1 TEMPERATURE DEPENDENCE OF HYSTERETIC HYDRATION/DEHYDRATION O F LAUMONTITE By GOKCE SENEM ATALAN A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGR EE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2013
2 2013 Gokce Senem Atalan
3 To my family
4 ACKNOWLEDG E MENTS I would like to truly thank my advisor Dr. Screaton for all her guidance support and under standing. Whenever I needed her she had always been there for me. I would like to extend those thanks to my former advisor Dr. Neuhoff for all the knowledge, support and help he provided during our studies. They both have been more than advisors to me and I feel blessed to have the chance to work with them. I would like to thank my committee members Dr. Martin and Dr. Zimmerman for their time and helpful comments; Laura Ruhl and Jie Wang for their help in the lab and for the scientific discussions. A ll my f riends through my studies, Jane, P.J., Kelly, Shawn, Mary Elodie, Alejandro, Laurel, Necibe, Esra Ece and all the others I could not name here ; the ones who were just a call away making me forget the thousands of miles between us Ilknur, Hediye and othe rs thank you for touching my life. I would like to give my sincere thanks to Debra Anderson, she was the one who always welcomed me with her warm smile, held my hand and pulled me out of rough times Many thanks to office staff for dealing with all the pa per work for me. I would like to thank National Science Foundation for supporting this project. I always believe family is the only real wealth in life; I am blessed in that manner. d like to thank my first love, my dad, and my true best friend, my mom, for everything they have done for me, for their never ending love support and encouragement. I always feel special to have parents like them always setting a great example of who I should be. I would like t o extend those thanks to my sister Sebnem, and my nephew Oytun, for bringing joy in my life cheering me up with all their positive energy L ast but not least I would like to thank my dear husband for being the best companion I could ask for T ogether w e have built
5 many memories, shared laughs, tears, joy and most importantly love. His love always warmed my heart made me feel like I am at the right place where I belong. Thank you for, no matter what, always be ing there for me.
6 TABLE O F CONTENTS page ACKNOWLEDG E MENTS ................................ ................................ ............................... 4 LIST OF FIGURES ................................ ................................ ................................ .......... 7 ABSTRACT ................................ ................................ ................................ ..................... 8 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 10 2 EXPERIMENTAL OBSERVATIONS ................................ ................................ ....... 15 Methods ................................ ................................ ................................ .................. 15 Results and Discussion ................................ ................................ ........................... 17 3 THERMODYNAMIC MODEL ................................ ................................ .................. 22 Model ................................ ................................ ................................ ...................... 21 Results and Discussion ................................ ................................ ........................... 26 4 CONCLUSIONS ................................ ................................ ................................ ..... 32 LIST OF REFERENCES ................................ ................................ ............................... 34
7 LIST OF FIGURES Figure page 2 1 Phase equilibrium observations of the total wa ter content of laumontite as a function of relative humidity obtained from salt buffer method. ........................... 20 2 2 Phase equilibrium observations of the total water content of laumontite as a function of re lative humidity obtained from thermogravimetric analyzer. ............ 21 3 1 Mole fraction of W1 versus activity of water ................................ ........................ 29 3 2 Gibbs energie s if mixing (G MIX ) as a function of mole fraction of W1 occupancy ................................ ................................ ................................ .......... 30 3 3 Solvus composition for the solid solution of fully hydrated laumontite and its partial dehydrate leonhardite ................................ ................................ .............. 31
8 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science TEMPERATURE DEPENDENCE OF HYSTERETIC HYDRATION/DEHYD R ATION OF LAUMONTITE By Gokce Senem Atalan May 2013 Chair: Elizabeth Screaton Major: Geology Laumontite is a common rock forming zeolite occurring world wide as a secondary mineral. Despite the geological importance of this mineral, its hydration/dehy dration behavior is poorly constrained. Among four distinct water binding sites in laumontite, the least energetic site, W1, w as previously shown to have reversible hysteretic hydration/dehydration behavior at room temperature but hysteretic behavior was n ot observed at higher temperatures. To study the influence of temperature on hysteresis, phase equilibrium observations were conducted across a range of temperatures (298 to 328 K) and as a function of relative humidity Water sorption was studied via two techniques, a salt buffer method and using a thermogravimetric analyzer. Prominent hysteresis was observed with both methods from room temperature to 323 K. As the temperature was increased, the observed hysteresis diminished and moved to higher humidity H ysteresis disappeared above 323 K. A solvus model of hydration/dehydration behavior was tested by finding the best fit equilibrium constant (K) and Margules parameter (W G ) for the experimental isotherm data. The thermodynamic calculation resulted in value s of log K and W G of 0.0495 and
9 5500 J/mol, respectively. These values together with thermodynamic data from the literature were used to predict the behavior of water during hydration/dehydration at room temperature and higher. Although fully matching the experimental results will require more complex modeling, the model reproduce s the basic fea tures of the observ ed isotherms, and thus, supports th e application of the solvus model to describe the temperature dependent nature of hysteresis observed during h ydration/dehydration of W1 in lau montite.
