1. Introduction

Temporary spent nuclear fuel (SNF) storage facilities at nuclear power plants in South Korea are projected to be saturated soon [1], making the permanent disposal of SNF an urgent issue in this country. SNF is hazardous to humans and the environment due to its long-lasting high levels of radiation and heat, requiring isolation for hundreds of thousands of years. The International Atomic Energy Agency has proposed deep geological disposal as the optimal disposal method for SNF [2].

The deep geological disposal method refers to disposal of SNF in an underground facility constructed in a deep rock mass using a multiple-barrier system composed of engineered and natural barriers. The former is an artificial structure built in the deep rock mass with a lifespan of several to tens of thousands of years, while the latter is the host rock itself that protects the former. The role of the latter becomes critical in the event of failure of the former. Therefore, understanding the geological characteristics of the rock mass at the disposal depth is a prerequisite for the safe disposal of SNF.

A hydrochemical investigation using a deep borehole is an essential component in understanding the geological characteristics of an SNF disposal site. The hydrochemical characteristics are directly related to the long-term underground migration and retardation of radionuclides. The hydrochemical conditions determine the chemical speciation, solubility, and mobility of radionuclides [3]. In addition, they affect the corrosion of the canister encapsulating SNF, making hydrochemistry an essential factor to consider when evaluating the long-term stability of engineered barriers [4]. Deep hydrochemical investigation consists of field water quality measurements and sampling of deep groundwater, which often exists at discrete depths, while strictly restricting disturbance, mixing, and atmospheric contamination.

Crystalline rocks have been selected as the repository host rock in countries such as Finland and Sweden. Deep hydrochemical field measurements and groundwater sampling in crystalline rocks have been performed extensively in these countries. The methods used include pumping using inflatable packers, tube sampling, and pressurized sampling using a Leutert positive-displacement sampler or PAVE (pressurized water sampling equipment) device. Pumping via the expandable packer method installs a double-packer in the water-conducting fracture zone in a borehole. Then, sensors at the surface measure the hydrochemical properties of groundwater flowing into a flow-through cell while continuously pumping the water [5]. The hydrochemical field variables that are typically measured include temperature, dissolved oxygen (DO), redox potential (Eh), electrical conductivity (EC), and hydrogen ion activity (pH). The tube sampling method uses a tube-type sampler in an open borehole [6]. For example, this method involves insertion of a polyamide tube divided into sections at 50-m intervals (up to 1,700 m in length) into a borehole to collect groundwater by fastening each tube section with a valve that is manually operable at the surface [7, 8]. This is an easy, convenient, and quick method used to investigate broad hydrochemical changes over the entire depth of the borehole, rather than those at the specific individual waterconducting fracture zones.

In South Korea, a multi-packer system using a borehole drilled in the Yuseong area was used to monitor longterm hydrochemical changes [9]. The values of the hydrochemical variables in the borehole were measured, and the groundwater was collected using the measuring and pumping port located between the packers. However, the amount of groundwater that can be collected by this method is limited. In addition, the sample collector is moved to an anoxic chamber for hydrochemical analyses, making it a time-consuming process, and determining if and how much drilling water remains in the collected sample is not straightforward.

Granite is a representative type of crystalline rock in South Korea. The groundwater flow in granite occurs mainly through fracture zones rather than diffusion through the rock matrix [10]. Hence, performing hydrochemical field tests precisely in the water-conducting fracture zone is critical to investigate granite rock properties. However, few studies have investigated the deep hydrochemical properties of granite in South Korea. It is challenging to conduct a deep hydrochemical field investigation in an isolated waterconducting fracture zone without disturbing the water quality in other zones. Therefore, it is essential to develop deep hydrochemical field testing technology and techniques on the water-conducting fracture zones.

Here, we carried out a set of hydrochemical investigations in a borehole 750 m deep drilled in a granite body in Wonju using a closed system composed of a double-packer, Waterra pump, flow cell, and portable water quality unit. The investigation consisted of locating water-conducting fracture zones, field water quality measurements, water sampling, and water quality measurements in the laboratory. The hydrochemical investigation method used is described in detail, along with the results. Additional considerations and suggestions for future improvements are also discussed.

