Sodium Dithiocuprate(I) Dodecahydrate [Na₃(H₂O)₁₂][CuS₂], the First Crystal Structure of an Exclusively H-Bonded Dithiocuprate(I) Ion, and Its Formation in the Alkaline Sulfide Treatment of Copper Ore Concentrates

Abstract

This article presents the single-crystal structure of the complex salt sodium dithiocuprate(I) dodecahydrate Na₃CuS₂·12(H₂O), i.e., [Na₃(H₂O)₁₂][CuS₂], which forms in the high-sulfide concentrations of the alkaline solutions used for arsenic separation from copper concentrates. It features a linear hydrogen-bonded dithiocuprate(I) anion, a novelty in crystallographically characterized thiocuprates. During the study of the alkaline sulfide leaching of Chilean copper concentrates, an analytical investigation of the solution led to the detection of this complex. This study aimed to understand the chemical behavior of the leaching solution by identifying existing ions, which facilitated the discovery of the complex using single-crystal analysis. The newly discovered complex was also synthesized from a modeling solution based on the leaching solution recipe for arsenic removal, allowing for further crystal characterization through Raman and XRD analysis. By estimating the sodium sulfide threshold concentration that enhanced the formation of the copper disulfide complex, this study defined the upper technical threshold limit of sulfide concentration for the economic development of alkaline sulfide leaching to remove arsenic.

1. Introduction

Arsenic poses a global threat to health and to the environment in the copper mining industry [1,2]. Therefore, there is a significant economic and environmental interest in developing a sustainable and economical process chain for removing arsenic from copper concentrates [1,2]. In this regard, the partial roasting of copper concentrates is the current industrial standard for arsenic removal [3]. In recent years, numerous investigations have focused on potential methods of hydrometallurgical treatment, such as alkaline sulfide leaching, to remove mineralogically bonded arsenic from copper concentrates by dissolving it into the alkaline sulfide solution [1,2].

A previous study addressed the investigation of the thioarsenate species formed in the leaching solutions by ⁷⁵As NMR spectroscopy [4], and thus allowed for deeper insights into the solution part of this process. The copper components were treated as essentially unaffected in terms of leaching. This is (at least in part) justified, as the chemistry of alkaline sulfide leaching involves the formation of chalcocite (Cu₂S) and covellite (CuS), as well as sodium tetrathioarsenate from enargite or tennantite. The sulfide forms tetrathioarsenate and, in combination with the high pH of sodium hydroxide, passivates the copper as copper sulfides, which remain in the solid leaching residues [5,6,7,8] (cf. Figure 1a).

Regarding the chemical interaction of copper with the alkaline sulfide solution, R. N. Gow [5] describes the interaction between copper sulfide from enargite and the sulfide in alkaline sulfide leaching solutions. Consequently, it is probable that copper from enargite (Cu₃AsS₄), as well as from covellite (CuS) and chalcocite (Cu₂S) in the resulting residues, forms further complexes depending on the sulfide concentration, with a typical potential of −0.5 V to −1.0 V and a pH of 13. With increasing sodium sulfide concentration, there is a higher probability of copper dissolving, although still in negligible amounts, up from a concentration of around 2 M NaOH [5].

The alkaline sulfide leaching medium typically operates at a pH of 11–14, depending on the sodium hydroxide concentration. In this study, a pH of 13 was achieved using 2.5 M NaOH as the alkaline medium. While the leaching solutions were prone to the crystallization of hydrates of sodium sulfide upon storage at room temperature, we investigated the solids formed by single-crystal X-ray diffraction upon random crystal picking. In addition to the well-known salt Na₂S·9(H₂O) [9], which was identified by unit cell determination, one sample contained small colorless crystals of a hitherto unreported structure, which must have formed in the leaching procedure (Figure 1b). Structure solution and refinement revealed the formation of the dithiocuprate(I) salt Na₃CuS₂·12(H₂O), which confirmed the formation of discrete [CuS₂]³⁻ ions (Figure 1c).

The following paper addresses this compound, Na₃CuS₂·12(H₂O), which unequivocally underlines the role of thiocuprate(I) formation as a potential means of the loss of copper during alkaline sulfide leaching and which represents the first crystallographic evidence of a simple dithiocuprate(I) salt with an exclusively hydrogen-bonded anion.

2. Materials and Methods

2.1. General Considerations

Leaching solutions were prepared using sodium hydroxide (VWR, Darmstadt, Germany, analytical grade 95–100%) and sodium sulfide trihydrate (VWR, Darmstadt, Germany, chemical analysis quality 58–64% Na₂S content). Solutions for the deliberate preparation of Na₃CuS₂·12(H₂O) were prepared with sodium hydroxide (Th. Geyer GmbH, Renningen, Germany, 98.8% analytical grade) and sodium sulfide hydrate (Carl Roth GmbH, Karlsruhe, Germany, 60% Na₂S content).

