Deliquescence of NaCl–NaNO3, KNO3–NaNO3, and NaCl–KNO3 salt mixtures from 90 to 120°C

Deliquescence of NaCl–NaNO3, KNO3–NaNO3, and NaCl–KNO3 salt mixtures from 90 to 120°C GEOCHEMICALTRANSACTIONS VOLUME6,NUMBER2 JUNE2005 DeliquescenceofNaCl–NaNO3, KNO3–NaNO3, andNaCl–KNO3 saltmixturesfrom90to120?C SusanCarroll,a)LauraCraig,andThomasJ.Wolery EnergyandEnvironmentDirectorate,LawrenceLivermoreNationalLaboratory,Livermore,California94550 ?Received11November2004;accepted26January2005;publishedonline11April2005? We conducted reversed deliquescence experiments in saturated NaCl–NaNO3–H2O, KNO3–NaNO3–H2O, andNaCl–KNO3–H2Osystemsfrom90to120?Casafunctionofrelative humidity and solution composition. NaCl, NaNO3, and KNO3 represent members of dust salt assemblages that are likely to deliquesce and form concentrated brines on high-level radioactive wastepackagesurfacesinarepositoryenvironmentatYuccaMountain,NV.Discrepancybetween modelpredictionandexperimentcanbeashighas8%forrelativehumidityand50%fordissolved ionconcentration.Thediscrepancyisattributedprimarilytotheuseof25?CmodelsforCl–NO3 andK–NO3 ioninteractionsinthecurrentYuccaMountainProjecthigh-temperaturePitzermodel to describe the nonideal behavior of these highly concentrated solutions. ? 2005 American InstituteofPhysics. ?DOI:10.1063/1.1872292? I.INTRODUCTION ityforitspuresaltcomponents.Saltdeliquescencedataare largely limited to 25?C for mixed salts3–5 but are widely YuccaMountain,NVisthedesignatedgeologicreposi- availableforsinglesaltsathighertemperatures.2 toryforpermanentdisposalofhigh-levelnuclearwaste.Cur- Therangeofbrinesformedbythedeliquescenceofhy- rentwastepackagedesigncallsfordoublewalledcontainers groscopic salts found in dusts can be calculated from mix- with an inner wall of stainless steel and an outer wall of turesofthepurephasesusingequilibriumthermodynamics, highly corrosion resistant Ni–Cr–Mo alloy, which are pro- because relative humidity is related to the activity of water tected with Ti shields to prevent rocks and seepage water 6usedthe EQ3/6 andsolutioncomposition.WoleryandWangfromcontactingthecontainers.1Ofconcernarethecorrosion geochemical code and the Yucca Mountain Project high- resistanceandlong-termintegrityofthesemetalbarriers.If temperaturePitzermodeltopredictthedeliquescenceofsalt theYuccaMountainsitelicenseisapproved,thewastepack- mixtures found in Yucca Mountain dust samples collected ageswillbeplacedintunnelsseveralhundredmetersbelow from exploratory tunnels. The modeling results predict that the ground surface in partially saturated volcanic tuff, but mixturesofNaCl,KNO3,and/orNaNO3arethemostpreva- still well above the groundwater table. A likely source of lent mineral assemblages, and that inclusion of KNO3?s? brinesthatmaypotentiallycorrodemetalcontainersanddrip lowersthedeliquescencerelativehumidityfromvaluesnear shields are those formed by the absorption of water by hy- 70%at25?Ctovaluesaslowas20%at160?C.Thisanaly- groscopic salts found in local and regional dust deposited sis implies that concentrated brines may contact the metal duringrepositoryconstructionandventilationstages. container and drip shield surfaces at low relative humidity Accurate prediction of brine formation is important for andhightemperature. thesafedisposalofradioactivewaste,becausebrinecompo- Inthispaperwecomparemodelpredictionsandexperi- sition is an indicator of the corrosiveness of the aqueous mentalresultsofrelativehumidityandsolutioncompositions environment and the relationship between deliquescence for saturated NaCl–NaNO3–H2O, KNO3–NaNO3–H2O, relativehumidityandtemperatureisanindicatorof‘‘reposi- and NaCl–KNO3–H2O systems from 90 to 120?C. These tory dryness.’’Deliquescence refers to the formation of an are benchmark experiments that can be used to assess the aqueoussolutionbytheabsorptionofwaterbyhygroscopic validityofthemodelcalculationsandPitzerparametersthat salt minerals. This process allows brines to form above accountfornonidealioninteractionsinthesehighlyconcen- 100?Catstandardatmosphericpressureof1.01325bar?or tratedsolutions. above96?Cand0.9barattherepositoryelevationof1039– 1107m;BSC,2004a,Sec.6.7.2.1?.Therelativehumidityat II.MODELCALCULATIONS which salts deliquesce is dependent on temperature and is characteristictoeachsaltmineralorassemblageofsaltmin- Relativehumidityisthermodynamicallytiedtosolution erals.Forexampleat90?C,MgCl2deliquescesat24%rela- compositionthroughtheactivityofwater.Forwater,itsac- tive humidity and KCl deliquesces at 78.5% relative tivity, aw, is the product of its mole fraction, xw, and its humidity.2Generally,thedeliquescencerelativehumidityfor molefractionactivitycoef,cient,?w: asaltmixtureislowerthanthedeliquescencerelativehumid- aw?xw?w. ?1? Themolefractionofwaterisdependentonthesolutioncom- a?Author to whom correspondence should be addressed; electronic mail: position: carroll6@llnl.gov 1467-4866/2005/6(2)/19/12/$22.00 19 ?2005AmericanInstituteofPhysics 20 Geochem.Trans.,Vol.6,No.2,June2005 Carrollet al. FIG.1. ?Color?DeliquescenceofNaCl–NaNO3 saltsat90?Cstartingfrominitiallydissolvedandinitiallysolidmixturesplottedas?a?