Environ. Sci. Technol. 2002, 36, 2826-2832
At-Sea High-Resolution Trace Element Mapping: San Diego Bay and Its Plume in the Adjacent Coastal Ocean BRADLEY K. ESSER* AND ALAN VOLPE L-231, Lawrence Livermore National Laboratory, Livermore, California 94550
We conducted at-sea experiments using separation chemistry and an inductively coupled plasma mass spectrometer to rapidly measure, in real time, trace element concentrations in the surface water of the San Diego Bay and the adjacent coastal ocean. The survey shows that surface water in the San Diego Bay is clearly enriched in Mn, Ni, Cu, Zn, Cd, and FDOM with respect to the coastal ocean. Trace element enrichment patterns were used to identify bay water as it enters the coastal ocean during tidal pumping. Quantifying the spatial distribution and temporal flux of pollutant trace elements are important steps in understanding coastal biogeochemical processes.
Introduction The San Diego Bay ranks as one of the most contaminated coastal areas in the United States. This ranking is based on high levels of organic and trace element pollutants in sediments and fish and demonstrated biologic effects in fish (1, 2). Quantifying the flux of trace elements out of the bay is important in determining the major sources and sinks of pollutant elements in the heavily impacted Southern California Bight. There have been studies of trace element distribution in the water column of the San Diego Bay (3-5), the adjacent coastal ocean (6-9), and in bay sediments and biota (1, 2, 10, 11). Recent studies have constrained the physical exchange between the San Diego Bay and the coastal ocean and demonstrated that mass fluxes are dominated by tidal pumping during flood and ebb cycles (12, 13). Still, the fate and dispersal of chemical components in plumes from the San Diego Bay, as well as the Mission Bay and the Tijuana River, have been little studied. These plumes are the primary mechanism for the exchange and transfer of nutrients, organisms, and trace elements between bays and the coastal ocean. Details of the rates of exchange of nutrients, pollutants, and trace elements in mixed waters, the beneficial or detrimental effects of these trace elements on biologic productivity, and the short-term fate of dissolved and particulate material as it flushes from the bays require further investigation. A first step in understanding these aspects of coastal biogeochemical processes is to quantify the dynamic distribution, both temporal and spatial, of important biologic and chemical species. In this context, we developed an enhanced ocean observational capability as a first step toward the full characterization of the impact of harbor, bay, and lagoon outflow * Corresponding author phone: (925) 422-5247; e-mail: esser1@ llnl.gov. 2826
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 13, 2002
on coastal waters along Southern California. At-sea science experiments in the San Diego Bay and Bight were conducted on September 13-17, 1999, aboard the R/V Robert G. Sproul with the collaboration of investigators from the Scripps Institution of Oceanography (SIO). The main purpose of this survey was to use an at-sea inductively coupled plasma mass spectrometer (ICPMS) to quantify elemental composition of waters in Southern California embayments and the coastal zone in conjunction with simultaneous sampling of invertebrate larvae (crabs, barnacles, bivalves). During the course of the survey, we mapped Mn, Ni, Cu, Zn, and Cd in the shipping channel of the San Diego Bay from Point Loma to the 24th Street Marine Terminal in National City and in the coastal ocean waters of the San Diego Bight.
