- Email: [email protected]
Polyacrylamide in hydraulic fracturing fluid causes severe membrane fouling during flowback water treatment
Journal of Membrane Science 560 (2018) 125–131
Contents lists available at ScienceDirect
Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci
Polyacrylamide in hydraulic fracturing fluid causes severe membrane fouling during flowback water treatment
T
Boya Xionga, Selina Roman-Whiteb, Bethany Piechowiczb, Zachary Millerb, Benjamin Farinab, ⁎ ⁎⁎ Travis Taskera, William Burgosa, Andrew L. Zydneyb, , Manish Kumara,b, a b
Department of Civil and Environmental Engineering, The Pennsylvania State University, Greenberg Building, University Park, PA 16802, United States Department of Chemical Engineering, The Pennsylvania State University, Greenberg Building, University Park, PA 16802, United States
A B S T R A C T
Sustainable wastewater management strategies are required to further minimize impacts of high-volume hydraulic fracturing (HVHF) as current practices such as reuse or direct disposal have long term limitations. Membranes can provide superior effluent quality in HVHF wastewater treatment, but the application of these systems is severely limited by membrane fouling. However, the key fouling components in HVHF wastewater have not yet been clearly identified and characterized. Here we demonstrate that fouling of microfiltration membranes by synthetic flowback water is mostly due to polyacrylamide (PAM), a major additive in slickwater fracturing fluids. A synthetic fracturing fluid was incubated with Marcellus Shale under HVHF conditions (80 ℃, 83 bar, 24 h) to generate synthetic flowback water. Different HVHF conditions and fracturing fluid compositions generated a fouling index for flowback water ranging from 0.1 to 2000 m−1, with these values well correlated with the peak molecular weight (MW) (ranging from 104 kDa) and the concentration of high MW components in the water. The lowest fouling index was 10 to 1.5 × observed when PAM was further degraded by ammonium persulfate under HVHF conditions, although this is infrequently used with PAM in current fracturing operations. These results highlight the importance of PAM and its degradation products in fouling of subsequent membrane systems, providing insights that can help in the development of effective treatment processes for HVHF wastewater.
1. Introduction In the past decade, the development of unconventional oil and gas using high-volume hydraulic fracturing (HVHF) has had a significant impact on the U.S. energy landscape. However, the environmental impacts of HVHF have generated tremendous social concerns, largely due to the production of large quantities of wastewater, referred to as flowback and produced water, that contain high salinity (220,000–340,000 mg/L), turbidity, organic constituents (1–5500 mg/L total organic carbon), and radioactivity (gross alpha 50–120,000 pCi/L) [1,2]. In some states, such as Texas, wastewater is primarily disposed of by deep well injection due to the ready availability of class II disposal wells. In Marcellus Shale gas wells in Pennsylvania, 90% of the wastewater (both flowback and produced water) is recycled and reused in subsequent HVHF operations because of the limited number of disposal wells [3]. The influence of such recycling after minimal treatment, particularly the effect of transformed organic additives, on well productivity has not been reported. More importantly, deep well injection
⁎
is not an environmentally friendly management strategy [4]; the opportunity for recycling will be limited when the number of new fracturing jobs declines. There is a clear need for developing processes that can provide cost-effective treatment of flowback and produced water. Membrane systems have been proposed by a number of investigators for treatment of flowback water [5,6], both to remove suspended solids and organics using microfiltration (MF) or ultrafiltration (UF) [7,8] and to remove high salinity/hardness using nanofiltration (NF) [9], reverse osmosis (RO) [10] and forward osmosis [11] or membrane distillation [12]. However, membrane fouling by hydrocarbon [10] and polymeric organics [9,13,14], inorganic scaling species [15], microbial [9] and particulate materials [7,14] in flowback water remains a challenge in efficiently treating the wastewater [16]. Our previous work with flowback water from Marcellus Shale gas wells showed a large variation in both water quality and fouling rates during microfiltration with no apparent correlation between the measured fouling index and either the total organic content or turbidity [14]. The exact nature of the key fouling constituents in these wastewaters remains unknown. This knowledge gap makes it
Corresponding author. Corresponding author at: Department of Chemical Engineering, The Pennsylvania State University, Greenberg Building, University Park, PA 16802, United States. E-mail addresses: [email protected] (A.L. Zydney), [email protected] (M. Kumar).
⁎⁎
https://doi.org/10.1016/j.memsci.2018.04.055 Received 28 February 2018; Received in revised form 30 April 2018; Accepted 30 April 2018 Available online 22 May 2018 0376-7388/ © 2018 Elsevier B.V. All rights reserved.
Journal of Membrane Science 560 (2018) 125–131
B. Xiong et al.
fracturing fluid and dissolved Fe2+ from shale at low pH [27]. There is a critical need to evaluate the fouling of degraded PAM characteristic of the materials in flowback / produced water from HVHF operations. The primary objectives of this work were to: 1) quantify the effect of downhole pressure and temperature on fouling indices of PAM compared to ‘raw’ fracturing fluid and flowback water; 2) evaluate the effect of combinations of PAM and other additives on the fouling index; and 3) examine the correlation between the fouling index and PAM size and concentration. These results aim to provide new insights into the fouling characteristics of flowback water, while also identifying possible strategies to reduce membrane fouling. The data also suggest possible concerns regarding plugging of the micro/nano-scale pores in shale formations (similar in size to those in porous membranes) during reuse of flowback water.
Nomenclature HVHF FR PAM MF UF MW SEC
High volume hydraulic fracturing Friction reducer Polyacrylamide Microfiltration Ultrafiltration Molecular weight Size exclusion chromatography
difficult to design treatment processes that are effective for managing fracturing wastewaters. The organics in flowback water come from both the chemical additives used in the hydraulic fracturing fluid and the hydrocarbons extracted from the shale. Over 1000 chemicals [17], some of which are proprietary, have been used in hydraulic fracturing, and many of these have been detected in flowback and produced water [18,19] as well as in contaminated surface water / groundwater samples [20–22]. Many of these analyses utilized advanced gas chromatography coupled with mass spectroscopy, which can identify hydrophobic and volatile organics such as hydrocarbons; however, components that are hydrophilic and larger than 1 kDa, such as polyacrylamide (PAM) used as a friction reducer (FR) and guar gum used as a gelling agent, cannot be detected in these analyses [21]. Two very recent studies found that 90% of the organic matter in fracturing wastewater is hydrophilic, and some wastewaters have been shown to contain 20–40% biopolymers [13,23]. These polymers, including high MW PAM, can be significant membrane foulants. For example, PAM was found to be the major contributor to the total membrane resistance in fouling tests performed with synthetic oil-field polymer flooding wastewater [24]. Wang et al. [25] reported that PAM used as a coagulant caused MF fouling predominantly by surface pore blockage, with the rate of fouling determined by the molecular weight and concentration of the polymer solution. Liu et al. [26] utilized atomic force microscopy to correlate intermolecular forces with the fouling resistance provided by hydrolyzed polyacrylamide during filtration through a polyvinylidenefluoride UF membrane. This study concluded that fouling was dominated by the resistance from the concentration polarization layer formed by intermolecular attraction, rather than the gel layer formed by polymer-membrane attraction, where both attractive forces are attributed to hydrophobic interactions involving the polymer backbone in combination with hydrogen bonding. However, it is not possible to directly extrapolate from these studies to the fouling characteristics of PAM in flowback water, particularly given the complex transformation of PAM that can occur under HVHF conditions due to interactions with the solid shale and with other additives present in the fracturing fluid. Recent work in our laboratory has demonstrated that PAM undergoes significant degradation via a free radical mechanism at the high temperatures, where the free radicals are generated by reactions involved dissolved oxygen present in the
2. Materials and methods 2.1. Synthetic fracturing fluid Stock chemicals of friction reducer (FR), biocide, corrosion inhibitor, crosslinker, and surfactant were provided by Weatherford Chemical, Inc. and are commercial products used in HVHF operations. Each additive contains a mixture of chemicals as provided by the supplier (see Table SI); the exact composition and concentration of individual components are proprietary. FR contains PAM and petroleum distillate. Vacuum incubation of a neat FR stock solution yielded a nonvolatile (polymer) portion of approximately 40% by mass. A previous study reported that raw FR fluid contained roughly 0.7 g/L PAM with a peak MW of 15 MDa based on size exclusion chromatography analysis [27]. Ammonium persulfate (breaker), citric acid (iron control), and potassium hydroxide and sulfuric acid (pH adjustment) were purchased from Sigma-Aldrich (St. Louis, MO) and were prepared using deionized (DI) water from a Barnstead Nanopure water purification system with a resistivity of > 18 MΩ cm. This work primarily considers a slickwater fracturing fluid, which mainly contains FR, biocide, corrosion inhibitor, surfactant and iron control agents, given that 84% fracturing operations in the Marcellus utilized slickwater frac based on a review of 100 drilling logs and 97% nationwide utilized FR based on a review of 750 drilling logs on FracFocs.org. Hybrid frac combines slickwater and gel frac, where both FR and a gelling agent (such as guar gum) are used. We also performed limited experiments simulating a hybrid fracturing fluid by adding breaker and crosslinker to the fluid; critically, gelling agent could be another significant source that contributes to membrane fouling, which will be discussed in detail in a separate paper. Gel frac without the use of FR is rare and thus was not considered in the current study. The concentrations of the fracturing additives used to prepare the synthetic fracturing fluid (Table 1) were based on FracFocus.org and literature sources [17,28]. Additive concentrations tend towards the high ranges reported in the literature. Synthetic fracturing fluids were prepared by mixing appropriate additives with DI water. The
Table 1 Chemicals used to prepare the synthetic fracturing fluid. Chemical additives
Specific compounds (provided by supplier)
Concentration
Friction Reducera Biocidea Corrosion Inhibitora
Polyacrylamide, petroleum distillate 2,2-Dibromo- 3-Nitrilopropionamide (DBNPA) Isopropanol, Ethylene glycol, N,N-Dimethylformamide, 2-Butoxyethanol, Cinnamaldehyde, Tar Bases, 1-Decanol,1-Octanol, Triethyl phosphate Sodium tetraborate pentahydrate，Glycerine, Potassium hydroxide Ammonium Persulfate Citric Acid Ethoxylated alcohol (C6-C12), Ethylene Glycol, Isopropyl Alcohol, D-limonene, 1-Octanol Potassium hydroxide and sulfuric acid
0.15% v/v 0.0017% v/v 0.0007% v/v
Crosslinkera Breaker Iron Control Surfactanta pH Adjustment a
0.03% v/v 0.011% w/v 0.0014% w/v 0.075% v/v pH adjusted to 7.2
indicates chemical is provided by Weatherford Chemical, Inc. Final concentrations are based on the volume fraction of the liquid chemical stock (e.g., % volume of chemical stock/volume of water). 126
Journal of Membrane Science 560 (2018) 125–131
B. Xiong et al.
2.4. Polymer size and concentration analysis by size exclusion chromatography
mixtures were blended for 30 s and the pH was then adjusted to 7.2 using potassium hydroxide or sulfuric acid as needed.
