Multianion analysis, applications of ion chromatography - Determination of Anions in Natural and Treated Waters - page 689

Table 12.27 Detection limits and calibrated range for sequential seven anion
determinations

Anion Detection limit, mg

L −1

Calibrated range, mg
L −1

Limit of linearity, mg
L −1

>20 a

Fluoride (F) 0.02 0.1-15.0

Chloride (CI) 0.02 0.1-30.0 >20°

Nitrite (NO 2 )

0.25 0.5-30.0 50

Orthophosphate
(PO 4 )

0.1 0.2-50.0 50-60

Bromide (Br) 0.1 0.5-50.0 >200>

Nitrate (NO 3 ) 0.1 0.2-30.0 I50

Sulphate (SO 4) 0.2 0.5-100.0 120

a Maximum concentration level tested
Source: Reproduced with permission from the American Chemical Society [78]

A variation in sensitivity of orthophosphate was noted with relation to concentration
range. Linear responses were obtained for concentration ranges 0-25mg L −1 phosphate
and 30-80mg L −1 phosphate. However, at a concentration between 25 and 30mg L −1
phosphate, the sensitivity changed. To establish a broad linear range, the data were
evaluated by using the method of linear regression. Up to a concentration of 55mg L −1
phosphate the actual phosphate values did not vary significantly from the linear
regression plot. Attempting to quantitate phosphate at concentrations above the 55mg L −1
level would result in significant positive error. The degree of error would increase as the
phosphate concentration increased. Because of the linearity problem the dynamic range
of the method was limited to 0-50mg L −1 phosphate. Calibration of the system was
established by using a concentration of phosphate standard (25mg L −1 phosphate) which
was mid-range for the method. Also, the phosphate concentration used for calibration fell
within the range where the change in sensitivity occurred. Precision and accuracy data
confirmed accurate calibration over the 0-55mg L−1 r a nge, using a 25mg L −1 phosphate
standard.
A decrease in retention time as concentration increased was noted for nitrate. The
average retention times observed during development of precision data are listed in Table
12.28. This phenomenon (peak migration) was a source of concern in the automated
method because the integrator identifies peaks on the basis of retention time. The
program entered into the integrator permitted only a 10% maximum deviation in retention
time between the sample peak and a standard calibration peak. If the sample peak
retention time deviated more than 10% from that of the calibration standard, the sample
peak would be identified by the integrator as an extraneous peak or would possibly be
incorrectly identified as some other anion having a similar retention time. A minimum