THE AMERICAN MINERALOGIST, VOL.47, MAY_JUNE, 1962

BENSTONITE, CazBa6(COr)tr,A NEW FROM THE BARITE DEPOSIT IN HOT SPRING COUNTY, ARKANSAS Fnrrrnrcs Lrrruenw,l Mineralogische Anstohen der Uniaersildl, Gdttingen,Germany.

Ansrnlcr Benstonite, (Ca,Mg,Mn)r(Ba,Sr)6(CO)s, was found at the barite mine in Hot Spring County, Arkansas, where it occurs in veins within the barite body associated with milky , barite and . It forms white cleavable masses which show faces up to 1 cm across.H:3-4; G 3.596 (meas.), 3.648 (calc.); good rhombohedral cleavagewith about the same angles as calcite. Uniaxial negative with o:1.690s and e:1.527. Single crystal r-ray studies revealed a rhombohedral lattice with the following characteristics: possible space groups RB or R3, the most probable being R3 because of -the symmetry of the CO3-groupand the available equipoints; hexagonal, o:18.28*0.01 A, c:8.67IO.02; rhombohedral, a,t:10.944, a:113.3o; volume 2509 Aa ftex.;, cell contents:3 [CazBao (COa)ul in the hexagonal unit. The index of the cleavageface is [3142] or [4132]. The hexagonal indices of the strongest r-ray reflections satisfy the relation Trzaftzlhk :13n, and they may be transformed to the indices of calcite reflections. The resulting "calcite-ceil" of benstonite has o!:a/t4i3:5.0? A, C':2c:17.34 A; volume 386.0 A3; calcite: a:4.990 A; c:17.06 A; volume 367.80 A3. Benstonite therefore has a structure which may be described as a superstructure derived from the calcite configuration. Ap- proximate parameters for the cation sites in benstonite are derived from the calcite ar- rangement, and four difierent models are proposed for the arrangement of the cations. The chemical analysis was carried out on selected material from which a small ad- mixture of calcite could not be separated. The formula Cae.lsMgo.zsrMno.ognBarzzsSro.zzz (COr)rs calculated after deduction of 2.23 weight per cent calcite may be generalized to CazBao(COa)raor possibly MgCaeBas(COr)n. Benstonite is named in honor of O. T. Benston. Malvern. Arkansas. who drew attention to the mineral.

IurnonucrroN In Hot Spring County, Arkansas, abofi 2l miles ENE of the igneous complex of Magnet Cove, a large barite deposit is mined in an open pit operation.The bedsof the depositlie betweenthe Arkansasnovacu- lite (Devonian-Mississippian) and the Stanley shale (Pennsylvanian) at the eastern end of a syncline that follows the strike of the Ouachita mountains. The deposit was describedby Parks and Branner (1933), and the geology of the region is summarized by Fryklund and Holbrook (1eso). The barite rock is fine-grained and has a gray color owing to varying amounts of interspersedshaly matter. In certain placesit contains irregu- Iar white veins, from an inch to a few inches in thickness, in which the new barium calcium carbonate, benstonite, is found. It is associated with milky qvartz, barite and calcite.

1 Present address: Grabbeplatz 12, Dortmund, Germany.

J6J FRIEDRICH LIPPMANN

Benstoniteis named in honor of O. J. Benston,born 1901,ore dressing metallurgist, Baroid Division, National Lead Company, Malvern, Arkansas. Mr. Benston had guessedfrom qualitative chemical tests and from the specificgravity of about 3.60 that the mineral might be either or . During a visit to the mine on New Year's Eve, 1954, he kindly furnished the specimenson which the presentstudy is based.

