M

OH + NO + M -> HONO + M, (11) HONO + hv -> OH + NO, (12)

The rate constants for all these reactions are reasonably well established, so that the concentrations of OH expected can be predicted with some degree of reliability. (This would not be true if the concentration of organics was high enough to perturb the NOx chemistry because of the many uncertainties in the mechanisms of the organic reactions.)

Line A in Fig. 16.12 shows the OH concentrations predicted using only the homogeneous gas-phase chemistry of reactions (1)—(13). Clearly, much larger concentrations of OH are observed than can be rationalized on the basis of gas-phase chemistry alone.

2 4 6 Irradiation time (hr)

FIGURE 16.13 Comparison of observed 03 concentration-time profiles (O) in two different evacuable chambers to predicted profiles if the heterogeneous production of HONO (reaction (14)) does not occur (curve A) and if reaction (14) with photoenhancement does occur (curve B). Results in (a) are results from the chamber of Akimoto et al. (1985) and those in (b) are from the evacuable chamber in Fig. 16.3 (adapted from Sakamaki and Akimoto, 1988).

2 4 6 Irradiation time (hr)

2 4 6 Irradiation time (hr)

FIGURE 16.13 Comparison of observed 03 concentration-time profiles (O) in two different evacuable chambers to predicted profiles if the heterogeneous production of HONO (reaction (14)) does not occur (curve A) and if reaction (14) with photoenhancement does occur (curve B). Results in (a) are results from the chamber of Akimoto et al. (1985) and those in (b) are from the evacuable chamber in Fig. 16.3 (adapted from Sakamaki and Akimoto, 1988).

As discussed earlier, one source of OH is the photolysis of HONO formed on surfaces by reaction (14),

If one assumes that 10 ppb of HONO was present initially, the OH radical profile would follow curve B. This is in agreement with the initial OH concentrations but not with the larger concentrations at later times. The shape of curve B is expected since any initial HONO present will rapidly photolyze, producing a "pulse" of OH; however, after the HONO has pho-tolyzed, lower concentrations of OH follow.

The shape of the observed OH profile suggests that some unrecognized source of OH must be present. Curve C is the predicted OH profile if it is assumed that a constant OH source exists during the run that produces OH at a rate of 0.245 ppb min-1. In contrast to curve B, this leads to agreement with the observed OH levels and curve shape at long reaction times, but not at short times. (The lower OH concentrations predicted at short reaction times are due to the net reaction of OH with NO while HONO is building up to its equilibrium concentration.)

Combining curves B and C (i.e., using the assumption of both an initial HONO concentration and a constant radical source), one obtains curve D, which matches the observations quite well.

While a constant flux of OH was assumed in generating curves C and D of Fig. 16.12, this is obviously only a simplified construct taken to represent some species that forms OH in subsequent reactions. The flux of this unknown species appears to depend on the particular chamber and chamber history and is roughly proportional to the light intensity. It was also found that it increases significantly with temperature, relative humidity, and NOz concentrations but is independent of total pressure and the NO concentration (Carter et al., 1982).

Glasson and Dunker (1989) used the oxidation of CO by OH to remove the chamber radical source. When an excess of CO is used, the OH is converted to HOz and then reacted with NO to give N02:

The OH concentration is then calculated from the rate of loss of NO, the CO concentration, and the known rate constants. In contrast to the observations of Carter et al. (1982), the chamber OH radical source was found to depend on light intensity and temperature, but not on N02.

The identity and mechanism of formation of the precursor(s) to OH are unknown and subject to some controversy (Killus and Whitten, 1981; Besemer and Nieboer, 1985; Carter et al., 1985; Leone et al., 1985; Sakamaki and Akimoto, 1988), although HONO appears to be the most likely candidate. The dependence of its flux on the particular chamber characteristics suggests the involvement of unrecognized heterogeneous reactions on the chamber surfaces.

Sakamaki and Akimoto (1988) have used a combination of computer kinetic modeling and the results of environmental chamber experiments to show that the ozone concentration-time profiles are consistent with known chemistry if the photoenhancement of reaction (f4) (Akimoto et al., 1987) is taken into account.

