Exposure generated by cough released droplets in the indoor environment a comparison among four ventilation systems

The knowledge gap is obvious.‖ Quantification of ventilation rates will influence ability of a particular ventilation system to control reduce airborne infectious disease transmission by

Chapter 1: Introduction

1.1 Background and Motivation

People in the developed countries spend more than 80 %of their time indoors (Robinson and Nelson, 1995) Many of the indoor pollutants are suspended particles in air The human respiratory tract handles 10 l/min of air and 3000 cycles of inhalation – exhalation per day Suspended droplets can penetrate into human respiratory system carried by inhaled air potentially causing acute or chronic health effects Viruses have been identified as the most common cause

of infectious diseases acquired within indoor environments, in particular those causing respiratory and gastrointestinal infection

Particles present in the indoor environments can originate from outdoors penetrating through building envelope drawn in by ventilation systems or can be emitted from the indoor source Indoor source can be cooking, building materials, consumer products or occupants Respiratory secretions from an infected person can be aerosolized through expiratory activities (breathing, talking, coughing, sneezing and vomiting) and dispersed through indoor environment Each of these expiratory activities produces different size distribution of droplets, amount of infectious agents and initial velocities Coughing and sneezing produces much higher number of droplets than breathing and talking although the latter two are much more frequent These expiratory pathogen laden respiratory droplets are believed to be responsible for the epidemic spread of several respiratory tract infections Epidemiologic studies implicate the droplet nucleus mechanism in the transmission of tuberculosis, measles, influenza, smallpox, chickenpox and SARS (Li et al., 2007) Vomiting can spread 107 virus particles per ml of vomit (Baker et al., 2001) Spread of viral infections through atomized vomit is a significant route of infection in

diseases which causes frequent vomiting such as Norwalk Like Virus (NLV), also vomiting by SARS infected person on the corridor of Metropol Hotel in Hong Kong is believed to have caused

a series of infections Infected individuals can release 1012 virus particles per gram of feces (Baker et al., 2001) or up to 105 CFU/ml of bacteria

Motion of infectious droplets in a ventilated room depends on ventilation air pattern, droplet size, density, number, pollution source location, etc Among these parameters the ventilation air pattern is the most important parameter influencing airborne infectious disease transmission in the indoor environment (Morawska 2006) Droplets in the air are subjected to Brownian forces, gravity, turbulent diffusion, inertial forces, RH, thermal gradients, electrical forces, electromagnetic radiation Depending on the droplet size above mentioned forces have different magnitude of influences on droplets dispersion

Several technologies have been used to reduce the amount of infectious droplets in the indoor air Ultraviolet radiation (UV) in the wavelength of 254nm is germicidal and has been used for air disinfection within indoor environment (Riley, 1972) Efficiency of UV lighting depends on: intensity of UV radiation, the species of organisms, and RH Wavelength of UV radiation is irritating to skin and eyes and it cannot be permitted to impinge on people in doses above the limit recommended by National Institute of Occupational Safety and Health (NIOSH, 1972) Due

to this restriction UV lighting is placed between occupants head and ceiling This influences the efficiency of UV lights and technology have not been applied outside medical facilities where risk of outbreak of airborne infectious diseases is high (Bloch et al., 1985; Gustafson et al., 1982, Catanzaro, 1982) Room air filtration is another technology applied for air purification and disinfection The effectiveness of in-room filtration depends on: single-pass filter efficiency, air flow rate through the filter and on other features of the indoor environment particularly the relative positions of the source and receptor and the indoor air flow patterns generated by air

conditioning and ventilation system (Miller-Leiden et al., 1996) Although this process have shown high efficiency to droplets with diameter less than 3μm when filter is placed close to the occupant, while for droplets with diameter of less than 1μm the efficiency is low when HEPA filter is not used Using HEPA filters increase energy consumption necessary for fan operation to overcome high pressure drop of filter This technology also has not found any application outside medical facilities Two main reasons preventing usage of UV lighting and room air filtration in office building are the low efficiency and high amount of energy they require for operation In the indoor environment where risk of infectious disease spread is not high usage of these technologies is not economically justified

Ventilation air patterns can be used to prevent airborne infectious disease spread Several ventilations systems are currently used in office environments These systems generate very different air patterns in the indoor environment Mixing Ventilation (MV) is the most commonly used ventilation system, which supplies the air with high momentum to induce mixing in the room Displacement Ventilation (DV) is ventilation system which supplies air with low velocity from the supply diffuser mounted on the floor Air supplied in this way moves on the floor until it reaches heat sources which entrain this air into their boundary layers and due to thermal buoyancy displaces upwards Under-floor Ventilation system (UF) supplies the air from diffusers mounted on the floor Air is supplied with high momentum to ensure sufficient upward momentum while air mixes inside indoor environment Personalized Ventilation (PV) is designed

to deliver conditioned (cool and clean) outdoor air to the breathing zone of the occupant The amount of inhaled personalized air has been shown to depend on the design of the Air Terminal Device (ATD), its positioning in regard to the occupant, the supply flow rate of the personalized airflow, as well as the difference between the room air and the Personalized Ventilation airflow temperature, size of target area, etc (Faulkner et al., 1999; Melikov et al., 2002) The optimal performance for the most of the ATD has not exceeded 50–60% of clean air in each inhalation

although it was pointed out by Bolashikov et al., (2003) that PV ATD can be designed to achieve100 % breathing zone air delivery effectiveness When Personalized Ventilation is compared to total volume mechanical ventilation potential for energy savings has been demonstrated (Melikov, 2004; Sekhar et al., 2005) This technology is suitable for commercial office buildings and has potential to reduce risk of airborne infectious disease transmission in the indoor environment while reducing overall energy consumption of air conditioning and mechanical ventilation system

Multidisciplinary literature review by Li et al (2007) concluded that: ―there is no evidence/data

to support specification and quantification of the minimum ventilation requirements in schools, offices, and other non-hospital environments in relation to the spread of airborne infectious diseases The knowledge gap is obvious.‖ Quantification of ventilation rates will influence ability of a particular ventilation system to control (reduce) airborne infectious disease transmission by reducing exposure of the healthy occupant(s) to the infectious agents released by expiratory activities of the infected occupant(s) Ability of the ventilation system to control (reduce) exposure to airborne infectious agents is protective ability of a ventilation system Since several ventilation systems are currently used protective ability of each one should be evaluated and compared to other systems In the present literature there is no methodology proposed for evaluation and comparison of the protective ability of different ventilation systems Several experimental studies (Qian et al., 2006; Qian et al 2008; Nielsen et al 2009; Nielsen et al., 2010)

or Computational Fluid Dynamic studies (CFD) (Li et al., 2005; Xie et al., 2009, Zhu et al., 2007; etc.) compared protective ability of two or more systems for a specific scenario, but no comprehensive comparison of protective ability of commonly used ventilation systems has been performed so far This identifies the necessity for development of the methodology for evaluation

of protective ability of ventilation system This methodology should be then applied to compare

different ventilation systems for the same supply flow rate as well as variation of protective ability of any of the system with variation of the supply flow rate Occupant density needs to be included into the analysis because disease propagation (number of new cases) will depend on the number of occupants in the indoor environment under consideration It is important to establish relationship between number of occupants in the indoor environment, ACH generated by the particular ventilation system and prevention of the airborne transmitted disease propagation

Previous studies examined ventilation efficiency of different system and provided knowledge regarding mechanisms (flow field generated) of air delivery to the breathing zone deployed by different systems These studies were performed under the steady state using tracer gas (SF6) to quantify different ventilation indices Cough release is episodic event which generates unsteady state in the environment When cough is released saliva droplets (with or without infectious agents) move through the environment due to high initial momentum (cough velocity can be up to

22 m/s, while average cough velocity is 10 m/s) before sufficient momentum decay occurs and ventilation streamlines reestablish and start to carry them There were no previous comprehensive experimental studies about protection mechanisms (flow field generated) engaged by different ventilation systems when cough release occurs in the indoor environment Since cough droplets have high initial momentum, distance between infected occupant (infector) and exposed healthy occupant are important parameters that impact on the flow field generated by cough that will cause exposure No previous studies have been conducted to establish influence of the infector-susceptible distance on the exposure generated by infectors cough release and flow fields generated to provide protection of the susceptible occupant using the multiphase flow

approach to simulate cough droplets Study performed by Melikov et al., (2009) investigated influence of distance between infected coughing person and exposed person using tracer gas to simulate cough release.

