Previous   |    Next

doi:10.2204/iodp.proc.345.102.2014

Inorganic geochemistry

During Expedition 345, chemical analyses of rock samples were performed using inductively coupled plasma–atomic emission spectroscopy (ICP-AES) for major and trace element analyses and gas chromatography for sulfur, CO₂, and H₂O. These samples were selected as representative of the rock recovered from Site U1415 by the Shipboard Scientific Party. A thin section was located next to each of the geochemistry samples to precisely determine their modal composition and degree of alteration because of the highly heterogeneous lithologies sampled at Site U1415 (see “Igneous petrology” and “Metamorphic petrology” in each hole chapter for the characterization of the lithologic units). The rock names in the resulting table (see Table T1 in the “Geochemistry summary” chapter [Gillis et al., 2014a]) were determined on the basis of the thin section descriptions (see “Igneous petrology”). A characteristic of Site U1415 gabbroic rock is the abundance of trace orthopyroxene observed in some holes. For simplicity, only samples with >2% orthopyroxene were described as orthopyroxene bearing in figures and in Table T1 in the “Geochemistry summary” chapter (Gillis et al., 2014a).

Sample preparation for geochemistry

Samples were prepared from 15 to 30 cm³ of rock for troctolites, gabbroic rocks, and basalts and 30 cm³ scooped material for drilling-induced disaggregated gabbro samples. The solid rock samples were cut from cores using a diamond saw blade. Whenever possible, a thin section billet was taken from the same rock fragment. Grain mounts were produced from the batches of drilling-induced disaggregated gabbro samples selected for geochemical analyses. Outer surfaces of rock samples were ground on a diamond-impregnated disk to remove saw marks and altered rinds resulting from drilling. Each cleaned sample was placed in a beaker containing isopropanol and ultrasonicated for 15 min. The isopropanol was decanted, and the samples were then ultrasonicated twice in nanopure deionized water (18 MΩ·cm) for 10 min. The cleaned pieces were then dried for 10–12 h at 110°C. During Expedition 345, this time was shortened to 6 h for the preparation of the igneous rock sampled in Hole U1415P.

The clean, dry whole-rock samples were crushed to <1 cm chips between two disks of Delrin plastic in a hydraulic press. The rock chips were then ground to a fine powder in a tungsten carbide mill in a SPEX 8510 shatterbox. A check on grinding contamination contributed by the tungsten carbide mills was performed during Leg 206 (Shipboard Scientific Party, 2003), and contamination was found to be negligible for major elements and most trace elements measured on board (Sc, V, Cr, Ni, Sr, Y, Zr, Nb, and Ba). A systematic analysis of the shipboard powders from Expeditions 304/305 indicated a possible contamination in Co during grinding (Godard et al., 2009); this element, although analyzed (Table T1), was eliminated from Expedition 345 data.

After grinding, a 5.00 ± 0.05 g aliquot of the sample powder was weighed on a Mettler Toledo balance and ignited at 1025°C for 4 h to determine weight loss on ignition (LOI) with an estimated precision of 0.02 g (0.4%).

The standard shipboard procedure for digestion of rock and subsequent ICP-AES analysis is described in Murray et al. (2000). The following protocol is an abbreviated form of this procedure with minor changes and additions. After determination of LOI, 100.0 ± 0.1 mg aliquots of the ignited whole-rock powder were weighed and mixed with 400.0 ± 0.5 mg of lithium metaborate (LiBO₂) flux, which was preweighed on shore. Standard rock powder and full procedural blanks were included with unknowns in each ICP-AES run. All samples and standards were weighed on the Cahn C-29 microbalance (designed to measure on a moving platform), with weighing errors conservatively estimated to be ±0.02 mg.

A 10 µL aliquot of 0.172 mM aqueous LiBr solution was added to the flux and rock powder mixture as a nonwetting agent to prevent the cooled bead from sticking to the crucible. Samples were then individually fused in Pt-Au (95:5) crucibles for ~12 min at a maximum temperature of 1050°C in a Bead Sampler TK-4100 (internally rotating induction furnace). After cooling, beads were transferred to 125 mL high-density polypropylene bottles and dissolved in 50 mL 10% dilution of concentrated trace-metal grade HNO₃ (hereafter referred to as 10% HNO₃), aided by shaking with a Burrell wrist-action bottle shaker for 1 h. After digestion of the glass bead, the solution was passed through a 0.45 µm filter into a clean 60 mL wide-mouth, high-density polypropylene bottle. Next, 1.25 mL of this solution was pipetted into a plastic vial and diluted with 8.75 mL of 10% HNO₃ to bring the total volume to 10 mL. The final solution-to-sample dilution factor for this procedure was ~4000. Dilutions were conducted using a Brinkman Dispensette.

