Background

Despite the enormous discrepancies between the size and duration of laboratory and field processes, column studies can encompass sufficient compositional, morphological and structural heterogeneity to approximate the continuum behavior of porous rocks. From prior column-scale compaction experiments on carbonates and other geologic materials, key findings include the limited role of mechanical compaction for porosity reduction (Shinn & Robbin 1983) and the essential role of aqueous solution chemistry (Rutter et al. 1972; Schmoker & Halley 1982; Zhang & Spiers 2005; Le Guen et al. 2007). More recent studies have used variations in calcite origin, composition and grain size, changes in temperature and effective pressure, and chemical species retarding calcite growth to infer the dominant mechanism and rate-limiting step (e.g., Zhang et al. 2010; Croize et al. 2010). However, as reviewed by Croize et al. (2013), additional long-term studies with additional in situ analyses are required to clearly establish the roles and interactions between pressure dissolution and subcritical crack growth, and to identify the controls on calcite reprecipitation.

Here we describe multi-modal laboratory studies that, combined with modeling, provide an enhanced ability to distinguish compaction mechanisms at the column scale. In particular, we propose to quantify the strongest constraint on the rate of chemically-mediated pressure dissolution: Gross dissolution and precipitation fluxes that facilitate chemical compaction. Although recent flow-through compaction studies have included the collection of some limited pore fluid geochemical data (Le Guen et al. 2007; Zhang et al. 2011), creep can occur close to chemical equilibrium, so the net rates of fluid–mineral elemental exchange approach zero. To address this challenge we will track the isotopic composition of fluids to provide a sensitive measure of gross mass exchange fluxes attending fluid–mineral reactions during laboratory compaction creep. In addition, time-lapse X-ray tomography, using instrumentation and analysis methods developed at LBNL, can provide quantitation of strain distributions in 3D and offers the possibility of capturing grain-resolved deformation, while the integration of acoustic emission (AE) sensors can correlate the evolution of physical properties with strain and chemical conditions. Acoustic measurements of the elastic moduli of a compacted granular material will be used to infer changes in bulk mechanical properties (Fabricius 2003).

The ultimate aim of most studies of compaction creep is the development of a constitutive creep equation to predict strain rates (cf. Croize et al. 2013). Reactive mechanics and transport modeling is necessary to fully interpret both steady-state creep and transient responses to altered compaction conditions. For example, transient drops in strain rate following a stress downstep were interpreted to indicate that calcite precipitation was a rate-limiting step (Zhang et al. 2010), a hypothesis that is testable using models for calcite growth developed in this group.

Proposed Work

In this project, we propose to perform long-term (months duration) studies of carbonate compaction in a custom triaxial cell under flow-through mode that enables chemical, isotopic, and mechanical analysis for the identification of microscale creep pathways. Initial studies will seek to balance interface exchange fluxes and pressure-dissolution creep. Subsequent studies will evolve through dialogue with work in other projects. As the pressure dissolution studies constrain molecular and asperity processes controlling interfacial phenomena, the long-term compaction experiment will provide an ongoing platform for establishing how influences on these processes affect transient and steady state macroscopic creep and properties. The experimental system and model will allow multiparameter testing of the molecular-scale processes governing interface and crack processes and their coupling with transport phenomena to generate the macroscopic strain and the mechanical properties of the column.

Experimental Design

Compaction will be performed in an X-ray transparent triaxial cell (Fig. 1) similar in design to recent studies (e.g., Neveux et al. 2014) capable of confining pressure (currently up to 24 MPa but higher capability in development), axial load and pore fluid pressure, temperature exceeding 100˚C and flow rate. The studies will use two calcite rocks (marble or limestone) that differ in their ⁸⁷Sr/⁸⁶Sr ratios. Each will be crushed, cleaned of fines, and size sorted. The material with lowest [Sr] will be equilibrated with an aqueous solution at the appropriate temperature to provide a saturated flow-through solution during the compaction of the highest [Sr] material. A dry compaction step followed by unloading will proceed the saturation of the column, flow through of several pore volumes and the application of pressure and experiment initiation.