Instrumentation Hydrographic Instrumentation. Continuous oceanographic data were collected using a towed body equipped with a Seabird Electronics (Bellvue, WA) SBE 911 underwater CTD system that included a Chelsea Aquatracka UV-fluorometer, a WETLabs miniature chlorophyll a fluorometer, and a SeaTech transmissometer. Temperature, salinity, fluorescent dissolved organic matter (FDOM), chlorophyll a and light transmissivity data were collected every 1.5 s during the survey and median-filtered after the survey to remove spikes. Details on the calibration, precision, and accuracy of the hydrographic instrumentation are described elsewhere (14, 15). A titanium/Teflon pump on the towed body delivered seawater to the shipboard analytical van for direct trace element analysis. Inductively Coupled Plasma Mass Spectrometry. A Hewlett-Packard (Agilent, Wilmington, DE) HP4500 inductively coupled plasma mass spectrometer was used in these experiments. The mass spectrometer was modified for use at sea, with shock mounting to isolate it from high-frequency shipboard vibration and by housing it in a mobile cleanroom (14). The instrument was run with Ni sampler and skimmer cones and a standard Babington nebulizer and Peltier-cooled double-pass spray chamber. The only change from standard tuning conditions was a doubling of the nebulizer peristaltic pump speed to 0.2 rps to reduce the analytical cycle and enhance detection limits. On-Line Preconcentration. A Hewlett-Packard (Agilent) integrated sample introduction system (ISIS) was used as the flow-injection accessory and comprised two low pulsation three-channel peristaltic pumps and two six-port two-way electrically actuated switching valves. For discrete samples (standards and calibration solutions described in this study), a Cetac ASX-500 autosampler was used. For continuous sampling of seawater in the field, seawater was pumped directly from a towed titanium/Teflon submersible pump to a Teflon manifold in the clean-room van (14). All seawater was filtered through acid-cleaned highcapacity 0.45-µm filters (Geotech, Denver, CO). Both the autosampler and sample manifold were connected to the ISIS unit by 1.4 m of 0.5 mm i.d. Teflon FEP tubing. Control of the autosampler and flow-injection module was fully integrated into the ICPMS operating software. Teflon FEP and PFA tubing and Teflon flangeless fittings were used to transfer solution (0.5 mm i.d tubing was used for general transfer; 0.8 and 2.0 mm i.d. were used for waste lines; and 0.3 mm i.d. tubing was used to transfer solution from the analytic column to the spray chamber). Tygon tubing was used for peristaltic pumps for its durability (1.02 mm i.d. two-stop tubing was used for the sample and wash lines; 0.25 mm i.d. tubing was used for the MQ water wash and 10.1021/es011222f CCC: $22.00
2002 American Chemical Society Published on Web 06/01/2002
internal standard lines (two-stop and three-stop, respectively); and 0.89 mm i.d. three-stop tubing was used for the eluant line). Slip fittings were used to connect the 0.25 mm i.d. peristaltic pump tubing to the small-bore Teflon transfer tubing; PEEK connectors (Upchurch Scientific, Oak Harbor, WA) were used to connect the larger diameter peristaltic pump tubing. During continuous field sampling with the sample manifold, seawater was filtered through a highcapacity 0.45-µm Versapor filter cartridge (Geotech Environmental Technologies).
Reagents Ammonium Acetate Buffer. A 2 N ammonia/ammonium acetate buffer was prepared by mixing 115 cm3 of concentrated ultrapure acetic acid (Seastar Chemicals, Sidney, British Columbia, Canada) with 1.7 mol of ultrapure ammonium hydroxide (Seastar Chemicals), and diluting to 1 L with 18 MΩ/cm water (Millipore MQ, Bedford, MA). The buffer solution pH was adjusted to pH 5.5 by the addition of acetic acid or ammonium hydroxide. Eluant and Wash. Trace elements were eluted from the column with 1 N nitric acid prepared from ultrapure concentrated nitric acid (Seastar Chemicals) and 18 MΩ/cm water. The column was conditioned prior to sample loading, and washed after sample loading with 18 MΩ/cm water. Ultrapure water has an approximate pH of 5.6 controlled by saturation with atmospheric carbon dioxide. Nitric acid (0.1% and 2%) rinse solutions for the autosampler probe were prepared by adding ultrapure nitric acid to ultrapure water. Calibration and Reference Solutions. The accuracy and analytical figures of merit for the technique were assessed using certified reference seawater solutions distributed by the National Resource Council of Canada. NASS-4 is an open ocean sample; CASS-3 is a coastal ocean sample. These solutions are acidified to pH 1.6 to stabilize trace elements in solution. Two sets of calibration standards were prepared. A set of aqueous multielement standards was prepared from NIST-traceable single-element stock solutions (Spex, Metuchen, NJ, and Inorganic Ventures, Lakewood, NJ). These solutions contained V, Mn, Ni, Cu, Zn, Mo, Cd, and Pb in the approximate proportions expected in seawater, were diluted with 18 MΩ/cm water, and were adjusted to a pH of less than 2 with ultrapure nitric acid. A second set of seawater standards was prepared by adding small aliquots of the multielement standard solutions to the NRC standards. All calibration solutions were made up in acid-leached highdensity polyethylene bottles. Instrument performance was monitored with an internal standard solution of Sc, Y, Ce, Tb, and Bi. Ion-Exchange Resin and Column. Toyopearl AF-Chelate 650M resin with a particle diameter of 65 µm/dm3 was pumped persistaltically into a column constructed of 1/8 in. o.d. and 1/16 in. i.d. Tefzel tubing and 1/8 in. natural PEEK fittings. The column fittings (Upchurch Scientific) consisted of a 1/4-28 super flangeless nut (modified to accommodate the short tubing length) and a 2-µm frit-in-a-ferrule on each end. The resultant 3-cm column contains approximately 50 µdm3 of resin, does not leak, will sustain flow rates in excess of 2 cm3/min at peristaltic pump pressures, and can be easily incorporated into or removed from the FIA system with standard 1/8 in. chromatography fittings. In addition to the analytic column, cleanup columns constructed in the same manner, were placed on the buffer line and carrier line to reduce reagent blanks. These columns were regenerated periodically with a 2% nitric acid solution.
