activity diagrams for clinoptilolite: susceptibility of …activity diagrams for clinoptilolite:...

American Mineralogist, Volume 75, pages 601419, 1990

Activity diagrams for clinoptilolite: Susceptibility of this zeolite to furtherdiagenetic reactions

Tnnrs.q, S. Bownns, Rocrn G. BunNsDepartment of Earth, Atmospheric, and Planetary Sciences, Massachusetts Institute of Technology,

Cambridge, Massachusetts 02 I 39, U.S.A.

Ansrn-q.cr

Clinoptilolite is the predominant zeolite in diagenetically altered volcanic rocks at YuccaMountain, Nevada, having formed by posteruptive reactions of ground water with vitrictuffs in the pyroclastic deposits. Compositional variations of clinoptilolites in the fracturedand zeolitized tuffs not presently in contact with ground water and the vulnerability ofzeolites to burial diagenesis raise questions about the long-term stability ofclinoptilolite.Equilibrium activity diagrams were calculated for clinoptilolite solid solutions in the sev-en-component system Ca-Na-K-Mg-Fe-Al-Si plus HrO, employing available thermody-namic data for related minerals, aqueous species, and water. Stability fields are portrayedgraphically on plots of log(ar^,/ar*) versus log(a*r-/(a".)2), assuming the presence of po-tassium feldspar, saponite, and hematite, and using ranges of activities for SiO, and Al3-defined by the saturation limits for several silica polymorphs, gibbsite, kaolinite, andpyrophyllite. Formation of clinoptilolite is favored by higher SiO, activities than allowedfor by the presence of quartz, thus accounting for the coexistence of clinoptilolite withopal-CT in zeolitized vitric tuffs. The clinoptilolite stability field broadens with increasingatomic substitution of Ca for Na, and K for Ca, reaches a maximum for intermediateactivities of dissolved Al, and decreases with increasing temperature. The thermodynamiccalculations show that ground water of the sodium-bicarbonate type, such as referenceJ-13 well water collected from fractured devitrified tuffs at the adjacent Nevada NuclearTest Site, is approximately in equilibrium at 25 "C with calcite and several zeolites, in-cluding heulandite and calcic clinoptilolite. Mg-rich clinoptilolites are stabiliz€d in groundwater depleted in Ca'?-. Decreasing Al3t activities result in the association of clinoptilolitewith calcite and opal-CT observed in weathered zeolitized vitric tuffs at Yucca Mountain.The activity diagrams indicate that prolonged diagenetic reactions with ground water de-pleted in Al, enriched in Na or Ca, and heated by the thermal envelope surrounding buriednuclear waste may eliminate sorptive calcic clinoptilolites in fractured tuffs and underlyingbasal vitrophyre.

INrnooucrroN

Clinoptilolite, ideally (Na,K,Cao r)usi3oAl6o72. 24H2O,is an abundant zeolite in diagenetically altered rhyolitetuffs, where it forms by posteruptive reactions of hydrat-ed glass shards (Hay, 1966; Hay and Sheppard, 19771,Iijima, 1975, 1980). Silicic air-fall and ash-flow tuffs arethe predominant rocks at Yucca Mountain, Nevada, thesite of the proposed repository for burial of high-levelnuclear waste (U.S. DOE, 1988). The repository horizonat this site is a densely welded and devitrified tuff unitunderlain by a basal vitrophyre and vitric to zeolitizednon-welded tuffs containing high proportions ofclinop-tilolite (Broxton et al., 1987). Because of its favorablecation exchange reactions, clinoptilolite is advocated asone mineral in the tuffaceous deposits capable of immo-bilizing several ofthe soluble cations and being an effec-tive barrier to radionuclide migration (Ogard et a1., 1984),

0003404x/90/0506-o601$02.00 601

should water entering the repository cause leakage offis-sion products.

Clinoptilolites analyzed in drill-core samples through-out Yucca Mountain and its vicinity display wide com-positional variations, particularly in fractures adjacent tothe repository horizon in the Topopah Spring Member ofthe Paintbrush Tuffunit and in underlying zeolitized vit-ric tuffs (Broxton et al., I 987). In the vadose zone beneathYucca Mountain, clinoptilolites with high Ca and Mgcontents line fractures in the Topopah Spring Member(Carlos, 1985; Broxton et al., 1986, 1987). In underlyingbedded tuffs, the clinoptilolites display regional and depthvariations (Broxton et al., 1986, 1987). On the westernside of Yucca Mountain the clinoptilolites are Na-K-bearing and become Na-rich with depth. To the east, theclinoptilolites are Ca-K-bearing and become Ca-rich withdepth. Such compositional variations are portrayed in

602 BOWERS AND BURNS: ACTIVITY DIAGRAMS FOR CLINOPTILOLITE

Na Ca+Mg

Fig. l. Triangulardiagram showingcompositionalvariationsof clinoptilolite and heulandites (modified from Broxton et al.,1986). Superimposed on data for these zeolites in the unsatu-rated zone at Yucca Mountain, Nevada, are compositions usedin the thermodynamic calculations: O, O western and easternparts of the repository block (Broxton et al., 1986); V liningfractures in and below the proposed repository horizon (Carlos,1985); * measured calorimetrically (Hemingway and Robie,I 98a); X different atomic substitutions used to calculate activitydiagrams in Fig. I 1.

Figure 1. Similar regional and depth variations occur inzeolitized tuffs at the adjacent Nevada Test Site (Hoover,1968). Such compositional variations of clinoptilolite notpresently in contact with ground water and the suscepti-bility of zeolites to further diagenetic reactions (Moncureet al., 1 98 I ; Smyth, 1982) suggest that the long-term ther-modynamic stability of clinoptilolite should be exam-ined.

Although the Topopah Spring Member tuffand under-lying zeolitized tuffs lie in the unsaturated or vadose zone,well above the present-day water table beneath YuccaMountain, these formations dip to the east so that at thelocation of the nearest water-supply well, designated J- I 3and located 6 km to the east at Jackass Flat on the Ne-vada Test Site, the Topopah Spring Member lies beneaththe water table. The major producing horizon for J-13well water is a highly fractured interval within the To-popah Spring Member (Delany, 1985). The chemicalcomposition of the ground water obtained from well J- I 3has been monitored for several years (Daniels elal.,1982;Bish et al., 1984; Kerrisk, 1987) and serves as a standardin laboratory experiments and geochemical modelingstudies for the Yucca Mountain exploration block (e.g.,Oversby, I 985; Delany, I 98 5; Knauss et al., I 985a, I 985b;Moore et al., 1986; Thomas, 1987; U.S. DOE, 1988, p.4-5 I ). Whether J- I 3 well water is of an appropriate com-position for the prediction of authigenic mineral reactions

in the unsaturated zone beneath Yucca Mountain re-quires critical evaluation.

In order to assess the stability limits of clinoptiloliteand its vulnerability to changes of ground water chem-istry relative to the composition of J-13 well water,equilibrium activity diagrams have been calculated forclinoptilolite solid solutions in the system Ca-Na-K-Mg-Fe-Al-Si-HrO, employing available thermodynamic datafor oxide and aluminosilicate phases. Although low-tem-perature processes are subject to many kinetically con-trolled variables, the assumption of thermodynamicequilibrium provides a reference from which to assessobserved mineral relations. Results reported here indi-cate that authigenic minerals such as clinoptilolite modifyand are modified by ground-water compositions.

C,l,lcuurroNs oF ACTTYTTY DTAGRAMS

Sources of thermodynamic data

The method for calculating activity diagrams is de-scribed in Bowers et al. (1984), who also tabulated ther-modynamic data for many of the phases considered here.Thermodynamic data for zeolites including clinoptilolite,heulandite, natrolite, scolecite, and mesolite are providedby calorimetric measurements made by Johnson et al.(1982, 1983, 1985) and Hemingway and Robie (1984).Table I lists formulae of all of the minerals consideredin this study. Note that several of these minerals (notablylaumontite, stilbite, and chabazite) were found not to ex-hibit stability fields within the pressure, temperature, andcompositional ranges of the calculations.

The clinoptilolite measured by Hemingway and Robie(1984) from altered tuffs of the Big Sandy Formation,Mohave County, Arizona (Sheppard and Gude, 1973) wasformulated by them as

Na. ruIQ *Ca, roMg, r,Feo 14,16 .sizso i 2' 22H2O,

with Si/Al : 4.33 and (Na + K) > Ca. The heulanditemeasured by Johnson et al. (1985) was formulated as

Cao.rrrSro ,rrBao ourNao ,83Iq r32Si6 835A12 165Or8 '6.0H2O,

with SilAl :3.16 and (Ca * Sr + Ba) > (Na + K). Notethat clinoptilolite and heulandite are isostructural, and acontinuous solid-solution series exists between the twominerals. By convention, the compositional boundary forclinoptilolite is determined by the ratios Si/Al > 4 and(Na + K) > Ca, whereas heulandite has Si/Al < 4 andCa > (Na + K). As indicated in Figure l, the composi-tion of the clinoptilolite for which thermodynamic dataare available resembles those of the (Ca + Mg)-rich zeo-lites that line fractures in the Topopah Spring Memberand are present in underlying zeolilized vitric tufs, par-ticularly beneath the northeastern block of Yucca Moun-tain (Broxton et al., 1987).

