THE HYDROTHERMAL CONVERSION of HORNBLENDE to BIOTITE 375 Cussedby Veblen & Buseck(1980)

Canadian Mineralogist Vol. 23, pp. 369-379(198s)

THE HYDROTHERMALCONVERSION OF HORNBLENDETO BIOTITE

GEORGE H BRIMHALL, CARL AGEE AND ROGER STOFFREGEN Departmentof Geologyand Geophysics,University of Califurnia, Berkeley,Califurnia 94720, U.S.A.

ABSTRACT (Roberts193, Moore & Cz?manskg1973, Titley 1982).Ces rdactions,ainsi que le remplacementde la titanite par anhy- Preliminary experimentson the biotitization of primary drite + oxydes de Fe-Ti, signaleraient les effets pr6coces igneous hornblende have been gompleted over a timpera- et trds r€pandussur les parois granitiquesdes solutions mag- ture range of 300 to 500'C at 1000 bars argon gas pres- matiques riches en sulfate qui circulent par convection prds sure. The experimentswere run with 0,10 m K2SOa-H2SOa des porphyres. Des 6tudes par diffraction X sur cristaux aqueous solutions and buffered by a solid a$"-lta6e oi uniques des produits de rfuction montrent que la transfor- . magnetite, microcline, quartz and pyrite. This buffer fixed mation solide-solide de hornblende en biotite mbne d une oxygenand sulfur fugacitiesat valuesappropriate for a fel- orientation pr6f6rentielle de la biotite sur la hornblende. sic intrusive rock undergoingpotassic alteration in the early Des tunnels parallElesd I'axe C de la hornblende seraient stageof the porphyry-copper environment. In theseexperi- des conduit pour les ions d diffusion rapide pendant I'alt6- ments hornblende was destroyed,producing biotite and ration. anhydrite presumably by the generalized reaction hornblende + K2SO4 + H2SOa* biotite+anhydrite+quanz; (Iraduit par la R€daction) the Caz+ liberated from the hornblende goes to make anhydrite. Biotitization of hornblende and re-equilibration Mots-clds: gisementde porphyre cuprifbre, hornblende,bio- of primary igneous biotite to lower Fe and Ti and higher tite, fluides acidesi sulfate, alt6ration potassique,bio- Mg contentshave beendescribed in high-temperaturealter- titisation de la hornblende, diffusion, tunnels. ation assemblagesaround the world (Roberts 1973,Moore & Czamanske1973, Titley 1982).These reactions, in addition to replacement of titanite by anhydrite and Fe-Ti INTRoDUCTION oxides, may represent the earliest and most widespread effects in granitic wallrocks, as magmatic sulfate-rich hy- Pervasive biotitic alteration, which entails the par- drothermal solutions begin convective circulation around tial or complete replacement of specific minerali by porphyries. Single-crystaldiffractioastudies of the products biotite in wall roci and in the porphyry host intru_ of experimentsindicate that the solid-solid transformation sive body, has been described irnumeious porphyry- of hornblende to biotite occurs with a preferential growt^h ;;;;rup.rttriiitf"y l9g2). Biotitization of horn_ of biotite in crystallographicorientation on the hornblenoe. ,-, I I . blende, the most common Tunnels paraliel to tie b axis of hornblende could be c# alteration of this type, oc- duits foifast-diffusing ions during alteration. curs in broad aureoles extending well beyond the limits of mineralization (Frg. l). The zone of biotiti- Keywords: porphyry-copper deposit, hornblende, biotite, zation is generally symmetrical with respect to zones acid sulfate fluids, potassium alteration, biotitization soalaining the earliest hydrothermal stages of depo- of hornblende,diffusion, tunnels. sition of copper and molybdenum sulfides, which have a fracture-controled disseminated occurrence (Roberts Sorrnao,ns 1973). In detail, near the outer fringes of this zone, where islands of unaltered wallrock per- sist, biotitization of hornblende can be shown on a effectu6,en exp6riencespr6liminaires, la transfor- to be mation en biotite d,une hornblendeprimaire-entre 300 et clearly related to the same veinlets that control al- 500oCi une pressiond'argon de 1000bars. Les exp6rien- teration halos. o-f the potassic or potassium silicate ces se font en pr€sence d'une solution aqueuse de type; these veinlets contain alkali feldspar, musco- K2SO4-H2SO4(0.10 m) tamponnde par l'assemblage vite, andalusite, biotite, quartz and anhydrite magn€tite-microcline-quartz-pyrite.ce tampon fixe les (Roberts 1973, Brimhall 1977). Using a variety of fugacitdsd'orygdne et de soufre d desvaleurs r€alistes pour geothermometers, a relativelv high raige of temper- alt6rationpotassique, a un irures is indicared sracre:1f]:',Tll1*i:'^:i:t-1-1*precoce crun environnement porphyre for the .intry nvatotrtermalsys- de cuprifbre. t.-, t9zi, noberts rg73, La hornblendeest d6truite dans ces expdriences, et lespro- tjzo_ioo.C tnri-rruu ,- --,-) o ^- _ l?J9).,.(-.i rr has beensuggested that the duils, biorite + anhydrite,seraient dus a ra reactionginl- If:P,1g,l-?ry raliseehornblenae i frsoo + H2SO4* Uiotltet ;rd- biotitization processis relatedto the initial develop- drite + quartz, danJ hquelleie ta2* , [b€r6 de ia ment of contrslling fracturesand representsthe first hornblende,formel'anhydrite. Labiotitisationdelahorn- hydrothermalstage in a long-livedsequence offrac- blendeet la rd-6quilibrationde la biotite primaireen bio- turing during which convective circulation of fluid tite appauwieen Fe et Ti et enrichieen Mg sont caracte- occurred (Iitley 1982),attended by an evolving com- ristiquesd'assemblages d'alt6rationi hautetempdrature position oitn"huia djrimnaU 1977).T:heclosegenet- 369 374 THE CANADIAN MINERALOGIST

