Open-File Report 2006–1350
U.S. GEOLOGICAL SURVEY
Open-File Report 2006–1350
Sampling sites were spaced throughout the reservoir to capture the areal distribution of trace element concentrations in sediment from the inlet to the Grand Coulee Dam (fig. 1). In addition, one site with minimal influence from the downstream movement of slag material, on the Sanpoil River, was selected to serve as representative of background conditions. Large landslides of terrestrial material along the shoreline, cited as causing anomalies in the chemistry profiles of previous sediment cores on the lake (Cox and others, 2005), were clearly visible and were avoided during the site selection process. Operation of the dam can result in changes in water level within the reservoir by as much as 24 m annually, and dewatering has been demonstrated to affect the biogeochemical cycling of metals and suggested as a mechanism that controls the population and diversity of the benthic community (Bortelson and others, 2001). To minimize this potential effect, sediment sampling locations were selected only where water depths were greater than 12 m and had been permanently watered for the previous 2 years.
One site per day (table 1) was sampled by boat between September 21 and 28, 2004, at water depths between 12 and 18 m using a 13.5 × 13.5 × 23-cm deep box corer, following the protocol for collection and processing described by the U.S. Environmental Protection Agency (2001). Four different types of sediment subsamples were collected at each site: (1) one 6.5-cm diameter sediment core (6.2 to 12.3 cm long) with intact overlying reservoir water for the sediment incubation experiments, (2) numerous 6.5-cm diameter plugs of the top 2 cm of sediment for extraction of porewater, (3) one 4.4-cm diameter core of up to 10 cm in length for evaluation of the vertical trends in elemental composition of the sediments, and (4) sediment within the top 10 cm taken from numerous box cores for the preparation of the large-volume composite sample (10 L). Sediment from the composite sample was used to analyze the grain-size distribution, loss on ignition (LOI) by heating to 450ºC for 4 hours, the elemental composition of the solid sediments, and the elemental composition of three operationally defined solid fractions sequentially and selectively extracted from the sediments.
Some difficulties were experienced during sampling. For example at site LR-2, the corer was not closing properly when sampling at a depth of 13 m because gravel-sized sediments jammed the closing mechanism; and the site was moved to deeper water (about 26 m) for all sediment collection. Obtaining a sealed core for the incubation experiment was difficult because the box corer usually penetrated the sediments only 4 to 6 cm at water depths of 13 m and the reservoir water often leaked at the corners of the cores. The sandy texture of the sediments at LR-1 prevented the box corer from forming an intact seal, and no incubation core was obtained at this site. Alternative sites LR-4A (about 30 m) and LR-5A (about 25 m) that allowed greater penetration of the sampler (table 2) were sampled in order to collect intact cores for the incubation experiments. One porewater sample and one sediment sample were analyzed at these two alternative sites. The coarse texture of the sediment from site LR-2 at 13 m, which is within the elevation range of drawdown, probably is a result of winnowing of finer materials by wave action during the annual reservoir drawdown as the water surface of the reservoir dropped below the elevation from which sediments were taken.
All plasticware (acrylic square and tubular core liners, styrene weighing pans, Teflon® spatulas, polypropylene spoons, beaker covers, centrifuge tubes and closures, all-plastic syringes, Teflon® syringe tips, and high-density polyethylene (HDPE) jars for sediments and bottles for water samples) for this study was cleaned with Liquinox®, rinsed with de-ionized water (DI), soaked in 5 percent trace-metal grade HCl, and rinsed with DI. Plasticware that came in contact with water samples analyzed for aqueous elemental concentrations was additionally soaked in 4 N trace-metal grade HNO3, rinsed with DI water, rinsed with 0.16 N Optima HNO3, rinsed with DI water, and dried in a laminar flow hood. The lot of acid-cleaned 25-mm Millex Durapore® in-line filters was pre-tested for contamination with the acid-cleaned, all-plastic syringes before executing the field sampling, and all results were acceptable.