10 CHAPTER 1 INTRODUCTION Minerals containing intracrystalline molecular water in their lattice are very scales (e.g., Stober and Bucher, 2004). Water content of these hydrous minerals and their response to changes in physical and chemical conditions is of great importance in a wide variety of environmental and technological applications (Breck D. W., 1974; Sadek and Mekhamer 2000, 2 001; Selvidge and Miaoulis, 1990). Reversible hydration and thermodynamic properties play an important role in the stability and the chemistry of these minerals in geologic systems (e.g., Carey and Bish, 1996; Neuhoff and Bird, 2001). Although some mineral s reversibly hydrate and dehydrate following the same path, other minerals exhibit repeatable hysteresis during hydration and dehydration. The water content of these minerals depends on whether the system undergoes hydration or dehydration. Such behavior h as been reported in clays (Ransom and Helgeson, 1994), mineral nanopores (e.g., Neimark et al., 2000), and more rarely in zeolites (Fridriksson et al., 2003b, Atalan and Neuhoff, 2006; Bish and Wang, 2010). Water molecules in these confined systems exhi bit a wide range of physical behavior s in response to various surface forces. The behavior of confined water depends on many factors such as the structure and the chemical composition of the confining medium, the presence of charged species, and the hydrat ion level. The effect of the interactions with pore walls and other ions produce localized restrictions which alter the behavior of confined water from that of bulk water. The cause s of hysteresis ha ve been debated vigorously in the literature. It has bee n attributed to kinetic factors, phase transitions, structural changes, surface
11 irregularities, and surface area changes. (e.g., Tamura et al., 2000). It has also been ascribed to differences in surface interactions with water molecules and the confining m edium during capillary condensation and evaporation in nanoporous materials. (e.g., Neimark and Vavitkovitch, 2001). Despite concerted investigation of hydration/dehydration processes, quantitative thermodynamic models of the degree of hydration as a funct ion of temperature (T), pressure (P) and chemical potential () remain elusive for many of these systems Laumontite, Ca 4 Si 16 Al 8 O 48 .nH 2 O, is a common rock forming zeolite occurring world wide as a secondary mineral in low grade metavolcanic and volcanicall y derived sedimentary rocks. It forms under a wide range of conditions, from near surface and diagenetic to metamorphic and hydrothermal environments. The presence of laumontite in these settings is very important for pressure temperature boundaries of for mation and defin es upper zeolite facies metamorphism (Coombs et al. 1959 ). The occurrence of laumontite near petroleum deposits can significantly reduce rock permeability, which may limit the production potential of a reservoir (Neuhoff and Bird, 2001). X ray diffraction (Yamazaki et al,. 1991; Artioli and Stahl, 1993; Fridriksson et al., 2003a) and neutron diffraction (Stahl and Artioli, 1993) studies have revealed the crystal structure of laumontite. Fully hydrated laumontite contains 18 water molecules per unit cell (Yamazaki et al., 1991; Stahl et al., 1996). The aluminosilicate framework is made up of chains of four membered rings of (Al,Si)O 4 tetrahedra linked to other adjacent chains across O atoms in the rings ( Stahl and Artioli, 1993 ).They bond to form 5.3 A channels. The charge imbalance created by the presence of Al tetrahedral sites is neutralized by Ca 2+ cations. Water molecules are distributed among four distinct water
12 sites : W1, W2, W5 and W8. Water molecules on the W2 and W8 sites solvate t he Ca 2+ cation. In contrast, those on the W1 and W5 sites are hydrogen bonded to framework oxygen and/or water molecules on the W2 and W8 sites (Fridriksson et al., 2003a). The water sites in laumontite dehydrate sequentially (Stahl et al., 1996; Yamazaki et al, 1991; Fridriksson et al.,2003a). W1 is the least energetic and is the first site to dehydrate and the last water site to be occupied upon hydration of laumontite. When the water at W1 site is lost via dehydration the remaining water sites remain ful ly occupied. This stepwise behavior contrasts to many other zeolites that show complex hydration/dehydration behavior and allows study of the individual water sites At ambient conditions fully hydrated laumontite, which contains 18 water molecules per f ormula unit (Ca 4 Al 8 Si 16 O 48 18H 2 O), loses part of its hydrogen bonded water molecules from the channels and becomes a partially dehydrated laumontite (Ca 4 Al 8 Si 16 O 48 1 4 H 2 O) (Artioli et al.,1989; Armbruster and Kohler,1992) which is refer red to as leonhardite Although leonhardite is not considered a valid mineral name any more (Coombs et al., 1997) it is still used in the literature to identify the partially dehydrated laumontite with one water site (W1) empty. Neuhoff and Bird (2001) showed that laumontite actually forms as leonhardite in many geological settings and only becomes fully hydrated at elevated pressures. Researchers have studied hydration and/or dehydration behavior of laumontite as a function of the relative humidity (Yamazaki et al.,1991), te mperature (Artioli et al.,1989; Stahl et al.,1996), pressure (Lee et al.,2004;White et al.,2004) and different exchangeable cations (Kiseleva et al. 1996; Rashchenko et al. 2012). These studies mostly focus on the structural changes and mechanisms upon hyd ration/dehydration
