THE AMERICAN MINERALOGIST, VOL 47, NOVEMBER_DECEMBER, 1962

ON THE CRYSTAL CHEMISTRY OF HISINGERITE BBNcr LrNoqvrsr aNn SvBw JlwssoN, Mineralogical-GeologicalInst., Uniaersity of Llppsala, Llppsal'a,Swed.en., Diaision oJ Chemicol Technology,The Royal Inst'itute of Technology,Stockhol'm, Swed'en. Assriecr In place of the current conception of hisingerite belonging to the montmoriilonite group, a tentative proposal is ofiered that it is principally a mineral with extensive substitutions of Fe for Si (the reason for the cryptocrystalline character) and with inter- layer hydronium ions.

INrnooucrroN Referring to the new data for hisingerite and neotocite newly pub- Iishedby Whelan and Goldich (1961),the writers would like to comment on this subject. This paper will deal with the t04ay ar'd electron diffrac- tion data of hisingeriteand associatedchlorite, but new aspectsof the of hisingerite will also be presented. The results are based upon an investigation of nine hisingerite samples-most of them being manganese varieties (neotocites)-from different localities in Sweden.With a few exceptionsthe samplesare characterizedmacro- scopicallyby a more or lessplaty appearance(Fig. 1). For those speci- mensmentioned in the text, the following notationshave beenused.

Frc. 1. Platy neotocite (Spec. K8) Photo: G. Andersson. 1356 EISINGERITE 1357

(MGIU : Mineralogical-Geological Inst., University of Uppsala) (RMA-Dept. of Mineralogy, Swedish Museum of Natural History, Stockholm) Spec. Hl: ferrous hisingerite (on pyroxene skarn) from the Sjdstriim Mine, Hofors, G?istrikland. Spec. 54: ferric hisingerite from the Solberg Mine, Elvestorp, Viistmanland (coll. MGrU 400/2). Spec. G5: neotocite from the Gillinge Mine, Sviirta, S

X-nay Drrnn,lcrroN Dara Compilations of available r-ray data for hisingerite show that the generalfeatures are in good agreement,in spite of their vagueness.In the present case,four broad and diffuse bands were recognizedin the powder diagrams of all the samples.Their interplanar spacingas measuredto the centresof the bandsvaried within the following limits: l) 4.3 to 4.6 A, 2) 3.55to 3.63A, ,2.57 to 2.63A, and 4) 1.54to 1.60A. The.e valuesare more or lessin accordancewith previousdata exceptthat the last figureis somewhattoo high. In some cases,additional lines were also found, i.a. at about 7.3 A. This line is also reported by Whelan and Goldich and is attributed by them to the formation of an "incipient chlorite structure." This is certainly an adequateexplanation. It will be shown, however, that the 3.6 A band is also a "false" one and dependson the presenceof chlorite. The occurrenceof a chlorite phasein the hisingeritematerial has been positively establishedfor many of the present samples.By mounting specially chosenmineral fragments large enough to furnish a usable plane surface, r-ray diffractometer traces typical for chlorite have been ob- tained. Owing to the enhancedpreferred orientation, the basalreflections are well recorded.All of the r-ray diffractometer tracesfoilow the scheme of Fig. 2, in which the weaknessof the first and third order basal reflec- tions is most prominent, indicating a high- or high-manganesechlorite. The peaks are generally more or less broadened but from the position of 002 reflectionsrecorded at up to 7.28A the chloritesoften seemto be of a rather "wide-spaced" variety, maybe approaching gonyerite, the manganesechlorite from Lingban (Frondel, 1955). However, no signifi- cant variations between chlorites of Mn-rich and Mn-poor hisingerites were reproducible. 1358 B. LINDQVIST AND S. "r,4/rSSO1v

It should be emphasizedthat the chlorite, giving no impression of being a secondarycoating, seems to be related to the hisingeritein a most intimate way. The broadnessof the 3.6 A line on the r-ray powder photo- graphsalso suggests a cryptocrystallinecharacter of most of the chlorite.