Molecular weight (MW) distributions of PAM in the fracturing fluid and flowback water were evaluated by conventional size exclusion chromatography (SEC) using a Waters HPLC 1515 system and a Shodex SB-806M HQ aqueous SEC column with 0.05 M Na2SO4 as the mobile phase at a flow rate of 0.5 mL/min. All samples were pre-filtered through 0.45 μm cellulose acetate syringe filters before injection (100 µL sample volume), with peak positions and concentrations determined by a refractive index detector. Nonionic PAM standards (American Polymer Standard Corporation, Mentor, OH) with a concentration of 0.5 mg/mL dissolved in 0.05 M Na2SO4 were used for calibration. Polymer concentrations were determined using the dn/dc coefficient determined experimentally as described in SI-Fig. 1 and Table SI−1.
2.2. Synthetic flowback water Synthetic flowback water was obtained by reacting the synthetic fracturing fluid with 25 g/L shale at 80 ℃ and 83 bar for 24 h simulating downhole conditions encountered in natural gas reservoirs. The variation in these conditions across different gas formations can be large [29], and its impact on polyacrylamide degradation was investigated previously. Marcellus Shale was collected from an outcrop near Frankstown, PA; it had a 6-meter overburden and had been exposed to the atmosphere. The shale sample was pulverized and sieved to a size of 0.3–2 mm using a method described previously [27]. Pulverized shale was mixed with freshly prepared fracturing fluid and was incubated in a 0.5 L T316 stainless steel Parr reactor with digital temperature control and nitrogen gas pressurization. After reaction, the resulting suspension was centrifuged at 13,700 g for 1 h at 4 ℃, with the supernatant collected as the synthetic flowback water. Filtration experiments were conducted within 24 h after centrifugation.
3. Results and discussion 3.1. Polyacrylamide used in FR is the dominant foulant among all the additives The friction reducer (FR) is one of the most important ingredients in slickwater HVHF and can potentially contribute to a large proportion of membrane fouling due to its large size. Filtration experiments with ‘raw’ unreacted combinations of the other additives shown in Table 1 (without FR), including breaker, crosslinker, surfactant, biocide, corrosion inhibitor and iron control agent, showed relatively little fouling, with the flux declining by less than 40% after a volumetric throughput of more than 100 L/m2 (Fig. 1A). Incubation of these additives (without FR) at HVHF conditions for 24 h (in the presence of 25 g/L shale, 80 ℃, and 83 bar), yielded a synthetic flowback water that also showed negligible fouling up to volumetric throughputs of 300 L/m2 (Fig. 1B). In contrast, the ‘raw’ unreacted FR showed a very rapid flux decline, with the flux decreasing to less than 0.05% of the initial value within 2 L/m2, corresponding to a fouling index of 1400 m−1. Incubation of the FR under HVHF conditions reduced the fouling index to 73 m−1, but this still corresponded to a flux decline of nearly 90% after 30 L/m2 of filtration. These values of the fouling index are similar to those reported previously for actual flowback water collected from the Marcellus shale play [14] (symbols labeled S1, S2, and S3 in Fig. 1B). Mixing FR with other additives resulted in a flowback water with lower fouling potential compared to only FR, which will be discussed in a later section. SEC analysis demonstrated that the molecular weight of PAM in FR
2.3. Membrane fouling experiments Fouling experiments were conducted using 0.2 μm pore size polyvinylidenefluoride (PVDF, EMD Millipore) membranes in a 10 mL Amicon 8010 dead-end filtration cell at a constant pressure of 0.27 bar (4 psi) at room temperature based on results from previous studies of MF fouling using actual flowback water from the Marcellus region [14]. Details of the filtration experiments were described previously [14]. The fouling rate was quantified using the initial membrane fouling index (m−1):
Fouling index = −
1 dJ J0 d V
() A
(1)
where J is the filtrate flux (L/m2 h), V/A is the specific throughput (L/m2), and J0 is the initial flux. The derivative was calculated using first order derivative formulas developed by a finite difference method accounting for the variable time steps [14]. This method does not rely on any assumptions about fouling mechanisms while providing a quantitative estimate of the initial fouling rate. Details on the calculations are presented in the SI.
Fig. 1. PAM in friction reducer (FR) is the major fouling component compared to other additives. Flux decline during microfiltration of: (A) Raw (unreacted) FR compared to other additives (no FR); (B) Flowback containing FR or other additives. Synthetic flowback was generated from incubation of raw fluid at 80 ℃, 83 bar with 25 g/L shale; S1-S3 are actual flowback water samples collected from the Marcellus shale play [14]. (C) Size exclusion chromatographs of raw (black solid) and flowback (red dash) of FR; and raw (dash blue) and flowback (solid green) of other additives (called “other”). Blue and green lines are completely overlapped. All FR fluids contain FR at a concentration of 0.15% v/v. Fouling experiments were conducted using 0.2 μm PVDF microfiltration membranes in a dead-end filtration cell at 0.27 bar. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) 127
Journal of Membrane Science 560 (2018) 125–131
B. Xiong et al.
Gel fracturing operations utilize a crosslinked polymer gel such as guar gum and its derivatives to enhance fluid viscosity in order to more effectively carry proppant throughout the fractures. During flowback, the viscous gel can impair fracture conductivity and hinder gas flow; thus, oxidative breakers such as peroxide and persulfate are commonly used to break down the polymer gel and enhance flowback recovery [35,36]. In slickwater fracturing [37], breaker is occasionally used in part due to recent concerns over potential formation damage by accumulated PAM [34,38]. Our review of 100 drilling logs (randomly selected) in the Marcellus play from 2015 to 2017 (FracFocus.org [39]) indicate that breaker was only used when polyacrylamide was mixed with gel fluid (giving a so-called ‘hybrid’ fracturing fluid); this was 16% of the total HVHF jobs. Thus most of the fracturing jobs utilizing only FR do not experience additional degradation by breaker. Among all the remaining additives, the presence of crosslinker and surfactant slightly increased the resulting MW distribution and fouling of the flowback water. The MW distribution and fouling of flowback with all the ingredients (FR+other) and FR with breaker, crosslinker, and surfactant were found to be similar (Fig. SI-4), suggesting that the remaining additives (i.e. iron control agent and biocide) had no measureable effect on flowback chemistry. The sample incubated at HVHF conditions with crosslinker (sodium tetraborate pentahydrate, 0.03% v/v) had a peak MW of 15 ± 3 kDa (blue line in Fig. 2B), which is slightly larger than the peak MW for the sample incubated without the crosslinker (8 ± 3 kDa). The sample incubated with crosslinker also had a slightly larger fouling index of 1.8 ± 1.8 m−1 compared to < 1 m−1 for the sample without crosslinker. The crosslinking chemistry associated with borate and PAM has not been previously studied. The lower degree of polymer degradation observed in the presence of the crosslinker may be the result of enhanced radical termination [30] caused by borate enhanced persulfate activation [40]. Surfactant is used in oil and gas extraction to reduce the interfacial tension between injected fluids and the formation which can enhance water recovery [41]. Limited experiments were thus performed after addition of surfactant (ethoxylated alcohol, 1-octanol and D-limonene [28], 0.075% v/v). The peak MW of the sample incubated with surfactant under HVHF conditions was 120 ± 7 kDa (labeled as FBCS in SIFig. 4), which is higher than the sample without surfactant
is much larger than that for any of the other additives as expected. Fig. 1C shows the raw chromatographs for the different samples examined in Figs. 1A and 1B; the largest components have the smallest elution volume since they are strongly excluded from the resin pores. The calculation of the peak MW from the SEC data is discussed in the SI. The PAM in FR had a peak MW as high as 1.5 × 104 kDa while the other additives had molecular weights less than 1 kDa (no peak shown on chromatograph of SEC). However, incubation of the FR for 24 h under HVHF conditions, leading to what we refer to as “FR flowback”, caused a reduction in the peak MW to approximately 200 kDa. This is discussed in more detail in the next section. 3.2. Reaction under HVHF conditions significantly reduced fouling index of FR SEC analyses suggest molecular weight reduction of PAM in FR occurred during HVHF reaction and that this is directly related to the reduced rate of fouling. In a previous study, we demonstrated that exposure to Marcellus Shale promoted both adsorption and a free radicalinduced chemical degradation of PAM, leading to a significant reduction in MW [27]. It was also found that elevated temperature and the presence of shale were the two critical conditions causing significant polymer degradation; in comparison, high pressure and salinity played a minor role. Additional filtration experiments were conducted using only FR incubated at room temperature (with shale under 83 bar pressure) and without shale (at 80 ℃ under 83 bar pressure). As shown in Fig. 2 (open circles and inverted triangles), these samples showed much greater rates of fouling than the FR exposed to shale at 80 ℃ (filled triangles and squares). In contrast, reaction at 1 bar generated a flowback water with very similar fouling behavior as that for the flowback water generated at 83 bar reaction pressure. These differences are also seen in the peak MW which were reported previously: the samples exposed to shale at 1 and 83 bar had a peak MW of 200–300 kDa while those treated at room temperature had peak MW of 4000 kDa. The small reduction in MW for the sample treated at room temperature was caused by adsorption of PAM to the shale; samples treated in the absence of shale had peak MW of 5000 kDa [27]. Similarly, salinity (3 M NaCl or 1.5 M CaCl2) had little impact on either polymer degradation or fouling (Fig. SI-2). These results suggest that different reservoir conditions across different shale formations could lead to flowback and produced waters with very different fouling potentials due to changes in the degree of PAM degradation and adsorption; this will be discussed in more detail in the next section. 3.3. Persulfate breaker reduces the fouling index by enhancing polymer degradation The fluid that mixed all other additives with FR had identical molecular size distribution as only FR (before reaction), and yet generated a flowback with peak MW of 85 ± 5 kDa (Fig. SI-3) and a fouling index of 3.3 ± 3.0 m−1 (Fig. SI-4B). Experiments with different fluid compositions revealed the important role of persulfate breaker among all the other additives in causing further polymer degradation and mitigating membrane fouling. The addition of 0.11 g/L ammonium persulfate significantly reduced the peak MW of our FR flowback to as low as 8 ± 3 kDa (after 24 h incubation of FR in the presence of ammonium persulfate under HVHF conditions) (Fig. 3B), consistent with the large reduction in apparent viscosity observed by Gao et al. [30]. As a result, the synthetic flowback water caused only 20% flux decline within 300 L/m2 (Fig. 3A), corresponding to a fouling index below 1 m−1. Persulfate decomposes to sulfate radicals (SO4•-), particularly at high temperatures [31,32]:
S2 O82 −→2SO4• −
Fig. 2. The fouling potential decreases after treatment at HVHF conditions. Flux decline during microfiltration of: Control (black square): synthetic FR flowback water generated from incubation at 80 ℃, 83 bar with 25 g/L shale; RT (red circle): flowback generated at room temperature, 83 bar with 25 g/L shale; 1 bar (blue triangle): flowback generated at 80℃, 1 bar with 25 g/L shale; No shale (green inverted triangle): flowback generated at 80, 83 bar with 0 g/L shale; All fluids contain FR at a concentration of 0.15% v/v. Fouling experiments were conducted using 0.2 μm PVDF microfiltration membranes in a deadend filtration cell at 0.27 bar. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
(2)
Sulfate radicals can then cause chain scission of the PAM backbone via free radical degradation [30,31,33,34]. 128
