Lower crust-derived granitic rocks provide constraints on the crustal reworking process and consequently give hints on the destruction mechanism of the cratons. The North China Craton (NCC) underwent extensive crustal melting in the Mesozoic. This study investigated granitic intrusions in the Dazeshan region of the Jiaodong Peninsula. Whole-rock major and trace element analyses and zircon U-Pb ages coupled with Hf isotopes were used to reveal the crustal reworking processes. Zircons separated from the quartz porphyry, Linglong granite, rhyolite porphyry, and biotite granite showed weighted mean ²⁰⁶Pb-²³⁸U ages of 119.2 ± 1.0 Ma, 140.2 ± 1.0 Ma, 120.6 ± 0.5 Ma, and 119.9 ± 0.7 Ma, respectively. The quartz porphyry, rhyolite porphyry, and biotite granite had high silica contents (SiO₂ = 74–77 wt.%) but low MgO, Co, and Ni concentrations. The calculated ε_Hf(t) values for the rhyolite porphyry and the biotite granite ranged from −18.3 to −20.0 and −17.8 to −20.2, respectively. These geochemical features imply ancient crust sources. The quartz porphyry showed distinct primitive mantle-normalized rare earth element (REE) patterns and was characterized by lower ΣREE content and lack of pronounced negative Eu anomalies. Whole-rock and zircon Dy/Yb ratios showed no correlation with whole-rock Rb/Sr ratios and zircon Hf contents, reflecting limited effects of crystal fractionation. The Ba/La ratios were also high (>150), but the Sr/Y and La/Yb ratios were low (Sr/Y < 50; La/Yb < 15). These features likely indicate that the quartz porphyry was generated by water-fluxed melting without differentiation. The rhyolite porphyry and biotite granite shared many geochemical similarities, denoting a unified source. The high La/Yb (>30) but low Sr/Y (<20) ratios and apparent negative Eu anomalies indicated plagioclase fractionation. Decreased zircon Dy/Yb with increasing Hf concentrations reflected noticeable amphibole fractionation. These two suites had fairly low Ba/La ratios. These data together point toward an identical source: dehydration melting of a relatively thickened crust. These melts experienced crystal fractionation after extraction. We propose that the intrusions were generated by the underplating of water-rich mafic magma, which provided both fluid and heat and finally induced two kinds of melting.

The North China Craton (NCC) has thick lithosphere in the Paleozoic (>200 km) but appears to be decratonized in the Mesozoic. However, the actual processes operated in the lower crust-deep mantle are still unclear. The Mesozoic granitic rocks can provide important clues to the NCC decratonization mechanism. Here, we conducted trace element modeling to check whether partial melting of the Archean lower crust can generate these Mesozoic magmatic suites. Meanwhile, zircon Hf isotope analysis was conducted to reveal crust-mantle interaction processes and further give constraints on the decratonization of the NCC. Zircon Hf isotope data of the Linglong, Guojialing, and Aishan suites, the mafic microgranular enclaves (MMEs) in the Guojialing suite, and mafic dykes display minor differences: the Linglong (160–150 Ma), Guojialing (~130 Ma), and Aishan (118–116 Ma) suites have zircon $εHft=–25.4$ to –14.5, –15.3 to –10.4, and –23.1 to –11.9, respectively. The Cretaceous mafic dyke (126 Ma) has a highly negative $εHft$ value (–22.8 to –17.7). Meanwhile, the MMEs (in the Guojialing granodiorite, DCW-2A, 129 Ma) have zircon $εHft=–13.0$ to –8.9. Temperature-pressure conditions calculated using amphibole compositions for both the Guojialing granodiorite and its MMEs are basically identical, implying possible magma mixing. Our modeling results show that certain trace elements (e.g., Tb, Yb, and Y) have to be retained in the source to match the composition of the Linglong suite, which requires substantial garnet residues (high-pressure melting) in the Jurassic. The Early Cretaceous garnet-dominated lower crust is Yb-/Y-enriched but depleted in elements like Sr and La. Therefore, it could not form geochemical features like high Sr/Y and La/Yb ratios akin to the Guojialing suite. Integrating the modeling results and zircon Hf isotope data, we propose that the crust in the eastern NCC had thickened and partially melted by dehydration to produce an eclogitic residue containing a large amount of garnet (>50% by weight) during the Jurassic (Linglong granite), whereas upwelling of hot and hydrous mafic magma from the asthenospheric mantle induced fluxed melting of both the lower crust and lithospheric mantle in the Early Cretaceous, during which the lithospheric mantle and part of the lower crust in the Jiaodong were removed by the convective mantle. About 10 Mys later while the Aishan suite formed, the crust was not thick anymore, and melting occurred under moderate pressure which does not necessarily require abundant garnet as the residue phase.

The North China Craton underwent extensive and widespread crustal reworking (or decratonization) during the Mesozoic. However, how the decratonization operated is not well understood. Zircon compositions are widely used by the scientific community to reconstruct crustal thicknesses. In this study, we sampled 13 magmatic rocks in the Jiaodong Peninsula and used zircon Eu/Eu* to constrain the crustal thickness of the Jiaodong area and reveal decratonization processes in the Mesozoic time. The reconstructed crustal thickness using zircon Eu/Eu* is approximately 70 km in the Jurassic, and this value is 89 km at around 130 Ma, after which the crustal thickness drops to 30–40 km at ca. 110 Ma. These results are generally compatible with or slightly higher than the calculation results using a whole-rock La/Yb proxy for the Jurassic and ∼130 Ma rocks. Crustal thickness estimated using a whole-rock La/Yb proxy for the ∼110 Ma rocks is thicker than 70 km, which is not consistent with the geological facts and the result given by zircon proxy. The whole-rock proxy failed in estimating crustal thickness because of amphibole fractionation for the ∼110 Ma rocks. The crustal thickening from Jurassic to ∼130 Ma was probably related to the westward subduction of the Paleo-Pacific slab. The thinning of the crust from 130 to 110 Ma is not a rapid process but occurs more slowly than expected, which might be explained by the chemical erosion process rather than a mechanical delamination model. The chemical erosion was most likely induced by a rollback of the subducting slab and an upwelling of the asthenosphere.