Review of Engineering Controls for Indoor Air Quality: A Systems Design Perspective

Abstract

This paper aims to review the engineering controls for indoor air quality (IAQ) from a systems design perspective. As a result of the review, we classify the literature content into three categories: (1) indoor air treatments, (2) dissemination control strategies, and (3) information technology. Indoor air treatments can be generally interpreted as the “cleaning” aspect, which covers ventilation and contaminant removal techniques. Dissemination control focuses on how contaminants generated in an indoor space can be transmitted, where four types of dissemination are classified. The category of information technology discusses IAQ sensors for monitoring, as well as the applications of the Internet of Things and IAQ data. Then, we further analyze the reviewed engineering controls by performing systems and functional analysis. Along with a discussion of IAQ functions, we suggest some systems design techniques, such as functional decoupling and design for flexibility/resilience, which are expected to promote more systems thinking in designing IAQ solutions.

1. Introduction

Indoor air quality (IAQ) is a traditional topic in the field of heating, ventilating, and air conditioning (HVAC). Since the COVID-19 pandemic, concerns about IAQ have risen to a high level (e.g., [1,2,3,4]). Emerging environmental issues, such as urban air pollution [5] and wildfires [6], also bring new challenges for IAQ. Along with other seasonal airborne diseases (e.g., influenza), IAQ remains an important driving force for building and HVAC design improvements.

IAQ is a complex topic since it involves subject domains across multiple disciplines. At a high level, there are three distinct yet related broad disciplines. The first one is environmental and natural sciences, which study the generation and dissemination of air pollutants in the indoor and outdoor environment. Top-level subjects include air pollution, environmental chemistry, and fluid mechanics. The second broad discipline is the medical sciences, which is the study of the effects of air pollutants on human health and the development of health-related policies. Top-level subjects include airborne infectious diseases, health impacts from exposure to air pollutants, and public health. The third broad discipline is architecture and building engineering, which considers the design elements of buildings to mitigate or improve IAQ issues.

By recognizing the disciplinary complexity of IAQ, this paper focuses on engineering controls, which are concerned with what engineers can design and develop for IAQ. As a background, we consider the “hierarchy of controls” from the Centers for Disease Control and Prevention and the National Institute for Occupational Safety and Health, for which a brief explanation can be found in [7]. At the top of this hierarchy of controls are elimination (e.g., banning indoor smoking) and substitution (e.g., changing gas stoves to electric stoves), which are considered the most effective forms of IAQ control. If elimination and substitution are not allowed, engineering controls are considered as the next level of controls. With this background in mind, engineering controls are only one approach among many other options for IAQ. They come with their own features and limitations, which we intend to highlight in this paper.

Though the scope of this paper is confined to engineering controls, the relevant content is still very extensive. For example, in response to the COVID-19 pandemic, researchers have identified and reviewed engineering controls and interventions with various focuses, as follows:

• Non-pharmaceutical measures, ventilation, and air purifiers [8];
• Control strategies, e.g., face masks, air distribution, filtration, and disinfection [9];
• Building operation guidelines/standards and energy use [10];
• Heating, ventilating, and air conditioning (HVAC) systems and ventilation systems [11,12];
• Natural ventilation and mechanical ventilation [13];
• The Internet of Things (IoT) and wireless sensor networks [14].

Notably, one category of engineering controls and interventions often delivers a specific function to maintain or improve IAQ. For example, ventilation can reduce the concentration of infectious bioaerosols but cannot effectively prevent close-contact transmission [15]. Thus, it is important to clarify the functions and limitations of individual engineering controls and explore how these controls can be integrated into a system-level solution. For example, some review papers promote the integration of ventilation with air cleaning or disinfection technologies and consider this a research trend [8,11,13,16]. Built upon these existing reviews, this paper intends to contribute to the IAQ research community in two ways. First, this paper offers an integrative framework of engineering controls for IAQ. This framework covers a wide range of IAQ controls, which are reviewed in Section 3, Section 4 and Section 5. Second, a systems design perspective is applied in Section 6 to analyze engineering controls in view of component-level versus system-level solutions, functional analysis, and systems design techniques (e.g., functional decoupling and design for flexibility/resilience). In our view, this systems design perspective has not been well explored for IAQ, and it is expected to support engineers and practitioners in exploring and developing a holistic or integrated approach to IAQ issues.

This review paper is organized as follows. Section 2 discusses our literature review procedure, which we used to search for papers and build an integrative framework for content organization. From Section 3, Section 4 and Section 5, we review the papers for three major topics, briefly outlined as follows: (1) air treatments, (2) dissemination controls, and (3) information technology. In Section 6, we apply the systems design perspective to analyze the reviewed engineering controls and suggest some systems design strategies for IAQ. Section 7 concludes this paper.

2. Literature Review Procedure

To prepare this review paper, we reference the guidance of [17], who classifies three types of reviews: systematic, semi-systemic, and integrative. As systematic reviews demand an exhaustive search of a well-defined field, they should not be applicable due to the vast volume of the relevant literature. In contrast, we choose an integrative review approach, which aims to develop an integrative or theoretical framework to summarize prior research work and suggest future research directions.

To develop the integrative framework, we perform functional decomposition by first reflecting the functions of two traditional IAQ controls: ventilation and pressure differential (or gradient) [11,13]. In general, the function of ventilation could be interpreted as supplying clean air and pressure gradients to control the flow of indoor air contaminants. These describe two distinct functions of IAQ controls, which can be further broken down into specific controls or interventions (e.g., what specific IAQ controls are intended to clean air).

Then, we initially used Google Scholar with various combinations of keywords (e.g., indoor air quality, ventilation, air distribution, airflow control, and COVID) to collect and study the papers. We collected references in English only, with a preference for journal papers and review papers published in recent years. We acknowledge that this approach could omit other non-English works on this issue, which is one limitation of this review. As a result, we propose the integrative framework, illustrated in Figure 1, which contains three major topics for organizing the review content.

The first topic is “Supply Air and Indoor Air Treatments”, whose duty is to provide clean air to indoor spaces by supplying outdoor air via ventilation or removing air contaminants using air-cleaning technologies. The second topic is “Dissemination Types and Control Strategies”. Even if the indoor spaces receive clean air, air contaminants generated within them can still disseminate and affect IAQ at various scales. In this topic, we will classify four dissemination situations and discuss their engineering controls. The third topic is “Sensors and Information Technology”, where the dashed lines in Figure 1 represent the information flow with two possible functions. The first function is monitoring, where sensors can collect the IAQ data from the indoor spaces. The second function is to utilize the IAQ data to facilitate air treatments and limit the dissemination of indoor air contaminants.

Using the integrative framework, we further refine and classify the sub-topics and identify their key review papers. The list of these key review papers is provided in Table 1. Then, we continued searching for and selecting papers that have cited these key review papers using Google Scholar. In this search process, we focused on identifying ideas for IAQ controls and interventions without aiming to exhaust all related studies and investigations.

In the literature, some papers have discussed the integration and systems design perspective of IAQ controls in various detail. For example, Shen et al. [18] evaluated the effectiveness of different IAQ controls at four scales (i.e., breathing zone, personal, room, and building). Bueno de Mesquita et al. [19] considered three scales (i.e., close-interactive, room, and building) to analyze engineering controls and exposure risks. Yan et al. [20] used simulation to compare the combined effects of different controls. In comparison, this review paper intends to provide another aspect by purposely separating air treatments (e.g., ventilation and air cleaning) from contaminant dissemination to review their individual functions. We expect this approach to offer insights into systems design (to be discussed in Section 6).

Table 1. Sub-topics and key review papers.

Sub-Topics	Key Review Papers
Ventilation rates	Persily [21]
Room air distribution	Yang et al. [22]
Demand-controlled and smart ventilation	Guyot et al. [23], Lu et al. [24]
Natural and hybrid ventilation	Chenari et al. [25], Sakiyama et al. [26], Zhang et al. [27]
Ventilation and health-related issues	Qian and Zheng [28], Shajahan et al. [29]
Air-cleaning technologies	Liu et al. [16], Kelly and Fussell [30]
Particle filtration	Fisk [31], Alavy and Siegel [32]
Removal of gaseous contaminants	Palliyarayil et al. [33], Weon et al. [34], Li et al. [35]
Disinfection of bioaerosols	Reed [36]
Dissemination in the same space	Ai and Melikov [37]
Sensors for IAQ monitoring	Chojer et al. [38], Saini et al. [39]
Wireless communications and the Internet of Things (IoT)	Saini et al. [40]

Table 1. Sub-topics and key review papers.

Sub-Topics	Key Review Papers
Ventilation rates	Persily [21]
Room air distribution	Yang et al. [22]
Demand-controlled and smart ventilation	Guyot et al. [23], Lu et al. [24]
Natural and hybrid ventilation	Chenari et al. [25], Sakiyama et al. [26], Zhang et al. [27]
Ventilation and health-related issues	Qian and Zheng [28], Shajahan et al. [29]
Air-cleaning technologies	Liu et al. [16], Kelly and Fussell [30]
Particle filtration	Fisk [31], Alavy and Siegel [32]
Removal of gaseous contaminants	Palliyarayil et al. [33], Weon et al. [34], Li et al. [35]
Disinfection of bioaerosols	Reed [36]
Dissemination in the same space	Ai and Melikov [37]
Sensors for IAQ monitoring	Chojer et al. [38], Saini et al. [39]
Wireless communications and the Internet of Things (IoT)	Saini et al. [40]

3. Supply Air and Indoor Air Treatments

This section will review the scientific and engineering approaches to treat supply and indoor air. At a high level, there can be two treatment strategies: ventilation and air cleaning. Ventilation can be viewed as an indirect cleaning technique that uses outdoor air to dilute the concentration of indoor contaminants. In contrast, air cleaning involves the direct cleaning of air, such as removing particulate matter (PM) from the air, neutralizing gaseous contaminants, and disinfecting airborne bioaerosols.

3.1. Ventilation

Ventilation is one early technique and remains the most popular practice as one engineering control for IAQ [21]. In this paper, we first cover the topics of ventilation rates and room air distribution, which are traditional in the design of ventilation systems. With the concern of energy and health, we then discuss the topics of smart ventilation, natural ventilation, and health-related studies involving ventilation.