Inductively coupled plasma–atomic emission spectroscopy

Analyses

Major and trace element concentrations of standards and samples were determined using a Teledyne Leeman Labs Prodigy ICP-AES instrument. The analyzed elements and the wavelengths used for sample analysis during Expedition 345 are provided in Table T1. Certified international rock reference materials, calibration and drift solutions, and chemical procedure blanks were included with the unknown samples for each sample run. Detection limits were calculated as three times the standard deviation of the mean for blank solution measurements.

The ICP-AES plasma was ignited at least 30 min before each sample run to allow the instrument to warm up and stabilize. After the warm-up period, a zero-order search was performed to check the mechanical zero of the diffraction grating. After the zero-order search, the mechanical step positions of emission lines were tuned by automatically searching with a 0.002 nm window across each emission peak using two solutions containing 10 ppm concentrations of set elements prepared in 1% HNO₃. During the initial setup, an emission profile was selected for each peak, using the multielement solutions, to determine peak-to-background intensities and set the locations of background levels for each element. The Prodigy data acquisition software uses these background locations to calculate the net intensity for each emission line.

The ICP-AES data presented in the “Inorganic geochemistry” sections for each hole were acquired using the Prodigy software. The intensity curve for each element is defined by 20 measurements within the designated wavelength window. The user selects the two background points that define the baseline. The Prodigy software integrates the area delineated by the baseline and the intensity curve. Each sample was analyzed four times from the same dilute solution within a given sample run. For several elements, measurements were made at two wavelengths (e.g., Si at 250.690 and 251.611 nm; Table T1). For each run, the wavelength yielding the best calibration line was identified and used for determining concentrations.

A typical ICP-AES run included

• A set of four certified rock standards, analyzed twice per run and chosen for their wide range of compositions in order to calibrate the analyses (basalt BCR-2; granodiorite JG-1A; peridotite JP-1; gabbro MRG-1);

• As many as 15 unknown samples;

• A drift-correcting sample (BHVO-2) analyzed every fourth sample position and at the beginning and end of each run;

• Procedural blank solutions run near the beginning of each run; and

• Two to three “check” standards chosen for their composition similar to that of the analyzed material (dunite DTS-1, gabbro JGb-1, and an ultramafic reference material from Core 147-895D-10W) run as unknowns.

A 10% HNO₃ wash solution was run for 90 s between each sample analysis. Each sample analysis is the average of four measurements. The check standards were used to test analytical accuracy and reproducibility of the obtained data.

Data reduction for inductively coupled plasma–atomic emission spectroscopy

Following each sample run, raw intensities were transferred to a data file, and all analyses were corrected for drift. A drift correction was applied to each element by linear interpolation between drift-monitoring solutions run every fourth analysis. After drift correction and subtraction of procedural blank, a calibration line for each element was calculated using the results for the certified rock standards. Concentrations used for the calibrations were compiled values from the literature recalculated on a volatile-free basis; the compiled values were from Govindaraju (1994) and from the GeoReM website (georem.mpch-mainz.gwdg.de; December 2012; Jochum et al., 2005) (Table T2). Element concentrations in the samples were then calculated from the relevant calibration lines.

Estimates of accuracy and precision of major and trace element analyses were based on replicate analyses of check standards, compared to values published in Govindaraju (1994) and downloaded from the GeoReM database for international rock standard DTS-1 and JGb-1 and to values published in Puchelt et al. (1996) for shipboard laboratory Standard 147-895D-10W. Results are presented in Table T3. During Expedition 345, run-to-run relative standard deviation by ICP-AES was generally ±1% for major elements except for MgO (±5%) and ±10% for trace elements. Accuracy was better than 2% for major elements and better than 5% for most trace elements, with the exception of low-concentration data for Cr, Zr, and Y.

Volatile measurements

Volatile concentrations were measured by gas chromatographic separation on a Thermo Electron Corporation CHNS analyzer (Flash EA 1112 Series). A new calibration strategy for the analysis of H₂O, CO₂, and S in mafic and ultramafic rocks by CHNS elemental analyzer was developed during Expedition 345. This calibration method involves measuring a series of international rock standards that approximate the composition of the unknown samples. All studied samples were powder splits of ICP-AES samples.

