Rayleigh-Brillouin Scattering in N 2, O 2, and Air

Transcription

1 Rayleigh-Brillouin Scattering in N 2, O 2, and Air Oliver Reitebuch 1, Benjamin Witschas 1, Ofelia Vieitez 2, Eric-Jan van Duijn 2, Willem van de Water 3, Wim Ubachs 2 1 DLR Oberpfaffenhofen, 2 Vrije Universiteit VU Amsterdam, 3 Eindhoven University Institut für Physik der Atmosphäre

2 A SPONTANEOUS RAYLEIGH-BRILLOUIN SCATTERING EXPERIMENT FOR THE CHARACTERIZATION OF ATMOSPHERIC LIDAR BACKSCATTER Ofelia Vieitez, Eric-Jan van Duijn, Wim Ubachs Laser Centre Vrije Universiteit Amsterdam, Netherlands Afric Meijer, Nico Dam, Institute for Molecules and Materials, Radboud University Nijmegen Spontaneous Rayleigh Brillouin scattering experiments Coherent Rayleigh Brillouin scattering experiments Willem van de Water Eindhoven University of Technology, Netherlands + RUN Theory and TENTI modelling; new code Ad Stoffelen, Jos de Kloe KNMI, de Bilt, Netherlands RB measurements on air, + water vapor Anne Straume, Oliver Le Rille ESA Benjamin Witschas, Oliver Reitebuch DLR Consequences for ALADIN Error budget Look up Tables Study was funded by ESA under ITT AO/1-5467/07/NL/HE Research stay of Witschas at VU Amsterdam was funded by EU FP7/

3 Molecular scattering in air - even more than 100 years after Rayleigh still some open issue IASI 5 nm resolution intensity [a.u.] anti-stokes vibration-rotation Raman lines Stokes vibration-rotation Raman lines [nm] intensity [a.u.] anti-stokes rotational Raman lines Stokes rotational Raman lines -1,4-1,2-1,0-0,8-0,6-0,4-0,2 0,0 0,2 0,4 0,6 0,8 1,0 1,2 1,4 [nm] intensity [a.u.] filter A Cabannes line (Hydrodynamic regime) Cabannes line (Knudsen regime) filter B -3,5-3,0-2,5-2,0-1,5-1,0-0,5 0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5 pm]

4 355 nm 2.1 µm 10.6 µm 3 km 10 km ±4 GHz ± 0.6 GHz ± 0.1 GHz B. Rye (1998), Appl. Opt., What is the exact lineshape in air? Fiocco and DeWolf (1968) pointed out the difference of Gaussian to Rayleigh-Brillouin lineshape to lidar community Lidar techniques using molecular backscatter and narrow instrumental bandwidths are affected, e.g. wind, T, HSRL Errors for wind retrievals of ADM-Aeolus would be 3% (10 km) to 10% (ground) if Gaussian is used (Dabas et al. 2008) and thus exceeding specification of 0.7% Widely used models for lineshape from Boley et al. (1972) and Tenti et al. (1974) - the Tenti S6 model - are valid for single species, but not for mixtures like air (N 2 +O 2 ) No experimental validation of Tenti S6 model for air; even Tenti S6 was not compared to N 2 for atmospheric pressures Most (or probably all) use Tenti S6 with N 2 gas parameters => Is there a difference between N 2 and air lineshape? What is the influence of the water vapour molecule (up to 4% in atmosphere)?

5 Setup for spontaneous Rayleigh-Brillouin scattering experiment at VU Amsterdam Choice 4 : No polarizing optics on detection PM Fabry-Perot etalon Choice 2: 90 o scattering L 7 scattered laser beam light L 1 cleaning L 6 S 2 L 5 L 4 L 3 Choice 1 Laser radiation nm narrowband L 2 M1 S 1 S 3 Choice 3 FSR ~ 7.4 GHz Resolution ~ 0.23 GHz light cleaning enhancement cavity reference laser

6 Measurements and Tenti models at 1000 hpa 3.5x10-4 N mbar 3.0x10-4 N 2 O 2 Air Gaussian Tenti S6 Tenti S7 3.5x10-4 O mbar 3.0x Tenti S6 Tenti S7 3.5x10-4 Air 1000mbar 3.0x Tenti S6 (Air model) Tenti S7 (Air model) Gaussian Intensity [normalized on area] 2.5x x x x x ±3.7 GHz Intensity [normalized to area] 2.5x x x x x Intensity [normalized to area] 2.5x x x x x x x x ±12 % Difference measurement-model [percentage] N mbar Tenti S6 Tenti S7 Gaussian Difference measurement-model [percentage] O mbar Tenti S6 Tenti S7 Difference measurement-model [percentage] 12 Air 1000 mbar Tenti S6 (Air model) Tenti S7 (Air model) Gaussian

7 Measurements and Tenti S6 model at higher pressures For higher pressure 3 spectral features become visible: the central Gross line and the Brillouin doublet Tenti S6 is able to resolve these features Gas parameter with highest uncertainty within Tenti model is bulk viscosity, which is obtained from sound absorption measurements (grey curve) Bulk viscosity was used as fit-parameter to minimize measurement-model difference in this study (black curve)

8 Is there a difference between lineshape of N 2 and air? Rel. difference between lineshape for N 2 and air from measurements and Tenti S6 model at 2000 hpa, 297 K, normalized to maximum of lineshape. Most (or all) use Tenti S6 model with N 2 parameters No measurements of lineshape were performed up to now for air (79% N 2, 21% O 2 ) There is a measurable difference between N 2 and air The difference can be modeled with Tenti S6 using appropriate gas parameters as input for molar mass shear viscosity bulk viscosity thermal conductivity

9 Is there an effect of humidity? intensity [a.u.] 4.0 4, , , , , , , ,5 0 % 0 humidity % humidity varying 100 % humidity varying humidity Atmosphere can contain up to 4 % water vapor (100 % rel. humidity at 37 C) Measurements showed no significant difference between lineshape of dry and humid air up to 2.6% water vapor 0.0 0, , ,5-2, , ,0-0, , , , , , , ,0 deviation [%] frequency [GHz] residual 0% - 100% residual 0% - varying% residual varying - 100% signal ~ sqrt(n) No need to consider water vapor within lineshape model for atmospheric conditions frequency [GHz]

10 Summary Spontaneous Rayleigh-Brillouin scattering was measured at nm, at a scattering angle of 90, at ambient temperature ( 300 K), and for pressures of 0.3 bar to 3 bar for N 2, O 2, and air by a setup at VU Amsterdam with a narrowband laser, a scattering cell within an enhancement cavity, and a confocal Fabry-Perot interferometer. For the first time the Cabannes lineshape was measured for dry air (N 2 +O 2 ) and for humid air with up to 2.6 % water vapor content. For the first time the lineshapes of N 2 and air were compared to Tenti S6 and S7 model Conclusion It was confirmed that Tenti S6 is (slightly) better than Tenti S7 Bulk viscosity parameter was varied as model input to minimize measurement to model difference => resulted in a factor of 2 higher bulk viscosity values as reported from sound absorption experiments in literature Deviations between Tenti S6 model and measurements are below +-2% for atmospheric pressures, if appropriate gas parameters are used as model input Water vapor does not influence lineshape - at least up to 2.6% content (compared to a maximum atmospheric content of 4%)