Previous   |    Next

doi:10.2204/iodp.proc.330.107.2012

Geochemistry

Igneous rocks

Concentrations of major elements and several trace elements were measured for 13 samples of igneous rocks from Site U1376 on Burton Guyot (Table T9) by inductively coupled plasma–atomic emission spectroscopy (see “Geochemistry” in the “Methods” chapter [Expedition 330 Scientists, 2012a] for information on analytical procedures, instrumentation, and data quality). All but three of the samples were collected from lava flows, lobes, and clasts in Units III and IV. Two samples were from aphyric dikes in Unit IV (Samples 330-U1376A-15R-2 [Piece 2, 35–37 cm] and 23R-1 [Piece 8, 98–102 cm]), and one was from a lava clast in the conglomerate of Subunit IIB (Sample 5R-2 [Piece 5, 128–130 cm]).

As at previous sites, total weight percentages for the major element oxides vary significantly, in this case from 91.06 to 102.37 wt%. Possible reasons for this variation are discussed in “Geochemistry” in the “Methods” chapter (Expedition 330 Scientists, 2012a). To better compare our results with one another and with data from the literature, we normalized the raw major element values to 100 wt% totals. The normalized values are presented below the raw data in Table T9 and are used in the figures and in the discussion below.

Weight loss on ignition (LOI) varies from 0.9 to 5.4 wt%, but only two samples (Unit III Sample 330-U1376A-7R-4 [Piece 6, 122–124 cm] and Unit IV Sample 21R-4 [Piece 4, 82–84 cm]) have values of >2.8 wt%. Values for the other samples indicate moderate overall levels of alteration (for comparison, unaltered basalt is typically characterized by values of <1 wt%; e.g., Rhodes, 1996) and are generally consistent with the petrography of the rocks (see “Alteration petrology” and “Igneous petrology and volcanology”). Alteration tends to affect K₂O more than it does any of the other elements analyzed (e.g., see “Geochemistry” in the “Site U1372” chapter [Expedition 330 Scientists, 2012b]). The range in K₂O at Site U1376 is 0.47–1.56 wt%, and a slight downhole increase is evident in that four of the five samples with K₂O > 0.90 wt% are in Unit IV, which corresponds with four of the six samples having LOI > 2.0 wt% in Unit IV. However, no overall correlation is present between K₂O and LOI. Indeed, the two samples with the highest LOI have among the lowest K₂O concentrations measured (0.47 and 0.63 wt%). Nevertheless, alteration has probably modified K₂O contents to varying amounts because K₂O, which is an incompatible element in mafic systems, does not correlate significantly with the alteration-resistant incompatible elements TiO₂ and Zr.

Data for Site U1376 largely overlap with those for Site U1372 on Canopus Guyot in a total alkalis (Na₂O + K₂O) vs. SiO₂ diagram, but not with Sites U1373 and U1374 on Rigil Guyot (Fig. F38). Because the Site U1376 samples have comparatively low total-alkali contents relative to their SiO₂ values, as a group they are the least alkalic rocks of Expedition 330. Data for most of the samples lie in the field of alkalic basalt, but four samples (Unit III Samples 330-U1376A-7R-4 [Piece 6, 122–124 cm] and 13R-4 [Piece 1, 0–3 cm] and Unit IV Samples 16R-2 [Piece 1, 10–12 cm] and 21R-4 [Piece 3, 82–84 cm]) have values that fall below the alkalic-tholeiitic dividing line. Like four similar samples from Site U1372, these rocks are classified as transitional basalt rather than tholeiites because either titanaugite or olivine were identified in their groundmasses (see “Igneous petrology and volcanology”).

Mg number (Mg# = 100 × Mg²⁺/[Mg²⁺ + Fe²⁺], assuming Fe₂O₃/FeO = 0.15) varies from 50.8 to 73.3 and averages 64.4, higher than for Sites U1372–U1374 (for which the averages are 50.4, 55.5, and 49.5, respectively). Three of the highest Mg numbers (71.7–73.3) are found in transitional basalt. The total range in MgO is considerable (4.91–16.79 wt%). However, the average value is high (10.74 wt%), and seven samples have MgO ≥ 11.85 wt%. Similarly, Ni varies from 69 to 741 ppm, with a high average value of 320 ppm. In an Al₂O₃ vs. MgO diagram (Fig. F39A) the Site U1376 data follow a broadly similar trend to other Expedition 330 samples. This trend is indicative of a dominant control on magmatic composition by olivine, with little influence from plagioclase. All of the Site U1376 samples with MgO ≥ 11.85 wt% contain abundant olivine phenocrysts and are likely to contain excess olivine. For example, within the 33 m thick massive basalt flow of Unit III, MgO varies from 12.06 to 15.84 wt%, probably largely as a result of local variation in the amount of olivine phenocrysts. Scandium concentrations at Site U1376 range from 20 to 33 ppm and increase slightly with decreasing MgO, whereas CaO/Al₂O₃ values are relatively high (≥0.78) with one exception (Fig. F39B). Although the Unit III rocks contain appreciable amounts of augite phenocrysts (see “Igneous petrology and volcanology”), fractionation of augite in its magma source appears to have been minor overall.

