Zhao-xian Zheng, Xiao-shun Cui*, Pu-cheng Zhu, Si-jia Guo

1 Institute of Hydrogeology and Environmental Geology, Chinese Academy of Geological Sciences, Shijiazhuang 050061, China.2 Key Laboratory of Groundwater Science and Engineering, Ministry of Natural Resources, Shijiazhuang 050061, China.

Abstract: Hydrogeochemical processes that would occur in polluted groundwater and aquifer system, may reduce the sensitivity of Sr isotope being the indicator of hydraulic fracturing flowback fluids (HFFF) in groundwater. In this paper, the Dameigou shale gas field in the northern Qaidam Basin was taken as the study area, where the hydrogeochemical processes affecting Sr isotope was analysed. Then, the model for Sr isotope in HFFF-polluted groundwater was constructed to assess the sensitivity of Sr isotope as HFFF indicator. The results show that the dissolution can release little Sr to polluted groundwater and cannot affect the εSr (the deviation of the 87Sr/86Sr ratio) of polluted groundwater. In the meantime, cation exchange can considerably affect Sr composition in the polluted groundwater. The Sr with low εSr is constantly released to groundwater from the solid phase of aquifer media by cation exchange with pollution of Quaternary groundwater by the HFFF and it accounts for 4.6% and 11.0% of Sr in polluted groundwater when the HFFF flux reaches 10% and 30% of the polluted groundwater, respectively. However, the Sr from cation exchange has limited impact on Sr isotope in polluted groundwater. Addition of Sr from cation exchange would only cause a 0.2% and 1.2% decrease in εSr of the polluted groundwater when the HFFF flux reaches 10% and 30% of the polluted groundwater, respectively. These results demonstrate that hydrogeochemical processes have little effect on the sensitivity of Sr isotope being the HFFF indicator in groundwater of the study area. For the scenario of groundwater pollution by HFFF, when the HFFF accounts for 5% (in volume percentage) of the polluted groundwater, the HFFF can result in detectable shifts of εSr (ΔεSr=0.86) in natural groundwater. Therefore, after consideration of hydrogeochemical processes occurred in aquifer with input of the HFFF, Sr isotope is still a sensitive indicator of the Quaternary groundwater pollution by the HFFF produced in the Dameigou shale of Qaidam Basin.

Keywords: Dameigou shale gas; Groundwater pollution; Hydraulic fracturing flowback fluids; Strontium isotope; Hydrogeochemical modelling

Introduction

Hydraulic fracturing flowback fluids (HFFF),which are produced by ubiquitous high-volume hydraulic fracturing, comprise injected and formation water released from the shale formations(Vengosh et al.2014). The toxic substances associated with HFFF have led to strong public concerns about their impact on groundwater during fracturing, storage, transport, and disposal (Huang et al. 2020; Mclntosh et al.2019; Zheng et al.2014). Therefore, a robust methodology to identify the HFFF spills is currently a key research topic with the global interest in shale gas extraction.Vengosh et al. (2015) reported that the HFFF produced from Marcellus shale gas are typically hypersaline (TDS＞50 000 mg/L) and characterized by a Cl-Na composition with high Br/Cl ratios(＞1.5×10-3). Chapman et al. (2012) stated that strontium isotopic ratios for HFFF from the Marcellus Formation fall within a narrow range of values (87Sr/86Sr = 0.710 148 to 0.712 119), and noticed that this isotopic range is distinct from most oil and gas brines in the Upper Devonian Venango Group associated with western Pennsylvania acid mine drainage. Warner et al.(2014) suggested that trace element ratios (B/Cl＞0.001, Li/Cl ＞0.002) and isotopic fingerprints(δ11B = 25‰ to 31‰, δ7Li = 6‰ to 10‰) for HFFF from the Marcellus and Fayetteville black shale formations were distinct in most water samples produced from conventional oil and gas wells. We observed that these studies identified the indicators and their characteristic values of the HFFF which are different from shallow groundwater and common pollution sources(such as traditional oilfield brines and coal seam water). However, few studies have focused on the hydrogeochemical processes that may occur after these indicators enter the aquifer with HFFF. The change of chemical composition and concentration of the HFFF-polluted groundwater can cause dissolution/precipitation, adsorption/desorption, cation exchange, chemical reaction,etc. and these processes may reduce the sensitivity of the HFFF indicators in the groundwater, and even lead to the HFFF indicators invalidate.