Experimental Variables

Following the establishment of steady-state compaction conditions, step changes in experimental conditions will be performed to study the following variables.

1. Flow rate: Although in principle changes in flow rate will not distinguish transport or reaction limits on pressure dissolution when the inlet fluid is perfectly saturated relative to the mineral (de Meer & Spiers 1997). Zhang et al. (2010) inferred a transition between diffusion- and precipitation-controlled pressure dissolution as a function of flow rate. Neveux et al. (2014) found the strain rate to be depending on a coupling between flow rate and stress when the influent was not pre-saturated. Flow rate dependences will be measured and interpreted with coupled reactive mechanics and transport modeling.

2. Fluid chemistry: The effect of solution species beyond Mg²⁺, PO₄^2-, NH₄Cl, will be studied through transient additions to influent solution. For example, analysis of marine carbonate cores with terrestrial inputs of organic matter and clays suggest that non-complexing organics and dissolved silica reduce carbonate–fluid exchange (Turchyn & DePaolo 2011).

3. Temperature: Carbonate compaction studies typically exhibit a small and not understood dependence on temperature. Complementary interface studies will inform these experiments.

4. Effective axial stress: Axial stress will be stepped up and down in long-term studies to record transients in creep and solution chemistry and monitor the steady-state creep rate as a function of experimental parameters and porosity.

In nature, contrasts in Sr isotopic compositions of different Earth materials reflect differences in Rb/Sr ratios and their ages as the radiogenic isotope ⁸⁷Sr is produced through the beta decay of ⁸⁷Rb (half-life of 48 Ga). The ratio of ⁸⁷Sr to the stable isotope ⁸⁶Sr in seawater has varied over geological time as different rock types, with different ages and Rb/Sr ratios, were exposed to chemical weathering (Faure 1986). The ⁸⁷Sr/⁸⁶Sr ratio in limestone rocks formed from marine carbonate biominerals thus reflects the composition of ocean water during formation and burial diagenesis and so with time have varied in ⁸⁷Sr/⁸⁶Sr ratio. The ⁸⁷Sr/⁸⁶Sr ratio of marble may be further altered by exchange with high-temperature fluids although a review of Sr isotopes in marble from the Mediterranean found the ratio to fall in narrow range, 0.7071 to 0.7092 suggesting metamorphism played a small role (Brilli et al. 2005).

Because the radiogenic ⁸⁷Sr/⁸⁶Sr ratio (corrected for mass fractionation) is not observed to be altered by calcite growth or dissolution, this isotopic parameter can serve as a quantitative tracer of fluid–mineral exchange even if the fluid is saturated relative to calcite. As an indication of the sensitivity of the technique, a simple mass balance calculation shows that with an initial calcite-fluid difference of 0.002 in ⁸⁷Sr/⁸⁶Sr, we can detect at 5 times the measurement precision a change in ⁸⁷Sr/⁸⁶Sr corresponding to a strain of ~5x10^-7. Along with the measurement of ⁸⁷Sr/⁸⁶Sr, we will measure the Sr and Ca concentrations of the effluent fluid using an isotope dilution method which will allow us to resolve changes in their concentrations at a level of better than 0.5% to provide constraints on calcite reprecipitation.

We have obtained a marble with [Sr] ~1060 ppm and ⁸⁷Sr/⁸⁶Sr ~0.7092 and we are collecting and measuring calcite spar and limestone samples to identify a second material with a contrasting ⁸⁷Sr/⁸⁶Sr ratio to use to produce the aqueous solution for the column experiments. The calcite growth model of DePaolo (2011) and Nielsen et al. (2013) that relates Sr/Ca partitioning into growing calcite to growth rate will be used in compaction modeling to constrain growth.

