fluid infiltration

American Mineralogist. Volume 78. pages 804-818, i,993

Fluid infiltration through metachert layers at the contact aureole of theBufa del Diente intrusion, northeast Mexico: Implications for wollastonite

formation and fluid immiscibilitv

Wrr,rrrcr,nr HnrNnrcrrInstitut fiir Angewandte Geophysik, Petrologie und Lagerstiittenforschung, Technische Universitiit Berlin,

EB 310, StraBe des 17. Juni 135, 1000 Berlin 12, Germany

AssrRAcr

Cretaceous sediments were contact metamorphosed by the Tertiary Bufa del Dientealkali syenite. Within the massive impure limestones, an aureole developed with a trem-olite, diopside, and wollastonite zone toward the contact. Numerous chert bands in thelimestones extend several hundred meters from the contact. Broad wollastonite rims occuraround the metachert layers into the tremolite zone. Isobaric (P = 1000 bars) univariantand invariant parageneses indicate that the fluid phase in the marbles was internally buf-fered and that the marbles were not subject to substantial aqueous fluid infiltration. Incontrast, the metachert beds acted as channelways for HrO-rich fluids under lithostaticpressure conditions. Fluid inclusions in recrystallized metachert quartz were studied fromsamples near the wollastonite isograd. Early highly saline brines tZr(V - L) : 500-580'C, ?"- (salt) : 400 "C, (NaCl + KCI)"' : 67 -73 wto/o, K,/Na ion ratios : 1-1.81 occur alongthe same microfractures as apparently one- or two-phase vapor-rich COr-bearing inclu-sions. The two fluids were simultaneously trapped in the two-fluid immiscibility field ofthe H,O-CO,-salt system at peak metamorphic conditions of 500-580 "C. Graphite-con-taining metacherts show similar features with highly saline brines coexisting with vapor-rich CHn-bearing inclusions. Late brine populations have f,(V - L) = Z* (salt) rangingfrom 400 down to 150'C. The K,/Na ion ratios of the brines decrease from 1.8 to 0.3 as?".ff ' L) decreases from 580 to 150'C. Trapping of brines channeled in the metachertsoccurred continually from peak metamorphic conditions throughout the cooling history.Their compositions were controlled by mineral-fluid equilibria in the cooling alkali syenitereservoir. The low density COr-rich fluids were generated by the reaction calcite + quarlz: wollastonite * CO, at the two-fluid solvus during flow of magmatic brine away fromthe heat source and probably escaped into the marble by grain-edge flow. The boundarybetween marble and wollastonite acted as a semipermeable membrane separating two flowpatterns in the aureole.

INrnooucrroN composition of fluids evolved during metamorphism. Theamount of infiltrated fluid can be determined by com-

Contrasting fluid infiltration patterns through adjacent paring the composition of evolved fluids with the valuestratigraphic units in both regional-metamorphic and of X"o, at equilibrium determined from mineral assem-contact-metamorphicr6gimes(e.g.,Hover-Granathetal., blages (Ferry, 1983, 1986, 1991). The results are time-1983; Nabelek et al., 1984; Ferry, 1987, 1988; Labotka integrated ratios of fluid to rock that can serve as fossilet al., 1988) showed channeling of fluid flow along indi- flux meters (Ferr),, I 987). Models linking estimated fluidvidual lithologic layers. In carbonate-rich metasedi- to rock ratios with heat and transport equations in porousments, calcareous argillite beds and quartz-rich layers can media have been presented by Labotka et al. (1988) andact as aquifers, whereas impure limestones may be im- Baumgartner and Ferry (1991). An important aspect ispermeable. In impervious units, equilibrium fluid com- the direction of fluid flow. Labotka et al. (1988) suggestedpositions are internally buffered (Rice, 1977; Rice and thatdecarbonationofperviousmetaargillitesattheNotchFerry, 1982) by prograde decarbonation and dehydration Peak aureole was driven by infiltrating fluids emanatingreactions and reflect the stoichiometric ratios of the vol- from the intrusion. In contrast, Baumgartner and Ferryatile components evolved by the rocks. In contrast, there (1991) argued that mineral parageneses in equilibriumis a complex interplay between the compositions and vol- with infiltrating fluid are devolatilized by fluids flowingumes of evolved and infiltrated fluids in permeable beds toward the heat source or toward the surface, whereas(e.g., Rumble et aI., 1982i Ferry, 1987; Labotka et al., flow away from the heat source would result in hydration1988). Mass balance calculations give the amount and and carbonation of the rocks. Their model does not per-0003-004x/93l0708-0804$02.00 804

HEINRICH: FLUID INFILTRATION IN METACHERT 805

mit decarbonation of calcareous beds by hot magmaticfluids unless decarbonation is not accompanied by a si-multaneous pressure drop along with fluid flow. Conse-quently, Ferry and Dipple (1992) used the data of La-botka et al. (1988) to show flow paths in the oppositedirection in the Notch Peak aureole.

The concept of fluid-rock ratios, calculated by the mea-sured progress of devolatilization reactions is based ontwo strict assumptions: (l) local mineral and mineral-fluid equilibrium occuffed everywhere in the flow system(Baumgartner and Ferry, l99l); (2) the corresponding de-volatilization reactions proceeded when the coexistingHrO-CO, fluid phase was homogeneous. Fluid inclusionstudies in both contact-metamorphic and regional-meta-morphic calcareous rocks (Sisson et al., l98l; Tromms-dorf et al.. 1985: Williams-Jones and Ferreira. 1989)confirmed that HrO-CO, immiscibility can be caused atmetamorphic conditions by the presence of dissolved saltsand that decarbonation produced COr-rich fluid thatmight have escaped (Trommsdorffand Skippen, 1986).Bowers and Helgeson (1983a) and Trommsdorff andSkippen (1986) demonstrated the intense bearing on thedevelopment of prograde mineral reactions when the flu-id-rock buffer system doesn't function because of CO,loss by effervescence.

This contribution presents a study of fluid infiltrationthrough contact-metamorphic metachert layers at the Bufadel Diente intrusion, northeast Mexico, using petrograph-ical and fluid inclusion observations. It will demonstratethat (1) fluid influx occurred in the direction ofdecreasingtemperature along metachert layers, (2) local equilibriumwas generally not maintained along metachert-marbleboundaries, and (3) decarbonation proceeded in the two-fluid region of the H,O-COr-salt system. Thus, boiling isdriven by external fluid influx during pluton emplace-ment. Fluid flux in marbles is estimated by calculatingthe amount of low densitv CO"-rich fluid lost into over-lying marble.

Gnor,ocrca,L sETTING

The Bufa del Diente intrusion is part of the TertiarySierra de San Carlos alkaline igneous complex (Hubber-ten, 1985) located in Tamaulipas, northeast Mexico, about180 km southeast from the city of Monterrey, Nuevok6n, Mexico. The geology of the Bufa del Diente areahas been described by Ramirez-Fernandez and Heinrich(1991). The Bufa del Diente stock is an ovoid-shapedalkali syenite that discordantly intruded undeformed Ear-ly to Late Cretaceous marine sedimentary rocks (Fig. l).The strike ofthe beds is parallel to the strike ofthe con-tact, resulting in a dome structure. The sediments to thesouth of the stock, where the dip is shallower than to thenorth, were brought to a higher level, reflecting asym-metric magma emplacement. Outcrops in the area areperfectly exposed. Beds can be traced along the strike forseveral kilometers.

The Early Cretaceous Tamaulipas inferior and Tamau-lipas superior Formations (Fig. 1) consist of thick-bedded

21.31 '

Fig. l. Generalized geologic map of the Bufa del Diente area,simplified from Ramirez-Fernandez and Heinrich (1991), show-ing the Tertiary Bufa del Diente alkali syenite, the Upper Cre-taceous Cuesta del Cura and Agua Nueva Formations, and theunderlying Lower Cretaceous Tamaulipas superior and inferiorFormations. Tremolite, diopside, and wollastonite isograds weremapped as occurring in the massive impure marbles. Black dotsrepresent samples from interstratified metacherts and meta-ar-gillites from Cuesta del Cura and Agua Nueva Formations.Numbers refer to sample numbers in the text.

micritic limestones. Interstratified calcareous argillites arerare and confined to the upper part ofTamaulipas supe-rior Formation. Argillite-rich material is also present instylolites. The overlying Late Cretaceous Cuesta del CuraFormation consists of limestones with chert nodules andnumerous interstratified chert bands. Calcareous argillitebeds occur in minor quantities. North of the aureole, in-dividual cheft strata arc 4-15 cm thick. These can betraced for several hundred meters directly to their contactwith the intrusion. At outcrops 8502 and 8503 there aresome 50 individual chert beds across the Cuesta del CuraFormation, which is about 70 m thick. In the overlyingAgua Nueva Formation, the number and thickness ofcalcareous argillite beds increase considerably. Towardthe top of this formation, limestone and argillite layershave a similar thickness of about 30 cm.

This contribution focuses on the fluid evolution in thelimestones as well as on metachert layers of the Cuestadel Cura Formation in a profile north of the Bufa delDiente stock (Fig. l). The contact-metamorphic evolu-tion of the calcareous argillites is considered elsewhere(Heinrich and Gottschalk, in preparation).