2. Study Site

We used a 750-m multipurpose investigation borehole drilled in Wonju Granite (Fig. 1). Wonju Granite was selected because it constitutes mineralogically and petrographically a typical granite body. Besides, its location allowed ready drilling and maintenance for long-term hydrochemical monitoring. Nonetheless, we point out that, although the study was conducted in the context of SNF geological disposal, it was strictly for research purposes, and the borehole location has nothing to do with an actual disposal site.

Fig. 1

The geologic map and the location of the drilling site in the Wonju area in South Korea [11].

Wonju Granite is fresh, coarse-grained, with uniform mineral composition, typical of Jurassic granite commonly found in Korea. Petrologically, it is classified as granodiorite. It did not go through a significant shear deformation. The surface survey suggested a wide distribution area (~50 × ~50 km). The borehole is located near the center of the granite body, providing rock core samples that can represent the characteristics of the rock body. The Wonju area is located to the west of the Taebaek Mountains. The middle of the northern part of Wonju is a rugged mountainous area composed of granite porphyry, quartz porphyry, and sedimentary rocks, whereas the southern part is mainly biotite granite [11]. In the north–central to north–south mountain range, the strike of the major joints in the granite body is parallel to the direction of the quartz grains [11]. The average elevations of the northern and southern areas are 350–400 and 120–150 m above sea level, respectively. Zircon U-Pb age dating using a sensitive high-resolution ion microprobe indicated that the granite in the Wonju area dates from 172.1 ± 1.5 Ma of Jurassic [12].

3. Deep Drilling in the Granite Rock

Considering their often complex and inhomogeneous nature, a multidisciplinary investigation using one borehole is beneficial to obtain an integrated understanding of crystalline rocks, such as granite [13]. In this granite deep drilling campaign, we drilled a borehole to a depth of 750 m and conducted a multidisciplinary investigation. The drilling depth of 750 m was decided based on the tradeoff of depth and expense. Assuming a final disposal depth between 300 and 500 m, a deeper depth will yield enough information for characterization and geoenvironmental modeling. However, the expense increases exponentially with depth.

In this drilling campaign, drilling down to 750 m was first completed. During the drilling process, a closed system was used to manage the drilling water labeled with a fluorescein dye (sodium fluorescein, uranine; referred to hereafter as dye). At depths greater than 520 m, an anionic polyacrylamide polymeric flocculant was added to the drilling water to collect the excessive amounts of debris in the water discharged from the borehole. After completion of drilling, multidisciplinary tests were performed in the borehole in the following order: geophysical well-logging, hydrochemical investigation, hydraulic tests, and geomechanical tests.

4. Hydrochemical Field Investigation Method

The hydrochemical investigation consisted of in situ measurements of the field hydrochemical parameters, groundwater sampling, and laboratory analyses. The dye added to the drilling water was used to identify the impact of drilling water mixing on the hydrochemical data. The drilling water data were also used to estimate the depths of the water-conducting fracture zones. Visual inspection of the rock core and geophysical well-logging data were then used to confirm the hydrochemical investigation depths.

4.1 Drilling Water Management

We designed and used a closed system composed of three interconnected storage water tanks (3 tons each) and a settling basin. The tap water labeled with 500 μg⸱L⁻¹ dye was used, recycled after settling, and reused in the drilling operation. The inflow and outflow drilling water were sampled and filtered through a 0.45-μm membrane, and the dye concentration was measured using a 10AU fluorometer (Fig. 2(a)). The dye was added daily as necessary to maintain it within 10% of the initial concentration of 500 μg⸱L⁻¹. The discharged drilling water was collected from the settling basin every hour, and the temperature, pH, and EC were measured using a portable multimeter (ProPLUS; YSI) and electrodes, along with the depth value. Groundwater from the borehole itself is a good drilling water source in that it minimizes perturbation of the chemical compositions of the groundwater. However, it is not as meaningful when the groundwater exists in discrete aquifers. In addition, it is not practical if there is insufficient groundwater in the borehole. In cases such as the current campaign, water with distinct chemistry can be used after labeling with a dye, and the dye concentration can be used to estimate mixing of the drilling water. In this study, we used tap water as the drilling water source.

Fig. 2

The schematics showing the hydrochemical investigation in the deep borehole: (a) the drilling water management system, (b) groundwater pumping system made of a double-packer and a Waterra pump, and (c) in situ measurements of hydrochemistry parameters and subsequent groundwater sampling.