Pnvsrc.q.rPnoppnrrBs Benstonite occursas cleavablemasses with cleavagefaces up to one centimeteracross. Cleavage fragments are more or lessperfect rhombo- hedra of the same shape as calcite and the other rhombohedral car- bonates. The cleavageis not quite as perfect as that of calcite. The cleavagefaces are slightly curved, and, especiallyon the larger pieces, they are composite, with mosaic structure as in some and .On the reflectinggoniometer extremely diffusesignals were ob- served.Values oI 44.7"and 44.3owere obtained for the polar distances (p angles) of the faces on two different cleavagerhombohedra (averages of a large number of readings).For these measuredp anglesthe calcu- lated cleavageangles (tr angles) are 75.I" and 74.4", respectively.(The corresponding literature values of p and tr for calcite are: 44.6" and 74.9".) The hardnessis between 3 and 4. The specificgravity as determined by the pycnometer method is 3.596. The color is snow white to ivory. When irradiated with long or short wave ultraviolet radiation benstonite fluorescesred, which is almost identical with that of a glowing cigaret. The same fluorescenceoccurs when the mineral is exposedto r-rays. Phosphorescenceis perceptibleat 20 kilovolts. At 50 kilovolts the afterglowlasts a coupleof minutes.

Oprrcer PnopBnrros In thin sectionbenstonite shows numerous bubbles of about a micron to a few microns diameter and other more irregularly shapedcavities of similar dimensions.These are responsiblefor the white overall appear- anceof the mineral which otherwisewould be translucent.Extinction of individual grainsis not uniform. Domains varying in sizefrom 0.1mm to over 1mm are displaceda few degrees.In somecases the boundariesare sharp, but usually the domains merge almost imperceptibly with each other. This phenomenon,which occurscommonly in siderite and dolo- mite correiateswith the composite nature of the cleavagefaces. It is similar to the undulatory extinction of deformedquartz grains in meta- morohic rocks. BENSTONITE 587

Benstonite yields a uniaxial negative interference figure. Even in thick sectionsthere is no indication that the mineral is biaxial with a small 2V. This is important for the distinction from alstoniteand barytocalcite,for which axial anglesof 6o and 150,respectively, are reported. The refractive indices were determined for sodium light by the im- mersion method:

X-nev CnysrarrocRAPHY A well terminated cleavagerhombohedron about 0.2 mm in size was mounted parallel to its trigonal axis. After orientation by the oscillation method, a rotation photograph and equi-inclinationWeissenberg photo- graphs ol zero, first, and secondlevels were taken with Ni-filtered Cu-K" radiation. All reflections observed satisfy the condition of rhombohedral centering-h+k+l:3nlor hk.l. The lattice levelsymmetries G of the zero Iayer and Cg of the first and secondlayers yield the diffraction sym- bol 3 R- which includesthe spacegroups R3 and ft3. The lattice con- stants weredetermined to be o: 18.4A, from the zero-layerWeissenberg photograph,and c:8.75 A, from the layer-linespacings on the rotation photograph. In indexing the powder data obtained with NaCl as an internal standardo:18.28*0.01 A and c:8.67 +0.02 A gavethe closestfit of calculatedand measuredd-spacings (Table 1). For thesehexagonal lattice constantsthe axialratio is: a:c:l:0.4743. The calculateddimensions of the rhombohedralcell are: a,n:10.94 A and a:113.3o. The cleavage rhombohedronmust have the index {3 | 4 2l or {4 1 3 2}1. For these indicesthe calculatedp angleoI 44.6"(I : 75.0') is within the rangeof the two measuredones. The indicesof the cleavageface are substantiatedby the fact that the corresponding powder diffraction line, which is the strongestof the whole pattern, is of greatly variable intensity with more or less preferred orientation. The cleavage and the r-ray diffraction properties suggestedthat the structure of benstonitemight be related to calcite.It was found that the strongest reflectionsin the powder pattern have hexagonalindices which satisfythe condition h2lk2lhh:13n. Thesemay be subdividedinto two groupswith 4hlk:13n and 3h-k:13n. If the first group is converted by the transformatior'4/fi t/13.0/-I/tt s/1t.0/0 0.2 the set of indices oI r-ray reflections of calcite is obtained. The same result is arrived at when the indices of the second group undergo the transformation

I Right- or left-handed, respectively. TAsr.n 1. X-R,w Powonn Dlta lon CAr.crtn (SweNsoN aNo Fuval, 1953) eNn ror BrxsroNrtn,r CazBao(COa)u,Hrxncoxl.t-, o: 18.2810.01 A, c: 8.67+0.02 A, Pnoeanlr Sp.a.cnGnoup R3 (PnnsrN:r Srunv) Strong reflections are underlined.

