1. SPECTRAL LINE PARAMETERS FOR THE 9 BAND OF ETHANE Malathy Devi & Chris Benner, W&M Rinsland & Smith, NASA Langley Bob Sams & Tom Blake, PNNL Jean-Marie Flaud, CNRS, France Keeyoon Sung & Linda Brown, JPL Arlan Mantz, Connecticut College

2. Why we did? What we did? How we did? Challenges

3. WHY ETHANE? WHY THE 9 BAND? 1. 12-µm emission features of ethane are seen in the spectra from outer solar system bodies of Jupiter, Saturn, Neptune and Titan. 2. The 9 band, especially the R Q 0 sub-band ( 0 =~ 822 cm -1 ) is of considerable interest due to its importance in molecular astrophysics and also because it is often used in remote sensing applications. 3. Laboratory measurements are required to convert the raw observational data of planetary observations into information useful for quantification,. 4. laboratory measurements normally include parameters such as line positions, intensities, pressure-broadened widths and shifts as a function of temperature. The purpose of the present investigation is to provide new and accurate measurements of individual spectral line parameters for R Q 0 AND several other Q, P, R sub-band transitions. Titan’s atmosphere consists predominantly of N 2 and measurable quantities of several organic molecules including ethane. Present investigations involve spectra of ethane and ethane broadened with nitrogen at various temperatures (298 K to 149 K), pressures and absorption path lengths.

4. Comparison of a Spectrum of Titan to our Laboratory Spectrum Titan’s Atmospheric Spectrum showing the C 2 H 6 features. Courtesy of: Henry Roe, Lowell Observatory A laboratory spectrum recorded with the Bruker 125HR FTS at JPL R Q(J, 2) Sub- band is shown in both figures.

5. 43 high resolution spectra with sample temperatures between 149 K and 298 K are fitted simultaneously. (17 pure C 2 H 6 and 26 C 2 H 6 +N 2 spectra). Temperature (K) Gas MixtureC 2 H 6 Volume Mixing Ratio Path (cm)Pressure Range (Torr) Number of Spectra 297.2C2H6C2H6 1.0324.0±2.00.31 298.2C2H6C2H6 1.020.00 ±0.024.5-365 297.2C 2 H 6 +N 2 0.011-0.045324.012-575 298.2C 2 H 6 +N 2 0.08-0.220.0030-1816 273.2C2H6C2H6 1.020.003.8- 163 273.2C 2 H 6 +N 2 ~0.220.0026-533 248.2C2H6C2H6 1.020.004-63 248.2C 2 H 6 +N 2 ~0.220.0026-503 223.2C2H6C2H6 1.020.004-63 223.2C 2 H 6 +N 2 ~0.220.0025-503 211.0C2H6C2H6 1.020.004-63 211.0C 2 H 6 +N 2 ~0.220.0025-503 148.3C2H6C2H6 1.020.387±0.0016.21 149.7C 2 H 6 +N 2 ~0.1720.38734.5541

6. Top: A spectrum of pure (99.5%) C 2 H 6 at 149 K with 6.2 Torr in a 20.387 cm cell Bottom: A low temperature (149 K) N 2 - broadened C 2 H 6 spectrum with a volume mixing ratio of 0.17

7. 1. We applied a multispectrum fitting technique a to simultaneously fit all 43 spectra of C 2 H 6 in the 9 band. High resolution (0.0016-0.005 cm -1 ) spectra were recorded with two different FTS (PNNL and JPL). 2. Intensities & separations for torsional split components, widths and their temperature dependences were constrained in the analysis to determine the line parameters for both torsional split components. No pressure shifts needed to fit. 3. Accurate line positions and absolute intensities were retrieved for over 1700 transitions of 9. N 2 - and self-broadened half width coefficients and their temperature dependences were also obtained for more than 1350 transitions at various sample temperatures between 149 K and 298 K. 4. Variations of the observed widths and their temperature dependences with respect to J, K quanta are discussed. Present results are compared with previously reported measurements and calculations. ====== a D. Chris Benner, C.P. Rinsland, V. Malathy Devi, M.A.H. Smith, and D.A. Atkins. JQSRT 1995;53:705-721.

