Micro-Raman spectra were collected on a sample of delchiaroite in the range of 100–4000 cm⁻¹ using a Horiba Xplora apparatus mounted on a confocal Olympus BX41 microscope (Dipartimento di Scienze della Terra, University of Pisa, Italy). The Raman signal was excited by an unpolarized green-diode-pumped solid-state laser (λ=532 nm) and measured by a CCD detector. The experimental parameters were as follows: 50× objective, 60 s exposure time, three exposures, 1200 lines per mm grating, and 2.5 mW laser power level. The minimum lateral and depth resolutions were set to a few micrometers. The system was calibrated using the 520.6 cm⁻¹ Raman band of silicon before each experimental session. The thermal damage of the measured points was excluded by visual inspection of the excited surface after measurement, by the observation of the possible decay of spectral features at the start of excitation, and by checking for thermal downshift of Raman lines.

Quantitative chemical analyses were carried out using a Cameca SX 100 electron microprobe (WDS mode) on a polished crystal surface at the National Museum, Prague, Czech Republic. Delchiaroite was unstable under the electron beam, and different conditions were tried. The better experimental conditions were the following: accelerating voltage of 15 kV, beam current of 2 nA, and beam size of 2 µm. Sulfur decreased during the experiment, and we used the time-dependent measurement methodics (S is measured for a total of 20 s, in separate intervals of 4 s, and the result is recalculated to the time 0). The standards (element, emission line) were as follows: chalcopyrite (Cu Kα, S Kα) and Tl(Br,I) (I Lα). Matrix correction by means of a PAP procedure (Pouchou and Pichoir, 1985) was applied to the data. The results of the chemical analyses (average of 12 spot analyses) are given in Table 2. As there is not enough pure material available for a direct determination of C and H, their amount was calculated based upon the assumed stoichiometry.

Intensity data were collected using a Bruker D8 Venture diffractometer equipped with an air-cooled Photon III CCD detector and microfocus Mo Kα radiation (CISUP, University of Pisa, Italy). The detector-to-crystal distance was 38 mm. Data were collected using ω and φ scan modes, in 0.5° slices, with an exposure time of 120 s per frame. A total of 1601 frames were collected. The frames were integrated with the Bruker SAINT software package using a narrow-frame algorithm. Data were corrected for Lorentz polarization, absorption, and background using the Apex4 software package (Bruker AXS Inc., 2022). Unit-cell parameters were refined on the basis of the XYZ centroid of 1326 reflections above 20 σ(I) with 4.813°<2θ<44.56°. They are a=16.924(10), b=4.099(2), c=5.572(3) Å, and V=386.5(4) ų.

The statistical tests on the distribution of |E| values (|E2-1| = 0.724) suggested the absence of a center of symmetry. According to systematic absences, the crystal structure of delchiaroite was solved in some different orthorhombic space groups using Shelxtl (Sheldrick, 2015a). A solution was found in the acentric Pmn2₁ space group, with unit-cell parameters transformed with respect to those given above according to the matrix [0 1 0 | 0 0 1| 1 0 0]. However, the structure refinement indicated the occurrence of racemic twinning, with a twin component ratio of 0.5(5). Moreover, a check with PLATON (Spek, 2009) suggested a higher symmetry due to the occurrence of a center of symmetry. For this reason, the solution in the centric space group Pmmn was proposed, and the crystal structure was refined using Shelxl-2018 (Sheldrick, 2015b). Neutral scattering curves, taken from the International Tables for Crystallography (Wilson, 1992), were used. The structure solution allowed us to locate the position of one I atom, two Cu atoms, one S atom, and one C one. After several cycles of isotropic refinement, R1 converged to 0.1873, with high-density residuals around I. Modeling the atomic displacement parameters of Cu, S, and I as anisotropic, the R1 factor was lowered to 0.1138. Some residuals around the C atom suggested the presence of H atoms, but the CH₃ group was modeled using the AFIX137 instruction (Sheldrick, 2015b). The R1 factor was slightly lowered to 0.1125. The single-crystal X-ray diffraction pattern of delchiaroite showed the contribution of several misaligned grains, and an examination of the list of the most disagreeable reflections revealed the occurrence of six reflections with Fo2 higher than Fc2. After the omission of the first two most disagreeable reflections, the R1 was lowered to 0.1001, further lowered by the omission of four further reflections. In the last stage of the crystal structure refinement, the positions of H atoms were fixed because their refinement did not allow for a correct convergence of the structural model, probably due to their disordered arrangement around the C atom. After several cycles of anisotropic refinement for all atoms but the H ones, the conventional R1 factor converged to 0.0962 for 360 unique reflections with Fo>4σF and 27 refined parameters. Two relatively high residuals around Cu(1) and I were interpreted as being due to the bad quality of the diffraction data (Rint=0.1227). Even if the model is presented with H atoms, it should be noted that these latter atoms are too close to C (around 0.96 Å instead of ∼ 1.08 Å – e.g., Marynick and Dixon, 1977), resulting in overbonding of C atoms (see below). Removing H atoms, the refinement converged to R1=0.0968.

