Scientific Background Information
ATMOSPHERIC CHEMISTRY AND SECONDARY PARTICLE PRODUCTION
Sulfate Aerosol Chemistry
Sulfate (SO42-) airborne particles are predominantly composed of sulfuric acid (H2SO4) and its salts, NH4HSO4, (NH4)2SO4, etc. Only a small fraction of the particles are emitted directly as particulate matter from the sources; most are produced in the atmosphere by the oxidation of sulfur-containing gases emitted into the air by both natural and anthropogenic sources. Sulfur dioxide (SO2) is by far the most important gas precursor to SO42- in the atmosphere over North America. The compounds dimethyl sulfide (CH3SCH3), hydrogen sulfide (H2S), dimethyl disulfide (CH3SSCH3), and related species are derived from natural sources and undergo rapid oxidation in the troposphere to provide a small background source of SO2, which amounts to about 6% of the SO2 found over North America (Placet and Streets, 1987). However, the major sources of SO2 are anthropogenic. Most important is the combustion of sulfur-containing fuels (primarily coal) by electric utility boilers, which account for almost three-fourths of SO2 emissions. Emissions from smelters and other combustion sources account for most of the remainder.
Gas-phase and liquid-phase oxidation processes are important in sulfate production. The major gas-phase oxidation pathway involves oxidation by the hydroxyl radical (HO), although other transient species can
also contribute (Calvert and Stockwell, 1984). Ozone (O3) is a major source of the HO radical through its photodecomposition to the highly reactive, excited oxygen atom, O(1D), which reacts with water vapor:
The major elementary reactions that occur in the homogeneous atmospheric oxidation of SO2 are well established (Stockwell and Calvert, 1983; Calvert and Stockwell, 1984; Margitan, 1984; Calvert et al., 1985; Gleason et al., 1987). The first of those, Reaction A-3, involves the HO radical formed in Reaction A-2:
Reactions A-4 and A-5 occur rapidly following the rate-determining step (A-3). When coupled with Reaction A-6, Reactions A-3 and A-4 constitute the elements of a HO-HO2 radical chain propagation sequence; that is, although an HO radical is removed in Reaction A-3, another is regenerated in Reaction A-6:
H2SO4 formed in Reaction A-5 condenses on airborne particles because of its low vapor pressure. Field data and theoretical analyses show that secondary SO42- produced by gas-phase reactions tend to accumulate on particles smaller than about 0.3 µm in diameter (Wilson and McMurry, 1981; Hering and Friedlander, 1982; McMurry and Wilson, 1983; John et al., 1990).
SO42- particles also can be produced by the oxidation of dissolved SO2 (largely present as bisulfite ion, HSO3-1) within cloud or fog droplets (Penkett et al., 1979; Martin, 1984; Schwartz, 1984; Hoffmann and Jacob, 1984; Calvert et al., 1985). The major oxidants, H2O2 and O3, are the secondary products formed in gas-phase reactions initiated by the irradiation of air masses containing NO, NO2, and hydrocarbons. Those compounds, as well as molecular oxygen in the presence of certain transition metal catalysts, can oxidize bisulfite to sulfuric acid and ultimately to the various partially neutralized species (principally ammonium bisulfate) that are the common aerosol SO42- (Hoffmann and Jacob, 1984).
The formation of sulfates by liquid-phase reactions occurs in several steps. First, airborne particles (cloud condensation nuclei) are activated to produce much larger (1-50 µm) fog or cloud droplets (Pruppacher and Klett, 1978). The SO2 and oxidant then dissolve in the droplet and react to produce SO42-. When the droplets evaporate, the resulting airborne particle is larger than the original particle due to the addition of the new SO42-.
