THn AMERICAN M IxERALOGIST

JOURNAL OF THE MINERALOGICAL SOCIETY OF AMERICA

Vo].51 JANUARY-FEBRUARY, 1966 Nos. 1 and 2

HUEMULITE,I NaaMgVroOzs.24HzO,A NEW HYDROUS SODIUM AND MAGNESIUM VANADATE FROM HUEMUL MINE, MENDOZA PROVINCE, ARGENTINA

C. E. Gonprrlo, E. LrNanrs, R. O. ToueBs, Comis'i6nNocionol d.eEnergio Atbmica, Argentina nNo H. Wrxcnnlr,, Department of Geology,Yale Uniaersity.

Aesrnlcr

Huemulite, a ner,vmineral u'ith the formula NaaMgVlsOzs. 24HzO, rvas found in several uraniferous ore bodies of the "sandstone-type deposits", in the southwestern part of Mendoza province, Argentina. Huemulite appears in botryoidal masses, thin films, or as interstitial filling in the sandstone that is the host rock. It is soft, yellowish orange to orange in color, with a dull luster and specific gravity oI 2.39 g/cm3. It is not fluorescent. The is easily soluble in cold n'ater, giving an orange solution of pH 5.5 to 6.5. From this solution the mineral recrystallizes giving tabular crystals elongated parallel to (010), with a perfect cleavageparailel to (001)and lessperfect parallel to (010).Theoptical properties determined in the recrystallized material show that the mineral is biaxiai nega- tive,2Y:25"-30o, pleochroic with X light yellow, Y golden yellow, and Z yellowish orange. The indices of refraction measured r,r,'ithsodium light are o:t.679, A:7.734, and t:1.7 42. The dispersionis strong with r ) v. Chemical analyses were performed on both natural and recrystallized material. A synthetic compound q'as prepared which gives a fi-ray powder, and single crystal data similarto thoseof huemulite Huemulite is triclinic, space group P1 or Pl. Unit cell data obtained on the recrystal- lized material are'. a:l1.770 A+.019, b:11.838+.008, c:9.018+.009 A, q:107" 13/+05/, A:112" 10/+06', z:101o 30'*05'; o;bic:0.9913i1:0.7618 Y:1o4o.67 A"+2.70. Cell contents NaaMgVlnO2s.24HzOgives a calculated density of 2.4Og/cms. The strongest lines of the r-ray powder diffraction pattern are: 7.6 (100), 10.6 (90), 9.1 (60), 10.2(ss),8.2 (3s),3.0s (30), s.26 (2s). The name huemulite comes from the Huemul mine. the most important de- posit of the area where the mineral was discovered.

INrnorucrroN This paperpresents the stud-vof a new hydroussodium and magnesium vanadate that was found in some of the uraniferous deposits of the Malargiie area, Mendoza province, Argentina. The discoverv of this

1 Approved by the Commission on Nerv and Mineral Names, IMA, May 19, 1965. 2 C. E. GORDILLO,E. LINARES, R. O. ]:OUBRSAND H. WINCIIELL speciesadds a new natural member to the salts of the isopolycomplexion (VroOza)6-,described by Evans et al. (1955;1964). V. Angelelli of the Geologl. Department of the Argentina Atomic Energy Commission(C.N.E.A.) first collectedhuemulite in 1959,in the underground levei - 18 of the ore-body known as Agua Botada. The sample was given to Linares for identification as a possible secondary mineral of uranium or vanadium. Optical and r-ray determinations showed that it did not correspondwith any of the uranium or vanadium minerals known, thus starting an investigation of its physical and chemi- cal properties. l\llore material was collected during other visits to the area and it was also identified in the ore bodiesHuemul, and Agua Botada Sur (Brodt- korb, 1963),which are adjacentto the original site of discovery. Preliminary chemical tests showed that the mineral is a hydrous sodium vanadate with uranium present only as impurity. The tests also showedthat it is easilysoluble in water and that it can be readily recrys- tallized from the solution by slow evaporationof the water. The optical and other physical properties were investigated mainly in recrystallized material whoseidentity with natural material was checkedby r-ray data. Chemical analyses were performed in both natural and recrystallized mineral; similar or identical synthetic compound was also prepared and studied. A preliminary report with all the data obtained was presentedto the II Argentine Geological Congressin 1963 (Linares et al., 1963).The follow- ing year the unit cell was measured by Linares and Winchell at Yale University and the data thus obtained refined using the facilities of the Yale Computer Center. During the same year the chemical and physical propertieswere more elaboratelychecked by Gordillo and Toubesin the laboratoriesof the C. N. E. A. The new mineral is named huemulite from the most important uranium mine of the Malargiie area.

