| Résumé | A broadband coherent anti-Stokes Raman spectroscopy (CARS) system optimized for single shot measurements in turbulent flames is described. The CARS spectrometer incorporates an intensified photodiode array (IPDA) with a P46-phosphor based intensifier that is shown to have an order of magnitude less image persistence than P20-phosphor based IPDAs, and thus largely eliminates the temperature errors that can result from this image persistence. The low and high signal nonlinearity of IPDAs incorporating both P20 and P46 phosphor based intensifiers is described. The CARS signal is dispersed with a single element concave holographic diffraction grating, which is fiber-optically coupled. To provide the necessary dynamic range (approximately 1000:1) to handle CARS spectra whose temperatures can vary from 300 K to adiabatic flame temperatures a fiber-optic splitter was employed and the nonlinearity of the IPDA detectors characterized. A method of determining the nonlinearity was developed that was convenient and the fiber-optic link was shown to provide a stable calibration of this nonlinearity. The problem of dye laser spectral drift, which is particularly severe in a long duration experiment, was addressed by constructing a compact high resolution spectrometer to continuously monitor the dye laser. The dye laser center frequency was maintained constant by manually adjusting an intracavity tuning filter. The accuracy of this technique and its effect on CARS derived temperatues was evaluated. The standard method of determining the instrument function by fitting to 295 K air spectra was compared to other methods and its use was shown to result in an uncertainty of less than 0.5% in CARS derived temperatures. The operation of this CARS instrument was demonstrated in a turbulent diffusion flame. The varying composition of these flames was accounted for by fitting C, tha ratio of the nitrogen mole fraction to the total third order nonlinear susceptibility. |
|---|