GENERALUED NORTH.SOUTHGEOLOGIC CROSS SECTION BUTTE MINING DISTRICT LOOKING WEST STtrffiD-TYPE QUARTZPOPHNY Sup6rlms Otidstio PRffiTAcE) ,/(EARLY

ENRICHMENT SPEROENE CHALCOCITE BLANKET EI{RICHIEI'IT BWKET

MINE VEINS LEVEL t teoo' I I I zood Htpogene Oridollon+ sultidqtion (l6oching+ 6nrlcimenll "--v i -28oo' /" lt I: *rH! -' r,ndlr*il*" *., * \ :..-fif-il*gq"* ol ri i3400 PRO'IORE - - \ i mE MAt! srreE \ !' ZOilEO PYRITE-rurcOmrE PROrcRE HtaH TffiERArn€ (6ooil@'c) (-sluGTE aLTgnATtoN \. t csod XAIN 9TAGE VEINS DOUIilAMLY CULCOtrRITE YITH SORIIIIE AD FYRITE \ 0 500 looo 2000 rFEET ffi METERS 500 Ftc. l. Generalizedvertical cross-sectionof the Butte District of Motrtana, showingthe subsurfacezone of completely biotitized hornblendefrom the work of Roberts (1973).Its relationshipto the various elementsin the ore deposit is shown, including supergenechalcocite enrichment blanket, Main SrageVeins, and pre-Main Stage molybdenite and disseminatedchalcopyrite-pyrite protore (Brimhall 1979,1980, Brimhall & Ghiorso 1983,Brimhall et ol. 1984).

ic associationwith intrusive phenomenais indicat- studies have shown that a number of persistent ed by the pervasive biotitization observed in mineralogical effects attend biotitization (Roberts magmatic-hydrothermal breccias at the apical por- 1973).Igneous hornblende is convertedpseudomor- tions of porphyry dykes,where hydrofracture den- phically to a mixture of shreddy hydrothermal bio- sity is highest (Brimhall 1977). The first introduc- tite, anhydrite and quartz, whereastitanite is replaced tion of copper into the wallrocks occurred in this by a mixture of ilmenite, hematite, quartz, rutile and environment in the Butte District of Montana. anhydrite. Igneousbiotite is rimmed by shreddybio- Although the space-time pattern, temperature, tite or alkali feldspar, and generally contains small and magmatic-hydrothermal affinity of biotitization exsolvedgrains of rutile. Both the igneousand the have been well documented,there is a lack of ex- shreddybiotite in these rocks are lower in Ti and perimental data pertinent to an understandingof (1) higher in Mg than biotite from unaltered samples. the chemicalcharacteristics of the fluids responsi- Pyrite and chalcopyrite occur as minute grains in ble, (2) the reaction chemistry of the hornblende-to- both biotitized hornblende and igneousbiotite. biotite transformation, (3) the crystallographic Using thesemineralogical observations, we have mechanismsof the solid-statetransformation, and designeda number of hydrothermal experimentsto (4) its relationship to deposition of chalcopyrite and understand how the biotitization of hornblende pro- pyrite, mainly at sitesof mafic minerals. ceeds, what the necessary solutes are in the hydrothermal fluid, what controls the extent of the PETROGRAPHY replacementreaction, and over what time scalethe solid trausformation proceeds. The Butte District of Montana provides an ideal opportunity to study the effects of biotitization, as EXPERIMENTAL PROCEDURE the quartz monzonite wallrock is equigranular and homogeneousthroughout the district. Petrographic Following heavy-liquid separation in acetylene THE I{YDROTHERMAL CONVERSION OF HORNBLENDE TO BIOTITE 371 tetrabromide, a nonmagneticfraction of fresh horn- and fluid was microfiltered and collected with a blende was extracted from a specimen of Butte digital pipette; solids were immediately washed quartz monzonite. Grains were sorted by dry screen- thoroughly with distilled water and left to dry at ing to a mean grain-sizeof 155 pm. Their habit is room temperature. The time betweeninitial quench prismatic; length and width of the prisms avaage2}l and final 'ry4shingof solidswas on the order of two and 109 pm, respectively.In order to remove fine to three minutes. particlesfrom hornblendesurfaces, the samplewas A quick initial check for the survival of magnet- washedultrasonically in distilled water. Additional ite in the buffering assemblagewas done with a hand reactants were selectedto act as a buffering assem- magnet,after which the solid products were castin blage during the experiment; quartz, pyrite, chal- epory grain-mounts.A polishedthin sectionwas pre- copyrite, rutile and microcline were taken from com- pared from the hardened casting, such that it could paratively pure natural specimensand crushed to a be used for both petrographic examinalion and fine grain-size in an agate crucible. Reagent-grade electron*microprobeanalyses. magnetite and anhydrite were used to complete the Careful inspectionof the solid products was per- buffering assemblage,simulating the pre-Main Stage formed with a petrographicmicroscope using both potassium silicate alteration-assemblage(Brimhall reflectedand transmittedlight. An inventory of the 1977,1979,1980, Brimhall & Ghiorso 1983). mineralsin the buffer assemblagewas made to con- A 0.10 molar H2SOaand 0.10 molar K2SOareac- firm that the starting quantities had been sufficient tant fluid was injected, using a positive displacement to insure a surplus throughout the hydrothermal micropipette, into 7.5-cm seamlessgold tubes con- reaction. Specialnote was made of the estimated taining the experimentalcharges. The tubes were amount of pseudomorphicreplacement of horn- crimped and welded shut using an oxyacetylene blende by biotite, and selectedgrains were chosen microtorch. The tubeswere kept at room tempera- for electron-microprobe analysis. All electron- ture by partial submergencein water during weld- microprobe work was done with an eight-channel ing. Immediately after welding, the capsuleswere Applied ResearchLaboratories model SEMQ, using insertedinto 0.96-cm-bore310 alloy stainlesssteel biotite standardsanalyzed by wet-chemicalmethods. cold-seal rod bombs fitted with filler rods. Experi- ments were carried out in horizontal Hoskins elec- ExpspJr4eNTaLCoNDITIoNS AND trical furnaces equipped with chromel-alumel ther- TsxruRAr. REsur-rs mocouples wired to Wheelco temperature controllers. A continuously recording Honeywell A total of ten experimentswere carried out, at tem- potentiometer monitored the temperature of the peraturesof 300to 500oC,with partial to complete experiments.Argon gas was used as the pressure biotitization of hornblendeobserved in all cases.The medium and maintained at a confining pressure of biotite produced in thesereactions, whether rimming 1000bars. or completelyreplacing the hornblende,is invaria- Following eachexperiment, quenching was done bly light yellow brown, has the characteristic grainy at pressureby submergingthe rod bombs in cold surfacetexture of biotite, and appearshomogene- water, and the gold capsulewas quenchedin cold ous in transmitted light, although compositional het- distilled water for five secondsupon withdrawal from erogeneityon a small scaleis apparentin microprobe the bomb. The contentsof the capsulewere removed, studies(see below). At relatively low valuesof the

