key: cord-0005489-o99op1qv
authors: Ai, Zhengtao; Mak, Cheuk Ming; Gao, Naiping; Niu, Jianlei
title: Tracer gas is a suitable surrogate of exhaled droplet nuclei for studying airborne transmission in the built environment
date: 2020-02-21
journal: Build Simul
DOI: 10.1007/s12273-020-0614-5
sha: 66f94278f2402d795a96d548dc6cd387484d7a9f
doc_id: 5489
cord_uid: o99op1qv

nan

Airborne transmission between persons is a valid and major respiratory transmission route for a number of infectious diseases (CDC 2007 (CDC , 2017 Hodgson et al. 2012; Li et al. 2015) . Worldwide there are approximately 3.0 million deaths annually caused by lower respiratory infections (WHO 2018). Airborne transmission has been shown to occur within an enclosed indoor space (Liu et al. 2017; Ai and Melikov 2018) , between different rooms on a same floor (Wu et al. 2016) , between different flats in the same building (Gao et al. 2009; Ai and Mak 2016; Mu et al. 2016) and even between adjacent buildings in high-density urban areas (Yu et al. 2004) . Understanding the transmission characteristics, high-risk routes and influential factors is of great importance for formulating effective control measures, which would in turn contribute to reduced number of deaths and illness leaves as well as lowered health care costs. The whole process of airborne transmission between persons involves the generation of infectious droplets from an infected person, the spread of infectious droplet nuclei in the air and the inhalation of infectious droplet nuclei by an exposed person (Nicas et al. 2005; Tang et al. 2006; Wei and Li 2016) . Human breathing, talking, coughing and sneezing can generate tens of thousands of droplets (Duguid 1946; Tellier 2006; Chao et al. 2009 ). Most of these droplets evaporate instantaneously to half of their initial size and become droplet nuclei (Nicas et al. 2005) . These droplet nuclei spread in the air with the influence of factors like gravity, inertia, ventilation flow, human body boundary layer flow and respiratory flow (Melikov 2015; Nazaroff 2016) . The spread of droplet nuclei in the air is usually evaluated by concentration distribution obtained using either experimental measurements or computational fluid dynamics (CFD) methods. The droplet nuclei are commonly simulated using a tracer gas or particles, which are termed as tracer gas simulation or particle simulation in this paper.

Compared with tracer gas simulation of the exhaled droplet nuclei, there are several deficiencies and difficulties associated with particle simulation. Firstly, in experimental contexts, the experimental space for particle simulation has to meet certain criteria so that the particles from other sources would not influence the measurements. Examples of these sources include particles in background and supply air and the resuspension of particles from surfaces of the space. Secondly, particle generators are normally stand-alone from a thermal manikin, where the influence of the presence of occupants and their thermal boundary conditions was not taken into account (Sze et al. 2009; Poon and Lai 2011; Pantelic and Tham 2013; Pantlic et al. 2015; Cao et al. 2015) . Very few studies have successfully dosed aerosols through the breathing system of a thermal manikin. The major reason reported was that the integration of an aerosol generator into the breathing system of a manikin would easily result in the blockage of the breathing system, due to the narrow "respiratory tract" and the relatively low expiratory speed of aerosol particles (Ai and Melikov 2018). Thirdly, the aerosol generators reported in past studies could not simulate accurately the number and size distribution of human expiratory droplet nuclei (Sze et al. 2009; Poon and Lai 2011; Pantelic and Tham 2013; Cao et al. 2015; Pantlic et al. 2015) , though most generators produce a range of particle sizes simultaneously (Sze et al. 2009 ). In addition, it is unknown to what extent the particles produced by these available generators are the same with expiratory droplet nuclei, in terms of quantity, size distribution, thermo-fluid properties and dynamics. Fourthly, the particle dynamics and fate are complicated (see Fig. 1 ), which is required in order to set up and perform the experiments properly and to analyze the data and explain the phenomena quantitatively. The movement of particles in turbulent flow is governed by gravitational force, thermophoretic force, Brownian force, and Saffman's lift force (Nazaroff 2004 (Nazaroff , 2016 . In addition, particles could deposit on surfaces, re-suspend from surfaces, coagulate, change phase etc. (Nazaroff 2004 (Nazaroff , 2016 . In numerical context, fully considering all dynamics and fates of particles in airflows in the built environment is impossible at the present. Knowledge on the relative importance of each force and fate in connection with particle size and particle volume loading in air is required in order to make reasonable assumptions and simplifications in numerical models. Fifthly, there are quite several uncertainties associated with the risk of cross infection, including the number concentration and size distribution of droplet nuclei generated by each respiratory process including coughing, sneezing and speaking, the frequency of the respiratory process, the infectivity of the virus concerned, the lifetime of the virus, the vulnerability of the exposed person. In addition, the droplet concentration and its size distribution can be influenced by many factors, such as body weight, gender, and age (Yang et al. 2007; Johnson and Morawska 2009; Zayas et al. 2012) . It is reported that, generally for a same person, among the processes of mouth exhalation, nose exhalation, coughing and speaking, coughing produces the largest droplet concentrations (Lindsley et al. 2012; Morawska et al. 2013 ) and nose exhalation the least (Papineni and Fig. 1 The dynamic dispersion behavior of droplet nuclei in the air is influenced by the interaction of flows (ventilation flow, human body convective flow and breathing flow), the combination of forces (gravitational force, thermophoretic force, Saffman's lift force and Brownian force) exerted on a nucleus, and some other effects (e.g., coagulation, deposition on and resuspension from solid surfaces, and phase change)

