key: cord-0886937-k4706do7
authors: Xu, Chunwen; Wei, Xiongxiong; Liu, Li; Su, Li; Liu, Wenbing; Wang, Yi; Nielsen, Peter V.
title: Effects of personalized ventilation interventions on airborne infection risk and transmission between occupants
date: 2020-05-30
journal: Build Environ
DOI: 10.1016/j.buildenv.2020.107008
sha: c6dce98d85980d22f25de3bf7358f5e15899e937
doc_id: 886937
cord_uid: k4706do7

The role of personalized ventilation (PV) in protecting against airborne disease transmission between occupants was evaluated by considering two scenarios with different PV alignments. The possibility that PV may facilitate the transport of exhaled pathogens was explored by performing experiments with droplets and applying PV to a source or/and a target manikin. The risk of direct and indirect exposure to droplets in the inhalation zone of the target was estimated, with these exposure types defined according to their different origins. The infection risk of influenza A, a typical disease transmitted via air, was predicted based on a dose-response model. Results showed that the flow interactions between PV and the infectious exhaled flow would facilitate airborne transmission between occupants in two ways. First, application of PV to the source caused more than 90% of indirect exposure of the target. Second, entrainment of the PV jet directly from the infectious exhalation increased direct exposure of the target by more than 50%. Thus, these scenarios for different PV application modes indicated that continuous exposure to exhaled influenza A virus particles for 2 h would correspond with an infection probability ranging from 0.28 to 0.85. These results imply that PV may protect against infection only when it is maintained with a high ventilation efficiency at the inhalation zone, which can be realized by reduced entrainment of infectious flow and higher clean air volume. Improved PV design methods that could maximize the positive effects of PV on disease control in the human microenvironment are discussed.

dose-response model. Results showed that the flow interactions between PV and the infectious 23 exhaled flow would facilitate airborne transmission between occupants in two ways. First, 24 application of PV to the source caused more than 90% of indirect exposure of the target. Second, 25

Exposure to pathogens from air has always been a major source of human morbidity and 2 mortality worldwide. Influenza, for example, is one of the most common diseases that 3 associate with human expiratory droplets and droplet nuclei, and the Centers for Disease 4 Control and Prevention (CDC) reports that millions of illnesses and thousands of deaths are manikins (d is the inner diameter of the duct, which is 75 mm). 8

Two PV nozzles were placed in front of the source (PVa) and the target manikin (PVb), 9

respectively, pointing to their nose tips at a relative distance of 0.36 m and at an angle of 23.6° 10 to the vertical direction. These dimensions were selected to ensure a relatively short distance 11 between the two manikins. The relative distance of 0.36 m was selected to satisfy both the 12 ergonomics considerations and a relatively high clean air volume. The PV jet was used to 13 provide clean air directly to the manikins' inhalation zone and was thus the complement to 14 MV. centerline were tested in [23] . The centerline velocity decay of the free nozzle jet with the 7 tested flow rates was measured with a hotwire anemometer (Swema 03+) and the velocity 8 profiles can be found in the Appendix (Fig. S1 ). 9

