The goal is to develop a sustainable approach for water management in the Marcellus Shale play, in which flowback water is economically treated on site and reused for hydraulic fracturing adjacent wells. Optimal treatment processes will be identified, and acid mine drainage (AMD) water will be examined as a potential supplement to flowback water used for hydraulic fracturing. Researchers will also investigate the effects of barium sulfate (BaSO₄) precipitation on well surfaces and fracture spaces.

University of Pittsburgh, Pittsburgh, PA 15260
Carnegie Mellon University, Pittsburgh, PA 15213

Hydraulic fracturing has enabled the economical recovery of natural gas from the Marcellus shale. The fracturing fluids used in development of the Marcellus consist mostly of freshwater withdrawn from local streams and amended with chemical additives. A single well hydrofracture in the Marcellus may require one to five million gallons of fracturing fluid, of which between 25 and 100 percent may be returned to the surface as "flowback" or "produced" water, which must then be disposed. In addition to chemical additives, flowback water from Marcellus hydrofracturing typically contains high levels of total dissolved solids (TDS) (ranging from 70,000 to 250,000 mg/L) hydrocarbons, and heavy metals. The presence of these constituents precludes untreated re-use, reinjection, or direct discharge onto land or into receiving streams, as they may adversely impact human health and environmental quality. The flowback water is not amenable to reinjection due to high concentrations of barium and strontium and the potential for calcite precipitation in the injection well.

Conventional treatment processes, such as reverse osmosis and distillation, are not likely to be utilized due to their high capital costs and energy requirements. Disposal by dilution into Publically Owned Treatment Works (POTW), the common method to date for handling Marcellus flowback water in Pennsylvania, is not sustainable either, as transportation costs are extremely high, and POTWs are limited as to how much water they can accept and treat. In response to high TDS levels measured in the Monongahela River in the fall of 2008, the Pennsylvania Department of Environmental Protection ordered a restriction on the amount of produced water (including flowback) disposal to POTWs in the basin. This restriction effectively halted gas drilling operations in some locations in western Pennsylvania and has limited disposal options for expanding shale gas production.

Preserving the favorable economics of gas development in the Marcellus play and maintaining responsible stewardship of the environment are the two drivers of future water management strategies in the Marcellus. This research provides a holistic approach to water management through re-use of flowback water and use of supplemental AMD water for subsequent hydrofracturing operations. The location, quantity, and quality of flowback and AMD waters will be evaluated and their chemical interactions will be examined in order to identify optimal treatment processes. A field demonstration of the treatment process will be conducted and key technical and cost parameters for field demonstration of a hydrofracturing operation using the mixture of flowback and AMD waters will be assessed.

The Marcellus shale represents one of the largest reserves of on-shore natural gas in the country. However, development of this resource will be limited by the high volumes of difficult-to-treat wastewater produced during well development. By providing technically and economically feasible approaches for the re-use of flowback water, this project will effectively reduce the amount of freshwater needed for Marcellus Shale development and minimize the disposal liability and costs associated with new well drilling. The use of locally available AMD as make-up water will also reduce the amount of freshwater use as well as the transportation costs associated with bringing make-up water to the project site.

Field test results showed that by adjusting the mixing ratio of flowback and AMD in the treatment system, operators can achieve desired final sulfate concentration that is accurately predicted by thermodynamic models. Average pH of the effluent was 7.4, turbidity and less than 3 NTU, iron was less than 0.1 mg/L, and sulfate was 75 mg/L. Control of supersaturation ratio and sludge recycle is required to produce barite suitable for drilling mud. Barite attachment to casing, proppant pack, or shale can be controlled with antiscalants.

Dr. Vidic identified a field site in Tioga County to demonstrate the field treatment process involving mixing AMD and flowback water. Testing was conducted at five gpm over three weeks, five days a week, eight hours a day.. 

The design of the treatment system was finalized. Flowback water and AMD was pumped from the frac and AMD tanks, respectively. The flow rate was monitored and adjusted to provide a desired mixing ratio of flowback water to AMD. The pH was adjusted with sodium hydroxide or hydrochloric acid to achieve optimum conditions as demonstrated in the bench-scale experiment. AMD was added to the recycled sludge stream in the mixing tank and fed to the mixing reactor. Suspended solids formed in the mixing reactor were removed by coagulation/flocculation and sedimentation processes. The system was operated in Tioga County, PA.

This project was a finalist in the 2014 Shale Gas Innovation Contest (http://www.sgicc.org/shale-gas-innovation-contest.html).

The design of the clarifier and thickener for pilot-scale testing in the field was developed based on bench-scale experiments that involved mixing AMD and flowback water. The blend ratio of flowback water and AMD is dependent upon the barium sulfate mass ratio in the mixture and the desired final sulfate concentration the finished water. The conventional coagulation/flocculation process is optimized based on the coagulant dosage, pH, and mixing conditions. Microsand and flocculant aids are also added to the process. Mixtures with higher TDS levels require higher dosages of coagulant and flocculant. A clarifier will be used to separate solids with a fraction of the solids recycling back to the reactor to provide barite crystal seed and accelerate barium sulfate precipitation. The remaining precipitate will be dewatered for disposal, and clarified effluent will be stored in surface impoundments or storage tanks for subsequent hydrofracturing operations. Discussions with exploration and production companies are underway to select the location for pilot-scale testing.

