Sample, load, effect precision

Another example often quoted is that there is no sample effect when measuring the length of a table. First, this is an incomplete definition of a measurand, since it is not clear what is meant by the length. But if, for example, the table forms a part of a measuring system to measure the precise shape of objects then the shape of the table can have an influence on the result. Then such factors as the material of the table, its temperature coefficient, the temperature gradients across the table, the loading of the table, the friction between its feet and the floor are sample effects that need to be taken into account. 

For infinite dilution operation the carrier gas flows directly to the column which is inserted into a thermostated oil bath (to get a more precise temperature control than in a conventional GLC oven). The output of the column is measured with a flame ionization detector or alternately with a thermal conductivity detector. Helium is used today as carrier gas (nitrogen in earlier work). From the difference between the retention time of the injected solvent sample and the retention time of a non-interacting gas (marker gas), the thermodynamic equilibrium behavior can be obtained (equations see below). Most experiments were made up to now with packed columns, but capillary columns were used, too. The experimental conditions must be chosen so that real thermodynamic data can be obtained, i.e., equilibrium bulk absorption conditions. Errors caused by unsuitable gas flow rates, unsuitable polymer loading percentages on the solid support material and support surface effects as well as any interactions between the injected sample and the solid support in packed columns, unsuitable sample size of the injected probes, carrier gas effects, and imprecise knowledge of the real amount of polymer in the column, can be sources of problems, whether data are nominally measured under real thermodynamic equilibrium conditions or not, and have to be eliminated. The sizeable pressure drop through the column must be measured and accounted for. 

Devices using a series of capil laries (13) or syringes (14,15,16) were employed with partial success in conventional TLC. However, none of the multicapillary or multisyringe devices proved to be suitable for the smaller dimensions of HPTLC. The main problems were precision of positioning and, last but not least, the effect that the workload of applying one sample after another manually is simply shifted to loading sample solutions one after the other into a miniaturized multidispensing device. 

---