Geochemical comparison of four cores from the Manson impact structure

Concentrations of 33 elements were determined in relatively unaltered, matrixrich samples of impact breccia at ∼3-m-depth intervals in the M-1 core from the Manson impact structure, Iowa. In addition, 46 matrix-rich samples from visibly altered regions of the M-7, M-8, and M-10 cores were studied, along with 42 small clasts from all four cores. Major element compositions were determined for a subset of impact breccias from the M-1 core, including matrix-rich impact-melt breccia. Major-and trace-element compositions were also determined for a suite of likely target rocks. In the M-1 core, different breccia units identified from lithologic examination of cores are compositionally distinct. There is a sharp compositional discontinuity at the boundary between the Keweenawan-shale-clast breccia and the underlying unit of impact-melt breccia (IMB) for most elements, suggesting minimal physical mixing between the two units during emplacement. Samples from the 40-m-thick IMB (M-1) are all similar to each other in composition, although there are slight increases in concentration with depth for those elements that have high concentrations in the underlying fragmental-matrix suevite breccia (SB) (e.g., Na, Ca, Fe, Sc), presumably as a result of greater clast proportions at the bottom margin of the unit of impact-melt breccia. The high degree of compositional similarity we observe in the impact-melt breccias supports the interpretation that the matrix of this unit represents impact melt. That our analyses show such compositional similarity results in part from our technique for sampling these breccias: for each sample we analyzed a few small fragments (total mass: ∼200 mg) selected to be relatively free of large clasts and visible signs of alteration instead of subsamples of powders prepared from a large mass of breccia. The mean composition of the matrix-rich part of impact-melt breccia from the M-1 core can be modeled as a mixture of approximately 35% shale and siltstone (Proterozoic "Red Clastics"), 23% granite, 40% hornblende-biotite gneiss, and a small component (<2%) of mafic-dike rocks. The SB unit is significantly different and more variable in composition than the IMB unit. Compared to the IMB in M-1, the SB is depleted in alkalis (K, Rb, Cs) heavy rare-earth elements (Yb, Lu), and high field strength elements (Ta, Th, U). For many elements, the "transition zone" (TZ) between the IMB and the SB is intermediate in composition, corresponding roughly to a 2:1 mixture of SB and IMB. There is no compositional discontinuity at the TZ-SB boundary, but concentrations of some elements (e.g., Ca, Zn, and Cs) decrease discontinuously at the IMB-TZ boundary. We interpret the M-1 TZ as a physical mixing zone, perhaps formed by adjustment between the IMB and SB during uplift of the central peak. Matrix-rich samples from IMB intervals in the other cores are generally similar in composition to those of the M-1 core; however, small but significant differences exist. Impact-melt breccias from the other cores are more variable in composition, in part because they have greater proportions of clasts. If the melt composition is the same for all cores, then the compositional differences among impact-melt breccias of different cores are caused largely by differences in the composition or abundance of clasts (plus some post-impact alteration effects) and the average composition of the clasts differs from that of the melt; this is most evident in M-7. Concentrations of Cs and Rb are significantly greater in the nonporous, relatively clast-poor IMB of M-1 than in the porous, clast-bearing or clast-rich IMB of the other cores. Concentrations of alkalis are also greater in the center of the IMB unit of M-1 than at the top and bottom of that unit. These observations suggest that alkalis were lost to hydrothermal fluids at the edges of the IMB unit of M-1 as well as from more porous IMB units elsewhere. The compositional characteristics of the M-1 TZ coupled with the fact that the TZ breccias have fragmental, porous matrices compared to the fairly impermeable overlying IMB suggest that the M-1 TZ may have been preferential for the flow of post-impact hydrothermal fluids. The different thicknesses of IMB units and variable mixing of IMB and SB indicated in the M-1, M-7, M-8, and M-10 cores reflect locally irregular thinning and adjustment of these units during central-peak uplift.