%relativehumidity and?b?solutioncomposition. The activity of water is commonly expressed in decimal xw?nw n ? n ?2? w i ? ? form and RH is commonly expressed as a percentage; thus ? ? i RH%?100?aw. where nw and ni are the number of moles of water and of In hydrologically unsaturated environments, the partial dissolvedionicconstituenti,respectively.Theconcentrations pressureofwatercannotexceedthetotalpressure.Thepar- ofthedissolvedconstituentsarelimitedbythesolubilitiesof tialpressureofwaterisrelatedtothevaporpressureofpure thethermodynamicallystablesaltminerals.Ifweusehalite 0 waterandtheactivityofwaterby pw?p aw. Boththeva- w ?NaCl?asanexample,thenthesolubilityproduct,KNaCl , at porpressureofpurewaterandthepartialpressureofwater agiventemperatureisdependentontheNa? andCl? con- overasaltsolutionincreasestronglywithtemperatureabove centrations and activity coef,cients according to the mass 100?C.Brinescannotexistattemperaturesabovethedry-out balancereaction: temperature, where the partial pressure of water equals the NaCl?s)?Na??Cl?, ?3a? totalpressure.Abovethedry-outtemperatureonlysolidsalt mineralsandwatervaporoccur.Sincethisprocessisrevers- ?3b? KNaCl??mNa??Na???mCl??Cl??, ible, the dry-out temperature is also the deliquescence tem- wheremiand?iindicatethemolalityandactivitycoef,cient perature for the same assemblage of salt minerals ?note: at ofspeciesi,withunitactivityimpliedforthesolid?NaCl?. muchlowertemperature,generally25?Corless,dry-outor Relativehumidity,RHfrac, isrelatedtoactivityofwater ‘‘ef,orescence’’is controlled by salt mineral nucleation in- stead of equilibrium thermodynamics.5 The deliquescence throughthepartialpressureofwatervapor,andisequivalent to temperature is also equivalent to the boiling point for the saturatedsolution. 0 RHfrac?pw/p , ?4? w Brines formed by the absorption of water by deliques- wherep isthepartialpressureofwatervaporoveranaque- w cent minerals are thermodynamically equivalent to brines ous solution and pw0 is the partial pressure of water vapor thataresaturatedwithrespecttothesamedeliquescentmin- overpurewater.Similarly,theactivityofwaterisequivalent erals. Figures 1–5 show the model calculations of relative to humidityandsolutioncompositionfrom90to120?Cversus 0 XNO3 or XNa for the NaCl–NaNO3–H2O, the aw?fw/f , ?5? w KNO3–NaNO3–H2O, and the NaCl–KNO3–H2O systems. where fw is the fugacity of water vapor over an aqueous FortheNaCl–NaNO3–H2Osystem,thisisdonebyadding 0 solution and f is the fugacity of water vapor over pure w sodaniter(NaNO3) toahalite?NaCl?saturatedsolutionun- water.Equatingfugacitywithitspartialpressure: til the solution is saturated with both soda niter and halite. 0 0 This represents the eutonic point for the NaCl–NaNO3 salt fw/f ?pw/p ?6? w w assemblage.Thecalculationforthisassemblageiscompleted yields: by a complementary run in which halite is added to a soda ?7? niter saturated solution until the same eutonic point is RHfrac?aw. Geochem.Trans.,Vol.6,No.2,June2005 Deliquescencesaltmixturesathightemperatures 21 FIG.2. ?Color?DeliquescenceofNaCl–NaNO3saltsat110?Cstartingfrominitiallydissolvedandinitiallysolidmixturesplottedas?a?%relativehumidity and?b?solutioncomposition. reached. An identical approach is used for the chargeandternaryinteractions(?M,M?,X,?X,X?,M ) involving KNO3–NaNO3–H2O and NaCl–KNO3–H2O systems. We threeionsincommon-cationandcommon-anionternarysys- use the EQ3/6 geochemical code7 and the Yucca Mountain tems. Pitzer parameters relevant to this study are listed in Projecthigh-temperaturePitzermodel.8Thismodelisbased TableI. on the Pitzer equations ?cf. Ref. 9? to account for the non- idealbehaviorofthebrinesolutions.Reactionpressurewas III.EXPERIMENTALMETHODS allowedtovarybymeansofthe1.013bar/steam-saturation A.Startingmaterials curve.ThePitzermodelisbasedontheavailableexperimen- tal includes binary interactions Analyticalgradesodiumchloride?NaCl,100.1%pure?, data and (0) (1) (2) ? (? ,? ,? sodiumnitrate(NaNO3, 99.7%pure?,andpotassiumnitrate M,X M,X ,C ) between two different kinds of ions of opposite charge ?cation M and anion X?, and also (KNO3, 99.5%pure?wereusedtosynthesizedissolvedand M,X M,X binary (?M,M?,?X,X?) interactions between ions of like drysaltmixtures.ImpuritiesinNaClincludelessthan0.01% FIG.3. ?Color?DeliquescenceofKNO3–NaNO3saltsat90?Cstartingfrominitiallydissolvedandinitiallysolidmixturesplottedas?a?%relativehumidity and?b?solutioncomposition. 22 Geochem.Trans.,Vol.6,No.2,June2005 Carrollet al. FIG.4. ?Color?DeliquescenceofNaCl–KNO3 saltsat90?Cstartingfrominitiallydissolvedandinitiallysolidmixturesplottedas?a?%relativehumidity and?b?solutioncomposition. B.Reversedeliquescenceexperiments Br; 0.005% K; 0.004% SO4, 0.003% ClO4, NO3, and Ba; 0.002% I and Ca; 0.001% Mg; 5 ppm PO4; 2 ppm heavy We measured brine composition at controlled relative metals,and1ppmFe.ImpuritiesinNaNO3includelessthan humidity for NaCl–NaNO3, KNO3–NaNO3, and 0.002%K;0.001%SO4,NO2,Ca,Mg,andR2O3 ;0.0005% 3 salt mixtures from 90 to 120?C as a function NaCl–KNOCl; 3 ppm PO4 and heavy metals, and 1 ppm Fe and IO3. ofXNO3 orXNa. Ourexperimentaldesignmimicsthemodel Impurities in KNO3 include less than 0.004% Na; 0.002% calculations in that one salt in the binary system will com- SO4; 0.001% Cl, and NO2, 0.0005% Ca; 0.0002% Mg; 3 pletelydissolveandthedissolutionoftheotherwillbelim- ppm heavy metals, 2 ppm PO4; and 1 ppm Fe and IO3. ited by its solubility. We approached the equilibrium brine Distilled and de-ionized ?18 M?? water was used to make compositionbyplacingidenticalmolefractionsofdissolved allsolutions. FIG.5. ?Color?DeliquescenceofNaCl–KNO3 saltsat120?Cstartingfrominitiallydissolvedandinitiallysolidmixturesplottedas?a?