3.1.1. Ventilation Rates

Ventilation rates can be viewed as one of the most important design parameters or criteria for engineers to decide on ventilation systems. The ANSI/ASHRAE Standard 62.1 [41] (Ventilation for Acceptable Indoor Air Quality) is an established and well-recognized standard that determines and regulates outdoor airflow rates in the breathing zones of different building spaces. For example, the ANSI/ASHRAE Standard 62.1-2019 specifies ventilation rates of office spaces at 2.5 L/s per person and 0.3 L/s per square meter of floor area. Persily [21] has reviewed the historical development of the ASHRAE Standard 62. One important point to note is that ventilation rates are input measures and do not directly assess IAQ performance or health-related outcomes. Persily [21] envisions future ventilation methods that can better consider multiple factors, such as contaminant sources and the presence of air-cleaning technologies. One example is the HealthVent project, funded by the European Committee, which incorporates health-related IAQ criteria from health authorities into ventilation guidelines [42,43].

In addition to outdoor airflow rates, ventilation rates have also been specified in terms of air change per hour (ACH). For example, Klein et al. [44] conducted a study on how varying ACH levels impact the concentrations of volatile airborne contaminants in laboratories, where fume hoods serve as a key IAQ control, and ACH is considered a more convenient measure in relation to exhaust rates. Importantly, ACH alone does not indicate the quantity of outdoor air supplied. For example, the ANSI/ASHRAE/ASHE Standard 170-2021 [45] (Ventilation of Health Care Facilities) intentionally defines “Outdoor ACH” and addresses air recirculation as part of their guidelines. Furthermore, in the building readiness document prepared by the ASHRAE Epidemic Task Force (at https://www.ashrae.org/file%20library/technical%20resources/covid-19/ashrae-building-readiness.pdf, accessed on 16 July 2022), the concept of “ACH of equivalent outdoor air” is used to combine the effects of using outdoor air and air-cleaning techniques. Risbeck et al. [46] employed the concept of equivalent outdoor air to analyze and compare the additional energy consumption costs associated with IAQ interventions against the reduced risk of airborne infection.

In view of engineering controls, the utilization of ventilation rates is relatively convenient as it directly specifies the quantities of outdoor air and exhaust rates for designing HVAC systems. However, since the energy consumption associated with ventilation can be high, especially for cold climate areas, it becomes a limiting factor preventing us from increasing ventilation rates indefinitely. Engineering solutions for energy-saving ventilation, which covers natural and hybrid ventilation techniques, will be discussed in Section 3.1.4.

3.1.2. Room Air Distribution

Room air distribution concerns how supply air is distributed and mixed with the air in a space. It is a traditional topic in both practice (e.g., 2021 ASHRAE Handbook–Fundamentals, Chapter 20; 2019 ASHRAE Handbook–HVAC Applications, Chapter 58) and research. For example, research has explored various analytical and experimental tools, such as computational fluid dynamics (CFD) [47,48,49,50], multizone airflow simulation [51], and thermal manikin [52,53,54]. Investigative studies of how ventilation design parameters impact IAQ are abundant, such as variable air volume (VAV) systems [55], volatile organic compound (VOC) removal performance [56], the development of design charts/databases to inform air distribution properties [57,58], underfloor air distribution [59,60,61,62,63,64,65,66], and personalized air distribution [67,68,69,70,71].

The topic of room air distribution has also been reviewed at different times. Cao et al. [72] classified six types of ventilation and compared them with different aspects, such as air exchange, occupant protection, heat removal, and energy saving. By examining the delivery approaches of supply air, Yang et al. [22] have classified and reviewed twelve (12) ventilation systems (e.g., wall attached and underfloor air distribution). Further review papers have examined ventilation systems in the context of using radiant heating/cooling [73], system adaptability for flexible space usage [74], and airborne infection control [9,13,28,75,76,77].

The issues surrounding room air distribution (or ventilation) approaches are complex and continue to be an active area of research. Their complexity can be explained in four aspects, along with recent research work, as follows.

(1)

Physical phenomena with multiple factors: The supply air distribution in a space involves multiple physical principles, such as turbulent airflow structure and buoyancy effects. Recent studies have included the interaction of airflow with indoor air recirculation devices (e.g., portable air purifiers) [78], unsteady airflow for air mixing [79], and the impact of air distribution on the performance of solar-based cooling systems [80];

(2)

Continual innovation of ventilation techniques: Recent studies continue to propose new ventilation ideas such as personalized ventilation for infection control [81,82,83,84,85], interactive cascade ventilation (the use of cooler supply air to suppress or interact with warmer supply air) [86,87], sidewall air supply to address the limitations of stratum and displacement ventilation [88], and attachment ventilation [89,90,91,92];

(3)

Diversity of “ventilation” functions and performance indices: The term “ventilation” can be confusing as it can imply different engineering functions (e.g., dilution/removal of contaminants and regulation of thermal comfort), thus requiring consideration of different factors and performance indices (e.g., [93]). Recent studies have focused on distinguishing contaminant removal effectiveness and air change effectiveness [94], consideration of outdoor PM air pollutants [95], and assessment of localized and overall infection risks [96];

(4)

Distinctiveness of occupancy requirements: As the IAQ requirements and concerns vary for different types of occupancy, new studies are motivated to examine ventilation in various building spaces such as hospitals [97,98,99,100,101,102], offices [103], oxygen supply for hypoxic areas [104,105,106], residential buildings [107], and poultry houses [108].

3.1.3. Demand-Controlled and Smart Ventilation

Instead of maintaining fixed ventilation rates for building spaces, smart ventilation considers real-time occupancy to adjust ventilation rates to save energy without compromising IAQ. Relevant review papers can be found for residential buildings [23], commercial buildings [24], and personalized ventilation [109]. One early application of smart ventilation is termed demand-controlled ventilation [110], which already outlines three typical control considerations:

(1)

Control parameters: Typical choices include carbon dioxide (CO₂) and relative humidity (RH) as common bio-effluents or surrogates of occupancy levels. As reviewed in [23,24], some indoor contaminants, such as VOCs, can be unrelated to occupancy levels and should not be overlooked if expected in specific spaces;

(2)

Control setpoints: Lu et al. [24] reviewed the changes in CO₂ concentrations as IAQ requirements in different standards (typically ranging from 500 to 2000 ppm), which can be used in the control design;

(3)

Control strategies: As discussed in [23], control actions can include modulating the exhaust rates only and a balanced approach (where supply air is also modulated). For example, Pavlovas [111] modulated the exhaust rate from 10 L/s to 30 L/s based on three CO₂ levels (800, 1000, and 1200 ppm) and three relative humidity levels (60%, 70%, and 80%).

Without aiming to exhaust all possible ideas, we discuss a few papers related to smart ventilation to illustrate the abundance of ideas on this topic. Beyond occupancy levels, Sherman and Walker [112] considered outdoor air quality and peak energy demand as additional factors for ventilation control. Young et al. [113] further considered temporary ventilation curtailment (without violating IAQ requirements) to reduce peak energy demand. Cheng et al. [114] modeled CO₂ concentration in the breathing zone (instead of the whole space) to determine the proper outdoor air ratio in stratum ventilation. Ng et al. [115] explored how real-time infiltration information could be estimated and used to save energy for residential ventilation. Walker et al. [116] compared smart ventilation performance with and without the zoning of a space using simulations (CONTAM and EnergyPlus). Song et al. [109] reviewed several air terminal devices that supply and remove air close to occupants.

Recent developments tend to advance control and information techniques. For example, Schibuola et al. [117] examined how parametric tuning in PID control can impact the performance of a demand-controlled ventilation system. Ganesh et al. [118,119] used model-based predictive control to determine ventilation rates with control parameters such as ozone, formaldehyde, and particulate matter (PM). Zhang et al. [120] applied the genetic algorithm to determine CO₂ concentration limits for constant air volume (CAV) ventilation to mitigate frequent on–off operations. Instead of rule-based and model-based control, Heo et al. [121] used the deep reinforcement learning algorithm to manage complex factors related to IAQ for smart ventilation. Notably, reinforcement learning (RL) and machine learning (ML) can be viewed as one trend of using advanced data analysis techniques for IAQ, and they have been reviewed in [109,122].

As reviewed in [23], the energy performance with smart ventilation can vary significantly due to various factors (e.g., climates, building types, and control techniques), making it difficult to draw conclusive statements. Case and empirical studies conducted after the review paper [23], and those not cited in it, can be found in residential homes or buildings in the United States [123], Poland [124], Denmark [125], and the Arctic [126], as well as in office buildings [127], schools [128,129], and underground parking garages [130].

3.1.4. Natural and Hybrid Ventilation

Natural ventilation is intended to dilute indoor air contaminants without using mechanical equipment (e.g., air-handling units and exhaust fans). Two principles drive air movement. One is a pressure gradient caused by wind. While the windward side of buildings receives positive pressure, the leeward side has negative pressure. Another is a temperature gradient, which causes buoyancy, with hot air moving upwards (e.g., stack effects). Due to the requirements of IAQ and energy saving, natural ventilation remains a popular topic in research and practice. In the next paragraph, we will outline some review papers.

Reviews by [25,131] have discussed different architectural elements (e.g., window-to-wall ratios and building orientation) and specialized designs (e.g., atrium/shaft, wind tower/catcher, and double-skin façade) for natural ventilation. Figure 2 shows simplified schematics of an atrium, wind tower, and double-skin façade to illustrate the airflow with these architectural elements. Subsequent reviews have focused on evaluation tools for natural ventilation, such as mapping evaluation tools to design stages [132], the use of tracer gas [133], and computational techniques (e.g., CFD, parameterization, and optimization) [26]. Other reviews have concentrated on specialized building features and natural ventilation, including balcony design, occupants’ perception [134], and underground buildings [135]. Recent review papers extend the scope of natural ventilation to address issues related to thermal comfort (and heatwave resilience) [136], the integration of natural ventilation technologies [27], and classrooms in response to COVID-19 concerns [137]. In this section, we intend to review key engineering ideas for natural ventilation, incorporating recent work that may not be covered by these review papers.

Single-sided ventilation refers to introducing outdoor air through openings (e.g., windows) from one side of the space. While it should be the most common approach, it tends to result in slow air exchange and shallow penetration [25,27]. Studies have been conducted to understand single-sided ventilation driven by wind [138,139], buoyancy [140], both [141,142], and periodic vortex shedding (or a pumping mechanism) [143,144,145]. The effects of window arrangements have also been studied [138,141,143,146].

