Comparison of optical transmittance and reflectance measurements on thin-media chromatograms

Abstract

This paper compares the relative merits of transmittance and reflectance procedures in the photodensitometric analysis of substances separated on thin-media chromatograms. The considerations used are based upon the equations of Kubelka and Munk, which describe in simplified terms the optical transfer properties of turbid media. The parameters used in the comparison are sensitivity, linearity and signal amplitude. The theoretical values of these parameters were calculated using a computer and represented in graphical form. For improved linearity a logarithmic transform of the primary signal was satisfactory in the case of transmittance. For reflectance an inverse representation gave the best linearity. Both transforms promise improvement only if the primary signal-to-noise ratio is reasonably high. When very low concentrations are to be measured, the primary signal provides adequate linearity and no transform is necessary. At larger concentrations transmittance with logarithmic conversion and reflectance with inverse representation give nearly equal performance for most media. With media of very high optical density the light intensity available in the transmittance mode may be insufficient compared to the electrical noise of the photodetector and reflectance is then the mode of choice. Reflectance may also have a slight advantage over transmittance with regard to optical noise if specularly reflected light can be kept away from the photodetectors. Any linearising transform has to be carried out before the total amount of substance is determined by integration. For full exploitation of the benefits of these transforms a flying spot scanner is essential. A method is also described which permits the approximate determination of the coefficients of scattering and absorption of the medium from simple photodensitometric measurements. Knowledge of these parameters is required if more sophisticated methods of linearisation than logarithm forming or inversion are to be employed, e.g., by using a computer. The conclusions obtained do not apply to methods based upon fluorescence.