%relativehumidity and?b?solutioncomposition. Geochem.Trans.,Vol.6,No.2,June2005 Deliquescencesaltmixturesathightemperatures 23 TABLEI. Pitzerinteractionparametersusedwith EQ3/6tocalculaterelativehumidityandsolutioncomposition insaturatedbrinescontainingNa,K,Cl,andNO3. Temperaturedependenceofeachoftheseioninteraction parametersisderivedfromEq.?11?:?(T)?a1?a2/(T?Tr)?a3ln(T/Tr)?a4(T?Tr). Pitzerinteractionparameters Ioninteractions Coef,cients a1 a2 a3 a4 Reference ?(0) 10 4.78?10?02 ?3.43?10?02 ?1.38 ?10?03 1.34K–Cl MX (1) 2.16?10?01 ?5.76?10?02 ?2.88 4.64?10?03 ? MX (2) ? 0.00 MX ?7.49?10?04 3.65?10?01 1.48?10?01 ?1.47?10?04 C?MX K–NO3 (0) ?8.16?10?02 9 ? MX (1) 4.94?10?02 ? MX (2) ? 0.00 MX C?MX 6.60?10?03 Na–Cl ?(0) 7.46?10?02 ?4.71?10?02 ?1.85 ?10?03 1.6610 MX (1) 2.75?10?01 ?5.21?10?02 ?2.88 4.71?10?03 ? MX (2) ? 0.00 MX 1.54?10?03 4.81?10?01 1.75?10?01 ?1.56?10?04 C?MX 11 (0) 3.57?10?03 ?7.03?10?02 ?3.35 3.98?10?03 Na–NO3 ? MX (1) 2.32?10?01 ?2.73?10?03 ?1.30?10?01 2.07?10?02 ? MX (2) ? 0.00 MX C?MX ?4.15?10?05 6.48?10?01 3.18?10?01 ?3.84?10?04 S?MM? ?3.20?10?03 1.40?10?01 9.09?10?13 ?2.66?10?15 K–Na 10 S?XX? 1.60?10?02 Cl–NO3 9 K–Na–Cl ?3.41?10?13 6.66?10?16 ?MM?X ?3.69?10?03 ?5.10 10 andinitiallysolidsamplesconvergeoverthetimeperiodof 2O) andsolidsaltmixturesinanenvi- ?withabout150gH theexperiments.Samplesweretypicallytakeneveryoneto ronmentalchamber?Ecosphere,Despatch?atcontrolledrela- tive humidity and temperature. Under these conditions the threedays.Calibratedtemperature??1??andrelativehumid- dissolvedsaltmixtureevaporatesconcentratingthesolution ity probes ?Vaisala Inc., model HMP243 or HMP233? were andprecipitatingoneofthetwosalts,andthesolidsaltmix- placed just above the solutions in the beakers, because mi- ture absorbs water dissolving the salts until equilibrium is croenvironments within the beakers differed from the envi- reached. ronmentalchamberrelativehumiditybyasmuchas5%rela- For each run, fourTe,on? beakers containing the dis- tive humidity. Relative humidity probe calibration checks solvedsaltsandfourbeakerscontainingthesolidsaltswere were conducted in saturated KNO3 solutions from 90 to placed into the chamber and sequential pairs of initially 110?C. Statistical analysis of these RH measurements to- aqueousandinitiallysolidbeakersweresampledovertime. gether with vapor pressure measurements for KNO3 ?Ref. Figure6showsthatsteady-staterelativehumidityandsolu- 12? yield an average standard deviation of about 1.6 RH tion composition are achieved with in the ,rst 200 h of re- units.Uncertaintyreportedinthetablesand,guresiscalcu- action and that solution composition of initially dissolved latedas2??3.1%RHunits?. FIG.6. Examplesofthebrine%rela- tive humidity and composition as a functionoftimeforthereverseddeli- quescenceexperimentsat90?Cfor?a? NaCl–NaNO3 and the ?b? KNO3–NaNO3 mixtures. 24 Geochem.Trans.,Vol.6,No.2,June2005 Carrollet al. TABLEII. Solution%RHandcompositionfromreverseddeliquescenceofNaCl–NaNO3 mixturesat90?C. Reportedvaluesrepresenttheaveragesteady-statesolutioncomposition. %RH Expt.ID T(?C) ?3.1 Namolal Clmolal NO3 molal ?NO3?/?NO3???Cl Initiallydissolvedsalt MS-9A 91 74.6 7.44?0.01 6.60?0.03 ?0.00 ?0.00 0.850.11?0.16 9.70MS-10A 90 69.3 5.29?0.01 4.50?0.02 0.46?0.00 MS-11A 90 57.6 19.09?0.09 1.83?0.00 17.30?0.01 0.91?0.00 MS-12A 90 59.6 16.74?0.98 3.56?0.27 13.31?0.62 0.79?0.00 MS-13A 90 55.8 18.14?0.41 2.79?0.10 15.10?0.13 0.84?0.00 Initiallysolidsalt MS-9B 91 74.7 7.74?0.02 6.48?0.00 ?0.00 ?0.00 1.270.16?0.25 9.75MS-10B 91 69.2 4.86?0.10 4.86?0.12 0.50?0.00 MS-11B 91 57.6 19.20?0.06 1.87?0.00 17.68?0.08 0.90?0.02 MS-12B 91 59.4 16.59?0.12 3.39?0.08 13.54?0.08 0.80?0.01 MS-13B 91 55.7 18.70?0.48 2.84?0.03 15.73?0.18 0.85?0.00 Samples were taken when there was at least 10 g of caluncertaintyfortheICP-AESandtheICislessthan2%. solution in each beaker. The exact amount of solution wasSolidswereanalyzedusingpowderx -raydiffractionforde- dependentonrelativehumidityandsolubilityofthesaltmix- tectionofhalite,sylvite,sodaniter,andniter.Traceamounts ture.About 1 g of solution was collected from each beakerofaspeci ,edsolidphaseindicatethattheintensityofmajor and,lteredthrougha0.45?msyringe-less,lterintoasam- peak was generally less than 5% of the most intense peak. plingbottle,thebottlewassealedandweighedtodetermine We suspect that the small amount of the minor component sampleamount,andthendilutedwithabout250gofdistilled represents solution trapped during the ,ltration process and and de-ionized water. All dilution factors were determined notanequilibriumsolid. gravimetrically. Each time a pair of samples was taken, the temperature and relative humidity probes were moved to IV.RESULTS monitorthenextpairofbeakers.Solidsfromthefourthpair The deliquescence of NaCl–NaNO3, KNO3–NaNO3, of samples were separated from the remaining solution by