Available online at www.sciencedirect.com
ScienceDirect Geochimica et Cosmochimica Acta 223 (2018) 520–536 www.elsevier.com/locate/gca
Adsorption isotope effects of water on mesoporous silica and alumina with implications for the land-vegetation-atmosphere system Ying Lin a,b,⇑, Juske Horita a,⇑, Osamu Abe c b
a Department of Geosciences, Texas Tech University, Lubbock, TX 79409-1053, USA Institute of Deep-sea Science and Engineering, Chinese Academy of Sciences, Luhuitou Road No. 28, Sanya, Hainan 572000, China c Graduate School of Environmental Studies, Nagoya University, Nagoya 464-8601, Japan
Received 7 June 2017; accepted in revised form 22 December 2017; available online 29 December 2017
Abstract Soil water dynamics within a vadose (unsaturated) zone is a key component in the hydrologic cycle, especially in arid regions. In applying the Craig-Gordon evaporation model to obtain isotopic compositions of soil water and the evaporated vapor in land-surface models (LSMs), it has been assumed that the equilibrium isotope fractionation factors between soil water and water vapor, a(2H) and a(18O), are identical to those between liquid and vapor of bulk water. Isotope effects in water condensation arise from intermolecular hydrogen bonding in the condensed phase and the appearance of hindered rotation/translation. Hydrogen bonding between water molecules and pore surface hydroxyl groups influences adsorption isotope effects. To test whether equilibrium fractionation factors between soil water and water vapor are identical to those between liquid and vapor of bulk water and to evaluate the influence of pore size and chemical composition upon adsorption isotope effects, we extended our previous experiments of a mesoporous silica (15 nm) to two other mesoporous materials, a silica (6 nm) and an alumina (5.8 nm). Our results demonstrated that a(2H) and a(18O) between adsorbed water and water vapor are 1.057 and 1.0086 for silica (6 nm) and 1.041 and 1.0063 for alumina (5.8 nm), respectively, at saturation pressure (po), which are smaller than 1.075 and 1.0089, respectively, between liquid and vapor phases of free water at 30 °C and that the differences exaggerate at low water contents. However, the profiles of a values with relative pressures (p/po) for these three materials differ due to the differences in chemical compositions and pore sizes. Empirical formula relating a(2H) and a(18O) values to the proportions of filled pores (f) are developed for potential applications to natural soils. Our results from triple oxygen isotope analyses demonstrated that the isotope fractionation does not follow a canonical law. For the silica (15 nm), fractionation exponents (17h) are 0.5361 ± 0.0018 and 0.5389 ± 0.0016 at p/po = 0.72 and 0.77, respectively. For the silica (6 nm), 17h values are 0.5330 ± 0.0011 at p/po = 0.65 and 0.5278 ± 0.0010 at p/po = 0.81. For the alumina (5.8 nm), 17h value is 0.5316 ± 0.0015 at p/po = 0.78. These values are greater than or equal to that of liquid-vapor equilibrium of bulk water (0.529 ± 0.001). Ó 2018 Elsevier Ltd. All rights reserved. Keywords: Adsorption isotope effects; Liquid-vapor equilibrium; Craig-Gordon evaporation model; Isotope fractionation factor; Isotope fractionation exponent
⇑ Corresponding authors at: The EDGE Institute, University of California, Riverside, CA 92521, USA (Y. Lin); Department of Geosciences, Texas Tech University, Lubbock, TX 79409-1053, USA (J. Horita). E-mail addresses:
[email protected] (Y. Lin), Juske.horita@ttu. edu (J. Horita).
https://doi.org/10.1016/j.gca.2017.12.021 0016-7037/Ó 2018 Elsevier Ltd. All rights reserved.