For single-crystal X-ray diffraction analyses, crystals were selected under an inert oil and mounted on a glass capillary (which was coated with silicone grease). Diffraction data were collected on a Stoe IPDS-2/2T diffractometer (STOE, Darmstadt, Germany) using Mo Kα radiation. Data integration was performed with the STOE software XArea (version 2.3). The structures were solved using SHELXT-2018/2 and refined with the full-matrix least-squares methods of F² against all reflections with SHELXL-2019/3 [10,11,12]. All non-hydrogen atoms were anisotropically refined, and hydrogen atoms were located as residual electron density peaks and were refined without restraints. For details of data collection and refinement see Appendix A,

Table A1. Crystallographic data from the data collection and refinement of the structure of Na₃CuS₂·12(H₂O) out of data sets collected from a crystal formed in a leaching solution (entry 1), from a deliberate synthesis from Cu₂O in a leaching solution (entry 2), and from a synthesis out of Cu₂O in a more concentrated Na₂S/NaOH solution (entry 3).

Parameter	1	2	3
Formula	Cu2H48Na6O24S4 1	Cu2H48Na6O24S4 1	Cu2H48Na6O24S4 1
Mr	825.64	825.64	825.64
T (K)	200(2)	160(2)	160(2)
Λ (Å)	0.71073	0.71073	0.71073
Crystal size (mm3)	0.15 × 0.10 × 0.05	0.42 × 0.38 × 0.10	0.35 × 0.30 × 0.08
Crystal system	trigonal	trigonal	trigonal
Space group	$R\overline{3}$c	$R\overline{3}$c	$R\overline{3}$c
a = b (Å)	8.5067(5)	8.4801(2)	8.4804(2)
C (Å)	38.618(4)	38.6136(13)	38.5146(11)
V (Å3)	2420.2(4)	2404.76(14)	2398.77(13)
Z	3 1	3 1	3 1
ρcalc (g·cm−1)	1.699	1.710	1.715
μMoKα (mm−1)	1.74	1.74	1.75
F (000)	1284	1284	1284
θmax (°), Rint	26.0, 0.0310	35.0, 0.0310	27.0, 0.0474
Completeness	99.8%	100%	100%
Reflns collected	6016	11,355	12,829
Reflns unique	534	1189	590
Reflns [I > 2σ(I)]	391	970	495
Restraints	0	0	0
Parameters	46 2	46 2	45 2
GoF	1.021	1.074	1.100
R1, wR2 [I > 2σ(I)]	0.0188, 0.0377	0.0163, 0.0378	0.0287, 0.0907
R1, wR2 (all data)	0.0311, 0.0398	0.0243, 0.0398	0.0355, 0.0954
Largest peak/hole (e·Å−3)	0.25, −0.23	0.22, −0.43	0.90, −0.76

¹ The formula and Z reported here are in accordance with the CheckCIF suggested values. In a more chemically reasonable way, the description of CuH₂₄Na₃O₁₂S₂ (i.e., Na₃CuS₂·12(H₂O) in the form of [Na₃(H₂O)₁₂][CuS₂]), which then corresponds to Z = 6, is used in the discussion. ² For the structures of entries 1 and 2, the additional parameter corresponds to the extinction parameter refined (0.0032(2) and 0.00094(12), respectively), whereas this parameter refinement is not applicable in case of 3 (in a test it converged to 0.0000(3)).

. Graphics of the molecular structures were generated with ORTEP-3 [13,14] and POV-Ray 3.7 [15]. CSD 2447939 (entry 1), 2447941 (entry 2), and 2447940 (entry 3) contain the supplementary crystal data for this article. These data can be obtained free of charge from the Cambridge Crystallographic Data Centre via https://www.ccdc.cam.ac.uk/structures/ (accessed on 30 April 2025).

Single point Raman spectra in the range of 100–3800 cm⁻¹ were measured at room temperature using a confocal Raman microscope (XploRA™ Plus, Horiba Jobin Yvon SAS, Villeneuve d’Ascq, Département Nord, France) with an excitation wavelength of 532 nm (P = 10 mW). For each measurement, an average spectrum was obtained from ten individual 10 s scans, collected by directing the laser beam through the outer wall of a glass vial containing the crystals using a 10× objective in backscattering geometry. The spectral resolution was approximately 1.4 cm⁻¹, achieved with a 1800 grooves/mm grating. Signal detection was carried out using a CCD detector (Syncerity OE, Horiba France SAS, Paris, France), cooled by the Peltier effect to −60.1 °C. Prior to measurement, external calibration was performed using a silicon standard (520.5 cm⁻¹). Data collection and export were performed using LabSpec6 (Version 6.5.2.11) [16], and graphics were generated with Origin(Pro) (Version 2019b) [17]. Apart from offsetting the y-axis for clarity, no data manipulation was applied to the individual spectra. For a stack plot for direct comparison, the spectra were normalized to the intensity at 184 cm⁻¹, and the y-axes were offset to improve clarity.