MrNsnAl, CoNTENT oF LIMESToNES

The progressive contact metamorphism of the impurelimestones results in the successive appearance of trem-olite, diopside, and wollastonite with decreasing distanceto the contact (Fig. l), resulting in four zones: calcite *dolomite + quartz (nonmetamorphic); calcite + tremo-lite + quartz (tremolite zone); calcite + diopside + quartz(diopside zone); and calcite + diopside + wollastonite(wollastonite zone). Since the limestones are relativelypure and dolomite deficient, calc-silicates rarely exceed

O 1 k m 2trenolite isoqrad

806 HEINRICH: FLUID INFILTRATION IN METACHERT

Tasle 1. Representative microprobe analyses of minerals from metamorphosed impure limestones (columns 1-6) and metachertlayers (7-9) from Cuesta del Cura Formation

Metacherts

MineralZoneNO

TrTr

819a7

frTr

81282

PhlFr

B1 261

DiDi

814-3

DiWo

811-3

KfsDi

8 1 1 0 4

H I

B502-75

Di in Ri Pc

8502 8502-211

sio,Tio,Al,o3V,O.FerO3FeOMnOMgoCaONaroKrOF

Total

Sir4tAlr6lAlTi

Fe3 *Fe, *MnMg

NaK

Total

56720 1 92.820.000.001 1 3o.27

22.48t J . o I

0280.230.260 . 1 1

97.882307 7470.2530 2000 020

0 1300.0314.5771.9910.o740 040

15.063

57.840 1 10 8 90 0 00.001.560.03

23.441 3 . 1 10.640.050.170.07

97.772307.9000.1 000.0430.011

0.1 780.0034.7731.9170.1 690 009

1 5.1 03

41 .990 . 1 9

12.480.000.002.880.06

26.440 . 1 60.918.980.090.04

94.141 1 02.9801.0200.o240.010

0.1710.0042.7980.0120.1250.813T . J T J

51.030.132.O70.000.00

1 1 . 5 41.209.03

24.590.280.000.000.00

99.876 01.9540.0460.0480.004

0.3700.0390.5161.0090.020

4.006

53.930.091.230.000.002.930.51

15.9725.430.350.000.000.00

100.446 01.9680.0320.0210.002

0.0900.0160.8690.9940.o24

4.016

65 480.00

18.440.000.000 0 00.000.000.000.83

14.990.000.00

99.748 03.012

1.000

58.1 I0 .120.010.080.002.320.31

22.237.814.203.633.91| .oc

1 0 1 . 1 52308.026

0.0010.0130.009

0.2680.0364.5711.1531.1240.638

15.839

55 030.080.072.250 0 04.010.07

14 9623.521 2 30.000.000.00

101 .216 02.000

0 0030.0020.066

0.1220.0020.8100.9160.086

4.007

53.060.00o.320.000.000.000.000.00

34.098.810.010.000.00

96.298.5 02.9720.021

2.O450.9570 0015.996

0.o740.8804 966

Note: Ri : richterite: Pc : Dectolite.. Cations on the basis of O atoms.

100/o of the rocks' mode. Talc- and forsterite-bearing as-semblages were not detected. Alkali feldspar and phlog-opite are occasionally present. Representative mineralanalyses are given in Table l. Within the mineral zones,isobaric univariant assemblages in the CaO-MgO-SiOr-COr-HrO system are common and the isobaric invariantassemblage dolomite + calcite + quartz * tremolite +diopside was detected at several locations close to thediopside isograd.

The asymmetry of the magma emplacement is reflectedin the greater width of the tremolite and diopside zonessouth of the intrusion. In the north, the tremolite anddiopside isograds are closer to the contact than in thesouth, reflecting a steeper dip in the north (Ramirez-Fer-nandez and Heinrich, 1991).

MrNnn-lL coNTENT oF METACHERTS

The progressive contact metamorphism of the chertlayers of the Cuesta del Cura Formation resulted in theextensive formation of wollastonite according to the re-action calcite + quartz: wollastonite * CO, along thechert-marble boundaries. Wollastonite rims along chertlayers occur at the contact with the intrusion and extendfar into the diopside zone. Some wollastonite rims werefound in the tremolite zone, al a distance of >400 m fromthe contact. Wollastonite either mav form as rims on both

sides of the chert-marble boundary or may separate thechert band into rounded quartzite nodules (Fig. 2A). Thethickness of the rims conelates roughly with decreasingdistance from the contact. It is generally not related tothe thickness of the metachert bands. Typically, in thewollastonite zone, chert layers of 4-6 cm thickness arecompletely reacted, resulting in pure wollastonite bands.In adjacent layers, 10-15 cm thick, thin quartzite bandsor nodules remained. However, there are exceptions tothis, as rare metacherts with thin wollastonite rims nearthe contact and others with thick rims exist in the diop-side zone. A rough estimate based on the rim thicknessofsome 50 individual layers gives an average ofabout 3cm3 wollastonite to I cm2 metachert layer at the wollas-tonite isograd, 2 cm3/cm'? at the central diopside zone,and 0.3 cm3/cm2 near the diopside isograd (outcrops 8502,Al7 ,B'4, respectively, in Fig. l).

Thin pegmatitic veins are common and can be tracedup to 100 m from the contact to the wollastonite isograd.They are confined exclusively to the metachert layers andare strictly conformable with bedding. Moreover, peg-matitic veins only occur along one of the two quartz-wollastonite boundaries (Fig. 2B), indicating that veincrystallization took place after wollastonite formation.Pegmatitic veins were not observed along wollastonite-marble boundaries. It is emphasized that vertical peg-


Frg.2. (A) Wollastonite-rimmed metachert band embeddedin diopside-zone marble from Agua Nueva Formation (SampleAl7). (B) Wollastonite-rimmed (wo) metachert band (ch) em-bedded in marble (m) from the boundary between the diopsideand wollastonite zones, Cuesta del Cura Formation. A pegma-titic vein (p) consisting of acmite rims (dark), perthitic alkalifeldspars (white), and eudialyte is strictly comformable with bed-ding and separates the upper wollastonite-quartz border along a

distance >60 m from the contact. (Sample B33, graphite-free).(C) Sketch of graphite-free metachert sample 8502 from the di-opside-wollastonite boundary. The wollastonite rim containsminor amounts of diopside (Di), alkali feldspar (Kfs), and albite(Ab). Remaining quartz nodules are recrystallized and associatedwith K-rich richterite (K-Ri), pectolite (Pc), and Kfs. They arethemselves rimmed by calcite (Cc) and Pc. (D) Detail from C:K-rich richterite (Ri) with inclusions of diopside and pectolite.

matitic or hydrothermal vein systems in the Cuesta delCura Formation do not exist. There is no field indicationofany vertical fluid or mass transport by fractures at thislevel ofthe aureole.

The wollastonite rims of the metacherts may includeaccessories such as alkali feldspar, albite, plagioclase(An,r_rr), diopside, titanite, Na-rich chlorine scapolite,vesuvianite, and grandite solid solutions, reflecting the lowargillite content of the chert protolith. These minor phasesform very thin (commonly = I mm) margins at the wol-lastonite-marble boundaries and at most make up a fewpercent of the metacherts' modal content. Thus, the re-action calcite + q\artz: wollastonite * CO, accountsalmost exclusively for the CO, produced in the metachertlayers.

The metacherts far from the contact are rich in organicmaterial. Graphite occurs up to about the wollastoniteisograd. From there to the contact, metacherts arebleached and are generally graphite-free. However, oc-

casionally graphite persists up to the contact. Wollaston-ite rims are free of graphite throughout the profile.Bleaching is accompanied by significant quartz recrystal-lization from grains <0. I mm in the diopside zone tograin sizes around I mm beyond the wollastonite isograd.

Remaining quartz nodules are themselves rimmed bythin pectolite (NaHCarSi.On, Table l) and calcite assem-blages (Fig. 2C). This can be observed at quartzite-wol-lastonite boundaries throughout the profiles. Pectolite alsooccurs inside the quartzite nodules along with smallamounts of diopside and alkali feldspar and traces offluorite (Fig. 2C). Throughout the wollastonite zone, al-kali-rich amphibole solid solutions occur inside the meta-cherts. These amphiboles typically enclose previouslyformed contact-metamorphic diopside (Fig. 2D). Theyare Al-free and F-rich (Table l). Cl was not detected.Amphibole solid solutions have compositions of 60-70molo/o potassium richterit e, 20-40 molo/o richterite, and0-20 molo/o tremolite component.

808 HEINRICH: FLUID INFILTRATION IN METACHERT

t

Fig. 3. Fluid inclusions in recrystallized quartz from meta-cherts from the diopside-wollastonite boundary. (A) Sample 8502.Typical HrO-rich highly saline four-phase fluid inclusions witha big sylvite and a smaller halite crystal exhibiting similar crystalforms and phase relations. (B) Sample 8502. H.O-rich highlysaline four-phase fluid inclusions (arrows) occurring along thesame trail as apparently one-phase vapor-rich COr-bearing in-clusions. (C) Sample B502. HrO-rich highly saline and vapor-rich CO, inclusions (arrows) associated with thin wollastonite

needles growing into quartz crystal. Vapor-rich inclusions revealvarying phase relations indicative of heterogeneous trapping ofthe two coexisting fluids. (D) Methane-rich fluid inclusions ingraphite-containing metachert sample B503. They are apparent-ly monophase (vapor) or two-phase (liquid + vapor) at roomtemperature. Some of these inclusions additionally contain oneor more salt crystals (long arrows), though often a liquid phaseis not visible. HrO-rich highly saline four-phase inclusions aredenoted by short arrows.

t

!

