Optical emission spectrometers are fundamentally based on comparative measurements. As a result, they need to be calibrated with standard reference materials to create a "calibration curve" before or during use. The analysis of a sample is based on this established curve. However, there can be a difference between the analysis data and the true value, which is why control samples are used to perform a re-calibration.
Let's look at the different types of standard samples for an optical emission spectrometer:
Calibration Standards: These are used to create the working curve for the spectrometer. A complete set of these standards is required.
High/Low Standards: Used for high/low point standardization of the working curve. Typically, at least two or more pieces are needed.
Control Samples: These are used to verify the accuracy of the spectrometer's working curve and to perform curve recalibration, with the ultimate goal of making the analysis of unknown samples more accurate.
The purpose of a control sample is to compensate for the differences between the analysis data and the true value. These differences can be caused by factors like curve drift, surface conditions, or variations in the sample's metallurgical and processing techniques. By using control samples to create a specific analytical program, these errors can be automatically compensated for.
When selecting a control sample, it's essential that it has the same metallurgical process and physical state as the analytical sample. Its elemental content should be accurate and reliable, with a uniform composition and no physical defects like sand holes, cracks, or pores. The elemental content of the control sample should also fall within the range of the calibration curve and be close to the elemental content of the sample being analyzed.
Experiments have shown that using self-melted and re-certified standard samples provides better analytical accuracy than using commercially available standards. Therefore, we recommend that well-equipped companies melt and re-certify their own samples to be used as control samples.