Relative d.oo,L fr Rel. int.a tlo,o, y ^fu do r. hh I3 calcite intensity2 measured synthetic calcite calculated cell

3 9.21 9.140 11.0 /.oJ 7 605 10 1 z 5.83 5.843 02.1 3 5.28 5.277 03.0 / 4.93 4.925 21.1 J 4.57 4.570 22.O t6 4.19 4.181 or.2 38 3.92 3.917 13.1 0t.2 3.86 I 3.80 3 .802 20.2 9 3.60 3 .601 40.r 3.511 12.2 4 3.45 3 .455 14.0 10 353 3 .350 32.1 95 3 085 3 085 3t.2 10.4 3 035 3.O47 33.0 T 2.98 2.974 05.1 2.923 0+.2 I 2.89 2.8q 00.3 00.6 2.845 2.828 24.r 2.784 232 2.755 2.756 11.3 2.702 51,1 I 2.637 2.639 06.0 2 2.557 2.557 50.2 28 2.536 2 535 25.O 11.0 2.495 2.535 03.3 1 2.495 2.493 43.1 1 2.46 2.462 Aa a 9 2.M5 2.M3 22.3 € N 2 .380 2.378 152 N 2.327 2.326 16.1 1 2.281 2.285 440

I Benstonite with NaCl as internal standard, Norelco X-ray spectrometer, ]o per minute, 1o slit, receiving slit 0.2 mm; CuKa radiation, Ni-filter. 2 Relative intensities are peak heights on spectrometer trace. They are distorted by preferred orientation. 3 Only the indices with positive numbers are listed for every difiraction line in the hk.l, col:trmr'. The other possible (enantiomorphic) index is given under hi'l only when necessary for the transformation to the "calcite-cel1" of a right-handed crystal. a The indices, intensities, and d-spacings of calcite reflections hk'|, with I odd are written vertical. They have no equivalent in benstonite because its c-spacing is about hali that of calcite. Tasr,e' 1 (continuerl)

htkt .It4 Relative dnntA du," f\ Rel. Int.a dv6' v A t hk.l3 hi.l calcite synthetic calcite intensity2 measured calculated cell

.) 2.233 2.231 34.2 10 2 217 2.2r7 r+.3 4 2.t89 2. 188 70.1 2.188 J5 I ? 2.r5 2.t48 10.4 23 2.t27 2.128 62.r 2.09s 4 2.t13 2.r09 6t.2 A 2.W7 2.@7 17.0 2.097 JJ.J 2.O9r 02.4 2.O42 2 .038 21.4 2.W6 2.005 53.2 J t.994 1.995 36.0 ? 1.973 |.974 54.1 10 | 957 1.95S 26.2 28.2 02.4 1.927 1.949 06.3 11 t.9M 13.4 144 01.8 r.913 1.929 08.1 1 905 1.906 25.3 11.6 1.875 1.901 0+.4 I .886 1.888 27.r 1.863 1.861 32.4 1.834 1.836 45.2 I .828 55.0 1.8t2 81.1 1.800 80.2 1.792 44.3 1 789 05.4 1.777 46.1 1.766 72.2 r.759 09.0 1.755 24.4 t.746 r.745 73r 1.728 r.727 28.O t.724 51.4 LaA 01.5 1.698 1.697 l/.J 1.692 r.694 20.5 r.670 1.675 64.2 N 1.664 t.666 12.5 1.645 1.648 37.2 1.642 47.O 1.642 36.3 1.630 65.1 ,7 1 630 1S.1 8 1.604 2 1.613 31 5 2 l.)d/ 6 1.550 56.2 q 1.525 590 FRIEDRICH LIPPMANN

Tesla 2. CourenrsoN BarwenN DrunNstotts on ,'Cerrrrn-CEtr,,, oF BntqsroNrrr lNl Srnucrunr Cnr,r, or C,r.r,crrn

