5 1 1

T hz Catadian M ine ralo g is tVol. 36, pp. srl-524 (t998)

A MICROBEAM XAFS STUDY OF AOUEOUS CHLOROZINC COMPLEXING TO Iti|O.CIN FLUID INCLUSIONS FROM THE KNAUMUHLE GRANITIC PEGMATITE,

SAXONIAN GRANULITE MASSIE GERMANY

ALANJ. ANDERSONI

Department of Geology, St. Francis Xavier University, P.O. Box 5000, Antigonish, Nova Scotia B2G 2W5

ROBERTA. MAYANOVIC

Department of Physics and Astronomy, Southwest Missouri State University, Springfield, Missouri 65804, U.S.A.

SASA BAJT'

Center for Advanced Radiation Sources, The University of Chicago, Chicago, Illinois 60637, U.S.A.

ABSTRAcT

The synchrotron X-ray microprobe (X26A) at the National Synchrotron Light Source (NSLS), Brookhaven NationalLaboratory was used to collect zinc K-edge absorption spectra from saline (ca. 36 wt% NaCl equiv.) fluid inclusions in quartzfrom the Knaumiihle granitic pegmatite, in the Saxonian Granulite Massif, Germany, at temperatures ranging between 30o and430"C. XAFS specra were also obtained from one fluid inclusion that was experimentally re-equilibrated at a high pressureof hydrogen. The temperature of the fluid inclusions was controlled during analysis with a programmable heating stage.Analysis of the XAFS data shows that ZnCl42- is the dominant aqueous zinc species in the fluid inclusions up to th^e averagetrapping temperature (430oC). Furthermore, the mean Zn-Cl bond length decreases uniformly from 2.3 1 + 0.0 I A at roomtemperature to 2.26 t 0.02 A at 430'C. The predominance of the tetrahedral chlorocomplex, rather than ZnClro, at hightemperatures is most probably due to the high chloride concentrations of the inclusion brine.

Keywords: fluid inclusion, X-ray absorption fine structure, metalliferous brine, chlorozinc complex, granitic pegmatite,synchrotron X-ray microprobe, synchrotron X-ray fluorescence, hydrothermal fluid.

Sovuene

La microsonde X (X26A) h rayonnement synchrotron situ6e au National Synchrotron Light Source (NSLS), BrookhavenNational laboratory a 6t6 utilis6e pour obtenir des spectres d'absorption X au seuil K du zinc dans les inclusions fluides salines(environ 36Vo NaCl equivalents, en poids) dans le quartz provenant de la pegmatite granitique de Knaumiihle, dans le massifgranulitique de Saxe, en Allemagne, d des temp6ratures allant de 30" b 430"C. Des spectres XAFS ont aussi 6t6 obtenus dansle cas d'une inclusion fluide que nous avons r6-6quilibr6 tr pression 6lev6e d'hydrogbne. La tempdrature des inclusions fluidesa 6t6 maintenue pendant I'analyse grdce I une platine chauffante programmable. Une analyse des donn6es XAFS montre quele complexe ZnCl+2* est dominant comme espbce aqueuse de zinc dans les inclusions fluides jusqu'd la temp6r^ature moyennede pi6geage, 430oC. De plus, la longueur moyeme de liaison Zn-Cl diminue de fagon uniforme, de2.31+ 0.01 A l temp6raoreambiante jusqu' d.2.26 iO.O2 A a +lO'C. La pr6dominance du complexe tdtra6drique chlor6, plut6t que ZnCl20, tr temp6rature6lev6e, serait probablement due aux teneurs €1ev6es de chlore dans la saumure pi6g6e dans ces inclusions.

(Traduit par la R6daction)

Mots-clds: inclusion fluide, structure fine de I'absorption X, saumure m6rallifldre, complexe h chlorozinc, pegmatite granitique,microsonde X A rayonnement synchrotron, fluorescence X d rayonnement synchrotron, fluide hydrothermal.

I E-mail address: aanderso@stfx.ca2 Present address: Lawlence Livermore National Laboratory, P.O. Box 808, L-395, Livermore, California 94550, U.S.A.

512 THE CANADIAN MINERALOGIST

lvtnooucuolt

Knowledge of the speciation and structure ofaqueous metal complexes at elevated temperatures isrequisite for a complete understanding of the transportand deposition of ore metals in hydrothermal systemsassociated with igneous intrusions (Helgeson 1964,Barnes 1979, Crerar et al. 1985). Studies of fluidinclusions have shown that the fluids derived fromshallow silicic intrusions are usually hypersaline(Roedder 1992,Cline & Bodnar 1991) and commonlycontain high concentrations of metals such as Mn, Fe,Ct,Zn and Pb (Anderson et al. 1989,Heinich'et aI.1992). The stability of aqueous metallic species inchloride-rich hydrothermal fluids has been the focusof extensive research (Barnes 1979, Seward l98l);however, our understanding of metal complexing inhighly saline fluids at temperatures above 350oCremains vague (Hemley et al. 1992, Cygan et al. 1994,Heinrich et al. 1996). This problem is due in part to thelack of solubility data from experiments conducted attotal chloride concentrations greater than 6 molal (i. e.,26wtVo NaCl) (Wood 1995).

hovisional estimates of the structure of concenffatedaqueous alkali metal halide solutions at supercriticalconditions have recently been made by Oelkers &Helgeson (1993) using Monte Carlo simulations. Theysuggested that polyatomic clusters become dominantin concentrated electrolyte solutions and that thissignificantly affects mineral solubilities as well as thethermodynamic properties of the fluid. Consequentlnthe practice of calculating solute speciation inhigh-salinity fluids by extrapolation of resulrs fromexperiments on dilute electrolyte solutions may not bevalid. Using X-ray and Raman scattering measruementsat 25oC on zinc chloride hydrate melts, Yamaguchiet al. (1989) have noted that polymerization oftetrahedrally coordinated species occurs in solutionsof concentrafions of solute above l0 M.

Various spectroscopic techniques have been used todirectly investigate solution speciation at elevatedtemperatures (e.9., Franck 1973, Buback 1981, Seward1984, Susak & Crerar 1985). One of the more importantrecent developments has been the application of X-rayAbsorption Fine Structure (XAFS) to measurement ofion hydration in high-temperature solutions (Pfundet al. 1994, Fulton et al. 1996, Seward et al. 1991).Anderson et al. (1995) demonstrated the potential ofmicrobeam XAFS for determining the structureof metal complexes within individual fluid inclusionsa[ room temperature. [n the present communication, wepresent the results of the first XAFS investigation ofaqueous metallic species in high-salinity brine in singlefluid inclusions at elevated temperatures, and extendthe preliminary report of our study presented elsewhere(Mayanovic et al. 1996).

The fluid inclusions selected for analyses occur inquartz crystals from a wg in the Knaumiihle pegmatitein the Saxonian Granulite Massif, Germany (Fig. l).Similar granitic pegmatites have been describedelsewhere in the region (Nurse 1993) and are associatedwith polymetallic sulfide hydrothermal veins thatcontain sphalerite Q.{eumann & Tischendorf 1986). Thepresence of coexisting vapor-rich inclusions and brineinclusions in the pegmatitic quartz indicate entrapmentof a two-phase fluid. The analyses of high-temperatureXAFS data from the highly saline aqueous inclusionsshow that a tetrahedral chlorozinc complex is dominantin the fluid inclusions up to the temperarure of trapping,and that there is a progressive decrease in the Zn-Clbond length with increasing temperature. The availablecompositional and microthermometric data on fluidinclusions from the nearby hydrothermal veins(Anderson et al. 1997) suggest that boiling, dilutionand interaction with an external reduced sulfur speciescontributed to the destabilization of the zinc complexand the precipitation of sphalerite in the quartz veins.