Upon bringing the box corer onboard with a sealed core, a tubular 6.5-cm diameter acrylic core liner was inserted into the middle of the 13.5 × 13.5-cm box-core liner in a manner in which the intact reservoir water was preserved. The water outside of the inner acrylic core liner was siphoned off and the remainder of the sediments in the box core outside the acrylic core liner was then removed to the compositing bowl. The bottom of the acrylic core was then sealed with Parafilm® held by a rubber cap. The gritty nature of the slag material collected at site LR-7 caused it to break through the Parafilm® seal and required that the incubation core liner be pre-sealed. Duplicate sediment samples from site LR-7 were collected in two 4.4-cm diameter core liners from two different box cores (LR-7 and LR-7R), which were then carefully inserted into the larger 6.5-cm diameter pre-sealed incubation core liners. After the smaller core liners were removed and the sediment settled, the larger core liners were each filled with different samples of surface water from the site. The tops of incubation core liners were sealed with Teflon® tape and polypropylene beaker covers that contained a center hole for the stirrer and two small holes for Teflon® syringe tips. The tops of the sealed cores were wrapped in Parafilm® and the cores were carefully placed in a carrying caddy in a cooler that was packed with ice for transport to the USGS field laboratory at Fort Spokane, Wash., where core incubation experiments began within 8 hours and continued for about a month.
Sediment for porewater extraction was obtained by inserting 4.4-cm diameter core liner into the middle of the box core to the bottom for core samples that were 4 to 6 cm long. After any reservoir water that had not leaked from the corners of the box core was siphoned off, sediment around the tubular core liner was moved aside. The 4.4-cm diameter plug of sediment was excavated by sliding a thin styrene tray under the core liner and removing the core liner containing the sediment plug from the box core. The tubular core liner was removed from the tray, and the top 2 cm of the sediment plug was packed into a 50-mL polypropylene centrifuge tube with a polypropylene spoon. Precautions to prevent sediment contact with atmospheric oxygen were not taken because the sediments had already been exposed when reservoir water leaked out of the poorly sealed box corer. After sealing the centrifuge tube closure with Parafilm®, centrifuge tubes from multiple box cores were double sealed in plastic bags and transported to the field laboratory on ice. Within 8 hours, the sediment was centrifuged at 2,000 revolutions per minute (rpm) for 20 min. Within a portable laminar flow hood, the supernatant in the centrifuge tube was withdrawn from the centrifuge tube into an all-plastic syringe, filtered through a 25-mm 0.22-µm pore size Millex Durapore® in-line filter into a 30-mL HDPE bottle, and preserved to a concentration of 0.16 N Optima HNO3. If sufficient water was extracted (a minimum of 6 mL), porewater from individual cores was analyzed. Otherwise, water from multiple cores was composited before analyses. Water from each site was collected either from a draining or sealed box corer, or from the surface of the water column (sites LR-5A and LR-7), filtered, preserved, and stored in a HDPE bottle to compare to porewater concentrations. Additional reservoir water was collected either from sealed box cores, draining box cores, or surface water in 500-mL HDPE bottles to replace water withdrawn for sample analysis during the incubation experiments, as described later in this report.
At sites LR-2, LR-3, LR-6, and SA-8, a second core was collected that was sufficiently long to warrant collecting sections of a vertical core. A 4.4-cm diameter core liner was inserted into the middle of the box core and all reservoir water was siphoned off. All sediment to a depth of 10 cm around the tubular core liner was added to the composite bowl. Parafilm® on an aluminum plate was slid under the tubular core liner and the assembly was placed on an extruder piston. The plate was removed and the sediment was sealed onto the piston by the Parafilm®. The core liner was manually moved down the piston in 1-cm increments, which exposed vertical core sections of sediments that were placed into plastic bags. One-cm sections from deeper sections of the cores at sites LR-2 (7-10 cm), LR-3 (5-6 cm), and LR-6 (3-4 cm), were combined before elemental analysis. Likewise, 1-cm sections (5‑7 cm) of a longer-than-average core (7 cm) from SA-8 were combined. The sediment remaining in the centrifuge tubes from 0-2 cm samples after the extraction of porewater was analyzed for elemental composition to assess variability in elemental composition of sediments at sites LR-1, LR-4, LR-5, and LR‑7, where the box core only penetrated about 4 cm.
