Long-Term Performance Of Concrete For Waste Disposal Facilities

K (Ken) E. Philipose P. Eng.

AECL Research, Chalk River, Canada

INTRODUCTION

The disposal concept of the low-level radioactive waste repository planned at AECL Research, Chalk River, relies on the durability of concrete to isolate the waste from the human environment for a minimum of 500 years.

The durability of concrete and the corrosion of reinforcements have been studied for some time. Most of the principal factors that reduce the long-term integrity of reinforced concrete elements are well understood. Unfortunately, few studies can help predict the longevity of reinforced concrete structures over hundreds of years. Very little information in the literature is helpful in relating in a quantitative way the rate of degradation of concretes subjected to the principal characteristics of the environment. For these reasons, a research program on concrete durability was initiated by AECL Research in 1987 as part of the repository licensing support. The objective of the research program was not to develop a single highly durable concrete, but to design a high-performance concrete for the waste repository, and assess the durability of a wide variety of concrete types and qualities subjected to different environmental exposure conditions. The program is jointly conducted by AECL Research and the Institute for Research in Construction of the National Research Council (NRC), Ottawa.

METHODOLOGY

Traditionally, concrete durability has been assessed by measuring either the strength or length change of specimens subjected to a corrosive agent. Usually, the corrosive agent is applied at the external boundary of the specimen, but sometimes it may be included within the mix as an internal agent. Results from these types of tests are not applicable to lifetime predictions and are not sensitive to the design geometry of structural members. Tests for sulphate resistance for cements, for example, which involve measuring the expansion of specimens in which sulphates have been integrally included, yield information on whether the cement is suitable for use. These tests give no information on the rate of deterioration. The same criticism can be made of tests of standard laboratory-size specimens exposed externally to a corrosive agent, where failure is indicated by a given loss of strength or a given expansion. The measure of durability in these cases would be strictly relative and insensitive to geometry

The service life of concrete is dependent on a slow rate of deterioration and is influenced by the quality of concrete and the service environment. Factors such as cement type, cement content and water-to-cement ratio can affect the diffusion rate of ionic species into concretes. In addition, service life will depend on the size of specimens and failure criteria adopted. After examining the major failure mechanisms for the repository concrete, corrosion of reinforcement was selected as the mechanism for the failure of the structure. Chloride ions in the presence of oxygen can initiate corrosion of reinforcement and failure of the reinforced concrete components. The failure criteria chosen for the concrete was the time taken for the aggressive ions to reach the reinforcing steel by diffusion through the concrete cover (75 mm thick). Based on these criteria, the rates of deterioration and hence an assessment of the longevity of concrete can be made.

The rate of penetration of aggressive ions into the concrete was evaluated by determining the reaction zone front with time of exposure in the solution baths. Prediction of long-term concrete behaviour involves the extrapolation of current data, based on the assumption that long-term processes will not invalidate the extrapolation. The durability prediction chosen for the study was based on the time-dependent depth of penetration of chloride and sulphate ions into the test concretes. Concretes were selected so that the effect of cement blends containing silica fume or blast furnace slag on the diffusion rate of chloride or sulphates in concrete could be investigated and compared with the diffusion in Type 10 cement concrete.

During a post-closure period of hundreds of years, the repository structure will be subjected to various aggressive elements in the environment. Different parts of the structure will be subjected to chemical elements that will differ with time. For example, the environment inside the repository will be influenced by the chemicals leaching out of the waste, or generated by the waste, whereas the external environment will be influenced by the changes in the anion and cation content of the precipitation, due to the changes in acid rain and the addition of road salt. On the basis of an analysis of the repository service environment, the following major degradation parameters were selected for laboratory testing of concrete specimens:

· sulphate ions,

· chloride ions,

· leaching of calcium hydroxide by water,

· carbon dioxide reactions, and

· several agents in combination.

Concrete Specimens

Two concrete prisms, 75 x 75 x 280 mm, were cast for each mix and each exposure condition. S1, S2 and S5 were moist cured for 7, 14 and 28 days, respectively. Prior to immersion in the test solutions, the prisms were coated with wax on all sides but one, to allow a unidirectional ingress of chloride or sulphate ions.