4.2 Selection of Hydrochemical Test Sections

A successful hydrochemical investigation starts with proper selection of the investigation depths. To select the hydrochemical investigation depths, we determined the locations of the water-conducting fracture zones using the drilling water data, well-logging data, and visual inspection of rock cores. First, we estimated the sections of the water-conducting fracture zones based on the drilling water monitoring data for the dye concentration, EC, pH, and temperature. The depths with anomalies in the water quality parameter values were identified as a function of depth (Fig. 3). The dye concentration in the injected drilling water decreases when diluted by the groundwater in the fracture zone. Thus, a decrease in dye concentration indicates the location of water-conducting fractures. In addition, the rates at which EC, pH, and temperature values change are higher in the water-conducting fracture zones. After completion of drilling, EC, temperature, and natural gamma values were obtained from the geophysical borehole well-logging and used to confirm the estimated sections. The rock cores were also inspected visually to assess whether secondary minerals filled the fractures (Fig. 4). By synthesizing the results, a total of eight sections were selected for the hydrochemical investigation, i.e., 127.5–130.5, 213.7–216.7, 338.8–341.8, 487.8–495.8, 514.0–522.0, 577.5–585.5, 676.3–684.3, and 698.7–706.7 m.

Fig. 3

Continuous hydrochemistry monitoring of the drilling water collected every hour during drilling: (a) dye concentration (μg⸱L⁻¹), (b) pH, (c) EC (mS⸱cm−1), and (d) temperature (℃).

Fig. 4

Pictures of the rock core samples recovered from the depths of (a) 127.5–130.5 m, (b) 487.8–495.8 m, and (c) 577.5–585.5 m.

4.3 Hydrochemical Field Measurements and Groundwater Sampling

The pumping and sampling system was composed of a double packer (Bimbar Packer; Petrometalic), a packer presser (Rothenberger), packer parts (packer rod, sleeve), a valve cylinder, a Waterra pump (Solinst), a gasolinepowered tubing actuator (Waterra power pack, PP1), a data logger (Level TROLL; In-Situ), a flow cell (YSI), a tripod, a winch, XL pipe (inner diameters of 12, 15, and 32 mm), and a rotation speed meter. A schematic of the groundwater sampling system is shown in Fig. 2(b). First, sleeves were mounted on both ends of the packer rod corresponding in length to the test section (3 or 8 m). One sleeve was connected to the XL pipe with an inner diameter of 15 mm, to which a valve cylinder was attached, and the 32-mm-innerdiameter XL pipe was connected to the other side. Then, a data logger (Level troll; In-Situ) was fitted into the valve cylinder. The tripod was set such that its center aligned with the borehole center. Next, the winch was placed on a flat surface near the tripod with its wire attached to the upper sleeve. The tripod supported the wire, and the assembled double packer connected to the wire was inserted into the borehole. The double packer and packer rod were then lowered to the test section using a rotation speed meter corrected for length. Finally, the Waterra pump was connected to the XL pipe with an inner diameter of 12 mm.

The groundwater in the test sections was pumped out to the surface by the Waterra pump placed at a depth of 50 m using the tubing actuator following the procedure outlined below (Fig. 2(c)). First, one test section volume between the packers was pumped out to remove the remnant drilling water. Next, the water layers were continuously pumped up with the upper packer released. The predetermined volumes were purged before collecting the target groundwater from the test depths. Opening the packer facilitates pumping because of the faster recovery to the stable groundwater level. Although not an ideal process, packer release is justified based on the consideration that it accelerates pumping without leading to direct exposure to the atmosphere or disturbance from the water from other depths. The dye concentrations in the collected groundwater samples were used to back-calculate and assess the mixing from the residual drilling water. The field measurements and groundwater sampling were performed at eight depths (Table 1).

Table 1

The depth, number of purging days, purging start date, volume of the 15A XL pipe, volume between the packers, planned purging volume, and actual volume purged for each test section

The hydrochemical parameters, temperature, pH, EC, Eh, and DO were measured using a multimeter (ProPLUS; YSI) and electrodes, which were calibrated and used in accordance with the standard methods of the Ministry of Environment or the standard procedures of the US Environmental Protection Agency (EPA). A spectrophotometer (10AU Fluorometer; Turner Designs) was used to determine the dye concentrations in the groundwater. The hydrochemical parameters were measured when the continuously monitored Eh values stabilized within ± 5%. The groundwater samples were then collected using three 2-L water bottles per test section: two for duplicates and a third for storage. Another 125-mL bottle was acidified using a highpurity concentrated nitric acid (Optima) and used for cation analysis. The entire measurement and sampling process was performed using a sealed flow cell without exposure to the atmosphere. The collected groundwater samples were stored at 4°C in the dark and then transported to the laboratory on the same or the following day via a courier service.