2.2. Leaching Experiments

The alkaline sodium sulfide solution for leaching (2.5 M NaOH/2.0 M Na₂S·3(H₂O)) was prepared by dissolving 100 g of sodium hydroxide (titer factor: 1.075) into 0.6 L of deionized water. The resulting sodium hydroxide solution was made up to 1 L with deionized water. This prepared sodium hydroxide solution was subsequently used to dissolve the weighed mass of sodium sulfide trihydrate to obtain a solution of 2 M of sodium sulfide trihydrate at 80 °C or 90 °C inside a drying cabin for the leaching tests, as previously described in the Copper 2022 conference paper [18]. These solutions were used for the investigation of leaching with different Chilean copper concentrates. The sample used herein was a weathered sample, which contained chalcopyrite, enargite, tennantite, and chalcanthite as the Cu-containing constituents in addition to greater amounts of pyrite and melanterite. The concentrate was prepared for initial analysis and leaching experiments by sieving and grinding (vibratory disk mill), passing through a 315 µm sieve (the sieve tower consisted of five sieves of 1 mm, 500 µm, 315 µm, 100 µm, and 63 µm; for leaching the combined fractions that had passed through the, ≤ 315 µm sieves were used). In the sieving procedure, 20 rubber balls were used in the steps from 1 mm to 500 µm and from 500 µm to 315 µm for the soft deagglomeration of the concentrate feed to prevent reagglomeration due to inner wetness.

The grinding was performed during each turn by vibration swing grinding (Haver) and by vibration sieving (Haver EML). The standard analysis sieves (d = 400 mm) for the vibration sieves were made of stainless steel, whereby each sieving turn was run for 5 min at amplitude 5. The upper limit of the 315 µm sieve threshold for the concentrate was chosen due to the measuring limit of the particle size measuring system.

The results from the leaching experiment of the concentrate are shown in

Table 1. Details of the leaching experiment of the alkaline sulfide leaching treatment. A leaching solution consisting of 2.5 M NaOH/2.0 M Na₂S·3(H₂O) was used for the leaching of the concentrate in an s/l ratio of 1/10. The content was measured by AAS and Cu by ICP-OES [7,8]—*[Cu_ICP-gel] in the post-evaluation.

Parameter	Value
[AsFeed-Conc.]/wt.-%	4.0
[CuFeed-Conc.]/wt.-%	14.3
[SFeed-Conc.]/wt.-%	40.8
Leaching yield of As (solid based)/%	97
Cu residual concentration of leaching residues/wt.-%	16.7
As residual concentration of leaching residues/wt.-%	0.15
Volumed feed/mL	200
Volume after leaching/mL	186
Mass of feed concentrate/g	20.0
Mass of leached concentrate/g	11.9
[AsICP-gel] in g/L	3.52
[CuICP-gel]/mg/L	126.2

. The leaching test was conducted in a 400 mL double-walled glass vessel connected to an external thermostat. Solid–liquid mixing was ensured by a Teflon stirrer driven by an overhead motor. The vessel lid featured a port for inserting the CrNi temperature sensor or the pH electrode. The total volume for all experiments was set to 200 mL. For the monitoring of the leaching progress, a 3 mL sample was taken every hour, transferred to a 5 mL centrifuge tube, and centrifuged at 3000 rpm for 1 min for solid–liquid separation. The arsenic leaching yield was based on the analytical determination of residual arsenic concentration with the digestion of 0.5 g of the solid in aqua regia and the AAS measurement of the diluted aqua regia solution filled up to 50 mL with deionized water.

The copper concentrate for leaching was treated with alkaline sulfide solution for 4 h at 80 °C. Afterwards, the residues were separated from the solution by centrifugation and washed with water. After the leaching experiment, the double-wall glass vessel was decoupled from the external thermostat, and the stirring unit and lid were lifted up and cleaned. In the course of ca. 30–40 min, the suspension consisting of the alkaline sulfide solution and the leaching residue was allowed to cool down from the reaction temperature to the temperature range of 30–40 °C to ensure a safe transfer of the suspensions to the 250 mL centrifugation bottles (PP plastic bottles) by a 50 mL syringe and 10 mL pipettes. At the time of centrifugation, the sample had attained room temperature, and the centrifuge temperature program was set at 21 °C using a cooling table centrifuge (type 5910 RI, Eppendorf, Hamburg, Germany).

Within ca. 1 h, the centrifugation procedure was carried out in four consecutive stages: In the first step (at 3000 rpm for 10 min), the sulfide leaching solution was separated from the leaching residues for the analytical determination of the content of the solution by AAS. Afterwards, the leaching residues were washed with deionized water (two cycles of washing with 200 mL of deionized water (SERAL-water systems) followed by centrifugation at 3000 rpm for 10 min; in a third cycle, centrifugation was carried out for 30 min at 3500 rpm) to eventually obtain sulfide-free washing solutions. Upon the removal of the supernatants (which were discarded into sulfide-containing residues), the leaching residues were dried in the PP bottle with an open lid inside a drying box at 60 °C. After 24 h in the drying cabin, the solid residues were used for analytical determination.

