Introduction

Sphalerite is one of the major industrial minerals not only for zinc, but also for dispersed metals, including cadmium, gallium, germanium, indium1, and sometimes silver and gold2. Solid-state processes, such as dynamic recrystallization for germanium3, or phase re-equilibration during diffusion for indium4 have been proposed to explain their redistribution in sphalerite. However, the role of aqueous process in the initial or secondary enrichment of these elements in sphalerite is still obscure. As a dispersed element with Clarke value less than 0.9 ppm, cadmium (Cd) tends to be enriched 103–104 times as trace impurity in the sphalerite lattice4,5. However, supernormal enrichment of Cd up to several wt%6, or as greenockite inclusions in sphalerite5 in sulfide ore assemblages has been frequently observed in some Mississippi Valley-type (MVT) Zn–Pb deposits. These suggest local upgrading of Cd in sphalerite exists, although the inherent mechanism has not been well constrained. In this work, we present nanoscale mineralogical evidences for detailed supernormal enrichment process of Cd in the sphalerite of the Jinding deposit via coupled dissolution–reprecipitation reaction (CDR). This work may facilitate understanding the redistribution of dispersed metals to form their own high-grade ores in the sediment-hosted Zn–Pb deposits.

Results

The Jinding deposit, supernormal Cd enrichment occurrence, and cadmiferous sphalerite

The Jinding deposit, within the Mesozoic-Cenozoic Lanping basin, Sanjiang Tethys Orogen7,8, hosts zinc ores rich in cadmium grading 0.01–0.2 wt%9. It is regarded as a subtype of MVT Zn–Pb deposit that is hosted in a thrust-triggered salt-diapiric mélange10,11. Three major stages of mineralization are identified in the Jinding deposit12,13, represented by (S1) fine-grained disseminated pyrite–marcasite–sphalerite–bitumen replacing pre-ore calcite cement of quartz sand, (S2) pyrite–galena halo that replaced pre-ore calcite and overprinted S1, and (S3) coarse-grained crystalline galena–sphalerite–barite veins.

Supernormal enrichment of Cd was recognized by scanning electron microscopy energy disperse spectroscopy (SEM–EDS) studies on mining products from a nearby concentrator plant, revealing the ubiquitous distribution of submicron- to micron-sized greenockite blebs and floccules in some cadmiferous sphalerite debris (Fig. 1a–d), suggesting potential existence of Cd-rich microsites. In situ, cadmiferous sphalerite with discrete greenockite inclusions was found within galena-dominated veins (Fig. 1e, f). These veins represent sub-types of S3 ore assemblages that have been reported by many researchers14. The cadmiferous sphalerite is dark greenish-gray crystalline, accompanied with coarse-grained galena and calcite (Fig. 1e). In the BSE image, it presents complex core–mantle–rim texture (Fig. 1f). The cores of the sphalerite are porous, showing mosaic and fan-shaped zoning patterns with bright contrast in the BSE image, which host discrete greenockite and galena inclusions (Fig. 1f). Two forms of inclusions are amorphous flocculent aggregates (Fig. 1g), and crystallized blebs (Fig. 1h), similar with those observed in zinc concentrates (Fig. 1a–d). These inclusions are related to hypogenic dissolution pores, and are mostly arbitrarily oriented, showing discordance with the growth direction of the sphalerite host (Fig. 1g–i). The mantles of the sphalerite have dark contrast in the BSE image, and are deficient in pores and inclusions (Fig. 1f).

Fig. 1: Hand specimen, microscopic, and SEM images of CdS phases in studied sphalerite.
figure 1

a–d CdS blebs, floccules, and associated cadmiferous sphalerite debris in zinc concentrates. e A typical Zn–Pb ore specimen from the Jinding deposit recording three stages (S1–S3) of sulfide mineralization. The location of the studied area in (e) is marked. f Greenockite and galena inclusions irregularly distributed within a cadmiferous sphalerite crystal with core–mantle–rim zoning textures. The Cd-rich cores that host sulfide inclusions are marked by white dashed lines. g–i Detailed FIB–SEM showing typical floccule-type and bleb-type CdS and PbS inclusions. The extraction positions of the FIB cuts are marked by black dashed rectangular. Abbreviations: Brt barite, Gck greenockite, Gn galena, Sp sphalerite.