13 and did not consider the distinct water sites separately. Fridriksson et al. (2003a ; b) studied the hydration/dehydration of laumontite for all four distinct water sites on laumontite. At room temperature, they found that hydration/dehydr ation of the W5 site was completely reversible without hysteresis, but hydration /dehydration of W1 exhibited distinctly hysteretic sorption behavior at room temperature, consistent with observations of Yamazaki et al. (1991). However, at temperatures abov e 298 K, Fridriksson et al. (2003b) observed su rprising ly low hydration levels o f the W1 site and did not detect any hysteresis. T he results of crystal structure analysis with X ray diffraction (XRD) by Fridriksson et al. (2003b) show ed that with in the hy steretic region at 298 K, peak amplitudes were split. This observation provides evidence that two phases with different water occupancies coexist during hydration and dehydration of W1 site and supports a solvus model for the hydration dehydration behavior This study presents results of integrated phase equilibrium experiments of the hydration/dehydration behavior of W1 site in laumontite over a range of temperature and relative humidity to map out the extent of the hysteresis and better understand the wa ter content and stability of laumontite under environmentally germane conditions. Hydration states of laumontite samples were measured using two different experimental techniques. The first technique, buffered batch experiments, has been traditionally app lied by many researchers (Chou et al., 2002; Yamazaki et al., 1991) and is based on creating test conditions and collecting water content data manually. The second technique employs a thermogravimetric analyzer (Carey and Bish 1996; Fridriksson et al., 20 03b). The experimental data gathered through these two techniques were compared and combined with calorimetric observations from previous studies to
14 develop quantitative thermodynamic models of solution behavior that can be applied to predicting water cont ent of the mineral at any temperature and relative humidity. The phase equilibrium observations were used to assess the thermodynamic properties of the reactions and the energetic s of mixing between hydrated and dehydrated phases The thermodynamic calcula tions presented in this study enabled testing whether a solvus model can explain the hydration/dehydration observations If confirmed, this model can be applied to other hysteretic systems and allows prediction of the hydration state of these systems over a range of temperature s and relative humidity conditions.
15 CHAPTER 2 EXPERIMENTAL OBSERVATIONS Methods Phase equilibrium observations were conducted on samples of powdered natural laumontite collected from Drain, Oregon. The crystal structure of this sam ple has previously been studied extensively (Fridriksson et al., 2003a). We can express the equilibrium between laumontite and leonhardite by the reaction: ( 2 1) The stoichiometry of the reaction is based on crystallographic studies of Yamazaki et al. (1991) and Stahl et al. (1996) and expressed per 12 framework oxygen molecules for simplicity Two experimental techniques, buffered batch experiments and thermogravimetric analy ses were applied. For batch experiments, a modified versi on of the salt buffer method described by Chou et al. (2002) was adopted. Approximately 0.2 g of samples was placed in preweighed 0.5 ml Eppendorf tubes. The open Eppendorf tubes were then set in closed shell vials within scintillation vials containing th e buffer salt solutions of known relative humidity (RH). They were equilibrated at constant temperature for 1 day, sequentially over a series of saturated salt solutions that buffered the water vapor pressure to a range of relative humidity from 32.7 to 95 .8% (saturated solutions of K 2 SO 4 KCl, NaCl, NaNO 3 KI, CoCl 2 MgNO 3 NaI and MgCl 2 ; Greenspan, 1977) with precision better than 0.12%RH. Mass changes of three samples were monitored at each step. After equilibration under the highest humidity conditions, the procedure was reversed to study the dehydration behavior of the sample. To study the influence of
16 temperature, hydration and dehydration observations were conducted at room temperature (298 K) and at higher temperatures (308 K, 313K, 318 K 323 K and 3 28 K) by submersion in a temperature controlled water bath. To test the repeatability of the hydration dehydration reaction, samples were first equilibrated at the lowest RH and after that at the highest RH and this process repeated several times. The same water content was observed at the same RH each time exhibiting repeatable hydration and dehydration. Isothermal the rmogravimetric analyses were conducted using a TA Instruments Q5000 sorption analyzer. For each experiment, approximately 30 mg of sample in a metal coated quartz crucible was placed on the calibrated balance and equilibrated at 298 K and 50% RH for 3 hours. For the first part of the experiments, temperat ure was held constant while RH was