AcrNowr-noGEMENTS Thanks are due by Linares, Gordillo and Toubes to our colleaguesof the ServicioLaboratorios (C.N.E.A.), and to V. Angelelli who encour- aged us in the study of this mineral, and also to Angelelli and P. N. Stipanicic for many helpful suggestions. Linares is also indebted to the Division of Raw Materials of the U. S. Atomic Energy Commission and to Dr. M. L. Jensenof Yale University, since the last part of this investigation was carried out during his post- doctoral research year at YaIe University, under a grant AT(30-I) 226I of the U.S.A.E.C. HUD.MULITE,A NEW VANADATE 3

OccunnBNcB Huemulite was found the first time in the level - 18 of the ore-body Agua Botada. It was later found elsewherein the samemine and also in the Huemul and Agua Botada Sur mines. The uraniferous district known as "Malargiie atear" where thesemines are located is in the southwesternpart of Mendoza province, about 450 km southwest of the city of Mendoza, Argentina (Fig. 1). The area is generally one of continental sediments of Middle Cretaceousage, called locally Diamantiano, with a total thickness of 900 m. Overlying it are marine aird continental sedimentsof Upper Cretaceousand Tertiary age. Middle Tertiary (Oligocene)basic volcanic bodies, predominantly horn- blendiferousandesites, cut through the sediments in the entire area. There are someoutcrops of small flat bodies of ignimbrite. The complex is coveredby Pleistoceneand Recent sediments. The uraniferous ore-bodiesare located in conglomerates,arkosic con- glomerates, and arkosic sandstone beds of the lower portion of the Diamantiano. The sandstones are generally enclosed between beds of siltstone and claystone; they are always associated with an asphaltic material originating, it is thought, from the oil in the adjacent small oil field of Pampa Amarilla. The average dip of these beds is 35o W., with a mean thicknessof 1.2 m and a grade of 0.25/6 TJzOa,2.07oof Cu and 0.270V2O1 is the principal uranium minerall it occurs with copper and iron sulfides such as chalcopyrite, chalcocite, covellite, bornite, idaite, pyrite, marcasite, and small amount of galena and sphalerite (Brodtkorb, 1963). No primary vanadium minerals are known yet. It is thought that the probable sourceof this element is the asphaltic material. In the alteration zone above the water-table, many hydrated yellow minerals like tyuyamunite, sengierite, , cuprosklodowskite, autunite, , andersonite, etc. are found in close associationwith copper minerals such as malachite, azurite and chalcantite. The three principal ore-bodiesof the area shown in Fig. 1 are Huemul, Agua Botada and Agua Botada Sur.

Pnvsrcar eNl Oprrcal PRopERTTESol HUEMULTTE Huemulite is yellowish orange to orangein color. It occurs as an aggre- gate of microscopic fibers, rounded massesforming thin films, and as interstitial filling in the sandstone. It is associatedwith the vanadates hummerite and rossite, and also with , and . The mineral is formed shortly after the levels and underground works are openedand is easily recognizedb_v its bright color. C. IJ. GORDILLO. E. LINARZS. R. O. TOUBES AND H. WINCIIELI.

gEe "fr\F ffii'r

o X8 z N ! @s E 86

dN v6

fE J* HMHffi >2

\

36

oq h< '= !6

E f,-r. o E :< o o 6C @ ) x">

o o

o 6 l- o _9

on67, iiiffilliliii,Lr""i"r HUEMULITE. A NEW VANADATE 5

Natural huemulite is too fine-grainedto permit the accuratedetermi- nation of its properties.However, it may be recrystallizedvery easilyby evaporatinga water solution made from it. The product gives the same r-ray powder pattern, and to the extent that other measurementscan be made on the fine-grainedmaterial, both recrystallizedand wholly-syn- thetic huemulite thus obtainedhave propertiesidentical with the natural mineral. The habit of the recr-u*stallizedmaterial and also of the synthetic prod- uct is tabular, similar to rossite(Palache et al., l95l), elongatedparallel to the c-axiswith perfect cleavageparallel to (001)and lessperfect paral- Iel to (010). The physical properties of the natural and recrl'stallized materialare the followins:

Natural Recrystallized

Color Yellowish orange Yellowish orange to reddish orange, depending to orange on the thickness of the crystals obtained. Yellow Yellowish orange Fluorescence None None Luster DulI Vitreous to subadamantine Hardness 2.5 to 3. Brittle Density 2 39 g/cm3

The optical properties were determined in the recrystallizedmineral and in the synthetic material, with very good agreementbetween both sets of data (Table 1). Under the microscope,thin crystalsof huemulite are yellow, pleochroic,biaxial negativewith 2V near 25oto 30o.The bire- fringenceis high, with colorsthat rangefrom blue to green.

Telr,r 1. Optrcer- Pnopnnrrcs or Hunuutrtr

Synthetic

Pleochroism X light yellow iight yellow Y golden yellow yellowish orange Z yellowish orange yellowish orange

Refractive indicesl d 1 679+.003 1.680 + .002 a 1.734+.003 |.734+.OO2 a 1.742+003 1.738+.002 2V, 20" to 25" 25" to 28" Dispersion / >x',strong />?,, strong

I The indices of refraction were measured by R. O. Toubes in the recrystallized mineral by C. E. Gordillo in the synthetic material, both using sodium light at 2O" C. C. E. GORDILLO, E. LINARES, R. C. TOUBI:.S AND H. WINCHELL

IN ,)"-I\:.-...-.

Frc. 2. Preliminarvontical orientation of huemulite.

X-ney Cnysr,lrrocRApHy Single cr1-stalsof the recrystallized mineral were studied using rota- tion, Weissenberg,and precessionmethods to determinethe dimensions of the unit cell. These data were then refined by using the powder data obtained on both natural and recrystallized material on a Norelco wide- range diffractometer. Zero-point corrections were applied to each diffrac- tometerchartbyreference to qtartz peaks(100) and (101)at20:20.840 and26.640respectively (CuKa:1.54050; anglecalculated from d-data of Swansonet al.,1953).The quartz was presentin the natural material as a minor impurit-v,and was added to the synthetic powdersbefore making the charts used for refinement. The single-crystal data permitted un- equivocalindexing of eight lines in the powder patterns; a least-squares solution using the 20 values for these eight lines permitted indexing several more lines and further refinement of the lattice constants. The final refinementgave an unreliability number Sa:0.00012, where Sq2is the sum of the squared differences between observed and calculated HAEMULITE, A NEW VANADATE

Tesm 2. Cnvster,r,oGn.rprrrcDere ol Huruur,nn aNn Svnrnnnc CoupouNo

[Iuemulite Synthetic

Symmetry Triclinic Triclinic a 11.770A+.0191 11.694A+.1231 b 11.838 +.008 11.869 +.053 c 9.018 +.009 9.049 + .061 q 107'13',+05', 107"28',+35', B 172"10,+06, llt" 52',+ 40', a 101'30',+05', l0l" 27'+ 3l' aibic 0.9943:l:0.7618 O.9853:l:07621 Volume 1040.67A+2.70 ro4l.26 A+ t7. So Space group P1 or PI Pl or Pl Unit cell contents NaqMgVroOzs.24HzO NaaMgVlqO23'24H2O Density: Calculated from NaaMgV16O2s.24HzO 2.404 2.403 Observed 2.39+.05

I Standard enor from least-squares solution; these estimates measure precision on the basis of internal consistency only. The least-squares refinement program is one written by C. W. Burnham of the Geo- physical Laboratory, Washington, D. C., to whom our thanks are due for a Fortran deck for compiling the program.