Cross ventilation requires openings on opposite sides of the space. To maintain effective cross ventilation, the length (L) of the opposite sides should not be too long relative to the ceiling height (H), where the ratio of L to H should be less than 5 [25,27,147]. Additionally, numerous studies have examined the effectiveness of cross ventilation with design features such as opening shapes [148,149], opening locations [150,151], roof features [150,152], exterior blockages [153,154,155], internal openings [151], and internal vegetation [156]. Window openings can impact natural ventilation and have been studied with regard to algorithmic control [157] and occupants’ behavior [158,159,160,161].

Stack ventilation is often promoted by specialized building features such as atriums and shafts to facilitate vertical air movement. Chenari et al. [25] and Zhang et al. [27] reviewed building features that utilize solar energy (e.g., solar chimney and double-skin façade) and wind energy (e.g., wind tower and wind catcher) to drive air movement. Monghasemi and Vadiee [162] reviewed various systems that integrate solar chimneys for building applications. Jomehzadeh et al. [163] reviewed different windcatcher systems. Notably, natural ventilation is particularly sensitive to outdoor air quality [164], and Xia et al. [165] proposed the use of nanofiber window screens to filter outdoor air as an innovative idea.

As a growing trend, ventilation system design tends to combine different principles and systems. For example, Chenari et al. [25] reviewed works on hybrid ventilation systems that incorporate mechanical components (e.g., fans) to assist natural ventilation. Zhang et al. [27] reviewed natural/hybrid ventilation systems that integrate multiple technologies (e.g., solar chimney + wind tower). Studies of natural/hybrid ventilation in practice can be found in the context of school buildings [166,167,168,169], public hospital wards [170], a university building [171], and an office [172].

3.1.5. Ventilation and Health-Related Issues

The role of ventilation and airborne infection control has been discussed in the literature, and this section is intended to capture some key ideas on this topic. After the epidemic of severe acute respiratory syndrome (SARS) in 2003, Li et al. [173] reviewed ventilation and healthcare studies with a panel of multidisciplinary experts and emphasized the importance of ventilation in infection control. The review by [174] distinguished large droplets and aerosols to explain their transmission mechanisms and ventilation control. Since then, studies of ventilation in healthcare settings have continued (e.g., [175,176]), and three control factors are generally summarized: ventilation rates, flow directions, and airflow patterns [28,177]. Shajahan et al. [29] systematically reviewed studies of indoor environmental parameters (e.g., indoor air temperature, relative humidity, and ventilation) and their effects on health outcomes. In an effort to establish ventilation guidelines and regulations, Carrer et al. [43] highlighted the importance of exposures (which cannot be easily quantified and controlled), and Carrer et al. [42] developed a framework to guide the determination of ventilation rates.

In the context of COVID-19, Li et al. [178] and Lu et al. [179] demonstrated evidence of long-range transmission due to inadequate ventilation. Lipinski et al. [180] reviewed ventilation strategies in the context of viral transmission mechanisms. Further reviews examined the effects of air filtration and recirculation [181], air cleaning techniques (e.g., ultraviolet light and a nano-porous air filter) [182], and ventilation practices in classrooms [183,184]. Examples of relevant studies include the analysis of ventilation and infection risk using the Wells–Riley model [185,186,187], the combined effects of social distancing and ventilation [188], the effects of air change rates in public transportation [189], the placements of supply/exhaust fans in a hospital room [190], the use of ventilation and portable air purifiers in dental clinics [191,192,193], and the use of heat recovery to mitigate the increasing energy consumption for ventilation in classrooms [194].

3.2. Air-Cleaning Technologies

According to the reviews by [30,195,196], we can identify several common air-cleaning technologies, such as mechanical air filtration, photocatalytic oxidation, and ultraviolet germicidal irradiation. These technologies demonstrate various cleaning capacities for different types of indoor air contaminants. However, it is important to note that research findings are abundant, with numerous possibilities for air-cleaning technologies, which cannot be easily organized and clarified in this review paper. To maintain an overview, we reference the summary tables by Liu et al. [16] (pp. 380–381) and organize the content of this section with respect to three types of air contaminants and their key (or most effective) air-cleaning technologies as outlined below.

• Particulate matters: mechanical and electrostatic filtration;
• Gas contaminants: adsorption and photocatalytic oxidation;
• Bioaerosols: ultraviolet germicidal irradiation.

3.2.1. Particle Filtration

In a nutshell, particle filtration can be viewed as a process that uses fibrous media as a physical means to trap particles in the air stream. The details of this process are complex, involving multiple mechanisms such as diffusion, interception, and inertial impaction [16,197,198]. In practice, the particle removal efficiency of filters is often standardized using the MERV (minimum efficiency reporting values) rating based on the testing procedure prescribed in the ASHRAE Standard 52.2 [199]. As the MERV rating specifies small particle sizes of 0.3–1.0 μm, Chen et al. [200] experimentally examined the efficiency of 17 filter classes, including ultrafine particles (less than 0.1 μm).

The electrostatic effect has been commonly utilized in the design of air filters to enhance filtration efficiency with fine particles without incurring higher pressure drop (or energy use). Recent examples include the development and analysis of electrostatically enhanced pleated air filters [201,202], new coatings for the collection part of an electrostatic precipitator [203], and the application of the electrostatic effect to a metal foam filter [204]. Recent investigations of electrostatic filters include the effects of moisture on their performance [205] and their efficacy in filtering the COVID-19 virus under various conditions [206]. Notably, it has been observed that non-thermal plasma (as another technology) may not be an effective approach for particle filtration when compared to traditional means [207].

Material research directly contributes to the performance and fabrication of filter media. Relevant reviews include carbon nanotubes in high-efficiency particulate air (HEPA) filters [208], the processing and properties of both fibrous and nonfibrous (e.g., cellular foams) filter materials [209], and nanofibers fabricated via the electrospinning process [198]. Studies have examined how carbon black can form on fibers and influence filtration properties [210].

Since air filters can lead to higher energy use and operational costs (e.g., filter change), research has been conducted to analyze their cost and benefits. Azimi and Stephens [211] conducted numerical analyses that demonstrated indoor air recirculation with filters could achieve equivalent ventilation using outdoor air with lower operational costs. In a field study, Zaatari et al. [212] noted that the slight increase in energy use by high-efficiency filters (MERV 13/14) could be easily justified by the improved quality of filtered air. Several numerical analyses have shown that the additional operational costs caused by air filters are considerably less than the monetized health benefits [213,214,215].

In addition, the operational aspect of particle filtration has been considered. For example, due to low system runtimes and the potential for fan overload, the benefits of high-efficiency filters could be questioned in residential applications [32,216,217]. Alavy and Siegel [218] showed from their in situ study that real-life filter performance could depend on other non-filter factors, such as ventilation rate and system runtime. Further approaches have been considered to improve the applications of air filters, such as a device that reduces PM_2.5 in air intake [219], coordinating natural ventilation and filtration [220], implementing different control strategies [221], and the use of stand-alone HEPA air purifiers [222,223,224,225].

One important aim of particle filtration is to reduce the indoor concentration of PM_2.5 (particulate matter with a diameter less than 2.5 μm), the impacts of which on humans are known by health and environmental professionals [226,227]. Fisk [31] reviewed sixteen (16) studies and generally summarized modest health benefits from particle filtration. Relevant studies have been designed with different contaminant factors, such as traffic-related particles [228,229] and second-hand tobacco smoke [230], and with different health concerns, such as asthma [229] and pregnancy [230]. Nevertheless, empirical studies do not always demonstrate positive health impacts. For example, the study by Brugge et al. [231] could not identify the health benefits of using HEPA filtration in low-income homes. The review by [232] further clarified two issues. While filtration can certainly reduce indoor PM_2.5 concentrations, the health benefit is less conclusive from empirical evidence due to the challenges of study design and the variability of environmental and health conditions.

In response to the COVID-19 pandemic, various studies related to particle filtration have been conducted. Using the ASHRAE 52.2 test procedure, Zhang et al. [233] examined the performance of different types of filters in view of bioaerosol removal. In their review paper, Mousavi et al. [181] reiterated the importance of air filtration in hospital buildings. To address the shortage of isolation rooms, Mousavi et al. [234] proposed the use of portable HEPA filters with plastic barriers to accommodate infectious patients. In the context of classrooms and a dentistry school, particle filtration using MERV 13 filters [235] and HEPA filters [191,236] was also found to be effective. Through a numerical study, Faulkner et al. [237] considered that MERV 13 filters are a balanced choice (compared to 100% outdoor air, MERV 10, and HEPA filters) for office buildings.

3.2.2. Removal of Gaseous Contaminants

Common indoor gaseous contaminants include VOCs (e.g., formaldehyde) and ozone, and the removal techniques typically follow two principles: adsorption by porous materials (e.g., activated carbon) and catalytic decomposition (especially photocatalytic oxidation, to be reviewed in this section).

With the application of adsorption, Palliyarayil et al. [33] have reviewed three general types of porous materials (i.e., carbon-based, silica-based, and zeolite-based). The review by Zhang et al. [238] focused on carbon-based materials and discussed three factors that affect the adsorption process (i.e., properties of carbon-based adsorbents, properties of VOCs, and environmental conditions). After a cost and benefit analysis, Aldred et al. [239] suggested using activated carbon filtration during the ozone season. Yang et al. [240,241] studied carbon nanotubes to filter particles and ozone simultaneously and to improve the removal of formaldehyde. Ligotski et al. [242] used operational data on VOC removal rates to estimate the time to replace carbon-based filters.

Catalytic decomposition is a chemical process that aims to transfer target gaseous substances (potentially harmful to humans) into less harmful substances (e.g., water vapor and carbon dioxide). One common technology is photocatalytic oxidation (PCO), which uses semiconductor materials as catalysts with ultraviolet (UV) light as the energy source. PCO is commonly used for VOC removal as it can handle the removal task at low VOC concentrations. Comprehensive reviews of PCO applications can be found in [34,243,244]. While many possible catalytic materials for PCO exist (as reviewed in [245]), the most common material is titanium dioxide (TiO₂), which has been specifically discussed in [246,247,248]. Haghighatmamaghani et al. [249] examined the performance of four TiO₂ products from the market with variations in operational parameters (e.g., VOC concentrations, relative humidity, and residence time). Sansotera et al. [250] proposed using a perfluoro polymeric coating to immobilize TiO₂ for its stability with UV effects.