8. A few sub bands near the prominent R Q 0

9. R Q 0 sub-band. (a) Low pressure C 2 H 6 spectrum illustrating the high density of lines. (b) N 2 -broadened C 2 H 6 spectrum. Features of Individual J transitions are completely obscured. Multispectrum technique allows fitting such spectra to great advantage. Only two of the 43 spectra used in the analysis are shown here.

10. 0.1 cm -1 wide interval in R Q 0 sub-band. a) low-pressure spectrum with L= 324 cm; P= 0.3 Torr C 2 H 6 at T= 297.2 K Each transition has two components with different J, symmetry species and intensity. (b) C 2 H 6 +N 2 spectrum with P(total)=180 Torr, L=20 cm; T=298.2 K.

11. Examples of Torsional Splittings & Statistical Weights

12. R Q 0 sub- band A few sample spectra of C 2 H 6 +N 2 recorded at various temperature s pressures and paths are shown

13. Different sub-band series ( P Q, R P, P P) Effects of temperature and pressure upon the relative strengths of the various transitions

14. SOME ANALYSIS DETAILS For the PNNL spectra, the calibrations of the wavenumber scales were done relative to OCS line positions from the NIST wavenumber Ref. data [Maki & Wells] For the JPL spectra the calibration was achieved with respect to positions of 2 water vapor transitions. The initial values for line positions, intensities, N 2 - and self-Widths taken from updates to HITRAN2008 database. 1. Positions, intensities, N 2 - and Self Widths, T-dependences of N 2 - and Self Widths measured for over 1330 transitions in 17 Q sub-bands (and several P P, R R, R P and P R sub-bands). 2. Standard Voigt line profile was sufficient (no line mixing or speed dependence) to fit individual manifolds in all sub-bands. 3. Measured N 2 - and Self-Widths vs. J and K quanta are studied. 4. Temperature dependence exponents for N 2 - and Self Widths vs. M (M = J′=J″ for Q sub-bands) are discussed.

15. In the global least squares fits, the pressure-broadened half-width coefficients were determined using the expression: b L (p, T) is the Lorentz half-width (in cm -1 ) of the spectral line at pressure p and temperature T. b L 0 (Gas)(p 0, T 0 ) is the Lorentz half-width coefficient of the line at the reference pressure p 0 (1 atm) and temperature T 0 (296 K) of the broadening gas (either N 2 or C 2 H 6 in this case).  is the ratio of the partial pressure of C 2 H 6 to the total sample pressure in the cell.

16. 43 spectra fitted simultaneously. Tick marks at the top correspond to all transitions (>430) included in the fit. HB=HOT BAND ( 9 + 4 - 4 ) No pressure- induced shifts, line mixing or speed dependence required to fit the spectra.

17. R Q 0 sub-band head. The top panel shows the fitted spectra and the bottom panel represents the weighted fit residuals on a magnified vertical scale. Overlap & blending of fundamental and hot-band transitions further complicates the analysis.

18. Well separated J,K transitions in this P P series vary from 7 to17(J″) and 1 to 6 (K″). The torsional split components overlap at high pressures. Hot-band transitions ( 9 +  -  ) are marked with (*) No pressure- induced shifts were required to fit even these well-separated lines.

19. LINE INTENSITIES IN C 2 H 6 9 BAND 1. Intensity measurements in the most recent HITRAN and GEISA databases are based upon analysis by J. Vander Auwera, N. Moazzen- Ahmadi, Jean-Marie Flaud. Astrophys J. 2007;662:750-757. Those values are used as initial input in the present analysis. 2. Line Intensities measured in this work are lower by ~15% from HITRAN 2008 values. 3. Line intensities in the databases were normalized to the band intensity from a medium resolution spectrum recorded at PNNL based on the integrated area under the entire region of the band. 4. Intensities are based upon 43 high-resolution spectra recorded by TWO different Bruker FTS (PNNL and JPL). Measurements are obtained fitting all spectra simultaneously. Absolute uncertainties in intensity measurements are estimated to about ±5%. Differences from the databases are probably due to the difficulty measuring the intensities in medium resolution spectrum?