TABIA L DETAIIS Otr'EXPERIUENT DESIGN INCII'DINC VOLIJUES OF SOIJI' AND AQUEOUS R;EACTANTSAIID BESULIING TErIUNES-8ERIES I.CAGB

DIPT. T DPT. g EKPT.4

HombleailE 0.0045 eEr 0,0045 cms 0.0124 ems quartz 0.0058 o.o042 o.0039 Pyrite 0.0062 0.oo83 0.0082 ItricrcliBe 0.0096 0.0094 o.0093 Xagnetito o.o358 0.o1?5 o.0190 Aahydrite 0.0048 0.0060 o.00stt

o.l0 llrso4 - &go4 1.308 cmg 1.151smg 0.886 qmg PRESSURE lKbc lKbu lKbu TEUPERATURE 300.c 400!c 5000c DURA1ION Zl dats el daF 21 days

FSSIttTnIG TEI(- ComDlete bioiitiza- Conpl€te biirtitiza- BioEte rim, hom' TURE tioD tion blolde cores 372 THE CANADTAN MINERALOGIST

TAII,E 2. DETAIIS OF EXPERMENT DASICN INCLUDINCVOLUUES OT SOIJD AND AQUEOUSREACTAIITSJERIES II.CACB

EXPT. 1 E(PT.4 EKPT.5

HornbleDde 0.0059 cm3 O.0080 cu 3 0.0148cm3 quartz 0.0069 0.0055 0.0035 Pyrite 0.0091 0.0084 0.o084 Misroclite o.oo?9 0.0103 0.0099 Magnetite 0.0042 0.0052 0.0039 A!hydrite 0.003e 0.0057 0.0054 Chalcopyrite 0.0111 0.0101 0.0132 Rutile 0.0094 o.oxzz 0.0084

0.10 H8s04 - Keso4 0.601 cE3 0.598cm3 0.59? cm 3 PRESSURE lKbd lKba 1Kb0 TEMPERATURE 500.c 500"c 500"c DTIRATION 18 days 18 daF 18 days

fluid,/hornblende ratio. SeriesII-CAGB and III- EXPT.I ETPT.A EXPT.4 EIPT.6 CAGB (Iables 2, 3), all run at 5@oC but with differ- ent values of the fluid/hornblende ratio, substan- SONNilIENDE vol.ul(g tiated the relationship betweenthis parameterand

