Origin of Flowback and Produced Waters from Sichuan Basin, China

Nov 13, 2018 - Produced waters from conventional natural gas wells from the ...... L. M. Environmental and Human Impacts of Unconventional Energy Deve...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/est

Cite This: Environ. Sci. Technol. 2018, 52, 14519−14527

Origin of Flowback and Produced Waters from Sichuan Basin, China Yunyan Ni,† Caineng Zou,† Huiying Cui,† Jian Li,† Nancy E. Lauer,‡ Jennifer S. Harkness,‡ Andrew J. Kondash,‡ Rachel M. Coyte,‡ Gary S. Dwyer,‡ Dan Liu,§ Dazhong Dong,† Fengrong Liao,† and Avner Vengosh*,‡ †

PetroChina Research Institute of Petroleum Exploration and Development, Beijing 100083, China Nicholas School of the Environment, Duke University, Durham, North Carolina 27708, United States § Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing 100029, China ‡

Environ. Sci. Technol. 2018.52:14519-14527. Downloaded from pubs.acs.org by MIDWESTERN UNIV on 01/13/19. For personal use only.

S Supporting Information *

ABSTRACT: Shale gas extraction through hydraulic fracturing and horizontal drilling is increasing in China, particularly in Sichuan Basin. Production of unconventional shale gas with minimal environmental effects requires adequate management of wastewater from flowback and produced water (FP water) that is coextracted with natural gas. Here we present, for the first time, inorganic chemistry and multiple isotope (oxygen, hydrogen, boron, strontium, radium) data for FP water from 13 shale gas wells from the Lower Silurian Longmaxi Formation in the Weiyuan gas field, as well as produced waters from 35 conventional gas wells from underlying (Sinian, Cambrian) and overlying (Permian, Triassic) formations in Sichuan Basin. The chemical and isotope data indicate that the formation waters in Sichuan Basin originated from relics of different stages of evaporated seawater modified by water−rock interactions. The FP water from shale gas wells derives from blending of injected hydraulic fracturing water and entrapped saline (Cl ∼ 50,000 mg/L) formation water. Variations in the chemistry, δ18O, δ11B, and 87Sr/86Sr of FP water over time indicate that the mixing between the two sources varies with time, with a contribution of 75% (first 6 months) to 20% (>year) of the injected hydraulic fracturing water in the blend that compose the FP water. Mass-balance calculation suggests that the returned hydraulic fracturing water consisted of 28−49% of the volume of the injected hydraulic fracturing water, about a year after the initial hydraulic fracturing. We show differential mobilization of Na, B, Sr, and Li from the shale rocks during early stages of operation, which resulted in higher Na/Cl, B/Cl, Li/Cl, and 87Sr/86Sr and lower δ11B of the FP water during early stages of FP water formation relative to the original saline formation water recorded in late stages FP water. This study provides a geochemical framework for characterization of formation waters from different geological strata, and thus the ability to distinguish between different sources of oil and gas wastewater in Sichuan Basin.

■

INTRODUCTION Technological advances in hydraulic fracturing and horizontal drilling have dramatically increased shale gas and tight oil extraction in the U.S. and Canada, with more global increases on the horizon.1,2 China has one of the largest global shale gas resources, with estimates ranging from 12.8 to 31.2 trillion m3 of natural gas.3−8 In 2016, the annual shale gas production from 356 active shale gas wells from the southern Sichuan Basin was 7.8 billion m3.9 China is aiming for a significant increase in shale gas production, with annual production goals of 30 and 100 billion m3 by 2020 and 2030, respectively.4,5,10−12 One of the key environmental issues that is associated with the operation of unconventional oil and gas development is the quality of the flowback and produced water (FP water) that is generated during hydrocarbon extraction.1,2,13,14 The ability to reuse, treat, and manage FP water that becomes wastewater is directly dependent on its quantity and its quality, which in turn requires a detailed assessment of the geochemistry of FP water.15,16 Evaluation of the geochemistry of the FP water can © 2018 American Chemical Society

also help in understanding the hydraulic fracturing processes and the origin of fluids in shale formations, yet most of the reported data on the geochemistry of FP water are restricted to unconventional shale basins from the U.S.,17−21 Canada,22 and northern Qaidam Basin in China.23 Here we present the first comprehensive geochemical data set for FP water from shale gas wells in Weiyuan gas field in the Sichuan Basin, China. The Sichuan Basin is the most productive and economically viable shale gas basin in China, where the Upper Ordovician Wufeng Formation−Lower Silurian Longmaxi Formation is the major shale gas resource. Several gas fields have been discovered in the Sichuan Basin (Figure 1), of which the Fuling, Weiyuan, and Changning fields are currently being exploited for shale gas.4,5,7−12 Received: Revised: Accepted: Published: 14519

August 6, 2018 November 8, 2018 November 12, 2018 November 13, 2018 DOI: 10.1021/acs.est.8b04345 Environ. Sci. Technol. 2018, 52, 14519−14527

Article

Environmental Science & Technology

Figure 1. Map of the gas fields in Sichuan Basin, China.