Full size image

EPMA point analyses of the cadmiferous sphalerite revealed that the Cd concentrations of the inclusion-bearing cores (1.47–8.67 wt% Cd, Supplementary Table 1) are significantly higher than those of inclusion-free mantles and rims (0.48–3.00 wt% Cd, Supplementary Table 1). Besides, the cores are more enriched in Ag (100–920 ppm) than the mantles (<detection limit), yet more depleted in Pb (<detection limit) than the mantles (50–2490 ppm). The greenockite inclusions are characterized by relatively high Zn concentrations (0.28–2.14 wt%). The galena inclusions are moderately enriched in Zn (0.04–0.98 wt%) and Cd (up to 0.85 wt%). Notably, high concentrations of Cd (1.83 wt%, Spot-40) and Ag (1.13 wt%, Spot-5) in galena inclusions were detected (Supplementary Table 1).

Discovery of various Cd-rich phases in the cadmiferous sphalerite

Focused ion beam (FIB) nano-sampling was conducted for subsequent transmission electron microscopy (TEM) studies. FIB cut #1 was extracted across floccule-type greenockite and galena inclusions and cadmiferous sphalerite (Fig. 1g). FIB cut #2 was extracted across a bleb-type greenockite inclusion and its sphalerite host (Fig. 1h). FIB cut #3 was extracted across high-Cd, low-Cd zones in sphalerite and a bleb-type galena inclusion (Fig. 1i). The high-angle annular dark-field (HAADF) images of FIB sections, positions of TEM observation, as well as related TEM–EDS spectra discussed below are shown in Supplementary Fig. 1. Caution has to be paid on the representativeness of sampling locations because of their small sizes and limited numbers, when compared to complex and spatially heterogeneous hydrothermal processes. The above-mentioned FIB cuts cover all types of Cd-anomaly phases that can also be seen ubiquitously in zinc concentrates, which minimize the uncertainty of results as much as possible.

The HAADF image on the FIB cut #1 shows that three Cd-rich phases are hosted in the dissolution pore of the cadmiferous sphalerite (Supplementary Fig. 1a). Their compositions are confirmed by qualitative TEM-EDS spectra (Supplementary Fig. 1a, spot 1). The bright-field TEM (BFTEM) image and selective area electron diffraction (SAED) pattern reveal that the floccule-type greenockite is polycrystalline and consists of randomly oriented cylindric nanoparticles (NPs) (Fig. 2a). The d-spacing of NPs is directly measured by high-resolution TEM (HRTEM) observation, confirming that they are poorly crystallized hexagonal-CdS NPs, with distorted areas in lattice along [002] zone axis (Fig. 2b). There are no other elements detected in the CdS NPs by TEM-EDS (Supplementary Fig. 1a, spot 1).

Fig. 2: STEM/TEM images, SAED patterns, and STEM-EDS elemental maps of Cd-rich phases from locations marked in the Supplementary Fig. 1.
figure 2

a, b Bright field TEM (BFTEM), high-resolution TEM (HRTEM) images, and SAED patterns of CdS nanoparticles. c, d BFTEM, HRTEM images, and SAED patterns of PbS nanoparticles. e, f BFTEM, HRTEM images, and SAED patterns of ZnS nanoparticles. g, h BFTEM, HRTEM images, and SAED patterns of the sphalerite host. Positions for SAED are marked by white targets. Typical NPs are marked by white dashed circles. Positions for HRTEM imaging in (b, d, and f) are marked by the yellow circles. i, j HRTEM images of phase boundaries between the crystalline sphalerite and greenockite. Phase boundaries are marked by white dashed lines. k–m HAADF images and atomic-scale TEM-EDS mapping of Cd-anomaly along a low-angle tilt boundary within the sphalerite lattice. Position for the high-resolution HAADF image in (l) is marked by the white solid rectangle. The lattice directions in (m) are marked by white arrows. Abbreviations: Gck greenockite, Gn galena, Sp sphalerite, c-PbS cubic PbS, c-ZnS cubic ZnS, h-CdS hexagonal CdS.

