masks, or to measure the flow rate of coughing with PIV (Particle Image Velocimetry). However, most of these studied the performance inside or on the surface of a mask, and research on particle emitting after wearing masks is insufficient. Therefore, in this study, the particle concentration leaked after wearing masks in the negative pressure isolation room experimental chamber was measured, using particle counters and low-cost sensors. A manikin was installed on a bed to release aerosols, and the tendency of particles to decrease after one cough was analyzed. As a result, most of the particles were removed within 10 minutes. The low-cost sensors and particle counters installed in the centre and right wall exhaust diffuser show that with N95 mask on, particles were removed the fastest, and with surgical mask on, particles leaked the most immediately after coughing and were removed the slowest.


Introduction
COVID-19, one of the SARS-CoV-2 viruses, has caused numerous casualties since the initial outbreak. It is mainly due to direct routes such as droplets released by sneezing, coughing, and conversation of infected people, and contact between people [1,2]. The droplets typically diffuse up to 2 m and float in the air for a limited time. However, in the case of SARS-CoV-2, it can float for up to 3 hours without damage. In addition, research has recently been conducted to prove the possibility of airborne transmission in the case of inhaling an aerosoltype virus rather than droplets. Liu et al. [3] revealed that transmission that occurs at a close distance from infected patients is caused by short-range air infection as well as droplets. Nissen et al. [4] sampled SARS-CoV-2 virus for HVAC systems in hospitals using a central air conditioning method. As a result, the positive reaction for viruses was confirmed in ducts and diffusers, and SARS-CoV-2 claimed that there is a possibility of long-distance transmission through air. According to the results of Korea Disease Control and Prevention Agency's survey, the rate of wearing masks to prevent air-borne infectious diseases has increased to 97.9%. Among them, the ratio of wearing disposable dental masks was the highest at 46.3%, and wearing KF94, KF80, and droplet blocking masks accounted for 44.7%. However, even if you wear a mask, a gap between face and mask may occur because it does not completely fit to your face, or the virus may spread through the airflow due to the poor performance of the mask filter. * Corresponding author: mksung@sejong.ac.kr Accordingly, many studies are being conducted to evaluate fitting rate, intake resistance, and leakage rate depending on the type of mask, or to measure the flow rate of coughing with Particle Image Velocimetry (PIV). However, most of these are studies that test the performance inside or on the surface of the mask through fit test equipment, and studies that visualize and analyze the airflow around the human body after wearing the mask are insufficient. It is important to check the tendency of the virus to enter the mask. However, research should be conducted to prevent airborne transmission by analyzing the tendency of the virus to leak to the outside. Therefore, in this study, the airflow around the manikin was analyzed by visualizing the airflow according to the types of seven masks and whether they were worn with a laser device.

Negative pressure isolation ward (NPIW) experimental chamber
The experimental chamber of the negative pressure isolation ward consists of an ante room and a ward ( Figure 1). A supply diffuser, an exhaust diffuser, and a make-up grill were installed in the ante room, and all diffusers were opened at the time of the experiment. Since there is no corridor, the ante room was set to positive pressure rather than the ward to prevent the leakage of pollutants in the ward. The ward has one ceiling-type supply diffuser, one ceiling-type exhaust diffuser, and two wall-type exhaust diffusers, and all diffusers were opened. Air change rate of the ante room was set to 6 ach. The temperature of the ante room and the ward was set to 21~22 °C, and the patient's bed was installed between the wall exhaust diffusers. The ward was required to maintain a differential pressure of -3 Pa or more compared to the ante room, and air change rate of the ward was set to 12 ach.  In this study, a smoke generator (Air Trace S, Concept), and a DPSS laser (PIV-RN-G5W, Intech) were used to visualize airflow, and digital temperature, humidity, and air flow meter (480, Testo) was used to measure the foot pump's flow rate. The experiment was conducted in the ward of the negative pressure isolation ward experimental chamber, and the setup diagram of the experiment is shown in Figure 2. Two manikins were placed to face each other 2 m apart, and the smoke generator and foot pump were connected to model 1 using a tube and valve. After putting a mask on Model 1, a smoke generator was operated to fill the acrylic box, and then a person coughed with a pump was implemented. Smoke emitted from the manikin by coughing was visualized with a laser sheet to confirm the airflow around the manikin. There are a total of 8 experimental cases, and they are classified according to whether or not to wear a mask and types of masks ( Table  1). The maximum wind velocity of the cough measured in the front of the respiratory tract of Model 1 is about 17 m/s, indicating a value like the cough flow velocity measured in the study of Wei&Li [5]. Using low-cost sensors and particle counter (11-D, Grimm), the number concentration of particles leaked out of the mask during coughing was measured depending on the types of masks and whether it was worn, and the experimental cases are shown in Table 1.

Particle concentration measurement
Each low-cost sensor has an embedded PM2008 (CUBIC) sensor. Specifications of the sensors are shown in Table 2 and 3.