Analytical method

Powders were dried for 12 h at 110°C to ensure evaporation of possible adsorbed moisture and kept in a desiccator prior to measurements. Powder samples were typically 40 mg. Samples were weighed on a Cahn Microbalance Model 29, mixed with 10 mg vanadium pentoxide (V₂O₅), an oxidizer used to improve sulfur detection, and packed into CHNS tin containers (Universal Tin Container “light”; Thermo Electron P/N 240-06400). A revolving autosampler dropped sample capsules into a 950°C resistance furnace where they were combusted in a reactor. Tin from the capsule creates a violent flash combustion within an oxygen-enriched atmosphere. The oxidized and liberated volatiles were carried by a constant helium gas flow through a commercial glass column (Costech P/N 061110) packed with an oxidation catalyst of tungsten trioxide (WO₃) and a copper reducer. Sulfide and sulfate minerals liberate SO₂ and SO₃, respectively; however, both were measured as S. When nitrogen oxide is present, the copper reducer in the reactor tube reduces it to N₂. The liberated gases were transported by the helium carrier flow to and separated by a 2 m packed gas chromatography (GC) column (Costech P/N 0581080). During the measuring time of 1000 s, the millivoltage at the detector was continuously recorded. CO₂, H₂O, and S separated by the GC column arrived at the thermal conductivity detector at approximately 94.0 s (CO₂), 250.0 s (H₂O), and 781.0 s (S) for the international gabbro standard JGb-1. During Expedition 345, we were not able to measure nitrogen because of the particularly low concentrations of this element in the analyzed mafic and ultramafic rock.

Calibration, blanks, and standards

The routine method for quantitative chromatographic geochemical analyses involves the preparation of a series of standard solutions (e.g., 80 mM L-(+)-cysteine hydrochloride). Because of the low volatile concentrations and different mineralogy in mafic to ultramafic plutonic rock, we chose an alternative method using a matrix-match calibration strategy based on international geostandards for CO₂ and H₂O. Because of the lack of silicate standards for the S calibration, S was calibrated using the geostandards and a series of increasing weights (0.1, 0.3, 0.6, and 1 mg) of sulfanilamide (18.62% S). Peak areas of the measured volatiles from the geostandard chromatographs were integrated and plotted weight-corrected as a function of their reference concentration published in Govindaraju (1994) and downloaded from the GeoReM database for the certified dunite DTS-1, syenite SY-2, gabbro JGb-1, and Cody Shale SCo-1 and to values from Puchelt et al. (1996) for the shipboard laboratory ultramafic peridotite Core 147-895D-10W. Procedural blanks were determined by using empty tin capsules (plus V₂O₅) and measured four times during each CHNS run of 8 samples (6 new samples plus 2 duplicates from the prior sample run). A measurement blank was performed after every sample to prevent cross-contamination within the chromatographic column. After weight correction, H₂O, CO₂, and S abundances were calculated by using the function resulting from the linear or polygonal functions of the calibration lines.

Peridotite JP-1 and gabbro MRG-1 were used as quality control by monitoring the analytical accuracy and reproducibility and as sensitivity drift check by replicate measurements. A typical CHNS run included a maximum of 8 unknown samples per run; this approach allowed for the frequent restandardization required for high accuracy. Results of the GC analyses for MRG-1 and JP-1 during Expedition 345 are presented in Table T4. Based on 10 runs, the reproducibility was better than 5%–14% for CO₂, better than 12%–13% for H₂O, and better than 13% for S for JP-1 and MRG-1. The average obtained concentrations for peridotite JP-1 and gabbro MRG-1 are in agreement with recommended values (Govindaraju, 1994) and downloaded from the GeoReM database.

In comparing the GC volatile analyses to the LOI results, it is also important to bear in mind that during ignition of the sample in an atmospheric oxygen fugacity, Fe²⁺ will change into Fe³⁺, which accounts for a weight gain of as much as 11.1% of the proportion of ferrous Fe within the sample. Also, ignition at or above 1000°C may induce a loss of K and Na, as these elements have boiling-point temperatures below 1000°C (759°C for K and 883°C for Na) (Lide, 2000). This can lead to discrepancies between LOI analyses and volatile concentrations determined by GC analysis.

Top of page   |    Previous   |    Next