Calibrated remote measurements of SO2 and O3 using atmospheric backscatter

Download Calibrated remote measurements of SO2 and O3 using atmospheric backscatter

Post on 09-Apr-2017

217 views

Category:

Documents

3 download

Embed Size (px)

TRANSCRIPT

<ul><li><p>Calibrated remote measurements of SO2 and O3 using atmospheric backscatterW. B. Grant and R. D. Hake Jr. Citation: Journal of Applied Physics 46, 3019 (1975); doi: 10.1063/1.321992 View online: http://dx.doi.org/10.1063/1.321992 View Table of Contents: http://scitation.aip.org/content/aip/journal/jap/46/7?ver=pdfcov Published by the AIP Publishing Articles you may be interested in Rate coefficient measurements for SO2+O=SO+O2 J. Chem. Phys. 73, 987 (1980); 10.1063/1.440748 Measurement of electron affinities of O3, SO2, and SO3 by collisional ionization J. Chem. Phys. 62, 3829 (1975); 10.1063/1.430941 Calibrated remote measurement of NO2 using the differentialabsorption backscatter technique Appl. Phys. Lett. 24, 550 (1974); 10.1063/1.1655049 Use of O2 for ESR Calibration for Quantitative Measurement of Gas Concentrations J. Chem. Phys. 44, 1715 (1966); 10.1063/1.1726918 Upper Atmosphere Temperatures from Remote Sound Measurements Am. J. Phys. 16, 465 (1948); 10.1119/1.1991145 </p><p> [This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:</p><p>150.135.239.97 On: Thu, 18 Dec 2014 01:52:33</p><p>http://scitation.aip.org/content/aip/journal/jap?ver=pdfcovhttp://oasc12039.247realmedia.com/RealMedia/ads/click_lx.ads/www.aip.org/pt/adcenter/pdfcover_test/L-37/884760910/x01/AIP-PT/Asylum_JAPArticleDL_121014/AIP-JAD-Cypher1.jpg/47344656396c504a5a37344142416b75?xhttp://scitation.aip.org/search?value1=W.+B.+Grant&amp;option1=authorhttp://scitation.aip.org/search?value1=R.+D.+Hake+Jr.&amp;option1=authorhttp://scitation.aip.org/content/aip/journal/jap?ver=pdfcovhttp://dx.doi.org/10.1063/1.321992http://scitation.aip.org/content/aip/journal/jap/46/7?ver=pdfcovhttp://scitation.aip.org/content/aip?ver=pdfcovhttp://scitation.aip.org/content/aip/journal/jcp/73/2/10.1063/1.440748?ver=pdfcovhttp://scitation.aip.org/content/aip/journal/jcp/62/9/10.1063/1.430941?ver=pdfcovhttp://scitation.aip.org/content/aip/journal/apl/24/11/10.1063/1.1655049?ver=pdfcovhttp://scitation.aip.org/content/aip/journal/jcp/44/4/10.1063/1.1726918?ver=pdfcovhttp://scitation.aip.org/content/aapt/journal/ajp/16/9/10.1119/1.1991145?ver=pdfcov</p></li><li><p>Calibrated remote measurements of 802 and 0 3 using atmospheric backscatter </p><p>W. B. Grant and R. D. Hake, Jr. </p><p>Stanford Research Institute, Menlo Park, California 94025 (Received 3 March 1975) </p><p>Remote measurements of calibrated samples of S02 and 0 3 have been achieved with a lidar using ultraviolet signals produced by a tunable dye laser and a nonlinear crystal. The operating wavelengths for these measurements were 292.3 and 293.3 nm for S02 and 292.3 and 294.0 nm for 0 3, The atmosphere in front of and behind the chamber acted as a distributed reflector to send laser light back through the chamber to a receiver near the laser. The laser measurements agreed well with in situ measurements. Integration of eight laser pulses at each of two wavelengths allowed the determination of S02 concentration with an uncertainty equivalent to 0.6 ppm in 100 m for low concentrations. For 0 3, the corresponding uncertainty limit was 1.2 ppm in 100 m. The measurement errors are primarily attributable to variations in atmospheric backscattering intensity during the experiment, since the different wavelengths were radiated sequentially rather than simultaneously. The sensitivity of a system transmitting more favorable wavelengths at intervals separated by less than 1 min is estimated to be near O.I ppm in 100 m for both S02 and 0 3, </p><p>PACS numbers: S7.60.P, 89.60., 42.6O.Q, 42.68.M </p><p>Recent experimental results have shown that it is possible to make remote measurements of N02 at con-centrations typical of urban environments. 