Effects of Droplet Generation by Respiration and Vocalization on Infection Risk

. Due to the COVID-19 pandemic, research on the quantitative evaluation of the risk of infection has become necessary. Because airborne transmission occurs by the inhalation of droplets from infected people, understanding the mechanism of droplet generation is important. In this study, the size distribution of droplets produced by various expiratory activities was investigated, the results were compared with those of previous studies, and the applicability of the simple measurement method was confirmed. The experiment was performed using an optical particle counter and a device that could continuously ventilate the generated droplets in a clean room with a low background concentration. Among the variables in the equation for calculating the quanta emission rate to evaluate the risk of infection, the droplet concentration and inhalation rates that could be measured were determined, and the relative risk of infection for each of the various expiratory activities was quantitatively evaluated. In addition to the cases identical to those in previous studies, conversations and vocalizations were conducted while wearing a mask. Particles smaller than 1 μm were analyzed based on the theory that viruses have a high proliferation rate and a high risk of infection. The concentration of droplets produced by the expiratory activity was dominated by particles with a number concentration of < 1 μm ; however, the mass concentration was only observed at a low rate. The risk of infection increased in proportion to the volume of voice, and was markedly higher in the case of loud voices. In addition, the risk of infection decreased when wearing a mask, and the extent of reduction varied depending on the method of wearing the mask.


Introduction
To evaluate the risk of infection via aerosols, it is important to understand the droplet concentration and size distribution of the aerosols in infected individuals. Therefore, we first examined the particle size distribution of droplets produced by various expiratory activities. Buonanno proposed the quanta emission rate, expressed in Equation (1), as a method to evaluate the risk of contagion [1]. (1) where cv is the viral load in the sputum (RNA copies mL -1 ), ci is the amount of virus expressed in terms of viral RNA copies corresponding to one quanta (quanta RNA copies -1 ), IR is the inhalation rate (m 3 h -1 ), Vd is the droplet volume concentration (mL m -3 ). In this study, we measure IR and Vd, which are measurable, and use the product (IR•Vd) to evaluate the relative risk of infection for various vocalization patterns with and without a mask.

Droplet measurement methods
To achieve this objective, a low background environment was required; therefore, we conducted our experiments in an ISO standard class 5 clean room. The measurement of droplets due to expiratory activities was performed by Morawska [2] using a small wind tunnel in which the subject could put their head and in which a built-in HEPA filter provided a clean environment. In this study, a zinc-based duct with a ventilation fan was installed in a clean room to enable simple measurements.  https://doi.org/10.1051/e3sconf/202339601032 IAQVEC2023 after the utterance. To confirm that the air in the duct was in a steady state, we measured the CO2 concentration, which indicated that 80-90% of the air was in a steady state at the position where the particle counter was measured. An optical particle counter was installed inside the duct to measure the steady-state concentration in the duct. The inlet of the particle counter was placed 15 cm from the mouth. The air mixed with the droplets sucked in by the fan was exhausted outside the room. Tab. 1 shows a comparison of measurement methods with Morawska.

Measurement protocol
This study was conducted with five subjects: two males and three females. To match the volume of the voice of each subject, the measurement was performed while constantly checking the value of the sound level meter placed immediately next to the face. The vocalizations were performed for 2 min, and the activity intervals were defined for each expiratory activity and measured using a metronome to avoid differences in the number of vocalizations between the subjects. Three major types of expiratory activity were performed, as shown in Tab. 2. The details of each activity are described below.

Comparison with previous studies
To examine the accuracy of a simple measurement method, we performed expiratory activities similar to those performed in Morawska's study. ・Nasal breathing: inhaling and exhaling through the nose. ・Breathing: inhaling through the nose and exhaling through the mouth. ・Whispered: counting in Japanese for 10 seconds in a whisper.

Vocalization method
・ Breathe in through the nose and out through the mouth (Breathe for 3.24 s per set for 2 min) ・Counting in Japanese for 10 s (Count from 1 to 10 in 3.75 s per set for 20 sets) ・Unmodulated (Vocalize "aah" for 3 s per set for 20 sets) Since the instrument reported the sum of 6 s, the duration of the vocalizations was determined by metronome rhythm, which could be performed in 6 s or less. When the subjects completed vocalizing, they held their breath until the next 6 s measurement began, and then took 20 6 s measurements, for a total of 2 min.
・ Breathe in through the nose and out through the mouth (Breathe at a natural pace for 2 min) ・Counting in English for 10 s (Alternate 10 s of counting and 10 s of natural breathing for 2 min) ・Unmodulated (Alternate 10 s of vocalization "aah" and 10 s of natural breathing for 2 min) The subjects were given a demonstration on how to vocalize, and the timing of the activity was determined using the second hand on an analog clock.

Calculation method
To compare with previous studies, the values obtained in this experiment were converted using the following equation.
Droplet number concentration per vocalization The concentration was obtained by multiplying the dilution factor (D) calculated from the following two equations to indicate the concentration in the upper respiratory during expiration ・Voiced: counting in Japanese for 10 seconds in a normal voice. ・Unmodulated: unmodulated vocalization with "aah." 2. Vocalization for conversation. The Japanese greeting "ohayougozaimasu" was uttered at three voice volumes. 3. Vocalizing while wearing a mask.
Six cases of nasal breathing, mouth breathing, and normal voice greetings were performed, with the mask firmly in place and with only the nose out. The masks were made of non-woven fabric with a VFE test of 99%, and all subjects used the same size.