Fr,urn TNCLUSToNS

Petrography

Fluid inclusions were studied in recrystallized quartzitenodules from four samples: B502, 833, 8503, and 8504,all taken from near the boundary between the diopsideand wollastonite zones (Fig. l). Only a few inclusionswere observed in wollastonite from the associated rims.Sample 8502 is graphite-free, B33 is graphite-free andassociated with a pegmatitic vein (Fig. 2B), and 8503 and8504 are graphite-bearing. Quartz recrystallization in thediopside- and tremolite-zone metacherts was minor, andthe fluid inclusions trapped in these zones were too small(-2 pm) for reasonable fluid inclusion measurements.

Samples B502 and B33 are extremely rich in fluid in-clusions, consisting predominantly of highly saline four-phase HrO-rich inclusions with a large sylvite crystal, asmaller halite daughter crystal, and a small vapor bubbleat room temperature (Fig. 3A). The inclusions are <10

pm in diameter. They were trapped mainly along healedmicrofractures and clusters and are considered to be ofsecondary origin. There is no unequivocal evidence ofprimary or pseudosecondary inclusions. Individual inclu-sions along a distinct trail generally exhibit similar neg-ative crystal forms and identical phase relations (Fig. 3A),thus defining a fluid inclusion population. A single qvartzgrain of I mm in size may be cut by up to 20 populations.These differ at room temperature, especially in the char-acteristic sizes of the daughter crystals. Populations withsmaller salt crystals as shown in Figure 3 are common.There are H, O-rich two-phase populations withoutdaughter crystals. However, freezing and heating cyclesshowed that daughter crystals are absent metastably upto salinities of 42 wto/o (NaCl + KCI)"'. This might be aneffect offast heating and cooling during sample prepara-tion (Sterner et al., 1988). In any case, H2o-rich halite-undersaturated fluid inclusions are not present in theserocks. Crosscutting relationships between the distinct


population trails and between trails and quartz grainboundaries suggest a chronological sequence throughoutthe samples. Very highly saline populations formed early,whereas populations with lower salinity developed later.

In samples B502 and 833, a second type, consisting ofvapor-rich COr-bearing inclusions, occurs. These are ex-tremely rare and at room temperature apparently onephase; sometimes a small amount of HrO is visible.Daughter crystals are not present. Sizes are typically 5-l0 pm and do not exceed 12 1rm. These inclusions arealways associated with highly saline HrO-rich inclusions.Figure 38 shows that apparontly one-phase vapor-richCOr-bearing inclusions exist along the same microfrac-tures as highly saline four-phase inclusions. From a tex-tural standpoint these inclusions were simultaneouslytrapped in the two-fluid immiscibility field of the HrO-CO,-salt system (Gehrig, 1980; Bowers and Helgeson,1983b; Trommsdorffand Skippen, 1986). Trapping waslargely homogeneous for both fluid types, which is a rarecase (Ramboz et al., 1982). However, there is also tex-tural evidence for heterogeneous trapping, resulting intrails with variable degrees of filling and also in the for-mation of vapor-rich COr-bearing inclusions containingsalt crystals, which indicate trapping of mixtures of thetwo immiscible fluids in various proportions (Ramboz etal., 1982). This is typically observed in rare fluid inclu-sion clusters that are associated with thin wollastoniteneedles growing into quartz crystals (Fig. 3C). Vapor-richCOr-bearing inclusions are rare, only appearing in about30lo of all highly saline populations. They represent about0.010/o of the total fluid inclusion population.

Fluid inclusions of the graphite-bearing samples 8503and 8504 have HrO-rich highly saline inclusion popula-tions. Likewise, halite-undersaturated HrO-rich inclu-sions do not occur. However, these samples contain manyvapor-rich, apparently one- or two-phase, CHo-bearinginclusions, often forming negative crystals to 40 pm. Theirtextural relationships with the highly saline inclusions arevery similar to those described above, again indicatinghomogeneous as well as heterogeneous trapping of twoimmiscible fluids. Furtherrnore, these inclusions may alsoform separate trails and clusters not associated with high-ly saline inclusions. Some of the vapor-rich CHo-bearinginclusions contain one or more salt crystals. Often noH,O is visible (Fig. 3D). The relative proportion of theCHo-bearing inclusions is about 400/0. COr-bearing inclu-sions were not found in these samples.

Methods

Microthermometric data were obtained on a modifiedLinkam TH600 heating and freezing stage calibrated atthe triple point of CO, (- 56.6 'C), the triple point of HrO(0 'C), the critical point of CO, (3 1.1 "C), and the criticalpoint of HrO (374.1 'C), using synthetic fluid inclusionssynthesized hydrothermally at Fachgebiet Petrologie, TUBerlin. Temperature accuracy is +0.3 'C at -56.6 'C,+0. I 'c at 0 "c , +0.1 "c at 31.1 "c , and +4"c at .374oC, and is believed to be + I "C at - 100'C. Phase changes

TraLe 2, Leachate analyses of bulk fluid inclusion samples

lon concentrations (ppm) lon ratios

Sample Mg l(Na Mg/Na CalNa

2.6 0.53 0.058.6 0.60 0.080 29 0.58 0 135 4 0 .14 0 .140.5 0 15 0 .1 1o.23 0 22 0.09

0.44 0 .180.61 0 .10

850211 7.212 3.513 071

833/1 3.71 2 4 213 2.1

8503/1 0218504/1 3.6

6.5J . O

0.70n o1 . 10.80 . 1 62 8

o.40290 . 1 0u.550.510 .190.040.24

0 2'l1 . 4 1o230.840.060.06

Notei Results are averages ot double determinations.

were repeated at least twice. Reproducibility is within therange of accuracy of the temperature determination.

Partial chemical analyses of fluid inclusion leachateswere obtained by crushing and leaching quartz grains fromthe quartzite nodules. Leachates were analyzed for K*,Na*, Mg+, and Ca2+, using standard AAS and ICP meth-ods. The results presented in Table 2 are averages ofdou-ble determinations.

Spectroscopic analyses of the constituent fluid specieswithin fluid inclusions of samples B503 and 8504 weredone on a Ramanor U 1000 Raman microprobe equippedwith an Ar ion laser (wavelength 514.5 nm).

Results

H,O-rich highly saline inclusions. Results of 600 HrO-rich highly saline fluid inclusions are presented in Figures4 and 5. Data include only distinct inclusion populationswith at least ten individual inclusions exhibiting similarphase relations. Any inclusions suspected ofhaving beenformed by heterogeneous trapping or by necking phe-nomena were excluded.

The salinities in these inclusions were determined us-ing microthermometric data for the vapor-saturated sys-tem NaCl-KCl-H,O (Sterner et al., 1988). Figure 4.A showsthe apparent KCI and NaCl contents of 214 inclusions ofsample B502 as obtained by the final melting tempera-tures of KCI and NaCl in the presence of vapor. Data for219 inclusions of sample B503 and of 8504 cover analmost identical range. Final salt melting temperaturesrange continuously from 120 to 400 'C, resulting in totalsalinities from 33 to 73 wto/o (KCI + NaCl)"o. There is agood correlation between total salinity and the apparentKCI/(KCI + NaCl) ratios (wto/o) with values from 0.69 to0.29 as the total salinity decreases from 73 to 33 wto/0.Recalculated bulk leachate analyses give average KCI/(KCl + NaCl) ratios (wto/o) of 0.41-0.47 (Table 2). A sim-ilar trend is seen in sample B33 (Fie. 4B). NaCl finalmelting temperatures (90-390 "C) are almost identical tothose of the other samples, although the KCI contents aresignificantly lower. KCI/(KCI + NaCl) ratios (wto/o) rangefrom 0.5 to values lower than 0.1 and roughly correlatewith apparent total salinities of 58-=30 wto/o (KCl +NaCl)"o. Average ratios from bulk leachate analyses havevalues from 0. l4 to 0.22 (Table 2). In samples 8502 and8503, the KCI contents at KCy(KCl + NaCD ratios <0.35

Hzo

8 1 0

nydro-halite

NqCt KCI

Fig. 4. Apparent NaCl and KCI contents of H'O-rich four-phase fluid inclusions of samples 8502 (A) and B33 (B) plottedon the vapor-saturated system NaCl-KCl-HrO (Sterner et al.,1988). The thick line denotes the NaCl/KCl ratio calculated frombulk leachate analyses (Table 2). In 8502, 23 inclusions have?"-(salts) > 7"(V - L) and are not plotted in Fig. 5.A. (see Fig.6,4,). For B33, at low KCI contents (shaded area) only NaCl finalmelting temperature is plotted. At this range, dots represent fiveindividual inclusions. Scale bar gives the uncertainty range forKCI contents. For further exolanation see text.

are not well constrained because of metastable processesduring keezing and heating cycles (for details see Sterneret al., 1988). In some inclusions only KCI crystals wereobserved, and sylvite melted in the absence of halite.However, in other inclusions of the same populations,sudden halite precipitation on the dissolving sylvite crys-tal was observed just before sylvite dissolved. It wastherefore concluded that in halite-absent inclusions, ha-lite failed to precipitate when the composition of the liq-uid crossed the cotectic line between the KCI and NaClfield (Fig. 4A) at temperatures lower than about 100 "C.These inclusions have been omitted in Figure 4. In B33,the KCI contents in the weakly saline region are also un-

HEINRICH: FLUID INFILTRATION IN METACHERT

Fig. 5. Last salt melting temperature vs. total homogeniza-tion temperature of HrO-rich highly saline four-phase inclusionsof sample B 502 (A) and B503 (B). At fh ry - L) between 500and 580'C (shaded area) many inclusions decrepitated beforereaching homogenization.

certain and may be very low, since KCI melting temper-atures could not be measured in the presence of hydro-halite or ice (Fig. 4B). Final melting of hydrohalite wasnot observed in the presence of sylvite or ice. Apparentinitial melting in most inclusions occurred at - 50 to -32

"C, indicating that CaCl, or MgCl, were present (Craw-ford. l98l).