The PV system was mounted inside the test chamber, and supplied with recirculated room air. 10 Specifically, recirculated isothermal room air was re-filtered by being passed through a 11 custom-made HEPA filter mounted in the air duct of the PV and then supplied to the nozzle as 12 clean particle-free air. The particle-free air chamber was well-sealed to avoid any inward 13 leakage of particles from the outside, thus recirculated and finely filtered room air was used to 14 supply the PV instead of external fresh air, to maintain the integrity of the test chamber and 15 also to simplify the experimental system. 16 The initial turbulence intensity of the nozzle jet (at flow rates of 3 L/s, 6 L/s or 9 L/s) was 17 measured with a hotwire anemometer (Swema 03+) and found to be less than 5%, and the 18 length of the clean air core was 3-4 times the diameter of the nozzle (50.8 mm), both of which 19 corresponded to our previous measurements [23] . It is preferable that a wider PV air terminal 20 device (ATD) with a longer clean air core is used, as this can reach an occupant's breathing 21 zone and therefore protect him/her from polluted room air. However, given that people vary 22 widely in sizes, shapes and seating distances, and that they may adjust the position of the PV 23 ATDs, it is unlikely that the use of PV can reliably offer optimal protection. Therefore, in this 24 study a small nozzle was used to represent a situation wherein PV ATD supplies 25 approximately 40% of clean air to the inhalation zone. 1 Fig. 3 presents the turbulent development of the PV flow with the same small nozzle aimed 2 with a horizontal placement at a real person, and aimed with a downward or upward incline 3 placement at the same manikin. It was found that with a nozzle-exit-to-target distance of 4 approximately 0.4 m, this PV setup (Fig. 3 ) achieved a clean air delivery efficiency of 5 approximately 40%, varying slightly with different placements of PV (horizontal, inclined 6 upward or downward). Increasing the flow rate of PV from 3 L/s to 9 L/s increased the PV 7 clean air-delivery efficiency by no more than 5% [23] . The horizontal placement of PV was 8 not used in this study as it occupied too much shared space between the two face-to-face 9 occupants. Instead, PV inclined upward or downward was used, as these placements achieved 10 similar clean-air delivery efficiencies as the horizontal case. 11 thermal plume [23] : (a) PV with a horizontal placement applied to a real person; (b) PV 1 applied to a thermal manikin inclined downward and (c) inclined upward. 2 To compare the effect of the positioning of PV, two scenarios were considered: nozzles placed 3 above (scenario 1, Fig. 1 (b) ) or below (scenario 2, Fig. 1 (c) ) nose height. In both scenarios, 4 PV may be performed with the PV ATD mounted on a movable arm-duct [12] attached to the 5 top or the bottom of the room. The PV ATD can also be integrated with the ceiling or the desk 6 to locate it close to the occupant. The concept of PV is to supply clean air directly to the 7 breathing zone of a sedentary person, and to allow the exposed occupant and the infector 8 some control of the flow rate and the flow direction. The infector may not be aware of his/her 9 infection, and/or may only use PV when his/her health condition permits. Ten cases under two 10 scenarios were tested (as shown in Table 2 ), which included the possibility of individual 11 control of the PV flow rate. 12 

The sensors and test equipment that were used and their specifications are listed in Table 3 . To 11 simulate multiphase flow consisting of expiratory droplets suspended in expelled air from the 1 infected person, a Collison nebulizer (3-jet, BGI, Inc., Waltham, MA) was used to generate 2 polydisperse droplets. With appropriate settings of the flow rate of supplied air and solution of 3 the liquid, this nebulizer can produce similar droplet size distribution profiles to human 4 breathing. The pressure of clean and compressed air supplied to the nebulizer was set as 10 5 psig (0.069 MPa), corresponding to an aerosol generation flow rate of 3.9 L/min. Sodium 6 chloride (NaCl) served as the core of droplet nuclei in human breathing [36] and 50% distilled 7 water mixed with 50% isopropyl alcohol by volume was used as the solvent to simulate the 8 evaporable components in human saliva. The density of the solution was 33.8 µg/mL. The 9

droplet size-distribution profiles between human breathing and the nebulized aerosols in this 10 study are compared in Table 4 . 11 The droplets released from the nebulizer were injected into the exhaled air of the source 7 manikin (manikin A) and were then expelled at a designed frequency and temperature (Table  8 1) from the manikin's mouth. It was expected that droplet deposition would occur in the 9 manikin's respiratory airway, which would result in particle loss before the exhalation flow 10 arrived at the manikin's mouth. However, as only droplets released from an occupant's mouth 11 would contribute to infectious particle transmission indoors, the deposition or the subsequent 1 re-suspension process occurring in the manikin's respiratory airway was ignored. Accordingly, 2 the particle concentration measured at the mouth opening was used as the initial concentration 3 for further analysis. However, the nebulizer-generated initial droplet concentration was found 4 to be higher than that from humans, and thus should be normalized by the average initial 5 concentration from human subjects' mouth breathing for further infection-risk assessment. 6

The time required for droplets to evaporate from their initial condition to nuclei was expected 7 to be very short, but depends on many factors [39] [40] [41] [42] . Based on the calculation suggested by 8