andNaCl–KNO3saltmixturesfrom90to120?Cissumma- ,ltration?0.45?mporesize?,dried,andstoredinadesicca- rized in Tables II–VI and in Figs. 1–5. Figures 1 and 2 toruntilanalyzedbypowderx-raydiffraction. compare the experimental results of the reversed deliques- cence experiments with the model calculations for the C.Analyticaltechniques NaCl–NaNO3–H2Osystemat90and110?C.Equilibriumis shownbytheconvergenceofthemeasuredrelativehumidity All aqueous samples were analyzed by inductively andsolutioncompositionfortheinitiallydissolvedandini- coupled plasma-atomic emission spectrometry ?ICP-AES? tially solid salt mixtures.At 90?C, we also show previous for sodium, and by ion chromatography ?IC? for chloride, nitrate,potassium,and?forsomesamples?sodium.Analyti- solubility data by Linke,13 which agrees with our results. TABLEIII. Solution%RHandcompositionfromreverseddeliquescenceofNaCl–NaNO3mixturesat110?C. Reportedvaluesrepresenttheaveragesteady-statesolutioncomposition. %RH ?NO3?/ Expt.ID T(?C) ?3.1 Namolal Clmolal NO3 molal ?NO3???Cl? Initiallydissolvedsalt MS45-A 110 52.3 23.25?0.44 ?0.00 20.75?0.06 ?0.01 2.340.903.17?0.01 MS46-A 110 56.3 19.23?0.20 16.01?0.25 0.83?0.00 MS47-A 110 60.3 15.96?0.20 3.63?0.01 12.64?0.15 0.78?0.00 MS48-3A 110 69.6 9.53?0.01 5.62?0.00 4.07?0.02 0.42?0.02 MS49-4A 110 65.6 12.54?0.02 4.65?0.01 7.97?0.03 0.63?0.01 MS50R-A 110 54.9 22.02?0.55 0.43?0.01 21.26?0.58 0.98?0.01 Initiallysolidsalt MS45B 110 51.5 22.84?0.03 ?0.00 20.56?0.03 ?0.01 2.530.89?0.01 3.06MS46-B 110 55.8 19.77?0.30 16.57?0.25 0.84?0.00 MS47-B 110 59.6 16.48?0.27 3.42?0.08 13.30?0.18 0.80?0.00 MS48-3B 110 69.0 9.99?0.00 5.25?0.01 4.79?0.01 0.48?0.01 MS49-4B 110 64.8 13.62?0.03 4.07?0.01 9.65?0.00 0.70?0.01 MS50R-B3 110 54.5 22.34?0.02 0.87?0.00 21.37?0.01 0.96?0.00 Geochem.Trans.,Vol.6,No.2,June2005 Deliquescencesaltmixturesathightemperatures 25 TABLEIV. Solution%RHandcompositionfromreverseddeliquescenceofNaNO3–KNO3mixturesat90?C. Reportedvaluesrepresenttheaveragesteady-statesolutioncomposition,unlessotherwisenoted. Expt. %RH ?Na?/ 3 molal T(?C) Namolal Kmolal NOID ?3.1 ?Na???K Initiallydissolvedsalt MS-14A1 90 65.5 3.35?0.01 ?0.04 22.02?0.076 ?0.00 18.150.16 ?0.13 18.27MS-14A 90 67.5 1.44?0.39 20.40?0.12 0.07 ?0.02 MS-15A 90 58.6 ?0.10 1.01?0.04 20.23?0.64 0.95 ?0.00 18.81 MS-16A 90 55.5 20.36?0.95 2.54?0.26 23.02?0.90 0.89 ?0.01 MS-17A 90 55.6 10.77?0.59 19.67?1.47 31.04?2.52 0.35 ?0.01 MS-18A 90 50.3 14.84?0.41 20.16?0.36 35.47?0.86 0.42 ?0.00 MS-19A 90 50.6 21.40?0.26 9.00?0.50 30.89?0.80 0.70 ?0.01 MS-20A 90 45.0 18.67?0.27 21.65?0.72 40.24?0.68 0.46 ?0.01 MS-21A 90 45.6 22.82?0.98 16.43?0.42 39.17?1.36 0.58 ?0.01 Initiallysolidsalt MS-14B 90 67.4 2.36?0.17 ?0.63 ?0.63 ?0.00 17.9320.820.116 MS-15B 90 58.3 18.53?0.79 1.12?0.23 20.47?1.32 0.943?0.01 MS-16B 90 55.2 20.32?0.98 3.33?0.30 23.36?0.73 0.860?0.01 MS-17B3 90 55.3 10.94?0.00 18.58?0.33 30.01?0.03 0.371?0.01 MS-18B3 90 50.4 14.44?0.05 19.72?0.05 33.66?0.02 0.423?0.01 MS-19B 90 50.7 21.66?0.60 9.80?0.90 31.84?0.66 0.689?0.02 MS-20B 90 45.5 19.71?0.46 22.53?0.40 42.48?0.71 0.467?0.01 MS-21B 90 45.9 22.54?0.58 16.37?0.15 38.95?0.94 0.579?0.01 Although there is good agreement between experiment and tive humidity by as much as 8% relative humidity and un- model for solutions XNO3?0.5, there are discrepancies for derpredictsthesolutioncompositionbyasmuchas8molal both solution composition and relative humidity for nitrate- ?or40%?.At90?C,theexperimentaldatasuggestadeliques- richsolutionsneartheeutonic.Themodeloverpredictsrela- cencerelativehumidityof56%RH(XNO3?0.85) compared TABLEV. Solution%RHandcompositionfromreverseddeliquescenceofNaCl–KNO3 mixturesat90?C. Reportedvaluesrepresenttheaveragesteady-statesolutioncomposition. %RH ?Na?/ Expt.ID T(?C) ?3.1 Namolal Kmolal Clmolal NO3 molal ?Na???K Initiallydissolvedsalt MS-24A3 90 62.4 2.68?0.00 16.29?0.02 ?0.00 ?0.01 ?0.00 2.5416.840.14?0.01 6.07MS-25A4 90 6.56?0.04 5.82?0.01 6.58?0.01 0.48?0.01 MS-26A 90 4.75?0.14 15.36?0.20 4.71?0.08 15.95?0.30 0.24?0.01 55.3 MS-27A 90 7.07?0.34 11.97?1.04 5.80?0.08 13.64?1.49 0.37?0.01 54.7 MS-28A 90 1.55?0.11 18.09?0.04 1.63?0.06 18.07?0.05 0.08?0.01 65.5 MS-29A 90 5.71?0.09 3.71?0.23 5.79?0.05 3.74?0.22 0.61?0.01 65.3 MS-30A 90 5.43?0.00 0.42?0.07 5.53?0.01 0.44?0.06 0.93?0.01 71.5 MS-31A 90 5.96?0.25 15.85?0.39 5.06?0.09 16.74?0.40 0.27?0.00 52.4 MS-32A4 90 7.1712.485.5114.460.3652.4 ?0.02 ?0.07 ?0.02 ?0.00 ?0.01 MS-35A4 90 56.4 6.61?0.07 9.90?0.03 5.93?0.00 10.29?0.05 0.40?0.01 MS-36A 90 6.98?0.43 16.68?0.38 5.35?0.06 17.91?0.05 0.29?0.02 51.5 MS-37A 90 50.7 8.00?0.06 14.91?0.07 5.56?0.04 17.59?0.12 0.35?0.00 Initiallysolidsalt MS-24B3 90 62.5 2.72?0.02 16.96?0.04 2.82?0.00 17.29?0.02 0.14?0.00 ?0.01 6.22MS-25B4 90 62.8 6.61?0.06 5.89?0.04 6.68?0.01 0.48?0.01 MS-26B 90 55.6 4.99?0.05 15.40?0.37 4.93?0.04 16.08?0.22 0.24?0.00 7.17?0.40 MS-27B 90 55.7 12.48?1.45 5.73?0.09 14.11?1.91 0.37?0.01 1.6418.101.8218.080.08MS-28B 90 65.7 ?0.09 ?0.20 ?0.09 ?0.21 ?0.00 MS-29B 90 65.6 5.81?0.02 4.09?0.21 5.80?0.01 4.08?0.23 0.59?0.01 5.54?0.02 MS-30B 90 71.9 1.03?0.04 5.60?0.02 1.05?0.03 0.84?0.00 MS-31B 90 52.7 6.57?0.16 16.04?0.04 5.34?0.11 17.28?0.07 0.29?0.00 MS-32B4 90 52.8 7.47?0.02 13.08?0.08 5.43?0.01 15.33?0.01 0.36?0.00 6.91?0.06 MS-35B4 90 55.8 10.57?0.07 5.60?0.02 11.54?0.02 0.40?0.01 MS-36B 90 50.9 7.53?0.20 16.50?0.22 5.52?0.11 18.18?0.28 0.31?0.00 MS-37B 90 51.1 15.12?0.11 5.58?0.05 17.70?0.18 0.35?0.00 8.06?0.07 26 Geochem.Trans.,Vol.6,No.2,June2005 Carrollet al. TABLEVI. Solution%RHandcompositionfromreverseddeliquescenceofNaCl–KNO3mixturesat120?C. Reportedvaluesrepresenttheaveragesteady-statesolutioncomposition. %RH ?Na?