doi:10.2204/iodp.proc.324.106.2010
Geochemistry
Major element oxide and trace element analyses
We analyzed eight samples of vesicular lava from stratigraphic Unit IV and five samples of autobreccia from stratigraphic Unit V in Hole U1349A for concentrations of major element oxides and several trace elements (Table T6) by ICP-AES (see "Geochemistry" in the "Methods" chapter for information on analytical procedures, instrumentation, and data quality). All of the samples were highly to completely altered (see "Alteration and metamorphic petrology"). Two splits were taken from Sample 324-U1349A-11R-5 (Piece 1A, 12–15 cm), which is an example of the mingling of two visually distinct types of lava observed in Unit IV (see "Structural geology"); one of the splits was from a zone of highly vesicular foamy lava and the other was from an immediately adjacent, red (oxidized), and much less vesicular zone.
Variation in total weight percentages for the major element oxides is somewhat smaller than at Sites U1346–U1348, but still significant (98.28–103.32 wt%), and we again normalized the raw major element values to 100 wt% totals. The normalized data are presented below the measured values in Table T6 and are used in the figures and in the discussion below.
Weight loss on ignition (LOI) is high in all samples, consistent with the extensive levels of alteration observed throughout the basement section at Site U1349. LOI values range from 10.03 to 34.76 wt% in samples from above ~200 mbsf (within Unit IV), with the exception of Sample 324-U1349A-9R-2 (Piece 11, 93–95 cm), which has only 4.63 wt% LOI. The highly vesicular (foamy lava) split of Unit IV Sample 324-U1349A-11R-5 (Piece 1A, 12–15 cm) yielded the 34.76 wt% value. This split is dominated by calcium carbonate, as evidenced by its very high CaO (65.97 wt%) and low SiO2 (15.37 wt%) contents, and we do not discuss it further. The split from the adjacent reddish, less vesicular zone has lower but still very high, LOI (14.47 wt%) and CaO (14.77 wt%). Three other samples from Unit IV contain significant amounts of CaCO3, as reflected in their high CaO (19.13–22.79 wt%) and CaO/Al2O3 (1.14–1.37) values. Despite the intense alteration in the lower part of the basement section (see "Alteration and metamorphic petrology"), samples from below ~200 mbsf (Unit V and the lower part of Unit IV; Sections 324-U1349A-12R-3 through 16R-6) yielded lower LOI values, from 2.37 to 6.51 wt%, except for Sample 324-U1349A-14R-1 (Piece 9, 38–42 cm) with 10.65 wt%.
Data for the four lowermost samples from Unit V lie in the field of tholeiitic basalt in a total alkalis versus silica diagram (Fig. F58A); these samples have four of the five lowest LOI values measured, between 2.37 and 4.99 wt%. Samples from shallower portions of the hole exhibit more alkalic, lower SiO2 compositions. As at Site U1346 (see "Geochemistry" in the "Site U1346" chapter), this behavior appears to be a result of alteration, for SiO2 shows a negative trend with LOI and a rough positive trend with K2O content (Fig. F58B). In contrast, Na2O defines only a weak negative trend with LOI; therefore, as at Site U1346, K2O is largely responsible for the increase in total alkali content with increasing LOI. The highest K2O value, 5.85 wt%, is reached in the reddish split of Sample 324-U1349A-11R-5 (Piece 1A, 12–15 cm), which has an LOI of 14.47 wt%. Increases in concentration with increasing LOI are also seen for CaO, P2O5, Sr, Ba, and Y, whereas MgO and Cu generally decrease with increasing LOI.
Contents of K2O in the three lowermost samples from Unit V are all very low (≤0.02 wt%). The LOI values of these samples (2.37–4.99 wt%), although small relative to values for Site U1349 overall, are nevertheless very large compared to LOI values for unaltered basalts. Given the very thorough but nonoxidative character of the alteration in Unit V (see "Alteration and metamorphic petrology"), we infer that these three samples have lost some potassium. Two of these samples and another sample from the lower part of Unit V also appear to have lost phosphorus, as they have very low P2O5 values (≤0.01 wt%; see Table T6).