Figure 16.13, for example, shows the concentration-time profiles for a run in the evacuable chamber shown in Fig. 16.3 and for one in the evacuable chamber of Akimoto et al. (1985). The calculation, which assumes no radical source, curve A, clearly underpre-dicts 03 by a large margin. However, inclusion of a photoenhanced production of HONO via reaction (14), curve B, matches the observations quite well (Sakamaki and Akimoto, 1988).

However, without knowledge of the source of the increased OH flux, extrapolation of the concentration-time profiles of both the primary and secondary pollutants observed in such smog chamber studies to real atmospheres becomes less certain. For example, the reactions leading to the unknown precursor(s) to OH may occur only in smog chambers. Extrapolation to ambient air would thus require subtracting out this radical source. On the other hand, the same reactions may occur in ambient air where surfaces are available in the form of particulate matter, buildings, the earth, and so on; if this is true, then the rates would be expected to depend on the nature and types of surfaces available and may thus differ quantitatively from the smog chamber observations.

While chamber contamination and the presence of unknown surface reactions are probably the most important problems in extrapolating smog chamber data to atmospheric conditions, other minor problems exist as well. These include the need to measure carefully and frequently a number of chamber-specific parameters such as the decay rate of 03 on the chamber walls and the initial formation of HONO. Such chamber-specific parameters raise the question again of how best to modify these parameters to describe ambient air.

However, despite these complications, smog chambers have proven extremely useful in studying the chemistry of photochemical air pollution under controlled conditions in which emissions and meteorology are not complicating factors. While there are some uncertainties and limitations in quantitatively extrapolating the results to ambient air, it may be that what appear to be chamber-specific complications may, in fact, apply in ambient air as well.

2. Isopleths for Ozone and Other Photochemically Derived Species

As we have seen, the chemistry connecting the primary emissions, VOC and NOx, to the concentrations of secondary pollutants such as 03, HN03, and particles is very complex, even when such effects as meteorology and new emissions into the air mass of interest are ignored. A result of this complex chemistry is that the concentrations of secondary pollutants are related to those of the precursors in ways that are, under many conditions, highly nonlinear. In principle, environmental chambers can be used to systematically examine these relationships. A first approach, dating back to the 1950s (e.g., see Haagen-Smit and Fox, 1954), is to examine the peak 1-h 03 formed when mixtures of known initial concentrations of VOC and NOx are irradiated in a laboratory chamber. The results are often displayed in the form of 2-dimensional isopleths such as those shown in Fig. 16.14a (Dodge, 1977a; Finlayson-Pitts and Pitts, 1993).

In practice, such isopleths are usually generated using computer models such as the EKMA (Empirical Kinetic Modeling Approach) model (vide infra) where the results of the model have been tested against environmental chamber data. Figure f6.f4a shows the peak 03 formed from the irradiation of mixtures of VOC and NOx at the initial concentrations shown on the axes. Figure 16.14b reflects the same data in three dimensions (Finlayson-Pitts and Pitts, 1993). The overall shape of the "ozone hill" in Fig. 16.14b is useful in examining whether VOC or NOx control, or both, would be most effective in controlling 03. Thus, at high VOC/NOx corresponding to point A, decreasing VOC alone at constant NOx along the AB line gives only slowly decreasing 03. However, decreasing NOx at constant VOC, i.e., along the AC line, is very effective in rolling down the ozone hill. Thus in this case, the chemistry of the polluted air masses is NOx-limited and NOx control is most effective. This region of high VOC/NOx is typical of suburban, rural, and downwind areas.

Low VOC/NOx ratios, e.g., point D, have been found to be typical of polluted air masses found in many major urban centers, e.g., downtown Los Angeles (DTLA). Here, reducing VOC at constant NOx along the line DE results in rolling down the ozone hill. However, reducing NOx at constant VOC, along the

Was this article helpful?

0 0

Post a comment