K.A.Savko·S.M.Piliugin·N.S.Bazikov

Experimental data for high-temperature decomposition of natural celadonite from banded iron formation

K.A.Savko1·S.M.Piliugin1·N.S.Bazikov1

Three experiments were set up to evaluate conditions for the high-temperature decomposition of celadonite from a banded iron formation in an alumina-free system and identify its decomposition products.It was estimated that at 650 and 750°C，with a NiNiO buffer and pressure of 3 kbar，celadonite completely decomposes and the decomposition products were tetraferribiotite，magnetite and quartz.Under more oxidizing conditions（hematite-magnetitebufferinsteadofNiNiO），ferrous potassium feldspar sanidine forms instead of magnetite. During the celadonite decomposition in oxidizing conditions more magnesian and aluminous tetraferribiotite，along with ferrous sanidine，are formed than at reducing conditions.

Experimental dissolution·Celadonite·

1 Introduction

The alkali-enriched banded iron formation（BIF）rocks are common and an important constituent in the Precambrian iron basins.We have identified these BIF as subalkaline type BIF.Rocks of this type are particularly common in the Paleoproterozoic basins of Hamersley，Western Australia（Trendall and Blockley 1970；Miyano and Klein 1983），Transvaal，South Africa（Beukes and Klein 1990）and in the Paleoproterozoic basin of the Kursk magnetic anomaly（KMA），Russia.Subalkaline BIF contains widespread riebeckite，aegirine，celadonite，tetraferribiotite，and Alfree chlorite instead of grunerite，stilpnomelane，minnesotaite，and greenalite，which are the usual minerals in BIF elsewhere.At the KMA iron deposits，BIF with alkali amphibole were metamorphosed at 370—520°C and 2—3 kbar（Savko and Poskryakova 2003）.

In the KMA，BIF green mica，which is responsible in composition to the celadonite with the formula KFe3+（Mg，F(xiàn)e2+）=［Si4O10］（OH）2，is quite abundant（Savko and Poskryakova 2003）.Celadonite forms emerald-green scales sizing from a few tenths to 1.5—2 mm，composing up to 30%—40%modal.It intimately intergrows with riebeckite，and the size of their crystals（plates，plaques）in the magnetite layers are larger than those in the quartz layers. The intergrowths of celadonite and riebeckite can form the edges around the coarse grains of carbonate，separating themfromthegranoblasticquartzaggregate.The riebeckite pseudomorphs occur on celadonite，and very small needles（‘seeds''）are found between the riebeckite grains of magnetite and celadonite.Often flakes of celadonite develop on carbonate.The magnetite-celadonite intergrowths form between carbonate and silica matrix. Compositionally，natural celadonite from BIF is extremely low in Al，moderately ferruginous（XFe=0.27—0.41）（Table 1）and is close to ferroceladonite of isomorphic celadonite—ferroceladonite series（Fig.1）.

In addition，BIF celadonite is found in hydrothermally altered mafic volcanicrocks，in sedimentary rocks，especially those with a tuffaceous component under diagenetic or very low-grade metamorphic conditions（Coombs 1954；Wise and Eugster 1964；Boles and Coombs 1975）corresponding to the zeolitefacies，andlesscommonlyintheprehnite-pumpellyite andlawsonite-albite-chloritefacies（Landis1974）；howeverit is not found in the greenschist facies（e.g.，Wise and Eugster1964）.Two new minerals of the celadonite family—ferroceladoniteandferroaluminoceladonite—werediscoveredinthe Triassic tuffs（New Zealand）and metamorphosed at the zeolite facies conditions（Li et al.1997）.

Table 1 Chemical composition（wt%）of natural celadonites and synthesized tetraferribioties

Fig.1 Compositions of celadonite from KMA BIF shown in a classification plot

Unfortunately，over the past twenty years，no new studies on celadonite behaviour during metamorphism and its thermal stability have appeared.This article is the first attempt to fill this gap.This topic is very important because currently it is unknown if the minerals，which are the low temperature equivalents of potassium Al-free mica，are unknown as minerals such as glauconite and nontronite contain alumina.In addition，it is not clear what minerals and（or）mineral parageneses are formed by their thermal decomposition due to metamorphism，because potassium pyroxenes，potassium Al-free amphiboles and purely ferruginous K-feldspars are unknown in nature.