GntiTSRAL Geolocy

The samples used in this study were obtained froma vug in a pegmatite in the Grimmscher granulitequarry, at a location known as Knaumi.ihle in theSaxonian Granulite Massif (Fig. 1). Both simpleabyssal pegmatites and complex rare-elementpegmatites occur in Proterozoic leucogranulites. Thegranulites are composed of quartz, mesoperthiticorthoclase and almandine, with subordinate amounts ofkyanite, sillimanite, oligoclase, rutile, corundum andhercynitic spinel.

The pegmatites at Knaumi.ihle consist mainlyof K-feldspar, qtraftz and biotite. Large crystals ofcordierite (up to 10 cm) are dispersed in the wall zone.The border zone contains cordierite and andalusitecrystals up to 5 cm in length. Enrichment of lithium,boron and fluorine in some of the more evolvedpegmatites is indicated by the presence of lepidolite,tourmaline (variety elbaite) and dumortierite. Therare-element pegmatites may be related to the nearby350 + 5 Ma Mittweida monzogranite (Fie. 1).

In the Hartmannsdorf quarry (.2 km northwest ofChemnitz; Fig. l), granitic pegmatites are spatiallyassociated with hydrothermal veins and stringers thatconsist of quartz, calcite and a complex association ofsulfides and sulfosalts. Jamesonite (PboFeSbuS,a) is thedominant sulfosalt in the veins, with minor amountsof arsenopyrite, l<illingite, pyrite, cassiterite, sphalerite,chalcopyrite, boulangerite, tetrahedrite and gold(Neumann & Tischendorf 1986). According to theseauthors, the hydrothermal veins are a continuation ofpegmatite mineralization.

CHLOROZNC COMPLEX IN FLUID INCLUSIONS : XAFS STUDY 5t3

Vl.XschisI envelope

N . . - - -v--^-4=---:-=- --,7,245$$SGi

7-4

t ' . ' . ' l

wNNNE

o#* .

Hc. 1. Geological map of the Saxonian Granulite Massif, Germany, modified after Franke (1993)..The asterisk designates thelocation of the granitic pegmatite sampled in this study.

gobbro

serpentiniteomphibolitecordierite gneissgronulitegronite

ExpeRryrENTAL MerHols

F luid inclusion p etro g raphy and microthermometry

Two types of fluid inclusions were observed indoubly polished wafers prepared from the quartz crystal.These are: l) high-salinity brine inclusions, which atroom temperature consist of a metal-rich aqueoussolution, a vapor bubble, and an assemblage of daughterminerals including halite and a high-birefringenceFe-rich phase (Fig. 2), and,2) vapor-rich inclusionswhich consist mainly of CO, with minor N2, as indicatedlaser Raman micro-analysis @. Thomas, pers. conrmun.).In some cases, the type-l fluid inclusions may containsylvite, calcite, wavellite and one or more unidentifiedphases. Both the brine- and vapor-rich inclusions attain150 pm in their maximum dimension, and occur withinhealed microfractures. The ratio of phases in type-1inclusions within a single fracture-plane appeais tobe nearly constant. In general, an individual fracturetrapped either brine or vapor-rich fluid inclusions. Lesscommonly, however, both types of inclusions occurwithin the same fracture-plane, indicating entrapmentof a two-phase fluid.

Volumes of fluid inclusions, and their deothsbeneath the polished surface ofthe quartz wafer, wereobtained using a modified spindle stage (Anderson1996) and a BIOQUANT OS/2 image analysis sysremat St. Francis Xavier University. Depth and volume

a\Ch€mnl tz

measurements were necessary for selection of suitablylarge, near-surface fluid inclusions, and for calculationof metal concentrations from X-ray fluorescenceyields.

Microtlermometric data were obtained usins aLinkham THMSG 600 programmable heating indfreezlng stage that was calibrated using syntheticfluid inclusion standards and the quartz cr-B ransitiontemperature. Apparent salinities of type-l fluid inclu-sions were estimated from the temperature of halitedissolution using the equation ofChou (1987), and thetemperatures of trapping were directly determined fromthe homogenization temperatures of cogenetic brineand vapor-rich inclusions.

Synchrot ron X-ray -fluore s c enc e microprobe analys is

Synchrotron X-ray-fluorescence (SXRF) spectrawere collected from individual fluid inclusions usingthe X-ray microprobe X26A at the National SynchrotronLight Source (NSLS), Brookhaven National LaboratoryNew York. Closely spaced tantalum slits were used tocollimate the X-ray beam to an incident-beam spot sizeof 12 x 9 pm. The inclusions were excited by a chro-matic (confinuum energy spectrum) beam covering therange of 3-30 keV. The X-ray fluorescence signal fromfluid inclusions was collected in air using a Si(Li)energy-dispersion detector positioned at 90" to theincident beam. X-ray maps of individual inclusions

ffi%fi'ffr'k

514 TI{E CANADIAN MINERALOGIST

''

Frc. 2. Tlpe-1 high-salinity fluid inclusion in quartz from the pegmatite.

were obtained by scanning the beam over tlle entireinclusion in 10 pm steps in both the X and Y directions.These maps were used to determine the disffibution ofFe,Zn,Pb and Br within a single fluid inclusion. Addi-tional details on the synchrotron X-ray fluorescencemicroprobe at the National Synchrotron Light Source(NSLS) can be found in Rivers et al. (1991).

Scanning electron microscopy - energy-dispersionarnlysis (SEM-EDA)

The SEM-EDA technique (Haynes et al. 1988) wasusefi.rl for partial analysis for light elements (Zbetween5 and 18) that could not be detectedby in situ X-raymicroprobe analysis of the fluid inclusions. Thinprecipitates of salt on the polished surface of quartz,directly above thermalty decrepitated fluid inclusions,were analyzed using an Electroscan EnvironmentalScanning Electron Microscope at the Bedford Instituteof Oceanography, Dartmouth, Nova Scotia. Spectrawere collected in the scanning and point modes atan accelerating voltage of 20 kV using a Si(Li) energy-dispersion spechometer with an ultrathin (0.3 nm thick)Novar polymer window. Data obtained from openedfluid inclusions are qualitative owing to uncertaintiesresulting from preferential loss of volatile species at thetime of thermal decrepitation.

Equilibration offluid inclusions at high pressures oJ

hydrogen

High hydrogen fugacities were generated in type-l

fluid inclusions in two quartz wafers prior to XAFS

analysis in order to assess the possible effects of

varying.(H) on zinc speciation in the inclusion brines.