Our previous study reported that87Sr/86Sr can be used as one of the indicators to differentiate HFFF from fresh groundwater and conventional oil-field brine in the continental Dameigou Shale gas field in the northern Qaidam Basin (Zheng et al. 2017;Cui et al. 2020a). Therefore, the present study takes the Quaternary aquifer located in the above Dameigou Shale gas field as the pollution receptor of the HFFF. Quantitative hydrogeochemical model was constructed for HFFF-polluted groundwater to assess the sensitivity of Sr isotope as HFFF indicator. The results of this study can provide a methodology of forensic evaluation using Sr isotope for groundwater pollution by HFFF produced from the Dameigou shale in the northern Qaidam Basin.

1 Study area

The Chai Ye 1 (CY1) well, which is located in Yuqia fault sag in the northern Qaidam Basin(Fig. 1), is the first continental shale gas exploration well in northwest China, and its target stratum for exploration is the Dameigou Shale (Cui et al. 2020a; Cui et al. 2020b; Zhou et al. 2016).The Yuqia fault sag has an arid continental climate with low precipitation (8.4-87.7 mm/year) and high evapotranspiration (2 065-3 040 mm/year).The CY1 well site was placed on the stream terrace of the Naoer River, which is a seasonal river and is characterized by high TDS (＞5g/L) and Cl·SO4-Na. The surficial geology of the study site is dominated by unconsolidated gravel, coarse sand,and fine sand of the Holocene and Upper Pleistocene (Fig. 1). The depth of unconfined groundwater in Quaternary aquifer is generally less than 5 m and drains as multiple springs along the Naoer River. In the study site, the groundwater flow direction is consistent with the surface water flow direction, and basically flows from northeast to southwest to Yuqia River, which is the lowest terrain in the region.

2 Method

2.1 Sample collection

Two categories of water samples were used in this study: (1) five samples of HFFF from the CY1 well that were sampled from the first to seventh flowback day (CY1-1d to CY1-7d) to identify the characteristics of the hydrochemical compositions in HFFF which is the pollution source of groundwater; (2) two samples of spring water discharged from Quaternary aquifer in CY1 well site were collected before fracturing to identify the characteristics of the hydrochemical compositions in background groundwater. After the fracturing, the HFFF demonstrated an artesian flowback process until the seventh flowback day due to the internal pressure of the Dameigou Shale. The sample sites of background groundwater were set downstream of CY1 well in order to simulate a real scenario of HFFF leakage. All the samples were collected in August 2014.

In the field, samples were filtered through a 0.45 μm nylon filter into precleaned HDPE bottles with no head space. Samples for trace element analysis (e.g. Sr) were acidified to pH 2 with ultrapure HNO3. All samples were kept on ice while in the field and refrigerated in the lab at 4°C until the analyses were completed.

2.2 Sample analysis

Fig. 1 (a) The topography of the study area, and locations of the CY1 well and Quaternary spring (b) Geologic map of the study area

Major cations and minor elements were analyzed by a PerkinElmer Inductively Coupled Plasma Optical Emission Spectrometer (ICP-OES) Model Optima 8 300 (precision, ±1%). The HCO3concentration was determined by phenolphthalein titration.Major anions, except HCO3, were analyzed by a Thermo Scientific Dionex ICS-4000 (precision,±1%). The hydrochemical compositions were analyzed at the National Research Center for Geoanalysis, Chinese Academy of Geological Sciences.The charge balance error of the hydrochemical analysis results is less than 5%. Strontium isotopes were measured using an Isotopx Phoenix Thermal Ionization Mass Spectrometry (TIMS) at the Center of Analysis, Beijing Research Institute of Uranium Geology. The average87Sr/86Sr ratio of NIST SRM 987 over the period of these analyses was 0.710 244±0.000 015 (n=30). To facilitate comparison, the87Sr/86Sr ratio is converted to εSr.The εSris the deviation of the87Sr/86Sr ratio from that of modern seawater (0.709 169) in parts per 104.

