The lillianite homologous series (LHS), defined by Makovicky and Karup-Møller (1977a, b), contains structures composed of (311)_PbS slabs of galena-like structure. Adjacent slabs are in a mirror-related orientation; in most cases these adjacent slabs are of the same thickness, but phases with different thicknesses of adjacent mirror-oriented slabs are known as well, resulting in their regular alternation (Ferraris et al., 2008; Makovicky, 2006).

The order of the homologue is determined by the number of cation coordination octahedra in a string running diagonally through an individual slab. The Pb–Ag–Bi sulfides start with N= 4, and known cases reach N= 11. The general formula of these homologues is Ag_xPb_N−1−2xBi_2+xS_N+2 (Z= 4), where N is the “chemical” value of N that for ideal chemical analysis is equal to (N1+N2)/2, and N1 and N2 are the structural values of N for the two sets of slabs related by (partial) reflection across composition planes and can be equal (N1=N2) or different (N1≠N2). The subscript x in the formula is the coefficient of Ag + Bi ↔ 2Pb substitution (Makovicky and Karup-Møller, 1977a). Hence, the lillianite homologue can be written as ^N1,N2L when N1 and N2 are known from structure study or as ^N_chemL when only chemical data are known. The molar fraction of the Ag–Bi end member (due to 2Pb²⁺ for the Ag++ Bi³⁺ substitution mechanism) is defined as substitution percentage (Subst. %). The general formula can be recast into an equation giving the value of Nchem, based on empirical contents of Pb(Cd), Bi(Sb), and Ag(Cu) as a tool for mineralogical determination of lillianite homologues (Makovicky and Karup-Møller, 1977a, b; Ferraris et al., 2008). When, as is often the case, only chemical data are available, the studied grain can be characterised as ^N_chemL_Subst. The analytical formulae for Nchem, Subst. %, and x are given in Makovicky and Karup-Møller (1977a). The above-mentioned observations by Makovicky and Karup-Møller (1977b) led to a limited spectrum of lillianite homologues known in the Pb–Bi–Ag–S system, ^4,4L, ^4,7L,^7,7L, ^5,9L, ^4,8L, and ^11,11L, which were either confirmed by structure determination or derived by strict geometrical considerations (Table 1.1 in Ferraris et al., 2008).

In the last decade, a synthetic ⁸L₇₀ phase was described by Topa et al. (2010), and a review of the LHS was given by Makovicky and Topa (2014). Recently, a new mineral, tarutinoite – a natural ^7.5L₅₄ member of the series – was described by Kasatkin et al. (2025).

During the detailed investigation of the material from the Erzwies mining area, we found a phase with chemistry and crystal structure close to those of the synthetic orthorhombic N= 8 member of the lillianite homologous series. The new phase described below increases the number of minerals with chemistry and crystal structure related to the lillianite homologous series.

The name erzwiesite (Erw) is for the middle-age gold mining place, the Erzwies mining area, Gastein Valley, Salzburg Province, Austria.

The mineral and mineral name have been approved by the Commission on New Minerals Nomenclature and Classification (CNMNC) of the International Mineralogical Association (IMA 2012-082, Topa et al., 2013). The holotype specimen, consisting of a hand specimen and a polished sample, was deposited in the mineralogical collection of the Natural History Museum, Wien, Austria, under catalogue no. O 2407.

Erzwiesite occurs as a very rare constituent in several samples from an unnamed prospect at 2390 m above sea level in the Erzwies mining district, Gastein Valley, Salzburg Province, Austria (47°5^′40^′′ N, 13°2^′15^′′ E). Au- and Ag-bearing mineralisation in this area is linked to the so-called “Kupelwieser” vein, which represents one of the NNE-striking structures of the “Tauerngold vein” type in the Rauris–Gastein mining district. The abovementioned structure is hosted by either Variscan granitoids (the so-called “Zentralgneis”) or metasediments of Jurassic age; the latter is characterised by numerous karst phenomena (Höfer, 2001, 2005). The “Kupelwieser” structure is accompanied by numerous NE–SW-striking quartz veins, which locally contain highly interesting Ag–Bi–Pb sulfosalt mineralisation (e.g. Paar, 2006).

