American Mineralogist, Volume 77, pages I 182- I 190, 1992

The free energy of formation of searlesite, NaBSirOr(OH)r, and its implications

WnN YlNc, Puu,n E. RosnNnnncDepartment of Geology, Washington State University, Pullman, Washington gg164-2812,U.5.A.

Ansrnlcr

The solubility of natural searlesite, NaBSiror(oH)r, has been determined at tempera-tures between 25 and 100 "C at I atm using distilled water, l-M NaCl, and various buffersolutions. The paths taken by different starting solutions to a steady state suggest thatequilibrium was attained. When necessary, reciprocal time (t o s) extrapolations of theactivity products (Q) were used to estimate the equilibrium constant (K) of the dissolutionreaction, thus permitting the calculation of the free energy of formation (AGf) of searlesite.I-og K values between 25 and 100 "c fit a linear equation, log K: 9.9502 - 2660.1/T(K); aGP of searlesite at 25 "C is - 2897.1 + 0.8 kJ/mol.

Activity diagrams (1og a"ruo, vs. log a.,or) for the system NarO-BrOr-SiOr-HrO at 25 "C,constructed using thermodynamic data, show that the searlesite stability field lies betweenthe saturation limits of quartz and amorphous silica, flanked by fields for magadiite andborax at higher and lower log 4sio2, respectively. when log a*^*/a* is fixed at 10.2 toapproximate the compositions of many natural brines, the minimum value of log a","o,necessary for searlesite crystallization is -3. Supersaturation, however, is not a suffici"enicondition for the crystallization of natural searlesite. Thermodynamic calculations suggestthat phillipsite may be altered by natural B-bearing brines to form searlesite and potassiumfeldspar at ambient temperatures. In nature the alteration reaction appears to be promotedunder diagenetic conditions.

INrnonucrroN

In order to predict the fate and mobility of B in naturalsystems, a detailed knowledge of the thermodynamicproperties of both the B-bearing aqueous and mineralspecies is required. Although the inorganic chemistry andthermodynamic properties of B in solution are relativelywell understood, few thermochemical data are availablefor the borosilicate minerals. Even the relative stabilitiesof these minerals at surface conditions are unknown. Be-cause of the lack of thermodynamic data, it is not possibleto predict accurately the geochemical behavior of B inmany natural processes.

Determination of thermodynamic data for borosilicateminerals has not been attempted because the two classicalapproaches for obtaining thermodynamic data for solidphases, calorimetric measurements and solubility studies,are difrcult to apply to these minerals. Calorimetric stud-ies are complicated by the inability to prepare or obtainpure compounds and by the large errors (in terms of logK) inherent in such measurements. Solubility studies aredifficult and time consuming because of the very slowapproach to equilibrium between the solid and the aque-ous phase, but they appeared to be the better techniquefor obtaining at least an estimate of the thermochemicalproperties of the borosilicates. Therefore, an experimen-tal investigation was initiated to determine the thermo-chemical properties of selected borosilicates using the sol-

Searlesite was selected for the first phase ofthis studybecause it was known to exist in alkaline lake evaporitesand, therefore, was thought to have a higher solubilitythan other borosilicates. Searlesite is the only commonauthigenic silicate that is restricted to highly saline, non-marine environments (Hay, 1966, p. 103). It occurs insediments derived from alkaline lakes, which are com-mon in the western United States, particularly in Cali-fornia (e.g., Searles Lake), and, consequently, it may bea component of some soils. Occurrences of authigenicsearlesite have been reported by many authors. Largeamounts of searlesite have been found associated withauthigenic potassium feldspar and zeolites (phillipsite andanalcite; Hay and Moiola, 1963) in the sediments ofSearles Lake, California, at depths of 180-875 feet.Searlesite is also associated with potassium feldspar inthe altered tuffs of Lake Tecopa (Sheppard and Gude,1968). In Teels Marsh, Nevada, searlesite is present inmudstone and ash beds, along with authigenic zeolites(phillipsite and analcite; Taylor and Surdam, l98l).Searlesite is the most widespread and abundant B mineralin the Green River Formation in Wyoming and Utah;reedmergnerite (NaBSirOr) occurs only in restricted areas(Milton and Eugster, 1959).

Thus, a solubility study of searlesite would not onlyprovide an estimate of the free energy of formation ofsearlesite, but it would also have direct application tosome alkaline lake and soil environments. Furthermore,B is in tetrahedral coordination in searlesite (Ghose andubility method.

0003-004x/92 / | | 12- | I 82$02.00 rt82

Treue 1. Molar stability constants (log K) of some B species al25-100 'c

l 1 8 3

Tlele 2. Analytical data (EPMA) for searlesite from Boron, Cal-

ifornia, and its ideal composition (wt%)

Oxide EPMA. ldeal

YANG AND ROSENBERG: FREE ENERGY OF FORMATION OF SEARLESITE

2345

ReaGtion' 25"C s0 "c 75 "C 1 0 0 ' c 13.12

o.o20.01

56.500 . 1 10.150.010.01o.02

74.96

15.190.000.00

58.910.000.000.000.000.00

17.068.83

100.00

9.248.92

20.1321.040.24

8.968.8s

19 6020.05o.25

8.888.83

19.2419.370.26*

8.848.84

18.9718.920.27' .

NarOFeOK"Osio,CaOAlro3Tio,MgoMnOBrO"'*Hrot'

Total

. Reactions and references: 1 : B(OH). + H*: H3BOB + HrO (Mes-mer et af., 1972\, 2 : 2B(OH); + Ht : B,O(OH); I 2H"O (Mesmer et al.,1972),3:38(OHF + 2H*: B.O"(OH); + sH,O (Mesmer et al., 1972),4 : 4B(OHF + 2H- : B4o5(OHE- + 7H,O (Mesmer et al" 1972)' 5:B(OH)4 + Na.: NaB(OH)l (Owen and King, 1943).

-. Extraoolated from data at lower temperatures.