Full size image

The BFTEM image and ring-like SAED pattern of floccule-type galena confirm that it is composed of numerous rectangular cubic-PbS NPs (Fig. 2c). They (~15 nm) are angular and well crystallized with perfect lattice planes, and without any defects in the HRTEM image (Fig. 2d). Detectable Cd concentration in the PbS NPs is confirmed by qualitative TEM-EDS spectrum (Supplementary Fig. 1a, spot 2).

ZnS NPs are recognized within the porous zone of the cadmiferous sphalerite near a dissolution pore (Supplementary Fig. 1a, spot 3), as confirmed by the BFTEM image and ring-like SAED pattern (Fig. 2e). The qualitative TEM–EDS spectrum revealed the major composition of ZnS, with detectable concentration of Cd (Supplementary Fig. 1a, spot 3). The d-spacing of NPs is directly measured by HRTEM observation, which suggests they are cubic-ZnS NPs (Fig. 2f). The ZnS NPs are generally distributed around nanoscale dissolution pores (Bright domains, Fig. 2e). Under the HRTEM observation, two cubic-ZnS NPs attach to each other via particle attachment, and show typical sphalerite-type {111} single stacking faults and {111} twin boundaries (Fig. 2f). The cadmiferous sphalerite host is crystallized, and is in transitional contact with the cubic-ZnS NPs, as confirmed by the BFTEM image and SAED pattern (Fig. 2g). The sphalerite host is highly distorted with enormous stack faults (Fig. 2g). A few thin wurtzite-like lamellae (<5 nm) are also recognized in the sphalerite lattice (Fig. 2h). The sphalerite host is also enriched in Cd (Supplementary Fig. 1a, spot 4).

There are nanoscale paragenetic relationships between different phases in the FIB cut #1 (Fig. 3). The HADDF image and TEM-EDS mapping confirm the accumulation of CdS and PbS NPs at the wall of nanoscale dissolution pores, or as fillings in intergranular space between ZnS NPs (Fig. 3a, TEM–EDS mapping). Such a phenomenon indicates earlier formation of the ZnS NPs is followed by later precipitation of the CdS and PbS NPs. The appearance of a nanoscale vein of PbS NPs crosscutting the ZnS NPs (Fig. 2e) and CdS NPs penetrating PbS and ZnS NPs (Fig. 3b, EDS mapping) also supports that the formation sequence of the NPs follows the order of ZnS, PbS, and CdS.

Fig. 3: HAADF images and STEM-EDS elemental maps from locations shown in supplementary Fig. 1.
figure 3

a Accumulation of CdS NPs at the wall of nanoscale dissolution pores, whereas PbS NPs accumulate as fillings in dissolution pores. Note the transitional relationship between ZnS NPs and crystalline sphalerite. b A nanoscale vein composed of CdS NPs penetrates into the interval of PbS NPs and ZnS NPs. c A dissolution pore hosting newly formed PbS and ZnS NPs. Note the abundant nanometric protrusions and etch pits appear around the dissolution pore. Phase boundaries in a–c are marked by white dashed lines. Abbreviations: Gn galena, Sp sphalerite.

