A new type of measuring and computing engine has been constructed which determines wave-lengths and wave numbers directly from spectrograms made by crossing the dispersion of a Fabry-Perot interferometer with that of a concave diffraction grating. The wave-length of a spectrum line is first determined to six figures from the grating dispersion by the method used in the author’s automatic comparator, and this wave-length is then converted to wave number by a mechanical continuous-function computing arrangement. The thickness of the etalon separator used having been pre-set into the machine, all wave-lengths are automatically divided into twice this thickness to give on dials the order of interference corresponding to the instantaneous wave-length setting of the comparator. The interference pattern. of each spectrum line is projected on a screen in front of the operator with a magnification proportional to the square root of the order of interference, and falls on a set of parabolic fiducial lines drawn to fit the optical system used to project the interference fringes on the slit of the grating spectrograph. The operator moves the screen carrying these parabolic lines up or down until they coincide with the density maxima of the interference pattern; the partial order of interference ε in the center of the pattern can then be read to three figures from a dial. By turning the plate control handle of the machine slightly the operator then makes the ε reading of the machine coincide with that of this dial; cams having introduced corrections for the dispersion of air and the variation of phase change with wave-length, the dials then give the wave-length and wave number of the line correctly to seven or eight figures, depending on its sharpness. The speed of reduction of Fabry-Perot patterns is by this means increased at least 20-fold over that of previous methods, and the precision of setting appears to be considerably improved also. Repeat readings on sharp lines are found consistent to within a few ten-thousandths of an angstrom. The precision of the wave-length values obtained is limited principally by the diffuseness of the interference pattern arising from the breadth of a line.
Improvements now under development include automatic photoelectric scanning of interferometer patterns, and the direct recording by an electric typewriter of eight-figure wave-length and wave-number values for each line. The new device is being used to determine standard wave-lengths needed for the more effective utilization of the original automatic comparator.
GEORGE R. HARRISON, "Direct Determination of Wave-Lengths from Fabry-Perot Interferometer Patterns," J. Opt. Soc. Am. 36, 644-651 (1946)