OCR Text |
Show Table 2 requires a few other clarifications. First, the technology to make laser diodes capable of operating in the spectral region where the OH radical absorbs has been lost, so the feasibility of such observations at this time is questionable. Second, the molecules most likely to react with surfaces in an extractive sampling system are marked, but this question must be evaluated for each system and molecule. Speedy conveyance of the gas sample into the multipass cell will clearly improve the accuracy of the measurement Finally, the concentrations listed for the major species CO2 and H20 are not in fact minimum detectable values, since that is not really the question of interest for them. Both molecules have very strong infrared absorption lines, and those lines could be used to detect tiny concentrations, but the concentrations of these gases in combustion exhausts will always be large. What is of interest in these cases is how accurately the absolute concentrations can be determined. This in turn depends not on noise in the detection system, but on accurate measurement of zero and full transmission values. Measurement of these quantities to better than 1 part in 100 is difficult, especially under field conditions, so a more meaningful feeling for the uncertainty of a measurement of a strong absorption is given by assuming a 1 per cent fractional error. More of these adjustments have been made in Table 3, which begins with a column of expected concentrations (appropriate to aircraft engine exhausts). When the minimum detectable concentrations derived as described above are more than 100 times smaller than the expected concentrations, 1 per cent accuracy error limits are substituted for them. Table 3 also includes columns for absorption measurements using FTIR and DOAS instrumentation. These values were scaled for path length directly from the values reported in Reference 28 for long path atmospheric observations, with the exception of the values in parentheses, which we estimated using literature values for absorption coefficients and the minimum detectable absorptions implied by the values reported by Reference 28. A review of a number of reports of long path measurements indicates that present FTIR systems are doing well when they approach minimum detectable optical depths of 10-3, and that this level is oarticularly difficult to achieve in regions with significant interferences such as water or CO2 lines. 10 |