4.4 Hydrochemical Laboratory Measurements

The collected groundwater samples were filtered using 0.45-μm membrane (Advantec®) filters in an anaerobic chamber filled with nitrogen in the laboratory. The filtered groundwater samples were measured for dissolved main components (cations and anions), trace elements, rare-earth elements, alkalinity, dissolved organic carbon, sulfide, ferrous, DO, and dye concentration.

The concentrations of major and minor cationic elements (K, Na, Ca, Mg, Si, Sr, Fe, Mn, Cu, Pb, Zn, Al, Sc, Ti, V, Cr, Co, Ni, and As) in the filtered groundwater samples were measured using an inductively coupled plasma-optical emission spectrometer (ICP-OES, Optima 8300; PerkinElmer) at the Korea Institute of Geoscience and Mineral Resources (KIGAM), according to the US EPA 6010C (2007). Trace elements (Cs, Ce, and U) were measured using a high-resolution inductively coupled plasmamass spectrometer (HR-ICP-MS, Element 2; ThermoFisher Scientific, Inc.) at the KIGAM. The concentrations of major anions (B^r−, PO₄³⁻, F⁻, Cl⁻, NO₃⁻, and SO₄²⁻) were determined by ion chromatography (IC, ICS-6000; Dionex) at the KIGAM, following US EPA 300.1 (1997) as the standard analytical method.

Alkalinity was titrated with 0.01 N hydrochloric acid using a pH meter and glass body combination electrode (Orion Versastar and Thermo Orion; ThermoFisher Scientific, Inc.) following the Gran titration method. The groundwater sample with a known volume was transferred to an Erlenmeyer flask, and the initial pH was measured after rigorous mixing on a stir plate. The groundwater was then titrated down to pH 4 by incremental addition of hydrochloric acid with continuous stirring. A linear plot was constructed based on the amount of acid added, and pH was measured to determine the alkalinity. Dissolved organic carbon was determined using a Vario TOC cube (Elementar; Chungnam National University Center for Research Facilities). A platinum catalyst was used at a furnace temperature of 850°C. The amount of sample injected was 0.2 mL, and the pH was adjusted within the range of 3–12 using HCl for total inorganic carbon removal. The non-purgeable organic carbon method was used. DO, ferrous iron, and sulfide levels were determined using the Multi-Analyte Photometer (V-2000; CHEMetrics) with the DO test kit (K-7553; CHEMetrics), iron kit (K-6203; CHEMetrics), and sulfide kit (K-9523; CHEMetrics), respectively.

5. In Situ Hydrochemical Parameters

The EC, Eh, pH, and temperature field measurements for the eight test sections are shown in Table 2. These values were measured when the Eh values were most stable (Table 2). In addition, the changes in Eh with time for the three test sections (127.5–130.5 m, 577.5–585.5 m, and 698.7–706.7 m) are shown in Fig. 5, while Fig. 6 shows the changes in DO with time for the two representative test sections (127.5–130.5 m and 698.7–706.7 m).

Table 2

Field measured hydrochemistry parameter values of the groundwater samples collected from each test depth

Fig. 5

Eh changes with time for the three representative test sections: (a)127.5–130.5 m, (b) 577.5–585.5 m, and (c) 698.7–706.7 m.

Fig. 6

DO variations with time for the two representative test sections: (a) 127.5–130.5 m and (b) 698.7–706.7 m.

Contrary to our expectations based on depth considerations, the Eh values were positive along the depths of the borehole (Fig. 5). The remaining drilling water indicated by the dye concentrations was likely to be at least partially responsible for the positive values. For example, the Eh values remained positive even after pumping out 385 L over 6 days, which is equivalent to three times the volume of the depth interval between 577.5 and 585.5 m. The corresponding dye concentrations in the water samples were approximately 15 ppb, indicating remnant drilling water. The large purging volume at this depth suggests that the inflow of the drilling water into the fractures was relatively fast. With continued pumping, the Eh value decreased to converge to +120.1 mV, although there were still some slight variations. Other depth intervals showed similar trends. At the deepest interval (698.7–706.7 m), the Eh value decreased gradually to +174.5 mV (Table 2) with continued pumping. Pre-pumping removed less than one section volume from this interval, likely due to slow groundwater flow. It is noteworthy that the dye concentration in this section was 309 ppb (Table 3). Considering the initial dye concentration of 500 ppb in the drilling water, this value indicated that the drilling water was the main constituent in the water sampled from this section. The field-measured parameter values from the two deepest intervals were significantly affected by the drilling water, likely due to the added flocculant chemical. Hence, the subsequent laboratory-measured water quality data from these intervals are less representative of the water quality from these depths.

