DEVELOPMENTS IN QUANTITATIVE ORGANIC ULTRAMICRO ELEMENTARY ANALYSIS

Numerous principles for the mineralization of samples and for the measure-ment of the elements have been tried since the days of Liebig, Dumas, Dennstedt and Pregl. Attempts to develop accurate and reliable ultramicro methods have been successful only in recent years. Almost absolutely quantitative recovery of the analyte from the mineraliza-tion process, freedom from risk of contamination, and very high sensitivity and specificity of the measuring method are the main requirements for the usefulness of a procedure for ultramicroanalysis. A number of excellent available methods are described and discussed with regard to these requirements. The most difficult problem is the quantification of hygroscopic and other difficult samples prior to analysis. This problem and a few solutions to different cases are discussed. For the quantitative determination of an element in a sample, the sample must be destroyed and the element isolated as a homogeneous, easily measurable compound. The most common methods for the destruction are heating in a flow of gas, which liberates the elements to be measured as volatile compounds; stationary combustion in a closed, oxygen-filled container, in which the combustion products can be absorbed and measured; and decomposition in hot oxidizing acids or alkalis, in which the analytes remain dissolved. METHODS USING DECOMPOSiTION iN A GAS FLOW Carbon, hydrogen, nitrogen A good example of the gas flow decomposition method is shown in Figure 1. This method which was described by Beicher et al. ' is adapted to deal with samples of between 5 and 50 rig. It was the first successful method for the determination of carbon and hydrogen with samples of below 100.Lg. When carbon, hydrogen and nitrogen are to be determined simultaneously the decomposition of the sample can be so arranged that carbon dioxide, hydrogen and nitrogen are the final products. They are all gaseous at low concentration and determination by gas chromatography has been attempted by several authors on the micro scale. Again Beicher and his co-workers