Biocides control Barnett shale fracturing fluid contamination

May 18, 2009
A team, composed of the operator and service company staff, tested various biocides to determine the best chemical for controlling bacteria in fluids used for fracturing Barnett shales in the Fort Worth basin of Texas.

A team, composed of the operator and service company staff, tested various biocides to determine the best chemical for controlling bacteria in fluids used for fracturing Barnett shales in the Fort Worth basin of Texas.

The selected biocide provided consistent performance in the quick kill of aerobic, fermentative, and sulfate-reducing bacteria as well as the long-term preservation of the produced fluid.

Operators commonly use hydraulic fracturing with varied proppant concentrations for extracting gas from shale formations. The fracturing fluid consists of water and polyacrylamide or sugar-based polymers.

Bacteria typically contaminate water obtained from local rivers, lakes, or oil field wastewater. Polyacrylamide and other organic compounds can serve as food for the bacteria. As a result, the frac fluid is a fertile breeding ground for sulfate-reducing bacteria (SRB) and acid-producing bacteria (APB).

If SRB and APB become established in the wellbore, production of hydrogen sulfide, iron sulfide, and microbial induced corrosion can damage production lines, surface equipment, and gas-gathering systems.

Bacterial contamination

Because the Barnett shale has very low hydraulic permeability, the water volumes used to frac wells have ranged from less than 1 million gal on early vertically completed wells to more than 5 million gal on recently completed horizontal wells.

The water used comes from several sources, including freshwater supply wells, chlorinated city water, rainwater, pond water, and lake water. Each water source contains some bacterial contamination.

During droughts or to reduce water-hauling expenses, operators on occasion will reuse flowback water from previous frac jobs. This practice usually increases bacterial contamination and solids loading.

The initial slick-water fracs used 500-bbl portable tanks for storing source water on location. These tanks provide a relatively easy way for controlling water quality because service companies could easily clean the frac tanks and water-hauling trucks and chemically treat the water in the tanks.

As frac jobs grew larger, it became uneconomical to use frac tanks to store water. The most common solution placed the frac water in lined or unlined earthen pits that are open to the atmosphere. As a result, dust, rain, and surface runoff can contaminate the water. The water can remain dormant in the pit for days to months before a frac job, leading to stagnant, highly contaminated water. Often, the jobs require a mix of several different water sources that may result in scale formation and increased bacterial activity.

Seasonal temperature fluctuations will affect bacterial activity in the frac fluids, resulting in higher bacteria concentrations in the warmer spring and summer months and lower activity in the cooler fall and winter months. Consequently, success in the winter may not translate to a successful frac biocide program in the summer without adjustments to the biocide loading rates.

Frac jobs pump large volumes of water downhole under high pressure, resulting in near-wellbore cooling that provides a favorable temperature for bacterial growth. In addition, the operator may shut in the wells for days to weeks following the frac job to await installation of surface processing equipment and flowlines.

Without a proper biocide treatment, frac-water bacteria can become established downhole and near the wellbore during the frac job and subsequent shut-in period. These bacteria then can contaminate separators, water tanks, flowlines, and disposal facilities downstream.

Bacterial contamination can produce biogenic sulfide (souring), form iron sulfide (black water), cause plugging, and result in corrosion failures of downhole equipment, surface separation and storage tanks, and flowlines.

Preventing contamination requires a good biocide program to prevent system souring, solids formation, and bacteria-related corrosion failures.

A good program has several key stages, including analysis of various water sources, biocide selection, implementation, monitoring, and optimization.

Source water analysis

Source water analysis requires proper sample collection. The Barnett shale program included collection of source water samples in clear, 8-oz plastic bottles from the source well, pond, lake or pit, flow-back water, separator, and produced water storage tank. The procedure involved:

  • Rinsing out the bottles with system water before collecting the sample.
  • Leaving head space in the bottles in the oxygenated source well, pond, lake, and pit water samples.
  • Filling the bottles to overflowing and capping them off immediately to maintain an anaerobic environment for the separator and produced-water tank samples.

