FIGURE 11.23 Schematic of principle of operation of denuders. G = gas, P = particles.

the tube and at distance x, respectively, D is the diffusion coefficient of the gas in air, and Q is the volumetric flow rate.

There are a variety of denuder designs, for example, ones incorporating a number of separate tubes in parallel or annular denuders in which the air is pulled through the annular space between two concentric tubes (e.g., Hering et al., 1988; Krieger and Hites, 1992; Koutrakis et al., 1993; Eatough et al., 1993). In a variant of this method, the "coating" is a stream of water that continuously flows along the walls of the denuder and is collected for analysis (Buhr et al., 1995), or, alternatively, a parallel plate with NaOH as the absorbing agent can be used (Simon and Dasgupta, 1995). The opposite approach is used in a diffusion separator developed for semivolatile organic compounds. in this case, the air containing the aerosol and gases flows along the outer walls of a tube, in the center of which is a core flow of clean air; only gases diffuse sufficiently rapidly to penetrate into the central core of clean air, which is sampled at the end of the tube onto a solid sorbent (Turpin et al., 1993). However, despite differences in design in each case, the fundamental principle of using rapid gas diffusion to separate gases and particles is common to all methods.

Denuders have been used in several different ways. One is to extract the walls of the denuder and measure the adsorbed gas directly by ion chromatography. Denuders have also been used as "difference denuders." For example, in nitric acid measurements, the combination of gas-phase HN03 and particle nitrate has been measured using a nylon filter or Teflon-nylon filter combination in one sampling train. In a parallel sampling train, particulate nitrate alone is measured by first passing the airstream through a denuder to re move gaseous HN03. The difference between the two gives gaseous HN03.

As discussed with respect to the measurement of individual compounds, different coatings are used for the collection and measurement of different compounds. The criteria used to choose these coatings and interferences that can occur in the application of denuders to ambient air measurements are discussed by Perrino et al. (1990).

c. Transition Flow Reactors (TFRs)

These operate in a manner similar to that of denuders except that the gas flow is in the transition regime rather than being laminar flow and only a fraction (F) of the gas of interest is trapped at the walls. As described by Durham et al. (1986), TFRs can be treated as if there is a stagnant film of air adjacent to the wall and a core of turbulent air passing through the center of the tube. Uptake of the gas can be thought of as molecular diffusion through the stagnant air film. The fraction of the gas taken up is then given by p = I - e(.-2Trr»x/Q\)

where x, D, and Q are as defined in Eq. (t), r is the radius of the tube, and A is the thickness of the stagant air film at the wall. F is typically about 10% and in practice is determined by independent calibrations. The advantages of this sampler are that it has a high gas transfer coefficient and samples a greater volume of air (Durham et al., 1986).

However, in at least one intercomparison study using diffusion denuders and transition flow reactors, different results were obtained for some important atmospheric gases such as S02, HN03, and H+, where the TFR values were about 30, 80, and 85% higher, respectively, than those from the denuder system (Sickles et al., 1989); the researchers attributed these differences to biases in the TFR measurements.

d. Mist Chambers and Scrubbers

Air is passed through a chamber where a mist of water or other aqueous solution is used to scrub out species of interest. The solution is then analyzed for the corresponding ions. As discussed shortly, this method has been used for several atmospheric gases, including HN03, carboxylic acids, and carbonyl compounds.

It has also been applied to measure inorganic chlorine gases and to differentiate HC1 from other inorganics such as CI2 and HOC1 (Keene et al., 1993; Pszenny et al., 1993). in this case, the first chamber has an acidic solution that scrubs out HC1, some Cl2 and HOC1, and other chlorine-containing species such as C1NO,

C1N02, and C10N02. The air then passes into a second chamber with an alkaline scrubbing solution, which absorbs most of the Cl2 and some HOC1. The two solutions are analyzed for chloride ion by ion chromatography. Differences in the chloride ion concentrations in the acid compared to the alkaline solutions provide a measure of chlorine-containing inorganics other than HC1.

4. Methods for, and Tropospheric Levels of, Specific Gases a. NO, N02, NOx, and NOy

As we have seen in earlier chapters, NO is the major form of nitrogen oxides emitted from combustion processes, but in the atmosphere it is oxidized to N02 and other oxides of nitrogen. The term NOx is used for the sum of (NO + N02). The term NO denotes the sum of NO, N02 (i.e., NOx), plus all other oxides of nitrogen where the nitrogen is in an oxidation state of +2 or greater:

+ HONO + PAN + higher peroxynitrates + alkyl nitrates + particulate nitrate ... (K)

The term NOz is also occasionally used in the literature. In these cases, it is defined by

Operationally, NO is defined by the measurement method used to measure it, as discussed in more detail in Section A.4.a(2). Since NO, NOr and NOx are commonly measured simultaneously using variants of the same techniques, these are discussed together in the following sections, and in that order, for reasons that will become apparent.

(1) NO Nitric oxide is most commonly measured using the chemiluminescence from its reaction with 03 described earlier. One such instrument designed for high-sensitivity (1- to 2-ppt detection limit in 10 s) is described by Ridley and Grahek (1990).

A second method is a two-photon laser-induced fluoresence (TP-LIF) technique (Bradshaw et al., 1985; Sandholm et al., 1990, 1997). Figure 11.18b illustrates the basis of this method. Ground-state NO (X2n) is pumped at 226 nm using a Nd:YAG pumped dye laser into the A2X state. This molecule is further excited by a second photon, A', in the 1.06- to 1.15-/j,m range into the D2X electronically excited state, from which it fluoresces, returning to the ground state. Because the fluorescence occurs at higher energies and shorter wavelengths (187-220 nm) than the two pumping steps, interference from the excitation lasers is minimal. While the simplest approach is to carry out the second step using a fixed (1.1 yu,m) wavelength (Sandholm et al., 1990), there are advantages to being able to tune the IR laser, such as increasing the selectivity of the measurements and optimizing the pumping efficiency from the A2X state to the D2X state (Bradshaw et al., 1985). The sensitivity of this method is ~20 ppt for a 1-s integration time and 0.4 ppt for fOO-s integration time at a signal-to-noise of 2:1 (Sandholm et al., 1990, 1997).

Intercomparison studies of these two measurement methods for NO generally show good agreement for levels of 25 ppt and greater (e.g., Hoell et al., 1987a; Gregory et al., 1990; Crosley, 1996). For example, Fig. 11.24 shows the results of one aircraft study in which the chemiluminescence method and the TP-LIF method were compared (Crosley, 1996). The slope of the plot in Fig. 11.24 was 0.94, with an intercept of —0.1 + 0.8 ppt and r2 = 0.90. For the data <25 ppt, although the slope was 0.989, the correlation was poorer, r2 = 0.66.

A technique that has been used in laboratory studies for oxides of nitrogen and shows promise for field measurements is resonance-enhanced multiphoton ionization (REMPI) (Guizard et al., 1989; Lemire et al., 1993; Simeonsson et al., 1994). For example, Akimoto and co-workers (Lee et al., 1997) have reported a REMPI system in which a (1 + 1) two-photon absorption of light at 226 nm by NO results in ionization (vide supra). They report a detection limit of ~16 ppt in their laboratory studies. Other oxides of nitrogen such as N02 and HN03 can also photodissociate in the

0 0

Post a comment