3 Results and discussion

3.1 Characteristics of hydrochemical compositions and Sr isotope in HFFF and background groundwater

The hydrochemical and isotopic results for the HFFF and the Quaternary groundwater are given in Table 1. The HFFF are characterized by a Cl-Na composition with high TDS up to 10 056 mg/L in the seventh flowback day. The concentrations of Na, Cl, and HCO3in HFFF ranges from 1 683 mg/L to 3 467 mg/L, 3 486 mg/L to 5 228 mg/L, and 415 mg/L to 1 764 mg/L, respectively, and all these concentrations gradually increased during the flowback period. The Quaternary groundwater are characterized by a Cl·SO4-Na composition with lower TDS (6 500-6 791 mg/L) than that in the HFFF. The concentrations of Na (1 531-1 796 mg/L), Cl (1 949-2 017 mg/L), and HCO3(341-370 mg/L) in the Quaternary groundwater are also significantly lower than those of the HFFF. These reflect that Na, Cl, HCO3, and TDS can be indicators of the Quaternary groundwater polluted by the HFFF. However, the conventional oil brine(COB) which is produced from the Neogene and Paleogene conventional oil reservoir in the study area, is another major source of groundwater pollution (Zheng, 2019). The COB is also characterized by a Cl-Na composition with high concentrations of Na, Cl, HCO3, TDS (Li et al.2014). Moreover, molar ratios of Na/Cl and(Ca+Mg)/SO4in the HFFF (0.75-31.23) are similar to those in COB (0.81-1.61 and 9.34-214.90) (Li et al. 2014; Li, 2007), respectively. Although molar ratios of HCO3/SO4(0.40-16.27) in the HFFF are higher than those in the Quaternary groundwater(0.24-0.28) and COB (0.01-0.32) (Li et al. 2014;Li, 2007), this molar ratio in the HFFF would significantly decrease due to the wide occurrence of sodium sulfate on the surface of the study area when the HFFF enter the aquifer by vertical infiltration from the waste pool. Hence, these hydrochemical indicators cannot distinguish groundwater polluted by the HFFF or the COB.

Table 1 Hydrochemical and isotopic results of HFFF and Quaternary groundwater

Sr concentration of the HFFF have similar values with those of the Quaternary groundwater, but εSrof the HFFF (55.64-62.71) are considerably higher than those of the Quaternary groundwater (41.89-43.77). Moreover, εSrof the HFFF are also significantly higher than those of the COB (28.81-39.05) (Li, 2007). These suggest that Sr isotope can be an indicator of HFFF polluted groundwater in the study area.

3.2 Hydrogeochemical processes affecting Sr isotope in the HFFF-polluted groundwater

Hyrogeochemical processes occurred during/after groundwater pollution by HFFF are the main factors affecting Sr isotope in the HFFF-polluted groundwater in addition to hydrochemically conservative mixing of HFFF and background groundwater. Sr isotopes are not fractionated by processes such as biological metabolism, phase separation, chemical speciation, and evaporation(Douglas et al. 1995; Ma et al. 2018; Zhang et al.2020). Thus, the change of87Sr/86Sr ratio in the HFFF-polluted groundwater would be a result of the mixing of Sr from solid phase of aquifer media.Therefore, two potential water-rock interactions that release Sr from solid phase during the groundwater pollution should be considered when modeling the Sr isotope of HFFF-polluted groundwater:(1) Dissolution Carbonates and silicates are the most common Srbearing minerals in aquifer. Carbonates are not found in the Quaternary aquifer of the study area.However, the aquifer is rich in silicate minerals such as quartz and feldspar. Because quartz is basically insoluble in groundwater under natural conditions (Li et al. 2018), the Sr from the incongruent dissolution of feldspar by weathering is the main Sr source of dissolution occurred in the aquifer. The conservative mixing model of the HFFF with the Quaternary groundwater is established for the purpose of discussing the feldspar stability in the HFFF-polluted groundwater and aquifer media system. The concentrations of major ions and trace elements, εSrin the HFFF increased with time. Thus, in this model, the CY1-3d HFFF which have the median values of hydrochemical and isotopic compositions of the HFFF, were set as the groundwater pollution source. According to the conservative principle,the QG1 groundwater which has smaller difference in εSrfrom the pollution source, was set as pollution receptor. The PHREEQC simulations show that QG1 is located in the stable state of illite and kaolinite in weathering systems of k-feldspar and albite, respectively (Fig. 2). Also, with the increase of the mixing volume ratio of the HFFF in polluted groundwater, the HFFF-polluted groundwater change to stable state of K-feldspar and Nasemectite, respectively (Fig. 2). Although the QG1 is located in the stable state of anorthite in weathering systems of anorthite, the HFFFpolluted groundwater is in the stable state of Casmectite when the mixing volume ratio of HFFF is higher than 10% of the polluted groundwater(Fig. 2). These suggest that feldspar dissolution would occur in the HFFF-polluted Quaternary groundwater and its aquifer media system.Lasaga(1984) reported that the mean lifetime in years of 1 mm crystals of feldspars is 5.2×105in solution at 25℃ and pH is 5. And the dissolution rate was negatively correlated with the pH of the solution(Schott et al. 1981). However, as the mixing ratio of HFFF increases, the pH variations of HFFFpolluted groundwater are in the range of 7.0-8.0.This indicates that feldspar in the Quaternary aquifer is dissolved at very low rates. Consequently,little Sr from the dissolution of minerals could affect the Sr isotope composition of HFFF-polluted groundwater.

