Tropospheric Ozone And Associated Photochemical Oxidants

Since Haagen-Smit and co-workers established in the late 1950s that the key ingredients in the formation of tropospheric ozone are organics, NOx, and sunlight, there has been a great deal of research devoted to developing and applying effective control strategies for ozone and associated photochemical oxidants. Their chemistry is very complex, sufficiently so that accurately predicting the impacts of controlling the precursor VOC and NOx on ozone and other photochemically derived species at a particular location and time is difficult. However, as we shall see in this chapter, research over the past four decades has led to the successful implementation of a number of effective control strategies.

Developing control strategies for ozone is very different than for relatively unreactive species such as CO. in the latter case, the concentrations in air are a direct result of the emissions, and all things being equal, a reduction in emissions is expected to bring about an approximately proportional reduction in concentrations in ambient air. However, because 03 is formed by chemical reactions in air, it does not necessarily respond in a proportional manner to reductions in the precursor emissions. Indeed, as we shall see, one can predict, using urban airshed or simple box models, that under some conditions, ozone levels at a particular location may actually increase when NOx is decreased! This complexity has resulted in many years of discussions about the best control strategies and the issues involved in developing them. For a detailed treatment, the reader is referred to the National Academy of Sciences 1991 document "Rethinking the Ozone Problem in Urban and Regional Air Pollution" and to the U.S. Environmental Protection Agency 1993 document "The Role of Ozone Precursors in Tropospheric Ozone Formation and Control." In addition, there is a major cooperative research effort, NARSTO (./Vorth American .Research Strategies for Tropospheric Ozone), by the United States, Canada, and Mexico to address the scientific understanding of ozone formation and transport in the troposphere and the implications for control strategy development in these three countries. Critical review papers prepared as part of this assessment will be published in Atmospheric Environment in 1999; the NARSTO Web site is listed in Appendix IV.

A number of different approaches have been taken to understanding the VOC-NOx chemistry and its application to control strategy developments. These include the use of environmental chambers, models ranging from simple linear rollback to complex Eulerian models, and field studies.

1. Environmental Chambers

Perhaps the most direct experimental means of examining the relationship between emissions and air quality is to simulate atmospheric conditions using large chambers. Measured concentrations of the primary pollutants are injected into these environmental (or smog) chambers, as they are called. These are then irradiated with sunlight or lamps used to mimic the sun, and the time-concentration profiles of the primary pollutants as well as the resulting secondary pollutants are measured. The primary pollutant concentrations as well as temperature, relative humidity, and so on can be systematically varied to establish the relationship between emissions and air quality, free from the complexities of continuously injected pollutant emissions and meteorology, both of which complicate the interpretation of ambient air data.

The results from such chamber studies are frequently used to test the chemical portion of various computer models for photochemical air pollution in order to provide a scientific basis for control strategies. While the interpretation of the results of smog chamber studies and their extrapolation to atmospheric conditions also have some limitations (vide infra), such studies do provide a highly useful means of initially examining the emissions-air quality relationship under controlled conditions.

In addition, smog chambers are useful as large reaction vessels to generate kinetic and mechanistic data on individual reactions believed to be important in the atmosphere. They have also been used extensively in the past as exposure chambers to study the effects of air pollution on people, animals, and plants.

a. Types of Chambers

Design criteria for these chambers are based on reproducing as faithfully as possible conditions in ambient air, excluding meteorology and the uncontrolled addition of pollutants. Although the general aims of all chamber studies are similar (i.e., to simulate reactions in ambient air under controlled conditions), the chamber designs and capabilities used to meet this goal vary widely. Thus chambers can differ in any or all of the following characteristics: (1) size and shape, (2) surface materials to which the pollutants are exposed, (3) range of pressures and temperatures that can be attained, (4) methods of preparation of reactants, including "clean air," (5) conditions (i.e., static or flow) under which experiments can be carried out, (6) analytical capabilities, and (7) spectral characteristics of the light source.

