Department of Civil Engineering

Mechanical and Viscoelastic Properties of Cementitious

Materials using Statistical Analysis Tools and

Nanoindentation

Hyuk Lee

This thesis is presented for the Degree of

Doctor of Philosophy

of

Curtin University

August 2016

Abstract

The research presents an investigation of mechanical and nanostructure characteristics of alter-native cementitious materials such as blended and alkali-activated cements using nanoinden-tation technology and statistical analysis tools. Statistical analysis of these mechanical andnano characteristics of cementitious materials results in materials properties which are useful instructural engineering application. The indentation analysis shows that indentation propertiessuch as modulus, hardness which are related respectively to elastic and strength properties withthe projected area during the indentation loading and unloading process. With microporome-chanics, the indentation modulus and hardness represents particle properties such as stiffness,Poisson’s ratio, cohesion, friction coefficient and packing density. The indentation analysis ofcementitious material identifies the link between the multiphase compositions, and elastic, mi-crostructure and viscoelastic properties. Statistical analysis tools are successfully being appliedas a useful tool in studying the influence of parameters in cementitious materials. The resultscan be analysed using the ANOVA technique to examine the variation in the measured prop-erties of cementitious materials. Moreover, the potential impact of indentation approach willencourage consideration of small scales examination to represent the large scale testing of civilengineering structural elements. Properties of hydration products of ordinary Portland cement(OPC) are determined as indentation modulus, hardness and packing density. The indentationmodulus, hardness and packing density of LD CSH are 16.787 ± 4.804 GPa, 0.704 ± 0.144GPa and 0.556 ± 0.01, respectively. The HD CSH has indentation modulus = 30.481 ± 4.257GPa, hardness = 1.415 ± 0.222 GPa and packing density = 0.595 ± 0.001. Creep compliancerate of OPC shows that capillary pores phase is the main phase that tends to increase creepcompliance. Also, indentation stress-strain curve shows that the strength failure can occur incapillary pore phase. Thus, porosity is an important factor to be considered when design OPCmixture. Based on statistic analysis of blended cements, an increase in fly ash content andsand to cementitious material ratio decreased the compressive strength development. The op-timization of the compressive strength of blended cement mixture is found to be 20% of fly ashcontent, 1.5 of sand to cementitious material ratio, 0.35 of water to cementitious material ratioand 0.2% of superplasticiser. Similarly, alkali-activated fly ash cement (AAFA) was investigatedbased on statistical analysis tools and nanoindentation. The results show that the increase insand has the greatest contribution to the increase in density. For compressive strength, normalpaste without SF, sand and SP with l/s of 0.6 gives the highest strength and the increase inSF significantly contributes to the adverse effect on compressive strength. For the indentationdata, the analysis using deconvolution technique confirms the four phases of reaction productsof AAFA. The main phase is sodium aluminosilicate hydrate (N-A-S-H), which is over 40%of the volume fraction. The microporomechanics of AAFA paste and mortar also demonstratethe relationships between the N-A-S-H volume fraction and strength; and activation degree andstrength. The creep behaviour of AAFA study revealed that partly-activated and non-activatedphases are the main reason for creep due to “block-polymerisation” concept. It is also foundthat liquid to solid ratio is the most affecting parameter on creep, an increase of liquid to solidratio leads to more creep.

ii

Acknowledgements

Firstly, my deepest gratitude goes to my supervisor, Dr. Vanissorn Vimonsatit, for hercontinuous guidance, encouragement and support during the pursue of my doctoral degree atCurtin University.

My gratitude also goes to my Research Guide Professor Prinya Chindaprasirt, Sustainabil-ity Infrastructure Development and Research Centre, Department of Civil Engineering, Facultyof Engineering, Khon Kaen University, Thailand, for providing a platform for completing thisresearch. I have a deep sense of gratitude and feel proud to have him as guide.

I would also like to acknowledge the assistance by Dr. Kornkanok Boonserm, Departmentof Applied Chemistry, Rajamangala University of Technology in Thailand, for her help withmicrostructural properties testing.

I would like to acknowledge the assistance and help this research by final year students HyoilKim and Yuyan Yang. I also thank Mick Elliss, Ashley Hughes, Dr Arne Bredin, Luke English,and Craig Gwyther, and others for their help with laboratory work at the Department of CivilEngineering, and special thanks to my friend, PengLoy Chow, who helped me to prepare in-dentation process.

I also want to express my special thanks to my father and mother for their continuous supporttill the completion of this Thesis. Lastly, this thesis would not be completed without the loveand support from my wife and son.

iii

Contents

1 Introduction 21.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.2 Research Aim and Objective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31.3 Research Significance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41.4 Outline of Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2 Literature Reviews 62.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.2 Construction Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2.2.1 Ordinary Portland cement . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.2.2 Fly ash . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82.2.3 Lime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

2.3 Alkali-Activated Cement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102.4 Alkali-Activated Lime-Pozzolan Cement . . . . . . . . . . . . . . . . . . . . . . . 142.5 Characteristic of Binding Cementitious Products . . . . . . . . . . . . . . . . . . 162.6 Engineering Statistics Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182.7 Chapter Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

3 Nanoindentation 213.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213.2 Homogeneous Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

3.2.1 Types of Indenter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233.2.2 Indentation Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

3.3 Heterogeneous Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253.4 Microporomechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273.5 Deconvolution Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293.6 Indentation Fracture Toughness . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303.7 Finite Element Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313.8 Time-depending Nanoindentation . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

3.8.1 Indentation Contact Relaxation Modulus . . . . . . . . . . . . . . . . . . 363.8.2 Indentation Contact Creep Compliance . . . . . . . . . . . . . . . . . . . 373.8.3 Numerical Analysis of Viscoelasticity . . . . . . . . . . . . . . . . . . . . . 39

3.9 Indentation Stress-Strain Curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423.10 Chapter Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

4 Properties of Cement Materials 474.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 474.2 Indentation Properties of Ordinary Portland cement . . . . . . . . . . . . . . . . 48

4.2.1 Sample Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 484.2.2 Statistical Indentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 494.2.3 Viscoelastic Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 534.2.4 Indentation Stress-Strain Curve and Fracture Toughness . . . . . . . . . . 554.2.5 Relationship between Indentation Properties . . . . . . . . . . . . . . . . 58

4.3 Statistical Analysis of Properties of Blended Cement . . . . . . . . . . . . . . . . 614.3.1 Experimental Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 614.3.2 Density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 634.3.3 Compressive Strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

iv

4.3.4 Water Absorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 734.3.5 Regression Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 754.3.6 High Temperature Exposure . . . . . . . . . . . . . . . . . . . . . . . . . 77

4.4 Chapter Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

5 Properties of Alkali-Activated Cement 865.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 865.2 Taguchi’s Design of Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . 875.3 Mechanical Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 895.4 Indentation Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

5.4.1 Sample Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 945.4.2 Statistical Indentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

5.5 Viscoelastic Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1085.5.1 Contact Relaxation Modulus . . . . . . . . . . . . . . . . . . . . . . . . . 1085.5.2 Contact Creep Compliance . . . . . . . . . . . . . . . . . . . . . . . . . . 1115.5.3 Indentation Stress-Strain Curve and Fracture Toughness . . . . . . . . . . 1185.5.4 Chapter Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

6 Strain-hardening Behaviour of Cementitious Composite 1226.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1226.2 Design of Strain-hardening Behaviour . . . . . . . . . . . . . . . . . . . . . . . . 1246.3 Single Fibre Pull-out . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

6.3.1 Numerical Model and Validation . . . . . . . . . . . . . . . . . . . . . . . 1286.3.2 Taguchi’s Design of Experimental . . . . . . . . . . . . . . . . . . . . . . . 1326.3.3 Effect of Parameters on Maximum Pull-out Force . . . . . . . . . . . . . . 1336.3.4 Single Fibre Pull-out Test with Polyvinyl Alcohol (PVA) Fibre . . . . . . 135

6.4 High-Performance Fibre Reinforced Cementitious Composite . . . . . . . . . . . 1376.4.1 Compressive Strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

6.5 Flexural Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1456.6 Chapter Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155

7 Conclusion and Future Research 1567.1 Summary of Main Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1567.2 Recommendation for Future Research . . . . . . . . . . . . . . . . . . . . . . . . 158

Appendices 159Appendix A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159Appendix B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

References 214

v

List of Figures

2.1 Powers-Brownyard model of two types of porosity (Powers, 1958) . . . . . . . . . 82.2 Descriptive model for alkali activation of aluminosilicate (Shi et al., 2011) . . . . 122.3 Effect of Na2SO3 dosage on strength development of LPC paste (Shi and Day,

1993b) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152.4 Relationship between the Ultimate strength and curing temperature (Shi and

Day, 1993a) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152.5 Effect of Lime content on Strength development of Lime-pozzolan cement cured

at 50◦C (Shi and Day, 1993a) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162.6 General process of design of experiment (Park, 2007) . . . . . . . . . . . . . . . . 18

3.1 Schematic of contact between a rigid indenter and a flat specimen with modulusE (adapted after (Fischer-Cripps and Mustafaev, 2000)) . . . . . . . . . . . . . . 22

3.2 Typical indentation load-displacement curve . . . . . . . . . . . . . . . . . . . . . 233.3 Different types of indenter (a) Spherical (b) Berkovich (c) Conical (d) Vickers

indenter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233.4 Berkovich geometry of contact . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243.5 Element volume of material in a microscopic structure: (a) heterogeneous (b)

local heterogeneous (c) homogeneous (adapted after (Dormieux et al., 2006)) . . 263.6 Schematic representation of grid indentation technique: (a) average composite

material with large indentation depth (h� D), (b) low indentation depth withone phase (h � D),(c) grid indentation technique giving several phases usinglow indentation depth (h� D) adapted after (Constantinides et al., 2006)) . . . 27

3.7 Berkovich indentation crack parameters (adopted after (Ling, 2011)) . . . . . . . 303.8 Indentation load (P ) and depth (h) curve represent energies (adapted after

(Cheng and Cheng, 2004)) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313.9 Indentation geometry probe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323.10 Viscoelastic behaviour (a) ideal, (b) actual viscous fluid . . . . . . . . . . . . . . 333.11 Relaxation and creep stress with time . . . . . . . . . . . . . . . . . . . . . . . . 343.12 Mechanical models of time depending properties of material (a) Maxwell model

(b) Kelvin-Voigt Model (c) Combined Maxwell-Kelvin-Voigt model . . . . . . . . 363.13 A conical indenter impress on specimen . . . . . . . . . . . . . . . . . . . . . . . 393.14 Meshing configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403.15 Normalised contact relaxation modulus with analytical, numerical solution and

Maxwell model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403.16 Numerical and Rheological (Maxwell) solution of contact creep compliance . . . . 413.17 Contact creep compliance rate with numerical analysis . . . . . . . . . . . . . . . 423.18 Schematic uniaxial and indentation stress-strain curve (adopted after (Martinez

et al., 2003)) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433.19 Indentation stress-strain curve on Fused silica with Spherical tip . . . . . . . . . 443.20 Indentation stress-strain curve on Fused silica with Berkovich tip . . . . . . . . . 453.21 The deformation regimes with different types of indenter tip (adapted after (Mar-

tinez et al., 2003)) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

4.1 Deconvolution technique of indentation modulus and hardness . . . . . . . . . . . 504.2 Typical indentation load-depth (P − h) curves . . . . . . . . . . . . . . . . . . . 504.3 Packing density (η) distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . 514.4 Deconvolution result of packing density . . . . . . . . . . . . . . . . . . . . . . . 51

vi

4.5 Contour map of hydration products of OPC paste. Image size is 180µm × 180µmwith 20µm grid spacing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

4.6 Normalised relaxation modulus . . . . . . . . . . . . . . . . . . . . . . . . . . . . 544.7 Typical logarithm curve fitting in creep phase . . . . . . . . . . . . . . . . . . . . 544.8 Contact creep compliance rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . 554.9 Deconvolution result of contact creep modulus . . . . . . . . . . . . . . . . . . . 564.10 Contact creep compliance rate of hydration phases . . . . . . . . . . . . . . . . . 574.11 Indentation stress-strain curves (a) MP (b) LD CSH (c) HD CSH (d) CH . . . . 584.12 Deconvoluted indentation properties with Power fit . . . . . . . . . . . . . . . . . 604.13 XRD pattern of (a) OPC (Type I) and (b) low calcium fly ash . . . . . . . . . . 624.14 Density with curing age . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 634.15 Density with curing age . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 644.16 Contribution of experimental parameters on density . . . . . . . . . . . . . . . . 654.17 (a) compressive strength development and (b) % of compressive strength devel-

opment between 7 to 28 days . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 674.18 XRD pattern of Mix 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 684.19 SEM image of Mix 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 684.20 Effect of parameters on compressive strength . . . . . . . . . . . . . . . . . . . . 704.21 Contribution of experimental parameters on compressive strength . . . . . . . . . 714.22 Effect of parameters on compressive strength gain . . . . . . . . . . . . . . . . . 724.23 Contribution of experimental parameters on compressive strength gain . . . . . . 734.24 Water absorption results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 734.25 Effect of parameters on water absorption results . . . . . . . . . . . . . . . . . . 744.26 Contribution of experimental parameters on water absorption . . . . . . . . . . . 754.27 Residual strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 784.28 Residual strength reduction factors (adopted after (de Normalisation, 2005)) . . 794.29 XRD pattern of Mix 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 804.30 SEM image of Mix 1 after exposed to high temperatures . . . . . . . . . . . . . . 804.31 Effect of parameters after exposed to high temperatures . . . . . . . . . . . . . . 814.32 Predicted residual strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

5.1 X-ray diffraction pattern of low calcium fly ash . . . . . . . . . . . . . . . . . . . 885.2 Effect of parameters on density and compressive strength at 28 days . . . . . . . 905.3 Contribution of experimental parameters on (a) density (b) compressive strength 915.4 Typical indentation load-depth (P − h) curve on AAFA . . . . . . . . . . . . . . 955.5 Packing density relationship distribution of AAFA on Mix 1 . . . . . . . . . . . . 955.6 Effect of parameters on reaction properties (a) Stiffness (b) Cohesion (c) Friction

angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 975.7 Effect of parameters on reaction properties (a) Stiffness (b) Cohesion (c) Friction

angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 985.8 Contribution of experimental parameters on N-A-S-H phase at 28 day . . . . . . 1025.9 Contribution of experimental parameters on activated degree and porosity . . . . 1045.10 Volume fraction distribution with varying silica fume of (a) reaction products

(b) porosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1055.11 Volume fraction distribution with varying liquid to solid ratio of (a) reaction

products (b) porosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1065.12 Relationship between compressive strength and volume fraction of N-A-S-H phase1075.13 Relationship between compressive strength and degree of activation . . . . . . . . 1075.14 Normalised relaxation modulus . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1105.15 Contact creep compliance rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1125.16 Effect of parameters on creep modulus . . . . . . . . . . . . . . . . . . . . . . . . 1135.17 Contribution of experimental parameters on creep modulus . . . . . . . . . . . . 1135.18 Effect of parameters on creep modulus of partly-activated and non-activated phases1165.19 Contribution of experimental parameters on creep modulus of (a) Partly-activated

slag (b) Non-activated slag (c) Non-activated compact glass . . . . . . . . . . . . 1175.20 Indentation stress-strain curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1185.21 Effect of parameters on fracture toughness . . . . . . . . . . . . . . . . . . . . . . 1195.22 Contribution of experimental parameters on fracture toughness . . . . . . . . . . 120

vii

6.1 The concept of stress-strain hardening and strain softening under tensile stress . 1236.2 HPFRCC stress-strain behaviour . . . . . . . . . . . . . . . . . . . . . . . . . . . 1236.3 Tensile stress-strain curve for HPFRCC (adopted after (Kanda et al., 2000)) . . . 1266.4 Idealised interface law in three stages for single fibre pull-out (adopted after

(Zhan and Meschke, 2014)) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1286.5 Fracture Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1296.6 Single fibre pull-out simulation model without inclined angle . . . . . . . . . . . 1296.7 Meshing configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1296.8 Equivalent stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1306.9 Validation of FE model with analytical model . . . . . . . . . . . . . . . . . . . 1316.10 S/N ratio of single fibre pull-out . . . . . . . . . . . . . . . . . . . . . . . . . . . 1336.11 Parameter contribution on fibre pull-out test . . . . . . . . . . . . . . . . . . . . 1346.12 Schematic of single fibre pull-out test . . . . . . . . . . . . . . . . . . . . . . . . . 1356.13 XRD pattern of fly ash . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1376.14 A schematic of flexural performance . . . . . . . . . . . . . . . . . . . . . . . . . 1386.15 Compressive strength development of AAFA composites,Series A . . . . . . . . . 1396.16 Compressive strength development of OPC composites, Series P . . . . . . . . . 1406.17 Typical stress-strain curves of AAFA composites, Series A . . . . . . . . . . . . . 1416.18 Typical stress-strain curve of OPC composites, Series P . . . . . . . . . . . . . . 1426.19 The effect of fibre volume ratio on strain and compressive strength, Series A . . . 1436.20 Flexural behaviour of AAFA composites in Group A of Series A . . . . . . . . . . 1466.21 Flexural behaviour of AAFA composites in Group B of Series A . . . . . . . . . . 1466.22 Flexural behaviour of OPC composite, Series P . . . . . . . . . . . . . . . . . . 1476.23 Effect of fibre volume fraction on deflection capacity in AAFA composites, Series A1486.24 Effect of fibre volume fraction on flexural strength in AAFA composites, Series A 1496.25 Effect of fibre volume fraction on toughness in AAFA composites, Series A . . . . 1496.26 Effect of fibre volume fraction on deflection of OPC composites, Series P . . . . . 1506.27 Effect of fibre volume fraction on flexural strength of OPC composites, Series P . 1516.28 Effect of fibre volume fraction on toughness of OPC composites, Series P . . . . 1516.29 Critical volume fraction against interfacial bond strength with AAFA composites,

Series A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1546.30 Critical volume fraction against interfacial bond strength with OPC composites,

Series P . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154

viii

List of Tables

2.1 General features of types of cement in ASTM C150 (Neville, 2011) . . . . . . . . 72.2 AS 3582: Grade Specified requirement (Australia Standard, 2009) . . . . . . . . 92.3 ASTM C618 : Chemical Requirements of fly ash (ASTM Standard C618, 2012) . 92.4 Major element chemistry of fly ash in selected countries (adopted (French and

Smitham, 2007)) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92.5 Application of Geopolymer material based on silica to alumina atomic ratio (Ran-

gan, 2008) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132.6 Terminologies of design of experiment . . . . . . . . . . . . . . . . . . . . . . . . 18

3.1 Project contact area, geometry factors and intercept factor for various types ofindenters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

3.2 Different indentation area function from numerical analysis (Sakharova et al.,2009) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

3.3 Indentation geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

4.1 General features of types of cement in ASTM C150 (Neville, 2011) . . . . . . . . 474.2 Properties and characteristics of cements in AS 3972-2010 (Australia Standard,

2010) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 484.3 Deconvolution results for indentation modulus and hardness . . . . . . . . . . . . 494.4 Deconvolution results for indentation modulus and hardness . . . . . . . . . . . . 514.5 Rheological properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 534.6 Deconvolution results of indentation modulus, hardness, contact creep modulus

and packing density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 564.7 Deconvolution results of indentation modulus, hardness, contact creep modulus

and packing density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 564.8 Variation parameters and levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . 614.9 Standard L9 orthogonal array . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 614.10 Chemical composition of OPC (Type I) and low calcium fly ash (wt. %) . . . . . 624.11 Density with curing age results (kg/m3) . . . . . . . . . . . . . . . . . . . . . . . 644.12 ANOVA results on density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 644.13 The results in based on compressive strength (MPa) . . . . . . . . . . . . . . . . 664.14 ANOVA results on compressive strength . . . . . . . . . . . . . . . . . . . . . . . 694.15 ANOVA results on density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 724.16 ANOVA results on density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 744.17 Experiment and Predicted density (kg/m3) and compressive strength (MPa) at

28 days . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 764.18 Regression analysis of test data . . . . . . . . . . . . . . . . . . . . . . . . . . . . 764.19 Residual strength of blended cement mixtures (Unit in MPa) . . . . . . . . . . . 784.20 ANOVA results on high temperatures exposure . . . . . . . . . . . . . . . . . . . 814.21 Predicted S/N ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 824.22 Coefficient of empirical relationship in Equation (4.3.3) . . . . . . . . . . . . . . . 824.23 Regression analysis of residual compressive strength . . . . . . . . . . . . . . . . 83

5.1 Variation parameters and levels for AAFA mixture . . . . . . . . . . . . . . . . . 875.2 Standard L9 orthogonal array . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 875.3 Chemical composition (wt.%) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 885.4 Mix proportion with density and compressive strength test result . . . . . . . . . 895.5 ANOVA on density and compressive strength at 28 days . . . . . . . . . . . . . . 89

ix

5.6 Experiment and Predicted density and compressive strength at 28 days . . . . . 925.7 Regression analysis of test data . . . . . . . . . . . . . . . . . . . . . . . . . . . . 935.8 Properties of the reaction products . . . . . . . . . . . . . . . . . . . . . . . . . . 955.9 ANOVA results on properties of reaction products . . . . . . . . . . . . . . . . . 965.10 Deconvolution results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1005.11 Deconvolution results of Volume fraction (f) (Unit in %) . . . . . . . . . . . . . 1015.12 ANOVA results on properties of reaction products . . . . . . . . . . . . . . . . . 1015.13 ANOVA results on properties of reaction products . . . . . . . . . . . . . . . . . 1035.14 Degree of activation and porosity (Unit in %) . . . . . . . . . . . . . . . . . . . . 1045.15 Rheological properties (Unit in GPa) . . . . . . . . . . . . . . . . . . . . . . . . . 1085.16 Average logarithmic coefficients and corrected coefficient . . . . . . . . . . . . . . 1115.17 ANOVA results on creep modulus . . . . . . . . . . . . . . . . . . . . . . . . . . . 1135.18 Specific creep . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1145.19 Deconvolution results of Creep Modulus (C) (Unit in GPa) . . . . . . . . . . . . 1155.20 ANOVA results on creep modulus of partly-activated and non-activated phases . 1155.21 Fracture energy release rate and toughness . . . . . . . . . . . . . . . . . . . . . . 1195.22 ANOVA results on fracture toughness . . . . . . . . . . . . . . . . . . . . . . . . 120

6.1 Variation parameters and levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1326.2 Standard L27 orthogonal array . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1326.3 Numerical studies of single fibre pull-out with Taguchi’s DOE . . . . . . . . . . . 1336.4 ANOVA of fibre pull-out force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1346.5 Maximum interfacial bonding strength . . . . . . . . . . . . . . . . . . . . . . . . 1366.6 Parameters for analytical modelling of fibre pull-out . . . . . . . . . . . . . . . . 1366.7 Maximum pull-out force between analytical and experimental results . . . . . . . 1366.8 Chemical composition of OPC (type I) and low calcium fly ash (wt. %) . . . . . 1376.9 Composition of mix proportions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1386.10 Compressive strength development (MPa) . . . . . . . . . . . . . . . . . . . . . . 1406.11 Compressive strain and strength test results . . . . . . . . . . . . . . . . . . . . . 1426.12 Modulus of Elasticity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1446.13 Flexural behaviours . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152

A1 Indentation test of OPC paste . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159B1 Indentation test of AAFA mix 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163B2 Indentation test of AAFA mix 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168B3 Indentation test of AAFA mix 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172B4 Indentation test of AAFA mix 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176B5 Indentation test of AAFA mix 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181B6 Indentation test of AAFA mix 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185B7 Indentation test of AAFA mix 7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190B8 Indentation test of AAFA mix 8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194B9 Indentation test of AAFA mix 9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198

1

Chapter 1

Introduction

1.1 General

In the last decades, consumption of construction materials such as concrete and mortar in the

world is highly increasing due to population and economy growth, which results in high de-

mand for buildings, roads, and infrastructures. Concrete is the most widely used construction

material because of its low cost, ease of placement, flexibility to form into shapes, and avail-

ability locally. Concrete requires cementitious materials that bind the construction materials

together. However, traditional cementitious materials such as Portland cement are now facing

environmental issues due to carbon dioxide emission during the cement production. Carbon

dioxide emission is mainly a result of limestone decarbonation which is necessary in the chem-

istry during the cement production (Flatt et al., 2012). It is estimated that cement production

is releasing approximately 5 to 8 percentage of the total carbon dioxide emission per year, and

is predicted to reach a rate of 3.5 billion tonnes per year by 2025 (Van den Heede and De Belie,

2012). Therefore, development of alternative cementitious materials to reduce the environmen-

tal impact caused by high carbon dioxide emission from cement production a major subject of

on-going intensive research worldwide.

A type of cementitious materials can be developed by using pozzolan as a partial replacement of

Portland cement (Neville, 2011). This type of cements is generally called blended cements, and

the common types of pozzolan used in blended cements are industry by-products such as fly ash

and ground granulated iron blast-furnace slag. Application of blend cements has been studied

particularly on the improved mechanical properties of concrete made with blended cements.

Despite the intensive research on different applications of blended cements, slow setting time

and early strength development due to poor pozzolans reaction at early age are negative effects

of using this type of cements. These negative effects thus limit the content of pozzolans to be

used. However, blended cements are still preferred with the degree of cement replacement as

high as 35 percent, depending on the properties of available pozzolans.

Another type of alternative cements is alkali-activated cement, which is made mainly of poz-

zolans and uses alkali activators to form synthetic binder product. Research on alkali-activated

cements for application of construction materials are currently under ongoing development.

Suitable pozzolans for alkali-activated cements must possess high quantities of reactive silicate

and aluminate compounds. Fly ash, ground granulated iron blast-furnace slag, and metakaolin

are most commonly used as alumina silicate sources. The main positive effect of alkali-activated

cements is that the source materials are not carbon bearing materials, thus, do not release vast

quantities of carbon dioxide compared to traditional cementitious materials such as Portland

2

cement. Other important advantages of alkali-activated cements are their superior mechanical,

chemical, and thermal properties compared to traditional cementitious materials. In spite of

advantages of alkali-activated cements, application of alkali-activated cements is still limited

mainly due to quality control of source materials and sensitivity to alkali concentration. These

disadvantages discourage their application in construction.

Several earlier studies have been focussed on development of blended cements and alkali-

activated cements as alternative cementitious materials. Notwithstanding a wide range of

investigation of these alternative cementitious materials, there is still a lack of knowledge of

properties in micro and nanoscale and the interaction between constituents during the binding

process, which govern mechanical properties such as strength, durability, and ductility. Tradi-

tional design of cementitious materials is aimed at the final properties of the binded products

which do not sufficiently provide fundamental knowledge of the binding properties. Recently,

however, industries have realised the need to evaluate the fundamental properties of cemen-

titious materials for manipulating materials to meet the needs of modern technology. A way

to establish a knowledge in evaluating properties of alternative cementitious materials is by

exploring the possibility of improving the knowledge mathematically. This method necessi-

tates selecting appropriate statistical analysis tools for evaluating the knowledge of properties.

These statistical analysis tools can provide a meaningful measurement of properties. Advance

in these cementitious materials, therefore, can be preceded almost every major technological

leap in civil engineering. This need, in turn, is driving research to develop and investigate

fundamental properties of cementitious materials. This research is to investigate fundamen-

tal and interaction properties between constituents of binding process in blended cements and

alkali-activated cements. The outcome will enhance a development of cementitious materials

and increase their potential applications in civil engineering and construction.

1.2 Research Aim and Objective

This research is aimed at investigating properties of cementitious materials to determine me-

chanical property, chemical property, microstructure, and nanostructure of blended cements

and alkali-activated fly ash based cements based on experiment and statistical analysis of test

data.

To achieve the aim, the objectives are identified in 4 parts, which are to determine various

properties of different cementitious materials as follows:

- Part 1: Nanostructure properties of cementitious materials:

◦ To assess the reaction products of blended cements and alkali-activated fly ash based

cements, and by using nanotechnology and statistical analysis tools to evaluate inter-

action between constituents binding process and govern mechanical properties such

as modulus, hardness, packing density and microporomechanics of the materials;

◦ To apply microporomechanics on the tested materials to identify interactive proper-

ties of the materials;

◦ To investigate nano properties of viscoelasticity such as relaxation and creep in the

tested cementitious materials based on short-term test, and to use statistical analy-

sis tools to use statistical analysis methods to determine long-term time-dependent

behaviour of the materials.

- Part 2: Properties of Portland cement and blended cements

3

◦ To determine nanostructure characteristics of Portland cement, which is traditional

cementitious material using nanotechnology.

◦ To determine properties of blended cements; these properties are density, compressive

strength, high temperature resistance and water absorption, and to use statistical

analysis tools to evaluate interactive relationships between constituent ingredients.

- Part 3: Properties of alkali-activated fly ash based cementitious systems

◦ To investigate properties of alkali-activated fly ash based cement, which are density

and compressive strength, using statistical analysis tools; and to identify the effect

of constituent ingredients on the alkali-activated fly ash based cementitious systems.

◦ To determine characteristics of alkali-activated fly ash based cement using nanotech-

nology and statistical analysis tools, and to evaluate the relationship between nanos-

tructure and overall alkali-activated fly ash based cementitious systems.

- Part 4: Analysis of properties of cementitious materials for use in structural engineering

application.

◦ To investigate the effect of cementitious properties on strain-hardening behaviour of

composite; and to examine the structural performance of composites with respect to

flexure and compressive strengths;

◦ To evaluate influence of nano characteristic on structural performance of composites.

1.3 Research Significance

This research contributes to the area of research relating to the development of sustainable,

alternative cements and nano characteristics of cementitious materials. The investigation of

nanostructure characteristics of cementitious materials will be a significant part of this research.

The outcome will pose numerous pertinent questions to guide future research. The main points

of significance are as follows:

- Contribute to evaluation of the fundamental properties of cementitious materials for ma-

nipulating materials to meet the needs of modern technology.

- Enhance the development of constituent ingredients of cementitious material systems by

statistical expression.

- Contribute the development of alternative cementitious materials.

- Recommend strategies and guidelines design of cementitious materials to achieve strain-

hardening composites.

1.4 Outline of Thesis

The thesis contains the following Chapters. Chapter 2 reviews the literature on background of

cementitious systems, statistical analysis tools for engineering, and experimental methodologies

for determining chemical and microstructure properties. In Chapter 3, a review of theory of

nanotechnology and developed statistical analytical tools is provided. Nanotechnology theory

is described based on nanoindentation method which is developed from contact mechanics.

The major challenge in this Chapter is applying nanotechnology to characterise cementitious

materials. In particular, this Chapter presents a development of analytical tools for determin-

ing viscoelastic properties which are relevant in civil engineering application. Chapter 3 also

provides a theoretical framework for solving contact mechanics problems of nanoindentation.

4

Chapter 4 contains two main parts. The first part focuses on the nanostructure characteris-

tics of Portland cement systems. Nano characteristics are determined through the developed

nanotechnology applications. The second part of Chapter 4 focuses on a series of experiment

on blended cements and analysis of test data using statistical tools. The results lead to un-

cover interactive relationship between constituent ingredients of blended cements. Chapter 5

deals with mechanical and nano characteristics of alkali-activated fly ash based cements by

using statistical tools to determine the interactive relationship between mechanical and nano

characteristics between constituent ingredients of alkali-activated fly ash based cements. Nan-

otechnology and microporomechanics are applied to determine nano characteristics, which lead

to meaningful information of reaction properties of the materials. Chapter 6 presents a study

on the application of the results of mechanical and nano characteristics of cementitious materi-

als in determining strain-hardening behaviours of the materials. A design guideline to achieve

strain-hardening composites is provided. Finally, Chapter 7 summarises the outcomes of this

research and provides suggestions for further research.

5

Chapter 2

Literature Reviews

2.1 Introduction

This Chapter presents a detailed background on cementitious materials. A considerable amount

of research has been undertaken into determining the properties of alkali-activated cement in

respect to constituents, activators, reaction products, and microstructures. A critical review of

findings reported in the literature forms a fundamental knowledge for this research.

2.2 Construction Materials

2.2.1 Ordinary Portland cement

Ordinary Portland cement (OPC) is made primarily from calcareous materials such as limestone

or chalk, and procedure alumina and silica found as clay or shale. The manufacturing process of

OPC consists of grinding raw materials and mixing them intimately in certain proportions and

burning at a temperature up to about 1450◦C in a rotary kiln. During heating, partial fusion

happens, and nodules of clinker are produced. The clinker is then mixed with few percentage

of gypsum which is to control the setting rate of binder product. Grinding and mixing of raw

materials can be done either in a wet or dry condition (Neville, 2011). Alite (C3S), Belite(C2),

Aluminate (C3A) and Ferrite (C4AF) are generally the major constituents of OPC: Common

chemical terms and their notations are: CaO = C, SiO2 = S, Al2O3 = A and Fe2O3 = F. C3S

and C2S are the most important constituents of OPC, their contents are 50-70% and 15-30%

of total, respectively (Neville, 2011). These main constituents of OPC when mixed with water

will produce Calcium Silicate Hydrate (CSH), Portlandite (CH), AFm (Al2O3Fe2O3-mono)

and AFt (C3A·3CaO·SO3H12). ASTM ASTM Standard C150 (2015) has designed six type of

Portland cement as Type I to V and White cement, which differ primarily in C3A contents

and fineness. Table 2.1 shows the general characteristics of types of cement classified by ASTM

ASTM Standard C150 (2015). CSH is the main hydration product which has around 50 to 70%

of the volume of total hydration products, and governs fundamental properties of concrete such

as strength. Various experimental techniques such as mechanical, chemical and microstructural

techniques have been used to understand behaviour over the last 100 years. The description

of CSH was up to morphology and relationship between the microstructure and macroscopic

properties. According to Neville (2011), gel-like morphology CSH was observed by Le Chatelier

more than 100 years ago. Taylor (1997) found that CSH is amorphous and has the properties

of a rigid gel and its structure is similar to that of natural minerals such as Tobermorite

(Ca5Si6O16(OH)24H2O) and Jennite (Ca9Si6O18(OH)64H2O). Two types of semi-crystalline of

6

Table 2.1: General features of types of cement in ASTM C150 (Neville, 2011)

Type Classification Characteristics

I General Purpose High C3S contentII Sulphate resistance Low C3A content (< 8%)III High early strength Fine groundIV Low heat (slow reaction) Low C3S content (< 50%) and C3AV High sulphate resistance Low C3A content (< 5%)

White White colour Low MgO, No C4AF

CSH, CSH (I) and CSH (II) are reported (Taylor, 1997) to be a structurally imperfect form of

Tobermorite and Jennite, respectively. From the commotional perspective, Jennite and CSH

(II) are closer to CSH gel than Tobermorite and CSH (I) (Taylor, 1997).

The Powers-Brownyard model (Taylor, 1997) is one of the models of CSH, which provides

quantitative information on CSH properties. This model assumes that properties of the CSH

gel of cement are relatively rigid and strong - as solid of internal surface. Also, the model

described the broad structure of the material by a model based largely on an indication from

total and non-evaporable water contents and water vapour sorption isotherms. According to

Powers-Brownyard model (Powers, 1958), volumetric quantities of porosity of two phases such

as capillary porosity (VCP ), and gel porosity (VGP ) can be expressed by:

VCP = 0.20α

[1− w/c

w/c+ ρw/ρc

](2.2.1a)

VGP = 2.12α

[1− w/c

w/c+ ρw/ρc

](2.2.1b)

The following symbols are used in Equation (2.2.1) as:

w/c = Initial water to cement ratio

ρc = Density of cement, typically 3150kg/m3

ρw = Density of water, typically 1000kg/m3

α = The degree of hydration at w/c ≤ w/c∗

Based on the Powers-Brownyard model, the relationship between volumetric porosity of three

phases in Equation (2.2.1) can be graphically presented in shown in Figure 2.1. This figure shows

that an increase of water to cementitious ratio leads to a decrease of the volume fraction of two

types of porosity. The mechanical strength of OPC is required for engineering structures. The

strength of concrete or mortar mostly depends on the cohesion of OPC paste and its adhesion

to the aggregate particles. There are several forms of strength tests such as compression and

flexure. The compression test is a most common method to determine the strength for structural

design. Similarly, flexural strength generally gives a good knowledge of the strength in tension.

There are several empirical relationships between porosity and strength. The most satisfactory

empirical strength equation is as (Taylor, 1997):

σ = σo (1− Cη) (2.2.2)

where σo is hypothetical maximum compressive strength attainable, η is mean of total porosity

and C is constant.

7

Figure 2.1: Powers-Brownyard model of two types of porosity (Powers, 1958)

2.2.2 Fly ash

Coal combustion products (CCPs) are inorganic solid particulates that are produced from

burning coal to generate electricity. CCPs include fly ash, bottom ash, boiler slag, flue gas

desulfurization gypsum, and others by-products from power plants. Use of CCPs in identified

applications can have substantial environmental benefits and fly ashes represent 80 to 90 per-

centage of the total CCPs volume (Guide to the use of fly ash in concrete in Australia, 2009).

The beneficial uses of CCPs are:

- Waste stream reduction and associated reduction in the requirement for the landfill,

- Conservation of resources such as the use of gypsum, limestone or fly ash as a replacement

in cement product, and

- Reduction of greenhouse gas emission when used as a cement replacement.

Bottom ash is formed in pulverised coal furnaces and agglomerated ash particles. It is too

large to be passed in the flue gases, which exit to the atmosphere by flue and impinge on the

furnace wall, or fall through an open grate to an ash hopper at the bottom of the furnace.

Bottom ash is generally grey in black colour. It has porous surface structure with an angular

shape. Bottom ash can be used as a replacement of aggregate and is usually well graded in

size. The porous surface structure of bottom ash particles makes this material less durable than

conventional aggregates and better suited for use in the base course and shoulder mixtures or

cold mix applications, as opposed to wearing surface mixtures. This porous surface structure

is also suitable in lightweight concrete applications (U.S. Environmental Protection agency,

2007). Fly ash has been used in concrete in low cost and improves properties of concrete,

therefore, usage of fly ash is gradually increasing. The main contribution of fly ash is an

improvement in workability, mix efficiency and improved concrete placement characteristics

(Guide to the use of fly ash in concrete in Australia, 2009). Also, Shi (1996) explained that

one of the main contributions to the strength of fly ash is pozzolanic reaction between fly ash

and CH. Generally, fly ash is separated from filter bags and flue gases by mechanical collector

and electrostatic precipitators. The main difference between electrostatic precipitators and

filter bags is electrostatic classify particles size. As fly ash is a by-product material, carbon

8

emission is virtually none compared with OPC base cement. Only carbon emission due to the

transportation of fly ash is to be considered but it is still considerably low (Strategies, 2007;

Heidrich and Woodhead, 2005). There are two grades of fly ash which are classified in AS

3582 (Australia Standard, 2009) as Normal and Special grade. In some areas of Australia,

there are limited quantities available of highly reactive fly ash (Special grade). In some areas

of Australia, there are limited quantities of highly reactive fly ash (Special grade). Special

grade fly ash complies with the requirements of fine grade as shown in Table 2.2 the relative

compressive strength at 28 days of curing should exceed 105% compared to the strength of

normal grade. ASTM ASTM Standard C618 (2012) classifies fly ash, as Class C and Class F,

which is according to the total percentage of Silica dioxide (SiO2), Aluminium oxide (Al2O3)

and Iron oxide (Fe2O3) as shown in Table 2.3. Generally, Class C fly ash has calcium oxide

contents more than Class F fly ash. Sometimes, Class C fly ash and Class F fly are calling “High

calcium fly ash” and “Low-calcium fly ash”, respectively. Also, ASTM ASTM Standard C618

(2012) classifies that raw or calcined natural pozzolan such as opaline cherts, shales, tuff and

volcanic ashes or pumicites as Class N. French and Smitham (2007) presented major elements

of fly ash from different countries with significant variation in major elements chemistry. Table

2.4 shows major chemical compositions of fly ash derived from different counties SiO2, Al2O3,

and basic oxides are calcium oxide (CaO), Magnesium oxide (MgO), Iron (III) oxide (Fe2O3),

Sodium oxide (Na2O) and Potassium oxide (K2O) approximately. The main effect of fly ash in

cement system is that on water demand and workability. Neville (2011) noted that between 5

and 15% of fly ash is the reduction in water demand of concrete for a consistent workability.

The main reason for a decrease in water demand of concrete is ‘ball-bearing’ effect which is

ascribed to fly ash due to its spherical shape. At the same time, another mechanism is involved

that fine fly ash particles become adsorbed on the surface of the cement particle by electrical

Table 2.2: AS 3582: Grade Specified requirement (Australia Standard, 2009)

Grade Fine Medium Coarse

Fineness∗ (% minimum) 75 65 55Loss on ignition (% maximum) 4 5 5Moisture content (% maximum) 1 1 1

SO3 contents (% maximum) 3 3 3

∗ by mass passing 45µm sieve

Table 2.3: ASTM C618 : Chemical Requirements of fly ash (ASTM Standard C618, 2012)

Class F Class C

SiO2 + Al2O3 + Fe3O3 (% minimum) 70.0 50.0SiO3 (% maximum) 5.0 5.0

Moisture content (% maximum) 3.0 3.0Loss on ignition (% maximum) 6.0 6.0

Table 2.4: Major element chemistry of fly ash in selected countries (adopted (French and Smitham,2007))

Components Australia US UK Japan

SiO2 55 to 60% 55 to 63% 43 to 48% 55 to 62%Al2O3 24 to 30% 18 to 28% 25 to 30% 23 to 28%

Basic Oxides 15 to 28% 18 to 33% 25 to 30% 14 to 18%

9

charges. Thus, the water demand for a given workability is reduced because fine fly ash particles

are covering the surface of the cement particles. However, the excess amount of fly ash would be

no benefit on water demand. Approximately up to 20% of fly ash content is reported to reduce

water demand in concrete. According to Ramezanianpour (2014), Class C fly ash has self-

hardening properties such as C2S, C3A, CaSO4, MgO and free CaO. The hydration behaviour

of C2S and C3A in Class C fly ash is the same as that in OPC, but the hydration rate of CSH

form is comparatively slow. Somehow, thus, Class C fly ash when mixed with water produces

hydration products in OPC as AFm, AFt and CSH. Class F fly ash, however, has little or no

self-hardening properties. Class F fly ash only hydrates when alkalis and CH are added. The

hydration products such as CSH, C2ASH8 and C4AH are formed and they are produced in the

later stage of the hydration process. Generally, fly ash in concrete is to be incorporated because

packing action of the fly ash particles at the interface of coarse aggregate particles would result

in reduced permeability. However, the chemical reactions of fly ash have an effect in improving

the microstructure of hydrated cement pastes which can lead to the strength development of

concrete. Papadakis (1999) studied the effect of Class F fly ash on Portland cement system.

They found higher strength in concrete when using Class F fly ash to replace aggregates or

cement. When Class F fly ash reacts with CH; it gives higher water content and lower total

porosity (Papadakis, 1999). Poon et al. (2000) reported that high strength of concrete with

45% Class F fly ash has a lower heat of hydration and chloride diffusivity than normal plain

cement concrete. Class F fly ash in concrete with lower water to binder ratio has better strength

contribution. Another research from Papadakis (2000) reported that Class C fly ash content

in concrete directly led to the strength enhancement after the mixing. The reaction of Class C

fly ash in hydrating cement makes it lower porosity due to the high content of reactive calcium

bearing phase in fly ash.

2.2.3 Lime

There are two types of Lime, Quicklime and Hydrated Lime (slake lime) which are commonly

used in practice. According to Shi (1992), quicklime with cement has higher strength than

hydrated lime with cement when having the same content of quicklime and undergoing the same

curing condition. Quicklime with water react quickly and form reaction production of calcium

hydroxide which is very fresh and high reactivity than commercial hydrated lime. Quicklime

is commonly known in the form of Calcium Oxide (CaO) or burnt lime. It is produced from

limestone (CaCO3) by calcination at high temperature to decompose the limestone to quicklime.

It is an alkaline product with more than pH 12, and normally in powder or granule with a density

greater than 1000 kg/m3. The process of progressive change of limestone to lime is shown as:

CaCO3heat−−−→ CaO + CO2 (2.2.3)

Hydrated lime is the chemical content of Calcium hydroxide (Ca(OH)2). It is produced by

quicklime (CaO) mixing with water. It is an alkaline product with pH¿12 and fine power with

450 to 780 kg/m3 of density (What is Lime?, 2005).

CaO +H2O → Ca(OH)2 (2.2.4)

2.3 Alkali-Activated Cement

Purdon Pacheco-Torgal et al. (2008) introduced a major development of alkali-activated cement

in 1940s. A highly concentrated alkali hydroxide solution or silicate solution reacts with a solid

10

aluminosilicate produces a synthetic alkali aluminosilicate materials called “geopolymer” (Davi-

dovits, 1994b; Joseph, 2011). These materials are classified as polymer because their structures

are large molecules formed by a number of groups of smaller molecules. Theses polymer are

formed by the reaction of an alkali solution and a source material which is rich in aluminosilicate

and includes organic minerals such as kaolinite and inorganic materials such as fly ash (Kumar

et al., 2005). Depending on the source material selection and recession conditions, geopolymer

exhibits a variety of properties and characteristics such as strength, shrinkage, setting time,

acid resistance, fire resistance and thermal conductivity. Thus, it is not necessarily intrinsic

to formulation but correct mix and processing design to optimisation properties and reduce

the cost for application (Duxson, Fernandez-Jimenez, Provis, Lukey, Palomo and Van Deven-

ter, 2007). Generally, ‘geopolymer’ is also commonly referred to as ‘alkali-activated cement’

(Shi et al., 2011). In this research, these synthetic alkali aluminosilicate materials are called

‘alkali-activated cement’ (AAC). According to Shi et al. (2011), AAC mainly consists of two

components, which are cementitious components and alkali activators. Cementitious compo-

nents can be taken from a variety of industrial by-products composed of silica and aluminium

such as granulated blast furnace slag, volcanic glass, coal fly ash zeolite, metakaolin, silica fume

and nonferrous slag (Shi et al., 2011). The composition of the cementitious components and

alkali-activated cement can be classified into five categories:

- Alkali-activated slag-base cement

- Alkali-activated Pozzolan cement

- Alkali-activated lime-Pozzolan/slag cement

- Alkali-activated calcium aluminate blended cement

- Alkali-activated Portland blended cement

Pacheco-Torgal et al. (2008) reported that blast furnace slag with sodium hydroxide as an

activator and process developed in two stages. The liberation of silica aluminium and cal-

cium hydroxide took place, forming of silica and alumina hydrate occurred as the regeneration

of the alkali solution. A number of researchers (Pacheco-Torgal et al., 2008; Roy, 1999; Shi

et al., 2011) reported that the difference between the compositions of traditional Portland

cement and fundamental rock-forming minerals of the earth crust. The calcium silicate hy-

drate (CSH) and calcium hydroxide (Ca(OH)2), tthe major reaction products, are containing

alkalis. The formation of major reaction products is raised by the probability of enhanced

durability. Based on investigations, it was developed as “soil-cement” which was a new type

of binder (Glukhovsky et al., 1980). TThe soil-cement was found from ground aluminosili-

cate mixed with rich alkalis industrial wastes. The general mechanism of alkaline activation

reaction of material into three stages, destruction-coagulation, coagulation-condensation, and

condensation-crystallisation, introduced by Glukhovsky (Shi et al., 2011). Recently, several

authors (Criado et al., 2005; Duxson, Fernandez-Jimenez, Provis, Lukey, Palomo and Van De-

venter, 2007; Shi and Fernandez-Jimenez, 2006) extended Glukhovky theories to explain the

polymerisation process. Figure 2.2 shows descriptive polymerisation mechanism of AAC. The

most important key process of the reaction mechanism of AAC is the transforming of alumi-

nosilicate source into a synthetic alkali aluminosilicate by fine grinding and heat treatment.

Rangan (2008) recommended heat treatment of AAC mixtures as being substantially assistant

to the chemical reaction of aluminosilicate source to form a synthetic alkali aluminosilicate.

11

Figure 2.2: Descriptive model for alkali activation of aluminosilicate (Shi et al., 2011)

Generally, the high temperature accelerates the polymerisation process compared to the am-

bient condition. Hardijito and Rangan (2005) indicated that alkali activated fly ash cement

(AAFA) did not harden immediately at room temperature. They reported that when the room

temperature was less than 30◦C, the hardening did not occur at least for 24 hours. AAFA pro-

duced in ambient temperature achieved lower strength in the early days as compared to heat

treated specimen. Kong and Sanjayan (2008) reported that longer heat curing regime did not

significantly affect the strength behaviour of AAFA. They found that most of polymerisation

were complete within the first 24 hours of heat curing.

Alkali activator solution is an important constituent in AAC. Alkaline activators produce a

dissolution of these source of materials. The alkaline condition activates the reaction between

raw materials in polymerisation process to form the polymer network. The aluminosilicate

materials can increase reactivity as a lower bonding energy of Al-O than Si-O. The alkali

activator breaks down the covalent bond Si-O-Si and Al-O-Al in aluminosilicate materials with

pH of alkali activator as follows (Shi and Day, 2000):

≡ Si−O − Si ≡ +3OH− → (SiO (OH)3)−

(2.3.1a)

≡ Si−O −Al ≡ +7OH− → (SiO (OH)3)−

+ (Al (OH)4)−

(2.3.1b)

The dissolved mono-silicate and aluminate form amorphous to semi-crystalline silico-aluminate

structure and its empirical formula is (Caijun and Della, 2006):

Mn (− (Si−O2)z −Al −O)n· wH2O (2.3.2)

where M is a cation such as sodium, n is the degree of ploy-condensation, and z is a poly

chain and ring polymers with oxygen; range from amorphous to semi-crystalline. However, the

reaction mechanism of AAC is still not fully understood.

The ratio of pozzolan to alkaline activator is critical in strength development and thermal

resistance of AAC (Kong and Sanjayan, 2008). The most common alkaline activators in AAC

are potassium hydroxide (KOH) with potassium silicate and sodium hydroxide (NaOH) with

sodium silicate (Na2SO4). It was proven that alkaline solution containing soluble silicate could

increase reactivity compared to alkali solution containing only hydroxide (Palomo et al., 1999).