nESln;l'INC Blottt€rlft+ Fataiblotite blout€rlG+ Faidtbioitb the efient of biotitization. This i$ illustrated in Figure lEfIUnU hombleade rts + large hombldd€ rims + larSB shows relative extent of biotitization as ods hombl€bdo 9ores hombleDds 2, which the ooFa ooM a function of fluid/hornblende ratio. In the products of experimentsI-CAGB-4, II-CAGB-5 and III- CAGB-I and -4, with valuesof the fluid/hornblende fluid,/hornblende ratio, biotite forms complete rims ratio betweenl19 and 54, the grainsshow an exten- grows preferentially around hornblende crystalsand siverim of biotite while still containing a core of fresh in cleavagesor fracfures in the hornblende.At higber hornblende.Experiments with "complete biotitiza- values, biotite completelyreplaces hornblende, but tion" all featurea fluid,/hornblenderatio above74. generally preserves cleavagetraces of the original Those with only an incipient rim of biotite present grain. (III-CAGB-3 and -5) have values of 13.38 and The degreeto which this biotitization reaction 13.24.It thus appearsthat a fluid/hornblende ratio progressed variesas a function of temperatureand between54 and7{representsthe regionin which the fluid,/hornblende ratici in the charge.Tables l, 2 and last primary hornblendeis altered to biotite at the 3 summarizetextural observations from the experi- specificconditions of 500oCand I kilobar pressure. ments, along with data on the initial fluid,/horn-- blende values. In the I-CAGB seriesof reactions Elr,crnoN-MIcRoPRoBEDATA (Table l), completebiotitization occurredat 300to OSTaTNED AI-oNc GRAIN TRAVERSES 400oC, but fresh hornblendecores were preserved in the 500'C experiment,which containeda lower The varying degreeof alteration makes it necessary to examineelectron-microprobe data in terms ofparticular regionsor zoneswithin eachhornblende grain. Rims, coresand some cleavaggsform areas mFsr t&ithdbn that are discretecompositional units. A particularly good exampleof the various regionsof alteration is in grain from experiment II-CAGB-S liolilo rlG & observed a bnl&itm (Frg. 3). Five zones,labeled A, B, C, D and E, have been encounteredin a microprobe traverse.Zones Y{y tdnt ric 8 lorgo d€ A and E are "biotite rims", B and D are "hornblendecores", and zone C is most likely a pennea- ble region of cleavageor fracture. The traversetaken is representedby a solid line, with centresof "horn- vdm fl ulds/volm llmblende blendecores" analyzedat l0 pm intervals,whereas areasof compositionaltransition were analyzedin Ftc.2, Experimentalresults showing degreeof biotitization as a function of fluid-to-hornblende volume ratio. intervals of 2 pm. Average values for fresh horn- T of experiments500"C, P I kbar, fluid 0,1 m H2SOa blende(l) and zonesA through E arelisted in Table - K2SOa, mean grain-size 155 pm. 4. There are significant differences between the THE HYDROTHERMAL COI{VERSION OF HORNBLENDE TO BIOTITE 373

TAEIA 4. ELECTRON-UICROPROBERESULTS. AlrlAIY$S (t) IS STAffiNc I{ORNBLENDE, A THROUCI{ E ARE RSACTION ZONES.

II cAGB-5(e)

'rA B CDE (o pt8) (25 pts) (u I pts) (e ptr) (3 pts)

SiOg 49.48 48.29 4?.98 s1.76 46,56 42.93 no8 o.48 t.t2 1.00 1.33 1.?8 1.51 Abos 3.85 5.81 6.79 6.e8 4.26 8.40 FeO 14.10 12.08 L1.52 13.85 15.43 r2.54 xso 13.60 1?.89 11.6? 12,21 11.?8 16.60 U!O o.55 o.49 0.60 0.51 0.60 o,50 CaO 12.10 o.93 11.51 1.4? 11.?0 1.9? &o o.B4 8.S0 1.01 7.32 1.05 8.31 Nas0 4.72 0,0? 1.09 0.04 1.15 o.02 I. o.3? o.z9 0.?8 o.22 0.45 o.z6 cl o.|),t 0.01 0.o? 0.00 o.03 o.00 sos 'j' 0.31 o.35 t2.42 o.74 o.55 Ni0 o.o1 0.01 0.03 0.03 0.05 CUO o.02 0.00 0.o0 0.00 0.00 ?nO 0.00 o.00 0.o1 o.00 0.oo BaO o.02 0.01 0.o1 o.o1 0.o1 TOTAT s5.8o 9S.82 95.48 96.45 98.8? 90.75

