Ar-Ar Geochronology for basalts dredged from seamounts in the south Tasman Sea during RV Investigator voyage IN2018_V08

Full description

This dataset presents Ar-Ar geochronological ages for mafic dredged rock samples from seamounts in the southern Tasman Sea and l’Atalante Basin. Rock samples were dredged during RV Investigator voyage IN2018_V08 in 2019, and the AGSO96 voyage in 1996. We interpreted recovered mafic samples as seamount edifice material. Representative sections of the freshest whole rock mafic samples, least affected by post-magmatic alteration, were selected for geochronology and geochemical analyses.
We selected samples based on thin section examination for 40Ar-39Ar incremental heating experiments. All basaltic rocks dredged from the seafloor exhibit some degree of low temperature alteration of primary minerals and groundmass to secondary minerals, such as clay and zeolite.
The data contains the sample id, seamount name, latitude, longitude, material analysed, age and error, %39Ar, MSWD, comment on the quality of the sample (green highlighted are the ages selected to contribute to a final seamount age), final seamount age.

Lineage

Progress Code: completed

Statement: Interpretation of 40Ar-39Ar age information from dredged samples from the seafloor is usually complicated by (1) the aphryric to glassy textures and low-K compositions typical of ocean basalts, (2) low-temperature alteration of groundmass and phenocrysts to clay and zeolite, and (3) nuclear reactor-induced effects. In ideal circumstances, incremental heating experiments on multi-phase, well-crystallized, unaltered basalts can produce (statistically) identical ages from the Ar-isotopic composition of gas, in sequential steps, released by increasing temperature-activated diffusion. The analytically concordant step ages are combined as a weighted mean (by inverse variance) to calculate a plateau age for the crystallization of the sample (York, 1978). The Ar-isotopic compositions of the gas steps, plotted either as 40Ar/36Ar vs 40Ar/39Ar or 36Ar/40Ar vs 39Ar/40Ar, will be linearly correlated with slope corresponding to the sample age and 40Ar/36Ar (or 36Ar/40Ar) intercept the initial (trapped) composition of Ar in the sample (at crystallization). These isochron ages are concordant with the plateau age. The fast neutron production of isotopes of Ar in experimental nuclear reactors is the basis for the 40Ar-39Ar incremental heating method (McDougall and Harrison, 1988). However, neutron capture is sufficiently energetic that atoms can move within crystals, from one crystal to another, and be lost from crystal margins (Onstott et al., 1995). This recoil is especially significant for the distribution of 39Ar (derived from 39K) and 37Ar (derived from 40Ca) in dating groundmass separates. Recoil movement of 39Ar from relatively K-rich sites to relatively K-poor sites will increase or decrease, respectively, 40Ar/39Ar (equivalent to measured age) subsequently degassed from those sites during step heating. Glassy or poorly crystallized matrix releases Ar gas at lower heating temperatures than plagioclase or clinopyroxene phenocrysts. The combined effects result in recoil-affected disturbance in age spectra characterized by older step ages at low heating temperatures, and younger step ages at higher heating temperatures. There may be preserved a plateau of step ages within the middle of the spectrum, or the whole spectrum may record an ever-decreasing (‘inverse staircase’), depending on distribution of K and grain size. Recoil of 37Ar operates in a similar way, except on the distribution of Ca, and affects the correction of 36Ar abundance. In cases where Ar-recoil dominates the age spectra, the best age estimate may be the integrated (total fusion) age. All basaltic rocks dredged from the seafloor exhibit some degree of low temperature alteration of primary minerals and groundmass to secondary minerals, such as clay and zeolite. Ar gas that accumulates from radioactive decay of K progressively diffuses from the mineral grain edges during alteration to more loosely bound structures (or seawater), leading to Ar-loss. Age spectra for which Ar-loss is a significant factor commonly produce low temperature heating step ages younger than the crystallization age, increasing up to the plateau age. This is because secondary minerals diffuse Ar at lower heating steps than primary minerals. This effect can be reduced by pre-heating the sample at low laser power to remove Ar that is loosely held on grain surfaces and in secondary minerals. In