In discussing these chambers, we need to keep in mind one major caveat. The use of chambers by necessity involves the presence of surfaces in the form of the chamber walls, and this is the single largest uncertainty in using them as a surrogate for ambient air studies. Contributing to this uncertainty are possible unknown heterogeneous reactions occurring on both fresh (i.e., "clean") and conditioned chamber surfaces. Additionally, the outgassing of uncharacterized reactive vapors either deposited there during previous experiments or released from the plastic films used to make the chambers (e.g., nonpolymerized organics) can have pronounced effects in certain reaction systems, especially kinetic studies of low-reactivity organics (e.g., see Lonneman et al., 1981; and Joshi et al., 1982). Another, less severe, problem is reproducing the actinic radiation to which pollutants are exposed in ambient atmospheres. Finally, variations in the kinetics in the initial reaction stages due to inhomogeneities in the reactant concentrations during the mixing should be taken into account (Ibrahim et al., 1987). The design of environmental chamber facilities attempts insofar as possible to minimize these variations from "real" air masses.

A general discussion of the nature and importance of these chamber characteristics, including "wall effects," follows. For detailed descriptions of various types of smog chamber facilities and their operation, one should consult the original literature, including, for example, indoor studies utilizing borosilicate glass cylinders (Joshi et al., 1982; Behnke et al., 1988), chambers made from Teflon (FEP) film with volumes up to

~250 m3 (Fitz et al., 1981; Spicer, 1983; Evans et al., 1986; Mentel et al., 1996), and two similar 6-m3 evacuable and thermostated chambers (Darnall et al., 1976; Winer et al., 1980; Akimoto et al., 1979). Fluo-rocarbon-film chambers range in size from 0.015-0.04 m3 (Lonneman et al., 1981) to 200-2000 L (Kelly, 1982; Kelly et al., 1985; Evans et al., 1986) to large outdoor chambers with capacities of 25-200 m3 (Jeffries et al., 1976; Kamens et al., 1988; Leone et al., 1985; Wangberg et al., 1997).

(1) Glass reactors Many studies have been carried out in borosilicate glass reactors similar to those used in typical laboratory studies of gas-phase reactions. These are usually relatively small, a few liters up to approximately 100 L (0.1 m3). However, Doussin et al. (1997) have developed a borosilicate glass chamber with a volume of 977 L by using four cylinders held together by flanges.

While glass reactors are convenient, inexpensive, and readily available, there are some problems associated with their use. For example, Pyrex glass absorbs light at wavelengths <350 nm (the cutoff depends on the thickness of the glass, Fig. 16.10), which we have seen is a critical region for atmospheric photochemistry. Thus unless separate windows are included, species such as 03 and HCHO which produce the free radicals OH and H02 are exposed to less short wavelength (e.g., 290-330 nm) radiation than is present under atmospheric conditions and hence undergo less photolysis.

In addition, such small vessels have high surface-to-volume (S/V) ratios, which may increase the relative contributions of reactions that occur on the surface. One such heterogeneous reaction is the decomposition on surfaces of 03 to form 02; obviously, the faster this decomposition, the lower the concentrations of 03 that will be observed during the chamber run.

A more important reaction that may occur on surfaces in smog chambers is one generating HONO:

As discussed in Sections B.3 and C of Chapter 7, this reaction has been shown to be too slow in aqueous solution to be significant in the atmosphere; it is faster on surfaces and has been proposed as a source of HONO in smog chambers (e.g., see Sakamaki et al., f 983; Pitts et al., 1984; Leone et al., 1985; and Chapter 7.C). Since HONO is a major OH source in the early stages of irradiation in smog chambers, it is important to understand the mechanism of its formation and to quantify its rate of production under various experimental conditions. Thus, if this reaction only occurs at a significant rate on the surfaces typically encountered in chambers, the effects it has on the overall reactions should be removed when the results of chamber studies are extrapolated to ambient air. In the case of HONO generation, studies carried out in a mobile laboratory where the surfaces were very different than in environmental chambers also showed HONO formation (Pitts et al., 1985), suggesting that this reaction may not be unique to smog chamber surfaces. Furthermore, as discussed in Chapter 7, there is evidence from field studies that HONO is formed from NOx heterogeneous reactions on other surfaces as well (e.g., see review by Lammel and Cape, 1996).

The belief generally has been that the smaller the S/V ratio (i.e., the larger the smog chamber), the less important such surface reactions will be, and hence the more representative of the ambient atmosphere the results. While there is doubtless some justification for this approach, it must also be kept in mind that there are a variety of surfaces present in real atmospheres as well. These include not only the surfaces of the earth, buildings, and so on but also the surfaces of particulate matter suspended in air (Chapter 9). If the heterogeneous formation of HONO occurs not only on chamber surfaces but also on those found in urban atmospheres as well, then it is important to include it in extrapolating the chamber results to ambient air. In this case, the effects on the kinetics due to the different types and available amounts of surfaces in air compared to chambers must, of course, be taken into account.