Bulk leachate analyses of all samples resulted in signif-icant Mg2+ and Ca2+ contents (Table 2). Because of thepoor reproducibility of replicates, the determined CalNaratios are ambiguous. If contamination by solid calciteinclusions is assumed to account for the high values of8502/2 and 833/1 (Table 2) it would seem that the actualCalNa ratios were low at values of around 0.2 (8502) and0.06 (833). Mg/Na ratios are in the range of 0.1 in bothsamples. Thus, the NaCl and KCI contents estimated fromthe microthermometric data using Figure 4A and 4B are

T6(s)oC

Hzo

HEINRICH: FLUID INFILTRATION IN METACHERT 8 l r

too high (Sterner et al., I 988), and the total salinities wereclassified as apparent. However, it remains unclearwhether the CalNa and Mg/Na ratios remained constantthroughout the populations or whether a correlation withthe total salinities exists like that detected for the K,/Naratios.

During the freezing and heating cycles, no phase tran-sitions could be detected in the vapor bubbles of the HrO-rich highly saline inclusions. Formation and melting ofclathrates could not be observed. The vapor bubbles inthe inclusions of samples 8503 and 8504 consist of pureCHo, as confirmed by Raman spectroscopy.

Homogenization temperatures to liquid cover a widetemperature range from 150 to 573 "C (Fig. 5,A., 5B). Nev-ertheless, the range ofZr(V - L) ofeach population doesnot exceed +10 "C. From 150 to 380 "C, fn(V' L) ofall inclusions is equal to the last salt melting temperature,corresponding to 33-66 wto/o (KCl + NaCl)"o in samplesB502 and 8503. This indicates that the fluid populationsat this temperature range were salt-saturated. There areno inclusions in which 7- (salts) is significantly higherthan Zr(V - L). This suggests that heterogeneous trap-ping of salt-saturated fluids with precipitated salt crystalsdid not occur. This is unlike developments in many otherhydrothermal and contact-metamorphic fluid systems(e.g., Ahmad and Rose, 1980; Williams-Jones and Fer-reira, 1989; Kelley and Robinson, 1990). In samples 8502and 8503, fluid populations with 7"rff - L) from 400 to580 "C were salt-undersaturated and displayed relativelyconstant salinities from 67 to 73 wto/o (NaCl + KCI)"". Atthis 7" range, apparent total salinities are not correlatedwith Z"(V - L). Above 500'C, many inclusions decrepi-tated before homogenization occurred, and none ofthemsurvived the transition from a to B quartz. Apparent KCI/(KCl + NaCl) ratios correlate significantly with the totalhomogenization temperatures. Recalculated apparentK,/Na ion ratios in samples 8502 and 8503 have maxi-mum values of 1.75 and 1.80 at fhff - L) of 500-573oC down to minimum values of about 0.3 at fh: 150'C. This is also true of sample B33, in which the apparentKCI content and the total salinities are lower (Fig. aB)and the apparent K./Na ratios shift to lower maximumvalues of 0.8 at ?"hff - L) : 450-530'C down to <0.13at fhry * L) : 95 'C. It is evident that the K/Na ionratio decreases at least half an order of magnitude in allsamples as 7"(V - L) decreases from maximum to min-imum values.

Vapor-rich COr-bearing inclusions. These inclusions areexclusively associated with highly saline inclusions hav-ing ?"n = 500 'C. Melting of solid CO, could be observedonly in l3 out of 60 inclusions from both B502 and B33.Z-(COr) ranges from -56.3 to -57.1 "C, indicative ofpure COr. The amount of liquid produced near the CO,triple point is estimated to be 10- I 5 vo10/0, which is justabove the optical resolution for a liquid film around avapor bubble. CO, homogenized to the vapor phase. Theapparent f"COr(L - V) range from about -50 to a max-imum value of 0 'C. If the optical resolution for the last

Tn {t-*v *v)

Fig. 6. Histogram illustrating the apparent homogenizationtemperatures to the vapor phase of the methane system of meth-ane-bearing inclusions in graphite-containing metacherts 8503and B504.

visible liquid is assumed to be l0 vol0/0, the true ?n"(L -

V) are higher. Using the curves of Walther ( 198 I ) relatingtemperature and volume, the apparent 7r (L - V) of 0oC corresponds to the true value of + 18", correspondingto the maximum density of 0.18 g,/cm3 for the CO' system(Kerrick and Jacobs, l98l). The mean value of 0.14 +0.04 g/cm3 covers all observed fhCO,(L - V). Theamount of the aqueous phase is estimated to be 5-10volo/o at room temperature. Owing to the small amountsof the liquid, final ice melting in the presence of clathratewas observed in only l7 inclusions. I- ice ranges between-4. I and - 10.5 oC, corresponding to salinities of 6-14.5wto/o NaCl.. (Potter et al., 1978). Clathrate melting wasobserved in a few inclusions, and the corresponding sa-linities are broadly coincident with those derived fromT^ice. H,O-CO, homogenization occurred to vapor. Theapparent ZnHrO-COr(L ' V) were preferentially esti-mated in inclusions with reentrants and range from 350to 460 oC. Due to difficulties in observing the final dis-appearance of liquid, these estimates are likely to be toolow with true Zn G - V) at least 100'C higher. Thus, thehomogenization temperatures of the vapor-rich COr-bearing inclusions broadly coincide with the peak ho-mogenization temperatures of the highly saline inclusionsassociated with them. Estimates for the bulk inclusiondensity depend on the amount of the aqueous phase atthe inclusion walls. On the basis of liquid-vapor ratiosand values of phase densities from Parry (1986) andBrown (1989), the vapor-rich CO,-bearing inclusions havebulk densities of 0.23 + 0.04 g/cm3 and contain between35 and 55 molo/o CO, and l-3 molo/o NaCl.".

Vapor-rich CHo-bearing inclusions. In these inclusions,methane homogenized to vapor between -85 and -90

"C (Fig. 6). The densities for the CHo system are 0.08 t0.02 g/cm3, as derived from Jacobs and Kerrick (1981).

oIt,

ct(,lJ-

(Na,K)Cl

8 1 2 HEINRICH: FLUID INFILTRATION IN METACHERT

HzO CHo

Frg. 7 . (A) Schematic isobaric isothermal section of the sys-tem H,O-CO,-(Na,K)CI at 1000 bars and 550'C (mo10/o) exhib-iting the two-fluid immiscibility field, based on experimentaldata of Frantz et at. (1992). The diagram is tentatively extrap-olated to high salinities, at which the exact position ofthe phaseboundaries is not known. Data for salt saturation on the HrO-rich side are from Stemer et al. (1988). (B) Schematic isothermalsection of the system H,O-CH.-(Na,KCI) at 550 'C and 1000bars, based on experimental work by Krader (1985) and tenta-tively extrapolated to high salinities (see above).

Raman spectrometry showed that the vapor phase is pureCHo. The spectral position of z, (CHo) of some inclusionsis at2914.l cm '. Applying the calibration curve of Chouet al. (1990) for the position ofz, as a function ofpressureresulted in an internal CHo pressure ofabout I l0 bars atroom temperature. Using the equation of state of Jacobsand Kerrick (1981), this corresponds to a CHo density of0.089 g/cm3, well within the range detected by microther-

mometry. Melting of clathrate was not observed. Theamount of aqueous liquid is estimated to be <10 vo10/0.Total homogenization temperatures into the vapor phaseare estimated to range from 250 to 500 "C. The salinityof the aqueous phase is highly variable. In many inclu-sions, such as that shown in Figure 3D, one or more saltcrystals are present. Melting of salt did not occur by heat-ing up to 600 "C, indicating that this fluid was trapped inthe presence of precipitated salt crystals. Bulk densitiesare estimated to be 0. l5 + 0.04 g,/cm', depending chieflyon the estimated amount and salinity of the aqueousphase. With the assumption of 5-10 volo/o HrO, the in-clusions contain between 40 and 65 molo/o CHo.

