Spectral techniques, relationship

Spectral features and their corresponding molecular descriptors are then applied to mathematical techniques of multivariate data analysis, such as principal component analysis (PCA) for exploratory data analysis or multivariate classification for the development of spectral classifiers [84-87]. Principal component analysis results in a scatter plot that exhibits spectra-structure relationships by clustering similarities in spectral and/or structural features [88, 89]. 

Both solution-state and solid-state NMR spectroscopy are important analytical tools used to study the structure and dynamics of polymers. This analysis is often limited by peak overlap, which can prevent accurate signal assignment of the dipolar and scalar couplings used to determine structure/property relationships in polymers. Consequently, spectral editing techniques and two- or more dimensional techniques were developed to minimize the effect of spectral overlap. This section highlights only a few of the possible experiments that could be performed to determine the structure of a polymer. 

Thus, it is seen that noncomputer, spectral results have been used in numerous investigations on vibrational spectra-structure relationships. When such complex molecules as carbohydrates, which are sensitive to the environment and reveal configurational and conformational changes, as well as intra- and inter-molecular hydrogen-bonding, are dealt with, the noncomputer techniques, even though more qualitative and less rigorous than the calculation methods, remain quite useful in practice. 

The advent of personal computers greatly facilitated the application of spectroscopic methods for both quantitative and qualitative analysis. It is no longer necessary to be a spectroscopic expert to use the methods for chemical analyses. Presently, the methodologies are easy and fast and take advantage of all or most of the spectral data. In order to understand the basis for most of the current processing methods, we will address two important techniques principal component analysis (PCA) and partial least squares (PLS). When used for quantitative analysis, PCA is referred to as principal component regression (PCR). We will discuss the two general techniques of PCR and PLS separately, but we also will show the relationship between the two. 

We shall conclude this chapter with a few speculative remarks on possible future developments of nonlinear IR spectroscopy on peptides and proteins. Up to now, we have demonstrated a detailed relationship between the known structure of a few model peptides and the excitonic system of coupled amide I vibrations and have proven the correctness of the excitonic coupling model (at least in principle). We have demonstrated two realizations of 2D-IR spectroscopy a frequency domain (incoherent) technique (Section IV.C) and a form of semi-impulsive method (Section IV.E), which from the experimental viewpoint is extremely simple. Other 2D methods, proposed recently by Mukamel and coworkers (47), would not pose any additional experimental difficulty. In the case of NMR, time domain Fourier transform (FT) methods have proven to be more sensitive by far as a result of the multiplex advantage, which compensates for the small population differences of spin transitions at room temperature. It was recently demonstrated that FT methods are just as advantageous in the infrared regime, although one has to measure electric fields rather than intensities, which cannot be done directly by an electric field detector but requires heterodyned echoes or spectral interferometry (146). Future work will have to explore which experimental technique is most powerful and reliable. 

In spite of the fact that both studies have reported important spectral features associated with the structure-property relationship, an Achilles heel of both approaches results from the necessity of using optically transparent films to allow infrared light to pass through the sample. New materials, such as fibers and composites, cannot be studied by transmission FT-IR techniques because they are often optically opaque. Thus, in order to monitor structural changes induced by external forces, it is necessary to utilize a method permitting the detection of infrared spectra on any material, regardless of its optical properties, shape or thickness. 

---