Several minerals of interest in this study, including cli-noptilolite, have measured heat-capacity and entropy data,but no Gibbs free energy of formation at 25 'C (AG8).When this is the case, free eneryies can be estimated by

BOWERS AND BURNS: ACTIVITY DIAGRAMS FOR CLINOPTILOLITE

Tlele 1. Minerals and formulas

603

OuartzAmorphous silicaGibbsiteDiasporeKaolinitePyrophylliteWdlastoniteGrossularAnorthiteGehlinitePrehniteMargariteCa end-member beidelliteLawsoniteWairakiteLaumontiteChabaziteCa end-member phillipsiteScoleciteEpistilbiteHeulanditeAlbiteNephelineParagoniteNa end-member beidelliteNa end-member phillipsiteAnalcimeNatroliteMordenitePotassium feldsparKalsiliteMuscoviteK end-member beidelliteK end-member phillipsitePhillipsiteH end-member beidelliteStilbiteMesoliteClinoptiloliteCa end-member saponiteHematite

sio,sio,A(OH)3Aro(oH)Al,siro5(oH)4Alrsi40,o(oH),CaSiO"Ca3Al25ieor2CaAlrSirogCarAlrSiO,Ca.Al,SLO,o(OH)eCaAl4Sirolo(OH),Cao r6A12(A10 sSi3 670,oXOH),CaAl,StOlOH),.H,OCaAl.SinO,..2HrOCaAlrSinO,.'4H.OCaAtSi401,.6H,OCaAlrSi501..5HrOCaAlrSi30,o.3HrOCaAlrSi60,6.5HrOCaAlrSirO,.. 6HrONaAlStOsNaAlSi04NaAl,(AlSioo,oXOH),Nao $Al2(Alo sSis nO,oXOH),NarAl2siso,4 5H2ONaAlSirOu.H.ONarAlrSi"O,o 2H"ONaAlSi5Orr gH?OKAtSirosKAtSi04KA|,(AtSi3O'oXOH),lG sAlr(Alo ssis 670,o)(oH),lcAlrsi5ol4.5HroNa, *lG roAl, *Siu,20,6. 6H2OHo sAlf Alossi3 dolo)(oH),NaCa"AluSi,rO*. 1 4H.ONa" ur"Gao *rAl, *Sio olOto 2.647HrO(Nao solG scal sMg, ,.[Al" rFeo.)Sioor2'22H2OCao rsMg3(Alo sSi3 670,oXOH),Fer03

various component-summation methods. The method ofChen (1975) was used here for Na end-member phillips-ite, K end-member phillipsite, and (Na,K) end-memberphillipsite. This method involves summing several setsof components, including oxides, simple silicates, andcomponents of similar structural type. The resulting free-energy determinations are f,t with an exponential curvethat asymptotically approaches a low value that is usedas AG!. Chen (1975) reported an estimated error in AG?obtained in this manner of less than 0.60/o for a numberof minerals when compared to their experimental values.An attempt was made to estimate free energies for theclinoptilolite formulated by Hemingway and Robie (1984)and for end-member heulandite, Ca end-member phillips-ite, stilbite, chabazite, and epistilbite formulated in TableI using Chen's (1975) method. Clinoptilolite and end-member heulandite were treated as separate phases onthe basis of different Si-Al ratios even though a contin-uous solid-solution series exists between the two min-erals. The independent estimates of AGP from summingvarious components for clinoptilolite and each of the end-member minerals listed above did not fall along expo-nential curyes, but decreased linearly such that an expo-nential flt could not be made. Chen (1975) noted a similar

difficulty with some minerals. The free energy of forma-tion for these minerals was therefore taken as the lowestvalue obtained from the various component-summationschemes and generally involved summing over zeolite andother hydrous components, including natrolite, scolecite,K end-member phillipsite, brucite, kaolinite, quartz, andhematite. The error associated with this method will besomewhat higher than the 0.60lo estimated by Chen (1975).All estimated AGP and AHP data used in this study are inTable 2. In addition, since alkali-rich mordenite fre-

TABLE 2. Estimated free energies and enthalpies of some zeo-lites"

acft298)(cal/mol) AHP(298)(cal/mol)

-2 003 469-2025 549-2 010 073-2 294 159

-2 234 763-1 866 861-5247 477-2 519 882-9 811 932

Na end-member phillipsiteK end-member phillipsiteCa end-member phillipsitePhillipsiteMordeniteEpistilbiteChabaziteStilbiteHeulanditeClinoptilolite

-1 851 376-1 871 460-1 860 596-2 112 851-1 465 647-2 065 242-'t 712 637-4 833 571-2 326 575-9 057 789

' Mineral formulas given in Table 1.

604 BOWERS AND BURNS:ACTIVITY DIAGRAMS FOR CLINOPTILOLITE

TABLE 3. Estimated free energies for compositionally variableclinoptilolite

AGP(298) (cal/mol)

Cornposition of ground water

Because the Topopah Spring Member tuffis the majorproducing horizon for water pumped from the J- l3 well,it has been generally assumed (Oversby, 1985) that thecomposition of J- 13 well water approximates the pre-vailing water chemistry of the proposed repository hori-zon in the same formation at Yucca Mountain eventhough the Topopah Spring Member there is in the un-saturated zone. As a result, J-13 well water continues tobe widely used as the reference aqueous phase for cali-brating numerous environmental parameters relevant tothe Yucca Mountain repository horizon (Oversby, 1985;Delany, 1985; Knauss et al., 1985a, 1985b; Moore et al.,1986; Thomas, 1987; U.S. DOE, 1988, p. 4-51). Thechemical composition of J-I3 well water has been mon-itored for several years (Daniels et al., 1982; Kerrisk,1987), and typical concentrations ofdissolved species init are summarized in Table 4. Small fluctuations in con-centrations over time have been recorded, but the vari-ations are minor compared with other variables in ex-periments in which J- I 3 well water was used (Daniels etal., 1982). However, during experiments in which J-13well water was contacted with tuff samples of the Topo-pah Spring Member taken from a drill core at the appro-priate region of main water production of the J- I 3 well,concentrations of many constituents changed slightly(Table 4), particularly Mg and Al, which decreased afterthree weeks at room temperature (Daniels et al., 1982).Moreover, filtration affected the compositions of someelements, particularly Fe, Al, and Mg, which were dras-tically reduced in samples passed through 0.05 pm Nu-clepore membranes compared to those obtained from 0.45pm Millipore filters (Daniels et al., 1982). The Al con-centration, for example, decreased from -40 mg/L Q.a5pm filter).to <0.01 mgll- (0.05 pm filter) (Daniels et al.,1982). The Al concentration of J-I3 well water (0.012mg/L) represents one of the lowest values reported fordrill holes throughout Yucca Mountain (Ogard and Ker-r i sk ,1984 ) .

Cation concentrations in solutions contacted with vit-rophyre samples from the Topopah Spring Member at152 "C showed significant increases of dissolved Si, Fe,Al, K, and Na and a decrease of dissolved Mg, whichwere attributed to dissolution of glass and precipitationof clays (Daniels et al., 1982). A specimen of zeolitizedtuff reacted with J- I 3 well water at | 52 t showed markeddissolution of clinoptilolite and disappearance of mor-denite and cristobalite (Daniels et al., 1982). In later ex-periments, Knauss et al. (1985a, 1985b) studied compo-sitional changes of J- I 3 well water after reacting it withcrushed tuff and polished wafer samples of the denselywelded, devitrified ash-flow tuffin a drill core taken fromthe respository level in the Topopah Spring Member. Themodal minerals of this horizon consist of -980/o micro-crystalline sanidine-cristobalite-quartz and accessory(<2o/o\ biotite-montmorillonite (Bish et al., 1984; Brox-ton et al., 1989). Reactions were performed over 2-3months at temperatures of 90, 150, and 250 "C and pres-

(NaosKoescar sMgl B)-(A167Feoo)Siao,r.22H20

Na-Ca substitutionNal 56Ca1 oNa2 scao5

Nd.r,Nao2BCa, s

Cd, r"K-Ca substitution

Kr e6oa1 oK2 e6oao s

K. r,Kosoa1 7n

C?, ,,Ca-llg substitution

C4'.MgroCa,7MQ, reCaroMgor"Car .Mgoo

Cazrs

-9 057 789

-9 052 5't7-9 047 244-9 041 972-9 059 265-9 060 742

-9 068 653-9 079 51 7-9 090 381-9 052 574-9047 142

-9 036 21 1-9 063 394-9 071 801-9 085 813-9 092 259

quently coexists with clinoptilolite in zeolitized tuffs atYucca Mountain (Sheppard et al., 1988), its AG! was alsoestimated by Chen's (1975) method. However, becauseno measured thermodynamic data are available for mor-denite, this zeolite was used only in activity diagramscalculated at 25 "C, although, as noted later, mordeniteoccurs at elevated temperatures in hydrothermal envi-ronments (Sturchio et al., 1989). The end-member heu-landite formulated in Table I was the only one used inthe calculations presented here because, as noted earlier,the specimen measured by Johnson et al. (1985) con-tained Ba and Sr components not included in this study.