Tlvo polished sections ofquartz containing type-l fluid

inclusions were placed in a gold capsule (25 mm

lengttt,4.4 mm OD,4.0 mm ID) containing CrN (from

K & K Lab. Inc.) and HrO. The capsule was sealed

using an arc welder and then weighed before placing it

into a cold-sealed pressure vessel fabricated fromStellite 25 dloy (from Haynes Stellite Co.). The vessel

was pressurized by CHa and held at2kbar and 600'Cfor a period of six days. The total pressure of the vesselwas maintained for the duration of tle run by daily

replenishment of Ctt to compensate for the hydrogenloit through the wall of tle pressure vessel. A high

hydrogen pressure was generated within the gold

capsule by the reaction of CrN with HzO at 600oC, andby ttre use of the graphite-methane buffer, which

significantly reduced the hydrogen fugacity gradient

across the gold capsule wall. At 600'C, hydrogenwithin the capsule will readily diffuse though thequartz and into the fluid inclusions, as observed by

Mawogenes & Bodnar (1994) in a similar experiment.

CHLOROZINC COMPLEX IN FLUID INCLUSIONS : XAFS STUDY 5 1 5

After six days, the pressure vessel was removedfrom the furnace and cooled slowly to room temperature.The gold capsule was then extracted from the pressurevessel, dried, and weighed a second time to ensure thatleakage had not occurred during the run. The polishedwafers were then removed from the capsule andwashed.

Collection of XAFS data

Temperature-dependent Zn K-edge XAFS spectrawere collected in the fluorescence mode from indi-vidual fluid inclusions situated befween 10 and 20 pmbelow the polished surface of the quartz wafer. XAFSspectra were also obtained at 400'C from a fluidinclusion that was re-equilibrated at high hydrogenpressures (see above).

Electrons within the NSLS X-ray storage ring weremaintained at beam current levels ranglng between 130to 230 mA, at 2.54 GeV. The X26A beam line setupused to make the XAFS measurements is as follows: awhite synchrotron X-ray beam entering from the frontend of the beam line was made monochromatic (4-20keV range) using a single Si(lll) channel-cut crysraland focused with a 8: I ellipsoidal Al @ coated) mirror,giving a flux of - I a I Qs photons/s at 230 mA of storedcurTent. The energy resolution using this setup was1.5-2 eV. Harmonic rejection was accomplished withthe focusing mirror, since it has an energy cutoff ofroughly 14 keV. The 30 mmz Si(Li) energy-dispersiondetector was used to collect the X-ray fluorescencesignal in air. A pair of parallel 15-cm-long metal plates,set I cm apart, were positioned near the back end of theHe-filled beam pipe. The detection of the ionization ofHe by X-rays passing through the gap between theplates was used to measure the intensity of the incidentbeam (I").

During XAFS measurement, the temperature of thefluid inclusion was controlled using a modifiedLinkham THMSG 600 programmable heating stage,which was mounted ontheX-Y-Z-O positioning stageof beam line X26A. The surface of the quartz waferwas precisely positioned uslngthe X-Y-Z-O stage andan optical microscope to give a 45o orientation to boththe incident X-ray beam and the detector. In order tooptimize tlte fluorescence yield, the size of the X-raybeam spot (as measured on the sample) was adjusted,using fine collimating slits and a pinhole mask, to theapproximate dimensions of the fluid inclusion (1.e.,6O x 24 pm). XAFS spectra were measured from theinclusion at temperatures ranging from 25 to 430oC(+loc).

XAFS spectra were also measured at roomtemperature from standard zinc chloride (t sodiumchloride) solutions having a chloride to zinc ratiobetween 2 and 8. The preparation of the solutions wasdescribed in Anderson et al. (1995). The solutions wereinserted into plexiglass sample containers with thin

TABLE l. THEZ8-CI BOND LENGffi G&e),C@RDINATIONNTJMBER (No)I AI{DMEAN-SQUARE

REI.ATTVE DIITORDER (d) RF,SULTS,AND fiIE WNDOW RAilIGES FOR CN-CUI.ATION

OF FOURIER TRAI.ISFORIVIS (K.RAI{GE) ANDTHE FTITING RANGES (R-RAT{GE) FOR THE XAFS SPFCTRA

MEASTJREDBETWEEN 25' AND 430'C

T R^o 1,{" dA A '

k-mgs R-rugBA 4 A

25t702503@430

2.30+ 0.012.29 * 0.012.29 t 0.V2224x0.012.26+O.92

4.0 * 0.4 0.0043 * 0.0004 3.5J.0 t.0-2.43.8 * 0.6 0.00et + 0.0@5 2.&{.75 t.+2.t54.OrO.4 0.@66+0.0005 2.7-9.5 0.65-2.154.t+0.4 0.0070+0.@6 2.94.3 0.612.254.1 + 0.5 0.0089 * 0.@07 2.w.1 r.2-2.15

Kapton windows in the walls to permit the transmissionof both the incident X-ray beam and the resultantfluorescence X-rays. Three sets of XAFS spectra wereobtained from the standard solutions and from the fluidinclusion at each temperature interval.

Analysis of MFS data

Data analysis was made using the University ofWashington software package (UWXAFS, version2.01) according to the principles outlined by Sayers &Bunker (1988). Isolated oscillations (r() of the XAFSdata were produced by removing the pre-edge back-ground using a simple second-order polynomial fit andby removing the above-edge background using a cubicleast-squares-spline approximant over the 28-340 eVrange. Fourier transforms were calculated from the kz1data using a modified Hanning function-type windowover the k-ranges shown in Table l. Quantitative struc-tural information for the Zn environment was derivedusing theoretical spectra generated with FEFF 6(Zabinsky et al. 1995, Mustre de Leon et al. l99l).Thereader is referred to Anderson et al. (1995) for anexplanation ofthe advantages ofthe XAFS techniquefor structural analysis of the aqueous species in fluidinclusions. Additional details on data collection andanalysis are given in Mayanovic et al. (1996).

Rnsu-rs

Microthermometry

Figure 3 shows the range of liquid-vapor homog-enization temperatures (Tht-u) versus the temperatureof halite dissolution (Tmruu.,) for high-salinify (type-l)fluid inclusions. The Th1*y values for these inclusionsaverage 90"C higher than the melting temperature ofhalite, and are therefore suitable for determinationof apparent salinity using the temperature of halitedissolution, Trnru.,. Assuming a simple HrO-NaCl

5r6 THE CANADIAN MINERALOGIST

ocooIJcoEDoEoToCLo?3:ctf

Ftc. 3. Temperarure of halite dissolution uersas liquid-vaporhomogenization temperature for type-1 fluid inclusions.The filled and solid symbols designate brine inclusionsthat belong to separate fluid-inclusion-decorated planesI to VII.

system, Tmpu.1 yields apparent salinities between 33and 37 wt.Vo NaCl equivalent.

The Th1-y for type-l fluid inclusions is between 320and 480"C. Howeveq fluid inclusions trapped within asingle healed fracture show a more restricted range oftemperatures (Fig. 3), indicating different "closuretemperatures" for different microfractures. The occur-rence of brine and vapor-rich inclusions within thesame healed fracture is evidence of entrapment offluid in the two-phase (liquid + vapor) field. Thereforetle temperature of homogenization is assumed to be thetemperature of trapping, with no pressure correctionrequired (Roedder 1984).

Using a temperature of trapping of 480"C (i.e., themaximum homogenization temperature) and an averagesalinity of 35 wt.Vo NaCl equivalent, an estimatedpressure of trapping of about 0.5 kbar is obtained usingthe vapor pressure curves for H2O-NaCl solutionshaving salinities from 30 to 70 wtTo NaCl of Bodnar &Vityk (199a). This is regarded as a minimum estimateof pressure because it does not account for the presenceof COr, which significantly increases the T-P range ofimmiscibiliry.

