Metavivianite and Kerchenite" a Review
MINERALOGICAL MAGAZINE, DECEMBER 1986, VOL. 50, PP. 687-91
Metavivianite and kerchenite" a review
K. A. RODGERS Department of Geology, University of Auckland, Private Bag, Auckland, New Zealand
ABSTRACT. Metavivianite has been shown to be not Sluys, 1971). Poullen interpreted his results as dimorphous with vivianite. Assuming the homogeneity showing that oxidation of monoclinic vivianite and uniformity of the original type samples, it is a triclinic resulted in a reduction of symmetry through pro- hydrated ferri-ferrous hydroxy phosphate whose formula duction of a ferri-ferrous hydroxy phosphate may be given as Fe~+xFe3+(po4)2(OH)x(8--x)H20 hydrate of the type where x > 1.4. The precise oxidation limits between which the triclinic lattice is stable are not known but the Fe23+ xVe3 + (POg)z(OH)x(8 - x)HzO structure persists close to total oxidation of all iron. The structure of metavivianite was established using a frag- assuming that compensation of the kind proposed ment of kerchenite whose formula as originally given is by Moore (1971) applied (and cf. Gamidov and covered by the general metavivianite formula. Assuming Mamedov, 1960). The X-ray diffraction signature of the homogeneity of the original (1907) samples,kerchenite this alteration product was that given by Ritz et al. and metavivianite appear to be identical. (1974) for metavivianite. Poullen (1979) noted, in KEYWORDS: metavivianite, vivianite, kerchenite. passing, that a similar alteration may occur in parasymplesite with partial oxidation producing a METAVIVIANITE is a triclinic iron phosphate mixed diffraction pattern containing prominent hydrate. It was first described by Ritz et al. (1974) reflections of triclinic symplesite. from the Big Chief pegmatite, South Dakota, In a series of papers, Chevalier et al. (1980), occurring as green crystals intimately intergrown Dormann and Poullen (1980), Fejdi et al. (1980), with kryzhanovskite in cavities in triphylite. These and Dormann et al. (1982) presented extensive authors adjudged the mineral to be dimorphous structural and chemical data in support of Poullen's with vivianite (Fe3(PO4)2 8H20) and isostructural conclusions. Many of these data were derived from with symplesite (Fe3(AsO4)2'8H20). These con- vivianite-metavivianite admixtures produced by clusions were reached on the basis of similarities in both artificial and natural oxidation of vivianite the X-ray diffraction patterns of metavivianite and and synthetic F%(POa)2 "8H20. While this in- symplesite and from direct electron microprobe formation is useful in defining the nature and limits comparison of the composition of metavivianite of the structural and compositional relationship with an unlocalized vivianite: 'metavivianite and that exists between vivianite and the redefined the vivianite are nearly identical in composition for metavivianite, it is data concerned with two meta- all elements with Z < 10 [sic] and therefore by vivianite samples identified as being mono- analogy with vivianite the formula of metavivianite mineralic that help characterize the ferri-ferrous is (Fe,Mn)3(PO4)2 8H20' (pp. 897-8). Regrettably, nature of the mineral. no separate analyses were reported for either The first of these was a specimen identified by Fe z +/Fe 3+ or water. Recent work has shown that Dormann and Poullen (1980, p. 634) as metavivian- the above conclusions are not correct and that ite from the Yukon, Canada, and containing MnO metavivianite is neither a ferrous phosphate nor an 2.83 ~, MgO 2.98 ~o, and CaO 0.72 ~. The M6ss- octahydrate and so cannot be dimorphous with bauer spectrum of this sample was interpreted as vivianite (e.g. Dormann et al., 1982; Rodgers and indicating a percentage occupancy of iron sites in Johnston, 1985). the lattice of Felz+ = 1~o, Fe3+ = 32~o, Fei] += Crystal chemistry. For some years X-ray diffrac- 2~, Fed+= 65~, (Fe3+ = 49~, Fe3ff = 16~o), tion data for vivianites have been reported which where sites I and II are similar to and derived from include prominent reflections of metavivianite (e.g. those specified by Mori and Ito (1950) for their Poullen, 1979; Henderson et al., 1984; Sameshima model of the vivianite structure, and where Fel3+ et al., 1985) although such reflections may not have and Fei3+ are identified as distinct types of Fell + been recognized as coming from metavivianite (e.g. sites within metavivianite(cf. Dormann et al., 1982); Minato et al., 1956; Zwann and Kortenburg van der the hyperfine spectral parameters of Fei3i+ are Copyright the Mineralogical Society 688 K. A. RODGERS close to those ofFer While no formula was given by TABLE I, Chemical and structural data For metavivianite and Dormann and Poullen (1980) their data were kerchenite calculated by Rodgers and Johnston (1985) as Fe2+ Fe~.