"Calcite-cell" of benstonite Structure cell of calcite

a-:5.O7 hexagonal" a' A a:4.990 i| \/13 C':2c:17.34 A c:r7.06 Ll 2V v,:_:386.0 A v:367.80 A3 IJ a':C':l:3,420 aic:1i3.4189 indexof cleavageiace: {tOf+} { 1or4} rhombohedrala',n:6.47 s A a,x:6375 A a':46.lo a:46.08o

1 Goldsmith and Graf, 1958.

-r/rc'O/t/rs s/ts +/rc-0/0 0.2. The inverse matrices are 3-I.0/l - 4. 0/00. 1/ z and 4 l. 0/ | 3. 0/ 0Q.1/2, respectively. If the indicesof the x-ray refl.ectionsof calcite with l: 2n are converted by one of the reverse transformationsthe indices of the strongestbenstonite spacings result. This meansthat the structure of benstoniteis basedon a pseudo-cellthe hexagonaldimensions a' ar'd c'of which are related to the structure cell dimensionsof benstonite, a and c, by at : a/ \/ 13 and c' : c. Two adja- cent pseudo-cellsin the c-axisdirection have total dimensionssimilar to the structure-cellof calcite. The numerical data of the "calcite-cell" of benstoniteare comparedwith the structure cell data of calcitein Table 2. The axialratios are almostidentical, but the "calcite-cell"of benstoniteis larger,as can be expectedfrom the larger sizeof the barium ions with re- spect to the calcium ions. From this similarity as well as from the simi- larity of the r-ray powder pattern of calcite and the strongest reflections of benstoniteit may be concludedthat the structure of benstoniteis of the same general schemeas the structure of calcite with the cations ordered to form a superstructure(and perhapsthe CO3-groupsrolated with respectto the calciteconfiguration). The pseudo-celi(and also the "calcite-cell") is rotated by 13.9'versus the real cell. The senseis counterclockwise,if the indices of benstonite with 4h*k:13n are transformed,or clockwise,if the indices with 3h-k :13n are involved. In the spacegroup R 3 both orientationsare identical if viewed from oppositedirections along the c-axis.In the spacegroup R 3 the two arrangementsare enantiomorphic.Since the cleavageface is the most prominent morphologicalfeature it wiil customarily be indexed BENSTONITE

Frc. 1. Calcite-type cation layer (calcite-cells: dashed lines, lattice constant o') with right-handed benstonite-type superstructure cell (full lines, lattice constant o'.t/13) containing 13 cations per unit in the layer. Equipoints are numbered according to the symmetry of R3 as in Table 3, column 1.

l3l42l in right-handed crystalsor |j132} in left-handedones. Whether this assignmentis really true could only be decidedby a determinationof the absoluteconfiguration. From the customaryindexing and the relation to the calcitestructure it follows that the strong reflectionsare assigned indices with 4h*k:13n in the right-handed enantiomorph and with 3h- k:13n in the left-handedone. The "calcite-cell" of benstonite (or two pseudo-cells)will certainly contain 6(RCOa) as in the hexagonalstructure cell of calcite. Since the transformation of the structure cell to the "calcite-cell" is accompanied by a changein volume by the factor 2f 13, there will be 39 carbonate units, i.e.39(RCOr, in the hexagonalstructure-cell of benstonite.In the spacegroups R 3 and R 3 this cell content implies for a structure similar to calcite that three cations and three carbonate groups have to be located at unique threefold positions.In the spacegroup R 3 thesehave the point symmetry 3. Since this is incompatible with the shapeof the carbonate group, R3 is ruled out. In the spacegroup R3 the three unique cationsand anionsare on positions(a) 0,0, z with z:0and z-l/2,re- spectively.The remaining36 cationsand 36 anionsmust be distributed to 4 general ninefold positions each, with z-0 and z-l/2 respectively. The available positions in one cation-layer as well as the pseudo-cellrelation are sketchedin Fig. 1. For a carbonatewith two different cationsthe fol- lowingf ormulaecan be expected:XrYu(COa)rs, XaYro(COs)u, X4Ye(COB)rl 592 FRIEDRICH LIPPMANN

