Assessment of Detector Nonlinearity in Fourier Transform Spectroscopy

In Fourier transform spectroscopy (FTS), nonlinearity amplitude distortions in the interferograms are mainly due to detector. In the case of photoconductive detectors, as for instance, the semiconductor Mercury-Cadmium-Telluride detector (MCT), that is the detector most frequently employed for high-performance infrared measurements, the nonlinearities have been often discussed in the literature. For other infrared detectors, such as bolometers, the effect is not studied with equal attention; nevertheless, also with these detectors some nonlinearity is expected to be present. Amplitude distortions in the interferogram induce, through Fourier transform, radiometric errors in the acquired spectra. In order to eliminate radiometric errors, the interferograms must be corrected before Fourier transformation, by using either a measurement or a model of the real response of the acquisition system. An assessment of nonlinearity errors in FTS can also be made after Fourier transformation using the harmonic features that are introduced in the measured spectra by the nonlinear response. Usually nonlinearity was detected by looking for a nonzero signal in the low-frequency region of the spectrum, where the spectral response should be equal to zero. This procedure may be difficult in presence of 1/f noise, that prevents an accurate identification of the low frequency feature. In this paper, a method for the detection and characterisation of small nonlinearity effects by way of an analysis of the amplitude of the second harmonic feature in the acquired spectral distribution is presented and discussed. The method is applied to real measurements obtained with a far IR interferometer with an 3He-cooled bolometric detector.