Fig. 2 Mineral stability diagrams in weathering systems of (a) k-feldspar, (b) albite, (c) anorthite for the HFFF (CY1-3d), Quaternary groundwater (QG1),and simulated polluted-groundwater

(2) Cation exchange

Sr can be released by cation exchange which is caused by the change of hydrochemistry during the pollution process of groundwater by HFFF, and the released Sr could affect the Sr isotope compositions of the HFFF-polluted groundwater. Clay minerals play a central role in cation exchange in the aquifer media. The Quaternary aquifer in the study area contains 0%-10% clay minerals such as kaolinite, illite, and smectite. Thus, the impact of cation exchange to Sr concentration and εSrshould be considered in the model for groundwater pollution. Na concentrations of the HFFF and Quaternary groundwater are higher than the other cations by one to two orders of magnitude. Therefore, the exchange of Na with the other cations would be the dominated process of cation exchange occurring in the HFFF-polluted groundwater and Quaternary aquifer system.

3.3 Model for Sr isotope in HFFFpolluted groundwater considering cation exchange

The complete pollution process can be divided into three steps to model the Sr isotope in the HFFFpolluted groundwater: (1) cation (mainly Na+, K+,Ca2+, Mg2+, and Sr2+in QG1 groundwater) exchange equilibrium between aquifer media and groundwater before pollution (Equation (1)); (2) consermixing for cations in the HFFF-polluted groundwater; (3) a new Sr exchange equilibrium between aquifer media and polluted groundwater as a result of the change in hydrochemistry during the pollution process of the HFFF and groundwater.

The species distribution at the cation exchange equilibrium in the first step is expressed as:

Where:KNaIis the cation exchange coefficient,which conforms to the Gaines-Thomas convention (Table 2), and can be checked from previous studies (Appelo and Postma, 2005); [Na+] and[I i+]1/iare the activities (mM) ofNa+andIi+in background groundwater;βIandβNais the equivalent fraction for ionI i+and Na+in solid phase of aquifer media; and CEC is the cation exchange capacity (meq/kg) of sedimentary material;meqNa-XandmeqI-Xiis the exchange capacity (meq/kg) of sedimentary material forNa+andI i+, respectively.

The CEC of aquifer media at Quaternary groundwater, containing up to 10% clay, is obtained by an empirical formula (Breeuwsma et al. 1986) that relates the CEC to the percentages of clay at near neutral pH:

ThemeqI-XiofI i+(K+, Ca2+, Mg2+, and Sr2+) in solid phase is obtained by substituting Equation (3)into Equation (2), and then converting ion activity to molar concentration (m) using the activity coefficient (γ):

Then, themeqNa-X, as well as eachmeqI-X, can be calculated using CEC:

Where:mmix Sr2+is the Sr concentration (mmol/L)of the conservative mixture of the HFFF and groundwater;mHFFF Sr2+andmQG1 Sr2+are the Sr concentration (mmol/L) of the HFFF and groundwater, respectively;xis the mixing volume ratio (%) of the HFFF in the HFFF-polluted groundwater;87Sr/86Srmixis the87Sr/86Sr of conservative mixture of the HFFF and groundwater;87Sr/86SrHFFFand87Sr/86SrQG1are the87Sr/86Sr of the HFFF and QG1 groundwater, respectively.

The conservative mixing in the second step was simulated using the Phreeqc (Version 3.4.0, USGS,Reston, USA) mixing module.

The final step is expressed as:

The released/absorbed Sr (ΔmSr2+) by cation exchange after the input of HFFF can be calculated by substituting Equation (10) into Equation (9) and using results of the first and second steps.