DrscussroN

Fluid inmiscibility

Textural relationships suggest that highly saline HrO-rich inclusions and vapor-rich COr- and CHo-bearing in-clusions were simultaneously trapped in the solvus of theHrO-COr-salt and HrO-CHo-salt systems. To verify thisinterpretation with microthermometric data one mustdemonstrate that (l) each group displays contrasting ho-mogenization modes (L + V - L;L + V' V), (2) totalhomogenization temperatures are coincident for eachgroup, and (3) compositions and bulk densities fit intoexperimentally determined solvi in the model systems.The homogenization temperatures and pressures of thetwo immiscible fluids then approximate their entrapmentconditions (Pichavant et al., 1982; Ramboz et al., 19821,Bottrell et al., 1990). Given that the vapor-rich CO,-bear-ing inclusions are associated only with highly saline brinesof fn = 500 "C and that the true Zn of COr-rich inclusionsmay exceed the apparent values of 350-460 "C by > 100"C (Walther, 1981), the data clearly meet requirements Iand 2. This supports the textural evidence and indicatesthat in samples 8502 and B33 two immiscible fluids havebeen trapped at temperatures of 500 to at least 573'C atthe wollastonite isograd. There is no evidence fortrapping of vapor-rich COr-bearing inclusions at lowertemperatures after the brine became salt-saturated. Insamples 8503 and 8504 two immiscible fluids in the HrO-CH.-salt ternary have been trapped in similar conditions.In the latter case, estimated homogenization tempera-tures as low as 250'C for the vapor-rich inclusions andthe presence of salt crystals in them indicate that trappingof two coexisting fluids occurred under salt-saturatedconditions.

Experimentally determined data on the solvus of theHrO-COr-NaCl system are available up to salinities ofabout 50 wto/o of the brine (Frantz et al., 1992). In thesystem HrO-CHo-NaCl, data are only available up to 8wto/o NaCl (Krader, 1985). Figure 7 shows schematic iso-baric isothermal sections of these systems at a tempera-ture of 550'C and 1000 bars tentatively extrapolated to(Na,K)Cl contents of 65-73 wto/0. Although the positionof the phase boundaries and the tie lines is not exactlyknown, it is evident that H,O-rich highly saline brines

l 'o* ; ,1I 55ooc I

tr.'-.til]I Na /Na .K 'o .5 I

( N a , K ) C l

HEINRICH: FLUID INFILTRATION IN METACHERT 8 1 3

coexist with low-density COr-rich vapors containing afew mole percent of salts (Pichavant et al., 1982). Theestimated contents in vapor-rich inclusions of between35 and 55 molo/o CO, (8502 and B33) and between 40and 65 molo/o CHo @503 and B504) are compatible withthe schematic phase diagrams. Isochores for vapor-richinclusions were calculated using the PVT data computedby Brown (1989). The estimated bulk densities and com-positions for the CO, inclusions correspond to internalpressures of 550-950 bars at 550 'C. Using correspondingdensities and compositions of the vapor-rich CHo-bear-ing inclusions of samples B503 and 8504 and applyinggraphical interpolations from Krader's ( I 98 5) experimen-tal data, internal pressures of480-990 bars are estimated.Hence, the maximum values of the estimated pressuresin both systems also agree with the interpretation of si-multaneous trapping on the two-fluid solvi. The high in-ternal pressures may account for the frequent decrepita-tion of hypersaline brine inclusions before homogenizationat temperatures above 500 'C. Given that sedimentarycover during the Oligocene was at least 2000 m and notmore than 3000 m (Ramirez-Fernandez and Heinrich,l99l), fluid pressure was nearly lithostatic and certainlynot hydrostatic.

T-X conditions of the formation of wollastonite rims

It has been shown that aqueous fluids trapped in themetacherts are represented by highly saline brines andthat vapor-rich COr-bearing inclusions are associated onlywith brines trapped above 500'C. Weakly saline fluidsare absent. The formation of wollastonite rims musttherefore have taken place at the HrO-rich side of thesolvus for the HrO-CO2-salt system. This occurred attemperatures given by the intersection point of the wol-lastonite-producing reaction with the two-fluid solvus orabove (Fig. 8). At 1000 bars and up to 35 wto/o NaCl,these isobaric invariant points always lie above 500'C atdistinct X.o, values, and it is most likely that this is alsovalid for very high salinities (Fig. 8). If one accepts thetentative extrapolation up to 70 wt0/o (shaded area in Fig.8). that would mean that wollastonite could have formedat temperatures at the solvus of >600 to about 500 "C atXco, of about 0.02. These temperatures coincide with thetemperatures at which the trapping of the two immisciblefluids occurred. This strongly suggests that the immisci-ble vapor-rich COr-bearing fluid was generated in situ bythe reaction calcite + euartz: wollastonite + COr. Thisis also supported by textural observations (Fig. 3C) andrenders an interpretation ofthe CO, fluid as generated byexsolution from the brine during cooling as highly im-probable. Trommsdorffand Skippen (1986) showed thatwhenever immiscible low-density CO.-rich fluid is gen-erated by a decarbonation reaction at the two-fluid sol-vus, it is preferentially removed from the system (see alsoYardley and Bottrell, 1988). The result is vapor loss byboiling (Trommsdorffand Skippen, 1986). Large amountsof wollastonite * CO, can be produced at the solvuswithout increasing the X"", of the brine. HrO is trans-

0.00 0.05 0.10H2O9.9669 NaC19.9331 (10wt%)

H2O9.9259NaClO.O715 (20wt%)

H2O9.g62gNaClO.i 172 (3O wt%)

0.15 0.20

Xcoz *

Fig. 8. Pseudobinary isobaric X.o,-Z sections at 1000 barsand low X.o, illustrating the position of the reaction calcite +qtrartz : wollastonite + COr relative to the two-fluid immisci-bility field at various NaCl contents. The reaction curves werecalculated using mineral data from Ge0-Calc (Brown et al., 1988)along with -f,o, values in the H,O-CO.-NaCI ternary from Bow-ers and Helgeson (1983b). The two fluid boundaries were graph-ically interpolated and extrapolated from Gehrig (1980) andFrantz et al. (1992). Circles denote maximum temperatures forfluid buffering by this reaction. The shaded area is a tentativeextrapolation to salinities of =70 wt0/0.

ferred to a minor extent from the brine into the low-density H,O-CO, fluid, and the salinity of the brine isincreased.

As regards cooling below 500 "C, the fluid compositionshould leave the solvus and should, in principle, be buf-fered by the reaction of wollastonite * CO, to calcite +quartz (Fig. 8). The presence of inner calcite rims betweenthe remaining quafiz nodules and the wollastonite rims(Fig. 2C) indicates that this back reaction occurred local-ly. However, fluid inclusion characteristics indicate thatthe fluid composition was not buffered during cooling andthat it did follow the two-fluid solvus down to low tem-peratures. The concomitant formation of pectolite rimsduring cooling (Fig. 2C) by the reaction wollastonite +HCI + NaCl"o : pectolite * CaCl,

"o might have pre-

vented retrograde fluid buffering.

Brine infiltration along metacherts

Within the marbles, the frequent occurrence of isobaricunivariant and invariant assemblages suggests that themassive limestones formed a closed or nearly closed sys-tem with respect to the fluid phase and that fluid com-position was internally buffered (Rice, 1977; Rice andFerry, 1982), at least in the tremolite and diopside zones.

tos-F

""qD-tx"w


t6 sboe-F

3500.1 0,2 0.3 0,4 0.5 0,6 0.7 0.0 0.s 1.0

xco2 *Fig. 9. Isobaric (P : 1000 bars) temperature-fluid composi-

tion diagram for tremolite-, diopside-, and wollastonite-produc-ing reactions in the massive marbles (Tr : tremolite, Di : di-opside, Wo : wollastonite, Cc : calcite, Do : dolomite, Qz :quartz, Kfs : alkali feldspar, Phl : phlogopite). Reactions in-volving Kfs and Phl contribute only to a minor extent to theformation of the index minerals and fluid evolution, respective-ly. Reaction curves are calculated for pure end-member com-positions of the minerals using GeO-Calc (Brown et al., 1988).The arrows indicate the X.or-T evolution in a closed system.

Figure 9 shows an isobaric T-Xco, section in the systemCaO-MgO-SiO,-H,O-CO, for tremolite-, diopside-, andwollastonite-producing reactions at 1000 bars, as esti-mated by maximum densities of vapor-rich fluid inclu-sions. Depending on the composition of the fluid, the firstoccurrence of tremolite occurs at approximately 350-400'C. With the assumption of a limiting case of a completelyclosed system, the buffering capacity of the system setsthe first occurrence ofdiopside at 460'C and X.o, : 0.9,the last occurrence of tremolite at490"C and &o, : 0.75,and the first occurrence of wollastonite at 610'C and X.o,: 0.75. Figure 9 also shows tremolite- and diopside-pro-ducing reactions involving alkali feldspar and phlogopite.Since bulk composition of the limestones are low in KrOand Al2O3, these minerals are rare; therefore, they werenot included in Figure l.

