How to avoid pitfalls in interpretation of Calibration plots

Calibration plots assume great significance in quantitative analysis. It is just not possible to carry out analysis without prior generation of calibration plots. However, it is important for you to adhere to the suggested guidelines so as to make the calibration data unquestionable and of highest reliability.

A single calibration plot for an analytical system is not valid

It is sufficient to prepare a single calibration plot for measurement of physical quantities such as length, weight, temperature or time. The calibration will be valid till there are marked changes in environmental parameters (temperature, humidity, etc) under which measurement devices are stored and handled .However, you should realize that such generalizations are not valid for chemical determinations as the variables apart from environmental controls in the laboratory get multiplied several fold. Examples of such variables are normal wear and tear of analytical instruments, fluctuations in operating conditions such as wavelength of light source, stray light changes, temperature of a chromatographic column, flow and pressure variations of mobile carrier phases, composition variation of mobile phases. Most important is to remember that a calibration plot will be dependent on concentration range over which determinations are to be made and interferences can result when same analyte is determined in different matrices. In short each determination under a different set of operating conditions requires a different calibration plot.

Three calibration points are not sufficient

It is a well known that three points can be connected to make a straight line. However, a calibration plot requires more than three points, say 6 to 8 points at least, to ensure reliability of calibration data.
Do not force the line to pass through all the points

It is a general tendency to force the calibration line through maximum if not all the points plotted manually on the graph paper. Your calibration plot may look impressive but it may lead to conclusions which could be questioned by others. It is best to use system software to display the calibration plots as such software use proven regression analysis and display the coefficient of variation values between the two variables.

Select the appropriate concentration range

The concentration range selected for generating a calibration plot should cover the concentration range expected in determination of a particular analyte in the sample. Making a single calibration plot over a concentration range will not hold for determinations over a different range of concentrations. In principle the measurements should cover at least between 80% to 120% of the expected concentration range and a blank should also be included.

Make use of traceable reference materials

The standards used for making calibration plots should be traceable to recognized global standard bodies and should be of highest purity. Working standards which are standardized against reference materials can be used but each time the solution should be prepared afresh.

Use calibrated volumetric glassware

Standards of highest purity will give erroneous results if calibrated volumetric glassware and pipettes or micropipettes are not used. Remember volumetric errors get multiplied in serial dilutions if the volume transferred at each stage is not error-free.

Use standard and validated software

It is important to use validated software from authorized sources for generating calibration plots. I rated softeners are capable of generating erroneous that a ritual for far outweigh the benefit gained in their lower cost

Carry out calibrations periodically

All analytical instruments should be calibrated regularly at intervals which are recommended by the suppliers. However, it is a good practice to generate a calibration plot every time the system is used for quantitative estimations over the required range of concentrations to be determined.