Table 3

The water quality data of groundwater samples from the eight test sections

The DO variation with time was greater than expected (Fig. 6(a)). When the groundwater was collected using the gasoline-powered tubing actuator, the inflow pressure seemed to be larger than ideal, and it subsequently interfered with the DO probe performance, which is sensitive to the flow rate. To compensate for this effect, we replaced the meter with a newer model and the flow cell with a larger model. Nonetheless, the DO value measured at the depth of 698.7–706.7 m did not change markedly and tended to decrease with pumping, similar to the Eh value (Fig. 6(b)).

Table 2 shows no significant changes in temperature in all eight sections. Temperature decreased gradually with depth. However, temperature was measured on the surface where the external temperature cannot be controlled. Hence, the temperature of the collected water was strongly correlated with the air temperature on the day of collection.

The pH of sections down to 577.5–585.5 m ranged between 6.87 and 7.63 (Table 2). In contrast, pH was distinctly higher at depths of 676.3–684.3 m and 698.7–706.7 m. However, we interpreted these observations as artifacts caused by the polyacrylamide polymeric flocculant used during the drilling process rather than an actual phenomenon.

There were no significant differences in the EC values in the sections investigated (Table 2). However, they were slightly higher in the 676.3–684.3 and 698.7–706.7 m sections than in the others. We interpreted this as also being due to the anionic polyacrylamide polymeric flocculant used during the drilling process.

6. Laboratory Water Quality Results

Table 3 shows the hydrochemical results of the groundwater samples measured in the laboratory. The charge balance values were obtained from the standard percent difference formula (Table 3). The percent difference was within 10% for the hydrochemical analyses of groundwater from depths of 127.5–130.5, 213.7–216.7, 487.8–495.8, 514.0– 522.0, and 577.5–585.5 m. On the other hand, it was higher in the sections at depths of 338.8–341.8 m, 676.3–684.3 m, and 698.7–706.7 m. These observations indicated that the error involved in the groundwater analysis data from these three sections was significant, and hence caution is required when using these data. The pumped water was less than one section volume from the section at depth 338.8–341.8 m, which may have affected the quality of the ion analysis results (Tables 1 and 3). The dye concentrations in the groundwater samples from 676.3–684.3 and 698.7–706.7 m were 241 and 309 ppb, respectively, indicating a significant contribution of the drilling water remaining in these sections. The drilling water used at these depths, in contrast to that at the other depths, was amended with polyacrylamide polymeric flocculant. However, the chemical composition of the polyacrylamide polymeric flocculant is a trade secret and has not been disclosed by the manufacturer. This observation suggests that additives to drilling water should be avoided as much as possible as they significantly affected the accuracy of the data.

7. Discussion

7.1 Locating the Water-conducting Fracture Zones

Identifying the location of water-conducting fracture zones is a critical step in conducting hydrochemical investigations. The in situ water quality parameters, pH, EC, and dye concentration of the drilling water measured every hour tended to show anomalies when drilling passed through a fracture zone (Fig. 3(a)). As the drilling water was collected and measured from the surface and was prone to be affected by the ambient temperature even in the closed system, the temperature did not reflect the fracture zones. We used the drilling rate to estimate the exact depth of the fracture zones from the surface-measured water quality data. We also used the visual inspection of rock cores and well-logging data to confirm the water-conducting fracture zones. The dye concentration gave the quantity of the residual drilling water mixed in the groundwater samples collected via an inverse calculation.

7.2 Hydrochemical Field Measurement Method

The groundwater sampling method used in this study requires a long time for the pre-sampling pumping period to remove the drilling water and groundwater from other depth zones. For example, in the section at depth 127.5– 130.5 m, pumping was carried out for a total of 8 days to remove 21.3 L (Table 1). From the depth of 338.8–341.8 m, 33.3 L was pumped out over 6 days (Table 1). There were sections in which the rate of groundwater recovery to a stable level was not fast enough for the waters from outside of the test sections to be removed by the Waterra pump located at 50 m. Hence, this configuration is more favorable when the groundwater flow rate is relatively fast.