20. Triangles and circles are the measured widths from multispectrum fit. The stars correspond to widths that are calculated from constrained n by empirical linear fits: n= a+b  (J-c)

21. Measured N 2 - and Self-Widths vs. M (M=J′=J″ for Q sub-bands). Units of widths are cm -1 atm -1 at 296 K. LEFT: r Q bands RED: Self-Widths BLUE: N 2 -Widths RIGHT: p Q bands RED: Self-Widths BLUE: N 2 -Widths Ratio of Self- Widths to N 2 - Widths =1.40±0.05

22. Measured values of n from multi- spectrum fit.

23. Measured T- dependence exponents (n) of N 2 - and Self-Widths vs. M (M=J′=J″) in 17 Q sub-bands. LEFT: r Q Sub-bands RED: Self-Widths BLUE: N 2 -Widths RIGHT: p Q Sub- bands RED: Self-Widths BLUE: N 2 -Widths n for Self- Widths < n for N 2 -Widths

24. Temperature dependence exponents For N 2 - and Self Widths obtained fitting an empirical linear equation

25. “Fitted” Temperature dependence exponents (n) for N 2 - and Self- Widths vs. M (M=J′=J″ for Q sub- bands). LEFT: r Q Sub-Bands RED: Self-Widths BLUE: N 2 -Widths RIGHT: p Q Sub- Bands RED: Self-Widths BLUE: N 2 -Widths n for self-Widths < n for N 2 - Widths

26. Measured N 2 - and Self-Widths vs. M (M=J′=J″ for all Q sub-bands). RED: Self-Widths BLUE: N 2 -Widths “Fitted” n using an empirical linear Equation: n= a + b  (J-c) BLUE: N 2 -Widths RED: Self-Widths Mean N 2 - to Self-Width = 1.40±0.05

27. (a)N 2 -Widths vs. K″ BLUE: p Q Sub-Bands RED: r Q Sub-Bands (b) Self-Widths vs. K” BLUE: p Q Sub-Bands RED: r Q Sub-Bands K″+0.05*(J″-K″) is used for pattern recognition Half-width coefficients are in Units of cm -1 atm -1 at 296 K

28. Values for a and b using n=a+b  (J-c), and the half-width coefficients for the highest J (=M=J″=J′ in Q sub-bands) for 9 measured r Q sub-bands. Sub-band For N 2 broadening For self broadening For N 2 & self broadening abWidth at J ab c and J p Q(J,K=9)0.884(5)0.0076(11)0.998(17)0.592(6)0.0092(12)0.731(18)16, 31 p Q(J,K=8)0.799(5)0.0096(12)0.933(18)0.591(6)0.0109(13)0.744(19)16, 30 p Q(J,K=7)0.841(3)0.0131(7)1.02(12)0.647(5)0.0115(8)0.809(13)16, 30 p Q(J,K=6)0.878(3)0.0158(6)1.10(10)0.694(5)0.0172(7)0.936(11)16, 30 p Q(J,K=5)0.787(3)0.0131(5)1.06(1)0.579(4)0.0137(6)0.867(13)13, 34 p Q(J,K=4)0.821(2)0.0087(3)0.987(7)0.650(4)0.0079(5)0.800(9)13, 32 p Q(J,K=3)0.878(2)0.0162(4)1.18(7)0.792(3)0.0085(5)0.964(9)13, 32 p Q(J,K=2)0.781(2)0.0054(3)0.895(7)0.690(4)0.0041(4)0.775(9)13, 34 p Q(J,K=1)0.808(2)0.0123(3)1.04(6)0.706(3)0.0074(4)0.847(9)13, 32 The half-width coefficients are in units of cm -1 atm -1 at 296 K.

