Analysis Of Experimental Measurement And Calculation Methods Of Detection Limit

With the continuous progress of experimental testing techniques,trace analysis has gradually become an important work for analysis andtesting, and it is also increasingly important for the study of trace analysismethods and some basic theories. In the trace analysis, detection limit isan important concept. Detection limit is possessed of a great deal ofinterest among analytical chemists, because it is not only of academiccuriosity but also of practical significance: detection limit is a widelyused criterion to evaluate analytical methods and the quality of analyticalinstruments. Presently, there are many means to calculate detection limit,which are based on different perspectives or different priorities, but thesecalculation methods can be divided into two categories: one is thecorrection method of single concentration, which based on blankvolatility; the other is the standard curve method, which based on thedeviation characteristics of calibration curve.The detection limit is a statistical concept, so it is inevitableemergence of condition of false positive value and false negative value.Through the research analysis, in this work, the detection limit is defined as: with at least99%confidence, the analyte concentration or amount isdetected by analytical instruments when the false positive probability ofmeasurement results is10%. The above theory takes both positivedeviation and negative deviation of measurement results intoconsideration. At the same time, it defines the probability of false positiveof quantification limit as5%. Firstly, it is needed to establish thecalculation procedure of detection limit; secondly, a variety of analyticalinstruments enter into trace analysis, and it takes time to obtain therespective detection limit; finally, detection limits obtained are analyzedand compared, trying to find out some regularity.In this work, we establish a complete set of calculation method ofdetection limit, and with the data model of linear function to fitexperimental results of various analysis methods, sequentially, a series ofrigorous and reliable detection limit could be received.