CaO,z(CaO + S03) ln zone C = .3?5 versus .419 for uhydrite. 374 THE CANADIAN MINERALOGIST altered "rim" regions and the altered region in zone of a grain as they can near edges.It is possiblethat C. Calcium and SO3 are abundant in zone C, chemical-potential gradients are smaller in these whereas"rims" are depletedin thesecomponents, regions, as the.diffusion distanceis greater. with biotite again being the dominant phase. This Figure 4 (top) shows the antipathetic relation relation is similar to that found in "completely bioti- betweenK and Ca observedin the traverseshown tized" grains, where Ca and SO, concentrationsare in Figure 3, which is consistentin the "hornblende higher in the centralregions, suggesting that biotite cores" (zonesB and D) and "biotite rims" (zones and anhydrite are two important secondary phases A, C and E). Zone C, however,shows erratic varia- in the initial stagesof alteration, and that as alteration in K and Ca and could represent biotite inter- tion progresses,Ca and SO3are 6sfuilized, whereas mixed with a calcium-rich phase (anhydrite?) on a biotite componentsremain fixed. Another possible relatively fine scale.Silica, as expected,is depleted explanation for high Ca values in areas such as in alteredzones (Fig. 4, middle), with the minimum altered "cores" or fractures could be that Ca2+ions SiO, concentration correlating with areas of abun- cannot diffuse with as much facility in central parts dant Ca and SO3. Consistentlylow Al2O3values in secondarybiotite point to a paucity of free Al3* in the experimen- t! tal system.Occupanry of octahedral and tetrahedral !t. I tt ii sitesby aluminum in the newly formed biotite is com- it. '"-t parable to aluminum site-occupanciesin the fresh t rr fF. lo hornblende, and could thus be considereda rernnant F iiq ? E, F of the original, double-chain structure, as the !r experimental fluid contained no Al3* at the start of ;7 xa HW the experiment.Sia+ dominates tetrahedral sites in Tc s-.o cio !i biotitized rims, whereas Mg2" fills the majority of oa q; octahedralpositions as is liberated (Fie, 4' fr l: "FeO" DI 4) into the coexistingaqueous fluid. Et ii bottom; Table 0 oa0a0aoao tto t X-Rav-DrrrnacnoN SruDIEs AND E!o I STRUcTRALMoosrs FoRTFmREAcnoN INTERFACE o Powder-diffraction patternsconfirm the identity of the secondarybiotite; in addition, a preliminary single-crystalstudy was performed in order to check H eto2 for structural relationshipsbetween biotite and horn- B-{ ^taog blende.Structural relations between crystals of biotite and hornblende can be determined from the Weissenbergand precessionX-ray patterns taken. Figure 5 showsa first-octant stereographicprojec- r.9 tion of crystal-faceorientation as encounteredin the experi- I tl pseudomorphicallyreplaced hornblende from ment III-CAGB-4. Nore that the (h00), (hlco)afi I .l Hto o ^t o-At- (00/), (0/.0 H the (0/d) of hornblendecorrespond to the 'It --E -o'rtrr4 (0k0) other faces shown are .s - eF ffit and of biotite, but that P r"-o' ;:ii ..'E'i -a bpI {s mismatched by two to five degrees. i o ,cq ti Veblen & Buseck(1980) have reported a compara- ta I cl z il I i4I i ble replacement-texture between coherent talc and I i'drr Ilr anthophyllite. There, talc is intergrown with o x anthophyllite so that its y axis is parallel or sub- I d o parallel to the anthophyllite 7 axis, and its e* is I parallel or nearly so to the anthophyllite x*. Using parallel Figure 6A showsa portion of o ao a0 t0 !0 roo 180 140 iao lto a argument, ols^ftoE acnoaaoiatr (atoro!.) the hornblende-biotiteinterface represented on the scaleof a TEM image. Noteworthy are the bound- Ftc.4. Electron-microprobe for K, Ca, Al, Mg data Si, along the hornblende(l@) and (ll0), where and Fe variation witl distanceacross a traversethrough ary aretu zonesA, B, C, D and E, as shown in Figure 3. Zone transitions of double sfuainsinto infinite-width chains C containssporadic anhydrite, resulting in the Ca var- (i.e., phyllosilicate structure) are accompaniedby iation shown. Elementsare plotted as oxides. gapsthat are predicted by the termination rules dis- THE HYDROTHERMAL CONVERSION OF HORNBLENDE TO BIOTITE 375 cussedby Veblen & Buseck(1980). These gaps, or three-dimensionaltunnels, are parallel to the amphibole e axis and could be conduits for fast-diffusing ions during alteration. Using the model of Veblen & Buseck(1980), the (010) of hornblendecontain- ing no largetunnels would have only an ordinary rate of diffusion, and formation of biotite there would be slow in comparisonwith that along the (100)and (110) interfaces. Figure 6B representsthe region surrounding one hypothetical tunnel along the biotite-hornblende reaction interface (from circled "tunnel" in Fig. 6A). 6sgslding to X-ray data and averagevalues for unit- cell dimensionsin amphibolesand micas (Deer et al. 1966),the width of the hornblendeaqd biotite T-O- T layersis approximately 9.5 and 9.2 A, respectively, giving a tqnnel diameter in this configuration of about 4.5 A. ttre M(4) and / sitesin hornblende, along with the interlayer K site of biotite, border directly on the tunnel and could be easily occupied or vacatedduring transport of ions along this conduit. In hornblendethe/ siteis sometimesoccupied by Na+ and K+, whereasM(4) sites usually hold lli.$;r-:- Caz* (Hawthorne 1983). From microprobe data Frc. 5. Stereographicprojection of the orientation of describedabove, depletion of Na+ and Ca2+during secondarybiotite with respectto the host crystal of horn- biotitization of hornblendeis strong, and therefore blende. evacuation along these reaction-interface tunnels would seema likely mechanism. Potassium enrichment can occur by occupationof the / site, which '*1 becomesequivalent to the interlayer K site in biotite, as double chains transform to infinite-width E chains. Tunnel diameters can easily accommodate :_o__ large cations such as K+ , Ca2* apd Na+, with atomicradii of 1.33,0.99 and 0.95 A, respectively (Pauling 1960), so that significant transport of I material is possible. OJ, Microprecipitatesof SiO2(Veblen & Ferry 1983) deposited in the interlayer planes of biotite could explain the complex chemical composition observed AA\ in certain altered grains. Alternatively, the presence of tunnels$uggests another possible explanation. If cation diffusion along the hornblende (010)interface is much slower than in borders with tunnels, then removal of C*r will also proceed more slowly. Therefore, likely areasfor formation of calcium-rich phasessuch as anhydrite might be in planes parallel to the hornblende (010) or in zonesdirectly adjacent to a (010) face of hornblende.