Sediment from multiple box cores (between 6 and 18 cores) was required to obtain 10 L of sediment to prepare the composite sample for toxicity testing by the USGS Columbia Environmental Research Center (CERC) laboratory in Columbia, Mo. (table 2). After examination by archaeologists for cultural resources using plastic utensils, sediment outside of the tubular core liners used to collect incubation cores, porewater samples, and vertical sections (to a depth of 10 cm) or sediment from the entire surface area of the box core to a depth of 10 cm was added to the compositing bowl. After a sufficient volume of sediment was collected, the composite was manually mixed for about 60 seconds. Aliquots from the compositing bowl were placed in a 500‑mL HDPE wide-mouth bottle for elemental analysis and a 200‑mL polypropylene jar for sequential selective extraction. A replicate composite sample (LR-7R) was collected at site LR-7 from additional box-core samples. Aliquots of the composite samples for elemental composition were shipped to the USGS Minerals Laboratory in Denver, Colo. Samples for analysis by sequential selective extraction were first shipped to the USGS Water Science Center in Austin, Tex. for freeze drying, and were then shipped to the USGS Minerals Laboratory for later analysis. Aliquots of composite samples (two 3.79‑L low-density polyethylene jugs) for analysis of grain size, loss on ignition, acid volatile sulfide (AVS), and sediment composition of selective elements were transported on ice first to the field laboratory, refrigerated at the field laboratory until the field exercise was completed, then shipped on ice to the USGS CERC laboratory.
Incubation experiments were designed to simulate potential release of aqueous elements under lake-like conditions. Upon returning to the field laboratory at Fort Spokane, Wash., each evening, the incubation core was placed in a holder in one of two incubators set at the temperature of the bottom reservoir water (approximately 18º C for sites LR-2, LR-3, LR‑4A, and SA-8; and approximately 14ºC for LR-5A, LR-6, LR-7, and LR‑7R. The polypropylene beaker cover was removed and site water was added to fill the core to the top. The rod of a polypropylene paddle was inserted through the center hole of the cover (leaving a gap less than 1 mm). The cover was resealed with Teflon® tape in a manner that positioned the paddle in the middle of the overlying water (fig. 2). The rod end of the paddle was attached to a motor set at 4 rpm, which was run continuously for the duration of the experiment. Two Teflon® syringe tips attached to all-plastic syringes were inserted to the depth of the paddle through press-fit holes. After about an hour, an 11-mL sample was withdrawn from the overlying water with the withdrawal syringe and filtered through a 25-mm 0.22-µm Millex Durapore® in-line filter into a 15-mL HDPE bottle under the confines of a laminar flow hood and preserved to a final concentration of 0.16 N Optima HNO3. After the withdrawal syringe was placed back on its tip, the replacement syringe was removed from the second inserted tip and filled with 11 mL of site water to replace the first 11-mL aliquot. The 11 mL of replacement water was then slowly added to the overlying water. This removal and replacement process was repeated at intervals of 2 and 4 days, and approximately 8, 16, and 32 days, with a duplicate sample being taken at approximately 16 days. In addition to the small gap between the cover and the stirring rod, the overlying water was in contact with the atmosphere only during portions of the 15 minutes required to process a sample through the 10‑cm‑long syringe tip.
At the end of the month-long incubation experiment (November 3, 2004), the reservoir water in each core was siphoned into a receptacle to measure the volume and pH. Each sediment core was then placed on an extruder, the bottom rubber cap cut away, and the top 2 cm was extruded into a 125-mL wide-mouth HDPE bottle and transported on ice to the USGS Washington Water Science Center, Field Service Unit in Lakewood, Wash., for the higher energy tumbling experiment. The lower portions of each core were placed in polypropylene jars, frozen, and transported to the USGS CERC laboratory for analysis of dissolved organic carbon in the porewaters.
The extent of release of metals from sediments is partially determined by the extent of mixing of water and sediments. The physical configuration of incubation cores and the slow speed of the paddle in the incubation experiment provided a low level of physical interaction between sediments and water, similar to a lake environment. The sediment from the top 2 cm from each incubation core was further exposed to water under a higher energy tumbling experiment, similar to a higher energy riverine environment. Approximately 30 mL of site water was added to each pre-weighed 125‑mL wide-mouth HDPE bottle containing the sediment and the bottle was reweighed. In addition, gritty sediment at the bottom of incubation cores LR-5A (4.7-6.2 cm) and LR-6 (4.2 and 5.2 cm) also were tumbled with site water. Two samples of approximately 150 grams of river sediment with slag collected by Cox and others (2005), and labeled as RM 743 also were tumbled with approximately 80 mL of LR-7 site water. The bottles containing the slurries were sealed and attached to a vertical 1-m wheel attached to a motor, which caused the bottles to be turned end over end seven times a minute. After 15 minutes, the wheel was stopped and the slurry was allowed to settle. Within a laminar flow hood, 11 mL of supernatant was withdrawn, filtered through a 25-mm 0.22-µm Millex Durapore® in-line filter into 15-mL HDPE bottles and preserved to a final concentration of 0.16 N Optima HNO3. The 125-mL bottles were placed on the rotating wheel for 43 days, at which time the samples were taken off the wheel and allowed to settle for one day before the supernatant was filtered, and preserved. The 125-mL bottles were then placed in an oven and the sediment dried at 105ºC to obtain the mass of the dry sediment and the percent moisture of the original sediment in the top 2 cm of the 6.5‑cm diameter core, and the water:sediment ratio of each slurry.