■

METHODS Sample Collection. Produced waters from conventional natural gas wells from the Sinian (n = 3), Cambrian (n = 4), Permian (n = 8) and Triassic (n = 20) formations were collected from different gas fields in Sichuan Basin (Figures 1and S1; Tables S3 and S4). FP water samples from unconventional shale gas wells were systematically collected on multiple occasions after the initial hydraulic fracturing from 3 shale gas wells (n = 65), in addition to samples from 10 shale gas wells that were collected at known times after hydraulic fracturing in Weiyuan gas field from Sichuan Basin (Figure 1; Tables S5 and 4; see more information on the FP production and operational of the 3 shale gas wells in SI). Fresh waters used for hydraulic fracturing and hydraulic fracturing fluids prior to injection (n = 13) were also analyzed (Table S7). All samples were filtered and preserved in high density polyethylene (HDPE) airtight bottles following USGS field collection protocols.24 Samples were filtered through 0.45 μm filters for dissolved anions, cations, and inorganic trace element isotopes (B, Sr). Trace metal samples were preserved in ∼1% Optima nitric acid. Samples were stored on ice or refrigerated and shipped to Duke University. Field data measurements included pH and electrical conductivity. Analytical Methods. Samples were analyzed for dissolved anions by ion chromatography on a Dionex IC DX-2100. The IC calibration was verified using a secondary Dionex 7-anion standard at varying concentrations. Only runs reporting values within 5% of the secondary standard were accepted. Major cations were analyzed by direct current plasma optical emission spectrometry (DCP-OES), whereas trace elements, including Li, B, and Sr, were analyzed by inductively coupled plasma mass spectrometry (ICP-MS) on a VG PlasmaQuad-3 after

This study aims to characterize the geochemistry and origin of the FP water from Weiyuan gas field in Sichuan Basin (Figure 1). Previous studies have shown that FP waters are typically derived from a mixture of formation water entrapped within the shale and returned injected water that is used for hydraulic fracturing.16,18,19,21 These studies have indicated that the relative proportions of these two water sources likely determine the overall salinity and water quality of the FP water. To address the source of FP water in Sichuan Basin, the chemistry of formation waters from the other geological formations in the Sichuan Basin must be evaluated, in addition to the direct analysis of FP water. In this study we provide a geochemical analysis of two major water sources: (1) produced waters extracted from conventional natural gas wells in different geological formations underlying and overlaying the Lower Silurian Longmaxi Formation in Sichuan Basin (Figure S1); and (2) FP water samples from Weiyuan gas field in the Sichuan Basin (Figure 1). We investigated the systematic geochemical variations of FP water through time (up to 470 days after hydraulic fracturing) in 3 wells, in addition to investigating FP water collected from another 10 shale gas wells during later time periods (400 to 1000 days after hydraulic fracturing). We used a wide range of geochemical tracers including major elements, trace elements, and the isotope systematics of oxygen, hydrogen, boron, strontium, and radium to identify the water sources and the mechanisms that control the chemistry of formation water and FP water. The extensive data set of both conventional formation waters (n = 35) and unconventional FP water (n = 65) provides the essential information for evaluation of the overall origin of formation water in Sichuan Basin, and specifically, the origin of the FP water. 14520

DOI: 10.1021/acs.est.8b04345 Environ. Sci. Technol. 2018, 52, 14519−14527

Article

Environmental Science & Technology

Figure 2. Box plots of the variations of chloride (mg/L) and Br/Cl ratios (molar), 87Sr/86Sr ratios, and δ11B (‰) of formation waters sorted by formation ages investigated in this study. For more detailed geological strata and formation names, see Figure S1. The data show that formation waters collected from geological formations overlying the Silurian shale (the Permian and Triassic) are characterized by higher salinity and Br/Cl ratios relative to those from the Silurian (FP water), Cambrian, and Sinian formations, inferring different sources of saline waters. The 87Sr/86Sr and δ11B variations reflect intensive water−rock interactions and modification of the original evaporated seawater.

a Conflo III flow adapter. Raw δ values were normalized offline against known vs measured isotope values for international reference waters VSMOW, VSLAP, and IAEA-OH16. The δ2H and δ18O values are expressed in per mil versus VSMOW, with standard deviations of ±0.5 ‰ and ±0.1‰, respectively. Boron and Sr isotope ratios were measured using Thermal Ionization Mass Spectrometry (TIMS) ThermoFischer Triton at Duke University. Samples for B analysis were pretreated with 10% Optima hydrogen peroxide. The 11B/10B ratios were measured on single Re filaments as BO2− ions in negative mode and normalized to the NIST SRM 951 standard. Boron isotope ratios are reported as

gravimetric dilution in 2% HNO3 to yield a total cation concentration comparable to that of NIST SRM1643F. To ensure negligible memory effects for all elements analyzed, especially boron with its well-known memory effect, sample rinse/washout and uptake times were evaluated in timeresolved fashion for the highest concentration samples and standards to be analyzed and extended to 200 and 80 s, respectively. At these timings, signals for all elements returned to analytical blank values by the end of the rinse/washout period and equilibrated for the next sample by the end of the uptake time. Lack of memory effects was further evaluated and confirmed through randomly analyzed analytical and fullprocess blanks. All concentrations are reported to the instrumental precision of ±0.1 mg/L for major elements and ±0.1 μg/L for trace elements. We calculated detection limits by multiplying the standard deviation of repeated blank measurements by three and dividing by the slope of the external standard. Relative standard errors (RSE) were less than 10% for randomized duplicates on the DCP-OES and ICP-MS. Charge balances were 1) 228Ra/226Ra ratios. The inverse correlation between δ11B and B/Cl and relationship to the evaporated seawater curve indicate that B in formation water can be derived from both mobilization of 10B from the host rocks (Sinian, Cambrian, Silurain, and Permian formation waters) or residual brine from which 11B is preferentially adsorbed into clay minerals (some of the Jurassic formation waters).