Dr. Vidic presented project results at the Petroleum Technology Transfer Council Innovative Water Management Workshop on August 22, 2013, in Morgantown, WV.

Initial experiments to test the affinity of barium sulfate for the sand proppant revealed that small barium sulfate particles are captured by the sand in the front section of the column. This results in an increase in pressure and a decrease in permeability. At higher flow rates, the effects are greater. Experiments at a lower velocity resulted in less loss to permeability. The experiment will be repeated with the proppant column in a vertical orientation.

Membrane filtration for separating precipitates in AMD and flowback water mixtures was evaluated. The severity of membrane fouling decreased with an increase in the age of the flowback water. (Note: The “age” of flowback water refers to its change in composition over time as it’s recovered during the course of drilling operations.) Lower levels of TDS in flowback water enable the presence of submicron particles whose stability is due to organic coating on the surface of colloidal particles. Produced water is not likely to cause severe membrane fouling due to relatively high TDS that facilitate aggregation of colloidal particles and minimize the presence of submicron particles that are typically associated with membrane fouling.

In the lab, conventional coagulation and flocculation processes were optimized using two Marcellus flowback water samples and locally available acid mine drainage. Final sulfate concentration levels are dependent on the barium content in the flowback water as well as the mixing ratio of flowback water to AMD. Effluents from conventional and ballasted flocculation lab tests are of comparable quality. Contact time between the flowback water and AMD is only 10 minutes for ballasted flocculation versus 60 minutes for conventional treatment. Lab test also showed that there will be no need for pre-mixing in the ballasted flocculation treatment process as barite precipitation was more than 90 percent complete in the first 10 minutes of water mixing.

Ballasted-sand flocculation was investigated as an alternative to membrane filtration for treatment of flowback water. Preliminary tests of AMD and flowback mixtures outlined an optimal pH range, and both rapid and slow mixing is required for optimum results. Anionic flocculant aids performed better than cationic aids. The footprint for a ballasted-sand flocculation system is smaller than that of conventional treatment systems.

Barium precipitation continues to affect hydraulic fracturing activities. The flow velocity of the proppant in the shale was studied to determine the range of velocities at which barium sulfate can flow through the fractures. A low velocity may cause the precipitate to settle and clog fractures.

Sub-micron particles in flowback water samples caused severe microfiltration membrane fouling during the early stages of filtration. Pretreatment of the flowback/AMD with coagulation and chemical oxidation has been shown to improve membrane performance by causing aggregation of the small particles. Ongoing research is focused on determining the optimal coagulant selection and dosing range needed to reduce membrane fouling.

The AMD database has been updated to include flow rates and chemical analyses over a period of time for 242 sites, so the variability in flow and contaminant levels can be taken into account. Users will be able to search for AMD sites based on desired flow rates, chemical analyses, and location with search results being displayed on a map.

Ceramic microfiltration experiments with a mix of flowback water and AMD were performed to determine sustainable operating conditions. Even under low transmembrane pressure and medium-high cross flow velocity, the membrane performance in terms of permeate flux is very poor. The membrane is rapidly fouled and neither barite nor calcite crystals that formed during mixing of the AMD and flowback water are responsible. Organic matter present in the flowback appears to be the main fouling agent.

Three chemical equilibrium models have been used to determine the effects of mixing simulated flowback water with solutions of sodium sulfate and sodium bicarbonate. These experiments will identify those metals that can be removed from flowback water when it is mixed with AMD. Site visits to select potential locations for field experiments involving flowback and AMD water have been completed, and discussions are ongoing to determine the location of the field test. Researchers have established working relationships with major natural gas developers, well drilling and hydrofracturing companies, the Pennsylvania Department of Environmental Protection (PADEP), water treatment companies, and watershed groups interested in the outcome of this project.

Site visits to select potential locations for field experiments involving flowback and AMD water have been completed, and discussions are ongoing to determine the location of the field test. Researchers have established working relationships with major natural gas developers, well drilling and hydrofracturing companies, the Pennsylvania Department of Environmental Protection (PADEP), water treatment companies, and watershed groups interested in the outcome of this project.

Flowback water locations and AMD sites in Pennsylvania were catalogued with assistance from the Appalachian Shale Water Conservation and Management Committee and the PADEP. To date, analyses of 160 samples of flowback water have been entered into their database. The average TDS was 106,000 mg/L (range of 680–345,000 mg/L) with a strong correlation between Cl and TDS concentrations. Water quality analyses of 140 AMD sites were conducted. AMD flow rates, proximity to gas wells, and levels of sulfate and other constituents were considered for use of AMD for flowback water treatment. The collected AMD samples appear, based on their chemistry, to be compatible for mixing with flowback water and re-use as fracturing fluid.

A Project Advisory Committee consisting of seven industry partners and government agencies was formed to help determine the most advantageous solution to the problem of produced water treatment and disposal in Pennsylvania.

The project has been completed and the final report is available below under "Additional Information".

$794,226

$284,010

NETL - Sandy McSurdy (sandra.mcsurdy@netl.doe.gov or 412-386-4533)
University of Pittsburgh - Dr. Radisav Vidic (vidic@pitt.edu or 412-624-1307)

Final Project Report [PDF-5.86MB]