29. Comparison of Temperature Dependences (n) of Self-Widths: This Study vs. Nguyen et al. (J. Mol. Spectrosc. 2008;39:429-434) a This study. The error bars for positions and temperature dependence exponents are twice the standard deviation. b Nguyen et al. Reported temperature dependence exponents were calculated from their measured self- broadened half-width coefficients at three different temperatures (242.2, 226.2 and 150.2 K). Line  (cm -1 ) a n (This work) Voigt profile a n Rautian profile b p Q(17,9)798.93257(1)0.602 ± 0.0120.636 ± 0.123 p Q(16,9)798.97339(1)0.592 ± 0.0120.620 ± 0.104 p Q(15,9)799.01179(1)0.583 ± 0.0120.621 ± 0.110 p Q(13,5)809.13646(1)0.579 ± 0.0080.659 ± 0.119 p Q(12,5)809.16832(1)0.566 ± 0.0080.707 ± 0.113 p Q(11,5)809.19773(1)0.552 ± 0.0100.643 ± 0.133 p Q(7,5)809.29055(1)0.497 ± 0.0120.669 ± 0.104 p Q(6,5)809.30729(1)0.483 ± 0.0140.742 ± 0.098 p Q(13,2)816.86769(1)0.690 ± 0.0080.735 ± 0.099 p Q(12,2)816.90092(1)0.686 ± 0.0080.758 ± 0.128 p Q(11,2)816.93136(1)0.682 ± 0.0080.664 ± 0.115

30. SUMMARY & CONCLUSIONS 1. 43 high-resolution (0.0016-0.005 cm -1 ) spectra of pure and N 2 -broadened C 2 H 6 are fitted simultaneously to retrieve: 2. Positions, absolute intensities, N 2 - and self-Widths and the temperature dependences of N 2 - and self- Widths measured for over 1300 J, K transitions in 17 Q sub-bands and several p P, r R, r P and p R sub-bands. 3. Mean Ratio of Self-Widths to N 2 -Widths is 1.40±0.05 4. No pressure-induced shifts, line mixing or speed dependence were detected in the spectra. (PTO)

31. SUMMARY & CONCLUSIONS T-dependence exponents for N 2 -Widths are larger than for self-Widths and their values as well as ratios vary with J, K quanta. The “n” values are fitted to empirical linear relationship: n = a+b ✕ (J-c), for each broadening gas in each sub-band structure. Line intensities from this work are found to be ~15% lower than those listed in the HITRAN 2008 Database. The reason for this discrepancy has recently been investigated from measurements on a (new) spectrum recorded at JPL and the new intensities confirm our present measurements.

32. C 2 H 6 : Keeyoon (JPL) recorded a (normal sample) ethane spectrum at room temperature in the 12 µm region and Linda Brown compared it with a synthetic spectrum using the HITRAN 2008 linelist (a rough estimation). The cell path of 14.93 cm needed adjustment to 13.25 cm to fit the observed spectrum with the calculated one.

33. Importance of Present Laboratory Investigation 1. Using the parameters determined in this study the spectrum of ethane can be computed at any temperature below 296 K at infinite resolution. 2. From there one simply applies the instrument line shape to compare to a spectrum of Titan. 3. The importance of the intensity problem is that it affects the retrieved mixing ratio on a one to one basis. 4. Lorentz widths are temperature and pressure dependent and we have described that dependence with great accuracy. 5. We have also got enough information (except for the unidentified lines) to describe the line intensity as a function of temperature-and no new lines will appear at lower temperatures than presently obtained (149 K).

34. ACKNOWLEDGMENTS The experimental spectra for the present study were recorded at the W. R. Wiley Environmental Molecular Sciences Laboratory, a national scientific user facility sponsored by the Department of Energy’s Office of Biological and Environmental Research located at Pacific Northwest National Laboratory (PNNL) and the Jet Propulsion Laboratory (JPL) in Pasadena, California. PNNL is operated for the United States Department of Energy by the Battelle Memorial Institute under Contract DE-AC05- 76RLO1830. NASA’s planetary atmospheres program supported the work performed at NASA Langley Research Center and the College of William and Mary. The research at the JPL and Connecticut College was performed under contracts and grants with NASA.