1. Introduction

A deep geological repository (DGR) has been considered the preferred technology in many countries for the disposal of high-level radioactive waste (HLW) [1,2]. DGRs consist of a multi-barrier system, including a metal container, engineered barrier, and natural barrier system, to maintain the physical integrity of spent fuel, accompanying the prevention of radionuclide migration [1]. In the engineered barrier system, compacted bentonite has been considered a potential buffer material that prevents the migration of radionuclides from spent nuclear fuel [3,4]. However, colloids can be generated by bentonite erosion due to the inflow of groundwater through the long-term evolution of the DGR system [4]. In addition, the colloid plays a role as a carrier to migrate radionuclides from the repository to the biosphere because radionuclides can be strongly adsorbed onto the surface of bentonite colloids [3,4]. Therefore, the assessment of the colloid-facilitated radionuclide transport in the underground environment should be conducted to validate the safety of the DGR system.

Several countries have studied in-situ migration experiments using tracers and colloids to investigate colloidfacilitated radionuclide transport in the underground environment. In particular, the National Cooperative for the Disposal of Radioactive Waste (Nagra) in Switzerland has performed in-situ colloid-facilitated migration experiments in the Alps since 1985 [5]. The influence of colloids on the radionuclide migration in crystalline rock has been investigated by sequential research projects, namely Colloid and Radionuclide Retention (CRR, 1998–2003) [6] and Colloid Formation and Migration (CFM, 2004–present) [5]. In Sweden, the mobility of radionuclides was investigated by the Tracer Retention Understanding Experiment (TRUE) in the Äspö Hard Rock Laboratory [7,8]. In this experiment, a dipole test using bentonite colloids was carried out to investigate the colloid mobility and stability in a Quarried Block [8]. In addition, a series of tracer tests for migration and diffusion of radionuclides were carried out for the safety assessment of the disposal system [7].

In South Korea, Korea Atomic Energy Research Institute (KAERI) has constructed a small-scale underground research facility, the so-called KAERI Underground Research Tunnel (KURT), for the fundamental research of HLW disposal [9]. However, the KURT has weak points due to the low depth (120 m) and the difficulty of using radionuclides [10]. Therefore, it is required to utilize foreign underground research laboratories to supplement the weak points of KURT. Based on the preliminary investigation, KAERI started cooperating with Nagra in 2008 because the research on the colloid-facilitated radionuclide transport was well established (CRR and CFM) in the Grimsel Test Site (GTS). Therefore, KAERI has joined the CFM project to investigate in-situ colloid-facilitated radionuclide transport within the CFM Phase Ⅱ since 2008. Consequently, KAERI has obtained the research outcomes and the experience of the URL operation from the CFM project for future research in the Korean URL.

This review discusses the current status of CFM international joint research in GTS, Switzerland. In particular, the results from two specific field studies, the long-term insitu test (LIT) and in-rock bentonite erosion test (i-BET), are briefly noted in this review. Hence, some suggestions would be offered for colloid research and domestic URL program in South Korea.

2. Research Status of CFM

The Grimsel Test Site (GTS) is located at an altitude of 1,730 m in the crystalline rock at the Aar Massif of central Switzerland. The depth of the GTS is approximately 450 m beneath the summit and is connected by an access tunnel [11]. The GTS tunnel is approximately 1 km long and was excavated using a tunnel boring machine [12]. The GTS has been operated as a center for underground research and development laboratories, supporting a wide range of geological disposal of HLW in stepwise phases. The sixth phase of the GTS research started with more than 17 international partners in 2003; this phase is focused on the implementation and demonstration of long-term safety in the repository system [5,13] (Fig. 1). In GTS, in-situ experiments using radionuclides are available because of the presence of an IAEA level B radiation-controlled zone; hence, it is possible to conduct in-situ research using radionuclides [14].

Fig. 1

Map of GTS with the location of the major experiments [13].