The region of Erzwies (“ore meadow”) was part of an important ore mining district in historical times. It is located north of the Baukarlscharte, on the western side of the Gasteiner Valley in Salzburg Province of Austria. The whole district (Siglitz–Bockhart–Erzwies) comprises an area of 5×2 km (Fig. 1). Mining for gold and silver dates back to the 14th century and even before that time (Gruber, 2006). In the 18th and 19th centuries, less important mining for lead, zinc, and iron occurred. Numerous adits, most of them collapsed, and extensive dumps are nowadays a testimony of extensive mining, which was done both from the surface and from underground.

The vein type mineralisation of the epi- to meso-thermal type was formed at the retrograde stage of the Alpine metamorphism at temperatures at and below 300 °C. A more detailed description is given in Paar (2006). Various ore types can be distinguished in the area of Erzwies. The prevailing mineralisation is (1) massive ore composed of sphalerite (coarse-grained, variously coloured); galena (with freibergite, polybasite, acanthite); and pyrite, particularly in banded sequences within a carbonate gangue (Ca–Mg–Fe–Mn carbonates). The other type (type 2) is composed of pyrite, arsenopyrite, and gold, frequently in brecciated textures together with quartz.

A third ore type (3) is represented by E–W trending sulfosalt-bearing veins of quartz which penetrate metagranitic rocks (the so-called “Central Gneiss”; Fig. 2). More than 70 veins and veinlets of this type are exposed at altitudes between 2200 and 2440 m a.s.l. and can be traced for at least several metres along strike. Their thicknesses can vary between a few centimetres and almost half a metre. A common associate is pyrite in cubes up to 10 cm and in a fine-grained, almost powdery variety. At least five of these quartz veins carry a rich sulfosalt mineralisation.

At Erzwies the sulfosalt minerals occur as needle-like single or bladed crystals and massive forms frequently intergrown as radiating aggregates. The size of the sulfosalt inclusions extend up to almost 10 cm. On the basis of a very detailed analytical study, the sulfosalt assemblage of these quartz veins (type 3) consists of bismuthinite, gladite, krupkaite, silver-rich members of the lillianite homologous series (such as vikingite, (?)treasurite, heyrovskýite, eskimoite, and ourayite), pavonite-type phases, cosalite, neyite, tetradymite, gold, and galena-matildite.

One of these veins contained dantopaite, a member of the pavonite homologous series (Makovicky et al., 2010; small exploration trench at 2275 m a.s.l.). Another quartz vein exposed at approx. 2390 m a.s.l. carries abundant sulfosalts, including the new mineral erzwiesite. A detailed treatment of the sulfosalt mineralogy at Erzwies is being prepared and will be published elsewhere.

Erzwiesite occurs together with other sulfosalts and sulfides within a single quartz vein of ore type 3 and was found only in a few specimens. This quartz vein is exposed at elevations between 2390 and 2395 m above sea level to the west of what had probably been the largest mine of this area, called “Blei Beul” or “Beim Bleibau”. The outcrop of the vein is on a steep slope, where it was explored for a few metres along strike and depth. The width of the vein varies between a few centimetres up to 12 cm. The quartz is highly fractured and limonite-stained. The ore mineralisation occurs irregularly distributed within the quartz, where it forms nests of intergrown sulfosalts up to a maximum size of 5 cm (Fig. 3). In hand specimens, the dominant minerals are bismuthinite and members of the lillianite and pavonite homologous series. Their distinction by the naked eye is based on the good cleavage of bismuthinite and the poor cleavage of the associated sulfosalts. Powdery pyrite is the dominating associate.

The accompanying sulfosalts are all Bi rich and belong to the aikinite–bismuthinite series (aikinite, krupkaite) and lillianite homologous series (gustavite, vikingite, heyrovskýite, eskimoite, ourayite). Minor associates are gold (electrum), tetradymite, and galena-matildite solid solution.

