THE UNIVERSITY OF MICHIGAN COLLEGE OF ENGINEERING High Altitude Engineering Laboratory Departments of Aerospace Engineering Meteorology and Oceanography Quarterly Report HIGH ALTITUDE RADIATION MEASUREMENTS 1 October.-1 971A -.7l January 1972 Fred -L. Bartman ORA PROJECT 036350 under contract with NATIONAL AERONAUTICS AND SPACE ADMINISTRATION CONTRACT No. NSR 23-005-376 WASHINGTON, D. C. administered through OFFICE OF RESEARCH ADMINISTRATION ANN ARBOR March 1972

As?9 -O-D:rC

Table of Contents Page Abstract v I. Introduction 1 II. High Resolution Measurements (L. W. Chaney) 1 A. 4. 3pm CO2 Band 1 B. 15.0umCO2 Band 3 III. Analysis of High Resolution CO2 Spectra (S. R. Drayson) 3 A. Bands of CO2 between 12 and 18im 3 B. The 4. 3MmBands of C02 4 IV. Medium Resolution Measurements of the 9. 6pm Absorption Band of 03(L. T. Loh and P. A. Titus) 4 V. Analysis of the Medium Resolution 03 Spectra (W.R. Kuhn) 5 VI. Study of the Spectra of CO2 Isotopes (J. B. Russell) 7 VII. References 8 iii

Abstract This report summarizes project activity during the period 1 October 1971 to 31 January 1971. Problems in the measurement of high resolution CO2 spectra at 4. 39m and 15. 0 um and medium resolution 03 spectra at 9. 6,um are noted. Data analysis for the 15prm CO2 are essentially completed. Only an initial inspection of the 4. 3pm C02 has been made, however the data looks precise 2 enough to allow absolute transmissivities to be obtained. The disagreement of the present 9. 6pm 03 data with Walshaw's results is discussed. Analysis of the 15 um CO2 isotope data is well underway. v

I. Introduction This is the 9th Quarterly Progress Report on Contract No. NSR 23-005-376, covering the period 1 October 1971 to 31 January 1972. The additional month has been added to this quarterly report period to cover the one month "no cost" extension of the contract from 1 October 1971 to 31 October 1971. The project effort during this time was divided among the following tasks. A. High resolution measurements (L. W. Chaney). B. Analysis of High Resolution CO2 Spectra (S. R. Drayson) C. Medium Resolution Measurements of the 9. 6p m Absorption Band of 03(L. T. Loh and P. A. Titus) D. Analysis of the Mediur Resolution 9. 6 m 03 Spectra (W. R. Kuhn). E. Study of the Spectra of CO2 Isotopes (J. B. Russell). II. High Resolution Measurements (L. W. Chaney) A. 4. 3mrn CO Band 2 -- At the end of the last reporting period the spectrometer was completely set up to make detailed line measurements. The spectral measurements required for the study have been completed. A total of 100 spectral scans were made at various pressures and broadened with various amounts of nitrogen. During the course of the measurements two significant problems occurred. First the last available carbon element for the water cooled infrared source burned out. The elements had been made in a university 1

instrument shop which no longer exists. Since the drawings available did not precisely describe the part, some experimentation was required to obtain a satisfactory replacement. An attempt was made to make a permanent element of molybdenum. This did not work because the "O" rings intended to separate the cooling water from the heated element were not completely effective. As a consequence, molybdenum pentoxide was formed and condensed on the focusing lens. It could be identified by its blue color. The effect could be modified but not eliminated by changing the argon purge. As a result of this experience, we learned that the water leakage on the carbon elements produced CO2 and CO. The normal argon purge eliminated the CO2, but a higher purge was required to eliminate the CO from the spectra being studied. The second problem concerned detector noise. The detector used for this study was a photo-voltaic indium antimonide detector obtained from Texas Instruments. The nominal bias under test conditions was zero millivolts. However, by placing a band pass interference filter on the cold shield the impedance of the detector increased considerably. The optimum bias for best signal to noise ratio was then increased to 6 millivolts. The exact voltage was very critical for a change of +1. Omv would produce a factor of 2 increase in total noise. Furthermore, the detector bias voltage was a critical function of the temperature of the spectrometer. Probably the slit jaws were viewed by the detector. A temperature change of 1 C produced a bias voltage change of about 1.2mv. Another complication was the fact that the detectivity of the detector changed with the spectromneter temperature. 2