~(PO4)z(OH)2.95.1 H20. 1 2 3 4 5 The second sample came from Kamysch-Burun Fe2+/Fe3+ 0.613 0.034 0.176 0.321 0.319 (USSR) and was used by Dormann et al. (1982) to FeO 23.85 9.50 9.49 Fe 03 13.55 32.89 32.965 determine the precise nature of the metavivianite Ni~ D.13 MnO 4.Z 1.83 1.99 1.84 structure. 'Le cristal utilis6 pour l'6tude structurale MgO 2.98 1,54 1.56 est un fragment de clivage pr61ev6 sur un 6chantil- CaO 0,5 0,72 0.49 0.46 P 0 28.4 28.19 28.21 Ion... est conserve sous le nom de "kerchenite ~" H~O 5 25,04 24.91 mais se compose en r~alit6 de deux phases Total 99.64 99.435 min6rales, l'une monoclinique et l'autre triclinique' a~ 7,81 7.84 9.08 9.11 (p. 148). In their earlier study, Dormann and 4.65 4.67 s ~ 94.77 95.04 Poullen (1980) had been able to distinguish and 6 ~ 97.15 96.94 separate vivianite and metavivianite in similar yO 107.37 107.72 samples using a binocular microscope: 'Cette i. Type metavivlanite, Big Chief, South Dakota. Ritz et al (1974), operation fut relativement ais6e, les limites des Rodgers and Johnston (1985). 2. Metavivianite, Yukon. Darmann and Poullen (1980). domaines monocliniques et tricliniques &ant 3. Metavivianite, Bamysch-Burun. Dormannetel (1982). toujours tr6s nettement marqu6es' (p. 634). 4,5. Kerchenite, Ketch. Popoff (1907). Presumably, the same technique was used to select the fragment of metavivianite in the 1982 study. The M6ssbauer spectrum of this fragment showed the structure of vivianite as being maintained in the 85 % of the total iron was Fe 3+ which must yield a oxidation sequence until 40 % of the total iron had formula (as above) Fe0.asF%.~5(PO4)/(OH)2.ss2 + 3 + been oxidized, corresponding to a formula 5.4H10 (but cf. Dormann et al. (1982, p. 154) Fe~.~Fe3+(po4)z(OH)I.26.8HBO. although the formula given by them does not use a simple Fe z + ~ Fe 3 + + OH- compensation as its They stress that in the triclinic lattice the Fe~ site is basis of calculation and, further, there should be completely oxidized while 'Le taux d'oxydation du 3 Fe atoms with a valency of + 2.66 and not 1 as site II varie de 20 % environ, valeur inf6rieure fi celle they indicate). des vivianites monocliniques les plus oxyd6es Subsequently, Rodgers and Johnston (1985) re- jusqu'& pratiquement 100% (m&avivianite de investigated type metavivianite from Big Chief by Yukon). L'esp~ce triclinique admet donc des M6ssbauer spectroscopy. They obtained a formula teneurs en Fe 3 +allant de 50 % ~ 100 % du fer total, of FeH4FeLs6(PO4)z(OH)l.866.1HBO2+ 3+ in which (p. 637). Further, M6ssbauer spectral studies were 62% of the iron was Fe 3+, a considerably lower unable to distinguish triclinic phase, Fe3§ proportion than in either the Kamysch-Burun or specimens from disordered X-ray amorphous Yukon specimens. Composition and structural material. parameters of the above samples are summarized in Dormann et al. (1982) appeared to have qualified Table I. the above on the basis of combined M6ssbauer, Compositional ran#e. Both Dormann and Poullen structural and thermal studies. They summarized (1980) and Dormann et al. (1982) were primarily their findings for natural crystals as: concerned with examining the oxidation sequence (i) limits of oxidation of monoclinic lattice of vivianite. Structural changes occurring in the 0~