Nicas et al. [42] , it was found that 0.1 s would be required for droplets with an initial size of 9 12 μm to evaporate to nuclei. Therefore, we expected that instantaneous evaporation occurred 10 before the aerosolized droplets reached the manikin's mouth via the tube inside the manikin's 11 body. Fig. 4 shows one of the examples of the sampled particle size distributions at the source 12 manikin's mouth. It can be seen that the geometric mean aerodynamic particle size is 13 approximately 0.7 μm, with more than 80% of droplets smaller than 1 μm, which means that 14 for short-range transmission, the deposition rate of aerosols from mouth breathing is conditions; when ε bz > 1, the proportion of exhaled droplets in the target's breathing zone is 7 greater than in the ambient environment; and when ε bz < 1, PV can protect the exposed 8 occupant from the exhaled droplets to some extent, which is the preferable condition. It may 9 be possible to reduce the inhaled concentration of droplet nuclei by using a high flow rate of 10 ventilation q 0 to dilute the overall infectious particle concentration in the entire system. 11

However, the localized concentration in the microenvironment between the occupants may 12 not be efficiently and effectively diluted by the general ventilation [10, 11] . 13

As shown in Fig. 5 , the localized concentration C bz in the inhalation zone of the target is not 14 only derived from indirect exposure C indirect but also from direct exposure C direct [46] . Indirect All of the measurements were conducted after a few hours of MV, when the particle 5 concentration in the exhaust had reached a relatively steady state. APS-based measurement of 6 C bz lasting for 10 minutes generated 600 samples of instantaneous particle concentration. 7 C indrect was obtained by measurement at the same sampling point as C bz , immediately after 8 shutting down the exhalation of the source. The sampled particle concentration in the 9 breathing zone was thus only from the air distribution around the target. 10 A 10-minute sampling period was also used for the measurement of C indrect . There might be a 11 decrease of C indrect over time without the release of particles from the source, but it can be 12 inferred that the overall concentration of tracer particles in the room would decay by less than 13 8% in an exponential function over time under an ACR of 2 h -1 . Given that the 14 microenvironment between the occupants might not be diluted efficiently by the general 1 ventilation, the decay of C indrect during the short sampling period was not considered. 2

Another index to evaluate infection risk under PV was used. The advantage of using tracer 4 particles as media to simulate pathogen transport is that the infection risk of certain diseases 5 such as influenza A can be evaluated by using the exposed droplet concentration in the 6 breathing zone of the susceptible person. Based on the dose-response model for an unsteady 7 imperfectly mixed environment [48], the probability of infection can be estimated using Eq. (7) 8 [49]: 9

where P 1 (t 0 ) is the probability of infection of the susceptible person, m is the total number of 10 size bins, v(t) j is the volume density of droplets of the jth size bin, and β j is the deposition 11 fraction of the infectious particles of the jth size bin. The exposure level is divided into 12 different size bins based on the consideration that the infectivity varies with particle size and 13 deposition site in the respiratory tract [50] . Furthermore, r is a fitting parameter to make P 1 (t 0 ) 14 equal to 0.5 when the infectious dose is ID 50 , c is the pathogen concentration in the respiratory The expiratory droplets evaporate and shrink to nuclei within a short time [39] [40] [41] [42] . The 20 concentration of droplet nuclei can be measured at the inhalation zone of the target manikin 21 and the droplet nuclei sizes are adopted for β j and r j . The nucleus sizes of aerosols should be 22 converted to their initial droplet sizes to calculate v(t) j , as c refers to the pathogen 1 concentration in the respiratory fluid before evaporation. 2

The volume density of expiratory droplets in the exhalation of the infector, v(t), can be derived 3 from the measured concentration in the breathing zone of the susceptible person, C bz [49] . As 4 the initial concentration of droplets released from the source manikin was higher than those 5 from real persons, the initial aerosol concentration, C 0 , measured at the mouth opening of the 6 source manikin (P1 Fig. 1 ) was normalized to an average initial concentration of 0.092/cm 3 , 7 which was in line with the data obtained by APS analysis of human subjects breathing [38] . 8

The C bz was then normalized with respect to the ratio between C 0 and 0.092/cm 3 , and was 9 used for further infection-risk assessment. Table 5 . 16 