Recent work has shown that two kinds of SO42- particle size distributions are found in the accumulation mode (Hering and Friedlander, 1982; McMurry and Wilson, 1983; John et al, 1990). Under dry conditions, when gas-phase reactions are likely to dominate, secondary SO42tend to accumulate in particles smaller than about 0.3 µm. In contrast, under humid conditions, when aerosol processing by clouds or fog is likely, SO42- particles tend to range in size from 0.5 to 1.0 µm. The investigators also have shown theoretically that SO42- produced by gasphase reactions accumulate in particles smaller than those produced by liquid-phase reactions. Thus, both theory and atmospheric observations show a close coupling between SO42- particle size distributions and the chemical mechanism of sulfur oxidation.
Nitrate Aerosol Chemistry
Nitrate airborne particles also can be produced by several mechanisms. One major mechanism of nitrate formation involves the gasphase reaction of NO2 with HO to produce nitric acid (Calvert et al., 1985):
Nitric acid also can be formed in heterogeneous chemistry, which occurs largely at night. That involves the reaction of gaseous dinitrogen pentoxide (N2O5) with aqueous aerosols:
About one in every 10–20 collisions between an aqueous aerosol and the N2O5 gas-phase molecule results in nitric acid generation (Mozurkewich and Calvert, 1988). N2O5 is formed through the reaction involving NO2 and O3:
During daylight hours the sequence of Reactions A-9, A-10, and A-8 is relatively unimportant since the transient NO3 radical is kept at very Iow concentrations through its decomposition by sunlight (Magnotta and Johnston, 1980).
The gaseous nitric acid can react at basic airborne particle surfaces to form nitrate salts (Seinfeld, 1986). For example, a particle containing calcium carbonate, can neutralize the nitric acid to produce calcium nitrate. It follows that nitrate size distributions depend, in part, on the size distributions of the particles on which they react. Reactions with
sea salt or soil dust material tend to produce coarse particle nitrates (particle diameter > 2 µm).
Ammonium Aerosol Chemistry
Gaseous ammonia (NH3), is the principal gas-phase neutralizing species in the atmosphere, and, as such, its distribution in the particle phase is closely linked to that of the predominant acid aerosol component, SO42-. Consequently, most of the particulate NH4+ is found in the submicron-particle-size fraction. Ammonium or acid SO42- are believed to be the most abundant inorganic components of cloud condensation nuclei. Nonetheless, the acidic aerosol components are seldom fully neutralized by NH4+, even in relatively remote areas; consequently, stoichiometric ratios that approximate NH4HSO4 are common.
When the levels of NH3 and HNO3 are sufficiently high (e.g., each above about 1 ppbv at 298 K), ammonium nitrate (NH4NO3) can be formed. Ammonium nitrate is often found in submicron particles in locations such as Denver or Los Angeles (Wall et al., 1988; Pierson and Brachaczek, 1988). For example, John et al. (1990) have shown that particulate nitrate measured in Los Angeles accumulates in three distinct modes with mass mean aerodynamic diameters of 0.2 ± 0.1 µm, 0.7 ±0.2 µm, and 5 ±1 µm. However, little submicron ammonium nitrate is typically found in continental air, where ammonia concentrations are iow and acid sulfate concentrations are high. When ammonium nitrate reacts with the acidic particles, such as sulfuric acid, the ammonium is retained and the nitric acid is released. The nitric acid tends to react with and be retained by coarse particle cations, including calcium and sodium (Wolff, 1984).
Organic Aerosol Chemistry
Secondary organic particle formation is poorly understood in comparison to the formation of SO42- and nitrates. There are several reasons for that. First, the problem is far more complex, involving many organic gas-phase particle precursors, each with a variety of possible particle formation pathways and products. Little is known about the contribu-
tions of primary and secondary organics and of anthropogenic and natural sources to ambient loadings, especially for remote areas.
Second, methods for sampling and analysis of organic gases and particulate matter are less well developed than those for nitrates and SO42-. The sampling of organic materials can be very difficult; organic particles that are collected on filters can lose volatile components during the collection period, and reactive particulate organics can undergo chemical changes after collection on filters, in flasks, or on adsorption media. Problems are especially severe in pristine areas where Iow concentrations lead to large relative errors because of sampling artifacts.