The solution was analyzed by AAS. Differences in arsenic content were related to the norming conditions of the two different analytical methods. Upon storage at room temperature, some crystalline solid formed in the leaching solution (mainly sodium sulfide nonahydrate, as identified by determination and comparison of the unit cell parameters [9]). By crystal picking, some crystals of Na₃CuS₂·12(H₂O) were detected in small amounts.

2.3. Deliberate Preparation of Crystals of Na₃CuS₂·12(H₂O)

Batch 1: A total of 80 mL of an NaOH/Na₂S stock solution was prepared from NaOH (8.60 g), Na₂S·3(H₂O) (21.13 g), and 79.0 g of water. Two test tubes were charged with this solution (10 mL each), whereupon some solid Cu₂O was added to the test tubes (10 mg and 25 mg, samples A and B, respectively). In both cases, a black precipitate formed immediately, and upon stirring at room temperature (using a glass rod), the precipitate did not dissolve. The tubes were sealed with glass stoppers and stored at room temperature. In the course of two days, colorless plates of Na₃CuS₂·12(H₂O) formed spontaneously at the bottom of sample B (inside the black precipitate). Even though crystallization did not commence in sample A upon further storage for one week, the addition of a tiny seed crystal from sample B also initiated the formation of some crystals of the target product in sample A. The crystals obtained from sample B contained spots of black impurities as they grew inside the precipitate, but the edges of the crystals were clear and equipped with crystal faces that indicated a trigonal or hexagonal crystal structure. A piece of such a crystal was used for the X-ray diffraction analysis listed in Appendix A, Table A1, entry 2.

Batch 2: A test tube was charged with a solution prepared from NaOH (1.15 g), Na₂S·3(H₂O) (2.00 g), and 5.55 g of water. At room temperature, Cu₂O was added to the solution in small portions of some milligrams, and upon stirring with a glass rod, the black precipitate, which had formed immediately upon addition, dissolved. The addition of Cu₂O was stopped when the copper sulfide precipitate did not dissolve completely any longer, which was achieved upon the addition of 13 mg of Cu₂O in total. The test tube was sealed with a glass stopper, and the almost clear and colorless solution obtained was stored at room temperature. In the course of two hours, bunches of thin colorless crystalline plates formed at the bottom of the sample tube. Single-crystal X-ray diffraction analysis then confirmed the formation of Na₃CuS₂·12(H₂O), but the different crystal shape (no sharp edges and faces), the slightly different unit cell parameters, and the results of structure refinement indicated faults in this structure (cf. Table A1, entry 3). Upon further storage at room temperature, large amounts of crystals of Na₂S·9(H₂O) formed in the test tube. They did not dissolve upon slight warming (to ca. 35 °C) for some minutes. Upon the further storage of the contents at room temperature for one week, well-shaped colorless hexagonal plates and prisms formed on the wall of the test tube above the level of the Na₂S·9(H₂O) blocks. Single-crystal X-ray diffraction analysis then confirmed their identity as a well-crystallized version of Na₃CuS₂·12(H₂O). With the aid of a spatula, a sample of those crystals and some of the supernatant were transferred into a small sample vial for Raman spectroscopic analysis. The crystals could be stored in the presence of the mother liquor. Any attempts at isolating crystals of Na₃CuS₂·12(H₂O) as a pure compound for further analyses failed. When removed from the mother liquor and dried with a thin strip of filter paper (on a Petri dish on a microscope), the crystals started to turn dark in the course of some minutes. Immediate transfer of the crystals into a sample tube followed by evacuating and setting the sample under an inert atmosphere (dry argon) resulted in immediate decomposition to a black solid. Finally, the whole sample in the test tube was stored in a warm water bath at 45 °C with gentle shaking to eventually dissolve the solid Na₂S·9(H₂O). Colorless crystals of Na₃CuS₂·12(H₂O) remained, which were filtered off with suction and immediately washed with ethanol. The latter caused a dark brown discoloration of the crystal surface. Yield: 20 mg. Upon transferring the isolated and washed crystals into a sample vial, followed by evacuating and setting under dry argon, the crystals turned black immediately.

3. Results and Discussion

3.1. Preparation of Crystals of Na₃CuS₂·12(H₂O)

The discovery of small crystals of Na₃CuS₂·12(H₂O) in a leaching solution, which aimed at extracting thioarsenates from enargite, gave a first indication that this salt could, in principle, be prepared in alkaline sodium sulfide solution from a suitable Cu(I) source. As the extractive decomposition of enargite (Cu₃AsS₄) and other copper-containing minerals with the formation of soluble thioarsenates and copper-rich sulfides may provide very fine copper(I) sulfide, the in situ preparation of Cu₂S from Cu₂O was chosen in order to also start with a fine powder of Cu₂S and to avoid the introduction of further anions in this deliberate preparation. In a test which mimicked the leaching conditions, a small amount of Cu₂O was added to a 2.5 M NaOH/2.0 M Na₂S·3(H₂O) solution in a test tube, and the immediate formation of a black precipitate (Cu₂S) occurred. Upon the storage of this batch at room temperature, some large colorless plates formed at the bottom of the test tube. X-ray diffraction analysis (cf. Section 3.2 and Appendix A, Table A1, entry 2) confirmed the identity of these colorless crystals as the target compound, Na₃CuS₂·12(H₂O). Resulting from the crystal growth inside the Cu₂S precipitate, these colorless plates contained inclusions of black particles, and prolonged storage of the batch at room temperature did not lead to the full conversion of Cu₂S into the target compound.