Inhalation rate measurement methods
The aerosol monitor AE-310S was used to measure the rate of inhalation. The measurement was performed by connecting a transducer to a specially attached mask. This device could acquire data every 0.1 s.

Analysis methods
In comparison with previous studies, the values obtained in this experiment were multiplied by the airflow rate of the fan ventilating the duct and divided by the inhalation rate to convert the concentration per exhalation to 1 cm 3 .
Santarpia performed an analysis on the presence of RNA and proliferative potential of the virus for patients with COVID-19. The presence of RNA was confirmed for all particle sizes, but statistical superiority of RNA replication was observed for particle sizes less than 1 µm, with a 90-95% confidence level in the 1-4 µm range, and no superior replication above 4.1 µm [3]. Morawska et al. also proposed that the size of the droplets depends on where they are produced, with 1 µm in the bronchi, 5 µm in the pharynx, and 50 µm in the oral cavity, and the risk of infection for particles less than 1 µm produced in the bronchi [4]. Based on these theories, we analyzed particles smaller than 1 µm to calculate the relative risk of infection.
To calculate the relative risk, the volume concentration of the droplets was calculated by assuming a spherical shape based on the average diameter of each particle size range, and the mass concentration was calculated by multiplying the volume concentration by the density of water, assuming that the droplets were water molecules.
Moreover, all results were calculated by subtracting the background concentration from the measured concentration, and the background concentration was measured with the face close to the duct and without exhaling, just as during vocalization. The background concentration measured with the mask was used to calculate the effect of the mask.

Comparison with previous studies
The measurement results of this study were converted to the concentration per exhalation of 1 cm 3 and compared with those of previous studies in Tab. 3. Consequently, the trend of values in this study and previous studies was similar, confirming the reliability of the data of the simple measurement method. The results for 0.8 µm and 1.8 µm in this study are larger than those in the previous study, but the error in the conversion of 1.8 µm is thought to be caused by the difference in classification between the measurement equipment used in this and the previous study. The results for each subject are shown in Fig. 2. This result shows that the droplet concentration varies between individuals and that some subjects generate more droplets than others. However, the result in Fig. 2 multiplied by IR to the number of droplets per hour (Fig.  3) shows that the result for another subject is the highest. Subject C had a high droplet concentration and a high inhalation rate, whereas subject B had a low inhalation rate in relation to the droplet concentration, which is probably the reason for the extremely high value. The inhalation rate would not be negligible because the number concentration per hour is related to the risk of infection.

Vocalization for conversation
The sum of the mass concentrations per hour of particles smaller than 1 µm is shown in Fig. 4. From this result, it can be concluded that the droplet concentration is roughly proportional to the loudness of the voice. Fig.  5 shows the sum of the mass concentrations of particles of all sizes. In this graph, the droplet concentration at 90 dB is prominent. This result suggests that, although there are individual differences, large-diameter droplets are generated when the voice is loud, and the generation of large-diameter particles has a significant effect on the results when converted in terms of mass concentration.  A comparison with the previous study in the previous section shows that particles smaller than 1 µm account for most of the total generation in terms of number concentration, whereas Fig. 6 shows that their ratio is negligible in terms of mass concentration.  Fig. 7 shows the droplet generation volume with the mask divided by the droplet generation volume without the mask. This indicates the extent to which the generated droplets penetrated the mask. Particles with a diameter of 3 µm or more were completely collected by the mask, indicating that the smaller the particle size, the more the virus leaked through the mask. In addition, when the mask was worn and the nose exposed, the transmission rate increased slightly in relation to that when it was completely covered.

Relative risk from expiratory activities
The average IR•Vd for nasal breathing without a mask was set to 1. The results calculated for the total particle size are shown in Fig. 8, and those calculated for particles smaller than 1 µm are shown in Fig. 9. For particles smaller than 1 µm, there was no difference in the risk of infection from activity in relation to those from particles of all sizes, but both results showed that vocalization without a mask was several times riskier than breathing, with the risk increasing with voice volume. Furthermore, wearing a mask reduces the risk of infection in all expiratory activities. For particles smaller than 1 µm, for which infection risk has been proposed, the reduction in infection risk by wearing masks is clear.

Conclusions
The results of this study show that even a simple method can generally identify trends in the droplet concentration and size distribution of aerosols. The droplet concentration during expiratory activities varied considerably among the subjects, but in all subjects, particles with a diameter of 1 µm accounted for a large percentage when converted to number concentration, and the percentage was very small when converted to mass concentration. The droplet concentration was proportional to the loudness of the voice for all particle sizes; however, the droplet concentration for large particle sizes was much higher when the voice was loud. The smaller the particle size, the more difficult it is for the mask to collect infectious particles; however, it was confirmed that the mask reduced the risk of infection for particles less than 1 µm in diameter. Moreover, to further reduce the risk of infection, it is important to correctly wear a mask.  Fig. 9. Relative risk of infection for each expiratory activities at particle sizes less than 1 µm with the mean value of all subjects in IR•Vd for nasal breathing set to 1.