In order to conduct a study of the cough released droplets exposure changes due to changes of the ventilation system experiments need to be conducted in the indoor environment capable of changing ventilation modes Cough release need to be simulated as a multiphase flow because decay of the cough velocity is much more intensive when only gas phase (commonly simulated using SF6) is used Liquid droplets are able to preserve higher momentum and reach larger distances from the source for the same boundary conditions compared to gas phase Simulated saliva or real human saliva needs to be used to properly simulate nonvolatile saliva characteristics Cough release droplets can evaporate only to droplet nuclei size which than cause exposure, but when, for example, water is used to simulate cough droplets evaporation is complete and there is no nonvolatile residue When gas (SF6) is used to simulate cough evaporation is neglected To assess the risk of airborne infectious disease transmission, exposure

of the occupants to cough released droplets need to be measured This might be achieved by using

a Breathing Thermal Manikin (BTM) to simulate convective boundary layer generated around heated body and inhalation – exhalation flow in the breathing zone Size distribution of cough droplet and droplet nuclei need to be measured in the breathing zone of the BTM to estimate the risk of infection because different droplet sizes have different deposition characteristics in different parts of the respiratory track Some of the previous experimental studies (Chao and Wan 2006; Wan and Chao 2007; Qian et al.; 2006; Cermak and Melikov 2007; Qian et al., 2008; Chao

et al 2008; Nielsen 2009) were conducted with some of the above mentioned equipment, but no study so far used all these equipment to study changes of the exposure to cough released droplets with the application of different ventilation systems

Epidemiological investigation of airborne infectious disease transmissions in the indoor environment (e.g Riley et al., 1978; Catanzaro 1982; Nardell et al., 1991; Nicas 2000) is usually faced with several uncertainties Occupants (infector and all susceptible occupants) move in the indoor environment and change distances, heights (sitting or standing) and position (facing each

other; back to back, or any other possibility in between) in respect to each other Changes of these parameters distance, height and position will cause changes of the susceptible occupants‘ exposure, but these changes are generally unknown When a particular air delivery system is used (e.g MV) type of supply diffusers (perforated, jet and multi-nozzle, conical diffusers, swirl and rectangular 4-way), size of return grilles, positions and number of supply diffusers and return grilles relative to each other will influence air patterns and turbulence levels generated in the indoor environment When cough is released these air patterns interact with multiphase cough flow and influence dispersion of potentially infectious cough droplets Different interactions among multiphase cough flow and generated air patterns will cause different levels of exposure of susceptible occupants Although generated air patterns under steady state environment can be known, due to changes of the distance, height and position of infector and susceptible occupants‘ exposures of the occupants are generally unknown Another important parameter influencing exposure is distribution of cough releases (frequency) throughout exposure time, which is generally also unknown In order to simplify investigation and overcome these uncertainties assumption of perfectly mixed environment and steady release of potentially infectious expiratory droplets are commonly used These assumptions can lead to large risk underestimation (Nicas, 1996) Wells-Riley approach (Riley et al., 1978) or modifications of the original equation incorporating other effects (Nazaroff et al., 1998; Rudnick and Milton 2003; Fisk et al, 2005; Noakes 2006; Noakes and Sleigh 2009) was used in several studies to calculate risk of airborne infectious disease transmission Although some of the studies (Noakes and Sleigh 2009) divided indoor environment into zones, each of the zones was treated as perfectly mixed while source was treated as steady constant release which is simplification of the episodic cough release occuing in the real environment Wells-Riley equation needs to be augmented to cater for heterogeneity and unsteadiness in the indoor environment in order to be applied for evaluation of risk of airborne infection for different air delivery systems and various supplied flow rates

1.2 Research Objectives

The aim of this research is to evaluate ability of different ventilation systems to provide protection to the susceptible occupant form the airborne infectious disease transmission due to cough released infectious agents The objectives are as follows:

1 Develop a methodology to evaluate the efficiency of reduction in concentration of cough released droplets in the breathing zone of the occupant and apply this methodology to evaluate protective ability of four ventilation systems at various air supply flow rates

2 Evaluate the validity of perfectly mixed assumption in the breathing zone commonly used in estimating the probability of getting infected for different ventilation systems at different air flow rates

3 Augment the Wells-Riley equation for unsteady heterogenic environment and demonstrating its applicability to different ventilation systems and infectious loads

4 Examine how the protective performance of different ventilation systems vary with the height of cough release (sitting and standing position of the infector) and distance between infected and susceptible occupant

5 Examine flow field characteristics of the interactive two-phase flows between cough released droplets and room air flow generated with different ventilation systems

1.3 Scope of Work

This study is in the field of ventilation but results from this study have application on medical problem of control of airborne infection disease transmission The scope of work and the structure of discussion in each chapter are described briefly as follows:

 Literature review This study is not independent of, but based on, previous research on

motion of expiratory droplets in indoor environment Chapter 2 provides useful information on experimental design used to study the dispersion of expiratory droplets and methods adopted to evaluate risk of airborne infectious disease transmission

 Research methodology The work comprised a series of three related studies: (i)

development and application of methodology for evaluation of overall protective ability

of different ventilation systems and reduction of assumption necessary to evaluate probability of getting infected using Wells-Riley approach; (ii) evaluation of distance between infector and susceptible on the protection from cough released infectious droplets achieved with different ventilation systems; (iii) investigation of flow field characteristics of the resulting flow field between cough released droplets and room air flow generated influencing reduction droplet concentration in the breathing zone The research methodology is described for the series of experiments, which includes experimental design (facility and instrument) and data analysis (data and statistical analysis) The discussion of research methodology is presented in Chapter 3

 Chapter 4 describes the overall influence of ventilation system on control of

airborne infectious disease transmission A evaluation methodology is proposed for the

calculation of overall averaged probability of getting infected for a susceptible occupant The method is adopted to two different risk assessment models for the indoor environment supplied with a particular ventilation system These results were than

compared to calculations conducted with the perfectly-mixed assumption, and Basic Reproductive number variations are shown for different levels of occupancy

 Chapter 5 describes the influence of distance between infector and susceptible on

the protection from cough released infectious droplets achieved with ventilation

The influence of protective performance of different ventilation systems on height (sitting

and standing position of the infector) and distance between infected and susceptible occupant was studied

 Chapter 6 (and Appendix 1 and 2) documents the potential exposure and flow field

characteristics generated by direct cough at different infector - susceptible distances

in the indoor environment supplied with various ventilation systems.The interaction between cough and the flow fields generated by different ventilation systems and their consequent impact on exposure from the direct cough at several infector – susceptible distances were studied

 Conclusion and Recommendation The objectives are reviewed and a summary of

significant findings is presented These include: (i) the contributions of the new methodology for evaluation of ability of ventilation system to control airborne infection disease spread; (ii) the validity of the proposed augmentation of Wells-Riley equation; (iii) evaluation of protective performance of different ventilation systems (overall and infector-susceptible distance dependant); and (iv) flow field characteristics responsible for protection of susceptible occupant Lastly, some suggestions for further research and the development of ventilation system for airborne infection transmission control are

given in Chapter 7

Chapter 2: Literature Review

Definition adopted by CDC (2003): ―Droplet transmission is, technically, a form of contact transmission, and some infectious agents transmitted by the droplet route also may be transmitted

by the direct and indirect contact routes However, in contrast to contact transmission, respiratory droplets carrying infectious pathogens transmit infection when they travel directly from the respiratory tract of the infectious individual to susceptible mucosal surfaces of the recipient, generally over short distances The maximum distance for droplet transmission is currently unresolved, but historically, the area of defined risk has been a distance of <3 feet around the patient and is based on epidemiologic and simulated studies of selected infections (Feigin, 1982; Dick, 1987).‖