Wan, 1976), as it is in most borosilicates, with the im-portant exception of tourmaline. Thus, determination ofthe free energy formation of searlesite should permit theapproximation of free energies of other borosilicates us-ing one of the published estimation methods (e.g., Cher-mak and Rimstidt, 1989).

Sor-urroN cHEMISTRY oF B

The solution chemistry of inorganic B has been exten-sively studied for many years (e.g., Owen and King, 1943;Mesmer et al., 1972\. The distribution of various B spe-cies in solution is a function of temperature, pH, and thetotal concentration of B. The stability constants of someimportant B polyanions and complexes at temperaturesfrom 25 to 100 "C are listed in Table l. The data ofMesmer et al. (1972), Owen and King (1943), and Rear-don (1976) were selected in this study because they in-clude the values of the equilibrium constant (K) at tem-peratures above 25 "C.

At a total B concentration less than 0. I M, H.BO! andB(OH); are the only significant species in solution; theyequilibrate according to the reaction

B(OH); * H*: H3BO3 + H'O. (l)

At 25 "C, log K, for this reaction is 9.24 (Mesmer et al.,1972), and, thus the activities of H.BO! and B(OH); willbe equal when pH : 9.24 (arro: l). At lower pH values,H3BO3 is the major species in solution, whereas at higherpH values, B(OH); predominates.

ExpnnruBNTAL AND ANALYTICAL PRocEDURES

Starting materials: Solids

Searlesite from Boron, California, which was used inthis investigation, occurs as veinlets, crusts, and clustersof radiating crystals in assemblages with colemanite,Ca[BrOo(OH).].H,O, ulexite, NaCa[B,Ou(OH)6]'5HrO,authigenic feldspar, and zeolites in the altered Saddle BackBasalt (Morgan and Erd,19691, Wise and Kleck, 1988).

The separation of searlesite from the matrix was ac-complished by crushing in a diamond mortar and hand-picking under a binocular microscope. Searlesite was alsoextracted using an isodynamic (magnetic) separator after

'Average of six Point analyses." B,O" and HrO cannot be analyzed by EPMA'

pulverizing the remaining material to a finer grain size.The small percentage of ulexite, found in the concentrate,was removed using gravity (heavy liquid) methods' Thecrystals were then ground under distilled HrO to less than50 pm and centrifuge-washed with distilled HrO. The grain

size fraction less than 0.2 pm was discarded and the re-mainder was dried at 60 "C.

Searlesite was used in the experiments without any fur-ther treatment. Due to the limited supply available' somematerial used in the earlier experiments was reused in thelater experiments. Searlesite was characterized by X-raypowder diffractometry and optical microscopy, as well asby electron microprobe (EPMA) analysis (Table 2). In-

asmuch as insignificant amounts of nonessential elementswere observed, the ideal formula has been used through-out this study. Other solid starting materials used in thisstudy without pretreatment include amorphous (fumed

colloidal) silica (ROC/RIC, Sun Valley, California) andgibbsite (Kiurick, 1966b).

Starting materials: Solutions

Various starting solutions, including distilled HrO, car-bonate-bicarbonate buffer solution, borax buffer solution,and l-M NaCl solution, were used in order to measurethe solubility of searlesite over a range of chemical con-ditions. The carbonate-bicarbonate buffer solution andborax buffer solutions fix the pH at values between 8 and10, depending on the proportions of their components(Perrin, 1974, p. 147-149). The pH of the distilled H,Ostarting solution increases during the first one or twomonths; in experiments of longer duration the solution isself-buffering at a pH ofabout 9.2.

As the behavior of the buffer solutions at elevated tem-peratures is not known, they were used only in experi-ments at 25 "C. At higher temperatures, all experimentswere carried out using distilled HrO or l-M NaCl solu-tions. Supersaturation with respect to searlesite was at-tained by adding various amounts of silica (200-1000ppm), B (80-6000 ppm), and Na (500-23000 ppm) tothe starting solutions, in the form of sodium metasilicate,boric acid, and NaCl, resPectivelY.

I 184

Laboratory procedures

Experiments at 25 "C were carried out in tightly closedpolyethylene bottles containing 80-150 mL of varioussolutions and 0.5-2 g of mineral powders. The bottleswere sealed using commercial Teflon (PTFE) thread seal-ant and subjected to continuous shaking for up to 2 yr.Solutions were sampled for analysis at regular intervals.Similar experiments were carried out at 50 .C in a HrObath and at 75 "C in an oven, without shaking.

At 100 9C, experiments were carried out in Teflon-lined,Al or stainless-steel pressure vessels that were heated inovens for up to four months and then quenched to roomtemperature in about 5 min using cold HrO (Sass et al.,1987; Aja,1989). The 20-mL Teflon liners usually con-tained l2-15 mL solution and 0.5-2 g of mineral pow-der.

Solution samples were filtered through 0.I-pm microfilters within 5 min after quenching; pH was measuredimmediately after filtration. For experiments at temper-atures higher than 25 .C, solutions were diluted for anal-ysis immediately after filtration and pH measurement toprevent precipitation. SiO, and B concentrations were de-termined within 5 h, whereas Na was analyzed within 24h of the termination of an experiment. Solid materialswere centrifuge-washed using distilled HrO and dried at60'C for examination by X-ray diffractometry and op-tical microscopy.

Analytical methods

Quantitative analysis of B was accomplished using theazomethine-H method (John et al., 1975). Some solutionsamples were also analyzed using specific-ion electrodetechniques; the results of the two methods are in closeaccord (Yang, 1990, p. 26). Silica was analyzed using themolybdosilicate method (Skougstad et al., 1979). Spec_trophotometric measurements for both B and silica werecarried out using a Bausch and Lomb Spectronic gg spec_trophotometer; the precision of the analyses is within 5ol0.

Na concentrations were determined by atomic absorp-tion spectrophotometry; the precision of these analyses isestimated to be +l0o/0. Solution pH was measured atroom temperature with a precision of +0.05 pH units.The overall error in the free energy of formation ofsearlesite due to analytical uncertainties is estimated tobe +0.8 kJlmol.