Details of the data collection and crystal structure refinement are given in Table 3. Atom coordinates and isotropic or equivalent isotropic displacement parameters are reported in Table 4, whereas Table 5 gives selected bond distances. Bond valence calculation, shown in Table 6, was performed using the bond parameters of Brese and O'Keeffe (1991).

Owing to the very small amount of pure material and its small size, X-ray powder diffraction data for delchiaroite were collected using a Bruker D8 Venture single-crystal diffractometer equipped with a Photon III CCD area detector (microfocus CuKα radiation) simulating a Gandolfi-like geometry. The observed X-ray diffraction lines are reported in Table 7, along with the calculated pattern based on the structural model discussed below. Unit-cell parameters, refined using the software UnitCell (Holland and Redfern, 1997) on the basis of 26 unequivocally indexed reflections, are a=16.8721(19), b=4.0992(4), c=5.5841(9) Å, and V=386.21(6) ų.

The Raman spectrum of delchiaroite is shown in Fig. 4. The strong band at 2902 cm⁻¹, associated with two weaker bands at 2809 and 2974 cm⁻¹, can be due to C–H stretching modes in CH₃ groups (e.g., Allkins and Hendra, 1966). No distinct bands related to O–H modes were observed. The band at 1308 cm⁻¹ is due to the symmetrical bending of CH₃ groups, whereas weak bands at 945 and 1082 cm⁻¹ can be related to CH₃ rocking modes (e.g., Scott and McCullough, 1958; Meyer, 1992). The relatively strong band at 685 cm⁻¹ is due to C–S single bonds (e.g., Scott and McCullough, 1958; Sexton and Nyberg, 1986; Meyer, 1992). Bands at wavenumbers lower than 500 cm⁻¹ are probably related to Cu–S and Cu–I modes, as well as to lattice modes. Weak bands at 332 and 464 cm⁻¹ may be due to Cu–S bonds (e.g., Mernagh and Trudu, 1993; Andrew et al., 1994). Finally, it is worth noting that the strong Raman band at 129 cm⁻¹ reminds us of that observed at 122 cm⁻¹ in the spectra of marshite, CuI (Lafuente et al., 2015) and in synthetic CuI (Li et al., 2022), as well as the band at 128 cm⁻¹ reported by Welch et al. (2016) in siidraite, Pb₂Cu(OH)₂I₃.

The collection of quantitative chemical data for delchiaroite by an electron microprobe was a very difficult task for several reasons. The first problem is related to the very limited amount of available material and its very small size, which made difficult the preparation of polished surfaces for chemical analyses. However, we were successful, and we obtained some very small flat surfaces on which it was possible to perform several spot analyses. Unfortunately, other issues remained. In some cases, owing to the extreme thinness of the acicular crystals of delchiaroite, the electron beam interacted with epoxy resin, thus lowering the total amount of the analyzed elements. Moreover, the mineral proved to be highly unstable under the electron beam, and several tests were conducted under various conditions, allowing us to find the better analytical conditions. However, S volatilization occurred, and a time-dependent measurement protocol was applied. Notwithstanding these issues, the agreement between the analytical Cu:I:S ratios and those derived from the structural data, together with the Raman evidence (C–H and C–S vibrational modes), lead to the overall consistency of the results.

