Abstract
Air pollution is due to emissions of pollutants in the atmosphere, which may be natural or of human origin. Thus, in order to understand air pollution, it is necessary to identify, characterize, and quantify those emissions. Furthermore, reducing air pollution requires either eliminating some of those emissions via a change in a product, process, or technology, or reducing those emissions using some control technologies. This chapter describes the main sources of air pollution and the technologies available to control those emissions. First, air pollutant sources are described. Next, the methods used to quantify the corresponding emissions and develop air pollutant emission inventories are presented. Finally, the main technologies used to control emissions of gaseous and particulate air pollutants are described.
Air pollution is due to emissions of pollutants in the atmosphere, which may be natural or of human origin. Thus, in order to understand air pollution, it is necessary to identify, characterize, and quantify those emissions. Furthermore, reducing air pollution requires either eliminating some of those emissions via a change in a product, process, or technology, or reducing those emissions using some control technologies. This chapter describes the main sources of air pollution and the technologies available to control those emissions. First, air pollutant sources are described. Next, the methods used to quantify the corresponding emissions and develop air pollutant emission inventories are presented. Finally, the main technologies used to control emissions of gaseous and particulate air pollutants are described.
2.1 Sources of Air Pollution
First, it is useful to recall the definitions of primary and secondary pollutants. A primary pollutant is a pollutant that is emitted directly in the atmosphere. A secondary pollutant is formed in the atmosphere via chemical reactions among other chemical species, which are called precursors. Some precursors may also be primary pollutants, and a chemical species may be both a primary pollutant and a secondary pollutant. Therefore, to understand air pollution, one must know not only the emissions of primary pollutants, but also those of precursors of secondary pollutants. Generally, precursors of secondary pollutants are considered to be an integral part of air pollution and are called air pollutants. It is the case, for example, in the United States (CFR, 2016) and in France (Code de l’Environnement, 2016). Therefore, this text will include both primary pollutants and precursors of secondary pollutants as air pollutant emissions.
Air pollutants may be emitted from anthropogenic sources (i.e., those sources related to human activities) and/or from natural sources. Examples of anthropogenic sources include transportation (on-road, rail, air, maritime, etc.), industry (fossil-fuel fired power plants, smelters, incinerators, refineries, etc.), agriculture (cattle, fertilizer use, etc.), and the residential, commercial, and institutional sector (heating, cleaning products, etc.). Examples of natural sources include emissions of volatile organic compounds (VOC) from vegetation and nitrogen compounds from soils, dust emissions due to wind erosion, ocean emissions, volcanic eruptions, geothermal sources, lightning (production of nitrogen oxides), and forest fires (however, those may also be due to human activities).
Four major categories of processes lead to air pollutant emissions: combustion, volatilization, mechanical processes (abrasion, resuspension, etc.), and natural processes that do not belong to one of the previous categories.
2.1.1 Combustion
Combustion may be the result of an anthropogenic activity (e.g., transportation, production of electricity, incineration of waste, heating) or a natural process (forest wildfires). Combustion leads to the production of heat, which can then be converted if needed into another form of energy (e.g., electrical, mechanical). The combustion process implies the presence of oxygen, which is available from the air, and carbon, which is the main component of fuels, such as coal, gasoline, diesel, and wood. This combustion occurs at high temperatures and leads to (1) the dissociation of oxygen (O2) and nitrogen (N2) molecules, both of which are present in the air, and (2) the oxidation of carbon. The dissociation of the O2 and N2 molecules leads to oxygen (O) and nitrogen (N) atoms, respectively. Then, the reactions among oxygen (O2, O) and nitrogen (N2, N) lead to the formation of nitrogen oxides (NOx), mostly nitric oxide (NO), but also a fraction (<10 %) of nitrogen dioxide (NO2). The complete oxidation of the fuel leads to carbon dioxide (CO2), a greenhouse gas, and to water vapor (H2O). However, combustion is generally not complete and carbonaceous compounds that are not completely oxidized are produced during combustion. Such compounds include, for example, carbon monoxide (CO), volatile organic compounds (VOC), soot particles (originating mostly from diesel engines and biomass fires), polycyclic aromatic hydrocarbons (PAH), and dioxins and furans. Some of those compounds are pollutants and some may even be carcinogenic (e.g., formaldehyde, soot particles from diesel engines, some PAH, dioxins, and furans). In addition, inorganic substances present in the fuel are released during combustion, often in their oxidized form due to chemical reactions occurring at high temperatures. Among substances present in coal, gasoline, and diesel, one may mention sulfur and mercury.