Journal of Membrane Science 560 (2018) 125–131
B. Xiong et al.
Fig. 3. Fouling behavior and polymer size of friction reducer (FR) decreases when breaker and crosslinker are added. (A) Flux decline during microfiltration and (B) size exclusion chromatographs of FR synthetic flowback water generated from FR alone (FR, red), FR plus breaker (blue), FR plus breaker and crosslinker (green). All incubations of raw fluid were conducted at 80 ℃, 83 bar with 25 g/L shale. Fouling experiments were conducted using 0.2 μm PVDF microfiltration membranes in a dead-end filtration cell at 0.27 bar. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
consistent with previous findings showing a correlation between molecular weight and transmembrane pressure during microfiltration of both hydrolyzed polyacrylamides (with MW between 1.5 and 5000 kDa) and poly(diallyldimethylammonium) chloride (100–500 kDa) solutions [25]. Samples with peak MW < 20 kDa fall in a low fouling region (< 1 m−1). Significant fouling was observed when the peak MW started to exceed 50 kDa, corresponding to a hydrodynamic size of approximately 45 nm, which is still well below the 200 nm nominal pore size for the membrane. This may suggest internal pore fouling, although it is also possible that the fouling was due to pore blockage associated with the larger components in the tail end of the size distribution and / or the simultaneous blockage by multiple polymer chains. This was analyzed further by plotting the fouling index as a function of the concentration of the high MW components in the distribution (details described in SI). The data (Fig. 4B) show a high degree of correlation, providing further evidence that the high MW components are the key foulants in these HVHF fluids. The highest fouling was seen with the raw fluids and with the fluids reacted without shale or at room temperature. Incubation under HVHF conditions lowered the fouling index by more than two orders of magnitude, with the fouling index influenced by the composition of the fluid, particularly the presence of breaker and crosslinker.
(15 ± 3 kDa, blue line in Fig. 2B) but comparable to the MW of only FR flowback. This sample gave a fouling index of 3.3 ± 3.2 m−1 compared to 1.8 ± 1.8 m−1 for the sample incubated under the same conditions without surfactant. We hypothesize that the surfactant alcohols can scavenge sulfate radicals via hydrogen abstraction, thereby reducing the degree of polymer degradation [42]. However, it is also possible that the surfactant interacts directly with the PAM; this requires further investigation.
3.4. Correlation between fouling index and polymer size and concentration The results from all of the fouling experiments using different raw and synthetic flowback fracturing fluids are summarized in Fig. 4. The left panel shows the fouling index as a function of the polymer peak MW (determined by SEC) while the right panel shows the same data plotted as a function of the concentration of high MW components in the water, which was determined from the area of the SEC peak above a critical size / MW. The critical size was chosen as 130 nm, corresponding to a MW > 330 kDa (based on the Fox-Flory equation [43]); this cutoff was chosen because it is only slightly smaller than the membrane pore size (200 nm) and it corresponds to the MW of one of the polyacrylamide standards that was used for calibration in the SEC analysis. Fig. 4A shows a large increase in fouling index with increasing polymer size over a nearly 4 order of magnitude range. This result is
Fig. 4. Fouling index of flowback water increases with (A) the peak molecular weight and (B) the concentration of high MW species (> 130 nm in diameter). Data from repeat experiments are presented using the same symbol. Black square indicates raw fluid (only FR without reaction) and the remaining data are synthetic flowback. FR: only FR; FB: FR plus breaker; FBC: FB plus crosslinker. The open circles (FR) and triangles (FB and FBC) are generated by reaction with shale at 83 bar, 80 ℃, 25 g/L. The diamonds are flowback of FR only generated using different reaction conditions: open diamond, FR-S + P + T: without shale (at 83 bar and 80 ℃); center dotted diamond, FR + S + P - T: room temperature (at 83 bar with shale); and center cross diamond, FR-S + P + T: atmospheric pressure (at 80 ℃ with shale, diamond with low fouling index). Overall fouling index of raw fracturing fluids and synthetic flowback waters can be roughly divided into three regions: a) green oval is low fouling region (≤ m−1) where less than 20% fouling (final flux over initial flux) occurs during the course of filtration (100–300 L/m2); b) the red oval is medium fouling region (20–200 m−1), corresponding to 50–95% fouling during the course of filtration (50–300 L/m2); c) the grey oval is high fouling region (> 1000 m−1), in which the flux declines to nearly zero within 5 L/m2. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) 129
Journal of Membrane Science 560 (2018) 125–131
B. Xiong et al.
HVHF wastewater treatment. Low molecular weight volatile organics that haven been extensively analyzed previously [2,19,21,45] do not directly cause initial flux decline. The fouling behavior observed with the synthetic flowback water containing degraded PAM is similar to that observed previously with high fouling flowback water obtained from actual Marcellus Shale wastewater [14]. Our results clearly demonstrate that the fracturing fluid recipe and the interactions between the PAM and shale under HVHF conditions directly affect the degree of polymer degradation and in turn the fouling behavior. Our results provide one possible explanation for the large variation in fouling index of different flowback and produced water samples collected from the Marcellus shale. Reservoirs at high temperatures (deep) and with highly reactive shale mineralogy are likely to generate a flowback water with low fouling tendency due to the significant degradation of the PAM. Furthermore, the addition of breaker significantly reduced the PAM size and its fouling index; however, breaker is currently used in only ~10% of slickwater fracturing operations. We also showed that microfiltration membranes with pores as small as 0.1 µm are unable to reduce the effective size of the PAM, suggesting that other methods of pretreatment such as coagulation or oxidation should be considered in place of simple prefiltration. These data suggest membrane treatability of flowback and produced waters can be enhanced by: 1) reducing the use or concentration of PAM or replacement of PAM with an alternative drag reduction polymer in HVHF operations; 2) increasing the dosage of oxidative breaker such as persulfate or peroxide during HVHF operations or possibly the addition of this readily available oxidant in a pretreatment step to reduce the PAM MW prior to membrane filtration; 3) reducing the use of crosslinker and surfactant; 4) utilizing drilling waste that contains reactive shale as oxidative treatment media to enhance degradation of PAM, or 5) exploring opportunities for PAM removal by coagulation.
3.5. The effect of prefiltration on membrane fouling To better understand the fouling characteristics of the degraded polyacrylamide, fouling experiments were performed with permeate samples obtained by pre-filtration of the synthetic flowback water through different polycarbonate track-etched membranes with uniform pores of different pore size. The prefiltration was performed using approximately 0.1 L of synthetic flowback water where 90% of the feed was collected as permeate. The polymer in the initial flowback water had a peak MW of 1500 kDa (hydrodynamic size of 350 nm), which was generated by incubation of 1.5% FR with 0.0001 g/L (0.0004 mM) ammonium persulfate with no shale at 80 ℃ for 24 h. As shown in Fig. 5A, solutions that were prefiltered through the 0.4 and 0.8 μm membranes had flux decline curves that were similar to that of the unfiltered solution, with the fouling index ranging from 90 to 110 m−1. In contrast, prefiltration through the 0.2 μm membrane reduced the fouling index to 43 m−1 while prefiltration through the 0.1 μm membrane reduced the fouling index to 13 m-1, with the normalized flux declining by less than 20% over the course of filtration of 100 L/m2. SEC chromatographs of the prefiltered solutions are shown in Fig. 5B. Prefiltration of the flowback water through the 0.2 and 0.1 μm membranes removed 50% and 97% of the polymer, respectively, consistent with the large reduction in fouling seen in Fig. 5A. Much of the polymer removal may have occurred through the heavily fouled polycarbonate membranes (Figs. 5C and 5D), with the fouling layer providing significant removal of PAM across the entire MW spectrum. Surprisingly, the chromatographs suggest that prefiltration through the various pore size membranes had little effect on the peak MW even when the prefiltration was performed using a membrane with 0.1 µm pore size. We speculate that the highly flexible linear polymer chains could stretch under the elongation flow field at the pore entrance and be able to pass through pores that are much smaller than the polymer hydrodynamic size (as determined by SEC). Similarly enhanced transmission has been observed during the ultrafiltration and microfiltration of linear DNA molecules [44]. This suggests that commonly used 0.45 or 0.22 μm microfiltration membranes might not be sufficient to selectively remove large polymer foulants, making it difficult if not impossible to use this type of membrane pretreatment to control fouling in subsequent treatment steps.
Acknowledgement This research was funded by a Penn State College of Engineering Innovation Grant and a seed grant through the Center for Collaborative Research in Intelligent Gas Systems (CCRINGS) program funded by General Electric (GE). Additional funding was provided by the Pennsylvania Water Resources Research Center Small grants program. The authors would like to thank Weatherford Inc. for providing the synthetic chemical additives. The authors acknowledge the Kappe Environmental Engineering laboratories for TOC measurement instrumentation and technical assistance by David Jones. The authors also thank Rajarshi Guha for discussions of the membrane fouling experiments.