Full size image

The HAADF image on the FIB cut #2 shows a negatively curved contact interface between a bleb-type CdS inclusion and its sphalerite host (Supplementary Fig. 1b). The d-spacing of the inclusion is directly measured by HRTEM observation, suggesting that it is crystallized hexagonal-CdS polymorph, namely greenockite (Fig. 2i). There is an obvious mis-orientation between lattices of the greenockite bleb and its sphalerite host as imaged by HRTEM observations (Fig. 2I, j). Edge dislocations occur at the sphalerite side near phase boundary, with the density of dislocations decreasing toward interior sphalerite lattice (Fig. 2j). The HAADF image of the FIB cut #3 shows a negatively curved boundary between a bleb-type galena inclusion and its sphalerite host (Supplementary Fig. 1c). A domain with brighter Z-contrast than that of surrounding areas appears within the sphalerite lattice near a porous zone in the HAADF image (Fig. 2k). Further HRTEM imaging and atomic-scale TEM-EDS mapping confirm that it is a Cd-anomaly zone distributed along a low-angle tilt boundary within the sphalerite lattice (Fig. 2l).

Discussion

A coupled dissolution–reprecipitation model for supernormal enrichment of Cd in sphalerite

Here we present a dissolution-reprecipitation model (Fig. 4), for successive Cd upgrading from lattice-bound impurity, defect-anomaly, sulfide NPs, to “greenockite diseases” (referring to speckling of micrometer-sized greenockite blebs). The initial enrichment of lattice-bound Cd critically depends on (1) Instantaneous fluid composition during crystal growth15. Hydrothermal replenishment from a basement-equilibrated fluid reservoir carrying Cd15,16,17, as well as potential fluctuation of the basinal-scale hydrodynamic regime during sphalerite crystallization, may contribute to various Cd content during sphalerite growth. (2) Physio-chemical parameters. From this perspective, the formation of Cd-rich sphalerite favors low temperature, low ∑Sred (SO42−-dominant) hydrothermal conditions18. (3) Crystallographic controls such as zoning pattern19, and degree of crystallization20. These factors emphasize that concomitant microscale variation in fluid composition at the fluid-crystal interface may play a key role in single crystal-scale Cd content variation in sphalerite. As for the Jinding sphalerite, crystallographic controls are negligible because of enormous variations of Cd content in both sphalerite crystals and fluid inclusions21,22.

Fig. 4: Cartoon illustrations of supernormal enrichment of Cd in sphalerite via coupled dissolution–reprecipitation reaction.
figure 4

a Primary precipitation of defective early-stage sphalerite during fluid mixing between metalliferous hot brines and reduced sulfur reservoirs. b Under reduced sulfur-deficient conditions, injection of highly evolved, oxidative, acidic, residual brines that are rich in Cd2+ and Pb2+ but depleted in Zn2+ triggers oxidative dissolution of primary sphalerite and formation of an interfacial aqueous layer where supersaturation of CdS and PbS occurs. The sphalerite host is further rich in Cd due to selective leach-out of ZnS components in its lattice and structural defects. c Supernormal enrichment of Cd as Cd-rich nanoparticles when CDR proceeds. d H+ consumption and sufficient reduced sulfur supply cease CDR, when late-stage sphalerite precipitation takes place. The sulfide nanoparticles experience coarsening and passivation. e A schematic diagram of an interface coupled dissolution–reprecipitation reaction for Cd-rich nanoparticle formation. The circles represent the sulfide nanoparticles. The yellow dashed line represents a Cd anomalous lattice defect.

Full size image

The oxidative dissolution of the cadmiferous sphalerite at the mineral-fluid reaction front to form “greenockite diseases” is based on the following facts: (1) The “greenockite diseases” are exclusively distributed within porous cadmiferous zones in sphalerite (Fig. 1a–d, f). (2) Chemical and textural anomalies of the sphalerite host, including nonstoichiometric (atomic sulfur <metal, Supplementary Table 1), high density of {111} stack faults, and wurtzite-like lamellas (Fig. 2g, h). These features suggest low fugacity of sulfur during sphalerite precipitation2,23, or more likely indicate partial oxidation of the sphalerite lattice by fluids with higher oxygen fugacity24. (3) The transition relationship between porous aggregates of cubic-ZnS NPs and defect-rich crystalline sphalerite indicates the pre-existence of a bulk sphalerite crystal that is partially disintegrated to form NPs and dissolution pores (Fig. 3a). The driving force causing sphalerite instability first orderly requires acidic and oxidative conditions in reacting fluids, while topographical kinetics of the reacting surface limits dissolution rates25,26. Studies by atomic force microscopy emphasize the formation of nanometric protrusions25, transient aqueous layers enriched in remobilized sulfur26,27, as well as lattice orientation transition between parent and product phases28 as direct nanoscale evidences of oxidative dissolution on sphalerite mediated by acidic fluids. In this work, TEM observations confirm that abundant nanometric protrusions and etch pits appear in the porous zones of sphalerite (Fig. 2g), simultaneously with fillings of sulfide NPs (Fig. 3a–c), as well as mis-orientation between ZnS NPs and the crystalline sphalerite host (Fig. 2e–f vs. 2h-i), which are in agreement with above-mentioned criteria.