Q-values for the lines used, divided by 1a (6 parameters were calculated using 20 lines measured on two diffractometer charts made with natural and with recrystallizedhuemulite). Parallel runs on data from the same charts and from these charts plus one made with wholly-synthetic material, with and without correction terms to allow for specimen displacement from the diffractometer axis (Kasper and Lonsdale,1959, p. 221), showed (a) that the systematic correction-term parameters were much smaller (t:.38 to .59) than their own estimated standard errors, (b) that the unreliability number Se was greater with than without correction-terms, and (c) that the wholly-synthetic powder, either alone or with the others, yields poorer results for reasons still unknown. AII of the lattice constants thus calculated agree within their respective standard errors, with the data in Table 2. In Table 3, the r-ray powder diffraction data for natural, recrystal- lized, and synthetic huemulites are compared with calculated values obtainedfrom the least-squaresrefinement.

DrrnBnBNrrAL THERMALANar-vsBs Samples of natural huemulite and of the synthetic compound were analyzedin a Deltatherm apparatus, model D-200, using chromel-alumel C. L. GORDILLO, E. LINAR]'S, R. O. TOUBES AND II, WINCHL,LL

Tenr,r 3. X-Rlv Pownn D,r.re l.on Hurvulrrn Natunal, Rncnvstaltrzrn, .lNn SvNtunrrc (CuKa : 1.54050)

Film Difiractometer Calcu- Synthetic Natural Recrystallized lated dA dA dA I dA

I 010 10.586 to.77 90 10 59 85 10 61 100 r0 65 2 100 10 205 55 10 20 50 t0 20 65 t0.12 3 110 9.122 S 9.11 60 912 50 9 11 40 905 + r01 8 211 M 822 8.22 35 820 10 8. 19 5 001 7.622 VS 7.62 t00 762 100 65 762 6 111 6.043 607 r.) 607 10 604 20 608 5.673\ 7 Itlr 5.62 7 563 5 566 10 565 lzor s.62oj 8 2IO 5 .398 9 541 8 545 I 020 5.293 11 532 9 5.32 11 5.34 10 011 5.250 25 526 lo 5.25 10 5.29 ll 211 5. 150 10 5.18 10 5 5.18 t2 220 4.561 4.57 l4 1.567 8 4.559 o 4.562

IJ 112 4 488 13 4.485 8 I 49r 5 4.502 f2zr 4.0301 11 4.022 IJ 4 000 9 4.024 4 4 015 | 116 3.eesl IJ 022 3 953 20 3.956 l5 3.952 IJ 3 973 3.8081 16 lz2t 3 793 3. 795 20 3 792 10 3.798 lrBr 3.8021 lttl 3.6e8l. 17 3 681 20 3 696 IJ 3 681 8 3.699 lozr 3.6e61 18 I12 3.600 3 581 20 3. 580 20 3 580 11 3.566 3.470.I 19 It2o 15 3.469 7 3 467 l.) 3.450 I 3ot 3.4fif 20 312 3.305 7 3 305 + 3 300 6 3.298

2l 3 2'16 l.) 3 268 7 3 264 20 3.259 \12 3.27oJ 22 012 3 193 20 3 182 6 .t I /5 15 1a 220 3.160 1J 3.147 10 3 148 15 3. 153 fisr 3.071l 2L 112r 3.046i 30 3 054 3.060 l2 3.062 l33o 3.046) 25 2.967 11 2 .963 10 2.961 10 2.964 26 2.905 35 2.833 2.882 10 2 886 27 25 2.843 20 2.843 1,2 2.849 28 2.8r7 10 2.813 7 2.812 11 2.817 29 2.750 10 2.748 8 2 753 30 2 714 25 2.703 20 2 704 6 2,696

JI t7 2.617 11 2 61,6 1J 2.612 32 2 569 t) 2.570 10 2 566 9 2.569 33 10 2 532 7 2 533 6 2 529 2.489 13 2.489 7 2 493 8 2.496 35 9 2.399 9 2 399 10 2 396 36 9 2 339 5 2.339 7 2.339 M 2 293 l5 2 296 2 296 7 2.294 38 2 196 9 2.187 39 2.132 14 40 2.O94 12 2.1,Or 4l 2 045 10 2.053 }IUT,MULITE,A NLW VANADATE 9 thermocouples,kaolin calcinedat 950oC. as inert material, with a rate of heating of 10'C./min-l10/6, in an air atmosphere. For the DTA analyses,300 mg of pure samplesof both huemulite and synthetic were ground to 30 microns to avoid any error attributable to particle size and also permit the comparisonbetween different curves. The samplesu,ere left in a desiccator,with a humidity of about 50/6,Ior 48 hours to attain equilibrium. The curves obtained (Fig. 3) are similar for both huemulite and syn- thetic, with endothermicpeaks at 85oC., 133oC., and 169" C. The areas

TEMPERATUREOC.