Theaimofthisexperimentalstudyistoevaluateconditions for the high-temperature decomposition of celadonite from BIF in an Al-free system and identify its decomposition products.

2 Research procedure

Fig.2 X-ray diffraction pattern of natural celadonite from KMA BIF.Anode type—Co

The basis of specimens for the experiments is celadonite monofraction from the borehole 466-R core，sampled at 433 m（Mikhailovsky iron deposit）.This ferruginous quartzite contains 25%modal celadonite.The sample was pre-crushed in the cast iron pounder，then the obtained powder was sieved on sieves（of diameter 0.316，0.2，0.15，0.1 mm）.Then the electromagnetic fraction was extracted from the 0.316 and 0.2 fractions and，finally，100 mg of celadonite was handpicked under a binomagnifier.Only 40 mg of celadonite was used for the experiments.The composition of natural celadonite is shown in Table 1，the X-ray difraction pattern in Fig.2.The mixture was thoroughly mixed and grinded to a powder with an alcohol in the agate pounder.Then it was put into the platinum capsule.The quantity of the initial components calculations were based on the coefficients of the estimated reaction：Sld+Qtz+Mag=Fe-Kfs+Gru（Hyp）+H2O，and the molecular weights of Fe，Si，O，K，Mg components. Distilled water was added，and its quantity was consequent to the physical parameters calculations based on the assumed reaction.The platinum capsules were welded and put into gold capsules with a larger diameter.To set the redox conditions，the powder of hematite or NiNiO was used as a buffer，and it was placed between the gold and platinum capsules（Fig.3）.The capsules were weighed，sealed and placed into the rig‘high gas pressure apparatus.''Three experiments were conducted：two of them were held with a NiNiO buffer at a pressure of 3 kbar and temperatures of 650 and 750°C.The values of the temperature and pressure remained constant for 8 days.The third experiment was conducted with a Hem-Mag buffer under the following conditions：T=700°C and P=1 kbar.The data of the temperature and pressure was constant for 4 days.Then the capsules were recovered from installation and were quenched and carefully opened.This material was extracted for further research.Some of this material was investigated with an X-ray diffractormeter DRON-4 to clarify the crystal structure of the mineral phases.High-qualityexperimentalpowderX-ray diffraction data for the compounds were obtained with an automatic diffractometer DRON-4（Russia）（fine-focus sealedtube，CoKα1radiation（λ=0.178897 nm），Johanson's Hybrid Ge｛111｝monochromator for the primary beam，Bragg—Brentano geometry）.The patterns were scanned in reflection mode，θ/2θ continuously scanned over the angular range of 3°—40°（2θ）with a step 0.01°（2θ）and counting time of 1000 s·step-1.Preferred orientation effects were reduced by grinding.Alignment and calibration were checked using Al2O3（SRM676）.Diffraction data was collected at room temperature（296 K）.The extraction of peak position for indexing was performed with the Pawley method.Unit cell parameters were refined by a least-squares fitting of Bragg's equation to the position of the diffraction lines.All calculations for the refinement of the diffraction patterns and the refinement of the unit cell parameters were performed using the complex programs available in PC software：‘HighScore Plus''supplied by PANalytical EMPYREAN ［Version：3.0.t（3.0.5），Date 30-01-2012.Produced by：PANalytical B.V.Amelo，The Netherland］.Another part of this material was used to prepare specimens in order to study the composition and morphology of the mineral individuals in the TESCAN VEGA II XMU electron microscope with an INCA X_SIGHT EDS analyzer at the Institute of Experimental Mineralogy，Russian Academy of Sciences.BSE images and spot analyses of minerals were obtained at a 20 kV accelerating voltage，a beam current of 1.2 mA，counting time of 70 s，and beam 1—3 μm in diameter.The ZAF corrections were introduced during the calculation of oxide concentrations and the installed software of the analytical system assayed the accuracy of the analyses.The accuracy of the analyses was systematically monitored by the replicate analysis of natural and synthetic standards.The cation proportions（crystal-chemical formulae）were calculated by normalizing to four oxygen atoms for magnetite，eight oxygen atoms for sanidine，and eleven oxygen atoms for tetraferribiotite and celadonite.