The empirical formula of delchiaroite, calculated on the basis of Cu + I + S = 6 atoms per formula unit, is Cu_3.07(8)I_1.01(7)(CH₃S)_1.92(8). The ideal, end-member formula is Cu₃I(CH₃S)₂, corresponding to (in wt %) Cu 46.30, S 15.57, H 1.47, C 5.83, and I 30.82.

The crystal structure of delchiaroite is shown in Fig. 5. It can be described as being formed by chains of Cu(1) and S atoms, running along b and decorated by CH₃ groups, giving the electroneutral formula [Cu(1)CH₃S]⁰. Every chain is bonded to a symmetrical one along a through Cu(2) atoms, forming ribbons with the chemical composition [Cu(1)₂Cu(2)(CH₃S)₂]¹⁺. These ribbons are connected, along c, by rows of tetrahedrally coordinated I atoms, forming electroneutral {100} layers of composition [Cu(1)₂Cu(2)I(CH₃S)₂]⁰. Finally, these layers are stacked along a, very probably through van der Waals interactions, which would permit a {100} perfect cleavage.

Copper is hosted at two symmetry-independent sites, namely Cu(1) and Cu(2). The former has a 3-fold, triangular coordination, forming two Cu–S (at 2.234 Å) and one Cu–I (at 2.617 Å) bonds (Fig. 6). Cu(2) displays a distorted tetrahedral coordination, with two Cu–S (at 2.267 Å) and two Cu–I (at 2.861 Å) bonds (Fig. 6). Copper–S distances agree with those observed in Cu(CH₃S), ranging from 2.23 to 2.28 Å (Baumgartner et al., 1993). Relatively short Cu–Cu contacts occur, i.e., Cu(1)–Cu(2) = 2.872(3) Å and Cu(1)–Cu(1) = 2.898 Å; these values can be compared with that reported by Welch et al. (2016) in siidraite, Pb₂Cu(OH)₂I₃, i.e., 2.782 Å. The bond valence sums at Cu(1) and Cu(2) are 1.01 and 0.96 valence units (v.u.), in agreement with the occurrence of Cu⁺ at both sites.

Carbon is hosted at the C site and is bonded to three H atoms and one S atom at 1.80 Å. In synthetic Cu(CH₃S), the C–S distance was fixed at 1.835 Å (Baumgartner et al., 1993). Carbon–H distances (∼ 0.96 Å) are too short with respect to the ideal one of ∼ 1.08 Å (e.g., Marynick and Dixon, 1977) reported in CH₃ groups. This results in a severe overbonding of C (5.44 v.u.); on the contrary, by using an ideal C–H distance of 1.08 Å, the bond valence sum at the C atom is 4.24 v.u., in agreement with the presence of C⁴⁺.

Iodine is tetrahedrally coordinated by Cu atoms hosted at two Cu(1) sites and two Cu(2) sites (Fig. 6), with an average bond distance of 2.739 Å. Its bond valence sum is 0.88 v.u. In marshite, the observed Cu–I distance is 2.625 Å (Cooper and Hawthorne, 1997), whereas Welch et al. (2016) found average < Cu–I > distances of 2.655 and 2.634 Å in siidraite.

Sulfur is bonded to three Cu atoms, i.e., two at Cu(1) and one at Cu(2), and to one C atom in a distorted S(Cu₃C) tetrahedral arrangement (Fig. 6), similarly to what was observed in synthetic Cu(CH₃S) by Baumgartner et al. (1993). As stated above, interlayer van der Waals bonding probably occurs between {100} layers, as suggested by H⋅⋅⋅S distances (3.11 Å) slightly longer than the sum of the van der Waals radii of these two chemical species (3.00 Å; e.g., Fargher et al., 2022).