Previous estimates of thermodynamic properties of cli-noptilolite, heulandite, and mordenite were made byKerrisk (1983). However, those calculations were madeprior to recently published calorimetric data (Johnson etal.,1982,1983, 1985; Hemingway and Robie, 1984). Di-rect comparisons between Kerrisk's (1983) estimates andthose presented here cannot be made because he esti-mated equilibrium constants of specific hydrolysis reac-tions rather than free eneryies of formation of the min-erals themselves.

Table 3 contains estimated free energies of formationfor compositionally variable clinoptilolites. Independentsubstitutions of Na for Ca, K for Ca, and Ca for Mg wereallowed, where constant charge balance and Si/Al ratioof 4.33 are maintained. Corresponding AG! values foreach of these derived compositions were estimated fromthe value given for clinoptilolite in Table 2 by a com-ponent-summation method using natrolite, scolecite, andHrO for Na-Ca substitution; K end-member phillipsite,scolecite, quartz, and HrO for K-Ca substitution; and sco-lecite, brucite, kaolinite, and quartz for Ca-Mg substitu-tion. These correction mechanisms resulted in lowerAG$ values for Ca over Na, K over Ca, and Ca over Mg-rich clinoptilolites.

BOWERS AND BURNS: ACTMTY DIAGRAMS FOR CLINOPTILOLITE

Tlele 4. Chemical composition of J-13 well water (mg/L)

605

LiNaKCaMg5rAIFesio,NO.FclHCO3SOopH

0.04243.95 . 1 1

12.51 .920.0350.0120.006

57.99.62.26.9

125.318.7

/ . b

0.999

539.02.37.2

178.818.37.27

58.55.586.460.315

1.64

148O E

2.47.4

61.018.56.97

444.4

1 32.0

598.72.2

1201 9

t . t

0.06455.3

1 1 . 51 .76

0.030.04

64.31 0 . 12.16.4

14318.16.9

0.05514.9

1 42.10.050.030.04

665.62.2I - 5

120227.1

0.0554.16.4

1 10.950.0020.010.004

71.6

55I - a

1 1 . 5'I .1

A. Delany ('1985).B. J-13 reacted with TS tuff at 90'C; Knauss et al. (1985a).C. J-1 3 reacted with TS tuff at 1 50 .C; Knauss et al. (1 985a).D. Moore et al. (1986).E. Ogard and Kenisk (1 984).F. Daniels et al. (1982).G. Daniels et al. (1982), after J-13 water reacted with TS tuff at 25.C.

sures of 90-100 bars. Results from the experiments at150 "C are in Table 4, where it can be seen that dissolvedSiO, concentrations increase and are close to the cristo-balite saturation value (Knauss et al., 1985a). Na alsoincreased during the experiments, Ca and Mg decreased,and Al and K both increased rapidly and then decreased.These effects were attributed to dissolution of montmo-rillonite and precipitation ofcalcite and illite. The exper-iments at 90 and 250'C produced similar results. Phasesidentified by scanning electron microscopy included illite,Mg-Ca or Fe-rich clays, gibbsite, calcite, and a pure SiO,phase considered to be cristobalite (Knauss et al., 1985a,1985b).

Studies to determine compositional changes of groundwater as it passes through the unsaturated zone in tuffa-ceous deposits have been conducted at Rainier Mesa, lo-cated 50 km to the north-northeast of Yucca Mountain(Benson, 1976; White et al., 1980). At Rainier Mesawelded and vitric tuffs overlie zeolitized. tuffs, resemblingthe sequence of ash-flow tufs at Yucca Mountain. Con-centrations of Ca and Mg in interstitial waters decreasedas a function of depth and were generally lower than thosein J- I 3 well water, whereas opposite effects were observedfor Na (Benson, 1976; White et al., 1980). The concen-tration of dissolved K was lower at depth, and that ofSiO, higher, than J- I 3 well-water compositions, whereasCl- decreased and HCO; increased with depth. The max-imum compositional variations of the interstitial vrateroccurred in alteration zones containing clinoptilolite andmontmorillonite (Benson, 1976; White et al., 1980). Waterseeping through fractures in tunnels beneath the zeoli-tized tuffs was HCOl-rich and had lower Ca, Mg, andSiO, contents, variable K concentrations, and higher Naconcentrations than J-13 well water (Benson, 1976; Whiteet al., 1980). The clinoptilolites along the fractures wereCa-Mg-K-rich, correlating with the depletion of these cat-

ions in the ground water, whereas the fracture-flow waterwas enriched in HCO; relative to the more Cl -rich in-terstitial water. Comparisons made with dissolution ex-periments on vitric and crystalline tuffs demonstrated therapid release of Na and SiO, but the retention of K inglass-bearing tuffs, whereas dissolution of crystalline tuffscontaining sanidine, qvart4 biotite, and clinopyroxenephenocrysts and sanidine-cristobalite groundmass result-ed in solutions rich in Ca,Mg, and HCO, (White et al.,1980). White et al. (1980) thus concluded that fracture-water compositions, such as J-13 well water, are domi-nated by dissolution of vitric tuffs but are modified byinfiltration through zeolitized tuffs.

The composition of J-13 well water not only plots inthe field of the Rainier Mesa interstitial and fracture-flowwaters, but analyses also lie in the middle of composi-tional ranges for ground water pumped from several wellsthroughout Yucca Mountain originating from variousdepths in the saturated zone (Ogard and Kerrisk, 1984)at temperatures rarely exceeding 40 'C. In the Yellow-stone hydrothermal environment, downhole water com-positions at 140 "C show appreciably higher concentra-tions of SiOr, HCOt, Na, and K, and lower Ca contentsthan the lower temperature J-13 well water (31 "C). Cli-noptilolites there are enriched in K and Na but are de-pleted in Ca (Keith et al., 1978; Sturchio et al., 1989).These results, and the experimental observations de-scribed earlier, clearly show that zeolitized rhyolitic tufsaffect ground-water chemistry, and they suggest that com-positional variability of clinoptilolites influence and areinfluenced by ground-water compositions. They also in-dicate that potassic clinoptilolites are stable to at least140 .c .

Recently attempts were made to measure compositionsof pore water extracted by triaxial compression of non-welded tuffs from the unsaturated zone at Yucca Moun-


tain (Yang et al., 1988). The samples of extracted waterhave much higher Ca, Mg, K, Sr, Cl , SOI , and SiO,concentrations than J-13 well water, and the composi-tions ofextracted water are influenced by changes ofaxialstress on the compressed tuffs. Unfortunately, pH,HCOt, and dissolved Al data were not reported for thesetuff samples. Consequently, the pore-water compositionscould not be used in our thermodynamic calculations.

Representation of activity diagrams

A number of activity diagrams are presented here, ini-tially for simple three-component systems and subse-quently for the more complex multicomponent systemsthat are required for depicting stability fields of clinop-tilolite in felsic rocks. The primary focus is on the cli-noptilolite composition highlighted earlier for whichthermodynamic data are available (Hemingway and Ro-bie, 1984). Although the chosen three-component sys-tems serve as a necessary reference for establishing sta-bility fields for clinoptilolite in the multicomponentsystems discussed later, they also may contain useful in-formation for diagenetic reactions involving other zeo-lites in felsic and mafic rocks.

Systems of three and four components plus HrO canbe readily represented in two dimensions. For example,in the system Ca-Al-Si-HrO, two components are selectedfor the x and y axes, all reactions are balanced with re-spect to a third component, and the activity of HrO isfixed, commonly equal to one. A reaction between amor-phous silica and heulandite can be represented by writinga hvdrolvsis reaction for each mineral:

SiO2("-): SiOr,"a

CaAlrSirO,r.6HrO + 8H*

: ca2+ + 2Al3+ + 7SiO2("q) + l0HrO (2)

Combining Reactions I and 2 and eliminating SiOr,",from the equations gives

CaAlrSirO,r.6HrO + 8H*: 7SiOr,.-) + Ca2+ + 2Al3- + lOHrO (3)

An equilibrium constant as a function of pressure andtemperature can be calculated from thermodynamic datafor this reaction and is expressed as

log K: 2log(ao,t-/(ar.)3) + log(a."2-/(as-)2) (4)

If the x and y axes are chosen as log(a^t-/(a,--)3) andlog(a*z-/(a"-)2), respectively, Reaction 4 is the equationof a line with a slope of -2 and a y intercept of log Kthat forms the boundary on an activity diagram betweenamorphous silica and heulandite (see Fig. 3a, discussedlater). Similar calculations are performed for all mineralpairs, and the resulting intersecting lines form the bound-aries of the phases that appear on the stability diagramspresented here.