Erzwiesite occurs as black anhedral crystals, with a black streak. It is non-fluorescent, opaque in transmitted light, and presents a metallic luster. It is brittle, no cleavage is observed, and the fractures are irregular. The calculated densities for empirical formulae (Z= 1) and unit cell parameters (from SCXRD studies) and for the crystal-structure refinement formula are 7.045 and 7.074 g cm⁻³, respectively. Mohs' hardness is 3–3.5 (VHN₅₀ ranges from 195 to 224, mean 210 kg mm⁻²). Erzwiesite has a greyish colour in reflected light. Pleochroism is distinct white to dark grey, and bireflectance is weak to distinct (in oil). Anisotropy is distinct in air and becomes enhanced in oil. Between crossed polars, rotation tints change from pale brown to pale bluish grey to dark brown and are thus very similar to that of other members of the LHS. Twinning is not observed.

Reflectance value measurements taken at extinction positions in air (WTiC standard, area measured 20×20 µm), acquired with a Leitz MPV-SP microscope photometer, are given in Table 1 and are plotted in Fig. 4.

Chemical analyses of erzwiesite and of the associated minerals were performed at the Central Research Laboratories of the Natural History Museum, Vienna, on a JEOL “Hyperprobe” JXA 8530F field-emission gun electron microprobe (FE-EPMA). Analytical conditions were WDS measuring mode, accelerating voltage of 25 kV and 20 nA, beam current, 1.5 µm beam diameter, and count time of 15 s on peak and 5 s on background positions. JEOL and a probe for EPMA software were used for acquisition and processing of data. The following emission lines and standards were used: AgLα (Ag metal), PbLα (galena), BiLα and SKα (Bi₂S₃), SbLα (stibnite), CdLα and TeLα (CdTe), CuKα (chalcopyrite), and SeLα (Bi₂Se₃). No other elements with Z>8 were detected. Under the analytical conditions described, the detection limits for the measured elements in the erzwiesite matrix were as follows (expressed in wt %): S and Cu ∼ 0.03; Ag, Cd, Sb, Te, and Se ∼ 0.06; and Pb and Bi ∼ 0.09.

Analytical data, as average chemical compositions for a specific number of measured points (NA) of four different erzwiesite grains from different areas (grain 1 to grain 4) of several polished sections, as shown in Fig. 5a, are presented in Table 2, together with the empirical formulae based on 76 apfu (36Me + 40S) and with the chemistry resulting from structural formula (SREFF) obtained from the SCXRD study. The ideal formula (IF) with resulting chemistry and calculated parameters, Nchem, Subst.(%), and x (Makovicky and Karup-Møller, 1977a), for all groups are also shown.

The amounts of Cu, Cd, Sb, Te, and Se are small but above detection limits. The empirical formula for the grain used for the SCXRD study is Ag_8.18Cu_0.05Pb_11.31Cd_0.12Bi_16.30Sb_0.01S_39.81Te_0.13Se_0.06, calculated based on 76 apfu. The assumption that Cu⁺, Cd²⁺, Sb³⁺, Te²⁻, and Se²⁻ can replace Ag⁺, Pb²⁺, Bi³⁺, and S²⁻, respectively, allows us to write the simplified formula as (Ag,Cu)_8.23(Pb,Cd)_11.43(Bi,Sb)_16.31(S,Te,Se)_40.03. The ideal formula, taking into account the simplified formula and the results of the SCXRD, is Ag₈Pb₁₂Bi₁₆S₄₀, which requires (in wt %) Ag 10.82, Pb 31.76, Bi 41.92, and S 16.08, totalling 100 00. Erzwiesite from Erzwies is represented by a range of compositions with a substitution percentage of ^8.01L_55.6 to ^7.91L₆₉.

Other minerals associated with erzwiesite in the polished section of Fig. 5 are treasurite ^6.01(15)L_61.15(50), with empirical formula Ag_4.51Cu_0.4Pb_10.55Cd_0.11Bi_12.88Sb_0.06S_31.88; heyrovskýite ^6.88(10)L_59.35(4), with empirical formula Ag_5.551Cu_0.35Pb_12.09Cd_0.07Bi_13.99Sb_0.07S_35.88; and galena, with empirical formula Ag_0.08Pb_0.85Bi_0.09S_0.99, carrying notable amounts of 3.5 Ag and 7.7 Bi (in wt %).