In situ Imaging of Grain Evolution During Compaction Creep

Synchrotron X-ray micro- and nanotomography will be performed to investigate the in situ evolution of grain and column structure. Axial deformation will be measured during experiments to calculate the axial strain and compare bulk deformation with local deformation as measured from time-lapse tomographic images. Porosity reduction and alteration to the connectivity of pore networks will be evaluated as compaction proceeds. Maps of strain heterogeneity in the grain pack will be used to identify areas of enhanced or reduced compaction compared to the bulk. Individual grains can be segmented out, and changes in shape and orientation will be quantified to separate the effects of particle re-arrangement, dissolution and cracking.

Laboratory Compaction Experiments

Laboratory benchtop compaction experiments will be run in parallel with in situ X-ray CT experiments. These experiments will be performed at the same conditions (axial pressure, temperature, flow rate, fluid chemistry) as the beamline experiments, but at a larger scale. The benefit of larger experiments is twofold. First, the larger size of the column will allow for significantly more reactive surface area for interaction between the granular calcite and the fluid. This will result in larger chemical signals and allow for more precise analyses with which to compare data from smaller experiments. Second, a larger experiment can be fitted with additional acoustic sensors, allowing for more complete monitoring of changes in acoustic properties as the column compacts.

Acoustic Measurements of Mechanical Evolution during Compaction

Acoustic data from compaction experiments can be inverted for the mechanical properties of the grain pack using rock physics models. As the column compacts, contacts between grains become more numerous and flat as a result of grain cracking, rearrangement and pressure dissolution of impinging faces. This results in changes in the acoustic impedance of the grain pack and manifests as changes in compressional and shear wave velocities. Using models (e.g. Digby 1981), the measured wave velocities can be inverted for effective grain contact parameters. High-resolution structural data from imaging experiments can be correlated with these parameters to provide insight into the fundamental mechanisms responsible for the changes in acoustic properties and make explicit links between chemically mediated deformation and physical properties.

Preliminary Experiment Results

A preliminary short-term compaction experiment was performed to assess the feasibility of the simultaneous measurement of strain and geochemical signatures of compaction (Fig. 3). The study clearly observed a correlation between strain and solution chemistry (Fig. 3c) but a simple recrystallization model indicated, unsurprisingly, that the strain was due to mechanical processes rather than dissolution–reprecipitation. X-ray tomography quantified macroscopic strain in excellent agreement with axial strain gauge, but voxel resolution was too coarse to track single-grain displacements. A modified cell is under construction with a smaller diameter (allowing for better resolution), temperature control up to 250˚C, and the addition of acoustic sensors.

1D Reactive Transport Modeling of Compaction Creep with Data Integration

Experimental results will be simulated using a 1D reactive transport approach in CrunchFlow (Steefel et al. 2015) that accounts for mineral dissolution and precipitation, diffusion and advection (Fig. 3d). After the model is constrained using these laboratory studies, it will form the basis for the geologic model for burial diagenesis.

The objectives of the reactive transport modeling are:

(1) quantifying the chemical processes as constrained by the bulk fluid chemistry, while providing additional insights on local dissolution versus precipitation and the impacts of flow and diffusion along the column

(2) disentangling the chemical and mechanical contributions to the measured strain rate at the column scale.

The model will focus on the relative amounts of carbonate dissolution and precipitation generating the observed isotopic and trace element signatures. The net mineral recrystallization contributes to the compaction of the column, and can be compared with the measured strain rate, which reflects the compound effects of chemical and mechanical change. Experimental factors such as the evolution of grain contact area and total surface area will be integrated with single-interface dissolution kinetics data to model the chemical and isotopic evolution of the pore fluid.

Further modeling work will investigate the control of flow rate and spatial heterogeneity on column compaction. The continuum scale model can be used to integrate the insights gained from the pore-discontinuum model that explicitly resolves pressure solution at grain-to-grain contact points, serving as the bridge between micro-scale conceptualizations and column-scale observations.

References