varied from 50% up to 98% (hydr ation) and from 98% to 5 0% (dehydration) in steps of 2% RH. The mass was monitored during the whole experiment. For each step, the sample was held at constant RH until the mass change was less than 0.001 % for 10 minutes to achieve complete equilibrium. This experiment was repeate d for every 5 K increment except 303 K, up to 328 K. The humidity was controlled by mixing N 2 dry gas and water vapor (100% RH) Humidity verifications for the TA sorption analyzer involved testing deliquescence points (defined as the percent RH where th e derivative of mass change with respect to the RH is zero) of salts. Various salt sample s were monitored whil e RH was decreased linearly from 98 to 5% at a rate of 0.2%RH/min. The verification was accomplished by obtaining deliquescence points which were consistent with the theoretical values. Uncertainty in the measurement of RH is about 2%.
17 To test if any condensation occurred, pure quartz was used as a sample under the same conditions as the laumontite. The mass change of quartz was no more than 0.01% indicating no condensation occurred at high humidity. The temperature in the sample chamber was measured using a calibrated thermocouple. A constant air flow of 200 cm 3 /min through the sample chamber was maintained throughout testing. Results and Discuss ion Results of phase equilibrium experiments obtained from the salt buffer technique and the sorption analyzer are plotted at six applied temperatures (Figure 2 1 and 2 2). For each temperature completely repeatable hydration/dehydration isotherms were obt ained. The u pper gr ay bar shows the stoichiometry of a fully hydrated unit cell of laumontite and the lower gray bar shows the stoichiometry of a partially hydrated leonhardite unit cell representing the R eaction 2 1 At 298 K hydration and dehydration of the W1 site are characterized by a reversible broad, prominent hysteresis loop between 40 and 84% RH. Within the region of hysteresis, the hydration state of W1 is constrained depending on the path followed Although there is no apparent hysteresis at 323 K from the limited data obtained with salt buffer method (Figure 2 1), high resolution data obtained by the sorption analyzer (Figure 2 2) show hysteresis between 80% RH and 98% RH. Along the hydration isotherm all samples start with the same water stoic hiometry and follow the same path. The hydration starts at 70 % RH at room temperature, and 80% RH, 85 % RH, 90 % RH, 92 % RH, 95 % RH at 308 K, 313 K, 318 K, 323 K and 328K, respectively. Figures 2 1 and 2 2 clearly show the temperature dependence of the hy dration/dehydration isotherms and the hysteresis loop associated with it. As temperature increases, the hydration and dehydration isotherm branches get
18 closer, and the hysteresis loop becomes narrower, moves toward higher relative humidity and disappears a t a critical temperature This general trend of hysteresis is similar to the hysteretical adsorption isotherms of nanopores (e.g. MCM41: Morishige et al., 2007; Neimark et al., 2000). The maximum water occupancy on W1 is strongly dependent on temperature. At room temperature 100% occupation of W1 was reached with the sorption analyzer. Because the salt buffer method has a lower maximum RH value, the occupancy does not reach 100%. The water content of W1 decreases systematically with increasing temperature and at 328 K maximum water content corresponds to only about 10% occupancy The salt buffer technique assumes that the samples reach equilibrium in a given time frame. On the other hand, the thermogravimetric analyzer is based on validating the point of equilibrium by continuously measuring the mass of the sample. As a result, the experiments using the thermogravimetric analyzer take less time. The agreement between the methods, as well as the repeatability of the reactions, suggests that both techniques are reliable. The thermogravimetric analyzer provided better resolution through the hysteretic region and thus provided better insight, which was especially needed for higher temperature data The results of experimental observations of this study for the hydration/dehydration of W1 site on laumontite agree well with previous observations of Yamazaki et al. (1991) and Fridriksson et al. (2003b) at ambient conditions. However Fridriksson et al. (2003b) found no evidence of significant hydration of W1 at tem peratures above 298 K, whereas the data obtained in this study clearly confirm
19 hysteresis with both techniques. The low observed values of hydration in the Fridriksson et al. (2003b) results suggest methodological issues It is possible that either equilib rium was not reached at higher temperatures or that there was a problem with the anticipated humidity and/or temperature values. It is important to note that the hysteresis behavior observed in these experiments is not a result of kinetic effects. It is c ompletely repeatable and time independent regardless of the technique used to obtain the isotherms. In order to confirm that equilibrium has been reached through the hystere tic region samples were left to equilibrate up to 10 days and the results were not affected The salt buffer technique was preferred for those measurements since it allowed the required time for slow processes.
