| Abstract |
- Extraction of natural gas from shale formations using the process of hydraulic fracturing (fracking) requires the use of thousands of cubic meters of fluid. Hydraulic fracturing fluids are pumped under pressure into shale formations, fracturing the shale and releasing pockets of trapped gases. When the pressure in a natural gas well is released, the previously trapped gases flow to the surface and are collected. As much as 80% of the fracking fluids return to the surface as well. These fluids, known as flowbback fluids or produced water, contain high concentrations of shale minerals including heavy metals and radionuclides. They also serve as a medium for microbial communities which produce hydrogen sulfide and other corrosive compounds. The storage, transportation and treatment of flowbback fluids increase the cost of natural gas production, and present an environmental health risk to local surface water systems. Further study of these fluids is required to constrain the process by which shale constituents are mobilized during hydraulic fracturing.
A fundamental component of any attempt to characterize the chemistry and microbiology of hydraulic fracturing fluids is effective sample collection. Even in ideal conditions, sampling these fluids consistently and with high frequency can be logistically difficult. Often periodic sampling, or spot sampling, is infrequent, occurring on the order of days to weeks. Spot samples offer a glimpse of fluid chemistry and microbiology at specific timed points, and can indicate that changes are occurring, but deliver little insight into the timescale or mechanisms of those changes. High resolution sampling, on the order of hours, has revealed that large changes in aquatic chemistry can occur on short timescales. In river systems, spot sampling has been shown to miss large volumes of nutrient influx as a result of heavy precipitation resulting in an underestimation of the concentration of these nutrients in the system. Once in the river, these nutrients are consumed and modified by microorganisms. Similarly, production rates of hydraulic fracturing fluids from natural gas wells are not constant and current spot sampling regimes may be missing changes in the fluids that are occurring on short timescales.
The purpose of this project was to build and test continuous, remote sampling systems for use in characterizing chemical and microbial changes in hydraulic fracturing fluids. To accomplish this, I adapted a design known as the osmosampler that has been used successfully in deep ocean research. Osmosamplers function entirely by osmotic pressure generated by separating a chamber of concentrated salt water from a chamber of deionized water. This pressure is used to collect a sample of fluid in a long coil of small diameter Teflon tubing. Upon retrieval of the sampler, the tubing is divided into sections that contain fluids sampled at different time points throughout the deployment. Analysis of these sections provides a highb resolution dataset regarding changes in fluid conditions over time. For this project, I constructed three samplers, one based on those used in deep ocean applications using rigid cartridge membranes, and two of my own design using thin film forward osmosis membranes. For 58 days, the samplers collected fluid out of a flask full of D.I. water that was periodically spiked with NaCl and fluorescent microspheres. After each addition, reference samples were collected from the flask using a pipette, and stored. The conductivity data, indicating the concentration of total dissolved solids collected from two of the samplers, closely matched the reference samples though on different timescales. The samplers using the thin film membranes pumped very quickly, one reaching capacity after only a few days and the other after a few weeks. Data retrieved from these samplers was not complete for the 58 days, but the subset of data they did collect was accurate when compared to the spot sample data. The sampler using the cartridge membranes pumped very slowly and did not reach capacity within the 58 day experiment. Data from this sampler did not accurately reflect changes in the reservoir fluid conditions due to an unpredictably variable pumping rate throughout the trial. This study will help to inform design choices for future osmosamplers to be deployed in the field. These samplers show great potential for use in characterizing the chemistry and microbiology of hydraulic fracturing fluids, as well as other aqueous environments. Further testing is required to ensure that the thin film samplers can withstand the harsh environment present in hydraulic fracturing fluids.
- Keywords: osmosamplers, shale formations, Hydraulic fracturing (fracking)
|
|---|