FIGURE 11.11 Schematic diagram of components of a DOAS system.
magnitude) through the atmosphere because of Mie scattering by the cloud droplets.
Surface-based instruments have also been developed for the application of DOAS to measure the integrated absorptions either over long direct path lengths or over folded light paths that give large total path lengths and hence high sensitivity but more closely approximate point measurements. There are two common approaches that have been used. In the earlier systems, a slotted-disk arrangement with a photomultiplier was used. These have been largely supplanted by the use of photodiode arrays.
in conventional spectroscopy, the grating of the spectrograph disperses the light so that the spectrum spreads out across the exit plane. The exit slit is stationary and wavelength scanning is achieved by slowly rotating the grating so that a series of wavelengths strike the exit sequentially and are detected by the photomultiplier. However, this is not suitable for ambient air studies where atmospheric turbulences with frequencies of < 10 Hz make it desirable that spectra be scanned at rates > 100 Hz. The use of the slotted disk, developed by Piatt, Perner, and co-workers, allows one to attain the high scan rates needed. In this technique, the conventional exit slit is replaced by a mask that allows a 6- to 40-nm segment of the dispersed spectrum to fall on a rotating wheel, with the central wavelength set by the spectrograph wavelength setting. The rotating wheel contains a number of narrow slits (typically 50) around its perimeter. As seen in Fig. 11.12, as the wheel rotates, the slits "scan" the portion of the spectum dispersed across the monochromator exit slit. The slits in the rotating wheel are sufficiently well spaced that only one rotating slit is in the aperture at one time and also sufficiently narrow that only the light from a small portion of the dispersed spectrum passes through the rotating slit to the detector.
The signal, detected using a photomultiplier, is measured at several hundred different locations of the rotating slit across the exit aperture (i.e., at several hundred different wavelength intervals), and these signals are stored in different channels of a computer for subsequent data analysis. The light barrier on the edge of the mask shown in Fig. If.12 triggers the computer so that as a rotating slit enters the mask aperture, data accumulation is started. As each rotating slit crosses the exit plane of the monochromator and performs one scan, the signals are added to the appropriate channels in the computer, resulting in many scans being superimposed; this signal averaging increases the signal-to-noise ratio.
As described in standard analytical chemistry books (e.g., Skoog et al., 1998), photodiode arrays consist of a series (typically 1024) of side-by-side semiconductor rectangular detectors, or pixels, in this second type of DOAS instrument, the exit slit of the spectrograph is replaced by the photodiode array detector (PDA). Light striking the spectrograph grating is dispersed onto the PDA. The particular range of wavelengths striking the PDA is determined by the rotation of the grating, and
the resolution, i.e., nanometers per pixel, by the entrance slit width. For example, a typical spectral range covered in one scan or set of scans is 40 nm, and with a 1024-element PDA, the resolution is then 40 nm/1024 pixels = 0.04 nm per pixel.
The advantage of using a PDA is that it records all wavelengths simultaneously, the so-called "multiplex" advantage. As a result, total photons detected are about 100-500 times greater in a given time period than for the slotted-disk arrangement, resulting in at least an order of magnitude increase in signal-to-noise (Stutz and Piatt, 1997). However, there are some complications with using PDA that must be taken into account. First, the response of each of the pixels is not identical, which must be taken into account, for example, using multichannel scanning techniques described by Brauers et al. (1995). Second, under atmospheric conditions, different angles of incidence of the light on the PDA can give rise to "residual structures" in the spectrum that remain after all of the true absorptions have been removed; these can be quite large, of the order of f0~2 absorbance units, thus limiting the sensitivity to an order of magnitude less than the slotted-disk instruments. The use of a quartz fiber mode mixer overcomes this problem by acting as a diffuser, providing even illumination of the PDA with relatively small losses (-20%) in the intensity (Stutz and Piatt, 1997).
(4) Typical DOAS spectra and detection limits Table 11.3 shows detection limits for some gases of atmospheric interest at a path length of 5 km for the slotted-disk and PDA techniques, respectively, and for the PDA at a path length of 15 km (Stutz and Piatt, 1997). Also shown are detection limits for a 5-km path length estimated by Plane and Smith (1995). With the improvements in the PDA method described by Stutz and Piatt (1997), the sensitivity is as good as, or better than, that using the slotted-disk approach. Detection limits for 15 km using the PDA vary from sub-ppt levels for N03 to about fOO ppt for HCHO.
Figure 11.13 shows a typical DOAS spectrum measured in air after correcting for atmospheric background light and an electronic offset (Stutz and Piatt, 1997). Below the spectrum are shown reference spectra for the gases that contribute to the atmospheric spectrum, scaled by the a, factors determined using Eq. (H). In this case, 03, N02, S02, and HCHO all contribute, leaving a residual spectrum with a peak-to-peak absorbance of 6 X f0~4.
DOAS has proven particularly useful for N03, for which other widely used methods are not available, and for HONO. In the latter case, denuder techniques have been applied, but a great deal of care must be exercised to recognize and, if possible, avoid artifacts (see later). Figure 11.14 shows the application of DOAS to the measurement of the nitrate radical during the night in Riverside, California. Since N03 photolyzes rapidly, it is only detectable at night. Bands at 623 and 662 nm can be seen growing in, peaking in this case at ~290 ppt around 8 p.m. local time (Platt et al., 1980b). As discussed in Chapter 7.D, the diurnal profile and time of the peak are quite variable, depending not only on its rate of formation but also on the scavenging processes.
DOAS has also been used for the measurement of the OH (see later) as well as BrO, CIO, and IO free radicals in the atmosphere (Piatt and Hausmann, 1994; Platt and Janssen, 1995; Tuckermann et al., 1997; Hebestreit et al., 1999; Alicke et al., 1999), all of which have absorption bands in the UV (see Chapter 4 and DeMore et al. (1997)). For example, Fig. 11.15 shows OH concentrations measured as a function of time using DOAS (Dorn et al., f 996). The OH bands clearly
Was this article helpful?