Fig.3 Scheme of ampoule filling with initial mixture of celadonite and water

Fig.4 Mineral individuals obtained during an experiment under 750°C：a，b，d—the obtained mineral；c—general view of the sample；e，f—mineral forms of tetraferribiotite in the sample.Point numbers correspond the numbers of analyses in Table 1

Fig.5 X-ray diffraction pattern of the sample obtained under 650°C

3 Experimental results

After experimenting at T=650 and 750°C（NiNiO buffer），a mixture of tetraferribiotite，quartz and magnetite was obtained，with the quantity of tetraferribiotite dominating over the other phases（Fig.4）.The tetraferribiotite d001values are 10.169 A°（at 650°C）and 10.189 A°（at 750°C）（see Fig.5，6）.The natural tetraferribiotite of Hamersley Basin BIF（Western Australia）is characterized by d00110.159，10.163 and 10.180 A°（Miyano and Miyano 1982）. After experimenting at T=700°C and 1 kbar（Hem-Mag buffer），a mixture of ferrous potassium feldspar，tetraferribiotite，quartz and magnetite was created（Fig.7）.

K-feldspar is made up of small grains of 5—10 μm（Fig.7）.In composition，this ferruginous potassium feldspar contains Fe2O3from 10.3 to 18.7 wt%，which corresponds to 40.7—76.6 mol%minal of ferrous potassiumfeldspar（Table 2）with 3.65 to 9.61 wt%Al2O3.The quantities of albite and anorthite components are 0.1—4.9 and 1.0—5.3 mol%，respectively.

Fig.6 X-ray diffraction pattern of the sample obtained under 750°C

Fig.7 The obtained mixture of ferrous K-feldspar，tetraferribiotite，quartz and magnetite.Point numbers reflect numbers of analyses in Tables 1 and 2

Table 2 Chemicalcomposition　　（wt%）ofsynthesized　　ferrous sanidines

X-ray diffraction study of K-feldspar（Fig.7）has shown that K-feldspar is made up of ferrous sanidine，for which the resulting diffraction pattern on the lattice parameters were calculated using UnitCell（Holland and Redfern 1997a，b）：a0=8.634（5），b0=13.276（5），c0=7.194（5）β=116.79（2）°，V=736.2These values are close to the unit lattice parameters for the refined structure，synthesized low sanidine which has the following composition：K0.93（Al0.75Fe0.27）Si3.01O8（Nadezhina et al.1993）：a0=8.627，b0=13.058，c0=7.209β=116.00°，V=730.0

Tetraferribiotite synthesized with Fe-sanidine appears in the shape of elongated 15—20 microns flakes（Fig.7）.On the diffraction pattern it is defined by a clear characteristic reflex of 10.234（Fig.8）.The tetraferribiotite obtained in the experiment is characterized by low-Al（Al2O3=0.59—0.68 wt%）and iron（0.243—0.339 at.u.），due to the significant prevalence of Fe3+over Fe2+in the structure（Table 1）. Tetraferribiotite synthesized at more reducing conditions（buffer NiNiO）has significantly more iron（0.696—0.740 at. u.）and alumina（Al2O3=2.10—2.62 wt%）（Table 1）.In composition，tetraferribiotites synthesized under more oxidizing conditions（Hem-Mag buffer）are closer to the natural tetraferribiotites in the KMA BIF（Savko and Poskryakova 2003），with higher concentrations of Fe3+and Mg and lower Fe2+，and approaching tetraferriphlogopite（Fig.7）.

4 Discussion

Fig.8 X-ray diffraction pattern of synthesized ferrous sanidine，tetraferribiotite and quartz

Tetraferribiotite，a fairly rare Al-free mica，with a formula of K（Mg，F(xiàn)e2+，F(xiàn)e3+）3［Fe3+Si3O10］（OH）2，was first discovered in 1955 during an unsuccessful experimental effort to synthesize K-OH amphibole（Veres et al.1955）.One member of the tetraferribiotite group，ferriannite，was synthesized in the experiments of D.R.Wones（1963）within the temperature range of 400—850°C，pressure range of 1035—2070 bar，and an oxygen fugacity between the hematite—magnetite and iron—wuestite buffers.He has suggested that mica of ferriannite composition should occur in magmatic Fe-rich rocks.But later ferriannite was described in the riebeckite-bearing rocks in the BIF of Western Australia（Miyano and Miyano 1982）and South Africa（Miyano and Beukes 1997）.