The results were obtained from the measurement point P3 (indicated in Fig. 1 transport of infectious particles. 14 3.1 Exposure risk assessment for scenario 1 15 Figs. 6 and 7 show the exposure risk ε bz and indirect exposure risk ε indirect for the scenario 1 16 test series in Table 1 , respectively. Fig. 7 depicts the indirect exposure without measuring 1 cases of PVab. It can be seen that C bz of the baseline case of MV was slightly higher than C e 2 with ε bz slightly higher than 1, indicating slightly higher droplet concentrations were present 3 in the microenvironment between the occupants than the ambient air. This where the indirect exposure ε indirect was comparable to ε bz with MV, accounting for an average 13 of 94.2% total exposure. 14 When the source manikin was subject to PV, indirect exposure was still dominant, with an 15 average proportion of 96.2-98.1% of the total exposure (Figs.8 and 9(b)). Fig. 6 shows that 16 the ε bz with PVa was slightly higher than that with MV and no significant correlation with the 17 flow rates of PVa was found. Due to the downward supply of PVa to the breathing zone of the 18 source manikin, the exhaled flow interacted with the PV jet immediately after it left the mouth 19 opening ( Fig. 9(b) ), and thus the mixing effect of the exhaled droplets with the forced 20 convection of PV was enhanced. The development of the exhaled flow was inhibited by PV, 21 especially at higher PV flow rates of 6 L/s and 9 L/s in which the centerline disappeared and 22 the centerline velocity was unmeasurable. However, the exhaled flow was able to penetrate 23 the 3 L/s PV flow, whose centerline velocity was measured to be 0.22 m/s and was 3.5 cm 24 higher than the mouth height at the P2 plane. This was because the penetration velocity was 25 significantly higher than the received PV velocity at the mouth. The maximum exhaled 1 velocity from the mouth was approximately 2.44 m/s and the PV centerline velocity decayed 2 to merely 0.56 m/s at a distance of 0.36 m from the nozzle [23] . As the PV was oriented 3 toward the nose, the velocity of the PV at the mouth away from the centerline would be even 4 less than 0.56 m/s. 5 Fig. 6 also shows that unexpected transmission of droplets could occur when PVb was used. 6

The ε bz and its fluctuations with PVb were significantly higher than those with MV or PVa. 7

The higher fluctuations of ε bz may be caused by the dynamic interactions of PV with the 8 periodic exhaled airflow, especially for PVb of 3 L/s and 6 L/s. The ε bz was elevated as the 9 droplet-laden exhaled flow grew with propagation distance, reached the accessible area of PV 10 flow and was then entrained by PV. The exhaled droplets were taken directly to the inhalation 11 of the exposed manikin by the PV air. 12

However, when the source manikin was inhaling, the droplets would not be directly loaded by 13 the PV air and the ε bz would be less. The ε bz reached a maximum value of 5.4 (Fig. 6 , PVb 7 14 L/s), but the averaged ε indirect for all PVb cases were less than 1 and even lower than those in 15 the MV case (Fig. 7) . This meant that the direct exposure largely contributed to the elevated 16 exposure risk via PVb. Fig. 3 and [56] . 4

The measurement results for scenario 2 are presented in Figs. 10-13. When the source 6 manikin used PVa, it was observed that the exhalation flow was able to penetrate the supplied 7 PV flow of 3 L/s. The centerline velocity of exhaled flow at P2 plane was found to be 0.24 8 m/s (9.8% of the initial velocity) and 7 cm above the mouth height, which also exceeded the 9 nose height of the target. The centerline of the exhaled flow was somewhat elevated by the 10 upward air stream of PVa. 11

When PVa was operated with 6 L/s or 9 L/s, the centerline of exhaled flow disappeared at the 12 P2 plane, and the mixing of exhaled droplets with PV air occurred in the breathing zone of the 13 source manikin. Therefore, the contamination of the inhaled airflow of the target manikin 14 largely depended on indirect exposure. The average ε bz of PVa was slightly greater than 1, 15 approximately the same as that under MV, indicating that complete mixing occurred in the 16 inhalation zone of the target. The proportion of indirect exposure to total exposed droplets was greater than 94%, slightly greater than that for 3 L/s. 1

