Trace element analysis of toxic heavy metals in soils and contaminated land using a portable XRF spectrometer.

Leian Grimsley, British Geological Survey, Nottingham, UK. Leiven Kempenaers, PANalytical BV, Almelo, The Netherlands.

Soil pollution is conventionally defined as adding chemical compounds, biological organisms or other materials to the soil, negatively effecting or altering how the soil functions. Common soil contaminants include heavy metals, hydrocarbons, biological pathogens and substances that can acidify and/or enrich the soil with nutrients.

In order to control this pollution of the soil, one has to be able to measure it: there is no point setting standards if its is impossible to measure the compliance. Once excellent method is X-ray fluorescence spectrometry. The main reasons for this are: that it is simple to prepare samples; the technique offers high accuracy and precision over a wide dynamic rance (ppm to percentages) and has good to excellent detection limits across a wide band of the periodic table (Na – U). Thus the method covers the needs of thos interested in soil sample anaysis for both agricultural and environmental reasons.

In order to illustrate the ability and accuracy of the process, a trial was undertaken to determine the ability to measure the presence of six heavy metals that are commonly found in soils from contaminated land: iron (Fe), nickel (Ni), copper (Cu), zinc (Zn), arsenic (As) and lead (Pb).

X-ray fluorescence spectrometry is a comparative technique; one compares the output from an unknown sample with that from samples of known compositions. For this trial, sixteen reference materials and secondary standards were obtained from BGS (British Geological Survey, Nottingham, UK). The precise concentrations of these elements in the secondary standards were obtained by various techniques including ICP-AES nd WDXRF at the British Geological Survey’s Geochemical Laboratories.

Portable and Practical for on-site use

The mearurements were performed using a PANalytical MiniPal 4 EDXRF spectrometer (see below), equipped with a rhodium anode tube, five tube filters, a helium purge facility, a high resolution Silicon Drift Detector and a 12-position removable sample changer with sample spinner. Only one set of measurement conditions was required to measure all the elements of interest (see Table 1). Iron, nickel, copper, zinc and arsenic were analyzed using

Ka fluorescent lines whereas for lead, the La fluorescent line was used. The total measurement time per sample was only 150 seconds.

Table 1. Measurement set-up conditions

With its small size and weight (only 28kg with sample changer!) the MiniPal 4 is a powerful compact bench top system and easily transportable. At BGS, the spectrometer is part of a mobile environmental laboratory and can be used to analyse soil samples at different locations in the U.K. The MiniPal 4 is equipped with a 12-position removable sample changer enabling unattended batch anaysis. Soil samples can show heterogeneity that affects the precision and accuracy of the analysis but, fortunately, rotating or spinning the sample in the X-ray beam can average thse out. For this reason, a robust sample spinner is an option for the MiniPal 4 system.

The MiniPal 4 inside the BCS mobile environmental laboratory. Permission kindly granted by the British Geological Survey © NERC

Calibration

Ten grams of each secondary standard powder, obtained from BGS, were prepared for analysis in the Minipal 4 in the standard way by putting the powder into a P1 "de Kat" disposable plastic cell, assembled with a 6um polypropylene supporting foil, and applying a weight of 500 grams to compress the loose powder and so reduce sampling error. The samples were then analyzed for 150 seconds each, and calibration graphs established using the regression model built in to the MiniPal software. Figure 1 shows the calibration plot for arsenic in contaminated soil. The plots illustrate visually that a good correlation can be obtained over a wide range of concentrations.

The calibration results for all the analyzed elements are shown in Table 2. The root mean square (RMS) error listed in Table 2 is a measure of the difference between the measured concentration and the certified chemical concentration and is therefore a measure of the accuracy of the method. However, the magnitude of the calibration RMS value is dependent on the range of concentrations in the calibration. A weighted error estimate, the K-value, is independent of concentrations and gives a more consistent indication of accuracy. As a "rule of thumb", a K-value less than 0.02 reflects a good calibration.

Figure 1. Calibration plot for arsenic in contaminated soil

Table 2. Calibration results for the six heavy metals.

Anayltical consistency

To test the analytical precision of the MiniPal 4 spectrometer, one polluted soil sample was measured fifteen times. The average concentration, RMS error, relative RMS error, and the relative counting statistical error (CSE) are presented in Table 3. The CSE is theoretically the minimum possible error. All the elements analyzed have precisions better than or equal to 3% relative. The variations in the measurements for zinc and lead are illustrated graphically in Figure 3.

Table 3. Results of the precision test

Figure 3. Repeatability of results for zinc and lead in a soil sample N.B Error bars for the selected elements are not visible within the point shown.

Detection limits

The detection limits for the elements studied in a typical contaminated soil sample are given in Table 4 for both the applied measurement time of 150 seconds. The Lower Limit of Detection (LLD) is calculated from:

Where:

S = sensitivity (cps/ppm) Rb = background count rate (cps) Tb= background counting time (s)

Table 4. Detection Limits

The LLD values quoted are typical for soil samples. However LLD values do vary for individual samples depending on the composition of the sample.

Conclusion

The results clearly demonstrate that the MiniPal 4 EDXRF spectrometer is well suited for the analysis of heavy metals such as iron, nickel, copper, zinc, arsenic and lead in soils and contaminated land. Good results have been obtained for the regressions, the analytical precision and the Lower Limits of Detection. High detector resolution powerful software correction models contribute to this result. Furthermore, the compact size and small weight of the spectrometer makes it an ideal system for on-site quality control of soils and contaminated land.