Rosenthal 1997). The error caused by these uncertainties may be comparable with that of tracer gas simulation of particles. Sixthly, some studies (Zhu et al. 2006; Rim and Novoselac 2009; Poon and Lai 2011; Yang et al. 2016; Liu et al. 2017 ) examined the dispersion of particles of different sizes, but most of them had no conclusion on the influence of particle size on the risk of cross infection, primarily because of the uncertainties mentioned above. In summary, the particle simulation cannot ensure an accurate modeling of the dispersion of expiratory droplet nuclei between persons, and the knowledge deficiency and the difficulties attached decrease the probability of obtaining reliable results. The statement in this paper is that tracer gas is a suitable surrogate of exhaled droplet nuclei for studying airborne transmission in the built environment. Most studies on airborne transmission between occupants in different spatial scales in the built environment were based on tracer gas simulation (e.g., Qian et al. 2008; Bolashikov et al 2015; Lipczynska et al. 2015; Yang et al. 2015; Villafruela et al. 2016; Liu et al. 2017; Ai et al. 2019) . The reasons for the popularity of tracer gas simulation should be attributed to not only the deficiencies and difficulties of particle simulation at the present stage, but also the well established tracer gas simulation techniques (Sherman 1990) , their simplicity and good accuracy (described in Section 3). With the consideration that the main purpose of engineering studies of airborne transmission is to reveal the dispersion characteristics and its influential factors, to identify the major and highrisk dispersion routes, and to develop control measures, the proper use of tracer gas technique, and its pros and cons are reviewed in this paper.

There are four major counter arguments as listed below:  Particle dynamics differs from gases due to the effects of gravity, inertia and deposition at solid surfaces.  Airflow with particles is a two-phase flow.  The coagulation and resuspension of particles cannot be simulated by tracer gas.  Particles have different aerodynamic diameters.