To implement the PCO process in practice, reactor design, which arranges the light source, catalyst, and airflow, becomes important [251]. An early review by [252] identified eight different designs with a discussion of three key design parameters: mass transfer rates (of contaminants), kinetic reaction rates (of PCO), and surface areas (for PCO reactions). Newer reviews have discussed the mathematical models (i.e., reaction kinetics and mass transfer models) to support the parametric design of PCO reactors [253] and the design requirements for PCO catalysts, such as selectivity and stability [251] (p. 547). Luo et al. [254] integrated lighting, mass transfer, and kinetics models with CFD to support reactor design. Besides using UV light as the energy source for the catalytic process, Xia et al. [165] investigated using thermal energy for the manganese oxide catalyst to decompose formaldehyde. In principle, PCO can also address other types of contaminants, such as particulate matter (PM) [255,256], inorganic gases (e.g., NOx, SOx, and CO), and odors [251].

The application of PCO for IAQ has been investigated for about two decades [257,258,259]. A pilot duct system has also been constructed for practical use [260]. Nevertheless, the recent review by [243] (p. 16) commented that large-scale or real-world applications for IAQ control are still limited. Reviews for advancements to practical applications have focused on intensifying the PCO process through better reactor design and synergistic photocatalytic reaction [261], utilizing light sources, and the long-term stability of catalysts [34]. Van Walsem et al. [262] demonstrated promising results from a prototype of the multi-tube reactor design, which is expected to be used in real-world applications. In our view, the PCO process remains at the component design level, and system-level integration will require further studies and investigations (e.g., how the PCO reactor can be integrated into the air distribution system or work with room air distribution).

In addition to adsorption and PCO, other technologies have been proposed to address gaseous contaminants. One example is non-thermal plasma (NTP) technology, which utilizes the ionization process to oxidize and decompose gaseous contaminants. The review by [263] covered the general issues of this technology, such as the reaction mechanisms with VOCs, the configurations of NTP reactors, the impact of humidity, and the use of catalysts. Further reviews tried to organize the design and operational parameters of NTP reactors [35,264]. Examples of relevant studies include the comparison of three metal–organic frameworks (MOFs) as catalysts for the NTP process [265], the comparison of VOC’s modular structures and their removal efficiency [266], and the comparison of three electronic air-cleaning technologies (i.e., PCO, NTP, and ozonation) using a test rig similar to a filter section of an air-handling unit [267].

Another example is phytoremediation, which utilizes indoor plants to neutralize the toxicity of gaseous contaminants mainly via metabolic processes. Three review papers have discussed this topic with various focuses on VOC removal mechanisms [268], the analysis of green walls [269], and CO₂ concentration reduction and psychological effects [270]. Jung and Awad [271] conducted a comparison study to show how palm pots can reduce CO₂ concentrations in classrooms.

Notably, the use of electronic air cleaning can produce ozone, which can be harmful to occupants [272,273]. The review by [274] examined three technologies for removing indoor ozone (e.g., active carbon adsorption, catalytic, and photocatalytic reactions). Additionally, as the PCO process has multiple stages of reactions, incomplete PCO reactions can potentially produce harmful intermediates [252] or by-products [248,260].

3.2.3. Disinfection of Bioaerosols

Bioaerosols can carry microbes and cause airborne transmission of diseases. Bioaerosols are airborne particles that can be removed via particle filtration. As particle filtration has been discussed in Section 3.2.1, we focus on disinfection technology, especially ultraviolet germicidal irradiation (UVGI), in this section.

Ultraviolet germicidal irradiation (UVGI) refers to a type of ultraviolet (UV) light with a wavelength of 200–280 nm, also known as UV-C, which can deactivate micro-organisms and viruses in air, water, and on surfaces. According to the review by Reed [36], the disinfection function of UV-C has been recognized and systematically studied since the 1930s. The principles and applications of UVGI have been documented in Kowalski’s book [275]. In response to the COVID-19 pandemic, UVGI has been reviewed and highlighted as one important technology for disinfection [276,277,278,279,280]. Concerning the fact that increasing ventilation rates are not desirable due to high energy impact, UVGI is considered an alternate approach to mitigate the infection risk [281] during the winter season [282] and in crowded spaces in hot/humid climates [283].

The UV lamp is one key component in the applications of UVGI. As discussed in [275], it can come with many types and specifications, and the traditional technology is a mercury-based UV lamp. Concerning the toxic effects of mercury on humans, one common alternative is the LED (light-emitting diode) UV lamp [284,285]. One study examined the performance of different materials used in high-frequency electrodeless light (HFEL) to generate UV light [286]. Besides the UV lamp, another important parameter is the UV rate constants, which describe a given microbe’s survival rates during UV exposure. Laboratory studies have been summarized to inform the UV rate constants in tables such as [275] (pp. 80–81) and [287] in response to the COVID-19 pandemic. UV rate constants, the ranges of the UV spectrum, and the species of microbes can then be used to estimate the disinfection efficacy of UV light.

With the UV lamp as the key component, there are multiple system configurations to apply the UV lamp in different contexts, such as standalone recirculation units, UV barrier systems, in-duct UV disinfection, and upper room systems [275], and the last two are more common in HVAC applications. In-duct UVGI systems place UV lamps in the air distribution systems to disinfect the airstream. Lee and Bahnfleth [288] studied the energy performance of in-duct systems with the variables of installation locations and climates. Yang et al. [289] showed that in-duct UVGI disinfection efficacy could be decreased with higher Reynolds numbers (or higher airflow velocity). Atci et al. [290] numerically analyzed four lamp array configurations in an in-duct UVGI system. Sarabia-Escriva et al. [291] developed a mathematical model to evaluate the killing ratio (of a pathogen) of in-duct systems with experimental validation. Luo and Zhong [292] reviewed and analyzed the in-duct UVGI systems in view of design factors (e.g., duct sizes, lamp arrangements, and UV dosage) and system performance (i.e., inactivation efficiency and energy consumption).

Upper-room UVGI systems install UV lamps on the upper wall or ceiling, and their design factors (e.g., sizing and air mixing) have been reviewed in [275] (Chapter 9) and [282] (p. 24). According to [36], the study by Well et al. [293] should be the first successful application of upper-room UVGI systems in an occupied room for infection control. Studies have been performed combining upper-room UVGI systems with different room air distribution approaches, such as a ceiling fan [294] and displacement ventilation [295]. Nunayon et al. [296] demonstrated that the rotating feature of the UV LED lamp can improve IAQ in poorly-mixed conditions. Notably, UVRI has also been applied to disinfect cooling coils [275] (Chapter 10.6) [297,298].

Overall, UVGI has several advantages, such as relatively simple installation and low energy use (compared to the equivalent effect using ventilation), and it can be used to supplement existing IAQ control methods [299]. At the same time, caution is needed when using UVGI for occupied spaces due to the possible hazards from UV light to occupants (e.g., potential injury to eyes and skin and generation of ozone) [275] (Chapter 12).

In addition to UVGI, another technology utilizes the disinfection properties of nano-structured materials, which can be applied to traditional air filters to inactivate pathogens. The review of such nano-structured materials for disinfection can be found in [300,301]. Some relevant papers include the comparison studies of disinfection materials applied to air filters [302,303], the making and evaluation of silver–polyacrylonitrile nanofibers [304], and the evaluation of the coating of a silver–silica composite on different air filters [305,306].

PCO has also been considered for disinfection [307,308]. Though it may not be considered a primary means, the COVID-19 pandemic has motivated discussions about using PCO for infection control. For general disinfection, two categories of photocatalysts have been reviewed: metal oxide [309] and graphene [310]. In the context of air cleaning, Truong et al. [311] generally reviewed different types of photocatalysts for degrading chemical contaminants and pathogens, as well as the fabrication techniques and applications. Poormohammadi et al. [312] summarized three disinfection mechanisms (i.e., chemical oxidation, attack of metal ions, and morphological damage of viruses), along with the discussion of environmental factors (i.e., airflow rate, relative humidity, and reactor temperature) affecting disinfection performance. The review by [313] focused on the synergy of using UV light and PCO for disinfection. Other technologies that utilize reactive species for disinfection include ionization [314,315,316], non-thermal plasma [317], and photoelectrochemical oxidation [318].

4. Dissemination Types and Control Strategies

Assuming that the supply air is clean, occupants can still be exposed to contaminants when these contaminants are generated within an indoor space and disseminated to other spaces. By analyzing the spatial relationship between contaminant generation and the occupants who may be exposed, we can classify four types of dissemination, as illustrated in Figure 3. The first type is source control, which arguably does not involve dissemination. However, for the purpose of comparison, we classify source control in this manner since it is a common practice. This section will explore these four types of dissemination and the relevant control strategies.

4.1. Indoor Generation of Contaminants and Source Control

If the sources of contaminants are known and expected to be generated in indoor spaces, it is often more effective to develop control strategies for these sources rather than relying solely on ventilation and air-cleaning approaches. Source control is a common strategy for improving IAQ, and common examples include kitchen ventilation (e.g., the 2019 ASHRAE Handbook–HVAC Applications, Chapter 34), laboratory fume hood (e.g., the 2019 ASHRAE Handbook–HVAC Applications, Chapter 17) [319,320], industrial local exhaust (e.g., the 2019 ASHRAE Handbook–HVAC Applications, Chapter 33) [321,322], and the use of an air curtain for emergency control [323,324].

As one example for elaboration, we can briefly discuss the topic of kitchen ventilation. The impact of cooking effluents on human health is well recognized [325,326,327]; the use of exhaust hoods is regulated in practice (e.g., the ASHRAE Standard 154 [328]), with the benefits well documented (e.g., [329,330]). As newer houses become more airtight, exhaust hoods play a crucial role in removing PM_2.5 particles generated from residential kitchens [331,332]. The performance of exhaust hoods can be measured through standardized tests [333,334,335]. In addition to using exhaust hoods, Zhao et al. [327] have reviewed how air distribution strategies around cooking areas and makeup air systems can affect the removal of contaminants. The study by [332] has also shown how occupants’ usage patterns of exhaust hoods can impact exposure to particulate matter related to cooking.