Definition adopted by CDC (2003): ―Airborne transmission occurs by dissemination of either airborne droplet nuclei or small particles in the respirable size range containing infectious agents that remain infective over time and distance Microorganisms carried in this manner may be dispersed over long distances by air currents and may be inhaled by susceptible individuals who have not had face-to-face contact with (or been in the same room with) the infectious individual (Coronado et al., 1993; Bloch et al., 1985; LeClair et al., 1980; Riley, 1959).‖

Li et al (2007) in a review co-authored by large group of engineers, microbiologists and epidemiologists on the role of ventilation on airborne transmission of infectious agents adopted this definition: ―airborne transmission refers to the passage of microorganisms from a source to a person through aerosols, resulting in infection of the person with or without consequent disease Aerosols are a suspension of solid or liquid particles in a gas, with particle size from 0.001 to over 100μm (Hinds, 1982) Infectious aerosols contain pathogens A droplet nucleus is the airborne residue of a potentially infectious (micro-organism bearing) aerosol from which most of the liquid has evaporated (Wells, 1934).‖

Droplet size is another variable which has been discussed Droplets traditionally have been defined as being >5μm in size Particles arising from desiccation of suspended droplets, know as droplet nuclei, have been associated with airborne transmission and defined as <5μm in size (Duguid, 1946) ―Large droplets‖ were first defined as droplets larger than 100μm in diameter by Wells (1934) Wells found that under normal air conditions, droplets smaller than 100μm in diameter (―small droplets‖) would totally dry out before falling to the ground 2 m away Xie et al (2007) conducted a study revisiting droplet size originally referred by Wells and proposed the following size definitions: ―large-droplet‖ diameter >60μm, ―small droplet‖ diameter <60μm and

‗droplet nuclei‘ diameter <10μm

2.1.2 Infectious agents transmissible by aerosols

Droplet nuclei can remain suspended in air for long periods and can travel considerable distances,

so close contact is not always necessary for transmission to occur Examples of infections transmitted in this manner include pulmonary TB, measles and chicken pox (Beggs, 2003) Respiratory viruses such as influenza and respiratory syncytial virus are mainly spread by droplet nuclei and droplet transmission (Ayliffe et al., 1999) Although respiratory viruses are transmitted through the air, other non-respiratory viral infections, such as chicken pox and measles (Riley et

al., 1978) are also spread by the airborne route The airborne route also contributes to the spread

of viral gastro-enteritis (Gellert and Glass, 1994)

There are only three respiratory diseases, i.e measles, chicken pox (varicella) and tuberculosis that are widely recognized as being primarily spread by airborne transmission, but many others, such as those due to the influenza virus and the respiratory syncytial virus, are also probably spread via airborne transmission (Yassi and Bryce, 2004) Tang et al (2006) in review on factors affecting aerosol transmission of infection points out that pathogens, such as parvovirus B19, enteroviruses and the organisms of atypical pneumonias, Chlamydophila pneumoniae, Coxiella burnetti and Legionella pneumophila, have the potential to be transmitted via aerosols as their life cycle involves replication at some point in the respiratory tract

One set of infection control guidelines for healthcare settings suggested that only tuberculosis

(Mycobacterium Tuberculosis,TB), measles (rubeola virus) and chickenpox (varicella zoster

virus, VZV) should be considered as ‗true‘ airborne infectious diseases (CDC, 2003) It is likely

that other infectious agents may also behave as ‗airborne‘, given a favorable environment, e.g

whooping cough (Bordetella pertussis), influenza virus, adenovirus, rhinovirus, Mycoplasma

pneumoniae, SARS coronavirus (SARS-CoV), group A streptococcus and Neisseria meningitides (Tang et al., 2006)

In multidisciplinary review of the 40 studies by Li et al (2007), the conclusion was that within the contemporary limitations of the conclusive studies chosen, there is strong and sufficient evidence to demonstrate the association between ventilation and the control of airflow directions

in buildings and the transmission and spread of infectious diseases such as measles, TB, chickenpox, anthrax, influenza, smallpox, and SARS

Guideline for Isolation Precautions: Preventing Transmission of Infectious Agents in Healthcare Settings 2007: ―Infectious agents to which this applies include Mycobacterium tuberculosis (Riley et al., 1959; Beck-Sague et al., 1992; CDC, 1994; Haley et al., 1989), rubeola virus (measles) (Bloch et al., 1985) and varicella-zoster virus (chickenpox) (LeClair et al., 1980) In addition, published data suggests the possibility that variola virus (smallpox) may be transmitted over long distances through the air under unusual circumstances and airborne infection isolation rooms (AIIRs) are recommended for this agent as well; however, droplet and contact routes are the more frequent routes of transmission for smallpox (Wehrle et al., 1970; Gelfand and Posch, 1971; Fenner et al.,1988).‖

2.1.3 Bioaerosol infectivity and virulence

The infectious disease process in an animal host is a function of microorganism concentration (infective dose) and virulence (disease promoting factors) that enable an agent to overcome the normal physical and immunologic defenses of the host (Cole and Cook, 1998) The infectious dose of a pathogen is the number of organisms required to cause infection in susceptible host

Data from research performed on biological warfare agents suggest that both bacteria and viruses can produce disease with as few as 1-100 organisms (e.g brucellosis 10-100, Q fever 1-10, tularaemia 10-50, smallpox 10-100, viral haemorrhagic fevers 1-10 organisms) (Franz et al., 1997) M tuberculosis may need only a single organism to cause disease, and as many as 3000 organisms can be produced by a cough or talking for 5 min, with sneezing producing many more (Fitzgerald and Hass, 2005) For many common agents, the infectious dose almost certainly varies between individual pathogens and their hosts, e.g immuno-compromised hosts may not only be more susceptible to infection with a lower infectious dose, but may also be a more infectious source, as the pathogen is poorly controlled by the defective immune system This may allow higher pathogen loads to be disseminated into the surrounding environment in some cases,

possibly leading to super-spreading events, such as described in some SARS outbreaks (Wong et al., 2004; Yu et al., 2005; Li et al., 2005; Yu et al., 2004; Li et al., 2005)

Measles is a highly contagious viral disease that is spread by the airborne route The infective dose is small, and as few as 4 doses per minute from an infected person can initiate an epidemic (Riley, 1980) Additionally, rubella (German measles) and varicella (chickenpox) viruses can be readily spread by aerosols in indoor air During major epidemics, influenza hospitalizations for persons at high risk may increase two to five times (CDC, 1992) placing health care workers at increased risk for infection Small infective doses are thought to be responsible because of the rapidity with which the disease spreads throughout a population The natural airborne transmission of respiratory infection with the coxsackie A virus type 21 was investigated and transmission of infection was demonstrated (Couch, 1981)

Inherent in the infection process initiated by the inhalation of infectious droplet nuclei is the area

of deposition within the respiratory tract Such deposition is influenced by hygroscopicity, which causes an increase in the size of inhaled aerosols through moisture take up as they move within the airways Knight (1993) estimates that a 1.5 μm hygroscopic particle a common size in coughs and sneezes increases to 2.0μm in diameter when passing through the nose and to 4.0μm in the saturated air of the nasopharynx and the lung He further theorizes that the effect of hygroscopicity and the resultant particle size change increase retention in the tertiary bronchioles and alveolar ducts, an effect that may be significant for viral aerosols, which are highly infectious for that part of the lung

2.1.4 Bioaerosol viability

When pathogenic microorganisms leave their host and are aerosolized, they are potentially injured during the generation process Additionally, once airborne they are outside of their natural

habitat and, depending on a variety of environmental factors, are increasingly subject to loss of viability with time Viability can be defined as the capability of a microorganism to reproduce Even if a microorganism remains alive, if it cannot reproduce it can be considered nonviable because it has lost the ability to reestablish a population within a defined microenvironment Factors influencing the survival of bioaerosols include their suspending medium, temperature, relative humidity, oxygen sensitivity, and exposure to UV or electromagnetic radiation