As the density of a particle is usually several orders of magnitude higher than air, a particle travelling in air is strongly influenced by its gravity and inertia (Hinds 1999; Nazaroff 2004 ). In addition, a particle could deposit and be collected by solid surfaces after deposition, which is one of the most important fates of a particle in the air (Nazaroff 2004) . A tracer gas in the air does not have such dynamic characteristics, and it is therefore no doubt that the gravity, inertia and deposition of a particle cannot be modelled using a tracer gas. In addition, the material of a particle, namely fluid or solid, determines the fact that a mixture of airflow and particles is typically a two-phase flow, where particles could departure from the main airstream and could even influence the airflow. Moreover, the coagulation and re-suspension are two important dynamic properties of particles (Thatcher and Layton 1995; Nazaroff 2004) . The coagulation occurs when particles collide with one another and adhere together, which shifts the size distribution of particles in the air. The most important mechanism causing coagulation is Brownian motion, which increases the chance of collision between particles. The important influence of coagulation on the concentration of particles emitted from unvented natural gas combustion was reported by Dennekamp et al. (2001) . Particles that have been deposited at solid surfaces may become re-suspended into the air by many mechanical activities including walking. A few studies have observed the increase of particle concentration indoors due to resuspension resulted by human activities (Thatcher and Layton 1995; Ferro et al. 2004) . Obviously, both coagulation and resuspension of particles cannot be simulated using a tracer gas. Furthermore, the droplet nuclei exhaled by human beings have different aerodynamic diameters, which could range from zero to dozens of micrometer (μm). The key attributes of particles, such as source, dynamic behavior and fate, are all strongly related to particle size (Nazaroff 2004) . However, the dispersion characteristics of droplet nuclei with different size bins cannot be investigated using a tracer gas.

All these counter arguments are correct. However, the importance of the particle dynamics mentioned above is strongly related to particle concentration and size distribution. For airborne transmission that is based on human expiratory droplet nuclei, they are less important and it is therefore believed that the tracer gas simulation is a suitable research method. The major reasons are listed below.

Firstly, the respiratory droplet nuclei were found to range in diameter from 0.25 to 42 μm, of which 97% between 0.5 to 12 μm and the most common between 1 and 4 μm (Duguid 1946) . Settling times for a 3-m fall are 17 minutes for 10 μm and 62 minutes for 5 μm, while droplet nuclei with a diameter less than 3 μm essentially do not settle (Tellier 2006) . The relaxation times for fine particles are much smaller than the typical time scales of indoor airflow; droplet nuclei with 10 μm in diameter take only 1.55 ms to reach 99.3% of their terminal speeds indoors. These three aspects may support the fact that the main driving force of the movement of the majority of expiratory droplet nuclei is airflow, rather than gravity (Nazaroff 2016 ). In addition, deposition at the solid surfaces is an important mechanism influencing the fate of droplet nuclei, only when the particle size is relatively large and the air exchange rate is small. It was reported that, for particles in the size range 0.5-1.0 μm, the deposition loss rate coefficient is 0.2-0.3 per hour for enclosed indoor environment (Thatcher et al. 2002; Nazaroff 2016 ). If, for example the air change rate is 3 h −1 , the airborne particles have a characteristic residence time of 20 minutes in the room air, during which the probability of deposition for 0.5-1.0 μm particles is only 0.067-0.1. Note that the air change rate of a high-risk space or during a high-risk period, such as at hospital wards during the outbreak of infectious diseases, can be much higher than 3 h −1 . Such a negligible deposition loss of fine particles was also supported by our CFD validation (Gao and Niu 2007) against a particle experiment (Chen et al. 2006) , where 96% of 1.0 μm particles was found at ventilation exhaust, meaning that the deposition loss rate is only 4%.

Secondly, the particle loading is the key parameter determining if the particle-flow interaction is important (Elghobashi 1994) . When the particle loading represented by volume fraction is higher than 10 −6 , the influence of particles on turbulent flow would be important and the two-phase flow should be considered. The concentration of droplet nuclei generated by respiratory processes is generally much lower than this threshold loading. It was reported that 1.1-6.7 mg of saliva were collected on a mask during a single cough, and 18.7 mg were collected while counting from 1-100 (Zhu et al. 2006; Xie et al. 2009 ). A study by Chao et al. (2009) shows that the estimated total number of droplets at all sizes ranged from 947 to 2085 per cough and 112-6720 for speaking. Equivalently, the estimated droplet concentrations for coughing ranged from 2.4 to 5.2 cm −3 per cough and 0.004-0.223 cm −3 for speaking, which are much smaller than the threshold loading. However, as the frequency of respiratory processes is case dependent, the accurate loading of droplet nuclei is difficult to estimate. Many past CFD studies (Chen et al. 2006; Gao and Niu 2007; Gao et al. 2009; Zhao and Zhao 2016; Xu and Wang 2017; Mei and Gong 2019) using Eulerian method that treats the fluid phase and the particulate phase as continuum reported good agreements with experimental results, indicating that the simplification of one-way coupling for human expiratory droplet nuclei is reasonable.