In the context of COVID-19 or infectious virus transmission, infected individuals can be considered the source of contaminants. Contaminant spread occurs during events such as coughing and sneezing [336,337,338], as well as talking and singing [339,340]. While face masks are often perceived as a protective measure for wearers, they can also be viewed as an effective means of source control [341,342,343]. In their studies, Lindsley et al. [344] compared the effectiveness of various face coverings (e.g., N95 respirator, cloth mask, and neck gaiter) as source control. Konda et al. [345] investigated the filtration efficiency of different cloth materials used for face masks. Cheng et al. [342] used the concept of “airborne virus abundance” to analyze the literature results and explain how face masks can be more effective in virus-limited conditions. The interactions of airflow, face masks, and the spread of aerosols involve several complex mechanisms in fluid mechanics, as analyzed and reviewed by various physicists [346,347,348]. Notably, there is abundant literature on face masks and COVID-19, and this review, focusing on engineering controls, cannot comprehensively cover all of this information. However, the papers reviewed in this paragraph align the importance of source control in the IAQ context with the findings of studies on face masks. For example, the study by Ueki et al. [343] showed that individuals wearing face masks as potential virus spreaders provide better protection than those wearing masks solely as receivers of the virus.

Engineering controls for isolating infectious patients can also be considered a form of source control. Airborne infectious rooms (AIIRs), which employ directional airflow and negative pressure, are commonly used to hospitalize COVID-19 patients [349]. Due to the increased demand for hospital beds, the conversion of existing facilities for COVID-19 patient care has been reported, utilizing various isolation techniques. These include headboard ventilation in alternative care sites [7], the creation of temporary anterooms with plastic barriers and portable air purifiers [234], personal portable booths with pressure differential control using fan-HEPA filter units [350], and the establishment of temporary negative-pressure rooms [351]. Another approach to source control is the use of aerosol boxes to protect healthcare staff during medical procedures (e.g., intubation) that generate aerosol from patients [352,353,354].

4.2. Dissemination in the Same Space

The dissemination of contaminants within the same space is closely related to the design of room air distribution, as discussed in Section 3.1.2. In this section, we review other factors and interventions that impact this type of dissemination. One pertinent topic is the airborne transmission of infectious droplets between two individuals within the same space. In the review conducted by Ai and Melikov [37], approximately seven (7) factors are discussed: air distribution, supply flow rate, relative distance, posture, breathing mode, particle size, and human movement. They also reviewed experimental techniques (e.g., thermal manikins) and computational methods (e.g., CFD) for research investigations, along with quantified indices to assess cross-infection risk.

Relative distance or social distancing is a crucial factor affecting the cross-infection risk between two occupants. Liu et al. [355] distinguished between direct exposure to infectious droplets (i.e., droplets from the infector to the receiver through an expiratory jet) and indirect exposure (i.e., droplets dispersed and taken by the receiver). They noted that indirect exposure is more closely related to ventilation rates and room air distribution. Anchordoqui and Chudnovsky [356] explained and demonstrated how convective airflow (which can be influenced by various factors such as interior objects and moving occupants) can carry airborne droplets beyond 6 feet. Feng et al. [357] examined how ambient wind and relative humidity can result in droplets traveling more than 6 feet and depositing higher fractions on objects.

The comparison of mixing and displacement ventilation has been explored by [355,358], demonstrating that the buoyancy effect in displacement ventilation could trap exhaled droplets in the breathing zone, leading to a higher risk of direct exposure. Ai et al. [359] investigated how stratum ventilation (or horizontal air distribution) could effectively mix exhaled air and make the cross-infection risk less dependent on the relative distance.

Interior partitioning can alter indoor airflow and impact IAQ. Ahn et al. [360] conducted a numerical investigation into how different ventilation types (mixing versus displacement) and the placement of supply air diffusers affect IAQ in two partitioned spaces. They observed that partitions can hinder effective air mixing. For instance, having an air supply and exhaust in the same partitioned space may lead to short-circuiting airflow.

In response to COVID-19, physical barriers have become common interventions to reduce the risk of cross-infection. In classroom settings, the efficacy of using partitions to protect students has been demonstrated both numerically [361] and experimentally [362]. Ren et al. [363] noted that the protective effect of physical barriers diminishes when a person is positioned far from the air exhaust (more than 4 m). Ye et al. [364] and Liu et al. [365] observed that partitions on dining tables can trap exhaled air from previous diners, thereby increasing the risk of cross-infection for new diners. The notion of direct and indirect exposure [355], discussed earlier, can probably explain some trade-offs associated with using partitions. While partitions are designed to protect individuals from expiratory jets or direct exposure, they may not be as effective in guarding against indirect exposure. Furthermore, partitions generally impede air mixing in the space [365], potentially leading to a higher risk of infection in specific situations.

Other interventions in response to COVID-19 include intermittent occupancy (intentional breaks between occupants entering and leaving the space, studied in classroom settings) [366] and the use of ceiling fans (to promote air mixing) [367]. Reviews and studies assessing infection risks in specific spaces encompass classrooms [183,368], aircraft [369,370,371,372,373], passenger cars [374], buses [375], enclosed or confined spaces [10,376], gyms [377], concert halls [378], and washrooms [379].

Notably, engineering controls (e.g., ventilation) tend to be less effective for close-contact transmission [15,380]. Furthermore, close-contact transmission is often associated with “short-term exposure events”, where physical mechanisms tend to be more transient (i.e., non-steady state) and dynamic (i.e., time-dependent) [381,382].

4.3. Dissemination between Adjacent Spaces

When indoor air pollutants are generated in one space, these pollutants can be disseminated to adjacent spaces through interior openings such as doors. In practice, the primary approach to address this issue is maintaining a pressure gradient (or differential) between spaces, with negative pressure relative to adjacent spaces being maintained in the “contaminated” space. This approach reduces the chance of pollutants unintentionally “spilling over” into adjacent spaces. Achieving this requires differentiating supply and return air flow rates and monitoring pressure values in different spaces. Early discussions revolved around determining the appropriate values for the pressure difference between adjacent spaces [383]. In the context of airborne infectious isolation rooms (AIIRs), Adams et al. [384] referenced the standards of their time and simulated the spread of droplet nuclei at pressure differences ranging from 2.5 Pa to 20 Pa, recommending higher pressure differences due to the movement of healthcare workers. For cleanroom design, Sun [385] stated that a pressure difference of approximately 5 Pa is sufficient between two spaces separated by an airlock door. To maintain the pressure gradient with variable supply air volume (e.g., demand-controlled ventilation for energy saving), Cheng et al. [386] proposed a new method to determine the variable residual air volume.

Several indoor activities (e.g., door motions and occupant movements) can compromise the intended pressure gradient. Some research studies have investigated how these activities affect IAQ between spaces. For example, several studies have examined the effects of two types of doors: hinged or swing doors versus sliding doors [387]. In the context of hospital applications, while studies tend to recommend sliding doors [388,389], door opening with foot traffic will still lead to undesired air exchange [390,391,392]. Considering the use of swing doors, studies have investigated the effects of door opening directions concerning positive/negative pressures [383,393], opening sizes [393,394], air temperature differences between two spaces [387,395], and door opening time and speed [396]. Several studies have also examined the effects of occupants passing through doors [389,390,393,397,398].

In addition to simple doors, anterooms and airlocks are commonly used as barriers between critical indoor environments and corridors. Critical indoor environments encompass cleanrooms [385,386], specific hospital spaces [399,400], and biosafety laboratories [401]. Each anteroom (or airlock) typically consists of two doors opened at different times to maintain the pressure gradient between the corridor and the critical space. In an anteroom, air can be filtered or displaced with pressure control to minimize cross-contamination. For design guidance, Sun [385] explained the contaminant migration rate as a performance measure for airlocks. Subsequently, Sun [385] described four configurations of different pressure gradients with airlocks (i.e., cascading, bubble, sink, and dual compartment) and the required time delay (RTD) between door operations. These can be considered as design parameters for airlocks.

Numerous studies have explored the effectiveness of anterooms under various environmental factors, including door motion [402], secondary infection (transmission from general patient rooms to isolation rooms) [403], transmission through shared anterooms [404], and the passage or movement of occupants [398,405]. Andalib et al. [400] conducted a systematic review of anteroom effectiveness in hospital applications and identified pressure gradients and air changes per hour as key influencing factors.

The dissemination of indoor air pollutants to adjacent spaces has also been discussed in other specific applications, such as control rooms of nuclear plants [406,407], the interface between houses and garages [408,409], and bathroom ventilation [410,411,412]. Multizone analysis has also been applied to study the spread of indoor air pollutants within buildings [399,413,414].

4.4. Dissemination between Non-Adjacent Spaces

Regarding dissemination between non-adjacent spaces, we aim to explore potential means or mechanisms around interior openings that can transport indoor air pollutants between two spaces. This review discusses two possibilities: dissemination through the ductwork/HVAC systems and the stack/wind/buoyancy effects.

If the sizes of indoor air pollutants are small enough (e.g., smoke, odor, aerosols, and PM_2.5), they can remain suspended in the air for hours (see the 2021 ASHRAE Handbook—Fundamentals, Chapter 11). When such pollutants are generated in one space, there is a possibility that they can be transported with return air (RA) to another space through the ductwork. As a result, it is not uncommon to encounter building codes that prohibit using return air (RA) from certain spaces (e.g., public corridors per the National Building Code of Canada 2020, Article 6.3.2.11). Additionally, maintaining pressure balance is crucial to maintaining desirable airflow in the return air duct (see the 2019 ASHRAE Handbook–HVAC Applications, Chapter 39). In the context of COVID-19, Manassypov [415] analyzed this type of dissemination, considering two mitigating control factors: the amount of ventilating or outdoor air (OA) and filter efficiency. Chirico et al. [416] conducted a rapid review to investigate evidence of viral transmission (primarily SARS, MERS, and COVID-19) through the ductwork/HVAC systems. In their literature analysis, Chirico et al. [416] classified spatiotemporal patterns (e.g., [417]) and computational models (e.g., [418,419]) as indirect evidence, while environmental samples from HVAC/ductwork were classified as direct evidence. They found that the only direct evidence came from a MERS-related study by Kim et al. [420], where one sample from the air exhaust damper tested positive. When proper ventilation and filtration systems are in place, the infection risk associated with return air should be low [421] or manageable [422].