■

with its immediate decay product 228Ac (t1/2 = 6.1 h). Following incubation, samples were counted on a Canberra Broad Energy 5030 Germanium Gamma detector surrounded by 10 cm of lead shielding. Samples were typically counted for 6- 48 h so that counting errors (2σ) were less than 10%. 226Ra activities were measured through the 351 keV energy peak of 214 Pb. 228Ra activities were measured through the 911 keV energy peak of 228Ac. The detector efficiencies were determined using a U−Th reference ore material (DL-1a) prepared by the Canadian Certified Reference Materials Project (CCRMP) that was packaged and incubated in a container identical to the samples. Background and efficiency checks were performed routinely prior to and during the time frame of sample analyses. During the course of this study, Duke Laboratory participated in an interlaboratory comparison, where oil and gas wastewaters with similar matrix to the Sichuan flowback and produced waters were investigated by 15 laboratories in the U.S. and Canada.25 The water major and trace elements (including boron) as well as radium nuclides data measured by Duke Laboratory were between ±5% and ±10% of the most probable value (MPV) obtained by the 15 laboratories, which verifies the integrity of the analytical methods conducted in this study.

RESULTS AND DISCUSSION Geochemistry of Formation Waters from Sichuan Basin. The variations of Cl, Na, Br, Ca, Mg, and Sr concentrations in produced waters from conventional gas wells collected from different geological formations in Sichuan Basin (Figure S1, Table S3) were compared to those of evaporated seawater (Figure S2). Some elements of the chemistry of the formation water mimic the composition of modified evaporated seawater.26−28 Formation waters from the Sinian and Cambrian formations (Cl up to 48 000 mg/L) had Br/Cl up to 3.5 × 10−3 (molar ratio), which reflects seawater evaporation up to ∼18-fold. In contrast, formation waters from the Permian and Triassic formations had a much wider chloride range (14 000 to 141 000 mg/L) and a higher Br/Cl ratio (up to 6 × 10−3; Figure 2), indicating a higher degree of original seawater evaporation, up to 30-fold, followed by dilution with meteoric water. We calculated the degree of evaporation based on the assumption that both Cl and Br in the formation waters are conservative elements and that the relatively high Br/Cl ratios of the Sichuan brines reflect the degree of evaporation in residual evaporated seawater as experimentally shown in McCaffrey et al.29 The dilution effect is also demonstrated by the variations of the stable isotopes of oxygen and hydrogen (Figure S2), which showed large dilution of the Permian and Triassic produced waters. The narrow salinity variations and relatively high δ18O and δ2H values 14522

DOI: 10.1021/acs.est.8b04345 Environ. Sci. Technol. 2018, 52, 14519−14527

Article

Environmental Science & Technology

Figure 4. Time series variations of Cl (A), Ca (B), δ11B (C), and 87Sr/86Sr (D) in FP water from Weiyuan gas field in Sichuan Basin. The distinction between early and late FP water suggest differential mixing proportions of the injected hydraulic fracturing water and the saline formation water, as well as geochemical modifications of the original saline formation water induced by interactions with the Silurian shale. For the variations of δ18O, Na, B, Li, and Sr in FP water see Figure S4.

lower 87Sr/86Sr ratios (0.7083), reflecting Sr contribution from a nonradiogenic rocks, most likely marine carbonate rocks with a lower 87Sr/86Sr ratio, typical of the Permian seawater (a 87 Sr/86Sr range of 0.7068−0.7080).33 The Triassic formation waters had intermediate 87Sr/86Sr ratios (medium = 0.7122), suggesting contribution from both carbonate and shale rock sources. The enrichment of Ra in formation waters (a range of 5 to 60 Bq/L of 226Ra and 3 to 52 of 228Ra; Table S2) was associated with high salinity (Cl up to 140 000 mg/L) and redox conditions (zero sulfate contents) that commonly characterize formation water, and are known to promote Ra mobilization from rocks to the aquatic phase.28 Similarly high Ra nuclides activities have been reported for formation waters and FP waters in oil and gas basins in the U.S.13,14,18,34 The Ra isotope variations seem to reflect the lithology of the host rocks; the 228Ra/226Ra activity ratios of the Cambrian and Silurian (FP water) waters were low (∼0.3; Figure 3), reflecting U-rich shale rocks that contribute 226Ra from the 238 U decay chain. Likewise, the Permian formation waters showed a predominance of 226Ra over 228Ra, most likely reflecting water−rock interactions with U-rich carbonate rocks, which is consistent with the low 87Sr/86Sr ratios observed in the Permian brines (Figure 3). In contrast, the Triassic brines had much higher 228Ra/226Ra ratios (median = 1.67; Figure 3), indicating a large contribution of 228Ra originating from the 232 Th decay series, and thus higher abundance of Th in the host rocks, which is common in sandstone rocks.35 We therefore were able to delineate the specific lithologies that interact with the various formation waters, including carbonate rocks (predominance of 226Ra), shale rocks (228Ra/226Ra ∼ 0.3), and sandstone rocks (228Ra/226Ra > 1).