In order to start the preparation from a clear solution of Cu(I) sulfidic species, Cu₂O was added to a more concentrated NaOH/Na₂S solution in a test tube. Upon the initial formation of a black precipitate, a clear solution was obtained upon brief stirring at room temperature, and a small amount of colorless crystals formed at the bottom of the test tube upon standing at room temperature over the course of some hours. X-ray diffraction analysis (cf. Section 3.2 and Appendix A, Table A1, entry 3) of these crystals basically confirmed the identity of these colorless crystals as the target compound, Na₃CuS₂·12(H₂O). In the course of prolonged storage at room temperature, large amounts of Na₂S·9(H₂O) formed in the test tube, which inhibited the isolation of the colorless crystals of the target compound. In an attempt at dissolving the crystals by heating (for recrystallization), complete dissolution was achieved, followed by the formation of a black precipitate upon further heating. The precipitate re-dissolved upon stirring as the solution had attained room temperature.

In a final attempt which basically repeated the previous procedure, upon the initial crystallization of Na₃CuS₂·12(H₂O) followed by the crystallization of Na₂S hydrates, the latter were dissolved in part by gentle warming, whereupon the batch was allowed to rest at room temperature for some days. In the course of storage, well-shaped crystals of Na₃CuS₂·12(H₂O) formed at the walls of the test tube above the crystalline blocks of Na₂S hydrates, which eventually allowed for the manual extraction of some of those crystals with a spatula. Their crystallographic analysis (not included in Table A1) confirmed a similar quality as in case of entry 2 in Table A1. The crystals were stable when stored under the original supernatant, but they decomposed (turned black) when exposed to air. Therefore, crystal preparation for X-ray diffraction analysis (for Section 3.2, preparation of crystals surrounded by a thin film of mother liquor under an inert oil) as well as Raman spectroscopic analysis of the crystals inside a sample vial in the presence of supernatant (Section 3.3) was possible. Removal of the film of mother liquor with, e.g., some filter paper resulted in the immediate dark discoloration of the crystals. Even without access to air, upon placing some crystals in a sample vial, drying in vacuum, and storing them under an inert atmosphere of dry argon, the crystals turned black. Thus, the procedure of drying itself, which may have initiated the loss of water of crystallization, can be regarded as the crucial step toward decomposition.

3.2. Crystallographic Analysis of Na₃CuS₂·12(H₂O)

The compound Na₃CuS₂·12(H₂O) crystallizes in the trigonal space group type R$\overline{3}$c. Upon the initial structure determination from a small crystalline piece that had formed in a leaching solution (cf. Appendix A, Table A1, entry 1), the attempts at the deliberate synthesis of this compound gave access to different crystals of essentially the same compound. From solutions which corresponded to the Na₂S/NaOH concentration levels of the leaching solutions, individual plates or prisms with well-shaped faces and trigonal or hexagonal appearance were obtained. Their size allowed for data collection at higher diffraction angles and the determination of this structure at an enhanced quality (Table A1, entry 2). Therefore, the data of the structure of entry 2 will be used for the discussion of molecular features. From more concentrated solutions, which allowed for the enhanced solubility of Cu₂S, clusters of thin irregular plates were obtained, the individual plates of which had a fish scale-like appearance without any characteristic edges, faces, and angles. Nonetheless, these crystals gave rise to similar diffraction data (single-crystal diffraction patterns, related unit cell dimensions), which allowed for structure refinement in the same space group type (Table A1, entry 3). Comparison of entries 2 and 3 in Table A1 (both data sets were collected at 160 K) revealed that the latter crystals exhibited a significantly shortened c axis (by 0.10 Å, which corresponded to a shortening of 0.26%). The temperature-dependent determination of the unit cell parameters of another crystal, which corresponded to the structure of entry 2 (

Table A2. Temperature-dependent unit cell parameters determined from a crystal of Na₃CuS₂·12(H₂O) in the same orientation and with the same set of ω-scans in increments of 1 deg. using an IPDS2T (STOE, Darmstadt, Germany) image plate detector distance of 95 mm, allowing for data collection up to 2θ = 60 deg. (φ-angle, ω-range: 0, 0–10; 0, 40–50; 90, 20–30).