Thirdly, coagulation due to particle-particle interaction is important for very fine particles with several sources, when particle concentrations are high (Nazaroff 2004 ) and air exchange rate is low (Rivas et al. 2015) . Kumar et al. (2011) reported that, depending on the duration of interests, particle coagulation may be neglected for particles larger than 0.01 μm, as the coagulation is too slow to substantially affect the number concentrations. According to Hinds (1999) , coagulation rate is proportional to the square of the number concentration. Assuming a diameter of 1.0 μm and a concentration of 10 particles per cm 3 , 1% loss of particle number concentration by coagulation requires about 500 days (Gao et al. 2009 ). Particle re-suspension from a solid surface is influenced by many parameters, such as particle size, surface material, and human activities. However, for fine particles, such as PM 2.5 , resuspension is very difficult and even human activities have a minor influence on it (Thatcher et al. 1995) . Considering also the low deposition rate of fine particles, the resuspension effect may also be neglected.

Fourthly, though the different aerodynamic diameters cannot be examined using a tracer gas, the most important and major portion of airborne droplet nuclei exhaled by human beings falls in the range of less than 5-10 μm, especially within 2-3 μm. In addition, a tracer gas certainly cannot simulate large particles like 10 μm in diameter. However, in this paper, we would like to claim that a tracer gas is suitable to simulate the commonest range of expiratory droplet nuclei that is less than 5-10 μm. Justification of this claim is described in detail in Section 3.

Tracer gas is a suitable surrogate of exhaled droplet nuclei for studying airborne transmission in the built environment, mainly because of the following reasons:  If a disease is found to be transmitted via airborne route, it is mostly the fine droplet nuclei generated by human respiratory processes that are the vehicles of pathogens.  Studies show that the movement of fine particles less than 3-5 μm can be well represented with tracer gas simulation.  Tracer gas simulation is of lower complexity and less demand on user knowledge and thus it is easier to obtain reliable results.

A study of ferrets found that transmission of influenza from infected to susceptible ferrets occurred despite the ferrets being separated by a long, straight air duct or by "S" or "U"-shaped ducts (Andrewes and Glover 1941). As large respiratory droplets could not move around the bends of the ducts whereas the smaller droplet nuclei could, this study may well indicate that fine airborne droplet nuclei were the major vehicles of pathogen, and that they could remain suspended in air for a prolonged period and be transported over an extended distance by airflows. Another study pointed out that the majority of droplets from human respiration activities are less than 5-10 μm in diameter (Nicas et al. 2005) , and this size range is used to differentiate between airborne transmission and droplet transmission. Another earlier study by Duguid (1946) reported that 97% of human respiratory droplet nuclei ranged in diameter from 0.5 to 12 μm and the commonest range between 1 and 4 μm (see Fig. 2 ). In addition, the size of viruses is mostly from 0.02 to 0.3 μm. For examples, an individual SARS coronavirus ranges from 0.075 to 0.16 μm, and an influenza virus is of a similar size (Morawska 2006) . Lindsley et al. (2010) measured influenza virus in droplet nuclei generated by coughing patients and reported that 42% of detected viruses were found in droplet nuclei less than 1 μm and 23% in droplet nuclei of 1-4 μm, which suggested that most viruses are contained within droplet nuclei in the very fine size range. Yang et al. (2011) confirmed that a substantial fraction of the detected influenza A viruses was associated with fine particles smaller than 2.5 μm. It is known that fine particles travel readily between persons and penetrate and deposit deeply in the respiratory tract. Based on a semi-empirical model (Yeh et al. 1996) , particles around 3 μm have the highest deposition efficiency in the respiratory tract (Nazaroff 2016) . 

