FTUV and FTIRHow to calibrate sigma?How to calibrate sigma?• Determine the absorption cross section (sigma)• Error sources in determining sigma• Limitations in Lambert-Beer’s Law•Resolution Effects•Resolution Effects• Saturation Effects• ApplicationsNov 20 2008Nov 20 2008CHEM 5161Methods to measure the bti ti(i)absorption cross section (sigma)• Pressure calibration• Kinetic calibration• Cross calibration with other method•Calculation from first principlesCalculation from first principlesGlyoxal (CHOCHO) background• Unstable molecule: Cuvet spectra not practicle•Wavelength calibration is not well known•Wavelength calibration is not well known • Only low resolution spectra are available•instrument function unknowninstrument function unknown• limits sensitivity•limits selectivity (NO2, Xe-lamp, Fraunhofer lines)limits selectivity (NO2, Xelamp, Fraunhofer lines)• Nonlinearities have been reported (but not yet explained)•up to 50% scatter in cross section valuesp%• Intercalibration of DOAS, FTIR, TDL, CDR not yet feasible⇒ High resolution sigma (UV/Vis and IR)⇒ Good wavelength calibrationVis absorption cross section1E 181E-181E-19on [cm2]1E-20 IUP HRross secti1E-21IUP convoluted MPI NCARPlum et alC3804004204404604805005201E-22Plum et al. Zhu et al.380400420440460480500520Wavelength [nm]Introduction GlyoxalOH•C2H2O2or CHOCHO• smallest α-dicarbonyl-type compound•Natural sources: fermentation (beer wine yogurt products)OH•Natural sources: fermentation (beer, wine, yogurt products), biomass burning, BVOC oxidation, (oceans ?)• Anthropogenic sources: emissions from mobile sources, AVOC oxidation, (energy sector, industrial processes ?)• In urban air: Airborne AVOC oxidation source >> direct emissions from mobile sources (ca. 70% aromatics, 20% (,alkenes, 10% acetylene, virtually no glyoxal from alkanes) • Residence time in the atmosphere: <1.2 hoursMjhl ihtli•Major gas-phase loss process is photolysis and OH-reaction (60% / 40%). Source for H2, CO, HCHO and HO2-radicals.Experimental setupLi ht•Light sources:– XBO 150W Xe-arc lamp–Glowbar• FTS optics (Table)• Detector (Table)Experimental conditions• White-cell coupled to the Bruker FTS 120 HRabsorption path: L=163 cmabsorption path: L 163 cmGlyoxal pressure: 0.005 – 13 mbar (MKS Baratrons)• Low resolution measurements:short integration times (10 min)Linearity of Lambert-beers law was demonstratedIR and UV integral cross-sections were simultaneously determinedIR and UV integral crosssections were simultaneously determined• High resolution measurements:long integration times (12 hours)IR spectra: unapodized spectral resolution of 0.009 cm-1UV spectra: unapodized spectral resolution of 0.06 cm-1high S/N levelhigh S/N level• Wavenumber accuracy checked from NO2spectra, linked to I2spectraUV/Vis absorption cross section1E-181E-19on [cm2]1E-20ross sectio1E-21Cr1E 2238000 36000 34000 32000 30000 28000 26000 24000 22000 200001E-22Wavenumber [cm-1]Corrections• Lamp drift F=1.01-1.25 (error <1-10%)• Leakage F=1.04 (<1%)• Photolysis F=1.02 (0.2%)• Wall deposition F=1.01 (0.2%)Characterization:• Purity (better 99%)• HCHO formation (<1%) • Reproducible Synthesis (0.2%)• Interpolation method (<2%)Example: 13 µbar glyoxal measured to 1.5% precision!