	T = 240 (2) K		T = 120 (2) K
a = b (Å)	8.5046 (7)		8.4672 (7)
∆a (Å) 1		0.0374
∆a % 1		0.44
C (Å)	38.665 (4)		38.594 (4)
∆c (Å) 1		0.071
∆c % 1		0.18
V (Å3)	2421.9 (4)		2396.2 (4)
∆V (Å3) 1		25.7
∆V % 1		1.01

¹ The differences (∆) of the parameters X (X = a, c, and V) are reported as X(T₁)–X(T₂) for T₁ > T₂, and the relative differences (∆%) are reported with reference to the average value of X.

) showed that even a variation in temperature in the range of 240–120 K merely caused a shortening of the c axis by 0.18%, and even at 120 K the c axis was still noticeably longer than the length of the c axis found with the crystal of entry 3. Moreover, structure refinement for entry 3 resulted in a structure of markedly lower quality (according to R-values). These features, in combination with the highest residual electron density peaks (they are aligned in the (0 0 1) direction on the same axis with Na, S, and Cu atoms and thus may indicate ghost peaks from the heavy atoms), indicated the stacking faults of this structure. Even though we did not observe any noticeable streaking in the diffraction pattern as an additional sign of stacking faults, some reflections of systematic absences were observed with significant intensity, thus indicating some kind of twinning (e.g., contributions of local primitive hexagonal lattice types of similar unit cell dimensions). Problems like these were also encountered with a silyl derivative of nitrilotris (methylenephenylphosphinic) acid [19].

The asymmetric unit of the crystals of compound Na₃CuS₂·12(H₂O) comprises the atomic sites which are labeled without asterisks (*) in Figure 2a. The two crystallographically independent H₂O molecules have full site occupancy, whereas the atoms on special positions have partial site occupancy in the asymmetric unit (Na1, Na2, S1, and Cu1 with 1/6, 1/3, 1/3, and 1/6, respectively). The molecular moieties, which eventually constitute the crystal, are the cation [Na₃(H₂O)₁₂]³⁺ and the anion [CuS₂]³⁻. The former features three hexacoordinate Na ions in rather distorted octahedral coordination spheres, which form a column of three face-sharing octahedra (cf. Figure 2b–e).

This type of cation has been reported with other salts of tri-anions that crystallize in a related manner (i.e., with the formation of columns of alternating cat- and anions along the crystallographic c axis of a trigonal lattice). For insights into the flexibility of the Na coordination spheres of this kind of cation, selected data on the structures of Na₃CuS₂·12(H₂O) (cf. entry 2 in Table A1), Na₃SPO₃·12(H₂O) [20], and Na₃TlCl₆·12(H₂O) [21] are collated in

Table 2. Selected crystal data and interatomic distances (Å) and angles (deg) of the [Na₃(H₂O)₁₂]³⁺ cation from the crystal structures of Na₃CuS₂·12(H₂O) (this work, cf. entry 2 in Table A1), Na₃SPO₃·12(H₂O) [20], and Na₃TlCl₆·12(H₂O) [21]. The atom distances (1, 2, 3, 4) and angles (α, β, γ, δ) reported in this table refer to those shown in Figure 3.

Parameter	Na3CuS2·12(H2O)	Na3SPO3·12(H2O)	Na3TlCl6·12(H2O)
Crystal
Crystal system	trigonal	trigonal	trigonal
Space group	$R\overline{3}$c	$R\overline{3}$c	$R\overline{3}$m
a [Å]	8.4801(2)	9.061(2)	10.345(5)
c [Å]	38.6136(13)	34.34(2)	18.007(5)
Atom distances
1	3.28	3.18	3.24
2	2.40	2.40	2.41
3	2.40	2.38	2.44
4	2.41	2.38	2.40
Angles
α	78.3	80.4	80.7
β	78.6	81.1	79.7
γ	81.1	90.6	89.9
δ	96.1, 104.7	93.3, 94.7	86.1, 106.0
∆δ	8.6	1.4	19.9

and Figure 3. As both reference structures featured severe disorders (in Na₃TlCl₆·12(H₂O), both the anion and the terminal H₂O sites of the cation were disordered by space group symmetry; in Na₃SPO₃·12(H₂O), the anion was disordered by symmetry), we limited the discussion of the distortion of [Na₃(H₂O)₁₂]³⁺ to the most striking difference: Whereas in Na₃TlCl₆·12(H₂O) and Na₃SPO₃·12(H₂O), the O-Na-O angles between the terminal H₂O molecules (γ) were close to 90 deg., the corresponding angle in Na₃CuS₂·12(H₂O) was markedly smaller (81 deg.). This deformation could be attributed to the O–H···X (X = Cl, S, O) hydrogen bonds formed with the respective tri-anion. Whereas in Na₃CuS₂·12(H₂O), the H-bond acceptor sites (S) are located on the same axis as the Na atoms of the tri-cation, the other two anions ([TlCl₆]³⁻ and [SPO₃]³⁻) featured H-bond acceptor sites (Cl and O, respectively) out of that axis. Moreover, the H-bonding to the anions perpendicular to the trigonal axis required different torsion of the opposing octahedral faces Na(OH₂^terminal)₃ and Na(OH₂^bridging)₃, which was reflected by the difference ∆δ within the individual pairs of H₂O^bridging-Na-OH₂^terminal angles (δ).