(2) Collapsible reaction chambers As a result of these problems, larger smog chambers with surfaces thought to be relatively inert have found increasing use. Thus conditioned FEP Teflon films, for example, have been shown to have relatively low rates of surface destruction of a variety of reactive species. In addition, they typically transmit solar radiation in the 290- to 800-nm region (Fig. 16.f) and have low rates of hydrocarbon offgassing.

Collapsible smog chambers are easily constructed using flexible thin films of this material. In addition to the low rates of destruction of reactive species (e.g., see McMurry and Grosjean, 1985; Grosjean, 1985; and Kuster and Goldan, 1987) and their transparency to actinic UV, they have the advantage that the size of the chamber can be easily varied. An additional advantage of such chambers is that they may be easily divided into two sections simply by putting a heavy divider (e.g., bar) across the middle of the bag. One can then use one side of the bag as a "control" and the other to study the effects of varying one parameter such as the injection of additional pollutants.

Wavelength (nm)

FIGURE 16.1 Typical absorption spectrum in the UV region of a 2-mil-thick Teflon film used to construct environmental chambers (spectrum taken by A. A. Ezell).

Wavelength (nm)

FIGURE 16.1 Typical absorption spectrum in the UV region of a 2-mil-thick Teflon film used to construct environmental chambers (spectrum taken by A. A. Ezell).

Figure 16.2 shows a schematic diagram of such a chamber (Fitz et al., 1981). Ports are included for the introduction of the primary pollutants and for sampling for product analysis. Because such bags do not have a rigid shape, they operate at atmospheric pressure. The volume of the chamber may be maintained during a run by introducing clean air at the same rate as sampling removes air from the chamber, thus diluting the mixture. Alternatively, these soft chambers can be allowed to collapse as air is removed for analysis; this maintains the pressure at f atm but results in an increasing S/V ratio during a run.

A potential problem with the use of these collapsible reaction chambers is contamination by outgassing of organics from the material used to fabricate the chambers. For example, the release of significant amounts of low molecular weight fluorocarbons from FEP Teflon has been reported (Lonneman et al., 1981). While this problem does not seem to occur with all samples (Kelly et al., 1985), investigators should clearly exercise caution with regard to possible contamination from this source.

Another possible problem is that the surfaces of Teflon films may be quite electrostatic. Transport to, and adsorption on, such surfaces can deplete the concentrations of particles in such chambers. For example, McMurry and Rader (1985) have shown that particles in the ~0.05- to 1.0-ynm size range are removed in Teflon chambers primarily by electrostatic attraction to the Teflon surfaces. This leads to a large portion

35-70%) of particles formed during smog chamber runs being deposited on the walls before the end of the experiment (McMurry and Grosjean, 1985). This problem can be minimized by the use of very large chambers, for example, the 256-m3 chamber at Jülich, Germany (Mentel et al., 1996), or the dual-chamber EUPHORE chambers at Valencia, Spain, each having a volume of ~200 m3 (Wängberg et al., 1997).

In any case, experience suggests that Teflon chambers first be conditioned by filling them with air containing 03 and leaving them in the dark for several hours and/or filling them with "clean" air containing added NOx and irradiating them for a period of time (Kelly, 1982). Such chamber "conditioning" and characterization studies appear to be an essential initial step in detailed chamber studies of the kinetics and mechanisms of atmospheric reactions.

Metal cover support

Metal cover support

FIGURE 16.2 Schematic diagram of typical outdoor 40-m3 collapsible bag environmental chamber (adapted from Fitz et al., 1981).

(3) Evacuable chambers Ideally, one would like to be able to vary the pressure and temperature during environmental chamber runs in order to simulate various geographical locations, seasons, and meteorology and to establish the pressure and temperature dependencies of reactions. Varying the pressure and temperature also allows one to simulate the upper atmosphere (e.g., to study stratospheric and mesospheric chemistry).

Temperature control has an additional advantage with respect to the problem of chamber contamination. After a smog chamber has been used, some hydrocarbons and nitrogen compounds may remain adsorbed on the chamber walls. These may desorb in subsequent runs and, in some cases (e.g., HCHO), act as free radical sources to accelerate the photooxidation processes. The ability to "bake out" smog chambers while pumping to low pressures is therefore useful in reducing chamber contamination effects.

