We performed chemical analyses on eight harzburgites, one dunite, and seven gabbros from Site 1270 selected by the shipboard scientific party, using inductively coupled plasma–atomic emission spectrometry (ICP-AES) for determining major and trace element concentrations and gas chromatography for H2O and CO2 (because of a technical problem, sulfur was not analyzed in most of the Site 1270 rocks). These 16 samples are representative of the rocks recovered from all four holes drilled at Site 1270 (see "Igneous and Mantle Petrology" and "Metamorphic Petrology" for the characterization of the lithologic units). We sampled two harzburgites from Hole 1270A, seven gabbros from Hole 1270B, one harzburgite from Hole 1270C, and five harzburgites (including one intruded by a gabbroic vein) and one dunite-harzburgite from Hole 1270D. The results for the major and trace elements, for both ultramafic and mafic rocks, are reported on a volatile-free basis in Table T4.
Site 1270 peridotites are characterized by high loss on ignition (LOI) values (>8.5 wt%). Sample 209-1270D-3R-1, 63–66 cm, which is a harzburgite intruded by a gabbroic dikelet, has the lowest LOI. The other Site 1270 peridotites have significantly higher LOI (13.0–14.4 wt%) and H2O (14.0–15.2 wt%), similar to that of lizardite (Fig. F94). These high LOI values and H2O concentrations are consistent with the visual core descriptions and XRD results (see "Metamorphic Petrology"), showing that Site 1270 peridotites are predominantly altered to serpentine. The gabbroic dikelet crosscutting Sample 209-1270D-3R-1, 63–66 cm, is also altered predominantly to talc (see "Metamorphic Petrology" and "Site 1270 Thin Sections"). Because talc is characterized by lower water content relative to lizardite, its presence in this sample along with serpentine leads to a lower LOI (8.5 wt%) and H2O (9.3 wt%) compared to the other peridotites.
Compared to Site 1268 peridotites, bulk rock compositions of Site 1270 peridotites show relatively restricted ranges for SiO2 (44.5–46.5 wt%), MgO (41.4–44.6 wt%), and Fe2O3 (8.15–9.44 wt%) concentrations (Fig. F94). These ranges are similar to what has been observed in peridotites from the MARK area at Site 920 (Casey, 1997). Sample 209-1270D-3R-1, 63–66 cm, is distinguished from other Site 1270 peridotites by its slightly higher SiO2 (49.4 wt%) and lower MgO (35 wt%) and Fe2O3 (7.40 wt%) concentrations (Fig. F94). However, its most distinctive characteristics are high Al2O3 (4.1 wt%) and CaO (3.48 wt%) concentrations (Fig. F95). The composition of this sample is displaced from the other peridotites because of the presence of an altered crosscutting gabbroic dikelet. This dikelet underwent replacement of plagioclase by talc and clinopyroxene by chlorite and amphibole (see "Site 1270 Thin Sections").
With the exception of Samples 209-1270D-3R-1, 63–66 cm, and 1R-1, 70–74 cm, the Site 1270 peridotites are characterized by low Al2O3 (0.33–0.85 wt%) and CaO (0.33–0.85 wt%) concentrations. These elements are concentrated in pyroxenes and, more precisely, in clinopyroxene for CaO. They can be used as a proxy for the degree of fertility of peridotites. The Al2O3 contents of Site 1270 peridotites are similar to those of the most enriched Site 1268 peridotites but are less than the Al2O3 contents of Leg 153 peridotites. This suggests that prior to alteration to lizardite Site 1270 peridotites were more refractory than Leg 153 peridotites. However, Al2O3 and CaO are not correlated in Site 1270 peridotites, as would be expected for residues of partial melting (Fig. F95). Site 1270 peridotites, in fact, display low CaO/Al2O3 ratios compared to Bulk Silicate Earth (Jagoutz et al., 1979; Wänke, 1981), ranging between 0.76 and 0.01. Site 1268 peridotites are also depleted in CaO relative to Al2O3. This depletion may stem either from preferential Ca loss during serpentinization and weathering or from the increasing importance of orthopyroxene and spinel in the Al2O3 budget at high degrees of melting (especially in the case of the Site 1270 peridotites). The extremely low CaO content of some of these rocks is consistent with only minor amounts of clinopyroxene remaining after melting.
An additional consideration is the effect of carbonate precipitation in some of the Site 1270 peridotites (see "Site 1270 Thin Sections"). This precipitation is most clearly identified by Ca enrichment (CaO = 3.49 wt%) in Sample 209-1270D-1R-1, 70–74 cm. In a diagram of CaO vs. CO2, this sample's composition is significantly displaced from that of the other peridotites analyzed from Site 1270 (Fig. F96). This displacement lies along a vector toward the composition of CaCO3, consistent with the presence of carbonate in this sample as described in thin section. Some Site 920 peridotites also lie along this trend (also see Fig. F45 and accompanying text in the "Leg 209 Summary" chapter).