fine-grained, multi-phase samples (groundmass or whole rock), Ar-loss and Ar-recoil effects can be superimposed, leading to age spectra that, with increasing temperature, start with high step ages that decrease rapidly, then rise up to a plateau before showing high temperature Ar-recoil (Fig.XX). Practitioners have proposed various criteria for interpreting age data from 40Ar-39Ar incremental heating experiments on altered basalts that produce disturbed spectra (Fleck et al., 1977; Jourdan et al., 2004; Sharp and Renne, 2005; Schaen et al., 2020). The goal is to extract a reliable crystallization age for the sample, which hinges on identifying steps that form a plateau age and also fit an isochron that gives a concordant age, and an intercept that provides the sample 40Ar/36Ar composition at the time of crystallization. Criteria include a minimum number of contiguous, plateau-forming step ages that account for a minimum fraction of the total gas released. In addition, statistical analysis must show that the step ages included in the plateau are from the same normal distribution; that is, variation in step ages is what is expected from the analytical uncertainties in those ages. Similar statistical treatment applies to the isochron age. Analytical improvements have led to high-sensitivity, low-volume, multi-collector mass spectrometers that offer the capability of much larger numbers of high-precision heating step ages on smaller, better prepared sample aliquots. Experiments of typically 30-50 heating steps now show gas release details previously undetected in experiments of 6-10 heating steps with poorer age precision. Hence, some samples that previously appeared to show acceptable plateaus can now exhibit subtle but significant Ar-recoil induced decreasing step ages across the ‘plateau’ heating interval. Higher precision step ages have also revealed that some well-behaved seafloor basalts produce non-atmospheric 40Ar/36Ar intercepts. This is due to the glassy nature of quenched mesostasis in many dredged rocks, which is known to trap mantle-derived (‘excess’) Ar at crystallization (Dalrymple and Lanphere, 1969). Trapped 40Ar/36Ar less than atmospheric occurs in vesicular rocks owing to Ar isotopic fractionation (XX). In these cases it is reasonable to use the measured 40Ar/36Ar intercept as the initial composition, to recalculate step ages and compositions (Heaton and Koppers, 2019). This methodology generally expands the number of step ages included in the plateau and isochron. Schaen et al. (2020) have updated criteria for reliability of plateaus as crystallization ages, appropriate for higher resolution multi-collector instruments: plateaus (1) should consist of 5 or more consecutive, concordant step ages consisting of greater than 50% of the total 39Ar released, (2) should not have a slope of increasing or decreasing step ages, and (3) have an isochron regressed through all steps with a 40Ar/36Ar intercept within analytical error of the atmospheric value. Samples that do not meet all the criteria for a most reliable crystallization age may still provide useful age information. Heaton and Koppers (2019) have proposed a high-moderate-low quality age rating scheme based on fraction of the total gas released forming the plateau (from greater than 50-30%), statistical confidence level of plateau and isochron fits (1 less than MSWD less than 2), and a spreading factor that assesses the step compositions to be at least 2% to more than 10% of the range between initial and radiogenic Ar (Jourdan et al., 2009). For other than high quality ages, reports of ‘age estimates’ are often given, meaning that geological and analytical disturbances have led to less confidence that the interpreted plateau or isochron age are crystallization ages. Often, other information such as multiple ages from the same dredge or nearby location, or in the case of drillcore, stratigraphic position can be used to assess the reliability of lower quality ages. Age analyses of mineral separates (plagioclase, clinopyroxene) from the same sample are particularly valuable in distinguishing age information from artifacts. Because these are single phases (generally uniform composition) they do not produce significant Ar-recoil effects, and because they are more resistant to alteration than the rock groundmass, they show less Ar-loss. Many of our dredged samples are aphyric to sparsely phyric, restricting available material to groundmass, but for several we are able to analyze both groundmass and plagioclase (and, rarely, clinopyroxene or hornblende).