Representation of Spectra

we will examine some of the basic features of an infrared spectrum. Early infrared instruments recorded percentage transmittance over a linear wavelength range. It is now unusual to use wavelength for routine samples and inverse wavelength units are used. This is the wavenumber scale, and the units used are cm i. The output from the instrument is referred to as a spectrum. Most commercial instruments present the spectrum with the wavenumber decreasing from left to right 

The infrared spectrum can be divided into three regions the far infrared ( 400cm ), the mid infrared (400CMOOcm 0 and the near infrared (14285-4000 cm i). Most infrared applications employ the mid-infrared region, but the near- and far-infrared regions can also provide information about certain materials. The majority of instruments are set up to scan only the mid-infrared range and this is the re on that we will largely concentrate on in later chapters. 

Generally, there are few infrared bands in the region between 4000 and 1800 cm but many bands between 1800 and 400cm i. Often the scale is changed so that the region between 4000 and 1800 cm is contracted and the region between 1800 and 400cm i is expanded to emphasise the features of interest. 

---

Figure 2.47 (a) 1RRAS spectra of CO adsorbed on Pt in 1.0 M HCI04 saturated with CO. The electrode potential was held constant at (i) 50 mV vs. NHE, (ii) 250 mV, (iii) 450 mV and (iv) 650 mV. (b) Difference spectra resulting from subtraction of the IRRAS spectra (iii) and (i). (c) EMIRS spectrum of Pt electrode in CO-saturated 1 M HCIO modulated between + 50mV and 450mV. (d) A schematic representation of spectra at two potentials which could produce an EMIRS spectrum similar to that shown in (c). The IRRAS spectra in (a) rule out this possibility. After Russell et al, (1982). Copyright 1982 American Chemical Society,... 

Organization of full-curve or reduced representations of spectra... 

Figure 8 Binary representation of spectra data (I - peak, 0-no peak). The sample has smallest number of XOR bits set with reference spectrum R4, and this, therefore, is the best match...

JCAMP-DX is a standardized file format for the representation of spectra, chromatograms, and International Union of Pure and Applied Chemistry (lUPAC). 

B.K. Alsberg, Representation of Spectra by Continuous Functions, Journal oj Chemometrics, 7 (1993), 177-193. 

A priori it cannot be assumed that the relationships between spectral data and structural data are linear. A lot of effort therefore has been spent to improve spectra classification by the implementation of non-linear methods, especially by neural networks. Typical results show that neural networks are sometimes better than linear discrimination methods but not considerably better. Obviously the representation of spectra by appropriate features (which is a domain of chemistry) is more important than the applied classification methods. 

---