بسم الله الرحمن الرحيم
Trace analysis
Trace analysis involves the determination of those elemental constituents of a sample that make up approximately 0.01% by weight of the sample or less. There is no sharp boundary between nontrace and trace constituents. The lower limit of detected concentration is set by the sensitivity of the available analytical methods and, in general, is pushed downward with progress in analytical techniques. A large number of different physical and chemical techniques have been developed for the measurement of the elemental composition at the microgram-per-gram and nanogram-per-gram level, thereby constituting the field of trace analysis.
Methods
All trace analytical methods can be divided into three component steps: sampling, chemical or physical pretreatment, and measurement. Depending upon the type of information desired in an analysis and the requirements of sensitivity, precision, and other performance figures of merit, an appropriate measurement technique is selected. Most frequently, this will be an instrumental approach. Therefore, it is essential to know the capabilities and limitations of the various methods of instrumental measurement available for trace determinations. Once they are known, appropriate steps can be taken in the sampling and pretreatment steps to provide sufficient amounts of the microconstituents that are free of interferences and in the appropriate form for the final measurement. In a number of methods the pretreatment step may be omitted, and in others the sampling and measurement occur simultaneously. In spite of possible deviations, these steps are interrelated and require different degrees of emphasis, depending upon the individual analytical situation. In many cases, the analyst uses a particular physical or physicochemical method in which manifestations of energy provide the basis of measurement. These methods are indirect in the sense that the emission or absorption of radiation or transformation of energy must be related in some way to the mass or concentration of the species that are being determined. The establishment of these relations almost invariably requires calibration, with the use of standards of known content of the constituent in question. As such, many of the available techniques do not provide absolute results.Trace analysis ranges from the more classical chemical methods of colorimetric and absorption spectrophotometric analysis to modern instrumental approaches. The wide diversity of methods is apparent from the number of approaches and the attendant classes and subclasses in the table. Within any one of these categories, there are techniques that are more specialized.Of the various criteria used in the selection of an appropriate trace analytical method, sensitivity, accuracy, precision, and selectivity are of prime importance. Other important considerations, such as scope, sampling and standards requirements, cost of equipment, and time of analyses, are of great practical significance.
Sensitivity and detection limits
For all analysis regimes, the measure x of some physical parameter is related to the concentration c of the analyte in a certain sample. The sensitivity S is the slope of the analytical calibration curve, which is the plot of the measure x versus the concentration of the analyte in a series of standards having known analyte concentrations. The term sensitivity must not be used to indicate either the limit of detection or the concentration required to give a certain signal.The standard deviation σ is given by Eq. (1),
(1) where xi is the individual measure, and x is the mean of n measurements. If n is less than about 10, σ is replaced by the term s. The relative standard deviation (RSD) is given by Eqs. (2).
(2) Precision is defined as the random uncertainty for the measure x of the corresponding uncertainty in the estimate of the concentration. It is often expressed as the relative standard deviation.Accuracy is a measure of the agreement between the estimated concentration and the true value. Accuracy can never be better than the precision of the analytical procedure. Bias (systematic errors) in the calibration procedure always causes the estimated uncertainties to disagree with the true value by an amount equal to the bias.The limit of detection cL is the smallest concentration that can be detected with a reasonable certainty. This concentration is given by Eq. (3),
(3) where xL is the limiting detectable measure and xb is the average blank measure. The value for xL is given by Eq. (4),
(4) where k is a factor determining a margin of confidence and sb is the standard deviation of a limited number of measures of the blank. Generally, k is taken as 3; thus, the value of cL can be determined by Eqs. (5).
(5) This means that any concentration resulting in a signal three times the background standard deviation is considered just detectable. To evaluate xb and sb, at least 20 measurements should be used. If the major sources of variation are electrical noises, sb can be replaced with Nb, the background noise level. If 3sb is chosen, the confidence level is 99.86% for a purely Gaussian distribution of errors. At low concentrations, broader and asymmetric distributions are likely, so 3sb corresponds to a practical confidence level of about 90%. It is noted that at the limit of detection the relative standard deviation is about 0.5; that is, there is equal confidence in deciding whether the analyte is detected or not.
As a result of widely varying pathways of development of many analytical techniques, there is a lack of consistency in the definition or specification of detection limits in the analytical literature. Consequently, it is difficult to make critical comparisons between methods for a particular element. However, since a particular element may be determined by a number of different techniques, depending upon the matrix in which it is being sought, a summary of the experimental values that have been published is pertinent