A new 3-step sequential separation chemistry for Sr, Nd and Pb from silicate samples, which is suitable for isotope analysis by MC-ICP-MS as well as TIMS, has been developed. The chemistry is designed to minimize the number of evaporation steps, enabling high throughput especially when MC-ICP-MS is employed. The sample solution in 0.5 mol l^-1 HNO₃ after digestion with HF-HClO₄, which is prepared for trace element analysis, can be directly used in the new chemistry. In the first column using Sr resin, Sr and Pb are collected. The recovered solution for Sr can be directly aspirated into MC-ICP-MS. The Pb solution is dried, re-dissolved, separated into two aliquots and measured by double spike MC-ICP-MS. Subsequently, LREEs are collected by the second column packed with cation exchange resin. Finally, Nd was purified from Sm using Ln resin in the third column. The recovered Nd solution can also be directly nebulized into MC-ICP-MS. The Sr and Nd isotope ratios for the standard material JB-3 (basalt) from the Geological Survey of Japan by MC-ICP-MS and TIMS were reported to verify the chemistry and mass spectrometry developed in this study.

View full abstract

Rock alteration processes can be studied through geochemical Mass Balance (MB) estimation in several ways. Due to the ease of its adaptability, especially its graphical approach, the ISOCON method of mass balance is extensively used for such studies. However the technique suffers from a serious limitation, in particular with reference to the graphical treatment involving arbitrary scaling, which is explained in this paper with the help of illustrative examples. The reference frame for MB obtained by best fit regression is biased to the element scaled upward that plot away from the origin. The scaling factors actually act as weighting factors which bias the slope of the ISOCON line because it is controlled by data points having higher numerical value in the plot. Since scaling of concentration data are unavoidable in ISOCON analysis, the results of MB may potentially be misleading. The slope of the best fit ISOCON line may vary within the range of slopes defined by individual conserved elements. In case, no scaling is done, the ISOCON shall weight in proportion to the abundance of the immobile elements. The co-linearity of the immobile elements with origin is also subjective in nature with the elements near origin apparently seem to satisfy a wide range ISOCON slopes than those that are away from origin. When equal weight is assigned to all elements, regardless of their concentration level, an average of the slopes of the immobile elements (i.e., concentration ratios in altered to unaltered) may provide better approximation for reference frame without recourse to graphical plot and ISOCON solution. The existing weighted least-squares or equal weight least-squares ISOCON method is found to be more appropriate in case of higher uncertainty in the protolith compositions. In any case, a proper identification of the conserved species is an essential prerequisite to successful implementation of mass balance computation minimizing the inherent errors of the ISOCON technique.

View full abstract

By using a step-wise heating method in conjunction with static vacuum mass spectrometry, we have measured at high temporal resolution (one month) the nitrogen isotopic ratio of a coral skeletal sample collected from Palau in the northern part of the Western Pacific. The δ¹⁵N value from the 900°C combustion step varies significantly, from +1.1‰ to +10.9‰ for the estimated time period from July 1995 to January 1998. The higher values occurred during winter and the lower values in summer. The δ¹⁸O and δ¹³C values of the skeleton were also analyzed for the same period, using conventional methods. The ranges of δ¹⁸O and δ¹³C variations are from -6.1‰ to -5.4‰ and from -1.6‰ to -0.9‰, respectively, which are significantly smaller than the δ¹⁵N variation. The δ¹³C change is possibly attributable to kinetic isotope effects, as deduced from a negative correlation between the δ¹³C value and mean sea surface temperature (SST) in the region. On the other hand, the δ¹⁵N variation could be derived from seasonal change of the relative ratio of two components utilized by symbiotic algae: a ¹⁵N-depleted component from N₂ fixation and ¹⁵N-enriched nitrate from the tropical oligotrophic open ocean.

View full abstract

Three standard minerals GA-1550 biotite, LP-6 biotite and Bern-4M muscovite used for K-Ar and ⁴⁰Ar-³⁹Ar dating have been investigated by the single grain Rb-Sr technique. Compared to the K-Ar and ⁴⁰Ar-³⁹Ar results previously reported, these mica standards nevertheless show complex characteristics in the Rb-Sr isotopic system. Analysis on eight individual grains of GA-1550 biotite gives a Rb-Sr isochron age of 98.7 ± 1.9 Ma, consistent with previous dating results of the K-Ar and stepwise heating and laser ablation ⁴⁰Ar-³⁹Ar methods. Although evidence of the laser ablation ⁴⁰Ar-³⁹Ar analyses in many previous studies demonstrates that LP-6 biotite is considerably inhomogeneous in the K-Ar isotopic system, Rb-Sr isotopic analyses on ten individual biotite grains show achievement of Sr isotopic equilibrium during crystallization. A Rb-Sr isochron age of 128.2 ± 2.2 Ma calculated from the analytical results is concordant with those previously reported K-Ar and ⁴⁰Ar-³⁹Ar ages obtained from c. 30 mg aliquots of biotite grains. Excellent age reproducibility on Bern-4M muscovite was previously achieved by K-Ar and ⁴⁰Ar-³⁹Ar dating by total fusion in c. 30 mg aliquots of muscovite grains, but single grain Rb-Sr analytical results in this study demonstrate considerable disequilibrium of Sr isotopic system during crystallization of muscovite. The discordance in the Rb-Sr and K-Ar isotopic systems of two standard minerals indicates complex factors that can influence isotopic disequilibrium of both isotopic systems.