REPRESET.ITATIONoF THE HoRNBLENDE-To-BIoTITE REACTIoN IN TERMS oF SoLUTToN CoMposrrroN Frc.6. A. Structural diagram of hornblende-biotite inter- The stoichiometry of the biotitization reaction face, showingposition of proposedthree-rlimensional described above and schematically represented as tubes in which large cations may be transported (after hornblende * K2SO4+ H2SOa*biotite + qrartz + Veblen& Buseck1980). B. Crystal structuresof horn- anhydrite clearly depends on the composition of the blende and biotite, illustrating the octahedral and tetra- hornblende and biotite involved. Because hornblende hedral sites. 376 THE CANADIAN MINERALOGIST in particular can show wide compositional variation, materials; however, it can be shown that the rela- the real biotitization reaction will be equally varia- tive amounts of iron and magnesiumin the solid ble. However,by a judicious choiceof end-member phasesdo not affect the dependenceof the reaction components,this reaction, along with 5imil4119ns- on the HrSO, and KrSOoactivities, the variables of tions that producechlorite as an alteration product interest in the current study. of hornblende or biotite, can be representedon a sin- The reactionsupon which Figure 7 is basedare gle activity-diagram that graphically illustrates the written conservingaluminum in the solid phasesand effect of the chemistry of the aqueous solution on assumingthe presenceof quartz and anhydrite. mineral equilibria. By consideringthe composition Thesereactions are: of the natural starting materials, the most appropriate amphibole end-member can be selectedfor use NaCa2Mgr(AlSi?Or(OH)2 + V2K2SO4+ 4H2SO4: on this activity diagram. edenite (aq) (as) Fresh hornblende from the Butte deposit is gener- KMgr(AlSi3OloXOH)2+ 4SiOr+ 2CaSOn ally intermediatebetween edenite and tremolite, a phlogopite quiltz anhydrite relatively low-aluminum hornblende. Edenite has + %Na2SO4+2MgSOa+4H2O (l) therefore beenselected as the hornblendecomponent (aq) (aq) in the activity diagram shown in Figure 7, as : present (composition 2NiCarMgr(AlSi?Or(OH)2 + I0H2SO4 appreciable K and Na are I in edmite (as) Table 4). In addition, stability fields of phlogopite Mg5Al(AlSi3Or0XOH\+ 4CaSOo+ Na2SOn and clinochlore are shown. These end members clinocblore mhydrite (as) neglect the presence of ferric iron in the natural + llSiO2+sMgSOo+8H2O Q) qustz (aq) = contoured tor Mg5Al(AlSi3oloXOHX+ K2Soo+ MgSoo+ 3SiO2 qtz, clinochlore (aO (ad quatz IncreaelngaNa2so4- + anhydr'lte 2KMg3(AlSi3OroXOH)2+2H2SO4 (3) pblogopite

i i phtosoptte Note that evenin this simplified system,four aque- tt i.': ous componentsstill must be consideredin represent- i i KMsg ing equilibrium relationsbetween these three solids. t, i, (o'srsoroxoHt In Figure 7, a two-dimensionalgraphical represen- tation of thesefour variableshas beenmade possible by combining HrSOaand MgSOaon one axis and contouring with respectto Na2SOoactivity. Increas- ing Na2SOaactivity will enlargethe stability field of A t t edeniterelative to both phlogopiteand clinochlore, s edenrre \ as shown by the dotted lines in the figure. The figure log also illustratesthat phlogopitemay be producedfrom aK2so4 NaCarMgU edenite by increasesin the activity of K2SOaor of cllnochlore HrSO. (or of both), a result that has also been (Alsl7o22)(oH)2 demonstratedby the experimentsdescribed above. Ms5A!(AtStg At lower activity of K2SOa, it is evident that clinochlorewould be the expectedproduct of replace- oro)(on), ment of the amphibole. Nnmerical valueshave not beenincluded in the figure owing to large uncertain- ties in the thermodynamicdata for the aqueousspecies and the solid edenite at the temperatures of the experiments. It is interestingto note that coefficientsfor the ' a \ '"'\'Hzso4,oq(a2 / 'Msso4/ -> aqueousspecies in thesereactions are dependenton the amount of aluminum in the amphibole. An increase in this aluminum content reduces the Frc. 7. Schematic activity-diagram showing relative stability-fields of edenite, phlogopite and clinochlore amount of H2SOainvolved in the reaction, and in with respect to activlties of aqueous speciesK2SO*, highly aluminous amphiboles,H2SOa and MgSOa H2SOa,Na2SOa and MgSOa, in the presenceof quartz actually switch sidesin the reaction, as in the case and anhydrite. Dashedlines show the effect of increas- of the amphibole pargasite,which may form phlogo- ing Na2SOaactivity. pite in the following reaction: THE HYDROTHERMAL CONVERSION OF HORNBLENDE TO BIOTITE 377

