METHODS

Methods for recovery of IW and details of the sample handling are described in Eberli, Swart, Malone, et al. (1997). All IW samples were double filtered. Samples were initially collected from the squeezer through 0.45-µm Gelman polysulfone disposable filters and into scrupulously cleaned 50-mL plastic syringes. The IW was subsequently refiltered through either 0.45-µm Gelman polysulfone disposable filters or acid-washed 0.2-µm Gelman polysulfone disposable filters. The latter was performed when collecting IW samples for shore-based trace element analyses. Trace element samples were stored in acid-washed plastic bottles, acidified with 50 µL of ultra-high purity HNO₃, sealed with parafilm and chilled until analysis.

Shore-based analyses of IW constituents were performed by atomic spectrometry and flow injection analysis (FIA). Dissolved Li⁺ and Rb⁺ were determined by atomic emission spectrometry (AES) using the method of standard additions on a Perkin-Elmer Model 603 double beam spectrometer. Lithium determinations were carried out at 670.8 nm using a slit width of 0.2 nm. The atomic emission of Rb was measured at 780 nm also using a 0.2 nm slit width. Because IW can show wide variations in matrix composition (e.g., salt concentrations), a background correction technique developed in our laboratory for analysis of geothermal and hydrothermal fluids was used (C. Fraley and E. De Carlo, unpubl. data). This method compensates for the large background absorption signals such as those encountered in high salinity recovered deep within the Bahamas Transect sites (Eberli, Swart, Malone, et al., 1997). Our offline background correction technique involves measurement of the emission intensity of the solution at the analytical wavelength, then measuring the emission signal 0.8 nm above and below the analytical line of the element of interest. Because atomic emission lines are very narrow (e.g., Robinson, 1990), the online measurement yields the emission intensity derived from the analyte and the matrix background, whereas the offline signal originates only from the matrix. Matrix absorption is typically of a broad-band nature, thus subtraction of the linearly interpolated background signal obtained during the two offline measurements from the online emission measurement yields a reliable analyte signal.

Dissolved Ba²⁺ and Sr²⁺ were determined by inductively coupled plasma-optical emission spectroscopy (ICP/OES) using a high-resolution Leeman Labs model PS1 echelle grating spectrometer (e.g., De Carlo, 1992). Instruments were calibrated using a series of dilutions of IAPSO seawater (Sr²⁺), or by preparation of single-element standards (Ba²⁺) in a NaCl matrix.

Dissolved Fe²⁺ and Mn²⁺ were determined by FIA with fluorescence detection (Resing and Mottl, 1992) and/or by ICP/OES. Standards for Fe and Mn analysis were prepared by spiking acidified surface seawater with appropriate amounts of certified single-element spectroscopic standards. IW samples were diluted with oligotrophic surface seawater as necessary to ensure that instrumental signals fell within the linear portion of the calibrations.

Data quality assurance (QA) was monitored by concurrent measurement of elemental concentrations in quality control (QC) U.S. Geological Survey (M-series and T-series) Standard Water Reference Samples (SWRS), certified reference materials from NRC Canada, and/or by analysis of IAPSO standard seawater. Agreement was always within the limits of the statistical range of the certified or recommended reference values and within the expected instrumental precision for the method utilized (3%-5% relative).