Frc. 3. Differential thermal analysis curves of : A-natural huemulite; B-synthetic huemulite. correspondingto thesepeaks, measured with a Coradi polar planimeter showed that of the total 24 moleculesof water, 2 are given off to produce the first peak, 20 for the middle one and 2 for the third one. The small endothermic peak that appears in the curve of the natural material at 97' C. and not detectedin the synthetic, correspondsto gypsum present as impuritv. The small exothermicpeak between340o C. and 360oC., corresponds to recrystallization of the anhydrous moleculervith evolution of heat. The endothermicpeak at 580' C. correspondsto the melting of the mineral. The thermogravimeter curve was performed in a Chevenard thermo- balance,with photographic register and a rate of heating of 300oC./hour. 10 C. E. GORDILLO, E. LINAR'S, R. O. TOABES AND II. WINCIIELL

TEMPERATURE"C.

-t'rc. 4. Chevenard thermobalance curve of huemulite.

Samplesweighing 200 mg of both natural and synthetic huemulite were used.The curve obtainedis shownin Fig. 4 and is the samefor both. The lossof weight took placein three steps.The total losscorresponds exactly with the 24 moleculesof water and is complete at 255" C. The weight loss up to 150oC. correspondsonly to 17 molecules. The poor coincidencebetween the data obtained with thesemethods

T.tnre 4. Crrcurc,rr,Axer,vsrs ol Hunlrur,rro(Ne:runer, Merenrer,)

Sample 1 Sample 2

I. Insoluble in cold water 25.87a 25.437a II. Solublein cold water VzOr JJ.J 40.21 CaOl J-O 3.53 SOar 7.2 4.45 KrO 0.4 0.52 Na:o 4.2 3.94 Mgo n.d. 1.18 MnO n.d. 0.02 UOs (10 ppm (10 ppm III. Water on naturai sample H'O- 11.5 12.O0 H'zO+ 8.0 8.80

Totals 98-2 100.08

1 The content of CaO and SO: corresponds to gypsum as an impurity which was easily checked by optical and r-ray determinations. HUEMULITE, A NEW VANADATE

O\+cOO\ I O ON€€ O bo o;-i ld @ \ONIQ "d >\ 6'

q> \OOrcO I EX @.r \oioD I 6N X' .l XO -9 \o -a dv . l<,--o t r d > I \O\ON\O I *f i I\oNoi '; I H =si I a I 9ag'b q .: 1o<{No I d l3:5fr | ^U a tFt I b0 E< >'o \Oo*D lD *ro6!+r l3 6-\ od.i d ld \oNlq Fr

orNo\ | 40N@ | r t, .i \o:oD I .l Xd f :'si oo O\Oe+\O :>, c\ \ONSN I .dE \oNoi I Za n 'l ad sl a

q loNNo, I d bo; n lB:5fr | XO a nI ir I d- m i'< \ONOO I cO O4o\\O lo a q:a9\ -\ .l O\

€o4:tr I d fr d> brN*r I \ONON I N N .l -O= al o\oi\oo d6 !2o o'Z>Eo EF

^.o-o ^ G E ni u :. mv.. : SZ>E i 6 12 C. E. GORDILLO,E. LINARLS, R. O. I-OUBL:SAND H. WINCIIELL should be attributed to the poor precisionof the thermobalanceand to the differentrates of heating.