T,tsr,B 3. Possrsrr Penewrnns or Cerror.rsrN a Rrcnr-Haronn Supnnsrnucrunn or BnlrsroNrrn Tvpn

x'y' z' yzr calcite- Model 1 Model 2 Model 3 Model 4 cell Benstonite-ceil

3 Car2 0,0,0 0, 0,0 Ca? 3Ca? 3 Caf 3Ca12 9Rrr 1,0,0 4/13,1/rs,o 9 Bar 9 Carr 9Carr 9 Carr 9Rrrr 1,1,0 3/t3,4/13,O 9 Barr 9 Bar 9 Carrr 9 Bar 9Rw 2,o,0 8/13,2/t3,0 9 Carr 9 Barr 9 Bar 9 Carrr 9Rv 1,2,0 2/13,7/13,O 9 Carrr 9 Carrr 9 Barr 9 Barr

1 All positionshave been expanded by the symmetryelements of R3. 2Or Mg for MgCaeBao(COa)ra. and X?Y6(COa)13.Of these, CaTBao(CO3)sagrees very closely with the chemicalanalysis (see below). Four difierent arrangements (models) are possible for one enantio- morphic form. They are listed in Table 3 for a right-handed crystal. The carbonategroups willbe on analogouspositionswith z-l / 2.Noprediction can be made regarding their orientations, which must be different from those in calcite on account of the difierence in the c-spacings.A detailed structure analysiswill have to decidebetween the possiblearrangements of the cationsand determinethe configurationof the carbonategroups.

Cururcar- CouposrrroN Selectedcleavage fragments were tested for uniform fluorescenceto in- sure the absenceof barite and quartz. About 10 grams of powder were ob- tained for analysis by wet chemical methods. When this was examined by *-ray difiraction an additional line at 3.03 A with an intensity like the medium-weaklines of benstoniteappeared. Since this line is absentin the other powderdiagrams prepared of benstoniteit had to be assignedto the strongest reflection of calcite. This mineral accompaniesbenstonite and may be intimately intergrown as was seenduring the optical study. In a float test with bromoform no weighable amount of calcite could be separated.Some excess of calcium carbonateis, however,to be expected in the chemical analysis. Samplesof one half to one gram were used for every determination or separation. After decomposition with hydrochloric acid, carbon dioxide was absorbed by soda Iime and weighed. Barium was separatedand weighed as barium chromate, precipitated from the acidic, acetate bufiered solution. Calcium and strontium were determined together as BIJNSTONITE 593

Tesr-E 4. Cnnlnclr, Aralvsrs ol BENSToNTTEMerpnr,ll Analyst: F. Lippmann

wt. 104- 96 wt.7ot

BaO 43.05 2807 BaCO: 55.40 SrO 4.O2 388 SrCO: .) . /.t CaO 19.52 3481 CaCOr 34.84 Mso r.69 419 N[gCO; 3.53 MnO 0.35 49 MnCOa 0.57

Sum of cations 100.07

COz 31.35 7t23 Total 99.98

1 Calculated from cation oxides. oxalates.In one instance both were precipitated as carbonates.These were dissolvedin nitric acid, and from the dried nitrates calcium nitrate was extractedwith a mixture of ether and alcohol.A value oI 4.021sStO was obtainedin this way. This value agreessatisfactorily with the 38070 SrO determinedby r-ray fluorescenceon the whole mineral with bromine (KBr) as an internal standard. Magnesiumand manganesewere each determined on separatesamples. In the magnesium determination, manganesewas first removed with ammonia and bromine. Then barium, strontium and part of the calcium were precipitated as sulfates.The remaining calcium was removed as oxalate. From the final filtrate magnesium was precipitated as mag- nesiumammonium phosphate.Before the manganesecould be oxidizedto permanganateby persulfatefor the colorimetricdetermination, barium, strontium and most of the calciumhad to be removedas sulfatesfrom the nitric acid solution. There was no weighableinsoluble residue.No lead and no iron were found with the usual wet chemicaltests. The resultsof the chemicalanalysis are given in Table 4. The values are averagesof closelyagreeing duplicates, except for SrO. Here the wet chemicalvalue is tabulated becauseit was deemedmore consistentwith the other wet chemicaldata. There is only a slight excessof the cations over carbondioxide as is evidentfrom the quotientsof weight percentages over molecular weights. Consequently,carbonate weight percentages calculatedfrom the cation oxidesand adjustedto 100/s wereused for the computationof the chemicalformula presentedin Table 5. 594 FRIEDRICH LIPPMANN