Systems with more than four components plus HrO arecalculated in a similar manner, but with the inclusion ofany additional minerals assumed to be at saturation to

constrain additional components. For example, the sys-tem Ca-Na-A1-Si-HrO might have Ca and Na on the axes,be balanced on Al, and have coexisting amorphous silica,as in Reaction 3. Alternatively, this four-component sys-tem could be balanced with respect to SiO, and have theAl component constrained by a saturation phase such asgibbsite:

CaAlrSiTOr8.6HrO + 2H*: 2A(OH)3 + 7SiO2("q) * Ca2. + 4HrO (5)

Gibbsite, however, provides a theoretical maximum ac-tivity of the Al3* component that may not be desirable inall circumstances. Saturation with respect to any Al-bear-ing mineral in the four-component system can be as-sumed, although, if the chosen saturation phase includesthe components plotted on the axes of the diagram, it willchange the topology of the other fields. At Yucca Moun-tain, drill-core samples in the vadose zone contain opaland smectite as coexisting authigenic SiO, and Al3--bear-ing phases, respectively, with some authigenic potassiumfeldspar and minor amounts of cristobalite, qlrartz, andcalcite (Broxton et al., 1987; Bish, 1989). These phases,together with the composition of J-13 well water (Table4), serve to define the ranges of silica and Al3- activitiesshown in Figure 2, which were used to construct the ac-tivity diagrams presented here. Thus, lines labeled D andH on Figure 2 represent the extremes of dissolved silicasaturation limits corresponding to quartz and amorphoussilica, respectively; cristobalite has an intermediate sat-uration level approximated by line E; line F correspondsto coexisting kaolinite and pyrophyllite; and line G is theactivity of dissolved SiO, in J-13 well water. Similarly,for dissolved A13., lines A, B, and C correspond to satu-ration values for coexisting amorphous silica plus pyro-phyllite, coexisting pyrophyllite plus kaolinite, and gibbs-ite, respectively, add line I represents an arbitrarily lowvalue of dissolved Al, which, as shown later, is consistentwith the coexistence of opal, calcite, and clinoptilolite.Note that, although gibbsite, kaolinite, and pyrophyllitehave not been reported at Yucca Mountain, their idealend-member compositions serve as useful constraints fordissolved Al3. and SiO, activities. Smectite, however,which is pervasive at Yucca Mountain (Bish, 1989), hasa sheet silicate structure resembling pyrophyllite. If smec-tite were used to constrain Al activities, its value wouldfall into the general range for pyrophyllite, depending onthe smectite composition. The reported analysis of Al inJ-13 well water, 0.012 mg,/L, which, as noted earlier, isone ofthe lowest concentrations reported for ground waterat Yucca Mountain (Ogard and Kerrisk, 1984), is toohigh to be equilibrium-controlled. The fluid speciationprogram used to calculate cation activities (rq3Nn ofWolery, 1983) indicates that at the measured pH of J-13well water (-7.5), dissolved Al occurs predominantly asA(OH);, with a calculated equilibrium value for [Al3-] of-2 x l0-" M. Using an activity coefficient of -0.6 forAlr gives a value for log(c^r-l(a"r)3) of 9.6, significantlyin excess of the gibbsite saturation value of approximate-

( l )

ly 7.9 (Fig. 2). This value is inconsistent with the equilib-rium coexistence of J-13 well water and any of the ob-served mineral assemblages at Yucca Mountain and otherareas for which dissolved Al concentration data are avail-able, and it cannot be used to constrain calculations pre-sented here. As noted earlier, a possible interpretation ofthis discrepancy is that the Al in J-13 well water containsunfiltered particulate matter that passes through mem-brane filters (Daniels et al., 1982). In calculating the ac-tivity diagrams, the activity of HrO is taken to be unity,and the calcite boundary is added to appropriate dia-grams by assuming a dissolved HCOt content equivalentto that of J- 13 well water (Table 4) and using the 90 "Cand 150 "C analytical data in the activity diagrams cal-culated at 100 .C and 150 oC, respectively.

Rnsur,rsDiagrams of systems of three components plus HrO

A series of diagrams of systems of three componentsplus HrO are shown in Figures 3 to 5 for the systems Ca-Al-Si, Na-Al-Si, and K-Al-Si, respectively. All of thesediagrams are balanced with respect to SiOr. Quartz hasbeen suppressed throughout the calculations describedhere in favor of amorphous silica because opal is thecommonly observed authigenic SiO, phase in zeolitizedash-flow tuffs at Yucca Mountain (Broxton et al., 1986,1987; Bish, 1989). As a result, each of the diagrams inFigures 3-5 has amorphous silica as the stable phase inthe bottom left corner. Amorphous silica occupies a rel-atively smaller stability field than would quartz,hadquartznot been suppressed. Figures 3a-3c illustrate the changesin phase relations for the system Ca-Al-Si with increasingtemperatures at 25, 100, and 200 "C with pressures cor-responding to the steam-saturation curve. Note that theheulandite field at 25 "C is replaced by Ca end-memberphillipsite at higher temperatures. The scolecite field de-creases in size with increasing temperature, and this zeo-lite is no longer stable at 200 "C (Fig. 3c). The field of Caend-member beidellite apparent at 200 "C may exist atlower temperatures as well but does not appear in Figures3a and 3b, possibly because of inaccuracies in the ther-modynamic data for Ca end-member beidellite or adja-cent phases. The stable limits of these and other activitydiagrams described later are delineated by the dashed lineslabeled gibbsite (or diaspore at 200'C) and calcite. HigherAl or Ca activities can only result from supersaturationof the fluid with respect to these phases. As noted earlier,J-13 well water is unconstrained on the Al axis becauseits measured Al content cannot be reconciled with exist-ing thermodynamic data. It is apparent from Figures 3aand 3b that in the simple Ca-Al-Si system J- I 3 well wateris somewhat undersaturated with respect to calcite at 25oC and would be slightly oversaturated at 100 'C.

Activity diagrams for the system Na-Al-Si are illus-trated in Figures 4a and 4b. The 25 "C diagram (Fig. aa)contains a stability field for mordenite. However, owingto lack of thermodynamic data for mordenite, this zeolite

607

1 2 3 4 5 6 7 t 9log Iaa1r*/(as *)3]

Fig. 2. Ranges of dissolved silica and aluminum activitiesused in the calculations of activity diagrams. Silica activitiescorrespond to amorphous silica (H), J-13 well water (G), coex-isting pyrophyllite-kaolinite-saturated solution (F), cristobalite-saturated solution (E), and quartz-saturated solution (D). Alactivities are those for solutions saturated by assemblages ofpy-rophyllite * amorphous silica (A), pyrophyllite-kaolinite (B), andgibbsite (C) and an arbitrary low value (I).

was not included in high-temperature calculations in Fig-ure 4b. Its absence from Figure 4b is not indicative ofstability limits for mordenite; indeed, mordenite occursat -170 'C in drill holes throughout the Yellowstone geo-thermal system (Keith et al., 1978; Sturchio et al., 1989).Thus, in the absence of mordenite the stability field ofalbite widens at 100 'C and Na end-member beidellitealso becomes stable (Fig. 4b). Figures 5a and 5b showactivity diagrams for the system K-Al-Si at 25 "C and100 "C in which K end-member phillipsite is joined bypotassium feldspar at high temperatures. J-13 well wateris consistent with equilibrium with respect to either mor-denite or albite (Fig. 4) and either K end-member phil-lipsite or potassium feldspar (Fig. 5).

Multicomponent diagrams

Figures 3 to 5 provide simplified activity diagrams usedas references for comparison with the more complex sys-tems of four and five components plus HrO necessary forplotting the stability flelds of clinoptilolite. Figures 6-9are activity diagrams for the system Ca-Na-K-Al-Si. Val-ues oflog(a""-/a"-) andlog(aor,/(ar*)'z) are plotted on thex- and /-axes, respectively, in each diagram. Either Al(Figs. 6 and 7) or Si (Figs. 8 and 9) has been used as thebalancing component. In each case, two additional com-ponents need to be specified. The component K+ is con-strained by assuming the presence ofpotassium feldspar,


g - 3aa2 - 4


4 5 6 7logla l ls+/(an +)31

ol+

+dC

at!!c

N

+

6t

+d6

U6to0

Fig. 3. Activity diagrams for the system Ca-Al-Si plus H,Obalanced with respect to Si (a) at 25 t, (b) at 100 .C, and (c) at200'C and 15.5 bars.

since it occurs as an authigenic mineral (Broxton et al.,1987) and as a phenocryst and groundmass mineral inthe rhyolite tuffs at Yucca Mountain. The other compo-nent, either Al3t or SiOr, is assigned a series of valuessuch as those shown in the plot oflog(a^t-/(a".)3) versuslog 4r,o, aI 25 "C in Figure 2. The activity diagrams inFigures 6a-6d are balanced with respect to Al at 25 "Cand have log 4116, specified by the lines labeled H, G, F,E, and D in Figure 2. The activity diagrams in Figures7a and 7b are balanced with respect to Al with amor-phous silica and quartz saturations, respectively, at 100

"C. Figures 8a-8d are balanced with respect to Si at 25'C and have Al3. concentrations constrained by valuescorresponding to lines labeled I, A, B, and C, respectively,on Figure 2. The diagrams in Figures 9a-9c are balancedwith respect to Si at 100 'C with Al constrained by py-rophyllite-amorphous silica, kaolinite-pyrophyllite, andgibbsite saturations, respectively. The 150 "C diagramshown in Figure 9d is also balanced with respect to Siwith Al constrained by pyrophyllite-amorphous silica.Clinoptilolite is included in each of the diagrams, wherestable, by considering it to be in equilibrium with Ca end-member saponite (smectite) and hematite to account for

grossular i t")

m a r g a r i t e {

I

scoleclI

I

I

. l

f l or ' E

aI t l- i € - -

- o !

e pyrophyltitc

6 7

m a r g a r i t e

Ca.phi l l ipsite

t 2 3 4l o g l a l l r + / ( a x + ) 3 I

a r g a r i t

Ca-phillipsite

amorphous silicaF ca-beidel l i te

l l

;

I kaoliniteII

€ pyrophyllite

9al

+ EI

J 7Qg 6!o

5

- 2 - l 0 ll o g l a a r r + / ( a u + ) 3 1

BOWERS AND BURNS: ACTMTY DIAGRAMS FOR CLINOPTILOLITE

- 2 - 1 0 1 2 3 4 5 6log laA l3+ / (aH+)31

Fig. 4. Activity diagrams for the system Na-Al-Si plus HrO balanced with respect to Si (a) at 25 {; and (b) at 100 €.