Sr concentrations in Site 1270 peridotites display a wide range of values (<5–530 ppm) (Fig. F97). Sr behaves as an incompatible element during partial melting. However, strontium is also highly mobile during alteration. Sr variations are correlated with Ca variations in Site 1270 peridotites. Probably as a result of secondary formation of carbonate during alteration, Sample 209-1270D-1R-1, 70–74 cm, is also the most Sr-enriched peridotite at Site 1270 (also see Fig. F46 and accompanying text in the "Leg 209 Summary" chapter).
Site 1270 peridotites are depleted in TiO2 (<0.01–0.02 wt%), Y (<2 ppm), and Zr (<2–3.6 ppm) and, to a lesser extent, in V (21–45 ppm) and Sc (5–9 ppm) compared to Leg 153 peridotites (Fig. F98). These elements behave moderately incompatibly, preferentially partitioning into the liquid during partial melting. Their concentration range suggests that Site 1270 peridotites underwent higher degrees of partial melting than Leg 153 peridotites. Cr and Ni behave as compatible elements, preferentially partitioning into the solid during partial melting. Site 1270 peridotites display high concentrations of Cr (1600–2930 ppm) and Ni (2000–2900 ppm). In both Site 1268 and 1270 peridotites, Cr is positively correlated with Al2O3. Because nearly all of the Cr must be concentrated in spinel in these rocks, the correlation of Cr with Al2O3 (Fig. F99) is evidence that spinel contributes a significant proportion to the bulk rock Al budget for the highly refractory peridotites from Sites 1268 and 1270, in addition to any pyroxene present. The relatively high Ni contents, in conjunction with low Al2O3 (Fig. F99), are explained by the relatively large modal proportion of olivine, in which nickel is compatible, in these peridotites prior to their alteration. Ti, Y, Zr, V, Sc, Cr, and Ni are usually considered to be relatively immobile during alteration (e.g., Hebert et al., 1990). The similarity of their concentrations for the Site 1270 and 1268 peridotites therefore suggests that prior to alteration Site 1268 and 1270 peridotites had similar compositions.
Sample 209-1270D-3R-1, 63–66 cm, which is intruded by a gabbroic dikelet, displays significantly higher incompatible element contents: TiO2 (0.13 wt%), Y (20 ppm), Zr (96 ppm), and Sc (10 ppm). More compatible Co, Cr, and Ni have similar concentrations in Sample 209-1270D-3R-1, 63–66 cm, and other Site 1270 peridotites (the slight difference in Ni and MgO concentrations reflects the low proportion of olivine in Sample 209-1270D-3R-1, 63–66 cm). For this particular sample, compatible elements are mainly concentrated in the peridotite, whereas incompatible element variations mostly reflect the composition of the crosscutting gabbroic vein. The high incompatible element content of Sample 209-1270D-3R-1, 63–66 cm, suggests that the gabbroic vein is incompatible element enriched compared to gabbroic rocks having similar modal compositions from Sites 1268 and 1270.
We analyzed seven gabbroic rocks from Hole 1270B. The recovered sequence of gabbros was divided into seven lithologic units and two major rock types (microgabbro and oxide gabbro) using visual core and thin section descriptions (see "Hole 1270B" in "Igneous and Mantle Petrology"). We selected two oxide gabbro samples from Unit I, three oxide gabbro samples from Unit VII, and a microgabbro from each of Units II and VI.
The gabbros can be subdivided into two groups that correspond to the main rock types assigned using visual core and thin section descriptions on the basis of their major and trace element geochemistry. Oxide gabbros have MgO (6.0–8.4 wt%), TiO2 (1.8–6.5 wt%), Al2O3 (10.7–15.8 wt%), and Fe2O3 (14–18.5 wt%). In contrast, the microgabbros effectively show no overlap in composition: MgO (10.7–12.5 wt%), TiO2 (<0.8 wt%), Al2O3 (15.5–17.5 wt%), and Fe2O3 (10–13 wt%) (Fig. F100).
The correlation between Fe2O3 and TiO2 (Fig. F101) shows that the variation in these two elements is related to oxide abundance in the rock, with the microgabbros having only a minor oxide content and therefore low amounts of both Fe2O3 and TiO2.