Mineralization is a process in which a variety of NPs can be formed29. However, sulfide NP’s formation, transport, and coarsening in the hydrothermal deposits are rarely documented. In a few cases, the formation of sulfide NPs can be attributed to solid-state exsolution from host sulfides by posterior modification, as well as supersaturation nucleation during hydrothermal vent emission30,31,32. Recently, the CDR has been considered an effective approach for synthesizing sulfide nanomaterials such as thiospinel mineral violarite (Ni,Fe)3S4 and synthetic roquesite CuInS233,34, although this process has rarely been reported in natural materials. The CDR generally involves disequilibrium between a solid phase and contact fluids when mineral dissolution takes place. An interfacial layer can then form and be supersaturated with respect to more stable phases that nucleate and grow along the reaction front35. As for this case, the critical requirements for ZnS, CdS, and PbS NPs’ formation via CDR have been met. The porous aggregate of cubic-ZnS NPs acts as a reaction front with high surface activity during cadmiferous sphalerite’s decomposition (Fig. 2g). The initial oxidative dissolution of sphalerite starts at Cd-rich zones because sphalerite with low Zn/Cd ratios is less stable than low-Cd sphalerite or Cd sulfides, and can break down at much faster rates under acidic and oxidative conditions36. It is also supported by the exclusive distribution of “greenockite diseases” within cadmiferous zones (Fig. 1f). Selective leach-out of Zn versus Cd along low-angle tilt boundary can also be observed in the sphalerite lattice, suggesting that structural defects may also serve as a privilege for mineral-fluid reaction (Fig. 2l, m).

As the oxidative dissolution proceeds, ions of Zn2+, Cd2+, Pb2+, and reduced sulfur would be remobilized or replenished from exterior fluids to form a distinct interfacial aqueous layer37, where sulfide NPs’ reprecipitation is attributed to the relative stability of sulfides, following the sequence of ZnS > CdS > PbS under oxidative acidic conditions38. The crosscut relationships between three different NPs are in accordance with such a process (Figs. 2e, 3a, b). Notably, the subsequent dissolution and reprecipitation processes may be self-sustainable because the newly-formed phases can initiate autocatalytic reactions coupling the dissolution and precipitation rates39. This is indeed the case of CdS and PbS NPs, The Cd- and Pb-dopped sphalerite exposed to oxygen and acid can form a series of sulfide voltaic cells, including ZnS–PbS and ZnS–CdS pairs. The ZnS component with lower electrochemical potential acts as an anode and is prone to be dissolved by oxidative acidic fluids, whereas the components of CdS and PbS act as cathodes with higher electrochemical potentials and are protected from further mineral-fluid interaction, remaining at reacting interfaces to form sulfide NPs40.