recorded in the produced waters from Sinian and Cambrian formations (Figure S2) suggest only a small degree of dilution with meteoric water, and hence a closed system in which the residual evaporated seawater was preserved. In contrast, the larger salinity variations and lower δ18O and δ2H values of brines from the Permian and Triassic formations indicate an open system and substantial mixing with meteoric water. Given that the δ18O and δ2H values of some of the diluted Triassic brines (−18‰ and −166‰; Figure S2) are much lower than those of modern meteoric water (−6‰ and −43‰; Table S7), we posit that the dilution occurred with fossil meteoric water during colder climate periods relative to modern conditions. Whereas Br/Cl and stable isotope data provide information on the original evaporated seawater source, all of the formation waters were enriched in Ca, Sr, and Ra relative to the expected seawater evaporation curves, reflecting intensive-water−rock interactions (Figure S2). The enrichment of Ca and Sr combined with Mg depletion relative to evaporated seawater (Figure S2) suggests modification of the original evaporated seawater through dolomitization and/or ion-exchange. This type of geochemical evolution and the formation of Cachloride brines has been documented in numerous sedimentary basins, including the Appalachian Basin in the eastern U.S.17−19,30−32 The Sr isotope ratios of the formation waters showed large variations (Figure 2); the Sinian and Cambrian brines had a large 87Sr/86Sr range (0.710 to 0.723), whereas Silurian waters (FP water, see discussion later) had a narrow range, with relatively high 87Sr/86Sr ratios (median = 0.719), which reflects contribution from a radiogenic Sr sources, most likely from the host shale rocks. In contrast, the Permian brines had much 14523

DOI: 10.1021/acs.est.8b04345 Environ. Sci. Technol. 2018, 52, 14519−14527

Article

Environmental Science & Technology

Figure 5. Bromide, Na, Ca, Li, B versus Cl and δ2H versus δ18O in FP waters from 13 shale gas wells and waters used for hydraulic fracturing (HF water) in Weiyuan gas field, Sichuan Basin. The high correlations between Br and Cl (R2 = 0.95; p < 0.001) and δ2H and δ18O (R2 = 0.75; p < 0.001) indicate that the FP water originated from mixing of the HF water and formation water. The higher B and Li contents and lower δ11B of FP water generated during the first ∼160 days relative to later-stages FP water suggest that B and Li are mobilized from shale rocks following hydraulic fracturing.

and (2) ∼30-fold evaporated seawater modified by water−rock interaction as well as dilution by (fossil) meteoric water in the Permian and Triassic formations. The relatively lower Br/Cl ratio of the FP water from the Silurian Shale is similar to the earlier phase of seawater evaporation recorded in the Sinian and Cambrian formations (Figure 1), and thus we suggest a similar source for the formation water entrapped in the overlying Silurian shale. This geochemical distinction is the base for our evaluation of the FP water in the next section. Origin of FP Water from Sichuan Basin. In Sichuan Basin, the water used for hydraulic fracturing is commonly composed of fresh water (∼90%) blended with ∼10% recycled FP water, with salinity of up to 500 mg/L (Table S7). In contrast, the salinity of the FP water following hydraulic fracturing was found to be much higher (Cl = 5000 to 37 000 mg/L; Table S5; Figures 3 and 4). Similar patterns of high salinity of FP water relative to the injected water have been reported in the U.S. and Canadian shale gas basins.18,19,22,32 These studies have suggested that the FP water originated from mixing of entrapped formation water in the shale and the injected hydraulic fracturing water. We tested this hypothesis by using conservative tracers of Br/Cl and stable isotopes. The Br−Cl and the δ2H−δ18O variations (Figure 5) confirm that the FP water in Sichuan Basin indeed originated from the blending of formation water with high Br/Cl, δ2H and δ18O values, which are similar (but not necessarily identical) to those of the Cambrian formation water, and the injected lowsaline and low-δ18O and δ2H water (Table S7). The Br/Cl of

The B isotope variations deviated from the expected evaporated seawater evolution curve26 and showed a systematic inverse relationship between δ11B and B/Cl ratios in most of the produced waters (Figure 3). This negative correlation indicates that boron is highly mobilized from the host rocks resulting in elemental B enrichment associated with lower δ11B relative to the expected B and δ11B of evaporated seawater. This was most noticeable in formation waters from the Sinian, Cambrian, Silurian (FP water; see further discussion below), and Permian formations. In contrast, several formation waters from the Triassic formations had relatively low B/Cl and high δ11B relative to evaporated seawater (Figure 3), reflecting either B isotope fractionation induced by selective uptake of 10 B, which typically occurs during boron adsorption onto clay minerals, or the composition of the original evaporated seawater with low B/Cl and high δ11B.25 Whereas the relatively lower δ11B values of the Sinian, Cambrian and Permian formation waters (10−26‰) are consistent with the composition of oil-field brines reported in other studies, the high δ11B of the Triassic produced waters (25−56‰) is similar to “Dead Sea” type brines observed in conventional oil and gas produced waters in the Appalachian basin.19,28 Overall, the geochemical data from produced waters collected from different geological formations in Sichuan basin (Figure S1) indicate the occurrence of two types of evaporated seawater remnants: (1) ∼18-fold evaporated seawater that was modified by water−rock interaction under a nearly closed system in the Sinian and Cambrian formations; 14524