After the problem was understood, the detector bias was set immediately before each scan, and no scans were taken before the instrument was stabilized to within 2 degrees of room temperature. The room temperature was held to within one degree during all scans, and usually after about two hours of operation the spectrometer would fall within this range. B. 15 Mm CO Band The modified Ge:Cu detector mounting described in the last report was tested in the spectrometer. The nitrogen cooled shield around the detector element reduced the drifiting of the detectivity as was expected. However, the thermal noise increased to almost twice its best value. Texas Instruments claimed that this might be due to contamination which is always a danger when the flake is exposed to the atmosphere. However, the detector is superior to the new Ge:Cu detector obtained from them. Although the detector was marginal for line width studies, a series of 25 scans were taken on two individual lines. The lines were broadened with various amounts of nitrogen. The data is being saved for possible analysis at a later date. The spectrometer is now being modified for other spectral studies. III. Analysis of High Resolution CO, Spectra (S. R. Drayson) A. Bands of CO2 between 12 and 18 Pm The analysis of the high resolution spectra of the samples of C02 of normal isotopic abundance has been virtually completed. Most of the remaining analysis is for the bands of the less common isotopic molecules 3

which are probably better left for the analysis of the enriched samples. A technical report is being prepared describing the experimental procedure and the method of analysis of the spectra. It will also contain the results of the analysis of each band, including the band center and rotational constants, with estimates of the errors associated with the determination of the constants. Selected spectra will be included. B. The 4. 3 pm bands of CO2 Only preliminary work has been done towards the analysis of the high resolution spectra. Calcomp plots have been made of each of the scans and an examination of these has determined that the stability of the background is sufficiently high to allow absolute transmittances to be obtained. IV Medium Resolution Measurements of the 9. 6/ m Absorption Band of 03. (L. T. Loh and P. A. Titus) During the initial part of this work period, the operation of the PDP-8 computer used for data handling became intermittantly bad. From time to time, the % of transmission vs wavelength curves obtained from the IBM cards differed from the analog plot recorded on the Perkin Elmer spectrophotometer. Since the trouble appeared very infrequently, its detection was not an easy task. Finally it was located within the computer itself. The better part of two weeks was required before the repair was completed. During the time that the computer was not working, alternative schemnes to keep the ozone work going without the use of this computer were studied. These schemes included manual data reduction using chart 4

record from Perkin - Elmer recorder, data acquisition using a Hewlett-Packard system we purchased a long time ago, or using a new system to be acquired. These deliberations were discontinued when the system became operable again. The ozone work was started again on November 10, and continued to January 27. Emphasis was put on measuring ozone spectra of very low ozone concentrations (of the order of 0. 01 atmosphere-cm) in the White cell 3200 cm absorption path. A study was made to find the relation between the accuracy of ozone measurement and percent transmission. This study showed that at low concentrations of a few hundreths atm - cm. of ozone in this White cell, the accuracy of the quantitative measurement of ozone would be poor. The situation can be improved by reducing the path length for the infrared spectra while maintaining that same path for the ultraviolet. It may also be necessary to determine ozone by the chemiluminescent method, which is adaptable to concentration of ozone similar to those in the upper atmosphere. V. Analysis of the M/edium Resolution O Spectra (W. R. Kuhn) Absorptivities forthe 9pum region are given intable 1. The total absorptivity as well as the individual contributions of the 9pm and 9. 6pm bands are given. Absorptivities were calculated for pressures from 7. 5 to 437 mm Hg, and for ozone mass paths from 0. 029 to 10. 04 atm-cm. Column 6 gives the 9. 6pm absorptivity as predicted from the empirical formula of Walshaw (1957), while the last column is the relative difference between the predicted Walshaw value and the present results. 5