As both the exhaled flow and the PVb air developed an upward direction as shown in Fig.  2 13(b), the collision between the PV jet and the end part of the exhaled flow or the remaining 3 suspended cloud of particles after exhalation was not as marked as that which occurred in 4 scenario 1. A smaller proportion of exposed droplet nuclei was derived from direct exposure 5 (56.7-72.7%) than that in scenario 1 (61.3-85%), but still more than 50% of the inhaled 6 droplet nuclei were directly delivered from the exhalation of the source. Thus, the use of PVb Therefore, the placement of PV for the susceptible person within the infectious 1 particle-exposure zone between the occupants should be avoided, as PV directed to this area 2 may actually lead to a heightened infection risk. Apparently, the separation distance of 0.86 m 3 was not sufficient for the use of PV for the two face-to-face occupants, and thus the minimum 4 distance required for the placement of PV to control infectious particle transmission should be 5 further investigated. 6

The application of PV to the infector would accelerate the mixing of infectious particles with 7 room air and indirect transmission would be the dominant mode of particle transport. In Fig.  8 10 a comparison of PVab with PVb indicates that PVb effectively protected the susceptible 9 person only when it was placed away from the high-concentration droplet flow and was 10 realized with a relatively large amount of clean air. 

Infection-risk assessment is a useful tool for studying disease transmission and evaluating the 6 effectiveness of infection-control measures [33] . Fig. 14 shows the predicted infection risk of 7 influenza A over an exposed time interval ranging from 10 min to 2 h. The results were 8 obtained based on the epidemiological data in Table 5 L/s had the highest influenza A infection risk in both scenarios. This was due to that fact that 1 the modeling of infection risk not only considers the absolute aerosol concentration C bz but 2 also takes into account the size distributions of both the inhaled droplet nuclei and the initial 3 expiratory droplets. 4

From the results in Fig. 14, it can be seen that PV has the potential to either increase or 5 decrease the infection risk of influenza A in both scenarios. The results also show that 6 increasing the PV flow rate does not always effectively control the infection risk. That is, the 7 increase of infection risk with PV compared with that with MV was caused by PV causing 8 increased indirect and/or direct exposure level to the susceptible person. The application of 9

PV to the infectious source may therefore in some circumstances accelerate the mixing of 10 infectious particles with room air and may increase the indirect exposure to the person who 11 shares the same room space, but in this situation the increase in infection probability was 12 estimated to be no more than 0.15 within the calculated exposure interval of 2 h. 13

When the PV flow developed in the infection-prone zone between the occupants, rapid 14 deterioration of the PV air would probably introduce even more infectious particles directly 15 into the inhalation, which obviously should be avoided in practical design. For a short 16 exposure time interval, e.g., 30 minutes, the infection probability was less than 0.1 and did not 17 vary much among different alignments of PV. With continuous contact with the infectious 18 particles for over 2 h, the infection probability could exceed 0.85, which equates to a high 19 possibility of infection for the susceptible person. If the intake dose was well controlled and 20 maintained at a low level, the infection risk over 2 h would be minimized to approximately 21 0.28 by use of PV, as shown in Fig. 14 (b) . This implies that PV may offer protection against Conversely, PV may aid in the transmission of infectious particles when it closely interacts 3 with the exhaled flow from the infectious source. For example, when PVa was 3 L/s, the 4 exhaled flow could penetrate the invading PV flow, and thus these PV conditions did not 5 protect the exposed occupant. 6

To summarize, the performance of PV used for infection-risk control depends largely on the 7 extent to which PV entrains infectious droplets from its working environment. That is, PV 8 used for the infector facilitated the transport of airborne particles but the increased indirect 9 infection risk was found to be limited in ventilated space. Nevertheless, direct entrainment of 10 an infectious exhaled flow by PV should be strictly avoided. 11

With respect to the design of PV required to protect against airborne transmission, the 12 following aspects should be considered: (1) PV should be carefully oriented away from an 13 adjacent occupant's breathing zone; (2) high efficiency PV with reduced entrainment of 14 ambient particle-laden air is preferable; and (3) 

As PV is a ventilation mode that directly interferes with a human microenvironment, the flow 7 interactions within this microenvironment should be carefully considered to identify the best 8 design for optimal infection control. Advanced PV designs should be developed to satisfy 9 requirements for effective interventions in airborne transmission pathways at any point within 10 In this study it was found that the effectiveness of PV in protecting against infectious particle 3 transmission largely depends on the extent to which PV entrains ambient infectious particles. 4