Hardijito and Rangan (2005) reported the alkaline solution containing soluble silicate, either

12

sodium or potassium silicate, had a high rate of reactions than that of alkaline hydroxides. In

addition, Xu and Van Deventer (2000) studied alkaline solution by adding Na2SO4 solution to

NaOH solution and found that it increased the reaction between the source materials. They

concluded that typically the NaOH solution has the higher extent of dissolution of minerals

than the KOH solution.

Rangan (2008) studied that failure behaviour and elastic properties of AAFA (Class F) concrete

are similar to OPC concrete. Also, he found that AAFA concrete has excellent compressive

strength, resistance to sulfate attack, good acid resistance, low drying shrinkage, and creep

behaviours. Hardjito et al. (2004) studied AAFA (Class F) concrete with varying concentration

of NaOH solution in morality, Na2SO4 to NaOH ratio, curing temperatures, and curing times.

They found that higher concentration in term of the morality of NaOH solution and Na2SO4

to NaOH ratio resulted in higher compressive strength. An increase of curing temperatures in

the range of 30 to 90◦C leads an increase of compressive strength. In term of curing times, the

increase in strength beyond 48 hours is not significant.

AAC has superior properties compared to ordinary Portland cement (OPC) (Demirel and

Kelestemur, 2010; Duxson, Provis, Lukey and Van Deventer, 2007; Kong and Sanjayan, 2010;

Xu and Van Deventer, 2000). Several properties which are superior to those in OPC are such

as less drying shrinkage, higher early age strength, and resistance to high chemical and tem-

peratures. Also, one of main source of OPC is CaO, which can be presented in the form of

carbonates (CaCO3). However, AAC does not contain CaCO3 because one of the main sources

is the aluminium. Therefore, alkali-activated pozzolan cement does not release vast quantities

of CO2 which can be described as due to the release of CaO from CaCO3. The application of

AAC has a wide range in the field of industries because of its superior performance and dura-

bility properties. According to Davidovits (1994a), the classification of the type of application

of AAC is described base on Si : Al ratio. A high ratio of Si : Al more than 15 provided a poly-

meric character to the polymeric material and low ratio of Si : Al of 1,2, or 3 (Rangan, 2008).

Alkali activator solution is an important constituent in AAC. The alkaline condition activates

the reaction between raw materials in polymerisation process to form the polymer network.

The ratio of pozzolan to alkaline activator is critical in strength development and thermal re-

sistance of AAC (Kong and Sanjayan, 2008). The most common alkaline activators in AAC

are potassium hydroxide (KOH) with potassium silicate and sodium hydroxide (NaOH) with

Table 2.5: Application of Geopolymer material based on silica to alumina atomic ratio (Rangan,2008)

Si:Al ratio Application

1Brick, CeramicsFire protection

2Low CO2 cement and concreteRadioactive and toxic waste encapsulation

3

Fire protection fibre glass compositeFoundry equipmentHeat resistant composites 200◦C to 1000◦CTooling for aeronautics titanium process

>3Sealant for industry 200◦C to 600◦CTooling for aeronautics SPF aluminium

20 to 35 Fire-resistant and heat resistant fibre composite

13

sodium silicate (Na2SO4). It was proven that alkaline solution containing soluble silicate could

increase reactivity compared to alkali solution containing only hydroxide (Palomo et al., 1999).

Hardijito and Rangan (2005) reported the alkaline solution containing soluble silicate, either

sodium or potassium silicate, had a high rate of reactions than that of alkaline hydroxides. In

addition, Xu and Van Deventer (2000) studied alkaline solution by adding Na2SO4 solution to

NaOH solution and found that it increased the reaction between the source materials. They

concluded that typically the NaOH solution has the higher extent of dissolution of minerals

than the KOH solution.

2.4 Alkali-Activated Lime-Pozzolan Cement

Lime Pozzolan cement (LPC) is one of the earliest building materials, widely used in the ma-

sonry construction during Roman times. Slow strength development with ambient temperature

curing condition is one of main disadvantage of lime-pozzolan cement. In the 19th century,

its use was significantly reduced because of the invention of Portland cement which was faster

setting and had high early strength. In the past 50 years, some developed countries have been

used for manufacturing construction products because of low cost and excellent durability than

Portland cement. Also, pozzolans are used for sustainability or mixing with Portland cement

for advantageous properties which are a reduction in cost and heat evolution, alkali-aggregate

expansion control, increased chemical resistance, reduced concrete drying shrinkage, improve-

ment of the properties of fresh concrete and decreased permeability (Caijun and Della, 2006).

Alkali activator is potentially possible to accelerate the early strength development. Shi (1992)

suggested that Na2SO4 is a most effective activator of LPC and 4% of Na2SO4 is the optimum

content as shown in Figure 2.3. Also, he determined that alkali activator in LPC paste can

increase early strength development (Shi and Day, 1993b) as shown in Figure 2.4. The following

products are recognised as the main hydration products of lime-pozzolan cement (Caijun and

Della, 2006).

Ca(OH)2 + Pozzolan + Water→ Hydration Products (2.4.1)

where hydration products are:

- Calcium silicate hydrate (C-S-H)

- Ettringite (C3A·3CaSO4 ·32H2O)

- Hydrated tetracalcium aluminate (C4AHx)

- AFm (C3A·CaSO4 ·12H2O)

- Hydrated gehlenite (C2ASH8)

- Hydrated calcium carbonaluminate (C3A·CaCO3 ·12H2O)

In lime-pozzolan cement, Ca(OH)2 is mixed with pozzolan in water without activator, the

solution reaches a high pH value which is approximately 12.5 at 20◦C.

Ca(OH)2 → Ca2+ + 2OH− (2.4.2)

In a high pH solution under OH−, pozzolan is dissolved and depolymerised into the solution

such as Ca2+, K+, and Na+. As Ca2+ ions which are in contact with those depolymerised or

dissolved pozzolan mono silicate and aluminate species, calcium silicate hydrate (C-S-H) and

14

Figure 2.3: Effect of Na2SO3 dosage on strength development of LPC paste (Shi and Day, 1993b)

Figure 2.4: Relationship between the Ultimate strength and curing temperature (Shi and Day,1993a)

calcium aluminate hydrate (C4AH13) form.

Y (SiO (OH)3)−

+XCa2+ + (Z −X − Y )H2O + (2X − Y )OH− → CxSyHz (2.4.3)

2 (Al (OH)4)−

+ 4Ca2+ + 6H2O + 6OH− → C4AH13 (2.4.4)

Lime-pozzolan cement has not shown significant strength after 3 days because dissolved monosil-

icate species diffuse more quickly than dissolved aluminate spices.

2 (Al (OH)4)−

+ 3SO2−4 + 6Ca2+ + 4OH− + 26H2O → C3A · 3CaSO4 · 32H2O (2.4.5)

6Ca2+ + 2 (Al (OH)4)−

+ 3CaO ·Al2O3 · 3CaSO4 · 32H2O + 10OH− (2.4.6)

→ 3 (3CaO ·Al2O3 · CaSO4 · 12H2O) + 5H2O

The reaction between Ca(OH)2 and Na2SO4 as Na2SO4 is added as expressed by (Litvan, 1986).

Na2SO4 + Ca (OH)2 + 2H2O → CaSO4 · 2H2O ↓ +2NaOH (2.4.7)

The components of quick lime contents more than 90% of Calcium Oxide (CaO). Thus, quick

lime with water would renovate into Ca(OH)2 and hydrated lime contents greater than 90%

of Ca(OH)2. Therefore, Na2SO4 act with Ca(OH)2 than transforms into Sodium hydroxide

(NaOH). Also, hydration products with Na2SO4 paste of AFt and AFm are as shown (Caijun

15

and Della, 2006):

2 (Al (OH)4)−

+ SO2−4 + 4Ca2+ + 6H2O + 4OH− → C3A · CaSO4 · 12H2O (2.4.8)

6Ca (OH)2 + 2Al2O3 + 3CaO ·Al2O3 · 3CaSO4 · 32H2O → 3 (3CaO ·Al2O3 · 12H2O) + 2H2O

(2.4.9)

Helmuth (1983) studied complete theoretical reaction between Limes to pozzolan ratio according

to the hydration products of lime-pozzolan cement. They assumed that

- Fly ash contains SiO2 = 50%, Al2O3 = 30%, Other compounds = 20%

- Average molar ratio of CaO/SiO3 of C-S-H is 1

- CaO/SiO3 weight ratio in Gehlenite hydrate (Ca2Al(AlSiO7)) is 0.55

- Additional lime required to produce AFt and AFm

It was concluded at least 45% hydrated lime should contain in lime-pozzolan cement as shown in

Figure 2.5. However, a rise in lime content will increase the water requirement in lime-pozzolan

cement and increasing the water will reduce the strength of harden paste. Over saturated

Ca(OH)2 in lime-pozzolan cement also will produce calcium carbonation (CaCO3) in paste

(Shi, 1996). The theoretical solubility of Ca(OH)2 in water is 0.178g/100mL at 20◦C. It is

impossible practically soluble Ca(OH)2 in water for cement mixture. Therefore, according to

the results of Shi and Day (1993b), 20% hydrated lime mixed with 80% of pozzolan was an

optimum mixture for lime-pozzolan cement.

Figure 2.5: Effect of Lime content on Strength development of Lime-pozzolan cement cured at 50◦C(Shi and Day, 1993a)

2.5 Characteristic of Binding Cementitious Products

There are several of techniques for characterising the composition and structure of the re-

action or the hydration products of binding cementitious materials, such as X-ray diffraction

(XRD), Scanning Electron Microscope (SEM), thermogravimetric analysis (TGA), nitrogen ad-

sorption/desorption (NAD), mercury intrusion porosimetry (MIP), and transmission electron

microscopy (DOH). Especially, XRD and SEM are widely used to identify the characteristics

of composition and structure of reaction or hydration products of binding cementitious mate-

rial because these are a rapid analytical technique. X-ray diffraction (XRD) provides evidence

of the periodic atomic structure of crystals by reflecting X-ray beams at certain angles of in-

cidence. The cement phase composition influences the performance characteristics of them

16

(Stutzman, 1996). XRD is a direct method for qualitative and quantitative characterisation of

fine-grained materials like cement and raw materials. Products in each phase have a unique

diffraction pattern that is independent of others and the intensity of each pattern is relative to

that phase concentration in the mixture. For Quantitative X-ray diffraction analysis (QXDA),

the most common method of analysing chemical composition in cement involves the addition

of a known amount of an internal standard having a controlled particle size. This internal

standard helps correct any matrix effects between specimens and time-dependent changes in

X-ray intensity (Stutzman, 1996). Scanning Electron Microscope (SEM) releases a beam of

electrons through the specimens and measures any signals resulting from the electron beam

interaction with the specimens. The images of topography could be used to study particle size,

shape, surface roughness and hydration products (Stutzman, 2001). In addition, back-scattered

electron (BSE) imaging has illustrated its potential in a study of cementitious materials. BSE

are electrons from the incident beam which are scattered over large angles that reflect from the

specimens (Scrivener, 2004).

According to (Zhang and Scherer, 2011), these techniques must be subjected to treatment

that arrests the reaction or the hydration products of binding cementitious materials. There

are two main methods for arresting hydration and removal of water from the samples, direct

drying and solvent exchange methods. In the direct drying methods, such as oven drying,

microwave drying, vacuum drying, freezing drying, D-drying and P-drying, water is removed

by convening into vapour. Microwave drying can be used for investigating total water content

and for accelerated curing but a rapid thermal extension of the pore liquid could cause damage.

D-drying, or dry ice method, is commonly recognised as the best standard drying technique. It

is assumed to remove all of physically unbound water in the cement paste pore and is considered

the best preserving microstructure drying method. The vacuum drying is generally done in the

chamber with less than 0.1 Pa pressure. It gives a similar result to D-drying method but

vacuum drying method cannot effectively arrest early hydration because the drying process is

slow. Freezing drying method causes less damage compared to oven drying method but it still

damages the microstructure and pore structure in the cement. Oven drying is probably the

simplest technique, thus widely used as the drying technique with temperature generally set

between 60 to 105 ◦C, and an atmospheric pressure of 101kPa, which will effectively remove

evaporable water. Oven drying at 105◦C for 24 hours will remove water from hydration process.

However, the oven dry method could damage the sample caused by the remaining structural

water and pore structure.

Solvent exchange method is to exchange water with isopropyl alcohol in binding cementitious

materials and most researchers (Zhang and Scherer, 2011; Scrivener and Nonat, 2011; Chen

et al., 2014) recommended this method for arresting reaction or hydration products. For X-

ray diffraction test, the presence of organic material may not matter, but the possibility of

decomposing or altering hydrated aluminate phases must be considered. The solvent to sample

ratio was suggested to be taken as 100:1 and oven or vacuum dry for 24 hours, the minimum

soaking time (t) also could be calculated by Equation , in which C0 is initial concentration, C1 is

constant surface concentration, c (r, t) is the average concentration in sample, a is characteristic

dimension, and D is diffusivity (m2/s) (Zhang and Scherer, 2011). The soaking time can be

solved using Equation , approximately 24 hours required for a soaking time when the time to

replace the water to 99.99% with a solvent in cementitious materials, 1mm of the granule of

sample size, and 1.0 × 10−11 m2/s of diffusivity of solvent assumed. Thus, binding cementitious

materials can arrest their reaction or hydration products using solvent exchange method with

17

24 hours soaking time.

c (r, t)− C1

C0 − C1=∞∑n=1

6

n2π2exp−n2π2Dt

a2≈ 6

π2expπ2Dt

a2(2.5.1)

2.6 Engineering Statistics Analysis

Civil Engineering experiments are performed to determine a characteristic of systems such as

materials and structures. These characteristics, thus, require to be evaluated appropriately

according to the objectives of the experiments which generally are to (Park, 2007):

- Evaluate characteristic values without statistical analysis.

- Identify experimental factors that are significant to the response and determine how large

the impact is.

- Determine statistically the affecting factors with small influence.

- Determine significant factors and then evaluate optimisation condition.

Design of Experiments (DOE) is a method to satisfy experimental objectives and various factors

that can be determined by statistical analysis of experimental results. Generally, the process

of DOE is as shown in Figure 2.6 and some terminologies of DOE are presented in Table 2.6.

DOE can be used by meta-models which are the model of a model. The approximated function,

which is a meta-model, can represent specific function values as vector b as follows:

DOE can be regarded by the approximated function y with vector b as:

y = g (b) (2.6.1)

Figure 2.6: General process of design of experiment (Park, 2007)

Table 2.6: Terminologies of design of experiment

Terminologies Description

Factor Influence the characteristic function or objective functionLevel Factor or design variable

Characteristic Response of the system

18

While the error in the approximation process is ε, the relationship between y and y is:

y = y + ε (2.6.2)

In DOE, a method is employed to define y and variable vector b is obtained (Park, 2007).

Therefore, DOE can be utilised to develop efficient experiments and for analysis of experimental

results.

Genichi Taguchi developed DOE method during the 1950s (Ranjit, 1990), and is now commonly

known as Taguchi’s method. Taguchi’s DOE approach to parameters design provides the design

engineer with a systematic and efficient method for investigating optimum design parameters

according to the required performance and cost. Taguchi’s DOE method identifies the ‘signal to

noise (S/N)’ factors. S/N factors are what causes a measurable product or process characteristic

to deviate from its target value (Ozbay et al., 2009). Target value might be:

Smaller is better: select when the goal is to minimise the response. The S/N ratio can be

determined as:

S/N = −10× log10

(1

n

n∑i=1

Y 2i

)(2.6.3)

Larger is better: select when the goal is to the maximum the response. The S/N ratio can be

determined as:

S/N = −10× log10

(1

n

n∑i=1

1

Y 2i

)(2.6.4)

Nominal is better: select when the goal is to target the response and it is required to base the

S/N ratio on standard deviations only. The S/N ratio can be determined as:

S/N = −10× log10

(1

n

n∑i=1

(Yi − Y0)2

)(2.6.5)

Yi is the measured value of each response in Equations (2.6.3 to 2.6.5) . When variability

occurs, it is because the physically active in the design and environment that promotes change

(Ozbay et al., 2009). S/N factors can be classified into:

- External noise factors, sources of variability that come from outside the samples.

- Unit to unit noise, due to the fact that no two manufactured components or products are

ever exactly alike.

- Internal noise, due to deterioration, ageing and wear incurred in storage and use.

Taguchi’s DOE requires creating a set of tables of numbers. These tables, known as ‘orthogonal

array’, are used to lay out particular factors and levels of constituents. Results from the

experiements are generally analysed using Analysis of Variance (ANOVA), which is to determine

the extent to which the effect of an independent variable is on the targeted tests. It is a general

procedure for isolating the source of variability in a set of measurements.

2.7 Chapter Summary

Review of cementitious materials and their binding cementitious products presented in this

Chapter outlined the chemical compositions and their reaction products. Calcium Silicate Hy-

drate (CSH) which governs fundamental properties such as strength, is important to binding sys-

tem of OPC based materials and needs attention to be fully understood. Due to beneficial uses

19

of fly in construction materials, its effects on the mechanical and chemical and microstructural

properties of hydrated cement systems are necessary to be explored. In particular, this research

will determine the possibility of establishing interactive relationships between constituents of

mixtures. Similarly, alkali-activated cement (AAC) is one of alternative cementitious mate-

rials which has superior properties such as mechanical, chemical, and thermal compared to

OPC-based binding cementitious products. However, the relationship between properties and

reaction products are also not well understood. This forms the background of this research as

further works are required to understand the reaction products and the interactive relationships

between selected constituents of mixture of AAC.

The literature review shows that the characteristics of materials and structures in civil engineer-

ing and construction require to be evaluated appropriately according to the objectives of the

experiments. Taguchi’s design experimental approach is one of DOE methods and introduced

that parameters design provides the design engineer with a systematic and efficient method

for investigating optimum design parameters for performance and cost. Analysis of Variance

can extend the knowledge on the effect of major independent variables of the data obtained

from Taguchi’s DOE. Further, use of statistical and systematical methods can help to draw

important conclusions on the properties of cementitious materials.

20

Chapter 3

Nanoindentation

3.1 Introduction

Nanoindentation is a technique used to obtain meaningful mechanical properties of materials.

This Chapter reviews a classical indentation analysis method of contact mechanism in homoge-

neous materials. This review defines the critical importance of the experimental investigation of

the fundamental properties of cementitious materials. Nanoindentation testing can determine

local mechanical properties of cementitious binder paste, mortar and concrete at a microstruc-

ture level. Indentation testing is done essentially by touching the material of interest, whose

mechanical properties such as elastic modulus, hardness, strain-hardening, fracture toughness

are unknown, using another material whose properties are known (Fischer-Cripps, 2011). The

main focus of this Chapter is to present details of analysis for determining indentation modulus

M , hardness H and related mechanical properties of indentation materials.

3.2 Homogeneous Materials

According to the literature (Cheng and Cheng, 2004; Constantinides and Ulm, 2007; Fischer-

Cripps, 1999; Oliver and Pharr, 2004), measurement of the elastic modulus and hardness of a

material can be obtained from indentation load-displacement data during one cycle of loading

and unloading. Figure 3 1 shows a contact between a rigid indenter and a flat specimen. This

figure presents behaviour of the modulus of elasticity of a material at the contact radius a, the

indenter load P , the indenter radius R, and indentation modulus M . Hertz contact equation

(Fischer-Cripps and Mustafaev, 2000) determines the indentation modulus M as:

a3 =4

3

PR

M(3.2.1)

where the quantity of M is expressed as the combined modulus of the indenter and the

specimen as:

1

M=

1− v2

E+

1− v′2

E′(3.2.2)

In the Equation (3.2.2), E is the elastic modulus of the specimen and v is the Poisson’s ratio

of the indented material, while the superscript prime denotes the corresponding properties of

the indenter, i.e., E′ is the elastic modulus and v′ is the Poisson’s ratio of the indenter. The

maximum tensile stress in the specimen occurs at the edge of the contact circle at the surface

as:

21

Figure 3.1: Schematic of contact between a rigid indenter and a flat specimen with modulus E(adapted after (Fischer-Cripps and Mustafaev, 2000))

σmax = (1− 2v)

(P

2πa2

)(3.2.3)

This tensile stress is acting in a radial direction on the outside of indentation surface. Com-

bination of Equations (3.2.2) and (3.2.3) provide the maximum tensile stress on the outside of

the contact circle, which can be expressed in term of the indenter radius R, as (Fischer-Cripps

and Mustafaev, 2000):

σmax =

(1− 2v

2π

)(4E

3

)2/3

P 1/3R−2/3 (3.2.4)

The mean contact pressure pm has the additional virtue of having actual physical significance

as:

pm =P

πa2(3.2.5)

Substituting Equation (3.2.1) into (3.2.3), the mean contact pressure pm is written as (Fischer-

Cripps and Mustafaev, 2000):

pm =

(4M

3π

)( aR

)(3.2.6)

The mean contact pressure pm is referred to as the “indentation stress”, the quantity of a/R

is the “indentation strain”. It is similar to the elastic condition of linear stress-strain response

that is commonly determined from conventional uniaxial tension and compressive test. The

mean contact pressure pm, which is defined as the “indentation hardness (H)” can be obtained

under fully developed plastic condition, which will be further described in Section 3.9.

Figure 3.2 presents a typical data set obtained with Berkovich indenter, where the parameter

P designates the load, and h the displacement relative to the initial undeformed surface. It

is important to measure the maximum load, Pmax, the maximum displacement, hmax, and

the elastic unload stiffness (or contact stiffness). Noting that S = dP/dh, which is the slope

of the upper portion of the unloading curve during the initial stage of unloading. The final

depth hr is the permanent depth of penetration after the indenter is fully unloaded which is

another important quantity. The most widely used method for determining the contact area is

developed by Oliver and Pharr (1992, 2004). The procedure for determining the contact area

begins with fitting the load-displacement curve acquired during unloading to the power-law

relation as:

22

Punloading = β · (h− hr)m (3.2.7)

where h is the penetration depth, β and m are empirically fitting parameters, and hr is the

final displacement as shown in Figure 3.2. The contact stiffness S is then established by (Oliver

and Pharr, 1992, 2004)

S = β ·m (h− hc)m−1 |h=hmax (3.2.8)

However, Equation (3.2.7) does not always provide adequate expression of the entire unloading

curves, thus, only the upper portion of the unloading around 25% to 50% of data is generally

sufficient.

3.2.1 Types of Indenter

The main achievement of nanoindentation technology is to extract elastic modulus and hardness

of materials from the load-displacement curve. Normally, the indentation hardness is related

to the measurement of the size of a residual plastic impression in the material. It provides

a measure of the contact area for a given indenter load. Thus, the penetration depth of the

specimen during indenting load is important to extract material properties of the specimen. The

known geometry of the indenter is therefore required to determine the contact area. Generally,

nanoindentation test can be conducted using several types of indenter as shown in Figure 3.3.

Spherical Indenter

The mean contact pressure pm is determined in the fully plastic zone in term of indentation

hardness H and the contact area of the spherical indenter Ac as:

Figure 3.2: Typical indentation load-displacement curve

Figure 3.3: Different types of indenter (a) Spherical (b) Berkovich (c) Conical (d) Vickers indenter

23

pm = H =P

Ac=

2P

πR2(3.2.9)

where R is the radius of the contact circle at fully loading. The contact depth as shown in

Figure 3.1 and the contact area can be calculated using the known geometry of the indenter,

thus the contact area of a spherical indenter is (Fischer-Cripps and Mustafaev, 2000):

Ac = π(2Rhc − h2c

)2(3.2.10)

Berkovich Indenter

The Berkovich indenter is a more precise control over the indentation process. The mean contact

area is normally obtained from a measure of the contact penetration depth hc as shown in Figure

3.4. The projected area of Berkovich contact is given as (Fischer-Cripps and Mustafaev, 2000):

Ac = 3√

3h2c tan2 θ (3.2.11)

where θ is 65.27◦.

Conical and Vickers Indenter

For a conical indenter, the projected area is obtained as:

Ac = πh2c tan2 θ (3.2.12)

The projected area of Vickers indenter is determined as (Fischer-Cripps and Mustafaev,

2000):

Ac = 4h2 tan2 θ (3.2.13)

3.2.2 Indentation Analysis

The indentation hardness H and modulus M of any axis-symmetric indenter can be estimated

by:

H =P

Ac(3.2.14a)

M =

√π

2

S

Ac(3.2.14b)

Figure 3.4: Berkovich geometry of contact

24

The indentation depth hr and hc are corresponding to the applied load; hr is the depth of the

residual impression, and hc is the contact depth which can be determined as:

hc = hmax − ξPmaxS

(3.2.15)

According to Oliver and Pharr (1992), the projected contact area Ac can be expressed by the

indentation depth hc as shown in Table 3.1. Sakharova et al. (2009) reported that areas of

indenters differ due to the imperfection at the tip. They presented the area function of the

indenter based on the numerical simulation of indentation process as shown in Table 3.2.

Another term in indentation analysis is the load frame compliance Cf which is the sum of the

compliance of the load frame that needs to be calibrated together with the area function of the

indenter. Oliver and Pharr (2004) developed a simple and accurate calibration procedure for

determining the total measured compliance C, which is defined as the inverse of the measured

stiffness and can be expressed by:

C = Cf +

√π

2M

1√Ac

(3.2.16)

The second term in Equation (3.2.16) is the contact which is two act like springs in series, and

the area function Ac can be determined from the measurement of the compliance as a function

of depth hc if Cf is known. The area function is proposed as (Oliver and Pharr, 2004):

Ac =8∑

n=0

Cn (hc)2−n

(3.2.17)

where Cn are constants, determined by curve fitting procedure and n is the number of constant.

Table 3.1: Project contact area, geometry factors and intercept factor for various types of indenters

IndenterProject contact

areaSemi-angleθ(degree)

Effectivecone angleα(degree)

Interceptfactor ξ

Spherical Ac ≈ 2πRhc - - 0.75

Berkovich Ac = 3√

3h2c tan2 θ 65.27◦ 70.3◦ 0.75

Vickers Ac = 4h2c tan2 θ 68◦ 70.3◦ 0.75

Cone Ac = πh2c tan2 θ α α 0.727

Table 3.2: Different indentation area function from numerical analysis (Sakharova et al., 2009)

Types of Indenter Area function A (µm2)

Berkovich A(h) = 24.675h2 + 0.562h+ 0.03216

Vickers A(h) = 24.561 (h+ 0.008)2

+ 0.206 (h+ 0.008)

Conical A(h) = 24.5 (h+ 0.011427)2

3.3 Heterogeneous Materials

Composite materials exhibit several types of heterogeneity which can be observed from the

perspective of different lengths scale. According to the literature (Constantinides et al., 2006;

Kanit et al., 2003; Pichler and Lackner, 2008; Zaoui, 2002), continuum micromechanics theory

25

offers a framework to distinguish constituents in heterogeneity. The idea of continuum microme-

chanics is a possibility of separating a heterogeneous material into phases with “on-average”

constant material properties. Continuum indentation analysis is based on spatially homoge-

neous mechanical properties. One of the most important elements of continuum approach is

the use of Representative Volume Element (RVE) to describe heterogeneous material statisti-

cally. The transition from heterogeneous material to homogeneous material can be ensured by

the characteristic size L of RVE (condition of separation) as:

d� L� (h,D) (3.3.1)

where d is the characteristic length scale of the local heterogeneities in RVE, h is the indentation

depth, and D is the characteristic size of the microstructure. Figure 3.5 illustrates a scale

transition to detect a homogeneous region of a heterogeneous material.In addition, general

nanoindentation tests are based on self-similarity that have no length scale limit. Thus, it

is important to select an appropriate indentation depth h and characteristic size L of RVE.

A good estimation is by using 3h for Berkovich indenter, and h for corner cubic, in order

to obtain an effective volume proportion (Constantinides et al., 2006). If h is much smaller

than the characteristics size of a microstructure D, i.e., h � D, then the indentation results

will yield a single phase of the material properties. On the other hand, if h is much greater

than D4, h � D, the average composite material properties will have to be obtained from

the analysed results. It should be noted that more tests (N � 1) need to be conducted

on a grid that has a spacing l larger than the effective volume proportion of the indenter

and the characteristic size of microstructure, i.e., l√N � D . This is to ensure that the

statistical analysis in each phase will present the surface fraction of the total phases. Figure

3.6 presents the important grid indentation technique in order to identify several phases of

different properties of a heterogeneous material. The grid spacing l should be defined to avoid

overlapping of each indentation impression as seen in Figure 3.6(c).

A cementitious material such as cement paste has several phases of the mechanical properties.

Jennings (2000), Tennis and Jennings (2000) proposed a model for determining two types of

calcium silicate hydrate (CSH), which are high density (HD CSH) and low density (LD CSH),

at different points of the specimen’s geometry. The content of CSH is determined in term of

its volume fraction of the indentation grid. Constantinides and Ulm (2004) determined two

types of CSH with Portlandite (CH), and clinker with nanoindentation. The result shows

that decalcification of the CSH phases is the primary source of nanometer-scale elastic mod-

ulus degradation. Nemecek et al. (2011) studied the reaction products of alkali activated fly

Figure 3.5: Element volume of material in a microscopic structure: (a) heterogeneous (b) localheterogeneous (c) homogeneous (adapted after (Dormieux et al., 2006))

26

Figure 3.6: Schematic representation of grid indentation technique: (a) average composite materialwith large indentation depth (h� D), (b) low indentation depth with one phase (h� D),(c) grid

indentation technique giving several phases using low indentation depth (h� D) adapted after(Constantinides et al., 2006))

ash-based cementitious composites (AAFA) with nanoindentation and environmental scanning

electron microscope (ESEM). Their AAFA samples were made using the ratio of activator to

solid of 0.531 and cured at 80◦C for 12 hours. They observed four peaks of the reaction products

of AAFA as:

- N-A-S-H

- Partly-activated slag

- Non-activated slag

- Non-activated compact glass (Raw fly ash particles)

The N-A-S-H phase is pure and is relative to the mechanical strength of AAFA matrix. The

contents of Si ions in N-A-S-H can be increased by the presence of Si ions in the raw materials.

It has been found that the increasing condensation degree of Si ions in N-A-S-H relates directly

to the mechanical strength gain (Fernandez-Jimenez and Palomo, 2003, 2005). The partly-

activated slag phase is intermixed with slag-like particles. The non-activated slag phase is

porous and contains non-activated slag-like particles. The non-activated compact glass phase

is solid, non-activated glass sphere.

3.4 Microporomechanics

The modulus and hardness can be representative of the particle properties such as particle

stiffness (λ), Poisson’s ratio (vs), cohesion (κ), friction coefficient (ζ) and packing density (η)

(Constantinides et al., 2006; Constantinides and Ulm, 2007; Dukino and Swain, 1992; Jennings

et al., 2007; Ulm et al., 2007):

M = λ×ΠM (vs, η, η0) (3.4.1)

H = κ×ΠH (ζ, η, η0) (3.4.2)

where ΠM and ΠH dimensionless scaling relating to stiffness and hardness. This scaling rela-

tionship can be determined by linear and non-linear microporomechanics theory. It is required

to form a hypothesis to apply microporomechanics to indentation analysis. In the present case,

the tested material is assumed to be porous. Pipilikaki and Beazi-Katsioti (2009) reported

that the porosity size of binding cementitious materials is smaller than 2.5nm. Thus, it can

27

be assumed that the characteristic size of the porosity is much smaller than the maximum

indentation depth (h) for scaling separability reason (Constantinides and Ulm, 2007).

The self-consistent or polycrystal micromechanical, model is similar to micro-elasticity of a

porous material whose solid is a granular material (Constantinides and Ulm, 2007). The self-

consistent scheme assumes that each particle of a given phase (pore or solid) reacts as if it is

embedded in an equivalent homogeneous medium (Dormieux et al., 2006). The solid percolation

threshold (η0) is the limit that the solid fraction requires for providing a continuous force path

through the system. Clear matrix-porosity morphology has continuous packing density in the

solid phase, i.e. 0 < η ≤ 1. However, a perfectly disordered porous material has a solid

percolation thread of η = 0.5 and below, i.e., if η < 0.5, then the values of the composite bulk

and shear modulus become negative, which will be considered unstable. The modulus function

is provided by the following relationship (Dormieux et al., 2006; Ulm et al., 2007):

ΠM (rs, η, η0 = 0.5) = G(9ηrs + 4G+ 3rs) (3rs + 1)

4 (4G+ 3rs) (3rs + 1)(3.4.3)

where rs = 2 (1 + vs) /3 (1− 2vs) > 0 and G is the composite bulk and shear modulus, defined

as:

G =1

2− 5

4(1− η)− 3

16(2 + η) +

1

16×[144 (1− rs)− 480η + 400η2 + 408rsη

2 + 9rs (2 + η)2]1/2 (3.4.4)

For solid Poisson’s ratio of vs = 0.5, Equation (3.4.4) becomes a linear scaling of modulus with

packing density (η):

ΠM = 2η − 1 ≥ 0 (3.4.5)

The hardness with packing density (η) for the scaling relation is introduced by non-linear

micromechanics:

ΠH (ζ) = Π0 ×[1 + (1− η) ζ − (d− eη) ζ2 − (f − gη) ζ5

](3.4.6)

where Π0 is a function of frictionless portions as:

Π0 =12η (a− bη) [(2η − 1) (2 + η)]

1/2

(1− cη) (2 + η)(3.4.7)

Based on microporomechanics, the coefficients suitable for Berkovich indenter are: a = 0.19567,

b = 0.03739, c = 0.77999, d = 20.3138, e = 31.5352, f = 52.1817 and g = 99.3465 with η0 = 0.5

(Ulm et al., 2007).

An inverse application to determine the solid properties and packing density is introduced by

Ulm et al. (2007) and Bobko and Ulm (2008). In an inverse application, the solid properties (λ,

vs, κ, ζ) and local packing density (η) are unknowns. An algorithm implemented as a quadratic

minimization problem between the experimental values and the theoretical scaling relationship

can be written as:

min

N∑i=1

[Mi − λ×ΠM (vs, ηi)− (Hi − κ×ΠH (ζ, ηi))]1/2

(3.4.8)

where N is the total number of the indentation tested points, Mi and Hi are as obtained from

Equations (3.4.1) and (3.4.2). Solving Equation (3.4.8) will provide the analytical results of the

28

scaling relationship between the indentation modulus and hardness versus the packing density

of the tested material.

Mercury intrusion porosimetry (MIP) is a common measurement for providing a valid estimation

of the pore size distribution of porous solid (Diamond, 2000). The measurement of porosity

with MIP, however, appears to be valid with limited application and difficulty in estimating

the porosity of cementitious based material. A way to determine porosity with statistical

indentation technique is by the new nonintrusive way (Ulm et al., 2007). Therefore, it is

possible to calculate the total porosity of the cementitious material from:

φ =N∑j=1

fj (1− nj) (3.4.9)

3.5 Deconvolution Technique

Heterogeneous materials or composite materials are generally complex. Thus, a statistical

technique is used for analysis of indentation results. It is computationally more convenient to

deconvolute the cumulative distribution function (CDF) rather than the probability density

function (PDF) (Ulm et al., 2007). The reason for using CDF rather than PDF is that the

generation of the experimental PDF usually requires a choice of bin size for histogram construc-

tion. Thus, the deconvolution technique starts with the generation of the experimental CDF.

In the following equations, let X be a type of the mechanical properties and, as before, N is

the number of indentation tests conducted on the specimen. The experimental CDF can be

written as:

Φe(X) =i

N− 1

2Nfor i ∈ [1, N ] (3.5.1)

in which, the following Gaussian distribution gives the theoretical CDF for each phase:

Φtj(X) =1

σj√

2π

∫ Xj

∞exp

(− (s− µi)2

2σ2j

)dt for i ∈ [1, N ] (3.5.2)

where µj is the mean, σj is the variance of the distribution and j is a phase present in the

sample. The theoretical CDF function can be found by minimising the following error function:

min

N∑j=1

∑(fiΦ

tj (X)− Φe (X)

)2(3.5.3)

Moreover, the total volume fraction can be expressed by a summation of all the single phases

fj as:

n∑j=1

fj = 1 (3.5.4)

where n is the total number of phases identifiable from the analysis. To avoid the overlapping

of each phase, the constraint of the minimisation problems is required as:

µi (X) + σj (X) ≤ µj+1 (X) + σj+1 (X) (3.5.5)

The mean and standard deviation of the results from the deconvolution technique can then be

used to determine the mechanical properties and surface volume fraction of each phase.

29

3.6 Indentation Fracture Toughness

Nanoindentation data can be used to determine the fracture toughness of materials to measure

the radial and lateral cracks indentation points (Chen and Bull, 2007; Ling, 2011). A measure

of fracture toughness using radial and lateral cracks is based on fracture mechanics. According

to Lawn et al. (1980), the indentation fracture toughness is formulated as:

Kc = ε

(E

H

)nP

c3/2(3.6.1)

where ε is empirical calibration constant, c is the radial crack length. A number of researchers

(Anstis et al., 1981; Dukino and Swain, 1992; Field et al., 2003; Laugier, 1987) conducted to

determine ε and n terms in Equation (3.6.1) using indentation test. It can be estimated from

indentation with Berkovich indenter that:

Kc = 1.073xv

( aL

)1/2(E

H

)2/3P

c3/2(3.6.2)

where Kc is fracture toughness, xv is 0.015, and a and L are measured by radial cracking length

as shown in Figure 3.7.

The indentation P−h curve can also be represented by a consideration of the energy transferred

during loading and unloading (Oliver and Pharr, 1992). The area under the loading (Ploading)

and unloading (Punloading) versus the indentation depth (h) curves represents the plastic energy

(Wp) and elastic energy (We), as shown in Figure 3.8. The total energy (Wt) is the sum of Wp

and We, or the total area under the P − h curve. Oliver and Pharr (1992) suggested using the

loading and unloading curve fitting with a power law function.:

Ploading = α · h2 (3.6.3a)

Punloading = β · (h− hr)m (3.6.3b)

where α, β and m are constants. The variation of m in the range 1.2 ≤ m ≤ 1.6, i.e., m = 1

for flat punch, m = 1.5 for parabolic and m = 2 for cone intender tip. By fitting loading and

unloading curve with a power law function, Equation(3.6.3), Wt and We can be obtained from:

Wt =

∫ hmax

0

Ploadingdh (3.6.4)

We =

∫ hmax

hr

Punloadingdh (3.6.5)

Figure 3.7: Berkovich indentation crack parameters (adopted after (Ling, 2011))

30

Figure 3.8: Indentation load (P ) and depth (h) curve represent energies (adapted after (Cheng andCheng, 2004))

The relationship between Wp/Wt and hr/hmax can be expressed as (Cheng and Cheng, 2004;

Chen and Bull, 2007):

Wp

Wt= 1−

[1− 3 (hr/hmax)

2+ 2 (hr/hmax)

3

1− (hr/hmax)2

](3.6.6)

The total indentation work Wt can be determined from:

Wt = Wp + Ufracture +We +Wother (3.6.7)

where Ufracture is the fracture dissipated energy, and Wother is another energy transferred such

as heat energy dissipation or thermal drift. However, in the present research, Wother from

the thermal drift is ignored in the calculation of Wt. The critical energy (Gc) and fracture

toughness (Kc) can be expressed as (Taha et al., 2010):

Gc =∂Wfracture

∂A=Ufracture

Ac(3.6.8)

Kc = (GcM)1/2 (3.6.9)

The approach will be used in determining fracture toughness based on the assumption that the

crack growth under indentation loading is stable.

3.7 Finite Element Analysis

The finite element (FE) method is commonly used for modelling different engineering problems.

According to some literature (Bhattacharya and Nix, 1991; Lichinchi et al., 1998; Stauss et al.,

2003), FE simulation could be an appropriate method for determining indentation properties.

Knowing that purely elastic deformation takes place only during the beginning of the indenta-

tion process, the Von-Mises yield criterion is applied to determine the occurrence of the plastic

deformation. The equivalent Von-Mises stress is given by:

31

σVM

=

√(σ1 − σ2)

2+ (σ2 − σ3)

2+ (σ3 − σ1)

2

2(3.7.1)

where σ1, σ2 and σ3 are the three principal stresses. When σVM reaches the yield strength σy,

the material begins to deform plastically. For the conical indentation in an elastic-perfectly

plastic behaviour material, indentation pressure is written as(Gaillard et al., 2003):

pm =2

3σy

[1 + ln

(M · tanα

3σy

)](3.7.2)

where pm is the indentation pressure (i.e. pm = Pmax/Ac), M is indentation modulus of the

sample and α is the conical indentation face angle, as already presented in Table 3.1. The

elastic-perfectly plastic behaviour is governed by the elastic and yielding properties of the

material.

3.8 Time-depending Nanoindentation

It is well known that materials exhibit viscous and elastic behaviours under deformation. Vis-

coelastic materials store energies under deformation, return to its original state upon removal

of stress, and can be under the influence of time-dependent stress-strain factors. Nanoindenta-

tion can be used to determine quantitative viscoelastic properties. Previously, the method of

analysis of nanoindentation was assumed that behaviours of material are in an elastic-plastic

manner. However, time-dependent viscoelastic properties, such as creep and relaxation, can

occur under indentation impressed (Fischer-Cripps and Mustafaev, 2000; Ling, 2011), therefore

it is necessary to assess these properties. The relationship between indentation load (P ) and

penetration depth (h) is derived as (Sneddon, 1965; Vandamme et al., 2012):

P (t) = γMoh(t)1+1/n (3.8.1)

where γ is:

γ =2

(√πB)

1/n

n

n+ 1

[Γ (n/2 + 1/2)

Γ (n/2 + 1)

]1/n

(3.8.2)

B can be defined by z = Brn, as shown in Figure 3.9 and Γ is the Euler Gamma function given

as:

Γ (X) =

∫ ∞0

tx−1 exp(−t)dt (3.8.3)

The instantaneous bulk (Ko) and instantaneous shear modulus (Go) are related to the instan-

Figure 3.9: Indentation geometry probe

32

taneous material modulus (Eo) and instantaneous modulus (Mo), given as:

Mo =Eo

(1− v2)(3.8.4a)

Go =Eo

2 (1 + v)(3.8.4b)

Ko =Eo

3 (1 + 2v)(3.8.4c)

where v is the Poisson’s ratio of the tested material. The coefficients n and B of the indentation

geometry in Equation (3.8.2) are defined according to the indentation shape as listed in Table

3.3.

As viscoelastic behaviour is the role played by time, if the material is ideally viscous fluid, the

stress can be instantaneously infinite under constant strain. However, actual material behaviour

shows that under constant strain, the stress generally decreases from its initial value rapidly and

later more gradually as shown in Figure 3.10 (Pipkin, 2012). This behaviour is known as stress

relaxation. Another important behaviour is creep, i.e., deformation of viscoelastic materials

under constant stress increases with time. Consider a rigid indenter of axisymmetric shape

impressing an infinite half-space is made of no aging linear viscoelastic material. The functional

Equation (3.8.1) can be applied to obtain the solution to a linear viscoelastic indentation in

the Laplace domain from the linear elastic solution by replacing the elastic constants in the

elastic solution. Figure 3.11 illustrates conventional relaxation and creep stress with time, which

present the behaviour of general viscoelastic material.

According to Riesz theorem (Stein, 1956), if the functional J is linear and equicontinuous which

is the creep function, strain (ε) is given as:

ε(t) =

∫ t

τ0

J(t− τ)dσ(τ) (3.8.5)

where τ0 is the time at rest without stress and strain, and τ is the time domain. Equation

(3.8.5) can be re-written by inverse relations as:

Table 3.3: Indentation geometry

Indenter n B

Cone or Berkovich 1 cot θ

Sphere 2 1/ (2R)

Flat punch →∞ 1/ (an)

Figure 3.10: Viscoelastic behaviour (a) ideal, (b) actual viscous fluid

33

Figure 3.11: Relaxation and creep stress with time

σ(t) =

∫ t

τ0

M(t− τ)dε(τ) (3.8.6)

where exchanging roles of stress and strain that gives M (t− τ) is the specific relaxation func-

tion, i.e., the stress response to a unit step of strain (Marques and Creus, 2012). In this case,

Equation (3.8.1) can be applied using Riesz theorem with indentation load and displacement

relation in time domain as:

P (t) = γ

∫ t

τ0

M (t− τ)dh

1/1 + n (τ)

dτdτ (3.8.7)

For relaxation test, the displacement can be described via a Heaviside step function:

h(t) = hmaxH (t) ; H (t) =

{0 for t < 0

1 for t ≥ 0

}(3.8.8)

Substituting Equation (3.8.8) into Equation (3.8.7) yields:

P (t) = γ

∫ t

τ

M (t− τ)h1/1 + n

max δ (τ) dτ (3.8.9)

where δ is Dirac delta function. This equation can be obtained in the Laplace domain as:

L{P} (s) = γsL{M} (s)h1/1 + n

max (3.8.10)

Moreover, application of the standard Laplace table yields in P (t) in the time domain:

P (t) = γM (t)h1/1 + n

max (3.8.11)

Thus, the contact relaxation modulus at any time t, Mc (t), can be determined as:

Mc (t) =1

γh1/1 + n

max

P (t) (3.8.12)

In this case, the functional Equation (3.8.1) also leads to the following time-dependent inden-

34

tation displacement under Riesz theorem as:

h (t)1/1 + n

=1

γ

∫ t

τ0

J (t− τ)dP (t)

dτd (τ) (3.8.13)

Heaviside step loading P (t) = PmaxH (t) is then applied to Equation (3.8.13), resulting in:

h (t)1/1 + n

=1

γ

∫ t

τ0

J (t− τ)Pmaxδ (τ) d (τ) (3.8.14)

This equation can be obtained in a Laplace transform as:

L{h}(s)1/1 + n =1

γsL{J} (s)Pmax (3.8.15)

Applying inverse Laplace transform to this equation gives the formulation of the contact creep

compliance Jc (t) as:

Jc (t) =γ

Pmaxh (t)

1/1 + n(3.8.16)

Moreover, the indentation creep compliance is linked to the contact relation modulus in the

Laplace domain as:

[sL{M} (s)]−1

= sL{J} (s) (3.8.17)

This material behaviour is time dependent and is generally obtained in term of a linear viscoelas-

tic model (Rheological model), such as Maxwell model, three-element Kelvin-Voigt model, and

combined four-element Maxwell-Kelvin-Voigt model, as shown in Figure 3.12 (Fischer-Cripps,

2004). Ideally, in a perfect linear elastic and massless condition, helicoidal spring represents

Hooke’s Law as:

σ (t) = Eε (t) (3.8.18)

The dashpot is an ideal viscous element that can be expressed by a rate proportion of the

applied stress:

σ (t) = ηε (t) (3.8.19)

where ε = dε/dt is the rate of strain and η is the unit of viscosity.

The case of a conical indenter for Maxwell model, the time-dependent of penetration depth

with constant indentation load Po can be expressed by:

h2 (t) = γPo

[(1

Mo+

1

ηMt

)](3.8.20)

Using a similar approach, the penetration depth with time for Kelvin-Voigt model is represented

as:

h2 (t) = γPo

[(1

Mo+

1

Mv

(1− exp(−Mv

ηvt)

))](3.8.21)

In the case of combined Maxwell-Kelvin-Voigt model, the penetration depth increases with time

according to:

h2 (t) = γPo

[(1

Mo+

1

Mv

(1− exp

(Mv

ηvt

))+

1

ηM

)](3.8.22)

35

Figure 3.12: Mechanical models of time depending properties of material (a) Maxwell model (b)Kelvin-Voigt Model (c) Combined Maxwell-Kelvin-Voigt model

Equations (3.8.20) to (3.8.22) can be applied to loading whereby the contact radius increases

with time and the term in the square brackets represents the time response of the materials

(Fischer-Cripps, 2011).

3.8.1 Indentation Contact Relaxation Modulus

The relaxation modulus is obtained due to decrease in stress under constant strain as shown in

Figure 3.11. From Equation (3.8.12), the contact relaxation modulus Mc (t) can be determined

for conical indenter as (Vandamme et al., 2012; Vandamme and Ulm, 2006):

Mc (t) =π cot θ

2h2maxP (t) (3.8.23)

For spherical indenter, the contact relaxation modulus Mc (t) is:

Mc (t) =3P (t)(

4√Rhmax

)3/2P (t) (3.8.24)

This equation can be used to obtain instantaneous modulus Mo, which is (Mc (t = 0)). However,

practically, the measurement of the indentation contact relaxation modulus in Equation (3.8.23)

or Equation (3.8.24) may be underestimated because of the effect of a sharp indenter tip such

as Berkovich on the time-independent and instantaneous plastic deformation. As a result, the

contact relaxation modulus relations take into account the occurrence of plasticity during the

loading phase, the measured load relaxation response P (t) can then be linked to the indentation

contact relaxation modulus function (Cao et al., 2010; Vandamme et al., 2012). Therefore,

normalised contact relaxation modulus (Mc (t)) is defined as the relaxation modulus divided

by instantaneous modulus (Mo) and depends only on the relaxation load, thus, the contact

relaxation modulus rate is:

Mc (t) =Mc (t)

Mo=P (t)

Pmax(3.8.25)

Prony series can describe the contact relaxation modulus as (Cao et al., 2010):

36

Mc (t) = M∞ +N∑i=1

Ri exp

(−tτMi

)(3.8.26)

where Ri is the relaxation coefficient and τi is relaxation time.

The relaxation modulus Mc (t) is greatly significant in engineering applications. Construction

materials with energy dissipation ability are required for vibration reduction and hazard miti-

gation of civil engineering structures. It is important to determine the damping behaviour of

construction materials, as it is generally required to determine the loss factor which is defined

by the ratio of the loss modulus Mcl to the storage modulus Mcs at a given frequency ω (Cao

et al., 2010; Nashif et al., 1985). The loss factor, tan δ, can be determined as:

tan δ =Mcl

Mcs=

N∑i=1

giωτgi

1 + ω2 (τgi )2

(1−

N∑i=1

gi +N∑i=1

giω2 (τgi )

2

1 + ω2 (τgi )2

)−1(3.8.27)

Using Equation (3.8.27), the damping capacity, which is the energy dissipated per cycle of

motion, can be determined as:

4U = 2πUmax tan δ (3.8.28)

where Umax is maximum potential energy or maximum kinetic energy.

