Minasgeraisite-(Y) discredited as an ordered intermediate between datolite and hingganite-(Y)

Abstract

Minasgeraisite-(Y) is discredited on the basis of it being an ordered intermediate between datolite and hingganite-(Y) (IMA-CNMNC Proposal 23-F). An idealised formula is (Ca₂Y₂)□₂(Be₂B₂)Si₄O₁₆(OH)₄, which corresponds to Ca₂□B₂Si₂O₈(OH)₂ (datolite) + Y₂□Be₂Si₂O₈(OH)₂ (hingganite-(Y)). The type material is rich in Bi, the Bi-richest portion yet discovered from the type locality is shown to be an intermediate member between datolite, hingganite-(Y) and a hypothetical end-member phase yet to be found of composition Bi₂□Be₂Si₂O₈(OH)₂. Minasgeraisite-(Y) has a different space group to datolite and hingganite-(Y). This lowering of symmetry to an acentric triclinic system is caused by different element occupancies on the A site of the gadolinite supergroup structure, which for minasgeraisite-(Y) becomes four individual sites. Such an order–disorder of elements is not considered as species-defining criteria despite the change in space group. Therefore, minasgeraisite-(Y) is discredited.

Introduction

Foord et al. (1986) described minasgeraisite [changed to minasgeraisite-(Y) to conform to the International Mineralogical Association (IMA) rules of nomenclature for rare-earth minerals (Bayliss and Levinson 1988)] as a new member of the gadolinite group, from the zoned complex of the Jaguaraçu granitic pegmatite, in the Mr. José Pinto quarry, Jaguaraçu, Minas Gerais, Brazil (IMA1983-90a). Using Inductive Coupled Plasma – Atomic Emission Spectroscopy (ICP-AES), Atomic Absorption Spectroscopy (AAS) and electron microprobe analyses, the empirical formula was given as:

$$\eqalign{& ( {{\rm Ca}_{0.45}{\rm Mn}^{2 + }_{0.20} {\rm Mg}_{0.08}{\rm Fe}_{0.05} {\rm Zn}_{0.02} {\rm Cu}_{0.01}\squ_{0.19}} ) _{\Sigma 1.00}\cr & ( {{\rm Be}_{1.55}{\rm B}_{0.21}{\rm Si}_{0.24}} ) _{\Sigma 2.00} ( {{\rm Y}_{0.72}{\it Ln}_{0.41} {\rm Ca}_{0.56}{\rm Bi}_{0.31}} ) _{\Sigma 2.00}\cr & ( {\rm Si_{1.95}{\rm P}_{0.08}} ) _{\Sigma 2.03}{\rm O}_{10}, \;} $$

where Ln expresses lanthanoids, La–Lu.

Their presented end-member formula was CaBe₂Y₂Si₂O₁₀ and their data suggested monoclinic symmetry. Since this time, a gadolinite-group mineral with this formula has been regarded as unlikely (Demartin et al., 2001, Bačík et al., 2014, 2017). The gadolinite supergroup (Bačík et al., 2017) includes mineral species with the general chemical formula A ₂MQ ₂T ₂O₈Φ₂, they can be silicates, phosphates or arsenates. They have monoclinic symmetry, space group P2₁/c. The currently recognised gadolinite-supergroup minerals are given in Table 1.

Table 1. Gadolinite-supergroup minerals.

The structure of gadolinite-supergroup minerals can be described as composed of two different layers parallel to (100) and alternating along the [100] direction (in the P2₁/c space group). One layer consists of TO₄ and QO₄ tetrahedra, and the other of AO₆Φ₂ polyhedra and MO₄Φ₂ octahedra.

At that time, the crystal structure of minasgeraisite-(Y) had not been determined, but to many, it seemed unlikely that the [6]-coordinated M-site would be dominantly occupied by Ca, which is assigned to the larger [8]-coordinated A-site in all other members of the supergroup.

A complicating factor in the ongoing discussion on the validity of minasgeraisite-(Y) as a mineral species, were several analyses performed in the original paper that indicated a high Bi content (Foord et al., 1986).