2.1.2 Volatilization
The volatilization of semi-volatile compounds consists in their transfer from a liquid phase to a gas phase, which may then be dispersed in the atmosphere. Volatilization affects, for example, hydrocarbons (e.g., oil, gasoline) during their storage and transfer and paints and solvents during their use. It also affects fuels contained in vehicles, and this volatilization process can contribute to a significant fraction of VOC emissions from vehicles when the ambient temperature is high or even moderate. Volatilization varies depending on the nature of the fuel, because it is a function of the physico-chemical properties of the hydrocarbons present in the fuel. Gasoline includes linear and branched alkanes (20 to 30 %), cycloalkanes (~5 %), alkenes (30 to 45 %), and aromatic compounds (30 to 45 %), as well as additives (such as ethanol). Laboratory chemical analyses lead to an average molecular formula for gasoline that is close to that of heptane (C7H16). Diesel includes mostly alkanes (linear, branched, and cyclic; >75 %) and aromatic compounds (<25 %). A theoretical average molecular formula of C16H29 may be used as representative of the ensemble of hydrocarbons present in diesel. (Note that although octane, C8H18, and cetane, C16H34, indices are used to characterize gasoline and diesel, respectively, these hydrocarbons represent only a small fraction of all the hydrocarbons present in these fuels and do not correspond to the average formula of these fuels.) Therefore, diesel is a fuel that includes VOC that are heavier, and therefore less volatile, than those of gasoline. As a result, the volatilization of VOC from vehicles pertains mostly to gasoline vehicles. Mercury may be emitted naturally from soils and oceans as elemental mercury, a form that is very volatile. In addition, reemission of semi-volatile pollutants (for example, some persistent organic pollutants, POP, such as PAH and pesticides) is a volatilization process.
2.1.3 Mechanical Processes
Among the mechanical processes leading to atmospheric emissions, one may mention anthropogenic activities such as construction activities, farming, and some industrial activities, as well as natural activities such as the emission of wind-blown dust and sea salt (aeolian emissions). Also, transportation is an important source of particles via processes such as braking (abrasion of the brake pads), driving (wear of tires, roads, metal wheels, railways, etc.), and the resuspension by traffic of particles present on roads.
2.1.4 Natural Processes
Natural processes other than the ones already mentioned include, for example, the metabolism of vegetation, which leads to the atmospheric emissions of VOC, and emissions associated with volcanic eruptions.
2.1.5 Summary of Global Emissions of Air Pollutants
Table 2.1 summarizes the global emissions of several major air pollutants: sulfur dioxide (SO2), nitrogen oxides (NOx), ammonia (NH3), carbon monoxide (CO), volatile organic compounds (VOC), and particles. Particles are represented here by PM10 (particles with an aerodynamic diameter less than 10 μm) and PM2.5 (particles with an aerodynamic diameter less than 2.5 μm, i.e., fine particles). Anthropogenic emissions were obtained from the Emission Database for Global Atmospheric Research (EDGAR) developed by the European Commission (EC, 2016) and are for the year 2010. Biomass fires may be from anthropogenic or natural origin and, accordingly, they are listed separately (Andreae and Marlet, 2001).
Table 2.1. Annual global emissions of selected major air pollutants (Tg/year). Data sources: EC (2016); Andreae and Marlet (2001); Bates et al. (1992); Bouwman et al. (1997); Gong et al. (2002); Guenther et al. (2012); Khalil and Rasmussen (1990); Logan (1983). (1 Tg = 1012 g.)