Tanr,r 5. Cer,cur,Arrol,r oF TrrE FoRMUT,Alon BnNstoNtrn

wt. Tatu wt.7o lu wt.'/"t Mol. lo Mol. wt Mol. wt. 532

BaCOr 55.37 28osi 5 273 :6:0.879 40.56 lstoz:o-stz SrCOs 5.72 387) o.727 :6:O.l2l 5.59 6,000 MgCOa 3.53 419 0.788 :7:0 .ll3 6.06 MnCOs 0.57 50 3947: 7:564 0.094 :7:0.013 0.72 combined CaCOs 3724:7. 532 6.118 :7:0.874 47.O7 223 7,000 free I I 34.81 34.81 3478 )

7139-233:6919

1 From last column of Tabie 4 adjusted to 100/y 2 The figures in the box were entered at a later stage of calculation. Formula:

CaousMgo zrsMnooslBas:zsSro z:z(CO:)ra or

(Cao szrMgou:Mno or)z(Bao srgSrotgr)r(COr)rr Average molecular weight of RCOr:141.35; 39(RCO3) per hexagonal unit cell V:2509 A3; G(calc.):3.648. 38.44 (RCO3) per unit cell (5.91 (RCOr) per "calcite-cell") for experimental density 3 596.

In all attempts to calculatea formula for benstonitefrom the chemical analysis,the large cations Ba and Sr and the smaller ones Ca, Mg, and Mn werecombined in two groups.The discoverythat the analyzedpow- der containeda few per cent of calciteas a discrete phasehad to be taken into account. No quantitative r-ray determination of the amount of cal- cite was attempted becauseof the complicationsto be expectedfrom pre- ferred orientation. Before the relation to the calcitestructure was recog- nized,a formula of the type XY(COt, was attempted in severalvaria- tions. First, 5/6 oI free calcite was subtracted for an approximate ratio 1:1 of the two groupsof cations.This procedure,however, either yielded strongly deviating calculated densities or else improbably high (about l5/) or negativeamounts of free calcitehad to be presumedfor 36 or for BENSTONITE

42 carbonate molecules, respectively, per hexagonal unit cell. The cell content of 39 derived from the pseudo-cellrelation necessitatesunequal amounts of small and large cations.Their ratio is 7.42:6, without deduc- tion of free calcite. For the ratio 7:6 demandedby the formula compatible with the unit ceII and its symmetry, 2.23 weight/o of CaCO3has to be subtracted from the analysis. The calculated amount of free calcite agrees with the relative intensity of 3.03 A reflection. The resulting formula for benstonite is given under Table 5. The calculated density of 3.648 aggreeswith the experimental one of 3.596, if the presenceof the microscopical cavities and the possible contamination with calcite are taken into account. The sum (Mgf Mn) is high enough to permit con- jecture about the concentration of these cations at the unique threefold positions of the unit cell. A structure analysis will reveal whether it is justified to write the formula MgCaoBao(COa)ra or in detail