DOI: 10.1021/acs.est.8b04345 Environ. Sci. Technol. 2018, 52, 14519−14527

Article

Environmental Science & Technology the FP water (∼3 × 10−3) was consistent with those of the Cambrian and Sinian formation waters, but lower than formation waters from the Permian and Triassic formations (Figure 2). This suggests that the formation water entrapped within the Silurian shale originated from similar evaporated seawater to Cambrian and Sinian formations in the Sichuan Basin (Figure 2). We therefore suggest that the FP water in Sichuan Basin originated from a blend of the injected hydraulic fracturing water and saline formation water. On the basis of Br/Cl mass balance calculations, and the similarity of the Br/ Cl ratios of the FP water to those of the Cambrian formation water, we suggest that the chloride content of the saline endmember (i.e., formation water entrapped within the Silurian shale) is similar to that of the Cambrian formation water (∼50 000 mg/L). The Cl variations in the FP water (Figure 4) reflect therefore a contribution of 25% (for Cl ∼ 10 000 mg/L) to 90% (Cl ∼ 37 000 mg/L) formation water at different times after the initial hydraulic fracturing operation. The salinity of FP water in the U.S. shale gas basins typically increases with time (typically 1−3 months) after hydraulic fracturing, which reflects a rapid increase in the fraction of the formation water in the FP water mix.18,19 In contrast, the chemistry and stable isotopes of FP water from Sichuan Basin showed relatively stable values with only small fluctuations during the first 5−6 months, with higher values only about a year after the initial hydraulic fracturing (Figure 4). This pattern is not consistent with a clear trend toward a predominance of saline formation water that has been observed in several time series profiles reported for the U.S. shale gas basins. Previous studies have suggested that the majority of the injected hydraulic fracturing fluids is retained through imbibition36 into the shale rocks following hydraulic fracturing, with estimates of up 90% retention.16,20 Using the chloride concentrations and the available volume data of the FP water, we conducted mass-balance calculations to evaluate the volume of the return injected water (see details in SI). We show that after a year, the accumulative volume of the FP water was 32% to 62% of the water volume used for hydraulic fracturing (Table S8). The calculated volume of the injected hydraulic fracturing fluid that was returned during this time varied between 28% to 49% of the volume of the injected hydraulic fracturing water (Table S8). The results of this model confirm a relatively smaller retention of the hydraulic fracturing fluids into the shale matrix in Sichuan Basin. Operational information indicates that the three investigated shale gas wells from Weiyuan gas field were not subject to refracturing, nor injection of cleaning or acidifying solutions (see SI for details). Therefore, artificial addition of fresh water cannot be the cause for the relative larger fraction of low-saline water in the FP water. Some of the wells were shut off for different time periods (see SI for details) that could induce some artificial effects on the FP water flow and composition. The available data resolution for wells 1 and 3 do not reveal a direct relationship between the shut-down intervals and salinity variation, although the large fluctuation in salinity observed in FP water from well 1 during late production stages (431−440 days after hydraulic fracturing; Figure S3) could be related to such artificial effects. Given these observations, we posit that the salinity and stable isotopes variations over time reflect differential arrival of the retained hydraulic fracturing fresh water from the shale matrix to the shale gas wells. Unlike the scenario of a constant and increasing fraction of exclusively formation water flow over

time (and thus increasing salinity until close to 100% contribution of formation water), it appears that both water sources (i.e., retained low-saline hydraulic fracturing water and formation water) flow at different proportions at different times back to the shale gas wells in Sichuan Basin. We do not know the source of the difference in this mechanism, nor what causes some of the hydraulic fracturing water that is retained within the Lower Silurian Longmaxi Formation in Sichuan Basin to return back to the wells after long periods of operation. Geochemical Variations over Time. The variations in water chemistry during the first 160 days as revealed by time series data from 3 wells was limited, followed by much larger fluctuations during later stages, up to 1000 days after hydraulic fracturing (Figure 4). For example, soon after fracturing, the chloride variations in the FP water were small (10 000−12 000 mg/L), which suggests contributions of ∼20% formation water. During later stages (>year), the salinity increased (up to 37 000 mg/L), reflecting a larger proportion (∼75%) of the formation water with an estimate salinity of 50 000 mg/L. Yet even after ∼600 days, some of the FP water showed a large fraction of returned hydraulic fracturing water with low Cl and δ18O values. The more drawn out of returned hydraulic fracturing water in the FP water blend means that the overall salinity and concentrations of inorganic constituents in the FP water from Sichuan Basin is much lower than FP water from shale formations like the Marcellus shale,16−18,30,32,37 and thus the FP water from Sichuan Basin could be more manageable for reuse and/or treatment. We distinguish between two types of FP water (Figure 5): (1) FP water with lower Cl, Ca, Sr, Na, higher B and Li, lower δ11B (23.8‰−27.6‰), and higher 87Sr/86Sr (0.7196−0.7200) that characterize the FP water during the first ∼160 days but also some wells from later stages; and (2) FP water with higher Cl, Na, Ca, Sr, lower B and Li (and their ratios to Cl), higher δ11B (up 33.5‰), and lower 87Sr/86Sr (as low as 0.71663) that characterize some of the late stage FP waters, over a year after hydraulic fracturing. We suggest that these variations indicate different types of water−rock interactions during the different stages of FP evolution. Given that Na/Cl and Br/Cl ratios covary in evaporated seawater,29 we use these ratios as indicators for the origin of the entrapped formation water in the shale rocks. The low Na/Cl (0.66) and Br/Cl (3.3 × 10−3) ratios in the late FP water match exactly the expected ratios of 17- to 18-fold evaporated seawater, which we show also in the Cambrian produced water and thus suggested to be the original saline source in both the Cambrian and Silurian shale rocks. Therefore, the high Ca, Sr, lower B and Li, higher δ11B and lower 87Sr/86Sr seem to reflect this original saline source composition. In contrast, the higher Na/Cl, lower Ca, Sr, higher B and Li, lower δ11B and higher 87Sr/86Sr of the early FP waters suggest modification through intensive water−rock interactions with the host Silurian shale. The relatively high B/Cl and lower δ11B of the earlier stage FP water is consistent with the FP water composition reported from other shale formations, such as the Marcellus Shale from the Appalachian Basin19 and the Dameigou Shale from Qaidam Basin.23 The inverse correlation between B/Cl and δ11B (Figure 3) and similarly lower Li concentrations of the late-stage and more saline FP water suggest that during the early stages of hydraulic fracturing, B and Li are controlled by desorption from clay minerals in the shale formation, resulting in high B and Li combined with low δ11B. Yet the arrival of a 14525