M tuberculosis is a hardy organism with a thick cell wall, and can survive for long periods in the environment (Fitzgerald and Hass, 2005) Data on human corona-virus (hCV) 229E from Ijaz et

al (1985) showed that, when airborne, this virus had a survival half-life of about 3 h at an RH of 80%, 67 h at an RH of 50% and 27 h at an RH of 30%, at 20oC, suggesting that high RH above 80% is most detrimental to survival of this coronavirus More recently, it has been shown that SARS-CoV can remain infectious in respiratory specimens for more than seven days at room temperature (Lai et al., 2005) Similarities with other viruses of nosocomial importance, i.e other RNA, lipid enveloped, respiratory viruses such as influenza, suggest that such organisms can survive for long enough in aerosols to cause disease, especially when associated with biological fluids such as mucus, faeces and blood

The studies by Meselson et al (1994) and Yu et al (2004) are striking, as they showed how the virus-laden aerosols could spread a few kilometers in a city or between buildings that are 60 m apart, due to wind flows

2.1.5 Bioaerosol size

Infectious bioaerosol particles may exist as (1) single bacterial cells or spores, fungal spores, or viruses; (2) aggregates of several cells, spores, or viruses; or (3) biologic material carried by other, nonbiologic particles (Navalainen et al., 1993) Microorganisms span wide size ranges In

general, infectious microorganisms range from 0.3 to 10μm for bacterial cells and spores, 2.0 to 5.0μm for fungal spores, and 0.02 to 0.30μm for viruses Specific pathogen sizes include 0.3 to 0.6 x 1 to 4μm for M tuberculosis; 0.3 to 0.90 x 2.0 to 20μm for L pneumophila, (Brenner et al., 1984) and 0.09 to 0.12μm for influenza virus (Murhpy and Kingsburg, 1991)

2.2 Dispersion of the contaminants released inside the indoor environment supplied with

total volume systems

2.2.1 Dispersion of exhaled contaminants in indoor environment supplied with total volume

systems

In a study of dispersal of exhaled air in displacement ventilated rooms (Bjørn and Nielsen, 2002) numerical simulation show that the simulated personal exposure is very sensitive to variations in the heat output of both the exposed person and the exhaling person, and to the cross-sectional exhalation area and pulmonary ventilation of the exhaling person The exhalation does not necessarily follow the boundary layer flow close to the body, but is able to ‗‗break free‘‘ and penetrate the breathing zone of other persons if persons are at distance up to 0.4 m Computational Fluid Dynamics (CFD) simulation by Gao and Niu (2007) findings performed for transient flow and experiment for steady state is in qualitative agreement these Air exhaled horizontally through the mouth results in much larger exposure than air exhaled through the nose Experiments show that air exhaled through the mouth can be ‗‗locked‘‘ in a thermally stratified layer, where concentrations can be several times higher than the return concentration If the vertical temperature gradient is larger than approximately 0.4-0.50C/m, this layer can settle in breathing zone height, but the boundary layer flow around the heated body can still offers some protection in this situation It should be noted that this phenomenon was only observed in cases where exhalation was conducted through the mouth With exhalation through the nose, the exhaled air was observed to follow the boundary layer flow and the thermal plume to the upper

part of the room, resulting in the typical two-zone concentration distribution Results gained through CFD simulation by Gao and Niu (2005) are in agreement with these experimental results These results indicate that stratification of the exhalation flow is not an acute problem in most normal situations

The potentially infectious droplets exhaled from the patients with pulmonary tract infection can cause cross infection especially in hospitals Performance of wall mounted and ceiling mounted downward mixing ventilation and displacement ventilation system in hospital ward was examined

by Qian et al (2006) using a personal exposure index (defined as the ratio between pollutant concentration in the breathing zone for the reference ventilation system (usually mixing ventilation) and ventilation system under investigation under steady state conditions) Qian et al (2006) established that Mixing Ventilation and downward ventilation systems have personal exposure index of 1 while Displacement Ventilation (DV) depending on the relative positions between the source and target show variations of personal exposure index When DV is used, it is possible to obtain a very high ventilation index for some source-target position, but it cannot be achieved in all situations (Nielsen 2009) Qian et al (2008) conducted the study of exhaled pollutant dispersion in hospital ward with downward ventilation and found that due to the interaction of the upward thermal plume and the downward supply airstream, which caused strong mixing, the supply air was incapable of pushing down the source patient‘s exhaled contaminants, and produced a downward unidirectional flow It was pointed out that results in the study by Qian et al (2008) are only qualitative and they only represent the trends since droplets are simulated by tracer gas, physics of large droplets was fully neglected as well as deposition, evaporation and coagulation of small and large droplets

2.2.2 Dispersion of particles released inside the indoor environment supplied with total

volume systems

Gao and Niu (2007) conducted a CFD study of particle dispersion inside the room with Mixing Ventilation (MV), DV and Underfloor system (UF) They concluded that particle concentration is almost uniform except at the points close to the ceiling where there is clean supply airflow The larger the particles are, the lower the mean indoor concentrations For particles smaller than 5μm, human exposure is nearly equal to the outlet concentration However for larger particles, human exposure is much lesser than the exhaust concentration For DV human exposure is greatly lower than that with MV for 0.1, 1.0, 2.5, and 5μm particles, but larger for 10 and 20μm particles With

UF the vertical stratification for 0.1, 1.0, and 2.5μm particle also appears and human exposure is lower than that with MV, but slightly higher than that with DV But owing to the weak carry-up effect, particles larger than 5μm are unable to be entrained to the upper level (Gao and Niu, 2007) With particle sources located within an internal heat source, desirable vertical concentration stratification appears in DV and UF for particles up to 10μm The advantageous principle of DV and UF that there is a less polluted occupied zone for non-passive gaseous pollutants is also applicable to particles whose diameters are less than a certain value which depends on the strength of the buoyancy force (Gao and Niu, 2007)

2.3 Personalized ventilation

2.3.1 General overview

Personalized ventilation (PV) is designed to provide clean and cool air close to each occupant (Melikov, 2004) Various ATD designs have been developed and studied previously Two small nozzles placed at the back corners of a desk generating two symmetrical jets or two linear diffusers placed at the front desk edge generating two jets, one toward the occupant‘s body and the second vertically, directed slightly away from the occupant (Sodec & Craig, 1990; Arens et al., 1991; Bauman et al., 1993; Faulkner et al., 1993, 1999, 2002; Tsuzuki et al., 1999; Cho et al., 2001; Melikov et al., 2002, 2003; Cermak & Melikov, 2003, 2004) ATD with a rectangular or circular opening mounted on a movable arm-duct which allows for changes of the distance

between the ATD and the person as well as the direction of the personalized flow (Melikov et al., 2002; Bolashikov et al., 2003), a flat ATD mounted on the top of a PC monitor allowing for change of personalized flow direction in a vertical plane (Melikov et al., 2002), a small nozzle integrated with the flexible support of a commercially available headphone supplying air very close to the mouth and the nose (Bolashikov et al., 2003), or combinations of some of these ATDs (Kaczmarczyk et al., 2004) have been studied Several other designs, such as a round nozzle attached to the chest blowing air against the face (Zuo et al., 2002), a displacement ATD placed below the desk (Loomans, 1999; Izuhara et al., 2002), a ventilation tower system (Hiwatashi et al., 2000), a partition integrated fan-coil unit (McCarthy et al., 1993; Jeong & Kim, 1999; Chiang

et al., 2002; Levy, 2002), Desktop Personalized Ventilation Air terminal Devices (Pantelic et al., 2009; Tham and Pantelic 2010) and textile terminals (Nielsen et al., 2007) have all been tested

Energy saving potential of PV has been studied by Sekhar et al (2003) The results indicate that the use of a secondary PV system in conjunction with a primary air-conditioning system not only enhances thermal comfort and IAQ acceptability but can reduce energy consumption by 15-30%