4. Conclusions To the best of our knowledge, this work demonstrates, for the first time, that both PAM and its degradation products generated under HVHF conditions are the primary cause of membrane fouling during
Fig. 5. (A) Flux decline of 0.2 μm PVDF membranes after prefiltration through polycarbonate membranes with pore size of 0.1, 0.2, 0.4, or 0.8 μm . (B) Chromatographs of polymer solutions after prefiltration. (C) and (D): Scanning electron microscopy images of polycarbonate track-etched membranes after filtration of the flowback water with pore size of 0.8 (top) and 0.2 μm (bottom) showed large aggregates as well as direct pore blocking (arrow). All samples were filtered by 0.45 μm syringe filters before SEC analyses. Prefiltration using 0.1, 0.2, 0.4 μm membranes (Whatman, Nuclepore™) was conducted at a constant pressure of 10 psi; prefiltration with 0.8 μm membranes (Millipore, Isopore™) was conducted at a constant pressure of 4 psi. The unfiltered solution was generated after incubating 0.15% FR fracturing fluid with 0.0001 g/L persulfate at 80 ℃ for 24 h, yielding a polymer solution with a peak MW of 1500 kDa (unfiltered). 130
Journal of Membrane Science 560 (2018) 125–131
B. Xiong et al.
Supporting information
surfactants, Anal. Chem. 86 (2014) 9653–9661. [20] G.T. Llewellyn, F. Dorman, J.L. Westland, D. Yoxtheimer, P. Grieve, T. Sowers, E. Humston-Fulmer, S.L. Brantley, Evaluating a groundwater supply contamination incident attributed to Marcellus Shale gas development, Proc. Natl. Acad. Sci. 112 (2015) 6325–6330. [21] B.D. Drollette, K. Hoelzer, N.R. Warner, T.H. Darrah, O. Karatum, M.P. O’Connor, R.K. Nelson, L.A. Fernandez, C.M. Reddy, A. Vengosh, R.B. Jackson, M. Elsner, D.L. Plata, Elevated levels of diesel range organic compounds in groundwater near Marcellus gas operations are derived from surface activities, Proc. Natl. Acad. Sci. 112 (2015) 13184–13189. [22] A. Butkovskyi, H. Bruning, S.A. Kools, H.H. Rijnaarts, A.P. van Wezel, Organic pollutants in shale gas flowback and produced waters: identification, potential ecological impact and implications for treatment strategies, Environ. Sci. Technol. (2017). [23] J. Rosenblum, E.M. Thurman, I. Ferrer, G. Aiken, K.G. Linden, Organic chemical characterization and mass balance of a hydraulically fractured well: from fracturing fluid to produced water over 405 days, Environ. Sci. Technol. 51 (2017) 14006–14015. [24] X. Wang, S. Chen, W. Dong, Z. Wang, L. Yang, X. Xi, Q. Zhang, L. Shi, Contribution of main pollutants in oilfield polymer-flooding wastewater to the total membrane fouling resistance, Sep. Sci. Technol. 47 (2012) 1617–1627. [25] S. Wang, C. Liu, Q. Li, Fouling of microfiltration membranes by organic polymer coagulants and flocculants: controlling factors and mechanisms, Water Res. 45 (2011) 357–365. [26] G. Liu, S. Yu, H. Yang, J. Hu, Y. Zhang, B. He, L. Li, Z. Liu, Molecular mechanisms of ultrafiltration membrane fouling in polymer-flooding wastewater treatment: role of ions in polymeric fouling, Environ. Sci. Technol. 50 (2016) 1393. [27] B. Xiong, Z. Miller, S. Roman-White, T. Tasker, B. Farina, B. Piechowicz, W.D. Burgos, P. Joshi, L. Zhu, C.A. Gorski, A.L. Zydney, M. Kumar, Chemical degradation of polyacrylamide during hydraulic fracturing, Environ. Sci. Technol. 52 (2018) 327–336. [28] T.L. Tasker, P.K. Piotrowski, F.L. Dorman, W.D. Burgos, Metal associations in Marcellus shale and fate of synthetic hydraulic fracturing fluids reacted at high pressure and temperature, Environ. Eng. Sci. (2016). [29] G.A. Kahrilas, J. Blotevogel, E.R. Corrin, T. Borch, Downhole transformation of the hydraulic fracturing fluid biocide glutaraldehyde: implications for flowback and produced water quality, Environ. Sci. Technol. 50 (2016) 11414–11423. [30] J. Gao, J. Yu, W. Wang, T. Lin, The accelerated degradation of aqueous polyacrylamide at low temperature, J. Appl. Polym. Sci. 69 (1998) 791–797. [31] H. Liu, T.A. Bruton, F.M. Doyle, D.L. Sedlak, In situ chemical oxidation of contaminated groundwater by persulfate: decomposition by Fe (III)-and Mn (IV)-containing oxides and aquifer materials, Environ. Sci. Technol. 48 (2014) 10330–10336. [32] P. Yin, Z. Hu, X. Song, J. Liu, N. Lin, Activated persulfate oxidation of perfluorooctanoic acid (PFOA) in groundwater under acidic conditions, Int. J. Environ. Res. Public Health 13 (2016) 602. [33] H. Sun, R.F. Stevens, J.L. Cutler, B. Wood, R.S. Wheeler, Q. Qu, A novel nondamaging friction reducer: development and successful slickwater frac applications, in: Tight Gas Completions Conference, in Society of Petroleum Engineers, 2010. [34] P.S. Carman, K. Cawiezel, Successful breaker optimization for polyacrylamide friction reducers used in slickwater fracturing, in: Society of Petroleum Engineers, 2007. [35] B.L. Gall, A.R. Sattler, D.R. Maloney, C.J. Raible, Permeability damage to natural fractures caused by fracturing fluid polymers, in Society of Petroleum Engineers, 1988. [36] R.L. Thomas, J.E. Brown, The impact of fracturing fluids on conductivity and performance in low-temperature wells, in: Society of Petroleum Engineers, 1989. [37] G.E. King, Hydraulic fracturing 101: what every representative, environmentalist, regulator, reporter, investor, university researcher, neighbor and engineer should know about estimating frac risk and improving frac performance in unconventional gas and oil wells, in Society of Petroleum Engineers, 2012. [38] C. Aften, W.P. Watson, Improved friction reducer for hydraulic fracturing, in Society of Petroleum Engineers, 2009. [39] What Chemicals Are Used, in, FracFocus Chemical Disclosure Registry. [40] B.V. Mork, J.M. Gravitt, Enhanced chemical oxidation, U.S. Patent Application 20140097381 A1, 2014. [41] D.B. Bennion, F.B. Thomas, R.F. Bietz, D.W. Bennion, Water and hydrocarbon phase trapping in porous media-diagnosis, prevention and treatment, J. Can. Pet. Technol. 35 (1996) 29–36. [42] G.P. Anipsitakis, D.D. Dionysiou, Radical generation by the interaction of transition metals with common oxidants, Environ. Sci. Technol. 38 (2004) 3705–3712. [43] M. Rubinstein, R.H. Colby, Polymer Physics, Oxford University, NEW YORK, 2003. [44] Y. Li, E.E. Borujeni, A.L. Zydney, Use of preconditioning to control membrane fouling and enhance performance during ultrafiltration of plasmid DNA, J. Membr. Sci. 479 (2015) 117–122. [45] S.J. Maguire-Boyle, A.R. Barron, Organic compounds in produced waters from shale gas wells, Environ. Sci.: Process. Impacts 16 (2014) 2237–2248.
The supporting information includes a detailed description of the calculation of fouling index and polymer MW and size distribution. It also includes additional data for raw fluids and synthetic flowback with FR with other additives. The Supporting Information is available free of charge on the ACS Publications website at http://dx.doi.org/10.1016/j. memsci.2018.04.055. Notes The authors declare no competing financial interest. Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.memsci.2018.04.055. References [1] N. Abualfaraj, P.L. Gurian, M.S. Olson, Characterization of Marcellus shale flowback water, Environ. Eng. Sci. 31 (2014) 514–524. [2] W. Orem, C. Tatu, M. Varonka, H. Lerch, A. Bates, M. Engle, L. Crosby, J. McIntosh, Organic substances in produced and formation water from unconventional natural gas extraction in coal and shale, Int. J. Coal Geol. 126 (2014) 20–31. [3] C. He, T. Zhang, X. Zheng, Y. Li, R.D. Vidic, Management of Marcellus shale produced water in Pennsylvania: a review of current strategies and perspectives, Energy Technol. 2 (2014) 968–976. [4] W.L. Ellsworth, Injection-induced earthquakes, Science 341 (2013). [5] Eureka-Resources, Standing Stone Facility, in, 2014. [6] A. Butkovskyi, H. Bruning, S.A.E. Kools, H.H.M. Rijnaarts, A.P. Van Wezel, Organic pollutants in shale gas flowback and produced waters: identification, potential ecological impact, and implications for treatment strategies, Environ. Sci. Technol. 51 (2017) 4740–4754. [7] C. He, X. Wang, W. Liu, E. Barbot, R.D. Vidic, Microfiltration in recycling of Marcellus Shale flowback water: solids removal and potential fouling of polymeric microfiltration membranes, J. Membr. Sci. 462 (2014) 88–95. [8] Q. Jiang, J. Rentschler, R. Perrone, K. Liu, Application of ceramic membrane and ion-exchange for the treatment of the flowback water from Marcellus shale gas production, J. Membr. Sci. 431 (2013) 55–61. [9] S.M. Riley, J.M.S. Oliveira, J. Regnery, T.Y. Cath, Hybrid membrane bio-systems for sustainable treatment of oil and gas produced water and fracturing flowback water, Sep. Purif. Technol. 171 (2016) 297–311. [10] S. Mondal, S.R. Wickramasinghe, Produced water treatment by nanofiltration and reverse osmosis membranes, J. Membr. Sci. 322 (2008) 162–170. [11] B.D. Coday, P. Xu, E.G. Beaudry, J. Herron, K. Lampi, N.T. Hancock, T.Y. Cath, The sweet spot of forward osmosis: treatment of produced water, drilling wastewater, and other complex and difficult liquid streams, Desalination 333 (2014) 23–35. [12] H.C. Duong, A.R. Chivas, B. Nelemans, M. Duke, S. Gray, T.Y. Cath, L.D. Nghiem, Treatment of RO brine from CSG produced water by spiral-wound air gap membrane distillation — a pilot study, Desalination 366 (2015) 121–129. [13] S.M. Riley, D.C. Ahoor, J. Regnery, T.Y. Cath, Tracking oil and gas wastewaterderived organic matter in a hybrid biofilter membrane treatment system: a multianalytical approach, Sci. Total Environ. 613 (2018) 208–217. [14] B. Xiong, A.L. Zydney, M. Kumar, Fouling of microfiltration membranes by flowback and produced waters from the Marcellus shale gas play, Water Res. 99 (2016) 162–170. [15] G. Chen, Z. Wang, L.D. Nghiem, X.-M. Li, M. Xie, B. Zhao, M. Zhang, J. Song, T. He, Treatment of shale gas drilling flowback fluids (SGDFs) by forward osmosis: membrane fouling and mitigation, Desalination 366 (2015) 113–120. [16] S. Alzahrani, A.W. Mohammad, Challenges and trends in membrane technology implementation for produced water treatment: a review, J. Water Process Eng. 4 (2014) 107–133. [17] M. Elsner, K. Hoelzer, Quantitative survey and structural classification of hydraulic fracturing chemicals reported in unconventional gas production, Environ. Sci. Technol. 50 (2016) 3290–3314. [18] K. Hoelzer, A.J. Sumner, O. Karatum, R.K. Nelson, B.D. Drollette, M.P. O’Connor, E.L. D’Ambro, G.J. Getzinger, P.L. Ferguson, C.M. Reddy, M. Elsner, D.L. Plata, Indications of transformation troducts from hydraulic fracturing additives in shalegas wastewater, Environ. Sci. Technol. 50 (2016) 8036–8048. [19] E.M. Thurman, I. Ferrer, J. Blotevogel, T. Borch, Analysis of hydraulic fracturing flowback and produced waters using accurate mass: identification of ethoxylated
131
Contents lists available at ScienceDirect
Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci
Polyacrylamide in hydraulic fracturing fluid causes severe membrane fouling during flowback water treatment
T
Boya Xionga, Selina Roman-Whiteb, Bethany Piechowiczb, Zachary Millerb, Benjamin Farinab, ⁎ ⁎⁎ Travis Taskera, William Burgosa, Andrew L. Zydneyb, , Manish Kumara,b, a b
Department of Civil and Environmental Engineering, The Pennsylvania State University, Greenberg Building, University Park, PA 16802, United States Department of Chemical Engineering, The Pennsylvania State University, Greenberg Building, University Park, PA 16802, United States
A B S T R A C T
Sustainable wastewater management strategies are required to further minimize impacts of high-volume hydraulic fracturing (HVHF) as current practices such as reuse or direct disposal have long term limitations. Membranes can provide superior effluent quality in HVHF wastewater treatment, but the application of these systems is severely limited by membrane fouling. However, the key fouling components in HVHF wastewater have not yet been clearly identified and characterized. Here we demonstrate that fouling of microfiltration membranes by synthetic flowback water is mostly due to polyacrylamide (PAM), a major additive in slickwater fracturing fluids. A synthetic fracturing fluid was incubated with Marcellus Shale under HVHF conditions (80 ℃, 83 bar, 24 h) to generate synthetic flowback water. Different HVHF conditions and fracturing fluid compositions generated a fouling index for flowback water ranging from 0.1 to 2000 m−1, with these values well correlated with the peak molecular weight (MW) (ranging from 104 kDa) and the concentration of high MW components in the water. The lowest fouling index was 10 to 1.5 × observed when PAM was further degraded by ammonium persulfate under HVHF conditions, although this is infrequently used with PAM in current fracturing operations. These results highlight the importance of PAM and its degradation products in fouling of subsequent membrane systems, providing insights that can help in the development of effective treatment processes for HVHF wastewater.