• Polimerization (<0.5%)• Column density (<0.5%)p⇒ error estimate: <5% UV / <3% IR⇒ not better than 2-5 10-22cm2Linearity of Absorptionyp3540ts] IR-Absorption ABS(IR) = -0.007 (0.014) + 67.7 (0.3) x Xts]100120UV-absorption (pure glyoxal)2530n [arb. unitn [arb. unit80100(gy) ABS(UV) = 0.04 (0.04) + 291.2 (1.1) x X UV-absorption (1013 mbar of N2)1520AbsorptionAbsorption6051015ntegral UV-Antegral IR-A20400.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.4005InInGlyoxal Pressure [mbar]0Glyoxal Pressure [mbar]Modelling deviations from LambertBeers LawLambert-Beers Law UV-vis –0.25 nm FWHM Infrared –1cm-1FWHMU titi i th t f l l li t fUncertainties in the spectroscopy of glyoxal explain part of the scatter among available literature yieldsGlyoxal absorption cross sectionOOHHTable 4: Comparison with available literature values. Reference Spectral resolutiona UV-region (350 nm) Vis-region (440 nm) Integral σ-ratio fit-coef.c shiftc [10-20 cm2] [10-20 cm2] [rel.units] [rel.units] [nm] UV / vis spectral ranges this work ~0.001 105 1.00 1.00 0.01 98 0.1 0.38 / 0.41b 53 1.0 0.39 26 SUNY [22] 0.01 0.63 SUNY [25] 0.01 0.51 SUNY [26] 0.01 46 0.48 0.44 0.56 EUPHORE [30] 0.17 33 0.89 0.86 0.10 MPI [24] 0.25 0.36 30 0.94 0.87 0.39 NCAR [23] 0.6 0.36 17 0.93 0.69 -0.03SAPRC [15] n.n. 0.01 25 0.83 0.64 1.08 IR spectral range NCAR [23] 1 cm-1 1.05 NIES [44] 1 cm-1 0.80 FORD [34] 0.25 cm-1 0.98 BUGH [45] 1 cm-1 0.74 a FWHM value in nm, unless noted otherwise. b Notation: σ350nm / σ351nm. For ref. [25] sigma at 351 nm needs be compared. c Least squares fit, after adjusting for resolution (see section 3.1). Volkamer et al., J. Photochem. Photobiol. A: Chemistry 172 (2005) 35.NASA recommended UV-vis absorption cross section (JPL 06-02).Sample applicationsSample applications1. UV spectra: glyoxal photolysis 2First direct detection in the atmosphere2.First direct detection in the atmosphere3. Global maps from satellite4. Mechanism development of hydrocarbonspy5. Source apportionment of HCHO6. Ocean sources of hydrocarbonsGlyoxal photolysis3003504004503003504004501x10-192x10-193x10-190.51.02x10144x10146x1014antum Yield [rel. units],rption Cross Section [cm2]Quantum-YieldActinic Photon Flux[ph cm-2 s-1 nm-1]sigmaActinic Photon FluxField measurements of J(Glyoxal) using spectral-00.0300 350 400 450 50001.0x10-51.5x10-5QuaAbsor sigmax Flux x QY s-1 nm-1]370-430nmPhotolysis: 20%290-365nm Photolysis: 80%radiometry requires good knowledge of sigma. 300 350 400 4500.05.0x10-6sigma x[ph Wavelength [nm]300 350 400 450150380 430nm330 375290 3251.2x10-4 J w/ sigma IUPJw/sigmaMPI005.0x10-61.0x10-51.5x10-550100150380-430nmPhotolysis: 20%Error: 18%330-375nm Photolysis: 40%Error: 60%290-325nm Photolysis: 40%Error: 22%sigma x Flux x QY [ph s-1 nm-1]40% of overall J-valuesigma variability [%], sigma [rel. units]60 10-58.0x10-51.0x10-4J w/ sigma MPIyoxal) [s-1]300 350 400 4500.002.0x10-63.0x10-64.0x10-650100 40% (available literature data)ma variability [%]lines this scalesigma x Flux x QY [ph s-1 nm-1]6% w/ sigma IUP12% w/ sigma