X-ray intensity data of a small (0.035×0.045×0.09 mm) needle-shaped crystal of erzwiesite were collected at room temperature using a Bruker AXS three-circle diffractometer equipped with a CCD area detector. The SMART (Bruker AXS, 1998a) system of programs was used for unit cell determination and data collection, SAINT+ (Bruker AXS, 1998b) for the reduction of the intensity data, and XPREP (Bruker AXS, 1997) for space group determination and empirical absorption correction based on pseudo-ψ scans. The space group is Cmcm, as proposed by the XPREP program, and is in accordance with the space group of all orthorhombic homologues of lillianite. The initial model was based on the coordinates and atom names of erzwiesite synthetic analogue ⁸L₇₀ (Topa et al., 2010), which contains five cation and six anion sites. The Pb2 and Bi1 sites are pure Pb and Bi sites, respectively, as in synthetic analogue ⁸L₇₀. The Pb1, Bi2, and Bi3 sites are mixed with Ag in a similar way as in synthetic analogue ⁸L₇₀ and were refined using the scattering curves of Pb and Bi against Ag.

The refinement of the structure against F2 using SHELXL (Bruker AXS, 1997) gave the Ag_8.64Pb_9.04Bi_18.32S₄₀ (Z= 1) formula with overestimated Bi and underestimated Pb general content and unbalanced charge values of 2.21 and 1.68 when atomic percent and formula coefficients were used, respectively. The Bi2 site is a mixed site with Bi_0.70(4)Ag_0.30(4) occupancy. Similar results were obtained for the (Bi,Ag)2 site of the synthetic analogue (Bi_0.742(7)Ag_0.258(7)) of erzwiesite (Topa et al., 2010), in which a certain Pb amount was suspected. Therefore, we assumed Bi_0.45(4)Pb_0.25Ag_0.30(4) occupancy for the erzwiesite Bi2 site [Bi_0.70(4)Ag_0.30(4)], in accordance with the inability to distinguish between Pb and Bi (very close atomic numbers) in the diffraction study, bond length, charge distribution, and bond-valance sum values. With that adjustment, the erzwiesite formula derived from the crystal-structure refinement (SREF) becomes Ag_8.64Pb_11.04Bi_16.32S₄₀, Z= 1, with a better charge balance value of -0.42 and close to the EPMA empirical formula (Table 2). The displacement parameters of cation positions were refined anisotropically, while the displacement parameters of anion positions were refined isotropically due to some negatively refined ones. The structure was refined to R1= 5.24 % for 308 data with Fo > 4σ(Fo) and 51 variable parameters. Experimental and refinement conditions are specified in Table 3. The relatively low resolution of the dataset (2θmax= 41.68°) reflects the improper strategy of data collection for the available crystal from Madoc. Despite this limitation, the refinement converged to a stable and chemically consistent structural model, supported by agreement with chemical data and bond-valence analysis. Fractional coordinates, site occupancies, anisotropic displacement parameters, charge distribution, and bond-valence sum values are compiled in Table 4. Table 5 reports selected cation–anion bond distances. Full details are given in the erzwiesite.CIF, erzwiesite_checkCIF.pdf, and structure factor files, all deposited as a Supplement (Supplement files S1, S2, and S3).

Powder X-ray data were not collected due to erzwiesite admixture with other members of the lillianite homologous series and galena but were calculated on the basis of the single-crystal-structure results. Data are listed in Table 6.

The asymmetric unit of the erzwiesite structure contains five independent coordination polyhedra of cations and six of anions (Fig. 6). It contains the Pb2 site with bicapped trigonal prismatic coordination (CN = 8) in a special position and placed on the reflection planes of the space group. The other four octahedral coordination polyhedra (the Pb1 site and the Bi2 site in the centre and Bi1 site and Bi3 site at the end) form PbS-like (311)_PbS slabs with the same width (Fig. 7). The centre of the slabs contains Pb and more Ag and less Bi (0.88 Pb, 0.67 Ag, and 0.45 Bi) than the Bi-dominant exterior of the slab (1.59 Bi and 0.41 Ag). Adjacent slabs are in a mirror-related orientation. The width of the slab is expressed by the eight metal sites in the chains of octahedra, which run diagonally across individual galena-like slabs and parallel to [011]_PbS (Fig. 7) and defines erzwiesite as the homologue number N= 8 (Table 7) of the lillianite homologous series (Makovicky and Karup-Møller, 1977a, b; Moëlo et al., 2008).

