Spectroscopic studies of low-pressure flames. II. Effective translational and rotational temperatures from CH bands By A. G. Gaydon and H, G. Wolfhard Chemical Engineering Dept, Imperial College, London (Communicated by Sir Alfred Egerton, F.R.S.—Received 2 February 1949) [Plate 1] Using a Fabry-Perot interferometer, the Doppler widths of lines of the CH bands emitted from the reaction zone of an oxy-acetylene flame have been examined. The experimental arrangement and method of computation are described. The effective translational tempera tures of the excited CH radicals are around 4000° K in the low-pressure flames, compared with theoretical maximum flame temperatures around 2700° K. At atmospheric pressure the translational temperature is near the expected flame temperature. The effective rotational temperatures have also been measured for the CH bands 4315 and 3900A and do not differ much from the theoretical maximum flame temperatures for any of the several flames ex amined. The method of excitation of the CH is discussed. It seems likely that the radicals are excited by collision with other active species in the flame; the high translational tem perature shows that we do not have normal thermal excitation, but there are also difficulties in attributing the effects to true chemiluminescence. For the weak CH band at 3143 A the effective rotational temperature is very high, especially at low pressure.
90 A. G. Gaydon and H. G. Wolf hard Strong 1948), this is probably the first detailed account of the use of Doppler broad ening for the measurement of temperatures in flames. The strong bands of CH around 4315 and 3900 A are among the strongest features of the emission spectrum of the inner cones of premixed flames of hydrocarbons and many organic com pounds, but they do not fit readily into most detailed theories of combustion pro cesses, as do the OH radicals, and it would be of interest if we could settle whether the strength of these bands resulted from the high concentration of CH radicals in the flames or was merely the result of chemiluminescence in some side reaction.
Spectroscopic studies of low-pressure . II 91 measurement, but by measuring the half breadth of the lines of CH it should at least be possible to determine whether or not there is any marked departure from expected flame temperature and to learn something of the time required for the translational energy of excited radicals to reach equilibrium with the surrounding gas. For obtaining a value for an effective translational temperature it is sufficient to measure the half-width of the line, but to find whether there is any departure from a Maxwell-Boltzmann distribution of translational energy among the molecules it is necessary to compare the intensity contour of the spectrum fine with that expected theoretically for pure Doppler broadening.
92 A. G. Gaydon and H. G. Wolfhard the instrumental line width as given by the interferometer a water-cooled mercury arc, A, was used; this was placed behind the flame and could be brought into focus on the slit by moving Lv The circular interference fringes formed in the interferometer were brought to a sharp focus on the slit by moving the lens since this lens was not achromatic it was necessary to take focus plates for near the wave-lengths required using the water-cooled mercury arc. The interferometer plates were adjusted for parallelism in the usual way, and the interferometer was set at a slight angle so as to throw the centre of the fringe system just above or below the slit which was then crossed by about ten horizontal fringes.
Sjpectroscojpic studies of low-pressure flames. 93 mercury; apart from main resonance lines, pressure broadening also appeared small in this arc; most of the mercury hues showed a definite hyperfine structure, but the' central component due to the isotopes of even atomic weight was much stronger than other components and no difficulty was found in deriving the instrumental line contour from measurement of suitable mercury lines.
94 A. G. Gaydon and H. G. Wolfhard contour we are unaware of any exact law, and the resultant seems to be between that given by the square and by the linear summation. Owing to varying effects of adjust ment, focus, and temperature change, the instrumental contour changes slightly from plate to plate. It was therefore necessary to make a detailed calculation for each plate.
Spectroscopic studies of low-pressure flames. II 95 Figure 2 a and b and figure 3 show respectively a typical group of observed in tensity contours for a CH line, the observed contour of the mercury line at 4077-6 A for deriving the instrumental contour, and in figure 3 three curves, one the deduced instrumental contour, the second the assumed fine contour for a selected value for the temperature, and the third the result of graphically integrating these two; points derived from the experimentally obtained curve are plotted along the calculated curve for comparison.
96 A. G. Gaydon and H. G. Wolfhard from stoichiometric and the theoretical flame temperature would be rather higher than the observed translational temperature in the reaction zone as obtained in this way. The value does, however, serve to show that even at atmospheric pressure the pressure broadening of the CH lines is comparatively negligible and that at this pressure the average translational temperatures of the excited molecules does reach approximate equilibrium with the gas temperature.
Spectroscopic studies of low-pressure flames. II 97 The observed intensity contour for the line of the 3900 A band was in satisfactory agreement with the contour calculated from the instrumental contour combined with an ideal Doppler broadening. The degree of agreement varied somewhat from plate to plate, largely because of the experimental error in allowing for the back ground intensity. In the curves reproduced in figure 3 it will be seen that the circles, derived from the observed line contour, follow the calculated contour quite well. The experimental error is such that the comparison between the observed and calculated curves cannot give high accuracy, but it does seem that there is no marked indication that the distribution of velocities in the excited CH radicals differs appreciably from a Maxwell-Boltzmann distribution at an effective temperature of about 4000° K.