In sharp contrast to the tri-cation [Na₃(H₂O)₁₂]³⁺, the anion of this structure represented a novelty. To our knowledge, this was the first crystallographic evidence of the anion [CuS₂]³⁻. As far as related motifs are concerned, in the structure of jalpaite (Ag₃CuS₂) [22], an atomic ensemble with linear S-Cu-S arrangement and a two-coordinate central Cu atom has been reported (it can be regarded as a [CuS₂]³⁻ ion), but each of its S atoms is bound to three silver atoms as nearest neighbors (at Ag–S distances of 2.53, 2.55, 2.55 Å), thus furnishing a polymetallic sulfide cluster rather than a salt of a genuine [CuS₂]³⁻ ion. Resulting therefrom, the Cu–S bond length at the two-coordinate Cu(I) in jalpaite (2.1822(2) Å) is markedly longer than the Cu–S bond in Na₃CuS₂·12(H₂O) (2.1188(2) Å). The latter is also shorter than the Cu–S bonds in di(thiolato)cuprate(I) salts such as [Et₄N][Cu(StBu)₂] (2.1380(7)–2.1434(6) Å, [23]) and [Ph₄P][Cu(SSiMe₃)₂] (2.1412(6), 2.1475(6) Å, [24]). In these cases, the longer Cu–S bonds can be attributed to the higher coordination number of the S atom (2 instead of 1). Moreover, higher coordination numbers at Cu(I) give rise to further Cu–S bond lengthening, for example, with tri-coordinate copper in [HN(CH₂CH₂)₃NH][Cu(SCN)₃] (2.2304(7) Å, [25]) and [Et₄N]₂[Cu(SPh)₃] (2.239(2)–2.258(2) Å, [26]) and with tetracoordinate copper in enargite Cu₃AsS₄ (2.291(3)–2.352(3) Å, [27]).

Even though the S atoms of the [CuS₂]³⁻ anion in Na₃CuS₂·12(H₂O) do not carry any further substituents but the Cu atom, the Cu–S bond length of 2.1188(2) Å is still influenced by further interactions in the crystal structure, i.e., a cage of 9 O–H···S hydrogen bonds about each of the S atoms. This large number of hydrogen bonds about a single sulfidic S atom is not unusual. For comparison, the S atom in Na₂S·5(H₂O) is surrounded by ten O–H···S hydrogen bonds in addition to the S-bound cation [28]. In the structure of Na₃CuS₂·12(H₂O), the O–H···S hydrogen bonds (O···S distances 3.218(1), 3.370(1), and 3.462(1) Å for O1–H1···S1, O1–H2···S1#1, and O2–H3···S1#2, respectively, with symmetry operations #1: −x + 1, −y, −z + 1 and #2: −y + 2/3, −x + 4/3, z−1/6) are formed by symmetry equivalents of H1, H2, and H3, whereas H4 and its symmetry equivalents are involved in O–H···O hydrogen bonding between adjacent [Na₃(H₂O)₁₂]³⁺ ions (O···O distance 2.813(1) Å for O2–H4···O1#3, with symmetry operation #3: −x + 5/3, −x + y + 4/3, −z + 5/6). Figure 4 shows the four different hydrogen bonds in this structure using the single O–H···O hydrogen bond as the central motif.

The role of H-bonding and the density of H-bonds about the [CuS₂]³⁻ anion are particularly visible in the results of a Hirshfeld surface analysis (Figure 5). This analysis was performed using the crystal structure associated with entry 2 in Table A1. (For comparison, a Hirshfeld surface analysis was performed on a structure refined against the same data set but with O–H bond lengths restrained to 0.98 Å, a more reasonable value in terms of O–H bond lengths determined by neutron diffraction [29]. Corresponding results from this comparison analysis are given as italicized numbers or italicized remarks in parentheses. They convey essentially the same picture.) In addition to the highly attractive interactions of the O–H···S and O–H···O hydrogen bonds, which sum up to 23.6% (23.6%) and 15.8% (16.5%), respectively, the remaining fraction of the Hirshfeld surface is connected with rather remote contacts of Cu···H (10.3% (10.5%), especially the blue equatorial “belt” around the anion and corresponding patches on the cation surface) and H···H (49.8% (48.7%), most of the blue areas on the surface of the cation).

3.3. Raman Spectroscopic Analysis of Na₃CuS₂·12(H₂O)

As the crystals of Na₃CuS₂·12(H₂O) decomposed upon their removal from the mother liquor, a sample of crystals with some drops of supernatant was filled into a small glass vial. For Raman measurements, the vial was held horizontally and the crystals were allowed to settle, whereupon the vial was turned 180 deg. about its axis to expose some crystals on the upper side of the vial which adhered to the glass wall in individual positions. The shape of the crystals (hexagonal plates or prisms) allowed for the visual identification of the positioning of the incident laser beam in the approximate (0 0 1) direction or perpendicular to this axis. Figure 6a–c give an impression of the positioning of three crystals used for three individual Raman measurements. Whereas in the case of positioning (c), the incident beam was directed along the trigonal axis, positionings (a) and (b) allowed for measurements in a rather perpendicular direction.

