Barite precipitation in porous media: Impact of pore structure and surface charge on ionic diffusion

Highlights

• Barite precipitation was studied in chalk and kaolinite samples.

• Micro-CT imaging showed two distinct barite precipitation patterns in chalk and kaolinite.

• Barite precipitation led to stronger HTO diffusivity reduction in kaolinite than chalk.

• Barite precipitation generated restricted diffusion of 36Cl ions than HTO.

• Zeta potential measurements showed generation of surface charge properties in precipitated zones.

Abstract

Several scientific fields such as global carbon sequestration, deep geological radioactive waste disposal, and oil recovery/fracking encounter safety assessment issues originating from pore-scale processes such as mineral precipitation and dissolution. These processes occur in situations where the pore solution contains chemical complexity (such as pH, ionic strength, redox chemistry, etc.…) and the porous matrix contains physical complexity (such as pore size distribution, surface charge, surface roughness, etc.…). Thus, to comprehend the participation of each physicochemical phenomenon on governing mineral precipitation, it is essential to investigate the precipitation behavior of a given mineral in different confined volumes. In this study, a counter-diffusion approach was used to investigate barite precipitation in two porous materials: micritic chalk and compacted kaolinite. The two materials present similar water and anionic tracer diffusivities and total accessible porosities but distinct pore size distributions with pore throats of c.a. 660 nm in chalk versus c.a. 35 nm in kaolinite.

X-ray tomography results obtained on the two materials showed a distinct distribution of barite precipitates: a 500 μm-thick homogeneous layer in chalk versus spherical clusters spread in a thickness of 2 mm in kaolinite. Mass balance calculations showed that barite precipitation led to a porosity decrease in the chalk reacted zone from 45% to 12% and in the kaolinite reacted zone from 36% to 34.5%. In contrast, water tracer diffusion experiments showed that diffusivity decreased by a factor of 28 in chalk and by a factor of 1000 in kaolinite. Such a discrepancy was attributed to the difference in the pore size distribution that would lead to the distinct barite precipitation patterns, capable of altering in a very different manner the connectivity within the reacted zone of the two selected porous media.

Such local alterations in connectivity linked to pore volume reduction would also magnify surface charge effects on ionic transport, as indicated by chloride diffusion experiments and electrophoric tests using zeta potential measurements. Indeed, ³⁶Cl⁻ was strongly more hindered than water, when diffused in reacted materials, with a diffusivity decrease by a factor of 450 in chalk and a total restriction of ³⁶Cl⁻ in kaolinite. These experiments clearly provide an insight of how local pore structure properties combined with mineral reactivity could help in predicting the evolution of pore scale clogging and its impact on water and ionic diffusive transport.

Introduction

In France, Switzerland and Belgium, argillaceous rock formations are considered potential host rocks for radioactive waste disposal facilities (Gaucher et al., 2004; NAGRA, 2002). Similar indurated rocks are also envisaged as cap rocks to seal anthropogenic CO₂ reservoirs for global carbon sequestration (GCS) (Bachu, 2002; Berthe et al., 2011; Fleury and Brosse, 2018). These rocks are selected due to their low permeabilities, which makes diffusion the main mass transport process, and their high concentrations of clay minerals, whose negatively charged surfaces are capable of adsorbing radionuclides under cationic form (Andra, 2005). However, the degradation of radioactive waste in disposal facilities and similarly, the use of different trapping methods to capture CO₂ in reservoir rock may generate some physicochemical imbalances that can enhance mineral dissolution and/or precipitation processes. Since these processes change the mineral composition of rock matrix, they can potentially alter the rock containment properties (Dagnelie et al., 2017; De Windt et al., 2008) and lead to safety assessment issues.

