Intensity Corrections For Gas Temperature Determinations At High Temperatures From FTIR Absorption And Emission Measurements

We are investigating species and temperature profiles in low-pressure pre-mixed flames using FTIR spectroscopy. As part of this project, we are seeking reliable methods of obtaining rotational temperatures. Because it is a common participant in most combustion reactions, we have chosen carbon monoxide for the initial st0;ies. Our calculational methods are extentions of the work of Anderson and Griffiths'' (for CO at T < 150°C (423°K)); to test our methods, we have measured CO absorption spectra from 40 to 600°C, and emission spectra from 100 to 600°C. Obtaining the rotational temperature of CO at high temperature (>423°K) from the structure of the v=0 - v=1 ro-vibrational band is complicated by three factors: The first and most important factor is the distortion of true peak intensities by the spectrometer resolution due to convolution of the spectrometer lineshape function and the true line profile. For absorption of a collisionally-broadened line convolved with an instrument lineshape gunction resulting from triangular apodization, the recorded transmittance is given by:.) T1 = I exp{-2.303 A pmam2/((v-vm) 2+a2m)}Dsin2(πD(v-v1))/(πD(v-v.))2 dv (1) where T1 = transmittance of line m when the instrument is set at frequency i, am = half width of half height of line m, vm = frequency at center of line m, v = frequency at which instrument is set (or reporting), D = slit width or optical retardation, and Apm = true peak absorbance of line m at vm . The second complicating factor is overlap of some lines with weak lines due to 13C16O;; this factor was taken into account by Anderson and Griffiths. The third factor which complicates temperature calculations is that some of, the lines in the v=0 - v=1 band of CO overlap with lines from the v=1 - v=2 transition.4 Figure 1 shows a portion of the CO absorption spectrum using 0.125 cm-1 resolution Visible in this spectrum are both "hot band" transitions as well as those due to 13C160. Because the "hot bands" are much stronger at high temperature (>300 or 400°C) and become more important as temperature increases than those bands due to 13C16O, we have decided to initially ignore l3CI160 bands. Figure 2 shows linecenter separation between "normal" and hot bands as a function of frequency for CO.