1. Introduction In the past decade, the development of unconventional oil and gas using high-volume hydraulic fracturing (HVHF) has had a significant impact on the U.S. energy landscape. However, the environmental impacts of HVHF have generated tremendous social concerns, largely due to the production of large quantities of wastewater, referred to as flowback and produced water, that contain high salinity (220,000–340,000 mg/L), turbidity, organic constituents (1–5500 mg/L total organic carbon), and radioactivity (gross alpha 50–120,000 pCi/L) [1,2]. In some states, such as Texas, wastewater is primarily disposed of by deep well injection due to the ready availability of class II disposal wells. In Marcellus Shale gas wells in Pennsylvania, 90% of the wastewater (both flowback and produced water) is recycled and reused in subsequent HVHF operations because of the limited number of disposal wells [3]. The influence of such recycling after minimal treatment, particularly the effect of transformed organic additives, on well productivity has not been reported. More importantly, deep well injection
⁎
is not an environmentally friendly management strategy [4]; the opportunity for recycling will be limited when the number of new fracturing jobs declines. There is a clear need for developing processes that can provide cost-effective treatment of flowback and produced water. Membrane systems have been proposed by a number of investigators for treatment of flowback water [5,6], both to remove suspended solids and organics using microfiltration (MF) or ultrafiltration (UF) [7,8] and to remove high salinity/hardness using nanofiltration (NF) [9], reverse osmosis (RO) [10] and forward osmosis [11] or membrane distillation [12]. However, membrane fouling by hydrocarbon [10] and polymeric organics [9,13,14], inorganic scaling species [15], microbial [9] and particulate materials [7,14] in flowback water remains a challenge in efficiently treating the wastewater [16]. Our previous work with flowback water from Marcellus Shale gas wells showed a large variation in both water quality and fouling rates during microfiltration with no apparent correlation between the measured fouling index and either the total organic content or turbidity [14]. The exact nature of the key fouling constituents in these wastewaters remains unknown. This knowledge gap makes it
Corresponding author. Corresponding author at: Department of Chemical Engineering, The Pennsylvania State University, Greenberg Building, University Park, PA 16802, United States. E-mail addresses: [email protected] (A.L. Zydney), [email protected] (M. Kumar).
⁎⁎
https://doi.org/10.1016/j.memsci.2018.04.055 Received 28 February 2018; Received in revised form 30 April 2018; Accepted 30 April 2018 Available online 22 May 2018 0376-7388/ © 2018 Elsevier B.V. All rights reserved.
Journal of Membrane Science 560 (2018) 125–131
B. Xiong et al.
fracturing fluid and dissolved Fe2+ from shale at low pH [27]. There is a critical need to evaluate the fouling of degraded PAM characteristic of the materials in flowback / produced water from HVHF operations. The primary objectives of this work were to: 1) quantify the effect of downhole pressure and temperature on fouling indices of PAM compared to ‘raw’ fracturing fluid and flowback water; 2) evaluate the effect of combinations of PAM and other additives on the fouling index; and 3) examine the correlation between the fouling index and PAM size and concentration. These results aim to provide new insights into the fouling characteristics of flowback water, while also identifying possible strategies to reduce membrane fouling. The data also suggest possible concerns regarding plugging of the micro/nano-scale pores in shale formations (similar in size to those in porous membranes) during reuse of flowback water.
Nomenclature HVHF FR PAM MF UF MW SEC
High volume hydraulic fracturing Friction reducer Polyacrylamide Microfiltration Ultrafiltration Molecular weight Size exclusion chromatography
difficult to design treatment processes that are effective for managing fracturing wastewaters. The organics in flowback water come from both the chemical additives used in the hydraulic fracturing fluid and the hydrocarbons extracted from the shale. Over 1000 chemicals [17], some of which are proprietary, have been used in hydraulic fracturing, and many of these have been detected in flowback and produced water [18,19] as well as in contaminated surface water / groundwater samples [20–22]. Many of these analyses utilized advanced gas chromatography coupled with mass spectroscopy, which can identify hydrophobic and volatile organics such as hydrocarbons; however, components that are hydrophilic and larger than 1 kDa, such as polyacrylamide (PAM) used as a friction reducer (FR) and guar gum used as a gelling agent, cannot be detected in these analyses [21]. Two very recent studies found that 90% of the organic matter in fracturing wastewater is hydrophilic, and some wastewaters have been shown to contain 20–40% biopolymers [13,23]. These polymers, including high MW PAM, can be significant membrane foulants. For example, PAM was found to be the major contributor to the total membrane resistance in fouling tests performed with synthetic oil-field polymer flooding wastewater [24]. Wang et al. [25] reported that PAM used as a coagulant caused MF fouling predominantly by surface pore blockage, with the rate of fouling determined by the molecular weight and concentration of the polymer solution. Liu et al. [26] utilized atomic force microscopy to correlate intermolecular forces with the fouling resistance provided by hydrolyzed polyacrylamide during filtration through a polyvinylidenefluoride UF membrane. This study concluded that fouling was dominated by the resistance from the concentration polarization layer formed by intermolecular attraction, rather than the gel layer formed by polymer-membrane attraction, where both attractive forces are attributed to hydrophobic interactions involving the polymer backbone in combination with hydrogen bonding. However, it is not possible to directly extrapolate from these studies to the fouling characteristics of PAM in flowback water, particularly given the complex transformation of PAM that can occur under HVHF conditions due to interactions with the solid shale and with other additives present in the fracturing fluid. Recent work in our laboratory has demonstrated that PAM undergoes significant degradation via a free radical mechanism at the high temperatures, where the free radicals are generated by reactions involved dissolved oxygen present in the
2. Materials and methods 2.1. Synthetic fracturing fluid Stock chemicals of friction reducer (FR), biocide, corrosion inhibitor, crosslinker, and surfactant were provided by Weatherford Chemical, Inc. and are commercial products used in HVHF operations. Each additive contains a mixture of chemicals as provided by the supplier (see Table SI); the exact composition and concentration of individual components are proprietary. FR contains PAM and petroleum distillate. Vacuum incubation of a neat FR stock solution yielded a nonvolatile (polymer) portion of approximately 40% by mass. A previous study reported that raw FR fluid contained roughly 0.7 g/L PAM with a peak MW of 15 MDa based on size exclusion chromatography analysis [27]. Ammonium persulfate (breaker), citric acid (iron control), and potassium hydroxide and sulfuric acid (pH adjustment) were purchased from Sigma-Aldrich (St. Louis, MO) and were prepared using deionized (DI) water from a Barnstead Nanopure water purification system with a resistivity of > 18 MΩ cm. This work primarily considers a slickwater fracturing fluid, which mainly contains FR, biocide, corrosion inhibitor, surfactant and iron control agents, given that 84% fracturing operations in the Marcellus utilized slickwater frac based on a review of 100 drilling logs and 97% nationwide utilized FR based on a review of 750 drilling logs on FracFocs.org. Hybrid frac combines slickwater and gel frac, where both FR and a gelling agent (such as guar gum) are used. We also performed limited experiments simulating a hybrid fracturing fluid by adding breaker and crosslinker to the fluid; critically, gelling agent could be another significant source that contributes to membrane fouling, which will be discussed in detail in a separate paper. Gel frac without the use of FR is rare and thus was not considered in the current study. The concentrations of the fracturing additives used to prepare the synthetic fracturing fluid (Table 1) were based on FracFocus.org and literature sources [17,28]. Additive concentrations tend towards the high ranges reported in the literature. Synthetic fracturing fluids were prepared by mixing appropriate additives with DI water. The
Table 1 Chemicals used to prepare the synthetic fracturing fluid. Chemical additives
Specific compounds (provided by supplier)
Concentration
Friction Reducera Biocidea Corrosion Inhibitora
Polyacrylamide, petroleum distillate 2,2-Dibromo- 3-Nitrilopropionamide (DBNPA) Isopropanol, Ethylene glycol, N,N-Dimethylformamide, 2-Butoxyethanol, Cinnamaldehyde, Tar Bases, 1-Decanol,1-Octanol, Triethyl phosphate Sodium tetraborate pentahydrate，Glycerine, Potassium hydroxide Ammonium Persulfate Citric Acid Ethoxylated alcohol (C6-C12), Ethylene Glycol, Isopropyl Alcohol, D-limonene, 1-Octanol Potassium hydroxide and sulfuric acid
0.15% v/v 0.0017% v/v 0.0007% v/v
Crosslinkera Breaker Iron Control Surfactanta pH Adjustment a
0.03% v/v 0.011% w/v 0.0014% w/v 0.075% v/v pH adjusted to 7.2
indicates chemical is provided by Weatherford Chemical, Inc. Final concentrations are based on the volume fraction of the liquid chemical stock (e.g., % volume of chemical stock/volume of water). 126
Journal of Membrane Science 560 (2018) 125–131
B. Xiong et al.
2.4. Polymer size and concentration analysis by size exclusion chromatography
mixtures were blended for 30 s and the pH was then adjusted to 7.2 using potassium hydroxide or sulfuric acid as needed.
