Previous   |   Next

doi:10.2204/iodp.proc.324.104.2010

Geochemistry

Major and trace element analysis

Forty samples of lavas from Hole U1347A were analyzed by ICP-AES for concentrations of major and several trace elements (see "Geochemistry" in the "Methods" chapter for information on analytical procedures, instrumentation, and data quality). Samples of all the igneous units recovered (stratigraphic Units IV, V, VII, IX, X, XII, and XIV–XVI) were included. All were taken from holocrystalline portions of the cores, except for one specimen from a rind of highly altered glass rimming holocrystalline rock at the top of Unit V (see below).

As with Site U1346, total weight percentages of the major element oxides in the Site U1347 analyses are quite variable, between 96.53 and 101.52 wt%. This variation may again partly be a result of the inability of the muffle furnace to attain a temperature above 1000°C during the ignition step of sample preparation (see "Geochemistry" in the "Methods" chapter). Some of the variation also may be a result of errors during weighing of the sample or flux powders. In any case, we again normalized the raw major-element values to 100 wt% totals for use in the figures and in the discussion below; the normalized values are presented below the raw data in Table T7.

Alteration is much less severe overall in the Site U1347 lavas than in those recovered from Site U1346 (see "Alteration and metamorphic petrology"). Weight loss on ignition (LOI) is 0.07–3.57 wt%. In comparison, the LOI range for Site U1346 is 3.12–13.85 wt%. Alteration effects on the elements we measured are correspondingly much smaller, in general, than they are among the Site U1346 samples. For example, whereas data points for only four Site U1346 samples (the four with the lowest LOI values) lie in the field of tholeiitic basalt in a total alkalis versus SiO₂ diagram, values for all but two of the Site U1347 samples fall within this field in Figure F50A. The exceptions are the samples with the highest LOI value, Samples 324-U1347A-22R-1 (Piece 9, 55–58 cm) and 13R-6 (Piece 1, 80–83 cm), which are from a segregation of dark gray basalt within lighter gray basalt near the base of Unit IV. Values for Sample 22R-1 (Piece 9, 55–58 cm) place it (barely) above the tholeiitic-alkalic dividing line in the figure, whereas Sample 13R-6 (Piece 1, 80–83 cm) has the highest SiO₂ value measured for Site U1347 (52.29 versus 47.68–50.67 wt% for the other samples), putting it just within the field of basaltic andesite. Although total alkali contents remain relatively low in the Site U1347 samples, alteration nevertheless appears to have variably increased K₂O concentrations, as K₂O shows a rough positive correlation with LOI (Fig. F50B). Even so, the range in K₂O measured for Site U1347 (0.05–0.45 wt%) is far smaller than for Site U1346 (0.10–4.77 wt%). Several other elements (e.g., CaO) show relatively minor effects probably related to alteration. One of the most altered Site U1347 samples is the altered glass from the top of Unit V (interval 324-U1347A-13R-7 [Piece 6, 28–30 cm]). This sample has 3.40 wt% LOI and one of the highest K₂O values (0.34 wt%). It also exhibits the lowest concentrations of Sr, P, Zr, Ti, Y, and Sc (e.g., Fig. F51A; Table T7); the second lowest Ba concentration; the lowest CaO and Al₂O₃ (e.g., Fig. F52A); and the highest Fe₂O₃^T (total iron calculated as ferric oxide) and MgO contents.

Downhole variations in element concentrations and interelement ratios do not lend themselves to any single, simple generalization. However, variations from unit to unit are in some cases smaller than variability within a unit. For example, chemical differences between Units XII and XIV are smaller, overall, than the intraunit variation among the subunits of Unit X or those of Unit VII (Fig. F53).