The contact relaxation modulus can be analysed in term of Rheological models in Figure 3.12.

In this research, a relationship is established between the contact relaxation modulus and creep

compliance, using Equation (3.8.17), and in respective of the three models. For Maxwell model,

it is proposed that the analytical contact relaxation is given as:

Manc (t) = Mo · exp

(−Mo

ηMt

)(3.8.29)

Similarly, the analytical contact relaxation modulus of Kelvin-Voigt model is proposed as:

Manc (t) =

MoMv

Mo +Mv+M2o · exp

(− (Mo+Mv)

ηvt)

Mo +Mv(3.8.30)

Also, the analytical contact relaxation modulus of Maxwell-Kelvin-Voigt model is proposed as:

Manc (t) =

MoMvηMMoηM +MvηM +MoMv

+M2o η

2M · exp

(−MoηM+MvηM+MoMv

ηv(Mo+ηM ) t)

(Mo + ηM ) (MoηM +MvηM +MoMv)(3.8.31)

3.8.2 Indentation Contact Creep Compliance

The creep properties of cementitious materials are still an enigma because the difficulty linked

to the timescale involved. Generally, the complex creep behaviour of cementitious material

is largely related to the viscoelastic response of the primary hydration or reaction products

and the binding phase of hardening. Creep testing involves applying a constant instantaneous

stress σ0 to the specimen and measuring strain as a function of time during stressing. The

creep compliance depends on the geometry of the axisymmetric indenter and maximum value

of control variables in the indentation process.

According to Vandamme and Ulm (2013, 2009), the measured indentation displacement re-

sponse h (t) is related to the contact creep compliance rate Jc (t) as:

37

Jc (t) =2a (t)

Pmaxh (t) (3.8.32)

where a (t) is the radius of the projected contact area, and h (t) is the change in the inden-

tation depth over the holding time, which can be expressed by a single power and logarithm

function as (Vandamme and Ulm, 2013, 2009):

h (t) = x1 ln

(t

x2+ 1

)+ x3t+ x4 (3.8.33)

where x1, . . ., x4 are constants. In the Equation (3.8.33), the material related term is only the

logarithmic term. Thus, h (t) can be replaced by the term of x1/t. From Equation (3.8.32), the

long-term contact creep compliance rate for short indentation creep experiment is given by:

Jc (t) =1

Ct, where C =

Pmax2acx1

(3.8.34)

where C is the contact creep modulus, and ac is the contact radius of maximum projected

area between the indenter and the indented sample just before the unloading phase, i.e., ac =√Ac/π. The higher contact creep modulus will lead to the lower logarithm creep of the material.

In Equation (3.8.32), it is required to input the radius of the projected contact area which cannot

determine directly from indentation test results. Therefore, in this research, an alternative

method is proposed by adapting to determine the contact creep compliance rate using directly

indentation load-displacement curve. Based on Equation (3.8.16), the contact creep compliance

rate for conical indenter can be described by:

Jc (t) =d

dt

(2 tan (a)

πPmaxh (t)

2

)(3.8.35)

The long-term creep compliance rate, h (t) can be replaced by x1/t, and substitute h (t) into

hmax then re-arrange equation to obtain the contact creep modulus (C) as:

Jc (t) =1

Ct, where C =

πPmax4 tan (a)hmaxx1

(3.8.36)

Similarly, the contact creep compliance rate for spherical indenter is given by:

Jc (t) =d

dt

(4√R

3Pmaxh (t)

2/3

)(3.8.37)

Thus, the long-term creep compliance is below:

Jc (t) =1

Ct, where C =

Pmax

2√Rhmaxx1

(3.8.38)

The contact creep compliance can also be described on Rheological models. The contact creep

compliance for Maxwell model is given as:

Janc (t) =1

Mo+

1

ηMt (3.8.39)

For Kelvin-Voigt model, the contact creep compliance is:

Janc (t) =1

Mo+

1

Mv

[1− exp

(−Mv

ηvt

)](3.8.40)

38

Similarly, the contact creep compliance for Maxwell-Kelvin-Voigt is:

Janc (t) =1

Mo+

1

Mv

[1− exp

(−Mv

ηvt

)]+

1

ηMt (3.8.41)

3.8.3 Numerical Analysis of Viscoelasticity

To confirm the validity of indentation viscoelastic properties in cementitious material, the com-

putation study was performed with a commercial finite element (FE) software package ANSYS

(ANSYS R○, 2015) in the present study. A 2-D axisymmetric model was employed for simula-

tion of the material behaviour in nanoindentation process as shown in Figure 3.13. A conical

indenter tip with half angle of 70.3◦, which is the effective angle of Berkovich as shown in Table

3.1. A fine mesh was defined under the contact area and near the indenter tip for a more

accurately as shown in Figure 3.14. The indented material model meshed with 4731 four-node

quadrilaterals. The contact between the rigid conical indenter and the indented sample was

considered frictionless contact. Indenter modulus was assumed 1141 GPa and Poisson’s ratio of

the indenter was 0.07. The FE indentation process was simulated with a non-linear geometry

method of displacement control.

Contact relaxation modulus

For contact relaxation modulus, the indented material was modelled using the Maxwell model

in the Equation (3.8.29) with Poisson’s ratio as 0.25 and Modulus of elasticity as 10 MPa. The

contact relaxation modulus of FE simulation and analytical and Maxwell model results are

shown in Figure 3.15. The results show that there is no difference between the numerical and

the proposed analytical results, but there is an 8.29% difference between the numerical results

and the original Maxwell model of the normalised contact relaxation modulus. This difference is

common in practice and a multiple correction factor is recommended to the Rheological model

for elastic indentation (Vandamme et al., 2012). Thus, Equation (3.8.25) is somewhat inaccu-

rate in the sense that it also fails to capture this radial contraction. As a result, indentation

relaxation modulus can be determined from nanoindentation test of cementitious material and

the time-dependent relaxation modulus can also be obtained.

Figure 3.13: A conical indenter impress on specimen

39

Figure 3.14: Meshing configuration

Figure 3.15: Normalised contact relaxation modulus with analytical, numerical solution andMaxwell model

40

Contact creep compliance

FE simulation of contact creep compliance was carried out using Maxwell creep model with

varying Poisson’s ratios as 0.25, 0.35 and 0.49. Figure 3.16 shows the results of the numerical

contact creep compliance and of using the Maxwell model. The errors between numerical and

Maxwell solutions are less than 1%. Therefore, the contact creep compliance can be determined

with indentation results.

For a logarithmic creep function, Equation (3.8.33) was used to obtain the change in indentation

displacement. With logarithmic creep function, a long-term contact creep compliance was

determined using the Vandamme Equation (3.8.34) and Equation (3.8.36), which is formulated

in the this research. Figure 3.17 shows the contact creep compliance rate based on the two

equations. It can be seen that there is a minor difference between the proposed and Vandamme

equations. For a sharp indenter such as Berkovich and conical indenter, it is hard to obtain a

projected area during the indentation process. Thus, the error between the two equations was

expected because the contact radius of the maximum projected area ac had to be assumed in

Vandamme solution, i.e., ac ≈√Ac/π because exact projected area ac physically impossible to

determine. As an alternative approach, Equation (3.8.36) is convenient to obtain the contact

creep compliance rate without knowing the contact radius of projected area ac. Therefore, the

proposed equation can determine the long-term behaviour without knowing the projected area

ac, and a nanoindentation creep experiment can determine the contact creep compliance. In

addition, the long-term contact creep compliance also can be obtained from a short period of

indentation test results.

Figure 3.16: Numerical and Rheological (Maxwell) solution of contact creep compliance

41

Figure 3.17: Contact creep compliance rate with numerical analysis

3.9 Indentation Stress-Strain Curve

Section 3.2 introduced the “Indentation stress pm” and “Indentation strain εi”. These indenta-

tion stress-strain curves can be measured by contact pressure pm, contact radius a, and indenter

radius R, as shown in Figure 3.18. The indentation stress-strain curve can be expressed by three

regimes of deformation that is the elastic, elastic-plastic and fully plastic regime. The elastic

regime is covered before the initial yield point. The region between the initial and stationary

yield point is described as the elastic-plastic regime, and over the stationary yield point is fully

plastic regime (Cao and Zhang, 2008). Determination of indentation stress pm requires indenta-

tion loading P and the projected contact area Ac with respect to the indenter tip and indented

material. It is noted that indentation stress pm also signifies the instant hardness H = P/A.

However, indentation tests cannot measure indentation contact depth hc, only measure the

maximum depth h. Therefore, it is required to determine the relationship between h and hc

as a function of hc = f (h) which depends on the indenter tip geometry and indented material.

The function hc = f (h) can be formulated by continuous measures of contact stiffness (CSM)

between indented material and indenter process. Knowing hc = f (h) and the instant contact

area of the indenter, the instant stress pm during the loading process is (Martinez et al., 2003):

pm =P

Ac(3.9.1)

The contact area function Ac can be obtained simply by curve fitting with hc as:

Ac =8∑

n=0

Cn (hc)2−n

(3.9.2)

The contact area function, in the Equation (3.9.2), is calibrated by a standard material such as

fused silica that has 72 GPa of elastic modulus and Poisson’s ratio of 0.18. The contact radius

and indenter tip radius are required to obtain indentation strain. The equivalent contact radius

42

Figure 3.18: Schematic uniaxial and indentation stress-strain curve (adopted after (Martinez et al.,2003))

a can be calculated directly from the contact area Ac as:

a =√

2Rhc − h2c for Spherical tip

a = 2√Ac/3

√3 for Berkovich tip

Cao and Zhang (2008) reported that if it is assumed that the blunt tip profile is approximately

circular, the maximum contact depth δ0, which depends on the indenter shape can be estimated

by:

δo = R (1− sin θ) (3.9.4)

where θ is semi-angle of the indenter tip. From contact mechanics analysis, the equivalent

indentation strain at shallow contact depth can be expressed as:

εi = Ka

Rwhen hc < δo (3.9.5a)

εi ≈ log

(hcδo

)when hc ≥ δo (3.9.5b)

where K is a constant, which can be determined by calibrating with a standard material. Thus,

the indentation stress-strain curve can be observed from indentation test results, and it will

display material elasticity and plasticity properties. According to Herbert et al. (2001), the

uniaxial tension stress-strain σt − εt are equivalent to indentation stress-strain pm − εI results,

and the yield strength σy can be approximated by:

σy ≈pm1.07

(3.9.6)

43

Indentation performed on the standard fused silica (E = 72 GPa) using the Spherical and

Berkovich indenter tip with CSM method can establish the validity of indentation stress-strain.

Firstly, the test was conducted using Spherical tip (150 µm of the radius). The relationship of

hc = f (h) is determined by the contact displacement hc and the instant maximum displacement

h using Equation (3.2.15). With Spherical tip, the indentation stress-strain curve on fused silica

is as presented in Figure 3.19. The result of the stress-strain curve shows that the modulus of

elasticity (initial slope) is 71.8±4.63 GPa, which is in good agreement with calibration constant

K = 7.445 ± 1.526. However the initial yield point σI

cannot be indicated, which means only

elastic regime can be observed with Spherical tip.

Similarly, the projected area function was indented with Berkovich tip (20nm of the radius)

on fused silica. The indentation stress-strain curve of fused silica is presented in Figure 3.20,

which shows elastic, elastic-plastic, and fully plastic behaviour of the material. The initial

yield point σI

is observed to be around 3.2 GPa, and the stationary yield point σII

is obtained

approximately 9 GPa. The stationary yield point σII

corresponding to its measured hardness

values was reported to be 8.5 GPa by Oliver and Pharr (1992). However, some literature (Cao

and Zhang, 2008; Martinez et al., 2003) reported that the initial yield point σI

is not clear to

observe because the fully plastic flow develops early in the loading stage with more than 20nm

of the radius of Berkovich tip.

Therefore, indentation method can provide indentation stress-strain curve, and this result can

be used to determine the yield strength of the indented material. The Spherical indenter with

150 µm of radius tip provides elastic regime only, but Berkovich tip can provide elastic to plastic

regime clearly. Martinez et al. (2003) studied the indentation stress-strain curve with different

indenters geometry to get a complete mechanical behaviour of the thin film as shown in Figure

3.21.

Figure 3.19: Indentation stress-strain curve on Fused silica with Spherical tip

44

Figure 3.20: Indentation stress-strain curve on Fused silica with Berkovich tip

Figure 3.21: The deformation regimes with different types of indenter tip (adapted after (Martinezet al., 2003))

45

3.10 Chapter Summary

The review of present tools of indentation analysis in this Chapter shows that indentation prop-

erties such as indentation modulus M and hardness H that are related respectively to elastic

and strength properties with the projected area during the loading and unloading process. This

projected area is generally estimated using Oliver-Pharr method. Based on the principle of mi-

croporomechanics, the indentation modulus and hardness represent particle properties such as

particle stiffness λ, Poisson’s ratio vs, cohesion κ, friction coefficient ζ and packing density η.

The statistical analysis tool of indentation of a multiphase material is characterised using inden-

tation of a material from the micro-scale to macro-scale. These tools for indentation analysis

of cementitious material identify the link between the multiphase compositions, and elastic,

microstructure and viscoelastic properties. The indentation process can determine indentation

stress-strain relationship of the indented material, which depends on the geometry of indenter

tip during the indentation procedure. The indentation stress-strain curves are useful for de-

termining the mechanical properties of the indented material. The analysis of nanoindentation

data presented in the Chapter has industrial and scientific benefits such as:

- Assessment of the viscous properties of the reaction products in binding cementitious

materials.

- Indentation results are comparable to conventional experimental results, such as Modulus,

hardness, relaxation modulus, creep and fracture toughens. Thus, one set of indentation

test results is representative of, and can replace, these conventional experiments.

- Indentation test identifies determinants of properties in nanoscale which enables an as-

sessment of the long-term macroscopic behaviour that is in the orders of magnitude faster

than what can be found by macroscopic analysis.

46

Chapter 4

Properties of Cement Materials

4.1 Introduction

This Chapter contains two main parts. The first part is on the properties of ordinary Portland

cement (OPC), particularly, based on nanoindentation technique. The second part is a series

of an experiment on blended cement using statistical analysis tools. Ordinary Portland cement

(OPC) is made primarily from calcareous materials such as chalk or limestone, and argillaceous

materials such as shale or clay. This material is burnt at a high temperature of around 1450◦C

in a rotary kiln to form clinker which is then grounded with a requisite amount of gypsum into

a fine powder known as Portland cement. ASTM ASTM Standard C150 (2015) has designed

six types of Portland cement as Type I to V and White, differ primarily in aluminate (C3A)

contents and fineness. The general characteristics of different types of cement are listed in Table

4.1. Two important constitutes in cement are Alite (C2S) and Belite (C3S) (Neville, 2011).

In Australia, available cement types are General purpose cement (GP), General purpose lime-

stone cement (GL), Blended cement (GB), and Special purpose cements such as high early

strength cement (HE), low heat cement (LH), sulphate resisting cement (SR), and shrinkage

limited cement (SL) as shown in Table 4.2 (Australia Standard, 2010).

OPC is the main binder of concrete thus a large amount of OPC is produced due to the high

volume of concrete used worldwide (Flatt et al., 2012). It is, therefore, necessary, as a reference,

to know the properties of OPC in detail so it can be optimised. In this research, nanoindentation

technique is used for this purpose.

The first part of this Chapter, Section 4.2, presents indentation properties of OPC paste, and

the second part, Section 4.3, presents the mechanical properties of blended cement, paste and

mortar. An experimental programme of blended cement mixtures was designed according to

Taguchi’s approach (Roy, 2010), which is an efficient method for investigating optimum design

Table 4.1: General features of types of cement in ASTM C150 (Neville, 2011)

Type Classification Characteristics

I General purpose High C3S contentII Sulphate resistance Low C3A content (<8%)III High early strength Fine groundIV Low heat (slow reaction) Low C3S content (<50%) and C3AV High sulphate resistance Low C3A content (<5%)

White White colour Low MgO, No C4AF

47

Table 4.2: Properties and characteristics of cements in AS 3972-2010 (Australia Standard, 2010)

TypeSetting time Temperature Expansion Shrinkage

min max max. rise max. at 16 weeks max. 28 days(min) (hours) (◦C) (microstrain) (microstrain)

GP 45 6 - - -GL 45 10 - - -GB 45 10 - - -HE 45 6 - - -LH 45 10 23 - -SR 45 10 - 750 -SL 45 10 - - 750

parameters for the required performances. Also, Analysis of variance (ANOVA) and regression

were conducted to understand the better statistical relationship between properties on different

parameters and influence of parameters that contribute to measured variation of properties.

4.2 Indentation Properties of Ordinary Portland cement

Nanostructure of concrete is controlled by the structure of calcium silicate hydrated (CSH)

which governs fundamental properties such as strength, relaxation, creep, and fracture be-

haviour of hydrated cement. This demands a detailed knowledge of the nanostructure and how

it relates to local mechanical properties. Thus, it is important to understand nanostructure of

OPC. Several techniques have been used to understand CSH structure. Thomas et al. (1998)

studied determine two CSH morphologies using nitrogen but results were varying parameters

such as internal water content. Similarly, Tennis and Jennings (2000), revised model of the

microstructure of OPC that quantitatively predicts the volume of various phases. This model

provides a mean for quantifying the volumetric proportion of major hydration products such as

capillary porosity (MP), low-density CSH (LD-CSH) and high-density CSH (HD-CSH). Con-

stantinides and Ulm (2004) determined two types of CSH with Portlandite (CH), and clinker

with nanoindentation. The result shows that decalcification of CSH phases is the primary source

of nanometer-scale elastic modulus degradation. Generally, the complex viscous characteristic

of OPC is according to the viscous behaviour of CSH, which is the primary hydration product

of OPC. Advances in cement technology have enabled to understand the properties of CSH

recently. However, CSH exhibits significant local variations and is still difficult to probe into

its overall characteristics directly. Thus, a study on viscoelastic properties of hydration prod-

ucts by nanoindentation is well-suited, making it possible to probe sub-micrometric volumes of

material (Vandamme and Ulm, 2013).

In this section, therefore, it will present an investigation of indentation properties of hydration

products of OPC using deconvolution technique, micro poromechanics, and time-dependent

properties using nanoindentation as introduced in Chapter 3.

4.2.1 Sample Preparation

General purpose (Type I) Portland cement available locally in Australia was used to prepare

OPC paste sample for nanoindentation test. Three specimens were cast in 50mm cubic mould

with 0.3 of water to cement ratio, based on workability of mixture. Specimens were cured in

ambient curing at 23◦C ± 3. The compressive strength of specimens was obtained around 94.6

MPa. Each specimen was cut using a diamond saw after 28 days of curing to get the core

part of 10mm long segment. The hydration of specimens was arrested using solvent exchange

48

and oven dry method (Zhang and Scherer, 2011). Miller et al. (2008) pointed out that it was

important to reduce the surface roughness of specimen to get an accurate nanoindentation

results because the specimen’s surface has a significant influence on the test results (ISO 14577-

1, 2002). Therefore, fine emery paper was used to grind all specimens to reduce the surface

roughness. After that, the surface was also polished by using a suspension solution ranging from

6.0µm to 0.1µm for 45 mins and with 0.05µm for another 15mins; noting that the polishing

effect would last for around 8 hours. Nanoindentation test was then carried out on three sets

of 10mm cubic specimens with XP testing method; each set was specified 10 × 10 grid (100

indentations) with 20µm grid spacing, i.e., approximately 300 points were impressed on the

sample. Maximum load of Pmax =0.5mN applied in 15 seconds, kept constant 15 seconds and

unloaded in 10 seconds. The Berkovich indenter tip was used and Poisson’s ratio was assumed

0.25.

4.2.2 Statistical Indentation

The deconvolution technique presented in Section 3.5 can be used to analyse the mean properties

and volume fraction of each phase from a grid indentation. This process was carried out using

MATLAB (2014). The deconvolution technique process requires input parameters such as initial

values and the number of material phase. It has been well presented that hydration products

of OPC have five significant mechanical phases: capillary pores (MP), two CSH, which are

low-density (LD) CSH and high density (HD) CSH, Portlandite (CH), and unhydrated clinker

(Constantinides and Ulm, 2007; Velez et al., 2001).

The indentation modulus (M) and hardness (H) are determined using five Gaussian deconvolu-

tion input parameters as presented in Section 3.2. The results of deconvolution are illustrated

in Figure 4.1 and given in Table 4.3. The mean value of the first peak shows M = 9.388

GPa , H = 0.297 GPa; the second peak, M = 16.591 GPa, H = 0.7 GPa, which are in good

agreement with literatures (Constantinides and Ulm, 2007; Velez et al., 2001). For the third

peak, M = 29.818 GPa, H = 1.412 GPa; the fourth peak M = 48.677 GPa, H = 10.495 GPa;

and the clinker, M = 112.378 GPa, H = 14.718 GPa. Therefore, it can be identified that

the hydration products of OPC are made of five mechanical phases and a deconvolution gird

indentation can be performed with five Gaussian deconvolution input parameters. The typical

indentation load-depth (P − h) curves of indented material are presented in Figure 4.2.

As mentioned in Section 3.4, with the hypothesis that OPC is a porous material, the quadratic

minimisation problem, Equation (3.4.8), can be solved. The quadratic minimisation procedure

of scaling indentation modulus and hardness exhibit the properties of the hydration products.

The packing density relationship distribution can be plotted by experimental and model values

as shown in Figure 4.3. The relative error between linear, Equation (3.4.6), and non-linear,

Equation (3.4.5), the scaling model and the experimental values of the minimisation are: eM =

Table 4.3: Deconvolution results for indentation modulus and hardness

PhaseM H f

mean StD mean StD %

MP 9.388 3.562 0.297 0.116 43LD CSH 16.591 4.712 0.700 0.147 34HD CSH 29.818 3.984 1.412 0.215 10

CH 48.677 10.495 2.352 1.239 8Clinker 112.378 29.557 14.718 8.027 5

49

Figure 4.1: Deconvolution technique of indentation modulus and hardness

Figure 4.2: Typical indentation load-depth (P − h) curves

50

14.042 ± 13.215 GPa for indentation modulus and eH = 9.681 ± 9.065 GPa for indentation

hardness. The solid properties of hydration products are: stiffness (λ) = 189.442 GPa, Poisson’s

ratio (vs) = 0.499, cohesion (κ) = 1.300 GPa, friction coefficient (ζ) = 0.566 and the friction

angle is calculated from friction coefficient (ζ) as θ = 29.544 degree. The solid properties

of hydration products can be understanding in the sense of the Druker-Parger strength and

Coulomb material models (Ulm et al., 2007). The deconvolution technique using Equation

(3.5.3), the indentation modulus (M), hardness (H), packing density (η) and volume fraction

(f) are obtained. The details are presented in Table 4.4 and the deconvolution results of packing

density are given in Figure 4.4.

CSH is a major hydration product which relates directly to mechanical strength (Berry et al.,

Figure 4.3: Packing density (η) distribution

Table 4.4: Deconvolution results for indentation modulus and hardness

PhaseIndentation Modulus Indentation Hardenss Packing density Volume fraction

M H η f

mean StD mean StD mean StD %

MP 9.467 3.604 0.298 0.118 0.531 0.01 44LD CSH 16.787 4.804 0.704 0.144 0.556 0.01 33HD CSH 30.481 4.257 1.415 0.222 0.595 0.01 10

CH 51.449 9.831 2.579 1.170 0.627 0.06 7Clinker 114.539 31.598 14.387 7.221 0.780 0.12 5

Figure 4.4: Deconvolution result of packing density

51

1990; Neville, 2011). The lowest indentation modulus, hardness and packing density which can

be observed are associated with a capillary pores (MP) (Constantinides and Ulm, 2007). In the

present case, it was found that pores phase has the volume proportion of 44% of the total volume

fraction of the overall hydration products, and has the indentation modulus M = 9.467± 3.604

GPa, hardness H = 0.298 ± 0.118 GPa, and packing density η = 0.531 ± 0.01. LD CSH

phase has a characteristic mean stiffness of indentation modulus M = 16.787 ± 4.804 GPa,

hardness H = 0.704 ± 0.144 GPa, and packing density η = 0.556 ± 0.01, with 33% of volume

fraction. The stiffness of indentation modulus, hardness, packing density and volume fraction

of HD CSH are: M = 30.481 ± 2.441 GPa, H = 1.415 ± 0.222 GPa, and packing density

η = 0.595±0.01, with 10% of volume fraction. The Portlandite (CH) phase has the indentation

modulus M = 51.449 ± 9.831 GPa, hardness H = 2.579 ± 1.170 GPa, and packing density

η = 0.627± 0.06 with 7% of the total volume fraction. The indentation modulus, hardness and

packing density of clinker are M = 114.539 ± 31.598 GPa, hardness H = 14.387 ± 7.221 GPa,

and packing density η = 0.780 ± 0.12 with 5% of volume fraction. Application of Equation

(3.4.9) to the indented test data yields total porosity (ϕ) = 0.357. The statistical indentation

technique provides a new nonintrusive way of determining the porosity of nanogranular material

(Ulm et al., 2007).

The deconvolution technique results proved the mechanical properties at a point of the inden-

tation grid. The information of these nodal values provides a morphological arrangement of the

different phases in the composite material (Constantinides et al., 2006). The contour map of the

indentation modulus of the hydration phases (MP, LD CSH, HD CSH, and CH) was constructed

using MATLAB (2014) as illustrated in Figure 4.5. The information on the contour map of the

mechanical properties provides the microstructural distribution of each phase in the specimen

with respect to the x-y plane. Showing localised effect in a small area is a good technique to

display the volume fraction of different phases. The contour map shows that higher modulus

is surrounding by lower modulus, i.e., HD CSH phase is surrounding by LD CSH phase, CH

phase is surrounding by HD CSH phase, and unhydrated phase (clinker) is surrounding by CH

phase.

Figure 4.5: Contour map of hydration products of OPC paste. Image size is 180µm × 180µm with20µm grid spacing

52

4.2.3 Viscoelastic Properties

A classical principle of elastic properties deals with the mechanical properties of elastic solids in

accordance with Hooke’s law, i.e., stress is directly proportional to strain in small deformation

and independent of the rate of strain (Ferry, 1980). Material behaviour exhibiting viscoelastic

characteristics has ability to store energies under load-deformation upon removal of stress and

return to its original state. The measurement of viscoelastic properties of the material can be

obtained considering the nature and the rate of configurational rearrangements and interaction

of properties. Nanoindentation is a new technique that can determine quantitative viscoelastic

properties of the tested material. The time-dependent properties of the material are generally

obtained in term of relaxation and creep properties using indentation results.

4.2.3.1 Contact Relaxation Modulus

Classical linear viscoelastic properties, also known as rheological models, such as Maxwell model,

three-element Kelvin-Voigt model and combined four-element Maxwell-Kelvin-Voigt model can

be obtained from indentation tests. Rheological models are associated with a part on the

nanoindentation load-displacement (P − h) curves, i.e., in the holding period at the maximum

load. In the present case, the P − h curves from the nanoindentation tests were fitted to the

Equations (3.8.20) to (3.8.22) using a non-linear least squares method in MATLAB (2014).

Based on the results, the corrected coefficients of Maxwell, Kelvin-Voigt and Maxwell-Kelvin-

Voigt model were observed as 0.902±0.046, 0.936±0.242 and 0.987±0.028, respectively. Table

4.5 summarises the Rheological properties of OPC. The results show that standard deviation

is significant because of results included total hydration products of Rheological element.

The relaxation modulus is greatly significant for engineering applications. The normalised

contact relaxation modulus, which is defined as the relaxation modulus divided by the instan-

taneous modulus represents important material properties of the linear viscoelastic material

(Cao et al., 2010). The mean normalised contact relaxation modulus is shown in Figure 4.6.

The contact relaxation modulus based on Maxwell model shows a decrease in normalised re-

laxation modulus because Maxwell model is a linear function. The results of Kelvin-Voigt and

Combined Maxwell-Kevin-Voigt of contact normalised relaxation modulus indicate that the

initial reduction of relaxation modulus is small (around 3.5% in 10 second). Also, normalised

relaxation modulus can be represented by the dynamic modulus, as discussed in Chapter 3.

The results indicate that initial reduction of elastic modulus while constant strain. It means

that the stress relaxation and reduction of elastic modulus can be negligible during early curing

ages of OPC pastes.

4.2.3.2 Contact Creep Compliance

The contact creep compliance Jc was introduced in Section 3.8. Generally, the complex creep

behaviour of cementitious material is largely related to the viscoelastic response of the vital

hydration or reaction products and the binding phase of hardening. The creep compliance

depends on the geometry of the axisymmetric indenter and maximum value of control variables

Table 4.5: Rheological properties

Maxwell Kelvin-Voigt Maxwell-Kelvin-Voigt

Mo (GPa) 14.532± 33.186 14.977± 34.347 15.074± 34.662Mv (GPa) - 143.412± 356.395 254.983± 574.723ηM (GPa s) 1795.889± 4497.288 - 3475.169± 8979.073ηv (GPa) - 566.122± 1192.053 427.100± 835.500

53

Figure 4.6: Normalised relaxation modulus

in indentation process (Neville, 2011).

For the long-term contact creep compliance, the change in depth of the creep phase was fit with

a logarithmic function in Equation (3.8.33). The curve fitting was performed by MATLAB

(2014). The average corrected coefficient of the fitted curve was observed as 0.986± 0.04. The

typical curve fitting in penetration depth versus time is presented in Figure 4.7. The average

error of x1 coefficient was captured as 4.858± 2.455 nm.

With the logarithm curve fitting results, x1, the long-term contact creep compliance rate can

be obtained as:

Jc =1

Ctwhere C =

πPmax4 tan (a)hmaxx1

(4.2.1)

Figure 4.7: Typical logarithm curve fitting in creep phase

54

where C is contact compliance creep modulus. As explained in Section 3.8, Pmax, hmax and

α are obtained from indentation test and geometry properties. The long-term contact creep

modulus was determined as 409.266 GPa. Figure 4.8 shows the average long-term contact

creep compliance rate. The result of creep compliance rate clearly shows that after a few days

of applying stress, the rate of the creep compliance sharply decreases. The results of creep

behaviour are useful in quantifying a unique mechanical response of time-dependent materials.

From the results of contact creep compliance rate of OPC, specific creep after one year was

determined by 18.32 microstrain/MPa.

As presented earlier in Section 4.2.2, the deconvolution technique identified material phases as

MP, LD CSH, HD CSH, CH and un-hydrated clinker. In addition, this deconvolution technique

can identify the contact creep modulus C of each phase. The greater the C value represents the

lower rate of the creep. Table 4.6 summarises the results of M , H, C and η with 5 Gaussian

deconvolution input parameters. For MP phase: C = 74.552 ± 37.754 GPa; LD CSH phase:

C = 162.201 ± 53.582 GPa; HD CSH Phase: C = 333.797 ± 46.979 GPa; and CH phase:

C = 663.387±188.444 GPa. The CDF and PDF results of contact creep modulus are presented

in Figure 4.9. Figure 4.10 shows a comparison between contact creep compliance rates of all

the hydration phase. MP phase has the highest creep compliance rate of OPC because it has

a highest creep modulus. It means that when the porosity increases, it tends to increase the

creep compliance.

4.2.4 Indentation Stress-Strain Curve and Fracture Toughness

The relationship between the maximum depth h and the contact depth, hc = f (h), depends on

the indenter tip geometry and indented material. The function, hc = f (h), can be measured by

contact stiffness measurement (CSM) between the indented material and the indenter process

using Equation (3.9.2). Knowing hc = f (h) function, and the geometry of indenter tip, the

instant contact area can be determined by Equation (3.9.1) which is the instant contact area

with respect to the instant stress for loading process. Using Equations (3.9.3) to (3.9.5), the

indentation stress-strain curve can be obtained as presented in Section 4.2. The results of

the indentation stress-strain curves of statistical deconvolution phases show that each phase

Figure 4.8: Contact creep compliance rate

55

Table 4.6: Deconvolution results of indentation modulus, hardness, contact creep modulus andpacking density

PhaseIndentation Indentation Contact creep Packing Volume

Modulus Hardness Modulus Density fractionM H C η f

mean StD mean StD mean StD mean StD %

MP 9.292 3.509 0.298 0.114 75.552 37.754 0.531 0.011 42LD CSH 16.326 4.510 0.696 0.146 162.201 53.582 0.556 0.014 33HD CSH 29.028 4.056 1.391 0.232 333.797 46.979 0.585 0.041 10

CH 45.755 11.384 2.107 1.261 663.387 184.444 0.599 0.028 8Clinker 106.234 34.349 13.160 8.810 3948.108 1388.759 0.783 0.088 6

Table 4.7: Deconvolution results of indentation modulus, hardness, contact creep modulus andpacking density

PhaseIndentation Indentation Contact creep Packing Volume

Modulus Hardness Modulus Density fractionM H C η f

mean StD mean StD mean StD mean StD %

MP 9.292 3.509 0.298 0.114 75.552 37.754 0.531 0.011 42LD CSH 16.326 4.510 0.696 0.146 162.201 53.582 0.556 0.014 33HD CSH 29.028 4.056 1.391 0.232 333.797 46.979 0.585 0.041 10

CH 45.755 11.384 2.107 1.261 663.387 184.444 0.599 0.028 8Clinker 106.234 34.349 13.160 8.810 3948.108 1388.759 0.783 0.088 6

Figure 4.9: Deconvolution result of contact creep modulus

56

Figure 4.10: Contact creep compliance rate of hydration phases

can be captured based on initial and stationary yielding using Berkovich indenter tip (20nm

of the radius). The indentation stress-strain curves clearly illustrate elastic, elastic-plastic and

plastic regimes. The initial yielding points are observed approximately when the modulus values

are 70 MPa, 79 MPa, 200 MPa, and 425 MPa for MP, LD CSH, HD CSH, and CH phases,

respectively. The stationary yielding points from indentation stress-strain curves are obtained

when the strength values reach 354 MPa, 420 MPa, 580 MPa, and 1200 MPa for MP, LD CSH,

HD CSH, and CH phases, respectively. Based on the indentation stress-strain curves, MP phase

has small strain capacity in the fully plastic regime, as seen in Figure 4.11(a). The indentation

stress-strain curves of LD CSH and HD CSH phases, Figure 4.11(b) and (c), illustrate that

indentation strain hardening increases in the fully plastic regime and fracture point cannot

be observed. Figure 4.11(d) presents the indentation stress-strain curve of CH phase which

shows that it remains stable in the fully plastic regime. With the indentation stress-strain

curve of each phase of OPC paste, it can be explained that the strength failure occurred in MP

phase because of the small strain capacity compared to other phases. Therefore, high capillary

porosity will lead to low compressive strength. Chindaprasirt et al. (2005) reported that fine

fly ash (grinded) contents in mixture resulted in low capillary porosity and higher compressive

strength than the original fly ash (not grinded) contents in the mixture. Also, an increase in

sand to cementitious material ratio can increase capillary porosity and lead to low compressive

strength. The present finding validates theses points.

The fracture toughness of a material is linked to the energy stored in the form of elastic or

plastic energy before fracturing. As presented in Section 3.6, a robust method for determining

fracture toughness using Equations (3.6.3) to (3.6.9) was proposed. The overall indentation

results of the fracture energy release rate (Gc) and fracture toughness (Kc) were obtained as

0.526 N/m2 and 0.368 MPa m1/2, respectively. Shah (1989) determined fracture toughens of

cement pastes by conventional test method (dye penetration technique) using varying water to

cementitious material ratio. The obtained results in the present case are within the range of the

fracture toughness of cement pastes determined Shah (1989), which are 0.3 to 0.5 MPa m1/2.

57

Figure 4.11: Indentation stress-strain curves (a) MP (b) LD CSH (c) HD CSH (d) CH

4.2.5 Relationship between Indentation Properties

This section presents the results to illustrate the relationship between the indentation properties,

which are indentation modulus, hardness, packing density and volume fraction, as shown in

Figure 4.12. The indentation modulus and hardness deconvoluted data align well along the

curve. The fitting of power function to the indentation modulus and hardness relation yields

M = 22.83H0.904 with corrected coefficient 0.991 (99.1%). It was found that the relationship

between indentation modulus and hardness is almost linear, i.e., an increase in modulus leads

to an increase in the hardness. The relationship between indentation modulus and packing

density has fitting power function M = 1956(η − 0.5)1.659. The corrected coefficient is 0.971

(97.1%). This relationship shows that an increase in indentation modulus leads to an increase in

packing density. The deconvolution data of indentation modulus and contact creep modulus has

a relationship as M = 0.4357C0.7174 with corrected coefficient 0.992 (99.2%). In addition, the

contact creep modulus in an exclusive manner with the packing density. The fitting with power

function to the deconvoluted data yields C = 709224(η − 0.5)3.0463 with corrected coefficient

0.926 (92.6%). It was found that an increase in indentation modulus leads to a decrease in

creep modulus. The results of the relationship between indentation properties show power

relationship of the indentation modulus and creep modulus have a unique manner with the

packing density. An increase in packing density provides an increase in indentation modulus

and creep modulus. Moreover, the result of an increase in modulus and creep modulus delivers

an increase in hardness which provides high strength; and a decrease in creep compliance offers

low creep behaviour. Thus, the relationship between properties can be adapted for use as

a guideline in the design of OPC mixtures, i.e., an increase in packing density is the most

important objective in design of OPC mixtures became an increase in packing density provides

high strength and low creep behaviour in OPC mixtures.

58

59

Figure 4.12: Deconvoluted indentation properties with Power fit

60

4.3 Statistical Analysis of Properties of Blended Cement

Blended cement is a hydraulic cement that contains general purpose cement and pozzolan such

as fly ash and ground granulated iron blast-furnace slag (GGBF) (Australia Standard, 2010).

Blended cement has been used because of its low cost and improved properties. Fly ash is

commonly used blended cement that the main improvement to the strength of blended cement

is due to reaction the between calcium hydroxide (Ca(OH)2) which is a by-product of hydration

of OPC. When blended cement mixes with water, fly ash reacts with water and Ca(OH)2 to

form Calcium Silicate Hydrate (CSH) (Guide to the use of fly ash in concrete in Australia, 2009;

Shi, 1996). This section presents a study on blended cement mixtures using statistical analysis

such as Taguchi’s Design of Experiment, Analysis of Variance (ANOVA), and regression to

investigate statistical relationship properties of blended cement with different parameters.

4.3.1 Experimental Program

Taguchi’s design experimental approach to parameters design provides design engineers with a

systematic and efficient method for investigating optimum design parameters for performance

and cost. In this research, the following parameters are considered in the mix proportions.

- Fly ash (FA)

- Sand to cementitious material ratio (s/c)

- Water to cementitious material ratio (w/c)

- Superplasticiser (SP)

Four parameters and three levels of test variable were selected as shown in Table 4.8. The

Standard L9 (34) orthogonal array (Ross, 1996) according to the selected parameters was used,

as presented in Table 4.9.

Table 4.8: Variation parameters and levels

Levels FA s/c w/c SP

1 0 % 0 0.3 0 %2 10% 1.5 0.35 0.1%3 20% 2.5 0.4 0.2%

Table 4.9: Standard L9 orthogonal array

Mix FA s/c w/c SP

1 1 1 1 12 1 2 2 23 1 3 3 34 2 1 2 35 2 2 3 16 2 3 1 27 3 1 3 28 3 2 1 39 3 3 2 1

In this experimental work, general purpose (Type I) Portland cement, and Class F (low calcium)

fly ash from Australia were used to prepare specimens. Table 4.10 shows the typical chemical

compositions of OPC and fly ash. The Loss on Ignition (LOI) of fly ash for the experiment

was 1.53% and the median particle size was 45µm. The sand (SSD: saturated surface dried

61

condition) with a specific gravity of 2.6 was used as fine aggregate. The XRD pattern of OPC

and fly ash shown in Figure 4.13. The specimens were cast in 50mm cubic mould. Blend cement

mixtures were cured in ambient curing at 23◦C ± 3 until testing. The compressive strength of

the specimens were tested at a loading rate of 900N/s with a universal testing machine at the

age of 28 days in accordance with ASTM ASTM Standard C109 (2011). Also, the density of

the specimen was measured at the age of 28 days in accordance with ASTM ASTM Standard

C642 (2013). A naphthalene sulphonate superplasticiser was used to improve workability. The

reported results are the average of five samples.

Table 4.10: Chemical composition of OPC (Type I) and low calcium fly ash (wt. %)

SiO2 Al2O3 CaO Fe2O3 K2O MgO SO3

OPC 21.1 4.7 63.6 2.7 - 2.6 2.5Fly ash 65.9 24.0 1.59 2.87 1.44 - -

Figure 4.13: XRD pattern of (a) OPC (Type I) and (b) low calcium fly ash

62

4.3.2 Density

Density is important in reducing the self-weight. Figure 4.14 and Table 4.11 show the average

density results of the tested blended cement mixtures. The results indicate that most of the

mixtures have fairly uniform density during the 7 to 28 days of curing ages. The paste samples

with fly ash, Mixes 4 and 7, have the lowest density of 1929 kg/m3 and 1776 kg/m3, respectively.

The mortar samples without fly ash, Mixes 2 and 3, have the highest density of 2164 kg/m3

and 2145 kg/m3, respectively. Overall, the density of all the mixtures slightly decreases after

curing for 28 days, with an exception that the mortar samples, Mixes 8 and 9, which contained

20% of fly ash to cement ratio, have a slight increase in the density during 7 to 14 days of

curing age. Based on these results, a statistical S/N ratio analysis was performed to determine

the effect of these parameters on the density of the mixtures, as shown in Figure 4.15, which

can be seen that sand to cementitious material ratio has the most effect on the density at 28

days of curing age. ANOVA was performed on the results from the nine mixtures. ANOVA

results indicate that sand to cementitious material ratio contributed 54% to the density at 28

days of curing age as presented in Table 4.12 and Figure 4.16. It was observed that increasing

fly ash contents resulted in decreasing the density. The contribution of fly ash on the density is

determined as 24%. The contribution of increasing the water to cementitious material ratio on

the density is 14%. Superplasticiser has a minor influence on the density at 28 days of curing

age; its contribution is obtained as 5%. As expected, the sand to cementitious material ratio

was the most important parameter on density.

Figure 4.14: Density with curing age

63

Table 4.11: Density with curing age results (kg/m3)

Mix 7 days 14 days 28 days

1 2082 2073 20702 2164 2153 21463 2145 2132 21264 1929 1934 19275 2056 2030 20276 2115 2111 21107 1776 1773 17628 2090 2111 20779 2118 2126 2091

Figure 4.15: Density with curing age

Table 4.12: ANOVA results on density

Source DFa SSb MSc Contribution %

7 days density

FA 2 29447 14724 23.84s/c 2 69882 34941 56.57w/c 2 17432 8716 14.11SP 2 6773 3386 5.48

14 days density

FA 2 22893 11477 18.86s/c 2 68161 34081 56.15w/c 2 23721 11861 19.54SP 2 6626 3313 5.49

28 days density

FA 2 29462 14731 24.82s/c 2 63513 31756 53.51w/c 2 20772 10386 17.50SP 2 4954 2477 4.17

adegree of freedom bsum of square cmean square

64

Figure 4.16: Contribution of experimental parameters on density

65

4.3.3 Compressive Strength

The compressive strength results are presented in Table 4.13 and Figure 4.17(a). The average

percentage increase in compressive strength at 7 days of curing age is 8% of 28 days compressive

strength. The paste mixture Mix 1 has the minimum compressive strength 87.8 and 94.5 MPa

at 7 days and 28 days of curing age, respectively. The mortar mixture Mix 5, with 10% of

fly ash, 0.15 of sand to cementitious material ratio, 0.4 of water to cementitious material ratio

and no superplasticiser has the minimum compressive strength of 43.6 MPa and 54.5 MPa at

7 days and 28 days of curing age, respectively. Figure 4.17(b) shows the compressive strength

development between 7 days to 28 days of curing age. Mix 4 has the highest compressive

strength gain between 7 to 28 days of curing age with 21.62 MPa. The compressive strength

gain in Mix 6 between 7 to 28 days is the lowest, which is 4.31 MPa. Therefore, Mix 5 has

minimum compressive strength but the highest compressive strength gains between 7 to 28 days

of curing ages.

Table 4.13: The results in based on compressive strength (MPa)

Mix7 days 14 days 28 days

fck,cubic StD fck,cubic StD fck,cubic StD

1 87.82 2.44 91.37 1.74 94.56 2.262 63.54 1.56 67.42 3.52 72.82 4.353 53.22 1.61 58.78 2.64 62.63 2.864 59.59 5.68 73.34 4.65 81.20 3.455 43.65 3.44 51.61 2.40 54.55 2.796 52.01 1.36 55.22 2.27 56.32 1.847 51.39 3.16 58.62 2.62 58.79 3.068 54.76 2.92 65.05 4.11 67.14 6.999 47.61 2.09 58.44 2.15 59.53 3.51

66

Figure 4.17: (a) compressive strength development and (b) % of compressive strength developmentbetween 7 to 28 days

67

The typical XRD pattern of OPC paste (Mix 1) is as presented in Figure 4.18. Highly notice-

able peaks are “Calcium Silicate Hydrated (CSH)”, which is main hydration products of the

cementing compound, “Portlandite (CH)” and “Dicalcium Silicate (C2S)”. Due to curing ages,

XRD pattern indicates that CSH phase increases. Scanning Electron Microscope (SEM) also

shows the growth of hydration products as fabric structures when curing proceeds as presented

in Figure 4.19.

Figure 4.18: XRD pattern of Mix 1

Figure 4.19: SEM image of Mix 1

68

The S/N ratio graphs of parameters for compressive strength at 7 days and 28 days of curing

age are shown in Figure 4.20. Fly ash has the most significant effect on 7 days compressive

strength. On 28 days compressive strength, however, sand to cementitious material ratio is the

most affecting parameter. ANOVA results show sand to cementitious material ratio has 41.56%

contribution on 7 days and 39.59% contribution on 28 days, as shown in Table 4.14 and Figure

4.21. The effect of superplasticiser on 7 days and 28 days compressive strength is 2.32% and

7.64% of contribution, respectively. The S/N ratio of fly ash and sand to cementitious material

ratio at 7 days and 28 days of curing age results show similar patterns. The S/N ratio of water

to cementitious material ratio shows that water to cementitious material ratio of 0.3 to 0.35 has

not significant effect on 28 days compressive strength. ANOVA results of sample containing

water to cementitious material ratio present mostly similar contributions of curing ages which

are in range 25.32% and 27.49%. Therefore, an increase in fly ash, sand to cementitious material

ratio and water to cementitious material ratio in blended cement mixtures lead to a decrease

in the compressive strength.

Table 4.14: ANOVA results on compressive strength

Source DFa SSb MSc Contribution %

7 dayscompressivestrength

FA 2 557.51 278.75 41.56s/c 2 394.68 197.34 29.42w/c 2 358.14 179.07 26.70SP 2 31.18 15.59 2.32

14 dayscompressivestrength

FA 2 295.47 474.02 25.36s/c 2 474.02 237.01 40.69w/c 2 3 20.24 160.12 27.49SP 2 75.28 37.64 6.46

28 dayscompressivestrength

FA 2 385.4 192.69 27.45s/c 2 555.9 277.95 39.59w/c 2 355.6 177.78 25.32SP 2 107.3 53.67 7.64

adegree of freedom bsum of square cmean square

69

Figure 4.20: Effect of parameters on compressive strength

70

Figure 4.21: Contribution of experimental parameters on compressive strength

71

The results of S/N on compressive strength gain between 7 days to 28 days of curing ages

indicate that superplasticiser is the most significant parameter as shown in Figure 4.22. Sand

to cementitious material ratio has a minor effect on the compressive strength gain. ANOVA

results reveal that the contribution of superplasticiser on compressive strength gain is 44.01%

and sand to cementitious material ratio is 9.19%, as shown in Table 4.15. The optimal mixing

design for compressive strength gain on blended cement mixtures is 20% of fly ash, 1.5 of

sand to cementitious material ratio, 0.35 of water to cementitious material ratio and 0.2% of

superplasticiser.

Figure 4.22: Effect of parameters on compressive strength gain

Table 4.15: ANOVA results on density

Source DFa SSb MSc Contribution %

Compressivestrengthgain

FA 2 21.7 10.85 11.19s/c 2 17.83 8.915 9.19w/c 2 69.09 34.546 35.61SP 2 85.37 42.687 44.01

adegree of freedom bsum of square cmean square

72

Figure 4.23: Contribution of experimental parameters on compressive strength gain

4.3.4 Water Absorption

Low permeability, especially resistance to freezing and thawing is one of the most important

properties of the good quality of cement binder. The permeability of cement binder differs

from absorption and relates to the size of pores, its distribution, and most importantly its

continuity. Therefore, good quality of binder cement has low water absorption and high density

(Neville, 2011). According to ASTM ASTM Standard C140 (2015), water absorption rate of

each specimen was determined. The results of water absorption test indicate that Mix 7 has

the highest water absorption, which is 18.02%. Mix 3 has the lowest water absorption rate of

7.45%, as shown in Figure 4.24.

Figure 4.24: Water absorption results

73

The results of S/N ratio on water absorption show that sand to cementitious material ratio

has a major effect on the water absorption, as shown in Figure 4.25. The increase of sand to

cementitious material ratio leads to the decrease of water absorption on blended cement mixture.

Superplasticiser is a minor important parameter on water absorption. ANOVA results show

that the contribution of sand to cementitious material ratio is 85.44% and superplasticiser is

1.92%, as shown in Table 4.16. Thus, an increase of fine solid such as fly ash and OPC could

lead to increase in water absorption. Moreover, high contents of superplasticiser also decrease

the water absorption. The optimised mixing proportion for water absorption is: no fly ash, 2.5

of sand to cementitious material ratio, 0.3 of water to cementitious material ratio and 0.2% of

superplasticiser.