(Mgo.zeaMno. osrCas n6) Ca6(Bas ersSro.Dr)o(COa)ra

Mrscprr,aNrous OBSERVATToNS When benstonite is crushed the odor of hydrogen sulfide is perceived. fhis compound must form part of the filling of the microscopic cavities. The presenceof hydrogen sulfide in the voids of benstonite means that the mineral formed in a reducing environment, which was probably caused by the overlying bituminous Stanley shale. It is likely that the barium of the benstonite came from the bedded barite, a small part of which was dissolved in the reducing environment, whereas calcium and magnesium were introduced by percolating waters. The red fluorescenceof benstonite is probably due to its manganese content. This conclusion is suggested by the following experiments: Benstonite is dissolved in hydrochloric acid. If the cations are pre- cipitated from the hot solution as a carbonate by addition of ammonia and ammonium carbonate, the carbonate obtained when exposedto X- rays fluoresceswith the same color as benstonite. The intensity however is weaker. If the manganeseis first removed by precipitation with am- monia and bromine and filtration, the carbonate precipitated from the filtrate doesnot fluoresce.No attempt was made to identify the nature of the precipitate. According to the experimentsof Terada (1953) a rhombo- hedral mixed crystal of barium calcium carbonate should have formed. Therefore the above described experiments are not a direct proof that manganesecauses red fluorescencein benstonite. They only show that manganese can make complex barium-calcium carbonates fluorescent. 596 FRIEDRICH LIPPMANN

Schulman el al. (1947)t have shown that red fluorescencein calcite re- quiresthe presenceof manganeseplus a secondactivator elementsuch as lead,which might be presentin benstonite,as a trace element.Otherwise barium could be the secondactivator. (Seenote, p. 598.)

DrscussroN Two compoundsof the formula BaCa(COa)zare known as : barytocaicite,monoclinic, and alstonite,orthorhombic. The structure of barytocalcitewas determined by Alm (1958).It is orderedwith respectto Ba and Ca, these cations occupying discretepositions in the structure. Alstonite is very similar to aragoniteand witherite both in morphology and lattice constants.It is not yet known whether alstonite is ordered with respectto Ba and Ca, but the weak additional layer lines observed by Gossnerand l{ussgnug (1930)on their rotation photographsmay be taken as evidencein favor of a superstructure. Rhombohedral mixed crystal carbonatesof barium and calcium were synthesizedby Terada (1952, 1953)with up to 55 mol 7o BaCOy They yieldedr-ray powder patternsvery much like thoseof calcite.Reflections (hk.l) with I odd were missing,if there were 20 or more mol/6 BaCOs. This meansthat the lattice constantin the c-directionis about half that of calcite, just as in benstonite.Apparently the presenceof enough Ba changesthe articulation of the CO3-groupsin calcite-type carbonates. For the composition of 40 mol 7o BaCOt and 60 mol /6 CaCO3, which comparesmost closelyto benstonile,a:5.06 A and 2c:17.69 A are calculated from the rhombohedral data of Terada. The o-value is slightly smaller and the 2c-vaIueis higher than for the "calcite-cell" of benstonite(at:5.o7 L, 2r:1734 h). The cell volume V:390 A3 is identical with that of the "calcite-cell" of benstoniteV':386 At *ithin the limit of error. If comparedwith the barium-calciumcarbonates of Terada,benstonite has a superstructure.One cause for the orderingof the cationsin benston- ite may be its slow rate of formation. If, however, it turns out that the Mg-ions in benstoniteoccupy the specialequipoints 0, 0, 0 of the hexag- onal cell, so that the formula is MgCa6Bau(COr)tr,it is more likely that the presenceof magnesiumis responsiblefor the formation of the super- structure. The intervention of magnesiumcould be the reasonwhy ben- stonite occurs at all, if one considersthe existenceof the other barium calciumcarbonates which do not seemto contain appreciableamounts of magnesiumaccording to the analysesreported in the literature. Whereasdolomite, CaMg(COe)2, and perhapsnorsethite BaMg(COr)z

1 Dr. Michael Fleischer kindlv directed the author's attention to this article. BENSTONITE 597

(Mrose et al. 196l), are superstructures derived from the calcite con- figuration by the distribution of the differentcations into alternatingcat- ion layers,the superstructureof benstoniteis formed by the ordering of the cationswithin the layers.This is accompaniedby a differentarticula- tion of the CO3-groupsas is evident from the c-valuewhich is about half that of calcite.Another exampleof a rhombohedralcarbonate structure with about half the c-value(7 .82 A) of calciteand with orderingof the cat- ions within the cation layersis 1\t[g3Ca(COr)a(Graf and Bradley, re62). It is interestingto note that complex calcite-typecarbonates, such as the mix-crystalsof Terada, norsethite,and benstonite,exist with a cat- ion as large as barium. This doesnot agreewith the current explanation that the crystallizationof CaCOsin the form of aragoniteis favored or even stabilized by the presenceof small amounts of large cations, e.g. strontium. A different mechanismmust be active in most caseswhere CaCOr crystallizesas aragonitein nature (seee.g. Lippmann, 1960 and Heinrich and Levinson, 1961,p. Aa\.