While glass reactors can be easily designed to include pressure and temperature control, they suffer from other limitations discussed earlier. In addition, the use of very large glass evacuable chambers at low pressures presents a potential safety problem. On the other hand, pressure and temperature are not easily controlled using collapsible reaction chambers.

As a result, some evacuable smog chambers have been designed and constructed to control both temperature and pressure. The one shown schematically in Fig. 16.3 is constructed with an aluminum alloy and the walls are coated with FEP Teflon. The end windows through which the mixture is irradiated are ultraviolet-

grade quartz to allow transmission of the actinic UV. The radiation cutoff and spectral distribution in the 290- to 350-nm region can be varied using filters between the irradiation source and the chamber windows. The pumping system used to evacuate the chamber is hydrocarbon free; back-diffusion of organics from the use of conventional pump fluids can produce VOC concentrations that by themselves exceed the VOC air quality standard! In the particular case shown in Fig. 16.3, the pumping system consists of a liquid ring roughing pump (the fluid is water) with an air injection pump, cryosorption roughing pumps, a diffusion pump, and a mechanical pump. (Both the diffusion and mechanical pumps use unreactive perfluorinated oils rather than hydrocarbons as the working fluid.) Ports are included in the chamber for air and pollutant injection, for sampling (e.g., for GC or GC-MS analysis), and for in situ analysis using optical absorption spectroscopy (Darnall et al., f 976; Winer et al., f 980).

This type of chamber satisfies most design criteria in that both pressure and temperature can be varied, the intensity and spectrum of the irradiation can be altered, and the surface can be coated with a relatively inert material to minimize heterogeneous reactions and pollutant adsorption and offgassing. In addition, ports for both in situ spectroscopic product analysis and sampling can be easily included. The disadvantages are that they are relatively expensive, varying the S/V ratio through changing the dimensions of the chamber is not practical, and changing the nature of the chamber surface is difficult and time-consuming. Three such

Haagen Smit Smog Chamber

for FTIR

FIGURE 16.3 Schematic diagram of the evacuable chamber at the Air Pollution Research Center, University of California, Riverside.

for FTIR

FIGURE 16.3 Schematic diagram of the evacuable chamber at the Air Pollution Research Center, University of California, Riverside.

chambers are described by Akimoto et al. (1979), Winer et al. (1980), and Doussin et al. (1997).

b. Preparation of Reactants, Including "Clean Air"

Different laboratories often use different sources and/or methods of preparation of the reactants NOx, VOC, and "clean air." Frequently, the desired compounds can be purchased commercially. However, care must be taken when using commercially supplied materials to be sure that trace impurities, which can alter the experiment, are not present or are removed by purification before use. For example, gaseous HN03 at concentrations up to several percent is commonly present in commercially produced NOz that is stored in a gas cylinder. Similarly, small concentrations of alkenes are frequently present in commercial cylinders of the alkanes.

By far the largest component in smog chamber studies is air. ft is especially imperative therefore that the air used to dilute the NOx and VOC is "clean." Ambient air and commercially supplied air generally contain sufficient organic and NOt impurities that extensive purification is needed to reduce these contaminants to fow-ppb levels or below. One such purification system is described by Doyle et al. (1977); others are described in the papers on chamber facilities cited earlier.

c. Light Sources

(1) The sun At first glance, it might appear that the sun would be the ideal irradiation source for environmental chambers, and, indeed, it has been used successfully in outdoor chamber facilities (e.g., see Jeffries et al., 1976; Leone et al., 1985; and Kamens et al., 1988). However, there are a number of practical problems. First, it requires either that the chamber be built outdoors or that the building housing the chamber have a suitable opening to admit the sunlight. Under these conditions, independent temperature control of the chamber becomes difficult. Second, the intensity of the sunlight can be altered by passing clouds in a manner that is difficult to measure and describe accurately. Third, experiments are limited to days of appropriate meteorology (e.g., rainy days are generally excluded). As an alternative, three types of lamps have been used to mimic irradiation from the sun. These are black fluorescent lamps, sunlamps, and xenon lamps.

(2) Black lamps A black lamp is a low-pressure mercury lamp whose envelope is covered with a phosphor such as strontium fluoroborate or barium disili-cate. The type of phosphor determines the spectral distribution of the lamp output (Forbes et al., 1976). Figure 16.4 shows a typical spectral distribution for A < 500 nm from a black lamp as well as the solar

0 0

Post a comment