Geochemical Properties of Confined Aqueous Solutions
Benjamin Gilbert, Abdullah Cihan and Piotr Zarzycki
Overview: This project will determine how the confinement of aqueous solutions in thin films and nanoscale pores alters water properties (e.g., density, dielectric properties) in order to predict solute properties (e.g., solute speciation, acidity) and mineral properties (e.g., solubility). Molecular spectroscopy and calorimetry, complemented by simulation, will elucidate how the hydrogen-bond (HB) network of water is altered by confinement by neutral and charged surfaces. Neural-network enhanced molecular simulation methods will identify the time-averaged and dynamic properties of HB networks that govern fluid thermodynamics and solute chemistry. Theoretical descriptions of aqueous solution thermodynamics, response functions and reactivity will be developed that are compatible with Crunch modeling of mineral-fluid processes.
Background
Knowledge of the physical and chemical properties of aqueous solutions is a necessity for predicting the evolution of any rock-fluid system. Indeed, the establishment of models for multicomponent aqueous solutions (Shock et al. 1992), including their extension to the extreme pressure and temperature conditions of the deep crust (Sverjensky et al. 2014), is a major accomplishment of modern geochemistry. Due to the innate complexity of water, however, significant challenges remain. It is presently not possible to construct sufficiently accurate descriptions of bulk aqueous fluids from experimental observations or computational predictions of molecular-scale structure and dynamics. Modern molecular methods have revealed increasingly detailed insights into chemical processes such as solvation, solute association, proton and ion transport, and mineral-fluid ion exchange. Nevertheless, molecular simulations cannot reliably predict parameters such as the pKa of an acidic solute or the equilibrium coefficient of an ion pair in dilute or concentrated solutions. Moreover, subsurface aqueous fluids are often confined to nanoscale pores, films and clay interlayers. Although many studies have demonstrated that solid-water and fluid-water interfaces perturb water structure and properties (e.g., Knight et al. 2018), we still lack conceptual and quantitative models for the impacts of confinement upon water properties: the equation of state and thermodynamic functions, electromagnetic and rheological response functions, and the self-ionization constant. All of these parameters are likely to be altered in confinement in ways that will affect the chemical–mechanical processes tackled throughout this proposal.
Static Dielectric Constant of Water
The static dielectric constant of water, 𝜺(0), is particularly important for equilibrium geochemical simulations of fluid-mineral systems because of the direct influence on the Gibbs free energy for the solvation of aqueous species. Prediction of 𝜺(0) for bulk water by ab initio molecular simulation at pressure and temperature regions inaccessible in the laboratory (Pan et al. 2013) extended the ability to predict the compositions of aqueous solutions in the deep Earth (Sverjensky et al. 2014). Yet, the appropriate values of 𝜺(0) for interfacial under ambient conditions remains highly debated. Recently, Fumagalli et al. (2018) reported strong and long-ranged suppression of 𝜺(0). This is surprising because the length scale exceeds the distance where experiment (Prost et al. 1981), theory (Shock et al. 1992) and simulation (Schlaich et al. 2016; Zhang et al. 2013; Zarzycki et al. 2020) predict that time-averaged water structure is perturbed. Interface-sensitive experimental studies are challenging. Although interface-sensitive non-linear spectroscopies have provided evidence of 𝜺(0) suppression (Boamah et al. 2018) and long-range changes in water dipole correlations (Duboisset et al. 2018), precise determination is ambiguous and model dependent.
Roles of Hydrogen Bonds
The challenge for developing first-principles descriptions of aqueous fluids is accounting for the structural and energetic contributions of the hydrogen bonds formed between the polar water molecules. Although some aspects of bulk aqueous solutions may be reproduced using isotropic models that neglect hydrogen bonding (Islam 2019), it is generally accepted that confinement acts by disrupting the hydrogen bond network. For example, solid surfaces both with or without sites that can form bonds to interfacial water cause significant density and orientational effects, yet the consequences for water chemistry remains debated.