Rodgers and Johnston (1985) found that 8 ~o of Palache et al. (1951) placed the name under a iron in Fe I was unoxidized while, despite their 'Variety' heading within their section on vivianite assertion to the contrary, Dormann and Poullen and equated it with 'Oxidized Vivianite', noting (1980) reported that 3 ~o of iron in Fel was ferrous. that while cleavable pseudomorphs may be ob- Clearly then, the above limits represent ideals and tained these 'give very diffuse X-ray patterns' (p. further study of natural specimens which are close 744). Hey (1962) indicated there was some doubt as to these limits is necessary to establish the composi- to the mineral's validity, regarding the varieties as tional field within which the triclinic lattice is stable. 'ill-defined' (p. 245), while Fleischer (1975) in his However, on the basis of the above discussion, the Glossary listed kerchenite as' = oxidation products formula of metavivianite can be written, to a first of Vivianite' (p. 82). Moore (1976) discounted the approximation, as term as not belonging to a true, crystalline mineral entity and decried the introduction of additional, Fe 2 +xFe3 +(PO4)2(OH)~(8- x)H20 where x > 1.4. trivial names for crystalline compounds of inter- Kerchenite. The Kamysch-Burun locality from mediate composition between end members. which the sample of Dormann et al. (1982) came, is Moore's comment highlights the fact that many one of several in the Kerch-Taman area of Crimea of the difficulties which kerchenite appears to have from which Popoff (1906, 1907, 1910a, b, and given mineralogists stem not from Popoff's (1907) elsewhere) described the mineral kerchenite which findings as summarized above, but from his asso- he found as forming by oxidation of vivianite. In so ciated interpretation of the relationship between far as the crystalline phosphates of Kerch were vivianite and kerchenite. For him the two were part concerned, the 1906 publication was essentially a of a solid solution series: Fe3P2Os-8H20... preliminary note; that of 1907 was far more com- nFe3P20 8 - mFePO4- PH20... FePO4" pHzO. He prehensive. In it he drew a distinction between the was unaware of the requirements of modern crystal crystalline vivianites and his new kerchenite. The chemistry such as Goldschmidt's rules and that former occurred as colourless to pale blue crystals continuous electrostatic compensation could be with a pale blue streak, a hardness of 2, and showed achieved in solid solution series by simultaneous cleavage on {010} only. Kerchenite occurred as substitutions such as Fe 2 + + H20 ~ Fe 3 + + OH-. green to black crystal pseudomorphs after vivianite, He became preoccupied with stoichiometry and it had a hardness of 3 and, in addition to cleavage was this that appears to have led him to become parallel to {010} of vivianite, also showed a less entangled in a tortuous argument whose main perfect cleavage parallel to {100} of that mineral thrust was to show that in such a series as the above, and a third parting perpendicular to the other two. a distinct species had formed whenever stoichio- Chemically, the mineral varied in its ferrous-ferric metry was satisfied. Since he had not identified ratio. The two analyses given by Popoff (Table I, OH- radicles amongst the components of the cols. 4, 5) had a high level of ferric iron and he gave a series, stoichiometry could not be achieved formula close to FeO" Fe20 3.P2OS.7H20. On continuously. the basis of the substance's comparative stability, its This unnecessarily complicated approach was stoichiometry, the differences in physical properties compounded in 1910. However, the title and con- and chemistry from those of vivianite, he argued cluding paragraph of the 1907 study make it quite that kerchenite was a distinct mineral species. The clear that as far as Popoff was concerned, in this subsequent fate of the species in the mineralogical 1907 study he was dealing with phosphates at literature has been one of fluctuating fortunes. Kerch that occurred in crystallized habits only, be Initial support for the mineral's validity ap- these the original crystals or their pseudomorphs. peared to come from further studies, e.g. P/opoff He noted that the more usual habit for the phos- (1910a, b) and Dvoichenko (1914) (cf. Hey, 1962). phates was as friable, earthy, heterogeneous, brown Mellor (1935) even provided representations of the materials. These were what he described in the 1910 structures of variously oxidized kerchenite varieties. publications and it was here that he specified the However, as workers found themselves unable to stoichiometric proportions for three subspecies of recognize such kercbenite varieties among the kerchenite with differing Fe2+/Fe 3+ ratios. The oxidation products of vivianite from areas outside heterogeneous brown materials he regarded as the Kerch and Taman peninsulas, using the criteria