Upon arrival at the USGS Minerals Laboratory in Denver, Colo., the aliquots from the composite samples, vertical sections, and sediments from the extraction of porewaters were oven-dried (table 2). The sediments were then digested using hydrochloric, nitric, perchloric, and hydrofluoric acids (lab code T20 as described in Briggs and Meier, 2002). This digestion procedure is not suitable for refractory elements. Laboratory quality-assurance procedures included digestion of the standard reference material (SRM) MAG-1 (duplicate), National Institute of Science and Technology (NIST) 2704, and SCO-1 and analysis of the digestion solution. A duplicate of standard National Research Council of Canada (NRCC) PACS-2 and the standard International Atomic Energy Association (IAEA) ‑405 were submitted as field samples. In addition, the total digestion of the eight composite samples using the four-acid digestion was compared to the strong-acid digestion used by the USGS CERC laboratory.
The sequential selective extractions procedure partitioned the trace elements within the sample into four operationally defined fractions that have been related to the bioavailability and mobility of elements. Discussion and assessment of sequential selective extraction techniques are summarized by Tessier and others (1979) and Chao (1984). Sequential selective extraction procedures of the composite sediment samples followed those of the European Commission of Standards, Measurements, and Testing (Rauret and others, 2001) and were evaluated by comparison to results of sequential selective extraction of the freshwater reference sediment BCR-701. The first fraction of the sequential selective extraction is presumed to contain elements weakly sorbed to the surface of sediment particles and elements associated with carbonate coatings. The second fraction contains elements associated with iron and manganese oxide coatings on the surface of sediment grains. The third fraction contains elements associated with organic matter and sulfide mineral coatings phases. Solid material remaining after the three extraction treatments is considered the recalcitrant residual material consisting largely of independent mineral grains.
Aliquots of the composite samples from each of the eight sites for analyses of sequential selective extracts were freeze-dried in Austin, Tex., beginning on October 1, 2004. The frozen samples were placed in a Labconco Freeze drier with condenser temperature of less than -40ºC. The atmosphere in the freeze-drier was evacuated to facilitate sublimation of the ice from the sediment-ice mixture. Samples were held in the freeze-drier for 4 days to complete the drying process. During the second, third and fourth days, progressive warming of the evacuated chamber up to 60ºC was applied to facilitate drying. When completely dry, the sediment samples were resealed in their original containers and forwarded to the USGS Minerals Laboratory for analysis.
A four-step sequential selective extraction procedure was performed on the aliquots of each of the composite sediment samples from the eight sites and the standard reference material BCR-701 in January 2005 at the USGS Minerals Laboratory. As part of the laboratory quality assurance, a process blank and wash blank were analyzed to assess bias and two pairs of duplicate sediment samples were used to assess variability in the trace-element concentration data for each step of the sequential selective extraction process. The process blank consisted of processing the appropriate reagent for each of the four steps through the extraction procedure in the absence of sediment. Following each step of the sequential selective extraction, the extract fraction was washed with DI water and the wash solution of each sample was composited to provide an average wash blank to quantify the amount of metals potentially lost during the rinsing process of each step. One of the duplicate pairs was the standard reference material BCR-701, for which certified concentrations of cadmium, chromium, copper, lead, nickel, and zinc have been determined for the sequential selective extraction procedures used in this study. The other sediment sample used for duplicate analysis was the aliquot from the composite sample from site LR-5.
