Versão online: http://www.lneg.pt/iedt/unidades/16/paginas/26/30/185 Comunicações Geológicas (2014) 101, 1, 71-74 ISSN: 0873-948X; e-ISSN: 1647-581X

A study on cement-poor concrete with talc for borehole sealing in rock hosting radioactive waste Estudo sobre betões com talco e pobres em cimento para a selagem de furos em rochas armazém de detritos radioactivos R. Pusch1*, L. Warr2, G. Grathoff2, A. Purbakhtiar1, S. Knutsson1, G. Ramqvist3 Recebido em 20/07/2012 / Aceite em 04/09/2012

Disponível online em Dezembro de 2014 / Publicado em Dezembro de 2014

© 2014 LNEG – Laboratório Nacional de Geologia e Energia IP

Abstract: The paper describes the performance of a concrete intended for sealing deep boreholes in the host rock of radioactive repositories. The concrete will form plugs where fracture zones are intersected and be located between very tight seals of smectite clay installed where the surrounding rock is tight. The concrete must be able to carry the clay segments after a couple of days but the bearing capacity does not have to be very high since the clay soon adheres to the rock and carries itself. The concrete contains talc as superplasticizer since ordinary organic additives for reaching high fluidity at casting are unwanted. It has a low cement content for maintaining its low porosity after dissolution in a long time perspective.

Keywords: clay, concrete, ductility, strength, talc. Resumo: O trabalho descreve o comportamento de um tipo de betão destinado à selagem de furos profundos em rochas previstas para servirem de armazém de detritos radioactivos. O betão deverá servir de tampão nas zonas intersectadas por falhas e ser colocado entre a selagem de argila esmectítica onde a rochas está muito compactada. O betão deverá ser capaz de suportar os segmentos da argila durante alguns dias mas a sua capacidade de suporte não precisa ser muito grande desde que a argila adira rapidamente à rocha. O betão contém talco como super-plastificante, uma vez que os aditivos orgânicos habituais que conferem alta fluidez à moldagem não são apropriados. Têm uma percentagem pequena de cimento com vista a manter baixa porosidade após dissolução numa perspectiva de longo prazo.

Palavras-chave: argila, betão, ductilidade, resistência, talco. 1Drawrite AB/Luleå University of Technology, Sweden 2Greifswald University, Geogr. & Geological Dept., Germany 3Swedish Nuclear Fuel and Waste Handling Co (SKB), Äspö Underground Laboratory, Sweden *Corresponding author/Autor correspondente: drawrite.se@gmail.com

1. Background

Deep investigation boreholes for site selection of repositories for high-level radioactive waste are proposed to be sealed with series of concrete and clay plugs for preventing released radionuclides to reach the biosphere. Laboratory and field investigations show that both materials undergo degradation where they are in contact and that reduction of the cement content in the concrete seals is desired. Their bearing capacity must still be sufficiently high to carry the clay seals, which consist of perforated tubes filled with compacted smectite-rich clay. Casting of the dense concrete requires use of superplasticizers. Organic ones are unwanted since they give off colloids that can move radionuclides to the biosphere. New types of concrete have to be developed and one with talc as superplasticizer has been investigated in this study

for finding out if it can serve acceptably. Following a principle proposed for sealing of long boreholes in

a rock mass hosting a repository for high-level radioactive waste (HLW), those parts of the holes where the rock has few fractures and a low hydraulic conductivity are preferably sealed with clay while those intersecting permeable fracture zones require mechanically strong concrete (Pusch and Ramqvist, 2003), cf. figure 1. The average number of clay and concrete seals in 500 to 1000 m long holes is on the order of 10-20. The concrete does not have to be completely tight since the fracture zones have a potential to undergo tectonically induced strain and seals placed in them will be sheared and eventually suffer loss of structural integrity.

The principle of placing clay plugs of the type mentioned in deep boreholes with 80 mm diameter and casting ordinary concrete over them has been successfully tested in a 500 m deep hole at Olkiluoto, Finland (Pusch and Ramqvist, 2007). Further testing is required for development of the technique so that it can be used in holes reaching down to 1000 m and more. Installation of clay seals in such holes will expose them to erosion by adflowing water, which can, however, be largely eliminated by coating with smectite/talc paste (Pusch, 2011).

0

100

200

300

400

500

600

700

800 m

Concrete

Clay

Concrete

ClayConcrete

Clay

ConcreteClay

ConcreteClayConcrete

Clay

Concrete

Fig.1. Proposed distribution of clay and concrete-sealed parts of a deep borehole at Forsmark (Borehole KFM07A, Pusch and Ramqvist, 2003). Fig.1. Proposta de distribuição da argila e do betão nas zonas de selagem num furo profundo em Forsmark (Furo KFM07A, Pusch and Ramqvist, 2003).

The function of the concrete is to provide axial support to the clay seals, which requires that it remains a mechanical buffer and does not suffer corrosion, i.e. a criterion that requires high density

Artigo original Original article

72 R. Pusch et al. / Comunicações Geológicas (2014) 101, 1, 71-74

and a minimum of cement. A further criterion is that it must not significantly degrade the contacting clay seals. Such impact is caused by the alkali nature of the concrete solution that originates from the alteration of Portland cement (Warr and Grathoff, 2010) as concluded from a three year field experiment (Fig. 2). This study indicated extensive alteration of the smectite clay by dissolution and cation substitution, with transformation to non-expanding clay minerals and amorphous silicious compounds. Such reactions led to significant loss in the strength of the concrete within a distance of a few centimetres from the clay/concrete contact.

Clay plug of basictype

Q/C concrete of KR-24 type

Sand/gravel56 mm borehole extending to water-bearing zone under pressure. The hole is filled with gravel

Note: The 80 mm hole is filledwith tap water before the clayplug is inserted, and the Q/C plugis cast 24 hours thereafter

0.8 m

1.5 m

2.5m

0.2 m

5 m

Packer

Fig.2. Field experiment with clay plug consisting of highly compacted smectite-rich clay in perforated copper over which concrete of quartz/cement (Q/C) was cast to create a mechanical support to the underlying clay plug. Fig.2. Ensaio de campo com um rolhão de argila consistindo em argila rica em esmectíte muito compacta, em recipiente perfurado de cobre, sobre o qual o betão com quartzo/cimento (Q/C) foi introduzido para criar um suporte mecânico para o rolhão de argila colocado inferiormente.

2. First results of a new concrete for sealing deep boreholes

2.1 Criteria

New types of concrete are under development and one is described in the present paper. It contains a small amount of Portland cement (<15%), has quartz as the dominant aggregate mineral, and the mineral talc that provides fluidity. The latter ingredient can be used to replace organic superplasticizers, such as Glenium 51, which is unsuitable due to the organic colloids that may be released. Talc is a soft, hydrated magnesium silicate (Mg3Si4O10(OH)2) with a sheet silicate crystal structure and it has a characteristic hydrophobic nature (Wallqvist et al., 2006). Addition of talc to mixtures of common concrete constituents, i.e. aggregates of granitic type (quartz, feldspars and heavy minerals), cement and water, would theoretically lead to a fluidity that is similar to that of concrete with organic superplasticizers when using the same water/cement ratio (w/c). The major aim of this pilot project was to see if the proposed principle of using talc for reducing the viscosity of freshly prepared concrete is practical and fulfils the following criteria:

the concrete must be fluid directly after preparation for casting in boreholes;

the concrete must be coherent and not undergo phase separation;

the concrete should harden to a compressive strength of at least 0.5 MPa in 48 hours;

the cast concrete must be physically stable under water after hardening.

Since a long-term function of the concrete is required for

100,000 years, according to criteria specified by the Swedish Nuclear Fuel and Waste Handling Co (SKB), (Pusch and Ramqvist, 2003, 2006) and the cement component will dissolve and be lost in a foreseen time period, the rest, i.e. the aggregate components must still provide continuing support for the neighbouring sections of clay seal. These components must therefore be very dense and have a granular composition to resist erosion.

2.2 Pilot tests

Exploratory tests with 12.4% ordinary Portland cement mixed with 23% rock flour (<50 m), 37% fine sand (50-250 m), 23% coarse sand (250-1000 m), and 4.6% talc (Univar Talc Haichen No1) was conducted for finding out if this hydrophobic mineral can serve as a superplasticizer. The water/cement ratio was 1.0. Load testing made 3 days after casting gave a compressive strength of 3.5 to 4.2 MPa. The appearance of the freshly prepared concrete is shown in figure 3 and the hardened samples loaded to failure in figure 4. The failure pattern indicates that it behaved as brittle material.

Fig.3. Semi-liquid concrete prepared with w/c 1.0 and a cement/aggregate ratio of 0.15. Fig.3. Betão semi-liquído preparado com w/c 1.0 e uma relação de cimento/agregado de 0.15.

Fig.4. Typical failure mode showing the formation of axial fractures under 1.0 MPa pressure. Fig.4. Aspecto da fissura típica mostrando as fracturas axiais sob uma pressão de 1.0 MPa.

2.3 Main tests

A main test series was made with concrete prepared according to table 1, employing considerably less cement than in the pilot test. The amount of water added to the materials was selected to represent: A) condition for filling forms without compaction or vibration; B) normal fluidity; C) condition requiring vibration for filling forms.

The samples had a cement/aggregate ratio of 0.07, i.e. about half the value for the samples in the pilot test series. They were stored in a moist environment and loaded after 2, 7 and 60 days for determining the unconfined compressive strength (Tab. 2). Sample

Cement-poor concrete with talc for borehole sealing in rock hosting radioactive waste 73

A had a compressive strength of 0.38 MPa after 2 days increasing to 1.0 MPa after 7 days, undergoing no further strengthening up to 60 days. Sample B with the highest w/c ratio (1.3) but nearly the same cement/aggregate ratio as Sample A had a compressive strength of 0.19 MPa after 2 days and 0.5 MPa after 7 days and 0.6 MPa after 60 days. The compressive strength of Sample C, which had the lowest w/c (0.6) but about the same cement/aggregate ratio as A and B was 0.5 MPa after 2 days and 1.7 MPa after 7 days. The data for all the tests are compiled in table 2, which shows that the strength approximately doubled after 5 days from the first compression event compared to what took place 2 days after casting. However, no or very limited additional improvement took place in the subsequent 7 weeks. The pH of the freshly prepared concrete was determined to be 12.

Table 1. Composition of second generation of talc/cement concrete.

Tabela 1. Composição da segunda geração de betão com talco/cimento.

1Univar Talc Haichen No1 2The material was the same as in the pilot test series

Table 2. Compressive strength results in MPa for the main test series.

Tabela 2. Resultados da força compressive em MPa, para as series dos principais testes.

Sample 2 days 7 days 60 days

A 0.38 1.0 1.0

B 0.19 0.5 0.6

C 0.50 1.7 1.2*

*Slightly non-parallel basal planes of compressed sample

The A- and C-samples were less ductile than B. The sample A) had a consistency required for filling forms without compaction or vibration, sample B) was sufficiently fluid for being pumped over longer distances, and sample C) was rather stiff but judged to be suitable for casting by use of vibrators. The fluid Sample B started stiffening already after about 6 hours and was somewhat ductile even after 1 day. Its creep properties after 60 days were tested by stepwise application of an axial load with recording of the strain for 5 minutes per load step as shown in figure 5. The large strain before failure that took place was due to the ductile behaviour caused by the talc.

2.4 Mineralogy

In order to test the effect of talc on the cementation reaction we analyzed the material prepared following the mixture of Case C (Tab. 1), with 5% of Portland cement, 11% talc of 95% purity, and a water to cement ration of 0.6. We analyzed the starting materials as well as the cement products that were set for 7 as well as 60 days with powder X-ray Diffraction (XRD). The results for the 60 day old cement are shown in figure 6. The minerals identified are ettringite, portlandite, larnite, brownmillerite, calcite and quartz. The talc-bearing cement contains exactly the same cement phases present in similar concentrations. Therefore the addition of talc appears to have no significant effect on cement phase development. The analyses of the 7 day old cement showed similar characteristics to that of the 60 day mineralogy.

0

1

2

3

4

5

6

7

Strain, multipliedby E5

Time, minutes

0 20 40 60 80 100 120 140 160 180

480 N

580 N

Failure

680 N

780 N

Fig.5. Creep strain of Sample B under stepwise increased uniaxial force. The material showed ductility but the failure, reached for a pressure of 1.1 MPa, was of the type shown by the failed sample in figure 3, i.e. indicating brittleness. Fig.5. Força de deslizamento da amostra B sob uma força uniaxial crescente. O material mostrou ductilidade, mas as fissuras, chegando à pressão de 1.1 MPa, foram do tipo mostrado na figura 3, i.e. indicando que é frágil.

Fig.6. Powder X-ray Diffraction patterns of two cements (without added aggregate) that set for 60 days. Top pattern is of solely Portland cement, the middle pattern is the cement with 11.7% talc. The bottom scan shows the relative difference between the two. The talc peaks are all labeled. The other peaks all belong to cement minerals including portlandite (Port), ettringite (E), brownmillerite (B), larnite (L) and others as well as quartz (Q) and calcite (C). The most intense ones are labeled. Instrument parameters: Cu-K α radiation, 40 kV 30 mA, Ni filter, Lynx Eye detector. Fig.6. Padrões de difracção de RX provenientes de pós de dois cimentos (sem junção de agregado) num intervalo de 60 dias. O padrão superior é só de cimento Portland, o padrão do meio é de cimento com 11.7% de talco. O padrão inferior mostra a diferença relativa entre os outros dois. Os picos do talco são todos assinalados. Todos os outros picos pertencem a minerais do cimento incluindo portlandite (Port), ettringite (E), brownmillerite (B), larnite (L) e outros como o quartzo (Q) e calcite (C). Os picos mais intensos estão assinalados. Parâmetros instrumentais: Cu-K radiação α, 40 kV 30 mA, filtro de Ni, detector Lynx Eye.

3. Discussion and conclusions

3.1 Role of the mineral constituents

In the case of underground storage of radioactive waste, concrete seals are necessary where deep boreholes intersect water-bearing fracture zones. To be effective the concrete must remain coherent when being cast, the puzzolanic component should be uniformly distributed and the cohesive strength sufficient. These properties are difficult to achieve without the help of superplasticizers, and here we propose talc as a safer alternative to the current organic additives used (e.g. Glenium 51). The challenge of finding alternative

74 R. Pusch et al. / Comunicações Geológicas (2014) 101, 1, 71-74

superplasticizers is particularly important when using concrete with very low cement content.

Microstructural investigation of the talc-cement by scanning electron microscopy (methodology described in Warr and Grathoff, 2011) revealed a number of features reflective of the plastic behaviour of an additional talc phase. The most obvious findings were the formation of skins (cutans) of talc particles oriented parallel to the surface of aggregate grains (Fig. 7), and of several tens of micrometer long shear zones of roughly oriented talc particles (Fig. 8). These features indicate the flow alignment of talc contributed to the plastic deformation of the localized cement matrix: a key process required for attaining low viscosity and eventual high compressive strength. The flow behaviour of the talc can be primarily attributed to its platy shape, hydrophobic character and very low value of frictional coefficient (Moore and Lochner, 2008; Prasad et al., 2010) that result in easy sliding and alignment of particles.

Fig.7. Residue of talc particles oriented around a spherical quartz grain (not shown) in the concrete. Fig.7. Resíduo das partículas de talco orientadas em redor de um grão de quartzo esférico (não mostrado) no betão.

Fig.8. Micrograph showing high degree of filling of the void between densely located aggregate grains and a shear zone of oriented talc particles. The about 100 m long zone starts in the lower left and extends to the upper right. Fig.8. Micrografia mostrando o alto grau de enchimento dos espaços vazios entre os grãos agregados e uma zona de cisalhamento com partículas orientadas de talco. Aquela zona tem cerca de 100 m de comprimento, começando em baixo à esquerda e indo até acima à direita.

The shape of the talc particles in the cement indicates that some dissolution has occurred. The edges are commonly irregularly curved and smooth (Figs. 7 and 8), in strong contrast to the straight crystal edges observed in the raw talc material (Fig. 9). Such dissolution would be expected to release more Mg and Si into solution, that then contribute to the crystallization of new cement minerals.

The effective axial pressure from the entire fill in a 1000 m deep borehole acting on a concrete seal at the bottom will not exceed 5 MPa according to silo pressure theory (Ravenet, 1981),

which gives a total axial compression of less than 0.1% over thousands of years even if the cement is dissolved.

Fig.9. SEM images of the raw talc material that was added to the Portland Cement. Fig.9. Imagem SEM de partículas de talco natural que foi adicionado ao cimento Portland.

3.2 Conclusions

Concrete for sealing boreholes can be prepared with sufficiently high density, at least 2200 kg/m3, corresponding to that of bottom morains, to resist axial compression.

Talc as mineral component serves as an alternative superplasticizer and as conditioner of the concrete making it cohesive and less sensitive to variations in water content than ordinary concrete.

Mixtures of the investigated types presented, giving concrete with a compressive strength of around 0.5 MPa after 2 days and 1.0 MPa after 7 days, can be taken as a basis for further development for constructing concrete seals without the need for organic superplasticizers.

The ductile behaviour of the matrix of the cement-poor talc concrete determines its properties in the preparation phase and during initial hardening. Its compressive strength is developed early, i.e in about a week, and then remains constant for the subsequent three weeks. The very dense matrix of aggregate grains indicates that even after complete loss of cement over a long time period it would not be expected to cause spontaneous compression of the material.

References

Moore, D. E., Lochner, D. A., 2008. Talc friction in the temperature range 25°-400 °C: relevance for fault zone weakening. Tectonophysics, 449, 120-132.

Pusch, R., Ramqvist, G., 2003. Sealing of investigation boreholes. Final Report, SKB Technical Report, Phase 3. SKB, Stockholm.

Pusch, R., Ramqvist, G., 2006. Cleaning and sealing of Borehole. Report of Sub-project 3 on plugging of borehole OL-KR24 at Olkiluoto and reference boreholes at Äspö. SKB IPR-06-30 Report, SKB, Stockholm.

Pusch, R., 2011. A technique to delay hydration and maturation of borehole seals of expansive clay. Engineering Geology, 121, 1-6.

Prasad, B. K., Rathod, S., Modi, O. S., Yadav, M. S., 2010. Influence of talc concentration in oil lubricant on wear response of bronze journal bearing. Wear, 269, 498-505.

Ravenet, J., 1981. Silo Problems, Bulk solids handing. Silos. Bins and Bunkers, 1, 4, 667-679.

Wallqvist, V., Claesson, P. M., Swerin, A., Schoelkopf, J., Gane, P., 2006. Interaction forces between talc and hydrophobic particles probed by AFM. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 277, 183-190.

Warr, L. N., Grathoff, G., 2010. Sealing of investigation boreholes: mineralogical and geochemical borehole plug analyses. SKB Report, 24.

Warr, L. N., Grathoff, G., 2011. Geoscientific applications of particle detection and imaging techniques with special focus on the monitoring clay mineral reactions. In: Handbook of Particle Detection. Springer, 669-683.