The water analysis used the serial dilution enumeration technique for semiquantitative enumeration of viable general aerobic APB, and SRB. The technique included serial dilutions with culture media adjusted to system water salinity in accordance with the NACE Standard Test Method 0194-2004.1

The salinity of the culture medium was adjustedto match the salinity of the water tested. The APB enumeration used modified phenol red dextrose while the SRB enumeration used a proprietary self-adjusted SRB medium. To simulate summer fracturing conditions, the culture media was incubated for 28 days at 85º to 90º F.

For quantitative enumeration of total bacteria, the analysis included the observation of samples through a microscope equipped with a fluorescence epi-illuminator in accordance with the NACE Standard TMO-194-2004.1

Preparation of the samples involved filtering them through black, 25-mm diameter, 0.2-µ pore diameter polycarbonate filters, and then staining the bacteria trapped on the filters with acridine orange, a chemical that binds to nucleic acids contained within all bacterial cells.

The tests used an epifluorescence microscopy at 1,000-times magnification to analyze the filters mounted on microscope slides.

Enumeration of bacteria involved averaging the number of bacteria counted over numerous microscopic fields and then calculating bacteria/ml of sample based on the amount of sample filtered, the filter size, and the microscopic field size.

Planktonic kill tests

To determine the best biocide for treating the fracturing fluid bacterial populations, the analysis studied the planktonic bacterial kill on several different chemically free-water sources.

The test involved inoculating the water with previously cultured indigenous bacteria, weighed out into clean 8-oz glass prescription bottles, and dosing with biocides at various concentrations. In addition, a control sample only had indigenous bacteria inoculated.

The analysis exposed the bacteria in each samples to the biocides for various contact times, such as 1 hr, 24 hr, 1 week, and 3 weeks. The longer contact times simulated the fracturing fluid water retained by the reservoir during the flowback period.

A serial dilution technique enumerated the surviving bacteria in each biocide-treated and control sample. The APB enumeration used samples diluted into a freshwater phenol red dextrose medium, while the SRB enumeration used samples diluted into a freshwater proprietary SRB medium.

To simulate summer conditions, the tests incubated the serially diluted culture vials for 28 days at 90º F.

The tests involved a six-vial serial dilution for the biocide-treated samples and an eight-bottle serial dilution series for the control samples.

The water source survey indicated high bacterial contamination, as shown under 1,000X magnification (Fig. 1).
Click here to enlarge image

The analysis also compared the results with other Barnett shale operator kill studies.

Water-source survey

A water-source survey indicated that each fracturing water source was highly contaminated with bacteria (Table 1 and Fig. 1). The general aerobic and acid-producing bacteria were the predominant bacterial population in most pits and ponds. Fig. 2 shows the quality variation of the samples.

Click here to enlarge image

Pits 2 and 4, however, contained very large SRB populations as well (greater or equal to 1 million SRB/ml). The amount of turbidity and discoloration in the water correlated with the SRB concentrations, with the highest SRB concentrations observed in the pit waters that were black and dark brown in color.

Sampled water quality varied widely (Fig. 2).
Click here to enlarge image

The pits and ponds that contained tan or opaque water had moderate concentrations of SRB (10-10,000 SRB/ml). Lined Pit 3 contained clear water and had the lowest concentration of APB and SRB. The natural pond samples had algae.

Click here to enlarge image

The pit with black water was a mixture of flowback and fresh frac water. This mixture provided the proper nutrient balance to stimulate SRB activity, resulting in biogenic sulfide production and black water.

Wellsite survey

The survey evaluated several sites at newly drilled and fractured gas wells. At each wellsite without a nearby naturally occurring pond, the operator drilled freshwater source wells for supplying water for the frac job. The job required placing the water from the source well, either pumped or hauled by truck, into a lined pit near the well location.

Click here to enlarge image

Upon completion of the frac job, the operator shut in the well from a few days to several months while waiting for installation of the surface equipment. Upon installation, flowback of the initial surge of frac water and sand was routed to the frac tanks.