The suggested pathways for secondary organic particle formation primarily involve gas-phase reactions with ozone, the hydroxyl radical (OH), and the nitrate radical (NO3). Grosjean and Seinfeld (1989) have summarized what is known about the potential for airborne particle formation of organic gases. They show that alkenes and aromatic hydrocarbons tend to be effective particle producers in photochemical systems, but alkanes and carbonyl compounds are relatively ineffective. Within any given chemical class, the particle-formation potential tends to increase with increasing molecular weight and increasing number of polar groups (such as nitrate, hydroxyl, etc.), because those products are less volatile.
The contributions of terpenes and isoprene emitted by plants to secondary organic particles are uncertain. However, it is important to attempt to judge the extent of that contribution, as it may contribute to the background particle loadings and to light scattering in national parks and wilderness areas. The particle-formation potential of the monoterpenes was recognized as early as 1960 (Went, 1960). More recent studies by Kamens et al. (1981, 1982), Hatakeyama et al. (1989), and Pandis et al. (1991) show that a significant fraction of the α-and ß-pinenes, which react with O3 or HO radicals in the atmosphere, can lead to organic particle formation, but isoprene forms very little particulate matter under similar conditions (Pandis et al., 1991).
Based on the study of the time dependence of the number and size distribution of organic airborne particles formed in isoprene and ß-pinene photooxidation at Iow, ambient concentrations, estimates were made by Pandis et al. (1991) to show that the potential organic particulate formed from the monoterpenes could be significant in three types of environments:
In poorly ventilated urban areas such as Los Angeles, which has extensive urban landscaping and brush-covered hills, natural hydrocarbons might be responsible for up to 50% of secondary organic particles.
In urban areas such as Atlanta, which has extensive wooded areas, natural sources could produce approximately 27 Mg of organic airborne particles per day, probably an order of magnitude greater than the anthropogenic secondary organic particles formed.
In highly wooded areas, typical of many national parks and wilderness areas, 2,000 µg/m2-hr of reactive organic hydrocarbons are emitted throughout the day (Hov et al., 1983). Pandis et al. (1991) have estimated that stagnant conditions in those areas could lead to organic particle concentrations of 30–39 µgC/m3 on the second or third night of stagnant conditions.
Although such estimates suggest that secondary organic compounds can contribute significantly to visibility impairment, there is little experimental evidence that permits quantitative estimates of the relative contributions of primary and secondary compounds to organic particle formation on a national scale. Gray et al. (1986) found that, on a long-term average, organic carbon particles in Los Angeles originate largely from primary emissions, even during the summer when there is extensive secondary particle formation by photochemical reactions. However, during smog episodes, secondary organic particles can be the largest contributor to secondary organic levels. Turpin and Huntzicker (1991) have reported that secondary particles constitute about 70% of the Los Angeles organic particle fraction during the afternoon on smoggy summer days. Similar studies have not been done in Class I areas.
Because organic carbon often contributes significantly to optical extinction in such areas (e.g., Ouimette and Flagan, 1982), present understanding of its primary and secondary origins should be improved. That will require improvements in the sampling and analysis of carbon-containing particulate constituents (see Appendix B) as well as a more satisfactory understanding of atmospheric processes.
In some areas, visibility degradation often is associated with the trans-
port of plumes, both individually and collectively as a polluted air mass, from specific source regions. The magnitude of the visibility degradation is controlled in part by the degree of dispersion of plumes. To assess the sources of pollutants that affect a region and to predict the degree of visibility degradation that might be expected, accurate measurements of a number of atmospheric characteristics are needed.