20 Figure 2 1 Phase equilibrium observations of the total water content of laumontite as a function of relative humidity at 298 K, 308 K, 313 K, 318 K, 323 K and 328 K obtained from salt buffer method. Error bars represent the average of three runs.
21 Figure 2 2 Phase equilibrium observations of the total water content of laumontite as a function of relative humidity at 298 K, 308 K, 313 K, 318 K, 323 K and 328 K obtained from thermogravimetric analyzer.
22 CHAPTER 3 THERMODYNAMIC MODEL Model Thermodynamic description of the stability of confined water has traditionally followed two approaches. The first one describes the wat er stability by means of the Kelvin or Laplace equations and variants. This approach describes the effect of pore size and energy of interaction between water and the pore wall and the condensation vapor pressure of water in the pore (e.g., Mercury and Tar dy, 2001; Mercury et al., 2000; Neimark and Ravitkovitch, 2001). It has been very successful in describing the macroscopic behavior of water in nm scale pores, including isotherm form, hysteresis, and pore size effects. The limitation of this approach is t hat it does not consider the thermodynamic properties and stability of coexisting bulk water phases or the confining material. The second approach adapts standard geochemical thermodynamic equation s to consider the hydrous and anhydrous (or partially hydro us) forms of the confining material as members of a solid solution. This approach has been successfully applied to hydration/dehydration systems (Ransom and Helgeson, 1994; Fridriksson et al., 2003b;Wang and Neuhoff, 2008; Vieillard et al., 2011) In this study, we employ the second approach and treat the partial hydration of the W1 site as a solid solution of completely hydrous (W1 site full, laumontite) and partially hydrated (W1 site empty, leonhardite) components which differ in composition only by the presence or absence of 1 mole of H 2 O. This enables thermodynamic analysis of experimental data and allows correlation of the thermodynamic consequences of hydration/dehydration at higher temperatures. The equilibrium constant, K for R eaction 2 1 can be expressed as
23 (3 1) where is the activity of H 2 O in the vapor phase, and represent the activities of leonhardite and laumontite. If we assume ideal gas behavior of water vapor under the low p ressure conditions of this study water activity can be taken as equal to water vapor pressure. Therefore, relative humidity conditions used to obtain isothermic data can be considered as the activity of liquid water at 298 K, 1 bar (Ransom and Helgeson, 1 994) The activities of the laumontite end member and leonhardite end member are related to their concentrations (expressed in terms of mole fractions, X) via the relations (3 2) (3 3) wh ere X laumontite is equal to the mole fraction of W1 occupancy and is the activity coefficient Since the total occupancy of the W1 site is unity X leonhardite is equal to (1 X laumontite ). Because the fully occupied W1 site contains 1 mole of water, the mole fractional occup ancy of W1 can be calculated for each data point by (3 4) where represents the water content of laumontite. The equilibrium constant of the reaction can be determined using equation 3 1 and utilizing the activities from the eq uilibrium observations.
