Samples considered by the shipboard scientific party to be representative of the various lithologies cored were analyzed for major oxide and selected trace element compositions with the shipboard ARL 8420 wavelength-dispersive XRF apparatus. Full details of the shipboard analytical facilities and methods are presented in previous ODP Initial Reports volumes (e.g., Legs 118, 140, 147, 153; Shipboard Scientific Party, 1989, 1992, 1993, 1995a). The elements analyzed and the operating conditions for Leg 176 XRF analyses are presented in Table T1.
After coarse crushing, samples were ground in a tungsten carbide shatterbox. Then 600-mg aliquots of ignited rock powder were intimately mixed with a fusion flux consisting of 80 wt% lithiumtetraborate and 20 wt% "heavy absorber" La2O3. The glass disks for the analysis of the major oxides were prepared by melting the mixture in a platinum mold in an electric induction furnace. Trace elements were determined on pressed powder pellets prepared from 5 g of rock powder (dried at 110ºC) mixed with a small amount of a polyvinylalcohol binder solution. The calibration of the XRF system was based on the measurement of a set of reference rock powders. A Compton scattering technique was used for matrix absorption correction for trace element analysis.
The total sum of the oxides determined on ignited powders was generally high. About 60% of the samples had "totals" between 101% and 103%. The systematic analytical error presumably reflects small inadequacies in the procedure for correction for matrix absorption or errors in the concentration of the "heavy absorber" in the fusion flux. A small set of samples was reanalyzed after the cruise by atomic emission and atomic absorption spectrometry at Leuven University, Belgium. It appeared that the error affects all the major element oxides, but it significantly affects only the result of the main component SiO2, and to lesser extent of Al2O3, CaO, MgO, and Fe2O3.
A second analytical problem concerned the trace element zirconium. Zirconium data for olivine gabbros obtained during Leg 176 were systematically higher than in equivalent rocks analyzed during Leg 118. To determine whether this was a real geochemical difference, a selection of Leg 118 shipboard powder samples was reanalyzed during Leg 176. Consistently higher Zr values were obtained. A postcruise analysis of control samples at the Universities of Copenhagen, Denmark, and Leuven, Belgium, showed that the lower values obtained during Leg 118 are the more accurate ones. Inadequacies in the correction for the interference of Sr X-rays on Zr X-ray peaks are the most likely source of analytical errors. Unfortunately, the problem could not be remedied during the cruise. As a consequence, the reported Zr data may be too high by a factor of two in the 15- to 30-ppm concentration range. At higher Zr concentration levels, the effect of the analytical error is minor.
Loss on ignition (LOI) for each sample was determined by the standard practice of heating an oven-dried (110ºC) sample to 1010ºC for several hours. To more fully investigate the results of LOI determinations of selected samples, gas chromatography of the expelled volatiles was performed on a Carlo Erba NA 1500 CHS analyzer; 20 mg of dried (110ºC) rock sample was combusted at 1010ºC, releasing water and carbon dioxide. Sulfur from sulfide minerals is oxidized to SO2 (by addition of V2O5), and SO3 is released from any sulfate minerals present in the sample. Gas-chromatographic separation was followed by quantitative determination of the respective gaseous oxides of carbon, hydrogen, and sulfur by a thermal conductivity detector. Reagent grade sulfanilamide (C6H8N2O2S) was used to calculate the bias factor of the analyzer for CO2, H2O, and SO3. Control rock samples covering the concentration ranges of hydrogen, carbon, and sulfur observed in the cored samples were repeatedly analyzed. These control samples were prepared by intimately mixing weighed quantities of a low-sulfur, low-carbon silicate rock (Gaby 153 Interlaboratory Standard prepared during Leg 153) and of sediment sample ODP6 (containing 3.2 wt% C, 0.56 wt% H, and 5300 ppm S). Because the concentration of total sulfur in most of the rock samples was close to the determination limit of the CHS analyzer, results of duplicate analyses for sulfur showed large scatter. Therefore, no results of sulfur analysis are listed in this volume.
After the cruise all the shipboard samples were analyzed for ferrous iron content at the University of Leuven, Belgium, using the redox-titration method of Shafer (1966). Sample aliquots weighing from 50 to 100 mg were dissolved in a H2SO4/HF mixture in a nitrogen atmosphere. Ferrous iron is titrated with potassium-dichromate solution using sodiumdiphenylaminesulfonate as a redox color indicator.