American Mineralogist, Volume 6E, pages 262-276, 19E3 Calciobetafite(new mineral of the pyrochlore group) and related minerals from Campi Flegrei, Italy; crystal structures of polymignyte and zirkelite: comparisonwith pyrochlore and zirconolite FronBNzo Mnzzt C.N.R. Cento di Studio per la Cristallografia Strutturale Istituto di Mineralogia dell'Universitd Via A. Bassi 4. Pavia, Italy aNo Roselse MUNNo Istituto di Mineralogia dell'Universitd Via Mezzocannone 8, Naples, Italy Abstract Polymignyte, zirkelite and zirconolite have often been considered the same mineral. Their occurrence, with calciobetafite (a new member of the pyroclrlore group), in a "sanidinite" from Campi Flegrei has allowed their crystal-chemical study and identifica- tion as three distinct phases.The three mineralsare polymorphs of the compound: (Ca,Na, REE,Th . )yrrtzrytr(Ti,Nb, . )Y\Fe,Ti)v,Ivo1a. The crystal structure of zirconolite (spacegroup C2lc)has been previously determined (Gatehouseet al., l98l) on synthetic crystals;those of polymignyte(e.g., Acam) and zirkelite (e.9., P3Q) are describedin the present paper. The crystal structures of polymignyte, zirkelite and zirconolite may be derived from that of pyrochlore. Chains, formed by distorted (Ca,REE . ) cubes alternating with (Ti,Nb . ) octahedra in pyrochlore, are replaced in the remaining minerals by either chains of Zr polyhedra with seven vertices or chains in which (Ti,Nb . ) octahedra alternate with distorted (Fe,Ti . ) tetrahedra or trigonal bipyra- mids. The different arrangementof thesechains gives rise to the diferent symmetriesof the three phases.The crystal structures of zirkelite and zirconolite are very similar, as they ditrer only in the stacking of identical pairs of layers of polyhedra. The difficulties in distinguishingthe three polymorphs by X-ray powder diagramsare discussed. Introduction nized on the basis of their morphology and some Calciobetafite,a new speciesof the pyrochlore optical features:(1) small (0.1-0.2 mm) octahedral group, and related minerals have been found at crystals, reddish brown in color and isotropic; (2) Monte di Procida (Campi Flegrei, Campania,Italy) elongated prisms (maximum 0.4 mm) enclosed in in a rock known as "sanidinite". This subvolcanic sanidine and sometimes in the interstitial glass rock is present in a phreatomagmatic explosion bordering the K-feldspar, dark red in color, show- breccia. The age of this pyroclastic formation, ing parallel extinction and very weak pleochroism; determined on an alkali-trachitic obsidian by the IU (3) platy crystals with roughly hexagonal outline; Ar method (Gillot, pers. comm.) is 0.035 m.y., luster resinous, brittle with splintery fracture; black whereas the absolute age of sanidinite, determined in color and reddish brown in very thin splinters, with the same method on the enriched feldspathic showing birefringence with extinction parallel to an fraction, is 0.084 (i0.008) m.y. (Civetta, pers. edge. comm.). The sanidiniteis composedof 75% sanidine By X-ray analysesthe three specieswere identi- (Oros), l6Vo plagioclase(An35), Mg-hastingsitic am- fied as a member of the pyrochlore group, polymig- phibole (occasionallywith a core of clinopyroxene), nyte and zirkelite respectively;the last mineral is biotite, magnetite, apatite and sphene.Minor inter- always the dominant phasein intergrowths of zirke- stitial glass is also present. Occasional small col- lite, zirconolite and pyrochlore. Conflicting data ored crystals were found scattered throughout the have been reported in the literature concerning rock: three possible specieswere tentatively recog- polymignyte, zirkelite and zirconolite, which have m03-004>u83/0102-0262$02.00 262 MAZZI AND MUNNO: PYROCHLORE AND RELATED MINERALS 263 often been considered the same mineral. As it will zirkelite and also gave the axial ratio a:c: be seen later, distinguishing among these three l:1.1647. Actually the zirkelitesfrom the two minerals is very dfficult: fortunately their occur- localities appear different with respect to their con- rence in the sanidinite, as well-developed crystals tents of Zr and Ti (Table l), even if the sum of the (not in their usual metamict form), allowed their two oxides and amounts of REE, Ca, Fe and Th are crystal-chemicalstudy and consequentlytheir iden- similar. Borodin et al. (1956) described the new tification as three distinct phases. mineral zirconolite, pseudo-cubic and metamict, similar to zirkelite in severalproperties, but consid- identification of the specimens Previouswork and ered it to be a different speciesbecause of the higher Polymignyte, found in a pegmatite from Fre- REE and lower U and Th contents of the former. dricksviirn (Norway) was first describedby Belzeli- These authors suggested the ideal