Click here to enlarge image

After flowing back for several days to the frac tanks, the operator rerouted the produced fluids to the separator for separating the produced water and condensate from the natural gas. The separated condensate and water go to storage tanks on location. From the tanks, trucks haul condensate to a sales point and the water to a saltwater processing facility for disposal into an isolated formation downhole.

Click here to enlarge image

The wellsite survey collected samples from the water source well, the fracturing water pit, the production separator, and the produced-water storage tanks (Fig. 3).

Despite the lining of the frac pit, the total bacteria count increased by 1.5 log units if the source-well water was in an open frac pit. The majority of frac pits had measurable SRB concentrations.

The separator sample showed high bacteria concentrations in the microscopic analysis, but bacterial culture media indicated that the majority of the bacteria observed by microscopy were not viable, with only 100 viable APB/ml recovered.

These microscopic photomicrographs from left to right are of the source well waste, pit water, and produced water storage (Fig. 4).
Click here to enlarge image

As the water proceeds downstream into the produced-water storage tank, bacteria concentration increased with the viable APB and SRB populations exceeding 1 million bacteria/ml. Microscopic analysis (Fig. 4) showed masses of bacteria surrounded by exo- polysaccharides (slime) indicating the formation of sessile (attached) bacterial biofilms in the tank bottoms.

Early Barnett shale operators experienced produced-water storage tank and flowline failures because of the stagnant conditions and solids deposition in the storage tanks. The water sat relatively stagnant in the storage tanks for weeks to months before removal and transportation to a water-disposal facility.

Click here to enlarge image

Fig. 5 shows the average bacterial concentration observed at each sample location during the field survey. Fig. 4 shows representative microscopic photomicrographs of each sample location analyzed during the field survey.

Biocide selection

In the last 4 years, Baker Hughes Inc. has run planktonic kill studies on at least 20 different Barnett shale fracturing water sources collected from six different operators. Because each bacterial control product contains different biocidal active ingredients, chemical activity levels, and product costs, our analysis evaluated a range of treating concentrations for each biocide in an effort to provide a cost-equivalent comparison of the different chemistries.

The kill study discussed in this article evaluated six different biocides at various concentrations. The study exposed bacterial populations in the frac fluid sample to the biocides for 1 hr and 24 hr before quantifying the concentration of viable APB and SRB in each sample as compared to an untreated control.

In this study, 50 ppm Biocide A and 100 ppm Biocide C provided a 5-6 log reduction in the SRB and APB populations after 1 and 24 hr of contact time with the biocides. Biocide B required 200 ppm to obtain similar reductions in the APB and SRB populations.

A 75-ppm concentration of Biocide A and 150-ppm concentration of Biocide C reduced the APB and SRB populations below detectable limits. No other biocides evaluated provided substantive reductions in the APB and SRB populations at the evaluated concentrations.

Click here to enlarge image

Table 5 shows a cost-equivalent comparison of 22 planktonic kill studies performed by Baker Hughes with Barnett shale frac waters. The products shown represent the biocide classes most commonly used for fracturing operations in the Barnett shale.

The average APB populations in the untreated control sample are four log units higher than the average SRB populations. The kill study results indicate that it is more difficult to kill APB than SRB. For the same relative cost, Product A provided 1,000 times greater kill of the APB populations than Product B and 100 times greater kill than Product C.

Bacterial populations in the frac water sources generally have become more difficult to kill over time, requiring increased concentrations of biocide. The analysis attributed this observation to increasing age and reuse of the pit waters.

Biocide A was the preferred biocide for the Barnett shale frac fluid treatment due to:

  • Consistent, cost-effective performance in the planktonic kill studies.
  • Compatibility with the frac fluid friction reducers and other frac fluid additives.
  • Passage of the Barnett shale core compatibility studies.
  • Absence of foam. Some operators have reported undesirable foaming issues with the addition of glutaraldehyde and quaternary ammonium-blend biocides.
  • Noncorrosive at typical frac loading rates.
  • Thermal stability at reservoir conditions.