The crystal data of erzwiesite (Table 7) showed same space group and similar unit cell parameters compared with the synthetic analogue ⁸L₇₀. The site occupancies for the Pb1, Bi2, and Bi3 sites in erzwiesite (Table 4) compared with the Pb1, Bi2, and Bi3 sites in the synthetic (Topa et al., 2010) phase ((Pb_0.53Ag_0.47)1, (Bi_0.74Ag_0.26)2, and (Bi_0.56Ag_0.44)3) (Topa et al., 2010) indicate more Pb and less Bi and Ag.

A comparison of ideal formulae, Nchem, the molar fraction of the Ag–Bi end member (Subst. %), and cell parameters for erzwiesite and N= 7.5 and N= 7 members of the lillianite homologous series is shown in Table 7. Aschamalmite (Callegari and Boiocchi, 2009) is an Ag-free (or Ag-low) ⁷L₀ monoclinic N= 7 member of the series, heyrovskýite is an orthorhombic N= 7 member of the series with chemistries varying from Ag-free ⁷L₀ (Pinto et al., 2011) to Ag-bearing ⁷L_52.8 (Makovicky et al., 1991), and eskimoite is an Ag-rich ⁷L₇₀ (Makovicky and Karup-Møller, 1977b) monoclinic N= 7 member of the series. Tarutinoite, recently described by Kasatkin et al. (2025), is monoclinic with a novel configuration of two different slab lengths of N= 7 and N= 8, with 7 (i.e. N1≠N2 and N= (N1+N2)/2 = 7.5) and chemistry characterised by ^7.5L_54.6, similar to the configuration N1≠N2 and N= (4+7)/2 = 5.5 found in vikingite ^5.5L_71.4. Although the ideal formulae (and also structure-derived formulae) for chemistry of erzwiesite (⁸L66.6) and its synthetic analogue (⁸L75) are relatively close to one another, their experimentally measured ranges are different. The synthetic analogue is homogenous (SH Table 2) with very limited variations (Topa et al., 2010), in contrast with the erzwiesite chemistry, which ranges from ^8.01L_55.6 to ^7.91L₆₉.

Besides the Erzwies mining area, erzwiesite was described from the Kutná Hora ore district, Czech Republic, by Pažout (2017), where it contains 1.95 to 2.96 Sb wt % and has a limited chemical variation expressed by ^8.68L_68.8 to ^8.06L_71.8. The degree of its substitution lies between those of synthetic analogue in the upper side and erzwiesite in the lower side of the range. Moreover, a long time after the publication of the sulfosalt chemistry from the Altenberg mining district, Salzburg Province, Austria, by Putz et al. (2003), we realised that regarding the group of chemistries showing N= 7 to N= 12 members of the lillianite homogenous series, there was a group characterised by ^8.09–8.15L_40.1–72.3, representing the first measurements of erzwiesite. Ag-free and Ag-high (more than synthetical analogue) erzwiesite chemistries were not found until now. The grains possessing low substitution values were difficult to extract for the single-crystal study, and, therefore we cannot predict the system of crystallisation and space groups of Ag-free and Ag-high erzwiesite chemistries.

A ternary plot of the system, (Bi + Sb) – (Ag + Cu) – (Pb + Cd), depicting the position of erwiesite (N= 8) and the other known members of the lillianite homologous series is given in Fig. 8. The chemistry of erzwiesite from Altenberg (grey background) exceeds both sides of the chemistry of type erzwiesite from Erzwies but remains under the chemistry of synthetic analogue of erzwiesite.

The plotting process in the ternary diagram of the chemistries obtained by the EMPA method, as well as the values of Nchem provided by them, will together give a first indication of the identification of the homologue number for a mineral belonging to the lillianite homologous series. The quality of Pb and Bi(Sb) element determination is crucial because the calculation of Nchem is based on them. The results of the single-crystal study will provide the real answer about the homologue number and will give the correct identification of the phase. For example, the calculated Nchem for the synthetic ⁸L₇₀ phase is 7.34 (Table 2), in contrast with the result of the single-crystal study, which clearly shows N= 8.