Figure 4.25: Effect of parameters on water absorption results

Table 4.16: ANOVA results on density

Source DFa SSb MSc Contribution %

Waterabsorption

FA 2 5.717 2.859 5.15s/c 2 94.8 47.4 85.44w/c 2 8.312 4.156 7.49SP 2 2.126 1.063 1.92

adegree of freedom bsum of square cmean square

74

Figure 4.26: Contribution of experimental parameters on water absorption

4.3.5 Regression Analysis

Regression analysis is a statistical tool for estimation of relationships between variables. One

of important reasons to determine the regression model is to uncover causes by studying the

relationship between variables. Sometimes, statistical relationship does not necessarily imply

causal relationship but the presence of relationship can give a good starting point for research.

If statistical confidence is indicated with the regression model, then values of the explanatory

variables can be used predict the output variables. Another reason for regression analysis is

to examine the test hypotheses (Seber and Lee, 2012). In this research, therefore, regression

analysis was used as a tool for estimating the relationship between variables and to predict

outside of parameters ranges of explanatory variables.

Through linear regression analysis, the empirical relationship of blended cement mixtures at 28

days curing ages was obtained as:

ρ28 = 1747 + 5478.9x1 + 44.308x2 + 1086.8x3 − 43947x4

+ 209.85x1x2 − 18486x1x3 − 73.835x2x3 (4.3.1a)

fck,cubic,28 = 199.11− 745.64x1 − 49.188x2 − 347.35x3 + 9736.9x4

+ 79.579x1x2 + 1716.3x1x3 + 106.24x2x3 (4.3.1b)

where x1, x2, x3, x4 denote fly ash, sand to cementitious material ratio, water to cementitious

material ratio and superplasticiser, respectively. As shown in Table 4.17, the empirical equation

is in good agreement with the experimental results.

Validation of predictive model can be carried out with corrected coefficient (Analla, 1998). The

information about regression analysis is presented in Table 4.18. It can be seen that corrected

coefficient values of density and compressive at 28 days are 0.980 (98%) and 0.983 (98.3%),

respectively. These corrected coefficients indicate that empirical model states the good fit of

the model being validated.

Several researchers (Ozbay et al., 2009; Srinivasan et al., 2003; Tanyildizi, 2013; Xu et al., 2012)

reported successful application of Taguchi’s design of experiment with ANOVA technique to

75

Table 4.17: Experiment and Predicted density (kg/m3) and compressive strength (MPa) at 28 days

MixExperiment Predicted Ratio

ρexp fck,cubic,exp ρpred fck,cubic,predρexp fck,cubic,exp

/ρpred /fck,cubic,pred

1 2070 94.56 2073 94.90 0.998 0.9962 2146 72.82 2112 69.40 1.016 1.0493 2126 62.63 2131 63.14 0.998 0.9924 1927 81.20 1941 82.57 0.993 0.9835 2027 54.55 2044 56.26 0.992 0.9706 2110 56.32 2131 58.37 0.990 0.9657 1762 58.79 1755 58.11 1.004 1.0128 2077 67.14 2068 66.28 1.004 1.0139 2091 59.53 2081 58.50 1.005 1.018

Average 1.000 1.000StD 0.008 0.026

Table 4.18: Regression analysis of test data

(a) Density at 28 days

Parameters Estimate Standard error T value

Intercept 1747.4 831.2 2.102FA 5748.9 5737.1 0.995s/c 44.308 270.45 0.164w/c 1086.8 2697.2 0.403SP -43947 62285 -0.706

FA×s/c 209.85 614.53 0.341FA×w/c -18486 16038 -1.523s/c×SP -73.835 909.15 -0.081

Corrected coefficient = 0.980

(b) Compressive strength at 28 days

Parameters Estimate Standard error T value

Intercept 199.11 82.509 2.413FA -745.64 569.49 -1.309s/c -49.118 26.846 -1.829w/c -347.35 267.73 -1.297SP 9763.9 6182.7 1.579

FA×s/c 79.579 61.002 1.305FA×w/c 1716.3 1592.1 1.078s/c×SP 106.24 90.247 1.177

Corrected coefficient = 0.983

76

cementitious-based binders. In this research, Taguchi’s design of experiment results are in good

agreement with the results in literature (Bye, 1999; Neville, 2011; Peiwei et al., 2001; Siddique,

2004), the increase in fly ash, sand to cementitious material ratio, water to cementitious ma-

terial ratio reduces the compressive strength. Thus, Taguchi’s design of experiment leads to

identifying the optimal mixing design with S/N ratio. ANOVA technique helps to determine the

parameters contribution in a measured variation of properties that indicate most considerable

parameters on mixing design.

4.3.6 High Temperature Exposure

Concrete can expose to high temperature during a fire or when it is near to furnace and

power reactions. During its exposure to high temperatures, the mechanical properties, such

as strength and elastic modulus, deteriorate and decrease significantly (Morsy et al., 2009).

Physical deterioration process affects the durability of the concrete structure, therefore, is an

important issue during and after high temperature exposure. The harmful effects of high

temperature on concrete can be minimised by taking a preventive measure such as choosing the

right material (Aydın, 2008). OPC is a common binder in concrete. However, when exposed

to high temperatures, spalling of OPC-binder concrete could occur, resulting in a rapid loss

of concrete cover, layer by layer, potentially exposing the main steel reinforcement. Therefore,

it is beneficial to improve the property of cement binder that carries good resistance to high

temperature in term of spalling resistance and strength loss (Kong and Sanjayan, 2010).

Chan et al. (1996) studied high strength and normal strength concrete exposed to high tem-

peratures in a range between 400 to 1200◦C. They found that temperature between 400◦C

and 800◦C was a critical range of a loss of concrete strength. The normal strength concrete

loses around 10 to 25% and 80% of its original compressive strength at 400◦C and 800◦C, re-

spectively. Many research papers reported that using pozzolan as partial cement replacement

could lead to an improvement in temperature resistant properties. Xu et al. (2001) studied the

influence of high temperature on fly ash in which OPC was replaced with 0%, 25% and 55%

of fly ash. The residual strength of concrete samples after exposing to 250◦C to 800◦C were

determined and found that a dosage of 55% fly ash in concrete led to high residual strength

than other dosages fly ash samples. Poon et al. (2001) conducted research to investigate the

effect of the residual strength and durability on normal and high pozzolanic concrete exposed to

high temperatures. They concluded that pozzolanic concrete has better performance than OPC

concrete after exposed to high temperatures. The mix containing 30% of fly ash replacement

has maximum relative residual strength. Also, most of spalling occurred between 400◦C and

600◦C but no spalling was observed in fly ash concrete. Mendes et al. (2008) used slag at 35%,

50% and 65% and determined the residual strength after exposing to 100◦C to 800◦C. It was

found that the samples with slag showed less damage under elevated temperatures. Although

numerous researchers studied the effect of high temperatures exposure on the properties of

blended cement mixtures, there is not enough on parameters contribution.

High temperature exposure test was conducted on selected five specimens of each mixtures as

shown in Table 4.9, after 28 days curing then dried in an oven at 105 ± 5◦C for 24 hours. After

drying, the specimens were placed in an electric kiln to be heated up to reaching maximum

temperatures of either 200, 400, 600 or 800◦C with 2 hours holding time. The heating rate

was 10◦C per minute. After exposed to temperature, the specimens were allowed to cool down

gradually to room temperature. Table 4.19 and Figure 4.27 present the residual strength of

blended cement mixture after exposure to high temperatures. The test results show that each

temperature range has a distinct pattern of strength gain or loss. It is observed that there is

77

a significant increase in strength around 18.3 MPa when mixtures are exposed to 200◦C. This

increase could be due to the hydration of unhydrated particles which were activated as a results

of temperature rise (Morsy et al., 2008). The maximum and minimum residual strength of Mix

1 and Mix 5 are 112.8 and 56.8 MPa, respectively, when exposed to 200◦C. With Mix 1, a

significant decrease in the strength is obtained when exposed to between 600 to 800◦C.

Table 4.19: Residual strength of blended cement mixtures (Unit in MPa)

MIX23◦ 200◦ 400◦ 600◦ 800◦

fck,cubic StD fck,cubic StD fck,cubic StD fck,cubic StD fck,cubic StD

1 94.5 2.2 112.8 5.7 97.5 3.3 57.0 4.9 - -2 72.8 4.3 81.3 3.6 66.2 2.7 44.2 5.8 25.7 2.73 62.6 2.8 66.7 2.5 49.9 4.6 38.7 2.7 21.3 1.94 81.2 3.4 86.3 3.7 76.1 5.9 48.1 7.6 31.7 5.15 54.5 2.7 56.8 3.4 42.4 1.6 28.7 3.4 16.4 1.46 56.3 1.8 61.9 5.9 49.2 6.7 36.2 5.2 19.9 3.37 58.7 3.0 67.2 1.9 64.4 2.7 35.7 4.3 24.8 1.48 67.1 6.9 84.9 4.3 71.8 4.1 51.1 3.3 29.6 2.79 59.5 3.5 66.0 3.1 52.7 6.0 35.4 3.3 20.3 2.4

Figure 4.27: Residual strength

78

European Standard (de Normalisation, 2005) provided residual strength reduction factors of

normal weight concrete at elevated temperatures, as shown in Figure 4.28. The results of resid-

ual strength reduction factors illustrate that experimental results have high residual strength

reduction factors compared to European Standard. However, the overall trend of residual

strength is the similar in a range of 1.12 to 0.33 exposed to between 600 to 800◦C.

The XRD pattern of Mix 1 is shown in Figure 4.29, where main peaks have been iden-

tified. Typical peaks are associated to Portlandite (Ca(OH)2), Calcium Silicate Hydrated

(Ca1.5·SiO3.5·xH2O) and Dicalcium Silicate(Ca2SiO4). After Mix 1 was exposed to high tem-

peratures, some of these peaks disappeared or reduced in the intensity. It is clear that in

samples exposed to 600◦C compared to 23◦C, the main hydration products such as Portlandite,

CSH disappeared. However, between 23◦C to 400◦C high temperatures, the intensity of main

peaks in XRD patterns are not clear. The reduction and dehydration of CSH can explain CSH

hydration product reform to Calcium Oxide (CaO) but less proportion (Alonso and Fernandez,

2004; Handoo et al., 2002; Xu et al., 2001). The SEM images of Mix 1 show that there is a

limited amount of fabric structures in the mixture after exposed to 600◦C compared to the

image of the mixture at 23◦C, as shown in Figure 4.30.

Figure 4.28: Residual strength reduction factors (adopted after (de Normalisation, 2005))

79

Figure 4.29: XRD pattern of Mix 1

Figure 4.30: SEM image of Mix 1 after exposed to high temperatures

80

The S/N ratio is analysed for residual strength at 200◦C, 400◦C, 600◦C and 800◦C and plotted

in Figure 4.31. It is observed that superplasticiser has the most effect on the residual strength

of blended cement after high temperature exposure. From ANOVA results, the contribution

of superplasticiser on the residual strength is determined to be 54.90%, as shown in Table

4.20. Sand to cementitious material ratio has the minor effect on residual strength with 5.79%

contribution. Fly ash is the second significant influence on the residual strength with 20.07%

contribution. The contribution of water to cementitious material ratio is determined as 19.24%.

Generally, the increase of water to cementitious material ratio decreases the residual strength.

This is well known as the release of water vapour could influence the chemical composition of

hydration products known as dehydration process (Handoo et al., 2002).

Figure 4.31: Effect of parameters after exposed to high temperatures

Table 4.20: ANOVA results on high temperatures exposure

Source DFa SSb MSc Contribution %

200◦Cresidualstrength

FA 2 566.3 283.14 23.59s/c 2 871 435.49 36.29w/c 2 808 404.01 33.66SP 2 155.1 77.54 6.46

400◦Cresidualstrength

FA 2 153.09 76.55 9.71s/c 2 916.17 458.09 58.10w/c 2 418.82 209.41 26.56SP 2 88.93 44.47 5.64

600◦Cresidualstrength

FA 2 124.79 62.39 19.06s/c 2 155.13 77.57 23.70w/c 2 287.23 143.62 43.87SP 2 87.54 43.77 13.37

800◦Cresidualstrength

FA 2 139.01 69.5 20.07s/c 2 40.08 20.04 5.79w/c 2 133.29 66.64 19.24SP 2 380.31 190.16 54.90

adegree of freedom bsum of square cmean square

81

However, it was found that reducing water to cementitious material ratio would reduce the

overall residual strength, especially after exposed to 800◦C high temperature. Between the

exposed temperature ranges 23◦C to 600◦C, reducing water to cementitious material ratio

would increase the residual strength but after exposed to 800◦C would decrease the residual

strength. Therefore,based on overall residual strength S/N ratio result of the residual strength,

the optimal mixing design of blended cement mixture is 20% of fly ash,1.5 of sand to cementitious

material ratio, 0.35 of water to cementitious material ratio, and 0.2% of superplasticiser. The

optimal mixing design can be determined with the predicted S/N ratio function as (Ross, 1996):

ηpredict = η +

f∑i=1

ηi − ηo (4.3.2)

where η = overall mean of S/N ratio, f = number of a factor, and ηi = the mean of S/N

ratio at the optimal level of each factor. The S/N ratio of optimal mix design and maximum

and minimum parameter levels are determined as presented in Table 4.21. With the predicted

S/N ratio, the residual strength is calculated by Equation (4.3.2), as shown in Figure 4.32.

According to this Figure, the residual strength of the minimum parameter levels (P2) has high

residual strength between 23◦C to 600◦C. After exposed to 800◦C, however, significant loss

in residual strength is indicated. Regression analysis was used to estimate the relationships

between residual strength and parameters. Through a linear regression analysis, the empirical

relationship obtained can be written as:

R(T ) = C0 + C1x1 + C2x2 + C3x3 + C4x4 + C5x1x2 + C6x1x3 + C7x2x3 (4.3.3)

where x1, x2, x3, x4 denote fly ash, sand to cementitious material ratio, water to cementitious

material ratio and Superplasticiser, respectively. The information about regression analysis is

presented in Table 4.23. It can be seen that corrected coefficient values of residual compressive

strength at 200◦C, 400◦C, 600◦C and 800◦C are 0.957 (95.7%), 0.955 (95.5%), 0.952 (95.2%)

and 0.952 (95.2%), respectively. These corrected coefficients indicate that the empirical model

represents the good fit of the model being validated.

Table 4.21: Predicted S/N ratio

Index FA s/c w/c SP S/N ratio

P1 Optimal 20% 1.5 0.35 0.2% 75.97P2 Min Level 0% 0 0.3 0% -93.01P3 Max Level 20% 2.5 0.4 0.2% 73.14

Table 4.22: Coefficient of empirical relationship in Equation (4.3.3)

CoefficientTemperature (T )

200◦C 400◦C 600◦C 800◦C

C0 325.51 329.07 165.98 -180.6C1 -475.71 -767.82 -153.39 270.16C2 -90.575 -85.594 -44.622 63.617C3 -706.61 -769.5 -361.94 603.37C4 4435.1 6791.2 4038.6 9894.7C5 -8.719 -15.832 -9.063 107.97C6 1420 2355.7 540.29 -1258.6C7 242.49 230.12 121.57 -217.66

82

Table 4.23: Regression analysis of residual compressive strength

(a) 200◦C

Parameters Estimate Standard error T value

Intercept 325.51 174.34 1.867FA -475.71 1203.3 -0.395s/c -90.575 56.725 -1.597w/c -706.61 565.72 -1.249SP 4435.1 13064 0.340

FA×s/c -8.719 128.89 -0.068FA×w/c 1420 3364 0.422s/c×SP 242.49 190.69 1.272

Corrected coefficient = 0.957

(b) 400◦C

Parameters Estimate Standard error T value

Intercept 329.07 174.64 1.884FA -767.82 1205.4 -0.637s/c -85.594 56.824 -1.506w/c -769.5 566.7 -1.358SP 6791.2 13087 0.519

FA×s/c -15.832 129.12 -0.123FA×w/c 2355.7 3369.8 0.699s/c×SP 230.12 191.02 1.205

Corrected coefficient = 0.955

(c) 600◦C

Parameters Estimate Standard error T value

Intercept 165.98 95.754 1.733FA -153.39 660.91 -0.232s/c -44.622 31.156 -1.432w/c -361.94 310.71 -1.165SP 4038.6 7175.2 0.563

FA×s/c -9.063 70.794 -0.128FA×w/c 540.29 1847.6 0.273s/c×SP 121.57 104.73 1.161

Corrected coefficient = 0.952

(d) 800◦C

Parameters Estimate Standard error T value

Intercept -180.6 98.353 -1.836FA 270.16 678.85 0.398s/c 63.617 32.001 1.988w/c 603.37 319.15 1.891SP 9894.7 7370 1.343

FA×s/c 170.97 72.716 1.485FA×w/c -1258.6 1897.8 -0.663s/c×SP -217.66 107.58 -2.023

Corrected coefficient = 0.952

83

Figure 4.32: Predicted residual strength

The main purpose of this section is to identify affecting parameters and optimal mix propor-

tion of blended cement for high temperature exposure. The proposed optimal mix proportion

when designed for high temperature exposure is: 20% of fly ash, 1.5 of sand to cementitious

material ratio, 0.35 of water to cementitious material ratio and 0.2% of superplasticiser. In or-

der to analyse the linear regression, the empirical relationship of residual compressive strength

was obtained. The proposed empirical models are generally higher than uniaxial compressive

strength (de Normalisation, 2005). The correct empirical cylindrical compressive strength can

be determined approximately by (de Normalisation, 2005):

fck = 0.4381 · f1.135ck,cubic + 3.123 (4.3.4)

where fck is a residual uniaxial cylindrical compressive strength, fck,cubic is a cubic compressive

strength.

4.4 Chapter Summary

This Chapter presents indentation properties of OPC and statistical analysis of properties of

blended cement. Properties of hydration products of OPC were determined using nanoinden-

tation. Deconvolution technique was used to determined indentation modulus (M), hardness

(H) and packing density (η) of hydration phases. M , H and η of LD CSH are 16.787 ± 4.804

GPa, 0.704± 0.144 GPa and 0.556± 0.01, respectively. The HD CSH has M = 30.481± 4.257

GPa, H = 1.415±0.222 GPa, and η = 0.595±0.001. Indentation fracture toughness and stress-

strain curves were generated from indentation measurement. Nanoindentation was also a useful

method for investigating the mechanical behaviour of OPC such as elastic, elastic-plastic and

plastic deformation regimes on each phase. Contact relaxation modulus and creep compliance

were also analysed. Using Rheological models with indentation test, the results of contact relax-

ation modulus indicated that stresses, under constant condition, reduce insignificantly during

the initial period within 5 second only. It can therefore be concluded that stress relaxation in

84

OPC paste can be negligible. The obtained creep compliance rate shows that MP phase is the

main phase that increases the creep compliance rate of the matrix. It means that the porosity

tends to increase creep compliance. Also, long-term creep compliance rate results show useful

metric that quantifies a unique mechanical response of creep behaviour. The strength failure

can occur in MP phase because MP phase has small strain capacity compared to another phase,

thus, high capillary porosity exhibits low compressive strength. Thus, porosity is an important

factor to be considered when designing OPC mixtures in order to optimise the compressive

strength and minimise the creep behaviour.

Based on the results of the investigation conducted on blended cement mixtures using Taguhi’s

design of experiment, Taguchi’s design of experiment and ANOVA results shows an increase in

fly ash content and water to cementitious material ratio leads to a decrease in the density of

blended cement. Superplasticiser has a minor effect on the density of blended cement mixtures.

The optimal mix design was obtained as: 20% fly ash, no content of sand to cementitious

material ratio, 0.4 of water to cementitious material ratio, and 0.1% of superplasticiser. For

compressive strength development, an increase in fly ash content and sand to cementitious

material ratio decreased the compressive strength development. The optimisation of the com-

pressive strength of blended cement mixtures was found to be 20% of fly ash content, 1.5 of

sand to cementitious material ratio, 0.35 of water to cementitious material ratio and 0.2% of su-

perplasticiser. XRD and SEM analysis confirms that the main hydration product of OPC paste

is CSH phase, which increases with curing ages. SEM image also show growth of hydration

products after curing processes. With respect to resistance to high temperature, superplasti-

ciser was identified as the most significant parameter and fly ash was the second most effect on

mixtures. Increasing in fly ash and superplasticiser was found to improve the overall residual

strength, and the optimisation of mix design was 20% of fly ash, 1.5 of sand to cementitious

material ratio, 0.35 of water to cementitious material ratio, and 0.2% of superplasticiser; XRD

and SEM analysis results also confirmed the dehydration process.

This study determined that identifying the characteristics of civil engineering materials. Taguchi’s

design of experiment approach was successfully applied as a useful tool in studying the influ-

ence of parameters in cementitious matrices. The results can be analysed using the ANOVA

technique to examine the variation in the measured properties of blended cement. Moreover,

the potential impact of indentation approach will encourage consideration of small scales exam-

ination to represent the large scale testing of civil engineering structural elements. In addition,

indentation application in civil engineering will enable formulation of composite materials such

as high-performance fibre reinforced cementitious composite.

85

Chapter 5

Properties of Alkali-Activated

Cement

5.1 Introduction

Alkali-activated cement (AAC) is a potential cementitious system for sustainable development

(Caijun and Della, 2006). Its main constituent is pozzolan, which can react with alkali activator

to form binder (Shi et al., 2011). The composition of cementitious components and alkali-

activated pozzolan cement are classified according to the type of pozzolans such as fly ash,

metakaolin, soda lime glass and natural pozzolan (Caijun and Della, 2006). A number of

researchers (Pacheco-Torgal et al., 2008; Roy, 1999; Shi et al., 2011) reported the difference

between the composition of traditional Portland cement and fundamental rock-forming minerals

of the earth crust. The common chemical compositions of Portland cement and fundamental

rocks are silica and alumina which can also be found in a number of industrial wastes.

Ground industrial wastes or by-products containing aluminosilicate when mixed with rich alkalis

could form a hydraulic binder which is a type of inorganic polymer called “geopolymer” or alkali-

activated cement (AAC). Davidovits (1994a) classified different types of geopolymer according

to Si to Al ratio in the mixtures for various industry applications. Lloyd et al. (2010) reported

that calcium content in geopolymer is important for alkali mobility that may be significant to

limit the durability of embedded steel reinforcement. A precaution should be taken as some of

the ground industrial wastes or by-products such as those containing calcium aluminate and

metakaolin geopolymer have loss of strength during long-term ageing (Provis and Van Deventer,

2009). Despite extensive research in this area, the entire polymerization process of AAC is not

totally understood. It is classified as a polymer because of its huge molecule formed by a number

of smaller groups of molecules. AAC has superior mechanical, chemical and thermal properties

compared to ordinary Portland cement (OPC) (Duxson, Provis, Lukey and Van Deventer,

2007). The main benefit of AAC is that the source material is not a carbonate-bearing material;

therefore, it does not release vast quantities of CO2 as in the case of Portland cement. Turner

and Collins (2013) reported that the carbon emission of AAC concrete is around 9% less than

comparable OPC concrete as alkali activators are high carbon footprint materials. Also, high

early age strength, high chemical durability and resistance to high temperature are beneficial

properties of AAC. Normally, cementitious materials have several phases that contribute to the

mechanical properties.

This Chapter presents the mechanical and micromechanical properties of Alkali-activated fly ash

86

cement (AAFA) paste and mortar. The experimental programme conducted has been designed

according to Taguchi’s method (Roy, 2010) an efficient approach for investigating optimum

design parameters as per the required performances. The investigated parameters are density

and compressive strength of AAFA samples containing varying proportions of constituents. In

addition to the density and compressive strength, microporomechanics of AAFA samples have

been explored using Nanoindentation and statistical analysis to determine viscoelastic proper-

ties, which are contact relaxation modulus and contract creep compliance. Further, statistical

analysis of nanoindentation based on contact stiffness measurements enables to produce inden-

tation stress-strain curves of multiphase of AAFA materials. The methods used in this Chapter

can quantify the effects of the test parameters on the mechanical and micromechanical prop-

erties of the materials, and the outcome leads to a guideline for design of AAFA mixtures to

achieve required properties.

5.2 Taguchi’s Design of Experiment

Taguchi’s design experimental approach to parameters design provides design engineers with a

systematic and efficient method for investigating optimum design parameters for performance

and cost. In this research, the following parameters are considered in the mix proportions:

- Silica fume (SF)

- Sand to cementitious material ratio (s/c)

- Liquid to solid ratio (l/s)

- Superplasticiser (SP)

Three levels of the test parameters were selected as shown in Table 5.1. The Standard L9 (34)

orthogonal array (Ross, 1996) is used as shown in Table 5.2.

Table 5.1: Variation parameters and levels for AAFA mixture

Levels SF s/c l/s SP

1 0 % 0 0.6 0 %2 2 % 0.25 0.65 2 %3 4 % 2.5 0.7 4 %

Table 5.2: Standard L9 orthogonal array

Levels SF s/c l/s SP

1 1 1 1 12 1 2 2 23 1 3 3 34 2 1 2 35 2 2 3 16 2 3 1 27 3 1 3 28 3 2 1 39 3 3 2 1

87

In this experimental work, Class F (low calcium) fly ash from Collie Power Station in Australia

was used to prepare AAFA paste and mortar samples. Sand having a specific gravity of 2.6

was used as fine aggregates, and silica fume was used as a very fine pozzolan. The chemical

composition of fly ash and silica fume is presented in Table 5.3. The Loss on Ignition (LOI) of

fly ash and silica fume were 1.5% and 3.8%, respectively. The median particle size of fly ash and

silica fume (condensed silica fume) were 45µm and 5µm, respectively. The XRD pattern of the

Class F fly ash is shown in Figure 5.1. The alkali activator was prepared by dissolving NaOH

pellet in water and left for 24 hours, and Na2SiO3 (water glass) was then added and again left for

24 hours before mixing. The chemical composition of the Na2SiO3 was: Na2O = 14.7%, SiO2 =

29.4%, and water content was 55.9% by mass. The effect of Na2SiO3 to NaOH solution by mass

on compressive strength was reported that higher Na2SiO3 to NaOH ratio would increase the

compressive strength (Hardijito and Rangan, 2005). Considering the economic cost of alkaline

liquid, Na2SiO3 to NaOH ratio of 2 was used to prepare all mixtures. A naphthalene sulphonate

superplasticiser was used to improve workability. Specimens were cast in 50mm cubic moulds

and cured in a chamber at 60◦C and 70% relative humidity for 24 hours. This temperature is

commonly used for curing alkali-activated binders (Duxson, Fernandez-Jimenez, Provis, Lukey,

Palomo and Van Deventer, 2007; Hardijito and Rangan, 2005). After that, the specimens were

placed in a 23◦C controlled room until the testing age. The compressive strength and density

of AAFA samples were tested at the age of 28 days in accordance with ASTM ASTM Standard

C109 (2011). The reported results are the average of five samples.

Table 5.3: Chemical composition (wt.%)

(a) Class F fly ash

SiO2 Al2O3 CaO Fe2O3 K3O Others

65.9 24.0 1.59 2.87 1.44 4.2

(b) Silica fume

SiO2 Na2 K2O Other

89.6 0.11 0.23 10.06

Figure 5.1: X-ray diffraction pattern of low calcium fly ash

88

5.3 Mechanical Properties

Nine mixtures of AAFA samples were prepared according to Taguchi’s design of experiment as

presented in the previous section. Details of the mixtures content and the experimental results

of the density (ρ28) and compressive strength (fck,cubic,28) at 28 days are presented in Table

5.4. Signal to Noise (S/N) ratio and analysis of variance (ANOVA) can be used to determine

the relative influence of factors and their interaction to the variation of the results. ANOVA

is a statistical test which analyses variance. It is helpful in formally testing the significance of

all main factors and their interactions by comparing the mean square against an estimate of

the experimental errors at specific confidence levels. In this study, the significant value of 0.5

was used. Graphs of S/N ratio and results of ANOVA analysis are presented in Figure 5.2 and

Table 5.5.

The obtained results clearly indicate that the density is an important parameter for reducing

the self-weight of structures. The density of AAFA at 28 days is between 1521 kg/m3 to 1717

kg/m3. The results of S/N on the density also indicate that the decreasing density is affected

by the increasing the liquid to solid ratio and superplasticiser. The ANOVA results show that

sand to cementitious material ratio is the most influencing parameter on the density of AAFA

with 79% contribution. The other influencing parameters are superplasticiser and liquid to solid

ratio with approximately 10% and 9% contributions, respectively, as shown in Figure 5.3(a). It

is observed that silica fume has the least influence on the density. The optimal mix proportion

for minimisation of density is when there is no sand with 2% of silica fume, 0.7 of liquid to solid

ratio and 4% of superplasticiser.

It is observed that the 28 days compressive strength of AAFA is in a range of 6.8 to 25.3 MPa.

Table 5.4: Mix proportion with density and compressive strength test result

Mixρ28 fck,cubic,28

(kg/m3) (MPa)

1 1570 20.302 1598 15.843 1615 12.454 1509 6.795 1586 10.406 1685 9.927 1521 8.828 1602 9.929 1717 13.25

Table 5.5: ANOVA on density and compressive strength at 28 days

Source DFa SSb MSc Contribution %

Density

SF 2 790.1 395.0 2.14s/c 2 29155 14577.5 79.14l/s 2 3270.0 1635.0 8.88SP 2 3624.5 1812.2 9.84

Compressivestrength

SF 2 133.4 66.7 55.48s/c 5.7 2.9 2.38 29.15l/s 31.2 15.6 12.99 12.99SP 70.1 35.1 29.15 2.38

adegree of freedom bsum of square cmean square

89

Figure 5.2: Effect of parameters on density and compressive strength at 28 days

90

From ANOVA results, silica fume has the most significant effect on the compressive strength

with 55% contribution as shown in Figure 5.3(b). It can be seen that there is a minimal effect

of sand to cementitious material ratio on the compressive strength of about 2% contribution.

The second most important parameter is the superplasticiser content with 29% contribution.

The optimum mix proportion for 28 days compressive strength is a normal paste, i.e., no silica

fume, no sand, no superplasticiser with liquid to solid ratio of 0.6.

The main reason to obtain regression model is to uncover causes of the outcome by studying

the relationship between variables. In this research, therefore, regression analysis was used as

a tool for investigating the relationship between variables and to predict parameters outside

of the ranges of variables. Through a linear regression analysis, the empirical relationship of

AAFA mixtures at 28 days curing ages was obtained as:

Figure 5.3: Contribution of experimental parameters on (a) density (b) compressive strength

91

ρ28 = 2004.1 + 621.43x1 + 190x2 − 721.43x3 − 8175x4

+ 182.86x2x3 − 1542.9x2x4 − 11286x3x4 (5.3.1a)

fck,cubic,28 = 72.244− 36.548x1 − 100.47x2 − 85.919x3 − 863.25x4

+ 146.97x2x3 + 237.43x2x4 + 1047.1x3x4 (5.3.1b)

where x1, x2, x3, x4 denote silica fume, sand to cementitious material ratio, liquid to solid ratio

and superplasticiser, respectively. As shown in Table 5.6, provides the results which match well

with the experimental results.

Validation of predictive model can be carried out with corrected coefficients (Analla, 1998). The

information about regression analysis is presented in Table 5.7. It can be seen that corrected

coefficient values of density at 28 days is 0.995 (99.5%). The corrected coefficient of compressive

strength at 28 days has low values compared with that of the density, as 0.855 (85.5%).

Table 5.6: Experiment and Predicted density and compressive strength at 28 days

MixExperiment Predicted Ratio

ρexp fck,cubic,exp ρpred fck,cubic,predρexp fck,cubic,exp

/ρpred /fck,cubic,pred

1 1570 20.30 1571 20.69 0.999 0.9812 1598 15.84 1588 12.70 1.006 1.2483 1615 12.45 1616 12.84 0.999 0.9694 1509 6.79 1514 8.36 0.997 0.8125 1586 10.40 1591 11.97 0.997 0.8696 1685 9.92 1690 11.49 0.997 0.8637 1521 8.82 1518 8.03 1.002 1.0988 1602 9.92 1599 9.13 1.002 1.0869 1717 13.25 1714 12.46 1.001 1.063

Average 1.000 0.999StD 0.003 0.139

92

Table 5.7: Regression analysis of test data

(a) Density at 28 days

Parameters Estimate Standard error T value

Intercept 2004.1 166.82 12.013FA 621.43 437.14 1.422s/c 190 477.98 0.399w/c -721.43 265.90 -2.713SP -8175 5974.8 -1.368

s/c×l/s 182.86 741.85 0.247s/c×SP -1542.9 1854.6 -0.832l/s×SP 11286 9273.1 1.217

Corrected coefficient = 0.995

(b) Compressive strength at 28 days

Parameters Estimate Standard error T value

Intercept 72.244 51.946 1.391FA -36.548 136.12 -0.269s/c -100.47 148.84 -0.675w/c -85.919 82.797 -1.038SP -863.05 1860.4 -0.464

s/c×l/s 146.97 231 0.636s/c×SP 237.43 577.5 0.411l/s×SP 1047.1 2887.5 0.363

Corrected coefficient = 0.855

5.4 Indentation Properties

A number of researchers (Criado et al., 2008; Fernandez-Jimenez and Palomo, 2005; Garcia-

Lodeiro et al., 2011; Xu and Van Deventer, 2000) studied the characterization of alkali-activated

materials by a variety of experimental methods including XRD, SEM, DTA and TGA. Jennings

(2000), Tennis and Jennings (2000) proposed a model for the determination of two types of

calcium silicate hydrate (C-S-H) viz., high density (HD) and low density (LD) C-S-H, at different

parts of the specimen’s geometry. The content of C-S-H is determined in term of its volume

fraction of the indentation grid. Constantinides and Ulm (2004) determined the two C-S-H

types, Portlandite (CH), and unhydrated clinker using nanoindentation method. The result

shows that decalcification of C-S-H phases is the primary source of nanometre-scale elastic

modulus degradation.

Recently, Nemecek et al. (2011) studied the reaction products of AAFA paste, which was made

from NaOH solution having the liquid to solid ratio of 0.531, and was cured at 80◦C for 12 hours.

Nanoindentation and environmental scanning electron microscope (ESEM) were used, and four

phases of reaction products were found. The four phases were identified as N-A-S-H (sodium

aluminosilicate hydrate), partly-activated slag (N-A-S-H gel intermixed with slag-like particles),

non-activated slag (porous non-activated slag-like particles), and non-activated compact glass

(solid non-activated glass spheres or their relicts). These findings also reveal that N-A-S-H

is the main reaction product which is linked to the atomic scale and nanostructure and is

independent of precursor material or the temperature curing regime. N-A-S-H phase is pure

and is related to the mechanical strength of AAFA matrix. The contents of Si ions in N-A-S-H

can be increased by the presence of Si ions in the raw materials. It has also been found that

93

the increasing condensation degree of Si ions in N-A-S-H relates directly to the mechanical

strength gain (Fernandez-Jimenez and Palomo, 2003, 2005). The partly-activated slag phase

is intermixed with the slag-like particles. The non-activated slag phase is porous and contains

non-activated slag-like particles. The non-activated compact glass phase is solid, non-activated

glass sphere.

This section presents the microporomechanics of AAFA samples using nanoindentation and

statistical analysis of nanoindentation data. A background on nanoindentation and how to

determine the indentation modulus, hardness, Poisson’s ratio, cohesion, friction coefficient and

packing density of materials have been presented in Chapter 3. Details on nanoindentation test

and statistical analysis results will be presented next.

5.4.1 Sample Preparation

For nanoindentation tests, each specimen after 28 days curing was cut using a diamond saw to

obtain 10 mm cube core part. The reaction of specimens was stopped using solvent exchange

and oven dry method (Chen et al., 2014; Zhang and Scherer, 2011). Miller et al. (2008) pointed

out that it was important to reduce the surface roughness of specimen to get an accurate nanoin-

dentation results because the specimen’s surface has a significant influence on the test results

(ISO 14577-1, 2002). Therefore, fine emery paper was used to grind all specimens to reduce

the surface roughness. After that, the surface was polished using a suspension solution ranging

from 6.0µm to 0.1mum for 45 mins and with 0.05µm for another 15 mins. Nanoindentation test

was then carried out on the four sets of 10 mm cube specimens with XP (all-purpose) testing

method (applied load: 0.5mN); each set was specified 10 ± 10 grid (100 indentations) with

20µm grid spacing. Berkovich indenter was used for the entire tests.

5.4.2 Statistical Indentation

The results of nanoindentation test using deconvolution technique in terms of the cumulative

distribution function (CDF) and probability density function (PDF) of the tested AAFA speci-

mens confirm the presence of the four phases in AAFA. This finding is similar to the previously

reported results (Nemecek et al., 2011). The typical indentation load depth (P − h) curves are

illustrated in Figure 5.4. The mean indentation depth is 240 ± 7 nm in the N-A-S-H phase,

143± 3 nm in the partly-activated slag, 82± 1 nm in the non-activated slag and 54± 4 nm in

the non-activated compact glass phases.

With the hypothesis that AAFA is porous, the quadratic minimisation problem, Equation

(3.4.8), can be solved using MATLAB (2014). The results of the quadratic minimisation proce-

dure of scaling indentation modulus and hardness exhibit the properties of the reaction prod-

ucts. A typical packing density distribution of AAFA in relation to the indentation modulus

and hardness is as shown in Figure 5.5. The results of the solid properties of reaction products

of AAFA from the quadratic minimisation are presented in Table 5.8. The ranges of stiffness,

cohesion and friction angle are 84.30 to 144.57 GPa, 0.25 to 1.46 GPa and 29.13 to 37.93 de-

gree, respectively. The properties of reaction products are to be understood in the sense of the

Drucker-Prager strength model and Coulomb material model (Ulm et al., 2007).

Analysis of S/N ratio shows the effect of the test variables on the properties of the reaction

products as shown in Figure 5.6. It can be observed that generally an increase in the value

of the parameters leads to a decrease in the stiffness of the reaction products. However, an

increase in the value of the parameters such as silica fume, sand to cementitious material ratio

and superplasticiser increases the cohesion property. It could mean that an increase in silica

fume, sand to cementitious material ratio and superplasticiser could lead to an increase in the

94

Figure 5.4: Typical indentation load-depth (P − h) curve on AAFA

Figure 5.5: Packing density relationship distribution of AAFA on Mix 1

Table 5.8: Properties of the reaction products

MixStiffness Poisson’s Cohesion Friction Friction angle(GPa) ratio (GPa) coefficient (degree)

1 144.57 0.50 1.46 0.62 31.852 87.06 0.50 0.48 0.62 31.783 84.30 0.50 0.39 0.67 33.934 84.40 0.50 0.32 0.63 32.315 103.80 0.50 0.48 0.62 31.816 87.93 0.50 0.38 0.64 32.647 84.87 0.50 0.43 0.56 29.138 84.81 0.50 0.40 0.62 31.939 94.95 0.50 0.25 0.78 37.93

95

bonding between particles to particles. Based on Mohr-Coulomb yield surface theory (Labuz

and Zang, 2015), friction angle can be used to understand Drucker-Prager yield surface. This

study reveals that relationship between stiffness and cohesion have an inverse relationship,

i.e., an increase in stiffness leads to a decrease in cohesion. ANOVA results show that the

contribution of superplasticiser, silica fume and sand to cementitious material ratio are the

most significant parameters on the stiffness, the cohesion and the friction angle, respectively,

as shown in Table 5.9 and Figure 5.7. The optimal mixing design to achieve high stiffness is

thus no silica fume, no sand content, 0.6 of liquid to solid ratio and no superplasticiser content.

The optimal mixing proportion to achieve high cohesion, based on these test results, is 4% of

silica fume, 0.5 of sand to cementitious material ratio, 0.65 of liquid to solid ratio and 4% of

superplasticiser. The optimal mixing design to achieve high friction angle is 4% of silica fume,

0.5 of sand to cementitious material ratio, 0.65 of liquid to solid ratio and no superplasticiser

content.

Table 5.9: ANOVA results on properties of reaction products

Source DFa SSb MSc Contribution %

Stiffness

SF 2 483.1 241.6 15.68s/c 2 411.8 205.9 13.37l/s 2 511.1 255.5 16.59SP 2 1674.6 837.3 54.36

Cohesion

SF 2 0.3217 0.1608 30.41s/c 2 0.2505 0.1252 23.68l/s 2 0.2625 0.1312 24.81SP 2 0.2232 0.1116 21.10

Frictionangle

SF 2 0.8509 0.4254 1.91s/c 2 23.4753 11.7376 52.62l/s 2 9.4317 4.7158 21.14SP 2 10.8536 5.4268 24.33

adegree of freedom bsum of square cmean square

96

Figure 5.6: Effect of parameters on reaction properties (a) Stiffness (b) Cohesion (c) Friction angle

97

Figure 5.7: Effect of parameters on reaction properties (a) Stiffness (b) Cohesion (c) Friction angle

98

Equation (3.5.5) was used to calculate the volume fraction of N-A-S-H, partly-activated slag,

non-activated slag, and non-activated compact glass phases of AAFA mixtures. The nanoin-

dentation deconvolution results of indentation modulus (M), hardness (H), packing density (η)

and volume fraction are obtained. As expected, N-A-S-H phase is the major reaction product

and relates directly to the mechanical strength gain (Fernandez-Jimenez and Palomo, 2003,

2005). It can be seen from the obtained results in the present case that the lowest indentation

modulus, hardness and packing density are with reference to pure N-A-S-H phase, the decon-

volution results of N-A-S-H phases vary from 4.44 to 16.78 GPa for indentation modulus, 0.11

to 0.75 GPa for indentation hardness, and 6 to 53% of volume fraction. Table 5.10 summarises

the results of the deconvolution procedure of the AAFA mixtures.

ANOVA results as shown in Table 5.11 and Figure 5.8 indicate that silica fume is the most

significant effect on the indentation modulus of N-A-S-H phase with 47% contribution, and

sand to cementitious material ratio has the minimal effect on indentation modulus of N-A-S-H

phase. The best mix proportion for the indentation modulus of N-A-S-H phase is 2% silica fume,

sand to cementitious material ratio of 0.5, liquid to solid ratio of 0.65, and 0% superplasticiser

content. For the indentation hardness of N-A-S-H, superplasticiser has the most effect, i.e.,

almost 63% contribution, while the liquid to solid ratio has the least effect, with just over 4%

contribution. The optimum mix design of indentation hardness of N-A-S-H phase is 4% silica

fume, sand to cementitious of 0.25, liquid to solid ratio of 0.6 and 4% superplasticiser content.

Based on these results, the relationship between indentation modulus and hardness of N-A-S-H

phase of AAFA is an inverse relationship. Silica fume is found to have the most adverse effect on

the volume fraction of N-A-S-H phase, with almost 76% contribution. This outcome correlates

well with the compressive strength results, i.e., the best mix for the maximum volume fraction

of N-A-S-H phase is a normal paste mixture.

99

Table 5.10: Deconvolution results

(a) Indentation Modulus (M) (Unit in GPa)

MixA B C D

µ σ µ σ µ σ µ σ

1 10.20 4.54 22.25 3.42 36.33 8.69 69.58 22.112 11.76 2.84 19.45 5.83 46.21 11.41 70.47 7.213 7.90 3.30 16.11 4.76 33.40 9.33 53.07 20.694 16.32 4.72 18.72 8.00 28.93 6.90 55.05 18.215 16.41 5.67 22.02 1.40 29.32 7.91 66.25 20.596 16.25 4.22 23.56 6.90 40.51 18.89 66.43 19.207 10.50 1.69 16.56 1.98 25.28 6.47 56.62 18.158 4.44 2.47 14.29 3.02 23.48 6.99 56.51 16.919 16.78 0.98 19.23 2.75 24.43 7.60 47.61 25.36

(b) Indentation Hardness (H) (Unit in GPa)

MixA B C D

µ σ µ σ µ σ µ σ

1 0.61 0.33 2.09 0.52 4.20 3.10 10.04 4.882 0.52 0.19 1.21 0.55 4.75 1.75 7.47 0.393 0.25 0.13 0.82 0.45 3.01 1.33 6.60 1.604 0.53 0.24 1.42 0.35 1.56 0.74 5.45 2.765 0.75 0.28 0.91 0.43 1.71 0.96 7.18 2.236 0.62 0.27 1.25 0.51 4.05 1.95 8.15 0.497 0.52 0.10 0.77 0.11 1.11 0.57 6.24 2.848 0.11 0.08 0.69 0.17 1.22 0.74 6.46 2.099 0.57 0.12 0.91 0.10 1.22 0.59 0.58 3.16

(c) Packing density (η)

MixA B C D

µ σ µ σ µ σ µ σ

1 0.55 0.02 0.61 0.01 0.61 0.12 0.72 0.052 0.58 0.02 0.63 0.03 0.77 0.05 0.83 0.013 0.56 0.02 0.61 0.03 0.71 0.05 0.80 0.044 0.60 0.03 0.64 0.02 0.69 0.02 0.82 0.075 0.60 0.03 0.62 0.01 0.66 0.04 0.82 0.056 0.61 0.03 0.63 0.04 0.75 0.07 0.86 0.017 0.58 0.01 0.62 0.01 0.65 0.04 0.84 0.078 0.53 0.02 0.61 0.02 0.64 0.05 0.83 0.059 0.60 0.03 0.62 0.01 0.63 0.04 0.76 0.08

A: N-A-S-H, B: Partly-activated slag

C: Non-activated slag, D: Non-activated compact glass

100

Table 5.11: Deconvolution results of Volume fraction (f) (Unit in %)

Mix A B C D

1 53 8 14 242 48 27 18 73 24 26 29 214 29 16 25 315 28 9 30 326 42 23 25 107 6 31 33 318 9 35 31 259 8 27 26 39

A: N-A-S-H, B: Partly-activated slag

C: Non-activated slag, D: Non-activated compact glass

Table 5.12: ANOVA results on properties of reaction products

Source DFa SSb MSc Contribution %

Modulus(GPa)

SF 2 74.06 37.1 46.95s/c 2 11.54 5.8 7.32l/s 2 34.66 17.4 21.97SP 2 37.50 18.8 23.77

Hardness(GPa)

SF 2 0.090 0.044 28.35s/c 2 0.014 0.007 4.43l/s 2 0.014 0.007 4.34SP 2 0.200 0.100 62.87

Volumefraction (%)

SF 2 0.190 0.100 75.83s/c 2 0.003 0.002 1.27l/s 2 0.036 0.020 14.26SP 2 0.022 0.010 8.62

adegree of freedom bsum of square cmean square

101

Figure 5.8: Contribution of experimental parameters on N-A-S-H phase at 28 day

102

The total activated reaction phases can also provide an estimation of the activation degree in

the form of:

ξ =

ϕ∑i=1

γfiε

(5.4.1)

where ϕ is the number of activated reaction phases, γ is the fly ash mass in the mixture, and

ε is the total mass of the mixture. Thus, the total activation degree in the present case is the

sum of the volume fraction of the reaction phases which are N-A-S-H and partly-activated slag

phase. Mercury intrusion porosimetry (MIP) is a common measurement for a valid estimation

of the pore size distribution of porous solid (Diamond, 2000). However, the measurement of

porosity with MIP appears to be valid with limited application and difficulty in estimating

the porosity of a cementitious based material. A way to determine porosity with statistical

indentation technique is by a new nonintrusive way (Ulm et al., 2007). Therefore, it is possible

to calculate the total porosity of the cementitious materials from:

φ =N∑i=1

fi (1− ηi) (5.4.2)

In the present case, the degree of activation and the porosity are found ranging from 23 to

61% and 30 to 40%, respectively which are good agreement with literature (Fernandez-Jimenez

et al., 2006). From ANOVA results as shown in Table 5.12 and Figure 5.9, silica fume has the

most significant effect on the activation degree and the porosity with 49% and 60% contribu-

tions, respectively. Superplasticiser has a minor effect on the degree of activation with 10%

contribution, and sand to cementitious material ratio has the least effect on the porosity with

less than 1% contribution. The optimum mix proportion in terms of the degree of activation of

AAFA is 4% silica fume, sand to cementitious material ratio of 0.25, liquid to solid ratio of 0.6,

and 0% superplasticiser content. For the porosity, the optimum mix is 2% silica fume, sand to

cementitious material ratio of 0.25, liquid to solid ratio of 0.7 and 4% superplasticiser content.

Table 5.13 shows a summary of degree of activation and porosity of AAFA mixtures.

Table 5.13: ANOVA results on properties of reaction products

Source DFa SSb MSc Contribution %

Degree ofactivation

SF 2 0.06697 0.033486 49.12s/c 2 0.02721 0.013603 7.32l/s 2 0.02831 0.014156 19.96SP 2 0.01384 0.006921 10.15

Hardness(GPa)

SF 2 0.004125 0.002063 60.39s/c 2 0.000055 0.000028 0.81l/s 2 0.001927 0.000964 28.22SP 2 0.000723 0.000361 10.58

adegree of freedom bsum of square cmean square

103

Figure 5.9: Contribution of experimental parameters on activated degree and porosity

Table 5.14: Degree of activation and porosity (Unit in %)

Mix Degree of activation Porosity

1 61 402 59 353 34 334 43 305 29 316 43 327 35 308 34 339 23 32

104

A refined analysis of the volume fraction is as shown in Figure 5.10 and Figure 5.11. These

Figures show the overall volume fraction of the four reaction products, N-A-S-H, partly activated

slag, Non-activated slag and Non-activated compact slag. The increase of the silica fume and

liquid to solid ratio entails a decrease in N-A-S-H phase. In turn, this decrease is due to a

decrease in the similar portion of porosity of N-A-S-H phase. While the silica fume and liquid

available for reaction decrease the amount of N-A-S-H, it also decreases the N-A-S-H porosity.

The increase in the amount of N-A-S-H phase favours the formation of loose packed N-A-S-H

phase.