C,lr,curarroNs AND EXTRApoLATIoNSCalculation of activity coefficients

Activity coefficients and pH at elevated temperatureswere calculated using the program Solveq (Reed, l9gl)by adding stability data for the B species (Table l) to thedata base. The stability constants of NaB(OH)! at 75 and,100 "C were extrapolated from lower temperatures be-cause expenmental data are not available at these tem_peratures. The concentrations ofB and Na solutions arenot very high (< | M), and,, therefore, the error caused byextrapolation to higher temperatures should be small.

YANG AND ROSENBERG: FREE ENERGY OF FORMATION oF SEARLESITE

Reciprocal time extrapolation

Dissolution of minerals in aqueous solutions is often aslow process. Solubility measurements are therefore dif-ficult for many minerals, especially silicates, which tendto be less soluble than carbonates and borates.

To circumvent this problem, Garrels et al. (1960) de-vised an empirical method for measuring the solubilitiesof carbonate minerals using free-drift dissolution rate dataand reciprocal time extrapolations. Using this method,quantities such as pH or log activities ofaqueous speciesare plotted as a function ofthe reciprocal square root oftime (r-os;. As the reaction time increases (l-os decreas-es), the data anay approaches linearity. The equilibriumvalues ofactivities are obtained by extrapolating the lin-ear portion of the array to infinite time (l o.s : Q).

Because of the slow rate of the searlesite dissolutionreaction at relatively low temperatures, equilibrium couldnot be reached in a reasonable time. Therefore, the recip-rocal time extrapolation method was used to obtain thesolubility constants (log K) of searlesite at 25, 50, and 75"C. At 100 oC, the duration of experiments was sufficientto bring the dissolution reaction to equilibrium, and,therefore, reciprocal time extrapolation was not neces-sary.

Reciprocal time extrapolation has been criticized be-cause it is not necessarily valid in all cases (Lafon, 1978).Although it is true that the method is purely empirical,it has been used successfully in many cases (e.g., Bricker,1965; Plummer and Mackenzie, 1974 Walter and Morse,1984) and is widely accepted. It has become one of theprincipal methods for determining the stabilities of bio-genic magnesian calcites (Bischoffet al., 1987). Further-more, Kittrick (1966a) and Routson and Kittrick (197 l)successfully used this approach to determine the free en-ergy of formation of kaolinite and illite, respectively. Inthis study, extrapolations of experiments carried out un-der various chemical conditions resulted in nearly iden-tical log K values, slggesting that this method may beappropriate for the searlesite dissolution reaction.

Calculation of free energy of formation (AGf)

Free energies of formation were calculated using thestandard-state convention of Helgeson et al. (1978),(AGf of elements defined as zero only at 298.15 K and Ibar) because the most recent thermodynamic data foraqueous species (Tanger and Helgeson, 1988; Shock etal., 1989) are based on this convention. The free energiesof formation of the aqueous species used in this study arelisted in Table 3.

The searlesite dissolution reaction may be expressed as

NaBSirOr(OH), * H*: Na* * H3BO! + 2SiOr,4. Q)

The equilibrium constant for this reaction, K", can becalculaled using the equation

log K, : log a",. * log 4rr"o, + 2log asioz - log ar- (3)

where a is the activity of the chemical species when thereaction is in equilibrium. The free energy of reaction,

TABLE 3. The AGP (kJ/mol) values of the aqueous species at TABLE 4.25-100 "C

rfc) siog H3BO3 HrO

I 1 8 5

Average values of log K, for the searlesite dissolutionreaction and the free energy of formation of searlesiteat 25,50,75, and 100 "C

rfc) log K, aG? (kJ/mol)

-2897.1-2901.5- 2905.3-2909.7

YANG AND ROSENBERG: FREE ENERGY OF FORMATION OF SEARLESITE

2550T5

100

255075

100Ref.'

-833.41-834.88-836.26-837.64

1

-968.76-972.70-976.7'l-980.86

1

-237.19-238.99-240.57-243.09

2

-261.92-263.42-265.05-266.73

3

0.050.691.321.83

* References: 1 : Shock et al. (1 989), 2 : Helgeson and Kirkham (1 974),3: Tanger and Helgeson (1988).

AGf, can be calculated from the log K, value using therelationship

AG? : -2.303 RI log K, (4)

where R is the gas constant and f is the temperature inkelvins. Since the AGP of the species in solution are known(Table 3), the AGfl*".,".,,. can be calculated by differencefrom the equation

AG8*-."n : AG8"^. + AGtH3Bo3 + 2Ac8s,o2..o- AGl". - aG9. (5)

ExpnmvrnNrAl RESULTS

Searlesite dissolution experiments were carried out at25,50,75, and 100 "C. Experiments at higher tempera-tures were not successful because concentrations ofSi andB in the final solutions were sufficient to cause the pre-cipitation ofa gel during the quench.

The analytical results are tabulated in Appendix I'whereas the calculated activities of chemical species, log

K and the free energy of formation of searlesite, obtainedin experiments from undersaturation, are listed in Table4. The average values of log K and the free energy offormation of searlesite at25,50,75, and 100'C are givenin Table 4.

The activity product, Qr, for the searlesite dissolutionreaction (Eq. 2) is given by the equation

log Qr: log a*^*/at. + log 4r,"or..o * 2 log 4.ror,"o (6)

where a represents the activities of the chemical speciesin solution. As the dissolution reaction progresses, log Q,changes toward a limiting value. When the system reach-es equilibrium ,log Q, : log K', where K is the equilib-rium constant.

A typical diagram showing a gradual increase inlog Q,with the duration of the experiments is shown in Figurela. Because a plateau was not attained, reciprocal timeextrapolation was used to estimate log Q, at infinite timeas a measure of log Kr. Figure lb shows the reciprocaltime extrapolation procedure used in this study. As theexperimental duration increases (/-o5 decreases), the re-lationship of log Q, vs. I o 5 approaches linearity; the ex-

trapolation of log Q, to infinite time (l-os : 0) is takenas log K, of the reaction. Because log Q, tends to varyirregularly during the first two to three weeks of the ex-periments, linear regression was carried out using only

the data points (three to six) representing the longest ex-perimental durations.

