Mix Design, Fresh and Hardened Properties and Microstructural Characterisation of Alkali- Activated Concrete Based on PFA/GGBS Blends

  • Ali Rafeet

    Student thesis: Doctoral ThesisDoctor of Philosophy


    Alkali-activated concrete (AAC) has recently emerged as a potential candidate to replace the conventional Portland cement concrete (PCC) in several applications due to its low CO2 footprint and promising mechanical properties. However, the robustness and the good reputation PCC has gained over the years make it difficult for this relatively new technology to be adopted at a larger industrial scale. The absence of a widely accepted standardised mix design method for AAC also limits its commercialisation. Moreover, some technical problems such as the need for elevated temperature curing and the lack of sufficient data on both hardened and fresh properties of AAC among others played a major role in limiting the industrial uptake of this green alternative. In this thesis, fly ash from coal fired power stations and ground granulated blast furnace slag from iron production have been investigated as binders to replace Portland cement. The influence of salient parameters such as percentage of GGBS in PFA/GGBS blend, paste content, binder content, water-to- solid ratio (w/s) on setting time, workability and compressive strength of AAC is investigated. The findings obtained from this investigation helped to suggest mix design guidelines for PFA/GGBS AAC to achieve a wide range of compressive strength (20-70 MPa) and consistency classes (S1-S5) with “optimized” binder content in the range of 320-350 kg/m3.

    Eventually, AAC building blocks were produced as a demonstration product to assess its competitiveness in terms of industrial production process, technical requirements and economical feasibility.

    Reaction kinetics studies and microstructural characterization of both low and high GGBS content were carried out to examine the effect of chemical dosages and GGBS content on the reaction kinetics, reaction products and morphology of neat PEA and both low and high GGBS content blends. Isothermal calorimetry (IC) analysis showed that lower GGBS systems (less than 20%) displayed very low reactivity at room temperature and the heat of reaction increased gradually with the increase in GGBS content. Furthermore, 100% PEA systems activated with higher activator dosages (M+ 11.5% AM 0.85) showed more compact and dense microstructure compared to those activated with lower dosages (M+ 7.5% and 8.5%). The main reaction product observed in neat PEA systems was low Ca geopolymer gel (N-A-S-H). Blends with 20% GGBS content and higher exhibited higher reactivity and the extent of reactivity increased with the increase in GGBS content. Higher GGBS blends i.e. 60/40 and 30/70 showed solid and compact microstructure compared to low GGBS blends in agreement with higher compressive strengths obtained in those blends. C-A-S-H binding gel was detected as the main reaction product in blends with 20% GGBS and higher. EDAX analysis showed the formation of both low Ca and high Ca gel which may suggest the coexistence of both N-A-S-H and C-A-S-H binding gels. The formation of a hybrid gel (C-N-A-S-H) with different amount of Ca depending on the amount of GGBS is also likely as detected by EDAX. Morphology of 30/70 blend activated with NaOH solution only (no soluble silicate) showed higher extent of unreacted particles suggesting lower degree of reactivity. This can be due to the lack of Si as it was found to affect the dissolution of A1 and hinder the availability of A1 in the gel nucleation. The result was low compressive strength of around 30 MPa which was around 3 times lower than that of the same blend activated with similar alkali dosage (7.5%) but with AM 1.25.

    The laboratory and factory trial production of AAC building blocks demonstrated the possibility of the transferability of this technology from the lab to the industry without any need for any special adjustments to the conventional production process used for PCC. The same equipment used for the production of PCC building blocks were used for the production of AAC blocks. The factory trials also can help to foster the relationship between academia and the construction industry and familiarize the industry with low CO2 footprint AAC products.

    Since commercial alkali activators were found to be responsible for the highest fraction of the unit cost and CO2 footprint of AAC mix, it is essential to source cheap and low embodied energy activators to enable the commercialization of this promising technology. In this study, the efficiency of alternative activators produced from silica rich feedstocks such as glass cullet (GC), rice husk ash (RHA) and microsilica was assessed. The efficiency of GC based activator with neat GGBS systems was around 70% compared to the commercial silicate activator. However, GC activator gave very low compressive strength when used with neat PFA systems. The efficiency of RHA and microsilica based activators with neat PFA systems was better and was around 50% and 90% respectively. Furthermore, the effectiveness of RHA and microsilica based activators was found to be similar to that of the commercial sodium silicate activator in both intermediate (35 MPa) and high (70 MPa) strength AAC. GC based activator gave similar efficiency in PFA/GGBS blends with strength lower than 50 MPa. However, 70 MPa AAC was not reached using GC activator; the highest strength obtained using GC activator was around 60 MPa.

    The cost analysis of AAC and PCC revealed that using commercial sodium silicate solution the cost of AAC was around 30% higher than that of PCC in intermediate strength concrete (35 MPa) and was similar in high strength concrete (70 MPa). By using alternative activators, the cost of AAC for intermediate strength concrete was similar to that of equivalent strength PCC. High strength AAC was found to be 20- 30% cheaper than similar strength grade PCC.
    Date of AwardJul 2016
    Original languageEnglish
    Awarding Institution
    • Queen's University Belfast
    SupervisorMarios Soutsos (Supervisor) & Wei Sha (Supervisor)

    Cite this