Environmental Engineering Reference
In-Depth Information
0.60
(a)
(b)
0.30
0.40
0.20
0.20
0.10
0.00
0.00
0.60
0.60
(c)
(d)
0.40
0.40
0.20
0.20
0.00
0.00
0.60
(e)
0.60
(f)
0.40
0.40
0.20
0.20
0.00
0.00
01234567
0123456
C (mg/L)
C (mg/L)
Figure 2.2
Correct and incorrect calibration curves: (a) Ideal, (b) Poor linearity, (c) Data too scattered,
(d) Insufficient data points, (e) Incorrect concentration range, and (f) Systematic error (Meyer, VR, 1997,
Reproduced with Permission, John Wiley & Sons Limited)
effort should be made to improve it prior to use. Data in (c) are too scattered, and
you should redo to correct the errors. The curve in (d) does not have enough data
points (five is normally the minimum). The concentration range for curve (e) is
not correct, either the sample has to be diluted or a wider range of concentrations
should be used. Calibration curve (f) does not run through the origin, implying the
presence of some systematic errors such as positive or negative matrix interference
(Meyer, 1997).
Practical tips
It is always a good laboratory practice to run the calibration curve along with
samples in the same batch, particularly when there is a considerable
instrument drift. It takes more time to prepare and run, but it will improve
the data validity and you can probably save time at the end.
When a dilution is made and sample (injection) volume is different from the
standard solution, corrections are needed to calculate sample concentration.
In this case, a plot of instrument response vs. analyte mass (rather than
concentration) is more appropriate.
Search WWH ::
Custom Search