Spectral Data Manipulation

P. Van Der Voort, Spectral Data Manipulation, ECC-Comett Program, Fourier Transform Infrared Spectroscopy, Antwerp, Belgium, 1992. 

Apart from the actual acquisition of the mass spectrum and its subsequent display or printout, the raw mass spectral data can be processed in other ways, many of which have been touched on in other chapters in thi.s book. Some of the more important aspects of this sort of data manipulation are explained in greater detail below. 

Very rarely, however, will a single mass spectrum provide us with complete analytical information for a sample, particularly if mass spectral data from a chromatographic separation, taking perhaps up to an hour, is being acquired. The mass spectrometer is therefore set up to scan, repetitively, over a selected m jz range for an appropriate period of time. At the end of each scan, the mass spectrum obtained is stored for subsequent manipulation before a further spectrum is acquired. 

Sample preparation for the common desorption/ionisation (DI) methods varies greatly. Films of solid inorganic or organic samples may be analysed with DI mass spectrometry, but sample preparation as a solution for LSIMS and FAB is far more common. The sample molecules are dissolved in a low-vapour-pressure liquid solvent - usually glycerol or nitrobenzyl alcohol. Other solvents have also been used for more specialised applications. Key requirements for the solvent matrix are sample solubility, low solvent volatility and muted acid - base or redox reactivity. In FAB and LSIMS, the special art of sample preparation in the selection of a solvent matrix, and then manipulation of the mass spectral data afterwards to minimise its contribution, still predominates. Incident particles in FAB and LSIMS are generated in filament ionisation sources or plasma discharge sources. 

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. 

Should at least a full spectral window be acquired for a 2Q-HoMQC spectrum, there are interesting possibilities for data manipulation through processing. 

The spectra of the peracetylated P-D-glucose are used as a reference in two ways. Firstly they serve in a comparative way to let you verify your results when studying the effect of different data manipulations and the influence of different processing parameters on the processing of the experimental raw data (FID). Secondly they serve as a reference of various NMR parameters (shifts, coupling constants,. ..) and give you valuable spectral information to help elucidate the unknown structure of the peracetylated oligosaccharide. 

The data manipulating capability of a computerized infrared spectrometer allows the spectroscopist to delve more deeply into the structural origin of the infrared absorptions by using data processing techniques to purify, manipulate, and correlate the spectra. If one can systematically vary the relative amounts of various structural contributions, absorbance subtraction can be used to isolate the spectral contributions of the structural components. 

Since the article by Spedding1 on infrared spectroscopy and carbohydrate chemistry was published in this Series in 1964, important advances in both infrared and Raman spectroscopy have been achieved. The discovery2 of the fast Fourier transform (f.F.t.) algorithm in 1965 revitalized the field of infrared spectroscopy. The use of the f.F.t., and the introduction of efficient minicomputers, permitted the development of a new generation of infrared instruments called Fourier-transform infrared (F.t.-i.r.) spectrophotometers. The development of F.t.-i.r. spectroscopy resulted in the setting up of the software necessary to undertake signal averaging, and perform the mathematical manipulation of the spectral data in order to extract the maximum of information from the spectra.3... 

No discussion has been devoted to the recent use of Fourier transform spectrometers rather than dispersion instruments. The ease with which the spectral data can be manipulated and background subtracted make the FT methods particularly useful for studies of surface species, particularly during catalytic reaction. Recently there has been a surge of interest in the coupling of computer subtraction techniques to conventional grating instruments. For many IR surface studies, where only limited frequency range is required, this... 

With this in mind, let us take a look at the design of the LC/MS, its operation, and the way mass spectral data are manipulated to produce chromatographic information and compound identification. This will be simply an... 

Smoothing and differentiation of spectral data The high geometric accuracy of OIDs provides the means for an accurate, software-preformed, smoothing (low-pass filtering) and differentiation (high-pass filtering) of raw spectral data. These manipulations facilitate the interpretation and identification of acquired spectra, Fig. 6 and 7. 

The current waveform on the anode of the photodetector is fed to the input connector of a transient digitizer/digital oscilloscope where it is converted into digital format and read into a PC. The resulting stored data can be processed for spectral/ dynamic content by one of several available commercially available data manipulation software packages. 

A wide variety of mathematical manipulation schemes are available to smooth spectral data, and in this section we shall concentrate on smoothing techniques that serve to average a section of the data. They are all simple to implement on personal computers. This ease of use has led to their widespread application, but their selection and tuning is somewhat empirical and depends on the application in-hand. 

---