/ Expt.ID T(?C) ?3.1 Namolal Kmolal Clmolal NO3 molal ?Na???K Initiallydissolvedsalt MS-34A?44? 120 43.0 8.35?0.00 17.19?0.32 ?0.00 ?0.05 ?0.01 6.1420.220.33?0.03 10.98MS-34A?41? 120 41.6 27.72?0.06 5.95?0.00 33.43?0.14 0.28?0.01 MS-34A?48? 120 48.1 8.18?0.06 16.55?0.25 6.09?0.02 19.27?0.06 0.33?0.01 MS-39A 119 47.3 9.14?0.07 20.33?0.13 6.09?0.04 23.85?0.09 0.31?0.01 MS-40A 119 47.9 5.54?0.05 30.26?0.08 5.53?0.00 30.38?0.13 0.16?0.00 MS-41A 119 45.7 9.86?0.05 22.09?0.03 6.22?0.02 26.44?0.06 0.31?0.01 MS-42A2 119 43.8 7.35?0.02 30.14?0.14 5.60?0.00 31.62?0.03 0.20?0.00 MS-43A2 119 43.3 10.65?0.03 24.90?0.04 6.00?0.03 29.29?0.17 0.30?0.00 MS-44A?40? 119 41.9 8.57?0.01 29.30?0.18 5.71?0.00 32.51?0.03 0.23?0.00 MS-44A?41? 119 42.9 8.19?0.01 29.28?0.09 5.45?0.00 31.94?0.05 0.22?0.00 Initiallysolidsalt MS-33B?48? 120 48.7 4.64?0.03 27.60?0.22 ?0.03 ?0.03 ?0.00 4.6027.940.14 MS-34B?44? 120 45.3 9.45?0.09 26.16?0.16 5.74?0.00 30.67?0.07 0.27?0.01 MS-34B?41? 120 41.0 10.89?0.03 28.53?0.11 5.92?0.00 34.26?0.06 0.28?0.01 MS-34B?48? 119 49.0 8.11?0.02 18.20?0.27 5.76?0.02 21.47?0.03 0.31?0.01 MS-38B 119 46.9 5.30?0.01 28.48?0.16 5.31?0.02 28.50?0.03 0.16?0.00 MS-39B 120 46.9 10.11?0.02 23.21?0.01 6.15?0.03 28.05?0.00 0.30?0.01 MS-40B 119 46.2 5.84?0.01 31.11?0.01 5.86?0.00 31.62?0.16 0.16?0.00 46.1 MS-41B 119 10.02?0.03 23.20?0.06 6.01?0.02 27.25?0.15 0.30?0.00 MS-42B2 119 7.37?0.01 31.58?0.16 5.83?0.00 33.11?0.01 0.19?0.00 MS-43B2 43.6 120 11.24?0.01 26.87?0.05 6.36?0.01 31.73?0.08 0.30?0.00 119 9.11?0.05 28.91?0.31 5.69?0.01 32.42?0.02 0.24?0.00 MS-44B?40? 41.3 MS-44B?41? 42.4 119 8.37?0.03 29.79?0.51 5.72?0.01 32.65?0.10 0.22?0.00 versed deliquescence experiments with the model calcula- tothemodelpredictionof59%RH(XNO3?0.92). Similarly tionsfortheKNO3–NaNO3–H2Osystemat90?C.Similar at 110?C, the experimental data suggest a deliquescence totheNaCl–NaNO3–H2Osystem,theconvergencebetween relative humidity of 52% RH (XNO3?0.9) compared to the themeasuredrelativehumidityandsolutioncompositionfor modelpredictionof56%RH(XNO3?0.97). initially dissolved and initially solid salt mixtures indicates AlthoughtherearediscrepanciesintheabsoluteRHand thatequilibriumwasachieved.However,thereispooragree- solution composition, both experiment and model exhibit similartrends.Relativehumiditydecreasesfromahighvalue ment between experiment and model for both the relative near75%atlowXNO3 toaminimumneartheeutonicpoint. humidity and solution composition. Experimental relative Abovetheeutonicpoint,therelativehumidityincreaseswith humidity values are as much as eight percentage points increasingXNO3asthedeliquescencepointofpuresodaniter higherthanthosepredictedbythemodelonthenitersideof is approached.The higher solubility of soda niter generates theeutonic(XNa?0.2). Trendsintheexperimentaldatain- nitrate concentrations that are substantially higher than the dicatethatthedeliquescencerelativehumidityofabout42% chloride concentrations above XNO3?0.5. Chloride concen- agrees with the model prediction, but yields a more KNO3 trations decrease with increasing XNO3, because chloride richbrine?aboutXNa0.5?thanpredictedbythemodel. solubilityislimitedbytheincreasingsodiumconcentrations A large discrepancy between experiment and model is fromdissolvingsodaniter?thecommonioneffect?. seen in the solution composition. Dissolved potassium, so- Thesolidsconsistedofhalitewithtracesodaniterbelow dium,andnitrateconcentrationsfollowsimilartrendsasthe the eutonic and of soda niter with trace halite above the model predictions, but the absolute concentrations are sig- eutonic.Traceamountsofsodaniterandhaliteprobablyrep- ni,cantlyhigher.Inthemostextremecase,solutioncompo- resentresidualsolutionthatwastrappedinporespacesdur- sitions are roughly twice the model prediction with experi- ing the ,ltration process when the salts were dried. Halite mental sodium?20 molal, potassium?22 molal, and should be the only solid phase present below the eutonic nitrate?42 molal. The high experimental molal concentra- because the solution is saturated with respect to halite and tionsarenotanartifactofderivingthevaluesfromsolution undersaturatedwithrespecttosodaniter.Abovetheeutonic, analyses ?see Sec. IIID?.Atest of the analytical methodol- sodanitershouldbetheonlysolidphasepresent,becausethe ogyusinghighlyconcentratedsolutionsshowedthatanalyti- solutionissaturatedwithrespecttosodaniterandundersatu- cally and gravimetrically determined values were generally ratedwithrespecttohalite.Onlyattheeutonic,whereboth within 4% of each other.The solids consisted of niter with mineralsaresaturated,wouldoneexpectto,ndbothhalite tracesodaniter,belowtheexperimentaleutonicandofsoda andsodaniter. niterwithtraceniter,abovetheexperimentaleutonic.Trace Figure 3 compares the experimental results of the re- amounts