Nomenclature of the Alunite Supergroup: Discussion

1295

The Canadian Mineralogist Vol. 38, pp. 1295-1297 (2000)

NOMENCLATURE OF THE ALUNITE SUPERGROUP: DISCUSSION

KEITH M. SCOTT¤ CRC LEME (Cooperative Research Centre for Landscape Evolution and Mineral Exploration), c/o CSIRO Exploration and Mining, P.O. Box 136, North Ryde, N.S.W. 1670, Australia

The aluniteÐjarosite minerals are defined as having review is timely. His systematic examination of miner- the general formula AB3(XO4)2(OH)6, where A is a large als relative to each different A-site-dominant cation is ion in 12-fold coordination (e.g., K, Na, Ca, Pb, REE), very clear and succinct. This careful work has led to a B is usually Fe or Al, and the XO4 anions are usually call for critical appraisal of currently used names, like SO4, PO4 or AsO4. Because of the potential for a large kemmlitzite, which plot outside their ideal fields, and a number of complex substitutions in the supergroup and cessation of the use of discredited names like “weilerite”. to avoid a proliferation of mineral names, Scott (1987) I concur with such a call. However, he also advocates a proposed that compositional limits be set on some pre- new system of classification based on a ternary system viously ill-defined aluniteÐjarosite minerals (viz., for the SO4, AsO4 and PO4 anions. Such a system is in alunite, jarosite, hinsdalite, corkite, goyazite, plumbo- accord with current general CNMMN recommendations gummite, florencite and the now discredited “lusun- for solid solutions (Nickel 1992), and could reduce the gite”). Those boundaries were set at 0.5 and 1.5 formula need for named members of a particular A-site-domi- units of (SO4) in the XO4 sites. That proposal was nant group of minerals from 10 to 6. For the whole accepted by the Commission on New Minerals and Min- aluniteÐjarosite family, this could potentially reduce eral Names (CNMMN) of the International Mineralogi- mineral names by 40%. This initially attractive propo- cal Association and used as the basis for a classification sition needs further evaluation. Does it significantly re- system for the aluniteÐjarosite minerals. That system is duce mineral names, and does it really help workers in based on 1) Fe or Al dominance in B sites for the initial the field? For minerals with cations other than Pb in A subdivision (because that is reflected crystallographi- sites, apart from the elimination of kemmlitzite (for cally and is easily measured by X-ray diffractometry), which the type material plots in the field of arseno- and then 2) the dominant cation in the A site, 3) the pro- goyazite in any case), only huangite would be eliminated portion of SO4, AsO4 and PO4 in XO4 sites with bound- if his proposal were accepted. For Pb-rich minerals, aries at 0.5 and 1.5 formula units of SO4, and 4) the use where 11 named species are currently recognized, five of adjectival modifiers to indicate significant amounts minerals could be potentially eliminated (Fig. 1). Thus of other ions in A or B sites. The system was intended to this Pb-rich grouping will be considered in more detail be a practical one which, at least in the initial subdivi- below. sion into Al- and Fe-rich groups (using differences in In the PbÐFe-rich members of the supergroup, the cell dimensions, which can be determined by X-ray classification proposed by Jambor (1999) leads to diffractometry), did not need full chemical analysis. corkite having a possible composition PbFe3H(PO4)2 However, it does not seem to have gotten widespread (OH)6 rather than PbFe3(PO4)(SO4)(OH)6, as currently acceptance, with most subsequent authors preferring to defined. His proposed classification implies that there retain the historical division into alunite, beudantite and is nothing special about having one divalent and one crandallite groups (e.g., Mandarino 1999, Gaines et al. trivalent anion in the XO4 sites. However, the work of 1997), i.e., grouping Fe- and Al-rich analogues together Blount (1974) and Radoslovich (1982) indicates that one and requiring compositional data for even preliminary of the PO4 anions in the “double phosphate” members 2Ð classification. Thus it is gratifying that the classifica- is actually protonated and divalent, i.e., (PO3OH) . tion outlined by Jambor (1999) relies upon the B-site Thus the major change in moving from a PbÐFeÐSO4 to occupancy, at least for its secondary divisions. PbÐFeÐPO4 mineral occurs where one of the sulfate Since my proposal, there have been many more groups is replaced. Ideally, the change can be envisaged members of the aluniteÐjarosite family reported; there as Pb0.5Fe3(SO4)2(OH)6 → PbFe3(PO4)(SO4)(OH)6 → are now more than 40 members, and so Jambor’s (1999) PbFe3H(PO4)2(OH)6, with the first step involving the