Figure 5.10: Volume fraction distribution with varying silica fume of (a) reaction products (b)porosity

105

Figure 5.11: Volume fraction distribution with varying liquid to solid ratio of (a) reaction products(b) porosity

106

Of particular interest is the relationship between the compressive strength and the volume frac-

tion of the N-A-S-H phase of the AAFA mixtures, as shown in Figure 5.12. The compressive

strength increases only slightly with the increase in the volume fraction of N-A-S-H phase from

0.05 to 0.42. When the volume fraction of N-A-S-H is over 0.42, the strength increases signif-

icantly. This finding is similar to those reported by other researchers (Fernandez-Jimenez and

Palomo, 2003, 2005), i.e., the N-A-S-H phase relates directly to the mechanical strength. From

Figure 5 12, the result of the relationship indicates that in order to obtain high strength AAFA,

the volume fraction of N-A-S-H phase should be greater than 0.50. The relationship between

the compressive strength and the degree of activation at the age of 28 days is also plotted as

shown in Figure 5.13. The compressive strength increases with the degree of activation. When

the degree of activation is greater than 0.5, the compressive strength increases exponentially.

Thus, to optimise the compressive strength of AAFA mixtures, the aim should be to ensure the

degree of activation of around 0.6.

Figure 5.12: Relationship between compressive strength and volume fraction of N-A-S-H phase

Figure 5.13: Relationship between compressive strength and degree of activation

107

5.5 Viscoelastic Properties

Material has viscoelastic characteristics when its behaviour exhibits both elastic and viscous

properties. The behaviour of viscoelastic materials can be seen when, upon loading beyond the

yield point, a slow and continuous increase of strain at a decreasing rate can be observed, and

when removing stresses, strain decreases following an initial elastic recovery path showing some

energy dissipation in a form of hysteresis loop. Viscoelastic materials have a strain rate that is

dependent on time (Findley and Davis, 2013), the effect of which can be critical when designing

materials for Civil Engineering applications. Nanoindentation is one of new advanced tools to

obtain quantitative viscoelastic properties such as relaxation and creep using indentation test

results. The details of the calculated viscoelastic properties of AAFA mixtures will be presented

next.

5.5.1 Contact Relaxation Modulus

Rheological models such as Maxwell, Kelvin-Voigt and Maxwell-Kevin-Voigt models, are clas-

sical linear viscoelastic properties, which can be determined from the indentation test of AAFA

samples. In the process of classical linear viscoelastic solutions, the holding regions at the

maximum load on the obtained nanoindentation load-displacement curves were fitted to the

Equation (3.8.20) to (3.8.22) using MATLAB with nonlinear least square method (MATLAB,

2014). Table 5.14 summarises the average classic linear viscoelastic properties instantaneous

modulus Mo, spring of stiffness Mv, and unit of viscosity η of AAFA mixtures.

Table 5.15: Rheological properties (Unit in GPa)

MixMaxwell Kelvin-Voigt Maxwell-Kelvin-Voigt

Mo ηM Mo Mv ηv Mo Mv ηM ηv

1 12.27 1432.35 15.20 147.05 912.96 15.25 304.16 53994.85 665.532 11.13 1229.94 13.91 123.32 818.98 14.10 257.13 45915.57 582.043 16.61 1875.54 20.41 184.39 1280.48 20.09 445.36 5503615.74 967.734 16.35 1948.27 19.48 178.23 1273.69 19.72 433.10 14047286.98 799.325 15.72 1820.17 18.72 167.18 1089.61 18.46 347.65 453469.74 817.956 18.46 347.65 16.09 170.15 979.85 16.03 314.00 442039.30 779.667 12.53 1395.27 14.73 136.50 767.62 15.08 260.73 170675.34 556.738 11.97 1326.49 15.45 139.25 882.61 15.36 293.61 1628124.46 597.879 13.59 1488.22 17.18 168.09 944.17 18.01 357.05 2399192.36 726.76

108

The normalised contact relaxation modulus, which is defined as the relaxation modulus divided

by the instantaneous modulus, represents an important material property of linear viscoelastic

materials. The mean normalised contact relaxation modulus of Combined Maxwell-Kelvin-

Voigt is as shown in Figure 5.14. The contact relaxation modulus based on the Combined

Maxwell-Kelvin-Voigt model of contact normalised relaxation modulus presents that AAFA

mixtures have a very small initial reduction of relaxation modulus within around the first 5

second period. These results generally indicate that AAFA mixtures undergo similar normalised

relaxation modulus to OPC. In Chapter 4, the results of the relaxation modulus of OPC also

show a small reduction in the relaxation modulus, which occurs in the initial period only. Thus,

the stress relaxation and reduction of the elastic modulus of AAFA are similar to that of OPC

and are negligible during the early curing ages.

109

Figure 5.14: Normalised relaxation modulus

110

5.5.2 Contact Creep Compliance

The creep behaviour of cementitious material is largely related to the viscoelastic response

of the vital reaction products and binding phases of hardening. For long-term contact creep

compliance, the change in depth of the creep phase was fit with logarithmic function in Equation

(3.8.32) using MATLAB (2014). The average corrected coefficient and coefficients of logarithmic

curve fitting obtained are as shown in Table 5.16. The results of the curve fitting show the

corrected coefficient is over 0.98 (98%). The long-term contact creep modulus C is one of the

important parameters to determine the long-term creep compliance rate. The long-term contact

creep compliance can be determined as:

Jc (t) =1

Ctwhere C =

πPmax4 tan (α)hmaxx1

(5.5.1)

As explained in Section 3.8, Pmax, hmax and α obtained from indentation test and geometry

properties. Mix 2 and Mix 4 have the lowest and highest contact creep modulus of 336.803 and

2172.179 GPa, respectively. Figure 5.15 shows the average long-term contact creep compliance

rate of AAFA mixtures.

Based on contact creep modulus, a statistical S/N ratio analysis was performed to determine

the effect of parameters on the creep modulus of the mixtures, as shown in Figure 5.16, which

can be seen that an increase of sand to cementitious material ratio in mixtures could decrease

the creep modulus. ANOVA was performed on the results of creep modulus from nine mixtures,

as shown in Table 5.17 and Figure 5.17. ANOVA results indicate that silica fume has 26.15%

contribution, sand to cementitious material ratio 20.65%, liquid to solid ratio 22.37%, and

superplasticiser 30.83%. Therefore, the affecting parameters on the creep modulus indicate

that four parameters contribute over 20% to the creep modulus of AAFA mixtures. An increase

in sand to cementitious material ratio clearly leads to a decrease in the creep modulus. The

optimal mixing design for the creep modulus of AAFA mixtures is, thus, 2% of silica fume, no

sand content, 0.65% liquid to solid ratio and 4% of superplasticiser.

From the test results of long-term creep compliance rate, the specific creep at one year was

determined as shown in Table 5.18. The values generally agree well with literature. Wallah and

Rangan (2006) used a conventional method to determine the specific creep of AAFA concrete

at one year and reported values being in a range of 15 to 29 microstrain/MPa for 40 to 67 MPa

of compressive strength of concrete.

The specific creep results generally show that AAFA mixtures undergo lesser creep compared

to OPC. In Chapter 4, the specific creep at one year of OPC paste was observed as 18.32

Table 5.16: Average logarithmic coefficients and corrected coefficient

Mixx1 x2 x3 x4 C

R-square(nm) (s) (nm/s) (nm) (GPa)

1 11.935 3.037 -0.005 0.136 386.756 0.9862 10.395 3.123 0.015 0.151 336.803 0.9883 17.626 4.437 0.035 0.145 504.181 0.9814 7.577 14.485 -0.058 0.151 2172.179 0.9825 9.159 71.840 -0.089 0.139 536.120 0.9836 8.236 3.412 -0.074 0.151 377.457 0.9827 6.365 2.160 -0.051 0.135 397.848 0.9878 14.230 2.508 0.117 0.152 503.422 0.9879 8.057 2.497 -0.051 0.113 486.692 0.984

111

Figure 5.15: Contact creep compliance rate

112

Figure 5.16: Effect of parameters on creep modulus

Table 5.17: ANOVA results on creep modulus

Source DFa SSb MSc Contribution %

Degree ofactivation

SF 2 706711 353356 26.15s/c 2 557892 278946 20.65l/s 2 604564 302282 22.37SP 2 833046 416523 30.83

adegree of freedom bsum of square cmean square

Figure 5.17: Contribution of experimental parameters on creep modulus

113

Table 5.18: Specific creep

MixSpecific creep after one year

(microstrain/MPa)

1 19.3892 22.2653 14.8734 3.4525 13.8676 19.8677 18.8488 14.8969 15.408

microstrain/MPa for mixture having the compressive strength of 95MPa. It seems that AAFA

paste (Mix 1) has higher specific creep at one year than OPC paste, but the compressive strength

is relatively low. Thus, the specific creep of at one year of AAFA paste is low while considering

a relative compressive strength. Similarly, Warner et al. (1998) presented that specific creep

of OPC concrete after one year was around 50 to 60 microstrain/MPa for 60MPa concrete,

30 to 40 microstrain/MPa for 80 MPa concrete, and 20 to 30 microstrain/MPa for 90 MPa

concrete. That means, generally, the specific creep of AAFA is less than the values reported in

the literature. According to Wallah and Rangan (2006), the reason for smaller creep behaviour

of AAFA compared to OPC may be due to “block-polymerisation”, that is, the silicon and

aluminium composition in the fly ash are not completely dissolved in the alkaline solution. The

polymerisation takes place only on the surface of the atoms and is adequate to form the blocks

to produce the polymer binder.

As presented earlier in Section 5.4.2, the deconvolution technique identified material phases

as N-A-S-H, Partly-activated slag, Non-activated slag, and Non-activated compact glass. In

addition, this deconvolution technique can identify creep modulus C for each phase. Table

5.19 shows results of M , H, C and η with four Gaussian deconvolution input parameters. The

deconvolution results of creep modulus show that N-A-S-H is mainly the phase that increases the

creep compliance of AAFA. Due to “block-polymerisation”, Partly-activated and Non-activated

phases in which the silicon and aluminium composition are not completely dissolved in the

alkaline solution show having smaller creep behaviour because of higher creep modulus than

that of the N-A-S-H phase. Therefore, Partly-activated and Non-activated phases are leading

the creep behaviour of AAFA due to high creep modulus.

The S/N ratio was analysed to understand the effect of parameters on the creep modulus of

Partly-activated and Non-activated phases. As shown in Figure 5.18, liquid to solid ratio is the

most significant parameter on creep modulus. An increase in liquid to solid ratio leads to a

decrease in the creep modulus. Therefore, an increase in liquid to solid ratio could be a cause

of increasing creep behaviour of AAFA. ANOVA was also performed on the results of creep

modulus of Partly-activated and Non-activated phases as shown in Table 5.20 and Figure 5.19.

The results show that contribution of liquid to solid ratio on creep modulus of Partly-activated

slag, Non-activated slag and Non-activated compact glass are 50.74%, 88.69% and 68.08%,

respectively. Also, it was found that sand to cementitious material ratio and superplasticiser

have a minor effect on creep behaviour of AAFA.

114

Table 5.19: Deconvolution results of Creep Modulus (C) (Unit in GPa)

MixA B C D

µ σ µ σ µ σ µ σ

1 61.597 38.354 200.611 78.705 1185.385 742.521 11946.131 5933.3132 78.028 35.566 177.069 80.719 658.149 403.671 5451.601 5933.3133 40.982 29.609 120.580 60.085 526.086 270.479 2757.519 1485.6644 63.828 28.705 212.993 89.196 853.733 508.367 6518.359 4303.5785 65.646 35.395 206.029 61.857 737.200 396.137 10523.844 6704.9336 98.314 60.987 296.108 119.977 1415.070 704.133 12209.248 6868.1917 92.288 37.273 195.745 53.218 628.075 303.435 4371.304 3105.2108 105.965 67.850 258.409 169.916 1488.145 830.584 11545.101 6116.9269 65.515 23.356 180.902 60.882 519.336 349.451 5281.275 4111.733

A: N-A-S-H, B: Partly-activated slag

C: Non-activated slag, D: Non-activated compact glass

Table 5.20: ANOVA results on creep modulus of partly-activated and non-activated phases

Source DFa SSb MSc Contribution %

Partly-activatedslag

SF 2 8017.5 4008.8 40.47s/c 2 344.6 172.3 1.74l/s 2 10051.4 5025.7 50.74SP 2 1395.8 697.9 7.05

Non-activatedslag

SF 2 68104 34052 5.99s/c 2 29827 14913 2.62l/s 2 1008916 504458 88.69SP 2 30730 15365 2.70

Non-activatedcompactglass

SF 2 16521193 8260597 15.19s/c 2 9059186 4529593 8.33l/s 2 74028816 37014408 68.08SP 2 9133741 4566871 8.40

adegree of freedom bsum of square cmean square

115

Figure 5.18: Effect of parameters on creep modulus of partly-activated and non-activated phases

116

Figure 5.19: Contribution of experimental parameters on creep modulus of (a) Partly-activated slag(b) Non-activated slag (c) Non-activated compact glass

117

5.5.3 Indentation Stress-Strain Curve and Fracture Toughness

The geometry of the indenter tip and the indented material are relating to the maximum

indenting depth h and contact depth hc. The function, hc = f (h), can be determined by contact

stiffness measurement (CSM) between the indented material and the indenter process using

Equation (3.2.13). Knowing hc = f (h), and the geometry of indenter tip, the instant contact

area can be determined by the instant stress from the loading process using Equation (3.9.1).

According to Equations (3.9.3) to (3.9.5), the indentation stress-strain curve can be obtained

from statistical deconvolution phases of AAFA paste (Mix 1) as presented in Section 5.4.2. The

indentation stress-strain curves of statistical deconvolution phases present that each phase can

be captured based on initial and stationary yield using Berkovich indenter tip with 20nm of the

radius. The results of indentation stress-strain curves clearly illustrate elastic, elastic-plastic

and plastic regimes. The initial yielding points are observed approximately when the modulus

values are 103 MPa, 87 MPa, 240 MPa and 987 MPa for N-A-S-H, Partly-activated slag, Non-

activated slag and Non-activated compact glass phases, respectively. The stationary yielding

points from indentation stress-strain curve are obtained when the strength values reach 412

MPa, 386 MPa, 1492 MPa and 1975 MPa for N-A-S-H, Partly-activated slag, Non-activated slag

and Non-activated compact glass phases, respectively. Figure 5.20(a) presents the indentation

stress-strain curve of N-A-S-H phase which shows that it remains stable in the fully plastic

regime. The indentation stress-strain curves of Partly-activated slag, Figure 5.20(b), Non-

activated slag, Figure 5.20(c), and Non-activated compact glass phase, Figure 5 20(d), illustrate

that indentation strain hardening increases in the fully plastic regime and the fracture point

cannot be observed.

Figure 5.20: Indentation stress-strain curve

118

The fracture toughness of a material is related to the energy stored in the form of elastic or

plastic energy before fracturing. A robust method for determining fracture toughness with

nanoindentation was introduced in Chapter 3.6, using Equations (3.6.8) to (3.6.8). The overall

indentation results of the fracture energy release rate (Gc) and the fracture toughness (Kc)

were determined. The results are as shown in Table 5.21.

Based on the fracture toughness results, a statistical S/N ratio analysis was performed to

determine the effect of parameters on fracture toughness of AAFA mixtures, as shown in Figure

5.21, which can be seen that increase of silica fume in mixtures could increase fracture toughness.

ANOVA was performed on the results of the creep modulus from the nine mixtures as shown

in Table 5.22 and Figure 5.22. ANOVA results show that silica fume has 29.60% contribution,

sand to cementitious material ratio 20.41%, liquid to solid ratio 22.28%, and superplasticiser

26.71%. Therefore, the affecting parameters on the fracture toughness designate that the four

parameters contribute over 20% to fracture toughness. The increase of silica fume clearly shows

an increase in fracture toughness. The optimal mixing design for fracture toughness is, thus,

4% of silica fume, 0.25 of sand to cementitious material ratio, 0.65% liquid to solid ratio and

2% of superplasticiser.

Table 5.21: Fracture energy release rate and toughness

MixGc Kc

(N/m2) (MPa m1/2)

1 0.960 0.5352 0.709 0.4003 1.176 0.6284 0.817 0.4815 0.744 0.4746 0.731 0.4467 0.752 0.4428 0.857 0.4249 0.659 0.438

Figure 5.21: Effect of parameters on fracture toughness

119

Table 5.22: ANOVA results on fracture toughness

Source DFa SSb MSc Contribution %

Fracturetoughness

SF 2 0.011415 0.005707 29.60s/c 2 0.008257 0.004128 21.41l/s 2 0.008594 0.004297 22.28SP 2 0.010300 0.005150 26.71

adegree of freedom bsum of square cmean square

Figure 5.22: Contribution of experimental parameters on fracture toughness

5.5.4 Chapter Summary

The results obtained from this investigation are based on Taguchi’s experimental design ap-

proach and statistical analysis of nanoindentation results of AAFA. Taguchi’s design of exper-

iment with ANOVA technique in a cement-based binder was successfully applied. The results

were then used to determine the effect of the test parameters on the compressive strength, den-

sity, and indentation properties such as deconvolution phases, viscoelastic, stress-strain, and

fracture toughness of AAFA. Based on the results of this study, the following conclusions can

be drawn:

- In terms of compressive strength, the normal AAFA paste, i.e., no silica fume, no sand,

no superplasticiser with liquid to solid ratio of 0.6 is the optimum mix. For the four

parameters investigated, silica fume has the most adverse impact on the compressive

strength. Increases of superplasticiser and liquid to solid ratio also contribute to the

decrease in compressive strength. There is no significant effect of sand to cementitious

material ratio on the compressive strength. In terms of density, the increase in the sand

content significantly increases the density of the mixture. The increasing of liquid to solid

ratio and superplasticiser dosage further decreases the density of AAFA mixture.

- An application of nanoindentation technology was applied for the micro poromechanics of

AAFA mixtures. Statistical analysis based on deconvolution technique was carried out to

determine the mechanical properties. AAFA was considered as the porous material. Thus,

the packing density of each phase could be estimated by solving a quadratic minimisation.

Properties of reaction products were determined based on Drucker-Prager strength and

Coulomb material models. Stiffness and cohesion of reaction products have an inverse

120

relationship.

- Four main phases of the reaction products viz., N-A-S-H, Partly-activated slag, Non-

activated slag and Non-activated compact glass phases were identified. Other properties of

the reaction products such as stiffness, Poisson’s ratio, and cohesion and friction coefficient

were also determined.

- The analysis confirms that N-A-S-H phase is the major reaction product of AAFA. The

compressive strength is shown to relate to the volume fraction of N-A-S-H phase. To

obtain high strength AAFA, the volume fraction of N-A-S-H phase should be larger than

0.42. The volume fraction of the N-A-S-H phase of 0.50 or over is recommended.

- The compressive strength is also shown to relate to the activated degree of AAFA. When

the degree of activation is greater than 0.5, the compressive strength increases exponen-

tially. Thus, it is recommended that the degree of activation should be around 0.6 to

obtain high strength mixture.

- The relaxation modulus of AAFA was found to have similar behaviour of that of OPC,

i.e., only small initial reduction occurred within a very short period (5 second). Thus,

the stress relaxation and the reduction of the elastic modulus of AAFA can be neglected

during early curing ages.

- The creep behaviour study revealed that Partly-activated and Non-activated phases are

the main reason for creep in AAFA due to “block-polymerisation” concept. It was also

found that liquid to solid ratio is the most affecting parameter on creep, an increase

of liquid to solid ratio leads to more creep. Sand to cementitious material ratio and

superplasticiser have a minor effect on creep behaviour.

- A robust method was used to determine fracture toughness. It was obtained that an

increase in silica fume leads to an increase in fracture toughness. Overall, the fracture

energy release rate and the fracture toughness of AAFA are 0.822 N/m2 and 0.474 MPa

m1/2, respectively.

In general, the study presented in this Chapter shows that mechanical and micromechanical

properties of AAFA can be obtained by means of statistical analysis of nanoindentation test

data.

121

Chapter 6

Strain-hardening Behaviour of

Cementitious Composite

6.1 Introduction

Previous Chapters presented cementitious materials properties such as mechanical, chemical,

microstructure and nano-scale properties. This Chapter presents a study on properties of High-

Performance Fibre Reinforce cementitious composite (HPFRCC).

Cementitious materials, such as mortar and concrete, generally show brittle tensile behaviour.

However, the brittle tensile behaviour could be significantly improved by adding discontinuous

fibres. Historically, general reinforcement in concrete has been in the form of continuous rein-

forcing bars, which should be in an appropriate location to resist the imposed tensile and shear

stresses. In fibre reinforced cementitious composite, fibres are discontinuous and are randomly

distributed throughout the cementitious matrix. They tend to be more closely located than

conventional reinforcing bars and are therefore better at controlling cracking. High-Performance

Fibre Reinforced Cementitious Composite (HPFRCC) is a type of material that exhibits pseudo

strain-hardening characteristic under uniaxial tensile stress among short discontinuous fibre re-

inforced cementitious composites. The “High-Performance” is the quality in fibre reinforced

cementitious composite based on the shape of its stress-strain curve in direct tension (Naa-

man and Reinhardt, 2008). Figure 6.1 illustrates strain softening and pseudo-strain hardening

(stress-strain hardening) responding behaviour after first cracking occurs.

HPFRCC can be generally classified by composite mechanics, energy and numerical approach.

One way to define the condition to accomplish strain hardening behaviour is that post-cracking

strength of the composites is higher than its cracking strength. It is, therefore, necessary

to understand some important parameters which are related to the shape of the stress-strain

relationship of HPFRCC (Li, 1997).

Generally, the stress-strain curve of HPFRCC is as shown in Figure 6.2, the first cracking

marked as A is defined as the first visible cracking or deviation from linearity as detected along

to an initial ascending portion of the stress-strain curve. During this time, several micro-cracks

might be developed in the member or has “percolated” through the structural tensile member

(Naaman and Reinhardt, 2008). The first cracking points are termed σI

and εI, respectively.

According to Li and Leung (1992), using energy approach and fracture mechanics to model

HPFRCC, stress at flaw tip of the cracking point is relative to the stress intensity factor of

KI , which could be derived from basic principle of elasticity by the principle of linear elastic

122

Figure 6.1: The concept of stress-strain hardening and strain softening under tensile stress

fracture mechanics (LEFM). The first cracking state is assumed to be equal to the steady-

state cracking stress. Two possibilities exist for fibre reinforced composite after first transverse

crack, which are strain hardening characterised by multiple cracking, or strain softening and

localisation characterised by the continuous opening of major crack during fibre pull-out. The

peak point at the end of the strain hardening branch marked as B on Figure 6.2 is at the

maximum post-cracking stress and strain coordinates in term of σII

and εII

, respectively, For

strain hardening composite, σII> σ

I, while for strain softening composite, σ

II< σ

I. After

peak point B, no more cracks could develop and one crack becomes critical defining that the

onset of crack localisation will open under increased deformation.

Figure 6.2: HPFRCC stress-strain behaviour

123

6.2 Design of Strain-hardening Behaviour

Naaman (1972); Naaman and Reinhardt (2008); Naaman et al. (1974) developed an analytical

model of the fibre reinforced composite. The number of fibre per unit volume and unit area of

the composite are important for deriving the tensile strength of composite at first cracking. It is

assumed that fibre diameter is df . The average number of fibres per unit volume of composite,

Nv is given as:

Nv =4Vfπd2fLf

(6.2.1)

where Vf is the fibre volume fraction and Lf is the length of the fibre. The number of a unit

area, Ns, is:

Ns =4Vfπd2fLf

α2 (6.2.2)

where α2 is one for unidirectional fibres, 2/π for fibres randomly oriented in planes, and 0.5 for

randomly oriented in space. If fibres are distributed randomly, the stress cracking of the matrix

in tensile prime can be illustrated by:

σI = σmu (1− Vf ) + ατVfLfdf

(6.2.3)

σII

= λτVfLfdf

(6.2.4)

where σmu is the tensile strength of the matrix, τ is the average bond strength between the

matrix and the fibre. The coefficients α and λ are to account for the fibre distribution, orien-

tation and bond efficiency. Developing a strain-hardening behaviour, the following condition

must be satisfied:

σII≥ σ

I(6.2.5)

Substituting Equations (6.2.3) and (6.2.4) into Equation (6.2.5) (Naaman, 1972; Naaman and

Reinhardt, 2008; Naaman et al., 1974):

λτVfLfdf≥ σmu (1− Vf ) + ατVf

Lfdf

(6.2.6)

Solving Equation (6.2.6) for Vf then (Naaman, 1972; Naaman and Reinhardt, 2008; Naaman

et al., 1974):

V criticalf ≡ Vf ≥1

1 + τ/σmuLf/df (1− Vf )

(6.2.7)

This equation shows the critical volume fraction of the fibre required to achieve the strain-

hardening condition of the fibre reinforced composites.

Similarly, Li and Leung (1992) studied the condition for steady-state and multiple cracking for

a 3-D randomly distributed discontinuous fibre reinforced composite. Based on this study, Li

et al. (1995) shows that the condition to ensure strain-hardening behaviour of composite is:

V criticalf ≡ Vf ≥12Gc

gτ (Lf/df ) δo(6.2.8)

where Gc is the composite critical energy release rate. The g term represents snubbing factor.

The crack opening at fibre bridging stress reaches a maximum, δo, when

δo =τLf

Efdf (1 + η)(6.2.9)

124

where η = (VfEf ) / (VmEm), fibre volume fraction and elastic modulus are Vf , Ef , respectively,

and Vm, Em are the matrix volume fraction and elastic modulus, respectively. The snubbing

factor g is defined in term of snubbing coefficient f as:

g ≡ 2

4 + f2

[1 + exp

(πf

2

)](6.2.10)

where

f =1

Φln(P)

(6.2.11)

The term Φ is inclining angle between fibre and matrix, P is normalised force of fibre. Li et al.

(1990) studied the snubbing friction with fibre pull-out test from cementitious matrix. They

found that average nylon fibre of snubbing coefficient f is 0.994 and 0.702 for polypropylene

fibre. With fibre pull-out test, snubbing friction increases the pull-out resistance and contributes

to the overall composite toughness.

Also, the ultimate strength of the composite σcu corresponds with the maximum bridging stress

σo when a brittle matrix presents strain-hardening behaviour as (Li et al., 1990):

σcu =1

2gτVf

(Lfdf

)(6.2.12)

In Equation (6.2.8), the composite critical energy release rate is given by (Budiansky and Cui,

1994; Li et al., 1995):

Gc = (1− Vf )(1− v2m

) K2m

Em≈ K2

m

Em(6.2.13)

where vm is the matrix Poisson’s ratio, and Km is matrix fracture toughness. The strain

hardening of fibre volume fraction is generally limited to few percentages. Thus, Vf term

in equations can be eliminated. Therefore, an approximate equation for critical fibre volume

fraction required to achieve strain-hardening behaviour is:

V criticalf ≈ 12Gcgτ2

d2fL3f

(6.2.14)

Kanda et al. (2000) studied the theoretical prediction of the tensile stress-strain curve of strain

hardening composite. According to their study, the relationship between tensile stress and

strain is characterised by two stages that are first crack state and the ultimate state, as shown

in Figure 6.3. It shows that tensile stress-strain relationships can be represented by the bilinear

behaviour. The first crack state is assumed to be equal to the steady-state cracking stress σfc,

which is the initially bend-over point and initiation of multiple cracking. The ultimate state is

the peak tensile stress state as multiple cracking terminates. Thus, a bilinear presentation of

HPFRCC can be illustrated as (Kanda et al., 2000):

σ (ε) =

Ecε when ε ≤ σfc/Ec

σi + Eiε when ε > σfc/Ec

(6.2.15)

where Eie and σi are:

Eie =σpeak − σfcεcu − σfc/Ec

, σi =

(1− Eie

Ec

)where σpeak and εcu are ultimate theoretical stress and respectively, and the composite elastic

modulus Ec. The steady-state cracking stress can be expressed with composite Poisson’s ratio

125

Figure 6.3: Tensile stress-strain curve for HPFRCC (adopted after (Kanda et al., 2000))

vc by (Kanda et al., 2000):

σfc = σog[√

2cs −cs2

](6.2.16)

where

σo =Vfτ

2

Lfdf

δ∗ =2τ

Ef (1 + η)Lf

df

Ktip =KmEcEm

Ec = EmVm + EfVf

K =Ktip

gσocoδ∗co =

(LfEc2Ktip

)2π

16 (1− v2c )2 cs =

√cs

ˆdelta∗ c =

csco

And the term of cs can be obtained by solving:

K =2cs√π

(√2cs3− cs

4

)(6.2.17)

The theoretical ultimate stress σpeak can be obtained as (Kanda et al., 2000):

σpeak =Vfgτ

2

Lfdf

(6.2.18)

The ultimate theoretical strain in terms of micromechanical parameters can be written as:

εcu =δpeak

xtheoryd

(6.2.19)

where σpeak is ultimate crack opening displacement (COD) and xtheoryd is theoretical unlimited

crack spacing that is the distance normal to the crack surface for the load in the matrix.

The σpeak was predicted considering the slip-hardening occurrence in the behaviour of a single

polymeric fibre pull-out of a cementitious matrix as (Kanda et al., 2000):

δpeak =3dfβ1 − Lfβ2 + Lfβ2

[(L2

fβ22−14Lfdfβ1β2+9d2fβ

21+32dfβ1

L2fβ

22

]1/24Lfβ2

(6.2.20)

where β1 and β2 are the first and second order non-dimensional hardening parameters from the

load-displacement relation observed in the single fibre pull-out test, respectively.

The xtheoryd can be predicted in terms of saturated crack spacing xd which depends on trans-

ferred stress via bridging fibre at cracking plane to non-cracked matrix plane and flaw size

126

distribution F (cc) at cracking stress reached maximum bridging stress. The theoretical ulti-

mate crack spacing can be expressed by (Kanda et al., 2000):

xtheoryd =xd

1− F (cc)(6.2.21)

The saturated crack spacing xd can be expressed by (Kanda et al., 2000):

xd =Lf2−

[L2f −

2VmdfLfσmu

gτVf

]1/22

(6.2.22)

where σmu is tensile strength of matrix. The flaw size distribution F (cc) can be given by

Weibull function as (Kanda et al., 2000):

F (cc) = exp

[− 1

λ

(cocc

)m](6.2.23)

where reference data fitting can determine scale factor λ, m is Weibull modulus that is typically

assumed to fall between 2 to 3 for concrete. The normalised flaw size at maximum bridging

stress cc can be obtained by solving:[2√

2cc3− cc

4

]+

√π

2

K

cc= 1 (6.2.24)

Moreover, reference crack radius co = cm which is:

cm =

√cm

δ∗cm =

cmco

cm =

[√π

2

Km

σmu

]Equations (6.2.15) to (6.2.24) present the tensile stress-strain prediction. This approach can be

validated with the experimental results. Based on the theoretical discussion, Li et al. (1995)

recommended the following guideline for desirable matrix properties.

- The lower the composite critical energy release rate Gc is to achieve ductile behaviour of

the composite easily.

- The fine aggregate contents achieve upper limit for the composite critical energy release

rate.

- A higher bond strength enhances composite performance in term of composite strength.

The minimum fibre-matrix interfacial bond strength should have 0.5 MPa.

- A lower first crack strength indicates a low fibre volume fraction required. The tensile

strength of the matrix is limited to 3.0 MPa for practical purpose.

In this study, Equations (6.2.8) to (6.2.13), proposed by several authors (Budiansky and Cui,

1994; Kanda et al., 2000; Li et al., 1995, 1990), were used to design mix proportion of HPFRCC.

6.3 Single Fibre Pull-out

It is important to note that the interface properties between the fibre and matrix influence

significantly the performance of composites. The interface properties are also important in

the fracture mechanism and the fracture toughness of composite. The failure process in a

composite material when a crack propagates is complex and involves matrix cracking. The

bonding strength between fibre and matrix is to be considered as a source of energy dissipation

of HPFRCC. The single fibre pull-out test is the most common method to understand the

127

interfacial strength. Generally, fibre pull-out test has three stages during debonding (Chen

et al., 2009a,b; Herrera-Franco and Drzal, 1992; Hsueh, 1990; Kullaa, 1996; Naaman et al.,

1991; Zhan and Meschke, 2014), as shown in Figure 6.4. Each stage of single fibre pull-out test

can be expressed by:

- The first stage, So: the fibre and matrix is bonded until reach the maximum interfacial

bond strength τmax,

- The second stage, So − S1: a crack propagation could occur along the interface between

the fibre and matrix which is leading to a complete debonding,

- The third stage, S1 − Sref : fibre is pulled out from the matrix and starts to slip.

Thus, the maximum pull-out force is the most important parameter of HPFRCC, which can

present maximum interfacial bond strength. The purpose of the study in this Section is to

investigate a single PVA fibre pull-out test by FE analysis using cohesive zone materials (CZM)

model.

6.3.1 Numerical Model and Validation

A numerical study was carried out using commercial finite element (FE) software package AN-

SYS (ANSYS R○, 2015). A 2-D axisymmetric model was employed for simulation of single fibre

pull-out process. In the developed model, PVA fibre with a radius Rf was embedded at the

centre of the cylindrical matrix and Ld was the total embedded length of the fibre. The bottom

of the model was constrained in both radial and axial directions. The interface properties were

used the bilinear CZM model (MODE II), which is established by fracture mechanic models such

as interface traction and the separations, as shown in Figure 6.5. The relationship between nor-

mal critical energy Gcn and tangential critical energy Gct can be expressed by maximum normal

contact stress σmax, maximum tangential contact stress τmax, complete normal displacement

δn, and complete tangential displacement δt as (ANSYS R○, 2012):

Gcn =1

2σmaxδn (6.3.1a)

Gct =1

2τmaxδt (6.3.1b)

Figure 6.6 presents the model of the FE single fibre pull-out test. The fibre and matrix

model were meshed with 122406 six node quadrilaterals elements, as shown in Figure 6.7.

Figure 6.4: Idealised interface law in three stages for single fibre pull-out (adopted after (Zhan andMeschke, 2014))

128

Figure 6.5: Fracture Models

This model was analysed with a non-linear geometry method and convergent displacement

control as illustrated in Figure 6.8. To confirm the validity of FE analysis of single fibre

pull-out, the analytical fibre pull-out test was conducted. An interfacial friction law for the slip

Figure 6.6: Single fibre pull-out simulation model without inclined angle

Figure 6.7: Meshing configuration

129

Figure 6.8: Equivalent stress

mechanism between the fibre and the matrix has been investigated by several authors (Kullaa,

1996; Naaman et al., 1991; Zhan and Meschke, 2014). Zhan and Meschke (2014) proposed the

model based on the interfacial law that model can capture the major mechanism involved in

various situations as follows:

τ (s) =

G · S ; S ≤ So bondend stage

τmax ; So < S ≤ S1 debonding stage

τo + (τmax − τo) exp [(S1 − S) /Sref ] ; S > S1 silding stage

(6.3.2)

where

τ : interfacial shear stress

τmax : maximum interfacial bonding stress

τo : asymptotic value of frictional stress

s : relative displacement which is a point

on the fibre axis with respect to the boundary of the matrix

G : relative modulus = E/ [df (1 + vm) ln ξ]

ξ : Rm/Rf

Rm : radius of matrix

Rf : radius of fibre

Sref : parameter controlling descending branch of curve

Based on this interface law, the values of the pull-out force F and displacement δ are obtained

at different stages as:

Fo =πdfτmax

λtanh (λLd) , δo = so =

τmaxG

; for bonded (6.3.3a)

Fmax = πdfτmaxLd, δmax = So +pidfτmaxL

2d

2AfEf; for debonding (6.3.3b)

F (δ) = πdf

[τo + (τmax − τo) exp

(δmax − δSref

)](Ld − δ + δmax) ; for sliding (6.3.3c)

130

where

λ =

√πdfG

AfEf

Fe =πdfτmax

λtanh [λ (Ld − Lr)]

Lr = remaining bonded section length

Af = cross setion area of fibre

For analytical fibre pull-out, Equations (6.3.2) and (6.3.3) were used to obtain the fibre pull-

out force. The results of analytical and FE model of single fibre pull-out are overall in good

agreement, as illustrated in Figure 6.9. Thus, FE simulation can be used for investigating

interface behaviour between fibre and matrix.

Figure 6.9: Validation of FE model with analytical model

131

6.3.2 Taguchi’s Design of Experimental

Taguchi’s design experimental approach, eight parameters and three levels of test variable were

selected according to literatures (Budiansky and Cui, 1994; Kanda et al., 2000; Kullaa, 1996;

Li et al., 1995, 1990; Naaman et al., 1991; Zhan and Meschke, 2014), as shown in Table 6.1.

The Standard L27 (313) orthogonal array (Ross, 1996) is used according to these parameters.

The detail of L27 orthogonal array is shown in Table 6.2. In this numerical study, the non-

linear geometry method of displacement control for convergent was used. The displacement

was applied as the value of δt.

Table 6.1: Variation parameters and levels

Parameter Level 1 Level 2 Level 3

Elastic modulus of matrix, Em (GPa) 20 25 30Diameter of matrix, dm (mm) 5 10 15Poisson’s ratio of matrix, vm 0.2 0.22 0.25Elastic modulus of fibre, Ef (GPa) 40 120 210Diameter of fibre, df (mm) 0.038 0.5 1Fibre embedded length, Ld (mm) 4 10 12Maximum tangential traction, τmaxt (MPa) 0.5 1 1.5complete tangential displacement δt (mm) 0.1 0.25 0.4

Table 6.2: Standard L27 orthogonal array

No. Em dm vm Ef df Ld τmaxt Gct

1 1 1 1 1 1 1 1 12 1 1 1 1 2 2 2 23 1 1 1 1 3 3 3 34 1 2 2 2 1 1 1 25 1 2 2 2 2 2 2 36 1 2 2 2 3 3 3 17 1 3 3 3 1 1 1 38 1 3 3 3 2 2 2 19 1 3 3 3 3 3 3 210 2 1 2 3 1 2 3 111 2 1 2 3 2 3 1 212 2 1 2 3 3 1 2 313 2 2 3 1 1 2 3 214 2 2 3 1 2 3 1 315 2 2 3 1 3 1 2 116 2 3 1 2 1 2 3 317 2 3 1 2 2 3 1 118 2 3 1 2 3 1 2 219 3 1 3 2 1 3 2 120 3 1 3 2 2 1 3 221 3 1 3 2 3 2 1 322 3 2 1 3 1 3 2 223 3 2 1 3 2 1 3 324 3 2 1 3 3 2 1 125 3 3 2 1 1 3 2 326 3 3 2 1 2 1 3 127 3 3 2 1 3 2 1 2

132

6.3.3 Effect of Parameters on Maximum Pull-out Force

Based on the numerical study with Taguchi’s design experimental approach, a statistical S/N

ratio analysis was performed to determine the effect of these parameters on Pmax as shown in

Table 6.3 and Figure 6.10. The S/N ratio shows that diameter of fibre has the most effect on

the fibre pull-out force. The elastic modulus of fibre and matrix has a minor effect on the pull-

out force. ANOVA was conducted and its results indicate that the contribution of diameter

of fibre on pull-out force contributed is 44.69% of in Table 6.4 and Figure 6.11 present the

summary of the results. It can be observed that increasing the elastic modulus of the matrix,

the diameter of the fibre, tangential traction and embedded length of fibre result in increasing

the pull-out force. The contribution of the elastic modulus of the matrix, tangential traction

and embedded length of fibre on the pull-out force are 14.48%, 8.92% and 9.47%, respectively.

Increasing the diameter of the matrix and Poisson’s ratio resulted in decreasing in the pull-out

force. It is observed that the diameter of the matrix, Poisson’s ratio, the elastic modulus of fibre

and complete tangential displacement are minor contributing parameters on the pull-out force,

approximately 2.5% of contribution. Through regression analysis, the empirical relationship

Table 6.3: Numerical studies of single fibre pull-out with Taguchi’s DOE

No. Pmax (N) No. Pmax (N) No. Pmax (N)

1 0.23 10 1.266 19 1.4262 15.508 11 9.421 20 9.4143 56.091 12 15.705 21 12.5654 0.239 13 1.09 22 1.6995 15.666 14 9.408 23 9.4146 55.945 15 15.585 24 12.5637 0.239 16 1.356 25 1.518 15.611 17 9.404 26 9.1929 14.614 18 15.692 27 12.556

Figure 6.10: S/N ratio of single fibre pull-out

133

obtained can be written as:

Pmax = −51.1871Em + 19.9782Ef − 3.0446Emdm − 0.2965EmEf − 9.3565Emτmax

− 4.8925EmLd + 4.2477Emδt + 0.1150dmEf − 9.7633dmdf − 8.5607dmτmax

+ 0.1457dmLd + 0.65233Efδt − 69.0918vmEf + 431.227vmLd − 0.1679Efdf (6.3.4)

− 0.3646Efτmax + 0.0465EfLd + 3.0134E2m + 3.0893d2m + 0.0047E2

f + 1.9292L2d

The correct coefficient (R-square) is observed as 0.999.

Table 6.4: ANOVA of fibre pull-out force

Source DFa SSb MSc Contribution %

Pmax

Em 2 737.6 368.8 14.48dm 2 127.2 63.61 2.50vm 2 129.3 64.65 2.54Ef 2 124 61.98 2.43df 2 2276.5 1138.27 44.69τmax 2 454.4 227.19 8.92Ld 2 482.5 241.24 9.47δt 2 126.7 63.36 2.49

Error 10 653.8 63.58 12.48

adegree of freedom bsum of square cmean square

Figure 6.11: Parameter contribution on fibre pull-out test

134

6.3.4 Single Fibre Pull-out Test with Polyvinyl Alcohol (PVA) Fibre

Polyvinyl alcohol (PVA) fibres have been used commonly on a large scale in the construction

as they have good mechanical and chemical characteristics. The monofilament PVA fibre was

used in this research, its diameter and length are 38µm (8 deniers) and 8mm, respectively.

PVA fibre has high chemical bond strength due to the hydrophilic nature of PVA fibre and high

alkali resistance characteristic. The physical tensile strength and elastic modulus of PVA fibre

are 1600 MPa and 40 GPa, respectively. The failure process in a composite material when a

crack propagates is complex and involves matrix cracking. The bonding strength between fibre

and matrix is to be considered as a source of energy dissipation. Thus, a single fibre pull-out

test was conducted with OPC paste matrix (w/c = 0.3) and AAFA paste matrix (l/s = 0.6)

using PVA fibre, as illustrated in Figure 6.12. Due to the limitation of the mechanical testing

machine, only failure force was captured. The embedded length Ld of fibre is around 4mm,

which is half of the total length, and diameter of fibre df was 0.038mm. Assuming uniform

bonding, the maximum interfacial bonding strength τmax can be expressed as:

τmax =PmaxπdfLd

(6.3.5)

where Pmax is a maximum pull-out force. With equation (6.3.5), the maximum interfacial

bonding strength τmax was determined as shown in Table 6.5. The average elastic modulus of

OPC and AAFA were obtained using indentation modulus results as presented in Chapters 4

and 5. Table 6.6 lists the parameters for the FEM pull-out of OPC and AAFA. Table 6.7 shows

a comparison of the maximum pull-out force between analytical and experimental results which

are good agreement.

Figure 6.12: Schematic of single fibre pull-out test

135

Table 6.5: Maximum interfacial bonding strength

TestOPC AAFA

Pmax (N) τmax (MPa) Pmax (N) τmax (MPa)

1 0.48 1.01 0.53 1.112 0.59 1.24 0.38 0.83 0.44 0.92 0.55 1.154 0.48 1.01 0.61 1.285 0.47 0.98 0.58 1.216 0.48 1.01 0.54 1.137 0.55 1.15 0.53 1.118 0.43 0.9 0.57 1.199 0.44 0.92 0.58 1.2110 0.47 0.98 0.41 0.86

Mean 0.48 1.01 0.53 1.11Std 0.05 0.11 0.07 0.16

Table 6.6: Parameters for analytical modelling of fibre pull-out

OPC matrix

Elastic modulus (MPa) 21218Poisson’s ratio 0.24

AAFA matrix

Elastic modulus (MPa) 28011Poisson’s ratio 0.24

PVA fibre

Elastic modulus (MPa) 40000Tensile strength (MPa) 1600

Poisson’s ratio 0.3Diameter (mm) 0.038

Interface

Embedded length (mm) 4mmMatrix to fibre size ratio 263.15

τmax (MPa) for OPC matrix 1.01τmax (MPa) ofr AAFA matrix 1.11

Asymptotic frictional stress -Reference slip (mm) (Zhan and Meschke, 2014) 0.25

Table 6.7: Maximum pull-out force between analytical and experimental results

FEM Experimental Ratio

Maximum Pull-outForce(N)

OPC 0.482 0.480 1.001

AAFA 0.530 0.530 1.00

136

6.4 High-Performance Fibre Reinforced Cementitious Com-

posite

An experimental study was conducted to understand HPFRCC in AAFA and OPC matrices

which led to the development of HPFRCC. In this research, Class F (low calcium) fly ash from

Australia and Local general purpose (GP) Portland cement according to Australia Standard

(AS 3972) (Australia Standard, 2010) were used for the cementitious matrix. The summary of

the chemical compositions of OPC and fly ash is presented in Table 6.8, XRD pattern of fly

ash is shown in Figure 6.13. The specimens were cast in 25mm cubic mould for compressive

strength test, the three prismatic plate specimens of 160 × 40 × 40 mm in dimensions for flex-

ural performance. AAFA specimens were cured 24 hours at 60◦C, which is a common curing

temperature for AAC (Joseph, 2011; Hardijito and Rangan, 2005; Rangan, 2008). After that,

the specimens were placed in a curing room at 23◦C ± 3 until testing. Compressive strength

test was conducted on 7, 14 and 28 curing days using 200 N/s of loading rate. The flexural per-

formance was conducted according to ASTM ASTM Standard C1609 (2012) with 0.05mm/min

of loading rate on each specimen. A schematic installation for the flexural performance is shown

in Figure 6.14.

The selected mixing proportion is the process of choosing suitable fibre volume fraction of AAFA

and OPC mixture as shown in Table 6.9. Series A is composite AAFA mixture, in groups A

and B notate that mixtures were without silica fume and with silica fume, respectively. Series

P is composite OPC mixtures, in group A and B design that mixtures were without fly ash and

with fly ash, respectively.

Table 6.8: Chemical composition of OPC (type I) and low calcium fly ash (wt. %)

SiO2 Al2O3 CaO Fe2O3 K2O MgO SO3

OPC 21.1 4.7 63.6 2.7 - 2.6 2.5Fly ash 65.9 24.0 1.59 2.87 1.44 - -

Figure 6.13: XRD pattern of fly ash

137

Figure 6.14: A schematic of flexural performance

Table 6.9: Composition of mix proportions

(a) AAFA composites

Series Group Index FA SF l/s SP Vf

A

A

AA1 1 - 0.5 0.02 -AA2 1 - 0.5 0.02 0.5%AA3 1 - 0.5 0.02 1.0%AA4 1 - 0.5 0.02 2.0%

B

AB1 1 0.2 0.5 0.02 0.5%AB2 1 0.2 0.5 0.02 1.0%AB3 1 0.2 0.5 0.02 2.0%AB4 1 0.2 0.5 0.02 -

(b) OPC composites

Series Group Index OPC FA w/c Vf

AA

PA1 1 - 0.4 1.0%PA2 1 - 0.4 2.0%

BPB1 1 0.2 0.4 1.0%PB2 1 0.2 0.4 2.0%

138

6.4.1 Compressive Strength

Compressive Strength Development

Figure 6.15 and Figure 6.16 show the average compressive strength development between 7 to 28

days of curing ages in the different composites. The average compressive strength development

results were obtained from the average of six specimens per mixture. It can be seen that

compressive strength of AAFA and OPC composites generally decreased by fibre volume ratio.

Also, it was observed that compressive strength development is not significantly increased by

fibre volume ratio except A2, which exhibited a high rate of compressive strength development

between 7 to 14 days. The test results indicate that compressive strength development is not

significantly affected by the fibre volume ratio in both AAFA and OPC composites. The test

results are presented in Table 6.10.

Figure 6.15: Compressive strength development of AAFA composites,Series A

139

Figure 6.16: Compressive strength development of OPC composites, Series P

Table 6.10: Compressive strength development (MPa)

Series Index7 days 14 days 28 days

fck,cubic,7 StD fck,cubic,14 StD fck,cubic StD

A

AA1 72.53 9.55 75.45 4.61 75.62 9.91AA2 54.09 10.14 73.99 3.35 71.45 5.47AA3 58.41 5.69 66.98 3.68 70.62 3.37AA4 56.34 3.73 57.88 3.16 61.01 1.69AB1 57.5 6.73 57.39 6.63 64.16 4.76AB2 54.05 4.85 58.07 3.44 61.7 3.33AB3 54.06 2.54 55.7 1.02 57.57 4.26AB4 49.46 2.45 54.5 2.21 56.42 4.10

P

PA1 41.64 6.93 43.86 5.3 44.84 1.79PA2 34.54 2.99 37.41 3.02 37.98 1.92PB1 28.77 2.09 30.37 2.81 34.28 1.41PB2 30.87 2.26 31.82 2.05 33.04 0.79

140

6.4.1.1 Compressive Failure Mode

The behaviour and ultimate compressive failure mode of composites are shown in Figure 6.17

and Figure 6.18. It is known that PVA fibre matrix can exhibit ductile behaviour after reaching

its ultimate compressive strength because of the transverse confinement effect of the PVA fibre

while normal AAFA mixtures without PVA fibres (AA1, AB1) present a significant decrease

in stress after reaching the ultimate compressive strength. However, OPC composites have

more ductile behaviour after reaching its ultimate compressive strength compared to AAFA

composites. The effect of the fibre volumetric ratio on compressive strength and strain capacity

of AAFA composites can be seen in Figure 6.19. Figure 6.19(a) shows the effect of the fibre

volume ratio on the strain at the ultimate compressive strength. It can be seen that the

compressive strain is not significantly affected by the fibre volume ratio. Also, the compressive

strain corresponding to the compressive strength is not meaningfully affected. As shown in

Figure 6.19(b), compressive strength generally decreases with increasing fibre volume ratio in

Figure 6.17: Typical stress-strain curves of AAFA composites, Series A

141

Figure 6.18: Typical stress-strain curve of OPC composites, Series P

the AAFA composites. Silica fume was added to achieve lower compressive strength in AAFA

composites. As discussed in Chapter 5 silica fume in AAFA matrix contributes to a significant

decrease in the compressive strength due to a decrease in the cohesion of the reaction products.