609

l 6

15

L4

l3

l2

+ l l

I t oz

Ss

1 2 3 4 5 6 7 8 9 1 0log [aa l r+ / (aga )3 ]

the small amounts of Mg and Fe found in the ctinoptil-olite specimen measured by Hemingway and Robie(1984). As noted earlier, limited thermodynamic data formordenite enable its stability field to be shown on dia-grams at 25 lC only. On all of the activity diagrams shownin Figures 6-9, calcite is plotted with a dashed line byassuming a bicarbonate ion content comparable to J-I3well water, and the circular symbol labeled J-13 cone-sponds to Ca and Na activities of this reference groundwater.

By comparing the activity diagrams shown in Figures6-9 through changing temperature, or for different activ-ities of Al or Si, trends in the relative stabilities of variouszeolite phases may be recognized easily. For example,Figure 6 shows that (l) clinoptilolite is stable at high ac-tivities of silica and its field decreases with decreasingsilica activity, (2) the heulandite field also diminishes withsilica activity but still remains when cristobalite is pres-ent, correlating with heulandite + cristobalite assem-blages found in some drill cores at Yucca Mountain, (3)the mesolite stability field, which exists at low activitiesof silica, narrows and then disappears with increasing sil-ica activity, and (4) at low silica activities, scolecite andalbite are stabilized relative to heulandite and mordenite,respectively.

The effects of temperature can be seen by comparingFigure 6a (solid lines) with Figure 7a (corresponding toamorphous silica saturation) and Figure 6d with Figure7b (quartz saturation), where it is apparent that both cli-noptilolite and mesolite have smaller regions of stabilityat 100 oC than at 25 oC, and that Ca end-member phil-lipsite replaces heulandite at low activities ofsilica.

The efects of variable Al activity at 25 'C can be ob-served in Figure 8. Clinoptilolite is not stable at high Alactivities corresponding to gibbsite saturation (Fig. 8d)but appears with decreasingAl activities (Figs. 8a-8c). Itsstability field maximizes at an intermediate Al activityconstrained by the coexistence of amorphous silica andpyrophyllite (cf. smectite) (Fig. 8b) and then becomessmaller with further decrease in Al activity (Fig. 8a). Sim-ilar less conspicuous trends exist for heulandite. The sta-bility field of mesolite, on the other hand, increases withrising Al activity and maximizes at gibbsite saturation,where the clinoptilolite, heulandite, mordenite, and phil-lipsite fields disappear (Fie. 8d). Note that circles repre-senting Ca and Na concentrations of J-13 well water plotclose to the join of mordenite, clinoptilolite, and amor-phous silica (Fig. 8a), and with heulandite, mordenite,and clinoptilolite (Fig. 8b), which are consistent withmineral assemblages observed at Yucca Mountain. Ef-

U

m o r d e n i t e

I

l k a o l i n i t eII

! 7

; 6z6'o ! -o 5

paragonite

k a o l i n i t e

pyrophyll ite

II

(a)

kals i l i te

o

u

muscoYite

amorphous silica

I

I kaoliniteI

e pyrophyllite


t2

l l

10

= 7

! . 6

-Ss

+

+v

o0

1 2 3 4 5 6 7 Elog la l t l + / (an+ )31

fects of temperature may be seen in Figures 9a-9c, whichshow that the clinoptilolite and heulandite stability fieldsdecrease with increasing temperature and appear only atlow Al activities. At 150 "C with Al activity correspond-ing to coexisting pyrophyllite and amorphous silica, cli-noptilolite, but not heulandite, still has a small stabilityfield (Fig. 9d). Thus, alkali-rich clinoptilolites are expect-ed to persist to at least 150'C, as indicated by their oc-currence in drill holes at Yellowstone (Keith et al., 1978;Sturchio et al., 1989), but under a significantly reducedrange of Al activities.

In Figures lOa and lOb, the system Ca-Na-K-Al-Si isrepresented at 25'C with log(4K-)/(4".) replacing log(a"")/(a"-) on the x-axis and albite replacing potassium feld-spar as the saturation phase. Figure lOa is balanced onAl, and amorphous silica is the saturation phase, whereasFigure lOb is balanced on Si with Al3t activity controlledby coexisting amorphous silica plus pyrophyllite. Thesetwo representative activity diagrams based on K* activ-ities are very similar to their Na-counterparts, except thatK--end-member-silicate phases replace Na<nd-mem-ber-silicate minerals. The stability field of clinoptiloliteis again largest at high silica activities, intermediate Al3.activities, and low temperatures.

1 2 3 4l og [a a ;s + / (a s + ;3 ]

Activity diagrams for clinoptilolites of variablecompositions

As noted earlier, clinoptilolite compositions vary con-siderably at Yucca Mountain (Fig. l)' Therefore, activitydiagrams were calculated for clinoptilolites of variableNa-Ca, K-Ca, and Ca-Mg contents (Fig' l), and the re-

sults are shown in Figure I l. The stability fields are iden-tified in Figure I la, which is related to the activity dia-gram in Figure 8b, where Si is balanced and Al activitiesare constrained by the assemblage pyrophyllite + amor-phous silica * potassium feldspar, which corresponds toconditions of maximum clinoptilolite stability at 25'C.

Figure llb shows that with increasing atomic substi-tution of Na in clinoptilolite, the clinoptilolite stabilityfield narrows. Conversely, the clinoptilolite stability fieldwidens for clinoptilolites with higher Ca contents andmerges into the heulandite field. Clinoptilolites more so-dic than Na" ruCa' 5 are no longer in equilibrium with J-13 well water, suggesting that ground water with higherNa and lower Ca concentrations, perhaps derived fromaltered vitric tuffs (White et al., 1980), is necessary tostabilize sodic clinoptilolites.

K has the opposite effect on the clinoptilolite stability

Fig. 5. Activity diagrams for the system K-Al-Si plus HrO balanced with respect to Si (a) at 25 "C; and (b) at 100 "C.

muscovite

amorphous silicaII

t kaoliniteI

e pyrophyllite


grossular

heulandite

J-

8

5

4

t

0 8'-4

l og [ap1+/8s+] log[ap1+/an+l

(c)

grossular

heulandite

calcitej-nO,

mordenite

kaolinitc

2 4 6 8 l 0 1 2 1 4 l 6 l uloglana+/an+] l og [eNs+ /aH+ ]

Fig. 6. Activity diagrams for the system Ca-Na-K-Al-Si plus H,O balanced with respect to Al at 25 t fot ditrerent silicaactivities: (a) amorphous silica and J- I 3 well water (dotted lines), (b) pyrophyllite-kaolinite, (c) cristobalite, and (d) quartz. Saturationphases also include potassium feldspar, hematite, and Ca end-member saponite.

6 l l

12

10

GI

+

ct

+daU

!a

l8

t5cl

; 1 4

* 1 2I

ag* to

mordenitcmordenite

24

22

20

It

t6

0

24

22

20

18

l6

6

4

,

cl

+

(l

d

I6t

o!

t4

t2

l0

t

6

4

at

?

d

I6l

o0


24

22

20

t2

10

d

+

6t

d

6

(J

oo

GI

+

6l

+d

O

o0

6 t l 0 1 2l og [aNs+ /aH+ ]

field (Fig. I lc), which widens considerably with increas-ing atomic substitution of K for Ca. Clinoptilolites lesspotassic than KorrCa, , would no longer be in equilibriumwith J-13 well water. In contrast, replacement of Ca byMg in clinoptilolite displaces its stability field to lowerCa activities (Fig. l ld). Clinoptilolites more magnesianthan Ca,rMg,o, would not be in equilibrium lvith J-13well water.