NaCa2MgoAl(AlrSi6Or(OH)2+ 5MgSO4+ MTNERALoGICALPnrg or,RnecrloN ANDCoMPAR. ptrgmite (aq) rsoN wrTH Nerunnr, SysreMs 3LK1SO4+ 3SiO2 * 6H2O= (aO qutrtz In porphyry-copperdeposits, hydrothermal bio- 3KMg,(AlSi3OroXOH)2+ 2CaSOa+ tite is typically more magnesianthan igneousbio- pblogopite mhydrite tite (Roberts 1973) and contains less Ti (Brimhall 1977). An overall test of the applicability of our experimental results, then, is to check our data 4H2SO4+%Na2SOn @) (aO (aq) againstthese ubiquitous effects. Paths ofhydrother- mal reactionsin natural biotite compositionshave been studied (Munoz & Swenson1981, Gilzean & However, for many cornmon igneousamphiboles, Brimhall 1984,Brimhall et al. 1983).In Figure 8 we reaction (l) is more appropriate, and this is certainly presenta compositediagram showing biotite com- the casein the experimentsdiscussed above. positions in terms of log (XF/Xoi versus log Reactionssimilar to (l) and (4), but written with (XM,/XE) octahedral.In the lower right corner we chloride instead of sulfate aqueous sDecies,are show the position of products from hydrothermal presentedby Eastoe (1982)in a discussionof the experiments.The averagedcompositions of rim bio- hornblende-to-biotitereaction inferred in the Pan- tite surrounding a hornblendecore are shown as hex- guna porphyry copper deposit. agons (experimentsII-CAGB-S average).The com-

UJ -F l- b o o HENDERSON A -|D cotoRADo .l- xo \ L \,x -l Ctl I g

MAGNETITE SIERRA JAmN* -2 SERIES NEVADA -0.5 0 o.5 tos(xM9/xrJB'otlrE

Ftc. 8. Plot of biotite compositions showing path of hydrothermal reaction in our experiments in relation to path of potassic alteration at Santa Rita, New Mexico. Three types of intrusive system are also shown: I) I-type granites relatedto porphyry-copper deposits,II) S- or l-type granitesof the irmeniteseries of Japan, relatid in iome cases to tin-tungsten deposits,and III) anorogenicMo-rich systems/e.g., Henderson,colorado). 378. THE CANADIAN MINERALOGIST

positions of biotite in completelybiotitized grains chalcopyrite, implyrng a link with the mineralizing with no hornblendecore are shown as solid circles process. (experiments.Il-CAcB-1 and -4). Thesecompletely Single-crystal-diffraction studiesindicate a nearly biotitized grains are more phlogopitic than the rim perfect alignment of the hydrothermal biotite with biotite and have a slightly higher log Xr/X6g. The the crystal structure of the original hornblende. path of hydrothermal reaction is shown by a solid. Three-dimensional tubes are inferred to occur black arrow. In the same systemsof co-ordinates betweenthe structure of both minerals, parallel to @ig. 8) we show compositionsof natural biotite from the C axis of the hornblende. These tubes may three distinct types of igneous system. Type I indi- representavenues for cation transport from the coex- catesthe field of biotite from I-type intrusive bod- isting fluid. ies, e.g., the Sierra Nevada suite (Dodge et al. 1969), the Boulder batholith of Montana (Brimhall el a/. 1983),the magnetiteseries of Japan(Czamanske el AcrNowt socsl'4ENTs al. l98l), and the SantaRita porphyry copperdeposit '(Jacobs1976, Jacobs & Parry 1979).The composi- The impetus for this study was the field work of tions of hydrothermalbiotite, part of the potassium the late StevenRoberts, friend and colleagueof the silicate alteration assemblageat Santa Rita, New senior author for years. His insight, enthusiasmand Mexico, forms a linear array with a positive slope wit have been sorely missed. NSF grant representingan isotherm (Munoz & Swenson1981) EAR-8115907has supported this study. Someof ttte for a processof high-temperaturealteration. Our electron-microprobeanalyses were performed by experimentaldata-points show that at 500oC, the David Sassani.Thoughtful reviewsby R.F. Martin direction and slope of reaction -" 1o*414 higher and G.M. Anderson improved the manuscript sig- valuesof the Mg/Fe ratio in the biotite, in accord nificantly. Their commentsare appreciated.We also with natural exchange-isotherms.Our experimental thank Adolf Pabst for his preliminary X-ray- results appear, therefore, to reflect the natural paths diffraction studies. of reaction of the biotitization of hornblende at Santa Rita. We show in Figure 8 two other magmatic-hydrothermal systemsin terms of biotite compositions. Type II representstypical S- or mixed REFERENCES S- and I-type granites, €.8., thellmenite seriesof Japan (Czamanskeet al. 1981).Type III are anoro- Brllvrnell, G.H., Jn. (1977): Early fracture-controlled genic intrusive bodies, e.9., the Hendersondeposit disseminated mineralization at Butte, Montana. in Colorado (Gunow et al. 1980). The ore deposits Econ. Geol. 72,37-59. for types I, II and III are copper, tin-tungsten and (1979): Lithologic determination of mass trans- (Munoz 1981, - molybdenum,respectively & Swenson fer mechanisms of multiple-stage porphyry copper Gilzean & Brimhall 1983). mineralization at Butte, Montana: vein formation by hypogene leaching and enrichment of potassium- CoNcr-usroNs silicate protore, Econ. Geol. 74,556-589.