Elemental analysis of fluid inclusions

Figure 4a shows a typical SEM-EDS spectrumobtained from salt precipitates on the quartz surfacedirectly above a thermally decrepitated type-l fluidinclusion. The SEM-EDS results consistentlv show the

presence of Na, Cl, K, Mn, Fe andZn; sulfur was notdetected in any of the precipitates. The SiKcr peak isgenerated from the quartz wafer. Figure 4b shows asynchrotron X-ray fluorescence spectrum obtainedfrom an unopened type-1 fluid inclusion. The spectrumreveals relatively high concentrations (x1,000 ppm) ofMn, Fe, Zn, Br, Sb and Pb. The estimated averageconcentration of zinc in this inclusion is 0.4 wt%o.Argon Ko and TiI(q peaks are from the air and quartz,respectively. Na, CI and K were not detected by SXRFowing to attenuation of their signal by the overlyingqnafiz.

Zn K-edge XAFS measurements

The X-ray absorption coefficient at the Zn K-edgefor standard solutions and a type-l inclusion over arange of temperatures is shown in Figure 5. Thetemperature at which each spectrum was collected isshown with each spectrum. Figure 6 shows the isolatedextended XAFS (kz2g) data of the Zn K-edge spectra inFigure 5.

Fourier transforms calculated from the k2X data arcshown as points in Figure 7. The solid lines in Figure 7represent the best-fit using FEFF 6 calculations

(b)

arsl rl

5 1 0 1 5

X-ray EnergY (keV)

Frc. 4. (a) A typical energy-dispersion spectrum from anevaporite produced by thermal decrepitation of a type-1fluid inclusion. (b) A synchrotron X-ray-fluorescence(SXRF) spectrum collected from a rype-l fluid inclusionsuch as in Fisure 2.

a lo t lA rrtA r vO vt r v rI vll

{04

't03

Q

E to"

6 to'

104

103

102

io t

o trottr ll

t oat ^^

ir '

Halite Dissolution CC)

(a)

Fe

Mn

VnCl2"] aqueous solutlon 25 "C

lZ:nC;l421 aqueous solutlon 25 "C

Fluld Incluslon 4il0 oC

Fluld Incluslon 290 oc

Fluld lncluslon 250 oC

Fluld lncluslon 25 oC

CHLOROZINC COMPLEX IN FLUID INCLUSIONS : XAFS STUDY 517

0.5

0.4

0.3

0.2

0 . 1

coo.t-oatl

k2-x

T = 4 3 0 o C

T = 3 0 0 " C

T=250oc

T = 1 7 0 o C

T = 25"C

0.0

-0.1

-0.210

9700 9800 9900

Energy (kev)10000

Frc.5.Zn K-edge XAFS spectra collected in the fluorescencemode from a type-l fluid inclusion at temperaturesranging from 25 to 430"C. Also shown ue Zn K-edgeXAFS spectra collected at room temperature from zincchloride (ZnCloz-1 and (ZnCl20) aqueous solutions.

assuming a ZnCloz- model. The Fourier transformsof the XAFS data are in excellent agreement withthe theoretically generated results, indicating that thedominant aqueous zinc species in the inclusion brineup to 430"C is ZnCloz . This conclusion is supportedby the close correlation between the XAFS spectraobtained from the fluid inclusion and those from thestandard solution containing ZnCloz- (Fig. 5). IdenticalXAFS spectra were also collected at 400'C from atype-l fluid inclusion that was re-equilibrated athighflH) conditions, indicating the predominance ofZnClo2 in this brine even at artificially high hydrogenpressures.

Table 1 lists the Zn-Cl bond lengths (Rrn-,),coordination number (N.,), and the mean-square rela-tive disorder o, (MSRD) for the dominant chlorozinccomplex in the inclusion brine over a range of tempera-tures, as determined from the fitting of the Fouriertransform data in R-space. The results show that thecoordination number of about 4 is constant over theentire range of temperature. There is, however, asignificant reduction observed in the Zn-Cl bondlength with increasing temperature by nearly 0.01 Aper 100oC (Fig. 8).

k (A-')

FIc. 6. The isolated extended XAFS (kr1) of the Zn K-edgeXAFS spectra shown in Figure 5.

DIScUSSIoN

Composition and origin of the fluids

The type- I fluid inclusions are dominated by NaCl,KCI and FeClr, with lesser concentrations of Mn,Zn,Pb, Sb, Br and Rb. Laser Raman analyses showvariable but low amounts of CO2 in the type-l fluidinclusions (Anderson et al. 1997). A significantdifference in composition between the present fluidinclusions and the highly saline fluid inclusions inporphyry Cu systems is the relatively high concentra-tions of Sb and the low concentrations of Cu (belowdetection). Campbell (1995) noted that fluids derivedfrom the Capitan pluton, New Mexico, were low in Cu,and suggested that little Cu was present in the melt atthe time of fluid separation.

Both brine- and vapor-rich fluid inclusions occurwithin healed fractures and are therefore presumed tobe pseudosecondary or secondary in origin. The highmetal concentrations in the fluid (i.e., Mn, Fe, Znand Pb) are in broad agreement with theoretical andexperimental predictions for chloride-rich brines thatequilibrated with granitic melts at high temperatures

5 r 8 THE CANADIAN MINERALOGIST

4.5

4.O

3.5

3.0

2.5

2.O

1 . 5

1 . 0

0.5

0.0

D ' I \Ir [fr'

FIc. 7. The magnitude of the Fourier transform (FT) (dots) generated from the yzy datashown in Figure 6, and the best-fit model (solid lines) of these data using the programFEFF 6 and aZnClaz model.

o!g

J

ltt-

oE'=ftro)(u=

hlJ.

(Eugster 1985, Whitney et aI. 1985). Furthermore, thehigh concentrations Rb and Cs in the type-l fluidinclusions (Anderson et al. 1997) is in agreement withfluid compositions associated with highly fracdonated,rare-alkali-enriched pegmatite-forming mels (Lagacheet al. 1995).

Secondary and pseudosecondary fluid inclusions ofsimilar composition and filling temperatures have beendescribed in other highJevel granite-pegmatite systems(e.9., Eadington 1983, Heinrich & Ryan L992);thesehave been interpreted to be "near-pristine" samples ofmagmatic fluids. The high chloride concentrations inthe present fluid inclusions relative to those foundin spodumene-bearing pegmatites of the rare-elementclass may be due in part to the lower pressure at whichthe fluids were separated from the melt at near-solidus

conditions (Cline & Bodnar 1991). Boiling also mayhave increased the salinity ofthe rype-1 fluid inclusions.Aqueous fluid inclusions in spodumene-bearingrare-element pegmatites, such as at Tanco, Manitoba(London 1986), tend to have lower salinities (ca.7 wt.VoNaCl equiv.).