The effect of the fibre volume ratio on compressive strain capacity of OPC composites was

also observed, as summarised in Table 6.11. The results show that fly ash contents of OPC

composites tend to decrease the compressive strain capacity.

Table 6.11: Compressive strain and strength test results

Series IndexStrain at Ultimate Compressive

Compressive strength strength

Strain (%) StD fck,cubic (MPa) StD

A

AA1 2.189 0.174 75.62 9.91AA2 2.226 0.102 71.45 5.47AA3 2.011 0.068 70.62 3.37AA4 1.866 0.125 61.01 1.69AB1 1.896 0.108 64.16 4.76AB2 2.000 0.159 61.70 3.33AB3 1.883 0.173 57.57 4.26AB4 2.020 0.391 56.42 4.10

B

PA1 2.401 0.276 44.48 1.79PA2 2.268 0.371 37.98 1.92PB1 1.578 0.177 34.28 1.41PB2 2.165 0.161 33.04 0.79

142

Figure 6.19: The effect of fibre volume ratio on strain and compressive strength, Series A

143

Modulus of Elasticity

According to Euro Standard 1992-1 (de Normalisation, 2015), the static modulus of elasticity

can be determined as:

Ecm = 22

(fcm10

)0.3

(6.4.1)

where fcm is the mean value of cylinder compressive strength at 28 days in MPa. The relation-

ship between fcm and cubic compressive strength (fck,cubic) can be approximated by:

fcm = 0.4381f1.135ck,cubic + 11.123 (6.4.2)

Similarly, the definition of elastic modulus from the compressive strength is (de Normalisation,

2015):

Eexp,cm =0.4fcmε0.4c

(6.4.3)

where ε0.4c is the compressive strain corresponding to the 40% of fcm. Using, Equations (6.4.1)

and (6.4.3), the average modulus of elasticity of each mix was determined, as shown in Table

6.12. The results of the modulus of elasticity from the experimental and Euro standard 1992

are different ratio in range of 9.5 to 10 for Series A and 11 to 14 for Series B. The reason of

these different could be due to the size effect of the specimens (Kim and Yi, 2002).

Table 6.12: Modulus of Elasticity

Series IndexEN 1992 Experimental Ratio

Ecm (GPa) Eexp,cm (GPa) Ecm/Eexp,cm

A

AA1 39.21 3.97 9.880AA2 38.71 3.89 9.950AA3 38.62 4.00 9.660AA4 37.42 3.98 9.400AB1 37.84 3.92 9.650AB2 37.51 3.79 9.900AB3 36.80 3.64 10.110AB4 36.57 3.51 10.420

B

PA1 33.99 2.74 12.410PA2 32.92 2.35 14.010PA3 32.27 2.90 11.130PA4 32.04 2.46 13.020

144

6.5 Flexural Performance

The flexural behaviour of composites will exhibit deflection-hardening, or softening behaviour

and loading capacity after first cracking. The first cracking point of the composite is defined

as Limit of Proportionality (LOP ), and the maximum equivalent flexural strength point of the

composite is defined as Modulus of Rupture (MOR). The equivalent flexural strength can be

determined by using Equation (6.5.1) with the loading capacity at LOP (PLOP ) and at MOR

(PMOR) provided by ASTM ASTM Standard C1609 (2012).

f =PL

bd2(6.5.1)

where L is the span length, b is the average width, and d is the average depth of specimen. In

addition, flexural toughness index T10 correlates approximately with the total energy absorption

capacity of the beams. T10 can be determined as the total area under the load-deflection curve

up to 10% of the maximum load (Ward and Li, 1991).

Figure 6.20 shows the flexural behaviour of typical AAFA composites in Group of Series A.

The flexural performance of AA1 mix shows typical deflection-softening behaviour, AA2 mix

shows quasi-deflection-softening behaviour, and AA4 mix shows deflection-hardening behaviour.

However for AA3 mix, some of the specimens have complex behaviour, which are deflection-

hardening and quasi-deflection-softening behaviour. The maximum loading capacity of AA4

mix is about 74% greater than that of other mixes and the deflection capacity of AA4 mix is

also greater than that of AA1, AA2 and AA3 mixes. Similarly, Figure 6.21 shows the flexural

behaviour of typical AAFA composites in Group B of Series A. The flexural performance of

AB1, AB2 and AB3 mixes shows typical deflection-softening behaviour, and AB4 mix presents

deflection-hardening behaviour. Maximum loading and deflection capacity of mix AB4 is around

65% and 85% greater than other mixes, respectively.

Similarly, the flexural behaviour of typical OPC composites is illustrated in Figure 6.22. The

flexural results of all mixtures show deflection-hardening behaviour. Maximum loading capacity

of PA2 has around 60% higher than other composites. The maximum deflection capacity of

PA2 composite mix is the highest compared to others.

145

Figure 6.20: Flexural behaviour of AAFA composites in Group A of Series A

Figure 6.21: Flexural behaviour of AAFA composites in Group B of Series A

146

Figure 6.22: Flexural behaviour of OPC composite, Series P

As the volume fraction of fibre content of AAFA composites increase from 0% to 2.0%, Figure

6.23 shows the effect of fibre volume fraction on the deflection capacity of different mixtures of

AAFA composites. These results show an increasing trend of deflection capacity at LOP as the

linear relationship, as shown in Figure 6.23(a), and an increasing trend of deflection capacity

at MOR, as the exponential relationship, as shown in Figure 6.23(b). The improvement of

deflection at MOR, in Group A of Series A was observed much higher deflection capacity

compared with Group B of Series A. As shown in Figure 6.24(a), the increase in fibre volume

fraction leads to an increase in flexural strength at LOP as a linear relationship. The increasing

trend of flexural strength at MOR was observed as the exponential relationship as shown in

Figure 6.24(b). The improvement of flexural strength at MOR in Group A of Series A was

obtained much higher flexural strength capacity at MOR while increasing fibre volume fraction.

Toughness (T10) was determined by trapezoidal numeral integration method with MATLAB

(2014). The influence of fibre volume fraction on toughness is presented in Figure 6.25. T10 of

mixtures with different fibre volume fraction ratio shows increasing trend as a power trend in

both Group A and B of Series A. The overall toughness in Group A of Series A is higher than

Group B of Series A was observed.

147

Figure 6.23: Effect of fibre volume fraction on deflection capacity in AAFA composites, Series A

148

Figure 6.24: Effect of fibre volume fraction on flexural strength in AAFA composites, Series A

Figure 6.25: Effect of fibre volume fraction on toughness in AAFA composites, Series A

149

Similarly, Figure 6.26 shows the effect of fibre volume fraction on the deflection capacity of

OPC composites having the volumetric fibre content 1.0 and 2.0%. The results show that

deflection capacity at LOP and MOR is increased by increasing the fibre volume content. It

can be noticed that fly ash content in OPC matrix does not improve the deflection behaviour

of OPC composites. As shown in Figure 6.27, an increase in volumetric fibre fraction indicates

an increase in flexural strength at LOP and MOR. The improvement of the flexural strength

at MOR in Group A of Series A is evident with an increase in fibre volume fraction. In Figure

6.28 shows the results of T10 with different fibre volume fraction ratio it can be seen that fly ash

content in OPC composites leads to an increase in toughness when the fibre volume fraction as

2.0%. The flexural response can be tested AAFA and OPC mixtures are summarised in Table

6.13.

Li et al. (1995) reported that adding fine aggregates in OPC composite can improve the pseudo-

strain hardening behaviours. However in AAFA composite, adding fine aggregates (SF) does

not improve flexural deflection and strength capacity. Adding Fine aggregate (SF) in AAFA

composite could decrease the flexural strength and increase the toughness of matrix containing

2.0% of fibre volume fraction. There is no improvement in the flexural deflection and the

strength capacity of the mixtures containing fly ash.

Figure 6.26: Effect of fibre volume fraction on deflection of OPC composites, Series P

150

Figure 6.27: Effect of fibre volume fraction on flexural strength of OPC composites, Series P

Figure 6.28: Effect of fibre volume fraction on toughness of OPC composites, Series P

151

Table

6.13:

Fle

xura

lb

ehav

iours

Ser

ies

Ind

exL

imit

ofP

rop

orti

onal

ity

(LOP

)M

od

ulu

sof

Ru

ptu

re(M

OR

)T

ough

nes

s

δ LOP

StD

f LOP

StD

δ MOR

StD

f MOR

StD

T10

(kJ)

StD

(mm

)(M

Pa)

(mm

)(M

Pa)

A

AA

10.1

410.

009

1.23

40.

111

0.14

10.

009

1.23

40.

111

--

AA

20.

176

0.0

641.

303

0.25

60.

176

0.06

41.

303

0.25

60.

048

0.01

9A

A3

0.20

30.0

611.

357

0.13

30.

527

0.31

51.

514

0.12

20.

266

0.08

5A

A4

0.24

20.1

031.

397

0.22

22.

647

0.33

45.

214

1.00

73.

438

0.86

5A

B1

0.1

460.

022

1.36

80.

161

0.14

60.

022

1.36

80.

161

--

AB

20.1

630.

028

1.63

10.

190

0.16

30.

028

1.63

10.

190

0.04

40.

013

AB

30.2

100.

076

1.72

50.

184

0.21

00.

076

1.72

50.

184

0.05

20.

013

AB

40.2

270.

111

1.89

00.

266

1.20

00.

380

4.62

50.

839

1.41

20.

401

P

PA

10.

454

0.13

64.

177

0.49

00.

966

0.13

06.

383

0.72

81.

536

0.31

4P

A2

0.59

50.1

195.

315

1.32

21.

860

0.16

911

.154

1.35

25.

073

0.72

5P

B1

0.50

40.

023

4.54

70.

379

0.95

00.

220

6.62

10.

797

1.47

80.

434

PB

20.

546

0.086

6.05

30.

615

1.18

80.

257

9.03

20.

744

2.87

71.

175

152

The tensile and compressive behaviour of a composite material can theoretically present the

flexural performance (Naaman, 1972; Naaman and Reinhardt, 2008; Naaman et al., 1974; So-

ranakom and Mobasher, 2008; Ward and Li, 1991). When the flexural behaviour of the compos-

ite is strongly associated with its tensile behaviour, the strain-hardening behaviour in tension

leads to a deflection-hardening behaviour in flexural (Kim et al., 2011). Therefore, the tensile

strength of the composite is related with the flexural performance of the composite. In this

study, the flexural test was conducted that failure mode of AA4 and AB4 composites has strain-

hardening behaviour and AA3 has a combination of quasi-strain-softening or strain-hardening

behaviour. Also, OPC composites show strain-hardening behaviour.

Based on the theoretical discussion in the previous section, the critical energy release rate (Gc)

and interfacial bond strength (τ) of the composite are important parameters to be considered

in matrix design to achieve the strain-hardening behaviour of the composite. The matrix

properties, such as elastic modulus and fracture toughness, are linked to the composite critical

energy release rate Gc and are affected by several parameters. In this research, the composite

critical energy release rate Gc of AAFA and OPC matrices were found to be 0.010 kJ/m2 and

0.005 kJ/m2, respectively. Base on nanoindentation test as presented in Chapter 3 and 4. In

Figure 6.29 and Figure 6.30, the test results of Series A and P composites are plotted against

the critical fibre volume fraction ratio and the corresponding strain-hardening behaviour of the

composite, i.e., f is snubbing coefficient which is in term of inclining angle between fibre and

matrix. It can be seen that with 2.0% of fibre volume fraction of AAFA composites, AA4 and

AB4 are in the region of strain-hardening, whereas with less than 0.5% of fibre volume fraction

of the AAFA composites, AA2 and AB2 are not in the region of strain-hardening. It can also be

noticed that with 1.0% of fibre volume fraction of AAFA composites, AA3 and AB3, are partly

in the region of strain-hardening and other parts are fall in the region of strain-hardening. This

is consistent with AA3 composites, which shows a combination of strain-hardening and quasi-

strain-hardening behaviour. It can be seen that the fibre volume fraction of OPC composites,

PA and PB, are in the region of strain-hardening. The fact that strain-hardening behaviour

was achieved in AAFA with 2.0% of fibre volume fraction in the matrix, and OPC with 1.0% of

fibre volume fraction in OPC matrix, the experimental results are consistent with the theoretical

estimation.

153

Figure 6.29: Critical volume fraction against interfacial bond strength with AAFA composites,Series A

Figure 6.30: Critical volume fraction against interfacial bond strength with OPC composites, SeriesP

154

6.6 Chapter Summary

The results of the experimental research on the effect of parameters on fibre pull-out test and

strain-hardening behaviour of the AAFA and OPC composite are presented in this Chapter.

Based on the theoretical discussion of HPFRCC and the properties of matrix presented in

Chapter 3 and 4, the following conclusion can be drawn.

- Interfacial bond strength was determined to be in a range of 0.8 to 1.0 MPa in both OPC

and AAFA matrices.

- An increase of fibre diameter and embedded length could increase interfacial bond strength

between the matrix and the fibre.

- Fine material (SF) in AAFA composites is not suitable. It could decrease the flexural

strength and strain capacity of AAFA composites.

- While achieving the strain-hardening behaviour of the composites, the compressive strength

was decreased.

- The composite critical energy release rate Gc in OPC and AAFA matrix was found using

nanoindentation to be approximately 0.01 kJ/m2. Theoretically, it is impossible to reach

strain-hardening behaviour when Gc is more than 0.015 kJ/m2. Thus, it is recommended

Gc should be less than 0.01 kJ/m2. The results of flexural performance show the strain-

hardening behaviour of AAFA and OPC composites with this value of Gc.

155

Chapter 7

Conclusion and Future Research

This research examines properties of cementitious materials such as Ordinary Portland cement

(OPC) and Alkali-activated fly ash cement (AAFA) and high performance fibre reinforced ce-

mentitious composite. With objective of understanding of the basic properties of cementitious

materials, statistical experiment were performed on OPC and AAFA pastes and mortar with

varying parameters. Indentation properties such as modulus, hardness, packing density, vis-

coelastic properties, stress-strain curves and fracture toughness of cementitious materials were

determined using nanoindentation. With these properties, high-performance fibre reinforced

cementitious composite was studied.

7.1 Summary of Main Findings

From statistical analysis such as Taguchi’s design of experiment, Analysis of variance and

regression, the effect of test factors on properties of cementitious materials were determined.

Based on this study, the following conclusion can be drawn:

a) Nanoindentation technology is used to obtain mechanical properties of reaction products

of blended cements and alkali-activated fly ash cement. Statistical analysis tools is success-

fully used for analysis of indentation results. This technology encourages consideration

of small scales examinate to represent the large scale of civil engineering construction

materials. It is associated with scientific benefits such as comparable to conventional

experimental results, assessment of the microporomechanics and viscous properties of re-

action products of blended cement and alkali-activated fly ash cement, and determinants

of properties in nanoscale enable an assessment of the long-term macroscopic behaviour.

b) OPC is the main binder material in civil engineering thus properties of hydration prod-

ucts of OPC paste were determined using nanoindentation. The structure of CSH govern

fundamental properties such as strength, relaxation, creep and fracture behaviour. The

average of modulus (M) and hardness (H) and packing density (η) of low density CSH

are 16.787 GPa, 0.704 GPa and 0.556, respectively. Indentation properties of high density

of CSH are M = 30.481 GPa, H = 1.415 GPa and η = 0.595. Also, stress relaxation

can be negligible in OPC paste and reducing of elastic modulus due to contact relax-

ation modulus of OPC occurs in very short period. The creep behaviour of OPC paste

mainly occurs capillary porosity which tends to increase creep compliance. From indenta-

tion stress-strain curves, strength failure of OPC paste arise in capillary porosity because

small strain capacity was captured compared to other phases and high capillarity porosity

exhibit low compressive strength. Based on statistical analysis on the results of blended

156

cement mixtures, an increase in fly ash content and a decrease of water to cementitious

material ratio lead to a decrease the density of blended cement. The optimal mix design

for density is 20% fly ash, no content of sand to cementitious material ratio, 0.4 of water

to cementitious material ratio, and 0.1% of superplasticiser. For compressive strength,

an increase in fly ash content and sand to cementitious material ratio decrease the com-

pressive strength development. The optimisation of the compressive strength design is

found to be 20% of fly ash, 1.5 of sand to cementitious material ratio, 0.35 of water to

cementitious material ratio, and 0.2% of superplasticiser. For residual strength of blended

cement, increasing in fly ash and superplasticiser improve the overall residual strength.

The optimal mix design for residual strength, 20% of fly ash, 1.5 of cementitious material

ratio, 0.35 of water to cementitious material ratio, and 0.2% of superplasticiser. The

regression model is proposed to estimate the properties of blended cement mixture.

c) From statistical analysis of AAFA mixtures, silica fume is the most adverse impact on the

compressive strength and increase of superplasticiser and liquid to solid ratio contribute

to the decrease in compressive strength. It is found that no significant effect of sand to

cementitious material ratio on compressive strength. The optimum mix of compressive

strength is no silica fume, no sand, and no superplasticiser with liquid to solid ratio

0.6. In term of density, the increase in liquid to solid ratio and superplasticiser dosage

further decreases the density of AAFA. Properties of reaction products are determined

to be understood Druker-Parker strength and Coulomb material models. It is found that

stiffness and cohesion of reaction products of AAFA have an inverse relationship, i.e., an

increase of stiffness leads to a decrease of cohesion. Four main reaction phases viz., N-A-S-

H, Partly-activated slag, Non-activated slag and Non-activated compact glass phases are

identified. It confirms that N-A-S-H phase is the major reaction products of AAFA and

its volume fraction relates to the compressive strength. The volume fraction of N-A-S-H

phase of 0.50 or over and 0.6 of activation degree is recommended to obtain high strength

of AAFA. Also, it is found that stress relaxation and reduction of elastic modulus of AAFA

can be negligible due to short period of reduction of relaxation modulus. Partly-activated

and Non-activated phases are the main phases an increase of creep behaviour of AAFA

due to “block-polymerisation” concept. Liquid to solid ratio has the most effect on creep

behaviour of AAFA and it leads to an increase of creep behaviour. Sand to cementitious

material ratio and superplasticiser have a minor effect on creep behaviour. In term of

fracture toughness, an increase of silica fume leads to an increase of fracture toughness.

The overall fracture toughness of AAFA is 0.474 MPa m1/2.

d) Based on properties of cementitious materials from statistic and indentation test results,

high performance fibre reinforce cementitious composite is studied. In term of fibre pull

out test, interfacial bond strengths is determined in the range of 0.8 to 1.0 MPa for

OPC and AAFA matrix. An increase of fibre diameter and embedded length leads to

an increase of interfacial bond strength between matrix and fibre. Fine aggregate (silica

fume) should be avoided in AAFA matrix due to decrease of flexural strengths and strain

capacity. It confirms that critical energy release rate of OPC and AAFA matrix are

0.010 kJ/m2 and 0.005 kJ/m2. These critical energy release rate can be used for design

high-performance fibre reinforce cementitious composite. To achieve strain-hardening

behaviour, fibre volume fraction is to be more than 1% in OPC, and 2% in AAFA.

157

7.2 Recommendation for Future Research

A better understanding of the properties of cementitious materials such as mechanical and in-

dentation properties is necessary to achieve the required performance. Though this research

provides valuable mechanical and indentation properties using statistical analysis, the outcome

of properties of cementitious material is still limited to the available technique used. Fur-

ther verification using different technique and source materials will complement the findings

presented in this research. Viscoelastic properties are important material characteristics for

civil engineering structure. This results provide a model and indentation properties to predict

viscoelastic properties of OPC and AAFA composite. The same procedure can be used vis-

coelastic properties of other construction materials. Other test methods are recommended for

further verification. High performance reinforced alkali-activated cement composites have been

achieved using PVA fibres. More studies are recommended using different type of fibre and

mixtures.

158

Appendices

Appendix A

This appendix contains the nanoindentation results of Ordinary Portland cement paste.

Table A1: Indentation test of OPC paste

Modulus Hardness Drift Correction Displacement Load

GPa GPa nm/s nm mN

33.074 1.731 0.067 174.2 0.972

16.426 0.729 0.018 261.326 0.959

13.99 0.353 0.013 358.844 0.971

17.04 0.239 -0.001 420.592 0.961

14.171 0.397 -0.014 339.852 0.965

13.067 0.467 -0.02 319.797 0.964

38.459 1.365 0.041 187.278 0.971

21.792 0.727 0.02 255.249 0.97

31.104 0.388 0.036 329.782 0.972

17.124 0.368 -0.013 345.814 0.96

6.589 0.172 -0.044 514.782 0.966

21.09 0.438 -0.007 315.91 0.96

18.867 0.443 -0.004 317.475 0.965

42.001 2.451 0.01 148.665 0.972

108.051 8.391 0.015 84.689 0.968

18.21 0.702 -0.001 262.456 0.962

27.672 0.874 0.004 231.626 0.97

11.883 0.45 -0.053 327.443 0.961

12.104 0.334 -0.034 370.976 0.969

11.115 0.195 -0.043 469.898 0.96

12.462 0.184 -0.029 481.824 0.967

12.566 0.425 -0.032 333.587 0.963

12.356 0.263 -0.046 410.048 0.965

16.384 0.573 -0.029 287.41 0.96

22.327 0.588 -0.032 277.083 0.962

16.215 0.478 -0.035 310.622 0.963

33.133 0.852 0.002 229.585 0.962

213.215 17.474 0.033 60.309 0.972

26.558 0.399 -0.068 326.468 0.966

52.961 0.418 -0.003 312.726 0.969

159

8.946 0.238 -0.076 437.513 0.964

26.619 0.36 -0.019 342.003 0.965

11.374 0.206 -0.064 457.253 0.958

12.328 0.355 -0.067 360.378 0.965

16.568 0.415 -0.036 330.647 0.971

12.003 0.27 -0.097 405.571 0.964

53.949 1.865 0.002 159.484 0.968

20.265 0.702 -0.042 259.491 0.96

15.481 0.153 -0.032 521.427 0.967

68.785 9.177 0.003 92.439 0.969

12.349 0.334 -0.064 368.991 0.962

13.99 0.604 -0.039 288.206 0.972

14.1 0.373 -0.015 350.29 0.971

11.921 0.276 -0.033 401.895 0.964

17.62 0.444 -0.021 319.097 0.968

22.269 0.635 -0.04 268.259 0.961

20.067 0.451 -0.042 314.747 0.971

8.644 0.209 -0.094 463.994 0.966

15.158 0.65 -0.062 276.524 0.964

63.959 3.216 0.017 126.631 0.966

143.139 14.904 0.028 68.426 0.969

7.411 0.153 -0.112 536.15 0.964

239.257 17.899 0.027 58.355 0.964

24.151 0.923 -0.038 228.51 0.962

26.952 0.489 -0.021 298.392 0.972

37.651 1.252 0.002 193.92 0.967

26.191 0.926 -0.016 226.361 0.962

19.715 0.714 -0.052 259.879 0.971

29.591 0.511 0.007 290.203 0.966

37.942 1.199 -0.024 197.466 0.969

31.629 0.811 0.001 236.664 0.973

103.127 4.041 0.034 109.931 0.971

14.997 0.259 -0.056 407.513 0.961

7.168 0.181 -0.092 499.726 0.963

13.939 0.579 -0.053 293.013 0.971

25.255 0.673 -0.015 259.375 0.963

13.971 0.309 -0.03 378.105 0.96

10.477 0.233 -0.056 436.979 0.964

64.489 2.698 0.03 135.47 0.971

88.459 2.33 -0.017 139.171 0.966

23.083 0.789 -0.03 244.861 0.964

14.457 0.423 -0.023 330.45 0.966

19.158 0.649 -0.047 269.017 0.959

15.494 0.387 -0.032 341.817 0.967

11.923 0.379 -0.059 351.685 0.965

17.901 0.559 -0.034 288.532 0.963

9.467 0.187 -0.07 482.605 0.958

160

9.538 0.158 -0.091 521.714 0.964

8.971 0.17 -0.075 504.661 0.958

171.068 5.412 0.02 65.214 0.482

14.832 0.438 -0.043 323.831 0.96

19.869 0.356 -0.028 348.71 0.964

14.404 0.289 -0.056 390.262 0.968

41.219 2.162 -0.017 155.553 0.968

320.036 10.325 0.036 67.801 0.965

28.846 1.164 -0.017 205.067 0.966

24.496 0.986 -0.016 222.939 0.967

13.219 0.272 -0.061 403.003 0.969

8.362 0.132 -0.087 571.744 0.968

12.007 0.316 -0.04 380.199 0.97

15.891 0.528 -0.038 298.526 0.963

52.297 2.211 0.024 149.551 0.967

19.428 0.715 -0.006 259.557 0.967

6.238 0.177 -0.118 509.159 0.962

21.962 0.587 -0.047 277.648 0.962

110.011 11.139 0.01 78.249 0.97

36.309 0.68 0.023 253.361 0.971

25.368 0.616 -0.003 270.461 0.971

14.701 0.449 -0.033 322.146 0.966

9.718 0.308 -0.113 390.98 0.97

11.152 0.292 -0.082 392.888 0.958

14.892 0.366 -0.054 350.543 0.965

11.53 0.306 -0.072 384.41 0.959

12.171 0.221 -0.027 442.43 0.96

12.41 0.252 -0.049 417.308 0.964

17.055 0.704 -0.046 263.813 0.96

38.74 0.708 -0.003 247.702 0.971

10.505 0.235 -0.043 436.334 0.969

16.616 0.539 -0.038 295.189 0.966

11.199 0.153 -0.067 524.921 0.96

9.26 0.231 -0.063 441.336 0.962

13.425 0.327 -0.047 370.267 0.962

25.675 0.831 -0.037 237.745 0.968

21.345 0.481 -0.021 303.031 0.96

41.704 1.641 0.007 172.136 0.966

7.386 0.151 -0.087 538.863 0.96

24.179 0.716 -0.013 254.505 0.97

79.862 3.888 0.022 114.452 0.961

21.219 0.303 -0.032 375.409 0.971

18.052 0.589 -0.027 281.758 0.96

22.789 0.71 -0.02 255.274 0.959

5.528 0.07 -0.064 775.834 0.963

10.864 0.26 -0.066 416.634 0.971

15.275 0.365 -0.037 349.343 0.96

161

14.225 0.375 -0.044 349.19 0.971

28.808 0.706 -0.02 251.911 0.965

7.374 0.166 -0.112 518.918 0.967

12.468 0.311 -0.045 379.939 0.96

11.23 0.201 -0.014 463.787 0.962

14.191 0.412 -0.075 335.125 0.969

18.941 0.559 -0.042 288.246 0.971

8.771 0.212 -0.082 461.641 0.967

18.38 0.17 -0.055 491.515 0.961

63.245 3.057 0.013 129.418 0.969

17.667 0.524 -0.059 296.336 0.96

29.298 0.735 -0.026 248.072 0.971

9.711 0.195 -0.085 475.577 0.967

15.325 0.528 -0.037 299.508 0.963

43.962 2.051 -0.001 157.428 0.971

23.143 0.72 -0.022 254.876 0.97

33.095 3.497 -0.007 139.882 0.97

13.509 0.351 -0.026 358.46 0.959

25.147 0.57 -0.024 279.74 0.969

28.73 0.731 -0.012 247.529 0.96

7.756 0.086 -0.227 697.816 0.967

21.549 0.636 -0.039 270.139 0.97

79.983 1.083 0.045 138.056 0.479

3.725 0.062 -0.349 832.771 0.959

37.571 0.193 -0.01 456.747 0.967

10.667 0.195 -0.06 471.423 0.963

4.883 0.124 -0.149 602.86 0.958

31.781 1.098 -0.009 208.404 0.972

11.456 0.391 -0.054 348.607 0.965

9.999 0.263 -0.058 416.348 0.968

11.216 0.26 -0.035 414.2 0.962

21.563 0.602 -0.062 274.754 0.959

48.034 0.959 0.033 213.681 0.969

13.806 0.223 -0.021 439.818 0.97

11.775 0.31 -0.064 384.224 0.97

10.375 0.093 -0.061 666.406 0.967

22.895 0.698 -0.034 257.493 0.963

15.67 0.321 -0.031 369.13 0.96

15.505 0.212 -0.039 447.999 0.971

5.276 0.066 -0.092 794.716 0.959

8.978 0.182 -0.075 492.337 0.969

18.608 1.202 -0.01 215.114 0.964

35.187 3.142 0.002 141.585 0.966

12.514 0.219 -0.112 443.412 0.961

24.955 1.069 -0.019 215.692 0.968

64.564 1.478 0.023 173.184 0.967

17.382 0.577 -0.05 286.407 0.969

162

14.347 0.352 -0.08 356.925 0.962

31.476 1.339 -0.037 191.877 0.962

26.368 0.844 -0.024 234.846 0.961

9.35 0.227 -0.051 445.24 0.964

26.293 0.277 -0.011 386.072 0.96

13.007 0.206 -0.038 454.588 0.959

14.618 0.264 -0.036 405.17 0.965

12.356 0.263 -0.046 410.048 0.965

16.384 0.573 -0.029 287.41 0.96

22.327 0.588 -0.032 277.083 0.962

16.215 0.478 -0.035 310.622 0.963

33.133 0.852 0.002 229.585 0.962

213.215 17.474 0.033 60.309 0.972

26.558 0.399 -0.068 326.468 0.966

52.961 0.418 -0.003 312.726 0.969

8.946 0.238 -0.076 437.513 0.964

26.619 0.36 -0.019 342.003 0.965

11.374 0.206 -0.064 457.253 0.958

12.328 0.355 -0.067 360.378 0.965

16.568 0.415 -0.036 330.647 0.971

12.003 0.27 -0.097 405.571 0.964

53.949 1.865 0.002 159.484 0.968

20.067 0.451 -0.042 314.747 0.971

8.644 0.209 -0.094 463.994 0.966

15.158 0.65 -0.062 276.524 0.964

17.901 0.559 -0.034 288.532 0.963

9.467 0.187 -0.07 482.605 0.958

9.538 0.158 -0.091 521.714 0.964

Appendix B

This appendix contains the nanoindentation results of Alkali-activated fly ash cement.

