(159a) Thermodynamic Modeling of Phase Behavior in Shale Media

Phase behavior in shale remains a mystery due to various complexities and effects. Methane in shale media may have a Langmuir type adsorption isotherm, but CO₂ adsorption may continuously increases with pressure. In conventional permeable media, once pore volume is known, the amount of fluid-in-place can be estimated accurately. This is because the fluid is locally homogeneous and pores are generally more than 100 nm, and surface adsorption will be negligible. In shale media, in addition to knowledge of pore volume, knowledge of pore size distribution, total organic content, and chemistry of the rock is also required. Currently, there is no viable model in the literature that can describe fluid phase behavior in shale rocks.

Fluid molecules in shale media can be found in three different states: 1-free molecules in the pores, 2-adsorbed molecules onto the pore surface, and 3-dissolved molecules in the organic matter. Of the three, the first two mechanisms are discussed in the literatures. In this work, we compute for the first time the amount of dissolved molecules. We present a comprehensive model for phase behavior in shale media that allows the computation of fluid-in-place and knowledge of where the fluid molecules reside.  

In order to compute the fluids in shale media, we divide the pores into sizes greater than 10 nm, and sizes less than 10 nm. In pores greater than 10 nm, the interface curvature affects phase behavior and fluid phases are homogeneous. Therefore, they can be described by conventional equations of state. Our calculations show that retrograde condensation increases in nanopores; the upper dewpoint increases, and the lower dewpoint decreases. These calculations are supported by experimental measurements.

In pores less than 10 nm, the fluids become inhomogeneous and the direct use of conventional equations of state cannot be applied even with adjusted critical pressure and temperature. We suggest the use of molecular modeling, such as density functional theory and Monte Carlo simulations. The molecular modeling will allow determination of adsorption isotherm for both pressure increase and decrease (hysteresis). We will also show that the use of Langmuir adsorption based on many assumptions in derivations cannot provide correct results.

Fluid dissolution may provide additional mechanism to fluid-in-place. Our modeling shows that dissolution in kerogen can be up to 20% of total fluid-in-place.

We use available data in shale media, which are mainly limited to excess adsorption of methane and carbon dioxide to compare to our thermodynamic model computations. Fluid adsorption in nanopores of shale is determined by the pore size distributions, but the contribution from fluid dissolution in kerogen is mainly dependent on the organic content. Based on our findings, we suggest two basic measurements in shale media: 1) pore size distribution, and 2) proportion of organic matter.