Response time 1sec
Time to first reading ≤ 8 seconds The manikin was installed on a bed in the chamber at the same height as the actual person's sitting height (about 0.8 m), and the oil syringe connected to the manikin was filled with an Atomizer (TSI 3079A) aerosol and then pushed the piston at a rate like the cough flow rate to release the particles (about 17.3 m/s). The measurement was conducted at positions shown in Figure 3. Both the low-cost sensors and particle counters were installed at the center of the ward and in front of right wall diffuser, so the values of the two devices were compared. After one cough, tendency of particles with a diameter of 0.3, 0.5, 1.0 μm (low-cost sensor), 0.298, 0.488, 0.943, and 2.982 μm (particle counter) to decrease was analyzed.  In the case of Case 1, the airflow was initially directed downward by the supply air, but then rose to reach model 2 along the airflow (Figure 4(a)). In Case 2, particles leaked from the front of the mask, nose, and chin, and the particles spread along the airflow to model 2 ( Figure 4(b)). In the case of Case 3, particles leaked from the front, side, and nose of the mask (Figure 4(c)). The particle amount of outflow was similar to that of Case 2, and the particles reached model 2 along the airflow. In Case 4, particles leaked from the front and nose of the mask (Figure 4(d)). The particle outflow decreased compared to the previous case but spread to model 2. In Case 5, particles flowed out from the front and nose of the mask, and it was confirmed that the amount of particle outflow decreased compared to the fish shaped mask (Figure 4(e)). The particles moved to 1/3 of the total distance and did not spread to model 2.

Visualization experiment
In the case of Case 6, particles leaked from the part of nose and jaw, and the outflow decreased compared to Case 5 (Figure 4(f)). Most of the particles fell at the same time as they were discharged and did not spread to model 2. In Case 7, particles leaked from the front, nose, and chin part (Figure 4(g)). The amount of particle outflow is like the fish shaped mask, and a small number of particles spread along the airflow to model 2. In Case 8, particles leaked from the nose of the mask, and the outflow was the smallest among the types of masks used in the experiment (Figure 4(h)). Most of the particles fell at the same time as they were discharged and did not reach model 2.

Particle concentration at the centre of the ward
As a result of measurement with the low-cost sensor, the concentration of particles released in the order of surgical, KF80 bird beak type, KF80 fish shaped, KF94 fish shaped, N95, KF-AD, and KF94 bird beak type was high ( Table 4). As a result of the particle counter ( Figure  5), high concentrations of particles were generated in the order of surgical, KF80 fish shaped, KF80 bird beak type, N95, KF94 fish shaped, KF-AD, and KF94 bird beak type. In the case of KF94 fish shaped, KF94 bird beak type, KF-AD, and N95, the particle concentration decreased at a high rate ( Figure 6). Also, the results of the two measurement devices were confirmed to be similar, proving the reliability of the low-cost sensor measurement results.      As a result of measurement with a low-cost sensor, the concentration of particles released in the order of surgical, KF80 fish shaped, KF80 bird beak type, N95, KF94 bird beak type, KF94 fish shaped, and KF-AD was high (Figure 7). As a result of the particle counter measurement, high concentrations of particles were generated in the order of surgical, KF80 fish shaped, N95, KF94 bird beak type, KF94 fish shaped, KF-AD, and KF80 bird beak type (Figure 8). The particle removal rate was fast in the order of N95, KF94 bird beak type, KF94 fish shaped, KF80 fish shaped, KF80 bird beak type, KF-AD, and surgical masks. When wearing a surgical mask, the largest number of particles were released, and the time taken to remove the particles that leaked out of the mask was the longest (Table 5). Likewise, the results of the two measurement devices were confirmed to be similar, proving the reliability of the low-cost sensor measurement results.

Conclusions
In this study, airflow visualization and concentration measurement around the manikin were performed according to the types of masks and whether they were worn in the experimental chamber of the negative pressure isolation ward (NPIW). The research results are as follows, and it is significant in that the movement of particles flowing out of the mask was analyzed. Most of the particles generated during coughing in the NPIW are removed within 10 minutes. Since aerosoltype particles are not precipitated and are likely to spread over a long distance depending on the airflow, the importance of wearing a mask can be seen through the case without wearing a mask. In the case of the surgical mask, particles passed through the mask and diffused more than 1~2 m. In addition, the concentration of leaked particles was the highest and the particle removal time was the longest. It was like the results of the study of Kim et al [6]. This is believed to be due to the absence of a filter in the mask. KF80 bird beak-type mask leaked a small concentration of particles compared to the fish shaped mask, and a small airflow spread. KF94 fish shaped mask had fewer diffused airflow than the bird beak type, but the particle outflow was similar.
Bird beak-type mask has a folding line in the center of the mask, so there is a possibility of leakage in the area. N95 mask had the lowest particle outflow and the shortest particle removal time. Therefore, it can be seen that the N95 mask has high fitting rate and is effective in preventing air-borne infectious diseases. When a high-grade mask was worn, the amount of particle outflow was reduced, and the particle removal time was shortened. In addition, when wearing a mask of the same grade, face fitting rate and wearing method are more important than the shape of the mask.