MULTICHANNEL SPECTROMETRY

Author Institution: Lick ObservatoryThe sensitivity of infrared detectors is already so close to fundamental limits that further large increases seem unlikely. Improvements in the performance of spectrometers are possible by increasing the area of the source that is used, either by increased dispersion or by multislit techniques. It is theoretically possible to increase the intensity of the source almost without limit if the random thermal processes of radiation can be replaced by a coherent mechanism such as occurs in a radio antenna. This principle is the basis of modern microwave spectrometry, but there is no immediate prospect of it being extended to infrared frequencies. Semicoherent spark techniques perhaps deserve more attention than they have been given in recent years. Further improvements must depend on using more efficient methods of observation. A scanning spectrometer is very inefficient in that at every instant it rejects all but one element of the incident spectrum. This loss is avoided in a spectrograph by having effectively a separate radiation detector for each spectral element. Sensitive infrared image detectors are not at present available, but the spectral elements can be measured simultaneously by a single detector if mutually orthogonal modulation patterns are impressed on the separate elements. A convenient method of producing the required modulation is by varying the path difference in a two-beam interferometer. No dispersing system, such as prism or grating, is then required. The theory of this method, which may be termed multichannel spectrometry, has been investigated in detail, and it has been confirmed experimentally that the method gives spectra with the expected resolution and that the theoretical increase in sensitivity is realized. The multichannel method is especially appropriate to astronomical spectrometry, where no improvement in intensity or area of the source is possible