7131. Two-stage vacuum pumps for high working pressures and high water vapor tolerances: W Kaiser,J Vac Sci Technol, A8, 1990, 2761–2763

Abstract

Impact melt and country rocks from the 23 km diameter Lake St. Martin structure in Manitoba were analysed for major and trace element abundances and for Rb-Sr isotopic compositions. Our aim was to better understand the formation of the melt rocks in this probable impact structure, to search for traces of the meteoritic projectile, to date the melt sheet in order to improve the chronological record of impact events on the Earth's surface, and to constrain the target, from which the melt rocks could have formed.Major element analyses are consistent with the observation by Simonds and McGee (1979) that many of the melt rocks from Lake St. Martin are chemically homogenized, to a degree similar to that of melt bodies from other impact structures. However, some samples are of slightly different composition. This can be explained by post-impact hydrothermal alteration (particularly near the present erosion surface), chemical exchange between melt-bearing breccia and basement rock near the bottom of the melt sheet, or limited mixing of different target lithologies. So-called “pseudotachylites” from the central uplift of the structure are chemically diverse and do not always represent the composition of their host rock. We conclude that such samples are not pseudotachylite but probably represent injected impact breccia.REEs and other lithophile elements are well homogenized in the melt in comparison with the country rock variation. There is a slight enrichment of the siderophile elements Co and Au and of Cr in the melt rocks, but further analyses are required to verify whether this is caused by a small contribution from the meteoritic projectile. The melt rocks are also homogenized isotopically, when compared with the spread of isotopic ratios determined for the country rocks. The melt rock isotopic composition, however, can only be achieved by mixing of various granitoid country rock types with another component of high 87Rb86Sr ratio. Furthermore, the melt rocks are isotopically reequilibrated: whole rock Rb-Sr isotope data yield an age of 212.4 ± 43 Ma (ISr = .7132 ∓ 3) at 2σ level. Whole rock and mineral separate data yield an age of T = 219 ± 32 Ma (ISr = .7131 ∓ 3). This age, consistent with previously reported K-Ar ages of 200 ± 25 and 250 ± 25 Ma, is thought to be the best approximation to the age of the melt rocks and thus of the impact event. However, the 2σ error of 32 Ma is too high to allow this age to become a significant addition to the data base of well-constrained impact ages for analysis of periodicity in the cratering record. An “age” of T = 2785.3 ± 157 Ma (2σ) for ISr = .6997 ∓ 11 was obtained for the basement rocks of the Lake St. Martin structure, but the true meaning of this “age” is still unclear. While the age obtained is comparable to other Superior Province data, it appears that the Lake St. Martin country rock data define a mixing line.The new chemical and isotopic results are consistent with an impact origin for the Lake St. Martin structure. Harmonic least-squares mixing calculations suggest that the melt rocks were formed from ca. 80% monzonitic and 15% granitic basement plus a small addition of ca. 6% carbonate rock. This is in good agreement with the results by Simonds and McGee (1979) and favors their suggestion that carbonaterich target rocks, even when they form the upper portion of the target stratigraphy, do not significantly take part in impact melt formation but are largely removed in a CO2-rich vapor phase.