Synthesis of brizziite

Brizziite was synthesized following the procedure of Ramirez-Meneses et al. (Reference Ramirez-Menses, Chavira, Domínguez-Crespo, Escamilla, Flores-Flores and Soto-Guzmán2007). Stoichiometric ratios (1:1) of Na₂CO₃ (4.3757 g, 0.0413 mol) and Sb₂O₃ (12.0400 g, 0.0413 mol) were mixed thoroughly using a mortar and pestle. Some 2.5 g of the mixture were placed in a fused silica boat and heated in a tube furnace at 1033 K ± 10 K for 120 hours. After cooling to room temperature, the product was obtained as a white crystalline powder, which was placed in aqueous 0.05 M Na₂CO₃ for 24 hours to dissolve any residual starting materials. The precipitate was then collected on GF/F grade filter paper, rinsed with water then acetone, and sucked dry.

Powder X-ray diffraction studies were carried out using a Philips PW1825/20 powder diffractometer (Ni-filtered CuKα₁ radiation, λ = 1.5406 Å, 40 kV, 30 mA). Traces were produced between 5–70°2θ, with a step size of 0.02° and a speed of 1.2°min^–1. Diffraction Technology Data processing software (Traces Version 6) and JCPDS-ICDD data base files were used to identify phases and purity.

Solubility of brizziite

Solubility studies were undertaken in sealed 250 cm³ conical Quickfit^R flasks in a 25.0±0.2°C thermostatted water bath. Measurements of pH were made using a Radiometer ION450 apparatus fitted with a combination electrode calibrated as per the manufacturer's instructions.

For the congruent dissolution of brizziite, the carbonate-washed product (~0.1 g) was added to a conical flask with 100.00 cm³ of standardized aqueous 0.0497 M Na₂CO₃. The flasks were left for 6 weeks whilst a paired flask was being monitored periodically for pH (~1 week) until no change was detected.

Equilibrium was achieved after 3 weeks but flasks were left longer to ensure no further changes occurred. Solutions were filtered using Whatman^R GF/F grade filter paper and the filtrates collected in clean PET bottles. Dissolved Sb concentrations were determined by inductively coupled plasma mass spectrometry (ICP-MS) using matched standards.

Synthesis of brizziite

High purity, single phase samples of brizziite were obtained in essentially quantitative yield using the methods described by Ramirez-Meneses et al. (Reference Ramirez-Menses, Chavira, Domínguez-Crespo, Escamilla, Flores-Flores and Soto-Guzmán2007). Unit-cell parameters for brizziite refined using LAPOD (Langford, Reference Langford1973) were a = 5.294(1), c = 15.949(3) Å, and agree well with those reported by Wang et al. (Reference Wang, Chen and Greenblatt1994), a = 5.2901, c = 15.9260 Å.

Solubility of brizziite

After equilibration in 0.0497 M Na₂CO₃, some solid brizziite remained and no other phase was detected by post-solubility powder X-ray diffraction (PXRD) analysis. Thus, for the purpose of solubility calculations we consider brizziite to dissolve congruently. Dissolved concentrations of Sb(V) determined by ICP-MS are reported in Table 1. The congruent dissolution can be expressed by equation 1.

(1)$$\eqalign{&{\rm NaSbO_3\left( s \right) + 3H_2O\left( l \right){\rm \mathbin{\lower.3ex\hbox{$\buildrel\textstyle\leftharpoonup\over {\smash{\rightharpoondown}}$}}} Sb\left( {OH} \right)_{6}^{\ndash} \left( {aq} \right)} \cr & \quad{\rm + Na^+ \left( {aq} \right)}}$$

Table 1. Dissolved metal concentrations for brizziite equilibrations.

The pH at equilibrium (10.890) was used in the determination of solution speciation using COMICS. A reliable lg K value for equation 2 of lg K(298.15 K) = –2.72 at I = 0 mol dm^–3 was reported recently by Accornero et al. (Reference Accornero, Marini and Lelli2008). Correction to I = 0.1491 mol dm^–3 by the method of Baes and Mesmer (Reference Baes and Mesmer1976), yields lg K(298.15 K) = –2.601. Substitution into equation 2 indicates only Sb(OH)₆^–(aq) under the experimental conditions employed. For equation 3 a reliable lg K value of lg K(298.15 K) = –14.2 at I = 0 mol dm^–3 was reported by Baes and Mesmer (Reference Baes and Mesmer1976). Correction to I = 0.1491 mol dm^–3 yields lg K(298.15 K) = –14.32. Thus negligible NaOH⁰(aq) is present. For equation 4 a reliable lg K value of lg K(298.15 K) = –3.67 at I = 0 mol dm^–3 was also reported by Baes and Mesmer (Reference Baes and Mesmer1976). Correction to I = 0.1491 mol dm^–3 yields lg K(298.15 K) = –4.026, indicating that, as expected, CO₃^2–(aq) is the only significant carbonate species in solution at the pH used for the equilibration runs.