At the diopside isograd, the isobaric invariant assem-blage calcite + dolomite + quartz * tremolite + diop-side flxes T and X.o, in marbles at about 470'C and 0.9.Wollastonite rims indicate that in metacherts X.o, was=0.02, irrespective of whether immiscibility occurred(Figs. 8, 9). Sharp boundaries between wollastonite rimsand marbles indicate that such large X.o, gradients oc-curred along distances of a few centimeters across theboundary between the wollastonite rim and the marble.

The absence of wollastonite in diopside-zone marblesshows that similar gradients existed also throughout thediopside zone. It is clear that a different fluid regime con-trolled the metamorphic evolution of the metachert lay-ers from that which prevailed in the limestones. Wollas-tonite rims along chert layers in the diopside and tremolitezones require infiltration of HrO-rich fluids, thus main-taining low X.o, and enhancing the decarbonation reac-tion, which in its turn is enhanced by fluid immiscibility.The metacherts acted as channelways for the brines.Limestones of the tremolite and diopside zones were im-pervious for highly saline fluids. Infiltration ofeven verysmall amounts of brines into the marbles must be ex-cluded, since that would have resulted in substantialprogress of the reaction calcite + qvartz: wollastonite+ CO, in the marbles at the H,O-rich side of the two-fluid solvus (Trommsdorff and Skippen, 1986). Phaseequilibria considerations are not conclusive as to whetherthe wollastonite-zone marbles were subject to brine infil-tration. However, stable isotope studies indicate that brineinfiltration was confined to a zone 7-17 m wide closelyadjacent to the intrusive contact (Heinrich and Hubber-ten, in preparation).

The peak temperature at the wollastonite isograd was610 "C at 1000 bars when derived from the wollastonite-producing reaction in a closed marble system (Fig. 9).Temperatures of 560-585 + 30 'C result when usingthermometry of the univariant assemblage garnet + qtarlz* plagioclase + wollastonite in neighboring metaargillitelenses (Heinrich and Gottschalk, in preparation). Giventhe pressure uncertainty from fluid inclusion densities,these temperatures broadly coincide. That fluid inclusionsamples from the wollastonite isograd have maximumhomogenization temperatures of at least 573 "C impliesthat trapping of two immiscible fluids in the metachertsactually started at peak metamorphic conditions of about600 "C and continued during cooling down to about500 "c.

Infiltration and K-Na metasomatism are further con-firmed by the presence of pectolite and potassium richter-ite in the previously alkali-poor metachert protolith. Theformation of potassium richterite at the expense of di-opside indicates that brine infiltration did not set in be-fore peak metamorphic conditions were achieved at thewollastonite isograd.

Origin and history of brines

The hypersaline brine populations, along with the va-por-rich COr-bearing inclusions, represent the earliestfluids detectable in the metachert system at the wollas-tonite isograd. Although hypersaline, their salt-undersat-urated nature at high ?n. (V - L) implies that they havenot been derived from boiling of salt-saturated weaklysaline fluid (Ramboz et a1., 1982). The hypersaline brinesof 66-73 wto/o (NaCl + KCI)"q and K/Na ion ratios ofapproximately I are therefore interpreted as having infil-trated as a fluid phase emanating from the crystallizingintrusion. The absence of a vertical fluid system, com-

NaCl


, 7

I

I 6\o

Za c

J

3 L3

Fig. 10. Isobaric isothermal section of the HrO-rich part ofthe system HrO-CO.-NaCl at 1000 bars exhibiting the compo-sitions ofsalt-saturated brines (solid circles) coexisting with CO.-rich vapor at various temperatures. The arrows indicate the evo-lution ofthe brine during isobaric cooling from 400 to 200'C.Data for the two-fluid solvus extrapolated from Gehrig (1980);salt saturation in the HrO-NaCl binary from Sterner et al. (1988).

bined with evidence for lithostatic pressure implies thatthis fluid was exsolved directly from the magma as a hy-persaline brine and does not represent a condensate (e.g.,Hein and Tistl, 1987). Decarbonation occurred ando as aresult of immiscibility, the low-density COr-bearing fluidmight have left the system. This implies that the originalbrine might have been less saline. The extent ofthis mod-ification would depend on the extent of the wollastonite-producing reaction.

The infiltrated brines did not escape into the marbles.One can reasonably infer that the brine in the metachertlayers shared the same cooling history as the aureole. Iftrapping had occurred continuously during cooling, brineinclusions would have evolved toward lower fh(V - L)at constant salinities along the H.O-COr-salt solvus untilsalt saturation had been attained. Then salinities wouldhave decreased, as would Z, (V - L), until 7" : 7- (salt).That should precipitate salt. Figure 54. shows preciselysuch a fluid evolution, indicating that the cooling brinesattained salt saturation at 350-400 "C and that trappingactually occurred continuously throughout the coolinghistory down to 150 "C. Inclusions of solid halite or syl-vite were not observed. It follows from the simplifiedHrO-COr-NaCl system (Fig. l0) that the CO, content ofthe salt-saturated brine is only slightly reduced duringcooling from 400 to 150 "C. Only small amounts of CO,were therefore exsolved. That may explain the absence ofvapor-rich COr-bearing inclusions at temperatures below500 'C. COr-rich fluids were trapped, obviously, only attemperatures above that for the decarbonation reaction.

L o g a ( K r / H + ) +

Fig. 1 1. Logarithmic activity diagram for a portion of thesystem NarO-KrO-AlrOr-SiOr-HCl at various temperatures forP : 1000 bars and at q\artz saturation. The position of thestability field boundaries were computed using Ge0-Calc (Brownet al.. 1988).

If this interpretation is correct, the brines in the meta-chert system must have been in contact with the largefluid reservoir of the cooling alkali syenite, and reequili-bration processes must have occulTed, at least partially(Labotka et al., 1988). The composition of this fluid res-ervoir is controlled by the mineralogy of the alkali syenite,which consists of about 900/o of the mode of perthiticalkali feldspar and is quartz-normative (Ramirez-Fernan-dez and Heinrich, 1991). Thus, under quartz-saturatedconditions, the activities of Na+ and K+ in the fluid co-existing with the feldspars are approximated by the equi-librium potassic feldspar * Na*/H* : albite + K+/H+,which is highly dependent on temperature (Fig. I l). Theratio of aKt/a""r is about I at 600 "C and decreases aboutone order of magnitude when the temperature drops to150 'C. The actual activity coemcients are unknown, butit is clear that the alkali concentrations in terms of K/Naion ratios should broadly follow the same trend. As shownabove, this trend is clearly seen in the evolution of thebrines. It is therefore concluded that a continuous outflowof fluids in equilibrium with the intrusion occurred alongthe metachert channels down to 150 "C. Their composi-tional characteristics, unless somewhat modified by theloss of COr-rich vapor during the process of decarbona-tion, were mainly controlled by mineral-fluid equilibriain the cooling intrusion.

Direction of brine flow

Ferry (1991), Baumgartner and Ferry (1991), and Ferryand Dipple (1992) argued that the channelized flow of


aqueous fluids in the direction of decreasing temperatureunder equilibrium conditions would produce a sharp re-action front at a single temperature and a single field point.In Bufa del Diente metacherts then, a band completelyreacted to wollastonite would change abruptly intoquartzite without wollastonite rims at some distance fromthe contact. In contrast, brine flow in the direction ofincreasing temperature would generate a wide zone inwhich calcite + qtJaftz * wollastonite would coexist. Theirarguments hold, even if immiscibility occurs. As shownabove, both features are not observed in the Bufa delDiente metacherts.

Wollastonite rims along metachert bands clearly rep-resent nonequilibrium textures (see Joesten, l99l). Thatimplies that equilibrium between calcite + quartz + wol-lastonite and infiltrating brines was not maintained overa few centimeters once wollastonite rims had formed. Itcan reasonably be inferred that the thin pegmatitic vein-lets (Fig. 28) trace the path of the infiltrating fluid. Theseindicate that flow occurred exclusively in the quartzitelayers and not in the developing wollastonite rims. Thedecarbonation reaction progress at any point is then con-trolled by a complex interplay of mass transport throughthe already developed wollastonite rim. This may en-hance in situ wollastonite formation. On the other hand,the advancing brine flow would cause wollastonite pro-duction with decreasing temperature away from the con-tact. Hydrothermal experiments (e.9., Tanneret al., 1985;Heinrich et al., 1989; Liittge and Metz, 1992) have shownthat mass transport through calc-silicate-armored relicsis a very slow process and may well be slower than hy-drous fluid fluxes under contact or regional metamorphicconditions (Labotka et al., 1988; Baumgartner and Ferry,199 l). Thus, the interpretation of the flow of the infil-trating brines in the direction of decreasing temperatureis compatible with the observed textures, as well as withthe occurrence of wollastonite formation from peakmetamorphism of about 600 down to about 500'C.