2.3.2 Dispersion of contaminants in the indoor environment supplied with personalized

ventilation

Non-uniformity in velocity and temperature field and differences in pollution generated in spaces will depend on the airflow interactions as well as the location of pollution sources This results in considerable variation in occupants‘ exposures (Melikov, 2004) Physical measurements have shown that a significant decrease of contaminant concentration in inhaled air with PV in comparison with Total Volume Systems (TVS) has been achieved but only at high supply flow rates of outdoor air (Faulkner et al., 1993, 1999, 2002; Melikov et al., 2002, 2003; Zuo et al., 2002; Cermak & Melikov, 2003, 2004; Cermak et al., 2004; Bolashikov et al., 2003) The airflow interaction, i.e., whether the personalized flow is transverse to, assisting or opposing the transient

flow of exhalation, the buoyancy driven boundary flow and the ventilation flow are of major importance for minimizing mixing in spaces With careful consideration of these factors affecting airflow patterns, pathogen laden droplets can be removed faster from the indoor air either exhausting them faster or increasing the deposition rate Available knowledge suggests that in rooms with mixing ventilation the use of PV will always protect the occupants from airborne transmission of infectious agents and will be superior to mixing ventilation alone (Melikov et al., 2003) In rooms with displacement ventilation, however, PV promotes mixing of the exhaled air with room air (Melikov et al., 2003; Cermak et al., 2004) A similar effect may occur in rooms with underfloor air distribution (UF) (Cermak & Melikov, 2003, 2004; Cermak et al., 2004) Cermak and Melikov (2007)reported that PV in conjunction with underfloor ventilation is more effective in protecting occupants from airborne pathogens released by exhalation than when an underfloor ventilation strategy is applied alone Owing to the non-uniform environment created

by PV, undesirable transport of exhaled pathogens can occur when an infected individual uses PV while the other occupants in the space do not use PV for protection However, these studies examined particular air terminal device (ATD) designs which create a particular flow field Nielsen et al (2007) reported that the protective ability of Personalized Ventilation (textile terminals) used in conjunction with vertical ventilation from ceiling-mounted terminals achieved

an increased efficiency of personal protection by factors of 5 up to 35

2.3.3 Dispersion of particles in room with personalized ventilation

In the study conducted by Faulkner et al (1993) room particle transport and the efficiency of removing tobacco smoke particles in the room with ceiling supply mixing ventilation and Personal Environmental Module was investigated The study concluded that only the high supply rate of outdoor air (40 l/s) directed to the breathing zone provided enhanced ventilation When 10 l/s were used particle removal efficiencies were essentially equivalent to those of a conventional ventilation system with thoroughly mixed indoor air

Distribution of particles with size of 0.5 – 10μm in a room ventilated by personalized ventilation system and ceiling supply system was investigated by the means of CFD by Zhao and Guan (2007) Particle source was assumed to be one person sitting indoor, with a generating rate of

5000 particles/s Zhao and Guan (2007) concluded that for the particles smaller than 2μm, a strong-enough personalized ventilation which can disperse the thermal plume is an effective ventilation mode to remove particles The personalized ventilation with less air supply volume has no obvious advantage compared to ceiling supply For the particles bigger than 7.5μm, Personalized Ventilation may not be the best ventilation mode It may have bigger particle concentration in breathing zone than that of ceiling supply ventilation

2.4 Expiratory droplet dispersion in indoor environment

2.4.1 Size distribution of expiratory droplets

Expiratory human activities such as: breathing, coughing, sneezing, or laughing result in droplet generation by wind shear forces The significance of each of these activities in the spread of infection depends on a number of factors, including: (1) the number of droplets it produces, (2) their size, (3) content of infectious agents, and (4) the frequency of its performance The content

of infectious agent expelled by an infected person depends, among other factors, on the location within the respiratory tract from where the droplets originate (Morawska, 2006)

Six studies that have reported size distribution of droplets emitted during coughing and sneezing were reviewed They are Duguid (1946), Louden and Roberts (1967), Papineni and Rosenthal (1997), Chao et al (2009), Morawska et al (2009) and Xie et al (2009) These studies used different measurement apparatus to measure droplets of different size bins and may be clustered into two groups: the first group that targeted the full range of released droplet size bins and the

second group that was focused on smaller droplet size bins, released by expiratory activity Among the first group, studies by Duguid (1946) and Chao et al (2009) reported very similar size distribution profile and mean droplet size while the results by Xie at al (2009) show a distribution shift into the larger droplet size range while maintaining similar shape of the distribution profile A full range of droplet size distribution is more realistic and this is adopted for this study

2.4.2 Dispersion of expiratory droplets generated by coughing

In the scenario studied using CFD by Gao and Niu (2006) where two occupants were facing each other at a distance of 1.2 while one occupant sneezed 1 s was enough for it to reach the breathing region of the exposed person Gao and Niu (2007) concluded that horizontal airflow (sneezing) is able to penetrate the protection of the boundary layer flow enclosing the exposed person and neutralize the positive effect of DV created by drawing fresh air from the lower level

Four CFD simulations were performed by Zhu et al (2006), examining dispersion of droplets particles with diameters of 30, 50, 100, 200, 300 and 500μm in the room with mixing ventilation (air conditioning device was placed on the wall close to the ceiling) In the case when two persons were sitting 1.54 m apart, simulation results for the droplets of 30μm suggest that after been coughed out the motion of these droplets were affected by the room airflow 30μm droplets tend

to remain airborne for prolonged periods and final point of deposition onto the surfaces inside the room cannot be predicted Some droplets of 50μm in diameter impacted on the face and neck of the person opposite, while the remainder passed by and then dropped to the floor For 100μm droplets simulation results suggested that more than 80% impacted on the lower neck area of the person opposite, while the remainder dropped to the floor soon after passing their body Saliva droplets of 200μm in diameter started to fall before reaching the person opposite, and nearly 100% of these droplets impacted on the chest of the person opposite However, the saliva droplets

of 300μm or more in diameter did not descend significantly, and nearly 100% of them impacted

on the neck of the opposite person Most saliva droplets of 500μm in diameter impacted on the face of the opposite person with little or no observable gravitational effect

Sun and Ji (2007) performed CFD simulation of the manikin coughing Four different cases were examined with mixing and displacement ventilation systems The CFD results show that droplets larger than 100μm generated by the sitting manikin body fall to the ground or onto the bed if the coughing burst are in the horizontal direction From a horizontal cough, expelled droplets of 80μm fall initially but dry up to nuclei while still some distance from the ground Therefore, droplets of 80μm can stay airborne and become the contaminant source from the lower part of the ventilated room Results from Sun and Ji (2007) also suggest that, droplets larger than 300μm fall

to the bed if the cough is directed upwards from a lying manikin From an upward cough from a manikin lying down, expelled droplets <100μm dry up to nuclei during their passage upwards and are then dispersed around the room by the air flow For droplets smaller than 50μm, because of the fast evaporation time, their transport and dispersion show similar characteristics to those of small passive particles Sun and Ji, (2007) showed that the ventilation set-up has a great influence

on removing the cough-expelled droplets of different sizes Mixing ventilation shows an almost equal ability to remove 1μm droplets and the nuclei of droplets with original sizes of 100 and 80μm, respectively However, displacement ventilation appears, in contrast, to remove 1μm droplets and the nuclei from large droplets of 100μm and 80μm, differently Small passive droplets show two-zonal distribution in a room with displacement ventilation, and the ventilation air removes them with very high efficiency The nuclei of the large droplets, however, are subject

to gravitational forces and show a tendency to settle down in the displacement ventilation the air flow of which has difficulty in carrying the nuclei upwards and removing them

Zhu et al (2006) also studied motion of cough droplets from a lying person while ventilation inlet was positioned on the wall near the ceiling above the head of the lying person CFD Simulation suggests that 30μm droplets were dispersed through out the room and deposit randomly to the surfaces The dispersion of droplets narrowed down with the increase in the droplet size from 50μm onward, and 300μm droplet dispersion was limited to the pillow area 500μm droplets impacted directly to the ceiling

Sun et al (2007) performed CFD study of droplet dispersion (50, 80 and 100μm) from a source at the middle of test room with no initial velocity in the long room with air supply was positioned at the wall close to the ceiling and exhaust was placed on the opposite wall close the floor Results show that for droplets less than 50μm, the dispersion feature is dominant due to their very short evaporation time and small settling velocity, therefore evaporating droplets of these sizes distribute in a similar manner as the neutral aerosol particles For droplets as large as 100μm, the settling feature is dominant due to the longer evaporation time and considerable large dropping velocity, their distribution consequently behave like large depositing particles in a room scale For droplets between these two sizes, the distribution tends to be at the lower part of the room than that of small neutral aerosol particles Within this size range, a lower initial position of the droplets in the room results in a higher deposition rate to the floor

