Physicochemical Properties of Test Compounds

Selecting the most appropriate set of test compounds for use in the evaluation of stationary phases is a key step of a chromatographic test. Only test compotmds possessing quite different physicochemical and structural properties will induce different interactions with the stationary phase, thereby allowing a good evaluation of its chromatographic performance. For this reason, it is highly recommended that a chemometric tool, such as Prindpal Component Analysis, is used for the selection of the most appropriate compounds. 

As already discussed by Euerby and Petersson in Chapter 2.1.3 (theory of PCA), PCA is applied to two-dimensional tables crossing individuals (test compounds) and quantitative variables (physicochemical properties of test compounds). The main objective of PCA is to evaluate the resemblance between individuals. Hence, two individuals are considered as being similar if their values are closer in comparison to the totality of the variables. In order to better illustrate this concept, an example will be given concerning the selection of the most suitable set of test compounds. 

In order to ensure that the selected test compounds (Table 1) are sufficiently different to evaluate chromatographic supports, PCA of some of their physicochemical properties was carried out  

Inspection of the obtained loading plot (Fig. 1) and of the contribution of variables (Table 2) to Principal Components (PC) is necessary to evaluate differences and similarities between basic compounds scores in Fig. 2). All variables are well represented in the selected two-dimensional space, as their length is close to 1.0 (see Chapter 2.1.3). By considering the direction of the and log D vectors and their contribution to PCI, compounds can be ranked along PCI in relation to 

When columns of different lengths and particle size are compared, h should be used instead of N. h = H/d = L/ N dp), where dp is the particle size, L the column length, and N the efficiency. 

---