Crystal growth during dike injection of MOR basaltic melts: evidence from preservation of local Sr disequilibria in plagioclase

Abstract

Profiles of a total of 23 plagioclase crystals erupted within the 1982–1991 and 1993 flows of the Coaxial segment of the Juan de Fuca ridge, the 1996 flow of the North Gorda ridge, and from the Western Volcanic Zone of the ultra-slow spreading Gakkel Ridge, have been studied for variations in major and trace element concentrations. We derive equilibration times for the relatively rapidly diffusing Sr in mid-ocean ridge basalt (MORB) plagioclase crystals of the order of months to a few years in each case. All crystals preserve diffusive disequilibria of strontium and barium. Crystal residence times at MORB magmatic temperatures are thus significantly shorter, of the order of days to a few months at most, precluding prolonged crystal storage in axial magma chambers and instead pointing to rapid crystal growth (up to ~10⁻⁸ cm s⁻¹) and cooling (up to ~1°C h⁻¹) shortly prior to eruption of these samples. Growth of these crystals is therefore inferred to occur almost entirely within oceanic layer 2 during dike injection. Crystals that grew at lower crustal levels or earlier in the differentiation sequence appear to have been excluded from the erupted magmas, as might occur if most of the gabbroic rocks in oceanic layer 3 formed an interlocking crystal framework, with viscosities that are too high to carry earlier formed crystals with the melt. The vertical extent of eruptible, crystal-poor melt lenses within the gabbroic zone is constrained to ~1 m or less by considering the width of local equilibrium growth zones, equilibration times, and crystal settling velocities. This lengthscale is consistent with field evidence from ophiolites. Finally, crystal aggregates within the Gakkel ridge sample studied here are the result of synneusis within the propagating dike during melt ascent.

Appendix

Laser ablation inductively coupled mass spectrometry

Major and trace element plagioclase rim-to-rim profiles of all samples were measured at the GeoForschungsZentrum Potsdam using a GEOLAS M Pro (Coherent, Germany) laser ablation device coupled to an ELAN DRC-e (PerkinElmer SCIEX, Canada) ICP-MS in dual detector mode. The LA device consisted of an excimer laser (COMPexPRO 102, Argon Fluoride 193 nm, Lambda Physik, Germany) with a maximum output energy of 200 mJ per pulse for repetition rates between 1 and 10 Hz, an optical beam path homogenizer to provide a homogeneous laser beam, an aperture mask with 10 circular holes with diameters of 0.125–4 mm (for ablation pit diameters of 5–160 μm), a petrographic microscope (Olympus BX 51), a 20 cm³ sample cell and a computer controlled x,y,z-stage. Samples were observed through the petrographic microscope with two lenses of different magnification (5 and 20 fold). During ablation, the sample image was continuously observed on a separate high-resolution monitor through a Schwarzschild objective (25-fold magnification) and a high-resolution CCD camera. Helium was used as carrier gas (1.0 L min⁻¹), and argon as plasma (15 L min⁻¹), auxiliary (1.2 L min⁻¹), and nebulizer gas (0.7 L min⁻¹).

Elements of interest were measured using an output energy of 100 mJ, with 10 ms dwell time and 3 ms quadrupole settling time. Background intensities from the gas blank were measured for 30 s (laser firing, shutter closed) followed by acquiring transient signals of the analytes for 60 s (laser firing, shutter open) or until the laser beam had drilled through the sample and reached the sample holder (in case of thin sections). Each analytical batch consisted of up to 20 analyses. Samples were measured using a repetition rate of 5 Hz, an energy density of 6 J cm⁻² and spot sizes of 16 or 24 μm. Rim-to-rim geochemical profiles were obtained for each crystal studied by spot analyses spaced between ~20 and ~100 μm apart, depending on crystal size. A reference glass standard (NIST 610) was measured twice at the beginning and twice at the end of an analytical batch for external calibration, using a repetition rate of 10 Hz, an energy density of 10 J cm⁻², and a spot size of 32 μm.

