Questa baseline and pre-mining ground-water-quality investigation. 16. Quality assurance and quality control for water analyses

Abstract

The Questa baseline and pre-mining ground-water quality investigation has the main objective of inferring the ground-water chemistry at an active mine site. Hence, existing ground-water chemistry and its quality assurance and quality control is of crucial importance to this study and a substantial effort was spent on this activity. Analyses of seventy-two blanks demonstrated that contamination from processing, handling, and analyses were minimal. Blanks collected using water deionized with anion and cation exchange resins contained elevated concentrations of boron (0.17 milligrams per liter (mg/L)) and silica (3.90 mg/L), whereas double-distilled water did not. Boron and silica were not completely retained by the resins because they can exist as uncharged species in water. Chloride was detected in ten blanks, the highest being 3.9 mg/L, probably as the result of washing bottles, filter apparatuses, and tubing with hydrochloric acid. Sulfate was detected in seven blanks; the highest value was 3.0 mg/L, most likely because of carryover from the high sulfate waters sampled. With only a few exceptions, the remaining blank analyses were near or below method detection limits. Analyses of standard reference water samples by cold-vapor atomic fluorescence spectrometry, ion chromatography, inductively coupled plasma-optical emission spectrometry, inductively coupled plasma-mass spectrometry, FerroZine, graphite furnace atomic absorption spectrometry, hydride generation atomic spectrometry, and titration provided an accuracy check. For constituents greater than 10 times the detection limit, 95 percent of the samples had a percent error of less than 8.5. For constituents within 10 percent of the detection limit, the percent error often increased as a result of measurement imprecision. Charge imbalance was calculated using WATEQ4F and 251 out of 257 samples had a charge imbalance less than 11.8 percent. The charge imbalance for all samples ranged from -16 to 16 percent. Spike recoveries were performed by spiking ground-water samples from SC2B, SC3A, SC3B, CC2A, and Hottentot with a mixed-element standard and then analyzing them by ICP-OES. The mean recovery for all the constituents by ICP-OES was 103 percent with a standard deviation of 16 percent. Fifteen surface- and ground-water sequential duplicates were collected from Straight Creek, Hottentot, and the Red River from 2002 to 2003. Except for chloride from well SC5B and low concentrations of iron (<0.05 mg/L) and aluminum (<0.01 mg/L), constituents of sequential duplicates are generally within 10 percent of each other. Analytical results from different methods and different laboratories, with rare exceptions, were within 10 percent. Chromium analyses were in poor agreement when comparing analyses from the USGS and a contract laboratory, but USGS analyses by ICP-OES and ICP-MS were usually within 10 percent for chromium concentrations above 0.03 mg/L and analyses by ICP-OES and GFAAS were usually within 15 percent for chromium concentrations as much as 0.1 mg/L. Filtration studies also were performed to study the effects of filtration apparatuses (Minitan, plate, capsule, and syringe), pore sizes, and timing on dissolved metal concentrations. Except for iron and aluminum, constituents with concentrations greater than about 0.05 mg/L were generally not affected by the filtration apparatus, membrane pore-size, and filtration delays. Iron, aluminum, and some dissolved metals concentrations less than about 0.05 mg/L, especially copper, were generally lowest in filtrates from the tangential flow Minitan system containing a filter membrane with a pore size of 10,000 Daltons. As part of a filtration timing study, grab samples were collected from two sites along the Red River and were processed immediately and then again 1 to 3 hours later. Aluminum and iron colloids formed during the delay in the sample collected at the USGS gaging station and, after the delay, 0.1-ìm filtrate aluminum and iron concentrations approached the ultrafiltrate (Minitan) concentrations. In the upstream site below Fawn Lakes, aluminum in the 0.1-ìm filtrate decreased but did not decrease in the 0.45-ìm filtrate, signifying that the colloids formed during the delay are between 0.1 and 0.45 ìm. Dissolved nickel and pH also decreased in both samples during the delay. Except for ferrous iron and barium, a sequential filtration study 2 demonstrated that water collected from the Red River at the gage did not affect dissolved metal concentrations with increasing sample volume passing through a plate filter with 0.45- or 0.1-ìm membranes. Barium and ferrous iron both slightly decreased in the filtrate from the 0.45-ìm filter.