Tungsten and tin fertility begins with different crustal sources

01

Why this matters

Tungsten (W) and tin (Sn) are economically and strategically important metals associated with evolved, commonly reduced granitic systems. However, some major W-Sn metallogenic provinces show a distinct spatial separation between W and Sn deposits. South China's Nanling belt is one example: Late Jurassic W deposits dominate the eastern domain, whereas most coeval Sn deposits occur in the west. Explaining this separation matters for exploration because a highly fractionated granite is not a unique indicator of either metal.

Previous models have explained this distribution through contrasting fluid behaviour, variable mantle input, or sequential low- and high-temperature melting of a broadly similar crustal source. Xia et al. (2026) instead test whether the granite magmas originated from different metasedimentary sources and whether these differences persisted through melting, emplacement and differentiation.

02

What the study found

The authors compiled whole-rock data from mineralisation-associated Mesozoic granites across Nanling and modelled fluid-absent melting of representative basement pelite and greywacke at 8 kbar.

W- and Sn-related granites show contrasting geochemical signatures. Vein-related W granites generally have low CaO/Na₂O, MgO, TiO₂ and F, but high Rb/Sr and Rb/Ba, consistent with clay-rich metapelitic sources. In contrast, Sn-related granites retain higher CaO/Na₂O, MgO, TiO₂ and F and are interpreted to derive mainly from sandstone-rich metagreywacke.

Phase-equilibrium modelling shows how source mineralogy and melting conditions influence metal fertility. Muscovite dehydration in pelite at approximately 690-750°C produces enough melt to mobilise W efficiently, generating liquids containing up to approximately 20-21 ppm W. In contrast, Sn in greywacke is hosted mainly in biotite and refractory Ti-bearing minerals. Temperatures above approximately 850°C and more extensive melting are therefore required to produce extractable Sn-rich melt containing up to approximately 18 ppm Sn. W- and Sn-rich magmas can thus form directly from contrasting sources and melting reactions, without requiring earlier extraction of a W-rich melt from the greywacke.

The W- and Sn-rich magmas also evolved differently after leaving their sources. W granites record more extensive differentiation and melt-fluid interaction than Sn granites, as indicated by lower Zr/Hf, Nb/Ta, Th/U and LREE/HREE ratios. Biotite barometry further suggests that W granites crystallised at greater pressures, commonly above 4 kbar. Xia et al. infer that the shallower Sn granites cooled more rapidly than the deeper W granites.

The proposed divergence of W and Sn during the magmatic-hydrothermal transition remains uncertain because the whole-rock trends are scattered and combine different plutons.

03

Our interpretation

The paper's strongest contribution is its separation of source fertility from later metal concentration. Metapelite and metagreywacke are not interchangeable sedimentary sources undergoing different degrees of the same melting process. They contain different mineral hosts for W and Sn, undergo different dehydration reactions and produce different melt volumes and metal ratios. Source composition establishes the direction of fertility, while melting temperature and degree determine whether enough of that metal enters an extractable melt.

The relative importance of source and fractionation is scale dependent. At the regional scale, basement composition and melting conditions most directly explain the east-west W-Sn division. Fractionation and fluid exsolution cannot generate a W- or Sn-rich system unless the relevant metal is first released effectively from the source. At district and deposit scales, however, source fertility alone is insufficient. The Triassic W granites provide the paper's clearest internal test: they share a broadly pelitic source with the more productive Late Jurassic W granites, but experienced less partial melting and weaker differentiation and are associated with much smaller deposits. Fertility must therefore be activated, transferred and focused.

Emplacement depth connects this source control with subsequent magmatic-hydrothermal evolution. Xia et al. propose that deeper W magmas cooled more slowly and underwent more extensive crystallisation and melt-fluid interaction, whereas shallower Sn magmas cooled faster and exsolved fluid earlier. However, the inferred degree of fractionation does not correlate consistently with the tonnage or grade of the regional W and Sn deposits.

A further uncertainty lies in the deeper history of the source. The metasedimentary basement formed from a broad Neoproterozoic-Ordovician succession that was metamorphosed during early Paleozoic orogenesis and remelted during Late Jurassic extension. A flysch or turbidite system could plausibly contain mud-rich and greywacke-rich domains, but the paper does not reconstruct basin architecture, depositional ages or provenance. The eastern and western sources may therefore represent different facies, stratigraphic packages or assembled terranes. This uncertainty does not invalidate the melting model, but the ultimate origin of the source heterogeneity remains unresolved.