24 (3 5) Once the equilibrium constant is known it can be used to determine the standard Gibbs r,T,P ). The equilibrium constant at a given temperature (T) and pressure (P) is related to r,T,P ) at those conditions by ( 3 6 ) where R is the gas constant (8.314 J/mol K) and T is temperature in Kelvin. is related to the standard entropy and enthalpy of reaction at a given T and P ( and respectively) by ( 3 7 ) T he properties of mixing between the hydrated and anhydrous states are represented by which relates degree of hydration to the respective activities. For systems like laumontite that show hysteretic behavi or models that assume ideal behavior are not appropriate and deviation from ideal behavior affects the thermodynamic properties. The Gibbs free energy of mixing ( G MIX ) can be described as the sum of two parts: the free energy associated from ideal mixing (G ideal ) and the energy arises from non ideal mixing. The difference between the property of a real solution and the property if the solution was ideal is called an excess property, G EX (3 8 ) is related to t he excess Gibbs energy of mixing (G EX ) via the relation : (3 9 )
25 The Gibbs free energy of the reaction was calculated from the experimental observations of reversible equilibrium between laumontite and leonhardite as a function of hu midity. Following the metho d of Ransom and Helgeson (1994), a n empirical one parameter mixing model wa s employed to describe the non ideal mixing of the solid solution between fully hydrated laumontite and its partial dehydrate leonhardite from the experim ental data. (3 10 ) where W G is an empirical temperature independent mixing parameter. (Margules parameter). According to the Margules formulation, activity coefficients can be written as (3 11 ) (3 12 ) Equations 3 11 and 3 12 can then be solved from the room temperature experimental data for and substituted into equations 3 1, 3 2 and 3 3 for the purpose of regression of K and W G The thermodynamic properties determined by the Margules equations were then used to predict the activity of confined water along the isotherm curves. T he t emperat ure variation of K is estimated from van t Hoff eq uation using an enthalpy of hydrati on ( H) of 8800 J/mol from Fridriksson et al (2003 b ) (3 13 )
26 Results and Discussion Figure 3 1 shows the simulated activity of water confined in laumontite with varying extends of hydration of W1 site as a function of RH. The curve is compared to the experimental data at (a) 298 K (b) 308 K (c) 313 K (d) 318 K (e) 323 K and (f) 328 K using the regressed values of log K and W G of 0.0495 and 5500 J/mol, respe ctively. The model yields a complex relationship between hydration and relative humidity. The S shaped curve bends backwards such that more than one possible X W1 exist s in some ranges of RH. The model is consistent with the hydration data along its lower p ortion and the dehydration data along its upper portion. As the temperature increases, the simulated S shape becomes less prominent, gets smaller and shifts to higher humidity range. The predictions are in good agreement with experimental data at low tempe ratures and fit more poorly at higher temperature. The calculated Gibbs energy of mixing (G MIX ) as a function of mole fraction of W1 (Figure 3 2) provides further insight on the hysteresis behavior. The two minima that appear in the Gibbs free energy curv e at low temperatures have lower values than the end members, prevent ing formation of compositions between them. This is because the tangent of the curve (dashed line) touches the curve at two points (a and b) and also means that two phases can coexist at the same time because the chemical potential of both phases have the same value. These points represent the coexisting stable compositions in the solution at a given temperature. This behavior gives rise to the presence of a solvus, with the energy region between the minima being compositionally prohibited. If the bulk composition of the system lies between these minima, the system will consist of a mechanical mixture of phases with compositions corresponding to the minima. Any composition between a and b is metastable with respect to a mixture of a
27 and b. As the temperature rises t he relative contribution of the excess free energy term dec reases and the curve gets smoother, causing the minima in the solution curve to move closer The two minima continue t o approach each other and form one minimum at the critical temperature At this critical temperature, the energy barrier no longer exists. T he minima merge and every point on the solution curve has a free energy lower t han that of any mixture phases and a single homogeneous solution is stable for every composition. This is consistent with the lack of hysteresis observed at 328 K. Simulated bimodal compositions calculated from the first derivative of the free energy curve s (Figure 3 2), are shown as a func tion of temperature in Figure 3 3. The boundary between the two phase and one phase region is the solvus (also referred to as the bimodal curve). There is a two phase immiscible region at lower temperatures and inside the solvus curve two compositions coe xist. This successfully explains the coexisting phases observed at 298 K by Yamazaki et al. (1991) and Fridriksson et al. (2003b). However, the single composition tends to be more stable at higher temperatures, explaining why the hysteresis loop becomes pr ogressively smaller with increasing temperatures. At and above the critical temperature, the two phases become fully miscible. Conversely, at lower temperatures the two coexisting solutions become less miscible and their compositions move apart. The value of the critical temperature is in good agreement with observed data showing no hysteresis at 328 K. Thermodynamic modeling of these results indicates that the solid solution between laumontite with W1 full and W1 empty can be represented by a solvus with a critical temperature (Figures 3 2 and 3 3). The model was able to predict the hydration and dehydration isotherms in reasonable agreement with the measured data, and to
28 predict the disappearance of the hysteresis loop above 328 K However, there are some discrepancies that could be related to simplificat ions that were made during the calculations. First it has been assumed that the Margules parameter was independent of temperature. Secondly, the regular solid solution has been considered as a symmetrical solution. The results and fit of the model to the data could potentially be improved with a two parameter asymmetrical solution model with temperature dependent parameters.
29 Figure 3 1 Mole fraction of W1 versus activity of water, the curve is model fit to the experimental salt buffer data A) 298 K B) 308 K C) 313 K D) 318 K E) 323 K F) 328 K.