The CFM project is ongoing international joint research with in-situ colloid-facilitated radionuclide transport conducted through the CRR of the GTS. It is dedicated to studying colloid formation by the erosion of bentonite buffer and investigating the transport of colloid-facilitated radionuclides in a fractured host rock. The aims and methods of CFM are summarized in Table 1 [15]. CFM was initiated in 2004 and is in the fourth research phase (2019–2023). In the first phase of the CFM project, site selection, characterization of geology, and preparation of the in-situ experiment were mainly carried out by Nagra. In particular, the objective of the research in phase I was to investigate the feasible water-conducting zone in the natural flow field of the GTS. From the examination, it was found that the migration shear zone (MI shear zone), that is located at the tunnel position of AU96, properly met the requirements of the in-situ experiment (Fig. 2). In this phase, several preliminary techniques such as cleaning the rock surface by sandblasting and pressure washing were carried out. Subsequently, the shear planes and entire tunnel surfaces were sealed by cement and resins, and two boreholes, CFM 06.001 and CFM 06.002 were drilled for testing in 2006. An observational inspection with a borehole camera showed that CFM 06.001 presented a lamprophyre intrusion, whereas no distinct fractures were observed in CFM 06.002. The hydraulic conductivity of CFM 06.001 was 10⁻⁹ m²·s⁻¹, that is slightly lower than that of the MI shear zone (10⁻⁸ and 10⁻⁶ m²·s⁻¹) [13,16].

Table 1

Aims and Method of CFM project [15]

Fig. 2

Map of the MI shear zone [13].

CFM Phase II research was conducted from 2008 to 2013 to prepare for long-term research on the colloid-facilitated radionuclide transport, the so-called Long-term In-situ Test (LIT). In this phase, a tracer test (Run 09-01) was carried out to investigate the natural flow field, the result showed that CFM 06.001 was an appropriate research site for the migration of colloids owing to the distinct flow, connectivity of the flow path in the shear zone, and suitability for overcoring [13]. The megapacker system, including rubber packers and the control system, was optimized in April 2010 (Fig. 3) [13]. The rubber packer successfully provided pressure control in the annular space between the packer and the wall of the tunnel during water level monitoring. To monitor the experiments, three monitoring boreholes, CFM 11.001, CFM 11.002, and CFM 11.003, were drilled and equipped with a triple-packer system in 2011. To perform the preliminary test of colloid-facilitated radionuclide transport, mobile laser-induced breakdown detection (LIBD) units were installed at the site by the Karlsruhe Institute of Technology (KIT).

Fig. 3

Illustration of the long-term in-situ test (LIT) experiment in the megapacker system [13,15].

CFM Phase III began in January 2014, and the primary purpose of this phase was the final preparation for the LIT experiment. For 4.5 years, the colloid-facilitated radionuclide transport through bentonite erosion was monitored. The overcoring of boreholes was finally conducted in phase Ⅳ (2019–2022). The details of the experimental setup and the results are summarized in the next chapter. In addition, a new in-situ experiment was designed to investigate bentonite erosion at the JGP site in CFM Phase III. After the characterization of the hydraulic field, two boreholes, JGP 11.003 and 11.001, were drilled, and double-packer systems were installed. Multiple sensors have monitored the bentonite source to detect the total pressure, pore pressure, relative humidity, and geochemical parameters, including fluorometers, pH, EC, Eh probes, and turbidity meters, and the monitoring of bentonite erosion is still ongoing.

3. Long-term In-situ Test

3.1 Overview

The primary purpose of LIT was to investigate colloidfacilitated radionuclide transport by an in-situ experiment. The LIT consists of a compacted bentonite source with radionuclide- containing vials in the water-conducting shear zone (MI shear zone). The LIT was initiated in 2014, and the swelling and erosion of the compacted bentonite generated colloids that migrated the radionuclides via the hydraulic gradient in the shear zone. The chemical properties and characteristics of colloids in groundwater were continuously monitored for 4.5 years and frequently sampled from three near-field monitoring boreholes and the tunnel wall outflow.