Changes in the concentrations of chemical species in

solution with the duration of experiments at 25 oC for a

typical experiment (W66, Appendix l) are illustrated in

Figure 2. The dissolution of searlesite is clearly incongru-

ent, inasmuch as the Si/B and Si/Na ratios are not stoi-

chiometric, as required by Equation 2. During the first

a-2a'\

l , l ,

0 .1 0 .2 0 .3

/ r r - 0 . 5rme(ooys /Fig. I . (a) The variation of log Q, with experimental dura-

tion. (b) Reciprocal time extrapolation; log Qrvs. t os lexperi-ment W56, Appendix l).

- 1

_ to)

- 4

0.4- 6--t-

0.0

I A copy of Appendix I may be ordered as Document AM-92-51 I from the Business Ofrce, Mineralogical Society of Amer-ica, ll30 Seventeenth Street NW, Suite 330, Washington, DC20036, U.S.A. Please remit $5.00 in advance for the microfiche.

i i r n p ( d n r r e \

oo

b.

YANG AND ROSENBERG: FREE ENERGY OF FORMATION OF SEARLESITE

Fig.2. The variations of SiO., B, Na, and H* concentrationsin solution during searlesite dissolution (experiment W66, Ap_pendix l).

several weeks, the concentrations of SiOr, B, and Na in-crease proportionally according to Equation 2. Later, theconcentration of SiO, becomes fixed at about 3.5 mmol/Lwhile the concentrations of Na and B continue to in-crease. Extrapolation to infinite time, using the procedureas described above, gives a -log SiOrM of 2.61, whichis close to the theoretical saturation limit of amorphoussllica (2.67, Walther and Helgeson, 197 7). Since searlesiteis the only crystalline product observed by means of X-raydiffractometry and optical microscopy, amorphous silicars presumed to have precipitated. The B/Na ratio remarnsstoichiometric after equalization for up to 2 yr, providingno evidence for the precipitation of B- or Na-bearing spe-cies.

At the outset of the experiments, the rate at which silicais added to the solution exceeds its rate ofprecipitation,owing to the rapid dissolution of searlesite. Thus, solu-tions become supersaturated with respect to amorphoussilica. During the course of the experiments, the rate ofaddition and subtraction ofsilica from solution equalizesas the rate ofsearlesite dissolution decreases.

Experiments with different proportions of solid and liq-

2 . 7

looo/T(K)Fig. 4. The variarion of log K, with 1000/Z (K) for the

searlesite experiments. Linear regression equation, log K, :8.9502 - 2660.1/T (K) (R : 0.99). Symbol size is approximatelyequal to analytical error.

uid were carried out to test the attainment ofequilibrium.If the log K, values obtained from experiments with dif-ferent solid/liquid ratios are similar, equilibrium is likelyto have been achieved. The changes inlog Q, with timefor two experiments with different solid/liquid ratios areshown in Figure 3. A linear extrapolation of the log Q,values for each experiment gives very similar values oflog K.. Thus, the solid/liquid ratio does not appear toinfluence the experimental results, provided that the ex-periments are of sufficient duration.

In Figure 4, the log K, values at different temperaturesare plotted against reciprocal temperature [1000/Z(K)].the relationship can be expressed by the equation log K,: 8.9502 - 2660.1/T(K), with a correlation coefficient(R) of 0.99. The apparent linearity suggests that equilib-rium was probably attained and that AIIi for the disso-lution of searlesite is approximately constant over thetemperature range of this study.

The paths taken by different starting solutions in ap-proaching equilibrium are shown on a diagram in Figure5 of log eNu*/eH- vs. log arruo, (arrows); log a.,o, is fixedat -3, the average value in these experiments. The solidtriangles represent experiments from undersaturation,whereas the open triangles represent experiments fromsupersaturation. The paths of the solution compositionsare indicated only for experiments from undersaturation,since the solution compositions in the experiments fromassumed supersaturation remained unchanged during thecourse of the experiments. The straight line is the inferredequilibrium boundary between searlesite and solution; itsslope is determined by the stoichiometry of the dissolu-tion equation (Eq. 2), whereas its intercept is based onthe experiments from undersaturation. The slope inferredfrom the experimental data conforms quite well to theslope calculated from Equation 2, again suggesting thatequilibrium was attained.

Efforts to approach equilibrium from assumed super-saturation were not successful; the concentrations ofthechemical species in solution did not change appreciably

3.J3 . 12.9

0 0

O - 1 o

o)o -z.o

- 4 .00 0 o. lz 0.4 0 '6 o. '8 r .b 1 '2

r ' / r \ - 0 . 5T t r n o l a l a \ / q- - ' , - )

Fig. 3. The variation of log Q. with reciprocal time (/ o 5) forexperrments W67 (open circles) and W68 (filled circles) (Appen-dix l) .

o oa

a

YANG AND ROSENBERG: FREE ENERGY OF FORMATION OF SEARLESITE I r87

9 V v

A A A

o 1 50 300 450 600 750

1 6

1 4

1 2

T

d - 1 0

)2B

d

3oJ

S ea r l es i t e

.tP $

Q)

mo6

dl

So lu t i on

1

o3o

J

- 1

-2

o - 6 - 5 - 4 - 3 - 2 - 1 0 1

Log @H.eo.

- 7

Fig. 5. The paths taken by solution compositions duringsearlesite experiments at 25 t. Filled triangles represent averagevalues obtained in experiments from undersaturation, and theopen triangles represent experiments from supersaturation.

during the course of these experiments. The variations oflog Q, with the duration of experiments from undersa-turation (W79, Appendix l) and from supersaturation(W80, Appendix l) are shown in Figure 6. Log Q'in theexperiment from supersaturation is essentially unchangedwith time.

Furthermore, evaporation of searlesite-seeded, dilutesolutions aI 25 "C up to very high concentrations ( I I 200ppm Na, 1870 ppm B, and 380 ppm sil ica, pH : 9.56)resulted in the precipitation ofborax and a silicious gel.