Figure 7 shows the full Raman spectrum recorded with the crystal positioning shown in Figure 6 a as a representative example. (The corresponding spectra for (b) and (c) are contained in the Supplementary Material.) The spectrum is dominated by strong bands of O–H stretch modes in the range of wavenumbers 3100–3500 cm⁻¹ (in this spectrum, two maxima are found at 3293 and 3393 cm⁻¹), and the sharp band around 1655 cm⁻¹ can be assigned to H–O–H deformation modes, which are expected features because of the cation [Na₃(H₂O)₁₂]³⁺. The range of lower wavenumbers features two distinct types of bands, i.e., sharp bands (in the range 100–200 cm⁻¹) and broad bands (in the range 400–900 cm⁻¹ as well as a shoulder at ca. 230 cm⁻¹), which can be assigned to lattice vibrational modes and modes of the [CuS₂]³⁻ anion, respectively. As to the latter, the isolated linear anion [CuS₂]³⁻ has point group symmetry D_∞h (like CO₂) and should thus give rise to Raman bands that correspond to the A_1g, E_1g, and E_2g modes. The positioning of the anion in the crystal may give rise to further splitting of the bands.

Figure 8 shows the lower wavenumber section of the spectra recorded from the three different crystal orientations shown in Figure 6. The direction-dependent relative susceptibility of the individual modes was in accordance with the axial symmetry of the crystal structure. (Direction-dependent Raman spectroscopic features have also been reported for other rather simple sulfides, e.g., wurtzite ZnS [31].) As to the overall position in the spectrum, the modes of [CuS₂]³⁻ are shifted to markedly higher wavenumbers with respect to bands arising from tetracoordinate Cu(I), like the CuS₄ environment in enargite and tennantite, where characteristic bands appear in the range 300–400 cm⁻¹ [32]. This shift to higher wavenumbers reflects the Cu–S bond shortening (or bond strengthening) in [CuS₂]³⁻.

4. Conclusions

The simple dithiocuprate(I) anion, [CuS₂]³⁻, combines both the thiophilicity of the HSAB soft d¹⁰ center Cu(I), which is also reflected in the abundance of sulfidic copper minerals such as enargite, and the tendency toward linear coordination, i.e., the formation of d¹⁰-ML₂ complexes. The latter is well known for gold (e.g., [AuCl(PMe₃] [33]), silver (e.g., [Ag(PBn₃)₂][BF₄] [34]), and also for copper (e.g., formation of the anion [CuCl₂]⁻ and salts thereof such as [NnBu₄][CuCl₂] [35] and thiolatocuprates such as [Cu(StBu)₂]⁻ [23]). In this regard, it is rather surprising that the anion [CuS₂]³⁻ had not previously been confirmed crystallographically. Now it complements other (anhydrous) sodium thiocuprates(I) such as NaCu₅S₃ [36], Na₂Cu₄S₃ [37], Na₄Cu₂S₃ [38], and NaCu₃S₂ [39]. In addition to CuS₃-coordination spheres (in NaCu₅S₃, Na₂Cu₄S₃, and NaCu₃S₂) and CuS₄-coordination spheres (in NaCu₃S₂), some of them also feature rather linear CuS₂ arrangements (NaCu₅S₃ and Na₄Cu₂S₃), but always complemented by a greater number of further Cu atoms and additional (more remote) S atoms. From the academic perspective, the finding of this simple anion is a novelty in general. Moreover, its initial finding in a product solution of an industry-related ore leaching solution rather than in a batch from a sophisticated laboratory synthesis clearly emphasizes the importance of further exploration of seemingly simple fundamental chemistry. Regarding the importance of the role of [CuS₂]³⁻ in alkaline sulfide leaching, attempts at the deliberate preparation of Na₃CuS₂·12(H₂O) have shown that this anion forms from Cu₂S and alkaline sulfide solution with dissolution at lower temperatures, whereas heating fosters its decomposition with the precipitation of Cu₂S. Thus, the temperature range applied for alkaline sulfidic leaching serves a kinetic purpose to accelerate the leaching of, e.g., arsenic compounds, and also plays a thermodynamic role in suppressing the formation of dithiocuprate(I) in greater amounts. While the formation of oligosulfidic Cu(I) complexes in aqueous solution is well known, e.g., NH₄CuS₄ [40,41], and their formation can be suppressed by avoiding the prerequisite of a greater excess of sulfur in the alkaline sulfide solution (higher oligosulfide solution), the finding of Na₃CuS₂·12(H₂O) provides new insights into the role of simple sulfide in the competitive dissolution of copper.