Zhao et al (2005) examined transport of 1μm droplets generated by respiratory system from a standing person in the room with mixing ventilation strategy using unsteady drift flux model Results from the study indicate that coughing may produce particles or droplets, which will transport over a long distance Coughing with outlet velocity of 20 m/s will cause the particles or droplets to transport over a distance longer than 3m, and larger outlet velocity will have even longer distance in shorter time At the distance of 3 m from the person‘s mouth concentration of 1μm droplets were 25%

Xie et al (2007) conducted a study of single droplet dispersion released by respiratory activity with exhalation modeled as a non isothermal air jet Droplets were released from 2m height which can be considered as standing human position Study confirmed that small droplets evaporate rapidly and large droplets fall to the ground quickly; however, the horizontally expelled large droplets can also penetrate a long distance At a low relative humidity, more droplets and droplet nuclei could suspend in air, increasing the probability of subsequent inhalation For respiratory exhalation flows, the critical size of large droplets was also between 60 and 100μm, depending on the exhalation air velocity and relative humidity of the ambient air Expelled large droplets were carried more than 6 m away by exhaled air at a velocity of 50 m/s (sneezing), more than 2 m away at a velocity of 10 m/s (coughing), and less than 1 m away at a velocity of 1 m/s (breathing)

In the study of respiratory droplets movement in indoor environment using CFD Lai and Cheng (2007), simulated two standing persons at 1m distance in mixing and displacement ventilation strategy In the CFD simulation 0.01 and 10μm droplet nuclei were used since it was adopted from previous studies (Nicas et al., 2005; Morawska 2006), evaporation times adopted was 0.5s and evaporation was simplified treating droplet nuclei as respiratory generated droplets 50m/s velocity was assigned to the expiratory droplets Under the well-mixed ventilation, the expiratory droplets takes approximately 0.5 s to reach the receptor‘s face, 2 s to reach the wall behind receptor, 5 s to the receptor‘s legs, 10 s to the wall behind the source The droplets are homogeneously mixed within 1 min, but after a longer time particles exhibit settling characteristics and size stratification (more 10μm particles are in the lower part of the room, but still in the inhalation zone) is observed in the room For the displacement ventilation droplet cloud takes approximately 7 s to reach the exhaust When the global airflow speed is lower, the

distinctive characteristics of coarse size start to appear 10μm droplets begin to settle at the lower region of the room under displacement ventilation

In the study of expiratory droplet dispersion in general hospital ward Wan et al (2007a) examined dispersion when droplets were introduced upwards (simulation patient lying on the back and coughing out droplets towards the ceiling) CFD was used with the experimental measurements to examine evaporating droplet movement Results suggest that two distinctive dispersion behaviors were observed according to the droplet initial size The small size group (initial size ≤45μm) exhibited airborne transmittable behavior The large size group (initial size 87.5 and 137.5μm) did not stay in the air long enough for airborne transmission due to heavy influence from gravitational settling Some droplet nuclei in the small size group remained airborne for more than 360 s while the large size group droplets suspended in air for less than 30s The small size group exhibited a two-stage lateral dispersion behavior with rapid dispersion in the initial dispersion stage and much slower dispersion in the later stable stage The large size group droplets or droplet nuclei had only initial dispersion stage behavior due to rapid settling Extraction by following the exhaust air stream was the major removal mechanism (over 50%) for the small size group and gravitational settling was the dominant removal mechanism (over 98%) for the large size group

Cough from a lying person was examined by Chao and Wan (2006) Study was performed in aerosol chamber with ventilation rate 138ACH, cough particles were injected at 0.8m from the floor towards the ceiling Conclusions from the study are that small aerosols (initial size >45μm) had settling time frame of <20s in downward flow but increased to about 32–80s in ceiling-return flow Due to heavy gravitational effect, the large aerosols (initial size >45μm) settled in <6s in both studied airflow patterns In the downward flow system, because of the short settling time frame, turbulence was the only significant driving mechanism, and lateral dispersion was limited

to short distances (<0.31 m from the center) In the ceiling-return system, where both bulk flow

transport and turbulent dispersion were effective, small aerosols were dispersed laterally throughout the entire width of the room (2.4 m from the center) Bulk flow transport was more significant than the turbulent dispersion, as measured by the overall dispersion coefficients The small aerosols had the overall dispersion coefficients of around 10–80 cm2/s in downward flow while the overall dispersion coefficients increased to about 200–600 cm2/s in ceiling return flow Although turbulent dispersion was considered minor to bulk flow transport, a considerable dispersion distance could be driven solely by turbulence with long settling time frame The small aerosols reached an actual dispersion distance of about 1 m in the z direction (vertical) of ceiling-return flow, which was under pure turbulent dispersion and with longer settling time Heavy gravitational and inertial effect limited the lateral dispersion of the large aerosols to less than a few centimeters, which could be considered negligible in building ventilation applications

Wan and Chao (2007b) performed study with the same conditions as mentioned in Chao and Wan (2006), except that unidirectional upward system and single-side-floor system were used Results from the study show that the unidirectional–upward system was more effective than the single-side–floor system in removing the smallest droplet nuclei formed from 1.5μm droplets by air extraction However, when the droplets and droplet nuclei size increased, the unidirectional–upward system became less effective while the single-side–floor system was shown to be more favorable in removing these large droplets and droplet nuclei The magnitudes of lateral overall dispersion coefficients for the small droplets nuclei initial size 45μm in the direction of lateral bulk airflow induced by the single-side–floor system were about an order higher than that in the unidirectional–upward system It indicates that bulk lateral airflow transport was a much stronger lateral dispersion mechanism than air turbulence

Dispersion of cough droplets in a hospital ward supplied with MV was studied by Chao et al (2008) and the conclusion was that the change of airflow supply rate had insignificant effect on

the transport and deposition of droplets with initial size dp > 87.5 μm while droplets with initial size dp < 45 μm had smaller lateral dispersion at lower flow rates Pantelic et al (2009) investigated the ability of Desktop Personalized Ventilation (DPV) to serve as a control measure against cough released infectious droplets and concluded that reduction of peak concentration and exposure time can be achieved with DPV compared to MV system, and that the ability of cough flow to blow off the protective DPV jet for the longest time occurs at a distance of 1.75 m, rendering DPV to be least effective at a distance of 1.75 m between the breathing zone of the susceptible and infected occupant

Removal efficiency of exhaled particles in the 6 bed infection isolation room with different arrangements of exhaust position was studied by Qian and Li (2010) using experiment in full scale airborne infection isolation room and CFD They reported that large differences of the concentration levels were observed with the variation of the location of exhaust points while supply diffuser positions were maintained constant They concluded that system with ceiling level exhaust remove gaseous pollutants and fine particles more efficiently than system with floor level exhaust or system with floor and ceiling level exhaust They also reported that floor level exhaust system was not much more efficient in removing large particles compared to ceiling level exhaust system They also observed that for large particles deposition played more important role than ventilation

Chen and Zhao (2010) performed CFD study of the dispersion of exhaled droplets in the ventilated rooms From the numerical results obtained authors concluded following: (i) evaporation can be neglected for dp≤ 100 μm; (ii) ACH generated by the ventilation system makes an important impact on dispersion of expiratory droplets with dp≤ 200 μm; (iii) ventilation patterns generated in the room make great impact of droplet dispersion; (iv) impact of ambient

temperature and relative humidity can be neglected; (v) initial velocity of the expiratory process (breath, cough, sneeze) had a great impact on exposure

Nielsen et al (2010) examined airborne risk of cross-infection in a two bed hospital wards with downward ventilation for three different layouts of return openings Experiments using racer gas was performed under steady state conditions Downwards ventilation generated displacement type of flow due to low velocity supply into the region of the room without heat source It was concluded in the study that risk of airborne cross-infection can be minimized using high air change rate and that position of exhaust openings played significant role in distribution of exhaled contaminants Among the three scenarios examined the lowest risk of airborne infection was observed when return openings were mounted on the wall behind the patients head (patient laying

in bed) at the location higher that recommended by centre for disease control (CDC)