Atmospheric transport can be estimated by (1) interpolating wind data obtained from meteorological stations; (2) extrapolating the measured wind field by means of a system of hydrodynamic equations, incorporating the appropriate initial and boundary conditions; or (3) using physical or chemical tracers characteristic of a specific source (or types of sources). The transport of visibility-reducing plumes often occurs within a layer located a few hundred meters or more above the surface of the earth. Consequently, wind data are required not only at the surface but also throughout the transport layer. The following sections outline the measurements needed to estimate plume transport through the interpolation of observations.
Surface-wind measurements are the easiest to obtain and can be useful in determining plume transport. Surface winds are most likely to be representative of the flow of air aloft in the afternoon when thermally driven convection couples the surface wind to the winds aloft. Normally, in pollution dispersion studies, unmanned meteorological stations are used to measure surface winds. The stations are set up in a relatively dense network. Wind data from the network are integrated with those obtained from existing weather service stations (such as those commonly located at airports), which provide information on a larger scale. Remote systems, such as the NCAR PAM stations, automatically collect information on wind speed and direction, humidity, and solar radiation. The data are periodically transmitted to satellites for relay to a primary data collection site. The stations can be powered by solar cells for use at remote locations.
Estimates of atmospheric transport based solely on surface winds are subject to interpolation errors in the horizontal direction and to extrapolation errors in the vertical direction. Horizontal interpolation errors are
particularly sensitive to topographic features; such errors are a major problem in visibility studies performed in national parks and wilderness areas, because those areas are often located in mountainous regions. Consequently, surface-wind networks should be regarded as a necessary, but not a sufficient, component in studies that evaluate pollutant transport.
Balloon studies are commonly used to measure winds above the surface. Measurements are made with tethered balloons, free-rise balloons, and constant altitude balloons. Tethered balloons, with attached tethersondes, can be used to measure winds, temperature, and humidity up to altitudes of about 2 km during light wind conditions; however, these balloons become unstable at wind speeds above roughly 8 m/sec. Free-rise balloons, including instrumented units (e.g., rawinsondes) and those that are tracked optically (pibals), can be flown in all weather conditions.
Constant altitude balloons (tetroons) ride at a fixed altitude and are used to simulate the trajectory of an air mass. Tetroons can deviate from their preassigned altitude, particularly during convection, thereby reducing their being representative of atmospheric transport. The accuracy of tetroons is also difficult to verify, as there is no independent measure of transport.
Although balloons are generally less expensive than other techniques, they are more labor-intensive. Another shortcoming of balloons is that they provide little information about the magnitude of atmospheric turbulence, an important factor in pollution dispersion.
Instrumented towers are used to obtain information about atmospheric transport and turbulence. The measurements are usually limited to heights of a few tens of meters; therefore, tower winds generally are not representative of the winds throughout the plume layer. However,
towers can furnish data on the variation in wind direction and speed with height; those data provide a measure of the dispersing capacity of the atmosphere.
Radar and lidar techniques can be used to provide data for wind measurements at rates that are orders-of-magnitude smaller than those available from operational measurement systems. Given the need for wind fields with high temporal resolution, particularly in areas of complex terrain, this technology is promising. However, it can only measure winds at one location, so spatial interpolation is still necessary. Also, the technique is expensive and requires a technically well-trained staff. With the continuing advances in computers and hardware, however, the cost of remote sensing systems will inevitably decrease.
Ideally, a meteorological measurement program would include the use of remote sensing to measure temporal variations of winds above the surface. Remote sensing would be complimented by balloons, either tethered or free-rise, depending on the problem being studied.
The extinction of light by hygroscopic airborne particles generally increases with increasing relative humidity because of the increase in particle size. Consequently, detailed information about humidity must be obtained during field programs. The relative humidity (or the dew point temperature) should be measured at all surface meteorological stations and in vertical profiles from balloons as well. Such measurements are usually inexpensive and can provide information for testing the ability of meteorological models to reproduce the vertical and spatial extent of water vapor transport; that information also is needed for modeling the wet removal of pollutants.