In some low-pressure systems, the final stages ofpegmatite formation are marked by the separation of anaqueous fluid from the melt and the development ofcrystal-lined pockets and vugs. Subsequent "pocketrupture" (Jahns 1982) resulting in a sudden reductionin pressure may have induced microfracturing of thecrystals and the unmixing of CO2 from the fluid.Entrapment of brine- and vapor-rich inclusions byclosure of the microcracks occurred over a temperaturerange of about 150"C (Fig. 3). The high concentrations

CHLORoZINC COMPLEX IN FLIIID INCLUSIONS ; XAFS STUDY 519

{ z.eooC)= 2.28

JJ.9O z.ze6

I

fi z.za

2.U

2.32

0 100 200 300 400 500

Temperature fC)

Ftc. 8. Variation in the Zn-Cl bond length of the tetrahedralchlorozinc complex as a function of temperature. Thedashed line is the linear fit of the data.

of Fe (percent level), Zn, Sb and Pb (several thousandppm each) in the fluid inclusions favor a direct rela-tionship between the pegmatites and the neighboringjamesonite-bearing quart z veins.

Neumann & Tischendorf (1986) proposed that thehydrothermal veins in the Hartmannsdorf quarry inthe Saxonian Granulite Massif are a continuation ofpegmatite mineralization, but suggested that the metals(1.e., As, Zn, Fe, Sb, Pb) were mobilized from thesurrounding host-rocks by retrograde processes ofVariscan metamorphism. This interpretadon is based inpart on a Pb isotopic model age of 354 Ma obtainedfrom jamesonite in the hydrothermal veins.

If the mineralized, qtartz veins are indeed acontinuation of pegmatite mineralization as proposedby Neumann & Tischendorf, then it follows that the oremetals were derived from the pegmatite-forming melt,and not directly from the host granulites. tn addirion, itis unlikely that the formation of, and metal enrichmentin, the rare-element-enriched pegmatite-forming meltsis due to partial mehing of a local granulite source-rockduring metamorphism in view of the low pressures andtemperatures indicated by t}te presence of: l) cordieriteand andalusite in the pegmatite (eernf & Hawthorne1982), and 2) fluid inclusions trapped in the two-phase(liquid-vapor) field. The inferred conditions areinsufficient for the generation of anatectic melts fromgranulite host-rocks.

In a study of sulfosalt-bearing veins of the Mid-European Saxothuringian zone, Dill (1985) rejected thegenetic model invoking remobilization of metals fromthe enclosing country-rocks because the veins bearingthe Pb-Sb sulfosalts cross-cut host rocks of differentage and lithology and typically have Sn as a majortrace element of the fahlores. Furthermore, Dill's Pb/Pbisotopic data suggest that the veins are related tocalc-alkaline granites. on the basis of the compositionof the fluid inclusions, we suggest that the fluids in therare-element pegmatites at Knaumiihle are geneticallyassociated with the hydrothermal quartz veins oftheregion. The source and evolution of these hydrothermalfluids are the subject of a detailed study currently inprogress (Anderson et al., in prep.).

Zinc complexing in the hydrothermalfluids

Susak & Crerar (1985) have emphasized thatknowledge of the molecular properties of hydrothermalfluids is essential to a complete understanding ofmineral solubilities and hydrothermal processes.However, owing to experimental difficulties such ascorrosion, there are essentially no experimental dataavailable on metal complexing in high-temperature,high-salinity brines such as those associated withmineralized felsic intrusions (Roedder 1984). At

TABLE 2. CALCUI,ATED AND APERJMENTAL BONI}ITNGTIISFOR AQUEOI'II ZNC.BEARJNG SPECIES AT R@M TEMPERATTJRE

speies Botrd Bold le$h (A) R€&@@Cslsldql Obsw€d

zi(Hro)? Pmhrdaar (l!}96)Tossell (190)

2.08 Otuhietql(lYlq

PKn6€dstal (l99OTossl & Valgbm (1992)

2.u M&!d^et aL (1996)

Prcbne@sraI (19OTocadl & Veghm (1992)

2.24 Mafd/set aL (196)

Prchemetal (195)Toesl & v@ghe (1992)

2.2842) Mafdzetsl(19p6,

Tose[ (190)PEc.h€dst4I (1996)

2.28, 2.30 Ksh& Sradtry(1962)2.2942) Mtfd,setaL(196)

Tosse[ & Veghs (1993)

Toecl & Vugha (1993)Tffi[ & Varghe (1993)

2.39 Helzeral (1993)

238 Dreiq&Rabe(19E6)

2.41 Drei6&R8bs(19E6)

73Cl-

Zackil

7^L\'

7aL\b

z($t !

z(srr),cl

TaBrf

A-srt

Z^4 2.147^4 2.@784

7^4 2.@7^4 2.297^4

h4 2.t37^11 2.21Zi4

h4 2.267*4 2.237e4

2.4

2.33

7i47^47^47^4

Z!-S

Zs-S7^4z^4

zi-Br

7a-Br

2.39

ZnClo,

zncrf,/

520

present, the most reliable data on aqueous speciesin hydrothermal solutions are biased toward lowtemperatures (r'.e., between 25 and 300"C) (Heinrich etal. 1996) and relatively low chloride concentrations(< 5 m Cl-) (Wood 1995). Furthermore, most experi-mental studies involve relatively simple solutions thatapproximate natural hydrothermal fluids and thusdo not take in to account the influence of minorcomponents and various volatile species on solubilitiesof ore minerals (Walther & Schott 1988, Wood 1995).

The results presented here demonstrate thatmicro-beam XAFS analysis of highly saline natural andsynthetic fluid inclusions at elevated temperatures canprovide much needed data on metal speciation in brinesof magmatic derivation. However, Hall et al. (1989),Morgan et al. (1993) and Sterner et al. (1995) haveargued that natural fluid inclusions may undergocompositional re-equilibration after trapping, whichmay influence some aqueous species in the originalentrapped fluid. Although it is generally accepted thatthe major constituents in fluid inclusions are preservedin most geological environments, there is strongevidence that hydrogen migrates in or out of fluidinclusions in quartz in response to a chemical potentialgradient between the inclusions and an ambient fluidreservoir. Hydrogen loss from a fluid inclusion mayinfluence the/(Or) and pH conditions of the originalfluid. Roedder (1984) suggested that the presence ofinsoluble daughter minerals, such as chalcopyritein fluid inclusions from porphyry copper deposits, isevidence that the original redox state of the fluidinclusion was modified by diffusion of hydrogen.

In order to assess the possible influence of naturalloss of hydrogen on zinc speciation in the fluid inclu-sions studied, XAFS spe€tra were obtained from type-linclusions that were experimentally re-equilibrated athigh hydrogen tugacities. The.flH) in the experimentallyre-equilibrated inclusions is, most probably, abnormallyhigh relative to that in the original trapped fluid(Mavrogenes & Bodnar 1994). Nevertheless, the XAFSresults show that the dominant zinc-bearing species inan inclusion subjected to highXHr) lsZnClo2-, which isidentical to that in the untreated type-1 fluid inclusions.Although the actualflH2) condition in the original fluidinclusion at the time of trapping is not known, thereseems little doubt that the comparatively small changesin /(Hz) expected to occur by natural diffusion ofhydrogen would not influence the stability of chlorozincspecies. Furthermore, the absence of insoluble solidphases and the dominance of aqueous Fe2* in the fluidinclusions studied (Mayanovic & Anderson 1997)indicate that the original redox state of the entrappedfluid was not significantly modified, as is the case inmany fluid inclusions from high-temperature porphyrysystems.