DOI: 10.1021/acs.est.8b04345 Environ. Sci. Technol. 2018, 52, 14519−14527

Article

Environmental Science & Technology

estimated up to ∼17- to 18-fold (inferred from Br/Cl ratio of ∼3 × 10−3) that was entrapped in the Sinian, Cambrian, Silurian, and Permian formations; and (2) a later stage with chloride up to 140 000 mg/L and evaporation up to 30-fold (Br/Cl ∼ 5 to 6 × 10−3) that was entrapped in the Triassic formations. The different lithologies of the hosting formations and different types of desorption/adsorption reactions are reflected by variations of 87Sr/86Sr, δ11B, and 228Ra/226Ra ratios. The FP water that is extracted with shale gas following hydraulic fracturing of the Silurian shale is derived from blending of the injected fresh water (with a small component of recycled FP water) and Silurian formation waters. The relative proportions of the two water sources vary in time, and we show that the injected hydraulic fracturing water can return to the wells even 2 years after hydraulic fracturing. The integration of geochemical and isotope tracers (87Sr/86Sr, δ11B, and 228Ra/226Ra) in this study provides diagnostic tools to delineate the different water sources, water−rock interactions, and the ability to distinguish between conventional and unconventional oil and gas wastewaters.

higher fraction of the saline end-member at later stages was associated with lower B/Cl and Li/Cl and higher δ11B, which apparently represent the composition of the saline endmember entrapped within the shale rocks. Likewise, the lower 87Sr/86Sr seems to reflect the distinction between the isotope composition of the saline end-member and the more radiogenic 87Sr/86Sr that is mobilized from shale rocks during the early stages of the formation of flowback water. It has been suggested that the introduction of low-saline water to the shale formation through the process of hydraulic fracturing enhances the mobilization of boron (and Li) from desorption sites on clay minerals as part of “freshening effect”,19 which we also see in the Sichuan Basin. We conclude that the reactions with shale rocks have induced inverse base-exchange reactions in which Ca and Sr were retained while Na, B, and Li were mobilized from the clay minerals during the migration of the Ca-rich saline formation water through the fractured Silurian shale. Geochemical Fingerprints of Conventional versus Unconventional Oil and Gas Wastewater. One of the intriguing questions related to the environmental effects of shale gas development is the actual impact of the recent unconventional shale gas and tight oil exploration relative to legacy contamination from historical operation of conventional oil and gas. The Sichuan Basin, like many of the shale gas sites in the U.S., has a long history of conventional natural gas development, and thus the distinction between the impact of current unconventional shale gas and legacy conventional gas activities could be important in cases of environmental disputes.1,19 Given the distinct isotope ratios of B, Li, and Sr in unconventional oil and gas wastewater relative to those of conventional produced waters, previous studies have utilized these isotopes as indicators for identification of possible migration of unconventional oil and gas wastewater in the environment.1,19,31,38 In Sichuan Basin, the combined utilization of Br/Cl, δ11B, 87Sr/86Sr, and 228Ra/226Ra tracers can provide an indicative method for distinction of wastewater derived from shale gas relative to conventional gas production. Our data show that these tracers can indeed be used to delineate the sources, but not in all cases. For example, produced waters from the Sinian and Cambrian formations have similar Br/Cl, 87Sr/86Sr and 228Ra/226Ra ratios to those of FP waters from the Silurian shale formation (Figures 2 and 4). Likewise, the δ11B values of the FP waters overlap with those from the Permian formation waters (Figure 4). In contrast, the Triassic produced waters had distinctive geochemical characteristics for all of the four parameters examined in this study (i.e., distinctive higher Br/Cl, δ11B, and 228Ra/226Ra combined with lower 87Sr/86Sr) that are different from those of the FP waters (Figure 3). We conclude that it would be possible to distinguish conventional from unconventional oil and gas wastewaters and their impacts in some cases, particularly where the conventional gas wastewater is originated from Triassic produced waters. Yet in other scenarios such a distinction might be limited due to the geochemical overlaps between the FP water and the Sinian, Cambrian, and Permian produced waters (Figures 3 and 4). Overall, data from produced waters from conventional natural gas wells and from FP waters from shale gas wells in Sichuan Basin suggest that the formation waters originated from residual evaporated seawater, followed by extensive water−rock interactions. We distinguish between two types of evaporated seawater remnants: (1) an early phase with chloride up to ∼50 000 mg/L and seawater evaporation

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.est.8b04345. Data on the volume of the FP water and operational of three shale gas wells, as well as mass-balance calculations for estimating the volume of the return hydraulic fracturing water (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*A. Vengosh. E-mail: [email protected]. Phone: 919-6818050. Fax: 919-684-5833. ORCID

Avner Vengosh: 0000-0001-8928-0157 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors gratefully acknowledge funding from PetroChina Research Institute of Petroleum Exploration and Development, Beijing, China. We thank three anonymous reviewers for their comments and insights that improved the quality of this paper.

