The Measurement of Arsenic in Drinking Water
By: By. Zoe Grosser, Lee Davidowski and Laura Thompson, PerkinElmer Inc.
Arsenic in our drinking water…it sounds like a plot line for the play Arsenic and Old Lace. But it does occur, naturally and from man-made pollution. How can we understand what levels may be safe to drink and how to measure at the concentrations required? News headlines describing the inadvertent exposure of the population in Bangladesh to high levels of arsenic in their well water have increased awareness of this issue and the diseases that can be caused. Drinking water in the United States can sometimes have higher levels of arsenic than desired, as shown in Figure 1. The arsenic arises mostly from natural mineral sources, with a small contribution from pollution. Headlines from a town in Texas show that even concentrations only slightly elevated above the safe levels are cause for concern to the general population. (1) A recent recall of bottled water because it contains unsafe levels of arsenic indicates that bottled water may not be a safer alternative! (2)
The US EPA has regulated the As concentration allowed in drinking water since 1975, under the Safe Drinking Water Act. At that time the amount permitted in drinking water before treatment is required was set at 0.05 mg/L. The World Health Organization has also recommended acceptable levels of arsenic in drinking water for many years and recently lowered their level to 0.01 mg/L in 1993. The US subsequently set a more stringent requirement of 0.01 mg/L (10 ppb) to protect human health and the environment and compliance was required beginning in January 2006. The state of California is considering a lower standard, perhaps as low as 4 parts per trillion, about 2,500 times lower than the new federal standard.
Arsenic measurement methods for drinking water are specified in the Code of Federal Regulations. The methods approved for measurement at the new, lower concentration level are shown in Table 1. The ICP method, previously approved when the higher concentration was regulated, has been removed because it doesn’t meet the requirements at the lower concentration.
Several current techniques that might be used to measure low concentrations of arsenic were evaluated. They do not all have EPA methods written to include the exact instrument configuration used to collect the data, but under method streamlining they may be considered variations of existing methods that might be acceptable.
Experimental
The following techniques were considered for arsenic measurement:
* ICP-MS using method 200.8
* ICP-MS with Dynamic Reaction Cell, method 200.8 (no reaction gas)
* GFAA using method 200.9
* Automated FIAS hydride AA, using method 3114B
* Axial ICP-OES, using method 200.7
* Axial ICP-OES, measuring As hydride, using 3114B variation
Method 200.7 for ICP-OES has been excluded from the approved list of methods to be used for low-level analysis; however, the detection limits in the published method are measured using a radially-viewed plasma, not an axially-viewed plasma, which is more sensitive. The last technique couples hydride generation with another form of detection.
The instrumental conditions were as follows:
ICP-MS: ELAN® 9000, PerkinElmer SCIEX, Inc. was set up using the following conditions
Hydride generation: The FIAS 100, PerkinElmer was used for the hydride AA measurements. A small mixing block was used to mix the reagents for the ICP-OES analysis. The standards were prepared as directed in the method. The pre-reduction (KI/Ascorbic acid for at least 45 minutes) step is critical for complete arsenic hydride generation, regardless of detector.
Results and Discussion
Methods were evaluated in a number of ways for potential utility. Although analytical capability is clearly the first criterion, other important considerations include capital investment, measurement time, and whether the technique can measure more than one element at a time.
Method detection limits were used as an indication of suitability for low-level arsenic measurement. The detection limits obtained are shown in Table 2. Generally detection limits ten times lower than the concentration used for a decision ensure confidence at the required limit.
The detection limits for all techniques, except the axial ICP-OES are well below the level needed to give confidence at the compliance level. The detection limit for ICP-MS with DRC was not optimized for arsenic, which would improve the detection limit that could be obtained. The DRC would provide additional benefits if a saline or more heavily matriced water were measured because detection limit degradation would not be observed. The axial ICP detection limits could be improved with preconcentration. Optimization of the instrumental conditions for arsenic may improve the detection limit also. Since arsenic was part of a multielement run, no special attempts were made to optimize the detection limit.
The overall performance of the different techniques was compared by measuring a variety of drinking water from public and well water supplies. Table 3 shows selected samples, spikes, and reference material results.
The precision of replicates is shown in the parenthesis and shows that measurement of arsenic at low levels can be done reliably by all of these techniques.
Once the suitability of the analytical capability is determined other considerations, such as the time for analysis versus workload in the laboratory can be evaluated. Table 4 shows the time for a single sample to be measured for arsenic.
Generally the entire suite of primary drinking water contaminants must be measured on each sample. So the time for analysis of one element may be misleading if the technique can measure more than one element at a time. A better comparison compares the analysis time for a full suite of elements to be measured, using different combinations of techniques found in a typical laboratory. The comparison in Table 5 shows the time compared for 11 elements to be measured as part of a 20-sample batch.
This analysis assumes that the drinking water sample meets the criteria for acidification only, no further sample preparation. Coupled with the capital investment information shown in Table 6, the optimum scenario for the lab can be determined, based on sample type and load.
Conclusion
The lower arsenic compliance level provides additional analytical challenges. However, arsenic is only one of a suite of elements that must be measured, so the entire set of measurements should be considered in choosing the appropriate analytical technique.
Techniques other than those directly listed in the CFR can provide sufficient analytical capability and offer additional flexibility to the lab managing a complex workload. The shift of arsenic from ICP-OES measurement may tip the balance towards an ICP-MS investment, depending upon the overall lab workload and analytical capability.
Even lower arsenic compliance levels may be in our not too distant future.
References
1. KRGV TV News story, http://www.newschannel5.tv/2007/4/5/966323/-Arsenic-Levels-in-Water-a-Little-High-
2. FDA Press Release, FDA Warns Again About Arsenic in Mineral Water, Five Brands Recalled Within Last Month, http://www.fda.gov/bbs/topics/NEWS/2007/NEW01594.html
|