of soda niter or niter probably represent residual Geochem.Trans.,Vol.6,No.2,June2005 Deliquescencesaltmixturesathightemperatures 27 solutionthatwastrappedinporespacesduringthe,ltration deliquescence of these salt mixtures. Before we discuss the processwhenthesaltsweredried. speci,cdataneedstoresolvediscrepanciesbetweenexperi- mental results and model predictions, we review the basic Figures4and5comparetheexperimentalresultsofthe high-temperaturePitzermodelusedinthesesimulations.8 reverseddeliquescenceexperimentswiththemodelcalcula- 3–H2O system at 90 and 120?C. tions for the NaCl–KNOThePitzermodelisderivedby,rstde,ninganexpres- sionfortheexcessGibbsfreeenergy(Gex) ofthetotalso- Similar to the other salt systems studied here, the conver- gence between the measured relative humidity and solution lution?Ref.9,Eq.?23??: composition for initially dissolved and initially solid salt ex Gmixturesindicatesthatequilibriumwasachieved.Valuesfor ?f?I?? ?ij?I?ninj ?? RTww i j thedeliquescencerelativehumidityareinagreementatboth temperatures.Thisagreementattheeutonicrelativehumidity appears to be fortuitous, because there is poor agreement ijkninjnk, ?8? ?? ??? i j k betweenexperimentandmodelforrelativehumidityandall solution composition. The extent of the mismatch is much where Gex is the difference or ‘‘excess’’in the Gibbs free greater at 120?C than at 90?C.At 90?C, the greatest mis- energybetweenarealsolutionandanidealsolutionde,ned match occurs in KNO3-rich solutions where the model un- on the molality composition scale, R is the universal gas derpredictsrelativehumiditybyasmuchas,vepercentage constant,Tistheabsolutetemperature,ww isthemass?kg? pointsandunderpredictssolutioncompositionbyasmuchas ? ofsolventwater,fisaDebye–Huckel functionthatdepends 2 4 molal ?about 30%?. In solutions dominated by NaCl on the ionic strength (I?1/2?imiz ), ?ij is a second-order i (XNa?0.5), there is reasonable agreement between model interaction coef,cient ?also a function of I?, ?ijk is a third- and experiment. At 120?C, the mismatch between experi- orderinteractioncoef,cient,ndenotesthenumberofmoles mentandmodelrelativehumidityissimilartothatat90?C. of a species, and i, j, and k denote solute species. Here m However,atthishighertemperature,themodelsigni,cantly denotesmolalityandzthechargenumber.Notethatmolality underpredictsthesolutioncomposition.Inthemostextreme isde,nedasmi?ni/ww. Uponsubstitutionofmiww forni case,solutioncompositionsareroughlytwicethemodelpre- ?andsoforthforthejandkcases?inEq.?8?,theionicsolute dictionwithexperimentalsodium?11molal,potassium?28 activity coef,cient (?i) and the solvent osmotic coef,cient molal, chloride?6 molal, and nitrate?33 molal.At 120?C, ???maybecalculatedasthepartialderivativeswithrespect atmosphericpressurelimitstheexperimentsto50%relative to the molality of the ionic solute and the mass of water, humidity, so we cannot determine if the model adequately respectively: predictsrelativehumidityandsolutioncompositioninNaCl- ln?i????Gex/wwRT?/?mi?nw, ?9? richsolutions. Analysisofthesolidsshowthatsylvite?KCl?isanim- portant solubility control near the experimental eutonic in ex/?ww?n /RT ??1????G ?10 ? ? i agreement with the results of model calculations at these mi temperatures.At 90?C, niter, sylvite and minor amounts of i ?Ref.9,Eqs.?34?and?35??.Theactivityofwaterisclosely halite were detected just to the left of the eutonic (0.28 relatedtotheosmoticcoef,cient(lnaw??(?imi/?)? ,where рXNaр0.31), andhalite,sylviteandminoramountsofniter ?isthenumberofmolesofwatercomprisinga1kgmass, were detected to the right side of the eutonic (0.35рXNa approximately55.51?. р0.40). As is expected, the solids consisted of niter (XNa SubstitutionofEq.?8?intoEqs.?9?and?10?followedby р0.25) and halite (XNaу0.60) on their respective limbs of differentiation yields the fundamental Pitzer equations for thephasediagram.At120?C,solutionsweresaturatedwith thesoluteactivitycoef,cientandtheosmoticcoef,cient?or repect to niter, sylvite, and halite between most of samples alternativeformsfortheactivityofwaterortheactivityco- Naу0.30. One sample at XNa?0.14 con- between 0.15рXef,cientofwater?.SeeBSC8orPitzer9forfulldetailsofthe tained only niter, two samples at XNa?0.16 and XNa?0.30 appliedformsofthePitzerequationsandthecorresponding containedniterandhalite,butnosylvite.Anytraceamounts practicalinteractioncoef,cients.Valuesforthepracticalin- ofsaltdetectedprobablyrepresentresidualsolutionthatwas teractioncoef,cientsaregenerallyobtainedby,ttingphysi- trappedinporespacesduringthe,ltrationprocesswhenthe cal property measurements, such as the osmotic coef,cient, saltsweredried. thevaporpressureofwateroversaltsolutions,or?lesscom- monly?mineralsolubilities. (1) (2) The model requires only two-ion (?MX(0) , ? , ? , V.DISCUSSION MX MX C?MX, S? and S? ) andthree-ion(? and? ) MM? XX? MM?X MXX? A.Na–Cl, Na–NO3, andK–NO3 high-temperature interactionparametersinthepresentstudyofionicsystems. Pitzermodels HereMdenotesacation,M?adifferentcation,Xananion, and X? a different anion. The S? MM?, S?XX? The comparison of model predictions and experimental , ?MM?X , and parameters occur only in mixtures of aqueous elec- resultsofrelativehumidityandsolutioncompositionsforthe ? MXX? NaCl–NaNO3–H2O, KNO3–NaNO3–H2O, and the troltyes?e.g.,MX–M?X,MX–MX?,ormorecomplexmix- NaCl–KNO3–H2Osystemsfrom90to120?Cindicatethat tures?. Within the framework of the standard Pitzer model, someparametersusedinthecurrenthigh-temperaturePitzer the values of the mixing parameters are independent of the modeldonotadequatelydescribebrinechemistryformedby possiblepresenceofothertypesofionsinthesolution,and 28 Geochem.Trans.,Vol.6,No.2,June2005 Carrollet al. FIG. 7. Comparison of predicted and measuredniter(KNO3) ?a?