mixtures of these subspecies in varying propor- of Popoff, kerchenite came to be regarded as being a tions. Subsequently, additional subspecies were mixture of vivianite and rather indeterminate oxi- recognized on the same basis, e.g. Dvoichenko dation products. This attitude was strengthened (1914) (cf. Palache et al., 1951; Hey, 1962). when some researchers reported failure to obtain A second difficulty in identifying kerchenite any new distinct X-ray signature from oxidized outside the type area appears to have arisen from vivianite, e.g. Chuov and Rudnitzkaja (1966). no worker encountering green to black crystals 690 K. A. RODGERS derived from oxidation of vivianite. Most re- assessed and the identity and homogeneity of all the searchers appear to have been concerned with material be established along with the applicability trying to interpret friable, earthy, heterogeneous, of all the published data on metavivianite as defined brown alteration products and, indeed, many above. Unfortunately, no further material was museum specimens labelled as kerchenite or one of available for further study from either of the the subspecies, appear to be no more than mixtures repositories of type metavivianite as given by Ritz such as alluded to by Popoff (1907) and described et al. (1974, p. 896). (1910a, b). It was from such a museum specimen If more than one phase does exist at Big Chief that Dormann et al. (1982) selected their cleavage then the question may arise as to which the name fragment of metavivianite for structural determina- metavivianite properly belongs. The name as re- tion. The condition of many museum specimens, defined by Poullen (1979) is now a misnomer in so partly as a consequence of their instability in the far as Ritz et al. (1974) chose it to 'emphasise the conditions under which they are stored, has pro- dimorphous relationship to vivianite' (p. 896). vided a lack of quality material for critical, com- While Popoff's attempt to establish a continuous parative study and has helped endorse the notion solid solution oxidation series between vivianite that all kerchenites, ipso facto, are mixtures. The and ferric phosphate hydrate was also in error, he descriptions Popoffgave of the samples he examined did describe kerchenite as forming by oxidation of in his 1907 paper gives little reason to assume that vivianite. The samples he used in his original they were other than monomineralic, whatever may description appear to have been homogeneous have been the nature of those described in later from the physical description he gave, which in- papers. cluded optical examination. The mineral had dis- Discussion. Metavivianite was incorrectly tinctive physical properties and its Fe 2 +/Fe 3 + ratio described by Ritz et al. (1974). Their Big Chief could vary over a wide range. While there appears specimen consisted of green crystals which ranged to be no justification for setting up separate names from transparent to opaque with the smaller grains for intermediate compositions between end mem- being transparent. The fragment from this same bers, the problem is to ascertain what are the locality examined by Rodgers and Johnston (1985) correct end members of the oxidation solid solution was khaki and near opaque. It was the same as series. It can be noted that all the various formulae X-rayed by Sameshima et al. (1985, Table I, col. 19) of Popoff's stoichiometric intermediate com- and yielded the same diffraction pattern as given by pounds are included within the general formula of Ritz et al. (1974) without showing any other metavivianite as given above, once H20 ~ OH- reflections. The Yukon and Kamysch-Burun speci- compensation is allowed for. For example, his mens were described by Dormann and Poullen so-called ~-kerchenite can be recalculated as (1980) and Dormann et al. (1982) as brown and Fel.66Fel.33(PO4)2(OH)l.336.66HzO.2+ 3+ Further, it opaque. Their identity and homogeneity were must be emphasized that the oxidized product of confirmed by X-ray diffraction. vivianite does give a positive X-ray signature (e.g. If it is assumed that all the data given by Ritz et Poullen, 1979; Dormann and Poullen, 1980; Dot- al. (1974) apply equally to the small transparent mann et al., 1982; Henderson et al., 1984; Same- crystals and to the opaque material, then it is clear shima