Molecular weight (MW) distributions of PAM in the fracturing fluid and flowback water were evaluated by conventional size exclusion chromatography (SEC) using a Waters HPLC 1515 system and a Shodex SB-806M HQ aqueous SEC column with 0.05 M Na2SO4 as the mobile phase at a flow rate of 0.5 mL/min. All samples were pre-filtered through 0.45 μm cellulose acetate syringe filters before injection (100 µL sample volume), with peak positions and concentrations determined by a refractive index detector. Nonionic PAM standards (American Polymer Standard Corporation, Mentor, OH) with a concentration of 0.5 mg/mL dissolved in 0.05 M Na2SO4 were used for calibration. Polymer concentrations were determined using the dn/dc coefficient determined experimentally as described in SI-Fig. 1 and Table SI−1.
2.2. Synthetic flowback water Synthetic flowback water was obtained by reacting the synthetic fracturing fluid with 25 g/L shale at 80 ℃ and 83 bar for 24 h simulating downhole conditions encountered in natural gas reservoirs. The variation in these conditions across different gas formations can be large [29], and its impact on polyacrylamide degradation was investigated previously. Marcellus Shale was collected from an outcrop near Frankstown, PA; it had a 6-meter overburden and had been exposed to the atmosphere. The shale sample was pulverized and sieved to a size of 0.3–2 mm using a method described previously [27]. Pulverized shale was mixed with freshly prepared fracturing fluid and was incubated in a 0.5 L T316 stainless steel Parr reactor with digital temperature control and nitrogen gas pressurization. After reaction, the resulting suspension was centrifuged at 13,700 g for 1 h at 4 ℃, with the supernatant collected as the synthetic flowback water. Filtration experiments were conducted within 24 h after centrifugation.
3. Results and discussion 3.1. Polyacrylamide used in FR is the dominant foulant among all the additives The friction reducer (FR) is one of the most important ingredients in slickwater HVHF and can potentially contribute to a large proportion of membrane fouling due to its large size. Filtration experiments with ‘raw’ unreacted combinations of the other additives shown in Table 1 (without FR), including breaker, crosslinker, surfactant, biocide, corrosion inhibitor and iron control agent, showed relatively little fouling, with the flux declining by less than 40% after a volumetric throughput of more than 100 L/m2 (Fig. 1A). Incubation of these additives (without FR) at HVHF conditions for 24 h (in the presence of 25 g/L shale, 80 ℃, and 83 bar), yielded a synthetic flowback water that also showed negligible fouling up to volumetric throughputs of 300 L/m2 (Fig. 1B). In contrast, the ‘raw’ unreacted FR showed a very rapid flux decline, with the flux decreasing to less than 0.05% of the initial value within 2 L/m2, corresponding to a fouling index of 1400 m−1. Incubation of the FR under HVHF conditions reduced the fouling index to 73 m−1, but this still corresponded to a flux decline of nearly 90% after 30 L/m2 of filtration. These values of the fouling index are similar to those reported previously for actual flowback water collected from the Marcellus shale play [14] (symbols labeled S1, S2, and S3 in Fig. 1B). Mixing FR with other additives resulted in a flowback water with lower fouling potential compared to only FR, which will be discussed in a later section. SEC analysis demonstrated that the molecular weight of PAM in FR
2.3. Membrane fouling experiments Fouling experiments were conducted using 0.2 μm pore size polyvinylidenefluoride (PVDF, EMD Millipore) membranes in a 10 mL Amicon 8010 dead-end filtration cell at a constant pressure of 0.27 bar (4 psi) at room temperature based on results from previous studies of MF fouling using actual flowback water from the Marcellus region [14]. Details of the filtration experiments were described previously [14]. The fouling rate was quantified using the initial membrane fouling index (m−1):
Fouling index = −
1 dJ J0 d V
() A
(1)
where J is the filtrate flux (L/m2 h), V/A is the specific throughput (L/m2), and J0 is the initial flux. The derivative was calculated using first order derivative formulas developed by a finite difference method accounting for the variable time steps [14]. This method does not rely on any assumptions about fouling mechanisms while providing a quantitative estimate of the initial fouling rate. Details on the calculations are presented in the SI.
Fig. 1. PAM in friction reducer (FR) is the major fouling component compared to other additives. Flux decline during microfiltration of: (A) Raw (unreacted) FR compared to other additives (no FR); (B) Flowback containing FR or other additives. Synthetic flowback was generated from incubation of raw fluid at 80 ℃, 83 bar with 25 g/L shale; S1-S3 are actual flowback water samples collected from the Marcellus shale play [14]. (C) Size exclusion chromatographs of raw (black solid) and flowback (red dash) of FR; and raw (dash blue) and flowback (solid green) of other additives (called “other”). Blue and green lines are completely overlapped. All FR fluids contain FR at a concentration of 0.15% v/v. Fouling experiments were conducted using 0.2 μm PVDF microfiltration membranes in a dead-end filtration cell at 0.27 bar. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) 127
Journal of Membrane Science 560 (2018) 125–131
B. Xiong et al.
Gel fracturing operations utilize a crosslinked polymer gel such as guar gum and its derivatives to enhance fluid viscosity in order to more effectively carry proppant throughout the fractures. During flowback, the viscous gel can impair fracture conductivity and hinder gas flow; thus, oxidative breakers such as peroxide and persulfate are commonly used to break down the polymer gel and enhance flowback recovery [35,36]. In slickwater fracturing [37], breaker is occasionally used in part due to recent concerns over potential formation damage by accumulated PAM [34,38]. Our review of 100 drilling logs (randomly selected) in the Marcellus play from 2015 to 2017 (FracFocus.org [39]) indicate that breaker was only used when polyacrylamide was mixed with gel fluid (giving a so-called ‘hybrid’ fracturing fluid); this was 16% of the total HVHF jobs. Thus most of the fracturing jobs utilizing only FR do not experience additional degradation by breaker. Among all the remaining additives, the presence of crosslinker and surfactant slightly increased the resulting MW distribution and fouling of the flowback water. The MW distribution and fouling of flowback with all the ingredients (FR+other) and FR with breaker, crosslinker, and surfactant were found to be similar (Fig. SI-4), suggesting that the remaining additives (i.e. iron control agent and biocide) had no measureable effect on flowback chemistry. The sample incubated at HVHF conditions with crosslinker (sodium tetraborate pentahydrate, 0.03% v/v) had a peak MW of 15 ± 3 kDa (blue line in Fig. 2B), which is slightly larger than the peak MW for the sample incubated without the crosslinker (8 ± 3 kDa). The sample incubated with crosslinker also had a slightly larger fouling index of 1.8 ± 1.8 m−1 compared to < 1 m−1 for the sample without crosslinker. The crosslinking chemistry associated with borate and PAM has not been previously studied. The lower degree of polymer degradation observed in the presence of the crosslinker may be the result of enhanced radical termination [30] caused by borate enhanced persulfate activation [40]. Surfactant is used in oil and gas extraction to reduce the interfacial tension between injected fluids and the formation which can enhance water recovery [41]. Limited experiments were thus performed after addition of surfactant (ethoxylated alcohol, 1-octanol and D-limonene [28], 0.075% v/v). The peak MW of the sample incubated with surfactant under HVHF conditions was 120 ± 7 kDa (labeled as FBCS in SIFig. 4), which is higher than the sample without surfactant
is much larger than that for any of the other additives as expected. Fig. 1C shows the raw chromatographs for the different samples examined in Figs. 1A and 1B; the largest components have the smallest elution volume since they are strongly excluded from the resin pores. The calculation of the peak MW from the SEC data is discussed in the SI. The PAM in FR had a peak MW as high as 1.5 × 104 kDa while the other additives had molecular weights less than 1 kDa (no peak shown on chromatograph of SEC). However, incubation of the FR for 24 h under HVHF conditions, leading to what we refer to as “FR flowback”, caused a reduction in the peak MW to approximately 200 kDa. This is discussed in more detail in the next section. 3.2. Reaction under HVHF conditions significantly reduced fouling index of FR SEC analyses suggest molecular weight reduction of PAM in FR occurred during HVHF reaction and that this is directly related to the reduced rate of fouling. In a previous study, we demonstrated that exposure to Marcellus Shale promoted both adsorption and a free radicalinduced chemical degradation of PAM, leading to a significant reduction in MW [27]. It was also found that elevated temperature and the presence of shale were the two critical conditions causing significant polymer degradation; in comparison, high pressure and salinity played a minor role. Additional filtration experiments were conducted using only FR incubated at room temperature (with shale under 83 bar pressure) and without shale (at 80 ℃ under 83 bar pressure). As shown in Fig. 2 (open circles and inverted triangles), these samples showed much greater rates of fouling than the FR exposed to shale at 80 ℃ (filled triangles and squares). In contrast, reaction at 1 bar generated a flowback water with very similar fouling behavior as that for the flowback water generated at 83 bar reaction pressure. These differences are also seen in the peak MW which were reported previously: the samples exposed to shale at 1 and 83 bar had a peak MW of 200–300 kDa while those treated at room temperature had peak MW of 4000 kDa. The small reduction in MW for the sample treated at room temperature was caused by adsorption of PAM to the shale; samples treated in the absence of shale had peak MW of 5000 kDa [27]. Similarly, salinity (3 M NaCl or 1.5 M CaCl2) had little impact on either polymer degradation or fouling (Fig. SI-2). These results suggest that different reservoir conditions across different shale formations could lead to flowback and produced waters with very different fouling potentials due to changes in the degree of PAM degradation and adsorption; this will be discussed in more detail in the next section. 3.3. Persulfate breaker reduces the fouling index by enhancing polymer degradation The fluid that mixed all other additives with FR had identical molecular size distribution as only FR (before reaction), and yet generated a flowback with peak MW of 85 ± 5 kDa (Fig. SI-3) and a fouling index of 3.3 ± 3.0 m−1 (Fig. SI-4B). Experiments with different fluid compositions revealed the important role of persulfate breaker among all the other additives in causing further polymer degradation and mitigating membrane fouling. The addition of 0.11 g/L ammonium persulfate significantly reduced the peak MW of our FR flowback to as low as 8 ± 3 kDa (after 24 h incubation of FR in the presence of ammonium persulfate under HVHF conditions) (Fig. 3B), consistent with the large reduction in apparent viscosity observed by Gao et al. [30]. As a result, the synthetic flowback water caused only 20% flux decline within 300 L/m2 (Fig. 3A), corresponding to a fouling index below 1 m−1. Persulfate decomposes to sulfate radicals (SO4•-), particularly at high temperatures [31,32]:
S2 O82 −→2SO4• −
Fig. 2. The fouling potential decreases after treatment at HVHF conditions. Flux decline during microfiltration of: Control (black square): synthetic FR flowback water generated from incubation at 80 ℃, 83 bar with 25 g/L shale; RT (red circle): flowback generated at room temperature, 83 bar with 25 g/L shale; 1 bar (blue triangle): flowback generated at 80℃, 1 bar with 25 g/L shale; No shale (green inverted triangle): flowback generated at 80, 83 bar with 0 g/L shale; All fluids contain FR at a concentration of 0.15% v/v. Fouling experiments were conducted using 0.2 μm PVDF microfiltration membranes in a deadend filtration cell at 0.27 bar. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
(2)
Sulfate radicals can then cause chain scission of the PAM backbone via free radical degradation [30,31,33,34]. 128
