1. On the impact of Vibrational Raman Scattering of N2/O2 on MAX-DOAS Measurements of atmospheric trace gases
Johannes Lampel1, Johannes Zielcke2, Udo Frieß2, Ulrich Platt2 and Thomas Wagner1
Max Planck Institute for Chemistry (MPIC), Mainz, Germany
Institute of Environmental Physics, University of Heidelberg, Germany
contact:
Theory
Abstract
Inelastic scattering results in a change of the energy of the scattered photon and of the scattering molecule.
The energy of the molecule E(v,J) is characterized by the vibrational (v) and rotational (J) quantum numbers.
The RRS and VRS cross-sections depend on molecular properties of N2 and O2, such as the polarisability a and the anisotropy γ and their respective derivatives a‘ and γ‘, which also directly determine the phase function.
Here the total cross-section for the Q-Branch (ΔJ=0;Δ v=1) is shown. Numerical results are shown in Table1.
Important for DOAS applications are the λ-4 dependency and the overall shift in wavelength. S is almost constant and depends only slightly on temperature.
A detailed description of Raman scattering can be found e.g. in [Long 2002], a more detailed description focusing on DOAS measurements in [Lampel et al. 2015].
In remote sensing applications, such as differential optical absorption spectroscopy (DOAS), atmospheric scattering processes need to be considered since they can modify the observed spectra. Inelastic scattering of photons by N2 and O2 molecules can be observed as additional intensity, effectively leading to filling-in of both, solar Fraunhofer lines and absorption bands of atmospheric constituents. The main contribution is due to rotational Raman scattering, which can lead to observed absolute optical densities of up to several percent. Measured optical densities are typically corrected for this effect (also known as Ring Effect).
In contrast to that Vibrational Raman scattering of N2 and O2 was often thought to be negligible, but can affect DOAS measurements in a similar way.
We present calculations of Vibrational Raman cross-sections for O2 and N2 for the application in passive DOAS measurements. Consequences of vibrational Raman scattering are red-shifted Fraunhofer structures, so called ‘Fraunhofer Ghost’ lines (FGL), in scattered light spectra and filling-in of Fraunhofer lines, additional to rotational Raman scattering. We also present first unequivocal observations of FGL at optical densities of up to several From our measurements and calculations of the optical density of these FGL, we conclude, that this phenomenon has to be included in the spectral evaluation of weak absorbers.
In agreement with calculated cross-sections the magnitude of the measured VRS(N2) cross-section is about 2.5% as large as the cross-section of RRS of air.
Fig.1: Phase function of RRS, VRS and Cabannes scattering.
Experimental Evidence from MAX-DOAS observations
VRS correction spectra can be calculated for DOAS applications by a Taylor expansion of the optical depth, assuming an additional intensity Ji for VRS of N2 and O2.
The Ring spectrum was calculated according to [Bussemer1993] in DOASIS [Kraus2006].
Fig.2: Cabannes (ΔJ=0;Δ v=0), RRS and VRS cross-sections.
To illustrate the difference to liquid water VRS, also its cross-section is plotted, scaled arbitrarily.
Fig.3: Intensity of scattered light due to Cabannes, RRS and VRS of N2, O2 and liquid water.
without VRS
Data recorded at Penlee Point, Plymouth, England on May 30th, :30 UTC, SZA=52°, RAZI=124°, Elevation 3°. 8 co-added elevation sequences, one minute per spectrum, using an Avantes ULS compact spectrometer with BG3 filter.
Detection of VRS below 360 nm
Easy detection of VRS at 430nm, due to pronounced Fraunhofer ghost lines of the strong Calcium lines (at 393 and 396nm)
The λ-4 wavelength dependence is the same for VRS and RRS. Despite less pronounced Fraunhofer lines in the UV range it should be nevertheless contribute significantly to observed optical depths in MAX-DOAS measurements, according to its cross-section shown in Table 1.
For very good fits (RMS<10-4), similar correlation plots of VRS(N2) with RRS as for the blue wavelength range are observed (for SZA<60°). The VRS(N2) correction spectrum was calculated from several noon-time zenith sky spectrum.