If low entrainment of ambient air by PV can be realized, PV can deliver clean air with high 5 efficiency directly to the breathing zone of the exposed occupant. PV ATDs with wider 6 diameter that can achieve a longer clean air core are recommended for this purpose. to 80% fresh air in the inhaled air with a supply flow rate of less than 3.0 L/s. 17 A desk-mounted round movable arm proposed by Bolashikov et al. [63] can also realize a high 18 clean-air delivery efficiency due to the short distance between the PV ATD and the BZ. 19 Nielsen et al. [18] reported an entrainment-minimized PV ATD to improve inhaled air quality; 20 this does not use the jet entrainment concept but locates the source of clean air in the 21 boundary layer close to the BZ. This idea was based on the body's close contact with textile 22

surfaces such as a neck support pillow or a backrest cushion of a seat. 23

Such designs would guarantee that PV would function with high efficiency at the BZ of the 24 protected person. The PV ATD used in this study can only obtain a clean-air delivery 1 efficiency of 40%, and was shown to not be very effective in protecting an exposed occupant 2 against infectious particles, especially at a low air-flow rate. Higher efficiency PV systems 3 may help to minimize the infection risk for an occupant. 4

Another way to increase the benefits of PV is to block the transmission route between two 5 occupants. From Figs. 9 and 13, it can be seen that direct exposure occurs when the 6 exhalation flow penetrates the exposed occupant's BZ. If a partition with sufficient height can 7 be set between the occupants to block the penetration of the exhaled flow, the possibility of 8 PV's direct entrainment of the exhaled flow from another occupant will be substantially 9 lowered. Such partitions are commonly used in office rooms, and thus could also be used in 

The breathing from a manikin could be different to that of humans [30] , which means that 8 these findings deviate from the reality in some aspects. In addition, it was assumed that the 9 average concentration of droplets reaching human subjects would all be infectious, but this 10 would not be true in reality because not all of the expelled droplets would contain infectious 11 particles, which would also lead to over-estimates of the infection risk of influenza A. For 12 mouth breathing, most measurements from human subjects have a downward flow-trend 13 angle rather than the upward flow from the manikin. Moreover, the propagation distance of 14 exhaled flow in humans was less than that of the manikin. These deviations would also 15 contribute to overestimation of exposure risk. 16

The alignments of PV with specific distances and angles in this study are not representative 17 configurations of all PV systems, and thus the conclusions are not generalizable to all PV 18 systems. The performance of PV in disease control could be affected by a number of factors, 19 such as background ventilation, the supplied air temperature and the relative distance between 20 occupants, and these should be investigated in future studies. 21

It was previously reported that PV protection technology was able to mitigate the spread of 23 airborne transmissible disease, but this research area is new. This paper was focused on 24 evaluating the effect of PV on the probability of cross infection between two occupants in 1 close proximity (<1 m). Two scenarios with different alignments of PV were compared. 2

Concerns about the possibility of PV facilitating the transport of exhaled pathogen were 3 examined in two ways . 4 In the first, the infector was subjected to PV, meaning that the PV flow interacted with the 5 exhaled flow from the source, resulting in accelerated mixing of exhaled droplets with room 6 air and occasionally increased exposure of the susceptible person to particles. Over 90% of 7 the droplets to which the susceptible person was exposed were derived from indirect 8 transmission when only the source manikin used PV. 9

In the second, the PV interacted with the infector's exhaled flow, which increased the direct 10 exposure level. When the PV flow fell within the area accessible to the exhalation of the 11 infector, the PV jet entrained more droplets, resulting in an increased direct exposure to the 12 susceptible person of more than 50%. 13

The infection risk of influenza A was evaluated based on the dose-response model. With a 14 continuous exposure time of 2 h, the infection probability varied from 0.28 to 0.85 according 15 to different PV alignments. PV displayed a potential to reduce infection risk only when PV 16 was operated with a high efficiency and successfully supplied clean air to the breathing zone 17 of the protected person. This was achieved when both the source and target were subject to 18 PV at a flow rate of 9 L/s in both scenarios, indicating that a higher clean air volume and no 19 direct entrainment from the infector's exhalation would aid in protection against disease 20 transmission to the exposed occupant. 21

The complex interaction of airflows in the microenvironment between occupants was found to 22 affect the efficiency of PV and the transmission of droplets. As PV is a ventilation mode that 23 directly perturbs a human microenvironment, the flow interactions between these airflows 24 should be carefully considered to achieve an optimal design for infection control. Use of PE 1 for infectious source control, measures to block the penetration of exhaled flow and designs 2 for improving the efficiency of PV ATDs would be beneficial to maximize the effective 3 application of PV for airborne disease control within the human microenvironment. 