Progress in understanding confined aqueous systems will likely require the application of a suite of laboratory capabilities and molecular simulation approaches for characterizing hydrogen bond contributions to fluid properties. Specifically, we are using dielectric relaxation spectroscopy (DRS; Keysight Network Analyzer installed 2016) which measures the complex dielectric response, 𝜺(𝜔), as a function of frequency, 𝜔. , complemented by molecular modeling (Zarzycki & Gilbert 2019). DRS can reveal how the rotational mobility of water is altered around solutes (Mamatkulov et al. 2018) but there have been relatively few applications to surfaces and confinement. DRS can also provide an estimate of the static dielectric constant but, despite some pioneering early studies, the potential of this approach has not been yet fulfilled. In addition, water adsorption isotherms (Micromeritics 3Flex installed 2018) quantify the free energy of vapor adsorption on surfaces and in pores. Coupling this device to a microcalorimeter (Setaram installed January 2020; prior collaboration with Dr. Ricardo Castro, UC–Davis provides differential heats of adsorption and the full thermodynamic properties of the sorbed water.
We have shown that multiple layers of water adsorbed to fine-grained clay mineral (pyrophyllite) and a nanoporous sedimentary silica rock (diatomite) are altered by interactions with surface sites (Fig. 1a). These interactions strongly affect the dielectric properties of water with, for diatomite, a striking shift in the Debye relaxation to higher frequencies (Fig. 1b) and a remarkably low inferred value of the static dielectric constant (scaled to account for diatomite porosity) of 𝜺(0) ~ 12.5.
Goals
The long-term goal of this effort is to develop quantitative descriptions of the properties of water and solutes confined in thin films and nanoscale pores, extending to hydrothermal conditions. The relevant aqueous solutions include ideal dilute solutions at the surfaces of minerals with low structural charge all the way to the non-ideal concentrated electrolytes that are typical in subsurface clay-rich media.
The primary 3-year goal is to establish an approach and theoretical framework for this effort by demonstrating and modeling the relationship between confinement and the dielectric properties of aqueous solutions in granular minerals and porous rocks, finishing work on low-charged layer silicates (pyrophyllite) and developing the study of calcite. These experimental and modeling approaches will also be applied to the studies of swelling clay minerals, as described in those projects.
A secondary effort will be to develop a quantum sensing method, equivalent to acquiring nuclear magnetic resonance (NMR) data at solid-water interfaces with submicron lateral resolution and 10–20 nm depth resolution. 1H relaxometry and low-field spectroscopy will provide physical and chemical properties of water and solutes at silicate–water–vapor and silicate–water–calcite interfaces.
Proposed Work
Thermodynamic and Dielectric Properties of Thin-Film and Confined Water
The interaction of water with all the minerals addressed in this proposal will be quantified. For example, high crystallinity sub-100-nm rhombohedral crystals of calcite will be synthesized under controlled conditions in the chemostat reactor developed by Lammers et al. (ref) and fully characterized. We will establish the full thermodynamic properties of adsorbed water by acquiring temperature-dependent (5–90˚C) water adsorption isotherms and water adsorption calorimetry, greatly expanding the data of Figure 1 and seeking to reduce uncertainty beneath 1 kJ/mol to discover changes in the enthalpy of condensation relative to bulk water (Ben-Naim and Marcus 1984). Combined with surface-area information from N2 BET, the water sorption data provide the integral thermodynamic properties, ΔG=ΔH-TΔS, as a function of (effective) film thickness, determined by water activity.