Two-fluid flow in the aureole

Infiltration of highly saline brine was confined to themetachert layers, and immiscibility of COr-bearing fluidproduced by decarbonation occurred within the meta-chert layers. The low-density Cor-rich fluid must thenhave infiltrated into overlying marble. This idea is sup-ported by observations made by Yardley and Bottrell(1988), who argued that where two immiscible fluid phasescoexist, the one produced by reaction tends to move outof the rock and leave the other behind. The rarity ofphenomena such as shown in Figure 38 confirms thatimmiscible fluids, in general, move separately (Yardleyand Bottrell, 1988). Experiments by Holness and Graham(1991) have shown that the fluid phase topology in mar-bles is a connected network for fluid only with X.o, closeto 0.5, permitting grain-edge flow in marbles only forsuch fluid compositions. Holness and Graham (1991)concluded "marbles would be impervious to the infiltra-tion of HrO-rich fluids and could therefore act as a semi-

permeable membrane losing locally generated fluid, butexcluding infiltrating fluid." This describes perfectly whathappened at the metachert-marble boundaries at the Bufadel Diente aureole. In contrast, quartz-dominated lithol-ogies permit pervasive flow of HrO-NaCl fluids, but notof H,O-CO, fluids (Watson and Brenan, 1987; Holnessand Graham, l99l). Vertical transport of immiscible low-density CO,-bearing fluid at the Bufa del Diente aureolewas further confirmed by stable isotope studies (Heinrichand Hubberten, 1991, and in preparation). Relativechanges in C and O isotopes along short profiles in over-lying, silicate-free marble perpendicular to the wollaston-ite-rimmed metachert layers at the wollastonite zone in-dicated that the fluid causing isotopic alteration in themarble was COr-rich.

If so, simple mass balance calculations will estimatethe amount of low-density COr-rich fluid generated at thechert-marble boundary that subsequently infiltrated theoverlying marble. If one assumes that all low-density COr-rich fluid is immediately lost, and one assumes an aver-age fluid composition of &o, : 0.5, X.or: 0.47, andX*.c,.. : 0.03 and an average density of 0.25 g/cm3 at 55OoC, every cubic centimeter of wollastonite produced cor-responds to approximately 6 cm3 COr-rich fluid infiltrat-ing the overlying marble. Given the average wollastoniterim thickness at the wollastonite isograd, the central di-opside zone, and the diopside isograd, and taking thevalues for fluid composition and density from above intoaccount along with the mean thickness of the Cuesta delCura Formation with about 50 chert layers, one finds thatabout 900, 600, and 90 cm3, respectively, of COr-richfluid infiltrated a marble column I cm x 1 srn x 70 mat the three outcrops 8502, Al7, and 84 (Fig. l). Withthe fluid flow distributed homogeneously along the wholecolumn, this would correspond to volumetric fluid-rockratios of 0. 13,0.08, and 0.013. Thus, less CO, infi l tratedthe marbles than was produced by them internally(Fig. e).

A striking point is that there is no indication of mixingbetween low-density CO, and CHo fluids. In graphite-freemetacherts, C must have been removed, probably by lossof CHo-rich fluid. However, the COr-rich inclusions mea-sured in 8502 and B33 are CH"-free. On the other hand,the metacherts 8503 and B504 do contain CHo fluidscoexisting with graphite, which are COr-free despite thefact that thick wollastonite rims indicate that largeamounts of CO, had been produced. A reasonable sug-gestion is that the two fluids never coexisted. Since thewollastonite-producing reaction follows a dissolution-precipitation mechanism (Tanner et al., 1985; Kerrick etal., l99l), CO, must have been generated at the dissolv-ing calcite crystals at the calcite-wollastonite boundaries.Fluid loss probably occurred immediately. In contrast,the fluid inclusion characteristics suggest that CHo for-mation by reaction of the brines with the C inside thequartzite nodules was sluggish and occurred over a widetime-temperature range. Removal of CHo-rich fluid mighthave been hindered by lack ofgrain boundary wetting by


CHo in the quartzite matrix. If so, persistence of a sepa-rate CHo-rich fluid in the quartzite possibly inhibited rap-id breakdown of graphite (see also Yardley and Bottrell,1 988) .

CoNcr,uoING REMARKS

It has been shown that the limestones of the Cuesta delCura Formation acted as aquitards and prevented for-mation of convection cells of hydrous fluids, as are com-monly observed in silicate-rich country rocks aroundshallow intrusions (Barton et al., 1991). Limestones areimpervious to grain-edge flow of aqueous fluids andquartz-rich beds of COr-rich fluids (Watson and Brenan,1987; Holness and Graham, l99l). Whenever a fracturesystem in an aureole is absent, convection cells cannotbuild up in metasediments containing specific strati-graphic horizons of these different lithologic types, nomatter which fluid type is present. Vertical fluid transporton a large scale in interstratified sediments is only pos-sible by fracturing. Fracturing processes induced by em-placement tectonics play an important role in the roofpendants of intrusions (Paterson et al., l99l; Barton etal., l99l). At the Bufa del Diente aureole, fracturing wasobserved in marbles of the Tamaulipas Formations southof the intrusion that were lifted to a much higher levelduring emplacement. The vast majority of aqueous fluidreleased by the alkali syenite may have escaped by thismechanism.

Experimental work by Johnson (1991) showed that theimmiscibiltiy field in the Hro-Cor-salt system expandsrapidly with pressure and that immiscibility occurs atgranulite facies conditions even at moderate salinities.Johnson (1991) pointed out that fluid immiscibility is awidespread phenomenon in metamorphic rocks. Thus, incalcareous metasediments, boiling off of COr-rich fluidproduced by decarbonation reactions at the two-fluid sol-vus may cause contrasting fluid flow patterns over a widerange of P-T conditions up to the granulite facies.

AcxNowr-nlcMENTS

I thank the members of the Facultad de Ciencias de la Tierra, Uni-versidad Aut6noma de Nuevo Le6n, Linares, M6xico, for field assistanceP. Metz and D. Mangliers, Mineralogisch-Petrographisches Institut,Universitit Tiibingen, and F. Galbert, ZELMI, Technische UniversitetBerlin, provided electron microprobe facilities. Thanks go also to A. Beck-er, SFB 69, Technische Universitit Berlin, who performed AAS and ICPanalyses, and to E Horn, IGDL, Universitiit Giittingen, for help with theRaman analysis. M. Gottschalk, F Lucassen, and A. Liittge read earlierversions of the manuscript. Very constructive reviews of V. Sisson andW. Lamb improved the paper substantially. Part of this work was sup-ported by a research grant of the Mexican Secretaria de Educaci6n Pfb-lica.

Rrrnnrucrs CrrnrAhmad, S.N., and Rose, A.W (1980) Fluid inclusions in porphyry and

skam ore at Santa Rita, New Mexico Economic Geology, 75,229-250.

Barton, M.D., Staude, J.M., Snow, EA, and Johnson, D.A. (1991) Au-reole systematics. In Mineralogical Society of America Reviews in Min-eralogy, 26, 723-847.

Baumgartner, L.P , and Ferry, J.M. (1991) A model for coupled fluid-flow

and mixed-volatile mineral reactions with applications to regional

metamorphism. Contributions to Mineralogy and Petrology, 106,273-

285 .Bottrell, S.H., Greenwood, PB, Yardley, B.W.D., Shepherd, T'J., and

Spiro, B. (1990) Metamorphic and post-metamorphic fluid flow in the

low-grade metamorphic rocks of the Harlech dome, north Wales Jour-

nal of Metamorphic Geology, 8, 131-143.

Bowers, T.S., and Helgeson, H.C. (1983a) Calculation of the thermody-

namic and geochemical consequences ofnonideal mixing in the system

HrO-COr-NaCl on phase relations in geological systems: Metamorphic

equilibria at high pressures and temperatures. American Mineralogist,

68 ,1059 - i 075 .-(1983b) Calculation of the thermodynamic and geochemical con-

sequences of nonideal mixing in the system H,O-CO,-NaCI on phase

relations in geologic systems: Equation of state for HrO-CO:-NaCl flu-

ids at high pressures and temperatures. Geochimica et Cosmochimica

Acta, 47 , 1247 -127 5Brown, P.E (1989) FLINCOR: A microcomputer program for the reduc-

tion and investigation of fluid-inclusion data. American Mineralogrst,

74,1390-1393.Brown. T H., Berman, R.G , and Perkins, E.H. (1988) Ge0-Calc: Soft-

ware package for calculation and display of pressure-temperature-com-

position phase diagrams using an IBM or compatible personal com-

puter Computers and Geosciences, 14, 27 9-289.

Chou, L-M., Pasteris, J.D., and Seitz, J.C. (1990) High-density volatiles

in the system C-O-H-N for the calibration of a laser Raman mrcro-

probe. Geochimica et Cosmochimica Acta, 54,535-543.

Crawford, M.L. (1981) Phase equilibria in aqueous inclusions In LS

Hollister and M.L Crawford, Eds , Short course in fluid inclusions:

Applications to petrology, p. 75-100 Mineralogical Association of

Canada, Toronto, Ontario, CanadaFerry, J M. (1983) Applications of the reaction progress variable in meta-

morphic petrology. Journal of Petrology, 24, 343-37 6.- (l 986) Reaction progress: A monitor of fluid-rock interaction dur-

ing metamorphic and hydrothermal events. In J V. Walther and B J.

Wood. Eds., Fluid-rock interactions during metamorphism, p 60-88

Springer-Verlag, New York-(1987) Metamorphic hydrology at l3-km depth and 400-550'C.