(2)$$\eqalign{& {\rm Sb\left( {OH} \right)_{5}^{0} \left( {aq} \right){\rm} + {\rm} H_2O\left( l \right){\rm \mathbin{\lower.3ex\hbox{$\buildrel\textstyle\leftharpoonup\over {\smash{\rightharpoondown}}$}}} Sb\left( {OH} \right)_{6}^{\ndash} \left( {aq} \right)} \cr & \quad {\rm + H^+ \left( {aq} \right)}}$$

(3)$$ \rm Na^ + \left( {aq} \right){\rm} + {\rm} H_2O\left( l \right){\rm \mathbin{\lower.3ex\hbox{$\buildrel\textstyle\leftharpoonup\over {\smash{\rightharpoondown}}$}}} NaOH^0\left( {aq} \right){\rm} + {\rm} H^ + \left( {aq} \right)$$

(4)$$ \rm CO_{3}^{2\ndash} \left( {aq} \right){\rm} + {\rm} H_2O\left( l \right){\rm \mathbin{\lower.3ex\hbox{$\buildrel\textstyle\leftharpoonup\over {\smash{\rightharpoondown}}$}}} HCO_{3}^{\ndash} \left( {aq} \right){\rm} + H^ + \left( {aq} \right)$$

Individual ion activity coefficients were calculated using the Davis extension of the Debye-Hückel equation for 298.15 K, lg γ = –0.5085z ²(√I/(1+√I) – 0.3I). For I = 0.1491 mol dm^–3, γ^2± = 0.335 and γ^± = 0.761; γ^o is taken to be unity. The activities of a(Sb(OH)₆^–) and a(Na⁺) were calculated from the solubility data, yielding a value of lg K for equation 1 of –6.70 ± 0.10 corresponding to ΔG_r^ө = 38.26 ± 1.1 kJ. Use of the appropriate data in Table 2 for equation 1 give ΔG_f^ө(NaSbO₃, s, 298.15 K) = –806.66 ± 2.5 kJ mol^–1. The estimated error takes into account the analytical error of the solubility experiments, errors quoted for the thermochemical data used, and an estimated error of ± 1.0 kJ mol^–1 for ΔG_f^ө(Sb(OH)₆^–, aq, 298.15 K).

(5)$$ \rm NaSbO_3\left( s \right){\rm} + {\rm} 3H_2O\left( l \right){\rm \mathbin{\lower.3ex\hbox{$\buildrel\textstyle\leftharpoonup\over {\smash{\rightharpoondown}}$}}} Na\left[ {Sb{\left( {OH} \right)}_6} \right]\left( s \right)$$

Table 2. Thermodynamic quantities (ΔG_f^ө) used in the calculations (T = 298.15 K).

The relationship of brizziite and mopungite can be described by equation 5. We reiterate that only the single known locality of brizziite reports an association of the two minerals, but mopungite is also reported from a further six localities. The association highlights the likelihood of kinetic effects being important for their coexistence.

Using the derived ΔG_f^ө value reported by Blandamer et al. (Reference Blandamer, Burgess and Peacock1974) for mopungite, ΔG_f^ө(Na[Sb(OH)₆], s, 298.2 K) = –1508.5 kJ mol^–1, ΔG_r^ө(298.2 K) for equation 5 is + 9.46 kJ. Thus, the thermodynamically stable phase at 298.2 K is brizziite and kinetic effects must account for the existence of mopungite, unless the stabilities of one or both of the minerals are influenced profoundly by temperature. In line with our previous findings of the Mg antimonate phases brandholzite and byströmite (Roper et al., Reference Roper, Leverett, Murphy and Williams2014), it must simply be the case that the rate of transformation of mopungite to brizziite is very slow at 298.2 K. The lack of brizziite in any of the six remaining localities hosting mopungite supports this notion.