Previous solubility studies predict that ZnClro is thedominant zinc species at temperatures greater than300oC for solutions in the 0.5 to 5.0 molal range oftotal chloride (Bourcier & Barnes 1987, Ruaya &Seward 1986, Cygan et al. 1994). The trend towardneutral-charge species is in agreement with a decreasein the dielectric constant of water with increasingtemperature (Brimhall & Crerar 1987). However, forhigh-salinity fluids, such as in tlte present fluid inclu-

7A

o/o ,o

@

log mCl-Frc. 9. The calculated preponderance of zinc species (ZnCl2o and ZnCl42-) in hydrochloric

acid media at275 and 350oC (afterRuava & Seward 1986).

CHLOROZINC COMPLEX IN FLUID INCLUSIONS : XAFS STUDY 521

sions, models developed for relatively dilute aqueoussolutions may not be valid, and very few experimentalstudies of solubility have been conducted at totalchloride concentrations above 6 molal ('Wood 1995).

The strucrure of aqueous chlorozinc complexes atroom temperature has previously been investigatedusing XAFS, X-ray scanering, Raman and infraredmeasurements . Table 2 lists calculated and measuredbondJengths for various aqueous zinc species at roomtemperature. Although the tetrahedral zinc complex isdominant up to 430oC, theZn-{) bond lengths show adecrease of nearly 27o compared to the length at 25"C(Table 1). This real effect is evidenced in part byincluding the third- and fourth-order terms (C3 and C4,respectively) in the cumulant expansion in the fittingof the spectra, which may account for any anisotropy oranharmonicity in the radial distribution of the Cl ligandions around the cenhal zinc ion. It was found that theseare very small; they were included in the fitting onlyfor the two highest temperatures (i.e., C3 = 3.8 x 10rA3 and C4 = 2.9 x 10{ A at 430"C). Simple calcula-tions show that the increase in internal equilibriumpressure of the fluid inclusion can only account for I or2Vo of the observed contraction in bond length withtemperarure.

Bond contraction may be associated with theprogressive reduction of the dielectric constant ofthe fluid with increasing temperature. The combinationof reduced hydrogen bonding and greater thermalagitation of water molecules causes a concomitantreduction of the solvent electric potential with tempera-ture at the complex site. This reduction of the solventpotential should induce greater Zn-Cl electron orbitalinteraction at the complex site, which in turn wouldresult in stronger and slightly contracted Zn-{l bonds.The observed contraction ofthe bond and its effect onthe stability of chloride complexes are the subjects of aparallel study that is currently in progress (Mayanovicet al., in prep.).

Figure 9 shows the calculated distribution of ZnClroand ZnCloz in hydrochloric acid media at 275 and350'C (after Ruaya & Seward 1986). At 350oC, Ruaya& Seward extrapolated t}te curves to Cl concentrationsof l0 molal. In contrast to our XAFS results presentedhere, the extrapolated curves predict that ZnClro isthe dominant species at the chloride concentrationsestimated in the inclusion brine (ca. 9 molal). It will benecessary to perform XAFS measurements on syntheticsolutions having high Cl to Zn ratios at elevatedtemperatures in order to determine if the complexcomposition of the inclusion brine is responsible forthe apparent enhanced stability of the tetrahedral zincchlorocomplex.

The XAFS results indicate that a tetrahedralchlorozinc complex was the dominant species at thetemperature of the fluid entrapment in the Knaumiihlepegmatite. Fluid-inclusion and wallrock-alterationstudies of the associated polymetallic sulfide-bearing

quartz veins in the region are currently in progressto determine the factors controlling the destabilizationofthe tetrahedral chlorozinc complex and the precipi-tation of sphalerite.

Minimum detection-limits for XAFS analysis of fluidinclusions

The minimum detection-limit for a particularelement within a fluid inclusion using XAFS is stronglycontrolled by sample parameters such as inclusionvolume and its depth beneath the polished surface ofthe host mineral. Therefore, judicious choice of a large,near-surface fluid inclusion is required to optimizethe X-ray fluorescence yield. At high temperatures, theXAFS spectra may be slightly dampened by atomicdisplacements due to thermal disorder (Crozier 1988).Nevertheless, this additional noise did not adverselyaffect data analyses in the present study. The sensitivityof the technique is also dependent on the atomicnumber of the absorbing ion. Relatively low-energyX-rays generated from light elements such as Na, Mg.K are severely attenuated by the host mineral, and arethus difficult to monitor. At present, the minimumdetection-limit for XAFS analysis of fluid inclusionsusing the X-ray microprobe (X26A) at the NationalSynchrotron Light Source precludes investigation of awide range of solutes that typically occur in lowconcentrations (<100 ppm) in hydrothermal solutions.We anticipate lower detectionlimits by using brighterX-ray sources than are now available.

CoNcr-usIoNs

A full understanding of the macroscopic behavior ofaqueous species in hydrothermal solutions can only beobtained from a detailed knowledge of their atomisticnature. We report the first XAFS study of metalcomplexing at elevated temperatures in natural fluidi nclusions, before and aft er experimental re-equilibrationat high fugacities of hydrogen. Our results indicate thatthe tetrahedral chlorozinc complex, ZnCloz-, it ,6"dominant aqueous zinc species at temperatures up to430"C, and that there is a significant decrease in thelength of the Zn-Cl bond with increasing temperature.These results suggest that zinc was transported asa tetrahedral chlorocomplex in the hydrothermalsolutions that seem involved in the deposition ofsulfide-bearing quartz veins in the region. The presenceof the tehahedral chlorozinc species in fluid inclusionsthat were experimentally re-equilibrated at high hydrogenpressures indicates that natural re-equilibration of fl Hr)most probably did not influence the chlorozinc speciesin the fluid inclusions studied. XAFS analysis ofnatural fluid inclusions, such as those used in thisstudy, provides the most direct evidence of metalcomplexing in paleohydrothermal solutions.

522 THE CANADIAN MINERALOGIST

AcKNowLEDGEMET.]TS

We are indebted to I-Ming Chou (U.S. GeologicalSurvey) for experimentally re-equilibrating naturalfluid inclusions at elevated pressures of hydrogen.Laser Raman analysis and information on the geologyof the Saxonian Granulite Massif were kindly providedby Rainer Thomas (Geoforschungzentnrm Potsdam).Frank Thomas and Kate Moran (Bedford Institute ofOceanography) are thanked for permission to use theirEnvironmental Scanning Electron Microscope. TracyCail assisted with microthermometric measurements.The manuscript benefitted from reviews by Mike Fleet,Bob Martin, Skip Simmons and Al Falster. This studywas supported by a NSERC research grant to A.J.A.,DOE grant (DE-FGO2-92ER14244) to S.B. and DOEgrant (DE-ACO2-76CHOOOI6) to BrookhavenNational Laboratory. We thank the NSLS staff forproviding synchrotron radiation.

REFERENCES

ANoEnsoN, A.J. (1996): A new oil immersion vessel for opticalspindle stage examination of fluid and solid inclusions./n PACROtrI VI, Sixth biennial Pan-American Conf. onResearch on Fluid Inclusions (Madison, Wisconsin: D.E.Brown & G.S. Hagemann, eds.), Program Abstr.

Cunx, A.H., MA, XrN-PEr, Pnlvnn, G.R.,MacARruun, J.D. & RoBnpBn, E. (1989): Proton-inducedX-ray and gamma-ray emission analysis of unopened fluidinclusions. Econ. GeoI. U, 924-939.

- MevnNovrc, R.A. & Berr, S. (1995): Determinationof the local structure and speciation of zinc in individualhypersaline fluid inclusions by micro-XAFS. Can.Mineral.33, 499-508.