Taking into account site multiplicity, the formula of delchiaroite obtained through the crystal structure refinement is Cu₃I(CH₃S)₂ (Z=2). This formula can be compared with the results of chemical analysis. The Cu:I:S atomic ratio – 3:1:2 according to the structural investigation – is in agreement with chemical data, giving the ratio 3.04:1:1.90.

A supergene origin of the delchiaroite-bearing assemblage also involving native copper, zincolivenite, lavendulan, and theisite is suggested, being formed at the expense of altered enargite. Supergene minerals found in the cavities of the Marble Formation originated through the interaction between primary sulfides and low-T meteoric fluids whose percolation is favored by widespread karstic phenomena. Following Pekov et al. (2011), low-T conditions are necessary to form iodide species (T<50–70 °C) because, at higher T, iodine is highly mobile. The coexistence of native copper, along with Cu²⁺-bearing lavendulan, zincolivenite, and theisite, as a result of the alteration of Cu⁺-bearing enargite indicates peculiar redox conditions. In fact, delchiaroite is a Cu⁺-bearing phase.

It is also worth noting that there is the occurrence of a Cl-bearing phase (lavendulan) in close association with delchiaroite, an indication of the circulation of halogen-bearing fluids. Our new evidence is not a single anomaly as four halogen-bearing species are currently known from the vugs of the Marble formation at Carrara: connellite, lavendulan, mimetite, and vanadinite (Biagioni et al., 2019, and references therein). These species contain Cl, whereas delchiaroite is the first I-bearing species.

Dealing with the possible origin of I, S, and organic components constituting delchiaroite, both intraformational and extraformational sources could have been involved based on the existing evidence.

The Carrara marble is characterized by the smell of rotten eggs when broken. Usually, such a smell is interpreted as being due to the occurrence of H₂S. This is supported by high sulfate concentrations measured in leachates of fluid inclusions hosted by the Carrara marble, as reported by Prochaska (2023), which may result from oxidation of H₂S (or other reduced-S species) by atmospheric O₂ during the preparation of the leachate. In addition, the halogen systematics of these leachates indicate extraordinary enrichment of included fluids in both Br and I (Fig. 7). This feature is likely to be related to thermal conversion of organic matter into graphite (present in some marble horizons) during regional metamorphism, which was associated with mobilization of I and Br (originally bound to organic matter) into metamorphic fluids. The host Carrara marble can thus represent a fertile local source of both I and S. On the contrary, it is probable that the contribution of a thermally immature external source is necessary for the methanethiolate anion because aliphatic thiols decompose at temperatures above ca. 170–250 °C (Yang et al., 2006), and the peak temperatures of the metamorphic overprint of the Carrara marble were between 350 and 450 °C (e.g., Carmignani and Kligfield, 1990; Molli et al., 2000, 2002).

Three potential sources of methanethiolate can be suggested in the study area. The first one can be represented by sedimentary rocks of the Tuscan Nappe, a tectonic unit overlying the Carrara marble. A hydraulic connection of both of these formations was already suggested by Costagliola et al. (1999) to explain the high salinity of late-stage fluid inclusions in hydrothermal veins hosted in the Carrara marble. Furthermore, the participation of late fluids dominated by meteoric waters during the formation of hydrothermal veins in Carrara marble (Costagliola et al., 1999; Molli et al., 2010) gives evidence for the long-lasting existence of migration paths from the surface into the marble body and supports the possibility of the transfer of some components leached from overlying sediments.

The second potential source of organic thiols could be represented by organic-rich soil layers (e.g., Rao et al., 2014) as these still cover marble outcrops outside the quarry (Fig. 1). Moreover, organic-rich soils can retain and store iodine as organo-iodine compounds (e.g., Keppler et al., 2004). These chemical species may then be leached by meteoric waters, percolate along karstified fractures, and interact with weathering enargite.