30 Figure 3 2 Gibbs energies if mixing (G MIX ) as a function of mole fraction of W1 occupancy for 298 K, 308 K, 313 K, 318 K, 323 K and 328 K
31 Fi gure 3 3 Solvus composition for the solid solution of fully hydrated laumontite and its partial dehydrate leonhardite
32 CHAPTER 4 CONCLUSIONS One of the goals of this study was to obtain reliable, high resolution data to map out the temperature dependenc e of hysteretic hydration/dehydration behavior of W1 site on laumontite. Careful measurements allowed quantification of the total amount of water adsorbed as a functi on of the %RH and temperature. The high quality, high resolution experimental data show th at hydration/dehydration of W1 site on laumontite is reversible and repeatable at all temperatures. The total water uptake of W1 site decreased with temperature from 100% at room temperature to 10% at 328 K. The hysteresis loop associated with hydration/de hydration shifted to higher humidity and diminished with increasing temperature before disappearing at a critical temperature. To enhance our understanding of hysteresis behavior, thermodynamic properties were derived from the experimental data to simulat e the temperature dependence of hydration/dehydration behavior. The model follows the hydration data along the lower limb and the dehydration data along the upper limb. The curve becomes less prominent as the temperature rises and appears at higher relati ve humidity following the same path with hysteresis observed. Thus, for any given temperature and relative humidity, it is possible to predict the number of water molecules on the W1 site. The solvus model nicely explains both the temperature dependence o f hysteresis and the presence of two coexisting phases. The calculated Gibbs free energy of mixing clearly exhibits an energy barrier prohibiting complete solid solution at lower temperatures, making the two coexisting solutions immiscible. At higher tempe rature the energy barrier becomes less distinguished and at critical temperature the energy barrier no longer exists.
33 The model provides a better understanding of the stability and reactivity of has a direct impact on the study of other hysteretic hydration/dehydration systems by permitting thermodynamic description of the stability of confined water. The methods and insights gained from this study can be used to develop new investigations of the relationship between hydration/dehydration processes and stability in other mineral systems.
34 LIST OF REFERENCES Armbuster T. and Kohler, T. (1992) Rehydration and dehydration of laumontite: a single crystal X ray study at 100 K. Neues Jahrbuch Fu r Miner alogie Monatshefte, 9, 385 397. Artioli G., Smith J.V., and Kvick A. (1989) Single crystal neutron diffraction study of partially dehydrated laumontite at 15 K. Zeolites, 9 (5), 377 391. Artioli G. and Stahl K. (1993) Fully hydrated laumontite: a structure study by flat plate and capillary powder diffr action techniques. Zeolites, 13 (4), 249 255 Atalan G., and Neuhoff P.S. (2006) Reversible conversion of tetranatrolite to paranatrolite under ambient conditi ons. Book of Abstracts, Zeolite International Conference on the Occurrence, Properties, and Utilization of Natural Zeolites), 50 51. Bish, D.L. and Wang, H. W. (2010) Phase transitions in natural zeolites and the importance of P H2O. Philosophical Magazine, 90 (17 18), 2425 2441 Breck D. W. (1974) Zeol ite Molecular Sieves: Structure, Chemistry and Use. John Wiley & Sons, London. Carey J. W. and Bish D. L. (1996) Equilibrium in the clinoptilolite H2O system. American Mineralogist, 81, 952 962 Chou I.M., Seal R.R.II, and Hemingway, B.S. (2002) Determina tion of melanterite rozenite and chalcanthite bonattite equilibria by humidity measurements at 0.1 MPa. American Mineralogist, 87, 108 114. Coombs D.S., Alberti A., and Armbruster T. (1997) Recommended nomenclature for zeolite minerals: Report of the subco mmittee on zeolites of the International Mineralogical Association, Commission on New Minerals and Mineral Names. Canadian Mineralogist, 35, 1571 1606 Coombs D.S., Ellis A.J., Fyfw W.S., and Taylor A.M. (1959) The zeolite facies, with comments on the inte rpretation of hydrothermal syntheses. Ceochim Cosmochim Acta, 17, 53 107 Fridriksson T., Bish D.L., and Bird D.K. (2003 a) Hydrogen bonded water in laumontite I: X ray powder diffraction study of water site occupancy and structural changes in laumontite d uring room temperature isothermal hydration/dehydration. American Mineralogist, 88, 277 287. Fridriksson T., Carey J., Bish D.L., Neuhoff P.S., Bird D.K. (2003 b) Hydrogen bonded water in laumontite II: Experimental determination of site specific thermodyn amic properties of hydration of the W1 and W5 sites. American Mineralogist, 88, 1060 1072.