■

REFERENCES

(1) Vengosh, A.; Jackson, R. B.; Warner, N.; Darrah, T. H.; Kondash, A. A Critical Review of the Risks to Water Resources from Unconventional Shale Gas Development and Hydraulic Fracturing in the United States. Environ. Sci. Technol. 2014, 48 (15), 8334−8348. (2) Vengosh, A.; Mitch, W. A.; McKenzie, L. M. Environmental and Human Impacts of Unconventional Energy Development. Environ. Sci. Technol. 2017, 51 (18), 10271−10273. (3) U.S. Energy Information Administration Technically Recoverable Shale Oil and Shale Gas Resources: China; U.S. Department of Energy: Washington, DC, 2015. (4) Dong, D.; Wang, Y.; Li, X.; Zou, C.; Guan, Q.; Zhang, C.; Huang, J.; Wang, S.; Wang, H.; Liu, H.; Bai, W.; Liang, F.; Lin, W.; Zhao, Q.; Liu, D.; Qiu, Z. Breakthrough and prospect of shale gas exploration and development in China. Natural Gas Indust 2016, 3 (1), 12−26. (5) Dong, D.; Zou, C.; Dai, J.; Huang, S.; Zheng, J.; Gong, J.; Wang, Y.; Li, X.; Guan, Q.; Zhang, C.; Huang, J.; Wang, S.; Liu, D.; Qiu, Z. 14526

DOI: 10.1021/acs.est.8b04345 Environ. Sci. Technol. 2018, 52, 14519−14527

Article

Environmental Science & Technology Suggestions on the development strategy of shale gas in China. J. Natural Gas Geosci 2016, 1 (6), 413−423. (6) Chang, Y.; Huang, R.; Masanet, E. The energy, water, and air pollution implications of tapping China’s shale gas reserves. Res., Conserv. Recycl. 2014, 91, 100−108. (7) Zou, C.; Dong, D.; Wang, Y.; Li, X.; Huang, J.; Wang, S.; Guan, Q.; Zhang, C.; Wang, H.; Liu, H.; Bai, W.; Liang, F.; Lin, W.; Zhao, Q.; Liu, D.; Yang, Z.; Liang, P.; Sun, S.; Qiu, Z. Shale gas in China: Characteristics, challenges and prospects (I). Petrol. Explor. Develop. 2015, 42 (6), 753−767. (8) Zou, C.; Dong, D.; Wang, Y.; Li, X.; Huang, J.; Wang, S.; Guan, Q.; Zhang, C.; Wang, H.; Liu, H.; Bai, W.; Liang, F.; Lin, W.; Zhao, Q.; Liu, D.; Yang, Z.; Liang, P.; Sun, S.; Qiu, Z. Shale gas in China: Characteristics, challenges and prospects (II). Petrol. Explor. Develop. 2016, 43 (2), 182−196. (9) Zou, C.; Ni, Y.; Li, J.; Kondash, A.; Coyte, R.; Lauer, N.; Cui, H.; Liao, F.; Vengosh, A. The water footprint of hydraulic fracturing in Sichuan Basin, China. Sci. Total Environ. 2018, 630, 349−356. (10) Ma, X.; Xie, J. The progress and prospects of shale gas exploration and development in southern Sichuan Basin, SW China. Petrol. Explor. Develop. 2018, 45 (1), 172−182. (11) Wang, S. Shale gas exploitation: Status, problems and prospect. Natural Gas Indus. B 2018, 5 (1), 60−74. (12) Liu, P.; Feng, Y.; Zhao, L.; Li, N.; Luo, Z. Technical status and challenges of shale gas development in Sichuan Basin, China. Petroleum 2015, 1 (1), 1−7. (13) Warner, N. R.; Christie, C. A.; Jackson, R. B.; Vengosh, A. Impacts of Shale Gas Wastewater Disposal on Water Quality in Western Pennsylvania. Environ. Sci. Technol. 2013, 47 (20), 11849− 11857. (14) Lauer, N. E.; Harkness, J. S.; Vengosh, A. Brine Spills Associated with Unconventional Oil Development in North Dakota. Environ. Sci. Technol. 2016, 50 (10), 5389−5397. (15) Kondash, A.; Vengosh, A. Water Footprint of Hydraulic Fracturing. Environ. Sci. Technol. Lett. 2015, 2 (10), 276−280. (16) Kondash, A. J.; Albright, E.; Vengosh, A. Quantity of flowback and produced waters from unconventional oil and gas exploration. Sci. Total Environ. 2017, 574, 314−321. (17) Engle, M. A.; Rowan, E. L. Geochemical evolution of produced waters from hydraulic fracturing of the Marcellus Shale, northern Appalachian Basin: A multivariate compositional data analysis approach. Int. J. Coal Geol. 2014, 126, 45−56. (18) Rowen, E. L.; Engle, M. A.; Kraemer, T. F.; Schroeder, K. T.; Hammack, R. W.; Doughten, M. W. Geochemical and isotopic evolution of water produced from Middle Devonian Marcellus shale gas wells, Appalachian basin, Pennsylvania. AAPG Bull. 2015, 99 (02), 181−206. (19) Warner, N. R.; Darrah, T. H.; Jackson, R. B.; Millot, R.; Kloppmann, W.; Vengosh, A. New Tracers Identify Hydraulic Fracturing Fluids and Accidental Releases from Oil and Gas Operations. Environ. Sci. Technol. 2014, 48 (21), 12552−12560. (20) Balashov, V. N.; Engelder, T.; Gu, X.; Fantle, M. S.; Brantley, S. L. A model describing flowback chemistry changes with time after Marcellus Shale hydraulic fracturing. AAPG Bull. 2015, 99 (1), 143− 154. (21) Rosenblum, J.; Nelson, A. W.; Ruyle, B.; Schultz, M. K.; Ryan, J. N.; Linden, K. G. Temporal characterization of flowback and produced water quality from a hydraulically fractured oil and gas well. Sci. Total Environ. 2017, 596−597, 369−377. (22) Osselin, F.; Nightingale, M.; Hearn, G.; Kloppmann, W.; Gaucher, E.; Clarkson, C. R.; Mayer, B. Quantifying the extent of flowback of hydraulic fracturing fluids using chemical and isotopic tracer approaches. Appl. Geochem. 2018, 93, 20−29. (23) Zheng, Z.; Zhang, H.; Chen, Z.; Li, X.; Zhu, P.; Cui, X. Hydrogeochemical and Isotopic Indicators of Hydraulic Fracturing Flowback Fluids in Shallow Groundwater and Stream Water, derived from Dameigou Shale Gas Extraction in the Northern Qaidam Basin. Environ. Sci. Technol. 2017, 51 (11), 5889−5898.