As with the Site U1346 lavas, only limited overlap is seen between the Site U1347 data and the field of OJP basalts in most variation diagrams. Instead, the Site U1347 data again lie within the field of East Pacific Rise basalts in many diagrams (e.g., Fig. F52A–F52D) and extend from values near those of lavas from Site 1213 (southern Tamu Massif) toward higher TiO₂, Zr, P₂O₅, and Ba and lower Ni, Cr, and Mg# (Mg# = 100 × Mg²⁺/[Mg²⁺ + Fe²⁺], assuming Fe₂O₃/FeO = 0.15). For example, Mg# ranges from 58.9 to 47.4 among the Site U1347 basalts versus 60.8 to 57.6 for Site 1213. These characteristics are consistent with the Site U1347 basalts representing more (and variably) evolved relatives of the Site 1213 lavas. The sample of the segregation near the bottom of Unit IV (Sample 324-U1347A-13R-6 [Piece 1, 80–83 cm]), in addition to having the highest SiO₂ content, is distinct in having higher Ba, P, Zr, and Y (Fig. F51A); lower V and Sc; and higher Zr/Ti than the other holocrystalline Site U1347 samples. These differences do not appear to be a result of alteration, as the sample (LOI = 2.53 wt%) is not visibly more altered than several others with similar LOI. We infer that these characteristics are related to details of the particular differentiation history of the material in this segregation.

Incompatible-element patterns of the Site U1347 basalts are also broadly ocean-ridge-like (Fig. F51B–F51F), with a marked relative depletion of Ba. However, the lavas exhibit notable differences from normal ocean-ridge basalts; in particular, they are slightly enriched in Ba, Sr, P, Zr, and Ti relative to Y, similar to so-called enriched (or E-type) ocean-ridge basalts. In this respect, they are similar to, and in some cases even slightly more enriched than, the Site 1213 lavas. The difference from normal ocean-ridge basalt patterns could reflect a difference in the chemical composition of the source mantle; that is, the source(s) of the Site U1347 basalts could have been slightly richer in the more incompatible elements than normal ocean-ridge mantle. Alternatively (or additionally), the Site U1347 lavas could represent smaller and/or deeper mean fractions of partial melting than do most ocean-ridge basalts; because Y is a compatible element in garnet (e.g., Salters and Longhi, 1999), low relative Y concentrations can be a sign of control by residual garnet during melting. Unlike the Site 1213 or typical ocean-ridge basalts, many of the Site U1347 patterns display small peaks at P relative to Zr and Sr. Although this may be a primary characteristic of the lavas, we cannot rule out the possibility that it is an artifact of the onboard measurement of P₂O₅ (see "Geochemistry" in the "Methods" chapter). The P peaks seem unlikely to be a result solely of alteration, as most of the Site U1347 samples define a rough positive correlation of P₂O₅ with alteration-resistant elements like Ti and Zr. Additionally, like Zr, the combined Site U1347 and Site 1213 data also show a rough increase in Sr with decreasing Mg# and increasing TiO₂ (e.g., Fig. F52E), whereas the Sr/Zr ratio displays no correlation with Mg# and only limited variation (mean = 1.7, standard deviation = 0.2, excluding the altered glass and segregation). Because Sr is compatible in plagioclase (e.g., Bindeman and Davis, 2000), these characteristics could suggest that plagioclase removal during differentiation was not significant enough to obscure the effects of any earlier (preplagioclase) stages of differentiation and/or small variations in amount of partial melting. Resolution of the effects of variations in partial melting and magmatic differentiation on the Site U1347 lavas awaits detailed onshore analysis of a larger set of elements.

Total carbon and carbonate carbon

Thirty-seven samples of sedimentary material from Cores 324-U1347A-3R through 17R were analyzed in replicate for carbonate content (percent carbonate). The carbonate values are baseline-corrected for CO₂ in equilibrium with acid and scaled to a 100% CaCO₃ standard (see "Geochemistry" in the "Methods" chapter for an explanation of analytical procedures and instrumentation used for carbonate, total carbon, and organic carbon measurement). The samples were also analyzed for total carbon, although not in replicate. The content of organic carbon was estimated by subtracting the percentage of measured carbonate from that of total carbon. Results are presented and discussed in "Sedimentology."

Top of page   |   Previous   |   Next