Crystal structure data

On the basis of an observation that minasgeraisite-(Y) might be a calcium rich hingganite-(Y) (Cooper and Hawthorne, 2018), Cooper et al. (2019) determined the crystal structure of a ‘Ca-hingganite-(Y)’ sample from the Heftetjern granitic pegmatite, located in southern Norway. They obtained a triclinic P1 unit cell with a = 9.863(4), b = 7.602(3), c = 4.762(2) Å, α = 90.002(15), β = 90.073(7), γ = 90.020(5)°, V = 357.1(5) ų and Z = 1, and expanded the general formula of gadolinite to 20 anions to show A-site cation ordering, with Ca and Y(Ln) ordered over four A sites (Fig. 1). The dominant constituent at the M site was determined as a vacancy (□), and no Ca was assigned to the M site. The structural formula presented was:

$$[ {( {{\rm Y}, Ln} ) {\rm Ca}{( {{\rm Y}, {\rm Ca}, Ln} ) }_2} ] ( {\squ , {\rm F}{\rm e}^{2 + }} ) _2( {{\rm Be}, {\rm B}, {\rm Si}} ) _4{\rm S}{\rm i}_4{\rm O}_{16}[ {( {{\rm OH}} ) , {\rm O}} ] _4.$$

Figure 1. Views of the two types of space groups. (a) Gadolinite-(Y) from the White Cloud pegmatite, Colorado, USA (Allaz et al., 2020). (b) Minasgeraisite-(Y) from Heftetjern granitic pegmatite, southern Norway (Cooper et al., 2019). Drawn using VESTA 3 (Momma and Izumi, 2011).

The refined formula was:

$$\eqalign{& ( {{\rm C}{\rm a}_{1.992}{\rm Y}_{1.873}Er_{0.135} } ) _{\Sigma 4.000}( {\squ_{1.216}{\rm F}{\rm e}^{2 + }_{0.784} } ) _{\Sigma 2.000} \vskip-6pt \cr & ( {{\rm B}{\rm e}_{2.24}{\rm B}_{1.58}{\rm S}{\rm i}_{0.18}} ) _{\Sigma 4.00} {\rm S}{\rm i}_4{\rm O}_{16}[ {{( {{\rm OH}} ) }_{2.382}} {\rm O}_{1.618} ] _{\Sigma 4.000}, \;} $$

where Er is not erbium, but an average based on all the lanthanides (57–71)

The simplified ideal formula reduces best to:

$$( {{\rm C}{\rm a}_2{\rm Y}_2} ) \squ ( {{\rm B}{\rm e}_2{\rm B}_2} ) {\rm S}{\rm i}_4{\rm O}_{16}[ {{( {{\rm OH}} ) }_4} ] $$

Cooper et al. (2019) did not address the discussion of defining a species name, but the ‘Ca-hingganite-(Y)’ studied corresponded to what would have been a minasgeraisite-(Y) at the time, and therefore could be regarded as the second occurrence of the mineral.

The chemical formula of Brazilian minasgeraisite-(Y) as reported by Foord et al. (1986) based on the structural data obtained by Cooper et al. (2019) becomes:

$$\matrix{ {{( {{\rm C}{\rm a}_{1.79}{\rm B}{\rm i}_{0.37}} ) }_{\Sigma 2.16}( {\rm Y}_{1.27}{\rm Y}{\rm b}_{0.35}{\rm E}{\rm r}_{0.09}{\rm L}{\rm u}_{0.08}{\rm D}{\rm y}_{0.06}{\rm T}{\rm m}_{0.03}}\hfill \cr {{\rm N}{\rm d}_{0.03}{\rm H}{\rm o}_{0.02}{\rm G}{\rm d}_{0.02}} {\rm L}{\rm a}_{0.01}{\rm C}{\rm e}_{0.01}{\rm P}{\rm r}_{0.01}{\rm S}{\rm m}_{0.01}{\rm T}{\rm b}_{0.01}) _{} {_{\Sigma 2.00}} \hfill \cr{( {\squ_{1.38}{\rm M}{\rm n}^{2 + }_{\;\;\;\;0.35} {\rm M}{\rm g}_{0.13}{\rm F}{\rm e}^{2 + }_{\;\;\;\;0.08} {\rm Z}{\rm n}_{0.04}{\rm C}{\rm u}_{0.02}} ) }_{\Sigma 2.00} \hfill \cr {{( {{\rm B}{\rm e}_{2.74}\squ_{0.89}{\rm B}_{0.37}} ) }_{\Sigma 4.00}} {{( {{\rm S}{\rm i}_{3.85}{\rm P}_{0.11}} ) }_{\Sigma 4.00}}\hfill \cr {[ {{\rm O}_{15.20}{( {{\rm OH}} ) }_{0.80}} ] }_{\Sigma 16.00}{( {{\rm OH}} ) }_{4.00}.\hfill}$$

The significant visual difference between the Brazilian and Norwegian minasgeraisite-(Y) is the colour, the lilac colouration of the Brazilian samples probably being due to Mn²⁺ and the brown–orange of the Norwegian due to the presence of Fe²⁺ at the M site.

Before the work on ‘Ca-hingganite-(Y)’, Cooper and Hawthorne (2018) had already determined the crystal structure of a ‘minasgeraisite-(Y)’ sample from the type locality. They used the name in single quotes because it was not the type specimen and, unfortunately, they were unable to chemically analyse it, choosing to apply the chemical data from Foord et al. (1986) instead.

They obtained a triclinic P1 unit cell (their figure 5) with a = 9.994(4), b = 7.705(3), c = 4.764(2) Å, α = 90.042(9), β = 90.218(14), γ = 90.034(9)°, V = 366.8(5) ų and Z = 1, and again expanded the general formula of gadolinite to 20 anions. They showed that Bi, Ca and REE were ordered over four different A sites, the dominant constituent at the M sites was a vacancy and Ca did not occur at the M sites. The structural formula given was:

$${\rm BiCa}( {{\rm Y}, Ln} ) _2( {\squ , {\rm M}{\rm n}^{2 + }} ) _2( {{\rm Be, B, Si}} ) _4{\rm S}{\rm i}_4{\rm O}_{16}[ {( {{\rm OH}} ) , {\rm O}} ] _4.$$

Appling this to the bulk sample chemical data presented in Foord et al. (1986), they acquired the empirical formula:

$$\matrix{ {{( {{\rm Y}_{1.42}{\rm C}{\rm a}_{1.19}{\rm B}{\rm i}_{0.81}Er_{0.58}} ) }_{\Sigma{4.00}} {( \squ _{1.37}{\rm M}{\rm n}^{2 + }_{0.63} ) }_{\Sigma {2.00}} {( {{\rm B}{\rm e}_{3.32}{\rm B}_{0.40}{\rm S}{\rm i}_{0.28}} ) }_{\Sigma{4.00}} } \hfill \cr {{\rm S}{\rm i}_4{\rm O}_{16}{[ {{( {{\rm OH}} ) }_{2.74}{\rm O}_{1.26}} ] }_{\Sigma {4.00}} .} \hfill \cr } $$

Applying the same methodology to the most Bi-rich analysis presented in Foord et al. (1986) gives:

$$\matrix{ {{( {\rm C}{\rm a}_{1.36}{\rm B}{\rm i}_{1.15}{\rm Y}_{0.94}{\rm Y}{\rm b}_{0.28}{\rm D}{\rm y}_{0.07}{\rm L}{\rm u}_{0.07}{\rm E}{\rm r}_{0.06}{\rm T}{\rm m}_{0.03}{\rm L}{\rm a}_{0.01}{\rm P}{\rm r}_{0.01}}{\rm N}{\rm d}_{0.01}}\hfill \cr {{\rm S}{\rm m}_{0.01}{\rm G}{\rm d}_{0.01}{\rm T}{\rm b}_{0.01}{\rm H}{\rm o}_{0.01}) }_{\Sigma 4.03} {( \squ _{1.26}{\rm M}{\rm n}^{2 + }_{0.47} {\rm M}{\rm g}_{0.14}{\rm F}{\rm e}^{2 + }_{0.09}} \hfill \cr {{{\rm Z}{\rm n}_{0.03}{\rm C}{\rm u}_{0.01}) }_{\Sigma 2}{( {\rm B}{\rm e}_{2.94}\squ _{0.67}{\rm B}_{0.41}) }_{\Sigma 4}{( {\rm S}{\rm i}_{3.84}{\rm P}_{0.16}) }_{\Sigma 4}} \hfill \cr {{[ {{\rm O}_{15.50}{( {{\rm OH}} ) }_{0.50}} ] }_{\Sigma 16}{( {{\rm OH}} ) }_4.} \hfill \cr } $$

A recent further study, the first which has studied both chemically and structurally a ‘minasgeraisite-(Y)’ sample from the type occurrence was published during review by the Commission on New Minerals, Nomenclature and Classification (CNMNC) of the IMA of the proposal relevant to deciding if minasgeraisite-(Y) required mineral species status. That study (Vereshchagin et al., 2023) confirmed much of the data of Cooper and Hawthorne (2018), but was unable to determine the lower symmetry due to lack of sample material. Their conclusion was that their sample, which was linked to the original description, but not confirmed as the type specimen was equivalent to a monoclinic Bi- and Mn-bearing hingganite-(Y). There was no additional discussion on the structural position of Ca other than noting excess Ca on the M site.

Conclusions

On the basis of the crystal structure and structural formula determined by Cooper et al. (2019) for ‘Ca-hingganite-(Y)’ and the chemical analysis determined by Foord et al. (1986) for the type specimen, the simplified ideal formula of minasgeraisite-(Y) is (Ca₂Y₂)□₂(Be₂B₂)Si₄O₁₆(OH)₄, which corresponds to Ca₂□B₂Si₂O₈(OH)₂ (datolite) + Y₂□Be₂Si₂O₈(OH)₂ (hingganite-(Y)).

On the basis of the crystal structure determined by Cooper and Hawthorne (2018) for a ‘minasgeraisite-(Y)’ sample from the type occurrence and the chemical analysis determined by Foord et al. (1986) on their ‘Bi-richest portion’, the simplified ideal formula for the richest Bi composition thus far reported of minasgeraisite-(Y) is (Y₂CaBi)□₂(Be₃B)Si₄O₁₆(OH)₄, which corresponds to Y₂□Be₂Si₂O₈(OH)₂ (hingganite-(Y)) + 0.5 Ca₂□B₂Si₂O₈(OH)₂ (datolite) + 0.5 Bi₂□Be₂Si₂O₈(OH)₂ (unknown phase).

Both minasgeraisite-(Y) and the Bi-richest minasgeraisite-(Y) are gadolinite-supergroup minerals with non-end-member compositions. The solid solution of more than two elements in the A sites causes a lowering of their symmetry to acentric triclinic resulted from cation ordering of Ca, REE (and Bi) at the A site, which become non-equivalent sites A1–A4 (Cooper et al., 2019). However, topologically similar polymorphs where such order–disorder relationships affect the symmetry of a mineral (different space group), without modifying the global topology do not define a new species (Nickel and Grice, 1998).

It is likely that, if the crystal structures of many gadolinite-supergroup minerals with non-end-member compositions were analysed accurately, P1 symmetry would be identified in the strictest sense, yet they would have P2₁/c average structures. As a consequence, minasgeraisite-(Y) should not be separately created based on its space group nor the presence of Bi. Should a Bi-dominant analysis overall on the A site(s), with ideal formula Bi₂□Be₂Si₂O₈(OH)₂, be discovered it would classify as a new species.

In summary, minasgeraisite-(Y) is discredited as an ordered intermediate between datolite and hingganite-(Y) (IMA-CNMNC Proposal 23-F, Atencio, 2023). The Bi-richest portion is an intermediate member between datolite, hingganite-(Y), and a hypothetical phase yet to be found.