2.5 Personal exposure and risk assessment

2.5.1 Wells-Riley approach

Wells – Riley equation (Riley et al., 1978) was used in several epidemiological studies (Fennelly

et al., 2004; Grammaitoni and Nucci, 1997; Barnhart et al., 1997; Catanzaro, 1982; Fennelly and Nardell, 1998; Nardell et al., 1991; Nicas, 2000; Riley and Nardell, 1989) to quantify the quantum generation rate of the infected person based on available epidemiological data Several assumptions were used by Riley et al (1978) to derive the Wells-Riley equation: infectious period of the infector, disease examined is exclusively airborne, infection confers lasting immunity, subclinical cases are neglected and biological decay of aerosolized microorganism was neglected, perfectly mixed environment, steady state concentration of airborne infection throughout the exposure period and neglecting other removal mechanisms except ventilation

Nazaroff et al (1998) and Fisk et al (2005) included the effects of deposition and control measures to the Wells – Riley equation Rudnick and Milton (2003) extended the Wells – Riley equation for unsteady state situations and used CO2 as a marker for exhaled breath to calculate the probability of getting infected based on inhaled fraction of air that has been exhaled by the infected occupant Wells-Riley equation modified by Rudnick and Milton (2003) was derived assuming: perfectly mixed environment, loss of viability, filtration, gravitational and diffusion deposition and other mechanism of removal are small compared to removal by ventilation, airborne infectious particles are droplet nuclei that remain suspended in air for a long time, probability of inhaling more than one quantum is negligible, each quantum inhaled by susceptible produces new case, CO2 in the room space only comes form occupants In the studies by Nazaroff et al (1998) and Fisk et al (2005) also used assumption of well mixed indoor environment, steady-state, infection is ―one-hit‖ phenomenon, all individuals equally susceptible

to the infection but deposition due to various mechanisms was taken into consideration

Model for evaluation of infection risk of secondary airborne infection by Nicas et al (2005) was proposed in such a way that expelled particles with da≤20μm have the ability to evaporate quickly and constitute the inhaled particles with da≤10μm Because the rate of particle removal from room air and the probability of particle deposition in the alveolar region depends on particle diameter, Nicas et al (2005) separately accounted for the particles in each of the four diameter bins The model proposed by Nicas (2005) is essentially the same as the traditional Wells- Riley ―one-hit‖

risk model for infection by M tuberculosis Asumptions used by Nicas et al (2005) are: well

mixed environment, steady-state, infection is ―one-hit‖ phenomenon, alveolar region is the target site of an infection, biological agents will be single or for a cluster that can be carried by a droplet

in the range of dmin to dmax,, particle diameters are uniformly distributed by count across the interval dmin to dmax

Noakes et al (2006) coupled Wells – Riley equation with classic epidemic models to include long-term dynamics of infection Noakes and Sleigh (2009) developed stochastic Wells-Riley model and coupled it with multi zonal ventilation model to account for proximity of the infector

to susceptible and to incorporate mixing in interconnected spaces with each of the zones which were considered as perfectly mixed

Qian et al (2009) integrated Wells-Riley model with CFD to investigate spatial variability of the airborne infection risk during nosocomial SARS outbreak in Hong Kong Computation was performed as a steady state with constant release of infectious agents

Well mixed assumption is satisfied if the indoor air is completely and rapidly mixed; however, mixing is not instantaneous and the rate varies with indoor air flow conditions (Drescher et al., 1995) The uniform mixing assumption (well mixed room, (WMR)) is particularly important when evaluating the effectiveness of general engineering controls, such as recirculation, filtration, ventilation, and upper-air ultraviolet irradiation; when a susceptible person is in close proximity

to an infectious person In the former case, the uniform mixing assumption can lead to either overestimating or underestimating efficacy (Miller-Leiden et al., 1996); in the latter circumstance, the uniform mixing assumption can lead to substantially underestimating exposure intensity (Nicas, 1996; Drescher et al., 1995) If well mixed assumption is applied to larger volume, like hospital ward, or entire building under epidemiological investigation well mixed assumption can lead to overestimation of risk (exposure) for the susceptible occupants (Noakes and Sleigh, 2009) Incomplete mixing is particularly important for the indoor environments in which Displacement (Brohus and Nielsen, 1996), Under-floor or Personalized Ventilation systems (Melikov, 2004; Pantelic et al., 2009) are employed

In general, steady-state conditions do not apply to indoor particle attributes (Nazaroff and Klepsis, 2004) Pathogen laden expiratory droplets released by breathing, talking, coughing and sneezing are episodic events; therefore steady state does not apply when they are considered as sources of indoor pollution

Cough droplets with an initial diameter larger than 80 μm tend to settle quickly onto the floor owing to gravity (Chao and Wan, 2006; Chao et al., 2008), and droplets smaller than 80 μm have the ability to evaporate to droplet nuclei size and have a very small response time which enables them to behave as fluid particles Neglecting particle depositions on room surfaces are not likely

to be valid According to a report that reviewed the results from eight studies (Thatcher et al., 2001), deposition rates can vary from 0.05 to 0.1, 0.1 to 1, and 0.5 to 3 h for 0.3, 1, and 3μm particles, respectively Thus, for particles greater than about 1μm, neglecting particle depositions

on room surfaces are not likely to be valid unless outdoor air supply rates are very high

The validity of the assumption regarding loss of infectiousness may depend on temperature and relative humidity Data suggests that loss of infectiousness in indoor air can be a very slow process with half-lives as long as 2–3 days (Harper, 1961; Ijaz et al., 1985)

Qian et al (2009) used CFD with Wells – Riley equation to predict concentration levels of pathogens in an hospital ward Chen et al (2009) coupled Wells – Riley approach with viral kinetics and the characteristics of the exhaled bioaerosols in the study of influenza transmission

2.5.2 Inhalation intake fraction

Although there is variation in the way these terms are defined across the many disciplines involved in environmental health, most accept that ―exposure‖ represents ―the contact between an agent and a target‖ In contrast, ―dose‖ is ―the amount of pollutant that is absorbed by a target‖ In

addition, the term ―dose‖ is not consistently defined among different health scientists Rather than recommending either of these terms, we propose ―intake‖ because it refers to ―the amount of pollutant that enters a target after crossing a contact boundary‖ (Zartarian et al., 1997)

A complementary approach connects emissions to intake more directly, using the intake fraction metric (Bennett et al., 2002) Intake fraction is defined as the attributable pollutant mass taken in

by an exposed population per unit mass emitted from a source It is important to stress that the intake fraction depends on several factors, including chemical properties of the contaminant, emission locations, environmental conditions (climate, meteorology, land use, etc.), exposure pathways, receptor locations and activities, and population characteristics (Bennett et al., 2002) Klepeis (2004) presented probability distribution functions of intake fractions for environmental tobacco smoke based on multi-zone simulations of residential exposures

Nazaroff (2008) proposed a concept of exposure evaluation in which the system begins with sources that emitted pollutants into air The pollutants are transported, dispersed, and transformed

to yield concentrations that vary in time and space People are exposed to pollutants when they encounter these concentration fields Intake represents the inhalation of contaminants owing to their presence in the breathed air For the inhalation pathway, the intake fraction is defined as the attributable mass of a pollutant inhaled per unit mass released

2.5.3 Imperfectly Mixed Indoor Environments

Incomplete mixing can be important in several circumstances, including these: (a) in buildings in which displacement (Brohus and Nielsen,1996) or personalized ventilation systems (Melikov, 2004) are employed, (b) in large buildings with multiple air-handling units, and (c) in assessing exposures of building occupants associated with their own activities, owing to close proximity to pollutant emissions (Nazaroff, 2008)

Experimental research shows that under certain conditions including convection flow (Baughman

et al., 1994) and horizontal plug flow (Yost et al., 1994) air pollutant concentration in the room is not uniform For the rooms that tend not to be perfectly mixed, equations that are derived with assumption of perfect mixing (WMR) are inappropriate descriptors of the contaminant concentration to which a room occupant may be exposed To account for imperfect mixing the

concept of mixing factor, m, is used to modify air supply flow rate (Soule, 1978; Jayjock, 1988)