1. INTRODUCTION The land surface system in arid regions of terrestrial environments is located at a very critical interface with land-vegetation-atmosphere continuum of the
Y. Lin et al. / Geochimica et Cosmochimica Acta 223 (2018) 520–536
energy-water balance and ecological systems. The partitioning of precipitation and soil water into fluxes of percolation to the subsurface, surface runoff, and evapotranspiration is accompanied by large changes in the stable isotope ratios (d2H and d18O), due to temporal and spatial variations in isotopic compositions of precipitation events and the isotope fractionation associated with the evaporation of water, both from open surface water and subsurface soil water (Gat, 2010). Soil water dynamics in a thick vadose (unsaturated) zone play a key role in the hydrologic cycle in arid regions. There is no fractionation of isotopes of water during water uptake by roots, due to the apoplastic free-diffuse pathway outside the cell membrane (e.g., Wershaw et al., 1966; Ehleringer and Dawson, 1992), except for some species that utilize the symplastic pathway across the cell membrane and through the endodermis (Lin and da Sternberg, 1993; Ellsworth and Williams, 2007). Thus, the isotopic compositions of soil water and water vapor, distinct from that of precipitation, are the key parameters for understanding the land-vegetation-atmosphere interface in arid regions (Allison et al., 1983; Gat, 2010) and have been incorporated into many land surface models (LSMs) (Fischer, 2006; Henderson-Sellers, 2006; Yoshimura et al., 2006; Risi et al., 2016). The Craig-Gordon model was initially developed to estimate the overall (equilibrium and kinetic) isotopic fractionation during the evaporation of water from open water bodies and has been successfully incorporated into global circulation models (GCMs) for simulating the dynamics of the global hydrologic cycle for present and past climates (Craig and Gordon, 1965; Horita et al., 2008). In the past decades or so, the Craig-Gordon model has been used in the isotopic studies of soil water, water transport in plants, and local-regional evapotranspiration (Horita et al., 2008). One of the important parameters in using the Craig-Gordon evaporation model to quantify the isotopic compositions of soil water and water vapor is the equilibrium isotope fractionation factor (a) between soil water and water vapor. a(2H) and a(18O) values for liquid-vapor equilibrium of bulk water were well determined experimentally from 0 to 350 °C, while a(17O) values were determined at 11.4–41.5 °C (Majoube, 1971; Horita and Wesolowski, 1994; Barkan and Luz, 2005). However, for the adsorption of water onto porous soils, the equilibrium fractionation factors could be different from those for bulk/free water liquid-vapor phase transition due to the interaction between water molecules and pore surfaces. Natural soils are complex and heterogeneous mixtures of minerals of different sizes, compositions, and textures, together with decaying organic matter and water. Few experimental studies have been conducted on the isotope effect of water adsorption on soil minerals and their analogs. Richard et al. (2007) found that equilibrium a(2H) values between adsorbed water on mesoporous silica (12–16 nm) and water vapor ranged from 1.030 to 1.055 when relative humidity increased from 10 to 87%, which are smaller than a(2H) of 1.085 between liquid and vapor of bulk water at 20 °C (Horita and Wesolowski, 1994). Stewart
521
(1972) reported that d2H values of water adsorbed on several clay minerals (kaolinite, illite, montmorillonite) are variably (0–100‰) lower than bulk, free water, but his data have considerable uncertainties. Oerter et al. (2014) reported oxygen isotope fractionation of water adsorbed to the interlayer cations (Na, K, Mg and Ca) of montmorillonite at different water contents. Although large changes (up to 3–3.5‰) in the d18O values of adsorbed water have been reported, their results are not conclusive because residual water likely present in the dried clay samples was not characterized and the degree of hydration of the clay samples was not determined. Zeolites, hydrated aluminosilicates of AlO4 and SiO4 tetrahedral framework filled with water molecules and exchangeable cations in channels (0.3–0.7 nm), can be considered as analogs of expandable clays. The channel water in zeolite minerals was found to be depleted in 18O by 0 to 3‰, 5‰, and 3‰ relative to the external liquid water near ambient temperature for wairakite (Ca8(Al16Si32O96)16H2O), analcime (NaAlSi2O6H2O), and stilbite (NaCa4(Si27Al9)O7228H2O), respectively (Karlsson and Clayton, 1990; Feng and Savin, 1993; Noto and Kusakabe, 1997; Karlsson, 2001). The adsorption of water on several types of organic matter (silage, litter, cotton and others) also showed smaller magnitudes of oxygen and hydrogen isotope fractionations, compared to liquid-vapor equilibrium of bulk water (Chen et al., 2016). The above studies reveal that our current knowledge on the isotope fractionation of water adsorbed on various soil materials is very limited and inconclusive. The objective of this study is to test the hypothesis that equilibrium isotope fractionation between adsorbed/porecondensed water within soils and water vapor differs significantly from that between liquid and vapor of bulk water (free water; non-confined water) at the same temperature, due to complex hydrophilic interactions between soil pore surface and water molecules. We use mesoporous silica and alumina materials as soil analogs because significant volumes of intraparticles with mesopores (2–50 nm) exist in soils (Hajnos et al., 2006; Chen et al., 2007), which becomes important when large interparticle pores are quickly drained and because amorphous to poorly crystalline oxides (Al/Fe-oxides) and silicates are important constituents of a clay fraction (