Tetraferribiotite occurs in the iron formation of the Mikhailovskoe mining district as small reddish-brown flakes，0.2—1.0 mm across，in association with magnetite，hematite，riebeckite，aegirine，celadonite，and carbonate. Compared with celadonite，the mineral occurs in subordinate amounts.It is preferably localized in magnetite-hematite layers as disseminated flakes，although it can occasionally also be composed of nearly monomineralic thin layers with merely minor amounts of celadonite and iron oxides.Tetraferribiotite sometimes forms inclusions in aegirine，although these rocks may not contain this mineral in the groundmass.

Tetraferribiotite in the KMA BIF is less aluminous（Al2O3=0.68—0.76 wt%）and more magnesian（MgO= 10.13—14.09 wt%）than natural ferriannite（Miyano and Miyano 1982；Miyano and Beukes 1997），whose Al2O3content is always higher than 1.3 wt%，usually about 4—5 wt%，and whose MgO concentration varies from 3.5 to 12.5 wt%（Fig.9）.TheSi/Alratioofthetetraferribiotiteisalways one order of magnitude higher than 3.0.Thus，the compositionsoftheKMABIFtetraferribiotiteareclosesttothoseof the end member of the annite-ferriannite series，and shift toward tetraferriphlogopite in the tetraferriannite-tetraferriphlogopite series（Fig.9）.

Fig.9 Compositions of tetraferribiotites shown in a classification plot.1—KMA BIF（Savko and Poskryakova 2003）；2—Penge iron formation，South Africa（Miyano and Beukes 1997）；3—Dales George iron formation，Western Australia（Miyano and Miyano 1982）；4—synthesized under 650°C（NiNiO buffer）；5—synthesized under 750°C（NiNiO buffer）；6—synthesized under 700°C（Hem-Mag buffer）

The results of the experiments carried out at 650—750°C and under different redox conditions（NiNiO and Hem-Mag buffers）suggest that，at smaller values of fO2（buffer NiNiO），tetraferribiotite and quartz are formed by celadonite decomposition.With higher values（Hem-Mag buffer）the breakdown products are tetraferribiotite，ferrous sanidine and quartz.The celadonite breakdown reactions in the system SiO2—Fe2O3—FeO—K2O—O—H for different fO2—（1）and（2）for NiNiO buffer and（3）to Hem-Mag buffer are shown below. andFe2+/（Fe2++Mg）decreasedaccordingly.Lower alumina content in tetraferribiotite synthesized on the Mag-Hem buffer is due to the simultaneous formation of ferrous sanidine，which contains from 3.6 to 9.6 wt% Al2O3（Table 2）.Thus，a small amount of the alumina from decomposed celadonite became included in the structure of the formed ferrous sanidine.

Similar results were found earlier for the thermal decomposition of synthetic celadonite under different redox conditions（buffers QFM，NiNiO，Mag-Hem）（Wise

The celadonite breakdown reaction（3）to form ferrous potassium feldspar and tetraferribiotite is illustrated on the diagram（Mg+Fe2+）—Si—（Al+Fe3+）（Fig.10）.

It is important to point out that，at more oxidizing conditions（Hem-Mag buffer）with Fe-sanidine，more Mg and Al enriched tetraferribiotite is formed than at reducing conditions（NiNiO buffer）.In addition，during celadonite decomposition（NiNiO buffer）ferrous feldspar was not obtained.Differences in the iron content in tetraferribiotite were due to the fact that，under more oxidizing conditions，the ratio of Fe3+/Fe2+increased and Eugster 1964）at pressures of 1—3 kbar.However，as noted by the authors，the experimentally determined upper limit of celadonite stability is about 410°C and doesn't correspond to the breakdown temperature of celadonite in natural systems.Moreover，they believed that，in nature，celadonite decomposes at temperatures below 400°C on the boundary of the zeolite and greenschist facies.

Fig.10 Celadonite breakdown with ferrous K-feldspar and tetraferribiotiteformation shownin the（Mg+Fe2+）—Si—（Al+Fe3+）diagram

The evaluation of metamorphic conditions for the Paleoproterozoic KMA BIF show that celadonite in the ferruginous quartzites was stable during metamorphis above 520°C at a pressure of 3 kbar（Savko and Poskryakova 2003）.