In collaboration with Dr. Larry Anovitz will use U/SANS to quantify the pore size distribution in dry samples and to identify film formation and pore filling as a function of rh across the adsorption isotherm. At selected values of film thickness, we will determine the GHz dielectric properties of the system to extract the water contribution and investigate the accuracy with which the static dielectric constant can be estimated from 𝜺'(𝜔⟶∞GHz). This will require careful volumetric correction and, if substrate relaxations contribute, may require the application of effective medium theory (e.g., Monecke et al. 1994). Complementary molecular simulations will be used to reproduce all the conditions and experimental observables. We showed that the full DRS spectra can be predicted to good accuracy using molecular dynamics simulation without assuming any relaxation form provided a suitable water model is chosen (Zarzycki & Gilbert 2019). Grand Canonical Monte Carlo provides good agreement in the water adsorption isotherm for pyrophyllite (Colla, Zarzycki, Gilbert et al. 2020).
Dielectric Properties of Hydrogen–Bond Networks from Molecular Simulation and Machine Learning
The properties of aqueous solutions emerge from the extended dynamical structures formed by transient hydrogen bonds. We hypothesize that confining surfaces affect the hydrogen bond networks in ways that predictably alter ensemble properties such as dielectric response. It is a major challenge, however, to identify characteristic ensemble configurations even if molecular simulation accurately predicts continuum properties. For example, we recently used large-scale molecular modeling to demonstrate that temperature dependence of the Debye relaxation time correlates closely to an ensemble definition of the hydrogen bond lifetime (Fig. 2). Although this was the first such correlation, it was not feasible to identify any single cluster trajectory accounting for the Debye relaxation. The development of Neural-Network enhanced molecular simulation will
Solute Speciation in Confined Aqueous Fluids
In order to examine the effect of confinement on solute chemistry we will use a classic method developed to investigate the acidity of clay surfaces (Mortland 1968; Liu et al. 2003). We will introduce volatile small-molecule acids, ammonia, pKa 9.25, (Dontsova et al. 2005) and boric acid, pKa 9.24, (Gaillardet et al. 2000) into adsorbed thin films and use spectroscopic methods, NMR or FTIR, to elucidate the protonation state as a function of film thickness. For montmorillonite, reducing the amount of interlayer water raises the apparent equilibrium coefficient written as pKaobs=[OH-][NH4+]/[H2O][NH3]. Although the findings imply that interlayer water is a stronger proton donor than bulk water, trends are complicated by contributions from surface- and counterion-associated waters. Our initial studies on low-charge minerals will simplify the analysis, as illustrated by a relevant recent study. Using pH-sensitive EPR spin labels tethered to silica nanoparticles in solution, Perelygin et al. (2019) used EPR titration to independently estimate surface potential, 𝜑, and interfacial 𝜺(0). However, the low values of 𝜺(0) detected by the spin labels were at least partly influenced by the high atomic weight molecules themselves.
Initial analyses will treat the confined solute pKaobs as differing from bulk solution Ka0 due to the effect of altered 𝜺(0) and 𝜑 through the linear relationship pKaobs= pKa0 + pKa𝜺 + pKa𝜑, where the bulk term and the electrostatic term, pKa𝜑=- e𝜑/ln(10)kT, can be subtracted to give pKa𝜺. We will check for self-consistency for the solutes differing in charge (i.e., NH4+ and BOH0) and compare the results with DRS measurements 𝜺(0).
Prediction of absolute pKa’s is a challenge for ab initio molecular simulation that typically achieve an accuracy better than 1 unit only with the use of empirical scaling factors (Klicić et al. 2002). Recent studies have demonstrated that the pKa’s of compounds can be predicted accurately via a careful selection of DFT methods and appropriate number of explicit water molecules without fitting any linear equations or thermodynamic cycles (Thapa & Schlegel 2017). Moreover, shifts in pKa's are expected to be predicted with significantly higher accuracy that absolute values.