Apart from one sample, all of the microgabbros and oxide gabbros have low LOI (<0.6 wt%), low H2O (<1.2 wt%) (Fig. F100), and low CO2 (0.06–0.09 wt%), attesting to their relatively low degree of alteration. A single sample of microgabbro (Sample 209-1270B-7R-1, 104–106 cm) has higher LOI (5.4 wt%), H2O (7.7 wt%), and CO2 (0.14 wt%) and also has distinctly lower CaO (7.34 wt%). Otherwise, there appears to be no systematic relationship between volatile species and other elements in the gabbroic rocks.
Compatible trace element geochemistry of Hole 1270B gabbros also reflects the two groupings defined by the major elements. The oxide gabbros have low concentrations of Cr (<121 ppm) and Ni (<165 ppm, the detection limit), whereas the microgabbros have significantly higher concentrations of Cr (~760 ppm) and Ni (200 and 240 ppm) (Fig. F102). Both of the gabbro types recovered from Hole 1270B have higher concentrations of Cr and Ni compared to those recovered from Site 1268 and overlap with the range in concentrations of Hole 735B and in the Leg 153 gabbros.
Zr and V show contrasting covariations with TiO2 in the Site 1270 gabbros (Fig. F103). The microgabbros have low TiO2 (~0.8 wt%) and V (200–240 pm) and relatively elevated Zr (22–35 ppm). In contrast, the oxide gabbros, with up to 6.5 wt% TiO2, display an increase in Zr with increasing TiO2, ranging up to 46 ppm Zr. Zr in the gabbroic rocks from Site 1270 is generally higher than that observed in the gabbroic rocks from Site 1268 and lies at the low end of the concentration range observed in Hole 735B and Leg 153 gabbros. The one sample of harzburgite with the altered gabbroic dikelet (Sample 209-1270D-3R-1, 63–66 cm, discussed above with the other peridotites from Site 1270) has the highest Zr (96 ppm) observed in the rocks from Site 1270. V in the oxide gabbros behaves differently from Zr (Fig. F104), showing a peak in V (1540 ppm) near 3 wt% TiO2. Above this, TiO2 and V contents decrease markedly (to 600 ppm V at 6.5 wt% TiO2).
Sc covaries with V in the Site 1270 gabbros (Fig. F104), showing relatively uniform Sc values (~65 ppm) in the oxide gabbros and lower Sc (~40 ppm) in the microgabbros. The Sc content of the microgabbros overlaps with that observed in Site 1268 and Leg 153 gabbros. Y covaries with Zr in the Site 1270 gabbros (Fig. F104), and both elements show a slight enrichment relative to the gabbros from Site 1268 but overlap with the range seen in gabbros from Hole 735B and Leg 153.
Geochemical data on Site 1270 and 1268 peridotites suggest that prior to alteration Site 1268 and 1270 peridotites had similar chemical compositions. The bulk rock analyses of Site 1270 peridotites show that all these rocks have been modified by alteration, predominantly to lizardite, leading to the addition of variable amounts of volatile constituents. The addition of carbonate has affected some of the Site 1270 peridotites. The effects of this additional alteration are most clearly identified in Sample 209-1270D-1R-1, 70–74 cm, which shows an enrichment in both CaO and CO2 compared to other peridotites analyzed from Site 1270 (Fig. F97). One other analyzed sample (209-1270D-3R-1, 63–66 cm) is composed of altered peridotite intruded by a plagioclase-rich gabbroic vein. This rock has markedly lower H2O, MgO, SiO2, and Fe2O3 (Fig. F94) and higher CaO and Al2O3 (Fig. F95). This rock is significantly enriched in trace elements such as Zr and Y, due to the presence of this vein.
Peridotites from Sites 1268 and 1270 show systematically lower Al2O3 compared to those from Leg 153 as well as most other abyssal peridotites (median = 1.4 wt%; Bodinier and Godard, 2003). In highly refractory peridotites such as those we have analyzed, the modal abundance of spinel, together with orthopyroxene, can exert a strong influence on the amount of Al2O3 present in the bulk rock. Because Cr is dominantly concentrated in spinel in these rocks, this effect produces a positive correlation between Cr and Al2O3. Prior to serpentinization, the peridotite protolith also underwent a relative large depletion in CaO, due to a pronounced depletion of clinopyroxene, probably during melting.
Oxide abundance appears to control many of the chemical systematics in the Site 1270 gabbros, evidenced by the good correlation between Fe2O3 and TiO2. The microgabbros have only a minor oxide component. For the oxide gabbros with TiO2 < 3 wt%, the visual estimate of modal oxide abundances agrees reasonably well with the estimates of the amount of magnetite present determined from bulk magnetic susceptibility (see "Discussion" in "Igneous and Mantle Petrology"). However, at higher TiO2 concentrations, the estimates of magnetite abundance from magnetic susceptibility measurements underestimates the amount of oxide observed in these rocks, which is consistent with an increased proportion of ilmenite in the oxide assemblage.