The failure of searlesite to crystallize from solution maybe due to a reaction between BoOr'z and SiOr."o, whichformed a relatively stable gel (see Iler, 1979,p. 190) rath-

er than a crystalline material.

Ps.l.sB RELATToNS IN THE SYSTEMNarO-BrO.-SiOr-HrO

Stability relationships

There are five minerals in the system NarO-BrO,-SiOr-HrO that are important to this study: searlesite [Na-BSiro5(OH)rl, reedmergnerite (NaBSirOr), borax(NarBoO,. l0HrO), kernite (Na,BoO,'4H'O), and maga-diite ([NaSi,O,3(OH)3'4HrO]. Borax and magadiite pre-

cipitate at the Earth's surface (Sonnenfield, 1984; Jonesetal., 1977), whereas the crystallization of searlesite andreedmergnerite appear to require diagenetic conditions(Hay and Guldman, 1987). Borax is stable up to about60 "C; at higher temperatures it dehydrates to form kern-ite (NarBnOr'4HrO) or tincalconite (NarBoOr'5HrO), thelatter being metastable with respect to the former (Bas-

sert. 1976).Although searlesite and reedmergnerite are both sodi-

um borosilicates, searlesite is a OH-bearing sheet struc-

T i m e ( d a y s )

Fig. 6. Variation of log Q with experimental time for exper-iments W79 (filled triangles) and W80 (open triangles) (Appen-

dix l).

ture (Ghose and Wan, 1976), and reedmergnerite is iso-

structural with albite (Clark and Appleman, 1960)' The

stability relations of these two minerals are of great in-

terest because they determine the fate of B in solution

under a wide range of geochemical conditions. Unfortu-

nately, data for reedmergnerite are not available in the

literature, and therefore its stability relationships are un-

known.Searlesite is frequently found in nature' although the

occumence of reedmergnerite is rare. Searlesite is present

in alkaline lake evaporites at Searles Lake (Hay and

Guldman, 1987), Teels Marsh (Smith and Drever, 1976),

and Lake Tecopa (Sheppard and Gude' 1968), whereas

reedmergnerite is found in significant amounts only in

the Green River Formation in Utah, where it occurs with

abundant searlesite (Milton and Eugster, 1959).

Like searlesite, magadiite [NaSirO,r(OH)r'4HrO] is an

important mineral in soils and sediments in certain high-

ly alkaline environments (McAtee et al., 1968). Thus, sta-

bility relationships of magadiite in the system NarO-BrOr-

SiOr-HrO are of considerable interest.With free-eneryies-of-formation values of - 5 5 l3'7 kJl

mol for borax (Felmy and Weare, 1986), -7373 kJ/mol

for magadiite (calculated from Bricker, 1969), -968'8 kJ/

mol for sassolite (Wagman et al-, 1982), and -2897 'l kJ/

mol for searlesite (Table 4), it can be shown that maga-

diite and borax are metastable with respect to searlesite

and should react according to the equation

2NaSi,O'r(OH)3'4H,O + 2.5NarBoO?' l0HrO

magadiite boru

: TNaBSi,O'(OH), + 3H3BO3 + 24.5H,O' (7)

ffirlesite gsslite

The free energy of reaction (AGp) equals -486'8 kJlmol

and, therefore, the reaction should proceed to the right'

As expected, the assemblage megadiite * borax has not

been reported in nature. Hoviever, the assemblage

searlesite + sassolite has not been reported either' pre-

sumably due to the high solubility of sassolite'

III

IIIII

IIII

iIIIIIIIII

I

iI

I 1 8 8 YANG AND ROSENBERG: FREE ENERGY OF FORMATION oF SEARLESITE

8-zTD(DI

do,o - 3

dEo -2I

€o,o

J

- 1-2-3-4-2-3-4

Log csioz

Fig.7 . Stability relations of some phases in the system NarO_BrO3-SiOr-Hr0 at 25 t as functions of log ar,_, and log a",o,,with the value oflog ar^-/ar* fixed at 7.5. Th; borax_searlesitiand searlesite-magadiite boundaries are drawn assuming that Nars conserved in the reactions. Triangles : searlesite experimentsat 25 "C (Appendix l); dashed lines : quartz and amorphoussilica saturation limits (Walther and Helgeson, l977).

The system NarO-BrOr-SiOr-HrO

Stability relationships in the system NarO-BrOr-SiOr_HrO are shown in Figure 7;logar^,/an- is fixed at7.5,an average value for the searlesite dissolution experi_ments (Appendix l). The borax-searlesite and searlesite_magadiite boundaries are drawn with the assumption thatNa is conserved in the reactions. The quartz and amor-phous silica saturation limits (Walther and Helgeson,1977) are shown by dashed lines.

Under the conditions specified for Figure 7, the searles-ite stability field is located entirely between the saturationlimits ofquartz and amorphous silica, and above log 4rr*,values ofabout -1. Borax is stable at log crr"o, valueihigher than -0.5, whereas magadiite is stable itiog a'o,values higher than -3.2.

The data obtained from searlesite solubility experi-ments from undersaturation at 25 "C (Appendix l) areshown in Figure 7 as triangles. Because the purpose ofthis diagram is to evaluate the stability of searlesite, theB activities for the experimental data have been adjustedin the diagram based on the searlesite dissolution reaction(Eq. 6) and the assumption that log e,^*/an*: 7.5. Ad-justment is necessary because log a*^./ar- is fixed at 7.5in Figure 7, but log a**/ar* for each experimental solu-tion is not necessarily 7 .5. If the log e*^*/ e* of an ex-perimental solution is not 7.5, the difference has beenadded to or subtracted from the log a".uo. value, in orderto keep the log Q, of the solution unchanged. The ad-

Log as;e,

Fig. 8. Stability relations of some phases in the system NarO-BrO3-SiOr-HrO, as functions of log a"ruo, and log a.,o, with thevalue oflog a"../an- fixed at 10.2. The borax-searlesite and sear-lesite-magadiite boundaries are drawn assuming that Na is con-served. Analytical dara for some natural brines are also shownon the diagram. Data sources: Searles Lake (filled circles), Smith(1979); Owens Lake (open triangles), Friedman et al. (1976);Lake Magadi (open squares), Jones et al. (1977); Teels Marsh(open circles), Taylor and Surdam (1981). See text for details.

justed results of the searlesite dissolution experiments liealong the metastable extension of the searlesite-solutionboundary in the stability field of magadiite (Fig. 7).

