Aqueous complexation has long been considered the only viable means of transporting gold to depositional sites in hydrothermal ore-forming systems. Here, we present direct evidence supporting an alternative hypothesis, namely, the transport of gold as colloidal particles. We observed nano-scale gold particles adsorbed on halloysite and micro-scale gold particles in altered rocks by TEM and SEM in the Shanggong orogenic gold deposit. Based on this evidence, we propose a feasible model for the origin of microscopic gold particles in alteration zones. In the early stage of ore-forming fluid, gold may migrate in the form of collaurum, which is maintained by supercritical CO₂ and colloidal silica. Low salinity and high pressure are conducive to the stable migration of colloidal gold. When the physicochemical conditions change, some collaurum is precipitated and adsorbed by the clay minerals produced by hydrothermal alteration, and some collaurum undergoes growth and evolves into micro-submicrometer-sized gold particles. This study highlighted the significance of collaurum in the formation of orogenic gold deposits.

The deep integration of machine learning technology with geological prospecting has brought to the forefront a key challenge: how to construct geological-mineralization models by fusing multi-source data, select model features with guidance from metallogenic factors, build multi-source metallogenic prediction models with geological constraints, and ultimately achieve a thorough integration of domain knowledge and machine intelligence. The Eastern-Central Tianshan region is one of China’s most important copper–nickel mineral resource bases, predominantly hosting magmatic copper–nickel sulfide deposits with significant resource potential. In this context, this paper proposes a metallogenic prediction model based on multi-modal feature fusion technology. The model employs a Residual Neural Network (ResNet) incorporating a Squeeze-and-Excitation (SE) attention mechanism and a Multi-Layer Perceptron (MLP) to extract features from different modalities. It integrates multi-source data, including geochemical information, geological metallogenic factors, and aeromagnetic data. A cross-modal feature interaction module, constructed using attention weighting and a gating mechanism, enables deep fusion of the features. After training, the model achieved a prediction accuracy of 97% on the test set. Compared to a unimodal model constructed using Random Forest, the confidence and discriminative capability of the training results were significantly enhanced, validating the effectiveness of multi-modal feature fusion. Applying the trained model to the study area, a total of 11 prospective metallogenic zones were delineated. These include 4 zones in the peripheries of known deposits and 7 zones in previously unexplored (blank) areas. Notably, some known mineral occurrences fall within the predicted blank-area targets, validating the feasibility and significant value of multi-modal feature fusion in mineral prediction. This work provides a novel methodology for the subsequent integrated processing of multi-source data.

Rare earth elements (REEs) are critical mineral resources that play a pivotal role in modern technology and industry. Currently, the global supply of light rare earth elements (LREEs) remains adequate. However, the supply of heavy rare earth elements (HREEs) is associated with substantial risks due to their limited availability. Ion-adsorption type rare earth deposits (IAREDs), which represent the predominant source of HREEs, have become a focal point for exploration activities, with a notable increase in global interest in recent years. This study systematically collected stream sediments and stream water samples from the Jiaping IARED in Guangxi, as well as from adjacent granitic and carbonate background areas, to investigate the exploration significance of geochemical surveys for IAREDs. Additionally, mineralized soil layers, non-mineralized soil layers, and bedrock samples from the weathering crust of the Jiaping deposit were analyzed. The results indicate that stream sediments originating from the Jiaping IARED and granite-hosted background regions display substantially elevated REE concentrations relative to those from carbonate-hosted background areas. Moreover, δEu values in stream sediments can serve as an effective indicator for differentiating weathering products derived from granitic and carbonate lithologies. Within the mining area, three coarse-grained fractions of stream sediments (i.e., +20 mesh, 20–60 mesh, and 60–150 mesh) exhibit REE concentrations comparable to those observed in both granite-hosted and carbonate-hosted background regions. However, the HREEs content in the finer -150-mesh stream sediments from Jiaping IARED is markedly higher than that in the two background regions. The (La/Sm)_N versus (La/Yb)_N ratios of -150-mesh stream sediments in the Jiaping IARED may reflect the mixing processes involving HREE-enriched ore layer, non-mineralized layer, and LREE-enriched ore layer. This observation implies that fine-grained (-150-mesh) stream sediments can partially inherit the REE characteristics of mineralized layers within IAREDs. Scanning electron microscopy (SEM) observations indicate that the enrichment of REEs in fine-grained stream sediments primarily originates from REE-rich accessory minerals derived from parent rocks and mineralized weathering crusts. A comparative analysis reveals that the concentrations of REEs in stream water collected during the rainy season are significantly higher than those collected during the dry season. Moreover, the levels of REEs, especially HREE, in stream water from the Jiaping IARED substantially exceed those in background areas. Collectively, these findings suggest that the geochemical signatures of REEs in rainy season stream water possess diagnostic potential for identifying IAREDs. In conclusion, the integrated application of geochemical surveys of stream water and -150-mesh stream sediments can effectively delineate exploration targets for IAREDs.