3.2 Experimental Layouts

3.2.1 MI Shear Zone

Detailed geological mapping is illustrated in Fig. 4 [13]. The MI shear zone in the GTS has a south–north strike and a steep SSE dip. The main foliation at this site was filled with feldspar and biotite. The average orientation of the foliation and shear planes was 155/75°. The MI shear zone has a twodimensional structure with a thickness of 0.16 and 0.90 m from the main access tunnel [6]. The MI shear zone was presumed to have undergone multiple deformation phases. In the early deformation stage, ductile structures are overprinted by brittle deformation, characterized by cataclastic overprinting, resulting in the formation of cohesionless fault gouge material filling the fault cores. The brittle horizontal feature in the MI shear zone has three steep (> 70°) shear planes with a south-north strike [6]. The main shear plane located in the north had older ductile features overprinted by a high degree of brittle deformation. The hydraulic properties of the main shear plane can be characterized by clear water flow conditions where surface packers have been installed, and the average transmissivity of the entire shear zone was very low (between 10⁻⁸ and 10⁻⁶ m²·s⁻¹) [6].

Fig. 4

Detailed structural geological mapping of the tunnel expression of the MI shear zone [13].

3.2.2 FEBEX Bentonite Emplacement in CFM 06.002

Initially, CFM 06.002 was drilled in October 2006 to investigate its hydraulic properties and tracers using a triple- packer system. Based on these boreholes, conservative tracer and radionuclide/colloid tests were performed from 2007 to 2012 (Tracer Test Run 12-02). In addition, three monitoring boreholes, CFM 11.001, CFM 11.002, and CFM 11.003, were drilled, and the boreholes were approximately parallel to CFM 06.002 (Fig. 5).

Fig. 5

Central and monitoring boreholes in CFM [13].

Sixteen rings of compacted bentonite for the LIT were emplaced in the packer system, as shown in Fig. 6. The thickness, dry density, and water content of the ring are approximately 0.5 mm, 1.63–1.66 kg·m⁻³, and 13.9wt%, respectively [13]. The four bentonite rings in the center of the packer system were emplaced by the mixed bentonite source using 10% Zn-labeled synthetic montmorillonite, that was mixed with 90% FEBEX bentonite [13]. Four central rings were positioned adjacent to the target fractures. Each bentonite ring contained four glass vials of clay paste with conservative tracers and radionuclides, and the volume of each vial was approximately 8 mL. Before mounting the bentonite to the mandrel, the weight of each ring was measured after grinding and drilling, and the final weight of the 16 rings was 2,613.56 g [13]. The effective dry density of bentonite is 1.34 kg·m⁻³, with an expected swelling pressure of 0.77 and 1.28 MPa. Assuming that the effective stress is equivalent to the swelling pressure would correspond to a dry density between 1.25 and 1.4 kg·m⁻³ [13].

Fig. 6

Bentonite source in the packer system [13].

3.2.3 Radionuclides

As noted in Section 3.2.2, radionuclides were injected into the vial as tracers during the LIT experiment. The emplaced radionuclides are summarized in Table 2 [13]. Initially, the bentonite suspension was equilibrated with strongly bound radionuclide tracers, such as ²⁴¹Am(III), ¹³⁷Cs(I), and ²⁴²Pu. The suspension was agitated for 26 h, and the solid phase was separated by centrifugation at 2,800 × g for 60 min. The paste was then stored in an Ar glovebox, and the weakly bound radionuclides ⁹⁹Tc(VII), ⁷⁵Se(VI), ⁴⁵Ca(II), ²³³U(VI), and ²³⁷Np(V) were added. Subsequently, Amino- G power (8.55 mg) and radionuclide-bentonite slurry (222 mg) were added to the glass vials (VWR International, 548-0347) [17].

Table 2

Radioactivity in the vials [13]

3.3 In-situ Monitoring

3.3.1 Monitoring Devices

Most monitoring devices are equipped with a megapacker; the schematic monitoring system is illustrated in Fig. 7 [13]. The monitoring system includes 1) a control board and pressure sensor of the surface packer system, 2) a glove box for protecting the sample, 3) a geochemical monitoring system for pH, EC, Eh, turbidity, and fluorometer, 4) a mobile LIBD for measuring colloids in the water, 5) a fraction collector for radionuclide analysis, and 6) a data acquisition system.

Fig. 7

Near-field monitoring system in CFM [13].

3.3.2 Results

After 100 days from initiation, a conservative tracer (amino-G acid, AGA) was detected by fluorescence measurements. The AGA concentration gradually increased until 500 days (Fig. 8) [18]. Owing to the increase in the extraction rate after 720 days, the concentration of AGA decreased, because the extraction rate caused water in the shear zone to dilute. When changing the extraction boreholes at the site, the AGA showed a high concentration owing to the diffusion effect. However, the concentrations of AGA sharply decreased in the subsequent sample extractions due to dilution with fresh water. From the results, the recovery of the AGA was obtained to be approximately 3.5% of the total mass [19]. In addition, although the results of the accelerator mass spectrometry showed the release of radionuclides from the source (Fig. 9), a significantly low concentration of radionuclides was found in the sample [19], suggesting that the glass vial was not completely broken owing to the low swelling pressure and the relatively short period of the experiment, and the radionuclides can remain in the bentonite source. The LIT in-situ experiment and monitoring finished in October 2018. In addition, epoxy resin was injected into the monitoring boreholes to isolate the borehole from the MI shear zone [18].

Fig. 8

Concentration of amino-G acid and total released mass from the source [18].

Fig. 9

Concentrations of ²⁴³Am and ²⁴²Pu after 600 days of injection [19].

3.4 Ongoing Step in LIT

The overcoring process at the LIT study site was conducted for postmortem analysis, considering environmental safety. Although the radionuclide migrates a few centimeters into the bentonite source, overcoring of the near field is required for postmortem analysis of the diffusion path from the bentonite. Because the overcoring sample informs the distribution of colloid particles in the faulted shear zone, postmortem analysis is an important process in field research [20]. Overcoring was designed to improve the probability of sample collection for the entire bentonite source from the MI shear zone because the bentonite colloidal particles potentially migrated into the shear zone. In general, the overcoring procedure consists of resin injection, drilling, grinding, annulus drilling, second resin treatment, and final overcore drilling. The two resin injections were dedicated to stabilizing the bentonite source in the near field to prevent damage during drilling. In the LIT overcoring procedure, a second resin injection was conducted after reaching 6 m depth and the engagement port in CFM 06.002. To obtain a flat borehole face, 6 m of the face was flattened using a custom-made grinder. After the completion of the overcoring process, a postmortem analysis is being carried out.

4. in-rock Bentonite Erosion Test (i-BET)

4.1 Overview

The study of bentonite erosion is inevitable for validating long-term stability in deep geological repositories [21] because bentonite erosion induces the degradation of the stability and integrity of the buffer material in a DGR [21,22]. In the LIT experiment, the research focused on colloid behavior in a single water-conducting fracture in the MI shear zone. Although bentonite and tracer sources were emplaced in the triple-packer system, the tracers migrated a few centimeters due to the low swelling pressure (0.6 and 1.5 MPa) and hydraulic conductivity. Therefore, Nagra introduced a new field experiment, i-BET, to characterize bentonite erosion under higher swelling pressures in a natural flow field than that of LIT. It focuses more on bentonite erosion than on colloid formation and the colloid-associated radionuclide migration for more than two years [23]; then, a bentonite source packer was emplaced in the 220 mm overcoring section of JGP 11.003 at the JGP site in the GTS. Furthermore, three monitoring boreholes were drilled and labeled CFM 18.001–18.003, as shown in Fig. 10 [23].

Fig. 10

Illustration of the boreholes in i-BET [23].

Based on the examination, the JGP site has been finally selected, as it was well characterized because the overcoring of the JGP 11.003 borehole was (220 mm) conducted to 11.5 m. Then, an extra borehole was prepared for the i- BET experiment. The larger diameter of the borehole offers an effective bentonite density accompanying high swelling pressures. In addition, the available structural and hydraulic characterizations indicate a favorable location with a waterconducting feature.

4.2 Experiment Layouts

Several previous studies have reported the characteristics of the groundwater in the GTS [24,25]. The results generally indicate that groundwater shows moderate alkaline pH (> 9), low ionic strength (approximately 1.2 mM), and low electric conductivity (approximately 100 μS·cm⁻¹) [17]. Groundwater is typically saturated with quartz, calcite, and kaolinite and is undersaturated with feldspars [26]. The partial pressure of CO₂ was below atmospheric conditions (approximately 5.5–10 bar). Therefore, GTS groundwater has the potential to form an excellent analog to freshwater events related to glacial activity, possibly entering a future deep geological repository in crystalline rock [17].

Detailed experimental layouts are summarized in the literature [23]. The hydraulic test between JGP 11.003 and JGP 11.002 demonstrated that the hydraulic connection was inappropriate for i-BET. Therefore, an additional investigation of the hydraulic field between JGP 11.003 and JGP 11.001 was conducted. The field test results showed that the steady outflow of JGP 11.001-i2 was approximately 80 mL·min⁻¹ in a steady dilution flow of JGP 11.003-i2 (3 mL·min⁻¹). The breakthrough of the tracer in JGP 11.001-i2 appeared 19– 100 h after starting the test, and the velocity of the groundwater in JGP 11.003-i2 was approximately 10⁻⁵ m·s⁻¹. Based on the test results, the study site would be appropriate for bentonite erosion. Compacted Wyoming bentonite rings were prepared for the i-BET experiments. The interval of the borehole is 1.05 m long, with a diameter of 0.22 m, and the volume of the interval is approximately 30 L. The interval is filled by the swelling of the compacted bentonite, and the expected saturation duration is 100–200 days [20].

4.3 Ongoing Step

The i-BET packer system was installed in December 2018 and has been monitored for over 700 days. During the monitoring, it was observed that the hydraulic pressure was relatively low owing to sampling and extraction from the near field of the boreholes. The piezometers in the compacted bentonite showed very low pressure but steadily increased with swelling of the bentonite. Most relative humidity sensors showed an early increase of over 90%. The monitoring will be completed by the end of 2023.

5. Concluding Remarks

KAERI joined the CFM project in 2008 to investigate in-situ colloid-facilitated radionuclide transport in the GTS. Currently, CFM is being carried out for Phase IV research within GTS Phase VI. In CFM Phase IV, two research programs, LIT and i-BET, are being conducted in a permeable shear zone. In the LIT, radionuclide-cocktail-containing compacted bentonite rings were installed in a triple-packer system for adjacent fractures in the shear zone. The LIT was conducted for 4.5 years since 2014 to investigate the colloid formation and migration in a flow field. However, the LIT showed that low hydraulic conductivity and swelling pressure did not result in detectable colloid migration during the experiment. Then, the injected radionuclides moved only a few centimeters inside the compacted bentonite, even for highly mobile radionuclides such as Tc and Se. Therefore, further research to investigate the partial migration of radionuclides and the infiltration of bentonite colloids into granite is being conducted after the overcoring step. Thus, an additional in situ experiment, i-BET, was designed to clarify colloid formation and migration in the shear zone. This experiment is focused more on bentonite erosion at a faster flow rate than that at the LIT research site.

The CFM results give some ideas for future field research in the URL of South Korea:

Hence, the performance and durability of the devices should be verified to obtain precise field data. Through the CFM project, KAERI has made efforts to conduct field research on the URL in South Korea. Although the construction of URLs in South Korea remains to be completed for at least a decade, practical knowledge through the full-scale study for the disposal of HLW will be obtained indirectly by participating in an international project like CFM. The lessons learned from participating CFM project will be contributed significantly to establishing Korean URL programs.