Other mechanisms can transport indoor air pollutants from one space to another non-adjacent space within the same building. One such mechanism is the stack effect, which represents indoor vertical and upward transmission. Lim et al. [423] analyzed and demonstrated how the stack effect could transmit infectious particles from lower to upper levels in a high-rise hospital using field measurements, multi-zone airflow simulation, and tracer gas. In contrast, another mechanism is outdoor vertical transmission between two units through windows driven by wind and buoyancy effects, also referred to as re-entry airflow [424]. Li et al. [418] used multi-zone modeling to analyze probable virus transmission between wind-driven flats and buoyancy effects. Gao et al. [425] used tracer gas simulation to analyze how the buoyancy effect could cause vertical transmission between flats in a high-rise building. Liu et al. [426] used the re-entry ratio to evaluate the risk of cross-unit contamination caused by wind and buoyancy effects. Additionally, the dissemination of air pollutants through cracks in building envelopes has been studied by [427,428]. Choi and Kang [429] further examined how leaky buildings could have poorer IAQ (in terms of PM_2.5) due to a higher infiltration rate.

5. Sensors and Information Technology

Compared to the use of thermostats and the control of thermal comfort, it seems that the applications of sensors and information technology for IAQ are less common in practice. One related topic is smart ventilation, which has been reviewed in Section 3.1.3. The COVID-19 pandemic has motivated more attention to these engineering techniques. For example, Omidvarborna et al. [430] discussed and reviewed five steps (i.e., sensor selection, calibration, deployment, data processing, and predictive modeling) to implement IAQ sensors for smart homes. One recent book [431] has collected articles on the applications of the Internet of Things (IoT) and artificial intelligence (AI) to support IAQ improvements. This section will review this topic in three aspects. First, we will review the sensors for IAQ monitoring. Then, we will review the applications of wireless communications and the Internet of Things (IoT) in the IAQ context. Lastly, we will review how the sensor data can be utilized for managing IAQ.

5.1. Sensors for IAQ Monitoring

The topic of IAQ sensors is broad, as reflected in multiple review papers. Morawska et al. [432] reviewed low-cost sensors for air pollution monitoring with use cases for outdoor air (stationary and mobile modes), indoor air, and personal monitoring. The review by Chojer et al. [38] and the study by Clements et al. [433] highlighted concerns about the quality of low-cost sensors (e.g., data reliability, long-term stability, and calibration information). Saini et al. [39] organized the review content of IAQ sensors to address six research questions (e.g., sensors’ types, specifications of measurements, and operating conditions). Schütze and Sauerwald [434] reviewed gas sensors and metal oxide semiconductor (MOS) sensors. Tran et al. [14] reviewed two-dimensional nanostructured materials and MOS, active research areas for improving sensor performance. Using an environmental chamber, Demanega et al. [435] examined eight consumer-grade IAQ monitors and eight single-parameter sensors. They considered that IAQ sensors can generally detect air-polluting events and reasonably track reference measures, though the quality of quantitative measures can be further improved.

To manage the scope of this review paper, we focus on sensors for three common types of indoor air pollutants: carbon dioxide (CO₂), particulate matter (PM), and volatile organic compounds (VOCs) [435]. In each type, we concentrate on the common sensing principles and review the technical capacity from an engineering standpoint.

5.1.1. Review Notes for CO₂ Sensors

One common sensing principle for CO₂ is non-dispersive infrared (NDIR), which leverages the absorption band property of CO₂ (4.2 μm) under an infrared light source (see the 2021 ASHRAE Handbook–Fundamentals, page 38.25). For example, as noted in the review by [38], the CO₂ sensors employing this sensing principle are based on NDIR [436,437,438,439,440,441,442,443], although there are some exceptions, such as a metal oxide semiconductor [444], electrochemical [445], and potentiometric [446] sensors. Demanega et al. [435] identified poor measurement results from one product (Foobot) that uses VOC data to correlate the CO₂ concentration.

In terms of engineering considerations, the reported accuracy of CO₂ sensors in review papers generally falls within the range of 50 to 100 ppm, which can be influenced by air temperature and relative humidity [435,447]. However, it is important to note that high sensing accuracy is not always necessary for IAQ control decisions. Therefore, we align with the discussion presented in [435], suggesting that the consumer-grade CO₂ sensors based on NDIR are generally sufficient. Other important factors to consider include the sensor’s response time and maintenance requirements (e.g., calibration and “automatic baseline correction”, as discussed in [435]).

For IAQ applications, CO₂ sensors are considered useful for measuring occupancy levels and have been used for demand-controlled ventilation [110,448]. Concerning COVID-19, while some studies have suggested using CO₂ concentrations to assess infection risk [449,450], the ASHRAE position document on filtration and air cleaning (at https://www.ashrae.org/file%20library/about/position%20documents/filtration-and-air-cleaning-pd-feb.2.2021.pdf, accessed on 16 July 2022) suggests that relying solely on CO₂ concentrations may not be sufficient due to the presence of other factors (e.g., occupant activities and use of filtration and air-cleaning techniques that do not affect CO₂ concentrations).

5.1.2. Review Notes for PM Sensors

One common sensing principle for PM is light scattering, in which the sensor cell detects light scattered by particles [451]. Examples from the review by [38] include [438,440,443,452]. While a particle size of about 2.5 μm (i.e., PM_2.5) is most common for measurement, other particle sizes have also been employed in studies, such as PM₁₀ [451] and PM₁ [453].

Due to the emergence of low-cost PM sensors, several studies have examined the performance of these sensors compared to more expensive (standard) techniques (e.g., the gravimetric method). These studies include laboratory assessments of the adequacy of low-cost PM sensors [454,455,456], comparisons between laboratory and field performance [457,458]), and examinations in residential environments [451,459,460]. Additionally, a comparison study for indoor and outdoor applications has been conducted [461]. Generally, low-cost PM sensors are adequate for event detection [453,460], but they may face challenges when measuring low PM concentrations [461], small-size particles (less than 0.3 μm) [451,459], and under varying relative humidity conditions [462]. Calibration needs have also been emphasized [455,463].

In contrast to CO₂, the literature has provided some evidence of a link between PM and COVID-19 infection [453,464]. There are two explanations for this. One is that PM could cause the overexpression of the alveolar ACE-2 receptor of the COVID-19 virus [465,466]. Another explanation is that infectious virus particles could attach to the surface of larger PM particles, potentially facilitating more effective airborne transmission [467].

5.1.3. Review Notes for VOCs Sensors

Volatile organic compounds (VOCs) are hazardous organic substances that exist in a vapor state under atmospheric conditions due to their low vapor pressure. Attention to VOCs in the context of IAQ can be traced back to a World Health Organization (WHO) report in 1989 (cited from the 2021 ASHRAE Handbook–Fundamentals, page 11.25), which listed and explained how VOCs in buildings could impact human health. Reviews by [434,468] have a more specific focus on VOC sensors. Without exhaustively covering all possible sensor technologies, this section highlights three common ones that can potentially be used for low-cost sensor deployment.

One common technology is the use of a metal oxide semiconductor (MOS), in which electrical resistance (or conductivity) can be changed in the presence of VOCs and other gaseous contaminants (e.g., carbon monoxide and nitrogen dioxide) (e.g., [440,469,470]). As commented by [14,434,468], research on MOS for gaseous sensing remains active, aimed at improving sensor performance for low-concentration detection, enhancing selectivity, and reducing maintenance requirements (e.g., energy source and routine calibration).

Another technology is electrochemical sensors, where the presence of VOCs can react with the electrodes, leading to changes in electrical charges (e.g., [471]). Environmental factors such as humidity and wind velocity could affect the performance of this type of sensor [468]. Although electrochemical sensors may be considered an “old” technology [468], materials research such as metal–organic frameworks (MOF) [472] has continually advanced the performance of this type of sensor.

The third technology to be discussed is photo-ionization detectors (PIDs), which use ultraviolet (UV) light to ionize gas molecules for detection (e.g., [473,474]). As commented by [468], PIDs tend to have low sensitivity, and their portable version can be expensive.

5.2. Wireless Communications and the Internet of Things (IoT)

It is common to employ wireless networks for IAQ sensors to maintain a low-cost infrastructure [475]. The options for wireless communications for IAQ have been reviewed in [443,476,477,478], and common networks include wireless sensor networks (WSN) (e.g., ZigBee) and wireless local area networks (WLAN) (e.g., Wi-Fi). Typically, a WSN consists of a set of sensor nodes (e.g., IAQ sensors), which are coordinated via a host node [479] or a base station [480]. The base station can communicate with a control center [479] or an application server [481] for data collection and management, with the potential development of a web interface [482]. Most examples of IAQ WSN track CO₂ (e.g., [446,475,480,483,484]). Based on the IEEE 802.15.4 standard, ZigBee is a common wireless protocol in these applications (e.g., [482,484,485,486]). The use of Wi-Fi communications tends to be more common when it comes to the Internet of Things (IoT) applications (e.g., [487,488,489,490,491]). Other wireless communications include mobile networks for personal IAQ applications [492], the long-term evolution (LTE or 4G network) [493], and low-power wide-area networks [478].

Extended from wireless communications, the Internet of Things (IoT) technologies feature their connection to Internet services, which allow for more diverse possibilities in data analysis and system control. For example, Chen et al. [494] described how cloud services can take IAQ sensor information for mobile apps and building operations. Benammar et al. [436] described an IoT-based IAQ monitoring system with error detection and data backup functions. Marques et al. [487] developed an IoT system to track indoor CO₂ to support “ambient assisted living.” Zhang et al. [495] used Raspberry Pi as a central hub to communicate with IAQ sensors and connect to Internet services. Coulby et al. [496] developed and examined a multimodal approach with IoT support for IAQ monitoring. Calvo et al. [497] adopted the edge–fog–cloud structure to develop an IoT-based IAQ monitoring system. Feng et al. [498] advanced the particle swarm optimization technique to locate indoor contaminant sources.

5.3. Applications with IAQ Data

The preceding sections of IAQ sensors and wireless communications can be seen as available engineering approaches for collecting IAQ data. In this section, we will explore how the collected IAQ data can be practically utilized.