¤ E-mail address: [email protected]

1295 38#5-oct.00-2188-16 1295 23/11/00, 13:18 1296 THE CANADIAN MINERALOGIST

FIG. 1. Minerals of the aluniteÐjarosite family with dominant Pb in the A sites (modified from Jambor 1999). * substantial Cu present in B sites in these minerals.

FIG. 2. Compositions of Pb-rich aluniteÐjarosite minerals, showing the restricted extent of PO4ÐAsO4 solid solution once SO4 exceeds 25% [modified from Jambor (1999), and with more data added]. Data points: black dot: Sejkora et al. (1998), black square: Rattray et al. (1996), open triangle: Scott (1987), open circle: Scott (unpubl. data, Cobar region).

1295 38#5-oct.00-2188-16 1296 23/11/00, 13:18 NOMENCLATURE OF THE ALUNITE SUPERGROUP 1297

introduction of the trivalent anion being balanced by the current names rather than describing them as plumbo- addition of 0.5 Pb in the A site, and the second step jarosite and plumbogummite, respectively (Jambor’s merely involving an interchange of divalent anions, SO4 proposal) is more informative about their compositions. and PO3OH. Alternatively, if beaverite, Pb(Fe,Cu)3 Of the various members of the aluniteÐjarosite fam- (SO4)2(OH)6, is taken as the starting material, the intro- ily, the Pb-rich are the best documented, with the full duction of the trivalent anion is balanced by the replace- range of compositions being recognized, and thus ment of the divalent Cu in the B site by trivalent Fe to Jambor’s (1999) proposal potentially has the most im- form corkite, PbFe3(PO4)(SO4)(OH)6.The current clas- pact upon their nomenclature. Therefore, in the above sification recognizes this significant change in the sub- argument, I have concentrated on them. His proposal stitution process, where approximately equal amounts hides the importance of the 1:1 divalentÐtrivalent anion of divalent and trivalent anions occur as corkite, but this mechanism of substitution in these minerals to form change is hidden by the system proposed by Jambor corkite and beudantite. In addition, the general paucity (1999). of published data for minerals containing significant Furthermore, consideration of naturally occurring amounts of both PO4 and AsO4, unless these anions to- material (Fig. 2) suggests that solid solution involving gether occupy 75% of the XO4 sites, favors the reten- the PO4 and AsO4 anions is not generally significant tion of a division at 25% SO4 (0.5 formula units of SO4), unless SO4 occupies less than 25% of the XO4 sites. at least until the importance of solid solution between Hence a boundary at 0.5 formula units SO4 may repre- minerals like corkite and beudantite is established. Thus sent a real compositional break that is recognized by I contend that the current system of nomenclature is Scott’s (1987) system but not by Jambor’s (1999) pro- actually more informative about the substitutions posed system (Fig. 2). Above this boundary, recogni- present within particular members of the aluniteÐjarosite tion of corkite and beudantite as PbÐFeÐPO4ÐSO4 and family, and should be retained despite the slightly PbÐFeÐAsO4ÐSO4 minerals, with 1:1 divalent and triva- greater number of mineral names it allows. lent anions and generally little incorporation of a second trivalent anion, thus appears useful. Hence, further ACKNOWLEDGEMENTS work to demonstrate extensive mixing of PO4 and AsO4 in minerals like corkite and beudantite is needed before Discussions with E.H. Nickel (CSIRO, Perth) and changes to the currently approved nomenclature should W.D. Birch (Museum of Victoria, Melbourne) are grate- be considered. Furthermore, even if additional work fully acknowledged. leads to the discovery of more compositions like that of the published corkite with significant AsO4 and Received June 18, 2000. hinsdalite with significant AsO4 (Fig. 2), retaining the

1295 38#5-oct.00-2188-16 1297 23/11/00, 13:18