The third possible source may be found in marine aerosols in rain. The Apuan Alps rise abruptly from the narrow coastal plain bordering the Ligurian Sea and are characterized by high precipitation (>2500 mm yr⁻¹; Rapetti and Vittorini, 1994). Oceanic phytoplankton is the source of dimethyl sulfide (CH₃SCH₃), which is one of the most important sulfur species released into the atmosphere (e.g., Martínez et al., 2000). However, it is probable that this trace chemical species would be largely oxidized during atmospheric transport, thus barely representing the actual source for the crystallization of delchiaroite. Nevertheless, the marine aerosols may have contributed both Cl and I (e.g., Sturges and Barrie, 1988); marine water as a possible source of halogens in the genesis of halides has been proposed by other authors (e.g., Whiterhead, 1919; Andreu et al., 2015).

In summary, available evidence supports the possibility of local sources for I, S, and hydrocarbon compounds. However, as delchiaroite is a very late-stage mineral, formed though supergene processes, the choice of a unique source for iodine and organic compounds is not straightforward, and further research is necessary.

A final consideration is as follows: iodides rapidly decompose under the effect of sunlight and atmospheric O₂ (e.g., Pekov et al., 2011). The specimen where delchiaroite was identified was collected by the quarryman Paolo Cagnoni during an exploitation activity and was not found in the quarry dump, where the mineral may have rapidly decomposed and/or oxidized. It is possible that this immediate sample collection after the extraction could have allowed for the preservation of delchiaroite. Considering the enrichment in I of the Carrara marble shown by the results of Prochaska (2023), as well as their possible ephemeral nature, the presence of so-far unrecognized iodide minerals among secondary mineral assemblages at Carrara cannot be excluded.

Natural iodide–chalcogenides are known, i.e., demicheleite-(I), BiSI (Demartin et al., 2010); hanauerite, AgHgSI (Pekov et al., 2023); mutnovskite, Pb₂AsS₃(I,Cl,Br) (Zelenski et al., 2006); perroudite, Hg₅Ag₄S₅(I,Br)₂Cl₂ (Sarp et al., 1987); and radtkeite, Hg₃S₂ICl (McCormack et al., 1991). Moreover, two copper iodides are known, namely marshite, CuI (Cooper and Hawthorne, 1997), and siidraite, Pb₂Cu(OH)₂I₃ (Rumsey et al., 2017).

Copper is also involved in the crystallization of several organic minerals (Table 8). Among the 12 currently known species, 10 have been described in the last 25 years, and 8 were found after 2010. However, none of them contain S as an essential constituent. Indeed, delchiaroite is the 13th organic mineral containing Cu, and it is the first natural Cu iodide–organochalcogenide, showing a novel crystal structure. Synthetic halo-thio-organometallic compounds are actively studied for their luminescence properties (e.g., Knorr et al., 2014; Rajput et al., 2015).

In addition, the synthetic phase [Cu(CH₃S)]_∞ is known (Baumgartner et al., 1993). It was obtained through thermal decomposition of [C₃H₇N][Cu₄(CH₃S)]₆ ⋅ CH₄O, whereas other thiolates described by Baumgartner et al. (1993) were obtained from solutions kept at 55 °C for 1 h and then evaporated at room temperature. Thus, all of these compounds seem to be low-T phases. The synthetic Cu(CH₃S) phase has a unit-cell volume of V=285.95 ų (Z=4). Delchiaroite, Cu₃I(CH₃S)₂, can be considered to be a combination of CuI + 2 Cu(CH₃S). Copper iodide is known as the mineral marshite, having a unit-cell volume of V=221.45 ų (Z=4). Thus, the volume Vmix of CuI + 2Cu(CH₃S) would be 1/4VCuI+2/4VCu(CH3S)=198.34 ų. Delchiaroite has Z=2, and, thus, the volume of this mineral (i.e., 386.5 ų) has to be compared with 2Vmix=396.7 ų.