1,2 The mea-surements in Refs. 1 and 2 were made USing mono static pulsed-laser radars employing the differential-absorp-tion-Udar (DIAL) technique. This technique uses two wavelengths for which the gas of interest has differing absorption coefficients. The measured difference in at-tenuation for the two wavelengths, determined by mak-ing time-resolved observations of the energy backscat-tered by the atmosphere, can be combined with the known difference in absorption coefficients to provide a range-resolved measurement of the absorbing species. </p><p>The N02 measurements were made using pulsed dye lasers operating in the visible near 450 nm. It has been suggested that a similar device operating in the uv could be used to monitor S02. 3,4 Total COlumn-content mea-surements of S02 have, in fact, been made using a retroreflector of uv energy to provide the return sig-nal. 5 This paper reports results of active remote mea-surements of S02 and 0 3 using atmospheric backscat-ter; it demonstrates the feasibility of using the DIAL </p><p>r - - --</p><p>5.3 x BEAM </p><p>technique for range-resolved measurement of S02 and 0 3 at uv wavelengths with an eye-safe lidar. </p><p>The equipment arrangement for this experiment is similar to that for N02 in Ref. 2. The principal changes are the use of a nonlinear crystal [ammonium dihydro-gen arsenate (ADA)]6 after the dye laser to generate uv wavelengths by nonlinear doubling, and the replacement of the Pyrex windows in the sample chamber with com-mercial-grade quartz windows (see Fig. 1). Other de-tails of the laser and receiver are given in Table L The ADA cr'ystal is temperature tuned to achieve 900 phase matching at the different wavelengths. The wavelengths employed were selected to give the highest available differential in absorption coefficients for S02 and 0 3 that could be achieved with operation of the ADA crystal in a reasonable temperature range. Use of a single ADA crystal prevented taking the measurements alternately at the two wavelengths at 8-s intervals, as was the case for N02 in Ref. 2. Instead, data were accumulated at one wavelength for about ~-h, with at least a i-h delay between different wavelengths required for changing and stabiliZing the crystal temperature. Thus, there was </p><p>~L EXPANDER QUARTZ PMT j _ WINDOWS </p><p>r=~ ~ fr~~ ) FIG,!. Block diagram of the ex-perimental apparatus used in this experiment. The equipment en-closed in the dashed line was not used in these measurements but could be added to provide nearly simultaneous transmission at two wavelengths (see text for discussion) . </p><p>3019 </p><p>MAGNETIC ,-----2.5 m----j TELESCOPE 12~ </p><p>8-BIT TAPE 9.6 MHz SIGNAL AEROSOLS </p><p>DIGITIZER rMI AND ~ MOLECULES </p><p>REAL-TIME DISPLAY </p><p>Journal of Applied Physics, Vol. 46, No.7, July 1975 Copyright 1975 American Institute of Physics 3019 </p><p> [This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:</p><p>150.135.239.97 On: Thu, 18 Dec 2014 01:52:33</p></li><li><p>TABLE I. Experiment parameters. </p><p>Laser (Refs. 7 and 8) </p><p>Energy Dye Beam diameter Beam divergence (far field) Pulse length Grating Linewidth Repetition rate </p><p>""15 mJ Rhodamine 6G (3.5 Xl(l"5 Mil in MeOR) ""2 mm (3 dB) 0.7 mrad (3 dB) </p><p>250 ns (3 dB) 316 l/mm, used in ninth order "'II,. 0.017Hz </p><p>Second-harmonic-generating crystal (Ref. 6) </p><p>Type Ammonium dihydrogen arsenate Length 2 cm Index matching fluid Fluorocarbon 77 Conversion efficien-cy for 15-mJ funda- 5% mental Pulse length '" 2 00 ns (3 dB) </p><p>Aperture </p><p>Filter </p><p>PMT ~tical efficiency from ADA crystal through PMT </p><p>Lamp </p><p>Filter </p><p>PMT Accuracy </p><p>Wavelength (rIm) </p><p>292.3 293.3 294.0 </p><p>Receiver </p><p>0.056 m2 </p><p>293 nm, 13-nm half-width, 27% peak transmission RCA 7200 </p><p>0.01 </p><p>Transmissometer </p><p>Deuterium 285.7 nm, 3-nm half-width, 17% peak transmission RCA IP 28 7 ppm </p><p>Spectral data </p><p>Absorption coefficient (cm-1 atm-1) (base e) </p><p>S02 0 3 (Ref. 9) (Ref. 10) </p><p>26 28 14 </p><p>22 </p><p>Atmospheric conditions </p><p>Time Visibility Molecular scattering coefficient (calcu-lated) Aerosol scattering (extinction) coeffi-cient (observed) Ambient S02' N02, and 0 3 concentra-tions (Ref. 11) </p><p>19 Dec. 1974, midnight to 5:00 a. m. 10-20 km </p><p>"" 0.14 km-1 atm-! </p><p>"" O. 5-0. 