Paper version: In stock and available from the USGS Store
A new analytical method for the determination of elements in filtered aqueous matrices using inductively coupled plasma-mass spectrometry (ICP-MS) has been implemented at the U.S. Geological Survey National Water Quality Laboratory that uses collision/reaction cell technology to reduce molecular ion interferences. The updated method can be used to determine elements in filtered natural-water and other filtered aqueous matrices, including whole-water, biota, sediment, and soil digestates. Helium or hydrogen is used as the collision or reaction gas, respectively, to eliminate or substantially reduce interferences commonly resulting from sample-matrix composition. Helium is used for molecular ion interferences associated with the determination of As, Co, Cr, Cu, K, Mg, Na, Ni, V, W and Zn, whereas hydrogen is used for Ca, Fe, Se, and Si. Other elements that are not affected by molecular ion interference also can be determined simply by not introducing a collision/reaction gas into the cell. Analysis time is increased by about a factor of 2 over the previous method because of the additional data acquisition time in the hydrogen and helium modes.
Method detection limits for As, Ca, Co, Cr, Cu, Fe, K, Mg, Na, Ni, Se, Si (as SiO2), V, W, and Zn, all of which use a collision/reaction gas, are 0.06 microgram per liter (?g/L) As, 0.04 milligram per liter (mg/L) Ca, 0.02 ?g/L Co, 0.02 ?g/L Cr, 0.04 ?g/L Cu, 1 ?g/L Fe, 0.007 mg/L K, 0.009 mg/L Mg, 0.09 mg/L Na, 0.05 ?g/L Ni, 0.04 ?g/L Se, 0.03 mg/L SiO2, 0.05 ?g/L V, 0.03 ?g/L W, and 0.04 ?g/L Zn. Most method detection limits are lower or relatively unchanged compared to earlier methods except for Co, K, Mg, Ni, SiO2, and Tl, which are less than a factor of 2 higher.
Percentage bias for samples spiked at about one-third and two-thirds of the concentration of the highest calibration standard ranged from -8.1 to 7.9 percent for reagent water, -14 to 21 percent for surface water, and -16 to 16 percent for ground water. The percentage bias for reagent water spiked at trace-element concentrations of 0.5 to 3 ?g/L averaged 4.4 percent with a range of -6 to 16 percent, whereas the average percentage bias for Ca, K, Mg, Na, and SiO2 was 1.4 percent with a range of -4 to 10 percent for spikes of 0.5 to 3 mg/L. Elemental results for aqueous standard reference materials compared closely to the certified concentrations; all elements were within 1.5 F-pseudosigma of the most probable concentration. In addition, results from 25 filtered natural-water samples and 25 unfiltered natural-water digestates were compared with results from previously used methods using linear regression analysis. Slopes from the regression analyses averaged 0.98 and ranged from 0.87 to 1.29 for filtered natural-water samples; for unfiltered natural-water digestates, the average slope was 1.0 and ranged from 0.83 to 1.22. Tests showed that accurate measurements can be made for samples having specific conductance less than 7,500 microsiemens per centimeter (?S/cm) without dilution; earlier ICP-MS methods required dilution for samples with specific conductance greater than 2,500 ?S/cm.
Additional Publication Details
USGS Numbered Series
Chapter 1. Determination of elements in natural-water, biota, sediment, and soil samples using collision/reaction cell inductively coupled plasma-mass spectrometry