A few studies compared the dispersion of tracer gas and particles, which show that the accuracy of tracer gas simulation of fine particles can be good enough. These studies are briefly described below. Results of two studies are also summarized in Fig. 3 . Bivolarova et al. (2017) investigated the influence of ventilation rate, free convection flow produced by a thermal manikin, and the presence of objects on the distribution of tracer gas (N 2 O) and particles (0.07, 0.7 and 3.5 μm) in a full-scale room. Three scenarios were examined, including an empty chamber, an office room with an occupant sitting in front of a table, and a single-bed hospital room. The concentrations of gas and particles were measured in the bulk room air, in the breathing zone of the manikin, and in S), where "3.5 ACH-Empty" denotes that the air change per hour was at 3.5 and the test chamber was empty, "Manikin off " denotes that one manikin without heat release was presented in the test chamber, and "Manikin on" means that the heating function of the manikin was switched on so that it simulated the heat release of an average person; (b) CFD results by Li et al. (2011 Li et al. ( , 2013 (reproduced with permission © 2011 Taylor & Francis and © 2012 John Wiley & Sons A/S, respectively), where "Intake fraction" is defined as the proportion of pollutant mass exhaled from the infected person that is then inhaled by the exposed person, and "MV", "UFAD" and "DV" denote mixing ventilation, under floor air distribution, and displacement ventilation, respectively the exhaust air. It was found that, within the breathing zone of the sitting manikin, the tracer gas was a reliable predictor for the exposure to all the three investigated particles, regardless of the change of ventilation rate and room surface area. Zhang et al. (2009) experimentally and numerically investigated contaminant transport in a section of half occupied, twin-aisle cabin mockup, where a tracer gas (SF 6 ) and particles (0.7 μm) released at the same location were used to simulate a gaseous and a particulate contaminant. The results show that the distributions of the gaseous and particulate contaminants were similar in most part of the cabin (except for the region near ceiling), suggesting that, in occupied zone, such very fine particles behave like a passive tracer gas. Noakes et al. (2009) experimentally compared tracer gas technique and tracer particle technique for evaluating the behavior of bioaerosols in hospital isolation rooms with mixing air distribution at 10 ACH (air change per hour). Both the tracer gas (N 2 O) and particles (including 3-5 μm) were released from a heated cylinder simulating a patient in bed. The results show that both N 2 O tracer gas and 3-5 μm particles compared well with bioaerosol data, suggesting both techniques gave a good representation of the behavior of bioaerosols. Gao and Niu (2007) modelled the particle dispersion and deposition in a room with typical office settings including a thermal manikin, a computer, a desk, ceiling lights etc. The room was conditioned by a mechanical ventilation system. Tracer gas simulation was performed for comparison purpose. The particle dynamics were treated by Eulerian approach in combination with a drift-flux model. It was found that the movements of particles not larger than 2.5 μm were like a tracer gas. Human exposure to this level of particles was also very close to a tracer gas, when the two were released at the same location. Li et al. (2011 Li et al. ( , 2013 simulated the spatial distribution of human respiratory droplet nuclei and the risk of cross infection between two face-to-face persons in a typical office room under different ventilation methods. Both particles (1, 2.5, 5, and 10 μm) and a tracer gas (CO 2 ) were examined. The results show that both the spatial distribution and co-occupant's exposure of particles not larger than 2.5 μm are very close to the tracer gas.

A study by Beato-Arribas et al. (2015) concluded that the distributions of CO 2 tracer gas and aerosolized Bacilus Subtilus bacteria are similar in a single isolation hospital mock-up at 12 ACH. A CFD study by Gao et al. (2009) showed that the dispersion characteristics and concentration distributions of CO 2 tracer gas and 1.0 μm particles are very close in a building with multiple stories.

In summary, these studies indicate that tracer gas simulation is accurate enough to study the dispersion of particles not larger than 3-5 μm. This size range is the dominant part of human expiratory droplet nuclei in terms of quantity.