Theoretically, minerals pass through a nanocrystal stage during formation29. However, this stage tends to be transitory and metastable and rarely to be preserved in geological settings41. In this work, both crystallized blebs of galena and greenockite and aggregates of PbS and CdS NPs distribute as diseases within one single sphalerite crystal, which suggests both NP coarsening and passivation took place during or after CDR. Studies on NP attachment have provided evidence for sphalerite crystallization via Ostwald ripening and/or oriented attachment by ZnS NPs42,43. The oriented attachment by cubic-ZnS NPs (Fig. 2f), as well as the high density of edge dislocations in sphalerite near the sphalerite–greenockite boundary (Fig. 2i) may document a healing history of sphalerite from ZnS NPs after extensive CDR. However, coarsening of CdS and PbS NPs is not observed, which is in paradox with the ubiquitous distribution of their crystalline counterparts. One possible reason for preservation of sulfide NPs is the presence of heterogeneously distributed natural organic matters in dissolution pores, especially carboxylic acids (C6H5O73−), which can act as passivators against NP coarsening44,45,46. Notably, hydrocarbon-bearing aqueous inclusions as well as solid bitumen are widely distributed in sphalerite-rich ores of the Jinding deposit47. Besides, large Cd-isotope fractionation in sphalerite in the Jinding deposit has been attributed to the presence of bacterial metabolic Zn-carboxylate in hydrothermal fluids22.

Redox and pH fluctuations in the fluid are essential for zinc-cadmium decoupling and discrete formation of Cd sulfides either at (1) hypogenic hydrothermal deposits such as skarn48, stratiform zinc-lead deposit49; (2) terrestrial sedimentary rocks such as lacustrine shale50, and (3) redox-sensitive modern systems such as periodically water-logged organic soil38, and polluted peatland51. Thus, we propose hydrothermal overprinting by late-stage oxidative acidic brines on early-formed Cd-rich sphalerite as a main trigger for supernormal enrichment of Cd via CDR. From hand specimen perspective, the dark-brown cadmiferous sphalerite with greenockite diseases represents an alternation front between S1 pale yellow disseminated sphalerite and S2 galena-dominated veins (Fig.1e). From single mineral perspective, the mantles and rims of the sphalerite are mismatched with cadmiferous cores, which have muti-center nature with irregular and vague contours (Fig. 1f). These features suggest the cores are relics of S1 Cd-rich disseminated sphalerite, rather than newly-formed sphalerite in galena-dominant veins.

The mixing of deeply-derived, oxidative (SO42−-dominant) and metal-bearing brines with shallow, H2S-rich, and reducible sedimentary formation waters is responsible for mineralization in the Jinding, as well as other Cd-rich MVT Zn–Pb deposits7,52. During brine ascending and cooling, the pH of the fluid shifts towards a more acidic condition, and the oxidation state increases above the hematite–magnetite buffer52. The massive precipitation of pyrite, marcasite as well as other sulfides at early stage can also produce superfluous H+ in the residual brine53 that is required for acid-mediated oxidative dissolution of early formed sphalerite. Besides, a previous study has confirmed that from S1 disseminated galena (109–432 ppm Cd, 56–66 Zn/Cd), S2 vein-type pyrite (23–1659 ppm Cd, 30–175 Zn/Cd), to S3 coarse-grained galena (1983 ppm, 20 Zn/Cd), the Cd concentrations and Cd/Zn ratios in fluid inclusions increase gradually21. Thus, a late-stage, oxidative, acidic, and highly evolved residual brine with low Zn/Cd ratios54 can serve as a candidate for CDR. We speculate that transient injection of metalliferous (Pb-rich, low Zn/Cd ratios) residual brines without sufficient sulfur supply from H2S reservoirs (Fig. 4a, b) will lead to CDR rather than sphalerite precipitation (Fig. 4c, e), and cause supernormal enrichment of Cd as discrete CdS phases (Fig. 4c). After H+ is consumed by mineral-fluid reaction or carbonate buffers, and supply of reduced sulfur is sufficient, sulfide precipitation predominates, and late-stage sphalerite with variable Cd content is formed (Fig. 4d). The Cd concentration in the newly-formed sphalerite is influenced by fluid composition and physio-chemical variations, rather than CDR.

