U-series disequilibria in MORB from the Garrett Transform and implications for mantle melting

Abstract

Here, we report ²³⁸U–²³⁰Th–²³¹Pa–²²⁶Ra disequilibria and ⁸⁷Sr/⁸⁶Sr measurements in 11 mid-ocean ridge basalt (MORB) glasses from the Garrett Transform (∼13°30′S latitude on the East Pacific Rise [EPR]) whose compositions range from primitive, depleted high-MgO basalts to evolved basalts. U and Th concentrations of samples range between 3 and 75 ppb and between 6 and 220 ppb, respectively, with a corresponding large variation in Th/U (1.5–2.9). (²³⁰Th)/(²³²Th) varies from 1.2 to 1.6 such that (²³⁰Th)/(²³⁸U) range from ∼15% excess ²³⁰Th in a high-Th/U evolved sample to ∼25% excess ²³⁸U in a high-MgO sample with low Th/U. Out of 11 samples, 7 have ²³⁸U excess, an unusual feature for MORB. All samples have ²²⁶Ra excesses, with (²²⁶Ra)/(²³⁰Th) varying between 1.3 and 3.8 constraining ages since eruption to <8000 years and the measured (²³⁰Th)/(²³²Th) to be within a few percent of its value at eruption. ⁸⁷Sr/⁸⁶Sr ratios range between 0.7022 and 0.7024 and poorly correlate with (²³⁰Th)/(²³⁸U).

Comparing the Garrett Transform to the Siqueiros Transform shows a remarkable correspondence between sample setting, composition and disequilibria systematics. Both settings produce linear trends of (²³⁰Th)/(²³⁸U) as a function of Th/U, consistent with mixing between two melts derived from different depths in the melting column. The mixing relationships are identical in both locations: The most incompatible rich samples with the highest Th/U and ²³⁰Th excess come from the ridge–transform intersection (RTI), whereas the most incompatible element poor basalts with the lowest (²³⁰Th)/(²³⁸U) and Th/U are erupted along leaky transform faults. Samples with intermediate Th/U and (²³⁰Th)/(²³⁸U) all come from within intra-transform spreading centers, consistent with the spreading centers acting to homogenize these diverse magmas. The cause of variation in Th/U could reflect either melting processes or different long-lived sources. No clear indication exists within these data.

Introduction

Mid-ocean ridges are the dominant crust-producing domains on Earth. Adiabatic decompression melting of upwelling mantle peridotite produces mid-ocean ridge basalt (MORB). The composition of the melt depends on the source material, degree of melting, depths of initiation and termination of melting, how melt–solid segregation occurs, and extent of fractional crystallization and magma mixing. As numerous studies have demonstrated, considerable chemical and isotopic variability exists in MORB erupted both on and off the ridge axis (e.g., [1], [2], [3]). Off-axis basalts (e.g., seamounts and transform basalts) are subject to similar melting processes as on-axis basalts. However, they are less dominated by magma mixing processes [4], [5], [6] and thus show an even greater degree of chemical variability [7], [8], [9], [10], [11], [12], [13], [14]. However, because only a few of these MORB suites have also been measured for their isotopic compositions of Hf, Nd, Sr, and Pb (e.g., [15]), the extent to which this chemical variability reflects variations in the melting processes, as opposed to variable mixing of melts from depleted (peridotitic) and enriched (pyroxenitic) sources is not well constrained.

There is little debate that compositional variability in some MORB reflects mantle source heterogeneity. Differences in long-lived isotopes occurring over short spatial scales support the proposition of a chemically heterogeneous mantle (e.g., [16], [17], [18], [19], [20], [21]). However, the role of melting a heterogeneous mantle in generating uranium series (U-series) disequilibium and trace element fractions in MORB, remains uncertain. Source heterogeneity and its role in melting has been recently addressed by Sims et al. [15] who, based on combined measurements of U-series disequilibria (U–Th–Ra and U–Pa), radiogenic isotopic compositions (Nd, Sr, Hf, and Pb) and major and trace element abundances, propose a mantle source beneath the 9–10°N segment of the East Pacific Rise (EPR) that is chemically homogeneous over the length scale of melting. Furthermore, recent 2D modeling also shows that significant fractionation of trace element ratios and U-series nuclides can occur as a result of melting processes without requiring source heterogeneities [22], [23], [24].

U-series disequilibria often are used to constrain the timescales of melt generation, melt ascent processes and differentiation (e.g., [2], [3], [15], [17], [20], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38], [39]). U-series data provide melt generation and ascent constraints because these magmatic processes operate over time scales similar to the half-lives of ²³⁰Th (t_1/2≈75,400 years), ²³¹Pa (t_1/2≈33,000 years), and ²²⁶Ra (t_1/2≈1,600 years). Thus, studies of disequilibria between daughter and parent isotopes ((²³⁰Th)/(²³⁸U), (²³¹Pa)/(²³⁵U), (²²⁶Ra)/(²³⁰Th)) can be used to make deductions about magma generation and evolution. Furthermore, because secular equilibrium should exist prior to melting for all the U-series daughter–parent pairs in a mantle source, regardless of enriched or depleted character, observed U-series disequilibria reflect the melting process and not simply source variations.

Although powerful for examining melting processes, U-series disequilibria alone cannot unambiguously distinguish melt generated from a homogeneous source versus that from a heterogeneous source. In particular, variations in Th/U, a ratio of two incompatible elements, could reflect mantle source or could reflect the melting process; discriminating these two alternatives requires complementary measurement of long-lived isotope systems (Sr, Nd, and Pb) indicative of the time-integrated parent–daughter ratios of the mantle sources. When used in concert with U-series isotopes, long-lived isotopic systems enable us to better distinguish between the effects of “source” variability and melting “process” (e.g., [15], [17], [25], [34], [40]).

Here, we focus attention on the genesis of melts from a location where magma chamber processes are less active, decreasing the role of magma mixing and accumulation in lessening the magma diversity of the melting column. Specifically, we investigate basalts from the Garrett Transform, which offsets the fast-spreading southern EPR at 13°30′S. We observe correlations between U-series disequilibria, trace element concentrations, and incompatible trace element ratios, which are consistent with mixing between two end-member melts. We also observe distinct ⁸⁷Sr/⁸⁶Sr isotopic ratios indicating mantle source heterogeneities. However, the link between U-series disequilibria and source heterogeneity is not clear. These observations closely parallel observations from another intra-transform domain, the Siqueiros Transform (9°N EPR), both in terms of chemical and spatial characteristics.