Acr

RBnotnNcos

Ar,u, KNur F. (1958), The of barytocalcite BaCa(COa):. Arki,o Mineral,. Geol.,2,399-410. Fnvrr,uwo, Vnnur C., Jn. .rxo Dnlw F. Hor,snooK (1950), Titanium ore deposits of Hot Springs County, Arkansas. Ark. Res. Dewl,. Comm. Dilis. Geol., BulL 16. Golosurtn, J. R. aro D. L. Gnel (1958), Relation between lattice constants and composi- tion of the Ca-Mg carbonates. Am. Mineral.,43, 8rt-101. Gosswnn, B. nxo F. Musscrwc (1930), Uber Alstonit und Milarit. Ein Beitrag zur Kennt- nis komplex gebauter Kristalle. Cent. Mi.neral,., A 1930, 220-238. 598 FRIEDRICH LIPPMANN

Gner, D. L. eNo W. F. Bnlor,rv (1962),The crystal structure of huntite MgaCa(COa)r. A cta Cryst., 15, 238-242. Hnrwnrcn, E. Wu. aNn A. A. LnvrNsoN (1961), Carbonatic niobium-rare earth deposits, Ravalii County, Montana. Am. Mineratr.,46, 142+-1447. Lrru.LttN, F. (1960), Versuche zur Aufkliirung der Bildungsbedingungen von Calcit und Aragonit. Fort. Mi,neral., 38, 15G161. -- (1961A'), Uber ein Barium-Calciumkarbonat aus Hot Springs County, Arkansas (abs.). F ort. IIinerd., 39, 81. - (19618), Benstonit, CazBa6(COa)3,ein neues Mineral. Di,e Natu,rwissenschaJten, 48, 550-551. Mnosr, M. 8., E. C. T. Cueo, J. J. Fermv AND C. MrLroN (1961), Norsethite, BaMg (COr)r, a new mineral from the Green River Formation, Wyoming. Am. Mineral.,46, 42H29. PAr,Acrre, C., H. Brnu.tll, exo C. Fnorropr, (1951), Dana's System of Mineralogy Vol. II(7th): New York, John Wiley & Sons, New York. P.tnns, Bnvln AND G. C. Bnannrn (1933), A barite deposit in Hot Springs County, Arkansas. Arh. Geol. Swrey, InJ. Circ. l. Scuutuer, J. H., L.W. Evars, R. J. GrNrnnn, ANDK. J. Munere (1947), The sensitized luminescence of manganese-activated calcite. Jour. Applied Physics,18r 732-739. SweNsoN, H. E. euo R. K. Fuver (1953), Standard. r-ray powder patterns. U. S. Nat. Bw. Stand..,Ci,rc.539, 5I-54. Tene.o,t, Jrrsuo (1952), Rhombohedral crystals of Ba-Ca and Sr-Ca double carbonates. fow. Phys. Soc. Jopon,7,432434. -- (1953), Crystal structure of the Ba, Sr and Ca triple carbonate. fow. Phys. Soc. fapan,8, 158-164.

Manuscript received.,August 9, 1961.

Note added in proof : Prof. K. H. Wedepohl, Gdttingen, kindly ran a spectrographic analysis for lead on the same material which was used for the chemical analysis of ben- stonite. The line of Pb at 2$3 A, which should still be visible at a concentration of as low as 5 ppm of Pb could not be detected. The benstonite therefore contains less than 5 ppm of lead. This renders more likely the explanation that barium is the second activator (besides manganese), as is perhaps also the case in bariumcyanoplatinate, where ob- viously barium and platinum act together.