formula unit: us (1824);more detailed studiesof its morphologi- CaZrTi2OTforzirconolite. Further datawere given by cal, physical and chemical properties were reported Borodin et al. (1961),who also describeda Nb-rich by Br<igger(1890). This material is metamict; Lima- zirconolite. In studying the system CaGZrOT- de-Faria(1958,1964), on the basisof X-ray powder TiOz, Pyatenko and Pudovkina (l9U) synthesized diagramsfrom the Fredricksvbrn mineral heated at the compound CaZrTi2OT, which they identified 700, 1000and 1300'C (Table 17), concluded that with the natural zirconolite (or zirkelite, which they polymignyte and zirconolite are synonymous. Pu- consideredsynonymous). This material is mono- dovkina et al. (1969)calculated approximate lattice clinic:a : 12.43,b : 7.26,c : 11.35A,F : 100.34". parameters(a : 10.31,b : 14.48,c: 7.41A)from They noted the structural similarities of zirconolite the axial ratios given by Brogger (1890)and the unit and pyrochlore. In his survey of the minerals of the cell volume, indirectly obtained from molecular pyrochlore group, Hogarth (1977)considered zirke- weight and density. lite and zirconolite as synonyms. Recently, Gate- The identification of the mineral from Campi house et al. (1981),working on syntheticcrystals, Flegrei with polymignyte was made through a com- determined the crystal structures of two zircono- parison of its properties with those given by Br<ig- lites: CaZr*Ti3-*O7.The lattice parameters are as ger: color, luster and particularly the [100] elonga- follows:a: 12.445;b :7.288; c = 11.4874,F: tion of the crystals, the orthorhombic symmetry 100.39"for x : 1.304,and a: 12.444;b :7.266; andthe axialratios (a:b:c :0.712 : I : 0.512for the c : ll34lA, F : 100.59ofor x : 0.850.Finally the Norwegian polymignyte) are the same; also the refinement of this monoclinic structure was carried stoichiometric unit calculated from the chemical out by Sinclair and Eggleton (1982) on natural analysis by Brogger agreeswell with that obtained crystals of "zirkelite" from Kaiserstuhl. from our crystals (Table l). Finally, the measured The examinationof the platy crystals from Campi lattice constants are similar to those calculated by Flegrei led to a clarification of some of the above Pudovkina et al. (1969). contradictions. This material always appearsas an The definitions of zirkelite and of zirconolite have oriented intergrowth of three crystalline phases: been a topic of debate and several studies have subordinateamounts of pyrochlore and of a mono- reached contradictory conclusions. According to clinic zirconolite phase (Pyatenko and Pudovkina, Hussak and Prior (1895),who first describedthe 1964and Gatehouse et al.,l98l) and, dominantly,a mineral from Jacupiranga(Slo Paulo, Brazil), zirke- trigonal phasewith an axial ratio a : c : | : 2.317: lite occurs as cubic crystals flattened parallel to an 1 :2 x 1.159,fairly similar to that given by Blake octahedral face, black in color, transparent in very and Smith (1913) for zirkelite from Ceylon. We tiny splinters, with a dark brown color. A complete suggest, therefore, the retention of the name zirke- chemical analysisof this material was subsequently lite for this trigonal phase. given by Prior (1897)(Table 1). Blake and Smith Chemical cornPosition (1913),after having reexamined the morphology of the Brazilian type material, considered some crys- Chemical compositionswere determinedby elec- tals from Walaweduwa(Ceylon) to be zirkelite. The tron microprobe analysis. However, the scarcity of latter were hexagonal(0001) plates or [0001]prisms, crystals, the large number of chemical components, brownish-black in color, opaque or sometimesred- and the common lack of suitable standards for dish-brown in thin section. Blake and Smith sug- unusual chemical elements made it difficult to ob- gested either hexagonal or trigonal symmetry for tain completely satisfactory results: the sum of the 264 MAZZI AND MUNNO: PYROCHLORE AND RELATED MINERALS Table l. Calciobetafite,polymignyte, zirkelite and zirconolite: chemical analysesand formula units Ca-betafite Polymignyte Zirkelite md Zirconolite ab Fr Bzk Czk Czk* czk** zi zi* CaO 16.0 16.0 7.6 8.6 6. 98 8.6 10.79 6,47 8.55 8. 18 11 .05 11 .0O Na2O o.4 o. 59 o.37 1.40 La203 o.7 0.7 n'a o.a 5.13 o.7 tl CeO2 4.O 4.0 n.a 3.7 6.20 2.6 2.64 I I 6.52 4.19 t2.64 lO-32 Pr2O3 n.a 0.6 n. a 0.6 n. a. tl NdzOo r.2 \,4 n.a 2.3 2.r tt Yzo: 0.6 0.6 n'a 2,5 2,26 2.4 0.2L 1.O8 ThO2 n.a 4.5 n.a 3.5 3,92 n.a 7.31- 20.44 8.33 o.23 o.58 2.90 uo^ n. a 4.8 n. a o.7 n.a 1 .40 r .o2 4.48 13.77 r.47 0.34 ZrOz 1.1 1.O 25.4 25.4 29.71- 23.7 52.49 30.73 34.19 35.27 32,44 25,00 MnO n.a O.9 n.a I .32 0.03 0.06 0.3a Mgo n.a O.2 n.a o. 16 o.22 2.34 1.33 1 .96 0 .45 Fe0 1.9 2.O 8 .7 8.0 2.OA 4,1 7 .72 4.O7 4.72 3.73 6.