The Eh values of the groundwater samples showed positive values (Table 2). In addition, the CHEMetrics spectrophotometer results demonstrated the presence of ferrous iron in the sections at 338.8–341.8, 487.8–495.8, 514.0– 522.0, 676.3–684.3, and 698.7–706.7 m (Table 3). The dye concentrations showed that a large amount of drilling water remained in the sections at 676.3–684.3 and 698.7–706.7 m. However, spectrophotometric analysis indicated that these sections contained 0.254 and 0.594 ppm Fe(II). This implies that the oxygen in the drilling water was likely to be consumed by the ferrous iron present in the section. In addition, spectrophotometric analysis demonstrated the presence of dissolved sulfide in all samples (Table 3).

7.3 Redox Conditions Along the Depths of the Borehole

The Swedish nuclear fuel and waste management company SKB [14] regards the absence of DO as a mandatory requirement for a site to host a high-level radioactive waste repository, meaning that the investigation must be stopped and the site abandoned if this condition is not satisfied. They consider a negative Eh value, presence of Fe(II) or S(-II), as indicating the absence of DO. Hence, the positive Eh and dysoxic DO in this study yielded contradictory evaluations against the presence of ferrous iron and sulfide ions in terms of whether the deep hydrochemistry in this model granite borehole has favorable qualities as a repository host rock. In contrast to common sense-based expectations, DO values ranged between 0.67 and 2.31 ppm, along the borehole depth from 127.5 to 706.7 m. Consistent with this, the measured Eh values were all positive, ranging from 75 to 302 mV.

However, the DO and Eh values must be handled with caution to avoid prematurely concluding that the hydrochemical conditions are not acceptable for deep geological disposal even at 700 m. The dye concentration provides useful information for interpreting these data. For example, the highest DO value was observed in the water sample from the deepest section, which is counterintuitive. Here, the corresponding dye concentration was 309 ppb. Assuming no dye loss, such as by sorption onto the geomedia or by dilution during pumping, we can conclude that the drilling water constitutes as much as 62% of the collected groundwater sample. Furthermore, if we assume a DO level of 8 ppm in the drilling water with no oxygen in the groundwater, we get up to 5 ppm in the groundwater sample collected due to the drilling water contribution. Hence, the measured DO of 2.31 ppm in the sample from 698.7–706.7 m does not necessarily indicate that this much DO is present in the unperturbed groundwater. Although based on a simplified calculation, this example demonstrates that DO requires multiple lines of evidence for application to siting processes. Second, the Fe(II) and S(-II) concentrations provide independent information regarding underground redox conditions. The dissolved Fe(II) in the two deeper samples ranged from 0.25 to 0.59 ppm, whereas S(-II) ranged from 0.59 to 0.66 ppm. These values, along with the dye concentrations, suggest that the oxygen in the drilling water would have been consumed by reaction with Fe(II) and S(-II), leading to a lower DO value. To improve the measured data quality and reduce uncertainties involved in transportation and handling of samples, we recommend in situ spectrophotometric determination of DO, Fe(II), and S(-II) at the field site simultaneously with groundwater sample collection.

The groundwater pumped to the surface in a closed circuit was passed through the flow cell and transferred to the sampling bottle. Various probes for hydrochemical measurements were inserted into the flow cell to measure the water quality parameters. Among them was the DO probe. As shown in Fig. 6, the measured DO value showed a wide degree of variation. The DO probe used in this study was an optical sensor type. According to Wei et al. [15], this type of probe is sensitive to the flow rate. Considering the relatively small volume (203 mL) of the flow cell used in this study, it seems that the pulse-type water inflow to the cell resulting from the Waterra pump operation was applying instantaneous pressure on the probe. The probe responded sensitively to the shock, leading to DO readings susceptible to the momentary flow rate changes. To compensate for this effect, we replaced the flow cell with a larger model, after which the DO values at 676.3–684.3 m and 698.7–706.7 m showed relatively stable trends. The DO probe likely received fewer fluctuations and thus gave relatively stable readings. To use the DO value as a reliable hydrochemical parameter, we recommend a design that prevents rapid flow rate changes in the cell. Alternatively, a DO probe that is less sensitive to flow rate would be favorable [15]. Spectrophotometrically determining the in situ DO concentrations in a portable anaerobic chamber will be worth considering, to complement the probe measurements.