ElncrnoN Drnrn.q.crroNDer.q. As it was felt that the *-ray data are too few to permit any unequiv- ocal interpretation, the attempt has been made to obtain electron diffraction patterns in the hope that additional information could be won. A pieceof hisingerite(Spec. 54) was ground to a fine powder which was pressedinto a reflection specimenholder or coilectedon a copper

717A [ 3 s?A li n l\ r\]\ril i' t.tait l\ r:sa / l r\ l \ zeoi 'n""* A / ' \q^^- ^*'{"rt*^^ tr"."*^/ \rr,^'J \4 -***/ n-u'"u '10 35 Degrees20 CuK*

Ftc.2. X-ray diffractometer curve from a plane surface of Spec. 54. electron-microscopygrid. Placedin the electronbeam of a Triib, Tduber & Co. electron diffraction apparatus, the specimen turned out to be highly insulating, which causedconsiderable trouble. With the help of a chargeneutralising device it was possibleto obtain at least one goodpat- tern characteristic for hisingerite. A pattern of silver was taken as standard for the measurements.As might be expected there is some sharpeningof the linesas comparedto the r-ray powder photograms.For further comparisonsTable 1 is to be consulted.The 3.6 A band is evi- dently proved to be alien to hisingerite and must be attributed to chlorite. Moreover, the resemblanceto the hectorite data is unmistak- able. For the hectorite Nagelschmidt (1938) has stated the lengths of o and b axesto be 5.24and 9.16A. Thus, the sameapproximate values are applicable to the hisingerite specimen.

Cnvsrer. SrnucrunB The similarity to where r-ray data are concernedIed Gruner (1935) to introduce hisingerite as a species.This similarity is strengthenedhere, as already stated. In the absenceof basal HISINGERITE 1359

Taern 1. Drllnlcrron Dnre

(1) (2) (3)

4 llq /-) 4.s3 (s) 4.57 (vs) 11,02 3.63(m) 2.s7(m) 2.s8 (s) 2 6s (s) 13,20 2.285(vw) 22,04 1.73 (w) 1.748] , 24,3l (m, 1.68ef l.) 1.s4 (m) 1.53(m) 1.527 (vs) 33,06 1.38(vw) 1.34(w) 1.320(s) 26,40 1 267 (vvu) 35,17 , 42 0 98 (viv) 0.990 (w) 19,46,53 0. 89 (vw) 0 878 (w) 39,60

(1) X-ray difiraction data of Spec 54. (2) Electron diffraction data of Spec 54 (3) Non-basal x ray reflections of Mg beidellite (hectorite), Hector, san Bernardino Co., California (Nagelschmidt, 1938)

reflections,however, the relatedseries of hkbandscannotactually be used as a criterion for montmorillonite in particuiar but revealsonly a layer Iattice with analogousdimensions within the rayers.Besides montmoril- lonite, the data could, for example,flt mica as weil as chrorite.Further- more, the absenceof basal reflectionsrenders the glycolation technique unsuitablefor diagnosticpurposes (Dietrich, 1961). when consideringthe structure,the main questionarises as to why the stacking of the hisingeritelayers is so extremerypoor that basal reflec- tions cannot be recognizedby any means. According to the writers a restrictedcrystallite growth of hisingerite(in the direction of the c-axis) can possiblybe dependenton a rather extensivesubstitution of Fe for si. on a small scalethis could be realizedin montmorilionite, but attention shouldalso be paid to the mica alternative with its increasedpossibilities of tetrahedral substitutions.rn the latter casethe rower charge of the tetrahedral layers must be balancedby the supply of interrayer hydro- nium ions as neither potassium, sodium, nor calcium are present in sufflcient amounts. rn order to study the compositional variations, the chemicar anaryses of twelve hisingerites have been recalculated. rn Table 2 they are pre- sented as incomplete structural formulas (interlayer ions and omitted). The calculationsare basedupon the hypothetical assumption that the intralayer cations, as in true mica, constitate 2l valences per basic formula. Higher chargewill accordingly raise all the figuresand at the sametime raisethe ratio of octahedralto tetrahedralions. rn no caseis a si-index typical of montmorillonite attained (even after 1360 B. LINDQVIST AND S. "/,4/rSSOff

