Calibration, generally curves

In many detectors the internal counting efficiency depends on the energy of the radiation. This must also be taken into account, if sources containing other radionuclides are used for calibration. Generally, calibration curves = f E) are deter-... 

A calibration/standard curve is the relationship between instrument response and the expected concentration of analyte. To adequately define the relationship between concentration and response, a sufficient number of standards/ calibrators should be used, typically >4. The reproducibility and accuracy of the standard curve across multiple ( 3) runs should be confirmed by calculating interassay precision (%CV) and accuracy (%recovery of observed concentration compared to expected concentration) for each standard. Generally, a well-defined and optimized calibration curve will demonstrate acceptable precision (%CV 20%) and acceptable accuracy (%recovery of 80-120%). 

For all four sets of PEO standards the coefficient of determination (R ) for the linear calibration curves for the four linear columns in water and in water/methanol are better than 0.99, except for the PL PEO standards and the TSK GM-PWxl column in water/methanol. The coefficient of determination for the TSK GM-PWxl column in general is not as good as the other three linear columns. The coefficient of determination for the TSK PEO standards showed the least dependency on columns and mobile phases. The TSK GM-PWxl column has a lower exclusion limit in the high molecular weight range than TSK GM-PW, Shodex SB-806, and SB-806MHQ columns. 

Visual and photoelectric colorimeters may be used as turbidimeters a blue filter usually results in greater sensitivity. A calibration curve must be constructed using several standard solutions, since the light transmitted by a turbid solution does not generally obey the Beer-Lambert Law precisely. 

With the multitude of transducer possibilities in terms of electrode material, electrode number, and cell design, it becomes important to be able to evaluate the performance of an LCEC system in some consistent and meaningful maimer. Two frequently confused and misused terms for evaluation of LCEC systems are sensitivity and detection limit . Sensitivity refers to the ratio of output signal to input analyte amount generally expressed for LCEC as peak current per injected equivalents (nA/neq or nA/nmol). It can also be useful to define the sensitivity in terms of peak area per injected equivalents (coulombs/neq) so that the detector conversion efficiency is obvious. Sensitivity thus refers to the slope of the calibration curve. 

Two general approaches have been used to overcome matrix effects (1) partial purification of the analyte prior to analysis by immunoassay ( cleanup methods) and (2) the use of a matrix blank when preparing the calibration curve. Both options are widely used, but each has its individual limitations. 

Peedc-to-peak resolution in SEC can be calculated by the ratio of peak separation at the peak maxiaut to the sum of the baseline peak widths. This general definition of resolution is less useful in SEC, where a measure of the ability of the column to separate solutes of different molecular weight is required. For this purposes, we define a new term, the specific resolution factor, R, which relates peak resolution to sample molecular weight, assuming all measurements are made within the linear region of the molecular weight calibration curve, equation (4.41)... 

Development of an industrial monitoring application for IMS requires extensive preparatory work, as well as optimal operational conditions for IMS, i.e. the nature of the reagent gas, calibration curves, evaluation of interferents, assessment of reliability. Analysis of mixtures with four or fewer components may be possible, but extension to more complex mixtures should be considered only in special cases, and generally would be unrealistic. Use of preseparators, such as GC columns, is the only known technical approach... 

For complicated samples where matrix or interelement effects are present, a linear calibration curve may not be valid, and one should consider using an empirical model for concentration correction. This usually requires a large set of standards of similar composition to the unknown, which generally makes analysis rather impractical. Inter-element effects can be calculated from a basic knowledge of physical parameters in combination with the appropriate use of samples of known composition, pure elemental standards or composite standards. 

---