As an alternative, tracer gas simulation is a well established research method both in experimental and numerical contexts. Experimentally, tracer gas simulation has been widely performed in different spatial scales, from a single room up to an urban area. For a single-room scale, tracer gas experiments do not have a special requirement on the cleanness of the experiment room, and the characteristics and area of the surfaces of the room do not obviously influence the experimental results. For an urban-area scale, there are very strong tracer gas sources to generate a large amount of tracer gas that allows the measurement at locations a few kilometers away from the source. Both the dosing system and the sampling system have been developed to a very high accuracy of control. The dosing system can be perfectly integrated with a breathing thermal manikin to include the influence of the presence of manikin body and its thermal boundary layer. The mixture of a tracer gas with air is a one-phase flow, which has nearly the same properties of air. There are no complicated dynamic processes, such as evaporation, condensation, coagulation, resuspension and phase change. Similarly, numerical simulation of tracer gas dispersion is much easier than particle simulation in terms of mathematical model complexity and practitioners' knowledge. In numerical context, a tracer gas is considered to be a passive scalar and only one more mass transport equation requires to be solved. The sensitivity of the influential parameters and constants (such as density, mass fraction, Schmidt number, and diffusion coefficient) related to this equation are easy to be tested. There are no assumptions and simplifications required for modelling the dispersion of tracer gas. Because of the lower complexity and less demand on user knowledge, it is easier for tracer gas simulation to obtain reliable results. This is especially important from the viewpoint of engineering practice.

Tracer gas is a suitable surrogate of exhaled droplet nuclei for studying airborne transmission in the built environment, because (a) if a disease is found to be transmitted via airborne route, it is mostly the fine droplet nuclei generated by human respiratory processes that are the vehicles of pathogens, and fine droplet nuclei have aerodynamics more close to a gas, (b) studies show that the movement of fine particles less than 3-5 μm can be well represented with tracer gas simulation, and (c) tracer gas simulation is of lower complexity and less demand on user knowledge and thus it is easier to obtain reliable results. Note that dispersion and deposition of droplet nuclei in the respiratory tract were not discussed in this paper.

As reviewed, tracer gas simulation and particle simulation of droplet nuclei have different error sources. Few studies have evaluated quantitatively the error of particle simulation. It would be interesting to assess whether or not the error of tracer gas simulation is smaller than the error of particle simulation, when compared to the real-life dispersion of exhaled droplet nuclei. In addition, few studies have compared the dispersion characteristics of different tracer gases, and it is still unknown how different are the experimental results obtained by using different tracer gases (i.e., different densities, diffusion coefficients and Schmidt numbers). Further studies are required.

Particle simulation of exhaled droplet nuclei may be necessary when the deposition becomes the dominating factor such as when the airflow passes through narrow passages (e.g., window or door gaps). In addition, particle simulation would potentially become an effective method for airborne transmission studies in the built environment, if (a) a particle generation system is developed for experimental studies to a level that the number and size distribution of the generated particles can be accurately controlled, (b) in experimental settings the factors including the background concentration that influence the measurement accuracy and repeatability can be conveniently eliminated, and (c) numerical models are advanced to be able to take into account the major dynamics and fates of particles.

Department of Health and Human Services

Characterization of expiration air jets and droplet size distributions immediately at the mouth opening

Modeling particle distribution and deposition in indoor environments with a new drift-flux model

Ultrafine particles and nitrogen oxides generated by gas and electric cooking

The size and duration of air-carriage of respiratory droplets and droplet-nuclei

On predicting particle-laden turbulent flows

Source strengths for indoor human activities that resuspend particulate matter

The airborne transmission of infection between flats in high-rise residential buildings: Particle simulation

Modeling particle dispersion and deposition in indoor environments

Aerosol Technology: Properties, Behavior, and Measurement of Airborne Particles

Airborne Infectious Diseases. ASHRAE Position Document

The mechanism of breath aerosol formation

Dynamics and dispersion modelling of nanoparticles from road traffic in the urban atmospheric environment-A review

Measurements of airborne influenza virus in aerosol particles from human coughs

Quantity and size distribution of cough-generated aerosol particles produced by influenza patients during and after illness

Thermal environment and air quality in office with personalized ventilation combined with chilled ceiling

Short-range airborne transmission of expiratory droplets between two people

Spatial distribution of human respiratory droplet residuals and exposure risk for the co-occupant under different ventilation methods