For data reduction, the LOTUS 123 macro-based spreadsheet programme LAMTRACE was employed. A short, general description of its capabilities is given by van Achterbergh et al. (2001), and its algorithms are described in Longerich et al. (1996). The programme performs background correction, correction for instrumental drift, internal calibration, choice of integration intervals, and calculation of element concentrations using external calibration. The Ca signal was used for the calibration of trace element concentrations. Repeat analyses of gem labradorite AMNH 95557 revealed that LA-ICPMS-derived Na/Ca ratios were consistently about 11% higher than those derived by electron probe microanalysis (EPMA). This discrepancy was corrected for in the calculation of LA-ICPMS-derived anorthite contents, which was based on plagioclase stoichiometry (cf. Zellmer et al. 2003). The same repeat analyses also provide the basis for assessment of relative uncertainties for individual elements, and these are 1.5% for X_An, 1.2% for Sr, and 2.6% for Ba (1σ, n = 16, cf. supplementary Table S1 for other elements).

Electron probe microanalysis

Following the LA-ICPMS analysis, the specimens were re-polished using a vibration polisher with 0.3 μm particle alumina compounds to remove the debris of LA, and then carbon coated for microprobe analysis. Quantitative chemical analyses and, in some cases, elemental distribution (mapping) analyses were made by a field emission electron probe micro analyser (FE-EPMA: JEOL JXA-8500F) equipped with five wave-length dispersive spectrometers (WDS) at the Institute of Earth Sciences, Academia Sinica, in Taipei. Secondary- and back-scattered electron images were used to guide the analysis on target positions of minerals. A 2-μm defocused beam was operated for quantitative mineral analysis at an acceleration voltage of 12 kV with a beam current of 5 nA. Groundmass analyses were undertaken on the same instrument with a 5 μm beam diameter, 15 kV accelerating voltage, and 10 nA beam current.

The measured X-ray intensities were ZAF-corrected using the standard calibration of synthetic chemical-known standard minerals with various diffracting crystals, as follows: wollastonite for Si with TAP crystal and Ca with PET crystal, rutile for Ti (PET), corundum for Al (TAP), chromium oxide for Cr (PET), fayalite for Fe with LiF crystal, tephroite for Mn (PET), pyrope for Mg (TAP), albite for Na (TAP), and adularia for K (PET). Counting times for peak and both upper and lower baselines were 20 s and 10 s, respectively, for each element. Relative standard deviations (RSD) for Si, Na, and K were less than 1%, and others were less than 0.6%. Detection limits were less than 500 ppm for all elements.

To aid cross-correlation of EPMA and ICPMS analyses and to evaluate the general characteristics of crystal zoning, elemental distribution mapping of selected plagioclase crystals was performed with 15 kV, 20–50 nA, and 1–3 μm for the acceleration voltage, beam current and beam size, respectively. X-ray intensities of Ca-Kα, Na–Kα, and K-Kα were counted for 0.02–0.03 s in each spot, at 1- to 3-μm intervals.

Comparison of EPMA-derived and LA-ICPMS-derived anorthite contents

EPMA- and LA-ICPMS-derived anorthite contents are compared for a number of plagioclase crystals in Fig. S1. Due to the small EPMA beam size and closely spaced sampling interval, a range in X_An is obtained over the distance of 16–24 μm covered by the laser ablation spot analysis (cf. Fig. S1c). In Fig. S1a, the average and the respective range in EPMA-derived X_An is provided for each LA-ICPMS data point. Data between the two dashed lines are within the combined analytical uncertainties of each technique. The results are largely consistent between the two analytical methods, with exception of a few outliers, for which LA-ICPMS appears to slightly underestimate X_An. These outliers are more closely inspected in Fig. S1b and S1c. During the laser ablation analysis of one location from crystal W9604-C3-5, situated at the edge of a calcic growth zone, there is a clear increase in the Na signal with ablation time, suggesting that ablation progressed into an underlying, more sodic growth zone (Fig. S1b). Another example is crystal 2792-4R-2, where rim LA-ICPMS analyses again appear to underestimate X_An. Here, the outermost sodic rim contributes to the material sampled during ablation but is not sampled during the spatially restricted microprobe analysis (Fig. S1c).

The above comparison indicates that (i) LA-ICPMS-derived anorthite contents that are based on plagioclase stoichiometry are reliable, because they are generally consistent with EPMA-derived anorthite contents, and (ii) in cases where direct comparison of anorthite content with trace element concentrations from the same spot is required (as in this study), the use of LA-ICPMS-derived X_An values is preferable, because they are based on exactly the same volume of ablated material, such that artefacts introduced by compositional heterogeneities within that volume affect both major and trace elements to a similar degree. We therefore use the LA-ICPMS-derived X_An values here.