Signal processing Fourier transforms

Apart from some special drift processes that we will treat separately, the noise in the measurements is expected to be the result of random processes much faster than the changes in the useful signal itself. Fourier transform spectral methods exploit this difference in frequency for separating the two components by considering a frequency-domain representation of the signal instead of its original time domain representation. 

By means of numerical convolution one can obtain Xg t) directly from sampled values of G t) and Xj(t) at regular intervals of time t. Similarly, numerical deconvolution yields Xj(t) from sampled values of G(t) and Xg(t). The numerical method of convolution and deconvolution has been worked out in detail by Rescigno and Segre [1]. These procedures are discussed more generally in Chapter 40 on signal processing in the context of the Fourier transform. 

Although the idea of generating 2D correlation spectra was introduced several decades ago in the field of NMR [1008], extension to other areas of spectroscopy has been slow. This is essentially on account of the time-scale. Characteristic times associated with typical molecular vibrations probed by IR are of the order of picoseconds, which is many orders of magnitude shorter than the relaxation times in NMR. Consequently, the standard approach used successfully in 2D NMR, i.e. multiple-pulse excitations of a system, followed by detection and subsequent double Fourier transformation of a series of free-induction decay signals [1009], is not readily applicable to conventional IR experiments. A very different experimental approach is therefore required. The approach for generation of 2D IR spectra defined by two independent wavenumbers is based on the detection of various relaxation processes, which are much slower than vibrational relaxations but are closely associated with molecular-scale phenomena. These slower relaxation processes can be studied with a conventional... 

Each signal sequence A or B will be processed separately by using the Fourier transform and CFAR target detection techniques. A single target with specific range and velocity will be detected in both sequences... 

Fig. 10.12. Pulse sequence for amplitude modulated 2D J-resolved spectroscopy. The experiment is effectively a spin echo, with the 13C signal amplitude modulated by the heteronuclear coupling constant(s) during the second half of the evolution period when the decoupler is gated off. Fourier transformation of the 2D-data matrix displays 13C chemical shift information along the F2 axis of the processed data and heteronuclear coupling constant information, scaled by J/2, in the F1 dimension.

A Fourier transform infrared spectroscopy spectrometer consists of an infrared source, an interference modulator (usually a scanning Michelson interferometer), a sample chamber and an infrared detector. Interference signals measured at the detector are usually amplified and then digitized. A digital computer initially records and then processes the interferogram and also allows the spectral data that results to be manipulated. Permanent records of spectral data are created using a plotter or other peripheral device. 

Modern IR spectrometers do not use light dispersion to acquire spectra. Rather, they utilize a device called an interferometer between the source and the sample. This design requires a signal processing circuit that performs a mathematical operation called a Fourier transformation to obtain the spectra. 

---