7 km-! </p><p>&lt; 0.03 ppm </p><p>at least a l-h delay between collection of comparable data at different wavelengths. </p><p>The current range-resolution limit is 70 m, deter-mined by the amplifier bandpass that was fixed at 2. 5 MHz to oversample these test returns. The resolution of a field system would be limited to 35 m by the 200-ns uv pulse duration. </p><p>3020 J. Appl. Phys., Vol. 46, No.7, July 1975 </p><p>The 802 was injected by syringe and had a l/e resi-dence time of about 14 min. The 0 3 was generated using a Welsbach ozonator. 12 It could fill the O. 9_m3 sample chamber with 30 ppm of 0 3 in 2 min. The gas content of the chamber was monitored by a dual-beam transmissometer. </p><p>Typical return signals with air and with 105 ppm 802 in the sample chamber are shown for 292.3 and 293.3 nm in Fig. 2. Each signal point is an average of eight consecutive return pulses. The initial peak near 100 m arises from the coaxial nature of the transmit/ receive geometry. The effect of the shadow created by the central obscuration decreases with distance and is negligible by about 150 ro. The peak near 300 m is the return signal from the quartz windows due to scat-tering caused by bubbles and surface reflections. Both peaks are artifically flattened by the data-acquisition system to extend the dynamic range of the 8-bit digitizer to low signal levels. Lidar data similar to those shown in Fig. 2 were obtained during four nights of operation. Minor system improvements were made between the various runs. </p><p>In order to make a comparison of the lidar data with the transmissometer data, a number must be extracted from the lidar return signal that is related to gas con-centrations in the sample chamber. The easiest way to accomplish this is to form a ratio of the lidar signal </p><p>' &gt; </p><p>~ </p><p>~ M cO en '" &gt;-f-in z w f-</p><p>"= ..J </p><p>'" in </p><p>1000 </p><p>100 </p><p>o AIR </p><p>o S02 - 105 ppm </p><p>I.., 0 I ... ---------~I A' </p><p>293.3 nm </p><p>o AIR </p><p>o </p><p> o </p><p>00 </p><p> S02 - 105 ppm o .... 'f0 0. 00 </p><p> ~ CbO oO </p><p> 0 ".0 </p><p> lL-~ __ ~ __ _L __ ~ __ ~_L_~L_~ __ -L __ ~ </p><p>1000 </p><p>100 </p><p>10 </p><p>a 100 200 300 400 500 600 700 800 900 1000 RANGE - meters </p><p>-</p><p>&gt; . ~ </p><p>M </p><p>'" ~ </p><p>&gt;-&gt;::: </p><p>"' ~ f-</p><p>"= ..J </p><p>'" Z '" in </p><p>FIG. 2. Plot of received signal vs range with air or 105 ppm S02 in the sample chamber. Each point is an average of eight laser pulses. The regions used for taking the ratios for calcu-lation of the gas concentration are indicated as A, A', and B. </p><p>W.B. Grant and R.D. Hake, Jr. 3020 </p><p> [This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:</p><p>150.135.239.97 On: Thu, 18 Dec 2014 01:52:33</p></li><li><p>11,---------,---------,--------,---------, </p><p>10 0 0 S02 </p><p>0.9 ~ 0 0 (a) 292.3 nm 10 </p><p>08 20 </p><p>0.7 0 </p><p>30 ~ 40 </p><p>Z </p><p>50 0 </p><p>&gt;= </p><p>06 0 0 </p><p>0: </p><p>60 a: ... 