Tnr:-;n 2. DrsrtrsurroN ol INtn-a.llvnn CerroNs rlt HrsrNcnnttrs es B.lsno rrpox l Mrcl Monnr-

(1) (2) (3) (4) (s) (6) (7) (8) (e) (10) (11) (r2)

Tetrahedral Si 3.09 3 03 2.91 3 05 3 0? 3.09 3.1 3 20 3 25 3.28 3'30 3.32 AI o.2g o.2o 0.14 0.04 0 i2 0 05 0 34 0.39 Fe?+ o 62 0.77 o 95 0.91 0 81 0 91 0.84 0.80 0 75 0 72 0.-16 0.29

Octahedral Fe3+ 0 75 1.53 1 55 1.62 1.79 | 93 | 47 1.27 0 52 1 65 1.01 1.13 Fe2+ 094 0.11 011 034 016 0.05 032 032 1.75 006 036 042 Mg 0 86 0.48 0 60 0.20 0 10 000 035 068 030 031 0.96 0.66 Mn 003 002 005 001 0.05 003 007

Sum of octahedral ions 2.58 2.20 2 26 2.16 2 05 | 98 2.1.9 2 28 2 62 2.02 2 36 2,28

(1) Spec.Hl (incompleteanalysis) (2) Spec.54 (SiOr35.2r1, TiOr 0.00, AbOa 2.00,FezOr 35.51, FeO 2.40,MnO 0 34, MgO 3'77, CaO 199, NarO 0.00.KzO 0.04.HzO+ 10.42,HrO 9.03,COr 0.00, sum 100.74) (8. Almqvist and B. Lindqvist, analysts) (3) Nicholson Mine, Saskatchewan,Canada (Bowie, 1955) (4) Parry Sound,Ontario, Canada (Schwartz,1924) (5) Riddarhyttan, Sweden (Clewe and Nordenskiitld, 1866) (6) Kawayama Mine, Japan (Sudo and Nakamura, 1952) (7) Montauban-les-Mines, Qu6bec, Canada (Osborne and Archambault, 1950) (8) Saksagan,Krivoi Rog, Russia (Nikolsky, 1953) (9) East Mesabi, Minnesta, USA (Whelan and Goldich, 1961) (10) Blaine Co, Idaho, USA (Hewitt and Schaller,1925) (11) and (12) Beaver Bay, Minnesota,USA (Whelanand Goldich, 1961) correction for higher charge).A structural formula of montmorillonite can only be derived by equating trivalent iron to divalent as done by Whelan and Goldich. If, however, the main component is montmorillonite, it could occur more likely interstratified with chlorite. To fit the chemical data the ratio of montmorillonite to chlorite in that caseshould be about 2:l as a rough estimate. The main reason to present the chemical data in the way demon- stratedin Tabie 2 is to illustrate the possiblefit to the mica structure.The Si-indexgenerally coincides with, or approachesthe valuesof (Si 3.0) or "" (Si 3.3-3.4). Under the given assumptionsit is seen from the last row of Table 2 that dioctahedral varieties would dominate but therewould alsobe examplesof pronouncedlytrioctahedral character. In Table 2, manganesespecimens are not included becauseof the analytical diffrculties involved in determining the valence states of manganese.It can be mentionedthat applicationof the samecalculation method upon the neotocite sample analyzed by D. Thaemlitz and C' O' Ingamells and cited by whelan and Goldich would give a si-index dis- tinctly below the range of Table 2. Finally, it must be admitted that no major conclusionscan be drawn from the chemical analyses.It seems,however, as if the range of chemical HISINGERITE L36l variations are more likely ascribedto isomorphism of a poorly crystallized silicatethan to the occurrenceof (non-siiicate)impurities.