The 9. 6m and 9/m absorptivities were determined by two different methods. The upper line in table 1 is the absorptivity for a division of the two bands at 9. 35gm which corresponds to the separation used by McCaa and Shaw (1967). In the second method the 1 band was considered to be symmetric about 9. 07pm, which implies that there is no contribution to the absorptivity from the 3 band beyond this wavelength. This method was discussed in the previous report and the results are given by the lower line for each pressure and mass path in table 1. The two methods give differences which are well within the experimental error for the v3 band; however, because of the smaller absorptivity of the v1 band, the two methods show appreciable differences being as large as 50% for the smaller mass paths. Clearly, both methods are somewhat arbitrary, and the v1 absorptivities can only be viewed as being very approximate. The present results for the 9.6gm band are consistently smaller than those predicted from Walshaw's formula, the relative difference being approximately 20%. It is interesting to note that this is a systematic differenc e as can be seen from Figure 1. Only those results have been plotted which fall within the mass path range used by Walshaw. It would appear that the present results indicate a larger ozone mass path than would have been found by Walshaw. This difference cannot be accounted for by different ultraviolet absorption coefficients since the present mass paths were adjusted for use in Walshaw's formulae, i. e. Walshaw's adopted ultraviolet absorption coefficients for the 2536 and o 2893 A lines were corrected to Hearn's (1961) values of 133. 9 and -1 17. 2 (rcm NTP ) respectively. A study is presently underway to ascertain the possible errors and quantitatively, their relative significance, 6

such as the temperature difference within and outside the White cell, ozone decay rates, and possible non uniform mixing of ozone within the cell. When these uncertainties have been resolved, additional measurements will be made for pressures and mass paths corresponding to typical atmospheric conditions. If these later results are compatible with those of McCaa and Shaw both their results and those of the present study will be used to develop an empirical expression for the absorptivity for the entire 9pm region foruse in stratospheric energy budget studies (Walshaw's empirical formula cannot be used for this purpose since it does not include the contribution from the vl band). If the discrepancy between the two sets of data persists, then a detailed comparison of the present experimental technique with that of McCaa and Shaw must be undertaken. A computer program is presently being developed which will estimate the variation in stratospheric flux divergence studies due to uncertainties in the experimental absorptivities. VI. Study of the Spectra of CO2 Isotopes (J. B. Russell) This work on the isotopic spectra of CO2, the PhD dissertation of J. B. Russell, is being done in absentia. Four of the six programs necessary to perform a line position analysis of the CO2 isotopes have been written to be compatible with both the University of Michigan Computer System and the U. S. Naval Avionics Facility, Indianapolis (NAFI) Computer System. The four are as follows. One program calculates line positions in the terms of cm from spectrometer drum turn positions and lines whose position in cm is known well enough already to be used as calibration lines. Another program analyzes the line positions of the Q-branches to determine the wavenumber of the band center and the two rotational constants B and D. The other two programs 7

analyze the P and R branches, one by analyzing them separately and thel( otlicr by analyzing 1.lhe sluml of the P and R line positions. The resutllt of these analyses is also the wavenumber of the band center and two rotational constants. Yet to be written are the programs which analyze the P and R lines in ir - 7r and higher order bands where a total of 5 unknowns are possible, the band center and four rotational constants (2 B's and 2 D's). Analysis of the 3C2O isotope has been started. Tlhe nmajor pa't of the work on this isotope should be completed by June 1. T'l'hl remriailing two programs will therefore also be completed by this date. Turn around time for computer runs at NAFI has improved greatly in the last few months. VII. References: 1. Hearn, A. G. 1961: Proc. Phys. Soc. 78, 932 2. Walshaw, C. D. 1957: Quart. J. Roy. Meteor. Soc. 83, 315. 3. McCaa, D. J., and J. H. Shaw, 1967: The Infrared Absorption Bands of Ozone, Sci. Rep. No. 2 Project No. 7670, Ohio State University, 94 pp. 8