For fraction 1, 40 mL of 0.11 M acetic acid was added to an acid-cleaned polypropylene centrifuge tube containing 1 g of sediment and the tube was shaken on a Fisher Scientific mechanical rotator, set at 18 rpm for 16 ± 0.5 hours at 23ºC. After centrifuging for 20 minutes at 15,000 rpm, the supernatant was decanted into an acid-cleaned polypropylene centrifuge tube and refrigerated until analysis by inductively coupled plasma-mass spectrometry (ICP-MS) within 10 days. The sediments were washed with 20 mL of DI water by shaking for 15 minutes in a centrifuge tube, the tube was centrifuged for 20 minutes at 15,000 rpm, and the supernatant decanted into an acid-cleaned polypropylene centrifuge tube for wash-blank analysis. Additionally, a process blank was prepared by submitting DI water to all processes of the fraction 1 extraction.
For fraction 2, 40 mL of 0.5 M hydroxylamine hydrochloride was added to the tube containing the washed sediment remaining after the procedure for fraction 1. The sediment then was shaken for 16 hours at 23ºC. The slurry was then centrifuged, the supernatant decanted into a separate bottle, and the sediments were then washed as in procedure for fraction 1 for a fraction 2 wash blank. A process blank also was prepared for fraction 2.
For fraction 3, the washed sediment remaining after removing fraction 2 was digested in 10 mL of 8.8 M H2O2 and stabilized to pH 2-3 for 1 hour at 22ºC. The digestion was continued for another hour at 85ºC, allowing evaporation to decrease the volume to 3 mL. Another 10 mL of H2O2 was added to the sediment and digested for a third hour at 85ºC, allowing the volume to decrease to 1 mL during the last 30 minutes. Fifty mL of DI water was added to the centrifuge tube, which was then shaken for 16 hours at 22ºC. After centrifugation, the supernatant was decanted into a separate bottle, and preserved to a concentration of 0.16 N HNO3 (pH of 2). The residual sediment was then washed as in the procedure for fraction 1 for a wash blank. A process blank also was prepared for fraction 3.
For fraction 4, the residual sediment was transferred to a Teflon® beaker for digestion of the residual fraction using a four-acid mixture of HCl, HNO3, HF, and HClO4. A process blank was also prepared for fraction 4. In some instances, the transfer of residual sediment from the extraction vessel to the Teflon® digestion beakers was incomplete with small amounts of residual sediment adhering to the 50 mL centrifuge tube used during the extraction procedure.
The four-acid digestion solutions of the sediment samples, laboratory controls, and blind SRMs were analyzed by ICP-MS on an Elan 6000 with dwell times between 3 and 50 milliseconds (Lab code T20 described in Briggs and Meier, 2002) at the USGS Minerals Laboratory. The ICP‑MS was calibrated to a 0.16 N HNO3 acid-blank solution and two multi-element standard solutions to cover the mass range and generate the mass response. Laboratory quality assurance included analysis of wash water, duplicate analyses of LR-1 composite at the beginning and at the end of the run, duplicate analyses of LR-1 in sequence at the end of the run, analysis of SRMs NIST 2704, SCO-1, and duplicate analyses of SRM MAG-1 during the run. In addition, PACS‑2 (analyzed in duplicate) and IAEA-405 were submitted from the field as blind samples. The relative percent difference (RPD) for analyses of the LR-1 composite and MAG-1 generally was less than 10 percent except for antimony, beryllium, cerium, lanthanum, molybdenum, scandium, selenium, thorium, titanium, and uranium (table 3; at back of report).
The average percent difference of laboratory controls MAG‑1, NIST 2704, and SCO-1 from the assumed true values suggested by Potts and others (1992) generally were less than 10 percent except for chromium (+10.8 percent), molybdenum (-29.5 percent), thallium (+17.4 percent), titanium (-14.0 percent), and yttrium (‑13.6 percent). Average results of the analyses of the SRM PACS‑2, submitted blind from the field, were within 10 percent of all certified values, except for uranium (‑23.3 percent), and results for the single analysis of the IAEA-405 SRM differed from the certified value for a number of elements: antimony (‑22.7 percent), chromium (+11.1 percent), lanthanum (-17.3 percent), lithium (+16.5 percent), and strontium (+23.7 percent). Analysis of field duplicates from LR-7 generally resulted in RPDs of less than 10 percent except for beryllium (11.8 percent), bismuth (47.1 percent), cadmium (15.4 percent), and thorium (17.7 percent). The concentrations and RPDs of the major elements and transition elements by the four-acid digestion generally were within acceptable ranges. Significant variations and deviations from certified or suggested values were found for the lanthanide and actinides elements and the results for these elements are provided for informational purposes only. There were considerable differences in the elemental concentrations of arsenic, cadmium, copper, lead, and zinc between the method using the four-acid digestion and analysis by the USGS Minerals Laboratory and the method using the strong-acid digestion and analysis of a different aliquot of the composite sample by the USGS CERC laboratory (table 4; at back of report). A four-acid digestion extracts more material phases than the strong-acid digestion, especially silicate-lattice minerals, which could explain the lower values obtained by the strong-acid digestion, especially for lead. The higher values returned by the strong-acid digestion were most likely a result of actual differences in the two aliquots taken from the composite sample due to its heterogeneity resulting from the coarse nature of the composite samples, such as the coarser material analyzed from site LR-7.