5 Conclusion

As a result of conducting an experiment at 700°C and 1 kbar with a Hem-Mag buffer，natural dioctahedral mica Alfree celadonite was completely decomposed to form ferrous sanidine（Fe—Fsp=0.407—0.766 at.u.）and trioctahedral Al-free mica（tetraferribiotite）.In the experiments at 650 and 750°C and 3 kbar，with a NiNiO buffer，the complete decomposition of celadonite gave tetraferribiotite，magnetite and quartz.Thus，according to the results ofourexperiments，wecandrawthefollowing conclusions：

（1） Ferrous sanidine，tetraferribiotite and quartz are formed during the thermal dissolution of celadonite at relatively high values of fO2—on hematitemagnetite buffer.At more reduced conditions（buffer NiNiO），potassium feldspar is not formed and the breakdown products are tetraferribiotite，magnetite and quartz.

（2） During celadonite decomposition in more oxidizing conditions（Hem-Magbuffer）with Fe-sanidine，more Mg and Al enriched tetraferribiotite is formed than at reducing conditions（NiNiO buffer）.

Beukes N，Klein C（1990）Geochemistry and sedimentology of a facies transition from microbanded to granular iron-formation in the earlyproterozoic transvaal supergroup，South Africa. Precambrian Res 47：99—139.doi：10.1016/j.bbr.2011.03.031

Boles JR，Coombs DS（1975）Mineral reactions in zeolitic Triassic tuff，Hokonui Hills，New Zealand.Geol Soc Am Bull 86：163—173

Coombs DS（1954）The nature and alteration of some Triassic sediments from Southland，New Zealand.Trans R Soc N Z 82：65—109

Holland TJB，Redfern SAT（1997a）Unit cell refinement from powder diffraction data：the use of regression diagnostics.Miner Mag 61：65—77

Holland TJB，Redfern SAT（1997b）UNITCELL：a nonlinear leastsquares program for cell-parameter refinement implementing regression and deletion diagnostics.J Appl Cryst 30：84

Landis CA（1974）Stratigraphy，lithology，structure，and metamorphism of Permian，Triassic，and Tertiary rocks between the Mararoa River and Mount Snowdon，western Southland，New Zealand.J R Soc N Z 4：229—251

Li G，Peacor DR，Coombs DS，Kawachi Y（1997）Solid solution in the celadonite family：the new minerals ferroceladonite，K2-Fe22+Fe23+Si8O20（OH）4，and ferroaluminoceladonite，K2Fe22+-Al2Si8O20（OH）4.Am Mineral 82：503—511

Miyano T，Beukes NJ（1997）Mineralogy and petrology of the contact metamorphosed amphibole asbestos-bearing Penge iron formation，Eastern Transvaal，South Africa.J Petrol 38：651—676

Miyano T，Klein C（1983）Conditions of riebeckite formation in the iron formation of the Dales Gorge Member，Hamersley Group，Western Australia.Am Mineral 68：517—529

Miyano T，Miyano S（1982）Ferri-annite from the Dales George Member iron-formations，Wittenoom area，Western Australia. Am Mineral 67：1179—1194

Nadezhina TN，Pushcharovskiy DY，Taroev VK，Tauson VI，Bychkov AM （1993）Crystal structure of ferrialuminosilicate low sanidine.Kristallografiya 38（6）：77—82（in Russian）

Savko ΚA，Poskryakova MV（2003）Riebeckite—aegirine—celadonite BIF at the Mikhailovskoe Iron Deposit of the Kursk Magnetic Anomaly：phase equilibria and metamorphic conditions.Petrology 11（5）：426—443

Trendall AF，Blockley JG（1970）The iron formation of the Precambrian Hamersley Group，Western Australia with special reference to the associated crocidolite.Western Australia，Geological Survey of Western Australia

Veres GI，Merenkova TB，Ostrovsky IA（1955）Artificial pure ferrous hydroxyl mica.Dokl Akad Nauk SSSR 101（1）：147—150（in Russian）

Wise WS，Eugster HP（1964）Celadonite：synthesis，thermal stability and occurrence.Am Mineral 49：1031—1083

Wones DR（1963）Phase equilibria of‘ferriannite''，KFe3+2Fe+3-Si3O10（OH）2.Am J Sci 261：581—596

10.1007/s11631-015-0066-2

5 March 2015/Revised：1 May 2015/Accepted：6 July 2015/Published online：17 July 2015

? N.S.Bazikov

nickolasss@yandex.ru

1Voronezh State University，Universitetskaya Square，1，Voronezh 394088，Russia

?Science Press，Institute of Geochemistry，CAS and Springer-Verlag Berlin Heidelberg 2015

Tetraferribiotite·Banded iron formation·Sanidine