Table B1: Indentation test of AAFA mix 1

Modulus Hardness Drift Correction Displacement Load

GPa GPa nm/s nm mN

5.704 0.091 0.507 688.717 0.971

16.203 0.451 0.505 318.436 0.963

58.588 7.23 0.608 101.607 0.969

73.975 9.24 0.538 90.616 0.974

15.113 0.813 0.456 254.66 0.965

22.489 4.181 0.482 151.609 0.964

35.025 5.37 0.486 125.149 0.962

17.595 0.637 0.416 274.252 0.965

14.962 0.468 0.392 315.691 0.965

163

15.583 0.525 0.373 300.505 0.967

11.498 0.445 0.348 330.411 0.964

14.291 0.616 0.345 284.745 0.967

11.379 0.501 0.298 316.735 0.969

10.983 0.213 0.33 452.874 0.965

32.541 3.128 0.362 144.07 0.963

12.981 1.035 0.312 241.274 0.967

66.244 7.507 0.355 97.567 0.971

10.003 0.374 0.199 360.351 0.969

10.17 0.302 0.248 391.119 0.962

46.893 2.42 0.334 147.097 0.974

37.716 1.859 0.319 166.227 0.967

29.234 1.961 0.298 169.136 0.961

51.216 9.077 0.316 101.91 0.97

51.159 4.88 0.297 115.792 0.972

13.221 0.647 0.224 281.825 0.965

18.985 3.932 0.278 162.671 0.962

32.811 3.961 0.229 135.604 0.963

10.654 0.321 0.184 380.206 0.963

64.88 4.695 0.255 110.996 0.963

10.922 0.478 0.164 324.384 0.97

11.546 0.409 0.161 341.244 0.96

8.549 0.238 0.172 439.631 0.966

28.237 0.86 0.207 232.779 0.971

14.489 2.052 0.219 197.308 0.963

29.53 3.809 0.216 141.03 0.965

52.621 8.321 0.221 102.348 0.972

41.732 3.582 0.215 131.944 0.972

74.724 10.121 0.209 88.551 0.97

13.804 0.675 0.156 275.132 0.96

28.558 2.619 0.18 155.618 0.961

18.465 0.847 0.154 243.607 0.961

67 3.06 0.199 128.458 0.97

49.957 7.726 0.198 105.362 0.971

9.732 0.487 0.123 326.706 0.969

41.785 3.39 0.192 133.389 0.963

23.528 1.011 0.149 222.507 0.973

9.847 0.459 0.073 333.025 0.966

11.896 0.51 0.102 312.051 0.963

11.555 0.325 0.112 376.524 0.968

7.873 0.364 0.082 373.051 0.962

21.323 0.494 0.116 300.352 0.965

11.172 0.438 0.081 332.996 0.96

16.085 1.324 0.143 214.716 0.967

21.405 1.017 0.153 224.444 0.972

14.255 0.567 0.081 293.379 0.963

15.493 0.638 0.101 277.193 0.96

164

13.782 0.579 0.089 292.228 0.964

10.286 0.453 0.045 332.83 0.966

89.355 11.026 0.166 82.612 0.968

33.881 1.848 0.129 169.076 0.967

17.661 0.71 0.074 261.911 0.96

32.409 3.774 0.142 137.743 0.965

38.14 3.295 0.132 136.955 0.961

16.85 0.789 0.09 253.176 0.962

29.689 1.637 0.15 180 0.967

17.75 0.69 0.082 264.976 0.962

37.877 6.391 0.153 118.566 0.962

35.052 1.395 0.12 187.291 0.969

17.913 1.056 0.091 226.739 0.969

18.989 0.878 0.068 240.687 0.97

12.915 0.536 0.065 304.527 0.971

16.508 0.694 0.061 266.16 0.96

13.373 0.734 0.048 268.74 0.965

20.424 2.373 0.092 173.302 0.962

51.285 6.919 0.131 106.563 0.97

25.475 1.71 0.115 182.214 0.972

6.306 0.084 -0.08 706.358 0.956

15.554 1.564 0.079 206.053 0.964

31.877 1.864 0.115 170.332 0.969

86.901 11.511 0.129 82.55 0.968

61.84 7.972 0.121 98.249 0.973

19.8 4.036 0.096 159.463 0.961

19.059 3.404 0.08 165.213 0.96

19.771 0.823 0.054 244.312 0.961

16.288 0.587 0.097 286.698 0.972

17.695 0.824 0.084 247.246 0.96

13.476 0.686 0.02 275.133 0.965

39.549 5.28 0.098 120.946 0.962

14.191 0.584 0.03 290.175 0.963

12.524 0.502 0.024 312.17 0.963

83.372 11.545 0.116 83.512 0.968

22.823 1.598 0.084 189.727 0.97

11.876 0.232 0.007 432.94 0.957

20.577 2.963 0.105 164.918 0.962

57.653 6.994 0.097 102.995 0.973

41.455 7.281 0.111 112.775 0.963

20.256 0.665 0.077 266.232 0.967

32.879 5.314 0.1 128.336 0.967

62.502 8.915 0.077 95.654 0.969

40.838 2.159 0.076 155.77 0.968

28.312 0.678 -0.081 256.181 0.961

8.371 0.128 -0.15 577.917 0.961

9.544 0.321 -0.15 384.198 0.964

165

19.071 0.559 -0.133 286.315 0.959

9.187 0.407 -0.162 351.491 0.967

18.287 0.82 -0.127 246.887 0.96

10.065 0.257 -0.179 420.25 0.967

54.82 6.093 -0.107 107.613 0.971

19.424 0.794 -0.096 248.523 0.963

10.585 0.727 -0.151 280.673 0.97

46.839 5.954 -0.091 112.846 0.971

8.945 0.396 -0.185 355.823 0.963

5.665 0.263 -0.254 439.899 0.964

6.699 0.414 -0.208 364.545 0.963

8.966 0.464 -0.203 335.776 0.967

14.219 0.321 -0.179 371.863 0.963

7.147 0.371 -0.212 374.483 0.96

19.353 0.79 -0.132 248.851 0.962

39.918 5.642 -0.115 119.537 0.97

11.511 0.466 -0.189 323.854 0.96

20.983 2.828 -0.141 165.662 0.964

7.168 0.275 -0.253 420.009 0.963

14.794 0.951 -0.149 241.182 0.96

10.627 0.636 -0.202 293.192 0.969

13.454 0.321 -0.188 373.005 0.963

10.161 0.538 -0.211 313.278 0.97

7.181 0.328 -0.232 393.617 0.968

3.45 0.081 -0.415 742.565 0.955

8.442 0.429 -0.228 347.608 0.962

56.197 8.192 -0.142 100.215 0.966

12.991 0.557 -0.216 298.56 0.963

58.295 6.962 -0.147 102.436 0.966

17.275 0.799 -0.198 250.836 0.959

13.72 0.437 -0.217 327.724 0.966

10.397 0.342 -0.238 370.367 0.96

35.333 2.637 -0.147 148.673 0.961

15.198 0.789 -0.19 257.65 0.969

54.871 7.037 -0.151 104.219 0.971

9.014 0.521 -0.233 320.871 0.96

7.713 0.236 -0.292 444.888 0.968

24.498 2.49 -0.149 163.907 0.967

10.037 0.499 -0.235 320.699 0.959

52.274 5.93 -0.147 109.642 0.971

91.062 9.486 -0.14 84.932 0.965

71.518 2.725 -0.171 133.454 0.972

47.205 6.056 -0.172 112.209 0.971

32.056 3.787 -0.154 138.288 0.969

5.768 0.17 -0.372 522.139 0.964

26.254 2.316 -0.163 164.012 0.96

63.736 7.203 -0.187 99.415 0.969

166

9.866 0.406 -0.288 348.612 0.964

33.525 3.882 -0.209 135.789 0.967

16.296 0.669 -0.224 270.313 0.959

12.468 0.832 -0.179 260.613 0.967

31.683 4.418 -0.18 134.322 0.969

12.452 0.433 -0.246 331.718 0.965

52.881 7.163 -0.18 104.959 0.971

25.954 3.577 -0.176 148.2 0.963

8.208 0.378 -0.293 366.883 0.968

23.716 1.425 -0.217 195.079 0.963

139.765 7.106 -0.156 85.716 0.968

34.183 2.403 -0.204 154.217 0.964

12.224 0.451 -0.261 326.649 0.963

50.104 2.689 -0.199 139.765 0.966

23.08 0.589 -0.239 275.883 0.959

13.718 0.607 -0.256 287.407 0.965

16.724 0.643 -0.238 275.057 0.967

7.105 0.377 -0.375 373.003 0.961

25.728 3.344 -0.235 150.421 0.96

9.484 0.436 -0.288 340.847 0.965

14.675 0.553 -0.267 295.123 0.962

13.668 0.813 -0.365 257.549 0.958

10.101 0.415 -0.291 344.345 0.962

8.099 0.472 -0.315 337.502 0.959

57.207 3.992 -0.215 119.872 0.969

31.063 1.624 -0.181 179.69 0.971

13.384 0.389 -0.306 344.001 0.964

46.9 4.241 -0.214 122.624 0.971

79.523 10.062 -0.19 86.726 0.962

73.954 7.35 -0.204 95.372 0.969

10.848 0.455 -0.31 328.698 0.957

21.023 0.427 -0.252 320.873 0.968

21.496 1.467 -0.302 196.502 0.964

36.271 4.368 -0.186 129.531 0.969

8.674 0.345 -0.336 377.207 0.966

10.147 0.504 -0.293 320.179 0.965

29.275 1.335 -0.224 194.062 0.965

14.61 0.625 -0.28 282.83 0.971

51.738 6.186 -0.218 108.891 0.971

14.636 0.492 -0.312 309.036 0.958

27.925 3.254 -0.271 147.927 0.96

10 0.439 -0.319 336.83 0.96

9.541 0.438 -0.337 339.262 0.961

23.334 1.586 -0.237 189.521 0.971

20.272 1.898 -0.229 183.428 0.959

18.315 1.057 -0.313 226.001 0.97

32.983 2.586 -0.222 151.853 0.965

167

2.135 0.011 -0.52 1937.884 0.946

5.809 0.048 -0.626 926.94 0.957

Table B2: Indentation test of AAFA mix 2

Modulus Hardness Drift Correction Displacement Load

GPa GPa nm/s nm mN

70.057 7.909 0.013 95.056 0.972

11.786 0.297 -0.06 389.661 0.962

5.084 0.106 -0.267 645.66 0.962

16.043 0.98 -0.079 235.638 0.96

18.313 0.996 -0.041 229.692 0.96

15.765 0.609 -0.085 282.547 0.965

73.772 7.817 0.004 94.095 0.972

19.343 1.082 -0.049 221.121 0.96

15.342 0.816 -0.066 253.515 0.964

38.128 4.976 0.075 123.596 0.96

12.606 0.7 -0.074 276.376 0.97

65.71 5.683 -0.014 104.961 0.97

14.653 0.758 -0.054 263.008 0.97

13.595 0.704 -0.062 271.729 0.962

27.637 1.162 -0.02 205.749 0.962

37.386 5.28 0.022 123.372 0.968

14.245 0.631 -0.032 282.134 0.967

13.401 0.597 -0.082 289.849 0.965

18.301 0.538 -0.055 293.602 0.97

13.972 0.728 -0.047 268.248 0.968

72.366 7.414 0.065 95.264 0.962

11.602 0.5 -0.069 315.343 0.963

10.292 0.427 -0.086 340.793 0.969

74.14 7.548 0.038 94.728 0.97

6.819 0.143 -0.202 556.557 0.965

6.214 0.064 0.062 811.325 0.972

13.131 0.662 -0.013 279.734 0.965

50.576 5.737 0.067 111.364 0.969

14.486 2.882 0.037 187.816 0.97

87.062 8.093 0.057 89.296 0.963

83.369 7.523 0.07 92.177 0.968

3.114 0.024 -0.101 1313.436 0.968

66.462 8.057 0.076 95.942 0.97

12.935 0.576 -0.017 294.366 0.959

79.539 7.56 0.059 92.646 0.96

19.814 1.676 0.034 192.194 0.971

11.496 0.484 -0.011 319.112 0.96

38.655 1.61 0.062 174.843 0.966

168

22.034 0.555 0.021 285.379 0.968

17.307 0.812 0.037 249.877 0.964

18.684 0.764 0.001 253.265 0.962

52.019 6.913 0.106 106.28 0.972

61.23 7.569 0.053 98.983 0.961

17.967 2.274 0.08 182.136 0.972

17.996 1.246 0.011 214.517 0.971

15.161 0.799 0.032 255.883 0.964

102.67 14.716 0.113 75.09 0.968

13.692 0.791 0.028 261.613 0.971

78.788 7.387 0.105 93.931 0.971

14.639 0.797 0.034 257.125 0.962

60.201 10.004 0.102 94.982 0.969

8.154 0.127 -0.058 581.488 0.967

9.924 0.28 0.061 406.578 0.971

12.608 0.558 0.033 300.662 0.97

24.817 1.602 0.112 186.667 0.969

92.827 9.685 0.122 84.396 0.971

13.561 0.676 0.017 277.269 0.971

16.243 0.407 0.029 333.728 0.97

34.818 3.647 0.09 136.797 0.971

11.158 0.396 0.038 348.436 0.971

8.581 0.519 -0.031 324.448 0.963

21.326 1.859 0.032 182.583 0.96

12.155 0.228 0.085 437.201 0.964

10.103 0.351 0.029 368.632 0.965

113.945 7.747 0.08 85.753 0.963

15.313 0.601 0.019 284.123 0.96

76.169 6.183 0.065 99.4 0.97

12.301 0.277 0.038 399.364 0.959

44.648 5.007 0.071 118.9 0.971

83.778 9.604 0.087 86.83 0.972

12.722 0.762 0.006 266.594 0.96

7.746 0.266 -0.041 422.987 0.964

10.879 0.676 0.024 284.909 0.961

12.767 0.621 0.033 288.16 0.97

21.25 1.226 0.065 209.498 0.967

14.928 0.738 0.056 263.803 0.963

14.039 0.722 0.044 268.102 0.963

11.175 0.379 -0.031 352.886 0.961

44.268 4.016 0.036 126.059 0.971

9.176 0.42 0.002 346.608 0.962

33.524 2.146 0.114 161.064 0.969

13.082 0.718 0.02 272.466 0.97

12.078 0.527 0.011 307.195 0.961

18.791 0.459 0.064 312.219 0.963

14.994 0.877 0.082 247.648 0.961

169

12.204 0.524 -0.013 307.986 0.963

13.611 0.655 0.015 280.073 0.97

18.095 1.012 0.023 229.115 0.963

8.915 1.037 0.051 262.787 0.964

37.562 2.701 0.009 145.786 0.96

12.76 0.52 0.008 308.253 0.97

13.74 0.7 0.035 273.064 0.97

12.793 0.663 0.014 279.735 0.959

12.452 0.641 0.014 285.003 0.965

9.363 0.272 -0.036 411.844 0.963

7.573 0.279 -0.019 415.918 0.965

62.457 2.179 0.138 148.011 0.973

14.331 0.772 0.023 262.171 0.971

12.457 0.543 0.041 302.423 0.96

36.815 1.921 0.059 165.097 0.97

15.075 0.951 -0.01 241.272 0.966

13.799 1.423 0.037 217.563 0.966

10.562 0.6 -0.022 297.963 0.959

12.604 0.589 -0.026 293.432 0.964

70.298 5.192 0.058 106.026 0.964

35.463 3.198 0.043 140.297 0.961

21.325 1.199 0.013 211.607 0.972

56.787 8.975 0.053 98.482 0.969

48.059 7.123 0.054 108.176 0.971

13.015 1.043 0.012 241.074 0.97

43.904 5.984 0.045 114.779 0.969

15.888 0.912 -0.012 242.081 0.961

30.716 1.72 0.039 176.371 0.971

9.925 0.553 -0.052 309.589 0.959

18.173 0.649 -0.035 270.707 0.96

7.33 0.345 -0.082 385.776 0.97

56.386 6.259 0.05 106.242 0.972

15.525 0.638 0.005 277.02 0.96

8.477 0.447 -0.032 342.921 0.965

35.99 2.77 0.033 145.838 0.96

3.242 0.105 -0.216 671.705 0.968

68.195 7.354 0.031 97.252 0.969

66.481 3.499 0.039 122.626 0.973

11.849 0.638 -0.031 288.19 0.969

11.8 0.899 -0.002 256.379 0.964

15.866 1.058 -0.012 231.249 0.97

9.645 0.562 -0.046 310.799 0.969

39.951 1.906 0.041 163.902 0.972

13.815 0.845 -0.012 254.428 0.964

22.247 1.305 0.013 204.068 0.972

33.433 4.371 0.035 132.611 0.971

10.935 1.535 0.01 227.766 0.965

170

12.449 0.619 -0.036 288.174 0.961

14.473 0.429 0 328.73 0.966

10.165 0.536 -0.053 311.944 0.959

37.529 3.693 0.041 134.033 0.973

14.409 0.71 -0.018 269.988 0.971

15.304 0.654 -0.019 276.559 0.972

19.946 1.008 -0.001 225.89 0.96

18.022 1.027 0.017 228.751 0.969

9.38 0.548 -0.049 313.954 0.964

10.766 0.581 -0.044 301.227 0.963

8.869 0.414 -0.044 350.526 0.965

10.478 0.56 -0.023 307.619 0.971

21.439 0.979 0.014 226.357 0.961

8.161 0.37 -0.053 368.821 0.961

11.432 0.57 -0.017 300.349 0.961

9.93 0.544 -0.021 311.623 0.961

62.457 7.207 0.043 100.009 0.97

36.013 5.051 0.023 125.875 0.969

23.295 0.744 0.038 251.052 0.968

8.844 0.445 -0.046 341.568 0.965

12.888 0.567 -0.03 296.917 0.964

15.29 0.988 -0.011 237.898 0.968

10.242 0.482 -0.03 323.994 0.959

64.098 7.306 0.057 99.028 0.971

46.246 6.329 0.035 111.962 0.972

20.875 1.812 -0.006 185.185 0.964

8.892 0.488 -0.068 329.854 0.965

33.728 5.077 0.033 127.886 0.962

10.353 0.517 -0.054 316.097 0.963

11.706 0.487 -0.027 317.618 0.959

19 0.547 -0.027 289.863 0.964

8.97 0.267 -0.081 417.503 0.968

13.32 0.682 -0.01 276.847 0.97

9.871 0.472 -0.032 328.21 0.959

14.96 0.868 -0.009 248.401 0.959

30.652 0.912 0.047 225.653 0.971

16.737 1.654 0.033 200.397 0.972

33.771 5.69 0.026 125.477 0.962

28.379 2.773 0.03 154.263 0.971

63.072 5.8 0.035 105.465 0.973

27.87 0.665 0.017 259.334 0.966

49.755 5.728 0.035 112.048 0.972

8.307 0.371 -0.059 367.687 0.961

9.602 0.381 -0.07 357.797 0.962

18.013 1.053 -0.005 226.815 0.969

9.67 0.479 -0.055 327.703 0.961

26.446 2.693 0.009 157.415 0.963

171

22.101 1.709 0.02 185.869 0.96

15.089 0.663 -0.022 273.732 0.959

12.021 0.524 -0.025 308.411 0.963

49.758 6.551 0.043 108.531 0.967

12.219 0.546 -0.032 304.246 0.971

14.34 0.658 -0.018 277.363 0.966

17.747 1.031 -0.008 229.143 0.97

96.322 8.689 0.064 85.572 0.961

17.267 1.07 -0.009 227.159 0.97

14.565 0.81 0.007 256.009 0.963

9.536 0.527 -0.045 317.524 0.963

20.322 1.507 0.004 197.504 0.971

17.719 1.063 -0.003 226.793 0.97

24.99 1.147 0.014 210.043 0.969

13.878 0.694 -0.017 272.373 0.962

13.314 0.66 -0.028 279.284 0.964

16.996 0.931 -0.007 237.782 0.96

11.956 0.583 -0.037 297.632 0.97

40.326 5.173 0.066 121.267 0.97

13.75 0.728 -0.025 268.603 0.966

Table B3: Indentation test of AAFA mix 3

Modulus Hardness Drift Correction Displacement Load

GPa GPa nm/s nm mN

35.431 7.804 0.302 118.979 0.967

39.971 1.613 0.367 174.523 0.97

33.502 2.328 0.342 156.291 0.963

24.274 1.229 0.408 205.169 0.966

12.984 1.318 0.336 225.857 0.971

13.808 0.805 0.262 258.374 0.961

5.534 0.116 0.087 619.297 0.969

30.513 2.214 0.324 162.24 0.971

11.591 0.585 0.28 297.969 0.967

12.91 1.028 0.326 242.787 0.972

49.166 3.989 0.356 122.987 0.962

63.594 6.82 0.339 100.991 0.973

38.36 3.918 0.32 131.157 0.971

12.799 0.39 0.243 346.331 0.971

19.485 1.146 0.302 217.546 0.969

74.815 10.449 0.359 87.764 0.965

82.589 8.796 0.321 88.777 0.969

35.887 1.905 0.402 166.343 0.972

74.261 6.877 0.302 96.897 0.968

17.718 2.383 0.27 180.674 0.968

172

30.651 3.949 0.279 138.299 0.963

96.834 11.528 0.308 79.809 0.962

31.078 1.989 0.299 166.738 0.963

85.463 10.477 0.267 84.36 0.964

79.529 8.029 0.284 91.507 0.966

18.627 2.695 0.254 173.846 0.969

12.072 0.78 0.171 266.916 0.961

55.114 10.135 0.247 97.651 0.967

22.937 0.952 0.231 227.97 0.969

32.48 3.038 0.214 145.885 0.971

10.262 0.449 0.145 333.112 0.96

41.233 4.448 0.236 124.863 0.972

5.862 0.062 -0.022 823.32 0.967

5.606 0.137 0.003 573.802 0.966

33.601 5.376 0.218 127.267 0.969

47.307 5.166 0.213 116.158 0.969

51.871 4.088 0.17 121.394 0.972

10.915 0.412 0.084 342.69 0.963

49.765 6.563 0.177 108.662 0.97

23.014 1.214 0.179 208.207 0.971

11.515 0.467 0.038 323.786 0.961

21.881 1.118 0.089 215.945 0.97

9.301 0.365 0.04 365.452 0.964

39.308 4.641 0.1 124.997 0.969

44.571 3.292 0.122 133.48 0.971

13.265 0.578 0.06 293.138 0.959

69.814 6.604 0.071 99.249 0.967

22.473 1.474 0.056 195.443 0.972

59.729 7.758 0.126 99.592 0.969

29.13 2.487 0.158 157.258 0.962

8.983 0.286 0.018 404.974 0.963

8.501 0.338 0.053 380.163 0.961

19.488 1.949 0.158 184.873 0.97

109.337 19.933 0.182 70.355 0.968

77.668 10.338 0.156 86.968 0.965

34.871 4.911 0.194 127.989 0.971

19.498 1.803 0.163 188.625 0.968

43.405 3.86 0.188 128.13 0.973

10.755 0.47 0.106 326.639 0.967

83.253 9.702 0.168 86.442 0.965

25.52 1.401 0.124 193.889 0.961

54.337 7.026 0.15 104.332 0.968

43.872 6.066 0.218 114.682 0.972

70.269 5.411 0.176 105.211 0.973

17.357 1.717 0.113 196.403 0.969

11.458 0.44 0.064 331.537 0.961

6.904 0.251 0.037 438.46 0.967

173

11.034 0.532 0.086 310.095 0.963

39.541 3.944 0.132 129.403 0.962

18.189 0.904 0.117 237.979 0.959

31.138 4.36 0.101 135.329 0.968

47.004 6.53 0.141 110.757 0.972

15.518 0.313 0.06 374.158 0.964

12.541 1.062 -0.023 240.709 0.963

20.083 1.756 0.126 188.435 0.964

11.323 0.876 0.166 261.551 0.972

30.641 3.568 0.073 141.644 0.965

19.92 2.435 0.131 173.246 0.96

82.665 6.213 0.191 97.314 0.963

5.665 0.136 0.002 574.172 0.96

30.332 2.738 0.069 152.535 0.972

51.755 1.602 0.131 170.805 0.973

84.298 9.782 0.185 85.942 0.964

20.334 0.403 0.076 329.461 0.965

20.219 1.673 0.039 191.082 0.967

9.261 0.196 0.021 473.853 0.962

3.581 0.032 -0.284 1130.572 0.959

3.355 0.029 -0.389 1189.193 0.951

95.944 11.162 0.127 81.008 0.972

15.021 1.293 0.04 219.45 0.969

5.571 0.114 -0.129 620.271 0.962

11.898 0.281 0.01 399.007 0.963

2.309 0.024 -0.576 1308.335 0.952

3.608 0.033 -0.467 1122.277 0.958

11.373 0.431 0.038 336.426 0.971

74.32 8.721 0.078 91.587 0.973

3.83 0.095 -0.291 686.706 0.956

59.449 10.962 0.144 94.104 0.967

9.452 0.338 0 375.827 0.96

14.03 1.499 0 213.309 0.961

20.597 0.71 0.081 258.267 0.964

53.261 6.77 0.072 105.739 0.968

22.249 2.467 0.11 167.698 0.96

14.551 0.866 -0.031 250.53 0.966

72.949 6.56 0.026 98.454 0.967

48.8 5.923 0.05 111.783 0.972

16.945 0.683 0.095 267.903 0.966

6.808 0.362 -0.078 380.869 0.961

13.565 0.733 0.059 269.334 0.972

9.743 0.348 -0.042 371.305 0.966

24.669 1.312 -0.004 200.598 0.972

117.801 16.296 0.063 70.591 0.964

4.76 0.164 -0.113 538.18 0.962

5.52 0.18 -0.082 510.863 0.96

174

5.703 0.08 -0.187 726.287 0.962

74.165 7.761 0.048 93.582 0.961

83.134 7.332 0.051 92.835 0.967

59.587 8.414 0.029 98.223 0.972

24.561 2.941 -0.002 157.363 0.968

14.465 1.422 -0.016 215.809 0.971

161.419 39.931 0.083 57.559 0.968

44.154 5.473 0.043 116.731 0.969

4.091 0.033 -0.291 1104.595 0.957

45.568 5.322 0.042 116.46 0.969

51.511 4.174 0.032 120.217 0.962

16.823 0.341 -0.045 359.524 0.97

12.494 1.036 0.003 242.32 0.96

21.802 0.811 0.014 243.136 0.961

42.035 8.169 0.032 111.024 0.971

10.909 0.345 -0.044 367.63 0.96

13.635 0.849 -0.107 253.736 0.958

11.801 0.781 -0.095 267.588 0.959

5.437 0.118 -0.137 612.275 0.958

92.704 9.763 0.046 83.982 0.965

23.592 1.638 -0.008 187.24 0.972

43.787 2.75 0.05 141.277 0.962

17.292 0.899 -0.034 240.695 0.964

9.569 0.295 -0.051 396.732 0.959

8.264 0.217 -0.121 457.229 0.961

18.399 0.956 0.035 233.022 0.961

15.829 0.589 0.007 285.238 0.96

15.454 0.614 -0.023 281.49 0.961

28.77 1.518 0.022 186.116 0.971

12.21 0.225 -0.026 440.185 0.968

35.495 3.006 0.048 143.55 0.971

6.556 0.233 -0.082 453.591 0.966

26.78 3.83 0.024 144.704 0.961

35.074 5.681 0.02 123.93 0.962

16.077 0.518 -0.011 301.644 0.97

8.974 0.27 -0.068 415.614 0.967

26.274 0.953 0.002 224.658 0.97

8.044 0.204 -0.089 470.853 0.962

4.857 0.107 -0.186 643.636 0.959

7.983 0.194 -0.128 481.099 0.961

7.958 0.425 -0.029 352.213 0.965

11.41 0.275 -0.065 404.723 0.968

19.881 0.812 0.002 245.429 0.961

20.374 0.631 0.03 270.942 0.961

21.576 0.634 -0.024 269.724 0.966

24.525 1.079 0.022 215.473 0.969

21.821 0.594 0.005 277.163 0.967

175

13.92 0.422 -0.007 332.065 0.968

50.219 4.739 0.055 117.21 0.972

24.729 4.943 0.041 143.468 0.966

9.694 0.299 -0.028 396.245 0.969

23.853 2.55 0.075 164.251 0.97

68.252 7.269 0.094 97.515 0.969

16.39 0.617 -0.04 279.318 0.961

17.22 1.445 0 206.11 0.965

25.727 1.681 0.028 182.287 0.966

14.376 0.943 0.041 243.577 0.964

67.139 7.679 0.045 96.719 0.971

35.295 6.072 0.031 122.416 0.962

57.022 6.27 0.031 105.549 0.966

40.265 4.263 0.045 126.904 0.971

13.231 0.862 -0.023 253.779 0.958

7.922 0.219 -0.098 456.155 0.959

18.211 0.496 -0.015 303.827 0.968

21.102 1.042 -0.001 221.511 0.96

15.155 0.53 -0.009 298.833 0.96

12.269 0.347 -0.019 362.847 0.96

13.693 0.486 0.006 221.492 0.484

6.696 0.186 -0.115 495.577 0.961

29.495 3.63 0.007 142.928 0.97

49.366 6.035 0.015 110.972 0.972

48.964 7.208 0.013 107.329 0.971

34.71 2.844 0.002 146.545 0.971

37.805 2.239 0.016 155.889 0.972

17.773 0.644 -0.021 272.066 0.96

8.109 0.3 -0.113 401.496 0.965

18.046 2.007 -0.066 186.6 0.965

26.651 2.679 0.012 157.151 0.961

20.503 2.749 -0.028 167.425 0.96

15.404 1.484 0.048 209.306 0.963

6.733 0.152 -0.12 539.703 0.959

9.221 0.179 -0.086 494.978 0.965

13.749 0.594 -0.065 288.763 0.959

6.54 0.129 -0.182 583.633 0.963

65.269 7.173 0.022 98.922 0.969

Table B4: Indentation test of AAFA mix 4

Modulus Hardness Drift Correction Displacement Load

GPa GPa nm/s nm mN

45.581 11.046 -0.286 104.457 0.967

26.999 1.811 -0.35 176.182 0.963

176

28.713 4.548 -0.329 137.094 0.958

78.123 6.879 -0.327 95.358 0.959

23.754 1.331 -0.36 199.447 0.96

14.825 0.918 -0.397 244.674 0.965

73.081 9.023 -0.349 91.131 0.968

29.318 1.619 -0.354 181.119 0.968

44.656 2.281 -0.402 150.657 0.965

53.878 1.627 -0.431 168.876 0.971

34.691 4.063 -0.368 133.262 0.969

70.313 8.491 -0.311 93.313 0.968

90.064 10.165 -0.367 83.861 0.967

47.573 4.774 -0.408 118.039 0.965

35.711 3.326 -0.346 139.143 0.969

20.003 0.892 -0.27 237.563 0.969

58.44 6.528 -0.342 103.886 0.966

18.433 1.497 -0.403 200.718 0.962

17.747 0.473 -0.429 309.541 0.963

23.119 0.593 -0.359 274.817 0.957

41.298 3.206 -0.297 136.317 0.968

4.719 0.111 -0.502 636.555 0.967

25.981 1.641 -0.376 183.702 0.968

70.803 10.652 -0.322 89.13 0.969

9.86 0.142 -0.438 548.271 0.968

51.404 2.151 -0.332 150.877 0.96

30.255 3.913 -0.368 138.792 0.959

82.742 10.401 -0.334 85.594 0.972

22.297 1.963 -0.323 177.995 0.959

25.594 1.378 -0.357 195.405 0.966

31.821 5.766 -0.328 128.413 0.97

27.078 2.2 -0.303 165.394 0.96

52.125 3.231 -0.314 129.965 0.96

39.39 3.544 -0.206 133.881 0.97

23.469 1.213 -0.399 206.771 0.962

12.753 0.876 -0.406 254.797 0.964

29.205 1.532 -0.208 184.686 0.967

34.916 2.453 -0.393 152.331 0.96

13.15 0.522 -0.365 306.083 0.964

7.37 0.205 -0.454 473.039 0.962

17.57 0.623 -0.322 275.855 0.959

54.549 5.452 -0.313 110.602 0.968

36.844 1.975 -0.339 162.92 0.964

6.713 0.312 -0.269 404.474 0.969

32.965 1.647 -0.307 177.129 0.969

114.513 9.979 -0.26 79.584 0.965

12.949 0.429 -0.275 330.54 0.958

15.897 0.863 -0.348 246.693 0.959

22.697 1.238 -0.272 207.083 0.97

177

121.047 6.284 -0.136 91.353 0.968

36.441 1.99 -0.295 162.543 0.962

9.171 0.127 -0.301 579.451 0.967

45.041 4.132 -0.244 123.931 0.961

15.423 2.283 -0.261 190.325 0.97

11.736 0.279 -0.327 400.671 0.964

18.349 1.225 -0.239 214.634 0.967

97.121 13.002 -0.194 78.245 0.972

19.144 0.919 -0.215 235.88 0.967

61.083 8.117 -0.216 97.908 0.967

18.865 0.309 -0.214 372.011 0.963

31.72 1.934 -0.203 168.323 0.969

20.136 1.85 -0.168 185.921 0.968

46.508 2.197 -0.186 151.475 0.96

101.228 11.438 -0.184 79.17 0.965

106.348 7.263 -0.216 88.631 0.965

75.86 7.749 -0.152 93.551 0.97

32.425 2.011 -0.175 165.077 0.964

23.801 0.654 -0.207 264.512 0.968

62.665 4.972 -0.169 109.88 0.964

14.126 0.2 -0.245 461.558 0.969

7.876 0.15 -0.361 539.581 0.964

36.399 2.061 -0.131 160.462 0.961

31.675 1.666 -0.218 177.464 0.97

49.336 4.745 -0.107 117.471 0.969

30.082 1.131 -0.118 206.907 0.97

42.246 2.643 -0.077 144.698 0.971

24.222 1.359 -0.16 198.508 0.971

22.243 0.188 -0.14 465.73 0.96

28.172 1.951 -0.132 171.193 0.969

94.767 10.34 -0.094 82.617 0.97

28.107 1.548 -0.141 185.561 0.972

48.804 4.018 -0.112 123.496 0.971

59.773 3.505 -0.069 124.482 0.972

30.508 3.909 -0.091 139.334 0.97

74.778 10.879 -0.054 87.241 0.967

77.949 8.584 -0.063 90.652 0.97

21.566 1.683 -0.087 187.982 0.964

27.341 1.765 -0.087 176.957 0.96

43.538 2.654 -0.054 143.419 0.966

31.394 1.885 -0.089 170.162 0.97

57.706 8.329 -0.058 99.102 0.966

66.628 4.447 -0.068 112.779 0.97

29.435 0.905 -0.063 225.884 0.961

18.196 0.804 -0.114 249.906 0.968

27.091 0.505 -0.068 293.054 0.965

63.425 5.204 -0.063 108.512 0.971

178

13.715 0.189 -0.112 472.357 0.961

76.808 12.46 -0.024 84.495 0.965

59.308 6.328 -0.056 103.984 0.961

12.45 0.237 -0.081 430.34 0.971

35.395 1.76 -0.01 171.534 0.972

19.386 1.032 -0.021 225.427 0.964

38.186 2.972 -0.009 141.899 0.971

13.883 0.521 -0.066 305.347 0.97

33.037 2.06 -0.007 163.905 0.972

11.459 0.444 -0.07 330.838 0.963

47.48 3.792 0.008 125.7 0.962

19.572 0.687 -0.042 263.65 0.968

31.831 1.563 -0.012 181.436 0.969

78.247 7.592 0.021 93.214 0.967

16.72 0.277 -0.047 394.788 0.97

47.064 4.634 0.008 119.49 0.969

9.527 0.471 -0.087 331.753 0.97

71.448 7.474 0.014 95.945 0.973

14.384 0.46 -0.042 318.711 0.961

36.418 1.379 0.011 187.716 0.973

57.082 3.655 0.008 123.649 0.972

18.709 0.596 -0.017 280.89 0.971

15.98 0.503 -0.029 304.635 0.965

28.319 1.076 0.003 211.488 0.962

60.467 4.824 0.009 112.093 0.972

24.392 0.884 -0.055 232.335 0.963

26.449 1.513 -0.016 188.428 0.969

80.78 8.214 0.026 90.695 0.967

83.929 7.71 0.006 91.189 0.963

24.083 1.959 -0.029 175.255 0.959

16.71 0.677 -0.031 268.524 0.96

12.782 0.437 -0.06 330.325 0.969

27.26 1.695 -0.028 179.679 0.962

35.963 4.124 0.014 131.084 0.963

81.91 12.199 0.007 83.035 0.965

58.744 6.895 0.009 102.796 0.973

25.061 1.624 -0.018 184.888 0.962

19.664 1.166 -0.029 214.984 0.96

18.146 0.657 -0.041 269.351 0.96

13.411 0.326 -0.04 370.609 0.963

17.979 1.172 0.008 218.781 0.969

12.415 0.552 -0.048 300.591 0.959

13.315 0.738 -0.051 268.451 0.965

12.443 0.387 -0.084 347.949 0.968

25.944 3.008 -0.016 153.834 0.962

16.157 1.245 -0.035 218.847 0.97

18.766 0.616 -0.029 275.594 0.96

179

17.463 0.877 -0.024 241.951 0.959

85.062 10.533 0.019 84.537 0.967

17.145 0.905 -0.015 239.965 0.961

19.637 1.126 -0.041 217.632 0.96

12.495 0.423 -0.059 334.179 0.961

45.237 4.805 0.025 119.369 0.966

43.725 6.549 0.009 112.778 0.965

23.723 0.831 -0.038 238.844 0.962

17.277 0.837 -0.046 246.802 0.961

21.792 0.798 0.009 246.066 0.972

14.645 1.647 -0.035 205.98 0.959

20.361 0.869 -0.019 238.735 0.964

16.479 0.788 -0.043 254.649 0.966

27.281 0.605 -0.05 270.688 0.967

63.568 6.592 0.025 101.691 0.97

35.855 4.447 0.009 129.506 0.97

23.448 2.933 -0.006 159.122 0.963

84.396 9.003 0.025 87.862 0.97

19.947 0.857 -0.018 240.545 0.963

12.152 0.361 -0.057 358.973 0.97

24.143 1.454 -0.012 193.598 0.967

23.902 1.733 0.006 182.804 0.965

9.875 0.584 -0.069 303.762 0.958

50.861 6.945 0.018 106.581 0.967

10.588 0.189 -0.125 478.04 0.959

13.491 1.205 -0.032 228.018 0.959

19.308 0.681 -0.033 264.988 0.968

26.401 1.271 -0.025 199.915 0.961

13.4 0.579 -0.051 293.13 0.963

20.821 0.771 -0.027 250.091 0.968

28.375 1.185 -0.032 204.182 0.967

51.905 9.853 0.007 100.286 0.969

70.384 7.753 0.026 95.075 0.965

14.26 0.831 -0.039 254.791 0.965

45.175 4.349 0.02 122.661 0.969

19.115 0.243 -0.053 416.49 0.969

65.244 8.86 0.031 94.657 0.972

96.004 8.802 0.022 85.737 0.97

43.921 6.511 0.012 112.961 0.969

12.414 0.459 -0.03 324.905 0.97

15.447 1.593 -0.012 206.081 0.97

52.086 2.776 0.017 137.887 0.973

48.841 3.891 0.025 124.718 0.972

20.917 0.628 -0.036 272.02 0.969

29.439 2.767 0.02 153.05 0.971

86.907 8.901 0.04 87.593 0.972

37.378 2.24 -0.007 156.106 0.971

180

15.683 0.533 -0.047 297.771 0.963

14.281 0.438 -0.035 325.306 0.961

67.64 8.874 0.015 93.511 0.969

36.847 1.429 0.017 184.247 0.966

103.801 11.17 0.03 78.991 0.963

51.247 5.429 0.015 112.613 0.972

7.45 0.453 -0.077 347.336 0.962

20.431 1.095 -0.018 218.511 0.96

18.703 1.134 -0.015 219.932 0.97

Table B5: Indentation test of AAFA mix 5

Modulus Hardness Drift Correction Displacement Load

GPa GPa nm/s nm mN

14.579 0.606 0.343 284.978 0.962

10.087 0.298 0.326 393.545 0.961

57.892 6.69 0.403 103.9 0.973

85.726 10.808 0.355 83.965 0.969

53.235 5.796 0.341 109.277 0.963

21.758 0.603 0.268 276.129 0.971

55.164 5.143 0.324 112.093 0.97

53.488 2.692 0.27 138.242 0.963

28.426 0.982 0.231 220.097 0.969

23.655 2.956 0.209 158.249 0.961

19.374 0.781 0.153 249.753 0.96

26.336 1.562 0.205 186.765 0.972

26.641 1.145 0.19 208.525 0.969

29.673 1.638 0.182 179.697 0.964

18.698 2.235 0.142 180.12 0.964

21.75 0.303 0.112 373.811 0.966

24.499 1.113 0.149 212.959 0.968

55.394 2.372 0.152 144.667 0.968

31.843 1.868 0.106 170.344 0.97

35.019 2.65 0.118 148.692 0.961

55.06 6.226 0.116 106.877 0.97

96.376 7.757 0.101 88.434 0.963

28.118 1.582 0.064 183.465 0.965

36.894 1.949 0.07 163.892 0.967

20.63 0.799 0.055 245.769 0.959

19.291 0.459 0.03 313.119 0.971

32.609 1.904 0.031 168.573 0.97

51.026 5.312 0 79.236 0.48

17.029 0.55 -0.014 292.433 0.967

124.101 14.06 0.057 71.754 0.966

28.537 2.106 0.03 165.923 0.96

181

38.051 4.401 0.022 127.058 0.96

23.022 1.092 -0.027 216.389 0.971

46.247 3.719 -0.009 127.693 0.971

22.722 0.731 -0.011 252.908 0.964

22.335 0.701 -0.03 258.547 0.97

23.097 1.561 -0.023 190.23 0.965

25.749 1.535 -0.023 188.553 0.971

21.608 2.287 -0.015 172.926 0.969

10.877 0.396 -0.07 347.553 0.959

133.529 27.407 0.003 63.267 0.965

96.278 9.914 -0.029 83.123 0.97

96.669 8.056 -0.029 87.732 0.97

21.831 1.105 -0.068 215.838 0.96

13.447 0.481 -0.094 314.931 0.962

26.536 1.116 -0.068 210.226 0.964

48.607 2.564 -0.022 142.367 0.961

25.234 0.75 -0.015 248.263 0.967

16.735 0.292 -0.019 386.313 0.972

12.21 0.769 -0.097 267.829 0.963

18.249 0.678 -0.045 265.972 0.962

14.866 0.407 -0.068 335.516 0.967

26.572 2.234 0.008 165.743 0.965

95.169 8.407 0.01 86.981 0.97

17.48 0.806 -0.013 249.798 0.96

20.199 1.12 -0.036 217.453 0.963

10.511 0.309 -0.09 387.403 0.968

20.685 0.974 -0.043 228.712 0.968

21.042 0.929 -0.052 232.48 0.97

28.877 1.043 -0.011 214.389 0.968

21.297 1.089 -0.036 218.155 0.964

20.146 0.992 -0.051 226.996 0.961

96.903 9.045 0.014 84.841 0.968

17.291 0.64 -0.079 274.066 0.964

55.237 6.608 0.005 104.857 0.961

34.102 3.577 -0.005 137.52 0.962

19.349 1.365 -0.049 205.459 0.969

14.049 0.405 -0.09 338.042 0.969

19.693 0.836 -0.08 243.021 0.962

62.179 8.965 0.008 95.579 0.967

27.071 0.987 -0.034 220.549 0.968

9.802 0.12 -0.161 590.404 0.959

82.413 7.196 0.001 93.476 0.967

36.772 3.075 0.004 141.558 0.971

128.272 12.673 0.017 72.918 0.966

84.147 7.726 0.003 91.277 0.967

89.721 8.157 0.016 88.7 0.967

31.723 0.986 -0.045 217.404 0.968

182

92.466 11.596 0.001 80.766 0.963

30.042 0.686 -0.022 254.201 0.964

80.891 8.198 -0.008 90.953 0.972

69.504 7.594 -0.004 96.045 0.969

35.97 3.52 -0.037 136.718 0.968

113.712 14.549 -0.009 72.94 0.966

19.228 0.824 -0.056 246.098 0.97

22.232 0.692 -0.021 260.044 0.97

10.323 0.142 -0.14 545.42 0.961

10.655 0.253 -0.075 422.051 0.969

70.514 11.18 -0.003 88.592 0.97

81.125 9.056 -0.014 88.621 0.97

11.972 0.127 -0.067 574.299 0.97

39.357 4.393 -0.03 126.51 0.968

9.21 0.311 -0.159 389.668 0.96

22.847 2.102 -0.062 173.803 0.961

14.272 1.153 -0.098 229.542 0.97

69.102 7.588 -0.014 96.334 0.972

85.129 7.843 -0.035 90.675 0.967

43.109 3.767 -0.031 128.818 0.967

23.258 1.397 -0.04 197.566 0.968

99.29 9.157 -0.024 84.106 0.968

36.136 2.671 -0.074 147.663 0.964

24.082 0.897 -0.087 231.663 0.966

14.481 0.747 -0.078 263.334 0.959

33.206 2.709 -0.044 149.972 0.97

35.525 1.926 -0.034 165.459 0.967

14.642 0.345 -0.105 359.15 0.961

20.386 1.096 -0.086 218.519 0.96

56.05 10.461 -0.029 96.671 0.966

47.904 1.858 -0.035 161.536 0.967

38.756 5.824 -0.025 119.946 0.97

73.822 5.133 -0.014 105.809 0.97

36.089 1.392 -0.054 186.448 0.965

25.266 0.788 -0.019 243.648 0.97

33.472 2.633 -0.015 150.378 0.962

28.088 1.506 -0.053 187.125 0.969

21.017 1.469 -0.019 197.588 0.968

48.868 5.4 -0.019 113.431 0.96

67.904 8.529 -0.011 94.04 0.967

50.159 6.511 -0.018 108.395 0.967

22.992 1.61 -0.047 188.485 0.965

75.223 8.144 -0.01 92.61 0.97

78.143 11.19 0.003 85.752 0.97

25.519 1.262 -0.055 201.405 0.961

36.102 1.976 -0.054 163.814 0.97

21.801 0.323 -0.057 362.891 0.965

183

50.741 6.597 -0.016 108.067 0.972

17.045 1.02 -0.076 230.113 0.959

18.185 0.513 -0.076 298.344 0.961

8.553 0.434 -0.044 345.573 0.962

17.888 0.691 -0.082 264.722 0.963

114.719 9.852 -0.008 79.611 0.96

22.97 1.752 -0.039 183.472 0.964

71.994 12.477 -0.007 86.46 0.967

48.902 3.9 -0.015 124.526 0.971

29.584 0.761 -0.032 242.95 0.961

29.102 1.144 -0.028 206.564 0.969

21.604 1.034 -0.046 222.29 0.967

23.626 1.049 -0.066 218.677 0.968

49.935 2.648 -0.035 140.216 0.961

23.449 2.069 -0.022 174.424 0.97

13.934 0.263 -0.123 407.269 0.967

28.401 5.383 -0.027 134.645 0.964

23.005 1.088 -0.071 215.54 0.96

28.088 1.747 -0.037 177.71 0.97

28.766 1.195 -0.052 203.217 0.968

60.298 7.578 0.002 99.386 0.961

36.234 1.655 -0.053 174.737 0.97

10.987 0.475 -0.078 324.786 0.969

38.86 1.185 -0.035 197.897 0.968

11.273 0.459 -0.093 326.313 0.959

97.549 10.023 -0.007 82.534 0.968

102.87 11.512 -0.004 78.721 0.965

25.156 1.379 -0.032 196.259 0.969

33.795 1.079 -0.053 207.944 0.965

26.62 1.246 -0.052 202.011 0.969

13.439 0.645 -0.106 282.145 0.97

51.108 4.662 -0.003 117.084 0.969

11.382 0.688 -0.109 280.744 0.958

36.308 5.14 -0.007 125.135 0.968

30.023 5.017 -0.017 133.147 0.961

16.504 0.702 -0.055 264.978 0.959

14.3 0.776 -0.076 261.03 0.966

36.539 2.966 -0.02 142.543 0.962

40.176 2.313 -0.005 152.699 0.971

43.412 6.516 -0.007 113.456 0.971

98.169 10 -0.002 82.45 0.968

61.083 6.805 -0.017 101.853 0.969

36.106 1.644 -0.026 174.925 0.966

15.886 0.808 -0.081 252.665 0.96

14.672 0.541 -0.078 298.612 0.967

23.726 1.183 -0.042 208.762 0.966

21.302 1.088 -0.06 217.79 0.96

184

16.088 0.618 -0.079 280.01 0.963

59.069 4.544 -0.004 114.489 0.969

65.589 7.872 0.004 96.846 0.972

30.809 2.822 -0.003 149.785 0.96

37.805 4.733 -0.043 125.925 0.971

14.358 0.53 -0.093 300.855 0.962

11.893 0.667 -0.079 282.039 0.959

31.221 1.986 0.006 166.991 0.966

14.308 0.812 -0.067 256.161 0.961

8.103 0.176 -0.158 502.425 0.966

62.979 6.881 -0.002 100.384 0.961

14.577 0.189 -0.145 470.901 0.958

21.091 1.467 -0.036 197.818 0.971

29.932 1.631 -0.026 180.488 0.972

31.009 0.687 -0.013 253.139 0.961

31.469 1.943 -0.034 168.373 0.97

61.176 6.529 -0.012 102.964 0.971

46.364 3.695 -0.005 127.453 0.964

92.446 9.399 0.017 84.821 0.965

59.17 4.254 -0.003 116.955 0.972

98.681 10.133 0.005 82.088 0.968

32.904 3.135 0 143.806 0.965

38.726 1.295 -0.009 189.998 0.96

23.29 0.836 -0.054 238.532 0.961

38.694 3.283 -0.006 136.672 0.961

26.36 1.437 -0.029 192.335 0.972

28.801 2.239 -0.008 162.661 0.962

Table B6: Indentation test of AAFA mix 6

Modulus Hardness Drift Correction Displacement Load

GPa GPa nm/s nm mN

105.555 9.816 0.262 81.19 0.961

85.377 7.989 0.273 90.489 0.974

17.516 0.758 0.232 255.864 0.961

81.418 11.417 0.311 84.284 0.968

40.845 3.073 0.305 138.576 0.97

44.011 3.754 0.316 128.738 0.972

54.077 5.572 0.348 110.296 0.97

22.625 1.121 0.292 214.729 0.97

50.659 7.791 0.335 104.682 0.97

18.08 1.665 0.293 196.309 0.97

34.047 2.696 0.307 149.553 0.972

16.095 0.569 0.248 289.352 0.963

52.151 6.112 0.316 109.038 0.972

185

11.978 0.675 0.26 282.461 0.972

7.944 0.211 0.158 466.38 0.971

15.349 0.712 0.254 267.009 0.967

43.046 2.086 0.273 156.592 0.967

72.251 8.571 0.299 92.56 0.971

17.105 0.69 0.215 266.429 0.965

24.414 0.602 0.229 272.993 0.964

40.984 4.956 0.252 121.761 0.969

89.759 9.439 0.281 85.392 0.966

6.429 0.367 0.116 383.636 0.968

15.608 0.834 0.198 251.052 0.965

28.428 3.26 0.243 147.512 0.964

102.181 8.596 0.246 84.879 0.964

73.174 8.056 0.252 93.563 0.971

7.431 0.181 0.1 498.972 0.963

17.412 0.561 0.19 289.319 0.967

26.023 1.353 0.215 196.925 0.972

11.963 0.545 0.108 304.397 0.966

28.988 1.642 0.179 179.974 0.962

20.199 1.022 0.179 224.553 0.961

11.872 0.414 0.106 338.538 0.961

10.662 0.237 0.108 432.709 0.964

52.869 6.957 0.197 105.66 0.973

18.468 1.053 0.137 224.854 0.96

15.842 0.332 0.124 364.155 0.964

56.3 6.035 0.175 107.231 0.973

29.534 2.263 0.171 161.362 0.962

21.254 1.033 0.134 222.782 0.966

15.586 1.404 0.164 211.778 0.961

36.102 4.775 0.151 127.357 0.971

84.025 7.855 0.171 90.855 0.966

21.769 0.76 0.096 249.527 0.961

26.392 1.701 0.114 181.171 0.97

77.953 10.949 0.149 86.164 0.97

78.831 6.682 0.164 96.413 0.968

181.27 11.682 0.15 69.691 0.966

14.429 0.592 0.055 288.357 0.965

18.507 1.352 0.086 207.924 0.97

21.24 1.21 0.094 211.111 0.973

15.943 0.339 0.082 360.287 0.962

17.891 0.452 0.068 316.385 0.967

16.694 1.585 0.115 202.819 0.972

18.277 0.889 0.086 239.723 0.963

59.145 7.672 0.126 100.293 0.973

23.833 0.933 0.083 229.029 0.971

21.971 0.703 0.069 257.758 0.964

23.678 0.834 0.073 239.199 0.968

186

21.031 0.991 0.066 226.948 0.97

12.929 0.512 -0.019 309.283 0.967

19.064 0.94 0.056 234.587 0.971

23.771 1.354 0.082 198.781 0.966

21.429 0.979 0.05 227.367 0.969

13.88 0.374 -0.034 348.83 0.962

97.875 8.751 0.115 85.593 0.972

45.629 7.263 0.083 109.49 0.968

30.795 1.736 0.067 175.14 0.965

47.018 4.629 0.077 119.725 0.972

80.225 9.056 0.095 88.872 0.97

22.964 0.969 0.049 225.79 0.964

26.138 0.872 0.051 231.966 0.962

25.409 1.43 0.046 192.964 0.965

14.923 0.686 0.01 271.605 0.966

26.976 0.856 0.022 234.236 0.971

25.569 1.051 0.043 216.183 0.965

15.44 0.869 0.042 247.467 0.962

13.575 0.717 0.008 270.597 0.966

32.361 1.097 0.033 206.952 0.961

14.112 0.54 0.018 300.359 0.969

93.249 10.113 0.078 83.277 0.968

43.244 2.089 0.081 156.865 0.973

17.766 1.414 0.056 206.826 0.972

39.92 2.587 0.079 147.209 0.972

18.185 0.919 0.032 237.359 0.965

87.933 7.883 0.073 89.743 0.963

12.122 0.211 -0.029 454.041 0.969

37.476 3.999 0.056 130.875 0.966

21.044 0.933 0.019 231.412 0.963

15.78 0.532 0.021 298.294 0.966

84.003 9.638 0.066 86.602 0.97

40.784 4.546 0.056 124.362 0.969

69.713 7.453 0.06 96.08 0.963

24.818 3.243 0.052 153.392 0.965

19.847 1.416 0.053 201.359 0.962

15.712 0.462 -0.048 315.128 0.959

17.92 0.731 0 259.09 0.964

16.491 0.681 0.008 268.191 0.959

17.514 0.81 0.096 250.552 0.97

22.426 1.105 0.08 216.187 0.97

32.078 3.531 0.094 140.257 0.964

36.262 3.296 0.093 138.44 0.961

25.19 2.126 0.08 169.734 0.962

29.976 2.005 0.059 167.155 0.961

24.828 2.228 0.072 168.715 0.97

67.804 8.198 0.077 95.075 0.971

187

43.991 4.671 0.076 120.707 0.961

97.399 9.679 0.078 83.265 0.967

27.545 1.007 0.06 218.828 0.971

18.19 0.215 -0.037 442.547 0.971

37.047 4.222 0.041 129.703 0.968

32.124 4.056 0.06 135.651 0.962

20.445 0.567 0.003 283.645 0.964

14.396 0.933 0.007 244.593 0.966

16.241 0.856 -0.01 246.566 0.959

27.479 1.126 0.009 208.295 0.96

18.871 0.785 -0.02 250.764 0.966

86.427 9.775 0.055 85.557 0.967

19.355 0.99 0.004 228.733 0.963

66.341 10.544 0.066 91.082 0.967

63.017 8.255 0.072 96.696 0.967

15.934 0.55 -0.003 294.166 0.967

18.201 0.885 0 239.915 0.959

36.292 5.023 0.049 125.274 0.962

58.614 8.447 0.061 98.371 0.966

51.687 3.096 0.052 132.189 0.963

33.591 3.172 0.035 142.269 0.96

16.132 0.638 -0.008 276.243 0.962

81.443 8.019 0.045 91.184 0.969

27.309 1.38 0.003 194.173 0.972

12.184 0.379 -0.029 351.988 0.97

48.881 1.75 0.033 165.395 0.971

59.681 8.161 0.047 98.603 0.969

85.804 10.765 0.051 84.063 0.97

11.007 0.79 -0.005 270.086 0.961

30.86 1.599 0.026 179.835 0.96

20.342 1.047 -0.018 222.123 0.959

18.099 1.164 0.009 219.143 0.97

16.311 0.409 -0.008 331.922 0.964

18.17 0.394 -0.04 335.441 0.968

14.953 0.839 -0.015 252.113 0.965

123.694 12.015 0.037 74.414 0.963

32.95 1.087 0.018 207.309 0.96

238.395 21.722 0.029 55.453 0.963

80.081 10.955 0.048 85.567 0.973

20.005 1.613 0.028 193.088 0.962

77.969 7.712 0.04 92.719 0.963

68.571 9.797 0.047 91.276 0.967

5.428 0.094 -0.084 681.235 0.968

69.005 7.935 0.05 95.371 0.972

59.691 5.369 0.032 108.995 0.972

14.523 0.698 -0.017 270.852 0.966

50.334 3.301 0.031 130.661 0.973

188

25.963 1.597 0.004 185.48 0.969

34.533 2.198 0.014 158.689 0.966

55.062 6.872 0.017 104.055 0.961

27.184 1.907 0.001 173.516 0.968

20.086 0.587 -0.004 280.858 0.969

33.662 2.913 0.033 145.623 0.96

22.421 1.791 0.016 183.193 0.965

24.583 1.29 0.006 200.628 0.96

27.026 1.274 -0.003 200.22 0.972

16.804 0.972 -0.036 234.713 0.961

23.681 1.155 -0.006 210.019 0.96

31.103 0.844 0.003 231.83 0.963

73.229 8.255 0.031 92.942 0.969

4.692 0.289 0.026 438.551 0.972

19.138 1.372 0.039 204.91 0.963

32.707 5.083 0.024 129.303 0.964

18.44 0.964 -0.012 233.387 0.97

47.695 6.366 0.029 110.426 0.965

24.755 1.246 -0.012 203.131 0.961

17.357 1.116 -0.028 222.73 0.961

91.537 6.035 0.028 96.933 0.973

94.679 8.969 0.037 85.583 0.97

51.28 5.71 0.032 111.234 0.972

13.592 0.444 -0.057 325.211 0.963

15.997 0.497 -0.019 306.439 0.966

21.856 1.223 -0.011 208.947 0.969

48.954 2.858 0.026 137.649 0.971

58.675 6.645 0.032 103.557 0.97

9.703 0.208 -0.007 461.978 0.968

33.399 2.289 0.041 157.911 0.971

17.86 0.933 -0.008 236.033 0.961

59.077 5.66 0.031 107.578 0.97

20.74 1.026 -0.028 223.197 0.959

8.726 0.264 -0.074 418.974 0.96

98.447 7.685 0.043 88.479 0.967

10.405 0.196 -0.095 469.513 0.957

12.547 0.263 -0.078 409.72 0.964

22.024 0.915 -0.009 232.822 0.972

13.469 0.728 0.004 268.521 0.96

20.295 0.802 -0.004 247.144 0.969

9.985 0.455 -0.043 333.414 0.965

16.535 0.75 -0.026 258.945 0.963

25.855 3.585 0.017 148.211 0.961

17.937 0.675 -0.024 267.304 0.964

12.619 0.667 -0.051 279.937 0.962

189

Table B7: Indentation test of AAFA mix 7

Modulus Hardness Drift Correction Displacement Load

GPa GPa nm/s nm mN

23.339 1.363 0.041 199.534 0.972

13.466 0.356 -0.052 356.684 0.961

68.245 10.411 0.048 90.461 0.967

17.314 0.87 -0.042 243.406 0.962

16.26 0.646 -0.064 275.262 0.966

31.659 4.191 0.001 135.131 0.96

10.203 0.488 -0.05 323.111 0.962

79.47 8.419 0.026 90.505 0.967

16.133 0.712 -0.043 265.868 0.971

20.831 1.154 -0.008 215.192 0.972

16.833 0.828 -0.04 248.761 0.962

49.675 5.141 0.022 114.876 0.969

36.088 5.472 0.014 123.524 0.962

12.479 0.577 -0.07 297.379 0.971

11.987 0.395 -0.104 345.633 0.965

7.637 0.426 -0.05 354.149 0.966

17.55 0.698 -0.054 264.746 0.966

16.491 0.61 -0.034 281.105 0.966

51.726 4.987 0.018 114.672 0.969

17.439 1.077 -0.051 226.439 0.971

66.858 7.989 0.022 95.969 0.97

11.115 0.454 -0.06 330.123 0.97

15.082 0.65 -0.058 276.952 0.966

17.085 0.72 -0.049 261.809 0.962

115.678 10.744 0.024 77.756 0.963

80.133 7.089 0.025 94.505 0.968

17.697 0.531 -0.074 295.268 0.965

16.308 0.9 -0.045 242.558 0.963

21.305 0.748 -0.031 251.859 0.963

18.54 0.588 -0.019 281.094 0.959

16.54 1.146 -0.029 222.968 0.964

30.584 2.831 -0.008 150.214 0.964

40.104 5.434 0.001 119.746 0.962

67.964 9.934 0.025 91.431 0.97

46.986 3.908 0.018 125.531 0.972

29.161 0.626 -0.027 264.651 0.961

38.512 3.545 0.008 133.887 0.961

107.601 11.42 0.028 77.883 0.963

15.191 0.793 -0.056 256.522 0.964

20.262 0.881 -0.042 237.142 0.96

14.078 0.715 -0.052 269.052 0.963

73.392 9.025 0.031 90.955 0.967

18.04 0.869 -0.035 243.113 0.97

190

16.849 0.788 -0.053 253.292 0.961

14.12 0.647 -0.051 279.219 0.963

13.173 0.611 -0.068 288.142 0.965

22.132 1.599 -0.021 190.55 0.969

75.509 9.188 0.023 89.911 0.967

21.507 1.035 -0.022 221.912 0.964

22.151 0.654 -0.053 265.127 0.962

9.465 0.117 0 598.35 0.963

42.163 3.195 0.071 136.191 0.971

24.601 0.676 -0.006 259.035 0.96

14.583 0.636 -0.046 279.876 0.963

15.453 0.683 -0.044 271.455 0.97

16.045 0.764 -0.04 258.157 0.964

63.08 3.937 0.048 118.524 0.97

12.398 0.692 -0.055 278.214 0.97

61.963 6.208 0.012 104.014 0.973

22.115 0.867 -0.023 236.4 0.961

74.116 10.166 0.023 88.679 0.97

16.027 0.832 -0.055 249.984 0.962

28.037 0.965 0.024 222.302 0.972

17.812 0.491 -0.038 305.01 0.965

16.213 0.664 -0.048 272.034 0.964

17.786 0.791 -0.048 250.84 0.959

19.017 0.921 -0.034 235.071 0.96

15.673 0.739 -0.052 262.084 0.963

12.745 0.663 -0.045 280.25 0.962

43.255 3.779 0.016 128.178 0.96

65.052 7.43 0.021 98.313 0.972

6.296 0.338 -0.123 396.316 0.969

14.694 0.691 -0.046 270.851 0.963

81.251 11.757 0.029 83.884 0.968

16.024 0.753 -0.055 259.544 0.964

58.129 9.122 0.026 97.346 0.967

20.826 0.773 -0.008 249.345 0.964

56.645 3.914 0.035 121.098 0.973

44.758 5.231 0.019 117.466 0.969

24.692 0.886 -0.004 231.621 0.961

78.219 13.682 0.034 82.977 0.968

16.037 0.406 -0.088 333.061 0.962

19.06 0.838 -0.039 243.485 0.959

38.09 7.143 0.018 116.851 0.968

30.269 2.259 0.029 161.452 0.971

9.24 0.25 -0.038 426.619 0.96

58.586 7.252 0.013 101.536 0.969

14.252 0.339 -0.044 362.301 0.958

39.215 2.049 0.01 160.023 0.972

24.949 1.656 0.021 184.624 0.97

191

10.669 0.215 -0.121 450.004 0.957

12.656 0.633 -0.054 285.81 0.965

6.504 0.196 -0.065 487.996 0.966

15.819 0.818 -0.043 252.536 0.967

16.325 0.653 -0.051 274.168 0.968

14.876 0.835 -0.032 251.867 0.959

18.907 0.7 -0.04 261.348 0.959

12.796 0.351 -0.046 360.354 0.962

14.271 1.007 0.02 238.701 0.964

70.532 14.388 -0.006 85.556 0.965

45.148 3.176 0.001 134.081 0.962

24.349 0.574 -0.055 278.497 0.963

23.342 1.145 -0.058 211.115 0.96

18.765 0.86 -0.06 241.527 0.959

69.832 8.318 -0.001 94.078 0.972

28.485 2.03 -0.011 168.45 0.966

17.543 0.822 -0.054 247.992 0.961

39.872 3.44 0.007 134.019 0.961

17.186 0.635 -0.056 274.894 0.962

26.746 0.73 -0.03 250.078 0.968

23.244 1.578 -0.04 189.966 0.971

11.157 0.53 -0.079 311.127 0.97