American Mineralogist, 7 2, 39-58-(1988) Contrasting mechanisms of fluid flow through adjacent

stratigraphic units during regional metamorphism, south-central Maine,

USA. Contributions to Mineralogy and Petrology, 98, l-lZ.-(1991) Dehydration and decarbonation reactions as a record of

fluid infiltration. In Mineralogcal Socrety of America Reviews in Min-

eralogy, 26, 35 1-393.Ferry, J M., and Dipple, G.M (1992) Models for fluid flow, mineral re-

action, and isotopic alteration during contact metamorphism: The Notch

Peak aureole, Utah Amencan Mineralogist, 77, 577-591.

Frantz, J D., Popp, R K, and Hoering, T.C. (1992) The compositional

limits of fluid immiscibility in the system H,O-NaCI-CO, as deter-

mined with the use of synthetic fluid inclusions in conjunction with

mass spectrometry. Chemical Geology, 98, 237 -25 5.

Gehrig, M. (1980) Phasengleichgewichte und Ptr4"-Daten terniirer Mi-

schungen aus Wasser, Kohtendioxid und Natnumchlorid bis 3 kbar

und 5 50'C. Hochschulverlag, Freiburg, Germany.Hein. U.F., and Tistl, M. (1987) Characteristics of fluid inclusions in the

porphyry copper deposit at La Granja, Peru Chemical Geology, 61,

I 83- I 92.Heinrich, W, and Hubberten, H.W. (1991) Infiltration, buffering, and

immiscibility of NaCI-KCI-H'O-CO' fluids at Bufa del Diente intru-

sion, NE-M6xico: Fluid inclusion and isotope studies. Terra Abstracts,

3 ( r ) , 4 3 8Heinrich. W, Metz, P, and Gottschalk, M (1989) Experimental inves-

tigation of the kinetics of the reaction I tremolite + 1 1 dolomite : 8

forsterite + 13 calcite + 9 CO, + I H,O. Contributions to Mineralogy

and Petrology, lO2, 163-173.Holness. M.B., and Graham, C M. (1991) Equilibrium dihedral angles rn

the system H,O-CO,-NaCl-calcite, and implications for fluid flow dur-

ing metamorphism Contributions to Mineralogy and Petrology' 108,

368-383.Hover-Granath, V C, Papike, J J., and Labotka, T.C (1983) The Notch


Peak contact metamorphic aureole, Utah: Petrology of the Big HorseLimestone Member of the Orr Formation Geological Society of Amer-ica Bulletin, 94, 889-906.

Hubberten, H W. (1985) The Sierra de San Carlos, Tamaulipas: An ig-neous complex ofthe eastern Mexican Alkaline Province Zentralblattfiir Geologie und Paliiontologie, Teil l, Heft 9/10, 1183-1191.

Jacobs, G.K., and Kerick, D.M. (1981) Methane: An equation of starewith application to the ternary H,O-CO,-CH.. Geochimica et Cos-mochimica Acta, 45, 607 -6 14.

Joesten, R.L. (1991) Kinetics ofcoarsening and diffusion-controlled min-eral growth. In Mineralogical Society of America Reviews in Miner-alogy,26, 507-582

Johnson, E.L (1991) Expenmentally detemined limits for H,O-CO,-NaCIimmiscibility in granulites. Geology, 19, 925-928

Kelley, D.S., and Robinson, PT. (1990) Development of a brine-domi-nated hydrothermal system at temperatures of400-500.C in the upperlevel plutonic sequence, Troodos ophiolite, Cyprus. Geochimica etCosmochimica Acta, 54, 653-661.

Kerrick, D M., and Jacobs, G.K. (1981) A modified Redlich-Kwong equa-tion for H,O, COr, and HrO-CO, mixtures at elevated pressures andtemperatures. American Journal of Science. 28 l. 7 35-j 6j

Kerrick, D.M., Lasaga, A.C., and Raeburn, SP. (1991) Kinetics of het-erogeneous reactions In Mineralogical Society ofAmerica Reviews inMineralogy, 26, 581-672.

Krader, T. (1985) Phasengleichgewichte und kritische Kuruen des terni-ren Systems HrO-CHo-NaCl bis 250 MPa und 800 K. ph.D. thesis,University of Karlsruhe, Karlsruhe, Germany.

Labotka, T C., Nabelek, P.I , Papike, J J. (1988) Fluid infiltration throughthe Big Horse Limestone in the Notch Peak contact-metamorphic au-reole, Utah. Amerian Mineralogist, 73, 1302-1324

Liittge, A., and Merz, P. (1992) Mechanism and kinetics of the reaction:I dolomite + 2 q\artz: I diopside + 2 CO, investigated by powderexpenments. Canadian Mineralogist, 29, 803-821.

Nabelek, P.I , I-abotka, T C , O'Neil, J.R., and papike, J J (1984) Con-trasting fluid/rock interaction between the Notch peak granitic intru-sion and argillites and limestones in westem Utah: Evidence from sta-ble isotopes and phase assemblages. Contributions to Mineralogy andPetrology, 86,25-34.

Parry, W T. (1986) Estimationof Xr,,,, P, and fluid inclusion volume fromfluid inclusion temperature measurements in the system NaCl-COr-H,O. Economic Geology, 81, 1009-1013.

Paterson, S R., Vemon, R.H., and Fowler, T K (1991) Aureole tectonicsIn Mineralogrcal Society of America Reviews in Mineralogy, 26,673-722.

Pichavant, M , Ramboz, C., and Weisbrod, A. (1982) Fluid immiscibilityrn natural processes: Use and misuse of fluid inclusion data. I phaseequilibria analysis: A theoretical and geometrical approach. ChemicalGeology, 37, l-27

Potter, R.W., II, Clynne, M A , and Brown, D L. (1978) Freezing pointdepression of aqueous sodium chloride solutions. Economic Geology,73,284-285.

Ramboz, C , Pichavant, M., and Weisbrod, A. ( I 982) Fluid immiscibilityin natural processes: Use and misuse of fluid inclusion data Il. Inter-

pretation of fluid inclusion data in terms of immiscibility. ChemicalGeology, 37 , 29-48.

Ramirez-Fernandez, JA., and Heinrich, W. (1991) Geology of the Ter-tiary Bufa del Diente intrusion and its contact aureole, Sierra de SanCarlos, Tamaulipas, M6xico. Zentralblatt fiir Geologie und PaHonto-l og ie ,6 , l 5 l 9 -1531 .

Rice, J.M (1977) Progressive metamorphism of the impure dolomiticlimestone in the Marysville aureole, Montana. American Journal ofScience,277, l-24

Rice, J.M., and Ferry, J.M. (1982) Butrering, infiltration, and the controlof intensive variables during metamorphism. In Mineralogical Societyof America Reviews in Mineralogy, 10,263-326.

Rumble, D ,III, Ferry, J.M., Hoering, T.C., and Boucot, A.J. (1982) Fluidflow during metamorphism at the Beaver Brook fossil locality, NewHampshire. American Journal ofScience, 282, 886-919

Sisson, VB., Crawford, M.L, and Thompson, PH. (1981) CO,-brineimmiscibility at high temperatures: Evidence from calcareous metased-imentary rocks. Contributions to Mineralogy and Petrology,78,371-378 .

Sterner, S.M., Hall, D.L., and Bodnar, R.J (1988) Synthetic fluid inclu-sions. V. Solubility relations in the system NaCl-KCl-H,O under va-por-saturated conditions. Geochimica et Cosmochimica Acta, 52,989-I 005

Tanner, S.B, Kemck, D.M, and Lasaga, A.C. (1985) Experimental ki-netic study of the reaction: Calcite + qtartz : wollastonite + carbondioxide, from I to 3 kilobars and 500" to 850 .C. American Joumal ofScience. 285. 5'77-620.

Trommsdorfr V., and Skippen, G. (1986) Vapour loss ("boiling") as amechanism for fluid evolution in metamorphic rocks. Contributions toMineralogy and Petrology, 9 4, 317 -322.

Trommsdorff, V., Skippen, G., and Ulmer, P. (1985) Halile and sylviteas solid inclusions in high-grade metamorphic rocks. Contributions toMineralogy and Petrology, 89, 24-29.

Walther, J. (1981) Fluide Einschliisse im Apatit des Carbonatits vomKaiserstuhl (Oberrheingraben): Ein Beitrag zur Interpretation der Car-bonatitgenese. Ph D thesis, University of Karlsruhe, Karlsruhe, Ger-many

Watson, E B, and Brenan, J M. (1987) Fluids in the lithosphere. I Ex-perimentally determined wetting characteristics of COr-HrO fluids andtheir implications for fluid transport, host-rock physical properties, andfluid inclusion formation Eanh and Planetary Science l-etters, 85,497-5 1 5 .

Williams-Jones, A E, and Ferreira, D R (1989)Thermal metamorphismand H,O-CO,-NaCI immiscibility at Parapedia, Quebec: Evidence fromfluid inclusions. Contributions to Mineralogy and Petrology, 1O2,247-254

Yardley, B.W.D, and Bottrell, S.H. (1988) Immiscible fluids in meta-morphism: Implications of two-phase flow for reaction history. Geol-ogy, 16, 199-202.

MeNuscrrpr RECETYED Mw 4, 1992M,qNuscnrpr AccEmED Mencs 5. 1993

fluid infiltration

Documents