| Sources | SO2 | NOxa | NH3 | CO | VOCb | PM10 | PM2.5 |
|---|---|---|---|---|---|---|---|
| Electric power and heat production | 49 | 31 | 0.1 | 6.5 | 0.7 | 4.9 | 3 |
| Oil refineries | 1.7 | 0.9 | 0 | 0.5 | 2.1 | <0.1 | <0.1 |
| Other industrial sources | 31 | 19.2 | 1.7 | 137.9 | 63.4 | 16.3 | 10.1 |
| Waste and wastewater treatment | 0.1 | 0.2 | 0.1 | 0.1 | 2.6 | 0.2 | 0.1 |
| Residential, commercial, and institutional sector | 8.2 | 6.2 | 5.1 | 232.5 | 36 | 30.2 | 18.2 |
| Agriculture | 0.4 | 6 | 47 | 75 | 4.4 | 10 | 8.3 |
| On-road transportation | 0.8 | 27 | 0.5 | 170 | 25.4 | 0.9 | 0.9 |
| Aviation | 0.3 | 2.9 | 0 | 0.5 | 0.1 | <0.1 | <0.1 |
| Maritime shipping | 11 | 17 | 0 | 5.3 | 1.2 | 1.9 | 1.8 |
| Other modes of transportation | 0.1 | 1.9 | 0 | 0.6 | 0.2 | 0.3 | 0.3 |
| Sub-total of anthropogenic sources | 103 | 112 | 54 | 629 | 136 | 65 | 43 |
| Biomass firesc | 3 | 25 | 6 | 413 | 251 | 49 | 36 |
| Natural sources | 19 | 53 | 16 | 140 | 1000 | 1690 | 460 |
| Total | 125 | 190 | 76 | 1182 | 1387 | 1804 | 539 |
(a) NOx emissions are expressed as NO2.
(b) VOC except methane, which is not chemically very reactive
(c) Except residential heating; biomass fires are mostly from anthropogenic activities, but there is also a natural contribution.
Natural sources vary depending on the pollutants. Natural emissions of SO2 result mostly from volcanic activities (degasing and eruptions) and they vary significantly from one year to the next depending on eruptions from active volcanoes (Bates et al., 1992). Natural emissions of NOx result from lightning and from soils, in similar proportions (Logan, 1983). Natural emission sources of ammonia include mostly the oceans, soils associated with natural vegetation, and the human population (Bouwman et al., 1997). Natural emissions of CO originate from vegetation and from the oceans (Khalil and Rasmussen, 1990). Natural emissions of VOC result from vegetation. They include isoprene for about 50 % (90 % from deciduous trees), monoterpenes for about 15 % (>80 % from deciduous trees), sesquiterpenes for about 3 % (from deciduous trees and evergreens), 2-methyl-3-buten-2-ol (MBO) for about 0.2 % (from evergreens), and other VOC, such as alcohols, aldehydes, ketones, alkenes, and carboxylic acids for the remainder (Guenther et al., 2012). Natural emissions of particles include aeolian soil erosion (mostly from deserts; Zender et al., 2003) and sea-salt emissions (Gong et al., 2002). Regarding sea salt, only fine particles are included here, because coarse particles have a lifetime of only a few hours and, therefore, do not have any long-range impacts. Volcanic eruptions are also a source of particles; this source is not included here because of its large interannual variability.
These global inventories do not include atmospheric chemical reactions. Some chemical reactions may be an important source for some of those pollutants, as described in Chapters 8, 9, and 10. For example, dimethyl sulfide (DMS) and hydrogen sulfide (H2S) are oxidized into SO2, VOC are oxidized and eventually form CO, and an important fraction of fine particles (PM2.5) is formed in the atmosphere via chemical reactions that involve SO2, NO2, VOC, and NH3.
Note that natural sources dominate the global inventory of VOC and particulate matter (PM10 and PM2.5). However, in the case of VOC, chemical speciation is essential, because VOC differ significantly in terms of chemical reactivity and toxicity. In the case of particles, they are regulated in terms of mass rather their chemical composition; nevertheless, their health impacts could depend on their chemical composition. Furthermore, natural sources are distributed widely over the globe, whereas anthropogenic sources are generally concentrated in or near areas where people live.
These global emissions have been evaluated for some pollutants using satellite data and inverse modeling. Such evaluations have been performed, for example, for SO2 emissions (Lee et al., 2011) and CO emissions (Kopacz et al., 2010). In both cases, the bottom-up global emission inventories were found to be consistent with the top-down emission estimates obtained from satellite data.