CuBlrrcar PnoppnrrBs eNl CourosrrroN Huemulite is easily soluble in cold water; the solution obtained is orange,vellowand the pH is 5.5 to 6.5, rvith no reducingpower. Thus all the vanadium must be as V5+. The mineral melts at 500" C. to a red liquid. The water solution of the natural material recrl'stallizesif evaporated slowly, giving the crystalsthat wereused for the single-crvstalstudy and for the final determination of the optical properties.In this casewe ob- tained cr.vstalsas big as 5 mm, weighing 15 to 20 mg. If the recrl.staliiza- tion is made in a more humid environment (by putting the Petri dish with the solution inside a larger container filled with water), instead of obtaining well formed crvstals, we got fibrous or radial aggregatesof acicuiar or needle-likecrvstals. Huemulite is easilysynthesized in a humid environmentb1,'putting the stoichiometricquantities of vanadium pentoxide,basic magnesiumcar- bonate and sodium carbonatein cold water. The mixture is heated in a water bath until the componentsdissolve and react with eachother, and CO: is eliminated.The solution is concentratedand the excessvanadic acid separatedb.v filtration. After this, the solutionis allowedto cool and cr1'stallize. Chemicalcompos,ition. The chemicalanalvses made on two impure sam- ples of huemulite are given in Tabie 4. The first chemicalanalvses performed on recrr.stallizedmaterial (Table 5) showed a deficit in the total; in the beginning of our study it was impossibleto synthesizea compound that would give the same r-ray diffraction data as the original.After severalruns and tests,we supposed that the deflcit might correspondto magnesium.A new chemicalanal1.- sis was perfomed and a s-v*ntheticprepared introducing magnesium in the composition.At this time we obtained triciinic crystals similar to huemulite and the r-rav data and optical propertiesof theseiast crops (H/5 and H/7) a"greever.v well with thoseof huemulite (Tables2 and 3). The chemicalanal.vses of the two different crops of s1'ntheticcrystals, of the recrvstallizedmaterial, and the theoreticalcomposition of huemul- ite are given in Table 5, which showsthat the unit-cell content is: NarMgVroOze.24H:O

RnlrruNcns

Bnoorrone, M. K. oB (1963) Mineralogia y g6nesis del yacimiento Huemul. fhesis, Univ. Buenos Aires (unpubl.). HUDMULITE, A NEW VANADATD IJ

ButNneu, C. W. (1963) An IBM computer program for least-squares refinement of Crystallographic Lattice Constant. Geophys.Lab. Carnegi'eInst.Washington, D C. EveNs, H. T. Jn., A. Gn,urllr Sw;rrr,ow, ANDW. H. Benwrs (1964) The structure of de- cavanadates.J our. Am. Chem. S oc. 86, 4209 - -- M E., Mnosn, AND R., ManvtN (1955) Constitution of the natural and artificial decavanaclates(abs.) . A m. M inerol., 4O,314. Kasrrn, J. S., eNn K. LoNsum (1959) (ed) International'I'ables for X-ray Crystallog- ' raphy, Vol. II,444. LrNenns, E. (1956) El yacimiento Cerro Huemul, Malargiie, provincia de Mendoza. 'I'hesis, Lrniv. Buenos Aires (unpubl ). ---C. E. Goporr.r.oaNn It O. Toutrs (1963) Huemulita, un nuevo mineral de vanadio hidratado de Ia Rep0blica Argentina. IL Iout. Geol.Argentinas, Salta (unpubl.). Palacun, C., H. BrnulN aNn C. FnoNoor. (1951)Dana's System of Mineralogy,4th. edit. SrlelNrcrc, P. N., O. L. Baurrns, F ILonnrco AND C. G. Menrrl'ns (1962) Dep6sitos uraniferos argentinos con control sedimentario. 4th. Interamerican Sltmp. Peoceful Appl. Nucleor l:.nergy, Mexico. Swrxsox, H Ir., R. K Fuyer .rNo G. M. UcnrNrc (i954) Standard x-ray difiraction polr''derpatterns. tI. S. Dept. Commerce,Nat. Bur. Sland. Ci.rc.539 (vol. 3), p. 24. Wnrxs, A., E A. CrsNBv eNn A. M. Snrnwoon (1950) Hummerite and montroseite, two newmineralsfromMontroseCounty,Colo. (.abs)Geol'.Soc.Am.Bull'.,6117573 -- ANDM. E. TrroupsoN (1954) Identification and occurrence of uranium and vana- dium minerals from the Colorado Plateau. I/.S. Geol.Suroey Bull'.IOO9-B'

Manuscr'i,ptreceited,June 1,1965; acceptd'-forpubli.cation, June30,1965.