HonN, I., Jacrsow, S. & THovres, R.(1997): Element paflitioning and ore-metal transpon in atwo-phase hydrothermal system: evidence from fluidinclusions in pegmatites and associated polymetallicsulfide veins in the Saxonian Granulite Massif, Germany./z ECROFI XIV, European Current Research on FluidInclusions (Nancy, France), Abstr.

Bannrs, H.L. (1979): Solubilities of ore minerals. InGeochemistry of Hydrothermal Ore Deposits (2nd edition;H.L. Barnes, ed.). Wiley-Interscience, New York, N.y.@c/-460]|.

BoDNAR, R.J. & Vnyr, M.O. (1994): Determination of fluidinclusion compositions. /n Fluid Inclusions in Minerals:Methods and Applications (B. De Vivo &M.L.Frezzotti,eds.). Short Course of the IMA working group "Inclusionsin Minerals" (Pontignano - Siena), 117-130.

BouRcrER, W.L. & BARNEs, H.L. (1987): Ore solurionchemistry. Vtr. Stabiliries of chloride and bisulfidecomplexes of zinc to 350"C. Econ. Geol.82,1 839-1863.

BRIMHALL, G.H. & Cnrnnn, D.A. (1987): Ore fluids:magmatic to supergene. /z Thermodynamic Modeling of

Geological Materials: Minerals, Fluids and Melts (I.S.E.Carmichael & H.P. Eugster, eds.). Reu. Mineral. 17,235-321.

BITBACK, M. (1981): Spectroscopic investigations of fluids. /nChemistry and Geochemistry of Solutions at HighTemperatures and Pressures (F.E. Wickman & D.T.Rickard, eds.). Pftys. Chem. Earth 13-14, 345-356.

CeptpnBr-r-, A.R. (1995): The evolution of a magmatic fluid: acase history from the Capitan Mountains, New Mexico. /nMagmas, Fluids, and Ore Deposits (J.F.H. Thompson,ed.). Mineral. Assoc. Can., Shart-Course VoL73, 139-152.

CrnrvY, P. & HAwTHoRNE, F.C. (1982): Selected peraluminousmine.rals. In Granitic Pegmatites in Science and Industry(P. Cernj, ed.). Mineral. Assoc. Can., Short-CourseHandbookS, 163-186.

CHou, I-MINc (1987): Phase relations in the system NaCl-KCI-HrO. III. Solubilities of halite in vapor-saturatedliquids above 445"C and redetermination of phaseequilibrium properties in the system NaCl-H2O to1000"C and 1500 bars. Geochim. Cosmochim. Acta 51.t965-1975.

Clnw, J.S. & BooNan, R.J. (1991): Can economic porphyrycopper mineralization be generated by a "typical"calc-alkaline melt'! J. Geophys. Res. 96,8 I l3-8126.

Cnrnen, D., Wooo, S.A., BRANTLEv, S. & Bocnnsly, A.(1985): Chemical controls on solubility of ore-formingminerals in hydrothermal solutions. Can. Mineral. 23,333-352.

Cnozrnn, E.D., REFR., J.J. & INGALH, R. (1988): Amorphousand liquid systems. In X-ray Absorption, Principles,Applications, Techniques of EXAFS, SEXAFS andXANES (D.C. Koningsberger & R. Prins, eds.). JohnMley & Sons, New York.

CycAN, G.L., Ibvrev, J.J. & D'ANcrlo, W.M. (1994): Anexperimental study ofzinc chloride speciation from 3@ to600oC and 0.5 to 2.0 kbar in buffered hydrothermalsolutions. Geochim- Cosmochim. Acta 58, 4841 4855.

DIll, H. (1985): Antimoniferous mineralization from theMid-European Saxothuringian Zone: mineralogy, geology,geochemistry and ensialic origin. Geol. RundschauT4,M7-466.

Dnnren, P. & Rnse, P. (1986): EXAFS-study of the Znz*coordination in aqueous halide solutions. J. Physique C8,809-8 1 2.

EeoNctoN, P.J. (1983): Afluid inclusion investigation of oreformation in a tin-mineralized granite, New England, NewSouth Wales. Econ. Geol. 78, 1204-1221.

EucsrER, H.P. (1985): Granites and hydrothermal ore deposits:a geochemical framework. Mineral. Mag. 49, 7 -23.

FRANCK, E.U. (1973): Concentrated elecrolyre solutions athigh temperatures and pressures. J. Solution Chem. 2,339-3s6.

CHLOROZINC COMPLEX IN FLIID INCLUSIONS : XAFS STUDY 523

Fnarxr, W. (1993): The Saxonian granulites: a metamorphiccore complex? Geol. Rundschau 82, 505-515.

FuLroN, J.L., PFUND, D.M., Wellru, S.L., Newvnm, M."SrERN E.A. & Mn, YaNluN (1996): Rubidium ionhydration in ambient and supercritical water. J. Chem.Phys. 105" 2161-2166.

Herl, D.L., STERNER, S.M. & BooNar, R.J. (1989): Experi-mental evidence for hydrogen diffusion into fluidinclusions in quartz. GeoI. Soc. Am., Abstn Program2l,A358.

Hemles, F.M., STERNER, S.M. & BoDNAR, R.J. (1988):Synthetic fluid inclusions in natural quartz. IV. Chemicalanalyses of fluid inclusions by SEM/EDA: evaluation ofmethod. Geochim. Cosmochim. Acta 52,969-977.

HuNnrcH, C.A. & RvAN, C.G. (1992): Mineral paragenesisand regional zonation of granite-related Sn-As{u-Pb-Zn deposits: a chemical model for the Mole Granitedistrict (Australia) based on PIXE fluid inclusion analyses.InProc.Tthlnt. Symp. on Water-Rock Interaction 2 (Y.K.Kharaka & A.S. Maest, eds.). Balkema, Rotterdam, TheNetherlands (1583- 1587).

WAr-srD, J.L. & llannorp, B.P. (1996): Chemicalmass transfer modelling of ore-forming hydrothermalsystenrs: current practise and problems. ore Geology Rev.10 ,319 -338 .

Hncrsox, H.C. (1964): Complexing and Hydrothermal OreDeposition. Pergamon, New York, N.Y.

Hnz, G.R., CHARNocK, J.M., VAUGHAN, D.J. & GanNen,C.D. (1993): Multinuclearity of aqueous copper and zincbisulfide complexes: an EXAFS investigation. Geochim.Cosmochim. Acta 57, 15-25.

HEMLEy, J.J., CycnN, G.L., FED{, J.8., RoBNsoN, G.R. &D'ANGELo, W.M. (1992): Hydrothermal ore-formingprocesses in the light of studies in rock-buffered systems.I. Iron-copper-zinc-lead sulfide solubility relations.Econ. Geol.87, I-22.

Jarws, R.H. (1982): Intemal evolution of pegmatite bodies. /nGranitic Pegmatites in Science and Industry @. Cern!,ed.). Mineral. Assoc. Can., Short-Course HandbookS,293-327.

KRUH, R.F. & STANDLEv, C.L. (1962): An X-ray diffractionstudy of aqueous zinc chloride solutions. Inorg. Chem. I,94r-943.

Lncacnr, M., DuroN, S.-C. & SEBASTTAN, A. (1995):Assemblages of Li-Cs pegmatite minerals in equilibriumwith a fluid from their primary crystallization until theirhydrothermal alteration : an experimental study. M ine ral.Petrol.55, l3l-143.