(6)$$ \rm 2NaSbO_3\left( s \right) + {\it M}^{2 +} \left( {aq} \right){\rm \mathbin{\lower.3ex\hbox{$\buildrel\textstyle\leftharpoonup\over {\smash{\rightharpoondown}}$}}} {\it M}^{2 +} Sb_2O_6\left( s \right){\rm} + {\rm} 2Na^ + \left( {aq} \right)$$

The relationships of brizziite with bystromite, ordoñezite and rosiaite are described by equation 6 where M ²⁺ = Mg²⁺, Zn²⁺ or Pb²⁺ respectively. Using the appropriate data from Table 2 gives ΔG_r^ө values of –8.38, –19.38 and –39.68 kJ and subsequent lg K values of + 1.47, +3.39 and + 6.95 for M ²⁺ = Mg²⁺, Zn²⁺ or Pb²⁺ respectively. In accordance with equation 6, when Na⁺(aq) activity = 10^–0.5 the formation of brizziite requires activities of Mg²⁺, Zn²⁺ and Pb²⁺ less than 10^–2.47, 10^–4.39 and 10^–7.95, respectively, to transform brizziite to the divalent metal antimonates. Thus, the rarity of brizziite is readily explained. Competing cations must be essentially absent from mineralizing solutions, save for Mg²⁺(aq). This is certainly the case for Pb²⁺(aq) (Roper et al., Reference Roper, Leverett, Murphy and Williams2014). Furthermore, local Ca²⁺(aq) concentrations would need to be quite low to prevent the formation of minerals of the roméite group, such as Ca₂Sb₂O₇, and at lower pH values than used in the solution experiments, Fe³⁺(aq) must be negligible to prevent the formation of tripuhyite, FeSbO₄ (Leverett et al. Reference Leverett, Reynolds, Roper and Williams2012).

A definite relationship exists between brizziite/mopungite and the roméite group of minerals, reported commonly as associated phases in the deposits listed above. However, due to lack of reliable thermochemical data for the end-members of the roméitegroup details of chemical relationships remain sketchy. It is also worth noting that in initial dissolution experiments 0.1 M HNO₃ was employed and this resulted in incongruent dissolution of brizziite and the formation of phases with the pyrochlore structure (roméite group), as determined by PXRD. It thus appears that pH values somewhat >7 are also necessary for the formation of either brizziite or mopungite.

Some further comment on associated phases is warranted here. It is likely that the oxidation of ottensite, a sodium-bearing thioantimonate(III) of formula (Na,K)₃Sb₆(SbS₃)O₉·3H₂O, gives rise to mopungite under oxidizing conditions, with brizziite eventuating, based on the rate of transformation kinetics. This suggestion is supported by the noted intimate association of ottensite and mopungite (Origlieri et al., Reference Origlieri, Laetsch and Downs2007) combined with rapid and localized oxidation characteristic of stibnite. The association therefore is likely to be an indicator of lower redox potentials than those which must have prevailed at Green prospect, Nevada, USA; here Williams (Reference Williams1985) reported associations with senarmontite, stibiconite, roméite and tripuhyite. Furthermore, the majority of deposits which carry mopungite also have an association with cetineite in the absence of the roméite-group minerals. Ottensite, the Na-analogue of cetineite, was described only recently (Sejkora and Hyršl, Reference Sejkora and Hyršl2007) and as such we suggest that reported associations of mopungite and cetineite may have overlooked in some cases the presence of ottensite. This is supported by the noted association of mopungite with cetineite from a number of Tuscan deposits, namely the Pereta mine (Marzoni Fecia Di Cossato et al., Reference Marzoni Fecia Di Cossato, Meacci, Orlandi and Vezzalini1987), Le Cetine Mine (Brizzi et al., Reference Brizzi, Ciselli and Santucci1988; Preite, Reference Preite1992) and the Tafone mine, (Meli, Reference Meli1999). The only reported Sb-bearing associate of mopungite at La Selva Mine, Tuscany, Italy, is native antimony (Brizzi and Meli, Reference Brizzi and Meli1990). In this case, either sample volumes were too small to identify any other Sb phases, or prevailing chemical conditions favoured only the formation of mopungite. It is interesting to note that of the two reported occurrences of mopungite at Le Cetine mine, neither noted the presence of brizziite. However, Le Cetine Mine is the type locality for brizziite, and Olmi and Sabelli (Reference Olmi and Sabelli1994) reported it to be associated with mopungite, stibiconite, cetineite and senarmontite; this adding significant weight to the suggested kinetic stability of mopungite.

Syntheses for ottensite and a number of other synthetic analogues of the cetineite group (Sabelli et al., Reference Sabelli, Nakai and Katsura1988; Wang and Liebau, Reference Wang and Liebau1998; Wang and Liebau, Reference Wang and Liebau1999) have been reported yielding single crystals. Replication of these syntheses indeed produced ottensite, however, this phase typically formed only a minor portion of mixtures containing a host of other Sb minerals as well as a number of unidentifiable phases by PXRD. As such, the products were deemed unsuitable for solubility experiments.

Brizziite is considerably rare due not only to the highly restricted conditions which serve to stabilize it, but also because of the kinetic stability of its hydrated precursor mopungite. We must conclude that brizziite has no significant impact on Sb mobility under typical conditions which persist in the oxidized zones of Sb-bearing ores.