View full abstract

Material sources of the northern deserts and the Loess Plateau in China are always one of the critical focuses in Quaternary and paleoclimate fields. In this paper, the method of Sr-Nd isotope geochemistry is applied to explore the relationship in material source between the Erdos desert and Chinese Loess Plateau. The ε_Nd(0) value of the <75 μm silicate fraction between -11.8 and -17.2 is more positive in the west than in the east, and its ⁸⁷Sr/⁸⁶Sr ratio varies from 0.719218 to 0.714824 without similar characteristics in the Erdos desert. In addition, there are Sr-Nd isotopic differences between different grain-size fractions of eolian sand in the Erdos desert. The coarse-grained fractions mainly originated from local parent rock weathering, but the fine-grained fractions were probably affected by the input of foreign materials. Geographical distribution of Sr-Nd isotopes (especially Nd isotopes) of <75 μm silicate fractions indicates that the input of foreign materials into the Erdos desert decreases gradually from its west to east. The ε_Nd(0) and ⁸⁷Sr/⁸⁶Sr values of eolian silicate fractions in the Chinese Loess Plateau vary from -9.2 to -13.56 and from 0.719957 to 0.714424, respectively in this study. The ε_Nd(0) and ⁸⁷Sr/⁸⁶Sr values of eolian silicate fractions are different between the western-central and eastern Loess Plateau. According to Sr-Nd isotopic data from previous and present studies, the ε_Nd(0) values of eolian silicate fractions in the western-central Loess Plateau mostly fall in a very narrow range of -9∼-11.5, suggesting a uniform source region. However, there is a large Sr isotope variation resulting from different pretreating methods employed by researchers. The ε_Nd(0) and ⁸⁷Sr/⁸⁶Sr values of eolian silicate fractions in the eastern Loess Plateau both change markedly. The differences in Sr-Nd isotopes of eolian silicate fractions between the eastern and western-central Loess Plateau indicate that there are different sources for eolian silicate materials in the two sub-areas. Sr-Nd isotopes, especially Nd isotopes of the <75 μm silicate fractions in the Erdos desert are clearly different from those of eolian silicate fractions in the western-central part, but roughly close to those of eolian silicate fractions in the eastern part of the Chinese Loess Plateau, which shows that the Erdos desert was perhaps a main source for the eastern Loess Plateau but not for the western-central Loess Plateau.

View full abstract

The potential for Se removal from mine drainage water using permeable reactive materials was evaluated by a laboratory column experiment. The column materials, organic carbon and zero valence iron (ZVI), were exposed to mine drainage containing 630 mg L^-1 SO₄^2-. The influent water was spiked with 40 mg L^-1 Se(VI) to assess the potential for Se removal. This high Se(VI) concentration was selected to ensure that there would be a sufficient mass of Se-bearing reaction products available for mineralogical characterization. The experiment was conducted in an anaerobic chamber to replicate the anaerobic conditions that prevail in permeable reactive barrier systems. After loading 10.8 pore volumes of input solution, the column effluent contained <0.002 mg L^-1 Se and <300 mg L^-1 SO₄^2-. After the column experiments was complete the reactive materials were sampled in the anaerobic chamber and examined using scanning electron microscopy (SEM) coupled with energy dispersion X-ray analysis (SEM-EDAX), X-ray photoelectron spectroscopy (XPS) and by Raman spectroscopy. Sulfate was reduced to sulfide and elemental sulfur, which accumulated on the surfaces of the column materials. Se-bearing precipitates were observed at the base of column. Scanning electron microscopy (SEM) showed the presence of elemental Se, suggesting that Se(VI) was partly reduced to metallic Se(0). The XPS results revealed that selenate was reduced mainly to iron selenide (FeSe and/or FeSe₂) on the surface of the column substances. These observations suggest that both chemical reduction and biologically mediated reduction of Se(VI) occurred.

View full abstract

Homogeneous partition coefficients (D) for Ra and Ba in calcite were determined from coprecipitation experiments using the Continuous Addition Method (CAM). The D-values derived for Ra and Ba are: D_Ra = (1.5 ±0.6) ×10^-1 and D_Ba = (1.6 ±1.1) ×10^-2. Although these data were derived at low precipitation rates (R) for calcite (R < 9 nmol mg^-1min^-1), the solutions were significantly oversaturated in calcite. Values of D should, ideally, be derived at equilibrium and hence the effect of oversaturation should be examined. Therefore, a coprecipitation experiment using the Free Drift Method (FDM) with slow degassing was also carried out. This experiment occurs under conditions that are only slightly oversaturated and the rate of coprecipitation was confirmed to be slow compared to dissolution and precipitation at the calcite surface. The resultant heterogeneous partition coefficients showed good agreement with homogeneous partition coefficients derived by CAM. This indicates that the oversaturation in CAM did not affect partitioning of these metals in calcite and that the derived D-values represented equilibrium partitioning between the solution and the solid phase. The derived value of D_Ra is one order of magnitude larger than that of D_Ba despite the larger ionic radius of Ra. This indicates that compatibility of ionic radius is not a dominant parameter for preferential incorporation of Ra and Ba in calcite.

View full abstract

We determined the concentrations of ⁹Be in JB-2 and JA-2 geological standard samples using an inductively coupled plasma mass spectrometer (ICP-MS). We determined ⁹Be abundance in standard samples of JA-2 and JB-2 as 2.26 ± 0.10 and 0.241 ± 0.010 ppm, respectively, after comparing and selecting ⁹Be concentration data obtained by the internal standard method, the standard addition method and the column separation method. In the course of comparison, we identified a problem in the internal standard method, which is usually used for abundance analyses using an ICP-MS. Repeated analyses of a basalt lava gave a standard deviation of ¹⁰Be abundance of 9% (1σ, n = 5). From the results, we estimated the uncertainty of ¹⁰Be/⁹Be ratio as 10% (1σ), considering a law of propagation of errors.

View full abstract