Mortar Specimens

Mortar specimens in the form of 75 mm diameter cylinders were made from the three blended cement systems. For all the mortar systems, the sand/cement ratios were 1.6, 2.0, 2.4 and 2.8 for Mix 1, Mix 2, Mix 3 and Mix 4, respectively--approximately the values for the concretes. The curing times were also the same as the concretes. The air content was measured using the standard test ASTM C185. All specimens were mixed according to ASTM C230 at a flow of 110 percent using sulphonated formaldehyde of naphthalene as a superplasticizer when needed. Darex was used as the air-entraining agent.

Solution Baths

Of the 25 baths used in this study, the worst-case scenario for the repository concrete was simulated in the laboratory in Baths 2, 3, 9, 13, 21 and 24, containing various aggressive ions and ionic combinations.

RESULTS AND OBSERVATIONS

Durability Aspects

1. Properties of mortars

(a) Ca(OH)₂ content: The results for the pastes increase in the following order: S5<S2<S1. System 5, containing 75 percent slag and Type 50 cement has a very low Ca(OH)₂, while system 1, 100 percent Type 10 cement, has values greater than 13 percent.

(b) Total porosity: Porosity generally increased with the water/cement ratio.

(c) Median pore-size diameter: The value increases with the various cement systems in the following order: S5<S2<S1. This is the same as the ranking for Ca(OH)₂ content.

2. Diffusivity and resistivity

Generally, diffusivity increases with an increase in total porosity and with cement systems S5<S2<S1. System 5 concrete has the lowest diffusivity coefficient, as it has the lowest total porosity percentage. Generally, the resistivity increases with decreasing porosity and increases with the various cement systems as S5<S2<S1.

The data obtained from the SEM indicate that concrete System 5 ranks the lowest with respect to permeability, and provides maximum resistance to chloride attack. On the basis of the physical test results and the diffusion test data, System 5 mix 2 was selected as the candidate high-performance concrete for the repository construction.

3. Service life predictions

The service life of reinforced concrete structures exposed to sufficient chloride ions to initiate the corrosion of embedded reinforcement is largely controlled by the rate at which the chloride ions penetrate the concrete.

Table 1 lists the time in years required for the chloride ions to penetrate a depth of 75 mm into the concrete Systems 1, 2 and 5, based on the assumption of constant diffusivity during that period of time. The correlation coefficient of 0.8 or above in Table 6 provides higher confidence in the analysis results. Unlike the laboratory test specimens, the repository concrete will undergo microcracking or cracking due to imposed mechanical loads and other effects. The influence of cracks in concrete on the rate of ionic ingress has to be taken into consideration for the final assessment .

Ionic profiles and depth-of-penetration measurements (determined by EDXA) in concrete show that reasonably accurate results can be obtained and predictions of ionic ingress made. There is some scatter in the experimental results, because of the difficulty of locating the reaction front in concrete test specimens, due to the tortuous path of ionic ingress through dense concrete. In addition, the rate of movement of the front can be speeded up by the rapid diffusion of ions in interfacial regions and in cracks. However, there is enough consistency and redundancy in the system to obtain fairly accurate results. The procedure following the diffusion path around the fine and coarse aggregate particles, using the scanning electron microscope and electron microprobe for analysis, has been successful.

CONCLUSIONS

The following can be concluded from the experimental test data:

1) Hydrated blended cements mortars have diffusivities up to 25 times lower than equivalent Type 10 hydrated Portland cement mortars. A 75% slag system generally yields the lowest values of diffusivity among the blends.

2) Median pore diameter and Ca(OH)₂ content are ranked in the same order for the three cement systems (S5<S2<S1), and are similar to the ranking for electrical conductivity and diffusivity.

3) Lower water-to-cement ratios in concrete systems decrease the diffusion rate of ions, and sulphate ions inhibit the rate-of-penetration of chloride ions.

4) The rates-of-penetration of ions increase with salt concentrations, and the durability of concrete increases with an increase in strength in a given system; however, strength may not be an indication of durability between concrete systems.

5) On the basis of experimental studies to-date, the System 5 concretes rank the lowest with respect to permeability, and provide maximum resistance to chemical attack and has a service of over 500 years.

Table 1: Time in Years for 75 mm Depth of Chloride Ion Penetration

by Extrapolation - Exposure in Bath 13

Concrete System	Mix 1	Mix 2	Mix 3	Mix 4
System 1 Concrete	50 (.9)	55 (.86)	10 (.98)	15 (.97)
System 2 Concrete	64 (.91)	1875 (.4)	25 (.95)	33 (.94)
System 5 Concrete	650 (.83)	2600 (.65)	170 (.92)	200 (.81)

Note: Figures in brackets represent correlation coefficients