Mg# ranges widely at Site U1349 (Mg# = 100 × Mg2+/[Mg2+ + Fe2+]), assuming that Fe2O3/FeO = 0.15). The samples with the smallest Mg#, between 38.1 and 49.3, are those with the greatest LOI (≥10.03 wt%), which is consistent with variable reduction of MgO content during alteration of these rocks. In contrast, the samples with LOI ≤6.51 wt% have distinctly larger Mg#, from 60.6 to 76.2; three of the Unit V autobreccia samples have values >70. As with the lower Mg# basalts, alteration probably modified Mg# in these samples. However, the highest Mg# values are likely to also reflect accumulation of olivine crystals (and perhaps clinopyroxene), consistent with the relatively abundant pseudomorphs after olivine (and, where not altered to the point of being unidentifiable, clinopyroxene) observed in thin section (see "Igneous petrology" and "Alteration and metamorphic petrology").
Titanium and Zr are two of the more alteration-resistant elements in mafic rocks (e.g., Humphris and Thompson, 1978) and are incompatible in olivine and clinopyroxene (e.g., Salters and Longhi, 1999). In contrast to Mg#, TiO2 variation is small at Site U1349, between 0.88 and 1.18 wt% (Fig. F59A). These values are lower than observed for lavas of the other Expedition 324 sites or Leg 198 Site 1213. Zirconium concentrations are correspondingly low, between 44 and 55 ppm. In Figure F59B, the combined Zr-TiO2 data form a small cluster at the lower end of the ocean-ridge basalt array and overlapping the Ontong Java field. Thus, the Site U1349 basalts appear to represent more primitive (less evolved) compositions than yet found elsewhere on Shatsky Rise. At least in terms of TiO2 and Zr, they are similar to primitive ocean-ridge basalts and have only slightly higher concentrations than the most primitive basalts recovered from the OJP (the Kroenke basalts; Mahoney, Fitton, Wallace, et al., 2001; Fitton and Godard, 2004). Strontium and Y contents, despite being affected variably by alteration, are also low among the lower LOI Site U1349 samples (50–94 and 9–26 ppm, respectively), again comparable to values in primitive ocean-ridge basalts.
Results for Cr and Ni are likewise consistent with relatively primitive magmatic compositions in that concentrations of both elements are high for basalt (Cr = 338–607 ppm; Ni = 124–350 ppm) (Fig. F59C, F59D). However, the data must be interpreted cautiously, as the severe alteration at Site U1349 has probably affected concentrations of both elements to some extent, particularly Ni (cf. Site U1346 Ni data in the figure). Equally important, Ni and Cr are compatible in mafic phases and in chromian spinel, present as inclusions in olivine (see "Igneous petrology") (e.g., Bougault and Hekinian, 1974; Leeman and Scheidegger, 1977; Skulski et al., 1994; Righter et al., 2006). Accumulation of these minerals thus raises bulk rock Ni and Cr concentrations above those in the melt phase.
Besides olivine accumulation, possible evidence for clinopyroxene accumulation in some Site U1349 basalts is provided by Sc, which tends to be a compatible element in clinopyroxene but incompatible in other common basaltic minerals (e.g., Bacon and Druitt, 1988). Variation in Sc is small (44–56 ppm), but concentrations range to slightly higher values than found for most ocean-ridge or Ontong Java basalts at similar TiO2 levels (Fig. F59E). If this feature is a pre-alteration characteristic of the lavas, it could be a signature of clinopyroxene accumulation.
In summary, seeing through the chemical effects of severe alteration and crystal accumulation is difficult, but the Site U1349 basalts appear to represent significantly less differentiated tholeiitic magmas than any of the basalts recovered at the previously drilled Expedition 324 sites or Site 1213.