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

Engineering control of respiratory infection and low-energy design of healthcare facilities

Characterizing transport and deposition of particulate pollutants in a two-zone chamber using a Markov chain model combined with computational fluid dynamics

Human body micro-environment: The benefits of controlling airflow interaction

Droplet fate in indoor environments, or can we prevent the spread of infection? Indoor Air

Indoor aerosols: from personal exposure to risk assessment

Wind tunnel tests of inter-flat pollutant transmission characteristics in a rectangular multi-storey residential building, Part A: Effect of wind direction

Indoor particle dynamics

Indoor bioaerosol dynamics

Toward understanding the risk of secondary airborne infection: emission of respirable pathogens

Adequacy of air change rate as the sole indicator of an air distribution system's effectiveness to mitigate airborne infectious disease transmission caused by a cough release in the room with overhead mixing ventilation: a case study

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

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

An experimental study quantifying pulmonary ventilation on inhalation of aerosol under steady and episodic emission

Dispersion of exhalation pollutants in a two-bed hospital ward with a downward ventilation system

Transport of particulate and gaseous pollutants in the vicinity of a human body

Outdoor infiltration and indoor contribution of UFP and BC, OC, secondary inorganic ions and metals in PM2.5 in schools

Tracer-gas techniques for measuring ventilation in a single zone

Experimental study of dispersion and deposition of expiratory aerosols in aircraft cabins and impact on infectious disease transmission

Factors involved in the aerosol transmission of infection and control of ventilation in healthcare premises

Review of aerosol transmission of influenza A virus

Effects of room furnishings and air speed on particle deposition rates indoors

Deposition, resuspension and penetration of particles within a residence

Influence of human breathing modes on airborne cross infection risk

The top 10 causes of death. WHO Newsroom. World Health Organization

On-site measurement of tracer gas transmission between horizontal adjacent flats in residential building and cross-infection risk assessment

Exhaled droplets due to talking and coughing

CFD modeling of particle dispersion and deposition coupled with particle dynamical models in a ventilated room

The size and concentration of droplets generated by coughing in human subjects

Concentrations and size distributions of airborne influenza A viruses measured indoors at a health centre, a day-care centre and on aeroplanes

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

Person to person droplets transmission characteristics in unidirectional ventilated protective isolation room: The impact of initial droplet size

Comparisons of calculated respiratory tract deposition of particles based on the proposed NCRP model and the new ICRP66 model

Evidence of airborne transmission of the severe acute respiratory syndrome virus

Cough aerosol in healthy participants: Fundamental knowledge to optimize droplet-spread infectious respiratory disease management

Experimental and numerical investigation of airflow and contaminant transport in an airliner cabin mockup

Numerical assessment of particle dispersion and exposure risk in an underground parking lot

Study on transport characteristics of saliva droplets produced by coughing in a calm indoor environment

This study was supported by the National Natural Science Foundation of China (No. 51908203).

Ai ZT, Mak CM (2016). Large eddy simulation of wind-induced interunit dispersion around multistory buildings. Indoor Air, 26: 259-273.

Zhengtao Ai is a full professor of building thermal environment at Hunan University. His area of expertise includes CFD simulation of flow and dispersion in coupled indoor and outdoor environment and climate chamber test of spread of human exhaled droplet nuclei indoors.Cheuk Ming Mak is a full professor of building environment at The Hong Kong Polytechnic University. His area of expertise covers building physics and urban physics. He has been for over 25 years focusing on CFD simulation of natural ventilation and related dispersion process in buildings. Naiping Gao is a full professor of thermal engineering and environment at Tongji University. His area of expertise covers building ventilation and indoor air quality, especially including CFD simulation of human microenvironment, human exposure, and particle dispersion and deposition in indoor environments.Jianlei Niu is a chair professor of building environment and energy at The Hong Kong Polytechnic University. His research interests are indoor air quality, thermal comfort and relevant energy-efficient solutions. He is one of the first researchers who investigate into the spread of SARS viruses in the built environment using field measurements, wind tunnel experiments and CFD simulations.