Calibration standards for the analyses of elements in leachates from fractions 1 and 2 of the sequential selective extraction were prepared in 0.11 M acetic acid and 0.5 M hydroxylamine hydrochloride, respectively, and both standards and samples were diluted 1:10 before analyses. The elemental concentrations in the wash blanks were less than the instrumental limit of detection (ILOD) or insignificant. The concentrations of the elements detected in the process blank water (table 5, at back of report) were near analytical reporting limits and typically less than 1 percent of the concentrations determined for the environmental samples, and thus generally not a significant source of bias in the sample data. The greatest degree of bias was found for chromium, for which the concentrations of seven samples were within twice the process blank for fraction 1 (0.3 µg/g), and for which concentrations of eight samples were within three times the process blank for fraction 3 (2.3 µg/g). Therefore, Cr concentrations for fraction 1 and 3 are rated as estimated. Lead was detected in the process blank of fraction 1 of the sequential selective extraction at 0.01 µg/g and of fraction 2 of the sequential selective extraction at 0.04 µg/g. Lead concentrations in process blank samples were less than 1 percent of reported sample concentrations in all instances, except for three samples of the first fraction where the process blank concentrations were less than 4 percent of the reported sample concentrations. Magnesium was detected in some process blanks at concentrations ranging from 2.4 to 5.5 µg/g; in all instances magnesium present in process blank samples was less than 1 percent of sample concentrations. The process blank for the analysis of residual sediment (fraction 4) contained concentrations of aluminum, barium, copper, gallium, lead, magnesium, potassium, rubidium, strontium, and zinc; all concentrations except copper were much less than 1 percent of reported concentrations. Copper was present in the process blank of fraction 4 at 0.2 µg/g, which was equivalent to about 3 percent of the lowest concentrations of the environmental samples.
Variability in the first three steps of the sequential selective extraction process was assessed by comparison of the concentrations measured in this study to the reported certified concentrations of cadmium, chromium, copper, lead, nickel, and zinc in the standard reference material BCR-701. These data (table 5) show that, except for the cadmium and lead determinations in fraction 3 and the lead determinations in fraction 1, recovery of the expected concentration typically was within 10 percent of the certified value; although concentrations had a slight tendency to be lower than expected (average recovery was 93 percent). In fraction 3 of the extraction process, the measured lead concentrations were only 50 percent of the certified values. Because the average percent differences from the certified value for the duplicate samples was 49.5 percent, the reported lead concentration data for fraction 3 in the extraction process also is likely to be biased low by a significant amount. Low recovery of lead in fraction 3 of the sequential selective extraction process, on the order of 50 percent, also was noted by Rauret and others (2001) in the concentration data used to obtain the certified concentrations.
Variability in the sample data set was assessed by the comparison of results from duplicate analysis of the same sample. Variability in duplicate samples from the sequential selective extractions typically was less than 10 percent. Variability, as measured by relative percent difference in concentrations determined on the residual sediment, following the sequential selective extraction however, was quite large, as much as 91.5 percent with an average of about 50 percent. Variability increased for the procedures for the subsequent fractions in the sequential selective extraction process. Average variability in relative percent differences in step 1 were about 3 percent, increasing to about 4 percent the second step and more than 5 percent in the third step.
