Estimating The CO2 Sequestration Capacity of Fractured Shale Formations Using Methane Production Rates: The Case of the Utica Shale

Abstract:

Fractured shale formations that have been drained of hydrocarbons could serve as attractive sites for geologic sequestration of CO₂. Shales preferentially sorb CO₂ enabling greater storage potential than would be expected based only on the pores vacated during CH₄ production. Sequestration in shales could have a variety of other benefits because the intrinsically low permeability of the rock could help mitigate leakage risks and infrastructure resources could be leveraged to minimize costs. Here a modeling framework developed by the authors to estimate the sequestration capacity of fractured shale formations based on CH₄ production rates was applied to the Utica Shale. The model is based on a unipore transport model, which assumes that diffusion of gases into and out of the kerogen matrix will control gas transport. The results from the Utica formation were compared to estimates for sequestration in the Marcellus shale to understand how the petrophysical characteristics of these two formations impact estimated sequestration capacity. A detailed sensitivity analysis was carried out to link modeling assumptions and key parameters with known physicochemical characteristics of these two shale formations. Modeling parameters were derived from published production data obtained from the state of Ohio.

The model was found to be most sensitive to the equilibrium sorption parameters of CH₄ and CO₂, for which there is good literature data available. Published values for CO₂ sorption varied considerably based on the composition of the shale. Improved experimental data is needed to provide the most accurate estimates of storage in different formations. Differences were observed in gas diffusivity estimates for the Marcellus and Utica shale that could be understood in terms of the petrophysical characteristics of the two formations. We also found important effects tied to the effective diffusion length out of an average pore in the kerogen. These results allow us to understand how this modeling framework, which captures the underlying physical transport processes, is computationally efficient and it can be run using readily available data sources, will compare to more complex models of sequestration in shales.