et al., 1985) despite statements to the that Poullen's contention that metavivianite is a contrary (e.g. Palache et al., 1951; Chuov and triclinic, hydrated, ferri-ferrous hydroxy phosphate Rudnitzkaja, 1966; Tien and Waugh, 1969; Vochten is upheld. The mineral has a variable Fe2+/Fe 3+ et al., 1979). The signature is that given by Ritz ratio and Mn can substitute for iron. Presumably, et al. (1974) for metavivianite. the variation in colour and transparency reflect this When Dormann et al. (1982) provided structural variability as do the slight variations in physical substantiation of Poullen's (1979) redefinition of parameters (Table I). metavivianite they did so using a cleavage fragment However, this assumption may be incorrect and of kerchenite. It would appear that kerchenite and two minerals may have been confused in the re-defined metavivianite are identical. original description. It may be that the green Consideration could be given to reinstating transparent crystals are one phase which oxidizes kerchenite as the name for triclinic ferri-ferrous to yield a ferri-ferrous opaque brown product hydroxy phosphate hydrate pending a re-examina- which is identical with that obtained by oxidation tion of both the Big Chief and Kerch localities to of vivianite. In the absence of ferrous/ferric analyses establish unambiguously the identity of the two in the original description this possibility should minerals and, in the case of Kerch, to confirm that not be overlooked. It would therefore seem prudent the prefixed varieties of kerchenite are intermediate that the original specimens of metavivianite and, if compositions within a solid solution series and not necessary, the original locality be carefully re- discrete mineral species. Such a re-examination METAVIVIANITE AND KERCHENITE 691 could help in defining the limits within which the McCammon, C. A., and Burns, R. G. (1980) Am. Mineral. triclinic lattice is stable as well as the nature of the 65, 361-6. reportedly X-ray amorphous brown product that Mellor, J. (1935) A comprehensive treatise on inorganic results from complete oxidation. and theoretical chemistry, 14. London. Minato, H., Kinoshita, K., and Okamoto, Y. (1956) Mineral. J. 1, 337 47. Acknowledgements. Thanks are owed to P. G. Embrey and Moore, P. B. (1971) Am. Mineral. 56, 1-17. the late Dr M. H. Hey of the Department of Mineralogy, --(1976) Mineral. Record, 7, 141. British Museum (Natural History) for their suggestions, Mori, H., and Ito, T. (1950) Acta Crystallogr. 3, 1-6. discussion, and advice--not all of which has necessarily Palache, C., Berman, H., and Frondel, C. (1951) Dana's been taken. System of Mineralogy, 2. John Wiley, New York. Popoff, P. (1906) Centralblatt Mineral. Geol. Paliiontol. REFERENCES 4, 112. --(1907) Bull. Acad. Sci. St. Pdtersbourg, ser. 6, 1, Chevalier, R., Gasp6rin, M., and Poullen, J.-F. (1980) C.R. 127 40. Acad. Sci. Paris, 291D, 661 3. --(1910a) Tray. Mus. Gdol. Acad. Sci. St. P~tersbour9, Chuov, F. V., and Rudnitzkaja, E. S. (1966) lsv. Akad. 4, 99. Nauk, 8, 51-6. --(1910b) Akad. Nauk. 4, 162-89. Dormann, J.-L., and Poullen, J.-F. (1980) Bull. Mineral. Poullen, J.-F. (1979) C.R. Acad. Sci. Paris, 289D, 103, 633-9. 51-2. --Gasp6rin, M., and Poullen, J.-F. (1982) Ibid. 105, Ritz, C., Essene, E. J., and Peacor, D. R. (1974) Am. 147-60. Mineral. 59, 896-9. Dvoichenko, R. (1914) Mere. Crimean Soc. Sci. Nat. 4, Rodgers, K. A., and Johnston, J. H. (1985) Neues Jahrb. 113-14. Mineral. Mh. 539-42. Fejdi, P., Poullen, J.-F., and Gasp6rin, M. (1980) Bull. Sameshima, T., Henderson, G. S., Black, P. M., and MinOral. 103, 135-8. Rodgers, K. A. (1985) Mineral. Mag. 49, 81 5. Fleischer, M. (1975) Glossary of mineral species. Tien, P., and Waugh, T. C. (1969) Am. Mineral. 54, Mineralogical Record, Tnscon. 1355 62. Gamidov, R. S., and Mamedov, K. S. (1960) Azerb. Khim. Vochten, R., Grave, E. de, and Stoops, G. (1979) Neues Zh. 4, 121 5. Jahrb. Mineral. Abh. 137, 208-22. Henderson, G. S., Black, P. M., Rodgers, K. A., and Zwann, P. C., and Kortenburg van der Sluys, G. (1971) Rankin, P. C. (1984) New Zealand J. Geol. Geophys. Scripta Geol. 6, 1-7. 27, 367-78. Hey, M. H. (1962) An index of mineral species and varieties arranged chemically. British Museum (Natural History), [Manuscript received 13 January 1986; London. revised 19 February 1986]