For the spectral range from nm, the shift of VRS(N2) also shifts e.g. ozone absorption into this window. This probably leads to erroneous results for SZA>60°.
MAX-DOAS observations from different campaigns were analyzed, using calculated VRS correction spectra: Using an Acton300i a clear signal was found in data from SOPRAN M91 (Peruvian upwelling, ship), when co-adding 16 elevation sequences. The magnitude of the VRS signal was correlated well with RRS (see above).
In order to minimize the possibility of cross-talk with liquid water VRS, the MAD-CAT campaign (Mainz, Germany) data recorded with a compact Avantes spectrometer was analyzed. Despite large NO2 absorption an agreement in relative cross-section size for VRS and RRS was found.
Impact on trace gases: All trace gas retrievals profit from reduced fit errors (up to 20%, at typ. RMS ). For IO ( nm) and Glyoxal ( nm) no significant changes in dSCDs are observed, for NO2 dSCDs ( nm) biases of up to molec/cm2 (Fig. 4) were found.
Fig. 6: VRS contribution to OD between nm. The data was filtered for an RMS < 10-4
Table 1: VRS(N2/O2) and RRS(air) cross-sections (integrated over all possible contributing transitions)) are calculated for different scattering angles θ (compare Fig. 1).
On the right side of the table experimentally determined values are shown from M91 and MAD-CAT. Within their respective errors they agree and are also in agreement with expected values from theory.
A possible work-around would be to calculate a VRS correction spectrum from each zenith sky spectrum (assuming a constant slit function) or correspondingly modelled spectra in order to account for the non-linear intensity variations below 330nm. Δ
Literature
Lampel et al (2015): The impact of vibrational Raman scattering of air on DOAS measurements of atmospheric trace gases, Atmospheric Measurement Techniques Discussions, 8, 3423–3469
Long, D. A. (2002). The Raman Effect: A Unified Treatment of the Theory of Raman Scattering by Molecules. Wiley Online Library
Haug, H. (1996). Raman-Streuung von Sonnenlicht in der Erdatmosphäre. Diploma thesis, Institut für Umweltphysik, Ruprecht-Karls-Universität Heidelberg.
Zielcke, J. (2015); Observations of reactive bromine, iodine and chlorine species in the Arctic and Antarctic with Differential Optical Absorption Spectroscopy, Dissertation, University of Heidelberg
Bussemer, M. (1993). Der Ring-Effekt: Ursachen und Einfluß auf die spektroskopische Messung stratosphärischer Spurenstoffe. Diploma thesis, Heidelberg University, Heidelberg, Germany.
Kraus, S. (2006). DOASIS - A Framework Design for DOAS. Dissertation, Heidelberg University.
Fig. 4: Impact of VRS on NO2 dSCDs for the M91 data set for telescope elevation angles below 10°.
The difference was calculated for a fit range from nm with and without correction spectra for VRS(N2/O2). Including the VRS correction increased the average dSCD for clear air measurements from to molec/cm2. The blue error-bar shows a typical measurement error.
Fig. 5: DOAS fit from Alert/Canada:
In the remote and pristine environment of the Arctic, the Fraunhofer reference spectrum was recorded at 90° elevation. 64 elevation sequences were co-added, resulting in a time resolution of 17 hours and a total exposure time of about 2 hours.
Here this individual fit already agrees reasonably well with expected values from the table above.
(See also [Zielcke 2015])
Acknowledgements:
We thank H. Haug for laying the foundations for this work in his diploma thesis (Haug, 1996). We thank the captain, officers and crew for support during research cruise M91. We want to thank the organizers of the Multi Axis Doas - Comparison campaign for Aerosols and Trace gases (MAD-CAT) in summer We thank the German Science foundation DFG for its support within the core program METEOR/MERIAN. We thank the German ministry of education and research (BMBF) for supporting this work within the SOPRAN (Surface Ocean Processes in the Anthropocene) project (Förderkennzahl: 03F0662F) which is embedded in SOLAS. We thank the authorities of Peru for the permission to work in their territorial waters. We thank GEOMAR for logistic support. We thank Mingxi Yang and Joelle Buxmann for providing the MAX-DOAS data from PML which was used for the results for VRS in the UV.