Hydrothermal experimentshave produced the (1980): Deep hypogene oxidation ofporphyry biotitization of hornblendeat temperaturesof 300 copper potassium-silicate protore at Butte, Montana: to 500oCat I kilobar. The main factors controlling a theoretical evaluation of the copper remobiliza- the completenessof the reaction are time and the tion hypothesis. Econ. Geol. 15, 384409. volume ratio of fluid to hornblende for a certain Culnnrcnem, A,B. & SrorrnecBN, R. (1984): range in size of starting hornblende crystals. Biotiti- Zoning in precious metal distribution within base zation reactions in K2SO4-H2SO4-H2Ofluids metal sulfides: a new lithologic approach using producebiotite compositionsvery near thoseof the generalized inverse methods. Econ. Geol. 79, natural hydrothermal phase.Mineralogical reaction- 209-226. paths in terms of increasingXF/Xqg^arld XMe/XFe are also similar to paths in biotite from porphyry- & Gsronso,M.S. (1983): Origin and ore- copper deposits. forming corNequencesof the advanced argillic alter- process by mag- The net chemical effect of the pseudomorphic ation in hypogene environments matic gas contamination of meteoric fluids' .Ecoz. replacementof hornblende by biotite is to fix Mg2*, Geol. 7E,73-90. SOa2-,and K+ in the mineral assemblageas a mixture ofbiotite and anhydrite. The aqueousfluid, in -, GttzsAN,M. & Bunxnarra,C.W. (1983):Mag- contrast, becomesenriched in Fd* and Ca2*. Sil- matic source region, protoliths and controls on ica, releasedby the destruction of hornblende, prob- metallogenesis:mica halogen geocherustry. Amer. ably precipitatesas quartz, and the iron as pyrite plus Geophys. Union Trans. 64,884 (abstr.) THE HYDROTHERMAL CONVERSION OF HORNBLENDE TO BIOTITE 379

Czevraxsrr,,G.K., Isumane,S. & A'rrn, S.A. (1981): Moonr, W.J. & CzevaNsxr,G.K. (1973):Composi- Chemistry of rock-forming minerals of the tions of biotites from unaltered and altered mon- Cretaceous-Paleocenebatholith in southwestern zonite rocks in the BinghamMining District, Utah. Japan and implications for magma genesis."/. Econ. Geol. 6E,269-274. Geophys.Ree. 86, 10431-10469. DErn,W.A., Howre,R.A. & Zussuen,J. (1966):An MuNoz,J.L. & SwsNsoN,A. (1981):Chloride-hydroxyl Introduction to the Rock-FormingMinerals, Long- exchange in biotite and estimation of relative man, London. HCI/HF activitiesin hydrothermalfluids. Ecoz. Geol.76.2212-2221. Dooos F.C.W., Svnu, V.C. & Mavs,R.E. (1969)l Bioutesfrom granitic rocks of the central Sierra PauuNc,L. (1960):The Nature of the ChemicalBond Nevada Batholith, California. J. Petrology 10, (3rd ed.). Cornell UniversityPress, Ithaca, N.Y. 2s0-27t. Easros, C.J. (1982):Physics and chemistryof the Rosrnrs,S.A. (1973):Pervasive early alteration in the hydrothermalsystem at the Pangunaporphyry cop- Butte District, Montana. Soc.Econ. Geol. Guide- per deposit,Bougainville, Papua New Guinea..Ecor. book, Butte Field Mtc. IilI1-HII8. Geol.77,ln-$A, Trrrrv, S.R. (1982):Mineralization and alterationin GrzBAN,M.N. & BmvseLt, G.H. (1983):Alteration porphyrycopper systems. Iz Advancesin Geology biotite chemistryand natureof deephydrothermal of the PorphyryCopper Deposits in Southwestern systembeneath Silverton District, Colorado.Geol. North America(S.R. Titley, ed.).University of Ari- Soc.Arner. Abstr. Programs15, 581. zona Press,Tucson, Arizona. GuNow,A.J., LuoncroN,S. & Muxoz,J.L. (1980): Fluorine in micasfrom the Hendersonmolybdenite Vesr-EN,D.R. & Busscr,P.R. (1980):Microstructures deposit,Colorado. Econ. Geol. 75, 1127-1137, and reactionmechanisms in biopyriboles.Arner. Mirteral. 65,599-623. HewlronrB, F.C. (1983):The crystalchemistry of the amphiboles.Can. Mineral. 21, 173-480. & Frnny, J.M. (1983): A TEM study of the Jecoas,D.C. (1976):Geochemistry of Biotite in the biotite-chlorite reaction and comparison with petro- Santa Rita and Hanover-Fierro Stocks, Central logic observations. Amer. Mineral. 6E, I160-1168. Minine District, Grant County,New Mexico. Ph,D, thesis,Univ. Utah, Salt Lake City, Utah. & Pennv,W.T. (1979):Geochemistry of biotite in the SantaRita porphyry copperdeposit, New ReceivedMay 15, 1984, revisedmanuscript accepted Mexico.Econ. Geol.74, 860-887. February13, 1985.