If the Sr concentration and isotope composition of two sources (Sr released by cation exchange and Sr in conservative mixing water) are known, their mixing ratio can be precisely calculated. The cation exchange is in dynamic equilibrium between aquifer media of Quaternary and groundwater.Thus, the Sr released by cation exchange had the same87Sr/86Sr ratio as background groundwater before HFFF pollution. Then, the Sr isotope ratio of mixing water for cation exchange is given by:

3.4 Sensitivity assessment of Sr isotope as indicator of HFFFpolluted groundwater

To assess the sensitivity of Sr isotope as indicator of HFFF-polluted groundwater, the groundwater pollution model has been calculated using the 3-day HFFF and QG1 groundwater as pollution source and pollution receptor, respectively. The results for the three steps of pollution processes for affecting Sr chemical and isotopic compositions in HFFFpolluted groundwater are given in Table 2 and Table 3. The simulation results show that the Sr is always released from solid phase of aquifer media by cation exchange when the HFFF pollutes the Quaternary groundwater. Also, the proportion of released Sr (0.1%-20.3%) in total Sr of the polluted groundwater continues to increase with the mixing volume ratio of the HFFF in polluted groundwater.Because input of Sr from cation exchange has lower εSr(43.77) than that of the HFFF (61.82), εSrof the polluted groundwater (43.79-57.43) consideringdering cation exchange is always lower than those of conservative mixture (43.79-60.92) of the HFFF reak and groundwater at different pollution intensities. This indicates that when the HFFF pollutes groundwater, the cation exchange occurred in Quaternary aquifer would reduce the effectiveness of Sr isotope as indicator of HFFF pollution. However, when the mixing volume ratio of HFFF in polluted groundwater is 10%, 30% and 95%, the difference of εSrbetween polluted groundwater (45.45, 48.45 and 56.82) and conservative mixture (45.53, 49.03 and 60.09) is less than 0.1, 1.0 and 4.0, respectively, and accounts for 0.2%, 1.2% and 5.4% of εSrin the conservative mixture, respectively. This suggests that the cation exchange has little impact on Sr isotope in slightly (HFFF volume proportion in polluted groundwater ＜10%) and moderately(HFFF volume proportion in polluted groundwater＜30%) polluted groundwater, and has limited impact on Sr isotope in severely polluted groundwater (HFFF volume proportion in polluted groundwater ＞50%) by HFFF.

Table 2 Cation exchange coefficients relative to Na+, cation molar concentrations and activities in QG1 groundwater, and exchange capacity for individual cation in Quaternary aquifer

Table 3 Summary of cation exchange process for calculating 87Sr/86Sr at different mixing volume ratios of HFFF proportions in polluted QG1 groundwater

Response curve of pollution intensity is also generated for the QG1 groundwater interacting with the 3-day HFFF (Fig. 3). For the scenario of groundwater pollution by HFFF, when the HFFF accounts for 5% (in volume percentage) of the polluted groundwater, the HFFF can cause detectable shift of εSr(ΔεSr=0.86) beyond the natural dynamic fluctuation ranges for the Quaternary groundwater; when the HFFF reaches 30% of the polluted groundwater, the εSrfor polluted groundwater would be significantly increased (ΔεSr=4.68).

Fig. 3 Response curve of Sr chemistry and isotope in groundwater under different pollution intensity

Note: The labeled tick marks on the curve indicate the volume proportion of HFFF in polluted groundwater.

4 Conclusions

(1) In the Quaternary aquifer of the study area where carbonate minerals are lacking, the inadequate dissolution of feldspar minerals is one of the main hydrogeochemical processes that can affect Sr isotopes in HFFF-polluted groundwater.Although feldspar dissolution would occur in the HFFF-polluted Quaternary groundwater and its aquifer media system, the very low dissolution rates of feldspar minerals lead to the result that little Sr from the dissolution could affect the sensitivity of Sr isotope as HFFF indicator in the Quaternary groundwater of the study area.

(2) Cation exchange is another major hydrogeochemical process which may affect Sr isotopes in HFFF-polluted groundwater. The Sr with low εSris constantly released to groundwater from solid phase of aquifer media by cation exchange when the HFFF enter the Quaternary aquifer in the study area. However, the shift in εSrof polluted groundwater resulting from the addition of Sr by cation exchange is at very low level. Therefore, cation exchange has little impact on the sensitivity of Sr isotope as HFFF indicator in the Quaternary groundwater of the study area.

(3) For the scenario of groundwater pollution by HFFF, when the HFFF accounts for 5% (in volume percentage) of the polluted groundwater, the HFFF can result in detectable shift of εSrin natural groundwater after consideration of cation exchange. This demonstrates that Sr isotope is a sensitive indicator of the Quaternary groundwater pollution by the HFFF produced in the Dameigou shale in the northern Qaidam Basin.

Acknowledgements

This study was supported by the National Natural Science Foundation of China (No. 41302192),Natural Science Foundation of Hebei Province of China (No. D2018504011), China Geological Survey (No. DD20190555), and the Ministry of land and resources of the People’s Republic of China (No. 201411052).