Natural brines in the systern NarO-BrOr-SiOr-HrOIn order to evaluate the stability of the minerals in

Figure 7 in natural brines, a phase diagram for the systemNarO-BrOr-SiOr-HrO has been constructed (Fig. 8) withlog a*^-/ar, fixed at 10.2, the average value for the naturalbrines (Searles Lake, Teels Marsh, Owens Lake, and LakeMagadi) shown on the diagram. The borax-searlesite andsearlesite-magadiite boundaries are drawn with the as-sumption that Na is conserved in the reactions involvingthese mineral pairs. At the higher log e*^*/er* of Figure8, the stability fields ofthe solid phases are enlarged andthe stability field of solution is reduced as compared withFigure 7. Because ofthe high concentrations ofdissolvedsolids in the natural brines (Fig. 8), the activities of thechemical species in solution were calculated using thePitzer equations (Pitzer, 1975), following the procedureof Felmy and Weare (1986). Inasmuch as the purpose ofthis diagram is to evaluate the degree of saturation ofthese brines with respect to searlesite, log a",uo" values ofthe brines have been adjusted in the light o? Iieaction 6and the assumption that log a*^./a^, equals 10.2; a sim-ilar adjustment was made for Figure 7.

Brines from all four ofthe lakes are supersaturated with

/ .-@Q)

:oo-

ca'D

o tea,r les Lake

l oi O w e n s L a k e

o 8 t r f ,ao^b a

F-,1o i a

Tee ls Marsh D^ oo l

i - d o: oqadii tdg lp "- l : : o

17 o Lake l , tasaoi

l t ot 9 0t ;t i

o n

i r@

respect to searlesite as well as magadiite. Searles Lakebrines and some of the Owens Lake brines are also su-persaturated with respect to borax. Although searlesitehas been observed in the sediments of Searles Lake andTeels Marsh, it has not been reported at Owens Lake orat Lake Magadi. Thus, supersaturation is not a sufficientcondition for searlesite precipitation in nature.

Precipitation of searlesite in nature

Precipitation of borates and borosilicates is known tobe an extremely slow pro@ss in nature (Sonnenfield, 1984,p. 221), despite high B concentrations in solution. Ex-perimental evaporation of sea water produced only a gel,which yielded anhydrous boracite [MgrCl(BrO,r)] upondehydration (Valyashko et al., 1969).

As an authigenic mineral, searlesite is almost alwaysrelated to tuffaceous rocks and alkaline, saline lakes (Hay,1966; Sheppard and Gude, 1968). It is usually found atdepth in ancient playa deposits. Precipitation ofsearlesiteunder surface conditions in modern saline lakes is un-known, suggesting that searlesite usually crystallizes un-der diagenetic conditions. Studies of the sediments atSearles Lake by Hay and Guldman (1987), Teels Marshby Taylor and Surdam (198 l), and Lake Tecopa by Shep-pard and Gude (1968) suggest that zeolite precursors,formed by the alteration of volcanic ash, may be neces-sary for the crystallization of searlesite. Searlesite oftenoccurs in assemblages with authigenic potassium feldsparand zeolites (Hay and Moiola, 1963). In the Pleistoceneand Recent sediments of Searles Lake, phillipsite forms80-900/o of some rhyolitic tuffhorizons in the upper I l0feet; the association potassium feldspar-analcime-searlesite is dominant at depths greater than 230 feet (Hay,1964). Furthermore, Sheppard and Gude (1968) providetextural evidence showing that potassium feldspar andsearlesite replace phillipsite. Therefore, it appears thatmost of the authigenic searlesite reported in the literatureformed by the alteration of zeolites when the activity ofB was sufficiently high, as suggested by Hay and Guldman( l 987) .

This inference may be confirmed by consideration ofthe reaction for the alteration of phillipsite to searlesite.The phillipsite in saline lakes and soils is usually verylow in alkaline-earth ions and has a SiOr/R O. ratio near5.0 (Hay, 1964). With the assumption that Al is con-served in the solid phases, the alteration reaction can bewritten as

NaKAlrSirO,o.5HrO + 3SiO2."" + H3BO3 + K*phillipsite

: NaBSirO,(OH), + 2KAlSi3O8 + 5HrO + H+ (8)searlmite lDtassium feldsptr

where

log K* : -log 4"r"o, - 3 log asioz - log(a"t/a"-): -AG?/(2.303 RO (e)

where -AGf is the free energy of reaction of Equation 8,

I 1 8 9

R is the gas constant (8.314 J/k'mol) and I is the tem-perature in kelvins.

The log K, value for Equation 9 is 1.4 at 25 "C, basedon the free energies of formation of the minerals andaqueous species in Equation 8. The free energy of for-mation ofphillipsite (-7798 + 27 kl/mol) was calculatedusing the method of Chermak and Rimstidt (1989); itsentropy and CB values were taken from the 1988 database of SUPCRT (Tanger and Helgeson, 1988). Free en-ergies of species other than searlesite are also from the1988 data base of SUPCRT.

Calculations by Hay and Moiola (1963) have shownthat log a*,/a.* and log 4.'o, in the Searles Lake brine areabout 9 and -4, respectively. Log 4rr"o, for the samebrine, calculated from the data of Smith (1979), is about- I .6 (Yang, 1990, p. 144-145). With these data, the cal-culated value of log Q" (l.a) is lower than log K' at 25'C (3.87), indicating that the reaction should proceed to-ward the riglrt and that searlesite * potassium feldsparis the stable assemblage. Thus, the crystallization ofsearlesite under surface conditions according to Equation8 is thermodynamically favorable but kinetically unfa-vorable at ambient temperatures.