Assessments of the long-term evolution of rock containment properties are generally performed using numerical chemistry transport codes. In these codes, the porosity change is estimated from mineralogical volume balance due to precipitation/dissolution. Feedback on diffusivity is then determined by the empirical Archie's relationship, which links the change in diffusivity to the porosity changes, i.e. $\frac{D_e}{D_o}={\varphi }^m$, where D_e is the effective diffusive coefficient in m² s^—1, D_o is the diffusion coefficient in bulk water in m² s^—1, ϕ is the total porosity of the porous sample, and m is a fitting factor also known as the cementation factor (Archie, 1942). However, very few lab-scale experiments are available to assess the validity of this relationship when used in numerical models that couple changes in transport process with mineral reactivity (i.e. chemical interactions and diffusive processes) via strong porosity changes (Chagneau et al., 2015a; Rajyaguru et al., 2019). Recently, Rajyaguru et al. (2019) showed that the distinct precipitation behavior of barite and gypsum in a chalk sample led to a very different impact on the diffusivity of a water tracer. The authors noted that this difference in precipitation behavior stems from the combined effect of mineral intrinsic properties (solubility, precipitation rate) and spatial variability in the properties of chalk (pore structure, reactive surface). Similarly, Singurindy and Berkowitz (2003) studied calcite dissolution by sulfuric acid and hydrochloric acid solutions and gypsum re-precipitation in a compacted sand matrix. This study showed that the flow rate and concentration of acid in the injected solution controlled the pattern of calcite dissolution channels and gypsum morphology. Other studies outlined that nucleation phenomena governed local mineral growth (Poonoosamy et al., 2016; Prieto, 2014), that pore pressure controlled crystal stability in confined media (Putnis and Mauthe, 2001; Andra, 2005), and that both processes collectively defined the possibility of precipitation in a given pore class (Emmanuel et al., 2010). One could infer from these studies that the nature of precipitation and its feedback on diffusion is potentially governed by the pore size distribution of the hosted material than by the porosity (volume of void) itself, contrary to the paradigm of the widely used Archie's law.

These studies also outline that local mineral distribution and morphology can be used to qualitatively decipher the average change in intrinsic properties of a reacted sample. Furthermore, formation of secondary minerals surfaces could also generate surface charge effects that were not visible in a porous material under intact form. In this case, ionic species diffusivity may differ from that of neutral species, such as a water tracer. Depending upon pH, electrolyte concentration, and the presence of potential determining ions, the sulfate alkali minerals may possess varying amounts of negative or positive surface charges (González-Caballero et al., 1988; Hang et al., 2009). For instance, Chagneau et al. (2015b) investigated the impact of celestite precipitation on transport properties through compacted illite. The authors showed that the reacted zone in illite allowed water tracer to diffuse, whereas ³⁶Cl^— diffusion was completely restricted. They proposed that the presence of additional negative surface charges on celestite, combined with surface charges on illite, resulted in total porosity clogging and restriction of ³⁶Cl^— diffusion through the reacted zone (Chagneau et al., 2015b). All of these experimental and theoretical observations indicate that reactive transport in porous media is a process that strongly couples chemical reactivity of minerals with local transport properties of the host pore structure.

The objective of the present work is twofold. The first goal is to experimentally investigate the role of pore size distribution on barite precipitation in two reference porous media having similar initial diffusive properties but distinct pore size distributions: chalk and compacted kaolinite. Barite was chosen as its formation is of concern in geothermal reservoirs and hydraulic fracturing (Mundhenk et al., 2013; Vengosh et al., 2014) and as it is capable of trapping radium in mill tailings of former uranium mines (Besançon et al., 2020; Chautard et al., 2020) and in nuclear waste disposal facilities (Brandt et al., 2018; Vinograd et al., 2018). The second objective is to quantify the impact of barite precipitates on diffusive transport properties. Neutral species (tritiated water) and ionic species (³⁶Cl^—) were used as reference tracers, since both the barite precipitates and the host pore surfaces were charged and therefore expected to generate a combined surface charge effect that could govern ionic species transport through the reacted samples. Results from these experiments may be used to improve safety assessments tests in sites of nuclear waste disposal and/or carbon sequestration, where robust numerical models are needed to predict the change in ionic transport properties (diffusivity) through host rocks in response to local clogging.