DrscussroN

The activity diagrams demonstrate that the formationof clinoptilolite is favored by SiO, activities higher thanallowed for by the presence of quartz. This is demonstrat-ed by Figure 6 and is achieved, for example, when cli-noptilolite coexists with opal in diagenetically alteredvolcanic glasses. Heulandite, however, can exist in thepresence of opal and cristobalite but not quartz. Suchassemblages are commonly observed in vitric tuff sam-ples from drill cores at Rainier Mesa and Yucca Moun-tain (Benson, 1976; White et al., 1980; Broxton et al.,1987) and from surface desert pavement and outcrop lo-cations (Burns et al., 1990).

Clinoptilolite has a maximum stability field at an in-termediate Al activity value but shrinks with either in-

(b)

grossular

rehnite

Ca-phillipsite / /nesolite/ /

J'13 C\/ / calcite- - v

paragonite

kaolinite

6 t 1 0 1 2log[aN6+/an+]

creasing or decreasing activities of Al. This is indicatedby Figure 8, in which the clinoptilolite stability field islargest when Al activities are controlled by the assem-blage amorphous silica + pyrophyllite (cf. smectite) (Fig.8b). Furthermore, since the composition ofJ-I3 well waterappears to be approximately in equilibrium with respectto calcite, the J-13 Ca2*/Na* points plotted in Figures 8band 8c suggest that if J-13 well water is in equilibriumwith clinoptilolite, then the Al activities should lie be-tween the values for kaolinite-pyrophyllite and pyro-phyllite-amorphous silica. Such Al activities also indi-cate that J-13 well water could be in equilibrium withother zeolites represented on the activity diagrams, in-cluding heulandite and mordenite (Fig. 6b) and possiblymesolite (Figs. 8c and l0b), particularly when the stabil-ity field ofNa-rich clinoptilolites is diminished (Fig. I lb).These observations for calcic clinoptilolite, heulandite,and mordenite correlate with the occurrence of these zeo-lites as fracture linings in welded and devitrified tuffs andin basal vitrophyres in the vadose zone throughout YuccaMountain (Broxton et al., 1987; Levy, 1984). Chabaziteoccurring with coatings offine-grained heulandite in frac-tured vitrophyre below the water table in the J-13 well(Carlos, 1989) does not appear to have a stability field

t6

l6l4

Fig. 7. Actiwity diagrams for the system Ca-Na-K-Al-Si plus HrO balanced with respect to Al at 100 € for different silicaactivities: (a) amorphous silica, and (b) quartz.

y r o p h y l I i t e

heulandite

____r : r1O_

clinoptilolite

w o l l a s t o n I t e

calcite

natrolite

phillipsite

mordenite

clinoptil$1

a m o r p h o u ss i l i c a

t2

tl

+

+da

c!0c

G|

+

6t

+d

C

6t

!0

BOWERS AND BURNS: ACTIVITY DIAGRAMS FOR CLINOPTILOLITE 6 1 3

6 t 1 0 t 2log[8xa+/8n+]

6 8 1 0 1 2log[aira+/aH+]

Fig. 8. Activity diagrams for the system Ca-Na-K-Al-Si plus H,O balanced with respect to Si at 25 € for different Al activities:(a) low Al activity corresponding to line I in Fig. 2, (b) pyrophyllite-amorphous sitica, (c) kaolinite-pyrophyllite, and (d) gibbsite.

Saturation phases again include potassium feldspar, hematite, and Ca end-member saponite.

6|+

d

L)

EO

heulendlte

m o r p n o u s

p y r o p h y l l i t e

clinoptilolite

l|

+

6

+d6

6t

!0

6 8 1 0 1 2log[ap1+/a u +]

II

kao l i n i t e l

y r o p h y l l i t e

6t4 BOWERS AND BURNS: ACTIVITY DIAGRAMS FOR CLINOPTILOLITE

Fig. 9. Activity diagrams for the system Ca-Na-K-Al-Si plus HrO at elevated temperatures balanced with respect to Si fordifferent Al activities: (a) pyrophyllite-amorphous silica at 100 'C, (b) kaolinite-pyrophyllite at 100'C; (c) gibbsite at 100 "C, and(d) pyrophyllite-amorphous silica at 150 t.

GI

+

cl

d

a

EO

al+

€+da

6t

00c

aa+

6

+da

oE]c

5 t l 0 1 2bg[eN1+/an+]

calclte

nephellne

morphous l I ts i l i c e | | |

* | = phillipsite

p y r o p h y l l i t e

gehlenite (b)

_ga l_c{e____

nepEeline

clinoptilotite ll | | e analcine

pyrophy l l

log[apa+/8n+]

t2

10

d

+Ec

+caag€!ao

l0

t

6 t 1 0 1 2lsg[at\1+/aH +]

8bg[aps+/an+t

- - - - g a E i l . - - -

nephcline

1€,-r:r1- -- - - - - + -

J-r3\'Ca-philllpsite

r m o r p n o u csilica +\p y r o p h y l l i t e


2 4 6 tloglax+/aH+l

Fig. 10. Activity diagrams based on K and Ca activities in the system Ca-Na-K-Al-Si plus HrO at 25 'C (a) balanced withrespect to Al with silica saturation by amorphous silica, and (b) balanced with respect to Si with Al saturation by pyrophyllite +amorphous silica. Saturation phases include albite, hematite, and Ca end-member saponite.

6 1 5

d

+d

(.)

oa

16d

i rc

* rzO

-* lo

t

- 2 0 2 4 6 E 1 0 t 2 r 4log [a6a /a6*1

and may be metastable with respect to mesolite or sco-lecite.

The clinoptilolite stability field decreases with increas-ing temperature between 25 and 150 "C (Figs. 7a, 9a, 9b,and 9d) and likely disappears by 200'C. This correlateswith hydrothermal experiments (Boles, l97l; Knauss etal., 1985a, 1985b; Hawkins et al., 1978) and observedgeological occurrences of clinoptilolite (Hay, 1966; Hayand Sheppard, 1977). Potassic clinoptilolites have beenidentified at 120-140 .C in drill cores from the Yellow-stone hydrothermal environment (Keith et al., 1978;Sturchio et a1., 1989), however, correlating with the in-creased stability field of K-rich clinoptilolites (see Fig.I lc) .

Znolite diagenetic zones have been suggested for alter-ation of vitric tuffs based on the appearance and disap-pearance of clinoptilolite in buried pyroclastic deposits(Iijima, 1975, 1980; Smyth, 1982).Zone I, forexample,is characterized by large-scale preservation ofglass in vit-ric tuffs above the water table and incipient alteration ofglass shards, particularly in groundmass, to smectite andopal. The Topopah Spring Member at Yucca Mountain,lying well above the water table, falls into Zone I. How-

ever, Ca-rich clinoptilolites occur in fractures throughlower welded tuff and vitrophyre horizons of the Topo-pah Spring Member and may be indicative of ground-water interactions, perhaps with microcrystalline devitri-fied tuffs, which produce relatively high concentrations ofdissolved Ca2*, Na*, and HCO; in fracture-flow water(White et al., 1980). The activity diagrams calculated for25 "C consistently show that calcic clinoptilolites are sta-ble in the presence of fracture-flow J- I 3 well water orig-inating from microcrystalline devitrified Topopah SpringMember tuffs, even though such zeolites have not beenobserved as fractureJining minerals at this level in J-13drill cores (Cados, 1989). The abundance ofdrusy quartzcoating fractures atthat level (Carlos, 1989) may depressthe silica activity below that necessary to crystallize cli-noptilolite (and heulandite).

Diagenetic ZnneII, which characterizes the bedded tuffsbelow the Topopah Spring Member, represents extensivezeolitization of vitric tuffs to clinoptilolite-bearing assem-blages and is promoted by water enriched in alkali cationsand by slightly elevated temperatures (Smyth, 1982). Pro-gressive hydration and dissolution reactions ofthe rhyo-litic vitric tuffs increase the concentrations of SiOr, Na*,

0_i

clinoptilolite

K-phillipsite

a m o r p h o us i I i ca

y r o p h y l I i t e

clinoptilolite

crlcltc

hculendlte J'

cllnoptllolltc

6 r 6 BOWERS AND BURNS: ACTIVITY DIAGRAMS FOR CLINOPTILOLITE

a\

+

a

+ala

!l

!0

6l

+

G

+CIqt

(J6

00

e{+tr6

+clrt

6t

00

- 4 0 4 tl og [ep1+/e g + ]

and ultimately K* in ground water (White et al., 1980)from which clinoptilolite-clay silicate-opal assemblagesare derived. The presence of Ca-poor, K-Na-rich clinop-tilolites in diagenetic Zone lI confonns with the activity

- 4 0 4 tlog[a pg +/a s + ]

- 4 0 4 El o g [ a N s + / a H + l

diagrams, which consistently show the clinoptilolite sta-bility field moving away from J-13 well-water composi-tions at elevated temperatures, as well as by increasedNa, but depleted Ca, concentrations in ground water.