According to Smith (1979), the average temperature atthe surface ofSearles Lake is about I 9 'C, while at a depthof 875 feet, the temperature is estimated to be 45 "C;searlesite and potassium feldspar appear at depths of I 80(about 26 "C) to 875 feet. Elevated temperatures at depthprobably promoted the alteration of zeolites to formsearlesite and potassium feldspar. Alteration of phillipsiteto searlesite and potassium feldspar (Eq. 8) at Teels Marshand Searles Lake buffers the activity of silica in brines atlower levels than those at Lake Magadi (Surdam andSheppard, l 978), which are lower in B and have not pre-

cipitated searlesite. Crystallization of searlesite apparent-ly leads to the formation of borax at Searles Lake, whereit is a relatively common mineral. In the deeper tufflay-ers, including the searlesite-bearing levels, the activity ofsilica in solution is controlled by equilibrium with quartz(Hay, 1964). Thus, solution compositions probably lie inthe stability field of borax (Fig. 8). At Owens Lake, mostof the ash is unaltered (Sheppard and Gude, 1968), littlephillipsite is present, and searlesite is absent despite thehlch B and silica contents ofthe brine (Fig. 8).

AcxNowr-rncMENTS

This research was funded by Battelle, Pacific Northwest Laboratories,under contract no. B-N4009-C-E. The authors thank Dhanpat Rai for hissupport and interest and J.A. Kittrick for his guidance in the early stagesof this study. We also thank Andrew Felmy and Eric Oelkers for reviewsofthe rnanuscript.

RnrnnrNcrs cITED

Aja, S.U. (1989) A hydrothermal study of illite stability relationshipsbetween 25"and 250 qC and P": PH,o, l9l p. Ph.D. thesis, WashingtonState University, Pullman, Washinglon.

Bassett, R.L. (1976) The geochemistry ofboron in thermal waters, 290 p.Ph.D. thesis, Stanford University, Stanford, Califomia.

Bischoff, W.D., Mackenzie, F.T., and Bishop, F.C. (1987) Stabilities of

YANG AND ROSENBERG: FREE ENERGY OF FORMATION OF SEARLESITE

I 190

synthetic magnesian calcites in aqueous solution: Comparison with bio-genic materials. Geochimica et Cosmochimica Acta. 51. 1413_1423.

Bricker, O.P. (1965) Some stability relarions in the system MnO,-H,O at25 t and I atmosphere total pressure American Mineralogist, 50,t296-r354

- (l 969) Stability constants and Gibbs free energies of formation ofmagadiite and kenyaite. American Mineralogist, 54, 1026-1033.

Chermak, J.A., and Rimstidt, J.D. (1989) Estimating the thermodynamicproperties (AG! and A}lll) of silicate minerals at 298 K from the sumof polyhedral contributions. American Mineralogist, 7 4, lOZ3-lO3 l.

Clark, J R., and Appleman, D.E. (1960) Crystal structure refinement ofreedmergnerite, the boron analog ofalbite. Science, 132, 1837-183g.

Felmy, A.R, and Weare, J.H. (1986) The prediction of borare mineralequilibria in natural waters: Application to Searles lake, California.Geochimica et Cosmochimica Acta. 50. 2771-2783.

Friedman, I., Smirh, G.I., and Hardcastle, K.G. (1976) Studies of qua-ternary saline lakes-Il Isotopic and compositional changes duringdesiccation ofthe brines in Owens Iake, California, 1969-1971. Geo-chimica et Cosmochimica Acta, 40. 53-7 2.

Garrels, R.M., Thompson, M.E., and Siever, R (1960) Stability of somecarbonates at 25'C and one atmosphere total pressure. American Jour-nal of Science. 268. 439-453.

Ghose, S., and Wan, C. (1976) Structural chemistry ofborosilicates, partII: Searlesite, NaBSi,O,(OH),: Absolute configuration, hydrogen loca-tions, and refinement of the structure. American Mineralogist, 61, 123_129.

Hay, R.L (1964) Phillipsite of saline lakes and soils. American Miner-alogist, 49, 1366-l 387.

-(1966) Zeolite and zeolitic reactions in sedimentary rocks. Geo-logical Society ofAmerica Special Paper, 85, 130 p

Hay, R.L., and Guldman, S.G (1987) Diagenetic alteration of silicic ashin Searles Lake, California. Clays and Clay Minerals, 35,449-457.

Hay, R. L., and Moiola, R.J. (1963) Authigenic silicate minerals in SearlesLake, C-alifornia, Sedimentology, 2, 3 12-332.

Helgeson, H C., and Kirkham, D.H. (197a) Theoretical prediction of thethermodynamic behavior of aqueous electrol).tes at high pressures andtemperatures. L Summary of the thermodynamic/electrostatic proper_ties ofthe solvent. American Journal ofScience, 274,1099-llgg.

Helgeson, H.C., Delany, J.M., Nesbitt, H.W., and Bird, D.K (1973) Sum-rnary and critique of the thermodynamic properties of rock-formingminerals. American Journal of Science, 278A, 229 p.

Iler, R.K. (1979) The chemistry of silica, 230 p. Wiley, New york.John, M.K., Chush, H.H, and Nenfeld, J.H. (1975) Applicarion of im-

proved azomethine-H method to the determination of boron in soilsand plants. Analytical Letters, 8, 559-568

Jones, FB., Eugster, H.P., and Retting, S.L. (1977) Hydrochemistry ofthe [.ake Magadi basin, Kenya Geochimica et Cosmochimica Acta,41,53-72.

Kittrick, J.A. (1966a) Free energy offormation ofkaolinite from solubilitymeasurements. American Mineralogist, 5 l, 1457 -l 466

- (l 966b) Free energy of formation of gibbsite and AI(OH); fromsolubility measurements. Soil Science Society of America proceedings,30, 595-589.

Lafon, G.M. (1978) Equilibrium criteria for two-component solids react-ing with fixed composition in an aqueous phase-Example: The mag-nesian calcites. American Journal ofScience, 278. 1455-1488.

