of Reinvestigation the 16O2 atmospheric A band by high-resolution Fourier transform spectroscopy Sophie Fally1 ([e-mail]), Univ. Libre de Bruxelles, Belgium
1
C. Hermans2, A. C. Vandaele2, Michel Carleer1 ([e-mail]), L. Daumont3 ([e-mail]), A. Jenouvrier3
SUMMARY
2 Institut
http://www.ulb.ac.be/cpm d'Aronomie Spatiale de Belgique, Belgium http://www.aeronomie.be 3 Univ. de Reims GSMA, France http://helios.univ-reims.fr/Labos/
WHAT? Line parameters of the oxygen atmospheric A band b1+g (v'=0) X 3-g (v"=0) (12800-13400 cm-1 or 780-745 nm). HOW ? A Fourier transform spectrometer (FTS) coupled to a T-controlled multipass cell 5 pure O2 & 2 O2+N2 mixtures spectra @ room T. HOW MANY ? A linelist of 58 lines including calibrated wavenumbers and linestrengths, also with self- and N2-broadenings and shifts for most of them. CONCLUSIONS: Interesting results for self- and N2- broadenings at high J" Still questionable results due to weakness and overlap of lines Large error bars and scatter of results precision and accuracy are not improved compared to existing data.
INTRODUCTION : THE PROBLEM ?
The scatter among published O2 line parameters is significantly larger than the sub-% precision required to improve spaceborne CO2 measurements by photon paths lengths retrievals [1, 2, 3]. Also, cloud retrievals results from satellite measurements can vary up to 20% depending on the chosen O2 dataset [4].
DISCUSSION: Comparison with literature data
n Hitran '04 Brown & Plymate, '00 Schermaul & Learner, '99 Cheah et al., '00 Ritter & Wilkerson, '87 O'Brien et al., '01 Hill et al., '03 Philips& Hamilton, '95 van Leeuwen et al., '04, Yang et al., '00 91 44 67 61 54 80 37 53 12 11 58 Wvnb 91 44 67 61 0 80 0 0 12 0 58 Intensity 91 44 65 59 54 0 0 0 12 11 58
0.040
gL self 91 44 65 0 54 0 0 0 12 3 53
A band b-X (0,0)
gL air 91 42 0 0 0 0 0 0 0 4 49
Pshift self 0 42 0 0 0 0 37 53 0 0 53
Pshift air 91 41 0 0 0 0 0 37 0 0 48
Brown & Plymate, 2000 Schermaul & Learner, 1999 Cheah et al., 2000 van Leeuwen et al., 2004 O'Brien et al., 2001 O'Brien + Brown This work
INTRODUCTION : THE AIM OF THIS WORK ?
Contribute to spectroscopic parameters improvements needed for atmospheric remote sensing applications. Attempt to validate reported values. Fulfill the need for new laboratory measurements in case of sparse existing data like pressure-shifts.
EXPERIMENTAL
Experimental conditions
Spectral range (cm-1) 8000-16000 Resolution (cm-1) 0.02 Path length (m) 61 Temperature (K) 293 & 220 O2 pressure (hPa) 20, 40, 80, 200, 400 N2 pressure (hPa) 320, 800 Lamp & detector W & Si diode Co-added scans 8 x 64
Set-up
FTS IFS120M
Cooling system
This work
Positions
Nu(HITRAN) nu(literature)
5 m long T-controlled multipass cell
0.030
Average 0.0004 0.0010 0.0028 -0.0042 -0.0022 0.0082 -0.0015
Std dev 0.0018 0.0078 0.0100 0.0095 0.0079 0.0011
-0.020 nu (HITRAN) - nu(literature)
Brown & Plymate, '00 Schermaul & Learner, '99 Cheah et al., '00 O'Brien et al., '01 O'Brien + Brown van Leeuwen et al., '04, This work
0.0013
0.020
0.010
Procedure & data processing
8 x Postzerofill Spectrum divided by Blank to eliminate the atmospheric contribution Spectrometer alignment regularly checked Voigt line shape and baseline fitting using WSpectra [5] Wavenumber calibration using I2 [6, 7] and atmospheric O2 lines Careful examination of spectroscopic data weak, unresolved, saturated lines excluded, outliers eliminated Line parameters determination using conventional equations [8] and linear least-squares fits [9]