7.4 Comparisons With Other Methods

The multi-packer system referred to by Bae et al. [9] is a method for long-term hydrochemical monitoring of multiple test sections in a borehole. This method uses a closed system of a set of multiple-packers with sampling ports in each test section. This method is relatively complex to install and is considered more of a permanent installation. Collecting large volumes of groundwater at once is difficult because of its design with sampling ports inside each section. Hence, this method is inefficient for investigating deep underground chemical characteristics within a short period. Granite is usually inhomogeneous in nature [13], and understanding its properties generally requires multiple borehole drillings with ideally multidisciplinary investigations in a single borehole. Thus, the duration of hydrochemical investigations is a crucial factor to consider under an often pressing schedule due to other tests. It should also ensure that potentially prolonged exposure to the atmosphere does not perturb the properties. The method used in this study was considered more suitable for conducting investigations to understand the chemical characteristics for a siting process because of its rapid and ready implementation. In addition, it has the advantage of enabling the collection of large volumes of groundwater at once.

The tube sampling mentioned by Nurmi and Kukkonen [6] is a method of inserting a tube-type sampler with multiple sampling ports into an open borehole and simultaneously collecting samples by depth. This method is considered suitable for use at sites where the water-conducting fracture zones have slow groundwater flows.

The U-tube sampler described by Liu et al. [16] collects groundwater by injecting non-reactive nitrogen gas into a test section in a closed system created by packers. Suction from a vacuum creates a driving force for groundwater sampling. The U-tube sampler can obtain fluid samples while minimizing fluid flow and keeping it at a slow, constant pressure to reduce the effect on the flow field so that it is suitable for monitoring requiring accurate fluid analysis. However, the samplers are fragile, while the lines are prone to be blocked when fluid turbidity is high [16]. The method has restricted sampling intervals because of the limits on the suction pressure. Compared with the U-tube, the approach used in the present study is applicable to various geological environments with few variables affecting groundwater collection, such as turbidity and interval length.

The deep hydrochemical investigations performed in this study provided insights into future improvements. First, good drilling water management, which uses no amended chemicals, is necessary for developing an investigation borehole. Simultaneously, relevant technologies and techniques are required to stabilize the borehole walls without influencing the hydrochemistry. In this study, the added anionic polyacrylamide flocculant influenced the groundwater quality data. Second, improvements in pump design are required to enable continuous groundwater pumping to the surface or in situ collection at the test intervals, especially when the groundwater flow is slow. Third, an improved flow cell design with an alternate type DO sensor would give more stable and reliable DO values. Finally, in situ spectrophotometric measurements of DO, Fe(II), and S(-II) will achieve independent, reliable redox-indicating parameters to complement the electrode-determined Eh and DO while suppressing the uncertainties involved in the sample logistics.

8. Conclusions

We carried out hydrochemical investigations using a 750-m borehole drilled in a granite body in Wonju, South Korea. We used a closed system that recycles and reuses the drilling water discharged out of the borehole to manage the drilling water. Every hour, we monitored the in situ hydrochemical parameters of the discharged drilling water, including pH, EC, temperature, and dye concentrations. Based on the data, we initially estimated the depth of the water-conducting fracture zones. Then, we used the well-logging data and visual inspection of the rock cores to determine the investigation depths. We installed a doublepacker system in the test sections to measure in situ water quality parameters and subsequently collected groundwater samples. In addition, we used the dye concentrations to quantify the amount of drilling water remaining in the collected groundwater samples.

The groundwater sampling system used in this study requires sufficient pumping time to promote near-complete removal of the drilling water from the test sections. However, the groundwater recovery rate was not fast enough for some test sections, leading to remnant drilling water. Nonetheless, the presence of Fe(II) and S(-II) and maintenance of the packer pressure demonstrated that the water sampling system functioned adequately and indicated that the deep geochemical conditions were promising in terms of siting. Reliable DO measurements require an improved flow cell design to reduce abrupt flow rate changes and subsequent wide variations in the measured values. This study demonstrated that the investigation must be planned carefully considering the drilling schedule and in coordination with other activities, such as drilling water management, hydrology tests, borehole logging, and hydrofracking. Therefore, concerted efforts are required among drillers, hydrogeologists, and hydrochemists.