18.804 0.767 -0.055 252.636 0.962

82.892 6.222 0.007 97.128 0.961

114.49 11.891 0.017 76.245 0.97

36.797 1.17 -0.025 199.182 0.961

12.413 0.423 -0.08 335.512 0.969

26.21 1.275 -0.035 200.696 0.969

15.527 0.742 -0.02 262.255 0.965

32.229 1.551 -0.012 181.795 0.97

39.776 1.189 -0.036 197.091 0.967

45.752 1.334 -0.012 185.593 0.966

37.885 3.469 0.002 135.917 0.971

36.956 1.098 0.007 205.075 0.967

29.114 4.284 -0.006 138.091 0.961

26.154 0.889 -0.016 231.359 0.972

67.431 7.051 -0.005 98.459 0.968

25.743 1.655 -0.014 183.781 0.972

17.446 0.789 -0.045 252.383 0.964

16.755 0.68 -0.053 269.335 0.97

28.624 0.584 -0.064 274.118 0.967

12.921 0.964 -0.056 246.71 0.965

23.943 1.516 -0.017 191.642 0.972

84.873 8.247 0.002 89.554 0.967

32.802 0.677 -0.043 254.934 0.969

67.128 9.055 0.012 93.473 0.972

18.648 0.79 -0.044 249.9 0.961

192

72.54 9.509 0.011 90.399 0.97

27.184 1.505 -0.003 187.4 0.962

105.707 11.759 0.019 77.697 0.963

37.827 1.508 -0.011 180.156 0.969

116.946 11.177 0.015 76.763 0.963

77.226 9.439 0.022 88.958 0.97

25.032 1.016 -0.024 220.072 0.969

93.498 8.437 0.013 87.108 0.967

32.2 2.197 -0.003 161.032 0.971

39.489 4.394 0.017 125.881 0.96

17.989 0.737 -0.058 258.014 0.963

48.283 5.58 0.035 113.047 0.963

19.842 0.963 -0.047 229.951 0.96

17.159 0.698 -0.058 265.456 0.966

11.314 0.484 -0.099 319.666 0.959

41.523 7.239 0.013 112.636 0.96

16.245 0.481 -0.064 309.193 0.961

44.35 6.18 0.014 113.879 0.972

19.229 0.716 -0.052 258.619 0.959

17.02 0.773 -0.046 254.827 0.962

20.347 1.044 -0.058 222.393 0.959

23.301 1.283 -0.027 203.414 0.968

28.825 2.694 -0.004 154.262 0.963

29.028 0.733 -0.023 247.247 0.961

14.62 1.102 -0.029 230.633 0.961

42.471 5.135 0.005 119.39 0.965

53 6.633 0.022 106.526 0.971

24.874 0.908 -0.032 229.264 0.961

61.385 7.873 0.028 98.493 0.969

19.705 0.798 -0.056 247.237 0.96

17.74 0.589 -0.059 282.698 0.964

29.863 1.386 -0.022 190.587 0.962

22.442 1.148 -0.037 213.212 0.971

22.974 1.293 -0.038 203.355 0.968

26.657 0.842 -0.05 234.983 0.962

50.777 7.812 0.022 104.041 0.961

29.393 1.982 -0.013 169.023 0.969

42.764 1.622 0.007 172.43 0.966

29.316 1.162 -0.033 204.249 0.961

17.629 0.849 -0.039 244.57 0.959

72.477 13.427 0.027 85.45 0.967

93.379 9.069 0.027 85.495 0.967

63.297 3.753 -0.017 119.8 0.961

66.74 8.148 0.029 95.67 0.972

15.705 0.762 -0.074 259.44 0.966

32.146 1.164 0.007 202.472 0.964

17.171 0.787 -0.038 254.105 0.971

193

18.199 0.787 -0.061 251.62 0.966

21.676 0.752 -0.032 250.884 0.962

20.846 0.792 -0.045 246.649 0.962

21.654 1.224 -0.038 209.431 0.971

16.987 0.836 -0.039 248.14 0.967

57.357 6.652 0.014 104.136 0.97

78.703 9.017 0.003 89.4 0.97

74.115 10.154 0.013 88.701 0.97

26.636 2.368 -0.012 162.444 0.96

14.63 0.601 -0.072 287.232 0.971

17.436 0.692 -0.035 265.008 0.96

17.488 0.848 -0.045 245.231 0.962

19.882 1.023 -0.054 225.964 0.97

15.38 0.703 -0.068 268.699 0.97

75.48 8.407 0.009 91.705 0.967

Table B8: Indentation test of AAFA mix 8

Modulus Hardness Drift Correction Displacement Load

GPa GPa nm/s nm mN

2.604 0.023 -0.871 1326.701 0.945

2.756 0.057 -0.409 876.466 0.949

6.352 0.137 -0.196 569.45 0.965

26.454 1.398 -0.001 193.682 0.967

16.865 1.076 -0.015 226.385 0.959

89.099 7.778 0.046 89.807 0.963

24.599 2.133 -0.015 170.23 0.96

37.871 1.88 0.043 165.898 0.972

103.24 15.213 0.048 74.439 0.966

17.424 0.659 -0.028 270.587 0.963

26.779 1.593 -0.012 185.037 0.972

74.091 8.301 0.042 92.428 0.967

18.63 0.832 -0.036 245.092 0.962

69.244 7.317 0.049 96.676 0.963

15.035 0.708 -0.042 268.795 0.971

73.703 6.096 0.046 100.479 0.97

18.179 1.039 -0.027 226.429 0.96

11.979 0.733 -0.052 273.123 0.963

7.071 0.652 -0.159 312.616 0.959

2.274 0.024 -1.32 1306.035 0.943

4.139 0.105 -0.285 657.908 0.965

10.223 0.651 -0.085 293.062 0.971

23.223 1.268 -0.014 204.786 0.972

21.527 0.751 -0.028 250.993 0.96

79.429 9.391 0.02 88.266 0.967

194

16.373 0.707 -0.052 265.406 0.966

91.29 15.333 0.032 77.278 0.963

63.479 13.8 0.022 89.802 0.971

24.474 1.98 0.005 175.081 0.97

13.191 0.166 -0.064 502.014 0.962

61.669 8.464 0.041 97.098 0.972

24.476 1.261 -0.038 203.528 0.97

85.647 10.219 0.025 85.156 0.973

16.594 0.839 -0.059 248.133 0.963

53.855 3.433 0.028 127.105 0.967

14.667 0.654 -0.052 277.123 0.966

15.086 0.292 -0.043 386.026 0.962

14.924 0.657 -0.05 276.106 0.966

26.892 3.013 0.002 152.459 0.964

34.7 4.843 -0.001 128.042 0.964

16.72 1.161 0 221.157 0.96

19.841 1.763 0.017 189.009 0.967

26.469 0.762 -0.05 244.746 0.959

30.545 0.913 -0.016 225.418 0.97

14.596 0.408 -0.072 336.091 0.97

85.858 8.638 0.016 88.27 0.967

5.165 0.029 -0.723 1175.529 0.96

12.649 0.569 -0.063 296.456 0.959

48.64 6.626 0.022 108.648 0.961

25.201 0.93 -0.033 226.66 0.96

24.834 1.172 0.008 208.643 0.97

111.63 11.291 0.028 77.826 0.973

24.337 1.691 -0.033 183.277 0.961

14.241 0.702 -0.048 271.227 0.968

37.714 1.546 0.006 178.627 0.971

19.209 0.843 -0.003 244.142 0.971

34.46 1.441 0.02 185.002 0.966

13.494 0.463 -0.059 319.301 0.959

9.687 0.261 -0.083 418.504 0.966

28.825 3.223 0.016 147.47 0.965

19.257 0.947 -0.053 233.414 0.969

15.73 0.905 -0.041 243.38 0.963

82.581 9.417 0.036 87.314 0.967

51.106 5.589 0.033 111.908 0.972

37.125 2.308 0.01 153.866 0.961

39.431 2.658 0.01 145.298 0.96

82.425 7.445 0.023 92.746 0.97

84.813 7.729 0.041 91.225 0.969

15.898 0.711 -0.032 265.772 0.966

11.28 0.438 -0.077 332.79 0.96

12.837 1.265 -0.041 228.167 0.964

4.561 0.037 0.031 1051.81 0.971

195

56.65 3.518 0.034 124.634 0.961

84.451 10.084 0.041 85.497 0.968

18.626 0.55 -0.038 289.503 0.964

14.269 0.567 -0.023 294.183 0.969

49.84 7.073 0.046 107.089 0.971

12.711 0.709 -0.033 273.681 0.962

5.589 0.033 -0.266 1109.641 0.955

14.787 0.598 -0.023 286.049 0.963

15.99 0.916 0.012 241.259 0.959

63.778 2.062 0.028 105.937 0.483

52.121 5.626 0.068 110.938 0.968

14.2 0.674 -0.02 274.787 0.963

27.502 1.396 0.055 192.851 0.969

12.394 0.622 -0.021 288.483 0.965

16.715 0.661 0.018 271.61 0.964

23.76 0.457 -0.006 309.755 0.971

4.778 0.251 -0.081 457.112 0.961

17.469 0.984 0.019 232.586 0.963

17.857 1.064 0 225.326 0.961

13.018 0.509 -0.004 310.498 0.971

16.881 0.796 0.009 252.595 0.964

34.077 3.416 0.072 139.367 0.965

5.752 0.086 -0.135 703.76 0.965

35.205 4.845 0.057 128.023 0.972

16.903 0.799 0.011 252.569 0.966

14.397 0.576 -0.003 291.645 0.966

33.463 1.372 0.014 189.808 0.973

12.463 0.422 -0.038 333.992 0.959

17.651 0.816 0.029 248.87 0.964

13.238 0.618 -0.043 286.141 0.962

20.58 2.555 0.031 169.842 0.96

19.577 0.617 -0.015 274.503 0.961

32.847 1.96 0.039 165.855 0.96

9.969 1.004 0.023 258.121 0.969

14.873 0.647 -0.031 277.248 0.962

74.408 8.374 0.07 92.19 0.968

22.606 1.11 0.017 215.475 0.97

103.923 16.877 0.07 72.905 0.961

14.131 0.68 -0.064 275.039 0.97

12.346 0.713 -0.043 275.507 0.97

19.86 0.802 -0.028 247.068 0.964

16.043 0.827 -0.018 251.181 0.967

11.175 0.454 -0.008 329.702 0.967

19.223 1.186 0.008 214.497 0.96

12.913 0.756 -0.017 267.77 0.968

69.478 8.943 0.06 92.535 0.967

9.739 0.333 -0.066 376.6 0.959

196

12.303 1.334 -0.019 228.063 0.97

52.037 6.981 0.043 105.915 0.97

15.579 1.059 -0.022 231.93 0.969

55.247 3.887 0.028 121.652 0.969

40.627 6.712 0.023 115.059 0.965

14.453 0.684 -0.044 273.759 0.972

5.755 0.296 -0.146 419.667 0.963

19.361 0.87 0.001 240.232 0.965

98.481 11.488 0.048 79.559 0.963

16.993 0.851 -0.022 245.544 0.959

34.052 2.061 0.008 162.093 0.96

15.233 0.362 -0.046 350.654 0.961

21.662 1.384 -0.002 199.777 0.962

13.72 0.67 -0.052 276.485 0.962

12.475 0.507 -0.053 310.87 0.961

60.787 7.587 0.037 99.668 0.971

12.601 0.478 -0.035 317.642 0.961

28.457 0.766 0.012 244.234 0.971

25.952 1.039 -0.014 216.423 0.962

15.677 1.112 0 226.772 0.959

82.171 12.044 0.04 83.359 0.97

69.979 10.43 0.059 89.71 0.967

44.78 8.732 0.049 107.01 0.961

36.906 2.176 0.039 158.104 0.972

15.297 0.842 -0.014 250.814 0.964

66.575 6.947 0.041 99.261 0.97

10.15 0.327 -0.03 378.58 0.963

20.354 0.918 0.041 233.335 0.96

101.598 14.55 0.239 75.483 0.968

77.807 6.98 -0.024 95.575 0.969

71.466 8.847 0.055 92.015 0.967

31.153 5.592 0.052 129.934 0.97

50.068 6.165 0.046 109.875 0.97

15.895 0.597 -0.029 284.123 0.963

30.848 3.751 0.037 140.323 0.972

20.56 1.235 0.032 209.769 0.965

88.049 9.103 0.057 86.609 0.967

30.722 1.276 0.042 196.818 0.969

79.4 7.756 0.068 92.083 0.961

14.207 0.567 -0.015 294.587 0.97

18.319 0.876 -0.023 241.125 0.964

89.129 11.561 0.069 81.884 0.968

54.254 9.388 0.053 98.932 0.961

24.439 0.718 0 253.999 0.97

11.681 0.575 -0.04 299.179 0.964

12.186 0.47 -0.041 321.638 0.965

25.836 0.819 0.018 238.496 0.964

197

52.156 5.346 0.064 112.57 0.972

79.816 8.288 0.073 90.758 0.967

97.352 9.56 0.136 83.603 0.969

48.244 4.493 0.103 119.955 0.971

67.995 8.592 0.083 93.876 0.967

16.674 0.581 -0.021 286.312 0.966

113.953 12.962 0.082 74.472 0.96

21.22 1.009 0.058 224.3 0.962

76.426 9.768 0.056 88.466 0.967

12.398 0.543 -0.037 302.971 0.962

22.862 1.406 0.056 197.253 0.964

37.925 1.134 0.037 201.711 0.965

12.588 1.376 0.012 158.049 0.482

12.6 0.635 -0.051 284.975 0.961

58.504 7.027 0.076 102.334 0.97

11.176 0.619 -0.024 293.833 0.969

97.688 8.465 0.072 86.395 0.971

12.681 0.548 -0.019 302.334 0.969

28.284 2.997 0.051 150.524 0.961

11.1 0.444 -0.012 333.198 0.97

26.711 1.681 0.039 181.151 0.965

35.328 2.615 0 148.961 0.96

23.845 0.879 0.007 233.336 0.962

9.537 0.425 -0.06 343.65 0.964

26.842 0.961 -0.01 222.198 0.96

14.695 0.807 -0.029 255.961 0.963

21.12 1.183 0.012 211.613 0.961

13.909 0.673 -0.076 194.352 0.482

67.108 4.325 0.006 79.541 0.48

65.696 9.293 0.037 93.374 0.967

75.214 7.043 0.006 95.99 0.968

14.248 0.628 -0.064 282.649 0.967

0.685 0.005 -0.392 2844.819 0.961

Table B9: Indentation test of AAFA mix 9

Modulus Hardness Drift Correction Displacement Load

GPa GPa nm/s nm mN

17.648 0.57 -0.047 286.114 0.96

20.956 1.056 -0.044 221.753 0.97

42.156 3.119 0.02 136.718 0.964

29.446 3.293 0.014 145.56 0.961

46.601 6.578 0.011 110.785 0.971

13.15 0.382 -0.088 346.625 0.96

17.955 0.893 -0.04 239.531 0.959

198

19.887 1.019 -0.031 225.466 0.963

39.806 3.596 0.006 133.109 0.971

84.331 9.248 0.026 87.259 0.968

22.026 1.308 -0.029 203.158 0.962

37.528 1.471 0.006 181.762 0.966

65.714 7.887 0.01 96.742 0.972

75.647 5.455 0.025 103.487 0.973

19.996 1.001 -0.036 226.896 0.964

17.728 0.898 -0.053 239.463 0.959

29.402 1.806 -0.01 173.838 0.963

21.2 0.849 -0.027 239.217 0.959

53.518 3.666 0.024 124.877 0.972

23.913 1.129 -0.014 212.358 0.968

17.793 0.559 -0.022 289.682 0.97

28.66 1.062 -0.013 213.436 0.972

28.551 1.251 -0.025 199.143 0.961

26.932 1.126 -0.028 209.867 0.971

58.538 9.788 0.016 96.057 0.966

40.294 1.915 -0.023 163.153 0.969

90.852 14.454 0.028 78.276 0.968

6.71 0.117 -0.256 610.431 0.968

61.285 7.455 0.019 99.748 0.97

16.944 0.836 -0.044 248.712 0.97

26.673 0.945 -0.026 223.904 0.961

8.891 0.036 -0.728 1054.105 0.95

22.457 2.692 0.027 163.96 0.961

14.719 0.977 -0.062 239.354 0.961

15.922 0.758 -0.066 260.109 0.971

26.246 1.174 -0.024 207.179 0.971

20.566 0.694 -0.039 260.462 0.962

12.823 0.42 -0.101 334.851 0.965

42.963 2.621 0.001 144.671 0.97

80.8 13.263 0.026 82.409 0.967

16.766 0.794 -0.038 253.311 0.966

36.002 2.514 -0.001 151.197 0.971

13.479 0.708 -0.048 271.493 0.962

26.271 1.704 -0.017 181.264 0.97

16.311 0.543 -0.054 294.194 0.961

16.211 0.545 -0.045 294.981 0.968

16.88 0.7 -0.037 264.732 0.96

27.021 1.387 -0.021 193.033 0.962

20.55 0.941 -0.035 230.859 0.96

68.348 7.537 0.026 96.746 0.972

24.389 2.014 0.023 173.608 0.962

40.906 0.955 -0.03 215.317 0.961

16.144 0.797 -0.045 254.067 0.965

17.912 0.906 -0.051 239.133 0.966

199

26.976 2.604 0.005 158.441 0.968

19.21 1.194 -0.029 215.134 0.97

18.249 0.868 -0.041 242.161 0.964

33.483 5.676 0.037 126.22 0.967

19.329 0.942 -0.047 233.091 0.964

21.667 0.906 -0.04 232.768 0.96

17.433 0.913 -0.04 238.482 0.959

20.168 0.99 -0.035 227.065 0.96

17.553 0.886 -0.043 240.872 0.959

34.649 2.085 0.003 161.628 0.968

18.013 1.007 -0.04 229.754 0.964

15.729 0.724 -0.068 264.251 0.965

22.781 1.393 -0.015 197.674 0.961

21.672 1.38 -0.013 199.898 0.961

11.155 0.456 -0.057 329.281 0.969

22.944 0.757 -0.024 249.858 0.97

22.111 1.305 -0.036 203.068 0.961

33.411 3.849 0.005 136.537 0.973

24.002 0.72 -0.037 253.068 0.962

19.309 0.865 -0.045 240.597 0.963

89.628 10.684 0.027 83.085 0.967

13.112 0.93 -0.037 248.407 0.962

13.93 0.466 -0.061 317.87 0.962

16.073 0.588 -0.064 285.904 0.966

7.525 0.07 -0.146 771.243 0.969

25.741 3.089 -0.002 152.983 0.959

20.362 0.888 -0.007 236.433 0.961

28.065 3.781 0.009 143.115 0.962

143.541 13.554 0.039 69.79 0.963

17.923 0.933 -0.047 236.058 0.962

15.506 0.525 -0.055 299.267 0.959

16.287 1.247 -0.011 218.078 0.967

15.184 0.716 -0.053 267.24 0.97

18.443 0.851 -0.038 243.635 0.964

18.835 0.954 -0.033 232.754 0.964

27.607 1.428 0.002 191.119 0.968

18.948 0.837 -0.057 244.701 0.967

52.906 8.584 0.035 101.647 0.971

21.576 1.251 -0.023 207.973 0.971

17.268 0.977 -0.044 233.715 0.964

31.322 1.847 -0.003 170.92 0.964

16.055 0.992 -0.028 235.977 0.971

26.763 0.976 -0.041 222.206 0.971

48.662 1.88 0.015 160.931 0.972

19.026 0.907 -0.03 236.349 0.959

17.686 0.964 -0.031 234.237 0.966

57.09 1.3 0.057 184.989 0.97

200

34.167 1.449 0.032 185.077 0.969

44.707 6.65 0.038 111.391 0.96

82.219 10.139 0.04 86.014 0.967

31.959 6.39 0.012 126.241 0.964

17.794 0.751 -0.036 256.361 0.962

18.123 0.857 -0.037 244.199 0.97

13.904 0.585 -0.039 291.991 0.972

74.684 9.264 0.03 90.136 0.97

13.064 0.591 -0.062 291.396 0.961

20.871 0.659 -0.013 265.593 0.961

19.332 1.135 -0.012 218.784 0.971

26.674 0.583 -0.019 274.972 0.964

54.924 5.218 0.031 111.921 0.972

51.741 4.168 0.028 120.557 0.969

10.286 0.429 -0.077 338.93 0.96

105.4 11.871 0.052 77.705 0.966

33.057 5.768 0.026 126.664 0.969

55.671 6.95 0.036 104.14 0.973

29.858 4.793 0.021 134.356 0.961

18.528 0.742 -0.028 256.98 0.966

91.605 10.26 0.034 83.364 0.968

16.097 1.204 -0.003 220.875 0.966

196.967 14.625 0.039 64.36 0.972

12.307 1.003 -0.029 245.927 0.964

20.69 0.35 -0.047 351.357 0.97

53.807 6.331 0.025 107.262 0.972

16.735 1.846 0.02 193.826 0.962

43.188 5.371 0.019 118.086 0.972

25.141 0.871 -0.016 233.08 0.963

76.295 12.222 0.035 85.159 0.97

30.397 1.453 -0.013 187.661 0.97

28.849 1.128 -0.006 207.499 0.965

26.32 4.899 0.026 140.308 0.966

41.85 9.217 0.025 109.867 0.971

69.5 5.517 0.03 104.679 0.97

76.379 9.685 0.036 88.759 0.97

47.133 5.489 0.02 114.751 0.972

36.948 2.448 0.026 151.011 0.961

77.601 11.108 0.024 86.045 0.97

19.206 1.012 -0.025 227.899 0.97

19.042 0.896 -0.029 237.915 0.964

18.562 0.882 -0.034 240.095 0.964

66.397 4.294 0.015 114.12 0.969

55.887 4.035 0.021 119.561 0.962

16.168 1.204 -0.048 221.125 0.97

13.794 0.827 -0.028 255.795 0.96

21.094 1.07 -0.035 219.327 0.96

201

20.16 0.91 -0.038 234.232 0.959

10.63 0.229 -0.024 439.763 0.966

16.705 0.767 -0.035 255.895 0.96

39.576 2.28 0.029 153.001 0.961

97.637 12.504 0.033 78.662 0.971

18.017 0.896 0.026 239.61 0.964

21.798 1.105 -0.025 217.052 0.971

40.88 4.646 0.003 123.797 0.971

30.742 0.949 -0.016 221.583 0.969

34.363 1.662 -0.012 175.656 0.97

44.575 2.703 0.019 142.346 0.971

60.003 4.328 0.033 115.468 0.962

18.637 0.978 -0.041 231.091 0.964

33.539 2.625 0.025 150.574 0.964

22.44 1.403 -0.017 197.879 0.963

24.868 1.289 -0.013 200.511 0.961

35.886 3.836 0.014 134.005 0.97

5.335 0.075 -0.174 753.471 0.961

19.621 1.033 -0.035 224.405 0.96

55.249 4.258 0.026 118.34 0.97

17.525 0.559 -0.018 288.472 0.96

17.545 0.986 -0.024 232.251 0.963

59.129 9.409 0.031 96.457 0.969

31.041 1.751 -0.002 174.378 0.965

19.559 0.995 -0.028 228.754 0.97

49.454 3.669 0.026 126.655 0.972

35.843 4.844 0.031 127.248 0.971

25.944 2.573 0.028 160.444 0.968

15.952 0.677 -0.042 270.433 0.964

20.022 0.955 -0.02 230.451 0.961

20.374 0.885 -0.002 236.535 0.96

21.303 0.872 -0.014 236.766 0.96

31.539 3.994 0.023 136.819 0.962

23.069 0.619 -0.031 271.415 0.968

57.987 8.356 0.034 99.163 0.972

60.215 5.575 0.02 107.704 0.972

26.605 1.672 0.007 181.974 0.969

32.72 2.619 0 152.045 0.972

100.943 10.151 0.048 81.6 0.968

28.892 2.225 0.032 163.637 0.971

19.538 1.061 -0.028 222.406 0.96

24.361 1.381 -0.009 197.06 0.969

29.878 2.54 0.017 155.949 0.967

61.818 6.565 0.042 102.498 0.97

26.042 1.018 -0.008 218.901 0.969

7.875 0.06 -0.209 825.304 0.964

133.226 12.14 0.044 72.994 0.963

202

88.916 13.686 0.046 79.502 0.968

15.226 0.718 -0.032 265.82 0.963

24.186 3.809 0.036 149.575 0.96

77.749 7.224 0.051 94.521 0.966

203

References

Alonso, C. and Fernandez, L. (2004), ‘Dehydration and rehydration processes of cement paste

exposed to high temperature environments’, Journal of materials science 39(9), 3015–3024.

Analla, M. (1998), ‘Model validation through the linear regression fit to actual versus predicted

values’, Agricultural Systems 57(1), 115–119.

Anstis, G., Chantikul, P., Lawn, B. R. and Marshall, D. (1981), ‘A critical evaluation of inden-

tation techniques for measuring fracture toughness: I, direct crack measurements’, Journal

of the American Ceramic Society 64(9), 533–538.

ANSYS R○ (2012), ‘Ansys mechanical apdl theory reference’, Documentation for Ansys .

ANSYS R○ (2015), ‘Release 16.2’, Documentation for Ansys .

ASTM Standard C109 (2011), Standard test method for compressive strength of hydraulic

cement mortars (using 2-in. or [50-mm] cube specimens), Standard, ASTM International

West Conshohocken, PA.

ASTM Standard C140 (2015), Standard test methods for sampling and testing concrete masonry

units and related units, Standard, ASTM International West Conshohocken, PA.

ASTM Standard C150 (2015), Standard specification of portland cement, Standard, ASTM

International West Conshohocken, PA.

ASTM Standard C1609 (2012), Standar test method for flexural performance of fibre-reinforced

concrete (using beam with third-point loading), Standard, ASTM International West Con-

shohocken, PA.

ASTM Standard C618 (2012), Specification for coal fly ash and raw or calcined natural pozzolan

for use in concrete, Standard, ASTM International West Conshohocken, PA.

ASTM Standard C642 (2013), Standard test method for density, absorption, and voids in

hardened concrete, Standard, ASTM International West Conshohocken, PA.

Australia Standard (2009), Supplementary cementitious materials for use with portland and

blended cement, Standard, Standard Australia.

Australia Standard (2010), General purpose and blended cements, Standard, Standard Aus-

tralia.

Aydın, S. (2008), ‘Development of a high-temperature-resistant mortar by using slag and

pumice’, Fire safety journal 43(8), 610–617.

Berry, E., Hemmings, R. and Cornelius, B. (1990), ‘Mechanisms of hydration reactions in high

volume fly ash pastes and mortars’, Cement and Concrete Composites 12(4), 253–261.

204

Bhattacharya, A. and Nix, W. (1991), ‘Finite element analysis of cone indentation’, Interna-

tional Journal of Solids and Structures 27(8), 1047–1058.

Bobko, C. and Ulm, F.-J. (2008), ‘The nano-mechanical morphology of shale’, Mechanics of

Materials 40(4), 318–337.

Budiansky, B. and Cui, Y. L. (1994), ‘On the tensile strength of a fiberreinforced ceramic com-

posite containing a crack-like flaw’, Journal of the Mechanics and Physics of Solids 42(1), 1–

19.

Bye, G. C. (1999), Portland cement: composition, production and properties, Thomas Telford.

Caijun, S. and Della, R. (2006), ‘Alkali-activated cements and concretes’, London and New

York: Taylor & Francis .

Cao, Y.-P., Ji, X.-Y. and Feng, X.-Q. (2010), ‘Geometry independence of the normalized relax-

ation functions of viscoelastic materials in indentation’, Philosophical Magazine 90(12), 1639–

1655.

Cao, Z. and Zhang, X. (2008), ‘Nanoindentation stress–strain curves of plasma-enhanced chem-

ical vapor deposited silicon oxide thin films’, Thin Solid Films 516(8), 1941–1951.

Chan, S. Y., Peng, G.-f. and Chan, J. K. (1996), ‘Comparison between high strength con-

crete and normal strength concrete subjected to high temperature’, Materials and Structures

29(10), 616–619.

Chen, J. and Bull, S. (2007), ‘Indentation fracture and toughness assessment for thin optical

coatings on glass’, Journal of Physics D: Applied Physics 40(18), 5401.

Chen, X., Beyerlein, I. J. and Brinson, L. C. (2009a), ‘Curved-fiber pull-out model for nanocom-

posites. part 1: Bonded stage formulation’, Mechanics of Materials 41(3), 279–292.

Chen, X., Beyerlein, I. J. and Brinson, L. C. (2009b), ‘Curved-fiber pull-out model for nanocom-

posites. part 2: Interfacial debonding and sliding’, Mechanics of Materials 41(3), 293–307.

Chen, X., Meawad, A. and Struble, L. J. (2014), ‘Method to stop geopolymer reaction’, Journal

of the American Ceramic Society 97(10), 3270–3275.

Cheng, Y.-T. and Cheng, C.-M. (2004), ‘Scaling, dimensional analysis, and indentation mea-

surements’, Materials Science and Engineering: R: Reports 44(4), 91–149.

Chindaprasirt, P., Jaturapitakkul, C. and Sinsiri, T. (2005), ‘Effect of fly ash fineness on com-

pressive strength and pore size of blended cement paste’, Cement and Concrete Composites

27(4), 425–428.

Constantinides, G., Chandran, K. R., Ulm, F.-J. and Van Vliet, K. (2006), ‘Grid indentation

analysis of composite microstructure and mechanics: principles and validation’, Materials

Science and Engineering: A 430(1), 189–202.

Constantinides, G. and Ulm, F.-J. (2004), ‘The effect of two types of csh on the elasticity

of cement-based materials: Results from nanoindentation and micromechanical modeling’,

Cement and concrete research 34(1), 67–80.

Constantinides, G. and Ulm, F.-J. (2007), ‘The nanogranular nature of csh’, Journal of the

Mechanics and Physics of Solids 55(1), 64–90.

205

Criado, M., Fernandez-Jimenez, A., Palomo, A., Sobrados, I. and Sanz, J. (2008), ‘Effect of

the sio 2/na 2 o ratio on the alkali activation of fly ash. part ii: 29 si mas-nmr survey’,

Microporous and Mesoporous Materials 109(1), 525–534.

Criado, M., Palomo, A. and Fernandez-Jimenez, A. (2005), ‘Alkali activation of fly ashes. part

1: Effect of curing conditions on the carbonation of the reaction products’, Fuel 84(16), 2048–

2054.

Davidovits, J. (1994a), ‘High-alkali cements for 21st century concretes’, ACI Special Publication

144.

Davidovits, J. (1994b), Properties of geopolymer cements, in ‘First international conference on

alkaline cements and concretes’, Vol. 1, pp. 131–149.

de Normalisation, C. E. (2005), ‘Eurocode 4: Design of composite steel and concrete structures.

part 1-2: General rules-structural fire design’, CEN ENV .

de Normalisation, C. E. (2015), ‘Eurocode 2: Design of concrete structures - part 1-1: General

rules and rules for building’, CEN ENV .

Demirel, B. and Kelestemur, O. (2010), ‘Effect of elevated temperature on the mechanical

properties of concrete produced with finely ground pumice and silica fume’, Fire Safety

Journal 45(6), 385–391.

Diamond, S. (2000), ‘Mercury porosimetry: an inappropriate method for the measure-

ment of pore size distributions in cement-based materials’, Cement and concrete research

30(10), 1517–1525.

Dormieux, L., Kondo, D. and Ulm, F.-J. (2006), Microporomechanics, John Wiley & Sons.

Dukino, R. D. and Swain, M. V. (1992), ‘Comparative measurement of indentation fracture

toughness with berkovich and vickers indenters’, Journal of the American Ceramic Society

75(12), 3299–3304.

Duxson, P., Fernandez-Jimenez, A., Provis, J. L., Lukey, G. C., Palomo, A. and Van Deventer,

J. (2007), ‘Geopolymer technology: the current state of the art’, Journal of Materials Science

42(9), 2917–2933.

Duxson, P., Provis, J. L., Lukey, G. C. and Van Deventer, J. S. (2007), ‘The role of inorganic

polymer technology in the development of green concrete’, Cement and Concrete Research

37(12), 1590–1597.

Fernandez-Jimenez, A., De La Torre, A., Palomo, A., Lopez-Olmo, G., Alonso, M. and Aranda,

M. (2006), ‘Quantitative determination of phases in the alkaline activation of fly ash. part ii:

Degree of reaction’, Fuel 85(14), 1960–1969.

Fernandez-Jimenez, A. and Palomo, A. (2003), ‘Characterisation of fly ashes. potential reactiv-

ity as alkaline cements’, Fuel 82(18), 2259–2265.

Fernandez-Jimenez, A. and Palomo, A. (2005), ‘Composition and microstructure of alkali acti-

vated fly ash binder: effect of the activator’, Cement and Concrete Research 35(10), 1984–

1992.

Ferry, J. D. (1980), Viscoelastic properties of polymers, John Wiley & Sons.

206

Field, J., Swain, M. and Dukino, R. (2003), ‘Determination of fracture toughness from the

extra penetration produced by indentation-induced pop-in’, Journal of materials research

18(06), 1412–1419.

Findley, W. N. and Davis, F. A. (2013), Creep and relaxation of nonlinear viscoelastic materials,

Courier Corporation.

Fischer-Cripps, A. (1999), ‘The hertzian contact surface’, Journal of materials science

34(1), 129–137.

Fischer-Cripps, A. (2004), ‘A simple phenomenological approach to nanoindentation creep’,

Materials Science and Engineering: A 385(1), 74–82.

Fischer-Cripps, A. C. (2011), Nanoindentation, Springer.

Fischer-Cripps, A. C. and Mustafaev, I. (2000), Introduction to contact mechanics, Springer.

Flatt, R. J., Roussel, N. and Cheeseman, C. R. (2012), ‘Concrete: An eco material that needs

to be improved’, Journal of the European Ceramic Society 32(11), 2787–2798.

French, D. and Smitham, J. (2007), Fly ash characteristics and feed coal properties, QCAT

Technology Transfer Centre.

Gaillard, Y., Tromas, C. and Woirgard, J. (2003), ‘Pop-in phenomenon in mgo and lif: obser-

vation of dislocation structures’, Philosophical magazine letters 83(9), 553–561.

Garcia-Lodeiro, I., Palomo, A., Fernandez-Jimenez, A. and Macphee, D. (2011), ‘Compatibility

studies between nash and cash gels. study in the ternary diagram na 2 o–cao–al 2 o 3–sio 2–h

2 o’, Cement and Concrete Research 41(9), 923–931.

Glukhovsky, V., Rostovskaja, G. and Rumyna, G. (1980), High strength slag-alkaline cements,

in ‘Proceedings of the 7th international congress on the chemistry of cement, Paris’.

Guide to the use of fly ash in concrete in Australia (2009), Report, Ash Development Association

of Australia.

Handoo, S., Agarwal, S. and Agarwal, S. (2002), ‘Physicochemical, mineralogical, and morpho-

logical characteristics of concrete exposed to elevated temperatures’, Cement and Concrete

Research 32(7), 1009–1018.

Hardijito, D. and Rangan, B. (2005), Development and properties of low-calcuim fly ash-based

geopolymer concrete, Report, Faculty of Engineering.

Hardjito, D., Wallah, S. E., Sumajouw, D. M. and Rangan, B. V. (2004), ‘On the development

of fly ash-based geopolymer concrete’, ACI materials journal 101(6).

Heidrich, C. and Woodhead, A. (2005), Prepared for the benchmarking module: Sustainablility

capcity building program, Report, Pavement Recycling and Stabilisation Association.

Helmuth, R. (1983), ‘Some questions concerning astm standards and methods of testing fly ash

for use with portland cement’, Cement, concrete and aggregates 5(2), 103–110.

Herbert, E., Pharr, G., Oliver, W., Lucas, B. and Hay, J. (2001), ‘On the measurement of

stressstrain curves by spherical indentation’, Thin solid films 398, 331–335.

207

Herrera-Franco, P. and Drzal, L. (1992), ‘Comparison of methods for the measurement of

fibre/matrix adhesion in composites’, Composites 23(1), 2–27.

Hsueh, C.-H. (1990), ‘Interfacial debonding and fiber pull-out stresses of fiber-reinforced com-

posites’, Materials Science and Engineering: A 123(1), 1–11.

ISO 14577-1 (2002), Metallic materials-instrumented indentation test for hardness and materials

parameters-part 1: test method., Standard, International Organization for Standardization

Geneva, Switzerland.

Jennings, H. M. (2000), ‘A model for the microstructure of calcium silicate hydrate in cement

paste’, Cement and Concrete Research 30(1), 101–116.

Jennings, H. M., Thomas, J. J., Gevrenov, J. S., Constantinides, G. and Ulm, F.-J. (2007),

‘A multi-technique investigation of the nanoporosity of cement paste’, Cement and Concrete

Research 37(3), 329–336.

Joseph, D. (2011), ‘Geopolymer chemistry & applications’.

Kanda, T., Lin, Z. and Li, V. C. (2000), ‘Tensile stress-strain modeling of pseudostrain hard-

ening cementitious composites’, Journal of Materials in Civil Engineering 12(2), 147–156.

Kanit, T., Forest, S., Galliet, I., Mounoury, V. and Jeulin, D. (2003), ‘Determination of the

size of the representative volume element for random composites: statistical and numerical

approach’, International Journal of solids and structures 40(13), 3647–3679.

Kim, D. J., Park, S. H., Ryu, G. S. and Koh, K. T. (2011), ‘Comparative flexural behav-

ior of hybrid ultra high performance fiber reinforced concrete with different macro fibers’,

Construction and Building Materials 25(11), 4144–4155.

Kim, J.-K. and Yi, S.-T. (2002), ‘Application of size effect to compressive strength of concrete

members’, Sadhana 27(4), 467–484.

Kong, D. L. and Sanjayan, J. G. (2008), ‘Damage behavior of geopolymer composites exposed

to elevated temperatures’, Cement and Concrete Composites 30(10), 986–991.

Kong, D. L. and Sanjayan, J. G. (2010), ‘Effect of elevated temperatures on geopolymer paste,

mortar and concrete’, Cement and Concrete Research 40(2), 334–339.

Kullaa, J. (1996), ‘Dimensional analysis of bond modulus in fiber pullout’, Journal of Structural

Engineering 122(7), 783–787.

Kumar, S., Kumar, R., Alex, T., Bandopadhyay, A. and Mehrotra, S. (2005), Effect of mechan-

ically activated fly ash on the properties of geopolymer cement, in ‘Proceedings of the 4th

World Congress on Geopolymer’, pp. 113–116.

Labuz, J. F. and Zang, A. (2015), Mohr–coulomb failure criterion, in ‘The ISRM Suggested

Methods for Rock Characterization, Testing and Monitoring: 2007-2014’, Springer, pp. 227–

231.

Laugier, M. (1987), ‘Palmqvist indentation toughness in wc-co composites’, Journal of materials

science letters 6(8), 897–900.

208

Lawn, B. R., Evans, A. and Marshall, D. (1980), ‘Elastic/plastic indentation damage in ceram-

ics: the median/radial crack system’, Journal of the American Ceramic Society 63(910), 574–

581.

Li, V. C. (1997), ‘Engineered cementitious composites (ecc)–tailored composites through micro-

mechanical modelling’, Fiber Reinforced Concrete: Present and the Future, Canadian Society

for Civil Engineering pp. 64–97.

Li, V. C. and Leung, C. K. (1992), ‘Steady-state and multiple cracking of short random fiber

composites’, Journal of Engineering Mechanics 118(11), 2246–2264.

Li, V. C., Mishra, D. K. and Wu, H.-C. (1995), ‘Matrix design for pseudo-strain-hardening fibre

reinforced cementitious composites’, Materials and Structures 28(10), 586–595.

Li, V. C., Wang, Y. and Backer, S. (1990), ‘Effect of inclining angle, bundling and surface

treatment on synthetic fibre pull-out from a cement matrix’, Composites 21(2), 132–140.

Lichinchi, M., Lenardi, C., Haupt, J. and Vitali, R. (1998), ‘Simulation of berkovich nanoinden-

tation experiments on thin films using finite element method’, Thin solid films 312(1), 240–

248.

Ling, F. F. (2011), Nanoindentation, third edition edn, Springer, New York.

Litvan, G. G. (1986), Mechanism of cement paste degradation due to chemical and physical

processes, National Research Council Canada, Institute for Research in Construction.

Lloyd, R. R., Provis, J. L. and Van Deventer, J. S. (2010), ‘Pore solution composition and alkali

diffusion in inorganic polymer cement’, Cement and Concrete Research 40(9), 1386–1392.

Marques, S. P. and Creus, G. J. (2012), Rheological Models: Integral and Differential Represen-

tations, Springer, pp. 11–21.

Martinez, E., Romero, J., Lousa, A. and Esteve, J. (2003), ‘Nanoindentation stressstrain curves

as a method for thin-film complete mechanical characterization: application to nanometric

crn/cr multilayer coatings’, Applied Physics A 77(3-4), 419–427.

MATLAB (2014), R2014a, The MathWorks Inc., Natick, Massachusetts.

Mendes, A., Sanjayan, J. and Collins, F. (2008), ‘Phase transformations and mechanical

strength of opc/slag pastes submitted to high temperatures’, Materials and structures

41(2), 345–350.

Miller, M., Bobko, C., Vandamme, M. and Ulm, F.-J. (2008), ‘Surface roughness criteria for

cement paste nanoindentation’, Cement and Concrete Research 38(4), 467–476.

Morsy, M., Rashad, A. and Shebl, S. (2008), ‘Effect of elevated temperature on compressive

strength of blended cement mortar’, Build Res J 56(2-3), 173–185.

Morsy, M., Shebl, S. and Rashad, A. (2009), ‘Effect of fire on microstructure and mechan-

ical properties of blended cement pastes containing metakaolin and silica fume’, Silicates

Industriels 74(3), 59.

Naaman, A. E. (1972), A statistical theory of strength for fiber reinforced concrete, PhD thesis,

Massachusetts Institute of Technology.

209

Naaman, A. E., Moavenzadeh, F. and McGarry, F. J. (1974), ‘Probabilistic analysis of fiber-

reinforced concrete’, Journal of the Engineering Mechanics Division 100(2), 397–413.

Naaman, A. E., Namur, G. G., Alwan, J. M. and Najm, H. S. (1991), ‘Fiber pullout and bond

slip. i: Analytical study’, Journal of Structural Engineering 117(9), 2769–2790.

Naaman, A. E. and Reinhardt, H. (2008), ‘High performance fiber reinforced cement com-

posites’, Naaman AE. High-performance construction materials: science and applications.

Singapore: World Scientific Publishing pp. 91–153.

Nashif, A. D., Jones, D. I. and Henderson, J. P. (1985), Vibration damping, John Wiley & Sons.

Nemecek, J., Smilauer, V. and Kopecky, L. (2011), ‘Nanoindentation characteristics of alkali-

activated aluminosilicate materials’, Cement and Concrete Composites 33(2), 163–170.

Neville, A. M. (2011), Properties of concrete.

Oliver, W. C. and Pharr, G. M. (1992), ‘An improved technique for determining hardness and

elastic modulus using load and displacement sensing indentation experiments’, Journal of

materials research 7(06), 1564–1583.

Oliver, W. C. and Pharr, G. M. (2004), ‘Measurement of hardness and elastic modulus by instru-

mented indentation: Advances in understanding and refinements to methodology’, Journal

of materials research 19(01), 3–20.

Ozbay, E., Oztas, A., Baykasoglu, A. and Ozbebek, H. (2009), ‘Investigating mix proportions of

high strength self compacting concrete by using taguchi method’, Construction and Building

Materials 23(2), 694–702.

Pacheco-Torgal, F., Castro-Gomes, J. and Jalali, S. (2008), ‘Alkali-activated binders: A review:

Part 1. historical background, terminology, reaction mechanisms and hydration products’,

Construction and Building Materials 22(7), 1305–1314.

Palomo, A., Grutzeck, M. and Blanco, M. (1999), ‘Alkali-activated fly ashes: a cement for the

future’, Cement and concrete research 29(8), 1323–1329.

Papadakis, V. G. (1999), ‘Effect of fly ash on portland cement systems: Part i. low-calcium fly

ash’, Cement and Concrete Research 29(11), 1727–1736.

Papadakis, V. G. (2000), ‘Effect of fly ash on portland cement systems: Part ii. high-calcium

fly ash’, Cement and Concrete Research 30(10), 1647–1654.

Park, G.-J. (2007), ‘Design of experiments’, Analytic Methods for Design Practice pp. 309–391.

Peiwei, G., Min, D. and Naiqian, F. (2001), ‘The influence of superplasticizer and superfine

mineral powder on the flexibility, strength and durability of hpc’, Cement and Concrete

Research 31(5), 703–706.

Pichler, C. and Lackner, R. (2008), ‘A multiscale creep model as basis for simulation of earlyage

concrete behavior’, Computers and Concrete 5(4), 295–328.

Pipilikaki, P. and Beazi-Katsioti, M. (2009), ‘The assessment of porosity and pore size distribu-

tion of limestone portland cement pastes’, Construction and Building Materials 23(5), 1966–

1970.

210

Pipkin, A. C. (2012), Lectures on viscoelasticity theory, Vol. 7, Springer Science & Business

Media.

Poon, C., Lam, L. and Wong, Y. (2000), ‘A study on high strength concrete prepared with

large volumes of low calcium fly ash’, Cement and Concrete Research 30(3), 447–455.

Poon, C.-S., Azhar, S., Anson, M. and Wong, Y.-L. (2001), ‘Comparison of the strength and

durability performance of normal-and high-strength pozzolanic concretes at elevated temper-

atures’, Cement and Concrete Research 31(9), 1291–1300.

Powers, T. C. (1958), ‘Structure and physical properties of hardened portland cement paste’,

Journal of the American Ceramic Society 41(1), 1–6.

Provis, J. L. and Van Deventer, J. S. J. (2009), Geopolymers: structures, processing, properties

and industrial applications, Elsevier.

Ramezanianpour, A. A. (2014), Cement replacement materials, Springer.

Rangan, B. (2008), Fly ash-based geopolymer concrete, Report, Faculty of Engineering.

Ranjit, K. (1990), ‘A primer on the taguchi method, competitive manufacturing series’.

Ross, P. J. (1996), ‘Taguchi techniques for quality engineering: loss function, orthogonal exper-

iments, parameter and tolerance design’.

Roy, D. M. (1999), ‘Alkali-activated cements opportunities and challenges’, Cement and Con-

crete Research 29(2), 249–254.

Roy, R. K. (2010), A primer on the Taguchi method, Society of Manufacturing Engineers.

Sakharova, N., Fernandes, J., Antunes, J. and Oliveira, M. (2009), ‘Comparison between

berkovich, vickers and conical indentation tests: A three-dimensional numerical simulation

study’, International Journal of Solids and Structures 46(5), 1095–1104.

Scrivener, K. L. (2004), ‘Backscattered electron imaging of cementitious microstructures: un-

derstanding and quantification’, Cement and Concrete Composites 26(8), 935–945.

Scrivener, K. L. and Nonat, A. (2011), ‘Hydration of cementitious materials, present and future’,

Cement and concrete research 41(7), 651–665.

Seber, G. A. and Lee, A. J. (2012), Linear regression analysis, Vol. 936, John Wiley & Sons.

Shah, S. P. (1989), Fracture toughness of cement-based materials, in ‘Fracture of Concrete and

Rock’, Springer, pp. 1–12.

Shi, C. (1992), Activation of natural pozzolans, fly ashes and blast furnace slag, Civil Engineer-

ing, University of Calgary.

Shi, C. (1996), ‘Strength, pore structure and permeability of alkali-activated slag mortars’,

Cement and Concrete Research 26(12), 1789–1799.

Shi, C. and Day, R. L. (1993a), ‘Acceleration of strength gain of lime-pozzolan cements by

thermal activation’, Cement and Concrete Research 23(4), 824–832.

Shi, C. and Day, R. L. (1993b), ‘Chemical activation of blended cements made with lime and

natural pozzolans’, Cement and Concrete Research 23(6), 1389–1396.

211

Shi, C. and Day, R. L. (2000), ‘Pozzolanic reaction in the presence of chemical activators: Part

iireaction products and mechanism’, Cement and Concrete Research 30(4), 607–613.

Shi, C. and Fernandez-Jimenez, A. (2006), ‘Stabilization/solidification of hazardous and ra-

dioactive wastes with alkali-activated cements’, Journal of hazardous materials 137(3), 1656–

1663.

Shi, C., Jimenez, A. F. and Palomo, A. (2011), ‘New cements for the 21st century: the pursuit

of an alternative to portland cement’, Cement and Concrete Research 41(7), 750–763.

Siddique, R. (2004), ‘Performance characteristics of high-volume class f fly ash concrete’, Ce-

ment and Concrete Research 34(3), 487–493.

Sneddon, I. N. (1965), ‘The relation between load and penetration in the axisymmetric boussi-

nesq problem for a punch of arbitrary profile’, International Journal of Engineering Science

3(1), 47–57.

Soranakom, C. and Mobasher, B. (2008), ‘Correlation of tensile and flexural responses of

strain softening and strain hardening cement composites’, Cement and concrete Composites

30(6), 465–477.

Srinivasan, C., Narasimhan, N. L. and Ilango, S. (2003), ‘Development of rapid-set high-strength

cement using statistical experimental design’, Cement and Concrete Research 33(9), 1287–

1292.

Stauss, S., Schwaller, P., Bucaille, J.-L., Rabe, R., Rohr, L., Michler, J. and Blank, E. (2003),

‘Determining the stress–strain behaviour of small devices by nanoindentation in combination

with inverse methods’, Microelectronic Engineering 67, 818–825.

Stein, E. M. (1956), ‘Interpolation of linear operators’, Transactions of the American Mathe-

matical Society pp. 482–492.

Strategies, E. (2007), ‘Review of co2-e emissions from concrete versus timber sleepers’.

Stutzman, P. E. (1996), Guide for X-ray powder diffraction analysis of Portland cement and

clinker, US Department of Commerce, Technology Administration, National Institute of Stan-

dards and Technology, Office of Applied Economics, Building and Fire Research Laboratory.

Stutzman, P. E. (2001), ‘Scanning electron microscopy in concrete petrography’, National In-

stitute of Standards and Technology 2.

Taha, M. R., Soliman, E., Sheyka, M., Reinhardt, A. and Al-Haik, M. (2010), ‘Fracture tough-

ness of hydrated cement paste using nanoindentation’.

Tanyildizi, H. (2013), ‘Variance analysis of crack characteristics of structural lightweight con-

crete containing silica fume exposed to high temperature’, Construction and Building Mate-

rials 47, 1154–1159.

Taylor, H. F. (1997), Cement chemistry, Thomas Telford.

Tennis, P. D. and Jennings, H. M. (2000), ‘A model for two types of calcium silicate hydrate in

the microstructure of portland cement pastes’, Cement and Concrete Research 30(6), 855–

863.

212

Thomas, J. J., Jennings, H. M. and Allen, A. J. (1998), ‘The surface area of cement paste as

measured by neutron scattering: evidence for two csh morphologies’, Cement and Concrete

Research 28(6), 897–905.

Turner, L. K. and Collins, F. G. (2013), ‘Carbon dioxide equivalent (co 2-e) emissions: a com-

parison between geopolymer and opc cement concrete’, Construction and Building Materials

43, 125–130.

Ulm, F.-J., Vandamme, M., Bobko, C., Alberto Ortega, J., Tai, K. and Ortiz, C. (2007),

‘Statistical indentation techniques for hydrated nanocomposites: concrete, bone, and shale’,

Journal of the American Ceramic Society 90(9), 2677–2692.

U.S. Environmental Protection agency (2007), ‘Waste-resource conservation-industraul material

recycling’.

Van den Heede, P. and De Belie, N. (2012), ‘Environmental impact and life cycle assessment

(lca) of traditional and greenconcretes: literature review and theoretical calculations’, Cement

and Concrete Composites 34(4), 431–442.

Vandamme, M., Tweedie, C. A., Constantinides, G., Ulm, F.-J. and Van Vliet, K. J. (2012),

‘Quantifying plasticity-independent creep compliance and relaxation of viscoelastoplastic ma-

terials under contact loading’, Journal of Materials Research 27(01), 302–312.

Vandamme, M. and Ulm, F.-J. (2006), ‘Viscoelastic solutions for conical indentation’, Interna-

tional Journal of solids and structures 43(10), 3142–3165.

Vandamme, M. and Ulm, F.-J. (2009), ‘Nanogranular origin of concrete creep’, Proceedings of

the National Academy of Sciences 106(26), 10552–10557.

Vandamme, M. and Ulm, F.-J. (2013), ‘Nanoindentation investigation of creep properties of

calcium silicate hydrates’, Cement and Concrete Research 52, 38–52.

Velez, K., Maximilien, S., Damidot, D., Fantozzi, G. and Sorrentino, F. (2001), ‘Determination

by nanoindentation of elastic modulus and hardness of pure constituents of portland cement

clinker’, Cement and Concrete Research 31(4), 555–561.

Wallah, S. and Rangan, B. V. (2006), ‘Low-calcium fly ash-based geopolymer concrete: long-

term properties’, Res. Report-GC2, Curtin University, Australia. pp pp. 76–80.

Ward, R. and Li, V. C. (1991), ‘Dependence of flexural behaviour of fibre reinforced mortar on

material fracture resistance and beam size’, Construction and Building Materials 5(3), 151–

161.

Warner, R. F., Rangan, B., Hall, A. and Faulkes, K. (1998), Concrete structures.

What is Lime? (2005), Report, Pavement Recycling and Stabilisation Association.

Xu, H. and Van Deventer, J. (2000), ‘The geopolymerisation of alumino-silicate minerals’,

International Journal of Mineral Processing 59(3), 247–266.

Xu, Y., Jiang, L., Xu, J. and Li, Y. (2012), ‘Mechanical properties of expanded polystyrene

lightweight aggregate concrete and brick’, Construction and Building Materials 27(1), 32–38.

Xu, Y., Wong, Y., Poon, C. and Anson, M. (2001), ‘Impact of high temperature on pfa concrete’,

Cement and Concrete Research 31(7), 1065–1073.

213

Zaoui, A. (2002), ‘Continuum micromechanics: survey’, Journal of Engineering Mechanics

128(8), 808–816.

Zhan, Y. and Meschke, G. (2014), ‘Analytical model for the pullout behavior of straight and

hooked-end steel fibers’, Journal of Engineering Mechanics 140(12), 04014091.

Zhang, J. and Scherer, G. W. (2011), ‘Comparison of methods for arresting hydration of cement’,

Cement and Concrete Research 41(10), 1024–1036.

214