ABSTRACT
The reproducibility of autosamplers for high-performance liquid chromatography can be checked by the injection of a sodium chloride solution without an installed column. The dispensed chloride plug is analyzed quantitatively by argentometric titration. If the titration can be performed with high enough precision, the autosampler performance is investigated without additional effects. Experiments with a 20-microL loop were performed. We found the following repeatabilities (relative standard deviations): titration, 0.07%; 10-microL "pull loop" injection (partial filling of the loop), 0.6%; and 20-microL "full loop" injection, 0.1%. In contrast to these data, the relative standard deviations obtained from chromatographic peaks include the steps of injection, separation, detection, and integration without any knowledge of the individual contributions. Such values are higher; we found for chromatographic peak areas of parabens 2% under pull loop conditions and 0.17% by full loop. The titration approach allows the determination of the contribution of the injection to the uncertainty budget of a liquid chromatographic analysis.
