For analysis of some matrices, improved accuracy may be obtained by using different preintegration or integration times for certain elements. For example, a method for determining both ‘soluble’ and ‘insoluble’ forms of elements (e.g. aluminium, titanium, boron) in steel depends on variable source conditions, pre-integration and integration times. Experimental determinations of intensity variations with time (referred to as burn-off curves) are carried out to establish whether this approach is useful.
5.2 CALIBRATION
5.2.1 Principle
With the spectrometer adjusted to the selected optimum settings, the intensities at the selected wavelengths for both calibration and normalization standards shall be measured. The measurements shall be repeated until acceptable repeatability is obtained for each standard.
A number of calibration standards shall be used to define the calibration curve over the specified concentration range. In general, the chemical and metallurgical composition of calibration standards and test samples should be similar. The resultant intensity measurements shall be used to establish calibration curves. This function is normally performed by the instrument computer and expressed as polynomial equations.
5.3 STANDARDIZATION
5.3.1 Principle
Standardization is also known as ‘normalization’. drift correction’. or ‘recalibration’. The measured intensities obtained from spectrometer channels for a specific set of conditions vary with time. To compensate for this variation and to ensure that the intensities being referred to the stored calibration are valid for that calibration, it is necessary to adjust the currently measured intensities for each channel.
The standardization is performed by referring intensities obtained currently on normalization standards to intensities obtained on the same standards at the time of initial calibration (Clause 52).
The frequency of standardization, which will depend on the stability of the instrument and the required precision of the determination, should be determined by each individual laboratory. The need for standardization can be determined by analyzing a known homogeneous reference sample and assessing the results obtained. As a guide for instruments in regular use, a reference sample should be analyzed and the results checked for each element, at least every four hours. If maintenance has been performed on the instrument. then the reference sample shall be analyzed and the instrument standardized, if necessary. If repeated standardizations fail to give satisfactory results for the reference sample, instrument performance should be checked in accordance with Clause 4; ultimately, a new calibration may need to be carried out.
NOTE: The stability of an instrument may depend on such short term factors as ambient temperature, electronic stability or change in source performance, or on long term effects such as change in optical alignment or gradual component deterioration.
5.3.2 Normalization standards
These standards shall be homogeneous and have concentrations covering all elements at each end of the concentration range. A suite of standards may be required to cover all elements.
5.3.3 Standardizing the measured intensities
During calibration, intensity measurements shall be obtained for the normalization standards. Whenever standardization takes place, a significant change in intensity measurements should be corrected (electronically or mathematically) back to those obtained during calibration. Thcsc corrcctions arc applicd to all samples subscqucntly analyzed.
5.3.4 Equation for corrections (see Figure 1)