35 Kiseleva I., Navrotsky A., and Belitsky I.A. (1996) Thermochemistry of natural potassium sodium calcium leonhardite and its cation exchanged forms. American Mineralo gist, 81 (5 6), 668 675 Lee Y., Hriljac J.A., Parise J.B., and Vogt T. (2004) Pressure induced migration of zeolitic water in laumontite. Physical Chemistry of Minerals, 31, 421 428 Mercury L., Vieillard P., and Tardy Y. (2000) Thermodynamics of ice polym orphs and 181. M ercury, L., and Tardy, Y. (2001) Negative pressure of stretched liquid water. Geochimica et Cosmochimica Acta, 65, 3391 3408. Morishige K., and Ishino M. (2007) Lowe r closure point of adsorption hysteresis in ordered mesoporous silica. Langmuir, 23, 11021 11026 Neimark A. V., Ravikovitch P. I., and Vishnyakov A. (2000) Adsorption hysteresis in nanopores. Physical Review 62, 1493 1496 Neimark, A.V., and Ravitkovitch P.I. (2001) Capillary condensation in MMS and pore structure characterization. Microporous and Mesoporous Materials, 44 55, 697 707 Neuhoff P.S., and Bird D.K. (2001) Partial dehydration of laumontite: thermodynamic constraints and petrogenic implicati ons. Mineralogical Magazine, 65 (1), 59 70 Rashchenko, S.V., Seryotkin, Y.V., and Bakakin, V.V. (2012) An X ray single crystal study of alkaline cations influence on laumontite hydration ability: II. Pressure induced hydration of Na,K rich laumontite. Micr oporous and Mesoporous Materials, 159, 126 131 Ransom, B., and Helgeson H.C. (1994) A chemical and thermodynamic model of aluminous dioctahedral 2:1 layer clay minerals in diagenetic processes: Regular solution representation of interlayer dehydration in smectite. American Jo urnal of Science, 294, 449 484. Sadek O.M., and Mekhamer W.K. (2000) Ca montmorillonite clay as thermal energy storage material. Termochimica Acta, 363, 47 54 Sadek O.M., and Mekhamer W.K. (2001) Na montmorillonite clay as thermal en ergy storage material. Termochimica Acta, 370, 57 63 Selvidge M., and Miaoulis I.N. (1990) Evaluation of reversible hydration reactions for use in thermal energy storage. Solar Energy, 44(3), 173 178. Stahl K., Artioli G. (1993) A neutron powder diffracti on study of fully deuterated laumontite. Eu ropean Journal of Mineralogy, 5 (5), 851 856
36 Stahl K., Artioli G., and Hanson J.C. (1996) The dehydration process in the zeolite laumontite: A real time synchrotron X ray powder diffraction study. Physics and Chem istry of Minerals, 23(6), 328 336 Stober I., and Bucher K. (2004) 151 Tamura K., Yamada H., and Nakazawa H. (2000) Stepwise hydration of high quality synthetic Smectite with various cations. Clays an d Clay Minerals, 48 (3), 400 404 Vieillard P., Blanc P., Fialips C.I.; et al. (2011) Hydration thermodynamics of the SWy 1 montmorillonite saturated with alkali and alkaline earth cations: A predictive model. Geochimica et Cosmochimica Acta, 75(19), 5664 5 685 Wang J. and Neuhoff P.S. (2008) Anomalous temperature dependence of the heat of hydration of natrolite. The Open Mineralogy Journal, 2, 60 65 White C., Ruiz Salvador A.R., and Lewis D.W. (2004) Pressure induced hydration effects in the eolite laumont ite. Angewandte Chemie International Edition, 43(4), 469 472 Yamazaki A., Shiraki T., Nishido H. and Otsuka R. (1991) Phase change of laumontite under relative humidity controlled conditions. Clay Science, 8, 79 86
37 BIOGRAPHICAL SKETCH Gokce Senem Atala n was born in May 1979 in Ankara, Turkey. She grew up in Balikesir, Turkey and attended Sirri Yircali Anatolian High School. Afterward, she attended Balikesir University and graduated with her Bachelo r of Science degree in p hysics e ducation. Gokce obtained her Master of Science degree in physics at Anadolu University where she worked as a teaching & research assistant She decided to continue her education at the University of Florida and was accepted by Department of Geological Sciences. She was awarded with a fellowship by the Turkish Higher Education Instit ute S he also worked as a teaching & research assistant at the department during her studies. She worke d in research projects under supervision of Dr. Philip S. Neuhoff and Dr. Elizabeth Screaton.