(24) U.S. Geological Survey. National Field Manual for the Collection of Water-Quality Data, 2011 (https://water.usgs.gov/owq/ FieldManual/) (accessed August 1, 2018). (25) Tasker, T. L.; B, W. D.; Ajemigbitse, M. A.; Lauer, N. E.; Gusa, A. V.; Kuatbek, M.; May, D.; Landis, J. D.; Alessi, D. S.; Johnsen, A. M.; Kaste, J. M.; Headrick, K.; Wilke, F. D. H.; McNeal, M.; Engle, M.; Jubb, A. M.; Vidic, R.; Vengosh, A.; Warner, N. R. Accuracy of methods for reporting inorganic element concentrations and radioactivity in oil and gas wastewaters from the Appalachian Basin, U.S. based on an inter-laboratory comparison. Environ. Sci.: Processes Impacts 2018, DOI: 10.1039/c8em00359a. (26) Vengosh, A.; Starinsky, A.; Kolodny, Y.; Chivas, A. R.; Raab, M. Boron isotope variations during fractional evaporation of sea water: New constraints on the marine vs. nonmarine debate. Geology 1992, 20, 799−802. (27) Bąbel, M.; Schreiber, B. C. Geochemistry of Evaporites and Evolution of Seawater A2 - Holland, Heinrich D. In Treatise on Geochemistry, Second ed.; Turekian, K. K., Ed.; Elsevier: Oxford, 2014; pp 483−560. (28) Vengosh, A. Salinization and Saline Environments, Holland, Heinrich D. In Treatise on Geochemistry, Second ed.; Turekian, K. K., Ed.; Elsevier: Oxford, 2014; pp 325−378. (29) McCaffrey, M. A.; Lazar, B.; Holland, H. D. The evaporation path of seawater and the coprecipitation of Br- and K+ with halite. J. Sediment. Res. 1987, 57, 928−937. (30) Barbot, E.; Vidic, N. S.; Gregory, K. B.; Vidic, R. D. Spatial and Temporal Correlation of Water Quality Parameters of Produced Waters from Devonian-Age Shale following Hydraulic Fracturing. Environ. Sci. Technol. 2013, 47 (6), 2562−2569. (31) Chapman, E. C.; Capo, R. C.; Stewart, B. W.; Kirby, C. S.; Hammack, R. W.; Schroeder, K. T.; Edenborn, H. M. Geochemical and Strontium Isotope Characterization of Produced Waters from Marcellus Shale Natural Gas Extraction. Environ. Sci. Technol. 2012, 46 (6), 3545−3553. (32) Haluszczak, L. O.; Rose, A. W.; Kump, L. R. Geochemical evaluation of flowback brine from Marcellus gas wells in Pennsylvania, USA. Appl. Geochem. 2013, 28, 55−61. (33) Korte, C.; Jasper, T.; Kozur, H. W.; Veizer, J. 87Sr/86Sr record of Permian seawater. Palaeogeogr., Palaeoclimatol., Palaeoecol. 2006, 240 (1), 89−107. (34) Rowan, E. L.; Engle, M. A.; Kirby, C. S.; Kraemer, T. F. Radium content of oil- and gas-field produced waters in the northern Appalachian Basin (USA)Summary and discussion of data; U.S. Geological Survey: Washington DC, 2011. (35) Vengosh, A. H.D.; Vinson, D. S.; Dwyer, G. S.; Raanan, H.; Rimawi, O.; Al-Zoubi, A.; Akkawi, E.; Marie, A.; Haquin, G.; Zaarur, S.; Hirschfeld, D.; Ganor, J. High Naturally Occurring Radioactivity in Fossil Groundwater in the Middle East. Environ. Sci. Technol. 2009, 43, 1769−1775. (36) Birdsell, D. T.; Rajaram, H.; Lackey, G. Imbibition of hydraulic fracturing fluids inot partially saturated shale. Water Resour. Res. 2015, 51, 6787−6796. (37) Harkness, J. S.; Dwyer, G. S.; Warner, N. R.; Parker, K. M.; Mitch, W. A.; Vengosh, A. Iodide, Bromide, and Ammonium in Hydraulic Fracturing and Oil and Gas Wastewaters: Environmental Implications. Environ. Sci. Technol. 2015, 49 (3), 1955−1963. (38) Vengosh, A.; Kondash, A.; Harkness, J.; Lauer, N.; Warner, N.; Darrah, T. H. The geochemistry of hydraulic fracturing fluids. In 15th Water-Rock Interaction International Symposium, WRI-15; Marques, J. M., Chambel, A., Eds.; Elsevier, 2017; Vol. 17, pp 21−24. (39) Vengosh, A.; Starinsky, A.; Kolodny, Y.; Chivas, A. R.; Raab, M. Boron isotope variations during fractional evaporation of sea waterNew constraints on the marine versus nonmarine debate. Geology 1992, 20 (9), 799−802.

14527

DOI: 10.1021/acs.est.8b04345 Environ. Sci. Technol. 2018, 52, 14519−14527