Unfortunately, mixing factor expression is incorrectly applied to a well mixed model room (WMR) albeit with an adjusted ventilation rate, to describe an imperfectly mixed room Depending on how mixing factor is estimated equations for concentration decay can seriously underestimate exposure intensity in the room‘s zone occupancy (Nicas, 1996) Mage and Ott (1994) established that modeling indoor concentrations using mixing factor is not consistent with mass balance concept

The simplest compartmental model that accounts for imperfect mixing is a two zone model Variations of this model have been previously applied to the room ventilation studies (Sandberg, 1981)

Related parameters that can be determined in tracer gas ventilation studies include the room mean age of air, the local mean age of contaminant, the room mean age of contaminant and the mean residence time of contaminant (Sandberg and Sjoberg, 1983) Various measures of ventilation efficiency based on these parameters have been proposed (Skaaret, 1986), but none permits to

estimate exposure intensity per se (Nicas, 1996)

Model that characterizes exposure concentration more accurately than well mixed room (WMR) model takes into consideration ―source proximity effect‖ (SPE) SPE model is based on the

premises that concentrations in close proximity to a source are higher than those predicted by well mixed model Futraw et al (1996)

2.6 Knowledge Gap and Hypothesis

Melikov (2004) in the review article on personalized ventilation identifies necessity for PV related research of airborne transmission of infectious agents and dispersion of large-droplet aerosols exhaled by people as a result of airflow interaction in rooms with Personalized Ventilation Zhao and Guan (2007) using CFD in their study concluded that for the particles smaller than 2μm, a strong-enough personalized ventilation (64m3

/h) is an effective ventilation mode to remove particles but with less air supply has no obvious advantage compared to ceiling supply Cermak and Melikov (2007)reported that PV in conjunction with underfloor ventilation

is more effective in protecting occupants from airborne pathogens released by exhalation than when an underfloor ventilation strategy is applied alone

Although previous studies explored PV protection from the exhalation flow carrying infectious droplets, it is still unknown if PV has higher protective ability from cough which has sufficient momentum to neutralize effect of PV flow or to cause such flow interaction that can increase

exposure compared to TVS The first hypothesis in this study is ―Desktop Personalized Ventilation is more effective than total volume ventilation as a control measure against airborne transmission of infectious diseases attributable to cough released droplets.”

Well mixed assumption can lead to substantially underestimating exposure intensity when source and receptor are close to each other (Nicas, 1996; Drescher et al., 1995), but when well mixed assumption is applied to larger volume (hospital ward, or entire building) it can lead to overestimation of risk (exposure) for the susceptible occupants (Noakes and Sleigh, 2009)

Incomplete mixing is particularly important for the indoor environments in which Displacement (Brohus and Nielsen, 1996), Under-floor or Personalized Ventilation systems (Melikov, 2004; Pantelic et al., 2009) are employed Chao and Wan (2005) compared four ventilation systems (unidirectional downward, unidirectional upward, single side ceiling, single side floor) different levels of deviation from the perfectly mixed condition were found

Although previous studies address effectiveness of well mixed assumption, they do not specify characteristics of the source especially when source can cause substantial changes of flow patterns in the environment (cough droplets release) Previous studies also recognize differences among flow patterns generated with different ventilation system but they do not address how interaction between cough flow and these flow patterns will influence exposure The second

hypothesis in this study is ―Well mixed assumption underestimates exposure due to cough released droplets in an indoor environment where release occurred.”

Wan & Chao (2005) compared the effects of four different types of supply–exhaust positions on the dispersion of cough-generated droplets and concluded that downward piston-type ventilation flow performed the best in controlling the transmission of cough droplets by reducing the number and lateral dispersion of droplets These studies indicate that airflow patterns with respect to the position of the source of expiratory droplets (person talking, coughing, sneezing and vomiting) have the most important role in the dispersion of cough droplets in ventilated rooms

Previous studies reported differences between air flow patterns generate with TVS and PV, but they did not examine influence of this differences on the protective performance (ability to reduce exposure) of different ventilation systems with the variation of distance between infected and

susceptible occupant The third hypothesis in this study is ―The effect of source-receptor distance increase on exposure to cough released droplets is different for DPV and TVS.”

Nazaroff (1998) when developing framework for evaluating measures to control nosocomial tuberculosis transmission pointed out that is possible to incorporate necessary effects into engineering risk assessment models to account for incomplete mixing Wan and Chao (2007) concluded study: ―The highly non uniform spatial distributions make the distance between the source and susceptible patients an additional concern in terms of exposure analysis These findings indicate the need for further developments in risk-assessment models incorporating the effect of different ventilation systems on the distribution of expiratory droplets and droplet nuclei.‖

The fourth hypothesis in this study is ―Wells-Riley equation can be extended for unsteady heterogenic environment and applied to evaluate the ability of different ventilation systems

to reduce transmission of airborne infectious disease due to cough released infectious droplets.”

After comparing dispersion effects produced by four different air patterns due to variation of supply and exhaust positions substantial differences were observed by Wan & Chao (2005) Wan and Chao (2005) concluded that air flow patterns are the most important factor influencing dispersion of droplets No variation of supply flow rate was explored in that study, hence it is not known if increase of supply flow rate will always generates airflow patterns that will reduce

exposure for all occupants in the environment The fifth hypothesis in this study is ―Increase of the ventilation flow rate for the total volume systems can in some special cases lead to increase of infectious droplets concentration in the breathing zone of the occupant.”

2.7 Study Design

In order to address the knowledge gap three independent but interrelated studies are used The first study (Chapter 4) addresses the first (H1), second (H2) and fourth hypotheses (H4); the second study (Chapter 5) addresses the third hypothesis (H3) and the third study (Chapter 6) addresses the fifth hypothesis (H5)

The first study (S1) whose details are described in Chapter 4, is designed to characterize infection risk from cough released droplets as a single value (overall probability of infection) The size resolved exposure measurement was used with Dose-response and Wells-Riley risk assessment approaches to calculate probability of infection Application of these two approaches requires usage of biological data (dose-response method), epidemiological data and size distribution of cough released droplets (Wells-Riley method) to characterize infection risk This data were used from previously published literature After obtaining overall probabilities of infection, variation

in the number of occupants in the indoor environment was used to investigate Basic Reproductive Numbers generated with different ventilation systems at two different flow rates (6 ACH and 12 ACH) (Figure 2.1)

The second study (S2) described in Chapter 5 is designed to characterize variation of infection risk with spatial variability characterized by direction and distance between infector and susceptible S2 just like S1 uses the size resolved exposure measurements, but risk assessment is performed using only dose-response method (Figure 2.1) Overall probability of infection characterizes indoor environment with a single value but spatial variability of infection risk are present due to various distances between infector and susceptible occupant(s) Result for spatial variability of airborne infection risk can be used to calculate overall probability of infection Averaging process need to take into account different contributions of various distances using method describe in Chapter 3 (Section 3.6)

The third study (S3) described in Chapter 6 is designed to investigate flow fields in the breathing zone of the exposed occupant when cough droplets are released The flow field examined in S3 represents flow field resulting from the interaction among: cough flow, room flow generated with air delivery system at a particular supply air flow rate, convective boundary layer around heated human body and inhalation/exhalation flow Results from S3 can be used to explain variations in overall probability of infection for different ventilation systems investigated in S1 and variations

of infection risk with spatial variability of the distance between infector and susceptible investigated in S2 (Figure 2.1)

S1

Ventilation systems:

MV, DV, UF and DPV

Supply flow rates for TVS:

6 ACH and 12 ACH DPV: 5 l/s Hypotheses: H1, H2 and H4

Biological data

Risk assessment models

Basic

Reproductive

Number

Overall Probability of infection

Supply flow rates for TVS:

6 ACH and 12 ACH DPV: 5 l/s i-s distances: 1m, 2m, 3m and 4m Hypothesis: H3

P=P(i-s distance)

S3

Ventilation systems:

MV, DV, UF and DPV

Supply flow rates for TVS:

6 ACH and 12 ACH DPV: 5 l/s i-s distances: 1m, 2m, 3m and 4 m Hypothesis: H5

Exposure profiles Velocity fields

Figure 2.1 Studies designed to address the knowledge gap