The Ti/V elemental ratio in oxide gabbros at Site 1270 shows a large range (10–60). In comparison, the microgabbros Ti/V (20) is similar to the value found in mid-ocean-ridge basalt (MORB) (Shervais, 1982). The igneous behavior of V is unique because its variable oxidation state (+3, +4, or +5) affects its partitioning behavior during magma evolution. At relatively low oxygen fugacity, such as is typically found in MORB systems, V is present in the +3 state and behaves as a relatively incompatible element, similar to Ti. Early stages of crystallization of olivine + plagioclase will produce a fractionation trend of relatively constant Ti/V ratios because these phases have low partition coefficients for both Ti and V. The crystal/liquid distribution coefficient (bulk D) for V in ocean ridge basaltic systems is usually greater than the bulk D for Ti, so extensive silicate fractionation can drive magma to higher Ti/V ratios (Shervais, 1982). Once magnetite becomes a crystallizing phase, Ti and V concentrations will decrease and there may be a dramatic increase in Ti/V ratio because V is highly compatible in magnetite (partition coefficient [KD] for magnetite in MORB = ~5–70, depending on oxygen fugacity) (Shervais, 1982). The variation of V with Ti in the oxide gabbros of Site 1270 (Fig. F103) is consistent with the presence of a significant quantity of oxide in these rocks (5–15 vol%) (see "Site 1270 Thin Sections"). The two oxide gabbros at Site 1270 that have high Ti/V ratios (20–50) also have high TiO2 (4–6.5 wt%), consistent with a higher proportion of ilmenite in these rocks.
The Sc budget in gabbroic rocks is largely controlled by the amount of clinopyroxene present. The Sc–V relationship shown in Figure F104 is consistent with the relatively uniform amount of clinopyroxene in the oxide gabbros (35–45 vol%) based on the igneous petrology thin section descriptions. However, Sc may also be compatible in magnetite, along with V. Notably, the Leg 153 gabbros show a sharp decrease in Sc content when V is above ~500 ppm.
An intriguing question is whether there is a genetic linkage between the gabbroic vein observed in the harzburgite analyzed from Hole 1270D and the gabbroic rocks from Hole 1270B. Despite the fact that no attempt was made to separate the vein and harzburgite lithologies prior to the analysis of Sample 209-1270D-3R-1, 63–66 cm, this sample has Zi and Y contents that are higher than those in the most enriched oxide gabbros from Hole 1270B. The approximate proportions observed in this sample were 3:1 to 4:1 (harzburgite:gabbro). For illustrative purposes, we have made an estimate of the composition of the gabbroic liquid by subtracting the mean composition of the other peridotites at Site 1270 from the analysis of this sample, renormalized to 100%. The composition of this reconstituted gabbro is, of course, sensitive to the proportion of harzburgite estimated in the bulk analysis, and this is uncertain. Nevertheless, the illustration demonstrates that any such gabbro composition has SiO2 near 60 wt%. To obtain reasonable MgO contents by this calculation (2–10 wt%) for the gabbro vein requires the proportion of harzburgite to be 75%–80%. This produces ranges of Zr (375–470 ppm), Y (80–100 ppm), V (43–47 ppm), Sc (20–23 ppm), TiO2 (0.5–0.6 wt%), and Fe2O3 (1.0–2.6 wt%) for the reconstituted vein. These values are shown in Figures F101 and F104 for comparison to the other gabbro results.
The above calculation reveals that trace element–enriched liquids may be present in the crust beneath Site 1270. Similar enrichments in Zr and Y have been observed previously in some of the Leg 153 gabbros, where Zr and Y range to values >1400 and 150 ppm, respectively. The concentrations estimated for Zr (375 ppm) and Y (80 ppm) for the gabbroic vein are also consistent with simple mass balance estimates of equilibrium liquid compositions for the gabbros analyzed from Site 1270. For example, taking bulk rock Zr and Y contents of 20 and 12 ppm, respectively (as observed in the microgabbros), assuming all the Zr and Y in the rock resides in clinopyroxene with a modal abundance of 40%, and assuming appropriate KD values for Zr (0.12) and Y (0.47) for clinopyroxene (Hart and Dunn, 1993), the equilibrium liquid will have values of Y (65 ppm) and Zr (400 ppm) quite similar to the values calculated independently for the reconstituted vein.