434 E. Eastwood and C. P. Snow Owing to lack of knowledge of the precise effect of electron exchange on the mobility of ions, it is not possible at present to use these results to deduce the forces between the ions and the gas atoms or molecules with which they are colliding. It is therefore proposed to extend the experiments to alkali ions in inert gases, since exchange processes do not then occur.
436 E. Eastwood and C. P. Snow Mulliken has built up an orbital structure of the C=C double bond in ethylene, and has assigned the bands of ethylene at X 1900 A to an electronic transition localized in the double bond and corresponding to a change of electric moment parallel to the axis of the molecule. (Owing to the absence of fine structure in the ethylene bands, this hypothesis, though very probable, is not definitely established.) In correspondence with us he has also suggested a structure for the >C =0 double bond in formaldehyde, which is quoted later in this paper. For this molecule the precise analysis of Dieke and Kistiakowsky shows definitely that the change of electric moment must be perpendicular to the axis of the mole cule. The number of possible upper states of formaldehyde is accord ingly limited: and further arguments, and an application of Dieke and Kistiakowsky’s complete analysis, give a unique choice of upper states and a transition which is not localized in the double bond.
438 E. Eastwood and C. P. Snow state to the combination levels. These series tend to overlap and so it may be said that the 1187 cm-1 frequency determines the broad features of the formaldehyde spectrum, the major bands and their associated subsidiary series imposing a definite periodicity of intensity on the spectrum.
Electronic Spectra of Polyatomic Molecules 439 even for acetaldehyde and propionaldehyde; it is only when the spectrum is compressed within a short space that the structure is readily visible. No evidence for the further resolution of these broad bands into the real ultimate bands was obtained, except for propionaldehyde and acetone when the commencement of the region of absorption became partially resolved with the E.l spectrograph. With acetaldehyde, however, the Frequency A—Acetaldehyde B—Propionaldehyde C—Butyraldehyde Fig. 2 band system was clearly resolved up to 2825 A as Schou had already reported.
Electronic Spectra of Polyatomic Molecules 441 Table I—(continued) Wave-length Frequency Frequency in air in A in vacuo in cm-1 separation Isobutyraldehyde 3242 30836 3132 31919 1083 3032 32972 1053 2942 33981 1009 2857 34991 1010 2776 36012 1021 2703 36985 973 2629 38026 1041 Mean frequency difference = 1027 cm-1 Isovaleraldehyde 3361 29744 3252 30741 997 3142 31818 1077 3042 32864 1056 2952 33865 1001 2868 34857 992 2787 35870 1013 Mean frequency difference = 1023 cm-1 Heptaldehyde 3228 30970 3123 32011 1041 3025 33048 1037 2934 34073 1025 2849 35090 1017 2770 36090 1000 2695 37095 1005 Mean frequency difference = 1021 cm-1 Acetone ? 2804 35653 2720 36754 1101 2640 37867 1113 Mean frequency difference = 1107? cm-1 for them in view of the widths of the bands (approximately 400 cm-1) and the vagueness attached to the selection of the centre points as the points defining the bands. The table is thus intended to be suggestive only, to indicate the great probability of the existence of the >C =0 valency vibration in the excited state with the approximate values given.
Electronic Spectra of Polyatomic Molecules 443 of the potential energy curves for the >C =0 group, and would indicate a pronounced bond-loosening, as the diminution in the >C =0 frequency also suggests. There is no passage into a region of continuous absorption corresponding to dissociation of the >C =0 group.