0.000
-0.010
Fig. 6 Fig. 6 & table: -0.030 0.E+00 1.E-24 2.E-24 3.E-24 4.E-24 5.E-24 6.E-24 7.E-24 8.E-24 This work shows a slight negative but systematic Intensity HITRAN difference with HITRAN, inferred to the multi-step calibration procedure The absolute difference is of the same order of magnitude of the uncalibrated positions given by Schermaul The scatter logically increases for decreasing line intensities
1.8
9.E-24
Intensities
Intensity ratio HITRAN / literature
A band b-X (0,0) 1.6
RESULTS Set of Measurements
Fig. 1: Overview of the oxygen A band absorption (arb. units) at the highest and the lowest pressure of pure oxygen (see table for experimental details). Fig. 1: 400 hPa & 20 hPa of O2
Fig. 7: Average ratios are written in ( ) Similarly to wavenumbers differences, the scatter logically increases for decreasing intensities
1.4
Brown line & Plymate, 2000 (1.00) Schermaul & Learner, 1999 (0.99) Cheah et al., 2000 (1.17) van Leeuwen et al., 2004 (1.12) Ritter & Wilkerson, 1997 (1.00) Yang et al., 2000 (1.02) This work (1.01)
1.2
1.0
0.8
1.5 1.4 1.3 1.2 1.1 1.0 0.9 0.8
0.6
Voigt versus Galatry line profile ?
Fig. 2: Comparison of line fitting obtained with Voigt (blue) and Galatry (green) line profiles. The observed spectrum @ 80 hPa (divided by blank) is in black. The Observed - Fitted residual is shown below (amplitude +/- 5%). This line corresponds to 88% of absorption. Voigt and Galatry fits are undistinguishable; The characteristic w-shaped residual due a Galatry profile modeled with a Voigt lineshape is not observable, and the line is not perfectly fitted neither by a Voigt, nor by a Galatry lineshape. Line intensity ratio Voigt / Galatry = 99.3% for this line, but values down to 97.1% have been measured for intense (Absorption> 85%) neighboring lines.
0.4
0.7 0.6
Fig. 7
0.2 0.E+00 1.E-24 2.E-24 3.E-24
0.5 0.0E+00
5.0E-26
1.0E-25
1.5E-25
2.0E-25
4.E-24 5.E-24 Intensity HITRAN
6.E-24
7.E-24
8.E-24
9.E-24
0.07
Self broadenings
Fig. 8: Still large errors for some lines This work extends Brown's data at high J" The divergence from the Hitran values is confirmed, in agreement with the data from Yang and Ritter & Wilkerson, also as well as with Yang's fit. Similar trends are observed for the PQ and RQ-type transitions as well as for the N2-broadenings although fewer literature data exist in the latter case.
0.06
A band b-X (0,0) P (-N") & R (+N") branches
Hitran, 2004 Brown & Plymate, 2000 Schermaul & Learner, 1999 Ritter & Wilkerson, 1987 Yang et al., 2000 Yang et al., 2005, fit This work
Fig. 2
gamma self (cm-1 atm-1)
0.05
P-induced effects
Fig 3: Pure O2 spectra at different pressures (0.02, 0.04, 0.08, 0.2, 0.4 hPa) clearly showing selfbroadening and shifting. Also shown is a blank spectrum containing the external atmospheric O2 contribution. Fig 4: N2-broadening and shift. Note the lineshape `degradation' at the highest O2+N2 pressure explaining the difficulty of obtaining high precision parameters. Fig. 5: Examples of good (RR(1), RR(3)) & less good (PP(17)) results for the pressure dependence of the Lorentzian linewidth.
0.08 hPa O2 + 0.32 hPa N2 0.2 hPa O2 + 0.79 hPa N2
0.04
0.03
Fig. 8
0.02 -45 -40 -35 -30 -25 -20 -15 -10 -5 0 N" 5 10 15 20 25 30 35 40 45
Fig. 4 Fig. 3
0.06 y = 1.20E-04x R2 = 9.99E-01 y = 1.08E-04x R2 = 9.96E-01
0.000 O2 N2-pressure shift for the A band (b
1 + g (0,0))
0.08 hPa O2
0.2 hPa O2
N2-shifts
Fig 9: Few literature data This work g...