As a group, the Site U1376 basalts tend to have slightly lower Al₂O₃ for their MgO contents than rocks from the other Expedition 330 sites. Unit IV lava Sample 330-U1376A-21R-4 (Piece 3, 82–84 cm), which has the highest MgO value, stands out in having even lower Al₂O₃ relative to its MgO content (Fig. F39A). Its Al₂O₃ concentration is the lowest encountered during Expedition 330, and its CaO/Al₂O₃ is the highest (Fig. F39B). This sample does not contain augite phenocrysts, so the high CaO/Al₂O₃ is not caused by augite accumulation. Aluminum is not very sensitive to alteration, but CaO contents can be modified by some types of alteration (e.g., see “Geochemistry” in the “Site U1372” chapter [Expedition 330 Scientists, 2012b]). Even though the sample has the highest LOI value measured for Site U1376, its high CaO/Al₂O₃ is not a result of high CaO but of low Al₂O₃. A low Al₂O₃ content in a high-MgO rock is suggestive of relatively small amounts of partial melting with residual garnet in the mantle source. This sample also has higher Fe₂O₃^T (total iron as Fe₂O₃) than the other high-MgO samples (Fig. F39C) and also is consistent with relatively high pressure melting in the garnet stability field. On the other hand, the sample has Zr/Y, Ba/Y, and Ti/Y ratios similar to those of the other Unit IV lava samples. Melts formed from a source with residual garnet should have higher values of these ratios than melts from garnet-free sources because Y, unlike Zr, Ba, or Ti, is compatible in garnet. Thus, the composition of this sample presently remains somewhat puzzling.

Incompatible trace element characteristics of the Site U1376 basalt generally resemble those of Sites U1372–U1374 (e.g., Fig. F40). The greatest overlap is with Site U1372. Within Site U1376, chemical distinctions can be drawn approximately along stratigraphic unit lines. Except for highly olivine-phyric Unit IV Sample 330-U1376A-21R-4 (Piece 3, 82–84 cm), Unit III has the highest Ni, Cr, and MgO contents and Mg numbers (e.g., Fig. F41A, F41B) and the lowest TiO₂, Sr, Zr, and Y concentrations (e.g., Fig. F41C). Moreover, variations in these quantities within Unit III, including both the three samples from the 33 m thick flow and the basalt from lithologic units above and below this flow, are small relative to those within Unit IV. The uppermost Unit IV lava Sample 330-U1376A-16R-2 (Piece 1, 10–12 cm) is chemically similar to Unit III in Mg number, Ni and TiO₂ contents, and Zr/Y ratio (e.g., Fig. F41A–F41D). The Unit II lava clast (an olivine-phyric Type 1 clast; see “Igneous petrology and volcanology”) broadly resembles the two lowest-MgO olivine-phyric Unit IV lava samples in some respects (e.g., Fig. F41A, F41B), but it has a distinctly lower Zr/Y ratio (5.92) than that measured for any sample from either Units III or IV (Fig. F41D). Values of Zr/Y for Unit III and the uppermost lava sample from Unit IV vary between 6.87 and 7.13, whereas values for the remaining Unit IV lava samples are even higher, varying between 7.61 and 8.05. This ratio, which is insensitive to fractional crystallization or crystal accumulation of olivine, clinopyroxene, or plagioclase, is an indicator of variation in the amount of partial melting and/or in mantle source composition. Thus, three distinct magma types appear to be present among the Site U1376 lava samples, and these correlate approximately with petrographically defined Units II–IV. Interestingly, downhole differences in Ba/Y among the lava samples of the three stratigraphic units are much less pronounced than those in Zr/Y. Yet in mafic and ultramafic systems, the difference in incompatibility between Ba and Y is greater than that between Zr and Y. Therefore, for a given range of partial melting, the Ba/Y ratio ordinarily varies more in magmatic rocks than the Zr/Y ratio. That this is not observed at Site U1376 suggests that the differences in Zr/Y may reflect source differences rather than variations in partial melting. Finally, the chemical characteristics of the aphyric basalt dike near the bottom of Unit IV are not significantly different from those of the two olivine-phyric Unit IV lava samples from Sections 330-U1376A-17R-3 and 19R-1. The dike near the top of Unit IV is similar in some respects but has significantly higher Ba/Y than any of the other Site U1376 basalts (Fig. F41E).

Carbon, organic carbon, nitrogen, and carbonate

No samples from Site U1376 were analyzed for carbonate, total carbon, total organic carbon, or total nitrogen content.

Top of page   |    Previous   |    Next