As IAQ is not a typical or default measure for building spaces in general, IAQ sensors and monitoring systems have primarily been applied to specific spaces of interest to investigate their IAQ conditions. In the context of IAQ and transportation, studies on IAQ in subway systems can be found in [499,500,501], covering topics such as multivariate statistics for IAQ data analysis and detecting faulty sensors. Zhao et al. [502] compared the IAQ of three types of egg production systems (i.e., conventional cage house, aviary house, and enriched colony house). De Gennaro et al. [503] investigated various factors (e.g., fuel quality and ventilation) contributing to IAQ in residential spaces using biomass for heating. Liu et al. [504] deployed their self-developed sensor to monitor the IAQ of a residential space. Firdhous et al. [505] measured the ozone level near a photocopier, representing a single contaminant source. Wang et al. [506] studied how temperature, humidity, and carbon monoxide concentration can be related to a student’s performance in a school. Kaduwela et al. [507] demonstrated how a low-cost sensor could provide information on IAQ issues (e.g., ventilation efficiency and distant wildfire) in classrooms. Palmisani et al. [508] used low-cost sensors to monitor CO₂, PM, and VOCs in a hospital environment.

When receiving measurement data from sensors, the data analysis functions for IAQ can be classified into two major types: interpretation and prediction.

For interpretation, data from sensors are interpreted for meaningful actions such as warning signals and system controls. For example, Ramalho et al. [509] examined how CO₂ is correlated with IAQ pollutants and how it can be used as a proxy of IAQ for health guidance. Dionova et al. [510] used fuzzy logic to determine an overall IAQ index from sensor data, and Pradityo and Surantha [511] used a similar technique to control different levels of an exhaust fan. Ha et al. [512] used the Kalman filter for IAQ data fusion. Mad Saad [513] used supervised machine learning to analyze sensor data and classify the sources of indoor air pollutants.

For prediction, the data analysis process aims to predict the trends of IAQ using current data so that proactive control actions can be planned. For example, Deleawe et al. [514] studied the predictive power of three machine-learning methods for indoor CO₂ levels. Yu and Lin [515] used a statistical technique (namely, an autoregressive integrated moving average) to predict CO₂ concentration based on data trends and historical random errors. Skön et al. [516] and Khazaei et al. [517] used air temperature and humidity to predict CO₂ concentration using neural networks. In their IAQ predictive analytic study, Mumtaz et al. [518] identified neural networks as an outperforming approach.

As outlined by Doukas et al. [519] in their “model’s philosophy”, sensor data and information technology have the potential to impact indoor environments and improve building operations significantly. One practical application is demand-controlled ventilation, which monitors indoor CO₂ levels to adjust ventilation rates for energy saving [110,448,520,521]. Zhang et al. [522] have integrated genetic algorithms and artificial neural networks to control ventilation actions and improve IAQ indices. To mitigate indoor infection risks associated with COVID-19, information-based technologies have been developed, such as social distance monitoring [523,524], cough detection [525], and the use of robots for disinfection [526,527].

Regarding the deployment of IAQ monitoring systems, several studies aim to improve various aspects, including the use of renewable energy sources [528], the design of sensor networks (e.g., determining the number and placement of sensors) [529], and the utilization of mobile IAQ data and crowd sensing [530,531,532].

With the emergence of low-cost IAQ sensors and information technology, we can anticipate increased adoptions of these technologies in practice. At the same time, we should note the gap between the availability and utilization of IAQ data. While IAQ data may become more economically accessible, we must learn how to effectively utilize real-time data for better building operations and occupant satisfaction. Since HVAC and buildings are complex systems, achieving this goal will require a deeper understanding and collaboration among electrical, software, and mechanical engineering disciplines.

6. Implications to Systems Design

Engineering controls for IAQ reviewed in this paper constitute a complex landscape that calls for more insights from the systems design perspective. To contribute to this direction, we first examine and distinguish component-level and system-level engineering controls. Then, we describe the concept of functions and perform a functional analysis of engineering controls. To demonstrate the application of functional analysis, we analyze the mitigation of infection risks and distinguish short- and long-range transmissions requiring different engineering controls. Lastly, we utilize systems design concepts and techniques to suggest some systems design directions for IAQ.

6.1. Component-Level and System-Level Engineering Controls

Component-level engineering controls can be referred to as components or sub-systems that deliver specific IAQ functions, the performance of which can be readily assessed in the laboratory or specified by some standards. Examples include air-cleaning technologies (e.g., air filters, photocatalytic oxidation, and ultraviolet germicidal irradiation) and IAQ sensors (e.g., non-dispersive infrared, light scattering, and metal oxide semiconductors). As reviewed in Section 3.2 and Section 5.1, the continual advancement of component-level engineering controls is expected through scientific research (e.g., advanced materials, nanotechnology, and improvements in related electrical/chemical/disinfection processes).

At the same time, we should also note the gap between the laboratory results and real-world applications, which demand more operational requirements such as ease of maintenance and operation, high standards of safety, and the cost of large-scale deployment. To illustrate this point, we will compare two examples of IAQ controls below.

(1)

Ultraviolet germicidal irradiation (UVGI): While the disinfection effect of UVGI has been well studied [275], the potential hazard of UV light to occupants has restricted its scope of applications and implied the need for a more complex design (e.g., the design of safety measures to mitigate the potential hazard);

(2)

Low-cost IAQ sensor: In contrast, low-cost IAQ sensors have become more popular in practice [432]. Besides their cost-effectiveness, they are considered low-risk and relatively easy to operate from the perspective of building owners and facility managers.

By comparison, system-level engineering controls require multiple components or sub-systems to deliver their primary functions. The performance of system-level controls tends to be sensitive to environmental or situational factors (e.g., outdoor weather conditions and occupancy details), which are less controllable from an engineering standpoint. To illustrate, three examples from this review are elaborated on below.

(1)

Smart ventilation (Section 3.1.3): Besides the traditional ventilation function (e.g., diluting the indoor air contaminants), it also requires a monitoring function to assess indoor conditions for delivering outdoor air to the spaces;

(2)

Natural ventilation (Section 3.1.4): While natural ventilation can be achieved by simply opening the windows, its effectiveness on IAQ and its impact on thermal comfort highly depend on the space’s geometry, the window’s size and orientation, and the weather conditions;

(3)

Dissemination controls (Section 4): While source control can be viewed as one component-level type, the effectiveness of other dissemination controls depends on less controllable factors, such as interactions between supply air/exhaust locations/interior partitions and the movement of doors/occupants. A system solution is needed to control the zone pressures of multiple spaces to achieve the intended pressure gradients.

As observed, both component-level and system-level engineering controls contribute differently to the overall IAQ goal. On the one hand, scientific and technological advancements for component-level controls are expected to be supportive and helpful in practice. On the other hand, the overall IAQ performance of a building depends on system-level controls, as one weak link can compromise the efforts of other components.

For example, suppose that the component-level control can ensure the cleanliness of the supply air. This cannot effectively eliminate the risk of indoor air contaminants generated and disseminated in the same space if the dissemination pathway does not interact with the supply air. This also highlights the importance of the systems design perspective in delivering a holistic IAQ solution. In the next section, we will analyze the functions of the IAQ controls further to help develop the systems design perspective.

6.2. Functional Analysis of Engineering Controls for IAQ

Function is a common concept in systems engineering, and it can be interpreted as “a specific or discrete action that is necessary to achieve a given objective” [533] (p. 86). From functions to design solutions, the notion of “form follows function”, originated from architecture, has been adopted in the generation of design concepts [534,535], advocating that designers should analyze the system’s functions first before developing the solutions (or forms). This practice should help designers clarify what they want to achieve first (in view of functions) without being restricted to specific solutions.

Based on the literature content reviewed in this paper, we conduct functional analysis, which initially results in three top-level functions. Figure 4 illustrates a simple tree structure to organize these top-level functions, briefly discussed below.

• Function 1: Reduce the concentration of indoor air contaminants. This function is associated with the “cleaning” aspect of IAQ and is related to the content in Section 3;
• Function 2: Hinder the dissemination of indoor air contaminants from indoor sources. This function is associated with the “dissemination” aspect of IAQ and is related to the content in Section 4;
• Function 3: Monitor and control IAQ. This function is associated with the “information” aspect of IAQ and is related to the content in Section 5.

Notably, these top-level functions somewhat indicate their independence in terms of what they aim to achieve. Therefore, we should not expect that a specific control or solution for one function would be effective in achieving another function. For example, while increasing the ventilation rate can reduce the concentration of indoor air contaminants (i.e., Function 1), we should not expect it to be an effective approach for preventing the dissemination of contaminants over short distances (i.e., Function 2). Although higher ventilation rates can weaken overall dissemination (due to exposure to lower concentrations of contaminants), this control should only be seen as a supportive intervention (but not a direct one) because it does not intentionally intercept or hinder the dissemination pathway.

At the next level of functional analysis, we can further decompose each top-level function to refine the details of solution approaches. Based on the literature content in this paper, we present the second-level functions and their potential solutions in

Table 2. Two levels of functions for IAQ engineering controls.

Top-Level Function	Second-Level Function	Examples of Engineering Controls
Function 1: Reduce the concentration of indoor air contaminants	Supply outdoor air	Determine ventilation rates; Specify ACH; Design air distribution to zones
	Remove airborne particles	Select filters; Determine filter locations and pressure drops
	Neutralize gaseous contaminants	Consider adsorption technology, photocatalytic oxidation, non-thermal plasma, and phytoremediation
	Disinfect bioaerosols	Consider ultraviolet germicidal irradiation and filters (to disinfect and trap bioaerosols)
Function 2: Hinder the dissemination of indoor air contaminants from indoor sources	Restrict the dissemination from the sources	Identify and contain the source of contaminants
	Block the dissemination in the same space	Design and control room air distribution; Promote social distancing; Use interior partitioning/intermittent occupancy/ceiling fans
	Disconnect the dissemination between adjacent spaces	Control pressure gradients; Design and control space entrances
	Identify and mitigate the dissemination between non-adjacent spaces	Analyze dissemination risks due to ductwork, stack effect, and buoyancy effect
Function 3: Monitor and control indoor air quality	Detect the concentration of air contaminants	Identify target contaminants and select sensors
	Execute the control actions	Send signals to occupants; Control HVAC systems
	Diagnose the IAQ situations	Determine acceptable IAQ conditions; Characterize events for poor IAQ
	Communicate and store the IAQ data	Select information and communication technology

. For each top-level function, we extend its second-level functions as follows.