McAtee, J.L., House, R., and Eugster, H.P. (1968) Magadiite from TrinityCounty, Califomia. American Mineralogist, 53, 206 l-2069.

Mesmer, R E., Base, C F., Jr., and Sweeton, F.H. (1972) Acidity mea-surements at elevated temperatures. Inorganic Chemistry, I l, 537 -543.

Milton, C., and Eugster, H.P. (1959) Mineral assemblages of the GreenRiver Formation. In P.H. Abelson, Ed., Researches in geochemistry,vol. l, p. I I 8- 150. Wiley, New York.

Morgan, V., and Erd, R.D. (1969) Minerals ofthe Kramer Borate District,California. California Division of Mines Geology, Mineral InformationService 22, 143-153, 165-172.

Owen, B.B., and King, E.T. (1943) The effect of sodium chloride uponthe ionization of boric acid at various temperatures. Journal of theAmerican Chemical Society, 65, 1612-1 620.

Perrin, D.D. (1974) Buffers for pH and meral ion control, 176 p Wiley,New York.

YANG AND ROSENBERG: FREE ENERGY OF FORMATION OF SEARLESITE

Pitzer, K.S. (1975) Thermodynamics ofelectrolltes: 5. Effects ofhigher-order electrostatic terms. Journal ofSolution Chemistry, 4,249-265.

Plummer, L N., and Mackenzie, F.T. (1974) Predicting mineral solubilityfrom rate data: Application to the dissolution of magnesian calcites.American Journal of Sciencn. 27 4. 61-83.

Reardon, E.T. (1976) Dissociation constants for alkali earth and sodiumborate ion pairs from l0 " to 50'C. Chemical Geology, 18, 309-325.

Reed, M.H. (1981) Calculation of multicomponent chemical equilibriaand reaction processes in systems involving minerals, gases and aque-ous phases. Geochimica et Cosmochimica Acta, 46, 513-528.

Routson, R.A., and Kittrick, J.A. (1971) Illite solubility. Soil ScienceSociety of America Proceedings, 35, 7 14-7 18

Sass, B.M, Rosenberg, P.E., and Kittrick, J.A. (1987) The srability ofillite/smectite during diagenesis: An experimental study. Geochimicaet Cosmochimica Acta. 5 l. 2103-2115.

Sheppard, R.A., and Gude, A.J. (1968) Distribution and genesis of au-thigenic silicate mineral in tuffs ofPleistocene Iake Tecopa, Inyo County,California. U.S. Geological Survey Professional Paper 597, 38 p.

Shock, E.L., Helgeson, H.C., and Sverjensky, D.A. (1989) Calculation ofthe thermodynamic and transport prop€rties ofaqueous species at highpressures and temperatures: Standard partial molal properties ofinor-ganic neutral species. Geochimica et Cosmochimica Acta, 53,2157-2 1 8 3 .

Skougstad, M.W., Fishman, M.J., Friedman, L.D, Erdman, D.E., andDuncan, S.S. (1979) Methods fordetermination of inorganic substancesin water and fluvial sediments In Techniques ofwater resource rnves-tigations of the United States Geological Suwey, p 495-496 U.S.Geological Suwey, Washington, DC.

Smith, C L., and Drever, J.I.(1976) Controls on the chemistry of springsat Teels Marsh, Mineral County, Nevada. Geochimica et Cosmochimi-ca Acta, 40, l08l-1093.

Smith, G.I. (1979) Subsurface strati$aphy and geochemistry, Quaternaryevaporites, Searles Lake, California. U S. Geological Survey Profes-sional Paper 1043, 130 p.

Sonnenfield, P. (l 984) Brines and evaporites, 6 I 3 p. Acadernic Press, NewYork.

Surdam, R.C., and Sheppard, R.A. (1978) Zeolites in saline, alkaline-lakedeposits In L.B. Sand and F.A. Mumpton, Eds., Natural zeolites: Oc-currence, properties, use, p. 145-174. Pergamon Press, New York.

Tanger, J.C., and Helgeson, H.C. (1988) Calculation of the thermodynam-ic and transport properties of aqueous species at high pressures andtemperatures: Revised equations of state for the standard partial molalproperties ofions and electrolltes. American Journal ofScience, 288,l 9-98.

Taylor, M.W., and Surdam, R.C. (1981) Zeolite reactions in the tuffaceoussediments at Teels Marsh, Nevada. Clays and Clay Minerals, 29, 341-352.

Valyashko, M.G., Zherebstsova, I.K., and l-avrova, A.N. (1969) Behaviorofsome trace elements during the eutonic-(bischofite) stage ofseawaterconcentration. Geochimiya, 5, 61 4-619.

Wagman, D.D., Evans, W.H., Parker, V.B., Schumm, R.H., Halow, I.,Bailey, S.M., Churney, K.L, and Nuttall, R.L. (1982) The NBS tablesof chemical thermodynamic properties: Selected values for inorganicand Cl and C2 organic substances in SI units. Journal of PhysicalChemistry Reference Data, I I (suppl. 2), 392 p.

Walter, L.M., and Morse, J.w. (1984) Magnesian calcite stabilities: Areevaluation. Geochimica et Cosmochimica Acta. 48. 1059-1069

Walther, J.Y., and Helgeson, H. (1977) Calculation ofthe thermodynamicproperties of aqueous silica and the solubility of quartz and its poly-morphs at high pressures and temperatures. American Joumal ofSci-ence .277 . l 3 l 5 -1351 .

Wise, W.S, and Kleck, W.D. (1988) Sodic clay-zeolite assemblages inbasalt at Boron, California Clays and Clay Minerals, 36, l3l-136

Yang, W. (1990) Solubilities of selected borosilicate minerals between 25.C and 250 € and P" : Pn,o, 159 p. Ph.D. thesis, Washington StateUnivenity, Pullman, Washington.

Mmuscnrpr REcErwD Aucusr l, 1991MaNuscnrrr AccEprED JuLv 16,1992