Basics of Environmental Sampling and Analysis - Fundamentals of Environmental Sampling and Analysis - page 18

calibration curve during routine laboratory operation'' (Popek, 2003). Sample PQLs
are highly matrix dependent. For example, in measuring polynuclear aromatic
hydrocarbons (EPA Method 8310), the matrix factor ranges from 10 (groundwater)
to 10,000 (high level soil/sludge by sonication).

2.1.5 Standard Calibration Curve

The calibration curve or standard curve is a plot of instrumental response
(absorbance, electrical signal, peak area, etc.) vs. the concentrations of the chemical
of interest. Five or more standard solutions of known concentrations are first
prepared to obtain the calibration curve - normally a linear regression equation:

y ¼ axþb ð2:11Þ

where y is the instrument response and x is the concentration of the chemical. The
slope factor (a) is the instrument response per unit change in the concentration and
it is termed as the calibration sensitivity. Since intercept b equals the instrumental
response when analyte is absent (x¼0), ideally b is a small number close to zero
(i.e., the calibration curve ideally will run through the origin (x ¼ 0

y ¼ 0)).
The concentration of the unknown sample is calculated using the standard curve
and the measured instrument response. The calibration curve is essential for
almost

;

all quantitative analysis using spectrometric and chromatographic

methods.
To obtain the regression equation (y ¼ axþb), one should be familiar with a
spreadsheet program such as Excel. Procedures for the regression analysis using
Excel are as follows:

Select Tools|Data Analysis
Select Regression from Analysis Tool list box in the Data Analysis dialog
box. Click the OK button
In the Regression dialog box

Enter data cell numbers in Input Y Range and Input X Range
Select Label if the first data cells in both ranges contain labels
Enter Confidence Level for regression coefficient (default is 95%)
Select Output Range and enter the cell number for regression output
Select Linear Fit Plots and/or others if needed
Click the OK button
In the regression output, r 2 , slope (a) and intercept (b) can be readily
identified.

Figure 2.2 illustrates several common outcomes when the calibration curve is
plotted. The calibration curve (a) is ideal due to its good linearity and the sufficient
number of data points. Curve (b) is nonlinear, which is less favorable and