CDC influenza programs protect

Aerosol transmission of influenza A virus: a review of new studies

COVID-19 epidemic statistics

The 2019-nCoV epidemic control strategies and future 20 challenges of building healthy smart cities

Airborne spread of infectious agents in the indoor environment

Bed with Integrated Personalized Ventilation for 12

Minimizing Cross Infection

Short-range airborne transmission of 15 expiratory droplets between two people

Short-range airborne route dominates 18 exposure of respiratory infection during close contact

Co-occupant's exposure to exhaled pollutants with two types of 21 personalized ventilation strategies under mixing and displacement ventilation systems

Human convective boundary layer and its 24 interaction with room ventilation flow

Impacts of airflow interactions with 2 thermal boundary layer on performance of personalized ventilation

Personalized ventilation: One possible solution for airborne infection 5 control in highly occupied space?

A review of advanced air 10 distribution methods-theory, practice, limitations and solutions

Improved inhaled air 13 quality at reduced ventilation rate by control of airflow interaction at the breathing zone 14 with lobed jets

Exposure of health care 17 workers and occupants to coughed airborne pathogens in a double-bed hospital patient 18 room with overhead mixing ventilation, HVAC&R Res

Control of the Free Convective Flow around the 21

Personalized Ventilation Unit

) Cleanrooms and associated controlled environments

Influence of air stability and 5 metabolic rates on exhaled flow

Airborne spread of expiratory droplet nuclei between the 8 occupants of indoor environments: A review

Effectiveness of a personalized ventilation system in 11 reducing personal exposure against directly released simulated cough droplets

Distribution of 14 exhaled contaminants and personal exposure in a room using three different air 15 distribution strategies

Influence of human breathing modes on 18 airborne cross infection risk

The size distribution of droplets in the exhaled breath of 21 healthy human subjects

Inhaling to mitigate exhaled bioaerosols

Size distribution and sites of origin of droplets 3 expelled from the human respiratory tract during expiratory activities

Evaporation and dispersion of respiratory droplets from 6 coughing

Dispersion of expiratory 8 aerosols in a general hospital ward with ceiling mixing type mechanical ventilation 9 system

How far droplets can move in indoor 12 environments: revisiting the Wells evaporation-falling curve

Towards understanding the risk of secondary 15 airborne infection: emission of respirable pathogens

Measuring the administered dose of particles on the facial mucosa of a realistic 19 human model

Aerosol emission and superemission during human speech increase with voice loudness

Risk of cross-infection in a hospital ward with 24 downward ventilation

Experimental variation of 2 the personal exposure in a hospital room influenced by wall heat gains

Airborne 5 cross-infection risk between two people standing in surroundings with a vertical 6 temperature gradient, HVAC&R Res

Aerosol technology

A methodology for 10 estimating airborne virus exposures in indoor environments using the spatial distribution 11 of expiratory aerosols and virus viability characteristics

Experimental tularemia in Macaca mulatta: relationship of 14 aerosol particle size to the infectivity of airborne pasteurella tularensis

Temperature-sensitive mutants of influenza virus. 3. Further characterization of the ts-1

A recombinant (H3N2) virus in man

Quantifying the routes of transmission for pandemic influenza

Human influenza resulting from aerosol 23 inhalation

A time-based analysis of the personalized 17 exhaust system for airborne infection control in healthcare settings

Performance evaluation of a novel 20 personalized ventilation-personalized exhaust system for airborne infection control

• Interactions between PV and infectious exhalation can facilitate the transport of exhaled pathogens.• PV used for the infector increases indirect exposure to the exposed occupant.• Direct entrainment of an infectious exhaled flow by PV should be strictly avoided.• Infection risk is lowered when PV delivers clean air to inhalation without direct entrainments from an infectious exhaled flow.

☒ 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.☐The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: