Transient tin mineralization from cooling of magmatic fluids in a long-lived system

Hydrothermal deposits derived from felsic magmatic reservoirs are primary sources of metals (e.g., Cu, Mo, Au, Ag, Sn, W) for an industrial society (Lehmann, 2021; Park et al., 2021). The growing demand for raw materials and the transition to an environmentally sustainable future require an improved ability to predict new deposits and environment-friendly mining techniques. A key knowledge gap is the fertility and longevity of hydrothermal systems. For example, fluid exsolution is inevitable during the cooling and crystallization of felsic magmas, but only in rare cases can these fluids form economically valuable deposits. As such, decoding the controls of fluid fertility is a top priority for shaping genetic models of ore formation and guiding mineral exploration. Additionally, while the longevity of hydrothermal systems now is well established within relatively long periods of at least a few to tens of thousands of years (Buret et al., 2016; Li et al., 2017), the effective time of mineralization (i.e., the time window when fluids can precipitate ore minerals) remains unknown. This not only hinders the accurate quantification of ore-forming rates and efficiency but also brings challenges for directly extracting metals from metalliferous fluids.

Addressing the aforementioned knowledge gap calls for an understanding of the evolutional trajectories of ore-forming hydrothermal fluids. This knowledge also provides insights into how the evolution of magmatic reservoirs regulates the transfer of ore-forming components from melts to volatiles and how fluids escape the reservoirs (Chelle-Michou et al., 2017; Parmigiani et al., 2017; Troch et al., 2022). Paragenetically constrained fluid inclusion assemblages are a key source of information on fluid evolution (Pettke et al., 2012), but quantifying the timing of fluids with high resolution via radiometric dating is challenging (Li et al., 2018). Another important source of information is the oxygen isotope composition (δ¹⁸O) of gangue minerals such as quartz, garnet, and tourmaline (D’Errico et al., 2012; Fekete et al., 2016; Harlaux et al., 2021; Li et al., 2022a), which can offer great detail when texture-controlled information is translated to temporal patterns using in situ techniques such as ion probe. However, assuming the co-precipitation of gangue and ore minerals and converting mineral δ¹⁸O to fluid δ¹⁸O bring additional complexities and uncertainties (e.g., minerals not co-precipitated, and the accuracy of temperature estimates can vary; Carr et al., 2017; Li et al., 2022b).

We simultaneously obtained δ¹⁸O values of paragenetically constrained gangue and ore minerals (quartz and cassiterite) for the first time using ion probe, which permitted the reconstruction of the thermal and isotopic history of ore-forming fluids with high temporal resolution. We then quantified the timing of fluid dynamics using aluminum diffusion in quartz chronometry and defined the effective time of mineralization. The data are used in discussion of the potential for extracting metals from metalliferous fluids.

Weilasituo (Inner Mongolia, China) is a veintype tin deposit hosted by Paleoproterozoic biotite-plagioclase gneiss and genetically linked to a ca. 138 Ma quartz porphyry. An extensive study of quartz-cassiterite veins (n = 14) revealed a relatively simple mineral paragenesis (using the samples of Han et al. [2022]), with three generations of quartz and one generation of cassiterite (Figs. 1A–1E). A core of stage 1 quartz (Q1 core) is enclosed by cassiterite and Q1 mantle and followed by Q1 rim (Fig. 1B). Textural evidence suggests that cassiterite precipitation started at the end of Q1 core precipitation and ceased at the beginning of Q1 rim precipitation. Stage 2 quartz (Q2) cuts Q1 and cassiterite as veinlets (Figs. 1C and 1D). Stage 3 quartz (Q3) encloses Q1 as shells (Figs. 1C and 1D). Given the identical paragenesis and mineral growth history for all 14 investigated veins, two representative vein samples were studied in detail for single fluid evolution analysis (Han et al., 2022) and in situ oxygen isotope analysis (this study).

Quartz and cassiterite crystals were analyzed on a CAMECA IMS 1280HR secondary ion mass spectrometer (SIMS) at the Beijing Research Institute of Uranium Geology with instrumental mass fractionation being corrected using matrix-matched reference materials (Tang et al., 2020; Li et al., 2022b). Aluminum in quartz was measured on a CAMECA NanoSIMS 50L at the Institute of Geology and Geophysics, Chinese Academy of Sciences (Beijing), using an imaging approach (Hao et al., 2016). To improve the accuracy and precision when measuring Al in quartz at high spatial resolution (i.e., ~0.1 μm), we reported relative Al abundances (i.e., Al/Si) assuming a consistent Si content in quartz at the micrometer scale. All data are presented in the Supplemental Material¹.

We modeled Al profiles (n = 9) as an initial step function modified by diffusional relaxation; hence, derived time scales represent maximum values (i.e., when the initial step function is unsatisfied). The growth time for core, mantle, and rim of Q1 was denoted as t1, t2, and t3, respectively. For core and mantle of Q1, our modeling considered their total diffusion time (t1 + t2 + t3, t2 + t3) following the thermal history in Figure 2A. Time scale and uncertainties from all sources were numerically calculated using the Diffuser software (Wu et al., 2022).

Magmatic quartz of the ca. 138 Ma causative quartz porphyry has a δ¹⁸O value of 10.4‰ ± 0.3‰, corresponding to a δ¹⁸O value of 8.6‰ ± 0.4‰ for primary magmatic fluids at 600 ± 50 °C (e.g., Fonseca Teixeira et al., 2022, and references therein). For hydrothermal quartz, Q1 core has homogeneous δ¹⁸O values of 11.4‰ ± 0.4‰ (Figs. 2A and 3). Q1 mantle shows a rimward increase in δ¹⁸O values from ~11.4‰ to 13.5‰, and co-precipitated cassiterite has a homogeneous δ¹⁸O value of 3.6‰ ± 0.4‰. Q1 rim has a homogenous δ¹⁸O value of 13.7‰ ± 0.4‰. δ¹⁸O values of Q2 fluctuate between 12.7‰ and 15.3‰ (Fig. 3) without obvious spatial trends. δ¹⁸O values of Q3 gradually decrease from 11.8‰ to 9.5‰ (Fig. 3). Following the cooling history in Figure 2B and using diffusivity of Tailby et al. (2018), Al diffusion in quartz chronometry yields growth time scales of ~50 k.y., 2–10 k.y., and 140–240 k.y. for the core, mantle, and rim of Q1, respectively (Fig. 4).

The rimward increase of δ¹⁸O values for Q1 mantle and the consistent δ¹⁸O values for coprecipitated cassiterite (Fig. 1) suggest that Q1 mantle and cassiterite both were precipitated from fluids with a δ¹⁸O value of 8.6‰ ± 0.4‰ (Fig. 3), with temperatures progressively decreasing from ~500 °C to ~390 °C (change in temperature ΔT = 110 °C; Fig. 2A). Because Q1 core formed immediately before cassiterite and Q1 mantle (Fig. 1E), its formation temperature was assumed to be the same as that at the beginning of the formation of Q1 mantle (i.e., ~500 °C). In a similar fashion, Q1 rim was assumed to have formed at ~390 °C. This tightly constrained cooling path along the isoδ18O path of cassiterite (Fig. 2A) with high temporal resolution is possible because of (1) the sensitive response of quartz-cassiterite δ¹⁸O values to temperature (Li et al., 2022b), and (2) the reading of mineral chemistry at high spatial resolution (i.e., 10 μm) by SIMS.

The thermal and isotopic history (Figs. 2 and 3) lend strong support to the interpretation that tin precipitation was triggered by cooling of pure magmatic fluids; mixing with external fluids and/or water-rock interaction are not evidenced. This is supported by single fluid inclusion microanalysis on the same sample set (Han et al., 2022), where homogenization temperatures gradually cooled from >450 °C to <350 °C with constant salinity. Recent cassiterite oxygen isotope studies (Li et al., 2022b, 2022c) also argue for negligible involvement of meteoric water during tin deposition. To summarize, we suggest that cooling was a primary trigger for tin mineralization at Weilasituo.

The formation temperatures of Q2 and Q3 were loosely constrained by homogenization temperatures of fluid inclusion assemblages (Han et al., 2022). Q2 marks the transition from lithostatic to hydrostatic conditions, with fluid δ¹⁸O values fluctuating between 4.9‰ and 7.1‰ at ~280 °C (Fig. 3), which indicates minor involvement of meteoric water. A further incursion of meteoric water was recorded by Q3, and the corresponding fluid δ¹⁸O values drop significantly from ~5.9‰ to <0‰ at ~230 °C.

The bulk tin at Weilasituo was deposited within 2 to 10 k.y., within a much longer time window (~50 k.y. plus 140–240 k.y.) when only gangue minerals were formed (Fig. 4). Hence only <5% of the hydrothermal system’s lifetime was effective for mineralization (Fig. 3), or even shorter if the growth time of Q2 and Q3 are considered. Therefore, ore-forming rates and efficiency might be much faster than previously thought and these require significant revision.

Much longer time scales (~3000 m.y., > 150 m.y., and >200,000 m.y. for core, mantle, and rim of Q1, respectively) were obtained when using the diffusivity of Jollands et al. (2020). These time scales are geologically implausible and could be attributed to (1) significant underestimation of formation temperatures; (2) the concentration profile prior to diffusion not being a step function (Till et al., 2015); and/or (3) the experiment of Jollands et al. (2020) being not directly applicable to hydrothermal quartz, e.g., due to coupled diffusion of Al with other elements (e.g., Li, Na, H) in a similar fashion to Li in zircon (Sliwinski et al., 2018). The thermal history defined by quartz-cassiterite oxygen isotope thermometry is 20–50 °C higher than that given by fluid inclusion assemblages (Han et al., 2022); hence, it is unlikely that our temperature estimates are underestimated. The extremely low Ti concentrations (<<1 ppm) in our quartz hinder the application of Ti diffusion in quartz as an independent test for the initial boundary condition, but it is unlikely that the initial step function has been extensively modified. If the diffusivity and/or the diffusion mechanism of Al were revised in future experimental work, then our results might need to be reevaluated, but the effective time of mineralization (i.e., < 5%) likely would remain unchanged.

The cooling rate during tin deposition was estimated at 0.01–0.07 °C/yr (ΔT/t2; Fig. 2B), which is significantly lower than that (1 °C/yr) in mafic systems (Hepworth et al., 2020) due to different cooling mechanisms. For mafic systems, cooling is much more efficient when hot magmas are emplaced into cold wall rocks. For relatively large and long-lived felsic systems, the crust is rather warm (i.e., a thermally mature system; Karakas et al., 2017), and conductive cooling is the primary means of heat loss in and around the magma reservoir. Hence, efficient advective cooling was, in the studied case, likely not achieved until meteoric water penetration was permitted.

Our in situ oxygen isotope analysis of paragenetically constrained cassiterite and quartz demonstrates that cooling is a primary driver of tin mineralization, and fluid mixing and waterrock interaction are not required. We also successfully quantified the effective time of mineralization to <5% of the total lifetime of the hydrothermal system, hence suggesting that the rate and efficiency of mineralization may be significantly higher than previously thought. These advances are possible only through an integrated application of novel in situ techniques to samples with robust petrographical constraints, which is readily applicable to other magmatic-hydrothermal systems. Our study highlights the importance of translating texture-controlled information into temporal patterns.

The synchronous growth of cassiterite and Q1 quartz (Fig. 1A) over several millennia requires a stable magmatic fluid source at a depth below the zone of mineralization, likely fed by fluid exsolution from a cooling magma reservoir. This prolonged time scale (greater than tens of thousands of years) for the existence of metalliferous fluids driven by cooling felsic magma reservoirs supports the extraction of critical metals directly from hydrothermal fluids (Blundy et al., 2021). However, as discussed herein, metal contents in the hydrothermal system are dynamic, and targeting the time window (or optimal conditions) when metal contents are at their highest, i.e., magmatic-hydrothermal systems in thermally mature crust with high degree of crystallization, is the key for success.

This research was funded by the National Natural Science Foundation of China (grants 42022022 and 92062220). We thank Elias Bloch, Matthieu Harlaux, and anonymous reviewers for their constructive comments, and editor Urs Schaltegger for his insightful suggestions.