70 ~ u 80 </p><p>z 0 </p><p>90 u </p><p>100 ~ </p><p>110 g </p><p>05 </p><p>0 </p><p>04 0 </p><p>00 0 </p><p>03 </p><p>02 5l 110 Z </p><p>iii 70 80 90 100 </p><p>12 ~ t;; </p><p>0 '" a: 0 w </p><p>0 &gt;-10 ~ 20 ~ </p><p>11 0 52 </p><p>10 (b) 293.3 nm </p><p>0 </p><p>30 ~ z 09 - 40 '" ;: </p><p>=150 </p><p>60 70 </p><p>0 0 00 </p><p>0 0 0 0 08 00 </p><p>0 </p><p>07 00 80 </p><p>90 100 </p><p>0.6 0 </p><p>110 OS </p><p>140 150 160 170 180 T1ME AFTER 00:35 PST - minutes </p><p>FIG. 3. Plots of lidar ratio B/(A +A') as a function of time for injection of 110 ppm 802 into the sample chamber. The ratio is normalized to 1 for air in the chamber at 100 min. The ppm scale on the right-hand side of each plot was calculated using Beer's law and the ratio for air in the sample chamber after the 802 was exhausted. The solid line is the transmissometer-determined value of the concentration; the points are ratios from an eight-pulse lidar averages. </p><p>from behind the chamber to the signal in front of the chamber. The ratio will decrease as the amount of gas in the chamber increases, in accordance with the expression </p><p>RL ex: exp(- 2kpl), (1 ) </p><p>where RL is the lidar ratio, k is the absorption coeffi-cient, p is the concentration, l is the length of the chamber, and the constant of proportionality includes range and atmospheric factors that are independent of wavelength and gas concentration in the test chamber. We have formed this ratio using the average signal in a 560-m segment starting 150 m behind the receiver, and the average signal in 50- and 75-m segments in front of the sample chamber. The segments used in front of and behind the chamber are denoted by A and A', and B, respectively, in Fig. 2. Lidar ratios, RL =B/(A+A'), obtained on 19 December 1974, were used to generate the data points shown in Figs. 3 and 4. </p><p>The points in Fig. 3 show a time history of the lidar ratio for S02' using sets of eight pulses at each of the two wavelengths. Gas injected into the sample chamber at a measured level of 110 ppm was allowed to leak out slowly for about 25 min, then an exhaust fan was turned on. The solid line in Fig. 3 gives the transmissometer-measured concentration of S02' referenced to the lidar ratio for air in the chamber after the S02 has been ex-hausted. The position and magnitude of the transmis-</p><p>3021 J. Appl. Phys., Vol. 46, No.7, July 1975 </p><p>someter concentration scale were adjusted so that the solid line indicates the actual material concentration and the expected lidar ratio as functions of time. Al-though there is fairly good agreement between the lidar ratios and the transmissometer concentration values, it is apparent that the return signals with air in the sam-ple chamber (after the S02 had been exhausted) are slightly higher than before the S02 was injected. Also, there are significant deviations of the data trends from the transmissometer-determined S02 concentrations Both effects arise from changes in atmospheriC condi-tions during the ~ h of data taking. </p><p>Time histories for the return ratios with 0 3 in the sample chamber are shown in Fig. 4. After the first few shots with air in the chamber, there are a few shots with 5-10 ppm of 0 3 in the chamber. Then the concen-tration was rapidly increased to about 30 ppm and al-lowed to decay slowly for about 25 min, after which time the exhaust fan was turned on. The uncertainty in the data is similar to that for S02. </p><p>Two kinds of fluctuations in the data are evident in Figs. 3 and 4: (i) long-term fluctuations [such as the difference between the clear-air shots at 5 and 45 min in Fig. 4(a)], which provide about a 15% uncertainty in the ratio; and (ii) short-term (point-to-point) fluctua-tions, which provide closer to a 3% uncertainty in the ratio. The long-term fluctuations in the lidar ratio are ascribed to variations in aerosol content along the lidar path as the local winds move different parcels of air through the scattering region. The slow changes in the amount of scattering in front of and behind the sample chamber cause the long-term excursions of the lidar ratio from the expected value determined from the clear-air calibration and the transmissometer values of gas concentration. </p><p>In a differential-absorption (DIAL) measurement, lidar ratios at two wavelengths are used to determine material concentration p as follows: </p><p>p == (1/26.kl )Q.nR Lv -lnRLP )' (2) </p><p>where 6.k is the differential-absorption c...</p></li></ul>

Recommended

View more >