%relative humidity and ?b? solubility as a func- tionoftemperature.Measureddataare from Kracek12 and Linke,13 respec- tively. Predictions were made using EQ3/6 version 8 geochemical code and theYuccaMountainhightemperatured Pitzer ion interaction thermodynamic database?Refs.6and8?. oncetheirvalueshavebeendeterminedforaparticularsys- Na–Cl–NO3, or K–Cl–NO3 ion interactions. Our experi- tem, the same values may be used for all other relevant mentaldataintheNaCl–NaNO3–H2Osystemindicatethat systems.9 temperature dependent parameters for Cl–NO3 and/or The high-temperature Pitzer model8 employed in the Na–Cl–NO3ioninteractionsareneededtodescribetherela- present study uses the following equation to represent the tivehumidityandthesolutioncompositionneartheeutonic temperaturedependenceofeachoftheseioninteractionpa- wheremaximumsolubilitiesareapproached.Theabsenceof rameters: temperature dependent parameters for K–NO3 ion interac- tionsintheYuccaMountainProjecthigh-temperaturePitzer 2 a3ln?T/Tr??a4?T?Tr?, ?11? model is the primary cause of the poor prediction of the ?a1? ??T??a?T?Tr? deliquescenceofsaltmixturescontainingKNO3 atelevated where?representsanyoftheabove-notedpracticalinterac- temperatures?Figs.3–5?aswellasthemeasured%relative tion parameters,13 Tr is the reference temperature ?298.15 humidity and solubility of KNO3 at elevated temperature K?,anda?Fig. 7?. Additionally, some of the mismatch in the 1, a2, a3, anda4 are,ttedcoef,cients.Avariety NaCl–KNOofotherformsareextantintheliterature?e.g.,Refs.10,14?. 3–H2O system may also be due to the absence Forthemodelusedhere,8datatakenfromothermodels?see of temperature dependent paramaters for Cl–NO3, the following? were re,t as necessary for consistency withNa –Cl–NO3, and/or K–Cl–NO3 ion interactions, in addi- the above temperature function. The differences associated tion to the absence of temperature dependent K–NO3 ion withre,ttingfromadifferenttemperaturefunctionareneg- interactionparameters. ligiblewithinthetemperaturerangesoftheoriginal,ts. Althoughthehigh-temperaturePitzermodelusedinthe present study is a fairly comprehensive one accounting for the nonideal behavior of highly concentrated electrolytes B.Implicationsforradioactivewastedisposal over a wide range of temperature ?nominally 0–200?C?, there are still signi,cant data needs for common ions. The It is important that geochemical calculations of the present model was founded on earlier high-temperature chemical environment at the waste package surfaces use a Pitzer models ?Refs. 10, 14, and references cited by themrobust high?, -temperature Pitzer ion interaction model for supplemented by parameter data from several other commonions,becausethedeliquescenceofaerosolsaltsand sources.dustislikelytobeaprimarysourceofbrinesthatcontactthe 11,15–24 It is also partly based on re,tting of param- etrizations from the published literature to the most widely wastecontainersanddripshields.Inadditiontothediscrep- used?standard?formofthePitzerequations.25 ancy identi,ed for relative humidity and solution composi- tion for salt mixtures containing KNO3 stemming from the The Yucca Mountain Project high-temperature Pitzer useofaconstanttemperaturemodelforK–NO3interactions, model contains robust thermodynamic submodels for the 3, Na–Cl, K–Cl, Na–K, andK–Na–Cl ioninterac- discrepanciesformagnesiumandammoniumsaltmixturesat Na–NO tions. The Na–NO3 model is based on data from ?37 to elevatedtemperaturesarelikelybecausethemodelincludes 152?Cconstanttemperaturemodelsformanyoftheirrespectiveion 11 that was re,t for consistency with the standard interactions.Pitzerform.Somedegradationin,tqualityresultsfromthis 8 Osmotic and activity coef,cients must be ex- re,tting, increasing the deviation in the osmotic coef,cient perimentallydeterminedasafunctionoftemperaturetode- rive the ion interaction parameters needed to describe the from about 0.01 to about 0.02.25 Similarly, two-ion and nonideal behavior of these concentrated solutions, because three-ionmodelsintheK–Na–Cl systemarebasedondata the Pitzer model is empirically based. Of the salt systems from 0 to 300?C10 ,t using the standard Pitzer form. In listedabove,robustmodelsforpotassiumsaltsareprobably contrast,theK–NO3andCl–NO3modelsarebasedononly 9 themostimportantathightemperature,becausemagnesium 25?Cdata andtherearenoparametersfortheNa–K–NO 3, Geochem.Trans.,Vol.6,No.2,June2005 Deliquescencesaltmixturesathightemperatures 29 FIG.8. Modelpredictionsofbrine?a? % relative humidity and ?b? solution compositionasafunctionoftempera- ture at the deliquescence point for a NaCl–NaNO3–KNO3 mineral assem- blage?Ref.8?.Predictionsweremade with EQ3/6 geochemical code and the