- Email: [email protected]
Modern method for sealing deep boreholes
Engineering Geology 202 (2016) 132–142
Contents lists available at ScienceDirect
Engineering Geology journal homepage: www.elsevier.com/locate/enggeo
Modern method for sealing deep boreholes Roland Pusch a,⁎, Gunnar Ramqvist b, Sven Knutsson a a b
Department of Civil, Environmental and Natural Resources Engineering, Lulea University of Technology, Sweden ElTekno AB, Oscarshamn, Sweden
a r t i c l e
i n f o
Article history: Received 28 July 2015 Received in revised form 11 January 2016 Accepted 13 January 2016 Available online 16 January 2016 Keywords: Borehole Canister Clay Concrete Disposal Radioactivity Sealing Supercontainer Waste
a b s t r a c t Deep, investigation boreholes for suitable location of radioactive repositories and holes with larger diameter for storing such waste at great depth must be effectively isolated from the biosphere. A method for sealing boreholes by use of dense smectite clay where the rock is low-permeable and has only very fine fractures, and by concrete cast where the holes intersect water-bearing fracture zones, has been used for sealing of up to 500 m deep exploration boreholes in Sweden and Finland. The paper describes the practical conditions for construction of seals, and the processes that lead to the required tightness of the clay and concrete materials. Application of this technique to even deeper boreholes requires use of mud instead of water for stabilization and for moderating the rate of maturation. Placeability requires that the maturation of the initially unsaturated clay is neither too fast nor too slow and means of controlling the rate are described. The proposed concrete material has a very low cement content, and talc as fluidizer, which gives slow but ultimately very significant strengthening and low solubility of the cement reaction products. At a few kilometer depth the temperature can be more than 60 °C, which has to be sustained by the sealing materials, and this makes saponite and mixed-layer smectite/illite clay possible alternatives to pure smectites. The concept is judged to be applicable also to sealing of abandoned boreholes used for fracking. © 2016 Elsevier B.V. All rights reserved.
1. Scope Deep investigation boreholes in rock considered for hosting a radioactive repository, and large-diameter holes for storing such waste at great depth must be effectively isolated from the biosphere by installation of tight seals. Systematic work for finding practically useful sealing methods and materials was started in Sweden by the Swedish Nuclear Fuel and Waste Management Co (SKB) in the eighties and a method has been worked out and utilized for sealing of cored boreholes of up to 500 m length (Pusch, 1994; Pusch and Ramqvist, 2007; Pusch, 2015). It is proposed for holes reaching down to at least 4000 m as described in the paper, implying that clay seals are installed where the holes are located in low-permeable rock, and that concrete is cast in reamed parts where fracture zones are intersected (Fig. 1). The role of the concrete seals is primarily to provide stable and erosion-resistant support to the clay seals placed over them. If clay is used here instead of concrete it would be dispersed, eroded and lost, which could ruin the overall tightness of the seal system.
in the host rock of repositories with highly radioactive waste, and for disposal of such waste: • the holes must be stable, rinsed and filled with electrolyte-poor water or clay mud, grouting of fracture zones for avoiding loss of clay from clay seals, and for avoiding build-up of hydraulic gradients that can damage placed but not yet matured1 seals, • the seals to be placed must not be stuck at installation because of obstacles, • maturation of the clay seals must neither be too rapid nor too slow. Quick maturation causing too early expansion can make it impossible to bring the seals down in deep holes, while too slow expansion can delay strengthening and their ability to carry concrete that is cast on top of them, • the hydraulic conductivity of the clay seals must be lower than of the surrounding rock and the swelling pressure must exceed 100 kPa for providing tight contact between seal and rock (Pusch, 2015).
2. Conditions A number of criteria must be fulfilled for successful installation and function of clay and concrete seals for up to 100,000 years in boreholes ⁎ Corresponding author. E-mail addresses: [email protected] (R. Pusch), [email protected] (G. Ramqvist).
http://dx.doi.org/10.1016/j.enggeo.2016.01.009 0013-7952/© 2016 Elsevier B.V. All rights reserved.
The first condition can be fulfilled by applying techniques that are common in deep drilling for petroleum and gas prospection and exploitation, i.e. use of heavy-duty rigs, bore muds, rinsing techniques, and grouting. 1 We will use “mature” for hydration (wetting) and homogenization for reaching a defined, stable state.
R. Pusch et al. / Engineering Geology 202 (2016) 132–142
133
Fig. 1. Clay/concrete seals in boreholes. Left: Schematic view of seals in boreholes intersecting a fracture zone. Right: Installation of perforated copper tube with dense clay blocks in the 500 m deep borehole KR-24 at Olkiluoto, Finland. Photo: Ramqvist.
Fig. 2. Technique for stabilizing boreholes. Left: Borehole intersecting fracture zone, Center: Reamed hole filled with concrete (UHCP-type developed by the Swedish Nuclear Fuel and Waste Management Co) between packers, Right: Re-boring creates a stabile concrete tube. After Torbjörn-Hugo Persson, SKB.
A technique for stabilization that has been tested in Sweden involves reaming, concrete casting and re-boring to the original borehole dimensions (Fig. 2). The second condition is particularly important for sealing of holes bored from drifts at large depths where the groundwater pressure is high (Fig. 3). They may intersect fracture zones with different water pressures, which can expose clay and concrete seals to high hydraulic gradients that can cause permanent channeling and malfunctioning by piping and erosion. The third, geometrical, condition means that clay seals and containers for casting concrete must not be longer than the straight parts of the commonly curved holes (Fig. 4). In practice, this means that they must not be longer than 10 to 12 m. The fourth condition, respecting the maturation rate, is particularly important for sealing very deep holes since slow installation of clay seals can cause significant dispersion, erosion and loss of clay moving out from the clay seals being placed. This important issue is further discussed below. The fifth condition concerns the geotechnical properties of matured clay seals and implies that their hydraulic conductivity should be lower than
Fig. 3. Cases where high hydraulic gradients can prevail and damage freshly prepared clay and concrete seals. A and B are fracture zones with high water pressure.
134
R. Pusch et al. / Engineering Geology 202 (2016) 132–142
with different densities are given in Table 2 that we will come back to later in the paper. 3. Description of the method for deep hole sealing 3.1. Clay seals
Fig. 4. Critical geometrical condition for placing clay seals called “plugs”.
that of the rock excepting fracture zones. It is estimated at E−11 m/s at 1–2 km depth, E−12 m/s at 2–3 km depth, and E−13 m/s deeper than 3 km. The conductivity and swelling pressure of smectite-rich clay seals
3.1.1. Function and installation The sealing potential of expansive smectite-rich clay is fundamental to the described borehole sealing principle, which involves installation of perforated tubes that are tightly filled with blocks of very densely compacted smectite granules in the holes to be sealed (Fig. 1). Holes deeper than some hundred meters should be filled with soft smectite mud in the boring and waste placement phases for stabilizing the rock and for controlling the rate of maturation of the clay seals in the installation phase. The dense clay blocks in the perforated tubes made of copper, Navy Bronze or titanium in rock hosting repositories for highly radioactive waste like spent reactor fuel, and steel for most other purposes, absorb water from the rock and expand out through the perforation and embed the tubes (Fig. 5). This process is slow if the blocks are pre-saturated with water and quick if the blocks have only 50–60% degree of saturation. The mud left in the holes between the perforated tubes and the rock is consolidated under the pressure from the expanding clay and ultimately becomes almost as dense and tight as the clay
Fig. 5. Early stage of maturation. Upper: Schematic model of a “skin zone” of clay being formed outside the perforated tube. Lower. 8 h laboratory experiment with clay plugs with and without silicon coating for delay of hydration after 8 h in distilled water. The small picture shows the jointing of clay seal units.
R. Pusch et al. / Engineering Geology 202 (2016) 132–142
135
Table 1 Full-scale field experiments with montmorillonite-rich clay in tubes with 50% perforation ratio and 2 mm wall thickness. Dry density = ρd. Pusch (1994). Case/start of test
Time of maturation, months
Borehole/tube diameter, mm
Borehole/tube length, m
Initial ρd of clay, kg/m3
ρd of clay in tube at end of test, kg/m3
ρd of clay ”skin” zone at end of test, kg/m3
1/1986 2/1986 3/1986 4/2006
30 12 N240 30
56/54 80/76 80/76 80/76
100/100 14/2 160/30 5/2.5
1760 1301 1580 1710
1600a 1250a In position 1350
1500a 1200a In position 1270
a
Axial inner tubes with cables to gauges reduced the cross section and gave low density.
remaining in the tubes. The technique has been tested under laboratory conditions and in Swedish and Finnish underground rock laboratories for sealing of initially water-filled holes (Pusch, 2008). Typical field test data are summarized in Table 1, which shows that the dry density of the clay skin formed around the perforated tubes was up to 90% of the dense clay remaining in the tubes at the end of the tests. Figs. 6 and 7 illustrate the appearance of a 76 mm clay plug shortly after placement in a water-filled 80 mm metal tube for testing. 3.1.2. Rate of maturation The factors affecting the rate of maturation of the clay seals and thereby of the time for reaching different degrees of sealing of the holes are: • ability of the rock to supply the clay plug with water for keeping up the maturation rate, • water pressure, • hydraulic gradients that can cause piping and erosion, • the mineral composition, dry density, and initial degree of water saturation of the dense clay in the perforated tubes, and of the mud, • geometrical conditions like the perforation ratio of the tubes and the size of the gap between tubes and rock, and between tubes and clay.
Looking first at the processes associated with installing dense, unsaturated clay seals, one has to consider the role of the water pressure, which is 10 to 40 MPa in the depth interval 1–4 km. It causes some elastic compression of the clay, followed by expansion back when it becomes saturated with water. The transport mechanism of water from the rock is by flow through winding microstructural channels with a diameter of a few tens of micrometers from which water molecules migrate into the clay matrix by diffusion (Pusch, 2008, 2015; Pusch and Kasbohm, 2001). The driving force is the groundwater pressure and porewater suction (−150 MPa) in the dense clay2 while the hydraulic conductivity of the successively saturated clay controls the rate of saturation. Using common lab data it has been found that the entire dense clay cylinder in the perforated tube will be saturated in 2 weeks to 3 months depending on the hydraulic conductivity of the clay (E − 13 to E − 10 m/s). The part of the dense clay that is next to the perforated tube starts to swell through the perforation as indicated by Fig. 5, forming little clay columns that expand laterally and unite to become an increasingly homogeneous skin in the gap between tube and rock. It is successively consolidated under the pressure of the expanding dense clay and its shear strength, which determines if the clay seals can be forced down to the intended depth, grows with time. The process of dense clay moving out through the perforation of the tube is almost independent of the groundwater pressure. Saturation of the dense clay can be considered as a diffusion process as shown in Fig. 7. The data agree well with actually measured densities. 3.1.3. Role of geometry on the ultimate density of clay seals Table 1 shows that the dry density of the clay in the recovered tubes was about 20–30% lower than of the clay that was initially in the tubes 2
Clay Technology AB, Lund, Sweden.
placed in the water-filled holes because of migration of clay from them to the gap between tubes and confining rock where a “skin zone” was formed. Fig. 8 illustrates the impact of the dimension of the perforated tube on the average density and thereby on the sealing function of the ultimately matured clay seals. In mud-filled holes the effect is naturally smaller. The gap is only a few millimeters wide in practice but the properties of the clay filling will control the physical performance of the clay seal as a whole. The diameter of the boreholes to be sealed is naturally important: for any initial dry density of the clay in slim holes it will become lower than in wider holes where the metal tubes occupy a smaller fraction of the cross section (Table 2). In deep slim holes the amount of clay in seals will be further reduced by dispersion and erosion when they are brought down in water-filled holes as demonstrated by Fig. 9. Such degradation will be much smaller in mud since no particle migration into or
Fig. 6. Appearance of clay seal unit early after start of wetting.
Fig. 7. Evolution of the degree of saturation of 76 mm clay plug in 80 mm borehole at unlimited access to electrolyte-poor water assuming the process to take place by diffusion of water with the diffusion coefficient D = 3E−10 m2/s. Lower curve: 12 days, central curve: 72 days, upper curve: 120 days.
136
R. Pusch et al. / Engineering Geology 202 (2016) 132–142
Fig. 8. Impact of the initial dry density on the average ultimate density of the matured clay in a sealed hole with defined geometry.
Table 2 Theoretical ultimate density and major geotechnical properties (K = hydraulic conductivity, ps = swelling pressure) of smectite-rich seals in boreholes with different diameters assuming that the initial dry density is 2000 kg/m3 and that the gap between tube and rock and the thickness of the tube wall are 2 mm. Pusch et al. (2015). Borehole diameter, mm
Initial dry density, kg/m3
Final dry density, kg/m3
Final density at saturation, kg/m3
Final density at saturation if the borehole is 3 mm wider hole
K, m/s for 3.5% CaCl2 water
Swelling pressure, MPa
56 76 80 100
2000 2000 2000 2000
1683 1760 1785 1830
2060 2110 2124 2150
1942 1995 2037 2083
5E−12 1E−12 7E−13 3E−13
1.0 1.5 9.0 12.0
out from the perforated tubes will take place in the short time of placement of the seals. The actual variation in borehole diameter that has to be considered in selecting tube diameters is exemplified in Fig. 10. 3.1.4. Practical role of the maturation rate The risk of too quick or too slow maturation of the clay plugs is of great practical importance. Taking as an example the case of installing a 12 m long set of jointed clay plugs with 76 mm diameter in an 80 mm borehole, weighing about 900 kg in air and about 440 kg under water, one finds that their own weight is not enough to let it sink if the “skin zone” has a higher shear strength than about 1.5 kPa. In practice, it will reach this value in less than 1 h and it will rise to 50 kPa in 6 h as indicated by Table 3, which shows the required amount of water for maturation of a clay seal unit of 2.5 m length in a 76 mm borehole and the geotechnical properties as functions of time. In very tight rock that does not provide the seal with enough water the maturation will be slower. For deep holes the high water pressure is believed to bring enough water to the clay despite the lower conductivity at depth. After 6 h of maturation in a hole with unlimited access to water the seal has to be loaded by 15 metric tons for forcing it down, which can be made by using a big and robust drill rig. After 8 h, which is a reasonable time for bringing such a set down by 1–2 km, the required load is estimated to be 30 tons, which may be critically high for getting it down unless a vibrator is used in combination with loading. While installation of clay seals may be hindered by their growth in shear strength, quick maturation is desired for making it possible for placed seals to carry concrete cast upon them where fracture zones are intersected. The bearing capacity can be estimated by use of the classical Prandtl failure theory for predicting the bearing capacity of loaded soil (Peck et al., 1973). It implies that failure takes place if the pressure exceeds about 6 times the undrained shear strength. For a clay plug that has matured for 6 h the pressure causing failure of the “skin” zone is hence 300 kPa, which corresponds to the pressure exerted by a 30 m
high concrete fill weighing 2200 kg/m3 in air and 1200 kg/m3 under water. Failure will cause concrete to penetrate the “skin” clay and displace it upwards along the hole. This phenomenon has been recognized in a field experiment with a clay seal of 76 mm diameter in an 80 mm borehole when concrete was filled over it shortly after installing the clay seal (Pusch and Ramqvist, 2007), One finds, in conclusion, that the rate of maturation and the time for installation of long plug sets must be seriously considered for successful installation of the clay seals. 3.1.5. Ways of retarding maturation of clay seals A proposed method is to coat the perforated tubes with a thin paste of mixed smectite clay and talc (Pusch, 2011). The dense clay in the perforated tubes sucks water from the coating and seals off the perforation holes, which retards migration of water into the clay in the tube (Fig. 11). Once the slow migration of water through the coating has initiated expansion of the dense clay in the perforated tube, causing the coating to break, its uptake of water accelerates as for uncoated tubes. The talc fragments are “submicroscopic” and disseminated and have insignificant impact on the geotechnical properties of the “skin zone”. Current research indicates that prewetting of the dense clay blocks to give them an initially high degree of water saturation is possible and effective since the suction potential and wetting rate are then lower (Pusch, 2008; Yang et al., 2015). 3.1.6. Long-term stability of clay seals The long-term chemical stability of montmorillonite-rich smectite clay, which is taken here as primary clay candidate because of its excellent sealing properties, has been investigated in detail since the beginning of the previous century. American and European mineralogists like Roaldset3 and Lindgreen (1991), who investigated a large number 3
Personal communication.
R. Pusch et al. / Engineering Geology 202 (2016) 132–142
137
Table 3 Calculated amount of water for maturation of a 2.5 m long clay plug of the type shown in Fig. with the initial dry density 1700 kg/m3 and geotechnical properties as a function of time.
Fig. 9. Impact of erosion on the average density of the matured clay seal in Fig. 8 with 1825 kg/m3 initial dry density.
of samples of Tertiary age from the North Sea basin buried at large depth, reached the conclusion that smectite in layers from 1500 to 1700 m below the present sea bottom have only undergone insignificant transformation to less expanding illite/smectite (I/S) forms since sedimentation. At 2800 m depth, where the temperature has been more than 100 °C for tens of millions of years there has been more extensive conversion to mixed-layer minerals I/S and illite (Fig. 12). Some cementation by precipitated silica has taken place as well. The processes followed the generally accepted conversion model of smectite-to-illite in Eq.(1).
Smectite þ Potassium N Illite þ Silica
ð1Þ
We conclude here that largely preserved mineral composition over 100,000 years of a smectite-rich clay seal at depths of up to about 4 km, where the prevailing temperature can be 100 °C can be counted on, but that silica cementation will reduce its expandability. The criterion that the clay shall stay ductile and exert a contact pressure on the borehole walls will therefore not be fulfilled except if the boreholes undergo enough radial convergence by creep. This needs to be proven.
Fig. 10. Example of diameter variation in SKB's borehole KLX21B with 76 mm nominal diameter. “Bhln” denotes borehole length.
Time, hours
Density of “skin zone”, kg/m3
Amount of sorbed water, liters
Hydraulic conductivity of “skin zone”, m/s
Shear strength of “skin zone”, kPa
6 12 24 48 96
1100 1150 1325 1400 1700
10 15 33 40 70
E−9 5E−10 5E−11 2E−11 E−12
50 200 400 550 700
3.1.7. Clay materials considered for use in deep holes Application of the criteria that the clay seals must not be more permeable than the boring-disturbed surrounding rock and be enough ductile to exert an active pressure on the rock of at least 100 kPa, sorts out practically useful clay seal materials. Data from comprehensive laboratory investigations collected in Table 4 has been the basis of the preliminary assessment that for a rock conductivity of E − 12 m/s all the smectite-rich clay candidates, with montmorillonite as number one, will do provided that the initial dry density is 1850–1900 kg/m3. Applying a reasonable margin the minimum dry density would be 1950 kg/m3. It should be mentioned, however, that hydrothermal testing has indicated that saponite is more chemically stable than montmorillonite (Gueven and Huang, 1990), which is still preferred because of its lower cost and acceptable sealing ability.
3.2. Concrete seals 3.2.1. Role and construction The role of concrete seals is essentially to fill the parts of the boreholes that intersect fracture zones with erosion-resistant material for supporting clay seals and serve as filter for preventing particles from adjacent clay seals to migrate into the fracture zone and be transported away by flowing groundwater. Casting of concrete in boreholes can be made by use of containers that are emptied just on top of the latest installed clay seal. The discharge of concrete displaces the much lighter clay mud and fills the hole (Fig. 13). The concrete is prepared with fresh-water in batches immediately before it is pumped down.
Fig. 11. Borehole seal of very dense montmorillonite-rich clay coated with a paste of mixed talc/clay with the dry mass ratio 3.3/1 and a water content of 110% submerged in distilled water for 3 h. No migration of clay from the tube had occurred in the pasted part while a soft clay gel had formed between the tube and the plexiglass container in the untreated part.
138
R. Pusch et al. / Engineering Geology 202 (2016) 132–142
3.2.2. Cement and concrete recipe An early recipe proposed in SKB's Borehole Sealing Program (Pusch and Ramqvist, 2007) had the following components in kg per cubic meter of concrete: Portland cement (60 kg), silica fume (60 kg), fine α-quartz (200 kg), fine cristobalite (200 kg), Glenium 51 superplasticizer (4.4 kg in dry form), granitic aggregate 0–4 mm (1700 kg), and water (244.3 kg). The compressive strength was 4.5 MPa after 2 days and 31.7 MPa after 28 days at room temperature. Such concrete, having pH N 12, was cast on top of a clay seal in a 76 mm borehole, and analyzed after about 3 years with respect to the mineralogy and chemical composition of both clay and concrete. This showed significant degradation of both, by dissolution and cation exchange from Na to Ca of the clay, and by dehydration and fissuring of the concrete (Warr and Grathoff, 2010). These observations and the probability that organic fluidizers will degrade or be eaten by microbes led to the following composition for the laboratory study: low-pH cement,4 quartz aggregate, and talc as fluidizer. The respective contents were: cement 6.5%, talc 9.5%, and aggregate 84% (100% b 4 mm, 93% b2 mm, 70% b1 mm, and 30% b 0.1 mm), (cf. Pourbakhtiar, 2012; Pusch et al., 2013). A suitable water/cement ratio can be 3–4 for this composition having an aggregate/cement ratio of 10–15. This gives 2070 kg/m3 density and pH = 10. Typical compressive strength data after 2–28 days of curing are summarized in Table 5, together with data for a corresponding concrete with Portland cement. The microstructural constitution of this type of concrete reflects the plastic behavior of the talc phase in the earliest curing phase: the formation of skin coatings of quartz grains by flaky talc particles resulting in easy glide of both (Fig. 14). The subsequent strengthening provided by the talc is caused by precipitation of cementing complexes formed by reactions between low-pH cement components and slowly dissolving talc (Fig. 15). The hydraulic conductivity was measured after 7 and 21 days giving K-values of 5E − 9 and 8E − 10 m/s, respectively (Mohammed et al., 2013). Major conclusions from compressive tests using talc as fluidizer are: • concrete with low-pH cement harden much slower than Portland cement concrete, and has about 50% of the strength of such concrete after 7 days, but is 2 to 3 times stronger after 28 days and • both concrete types are brittle but the concrete with low-pH cement maintains a certain ductility and tendency to undergo creep strain in the curing time, implying tighter physical interaction with the rock than concrete with Portland cement.
Fig. 12. Smectite-to-illite conversion via mixed-layer I/S formation and/or direct precipitation of illite (Pusch, 1994). The dark contours represent precipitations of silica and/or illite.
It can be prepared by mixing the cement and aggregate with electrolytefree tap water to the water/cement ratio prescribed by the selected recipe or by mixing in “Dry water” in the mass of solid components (Forsberg et al., 2014). Such water consists of microscopic droplets confined in extremely thin silica “shells” that break and release water when compacted. It will be of no importance if some small amount of mud remains since it would react with the cement component of the concrete and stiffen.
3.2.3. Long-term performance of concrete The described grain-size distribution of the aggregate and the very low cement content gives a tight microstructure that is similar to bottom moraine and minimizes changes in physical performance of the concrete in the unlikely case of complete loss of the cement (Mohammed et al., 2013). Thus, the concrete has a very low compressibility and a pore size distribution that minimizes the risk of infiltration and through-transport of clay particles from contacting clay seals. The longevity of the magnesium/silicate cement reaction products from talc/cement interaction is expected to be very significant, hence making the risk of total loss of cement improbable. Table 6 summarizes the element concentrations in such concrete with and without talc after 7 days maturation in initially electrolyte-free water. The major difference between cement with and without talc is a loss in F, Al, S, and K and an increase in Na and Mg in the presence of talc. Compared with virgin cement the significant increase in Ca, Mg and Si and the decrease in Al were the most obvious findings. They indicate that the talc component was partly dissolved and the elements incorporated into new cementing compounds that formed in the first week of reaction. Further reactions caused the obvious strengthening recorded 4
MERIT 5000 which has 34% SiO2, 13% Al2O3, 17% MgO and 31% CaO (Pusch et al., 2003).
R. Pusch et al. / Engineering Geology 202 (2016) 132–142
139
Table 4 Clays for sealing 80 mm boreholes with 76 mm tubes down to 2 km depth in salt, Ca-rich groundwater. Geometrical changes associated with maturation are according to Fig. 3 but no chemically induced degradation is assumed. Clay
Initial dry density, kg/m3
Density of saturated homogeneous seal, kg/m3
Hydraulic conductivity of clay seal, m/s
Swelling pressure of clay seal, kPa
Montmorillonite Montmorillonite Saponite Saponite Mixed-layer montm./illite, 25–35% expandable Mixed-layer montm./illite, 25–35% expandable Mixed-layer montm./illite, 60–70% expandable Mixed-layer montm./illite, 60–70% expandable
1850 1500 1850 1500 1850 1500 1850 1500
2000 1800 2000 1800 2000 1800 2000 1800
8E−13 6E−11 7E−13 3E−12 2E−11 7E−11 5E−12 3E−11
4700 1000 5000 2000 700 200 2000 600
after 28 days. The results also imply that the dissolution of talc provided additional Mg for the cementing minerals. It is estimated that complete curing, requiring dissolution, diffusive migration of mobile elements and successive precipitation of reactive elements in the chemically transient porewater, can take years and decades and even longer time in big seals in large-diameter holes. 3.3. Performance of contacting concrete and clay seals — exploratory studies Field experiments running for almost 3 years have shown how concrete with Portland cement interacts with smectite-rich (montmorillonite) clay and what the impact is on the physical properties of the respective material (Pusch et al., 2003; Warr and Grathoff, 2010). The major finding was that the two materials caused significant changes within several centimeters from their contact, the most obvious reaction being partial loss of expandability of the clay and loss of strength of the concrete. Subsequent 8 week hydrothermal tests in the laboratory with the Danish mixed-layer I/S “Holmehus clay” (Pusch et al., 2015) of
Tertiary age, composed of smectite (60% montmorillonite), glauconite, quartz, feldspars, calcite, glass, kaolinite, Ti oxide and pyrite, in contact with concrete of the described type with low-pH cement, have shown the impact of temperature. The test arrangement, illustrated in Fig. 16, consisted of two completely but differently fluid-saturated clay samples with a dry density of 1500 kg/m3 in the hydrothermal cell, separated by concrete cast over the lower clay sample. The confining filters were connected to vessels with distilled water and 3.5% CaCl2, respectively. One test was conducted at room temperature, one at 100 °C and one at 150 °C. Some tests were made at higher temperature for accelerating the chemical processes and for simulating the conditions at depth in rock and they indicated, in contrast with the unheated materials, obvious mineralogical changes as indicated by the XRD diagram in Fig. 17 and scanning electron images in Fig. 18. These changes were also mirrored by the uniaxial compression tests reported in Fig. 19 and evaluated in the form of the stress/strain diagram in Fig. 20 of the room temperature sample. The 100 °C samples showed similar behavior while the 150 °C sample had about 25% lower peak strength, much higher stiffness, and obvious brittleness (Pusch et al., 2010). The clay samples prepared with calcium chloride solution behaved as the ones with distilled water but showed somewhat higher peak strength values and stiffness (Mohammed et al., 2014). An indication of the relative amounts of important elements, expressed in oxide form, in the contacting clay and concrete is given by Table 7. Compared with untreated I/S Holmehus clay the ones in contact with concrete had become significantly richer in SiO2 and poorer in Al2O3, hence being less expandable.
Fig. 13. Casting concrete by extrusion from a tube attached to the drill string with its open end held at the top of the latest placed clay seal.
Table 5 Average compressive strength in MPa of concrete prepared with the same low cement content. Pourbakhtiar (2012). Duration (days)
Portland concrete
Merit 5000 concrete
2 7 28
0.5–0.6 0.6–0.7 0.8–0.9
0.01–0.02 0.1–0.4 1.8–2.7
Fig. 14. Accumulation of talc particles oriented around a rounded quartz grain (lost) in the concrete (Institute of Geography and Geology, University of Greifswald, Germany).
140
R. Pusch et al. / Engineering Geology 202 (2016) 132–142
Fig. 16. Set of clay/concrete disks for hydrothermal treatment under isothermal conditions. Fig. 15. Talc particles in MERIT concrete. The smooth, slightly rounded edges of the talc indicate dissolution (Institute of Geography and Geology, University of Greifswald, Germany).
As to the clay, the pressure at failure was about half the swelling pressure, 1.5 MPa, of “virgin” Holmehus clay saturated with 3.5% CaCl2 solution. This indicates weakening of the clay by dissolution in contact with concrete. For the concrete the compressive strength 4.2 MPa was slightly higher than the average value for a 28 day samples (Table 5). This strengthening can have been due solely to continued talc/cement reactions. Random powder diffractograms of the clay samples that were in contact with concrete and of “virgin” salt clay had main dominant reflection in the clay of illite (10 Å) and of gypsum, which appeared to be abundant. Minor amounts of kaolinite and chlorite were present as well but less abundant than in the untreated “virgin” clay, which showed typical I/S peaks and a dominant quartz reflection. The mineralogical changes do not explain the recorded weakening, the most probable reason being partial conversion of crystalline matter to amorphous. 4. Conclusions The overall conclusion of the present exploratory study is that the proposed borehole sealing techniques and materials are feasible: very dense smectite clay in perforated metal tubes, and talc-based concrete cast where fracture zones are intersected. Application of the method can be of great importance for isolating deep repositories for highly radioactive waste and for preventing chemically contaminated water from “fracking” to reach up to shallow acquifers. The major findings of the study can be summarized as follows:
if the “wall friction” is too high. Too slow maturation can delay utilization of the isolation potential of the clay seals. The maturation rate of clay seals can, however, be effectively controlled, • concrete is required for providing stable support to the clay seals. It does not have to be low-permeable – the actual hydraulic conductivity being E−11 to E−10 m/s – but must have a microstructural constitution that prevents clay particles from adjacent clay seals to migrate into it and be washed out by flowing groundwater (Mohammed et al., 2013). Traditional binders like Portland cement are expected to last only for some decades and a new concrete, with talc instead of degradable organic fluidizer and with higher chemical stability, is proposed. The curing rate of the concrete seals determines the time until the next clay seal can be installed and the study indicates that placement of a clay plug over a concrete seal can be made after about one week. The density and granular composition of the concrete is such that even total loss of the small cement content does not jeopardize its capacity to provide support to the clay plugs and prevent migration and loss of clay particles into the intersected fracture zone, • the sealing effect of montmorillonite-rich clay seals in contact with the proposed concrete is sufficient in a 100,000 year perspective according to the exploratory study for up to 100 °C. For very deep holes heated to 150 °C by highly radioactive waste other smectite clays, like the Mg-rich saponite, may be preferable,
• borehole sealing by means of perforated copper tubes with montmorillonite-rich smectite clay with high density for long-term function can be effective as demonstrated by pilot field experiments with up to 500 m deep, steep boreholes and 100 m long subhorizontal boreholes, • in very deep holes the maturation of clay seals can be too quick or too slow, the first mentioned causing difficulties in getting the seals down Table 6 Element concentration (wt.%) of MERIT 5000 cement with and without talc, measured by energy dispersive X-rays (EDX). Warr and Grathoff, G (2011). Material
F
Na
Mg
Al
Si
S
K
Ca
Ti
Fe
Raw cement Cement without talc Cement with talc
– 1.12
– 0.07
2.63 1.14
1.57 1.11
4.46 6.28
0.11 0.32
– 0.13
1.37 8.84
– 3.00
– –
0.08
0.60
3.06
0.60
7.66
0.08
0.03
8.15
2.47
0.23
Fig. 17. Generalized X-ray diffraction pattern (red) of the Holmehus clay sample in the clay/concrete set compared to that of untreated (“virgin”) clay. Co-Kalpha radiation, d-values are in Å units. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
R. Pusch et al. / Engineering Geology 202 (2016) 132–142
141
Fig. 18. Scanning electron micrographs of clay at the contact with concrete (A), and of the concrete at this contact (B); (Institute of Geography and Geology, University of Greifswald, Germany).
Fig. 19. Compression stages at room temperature. Initial failure took place in the upper clay sample at 519 kPa pressure, corresponding to a shear strength of about 260 kPa of the clay saturated with distilled water. The concrete remained intact until the pressure was 4.2 MPa and the total compressive strain had become about 50%. Mohammed et al. (2014).
• the described sealing method has a potential to be used also in very deep holes – reaching down to 4 km – for storing highly radioactive waste (“VDH” concepts). The upper 2 km of such holes will not be
heated to more than 60–70 °C and can well be sealed by use of the described method while the evolution of the seals in the lower 2 km part heated up to 150 °C has to be further examined.
Acknowledgments The authors express their gratitude to Dr. Mohammed Hatem Mohammed, formerly at the Luleå University of Technology, Luleå, Sweden, for the assistance in preparing the report and for providing results and diagrams from his scientific work. Table 7 EDX compositions (weight of oxides, n = 30) of illite/smectite and cement phases across the clay/concrete contact at room temperature. Na2O
MgO
Al2O3
SiO2
K2O
CaO
TiO2
FeO
Sum
Smectite clay particles 21 °C 1.3
2.9
19.2
63.1
3.8
0.6
1.6
7.5
100
28.7
2.4
65.7
0.1
1.1
0
1.4
100
Temperature
Fig. 20. Stress/strain diagram for the compressed sample set at room temperature. A–D = compression to failure of clay with distilled water. E = compression to failure of concrete.
Cement phases 21 °C 0.4
142
R. Pusch et al. / Engineering Geology 202 (2016) 132–142
References Forsberg, T., Pusch, R., Knutsson, S., 2014. Rational method for preparing dense clay blocks with desired degree of water saturation. Engineering Geology, Technical Note. Vol.X, No.Y (pp.00). Gueven, N., Huang, W.-L., 1990. Effects of Mg2+ and Fe3+ Substitutions on the Crystallization of Discrete Illite and Illite/smectite Mixed Layers. Dept. Geosciences Texas Tech University, Exxon Production research Co, Houston, Texas, USA. Lindgreen, H., 1991. Elemental and Structural Changes in Illite/smectite Mixed-layer Clay Minerals During Diagenesis in Kimmeridgian–Volgian (–Ryazanian) clays in the Central Trough, North Sea and the Norwegian–Danish Basin. Bull. geol. Soc. Denmark Vol. 39, pp. 1–82 Copenhagen, October 25th 1991. Mohammed M. H., Pusch R., Al-Ansari N., Knutsson S., Emborg M., Nilsson M., and Pourbakhtiar A., 2013. Talc-based concrete for sealing borehole optimized by using particle packing theory. J. Civil Eng. Archit., Apr. 2013, Volume 7, No. 4, pp. 440–455. Mohammed, M.H., Pusch, R., Warr, L.N., Kasbohm, J., Knutsson, S., 2014. Interaction of Clay and Concrete in Deeply Disposed High-level Radioactive Waste (Manuscript Submitted to Construction and Building Materials). Peck, R.A., Hanson, W.E., Thornburn, T.H., 1973. Foundation Engineering. second ed. John Wiley & Sons, Inc. Pourbakhtiar, A., 2012. Pilot Study of Method for Constructing Concrete Seals and Fracture Grouts in Deep Boreholes and Cementitious Backfills in Tunnels, Drifts and Shafts in Crystalline Rock (M.Sc Thesis) Department of Civil, Mining and Natural Resources Engineering, Lulea University of Technology, Sweden. Pusch, R., 1994. Waste Disposal in Rock, Developments in Geotechnical Engineering, 76. Elsevier Publ. Co. (ISBN: 0–444-89449-7). Pusch, R., 2008. Geological Storage of Radioactive Waste. Springer-Verlag, Berlin, Heidelberg (ISBN: 978–3-540-77332-0).
Pusch, R., 2011. A technique to delay hydration and maturation of borehole seals of expansive clay”. Eng. Geol., Vol. 121, 2011, (pp. 1–6). Pusch, R., 2015. Bentonite Clay. CRC Press (Taylor & Francis Group) (ISBN 13:978–1-48224343-7). Pusch, R., Kasbohm, J., 2001. Can the Water Content of Highly Compacted Bentonite Be Increased by Applying a High Water Pressure? Technical Report TR-01-33, Swedish Nuclear Fuel and Waste Management Co (SKB) Stockholm. Pusch, R., Ramqvist, G., 2007. Borehole Projects — Final Report of Phase 3. SKB R-07-58, Swedish Nuclear Fuel and Waste Management Co. Svensk Kärnbränslehantering AB, SKB. Pusch, R., Zwahr, H., Gerber, R., Schomburg, J., 2003. Interaction of cement and smectite clay — theory and practice. Appl. Clay Sci. 23, 203–210. Pusch, R., Yong, R.N., Nakano, M., 2010. Stiffening of smectite buffer clay by hydrothermal effects. Eng. Geol. 110, 21–31. Pusch, R., Warr, L., Grathoff, G., Pourbakhtiar, A., Knutsson, S., Mohammed, M.H., 2013. A talc-based cement-poor concrete for sealing boreholes in rock, 2013. Eng. Geol. 5, 251–267. Pusch, R., Kasbohm, J., Hoang-Minh, T., Knutsson, S., Nguen-Thanh, L., 2015. Holmehus clay — a Tertiary smectitic clay of potential use for isolation of hazardous waste. Eng. Geol. 188, 38–47. Warr, L.N., Grathoff, G.H., 2010. Sealing of investigation boreholes: mineralogical and geochemical borehole plug analyses. SKB Report. Ernst-Moritz-Arndt-Universität, Greifswald, Germany. Yang, T., Pusch, R., Knutsson, S., Liu, X., 2015. Lab testing of method for clay isolation of spent reactor fuel in very deep boreholes. Proc. World Multidisciplinary Earth Sciences Symposium — WMESS 2015″, September, pp. 7–11 (Prague, Czech Republic).
Contents lists available at ScienceDirect
Engineering Geology journal homepage: www.elsevier.com/locate/enggeo
Modern method for sealing deep boreholes Roland Pusch a,⁎, Gunnar Ramqvist b, Sven Knutsson a a b
Department of Civil, Environmental and Natural Resources Engineering, Lulea University of Technology, Sweden ElTekno AB, Oscarshamn, Sweden
a r t i c l e
i n f o
Article history: Received 28 July 2015 Received in revised form 11 January 2016 Accepted 13 January 2016 Available online 16 January 2016 Keywords: Borehole Canister Clay Concrete Disposal Radioactivity Sealing Supercontainer Waste
a b s t r a c t Deep, investigation boreholes for suitable location of radioactive repositories and holes with larger diameter for storing such waste at great depth must be effectively isolated from the biosphere. A method for sealing boreholes by use of dense smectite clay where the rock is low-permeable and has only very fine fractures, and by concrete cast where the holes intersect water-bearing fracture zones, has been used for sealing of up to 500 m deep exploration boreholes in Sweden and Finland. The paper describes the practical conditions for construction of seals, and the processes that lead to the required tightness of the clay and concrete materials. Application of this technique to even deeper boreholes requires use of mud instead of water for stabilization and for moderating the rate of maturation. Placeability requires that the maturation of the initially unsaturated clay is neither too fast nor too slow and means of controlling the rate are described. The proposed concrete material has a very low cement content, and talc as fluidizer, which gives slow but ultimately very significant strengthening and low solubility of the cement reaction products. At a few kilometer depth the temperature can be more than 60 °C, which has to be sustained by the sealing materials, and this makes saponite and mixed-layer smectite/illite clay possible alternatives to pure smectites. The concept is judged to be applicable also to sealing of abandoned boreholes used for fracking. © 2016 Elsevier B.V. All rights reserved.
1. Scope Deep investigation boreholes in rock considered for hosting a radioactive repository, and large-diameter holes for storing such waste at great depth must be effectively isolated from the biosphere by installation of tight seals. Systematic work for finding practically useful sealing methods and materials was started in Sweden by the Swedish Nuclear Fuel and Waste Management Co (SKB) in the eighties and a method has been worked out and utilized for sealing of cored boreholes of up to 500 m length (Pusch, 1994; Pusch and Ramqvist, 2007; Pusch, 2015). It is proposed for holes reaching down to at least 4000 m as described in the paper, implying that clay seals are installed where the holes are located in low-permeable rock, and that concrete is cast in reamed parts where fracture zones are intersected (Fig. 1). The role of the concrete seals is primarily to provide stable and erosion-resistant support to the clay seals placed over them. If clay is used here instead of concrete it would be dispersed, eroded and lost, which could ruin the overall tightness of the seal system.
in the host rock of repositories with highly radioactive waste, and for disposal of such waste: • the holes must be stable, rinsed and filled with electrolyte-poor water or clay mud, grouting of fracture zones for avoiding loss of clay from clay seals, and for avoiding build-up of hydraulic gradients that can damage placed but not yet matured1 seals, • the seals to be placed must not be stuck at installation because of obstacles, • maturation of the clay seals must neither be too rapid nor too slow. Quick maturation causing too early expansion can make it impossible to bring the seals down in deep holes, while too slow expansion can delay strengthening and their ability to carry concrete that is cast on top of them, • the hydraulic conductivity of the clay seals must be lower than of the surrounding rock and the swelling pressure must exceed 100 kPa for providing tight contact between seal and rock (Pusch, 2015).
2. Conditions A number of criteria must be fulfilled for successful installation and function of clay and concrete seals for up to 100,000 years in boreholes ⁎ Corresponding author. E-mail addresses: [email protected] (R. Pusch), [email protected] (G. Ramqvist).
http://dx.doi.org/10.1016/j.enggeo.2016.01.009 0013-7952/© 2016 Elsevier B.V. All rights reserved.
The first condition can be fulfilled by applying techniques that are common in deep drilling for petroleum and gas prospection and exploitation, i.e. use of heavy-duty rigs, bore muds, rinsing techniques, and grouting. 1 We will use “mature” for hydration (wetting) and homogenization for reaching a defined, stable state.
R. Pusch et al. / Engineering Geology 202 (2016) 132–142
133
Fig. 1. Clay/concrete seals in boreholes. Left: Schematic view of seals in boreholes intersecting a fracture zone. Right: Installation of perforated copper tube with dense clay blocks in the 500 m deep borehole KR-24 at Olkiluoto, Finland. Photo: Ramqvist.
Fig. 2. Technique for stabilizing boreholes. Left: Borehole intersecting fracture zone, Center: Reamed hole filled with concrete (UHCP-type developed by the Swedish Nuclear Fuel and Waste Management Co) between packers, Right: Re-boring creates a stabile concrete tube. After Torbjörn-Hugo Persson, SKB.
A technique for stabilization that has been tested in Sweden involves reaming, concrete casting and re-boring to the original borehole dimensions (Fig. 2). The second condition is particularly important for sealing of holes bored from drifts at large depths where the groundwater pressure is high (Fig. 3). They may intersect fracture zones with different water pressures, which can expose clay and concrete seals to high hydraulic gradients that can cause permanent channeling and malfunctioning by piping and erosion. The third, geometrical, condition means that clay seals and containers for casting concrete must not be longer than the straight parts of the commonly curved holes (Fig. 4). In practice, this means that they must not be longer than 10 to 12 m. The fourth condition, respecting the maturation rate, is particularly important for sealing very deep holes since slow installation of clay seals can cause significant dispersion, erosion and loss of clay moving out from the clay seals being placed. This important issue is further discussed below. The fifth condition concerns the geotechnical properties of matured clay seals and implies that their hydraulic conductivity should be lower than
Fig. 3. Cases where high hydraulic gradients can prevail and damage freshly prepared clay and concrete seals. A and B are fracture zones with high water pressure.
134
R. Pusch et al. / Engineering Geology 202 (2016) 132–142
with different densities are given in Table 2 that we will come back to later in the paper. 3. Description of the method for deep hole sealing 3.1. Clay seals
Fig. 4. Critical geometrical condition for placing clay seals called “plugs”.
that of the rock excepting fracture zones. It is estimated at E−11 m/s at 1–2 km depth, E−12 m/s at 2–3 km depth, and E−13 m/s deeper than 3 km. The conductivity and swelling pressure of smectite-rich clay seals
3.1.1. Function and installation The sealing potential of expansive smectite-rich clay is fundamental to the described borehole sealing principle, which involves installation of perforated tubes that are tightly filled with blocks of very densely compacted smectite granules in the holes to be sealed (Fig. 1). Holes deeper than some hundred meters should be filled with soft smectite mud in the boring and waste placement phases for stabilizing the rock and for controlling the rate of maturation of the clay seals in the installation phase. The dense clay blocks in the perforated tubes made of copper, Navy Bronze or titanium in rock hosting repositories for highly radioactive waste like spent reactor fuel, and steel for most other purposes, absorb water from the rock and expand out through the perforation and embed the tubes (Fig. 5). This process is slow if the blocks are pre-saturated with water and quick if the blocks have only 50–60% degree of saturation. The mud left in the holes between the perforated tubes and the rock is consolidated under the pressure from the expanding clay and ultimately becomes almost as dense and tight as the clay
Fig. 5. Early stage of maturation. Upper: Schematic model of a “skin zone” of clay being formed outside the perforated tube. Lower. 8 h laboratory experiment with clay plugs with and without silicon coating for delay of hydration after 8 h in distilled water. The small picture shows the jointing of clay seal units.
R. Pusch et al. / Engineering Geology 202 (2016) 132–142
135
Table 1 Full-scale field experiments with montmorillonite-rich clay in tubes with 50% perforation ratio and 2 mm wall thickness. Dry density = ρd. Pusch (1994). Case/start of test
Time of maturation, months
Borehole/tube diameter, mm
Borehole/tube length, m
Initial ρd of clay, kg/m3
ρd of clay in tube at end of test, kg/m3
ρd of clay ”skin” zone at end of test, kg/m3
1/1986 2/1986 3/1986 4/2006
30 12 N240 30
56/54 80/76 80/76 80/76
100/100 14/2 160/30 5/2.5
1760 1301 1580 1710
1600a 1250a In position 1350
1500a 1200a In position 1270
a
Axial inner tubes with cables to gauges reduced the cross section and gave low density.
remaining in the tubes. The technique has been tested under laboratory conditions and in Swedish and Finnish underground rock laboratories for sealing of initially water-filled holes (Pusch, 2008). Typical field test data are summarized in Table 1, which shows that the dry density of the clay skin formed around the perforated tubes was up to 90% of the dense clay remaining in the tubes at the end of the tests. Figs. 6 and 7 illustrate the appearance of a 76 mm clay plug shortly after placement in a water-filled 80 mm metal tube for testing. 3.1.2. Rate of maturation The factors affecting the rate of maturation of the clay seals and thereby of the time for reaching different degrees of sealing of the holes are: • ability of the rock to supply the clay plug with water for keeping up the maturation rate, • water pressure, • hydraulic gradients that can cause piping and erosion, • the mineral composition, dry density, and initial degree of water saturation of the dense clay in the perforated tubes, and of the mud, • geometrical conditions like the perforation ratio of the tubes and the size of the gap between tubes and rock, and between tubes and clay.
Looking first at the processes associated with installing dense, unsaturated clay seals, one has to consider the role of the water pressure, which is 10 to 40 MPa in the depth interval 1–4 km. It causes some elastic compression of the clay, followed by expansion back when it becomes saturated with water. The transport mechanism of water from the rock is by flow through winding microstructural channels with a diameter of a few tens of micrometers from which water molecules migrate into the clay matrix by diffusion (Pusch, 2008, 2015; Pusch and Kasbohm, 2001). The driving force is the groundwater pressure and porewater suction (−150 MPa) in the dense clay2 while the hydraulic conductivity of the successively saturated clay controls the rate of saturation. Using common lab data it has been found that the entire dense clay cylinder in the perforated tube will be saturated in 2 weeks to 3 months depending on the hydraulic conductivity of the clay (E − 13 to E − 10 m/s). The part of the dense clay that is next to the perforated tube starts to swell through the perforation as indicated by Fig. 5, forming little clay columns that expand laterally and unite to become an increasingly homogeneous skin in the gap between tube and rock. It is successively consolidated under the pressure of the expanding dense clay and its shear strength, which determines if the clay seals can be forced down to the intended depth, grows with time. The process of dense clay moving out through the perforation of the tube is almost independent of the groundwater pressure. Saturation of the dense clay can be considered as a diffusion process as shown in Fig. 7. The data agree well with actually measured densities. 3.1.3. Role of geometry on the ultimate density of clay seals Table 1 shows that the dry density of the clay in the recovered tubes was about 20–30% lower than of the clay that was initially in the tubes 2
Clay Technology AB, Lund, Sweden.
placed in the water-filled holes because of migration of clay from them to the gap between tubes and confining rock where a “skin zone” was formed. Fig. 8 illustrates the impact of the dimension of the perforated tube on the average density and thereby on the sealing function of the ultimately matured clay seals. In mud-filled holes the effect is naturally smaller. The gap is only a few millimeters wide in practice but the properties of the clay filling will control the physical performance of the clay seal as a whole. The diameter of the boreholes to be sealed is naturally important: for any initial dry density of the clay in slim holes it will become lower than in wider holes where the metal tubes occupy a smaller fraction of the cross section (Table 2). In deep slim holes the amount of clay in seals will be further reduced by dispersion and erosion when they are brought down in water-filled holes as demonstrated by Fig. 9. Such degradation will be much smaller in mud since no particle migration into or
Fig. 6. Appearance of clay seal unit early after start of wetting.
Fig. 7. Evolution of the degree of saturation of 76 mm clay plug in 80 mm borehole at unlimited access to electrolyte-poor water assuming the process to take place by diffusion of water with the diffusion coefficient D = 3E−10 m2/s. Lower curve: 12 days, central curve: 72 days, upper curve: 120 days.
136
R. Pusch et al. / Engineering Geology 202 (2016) 132–142
Fig. 8. Impact of the initial dry density on the average ultimate density of the matured clay in a sealed hole with defined geometry.
Table 2 Theoretical ultimate density and major geotechnical properties (K = hydraulic conductivity, ps = swelling pressure) of smectite-rich seals in boreholes with different diameters assuming that the initial dry density is 2000 kg/m3 and that the gap between tube and rock and the thickness of the tube wall are 2 mm. Pusch et al. (2015). Borehole diameter, mm
Initial dry density, kg/m3
Final dry density, kg/m3
Final density at saturation, kg/m3
Final density at saturation if the borehole is 3 mm wider hole
K, m/s for 3.5% CaCl2 water
Swelling pressure, MPa
56 76 80 100
2000 2000 2000 2000
1683 1760 1785 1830
2060 2110 2124 2150
1942 1995 2037 2083
5E−12 1E−12 7E−13 3E−13
1.0 1.5 9.0 12.0
out from the perforated tubes will take place in the short time of placement of the seals. The actual variation in borehole diameter that has to be considered in selecting tube diameters is exemplified in Fig. 10. 3.1.4. Practical role of the maturation rate The risk of too quick or too slow maturation of the clay plugs is of great practical importance. Taking as an example the case of installing a 12 m long set of jointed clay plugs with 76 mm diameter in an 80 mm borehole, weighing about 900 kg in air and about 440 kg under water, one finds that their own weight is not enough to let it sink if the “skin zone” has a higher shear strength than about 1.5 kPa. In practice, it will reach this value in less than 1 h and it will rise to 50 kPa in 6 h as indicated by Table 3, which shows the required amount of water for maturation of a clay seal unit of 2.5 m length in a 76 mm borehole and the geotechnical properties as functions of time. In very tight rock that does not provide the seal with enough water the maturation will be slower. For deep holes the high water pressure is believed to bring enough water to the clay despite the lower conductivity at depth. After 6 h of maturation in a hole with unlimited access to water the seal has to be loaded by 15 metric tons for forcing it down, which can be made by using a big and robust drill rig. After 8 h, which is a reasonable time for bringing such a set down by 1–2 km, the required load is estimated to be 30 tons, which may be critically high for getting it down unless a vibrator is used in combination with loading. While installation of clay seals may be hindered by their growth in shear strength, quick maturation is desired for making it possible for placed seals to carry concrete cast upon them where fracture zones are intersected. The bearing capacity can be estimated by use of the classical Prandtl failure theory for predicting the bearing capacity of loaded soil (Peck et al., 1973). It implies that failure takes place if the pressure exceeds about 6 times the undrained shear strength. For a clay plug that has matured for 6 h the pressure causing failure of the “skin” zone is hence 300 kPa, which corresponds to the pressure exerted by a 30 m
high concrete fill weighing 2200 kg/m3 in air and 1200 kg/m3 under water. Failure will cause concrete to penetrate the “skin” clay and displace it upwards along the hole. This phenomenon has been recognized in a field experiment with a clay seal of 76 mm diameter in an 80 mm borehole when concrete was filled over it shortly after installing the clay seal (Pusch and Ramqvist, 2007), One finds, in conclusion, that the rate of maturation and the time for installation of long plug sets must be seriously considered for successful installation of the clay seals. 3.1.5. Ways of retarding maturation of clay seals A proposed method is to coat the perforated tubes with a thin paste of mixed smectite clay and talc (Pusch, 2011). The dense clay in the perforated tubes sucks water from the coating and seals off the perforation holes, which retards migration of water into the clay in the tube (Fig. 11). Once the slow migration of water through the coating has initiated expansion of the dense clay in the perforated tube, causing the coating to break, its uptake of water accelerates as for uncoated tubes. The talc fragments are “submicroscopic” and disseminated and have insignificant impact on the geotechnical properties of the “skin zone”. Current research indicates that prewetting of the dense clay blocks to give them an initially high degree of water saturation is possible and effective since the suction potential and wetting rate are then lower (Pusch, 2008; Yang et al., 2015). 3.1.6. Long-term stability of clay seals The long-term chemical stability of montmorillonite-rich smectite clay, which is taken here as primary clay candidate because of its excellent sealing properties, has been investigated in detail since the beginning of the previous century. American and European mineralogists like Roaldset3 and Lindgreen (1991), who investigated a large number 3
Personal communication.
R. Pusch et al. / Engineering Geology 202 (2016) 132–142
137
Table 3 Calculated amount of water for maturation of a 2.5 m long clay plug of the type shown in Fig. with the initial dry density 1700 kg/m3 and geotechnical properties as a function of time.
Fig. 9. Impact of erosion on the average density of the matured clay seal in Fig. 8 with 1825 kg/m3 initial dry density.
of samples of Tertiary age from the North Sea basin buried at large depth, reached the conclusion that smectite in layers from 1500 to 1700 m below the present sea bottom have only undergone insignificant transformation to less expanding illite/smectite (I/S) forms since sedimentation. At 2800 m depth, where the temperature has been more than 100 °C for tens of millions of years there has been more extensive conversion to mixed-layer minerals I/S and illite (Fig. 12). Some cementation by precipitated silica has taken place as well. The processes followed the generally accepted conversion model of smectite-to-illite in Eq.(1).
Smectite þ Potassium N Illite þ Silica
ð1Þ
We conclude here that largely preserved mineral composition over 100,000 years of a smectite-rich clay seal at depths of up to about 4 km, where the prevailing temperature can be 100 °C can be counted on, but that silica cementation will reduce its expandability. The criterion that the clay shall stay ductile and exert a contact pressure on the borehole walls will therefore not be fulfilled except if the boreholes undergo enough radial convergence by creep. This needs to be proven.
Fig. 10. Example of diameter variation in SKB's borehole KLX21B with 76 mm nominal diameter. “Bhln” denotes borehole length.
Time, hours
Density of “skin zone”, kg/m3
Amount of sorbed water, liters
Hydraulic conductivity of “skin zone”, m/s
Shear strength of “skin zone”, kPa
6 12 24 48 96
1100 1150 1325 1400 1700
10 15 33 40 70
E−9 5E−10 5E−11 2E−11 E−12
50 200 400 550 700
3.1.7. Clay materials considered for use in deep holes Application of the criteria that the clay seals must not be more permeable than the boring-disturbed surrounding rock and be enough ductile to exert an active pressure on the rock of at least 100 kPa, sorts out practically useful clay seal materials. Data from comprehensive laboratory investigations collected in Table 4 has been the basis of the preliminary assessment that for a rock conductivity of E − 12 m/s all the smectite-rich clay candidates, with montmorillonite as number one, will do provided that the initial dry density is 1850–1900 kg/m3. Applying a reasonable margin the minimum dry density would be 1950 kg/m3. It should be mentioned, however, that hydrothermal testing has indicated that saponite is more chemically stable than montmorillonite (Gueven and Huang, 1990), which is still preferred because of its lower cost and acceptable sealing ability.
3.2. Concrete seals 3.2.1. Role and construction The role of concrete seals is essentially to fill the parts of the boreholes that intersect fracture zones with erosion-resistant material for supporting clay seals and serve as filter for preventing particles from adjacent clay seals to migrate into the fracture zone and be transported away by flowing groundwater. Casting of concrete in boreholes can be made by use of containers that are emptied just on top of the latest installed clay seal. The discharge of concrete displaces the much lighter clay mud and fills the hole (Fig. 13). The concrete is prepared with fresh-water in batches immediately before it is pumped down.
Fig. 11. Borehole seal of very dense montmorillonite-rich clay coated with a paste of mixed talc/clay with the dry mass ratio 3.3/1 and a water content of 110% submerged in distilled water for 3 h. No migration of clay from the tube had occurred in the pasted part while a soft clay gel had formed between the tube and the plexiglass container in the untreated part.
138
R. Pusch et al. / Engineering Geology 202 (2016) 132–142
3.2.2. Cement and concrete recipe An early recipe proposed in SKB's Borehole Sealing Program (Pusch and Ramqvist, 2007) had the following components in kg per cubic meter of concrete: Portland cement (60 kg), silica fume (60 kg), fine α-quartz (200 kg), fine cristobalite (200 kg), Glenium 51 superplasticizer (4.4 kg in dry form), granitic aggregate 0–4 mm (1700 kg), and water (244.3 kg). The compressive strength was 4.5 MPa after 2 days and 31.7 MPa after 28 days at room temperature. Such concrete, having pH N 12, was cast on top of a clay seal in a 76 mm borehole, and analyzed after about 3 years with respect to the mineralogy and chemical composition of both clay and concrete. This showed significant degradation of both, by dissolution and cation exchange from Na to Ca of the clay, and by dehydration and fissuring of the concrete (Warr and Grathoff, 2010). These observations and the probability that organic fluidizers will degrade or be eaten by microbes led to the following composition for the laboratory study: low-pH cement,4 quartz aggregate, and talc as fluidizer. The respective contents were: cement 6.5%, talc 9.5%, and aggregate 84% (100% b 4 mm, 93% b2 mm, 70% b1 mm, and 30% b 0.1 mm), (cf. Pourbakhtiar, 2012; Pusch et al., 2013). A suitable water/cement ratio can be 3–4 for this composition having an aggregate/cement ratio of 10–15. This gives 2070 kg/m3 density and pH = 10. Typical compressive strength data after 2–28 days of curing are summarized in Table 5, together with data for a corresponding concrete with Portland cement. The microstructural constitution of this type of concrete reflects the plastic behavior of the talc phase in the earliest curing phase: the formation of skin coatings of quartz grains by flaky talc particles resulting in easy glide of both (Fig. 14). The subsequent strengthening provided by the talc is caused by precipitation of cementing complexes formed by reactions between low-pH cement components and slowly dissolving talc (Fig. 15). The hydraulic conductivity was measured after 7 and 21 days giving K-values of 5E − 9 and 8E − 10 m/s, respectively (Mohammed et al., 2013). Major conclusions from compressive tests using talc as fluidizer are: • concrete with low-pH cement harden much slower than Portland cement concrete, and has about 50% of the strength of such concrete after 7 days, but is 2 to 3 times stronger after 28 days and • both concrete types are brittle but the concrete with low-pH cement maintains a certain ductility and tendency to undergo creep strain in the curing time, implying tighter physical interaction with the rock than concrete with Portland cement.
Fig. 12. Smectite-to-illite conversion via mixed-layer I/S formation and/or direct precipitation of illite (Pusch, 1994). The dark contours represent precipitations of silica and/or illite.
It can be prepared by mixing the cement and aggregate with electrolytefree tap water to the water/cement ratio prescribed by the selected recipe or by mixing in “Dry water” in the mass of solid components (Forsberg et al., 2014). Such water consists of microscopic droplets confined in extremely thin silica “shells” that break and release water when compacted. It will be of no importance if some small amount of mud remains since it would react with the cement component of the concrete and stiffen.
3.2.3. Long-term performance of concrete The described grain-size distribution of the aggregate and the very low cement content gives a tight microstructure that is similar to bottom moraine and minimizes changes in physical performance of the concrete in the unlikely case of complete loss of the cement (Mohammed et al., 2013). Thus, the concrete has a very low compressibility and a pore size distribution that minimizes the risk of infiltration and through-transport of clay particles from contacting clay seals. The longevity of the magnesium/silicate cement reaction products from talc/cement interaction is expected to be very significant, hence making the risk of total loss of cement improbable. Table 6 summarizes the element concentrations in such concrete with and without talc after 7 days maturation in initially electrolyte-free water. The major difference between cement with and without talc is a loss in F, Al, S, and K and an increase in Na and Mg in the presence of talc. Compared with virgin cement the significant increase in Ca, Mg and Si and the decrease in Al were the most obvious findings. They indicate that the talc component was partly dissolved and the elements incorporated into new cementing compounds that formed in the first week of reaction. Further reactions caused the obvious strengthening recorded 4
MERIT 5000 which has 34% SiO2, 13% Al2O3, 17% MgO and 31% CaO (Pusch et al., 2003).
R. Pusch et al. / Engineering Geology 202 (2016) 132–142
139
Table 4 Clays for sealing 80 mm boreholes with 76 mm tubes down to 2 km depth in salt, Ca-rich groundwater. Geometrical changes associated with maturation are according to Fig. 3 but no chemically induced degradation is assumed. Clay
Initial dry density, kg/m3
Density of saturated homogeneous seal, kg/m3
Hydraulic conductivity of clay seal, m/s
Swelling pressure of clay seal, kPa
Montmorillonite Montmorillonite Saponite Saponite Mixed-layer montm./illite, 25–35% expandable Mixed-layer montm./illite, 25–35% expandable Mixed-layer montm./illite, 60–70% expandable Mixed-layer montm./illite, 60–70% expandable
1850 1500 1850 1500 1850 1500 1850 1500
2000 1800 2000 1800 2000 1800 2000 1800
8E−13 6E−11 7E−13 3E−12 2E−11 7E−11 5E−12 3E−11
4700 1000 5000 2000 700 200 2000 600
after 28 days. The results also imply that the dissolution of talc provided additional Mg for the cementing minerals. It is estimated that complete curing, requiring dissolution, diffusive migration of mobile elements and successive precipitation of reactive elements in the chemically transient porewater, can take years and decades and even longer time in big seals in large-diameter holes. 3.3. Performance of contacting concrete and clay seals — exploratory studies Field experiments running for almost 3 years have shown how concrete with Portland cement interacts with smectite-rich (montmorillonite) clay and what the impact is on the physical properties of the respective material (Pusch et al., 2003; Warr and Grathoff, 2010). The major finding was that the two materials caused significant changes within several centimeters from their contact, the most obvious reaction being partial loss of expandability of the clay and loss of strength of the concrete. Subsequent 8 week hydrothermal tests in the laboratory with the Danish mixed-layer I/S “Holmehus clay” (Pusch et al., 2015) of
Tertiary age, composed of smectite (60% montmorillonite), glauconite, quartz, feldspars, calcite, glass, kaolinite, Ti oxide and pyrite, in contact with concrete of the described type with low-pH cement, have shown the impact of temperature. The test arrangement, illustrated in Fig. 16, consisted of two completely but differently fluid-saturated clay samples with a dry density of 1500 kg/m3 in the hydrothermal cell, separated by concrete cast over the lower clay sample. The confining filters were connected to vessels with distilled water and 3.5% CaCl2, respectively. One test was conducted at room temperature, one at 100 °C and one at 150 °C. Some tests were made at higher temperature for accelerating the chemical processes and for simulating the conditions at depth in rock and they indicated, in contrast with the unheated materials, obvious mineralogical changes as indicated by the XRD diagram in Fig. 17 and scanning electron images in Fig. 18. These changes were also mirrored by the uniaxial compression tests reported in Fig. 19 and evaluated in the form of the stress/strain diagram in Fig. 20 of the room temperature sample. The 100 °C samples showed similar behavior while the 150 °C sample had about 25% lower peak strength, much higher stiffness, and obvious brittleness (Pusch et al., 2010). The clay samples prepared with calcium chloride solution behaved as the ones with distilled water but showed somewhat higher peak strength values and stiffness (Mohammed et al., 2014). An indication of the relative amounts of important elements, expressed in oxide form, in the contacting clay and concrete is given by Table 7. Compared with untreated I/S Holmehus clay the ones in contact with concrete had become significantly richer in SiO2 and poorer in Al2O3, hence being less expandable.
Fig. 13. Casting concrete by extrusion from a tube attached to the drill string with its open end held at the top of the latest placed clay seal.
Table 5 Average compressive strength in MPa of concrete prepared with the same low cement content. Pourbakhtiar (2012). Duration (days)
Portland concrete
Merit 5000 concrete
2 7 28
0.5–0.6 0.6–0.7 0.8–0.9
0.01–0.02 0.1–0.4 1.8–2.7
Fig. 14. Accumulation of talc particles oriented around a rounded quartz grain (lost) in the concrete (Institute of Geography and Geology, University of Greifswald, Germany).
140
R. Pusch et al. / Engineering Geology 202 (2016) 132–142
Fig. 16. Set of clay/concrete disks for hydrothermal treatment under isothermal conditions. Fig. 15. Talc particles in MERIT concrete. The smooth, slightly rounded edges of the talc indicate dissolution (Institute of Geography and Geology, University of Greifswald, Germany).
As to the clay, the pressure at failure was about half the swelling pressure, 1.5 MPa, of “virgin” Holmehus clay saturated with 3.5% CaCl2 solution. This indicates weakening of the clay by dissolution in contact with concrete. For the concrete the compressive strength 4.2 MPa was slightly higher than the average value for a 28 day samples (Table 5). This strengthening can have been due solely to continued talc/cement reactions. Random powder diffractograms of the clay samples that were in contact with concrete and of “virgin” salt clay had main dominant reflection in the clay of illite (10 Å) and of gypsum, which appeared to be abundant. Minor amounts of kaolinite and chlorite were present as well but less abundant than in the untreated “virgin” clay, which showed typical I/S peaks and a dominant quartz reflection. The mineralogical changes do not explain the recorded weakening, the most probable reason being partial conversion of crystalline matter to amorphous. 4. Conclusions The overall conclusion of the present exploratory study is that the proposed borehole sealing techniques and materials are feasible: very dense smectite clay in perforated metal tubes, and talc-based concrete cast where fracture zones are intersected. Application of the method can be of great importance for isolating deep repositories for highly radioactive waste and for preventing chemically contaminated water from “fracking” to reach up to shallow acquifers. The major findings of the study can be summarized as follows:
if the “wall friction” is too high. Too slow maturation can delay utilization of the isolation potential of the clay seals. The maturation rate of clay seals can, however, be effectively controlled, • concrete is required for providing stable support to the clay seals. It does not have to be low-permeable – the actual hydraulic conductivity being E−11 to E−10 m/s – but must have a microstructural constitution that prevents clay particles from adjacent clay seals to migrate into it and be washed out by flowing groundwater (Mohammed et al., 2013). Traditional binders like Portland cement are expected to last only for some decades and a new concrete, with talc instead of degradable organic fluidizer and with higher chemical stability, is proposed. The curing rate of the concrete seals determines the time until the next clay seal can be installed and the study indicates that placement of a clay plug over a concrete seal can be made after about one week. The density and granular composition of the concrete is such that even total loss of the small cement content does not jeopardize its capacity to provide support to the clay plugs and prevent migration and loss of clay particles into the intersected fracture zone, • the sealing effect of montmorillonite-rich clay seals in contact with the proposed concrete is sufficient in a 100,000 year perspective according to the exploratory study for up to 100 °C. For very deep holes heated to 150 °C by highly radioactive waste other smectite clays, like the Mg-rich saponite, may be preferable,
• borehole sealing by means of perforated copper tubes with montmorillonite-rich smectite clay with high density for long-term function can be effective as demonstrated by pilot field experiments with up to 500 m deep, steep boreholes and 100 m long subhorizontal boreholes, • in very deep holes the maturation of clay seals can be too quick or too slow, the first mentioned causing difficulties in getting the seals down Table 6 Element concentration (wt.%) of MERIT 5000 cement with and without talc, measured by energy dispersive X-rays (EDX). Warr and Grathoff, G (2011). Material
F
Na
Mg
Al
Si
S
K
Ca
Ti
Fe
Raw cement Cement without talc Cement with talc
– 1.12
– 0.07
2.63 1.14
1.57 1.11
4.46 6.28
0.11 0.32
– 0.13
1.37 8.84
– 3.00
– –
0.08
0.60
3.06
0.60
7.66
0.08
0.03
8.15
2.47
0.23
Fig. 17. Generalized X-ray diffraction pattern (red) of the Holmehus clay sample in the clay/concrete set compared to that of untreated (“virgin”) clay. Co-Kalpha radiation, d-values are in Å units. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
R. Pusch et al. / Engineering Geology 202 (2016) 132–142
141
Fig. 18. Scanning electron micrographs of clay at the contact with concrete (A), and of the concrete at this contact (B); (Institute of Geography and Geology, University of Greifswald, Germany).
Fig. 19. Compression stages at room temperature. Initial failure took place in the upper clay sample at 519 kPa pressure, corresponding to a shear strength of about 260 kPa of the clay saturated with distilled water. The concrete remained intact until the pressure was 4.2 MPa and the total compressive strain had become about 50%. Mohammed et al. (2014).
• the described sealing method has a potential to be used also in very deep holes – reaching down to 4 km – for storing highly radioactive waste (“VDH” concepts). The upper 2 km of such holes will not be
heated to more than 60–70 °C and can well be sealed by use of the described method while the evolution of the seals in the lower 2 km part heated up to 150 °C has to be further examined.
Acknowledgments The authors express their gratitude to Dr. Mohammed Hatem Mohammed, formerly at the Luleå University of Technology, Luleå, Sweden, for the assistance in preparing the report and for providing results and diagrams from his scientific work. Table 7 EDX compositions (weight of oxides, n = 30) of illite/smectite and cement phases across the clay/concrete contact at room temperature. Na2O
MgO
Al2O3
SiO2
K2O
CaO
TiO2
FeO
Sum
Smectite clay particles 21 °C 1.3
2.9
19.2
63.1
3.8
0.6
1.6
7.5
100
28.7
2.4
65.7
0.1
1.1
0
1.4
100
Temperature
Fig. 20. Stress/strain diagram for the compressed sample set at room temperature. A–D = compression to failure of clay with distilled water. E = compression to failure of concrete.
Cement phases 21 °C 0.4
142
R. Pusch et al. / Engineering Geology 202 (2016) 132–142
References Forsberg, T., Pusch, R., Knutsson, S., 2014. Rational method for preparing dense clay blocks with desired degree of water saturation. Engineering Geology, Technical Note. Vol.X, No.Y (pp.00). Gueven, N., Huang, W.-L., 1990. Effects of Mg2+ and Fe3+ Substitutions on the Crystallization of Discrete Illite and Illite/smectite Mixed Layers. Dept. Geosciences Texas Tech University, Exxon Production research Co, Houston, Texas, USA. Lindgreen, H., 1991. Elemental and Structural Changes in Illite/smectite Mixed-layer Clay Minerals During Diagenesis in Kimmeridgian–Volgian (–Ryazanian) clays in the Central Trough, North Sea and the Norwegian–Danish Basin. Bull. geol. Soc. Denmark Vol. 39, pp. 1–82 Copenhagen, October 25th 1991. Mohammed M. H., Pusch R., Al-Ansari N., Knutsson S., Emborg M., Nilsson M., and Pourbakhtiar A., 2013. Talc-based concrete for sealing borehole optimized by using particle packing theory. J. Civil Eng. Archit., Apr. 2013, Volume 7, No. 4, pp. 440–455. Mohammed, M.H., Pusch, R., Warr, L.N., Kasbohm, J., Knutsson, S., 2014. Interaction of Clay and Concrete in Deeply Disposed High-level Radioactive Waste (Manuscript Submitted to Construction and Building Materials). Peck, R.A., Hanson, W.E., Thornburn, T.H., 1973. Foundation Engineering. second ed. John Wiley & Sons, Inc. Pourbakhtiar, A., 2012. Pilot Study of Method for Constructing Concrete Seals and Fracture Grouts in Deep Boreholes and Cementitious Backfills in Tunnels, Drifts and Shafts in Crystalline Rock (M.Sc Thesis) Department of Civil, Mining and Natural Resources Engineering, Lulea University of Technology, Sweden. Pusch, R., 1994. Waste Disposal in Rock, Developments in Geotechnical Engineering, 76. Elsevier Publ. Co. (ISBN: 0–444-89449-7). Pusch, R., 2008. Geological Storage of Radioactive Waste. Springer-Verlag, Berlin, Heidelberg (ISBN: 978–3-540-77332-0).
Pusch, R., 2011. A technique to delay hydration and maturation of borehole seals of expansive clay”. Eng. Geol., Vol. 121, 2011, (pp. 1–6). Pusch, R., 2015. Bentonite Clay. CRC Press (Taylor & Francis Group) (ISBN 13:978–1-48224343-7). Pusch, R., Kasbohm, J., 2001. Can the Water Content of Highly Compacted Bentonite Be Increased by Applying a High Water Pressure? Technical Report TR-01-33, Swedish Nuclear Fuel and Waste Management Co (SKB) Stockholm. Pusch, R., Ramqvist, G., 2007. Borehole Projects — Final Report of Phase 3. SKB R-07-58, Swedish Nuclear Fuel and Waste Management Co. Svensk Kärnbränslehantering AB, SKB. Pusch, R., Zwahr, H., Gerber, R., Schomburg, J., 2003. Interaction of cement and smectite clay — theory and practice. Appl. Clay Sci. 23, 203–210. Pusch, R., Yong, R.N., Nakano, M., 2010. Stiffening of smectite buffer clay by hydrothermal effects. Eng. Geol. 110, 21–31. Pusch, R., Warr, L., Grathoff, G., Pourbakhtiar, A., Knutsson, S., Mohammed, M.H., 2013. A talc-based cement-poor concrete for sealing boreholes in rock, 2013. Eng. Geol. 5, 251–267. Pusch, R., Kasbohm, J., Hoang-Minh, T., Knutsson, S., Nguen-Thanh, L., 2015. Holmehus clay — a Tertiary smectitic clay of potential use for isolation of hazardous waste. Eng. Geol. 188, 38–47. Warr, L.N., Grathoff, G.H., 2010. Sealing of investigation boreholes: mineralogical and geochemical borehole plug analyses. SKB Report. Ernst-Moritz-Arndt-Universität, Greifswald, Germany. Yang, T., Pusch, R., Knutsson, S., Liu, X., 2015. Lab testing of method for clay isolation of spent reactor fuel in very deep boreholes. Proc. World Multidisciplinary Earth Sciences Symposium — WMESS 2015″, September, pp. 7–11 (Prague, Czech Republic).