• In Function 1, four second-level functions are identified, where the function “supply outdoor air” can be viewed as a general strategy. In contrast, other second-level functions specifically target three types of contaminants: airborne particles, gaseous contaminants, and bioaerosols;
• In Function 2, the second-level functions are classified to address four dissemination cases discussed in Section 4: containment of sources, dissemination in the same space, between adjacent spaces, and between non-adjacent spaces;
• In Function 3, the second-level functions are defined to cover the typical functions in control systems (e.g., the use of sensors and control signals) and information technology (e.g., IAQ data analysis and communication).

Second-level functions can be either duplicative or coordinative in achieving the top-level functions. Functions are considered duplicative if they can be replaced with each other in principle. For example, two second-level functions, “supply outdoor air” and “remove airborne particles”, can be viewed as duplicative since the use of air filters can reduce the need for the increased supply of outdoor air to maintain an adequate level of IAQ (e.g., [211]). In contrast, functions are coordinative if they need to work together to deliver the top-level functions. For example, the second-level functions associated with IAQ sensors and control actions must be coordinated to achieve the top-level Function 3 (i.e., monitor and control).

The functional analysis results can support the development of the systems design for IAQ in two ways. First, the functions (both top-level and second-level) can assist designers in reviewing system requirements and clarifying the system’s capacity limits. For example, if designers recognize the limitations of current engineering solutions for dissemination control within the same space, it can motivate them to recommend non-engineering interventions (e.g., policy and regulation) to address certain situations. This approach also encourages a more holistic perspective beyond engineering controls to enhance IAQ for occupants.

Second, the articulated functions (especially Function 3) illustrate opportunities for them to collaborate and achieve desirable system properties such as resiliency and flexibility. One example is demand-controlled and smart ventilation (Section 3.1.3), which combines Functions 1 and 3. Another example is that while dissemination control within the same space (i.e., Function 2) can be challenging from an engineering perspective, certain information technology tools (i.e., Function 3) like social distance monitoring [523,524] and cough detection [525] can assist the system and occupants in taking mitigating actions, such as warning signals to occupants.

Table 2 has listed twelve (12) second-level functions, some of which can be further refined for specific IAQ technologies. It is important to consider that some of these functions must work together to deliver higher-level functions. This illustrates that systems design for IAQ can be complex in practice when aiming to develop a holistic solution approach. The next section will demonstrate how to apply this functional analysis to examine mitigating infection risks associated with IAQ.

6.3. Functional Analysis Example: Mitigation of Infection Risks

Infection risk is a multi-disciplinary issue. We initiate the analysis with reference to a review paper by Leung [536] from the public health perspective. In this paper, Leung [536] explained the concept of disease transmissibility, which involves five factors. We illustrate these factors in Figure 5. The first three factors, namely the contagiousness of infected persons, the infectivity of pathogens, and the susceptibility of exposed persons, are more associated with medical knowledge but less with the built environment. Therefore, engineering controls are not considered effective for addressing these factors. In contrast, the fourth factor is related to the environmental stress imposed on the pathogen and can be connected to disinfection controls (e.g., Section 3.2.3). The fifth factor pertains for contact patterns between infected and exposed individuals, and it is more closely related to health and public policy (e.g., social distancing and limiting indoor occupancy). As observed in this discussion, the ease with which an exposed person can become infected in an indoor environment depends on various factors, where engineering controls can only be effective to some of them.

Further, Leung [536] explained two types of transmission: short-range and long-range. Focusing on airborne transmission, short-range transmission is primarily driven by droplets with particle sizes roughly larger than 5 μm. In contrast, long-range transmission is driven by aerosols, which tend to travel longer distances due to their smaller particle size. Based on this information, we interpret the mitigation of infection risks in three aspects: (1) environmental stress on the pathogen, (2) short-range transmission, and (3) long-range transmission. To keep the functional analysis concise, we select three second-level functions from Table 2 that are common for infection control: (1) supply outdoor air, (2) disinfect bioaerosols, and (3) block dissemination within the same space.

Table 3. Effectiveness of second-level functions in view of infection risk mitigation.

	Environmental Stress on the Pathogen	Short-Range Transmission	Long-Range Transmission
Supply outdoor air	No	No	Yes
Disinfect bioaerosols	Yes	No	Yes
Block the dissemination in the same space	No	Yes	Yes

summarizes the effectiveness of each second-level function in addressing the three aspects of infection risks, which are elaborated as follows:

(1)

Since environmental stress on the pathogen directly neutralizes the pathogen, the function “disinfect bioaerosols” is considered the only effective function for this aspect;

(2)

Regarding short-range transmission, the functions “supply outdoor air” and “disinfect bioaerosols” are not often applied in close proximity to the infected/exposed person, making them ineffective in intervening in infection events at short distances. In other words, typical configurations of ventilation and UVGI cannot effectively prevent short-range transmission. An exception may be personalized ventilation (e.g., [81], reviewed in Section 3.1.2), arguably not commonly practiced. While the function “block the dissemination in the same space” can be effective, the effect of space partitioning is not straightforward, as discussed in Section 4.2. The challenge of controlling short-range transmission corresponds to the discussions on challenging issues related to close-contact transmission [15,380] and short-term exposure events [381,382];

(3)

Long-range transmission is associated with the average concentration of bioaerosols in space, and this aspect can be addressed by the functions “supply outdoor air” and “disinfect bioaerosols”. Additionally, the function “block the dissemination in the space” should effectively make bioaerosol concentrations evenly distributed in the space (or avoid some high-concentration zones), making it suitable for addressing long-range transmission.

Mitigating infection risks has evidently remained a long-standing challenge for healthcare and engineering professionals. This sub-section aims to illustrate how functional analysis can assist in clarifying the scopes of engineering controls to address infection control issues. To mitigate these risks effectively, systems design is required to deploy engineering controls for specific aspects (e.g., short-range versus long-range). In the next section, we will discuss some systems design techniques that can help manage system complexity and promote desirable system properties while developing IAQ design solutions.

6.4. Systems Design Techniques for IAQ Engineering Controls

In this section, we will discuss three system properties, complexity, flexibility, and resilience, in the context of IAQ engineering controls. System complexity relates to the interconnectivity of engineering controls and building performance. For example, increasing the ventilation rate (e.g., for infection control) can affect thermal comfort or introduce another IAQ issue when outdoor air quality is poor. One technique to mitigate system complexity is functional decoupling [537]. The basic idea is to reduce interconnectivity using one primary engineering control (or solution) to deliver one function approximately. Figure 6a,b illustrate the difference between coupled and decoupled systems in view of the mapping between functions and controls.

One example of coupled systems is traditional all-air systems, which must fulfill two functions: heating/cooling and ventilation. In functional decoupling, dedicated outdoor air systems (DOAS) can be used to provide the ventilation function, while the heating/cooling function can be addressed by other controls (e.g., radiant panels [538]). As studied in [539,540], the decoupling strategy allows designers to adjust operating parameters more flexibly to meet various functional requirements (e.g., cooling, humidity control, and ventilation in their cases) while benefiting from energy savings.

System flexibility can be understood as the ability of a system to adapt to various situations. In the context of IAQ, flexibility is a desirable property due to the dynamic changes in weather and occupancy conditions. It is worth noting that functional decoupling can support system flexibility since designers can apply independent controls for new functional requirements. Additionally, we can consider two approaches to achieve system flexibility: intentional duplication and information technology.

In intentional duplication, we purposefully design more than one control to serve a single function, as illustrated in Figure 6c. Hybrid ventilation (Section 3.1.4) is one example where the ventilation function can be achieved through a range of active (mechanical ventilation) to passive (natural ventilation) strategies, and the system can choose which one to apply according to the specific situation.

The review of information technology (Section 5) has illustrated the emerging trend of low-cost IAQ sensors, which can help detect dynamic and real-time conditions for flexible actions. The system can also analyze Internet and real-time data for predictive control (e.g., [119]), where intentional duplication can provide more available and controllable actions.

Resilience can be defined as a system’s ability to recover its functions after a major disruption [541] (p. 89). In the context of IAQ controls, resilience can be interpreted in two aspects. First, some engineering controls can be disrupted or become unavailable while the need to maintain IAQ standards persists. An example is occasional high levels of outdoor air pollutants (e.g., wildfire smoke and ozone season) when outdoor air cannot be used directly for ventilation. In the second aspect, the original IAQ standard may need to be exceeded beyond the design conditions using existing engineering controls. An example can be found in the pandemic conditions when occupants wish to reduce bioaerosol concentrations further using the existing system. It is worth noting that resilience and flexibility are related in the sense that a resilient system should also be flexible in responding to disruptions.

Importantly, technical limitations of engineering controls (e.g., filter efficiency) may not be the restricting factor in achieving resilience. Instead, when disruptions occur, there is typically very little time available to coordinate existing engineering controls and develop mitigating solutions. Thus, one strategy to achieve resilience is to plan for uncommon events (e.g., wildfire smoke and influenza season) and develop different modes of IAQ operations (e.g., the pandemic mode of operations), as illustrated in Figure 6d.

7. Conclusions

In this paper, we have reviewed engineering controls for IAQ from a systems design perspective. One major proposition is to distinguish between the functions of indoor air treatments (reviewed in Section 3) and dissemination control strategies (reviewed in Section 4). This distinction helps us analyze IAQ issues (e.g., concerns regarding the cleanliness of supply air and/or close-contact transmission) and identify relevant engineering controls. The review of sensors and information technology (Section 5) shows that low-cost IAQ sensors and the Internet of Things can facilitate the development of smart systems to support IAQ. To promote systems thinking for IAQ, we have conducted a functional analysis (Section 6.2) for the engineering controls reviewed in this paper. Additionally, we have suggested some systems design ideas for IAQ (Section 6.4), including functional decoupling and design for flexibility and resilience.

Based on this review, three directions for systems design can be suggested for IAQ research. First, as discussed in [430,431], we can encourage greater integration of traditional IAQ controls (e.g., ventilation and zone pressure gradients) with information technology. While smart ventilation is one existing example, the concept of smart features can also be considered for dissemination control.

Second, flexible IAQ systems are crucial to addressing dynamic changes in environmental conditions and occupants’ needs. We should explore and develop more creative approaches that utilize and coordinate existing engineering controls to achieve flexibility for traditional IAQ functions. One example is hybrid ventilation, which can flexibly utilize natural ventilation to improve IAQ.

Third, systems design can incorporate considerations of uncommon events (e.g., pandemics) and develop corresponding operational modes and strategies for IAQ. In this way, switching operational modes can demonstrate the system’s resilience when uncommon events occur.