Swelling and Fluid Transport of Re-sealed Callovo–Oxfordian Claystone

Abstract

Several experimental methods are used to investigate the swelling capacity and transport properties of re-sealed macro-cracked Callovo–Oxfordian (COx) claystone, particularly in the absence of confining pressure. Six COx claystone samples from four different geological locations of the Bure basin (France) are tested, five of which are macro-cracked and one remains intact. Sample swelling occurs, during re-hydration with liquid water, leading to the measurement of an apparent swelling pressure. The latter is continuously recorded with a dedicated device. The asymptotic apparent swelling pressure of macro-cracked UT (transitional unit) COx is approximately 1 MPa, while it varies from 3 to 5 MPa for macro-cracked UA (clayey unit) COx. Quantitative X-ray diffraction (QXRD) analysis demonstrates that the amount of smectite, which is a swelling clay, is weakly correlated with apparent swelling pressure. Surprisingly, the interstratified illite/smectite with lower smectite content is highly correlated to apparent swelling pressure. Nitrogen isotherms data imply that the Gurvich total pore volume (V_Gurvich) and specific surface area (SSA) are highly linearly correlated with the low smectite content interstratified phase. This means that the distribution of smectite strongly affects the swelling capacity of COx. Moreover, nitrogen sorption is an easier and more effective technique than QXRD for assessing COx swelling capacity, since both V_Gurvich and SSA have been proven as effective indicators. For both UT and UA COx, self-sealing can cause significant reductions in water permeability (Kw). In particular, UA COx shows higher sealing efficiency and faster kinetics compared to UT COx. After sealing, the equivalent crack aperture (calculated from Poiseuille’s law) decreases from tens of microns to less than 1 micron. According to the gas breakthrough tests, the gas breakthrough pressure (GBP) of re-sealed macro-cracked COx is of the same order of magnitude as the equivalent capillary pressure of residual crack. This indicates that the gas migration in the re-sealed cracked COx claystone mainly occurs through the residual crack and is ‘a priori’ controlled by capillary processes.

Highlights

• A dedicated device was designed to study the transport and swelling properties of a sealed COx claystone under low confinement.

• The apparent swelling pressure and water permeability were continuously recorded during the sealing test due to water injection in fractures.

• Water permeability of samples decreased by orders of magnitude to almost meet the intact material permeability.

• Gas breakthrough experiments performed on sealed samples evidenced gas flow through the residual cracks.

Appendix

1.1 Quantitative X-ray Diffraction (QXRD)

To determine the bulk mineralogical composition, first, the representative sample of claystone are crushed manually into powders, combined with internal standard (0.2 g/g Al₂O₃ serving as the accuracy control). The powders are subsequently micronized, milled in ethanol and spray dried. After that, the powders are loaded in a sample holder by side loading and measured by X-ray diffraction (in this study, a CuKα radiation was used). The subsequent weight quantification of each mineral can be performed based on the Rietveld method (Zhou et al. 2018). For the detailed 2:1 clay minerals analysis, a chemical treatment was first used to remove cementing agents of powders, the powder size is selectively separated to below 2 µm by means of centrifugation. Each type of powders will be saturated with Ca²⁺. Thus, all expandable clay mineral cations are exchanged to Ca²⁺ and this greatly facilitates identification and quantification of the different clay mineral species. The oriented preparations are made by sedimentation on a porous ceramic plate. The preparations yield highly oriented clay particles. They are subsequently analyzed by X-ray diffraction for the measurement of the first reflections (001) of the clay minerals.

Generally, the accuracy of non-clay mineral composition can be controlled within 1%, while it is difficult to assess the accuracy of clay mineral composition due to the lack of standard materials. The bias of XRD results is mainly associated with the mistaken of quantifying illite as smectite (Seemann et al. 2017).

1.2 Cation Exchange Capacity

The representative part of samples (around 5 g) is retrieved from different cores, and first dried at 110 ◦C and then weigh noted. All samples were subsequently stirred and dispersed in ultrapure water with Co (III)-hexamine, following the method described in (Meier and Kahr 1999). The exchange progress was accomplished by means of placing the sample and solution in an overhead shaker until the exchange reaction was completed. All samples were then homogenized after ultrasonic treatment and the absorption of the supernatant liquid was measured using a spectrophotometer. The cation exchange capacity of each sample was derived by the difference of concentration between the solution and supernatant liquid. For each sample, the procedure was repeated several times to increase the accuracy of the results. The detailed methodology was elaborated interpreted in the literature, for instance, (Meier and Kahr 1999). The value of CEC is quantified in milligram equivalents (meq) of exchangeable cations per 100 g of the dry sample or in millimoles of exchangeable cations per grams (mmol/g) of dry sample (Song et al. 2017).

1.3 Nitrogen Physisorption

For each measurement, the representative COx sample (2–4 g) is first manually crushed and sieved to obtain fragments of 63–400 µm size. All the treated sample fragments are degassed for 12 h, then heated in a vacuum environment at 130 ℃ for at least 12 h. The adsorption/desorption measurement is carried out on a Micromeritics ASAP2020 machine. Using nitrogen at a constant temperature of 77 k, 31 relative pressure points during adsorption and 24 relative pressure points during desorption are recorded, within a relative pressure interval of 0.001–0.995 p/p⁰. The equilibration is regarded as the pressure change over 10 s is less than 0.01% of the average pressure during the latter interval. And the saturation pressures (i.e., p⁰) are re-calibrated independently for each step.