Geochemistry of impact glasses in the Chang’e-5 regolith: Constraints on impact melting and the petrogenesis of local basalt

Lunar impact glasses can provide important information on the bulk compositions of their sources and the impact history of the Moon. Here, we report the chemical composition of fifty-four clean glass spherules containing neither relict clasts nor crystals from the Chang’e-5 (CE5) regolith. They can be subdivided into three compositional groups: (1) mid-Ti basaltic (TiO2 = 4.1 ∼ 6.5 wt%), (2) low-Ti basaltic (TiO2 = 1.3 ∼ 3.9 wt%), and (3) high-Al (Al2O3 > 15 wt%). Fifty-one glasses (∼94 %) are mid-Ti basaltic, which form a loose compositional cluster for most major and trace elements. These glasses exhibit considerable variations in SiO2 (35.3 ∼ 45.3 wt%). Their TiO2, Al2O3, MgO and CaO show negative correlations with SiO2, while the Na2O, K2O and P2O5 positively correlate with SiO2, also yielding a positive correlation between the CIPW normative plagioclase and olivine. These variations likely result from differential vaporization of SiO2, strongly suggesting an impact origin of these glasses. Their major and trace element compositions are averagely similar to the bulk-rock, in turn indicating that they were formed from the local regolith. The remaining three glasses, including two low-Ti basaltic and one high-Al variety, exhibit distinct major and trace elements from the regolith, indicating an exotic source. In addition, the mid-Ti basaltic glasses provide another approach for estimating the average composition of the CE5 basalt other than directly measuring the small basalt fragments assuming that the exotic materials in the CE5 regolith were limited. This estimation reveals critical trace element characteristics of the CE5 basalt, e.g., it has higher La/Yb (3.71), Sm/Yb (1.76), Sr/Yb (31.6), and (Eu/Eu*)N (0.45) than KREEP, indicating that CE5 basalt must derive from a non-KREEP source. Chemical modeling indicates that the contribution of KREEP-rich materials in the mantle source should be less than 0.3 %. The trace element characteristics of the CE5 basalt can be reproduced by extensive (80 %) fractional crystallization after low-degree (2 %) melting. We propose that this fractional crystallization process might occur at depth, implying vast igneous underplating (7,250 ∼ 11,750 km3) beneath the CE5 landing area. This study also suggests that the high Th concentration (5.43 ppm) is an inherent property of the CE5 basalt resulting from extensive fractional crystallization. Thus, high Th detected by remote sensing may not be associated directly with a KREEP component but rather with highly fractionated basalts.