Site water, porewaters, overlying water of the incubation cores, and supernatants from the tumbling experiment were diluted by a factor of 5 with 0.16 N HNO3 before analysis by ICP-MS on a Perkin-Elmer Elan 6000 with 35 sweeps in the peak-hopping mode, with dwell times ranging from 10 to 100 milliseconds. The instrument was calibrated with standards prepared in 0.16 N HNO3 (Lab codes T21 and T23 in Lamothe and others, 2002). Internal standards (500 µg/L lithium-6, 20 µg/L rhodium, and 10 µg/L iridium) were mixed in a 1:1 ratio with samples using a dual-channel peristaltic pump equipped with a mixing manifold and coil. The isotopes used to measure elemental concentrations were selected to minimize isobaric overlap from other elements, and internal calculations correct for isobaric interferences for arsenic, cadmium, dysprosium, gadolinium, germanium, iron, molybdenum, tungsten, selenium, and vanadium. Laboratory quality-control samples that are not included in this report include analysis of solution blanks, duplicates, and NIST SRMs 1640 and 1643d. Field quality-control samples included analysis of field-laboratory filtering blanks, duplicates of the incubation samples at about 16 days, and eight samples of five reference samples standards (NRCC SLRS-3, USGS Standard Reference Sample T-171 and T-179, and NIST 1640 and 1643e).
The average concentration of only manganese in the DI water-filter blanks (after passing through pre-cleaned in-line filters) was greater than the ILOD and the reporting limit for manganese was adjusted to 5 µg/L, based on 3 times the standard deviation of concentration of the field laboratory filter blanks (table 6; at back of report). ILODs are calculated based on the variation of the instrumental signal and could be a result of a number of factors, including polyatomic interferences, impurities in reagents, or random contamination in the laboratory or field.
The decision to increase the reporting limit by a factor of 5 times the ILOD to account for the 5-fold dilution of samples prior to analyses was based on: (1) the replication and recovery of the double blind standards as a function of increasing concentration, (2) the replication of field duplicates as a function of increasing concentration, (3) the internal consistency of temporal trends in the incubation samples, and (4) the amount of data with concentrations between 1 and 5 times the ILOD. The reporting limits of all elements for which SRM data were available were increased by a factor of 5, except for antimony, arsenic, copper, molybdenum, nickel, and sodium (table 7; at back of report).
The recovery of copper in the low-concentration SRM NRCC SLRS-3 was low (57 percent), and the RPD of the nickel analysis was 50 percent. Additionally, recovery of magnesium in SLRS-3 was very low (42 percent) and recovery of manganese was very high (260 percent). Concentrations of all elements for which there were no reported SRM values for these five standards are provided for ancillary information purpose only. The original ILOD was accepted if the RPD at concentrations near the ILOD were low. If the RPDs for duplicates at concentrations between 1 and 5 times the ILOD were unacceptable or inconsistent, the reporting limit was increased by a factor of 5. The assessment of the quality-control data for each element is provided in table 7. The results for single elements from various samples were flagged as outliers in the results tables and the disregarded values are provided in table 8 (at back of report). In addition, the major cation data from one of the duplicate samples from the LR‑5A incubation core at about 19 days (LR-5AINC19A) was inconsistent with the major ion data from other samples of the overlying water and the results from the entire sample were disregarded.
The uptake or release of elements by the apparatus for the incubation experiments was tested by running a control experiment (no sediment) containing site water from the lower reservoir held in a pre-sealed core liner with the stirring paddle and syringe tips inserted through the sealed cover. Triplicate samples were withdrawn and filtered after 1 hour and after 34 days. The concentration of alkali and alkaline earth elements and silica increased within a narrow range between 20 and 33 percent (table 9; at back of report), but these concentrations were statistically the same (p values between 0.07 and 0.32). For example, concentrations of magnesium increased from 5.0 ± 0.1 mg/L to 6.0 ± 1.1 mg/L. The consistent increase over the 34 days may have been a result of evaporation within the incubator and loss of water vapor through the 1 mm gap between the cover and the rod of the stirring paddle. The low concentrations of aluminum and antimony at the beginning of the control test did not increase beyond the precision of the analyses (approximately 10 percent). At the end of the control experiment, copper was detected in one sample at the reporting limit and in one sample at twice the reporting limit in one sample. At the end of the control experiment, the average zinc concentration was 3.3 µg/L and the reporting level for zinc values for only the incubation experiments was set at 3.3 µg/L.
U.S.
Department of the Interior | U.S. Geological
Survey
Persistent URL: http://pubs.water.usgs.gov/ofr20061350
Page Contact Information: Publications Team
Page Last Modified: Saturday, 12-Jan-2013 22:20:38 EST