2.2 Emission Inventories
Emission inventories are needed to track the temporal evolution of air pollutant emissions. For example, countries in Europe and states in the United States must report their emissions on a regular basis to the European Union and the federal government, respectively. Greenhouse gas emissions may also be included in such emission reporting. In addition, numerical modeling of air pollution, which is conducted for air quality impact assessments, emission scenario simulations, and air quality forecasting, requires spatially distributed and temporally resolved emission inventories. Methods that are used to develop emission inventories are briefly described in this section.
The fundamental equation for the quantification of most air pollutant emissions is as follows:
where Sij is the rate of emission of air pollutant i from source j (in g s−1), EFij is the emission factor for air pollutant i emitted from a source category corresponding to source j (in g per activity unit), and Aj is the activity of source j (in activity units per second).
The activity of a source is defined in different ways depending on the source type. For example, it may be defined in terms of vehicle km per hour for on-road traffic, energy production per unit time (for example, MW) for power plants, and the amount of fuel used per year for residential heating.
Emission factors are expressed in units that are consistent with the unit of the corresponding activity. They may be obtained in several ways. For some sources, emission measurements may be performed at the source. The emission factor obtained for a specific source may then be used more generally for the source category (i.e., for other similar sources). For example, vehicle (or engine) emission measurements are performed on a dynamometer to obtain emission factors for on-road traffic (see Section 2.3.4). For some pollutants, a mass balance may be performed on the emission process. For example, in the case of sulfur present in a fuel (e.g., coal, gasoline, diesel), the sulfur content of the fuel can be used to estimate the emission of sulfur compounds (mostly sulfur dioxide and sulfuric acid), since the sulfur mass is conserved during the combustion process. In a few cases, a simulation of the process may be performed to obtain the chemical speciation of some pollutants (for example, the relative fractions of elemental and oxidized mercury emitted from coal-fired power plants).
For some emissions, the process may be more complex (it may depend, for example, on meteorology) and a parameterization must then be used. This is the case for biogenic VOC emissions from vegetation, which depend on ambient temperature and solar radiation, for VOC volatilization from gasoline vehicles, which depends on ambient temperature, and for wind-blown dust emissions from desert areas and sea-salt emissions from oceans, which depend on wind speed. Models have been developed to estimate those emissions as a function of meteorology. Some models used to estimate VOC emissions from vegetation are mentioned at the end of this chapter. Some models used to estimate aeolian emissions are presented in Chapter 11.
The development of an emission inventory typically requires some method to organize the various source categories. Examples are provided here for the United States and France.
In the United States, the National Emissions Inventory (NEI) is developed by the U.S. Environmental Protection Agency (EPA) from data reported by the states. The NEI uses several types of codes to classify source categories, industrial facilities, geographical regions, pollutants, and emission control equipment. Sources are classified according to Source Classification Codes (SCC). An SCC is specific to an item of equipment, an operation, or a practice that is a source of air pollutants. These codes include eight digits for large point sources (such as power plant stacks) and ten digits for other sources. The North American Industry Classification System (NAICS) is used to identify the primary activity of an industrial facility. The Federal Information Processing Standards (FIPS), state and county codes, and tribal codes are used to identify the state, county, territory, or tribe area where the source is located. Seven-digit numerical codes are used to identify specific pollutants. Finally, three-digit codes characterize the type of emission control equipment used on a specific source. Other codes are used to identify the emission calculation method and the type of reporting period (e.g., seasonal, annual). The use of such codes facilitates the retrieval of specific information on the method and data associated with the development of the emission inventory. More information on the U.S. EPA emission inventory system is available at www.epa.gov/air-emissions-inventories.
In France, CITEPA (“Centre interprofessionnel technique d’études de la pollution atmosphérique”) is the organization responsible for the development of the national emission inventories for the French ministry in charge of the environment. CITEPA uses the SNAP 97 c (Selected nomenclature for air pollution 1997, corrected version) classification for source activities and NAPFUE 94 c (Nomenclature for air pollution and fuels 1994, corrected version) for fuels. SECTEN (“Secteurs économiques et énergie”) and SNAP 97 c are generally used in France for the emission inventory output formats. Other formats, such as NFR (Nomenclature for reporting) and CRF (Common reporting format) are occasionally used for international reporting in Europe. Tables have been developed to convert emission inventories from one format to another (e.g., www.citepa.org).