Journal of Membrane Science 560 (2018) 125–131
B. Xiong et al.
Fig. 3. Fouling behavior and polymer size of friction reducer (FR) decreases when breaker and crosslinker are added. (A) Flux decline during microfiltration and (B) size exclusion chromatographs of FR synthetic flowback water generated from FR alone (FR, red), FR plus breaker (blue), FR plus breaker and crosslinker (green). All incubations of raw fluid were conducted at 80 ℃, 83 bar with 25 g/L shale. Fouling experiments were conducted using 0.2 μm PVDF microfiltration membranes in a dead-end filtration cell at 0.27 bar. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
consistent with previous findings showing a correlation between molecular weight and transmembrane pressure during microfiltration of both hydrolyzed polyacrylamides (with MW between 1.5 and 5000 kDa) and poly(diallyldimethylammonium) chloride (100–500 kDa) solutions [25]. Samples with peak MW < 20 kDa fall in a low fouling region (< 1 m−1). Significant fouling was observed when the peak MW started to exceed 50 kDa, corresponding to a hydrodynamic size of approximately 45 nm, which is still well below the 200 nm nominal pore size for the membrane. This may suggest internal pore fouling, although it is also possible that the fouling was due to pore blockage associated with the larger components in the tail end of the size distribution and / or the simultaneous blockage by multiple polymer chains. This was analyzed further by plotting the fouling index as a function of the concentration of the high MW components in the distribution (details described in SI). The data (Fig. 4B) show a high degree of correlation, providing further evidence that the high MW components are the key foulants in these HVHF fluids. The highest fouling was seen with the raw fluids and with the fluids reacted without shale or at room temperature. Incubation under HVHF conditions lowered the fouling index by more than two orders of magnitude, with the fouling index influenced by the composition of the fluid, particularly the presence of breaker and crosslinker.
(15 ± 3 kDa, blue line in Fig. 2B) but comparable to the MW of only FR flowback. This sample gave a fouling index of 3.3 ± 3.2 m−1 compared to 1.8 ± 1.8 m−1 for the sample incubated under the same conditions without surfactant. We hypothesize that the surfactant alcohols can scavenge sulfate radicals via hydrogen abstraction, thereby reducing the degree of polymer degradation [42]. However, it is also possible that the surfactant interacts directly with the PAM; this requires further investigation.
3.4. Correlation between fouling index and polymer size and concentration The results from all of the fouling experiments using different raw and synthetic flowback fracturing fluids are summarized in Fig. 4. The left panel shows the fouling index as a function of the polymer peak MW (determined by SEC) while the right panel shows the same data plotted as a function of the concentration of high MW components in the water, which was determined from the area of the SEC peak above a critical size / MW. The critical size was chosen as 130 nm, corresponding to a MW > 330 kDa (based on the Fox-Flory equation [43]); this cutoff was chosen because it is only slightly smaller than the membrane pore size (200 nm) and it corresponds to the MW of one of the polyacrylamide standards that was used for calibration in the SEC analysis. Fig. 4A shows a large increase in fouling index with increasing polymer size over a nearly 4 order of magnitude range. This result is
Fig. 4. Fouling index of flowback water increases with (A) the peak molecular weight and (B) the concentration of high MW species (> 130 nm in diameter). Data from repeat experiments are presented using the same symbol. Black square indicates raw fluid (only FR without reaction) and the remaining data are synthetic flowback. FR: only FR; FB: FR plus breaker; FBC: FB plus crosslinker. The open circles (FR) and triangles (FB and FBC) are generated by reaction with shale at 83 bar, 80 ℃, 25 g/L. The diamonds are flowback of FR only generated using different reaction conditions: open diamond, FR-S + P + T: without shale (at 83 bar and 80 ℃); center dotted diamond, FR + S + P - T: room temperature (at 83 bar with shale); and center cross diamond, FR-S + P + T: atmospheric pressure (at 80 ℃ with shale, diamond with low fouling index). Overall fouling index of raw fracturing fluids and synthetic flowback waters can be roughly divided into three regions: a) green oval is low fouling region (≤ m−1) where less than 20% fouling (final flux over initial flux) occurs during the course of filtration (100–300 L/m2); b) the red oval is medium fouling region (20–200 m−1), corresponding to 50–95% fouling during the course of filtration (50–300 L/m2); c) the grey oval is high fouling region (> 1000 m−1), in which the flux declines to nearly zero within 5 L/m2. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) 129
Journal of Membrane Science 560 (2018) 125–131
B. Xiong et al.
HVHF wastewater treatment. Low molecular weight volatile organics that haven been extensively analyzed previously [2,19,21,45] do not directly cause initial flux decline. The fouling behavior observed with the synthetic flowback water containing degraded PAM is similar to that observed previously with high fouling flowback water obtained from actual Marcellus Shale wastewater [14]. Our results clearly demonstrate that the fracturing fluid recipe and the interactions between the PAM and shale under HVHF conditions directly affect the degree of polymer degradation and in turn the fouling behavior. Our results provide one possible explanation for the large variation in fouling index of different flowback and produced water samples collected from the Marcellus shale. Reservoirs at high temperatures (deep) and with highly reactive shale mineralogy are likely to generate a flowback water with low fouling tendency due to the significant degradation of the PAM. Furthermore, the addition of breaker significantly reduced the PAM size and its fouling index; however, breaker is currently used in only ~10% of slickwater fracturing operations. We also showed that microfiltration membranes with pores as small as 0.1 µm are unable to reduce the effective size of the PAM, suggesting that other methods of pretreatment such as coagulation or oxidation should be considered in place of simple prefiltration. These data suggest membrane treatability of flowback and produced waters can be enhanced by: 1) reducing the use or concentration of PAM or replacement of PAM with an alternative drag reduction polymer in HVHF operations; 2) increasing the dosage of oxidative breaker such as persulfate or peroxide during HVHF operations or possibly the addition of this readily available oxidant in a pretreatment step to reduce the PAM MW prior to membrane filtration; 3) reducing the use of crosslinker and surfactant; 4) utilizing drilling waste that contains reactive shale as oxidative treatment media to enhance degradation of PAM, or 5) exploring opportunities for PAM removal by coagulation.
3.5. The effect of prefiltration on membrane fouling To better understand the fouling characteristics of the degraded polyacrylamide, fouling experiments were performed with permeate samples obtained by pre-filtration of the synthetic flowback water through different polycarbonate track-etched membranes with uniform pores of different pore size. The prefiltration was performed using approximately 0.1 L of synthetic flowback water where 90% of the feed was collected as permeate. The polymer in the initial flowback water had a peak MW of 1500 kDa (hydrodynamic size of 350 nm), which was generated by incubation of 1.5% FR with 0.0001 g/L (0.0004 mM) ammonium persulfate with no shale at 80 ℃ for 24 h. As shown in Fig. 5A, solutions that were prefiltered through the 0.4 and 0.8 μm membranes had flux decline curves that were similar to that of the unfiltered solution, with the fouling index ranging from 90 to 110 m−1. In contrast, prefiltration through the 0.2 μm membrane reduced the fouling index to 43 m−1 while prefiltration through the 0.1 μm membrane reduced the fouling index to 13 m-1, with the normalized flux declining by less than 20% over the course of filtration of 100 L/m2. SEC chromatographs of the prefiltered solutions are shown in Fig. 5B. Prefiltration of the flowback water through the 0.2 and 0.1 μm membranes removed 50% and 97% of the polymer, respectively, consistent with the large reduction in fouling seen in Fig. 5A. Much of the polymer removal may have occurred through the heavily fouled polycarbonate membranes (Figs. 5C and 5D), with the fouling layer providing significant removal of PAM across the entire MW spectrum. Surprisingly, the chromatographs suggest that prefiltration through the various pore size membranes had little effect on the peak MW even when the prefiltration was performed using a membrane with 0.1 µm pore size. We speculate that the highly flexible linear polymer chains could stretch under the elongation flow field at the pore entrance and be able to pass through pores that are much smaller than the polymer hydrodynamic size (as determined by SEC). Similarly enhanced transmission has been observed during the ultrafiltration and microfiltration of linear DNA molecules [44]. This suggests that commonly used 0.45 or 0.22 μm microfiltration membranes might not be sufficient to selectively remove large polymer foulants, making it difficult if not impossible to use this type of membrane pretreatment to control fouling in subsequent treatment steps.
Acknowledgement This research was funded by a Penn State College of Engineering Innovation Grant and a seed grant through the Center for Collaborative Research in Intelligent Gas Systems (CCRINGS) program funded by General Electric (GE). Additional funding was provided by the Pennsylvania Water Resources Research Center Small grants program. The authors would like to thank Weatherford Inc. for providing the synthetic chemical additives. The authors acknowledge the Kappe Environmental Engineering laboratories for TOC measurement instrumentation and technical assistance by David Jones. The authors also thank Rajarshi Guha for discussions of the membrane fouling experiments.