Newly built dedicated frac chemical-injection trucks ensured that the frac fluids were treated with the appropriate biocide concentrations. The frac chemical trucks tie into the blender and flowmeter of the fracturing operation, allowing for injection of the biocide and other production chemicals, such as oxygen scavengers and scale inhibitors, on the fly into the fracturing fluid as it is pumped downhole.

This procedure ensures treatment of all fracturing fluid at the desired biocide concentration. Separate measurement of the chemical additives through a closed manifold system ensures adequate mixing of the chemicals with the frac fluids and minimizes oxygen entry into the frac fluids during chemical injection.

The jobs adjust chemical loading rates based on prefrac and postfrac monitoring results and visual water quality.

Biocide monitoring

Table 6 breaks down the frac biocide program monitoring results for 300 wells treated during the last 4 years. It categorizes the wells by operator, biocide-loading rates, lining of the frac pits, and county. Columns 2 and 3 show the percentage of wells in each category that have low levels of viable SRB or APB. This study defines low level as 1,000 or less viable SRB or APB/ml in the frac flowback sample.

Click here to enlarge image

Operator 2 was one of the first operators in the Barnett shale. In its early frac biocide treatment programs, Operator 2 used 37 ppm of Biocide A. Based on the poor monitoring results (only 35% and 48% efficacy rates for SRB and APB, respectively), it increased its biocide-loading rate to 60-100 ppm with the loading rate dependant on the visual water quality and seasonal temperature variation.

Click here to enlarge image

Operator 1 had the best overall biocide performance, despite using identical or lower biocide concentrations than the other operators. Figs. 6 and 7 break down the frac biocide program monitoring results for Operator 1. One can attribute this high-level performance mostly to lined frac pits, which will provide lower sediment loading in the water.

Click here to enlarge image

In addition, Operator 1 has its flowback sample monitoring point at the wellhead. Many other operators do not have a sampling point at the wellhead. Therefore, they have to take the flowback sample from the separator, which can result in aberrant readings due to stagnant fluid or solids deposition in the separators.

The first area of the Barnett shale developed was in Denton and Wise counties. Because of this, this area has older frac pits and a pond that have been in use for a much longer time and have had much more fluid mixing.

These factors eventually led to higher bacterial loading and poorer water quality in these pits. A comparison of equivalent biocide loading rates from different operators shows that the frac biocide performance in Denton and Wise counties is slightly lower than for the other counties, such as Johnson and Hood counties where development has just started.

Click here to enlarge image

Table 7 compares the effect of seasons on biocide performance for the same 300 wells. The poorest biocide performance (54% and 38% for APB and SRB, respectively) is in the warmest quarter of the year, an average of 82.5º F. during July to September. This information prompted recommendations to increase the biocide loading rates in the warmer late spring, summer, and early fall months.

Control program optimization

Design of a proper bacterial control program requires a thorough understanding of system operations. Gaining this knowledge requires a survey of the various types of fracturing water sources and the frac fluid flowback process to assess the severity of the microbial contamination. One cannot overlook a bacteria problem because a frac water source is chlorinated or of drinking-water quality.

Also needed is the assessment of the bacteria-related problems encountered during standard operating procedures and recognition, simulation, and evaluation of the situations that cause deviations from procedures so as not to overlook additional bacterial and solids loading.

Mixing of frac waters from different sources and reuse of flowback waters can cause elevated bacterial contamination and scaling issues. Eliminating these problems requires compatibility mixing studies with various waters before blending them in the field. The study should evaluate each frac water source multiple times throughout the year to assess how seasonal variations affect bacterial activity and water quality.

These practices provide information on the severity of bacterial contamination and solids loading in the frac water sources, allowing for an adjustment of the biocide loading rates.