Table I. Comparison of Absorptivities Pressure (mm Hg) Mass path (cm STP) Walshaw 9. u m Walshaw-Present Walshaw 7. 5 11.3 11.5 11. 9 12.4 12.5 13.8 16.4 21. 6 23. 3 24. 0 27.0 32. 34. 9 35. 7 40. 1 40. 8 44. 5 47.2 47. 5.141 588 069.500 029 9. 4 1.097 1.002 440.213.103 296.900 1.450 ~ 630.096 351 2. 5 2. 37 1.878 9. 6tm 9. 0 m 10.1.93 10.3.70 26.8 2.1 26.6 2.3 6. 6.31 6.6.26 28.0 2.5 27.8 2.8 2.9.20 2.8.24 91. 7 31.8 93.0 38.2 39.8 3.6 39.6 4.1 43.6 5.1 43.3 5.7 29.5 1.7 29.3 1.9 19.6.58 20.0.24 11.9.55 12.2.26 24.3 1.1 24.3 1.1 43.3 3.7 43.3 4.0 55.2 6.3 54.9 7. 3 38.4 2.4 38.3 2.6 13.1 1.6 13.4 1.3 30.9.13 31.0 67.2 10.1 67.2 11.6 67.5 11.0 67.6 12.8 63.6 8.3 63.4 9.6 11.0 28. 9 6.89 30. 5 3. 1 123. 5 43. 4 48. 6 31.2 20. 2 12.4 25. 4 47. 0 61.5 40. 9 14. 7 31.1 77. 3 78. 6 71.8 13.02 31.07 10. 42 30.48 5. 78 46. 32 35.82 24. 6 16. 35 30. 61 52. 18 46. 66 17.94 38. 22.22.14 37.08.50.14.18.20.27.21.17.18.27.19 Total

Table I (Continued) 69.7 11.5 51.7 2.6 69 7 1 81.2 69.8 13. 4 54.8 3. 79 52.3 1.310 837 58.6 63.73. 14 41.8.92 53.5.599 419 42.7 49.4.15 41. 9 75 55. 2 5. 2 60.4 1.330 522 60.4 65.13.15 55.1 5.7 74.8 13.5 67.3 3.0 7. 15. 88. 3 75. 0 15.8 729 11.7 73.7 2.63 729 117 84. 6 72. 9 13. 7 74.2 1.016 53 3.3 56. 5 62.90.15 53. 3 3. 2 72.5 12.2 76.5 3.4 72184.6 72. 6 14.60 72. 9 12. 3 76.6 2.68 73. 12 85.2 73.0 14.2 64. 8 7. 9 83.9 1. 97 648 72. 7 64.8 9.079.2. 84.7 2.4 703 878.3 70. 1 909.0 77.9 16.2 91. 4 3. 8 94. 1 91.4 3.8 78. 2 19. 2 94 75.3 14.0 113. 3.2 75.589. 2 75. 5 ~16. 3 119. 4.57 81.9 189 100.8 82. 1 22.8 121. 5.23 84. 9 2 103.2 84. 7 22.4 82.3 17.6 99 9 132. 4.8 82.7 99 9 82. 7 20.8 134. 5.4 814 18.6 100. 0 81.8 22.0 146. 2.65 7149 2 80.6 71.3 10.6 81.5 19.1 163. 5.1 81 100. 7 81.6 22. 9 193. 6.7 874 247 112.1 88. 1 29.4 90. 1 27. 2 197. 8.4 91 117. 3 191. 2 32.4 113 198. 6.8 88.5 24.7 113.2 89.1 29.6

Table I (Continued) 202. 226. 5 258. 282. 8.5 89.0 90.1 44. 6.440 44. 5 9.4391 7 93. 90. 9 7 52 9 9 91. 9 7. 9907 91.6 85. 4 4. 6885 85.5 93. 6 10. 94.8 10 04 94.8 92. 7 10. 04 93. 7 27. 0 32.2 2. 7 3. 0 31.8 38.2 27. 1 32.4 28. 7 34. 3 16.6 19. 9 35. 2 42.4 34. 6 41. 9 116. 1 47. 3 123. 5 118. 119.4 102. 128. 8 127. 3 303. 328. 332. 432.

70 60 E 50 0 To _ 40 ' 20 D 10 </> 0 -n2 - 0o 10 20 30 40 50 60 70 Absorptivity (Walshaw), cm1 Figure 1. Comparison of absorptivities

UNIVERSITY OF MICHIGAN 3 9015 02227 1442