TnBnulr, Der.q. SeveralDTA curves have been published,but the reactionsrecorded have been interpretedin somewhatdifferent ways, and the data are not particularly illustrative. The writers have preferred to follow the thermal behaviorby meansof weight losscurves and r-ray phaseanalyses. In this way it has beenshown that the thermal stability of hisingeriteis remark- ably lower than what is to be expectedfor a montmorillonite. As to the dehydration stageit appearsthat almost all water (inclusive of hydroxyl) is driven off already on heating to 500oC. Contemporane- ously the crystalline structure is completely destroyedas witnessedby the vanishedhkbands. Thus, the destructiontemperature is considerably lower than of any known montmorillonite or mica, which might be cor- related to the exceptionalconditions of substitution. The recrystallizationprocess starts at about 800oC., principally with the formation of ferrites(and braunite).

C,s.rroNExcrraNcn C,\pA.crry As far as is known, the exchangecapacity of hisingeritehas never been investigated.Some preliminary testsof this pertinent property will there- fore be presentedhere. The powdered sampleswere Ca-saturatedby treating with calciumacetate solution (pH 8) and the amount of calcium subsequentlyreplaced by NHa+ ions was determined.By electrometric titration the number of hydrogen ions replaced during the first process also was determined.Perhaps there are some objectionsto the method used,but the following results will reasonablyindicate the right magni- tude. The figuresrefer to the total cation exchangecapacity and to re- leasedhydrogen (within parentheses),both expressedin meq/100 g. Spec. E9: 20 (10) Spec K8: 25 (s) Spec. 54: 25 (2s) Spec.G5: 50 (1s) Apparently the values lie within the range typical for illite and chlorite. The very high cation exchangecapacity of montmorillonite, consideredto be one of its most unique properties,has not been encounteredin any case.

CowcluorNc RBuanrs Even if thereare great difficultiesin obtainingpure hisingeritesamples, the amount of uncontroilable, amorphous impurities which may be 1362 B. LINDOVIST AND S. "r,41rSSOlr presentdoes not seemto be large enoughto affect the chemicalpattern. The fact that hisingeriterepresents a layer silicate cannot be doubted, but there are no unequivocalindications of it being a montmorillonite mineral, either from structural or thermal data, or from the study of infrared absorption spectra performed by Whelan and Goldich (1961). There are, on the contrary, some features which are incompatible with the montmorillonite theory, as for example, the low cation exchange capacity. It is implied above and here tentatively proposedthat hisingeriteis built principally upon a mica model.This would require that there are Fe proxying for Si and hydronium for interlayer alkali ions. The rather ex- tensive incorporation of iron in the tetrahedral layer must create a strained lattice which could be the explanation for the extremely poor crystallinity of hisingerite. It is not claimedhere that this hypothesisis necessarilythe correctone, but the aim has been merely to accentuate the fact that hisingerite definitely warrants further study.

AcxivowrnncMENTS Thanks are due to Prof. E. Norin (Mineralogical-GeologicalInst., Uni- versity of Uppsala) for laboratory facilities placed aL our disposal,and to Prof. F. E. Wickman (Dept. of Mineralogy, Swedish Museum of Natural History) for furnishingmineral specimens.

Rnnrnorgces

Bowrr, S. H. U. (1955) Thucholite and hisingerite-pitchblende complexesfrom Nicholson Mine, Saskatcheu'an,Canada. Bril'|. GeoI.Surt. Gr. Britain, 10, 45-57. Cr.nvr, P. T. and A. E. NomnNsrtcir.o (1866)Om sammansd.ttningenaf jernhaltiga kolloid- silikater. K gl. V etensk.A k ad..F brlt,. 23, 769-783. Drnrnrcrr, R. V. (1961) Hisingerite in "dark" larvikite Norsk Geol. Ti'dskriJt,41, 95-108. FnoNrrr-, C. (1955)Tr-o chlorites:gonyerte and melanolite. Am. Mineral.40, 1090-i094. Gnuwnn, J. W. (1935) The structural relationship of nontronites and montmorillonite. Am. Mineral. 20, 475 483. Hrwetr, D. F. and W. T. Scu,c.Lr-Bn(1925) Hisingerite from Blaine Co.,Idaho, Am. Iour. Sci.,2lO,29-38. Nlcrr.scultrnr, G. (1938) On the atomic arrangement and variability of the members of the montmorillonite group. Mineral. XIag.25, 140-155. Nrr

Monuscri.pt receiaed,,June 18, 1962.