When designing a biocide program, one should consider several different factors. First, it is imperative that biocide selection testing include waters representing each type of frac water source. One should only consider broad-spectrum biocides to ensure control over the heterogeneous microorganism population present, such as SRB, APB, algae, and fungi.

After the generation of a list of products that perform well against the frac water microbial populations, it is important to assess if the frac water contains components such as H2S, iron sulfide, ammonia, or dissolved oxygen that might degrade biocide performance. It is essential to assess compatibility of the biocides with the frac-fluid additives, such as polymers and any other chemicals including oxygen scavengers, scale inhibitors, corrosion inhibitors, and friction reducers.

Planktonic kill studies provide two important pieces of information required for killing bacteria:

  1. Concentration of biocide required to kill the target microorganisms in the water source.
  2. The amount of contact time required for the biocide to kill the target microorganisms.

Controlling bacteria in frac programs requires storage and treatment of all frac fluid at the predetermined biocide dosage rate in frac tanks or application of the biocide at the desired concentration continuously on the fly, as the frac fluid is pumped downhole.

In frac-tank applications, the jobs should retain the frac fluid in the tanks for the required biocide contact before the fluid is pumped downhole. In on-the-fly applications, one should ensure that the biocide-treated fluid is shut in downhole for the required amount of contact time.

For Barnett shale applications in which the biocide is injected on the fly at the blender, the operator should shut in the wells for at least a 24-hr contact time before bringing the well online.

The jobs should use EPA-registered biocides according to the label specifications. For example, application of an EPA-registered biocide into an open pit is illegal because of the potential for leaching into the groundwater and other surface waters, aquatic organism toxicity, and terrestrial hazards due to animal consumption.

One needs to consider the biocide half-life and recalcitrance to ensure that the biocide will maintain bacterial control in the frac waters throughout fracturing, shut-in, and flowback.

An aggressive monitoring program is instrumental in assessing the performance of the biocide program. Monitoring can allow an operator to verify the merit of various costly operational practices, such as lining of pits.

Monitoring can reveal issues with sampling points. For this type of application, a sampling point at the production wellhead is necessary to assess properly the biocide performance and monitor product residuals. One should use the information gained from the monitoring program to optimize a biocide program and assess system conditions that would require an adjustment of the biocide-loading rate.

When frac pits are filled at the last minute, bacteria enumeration techniques are unable to provide information on bacterial activity before proceeding with the frac job. When flowback waters are reused and mixed with fresh waters as the water source for fracturing operations, a water chemistry check must be done first to insure that the salinity of the culturing media matches the salinity of the frac source water.

Operators have used several different biocides in the Barnett shale gas play. They have incorporated the biocide into slick-water fracturing packages to reduce the chance of bacterial contamination downhole that could cause microbial-influenced corrosion and biogenic souring of the reservoir.

The risks associated with an ineffective microbiocide program reach across all phases of production operations.

Acknowledgment

The authors acknowledge the assistance of Trey Femihough of Baker Hughes.

Reference

  1. NACE Standard Test Method 0194-2004, “Field Monitoring of Bacterial Growth in Oil and Gas Systems,” 2004.

The authors

Click here to enlarge image

Jennifer K. Fichter ([email protected]) is technical specialist, water management and microbial control, for Baker Hughes. Her expertise encompasses oil field microbiology, cooling and boiler system microbiology, and waste water microbiology. Fichter holds an MSc and BSc in microbiology from Iowa State University.

Click here to enlarge image

Karl French ([email protected]) is oil field chemicals account manager for Baker Hughes. French has more than 26 years of sales experience and currently specializes in frac chemical sales in the Barnett shale area. He holds a BS in secondary education from West Texas State University, Canyon, Tex.

Click here to enlarge image

Kelly Johnson ([email protected]) is oil field chemical sales district manager for Baker Hughes with expertise in oil field chemicals. Johnson holds a BS in industrial engineering from Texas A&M University.

Click here to enlarge image

Ron Oden ([email protected]) is operations manager of the Fort Worth North Division for EOG Resources. He holds an ME in petroleum engineering from Texas A&M University.