GI

c

+6ln

ca0

- 4 0 4 tlog[a pg +/a H + ]

Fig. I I . Activity diagrams for clinoptilotites having variable cation compositions. The calculated stability fields at 25 € cor-respond to aluminum saturation by pyrophyllite plus amorphous silica and are balanced with respect to Si (compare Fig. 7b). (a)reference diagram for the clinoptilolite composition Na" rul$ rrCa, ,oMg, ,r; (b) Ca-Na variations: -. . -. '- : NaoCa, r., -'-'- : Nao rrCa, un,-: NaoruC?rs, ----- : Na,ruCar6, NarruCaor. (c) K-Ca variat ions: Kr,orCaot, I(2e8caoi, - ' - ' - :

K , r r C a , o , - : K { s c a r 5 , - - - : I G r C a , 7 4 , . . . . . : & C a , n r . ( d ) C a - M g v a r i a t i o n s : . . . ' . : C a r r r M g o , - - - - - : C a r . r M & . r r , - . - ' -: CaroMgorr, Ca, ,Mg, or, - : Ca*Mg, ,r, Cao trMg2o.

N ooC a r .z t

o9"'-'"I::;F

*?:'i- i . .

.gt'.,"ar$

BOWERS AND BURNS: ACTIVITY DIAGRAMS FOR CLINOPTILOLITE 6 t l

Deeper drill cores through Yucca Mountain have yield-ed analcime instead of clinoptilolite, which is indicativeofdiagenetic ZonelII, whereas ZoneIY is represented bythe breakdown of analcime to albite. It has been suggest-ed that the clinoptilolite-analcime transition depends onthe concentration of dissolved Na, and the Zonell4oneIII boundary was estimated to lie between 100 and 150"C for present-day ground-water compositions at YuccaMountain (Smyth, 1982). However, Kerrisk (1983) con-cluded from reaction-path calculations that the clinoptil-olite-analcime transition is controlled by the activity ofdissolved silica. Mordenite coexists with clinoptilolite inzeolitized tufs at Yucca Mountain. Textural evidence(Sheppard et al., 1988) suggests that some mordeniteshave formed at the expense of clinoptilolite during late-stage dissolution, and this observation is consistent withour activity diagrams. However, mordenite also persistsat higher temperatures (- 170'C) than clinoptilolite (- 140'C) in drill holes through the Yellowstone hydrothermalenvironment (Keith et aI., 1978; Sturchio et al., 1989),but this observation cannot be correlated with activitydiagrams calculated at elevated temperatures because oflack of thermodynamic data for mordenite. The inabilityto include mordenite in such activity diagrams may beresponsible for the apparent stability fields of albite orphillipsite abutting the clinoptilolite field instead of an-alcime in the diagrams calculated at 100 "C (Figs. 7a and9a) and 150 "C (Fie. 9d).

Effects of temperature on the calcic clinoptilolite andheulandite stability fields, together with occurrencesof K-Na-rich clinoptilolites at elevated temperatures(Sturchio et al., 1989), suggest that calcic zeolites are mostvulnerable to thermal decomposition (cf. Figs. 8a and 8d).Such Ca-rich zeolites occur in fractures (Carlos, 1985) andin the vitrophyre at the base of the Topopah Spring Mem-ber tuffat Yucca Mountain (kvy, 1984). The density offractures is highest in the densely welded tuffhorizon inthe Topopah Spring Member (Scott et al., 1983) wheresiting ofthe proposed nuclear waste repository is planned.In the vicinity of the heat envelope that would surroundradioactive waste buried there, fluid temperatures couldexceed 150 "C within a radius of l0 m from the proposedrepository horizon (Travis et al., 1984; U.S. DOE, 1988,p. 4-l l8), and thereby affect the thermal stability of cal-cic zeolites in the fractures and underlying basal vitro-phyre.

Several observed reactions suggested by phase assem-blages in weathered vitric tuffs (Benson, 1976; White etal., 1980; Burns et al., 1990) can be demonstrated on theactivity diagrams. For example, the reaction of glass *clay silicates to clinoptilolite + opal plots at the intersec-tion of amorphous silica + pyrophyllite * mordenite *clinoptilolite in Figures 6a and 8b, but requires lowercalcium activities in the coexisting fluid than that of J- I 3well water. Low Ca and slightly reduced K activities ofrainwater would account for the assemblage glass * claysilicates * opal + clinoptilolite * authigenic potassiumfeldspar forming on weathered vitric tuffs at outcrops and

in detritus forming desert pavement (Burns et al., 1990).The assemblage of clinoptilolite * calcite * opal alsofound in weathered vitric tuffs is represented on Figure8a, requiring, however, very low activities of A1.

CoNcr,usroNs

The calculated activity diagrams presented here quan-tifi, observed field occurrences and verifu deductions madeabout the stability of clinoptilolite during burial diagen-esis of vitric tuffs. The coexistence of clinoptilolite withopal correlates with its calculated wide stability field inaqueous solutions saturated with amorphous silica. Cli-noptilolite-smectite assemblages indicate that the zeolitecrystallized from ground water with dissolved Al concen-trations lower than saturation values with respect togibbsite and lower than the reported measurements ofJ-13 well water. Calcic clinoptilolites associated with cal-cite are consistent with crystallization from fracture-flowground water containing Ca2* and HCO; derived fromincipient dissolution of microcrystalline devitrified tufs.Alkali-rich clinoptilolites, on the other hand, correlatewith ground water having elevated Na* and K* but de-pleted Ca2+ concentrations, which are associated with al-tered vitric tuffs. Although the crystallization of clinop-tilolite may be promoted by ground water enriched insilica and alklai metals, the clinoptilolite stability fielddiminishes appreciably between 25 and 150 "C, correlat-ing with burial diagenetic reactions, and confirming doubts(Smyth, 1982) about the thermal stability of calcic cli-noptilolites close to buried radioactive waste. Questionsremain, however, concerning the stability of calcic versusK-Na-rich clinoptilolites at elevated temperatures and theeffects of compositional variations in heulandite andmordenite. Recently revised thermodynamic propertiesof heulandite (Johnson et al., 1989) and limited infor-mation on thermal stability limits of mordenite (Kusaka-be et al., l98l) may constrain our ongoing calculationsof zeolite activity diagrams.

AcxNowlnocMENTs

We gratefully acknowledge the assistance of Drs. M. E. Morgenstein,C L. Johnson, and D. L. Shettel, who reviewed early drafts of the manu-script and provided information about ground-water chemistry at YuccaMountain. Helpful comments on the original manuscript were receivedfrom Drs. J. A. Apps, D. L. Bish, J. G. Liou, W. M. Murphy, J. K. Bohlke,and an anonymous reviewer. Dr. J. F. Kerrisk kindly provided his reporton clinoptilolite stability field calculations. The study was funded by theState of Nevada, Nuclear waste Project Ofrce, under a Department ofEnergy grant from the Nuclear Waste Fund.

RnrnnrNcrs cltnoBenson, L.V. (1976) Mass transport in vitric tuffs of Rainier Mesa, Nye

County, Nevada. U.S. Energy Research and Development Administra-tion, Nevada, Report I.IVO-1253-10.

Bish, D.L. (1989) Evaluation ofpast and future alterations in tuffat YuccaMountain, Nevada, based on the clay mineralogy of drill cores USWG-1, G-2, and G-3. U.S. National Technical Information Service, Re-port LA-10667, 40 p.

Bish, D.L., Ogard, A.L., Vaniman, D.T., and Benson, L. (1984) Miner-

6 1 8

alogy-petrology and groundwater chemistry ofYucca Mountain Tufs.Materials Research Society Synposium Research hoceedings, 26,283-2 9 t .

Boles, J.R. (1971) Synthesis ofanalcime from natural heulandite and cli-noptilolite. American Mineralogist, 56, 17 24-17 34.

Bowers, T S., Jackson, K.J., and Helgeson, H.C. (1984) Equilibrium ac-tivity dia$ams for coexisting minerals and aqueous solutions at pres-sures and temperatures to 5 kb and 600'C. Springer-Verlag, New York.

Broxton, D.E., Bish, D.L., and Warren, R.G. (1987) Distribution andchemistry of diagenetic minerals at Yucca Mountain, Nye County, Ne-vada. Clays and Clay Minerals, 35, 89-l 10.

Broxton, D.E., Warren, R.G., Hagan, RC., and Luedemann, G. (1986)Chemistry of diagenetically altered tuffs at a potential nuclear wasterepository, Yucca Mountain, Nye County, Nevada. U.S. NationalTechnical Information Service, Report LA-10802, 160 p.

Broxton, DE., Byers, F.M., Jr., and Warren, R.G. (1989) Petrographyand phenocryst chemistry of volcanic units at Yucca Mountain, Ne-vada: A comparison of outcrop and drill hole samples. U.S. NationalTechnical Information Service, Report LA- I I 503, 65 p.

Burns, R.G., Bowers, T.S., Wood, V.J., Blundy, J.D., and Morgenstein,M.E. (1990) Reactivity of zeolites forming in vitric tuffs in the unsatu-rated zone at Yucca Mountain, Nevada. In Proceedings ofnuclear wasteisolation in the undersaturated zone, Focus '89, p. l0l-1 12. AmericanNuclear Society, La Grange Park, IL.

Carlos, B.A. (1985) Minerals in fractures of the unsaturated zone fromdrill core USW G-4, Yucca Mountain, Nye County, Nevada. U.S. Na-tional Technical Information Service, Report LA-10415, 55 p.