In terms of a geographical coordinate system, several options are available, depending on the need of the user. For example, the Lambert, UTM (Universal Transverse Mercator), and latitude-longitude systems are widely used.
Once the emission rates have been calculated for all the identified sources of air pollution, these emission rates must be distributed spatially and temporally, if they are to be used in a numerical modeling study.
The spatial distribution is performed differently depending on the source type. Typically, sources are grouped as point sources (e.g., large stacks), area sources (lumping sources that are too small to be treated individually, such as residential heating), line sources (representing, for example, major roadways), and volume sources (used, for example, to represent industrial fugitive sources). In a standard air quality simulation model (see Chapter 6 for a discussion of different types of air quality models), emissions are only represented by means of point sources and area sources. However, these emissions are released in three-dimensional grid cells, and the corresponding sources are, therefore, equivalent to volume sources. In an air quality model that provides a multi-scale treatment of air pollution, it is possible to treat emission sources with greater detail, using point sources (e.g., for tall stacks), line sources (e.g., for major roadways), and volume sources (e.g., for fugitive emissions at industrial sites). Similarly, atmospheric dispersion models may treat individual sources of various types (see Chapter 6) and the four categories of sources may then be used.
The locations of point sources are identified exactly. Area sources represent a large amount of small sources, which cannot be identified individually exactly. Therefore, one must use a surrogate variable to distribute spatially the emissions of that source category. For example, population density may be used to treat residential heating so that the emissions can be distributed spatially over a city, a district, or a region. Line and volume sources correspond generally to specific sources (roadways, industrial sites), which can be localized precisely.
The temporal resolution of emissions is generally hourly for air quality simulations. In some cases, emissions are available with some temporal resolution. This is the case, for example, in Europe and North America for some industrial sources (e.g., power plants) that are required to monitor their emissions for some regulated pollutants (e.g., NOx and SO2). However, in most cases, no specific information is available. Then, one must use temporal distribution factors obtained from other databases. These factors may include different temporal scales. For example, in the case of on-road traffic, temporal distribution factors may include daily, weekly, monthly, and seasonal distribution factors. Daily factors may also vary depending on the day of the week.
Air pollutant emissions include generally gases, particulate matter, and greenhouse gases. For some gaseous pollutants, it is necessary to obtain a chemical speciation: this is the case for nitrogen oxides, which must be categorized as nitric oxide and nitrogen dioxide, and for VOC, which must be distributed among a large number of specific organic molecules. For particulate matter, the chemical composition is needed (black carbon, organic matter, sulfate, etc.). In addition, the particle size distribution is essential because (1) the regulations pertain to specific particle size ranges (PM2.5 and PM10, see Chapters 9, 12, and 15) and (2) the dynamics of particles in the atmosphere depends on their size.
Emission factors and methods available to calculate emission rates for major source categories may be obtained from the following list of selected organizations:
– In the United States: AP-42, Compilation of air pollution emission factors (www.epa.gov)
– In Europe: EMEP/EEA air pollutant emission inventory guidebook (www.eea.europa.eu/publications)
– In France: “Organisation et méthodes des inventaires nationaux des émissions atmosphériques en France” (OMINEA), available from CITEPA (www.citepa.org)
While air pollutant emission inventories are generally developed by international, national, or regional organizations, there are some countries and regions for which no emission inventory is available. In such cases, it is possible to develop an emission inventory using information on source activities available locally and emission factors available in the references provided above. The development of an emission inventory for Lebanon, and more specifically for its capital, Beirut, exemplifies such an approach (Waked et al., 2012).
In addition, one should note that some emissions are particularly difficult to estimate accurately. For example, biomass fires are highly variable from one year to the next. The use of satellite data for burning and burned areas may help develop emission inventories pertaining to biomass fires (e.g., Mieville et al., 2010).