Properties of Interfacial Aqueous Films by Quantum Sensing
The Geologic Quantum Sensing method is anticipated to provide new approaches for studying the chemistry of nanoscale aqueous films. The principal goal of this effort will be to characterize the physical and chemical properties of water and solutes in thin aqueous films on oxide- or carbonate-terminated single-crystal diamond surfaces (Fig. 3) as a function of rh and hence film thickness. The group of Dr. Miguel Salmaron at LBNL has recently shown that plasma enhanced chemical vapor deposition at the Molecular Foundry can create ultrathin films of Al2O3, TiO2 and other oxides on graphene (Salmeron personal communication; Morales, Salmeron et al. submitted) while Dziadkowiec et al. (2018) used atomic layer deposition to create CaCO3, films on mica for SFA studies. We have a Foundry User Proposal to develop thin-film fabrication on single-crystal diamond prepared for quantum sensing.
We will apply NMR methods that have previously been implemented by diamond NV––based systems (Glenn et al. 2018), in particular the use of 1H relaxometry to quantify the effective viscosity of thin films (Staudacher et al. 2015).
A further effort will involve the measurement of the protonation state of oxyanions to test for changes in proton activity as a function of film thickness, composition and temperature. Current efforts to use conventional NMR on bulk geologic samples have been unsuccessful due to the presence of paramagnetic impurities present in all natural materials. Using approaches developed by Tokunaga in prior BES Geochemistry research (Tokunaga et al. 2017), we will create nanometric thin surface films in equilibrium with a bulk fluid reservoir.
We are collaborating with the UC-Berkeley group of Dr. Jeff Reimer on the interpretation and simulation of the low-field NMR data that can be acquired using the diamond NV––based method. In particular, the Spinach software can be used to distinguish contributions from chemical shifts (interactions between nuclear spins and the valence electron density) and J-couplings (nuclear-nuclear spin interactions) (Göggler et al. 2011; Biternas et al. 2014).
References
Ben‐Naim, A., & Marcus, Y. (1984). Solvation thermodynamics of nonionic solutes. The Journal of Chemical Physics, 81(4), 2016-2027.
Biternas A, Charnock G, & Kuprov I (2014) A standard format and a graphical user interface for spin system specification. Journal of Magnetic Resonance 240:124-131
Boamah, M. D., Ohno, P. E., Geiger, F. M., & Eisenthal, K. B. (2018). Relative permittivity in the electrical double layer from nonlinear optics. The Journal of Chemical Physics, 148(22), 222808.
Duboisset, J., & Brevet, P.-F. (2018). Salt-induced Long-to-Short Range Orientational Transition in Water. Physical Review Letters, 120(26), 263001.
Fumagalli, L., Esfandiar, A., Fabregas, R., Hu, S., Ares, P., Janardanan, A., . . . Geim, A. K. (2018). Anomalously low dielectric constant of confined water. science, 360(6395), 1339.
Gaillardet, J., Lemarchand, D., Göpel, C., & Manhès, G. (2001). Evaporation and Sublimation of Boric Acid: Application for Boron Purification from Organic Rich Solutions. Geostandards Newsletter, 25(1), 67-75.
Glenn, D. R., Bucher, D. B., Lee, J., Lukin, M. D., Park, H., & Walsworth, R. L. (2018). High-resolution magnetic resonance spectroscopy using a solid-state spin sensor. Nature, 555(7696), 351-354.
Glöggler S, Blümich B, & Appelt S (2011) NMR spectroscopy for chemical analysis at low magnetic fields. Modern NMR Methodology, (Springer), pp 1-22.
Islam, S. M. T., & Ashok, A. (2019). Empirical Characterization of Vertical Channel in Underwater Visible Light Communication. Paper presented at the Proceedings of the International Conference on Underwater Networks & Systems, Atlanta, GA, USA.
Klicić, J. J., Friesner, R. A., Liu, S.-Y., & Guida, W. C. (2002). Accurate prediction of acidity constants in aqueous solution via density functional theory and self-consistent reaction field methods. The Journal of Physical Chemistry A, 106(7), 1327-1335.
Knight, A. W., Tigges, A. B., & Ilgen, A. G. (2018). Adsorption of copper (II) on mesoporous silica: the effect of nano-scale confinement. Geochemical Transactions, 19(1), 13.