4. Conclusions To the best of our knowledge, this work demonstrates, for the first time, that both PAM and its degradation products generated under HVHF conditions are the primary cause of membrane fouling during
Fig. 5. (A) Flux decline of 0.2 μm PVDF membranes after prefiltration through polycarbonate membranes with pore size of 0.1, 0.2, 0.4, or 0.8 μm . (B) Chromatographs of polymer solutions after prefiltration. (C) and (D): Scanning electron microscopy images of polycarbonate track-etched membranes after filtration of the flowback water with pore size of 0.8 (top) and 0.2 μm (bottom) showed large aggregates as well as direct pore blocking (arrow). All samples were filtered by 0.45 μm syringe filters before SEC analyses. Prefiltration using 0.1, 0.2, 0.4 μm membranes (Whatman, Nuclepore™) was conducted at a constant pressure of 10 psi; prefiltration with 0.8 μm membranes (Millipore, Isopore™) was conducted at a constant pressure of 4 psi. The unfiltered solution was generated after incubating 0.15% FR fracturing fluid with 0.0001 g/L persulfate at 80 ℃ for 24 h, yielding a polymer solution with a peak MW of 1500 kDa (unfiltered). 130
Journal of Membrane Science 560 (2018) 125–131
B. Xiong et al.
Supporting information
surfactants, Anal. Chem. 86 (2014) 9653–9661. [20] G.T. Llewellyn, F. Dorman, J.L. Westland, D. Yoxtheimer, P. Grieve, T. Sowers, E. Humston-Fulmer, S.L. Brantley, Evaluating a groundwater supply contamination incident attributed to Marcellus Shale gas development, Proc. Natl. Acad. Sci. 112 (2015) 6325–6330. [21] B.D. Drollette, K. Hoelzer, N.R. Warner, T.H. Darrah, O. Karatum, M.P. O’Connor, R.K. Nelson, L.A. Fernandez, C.M. Reddy, A. Vengosh, R.B. Jackson, M. Elsner, D.L. Plata, Elevated levels of diesel range organic compounds in groundwater near Marcellus gas operations are derived from surface activities, Proc. Natl. Acad. Sci. 112 (2015) 13184–13189. [22] A. Butkovskyi, H. Bruning, S.A. Kools, H.H. Rijnaarts, A.P. van Wezel, Organic pollutants in shale gas flowback and produced waters: identification, potential ecological impact and implications for treatment strategies, Environ. Sci. Technol. (2017). [23] J. Rosenblum, E.M. Thurman, I. Ferrer, G. Aiken, K.G. Linden, Organic chemical characterization and mass balance of a hydraulically fractured well: from fracturing fluid to produced water over 405 days, Environ. Sci. Technol. 51 (2017) 14006–14015. [24] X. Wang, S. Chen, W. Dong, Z. Wang, L. Yang, X. Xi, Q. Zhang, L. Shi, Contribution of main pollutants in oilfield polymer-flooding wastewater to the total membrane fouling resistance, Sep. Sci. Technol. 47 (2012) 1617–1627. [25] S. Wang, C. Liu, Q. Li, Fouling of microfiltration membranes by organic polymer coagulants and flocculants: controlling factors and mechanisms, Water Res. 45 (2011) 357–365. [26] G. Liu, S. Yu, H. Yang, J. Hu, Y. Zhang, B. He, L. Li, Z. Liu, Molecular mechanisms of ultrafiltration membrane fouling in polymer-flooding wastewater treatment: role of ions in polymeric fouling, Environ. Sci. Technol. 50 (2016) 1393. [27] B. Xiong, Z. Miller, S. Roman-White, T. Tasker, B. Farina, B. Piechowicz, W.D. Burgos, P. Joshi, L. Zhu, C.A. Gorski, A.L. Zydney, M. Kumar, Chemical degradation of polyacrylamide during hydraulic fracturing, Environ. Sci. Technol. 52 (2018) 327–336. [28] T.L. Tasker, P.K. Piotrowski, F.L. Dorman, W.D. Burgos, Metal associations in Marcellus shale and fate of synthetic hydraulic fracturing fluids reacted at high pressure and temperature, Environ. Eng. Sci. (2016). [29] G.A. Kahrilas, J. Blotevogel, E.R. Corrin, T. Borch, Downhole transformation of the hydraulic fracturing fluid biocide glutaraldehyde: implications for flowback and produced water quality, Environ. Sci. Technol. 50 (2016) 11414–11423. [30] J. Gao, J. Yu, W. Wang, T. Lin, The accelerated degradation of aqueous polyacrylamide at low temperature, J. Appl. Polym. Sci. 69 (1998) 791–797. [31] H. Liu, T.A. Bruton, F.M. Doyle, D.L. Sedlak, In situ chemical oxidation of contaminated groundwater by persulfate: decomposition by Fe (III)-and Mn (IV)-containing oxides and aquifer materials, Environ. Sci. Technol. 48 (2014) 10330–10336. [32] P. Yin, Z. Hu, X. Song, J. Liu, N. Lin, Activated persulfate oxidation of perfluorooctanoic acid (PFOA) in groundwater under acidic conditions, Int. J. Environ. Res. Public Health 13 (2016) 602. [33] H. Sun, R.F. Stevens, J.L. Cutler, B. Wood, R.S. Wheeler, Q. Qu, A novel nondamaging friction reducer: development and successful slickwater frac applications, in: Tight Gas Completions Conference, in Society of Petroleum Engineers, 2010. [34] P.S. Carman, K. Cawiezel, Successful breaker optimization for polyacrylamide friction reducers used in slickwater fracturing, in: Society of Petroleum Engineers, 2007. [35] B.L. Gall, A.R. Sattler, D.R. Maloney, C.J. Raible, Permeability damage to natural fractures caused by fracturing fluid polymers, in Society of Petroleum Engineers, 1988. [36] R.L. Thomas, J.E. Brown, The impact of fracturing fluids on conductivity and performance in low-temperature wells, in: Society of Petroleum Engineers, 1989. [37] G.E. King, Hydraulic fracturing 101: what every representative, environmentalist, regulator, reporter, investor, university researcher, neighbor and engineer should know about estimating frac risk and improving frac performance in unconventional gas and oil wells, in Society of Petroleum Engineers, 2012. [38] C. Aften, W.P. Watson, Improved friction reducer for hydraulic fracturing, in Society of Petroleum Engineers, 2009. [39] What Chemicals Are Used, in, FracFocus Chemical Disclosure Registry. [40] B.V. Mork, J.M. Gravitt, Enhanced chemical oxidation, U.S. Patent Application 20140097381 A1, 2014. [41] D.B. Bennion, F.B. Thomas, R.F. Bietz, D.W. Bennion, Water and hydrocarbon phase trapping in porous media-diagnosis, prevention and treatment, J. Can. Pet. Technol. 35 (1996) 29–36. [42] G.P. Anipsitakis, D.D. Dionysiou, Radical generation by the interaction of transition metals with common oxidants, Environ. Sci. Technol. 38 (2004) 3705–3712. [43] M. Rubinstein, R.H. Colby, Polymer Physics, Oxford University, NEW YORK, 2003. [44] Y. Li, E.E. Borujeni, A.L. Zydney, Use of preconditioning to control membrane fouling and enhance performance during ultrafiltration of plasmid DNA, J. Membr. Sci. 479 (2015) 117–122. [45] S.J. Maguire-Boyle, A.R. Barron, Organic compounds in produced waters from shale gas wells, Environ. Sci.: Process. Impacts 16 (2014) 2237–2248.
The supporting information includes a detailed description of the calculation of fouling index and polymer MW and size distribution. It also includes additional data for raw fluids and synthetic flowback with FR with other additives. The Supporting Information is available free of charge on the ACS Publications website at http://dx.doi.org/10.1016/j. memsci.2018.04.055. Notes The authors declare no competing financial interest. Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.memsci.2018.04.055. References [1] N. Abualfaraj, P.L. Gurian, M.S. Olson, Characterization of Marcellus shale flowback water, Environ. Eng. Sci. 31 (2014) 514–524. [2] W. Orem, C. Tatu, M. Varonka, H. Lerch, A. Bates, M. Engle, L. Crosby, J. McIntosh, Organic substances in produced and formation water from unconventional natural gas extraction in coal and shale, Int. J. Coal Geol. 126 (2014) 20–31. [3] C. He, T. Zhang, X. Zheng, Y. Li, R.D. Vidic, Management of Marcellus shale produced water in Pennsylvania: a review of current strategies and perspectives, Energy Technol. 2 (2014) 968–976. [4] W.L. Ellsworth, Injection-induced earthquakes, Science 341 (2013). [5] Eureka-Resources, Standing Stone Facility, in, 2014. [6] A. Butkovskyi, H. Bruning, S.A.E. Kools, H.H.M. Rijnaarts, A.P. Van Wezel, Organic pollutants in shale gas flowback and produced waters: identification, potential ecological impact, and implications for treatment strategies, Environ. Sci. Technol. 51 (2017) 4740–4754. [7] C. He, X. Wang, W. Liu, E. Barbot, R.D. Vidic, Microfiltration in recycling of Marcellus Shale flowback water: solids removal and potential fouling of polymeric microfiltration membranes, J. Membr. Sci. 462 (2014) 88–95. [8] Q. Jiang, J. Rentschler, R. Perrone, K. Liu, Application of ceramic membrane and ion-exchange for the treatment of the flowback water from Marcellus shale gas production, J. Membr. Sci. 431 (2013) 55–61. [9] S.M. Riley, J.M.S. Oliveira, J. Regnery, T.Y. Cath, Hybrid membrane bio-systems for sustainable treatment of oil and gas produced water and fracturing flowback water, Sep. Purif. Technol. 171 (2016) 297–311. [10] S. Mondal, S.R. Wickramasinghe, Produced water treatment by nanofiltration and reverse osmosis membranes, J. Membr. Sci. 322 (2008) 162–170. [11] B.D. Coday, P. Xu, E.G. Beaudry, J. Herron, K. Lampi, N.T. Hancock, T.Y. Cath, The sweet spot of forward osmosis: treatment of produced water, drilling wastewater, and other complex and difficult liquid streams, Desalination 333 (2014) 23–35. [12] H.C. Duong, A.R. Chivas, B. Nelemans, M. Duke, S. Gray, T.Y. Cath, L.D. Nghiem, Treatment of RO brine from CSG produced water by spiral-wound air gap membrane distillation — a pilot study, Desalination 366 (2015) 121–129. [13] S.M. Riley, D.C. Ahoor, J. Regnery, T.Y. Cath, Tracking oil and gas wastewaterderived organic matter in a hybrid biofilter membrane treatment system: a multianalytical approach, Sci. Total Environ. 613 (2018) 208–217. [14] B. Xiong, A.L. Zydney, M. Kumar, Fouling of microfiltration membranes by flowback and produced waters from the Marcellus shale gas play, Water Res. 99 (2016) 162–170. [15] G. Chen, Z. Wang, L.D. Nghiem, X.-M. Li, M. Xie, B. Zhao, M. Zhang, J. Song, T. He, Treatment of shale gas drilling flowback fluids (SGDFs) by forward osmosis: membrane fouling and mitigation, Desalination 366 (2015) 113–120. [16] S. Alzahrani, A.W. Mohammad, Challenges and trends in membrane technology implementation for produced water treatment: a review, J. Water Process Eng. 4 (2014) 107–133. [17] M. Elsner, K. Hoelzer, Quantitative survey and structural classification of hydraulic fracturing chemicals reported in unconventional gas production, Environ. Sci. Technol. 50 (2016) 3290–3314. [18] K. Hoelzer, A.J. Sumner, O. Karatum, R.K. Nelson, B.D. Drollette, M.P. O’Connor, E.L. D’Ambro, G.J. Getzinger, P.L. Ferguson, C.M. Reddy, M. Elsner, D.L. Plata, Indications of transformation troducts from hydraulic fracturing additives in shalegas wastewater, Environ. Sci. Technol. 50 (2016) 8036–8048. [19] E.M. Thurman, I. Ferrer, J. Blotevogel, T. Borch, Analysis of hydraulic fracturing flowback and produced waters using accurate mass: identification of ethoxylated
131