Discussion
The primary difficulty in 40Ar/39Ar analyses of very fine-grained materials, such as pseudotachylyte and ultramylonitic biotite, is the high probability of incorporating undesirable minerals or cataclastic fragments. Laser beam diameters of 30-42µm enabled the ablation of a spot less than 50µm in width, which was either rastered over a square region or tracked along predefined lines. This flexibility and ability to predetermine the path of the laser made avoiding optically-visible foreign materials fairly straightforward. The advantages of this technique are apparent in the close precision among individual analyses of a given sample, and this precision is reflected in small 2σ error constraints on weighted mean ages (Table 1). However, the size distribution of cataclastic fragments within pseudotachylyte veins can include submicroscopic fragments (Figure 2c). Many of the analyzed raster squares likely contained a few submicroscopic clasts, which produced significant errors in some of the individual analyses (e.g. analyses 6, 7, 15 of sample 9819, Table AP2).
The fine-grained matrix of the ultramylonite samples is composed primarily of biotite but also contains plagioclase and occasional quartz grains (Figure 2d). Although the laser beam was tracked through predominantly biotite regions, the heterogeneous, interlocking fabric of the matrix meant that incorporation of some plagioclase±quartz was unavoidable. Other complications included occasional mistracking of the laser when it reached a turning node in the predetermined path, which sometimes resulted in the partial ablation of undesirable minerals. This is reflected in large errors on some of the individual ultramylonite analyses (e.g. analyses 1, 6, 7 of sample 006, Table AP1)).
The use of a small diameter laser beam for spot fusion makes it difficult to generate Ar gas in sufficient volumes to achieve a high signal-to-noise ratio (Roberts et al., 2001). The very fine-grained target materials in these samples necessitated the use of 30-42µm beams, which required high current intensities (28-30kV) for long durations (8-12 minutes) to enable the accumulation of sufficient Ar gas for robust analyses. Analyses that fell below a threshold of 70% 40Ar lacked sufficient gas for robust results and therefore were discarded (see asterisk-marked analyses, (Table AP1 and Table AP2). Low gas yields on several of the individual analyses above the 70% threshold still produced significant errors greater than ±30 Ma (e.g. analyses 004: 1; 006: 1, 6, 7; and 9819: 6, 7, 12, 15; Table AP1 and Table AP2). These low yields likely resulted from a poor choice of analysis area and/or occasional mistracking of the laser, which allowed undesirable clasts or minerals to be partially ablated.
Nuclear recoil during sample irradiation can result in the transfer of 39Ar from K-rich minerals to adjacent K-poor minerals (Turner and Cadogan, 1974, Onstott et al., 1995), which would artificially increase the apparent ages of analyzed K-rich minerals. Recoil distances have been estimated at approximately 0.08µm across a mineral boundary (Villa, 1997), and thus can be a factor for concern when dating fine-grained minerals such as the 10-20µm biotite in the ultramylonite samples. The possible recoil influence was minimized in these analyses by using an excimer (uv) laser to target the centers of biotite-rich areas, and avoid marginal regions which are more susceptible to 39Ar loss (e.g. Roberts et al., 2001). In the ultramylonite samples the matrix is predominately biotite, so most 39Ar transfer due to recoil would simply exchange 39Ar between adjacent biotite grains, and therefore result in little net loss. A small percentage of K-poor minerals, such as plagioclase±quartz, occur within the ultramylonitic matrix and may have absorbed a fraction of the recoil-produced 39Ar, thus yielding artificially older apparent ages. However, the generally good agreement among individual analyses within each sample suggests that nuclear recoil was not a significant factor in the weighted mean age results.