Implications on dispersed element targeting and nanomaterial synthesis

A large proportion of the dispersed elements resources Cd, Ge, Ga, and In is associated with sediment-hosted Zn-Pb deposits. However, exploitation of these metals is rather challenging because of spatial heterogeneity in their enrichment in orebodies, as well as low-grade concentration in specific host minerals. Thus, understanding the remobilization and redistribution of these metals to form their own high-grade ores in the deposit scale will lead to enhanced extraction and recovery efficiency3. Recent studies emphasize solid-state exsolution or diffusion as the dominant mechanism for dispersed elements’ redistribution at the microscopic scale, as exemplified by (1) Ge redistribution from trace impurities to briartite Cu2(Zn, Fe)GeS4 NPs by sphalerite deformation55, and (2) formation of basket-weave indium-rich domains in the sphalerite lattice by solid-state phase re-equilibration56. These observations give instruction to targeting enrichment microsites of these metals at certain brittle–plastic deformation zones or temperature gradients of certain deposits.

As a common process causing element redistribution and NP formation in both hypogene and supergene conditions57, mineral replacement reaction in aqueous fluids via CDR has long been overlooked as one potential mechanism for dispersed metals’ redistribution. We provide robust nanoscopic evidence that confirms the essential role of CDR during supernormal enrichment of Cd in MVT Zn–Pb deposits. Future targeting for Cd-rich zones in this kind of deposit may focus on Zn–Pb orebodies where intense hydrothermal overprinting between different ore-forming stages occurs. Except for this work, the observation of Ge redistribution and concentration into hemimorphite at the replacement front of sphalerite in zinc mine tailings58 suggests that CDR may also play a key role in the secondary enrichment of Ge under supergene conditions. Thus, future work should also pay efforts on supergene enrichment of dispersed metals via potential CDR approaches.

The hexagonal-CdS NP formation via CDR may also shed light on manufacturing II-VI semiconductor materials, such as ZnS, PbS, and CdS nanomaterials. Although not been massively applied yet, the CDR routes are confirmed effective for synthesizing nanomaterials that are often difficult to acquire via traditional methods or nanomaterials with unique properties33,34. Thus, detailed studies on sulfide NPs formed in nature via CDR may enhance artificial syntheses of these materials in the laboratory.

Methods

Geological sampling

Ore sampling was carried out by the lead author from the 2560 mining platform, the Beichang open pit, Jinding deposit. The World Geodetic System 1984 (WGS84) geographic coordinates for the location of the sample used in this study are 99°25′35.93″N, 26°24′42.33″E.

Electron probe microanalyses

EPMA analyses were carried out on studied sphalerite, greenockite, and galena using EPMA-1720 of SHIMADZU housed in the State Key laboratory of Geological Processes and Mineral Resources, China University of Geosciences, Beijing, to determine major element concentrations and elemental distributions. Elements including S, Zn, Pb, Cd, Ag, Fe, Mn, Sb, Co, Cu, and As were measured at 15 kV of accelerating voltage, 20 nA of current, 1 μm spot size, 30 s peak time and 15 s background time. The following standards were used: chalcopyrite (S-Kα, Fe-Kα, Cu-Kα), sphalerite (Zn-Kα), galena (Pb-Mα), Cd metal (Cd-Lα), Ag metal (Ag-Lα), rhodonite (Mn-Kα), antimony telluride (Sb-Lα), and cobaltite (Co-Kα, As-Lα). The EPMA analyzing results are shown in Supplementary Table 1.

TEM/STEM–EDS observations and analyses

The thin sections for the TEM study were prepared using FIB–SEM by cutting across the sphalerite host and greenockite and galena inclusions. The sampling locations are shown in Fig. 1g–i. Then they were placed on Cu grids and thinned. HRTEM imaging and element X-ray EDS mapping were performed by using an FEI F200X Talos field-emission TEM with an accelerating voltage of 200 kV and with a 16 MB ceta camera. Ultrahigh-resolution HAADF imaging and atomic EDS mapping were conducted on a Spectra300 AC-TEM with an accelerating voltage of 300 kV and with a SuperX detector in Sinoma Institute of Materials Research, Guangzhou, China. Diffraction measurements were carried out by using DigitalMicrograph 3.11.1 and Winwulff 1.4.0 software. Data from the American Mineralogist Crystal Structure Database were used for indexing electron diffractions.