Lammers, L. N., & Mitnick, E. H. (2019). Magnesian calcite solid solution thermodynamics inferred from authigenic deep-sea carbonate. Geochimica et Cosmochimica Acta, 248, 343-355.
Liu, H., Liu, D., Yuan, P., Tan, D., Cai, J., He, H., . . . Song, Z. (2013). Studies on the solid acidity of heated and cation-exchanged montmorillonite using n-butylamine titration in non-aqueous system and diffuse reflectance Fourier transform infrared (DRIFT) spectroscopy. Physics and Chemistry of Minerals, 40, 479-489.
Mamatkulov, M., Yudanov, I. V., Bukhtiyarov, A. V., Prosvirin, I. P., Bukhtiyarov, V. I., & Neyman, K. M. (2019). Pd Segregation on the Surface of Bimetallic PdAu Nanoparticles Induced by Low Coverage of Adsorbed CO. The Journal of Physical Chemistry C, 123(13), 8037-8046.
Monecke, J. (1994). Bergman spectral representation of a simple expression for the dielectric response of a symmetric two-component composite. Journal of Physics: Condensed Matter, 6(4), 907-912.
Mortland, M. M., & Raman, K. V. (1968). Surface Acidity of Smectites in Relation to Hydration, Exchangeable Cation, and Structure. Clays and Clay Minerals, 16(5), 393-398.
Pan, Y., Birdsey, R. A., Phillips, O. L., & Jackson, R. B. (2013). The Structure, Distribution, and Biomass of the World's Forests. Annual Review of Ecology, Evolution, and Systematics, 44(1), 593-622.
Perelygin, D. N., Boichuk, I. P., & Gorlov, A. S. (2019). Rational design of low-tonnage technological complexes for fine and ultrafine grinding of materials. IOP Conference Series: Materials Science and Engineering, 560, 012118.
Schlaich, A., Knapp, E. W., & Netz, R. R. (2016). Water Dielectric Effects in Planar Confinement. Physical Review Letters, 117(4), 048001.
Shock, E. L., Oelkers, E. H., Johnson, J. W., Sverjensky, D. A., & Helgeson, H. C. (1992). Calculation of the thermodynamic properties of aqueous species at high pressures and temperatures. Effective electrostatic radii, dissociation constants and standard partial molal properties to 1000 °C and 5 kbar. Journal of the Chemical Society, Faraday Transactions, 88(6), 803-826.
Sposito, G., & Prost, R. (1982). Structure of water adsorbed on smectites. Chemical Reviews, 82(6), 553-573.
Staudacher, T., Raatz, N., Pezzagna, S., Meijer, J., Reinhard, F., Meriles, C. A., & Wrachtrup, J. (2015). Probing molecular dynamics at the nanoscale via an individual paramagnetic centre. Nature communications, 6(1), 8527.
Sverjensky, D. A., Harrison, B., & Azzolini, D. (2014). Water in the deep Earth: The dielectric constant and the solubilities of quartz and corundum to 60kb and 1200°C. Geochimica et Cosmochimica Acta, 129, 125-145.
Tokunaga, T. K., Finsterle, S., Kim, Y., Wan, J., Lanzirotti, A., & Newville, M. (2017). Ion Diffusion Within Water Films in Unsaturated Porous Media. Environmental Science & Technology, 51(8), 4338-4346.
Zarzycki, P., Colla, C. A., Gilbert, B., & Rosso, K. M. (2019). Lateral water structure connects metal oxide nanoparticle faces. Journal of Materials Research, 34(3), 456-464.
Zarzycki, P., & Gilbert, B. (2020). Temperature-dependence of the dielectric relaxation of water using non-polarizable water models. Physical Chemistry Chemical Physics, 22(3), 1011-1018.
Zhang, C., Gygi, F., & Galli, G. (2013). Strongly Anisotropic Dielectric Relaxation of Water at the Nanoscale. The Journal of Physical Chemistry Letters, 4(15), 2477-2481.