hightemperaturePitzerioninteraction thermodynamic data base. Triangles, circles, squares, and diamonds repre- sent NO3, Na, K, and Cl concentra- tions. concentrations are likely to be limited by insoluble silicate 3–H2O, KNO3–NaNO3–H2O, and NaCl–NaNOminerals26 and ammonium concentrations are likely to be NaCl–KNO3–H2Osystemsfrom90to120?Cshowthatuse limitedbygasvolatility. of 25?C parameter values for K–NO3 and Cl–NO3 ?and RecentmodelingeffortsbytheYuccaMountainProject possiblyforthree-ionparametersintheNa–K–Cl–NO 3sys- 3 and/orNaNO3 arethe predictthatmixturesofNaCl,KNOtem?intheYuccaMountainProjecthigh-temperaturePitzer most prevalent mineral assemblages that may deliquesce in model do not accurately predict the equilibrium solubility repository environments using the current high-temperature and corresponding relative humidity of KNO3 salt mixtures Pitzer model.8 As the temperature increases, the deliques- andofNaCl–NaNO3mixturesneartheeutoniccomposition. cencerelativehumidityandtheNO3concentrationsincrease ?Fig.7?.Weexpectonlysmalldifferencesbetweenpredicted ACKNOWLEDGMENTS andactualdeliquescencerelativehumidityandbrinecompo- We thank Greg Gdowski, Joe Rard, MaureenAlai, and sition at temperatures below 60?C, because the model ad- QueAnhNguyenfortheircontributionstotheexperiments. equatelypredictsnitersolubilityatthesetemperatures.How- This work was performed under the auspices of the U.S. ever, at higher temperatures, much larger differences Department of Energy by the University of California, betweenpredictedandactualenvironmentareexpected,be- Lawrence Livermore National Laboratory under Contract causethemodelunderpredictsnitersolubilitybyabout200% No.W-7405-Eng-48. at135?C?Fig.7?b??.AtthesetemperaturesthecurrentPitzer model will signi,cantly underpredict brine NO3 brine com- position because the K–NO3 interactions will be more im- 1G.M.Gordon,Corrosion58,811?2002?. 2L.Greenspan,J.Res.Natl.Bur.Stand.,Sect.A81,89?1977?. portant.ThisisclearlyillustratedinourKNO3–NaNO3 and 3Z. Ge, A. S. Wexler, and M. V. Johnston, J. Phys. Chem. A 102 , 173 NaCl–KNO3 deliquescence experiments, where NO3 con- ?1998?. centrationscanbetwicethepredictedconcentrations.Addi- 4I. N. Tang and H. R. Munkelwitz, Atmos. Environ., Part A 27A , 467 tionally,itislikelythatcalculateddeliquescencerelativehu- ?1993?. 5I.N.TangandH.R.Munkelwitz,J.Appl.Meteorol.33,791?1994?. midities for the three salt system are of limited accuracy, 6BSC?BechtelSAICCompany?,ANL-EBS-MD-000001REV01,LasVe- becausethecurrentmodelforKNO3isoflimitedaccuracyat gas,Nevada,2004. high temperature ?Fig. 8?. Therefore calculated dry-out or 7T.J.WoleryandR.L.Jarek,2003.EQ3/6,Version8.0,SoftwareUser’s deliquescence temperatures are uncertain. Dry-out or deli- Manual, Software Document Number: 10813-UM-8.0-00, U.S. Depart- mentofEnergy,Of,ceofCivilianRadioactiveWasteManagement,Of,ce quescencetemperaturesmadeassumingatotalpressuresimi- of Repository Development, 1261 Town Center Drive, Las Vegas, NV lar to current atmospheric pressure ?0.90 bar? predict that a 89144. brine saturated with NaCl, NaNO3, and KNO3 would boil 8BSC ?Bechtel SAIC Company?, inANL-EBS-MD-000045 REV02. Las justabove135?C.6 Vegas,NV,BechtelSAICCompany,2004b. 9K. S. Pitzer, in Activity Coef,cients in Electrolyte Solutions, 2nd. ed., editedbyK.S.Pitzer?CRCPress,BocaRaton,FL,1991?,pp.75–153. VI.CONCLUSIONS 10J. P. Greenberg and N. Møller, Geochim. Cosmochim. Acta 53 , 2503 ?1989?. Adsorptionofwaterbydeliquescentsaltmineralsfound 11D.G.Archer,J.Phys.Chem.Ref.Data29,1141?2000?. inaerosolsandduststhatmaybedepositedonwastepack- 12F.C.Kracek,inInternationalCriticalTablesofNumericalData,Physics, agesurfacesduringtheconstructionandventilationstagesof Chemistry and Technology, edited by E. Washburn ?1928?, Vol. 3, pp. 351–374. ahigh-levelradioactivewasterepositoryatYuccaMountain, 13W.F.Linke,Solubilities:InorganicandMetal-OrganicCompounds.V.2, NV, will be a primary source of brines that might lead to K-Z., 4th ed. ?American Chemical Society, Washington, DC, 1965? , p. corrosion of the waste package surfaces.Although the deli- 200. 14N.Møller,Geochim.Cosmochim.Acta52,821?1988?. quescence relative humidity of most pure salt minerals is 15H.F.Holmes,R.H.Busey,J.M.Simonson,andR.E.Mesmer,J.Chem. knownoverarangeoftemperature,thebehaviorofsaltmix- Thermodyn.19,863?1987?. turesatelevatedtemperaturesisunknown.Ourreverseddeli- 16R. T. Pabalan and K. S. Pitzer, Geochim. Cosmochim. Acta 51 , 2429 quescent experimental results in saturated ?1987?. 30 Geochem.Trans.,Vol.6,No.2,June2005 Carrollet al. 22S. L. Clegg, S. Milioto, and D.A. Palmer, J. Chem. Eng. Data 41 , 455 17S.L.CleggandP.Brimblecombe,J.Phys.Chem.94,5369?1990a?. 18S.L.CleggandP.Brimblecombe,Geochim.Cosmochim.Acta54 ,3315 ?1996?. ?1990b?. 23H.F.HolmesandR.E.Mesmer,J.Chem.Thermodyn.30,723?1998?. 19W.E.ThiessenandJ.M.Simonson,J.Phys.Chem.94,7794?1990?. 24C.S.Oakes,A.R.Felmy,andS.M.Sterner,J.Chem.Thermodyn.32 ,29 20S.HeandJ.W.Morse,Geochim.Cosmochim.Acta57,3533?1993?. ?2000?. 21A.R.Felmy,C.C.Schroeder,andM.J.Mason,PNL-SA-25345,Paci,c 25J.A.RardandA.M.Wijesinghe,J.Chem.Thermodyn.35,439?2003?. 26M.Alai,M.Sutton,andS.Carroll,Geochem.Trans.?submitted?. NorthwestNationallaboratory,Richland,WA,1994.