LoNDoN, D. (1986): Magmatic-hydrothermal transition inthe Tanco rare element pegmatite: evidence from fluidinclusions and phase-equilibrium experiments. Arz.Mineral.7l,376-395.

Masoa, M., Iro, I., Honl, M. & Jouaxssou, G. (1996): Thestructure of zinc-chloride complexes in aqueous solution.Z. Naturforschung Section A-A Journal of PhysicalSciences. 51,63-70.

MAvRocENES, J.A. & BonneR, R.J. (1994): Hydrogenmovement into and out of fluid inclusions in quartz:experimental evidence and geologic implications.Geochim. Cosmochim Acta 58, 141- 148.

MAyANovIC, R.A. & Ar'nsnsoN, A.J. (1997): Micro-beamXAFS studies on fluid inclusions under hydrothermalconditions.1z ECROFI XIV, European Current Researchon Fluid Inclusions (Nancy, France), Abstr. (203-204).

& BArr, S. (1996): Microbeam XAFSinvestigations on fluid inclusions. 1n Applications ofSynchrotron Radiation Techniques to Materials Sciencem G.J. Terminello, S.M. Mini, H. Ade & D.L. Perry,eds.). Materials Res. Soc., Symp. Proc. Vol. 437 ,201-206.

MonceN, G.B., VI, Csou, I-Mnqc, Prrstents, J.D. & OtsEN,S.N. (1993): Re-equilibration of CO, fluid inclusions atcontrolled hydrogen fugacities. J. Metamorphic Geol. ll,t55-164.

Musrnr DE LEoN, J., REHR, J.J., ZABINSKY, S.I. & ALBERS,R.C. (1991): AD initio cwved-wave X-ray-absorption finestructure. Phys. Rev. B M, 4146-4156.

NEUMANN, W. & TtsclnNDonn, G. (1986): Genesis of sulfidemineralization in pyriclasites of Hartmannsdorf, GranuliteMassif. /z Proc. Conf. Metallogeny of the Precambrian (2.

Pouba, ed.). Int. Geol. Correlation Program, Project9l,8 I -84.

Nunsp, R.E.G. (1993): Paragenetische und thermo'barometrische Untersuchungen an strukturgebundenenMineralisationen im Siichsischen Granulitgebirge.Dissertation, Freiberg University of Mining andTechnology, Freiberg, Germany.

Oprrnns, E.H. & HercpsoN, H.C. (1993): Multiple ionassociation in supercritical aqueous solutions of singleelecfolytes. Science 261, 888-89 l.

Onrerr, H., YavecucHt, T. & Meroe, M. (1976): X-raydiffraction studies of the structures of hydrated divalenttransition-metal ions in aqueous solutions. BUII. Chem.Soc. Japan 49,701-708.

PARcnMENT, O.G., VINcENT, M.A. & HrLr-nn, I.H. (1996):

Speciation in aqueous zinc chloride. An ab initio hybridmicrosolvation/continuum approach. J. Phys. Chem. 100,9689-9693.

PruNp, D.M., Denas, J.G., FulroN, J.[-. & Ma, YANntN(1994): An XAFS study of strontium ions and krypton insupercritical water. J. Phys. Chem. 98, 13102-13107 .

RrvERs, M.L., SurroN, S.R. & JoNes, K.W. (1991):Synchrotron X-ray fluorescence microscopy. Synchn Rad.News 4.23-26.

524

RoEDDER, E. (1984): Fluid inclusions. Rev. Mineral. 12.

(1992): Fluid inclusion evidence for immiscibiliryin magmatic differentiation. Geochim. Cosmochim. Acra56,5-20.

RuAyA, J.R. & Srwano, T.M. (1986): The srability ofchlorozinc(II) complexes in hydrothermal solutions upto 350'C. Geochim. Cosmochim. Acta 50,651-66I.

Sevsns, D.E. & BuNrcn, B.A. (1988): Data analysis. In X-rayAbsorption, Principles, Applications, Techniques ofEXAFS, SEXAFS and XANES (D.C. Koningsberger& R. Prins, eds.). John Wiley & Sons, New York, N.Y.

Sewano, T.M. (1981): Metal complex formarion in aqueoussolution at elevated temperatures and pressures. 1nChemistry and Geochemistry of Solutions ar High Tem-peratues and Pressures (F.E. Wickman & D.T. Rickard,eds.). Plzys. Chem. Eanh 13-14, ll3-129.

(1984): The formation of lead(Il) chloride com-plexes to 300oC: a spectrophotometric study. Geochim.Cosmochim. Acta 48, 121-134.

HENDERSoN, C.M.B., CHaruvocx, J.M. & DonsoN,B.R. (1997): An X-ray absorption (EXAFS) specrroscopicstudy of aquated Ag+ in hydrothermal solutions to 350oC.G e oc him. Cosmochim. Acta 60, 227 3 -2282.

SrrnNEn, S.M., Harr-, D.L. & Keerr-nn, H. (19951:Compositional re-equilibration of fluid inclusions inquafa. contrib. Mineral. Petrol. ll9,l-15.

Suser, N.J. & Cnnnan, D.A. (1985): Specrra and coordinarionchanges of transition metals in hydrothermal solutions:implications for ore genesis. Geochim. Cosmochim. Acta49, sss-564.

TossrLl, J.A. (1990): Ab initio calculation of the structures,Raman frequencies and Zn NMR spectra of tetrahedralcomplexes of Znz*. Chem. Phys. Leu. 169, 145-149.

& VnucHu, D.J. ( 1992): TheoreticaL Geochemis-try: Applications of Quantum Mechanics in the Eanh andMineral Sciences. Oxford University Press, New York,N.Y.

& - (1993): Bisulfide complexes of zincand cadmium in aqueous solution: calculation of structure,stability, vibrational, and NMR spectra, and speciationon mineral surfaces. Geochim. Cosmochim. Acta 57.193s-1945.

War-rrmn, J.V. & ScHorr, J. (1988): The dielectric constantapproach to speciation and ion pairing at high temperatureand pressure. Nature 332,635-638.

WnrtNny, J.A., HEMLEy, J.J. & Slwrow, F.O. (1985): Theconcentration of iron chloride solutions equilibrated withsynthetic granitic compositions: the sulfur-free system.Ec on. G eo l. 80, 4 4 1 -460.

Woop, S.A. (1995): Solubility constraints on the size ofhydrothermal ore deposits. In Proc. Second Giant OreDeposits Workshop, Controls on the Scale of OrogenicMagmatic-Hydrothermal Mineralization (A.H. Clark,ed.). QminEx Associates, Kingston, Ontario (262-299).

YeuecucHr, T., Heyessr, S & OHre<r, H. (1989): X-raydiffraction and Raman studies of zinc(II) chloride hydratemelts, ZnClr.HrO (R = 1.8, 2.5,3.0,4.0, and 6.2). J. Phys.Chem. 93, 2620-2625.

ZasrNsry, S.1., RBsn, J.J., ANKUDTNoV, A., Alsens, R.C. &ELLsn, M.J. (1995): Multiple-scattering calculations ofX-ray absorption spectra. P hy s. Rev. B.52, 299 5 -3009.

Received September 5, 1996, revised, manuscript acceptedFebruary 2, 1998.