- (19E9) Fracture-coating minerals in rhe Topopah Spring Memberand Upper Tuff of Calico Hills from drill hole J-13 U.S. NationalTechnical Information Service, Report LA-l 1504, 20 p.

Chen, C.-H. (1975) A method of estimation of standard free energies offormation of silicate minerals at 298.15 "K. American Journal of Sci-ence,275,801-817.

Daniels, W.R., et al. (29 coauthors) (1982) Summary report on the geo-chemistry of Yucca Mountain and environs. U.S. National TechnicalInformation Service, Report LA-9328.

Delany, J.M. (1985) Reaction of Topopah Spring tutrwith J-13 water: Ageochemical modeling approach using the EQ3/6 reaction path code.U.S National Technical Information Service, Report UCRL-53631.

Hawkins, D.B., Sheppard, R.A., and Gude, A J., nI (1978) Hydrothermalsynthesis ofclinoptilolite and comments on the assemblage phillipsite-clinoptilolite-mordenite. In L.B Sands and F.A. Mumpton, Eds., Nat-ural zeolites, p. 337-349. Pergamon Press, New York.

Hay, R.L. (1966) Zeolites and zeolite reactions in sedimentary rocks. Geo-logical Society ofAmerica Special Paper 85, p. l3G-146.

Hay, R.L., and Sheppard, R.A. (1977) Zeolites in open hydrologic sys-tems. In F.A. Mumpton, Ed., Mineralogy and geology of natural zeo-lites. Mineralogical Society of America Reviews in Mineralogy, 4, 92-102.

Hemingway, B.S., and Robie, R A (1984) Thermodynamic properties ofzeolites: Low temperature heat capacities and thermodynamic func-tions of phillipsite and clinoptilolite. Estimates of the thermochemicalproperties of zeolitic water at low temperatures. American Mineralo-gist, 69, 692-700.

Hoover, D.L. (1968) Genesis of zeolites, Nevada Test Site. GeologicalSociety of America Memoir 110,275-284.

Iijima, A. (1975) Effect of pore water to clinoptilolite-analcime-albite re-action series. Journal of the Faculty of Science, University of Tokyo,Series II, 19.133-147

-(1980) Geology of natural zeolites and zeolitic rocks. In L.V.C.Rees, Ed , Proceedings of the Fifth Intemational Conference on Zeo-lites, p. 103-1 18. Heyden & Co., London.

Johnson, GK., Flotow, H.E., O'Hare, P.A.G., and Wise, W.S. (1982)Thermodynamic studies of zeolites: Analcime and dehydrated anal-cime American Mineralogist, 67, 736-748.

-(1983) Thermodynamic studies of zeolites: Natrolite, mesolite, andscolecite. American Mineralogist, 68, I l3,t-1145.

-(1985) Thermodynamic studies of zeolites: Heulandite. AmericanMineralogist, 70, 1065-107 L


-(1989) Thermodynamic studies of zeolites: Heulandite: Erratum.

American Mineralogist, 74, 697.Keith, T.E.C., White, D 8., and Beeson, M H. (1978) Hydrothermal al-

teration and self-sealing in Y-7 and Y-8 drill holes in northern part of

Upper Geyser Basin, Yellowstone National Park, Wyoming. U'S. Geo-logical Survey Professional Paper 1054-A, A'l-426.

Kerrisk, J.F (l 983) Reaction-path calculations of goundwater chemistry

and mineral formation at Rainier Mesa, Nevada. U.S. National Tech-

nical Information Service, Report LA-9912-MS, 41 p.

Kerrisk, J.F. (1987) Groundwater chemistry at Yucca Mountain, Nevada,

and vicinity. U.S. National Technical Information Service, Report LA-

10929,1r3 p.Knauss, K.G., Beiringer, W.J., and Peifer, D.W. (1985a) Hydrothermal

interaction of crushed Topopah Spring tuf and J-1 3 water at 90, 1 50,

and 250 € using Dickson-type, gold-bag rocking autoclaves. U.S' Na-

tional Technical Information Service, UCRL-53630,27 p.

K-nauss, K.G., Delany, J.M., Beiringer, W.J., and Peifer, D.W. (1985b)

Hydrothermal interaction ofTopopah Spring Tuffwith J-13 water as

a function oftemperature. Materials Research Society Symposium Pro-

ceedings, 27, 539-546.Kusakabe, H., Minato, H., Utada, M., and Yamanaka, T. (1981) Phase

relations of clinoptilolite, mordenite, analcime and albite with increas-

ing pH, sodium ion concenffation and temperature. Scientific Papers

ofthe College ofGeneral Education, University ofTokyo, 31, 39-59'

krry, S.S. (1984) Studies ofaltered vitrophyre for the prediction ofnu-

clear waste repository-induced thermal alteration at Yucca Mountain,Nevada. Materials Research Society Symposium Proceedings, 26, 959-

966.Moncure, G.K., Surdam, R.C., and McKague, H.L. (1981) Zeolite dia-

genesis below Pahute Mesa, Nevada Test Site. Clays and Clay Minerals,29,385-396

Moore, D.E., Morrow, C.A., and Byerlee, J.D. (1986) High-temperaturepermeability and groundwater chemistry of some Nevada Test Sitetuffs. Joumal of Geophysical Research, 82, 2163-217 l.

Ogard, A.E., and Kerrisk, J.F. (1984) Groundwater chemistry along flowpaths between a proposed repository site and the accessible environ-ment. U.S. National Technical Information Service, Report LA-10188.

Ogard, A.E., Wolfsberg, K., Daniels, W.R., Kerrisk, J., Rundberg, R.S.,and Thomas, K.W. (1984) Retardation of radionuclides by rock unitsalong the flow path to the accessible environment. Materials ResearchSociety Symposium Proceedings, 26, 329-336.

Oversby, Y.M. (1985) The reaction of Topopah Spring tuff with J-13water at 90 'C and 1 50 'C-Samples from drill cores USW G- 1, USWGU-3, USW G-4, and UE-25h#1. U.S. National Technical lnforma-

tion Service, Report UCRL53629,26 p.

Scott, R.8., Spengler, R.W., Diehl, S., Lappin, A.R., and Chornack, M.P.(1983) Geologic character of tufs in the unsaturated zone at YuccaMountain, southern Nevada. In J.W. Mercer, P.S.C. Rao, and I.W.Wendell, Eds., Role ofthe unsaturated zone in radioactive andhazard-ous waste disposal, p. 289-335. Ann Arbor Science, Ann Arbor, Mich-rgan.

Sheppard, R.A., and Gude, A.J., III (1973) Zeolites and associated au-

thigenic silicate minerals in tuffaceous rocks of the Big Sandy forma-

tion, Mohave County, Arizona. U.S. Geological Survey ProfessionalPaper 830.

Sheppard, R.A., Gude, A.J.,III and Fitzpatrick, J.J. (1988) Distribution,characterization, and genesis of mordenite in Miocene silicic tuffs atYucca Mountain, Nye County, Nevada. U.S. Geological Survey Bul-lelin 1777, 22 p.

Smyth, J.R. (1982) Zaolile stability constraints on radioactive waste iso-lation in zeolite-bearing volcanic rocks. Journal ofGeology, 90, 195-201.

Sturchio, N.C., Bohlke, J.K., and Binz, C.M. (1989) Radium-thoriumdisequilibrium and zeolite-water ion exchange in a Yellowstone hydro-thermal environment. Geochimica et CosmochimLc Acta,53, 1025-1034

Thomas, KW. (1987) Summary of sorption measurements performed

with Yucca Mountain, Nevada, tuffsamples and water from \ilell J-13.U.S. National Technical Information Service, Report LA-10960, 99 p'

Travis,8.J., Hodson, S.W., Nuttall, H.E., Cook, T.L., and Rundberg, R.S.(1984) Numerical simulation of flow and transport in fractwed tuf.Materials Research Society S).rnposium Proceedings, 26, 1039-1046.

U.S. Department of Energy (1988) Site characterization plan, YuccaMountain Site, Nevada, research and development area, Nevada, Re-port DOE/RW-0160, 8 vols.

White, A.F., Claassen, H.C., and Benson, L.V. (1980) The effect of dis-solution ofvolcanic glass on the water chemistry in a tuffaceous aquifer,Rainier Mesa, Nevada. U.S. Geological Survey Water-Supply Paper1535 -Q , 33 p .

619

Wolery, T.J. (1983) EQ3NR, a geochemical program for geochemical

aqueous speciation-solubility calculations. U.S. National Technical In-

formation Service, UCRL-5, Distribution Category UC-70.Yang, LC., Turner, A.K., Sayre, T.M., and Montazer, P. (1988) Triaxial-

compression extraction of pore water from unsaturated tufl YuccaMountain, Nevada. U.S. Geological Survey Water-Resources Investi-gations Report 88-4189, 68 p.

MANUscRJpT RrcHvED NoveMsER 28, I988MANUscRrpr AccEprED FBnnumY 2, 1990


activity diagrams for clinoptilolite: susceptibility of …activity diagrams for clinoptilolite:...

Documents