The lunar basalt samples returned by the Chang’e-5 mission erupted about 2.0 billion years ago during the late period of the Moon’s secular cooling. The conditions of mantle melting in the source region and the migration of magma through the thick lithosphere that led to this relatively late lunar volcanism remain open questions. Here we combine quantitative textural analyses of Chang’e-5 basaltic clasts, diffusion chronometry, clinopyroxene geothermobarometers and crystallization simulations to establish a holistic picture of the dynamic magmatic–thermal evolution of these young lunar basalts. We find that the Chang’e-5 basalts originated from an olivine-bearing pyroxenite mantle source (10–13 kbar or 250 ± 50 km; 1,350 ± 50 °C), similar to Apollo 12 low-Ti basalts. We propose these magmas then ascended through the plumbing system and accumulated mainly at the top of the lithospheric mantle (~2–5 kbar or 40–100 km, 1,150 ± 50 °C), where they stalled at least several hundred days and evolved via high-degree fractional crystallization. Finally, the remaining evolved melts erupted rapidly onto the surface over several days. Our magmatic–thermal evolution model indicates abundant low-solidus pyroxenites in the mantle source with a slightly enhanced inventory of radioactive elements can explain the prolonged, but declining, lunar volcanism up to about 2 billion years ago and beyond.

嫦娥五号任务返回的月球玄武岩样本在大约 20 亿年前月球长期冷却的晚期爆发。源区地幔熔化的条件以及导致这种相对较晚的月球火山活动的岩浆通过厚岩石圈的迁移仍然是悬而未决的问题。在这里，我们结合嫦娥五号玄武岩碎屑的定量结构分析、扩散计时法、单斜辉石地质温度计和结晶模拟，建立了这些年轻月球玄武岩动态岩浆-热演化的整体图景。我们发现嫦娥五号玄武岩起源于含橄榄石辉石岩地幔源（10–13 kbar 或 250 ± 50 km；1,350 ± 50 °C），类似于阿波罗 12 号低钛玄武岩。我们建议这些岩浆然后通过管道系统上升并主要聚集在岩石圈地幔的顶部（~2-5 kbar 或 40-100 公里，1,150 ± 50 °C），在那里它们至少停滞了数百天并通过高度分级结晶。最后，剩余的演化熔体在几天内迅速喷发到地表。我们的岩浆-热演化模型表明，地幔源中含有丰富的低固相辉石岩，放射性元素存量略有增加，这可以解释长达约 20 亿年前及以后的月球火山活动持续但下降的现象。剩余的演化熔体在几天内迅速喷发到地表。我们的岩浆-热演化模型表明，地幔源中含有丰富的低固相辉石岩，放射性元素存量略有增加，这可以解释长达约 20 亿年前及以后的月球火山活动持续但下降的现象。剩余的演化熔体在几天内迅速喷发到地表。我们的岩浆-热演化模型表明，地幔源中含有丰富的低固相辉石岩，放射性元素存量略有增加，这可以解释长达约 20 亿年前及以后的月球火山活动持续但下降的现象。