Procedure Flow-Injection Protocol. The flow-injection protocol is described in detail elsewhere (16). The ICPMS peristaltic
pump, which pumps eluant solution through the analytical column or directly into the nebulizer and spray chamber, ran at a constant speed of 0.2 rps throughout the entire cycle. The eluant solution was mixed on-line with an internal standard solution (producing a diluted internal standard element concentration of 7 ng/cm3) before being introduced to the nebulizer and spray chamber. The sample solution was buffered on-line with the 2 N ammonia/ammonium acetate to a pH of 5.6. After loading the sample onto the resin, the resin was washed with ultrapure water to remove residual salts and eluted into the ICPMS with 1 N nitric acid. At 0.4 rps, the flow rate through the analytical column was 2 cm3/min; the 90 s loading cycle processed 3 cm3 of sample. The entire procedure took 4 min. Data Collection. Data were collected in time-resolved analysis mode. The central channel of each isotope was integrated for 0.1 s (or 0.2 s for Mn and Cd) in each 1.32 s time step. Analyte signals were integrated over 30 s time periods (14-44 s) after baseline subtraction. Vanadium had a longer elution curve and was integrated between 14 and 54 s. The remaining time in the 75 s elution step was used to thoroughly clean the column. Analyte peaks were chosen to minimize the potential for polyatomic interferences: 60Ni, 65Cu, 66Zn, 111Cd, and 208Pb. Manganese (55Mn) and vanadium (51V) are monoisotopic. Analyte signals were also normalized against internal standards: V, Mn, Ni, Cu, and Zn against 45Sc; Cd against 89Y; and Pb against 208Bi. Concentrations were determined by external calibration using a simple linear regression of 5-8 spiked seawater standards. The unspiked reference standards were not used to determine the regression. Aqueous standards were run to compare uptake and response between dilute aqueous solutions and saline seawater solutions. Blanks were determined by running ultrapure water or ultrapure water acidified to a pH of less than 2 with ultrapure nitric acid. San Diego Bay and Bight Survey. Trace element surveys of the San Diego Bay and the adjacent coastal ocean in the San Diego Bight were conducted on September 13-17, 1999, aboard the R/V Robert G. Sproul. The survey was divided into four sequential segments, corresponding to B1-B4 in Figure 1. The locations of bottle sampling for postcruise laboratory validation are also indicated (open squares). Water was collected using a submersible titanium/Teflon pump on a towed body off the starboard side of the ship, 1.4-2.3 m below the surface. Postcruise Laboratory Validation. Discrete bottle samples were collected during the survey into acid-leached highdensity polyethylene bottles and acidified to less than pH 2 with high-purity (Seastar) nitric acid. These samples were analyzed after the cruise using the same procedure as that used during the survey. Quantitation was accomplished by both external calibration with spiked NRC seawater standards (as was done at sea) and by standard additions.
Results Figures of Merit. Method blanks were determined between survey legs to assess the ability of the procedure to control memory effects. Method blank levels and limits of detection (LOD) are shown in Table 1. The reported blank concentrations are the mean of 14 blank determinations. Limits of detection for each element were calculated as equal to 3 times the standard deviation of the 14 method blanks. Trace element limits of detection are in the low pg/cm3 range, which are adequate for analysis of unpolluted open ocean water. The high background at a mass/charge ratio of 55 in dilute nitric acid solutions (from the ArNH+ polyatomic ion) resulted in relatively higher detection limits for 55Mn and made determining the true Mn blank difficult. Measured and certified concentrations for analyte elements in the seawater reference standards are shown in Table 2. Both open ocean VOL. 36, NO. 13, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2827
FIGURE 1. R/V Sproul ship track showing surveys conducted in and around the San Diego Bay during September 14-15, 1999. Legs B1 (dotted line), B2 (black line), and B3 (gray line) represent separate survey transects. Open squares show bottle sample locations.
TABLE 1. Method Limits of Detection, Blanks, and Yieldsa element
LOD (pg/cm3)
blank (pg/cm3)
relative response
V Mn Ni Cu Zn Cd Pb
5.3 29 2.2 2.8 26 1.0 0.3
2.5
