USGS-NWQL: I-2668-98:  Selenium in water by graphite furnace-atomic absorption spectrometry, dissolved

  • Summary
  • Analytes
  • Revision
  • Data and Sites
Official Method Name
Determination of Arsenic and Selenium in Water and Sediment by Graphite Furnace Atomic Absorption Spectrometry
Current Revision
1998
Media
WATER
Instrumentation
Graphite Furnace-Atomic Absorption Spectrometer
Method Subcategory
Inorganic
Method Source
  USGS-NWQL
Citation
Jones, S.R., and Garbarino, J.R., 1998, Methods of analysis by the U.S. Geological Survey National Water Quality Laboratory-- Determination of arsenic and selenium in water and sediment by graphite furnace-atomic absorption spectrometry: U.S. Geological Survey Open-File Report 98-639.
Brief Method Summary
The determination of selenium by GF-AAS requires that a small (microliter) volume of sample be pipetted into a graphite tube. The tube is held between two graphite rings with quartz windows at each end, which produces a nearly closed environment to enhance the absorbance signal. The tube is pyrolytically coated with high-density carbon to reduce the formation of nonvolatile carbides to prevent surface adsorption of the sample onto the walls of the graphite tube, and to increase tube life (Ghe and others, 1983). The sample is evaporated to dryness, charred, and atomized by using specified temperatures and high-temperature ramping. The absorbance signal is measured and compared to standards.
Scope and Application
This method is used to analyze filtered water samples for the determination of dissolved selenium. Using a 30-uL sample injection, the linear analytical range is 1 to 50 ug/L. Samples that contain selenium concentrations exceeding the upper limit of the analytical range need to be diluted and re-analyzed or analyzed by an alternate method.
Applicable Concentration Range
1 to 50 ug/L
Interferences
Two types of interferences affect GF-AAS analyses: spectral and nonspectral. The most common type of spectral interference is background absorption where undissociated molecular forms of matrix materials produce broadband absorption spectra. Zeeman background correction compensates for broadband absorption by using a strong magnetic field to shift the electronic energy levels of an element's atom. The shifted atomic spectrum can then be differentiated from the background spectrum (Beaty, 1988). Atoms having absorption wavelengths unresolved from the elemental absorption wavelength cause either positive or negative errors when measuring the element concentration profile (Flajnik-Rivera and Delles, 1996).
Sufficient energy must be available during atomization to dissociate the element and create free atoms. The composition of the sample matrix interferes with this process and results in either positive or negative errors. Matrix vapor condensation takes place not only in end-heated graphite atomizer-type furnaces, but also in spatially isothermal, transverse heated graphite atomizer-type furnaces (Frech and L'Vov, 1993). It has also been shown that atoms can be trapped on the surface of condensed matrix particles and deposited at the cooler tube ends, resulting in a depression of the elemental signal. Introducing a low-flow of inert gas during the atomization step can eliminate this type of interference. However, using low-flow causes some loss of sensitivity because some of the atoms are swept from the chamber (Frech and L'Vov, 1993). Nevertheless, low-flow conditions are recommended for routine sample analysis at the NWQL.
Lindberg and others (1988) have reported spectral and volatilization interferences on selenium. Selenium may be lost either by volatilization or by the formation of decomposition products, such as hydrogen selenide and selenium monoxide, which are not dissociated during the ashing and atomizing steps. Different modifiers, such as palladium, have been used to reduce volatilization losses (Lindberg and others, 1988).
Quality Control Requirements
Analyze a quality-control sample immediately following calibration and after every tenth sample. Minimum quality-control requirements per set must include analysis of a laboratory reagent blank and quality-control samples, such as SRWS, and may include check standards, sample duplicates, and sample spikes. Field spikes and duplicates are suggested as additional checks. Correlation coefficients for calibration curves must be at least 0.999. QC samples must fall within 1.5 standard deviations of the mean value. If all of the data-acceptance criteria in the SOPs are met, then the analytical data are acceptable.
Sample Handling
Use 250 mL polyethylene bottle, acid-rinsed. Filter through 0.45-um filter, use filtered sample to rinse containers and acidify sample with nitric acid (HNO3) to pH <2.
Maximum Holding Time
180 days
Relative Cost
Less than $50
Sample Preparation Methods