2.3 Emission Control Technologies
2.3.1 Gaseous Pollutants
The technologies available to control gaseous pollutant emissions may be summarized according to the following major categories (Flagan and Seinfeld, 1988; Wang et al., 2005):
Absorption in a Liquid
This approach may be used to reduce emissions of air pollutants that are very soluble, generally in water, for example hydrogen chloride (HCl) and hydrogen fluoride (HF). Dissolution in water occurs according to Henry’s law, and the efficiency of this emission control process depends on the solubility of the pollutant in water. In some cases, it is possible to increase this efficiency by displacing the gas/water equilibrium toward the aqueous phase. For example, the gas/water equilibrium of sulfur dioxide (SO2) may be displaced toward the liquid phase by using an alkaline solution (i.e., a basic solution with a pH greater than 7), because SO2 is a weak acid and its dissolution increases with pH (see Chapter 10). The efficiency of the absorption process may also be increased by adding a chemical transformation to displace the transfer of the substance toward the liquid phase (see the section on chemical transformation).
Adsorption on a Solid
This approach is based on the formation of a bond between a gas molecule and the solid surface. This phenomenon includes various processes that may be categorized as follows:
– Adsorption on a non-polar solid, such as activated carbon, which is a carbonaceous substance with high porosity, thereby allowing a large surface area for interaction with the gas phase. This method is used, for example, to reduce the emissions of persistent organic pollutants (POP) and mercury from incinerators.
– Adsorption on a polar solid (alumina, silica …); however, this type of solid will also adsorb water, which can significantly reduce the efficiency of the adsorption process when the emission effluent contains large amounts of water.
– Chemisorption, which corresponds to a chemical reaction of the adsorbed substance with the solid and which may lead to desorption of the (potentially less harmful) reaction product. Heterogeneous catalytic reactions (i.e., chemical reactions taking place on a solid catalyst) may be included in this category. For example, catalytic converters, which use a catalyst to convert carbon monoxide (CO) into carbon dioxide (CO2) and unburned hydrocarbons into CO2 (two-way catalytic converters), may be included in this category. Three-way catalytic converters also convert a fraction of nitrogen oxide emissions (NOx) into molecular nitrogen (N2) (see the description of emission control systems for on-road vehicles in Section 2.3.4).
In all these cases, the solid that becomes laden with the adsorbed pollutant must be disposed of in a safe and environmentally sound manner.
Chemical Transformation
Chemical transformation may be used to form a pollutant that is more easily controlled or to form a product that is not a pollutant (or at least a pollutant that is less harmful than the original pollutant). Two examples may be mentioned: the control of sulfur dioxide (SO2) and nitrogen oxides (NOx) emissions from coal-fired power plants. These emission control technologies have been implemented, for example, to reduce acid deposition in the United States (reduction of emissions of SO2 and NOx, which are precursors of sulfuric and nitric acids, respectively, see Chapters 10 and 13), as well as ozone levels (NOx are precursors of ozone, see Chapter 8).
To reduce SO2 emissions, SO2 may be transformed into sulfate by reaction with calcium carbonate after absorption in a scrubber (i.e., partial dissolution in an aqueous phase):
Calcium sulfite may be oxidized to sulfate:
The oxidation of SO2 by calcium carbonate, in the form of limestone, leads to the formation of calcium sulfate (CaSO4), also called gypsum, which precipitates as a solid. Gypsum must be removed from the scrubber to avoid its clogging. Typically, gypsum precipitation is minimized by maintaining the pH above 6. Gypsum may be sold as building material, if it does not contain too many toxic substances (metals …). The oxidation of SO2 may alternatively be performed with calcium oxide, also called quicklime:
The oxidation of SO2 by quicklime is more efficient than that by calcium carbonate, but quicklime is more expensive. These reactions may take place in a scrubber where droplets are sprayed into the effluent that contains SO2. Calcium sulfate particles are formed following the evaporation of the droplets, and those particles must be captured by filtration. These emission control systems for SO2 are typically called flue gas desulfurization systems (FGD). The efficiency of an FGD is on the order of 75 to 95 % for SO2 emission control. Dry FGD processes are also available; however, they are not as widely used as wet FGD (see Srivastava and Jozewicz, 2001, for a review of FGD technologies).
To reduce NOx emissions, NOx may be reduced to molecular nitrogen (N2), which is the main constituent of the atmosphere, by using ammonia (NH3). The corresponding chemical reactions are the following: