Microstructure Development and Transport Properties of Portland Cement-fly Ash Binary Systems : In view of service life predictions

Abstract

Fly ash is a by-product of burning coal in electric power generating plants. It is commonly known that owing to its pozzolanic properties fly ash is widely used as a partial replacement for Portland cement in concrete. The use of fly ash in concrete not only reduces the landfill costs of fly ash, but also reduces the use of Portland cement in concrete, consequently reduces CO2 emission per ton concrete. More important, the presence of fly ash improves the durability of concrete and extends the service life of concrete structures. Today, there is a demand for concrete structures with a service life of 80, 100, or even 200 years. In many cases chloride-induced rebar corrosion is assumed to be the dominant mechanism determining the service life of reinforced concrete structures. It is commonly believed that fly ash concrete has a better resistance to chloride penetration than Portland cement concrete, since the microstructure development of Portland cement-fly ash binary systems is different from that of pure Portland cement system. The resistance of concrete to chloride penetration is highly related to their microstructure. The studies on Portland cement-fly ash binary systems (concrete or paste) have been carried out for many years. Most studies are based on experimental results at a relative short curing period (i.e. 3 months) or from different concrete mixtures with different fly ash and Portland cement. The advantage of using fly ash, however, becomes evident at later ages, i.e. beyond 90 days. Systematic long-term investigations on Portland cement-fly ash binary systems are still limited. In this thesis the research on these binary systems starts from the hydration process (chapter 3), the microstructure development (chapter 4) to transport properties (water permeability and chloride penetration) (chapter 5- chapter 8) in view of service life predictions of concrete structures made with fly ash-blended cements. In a fly ash cement paste there are two types of chemical reactions: hydration of cement and pozzolanic reaction of fly ash. The pozzolanic reaction of fly ash needs calcium hydroxide (CH), produced by the hydration of cement, to occur. The evolution of the amount of CH with time reflects the rate of hydration of cement and pozzolanic reaction of fly ash in binary systems. As discussed in chapter 3 at early ages, i.e. before 7 days, the CH content of blended cement paste was higher than that of Portland cement paste. It indicated that the presence of fly ash leaded to faster hydration of cement in binary systems. After about 7 days, the CH content in blended cement paste decreases significantly. It suggests that in binary systems the rate of the pozzolanic reaction of fly ash (consuming CH) is faster than that of the hydration of cement (producing CH). At later ages, i.e. beyond 180 days, the CH content in blended cement paste stays at a constant low level. It is inferred that beyond 180 days the rate of the pozzolanic reaction of fly ash in binary systems becomes very slow. The pozzolanic reaction of fly ash results in a different microstructure development of blended cement paste compared with pure Portland cement paste. In chapter 4 the evolution of the pore structure of Portland cement paste and blended cement paste was investigated at ages up to 3 years. The porosity of blended cement paste was higher than that of pure Portland cement paste, even at an age of 3 years. At later ages, i.e. after about 28 days, the presence of fly ash results in the formation of a large amount of small capillary pores in the range between 10 and 100 nm. At later ages, i.e. after 180 days, blended cement paste had a lower connectivity of the pores than Portland cement paste. The pore structure of blended cement paste was refined at later ages while the porosity of blended cement was still higher than that of Portland cement paste (at ages up to 3 years). The microstructure of paste determines the transport properties. In chapter 5 the water permeability of Portland cement paste and blended cement paste was studies. At early age the pastes containing fly ash exhibit a higher capillary porosity than pure Portland cement paste. The initial water permeability of blended cement paste is higher than that of Portland cement paste. However, after about 180 days blended cement paste is less permeable than pure Portland cement paste, even though the capillary porosity of blended cement paste is higher than that of Portland cement paste. The water permeability of pure cement paste and blended cement paste depends on the connectivity of the pores. At later ages, i.e. after 180 days, the connectivity of the pores of blended cement paste is lower than that of pure Portland cement paste, resulting in a less permeable microstructure. In chapter 6 the resistance of Portland cement concrete and fly ash concrete to chloride penetration was investigated. Under moist curing conditions the DRCM values of Portland cement concrete made with different w/c ratios (0.4, 0.5 and 0.6) decrease with time at early ages, i.e. from 28 to 180 days. After that the DRCM values of Portland cement concrete increase and then turn to decrease again after around 1 year (Figure 6.6). The possible reason might be the delayed ettringite formation in Portland concrete when limestone powder (as filler) is blended with Portland cement clinker and when it is cured under moist conditions (see chapter 7). The DEF results in a change of the microstructure of hydrated cement paste and an increase of DRCM at later ages. At ages beyond about 28 days the concrete mixtures made with fly ash have better resistance against chloride penetration than Portland cement concrete. Ettringite forms in fly ash concrete at later ages. This ettringite is found in voids initially present in the paste and in the spaces left after the reaction of fly ash particles. Formation of ettringite in empty spaces explains why DEF in fly ash concrete does not lead to expansion and micro-cracking and an associated increase of the DRCM values as observed for Portland cement concrete. Based on the measured DRCM values for Portland cement concrete and fly ash concrete, the ageing factor n was determined (see chapter 8). It represents how rapidly the chloride migration coefficient of the concrete decreases with time. In DuraCrete the 28 days values of DRCM and n are two important input parameters to predict the service life of concrete structures. An important question is whether new n-values, as those determined in this study, can directly be adopted in the currently used version of DuraCrete for service life predictions. In DuraCrete it is assumed that the chloride diffusion (migration) coefficient of concrete, D(t), decreases considerably with increasing age of the concrete. This decrease is quantified with a constant value of n. It means that D(t) would go to zero as time tends to infinity (t ? ?), which is known not to be realistic. In reality, the chloride diffusion coefficient D(t) is directly determined by the microstructure of concrete. In fact, the decrease of the diffusion coefficient cannot be described adequately with a constant value of n. A more accurate description of the evolution of the diffusion coefficient D(t) with an not constant value of n, however, will affect the consistency of the currently used version of DuraCrete. A reconsideration of n-values should be accompanied by reconsidering the values of other model parameters values (e.g. environmental factor ke and curing factor kc in DuraCrete), since these parameter, ke, kc and n, are mutually interdependent.